Materials

and Processes
for Spacecraft and High
Reliability Applications

Barrie D. Dunn
Springer Praxis Books

Astronautical Engineering
More information about this series at http://www.springer.com/series/5495
Barrie D. Dunn

Materials and Processes

for Spacecraft and High Reliability
Applications

123
Barrie D. Dunn
School of Engineering
University of Portsmouth
Portsmouth
UK

Published in association with Praxis Publishing, Chichester, UK

ISSN 2365-9599 ISSN 2365-9602 (electronic)

Springer Praxis Books
ISBN 978-3-319-23361-1 ISBN 978-3-319-23362-8 (eBook)
DOI 10.1007/978-3-319-23362-8

Library of Congress Control Number: 2015948763

Springer Cham Heidelberg New York Dordrecht London

Springer International Publishing Switzerland 2016
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is
concerned, specically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction
on microlms or in any other physical way, and transmission or information storage and retrieval, electronic
adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not
imply, even in the absence of a specic statement, that such names are exempt from the relevant protective laws and
regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed
to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty,
express or implied, with respect to the material contained herein or for any errors or omissions that may have been
made.

Cover design: Jim Wilkie

Cover images: Front cover topThe Falcon 9 rocket streaks towards space from Floridas Cape Canaveral Air Force
Station containing supplies, including the rst 3D printer in space and a troop of 20 mice, for the International Space
Station (Courtesy SpaceX). Front cover lowerthe assembly and integration of a satellite in SSTLs clean-room
(Courtesy of Surrey Satellite Technology Ltd.). Rear coverVega VV05 in its mobile gantry prior to launch at
Europes Spaceport in Kourou, French Guiana (Courtesy ESA-M. Pedoussaut).

Printed on acid-free paper

Springer International Publishing AG Switzerland is part of Springer Science+Business Media

(www.springer.com)
Talking of education, People have now a-days, (said he,) got a
strange opinion that everything should be taught by lectures.
Now, I cannot see that lectures can do so much good as reading
the books from which the lectures are taken. I know nothing that
can be best taught by lectures, except where experiments
are to be shewn. You may teach chemistry
by lecturesYou might teach making of shoes by lectures!

Samuel Johnson, 1766

(from Boswells Life)
This book is dedicated to Cato and Dennis
Preface

This book, as implied by the title page, is an extensively revised version of the former
Metallurgical Assessment of Spacecraft Parts, Materials and Processes published in 1997.
The present title has been modied to set it apart from the previous work and describe its
expanded content. The book has become more voluminous, this reflects the huge advances
made during the past 20 years when we have witnessed the increased usage of modern
materials and manufacturing techniques that were unforeseeable when the former book was
written. Also, the number of case studies and amount of general information has been
extended to become a source for engineers, space scientists, laboratory experimenters and
technicians. Although much of the book considers metallurgical aspects of spacecraft engi-
neering, there is now basic advice covering organic and ceramic materials as well as tech-
niques available for assembling them into essential sub-systems, reliable parts and structures.
A good number of the original illustrations are retained but many new ones have been
added. Several images reflect the quite remarkable outcomes of space projects. These include
high resolution images of Earth taken by satellites which are relevant for surveillance and the
forecasting of weather. Also included are fly-by images of enigmatic little moons and comets
captured by spacecraft after many years of voyaging in search of life and the origins of water
in our own Solar System. Equipment on-board the International Space Station and
satellite-based communications are mentioned. These have all been made possible by
breakthroughs in materials, processes and electronic-engineering.
Plato saw engineers as doers not thinkers. From ancient times no one expected engi-
neers to question what they were asked to build and consider the consequences of such
achievements. Nowadays engineers are more condent in their social role and have learned to
say no when the products are questionable or environmental damage may occurthe
generation of space debris is one pertinent example. Hopefully, some lessons learnt guid-
ance may ensue from the case studies and failure analyses recorded in this book. In 1986
engineers said go to the Challenger launchother engineers said no but were over-ruled
and the space shuttle exploded shortly after lift-off. It is only in hindsight that we understand
that decision making can be extremely difcult, but such decisions must consider input from
all engineering disciplines and the recognition of material properties is vital.
A casual review of the Contents and Index will suggest to the reader that the subject matter
is likely to be of interest not only to spacecraft engineers, but in the broader sense, to workers
in quite different areas where metals, organic materials, composites, ceramics and glass are
used under terrestrial conditions or within high vacuum systems. Advancements in technology
always produce questions related to the reliability of new systems. Materials testing to agreed
codes of practice have been shown to help maximise the reliability of new materials, pro-
cesses, and applications. Metallography (or materialography) has led to an increased
understanding of failure modes. Much emphasis of this book has been placed on failure
analysis investigations. Each case must be developed in a logical mannerlarge-scale

ix
x Preface

(macroscopic) features are initially investigated, then the microscopic features of the materials
involved. Test specimen or samples of spacecraft hardware must be meticulously prepared,
then examined using both light and electron microscopy. It is amazing how these techniques
have evolved and how the recording of images has progressed. The author and his metallurgist
contemporaries may well remember early student days when contributions to reports were
exquisitely detailed hand drawn micrographs or images captured on photographic plates. The
digital revolution has now enabled all levels of detail to be recorded using super-resolution
microscopes and the future seems to be heading towards 3-dimensional microscopy.
In this book I have endeavoured to achieve a reasonable balance between general back-
ground knowledge and in-depth technical information. An elementary understanding of metals
and materials on the part of the reader is assumed. I have deliberately excluded a compre-
hensive account of the techniques employed in modern materials laboratories (unless
specically related to unusual space material test methods). Many texts are available and cited
in the Reference section. The Appendices have been extended and include many Tables
related to: spacecraft materials properties; alloy comparisons as they may be procured in
different countries; a simplied M&P management guideline for universities; and, examples of
Declared Materials and Processes Lists.
The space industry is a key sector in driving economic growth and creating new jobs. By
2030, the global space economy is predicted to be worth 400 billion per annum. At the time
of writing, the European space manufacturing industry alone has an unprecedented overall
turnover at 6 billion and a total direct employment of 38,000 persons. New spaceports will be
established and spaceplanes are most likely to be the next generations means for transporting
commercial and scientic payloads into orbit. Many future spacecraft engineers, space sci-
entist and technologists, all specialists in their own elds, may be aghast that some funda-
mental, old-hat information is contained in this book. But it is the lessons-learnt scenarios
that have brought us to where we are today. The industry is expanding and new employees
need to learn from our past mistakes and, at least, understand why certain design rules exist.
The wide acceptance of the previous book has been most welcome, and I hope the new
changes and additions will also nd approval by my colleagues in the space industry and
others in the wider engineering community.

Bosham, West Sussex Barrie D. Dunn

December 2015
Acknowledgments

This book has been brought about by the blending of various published research and inves-
tigation projects that I have undertaken as a metallurgist for the European Space Agency, from
some written works of others and from personal friends. I am especially grateful to the late Dr.
Jacques Dauphin my former Division Head at ESA who gave the encouragement to undertake
the writing of the earlier book. He was a native of the French province of Lorraine, where the
motto is Qui sy frotte sy pique which loosely translates to gather thistles, expect prick-
lesquite an apt maxim for those of us who have been involved with failure investigations.
I also acknowledge the help received from my former ESA colleagues: Dr. Ton de Rooij, Jack
Bosma, Guy Ramusat, Adrian Graham, David Collins and David Adams. Special thanks are
also given to Dr. Ernst Semerad, Dr. A. Merstallinger, Grazyna Mozdzen and Markus Fink
of the Aerospace and Advanced Composites GmbH (formally ARC), Wr. Neustadt, Austria,
with whom I have had many years of professional collaboration. As previously stated, there
has been a marked progress in this eld of materials technology, resulting in signicantly more
citations to references in this Edition, but even so, the bibliographic information certainly is
not complete. Where I have forgotten to cite a reference or credit an image I hope the author
will forgive my oversight.
I am also grateful to ESA and NASA for some of the illustrations used in the book. It
should be noted that the opinions expressed in this book are those of the author and do not
necessarily reflect the policy of the European Space Agency.
Let me add a special note of thanks to my late wife, Hanneke, my son, Martin, and my
daughter Harriet, for their patience through the spare-time hours that went into the making
of the previous Edition. Also, to Anne for her unswerving support and help editing this present
book. Stephen Hulcrofts assistance at BlueFish Computer Services, Chichester is appreciated.
I also wish to thank Clive Horwood, and the staff at Springer Praxis Books in Germany
(Ms. Janet Sterritt) and India (Mr. Antony Raj Joseph and Ms. Sivajothi Ganesarathinam), for
their assistance during the publication of this book.
The author would like to thank all his colleagues and friends at the following organisations
who kindly supplied new information, reference material and photographs:
Torbjrn Lindblom, Celsius Materialteknik, Karlskoga, Sweden.
Dr. Michael Osterman, The Centre for Advanced Life Cycle (CALCE), University of
Maryland, MD, USA.
S. Clment, Centre National dEtudes Spatiales, Toulouse, France.
Dr. H. Boving, Centre Suisse dElectronique et de Microtechnique SA, Neuchtel,
Switzerland.
H. Papenberg, DASA-ERNO Raumfahrttechnik GmbH (now Airbus Industries), Bremen,
Germany.
D. Bagley, ERA Technology, Leatherhead, UK.
Dr. A. Feest, The Harwell Laboratory, Metals Technology Centre, Harwell, UK.
W. Feuring, Heraeus GmbH, Hanau, Germany.
Massimo Bonacci, High Technology Center (HTC), Foligno, Italy.

xi
xii Acknowledgments

Poul Juul, Hytek, Aalborg, Denmark.

Messrs G. Kudielka and W. Maier, IFE, Oberpfaffenhofen, Germany.
Luca Moliterni and Gianluca Parodi, Italian Institute of Welding (IIS), Genoa, Italy.
Norio Nemoto, Japan Aerospace Exploration Agency (JAXA), Tsukuba, Japan.
Dr. Suman Shrestha, Keronite International Ltd., Haverhill, UK.
P. Fletcher, Airbus (formally MMS-UK), Portsmouth, UK.
Dr. Christopher Hunt, Martin Wickham and Ling Zou, The National Physics Laboratory,
Teddington, UK.
Dr. David Bernard, Nordson DAGE, Aylesbury, UK.
Jo Wilson and Bob Hussey, RJ Technical Consultants, Juicq, France.
Messrs Jrgen Svensson, U. Berg and Hans Ollfors, RUAG (formally Saab Ericsson
Space), Gothenburg, Sweden.
M.P. Hayes, The Spring Research and Manufacturers Association, Shefeld, UK.
Ian Turner, Cathy Barnes and Malcolm Snowdon, Spur Electon Ltd., Havant, UK.
Dr. R. Eckert, Standard Elektrik Lorenz, Stuttgart, Germany.
Dr. P. von Rosenstiel, Stichting Geavanceerde Metaalkunde, Hengelo, The Netherlands.
Luca Soli and Ulisse Di Marcantonio, Thales Alenia Space Italia, Milan, Italy.
Dr. J.M. Motz, Thyssen Guss AG, Mlheim a.d. Ruhr, Germany.
Stephen Kyle-Henney, TISICS Ltd., Farnborough, UK
Bill Strachan and Dr. Asa Barber, The University of Portsmouth, Portsmouth, UK.
K. Ring, Zentrum fr Verbindungs Technik, Gilching, Germany.
Robert Wm. Cooke, NASAJohnson Space Center, Houston, TX, USA
Pablo D. Torres, NASAMarshall Space Flight Center, Huntsville, AL, USA
Dr. Fabiola Brusciotti, Tecnalia, San Sebastian, Spain
Contents

1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

2 Requirements for Spacecraft Materials . . . . . . . . . . . . . . . . . . . . . . . . . ... 7

2.1 General Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 7
2.2 Considerations for Materials and Processes . . . . . . . . . . . . . . . . . . . ... 10
2.2.1 General Considerations During the Selection
of Materials and Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.2.2 Some Futuristic Ideas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2.3 Some Basic Considerations Regarding Corrosion Prevention . . . 17
2.2.4 Space Projects Phases and Management Events . . . . . . . . . . . . 20
2.3 The Effect of a Space Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.4 Materials for Space Launch Vehicles . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.5 Non-metallic Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.5.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.5.2 Classes of Non-metallic Materials . . . . . . . . . . . . . . . . . . . . . . 42
2.5.3 Novel Non-metallics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.6 The Potential for Welding and Joining in a Space Environment . . . . . . . . 49
2.6.1 Background Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2.6.2 Potential Joining and Cutting Processes . . . . . . . . . . . . . . . . . . 50
2.6.3 Expectations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

3 The Integration of Materials into Product Assurance Schemes . . . . . . . . . . 55

3.1 General Product Assurance and the Role of Materials . . . . . . . . . . . . . . . 55
3.1.1 Product Assurance Management . . . . . . . . . . . . . . . . . . . . . . . 55
3.1.2 Quality Assurance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.1.3 Reliability and Safety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3.1.4 Materials and Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.1.5 Component Part Selection, and Procurement . . . . . . . . . . . . . . 61
3.1.6 Control of Ground-Handling Facilities. . . . . . . . . . . . . . . . . . . 63
3.2 The Materials Laboratory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
3.2.1 Major Objectives of Laboratory . . . . . . . . . . . . . . . . . . . . . . . 66
3.2.2 Facilities and Instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . 67
3.2.3 The Use of New Laboratory Techniques for NDT . . . . . . . . . . 85
3.2.4 Organic Chemistry and Environmental Test Laboratories . . . . . . 98
3.3 Preparation of Materials and Metallographic Evidence. . . . . . . . . . . . . . . 100
3.3.1 The Metallographer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
3.3.2 Laboratory Records and Reports . . . . . . . . . . . . . . . . . . . . . . . 101
3.3.3 Report of Materials Data to Spacecraft Projects . . . . . . . . . . . . 101
3.3.4 Training of Materials Engineers and Laboratory Staff . . . . . . . . 103
3.3.5 Ethical Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104

xiii
xiv Contents

3.4 The Future for Materials Failure Investigations. . . . . . . . . . . . . . . . . . . . 104

3.4.1 The Larger Company . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
3.4.2 The Smaller Company. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
3.4.3 Product Liability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
3.5 Greener Spacecraft . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
3.6 The Potential for Recycling Electronic Waste. . . . . . . . . . . . . . . . . . . . . 111
3.6.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
3.6.2 Elemental Distribution for Spacecraft Electronic Box . . . . . . . . 111

4 Spacecraft ManufacturingFailure Prevention and the Application

of Material Analysis and Metallography . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
4.1 Sources of Failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
4.2 Drawings and Workmanship . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
4.2.1 Design and Manufacturing Drawings. . . . . . . . . . . . . . . . . . . . 115
4.2.2 Workmanship Standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
4.3 Mechanical Damage Revealed by Microstructure . . . . . . . . . . . . . . . . . . 122
4.4 Hydrogen Embrittlement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
4.4.1 Interaction of Metal with Hydrogen . . . . . . . . . . . . . . . . . . . . 122
4.4.2 Hydrogen Embrittlement of Spring Steel . . . . . . . . . . . . . . . . . 123
4.4.3 Blistering of Plated Aluminium Alloy . . . . . . . . . . . . . . . . . . . 124
4.4.4 Examination for Titanium Hydride Precipitates. . . . . . . . . . . . . 125
4.4.5 Embrittlement of Copper . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
4.4.6 Future Developments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
4.5 General Corrosion Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
4.5.1 Bimetallic Corrosion-Related Failures . . . . . . . . . . . . . . . . . . . 128
4.5.2 Corrosion Resistance of Anodic and Chemical Conversion
Coatings on Al 2219 Alloy . . . . . . . . . . . . . . . . . . . . . . . . .. 132
4.5.3 Evaluation of Alodine Finishes on Common Spacecraft
Aluminium Alloys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
4.5.4 Cleaning, Passivation, and Plating of Spacecraft Steels . . . . . . . 137
4.5.5 Launch Site Exposure and Corrosion. . . . . . . . . . . . . . . . . . . . 138
4.6 Stress-Corrosion Resistance of Metals . . . . . . . . . . . . . . . . . . . . . . . . . . 139
4.6.1 Stress-Corrosion Cracking . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
4.6.2 SCC Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
4.6.3 The Properties of Spring Materials . . . . . . . . . . . . . . . . . . . . . 144
4.6.4 Bearing Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
4.7 Control of Printed Circuit Boards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
4.7.1 Chemical Composition of Tin-Lead from Microstructure . . . . . . 148
4.7.2 Grainy Solder Coverage on PCBs and the Effects of Rework . . . 150
4.7.3 Evaluation of Multilayer Board Internal Connections. . . . . . . . . 155
4.7.4 Flexible Circuits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
4.7.5 Hot-Air-Levelled Circuit Boards . . . . . . . . . . . . . . . . . . . . . . . 160
4.7.6 Solder Assembly of Component Packages onto Multilayer
Boards with High Heat Capacity . . . . . . . . . . . . . . . . . . . . . . 161
4.8 Control of Composite Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
4.8.1 MetalMatrix Composites for Space Structures. . . . . . . . . . . . . 161
4.8.2 Composite Contact Devices . . . . . . . . . . . . . . . . . . . . . . . . . . 164
4.8.3 Fibre-Reinforced Plastic Composites . . . . . . . . . . . . . . . . . . . . 166
4.8.4 Fibre-Reinforced Glass Ceramics . . . . . . . . . . . . . . . . . . . . . . 170
4.8.5 CarbonCarbon Composites . . . . . . . . . . . . . . . . . . . . . . . . . . 170
4.8.6 Metal Matrix Composites for Spacecraft Pressure Vessels . . . . . 172
Contents xv

4.9 Control of Capillary Screens . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 172

4.10 Examination of Electroless Nickel Deposits . . . . . . . . . . . . . . . . . . .... 173
4.10.1 Microcracked Electroless Nickel . . . . . . . . . . . . . . . . . . . .... 173
4.10.2 Electroless Nickel Plating of Aluminium
Electronic Housings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
4.11 Control of Electroforming Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
4.12 Dip Brazing of Aluminium Alloys . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
4.13 Considerations for the Assembly of Subsystems by Welding . . . . . . . . . . 181
4.13.1 General Welding Methods and Controls . . . . . . . . . . . . . . . . . 181
4.13.2 Electron Beam Welding. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
4.13.3 Laser Beam Welding. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
4.13.4 Explosive Welding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
4.13.5 Welding of AluminiumLithium Alloys. . . . . . . . . . . . . . . . . . 187
4.13.6 Welding of Thermoplastics for Space Applications . . . . . . . . . . 188
4.14 Control of Power System Weldments . . . . . . . . . . . . . . . . . . . . . . . . . . 189
4.14.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
4.14.2 Welded Solar Arrays. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
4.14.3 Suitability of Welded Battery Cells . . . . . . . . . . . . . . . . . . . . . 193
4.15 Problems Associated with Residual Stresses in Weldments . . . . . . . . . . . 195
4.16 Electromagnetic Emission from TIG Welding Equipment . . . . . . . . . . . . 195
4.17 Titanium Aluminides for High-Temperature Applications . . . . . . . . . . . . 196
4.18 Shape-Memory Alloys for Spacecraft Devices . . . . . . . . . . . . . . . . . . . . 197
4.19 Foamed Aluminium for Damping Purposes . . . . . . . . . . . . . . . . . . . . . . 202
4.20 Superplastic Forming and Diffusion Bonding of Metals. . . . . . . . . . . . . . 203
4.20.1 Forming of Propellant Tanks . . . . . . . . . . . . . . . . . . . . . . . . . 203
4.20.2 Diffusion Bonding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
4.20.3 Superplastic Forming and Diffusion Bonding
in One Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
4.21 Cleaning of Mechanical Parts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
4.21.1 General Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
4.21.2 Metallic Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
4.21.3 Cleaning of Individual Parts. . . . . . . . . . . . . . . . . . . . . . . . . . 210
4.21.4 Cleaning of Metallurgically Joined Assemblies . . . . . . . . . . . . . 213
4.21.5 Maintenance of Cleanliness . . . . . . . . . . . . . . . . . . . . . . . . . . 216
4.21.6 Cleaning of Silicone Contamination . . . . . . . . . . . . . . . . . . . . 219
4.22 Novel Thermal Management Materials . . . . . . . . . . . . . . . . . . . . . . . . . 220
4.23 Cold Sprayed Coatings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
4.24 Advanced Plasma Electrolytic Oxidation Treatment for Aluminium,
Magnesium and Titanium Alloys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
4.24.1 General Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
4.24.2 Characteristics of PEO Coatings . . . . . . . . . . . . . . . . . . . . . . . 225
4.24.3 Applications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
4.25 Joining by Friction Stir . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
4.25.1 Friction Stir Welding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
4.25.2 Friction Stud Welding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
4.26 Selective Brush Electroplating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
4.27 Control of Coatings and Bonded Items by Tape Testing . . . . . . . . . . . . . 237
4.28 The Application of EB Welding Machine for Reflow Brazing . . . . . . . . . 239
xvi Contents

5 Metallography Applied to Spacecraft Test Failures . . . . . . . . . . . . . . . . . . . 247

5.1 Application of Electron Microscope . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
5.1.1 SEM Examination of Fracture Surfaces . . . . . . . . . . . . . . . . . . 247
5.1.2 TEM Examination of Metallic Failures . . . . . . . . . . . . . . . . . . 250
5.2 Fasteners . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
5.2.1 Spacecraft Fasteners . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
5.2.2 Fastener Failure Due to Forging Defect . . . . . . . . . . . . . . . . . . 254
5.2.3 Laps and Surface Irregularities in Threads . . . . . . . . . . . . . . . . 255
5.2.4 Hydrogen Embrittlement of Steel Fasteners . . . . . . . . . . . . . . . 255
5.2.5 Embrittlement of Titanium Alloys. . . . . . . . . . . . . . . . . . . . . . 255
5.2.6 Galvanic Corrosion of Fasteners . . . . . . . . . . . . . . . . . . . . . . . 257
5.2.7 Contamination and Organic Fastener Lubrication Systems . . . . . 257
5.2.8 Metallic Particle Generation . . . . . . . . . . . . . . . . . . . . . . . . . . 258
5.2.9 Quality Assurance Controls for Fasteners. . . . . . . . . . . . . . . . . 261
5.3 Thermal History from Microstructure . . . . . . . . . . . . . . . . . . . . . . . . . . 262
5.4 Effect of Inclusions Within the Microstructure of Explosively
Deformed Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264
5.5 Degradation of Passive Thermal Control Systems . . . . . . . . . . . . . . . . . . 266
5.5.1 General Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266
5.5.2 Low-Emissivity Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268
5.5.3 High-Absorption Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
5.5.4 Rigid Optical Solar Reflectors . . . . . . . . . . . . . . . . . . . . . . . . 270
5.5.5 Flexible Second Surface Mirrors. . . . . . . . . . . . . . . . . . . . . . . 271
5.6 Sublimation of Metals. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
5.6.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
5.6.2 Sublimation of and Condensation of Cadmium and Zinc . . . . . . 274
5.6.3 Heater Sublimation Problem Associated with Thruster Motor . . . 276
5.6.4 Sublimation of Klystron Cathode-Heaters . . . . . . . . . . . . . . . . 276
5.6.5 Sublimation of Rhenium . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278
5.7 Beryllium for Spacecraft Applications . . . . . . . . . . . . . . . . . . . . . . . . . . 280
5.7.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280
5.7.2 Health and Safety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
5.7.3 Integrity of Machined Beryllium. . . . . . . . . . . . . . . . . . . . . . . 283
5.7.4 Thermal Cycling on Work-Hardened Beryllium . . . . . . . . . . . . 284
5.7.5 General Etching Solutions for Beryllium . . . . . . . . . . . . . . . . . 285
5.7.6 Investigation of Microcracked Thin-Foil Detector Windows . . . . 286
5.7.7 Aluminium-Beryllium Alloys . . . . . . . . . . . . . . . . . . . . . . . . . 288
5.8 Deactivation of Catalyst Particles for Hydrazine Decomposition . . . . . . . . 288
5.8.1 Testing Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
5.8.2 Material Investigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
5.8.3 Mechanism of Particle Deactivation . . . . . . . . . . . . . . . . . . . . 290
5.9 Cathode Emitter Degradation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
5.10 Investigation of a Failed Spacecraft Antenna . . . . . . . . . . . . . . . . . . . . . 293
5.11 The Wear of Ball Bearings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296
5.12 Cold Welding of Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304
5.12.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304
5.12.2 Cold Welding Due to Cyclic, Impact Loading . . . . . . . . . . . . . 306
5.12.3 Cold-Welding Due to Fretting . . . . . . . . . . . . . . . . . . . . . . . . 307
5.13 Defective Black-Anodized Electrical Connector . . . . . . . . . . . . . . . . . . . 308
5.14 Contaminant ParticlesIdentification of Their Sources . . . . . . . . . . . . . . 309
Contents xvii

5.15 Silicone Contamination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310

5.15.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
5.15.2 Contamination of Black-Anodized Finish. . . . . . . . . . . . . . . . . 311
5.15.3 Contamination of Invar Moulding Tool . . . . . . . . . . . . . . . . . . 312
5.15.4 Removal of Silicone Polymers . . . . . . . . . . . . . . . . . . . . . . . . 314
5.15.5 Contamination of Aluminium Tubes for Vacuum Pinch-Offs . . . 317
5.16 Magnetic Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
5.17 Thermal Stress-Induced Dimensional Changes . . . . . . . . . . . . . . . . . . . . 319
5.17.1 General Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319
5.17.2 Stress-Relaxation by Thermal Gradients. . . . . . . . . . . . . . . . . . 319
5.17.3 Thermally Induced Vibrations . . . . . . . . . . . . . . . . . . . . . . . . 321
5.18 Defects in Titanium Piece-Parts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323
5.18.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323
5.18.2 Alpha-Case Embrittlement . . . . . . . . . . . . . . . . . . . . . . . . . . . 323
5.18.3 Titanium Hydride Embrittlement. . . . . . . . . . . . . . . . . . . . . . . 324
5.19 Leaking Water Tank on Launcher. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325
5.20 Compatibility of Liquid and Solid Propellants with Components
and Subsystems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 326

6 Failure Analysis of Electrical Interconnections and Recommended

Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
6.1 Material Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
6.2 Welded Lead Wire Interconnections . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
6.3 Purple Plague . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332
6.4 Mechanical Electrical Connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
6.4.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
6.4.2 Wire-Wrapped Connections . . . . . . . . . . . . . . . . . . . . . . . . . . 337
6.4.3 Crimped Joints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
6.5 Soldered Interconnections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340
6.5.1 Introduction to Soldering . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340
6.5.2 Inspection of Soldered Joints . . . . . . . . . . . . . . . . . . . . . . . . . 341
6.5.3 The Effect of Thermal Fatigue on Solder-Assembled
Leaded Components . . . . . . . . . . . . . . . . . . . . . . . . ....... 344
6.5.4 Effect of Thermal Fatigue on Leadless Components . . ....... 351
6.5.5 The Effect of Thermal Fatigue on Semi-rigid
Cable Connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353
6.6 Problems Associated with Coatings for Soldering Applications . . . . . . . . 357
6.6.1 The Need for Coatings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357
6.6.2 Surfaces that Can Be Soldered To . . . . . . . . . . . . . . . . . . . . 357
6.6.3 Surfaces that Can Be Soldered Through . . . . . . . . . . . . . . . . 359
6.7 The Use of Indium Solder Alloys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363
6.8 Wires and Cables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369
6.8.1 Selection of Plated Finish on Copper Conductors . . . . . . . . . . . 369
6.8.2 Effect of Ageing on the Solderability of Tin-Plated
and Silver-Plated Wires . . . . . . . . . . . . . . . . . . . . . . ....... 371
6.8.3 Red Plague Corrosion of Silver-Plated Copper,
and Plagues on Other Plated Stranded Wires . . . . . . . . . . . . . . 375
6.8.4 Manganin Wire . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 379
6.8.5 High-Voltage Wires, Cables, and Connections . . . . . . . . . . . . . 380
6.8.6 Cold Welding of Stranded Wires and Cables . . . . . . . . . . . . . . 380
xviii Contents

6.9 Problems Associated with Soldering Fluxes . . . . . . . . . . . . . . . . . . . . . . 380

6.9.1 Purpose of a Flux . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 380
6.9.2 Heat-Shrinkable Sleeves Containing Solder Preforms . . . . . . . . 381
6.9.3 Stress Corrosion of Component Lead Material . . . . . . . . . . . . . 383
6.9.4 Flux-Corrosion of Silver-Plated Stranded Wires . . . . . . . . . . . . 383
6.9.5 Selection of a Soldering Flux or a Solderable Finish . . . . . . . . . 386
6.9.6 Control of Galvanic Corrosion . . . . . . . . . . . . . . . . . . . . . . . . 389
6.9.7 Cleaning of Flux-Contaminated Surfaces . . . . . . . . . . . . . . . . . 389
6.9.8 Flux Residues, Their Ingress into Top-Coat of PCB Surfaces,
and Bake Out After Cleaning . . . . . . . . . . . . . . . . . . . . . . . .. 391
6.9.9 Conductive Anodic Filament (CAF) Formation
and Particulate Contamination . . . . . . . . . . . . . . . . . . . . . . . . 394
6.9.10 Potential Health Hazards in the Electronic Assembly Area . . . . . 398
6.10 Problems Associated with Brazing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 399
6.10.1 Design Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 399
6.10.2 Brazeability of Materials and Braze Alloy Compositions . . . . . . 400
6.10.3 Brazing Fluxes and Their Removal . . . . . . . . . . . . . . . . . . . . . 403
6.10.4 Atmospheres for Brazing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 404
6.10.5 Safety Precautions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 405
6.10.6 Produce Assurance Applied to Brazing Operations . . . . . . . . . . 405
6.10.7 Inspection Criteria for Brazed Aluminium Alloy
Waveguide-to-Flange Joints . . . . . . . . . . . . . . . . . . . . . . . . . . 406
6.11 Diffusion Soldering/Brazing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 408
6.12 Effects of Rework and Repair on Soldered Interconnections . . . . . . . . . . 408
6.12.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 408
6.12.2 Cosmetics of Solder Fillets . . . . . . . . . . . . . . . . . . . . . . . . . . 410
6.12.3 Effect of Rework Electronic Components . . . . . . . . . . . . . . . . 410
6.12.4 Effect of Rework on Plated-Through Holes . . . . . . . . . . . . . . . 410
6.12.5 Effect of Rework on Composition of Joint . . . . . . . . . . . . . . . . 412
6.12.6 Recuperation of Unsolderable PCBs and Component Leads . . . . 413
6.13 Electrical Conductive Adhesives. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 413
6.14 Training and Certification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 415
6.14.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 415
6.14.2 Certification for Electronic Assembly Techniques . . . . . . . . . . . 417
6.14.3 Understanding Process-Induced Failures
and the Importance of Workshops. . . . . . . . . . . . . . . . . . . . .. 418
6.15 Verification of Surface-Mount Technology and Prevalent
Failure Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 419
6.15.1 Verification Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 419
6.15.2 Failure Under Mechanical Overloading . . . . . . . . . . . . . . . . . . 422
6.15.3 Failures Due to Board Flatness Problems. . . . . . . . . . . . . . . . . 422
6.15.4 Failure Due to Co-planarity Problems . . . . . . . . . . . . . . . . . . . 423
6.15.5 Solder Joint Failure Due to Thermal Mismatch
Between SMD and Substrate . . . . . . . . . . . . . . . . . . . . . . . . . 425
6.15.6 Conductor Track Failure Due to Thermal Mismatch . . . . . . . . . 428
6.15.7 Failure of RF Cables Connected by SMT . . . . . . . . . . . . . . . . 428
6.15.8 SMT Solder Joint Failure Due to Conformal Coatings. . . . . . . . 428
6.15.9 SMT Problems Related to Flux and White Residues . . . . . . . . . 432
6.15.10 Area Grid Array (AGA) Packaging . . . . . . . . . . . . . . . . . . . . . 434
Contents xix

6.15.11 High Voltage Interconnections and Influence

of Geometry (Workmanship) on Corona Discharge . . . . . . . . . . 442
6.15.12 Tin Pest. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 448
6.15.13 Mechanical and Electrical Properties of Electronic
Materials at Temperatures Down to 4.2 K . . . . . . . . . . . . . . . . 451

7 Whisker Growths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 461

7.1 The Problem of Whisker Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 461
7.2 Analysis of Failures Due to Whisker Growth . . . . . . . . . . . . . . . . . . . . . 462
7.2.1 Molybdenum Whiskers on Metallized Miniature Circuits . . . . . . 462
7.2.2 Tungsten Whisker Growth Within Travelling Wave Tubes. . . . . 466
7.2.3 Metal Oxide Whisker Precipitation in Glass Seals. . . . . . . . . . . 466
7.2.4 Integrated Circuit Failure Modes Due
to ElectromigrationAluminium Whisker Growth
and Solder Joint Voiding . . . . . . . . . . . . . . . . . . . . . . . . .... 468
7.3 Tin Whisker Growths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 472
7.3.1 Tin Whisker Growth on a Plated Steel Housing . . . . . . . . .... 472
7.3.2 Tin Whisker Growth on PCB and Other Electronic
Materials During Thermal Cycling . . . . . . . . . . . . . . . . . .... 474
7.3.3 Tin Whisker Growth on Crimp Termination Devices. . . . . .... 479
7.3.4 The Nucleation, Growth and Mechanism of Growth
of Tin WhiskersResults from a C-Ring Test Programme . . . . . 481
7.3.5 Some Properties of Tin Whiskers . . . . . . . . . . . . . . . . . . . . . . 485
7.4 Precautions to Avoid General Whisker Growths . . . . . . . . . . . . . . . . . . . 491
7.5 The Creation of Lead-Free Control Plans. . . . . . . . . . . . . . . . . . . . . . . . 494
7.5.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 494
7.5.2 Methods for Reprocessing Pure Tin Terminations . . . . . . . . . . . 495
7.5.3 Mitigation Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 498

8 Assessment of Post-flight Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 501

8.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 501
8.1.1 Hardware Return from Space . . . . . . . . . . . . . . . . . . . . . . . . . 501
8.1.2 Raw Materials from the Moon . . . . . . . . . . . . . . . . . . . . . . . . 501
8.1.3 Recent Investigations Using Retrieved Materials . . . . . . . . . . . . 503
8.2 Space Environmental Effects from Vacuum and Radiation. . . . . . . . . . . . 503
8.2.1 Organic Materials and Lubricants . . . . . . . . . . . . . . . . . . . . . . 503
8.2.2 Radiation Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 507
8.2.3 Effects of Vacuum on Metals . . . . . . . . . . . . . . . . . . . . . . . . . 508
8.3 Temperature Cycling. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 509
8.4 Micrometeoroids and Debris . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 509
8.4.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 509
8.4.2 Debris Emanating from Catalytic Bed Thruster Motors . . . . . . . 512
8.4.3 Returned Hardware . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 514
8.4.4 Protection Shields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 515
8.5 Effect of Atomic Oxygen on Materials . . . . . . . . . . . . . . . . . . . . . . . . . 517
8.6 Decelerators and Heat Shield Materials . . . . . . . . . . . . . . . . . . . . . . . . . 524
8.6.1 General Examples. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 524
8.6.2 Beryllium as a Heat Shield . . . . . . . . . . . . . . . . . . . . . . . . . . 528
8.6.3 Alternative Heat Shield Materials . . . . . . . . . . . . . . . . . . . . . . 531
8.6.4 High-Temperature Fasteners . . . . . . . . . . . . . . . . . . . . . . . . . . 533
xx Contents

8.7 Manned Compartments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 535

8.7.1 General Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 535
8.7.2 Solder Assembly Defects . . . . . . . . . . . . . . . . . . . . . . . . . . . . 538
8.7.3 Inspection of Spacelab Post-flight Hardware. . . . . . . . . . . . . . . 542

Appendix 1: Coefficient of (Linear) Thermal Expansion for Selected

Materials (COE or CTE) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 557

Appendix 2: Properties of Printed Circuit Laminates . . . . . . . . . . . . . . . . . . . . 559

Appendix 3: Reagents for Microetching Metals and Alloys . . . . . . . . . . . . . . . . 561

Appendix 4: Conversion Table for Mechanical Properties . . . . . . . . . . . . . . . . . 565

Appendix 5: Aluminium Alloy Temper Designations . . . . . . . . . . . . . . . . . . . . . 567

Appendix 6: Metal Alloy Comparison Tables . . . . . . . . . . . . . . . . . . . . . . . . . . 571

Appendix 7: Variation of Standard Free Energy of Formation

of Oxides with Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 613

Appendix 8: Simplied Procedure for the Management of Materials,

Processes and Mechanical PartsPossible Guidelines
for a Cubesat or Small University Spacecraft . . . . . . . . . . . . . . . . 615

Appendix 9: Materials and Processes Standards Related to Space

(Released by ECSS, JAXA and NASA) as of 2015 . . . . . . . . . . . . . 619

Appendix 10: Examples of Declared Process Lists (DPL). . . . . . . . . . . . . . . . . . 621

Appendix 11: Examples of Declared Materials Lists (DMLs) . . . . . . . . . . . . . . . 625

Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 629

References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 639

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 655
 Introduction
It is always impressive to look at the hardware of space
ventures, whether in the form of launch vehicles culminating
with the successful launch and landings of the Space Shuttle,
of land-sited test chambers, of satellites being tested, of large
antenna dishes, or of complex electronic circuitry under high
magnication. But this is not the real space capability. The
real capability lies in the people, in their technical compe-
tence, and in their manipulations of the metals and materials
which have made space communication programmes and
space science programmes possible.
Every 24 h an average of 45,000 storms break around the
world. Meteorology is probably one of the oldest sciences,
but it is presently one of the least accurate even though the
four elements which enable the weather to be forecast
(clouds, atmospheric pressure, temperature, and wind) are
perfectly measurable. The various space programmes
undertaken by the US, the USSR, Europe, and more recently
Japan, India, China, South Korea and Brazil, are providing
considerable amounts of information about the environment.
The weather develops in a restricted area consisting mainly
of the troposphere, the lowest layer of the atmosphere, which
never exceeds 16 km in depthbarely twice the height of
Mount Everest. But no part of the atmosphere will act
independently of the other, and there is a continuing need to
answer such questions as, how stable is our climate, or, how
much additional atmospheric or water pollution can be tol-
erated without drastically altering it? Such knowledge comes
in many ways, and no more dramatically than in recent years
from sensors in satellites orbiting the Earth, as illustrated in
Figs. 1.1, 1.2, and 1.3, from observations and photographs
made by astronauts, and even by correlation with the
atmospheres of Mars, Venus, and Saturn which have also
been investigated recently by space probes and landers.
Wonderful images made by the Hubble Space Telescope
have focused on galaxies in an innite universe, but this
orbiting observatory has also assisted in the expansion of
knowledge of our own planet. The greenhouse effect, in
which pollutants put into the atmosphere by the burning of
Springer International Publishing Switzerland 2016
B.D. Dunn, Materials and Processes, Springer Praxis Books,
DOI 10.1007/978-3-319-23362-8_1
1
fossil fuels and industrial processes trap heat and lead to the
harmful warming of the planet, can be evaluated by using
the new Earth resources satellites. These remote sensing
spacecraft can also study deforestation, damage to the rain
forests of South America, Africa, and Asia, and the effect of
sewage and industrial waste on our oceans and shorelines.
Meteorology, Earth phenomena observations,
space-based satellite navigation systems (GPS) and com-
munications by satellites are four major areas of more
immediate interest where space technology can be applied
usefully for the service of mankind. GPS saves mankind
billions of dollars per year in wasted fuel for cars, aircraft
and wear/tear! The Galileo global navigation satellite system
(GNSS) should see 30 satellites in operation by 2019, pro-
viding users with horizontal and vertical position measure-
ments within a 1-m precision. Telecoms satellites in a
geostationary orbit, 22,236 miles above the Earths surface
enable internet signals to travel from one user to another
anywhere in the world within 700 ms. These satellites allow
us to speak and to see from one country or continent to
another; to receive the same television picture simultane-
ously in Britain, Luxemburg, America, Russia, or New
Zealand will educate and entertain the people of the world
we may hope that the so-called telecom satellites will help to
abolish the frontiers of misunderstanding and ignorance.
This could be the key to the continuation of our civilization.
The aforementioned attributes to space also support the
United Nations goals of enhancing the literacy of everyone
on the plant, eradicating poverty and improving health.
Of particular concern to the space industries is the
unpredictability of the Suns weather. It has been predicted
that after years of relative calm, the years 2015 onwards may
see a period of more intense activity such as solar flares,
coronal mass ejections and magnetic storms. This so-called
space weather has been added to the UKs National Risk
Register as neutron-storms risk degrading satellite data and
hardware, as well as on-Earth facilities that use sensitive
electronic components such as metal-oxide-semiconductor
1
2 1 Introduction

Fig. 1.1 Example of visible image: clouds appear white and space is seen as black (courtesy NASA)

eld-effect transistors (MOSFETs). Five spacecraft, in the
project called High Energy Solar Physics Data in Europe,
will continue to observe solar flares, explosions on the Sun
and the magnetic phenomena that flings radiation and par-
ticles across the Solar System.
Permanently manned space stations such as the Rus-
sian MIR and more substantially, the International Space
Stationexpected to remain operational until 2024have
been realized. Chinese scientists are planning an orbital
space station where the core module will be launched in
2018, followed by two laboratory modules in 2020 and
2022. Settlements on the Moon are periodically proposed by
NASA and other space agencies, but since the discovery of
water-ice close to one of the lunar poles, this concept may
become a reality. The remarkable Mars exploration projects
involving orbiting spacecraft and rovers are establishing the
1 Introduction
Fig. 1.2 Example of infrared image of the Earth from space; cold and high clouds appear white. Warmer (low) clouds grey (courtesy ESA)
habitability of that planet. Now that it has been established
that water exists below the Martian surface, the future
presence of manmade colonies on Mars may not be just the
fantasy of science-ction writers.
The advancements in space technology have been made
possible by many specic breakthroughs in materials science,
manufacturing processes and novel technological advance-
ments. Novel organic and inorganic materials have been
developed by new methods of synthesis and metallurgical
3
processing and these are facilitating the development of
highly sophisticated spacecraft subsystems and electronic
devices. As these developments have continued it is perhaps
surprising to note that, in the main, it is the ordinary metals
and plastic materials and their related manufacturing or
assembly processes that nd their way into the construction
of space hardware. Consequently we have now determined
that these ordinary materials produce the greatest prepon-
derance of failures during the various stages of spacecraft
4 1
Fig. 1.3 Example of water-vapour channel image. Taken by Meteosat Second Generation spacecraft in 2006 (courtesy Eumetsat)
programmes. The understanding of material and component
limitations is a real requirement during the development of
spacecraft systems in order that materials engineers can
predict realistic margins of safety.
Failures may occur as a result of over-testing, overload-
ing, or over-pressurization when it can be demonstrated that
no fault is attached to the material itself. Other failures occur
because of poor choice of material, shortcomings in design,
mistreatment during construction, or when the part was not
adequate to withstand a particular fatigue or corrosive
environment. Spacecraft failures occur during their fabrica-
tion, assembly, integration, and environmental testing, and
the generally short period of approximately four years
between design and launch necessitates that failure analyses
are rapidly performed to identify failure modes and their
causes. This can be achieved only when sufcient informa-
tion is available about the history of the failed part, from its
initial composition and heat treatment, through manufac-
turing details to records of all post-manufacture storage and
testing. Fortunately, the high degree of surveillance by the
Product and Quality Assurance teams of space hardware
contractors usually enables proper documentation of most
details pertinent to a failure. This knowledge of the precise
Introduction
operating stresses and environmental conditions is of great
help in the diagnosis of a failure mode.
There is a grave danger that the line of development of
space equipment and instrumentation may be lost if care is
not taken to preserve documentation related to past failures.
Much information is contained in in-house laboratory reports
which are often led and forgotten. This book may go a little
way to collate a small percentage of the examinations per-
formed by the author in order that similar design or pro-
duction problems may be minimized in the future. The
classical failure modes of fatigue, stress corrosion cracking,
hydrogen embrittlement, and the degradation of polymers by
ageing, outgassing radiation, etc. have frequently been
associated with spacecraft hardware. It is expected that
chemists, metallurgists and other material or manufacturing
engineers, who have metallic failures to contend with, will
be able to draw parallel examples to their work from the
illustrated case histories included within this book. Many of
these examples necessitate that some of the material
requirements for space flight are understood, as these are
frequently directly attributable to particular failure modes.
An overview of the specic requirements of spacecraft
materials is given in Chap. 2, and the role of material
1 Introduction
evaluations vis--vis spacecraft product assurance schemes
is described in Chap. 3.
The case histories have been selected from a large num-
ber of examinations involving standard material testing
techniques, and they are divided into three characteristic
groupings. Chapter 4 concerns problems encountered during
certain spacecraft manufacturing phases; Chap. 5 relates to
failures which possibly occurred during testing; and Chap. 6
concerns failures which may cause deterioration of electrical
interconnections.
Chapter 7 deals with the many aspects of whisker growth
as they might affect practising materials engineers associated
with structural materials, electronics, and space-related
industries. The case histories presented are limited to elec-
trical device failures that have resulted directly from the
growth of whiskers. Corrective actions are proposed for each
case, and it will be evident that the published literature does
not hold a solution for every situation.
Chapter 8 is a short overview of the effect that the space
environment has on materials. Here, examples are given
from the large number of analyses made on materials
returned from Low Earth Orbit (LEO). The effects of out-
gassing, temperature, micrometeoroids, and atomic oxygen
are illustrated from data amassed by Space Shuttle flight
experiments, the Long Duration Exposure Facility, Eureca,
Medet and materials retrieved from the Hubble Space
Telescope during its various repair missions.
The major tool utilized for each investigation is that of
microscopical metallographya technique developed by H.
C. Sorby in the 1860s for the examination of geological
samples. However, the number of tools available to materials
engineers for the examination of polymeric molecules,
microstructure and the like, has increased enormously over
the last decade. It is often necessary to utilize advanced
instruments such as ESCA and Auger spectrometers, laser
5
microprobe analysers, high-denition radiographic units,
scanning-laser acoustic microscopies, infrared spectroscopy,
etc. as diagnostic tools for failure analysis. These are fre-
quently available on loan, or can be rented from local uni-
versities or research establishments. It must be emphasized
that the majority of these tools give very limited information
to the inexperienced. Only with sufcient practical experi-
ence or on-the-spot guidance will the investigator be able to
piece together information gained from various stages of
nondestructive testing, supercial binocular microscope
observations, physical and analytical tests, and
metallography.
In the scope of this book it is not possible to include a
description of the equipments selected for particular inves-
tigations, nor is there place for detailed accounts of the
methods which have been chosen. The traditional guidelines
applicable to all engineering failures have been followed by
the author who is indebted to his highly professional col-
leagues as well as access to well-equipped materials labo-
ratories and test facilities. As will be discussed, the majority
of organic material and metallurgical investigations are made
during hardware production or after equipment level testing.
Some defective items also originate from units or structures
which have been installed on-board engineering or quali-
cation model spacecraft, so that results and recommenda-
tions from material and failure review boards can be fed
back to project designers and engineers. This procedure will,
we hope, eliminate future problems with flight model
spacecraft. This book deals not only with failure analysis but
also with the measures which may be taken for failure pre-
vention by improving product reliability. Finally, the term
failure can be construed to have many meanings ranging in
scope from the trivial to the calamitous. Throughout this text
failure is employed as a technical term meaning cessation
of function or usefulness.
 Requirements for Spacecraft Materials
2.1 General Background
The Space Age began in 1957, with an 83 kg Russian
Sputnik satellite bleeping greetings to a surprised world.
Since that spectacular beginning, intensive effort has gone
into the scientic exploration of space, exploration of the
Moon and distant planets, manufacturing of materials in
space laboratories, and exploiting orbiting satellites for
communication, navigation and observation of the Earth.
The early steps have passed into history, and most equip-
ment and instrumentation has been and will continue to be
replaced by lighter and more complex substitutes. The
remarkable achievements of the Apollo Lunar Exploration
Programme two decades ago still tend to overshadow the
unmanned automated satellite flights, and it is not always
realized that spacecraft orbiting above all continents of the
world have already revolutionized global communications,
maritime navigation, and worldwide weather forecasting.
These satellites are now vital links in a global network. They
would not have been economically or technically feasible
before the advent of near-Earth space explorations.
Satellite communications started on a commercial basis
with the launch of Early Bird in 1965, less than eight years
after the launch of the rst Sputnik. This was the rst
satellite to remain stationary over the Earth, and it was able
to provide a continuous connection between any two Earth
stations. Until comparatively recently these so-called ap-
plications satellites were merely assemblies of separately
designed components rather than thoughtfully integrated
systems. Often component interfaces failed to match,
reducing the overall system performance.
These satellites, and to a more limited extent the scien-
tic satellites, are now incorporating standardized subsys-
tems in an attempt to optimize performance factors including
weight, reliability, and cost.
It seems likely that the spacecraft designer has placed
greatest emphasis on mass, as this is usually set by the
Springer International Publishing Switzerland 2016
B.D. Dunn, Materials and Processes, Springer Praxis Books,
DOI 10.1007/978-3-319-23362-8_2
2
capabilities of the assigned launch vehicle which will take
the satellite from the Earths surface and inject it into the
desired orbit. The lighter the satellite, the cheaper will be the
launch costs. Another major performance factor, reliability,
can also be purchased if money is preferentially funneled
into reliability and test programmes rather than launch
vehicles. The important point is that performance factors of
weight, reliability, and cost are all interrelated. The designer
of an applications satellite will be more willing to pay for the
reliability level that would give him 10 years of operation
than the designer of a scientic satellite designed to shut off
transmission after only one year when the mission objectives
are attained.
One of the major aims of the European satellite manu-
facturer has been to set up a European communications
programme which will develop and launch long-life satel-
lites. A supporting technology programme has been under-
taken to develop and qualify most of the critical subsystems
that will enter the design of future operational satellites. An
experimental satellite (Orbital Test SatelliteOTS) was
launched in 1978 to evaluate and test the performance of the
various subsystems of future European communication
satellite systems. OTS and its launcher are illustrated in
Figs. 2.1a and 2.2. Its major subsystems under evaluation
included:
communicationsto relay information (data and com-
mands) between Earth and satellite and, in concept, to
and from other spacecraft.
power supplyto provide electrical power to all satellite
subsystems.
on-board propulsionto provide thrust for orbit changes,
station-keeping, and deorbiting.
Environmental controlto maintain specied tempera-
tures, radiation levels, electromagnetic environment, etc.
structureto support and maintain satellite conguration
on the ground, during launch and in orbit.
7
8 2 Requirements for Spacecraft Materials

Fig. 2.1 a OTS structural model during vibration testing in 1975. France, 15 March 2013. This communications satellite is 7.1 m high
Thermal blankets are not yet tted. This is the rst ESA communication (ESA). c These 9 m-high spike-lined walls enclose the hushed interior
satellite and has a height of 2.5 m (ESA). b View of the Alphasat of ESAs Maxwell test chamber, which isolates satellites from all
satellite, after tests in the Intespaces anechoic test chamber, Toulouse, external influences to assess their electromagnetic compatibility (ESA)
2.1 General Background 9

b Fig. 2.2 a Launch of the European orbital test satellite (OTS-2, in

1978) on a Thor Delta rocket at Cape Canaveral. The 2 TV channels
and 5000 telephone circuits operated without defects between 52
ground-stations (between Norway and Egypt) (ESA). b Launch of
Alphasaton the 25th July 2013, an Ariane 5 lifted off Europes
largest telecommunications satellite (ESA)

The general development plan for a new satellite type

such as OTS involves the building of several test models
such as a structural model, thermal model, and engineering
model (refurbished from the thermal model), before con-
structing the qualication model and nally a flight
spacecraft.
Alphasat, shown in Fig. 2.1b is a high-power telecom
satellite built by Astrium, through a publicprivate partner-
ship between ESA and UK operator Inmarsat. It is based on
the mighty Alphabus, the new European telecom platform
developed by Astrium and Thales Alenia Space under joint
contract from ESA and the French space agency, CNES.
Alphabus is Europes response to increased market pressure
for larger telecom payloads for direct-to-home TV broad-
casting, digital audio broadcasting, broadband access and
mobile services. Alphabus incorporates innovative tech-
nologies including:

electric propulsionto optimise the satellites mass in

favour of payload
modular payloadincluding an antenna module which
can be adapted for different missions
star trackersensure highly accurate attitude and orbit
control
lithium ion cell batteriescharged from high-perfor-
mance solar cells

Whereas OTS generated 1260 W from its pair of solar

panels, feeding to two 24 Ah NiCd batteries and had a weight
of only 1490 kg, Alphasat can accommodate missions with
up to 18 kW of payload power and has a weight of 6000 kg.
Organisations such as the EU, ESA and NASA use mea-
sures to assess the maturity of evolving technologies which
can be related to devices, materials, components, etc.
Regarding materials, mechanical parts and manufacturing
processes, a new breakthrough or invention will not be suit-
able for immediate use and some basic research will have to be
conducted. This may lead to the technology being assessed for
feasibility, for development and later the technology may be
demonstrated in a laboratory environment. Validation of the
new material may be made according to certain test methods.
Mechanical parts may be qualied and manufacturing pro-
cesses may be veried by the testing of technology
10 2 Requirements for Spacecraft Materials

samplesthese are the steps usually taken in order to get projects. At the other end of the spectrum, manned flight
approvals for space use by authorities (ECSS-Q-STD-70 safety management will differentiate between systems
2014). As a guide, the following listing can be used to assess safety and payload safety. Space systems safety will be a
the level of readiness of any materials technology: trade-off between complex project elements using flight
proven technologieshere astronaut safety must be of
paramount importance. Payload safety will consider the
Technology Readiness Description materials, mechanical parts and manufacturing processes and
Level whether the payload is essential for flight operations and crew
TRL 1 Basic principles observed and safety. Payloads and experiments can fail and not cause a risk
reported to the astronauts. However, the materials from which they are
TRL 2 Technology concept and/or manufactured will be of particular concern as these may
application formulated operate beyond their intended temperatures; it is essential that
TRL 3 Analytical and experimental these, usually non-metallic materials, do not release toxic
critical function and/or charac- substances by off-gassing, nor any re hazard because of the
teristic proof-of-concept flammable nature of the piece-part.
TRL 4 Component and/or breadboard
validation in laboratory
environment 2.2 Considerations for Materials
TRL 5 Component and/or breadboard and Processes
validation in relevant
environment 2.2.1 General Considerations During
TRL 6 System/subsystem model or the Selection of Materials
prototype demonstration in a and Processes
relevant environment (ground or
space) The change of emphasis in Europe from building scientic
TRL 7 System prototype demonstration satellites during the 1970s with designed mission lives of
in a space environment one or two years to the production of a new generation of
TRL 8 Actual system completed and application satellites, which must be assured for periods of
Flight qualied through test greater than twenty years in a somewhat hostile space
and demonstration (ground or environment, has necessitated that a greater effort is placed
space) on conrming the reliability of many materials and tech-
TRL 9 Actual system Flight proven nologies which have previously been accepted as virtually
through successful mission fault-free. Additionally, the new modular approach and the
operations drive to standardize subsystems for easy and economical
adaptation for different satellite missions has led to long
The reader may consider the above Technology Readiness ground storage periods. This can cause material degradation
Levels (TRLs) during the selection of materials and processes problems, particularly the decay of liquid and solid fuels and
for a new application intended for use on board a spacecraft the general corrosion of sensitive surfaces and even stress
or even during the construction of a ground station (launch corrosion of structural elements. A listing of materials
site). Obviously for any technology: the lower the TRL the approved and utilized for the fabrication of a satellite such as
more time and effort will be required before the approving the aforementioned Orbital Test Satellite in 1975 included at
authority can give authorisation for its incorporation into a least 500 different organic and inorganic materials. Each was
space system. The concept of TRLs will not be addressed preliminarily approved for use in a given application, bear-
during the following chapters of this book as every approval ing in mind the environmental conditions it has been
of a space material, mechanism and process will depend on designed to withstand. The Declared Materials Lists asso-
the very precise requirements of a given space project. For ciated with multipurpose space platforms for large
some projects there may be an accepted higher level of risk telecommunications payloads, such as the 6.6 t Alphasat
involved during the selection of technologies. Low budget launched in 2013, involve more than 1000 different mate-
space flight experiments, providing they do not constitute a rials. Until the late 1980s metallic materials have been the
risk to the overall project, might choose to fly breadboard basic building materials of all satellites and launch vehicles
models that can give sufcient data return to university with only a limited number of inroads from carbon bre
2.2 Considerations for Materials and Processes
reinforced plastics (CFRP). Because of their exact alignment
requirements some solar panels, dish antennas, and antenna
platforms are fabricated from CFRP which, because of its
small coefcient of expansion, will retain dimensional
accuracy under the changing temperature conditions of an
orbit (160 to +180 C). Launch vehicles, satellites, space
probes and manned modules are predominantly built by
industrial concerns engaged in aircraft manufacture (e.g.
prime contractors such as Boeing, Airbus, Lockheed
Martin, Alenia, Aerospatiale and Astrium). Because of this,
designers will prefer to choose structural and mechanical
parts from traditional metal alloys and composites, and will
limit manufacturing to joining and nishing technologies
which already exist in their respective plants. When com-
pared to a mass production industry there is often little
incentive to promote the use of advanced materials and
alloys which may improve reliability and be weight-saving
but will suffer the drawback of requiring costly fundamental
testing and qualication before being incorporated into space
hardware.
Discussions between customers, prime contractors and
their sub-tier suppliers involve contract requirement negoti-
ations related to Materials and Processes issues and a con-
siderable number of reviews will take into account such topics
as design, materials selection, and fabrication processes.
These are held throughout the various stages of every space
project, from inception on the drawing board until the envi-
ronmental testing and qualication of manufactured hardware
prior to launch. However, it is not until the actual hardware is
seen that one is struck by the results of cooperation between
the many engineering disciplines. It is probable that the
introduction of computer-aided design now means that
spacecraft subassemblies and piece-parts are being fabricated
to the closest tolerances ever achieved. The optimization of
structural weight and the smaller design margins mean that a
thorough knowledge of the materials selected for the appli-
cation must be well established. This is particularly true for
new, advanced materials, as the small design margins means
there is no longer a reserve of strength built into the structure,
as was the case for earlier spacecraft, to cover ignorance of
design loads or stress intensities. The safety margins required
of materials are real, but the over-conservative designs orig-
inating from so-called gloom factors or scatter in materials
properties should be a thing of the past.
To illustrate the accuracy demanded of modern machin-
ing capabilities one can consider the unfortunate situation of
the flawed primary mirror of the Hubble Space Telescope
(HST). The prime objective of the HST mission was to
obtain images of astronomical objects in approximately ten
times sharper detail than that obtained by ground-based
telescopes. The HST 2.4 m mirror was designed to be a
precisely calculated hyperboloid. Although the mirror is
actually smooth to a precision of 1/64 the wavelength of
11
light (or one-millionth of an inch), a calculation error caused
the mirror which was originally launched to have been
fabricated with a curvature that was too shallow with a total
centre-to-edge error of about 2 m (or 1/50 thickness of a
human hair). The result was that light rays hitting the mirror
edges eventually made focus to a point that was slightly
away from where light rays from the centre of the mirror
focused: a defect called spherical aberration. The HST,
delayed three years by the Challenger disaster, was launched
in April 1990. Despite the flawed mirror, which rendered
many of Hubbles initial observations fuzzy, the new
spaceborne telescope quickly demonstrated the advantages
of an orbiting platform free from the interference of the
Earths atmosphere. After the dramatic December 1993
repair mission, using astronauts from Space Shuttle
Endeavour to correct the mirror and solar panel (see Sect. 8.
2) problems, Hubble began to demonstrate its full potential
to peer into the universe.
2.2.2 Some Futuristic Ideas
Advanced materials are nding more and more applications
in new designs, and this is particularly true of reinforced
polymers based on carbon or Kevlar bres, clean materials
(with low outgassing), and several new types of lightweight
metal alloys. The microminiature electronic circuits so
important for the relay of enormous volumes of data within a
fraction of a second are also incorporating new materials
with unique physical characteristics. Microdevices continue
to be designed and prototyped (David 1996)today these
are termed micro-electro-mechanical systems (MEMS).
Although many MEMS devices have been manufactured, to
date, the only devices that have flown are accelerometers and
gyroscopes (de Rooij 2009). In the USA, the JPL Centre for
Space Microelectronics Technology has already produced a
micro seismometer having a diameter of 12 mm and a
camera on a chip about the size of a ngernail. These kinds
of advancement will certainly lead to smaller, lighter, and
less costly spacecraft for the future. Even the so-called nano
satellite, weighing about 12 kg, is thought to be feasible
due to breakthroughs in small-scale engineering of MEMS.
The cost of launching a satellite into LEO by the Space
Shuttle was about 14,000 per kilogram and now, 5000 to
12,000 per kilogram when an ELV is selected. Costs to
place a spacecraft into a geosynchronous transfer orbit
(GTO) are estimated to be between and 20,000 per kilo-
gram. Either launch vehicles should become less expensive,
or satellites need redesigning to become far smaller and
lighter so that multiple payloads (or even nanosats) can be
launched simultaneously. A proliferation in the number of
miniaturised satellites (often referred to as CubeSats) have
been built by companies such as Clyde Space and SSTL
12
(now part of Airbus) but probably the majority of those
presently in orbit originate from schools and universities.
These have low construction costs combined with fewer
materials and processes requirements. Many have applica-
tions beyond those of academic research or technology
demonstration, and are used for Earth observation and
defense purposes.
The future will see more advanced manufacturing pro-
cesses involved with the construction of space hardware
even traditional methods such as casting and forging will
probably be to closer specication and under more highly
inert atmospheres. The autoclave curing of composites will
be done under clean conditions without the use of low
volatile organic materials and any mold release agents will
not contain silicones as they are difcult to remove prior to
painting. Friction stir welding and FricRiveting can be
envisaged for joining metals to thermoplastics; and laser
materials processing will involve localized, intense heating
of solid targets and components by laser, to achieve ultrafast,
novel and economic joining and surface engineering.
It used to be impossible to select a material without a full
knowledge of how it might best be processed into a nal
piece-partbut today it seems that 3D printing has opened
up a world of endless possibilities for designing and creating
everything from a complex space mechanism to printed
chocolates and foods for astronauts! 3D printing, also known
as additive manufacturing (AM), produces three-
dimensional items by printing them, layer-by-layer from
raw ingredients consisting of powdered plastics, aluminium
alloys, titanium alloys, low expansion alloys and other
spacecraft materials. The most usual processing method is to
introduce metal powder into a laser beama precise depo-
sition of either sintered or melted powder is directed onto a
flat table. The laser beam is controlled by CAD programs to
raster across the sintered metal powder layer so consolidat-
ing the deposit before another layer is added. The item is
then built up, layer-by-layer to create a net-shaped part. This
revolutionary rapid protyping process can now be used to
create nalized spacecraft parts with a 40 % weight saving.
Some improvements in the chemical purity of the powders
used should increase yields to 100 % and, already, a 3D
printer is in operation on the International Space station.
This rst 3D printing in space was performed in December
2014 by Butch Wilmore as part of a Zero gravity demon-
strationengineers up-linked a custom-made digital design
le of a ratchet wrench to the 3D printer and produced a tool
measuring 11.4 cm in length. This process will enable fragile
items to be manufactured in space without the need to
incorporate robustness and extra weight for surviving the
shocks, vibration and mechanical loads encountered during
launch. The deposition of materials can be so-called, func-
tionally graded, permitting one face of the deposit to have
totally different properties when compared to the opposite
2 Requirements for Spacecraft Materials
face. Similarly, custom compositions can be 3D printed so
that undesirable compositions (possibly brittle intermetallic
compounds, magnetic phases or corrodible compounds) in
any binary or ternary phase diagram can be avoided by only
depositing the useful compositions. The waste-, weight- and
money-saving attributes of AM have already attracted
manufacturers in all elds of advanced technology to
incorporate AM into the production lines for their cus-
tomized parts. These include heat-resistant bosses for turbine
cases, large bearing housings, rocket engine injectors,
landing gear support struts, and numerous spare parts.
Creative endeavors, like that noted in Fig. 2.3, may even
enable additive layer manufacturing (ALM) to print a hab-
itable structure on the Moon (Redahan 2014). Figure 2.3a
illustrates how functional habitation modules could be
brought from Earth; the surface of the thin-walled inflatable
structures would then be coated by 3-D printing a powder
made entirely of regolith, having a particle size of around
200 m, onto the thin-wall until a sufciently large wall
thickness could be built up to protect human space-workers
from radiation and micrometeoroids. Regolith is the name
given to lunar dust and this local resource has already been
encountered by humans and analyzed with respect to particle
size and composition (see Fig. 2.3b, c). Quantitative optical
and electron-probe studies by the UK Institute of Geological
Sciences (Simpson 1970) have shown that lunar samples can
contain ilmenite, pyroxene, chrome-titanium spinel, troilite,
native iron, iron-nickel alloy, and even native copper (as
shown in Fig. 8.1). This concept for constructing a human
outpost on the Moon using lunar soil, and ways to monitor
the buildings progress from Earth by means of an industrial
CCD camera positioned on the printer, have been described
by Ceccanti (2010) and Colla (2014).
Another rapid prototyping process that demonstrates
great promise has been described by Maxwell et al. (2013).
Known as Hyperbaric Pressure Laser Chemical Vapour
Deposition (HP-LCVD), this rapid prototyping process
incorporates a mixture of reactive gases into which laser
beams penetrate for the growth of materials from atomic
level to large structures by means of thermally- or
photolytically-induced decomposition of the gaseous pre-
cursor. Exploitation of this HP-LCVD process, as prescribed
by Dynetics and the NASA Marshall Space Centre (Maxwell
et al. 2013), may enable 3D rapid prototyping in-space for
the fabrication of components, replacement parts and even
nuclear thermal propulsion systemsby the use of precursor
gases and raw materials found, often in abundance, within
our own Solar System.
Numerous advanced materials and manufacturing tech-
niques will be individually described in Chap. 4.
2.2 Considerations for Materials and Processes
Fig. 2.3 a An artists impression of an igloo, built on the Moon by
means of a 3D printer attached to the arm of the robotic vehicle seen to
the right. Printing powder material is Moon dust (regolith), processed
into a cell-like structure of high strengththe idea is to initially inflate
a thin folded dome brought from Earth and protect it with a cellular
shell using the 3D printerthe pressurised enclosure so sheilds
astronauts from solar radiation, micrometerites and severe thermal
13
cycling. Concept and illustration courtesy of ESA and architects Foster
+ Partners. b Apollo 11 astronauts footprint in lunar soil, made up of
small, dust-like particles of regolith (courtesy NASA photo
AS11-40-5877). c Grain size distribution of lunar soil from three
different sites; about 50 % is greater than 100 m from Heiken et al.
(1974)
14 2 Requirements for Spacecraft Materials

Table 2.1 Static corrosion potential of metals and alloys (de Rooij 1989a)
Material EMF Potential
The metals having the greater negative EMF will tend to corrode and form EMF between a calomel electrode and a 3.5 % NaCl water
oxides solution (V)
Platinum +0.17
Carbon +0.15
Gold +0.15
Rhenium +0.08
Rhodium +0.05
Tantalum +0.04
Silver 0.03
Ag10Cu braze alloy 0.06
A286 (15Cr, 25Ni, Mo, Ti, V) passive 0.07
AISI 316 (18Cr, 13Ni, 2Mo, rem Fe) passive 0.07
AISI 321 (18Cr, 10Ni, 0.4Ti) passive 0.08
AISI 347 (18Cr, 12Ni, +Nb, rem Fe) passive 0.08
AISI 301 (17Cr, 7Ni) passive 0.09
AISI 304 (19Cr, 10Ni, rem Fe) passive 0.10
Hastelloy C (17Mo, 15Cr, 5W,6Fe, rem Ni) passive 0.10
Nichrome (80Ni, 20Cr) passive 0.10
Monel 60 (65Ni, 0.2Fe, 3.5Mn, 2Ti, 27Cu) 0.10
Inconel 92 (71Ni, 16Cr, 7Fe, 3Ti, 2Mn) passive 0.11
17-7PH stainless st. (17Cr, 7Ni, 1.1Al) passive 0.11
AISI 309 (23Cr, 13Ni) passive 0.11
Titanium 0.12
Monel 400 (32Cu, 2.5Fe, 2Mn, rem Ni) 0.12
CDA 442 (71Cu, 1Sn, 38Zn) 0.12
CD A 715 (70Cu, 30Ni) 0.12
Molybdenum 0.12
MP35N (Ni, 35Co, 2.0Cr, 10Mo) passive 0.15
CDA 510 (96Cu, 4Sn, P) phosphor bronze 0.16
AISI 420 (0.35C, 13Cr, rem Fe) passive 0.17
AISI 434 (0.12C, 17Cr, 1Mo, rem Fe) passive 0.17
Bismuth 0.17
Waspaloy (59Ni, 19.5Cr, 13.5Co, 4Mo) passive 0.17
Nickel passive 0.18
Monel 67 (67.5Cu, 31Ni, 0.3Ti, 0.5Fe) 0.18
Copper phosphorus (4.5P, rem Cu) 0.18
Copper phosphorus (8.5P, rem Cu) 0.19
Copper phosphorus (10.5P, rem Cu) 0.20
Copper 0.20
CDA 110 (electrolytic tough pitch) 0.20
CDA 172 (2Be, rem Cu) 0.20
Gold-germanium solder (12Ge, rem Au) 0.20
Copper-Gold (25Au, rem Cu) 0.20
AISI 440B (17Cr, 0.5Mo, rem Fe) passive 0.23
(continued)
2.2 Considerations for Materials and Processes 15

Table 2.1 (continued)

Material EMF Potential
The metals having the greater negative EMF will tend to corrode and form EMF between a calomel electrode and a 3.5 % NaCl water
oxides solution (V)
Ti6A14 V (6A1.4 V, rem Ti) 0.24
Silicon 0.24
Tungsten carbide (94WC, 6Co) 0.25
CDA 240 (80Cu, 20Zn) 0.25
CDA 220 (90Cu, 10Zn) 0.25
CDA 752 (65Cu, 18Ni, 17Zn) 0.25
CDA 180 (60Cu, 40Zn) 0.26
CDA 464 (60Cu, 1Sn, 39Zn) 0.26
CDA 270 (63Cu, 37Zn) 0.26
CDA 298 (52Cu, 48Zn) 0.27
Nichrome 80/20 (80Ni, 20Cr) active 0.27
CDA 521 (7Sn, rem Cu) 0.27
CuA110Fe (10A1.3Fe, rem Cu) 0.27
Armco 21-6-0 (22Cr, 12Ni, rem Fe) 0.27
Inconel 92 (71Ni,16Cr, 7Fe, 3Ti, 2Mn) active 0.28
CuA112 (12Al, rem Cu) 0.29
Niobium (1Zr, rem Nb) 0.30
Tungsten 0.30
Nickel active 0.30
Kovar, Nilo K (29Ni, 17Co, rem Fe) 0.30
Chromium active 0.31
Cobalt 0.32
Nitinol (45Ti, 55Ni) 0.33
Invar (36Ni, rem Fe) 0.38
Cerrotric (42Sn, rem Bi) 0.39
SnAg4C3.5 solder 0.42
Sn95Ag4.9In0.1 solder 0.43
SnAg4 solder 0.46
SnAg5 solder 0.46
Tin 0.46
Sn10Sb solder 0.48
Indalloy no. 10 (75Pb.251n) solder 0.48
Indalloy no. 7 (50Pb,501n) solder 0.49
Lead 0.50
Sn63 (63Sn, 37Pb) solder 0.51
Sn60 (60Sn, 40Pb) solder 0.51
Sn62Ag2 (62Sn, 36Pb, 2Ag) solder 0.51
Sn59Sb2 (59Sn, 39Pb, 2Sb) solder 0.51
Sn60Sb5 (60Sn, 35Pb, 5Sb) solder 0.51
Sn60Sb10 (60Sn, 30Pb, 10Sb) solder 0.51
Sn60Pb39.5Cu0.12P0.9 solder 0.51
PbSn5Agl.5 solder 0.51
Mild steel 0.52
(continued)
16 2 Requirements for Spacecraft Materials

Table 2.1 (continued)

Material EMF Potential
The metals having the greater negative EMF will tend to corrode and form EMF between a calomel electrode and a 3.5 % NaCl water
oxides solution (V)
AISI 304 (19Cr, 10Ni, rem Fe) active 0.52
AISI 420 (0.35C, 13Cr, rem Fe) active 0.52
AA 2219-T3.T4 (6.3Cu, 0.3Mn, 0.18Zr, 0.1V, 0.06Ti, rem Al) 0.56
AISI 440B (17Cr, 0.5Mo, rem Fe) active 0.59
AA 2014-T4 (4.5Cu, 1Fe, 1Si, 0.15Ti, rem Al) 0.61
AA 2017-T4 (4Cu,1Fe,1 Mg,0.1Cr,rem Al) 0.61
AA 2024-T3 (4.5Cu, 1.5 Mg, 0.6Mn, rem Al) 0.62
AA B295.0-T6 (2.5Si, 1.2Fe, 4.5Cu, rem Al) casting 0.63
In75Pb25 solder 0.64
Indalloy No. 1 (50In, 50Sn) solder 0.65
AA 380.0-F (8.5Si, 2Fe, 3.5Cu, rem Al) casting 0.66
AA 319.0-F (6Si, 1Fe, 3.5Cu, rem Al) casting 0.66
AA 333-0-F (9Si.1Fe, 3.5Cu, rem Al) casting 0.66
Indium 0.67
AA 2014-T6 (4.5Cu, 1Fe, 1Si, 0.15Ti, rem Al) 0.69
Cadmium 0.70
AA 2024-T81 (4.5Cu, 1.5Mg, 0.6Mn, rem Al) 0.71
AA 2219-T6,T8 (6.3Cu, 0.3Mn0.18Zr, 0.1V, 0.06Ti, rem Al) 0.72
AA 6061-T4 (1 Mg, 0.6Si, 0.25Cu, 0.2Cr, rem Al) 0.72
AA 4043 (12Si, 1Cu, 1Mg, rem Al) 0.74
AA 6151 (1 Mg, 1Fe, 0.25Sn, 0.15Ti, rem Al) 0.74
AA 7075-T6 (5.6Zn, 2.5Mg, 1.6Cu, 0.3Cru.03Cr, rem Al) 0.74
AA 7178-T6 (6.8Zn, 32Mg, 2Cu, 0.2Ti, rem Al) 0.74
AA 1160 (98.4Al) 0.75
Aluminium 0.75
AA 5356 (5Zn, 0.1Ti, 0.1Cr, rem Al) 0.75
AA 5554 (5 Mg, 1Mn, 0.25Zn, 0.2Cr, rem Al) 0.75
AA 1050 (99.5Al) 0.75
Al-3Li 0.75
AA 1100 (99.0Al) 0.75
AA 3003 (1.2Mn, rem Al) 0.75
AA 6151 (1Mg, 1Fe,0.8Mn, 0.25Zn, 0.15Ti, rem Al) 0.75
AA 6053 (1.3Mg, 0.5Si, 0.35Cr, rem Al) 0.75
AA 6061-T6 (1Mg, 0.6Si, 0.25Cu, 0.2Cr, rem Al) 0.75
AA 6063 (0.7Mg, 0.4Si, rem Al) 0.75
Alclad 2014 (4.5Cu, 1Fe, 1Si, 0.15Ti, rem Al) 0.75
Alcald 2024 (4.5Cu, 1.5Mg, 0.1Cr, rem Al, Al-clad) 0.75
AA 3004 (1.5Mn, rem Al) 0.76
AA 1060 (99.6Al) 0.76
AA 5050 (1.5Mg, rem Al) 0.76
AA 7075-T73 (5.6Zn,2.5 Mg,1.6Cu,0.3Cr,rem Al) 0.76
AA 5052 (2.5Mg, 0.25Cr, rem Al) 0.77
AA 5086 (4Mg, 0.5Mn, rem Al) 0.77
(continued)
2.2 Considerations for Materials and Processes 17

Table 2.1 (continued)

Material EMF Potential
The metals having the greater negative EMF will tend to corrode and form EMF between a calomel electrode and a 3.5 % NaCl water
oxides solution (V)
AA 5154 (3.Mg, 0.25Cr, rem Al) 0.78
AA 5454 (2.8Mg, 1Mn, 0.2Ti, 0.1Cu, 0.2Cr, rem Al) 0.78
AA 4047 (12Si, rem Al) 0.78
Al-C 0.78
AA 5056 (5.2Mg, 0.1Mn, 0.1Cr, rem Al) 0.79
AA 7079T6 (4.3Zn, 3.3Mg, 0.6Cu, 0.2Mn, 0.2Cr, rem Al) 0.79
AA 5456 (5Mg, 0.7Mn, 0.15Cu, 0.15Cr, rem Al) 0.79
AA 5083 (4.5Mg, 0.7Mn, rem Al) 0.79
AA 7072 (1Zn, 0.5Si, 0.3Cr, rem Al) 0.87
Beryllium 0.97
Zinc 1.03
Manganese 1.21
Erbium 1.34
Electron (4Zn, 0.7Zr, rem Mg) 1.55
ZW3 (3Zn, 0.5Zr, rem Mg) 1.57
AZ61 (6Al, 1Zn, 0.3Mn, rem Mg) 1.57
AZ31B (3Al, 1Zn, rem Mg) 1.60
Magnesium 1.60
HK31A (0.7Zr, 3Th, rem Mg) 1.61
Notes to Table 2.1
Compatible material couples are considered to have a maximum potential difference of
0.25 V for non-cleanroom environments
0.50 V for cleanroom or hermetically sealed environments
This galvanic series chart is useful for a rst approximation in selecting materials for corrosion control, but may be too simplistic for further
dependence. It provides no information concerning corrosion rates or what will happen when three or more metals are electrically coupled. Service
conditions such as ionic concentration, aeration, metal purity, etc. can change relative positions. Reversals, especially with metals that are very
close in the series, such as steel and aluminium, can cause serious service problems, and specialized polarization studies are then recommended
The majority of alloys present in this Table can be referred to in Appendix 6 which lists specication number, composition limits and equivalent
British, French, German and US standards

2.2.3 Some Basic Considerations Regarding
Corrosion Prevention
It is necessary to ensure that any newly selected material will
retain its functional properties during all stages of the
spacecrafts designed life, up to the end of the mission.
During manufacturing, the material must not degrade
because of contamination from processing steps such as the
release agents used for items moulded from CFRP, or by
cutting oils used in the machining of alloys. Galvanic and
general surface corrosion must be avoided during environ-
mental testing and ground storage by the correct selection of
surface nishes such as anodic lms, chemical conversion
lms, and paints. When electrical grounding is required,
only contacts having a compatible coupling of less than
0.5 V should be chosen. The static corrosion potential for a
large number of metals and alloys has been established (de
Rooij 1989a) and is presented in Table 2.1. However, de
Rooij has simplied this Table into Groupings of metallic
alloys and the modied tabulation now appears as shown in
Table 2.2. The material may need a high resistance to Stress
Corrosion Cracking (SCC) before launch, and in such cases
can be selected from those alloys listed in Table 2.3.
The primary and secondary structures will be made from
light alloys based on aluminium and magnesium, together
with titanium and to a very limited extent beryllium. Nickel
alloys are often selected for their high-temperature perfor-
mance and oxidation resistance; they are often known by
trade names, rather than by their specication code numbers.
Commercially pure nickel, easy to form into complex
shapes, is used in the construction of spacecraft electronics
where its electrical and magnetic properties are crucial.
Mechanical designers often select Inconel alloys 600 and
625 because they appear in Table 2.3, but it has recently
18 2 Requirements for Spacecraft Materials

Table 2.2 Suggested compatible couples for bimetallic contacts (after de Rooij, based on Table 2.1)

Key
0Can be used without restriction
1Can be used in a non-controlled environment (e.g. assembly area and general non-clean room environmnet)
2Can be used in a clean room environment
3Need specic measures to avoid corrosion when these combinations are selected
2.2 Considerations for Materials and Processes 19

Table 2.3 Alloys with high resistance to stress-corrosion cracking Table 2.3 (continued)
Steel alloys Aluminium alloys
Alloy Condition 7050 T73 A712.0, C712.0 As cast
Carbon steel (1000 series) Below 180 ksi 7475 T73
UTS Copper alloys
Low alloy steel (4130, 4340, D6AC, etc.) Below 180 ksi
CDA No.h Condition (% cold rolled)i
UTS
110 37
Music wire (ASTM 228) Cold drawn
170 AT, HT.j
1095 spring steel Tempered
172 AT, HT.j
HY 80 steel Tempered
194 37
HY 130 steel Tempered
195 90
HY 140 steel Tempered
230 40
200 series stainless steel (unsensitized) All
280 0
300 series stainless steel (unsensitized)a All
422 37
400 series ferritic stainless steel (404, 430, All
444, etc.) 443 10
Nitronic 32 Annealed 510 37
b 521 37
Nitronic 33 Annealed
b 524 0
Nitronic 40 (formerly 21-6-9) Annealed
A-286 stainless steel All 606 0
AM-350 stainless steel SCT 1000 and 619 40 (9 % B phase)
above 619 40 (95 % B phase)
AM-350 stainless steel SCT 1000 and 655 0
above
688 40
AM-362 (Almar 362) stainless steel 3 h. at 1000 F
704 0
Carpenter 20Cb-3 stainless steel All
706 50
Carpenter 20Cb-3 stainless steel All
710 0
Custom 450 stainless steel H1000 and above
715 0
Custom 455 stainless steel H1000 and above
725 50 Annealed
15-5PH stainless steel H1000 and above
Nickel alloys
PH15-7Mo stainless steel CH900
Alloy Conditions
17-7PH stainless steel CH900
Glass Seal 52 CR (51Ni-49Fe) All
Aluminium alloys
Invar 36 (36Ni-64Fe) All
Wrought Cast
Hastelloy B Solution heat treated
Alloyc Temperd Alloye Temper
Hastelloy C All
1000 series All 319.0, A319.0 As cast
Hastelloy X All
2011 T8 333.0, A333.0 As cast
Incoloy 800 All
2024 rod, bar T8 355.0, C355.0 T6
Incoloy 901 All
2219 T6, T8 356.0, A356.0 All
Incoloy 903 All
2418 T8
Inconel 600 Annealed
2618 T6 357.0 All
Inconel 625 Annealed
3000 series All B358.0 (Tens-50) All
Inconel 718 All
5000 series Allf, g
359.0 All
Inconel X-750 All
6000 series All 380.0, A380.0 As cast
Monel K-500 All
7049 T73 514.0 (214) As castg
Ni-Span-C 902 All
7149 T73 518.0 (218) As castg
Rene 41 All
7050 T73 535.0 (Almag.35) As castg
(continued)
(continued)
20 2 Requirements for Spacecraft Materials

Table 2.3 (continued) All the classical assembly methods are employed: weld-
Nickel alloys ing, brazing, soldering, riveting, bolting, and adhesive
Unitemp 212 All bonding. It is important to ensure the joining processes
themselves have not degraded the materials surface or stress
Waspaloy All
corrosion resistance. (Heating can modify an alloys
Miscellaneous alloys
microstructure, weld metal and heat affected zones will be
Alloy Conditions different to the parent metal, braze metal may be noble to the
Beryllium S-200C Annealed remaining surfaces which can preferentially corrode,
HS25 (L605) All mechanical joints can have re-entrant faces that retain water
HS 188 All and cause pitting, even cured resins may release acids that
MP35 N Cold worked and aged damage the surrounding surfaces). Generally, aircraft
MP159 Cold worked and aged
industry manufacturing standards are followed, and much
attention is given to process control and there is a need to
Titanium 3Al-2.5 V All
evaluate all process used to join together structural and
Titanium 5Al-2.5SN All
electrical parts.
Titanium 6Al-4 V All
Titanium 10Fe-2 V-3Al All
Titanium 13V-11Cr-3Al All 2.2.4 Space Projects Phases
Titanium IMI 550 Al and Management Events
Magnesium MIA All
Magnesium LA141 Stabilized
It is important to note that before a satellite becomes fully
operational in orbit its subsystems, mechanisms, and elec-
Magnesium LAZ933 All
a
tronics will have been subjected to the following main
Including weldments of 304L, 316L, 321 and 347
b environmental conditions:
Including weldments
c
Including weldments of the weldable alloys
d
Mechanically stress relieved (TX5X or TX5XX) where specied (A) Ground activities:
e
The former designation is shown in parentheses where signicantly Operation for test and checkout
different. See Appendix 5 for temper designations
f Handling
High magnesium alloys 5456, 5083, and 5086 should be used in
controlled tempers (H111, H112, H116, H117, H323, H343) for Transportation
resistance to SCC and exfoliation Storage
g
Alloys with magnesium content greater than 3.0 percent are not Exposure to the elements
recommended for high-temperature application, 66C (150F) and
(B) Subjection to launch and ascent:
above
h
Copper Development Association alloy number Acceleration and shock
i
Maximum percent cold rolled for which SCC data are available Vibration and acoustic noise and possibly contact with
j
ATannealed and precipitation hardened reactive fuels, oxidizers, and temperature extremes
HTwork hardened and precipitation hardened
Pyrotechnic shock
Notes to Table 2.3
Data are compiled from NASA MSFC Spec 522B and (C) Transfer to operational orbit position:
ECSS-Q-ST-70-36. Recent issues of these documents should be Thermal cycling due to exposure to the Sun and eclipse
consulted for classication of alloys with both a moderate and a low Ultrahigh vacuum
resistance to stress corrosion cracking. Appendix 5 describes the
Radiationelectromagnetic and penetrating particles
aluminium alloy temper designations. A search through Appendix 6
will indicate similar European alloys Zero gravity

been acknowledged that these alloys soften, and can suffer
from SCC in pure water, at temperatures above 300 C
(steam generators for nuclear plants). It may be wiser to use
a development of the 625 alloy when high temperatures may
be expected, for instance in propulsion systems. This
development involved the additions of molybdenum and
niobium to 625 to impart solid solution hardening and the
formation of Ni3Nb, a very effective hardening precipitate.
This is known as the super alloy Inconel 718 and has
become the most widely used high temperature nickel alloy.
The new millennium saw a great increase in the size and
complexity of spacecraft. This has necessitated ground test-
ing facilities to become modernized, physically larger, and
more sophisticated for the exposure of space hardware to the
environments listed in A to C above. The Test Facility within
space agencies are comprised of high capital investments
such as Large Space Simulators that reproduce the vacuum,
certain radiations and the cryogenic-to-elevated temperatures
encountered by space hardware. Mechanical and acoustic
tests simulate the launch environment, the magnetic
2.2 Considerations for Materials and Processes
characteristics of spacecraft are evaluated and all tests are
performed under various classes of cleanroom conditions.
As mentioned previously, a development plan for a new
spacecraft design will involve a model philosophy where
models of the craft will be dynamically testedwithout
testing the risk of failures is too great. In the early days it was
not uncommon for space authorities to build four models for
testing prior to actually building a flight spacecraft. The
model philosophy will be accounted for in the next para-
graph, but it is emphasised that as the space industry has
matured, the design margins have become established (for
structures and electronic systems) so that there is now more
focus on analysis and less on actual testing. It is now the
norm to build only one prototype for testingoften this
build and test is completed only six months before work
commences on manufacturing the flight spacecraft.
It is interesting to remember the aims of the model
philosophy for the early (70s and 80s) European
telecommunication satellite projects. The structural model
was subjected to a programme of tests which exceeded the
expected launch environment conditions (each type of
launch vehicle has its own characteristic levels of vibration
and acoustic noise). Typical test congurations are shown in
Fig. 2.1ac. The enormous amount of energy released during
launch can be witnessed from Fig. 2.2. Weakness in designs
may be exposed by this model, such as failures resulting
from fatigue of welds, struts, electronic box hold-down
points, and the like. The thermal model was subjected to
solar simulation and thermal balance testing to conrm and
update previously determined mathematical models. During
these tests, decient designs may promote several material
failures related to thermal fatigue, overheating, and embrit-
tlement of incompatible joining techniques and metal alloys.
By simply modifying the paint nish of the spacecraft sur-
face, or by the attachment of reflective mirrors, it was found
possible to adjust and reduce the local temperature envi-
ronment of each subsystem or equipment and reduce the
chance of thermally induced failures. Workmanship prob-
lems abounded, with non-conformances relating to open
circuits in cable harnesses due to wires separating from
crimp barrels and cold soldered joints on circuit boards (such
events became less frequent once operator and inspector
training schemes were introducedsee Sect. 6.14). The
engineering model ensured that integration and performance
could be achieved. A review of the old project and labora-
tory failure investigation reports was made by the author (at
the time of ling and archiving as a result of which they were
lost forever!). This revealed that several material problems
only came to light during integration, particularly at the
mechanical interfaces between equipments and the structure
(e.g. failure of springs and bolting devices due to incorrect
plating processes which cause delayed failure by hydrogen
embrittlement; the over-torqueing and fracture of lock-nuts
21
and other operator handling errors, etc.). It was seen that the
reliability of electrical interfaces between equipments, and
the mutual compatibility between the constituent subsys-
tems, was seriously jeopardized by fundamental oversights
(e.g. high electrical resistance between gold-plated and
aluminium-nished interfaces due to galvanic corrosion;
migration of silver to produce short circuits; the use of
austenitic steels having work-hardened and therefore slightly
magnetic surfaces in locations required to be magnetically
clean; etc.). Remedial actions were taken by suppliers and
assemblers as a result of failure review boards (FRBs).
Lessons learnt documents (nowadays called internal problem
notication documentsIPNs) were written and circulated.
Finally, a fully assembled qualication model was built and
subjected to a comprehensive series of environmental
ground tests. These usually included: sine vibration, spin,
acoustic noise vibration, centrifuge acceleration, and solar
simulation. Each of these major test phases was preceded by
an integrated systems test in order to verify that the func-
tional behaviour of the satellite was correct. It was not
uncommon that the qualication model would be recognized
as a flight spare in case of launcher failure.
Flight model test programmes were, and continue to be,
more limited than those used on the qualication model
(most customers now refer to this as the protoflight model).
It may be assumed that some subsystem parts that will
operate for relatively short periods after launch can accu-
mulate several hundred hours of test operation before the
actual time of launching. Materials selected for use under the
vacuum conditions of space may therefore have to operate for
periods under normal atmospheric conditions. This may
create special problems. One is immediately reminded that
very thin (tens of ngstroms1) lms of lead, or molybdenum
disulphide, for the lubrication mechanisms, can rapidly oxi-
dize under terrestrial conditions and become the cause of
malfunction. This goes to illustrate the need to know the
effects that ground testing may have on delicate surfaces.
A nal concern of the writer relates to the participation of
materials experts on spacecraft project review boards. It is
paramount that an experienced materials engineer is incor-
porated into each of the four major reviews during the design
and construction of individual spacecraft. Whenever ECSS
Q-ST-70 is included as a contractual document this become
a requirementas a minimum, the materials engineer will
manage the steps taken for the project-approval of declared
lists for every flight material, mechanical part and their
related processes (i.e. DML, DMPL and DPL). Tasks should
also include cleanliness and contamination control, the
testing and validation of new materials, assistance in the
1 equals 1010 m.
1
22
qualication of mechanical parts and the verication of new
or critical processes.
Spacecraft, such as CubeSats and university flight
experiments may follow less rigorous requirements and a
reduced M&P programme is suggested in Appendix 8.
Reviews are usually contractual milestones and essen-
tially question: will the design, hardware, software, and
operational approach satisfy the mission objectives? The
review names and their main objectives are as follows:
PDR (preliminary design review): after evaluation of thermal
and/or engineering models, to approve and release the
preliminary design, including materials and processes.
CDR (critical design review): this establishes the nal design
and agrees that flight hardware manufacturing can com-
mence (all declared materials being approved and pro-
duction process, when required, are veried as being
suitable).
QR (qualication review): assess that all qualication
activities on subsystems are completefor certain projects
a qualication model is built.
LRR (launch and operations readiness review): this checks
that all lower-level acceptance reviews have been suc-
cessfully completed, the flight model spacecraft accepted,
and it is then authorized to be launched.
2.3 The Effect of a Space Environment
(a) General
The purpose of this section is to provide the reader with an
overview of the salient points concerning the effect of a
space environment on spacecraft parts and materials. Far
greater discussion with examples will be made of these
effects throughout the remainder of this book.
Each spacecraft material will be required to suffer no
vibrational fatigue damage during the launch. In orbit, it
will need to survive the space environment (see
Table 2.4) and, in particular, to possess low-outgassing-
under-vacuum properties, whether it be a lubricating
grease or a structural plastic. Radiation and thermal
cycling must not degrade thermal-control surfaces or
joints in materials possessing different coefcient of
expansion. The presence of atomic oxygen in low Earth
orbits, a relative newcomer in environmental effects, has
been seen to lead to the erosion/corrosion/oxidation of
many material surfaces, and more coatings with good
atomic oxygen durability need to be developed. These
aspects will be detailed in Chap. 8.
2 Requirements for Spacecraft Materials
(b) Sublimation and evaporation
The minimum altitude for an Earth-orbiting satellite is
200 km (125 miles), and, once this altitude has been reached,
appreciable changes can be produced in common engineer-
ing materials, whether they be metals, plastics, or ceramics.
The vacuum in space is very high, the pressure falling from
106 mm Hg at 200 km to less than 1012 mm Hg beyond
6500 km. Some polymers will decompose and some metals
will tend to sublimate under vacuum. The rate at which the
molecules or atoms leave a surface in vacuum will rise
rapidly with an increase in temperature according to the
equation:
G 5:04  103 PM=T 1=2
where
G grams of material evaporated or sublimated per square
centimetre per day
P Vapour pressure of the evaporating species in torr2
M Molecular weight of the material
T Absolute temperature, K
Temperature has an enormous effect on the amount of
metallic material that is sublimated. As examples, cadmium
and zinc have sublimation rates that increase by a factor of
ten for roughly every 30 C rise in temperature:
Cadmium has a vapour pressure of 108 at 70 C
(approximately)
107 at 90 C
106 at 120 C
105 at 150 C and
104 at 180 C
The relationship between the vapour pressure of a metal
(P) and temperature (T) is given by the following equation:
P P/ eE=RT
where P = a constant (i.e. vapour pressure at T = ),
E = heat of evaporation (e.g. joule mole1), and R = gas
constant (8.3 J mol1 K1).
The temperatures for given metallic sublimation rates
are listed in Table 2.5. The thermal environment in space is
completely different from thermal conditions on Earth.
Without an atmosphere, the only means of exchanging
thermal energy is by thermal radiation and conduction.
Certain parts of satellites have been calculated to follow
2
The term torr is generally used instead of mm Hg by international
agreement of several vacuum societies. Torr honours the name of
Torricelli, who discovered atmospheric pressure in 1643.
2.3 The Effect of a Space Environment
Table 2.4 Characteristics of the space environmentorder of magnitude only (Dauphin 1984)
Altitude Pressure (mm Kinetic Gaseous density
(km) Hg or torr) temperature (particle cm1)
(K)
Sea 760 300 2.5 1019
level
30 10 4 1017
6
200 10 1200 1010
800 109 1300 106
6500 1013 103
22,000 1013 101102
thermal excursions from approximately 160 to +180 C.
The actual temperature attained will differ from one space-
craft to another, the major temperature effect arising from the
spacecrafts spin rate. Surfaces of non-spinning satellites
exposed to direct solar radiation may be unable to dissipate
thermal energy efciently, and will reach higher tempera-
tures than spinning satellites. Temperature variations will
also depend upon the amount of albedo radiation and the
amount of thermal radiation to space from the spacecraft.
Both active and passive thermal control systems are
employed on satellites in order to restrict the oscillating
temperature extremes. The active systems have made use of
thermostatically controlled heaters. Passive systems involve
the surface absorptance/emittance, /, properties of material
nishes. Solar reflectors have low / ratios, being generally
white paints or clear anodized aluminium. Black paints and
inorganic black anodized aluminium wave / values of
approximately 1. Solar absorbers have an / value greater
than 1, and these are generally polished metals since the
emittance values of uncoated metals are very low (<0.l).
Examples are described in Sect. 5.5.
The majority of metals do not sublimate at normal
spacecraft temperatures. However, as can be seen in
Table 2.5, cadmium and zinc must be excluded from use as
they will readily sublimate and could condense in the form
of thin conductive deposits on electrical insulators, or opa-
que deposits on optical components which may be situated
within the satellite or on its external surfaces. All cadmium,
zinc, or tin-plated surfaces, such as the protective nishes on
equipment or components including commercial connectors,
must be avoided, because, as will be described in Chap. 7,
they are known to grow single-crystal whiskers exceeding
2 cm length in vacuum. Extreme care must be taken to
ensure that these metals are not used in the corrosion
23
Composition Ultraviolet Particle radiation
radiation (particles cm2 s1)
78 % N2, Section of solar
21 % O2, 1 % A spectrum 0.3
N2, O2, A Absorption zone
N2, O, O2, O +
Full solar
spectrum
O, He, O+, H Full solar
spectrum
H+, H, He+ Full solar 104 protons 35 MeV
spectrum 104 electrons 40 keV
85 % H+, Full solar 108 protons 5 MeV
15 % He2+ spectrum
108 electrons 40 keV
104 electrons 1.6 MeV
proong of any of the spacecrafts components. Magnesium
parts could pose sublimation problems after long exposure to
vacuum at temperatures greater than 125 C, and experi-
ments have shown (Frankel 1969) that magnesium sheets
held at 230 C and 1 107 mm Hg for only 168 h (one
week) became severely pitted and dramatically decreased in
static strength properties. Because of their relatively high
strength-to-weight ratios, magnesium alloys are often
employed as structural parts, but it is essential that these
parts are nished with an adequate plating or chemical
conversion coating which will prevent both the corrosion of
the part before launch and the subsequent sublimation
problem in orbit. Tinlead alloys, such as those employed
for soldering electrical components, have not been seen to
sublimate under spacecraft environments as they are
restricted for use in areas which are thermally controlled to a
maximum of about 80 C. Solder alloys are used for joining
silver-plated molybdenum interconnector strips between
solar cells on the solar arrays of spinning satellites. As such
satellites rotate, the maximum temperature of the arrays does
not degrade the soldered joints. The stationary communica-
tion satellites will not be able to dissipate the absorbed
thermal radiation on the solar arrays efciently, so that
welded interconnectors are necessary.
The events of sublimation and evaporation cause the
release of metal atoms which travel and are capable of
recondensing on cooler surfaces. They are readily ionized
and may be contributors to corona and arcing phenomena.
These metallic ions may also cause the complete failure of a
satellite mission by recondensing between slip rings, causing
electrical short circuits, or recondensing on optical surfaces
causing the loss of a specic wavelength transmission. When
such ions recondense on the spacecrafts highly reflective
thermal control surfaces the thermal balance can be so
24 2 Requirements for Spacecraft Materials

Table 2.5 Sublimation of metals and semiconductors in high vacuum

Sublimation rate 1000 a/year 103 cm/year 101 cm/year
(0.0004 in/year) (0.04 in/year)
Temperature C C C
Cadmium 40 80 120
Selenium 50 80 120
Zinc 70 132 180
Magnesium 110 170 240
Tellurium 130 180 220
Lithium 150 210 280
Antimony 210 270 300
Bismuth 240 320 400
Lead 270 330 430
Indium 400 500 610
Manganese 450 540 650
Silver 480 590 700
Tin 550 660 800
Aluminium 550 680 810
Beryllium 620 700 840
Copper 630 760 900
Gold 660 800 950
Germanium 660 800 950
Chromium 750 870 1000
Iron 770 900 1050
Silicon 790 920 1080
Nickel 800 940 1090
Palladium 810 940 1100
Cobalt 820 960 1100
Titanium 920 1070 1250
Vanadium 1020 1180 1350
Rhodium 1140 1330 1540
Platinum 1160 1340 1560
Boron 1230 1420 1640
Zirconium 1280 1500 1740
Iridium 1300 1500 1740
Molybdenum 1380 1630 1900
Carbon 1530 1680 1880
Tantalum 1780 2050 2300
Rhenium 1820 2050 2300
Tungsten 1880 2150 2500
l = 1010 m
a

Based on data by Jaffe and Rittenhouse, California Institute of Technology

degraded that severe overheating will promote malfunctions.
Cesium is an uncommon metal with a melting point of 28 C
and a boiling point of 671 C but it has been proposed as a
propellant for eld effect electronic propulsion (FEEP) when
very low thrust applications are required. Thrusters have
been developed using this metal, but CNES evaluations have
determined that cesium can easily contaminate spacecraft
surfaces due to its low vapor pressure and re-evaporate from
surfaces warmer than 30 C to cause further contamination
and react chemically with polymers or some oxides to
2.3 The Effect of a Space Environment
change their optical properties (Tondu 2011). The
BepiColumbo mission dedicated to the study of Mercury is a
challenging ESA project where temperatures may reach
300 C at all external surfaces and metallic sublimation is of
major concern (as is the outgassing of organic materials).
A reference to Table 2.3 will immediately show that mate-
rials containing zinc, cadmium (always prohibited in space
applications) and lead are unsuitable candidates. Sublimation
and cold welding of solar array drive mechanisms will
require strict controls for the bearings, slip-rings and cables
(Fink et al. 2009a, b).
For additional examples of the effect of sublimation on
spacecraft hardware, the reader may refer to Sect. 5.6.
(c) Radiation and particle damage
Organic materials, such as those used for electrical insula-
tion, may be damaged by ionization due to protons and
electrons from radiation belts, solar emissions, and cosmic
rays. The Van Allen radiation belt is especially damaging to
organic materials and even inorganic materials which make
up optical lenses, ceramic insulators, and sensitive electronic
components. The metallic materials of most Earth-orbiting
spacecraft, as well as deep-space probes, are unaffected by
small particle radiations, and are only slightly degraded by
erosion from a cloud of meteoric dust which surrounds the
Earth and other planets. Larger particle damage was, how-
ever, a major problem for the Halleys Comet yby mission
called Giottowhich encountered that comet in 1986.
Special armour plating was developed to surround that
spacecraft, which was impacted by rock debris travelling at
hypervelocities of 1070 km/s. These aspects are reviewed
in Chap. 8.
(d) Friction and wear
One of the major material problem areas for advanced
spacecraft is that of friction and wear of surfaces which must
rub or slide over each other under conditions of temperature
cycling and high vacuum. This may be encountered in the
operation of hinges, gears, bearings, and electrical contacts
used in a vast number of spacecraft mechanisms. During
sliding under normal terrestrial conditions most contacting
metallic surfaces are protected by a surface lm of oxide,
oil, grease, or other contaminant which will act as a shear
layer and prevent binding. Under vacuum conditions such
contaminants outgas, and oxides, once disrupted or
removed, are unable to reform. Also, the minute junction
spots which carry the full load between contacting metallic
surfaces will usually have vastly increased friction coef-
cients, and probably high rates of wear, so that many
metallic couples will tend to cold-weld. It is usually
25
impracticable to enclose such moving parts within hermet-
ically sealed containers, so special lubricants must be found
which do not decompose or sublimate under vacuum. Films
of low shear strength, such as molybdenum disulphide or
vacuum deposited soft metals (e.g. gold, lead, or silver), are
most efcient, particularly when designed to be situated
between hard substrates which will support the load and
keep the contact area small.
The possibility of contacting metallic surfaces becoming
cold welded to each other during, for instance, the operation
of spacecraft mechanisms or loading of threaded fasteners
under vacuum conditions depends on a number of factors;
there will be a greater chance of cold welding if
(a) the relative phase diagram indicates that the contacting
metals or alloys form a solid solution with each other,
(b) the metals are soft and have the same crystal structure,
(c) contact surfaces are clean, or possess easily
damaged/removable oxide lms. Surface oxides are
normally more brittle than metals and are therefore
more likely to crack and expose underlying metal if it is
too soft to provide a rm support under load, and
(d) contact pressures are higher. The rst step in order to
minimize the possibility of cold welding is to select
those metal combinations known to resist adhesive
wear, as for instance those shown in Fig. 2.4a. The next
step is to consider the surface nish.
Clean metallic surfaces react with the surrounding
atmosphere to form oxides, nitrides, or other compounds that
are held together by either strong chemical bonds, or weak
van der Waals forces. These surface lms reduce the pos-
sibility of metalmetal adhesion that would otherwise occur
on intimate contact; they can be considered as naturally
formed lubricating lms. Under the space environment such
lms are unlikely to be self-healing if they become dis-
placed. For this reason protective lms of PTFE (Teflon),
graphite, and molybdenum disulphide are frequently selec-
ted to prevent wear and cold-welding.
Suitable non-lubricated pairs of engineering alloys for
sliding wear situations can be selected from engineering
alloys having very different hardnessestraditional refer-
ence books may be consulted (Brandes 1983; Lansdown and
Price 1986), or for very precise data, the recent work of
Merstallinger et al. (2009) can be checked for evidence of
fretting wear and cold welding (it is intended that this ref-
erence work is maintained on the internet and updated as
new pairs of engineering material pairs become tested).
In general, steels can be coupled with either copper alloys
or steels of a different alloy type and hardness. Copper alloys
can be coupled with chromium plating, high-chrome steels,
and tungsten steels. Austenitic stainless steels have a
26
Fig. 2.4 a Depicts the theoretical possibility of clean metal surface
pairs becoming cold-welded upon contact. Choice of a metal to resist
adhesive wear with another specied metal. The blacker the circle,
the better will be resistance to adhesive wear and cold welding under
vacuum (Lansdown and Price 1986). Same-to-same metal contacts will
result in solid solution, sticking and cold welding. b Demonstration of a
case where same metal contacts have become cold welded (a form of
solid state diffusion). By applying lead or silver coatings (e.g. as solid
great tendency to cold-weld to each other (see Fig. 2.4) as
they are both soft and unable to form thick protective
chromium oxides. Alternatively, with very hard substrates
such as chromium plating, surface deformation is so small
that the surface chromium oxide lm is never disrupted (see
Fig. 2.5). Materials suitable for spacecraft bearing applica-
tions are titanium-carbide coated balls located in raceways
fabricated from 440C or SAE 52100. A particularly good
anti-wear surface, as for instance in gears, bushes, and piv-
ots, is plasma-nitrided steel (Rowntree and Todd 1988).
Thermal spraying is also a novel process for the coating of
spacecraft subsystems enhancing wear resistance and as a
thermal barrierhere thermal spraying can deposit both low
and high melting point materials such as polymers and
ceramics and metallic layers such as aluminium onto CFRP
substrates such as antennae face-skins (Sturgeon and Dunn
2006; Saber-Samandari and Berndt 2010). Case histories
related to wear are detailed in Sects. 5.2.7, 5.11, and 5.12.
A large number of rules and design recommendations for the
avoidance of wear and cold welding (for instance at mech-
anism end stops, hold-down and release springs, sliding
contacts, and ball bearings) have been listed in the form of a
2 Requirements for Spacecraft Materials
lubricants) such sticking will be avoided. The very thin oxide lm on
austenitic stainless steel is easily ruptured under the sliding conditions
of torquing-up nuts and bolts. High vacuum operation has contributed
to the complete seizure by cold-welding of this 316 alloy vacuum
chamber support xture. The 45 mm diameter bolt was initially cut to
release the threaded portion which was later cross-sectioned. The
pointer shows the main region of cold-welded asperities
standard (Labruyre and Urmston 1995; Doyle and Hubbard
2010; ECSS-E-ST-32-08 2013).
It should be noted that all dry lubricants wear, and will
possess a nite life. However, improved wear characteristics
can be achieved by carefully selecting the process of applying
the dry lubricant. Ion-plated lead and sputtered molybdenum
disulphide are now well proven, having low coefcients of
friction and a long life. Burnished or spray-bonded molyb-
denum disulphide has inferior friction properties. It is
important to store the dry lubricated spacecraft mechanisms
in a dry inert gas in order to prevent moisture pick-up, and as
neither lead nor molybdenum disulphide perform well in air
the number of operations in normal atmosphere should be
restricted (Rowntree and Todd 1988).
(e) Cryogenic temperatures
All spacecraft structural metals will undergo changes in
properties when cooled from normal ambient temperatures
to temperatures in the subzero range encountered during
solar eclipse periods or when voyaging on deep-space mis-
sions. This will be an important factor when liquid helium
2.3 The Effect of a Space Environment
Fig. 2.5 Example of 300 mm long Spacelab pallet trunnion
machined from Inconel 718 and hard chromium plated. The insert
shows plating to be 10 m thick and well bonded to the etched substrate
cryostats form a major part of a payload, as for instance has
been designed for the Infrared Space Observatory, where
sophisticated instruments are located in a 60 cm telescope
cooled to 2 K. The greater changes involve the embrittle-
ment of metal alloys, particularly carbon steels. Space
vehicles must be fabricated from materials with high
strength-to-weight ratios. They must also be required to
retain high levels of fracture toughness at all service tem-
peratures to ensure fail safe lifetimes. In general, yield
strengths, Youngs modulus, and tensile strengths increase
as the exposed temperature is decreased. The effect of
low-temperature exposure on ductility and toughness is,
however, dependent on alloy composition, and for specic
alloy data special handbooks should be consulted (Campbell
1980; Reed and Clark 1983).
(f) Corrosion
It should be emphasized that several effects of the space
environments are benecial to metallic materials. Before
launch many criteria have to be set forth for the selection of
spacecraft materials, so that failure resulting from corrosion
and particularly stress-corrosion cracking will be prevented.
27
With the exception of pressure vessels, plumbing lines, liquid
fuel cells, and galvanic battery cells, these problems of cor-
rosion are not evidenced in the vacuum environment of space.
(g) Material fatigue
The low and high cycle fatigue lives of parts fabricated from
most steels, aluminium and titanium alloys are impressively
extended under vacuum conditionsthis is particularly
welcome as many spacecraft parts will be subjected to
extensive mechanical and thermal fatigue during their
operational lives. An analysis of the results from extensive
test programmes (Grinberg 1982) strongly indicates that the
vacuum environment produces a change in the plastic strain
intensity in the near-fatigue crack region of all alloys, and an
increase in the plastic zone depth of ductile materials. This
results in a decrease in crack propagation rate, due in part to
the absence or a considerably reduced effectiveness of oxide
or chemisorbed lms on fresh crack surfaces.
(h) Spacecraft charging
In orbit or in deep space, spacecraft and space vehicles can
develop an electric potential up to tens of thousands of volts
relative to the ambient extraterrestrial plasma (the solar
wind). These large potential differences (called differential
charging) can also occur on the external surface of a launch
vehicle. The main consequences of spacecraft differential
charging are the phenomena of electrical discharge (coro-
nawhich produces a damaging glow around conducting
materials at high potential) and arcing (a luminous bridge
formed by discharge between spacecraft electrical conduc-
tors). Similar discharges may also be observed when
high-voltage equipment, such as travelling wave tubes and
electronic power supplies, operates on board for the trans-
mission of signals from the spacecraft back to Earth. Many
factors contribute to spacecraft charging, including the
spacecraft conguration, its structural and surface materials,
how correctly these materials are grounded, whether the craft
is operating in sunlight or shadow, its altitude above Earth,
and the flux density of high-energy solar particles or level of
magnetic storm activity. Many possibilities exist to neu-
tralize the spacecraft potential: where possible all intercon-
necting parts, particularly at the surface, should be
electrically grounded to ensure sufcient electrical conduc-
tivity between interfaces. This will include solar call cover
glasses and optical solar reflectors (see Sect. 5.5.4). Alter-
native, new methods that reduce surface potentials (partic-
ularly for scientic spacecraft designed to measure plasma
and electric elds in the space environment) are active sys-
tems that release a sufcient amount of charged particles
across the external surfaces. These particles are ions, emitted
28
by eld emission from a liquid metal source that can be
indium. Ion release is usually for a short time, until the
spacecraft potential is reduced, or reaches zero.
(i) Spacecraft in hibernation
The Rosetta spacecraft was manufactured by a European
consortium during the 19982003 time period. It was laun-
ched in March 2004 with the objective of making a ren-
dezvous with the comet Churyumov-Gerasimenko,
otherwise known as 67P. This craft achieved a new rst in
human history, by reaching its destination in August 2014
and began orbiting 100 km above the surface of the icy
comet, taking images and then descending to a height of
30 km before detaching its lander, named Philae. Much of
Rosettas 10 year journey was in hibernation mode. The
journey covered 6.4 bn km through the Solar System (three
times around the Earth, once around Mars, once close to
Jupiter and ve times around the Sun). Consequently this
spacecrafts hardware was subjected to most of the space
environments compiled into (b) to (h) above. Close to the
Sun the problem of overheating was solved by using radi-
ators to dissipate heat into Space. Conversely, close to
Jupiter, the hardware and experiments (20 in all) were kept
warm by multi-layer insulation blankets and heaters located
at strategic points such as fuel tanks, pipework and thrusters.
The Philae lander was separated from Rosetta by means of a
small pyro (explosive) cable cutter which activated the
release of a large compressed spring. Much metallurgical
work was conducted prior to launch to ensure that the Car-
penter spring steel would not become embrittled at the very
low (160 C) outer Solar System temperatures, or become
cold welded to its mated structural surfaces, during its pas-
sage close to the Sun. Philaes landing gear was also mate-
rially demanding due to the low temperatures encountered
and low power budget (Thiel et al. 2003). The main com-
ponents being harpoons to anchor to the comets surface: a
copper beryllium projectile, pyrotechnic expansion system,
cable magazine and a rewind system (AA 7075-T7351)
driven by a brushless motor having plain bearings machined
from MoS2-lled polyimide (Vespel SP3).
Rosetta is supplied by power from two 14-m-long solar
arrays having a total area of 64 m2. The Si solar cells used
are 200 m thick, of low intensity, low temperature type,
approximately 38 62 mm size. The cover glasses are
100 m thick ceria doped micro-sheets. Four 10 Ah NiCd
batteries store the power to supply the 28 V bus lines.
Many other scientic spacecraft covering great distances
use power from radioisotope thermoelectric generators
(RTG). For instance, the New Horizons spacecraft, launched
in January 2006 has been in hibernation for two thirds of its
flight time, and will reach Pluto in mid-2015. On reaching
2 Requirements for Spacecraft Materials
the most outer bodies known to orbit the Sun, the RGA
power system will be turned on and seven science instru-
ments activated.
2.4 Materials for Space Launch Vehicles
At present the only way that satellites, people, and cargo can
be carried off from the Earth into the environment of space is
by the use of rocket-propelled vehicles. Expendable launch
vehicles (ELVs) are used only once. Many nations are
involved with the construction and launch of ELVs. The
most well known of the several hundred launch vehicles to
have boosted spacecraft from Earth are listed in Table 2.6.
The rst European telecommunications satellite
(OTS) was lost when its ELV exploded during launch,
probably due to a defective solid rocket motor case. The
failure review established that the steel case material had
been incorrectly heat-treated. Parts of the exploded case
were retrieved from the Atlantic Ocean by submarine.
Metallographic evidence determined that the large cylindri-
cal piece-parts had received an austenitizing time or tem-
perature which was insufcient to solution treat the AISI
4130 (0.3C, 0.95Cr, 0.2Mo rem. Fe) steel. This was apparent
from the presence of large-sized spherical carbides in the
microstructure of the solid rocket motor case. The flight
hardware case had not achieved peak hardness during sub-
sequent quenching and normalizing. Incidentally, the in-line
process control sample that had accompanied the flight case
did have adequate mechanical and microstructural properties
this was because of the small mass of the test piece which
responded well to the time-temperature prole. A replace-
ment spacecraft, OTS 2, was launched successfully 8 months
later in 1978 on Thor Delta number 141this event can be
seen in Fig. 2.2. The Delta ELV continues to be one of the
most successful US launchers.
A schematic diagram of the main features of a typical
ELV is shown in Fig. 2.6. This shows the Titan rocket which
was initially developed in the USA during the 1950s and
continues to be launched today as a stretched version (Titan
in and IV). This ELV, together with the Delta rocket, was
complementary to the Space Shuttle fleet, particularly for the
launch of heavy payloads. The fleet of Space Shuttles were
retired in 2011 and NASA has selected two spacecraft to
potentially replace the shuttles. It is intended that the US will
take astronauts to the International Space Station (ISS) in
2017 by means of reusable capsules, the SpaceX Dragon and
Boeings CST-100. Each capsule can be placed on
single-use rockets, such as the Falcon 9 Heavy or Atlas 5
series, and they have been designed to carry up to seven
astronauts at once. At the time of writing, only Russia is able
to transport astronauts to and from the ISS by means of its
2.4 Materials for Space Launch Vehicles 29

Table 2.6 Selected launch vehicles

Country of origin Launch vehicle (latest known version) Typea Payload into GTOb (kg)
China Long March 3 (1992) ELV 2500
Long March 3B (2007) ELV 11,500 (LEO)
5500
European Space Agency Ariane 4 (1990) ELV 2600
Ariane 5 (1996) ELV 6800
Ariane 5 (2014) ELV 10,500
VegaItaly (2014) ELV 1500 (LEO)
India Vehicle 3 (1979) ELV 40 (LEO)
Polar Satellite Launch Vehicle (2014) ELV 3250 (LEO)
Israel Shavit (1988) ELV 160 (LEO)
Japan H-1 (1986) ELV 1100
H-IIB (2009) ELV 16,500 (LEO)
8000
USA Scout (1979) ELV 5400
Atlas 2 (1991) ELV 2700
Atlas 531 (2014) ELV 17,000 (LEO)
Thor Delta (1992) ELV 2000
Delta 2 (2012) ELV 2500
Saturn V ELV 10,000
Titan III and IV (1989) ELV 5000
Falcon 9 (2014) ELV 13,150 (LEO)
Space Shuttle (1990, retired 2011) AV 25,000 (LEO)
Former USSR Vostok (1960) ELV 5000 (LEO)
Proton (1968) Russia ELV 5500
Soyuz-2.1b (2014) Russia ELV 3000
Soyuz-2.1b (2014) Russia ELV 8500 (LEO)
Soyuz-2.1v (2013) Russia ELV 2800
Zenit 3SL (2002) Ukraine ELV 6000
Buran Retired AV 30,000 (LEO)
Key
a
ELV expendable launch vehicle; AV aerospace vehicle
b
Geosynchronous transfer orbit [flight performance to Low Earth Orbit (LEO) is usually more than twice this payload weight]

Soyuz rocket. Supplies of equipment and food are trans-
ported to the ISS using either the Russian Progress space-
craft of the European Automatic Transfer Vehicle. Since its
rst flight in 2008, the ATV has played a vital role in ISS
logistics serving as a cargo carrier, space tug and storage
facility. ATV will evolve into the European Service Module
(ESV) designed to support the NASA Orion spacecraft. The
deep-space missions envisaged by NASA will rely on the
Orion spacecraftit has been flown on a successful test
flight aboard a Delta 4 Heavy booster launched from the
Kennedy Space Centre at the end of 2014.
Propellants for launch vehicles are regarded as materials.
All rockets are propelled into the vacuum of space by
utilizing only the liquid or solid materials on-board; at pre-
sent, there is no possibility of using atmospheric oxygen.
The term propellant is used to denote the two chemical
products, the oxidizer and the fuel, contained inside
conventional rockets. The fuel is burnt with the oxidizer in
order to achieve the enormous amounts of energy needed for
liftoff. The propellants can be in either liquid or solid state.
Modern rockets are powered by cryogenic propellants:
liquids which at atmospheric pressure have boiling points
below 0 C. Examples are liquid oxygen (183 C) and
liquid hydrogen (253 C).
The most simple solid propellant is prepared from a
mixture of nitrocellulose and nitroglycerinethis is the fuel.
30
Fig. 2.6 View of the main structural parts of an ELV (based on a
Titan III design)
The oxidizer is prepared separately from either ammonium
nitrate, ammonium perchlorate, or potassium perchlorate.
The fuel and oxidizer are made into powder form and then
mixed with binder such as polyvinyl chloride or a poly-
urethane. The resulting substance is poured into the solid
rocket motor casing, where it sets hard. The case is usually
made of steel, as discussed previously for the Thor Delta
solid rocket motor, but composite materials based on gra-
phite bres are also used. More complex solid propellant
chemistry is based on polybutaidene acrylonitrile (PBAN)
and hydroxytelechelique polybutadiene (HTPB) propellant.
The HTPB propellant for Ariane V solid boosters is, typi-
cally, 68 % ammonium perchlorate, 14 % polybutadiene,
and 18 % aluminium powder. Much progress has been made
during the last decade in the development of improved solid
propellants. One propellant is identied as GAP/Al/HNF
[glycidyl azide polymer being the binder, aluminium powder
and a new form of powerful oxidizer with the chemical
composition of hydrazinium nitroformate (Schoyer 1996)].
2 Requirements for Spacecraft Materials
Comparative tests were made between this new propellant
and the best-performing HTPB-containing propellant. The
ndings included an increase in characteristic velocity of
about 8 % and, importantly, an ecologically benign exhaust,
free of chlorine, where the combustion products are nitrogen,
water, carbon dioxide, nitrogen oxide, and aluminium oxide.
Liquid propellant motors are usually fed from two tanks,
one containing the liquid fuel (such as kerosene, liquid
hydrogen, or hydrazine (N2H4)), the other containing the
oxidizer (usually liquid oxygen or nitrogen tetroxide (N2O4)).
Infrequently, fuming red nitric acid or nitrogen peroxide is
used. The propellants are injected into the rocket motors
combustion chamber, where ignition and combustion occur
with a great release of thermal energy. The combustion gases
are then forced through the exit nozzle of the motor. Here, the
kinetic energy is absorbed by the nozzle as the gas velocity
increases and then decreases through the throat of the
nozzle. A drawing of the Ariane IV launcher vehicle parts is
shown in Fig. 2.7. The rst two stages of this ELV use
hydrazine and nitrogen tetroxide. The third stage is propelled
by a cryogenic engine fed by liquid hydrogen and liquid
oxygen. As illustrated in the gure, Ariane IV has the pos-
sibility to incorporate either liquid or solid propellant booster
motors to assist in the rst stage liftoff.
In contrast to Ariane, the Space Shuttle was, as its name
implied, reusable and of extreme importance for its role in
manned, near-Earth activities. The Shuttles primary
propulsion consisted of a large external tank (47 m in length
and 8.7 m in diameter) which contained compartments for the
liquid oxygen/liquid hydrogen propellantsthese were fed
to the three main engines, and two strap-on booster motors
which contained a solid composite propellant (polybutadiene,
acrylic binder, and ammonium perchlorate oxidizer).
Rocket structural materials are usually based on the
Duralumin series of aluminium alloys (4 %Cu, 2 %Mn, rem.
Al). These alloys have high strength-to-weight ratios and are
detailed in further sections of this book. Concerning Ari-
ane IV, much use has been made of the aluminium alloys
AU4GN (AA2024), AZ5GU (AA7075), and AZ5G
(AA7020). These are the French alloy designations with the
US Aluminium Association designation in parentheses.
Further cross-references to national alloy specications are
given in Appendix 6. Some locations for these alloys are
indicated on Fig. 2.7. Unfortunately, several of their
heat-treatment conditions are susceptible to stress corrosion
cracking (SCC). A particular problem resulting from SCC is
discussed in Sect. 4.5. All the major structural materials used
for the construction of Ariane and its motors are identied in
Fig. 2.7. The more modern rocket motor and booster bodies
containing solid propellants are machined and welded from
maraging steels, which are based on iron with large
amounts of nickel, cobalt, and molybdenum. These steels
2.4 Materials for Space Launch Vehicles
Fig. 2.7 Launch vehicle inboard
prole with main structural
materials indicated (Ariane IV, 42
LP)
can be easily rolled and formed into complicated shapes,
then weldedafter this a suitable heat treatment is made
which produces very hard and tough material properties.
Liquid fuels and oxidizers are usually stored in pressure
tanks which may be made of titanium or aluminium alloys,
as shown in Fig. 2.8. These inner surfaces which make
contact with the liquid fuels must be tested and found
compatible and non-igniting with the liquid. For instance,
the titanium alloy Ti6A14V is known to be compatible with
hydrazine, and aluminium alloys compatible with liquid
31
oxygen. Similarly, any pressure vessel liner materials based
on organic materials need to be compatibility testedthis is
more difcult, only one resin system, Torlon Al-10, being
found to be compatible with liquid oxygen (Healy et al.
1995).
The Zenit ELV of the former Soviet space agency was
formally announced in 1989. The vehicle is a two-stage
liquid oxygen and kerosene rocket which, like all CIS
launchers, is assembled horizontally. Its launcher assembly,
payload integration, and launch preparation phases have
32
Fig. 2.8 A liquid propellant rocket engine
been described as highly automated (Isakowitz 1995).
General views of the Zenit assembly hall and its kerosene
LOX engine are seen in Figs. 2.9 and 2.10 respectively.
This ELV has recently been proposed for launching satellites
from the sea. The novel idea is to utilize a covered
semi-submersible oil rig as the ring platform. This would
be anchored in the Pacic, at the Equator, so making full
advantage of the Earths maximum rotational velocity (about
1600 kph). The Zenit rocket, standing 62 m tall, would face
eastward in a direction avoiding any inhabitable landfall. All
other major world space launch sites are rather far north
from the Equator with the exception of ESAs facility at
Kourou, French Guiana, which is less than 500 km north of
the Equator. The uniquely flexible Sea Launch site will
undoubtedly require the implementation of a comprehensive
corrosion protection scheme for all the associated spacecraft
materials.
The Ariane V development was initiated by the European
Space Agency in 1985. It is designed for commercial mis-
sions to launch satellites and cargo for the future space
stations. The proposal for launch of the European Hermes as
a reusable winged manned space vehicle atop the Ariane V
was cancelled in the mid 90s. However, it has flown several
cargo vehicle [Automated Transfer Vehicle (ATV)] and is
designed for a Crew Transfer Vehicle (CTV). An illustration
of the Ariane V in a dual launch conguration is shown in
Fig. 2.11. Ariane V has a length of 54 m, a gross mass of
710,000 kg, and a designed thrust at liftoff of 15.9 MN. The
2 Requirements for Spacecraft Materials
Fig. 2.9 General view of the Zenit assembly hall
Fig. 2.10 Installation of Zenit rst stage keroseneLO2 engine
Vulcain engine powers the Ariane V main cryogenic stage.
This engine consists of a gas generator cycle, in which tur-
bopumps driven by a gas generator fed by propellants tapped
from the main supply system feed fuel and oxidizer to the
combustion chamber. Liquid oxygen (oxidizer) and liquid
hydrogen (fuel) are sprayed into the combustion chamber.
Because of the extremely high combustion temperature,
reaching 3600 C and with about 1600 C at the internal wall
of the chamber, it is necessary to cool the chamber. This is
done by machining channels into the chamber wallthere
are 360and passing liquid hydrogen through them as the
engine is red. The chamber is fabricated from a wrought
high-strength copper alloy (Narloy-Z) with an outer band of
nickel. The combustion chambers of the Space Shuttle main
engines were also made from Narloy-Z (Cu, 3Ag, 0.5Zr wt%
with O2 of approximately 50 ppm) and, when correctly heat
treated, this alloy was ideal for combustion chambers oper-
ating from 252 to 540 C but above this temperature the
hot wall mechanical properties were degraded because of
grain growth, grain boundary sliding and bulging that
2.4 Materials for Space Launch Vehicles
Fig. 2.11 The main system of Ariane 5. Note For Ariane 5, all the
welded aluminium AZ5G (AA 7020) alloy previously selected for the
structure of Ariane 4, has been replaced with AA2219 owing to this
alloys improved resistance to stress corrosion cracking. The painted
surfaces of Ariane 5 can be seen in Fig. 4.113
resulted in cracking (Singh 2005). The liquid hydrogen
passing through the coolant passages at 252 C not only
cools the chamber and prevents the Narloy Z from melting,
but it is itself heated, and in turn injected into the combustion
chamber. A cross-sectional view of the chamber wall is
shown in Fig. 2.12 and is described in its caption. The
Vulcain engine develops 1.12 MN of thrust by ejecting some
250 kg s1 of gases at very high speed. These gases
harmless to the environment as they are basically water
vapourare accelerated by the nozzle to a supersonic speed.
Like the combustion chamber, the bell-shaped nozzle is also
cooled by a flow of hydrogen. Hundreds of test rings have
been made with many Vulcain engines in order to demon-
strate that the engine meets its performance specications.
Some rings (see Fig. 2.13) have exceeded 18,000 s of
operation so that engine material and equipment limits
(endurance, failure modes, etc.) can be determined. It is here
that work involving metallurgical inspections and detailed
failure evaluations is importantthe reporting of these tasks
33
Fig. 2.12 Cross-sectional view of combustion chamber wall showing
the rectangular coolant channels machined into the chamber lining
(Narloy-Z, a copper-silver-zirconium alloy with signicantly greater
strength than pure copper but with only slightly lower thermal
conductivity). The close-out band is made of thin electroplated copper
followed by electroplated nickel. This design is incorporated in the
Space Shuttle main engine and the Ariane HM60 and Vulcain engines
forms part of the quality assessments and assists in the
generation of reliability ratings.
An interesting new development is the 3D printing of a
full size combustion chamber by engineers at NASA-MSFC.
The cooling channels, similar to those seen in Fig. 2.12,
consist of 200 intricate conduits with complex internal
geometries. They are located between the inner and outer
liner walls and although difcult to accommodate, were
successfully built in by additive manufacturing. The copper
alloy powder used for the 3D printing is GRCo-84, a
material developed by NASA Glen Research Centre. It has
been stated (Hipolite 2015) that the selective laser melting
machine at MSFC fused 8255 layers of the alloy powder to
make the chamber in under 11 days.
Whereas the main structural alloy for the Ariane IV
launcher was aluminium alloy AZ5G (AA7020), a major
decision was made during the design stage of Ariane V to
change to the alloy AU6MT (AA2219) as shown in
Fig. 2.11. The liquid oxygen (LOX) tank of the main
cryogenic stage is made from two hemispherical sub-
assemblies machined from AA2219 dye forging. An equa-
torial weld unites these parts. During the ground-hold period
just prior to launch of Ariane V the LOX tank is pressurized
with helium, initially loaded into the tank as liquid helium.
An internal pressure of 2123 bar is maintained during the
flight of Ariane V. This high pressure and the operation of
the LOX tank at cryogenic temperatures under the enormous
vibrational and flight loads, requires that the AA2219
material and its weldments are extremely well controlled by
metallurgical techniques.
As a further example, the Ariane V upper stage, storable
propellant system (EPS), consists of 4 EPS tanks for the
storage of N2O4 and monomethyl hydrazine. These tanks are
also constructed from AA2219 material by direct current,
34
Fig. 2.13 Test ring of the Vulcain-2 engine (courtesy of SNECMA,
France). The nozzle exit diameter is 2.5 m. Liquid hydrogen fuel is
used to cool the engine, flowing through a jacket surrounding the thrust
chamber. The heat absorbed this way is necessary in prolonged rings
and enhances the initial energy content of the LOX and LH2
propellants prior to injection into the combustion chamber where the
highly exothermic reaction produces steam (2H2 + O2 2H2O) and
more than 1000 kN of thrust. The nozzle supports rectangular tubes
made from Inconel and welded together to form a helical structure
through which the hydrogen coolant flows (Suslov et al. 2010)
variable-pulse TIG welding. They have a volume of 2335 1
each, made up from two spin-formed hemispheres and an
equatorial ring. The initial 2219 is procured as sheet in
temper 0 A blank is precontoured on a lathe and this is
convex spin-formed in 21 steps at a temperature of 200 C.
The alloy is then solution heat-treated and quenched (535 C
for 50 min, then water-quenched within 10 s). The parts are
then stretched by 13 %, which results in the temper T-31.
The parts are welded and subsequently warm-aged at 177 C
for 18 h in order to achieve high strength properties.
Metallurgical assessments are made throughout these pro-
cess steps in order to control against grain coarsening of the
2219. Heating temperatures for spin-forming must not be so
high as to induce grain growth. Also, strain-hardening steps
2 Requirements for Spacecraft Materials
have had toe optimized in order to achieve the tanks
required margin of safety.
The large-diameter Ariane V solid propellant booster
motor cases (EAP) are seen in Fig. 2.11. These are
thin-walled tubes formed by the novel process of
counter-roller flow forming. The material is the steel
48CDNV4 (American D6AC), for which a large amount of
thermomechanical processing and property data are avail-
able. Conventional tube spinning is to place a cylindrical
tube blank on a solid cylinder mandrel and roll the outer
surface of the tube so that it becomes thinner and longer. The
special flow-forming process applied by the German MAN
company is to place a cylindrical blank tube of D6AC
between four roller-pairs. The roller-pairs are separated by a
xed distance that will be the nal tube diameter. The tube is
initially attached to a face plate, which can rotate. The tube
wall thickness is forced into each roller-pair. As the tube is
rotated, the roller-pairs are fed down the length of the tube,
causing it to elongate and become thinner. Metallurgical
assessments making use of mechanical testing and metal-
lography were used during this process development in order
to optimize the flow compression stresses in the deformation
zone and understand their effects on the steels
micro-structure. The motor thrust frame on the main stage of
Ariane 5 is essentially an unpressurised conical structure
consisting of many integrally machined parts made from AA
7075. These are mainly interlocked by manually installed
Hi-lokTM fasteners, but these lap-joined parts have also been
proven to be reliable when joined by friction stir welding
(FSW) methods (Brooker 2001).
Several companies are considering designs and the use of
new materials for future reusable launch vehicles. Single
stage to orbit (SSTO) launch systems will use advanced
metallics and advanced composite materials, and it is the
re-entry phase for such craft which will generate the highest
aerodynamic heating rates. In the early 1970s many data had
been established for a wide range of quality refractor metals
based on molybdenum and molybdenumrhenium alloys
that were covered with protective coatings. These heavy
alloys were abandoned in the mid-1970s once NASA had
decided to select a ceramic reusable material for the thermal
protection of the Space Shuttle.
There are several kinds of heat shield. The system
designed for the Shuttle is the most advanced to have flow to
date. On the Mercury, Gemini, and Apollo capsules, the
shield consisted of an epoxy-resin compound that removed
heat by flaking off as it passed through the atmosphere. This
kind of ablative shield can be used only once. The newer
thermal protection material was developed by NASAs Ames
Research Laboratory and the Lockheed Missiles and Space
Company. It is very lightweight, relatively easy to manu-
facture into small square tiles, and impervious to heat (one
2.4 Materials for Space Launch Vehicles
Fig. 2.14 a Night launch of Space Shuttle. b Illustration of the thermal
protection systems (TPS) for the Space Shuttle fleet. Key: RCC
reinforced carbon/carbon; HRSI and LRSI high- and low-temperature
ceramic tiles; FRSI felt reusable surface insulation. c Atlantis
can hold a six-inch square tile at its edges and not feel any
heat when the centre of the tile is glowing red-hot). There are
about 31,000 tiles on each Shuttle, each moulded and cut to
t the orbitors contours. The basic material for the Shuttle
insulation is a very ne glassy bre of pure silica (SiO2).
Figure 2.14a shows the approximate layout of the various
Shuttle thermal protection systems (TPSs). The reinforced
carbon-carbon (RCC) regions can reach temperatures of up
to 1500 C during descent through the atmosphere. The
reusable felt surface insulation may reach 500 C during
ascent. The HRSI and LRSI ceramic tiles differ only in their
surface coatings (Vaughn 1985). The high-temperature
(HRSI) tiles are coated with a high-emissivity black reac-
tion cured borosilicate glass (to re-radiate the heat efciently
during re-entry). The low-temperature material is coated
with a white silica/alumina coating designed to reflect the
suns radiation as the craft is in orbit. Figure 2.14b shows the
microstructure of a ceramic tiles interior. Here, the bres are
35
photographed at the KSC Visitors Centre after retirement; showing
the triple paned, optical quality windows, the LRSI ceramic tiles and
beryllium is associated with the window frames. d Scanning electron
image of Shuttle tile (silica bers in unglazed condition) 1000
seen to vary in diameter from 1 to 10 m. This material has a
density of 136 kg m3, but more rugged varieties are denser
and some contain boron, which causes the bres to weld
together. The thermal protection surfaces developed from
Shuttle experience are now very robust. The
alumina-enhanced thermal barrier tiles overcoated with
reacting cured glass have been produced that can withstand
the impact of 800 kph raindrops (Healy et al. 1995). The tiles
are bonded onto the aluminium skin of the Shuttle with a
silicone adhesive known as RTV 560, which is a somewhat
soft material that avoids cracking of the rigid tiles as the
aluminium airframe deflects in flight. Thermal protection is
not needed for lower surface temperatures; for instance,
Inco-617 could be selected for temperatures under 980 C
and Ti-1100 for temperatures below 700 C (Baumgartner
and Elvin 1995).
The retirement of the Shuttle fleet in 2011 has meant that
crew members from the ISS can only return to Earth by
36
Fig. 2.15 Engineers completing the installation of the heat shield on
NASAs Orion spacecraft. The Avcoat ablative material is covered with
a silver reflective tape that protects it from the extreme cold
temperatures of space. The heat shield has been shown able to protect
the Orion crew from temperatures in the order of 2800 C during
re-entry and ocean splashdown (courtesy NASA)
means of the Russian Soyuz Descent Module. This vehicle
can hold up to three astronauts/cosmonauts and takes about
3.5 h to return and land on the flat grassy plains of Kaza-
khstan in central Asia. Four parachutes dramatically reduce
the speed of descent 15 min before landing, followed by the
nal main parachute which reduces the speed of the craft to
24 feet per second and at this speed, one second before
touchdown, the vehicle is slowed down to make a soft
landing by the ring of two small engines located at the
bottom of the vehicle.
NASAs Orion crew module (Fig. 2.15) is designed to
accomplish manned missions to the Moon, an asteroid and
even Mars. On its return to Earth this spacecraft will descend
through our atmosphere much like the Apollo capsules. The
rst Orion was launched in December 2014 atop a Delta 4
Heavy rocket in a test flight to check many of the systems
critical to safety. It has shown that advances in heat-shield
design and materials technology can tolerate temperatures in
the order of 2800 C during decent phases of such missions.
The Orion heat shield consists of a titanium skeleton covered
with a carbon-bre skin. This skin is later covered with a
honeycomb layer composed of ber-glass and phenolic
resin. The honeycomb cells are then lled using a hand-held
dispensing gun with a material called Avcoat, a mid-density
ablator that is similar to the ablative material used on the
Apollo command module (Reuther 2010). Avcoat is an
epoxy novalic resin, designated by NASA as 5026-39-HC/G
(with a revised composition that contains less carcogenic
compounds than the Apollo shield chemistry).
The Falcon 9 rocket is an evolved expendable two-stage
launch vehicle manufactured in the US by the SpaceX com-
pany. It made history in 2012 by launching the Dragon
spacecraft (somewhat similar to Orion) into the correct orbit
2 Requirements for Spacecraft Materials
for rendezvous with ISS. Falcon 9 has nine rst-stage engines
and a single second stage engine, this conguration is con-
sidered to reduce the number of possible separation events.
Falcon 9s faring is manufactures in two halves from sheet
carbon bre bonded to an aluminium honeycomb core. The
rst stage tank is manufactured from aluminium-lithium alloy
by means of friction stir welding and it contains two alu-
minium tanks that are capped by aluminium half-domes for the
liquid oxygen and RP-1 propellant respectively. The second
stage is manufactured similarly to the rst stage from AlLi
and the interstage utilizes a composite structure made from
carbon-bre lay-uped onto an aluminium core. Unusually, a
pneumatic separation is employed for the faring and for both
interstage separations, in order to reduce the shock forces that
accompany the ring of pyrotechnic devices. The follow-on
Falcon Heavy, presently in design phases, is expected to be the
most powerful rocket flyingonly the Saturn V moon rocket
last flown in 1973 could deliver more payload into space.
A recent addition to the European family of launchers is
Vega (see Fig. 2.16) developed initially by the Italian Space
Agency (ASI) and Italian industry (see Table 2.6). Vega
became ofcially an ESA programme in 1989 and the rst
Vega lifted off from the Guiana spaceport in 2012 on a flawless
qualication flight. The three subsequent launches were fully
successful and nine further launches are planned post-2015.
A breakdown of the advanced materials used to construct
Vega (2006, VEGA) can be listed in a similar way to those
selected for Ariane and illustrated in Fig. 2.7.
Payload Fairing:
Structure comprises of two halves of sandwich panel
CFRP sheets and Al honeycomb core that has been already
used on Ariane 4.
AVUM liquid rocket (UDMH and nitrogen tetroxide)
Upper Stage:
Carbon-epoxy cylindrical case with propellant stored in
two identical titanium tanks as seen in Fig. 2.17.
First Stage (with solid rocket motor SRM P80FW), Sec-
ond Stage (with SRM ZEFIRO 23) and Third Stage (with
SRM ZEFIRO 9):
Structures are graphite-epoxy lament wound monolithic
motor case protected by low-density thermal insulation
packed with microspheres (EPDM). The interstages are
forged rings of either 7075 or 7175Al alloy in the T7351
heat treatment condition (for each of the 1/2, 2/3 and the
3/Avum interstages) and these are chromic acid anodized for
corrosion protection. Some separation springs are made from
17-7PH steel. The solid propellant for the solid rocket
motors is monolithic Finoxil grain shaped with a star shape
on the nozzle side and EG1LDB3 low-density EPDM based
rubber insulant. The motor nozzles are made from: 3-D
carbon/carbon throat and carbon phenolic exhaust cones.
In 2015 the Vega rocket successfully launched a 5 m-long
wingless reusable spacecraft, weighing 2 t, into a suborbital
2.4 Materials for Space Launch Vehicles
b Fig. 2.16 a Evening view of a mock-up Vega launch vehicle within
path. The craft, named Immediate Experimental Vehicle
(IMV), separated from Vega at a height of 450 km and
reached a speed of 7.5 km/s, later demonstrating maneu-
verability using thrusters and aerodynamic flaps to reduce
speed from hypersonic to supersonic. It then glided through
the atmosphere before deploying a parachute and landing in
the seathe craft was then picked up by a recovery vessel
and shipped to ESA-Estec for analysis. Of key concern will
37
the mobile gantry in Kourou. Vega benets from the location of the
European Spaceport, GSC, at Kourou in French Guiana as it lies close
to the Equator. The vehicle will take small payloads into a variety of
orbits including Low Earth, Sun Synchronous and Polar. Note the four
tall lightning rod towers stretching far above the mobile gantry and the
housed launch vehicle. Lightning is very common at both KSC and
GSC sites so it is important that launch vehicles are protected against
the ionized plasma that constitute the branches of lightening discharges.
The tops of four towers are interconnected by stainless steel catenary
cables which form effective Faraday shielding. b Kennedy Space
Center in Cape Canaveral, Florida. Space shuttle Endeavour sits on
Launch Pad 39A as a storm passed prior to the rollback of the Rotating
Service Structure on 28th April 2011. The xed service structure
(FSS) seen here, is 106 m high to the top of its lightning mast. The
following day Endeavour launched the STS-134 mission to deliver a
high pressure gas tank, spectrometers, spare parts and antennae to the
ISS. 39A has been used for the manned Apollo-Saturn V launches.
SpaceX, the new commercial space company now launches its Falcon 9
and Falcon Heavy from this same pad
be the effect of the flight prole on IMVs structural mate-
rials and the hypersonic aerothermodynamics phenomena
whereby air (O2 and N2) molecules are dissociated at high
velocities and temperatures. A wide variety of materials are
being tested during the IMV test flights into low Earth orbit.
These include different types of ceramic matrix composites
used on the nose, hinges and flaps and ablative thermal
protection materials such as Portuguese cork and
silicon-based materials to withstand the severe re-entry
environment up to 1700 C.
Silicon carbide bre reinforced metal matrix composites
are of increasing interest for aerospace and space applica-
tions. The surge in interest in the late 1980s was driven by
interest in the USA and Europe in hypersonic vehicles
requiring high specic strength and stiffness materials with
elevated temperature capability. Further aerospace work has
focused on actuators and landing gear where the very high
compressive strength (2.54 GPa) coupled with low mass
and corrosion resistance make them ideally suited to replace
steel and other metallic parts with 3040 % weight reduc-
tions typical (Fig. 2.18a). The increased aerospace devel-
opment has re-established interest from the space sector with
the low mass high compression strength and potential for
welded joints being of particular interest for space craft
structures and robotic systems.
The Reaction Engines Skylon spaceplane has identied
titanium MMCs as a suitable material for the fuselage truss
structure and other static components (Fig. 2.18b, c). This
truss structure needs good tension and compression proper-
ties and potential for elevated and cryogenic temperature
operation. The truss structure also needs a low mass and
easily assembled jointing for the MMC tubes which can be
welded and inspected with upwards of 60,000 tubes over the
80 m fuselage. Other materials incorporated into this
spaceplane are seen in Fig. 2.18d.
38
Fig. 2.17 Installation of the nal stage of the second Vega (VV02)
launch vehicle inside the mobile gantry at GSC. This upper stage is
called the AVUM (Attitude Vernier Upper Module) contains the
Propulsion Module that burns pressure-fed UDMH and nitrogen
tetroxide as propellants (built by the Yuzhnoye Design Bureau) and
the Attitude Control System (courtesy ESA)
2.5 Non-metallic Materials
2.5.1 General
Non-metallic materials appear throughout the chapters of
this book as they are selected during the design and manu-
facturing stages of every spacecraft subsystem. Once, only
metallic materials, specically alloys, were used for indi-
vidual structures, housings and tanks. Nowadays, as the
physical and mechanical properties of organic and composite
materials have been developed for terrestrial use and rened
for the space environment, spacecraft are often constructed
from more (both volume and weight) non-metallics than
metallic materials.
Spacecraft utilize plastics for components, elastomers for
propellant diaphragms, lubricants for bearings, polymers for
circuit boards and paints, textiles for astronauts spacesuits,
adhesives for structural bonding, ceramics for optical mirrors
2 Requirements for Spacecraft Materials
and composite materials for an equally exhaustive list of
applications.
Nonmetals are often dened as not containing metals
and that they are able to combine with hydrogen, carbon and
oxygen to form stable compounds. This is a very loose term
as most non-metallic materials do contain metallic elements.
Polymers and elastomers are formed when monomers and
other mers are combined together and this is often referred
to as polymerization. These reactions are often not completed
and some material remains as an unreacted fragment. Simi-
larly, the reaction may be complete but there remains a
byproduct that may be unstable, particular under vacuum.
Other polymers or organic materials will often contain
deliberate additions of chemicals and metals which impart
special properties such as flame retardation, anti-oxidation,
colouration, and electrical or magnetic properties.
Sublimation, or the transition of solid metals directly to a
gaseous phase (without passing through an intermediate
liquid phase) has been discussed in Sect. 2.3b. Certain
metals and alloys are specically forbidden from space use
as they can sublimate from one surface on the spacecraft and
later condense on another, not necessarily cooler, surface to
cause problems such as optical degradation of mirrors and
electronic short circuits. The behavior of organic materials
under vacuum is more complex than that of sublimation.
Organic materials are said to outgas under vacuum. Here
there can be an immediate loss of species such as H2O, CO2
and N2 from a surface, followed by constituents having a
low molecular weight, a process akin to distillation. Raising
the temperature causes more molecules to outgas and there
may be some breakdown of chemical bonds resulting in
further outgassing and reduction in weight of the organic
material. With time, there can be solid or gaseous diffusion
from the materials interior to the surface and this can cause
the material to further degrade in properties or lose weight in
a vacuum or space environment.
Outgassing is one of the most problematic behavior of
spacecraft materials as the outgassed species can recondense
on thermal control surfaces, optical surfaces, electrical
contacts (rendering them open-circuits) and cause corona
effects. Each of these problems are described as case studies
or failure modes within this book and can be located via the
Index.
Most of the Space Agencies have performed tests on
organic materials during the screening of potential spacecraft
materials and their processes. The main test procedures are
those dened by the American Society for Testing of
Materials (ASTM 2007) and ECSS (ECSS-Q-ST-70-02),
these are considered important reading for those involved
with material selection. Basically, the test is performed on
very accurately pre-weighed samples (between 100 and
300 mg) positioned in a specially designed holder. Each
sample is conditioned (24 h at 22 C at 55 % RH). The
2.5 Non-metallic Materials 39

Fig. 2.18 a Illustration of a

landing gear side stay made from
silicon carbide bre reinforced
titanium alloy matrix highlighting
its light-weight property. This is
30 % lighter than the forged
aluminium baseline. Courtesy of
TISICS Ltd., Farnborough. b The
Skylon spaceplane is unpiloted,
reusable and intended to transport
up to 15 t of cargo into space. It
will use the SABRE engine that
combines air-breathing and rocket
cycles so that take-off is possible
from a runway, flight is direct into
Earth orbit and return is, like an
aircraft, back to a runway.
c These Skylon truss struts have
been tested in compression up to
50 kN exceeding the 30 kN
design load. Actually a couple of
struts did fail around 50 kN. In
tension they fail close to 30 kN
indicating that there is room for
further improvement in the cuff
region. d The materials proposed
for construction of Skylon are
silicon carbide bre reinforced
titanium for fuselage and wing
spars, aluminium alloy for the
propellant tank, the external
aeroshell is made from bre
reinforced ceramic, and the
landing gear will possibly use
silicon carbide bre reinforced
titanium matrix composite.
Courtesy of Reaction Engines
Limited
40
holder is transported into a vacuum chamber having speci-
ed dimensions. Testing is made under vacuum at 1 Pa, at a
temperature of 120 C, for 24 h. A collector plate is main-
tained at 25 C and situated at 10 mm above the sample
during the test period. Figure 2.19 illustrates the basic con-
cept of the outgassing test method. After the samples
temperature/vacuum/time exposure, the sample is again
weighed. It is then re-conditioned at 55 %RH (in order to
re-absorb any water) and re-weighed. The mass of con-
densed material on the collector plate is also measured by
accurate weighing.
The following terms should be understood:
collected volatile condensable material (CVCM)quantity
of outgassed matter from a test specimen that condenses on a
collector maintained at a specic temperature for a specic
time
NOTE CVCM is expressed as a percentage of the initial
specimen mass and is calculated from the condensate
mass determined from the difference in mass of the col-
lector plate before and after the test.
recovered mass loss (RML)total mass loss of the specimen
itself without the absorbed water
NOTE 1 The following equation holds: RML = TML
WVR.
NOTE 2 The RML is introduced because water is not
always seen as a critical contaminant in spacecraft
materials.
The WVR is the water vapour regained by the sample after
reconditioning.
total mass loss (TML)total mass loss of material outgassed
from a specimen that is maintained at a specic constant
temperature and operating pressure for a specied time
NOTE TML is calculated from the mass of the specimen
as measured before and after the test and is expressed as a
percentage of the initial specimen mass.
The following Accept Criteria are generally contractual
requirements set by the customers procuring space hardware,
but it is accepted that many spacecraft units can be baked in
order to satisfy these materials requirements:
As a minimum, the outgassing screening parameters for a
material selection shall be as follows:
1. RML < 1.0 %;
2. CVCM < 0.10 %.
NOTE 1 For materials used in the fabrication of optical
devices, or in their vicinity, the acceptance limits can be
more stringent than those stated below.
NOTE 2 It is nowadays becoming standard practice to
bake critical hardware (such as structural parts, harness,
electronic boxes and thermal blankets) to the highest
2 Requirements for Spacecraft Materials
permissible temperature for a few days in order to
remove residual contaminants, process contaminants and
handling contaminants.
Several compilation tables of outgassing data of materials
intended for spacecraft use have existed since the 1960s. Of
course, much of that data is now obsolete as material man-
ufacturers and suppliers update their products and withdraw
some from the market. Importantly, some manufacturers
may have retained their products Trade Name, but changed
the procedure by which it is made. This makes it important
to check outgassing data on every batch of production
unless safeguards are made to ensure that products conform
to certicates of compliance, or that process identication
documents (PIDs) are established and audited by the
procuring agent.
Websites can be visited to acquire some preliminary data,
unfortunately they are often not kept up-to-date, and include:
http://outgassing.nasa.gov/
http://TEERM.nasa.gov/links/html
http://esmat.esa.int/Services/outgassing_data/outgassing_
data.html
Testing for outgassing may be made at centres such as:
Integrity Testing Laboratory, Markham, Ontario,
Canada; AAC, Wiener Neustadt, Austria; ESA-Estec,
Noordwijk, the Netherlands; INTA, Spain; TS-Space Sys-
tems, UK, and the Goddard Space Flight Centre, Greenbelt,
MA, USA.
Outgassing is distinctly different from the offgassing of a
material. Outgassing, as previously discussed, occurs under
a reduced pressure or vacuum. When materials emit volatile
compounds in an ambient atmosphere this is termed off-
gassing. It can be likened to the smells that are emitted from
the interior of a new car. In the 1970s the offgassing of art
objects in a museum were investigated and found to be
related to the degradation of both antiquities and their
packaging materialsa standard Oddy test method was
devised using gas chromatography-mass spectrometry
(GCMS) analysis. Many offgassed species were found to
degrade stored antiques, they were based on alcohols, acids,
amines and siloxanes (Tsukada 2012). In one specic case,
amines were found to volatilise from polyester polyurethane
foams that were used to cushion and padding within storage
cases. The amines released by these foams had severely
damaged Ming Chinese jars by causing thick white
efflorescent deposits to form on these ancient pieces. Very
similar contamination from offgassed foams has resulted in
the corrosion of stored spacecraft electronic components:
leads nished with silver, nickel and also tin-lead have
become tarnished and sometimes so oxidised that they are
2.5 Non-metallic Materials 41

Fig. 2.19 a Photograph of a Micro-VCM test equipment for the determination of outgassing properties. b Schematic drawing to show the heater
bar and location of the plate onto which collected volatile condensable materials (CVCM) will deposit ESA
42
rendered non-solderable, even with activated solder flux. It is
strongly recommended to avoid the following packaging
foams and sheets: polyvinyl chloride, Styrofoam, rubber,
paper, wool, silk, nylon and some untested polyurethanes.
Better long-term storage media, some of which provide for
anti-static discharge, would be glass, stainless steel, poly-
ethylene, Dow Ethafoam 220 or similar, and Azote poly-
olen foams and Zotek polymer foams from Zotefoams.
Offgassing is associated with the safety of human life and
is addressed during the assessment of materials and specic
equipment that will be used in a manned environment such
as the International Space Station. European standards for
offgassing and toxicity were developed in the 80s during the
construction of the ESA Spacelab in order to avoid causing
harm to both astronauts and to the hardware in these manned
compartments. In fact for such projects, as well as the var-
ious equipment and experiments that are supplied to the ISS,
it is mandatory that all organic materials are tested for off-
gassing, toxicity, bacterial and fungal growth as well as
flammability. The relevant standards include ECSS-Q-ST-
70-21 and -70-29 and NASA-STD-6001.
2.5.2 Classes of Non-metallic Materials
The space community has adopted several classes of
non-metallic materials. As a brief summary some of the vast
number of suitable materials are listed hereunder. However,
it must be recognized that every material will require to be
tested to ensure it is suitable for a given project application.
Some of the products listed are contained in the datasheets
annexed to ECSS-Q-ST-70-71.
Optical materials such as organic glasses, oxide ceramics
and amorphous inorganic glasses (need to consider radiation
environment to avoid damage and yellowing): Makrolon
(polycarbonate) from Bayer; and Spectrosil fused silica from
Optik Heraeus. Zerodur is selected by many companies for
the manufacture of space and ground-based telescopes, it is a
lithium aluminium silicon oxide glass ceramic characterized
by evenly distributed nano-crystals within a glass phase
matrix having an extremely low expansion coefcient.
Optical bres have come into a class of their own and are
more frequently selected for box-to-box optical communi-
cations within spacecraft. Each bre, generally made of silica
sometimes containing a germanium dopant, is an optical
waveguide that can transmit light along its axis. Optical bres
are manufactured as a core inside a cladded layer and an outer
protective coating. They have advantages over copper elec-
trical wires in that they are lighter, chemically inert, do not
corrode and can be easily joined (see also Sect. 6.14).
Important for space use, they are immune to electromagnetic
interference, and are electrically non-conductive so avoiding
short circuits and grounding problems. One drawback is the
2 Requirements for Spacecraft Materials
radiation induced damage to these bresbut this will
depend on the orbital altitude and proximity to the Van Allen
belt. The Long Duration Exposure Facility (LDEFsee
Fig. 8.12) demonstrated that after 69 months in LEO no
damage ensued except for one cable hit by a micrometeorite.
Adhesives, coatings and ATOX-resistant paints:
Araldite AV 138 with hardener epoxy from Huntsman;
DC93-500 space grade encapsulant from Dow Corn-
ing; RTV 566 silicone from Momentive Performance
Materials; MAPSIL 213 silicone potting and MAP ATOX
41-8 varnish for atomic oxidation protection, also electri-
cally conductive brown adhesive QS 0225 EA83, thermally
conductive white adhesive QS 1123 TA66 and black general
adhesive QS 1123 CEIT, all from MAP Technology.
Adhesive tapes (for flight or temporary use): Scotchtape
No. 5 from 3 M Company; Eccoshield PST CA from
Emerson and Cummings.
Paints (can be epoxy-, silicone- and polyurethane-based,
with or without pigments such as zinc oxide and titanium
oxide (white) or carbon black (black)): MAP SG121FD
white non-conductive flexible; New paints compliant with
REACH, being water-based, and having no ITA restrictions
are: AQPU1 a non-conductive paint, and AQPUK a con-
ductive paint; all from MAP Technology; Aeroglaze Z306
black polyurethane with high thermal absorptivity from
LORD Corporation.
Lubricants: Dry MoS2 (alone or within a Teflon or
polyimide binder), Fomblin Z25 silicone oil from Solvay
Plastics; see also organic lubricants listed in Table 5.2.
Potting compounds, sealants and foams. It should be
noted that there is a great variation in the effectiveness of
sealantsthat is, the time for moisture to permeate various
sealant materials. Generally silicones are poor sealants for
moisture, epoxies are better followed by fluoro-carbons, but
for truly hermetic seals only glasses or metals can be con-
sidered: Solithane 113300 two part urethane from EV
Roberts; Stycast 2651-40 dielectric epoxy, and other low
outgassing products from Henkel or Emerson and Cuming.
Reinforced plastic composites (usually carbon, glass,
Kevlar, Zylon or boron bres, whiskers or chopped bres, in
a polymer matrix). These are used for both structural and
electronic applications: Applications for reinforced plastics
can be found throughout this book by making an index
search. Fasteners such as bolts, screws and pins give ultimate
weight saving choices for applications with moderate loads,
with the weight of aluminium and 1/5th the weight of steel
they are resistant to harsh chemicals. With other com-
posite materials they have a matched CTE and superior
galvanic corrosion properties while being electrically and
magnetically transparent. Various fastener products from
Click Bond, Inc.
Rubbers and elastomers. Used for damping systems,
clamps, seals and bladders for fluids: Viton B-910 from
2.5 Non-metallic Materials
DuPont; SIFA-35 and AF-E-332 from PSI Corp. It is
essential to know the properties of these materials over a
wide range of temperatures as will be highlighted by the
explosion, and loss of life, of the space shuttle Challenger
(STS-51). This accident was caused by a sequence of events,
but the engineering cause related to the low temperature
mechanical properties of the elastomeric O-ring seals that
interfaced the joint between the two cylindrical lengths of
each booster. The seals were designed to be compliant
between the two metallic joint faces in order to prevent hot
gas, under huge pressure, escaping during take-off. The
STS-51 launch took place during an unusually cold period at
KSC which caused the elastomer to shrink and loose elas-
ticitythis caused the hot combustion gases to escape, burn
the seal and release enough gaseous fuel to cause melting of
the case and the catastrophic explosion.
Thermoplastics. Often used for multi-layer insulation
(MLI), second surface mirrors, electronic wire insulation and
sleeving: a multitude of products such as Kapton H, FEP,
PTFE are produced by companies such as BASF, ICI and
Aventis.
Thermosetting plastics. Usually thermosetting polymers
consisting of base, hardener and catalyst which when mixed
have a determined shelf life: products include Araldite
CT205 as used for printed circuit board lamination; and,
Epikote-828.
Obsolescence can be a problem with inorganic materials.
Production lines may have ceased due a materials lack of
demand. The space market is very small compared to other
industries and some costly production processes may cause
the manufacture of space-approved products to be unprof-
itable. For instance, RT/Duroid 5813 was extensively use in
the space industry for ball bearing cage applications, but
manufacture of this precision bearing material stopped in the
mid-90s. Stocks were exhausted just before a replacement
material was foundidentied as PGM-HT (a composite of
PTFE, glass bre and MoS2). Although thought to be a
direct replacement for Duroid, the PGM-HT was found to
shrink on returning to room temperature after heating under
vacuum. After much more searching and testing at the
European Space Tribology Laboratory (Buttery 2011)
another material was found: Sintimid 15M/Tecasint 1391
which is currently considered to be a viable alternative to the
recent, US-sourced, Vespel SP-3. Incidentally, the
RT/Duroid laminates produced by Rogers Corporation for
space electronic printed circuit boards (high-power and
microwave) continue to be produced, they consist of PTFE
lled with random glass or ceramic micro-bres.
43
2.5.3 Novel Non-metallics
(a) Vespel (DuPont) has been utilized more frequently for
spacecraft applications during recent years as inferred
by the previous paragraph. It is a polyimide-based
plastic and has particularly good space properties,
such as a low outgassing under vacuum, even at high
temperatures. The basic material (SP-1) has a low wear
rate and low coefcient of friction in vacuum and
moisture-free environments. The tribological properties
of Vespal can be further enhanced by the additions of
15 % graphite (SP-21), 10 % PTFE (SP-211) or 15 %
molybdenum disulphide (SP-3). The low thermal
expansion of all Vespel variants are between 35
(SP-21) and 55 (SP-211) ppm/C, relatively low for
polymers (when compared to values listed in Appen-
dices 1C and 1E). Vespel has been used in a spacecrafts
liquid helium cryostat, as a valve seat, selected for its
stability at cryogenic temperatures (see Fig. 5.71). As an
example of its versatile temperature use, Vespel has also
been demonstrated as a successful non-metallic inter-
face material against aluminium in the design of the
electric propulsion pointing hold down and release
mechanism of Bepi, qualied from 40 to +150 C
under contact loads of between 15 and 60 MPa (Janu
et al. 2009).
(b) Graphine. This is the thinnest compound known to
man and comprises of a thin layer of pure carbon. The
carbon atoms are bonded together in the form of a
hexagonal, honeycomb lattice. It is a two-dimensional
carbon monolayer and being only one atom in thick-
ness, it is the lightest material known to man (1 m2
weighs about 0.77 mg). Graphine is also the most
electrically conductive material available and although
it has the potential to be used for spacecraft commu-
nication equipment, with vast weight savings, it can
only be produced in very small sizes and quantities.
Applications include durable display screens, electrical
circuits and solar cells. Graphine is produced by
chemical vapour deposition (CVD) and this process
deposits carbon in the form of multi-crystals having a
single-layer sheet of carbon atoms. It is these grain
boundaries that are thought to affect the larger-scale
properties of the sheet, such as thermal and electrical
conductivity. The grain orientations and surface
topology of graphine has been observed by
atomic-resolution transmission microscopy (Chuvilin
et al. 2014). Possibly graphine could also be used to
44
(c) Carbon nanotubes (CNT) are, like graphine, a great
(d) Graphite bres having a thermal conductivity of over
(e) Shape memory polymers or so-called SMPs (as
Fig. 2.20 a Depiction of the carbon nanotube structure; properties can
be used to create the strongest known bres, they can be very flexible
and foldable. b Illustration of an ultra-low density inorganic foil which
can be made from carbon bre mesh, open-cell SMP foams and
aerogels. Once unfurled, they are stable, moderately heat-resistant and
could be used for solar sailing. They could overcome current
deployment limitations by being self-deployable and very large
structures can be envisaged (Dunn 2001)
2 Requirements for Spacecraft Materials
build a large solar sail; it would have the smallest mass
per unit area and, if the material could be manufactured
into large sheets, it could sail using solar radiation
(photons) from the Sun. By adjustment of the sails
angle relative to the Sun, the sail could be steered so as
to tack towards or away from the Sun and could pro-
gress to the edge of the Solar System. To date, only
beryllium sails have been considered possible as they
can be produced to a thinness of about 40 nm. Plastic
thin lms are likely to vaporize once they are unfurled
and exposed to ultraviolet radiation. Similarly, the
high-energy, ionizing solar photons may degrade and
make the properties of the graphine sail unobtainable
(Matloff 2013).
challenge to engineers who are attempting to utilize
these cylindrical rolls of carbon molecules (Fig. 2.20a).
As CNTs have a density 50 % less than aluminium and a
tensile strength of 1060 GPa (an order of magnitude
greater than the strongest steel, or Kevlar) the material
could be formed into a composite for highly efcient
space structural applications. Carbon nanotube based
composites and carbon-carbon are currently under study
as structural materials for space telescopes, optical
benches and large mirrors. CNTs may have tune
able electronic properties from metalloid to
semi-conductors.
800 W/m-K, are now being bundled with a polyester
cord to form Thermal Straps that can be used as flexible
thermal links (or heat straps). The US company,
Technology Applications Inc. is the only producer of
these fascinating straps. They are extremely light,
flexible and up to one tenth the weight of equivalent
copper straps. These Thermal Straps have been used as
heat pipes in spacecraft applications where extremely
low outgassing is required. They have a variety of end
ttings; aluminium and copper can be selected to avoid
galvanic couples or corrosion.
opposed to SMAs, the alloys described in Chap. 4.18)
are only recently being considered for industrial
applications; these are based on covalently cross-linked
polymers, and their shapes can be changed by the
application of heat and electric elds. Damage to
spacecraft antennae or structures caused by space
debris may be mended and accurate proles restored by
realigning those surfaces to face solar radiation from
the Sun (Athimoolam 2012). These novel, and
gossamer-like materials may also be used as ultra-light
materials for self-deploying structures as shown in
Figs. 2.20 and 2.21.
2.5 Non-metallic Materials
Fig. 2.21 Illustration of the InflateSail 3U CubeSat with 1 m long
inflatable boom and 10 m2 deorbiting sail proposed by Surrey Space
Centre at the University of Surrey. The primary goal of this concept
project is to attach such a sail to a low Earth orbiting satellite at the end
of its life and use the sail to remove it from orbit (image credit SSC)
Many other new materials are under investigation for
space use. ArallR is a bre-metal laminate having ara-
mid bre-prepreg sandwiched between sheets of alu-
minium alloy. GlareR consists of aluminium alloy
sheets bonded by glass bre prepreg. Metal matrix
composites and ceramic matrix composites will be
referred to in subsequent chapters.
(f) So-called self-healing materials have been considered
for the repair of military vehicles and battle jackets,
ships and airplanes during the past twenty years. They
have also been studied for spacecraft applications
(Semprimoschnig 2007). Probably the most promising
system is based on fragile microcapsules. These can be
incorporated into an organic coating and, for instance,
when the coating is damaged due to impact or fatigue
cracking, the microcapsules in the proximity of the
defect will break and release a mix of polymerizing
liquids that then cure and heal the damaged area.
A similar concept, but based on hollow bres (see
Fig. 2.22), has been developed by the University of
Bristol where healing occurs when an epoxy
resin-solvent is released from the bres and flows onto
a complex catalyst contained within the matrix and is
exposed at a crack front (Coope et al. 2014). Clearly,
self-healing materials are of great interest for space
missions where there are threats from impacts by
micro-meteoroids and space debris; remoteness and the
45
Fig. 2.22 Example of an early hollow bre repair sample seen by
electron microscopy. The release of the repair agents from within these
storage reservoirs mimics the bleeding mechanism in living creatures.
Here, alternate hollow bres contained a resin and a hardener. Work by
Dry, C. reproduced for the STFC Workshop (Semprimoschnig and
Eesbeek 2007)
inability to make effective repairs in space will spur on
further developments in this discipline of materials
science.
More often scientists and materials engineers are
looking to nature for ideas as how to make structures
stronger to improve their performance or lighter so that
they use less materialthese aims have been
expounded by Professor Barbers team at the Univer-
sity of Portsmouth in their study of limpet teeth (Barber
et al. 2015). These small teeth were found to be com-
posed of a composite structure containing goethite
nanobers, -FeO(OH), distributed within a softer
matrix of a protein phase. The tensile strength of
miniature dog-bone samples, removed from the limpet
teeth by cutting with a focused ion beam in vacuum,
was found to range from 3.0 to 6.5 GPa. Each sample
was milled using a focused ion beam attachment to a
scanning electron microscope and the results can be
seen in Fig. 2.23. The strain behavior at fracture being
between 4 and 8 %. These values are equivalent to the
strongest man-made bres such as Toray T1000G
carbon bres with a tensile strength of 6.5 GPa.
(g) Another interesting development is the in-laboratory
manufacture of Gecko biomimetic adhesive tape for
space applications such as holding down Multi-Layer
Insulations (MLI) and other reflective or protective lms.
Microscopy has determined that a geckos foot has about
ve hundred thousand keratinous hairs, each up to
130 m in length and having a diameter of one-tenth
that of a human hair. The rst direct measurements of a
single hairs adhesion property was made using a
two-dimensional micro-electro-mechanical systems
force sensor (Autumn et al. 2000). The adhesive forces of
46
the hairs were thought to be governed by the
ultra-efcient application of van der Waals forces. The
mimicking of gecko hair properties has been attempted in
a large Italian study (Gregoratti et al. 2013). Vertically
aligned multiwall carbon nanotubes (VA-CNTs) were
grown on both rigid and flexible metal substrates using
lithographic techniques: some gecko adhesive properties
were achieved but it seems more questions than answers
are recorded at this early stage in the study (Gregoratti
et al. 2013): How the do the gecko properties respond to
changes in temperature and humidity? What is their
long-term reliability in specic environments? Can such
a material surface be made robust and reliable? And,
most importantly, can they ever be fabricated
cost-effectively over large areas?
(h) Although not entirely a new material to the space
materials engineer, cork is nding more interesting and
novel applications. Owing to the chemical and physical
structure of corkthe natural bark growing on the trunks
of cork oak trees (quercus suber)this materials has
found important uses as ablative material on re-entry
vehicles. As listed in Table 2.7, cork has a low density
and low thermal conductivity. Corkcalled natures
foam by NASAis 90 % gas and is essentially a struc-
ture made up of dead cells, each having a polyhedron of
only 3040 m in diameter (Coelho 2009). Ablative
materials protect space vehicles from the severe aero-
dynamic heating during launch and during re-entry into
the atmosphere at velocities greater than 7 km/s.
The ablative materials used on the re-entry capsules of
rst US manned missions, Mercury, Gemini and Apollo,
were based on cork. It was used on the solid rocket
boosters of the Shuttle and applied to the booster
nose-cone and frustum of Delta launch vehicles. In
Europe, cork-based ablative materials were applied to
the front and back covers of Beagle-2, are used on
Ariane-5, the Atmospheric re-entry demonstrator and
Mars Express (Bouilly et al. 2006, 2013). Illustrations of
cork being applied to space hardware are shown in
Figs. 2.24 and 2.25.
Cork is re resistant and chars rather than burns. As a
thermal protection system (TPS) it undergoes several
physiochemical changes each of which are
heat-absorbing. Initially it heats up, then the front
exposed surface is thermally degraded (charred), gasses
released during pyrolysis pass though the charred layer
which becomes thicker. Gas reaction with the carbona-
ceous layer creates channels through which gas and
reaction products are released as ablation vapours. This
releases thermal energy into the air flowing past the
structure. The cork composite is formulated so that as it
chars, the interface between the advancing char and the
underlying (virgin) material causes chemical reactions
2 Requirements for Spacecraft Materials
that are either exothermic (at the lower temperatures) or
endothermic (as the material sublimates). The charred
layer is a good insulator, it has a good emissivity and, as
its surface is eroded in the flow of passing air, further heat
is dissipated away from the structure. Extensive studies
into chemical reaction, ablation mechanism and mor-
phology of the resultant black charred layer include those
of Asma et al. (2010). A recent EU-funded project called
Aerofast has provided promising results for two new
families of TPS materials fabricated from cork supplied
by Norcoat-Liege and Amorium Cork Composite
(ACC). The work (Pinaud et al. 2014) also validated a 3D
charring ablation model taking into account the decom-
position of the TPS at various temperatures, its decrease
in thickness and required dimensions for various aero
dynamical shapes.
The protective cover attached to US spacecraft bound
for Mars is called an Aeroshell. The Aeroshell is
constructed from an aluminium honeycomb, covered
with graphite-epoxy face-skins that are in turn covered
with a phenolic honeycomb containing ablative mate-
rial. The ablator is a unique mixture of specially graded
cork granules, binder and silica glass microspheres. The
chemical formulation of this mix is designed so that as
the probe or landing craft enters the Martian atmo-
sphere, friction causes the ablator to undergo chemical
reactions and char. This has the effect of dissipating heat
which is taken away by the hot wake of gas behind the
vehicle. Thermal protection system must be able to
resist temperatures which could reach 1600 C during
entry into the Martian atmosphere, and yet maintain a
temperature of less than 125 C inside the probe. The
tiles can be formed and individually machined on all six
faces, then bonded to the probes surfaces with silicone
blue, ESP 495.
(i) A material nding many new applications is expanded
polytetrafluoroethylene (ePTFE). The microstructure
mimics some cell structures found in nature, as seen in
Fig. 2.26. The main usage of ePTFE in space applica-
tions is as a low dielectric constant (1.3) insulation
material for high-speed digital coax cables. Here,
ePTFE provides high signal quality and optimum per-
formance in miniaturized systems where conductor
sizes as small as AWG 42 can be manufactured. As this
material is inherently inert, thermally resistant and
hydrophobic, it can also be processed into
high-flowrate microporous membranes. Applications
will be as lters within manned spacecraft modules for
air and potable water purication systems. ePTFE l-
aments have recently been incorporated into a
bre-reinforced matrix (such as PEEK) as a reservoir
for solid lubricants. The porosity of ePTFE, as seen in
Fig. 2.26, can be processed to vary from 5 to 90 %.
2.5 Non-metallic Materials 47

Fig. 2.23 Scanning electron

micrographs showing (a) the
limpet tooth prior to FIB milling,
and (b) milling to thin the sample
towards a dog bone tensile
sample geometry (c) shows plots
of stress-strain for individual
limpet samples with a variety of
lengths that were tested to failure
using an atomic force microscope
(Barber 2015)

Table 2.7 Three varieties of thermal protection systemstheir sizes and properties (by permission of Amorim Cork Composites ACC, Mozelos
VfR, Portugal)
P45 P50 P60
Cork particle size (mm) 12 0.51 0.51
Sheet dimension (mm) 1270 710 1270 710 1000 500
Sheet dimension (in) 50 28 50 28 40 20
Density @ 20 Ca 0.32 0.48 0.45
b
Tensile strength (psi) 110 250 160
Tensile strength (MPa)b 0.76 1.50 1.10
b
Elongation (%) 30 13 7
Thermal conductivity (Btu in)/(h ftb F)c 0.45 0.50 0.55
c
Thermal conductivity (W)/(mK) 0.06 0.07 0.08
Specic heat (Btu/lb F) 0.6 0.5 0.4
d
Specic heat (KJ/Kg/K) 2.5 2.1 1.9
Substrates to bond Metals and composites
a
ASTM F1315
b
ASTM F152, method B
c
ASTM C177
d
ASTM C351

Work by the University of Florida (Vail et al. 2011) is pores of the included ePTFE laments onto the con-
on-going: to regulate wear in tribological applications tacting surfaces of mechanisms. As all materials under
by the secretion of very thin lubricant lms from the evaluation appear to be stable in a vacuum/space
48 2 Requirements for Spacecraft Materials

Fig. 2.24 The Beagle 2 Mars entry probe tted with cork TPS tiles.
a Front-shield and back cover inspection in assembly work-
shop. b Front-shield tile being bondedprior to packing in bio bags
and sterilization treatment

environment, such novel lubrication methods might

eventually be applied to spacecraft mechanisms.
(j) Liquid crystal polymers (LCPs) have been evaluated
by NASA Glen and are said to out-perform conven-
tional polymeric materials for application as flexible
printed circuit substrates and large flex-antenna arrays
that can be rolled into cylindrical shape and launched
into space (Kingsley 2008). These LCPs are heat
resistant, strong and light-weightideal properties for
space applications. LCPs have been commercially
available for two decades, but modern process have
enhanced their strength and ability to be formed into
Fig. 2.25 ac Cork is applied as a TPS onto fairings and nose cones of
exceedingly thin sheet and ne wires. The wires have
modern launch vehicles such as Ariane 5, Vega and the SpaceX Falcon.
been compared to spiders silk, being of exceptional Its long heritage includes successful use on Atlas V, the former Space
strength, and comparable to steel-wire. LCP sheet can Shuttles and Apollo missions
be made to melt at 290 C or 315 C, at the lower
temperature it is called bondply and at the higher tem- components. The material can be made with a thickness
perature, the core. So, by utilizing bondply between two of 25 m, it has a low water absorption of up to 0.04 %
cores, and applying heat to slightly above 290 C, an (but is permeable to water) and resists degradation due
inseparable bond is achieved. Multiple stacking can be to radiation in space. The Japan Aerospace Exploration
made, rather like a flexible multilayer board, with metal Agency (JAXA) has used very thin LCP sheet material
vias, printed interlays of metal, and even embedded for solar sail applications, several have been successful.
2.5 Non-metallic Materials
Fig. 2.26 Electron microscope image of expanded PTFE membrane.
This is seen as a brillated form of PTFE; with a porous network of
brils connected to dense nodes of PTFE. Magnication of 40,000 times
One named Ikaros recently passed Venus and JAXA
are now considering future sail missions, powered only
by the Suns light particles, to Jupiter and the asteroid
belt.
Returning to LCP wire, Vectran is said to be the only
commercially available melt-spun LCP bre in the world;
weight-for-weight this bre is ve times stronger than steel,
has a high abrasion resistance, cut resistance and has good
mechanical property stability over a wide range of temper-
atures. These bres can be coated with copper, nickel or
silver and is available (Liberator Conductive Fiber) in
bundles comprising of 20, 40 and 80 individual laments.
This is a new product and has been used as a woven braid
material for cables. The new braid material will be of interest
to spacecraft harness fabricators due to its substantial weight
savings over copper braid as illustrated in Fig. 2.27.
2.6 The Potential for Welding
and Joining in a Space
Environment
2.6.1 Background Considerations
Large dimensions and mass mean that structures such as the
International Space Station (ISS) cannot be transported as
one complete structure into low Earth orbit. They must be
launched in sections and assembled in space (and as already
49
Fig. 2.27 Braided sleeves, each 11 mm wide, for coaxial cables (a) is
standard nickel-plated copper braid, (b) is a novel braided sleeve
utilizing Liberator nickel-clad conductive LCP bres. These novel
bres are 15 times stronger than copper and 86 % lighter; for a RG142
coax cable the overall cable is 40 % lighter than a standard shielded
cable
discussed, in future missions assembly may occur on the
surface of our Moon and possibly other planets). So far, little
work has been specically dedicated to intra-vehicular and
extravehicular activity (IVA and EVA) joining and cutting
issues in space. This section will revisit some of the ndings
of Dunkerton et al. (2001) and include more recent welding
developments having relevance to space applications. By far
the largest contributions made to date, on the development of
welding and brazing technologies for use in space (Dzha-
nibekov et al. 1991; Paton 2003), have been made by Russia:
1969Vulkan, for electron beam (EB), low-pressure plasma
and consumable electrode welding
1979Isparitel, the depositing of thin coatings by conden-
sation and evaporation
1983Isparitel M, deposition of braze metals
1984development of a versatile hand-held tool for cutting,
brazing, welding, heating and coating
1988YantarEB coating and welding
1990Universal Hand-held 3 kW manual EB gun used on
Mir Space Station. (Figs. 2.28 and 8.33)
Reference to the numerous metallurgical experiments
performed in the space environment by astronauts and cos-
monauts, for instance on Skylab, Soyuz, Mir and the Salyut
spacecraft, have been compiled by Flom (2005) and by
Paton in his reference book (2003). The majority concerned
the melting and joining of metals in space. Flom (2005,
2006) noted that the vacuum of space is higher than a typical
Earth-situated vacuum chamber can provide so that electron
beam welding would require very little power and cathode
50
Fig. 2.28 ESA astronaut Claude Nicollier at work in December 1999
making repairs to the Hubble Space Telescope with a mechanical
fastener tool during the refurbishment mission (courtesy NASA)
life would be enhanced; he further describes the
NASA GSFC study for electron beam brazing. These were
not pursued during the construction of the ISS which is
based on modular assembly; the interconnection of its units
uses rather traditional mechanical joints designed to be
compatible with EVA astronauts. For instance, truss joints
are assembled by using tubular parts that are interlocked by
rst insertion, twisting and locking. Mechanical fastener
tools have been used for more than two decades, one oper-
ation is shown in Fig. 2.28.
The use of mechanically assembled trusses for assem-
bling large structures is described by Watson et al. (2002),
but in a later NASA work (Dorsey et al. 2012) proposes that
although astronauts and robots can efciently construct
space structures with mechanical connections, it is the use of
custom-built welded structures that will be the game-
changers of the future.
The majority of spacecraft materials are coated (e.g.
inorganic paints, silicones, specially conducting indium tin
oxides, alumina, silica, platinum, gold, and chemical con-
version coatings such as Alodine and the anodisation of
aluminium alloys). This provides protection from terrestrial
corrosion prior to launch. Coatings are also designed to
ensure that surfaces have stable features, such as adequate
resistance to atomic oxygen and constant solar absorption
and emission properties in orbit. These coatings may affect
subsequent joining/cutting operations. Therefore there is a
2 Requirements for Spacecraft Materials
certain antithesis between coating of the structure to ensure
its stability and resistance to degradation, whilst still
ensuring that the joining/cutting process is viable.
If one considers using welding or other processes for
repairing manned habitats it is essential that the Whittle
shields (to protect habitats from micrometeorites and
man-made debris as described in Sect. 8.4.4) are removed in
order to access the modules wall. Shields and thermal
blankets if not detachable, can form a physical barrier to
EVA joining/cutting activities in terms of accessibility
and/or temperature constraints.
If joining techniques are to be applied for construction
activities in space it will be important to consider the fol-
lowing list of critical factors:
Range of material compositions, their coatings and form,
including thickness
Portability/ease of use/stability of the tool under vacuum
(e.g. lubrication)
Health and safety
Energy requirements
Surface preparation/t-up/jigging
Metallurgical implications and joint performance
Processing debris must be avoided
Intra or extra vehicular process application
Equipment maintenance
Gravity effects
Joint hermeticity.
2.6.2 Potential Joining and Cutting Processes
The following non-mechanical processes can be
considered:
Adhesives, Arc welding, Brazing, Electron beam weld-
ing, Friction welding and Laser welding.
These processes can be considered against the criteria listed
in Sect. 2.5.1, to select the most promising techniques as noted
below. However, it must be noted that these considerations are
subjective and would need extensive further evaluation:
Organic Adhesives
These offer a nominally simple means of bonding two
materials together, with fairly simple equipment and at rea-
sonable cost. They might be the most suitable means for
joining carbon bre reinforced plastic components that have
been designed to incorporate lap-joints into which adhesive
resin can be injected. However, adhesives are not considered
to be suitable for general application during IVA or EVA
activities since such joints suffer from poor hermeticity,
usually high outgassing rates prior to curing under vacuum
and unknown reliability over time.
2.6 The Potential for Welding and Joining in a Space Environment
Arc Processes
The fundamental argument against using arc fusion pro-
cesses in space is that an ionised gas, or plasma, is
required for any of the processes to operate. This means
that, generally, arc processes exhibit a lack of vacuum
stability and require copious amounts of gas. Gas is
supplied either from pressurised gas cylinders, or by a flux
coating. However, a variant of gas tungsten arc welding
(GTAW) However, a variant of the GTAW process,
which uses a hollow tungsten electrode, was rst devel-
oped for welding in a vacuum during research activity
starting in 1965 at the E O Paton Electric Welding
Institute. This equipment was tested in October 1969 on
the Soyuz 6 spacecraft (Fig. x) but the process did not
full expectations. GHTAW (gas hollow tungsten arc
welding) requires only 1/1001/1000 the flow rate of gas
normally used at atmospheric pressure, but even this
concentration of plasma-forming gas in the arc gap zone
could not be maintained. Development work continued
and in 1973 Shiganov (1973) reports that a highly stable
arc discharge could be maintained in an evacuated
chamber on Earth. In 1986, Rocketdyne, a division of
Rockwell International, began development of similar
hollow electrode technology in the US. In 1989, Rocket-
dyne prepared a report for NASA proposing an extrave-
hicular activity welding experiment and trails, made in
drop-shafts and in low gravity aboard an aircraft, gave
encouraging results (Watson and Schnittgrund 1989).
The GTAW process offers the following advantages:
Operation in both the vacuum environment in space and the
pressurised environment inside the space vehicle is possible.
Relatively wide tolerance in process variables (travel speed,
position and arc length) enables manual operation. This
process is easily automated and relatively safe.
GTAW is widely used to weld all common aerospace
alloyse.g. steels, Al and Ti alloys and is an established
technology.
However, there are a number of notable disadvantages to
this process, namely: a gas supply required (heavy cylinders,
need relling or replacing); is not a cutting technology.
Electrode damage due to vaporisation in vacuo (need to
regrind or replace) and vaporised products may condense
and contaminate astronauts visor and other critical surfaces.
Also, electromagnetic interference (EMI) may interfere with
electronics systems, plus high voltage or HF surges on arc
initiation could draw power away from other systems.
Therefore, it is considered that, whilst a useful technique,
there may be other welding methods of greater merit.
51
Brazing
Brazing is of limited use for the assembly of large structures
in space, especially if they are made from alloys such as
AA2219, due to the lack of suitable brazing alloys. The
removal of oxidation layers (from ground storage) or pro-
tective coatings would be required prior to any brazing
operation. Presently the most suitable/versatile method for
the heating of braze alloys is via an electron beam. This
technique is difcult to adapt for the repair/patching of
damaged structures. Flom (2006) has, however developed a
low temperature brazing process, where heat is delivered
from an electron gun and, under high vacuum, braze alloys
based on Au-Sn can easily be drawn by capillary action into
specially pre-prepared joints. [Flons EB brazing process
was used in the refurbishment of an Italian spacecraft elec-
trode housing, see Sect. 4.28].
Electron Beam Welding
Electron beam welding is particularly suitable for use in
space. EB can be used as a very controllable heat source for
welding, brazing, soldering, cutting, original manufacture,
repair etc. on electrically conductive substrates. The material
heated during the welding process is minimallimited to a
very narrow fusion zone (of say 1 mm for a joint of 3 mm
thickness) and the heat affected zone (extending say 0.3 mm
beyond the fusion zone). As little material is heated, the
process is extremely efcient compared with arc welding
techniques. The method is utilised throughout the space
industry for the fabrication of space hardware, as described
in Sects. 4.2 and 4.13. Although narrow welds may be
produced with some lasers, the efciency of electron beam
generation can be in excess of 95 % compared with typically
10 % for the most efcient laser. The Russian hand-held low
accelerating voltage (<8 kV) electron beam gun (named the
Universal Hand Tool)as described by Paton (2003) and
illustrated in Fig. 2.29has already been demonstrated in
space EVAs for EB welding, heating and cutting.
There are two main potential problems associated with EB
equipment. First, the process generates X-rays at the point of
impingement of the beam on the substrate. By choosing a low
*10 kV voltage, as for the Universal Hand Tool, the addi-
tional contribution to radiation exposure during EVA is likely
to be small. Secondly, if the welding beam should impinge on
(say) an astronaut during EVA, the risk of perforating the
spacesuit is signicant. In the past this risk was deemed
unacceptable by NASA, but recent US activities (Flom 2006,
Dorsey et al. 2012) indicate that safety considerations are
now acceptable, particularly when it is robots that perform
the handling, manipulation and EB welding tasks.
52
Fig. 2.29 Universal hand-held 3 kW manual EB gun used on Mir
space station (source from 1999, Avtomaticheskaya Svarka, 10)
The EB processes and EB equipment which can be uti-
lised in space are by no means perfect for every application,
but they currently offer the best compromise for many
applications as well as signicant scope for further rene-
ment and diversication in the future.
Friction Welding
Unlike some processes, there is no published data on friction
welding in space. The principle of the friction stir welding
(FSW) process is described in Sect. 4.25 and characterised
by Fig. 2.30. This process is used extensively during the
manufacture of spacecraft hardware in many countries.
Briefly, the heat is generated by a rotating tool, which
consists of a shoulder, which rests on the surface of the
material to be welded, and a pin, which penetrates almost the
complete thickness. Frictional heat causes material to soften,
and the forward translation of the rotating tool causes
material to be pumped from the front of the tool to the rear,
where it is consolidated under pressure to form a solid state
weld. It is emphasised that there is no fusion in this process.
Modications to the FSW process are termed friction
taper plug and stud welding, these have also been adapted to
2 Requirements for Spacecraft Materials
Fig. 2.30 Schematic drawing of friction stir welding
ll holes. There are several distinct processes, in which the
process mechanisms differ fundamentally. All of the pro-
cesses work better on fully penetrating holes, as it is often
difcult to completely eliminate defects at the bottom of
blind holes. Perhaps the most important process is friction
taper plug welding, where a tapered plug is friction welded
into a tapered hole. The process has been scaled up to make
plugs in material of up to 38 mm, but clearly the equipment
becomes more robust and heavier, and the power demands
greater. However, for aluminium materials of 10 mm or less,
such as the AA2219 alloy plate used on Spacelab and ISS, it
is feasible to consider this process.
Friction stud welding equipment is generally light and
easily transportable, and the intermittent use of the equip-
ment suggests that overheating of the motor in vacuum is not
a major issue. Although it is only capable of welding studs to
other structures, this can be developed with a little imagi-
nation to enable many construction and possible repair tasks
to be undertaken. It is probable that stud diameters should be
restricted to about 5 mm in order to reduce mass, and
increase ease of use in space, but this will also require an
increase in the rotation speed of the motor. Friction stud
welding might be envisaged to plug damage caused from
micro-meteoroid impacts on pressurised spacecraft as illus-
trated in Fig. 2.31. Threaded studs might also be attached to
space structures and enable the further integration of hard-
ware by mechanical screw fastening in space. Similar EVA
was performed by an ESA astronaut (Fig. 2.28).
Laser Welding
Lasers are a very attractive tool for application in space due
to their versatility to weld, cut and repair metals and to a
certain extent, non-metals. Diode lasers appear to be the
most promising candidates due to their high efciency, small
size and low maintenance requirements. However, the
2.6 The Potential for Welding and Joining in a Space Environment 53

Nevertheless, it is expected that developments in the

brightness of diode lasers will increase the available power
density to be suitable for welding and repairing the materials
of interest for space applications.

2.6.3 Expectations

1. Electron beam (EB) welding is considered today, to be

2. Friction welding has much potential for use in a space
Fig. 2.31 Microsection of friction stud weld made to 3 mm thick plate
that contained a simulated impact crater. Note that no material debris
was released and a proper metallurgical joint exists between stud and
plate (Dunn unpublished)
current developments of diode lasers are not sufcient in
terms of power density, to produce welds or cuts within the
range of materials and thickness used in space applications.
the most applicable process for both joining and cutting
operations. In particular is the ease with which the pro-
cess can be adapted and its ability to be operated in a safe
manner. Extensive work has already been carried out on
EB welding in space and suitable, commercially avail-
able equipment is already on sale.
environments, although it is more suited to joining than
to cutting. It is a versatile process, which can be used in
arduous environments and suitable for the types of
materials used in space construction.
 The Integration of Materials into Product
Assurance Schemes
3.1 General Product Assurance
and the Role of Materials
3.1.1 Product Assurance Management
The role of a Product Assurance Manager is to support the
spacecraft project by providing technical management
leadership in applying a number of technical disciplines to
the design, manufacture, testing, handling, and operation of
the project. At the same time he needs to ensure that each
discipline has the required impact to produce high-quality
hardware for the amount of money spent. Product Assurance
(PA) will be applied to each of the spacecraft models
described in Sect. 3.2.2 (viz. thermal, structural, engineering,
and flight models). The actual range of PA disciplines may
vary from one project to another, but generally comprises:
1. Quality assurance (QA) and Software assurance
2. Reliability engineering and assessment
3. Safety
4. Materials and processes
5. Component part selection and procurement
The majority of these ve PA activities in the space
industry do not differ from similar activities in other indus-
tries where high quality and high reliability are essential
requirements. However, it is probable that in the manufac-
ture and fabrication of spacecraft their application is more
rigorous because once a spacecraft is launched the possi-
bility of repairs/replacements is usually remote (Space
Shuttle did permit limited repairs to be performed in space,
and projects such as the Space Telescope were designed so
that spacecraft systems such as the solar arrays could be
periodically returned to Earth for refurbishment). Thus all
systems must function and continue to function satisfactorily
for their design-lives, which, as already discussed, are cur-
rently up to 15 years for a spacecraft providing a commercial
communications service.
Springer International Publishing Switzerland 2016
B.D. Dunn, Materials and Processes, Springer Praxis Books,
DOI 10.1007/978-3-319-23362-8_3
3
The foregoing description of PA management referred to
ve disciplines. It is essential for any materials engineer,
chemist and metallurgist working in the space eld to fully
understand the activities and functions of these disciplines.
They are to some degree overlapping, but should never be
considered solely in the context of the old-fashioned term
quality control (QC). In fact QC is a superfluous term
which can always be replaced by a more explicit word such
as inspection, radiography, test, etc. The following subsec-
tions will detail each of the ve PA disciplines and make
reference to the important role of the materials laboratory
and metallographic services. In fact the variety of different
materials investigations selected and reported in this book
may be individually ascribed to one of these disciplines.
The specialist engineers working in the elds of QA,
reliability, materials and processes, and electronic compo-
nents might well appreciate the usefulness of a laboratory,
but unless the unique assets of this service are appreciated in
the company boardroom, funding and support to this labo-
ratory will fall off and will be accompanied by an inevitable
fall in the quality and reliability levels of the companys
product. Equally, there will be little ground for establishing
condence in new technologies developed by industrial
contractors for future space applicationsthis aspect,
including the need for destructive physical analysis, failure
mode analysis, etc., will be discussed later on. It is essential
that top management policy should reflect the PA managers
support for a strong pool of materials experts and a labora-
tory able to give rapid functional support to the various
spacecraft projects; this must be an integral part of every
spacecraft contractors product assurance scheme.
3.1.2 Quality Assurance
QA embraces all the activities concerned with quality. It is a
management system, and is equally concerned with the
functions of design, calibration of process equipment,
workmanship standards for process operators, heat-treatment
55
56
control, and inspection and testing. Standard metallurgical
processes rely on the generation of pre-production test
samples and metallography for the setting up of optimum
process variables, generation of workmanship standards, and
the determination of accept/reject criteria for porosity or
inclusions within materials and joints. Inspection functions,
frequently referred to as quality control, might utilize met-
allographic techniques for establishing the cause and possi-
ble acceptability of mechanical damage. Heat-treatment
condition, approximate chemical composition, dimensional
tolerances for machined or electroformed hardware having
complex shapes, etc. can each be evaluated or inspected on a
sampling basis by means of metallography. Similarly,
organic materials and their processes will be assessed by the
kinds of tests discussed in Sect. 3.2.5. Material traceability is
important; this will usually start from the procurement
source and be controlled by documentation to stores and
then to the assembly area. Materials with a limited shelf life,
such as adhesives, paints, potting compounds, and confor-
mal coatings should be kept in containers that have been
labeled by QA staff to show product name and code, date of
manufacture and expiry date. In addition these materials
should be linked to specications that detail how they should
be stored (temperature limitations and possibly refrigeration)
and this documentation may specify how, or if, measures can
be made to extend their shelf life.
Critical materials will probably be those requiring out-
gassing and flammability testing or stress corrosion testing
so that they can be validated for a given space project.
Critical processes are those that are irreversible (product
cannot be repaired) and when that process is applied to very
expensive and complex assemblies. They may be considered
critical if the processed part or sub-assembly cannot be
inspected. Processes such as coating, soldering are said to be
veried if technology samples are evaluated by testing and
found acceptable.
As an aside, it is interesting to observe that in European
industry, many materials engineers, who considered their
careers to be limited by their purely technical role, have
turned to QA engineering. It has been remarked that they have
a reasonableness with regard to taking into account many
ill-dened influences when making a decision, and therefore
make good QA managersa somewhat misguided inference
to materials engineering as being an inexact science!
All companies contracting to the space industry, whether
as manufacturers, suppliers, or test laboratories, are aware of
the fact that quality is synonymous with survival. The
implementation of slogans such as higher productivity,
lower costs, on-time deliveries, and increasing product
quality are strategies considered essential in order to com-
pete in the international market. One way to help achieve
these goals is to apply for registration, or seek to be qualied
according to the ISO Standard 9000 with a registrar (e.g.
3 The Integration of Materials into Product Assurance Schemes
Lloyds of London or the Underwriters Laboratories). ISO
9000 is an auditing type of standard that requires companies
to document all aspects of their operation. A company
normally begins by documenting all their procedures as
standards. They write, or rewrite, their quality manual to
cement their newly established procedures, which should
enable them to consistently produce products to the quality
levels demanded by their customers. All changes to manu-
facturing and testing procedures must be properly docu-
mented. In some companies where material processes have
been performed for many years, the skills of the process are
often passed down from one operator to another. For
instance, the precise times and temperatures for a brazing
operation may not be documented in a controlled procedure
and, instead, the writer has seen a broad range of tempera-
tures and times handwritten on a piece of paper glued to the
side of a console above the brazing furnace. Similarly, the
cleaning operation may not be documented and different
operators are observed to clean printed circuit board
assemblies in a bath of solvent for different times, some only
swilling the board around while others use brushes to
remove contaminants from difcult-to-reach surfaces of the
board beneath components. And no-one really knows when
the solvent is so contaminated that it has to be replaced. In
these cases the procedures manuals will be difcult to write,
as these will set the standards to which all personnel must
work, and there can be many points of conflict for individ-
uals. We do a good job, have done for ten years and no-one
complained before, It was done for OTS-2 and that satellite
was still working 11 years after its two-years designed life!,
etc. These comments come from people who have yet to
appreciate the concept of QA. Tension can certainly run high
when the rst set of procedures is produced, despite all the
assurances that it is the starting point and requires feedback
from staff and operators in rening the wording and content.
Some operators will think it is an administrator in the ofce
that is telling them how to suck eggs, or just wishing to
create paperwork for the sake of it. Again, these people have
preconceived ideas and are reluctant to see what something
new may be able to do for them. Someone has to produce the
written procedure, and it is seldom that this is written by the
doers. They are, however, more likely to provide an input
to the draft when they can see what is intended. QA is not
about checking or everyone following one set of rulesit is
about effective management which will enable problems to
be identied, analysed, and rectied. It does not mean that
there will never be mistakes, but that they will be minimized
and have less chance to cause schedule and launch delays.
Materials and Failure Analysis laboratories can be audited
and certied to the international management standard, ISO
9001. This is widely used by space manufacturing compa-
nies for managing the quality of the lab service. However,
this standard does not evaluate the technical competence of
3.1 General Product Assurance and the Role of Materials
the laboratory or that its test procedures, inspection methods
or calibration are suitably accurate. A more suitable standard
to address these aspects is ISO/IEC 17025; this standard will
also assess the competence of the lab staff, whether the test
methods are valid, equipment is calibrated to national stan-
dards and the suitability of the labs own QA system.
Implementation of the system of auditing required by ISO
has highlighted and then strengthened some weaknesses in
the smaller space manufacturing companies. It is also an
effective tool to build trust and a higher degree of coopera-
tion between management and the shop floor because it
keeps all procedures up to date, everything is written and
documented, and employees know what their job is and
where their boundaries and responsibilities lie. Another
advantage is that the number of customer visits for quality
audits is reduced when the ISO certication is accepted in
place of the visit. Machining rms, heat-treatment, plating
shops and painting facilities are the kind of companies likely
to benet as more emphasis is made of in-line process
controls. This has been found to reduce the scrap rate at nal
inspection and thus offer quality products at lower prices.
In the space manufacturing industries, the introduction of
ISO 9000 and its related specications, as well as the stan-
dard space project contractual documents such as the
European Cooperation for Space Standardization, ECSS
series, has helped not only quality but has also ensured that
state-of-the-art technology has been exploited. In recent
years this has included computer-aided design and analysis
(to model stresses in a system, estimate load-bearing capa-
bilities, etc.), manufacturing techniques and tooling (CNC
machining cuts steps between design and manufacturing,
computer-controlled lathes are accurate, flexible, and fast to
set up), computer programs (for statistical process control
and computerization of test and inspection equipmentthis
can measure tighter tolerances), and materials (e.g.
high-alloy steels with better cleanliness, fewer inclusions,
and a lower gas content to enable higher strength levels).
3.1.3 Reliability and Safety
Reliability activities are directed toward ensuring that the
design, selection of materials, and choice of fabrication
methods for the various spacecraft subsystems are based
upon sound principles of reliability engineering and analysis.
In recent years the concept of reliability has acquired a new
dimension which encompasses the generation of numerical
predictions of the reliability of a spacecraft system. This is
made possible by the study of individual components, metal
joints, etc. under environmental testing or during real life so
that their life times can be predicted. Single-point failures
are generally not permitted on spacecraft; they will be
replaced by redundant electronic circuits, mechanical
57
devices, and additional structures, to improve the overall
system design.
The reliability of electrical and mechanical components,
metallic joints, and structures are calculated from test pro-
grammes that incorporate the stresses and environments
expected in spacecraft service. Such programmes rely
heavily on the results of materials lab work which are
brought together in report forms referred to as Destructive
Physical Analysis (DPA) and Failure Mode Analysis
(FMA). The objectives of the DPA are to catch unapproved
design changes or unacceptable workmanship. Normally, a
small number of parts are selected at random from a batch of
mechanical or electronic parts and are sent for evaluation.
They provide the baseline for the quality of the item which
will then be submitted to both static and dynamic tests (e.g.
vibration, shock, thermal cycling under vacuum) as part of
the reliability test programme. The FMA will subsequently
be made to identify the various actual failure modes expe-
rienced or to pinpoint possible failure modes. Several
examples of these kinds of materials examination will be
illustrated in subsequent chapters; they relate to the relia-
bility of spacecraft structures, fasteners, battery cells, elec-
tronic components, and the like.
In the case of manned programmes, such as the
now-retired Space Shuttle Orbiters, Spacelab, and the Inter-
national Space Station (see Fig. 3.1a, b), it is also essential to
evaluate all materials used from the safety viewpoint, and
tests are conducted for flammability, toxicity, and odour.
Hazardous metallic materials such as beryllium and mercury
are generally excluded from designs, but with additional
safety precautions they might be permitted for applications
such as X-ray windows or charged particle thruster engines.
Each of these materials would be tested in its application
set-up. Another aspect of safety includes the testing and
rejection of metal alloys found to be susceptible to stress
corrosion cracking (this test is described in Sect. 4.5; it
requires that specimens, stressed and exposed to cycling
immersion in a salt solution and drying, must be metallo-
graphically examined and compared with unstressed con-
trols). The maintenance of reusable spacecraft, like that of
aircraft, has always been considered superior to the mainte-
nance of vehicles for other modes of travel. The amount of
time and effort expended by spacecraft designers, mechanics,
and inspectors in providing a high level of safety is unsur-
passed by any other industry. Yet, although few in number,
there have been occasions where mistakes were made which
resulted in accidents. No matter how small the calculated
probability is that an adverse event will occur, there is still
that chance that it will happen. To increase the safety margin
against structural failure, spacecraft designers employ
redundant load paths and redundant systems. Finally, safety
also includes the so-called fracture control procedures. Since
all engineering materials contain flaws introduced during
58 3 The Integration of Materials into Product Assurance Schemes

Fig. 3.1 a Early on-board

activities during the Spacelab-1
mission. Crew members Robert
Parker and Ulf Merbold carry out
a ballisto cardiography
investigation. The brush Alodined
weld zones on the aft end cone
are arrowed (see also Fig. 8.42).
Post-flight inspections of these
accessible surfaces showed no
evidence of degradation by
corrosion. (Photo courtesy of
NASA.) b Back-dropped by the
rugged Earth terrain, the
International Space station is
photographed by an STS-130
crew member on the space shuttle
Endeavor after the station and
shuttle began their
post-undocking relative
separation. Credit NASA

processing of the basic material, or fabrication of piece-parts,
it is essential that analyses and tests are performed to establish
the limits of detection of these flaws, using nondestructive test
(NDT) methods such as eddy currents, ultrasonics, dye pen-
etrants, and radiography. Again, metallography is employed
to quantify the actual sizes and forms of cracks under NDT
investigation. Fracture-control screening is generally per-
formed for the complete spacecraft structure. The structural
items are those parts and assemblies that have been designed
to sustain loads or pressures, provide stiffness and stability, or
provide support or containment. These may be: pressure
vessels, composite structures, joints, and other load-bearing
3.1 General Product Assurance and the Role of Materials
parts. ESA space vehicles, and, when implemented, are also
able to satisfy the NASA approval authorities (by means of
the standard that denes the project safety programme and its
safety technical requirements; ECSS-Q-ST-40).
3.1.4 Materials and Processes
Spacecraft project management will follow the contractual
requirements formally agreed between customer and usually
the prime contractor. Sub-tier contractors will often have those
same requirements flowed down in order to ensure a coherent
approach. Materials and processes will be well-dened and
cover such topics as selection criteria and rules; evaluation,
validation and qualication testing; the generation and
maintenance of Declared Materials, Mechanical Parts and
Processes Lists; and procurement and receiving inspection
requirements. In Europe, these requirements often follow the
ECSS-Q-ST-70 standard, the latest issue should be consulted
for details. Full adherence to the ECSS-Q-ST-70 requirements
is preferable, but may be too demanding for some low cost
projects such as those carried out by universities where cost is
the driving factor rather than highest reliability. Appendix 8 is
proposed for the basic management of low cost Materials and
Processes activitiesthis is a rudimentary scheme that may
need to be enhanced as the project develops.
Examples of ctitious Declared Materials Lists and
Declared Processes Lists are shown in Appendices 10 and 11.
The utilization of such DMLs and DPLs are considered
essential by the writerthey follow the hereditary of the
space subsystem and can be immediately recalled in case of
failures when material composition will need to be identied.
The majority of materials used in spacecraft manufacture
are selected from commercial metallic alloys, ceramics and
polymers that can be processed using conventional tech-
niques. This is due to four main reasons:
Contractors prefer to stick to tried and trusted
technologies.
The validation of new state of the art materials, or the
verication of new processes for a given space pro-
gramme, usually require much additional effort (and
costs) in terms of generation of quality documents, pro-
cess controls, reliability test data, etc.
The quantities of materials utilized on spacecraft are
measured in kilograms and not tens of thousands of tons
as would be preferred by commercial material producers.
The time from the start of a new space project until its
launch is about 4 or 5 years, and contracts may be too
late for any new material development.
Nevertheless, during the past decade we have seen the
development of new, advanced materials for use in structural
59
applications, mechanical devices, propulsion systems, and
electronics. As previously stated, the ordinary materials
known and well-proven in aircraft applications are most often
selected, but it must not be overlooked that they are now
being fabricated by revolutionary processes that develop near
net shape, by 3-D printing, by special heat treatments for
optimized grain size and with unusual surface protection
layers. Many changes can be seen during visits to the fac-
tories and workshops involved with the manufacture of space
hardware. Often the remnants of outdated methods are visi-
ble, capstan and turret lathes lie idle in corners or are seldom
used by the skilled persons needed to operate them. The
newest CNC (computer numerical control) turning machines
are now in the spotlight with driven tools, automated by
gantry robot-handling systems and many other newly avail-
able accessories for unmanned operations. These facilities
have been purchased not for high-volume runs, where they
can be left permanently set up for one part, but for
multi-operations where lathes have great flexibility and can
produce the large variety of turned parts required for space-
craft. The tool magazines may contain up to 90 different tools
and, by means of interactive programming, even inexperi-
enced operators are able to soon learn how to programme
complex work pieces. The small batches of turned compo-
nents having many shapes, as are required for space projects,
can now be made on a single turning machine where the
metal input is in bar or billet form. In conventional machining
the use of multiple machines has become inefcient for small
batches. Furthermore, when high surface nishes and close
tolerances are required for spacecraft mechanisms, grinding
can be made with a new generation of tools for instance
having CBN (cubic boron nitride) inserts. Ideally, the com-
plete grinding sequence can be done in one operation to
reduce time and handling. Swiveling grinding heads are
available for aerospace parts having powerful 40 kW grind-
ing spindles which offer a speed range of up to 9000 rpm.
A compensating system is available for high-accuracy parts,
which negates the effect of thermal expansion. These are
being used for the manufacture of lightweight spacecraft
mirrors. Depending on the alloy being turned or ground, a
standard coolant or oil can be used. This is supplied under
high pressure, and the coolant vapour is collected within the
enclosure of the working area by a vacuum system. It can be
said that modern technology has redened the lathe and
grinder, with 8760 working hours every year these new,
automatic machines can continued without the need for
sleep. Joining techniques have also advanced and whilst, for
instance, the main structural alloy All 2219-T851 for the ESA
Spacelab project (Fig. 3.1a) was generally TIG-welded or
electron-beam-welded in the early 1980s, the same wrought
alloy is being jointed by the recently evaluated Variable
Polarity Plasma Arc (VPPA) welding, providing far superior
weld quality for the Columbus/Space-Station elements
60
(Fig. 3.1b) and the Ariane V propellant tanks (see also
Fig. 4.57). It was thought likely that VPPA welding would be
the only suitable process for welding the wide range of alu-
miniumlithium alloys that have recently become commer-
cially available (Ilyushenko 1993). Conventional welding
techniques for AlLi alloys produce unacceptable levels of
porosity due to the presence of lithium compounds that may
be hydrides, hydrocarbonates, and hydrated oxides. During
welding, such compounds decompose with the evolution of
gas bubbles. They are entrapped by the metal being solidied
and form pores. Nowadays Friction Stir Welding (FSW) has
become the preferred method for joining AlLi space struc-
tures. The longest such weld being 445 inches, by Lockheed
Martin on the Orion spacecraft (for joining the forward cone
assembly to the aft barrel assembly). Developments in FSW
at the NASA Marshall Space Center have substantially
reduced weld defects so that costly and risky repairs on Al
Li, superlight spacecraft tanks and structures (Russell 2014).
FSW is termed a key technology for primary aerospace
structures in Europe and this clean, cost effective joining
technology has been transferred to the manufacture of many
European spacecraft tanks and structural systems (such as
AlLi cryogenic LH2 and LO2 tank demonstrators, the upper
stage of Ariane V and its main motor thrust frame (Brooker
2001)). FSW is described in more detail in Sect. 4.25.
The following is a generalized account concerning the
acceptance of a new alloy product for application in a
spacecraft:
If the new metal alloy, no matter how promising, is
developed by an engineering company, it will not be used
on a space project unless an individual, or team of mate-
rials and design engineers related to the project, recognizes
its potential merits. They need the courage to evaluate it,
and use it if the new alloy lives up to its promise. That
alloy, as with all previous metals selected for space use,
would have been subjected to a sequence of evaluation
tests specically called up by the various technical disci-
plines of the projects PA team. The Materials and Pro-
cesses (M&P) engineer reviews the evaluation test results
and also the conclusions and recommendations given by
the other PA engineering disciplines. The new materials
will only be approved by the M&P engineer once he is
convinced that the alloy is totally suited for its application
and that its method of fabrication could not adversely
affect the spacecrafts integrity. Metallography will be
utilized during all the evaluation tests (probably involving
stress corrosion tests, accelerated fatigue, testing for
hydrogen embrittlement), and this technique will often
establish certain test acceptance criteria.
The imposition of an M&P policy is occasionally met with
some reluctance by industry, and only the larger space
3 The Integration of Materials into Product Assurance Schemes
projects employ dedicated materials engineers for this
function. To be effective, each spacecraft material and
process must be documented and then approved by the
M&P engineerthis task will be facilitated by the exis-
tence of a specialist support team having access to an
adequately equipped and manned materials test laboratory.
Concerning processes, it is probable that the introduction
of computer-aided design now means that spacecraft
sub-assembles and piece-parts are being fabricated to the
closest tolerances ever achieved. To illustrate the accu-
racy demanded of modern machining capabilities one can
consider the unfortunate situation of the flawed primary
mirror of the Hubble Space Telescope (HST). The prime
objective of the HST mission was to obtain images of
astronomical objects in approximately ten times sharper
detail than that obtained by ground-based telescopes. Or,
in other words, HST is designed to point on a celestial
target with a stability measured in tens of nano-radians:
roughly equivalent of keeping the HST line-of-sight on a
human eyeball at a distance of 320 km. The HST 2.4 m
mirror was designed to be a precisely calculated hyper-
boloid. Although the rst mirror flown was actually
smooth to a precision of 1/64 the wavelength of light (or
one-millionth of an inch), a calculation error had caused it
to have been fabricated with a curvature that was too
shallow with a total centre-to-edge error of about 2 m
(1/50 thickness of human hair). Results from the mirror,
during the period 19901993, showed that light rays
hitting the mirror edges eventually focused to a point that
was slightly away from where light rays from the centre of
the mirror focus: a defect called spherical aberration. The
shuttle mission to revisit the HST and to install new solar
arrays and new instruments, and to correct the defective
mirror was made successfully in December 1993.
The space industry places extraordinary demands on the
properties, performance and cost of materials as these are
intertwined with considerations of safety and reliability. In
this book, the testing and evaluation of materials will mainly
focus on the various phases of selection of individual
materials for a particular design based on properties and their
performance in a particular environment. Cost may be
important, but generally the best material for the job will be
selected bearing in mind that the price of procured materials
for a complete space project will probably only lie within 1
and 10 % of overall costs. As will be noted in the chapters
relating to failure analyses, it is often the environment that is
ultimately responsible for any particular hardware failure.
These environmental factors can be extraordinary difcult to
comprehend and it is essential that engineers from all dis-
ciplines collaborate to both prevent, and where necessary,
solve problems associated with failures. The complexities
3.1 General Product Assurance and the Role of Materials
associated with launcher and spacecraft failures necessitate
an interdisciplinary approach to failure analysis where the
materials engineer will play an important and active role.
Two immediate examples come to mind:
The Space Shuttle Challenger accident in 1986 was ini-
tially observed to result from puffs of black smoke
originating from a leaking lower seal joint on its
right-hand side metallic Solid Rocket Booster (SRB).
The seal, actually a complex design incorporating an
elastomeric O-ring, is needed to prevent hot gasses
escaping from the highly pressurized rocket motor during
take-off. Within a minute of launch a flame grew from the
leaking pressure seal, the flame impinged on the sur-
face of the External Tank causing it to breach and release
hydrogen. Total failure occurred within milliseconds
once the adjacent oxygen-containing Intertank was also
fractured. The cause of the accident was determined to
result from a faulty design unacceptably sensitive to a
number of factors: out-of-round dimensions between the
two metallic segments of the failed joint (they has grown
due to previous useage of this reusable SRB, but were
still within tolerance); the elastomeric O-ring seals
resilience at an unusually low launch temperature (the
Florida launch site was at 1 C, whereas most launches
occur at 24 C when the O-ring is ve times more
responsive to return from its compressed to its uncom-
pressed shape); the possible presence of ice within the
joint volume (it had been raining before launch and later
tests showed ice formation to be a possibility); stresses
during SRB ignition and ring can slightly open the
seals tang and clevis and, at the low temperature, this
gap may not have been lled by the O-ring due to its
lower response time to reach the uncompressed dimen-
sion. Elastomers shrink and loose elasticity at low tem-
peratures. The mechanical properties of the O-ring
material were known for all temperatures of operation.
However, the importance of the change in seal gap
dimension, together with the reduced elasticity of the
O-ring material to ll the gap at the low environmental
temperature, were not realised by all concerned in the
launch decision.
The other example, in 2002, concerns the failure of the
Ariane 5 ECA rocket (Flight 157) when it suffered major
trouble with its Vulcain 2 main stage engine before going
out of control and exploding over the Atlantic Ocean.
Failure analyses reported by Drogoul (2005) found the
problem to have been caused by a deformation to the
engines nozzle (Fig. 2.13 shows the engine and location
of a helical structure of welded Inconel tubes through
which hydrogen flows to cool the nozzle). Ground test-
ing, as shown in Fig. 2.13, was of course conducted at
61
atmospheric pressure and, during such tests, the nozzle
was found to be subjected to axially tensile forces. The
failure mode that led to the loss of Flight 157 was a
decrease in pressure around the nozzle as the rocket
gained altitude. That loss in pressure during launch
placed the nozzle under gradually increasing axially
compressive forces and this resulted in a deformation of
the upper portion, causing it to buckle. Distortion of the
nozzle increased the thermal flux in a localised area until
the nickel-based alloy melted and the cooling system was
perforated. This launcher was immediately destroyed by
a command from the range safety ofcer. The main les-
son learned from this experience is the fact that while the
materials may have been correctly selected, qualication
authorities were unable to reproduce the flight environ-
ment. The reducing atmospheric pressure could not be
included as part of a launch simulation.
3.1.5 Component Part Selection,
and Procurement
The work of the component engineer is analogous to that of
the M&P engineer in that electronic components will be
selected and procured from the types already listed as either
having been tested and qualied for space use, or those
nonstandard devices that will require qualication by testing.
The major objective of the component engineer will be to
ensure that electric, electronic and electromechanical parts
(EEE parts) are selected and procured for space projects,
whether these are launch vehicle, space probes or satellites.
The format for compiling component data as a basis for
approval of a component, or EEE part, for a space project is
by means of a continually updated and assessed Parts
Approval Document. Analogous to the Declared Materials
List, components engineers will review each projects
Declared Components List (DCL), identifying all compo-
nents needed for the project, their full specications, quali-
cation data and approval rating. Screening and Lot Approval
testing will be specied and particular attention will be given
to radiation verication testing (RVT) with respect to the
anticipated dose calculated for the component during its
particular in-orbit or exposure life to electromagnetic radia-
tion (i.e. solar radiation consists of ions and electrons plus
trapped electrons and protons in the van Allen Belts and in the
South Atlantic Anomaly, see also Sect. 8.2.2). These and
other forms of radiation data, including cosmic rays and solar
proton events, both able to knock out spacecraft, can be read
online from sites such as http://spaceweather.com and http://
www.spenvis.oma.be/. Excepting the specic electrical tests,
most components will be subjected to the same vibrational,
shock, and thermal cycling tests as were described in
62
Sect. 3.2.2. As mentioned, tests may be required involving an
assessment of electromechanical devices when degradation
resulting from radiation exposure may be expected (the
in-orbit levels of electromagnetic, cosmic, and particle radi-
ation do not affect metals). All classes of components ranging
from the relatively simple resistor to the highly complex
microprocessors and memories will require to be evaluated
by a comprehensively equipped laboratory. In addition,
advanced test equipment is needed for the characterization of
very large-scale integrated circuits.
Space-qualied components are generally covered by
ESA SCC specications. These specications also control
the selection of peripheral parts for a spacecraft, such as
connectors, wires, cables, and fuses. It is no use sending a
multi-million pound spacecraft into orbit if, during its launch
phase, connector pins break, fall out, or wires fracture. Here
the materials specialists will assist the components engineers
in the choice of materials and processes best suited to sur-
vive long-term duration in space. With the ever-increasing
complexity of space systems and the demands for a reduc-
tion in the size and weight of interconnection devices (e.g.
for the ESA Cluster spacecraft the 50 km of harness was
made from AWG 28 stranded wires) there has been a need to
make even more stringent materials tests to ensure the
durability and electrical integrity of interconnectors.
Micro-D connectors have been developed specially for the
space market. They are small and light, and have higher
density contact congurations than are available in tradi-
tional rectangular connectors. Some have aluminium alloy
shells with a nal nish of gold plating; these are nonmag-
netic and offer electromagnetic and radio frequency shield-
ing (beware of commercial connectors which are totally
unsuited for space applications owing to their
cadmium-plated steel shell, sublimation, and magnetism
problems). The new designs also possess a silicone elas-
tomer compression interfacial seal to provide a moisture and
humidity seal between each contact and the contacts and the
shell (the silicone passes outgassing requirements whereas
the organic insulants incorporated in commercial connectors
can outgas under vacuum).
The design and selection of materials for electronic boxes
rest with the industrial companiesand the specic space
project contractual requirements. Usually the housing is
machined and/or assembled from a stress corrosion resistant
alloy listed in Table 2.2. Care must be taken to avoid
unwittingly incorporating nickel plating on the surfaces of
any electronic box designed for magnetically clean scientic
spacecrafta costly problem involving the use of electro-
less nickel is described in the case study presented in
Sect. 4.10.2. Electronic boxes will contain runners, frames
and spaces that locate component-populated printed circuit
boards. Boxes may only house 3 boards, but some complex
boxes have been known to house up to fty boards
3 The Integration of Materials into Product Assurance Schemes
together with connectors, hook-up wires and cables and
coating/potting materials. Such boxes are heavy and require
special handling procedures. Boxes have been known to be
constructed from aluminium, magnesium and steel materi-
als. The thickness of the walls will need to satisfy
mechanical design requirements to ensure no damage during
vibration and other loading. Radiation analysis should also
be madeoften the wall thickness is selected so that total
radiation within the box will not exceed 10 kRad(Si).
Typically 2 mm thick aluminium is presently used as the
norm for shielding electronic componentsas some FPGA
packages can be badly affected with less than 100 kRad
dosages. Shielding studies are continually performed in an
attempt to nd a light-weight solution because of the
inherently high mass of traditional aluminium boxes. Some
ideas are to use CFRP composites that contain carbon
nanotubes together with diamond like coatings deposited
onto tungsten foils. Aluminium alloy boxes are often surface
treated with Alodine 1200 and later, nally painted black or
white with an Aeroglaze or MAP product. Electrostatic
precautions are always necessary to avoid ESD and
operators/handlers need themselves to be grounded by
means of wrist straps and conductive shoes.
Wires and cables used in electrical applications such as
cable harnesses and wires within electronic boxes are strin-
gently controlled for space applications. A multitude of
variants exist and the space-approved types are specied by
ECSS documents. These specications dene the require-
ments for a particular family of wires or cables. They include
the maximum rating, physical characteristics, design (i.e.
shielded or not, AWG wire size, number of strands and
physical dimensions). Extensive qualication tests (involv-
ing accelerated ageing at 250 C, mechanical tests such as
abrasion and notch resistance, solderability, possibly
flammability, etc.) are performed according to the relevant
detailed specication in the ESCC 3901/xx series. The
most-use conductors for larger gauge wires (AWG 1624)
are silver-coated, soft or annealed oxygen-free high con-
ductivity copper. The smaller gauges (AWG 2630) usually
require silver coated high strength (alloyed) copper strands.
The European space standards require that the silver has a
minimum thickness of 2 m in any one location; this is to
avoid the problem of red plague as will be discussed in
Sect. 6.8.3. In recent years silver-plated aluminium con-
ductors have been selected for power distribution systems on
satellites where weight saving has been essential. Insulation
materials are generally, polyimides, cross-linked ETFE, and
expanded-PTFE. For low Earth orbit (LEO) the cable jacket
must be resistant to atomic oxygen and a dual insulation of
polyimide/fluorothermoplastic has been found suitable.
In recent years, as system integrators attempt to manage
higher reliability from their system components together
with the need for greater cost-savings, there has been much
3.1 General Product Assurance and the Role of Materials 63

interest in the use of commercial off-the-shelf components

(COTS). Components utilized on small satellites, cubesats,
and for the assembly of university experiments being flown
on the ISS and on other spacecraftwhere failure of the
experiment will not jeopardise the host crafts missionare
reliant on COTS devices.
Similar cost restrictions are met by the use of SCOTS
(Spacecraft commercial off the shelf) devices for high
reliability space-systems. An example are the sun-sensors
assembled by one SME; these must retain a very high reli-
ability and a new philosophy is emerging, so far with an
excellent prognosis. Instead of using commercial compo-
nents, high-rel EEE parts are selected and specially designed
quad photodiodes are procured. It is these sun-sensors that
are manufactured in large production numbers, as seen in
Fig. 3.2, with the ultimate aim of lowering costs for an
expected market (Leijtens 2015). The units will be stored in
clean-room conditions within special containers, under an
inert gasalthough not literally off-the-shelf, they can be
ordered as an existing product and dispatched within a
couple of hours. In producing multiple units the company is
required to undergo enhanced reliability and qualication
testing. Environmental testing is made at comparatively high Fig. 3.2 a BiSon 64 spacecraft sun-sensor (providing 64 degrees eld
levels and on many more samples. As a brief example: of view) is a recurring product manufactured at low cost, in high
numbersconsidered a hi-rel SCOT. b Consideration of total costs
Radiation testing includes: a total dose of 1 Mrad (on versus unit costs for increasing number of units manufactured in the
same time period (courtesy of Lens R&D, the Netherlands)
bare diodes) and 1016 1 MeV electrons (on bare diodes).
Mechanical environment includes constant acceleration, It is frequently noted that in industry both Component
transient, sinusoidal and random vibration, acoustic noise and Materials laboratories are physically located adjacent to
and pyro-shock loading such as: 30 g random 3-axis each other. Each laboratory will possess its own dedicated
(20 mm minimum 5 and 100 Hz, 1 oct/mn up and down), equipments, but the one common facility to be utilized by
37.9 g random 3-axis (20100 Hz +6 dB, 1 g/Hz, 1000 both disciplines is that concerning metallography (scanning
2000 Hz 6 dB for 3 min); pyro-shock 10.000 g 0.5 ms; electron and optical microscopes, mounting and polishing of
Eigen-frequency measured to be >2000 Hz (rigid body). sample cross-sections, etching to reveal microstructure, etc.).
Thermal cycling: 1,000 cycles between 45 C and +85 All component types and wires/cables will, at some stage, be
C, (or 30,000 double sine testing, where the base cycle is submitted for metallographic evaluation, either as part of a
from 20 C to +60 C and there is a modulation of +20 C verication DPA or as part of a failure analysis. It is here
and 20 C on top of the base cycle in order to mimic real that the materials engineer will work with the electronics
operating conditions where a spacecraft might encounter a engineer in solving problems in an effort to bring better
slow cycle over a year and a fast cycle every day). quality and reliability into a product.
Mate and de-mating of connector 3000 times (usual
specication is 500 times).
3.1.6 Control of Ground-Handling Facilities
Once these samples have been tested they will be further
evaluated in a Components/Materials laboratory by physical One of the most critical safety items associated with the
destructive analyses (DPA). The onset of damage to com- ground handling of spacecraft and their heavy subsystems is
ponents, interconnections and housings may be found, and if the wire ropes installed on hoists, drum winders, and cranes.
items have fractured some failure mode analyses will be These ropes suspend the costly space structures and transfer
documented. Such inputs will be essential to specialist them from integration halls onto transportation platforms
component and reliability engineers attending review boards (pallets), then further to test facilities and nally launch sites.
covering qualication and flight acceptance of the tested The hoisting of some large spacecraft structures at the
product. ESA-ESTEC test facility is seen in Figs. 3.3, 3.4, and 3.5.
64 3 The Integration of Materials into Product Assurance Schemes

Fig. 3.3 Hoisting of the Eureca

spacecraft into the large space
simulation chamber at ESTEC.
Note that the steel ropes are
covered with plastic as control
against contamination

Fig. 3.4 Ground handling of the

Hermes spaceplane dummy
3.1 General Product Assurance and the Role of Materials
Fig. 3.5 Preparation of the
Ariane V half-fairing
qualication model, for acoustic
testing at ESTEC
The most usual lifting gear uses steel wire, although some
systems requiring less load-bearing capability favour the use
of synthetic bre ropes and plastic-covered wires. Synthetic
ropes are made of Kevlarthese now compete with steel
wire and cable, as they have a lower weight and can be
operated using decreased motor and hoist hardware sizes.
Similar, braided Kevlar is used for tethering small satellites
or space experiments to the Space Shuttlethis tether has a
diameter of less than 2.5 mm and can withstand loads of
180 kg (approximately 700 MPa) as the hardware is
deployed from the Shuttle and back by means of a simple
reel (Scala 1996). In comparison, the steel wire strands that
are in common use for hoisting applications will have
strengths in the order of 2000 MPa, although ropes made of
even higher strengths are available (Shipley 1988).
Wire rope is constructed from twisted bundles containing
strands of small wires having the same cross-sectional area
of metal. Whereas a solid wire, for example having a
diameter of 10 mm, is very stiff and will stand relatively little
back-and-forth bending before breaking, the same aggre-
gate sectional area of metal in stranded form will be found to
have gained greatly in flexibility, tensile strength, and life.
65
By changing the number and arrangement of wires in each
bundle (for instance by making ropes with ne metal bre
cores, by laying the bundles helically with either right lay or
left lay, or selecting higher-strength steels for the outer
bundles in order to achieve optimum wear resistance) it can
be seen that the number of possible wire rope combinations
becomes profuse. However, only a small fraction of these
constructions are sufcient for industrial purposes in order to
provide strength and stiffness with flexibility in bending.
Expensive alloy steels with their more complex heat treat-
ments are seldom selected for slings and holding devices.
However, they are seen in the cleanrooms of spacecraft
integration halls, as these alloys do not corrode and are less
likely to generate dust or wear particles. In less clean envi-
ronments these wires are essentially a plain carbon steel of
carefully controlled composition and a carbon content gen-
erally close to the eutectoid. The patenting process of
isothermal transformation and work hardening resulting
from successive wire drawing is largely responsible for
achieving the 2000 MPa strengths.
The spacecraft and materials handling systems such as
hand hoists, power hoists, wall and floor cranes, grabs,
66
slings, and holding devices must be fully maintained to
written procedures that also include inspection steps for
signs of wire spreading, wear damage and overloading.
Maintenance schedules are attributed to wire ropes which
will be checked possibly at least once per month for signs of
marring, stretch, and other deformations. The nature of rope
construction provides many parallel load paths, giving it a
high level of redundancy. The essentially helical nature of
rope construction usually causes the outer bundles or strands
to be marginally greater stressed in service than the inner
strands and these will be preferentially broken in situations
of overload. Strands in the bundles then slide slightly under
surface shear forces to take up the applied load from the
broken strand. There is no stress concentration to promote
the fracturing of adjacent strands, and the location of the
break can be readily found during routine visual inspections.
Such inspections will also check for damage from wear,
crushing, corrosion, nicks, and cuts. Owing to the often
harsh duty cycle of hoisting ropes it is inevitable that some
degradation will occur over time. It is the ground-handling
managers responsibility, guided by the materials engineer,
to monitor degradation and remove rope and lifting gear
before it becomes a risk to both the facility operators and the
spacecraft hardware. The author is not aware of any space-
craft handling failures attributable to rope or cable structures,
even though these are subjected to high load capacity
combined with flexibilityprobably due to the inherent
redundancy and damage tolerance of the numerous single
woven wires. Other commercial industries have suffered
from catastrophic steel cable failures, most often attributed
to corrosion and local fatigue (Piskoty 2009).
3.2 The Materials Laboratory
3.2.1 Major Objectives of Laboratory
The company materials laboratory can represent over one
million pounds of investment in facilities that enable the
provision of services to spacecraft projects and associate
R&D programmes. It is preferable that the work is con-
ducted as an integral part of the PA effort described in the
previous section, and this will include such topics as the
selection and approval of metallic materials for spacecraft
structures, the assessment of electronic materials for specic
applications, the qualication of critical spacecraft pro-
cesses, the testing and approval of novel fabrication and
surface-coating techniques, corrosion control, and the relia-
bility testing of electrical interconnections.
3 The Integration of Materials into Product Assurance Schemes
The laboratories, which perform testing and failure
analyses, are run to give rapid functional support particularly
in the form of reporting to Critical-Design Reviews, Material
Review Boards, and Failure Review Boards. The items of
spacecraft hardware that may be submitted for evaluation
can range in size from thermocompression weldments that
interconnect 25 m diameter gold wires within miniature
electronic circuits, to the massive 4 m diameter forged and
heat-treated Spacelab structural rings, the largest to have
been fabricated in Europe from a high-strength aluminium
alloy.
The role of metallurgical tests in the somewhat overlap-
ping range of spacecraft project PA activities can be illus-
trated by the following list:
Quality assurance (QA): Metallography and mechanical
testing for the setting up and control of optimum process
variables (such as those encountered during welding, braz-
ing, heat treatment, electroplating, etc.), and for the gener-
ation of workmanship standards. In-depth testing and
qualication programmes for the approval of industrial
fabrication lines (such as for double-sided and multilayer
printed-circuit-board production, automatic machine solder-
ing, and critical process assessment).
Reliability and safety: Hardware or unit-level testing
directed toward ensuring that design, selection of materials,
and choice of fabrication methods are based on sound
principles of reliability, engineering, and safety analyses.
Prediction of lifetimes, failure-mode analyses, and identi-
cation of single-point defects. Hardware is subjected to
both static and dynamic tests (e.g. high-temperature ageing,
thermal cycling under vacuum, vibration, and shock) fol-
lowed by Destructive Physical Analyses (DPAs), as part of
the reliability test programme.
Materials and processes (M&P): Development and test-
ing of new materials and processes before their being
ascribed to spacecraft production, screening of metals for
specic applications, and the review of certain QA functions
(listed above). Metallography is utilized during the majority
of evaluation tests (possibly following material or intercon-
nection exposure to corrosion, mechanical loading, acceler-
ated fatigue, testing for hydrogen embrittlement, or a
combination of these environments).
Components part selection: This activity is analogous to
that of M&P in that electronic components are selected from
those previously tested and approved for space use and those
nonstandard devices that need to be qualied by testing.
Metallurgical evaluations complement work performed by
components laboratories in areas related to glass-to-metal
sealing, packaging, microwire bonding, solderability, elec-
tromigration, and the formation of whiskers.
3.2 The Materials Laboratory
3.2.2 Facilities and Instrumentation
3.2.2.1 Macroscopic Examination
Macro-examinations relate to material features that are
apparent to the naked eye, although in practice magnica-
tions of up to ten times may be used. A number of
stereo-zoom microscopes and cameras exist to reveal and
record surface morphologies such as cracks, voids, inclu-
sions, deformation, blistered coatings, de-wetted soldered
joints, contaminants, dust particles, and the like. An example
of the use of macro-examinations is shown in Fig. 3.6; the
morphology of the titanium propulsion tanks internal weld
bead was initially macrophotographed before being sec-
tioned and polished to reveal the weld macrostructure.
A good magnier and light source are needed to perform
inspections efciently and to incur as little eyestrain as
possible. Traditional overhead lighting, either fluorescent or
incandescent, is not appropriate for most inspection tasks
because each produces glare that interferes with the viewing
of small parts and details. The best light source is adjustable,
on an articulated arm, so it can be directed to enhance
contrast and colours. It should be independent of the mag-
nier lens to reduce glare and reflection from metallic
specimens. The magnier having a circular lens works well
for small objects, but if large numbers of production items,
such as printed-circuit-boards, need to be inspected it is
preferable to initially scan each board for defects under a
rectangular or stadium-shaped lens, which permits more
visibility of the board at one time. Areas of interest can be
identied with detachable plastic markers before the board is
transferred to a stereo-zoom microscope or camera. At the
specialist end of inspection, 3-D imaging capabilities are
available. This enables volumes to be measured, and the 3-D
Fig. 3.6 Girth weld of a hydrazine propulsion tank. These macropho-
tographs highlight: a the root bead of the electron beam weld and
associated weld sputter particles, and (b) the Ti6A14 V weld
67
systems are particularly useful for inspecting for cracks,
voids, and solder paste volumes.
3.2.2.2 Microscopic Examination
A microscopic examination is performed at magnications
greater than ten diameters. It is far broader in scope than in
macroscopic examination, and consequently much addi-
tional information can be generated. Both fractography (the
microscopic examination of the irregular surfaces produced
when a piece of material is broken) and metallography (the
examination of the microstructure of metals and alloys as
revealed by cutting, polishing, and etching a sample through
the plane of interest) are used. Correct observation and
interpretation of the etched microstructures depend heavily
upon the ability of the examiner to interpret what is revealed.
An excellent reference dedicated to the systematic prepara-
tion of a vast range of material surfaces and a comprehensive
overview of microstructural analysis has been prepared by
Bouseld (1992, 1997). Reagents of microetching metals
and alloys are detailed in Appendix 3.
The metallurgical laboratory microscope shown in
Fig. 3.7a) is supporting part of a multilayer printed-circuit-
board that has been mounted in a room-temperature-curing
plastic and then ground, polished, and etched to reveal the
cross-section of a plated-through hole. The area of interest
can be viewed through the eyepieces or, as shown, projected
onto a viewing screen. Attachments exist for making digital
imaging and storage, 35 mm slides or photographs and, now
obsolete, Polaroid instant prints. Other accessories available
for these optical microscopes include an interferometer,
polarized-light illumination, and a micro-hardness tester.
The traceability and recording of all laboratory samples
such as mechanical test pieces, metallographic mounts, etc.
macrostructure, revealed in cross-section after grinding, polishing,
and etching. Vibration could cause particle detachment, resulting in
blockage of this spacecraft propellant system
68 3 The Integration of Materials into Product Assurance Schemes

Fig. 3.7 a An iconic 1970s

Reichert projection microscope
with attachments continues to be
capable of producing
magnications in the range
of 10 to 2000. b Example of
good metallurgical laboratory
practice where all documents are
traceable to these various
weld-related samples. The mounts
and samples are engraved with a
vibrating tool to permanently
show unique reference numbers
(courtesy of HTC, Foligno, Italy)
3.2 The Materials Laboratory
is extremely important. They are needed for laboratory test
reporting, and may be requested for re-examined during
future yearsparticularly if failures related to evaluated
spacecraft parts are encountered during testing of during the
long life-times required of modern spacecraft. The departure
of lab staff and re-hiring of new personnel means that the
transfer of knowledge may only be possible if good house-
keeping and recording is effective. Examples of good data
logging, and sample traceability are illustrated in Fig. 3.7b).
This subject will be further discussed in Sect. 3.3.3.
High-temperature optical microscopy is available at a
number of laboratories. This can be a very useful technique
for the study of diffusion and phase changes, particularly
those that may occur during the temperature excursions
experienced by spacecraft materials during exposure to solar
heating or during re-entry into the Earths atmosphere. The
sample is prepared, usually as a polished section, and placed
on the hot stage of the microscope. Some microscopes are
advertised as operating to temperatures of 1500 C; such
temperatures are essential for studying refractory metals, but
in practice the instruments rarely reach such temperatures
without the need for extensive redesigns. A full run can take
several hours so that it is advisable to record the optical
image on video, allowing repeated study of events. Investi-
gations where the hot-stage microscope has provided useful
data have included: accelerated growth of the intermetallics
that form between gold and tin when solder alloy is applied
to gold-plated terminations; the effect of dehydration on the
densication and cracking of chemical conversion coatings
such as Alodine 1200 on aluminium substrates; viewing the
crack-opening displacement during the thermal fatigue of
solder joints resulting from z-axis expansion of
printed-circuit boards (see also Fig. 6.17); and observing the
preponderance of adjacent phases in a metallurgical system
to tarnish, oxidize, and become less reflective.
In many situations even greater magnications are useful.
This capability is provided by the Scanning Electron Micro-
scope (SEM) shown in Fig. 3.8, which can magnify to 200
000 with good resolution. The great depth of eld of this
instrument has made it particularly useful for examining
rough fracture surfaces to determine the cause of a component
failure. The majority of SEM investigations involve electri-
cally conductive objects which need no special preparation.
However, when electrically nonconducting parts are exam-
ined in the SEM they can become charged and this causes the
image quality to suffer. It is therefore necessary to coat such
objects with a very thin layer of conductive material, either by
sputtering or by vapour depositiongold is most often used
for this purpose. Recently designed SEMs can operate at a low
cathode voltage, down to 1 kV or less, and with renements in
controlling the equipments beam path, this has the effect that
the electron beam hardly penetrates the surface of the material
69
under investigation (Pohl 1996). With these improvements it
is found that even nonconducting surfaces can be examined,
without the need for coatings. Ceramics, plastics, and com-
posites can now be examined for minute surface features
having a good lateral resolution down to 4 nm. Occasionally
there will be the need to resolve precipitate particles within the
matrix of a metallic sample, grain boundaries, and dislocation
defects. These can be viewed by means of the Transmission
Electron Microscope (TEM); either the sample must be
thinned by special electrochemical thinning techniques, or
thin lms of carbon or polymeric material must be used to
produce a replica of the sample surface and it is the replica that
is used for the examination. The TEM has a very high
resolving power, as illustrated in Fig. 3.9a. More recently, the
Scanning Transmission Electron Microscope (STEM) has
been developed for thicker samples, and this instrument can
also produce diffraction patterns, so enabling the characteri-
zation of crystal structures.
There are several occasions when the materials laboratory
may be requested to examine organic materials and even
sample types such as bacteria, algae, spore, fungal growths,
and human hair. The latter types may occur as particulate
contaminants inside cleanrooms, and even if present in low
levels, they can be harmful during the production of
advanced electronics such as integrated circuits (ICs). They
may be present as human-sourced contamination capable of
degrading wafers and producing on-die contamination so
reducing the reliability of ICs and hybrid packaged elec-
tronics. As part of the Space Station life support activities,
research is progressing to nd alternative sources of oxygen
for the manned modules. One source is the Microbial Eco-
logical Life Support System Alternative. In this system,
different bacteria and cyanobacteria are grown in a culture
and used to convert organic waste and carbon dioxide into
food and oxygen. The bacteria are being examined in the
authors laboratory, optical microscopy is satisfactory for
making size and shape determinations, but SEM examination
is needed to see details (see, for instance, Fig. 3.9). Here, the
bacteria are collected on a 0.4 m cellulose membrane lter,
washed with water-soluble solvent, and then allowed to dry.
The nal stage is to sputter-coat the lter with a thin layer of
gold and examine using low-energy electrons (about 5 kV) so
as to inflict minimal damage to the organic specimen.
Recent experiments (Thiel 2014) have surprisingly
revealed that DNA can survive the extreme conditions of
space and the high temperatures (1000 C) of re-entry to
Earth through the atmosphereraising the possibility of
molecules of extraterrestrial life arriving from space. Plas-
mid DNA, circular strands found naturally in bacteria, was
formed into 12 samples and applied to the surfaces of a
sounding rocket payload and its screw threads. These sam-
ples were examined before and after flight using a scanning
70
Fig. 3.8 Scanning Electron Microscopes have evolved, but early
models, a are still put to good use in laboratories where they have been
the workhorse in materials investigations since the 1960s. Very
modern versions such as the EVO, b can be used in studies of organic
materials and life science samples at nano-scale interaction (at different
temperatures, pressures and humidities), they capture outstanding
topographical details at low voltages. Such state-of-the-art SEMs and
their sophisticated electron columns, are useful for electronic materials
work: they enable high resolution imaging up to 0.8 nm (scanning
transmission mode) and nano-point analysis. Space materials scientists
can gain maximum information from a sample with the on axis in-lens
electron microscope at the University of Zurich. 35 % of the
DNA retained its biological function. Clearly, these results
suggest that even greater sterilization and other safety pre-
cautions should be taken to prevent carrying terrestrial DNA
to other moons and planets. They also suggest that DNA
from outer space, could enter the atmosphere by means of
meteorites.
New techniques for the microstructural examination of
metallographic specimens are continually being developed.
Two recent methods are the Scanning Tunnelling Micro-
scope (STM) and the Atomic Force Microscope (AFM).
The STM functions by scanning a ne-pointed needle across
the specimen surface with a distance from the surface of a
few tenths of a nanometre (nm). A potential is applied
between the specimen and the point, and the resultant current
flow is measured. The STM measures the height of the
specimens surface and can ideally examine grain bound-
aries; with multiphase materials, the aim is to remove some
nanometres less of one phase than of the others during
etch-preparation of the polished specimen.
3 The Integration of Materials into Product Assurance Schemes
Duo detector, which can be switched between ultra-high resolution
in-lens secondary electron (SE) imaging and on energy selective
backscattered detection (EsB) capable of detecting extremely small
differences in materials composition. For instance, c shows the rear-side
of a focused ion beam aperture after a few 100 h exposure, 4000.
d shows the appearance of the fused near-eutectic tin-lead nish on the
track of a printed circuit boardthe microstructural features are
contrast-enhanced by use of the SEM in back-scattered electron mode
(lead being the white dendrites), X400. bd are courtesy of T. Simpson
UWO, and Zeiss
The AFM gives similar microstructural results to the
STM. Also, a needle probe scans over the specimen in a
raster pattern with great precision. Unlike optical micro-
scopes that use mechanical gears to position the
specimen-stage, the AFM uses piezoelectric forces to posi-
tion and move the tip over the sample. The precision of these
movements is in the order of 0.01 nm. The needle-tip is
actually mounted onto a soft spring, known as a cantilever.
As the tip scans across the surface it may be pushed away
from very local areas of that surface by a repulsive force.
The AFM detects the springs movement by sensing changes
in laser light that reflects from the cantilever into a photo-
detector. The photodetector then initiates a piezoelectric
force on the cantilever to push the cantilever and needle-tip
back into its original position above the specimen. This
piezoelectric signal is converted, by a computer in the AFM,
into a topographical image or picture. The STM and AFM
are becoming increasingly used in the semiconductor
industry because they can detect surface features on semi-
conductor devices with a resolution of better than 1 nm.
3.2 The Materials Laboratory
Fig. 3.9 a Relationship between particle size, diameter, and number of
atoms for the light and electron microscope range, from McCrone and
Delly (1973). The choice of microscope for particle size measurement
is indicated. Bacteria can be examined by both SEM and optical
microscopes. b Illustration of a large cluster of bacteria under
investigation for application in manned space missions (courtesy of
D. Adams)
The AFM has evolved into a powerful tool for the
nano-scale evaluation of polymers and macromolecules
(Vancso 2009). An adaptation of this technique is the
nano-hardness tester described and demonstrated at the end
of Sect. 3.2.2.4.
Specialist laboratories involved with spacecraft quality
and failure investigations are now using modern preparation
methods in order to determine a material or components
grain structure, micro-constituents and micro-defects. They
avoid the use of a microtome for physically slicing through a
sample to reveal its cross-sectional appearance and chemical
make-up (as for instance sectioning cathode emitters seen in
Fig. 5.54). These methods also avoid the mechanical pol-
ishing of samples which can be laborious and, with certain
combinations of materials, cause damage when a hard
71
abrasive might become embedded in a void, crack or in a
soft material (such as plastic, friable coating, indium and
lead). Mechanical polishing can also introduce sub-surface
strains which, during optical examination, can result in
micro-creep causing some polished grains to lift or tilt and
produce unwanted artifacts on resulting micrographs. The
focused ion beam (FIB) microscope rasters a beam of ions
over the sample within an instrument very similar to the way
electrons are rastered in the SEM. The FIB has become a
powerful tool for the preparation of nano-samples and for the
fabrication of micro-mechanical parts. One example of a
micro-tensile test piece was seen in Fig. 2.22. Several trea-
tises covering FIB microscopy and micromachining are
given by Volkert and Minor (2007), Hassan (2014). Another
recently marketed instrument is the Argon Beam Cross
Section Polisher, here, an argon gun operating at an accel-
erating voltage of between 2 and 6 kV uses a beam of low
energy argon ions to mill the sample. The instruments stage
holds the sample and the area for sectioning is initially lined
up for the beam under an optical microscope. After the
surrounding area on the sample is masked, the sample
chamber is evacuated, the beam activated and cross section
polishing progresses for pre-determined times. This proce-
dure is semi-automatic, causes no damage to the sample
during milling, polishes composites materials with different
hardness values and has low running costs; it is said to be
more convenient than FIB sectioning as it cross sectioning
area is larger (an example is shown in Fig. 6.9). The FIB and
Cross Section Polisher instruments were recently used (Kim
2011) for the sectioning of porous WC-Co hard coatings
when mechanical polishing was found to be unsatisfactory.
Both methods gave different but complimentary results:
argon ion milling revealed the WC-Co microstructure, its
porosity and distribution of pores without damage, smearing
or deformation of this composite material. The FIB method
determined the internal interfacial microstructure between
individual particles and the softer matrix and could prepare
thin samples for further transmission electron microscopy.
Argon Ion Milling has been particularly useful for the
preparation of silver-to-aluminium bonds. Such welded
interconnections on microcircuits can easily corrode due to
humidity during service. Unfortunately they also degrade by
galvanic corrosion of the aluminum during microsectioning
when liquids cause local dissolution. As part of an extensive
destructive physical analysis (DPA) the Ag-Al bond quality
and degradation mechanism in a packaging module was
studied (Chen 2014). It was only after dry Ion Milling that
cross-sections could then be further evaluated using: Scan-
ning Electron Microscopy (SEM), Dual Beam Focused Ion
Beam-Scanning Electron Microscopy (FIB-SEM), Scanning
Transmission Electron Microscopy-Energy Dispersive X-ray
Spectroscopy (STEM-EDS) and Time-of- Flight Secondary
Ion Mass Spectrometry (TOF-SIMS)! Galvanic corrosion
72
was determined to have caused these bonds to fail in the eld
and, by Ion Milling, specic attention could be given to
microstructures and the corrosive ions that degraded the
bonds.
3.2.2.3 Chemical Analysis
(a) General Analysis in the Metallurgical Laboratory
An energy-dispersive X-ray analyser (EDAX) attachment to
the SEM provides instant elemental analysis of materials.
The instrument will identify some elements in concentra-
tions of only 20 parts per million. Elements from sodium to
uranium are readily identied. The method can be used to
analyse only small areas, but is nondestructive. The results
are generally qualitative, but a computer program can be
developed to generate quantitative data by means of standard
specimens. The primary use of this technique is the exami-
nation of unknown materials, batch comparisons, failure
analysis, and contamination analysis. Spacecraft microelec-
tronic devices, microwelded solar cell interconnections,
electrical slip rings, and optical lenses all need to be
assembled or processed in highly clean areas. The writer has
been engaged in the analysis of ionic and particulate con-
taminants that have severely degraded the quality of each of
the aforementioned space hardwarendings were that
these contaminants, and EDAX analyses determined the
chemical composition. Dried ngerprints contained sodium,
potassium, and chlorine; human hairs contained carbon and
calcium; cosmetics contained silicon, sulphur, and titanium;
and dandruff contained calcium and sulphur.
Energy-dispersive spectroscopy provides the main chemical
analysis and raises the awareness level of everyone involved
in the handling of very small spacecraft parts of the relia-
bility hazards posed by human-sourced contaminants. To
enable a more denite identication of contaminants, addi-
tional modern methods are required. A wide range of
modern chemical analytical methods, such as mass spec-
troscopy, auger surface analyses, X-ray photo-electron
spectroscopy, etc., are generally available at the sites of
larger companies, but may be accessible when required at
local universities and research institutes (for additional
information see Sect. 3.2.2.6). For a complete coverage of
the techniques for, and the chemical analysis of small par-
ticle identication, the reader is directed to the work of
McCrone and Deley (1973).
(b) Instrumental Spectroscopic Methods (Reference to Al
Li Alloys)
Whenever metal alloys are procured, it is important to ensure
that the delivered material conforms to its raw material
3 The Integration of Materials into Product Assurance Schemes
specication. This can be done as part of the procuring
companys goods inwards inspection when a certicate of
conformance or a chemical analysis record, provided by the
metal supplier, must be compared against the purchase order
requirements. In some instances when certicates of con-
formance are missing, when there is a likelihood that
materials or batches might have lost their identication and
when the metal alloy is to be used in a very critical appli-
cation or submitted to qualication testing (e.g. expensive
fracture mechanics tests and stress corrosion cracking tests),
it will be necessary to perform some simple mechanical
tensile testing and chemical composition analysis on samples
cut from the delivered alloy.
The method by which the chemical analysis is made must
be well dened and agreed with the laboratory performing
the work. As illustrated by Table 3.2 on page 102, the dif-
ferent families of the new aluminiumlithium alloys have
rather similar compositions. Clearly, the analytical technique
must be highly accurate and reproducible in order to dis-
tinguish between these alloys.
A review of the various methods for chemically analysing
aluminiumlithium alloys was made with Mr K. David of
MTS Daventry Ltd. (UK). Virtually all chemical analysis of
aluminium alloys is now carried out by instrumental spec-
trometric methods, the longer, time-consuming classical wet
methods having been superseded. Atomic emission, atomic
absorption, and X-ray fluorescence methods are most com-
monly used nowadays for metallurgical analysis, but there
are limitations to each, and the analysis of aluminium
lithium alloys does present particular difculties as sum-
marized below:
X-Ray fluorescence spectrometry. This technique is lim-
ited by the inability to analyse for elements of low atomic
weight. Although modern instruments can now analyse for
beryllium and magnesium, lithium is too light an element to
be determined by this method.
Atomic absorption spectrometry. This technique has the
advantage of being comparatively cheap, but some elements,
e.g. boron, which is found in some aluminiumlithium
alloys, cannot be determined by this method.
Atomic emission spectrometry by direct-reading instru-
ments. This is probably the most common method for the
analysis of aluminium alloys, in which a solid sample is
excited by sparking, and the resulting emission spectrum
analysed. However, instruments need to be calibrated by the
use of reference materials of known composition, and as yet
there are no suitable aluminium-lithium certied reference
standards commercially available.
Atomic emission spectrometry by inductively coupled
plasma excitation. This is almost certainly the only suitable
method for the analysis of aluminiumlithium alloys. The
technique involves dissolving samples in acid, and pumping
the solutions into a plasma, again interpreting the resulting
3.2 The Materials Laboratory
spectrum. Instruments still need to be calibrated with suit-
able standards, but as the technique involves solutions rather
than solids, these standards can be prepared as cocktails of
pure elemental solutions. These can be accurately made up
to cover a wide range of different elemental concentrations.
Modern instruments, particularly if coupled with mass
spectrometry, offer lower and lower detection limits.
3.2.2.4 Hardness Testing
A number of hardness testers are available for the testing of
metals, ceramics, rubbers, and plastics. The metals micro-
hardness tester employs a square-based, diamond pyramid
penetrator that can be selectively loaded with deadweights.
The indentation made in the test specimen by this penetrator
is measured with the aid of a microscope. The Vickers
hardness number (VPN) is calculated by dividing the load by
the area of indentation, but in practice it is usual to measure
the diagonal of the square and then consult tables.
There are two elds in which hardness measurements are
particularly important: in establishing standards and in
maintaining product uniformity (e.g. a simple hardness
check can be used to inspect the uniformity of a particular
treatment such as a forming operation, a heat-treatment
cycle, or a series of surface-hardening processes).
Hardness testing is performed at some stage during the
production of the majority of spacecraft metallic parts and
subsystems. It may be carried out as a rough check that the
correct material is being withdrawn from the stockroom, or
later during fabrication of a part, as a process control. Along
with the concept of total quality management, it has been
found useful to include the measurement of hardness in
statistical process control. Spacecraft parts range in size from
small bearings and micro-ne gold bonding wires to pres-
sure vessel weldments and very large booster motor cases.
Each of these parts must possess correct thermomechanical
treatments and surface processing to achieve the required
hardness. This is as crucial as machining operations for
achieving correct dimensional t. Hardness is related to the
tensile strengths of some metals including steel, it can be an
indication of their wear properties, it can assess the degree of
temper as a result of heat-treatment, and will conrm whe-
ther a material is homogeneous.
Large items will probably be hardness tested and the
results recorded in terms of VPN. A typical investigation is
to assess the suitability of spacecraft gas bottles, reaction
control system tanks (Fig. 3.6), and payload support struc-
tures (Sect. 4.4.4) which are constructed from Ti6A14 V by
means of electron beam and TIG welding. Test samples will
be made and subjected to various post-weld treatments, such
as annealing or mechanical stress-relief or solution treatment
followed by ageing. Hardness test samples are small and
inexpensive to investigate. They can replace the larger ten-
sile test samples which are both expensive and unsuitable for
73
use on welded pressure vessel material having a curved
prole. A typical report (Thomas et al. 1993) that incorpo-
rated much hardness testing across the weld pool and
heat-affected zones of Ti6A14 V concluded that as-welded
spacecraft components were able to meet the mechanical
property requirements and there was no special reason to go
in for any elaborate post-weld treatments.
The hardness of thin wires, coatings, and the individual
microconstituents of alloys are usually measured on pol-
ished cross-sections by means of a Vickers microhardness
tester (see Sect. 6.3 for examples) or a Knoop microhard-
ness tester (having a rhomboidal indentor). In some of the
latest hardness test machines, a means of electronic cali-
bration can be provided to maintain the accuracy to
test-block standards. However, these machines can only
take measurements on surfaces greater than about 5 m
wide. Many of the current anti-wear surface coatings, such
as the TiC-coated bearing steels discussed in Sect. 5.11,
have thicknesses in the order of 2 m. In response to the
demand to check the hardness of such thin layers, a
Nano-hardness tester has been developed.
The principle of the Nano-hardness tester is described by
its producer, the Swiss rm CSEM. Nanoindentation mea-
surements are performed with a standard Vickers indenter,
mounted on a cantilever spring. The indenter tip and can-
tilever are driven to the precise surface location by piezo-
electric actuators. The applied force is measured via the
cantilever beam by a sensorthis operation is fully software
drivenas the indenter penetrates the samples surface. The
microhardness and Youngs Modulus of the sample are
automatically recorded by means of the machines software,
and a 3-D topographic inspection of the indentation area,
prior to and after the indent, can be made by scanning atomic
force microscopy. The image and a typical force-penetration
curve obtained during the microhardness testing of a thin
gold-coated specimen are shown in Fig. 3.10.
3.2.2.5 Mechanical Testing
The mechanical behaviour of materials is characterized by
relations between the stresses and strains imparted to the
material under service conditions. A tensile test machine
(Fig. 3.11a) enables standard test specimens to be pulled
apart at different strain rates by measured loads. Load cells
may be changed so that full-scale chart readings of between
2 N and 100 kN (this is generally the maximum capacity)
can be achieved in both tension and compression (see also
Fig. 3.11b). This equipment can be used to investigate the
mechanical properties of all materials at temperatures rang-
ing from 150 C to +300 C (a custom-built environmental
chamber may be suspended from rails attached to the
machines xed upper crosshead). The equipment has the
ability to make fatigue tests by applying repeated loads of
the same magnitude to a specimen or component.
74
Fig. 3.10 a Nano-indentation made in gold-plated layer; b the
corresponding force versus penetration curve. The microhardness is
determined by extrapolation of the unloading curve to the x-axis. The
modules of elasticity is determined by the slope of the unloading curve
(courtesy of CSEM)
Specimens of various standard shapes, including
dumb-bell shapes stamped from sheet or foil, may be tested
simply, but with great accuracy, to provide fundamental
material information such as yield strength, ultimate tensile
strength, elongation and reduction in area, number of cycles
to fatigue, etc.
Alternative uses of the tensile test machine can include
the proof loading of cables, testing of the mechanical
properties of recording tapes and drive bands, and the fatigue
testing of soldered joints and adhesive bonds.
3.2.2.6 Thermal Cycling Systems
Systems are available for the screening of materials and
small assemblies by thermal cycling between temperature
extremes of 150 C and +150 C (Fig. 3.12). The test
method evaluates the effects of cycling temperature varia-
tions on individual parts and materials (e.g. due to power
dissipation, daily in-orbit temperature variations, and power
on/off with or without ambient thermal cycling, etc.).
Heating and cooling rates may be varied, but a rate of
10 C/min has been found effective for the accelerated
testing of metallic materials. Some systems operate under
high vacuum (1 106 mm Hg) and have been provided
with viewing ports and electrical feedthrough connections so
that the effects of thermal fatigue and embrittlement can be
monitored both in the camera and by electrical signals.
3 The Integration of Materials into Product Assurance Schemes
Thermal cycling has been responsible for the failure of
much spacecraft development hardware including
printed-circuit-board assemblies, spacecraft thermal-control
coatings, carbon-bre-reinforced epoxy structures.
glass-to-metal seals, solar-cell interconnector weldments,
and metallized pads on solar reflectors. Material lab inves-
tigations have shown that the primary degradation mecha-
nism is caused by thermal stress and low
cycle-strain-controlled fatigue resulting from the unequal
thermal expansions of the interconnecting materials; two
samples are shown in Fig. 3.14. Alternative materials and the
incorporation of stress-relief bends may be recommended to
project engineers as means for the rectication of these
problem areas.
Many material failures result from heating and cooling
processes encountered in the manufacture, assembly pro-
cesses, acceptance testing, and service life of spacecraft
systems. The thermomechanical failures illustrated in
Fig. 3.14 are due to thermal strains imposed on the material
microstructures due to both steady-state and transient com-
ponents. The steady-state component strains result from the
different thermal expansion characteristics of the individual
microconstituents of the article, while transient strains result
from thermal gradients across the article.
Accelerated thermal cycling tests are employed to predict
the lift of spacecraft components in a wide variety of pro-
cessing or service environments, and these need to be well
understood before selection of the test conditions. Severe
thermal shock may be imposed, for instance, by repeatedly
immersing a printed-circuit-board sample into molten solder
to simulate a number of soldering repair operations in the
vicinity of plated-through holes. Thermal proles may also
be designed to long-established specications such as those
shown in Fig. 3.12b, c.
Thermal cycling rather than thermal shock is considered
to be a more suitable test for the evaluation of manufac-
tured spacecraft electronic systems. The slower rate of
heating and cooling experienced in Fig. 3.12b will provide
better temperature stabilization and generate transient
strains, commensurate with a non-spinning satellite
emerging from solar eclipse. Also, the longer dwell times
will permit some, if not all, of the imposed stresses to relax
by means of a creep mechanism, and, once the half cycle is
reached, there will be larger stresses imposed in the
material than if the impressed temperature extremes are
achieved by the (c) thermal shock prole. It should also be
stated that the incidence of soldered joints operating at a
temperature in excess of 70 C is extremely rare, and when
soldered interconnections, particularly those based on
indiumlead solders, are held at temperatures of the order
of 125 C they become articially degraded by metallur-
gical reactions not experienced during normal spacecraft
operations.
3.2 The Materials Laboratory 75

Fig. 3.11 a General view of a

tensile test equipment with
custom-built environmental
chamber, chart recorder, and
control panel. b Same equipment,
but operated in compression
mode. Kevlar sandwich sample
undergoing four-point bending
test for the Giotto project (this
material was selected for the
Giotto spacecrafts protective
bumper shield)

Extensive thermal cycling of component-assembled
printed circuit boards will eventually produce failure
mechanisms in different locations of the circuit. The results
of numerous thermal cycling programmes have been anal-
ysed and it has been found possible to generate a fatigue-life
curve relating the magnitude of the thermal cycle (equivalent
to the stress range) to the number of cycles necessary to
produce the rst failure (Halpin 1985). Figure 3.13 illustrates
the concept of a thermal fatigue curve for component
assembled multilayer boards manufactured to military
76 3 The Integration of Materials into Product Assurance Schemes

Fig. 3.12 a. VMDI system for

thermal cycling under vacuum.
b Thermal cycling prole 55 C
to +100 C, t 6 C to 10 C/min
at each extreme total cycle
approx. 80 min. Based on
ECSS-Q-ST-70-04, and
Q-ST-70-08 specs. c Thermal
shock prole 55 C to +125 C,
t is fast. Dwell of 10 min at each
extreme. Total cycle approx.
25 min. Based on MIL STD
883 C, method 1010 condition B
3.2 The Materials Laboratory
Fig. 3.13 Thermal fatigue
failures in printed-circuit-board
assemblies and the concept of
using thermal cycling as a
verication test based on US Air
Force report (Halpin 1985)
quality. As part of a spacecraft new technology evaluation a
circuit could be submitted to verication tests following the
thermal cycling prole shown in Fig. 3.12b. The results can
be placed on the thermal fatigue curve and test circuits that
survive are assumed to have fatigue curves no worse than the
dashed curve shown on Fig. 3.13. When the new technology
is put into spacecraft service the operating stresses are less
than those on the dashed fatigue curve and the length of time
before failure can be expected to exceed the spacecrafts
designed life. The same curves are applied as a rule of thumb
by the author during the verication testing of surface-
mounted leadless components. The number of cycles nec-
essary to produce the rst leadless component failed
soldered joint may be 3750 for a 180 C temperature range
and 1250 for a 215 C temperature range (Frear et al. 1993,
Table 9.1).
3.2.2.7 Corrosion Testing
Corrosion tests are performed on spacecraft materials in
accordance with several international specications. Corro-
sion laboratory equipment such as that shown in Fig. 3.15a, b
will evaluate a wide range of features, including the suscep-
tibility of structural materials and springs to stress-corrosion
racking, the resistance of anodic and chemical conversion
coatings to surface corrosion, the intergranular corrosion of
weld-heat-affected zones, the protective properties of paints,
and the compatibility of bimetallic contacts. Such electronic
devices as component packages, battery cells, and. wires and
cables should be screened against corrosive environments
that include humidity, the residues of soldering fluxes, and
salt spray.
Large-scale stress-corrosion tests necessary for the eval-
uation of new spacecraft alloys (such as aluminiumlithium
alloys and particular-reinforced alloys) are usually made
under constant-load conditions by alternate immersion
cycles of 10 min in a solution of 3.5 % NaCl and 50 min
drying in air; typical apparatus is shown in Fig. 3.15c.
77
It may be noted that only static SCC test methods have
been described; these are considered to be most suited for
the classication of metal alloys according to their resistance
to SCC (see also Sect. 4.6.1). Dynamic SCC tests are fre-
quently performed by laboratories, and data can be achieved
within hours rather than the weeks (30 of 80 days) required
for the constant load test method. When cylindrical tensile
specimens or precracked specimens are evaluated by vari-
able strain rate techniques the results should only be used to
compare batches of material within the same family of alloy
composition. At high strain rates the electrochemical reac-
tions associated with SCC do not have sufcient time to
occur, and the deformation process will depend on disloca-
tion movements occurring throughout the specimen to cause
necking, a high elongation, and, ultimately, transcrystalline
breakage of the metal grains. As the strain rate is decreased
while the specimen is exposed to the corroding solution, for
instance a saline solution, the susceptibility of the specimen
to SCC may or may not increase. This uncertainty depends
on several factors, but mainly on the ability of the metallic
surface to form a protective passive lm. If the lm does
form it will rupture under further strain to form a crack, so
producing localized anodic dissolution of the metal under
test. If the strain rate is sufciently slow the repassivation
process will have time to ensure a new lm above the crack
before dislocation movement causes the emergence of the
next slip step. The ability of the alloy to repassivate is the
predominant factor controlling the propagation of the stress
corrosion crack under dynamic testing; other factors are
associated with hydrogen-induced cracking. The existence
of a different threshold strain rate for the repassivation of
each metal alloy under test renders the results of dynamic
SCC testing difcult to interpret for classication against
SCC susceptibility. Selection of the most appropriate
method for performing SCC testing must taken into account
a review of the spacecraft hardware material under consid-
eration and its applicable in-service environment. The
78 3 The Integration of Materials into Product Assurance Schemes

Fig. 3.14 Examples of thermal

fatigue cracking (as revealed by
metallographic inspection of
polished cross-section).
a Open-circuit failure of a
non-approved soldered joint. The
crack propagates from the corner
of plated-through hole and across
the solder llet (200 cycles
between +100 C and 60 C).
Arrows detail intermetallic
growth. b Transverse section
through CRFP structural tube
after thermal cycling; microcracks
have formed in the thicker
longitudinal lay-ups

majority of components will be exposed to static loads, as
the period of launch is small in comparison to their required
lifetimes, hence the constant load method is to be preferred.
To duplicate the stress corrosion failure of spacecraft
battery cell terminals and the leads of electronic component
packages, fabricated from Nilo 42 and Nilo K respectively,
testing is made in the laboratory and under marine atmo-
spheric exposure conditions. Since the failure of both alloys
in service had been in situations where plastic strain (from
bending and forming) and elastic strain were present, stress
corrosion tests were carried out by the loop test devised by
Thompson (1961). The marine test sites and test samples are
shown in Fig. 3.16.
A review of the methods for testing ferrous and nickel
alloys and their weldments for a wide range of corrosive
environments (general, pitting, intergranular, crevice, SCC
and galvanic corrosion) has been compiled by Pettersson
(2007). This extensive paper discusses the applications of
stress corrosion standards. The Constant Load Test
(CLT) which is preferred by the European space community
seems to be the most used test for the oil and gas industries.
3.2.2.8 Fracture Mechanics Testing
The concepts and principles of fracture mechanics are rela-
tively new, but when the discipline is applied it can offer a
system-type approach to the failure prevention of launcher
casings and pressure vessels. In view of the high cost of
launching materials, it is increasingly pertinent to optimize
the use of metal alloys and their heat treatment with per-
formance. Damage tolerance is generally applied to all
manned spacecraft piece-parts where a failure could cause a
catastrophic hazard. In applying the damage tolerance phi-
losophy the designer must assume that crack-like defects
exist in each safe-life piece of hardware, and that, during
dened intervals such as the mission duration or inspection
interval, these cracks will not propagate to cause failure.
This activity is greatly assisted by the establishment of
empirical fracture mechanics tests and fatigue tests on rep-
resentative materials or proof tests on the actual fabricated
item. It is important to realize that both fracture and fatigue
testing for the generation of crack growth, material tough-
ness, and fatigue properties are influenced by many param-
eters such as the precise alloy composition, heat treatment,
3.2 The Materials Laboratory 79

Fig. 3.15 a, b Typical corrosion

laboratory equipment.
a Salt-spray cabinet (foreground)
and humidity chamber, both of
which have temperature and
humidity regulators. b Carousel
for the alternate immersion of
specimens into a 3.5 % NaCl
solution. c General photograph
showing a laboratory set-up for
constant-load stress-corrosion
testing
80
Fig. 3.15 (continued)
grain direction, and test environment. Only by extensive
testing can basic material data be derived for spacecraft
fracture mechanics analysis. This is also particular relevant
for composite materials; life predictions for composite
structures are more frequently made because of the increased
use of these materials in aerospace applications.
Fracture toughness tests and crack growth tests are made
with specimens that have been precracked with either sur-
face or internal cracks. The ESA fracture control require-
ments, (ECSS-Q-ST-40, 2008) dene when fracture
toughness tests are performed to determine the KIC (plane
strain fracture toughness). The test specimen depends on the
form, the strength, and the toughness of the material under
test. The ASTM Method E399 is generally followed using a
rectangular cross-section test-piece in which a notch is
machined. The notch extends from one edge through half the
depth of the specimen, and the specimen width is half the
depth. The specimen is rst subjected to fatigue loading to
cause a real crack to form at the root of the notch, and loaded
until the crack has grown to 5 % of the overall notch length.
The crack length is monitored with a travelling microscope,
and measured accurately after test. Test samples made from
AA2219 forgings are shown in Fig. 3.17. The samples were
loaded in tension. The load versus displacement is measured
and the loading rate determined as being within specica-
tion. The KIC values are assessed from tables. The methods
of testing are becoming well dened in recognized
3 The Integration of Materials into Product Assurance Schemes
specications. The testing is generally performed with the
following equipment:
Servo hydraulic testing machines with capacity in the
range 100500 kN (computer-controlled input capabili-
ties for variable amplitude load spectra to simulate tran-
sient and random service loading).
Facilities for measuring crack growth.
Environmental temperature testing, high pressure or
testing in vacuum, and possibly testing in an aggressive
gas or liquid environment.
Fatigue tests are performed to determine the number of
load cycles that a material can withstand without fracture
the tests load the specimen for a denite stress cycle, applied
at a dened frequency. For all metals there is a limiting
range of stress below which fracture will not occur for any
number of loading cycles, however large. When there is
fatigue crack growth it is possible to assess the growth rate
by inspecting the polished edge of the specimen (as for
instance was done for producing the fracture toughness
specimen in Figs. 3.17a and b. Fatigue fracture surfaces are
illustrated in future chapters (a classical fatigue failure is
shown in Fig. 5.2), but for modern metallic alloys it is the
microstructure of the alloy that provides for their low fatigue
crack growth rate (FCGR). The superior FCGR resistance of
3.2 The Materials Laboratory 81

Fig. 3.16 Corrosion test sites,

from small to great investments.
a General view of the UKs
Hayling Island test site at low
tide. b Constant strain loop test
specimens exposed under marine
atmospheric conditions Loop
specimens after 18 months on
test. c Fissuring in Nilo K
specimen after 18 months on test
(400). d Overview of the
specialist sea coast corrosion test
site at NASA Kennedy Space
Center, Florida, and e the
seawater spray rig for the testing
of coated panels (courtesy Dr Luz
Marina Calle, NASA)

aluminiumlithium alloys can be traced to their micro-
structure. This is shown in Figs. 3.17c and d, where the
fatigue crack has a tortuous crack path resulting from the
alloys highly planar slip characteristics and the presence of
texture. This tortuosity increases the energy required to
propagate the crack and impedes crack growth. The facets on
the fatigue crack surface have been studied in detail by Wu
et al. (1994); they are separated by 109.5 and this angle of
separation coincides with the angle between the favourable
slip plane of aluminiumlithium{III} and the plane that is
parallel to the rolling direction of this plate material.
The present standards for fracture mechanics testing were
originally developed for the testing of homogeneous metallic
alloys. Future modications must be made to the test pro-
cedures when ceramics, metal matrix composites, and
metallic weldments are tested.
Data on the strength of materials have been derived over
the years by industry and national agencies. A computerized
materials database is available for compiling this data.
FRAMES-2 contains raw data taken from measurements
performed on test specimens (such as force, displacement,
crack length, and number of load cycles), deduced data
produced from raw data by numerical methods such as curve
tting or differentiation, and design data intended for veri-
fying a structural design, such as lower-bound values.
Industrial companies can access the ESA FRAMES-2 data-
base as an aid in their selection of structural materials.
3.2.2.9 Printed-Circuit-Board and Component
Assembly Evaluations
Printed-circuit-board (PCB) evaluations are made in metal-
lurgical laboratories during the quality assessment of PCB
manufacturing lines. New PCB laminates and fabrication
techniques can be studied for possible use on future space-
craft. Equipment exists for the visual inspection of circuits,
peel-strength testing of conductor tracks, and the solder-
ability testing of terminals, whether they be pads or
plated-through holes.
Component assembly on PCBs and wire/cable inter-
connection by manual soldering, automatic soldering,
microwelding, and crimping are some of the most critical
spacecraft processes. Standard test equipment exists in the
82 3 The Integration of Materials into Product Assurance Schemes

Fig. 3.17 a and b Plane strain

fracture toughness specimen after
test. Note the initial fatigue crack
arrowed at the base of the notch.
The toughness property is the
critical value of the stress
intensity factor at which rapid
propagation of a crack occurs
during static loading. (c and d).
c Microstructure of aluminium
lithium alloy in a plane
longitudinal to the plate rolling
direction. Superimposed is a
three-dimensional schematic
showing the relative orientation
of favourable slip planes in a
compact tension specimen;
d fracture surface shows
well-dened slip band facets

laboratory to evaluate the material properties and metal-
lurgical characteristics of samples produced by these join-
ing techniques:
Mounting, grinding, and polishing equipment, with
ultrasonic cleaning baths for the metallographic prepa-
ration of soldered joints, plated-through holes, etc.
Photographic camera for the general photography of
incoming samples.
Solder globule solderability tester for wetting/dewetting
checks on component leads, wires, terminal posts, etc.
Rotary dip machine for solderability and solder-float
testing of PCB pads and plated-though holes. Some
effects of this test on poor-quality products are illustrated
in Fig. 3.18.
Peel test made by using attachment to tensile test
machine (see Sect. 3.2.2.5).
Thermal cycling systems (see Sect. 3.2.2.6).
Laminar flow bench providing a dust-free working space
for the assembly and inspection of spacecraft hardware
under ideal lighting conditions and a Class-100 envi-
ronment (US Fed. Std 209a).
Stereo-zoom binocular microscope and metallurgical
microscopes for quality inspection of electronic
hardware.
In the most simple case a PCB consists of an
epoxy-impregnated breglass-laminated weave with copper
conductor lines bonded to the epoxy laminate surface by
means of a further epoxy adhesive layer (called the top
coat). Electronic components are soldered to copper pads of
plated-through holes. More modern boards are made of
polyimidequartz laminate, and these are particularly
well-suited for the assembly of leadless surface-mount
components such as ceramic chip capacitors and resistors,
leadless ceramic chip carrier packages, and the like. In more
complex, and more expensive, PCB assemblies, the lami-
nated substrate consists of a ceramic dielectric with a
three-dimensional refractory metallized conductor network
with pads that have a series of metallized terminations onto
which electronic components can be soldered. Multichip
modules (MCMs) can be mounted onto these
ceramic-laminated structuresthis is essentially the
mounting of integrated circuits onto a ceramic board that has
incorporated a wiring circuit with signal layers, power
planes, and ground planes built in. Simply put, the chip
carrier becomes the ceramic printed circuit, or multilayer
board. Obviously, these MCMs offer a decreased size and
weight and a faster performance (up to 34 times as fast
traditional PCBs), and they are presently under consideration
for qualication to spacecraft test standards.
3.2 The Materials Laboratory 83

Fig. 3.18 a Schematic

representation of major steps in
multilayer board fabrication.
b Section through multilayer PCB
showing delamination next to
high thermal mass of
ground-planes. Areas of resin
recession are also arrowed
(after solder-float test for 20 s at
288 C) (courtesy of Jane Jellison
and Karen Houlberg)

Ceramics such as alumina oxide (A12O3), aluminium
nitride (AIN), silicon (Si), and beryllia (BeO) are a great deal
more fragile and susceptible to microcracking than metals
and PCB materials. Consequently, all metallographic
preparations must be made extremely carefully and by using
methods that distribute compressive stresses evenly across
the ceramics surface. Samples should be potted in a hard
plastic and slowly ground down to the plane of interest.
Sawing, even with diamond-coated blades, is unsuitable, as
microcracks having long cleavage paths will be generated
and it will often be impossible to grind back sufciently to
eliminate these cracks. When conductor thicknesses are to be
measured it is essential to over-plate the metallizations with
a hard layer of nickel or thick copper in order to prevent
edge-bevelling during the polishing of microsections. The
evaluation of ceramic PCBs typically does not require any
special operator training beyond an emphasis on extra care in
handlinga ceramic board when dropped will very often
shatter.
3.2.2.10 Conventional Nondestructive Testing
(NDT)
A large number of NDT techniques are available to the
materials engineer, and spacecraft piece-parts and small
components are frequently inspected for defects such as
microcracks in brittle materials and porosity in welded
structures. The defects can be generally categorized as:
raw material defects (e.g. rolling defects, blow holes in
castings, seams, segregation of alloying elements),
spacecraft processing defects (e.g. brazing and welding
anomalies, incorrect location of encased units, delami-
nation, machining defects, misalignment),
service defects resulting from overtesting, incorrect
choice of materials, fatigue cracking.
Conventional NDT techniques require trained and com-
petent operators and methods for record keeping. A good,
clear, enlarged photograph or radiograph is worth many
84
words, and occasionally video tapes may be useful to record
visual examination. The most important methods are:
Visual examinationThis is probably the single most
important method of inspection, and relies on a trained eye,
magnifying devices, good lighting, endoscopes, mirrors, and
brescopes.
Liquid penetrant testsPenetrants may be simple mix-
tures of alcohol and water so that, after a suitable penetration
time, the excess liquid is removed and the surface dried.
Liquid seepage out of crack-like defect may be made more
apparent by means of dye, fluorescent additives, and thin,
white surface coatings which become stained by the seeping
liquid. Penetrants can contaminate spacecraft if not com-
pletely removed.
Ultrasonic testingAll spacecraft metals readily transmit
ultrasonic vibrations which are reflected or scattered (atten-
uated) by flaws or inclusions. Methods include transmission,
pulsed echo, and acoustic imaging. Composite materials are
more difcult to assess and should be double checked by
radiography particularly in the case of honeycomb sandwich
panels.
RadiographyRadiographs can be produced by means
of X-rays and gamma-rays that pass through the item under
investigation to produce a photographic image. X-ray
absorption will occur when dense material and refractory
inclusions are present, giving rise to light areas. Voids,
porosity, and cracks, being hollow, show up as dark areas.
Organic composite materials may be more difcult to inspect
because the X-ray absorption of resin and graphite bres is
similar. Penetrant-enhanced radiography can be used to
detect disbonds within honeycomb skin lay-ups.
Coin or tap testGross defects can be located by lis-
tening to the ring of a metallic item; flawed parts may emit a
dull note when impacted with a coin or small hammer.
Disbonds between face skins and honeycomb panels can be
easily detected by carefully dropping a coin, edgeways, onto
the face skin surface. A dull rather than a harsh response will
result above areas of disbond. Listening sensitivity can be
improved by means of a stethoscope.
3.2.2.11 Tribological Test Methods
Spacecraft mechanisms are generally designed to incorporate
a lubricant that does not degrade either during ground testing
in air, or testing and operation under vacuum. Metal to metal
contacts are composed of dissimilar metals as discussed in
Sect. 2.3(d). A common failure mode seen during the testing
and operation of spacecraft is termed cold welding. Euro-
pean laboratories refer to this as adhesion, sticking or actuator for frequencies between 0 and 300 Hz and ampli-
stiction. Cold welding between two contacting surfaces can
occur under conditions of impact or fretting. These surfaces
may be bare metals, or inorganically or organically coated
metals and their alloys. One documented example of a
failure due to cold welding after fretting occurred on the
3 The Integration of Materials into Product Assurance Schemes
Galileo spacecraft (Johnson 1989), when the high-gain
antenna could not be fully deployed. The ribs of the
umbrella-shaped antenna were locked for launch, but failed
to open. Investigations have shown that fretting during
transport and lift-off caused the ribs to cold weld together in
the launch position. Further more recent examples will be
given in Sect. 5.1.2.
The aim of a recently developed tribology test (Mer-
stallinger 2009) is to provide designers and engineers with
data on adhesion forces that have to be considered in design
of mechanisms, in order to assess, on the one hand, cold
welding on bare metal contacts and, on the other, the per-
formance of coatings in preventing adhesion. The test
enables comparison of results for different material pairings,
and the collection of data for entry to a database.
The impact test facility, for loads up to 100 N, enables the
measurement of adhesion forces under vacuum between
contact points after cyclic contacts, varying from static con-
tacts (long-term compression) to slow closing cyclic contacts
(static adhesion), and after impacts with impact energies up to
0.02 J. The impact test facility consists of an ultra high
vacuum system (108 mbar after bake-out at 130 C) with an
ion getter pump and an air damping system for vibration-free
measurements. The contact is made between a ball and a flat
disc. These are the two metals under evaluation. The ball is
mounted on a pushrod, which is driven by an electromagnet.
A low-friction loading system enables an accuracy of 1 mN
(0.1 g) for the adhesion force measurement, which is made
directly above the pin in the vacuum chamber using a piezo
force transducer. The transducer measures the impact force as
well as the adhesion, i.e. the force needed to separate the two
materials.
Cyclic loading may be done either slowly (static) or by
impact (dynamic). By varying the ball radius (typically 2
20 mm), the contact pressure can be adjusted to the yield
strength of most materials. As an option, the contact surfaces
can be cleaned in situ by glow discharge before the test.
The fretting facility allows investigations of the influence
of fretting on the tendency to cold welding of materials.
After a certain number of fretting cycles the adhesion force
between contacts is measured. The loading mechanism is
similar to that of the impact test device described above. The
loading and adhesion forces are measured by the z-direction
of a 3-axis piezo transducer mounted directly below the disc
under vacuum. The friction force due to the fretting move-
ment is measured in the x-direction. The fretting movement
(sine, triangle or square wave) is generated by a piezo
tudes up to 100 m. This lateral movement is transduced to
the pin via a CuBe plate and controlled at the contact by a
triangulation sensor.
This facility has allowed for the rst time the simulation
of high-frequency vibrations resulting from bearings, for
3.2 The Materials Laboratory
example, combined with measurements of the adhesion
force.
3.2.3 The Use of New Laboratory Techniques
for NDT
3.2.3.1 Holographic Interface Tester
The metallographic methods employed for the case investi-
gations detailed in this book are based on both optical and
electron microscopy, and X-ray analysis. In recent years, the
continued demand for better monitoring of quality and
reliability has led to the introduction of more modern
inspection methods and laboratory techniques than those
outlined in the previous section. They can supplement con-
ventional procedures by providing remarkably accurate
nondestructive data. This can be useful for monitoring the
growth of defects and for determining the optimum locations
for making microsections.
The usefulness of holographic NDT inspection will be
demonstrated in Sect. 5.10 as part of the failure analysis of a
defective spacecraft antenna. Debonded and fractured parts
of the CFRP face-skins were uniquely identied by the
Holographic Interface Tester (as developed by Schliekel-
mann 1972), and the extent of the damage could be photo-
graphically recorded before progressing with the destructive
metallographic failure mode assessment.
3.2.3.2 Acoustic Microscopy
(a) Scanning-laser acoustic microscopy
The Scanning Laser Acoustic Microscope (SLAM) produces
images of the internal structure of metals, ceramics, and even
biological tissues, and these can be displayed in real time on
a television monitor, as shown in Fig. 3.19. The SLAM uses
very-high-frequency ultrasound to produce images.
The basic transducer and detection scheme employed by
the SLAM is illustrated in Fig. 3.19b. In this case solar-cell
interconnector welds are under evaluation. The samples are
hand-held and are easily manipulated on a simple specimen
stage. As shown in the sketch, acoustic waves are trans-
mitted into the bottom face of the solar-cell sample by means
of a distilled-water coupling medium.
Ultrasonic waves will not propagate across an air gap,
and are therefore attenuated in areas where there is lack of
bonding. The waves will pass through where solar-cell
interconnector weldment is good. They are then reflected at
the samples top surface. The transmitted acoustic energy
imparts pressure waves on the top surface of the specimen.
The oscillations have the same frequency as the incident
wave, but the variation in their amplitude depends on the
85
acoustic attenuation and absorption properties of the
underlying material.
The surface waves are detected by a scanning laser beam
with a small spot size, which permits an imaging resolution
of the order of 10 m. The acoustic image is displayed on a
TV monitor, where white regions correspond to areas of the
sample with good acoustic transmission properties, and
darker regions denote areas where ultrasonic attenuation is
higher.
The SLAM uses a video-signal processor to integrate the
signals from laser scans to a number of grey levels, each of
which is assigned a different colour. This is demonstrated in
Fig. 3.21. Which shows an example of SLAM inspection of
a silicon chip attached to a ceramic carrier by means of a
solder preform. Although the colour images contain the
same amount of information as the black and white ones,
they are easier to interpret.
Photographs, termed acoustic-amplitude micrographs,
are taken of the acoustic TV image as seen in Fig. 3.21b. In
addition to displaying the acoustic-amplitude distribution
throughout the eld of view, the SLAM provides an acoustic
interference mode of operation. Acoustic interferograms
(Fig. 3.21c) exhibit a series of alternating light and dark
bands. For acoustically homogeneous samples, these bands
(interference fringes) are arranged parallel to one another
and are equally spaced.
The rst SLAM purchased for a European laboratory was
in 1980. This instrument has provided decades of useful
serviceit was found (Dunn 1982) to be a unique tool that
can accurately judge the initial quality of electrical resistance
welds that joint together tens of thousands of individual solar
cells making up a spacecraft solar array. These welds are
made by a process requiring close control of every process
parameter, particularly for the electrode pressure and current
pulse. Weld quality could not be determined by conventional
nondestructive methods including high-denition radiogra-
phy, and reference to workmanship standards also failed to
provide information about these very small weld zones. This
problem was solved by the SLAM method, which managed
to locate and determine the extent of weld zones between
solar-cell metallization and interconnectors. The progress of
thermal fatigue within the welded zones can also be fol-
lowed. The accuracy and high resolution of the SLAM
method have been conrmed by detailed metallography,
particularly or interconnected samples that had been sub-
mitted to 30,000 cycles between 100 C and +100 C as
part of a weld reliability programme for a low-orbiting
spacecraft.
The SLAM instrument is no longer manufactured having
been replaced by a series of acoustic microscopes without
scanning laser features. The latest version is shown in
Fig. 3.20 together with images made using two modes of
operation. The reflected mode, where pulsed ultrasound has
86 3 The Integration of Materials into Product Assurance Schemes

Fig. 3.19 a The rst Scanning

Laser Acoustic Microscope
(SLAM) to be operated in
Europe, for decades the
laboratory work-horse; and b set
up for the evaluation of solar-cell
interconnector weld zone (not to
scale)

been directed into the base of a leadless ceramic chip carrier, two transducers, above and below the component. Hori-
uses the return echoes to produce high-resolution images of zontal slices can be made from the top to the bottom faces of
large delaminated areas; these have been highlighted in red the chip capacitor example seen in Fig. 3.20. Precise
(the colour choice is arbitrary). The through scan mode uses knowledge of the extent and location of defects within this
3.2 The Materials Laboratory 87

Fig. 3.20 Photograph of the

recent Sonoscan Gen6, together
with acoustic microscope images
taken in different modes:
reflected-mode to reveal an
LCCC device with internal
delamination (area in red); and
through scan mode (transducers
either side of sample) to show
consecutive horizontal slices
through a ceramic chip capacitor
with internal voiding (courtesy of
Richard Carr, Sonoscan Europe)

capacitor have been made possible by the recording 50
images, each taken at an equal, small increment throughout
the body thickness. As major defects have appeared (the
white zones at increments 15 to 18) they were highlighted by
the yellow arrows, but also different material densities can be
seen by the changing shades of grey. The C-SAM is
described in the next section.
(b) C-mode scanning acoustic microscope (C-SAM)
The C-SAM analysis involves placing a single, focused
acoustic lens over the sample. Suitable samples are usually
no larger than 4 cm long and up to 3 mm thick. The lens
mechanically rasters a small dot of ultrasound over the top
surface of the sample. As the ultrasound is pulsed into the
surface, a reflected wave (echo) is generated at each interface
within the volume of the sample and returned to the trans-
ducer for processing. A high-speed digital signal processing
unit permits information to be gathered from all depths
within the sample. Images can be generated from specic
depths or through the complete sample thickness. The
high-frequency ultrasound (10100 MHz) is very sensitive
to internal discontinuities in materials. The method has been
developed for many NDT applicationssee for instance
Fig. 3.22including those described for the SLAM in the
preceding paragraphs. In some circumstances the technique
can be used instead of making destructive microsections.
Metals, polymers, ceramics, and certain composite materials
have sufcient acoustic transmissivity to permit imaging.
Defects that are positioned beneath other discontinuities are
not detectable by C-SAM imaging (Martel and Adams 1996;
Rameriz and Adams 2002).
3.2.3.3 High-Definition Radiography
(a) X-ray radiography
Conventional X-ray units are primarily designed to provide
the sharpest of images in the shortest of times. These radi-
ological examinations are based on the fact that X-rays and
gamma-rays can pass through materials which are optically
opaque. The quality or denition of a radiograph will
depend on factors that affect the transmission of radiation
through the specimen, and these will include material
thickness, composition and homogeneity of the material,
and intensity of the initial radiation emanating from the
equipments focal spot. Modern equipments will detect the
88 3 The Integration of Materials into Product Assurance Schemes

Fig. 3.21 Nondestructive SLAM inspection of chip attachment.

a Optical photograph of chip soldered to ceramic carrier. b SLAM
acoustic-amplitude micrograph of same area (colour-enhanced).
c SLAM acoustic interferogram revealing poor solder bonding of
device (colour-enhanced). Length of chip is 4 m

transmitted radiation, making use of both a fluorescent

screen (fluoroscopy) and a photographic plate (radiography).
Video signal recorders and sensitized lm provide for per-
manent records. Cracks and cavities, being hollow, show up
as dark areas, whereas heavy metals and inclusions within
materials appear as light areas since they usually absorb
more radiation.
X-ray units operating at up to 100 kV with radiographic
sources emanating from extremely small focal spots having
approximate diameters of 15 m are now available for
very-high-denition radiography. They enable ne detail to
be resolved by the direct projection of transmitted radiation
to magnications of 20 as shown in Fig. 3.23.
Electrical, or in-circuit testing, is the standard way to
detect short circuits and opens on PCB assemblies. As the
density of components rises and the number of layers within
a multilayer PCB increases (some have 20 layers), physical
test access becomes unreliable, and in some cases impossi-
ble. More use is being made of X-ray (microfocus) methods
for inspection of such space hardware. The method is also
used for measuring the quality of a soldered joint. The solder
lowers the intensity of the X-ray beam much more than any
other material present, such as copper, silicon, or the glass
epoxy PCB. The X-ray image can rapidly detect defects such
as porosity (internal to the solder joint) and solder-balling
beneath component packages. Figure 3.24 details some
X-radiographs of solder joints containing spherical pores.
These soldered connections were made by wave-soldering
without any pre-tinning of the component leads. The
microsections made of the same joints conrm the presence
of voids. A similar evaluation was made with component
packages from the same batch, but which had been
pre-tinned according to ECSS-Q-ST-70-08. In this case the
joints contained no porosity and by microsectioning were
judged to be of excellent quality. A large number of elec-
tronic failure analyses involving the use of microfocus
radiography are described by Marchaise and Glodowski
(1991), some of which illustrate the effect of gold contam-
ination in solder joints and the resulting open circuits.

(b) Neutron radiography

Neutron radiography is also capable of producing images of

components and materials on lm or by real-time imaging
devices; here neutrons are used as the penetrating radiation.
Neutron radiography is particularly useful for the evalu-
ation of structures and spacecraft subassemblies fabricated
3.2 The Materials Laboratory
Fig. 3.22 C-SAM detection for cracked, delaminated, voided, and
internally flawed components. The screening of production lots prior to
surface mount assembly is made to MIL-C-123, Appendix C. The
acoustic micrograph shows six ceramic multi-layer chip capacitors.
Four components have internal defects (courtesy of Sonoscan, Inc.)
from composite materials, as the thermal neutrons are highly
attenuated by the elements hydrogen, carbon, and boron.
Adhesively bonded face-skins on aluminium honeycomb
can be easily inspected by neutron radiography because of
the high neutron capture rate of the hydrocarbon adhesive.
Areas supporting very thin adhesive and low-volume adhe-
sive llets will appear darker than those areas covered in a
thick layer of the organic materials which may become
completely opaque to the neutron beam. Neutron radiogra-
phy has the particular characteristic that neutrons can pen-
etrate heavy metals such as steel, uranium, and even lead. It
has the unique ability to image many light materials such as
Fig. 3.23 High-denition
radiographs of satellite thruster
heaters after life-testing. The steel
cylindrical casing has a 300 m
wall thickness. This technique
readily exposes the coiled 50 m
diameter heater lament, and can
detect local regions where it may
have oxidized or sublimated
89
plastics and materials containing hydrogen, carbon, and
boron, even when they are encased in the dense metals.
Neutron radiography is used for the NDT of several
critical spacecraft components, such as
High-voltage electrical connectors located between
travelling-wave tubes and their electric power supply unit
(to control for air bubbles within the epoxy resin back-
lled connectors, as these lead to corona breakdown).
Control of cable-cutters and separation bolts as shown in
Fig. 3.25.
(c) Computerized X-ray tomography
Complicated assembled spacecraft hardware such as elec-
tronic boxes containing up to twenty component-assembled
PCBs that have been totally embedded in foam materials to
ensure against vibration damage during launch can be
inspected by computer-aided tomography, using apparatus
designed and extensively employed for medical examina-
tions. The same automated radiographic method can be used
for composite structures where different elements may not be
of sufciently different density to make them distinguish-
able, and/or their images may overlap in the direction of
view. Digitization of the X-ray image and the enhancement
of image quality by computer methods can reveal the precise
location of defects by inspection of through-the-thickness
slices.
The US space agency NASA uses highly reliable multi-
layer PCBs having up to 50 layers. The plated-through holes
are inspected regularly for solder-ll by laminographya
type of cross-sectional X-ray test. Transmissional X-ray
images, such as the one shown in Fig. 3.24a cannot separate
the top of the barrel solder llet from the one at the bottom.
Laminography overcomes this by taking a series of slices
through the board, parallel to its surface. Further examples
are shown in Sect. 6.15.11.
90 3 The Integration of Materials into Product Assurance Schemes

Fig. 3.24 Examples of

microfocus x-radiography.
Polyimide multi-layer board
assembled with leaded flat
packages by wave soldering.
Components had been mounted
either without pre-tinning or with
pre-tinning. These X-radiographs
(a, b) and microsection (c) reveal
the negative effect of omitting the
pre-tinning stage. d Clearly
reveals void formation beneath a
IC chip. e is a standard
radiograph, but f shows the same
flat package solder with far more
detail of voiding between leads
and pcb pads. Courtesy of Hytek,
Denmark

3.2.3.4 Acoustic Emission to Monitor Damage
Accumulation in Materials
and Welds
Advanced spacecraft structures and certain mechanisms can
be constructed from a number of composite materials in
order to achieve increased specic strength when compared
to conventional monolithic alloys. Composites can also
contribute to other improved characteristics such as better
wear resistance or a specially controlled coefcient of
expansion. The mechanical properties of intermetallic com-
posites and metal matrix composites (IMCs and MCCs) as
well as bre-reinforced plastics (FRPs) depend greatly on
the integrity of the bond between the matrix and the
reinforcement. A rather modern technique of detecting the
acoustic emission (AE) from materials under stress is being
developed which can provide a direct method for studying
mechanical property degradation, interfacial disbonding, and
the development of microcracks. An overview summarizing
the method has been written by Hinton (1995). AE refers to
relatively low-amplitude sound waves that are generated by
various dynamic energy transfer mechanisms from micro
areas in a material. According to ASTM E610-82, the de-
nition of AE is: Acoustic emissions are transient elastic
waves generated by the rapid release of energy from energy
sources within a material. In metals these can be the
cracking or decohesion of inclusions and the interaction of
3.2 The Materials Laboratory
Fig. 3.25 High-denition
neutron radiographs of spacecraft
pyrotechnic cutters. Both designs
contain two electric
current-ignited pyrotechnic
cartridges for greater reliability.
These devices have been red. Of
particular interest are the sealing
systems to keep the spent
explosives from escaping and
causing damage (see also
Sect. 5.4)
slip lines, as shown in Fig. 3.26. The formation of twin
bands and grain boundary cracking also emit AE signals (an
extreme case is that of pure, polycrystalline tin which
releases an easily audible sound, or cry, as it is being
deformed). In composites the AE is generated by bre
breakage, matrix damage and, as previously stated, the
delamination and pull-out of reinforcement particles from
the matrix.
The detection of AE is conventionally performed by
placing a sensor either directly, or by means of a couplant,
onto the surface of the item under investigation. Sensors are
piezoelectric crystals such as Piezo-Zirconate-Titanate
(PZT), mounted as transducers, and these are extremely
sensitive to acoustic waves having an amplitude as low as
25 pm (Pollack 1989). A variety of waveguides, electronic
ampliers, and electronic lters may be used before the
signal is processed by a secondary amplier and nally
displayed on an oscilloscope. By using multiple transducers
attached to the surface of a device or structure it is possible
to precisely locate the source of a defect within the volume
of the material (this can be considered as analogous to
nding the epicentre of an earthquake).
91
Many aerospace materials have been evaluated by AE
(Hamstad 1986), but because it is not possible to distinguish
damage modes or failure mechanisms by this technique
alone, the method has received a somewhat lukewarm
reception by materials engineers. However, AE response
behaviour has been characterized for a welded maraging
steel and acceptance criteria determined for the control of a
spacecraft pressure chamber and undergoing a proof pres-
sure test (Chelladurai et al. 1995). AE has also been used to
monitor fatigue damage in samples of Inconel 901 and to
predict the onset of crack initiation (Fang and Berkovits
1995). The same authors found that the threshold stress
intensity factor values determined by AE tests were com-
parable to those obtained from conventional fracture
mechanics tests, but for a much reduced time and cost.
The major asset of AE is that it is now a sufciently
mature technique to assist in both the production and the
quality controls for composite materials. For instance, under
proof loading, spacecraft hardware can be AE monitored for
First Damage Detection, Flaw Detection, and Failure
Warning. Also, the emissions that are originating from the
relief of residual stresses can be monitored in situ during
either thermal or mechanical processing.
92
Fig. 3.26 a Polished and etched
microstructure of a 0.33C
pressure vessel steel containing
ferrite, pearlite, and silica
inclusions. Under proof loading
(long before crack formation) the
silica inclusions break, between
arrows, on (b). This event
produces acoustic emission
signals shown simplied in (c).
Loading of the polished steel
sample was stopped at a 7 %
elongation, when slip lines and
Luder bands interact, such a
dislocation movement also
producing AE signals. (Dunn,
unpublished report)
AE was applied by industry for monitoring materials used
in the construction of the Concord supersonic planes in the
1960s and is now being used to study microssures in space
materials such as hybrid glasscarbon and Kevlar laminates
(Perami et al. 1992), aluminium and titanium matrix com-
posites (Roman et al. 1992), and graphite-epoxy which is
wound onto a liner/mandrel pressure vessel consisting of
two aluminium hemispheres that had been tungsten arc
welded at the equator (Downs and Hamstad 1995). AE was
also applied in a manner similar to that presented in the last
reference during the development of the lament-wound
MAGE apogee boost motors discussed in Sect. 4.8.3. Here,
the optimization of manufacturing process parameters were
partly determined by AE responses generated during pres-
surization of these vessels up to both the proof load and their
actual burst pressure. In-process quality controls utilizing
AE are also applied during the fabrication of the Ariane IV
launch vehicles various payload fairing (SPELDA) designs.
Each SPELDA (see Fig. 2.7) takes up the stresses during
multi-satellite launches and provides the aerodynamically
shaped external protection skin that is released by the det-
onation of pyrotechnic devices just prior to the spinning and
ejection into space of the individual satellites. The SPEL-
DAs are constructed from thin carbon-bre reinforced plastic
skins which are laid up and bonded to aluminium
load-bearing rings. These bonded joints are difcult to
control by nondestructive methods such as radiography or
ultrasonics owing to the bre content and the presence of
small voids, but they are assessed by AE. The test setup for
this part of the European launcher is described by Cole and
Thomas (1989); the locations of attached sensors are shown
in Fig. 3.27. In fact, it was found that the small voids have
no effect on the strength of the joint, as it is the interfacial
3 The Integration of Materials into Product Assurance Schemes
bond between the CFRP and the aluminium surface which is
most critical. Numerous sensors are attached to surfaces both
in the region of the aluminium-to-CFRP bonds and the
CRFP panel-to-panel joints. The AE testing is mainly con-
ducted with the SPELDA in compression (to simulate the
compressive forces expected from the payload and fairing
during launch). Any unacceptable AE can give an early
warning of failure so that testing is halted before ultimate
failure occurs and the problem areas can be pinpointed and,
after being subjected to a material review board, there is the
possibility of reworking the structure to an acceptable level
of condence. The base line criteria for this condence were
gained from AE testing on development and qualication
models which were loaded to failure. The typical failure
criteria for one-shot launch vehicles are very different to
those applied to aircraft or space shuttle vehicles with
20-year design lives, and may include points such as: a
Felicity ratio (FR) of greater than 0.95 measured at 90 % of
the loadthe FR being dened as the load at which the
onset of signicant emissions occurs on a subsequent
loading divided by the previous maximum load (see
ASTM E 1316 for details), and no rapid increase of AE with
stress.
Spacecraft materials are often subjected to severe tem-
perature conditions and gradients of fast heating and cooling.
This leads to thermal microcracking and delamination due to
the mismatch in thermal expansion coefcients of the vari-
ous constituents of a composite material. Potential materials
for a reusable aerospace vehicle such as carbon-reinforced
bismaleimide or polyimide thermostable matrix composites
have been studied (Favre and Raud 1995) by thermal cycling
between +250 and 110 C with low and high (to 15 C per
min) cooling rates. Special heat-resistant detectors were
3.2 The Materials Laboratory
Fig. 3.27 Acoustic emission sensors located along SPELDA top ring
during compression testing in the Load Rig (courtesy of Physical
Acoustics Ltd., UK)
applied to the samples. AE data were collected throughout
the test period, and it was found that AE was concentrated to
the cooling down part of the cycle, and when the samples
were reheated from 110 C these emissions were seen to
stop. This research was also followed by quantiable
radiography and metallography, but the overall results
demonstrate that very complex time- or environment-
dependent parameters exist, and a much better understand-
ing of these materials is needed before they can be used in
actual applications.
Other applications related to AE testing of spacecraft
parts include:
SMART tanks with embedded sensors and on-board
monitors to provide for safe, high-pressure gas or fuel
storage tanks during spaceflight;
an ESA study of a crack detection system for space
station using projectile impingement onto the internal
surface of its aluminium module and recording of nom-
inal beginning of life characteristics from AE sensors
situated in critical locations around the structure.
Unusually high service loads or any crack initiation is
supposed to alter the AE characteristics of the module,
and digital analysis should indicate where the defect is
93
located, even if it is inaccessible due to panels and
equipment;
the detection of cracking in ceramic electronic compo-
nents during thermal cycling or under bending of PCB
assemblies (Prymak and Bergenthal 1995).
3.2.3.5 Methods for Determining Residual Stress
(a) Causes of residual stress
Residual stresses are the macroscopic stresses that are set up
within a metal as a consequence of the introduction of
non-uniform plastic deformation. Usually these stresses are a
result of cold working the metal, or originate from steep
temperature gradients during casting, heat-treatment, or
welding. The residual stresses may also occur in spacecraft
hardware owing to the heat dissipation from electronic
component packages during operation or by in-orbit solar
heating. Here, the stresses may originate from the differences
in coefcient expansion of joined materials and can, in the
case of soldered interconnections, lead to creep and thermal
fatigue failures. The presence of residual stresses may ini-
tiate stress corrosion cracking if susceptible alloys are
exposed to damp environments. They can also promote the
precipitation of hard phases within the grains of metals; for
instance, titanium and its alloys can be embrittled by the
formation of titanium hydrides in the presence of macros-
tress and interstitial hydrogen atoms. These, and many other
examples of the damaging effect of residual stresses, are
presented throughout this book.
Welding is the biggest contributor to unwanted residual
stresses in metals due to the large thermal gradients associ-
ated with this process. Some idea of the two-dimensional
stresses caused by welding can be seen from the sketch
shown in Fig. 3.28. The most critical space hardware con-
sists of welded structures. Consider the long-life require-
ments coupled with no leak permitted in the pressurized
structures for manned space vehicles; these are mainly
welded constructions from the aluminium alloy 2219. Also
the pressure vessels are designed to store different fluids and
propellants for launch vehicles and satellites, made typically
from either AA2219 or the titanium alloy Ti6A14 V. The
welding operation not only changes the microstructure of
these alloys and modies their mechanical properties, but it
also introduces stresses that can be either positive or nega-
tive and of such high levels that the materials yield point is
exceeded, so causing the part to deform. In the case of a
titanium propellant tank, it is important that the total stress
levels in the weld region are well known so that engineers
can ensure the safety of the tank during its storage, as it is
dynamically stressed in the launch phase, as well as during
operational life. The total stress in the welded region (parent
94 3 The Integration of Materials into Product Assurance Schemes

Fig. 3.28 a Temperature

distribution at different times in a
butt-welded plate and the related
shrinkage process. b Distribution
of longitudinal (1) and (t)
residual stresses in same
butt-welded plate

metal, heat-affected zone (HAZ), and weld bead) will be the
summation of the residual stress, and those stresses con-
tributed by the weight of the tank and its propellant.
In the case of 33 mm thick forged AA2219 plate that had
been multipass TIG welded with 2319 ller metal, extremely
high residual stresses caused the plate to bend, conrming
that the 280 MPa yield point of the ller metal had been
exceeded (Verderaime and Vaughan 1995).
With thin-walled titanium pressure vessels (yield
point = 820 MPa) it is known that in the residual stresses in
the HAZ of the girth welds are hoop tension and longitudinal
compression. Immediately after welding the hoop stress can
reach 375 MPa. This value was obtained from hole-drilling
strain-gauge techniques described in (b) below. After a
suitable stress-relief heat-treatment there is a remarkable,
order of magnitude reduction in residual stress down to
about 37 MPa (Haupt 1996). Neutron diffraction mea-
surements for a similar tank design, but produced by a dif-
ferent manufacturer (see (d) below), indicated lower
3.2 The Materials Laboratory
as-welded tensile stresses and a less effective stress-relief
treatment.
The main methods for measuring residual stress, in their
order of industrial popularity, are described below.
(b) The hole-drilling strain-gauge method
This is dened by the ASTM method E837 (1994) and
consists of attaching a special three-element strain-gauge
rosette to the surface of the metallic piece under investiga-
tion. As the gauge is quite small it can be used on surface
positions in the weld metal, HAZ, or anywhere on the parent
metal. It is usual to initially dress the weld bead in order to
achieve a reasonably flat surface for strain-gauge attachment.
The three gauge grids are wired and electrically connected to
a static strain indicator. A small, shallow hole is then drilled
through the centre of the rosette in small increments. The
introduction of a hole into the residually stressed metal
relaxes the stresses at that location, causing the local strains
on the surface to change correspondingly. Readings are
made of the relaxed strains and, by means of special charts,
these are converted into values of residual stress (in three
angular orientations). The accuracy of this method is con-
sidered to be 10 %. The method is widely applied in
Europe and the USA. An interesting case study related to the
NASA MARS Global Surveyor propellant tanks has been
described by Haupt (1996), who concluded that this was a
practical method, being cost-effective and the least destruc-
tive technique for gathering residual stress data.
(c) X-ray diffraction method
When external or residual stresses are applied to a structure
made of metallic grains, the crystalline lattice is distorted
and the interatomic distances are changed. As the stress is
increased, so the interatomic spacing will be proportionally
increased. Monochromatic X-rays are directed onto the
sample under investigation, and these rays are then reflected
from successive parallel planes of atoms in the metals grain
structure. The reflected rays are reinforced in one direction,
which is called the diffraction direction. The angle between
the incident or reflected beams and the surface of the
reflecting planes is measured by either lm or goniometer
techniques. Braggs law denes the condition for diffraction,
and by means of a simple equation the interplanar spacing,
d, can be calculated. In order to evaluate the strain from
the interplanar distances or from the lattice parameter, a
reference sample free of strain is needed. The standard way
to achieve this is to grind material from the various parts of
the sample, particularly the weld metal and the parent metal,
as these may have different compositions. The powder pro-
duced is thermally treated and slowly cooled to room
95
temperaturethis is considered to be the stress-free refer-
ence. X-rays have a low depth of penetration, and only very
slightly subsurface values for residual strain can be made
with this method. A full description of the X-ray diffraction
method for residual stress measurement, together with many
case studies, is given by AMS International (1991).
(d) Neutron diffraction method
The principle is similar to that of the previous X-ray method
except that neutrons have a 1000 times greater penetration
depth. Also, absolute strains in materials can usually be
calculated from measurements made at any depth from the
samples surface. In the case of neutron diffraction the
specimen has to be taken to a nuclear reactor. Stresses, for
instance in the various local regions of a weld, can be
mapped by studying how the neutrons are scattered by the
planes of atoms in the solid weld material. The neutron beam
has a xed wavelength (usually 1.65 ), selected to make
the scattering angle close to 90. The strains are again
determined by observing the diffraction angles and calcu-
lating the lattice spacing.
Peaks in the diffracted intensity are observed when the
Bragg condition is satised. This condition is given by the
expression
k 2d sin h
where is the neutron wavelength, d is the interplanar
spacing, and 2 is the angle between the incident and dif-
fracted beams.
In a state of tension the interplanar spaces are expan-
ded. The strain, , is a dimensionless quantity, the frac-
tional change of d from that in a stress-free reference
sample, d0:
 
d
e 1
d0
Thus, careful determinations of 2 and give direct mea-
surements of d and the strain.
A three-dimensional map of strains in a series of volume
elements in the weld, the HAZ, and the base metal can be
generated. From this the residual stresses in each of the three
principal directions can be calculated. Further information is
contained in Hutchings and Krawitz (1992).
The technique has been successfully applied to calculate
the residual stresses and strains in several spacecraft appli-
cations. Titanium alloy propellant tank weldments were
measured; initially residual stresses approaching 250 MPa
were recorded, post-weld heat-treatments reducing this to
about 50 MPa. AA 2219 welded plates and AISI 304
stainless steel components were also evaluated. These
96
examples are shown in Figs. 3.29a and b, courtesy of Prof.
F. Rustichelli, University of Ancona, Italy.
The Rustichelli work has been extended in order to
measure the residual stresses in Ti6Al4 V fuel tanks fabri-
cated for space use by the welding together (TIGtungsten
inert gas) of two cold drawn half-cylinders supplied by the
author (Bruno 2000). Three welded curved samples were
investigated, each having different thicknesses. The welded
6 and 2 mm samples behaved similarly (hoop stresses being
300 MPa in the weld, and 100200 MPa in the parent metal),
whereas the thinnest 1.6 mm sample possessed lower,
180 MPa hoop stress in the welded zone, and zero residual
stress in the parent metal.
Very little work has been done on Ti6Al4 V welded
spacecraft tanks so the investigations were resumed on two
further 2 mm thick curved samples and the results compared
to the weldment microstructures. These samples were stress
relieved after TIG welding (485 C for 4 h in a 104 torr
vacuum). The main result found that this stress relief treat-
ment reduced the weldments hoop stress from 300 MPa
down to about 160 MPa (Bruno 2004).
Thermal spraying is widely used in industry to apply
surface protection to engineering components. This process
has been considered for space vehicle applications, including
composite, ceramic and metallic coatings for protection
against heat as encountered during re-entry into the Earths
atmosphere, or entry through the atmospheres surrounding
distant planets and their moons. Such coatings have been
evaluated (Luzin 2010) for residual stresses by means of
neutron diffraction; the evolution of residual stresses during
deposition were established for thick deposits (1030 mm)
down to thin deposits having a thickness of 0.5 mm. Anal-
yses found stress distributions through the deposits due to
the thermal mismatch that occurred between the substrate
materials and the coating. This is complicated by the
quenching stresses created by the molten particles as they
rapidly solidify on impact. The interfacial stresses can be
monitored and when high, spalling of the coating from the
substrate can be expected. However, spalling, or detachment
may not occur if there is some solid solubility between the
coating material and the substrate (without the formation of
brittle intermetallics) and here, for metals, the relevant phase
diagrams need to be consulted.
(e) Less common methods
These have been briefly covered by Wedgewood (1994),
but are unlikely to nd wide applications. An ultrasonic
method depends on the small changes to the ultrasonic
velocity as ultrasound is passed through the material. The
change in velocity is very small, about 10 ppm, and is
caused by the strained atoms as they are pulled from their
3 The Integration of Materials into Product Assurance Schemes
equilibrium positions. Other methods depend on the material
being ferromagnetic, so they may be useful for steels and
nickel alloys. Stress has been shown to affect magnetic
parameters including a materials permeability, coercivity,
and redistribution of magnetic domainsall changes to
these parameters from values made when the material is in a
stress-free condition can be related to both applied and
residual stress levels.
3.2.3.6 Modern Techniques for Surface Analysis
A great number of surface analytical techniques have been
developed since the mid-1960s. All rely on the evaluation of
samples under ultra-high vacuum conditions in order to
eliminate the effects of surface contamination by residual
gases. The scanning electron microscope (SEM) has become a
common laboratory tool, and it has largely replaced the
transmission electron microscope owing to its ease of opera-
tion. The characteristic X-rays generated at the surface of a
specimen by the incident electron beam in the SEM can be
utilized to display or analyse the chemical elements present at,
and slightly below, the materials surface. The signals are
usually detected by an energy-dispersive X-ray analytical
(EDAX) system. Several operational modes exist, and these
have been detailed by Goldstein et al. (1981). The
EDAX-system (see also Sect. 3.2.2.3) has an elemental sen-
sitivity of approximately 0.5 % if calibrated against good
standards, a lateral resolution of 50100 1010 m. This
method is virtually nondestructive to the specimen surface,
and under normal operation local heating effects will be
introduced only if electrically insulating materials are
examined.
A battery of alternative methods for the surface analysis
of materials is available to the materials engineer; each has
its own merits, and although Table 3.1 is introduced as a
means for the initial selection of certain analytical tech-
niques, it is stressed that expert advice must be solicited for
any particular application. Fourier transformation infrared
(FTIR) spectrometers are particularly useful for analysing
organic residues and examples of this technique are given in
Fig. 3.30, and in Fig. 5.76. The FTIR equipment described
by Smith (2007) has uncovered many examples where
organic species have had a detrimental effect on electronic
equipment at various stages of their livesthe case studies
are a recommended read and cover: acrylic adhesive and
paint residues found to damage ball bearing races in an
electro-optic system and stepper motors; brown residues of
rosin on printed circuit boards which were identied as rosin
derived from pine tree resin rather than synthetic resin;
methyl cellosolve transferred during pcb cleaning to pene-
trate a connector and cause electrical insulation breakdown
between the connectors pins; degradation of greases due to
long term storage; and, carbonate deposits on electronics
operating in a sea-coast environment.
3.2 The Materials Laboratory 97

Fig. 3.29 a Behaviour of the

lattice parameter of the
aluminium alloy 2219 across a
welded plate. The 7 mm thick
plate had been welded by the
variable pulse plasma arc method
being used for manned spacecraft.
The weld bead and root have been
machined off and polished flat.
This 50 mm trace, running normal
to (across) the weld path, depicts
the lattice constants at a depth of
3 mm from the top (head) side of
the plate. The highest residual
stresses are compressive and exist
at the weld metal-to-HAZ
interface. (Methodology to be
publishedRustichelli and Dunn
1997.) b. Multipass welded AISI
304 stainless steel plate, having a
thickness of 35 mm (the weld
head is intact). The colour image
displays a map of the iso-strain
levels in the <111> direction of
the crystallographic planes
98 3 The Integration of Materials into Product Assurance Schemes

Table 3.1 Summary of the characteristics of certain techniques for the analysis of surfaces
Characteristics Analytical technique
SIMS EPMA PIXE TEM ESCA AES LEIS HEIS (RBS) LMP LOES
Element range * *
Detection limit *
Quantitative analysis * * *
Information depth STa *
Dynamic range *
Instructivityb
Lateral resolution *
3D-imaging *
Compound information *c *
Atom location in lattice * *
Isotope separation * *
Organic samples
Courtesy of Mr H. Werner and Dr P. von Rosenstiel of the Stichting Geavanceerde Metaalkunde, Hengelo, The Netherlands
Key to table
Performance indicator:Not reported or not possible: Poor; Moderate; Good; * Excellent
a
ST = limited by sample thickness
b
Performance indicator for destructivity: intrinsic, moderate, weak
c
By electron diffraction
AES Auger Electron Spectroscopy
EPMA Electron Probe Microanalysis
ESCA Electron Spectroscopy for Chemical Analysis
HEIS High Energy Ion Scattering = RBS = Rutherford BackScattering
LEIS Low Energy Ion Scattering = ISS = Ion Scattering Spectroscopy
LMP Laser Microprobe
LOES Laser Optical Emission Spectroscopy
PIXE Proton Induced X-ray Emission
SIMS Secondary Ion Mass Spectrometry
TEM Transmission Electron Microscopy

3.2.4 Organic Chemistry and Environmental
Test Laboratories
These laboratories are essential support facilities for the
testing and evaluation of space materials. Most space
agencies and prime contractors will provide laboratory ser-
vices and possess equipment to follow standard test meth-
ods, but the true space labs will have dedicated systems for
more exotic tests, those that produce accelerated ageing and
those providing for the simultaneous exposure of samples to
components of the space environmentso-called syner-
gistic testing. Analytical laboratories will involve special
testing techniques and instrumentation. These sophisticated
procedures necessitate that trained and learned operators are
available in order to obtain accurate, optimized and relevant
results that will be useful to the projects requesting data.
Numerous reference books can be consulted for detailed
descriptions and space applications for these techniques
(Dauphin 1992; Bosma 2009; Pretsch 2009; Gross 2011;
Lund 2011; Kleiman 2013; Voigtlander 2015).
The primary functions of an analytical chemistry labora-
tory are to determine thermal, optical and chemical properties
of organic materials. Activities and analyses will include:
Optical Spectroscopy: Fourier Transform Infrared Spec-
trometry (FT-IR) for surface analysis, high resolution
infra-red-imaging, depth proling, reaction kinetics, quality
control, quantitative determination of chemical species, a
collection of IR-reference spectra for structural
assignment/conrmation will be available from databases.
Liquid and Gas Chromatographic Techniques: determi-
nation of the type of monomer in polymers, structure elu-
cidation, quantitative determination, quality control, Mass
Spectroscopy reference spectra will be available from
databases.
Elemental Analysis: by Atom Adsorption Spectrometer,
X-ray fluorescence and X-ray diffraction.
Environmental test and screening facilities have been
described elsewhere in this book, for instance the micro-
VCM test which provides quantitative gravimetric measure-
ment of outgassing and condensable materials was illustrated
3.2 The Materials Laboratory
Fig. 3.30 Analyses of human contaminants. Fourier transformation
infrared (FT-IR) spectrometers are able to provide a rapid analysis of
various organic and polymer contaminants which may be found on the
surfaces of critical spacecraft hardware. This has been made possible by
developments in the speed and power of small, affordable computers.
Libraries of standard spectra are stored in the computer memory and
may be compared against spectra from contaminant species by
relatively simple search routines. Two case histories are illustrated
(a) shows the optical photograph of a ngerprint on a polished steel
mirror together with the FT-IR spectrum and its analysisthe nger
marks contain some water and typical skin secretions. A threaded steel
in Sect. 2.5. The condensable materials that are collected
from the outgassing test can be chemically analyzed by
comparison to spectra available in databases as shown in
Fig. 3.31. The dynamic outgassing testing of samples,
material pieces or individual components, can also be mea-
sured by specialized equipmenthere the sample is con-
tinually monitored for mass loss against time, rstly at 25
C, then by steps of 25 C every 24 h to a temperature of
125 C. This test can provide data about the long term effect
of the space environment on a material, so supplementing
the more simplistic VCM screening test (Rampini 2009).
The effect of Atomic Oxygen on materials and parts will
be reviewed in Sect. 8.5 and have been extensively reported
by Banks (2003). The atomic oxygen existing in the low
Earth orbital environment has formed by the
photo-disassociation of diatomic oxygen. These oxygen
atoms, O+, are highly corrosive with most of the materials
they contact. When a spacecrafts orbital velocity is
99
fastener was also submitted to FT-IR analysis following the discovery
of particle contaminants within its threads. The SEM image is shown in
(b) together with the IR spectrum of the white particlesa skin-flake
reference spectrum is also shown for comparison. EDAX analysis also
conrmed the presence of the elements: S, CI, K, Ca, and Na-detectable
ingredients of skin for this method. Further investigations indicated
these to be small scales of dead skin (dandruff) (courtesy of Torbjorn
Lindblom, Celsius Materialteknik, Karlskoga, Sweden.) a Fingerprint
contamination on polished mirror and analysis. b Skin flakes in bolt
thread and analysis
considered to be 8 km/s the stream of atomic oxygen on the
front surface has a charge energy of approximately 5 eV.
The interaction of atomic oxygen causes an optical glow
close to the spacecrafts surface. Few laboratories are able to
produce an atomic oxygen (AO) flux and the majority of
realistic assessments of AO on materials have been made by
the examination of samples returned from space, as will be
shown in Sect. 8.5. Ground laboratory Atomic Oxygen
testing generally follows the ASTM E2089 (2014) standard
practice. Facilities designed and commissioned for
on-ground AO exposure include ESA-Estec, Noordwijk;
Physical Sciences Inc., MA, USA; Integrity Testing Labo-
ratory, Markham, Canada; and, SAS Lab., Rome. AO pro-
duced by methods that include the dissociation of molecular
oxygen using a pulsed laser source. The lab typical AO
energy is about 5.5 eV and temperature of sample can be
controlled between 150 and +150 C in a chamber under
vacuum of less than 105 mbar.
100
Fig. 3.31 Example of chemical analysis by infrared spectroscopy.
a Photograph showing the collected volatile condensable material as it
has condensed onto the central area of a standard test plate (ECSS
Q-ST-70-02 denes the outgassing-under-vacuum test). b Sample of an
IR spectroscope reading, showing the condensed species peaks around
3500, 2200 and 1600 cm1 on the horizontal axis. These peaks
3.3 Preparation of Materials
and Metallographic Evidence
3.3.1 The Metallographer
Metallography is sometimes thought of as black magic
by project management, but to the innovative engineer the
results are actually a combination of art and science. Metal-
lography has come to embrace ceramics, polymers, and many
inorganic substances as will be examined in such diverse elds
as mineralogy, forensic science, and the study of particular
matter in the pollution eld. To this extent metallography
might more aptly be termed materiallography, the practice of
gaining information on composition and properties, as well as
material structure. The dexterous optical-metallographer will
note the relative hardness of the phases in an alloy from their
behaviour during grinding and polishing, will note their
chemical behaviour during immersion in different etchants,
and will observe their anisotropy under polarized light. He will
also have a knowledge of optical systems and the application
of the metallurgical microscope as a working tool in the pro-
duction of photomicrographs. When the eye is functioning at
its best it can separate ne detail as small as 0.1 mm. By
employing the optical microscope with a monochromatic light
source, detail of the order of 0.2 m can be resolved. The
electron microscope has enabled particles down to 0.1 nm to
be seen, as illustrated in Fig. 3.9.
3 The Integration of Materials into Product Assurance Schemes
correspond to the database stored spectrum for a commercial epoxy,
they have no resemblance to the fluorocarbon bre contained in the
sample (which had not outgassed/condensed). ECSS Q-ST-70-05
denes the method for detecting organic contamination of surfaces
by IR spectroscopy (courtesy ESA)
The metallographer should also be able to produce pho-
tomicrographs of the highest quality, combining technical
skill with artistic visionthus producing micrographs which
are scientic and educational expressions. Success in the use
of metallographic techniques in the spacecraft eld will, as
for many other advanced industries, depend on a knowledge
of several other disciplines. A correct interpretation of the
microsection and its micrographs will necessitate an
exchange of views between metallurgists, composite mate-
rial engineers, electrical engineers, structural engineers, and
so forth.
Objectivity is one of the two most important ingredients
in any failure investigation (experience is the other), and this
will mean that the investigator/metallographer must keep an
open mind and an absence of emotion. It is important that
personal opinions are not held at the beginning of an
investigation, as this might result in the seeking of infor-
mation to conrm that opinion. Emotion can run high if
personal injuries or substantial spacecraft damage have
occurred to persons or equipment known to the investigator
before the failure. Also, during the course of the metallur-
gical troubleshooting it is important that the investigator
does not pass on conjecture, initial ideas, or preliminary
conclusions to project personnel. He will be held responsible
for the nal report and must not need to justify interim
opinions that, in a complex failure, may become modied as
the investigation proceeds and could have no bearing on the
nal conclusions.
3.3 Preparation of Materials and Metallographic Evidence
3.3.2 Laboratory Records and Reports
One of the most important documents generated during
laboratory studies, whether for a PA function or as a basic
research and development activity, is a properly kept labo-
ratory notebook. The steps taken in the DPA, failure anal-
ysis, or routine investigation need to be accurately recorded
to prevent the loss of ndings, to validate the quality of
experimentation, to enable the work to be repeated in an
identical manner by other laboratories, and to prevent
repeating the work at some time later because it was not
known that the work had been tried and abandoned.
In the space industry it is common that one of the QA
tasks is to ensure that laboratory notebooks are kept updated
by technical staff. The records of laboratory activities,
together with the records of meetings with engineers from
other disciplines, will be periodically required for compila-
tion into a wide range of documents ranging from monthly
reports to ofcial technical reports, working papers, papers
for publication in journals or at conferences, and nally as
data to be presented at material review boards and failure
review boards.
The materials and processes tasks performed in the space
industry are sometimes of a routine nature, but it is the
authors experience that certain test failures, or last-minute
malfunctions, can be diagnosed only by a crash laboratory
investigation when operators may be required to work
overtime at night or during weekends. It is upon these lab-
oratory ndings that a spacecraft project manager will base
important decisions. They may determine whether a
multi-million pound spacecraft will be launched as sched-
uled, or postponed pending further work or equipment
changes. The accurate and extensive recording of experi-
mental or test data in laboratory notebooks is therefore
essential. Sample traceability has been coved previously in
Sect 3.2.2.2an example of good met lab practice was
illustrated in Fig. 3.7b).
The author has found it necessary to apply the following
guidelines. When adhered to, they should obviate continuity
problems that might arise when staff are absent on leave or
sickness:
(a) Laboratory notebooks should contain bound,
pre-numbered pages (to assure management that vital
records are not missing because they were not led).
(b) All entries must be clearly written, and in a multi-
national organization, preferably in the English or
French language (a notebook is of no use if an entry
cannot be read).
(c) The use of permanent ink is preferred (this will also
facilitate photocopying).
101
(d) Data should not be obliterated or erased. It is better to
cross out an entry with a single line so that it is still
readable (the original readings may be required again if
instrument fluctuations are found at a later date etc.).
(e) The initiator of work project, all reference to material
history, environmental test parameters, time of failure,
etc. should be recorded.
(f) All preliminary examinations of the item and visual
inspection results should be recorded, and special sur-
face features photographed; then possible nondestruc-
tive tests performed and their results recorded; nally, a
record should be kept of how the item was cut to aid
further SEM and metallographic examination (e.g.
precise orientation of cross-section), the mounting
media (particularly if under pressure, or with exother-
mic resin), polishing media, magnication of micro-
graphs, etc. (this will permit a similarly qualied
worker to reconstruct and repeat the work as originally
performed).
(g) Supporting records should be glued, stapled, or taped
into the laboratory notebook (particularly
digitally-stored photomicrographs, graphs, hardness
results, etc.).
(h) Adequate references to procedure specications, other
records, drawings, sketches, negative numbers (this
will facilitate upgrading of results into the format of an
ofcial report) should be made.
3.3.3 Report of Materials Data to Spacecraft
Projects
A book concerning the uses of applied materials science and
techniques in the spacecraft manufacturing industry would
not be complete without a short account of the role of report
writing. A metallurgists or chemists work is never nalized
until he/she has described what has been accomplished by
the all the practical examinations, and this will necessitate
the writing of a nal ofcial technical report for his/her
organization. The impact of the report will be reflected by its
contents and the format in which it is presented. It is the
authors experience that materials examinations and studies
for a spacecraft project are usually performed to establish the
quality of a product, to determine a mode of failure, or to
assess the suitability of a materials or device for space use.
Ofcial reports will be based on the detailed entries con-
tained in the laboratory notebook, the contents of which
were discussed in the previous subsection. When novel
experimental or test techniques were employed, a full
description is not required, but it is preferable to list specic
102 3 The Integration of Materials into Product Assurance Schemes

references that readers could consult for more detailed
information. As internal reports, they will need to be written
for internal reasons, and only a summarized nal document
may be required. This will be circulated so as to feed back
information to other departments, to justify design changes,
and for PA archival purposes. Long records of details will
not be needed, and in most cases the recipients will only be
interest in, or have the time to read about, the state of the
engineering problem in the introduction and what has been
done to solve it in the conclusion or summary.
In any presentation, photographs, sketches, and photo-
micrographs are extremely helpful. When they are suitably
captioned they can reduce the need for long descriptive
narratives. They will also be useful during oral presentations
when materials or metallographic data will need to be shown
and discussed at material review boards and failure review
boards.
Frequently it will be necessary to dispatch material to
chemical laboratories, (e.g. Sect. 3.2) for testing or accurate
analyses of the constituent polymers or alloys. The presen-
tation of the major alloy elements may be important to
identify the alloy designation and, from that, the supplier.
Aluminiumlithium alloys may soon nd some spacecraft
applications. It can be seen from Table 3.2 that AlLi fully
heat-treated sheets can have large density and mechanical
property variations due to very slight modications in their
elemental composition.
Materials review board (MRBs) (increasingly termed
NRB, non-conformance board) consist of members respon-
sible for engineering production. P A, and materials or
components specialists. Such boards meet to critically
examine new materials, new fabrication processes, and a
wide assortment of non-conformances. The MRB will
frequently draw upon the metallographic ndings of a lab-
oratory study when considering the dispositioning of articles
or materials. These dispositions will normally fall into four
categories, such as:
suitable for use on spacecraft,
suitable only after article has been repaired or
refurbished,
not acceptable, return to suppliers, and
not acceptable, scrap.
The failure review board (FRB) is convened to review
failures, evaluate failure analyses of the kind illustrated in
this book, and establish corrective action plans. Again,
specialist reports are discussed, and the recommendations
made as a result of the metallographic studies are usually
accepted. Corrective actions might include engineering
changes, materials changes, and inspection or test procedure
changes.
The metallographic presentations at both MRBs and
FBRs are required by QA and reliability disciplines, but they
will always be important to the design and production
engineers who were responsible for the choice of material
and manufacture of the part. Care must be taken to avoid
laying the blame for a problem or failure on any individual
person. The oral presentation or report must be conned to
metallurgical and other technical considerations, and it is
preferable to avoid reference to political areas. This will (or
should) be taken up at a higher level in the organization. It is
also advisable to give maximum credit to the production
people when metallographic examinations nd items have

Table 3.2 Compositions and properties of AlLi alloys listed according to density and compared against the main spacecraft structural alloy, AA
2219
Alloy Elementsa (mass%) Grain reners E Rm Rp0.2 A
Li Cu Mg in g/cm3 in GPa in MPa in MPa in %
1420 1.92.3 0.05 4.56.0 Zr: 0.080.15 2.47 74.8 420430 260270 1012
8090 2.22.7 1.01.6 0.61.3 Zr: 0.040.16 2.56 77.9 410430 290330 912
1440 2.12.6 1.21.9 0.61.1 Zr: 0.10.2 2.56 79.4 450470 380410 1114
1430 1.51.9 1.41.8 2.33.0 Zr: 0.080.14 2.57 79.0 430450 320350 913
2090 2.12.2 2.52.75 0.0170.02 Zr: 0.110.12 2.57 79.0 448540 390490 45
2091 1.72.3 1.82.5 1.11.9 Zr: 0.040.16 2.58 77.2 430460 330370 1014
2090 1.92.6 2.43.0 0.25 Zr: 0.080.15 2.59 79.3 530560 490520 58
1460 2.02.5 2.63.3 0.05 Zr < 0.15, Sc < 0.14 2.60 81.8 530550 480500 79
b
Weldalite 049 1.3 4.06.3 0.4 Zr: 0.14, Ag: 0.4 2.71 77.9 560590 520540 57
2195 0.81.2 3.74.3 0.250.80 Zr: 0.080.16 2.72 76.0 525615 490580 912
2219 None 6.3 0.02 Zr: 0.100.25 2.85 73.5 435440 320330 78
a
Al-2095. All alloys contain Ti < 0.15 %
b
Maximum density and typical tensile properties for sheets after full heat treatment are given
Weldalite series includes 2049, 2095 and 2195
3.3 Preparation of Materials and Metallographic Evidence
been well-made (spacecraft parts are often one-off items and
may consist of complex shapes made from unusual,
difcult-to-machine materials). It is important that the
materials engineer does not take on the reputation of being
an over-critic. Full cooperation between all personnel will be
needed during future troubleshooting exercises, and these
are hindered when police tactics are used by PA laboratory
staff.
In summary, keep clear and legible records; never spec-
ulate publically as to the potential causes of a failure, par-
ticularly if the work has not been completed; never make
casual or flippant remarks on any aspect of your work; do
not criticize colleagues. In the space business if there is an
accident, the engineer is keen to discover the cause. It is the
writers (fortunate) experience that spacecraft accidents do
not result in disputes, arbitration, and costly litigation
involving the legal profession. This would have a negative
effect, as lawyers want to determine who was responsible
and this could cause conflict. In other elds, engineering
disputes and claims are part of business and, here, the
reporting of materials data may be unfavourable to the client
who may decide to bury the report and look elsewhere for
another opinion (Allen and Jubb 1991).
3.3.4 Training of Materials Engineers
and Laboratory Staff
Visits to different materials or metallurgical laboratories in
Europe have shown some national differences concerning
the deployment of staff. Certainly differences exist which
varied according to the size of the facility; however, it is felt
that continental companies, particularly in Germany and
France, placed a much greater emphasis on machine-shop
crafts, understanding of laboratory equipment, and inter-
mediate skills than is experienced in Britain or Scandinavia.
In all laboratories there is an amplied role for technicians,
and whereas some continental companies will upgrade their
technician staff by sponsorship through technical university
courses, the converse is also seen, whereby university
graduate engineers are deployed in technician roles. Many
graduates are too specialized or research-oriented to become
easily socialized into an industrial or commercial labora-
tory, and they have difculty coping with the upkeep and
repair of modern laboratory equipment, liaison with service
departments such as machine-shop personnel, or even
operation of simple lathes for the preparation of specimens.
When graduates are flexible and do integrate well into a
company laboratory it is found that small rms provide for
more congenial surroundings, and this promotes a higher
level of loyalty and more successful on-the-job training than
the larger, more anonymous rms with their systematic
training programmes. Materials technology students are
103
often found industrial placements as part of their university
course (Puig 2006). Usually material student trainees will
nd themselves somewhere between engineering, spacecraft
integration and quality assurance teams. They will spend
vacations or up to one year receiving such training in prime
and sub-tier space companies, being paid and supervised so
as to make full benet of the study or managerial function
undertaken. Topics covered can range from wire harness
assembly methods to metallurgy, composites, surface treat-
ments, bonding and testing materials in a simulated space
environment. As observed by Puig, this is a win/win sit-
uation; often the student obtains enough information and
laboratory data to be able to submit his/her thesis, while the
company may benet from a well-researched report, better
access to local university facilities and an insight as to
whether the student could be a potential candidate for
post-graduate full employment.
The use of a smaller test-house laboratory generally gives
a better customer service, being more competitive and more
personalized. Small laboratories may, at the same time and
on account of their size, be in a weaker position than the
larger laboratories as they need to concentrate on the certain
research or test disciplines that they do best, and a contra-
dictory situation will exist if they attempt to cover them-
selves too widely.
When complex spacecraft technologies need to be
assessed in a materials laboratory there is a need for both
good technician resources and graduate materials specialists.
So-called generalists may also be recruited to coordinate
the various work orders or investigative projects; these staff
may know less of the ingredients of the different disciplines
but more about how they relate together. The skilled tech-
nicians must understand the objectives of the laboratory
work, particularly if it is part of a technology assessment or a
materials failure analysis, so that they can intervene and alert
the specialist during an investigation whenever a new, or
unexpected, situation arises. It is particularly important that
materials lab staff understand the full history of the sample
they are preparing so that facts are not lost during, for
instance, the specimen cleaning stage, or during
hot-mounting when microstructural evidence can be easily
destroyed by the introduction of heat and pressure. The
integrity of the material sample preparation method is also
receiving more attention, particularly when quality controls
and failure investigations question aspects related to coating
porosity, the particulate congurations within metal matrix
composites, and the presence of microcracks within envi-
ronmentally tested electronic components and interconnec-
tions. The principles of material investigations, such as
metallography, chemical analysis, fracture mechanics test-
ing, etc. are taught at Technical Universities and Colleges,
and an extensive library of video courses has been produced
by the American Society of Metals International so that rms
104
can run on-site training with the help of acknowledge
experts. It is also clear that the staff of test laboratories all
around the world are making a growing use of the so-called
information highway afforded by Internet. These staff,
including those of the author, are nding other materials
engineers and scientists with similar interests.
There is a need for laboratory staff to be retrained every
ve or ten years. The active life of a rst university degree
can be as short as ve years according to Professor Goodhew
of Surrey University, and owing to the rapid pace of tech-
nological advance a materials graduate may need to attend
refresher courses six times during his/her entire career. New
material developments are continually occurring, particularly
in the elds of carbon bre composites, high-performance
oxide ceramics, high-temperature superconducting ceramics
particularly for large passive electronic devices such as
RF cavities and antennae, super alloys manufactured by inert
gas automization, and hot isostatic pressing, etc. Engineers
and technicians must also be kept up-to-date with new lab-
oratory techniques that may be available only for loan or hire
at universities or large institutes. These methods may include
assisted operation of a scanning tunnelling microscope, laser
acoustic microscope, the mechanical testing of spacecraft
metals and components made at 4.2 K in specially con-
structed cryostats, or the use of radioactive trace techniques
for the investigation of absorption rates of, for instance,
water, by the surfaces of materials following exposure to
space simulation environments. New concepts in metallo-
graphic preparation may also need to be established, par-
ticularly when small, unique samples are presented for
failure analysis. Here there is no opportunity for the tradi-
tional trial and error approach, and a systematic sequence
of grinding and polishing will require the selection of pol-
ishing cloths, abrasive materials and their size, speed of
polishing, and pressure for different groups of difcult
materials (Bouseld 1988, 1992). In short, read the litera-
ture, go to the talks and conferences, and when and where
necessary augment formal training with comprehensive
courses.
Finally, it is also important that there should be a strong
interaction between the metallurgist, chemist or materials
engineer and the mechanical, electrical, or aerospace engi-
neer responsible for the hardware being evaluated. This will
enable the problem-oriented topic to be adequately addres-
sed with a good mutual understanding of design, product
processing, or service operational environment requirements
an extremely useful form of on-the-job training.
3.3.5 Ethical Issues
Professional engineers and laboratory staff, as will be noted
in Sect. 3.5, are expected to pay regard to the environment
3 The Integration of Materials into Product Assurance Schemes
and sustainability of resources. But also during their work
which can involve dangerous lab chemicals, hazardous
materials at the launch siteconsiderations must be given to
personal welfare, health and safety. It may be necessary to
make a risk assessment before performing dangerous tasks.
For instance, take care/precautions while sawing and pol-
ishing beryllium, and handling vessels that might contain
residuals such as hydrazine. Materials and processes engi-
neers hold a privileged and trusted position in society and
therefore should only perform services within their areas of
competence. As previously stated it is essential that their
hierarchy understand the need to keep up their skills and
knowledge bases. Evidence gained from lab examinations,
studies and failure investigations should be interpreted and
reported honestly, accurately and without bias. These state-
ments may seem to be obvious, but under pressure from
project or clients management, engineers and lab staff need
to avoid improper influences that may lead to cutting corners
and making conclusions based on insufcient test or his-
torical evidence.
3.4 The Future for Materials Failure
Investigations
3.4.1 The Larger Company
The majority of the larger companies engaged in the man-
ufacture of aircraft structures and components, marine sys-
tems, road vehicles, and the construction of power stations
and telecommunications equipment, will have already
invested in the capital equipment required for the smooth
running of a metallurgical failure laboratory as described in
Sect. 3.2. Failure analysis is becoming increasingly impor-
tant for a number of reasons. Some failures are very costly,
and not only can valuable equipment become destroyed but
high-volume product output may be interrupted. Downtime
can cost hundreds of thousands of pounds per day whether it
be associated with an aircraft, power station, or satellite
communications link. It is probable that in the future,
material costs for heavy engineering units will be dispro-
portionately more expensive than they are today so that
investment in the tools for performing failure investigations
will be increased. In this case the cost of determining a
failure can be justied.
On the other hand, nancial constraints now necessitate
that commercial electronic systems, such as television
receivers, computer hardware, and even land-based data
relay stations, incorporate the technique of redundant circuits
and throw-away modules. This means that there is seldom a
requirement for engineers to investigate the mechanisms that
result in failures. Other factors in this argument include the
3.4 The Future for Materials Failure Investigations
growing shortage of skilled maintenance technicians and the
wide availability of mass-produced low-cost electronic cir-
cuits: thus the throw-away maintenance philosophy could
become more and more appealing.
3.4.2 The Smaller Company
In general, it is the authors view that, within the smaller
company (e.g. small, medium enterprisesSMEs), top
management is not giving the M&P function all the attention
it deserves. One particular issue that may be especially
important in the space industry is the complaint that the
smaller spacecraft contractors, particularly those manufac-
turing mechanical devices and those in the electronics eld,
have insufcient capital to invest in all the related areas of
PA. Another reason is that frequently top management is
unaware of the need to employ a team of materials or
component parts specialists to perform DPAs and FMAs. It
is not until costly failures occur that consideration is given to
the need of an on-site materials laboratory for performing
controls on raw materials and processes, and the occasional
defect or failure analysis work.
The many cases the occurrence of failures has meant
resorting to lab testing in order to answer the immediate
question, Why?. This is not simply to satisfy curiosity but
to avoid repeating the same mistake. Certain cases have
demonstrated the forensic engineering nature of the inves-
tigation. Here, by reconstructing the sequence of events
leading to the material failure it should be possible to
establish and report on the underlying cause of the problem.
Against this background it is important that top management
in SMEs realise the contribution that laboratory work can
make to the success of spacecraft projects. If a smaller
company cannot afford to employ materials/components
specialists, or have available materials/components labora-
tories, other alternatives must be found. It is often experi-
enced that national research establishments and local
universities possess excellent facilities, and their staff are
more than willing to participate in solving space materials
problems. These experts can give advice for the choice of
materials and production methods, and they will perform
occasional metallographic inspections and other tests on
technology samples.
The extreme end of an SME-dened company includes
the successful sun-sensors specialist rm Lens R&D (foun-
ded from the ESA Business Incubation Centre and based on
technologies developed for space application within the
Dutch Micro Technology development programme). This
two- to three-man company, operating in the Netherlands,
works together with highly specialized partners (not called
suppliers). Rather than investing themselves in costly
machines which could remain idle during the well-known
105
peaks and troughs encountered by space hardware pro-
curement cycles, Lens part-nance the purchase of state of
the art equipment for their partners who, in turn, manufac-
ture and deliver piece-parts (Leijtens 2015). Engineering
design, together with electronic and mechanical assembly, is
made in-house. Product verication and space-qualication
testing (e.g. the complete range of environmental testing) is
performed in external facilities. The tasks of material and
component analyses, destructive and failure analyses, as well
as reporting, are also out-sourced as a joint collaboration
with internationally accredited laboratories. Although unu-
sual, this philosophy of working has been approved by
various space agencies.
3.4.3 Product Liability
Products made for the domestic consumer have, in the
course of their use, proved to be defective, fail, and cause
injury. It is becoming more common that Product Liability
Laws are transferring the nancial consequences of a failure
more directly onto the producer. This will, in turn, cause the
legal profession to make more use of the failure analysis
laboratory to unravel the clues, and then place responsibility
for the failure. Product liability is becoming a serious matter
in the commercial world, and it seems that these matters may
also become evident in the space industry (Thoma 1981).
The space environment is ideal for some metallurgical
processes, such as pressure or diffusion welding under vac-
uum, but the working risks will be high, so that it is unlikely
that astronauts will be called to perform routine repairs in
low Earth orbit. Secondly, spacecraft hardware that is not
situated in very high equatorial orbits may be retrievable by
means of the crew of a space plane such as the former fleet
of Shuttles or Hermes, and laboratory analyses could then
identify those individual manufacturers who produced
defective hardwarepossibly responsible for breakdown of
a multi-million pound commercial Earth resources satellite
or space observatory. It is evident that PA will continue to
play an essential role in future projects, and that product
safety policies will place great importance on laboratory
techniques to ensure a higher product reliability.
3.5 Greener Spacecraft
As a consequence of mans actions, the oceans, air, and land
suffer from ever-increasing exploitation and contamination.
This has led to the implementation of various national and
international measures, laws, and regulations. These range
from requirements that operators continuously monitor par-
ticulate emissions from metallurgical industries such as
106
foundries during metal production, to the partial or total
banning of hazardous materials and ozone-depleting chem-
icals. Until recently, large-scale monitoring was possible
only by local, on-site detectors such as those situation close
to chimney stacks. Aircraft are also used to routinely check
for oil pollution in coastal areas, air pollution above indus-
trial sites, or monitor for any increase in the toxic chemicals
that seep from landll sites into waterways and the soil. With
the advent of Earth observation satellites, extensive and
repetitive surveys have become possible and in some
instances these can be more effective and cheaper than
routine airborne surveys. Radar sensors overcome obscuring
clouds and haze by using the microwave spectral bandwidth
for collecting surface information independent of weather
and viewing conditions. Often optical satellite imagery is
sufcient to show the required information (see, as exam-
ples, the images presented in Fig. 3.32).
The continuity of satellite observations will be ensured by
the second generation of remote sensing satellites being
designed, fabricated, and launched. These include the
European ERS-2, Envisat and Sentinal spacecraft which
have demonstrated highly sensitive detectors for global
ozone monitoring, the US Upper Atmospheric Research
Satellite, Landsat, and NOAA, as well as a multitude of
instruments on-board the International Space station.
The constant developments in materials and processes
ensure that all forms of spacecraft become lighter and more
accurate. Advances in their various technologies are main-
tained owing, to a large extent, to the existence of worldwide
Fig. 3.32 a. SAR multitemporal image made by ERS of an
85 85 km area around Rome. The technique requires several satellite
passes over the same area and reveals pollution and oil-spills in the sea,
wave patterns, some topographical relief, and changes in soil moisture
(courtesy of ESA.) b. A satellite optical photograph of Europe lit up at
night, evidence of the wasteful use of energy resourcesto some
3 The Integration of Materials into Product Assurance Schemes
competition between international companies. At the present
time there is an attempt to curb massive exploitation of our
environment by the worlds large manufacturing industries.
It is therefore very important that the space industry should
also monitor its own operations and ensure that it is not
contributing unnecessarily to the range of lethal fumes and
toxic wastes discharged into the surroundings, and that it too
utilizes the Earths material resources in the most sustainable
way. Most engineers also consider themselves to be envi-
ronmentalistswe care about the environments of both
Earth and space, and the need to protect them for the future
of our children. It is up to engineers to provide the tech-
nology for the solutions to environmental problems. The
German word for engineer is ingenieur, which comes from
the Latin word meaning ingenious. Landll or incineration
disposal of scrap, byproducts, ground station gantries (see
Fig. 3.33) and unused spacecraft hardware is optimal.
Environmental regulations may eventually lead to a ban on
both so recycling, which has the potential to offer revenues
long beyond the space projects end of life, is certainly the
best option. The recycling of electronic items will be further
discussed below, and an assessment of the metals, organic
compounds and oxides available to be recycled or extracted
from waste spacecraft electronic boxes will be given in
Sect. 3.2.2.
However, the arguments and balances which determine
the selection of materials and processes when environmental
factors are introduced to the equations are complex. For
instance, it is widely known that weight-saving is a critical
astronomers light pollution should be made an offence. The scourge
of light pollution may be one of the most detrimental spin-offs of our
technological age; excluding younger urban generations from the
experience of a truly dark sky and the stars. (Picture courtesy
International Dark-Sky Association, Tucson, Arizona)
3.5 Greener Spacecraft
Fig. 3.33 ESAs launch complex ELA-2 as seen in August 2011. It is
extensively damaged by wide-spread corrosion of its painted
zinc-plated-steel support structure, its roof structure and internal
platforms. This facility launched 119 space missions of the Ariane 2,
3 and 4 series over the period from 1986 to 2003. It is split into two
zones, the launcher pad itself (to the left of picture) and the mobile
gantry, seen here 950 m to the right, which provided a protective
factor in the selection of satellite materials, particularly when
the cost of taking one kilogram into a low Earth orbit is very
approximately 2000. A kilogram saved by selecting
beryllium in place of an aluminium alloy for a structural part
might seem attractive. However, the extra cost of imple-
menting the safety requirements which control the toxic dust
and scrap generated during machining of beryllium and
disposing of this waste in an environmentally friendly way
may mitigate against this advanced material. This is despite
the fact that when weight-saving factors are translated into
overall launch vehicle weight, there should have been a
saving of fuel or an increase in the number of payloads,
which would have in turn meant less pollution from the
propellants being red during launch.
High-quality workmanship and a good standard of living
both at home and at work have been achieved in the modern
world, and to some extent this will be maintained by the
implementation of internationally agreed specications, such
as ISO 9000. This document has brought into focus the need
for effective working standards and defect-free products. The
document also aids in the quest for greener products. The
107
building for the vehicle and integration of its satellite payload. The
mobile gantry was rolled back to its parked location prior to the
launchers lift-off. One month after this picture was taken, ELA-2 was
brought down by a controlled explosion and dismantled. The metal
structure and electronic contents were stripped down, compressed and
returned by ship to Europe for recycling (photograph, Dunn)
manufacturing industry is forever changing because of reg-
ulations and rules, many emanating from the EU. Most often
observed are the restriction, or compliance, of materials
resulting from REACH (pertaining to restriction and
authorization of all dangerous chemicals) and RoHS (ban-
ning dangerous substances that are used in electronic prod-
ucts). The interaction between REACH and RoHS is
complex but complementary, and both can be related to
WEEE (waste electrical and electronic equipment). These
regulations impact the space industry, but are beyond the
scope of this book. Already the author is aware of several
large organizations and companies in the space industry who
are making studies to nd alternative substances for an
arbitrary table of which might be called, The most haz-
ardous materials to be presently used in the manufacture and
launch of spacecraft. This table contains the following
materials:
Ozone-depleting chemicals (ODCs) such as the
chlorofluorocarbons (CFCs) used most extensively as
solvents in the cleaning processes for metals and
108
electronic assemblies. The most common such chemical
which was used in the space industry and has now been
almost totally replaced is trichlorotrifluoroethane, also
known as CFC-113. This substance was also used in the
simulation of more hazardous liquids, the fuel hydrazine
and its oxidizer nitrogen tetroxide, during the ground
testing of propellant tanks and associated reaction-control
equipment. See also Sect. 4.21.1.
Chromate primers which passivate metallic surfaces
including aluminium and magnesium alloys and silver by
simply immersing the parts in a chromate solution to
form a coating of chromate on the base metal. These
conversion coatings are widely specied as very
cost-effective means of enhancing the corrosion protec-
tion of spacecraft surfaces (see Sect. 4.5.3) and increas-
ing the adhesive strength of paints or adhesive joints to
structural surfaces (see Sect. 5.10). However, inhalation
of the chromate ion is recognized as potentially car-
cinogenic, leading to serious health and safety concerns
about its use in the workplace. Unless properly handled,
chromates can cause ulceration and environmental con-
tamination. As yet, despite extensive testing of new
products (e.g. Alodine 5700, Nabutan STI/310 and Iridite
NCP), no effective, low-toxicity alternatives appear to be
commercially available (Pereira 2008). The most com-
mon chemical conversion coating (CCC) currently used
for the improvement of corrosion resistance on alu-
minium alloys is Alodine 1200, manufactured by Henkel.
This is a chromium (V1)-based CCC. Due to the envi-
ronmental impact of hexavalent chromium this product is
likely to be forbidden for space-use in future years. An
important property of Alodine 1200 is that has a low
electrical resistance and can comply with spacecraft
requirements for electrical bonding (to ensure the craft
has proper electrical grounding, an equi-potential
between sub-systems and avoidance of corona dis-
charge under vacuum). A resistance of less than 0.0025
ohms is possible for 3xxx, 4xxx, 5xxx and 6xxx-series
alloys is achievable with Alodine 1200 coatings. The
high-strength alloys, 2xxx and 7xxx with the same
coating, have electrical resistances below 1.0 and often
below 0.1 ohms which is suitable for removable contacts
(for instance, between an electronic housing and a
structure when both are Alodined).
The ground equipment associated with launch sites and
tracking stations should avoid the use of cadmium as a
corrosion protection nish on mild and low-alloy steels
owing to the highly toxic nature of cadmium and its
by-products. Even the pickling of cadmium-plated parts
can cause signicant quantities of cadmium to enter the
effluent system. A number of platers have already been
penalized for discharging cadmium. Similarly,
silver-cadmium brazing alloys and the traditional
3 The Integration of Materials into Product Assurance Schemes
cadmium-containing silver solders can pose a serious
health hazard to those making these joining operations,
even when fume extraction is provided. Alternatives are
available, such as zinc plating and cadmium-free joining
alloys. Note, both cadmium and zinc are prohibited from
items exposed to the space environment owing to prob-
lems associated with sublimation and whisker growth, as
described elsewhere in this book.
Lead-bearing solders have been banned from use in
domestic plumbing to prevent ingestion of lead from
drinking water. Lead and lead-containing substances are
very toxic and pending legislation and U.S. Environ-
mental Agency regulations threaten to ban the use of lead
in the solders used widely in the electronics and space
industry. The main reason for this is the belief by some
politicians that by discarding domestic electrical goods
into landll sites there will be an ecological problem in
years to come due to the seepage of lead-bearing com-
pounds into our water supplies. The long historical use of
tinlead solder, reaching back to artisans in Mesopotamia
in 4000 BC, means that there is an extensive database of
solder properties. At the present time both tinlead and
indiumlead solders are the preferred solders for space-
craft electronic assembly. Alternatives, having alloy
constituents that are essentially nontoxic, with a good
oxidation resistance, and an ability to wet and form
metallurgical bonds to the wide range of metals used in
the electronics industry, are still being evaluated. Early
works found none of the lead-free alloys to possess the
overall good properties and low cost of eutectic tin-lead
alloy (Akinade et al. 1995; Jacobson and Humpston
1995). Little progress has been made and it is certain that
the reliability of lead-free products will depend on the
component package type and circuit board material and
design. Certainly, due to the higher melting temperature
and higher strength of lead-free solder alloys versus
eutectic tin-lead, the reliability of reworked and repaired
electronic circuits will be severely reduced (Ma 2012; Ma
2013; Konoza 2012). Compared to tin-lead, the lead-free
processing windows are reduced when assembly and
rework are considered, there is less margin for error. Pcb
assemblies that would have been acceptable with Sn-Pb
process parameters are at risk of delaminating and failing
during operation when lead-free solders are employed
(Reid 2007). For these, and many other reasons, ESA has
advocated the application of Lead Free Control Plans for
the space industry (Dunn 2012). One might question the
impact that the electronics industry has had on the
worldwide usage of lead? In the USA it is estimated that
only 0.6 % of all lead consumption is in the electronics
industry, whereas 80 % is used by the car battery industry
(Craig 1994). These values are also probably reflected in
European industrial consumption. 100,000 km of lead
3.5 Greener Spacecraft
sheet was used in the UK during 2010, mostly, for house
roofs (Anon 2011). The real conundrum is that the same
persons who are attempting to ban lead in electronics also
want to reduce the use of fossil fuels and cut down on the
number of petrol-driven vehicles by replacing them with
battery-driven electric cars (based on lead acid
batteries!).
Lead in brass and bearing materials, can be thought of
as an essential ingredient. It is normally added as
between 2.5 and 3.5 %wt. The lead is present as discrete
islands within the copper-zinc matrix, usually at grain
boundaries. It is trapped within the brass dual grain
microstructure, being insoluble in both alpha and beta
grains. Lead ensures less tool wear during machining as
it acts as a solid lubricant, it provides for a more uniform
surface nish compared to lead-free brass, and piece
parts can be machined to better dimensional tolerances
with less cutting force, and chip size (Pantazopoulos
2008).
Volatile organic compounds (VOCs), such as the sol-
vents that are mixed with paints, are harmful to operators
and must be extracted from the work-area. They are also
environmentally hazardous, so cannot be simply emitted
into the outdoor atmosphere. Regulations in European
and other countries mean that all VOC-emitting plants
must be registered and monitored for environmental
pollution by a number of authorities. These VOC sol-
vents have little productive use once the paint has been
deposited on the part. The space industry utilizes many
different paint products (for both corrosion protection and
thermal control). The answers are to use abatement
equipment together with higher solid-content paints, with
up to 75 % solids.
Green propellants, how they can be identied and uti-
lized as launcher and thruster motor fuel, is one of the
greatest challenges for propulsion engineers. We are
presently almost totally dependent on the family of
hydrazine-based liquid fuels, and ammonium
perchlorate-based oxidizer (AP) for solid fuels, both of
which are highly toxic and carcinogenetic. Legislation
presently under discussion in the EU, and a common
desire in the space business, points to the need to steer
away from these substances. If hydrazine is replaced,
then its oxidizing agent dinitrogen tetroxide will proba-
bly also be abandoned. Apart from the toxicity of this
liquid propellant combination, both agents have good,
well proven, long-term storage capabilities on ground
and in space; in operation they have a very high specic
impulse; and once ignited combustion cannot fail.
Alternatively, hydrogen peroxide (H2O2) is a green oxi-
dizer (Meaker 2006) and has been used with kerosene for
the main propulsion of the UK designed and manufac-
tured Black Arrow launcher seen in Fig. 3.34. Presently
109
prototype thrusters are being tested in Italy using
monopropellant H2O2 and bi-propellant H2O2 with
ethane. Another material, that is said to have the potential
to replace both AP and hydrazine is ammonium dini-
tramide (ADN), NH4N(NO2)2. Swedish tests on a solid
AND propellant and an AND-based liquid monopropel-
lant have been successfully tested (Larsson 2011).
Another, so-called green propellant is Hydroxyl
Ammonium Nitrate fuel/oxidizer mix, also is known as
AF-M315E. This was developed at the Air Force
Research Laboratory at Edwards Air Force Base in
California and is proposed to replace hydrazine in
thruster motor applications (Hawkins 2010).
Brominated flame retardants, (BRFs) and less fre-
quently chlorinated flame retardants are added to ther-
moplastics, electronic packaging and printed circuit
board materials to achieve a reduced or zero possibility
of re. These substances do have a damaging effect on
health as they release dioxins when disposed of by
incineration. They may be prohibited, but in removing
BRFs from electronics, there can be a signicant negative
impact on human health as the risk of re may cause
injury or death.
Particulate emissions, particularly during the machining
and grinding of toxic spacecraft materials, must be con-
tinuously monitored to satisfy environmental legislation.
The monitors for particle emission may utilize the prin-
ciple of particle impingement (triboelectricity) whereby
the collision of a particle with a metallic surface (here the
sensors probe) causes a charge transfer to occur between
the sensor and the probe. This charge generates a signal
which is recorded in the sensor electronics. A data logger
can store and display the dust particle numbers collected
daily or weekly, and this information will satisfy the
Environmental Protection Act in the UK. It is especially
necessary to monitor the dust emitted during the
machining and shot blasting of materials such as beryl-
lium and beryllia, which are toxic in the nely divided
form; such particles are shown in Figs. 5.15 and 5.41.
Advanced materials are slowly becoming accepted into
the commercial manufacturing industry, but there are
some unlikely factors which might limit this trend. One is
the desire that it should be possible to recycle all of these
materials. Although this argument must be taken into
account, the most important reason for material selection
is its performance during service life. In the case of the
expensive beryllium alloys referred to above, it is clear
that scrap material or unused spacecraft parts can be
recycled. But this involves the careful handling, storage,
and traceability of scrap which can then be sold back to
the metal producer. More problems have been experi-
enced with the aluminiumlithium alloys earmarked for
aircraft structures where there can be extreme wastage.
110
Fig. 3.34 Photograph of the
UKs black arrow satellite
launchers successful launch from
Woomera, Australia in 1971. The
rst two stages relied on
hydrogen peroxide and kerosene
(courtesy of the Science Museum,
London)
Between 80 and 93 % is lost during the milling of parts
for the Tornado project. These aluminiumlithium alloys
are 10 % lighter and 10 % stiffer than current aluminium
alloys; however, they are also between 3 and 4 times the
cost of the alloys they are designed to replace, so it is
clear that their recyclability is an important factor during
material selection. Various approaches are being carried
out by the manufacturers of these alloys, but it is now
clear that great care must be taken to avoid the mixing of
scrap AlLi with other aluminium alloys, as this would
mean that the secondary metal chain would become
3 The Integration of Materials into Product Assurance Schemes
contaminated with lithium, and even small amounts
(30 ppm according to Cleave and Morton 1989) will give
unacceptable levels of porosity in conventional alu-
minium alloy castings.
The electrodeposition of metals such as gold, silver, and
copper is made from cyanide baths for purposes such as
electrical contacts, conductors, and solar reflectors. This
calls for extreme care when operators are making cyanide
additions to the tanks. Management must be sure that
spent plating solutions and rinse water solutions are fed
into an approved waste treatment process. Several
3.5 Greener Spacecraft
companies have decided to attempt the change to
noncyanide-plating baths. Only copper plating was suc-
cessfully switched to a noncyanide, alkaline electrolyte
for both rack and barrel installations. Some obstacles and
difculties may be encountered which are overcome by
changing to a different pre-plate chemistry, and more
frequent in-house bath analyses are needed until a com-
fort level is established for the consistency and reliability
of the product. It has been found that the noncyanide
copper gives better plating coverage in
low-current-density areas than the cyanide copper, but
overall it has a 1015 % slower deposition rate.
As will be seen in Sect. 3.6, it may be more cost effective
to retain the use of some banned substances and utilize
mechanical processing in the recycling of waste elec-
tronic equipment.
3.6 The Potential for Recycling Electronic
Waste
3.6.1 General
The prospect of recycling electronic waste materials became
intriguing when the Institute of Materials, Minerals and
Mining declared in 2012 that the urban gold mine will be up
to 50 times more lucrative in the coming years because of
the 320 tons of gold, 7500 t of silver and immeasurably more
tons of tin going annually into products such as computers,
televisions and phones. Until now only 15 % of these
materials are recycled and recovered, but this will become a
lucrative business for urban minerswww.gesi.org.
In particular, gold and the rare earth metals, often the
subject of political bartering, should be recycled because
although their concentration is low, it is still higher than in
mineral ores. Electronic equipment can be crushed, shred-
ded, magnetically separated and by means of hydro- and
extraction metallurgy, these valuable metal elements can be
recovered; far better than diverting potentially toxic mate-
rials to land-ll sites. Mechanical processing of pcb assem-
blies allows for the recovery of 80 % of the constituent
metals, particularly copper which represents 75 % of the
metallic fraction (Veit 2002). Pyrolysis by conventional and
microwave heating can be used to separate the majority of
organic compounds from the inorganic materials, and the
resultant gasses can be used to fuel the further pyrolysis of
printed circuit board assemblies (Hall 2007; Sun 2012). The
isolated metals can then be recycled following traditional
routes for scrap metals.
111
3.6.2 Elemental Distribution for Spacecraft
Electronic Box
Background
Some years ago European space scientists proposed and
designed a spacecraft with the scientic objective of per-
forming a complete global survey of the Moons physical
and chemical properties. This complex programme, which
was later abandoned, relied on a low attitude lunar orbiter in
combination with a relay satellite. The project was intended
to supplement the wealth of data already obtained from the
six landing sites provided by the US NASA Apollo missions
(see also Sect. 8.1.2). Spectrometers and multi-spectral
stereo-imaging would produce a map of the elemental
abundances over the Moons surface and also map its min-
eralogical composition in conjunction with images for
photo-geological investigations.
The major rock-forming elements on the lunar surface are
Mg, A1, Si, Ca, Ti, and Fe. It was proposed that a chemical
map would be of great interestthis should indicate the
concentrations of these elements to an accuracy of at least
20 %, across the entire lunar crust and so establish the lateral
and vertical heterogeneity of the Moons surface.
Chemical analyses would be made using two instruments:
X-ray spectrometer
The experiment observes x-rays emitted by the lunar surface
as a result of solar x-ray excitation.
Gamma-ray spectrometer
Chemical mapping to a depth of about 30 cm from the
Moons surface might be achieved by observing the
gamma-rays produced from the excitation of nuclei by cos-
mic rays (mainly secondary neutrons). In polar regions, H2O
was thought detectable down to a concentration of 0.4 %.
It had been proposed that both spectrometers should be
mounted on booms in order to minimize the effect of
spacecraft-produced background emissions that would
reduce the accuracy of the lunar surface analyses.
Under ideal circumstances the spacecrafts own back-
ground emissions could be characterized by operating the
two spectrometers during their flight towards the Moon.
A knowledge of this elemental distribution could be con-
sidered as a perturbation factor which, when computer
modelled, would be subtracted from the crafts own analysis
of the lunar surface. Studies of the Moons gamma radiation
have been made by Russian Lunas spacecraft, the US
Apollos 15 and 16 as well as numerous other craftthe
gamma-ray spectrometers and X-ray fluorescence spec-
trometers together with detailed presentations of those
112 3 The Integration of Materials into Product Assurance Schemes

Table 3.3 Elemental distribution of parts and materials making up a typical spacecraft electronics box (percent by weight)
Part White paint Aluminium Potting Printed Soldered joint Connector Steel Label
on box box foam circuit and plated screws
board through hole
Material DC 92-007 AA 2014 Eccofoam G10 (tin Sn63 and Cannon (FR-7507) AISI AA3003
type FPH lead on copper barrel DDM-SOP-NMC-PR 316
copper 103A
tracks)
Weight 0.02 gcm2 2.8 gcm3 0.5 gcm3 2.0 gcm3 0.3329 g 20.705 g 1.0 g 1.0 g
or (100 each each
density thick)
of each
H 4.60 6.49 2.46 2.50
B 1.00 0.20
C 18.30 59.64 20.10 29.70 0.08
N 9.05 0.20
O 28.60 23.37 35.40 4.20
F 0.05
Na 0.05 0.03
Mg 0.50 0.30 0.12
Al 12.20 93.60 9.00 0.03 1.50 98.80
Si 6.40 0.70 18.70 0.10 3.50 0.50
P 0.04
S 0.02
Cl
K 0.01
Ca 2.50 0.05 1.20
Ti 20.10 0.20
Cr 0.02 0.03 17.00
Mn 0.80 1.00 1.20
Fe 0.04 0.02 0.02 68.40
Co
Ni 1.40 0.15 12.00
Cu 0.06 4.40 5.50 75.00 52.00
Zn 2.10 17.00
Ge
As
Sr
Zr
Mo 2.50
Rh
Pd
Ag 0.01
Sn 3.60 12.00
Sb 0.70
Ba 0.02
Nd
Ta
(continued)
3.6 The Potential for Recycling Electronic Waste 113

Table 3.3 (continued)

Part White paint Aluminium Potting Printed Soldered joint Connector Steel Label
on box box foam circuit and plated screws
board through hole
Material DC 92-007 AA 2014 Eccofoam G10 (tin Sn63 and Cannon (FR-7507) AISI AA3003
type FPH lead on copper barrel DDM-SOP-NMC-PR 316
copper 103A
tracks)
Weight 0.02 gcm2 2.8 gcm3 0.5 gcm3 2.0 gcm3 0.3329 g 20.705 g 1.0 g 1.0 g
or (100 each each
density thick)
of each
Au 0.50
Pb 3.00 12.00 0.10
Total 92.42 100 98.55 101.94 100.64 113.57 101.54 100

Table 3.4 Presence of oxides (approximate weight percent) within electronic box materials and components
Si02 AI203 Ti02 MgO BaO CaO
White paint on box 2 23 33
Printed circuit board 40 18 <1
Large resistor 38 9 25 3
Small resistor 21 5 1
Glass-bodied diode 52 1
Small diode 5 <1
Large rectangular capacitor 70 3
Large cylindrical capacitor 43 16 4 17
Small capacitor 50 6 2 2 21
Transistor >1 >1
Flat-package (kovar body) 9 1 >1
Flat-package (ceramic body) 10 42 3

results are detailed in the books written by Surkov (1997)
and Schrunk (1999).
Special attention would have been paid to the materials
selected for the construction of the spacecraft in order to
reduce this chemical background. For instance, the struc-
ture could be made entirely from aluminium, beryllium or
titanium. It was recognized that a major part of the space-
crafts mass comprises of electronic boxes, each loaded with
high-density printed circuit board assemblies.
The space scientists requested that a computer model be
made pertaining to the global chemical (elemental) content
of a typical spacecraft electronic box.
Method and Results
A discarded flight-spare unit from a successful space project
was used for this work (Dunn and Nicholas 1998). Initially
the box was opened, the printed circuit board assemblies
were de-mounted from their supports, potting foam was
removed and the electronic components were de-soldered.
Some chemical analysis data could be established from
the certicates of conformance issued by the material sup-
pliers. Here full analysis records were available for the actual
aluminium billet used to make the electronic housing. Sim-
ilarly, analyses were available for purchased batches of
aluminium foil and stainless steel screws.
The remaining items, such as the external paint, potting
foam and the various electronic components were individ-
ually crushed, pulverized and their material constituents
measured by various methods of chemical analysis. Special
efforts were made with the soldered joints. Here the
copper-plated through holes were mechanically removed
from several multi-layer circuits and submitted for analysis.
Full chemical analyses have been attempted. Fluorine,
chlorine, bromine, iodine, sulphur, phosphorus, etc., have
been detected by emission spectroscopy, X-ray fluorescence
spectroscopy and activation analysis. Gas chromatography
and infrared absorption spectroscopy have been used to
separate and determine the volatile organic components.
114
Non-volatile components have been identied by both ionic
and non-ionic chromatography on ion exchange columns.
The metallic elements were determined by Spark Source
Mass Spectroscopy (SSMS) of the pulverized powders.
Further analysis was made by dissolving some of the
metallic particles in mixed acids and measuring the metal ion
content of the solution.
An attempt has been made to compile all the chemical
analysis results and this is shown, based on elemental dis-
tribution, in Table 3.3. The accuracy of the tabulated results
3 The Integration of Materials into Product Assurance Schemes
is somewhat limited by the fact that different analytical
techniques were used to establish the weight percent of the
constituent materials. The identication of organic and
intermetallic compounds has not been recorded. The main
oxides discovered from analysis are listed in Table 3.4.
The recording of these results might be useful for future
scientic projects. They may also be of interest to organi-
zations involved in the recycling of electronic hardware.
 Spacecraft ManufacturingFailure Prevention
and the Application of Material Analysis
and Metallography
4.1 Sources of Failure
Potential sources of failure are frequently identied during
the manufacturing and assembly stages of spacecraft pro-
duction. They may related to one or more of the following:
unsuitable aspects of design,
incorrect material selection,
human error (e.g. machine tool operator becomes
careless),
materials imperfections (e.g. internal porosity or harmful
inclusions),
defects introduced during reworking, repair or modi-
cation of existing structures,
corrosion damage during storage.
Whether materials are custom-made for the space indus-
try or bought off-the-shelf from general commercial markets,
there will be a need for careful planning and the use of a
wide range of quality control techniques. These will monitor
whether fabrication processes have been under control and
verify that the quality of the nal product is acceptable for its
design application. Metallography is frequently applied as a
quality control tool and in a unique way can supplement
both destructive techniques, such as mechanical testing, as
well as nondestructive investigations like radiography, dye
penetrant testing, and ultrasonic testing. Each of the potential
material failure sources listed above can be evaluated by
metallographic inspection either by conventional optical
microscopy or by more sophisticated procedures such as
scanning electron microscopy, and scanning laser acoustic
microscopy.
Failures are occasionally caused by human error and
may relate to the overtesting of perfectly spaceworthy
equipment. This can be illustrated by the following anecdote
that was brought to the authors attention. Several electronic
equipment units were to be qualication tested and had been
heavily instrumented with vibration and shock accelerome-
ters. Early in the programme the vibration range for the test
Springer International Publishing Switzerland 2016
B.D. Dunn, Materials and Processes, Springer Praxis Books,
DOI 10.1007/978-3-319-23362-8_4
4
was set by one of the top dynamics engineers at 10 gRMS.
Although this was based on good judgement and previous
experience, the equipment manager sought out a transducer
expert for corroboration. The channel setting was raised
from 10 gRMS to 15 gRMS by the transducer expert (who was
conservative by nature). The qualication plan was then
given to the project management and another group of
measurement experts, who added their margins to the test. It
is not surprising that, after further stops along the organi-
zational chain, the test ended up at 50 gRMS.
The relationship between human issues and engineering
failures has been addressed by Le May and Deckker (2009)
and researching routes to failure-avoidance found that
human causes were responsible for 80 % of accidents,
so-called unintentional mistakes. These issues can be caused
by: organizational change, where staff may change their role
in a company without having the necessary re-training;
fatigue and loss of alertness; failure to maintain equipment
and perform calibration tasks; and, failure to write and/or
comply with manufacturing procedures or process identi-
cation documents.
4.2 Drawings and Workmanship
4.2.1 Design and Manufacturing Drawings
The design of spacecraft is based on the practices developed
initially for the aircraft industry. The main goal of the design
engineers is to minimize the structural mass without jeop-
ardizing the integrity of the spacecraft structure during the
phases of test, launch, zero-gravity, and possibly re-entry.
Almost equally important are the designs of the many sub-
systems for communication, power, attitude control, and the
like, which must also survive extensive testing and launch
loads. The design of the individual piece-parts that are used
to construct the nal subsystems may undergo several
changes from the initial concept of the spacecraft design.
Depending on the model philosophy, these changes might be
115
116
needed following the testing of a thermal model, engineering
model, or qualication model. As discussed in Sect. 2.2, the
freezing of the design should occur at the time of the critical
design review, but some changes do occasionally happen
after this time and will need to be qualied prior to the
launch readiness review. An overview of various spacecraft
structure designs is given by Turner (1995), and is outside
the scope of this book, but, as will be illustrated in the
following chapters, initial designs are sometimes totally
inadequate owing to the complex interactions between
materials, their surface nishes, and their exposure to sim-
ulated space environments.
An engineering drawing will be assigned to each single
part. It is important that these drawings are closely controlled
(under conguration control) and that they have a standard-
ized format. The large numbers of spacecraft manufacturing
contractors, combined with the potential for numerous
changes to dimensional detail, material nish, etc., necessi-
tate that drawings originating from one design ofce are also
fully understood by designers and production persons, not
only in other companies, but in other countries and on dif-
ferent continents. Drawings will be prepared to contain an
identication number, an issue date, and details that cover the
complete manufacturing picture of the product. It is probable
that each engineering drawing will be approved by the per-
sons responsible for design, selection of materials, produc-
tion, and product assurance. All materials and processes will
be later submitted for approval to the customer by means of a
Declared Materials List (DML) and a Declared Processes List
(DPL). Examples of DMLs and DPLs are given for ctitious
spacecraft systems in Appendices 10 and 11. This approval
activity occurs both at the preliminary design review and,
again, after modications, at the critical design review. These
reviews will consider factors such as suitability of materials
for structural use (mechanical properties, fracture mechanics,
resistance to stress corrosion cracking), operation under
vacuum (sublimation and outgassing), and suitability for
production (solderability of electronic materials, weldabil-
ity). The processes identied on the engineering drawings,
and later tabulated on the projects DPL, will also be scru-
tinized to see if they have been suitably veried or qualied
and whether the process documentation, which ensures
reproducibility, has also been approved.
The engineering drawings are therefore of great impor-
tance during design, production, approval reviews, and,
should failures occur, during the failure investigation.
A summary of the useful data that must be recorded on the
drawings is as follows:
(a) Drawing number and revision number (contained in a
box at the lower right-hand corner). The number will
usually be a long string of individual numbers and
4 Spacecraft ManufacturingFailure Prevention
letters that refer, sequentially, tothe part family (e.g.
electronic housing, support strut, bearing, thruster
motor, or electrical drawing);the spacecraft job
number;the assembly area location;and the draw-
ing serial number and issue number.
(b) The accurately drawn picture of the part with all its
views, dimensions, and surface nishes. This must be
fully understood by all operators both in the parent
company and those located at sub-contractors sites.
Dimensions are usually in millimetres. Tolerances,
where necessary, need to be specied for radii (changes
in section can lead to stress concentrations), surface
roughness, and interlocking parts. Abbreviations
should only be used if they are universally accepted.
(c) A list of materials needed to make the part must be
shown on the drawing.
(d) The manufacturing processes must be listed in the
sequence of production. This may commence with saw
cut, and continue through dip braze, cleaning,
nondestructive inspection, stress relieve, shot
blast, chemical conversion coat, paint-colour, and
bake. Each process identied for flight hardware
should be accompanied by a process identication
document reference number and its issue number. The
issue number is important. It must reflect the actual
process document and process parameters utilized for
the production of qualication hardware.
4.2.2 Workmanship Standards
Bitter experience has shown that things do not work if a
certain number of manufacturing precautions and controls
are not exercised. This is particularly true for automated
processes such as resistance spot welding or the making of
mechanical joints such as crimped electrical connections.
Here all process variables must be evaluated, and, once
optimized process schedules have been attained, they must
be preset and maintained during production runs. Unfortu-
nately, the quality of most metallurgical joints is difcult to
assess by visual inspection alone. In these instances it is
important to produce so called technology samples during
preproduction runs, and these will be inspected and then
subjected to destructive tests such as mechanical pull and
metallography. Duplicate samples can also be used to gen-
erate workmanship standards or visual aids to illustrate the
quality characteristics of joints by reference to various
machine settings (e.g. pressure, time, current).
Workmanship (Dunn 2008) can be dened as, the skill
with which something was made or done. It is the product
of effort or endeavour and relates very much to the skills of
the person involved. These might be a craftsman able to
4.2 Drawings and Workmanship 117

fashion jewellery from precious metals, or an operator on the
shop floor assembling spacecraft hardware. Such skills are
acquired by training, and it is the physical skills, such as the
handling of a blow-torch or soldering iron, which may result
in either excellent workmanship, or shoddy workman-
ship. See also Sect. 6.14.2 concerning training schools.
A set of photomicrographs, photographs and sketched
illustrations may be used as workmanship standards and, as
illustrated in Figs. 4.1, 4.2, 4.3 and 4.4, they can dene
process acceptance and rejection criteria. When required
during production, in-line quality control samples will be
taken at regular intervals and subjected to rapid metallo-
graphic preparation for comparison against these standards.
This will ensure that the variable settings on automated
process equipment have not altered since they were initially
set up, thereby giving added condence that the overall
quality levels of flight equipments will remain consistently
high.
Optical microscopy is also widely used to examine pol-
ished metallographic samples for inclusion content. The

Fig. 4.1 Metallographic

workmanship standards for
resistance spot welds; each
sample had a similar external
appearance. Initially, welding
parameters such as applied
current, time, and pressure are
optimized against mechanical pull
strength and metallographic
appearance. This example
illustrates an electrical
interconnection within spacecraft
nickelcadmium battery cells
which must operate for at least
eight years. Two nickel-plated
plain carbon steel ribbons are
welded to a nickel pin. The
gures illustrate: a marginally
insufcient weld penetration,
b marginally excessive weld
penetration, c optimized
parameters
118 4 Spacecraft ManufacturingFailure Prevention

Fig. 4.2 End closure weld in

nickel alloy pressurized housing.
Metallography serves as the best
means for evaluating this tungsten
inert gas (TIG) welded assembly.
Welding parameters are assessed
against workmanship standards.
Device hermeticity cannot be
guaranteed with a insufcient
weld bead, or b excessive
porosity in weld bead. The
optimum weld is shown in (c)
4.2 Drawings and Workmanship
Fig. 4.3 Workmanship standards exist for a wide range of welding
technologies and specic standards need to be consulted that relate to
the metal alloy, the design and the welding equipment. Spacecraft
weldments can be visually controlled to ECSS Q-ST-70-39 and this
standard also contains workmanship guides for accept/rejection of
surface oxidation on post-welded titanium and stainless steel parts. The
illustrations shown here dene acceptable and unacceptable weld
proles for tube welds. They depict what might be observed on the
polished and etched surfaces of microsections made either during weld
parameter set up, or during in-line quality control sampling. The lower
set of illustrations depict visual inspection criteria for acceptable and
simplest method is to survey the as-polished sample and
compare the number and form of inclusions with standard
charts. This method is employed for the assessment of
thin-walled pressurized devices which might otherwise leak
if an inclusion were to occupy the wall thickness. The
assessment of size, shape, and distribution of inclusions has
also been successful in the selection of material for space-
craft pyrotechnic devices as will be discussed in Sect. 5.4.
Finally, workmanship standards exist for all the designs
of solder joints employed for spacecraft electronics. They are
used to determine the quality of solder joints through visual
criteria, as there is a degree of correlation between surface
appearance and shape of the solder llet with the
119
unacceptable weld bead width and the degree of permitted meander.
These illustrations are reprinted from ASME BPE-2014 (pp. 132 and
135), by kind permission of The American Society of Mechanical
Engineers. All rights reserved. No further copies can be made without
written permission from ASME. a Acceptable, b misalignment
(mismatch), c O.D. concavity, d I.D. concavity (suckback), e lack of
penetation, f convexity. A Acceptable weld bead, B excessive weld
bead width variation, C excessive weld bead meander. General note
Applies only to nonproduct contact surfaces and only if weld on
product contact surface cannot be inspected
metallurgical structure that exists in the joints. It should be
noted, however, that satisfactory surface appearance does
not guarantee a sound metallurgical bond. Solder defects are
usually classied as major, minor, or cosmetic. Major
defects could reduce the operational life of the electronic
hardware and must always be repaired or reworked (see
Sect. 6.12.3). Minor defects, such as borderline solder vol-
umes (either excessive or insufcient), generally do not
cause failuresfor high-reliability electronic circuits they
would be reworked. Cosmetic defects do not affect the
performance or operating life of the product (e.g. component
tilting, small folds on the solder llet); they do given an
undesirable appearance and indicate that the production
120 4 Spacecraft ManufacturingFailure Prevention

Fig. 4.4 Metallographic

workmanship standards of
crimped electrical connections.
These photomicrographs are
selected to show: a stranded wire
volume contains unacceptably
more than 10 % voids by
cross-sectional areaundercrimp
(tool setting-2 positions under
optimum, wire not deformed),
b use of an oversized crimp barrel
for this wire gauge, c optimized
process providing a gas-tight
joint, d, e details of the damage
caused by a cracked barrel (from
visual inspection alone it is not
possible to assess if cracks in the
plated layer extend into the crimp
barrel), and f examples of ECSS
acceptable lug and splice wire
crimps
4.2 Drawings and Workmanship 121

Fig. 4.5 a General view of (a) (c)

flexible waveguide showing
position of pinhole (arrowed).
b Full cross-section of defective
region showing solder llet and
location of pinhole (arrowed).
c Etched micro-structure of
beryllium copper in region of
pinhole. Grains are generally
equiaxed, but have deformed
beneath abraded region and
contain a high density of slip
bands

(b)
122
assembly process needs to be better controlled before real
problems occur. Solder contact angles less than 90 (positive
wetting) are acceptable for solder joints; those greater than
90 (negative wetting) are unacceptable. The goal of per-
forming visual inspection against documented workmanship
standards is to maintain process control and minimize
rework.
Workmanship standards related to Area Grid Arrays
(AGAs) are reviewed in Sect. 6.15.10. AGAs are particularly
difcult to assess because the terminations, all present on the
underside of these packages, are often impossible to visually
inspect and some newer devices have more than 1000
interconnections. X-ray systems have been evaluated
(Wickham et al. 1999) and many of the resulting
X-radiographs were incorporated into the ECSS standard for
surface mount technology (ECSS Q-ST-70-38). Similar
workmanship standards have been published by the Institute
for Printed Circuits (IPC 610D and IPC 7095) and the Joint
Electron Devices Engineering Council (JEDEC-217). The
major defects addressed in these standards include:
misalignments, missing solder balls or columns, presence of
voids in the solder, non-wetted pads on the pcb, and the
presence of small unwanted solder balls created during the
soldering process that might eventually cause short circuits.
Such NDT inspection is tedious and some advancements in
the automated X-ray inspection of AGAs are available for
this work in order to reduce variability and decrease the time
needed for inspection (Said 2012).
4.3 Mechanical Damage Revealed
by Microstructure
Figure 4.5a shows a flexible waveguide from a batch which,
at nal inspection, was rejected owing to the occurrence of
small pinholes adjacent to the tube-to-flange soldered joints.
These defects were visible only at 10 magnication. The
tube material was solution-treated beryllium copper and the
pinholes were limited to the rst two tube convolutions.
At the start of the investigation, two causes for the defect
had been proposed by the waveguide manufacturer: rst, the
tubes had been supplied with drawing defects which had
been opened out by the heat dissipated during the soldering
operation, and, second, the strong soldering flux required to
produce smooth and bright solder llets had chemically
corroded the tube. Another opinion was that the soft
solution-treated alloy had been penetrated by sharply cor-
nered tooling jigs used to bend the tube into its spacecraft
conguration. After forming, the assembly had been
precipitation-hardened at 315 C for 2 h and then soldered to
the flange at 240 C.
The defective region was microsectioned and polished
(see Fig. 4.5b). It can be seen that at least two of the
4 Spacecraft ManufacturingFailure Prevention
convolutions were damaged, and that in these regions the
section thickness of the tubing was greatly reduced and
hence mechanically weakened.
Complete beryllium copper removal had occurred at the
pinhole site. The detail of one of the defective areas (etched,
Fig. 4.5c) shows the damaged internal wall of the tube to
have slightly collapsed. This indicates that a compressive
force had been applied to the tube surface. Reagents for
microetching metals and alloys are detailed in Appendix 3.
Also, the microstructure of the undamaged tube is shown
to consist of non-deformed equiaxed grains. In the damaged
region, the grains have been slightly deformed; they contain
a high density of parallel clip bands and mechanical twins
which change direction from grain to grain. Such features
are characteristic of mechanical damage that must have
occurred subsequent to the precipitation treatment because,
at 300 C, they would have been annealed out.
At this stage, it was possible to rule out the initial
judgement that penetration had been caused by forming or
corrosion. A visit to the manufacturers shop-floor revealed
that an inexperienced soldering operator had assembled this
particular batch of waveguides. Many of his llets had
contained more solder than was normal. By means of pulling
a ne cord, loaded with abrasive grit, around the llets, he
was able to remove the excess solder and produce a clean
bright llet. In doing so, this action had also produced wear
and excessive thinning of the waveguide tube. The charac-
teristic slip lines and deformation twins in the microstructure
(Fig. 4.5c) had accurately indicated that abrasion and pinhole
formation had been introduced at a stage after the nal heat
treatment of the tube.
4.4 Hydrogen Embrittlement
4.4.1 Interaction of Metal with Hydrogen
Since the early days of alloy development there has been the
challenge of hydrogen. Owing to their smaller diameter,
hydrogen atoms can be easily absorbed into metal lattices,
particularly those with large interstices like Al, Pd, Ir, Mo,
Ta, W, Zr, and Ti, and with difculty into metals with small
interstices such as Cu, Ag, Au, Ba, and Cr. The reaction of
hydrogen with metals may result in the formation of
hydrides, molecular hydrogen, and gaseous products of the
reaction between hydrogen and inclusions or alloy phases
held within the metal matrix. In all these cases hydrogen has
a harmful effect on the mechanical properties of the metal
and causes embrittlement or brittle fracture (Raymond 1988;
Robertson and Teter 1996).
Several theories have been developed to explain the
interactions between hydrogen and metals. They presume a
generation of very high internal pressures caused by the
4.4 Hydrogen Embrittlement 123

recombination of atomic interstitial hydrogen to form
hydrogen gas, a trapping of dislocations and hence a
reduction in mechanical deformation by the increased solu-
bility of hydrogen in the strained zones of the material, a
reduction in the bonding forces between metal atoms in the
lattice, and any combination of these proposals.
A continuous lm of oxide on the surface of a metal will
often prevent the dissociation of molecular hydrogen, but
numerous examples exist where spacecraft hardware, after
successful trials, will suddenly exhibit hydrogen embrittle-
ment. Examples are the methanol embrittlement of titanium,
the hydrazine embrittlement of certain stainless steels, and
the hydrogen embrittlement of battery cells.
Laboratory failure investigations have highlighted the fact
that atomic hydrogen may be introduced into spacecraft
piece-parts during their production by the forging of end
domes and barrel sections of titanium propellant tanks; the
heat treatment of steel mechanical parts and titanium bolts;
the welding of apogee boost motors made of titanium alloy;
the pickling and plating of aluminium waveguides and steel
springs; and the brazing of commercial copper component
parts in a hydrogen furnace. The following subsections
represent typical investigations associated with the hydrogen
embrittlement of spacecraft equipment.
4.4.2 Hydrogen Embrittlement of Spring Steel
Several small springs which had been installed for 2 days
within the equipment of an engineering model satellite were
heard to snap into two halves before equipment level testing.
The metallurgical analysis found the failed springs to have
been fabricated from a standard 0.4 % carbon low-alloy
spring steel which had been hardened and tempered, des-
caled, and nally plated with, rst, copper and then nickel.
The spring fracture surfaces were examined in a SEM
(Fig. 4.6a), and seen to consist of delaminated bands sepa-
rated by dimpled rupture regions. The fracture path direction

Fig. 4.6 a SEM fractograph of

prematurely failed steel spring
showing plated layers and
delaminated appearance of
matrix. b Longitudinal
microsection made through
fracture surface presented in (a).
c Extremely ne crack beneath
plated layers. d Internal
microcracks formed by hydrogen
and probably initiated adjacent to
elongated inclusions in the rolled
steel (lightly etched in 2 % Nital)
124 4 Spacecraft ManufacturingFailure Prevention

had followed along machine marks and some deep scores

which existed on the nickel surface which are detailed in the
microsection shown in Fig. 4.6b.
Based on a review of the spring processing procedure in
conjunction with examination of the highly magnied steel
microstructure (Fig. 4.6c, d) the primary failure of the springs
was attributed to hydrogen embrittlement. These springs had
been conditioned by heating to 900 C, oil-quenched, and
then immediately tempered at 250 C to produce the required
hardness and spring properties. The subsequent cleaning and
plating process stages generated hydrogen in its atomic form
at the steel surfaces. Atomic hydrogen will rapidly diffuse
through the metal lattice, and, when uniformly distributed, is
nondamaging. However, once the springs were installed on
the spacecraft, they became loaded and, under the influence
of applied stress, the hydrogen diffused to areas of stress
concentration (e.g. internal sites of inclusions and disloca-
tions). In these locations, the atomic hydrogen recombines to
form molecular hydrogen as shown in Fig. 4.6d. This
molecular hydrogen generates exceedingly high pressures
which facilitate cracking.
In most rolled-steel sheets, banded structures containing
ne, elongated inclusions are common. It is probable that the
microcracks or delaminations visible in this micro-structure
result directly from the combined interaction of stress and
the generation and accumulation of gaseous hydrogen at the
inclusion-to-matrix interfaces. Once the density of microc-
racks had impaired the load-bearing capability of the spring
sufciently, the nal failure occurred by overstress with the
regions between the microcracks tearing in a ductile manner
as shown by the dimpled zones on the fractograph.
Hydrogen embrittlement can be eliminated by a simple
baking procedure which rattles out any absorbed atomic
hydrogen before it can recombine as gaseous hydrogen. A rec-
ommended practice for hydrogen bake-out was successfully
employed in the process of fabricating replacement springs.
This consisted of heating the parts for 3 h at 190 C within an
hour of the plating operation and taking all necessary precau-
tions to avoid flexing the articles before they were baked.

4.4.3 Blistering of Plated Aluminium Alloy

Special rotors for the switching of RF waves between dif-

ferent waveguides were installed within precision-machined
cylindrical cavities that provided for a 2 m air gap during
switch rotation. The rotors were machined from an AlMg
Si alloy and electroplated, for special electrical properties,
with 8 m of nickel followed by a 6 m silver nish. Initial
checks ensured that rotors were matched with different cav-
Fig. 4.7 Examination of waveguide switch (courtesy of the GEC Hirst
Research Centre). a General view of waveguide switch. Blisters
normally occurred at location (1), the edge of the waveguide entrance,
or at (2), the flange region. b Section through blister at location (1)
(50). c As (b) but after etching in sodium hydroxide (500)
4.4 Hydrogen Embrittlement
ities, so enabling unrestricted rotation and low signal loss.
The rotors were then withdrawn from their cavities, so that
small permanent magnets could be adhesively bonded to their
bases. The heat cycle used to cure the epoxy adhesive was
seen to produce blisters in the plated layers. Hydrogen atoms
absorbed during plating were thought to have recombined to
form molecular hydrogen at the plating-to-substrate interface.
As shown in Fig. 4.7 the disbonded plating is raised locally to
form a surface blister. The blisters prevented reinsertion of
each rotor into its matched cavity as it had blocked the 2 m
air gap. Future batches of rotors were hydrogen relieved by
careful baking treatments involving slow heating cycles to
200 C for 4 h between each plating operation. This redis-
tributed or released any absorbed hydrogen to give a
blister-free surface.
4.4.4 Examination for Titanium Hydride
Precipitates
A manned spacecraft payload support structure has been
constructed of Ti6A14V alloy struts. Fitting and tubular
parts are machined from wrought alloy bar, whereas conical
parts are made as alloy castings. These parts have been
tungsten-inert-gas (TIG)-welded. Unalloyed titanium-ller
metal rod was inadvertently used in making each weld.
These 0.045 in. diameter (1.14 mm) rods had been supplied
with a certicate of analysis purporting the material to be Ti
6A14V. Spectrographic analyses made after samples and
flight equipment had been built showed that the user ller
rods were in fact pure titanium. Interstitial elements (i.e.
hydrogen, nitrogen, and oxygen contents) were not deter-
mined. Following these welding activities all remaining
noncompliant welding rods were removed from store and are
no longer available.
The use of unalloyed titanium ller metal in the welding
of titanium Ti6A14V and other titanium alloys is common
practice in the aerospace industry. However, the materials
and process must be extremely well controlled in order to
prevent embrittlement of the nal weldment from interstitial
elements (i.e. hydrogen, oxygen, nitrogen). The solid solu-
bility of hydrogen in titanium alloys, with the exception of
beta-alloys, is very low at room temperature and this results
in the precipitation of a hydride phase (Robertson and Teter
1996). The hydrides are brittle and have a detrimental effect
on the mechanical properties. As a consequence of titanium
embrittlement problems encountered by NASA [most nota-
bly the failure of a helium pressure vessel which led to the
destruction of a Saturn rocket during a test sequence (Wil-
liams et al. 1979)], the welding of commercially pure tita-
nium to alloyed titanium is now prohibited.
125
A comprehensive programme of tensile testing, bend, and
sustained load tests were made on welded material from the
flight batch. Chemical analyses were made together with
extensive metallographic inspections for the presence of
titanium hydrides in order to acquire an acceptable con-
dence level in the flight unit.
The investigation results showed that the weld tensile
strengths lay between the values expected for pure titanium
and Ti6A14V, and the welds showed a good ductility, far
greater than that experienced in the Saturn rocket tank fail-
ure. The sustained load test in this programme (at
220 N/mm2 for 14 days) was designed to induce delayed
degradation of the weld sample due to additional migration
of hydrogen and secondary precipitation of hydrides, as was
observed in the aforementioned helium tank failure. No
delayed embrittlement was encountered, and no titanium
hydrides were associated with the fracture path. Also, the
chemical analysis showed that interstitial elements were
lower than required by the raw material specications.
The entire spectrum of possible microstructures, indicated
by the phase diagram inserts (see Fig. 5.18), has been
observed on the polished and etched specimens. These are
shown and detailed in Fig. 4.8ad. The captions describe the
detailed photomicrographs. Titanium hydrides were
observed in small isolated clusters set in the weld bead face
run, but not with the weld root run. No hydrides were pre-
sent in the HAZ or parent Ti6A14V material. No hydrides
were associated with any of the fracture paths seen by
metallography of the mechanical test specimen.
It should be noted that titanium and its alloys are more
difcult to prepare by metallography than steels. Of partic-
ular importance when investigating for hydrides, is that the
mounting material is a low-viscosity epoxy resin that cures
slowly without much exothermic reaction, so that the tita-
nium sample is not heated by more than 10 C. Some
acrylic, fast curing resins can reach very high temperatures
which, like hot-mounted specimens in a press, can reach up
to 150 C and these temperatures can modify the morphol-
ogy of any precipitated TiH within sections microstructure.
A good procedure for the specimen preparation of titanium
and its alloys is given by Vander Voort (2008).
The macrographs of each weld show the root weld to
have become particularly well alloyed with the parent metal
forging. The presence of beta-stabilizing vanadium in this
portion of the weld will absorb interstitial elements to a far
greater extent than pure alpha titanium. The lack of titanium
hydrides within the rst properly alloyed weld deposits is
explained by the presence of vanadium. As the strength of
the tested weld sample was at least six-fold better than the
expected maximum service stress, a decision was made to fly
the payload support structure.
126 4 Spacecraft ManufacturingFailure Prevention

Fig. 4.8 a Sample showing

parent metals, heat-affected zone,
and weld pool. Positions of
detailed photomicrographs are
shown (10). b Face weld run
showing isolated very small
titanium hydride needles
(arrowed). Microstructure shows
serrated alpha (primary) in matrix
of secondary alpha. c Mid-weld,
mid-thickness. Indication of
hydride needles. Microstructure
shows serrated alpha (primary) in
matrix of acicular alpha. d Root
run. No hydrides within this
martensitic alpha matrix

The particular case-history from the European Spacelab
project resulted from a breakdown in raw materials control
by the relevant Materials Quality Engineer. The welding rods
were incorrectly identied by the vendor as having the
composition Ti6A14V, both on the packaging and typed takes also dogged the USAs Space Shuttle projectthey
on the delivered certicate of conformance. The user com-
pany accepted the vendors paperwork and did not bother to
forward a sample of welding rod to the Materials Acceptance
laboratory for chemical spectrographic analysis. Had this
been done, the welding rods would have been seen to have
the wrong composition (unalloyed titanium) and they would
have been rejected and returned to the vendor. Instead, the
rods were used to weld flight equipment. Fortunately, this
equipment was found to be metallurgically acceptable but the
material review board exercise which led to the decision to
use as is was extremely costly. Alternatively, any decision
to reject the nished hardware would have led to even greater
costs and some delay to the launch date. Welding wire mis-
caused failure of the main propulsion engine and damage to
its test stand when improper welding wire was used to weld
together the steer horn hydrogen lineinstead of Inconel
718, a weaker weld wire, Inconel 600, was used, and here the
mistake was also blamed on a supplier.
As concluding remarks, titanium has a reputation for
being a difcult material to weld. This is largely due to
unfamiliarity with the material and the necessity to take
anti-contamination precautions. The surfaces to be joined
4.4 Hydrogen Embrittlement
need thorough cleaning, even ngerprint greases can lead
to weld porosity. The shielding gases (usually argon or
helium, or a 75 %Ar25 %He mixture of the two) must
have a very low dew point (below 50 C) in order to
prevent rstly the pick-up of hydrogen by the liquid weld
metal, and later the formation of titanium hydrides. Rubber
hose for transportation of the shielding gases is to be
avoided, as this can release moisturevinyl plastic is
preferred. Also, the surfaces of titanium in the heat-affected
weld zone can become contaminated by interstitial gases
(O, N, H) when held in the region of 500 C. The thin
oxides that form on the metallic surfaces during cooling
generate interface colours ranging from light straw to dark
blue and grey. A rule of thumb classication for the
acceptability of titanium welds is given by Ellis and Gittos
(1995) for the colours of the weld zone. Workmanship
standards that compile acceptable and reject surface
colourations are now illustrated in the latest, 2015, ECSS
welding standard, Q-70-39. The following is a guide:
Fig. 4.9 The brazeability of copper is generally excellent, but
difculties may occur during furnace brazing in hydrogen atmospheres
that have a reducing potential. Oxygen-containing or tough-pitch
phosphorous deoxidized coppers are subject to embrittlement when
heated above 450 C. a, b show the effect of heating tough pitch copper
to 850 C in a hydrogen atmosphere. Grain boundary embrittlement and
internal cavitation are caused by the reaction of copper oxide and the
127
silvercorrect shielding, satisfactory
light straw and dark strawslight contamination, but
acceptable
dark blueheavier contamination, may be acceptable
depending on service
light blueheavy contamination, unlikely to be acceptable
grey-blue, grey and white (loose deposit)very heavy
contamination, unacceptable.
4.4.5 Embrittlement of Copper
Tough-pitch copper can, if insufciently deoxidized, contain
up to 1 % cuprous oxide distributed throughout the copper
matrix as small discrete included particles. Such material
was machined to form electronic component piece-parts,
including items for use within travelling wave tubes that
were subsequently furnace heat-treated and brazed with
formation of high-pressure steam. A sample of oxygen-free high
conductivity copper was simultaneously heated in the same atmosphere
to produce the fully annealed grain structure shown in (c, d). Note It is
difcult to obtain a scratch-free micrograph with the tough-pitch copper
as the porosity traps abrasive with mechanical polishing and chemical
polishing pits the pores and enlarges them. Courtesy of Mr T. Pinder
128
silvercopper eutectic ller metal. Hydrogen-reducing
atmosphere furnaces were employed. At temperatures
above approximately 450 C absorbed hydrogen reacts with
the inclusions to form steam in the microstructure:
Cu2 O + 2H ! 2Cu + H2 O steam)
Owing to this reaction, local high pressures cause the
ssuring of the copper grain boundaries, leading to the
complete break-up of the part by intergranular failure.
Only oxygen-free copper grades can be considered for
such piece-parts, particularly if brazing, welding, or
heat-treatment is to be performed in a reducing atmosphere.
A routine test may be performed to assess the suitability of
copper grades. It is recommended in ASTM B379, and it
involves annealing off-cuts from the copper batch at 800875
C for 20 min in atmospheres containing at least 15 % hydro-
gen. Typical results are shown in Fig. 4.9ad. The test speci-
men must then pass a subsequent bend test without cracking.
Brazing procedures are discussed further in Sect. 6.10.
4.4.6 Future Developments
The design of hypersonic vehicles fuelled by hydrogen will
be exposed to a variety of hydrogen environments. Pressure
vessels and fuel management devices will be exposed to
liquid hydrogen at cryogenic temperatures. Motors and
exhaust nozzles will experience high hydrogen temperatures
at great pressures. At temperatures above about 650 C most
structural materials will become saturated with atomic
hydrogen, and, as such, these materials can be considered as
hydrogen-alloyed (Nelson 1988). On cooling, the matrix
may become supersaturated with hydrogen, leading to
property degradation due to hydride formation or the cre-
ation of molecular hydrogen. Further damage may result
from a reheating cycle when dissociation of metal hydride
particles causes a localized hydrogen supersaturation.
Attention must also be drawn to the fact that hydrogen in
its atomic form will react with graphite at ambient tempera-
tures to form methane gas. Metalmatrix composites con-
taining graphite and silicon carbide bres may readily react
with hydrogen to form cavities of high-pressure methane gas.
4.5 General Corrosion Problems
4.5.1 Bimetallic Corrosion-Related Failures
Corrosion has to be considered during all manufacturing and
pre-launch phases of spacecraft. Certain equipment may also
4 Spacecraft ManufacturingFailure Prevention
have to be designed to withstand strongly corrosive envi-
ronments during its useful life of possibly 40 years in orbit.
Such equipment includes battery units, heat pipes and sys-
tems, and pressure vessels containing liquid fuels. Excepting
emergency conditions, space hardware on-board proposed
spaceplanes, will not be exposed to uncontrolled terrestrial
environments. However, pre-launch phases and ground
storage periods for single shot rockets and reusable space-
craft may be long and it has been necessary to develop and
implement a plan which ensures that all structural or
load-bearing metals are adequately protected from general
surface corrosion and stress corrosion cracking. Manned
space modules such as those of the International Space
Station will also be subject to humidity and condensation on
internal cold surfaces. Assessments of corrosion protection
nishes are generally made by accelerated testing, using
saline solutions (e.g. 3.5 % NaCl) for structural hardware,
and temperature/humidity exposure for electronics. Specic
cases and test environments will be described in other Sec-
tions of the book.
For electrical connections, it is highly desirable to select
metallic combinations from the approved couples detailed in
Tables 2.1 or 2.2. However, it is also important to note that
many of the less noble metals shown in this tabulation
require additional protection from general surface corrosion
in the form of platings, conversion coatings, anodic lms,
paints, etc., and additional polarization testing is often nec-
essary. Metals never considered for use in structural systems
of a spacecraft are widely used in the fabrication of its
electronic equipment. Some of the rarer metals are found in
transistors and micro-miniature circuits. As an example of
the potential problems to be encountered from dissimilar
material corrosion some of the different alloy systems used
in the construction of various spacecraft electrical systems
are indicated in the following list:
1. Iron and steel (ferrous alloys) are used as component
leads, magnetic shields, transformer cores, brackets,
racks, and general hardware.
2. Aluminium and aluminium alloys are widely used in sys-
tems as antennas, structures, chassis, supports, and brackets.
3. Magnesium alloys are occasionally used as equipment
housings.
4. Stainless steel is used for mounting racks, brackets, and
hardware.
5. Copper and copper-based alloys are generally used in
systems as contacts, springs, leads, connectors,
printed-circuit-board lands, and wire.
6. Fibre-reinforced plastics are used in waveguide fabrica-
tion (graphite bres), circuit boards (glass bres), and
4.5 General Corrosion Problems
component support adhesives (metal and oxide-loaded
epoxies).
7. Nickel and tinlead plating are used for protective coat-
ings and for solderability purposes. The use of tin in
solder is a well-known application; however, tin-plating
is also common on RF shields.
8. High-purity electrodeposited gold has wide application in
electrical connectors, printed-circuited runs and edge
connectors, miniature coaxial connectors, semiconduc-
tors, leads, and contacts.
9. Silver is used normally as a plating material over copper
in waveguides, miniature and microminiature circuits,
wires, contacts, high frequency cavities, and RF
shielding.
From this listing it is evident that the potential for gal-
vanic corrosion is inherent in the design of such equipments;
this will become apparent in later chapters.
Until recently, magnesium alloys with their high
strength-to-weight ratios have been frequently selected for
the construction of spacecraft structures and the housing of
electronic systems. General spacecraft requirements now
stipulate that all units must be electrically grounded to their
structures, and, at the same time, possess an adequate surface
corrosion protection system. This has led to the virtual
exclusion of magnesium from present spacecraft design
(however, selective copper plating, see Sect. 4.28, small
areas of the magnesium skin can be solder-connected with
wires to the grounding structure in alternative designs where
the entire remaining surface will be painted). Commercial
conversion coatings, such as DOW 7, may be suitably
conductive but, being thin, they frequently give rise to cor-
rosion problems (see Fig. 4.10a). Figure 4.10a shows the
effect of slight condensed moisture on an experiment hous-
ing constructed of magnesium alloy after equipment quali-
cation by thermal cycling. A non-conformance review
board considered that the housing had been removed from
the thermal chamber while still very coldleading to water
condensing onto all surfaces. The white, dust-like magne-
sium oxide corrosion product was also considered to be a
contamination hazard.
New, high-strength magnesium alloys and metal matrix
composites having high strength-to-weight ratios may nd
new applications on launch vehicles. Here, corrosion must
be avoided by the strict control of surface contamination,
flux inclusions from welding operations, galvanic couples,
and ensuring that surface protection schemes are suited to
the application. Chromate conversion coatings will delay the
onset of natural surface oxidation for a short time, and it is
preferable that they are coated with resin-based paints. One
chemical conversion coating free of hexavalent chromium
(Cr-free) is based on molybdate; recent work shows this
129
nish to provide excellent corrosion protection to the mag-
nesium alloy AZ31 (Yao & Basaran 2013). Another Cr-free
chemical conversion coating for magnesium alloys is the
MAGNOLYA treatment (European Patent pending) being
process-developed by Tecnalia. This coating can be applied
by immersion of parts into the solution, or by brush. It leaves
a thin coating on the magnesium alloys so far tested
(EV31A, AZ91 and AM60). The coating process involves
pickling and activation pre-treatments, followed by
phosphate-permanganate-rare-earth conversion coating and
drying. It appears to be suitable for parts with very close
tolerance mating surfaces and may be a candidate for
spacecraft housings, such as the one shown in Fig. 4.10a.
The bare MAGNOLYA coating on EV31A meets one of the
space industrys requirement standards, ASTM B-117, as
shown in Fig. 4.10be. The test samples shown in Fig. 4.10
(Brusciotti et al. 2014) were exposed to salt spray at 6 off
from the vertical and exposed for 168 h to the salt spray.
Additional salt spray test exposures were made on EV31A
samples with resin and primer coatings applied to both the
Cr-free coating and its reference Cr6+/Mn coating; the results
found the MAGNOLYA coated system to be comparable or
superior to the presently-used system.
Of the many hard anodizing treatments developed for
magnesium, the HAE and DOW 17 processes are the most
commonly used, and, when present as a thick lm, they offer
excellent abrasion resistance. A useful highly protective
nish for spacecraft (even when exposed to tropical launch
site environments) was the application of a sealing layer to
the anodic lm comprising a thin, clear high-temperature
baking epoxy resin, followed by a chromated epoxy primer
and epoxy top coat. It would appear that such nishes are
resistant to temperatures of 200 C (Rendu and Tawil 1988).
More recently, the plasma oxidation treatment of magnesium
alloys, which utilizes micro-discharges to change to surface
of the alloy into hard ceramic layers, has become wide-
spread. Here the magnesium is both protected from corro-
sion and this coating also possesses extremely good wear
properties (Blawert and Srinivasan 2010).
Contamination of an aluminium alloy cooling loop by a
few minute particles of copper swarf (Fig. 4.11) has been
identied as promoting the failure of this system. Connec-
tion of copper, or its alloys, to aluminium is a most unfa-
vourable galvanic couple, particularly in aqueous
environments. In this instance, a small number of hemi-
spherical pinholes, originating from the inner surface of the
channel, have been produced by the localized corrosion cell
set up between the cathodic copper particle and the alu-
minium. Away from the pinholes, the aluminium surfaces
have not been corroded.
Figure 4.12 details a severe corrosion pit within a
camera-pointing mechanism constructed of aluminium alloy
130 4 Spacecraft ManufacturingFailure Prevention

Fig. 4.10 a Magnesium

experiment housing of 25 cm
diameter nished with DOW 7,
showing corrosion along bottom
surface of ring (A) and around the
threaded stainless steel helicoil
inserts (B). A tape test conrmed
these white corrosion products to
be loose and powder-like. b A
typical salt spray cabinet for the
assessment of surface corrosion
resistance of metals according to
ASTM B-117. Here, magnesium
alloy EV31A, with the new
Cr-free MAGNOLYA chemical
surface treatment, is being
evaluated against an existing Cr6
+
/Mn chemical corrosion coating
presently used in the aerospace
industry (Brusciotti 2014). c,
d show respectively, the
MAGNOLYA coated magnesium
alloy before and after exposure to
168 h salt spray and the resulting
absence of corrosion. e, f are
before and after pictures of the
Cr6+/Mn reference sample
showing extensive corrosion sites
[courtesy of Tecnalia, Parque
Tecnolgico de San Sebastin
(Spain)]

2024-T6. This part had been nickel-plated and subjected to a surface protection system, but, like many nickel plates, it
saline solution for 30 days as part of a qualication pro- contained microscopically ne cracks resulting from high
gramme. The nickel-plating was intended to provide the residual stresses in the plating, and caused the substrate to
4.5 General Corrosion Problems
Fig. 4.11 Copper (cathodic) surface contamination has resulted in
severe pitting and leakage of an aluminium cooling channel
Fig. 4.12 Pitting corrosion of aluminium alloy 2024-T6 enhanced by
local porosity in 8 m thick nickel plating, after 30 days in saline
solution
become exposed. In these locations, galvanic corrosion of the
aluminium is enhanced by the large nickel area (EMF dif-
ference between Al-2024 and nickel is approximately
0.41 V). This particular alloy also contains a high copper
content, and the deep lamellar corrosion channels observed in
Fig. 4.12 result from selective galvanic attack between
undissolved cathodic particles of CuA12 and the alloy matrix.
The choice of metals for specic corrosive environments
may be made by referring to corrosion handbooks. Unfor-
tunately, important details are occasionally overlooked. Two
historical accounts are now given to illustrate the use of
incompatible metals in contact with corrosive environments.
Problem no. 1 was caused by nickel-cadmium battery
cells which had been installed in a satellite and were found
131
Fig. 4.13 a White crystalline potassium carbonate on surface of
leaking ceramic-to-metal seal on terminal of nickelcadmium battery
cell (potassium hydroxide electrolyte). b SEM photomicrograph of
section made through leaking battery terminal. The energy dispersive
X-ray analysis of marked area identies the presence of Al, K, Ni, Mn,
and low Mo. The displayed low-wavelength distribution highlights two
large potassium peaks
to have developed slight leaks (Fig. 4.13a, b). The potassium
hydroxide electrolyte had seeped out onto the cell tops and
formed a white deposit which was analysed to be potassium
carbonate, the reaction product between the electrolyte and
carbon dioxide from the atmosphere. The central terminal
posts of these battery cells are separated from their case
bodies by a ceramic insulator. The cell construction mate-
rials had been designed to ensure proper matching of
132
expansion coefcients. A strong metallization system which
could be painted onto the ceramic insulator had been
selected. This would then be red and subsequently brazed
to form a reliable ceramic-to-metal seal. The metallization
chosen for the alumina ceramic was molybdenumman-
ganese metallic powder. During ring, the powder sinters
together; the manganese component is oxidized, and this
reacts with the alumina and its secondary siliceous phases to
form a fluid glass (in the MnOAl2O3SiO2 ternary system).
This liquid then penetrates between the grains of the alumina
and encompasses the molybdenum metallic particles. The
exposed, electrically conductive molybdenum particles
could then be plated with nickel. Coppersilver eutectic
preforms were used subsequently to braze the ceramic into
the Kovar cell cap. To satisfy long-life requirements, it was
necessary for these seals to be completely inert to the cell
electrolyte, a solution of potassium hydroxide, in the pres-
ence of oxygen, hydrogen, and applied potentials up to
1.5 V. Short-term, accelerated tests on the seals produced
satisfactory results, but did not promote the actual failure
mode which had initiated long-term cell leakage. The failure
analysis of the leaking cells (Fig. 4.13) revealed that the
minute amount of free molybdenum in the metallization was
being chemically dissolved and leached out by the potassium
hydroxide electrolyte. The leak path had a width of between
0.5 and 2 m, and complete penetration of these caps had
occurred within 18 months after nal assembly. Additional
cells were developed utilizing a titanium active metal met-
allization in conjunction with the silvercopper braze and
these have remained hermetically sealed for more than
ten years. In active metal brazing, titanium is added, between
5 and 10 %, to 72Ag28Cu entectic. This is done mechan-
ically to produce sheets and braze metal pre-forms. During
brazing the titanium is active and changes the surface
composition of the alumina ceramic so rendering it wettable
by the liquid braze metal. The reactivity of these titanium
coppersilver alloys requires the brazing operation to be
carried out in a vacuum or a very good inert gas environment
(see also Fig. 6.53).
Problem no. 2 concerns incompatible materials which
may be introduced unwittingly into fully satisfactory units.
Reaction control systems, which are based on hydrazine
fuel, suffer from severe material limitations because very few
metals have been found to be compatible with hydrazine or
its decomposition products. Qualied thruster designs were
noted, during particular service tests, to operate at an unac-
ceptably high temperature. All the existing thruster con-
struction materials were known to be compatible with
hydrazine; however, to increase the thermal dissipation of
heat from thruster chambers, a small modication was made
to existing hardware. This consisted of tagging thin copper
ribbons onto the outside surface of each thruster chamber by
means of a very small electron beam weld. Unfortunately,
4 Spacecraft ManufacturingFailure Prevention
Fig. 4.14 0.8 mm thick copper ribbon electron beam welded to
superalloy. Note extent of alloying and depth of grain boundary
penetration (in box)
the depth of weld penetration was far greater than imagined,
as shown in Fig. 4.14. Copper is seen to alloy with, and
penetrate along, the grain boundaries of the chamber mate-
rial. Local penetration of copper caused it to appear on the
inside surface of the chamber. Thruster rings introduce
hydrazine into the copper-contaminated areas and, within a
few minutes, the chambers were observed to leak owing to
the rapid dissolution of copper by the reactive hydrazine.
Armed with this knowledge, satisfactory modications
were implemented by relocating the position of the copper
ribbons to less critical parts of these thrusters.
4.5.2 Corrosion Resistance of Anodic
and Chemical Conversion
Coatings on Al 2219 Alloy
4.5.2.1 The Alloy 2219
The Al-2219 alloy is a heat-treatable wrought alloy devel-
oped by Alcoa in 1954. It provided industry with a material
which had elevated-temperature (260300 C) properties
exceeding those of all other aluminium alloys. The weld-
ability of this alloy is excellent. The mechanical properties of
both wrought and welded 2219 are also excellent at tem-
peratures down to 250 C.
As with all AA 2000 series alloys, 2219 has somewhat
less resistance to atmospheric corrosion than the
lower-strength AlSiMg (AA 6000 series) and AlMg (AA
5000 series) wrought alloys. Inhomogeneities frequently
initiate localized surface corrosion attack, and for maximum
resistance to corrosion the composition of each aluminium
alloy must be kept as homogeneous as possible. This prin-
ciple applies to all the AA 2000 series of alloys, of which
2219 is typical. The high copper content (5.86.8 % by
4.5 General Corrosion Problems
Fig. 4.15 Al-2219-T851 (forged rings, no surface protection) follow-
ing exposure to a 3.5 % NaCl solution for 30 days. This micrograph
clearly highlights pitting corrosion adjacent to the coarse network of
electro-positive intermetallic particles (red-orange, arrowed). Most of
these large inclusions are shown up by polarized light as predominately
CuAl2 which, owing to the high copper content of this alloy, was
unable to dissolved completely into the saturated solid during the
weight) of 2219 generally depresses the electrode potential
of aluminium in the cathodic (more noble) direction, as
shown in Table 2.1.
From both a mechanical strength and an optimum corro-
sion resistance stand-point it is essential that, during alloy
fabrication, copper should be fully dissolved into the alu-
minium. It is the solid-solution, or homogenizing, treatment
that dissolves the copper, and this is performed in the tem-
perature range 535 5 C. Cooling from this temperature must
be rapid to prevent the formation of the intermetallic com-
pound CuAl2 at the grain boundaries. If excessive numbers of
CuAl2 intermetallics do form at grain boundaries they are
surrounded by adjoining volumes of alloy depleted in copper,
and this will facilitate corrosion attack as shown in Fig. 4.15.
The decisions to select the structural alloy Al-2219 for
manned spacecraft such as the Space Shuttle, Space Station,
Spacelab, the Columbus Module and Ariane 5 and 6, were
133
solution treatment. These CuAl2 particles have no doubt been broken
up and redistributed during the forging treatment, but they exist in a
displaced network that had originally formed during solidication of
the cast metal. The micrograph also shows both Al2(OH)3 (white,
arrowed) and green CuCl corrosion products seeping from corroded
micro-channels. See also Fig. 4.19d for further explanation. 500
magnication (Dunn 1984)
based on the need for a material having an optimum com-
bination of properties including those of mechanical strength,
fracture toughness, and resistance to general corrosion and
stress-corrosion cracking. This material is commercially
available as sheet and plate, extruded rod, and bar; it can be
forged and made available as Alclad sheet and plate.
In the 1980s, Spacelab (see Fig. 3.1) became the largest
joint space programme undertaken by Europe and the USA.
Since its rst flight in 1983 and 1998 there have been 22
Spacelab missions flown by the Space Shuttle, and the
overall flight performance of Spacelab has been excellent
(see also Sect. 8.7.3). The welded 2219 module structure has
remained leak tight in orbit even though it has been
assembled and disassembled many times for experiment
payload changes. It has outlasted the design requirement for
10 years life (even though the number of missions during
this time has not reached the 50 which were also expected).
134 4 Spacecraft ManufacturingFailure Prevention

4.5.2.2 Surface Protection Treatments

The Alclad forms of 2219 have a very high inherent resis-
tance to corrosion and may be used without the application
of further protective coatings. The most effective protection
of machined structural parts is obtained by the application of
an epoxypolyamide paint to shot-peened or electroplated
surfaces of the alloy. The following nishes are generally
selected for space use:

Alclad products A core of 2219 sheet material is sand-

wiched between thin sheets of either pure aluminium or
7072 alloy, then hot-rolled to effect bonding. Surface has
a high resistance to corrosion and is sufciently anodic to
the 2219 core to afford electrochemical protection.
Mechanical nishes Sandblasting or shot peening gives a
rough matt nish which causes surface of 2219 to be
under slight compression. This may slightly reduce sus-
ceptibility to general and stress corrosion but must be
covered with an organic coating.
Anodizing (see Fig. 4.16) Anodic coatings formed by
electrolysis in sulphuric acid or chromic acid baths.
Sulphuric acid provides thickest and most corrosion
protective nish (225 m depending on anodizing time)
for 2219. This has been conrmed by a more recent study
(Danford 1994) for AA 2024, 7075 and 6061.
Chemical conversion coatings (see Fig. 4.17) Processes
such as Alodine and Irridite provide only temporary cor-
rosion protection on 2219. They are an excellent base for
paints.
Electroplating 2219 can be immersed in sodium zincate
Fig. 4.17 Welded 2219 plate following Alodine 1200 protection and
of controlled composition. This is a satisfactory base for
exposure to 14 days in salt spray. Corrosion pitting is evident to depths
depositing nishes of copper then nickel, chromium, of 12 m (etched in Kellers reagent to reveal grain structure). a Parent
gold, silver, etc. metal. b Weld pool

Vapour deposition 2219 can be coated with pure alu-

minium by an ion vapour deposition method. This
coating is smooth, ne grained, and very adherent.
Painting Oil and oxides on 2219 must be removed by
dipping 2219 in solvents then phosphoric acid (room
temperature). Surface can be mechanically treated,
primed with a conversion coating or special primer, then
painted with epoxypolyamide or polyurethane resins.

4.5.3 Evaluation of Alodine Finishes

on Common Spacecraft
Aluminium Alloys

Our rst programme to assess the performance of a chem-

Fig. 4.16 Sample of chromic acid anodic oxidation following a
chemical etching treatment. The rough surface represents a worst case
situation. The anodic lm is continuous and remains uncorroded after
the standard 30-day salt-spray test
ical conversion coating on three common spacecraft alloys
(Al-2024, Al-6061 and Al-7075) has been performed (Scott
1985). The conditions imposed were SALT SPRAY
(CORROSION TEST) to simulate the extensive ground
4.5 General Corrosion Problems
storage of spacecraft equipment in an uncontrolled sea-coast
environment and THERMAL CYCLING UNDER
VACUUM, an in-orbit space environment known to degrade
surface coatings due to a mismatch in the coefcient of
expansion between the coating and substrate. The properties
of the alloys and their coatings were assessed by measuring
infrared reflectivity, solar reflectivity, resistivity, optical
properties (SEM and photographic recording), outgassing
rates, and infrared absorptivity.
The general corrosion properties of the unprotected alloys
are as follows:
Al-2024 Poor resistance to marine environments. More
susceptible to corrosive attack than other groups
of wrought-aluminium alloys. Nature of attack
influenced by thermal treatment.
(a)
(b)
2.0 m
Fig. 4.18 a SEM views of Alodine 1200 coating on Al-2024-T3 after
thermal cycling under vacuum to ECSS Q-ST-70-04 (5000), and b on
Al-2219 (as-coated, no environmental testing) in a region containing
large copper-rich CuAl2 inclusions (8000)
135
Al-6061 Good resistance to marine environments. Nature
of attack influenced by thermal treatment.
Al-7075 Attacked by a marine environment and generally
less corrosion-resistant than other wrought alloys.
All specimens were obtained from rolled plate and
machined to a standard surface roughness before being
Alodined for three different periods to produce three layer
thicknesses. A typical coating is seen in Fig. 4.18a. The
conclusions from this study were as follows:
(a) The surface structure of the chromate conversion
coatings was similar for all alloys.
(b) The corrosion properties were found to depend on the
copper content of the substrate and thickness of the
Alodine layer. As the copper content increases, so the
corrosion resistance decreases. The thicker coatings
always resulted in better corrosion protection.
(c) The effect of thermal cycling under vacuum on the
specimens was to make the coatings darker because of
water loss and to reduce the corrosion resistance of
alloy Al-2024. Alloys Al-6061 and Al-7075 were rel-
atively unaffected by the thermal vacuum.
(d) Infrared reflectivity was dependent on the thickness of
the coating, and a reduction in reflectivity with
increasing thickness was noted. Reflectivity increased
after thermal vacuum, and this is again related to the
water loss from the coatings by the environmental
effects of temperature and vacuum. The corrosive
environment had little effect on the reflectivity when
the specimens did not corrode. If corrosion did occur,
then reflectivity decreased.
(e) Electrical resistance was also dependent on coating
thickness and increased as coating thickness increased.
The resistance of all the uncoated alloys dramatically
increased after corrosion, but with coated alloys there
was little change. The surface resistance was found to
be proportional to the degree of corrosion.
Our second programme was initiated when EU legislative
regulations indicated that Alodine 1200 might become a pro-
hibited substance as it was known to be composed of hex-
avalent chromium (Cr6+). It was agreed that this study would
consider alternative, industrially available conversion coatings
that contained no Cr6+. The objective was to nd an alternative
having good corrosion protection, adequate electrical proper-
ties for grounding purposes and thermal properties that would
not degrade in a space simulated environment (thermal cycling
under vacuum). The substrate alloys selected were: the high
strength aerospace alloys, Al-7075 and Al-2219, and the
non-structural alloy Al-5083. The chemical conversion coat-
ings nally selected were: Alodine 5700 (organometallic zir-
conate complex chemistry from Henkel), Nabutan STI/310
136 4 Spacecraft ManufacturingFailure Prevention

Fig. 4.19 Schematic diagram of

a an intermetallic, after polishing,
embedded on the surface of an
Al-2219 wrought product,
b corrosion reactions in the
sea-salt environment (i.e. NaCl
solution), c after long exposure to
the solution and, d schematic
diagram of the reactions inside
the pit that lead to the deepening
of crevices as seen in Fig. 4.15
(Grilli et al. 2010)
4.5 General Corrosion Problems
(Fluorotitanic acid chemistry from Chemo-Phos), and
Iridite NCP (thought to be fluorotitanate chemistry from
MacDermid) together with Alodine 1200S (chromic acid with
complex fluorides from Henkel) as the control coating.
Chemical cleaning was also applied by different methods prior
to the coating process in an attempt to achieve a clean,
defect-free initial surface. The extensive report (Pereira et al.
2008) concluded that only Alodine 1200S survived thermal
cycling and that the other coatings had a tendency to spall-off
(producing contaminant particulates). The electrical resistivity
values of the other samples were very similar, even after
thermal cycling. Only Alodine 1200S survived the salt spray
testing on the high strength alloys. It was recommended that
space projects should continue with this well-proven product,
and to continue the search for an alternative.
The four chemical conversion coatings are likely to be
suitable for most non-critical applications when applied to
the lower-strength alloys such as Al-1100, Al-3003, Al-5052
and alloys in the Al-6xxx series. These, and the high strength
Al-2xxx and Al-6xxx series of alloys will also benet from
all coatings when they are applied as a base for nal nish
painting.
A third programme was designed to investigate the
manner in which a conventional Cr6+ conversion coating
reacts with a bare aluminium alloy containing undissolved
CuAl2 inclusions, such as Al-2219, (Grilli et al. 2010). By
combining scanning auger microscopy (SAM) and EDX
images it was possible to deduce the distribution of the
conversion coating as a function of alloy microstructure. The
conversion coating is thick and uniform all over the alloy
matrix (and XPS indicated this was a mixture of Cr3+ and
Cr6+ species), cracking occurs during, or shortly after the
completion of the Alodining process to expose the alu-
minium substrate (Fig. 4.18b). On the intermetallic inclusion
surfaces the chromium concentration is much lower,
implying that the conversion coating is not able to protect
these more noble areas of the surface.
A fourth programme involved investigating samples of
Al-2219 having the titanium based Nabutan STI/310 coating.
SEM and SAM analyses conrmed the TiF based coating
was also generally uniform across the surface, but as with the
Alodine samples, the CuAl2 inclusions only supported traces
of a titanium compound. Again, this indicated a reduced
corrosion protection for the inclusions. Three precipitates
were selected on the samples surfaceeach had a different
size. The Ti- chemical conversion coated sample was then
exposed to a 3.5 %NaCl solution for progressive exposures
of 15, 45 min, 2, 4, and 8 h. At the end of each exposure the
sample was re-entered into the vacuum chamber and exam-
ined by SEM, AES and EDX analyses. Out ndings (Grilli
et al. 2010) revealed the fact that the coating had deposited
preferentially on the aluminium matrix, but, because of the
cathodic nature of the intermetallics with respect to the
137
matrix, it was the coating itself and the matrix that suffered
degradation. Corrosion products rich in aluminium and
oxygen were seen to have deposited around the intermetallic,
and their volume increased with time of exposure to the NaCl
corrosive solution. For the very small precipitates (<1 m
diameter) corrosion had penetrated underneath the precipi-
tate, so releasing it from the samples surface.
The workhorse, high strength alloy Al-2219, in the T6
and T8 heat treated condition has been selected for Spacelab,
the Columbus module of the International Space Station,
Arianes 5, 6 and many other US and European space pro-
jects. Its corrosion properties, discussed within this Chapter,
are of particular importance due to the airborne salt particles
that will deposit onto rockets and other space hardware
stored at sea-coast launch sites. Al-2219 has a matrix of
aluminium containing a Cu content of up to 6.8 %, Fe 0.3 %,
Mn 0.4 %, and Ti 0.1 %. Our analyses have shown that the
undissolved and precipitated second phase intermetallics
contain mainly AlCu (as CuAl2) with small fractions of Fe
and Mn (Al-2219 contains a saturated solid solution of
copper, designed for maximum precipitation hardening
properties so there will be inevitable undissolved CuAl2
within the matrix microstructure). The cathodic nature of the
intermetallics with respect to the aluminium matrix will
cause preferential oxidation of the less noble matrix as was
seen in Fig. 4.15. In the absence of an effective coating,
sever pitting corrosion is known to be prevalent at sea-coast
launch site environments. The schematic diagrams in
Fig. 4.19 are thought to represent the corrosion mechanisms
most likely to occur. Progressive corrosion leads to exten-
sive pitting, or crevice corrosion, and these defects could
becoming stress raisers. For alloys susceptible to stress
corrosion cracking this can lead to catastrophic fractures (see
for instance Figs. 4.21 and 4.22).
4.5.4 Cleaning, Passivation, and Plating
of Spacecraft Steels
The preferred steels for spacecraft construction have been
tested and found to possess a high resistance to
stress-corrosion cracking; these are listed in Table 2.3. This
listing contains a complete range of steels from the
low-strength plain carbon family to complex, high-alloy
duplex steels which have a very high strength if they are
correctly precipitation-hardened. The plain carbon steels and
the carbon, low-alloy steels all have a low resistance to
corrosion. That is, they produce a loose oxide which breaks
off under very little stress, allowing further corrosion to take
place. Such steels have few spacecraft applications but may
be used as shear blades in pyrotechnic cutters, as springs (see
Sect. 4.4.2) and as component housing. They must always be
plated to prevent the occurrence of corrosion. They may be
138
forbidden from some scientic experiments owing to their
strong magnetic properties.
Steels that have poor resistance to corrosion (1000 series
and 4000 series) are vapour-degreased and pickled (possibly
in hydrochloric acid of strength 15 %, or in a 75 % solution
of sulphuric acid in water) for a short time in order to remove
light oxides. The plating is applied after all basis metal
heat-treatments, machining, welding, brazing, and forming
have been performed. An underplating of copper having a
thickness of between 5 and 10 m is often initially applied.
This is usually followed with an overplate of 1020 m of
electrodeposited nickel. There is a hazard that hardened and
tempered cold-worked or cold-straightened parts might crack
during cleaning and plating. Such parts need to be stress-
relieved at 180200 C for three or more hours prior to
plating. The same heat-treatment must be performed again
after the plating operation (and within 4 h of the plating) to
provide for hydrogen-embrittlement relief.
Steels that are corrosion-resistant are usually not plated,
but given a so-called passivation treatment. Passivation is a
post-fabrication method of maximizing the inherent corro-
sion resistance of stainless steels, but care should be taken to
apply the correct method to the particular steel, as will be
stated below. This treatment forms a protective oxide on the
steel surface in a controlled and repeatable way. Two solu-
tions based on nitric acid are applied separately to two dif-
ferent steel groupings:
Passivation treatment A, made up of 2.5 % (by weight)
sodium dichromate and 25 % (by volume) nitric acid, at 48
50 C for 20 min, for types
Precipitation hardening: A-286, 15-PH, PH15-7Mo, 17-7,
AM-350, AM-355
Free machining: 303, 347, 430
Ferritic 400: 405, 409, 429
Martensitic 400: poor SCC resistance 440C, 410, 414, 416,
420, 431
Passivation treatment B, made up of 2545% (by volume)
nitric acid, at 2030 C for 30 min, for types
Austenitic 200: 201, 202
Austenitic 300: 301, 302, 304, 310, 314, 316, 316L, 321,
329, 347
After passivation the parts should be rinsed clean in hot
water, then dried.
Carburized surfaces cannot be passivated because the
carbon combines with the chromium on the surface to form
chromium carbides.
Nitrided stainless steel should not be passivated, as the
solutions will severely corrode the nitrided case.
4 Spacecraft ManufacturingFailure Prevention
Martensitic high-strength alloys such as 440C are subject
to hydrogen embrittlement and intergranular attack when
exposed to acids; they also have a low resistance to SCC.
Quality control tests are usually required for the in-line
testing of spacecraft items.
For plated layers the thickness can be measured
destructively on samples by microsectioning. Nondestructive
eddy current measurements can be made using equipment
calibrated to known standards. Plating adhesion is tested by
clamping and bending the sample and inspecting for ruptures
and delamination. To check for embrittlement, samples are
held for 200 h at a load of about 115 % of the yield load, and
then examined for cracks or fracture.
Passivated parts can be checked for continuity of oxide
lm by exposing samples to the copper sulphate test. The
test solution is made up of 40 g cupric sulphate, which is
dissolved into a glass beaker containing 900 ml of distilled
water to which has been added 100 g of sulphuric acid
(having a specic gravity of 1.84). The parts can be
immersed into the solution or swabbed continually for 6
8 min. Visual inspection shall verify that the tested surfaces
are free of copper deposition. The test solution must be
removed from useful parts by immersing them into cold
running water, then re-passivate in one of the aforemen-
tioned solutions.
Several specications exist covering the chemical passi-
vation of stainless steel parts. The favoured ones are: ASTM
A380M, ASTM A967M and for electroplating such parts
ASTM B254-92 (all dated 2014).
Attempts to utilize greener passivation processes have
been successfulthey are based on the use of citric acid
passivation; increasingly popular with fabricators wishing to
avoid the use of mineral acids or sodium dichromate. Citric
acid is a bio-based material that releases no toxic fumes
during passivation, requires low concentrations and is easily
disposable. Parameters for passivation with citric acid for a
wide range of stainless and ferritic-FM steels is provided by
DeBold and Kosa (2004). This nish has been fully studied
by NASA, including exposures at the KSC Beachside cor-
rosion test facility (see Fig. 3.16) where samples showed no
signs of corrosion after 6 months (Lewis et al. 2013).
4.5.5 Launch Site Exposure and Corrosion
Launch complexes such as the Ariane facilities at the ESA
Guianese Space Centre in Kourou and the NASA Kennedy
Space Centre on Cape Canaveral suffer from severely cor-
rosive environments owing to their close proximity to the
Atlantic Ocean. These site locations and the usefulness of
corrosion testing has been presented in Sect. 2.4. Very high
salt-falls are deposited on these coastal sites throughout the
year as they are exposed to both northeast and southeast
4.5 General Corrosion Problems
trade winds. The chloride concentration in precipitation in
the Guiana region was measured at the turn of the twentieth
century to be equivalent to 193 lb/acre/year (216 kg/ha/year,
or 60 mg/m2/day of deposited salt) (Miller 1906). The author
could not uncover more recent data! Similar precipitation
can be expected at the KSC. Corrosion rates of carbon steel
at KSC are almost four times those further along the coast at
Daytona Beach and forty times those inland in Pittsburg PA
(Calle 2014).
Wind moving over breaking sea waves will pick up small
droplets of saltwater. When these droplets and tiny particles
or airborne salt accumulate on exposed metal surfaces, cor-
rosion can occur, particularly when the atmospheric relative
humidity is high. Sub-micron-sized salt particles may deposit
within the cleanrooms of spacecraft integration halls, and
later, when atmospheric conditions of temperature and
humidity permit, will absorb water and form a highly con-
ductive solution. The science of breaking waves and the
release of salt particles has become a science in its own right
and Sharkovs book (2007) contains astounding amounts of
information, mostly gained from Russian Cosmos satellites,
and interpretation of data by the Russian Academy of Sci-
ence. Small droplets (aerosols) of salt water produced by surf
tend to be coarse (520 m) and those produced by whitecaps
are generally smaller (0.53 m). Due to gravity, the large
droplets will deposit close to the shore-line, whereas the small
aerosols can be transported considerable distances (Cole et al.
2004). The Cole paper includes salinity maps of Australia.
Here, very few coastal regions reach the 60 mg/m2/day salt
deposition rates of the Guianese Space Centre. It may be
remembered that the proposed Cape York Spaceport project
was to have transformed the northern tip of Queensland into a
high tech, international space centre. According to Coles
map, the proposed location appears to have appreciably low
airborne salinity levels (approximately 4 mg/m2/day).
Plans for the UK to become the home of Europes rst
spaceport will be developed in 2016 when a detailed technical
specication is due to be published. Interestingly, each pro-
posed site has a different environment and will suffer from
different corrosion rates (based on the average zinc corrosion
rate in m/year). Urban coastal areas and areas of high salinity
can differ between 2 and 8 as detailed in the Zinc Millennium
Map, published in the excellent design and metallurgical
publication: 2014 Engineers & Architects Guide.
Corrosion by electrolytic action can result when two
dissimilar metals are involved. Also, corrosion of a single
metallic part can occur when, for instance, a saline solution
reacts chemically with the parts surface. Impingement of
salt laden aerosols can of course, provide a conductive
electrical path and short-circuit electrical equipment.
Acceptance of the laboratory corrosion tests described in the
previous sections as an exact representation of the corrosion
occurring at a specic spacecraft assembly location, or
139
launch site, may result in erroneous conclusions. It is rec-
ommended that where corrosion by the atmosphere is an
important factor, on-the-spot tests are needed. A simple test,
such as the samples wire-on-bolt test (Doyle et al. 1963)
can be conducted on the site, with test samples located at
various heights above the ground. However, facilities such
as those depicted in Fig. 2.16 will provide far better data,
and give reliable predictions as to a materials suitability for
the launch site environment.
The various launch sites (see Figs. 2.2 and 2.17) depict
several metallic structures such as the Mobile Gantries, and
the lightening masts that give protection the launch vehicle in
case of tropical storms Pylons and masts are slender and have
a certain amount of flexibility under high wind conditions.
The structures are usually constructed from mild steel and are
heavily covered with hot-dipped galvanization having often a
composition of 0.3 % lead, 1.0 % tin and some bismuth, rest
zinc. Metallurgical considerations have been given to such
structures, as their bases are often welded and can suffer from
wind fatigue, resulting in fracture. The study of low cycle
fatigue of galvanized lightening masts has been performed at
the University of Genoa and calculation procedures are pre-
sented for elimination such failure (Solari & Repetto 2014).
An associated problem with welded galvanized steel is the
phenomenon of liquid metal embrittlement (LME), and liquid
metal induced stress corrosion. These cracks occur when
liquid zinc penetrates between the steel grainsit reduces
ductility and causes high localized stresses between the
grains. Trained and certied welding operators will know of
these problems and, particularly if the structure is under stress
during welding, they may wish to try and unload the parts or
take measures to reduce the stress levels. Unfortunately these
cracks are not often visible from the surface, which is coated
with zinc. Dye penetrants or X-ray inspections are not always
useful, but eddy current testing can detect LME cracks. New
welding wires have been developed to avoid or reduce LME;
salt spray testing of such weldments appear to show
promising results (Kodama et al. 2010).
4.6 Stress-Corrosion Resistance of Metals
4.6.1 Stress-Corrosion Cracking
The stress-corrosion cracking (SCC) of metals is initiated at
several sites by localized corrosion, by pit formation, or by
intergranular attack. The influence of stress on the corrosion
rate is not very specic above a particular threshold stress,
and the other principal factors involved in a failure are
environment, time, grain orientation, and heat-treatment
condition. The stresses that will produce failure by SCC in
service may be due to the residual stresses in the metal caused
by fabrication (quenching, forging, differential expansion),
140
by assembly stresses (resulting from improper tolerances
during t-up, overtorquing, press ts), or a combination of
residual and operating stresses induced by service loads.
SCC at the launch site (see also Sect. 4.5.5) is a major
contribution to failures in ground support equipment. At the
NASA Kennedy Spaceflight Centre in Florida, the problem
has been highlighted by Majid (1988), who considered that
75 % of all SCC failures could have been prevented if those
persons responsible for design had been sufciently aware of
the problem. In designing piece-parts the following must be
seriously considered:
1. Tensile properties of metals.
2. The materials susceptibility to SCC.
3. Fracture toughness.
4. Type of corrosive environment.
Testing for stress-corrosion resistance may be undertaken
by plotting applied stress against specimen life in a particular
corrosive environment. The shapes of such curves will pre-
sent data on stress-corrosion susceptibility. The sample
conguration may have the form of uncracked tensile
test-pieces having a cylindrical or dumb-bell conguration,
or precracked samples that resemble fracture toughness
test-pieces.
In testing it is common practice to load specimens for
30 days at a single stress level, that of 75 % of the mean
yield stress, although there have been some instances where
a higher value, up to 90 % of the yield stress, has been
favoured. The SCC resistance of metals, particularly alloys
of aluminium, is always less when tension is applied in a
direction transverse to the grain orientation. It is least for the
short transverse direction. The criteria used in the selection
of materials for spacecraft and associated equipment and
facilities so that SCC will be prevented are detailed in
European specication ECSS-Q-ST-70-36. The constant
load method is most favoured (i.e. specimens stretched in
spring-loaded devices) over other methods involving con-
stant strain loading (i.e. specimens stretched or bent in jigs to
a xed position). It is accepted that specimens having short
life will fail in a similar time when tested by either the
constant stress or the constant strain methods. However,
materials with a high resistance to SCC will give longer lives
under constant strain owing to relaxation of the stress by
creep (plastic deformation). Structural material failures in the
early USA space programmes were most commonly caused
by SCC, resulting in much loss of hardware, time, and
money (Johnson 1973). Of particular concern were complex
parts that had been machined from thick sections. Particu-
larly vulnerable were the regions exposed to short transverse
loads, because SCC threshold stresses are lowest in this
transverse direction. Also, of critical importance is the sus-
ceptibility of welds to SCC. This applies to all the main
4 Spacecraft ManufacturingFailure Prevention
classes of structural alloys. Surprisingly little is known about
such weldments as, in general, it is the parent metals that are
thoroughly tested (Irving 1992). The Ariane 4 ELV has also
suffered from SCC (see Sect. 5.19) and, even when alloys
are listed in Table 2.3, they can fail by SCC when exposed to
non-saline environments. The auxiliary power units of the
Space Shuttle Columbiawhich supply power needed for
activation of the rudder, speed brake, and elevonsfailed
owing to stress-corrosion cracking of the Hastelloy B
injector tubes. Here the environment was exposure under
stress to ammonium hydroxide vapours, resulting in the
leakage of liquid hydrazine (Korb et al. 1985).
Probably the most incongruous failures are those which
are life threatening; for instance, the selection of an AA
2014-T6 forging for a naval helicopter main rotor blade
grip. This alloy is known to have a poor resistance to SCC,
yet it was specied on the manufacturers drawingthe part
failed by SCC in the short transverse plane (Parker 1991).
Whereas the paper by Johnson (1973) lists many NASA
failures due to the SCC of aluminium alloys, there exists an
equally interesting early report (Stanley 1968) that describes
NASA launch vehicle failures due to the SCC of steel parts
including 17-4 PH, H-11, and 440C.
The ideal material is one that suffers from no stress-
corrosion crack propagation below the maximum stress
levels that may occur during the service life of the compo-
nent. Numerous papers have been devoted to the topic of
SCC and a good overview is given by Bussu (2009). There
appears to be a growing consensus that regardless of the
specic mechanism, hydrogen is involved with the SCC
propagation of both ferritic and austenitic stainless steel
(Hehemann 1985), and high-strength aluminium alloys
(Gruhl 1984; Newman 1990).
4.6.2 SCC Evaluation
An important design goal is to achieve a highly efcient,
low-weight spacecraft. This is possible only if materials are
selected from high-strength alloys such as those listed in
Table 2.3 with a high resistance to SCC. In Europe, space-
craft construction, materials, and their weldments are gener-
ally evaluated for SCC by the procedure detailed in ESA
specication ECSS-Q-ST-70-36 Determination of the sus-
ceptibility of metals to stress corrosion cracking. The method
is particularly applied to all new structural metals and their
special heat treatments. It is used whenever metals have been
welded or unusually processed and is applied as a quality
control during production when in-line samples are tested.
Many of the metals characterized as having a high resistance
to SCC, as listed in Table 2.3, were evaluated as follows: The
tests were carried out under alternate immersion conditions in
3.5 % sodium chloride over a thirty-day (720 h) exposure
4.6 Stress-Corrosion Resistance of Metals 141

Table 4.1 Two-pass TIG-welded 2219-T851 alloy specimens results of mechanical and stress-corrosion tests
Specimen Description Mechanical test results Stress-corrosion test results
number 0.2 % PS TS Elongation % on Time to fall Max. depth of
(N/mm2) (N/mm2) 50 mm (h) corrosion visible in
microsectiona (m)
Parent Weld
haz
B1 Tensile test specimen 148.6 279.8 7 Not exposed
B2 149.2 278.2 6 Not exposed
B3 151.0 272.5 6 Not exposed
Average 149.6 276.8 6
B8 SC specimen Not pulled U 364 192 None
B9 166.4 264.0 4 U 245 175 None
B10 164.3 271.6 4 U 380 400 None
B4 Unstressed control Not pulled U 280 375 None
B5 (720 h) 141.4 265.0 5 170 200 None
B6 144.3 267.0 5 308 500 None
2
Stress-corrosion specimens were stressed at 75 % of 0.2 % PS of weldment = 112 N/mm
U = not failed after 30-day test (720 h)
a
No stress-corrosion cracking was observed on these specimens

period. Unstressed control specimens were exposed to the
same environment to provide a basis for comparison in
assessing stress-corrosion susceptibility of alloys that survive
30 days in the test. The susceptibility is assessed by tensile
tests to compare the residual strengths of the specimens
exposed stressed and unstressed, and by metallographic
examination of microsections from stressed and unstressed
specimens to distinguished between stress-corrosion and
intergranular corrosion or pitting occurring independently of
stress. A typical test rig is shown in Fig. 3.15c.
The 0.2 % proof stress was established for each type of
sample from an average of three tensile tests. These values
would establish the stresses at which the stress-corrosion
tests were to be conducted and they provide reference values
of tensile strength and elongation for comparison with
specimens to be tensile tested after surviving 30 days in the
stress-corrosion test.
The stress-corrosion tests were carried out in triplicate,
the specimens being loaded in tension by calibrated springs.
The time to failure was recorded automatically for the

Table 4.2 Single-pass EB-welded 2219-T851 alloy specimens results of mechanical and stress-corrosion tests
Specimen Description Mechanical test results Stress-corrosion test results
number 0.2 % PS TS Elongation % on Time to fall Max. depth of corrosion
(N/mm2) (N/mm2) 50 mm (h) visible in microsectiona
(m)
Parent Weld
haz
J1 Tensile test specimen 179.3 287.0 4 Not exposed
J2 182.1 296.8 3 Not exposed
J3 180.9 300.1 3 Not exposed
Average 180.8 298.0 3
J4 SC specimen Not pulled U 520 None None
J5 321.6 3 U
J6 282.8 3 U
B7 Unstressed control Not pulled 500 None None
B8 (720 h) 242.8 1
B9 264.8 2
Stress-corrosion specimens were stressed at 75 % of 0.2 % PS of weldment = 136 N/mm2
U = not failed after 30-day test (720 h)
a
No stress-corrosion cracking was observed on these specimens
142 4 Spacecraft ManufacturingFailure Prevention

Table 4.3 BS L93 alloy plate specimens taken in short transverse direction results of mechanical and stress-corrosion tests
Specimen Description Mechanical test results Stress-corrosion test results
number 0.2 % PS TS (N/mm2) Elongation % on Time to fall Max. depth of corrosion
(N/mm2) 50 mm (h) visible in microsectiona
(m)
Pitting Pitting plus
cracking
E1 Tensile test 334.1 373.9 4 Not
specimen exposed
E2 326.4 372.7 4 Not
exposed
E3 337.4 375.3 4 Not
exposed
Average 332.6 374.0 4
E8 SC specimen (15 % Not pulled U 370 nc
E9 PS) 360.0 360.0 1 a
U 980 nc
E10 361.7 372.0 1a U
E4 Unstressed control Not pulled 1000 nc
E5 (720 h) 341.7 342.0 2a
E6 328.3 365.0 2 860 nc
ST4 SC specimen (25 % Not pulled U 900 850
ST5 PS) 326.9 1 U 650 800
ST6 251.4 1 U 630 400
ST7 Unstressed control Not pulled 400 nc
ST8 (720 h) 267.8 1 400 nc
ST1 SC specimen (25 % Failed in SC 625 800 1450
ST2 PS) test 504 650 670
ST3 504 370 500
ST9 Unstressed control 348.9 1 600 nc
(435 h)
ST10 Unstressed control 341.1 1 324 nc
(504 h)
Mark E specimens had 60 rnm2 cross-section area
Mark ST specimens had 36 rnm2 area
U = not failed after 30-day test (720 h)
nc = no cracking
a
Specimen broke outside gauge length

stressed samples. They were then removed together with the
corresponding unstressed sample. After removal from the
salt environment, and at the end of the thirty-day period, all
specimens were washed in warm water, with gentle scrub-
bing to remove most of the salt and corrosion products on
them. They were dried in warm air and stored in a desic-
cator. The samples were then tensile tested and subjected to
metallographic examination in accordance with the
ECSS-Q-ST-70-37 test procedure.
The ESA constant-load test overcomes the repro-
ducibility problems associated with constant-strain tests,
which suffer owing to relaxation of the stressing jig; or after
the onset of SCC in small specimens, there is plastic strain
and creep with an attendant reduction in the initial load. Once
the SCC ratings have been determined (i.e. highly resistant
moderate, or low resistance to SCC) it may be necessary to
perform fracture mechanics testing. The additional work is a
requirement for alloys having a moderate or low rating which
will be used in any structural application. The ESA test is
similar to a screening test for alloys and weldments, and
indicate that below 75 % of the materials 0.2 % proof stress
there will be no onset of SCC. The fracture mechanics tests
involve pre-cracked specimens (see Sect. 3.2.2.8), and it is
necessary to obtain a threshold value, at a crack tip, when the
stress-intensity factor (known as KISCC) will not cause the
crack to propagate while the sample is being exposed to the
liquid test environment. Note that the usual test liquid is a
3.5 % sodium chloride solution, but tests are also performed
4.6 Stress-Corrosion Resistance of Metals 143

Table 4.4 a AlLi alloy plate, 8090-T8771 (short transverse) results of mechanical and stress-corrosion tests. b AlCuLi alloy platea,
2195-T8P4 (longitudinal direction) results of mechanical and stress-corrosion tests
Description Mechanical test results Stress-corrosion test results
0.2 % PS TS Elongation % on Time to Max depth of cracking in
(N/mm2) (N/mm2) 15 min fail (h) microsectionb (m)
a
Tensile test specimen 305.3 455 2 Not
exposed
302.7 396b 0b Not
exposed
310.3 464 1 Not
exposed
Average 306 460
Stress-corrosion specimens loaded 214.5 1500
at 150 N/mm2 223 1500c
242 1500
Corresponding unstressed controls 302.9 441 0 U 0
N.T. N.T. N.T. U 0
Stress-corrosion specimens loaded 37.5 75
at 175 N/mm2 99.5 140
196.5 150
Corresponding unstressed controls 295.8 460 0 U 0
307.7 432 0 U 0
N.T. N.T. N.T. U 0
b
Tensile test specimen 604 635 6.4 Not
exposed
603 636 6.9 Not
exposed
604 638 14.0 Not
exposed
Average 604 636 9.1
Stress-corrosion specimens loaded 557 584 5.9 U 394
at 453 N/mm2 562 586 3.9 U 349
U 446
Average 559 585 4.9 242
Corresponding unstressed controls 563 586 3.7 327
561 585 4.9 437
460
Average 562 586 4.3 0
No evidence of stress corrosion cracking was found. Classied Class 1 per ECSS-Q-ST-70-37A
U = Not failed after 30-day test (720 h)
N.T. = Not tensile tested
Note All stressed specimens showed some degree of intergranular attack associated with the stress-corrosion cracks
a
Composition Al3.9Cu0.04Fe0.9Li<0.01Mn0.03Si0.02Ti0.02Zn0.14Zr0.32Ag
b
Specimen failed prematurely owing to large inclusion in gauge length
c
See Figs. 4.22 and 5.4

in liquids such as hydrazine, nitrogen tetroxide, cleaning Each test programme is reported in detail but a standard
solvents, or any other liquids that might come into contact summary table is produced as illustrated by Tables 4.1, 4.2,
with the metal during its service life. 4.3 and 4.4. Corresponding photomicrographs of the
144
Fig. 4.20 a TIG-welded 2219 stressed to 75 % of PS. No SCC was
noted, particularly no cracks in the detailed inspection of the areas
arrowed (6.5). b Single-pass EB-welded 2219 stressed to 75 % of PS.
No SCC was noted, particularly no cracks in the detailed inspection of
the areas arrowed (6.5)
corrosion features appearing on the SCC test samples are
also shown in Figs. 4.20, 4.21 and 4.22.
The current development of aluminiumlithium alloys,
listed in Table 3.2, by the main aluminium producers is
being followed with great interest. Spacecraft structures
Fig. 4.21 BS L93 alloy plate, stressed to 25 % of PS showing a pit at surface and b detail of the crack tip
4 Spacecraft ManufacturingFailure Prevention
could benet from these alloys, which have a 520 %
increased modulus of elasticity and an 85 % reduced den-
sity compared with the present aluminium alloys employed.
It is unfortunate that stress-corrosion testing has shown both
AA 2090 and AA 8090 to be susceptible to SCC, and it
would appear that, in direct contrast to the benecial effects
of thermal treatment on the SCC resistance of conventional
aluminium alloys, the AlLi alloys are almost insensitive to
an improved SCC behaviour by the application of ageing
treatments (Lumsden and Allen 1988; Buis and Schijve
1992). These results are conrmed by the stress-corrosion
cracks illustrated in Fig. 4.22. The cladding of AlLi alloys
with pure aluminium by pressure-rolling processes to form
composite sheet may be one method of avoiding the SCC
limitations when general structural parts are considered, but
these alloys can be used for many nonstructural applications
such as electronic housings, face-skins of a platform, cor-
rugations of a thrust cylinder, etc. Some recent developments
in vacuum rening are shown to improve resistance to SCC
(Webster 1994) but in the main potential users remain cau-
tious against a wide acceptance of the AlLi range of alloys
(Nurse 1996). To date, no AlLi alloys are listed as Table I
alloys having a high resistance to SCC, or Table II, moderate
resistance to SCC (MSFC-STD-3029 rev. A, Feb. 2005).
Al-2195 was used for the Shuttle external tanks, the Orion
capsule and is being evaluated by ESA. Similarly, Al-2099
has many aerospace applications, from structures to cryo-
genic tanks but great care must be taken to avoid loading
these alloys, or weldments, to beyond the K1SCC value for
the alloy and its heat treatment condition.
4.6.3 The Properties of Spring Materials
Table 2.3 gives details of metals that have good resistance to
stress-corrosion cracking. A review of the suitability of these
metals for springs has been performed by the Spring
Research and Manufacturers Association, with particular
4.6 Stress-Corrosion Resistance of Metals 145

Fig. 4.22 a AlLi alloy 8090-T81 3 mm sheet, stressed in the long
transverse direction at 75 % PS. Severe SCC causes specimen failure
after only 3 days alternate immersion. Etched microsection (50). b Al
Li alloy 8090-T8771 plate, stressed in the short transverse direction at
50 % PS. Failure, and severe SCC is noted after only 9 days alternate
immersion. This alloy is categorized as being Class III per ECSS
Q-70-37 (the SCC test method standard). Picture shows the as-polished
microsection (200). c illustrates a typical example of how sustained
stresses are introduced into a mechanical connectionexcessive
clearance between lugs result in normal stresses parallel to the
short-transverse (ST) direction of the materialsuch a design must
not incorporate AlLi 8090-T8771 thick plate, or any other Table III
alloy per ECSS Q-ST-70-36 (the standard listing alloys as having a
high, medium and low resistance to SCC), (Bussu and Dunn 2002),
courtesy of ESA

Table 4.5 Metals not suitable or seldom used for spring manufacture
Steel alloys Condition Nickel alloys Condition Copper condition alloys Miscellaneous alloys Condition
2
Carbon steel (100 Below 125 kg/mm Hastelloy X All CDA 110 37 % reduction Beryllium S-200 C Annealed
series)
Low-alloy steel Below 125 kg/mm2 Incoloy 800 All CDA 194 37 % reduction HS 188 All
D6AC, H11 Below 148 kg/mm2 Incoloy 901 All CDA 195 40 % reduction Titanium 3Al-2, 5 V All
HY80, HY130, Quenched and Incoloy 903 All CDA 230 40 % reduction Titanium 13V-11Cr, All
HY140 tempered 3Al
AM 350 SCT 1000 and Inconel 600 Annealed CDA 422 37 % reduction Magnesium alloys
above
AM 355 SCT 1000 and Inconel 625 Annealed CDA 443 10 % reduction
above
Almar 362 H 1000 and above CDA 619 40 % reduction
Nitronic 33 All CDA 688 40 % reduction
CDA 706 50 % reduction
CDA 725 50 % reduction
(annealed)
146 4 Spacecraft ManufacturingFailure Prevention

attention paid to metallurgical condition, mechanical prop-
erties, and the temperature limits of these materials (the
SRAMA as originator is gratefully thanked for permission to
publish this information).
Some of the metals in Table 2.3 are not capable of being
hardened sufciently for springmaking, and others are
seldom, if ever, used for springs. These metals have been
listed in Table 4.5, and are not considered further.
A full evaluation of all other alloys for spring manufacture
is given in Tables 4.6, 4.7, 4.8 and 4.9. Where appropriate the
relevant UK, USA, and/or DIN specication that applies to
these metals when used for springs has been indicated,

Table 4.6 Steel alloys generally considered suitable for spring manufacture
Music wire (ASTM 228)
Music wire is supplied in the cold-drawn pearlite condition within the size range 0.16.0 mm. The tensile strength of the wire supplied for
spring-making depends on size, but will be within the range 1600300 N/mm2. The temperature limit for music wire is 150 C, above which
excessive spring relaxation will occur, and this material can be safely used at sub-zero temperatures. Music wire may be supplied to BS 5216 M
or DIN 17223 grade D, which are approximate equivalents to ASTM A228
1095 spring steel (quenched and tempered)
1095 steel is generally used for springs in strip form. The mechanical properties of this material are achieved either by hardening and tempered or
austempering, and the appropriate hardness depends on the material thickness. A guide to the appropriate hardness values is given in BS 5770
part 3 under the heading CSHT 95. This material should not be used for springs which are likely to operate above 170 C, and it is prone to brittle
failure at sub-zero temperatures. It should be noted that the stress-corrosion resistance of 1095 steel is poor in the quenched and tempered
condition, but is not so poor if the material is austempered
300 series stainless steel
300 series stainless steels are only strong enough to be used for springs in the hard-drawn (wire) or hard-rolled (strip) condition. 302 and 316
types are the principal grades used for both strip and wire, and strip is also supplied to 301 specication. 305, 321, and 347 grades are also used
occasionally in the form of hard-drawn wire. The mechanical properties of these materials depend upon the size supplied, and are specied in BS
2056, ASTM A313, or DIN 17224 for wire and BS 5770 part 4, ASTM A117, or DIN 17224 for strip
The temperature limit for 300 series stainless steel springs is 250 C, and these materials may be safely used at sub-zero temperatures. The
stress-corrosion resistance of 316 is better than that for 301 and 302 stainless
21-6-9 stainless steel
21-6-9 is a nitrogen-solid-solution-strengthened stainless steel and may be used for springs in the 30 % cold-drawn/rolled condition with a tensile
strength of 1200 N/mm2 approximately. The maximum spring operating temperature is 250 C
Carpenter 20 Cb and 20 Cb-3 stainless steels
Carpenter 20 Cb-3 has replaced 20 Cb and is used in the cold-drawn condition for springs. The mechanical properties are similar to those of 316
stainless, and the maximum spring operating temperature is 250 C
A 286 stainless steel
A 286 is an age-hardening stainless steel that is used for springs either in the cold-drawn/rolled and aged, or in the solution-treated and aged
condition. The mechanical properties after age-hardening are approximately 1000 N/mm2 for the solution-treated and 1500 N/mm2 for the
cold-worked material as shown in AMS 5525. The maximum spring operating temperature is 400 C
Custom 455
Custom 455 is used for springs in the cold-drawn and age-hardened condition, but is prone to stress-corrosion cracking in the H 900 condition
specied in ASTM A-313. In the H1000 condition, the tensile strength is 1400 N/mm2 minimum. The maximum spring operating temperature is
about 300 C
15-5PH stainless steel (H1000 condition and above)
15-5PH and, the similar, 17-4 PH grades are used for springs in the H1150 M condition, which is the averaged condition in order that they are
resistant to sulphide stress cracking. The tensile strength in this condition is typically 862 N/mm2. The maximum operating temperature is
approximately 300 C
PH 14-8 Mo stainless steel
PH 14-8 Mo is another precipitation-hardening stainless steel, which may be used for springs in the CH 900 condition and above
PH 15-7 Mo stainless steel
PH 15-7 Mo is a higher-strength precipitation-hardening stainless steely which is generally used in the cold-drawn and aged condition. The
tensile strength of 2.35 mm diameter wire is typically 2000 N/mm2 in this condition. The maximum spring operating temperature is 300 C
4.6 Stress-Corrosion Resistance of Metals 147

Table 4.7 Nickel alloys generally considered suitable for spring manufacture
Hastelloy C
Hastelloy C-276 is the principal Hastelloy grade used for springs, and is used in the cold-drawn/rolled and stress-relieved condition. The
mechanical properties are indicated in ASTM B575 specication
Inconel 718
Inconel 718 is an age-hardenable nickel alloy that is supplied for springs in the 1550 % cold-drawn/rolled condition. The mechanical properties
after age-hardening as specied in AMS 5597A are typically 1400 N/mm2. The safe spring operating temperature range is 196 to 550 C
Inconel X 750
Inconel X 750 is an age-hardenable nickel alloy that is supplied for springs in one of two conditions: 15 % cold-reduced (No. 1 temper) to AMS
5698D and 65 % cold-reduced (spring temper) to AMS 5699C. The 15 % cold-reduced product is age-hardened at 735 C for 16 h to give a
tensile strength of approximately 1250 N/mm2. The alloy in this condition has a maximum spring operating temperature of 500 C and is known
to operate safely down to 196 C. The spring temper material may be age-hardened for maximum strength by ageing at 650 C to give a tensile
strength of 1800 N/mm2 minimum, but at this strength the maximum safe spring operating temperature is 370 C. For the maximum spring
operating temperature of 600 C, the spring temper material should be triple heat treated as specied in AMS 5699C
Monel K-500
Monel K-50 is an age-hardening copper-nickel alloy used for springs in the solution-treated and aged or cold-drawn/rolled and aged condition.
Typical mechanical properties are given in BS 3075 (wire) and BS 3076 (strip). The maximum spring operating temperature is 260 C
Nispan C-902
Nispan C is a constant modulus nickel-iron alloy for which the modulus is virtually constant over the temperature range 50 to 150 C. The alloy
is generally supplied in the cold-drawn/rolled condition and is age-hardened after springmaking to give the optimum spring properties for the
application
Rene 41
Rene 41 is an age-hardenable nickel-based alloy generally used in the solution-treated and age-hardened or lightly cold-drawn and age-hardened
condition for springs. In strip form ASM 5545 is applicable. The tensile strength of Rene is 15002100 N/mm2, depending on the amount of cold
work. The maximum spring operating temperature is 550 C
Waspaloy
Waspaloy is another age-hardening nickel alloy generally used in the solution-treated and age-hardened condition for springs. In sheet form AMS
5544 C is applicable. The maximum temperature is 500 C

Table 4.8 Copper alloys generally considered suitable for spring manufacture
CDA 170 annealed and precipitation hardened
CDA 170 is a copperberyllium alloy which is age-hardened at 335 C for 2 h after springforming to give a tensile strength of 1050 N/mm2
minimum. The maximum spring operating temperature is 125 C. The applicable specications for CDA 170 and 172 are BS 2873 CB101,
ASTM B197, and DIN 17682 Cu Be 2 for wire and BS 2870, ASTM B194, and DIN 17670 for strip
CDA 172 work hardened and precipitation hardened
CDA 172 is a copperberyllium alloy which is supplied for springmaking in the cold-drawn/rolled condition. After age-hardening at 335 C for
2 h, the tensile strength is 1240 N/mm2 minimum. The maximum spring operating temperature is 125 C
CDA 510 and CDA 521 37 % cold-rolled
CDA 510 and CDA 521 are respectively 5 and 9 % phosphor bronze alloys, which are used for springmaking in the cold-drawn/rolled condition.
The mechanical properties depend largely on the percentage reduction in area, but CDA 510 is available up to 900 N/mm2 and CDA 521 up to
1000 N/mm2. The maximum spring operating temperature is 80 C. The applicable specications are BS 2873 BP 102 and ASTM B159 for CDA
510 wire, BS 2870 and ASTM B103 for CDA 510 strip, DIN 17682 Cu Sn 8 for CDA 251 wire, DIN 17670 for CDA 521 strip

Table 4.9 Miscellaneous alloys generally considered suitable for spring manufacture
HS 25 (L605)
HS 25 is a cobalt alloy used for springmaking in the solution-annealed or lightly cold-drawn condition (15 or 30 % reduction in area). The tensile
strength of the alloy is between 1000 and 1500 N/mm2 depending on the amount of cold work. The maximum operating temperature is thought to
be about 500 C
MP35N
MP 35N is a multiphase nickel, cobalt, chromium, molybdenum alloy which is used for springs in the cold-drawn condition. The tensile strength
depends on wire size, but typically a minimum alloy strength of 1856 N/mm2 is available at 0.76 mm diameter and 1650 N/mm2 at 4 mm
diameter. This alloy is known to work satisfactorily down to 196 C, and the maximum spring operating temperature is about 500 C
Titanium 6A14V
This titanium alloy is usually given the designation Titanium 318 and is used for springs in the cold-drawn condition. The tensile strength in this
condition is approximately 1400 N/mm2, and the maximum spring operating temperature is 150 C
148
together with the metallurgical condition in which the alloy is
used. Many of the alloys have mechanical properties that
depend on section size, and in these cases it will be necessary
to refer to the specication for these data. The spring oper-
ating temperature limits indicated are those known to be safe.
The maximum temperature is generally governed by spring
relaxation properties. Where information is known about an
alloys usage below 0 C, this has been given.
4.6.4 Bearing Materials
Bearings for spacecraft mechanisms consist of balls, which
are the rolling elements, and the rings which form the
raceways. Where very high radial and thrust loads are
expected, then the contact area is increased by using taper
roller bearings in place of balls. Few bearing materials are
found to possess the performance required for spacecraft
bearings, these are based either on high alloy steels, or
occasionally, ceramics are used for special applications (for
instance the Shuttle and Ariane cryogenic bearings for
pumping liquid hydrogen have silicon nitride rolling ele-
ments). Commercial bearings are often made from AISI
52100, a low alloy steel containing 1.4 % chromium (see
Appendix 6 for some European equivalent designations).
Aerospace steels contain prescribed amounts of
high-chromium, carbon and molybdenum (1417 %Cr, 0.6
1.1 %C and 0.75 %Mo). A common corrosion resistant
steel is designated 440C; this has been selected for most
space bearings because of excellent performance and a
successful heritage. The combination of chromium and
carbon ensure that heat treatment, mechanical deformation
(forging, etc.) and induction hardening, provide a mixture of
ne M7C3 and M23C6 carbides (see also Fig. 5.7) in the
steels tempered martensitic microstructure. This composi-
tion provides the surface hardness required of the bearing,
usually between 59 and 61 HRC. The chromium gives the
steel some corrosion resistance, at least 12 %Cr is required
at the surface to form a thin, insulating chromium oxide
surface lm. Due to the formation of chromium carbides the
initially high-chromium available at the surface becomes
depleted, but the presence of molybdenum avoids the critical
depletion to below 12 %Cr.
Due to the critical surface condition care must be taken
during the handling of products made from 440C. This is not
an austenitic steel, such as stainless AISI 304 or 316,
possessing thick, very adherent chromium oxide nishes.
The thin chromium oxide lm on 440C can easily become
degraded and render the surface at risk to rusting and even
stress corrosion. In fact 440C is known to be a Table III
alloy, susceptible to stress corrosion cracking per ECSS and
NASA standards. The strict post-manufacture controls
should include:
4 Spacecraft ManufacturingFailure Prevention
cleaning after part manufacture (see Sect. 4.21)
passivation and control that surface is completely passi-
vated (see Sect. 4.5.4)
application of oils or greases (see Table 5.2)
Storage of bearings, mechanisms and similar corrosion
sensitive parts in sealed, leak-tight bags containing dry
nitrogen.
These preventative measures should avoid the damage to
bearings or similar mechanisms. They are based on the lessons
learned from some diligent failure analysis work (Saltzman
et al. 2009) performed on a solar array drive actuator where the
root cause was concluded to be: stress corrosion due to
incomplete cleaning, the presence of small amounts of chlorine,
sodium, water vapour and very long storage times.
4.7 Control of Printed Circuit Boards
4.7.1 Chemical Composition of Tin-Lead
from Microstructure
It is generally accepted in the space business that
printed-circuit boards (PCBs) to be assembled by soldering
support eutectic composition tinlead nishes which have
been electro-plated and then fused onto their high-purity
electrolytic copper conductors.
At least eight other pcb nishes are available but it seems
unlikely that these will nd extensive use for space boards.
Briefly, these are:
HASLhot air solder level, poor shelf live as solderability
degrades with time (Fig. 4.37).
OSPorganic solderable preservative, not possible to use
test probes without damaging coating and coatings probably
not stable under vacuum
ENIGelectroless nickel immersion gold, tarnishes and can
be susceptible to black pad.
Immersion tingood for ne pitch packages but may suffer
from whisker growth.
Nickel palladium with immersion silver or goldtarnishes,
or gold tin intermetallic problem.
Immersion silvertarnishes and may suffer from electro-
chemical migration.
Reflowed tin-leadpresently preferred, good solderability,
(nothing solders better than solder), solder slumping may
occur but, as will be shown below, many process precautions
and quality controls are needed.
From experience to date, hot oil fusing of tinlead n-
ishes by means of a two-stage heating cycle (see also
Fig. 4.24) has proved to be a controllable process (Dunn
1980). One supplier of hi-rel PCBs has, however,
4.7 Control of Printed Circuit Boards 149

Fig. 4.23 a SEM photograph of

PCB track surface showing the
dendritic appearance and
associated porosity (A) within the
fused tinlead nish. b The
microsection details the porosity
(A) seen on the track surface; a
thin band of intermetallic is just
visible (B). c Porosity also exist
within the plated-through hole of
this reject board (A), surrounded
by primary tin-rich islands
(B) and eutectic matrix (C).
d Properly plated and fused holes
should contain a fully eutectic
tinlead microstructure

encountered systematic problems resulting in rough, fused
tin-lead nishes containing a large degree of porosity
(Fig. 4.23a). A metallurgical investigation was instigated to
determine the severity of the problem and to recommend a
procedure which would eliminate the observed defects.
Microsections (Fig. 4.23b, c) were analysed by EDAX and
line-scan attachments to a SEM. Wet chemical analysis
conrmed the presence of slight copper contamination and a
hypereutectic composition of the tin-lead, tin being 69.4 %
whereas the eutectic composition of pure solder is 62.9 %.
The detailed photomicrographs highlight several large,
partially sealed pores in the cross-sectioned coating; the
normal coppertin intermetallic layer is just visible (A and B
of Fig. 4.23b, respectively). The microstructure of the
plated-through hole shown in Fig. 4.21c also details severe
porosity (A) and a ne laminated eutectic structure (C). This
phase distribution would be expected in a tinlead alloy
containing 69 % tin, the relative volumes of islands and
eutectic being given by the tinlead binary phase diagram.
The surface cavities and spherical pores associated with
the rejected boards are considered to result from the
volatilization of occluded organic materials during the hot
oil fusion process. At some time during solder fusion, before
the alloy had cooled to its solidus temperature, the presence
of tin-rich dendrites has entrapped gas pockets which would
otherwise have been released at the liquid solder-to-oil
interface. A revised plating procedure eliminated all trace
contaminants and, once fused, provided a microstructure of
almost exactly eutectic composition (Fig. 4.23d). No surface
imperfections were observed on PCBs made to this revised
procedure.
It was found to be essential that tinlead fusing trials
should be performed to establish the optimum fusion
schedule for both double-sided and multilayer boards in
order that they retain a good solderability after long term
storage. Research has shown that the metallographic
inspection of test coupons from boards can adequately
control the quality of nished boards.
150 4 Spacecraft ManufacturingFailure Prevention

4.7.2 Grainy Solder Coverage on PCBs on the same manufacturing line. Three of the boards had
and the Effects of Rework been reworked by a second fusing operation. An attempt was
made to establish the acceptability of each board for
spacecraft use.
Visual defects
Under ideal situations the surface of a fused tinlead printed
Tinlead coating procedure
circuit board (PCB) should be bright, shiny, and without any
The processing procedure leading to fluoroborate tinlead
graininess or roughness. Unfortunately, the real-life sit-
plating of the submitted boards followed conventional
uation is that such nishes can vary to show very different
PCBN manufacturing methods. Before tinlead plating the
optical surface effects related to the microstructure and
copper plating is cleaned by slight surface etching in sul-
metallurgical reactions which occur within the plated layer
phuric acid; there is no danger of embedment of particles
during the tinlead fusion process. This can give difculties
onto the copper surface as no abrasives are used. The
to inspectors responsible for the goods-inwards controls of
manufacturers intention is to plate tin-lead with a eutectic
high-quality products. The visual appearances have been
composition of 63 % tin, 37 % lead. However, even with
noted to change not only from one manufacturer to another,
daily analytical checks this is difcult to meet, and it has
but also between matches manufactured on the same process
often been found that variations in current plating density
line.
in different locations on the same boards give rise to
The visual aspects described in various PCB specica-
deviations from the eutectic composition by 8 %. Before
tions list major and minor defects which can occur in the
fusing, the tinlead surfaces are chemically etch-cleaned.
metallized PCB pattern nish. Apart from the more obvious
The oil bath preheating and fusion schedules are 170
types of defects such as lifted conductors, short-circuits, or
C/15 s and 190 C/5 s respectively. This is similar to the
out-of-tolerance pads and holes, there are more subjective
asymptotic heating recommended for preheating, as illus-
aspects such as dewetting on areas to be soldered, granular
trated in Fig. 4.24, and is particularly important in the case
aspects, and supercial scratches. The following subsec-
of multilayer boards having heat-sink zones when it might
tions describe the evaluation of ve different PCBs, each
otherwise be difcult to maintain the required 2 m of tin
having a unique surface appearance, which were produced
lead coverage around the corners of plated-through holes.

Fig. 4.24 Two methods of heating for fusion of tinlead plating on temperature. b Excess temperature heating strategy with short heating
PCBs which support zones of high (H) and low (L) thermal capacity. period but danger of overheating zone L
a Asymptotic heating strategy with excellent control over maximum
4.7 Control of Printed Circuit Boards
Table 4.10 Identication of samples
Board identication number Process history
1A/B Reworked by a second
2A/B Fusing operation
3A/B Reworked by a second fusing operation
4A/B As-fused
5A/B
a
All boards were rejected by the incoming inspection personnel of the procuring company. This was based on visual examination. A/B designate
respectively the top side and bottom side of each printed circuit board
Oil fusion was followed by cleaning in isopropyl alcohol
and bake drying for 2 h.
Five boards were submitted for this investigation. They
had been fused according to the aforementioned procedures
and, in addition, three were given a further fusion treatment
as a rework operation (210 C until thought re-melted by
the operator) in an attempt to improve their surface visual
appearances. All the boards were subsequently rejected at
incoming inspection by a company engaged in the compo-
nent assembly of spacecraft circuit boards, as detailed in
Table 4.10. The reasons for rejection were that each board
had an irregular surface appearance. Only two boards were
later dispositioned as rejectable by the PCB manufacturer.
Method of evaluation
The submitted boards were evaluated as follows:
Visual inspection and photography with oblique lighting.
Samples removed for microsectioning.
Samples removed for rotary dip solderability testing of
plated-through holes (ECSS-Q-ST-70-10 requires wet-
ting in 3 s and no dewetting after 8 s).
Component assembly soldering trials with a 2 mm diam-
eter soldering-iron bit held at either 270 C or 300 C.
Soldering was made according to spacecraft procedures
but with soldering times of both 2 and 8 s. Each resultant
joint was then visually inspected and later microsectioned.
The metallography was performed by using standard
techniques. Each board sample was mounted in a room
temperature curing, low exotherm resin, then ground and
polished to a 0.5 m nish so as to bisect each hole.
Microstructures of the various copper layers were revealed
by lightly etching the polished mounts in a mixture of 9 parts
concentrated ammonia and 1 part 30 % hydrogen peroxide.
The tinlead microstructure was visible in the unetched
condition, and the intermetallic bands of Cu3Sn and Cu6Sn5
were later highlighted by nal etching of the mount in a
mixture of 98 parts isopropyl alcohol and 2 parts concen-
trated nitric acid. Colour photomicroscopy was made at
magnications of up to 1350 with an oil immersion lens
attachment to a projection microscope.
151
Disposition of PCB manufacturera
Reject
Reject
Accept
Accept
Accept
Results and discussion
The ve submitted boards were noted to possess totally
different topographic features, as can be seen from
Figs. 4.25, 4.26, 4.27, 4.28, 4.29, 4.30, 4.31, 4.32, 4.33 and
4.34. The captions describe features ranging from, lumpy
solder islands with high contact angles and bright flat
dewetted regions, to, matt, dark, gritty nishes with large
dull islands. Detailed photographs highlight the pads around
plated-through holes; they were taken under identical par-
allel lighting conditions which accentuated the observed
features. These pad areas represent the most critical board
surfaces; they are required to have excellent solderability to
facilitate component assembly with both nonactivated fluxes
and low soldering iron temperatures.
Only the illustrations associated with Board no. 1 will be
reproduced in Figs. 4.26, 4.27, 4.28, 4.29, 4.30, 4.31, 4.32,
4.33 and 4.34 in order to demonstrate the microsectioning
technique and observations. Table 4.11 summarizes every
result of the various thickness measurements and solder-
ability tests. All boards, with the exception of Board no. 1,
had adequate tinlead coverage around the corners of each
plated-through hole.
Fig. 4.25 Board no. 1, bottom side (B). Pad and track surfaces feature
lumpy solder islands, having high contact angles with flat, bright
dewetted regions
152
Fig. 4.26 a Board no. 1 (B side). Detail of soldering pad from
Fig. 4.25. Microsection was made across line YX. b Board no. 1 (B).
Section through pad shown in (a), at the same magnication
Fig. 4.27 Board no. 1 (B), corner of plated-through hole supports a
total solder coverage of 34 m, but up to 1.5 m is coppertin with dark surface appearances possessed an overall lead-rich
intermetallic. This is a detail from Fig. 4.26b
The most suitable composition to ensure good tinlead
fusing and retain an adequate coverage at plated-through
hole corners is marginally hyper-eutectoidal, being 65 % tin
and 35 % lead. It is noted that in this investigation the boards
4 Spacecraft ManufacturingFailure Prevention
Fig. 4.28 Board no. 1 (B). Detail of dewetted area of pad shown in
Fig. 4.26b. This supports uncoated intermetallic having a thickness of
1.5 m. The solder island is 58 m thick
Fig. 4.29 Board no. 1, top side (A)
Fig. 4.30 Board no. 1 (A), hole corner is covered with 1.5 m
intermetallic and has dewetted
microstructure, whereas bright nishes are present on boards
having a tin-rich composition. These composition estimates
were made by comparison of the relative volumes of tin and
lead appearing in the microstructure.
The solderability test results are illustrated to show the
worst conditions of solder flow through the plated holes.
4.7 Control of Printed Circuit Boards 153

Fig. 4.31 Typical microsections

of Board no. 1 after solderability
test. a 3 s rotary dip. Solder wets
hole up to the B-side corner. b 8 s
rotary dip. No wetting around
B-side corner. c Detail of high
contact angle where solder is
unable to wet the intermetallic
ring around bottom-side corner
(1350)

Boards which have been reworked by a second fusing

Fig. 4.32 Soldering by skilled operator to Board no. 1, side (B).
Solder had been applied from top side (A). Soldering iron at 270 C for
23 s (joint 1) and 8 s (joint 2). Note complete dewetting of pad on joint
1. The microsections of these joints appear in Figs. 4.33 and 4.34
operation are noted to take longer for solder capillary flow
within each hole. Generally there is suitable flow after 8 s,
but, in the case of Board no. 1, when a thick band of
intermetallic is present around the hole corners, it is
impossible to achieve solder wetting, and a solder plug
features on the top side of the board (see Fig. 4.31). The
boards which have a poor surface appearance but which
were not reworked by the second fusing operation show
good solderability by this rotary dip method.
The soldering trials were made by a certied soldering
operator. Cored solder containing a nonactivated flux was
applied to pre-tinned component leads which had been
inserted into degreased plated-through holes. Locations to be
soldered were chosen so as to possess land masses of equal
sizes (i.e. similar heatsink effects), and additional liquid flux
was not applied during soldering. Under controlled
154
Fig. 4.33 Microsections through joints 1 and 2. Both joints were difcult to make. The poor solder flow caused excess solder on top llets
Fig. 4.34 Detail showing
nonwetting of solder around hole
corners, joint 1
conditions the operator was in a position to judge and
compare the ease of making each joint. Table 4.11 compiles
the results of these investigations, and it can be seen that all
board which had been reworked by a second fusing opera-
tion had potential soldering problems.
Photographs and microsections of the hand-soldered
joints made to Board no. 1 shown in Figs. 4.32, 4.33 and
4.34. As with the solderability test, it is noted that solder is
4 Spacecraft ManufacturingFailure Prevention
prevented from flowing around the plated-through hole
corners, and in addition there are signs of severe dewetting
on the bottom (B) pad surfaces. Severe solderability prob-
lems will exist when exposed coppertin intermetallics are
present at the knees of plated-through holes (Hagge and
Davis 1985). The formation of coppertin intermetallic does
not adversely affect the soldering as long as there is a coating
of tinlead to protect the intermetallic from oxidation. At
4.7 Control of Printed Circuit Boards 155

Table 4.11 Results of various evaluations

Board no. Process history Thickness measurements from micro sections (m) Solderability Soldering trials
Pad Corners Tests 270 C 300 C
Sn/Pb (min) Sn/Pb (min) Intermetallic (max) 3s 8s 23 s 8s 23 s 8s
1 Reworked 1.58.0 1.54.0 1.5 Fail Fail Fail D D Fail
2 Reworked 6.08.0 6.08.0 1.52.0 P.F. Pass OK OK D D
3 Reworked 8.0 8.0 1.5 P.F. Pass OK OK OK OK
4 As-fused 7.0 9.0 0.7 Pass Pass OK OK OK OK
5 As-fused 3.05.0 2.05.0 0.7 Pass Pass OK OK OK D
P.F. = Partial lling but adequate wetting, considered to pass test
D = Difcult (poor flow of solder causes excessively large llets)
OK = Acceptable

least 1.5 m of coating was recently determined to be critical
to ensure solderability (Ray et al. 1995). However, the
0.7 m of intermetallic that forms during the hot oil fusion
will continue to grow at a rate of about 1.0 m per year
dependent on storage temperature and humidityso it can
be seen that all tinlead coatings will have a nite useful life
(Craven 1993). It should be noted that in both the rotary dip
test and the soldering trials the solder was applied from the
top side (A) of each board. Joint inspection was made to
accept/reject criteria of spacecraft soldering specications,
and no attempt was made to rework the defective joints.
Although the hand-soldered joints depicted in Fig. 4.32 do
not meet visual inspection requirements, it can be argued
that the mechanical and electrical integrity of these con-
nections are probably not impaired by the absence of solder
on the PCB bottom side llet. Unfortunately it is difcult to
make nondestructive tests on soldered joints so that much
importance is made of the cosmetic appearance of solder
llets and wetting angles.
Conclusions
The reworking of a high-rel spacecraft PCB by a
second fusing operation is considered unacceptable. The
additional reflow treatment reduces the solderability of
each through-plated hole and increases the band of
coppertin intermetallic from a thickness of 0.7 m
(maximum) to over 1.5 m. These thicker layers of
intermetallic produce dewetting of surfaces to be sol-
dered (e.g. at the corners of plated-through holes and on
termination pads). Unacceptable solder joint quality can
result if an attempt is made to mount components to
such boards.
The unusual surface features of as-fused PCBs are
undesirable and will cause problems during incoming
inspection (Bulwith 1986). No soldering problems were
encountered on non-reworked tinlead nishes, but by
visual inspection alone it is impossible to establish if a
surface feature is a cosmetic or an actual defect.
Destructive evaluations can be made on the test coupons
attached to delivered PCBs; they must include solder-
ability tests, pad adhesion tests, and microsectioning.
In the non-heat-treated condition, as-plated tinlead is
present as a duplex structure consisting of almost pure
lead grains which are dispersed within a tin-rich matrix
(Dunn 1977). Each zone will have a theoretical melting
point of 327 C for lead and 232 C for tin. It is therefore
desirable for the as-plated duplex grain size to be small in
order to facilitate solid-state diffusion and interalloying.
Under these conditions eutectic liquidation will be
achieved during a shorter oil fusing time (for any given
oil bath temperature).
4.7.3 Evaluation of Multilayer Board Internal
Connections
Plated-through holes provide the electrically conductive paths
from one side of a double-sided PCB to the other side. In the
case of multilayer PCBs the plated-through hole may inter-
connect more than 20 layers of conductor patterns. The ther-
mal expansion coefcient for solder and copper in a multilayer
board assembly is approximately 1517 m/mm/C. In con-
trast, the epoxyglass board substrate expands at approxi-
mately 100340 m/mm/C in its Z-axis and is nonlinear in its
temperature increase between room temperature and the sol-
der melting temperature of 183 C (see also Table 6.1). From
this it can be seen that during solder fusing, soldering opera-
tions, and thermal cycling under service conditions the
epoxy-glass expands more than the copper plating within the
hole. This will cause mechanical stressing at the point where
the internal pads connect with the copper plated-through hole.
Work hardening and subsequent fatigue cracking can result at
these junctions, as can be seen by the examples illustrated in
Fig. 4.35ad. These illustrations show how inherently defec-
tive multilayer boards (from nonapproved sources) can pass
electrical continuity tests on receipt at a spacecraft contractors
156 4 Spacecraft ManufacturingFailure Prevention
 4.7 Control of Printed Circuit Boards
b Fig. 4.35 a Photomicrographs of cracked pth internal connection.
Cross-section made across top part of plated-through hole. Note crack
(between rst internal copper conductor and the plated-through copper)
extends completely around hole. Detail of crack at the interface
between internal copper conductor and plated-through copper. Electri-
cal contact is made at only one small point (arrowed). Additional
cracking across internal conductor at a lower circuit. b Photomicro-
graphs to illustrate manufacturing defects on multilayer board (8 layers)
supplied to experimenter from non-approved source. Exposure to high
site. It is important that destructive microsections are made
both by the circuit board manufacturer as part of his nal
inspection and by the customers incoming inspection
department; this is done on the test coupons attached to flight
quality boards. It is now recognized that this microsectioning
task must be performed excellently in order to locate the
presence of the very ne microcracks present in Fig. 4.35b.
Often, grinding and polishing will cover over the microcracks.
Alternatively, when the copper is etched, an uninformed eye
may mistake the dark line between the copper laminate and the
plated copper as a crack when it could be a series of grain
boundaries. The thermal shock caused during the component
lead soldering operation has immediately opened the
defective interfaces shown in Fig. 4.35c. Alternatively, ther-
mal cycling at unit level testing can thermally fatigue these
connections. Worse still would be the generation of open
circuits caused by in-orbit thermal fatigue due to component
heat dissipation or solar heating.
Other defects may be shown up during thermal cycling,
such as the delamination and resin recession referred to in
Fig. 3.19b. Delamination can be prevented by black-oxidizing
the copper foils before laminating the multilayers in order to
increase the available surface area for bonding. Processes
such as etching the copper to a minimum depth of 1.25 m,
control of the etchant composition, and strict control of the
black copper oxide thickness can be assessed by standard peel
strength and solder float test methods followed by metallog-
raphy (Jackson 1987). It is essential to avoid delamination if
the thermal fatigue of internal connections is to be avoided.
The plated-through copper must also have a very high purity
and a suitable grain structure to enable a certain amount of
plastic deformation during temperature excursions. Trace
elements or impurities in the electrodeposit interfere with
dislocation motion, and very ne grain structures prevent
extensive dislocation movement during plastic deformation,
causing the plating to be hard. If dislocations cannot move, the
applied loads caused by thermal expansion mismatches cause
the plated hole to fracture, resulting in intermittent open cir-
cuits. Suitable copper electrodeposits have a low hardness
(less than 100 kp/mm2) and an elongation at fracture of greater
than 10 %.
The fatigue life of multilayer boards, when designed to
withstand thermal cycling, can be markedly improved by:
increasing the diameter of the plated-through holes; by
157
temperature or thermal fatigue will result in these cracks growing and
causing electrical open circuits. c Usage of the multilayer board shown
(as-received) in (b) has resulted in open circuits. Thermal shock created
during the soldering operation has caused wide cracks at the internal
conductor to plated-through hole interfaces (courtesy of ZVE).
d Beginning of a corner crack in the electroplating of a
plated-through hole caused by resin smear on the edge of the copper
cladding (courtesy of IFE)
increasing the copper plating thickness; and by locating the
functional pads at the centre of the board thickness (i.e.
internal connections at the neutral axis). By optimizing these
design parameters, impressive improvements have been
demonstrated by Yoder et al. (1993), with no failures when
plated-through holes were subjected to 2000 thermal cycles
(55 to +100 C) over a sixteen-month period as part of a
NASA space environment qualication programme.
Specications for the procurement of multilayer boards
rely heavily on metallographic examination of
plated-through holes from the test coupons representative of
the production boards in order to ensure quality and relia-
bility. These examinations are made after coupons have been
exposed to thermal excursions that simulate conditions that
the board might be exposed to during initial assembly
operations and repair or reworking operations. Separation of
internal connections and the presence of resin smear are
examined for in the as polished condition. The microsec-
tion is then lightly etched to detect cracks in the copper
plating or foil, etching conditions, lamination, drilling
parameters, hole cleaning, and etchback processes.
The EDAX attachment to the SEM and Auger Electron
Spectroscopy can be used to detect contaminants, especially
in cases when internal layer separations are observed.
Figure 4.36 shows how microsectioning was used to
ensure the integrity of multiple component replacement
activities and the effects of these operations on the quality of
the internal connections within the multilayer board to which
they had been assembled. Several components including this
capacitor were assembled and desoldered three times, using
the flight board test coupon. Checks were also made to
ensure that the capacitors internal solder (low melting
silver-loaded tinlead alloy) had not melted to produce either
open circuits due to lead separation, or short circuits due to
the presence of reflowed solder particles between the elec-
trodes. Figure 4.36b, c conrms the absence of cracked
platings and a good internal layer connection following the
thermal stressing. The reliability of inner layer junctions can
be assessed by submitting samples to thermal shock by
floating them on the surface of molten solder. This test is
recommended if solder joint reworking or extensive com-
ponent replacement is envisaged. The rework of assembled
multilayer boards can be a problem, and should be attempted
only as a last resort (Grey 1989). A further description of
158 4 Spacecraft ManufacturingFailure Prevention

Fig. 4.36 Composite photograph of capacitor A in microsection. (4) reflow of capacitor solder, (5) cracking of capacitor ceramic etc.
a This general view shows locations where detailed examinations were capacitor A. Rework has not degraded the multilayer board: b no
made (arrowed) to check for (1) delamination of pad lifting in mlb, cracked path barrels, c no internal separations at plated joints
(2) plated-through hole cracks, (3) internal connection separation,
4.7 Control of Printed Circuit Boards
reworking activities and their effect on PCB reliability will
be presented in Sect. 6.12.
A note must be made to readdress epoxy smearing. This
is the drill smear of epoxy resin which has been transferred
to the edges of copper foil in the holes of multilayer boards.
It has been seen both to be a uniform coating and to be
present in scattered patches. It is supposed to be totally
removed in the etchback process of mechanically and/or
chemically cleaning the hole prior to through-hole plating.
The epoxy smear is undesirable because it can electrically
isolate the conductive internal layers from the plated-through
hole. The epoxy smear can be caused by improper drill
feed-to-speed ratio, blunt drills, incomplete cure of the B
stage epoxy or smear induced by overfeeding the drill into
the backing plastic sheet. Almost any hole drilled in a
multilayer board may contain epoxy smears; it is therefore a
crucial task of the processor to control for each possible
cause. Smears are impossible to detect without resorting to
microsectioning (see, for instance, Fig. 4.35d) and even then
it can only be detected if the mount is very carefully pol-
ished. It is impossible to predict the number and location of
smears, and they may miss detection during the in-line
quality control steps associated with the test coupon.
An illustrated handbook containing the rationale and
procedure for the evaluation of multilayer PCBs has been
written by Jellison (1986). It places great emphasis on the
problems addressed in this section and can be used as a
practical aid for those in the space industry. The handbook,
prepared from NASA experience, is certainly still appli-
cable today, as it contains photomicrographs of
cross-sectioned holes to illustrate defect types, acceptable
and unacceptable conditions, and methods of measurement
that are also similar to European PCB requirements
(ECSS-Q-ST-70-10).
Finally, attention is made to a collation of ber and par-
ticle contamination defects seen during the inspection and
testing of spacecraft pcbs (Heltzel 2014). His presentation
details failure mechanisms caused by electrochemistry, dis-
continuities in the laminate and discusses the sources of the
contamination and relates the problem to pcb design. The
included viewgraphs would make excellent Workmanship
Standards for pcb inspection and quality tasks.
4.7.4 Flexible Circuits
Electrical power lines and data transfer conductors are often
constructed with flexible circuitry. The range of construc-
tions for these spacecraft flexible circuits is as wide as the
variations in materials selected for rigid PCBs. The main
difference is that electronic components are very seldom
assembled onto flexible circuits and, in the rare event that
this is a special design feature, special qualication testing
159
will be required to assess the effect of vibration fatigue on
the soldered connections.
Copper is the most common conductor material. The
preferred type is rolled annealed (RA) copper, as this has the
greatest flexibility and a high resistance to mechanical fati-
gue cracking than the cheaper, electrodeposited copper.
When using RA copper it is important that electrical
designers are aware that the copper grain direction should
run parallel to the length of the flexing section of the circuit.
The thinner coppers are more flexible. They can be pur-
chased with the same nomenclature as is used for the
thickness of copper on rigid PCB substrates. This is stated in
ounces per square foot but can be related to copper thick-
ness; for example, 1 oz/ft2 equals a thickness of 35 m.
Other electrical conductors can be silver, nichrome, or
berylliumcopper alloy.
The most common spacecraft dielectric selected for
flexible circuits is polyimide (e.g. Kapton), as this has
excellent mechanical, electrical, and chemical properties.
A usual thickness is 50 m. Fluorinated ethylene propylene
(FEP) is selected for spacecraft microwave applications; this
is a thermoplastic with an upper temperature limit of 270 C.
However, it is extremely expensive and difcult to process.
Polyesters have high outgassing properties, and whilst they
are the most cost-effective commercial materials, they are
unsuitable for space use.
Flexible circuits are often used as the umbilical electrical
connection to mechanical devices that operate under elec-
trical power, particularly in the close proximity to hinged
subassemblies. They form part of heating devicesfor
instance, it is important that the titanium pressurized tanks
which contain hydrazine fuel are not permitted to fall in
temperature when the spacecraft passes into eclipsehere
nichrome elements are bonded inside the flexible circuit
which, in turn, can be adhesively bonded to the surface of
the titanium tank. Where rigid multilayer boards (Sect. 4.7.3)
are used in conjunction with flexible laminates, the con-
struction is called a flexi/rigid circuit. In this case the
flexible polyimide circuit passes into, and is sandwiched
between, the layers of the multilayer board. The flexi/rigid
circuits are used to provide a total package combining a
substrate to which electronic components can be conven-
tionally assembled by soldering and the flexible intercon-
nection which can pass on to other circuits or connectors that
may even be in a different plane to that of the mother board.
The flexible circuit should have an oxidation protection
layer such as gold-plating on the copper conductor lines.
Tinlead should be present on termination areas where
joining by soldering is envisaged. Fused tinlead is not
recommended, as the hot oil fusion process can cause the
flexible circuit to delaminate. When the flexible circuit is
located on the outside of a spacecraft designed to operate in
low Earth orbit, it is important to protect the polyimide
160 4 Spacecraft ManufacturingFailure Prevention

substrate from attack by atomic oxygen. This is suitably

achieved, for instance on solar array flexible substrates, by
painting the surface with a silicone coating such as DC
93500 (Kuchler 1995).

4.7.5 Hot-Air-Levelled Circuit Boards

A common space requirement pertaining to PCBs is for a

fused tinlead nish having a thickness of greater than 1 m
at the knees of plated-through holes and greater than 8 m
on termination surfaces. These objectives are met by elec-
troplating the tinlead and then fusing it in hot oil (as
described at the beginning of this section). An alternative
method of coating the copper conductors on a printed circuit
is to selectively cover the copper with a solder mask, leaving
only the solderable termination areas exposed. The board is
then dipped into a large bath of molten solder and then, as
the board is withdrawn, it is blasted with hot air knives
which remove excessive solder and unblock plated-through
holes. Experience has shown that whilst this solder coating
method is quick and relatively cheap, it has several draw-
backs for space hardware production lines. The main prob-
lem is that the solder thickness does not meet the required
specication limits. These hot-air-levelled coatings may
have an initially acceptable solderability, but they do
become nonsolderable after several months of storage in
clean room environments.
A spacecraft breadboard circuit was found to be impos-
sible to solder using standard practices and the application of
low-activity liquid flux. This board was submitted to a rotary
dip solderability test which resulted in large areas of
dewetting, as shown in Fig. 4.37a. Microsections determined
that much of the hot-air-levelled surface consisted of less Fig. 4.37 Defects associated with hot-air-levelled circuit boards.
a Illustrates thin, flat areas of tinlead coating and associated islands of
than 1 m of solder and that this was probably solely cop- dewetted solder. The outgassing of water vapour from certain
pertin intermetallic. The same dewetting aspect has been plated-through holes, b has resulted in the formation of large bubbles
evaluated by Lobley (1990) and Banks (1995), who attrib- of porosity during a solderability test (courtesy of IFE)
uted the problem to be due to oxidation of the exposed
intermetallic, the very low shelf life of such boards (several defects. Investigations found that the copper within the hole
weeks) suggesting a nonoptimized hot-air-levelling process. was discontinuous and that hot-air-levelling had caused the
The explanation for this dewetting originates from the time glassepoxy substrate to outgass when the board was sub-
when excess solder was being cleared out of the jected to the thermal shock of being plunged into the coating
plated-through holes by blowing hot air from the knives equipments bath of liquid solder. This same effect was
across the board surface. So much solder was removed from reproduced under the controlled conditions of a rotary dip
some areas of the board that only intermetallic remained. solderability test (see Fig. 4.37b). The outgassing is caused by
Even hot-air-levelled boards with optimized plating thick- a release of water from the epoxy material which, in turn,
nesses only have shelf lives of typically 12 months (Parquet forms bubbles of steam within the liquid solder. This effect
and Boggs 1995), so such nishes are quite unsuitable for can be minimized by preconditioning the boards at 125 C for
spacecraft boards, as these may be subjected to several years a minimum of 6 h. Similar preconditioning treatments are also
of assembly processes and modications prior to launch. recommended before the commencement of any soldering
The hot-air-levelling process is sometimes applied to process, particularly those involving machine soldering
boards that have a poor quality of through-hole copper plat- operations, as this will reduce the number of blow-holes and
ing. One board was noted to have many plated-through-hole porosity inside the resultant solder joints.
4.7 Control of Printed Circuit Boards
4.7.6 Solder Assembly of Component
Packages onto Multilayer Boards
with High Heat Capacity
A common problem encountered when hand-soldering
components onto thick, up to 24-layer boards, having
back- or ground-planes is that the solder does not flow from
the soldering side of the board to the top surface. It will not
be possible to only use a higher soldering temperature
(contravening the ECSS Q-ST-70-08 maximum of 360 C
for difcult boards) as this can easily cause delamination,
fracture of internal plated hole connections, pad lifting and
may exceed the recommended maximum soldering temper-
ature listed on a data-sheet. Wave soldering for these board
designs was successful on one project. It should be noted
that recent soldering standards have relaxed the visual
inspection requirements for such boards75 % hole ll with
25 % circumferential llets may now be permitted.
Soaking the board in an oven at 125 C and then sol-
dering is possible, but dangerous for the operator. A study at
one of the ESA training schools (Strachan 2000) attempted
three pre-heating methods: IR lamp; Hot plate with hot air
stream; and, the application of a hot air gun.
Results indicated the IR lamp (Martin HB02 Hot Beam)
gave the best resultsheating being on the component side.
This allowed the soldering side to reach 100 and 130 C on
the component side. No damage ensued, and the process was
comfortable and convenient. However, some component
bodies could shadow the joint being soldered.
The hot plate method (Pace HeatWave) gave good results
with solder side at 100 C and the component side at 140 C.
Some poor visibility of the soldering area was noticed.
The hot air system (Leister 7A1) gave a very fast heat up
rate which could lead to destructive temperatures being
reached. However, this system was best for rework and
visability.
4.8 Control of Composite Materials
4.8.1 MetalMatrix Composites for Space
Structures
The eld of composite materials is relatively new, with the
greatest expansion in growth occurring during the last dec-
ade. The impetus in growth has been to achieve high
strength-to-weight ratios in materials. An ever increasing
number of papers are being published on this subject, and it
may be quite frustrating for potential users of such materials
that, despite the vast amount of work that has been com-
pleted in research and development, there is still a striking
lack of technically and economically available materials for
161
commercial industry. This is particularly true of metalma-
trix composites which seem to be just emerging from the
research and development stage.
The word composite in composite materials means two
or more materials, called constituents, that are combined in
a certain conguration to form a new material with specic
desired properties. Usually, the strengthening phases are
whiskers or discontinuous bres embedded in a matrix. In
the utilization of composite materials a number of conditions
have to be met. These conditions are related to the properties
of both bre and matrix. Primarily, the bres are expected to
bear the tensile load without yielding. At the same time the
matrix should distribute the load evenly over the bres,
protect their surfaces from oxidation, and diminish any
possible crack propagation.
Although such research has been done on compounds
such as SiO2, Al2O3, and SiC, the main interest to day seems
to be focused on graphite and boron bre with strengths
approaching their theoretical values of 210 GN m2.
If we turn our attention to the matrix material it will be
clear that although it is not required to bear the load, the
matrix must possess the following minimum properties:
a good bondability to the bre,
a compatible coefcient of expansion to avoid thermal
fatigue problem,
a protective layer against corrosion, and
suitable elasticity and toughness.
The bondability is necessary to transmit and redistribute
shear stresses, as a result of loading, to all active bres. It is
also important that there is minimal reaction between matrix
and bres, thus avoiding the growth of brittle intermediate
phases, cracks, or internal residual stresses. Unfortunately, it
is found that structurally perfect, strong interfaces are dif-
cult to form.
The practical problems associated with incorporating
reinforcing whiskers into a metal matrix are shown in
Fig. 4.38a. These aspects were addressed by Calow (1974)
with particular reference to the composites Ni/Al2O3 for use
at elevated temperatures and Al/C, Al/SiC, and Cu/C for use
at ambient temperatures. It would seem that the predictions
from this early work are still relevant today in that the future
for structural composite materials will rest entirely on eco-
nomic factors, dependent on lowering the cost of producing
bre materials and also for thermally cycling environments,
achieving a more permanent interfacial bond between bre
and matrix.
Much work has been made into investigations for
replacing the very heavy high-temperature nickel-based
superalloys by SiC bre reinforced titanium alloy Ti6A14V.
The bres are initially coated with the Ti6A14V-alloy matrix
162 4 Spacecraft ManufacturingFailure Prevention

Fig. 4.38 a Problems associated with achieving optimal composite pushed-in appearance. d 100 m diameter bre without a protective
material strength. b General view of 100 m diameter SiC bres coating showing a high (780 MPa) shear strength. (Reproduced by kind
embedded in a matrix of Ti6A14V. c 140 m diameter bres with permission from Dr. Hemptenmacher, Deutsche Forschungsanstalt ir
protective carbon coating after macroindented test. Pushed-out and Luft-und Raumfahrt e.V., Linder Hhe)
4.8 Control of Composite Materials
material by condensation from the gas phase (PVD). The
coated bres are then bundled, encapsulated, evacuated, and
hot-isostatically pressed at a temperature of about 900 C
and a pressure of 1900 bar. The material composite is
formed by the Ti-matrix flowing into the initial cavities
between the coated bres. The porosity-free composite is
conrmed by metallography as shown in Fig. 4.38b. The
shear strength between the bres and the matrix of this
MMC was determined by Hemptenmacher et al. (1994) us-
ing a method developed by Grande et al. (1988). Here,
individual bres were pushed out of a flat parallel test-piece
using a hardness test indenter. The shear strength is deter-
mined by the force required for bre debonding, divided by
the surface area of the cylindrical bre under test. A bre
with a protective carbon coating has a far lower bond
strength to the matrix than a bre with no coating, as seen in
Fig. 4.38c, d. However, as noted by Hemptenmachter et al.,
the shear strength depends also on the temperatures to which
the composite has been exposed. A special review of the
interfacial phenomena in MMCs, how they can be examined,
and how they affect performance, has been prepared by Feest
(1994). Ti6A14V/SiC MMC with 32 % volume fraction SiC
bre has a tensile strength of 1450 MPa parallel to the bres
and 450 MPa transverse to the bres.
One of the most promising metalmatrix composites is
the newly developed graphite bre/magnesium composite,
which offers great potential in terms of high strength and
stiffness combined with light weight (Goddard 1984). A new
graphite coating system that allows molten magnesium to
wet the bres has been developed. This means that it is
possible to lay up the coated graphite bres in a mould of
any desired conguration, and then cast molten magnesium
onto them. The result is a metal composite cast to nearly the
net shape which can be subsequently machined by conven-
tional tooling practice. In addition, as the composite follows
the rule-of-mixtures behaviour, each casting can be tailored
for a particular application. This is demonstrated by con-
sidering an article that requires a good longitudinal strength
and a reasonable transverse strengthin this case the casting
could be made up of 30 % bres (by volume) oriented
longitudinally and 10 % oriented transversely. Of particular
interest to spacecraft designers is the fact that this material
possesses zero, or near-zero, thermal expansion over a wide
range of temperatures. Special bres giving more options in
strength, stiffness, light weight, and endurance against heat
have been developed (Klein 1988). They range in diameter
from 150 m to 2 nm.
An alternative method for producing carbonmagnesium
MMCs for lightweight space structures has been developed
at Aerospatiale (Cornec and Clariou 1994), which is based
on an organic prepreg route (solid densication) or on a
liquid phase inltration route. A wide range of mechanical
163
properties can be achieved depending on the specic lay-up
design. For unidirectional plates the room temperature ten-
sile strength can reach 1000 MPa with a tensile modulus of
300 GPa. For thick plate with 0/45/90 angle plies the
strength is 240 MPa and modulus 120 GPa. These carbon
magnesium MMCs have a coefcient of expansion ranging
from 0 to 4 ppm, dependent on the bre lay-up direction.
Such MMCs are suitable for applications where a low
expansion is essential (e.g. optical bench) but they will
require suitable corrosion-resistant surfaces (see Sect. 4.5.1).
Large space structure prototypes have been constructed in
the USA incorporating graphitemagnesium struts and t-
tings (Paprock 1988); full metal inltration was demon-
strated by metallography.
Silicon carbide particles are also used to reinforce mag-
nesium and magnesium alloy composites in order to achieve
low density and high stiffness-to-weight ratio products for
applications such as aerospace equipment. These MMC
composites may be fabricated by both powder metallurgy
and ingot metallurgy routes (Krishnadev 1993), but it is
found that the former route results in material having the best
properties for the Mg/10 vol.% SiC composites.
Low-expansion materials can also be produced from
aluminium MMCs. A special gas pressure inltration pro-
cess has been developed by AMAGFVA in Austria for the
production of continuous-bre reinforced mouldings for
high-strength and thermal stability. These aluminium MMCs
are comparable to those described by Cornec and Clariou
(1994) and mentioned above for magnesiumMMC. How-
ever, AlMMCs have a greatly improved surface-corrosion
resistance. A wide variety of bres and aluminium compo-
sitions are produced by AMAGFVA. A typical plate pro-
duct is composed of 60 % high-modulus carbon bres in a
matrix of 40 % high-purity aluminium (Keramol
1085/C-HM/60f). This plate material has a tensile strength at
room temperature of 360 MPa, with a tensile modulus of
95 GPa. The tensile strength and modulus can be increased
by designing products having a thinner cross-section (the
so-called bre paradox). This MMC has a coefcient of
thermal expansion of 0.4 ppm and a density of 2.2 g/cm3.
The revolutionary concept of metal-matrix composites,
combined with their complex requirements for structural
applications (particularly for aircraft and spacecraft), pre-
sents a formidable uncertainty impeding acceptance by
design engineers. Some problems are highlighted in
Figs. 4.38 and 4.39. Because of this, designers of today will
prefer to choose structural and mechanical parts from tra-
ditional metal alloys, and will limit manufacturing to joining
and nishing technologies which already exist in their
respective plants. Whenever MMCs are selected it is
advisable to discuss the material requirements face to face
with the producer.
164
Fig. 4.39 Micrographs depicting Avco SiC monolaments within an
AA 2014 matrix. This metal matrix composite (MMC) has been
recently identied as a potential advanced spacecraft structural
material; it is produced by multistage processing. A carbon core, made
from a polymer precursor, is coated with SiC by chemical vapour
deposition initially as a carbon-rich deposit and then as a columnar
deposit possessing grains of preferred orientation and a stoichiometric
composition. The SiC monolaments are coated with a carbon coating:
the laments are then wound onto AA 2014 foil. Additional aluminium
is added to the foil by plasma spraying. Hot-pressed layers of these foils
4.8.2 Composite Contact Devices
The insertion of bres of a high-melting point metal into a
matrix of a highly conductive metal for the purpose of
producing contact materials was rst described by J.K. Ely
in a patent of 1942. Fibres were of tungsten or molybdenum
in a matrix of a coppersilver alloy. The bres imparted
mechanical strength, reduced mechanical wear, and pre-
vented particles of the contact from crumbling away. Similar
contact materials are still being employed even in spacecraft
electronic systems of today. They unfortunately suffer from
the phenomenon of metal migration, i.e. silver transfer
4 Spacecraft ManufacturingFailure Prevention
produce a compact MMC plate (by courtesy of the Metals Technology
centre, Harwell Laboratory, UK). a General optical micrograph
showing the etched AA 2014 matrix structure of the MMC plate.
The bres possess a graphite core surrounded by silicon carbide and a
thin layer of carbon. b TEM view showing internal structure of SiC at
the transition between the inner carbon-rich and the outer stoichiomet-
ric SiC regions (at point arrowed on the optical micrograph). c SEM
view of a tensile fracture face showing lament pull-out (minus the
carbon layer), and a predominantly ductile matrix failure
between electrical contacts. Typically, metal will transfer
from the anode to the cathode in a d.c. circuit relay after
several thousands of switching operations. This occurrence
becomes serious when microscopic particles of detached
deposits break loose within spacecraft devices. They can
cause short-circuits between adjacent conductor paths.
Typical particles from worn contacts are shown in Fig. 4.40.
Alternative metalmatrix composite materials are being
considered for advanced contacts. The silvernickel com-
posite and the copperpalladium system appear to be suit-
able; they have been developed and tested in Germany
(Kocker and Stockel 1979) and are noted to possess a high
4.8 Control of Composite Materials
Fig. 4.40 Morphology of wear particles generated from sliding
contacts (elongated particles of MOS2 set in a silver alloy matrix).
They can cause electrical short circuits under zero-gravity conditions.
a Silver alloy debris. Most of the particles are scale-shaped; some are
fragmented deformation tongues. All are typical of sliding wear.
b Detail showing laminated structure and shear fracture on this wear
particle
thermal conductivity, low electrical resistivity, a resistance
to the material-transfer phenomenon, and a particularly good
formability. It is interesting to note that the relationship
between the physical properties (including electrical con-
ductivity) and the composition of the alloys of AgNi and
CuPd varies according to whether they exist in homoge-
neous or heterogeneous states. Where the components are
completely mutually soluble in the solid state there is a clear
conductivity minimum produced by the electrical interaction
of the two statistically distributed nuclides, and this has been
demonstrated for the copper-palladium alloy system. The
phase diagram of the CuPd system, together with a plot of
conductivity versus composition, is shown in Fig. 4.41. The
165
Fig. 4.41 Phase diagram and electrical conductivity for the copper
palladium system and examples of industrial composite contact rivets
(courtesy of G. RAU GmbH & Co., Pforzheim)
articially produced CuPd composite shows a linear
relationship between electrical conductivity and composi-
tion. Here, the whiskers of Pd are manufactured by plating
thin wires of Pd with Cu and drawing them down to a 99 %
reduced cross-sectional area. The microscopic structure of
circular rods of Pd embedded in a copper matrix appears as a
completely insoluble duplex alloy, such as copper and
tungsten, where conductivity versus composition is descri-
bed by the law of mixtures (i.e. a straight line). On the other
hand, Fig. 4.41 also shows that the real alloy undergoes a
distinct drop in copper conductivity as the palladium content
166
Fig. 4.41 (continued)
is increased in the solid solution. There is some increase at
60 % Pd if the alloy is transformed to the ordered state.
The results of extensive testing have shown that contacts
made of the 30 % by weight palladiumcopper metal matrix
composite produce the most suitable contact materials. In
addition to the aforementioned properties, Fig. 4.41
demonstrates that this alloy has the same conductivity as the
homogeneous alloy 60Pd/40Cu, and, even considering the
extra processing steps, it is appreciably cheaper.
More modern spacecraft d.c. motors incorporate
spring-loaded brushes which slide against berylliumcopper
alloy commutators. The brush materials vary in composition
but are usually around 50 % copper and 49 % carbon together
with a small amount of adjuvant (lubricant) such as molyb-
denum disulphide or barium fluoride. The presence of a
lubricant, usually molybdenum disulphide, is essential if
brush operation is performed in vacuum. In damp air, the
water vapour promotes sliding between the graphite (carbon)
plates. Under vacuum, or in dry nitrogen, this mechanism does
not exist and it is found that the wear rate of graphite increases
by a factor of 1000 with a friction increase of 4 (Buckley
1981); for this reason, MoS2 should be present in the carbon
containing brush composite materials for all space motors.
4.8.3 Fibre-Reinforced Plastic Composites
Spacecraft have continued to make increased use of ther-
mosetting and, to a much lesser extent, thermoplastic
bre-based composites to reduce the weight and increase the
stiffness of structural components. Such materials as unidi-
rectional woven carbon bre (graphite), Kevlar, and glass bre
are procured in roll form and cut by special machines, often to
the geometry of the parts developed in a CAD system.
Guidelines on such composite applications are now well
established, although important aspects such as design
4 Spacecraft ManufacturingFailure Prevention
allowables, the signicance of defects, damage tolerance, and
failure criteria need further evaluation. Up to now, this has not
been a major problem, but it will become important for man-
ned missions, particularly when the different coefcients of
expansion and stiffness of bres and matrices induce microc-
racking in the material under thermal cycling.
Carbon-bre-reinforced thermosetting resins nd the
greatest number of composite applications. The large
payload-support structure for microwave instruments on the
ERS-1 spacecraft is constructed from CFRP (Fig. 4.42).
Both the primary structure of the Shuttle Pallet Satellite
(SPAS-01) and the European Retrievable Carrier are formed
essentially of lament-wound carbon epoxy tubes (e.g.
Toray M 40 and Araldite 209/HT 972) adhesively bonded to
titanium ttings to form a framework-type structure. Launch
vehicles also incorporate CFRP in certain locations, such as
the Ariane-4 Interstage-2/3 stiffened skin (Fig. 4.43) and part
of this vehicles SPELDA structure (one of the largest CFRP
structures to be built in Europe).
Space usage of carbon bres is based on polyacrylonitrile
(PAN) precursor bre material. The PAN bres undergo a
series of well-dened, rather complex processes. These
include a pre-treatment to stretch the bres in steam, fol-
lowed by further stretching with oxidation in air at 300 C,
carbonization in an inert gas such as N2 at about 1200 C to
produce type III bres (low heat treatment), or in Ar up to
1700 C to produce type II intermediate heat treatment
(IHT) bres. A further graphitization treatment is sometimes
given up to 2800 C in an inert gas to produce type I bres
(high heat treatmentHHT). Type I bres are associated
with possessing a high modulus up to 600 GPa, Type II with
high-strength, and Type III with low modulus and low
strength. Generally, Type I high modulus bres are selected
for satellite and space probe applications, and Type II high
strength bres for launchers.
Continuous-bre-reinforced thermoplastic matrix com-
posites include systems favoured for high-temperature
applications, as they have excellent mechanical property
proles and glass transition temperatures above 200 C.
These so-called aromatic polymer composites include
bre-reinforced PEEK (Dauphin 1986).
Kevlar-49 Reinforced Plastic (KRP) and polyurethane
shock absorbing foam in a sandwich conguration has been
successfully used for the rear shield of the Giotto
dust-protection system (the outer and inner shields being
made of aluminium alloy sheet). This advanced material
sandwich, seen in Figs. 3.11b and 4.44, suitably protected the
Giotto spacecraft from the impact of hypervelocity mete-
oroids travelling at 60 km/s. KRP is also used in the fabri-
cation of solid apogee-boost motors, known as MAGE
motors, for European satellites. The Kevlar-reinforced case
developed by MAN (Germany) is a good example of
lament-winding technology (Fig. 4.44). Similar
4.8 Control of Composite Materials
Fig. 4.42 Set-up for vibration
testing of the ERS-1 spacecraft
(the payload-support structures
are arrowed) (ESA)
apogee-boost motors have already placed seven satellites in
their required orbital positions.
Special surface pretreatments of graphite bres have
resulted in exceptionally high bond strengths with epoxy
matrices. Such composites have been successfully fabricated
into complex spacecraft antenna face-skins that must survive
in excess of 10,000 thermal cycles between +100 and 150 C
during a ten-year life in orbit around the Earth. Despite the
superior properties of such materials the author has found that
the human factor in the physical lay-up of such structures can
167
cause low bond strength between lay-ups, so that low buckling
or twisting forces will easily delaminate this composite
material. As with all composites, full quality assurance
surveillance, including the testing of standard-in-line samples,
must be made during all steps of manufacture, and similar
samples may be submitted to thermal cycling tests (see Fig. 3.
14b). The ultra-high-modulus bres used in spacecraft have
been found to possess a considerable variation in properties
from one batch to another, so that particular care is needed in
the interpretation of the results of materials tests.
168
Fig. 4.43 Ariane-4 interstage
skins (photo courtesy of Fokker
BY)
The environmental effects on spacecraft components con-
structed of composites have been reviewed (King 1988).
Atomic oxygen has been seen to cause carbon breepoxy
systems to lose weight by surface erosion, whereas Teflon and
silicones are almost unaffected. Structures manufactured of
CFRP systems in low Earth orbit will require some form of
protection, as atomic oxygen has been calculated to erode
surfaces to a depth of 3.3 mm during a 30-year period,
thickness loss being made up of both bre and matrix degra-
dation. Coatings of metals and oxides afford some protection.
Metallized CFRP waveguides have proved to be enor-
mously successful on both the ERS 1 and the ERS 2
spacecraft. This technology is sure to nd increased usage
during the next millennium when low attenuation and phase
stable power transmission requirements will be more
demanding. Typical commercial spacecraft waveguides are
4 Spacecraft ManufacturingFailure Prevention
made of either brass (1900 g/m length of waveguide) or
aluminium alloys (600 g/m length). The developments made
in Germany with thin-walled lightweight (less than 100 g/m
length) CFRP waveguides metallized on their internal sur-
faces with either copper, silver, or gold have also led to a
similar fabrication line for complex components having
different shapes, such as those needed for choke flange
connections, couplers, and bends (Wagner 1992). For
straight waveguides, mechanically removable and reusable
mandrels are selected. The mandrel material is secret, but it
can be machined and polished to a very high accuracy. This
surface is then plated with one or two microns of copper,
silver, or gold. An adhesion treatment is applied and then
carbon bres are wound around the plated layer. Special
winding directions and layer congurations have been
developed to avoid bending, during thermal cycling service
4.8 Control of Composite Materials
Fig. 4.44 The MAGE motor
being integrated into the Giotto
spacecraft. The rear shield of the
dust-protection system is made of
Kevlar-49 reinforced plastic and
polyurethane shock-absorbing
foam (ESA)
operation, due to mismatches between the expansion of the
metallization and that of the carbon bres. For complex
shapes, sacricial mandrels are employed; after curing of the
CFRP part they are chemically removed. These parts are
subjected to many of the quality assurance activities
described in Chap. 3 of this book. These include tests on
samples manufactured in-line with flight parts, such as
microsectioning (to establish complete densication of the
laminate and the thickness of the metallization); peel
strength tests to check the adhesion between metallization
169
and CFRP; torsion mechanical analysis (TMA) to determine
the glass transition temperature and ensure complete curing
of the laminate; X-ray inspection to determine faults in the
metallization; and, ultrasonic C-scan inspection of the CFRP
laminate for voids and delaminations. The qualication
testing of these waveguide constructions included more than
10,000 thermal cycles, in the temperature range 70 to
+115 C, without any degradation of peel strength or elec-
trical attenuation. In orbit they have performed faultlessly for
more than six years.
170
The CFRP processing advances are very important for
construction of space structures having a high precision. The
antenna reflector for the FIRST project will consist of a 3 m
diameter parabolic antenna with a focal distance of 1.58 m and
a surface accuracy of 7 m RMS (root mean square). It will be
made totally of CFRP, comprising a sandwich core and two
face-skins of high stiffness and low weight (Ehmann 1994).
One of the most important factors to understand when
employing CFRP is the effects of a humid environment
(moisture) on the dimensional stability, strength, and stiff-
ness of the laminated composite. Moisture will diffuse
through the resin of the CFRP regardless of the resin type.
The diffusion process is similar for both thermoplastic and
thermoset resinmatrix composites (Lucas and Zhou 1993).
Some dimensional changes can be expected due to swelling
of the resin; this is not detectable in the direction of the
bres, but may be extensive in the width and thickness
directions. It is possible to optimize the winding directions
of carbon bres so that the net effect of expansion of the
resin matrix will have no influence on the nal structure.
This has, in part, been demonstrated by Hammesfahr et al.
(1995) in association with the considerations for an X-ray
mirror assembly for the XMM project which was based on
carbon-bre technology. Initially the CFRP solution was
preferred for the construction of the 58 Wolter type cylin-
drical mirrors, their supporting structure, and a thermal
baffle. In 1996 the decision to utilize nickel mirrors, even
though they are much heavier (320 kg vs. 137 kg for CFRP),
appears to be due to manufacturing difculties caused by the
presence of moisture.
4.8.4 Fibre-Reinforced Glass Ceramics
Glass ceramics are formed in two main processes. Initially, the
constituents of the glass are melted then press-formed into
shapes. This is followed by heat treatments that nucleate cer-
tain phases in the glass microstructure and cause crystalliza-
tion. The most common glass ceramic systems are
lithia-alumino-silicate (LAS), zinc-alumino-silicate (ZAS),
and magnesia-alumino-silicate (MAS). They are used to make
radomes, electronic devices and the blanks for mirrors. An
excellent overview has been presented by Hanson and Fernie
(1993). Important characteristics of the glass ceramics are their
ability to form strong ceramic-to-metal bonds, their
tailor-made coefcients of expansion (020 ppm), a high
strength, and their high softening temperatures (up to 1400 C).
A glass ceramic matrix which combines the oxides of
magnesium, aluminium, silicon, and lithium named MAS-L
has recently been processed. It was selected not only for its
high elastic modulus, low coefcient of expansion, and up to
1000 C creep strength, but also for its good compatibility
and bonding to silicon carbide Nicalon bres. Several
4 Spacecraft ManufacturingFailure Prevention
constructions of SiC/MAS-L have been fabricated and later
characterized (Peres et al. 1994). Unitape plies are made by
winding (solvent, resin, glass powder) slurry-impregnated
SiC yarn onto a mandrel. The plies are stacked into a furnace
and the resin binder is burnt out. Hot-pressing at a high
temperature permits the glass ceramic to flow into the inter-
stices between individual bres to form a composite densied
up to 98 % of the theoretical density. The microstructure of
the composite sheets shows a 2-D pattern of seven layers
where the bre plies are oriented at 0 and 90. A few
residual pores are visible within the as-manufactured sheets.
1000 thermal cycles between 100 and +100 C cause a
small amount of microcracks which run parallel to the bre
lay. Thermal shocks from 1000 C into water at 20 C cause
microcracks that run at a 45 inclination to the bre lay.
Thermal cycling caused a slight reduction in room tempera-
ture tensile strength of the composite from 179 MPa down to
165 MPa. Thermal shock was more dramatic in reducing this
value to 68 MPa. However, in compression and in bending,
the strength decreases caused by thermal cycling and thermal
shock were not so great, particularly when the mechanical
testing was made at elevated temperatures up to 1000 C. It
was concluded that this new class of material, with its very
low coefcient of expansion, should be considered for
high-stability structural applications where thermal cycling
might be present between 100 and +100 C.
4.8.5 CarbonCarbon Composites
Carboncarbon composites are employed almost exclusively
in high-temperature applications. Occasionally, very thin
carboncarbon face-skins are selected and bonded onto
honeycomb spacers for lightweight panel applications. The
main uses are associated with propulsion systems and
re-entry vehicles. During the combustion of the solid chem-
ical propellants (described in Sect. 2.4) high pressures are
generated together with the liberation of large quantities of
energy in a small volume. The subsequent expansion of these
high-temperature products of combustion through a nozzle
converts thermal energy into directed kinetic energy for
spacecraft propulsion. Woven carboncarbon composites
have been found to be ideal materials for the construction of
rocket motor nozzles as they have a low density, a high
strength at operational temperatures, and little chemical
reaction with the exhaust gases during propellant burn time.
Nozzles for the Ariane launch vehicles are shown in Figs. 2.7
and 2.11. A similar external view of a solid propellant apogee
motor is shown in Fig. 4.44. Carboncarbon is also an
exceptionally good material for thermal protection systems
and is selected for the highest temperature regions of re-entry,
vehicles, the nose cones and leading sections of wings (see
Figs. 2.14 and 3.4).
4.8 Control of Composite Materials
The fabrication route for carboncarbon composite parts
is complex and involves many variable parameters. For this
reason it is useful to follow production steps by means of
microsectional controls. In-line control samples can be cut
from the product at different stages during its production.
Fig. 4.45 Propellant tanks manufactured from Ti6Al4V; spun domes
are diffusion bonded to SiC bre-reinforced titanium matrix composite
tubular lengths, a before and, b after burst testing. c Tungsten cored
silicon carbide monolament selected for the tubular lengths showing
darker laminar carbon coating interface to the titanium alloy matrix.
171
Tight processing parameters must be maintained and the
end-product should contain sufcient offcuts for mechanical
performance tests.
A carboncarbon composite consists of two phases of the
element carbon, a carbonaceous matrix that is reinforced by
d Fractograph SEM image showing angle ply lay-up of SiC monol-
aments in an aluminium matrix (this is a high magnication fracture
surface of an angle ply 100 m bre in aluminium matrix, an
alternative choice of material to the titanium matrix seen in (c) (Cour-
tesy of TISIC, Farnborough)
172
carbon (or graphite) bres. The carbon bres may be woven
and physically formed into the required shape of the
piece-part. For low-cost products, where little mechanical
loading will be encountered, nonwoven bres are used. The
carbon bres are impregnated with a polymeric resin which
is then cured. The composite is then subjected to its rst
carbonization, when it is heat-treated to temperatures
between 500 and 1000 C, and in this stage the polymer
matrix is pyrolysed to form the secondary carbon phase,
which is porous and may contain transverse matrix cracking.
Microsectioning of control samples, together with nonde-
structive testing of the remaining material, will establish if
defects such as voids and microcracks are in sufcient
numbers to render the material as being unsuitable for fur-
ther processing, i.e. workmanship standards would identify
rejectable situations. This is an important step, as further
proprietary processing operations that involve a second
heat-treatment can cause the isolated defects to develop into
gross delaminations which cannot be densied by further
processing.
4.8.6 Metal Matrix Composites for Spacecraft
Pressure Vessels
Metal matrix composites for structural applications were
described in Sect. 4.8.1. The development of bre reinforced
pressure vessels and propellant tanks by TISICS Ltd.,
Farnborough started with a xenon gas tank design based on
requirements for the SSTL Tech-Demo-Sat-1 satellite.
TISICS is the only commercial plant outside of the US
producing CVD (chemical vapour deposition) silicon car-
bide monolaments. Pressure vessels were designed and
fabricated with their own SiC bre-reinforced titanium
matrix composite in the tubular region. These tanks (see
Fig. 4.45) were tested and passed against flight release test
schedules. Testing to burst at greater than 4-times the
maximum expected operating pressure resulted in a rupture
failure similar to a metallic rather than composite failure
mode. Quality controls included metallography of in-line
test samples and X-ray tomography of the composite region
and the interface-to-monolithic titanium diffusion bond.
Further work is underway to develop thin wall titanium and
aluminium composite propellant and pressure tanks for the
new green propellants (Fig. 4.45c, d). Fibre performance
and increased production efciency has been achieved for
both the 100 m bres required for thin wall angle ply
systems, and the stronger and stiffer 140 m bres for most
other applications.
4 Spacecraft ManufacturingFailure Prevention
4.9 Control of Capillary Screens
Present spacecraft liquid propulsion systems concentrate on
monopropellant hydrazine for in-orbit propulsion. These
systems use catalytic and electrothermal decomposition
techniques that convert liquid hydrazine into gaseous
hydrogen and ammonia. Short bursts of these gases from
small thruster motors enable spacecraft manoeuvres by
three-axis attitude control. The large titanium vessels that
contain liquid hydrazine fuel onboard satellite presently
incorporate a rubber bladder that is pressurized by a gas, so
forcing the fuel to flow to the thruster motors. However, such
organic membranes rupture after relatively few expulsion
cycles and can be dissolved to cause contamination of the
hydrazine fuel as described in Sect. 5.8. A new flow system
should make obsolete the perishable bladderit is known as
the surface-tension tank. See also Sects. 5.8 and 5.20.
Spherical surface-tension tanks manufactured in Europe
are flying on Frances Spot spacecraft, and have been pro-
cured in the USA for Olympus. These tanks are fabricated
from double-melt Ti6Al4V billets that are forged and then
machined into thin-walled hemispheres. Each tank has a
Propellant Management Device (PMD) which enables
gas-free liquid propellants to be fed to the engines (e.g.
thrusters). The PMD comprises an internal assembly of tubes,
vanes, and ne-mesh screens which, by capillary action,
perform the propellant separation and gas-free feeding.
Once the PMD is incorporated into the tank, the two
hemispheres are either TIG-welded or EB-welded along their
equatorial girth. Present PMD capillary screens are made of
304L stainless-steel mesh, woven to provide 5 m diameter
interstices. To achieve greater weight savings, the PMDs of
the future will be constructed of titanium alloy (only a few
materials, including titanium and stainless steel, are fully
compatible with the hydrazine propellant), and efforts are
being made to acquire titanium capillary screens. Recent
powder metallurgy techniques have produced a Ti6A14V
alloy with a controlled porosity level of around 40 % by using
a combination of high-temperature and high-argon backll
pressure. This could nd applications within PMD designs.
The pores have a cubic morphology (Fig. 4.46a) and can be
made either interconnecting or ne and discrete
(non-connecting). Although in their infancy, titanium screens
made from 5 m diameter bres (Fig. 4.46b) are being
evaluated. Titanium wires are initially covered with protec-
tive alloy sheaths and drawn through dies until a minimum
diameter is reached. The drawn wires are annealed by heat
treatment, bundled together again, and redrawn. These steps
are repeated until the desired nal bre dimension is
4.9 Control of Capillary Screens
Fig. 4.46 a Material being studied for PMD applications. Microstruc-
ture of Ti6A14V porous material (courtesy of IMI Ltd.). b Mat of
titanium bres for PMD capillary screens (see also Fig. 5.87)
achieved. The last step is to remove the sheath material by
chemical leaching, which results in a bundle of whisker-like
strands that can be built up into a screen (US Patent 3277564).
Both SEM examination and metallography are required to
establish the actual mesh size of these capillary screens.
4.10 Examination of Electroless Nickel
Deposits
4.10.1 Microcracked Electroless Nickel
Several problems associated with the cracking of electroless
nickel deposits on spacecraft components have been evalu-
ated. Common defects have resulted from a number of
factors:
173
Fig. 4.47 Problems associated with electroless nickel platings. a Micro-
section to illustrate the lack of adhesion and cracking of EN-to-brass
interface on part of a failed terminal pin. b Crazing and microcracking of
diode lead surface; here, the Kovar lead material has been overplated
with copper, electroless nickel, and a gold flash. The diodes were being
assembled onto flight electronics when operators noted that lead forming
prior to solder dipping caused the brittle EN to flake off
Bad machining of the substrate; nickel had deposited on
burrs which later became detached to form particle
contamination of the mechanical system.
Incomplete cleaning pretreatment prior to electroless
nickel (EN) deposition; crimp buckets and PCB terminal
pins suffered from poor nickel adhesion resulting in
electrical open circuits (Fig. 4.47a shows the detachment
of a solder-covered terminal pin along the EN interface).
Voids and thinning of EN deposits leading to corrosion
of aluminium alloy; inclusions and mixed grain structure
were thought to be contributors.
Cracking of EN deposits on the leads of electronic
packages resulting from hard deposits followed by
lead-bending operations; this is a frequent problem
(Fig. 4.47b).
174
Fig. 4.48 SEM view of cracked knee on bent package lead: a general
appearance; b the base Kovar is exposed beneath the cracked
gold-plated EN
The following case history concerns cracked leads on a
thick-lm carrier package. The Kovar leads had been
glass-to-metal sealed into the package. They were then EN
coated, gold plated, and stress relieved by heat treatment at
200 C. After lead bending, inspection revealed that all outer
bend radii contained cracks (see Fig. 4.48). One row of leads
was de-golded and pretinned with eutectic tinlead solder in
an unsuccessful attempt to cover the cracks (see Fig. 4.49)
and minimize the chance of causing failure by stress corro-
sion (see also Sect. 6.9.3) or mechanical fatigue at spacecraft
vibration levels. The damaged leads were microsectioned,
and cracking was seen to be limited to the plating thick-
nesses. This is highlighted in Fig. 4.50; the 5 m thickness
gold nish is suitably ductile, but it is the 4.5 m thick EN
that has initiated the cracks. The EN was found to be hard,
possessing a microhardness value of 500 VPN, whereas the
Kovar lead material is an annealed condition (250 VPN).
4 Spacecraft ManufacturingFailure Prevention
Fig. 4.49 SEM view of de-golded and solder-tinned packaged lead:
a general appearance; b solder has not covered volume of crack
The ductility of EN is low; elongation is 0.51.5 % com-
pared with 1015 % for electroplated nickel (Tulsi 1986).
EN is actually an alloy of nickel-phosphorus. In the
as-deposited form, NiP deposits are amorphous, since
X-ray diffraction studies did not indicate the presence of any
crystalline structure. Such deposits are, in fact, supersatu-
rated solid solutions of phosphorus in a microcrystalline
(less than 100 ) nickel matrix (Puttlitz 1990). At 200 C the
precipitation of nickel phosphide (Ni3P) begins. The main
effect of precipitation is to harden the EN deposit, which
causes a further reduction in elongation. EN coatings are
readily soldered if mildly activated fluxes are used. How-
ever, no flux has been found to successfully solder deposits
which have been heat-treated. Phosphorus contents as low as
3.5 wt% do not preclude this mechanism. These EN deposits
are therefore considered unsuitable for spacecraft applica-
tions, and electrolytic nickel-barrel plating of these elec-
tronic packages is recommended.
4.10 Examination of Electroless Nickel Deposits
Fig. 4.50 a Detailed micrographs taken with oil immersion lens from
cracked regions shown in Fig. 4.48. Note (1) gold plate has good
ductility as evidenced by necking above crack; (2) nickel is extremely
brittle and shows no ductility, there being no evidence that it can
accommodate bending. b Transmission electron micrograph of gold
(black)- and nickel-plated layers. This highlights the crystal structure of
the phosphorus nickel layer which contains small nickel phosphide
(dark) precipitates. The white magnication bar is 50 nm. (TEM
courtesy of IVF, Sweden.) See also Sect. 6.6.2
4.10.2 Electroless Nickel Plating of Aluminium
Electronic Housings
The most usual method of applying plating to aluminium
alloy electronic housings consists of the use of zinc
immersion coating followed by copper strike, copper plate,
silver plate, and/or gold plate. There are some advantages in
the use of electroless nickel for plating of aluminium in
electronic equipment applications, and it has replaced the
conventional copper strike and copper plating owing to its:
improved adhesion to aluminium up to 350 C,
uniform thickness of coating in recessed places,
uniform deposition rate (e.g. up to 20 m per hour at
90 C), and
improved corrosion resistance on parts with surface
discontinuities.
175
The need to use plating xtures to deposit adequate
coating thicknesses in recessed areas of electronic equipment
is eliminated when electroless nickel is employed, and this
leads to cost and labour savings. However, it is essential to
clean and then pretreat the aluminium with a preliminary
zinc immersion coating before electroless nickel plating.
Electroless nickel is usually deposited from a solution con-
sisting of 30 g/l nickel chloride, 50 g/l sodium glycollate and
10 g/l sodium hypophosphide at an elevated temperature of
8090 C. The bath wall can be stainless steel or
PTFE-lined. The preferred rate of deposition, at 80 C and a
pH of 46, is about 610 m per hour. A good treatise on
this process is given by Schlesinger (2010).
Many spacecraft electronic housings have been machined
and dip braze assembled from alloys such as 6061 and 6063.
Some items contain difcult recessed areas, with critical
sharp edges, tapped tuning-screw holes, and small blind
holes which must be completely plated for corrosion pro-
tection and stable electrical characteristics. It was impossible
to conventionally electroplate these parts and provide com-
plete coverage in the recessed areas, so that electroless
nickel, having a thickness of 10 m, was used in combina-
tion with a silver nish plate.
Electroless nickel-plated aluminium electronic items are
found in all applications spacecraft. However, scientic
spacecraft have the additional characteristic to demand spe-
cial performances that are not normally found in other engi-
neering applications. Scientic experiments often require
spacecraft hardware to perform measurements in a wide
variety of orbital conditions, and great care must be taken to
ensure that there is no perturbative effect by the presence of
the instruments themselves, or by the vehicle which carries
them (Delahais 1985). When one unit of a recent scientic
satellite was submitted to subsystem magnetic tests it was
found to be signicantly magnetic. All materials had been
carefully selected to ensure a high degree of magnetic
cleanliness, so that the very small magnetic elds to be
measured in orbit would not be masked by ambient elds
introduced by the spacecraft. After considerable investigation
the problem was traced to the electroless plating process
applied to some electronic housings. Owing to the complexity
of the unit it was impossible to replace before launch. In this
case it is clear that the standard electroless nickel cannot be
utilized, and some extra controls will be necessary to ensure
that the plating is nonmagnetic. This can be done by adjusting
the nickel-to-phosphorus ratio of the electroless plating
solution, and, with tight limits on temperature and pH
deposits, a nonmagnetic composition can be guaranteed. Pure
nickel plating between 3.5 and 7.0 % phosphorus will have an
as-plated magnetic coercivity of from 30 down to 2 Oersteds.
As-plated deposits with 8 % phosphorus can be considered as
being nonmagnetic, but there can be an increase in magnetic
coercivity if such deposits are heat treated. As-plated deposits
176 4 Spacecraft ManufacturingFailure Prevention

containing 12 % or more phosphorus are said to retain their
nonmagnetic properties even after heat treatment at 300 C
(Tulsi 1986). Experience has shown that heat treatments at
200 C, as for hydrogen embrittlement relief, result in an
improved corrosion resistance. Heat treatments made over the
range 250400 C cause a signicant reduction in corrosion
resistance due to the precipitation of nickel phosphide. This
causes the EN deposit to harden and can promote microc-
racking because of anisotropic dimensional changes. Fig-
ure 4.50b illustrates the crystal structure of the phosphorus
nickel layer. Possibly brittle cleavage fracture occurs at the
boundaries of these columnar-shaped crystals. The internal
volume of these crystals appears to contain nano-sized pre-
cipitates of nickel phosphide. Differential thermal expansion
between the EN deposit and its substrate will further impair
deposit integrity. The problem related to black pad will be
further discussed in Sect. 6.6.2.
As mentioned above, heat treatment after plating is often
carried out; this ensures a good adhesion of nickel to sub-
strate, and aims to expel any hydrogen from the work-piece.
It has been proven to improve corrosion-resistance. Using
furnace treatments, the entire work-piece has to be heated to
the relatively high temperature of 200300 C. This process
may be damaging for the substrate, it may alter the substrates
heat-treatment condition, and some parts may be restricted
from entering a furnace due to their large size. A relatively
new process which can obviate the need for an oven or
furnace, is that of laser annealingthis readily achieves
nanocrystallisation of the electroless nickel, without affecting
the properties of its substrate material. A recent study into the
utilisation of lasers for modifying NiP and Ni-W-P elec-
troless coatings, in an attempt to increase the corrosion pro-
tective properties of these coatings, has been published by
Liu et al. (2014). Laser heat treatment of a Ni-3.9W-13.4P
(wt%) as-plated amorphous coating causes nanorecrystalli-
sation, and the precipitation of the Ni3P phase, this is a stable
compound in a corrosive medium. Corrosion (preferential
dissolution) of nickel occurs, leading to an enriched phos-
phorous surface layer and a surface coated with adsorbed
hypophosphite anions which blocks further hydration, or
corrosion of nickel. For the as-plated, amorphous coating,
this corrosion protection mechanism is weak. It is only after
heat treatment, with the formation of Ni3P, that the best
corrosion resistance is achieved. Such desirable nishes can
be attained (Liu et al. 2014) without damaging underlying
material, by the laser annealing/surface modication process.
4.11 Control of Electroforming Processes
A wide range of spacecraft piece-parts have been success-
fully fabricated by electro-forming. They include electro-
formed copper waveguides, CFRP waveguides supporting
silver-plated internal faces, microwave horns, and a

Table 4.12 Mandrel materials for electroforming

Material type Preparation for plating
Nonconductive
(a) Plastic The surface is rst cleaned then primed
by immersion into 10 % stannous
(b) Wax (e.g. Ceresin chloride. The mandrel is then made
which melts at 6070 conductive by silver reduction. The water, then clean internal
C) silvering operation takes 20 min at a
solution temperature of 21 C; the
solution contains Rochelle salt, silver
nitrate, and ammonia. The thin silver
mirror is further plated with copper and
other metals. Good throwing power
Electrically conductive
(c) Aluminium Clean in tribasic sodium phosphate, then
bright dip in phosphoric acid prior to
plating with copper (or zincate)
(d) Stainless steel Bright pickle in hydrochloric acid then
electroplate with copper
Electrically conductive
(e) Low-expansion Bright dip in 25 g oxalic acid, 13 g
metal (e.g. Invar, hydrogen peroxide, 0.1 g sulphuric acid,
Nilo-K) and 11 water solution. Then immediately
copper plate
(f) Low-melting alloys After machining, degrease and
(e.g. Woods alloy melts immediately plate
at 70 C; eutectic
solder at 183 C)
Method for parting
Soften by heat, then withdraw
plastic from the electroform
Wax may be melted out in hot
electroplating solution must not be at
surface with organic solvent
Aluminium and its alloys are
readily soluble in a hot
sodium hydroxide solution
Impact the electro form by a
sudden hammer blow, or push
mandrel out by a hydraulic
ram
Heat electroform with a torch
or hot oil bath to take
advantage of deposits larger
coefcient of expansion
These are melted and shaken
out; zinc alloys can be
dissolved in a hydrochloric
acid solution
Notes
Plastic is best for permanent moulds
Wax can be cast into metallic moulds;
elevated temperature due to high
expansion coefcient of wax and
softening (i.e. loss of dimensional
accuracy)
Electroformed parts can be hard (nickel
etc.) or soft (gold). Silver is difcult to
plate
Heating may help, or cooling in dry ice
(solid CO2) and acetone depending on
coefcients of expansion differences
Good for nickel-plated forms when
stress-free deposits are electroplated
from chloride-free sulphamate baths
Suitable for soft platings, but avoid
tinning of the plated internal surface by
graphiting mandrel
4.11 Control of Electroforming Processes 177

Fig. 4.51 a General view of microwave horn. b Showing electroformed internal surface (dark area of damaged castellations are arrowed)

combustion chamber containing a complicated system of

cooling channels for the Ariane launcher. These electroforms
have been produced by the electrodeposition of metallic
layers onto mandrels made of stainless steel, aluminium
alloys, low-melting alloys, plastics, and wax (Suchentruck
1986). Some processing details and experiences gained by
the author are presented in Table 4.12.
Highly polished mandrel nishes can provide the basis for
reflectors where the mandrel is made the mirror image of the
internal surface of the piece being fabricated. In this case
stainless steel is often used for the reusable mandrel which,
after plating, can be removed by tensile forces or compression.
Aluminium and low-melting-point alloys are mainly used
when complex shapes are to be made, such as those having
undercuts and varied internal dimensions. Because of these
shapes the mandrel can be removed only by destroying it
completely by chemical etching or melting. Some useful
electrically conductive waxes are available which can be
formed and integrated onto an existing metal structure.
These waxes have a good wettability in the electrolyte and
are not chemically attacked during the plating operation.
Upon heating they are easily removed from nished channel
systems. Fig. 4.52 Schematic drawing of cross-sectioned microwave horn.
a Electrodeposition of thin gold then thick copper onto aluminium
Waveguides can be fabricated by using disposable man-
mandrel to provide internal castellations. b Poor plating procedures
drels with long rectangular shapes. A metal having a good result in the incomplete lling of the aluminium channels. The
electrical conductivity such as silver is plated in a thin layer aluminium is then chemically removed by etching
178 4 Spacecraft ManufacturingFailure Prevention

Fig. 4.53 The internal

castellations have a height and
width of 0.14 mm. Owing to the
lack of copper to back up the thin
gold plate some castellation tear
to form dark areas. The initial
gold plating appears to have had a
reduced throwing power into the
aluminium cut-outs and with a
thick build-up at the corners there
is little chance to complete the
process by fully lling the
castellations with copper plating

onto the mandrel. The next step is to apply a thicker back-up
metallic layer to strengthen the waveguide. The back-up can
be solid copper or nickel, and this may later be machined to the
correct outer dimensions of the nished waveguide. Complex
multi-horn feeds for the British Orian satellite (Teare 1994)
could only be manufactured by electroforming, owing to the
complex shape of the horns and the stringent geometric
requirements. A one-piece aluminium mandrel was produced
using 5-axis numerical control milling machines. The mandrel
was then silver-plated following a proprietary pretreatment.
This was followed by nickel plating from a nickel sulphamate
electrolyte (an alternative plating solution based on nickel
sulphate/nickel chloride was unsuitable because it produced
nickel deposits having a high internal stress level). The
mandrel was removed by conventional dissolution in sodium
hydroxide without damaging either the internal silver or the
4.11 Control of Electroforming Processes
external nickel layers. Qualication testing on nished
multi-horn feeds involving thermal cycling and vibration
followed by PIM tests and RF tests revealed no degradation in
the electrical performance. Lightweight waveguides can also
be produced by using a somewhat thinner copper plating to
which a bre-reinforced composite system is wound. After
resin impregnation a curing cycle will harden the waveguide,
and the mandrel can then be removed by chemical etching or
high-temperature melting.
Metallographic control of the electroformed items may be
necessary to control the internal dimensions of complicated
parts or to account for production defects. Figure 4.51a
shows a general view of an electroformed microwave horn.
The internal castellations must have an extremely high
reproducibility, but, as highlighted in Fig. 4.51b, there are
dark, disturbed areas on the internal surface. The item was
microsectioned in order to assess the damaged region. This
microwave horn had been made onto an aluminium mandrel
as shown in Fig. 4.52. Owing to its complex shape the
mandrel could be removed only by destroying it completely
by etching in sodium hydroxide. Upon closer examination
the released castellations are seen (Fig. 4.53 in cross-section)
to contain voids resulting from the use of incorrect elec-
trolyte systems. Special-purpose electroplating baths were
purchased to achieve the desired solid geometry.
4.12 Dip Brazing of Aluminium Alloys
Dip brazing is used successfully in the manufacture of alu-
minium alloy waveguides and heat-exchangers for space-
craft applications. Components must be correctly held in
position by some form of jigging. Tabs and mating slots are
often used for self-jigging. 304 type stainless steel, Nimonic
75, and Inconel X-750 have been used as spring-held jigging
materials, being resistant to the corrosive environment dur-
ing both dip brazing and post-cleaning operations.
Dip brazing equipment comprises an electrically heated
forced air circulation furnace to preheat the assemblies to
500600 C. A ceramic-lined salt bath, heated electrically by
submerged electrodes of high-grade graphite paired across
the bath near the floor, gives circulation to the molten salts.
The typical analysis of a molten salt is sodium chloride
23.0 %, potassium chloride 47.5 %, lithium chloride 24.7 %,
and sodium fluoride 4.8 %. Care must be taken to avoid
contaminating the salt with iron, as this lowers the activity of
the salt and degrades the properties of the braze ller metal.
The waveguide assembly seen in Fig. 4.54 is made of
alloy 6061. It would completely fail the set of inspection
criteria for brazed waveguide flanges that are listed in
Sect. 6.10.7. Spot-welding at the corners of the waveguide
tube and the flange was an effective means for jigging. After
vapour degreasing, a brazing paste containing BAlSi-4 ller
179
Fig. 4.54 Side view of joint between alloy 6061 waveguide and flange
after vacuum exposure. Note severe salt contamination and corrosion
adjacent to dip brazed llet (between tack-welded corner points)
metal (aluminium-silicon, melting range 565600 C) was
applied to the outer side of the braze path. The assembly was
then preheated in the electric oven until soaked to a tem-
perature of 540 C for 1 h to avoid distortion, and then
slowly lowered into the molten salt held at 600 C. The
brazing process required 10 min, and when the assembly
was hoisted from the salt bath it slowly cooled to about 150
C before being air blasted and immersed into hot water for
the removal of the salts. The dark surface discoloration seen
in Fig. 4.54 resulted from the incomplete cleaning of salt
residues entrapped within small pores situated along the
braze llet. Such salt contamination leads to surface corro-
sion, and is unacceptable for high-vacuum operation.
Improvements to the salt bath included heating the salt to
700 C and then dropping in a coil of pure aluminium. This
had the effect of removing water picked up from the surface
of previous loads. Molten salts containing a few-parts per
million of water can render aluminium alloys nonbrazable.
By heating to 700 C the water combines with pure alu-
minium to form hydrogen and aluminium oxide.
The aluminium piece-parts were then prepared for the
brazing process. Grease, fatty acids, and possible oils were
removed by vapour degreasing. Oxide lms were removed
chemically by a 1 min dip into a 5 % sodium hydroxide
solution and water kept at 90 C. The oxide removal step
was followed by a rinse/desmutting operation which
entailed immersion for 30 s in a room temperature mixture
of 20 % nitric acid with 1 % hydrofluoric acid. Finally, the
parts were washed in hot water, dried, and stored in clean
plastic bags.
Further brazing operations within 24 h of the cleaning
steps produced virtually pore-free surfaces. An ultrasonic
nal clean in hot distilled water was employed. The chloride
content of the cleaning water was monitored by a silver
nitrate test to ensure a contamination-free flight unit.
After brazing, the alloy 6061 is effectively in the
solution-treated condition. It can be carefully straightened, if
there has been any distortion from dip brazing, and then
articially age-hardened for 8123 h at 160 C.
A major difculty inherent in the design of piece-parts for
dip brazing using brazing pastes, such as aluminiumsilicon
180 4 Spacecraft ManufacturingFailure Prevention

Fig. 4.55 Section through

defective electromechanical
device assembled, using brazing
alloy foil. Incomplete brazing
between interfaces [arrowed on
(a)] resulted from poor jigging
techniques. Quality controls
inspected only for a satisfactory
external llet as detailed in (b).
a Section through device showing
a 90 % incomplete braze path
along interfaces between braze
shim and half bodies. Unetched.
b Detail of external llet showing
the typical AlSi eutectic
structure. Unetched
4.12 Dip Brazing of Aluminium Alloys
ller metal, has been the slower heat-up rates of the thicker
components in the assembly. This can lead to insufcient
braze metal flow and the formation of unsatisfactory joints.
Recent developments have shown that brazing sheet or foil
can be fabricated with a core of 3003 alloy, clad on both
sides with AlSi brazing ller metal produced from
high-purity components. The brazing sheet can be tack
welded between castings, and plates made from different
aluminium alloys. After special cleaning, the assembly is
preheated to 540 C at a rate not exceeding 250 C per hour
and then immersed for 2 min in a brazing bath at 607 C
(possibly only 14 C below the melting point of the casting,
so that temperature control is critical). The assembly is
allowed to cool slightly so that the brazing salt can be flu-
shed away and then slowly cooled to room temperature.
Large parts for the USA Space Shuttle Orbiter have been
successfully fabricated in this manner (Anon 1981). They
are later radiographically inspected and nally heat-treated
and age-hardened.
It should be noted that a competing process to dip brazing
is known as controlled atmosphere brazing (CAB). Only in
recent years has CAB has become a state of the art tech-
nology; it has been developed for the production of alu-
minium heat exchangers used in the automobile industry.
CAB is not known to have been applied to the production of
spacecraft parts as it involves the use of specialized costly
equipment, ideal for mass production but not for the few
parts used in the space industry. Additionally, the actual
furnace atmosphere is a well-kept secret. It is known that
successful ller alloys are in the AWS BAlSi series, and that
a suitable flux, containing potassium fluoroaluminate com-
pounds, has been developed by Solvay Fluor GmbH having
the name Nocolok. This flux is considered to be
non-corrosive before, and as a residue after brazing, being
only active at the brazing temperature whilst in the protec-
tive inert CAB atmosphere (Zhao et al. 2013). Obviously
such fluxes would need to be entirely removed for any
considered space part assembly as is the case with post-dip
brazing processes.
Clad brazing sheet can also be used to dip-braze assemble
miniature parts that have been machined from aluminium
alloys. Figure 4.55 illustrates the importance of accurate
jigging. Here, brazing sheet has been stamped into a com-
plex shim to interface between the two half-bodies of a
spacecraft electromechanical device that would, when
brazed, incorporate several externally sealed cavities. Non-
axial compressive forces applied by an unsuitable jigging
arrangement have caused tilting at one side of the device,
and, although a satisfactory bond has been formed along the
lower side of the brazing shim, there has been only partial
contact between the upper surfaces. All external braze llets
were inspected, assessed to be acceptable, and the device
was then electroplated with silver. The failure of one device
181
under low loading conditions was attributed to the presence
of zones of incomplete braze penetration, as highlighted in
Fig. 4.55. The manufacture of automobile semi-nished
products by means of brazing and the usage of braze clad-
ding is fully described in the useful reference by Sicking
(2005).
Finally, some general precautions must be taken during
the selection of a suitable aluminium alloy for flux bath dip
brazing:
The solidus temperature of the alloy being joined must be
above the liquidus temperature of the braze alloy.
Chemical composition limits should not exceed 0.5 %
copper, 1.5 % silicon, and 1.75 % magnesium (wrought
sheet and plate meeting these requirements are: pure
aluminium 1050A, 1080A, 1200, 1350;
non-heat-treatable alloy 3103; and the heat-treatable
alloys 6061, 6063, and 6082).
Non-heat-treatable materials are brazed with Al 7.5 %Si,
which is fully molten at 615 C; in the case of the
heat-treatable alloys the solidus of the parent metal
determines the choice of braze alloy, and an Al 1013 %
Si braze alloy having a melting range of 595585 C is
used (additions of 25 % copper to the Al 1013 %Si
alloys will lower the braze metal solidus temperature to
approximately 550 C).
As the brazing temperature is higher than the recrystal-
lization temperature of aluminium and its alloys, anneal-
ing takes place during brazing, and much mechanical
strength is consequently lost; the heat-treatable alloys
6061, 6063, and 6082 can be strengthened by quenching
after brazing followed by natural or articial ageing.
The sublimation of aluminium alloys and their brazing
alloys needs to be considered when applied to spacecraft
parts designed to operate under vacuum and at high
temperatures; here, the reader is directed to Sect. 5.6.1 for
an assessment of sublimation data.
4.13 Considerations for the Assembly
of Subsystems by Welding
4.13.1 General Welding Methods and Controls
Automatic welding processes are becoming much more
widely used by both the larger and the smaller companies
engaged in the assembly of spacecraft subsystems. However,
manual tungsten inert gas (TIG) arc welding still nds a
place owing to its versatility, and novel innovations may be
found for applications when the occasional welding of parts
is performed by the contractor. For instance, a range of
flexible welding enclosures have been designed for
182 4 Spacecraft ManufacturingFailure Prevention

applications where a rigid chamber may not be economically

Fig. 4.56 Argweld enclosure for the occasional welding of titanium
and stainless steel parts. This technique eliminates the need for
expensive post-weld cleaning to remove oxide discolorations (courtesy
of Huntingdon Fusion Techniques Ltd.)
viable, as shown by Fig. 4.56. Here the enclosure contains
an argon entry hose and an exhaust valve to vent displaced
gas to the atmosphere. Although manual welding will always
have a place in spacecraft assembly, and particularly during
the repair or modication of flight hardware, experience to
date indicates that there are considerable economies to be
made by using automatic processes which are less
labour-intensive and capable, with optimized control of the
equipment parameters, of producing more consistent and
higher-quality welds. Product assurance schemes will ensure
that these critical processes are overseen by experienced
operators, and test programmes involving the manufacture of
technology samples for destructive mechanical and metal-
lographic evaluation will correct and optimize the process
parameters. True automation, meaning complete elimination
of the operator as has been seen in the automotive industry,
is not viable for the assembly of spacecraft hardware due to
the complexity of the designs and the low number of
piece-parts being made. The wide range of welding pro-
cesses utilized for spacecraft assembly, both of subsystems
and of electronics is given in Table 4.13.
Oxyacetylene gas welding, which has either an oxidizing
flame temperature of about 3500 C, or a reducing flame
giving a temperature of 3000 C, is very seldom used in the
space industry. By far the most popular technique involves
helium or argon arc welding in which a tungsten electrode is

Table 4.13 Principal welding processes for spacecraft assembly

Process Typical applications Typical appearance or
microstructure
Fusion welding
TIG including plasma and Piping systems, RCS tanks and pressure vessels, Ariane V and space station Figures 4.3, 4.19 and 4.62
pulse structures
Manual metal arc For nonstressed applications only (electrical miniature interconnections) Figures 5.31 (torch braze),
Oxyacetylene 5.78
Electron beam High-technology space structures and microwelding electronic equipment Figures 3.6, 4.57, 4.59 and
5.71
Laser beam Precision microwelding for electronics equipment, small structural parts Figure 4.60
Electrical resistance welding
Spot Thin sheet metal work, electronic boxes, solar cell interconnectors Figures 4.1, 4.54, 4.65 and
6.2
Flash Wire joining for electrical conductors Figure 6.4
Solid phase bonding
Friction Dissimilar metal joints Al/Cu, electrical conductors especially lead wires to
component packages
Explosive Al/Cu, Al/steel, electrical, transition joints, launch vehicles Figure 4.61
Ultrasonic Lap joints in foil, thin sections, thin to thick, electrical microne wires
Thermal-diffusion Microwire and compression bonds Figures 4.71, 4.81, 6.56.9
4.13 Considerations for the Assembly of Subsystems by Welding 183

employed while the inert gas is provided through a separate

nozzle to shield the molten metal from atmospheric oxida-
tion; a bare metal ller rod is often used to provide addi-
tional material. In this instance the tolerance in joint gap and
arc voltage, which can be accommodated without loss of
weld integrity, are well known for the standard aerospace
alloys. Once the detailed design of the piece-part has been
completed the next consideration is the procurement of the
base material as a semi-nished product. The availability of
this material in a convenient form within a known time scale
often presents spacecraft project engineers with large prob-
lems due to the fact that small quantities will be ordered and
the alloy may be of an unusual composition and heat treat-
ment. Consideration must be given to this at an early date.
More specically, for critical applications such as pressur-
ized electronic housings, battery cell envelopes, etc.,
obtaining this material to acceptable homogeneity and con-
sistency of quality will be signicant. In-house testing will
usually be mandatory and the failure of procured welding
material to satisfy acceptance criteria such as chemical
composition, surface roughness, hardness, and the like can
result in the rejection of the material and the need to re-order.
This will aggravate the problem of delay and cause inter-
ruption to planned hardware production. Case histories
related to the supply and use of welding rods having a wrong
composition were discussed in Sect. 4.4.4.
The basic composition and heat treatment for piece-parts
that are designed to withstand high structural loads are
selected from the metallic alloys listed in Table 2.3.
Spacecraft engineering often demands that a compromise
must be effected to satisfy the designers need for base
material strength and the need to join the part by a welding
process. Welding will usually reduce the strength of the
material, as shown in Figs. 4.57 and 4.58. It may also have
an influence on the ease of performing a quality control task
or inspection. Special attention must be given to the welding
of thin metallic curved shells which are assembled for use as
cryogenic tanks for conventional launchers. Factors such as
stress-corrosion resistance of the weldment, and in particular
at its heat-affected zone, need to be known. The design must
permit for nondestructive inspection and acceptance criteria
for porosity, and depth of weld penetration should be well Fig. 4.57 Strut intended for satellite structure. Tube and machined lug
established. Clearly such thin structures will be very sensi- (both AA2618) joined by EB welding. This strut had been stored
tive to material defects as well as fatigue loading. When nonloaded for 6 months. After integration, a similar strut failed along
the path denoted by dotted line
reusable tanks are being designed, as for instance envisaged
for future European space launchers, such welded structures
must conform to damage tolerance criteria throughout the
expected service life. Representative samples will be The majority of weld failures result from the vibration
designed and laboratory tested. Specialized equipment exists fatigue of spacecraft models either during their qualication
that can simulate the combined mechanical and thermal or at their acceptance level testing. Because of the relatively
loads that would be imposed on a welded cryogenic tank slow crack growth, detection systems such as accelerometers
during repeated ascent and descent cycles (see also attached onto many different positions around the spacecraft
Fig. 4.92). during vibration testing will warn of distortions and the test
184 4 Spacecraft ManufacturingFailure Prevention

Fig. 4.58 Photomicrographs progressively detail a continuous network of hairline cracks which surround and propagate into the weld pool; several
intergranular cracks also exist in the HAZ. These are polished and then etched details of the weld seen in Fig. 4.57. Residual stresses in the HAZ have
caused the cracking and poor strength of this weldment as will be described in Sect. 4.15

can be stopped before catastrophic (and not inexpensive)
failures occur. The failure of welds by brittle fracture is more
serious in its consequence but rarely has been seen to occur.
The fatigue life of a structural part is determined by the
number of stress fluctuations, the magnitude of these fluc-
tuations, and the kind of stress concentration in the part (see
Fig. 5.2 for a typical fatigue fracture). The designer has
control over the last two and is able to plan to avoid sudden
changes in dimension and, by the choice of weld details, to
prevent cracks and porosity caused by improper processing.
There are large differences in the fatigue performance of
various welded geometries, and the problems of selecting
optimum designs cannot be overstated. Similarly, the
harmful effect of a last minute, uncontrolled, attachment
weld is all too often overlooked. This is illustrated by the
failure of a hydrazine thruster chamber (see Fig. 4.14) which
was modied by an electron beam weld that was never
properly assessed several months prior to the launch of a
telecommunications satellite. Although the modication
overcame one problem, it generated another, which even-
tually caused the device to fail. The launch was delayed and
the defective thrusters were replaced at the launch site. Our
knowledge about the influence of weld defects is always
increasing. It is important that both designers and process
engineers have access to these data if they are to raise quality
and prevent costly failures.
Many descriptions of the commonly used TIG welding
process are described throughout this book. Not described
are the precautions that need to be taken to avoid adversely
affecting the health of the welding personnel. For instance,
arc welding can give rise to ozone emissions and fumes that
exceed the permitted levels dened by national standards
(Anderson and Wiktorowicz 1995; Sterjovski et al. 2010)
this does produce respiratory problems and can even con-
taminate and cause damage to spacecraft electrical hardware.
The eyes must also be completely protected against injury
from directly viewing bright light emitted from welding arcs.
Some case studies can be presented for the more recently
developed welding techniques such as electron beam and
laser welding methods, and these will be presented in the
following paragraphs.
4.13.2 Electron Beam Welding
The fusion process that is accomplished by electron beam
welding (EBW) occurs when a beam of electrons is accel-
erated up to nearly the speed of light and then focused onto
4.13 Considerations for the Assembly of Subsystems by Welding
the metallic pieces to be joined. The kinetic energy of the
electrons is converted into thermal energy, causing the metal
to melt and vaporize. The very high power density can be
focused onto a spot having only about 0.25 mm diameter.
The depth of penetration of the electron beam is dependent on
how well it is focused, its power density, and the type of
material being welded. Copper can dissipate heat rapidly
(Fig. 4.14) and has a wider weld pool than titanium (Fig. 3.6)
or aluminium. The greatest advantage of EBW is its ability to
produce welds that are far deeper and narrower than TIG arc
welds (compare Fig. 3.6 with Fig. 4.8 for similar wall
thicknesses of Ti6A14V alloy, and Fig. 4.20a with b for
Al-2219 alloy plate). Also, again illustrated by Fig. 4.20a, b,
the EBW eliminates the need for multiple-pass welds. Weld
pools formed by conventional arc welding are not only wide,
but are associated with a wide heat-affected zone, as shown in
Fig. 4.8. A similar EBW can have virtually no heat-affected
zone and this results in minimized distortion, enables welding
to be made in close proximity to sensitive areas such as
glass-to-metal seals, and of most importance, the overall
mechanical properties of the joined parts are virtually the
same as the parent materials. Figure 4.59 was a detail from a
very deep EBW, made in a Ti6A14V housing, situated
adjacent to an electro-optical sensor. The fragile device was
not damaged and, as shown in the microsection, the
microstructure of the parent material is not affected by this
butt-weld. The lack of any heat-affected zone adjacent to the
weld pool was conrmed by a microhardness traverse across
the weld.
Fig. 4.59 Microsection across an electron-beam-welded titanium alloy
housing for an optical payload. The total weld depth is 7 mm and the
weld pool has a diameter of only 0.8 mm. The remarkable feature is the
total lack of any heat-affected zone. Microhardness readings (Hv = 500 g)
in the ne-grained parent plate were constant at 332, and increased
sharply to 370 in the columnar structure of the weld pool (transformed
beta-martensite, as illustrated in Fig. 5.18)
185
EBWs are considered to be the most favorable welding
techniques for spacecraft piece-parts. Much experience has
been gained in recent years and unusually combinations of
metallic materials may be successfully welded when other
welding methods have failed. This is true for titanium joined
to beryllium, stainless steel to copper, and deep-drawn alu-
minium to itself (ingrained lubricating oils were outgassed in
the EBW vacuum chamber, whereas under arc welding the
occluded oil produced massive porosity in the weld pool).
Some care is needed when magnetic materials are being
joined, as the electron beam can meander off-centre owing to
magnetic eld effects. Also, because of the high-energy
density of the electron beam, care must be taken to avoid the
kind of unbeaded spatter as was shown in Fig. 3.6. The
notoriously difcult combination of titanium to stainless steel
could not be joined by EBW, as very brittle irontitanium
intermetallics are formed in the weld pool. This interfacing of
metals is often designed out of RCS fuel lines although some
attempts to use alternative joining methods have been
attempted (see Sects. 4.13.4, 4.18, and 6.10).
There are two important disadvantages associated with
EBW. The main one is that this process needs a vacuum
chamberhere, the size of the chamber controls the dimen-
sions of the parts to be joined. The second problem is that
X-rays are generated at the electron beam-to-workpiece inter-
face and therefore operations such as the lid sealing of elec-
tronic packages or the use of EBW close to electronic boxes
must be prohibited as this radiation may damage components.
At present there are no denitive specications for the
control of EBWs or laser welds. A draft version of EN 12185
dated 1996 is considered suitable for controlling the stringent
level-B welds associated with spacecraft parts. This norm
only covers aluminium alloys but its diagrams could be
applied to other metals. The thickness of aluminium alloys
considered as being suitable for conventional EBW is 50 mm,
but up to 200 mm is achievable. For laser welding the method
is limited to material having a maximum thickness of 12 mm.
4.13.3 Laser Beam Welding
Light amplication by simulated emission of radiation
(laser) welding results from the impingement of a coherent
light beam onto the surfaces of the materials being joined.
The laser beam is a stream of photons that is usually focused
by mirrors and lenses to produce a high-density spot size.
Unlike EBW, this method produces no X-radiation and can
therefore be used close to electronic components. It has been
used to hermetically seal spacecraft hybrid packages where
side-walls made of Kovar containing delicate glass-to-metal
seals are laser welded to lids cut from stainless steel or iron
nickel sheet. The process does not require vacuum and it is
possible to optically transmit the beam over large distances
186
Fig. 4.60 Polished and etched microsections of laser-welded space-
craft axle shaft. Different welding parameters were selected to optimize
the strength of this structural joint made of titanium alloy. Mechanical
testing has revealed the paths of weakness (Dunn to be published)
to produce welds in locations difcult to access by other
welding techniques. The author has also used a small pulsed
laser to repair solar cell interconnection welds that were
present beneath glass cover slips. The versatility of the laser
has led to many applications associated with welding. It has
4 Spacecraft ManufacturingFailure Prevention
been used to cut both aluminized plastic sheet (to produce
solar reflectors for the Hubble space telescope) and metallic
sheet, which was later joined by laser welding. Most appli-
cations have used carbon dioxide lasers, as they offer the
necessary power and speed of welding, but Nd/YAG lasers
are becoming more important as developments have
increased their power.
Laser welding equipment has been used to both cut and
successfully weld sheets (from 3 to 6 mm thickness) of AISI
316 stainless steel, Inconel 600, various aluminium alloys,
and titanium6A14V. Butt-welding designs are seldom used
but it is possible to introduce ller wire to the welds to
produce a more voluminous bead. More usual is the selec-
tion of interlocking, or overlapping designs. The axle shafts
for a solar array drive mechanism have been recently con-
structed from laser-welded Ti6A14V. The test plan for
optimization of these welds involved the manufacturing of
many test specimens using a wide range of weld parameters.
The specimens were subjected to tensile testing and metal-
lography (see Fig. 4.60). Welds having a penetration depth
of only 30 % were seen to fail in the weld pool. Welds
having a greater penetration than 50 % could be seen to
initiate cracking just inside the weld pool with further
propagation through the heat-affected zone. The nal failure
is noted in Fig. 4.60c to be quite ductile and in the parent
material. These photomicrographs also show that the
heat-affected zone of laser welds is intermediate between the
wide zones caused by arc welding and the absence of such
zones adjacent to EBWs. Because of the slightly lower
energy of the laser weld when compared to an EBW, there is
no underbead spatter and complete fusion is achieved.
Expert opinion is that there are a vast number of com-
ponents still being conventionally welded which could be
more effectively laser welded (Holt 1995).
4.13.4 Explosive Welding
Explosive welding has been known for at least 20 years. It
was initially used to bond corrosion-resistant liners to the
internal surfaces of steel pressure vessels. Newer uses have
found application in the space industry. One use is for the
joining of Ti6A14V to stainless steel. This interconnection
has been demonstrated as an aerospace transition joint
between stainless steel plumbing lines and thin-walled tita-
nium alloy pressure vessels. The steel tube is initially
inserted into a titanium alloy ring supported by a
thick-walled outer mandrel. A small stand-off gap is needed
between the surfaces being joined. Explosive material is
placed within the tube. So-called cladder material surrounds
the explosive. When the explosive is detonated a shock wave
passes through the wall of the steel tube so that it plastically
deforms and makes contact with the inside surface of the
4.13 Considerations for the Assembly of Subsystems by Welding 187

titanium ring. The yield strengths of both materials are

exceeded and a wave passes down the interface. At the peaks
and troughs of these waves there are discrete miniature zones
of melting and alloying of the interfaces. An alternative
method is to surround the outer ring with explosive charge
and ll the inside of the steel tube with a solid or liquid
mandrel. Once this transition joint has been cleaned and
re-machined to closer tolerances, the titanium ring can
subsequently be joined into a titanium alloy pressure tank by
means of EBW or arc welding.
Explosive welding is regularly employed for joining
aluminium sheet rings to titanium sheet rings of a larger
diameter. This combination of materials is used to form
transition rings for the fabrication of spacecraft detectors.
Very thin, circular beryllium windows are EB welded to the
aluminium part of the explosively welded transition ring.
A concentric U-shaped bend is impressed into the flat sur-
face of the titanium ring; this acts as a stress-relief bend
when the detector is subjected to thermal cycling in orbit.
This titanium ring is later welded into the detectors light-
weight titanium alloy housing.
The Italian SAX spacecraft (Satellite for Astronomy in
X-rays) was a complex observatory designed to characterize
a variety of stellar and galactic objects. The spacecraft was
very successful and de-orbited in 2005. However, one
manufacturing challenge had to be overcome. The author
participated in this workhow to bond the large beryllium
X-ray spectrometer window to its titanium alloy housing.
The solution is proprietary but incorporated explosive
welding the piece-parts together in the Californian desert
explosive charge parameters were optimized by using test
samples, and the nal flight detector was successfully pro-
duced having strength, hermeticity and sufcient robustness Fig. 4.61 Explosive weldingmicrosections of stainless steel joined
to survive extensive environmental testing. with an intermediate layer of brass to a thick copper substrate. The
Intermediate layers may be placed between the surfaces wavy interfacial bonds are evident in (a). The melted islands of alloying
material are detailed in (b) at the vortices of the waves
of parts being joined by explosive welding. Figure 4.61
illustrates such an intermediate lay of brass that has been
successfully sandwiched between stainless steel and copper. 718copper, copperbrassstainless steel, the combination
The detailed photomicrograph highlights the wavy interface of aluminiumTi6A14V and copper to tantalum (Maliutina
of the steel to brass bond. Islands of melt are present at the et al. 2014). Many of these combinations have no mutual
vortices of the waves; these could be brittle intermetallics, so solid solubility and it has been shown that a thin volume of
it is important that their size and continuity are minimized to hetrophase material will form at these interfaces (Greenberg
achieve optimum bond strengths. In order to form a quality et al. 2012).
weld and be able to reproduce it, the variables that affect the
weld formation must be controlled. There are three major
parameters which need to be controlled: the explosive det- 4.13.5 Welding of AluminiumLithium Alloys
onation velocity, the explosive load, and the interface
standoff spacing. Mechanical testing combined with metal- The space industry has been slow to accept the new series of
lographic investigations are used to monitor the quality of aluminiumlithium alloys for any major applications.
these items. The following systems are some that have been Despite the large amount of research that is reported in the
successfully bonded by explosive welding: titaniummild literature concerning various properties of these alloys there
steel, titaniumcopper304 stainless steel, aluminium304 has been no possibility to gain any user experience. This
stainless steel, aluminiumsilverstainless steel, Inconel may be due to the fact that the poor stress-corrosion
188 4 Spacecraft ManufacturingFailure Prevention

resistance of the lightweight alloys has precluded their

selection for structural items and that these alloys are rela-
tively difcult to weld by conventional methods. It appears
likely that the large external tank of the Space Shuttle will be
fabricated from a new generation of Al-Li alloys designated
Al-2095 and 2195 (see also Table 3.2). These alloys will
replace the present tanks, which are constructed with
Al-2219 and have been successful since their introduction in
1980. The Al-2095 and 2195 are resistant to stress corrosion,
they can be protected against surface corrosion by chemical
conversion coatings, and they are weldable using methods
similar to those for Al-2219. The main drawback is their
densitywith up to 6 % copper and a lithium content of
only up to 1.5 %, the potential weight-savings are not as
good as other alloys presented in Table 3.2.
The desirable lightweight AlLi alloys have a poor
weldability. This is evident from mechanical testing and
metallographic evaluations which reveal severe defects such
as weld metal porosity, hot cracking and weakening of the
heat-affected zone. These effects have been summarized by
Ellis (1996). It would appear that the problem of porosity
may be due to interstitial elements such as oxygen and
hydrogen present in the material, together with chemical
compounds associated with the very active lithium con-
stituent. The compounds are probably lithium hydride,
lithium carbonate, and lithium hydroxide; these will all
decompose during the welding process and release large
volumes of hydrogen, carbon dioxide, and steam (Ilyush-
enko 1993). The hot cracking susceptibility was investigated
by Novikov and Grushko (1995) and was shown to occur as
the liquid weld pool solidied. The solid phase in the weld
pool increased by dendritic growth as the weld cooled and Fig. 4.62 Photomicrographs of polished and etched TIG-welded
was accompanied by a certain contraction. This contraction aluminiumlithium alloy plates. a Shows AA 2090 welded with AA
could not be lled by any flow of the last volume of weld 2319A ller (AlCu), and b AA 8090 welded with AA 2319A ller
metal to solidify, as is the case for Al-2219 when liquid (bottom edge of weld bead shows slight porosity) (magnication 40)
metal of eutectic composition lls these shrinkage gaps, and
this results in microcracking of the weld. 4.13.6 Welding of Thermoplastics for Space
An example of a trial spacecraft weldment is shown in Applications
Fig. 4.62. The porosity problem can be reduced by
mechanically removing up to 0.5 mm from the surfaces of The number of thermoplastics selected for space use has
the wrought AlLi stock by milling before welding. increased over the past twenty years. They are used in
Chemical milling in alkaline solutions of vacuum degassing electronic circuits, as non-adhesive tapes and foils; particu-
followed by milling will also reduce porosity. The problem larly multilayer insulations (MLI) used for thermal control
of hot cracking can be avoided if a nonmatching, purposes and flexible second surface mirrors (e.g. solar
crack-resistant ller wire is used during TIG welding of the reflectors) as well as sleeves, tubing and small mechanical
AlLi alloy, and some examples are given by Ellis (1996). parts. Polystyrene softens from about 80 C and above and is
Both laser and electron beam welding can produce not space approved, whereas polytetrafluoroethylene (PTFE)
defect-free joints in some of the Al-Li alloys. The majority will only soften above 300 C. Recent developments in the
of alloys will respond to a suitable post-weld heat treatment welding of thermoplastics may see these joining processes
to increase their tensile strengths. being applied more often in spacecraft assembly areas.
4.13 Considerations for the Assembly of Subsystems by Welding
Complex adhesive bonding processes that require surface
preparation of the parts to be joined coupled with heat cycles
for curing that may damage surrounding areas.
In some ways the welding of thermoplastics is similar to
welding of the light metals. Methods include laser welding,
friction stir welding and ultrasonic welding. Only like-to-like
plastics can be welded and torch or hot gas welding cannot
be applied to space parts. Laser welding appears to be the
most satisfactory method and relies on the upper part
being transparent to the laser beam. Usually the parts are
placed together by a pressure xture so enabling weld widths
of between 10 and 100 m. Specialist papers should be
consulted when specifying methods for non-destructive
inspection and mechanical testing of these connections
(Troughton et al. 2013).
4.14 Control of Power System Weldments
4.14.1 General
The two most important elements of every spacecraft power
system are the energy sourceusually a solar arrayand the
energy storagemainly nickelcadmium or nickel-hydrogen
batteries used to provide basic power in eclipse, or peak
power in sunlight. The charging and discharging of the
spacecrafts batteries will be determined by the routine
interruption of the power source as the Earth passes between
the orbiting satellite and the Sun. The characteristics of the
orbit can be drawn into four main groupings:
low Earth orbit (LEO): close to the Earth the eclipses last
about 0.5 h during each 1.5 h orbit
geostationary orbit (GEO): 36,000 km above the Earth,
with a 1.2 h eclipse for a 24 h orbit
highly elliptical orbit (HEO): perigee close to the Earth
and apogee far removed, which gives rise to variable
eclipse/orbit ratios
deep space orbits: result in very few eclipses, but large
variations in solar energy; in this case a radioisotope
generator may be the energy source.
The most demanding orbit listed above is the LEO, with
its high eclipse-to-sunlight ratio of about 30 %. This means
that a heavy discharge of power during the 0.5 h eclipse will
need a full battery replenishment in a period of less than 1 h.
Other orbits will provide for longer charging cycles, but there
can be quite different power requirements imposed by the
mission. Science missions have various orbits and generally
require lower power levels (200 W1.5 kW). The Hubble
Space Telescope (HST) array, with a surface area of about
130 m2, is providing 4.6 kW of electrical power of 34 V.
Telecommunication satellites will be in GEO orbits and may
189
need power levels up to 3 kW during eclipse and up to 10 kW
sunlight power. Earth resources satellites may also need
10 kW but, like the HST, they are in LEO where there will be
greater stresses on the arrays and batteries. Manned missions
such as Spacelab and the International Space Station are also
in LEO and need power levels of between 3 and 50 kW.
4.14.2 Welded Solar Arrays
The current delivered by a solar array is essentially propor-
tional to the solar energy received by the array. This current
will vary owing to the pointing angle of the array to the Sun
(cosine factor) and the distance between the Sun and the
spacecraft (inverse square rule). By similar rules, the maxi-
mum heating effect of the Sun on the solar array will depend
on the arrays distance and orientation of the Sun. Heat
transfer from the array is by radiation to deep space during
periods of eclipse. Spacecraft in LEO, such as the Hubble
Space Telescope (HST), are at this moment surviving about
30,000 thermal cycles between 100 and +100 C during a
ve-year period of operation. A more benign orbit may cause
temperature cycling between 80 and +80 C whereas a
severe cycle would be from 150 to +100 C. The HST has
two solar panel assemblies as shown in Fig. 4.63 (each panel
is made up of ve power-generating sections).
Every cycle between hot and cold results in the array
components expanding and contracting according to their
individual coefcients of thermal expansion (CTE). The sin-
gle strings of solar cells are interconnected by means of flex-
ible foils that have been formed to include a stress-relief bend.
One design of solar array is shown schematically in Fig. 4.64;
here, the solar array has a particularly high electromagnetic
cleanliness due to the presence of grounded cover glasses.
In the past 30 years, monocrystalline silicon solar cells
have been used on European space solar generators. The
cells are connected by silver mesh or silver foil. Early joints
between the silver interconnectors and the silver-plated front
or back face of the cell were achieved by soldering. The
cyclic temperature range and high upper temperature caused
the microstructure of the solder joints to coarsen and rapidly
fatigue. Welded interconnector-to-cell joints were developed
and accelerated testing together with actual flight experience
has proved that these joints will withstand 10 years in LEO
and 15 years in GEO. The gradual degradation of the welded
joints can be followed nondestructively by means of scan-
ning laser acoustic microscopy (SLAM, described in Sect. 3.
2.3.2) or by destructive metallography. The method is shown
in Fig. 3.15 and the captions of Figs. 4.65, 4.66, 4.67 and
4.68 describe actual cases where welding processes were
assessed for quality and reliability.
Large thermal stresses build up on the welded joints
because of the different CTE behaviours. With poor-quality
190 4 Spacecraft ManufacturingFailure Prevention

Fig. 4.63 Post-flight inspection of the Hubble Space Telescope (HST)
solar array shortly after retrieval. This shows one of the two HST solar
array wing assemblies being opened out at the MMS-UK facility where
it had originally been fabricated. All outer surfaces were photographed
to assess the effects of 3.6 years operation at 600 km above the Earth.
This is the largest roll-out solar generator built to date, employing a total
of 48,760 BSFR silicon solar cells to provide more than 5 kW of
electrical power. During inspection (see also Fig. 8.4), 704 large
hypervelocity impacts were photographed. 148 impacts had completely
penetrated the arrays thickness but they did not cause functional failure
of the array (a total of 3862 impacts was counted on the front side of one
wing and 378 on the rear side. 738 of the 24,380 cells were damaged)

Fig. 4.64 Cross-section through a solar array for a scientic satellite (GEOS) showing location of various materials and the stress-relief loop of
the silver cell interconnect. Note this assembly supports quartz cover slips that are also grounded by interconnection
4.14 Control of Power System Weldments 191

Fig. 4.65 Solar cell assembly from a large communications satellite weldments are located between arrows. b Detail of weld electrode
solar array after 30,000 thermal cycles between 100 and +100 C. impressions on silver mesh from an area arrowed on (a). These
a Photograph shows rear of cell after dissolution treatment and peeling impressions are very shallow and are only just discernible along the
away of substrates. The silver interconnector-to-silver-metallization lines xx and yy. Illuminated under glancing light

Fig. 4.66 Examination of the resistance-welded silvermesh intercon- acoustic-amplitude micrograph. The weld points have been superim-
nector (shown optically in Fig. 4.65) by SLAM identies discontinuous posed as zones xx and yy. Also superimposed are the foil perforations
bonded regions beneath the welding electrode pressure points. These and the bond areas
composite micrographs depict a the acoustic interferogram and b the
192 4 Spacecraft ManufacturingFailure Prevention

Fig. 4.67 Composite

photomicrographs of as-polished
sections made across the silver
mesh interconnector detailed in
Fig. 4.65. Plane of sections runs
almost parallel to weld
impression yy. Slice a runs along
boundary of weld impression y1
y1; only two bonds are identied:
under C and to the left of B.
Slice b is along midsection of
weld impression. Bonds exist
under A, B, and C. Other
parts of the mesh show no
evidence of contact. These
locations correspond exactly with
the zones identied by SLAM in
Fig. 4.66
4.14 Control of Power System Weldments
Fig. 4.68 Microsection of
silver-coated molybdenum
interconnector weld from the
Hubble Space Telescope (see also
Fig. 4.63) after 30,000 thermal
cycles between 100 and +100
C. The weld zone has no fused
nugget, which is commonly
observed between
resistance-welded parts, but the
evidence of grain growth
(arrows) does indicate that the
temperature, time, and pressure of
welding have produced
strain-induced, grain-boundary
migration. There is no sign of
thermal fatigue
welds, the CTE mismatches between flexible substrate, cell,
and interconnect foil can produce low-cycle fatigue where
the bonded silver surfaces become detached. A far stronger
weld occurs when silver-plated molybdenum foils are used
as interconnectors. In this instance electrical resistance
welding causes a far greater temperature than for the pure
silver foils. This is due to the greater electrical resistivity of
the molybdenum. It has been seen (Dunn 1982a, b) that the
plated molybdenum foil forms a true weld between its silver
coating and the silver coating on the silicon cell, as there is
recrystallization and grain growth across these interfaces.
The silver-coated molybdenum interconnectors were selec-
ted for the HST. In December 1993 one of the two flexible
HST solar array wings was successfully retrieved from
space, after 1320 days in a 600 km orbit. Examination of the
welded interconnectors from this array using SLAM and
metallography showed that the 21,000 thermal cycles
between 100 and +90 C had only slightly reduced the size
of the weld areas (de Rooij and Collins 1995).
In order to maintain a high quality of welded intercon-
nections during a spacecraft array production run, it is
essential to control the dimensions and the physical prop-
erties of materials that make up a particular solar-module
193
design. All can indirectly affect the nal weld quality, and
the following controls are essential:
the interconnector material for chemical purity, hardness,
thickness, and form;
the solar cell for signs of contamination of the silver
metallized layer (e.g. sulphides and oxides), thickness,
and surface roughness; and
the repeatability of all welding parameters, such as
electrode dimensions, pressure, voltage, and duration of
weld pulse.
Even if these joints are made to the most optimized
processing parameters, they will have a denite end-of-life
resulting from thermal fatigue. This wear-out is caused by
the large difference between the CTE of silicon and that of
silver (respectively, 3 and 19 ppm). Molybdenum has an
intermediate CTE of 4.9 ppm, and this certainly improves
the joint integrity. It will also be seen, in Chap. 8, to be
resistant to the effects of atomic oxygen in low Earth orbits.
4.14.3 Suitability of Welded Battery Cells
As might be expected, the batteries also pose a problem for
materials engineers, as there will be a consistent charging
and discharging whilst spacecraft orbit the Earth. As a
general rule, when a large number of low-depth discharges
are expected, for instance in a LEO operation, then
nickel-cadmium cells are selected. When the power system
encounters only a few, deep discharges in a GEO, then
nickel-hydrogen cells will be preferred. The chargedis-
charge rate of the battery is controlled by current regulation.
Ways to sense the charge-state are to measure the cells
voltage and the gas pressure within the cell case.
An old production NiCd battery cell was shown to be
leaking in Fig. 4.13. The leak was caused by corrosion and
resulted from the wrong selection of metallization material.
Corrosion tests are now performed on all new cell materials to
ensure that they are fully compatible with the cells elec-
trolyte. Mechanical tests are also performed and in order to
verify the structural integrity of the battery cells during sim-
ulated charge/discharge cycles, the cans are pressured-cycled
between 5 and 1 bar using dry air. A common test is to
pressure-cycle the empty cans at a rate of 12.5 cycles per hour
up to a total number of 6000 cycles. A batch of battery cans
having a new, lightweight design was submitted to this test.
The cover/base plate of the cans was made of Kovar and the
can wall was formed from a nickel alloy. Autogenous TIG
194 4 Spacecraft ManufacturingFailure Prevention

Fig. 4.69 a Microsection made through leaking battery cell can after
pressure cycle failure. Wall thickness is only 0.42 mm thick. b Note
initiation of fatigue cracks arrowed in the TIG weld

welds joined the can wall to lips formed into its cover and base
(corner flange weld design). During pressure-cycling tests
these cans began to fail by losing their hermeticity in the
welded regions after only 1200 cycles. The failed welds were
inspected by scanning electron microscopy and metallogra-
phy. Three distinct modes of failure were seen on the fracture
surface of the weld metal (see Figs. 4.69, 4.70). The initial
crack in the weld pool is caused by mechanical fatigue, and
this has increased in depth until overloading occurred with
shear tearing and nal fracture by tensile overload. The
quality of the weld and the suitability of the selected metals Fig. 4.70 a Shows SEM view of fractured weld in region of leak
are not in doubt, but the cross-sectional area of the can was of following battery case pressure cycling test. Arrow indicates inside
insufcient dimensions to withstand the pressure cycles. The surface (same position as crack initiation site shown in Fig. 4.69). There
cans were redesigned with a 50 % increased wall thickness. are three distinct fracture zones. b Detail of zone A showing
mechanical fatigue striations running parallel to edge of weld pool.
With this modication, further cyclic-pressure tests and burst c Zone B shows tear overload followed by tensile overload and ductile
tests were completed successfully. dimples
4.15 Problems Associated with Residual Stresses in Weldments
4.15 Problems Associated with Residual
Stresses in Weldments
Electron beam (EB) welding is frequently employed during
spacecraft construction. It is particularly suitable for weld-
ments that are bodies of rotation, e.g. pressure vessels and
tubular assemblies. The welding motion is mechanized, and
generally narrower weld widthsand therefore reduced
distortion and heat-affected zones (HAZ)are produced. It
should be remembered that even EB welds will contain
certain residual stresses which are caused by expansion
differentials, shrinkage of the weld metal as it solidies, and
subsequent contraction due to uneven cooling. The design of
one spacecraft structure included a strut manufactured by EB
welding an aluminium alloy tube to an end-piece lug which
had been machined from the same high-strength
precipitation-hardening alloy Al 2618.
Figure 4.57 shows the appearance of this painted cylin-
drical weldment; the section shows that adequate weld
penetration has fused and welded the end of the tube to the
end-piece without the need for a ller metal. During inte-
gration, this weldment was observed to fracture under very
low loading conditions. The crack path had propagated
through the weld region in a manner superimposed on the
micrograph.
As part of this failure investigation, an as-welded, but
never loaded, strut was microsectioned. This sample was
lightly etched (Fig. 4.58) to reveal the presence of a ne
network of cracks at the weld pool-to-parent metal interface.
Of great importance is the evidence of multipass welding as
shown by the liquation zones present in the etched weld
pool. Some large, secondary intergranular cracks have also
formed in the HAZ. It was later conrmed that cracks did not
exist in newly made weldments. In fact, they had initiated
and grown at some stage during the struts 6-month storage
period. In this case, the reasons for crack initiation and
growth are believed to be twofold. The strut material is a
rather complex precipitation-hardening alloy which under-
goes various submicroscopic changes during fabrication.
The alloy adjacent to the weld pool (HAZ) heats up during
the welding operation and undergoes a localized solution
treatment. During the second, third, or subsequent fusion
there is extreme thermal expansion and contraction in the
base of the weld. The local stresses may be so great as to
cause yielding and bending of the material (Verderaime and
Vaughan 1995). At room temperature, the residual stresses,
which have been set up within the material as a result of
weld metal contraction, cause the solution-treated HAZ to
gradually harden because of natural strain ageing. Unfortu-
nately, such a condition may become embrittled, particularly
if exposed to a mildly corrosive environment. It should be
noted that, following the welding process, these struts had
been chemically cleaned, alodined, and nally painted. It is
195
likely that these corrosive processing environments caused
crack initiation sites which later were to grow by a
stress-corrosion mechanism.
Based on the failure analysis results, subsequent struts were
carefully heat-treated in a furnace immediately after the
welding operation. The residual welding stresses, which can
reach levels as high as the yield strength of the welded struc-
ture, were sufciently relieved during this heat cycle. At the
same time, the solution-treated HAZ was modied to a state of
controlled higher strength owing to precipitation-hardening.
No further cracking problems were experienced after the
adoption of this post-weld heat-treatment procedure.
4.16 Electromagnetic Emission from TIG
Welding Equipment
Arc welding equipment is known to produce a high level of
electromagnetic emission which can cause interference to
electrical communications systems. Laws covering the
electromagnetic compatibility of industrial welding equip-
ment came into force within the European Union in 1996.
This legislation makes it an offence to produce excessive
electromagnetic emission. The emission (mainly radiation)
can be produced by both the welding power source and the
arc itself (Lucas 1995). Unfortunately it is impossible to
eliminate these emissionsthey should also be considered
when welders bodies may be exposed to locally conned
high magnetic eld strengths, such near the welding torch,
its cables and power source (Mair 2005). They begin at the
moment the welding arc is started. The space between the
welding electrode and the metal workpiece is ionized by
high-voltage sparks (540 kV at a frequency within the
range 101000 kHz). Further sparks are generated during
welding owing to the presence of surface oxides on the
piece-parts being joined; this has an effect similar to light-
ning as miniature flashes propagate through the shielding gas
(e.g. argon). In addition, it has been observed that the
welding cables connecting the weld head to the power
supply/controller, which may be 10 m away, will act as
antennae, causing the level of radiation to increase.
Many electronic devices are sensitive to electromagnetic
radiation. Components such as CMOS packages are partic-
ularly vulnerable and are known to have failed owing to
radiation emitted from TIG welding equipment which was
being used, in close proximity to the electronic circuitry, to
carry out repair welds on both the Voyager and Galileo
spacecraft (Lumsden and Whittlesey 1981). Concerns that
similar damage to electronic devices might occur were
expressed during the replacement of hydrazine propulsion
system tubing (RCS thruster pipes) on board flight spacecraft
of the Cluster project. These pure titanium tubes were
thought to be over-etched and had shown signs of weakness
196
when incorrect xation procedures had caused a length of
unsupported tube to fracture during ground transportation of
the spacecraft.
A specialized circumferential in-place pipe welding head,
developed for aerospace applications, was used to both cut
the defective pipes (3.17 mm diameter, 0.25 mm wall thick-
ness) and, later, weld together the replacement pipes. Some
manual saw cutting of the pipes was also undertaken using
nitrogen purging and special contamination controls to ensure
that slivers of metal were not introduced into the propellant
system. Welding was performed by rotating the TIG electrode
around the joint on a ring gear. The butt joints were welded
autogenously with arc voltage and other parameters con-
trolled by a microprocessor. Experiments were made to
ensure that minimal electromagnetic emission was produced
during the weld strike. This work followed the practices
outlined by Lumsden and Whittlesey (1981) at the Jet
Propulsion Laboratory. Here, the use of a whisker arc start
instead of the usual transient-arc method was noted to reduce
the initial burst of radiation. A 50 m diameter pure titanium
metal whisker is placed between the electrode and workpiece;
it is vaporized by the weld current, leaving a plasma to initiate
the arc. The spark gap was reduced to 0.5 mm, to reduce the
spark voltage. Low-impedance braid (less than 10 m) was
used to ground the spacecraft structure and the welder chassis
to the facility earth. The welder-to-weld head cable was
completely shielded with additional aluminium foil and
Eccoshield tape wrap and grounded to the welder chassis.
Some of the sensitive electronic boxes were removed prior to
the repair welding operation, and others were covered with
aluminium foil and rectangular ferrite pieces to isolate them
from the electromagnetic radiation, albeit reduced by the
whisker-wire start. Prior to spacecraft repair welds being
made, test specimens were produced and controlled by
X-radiography, bend testing, microhardness checks, and
metallography. No radiation damage was caused during these
operations and the resulting flight weldments are considered
to be of excellent metallurgical quality.
4.17 Titanium Aluminides
for High-Temperature
Applications
Structural metals for spacecraft applications are predomi-
nantly selected from either aluminium alloys or titanium
alloys. The latter are well established as lightweight engi-
neering materials and can be developed to withstand tem-
peratures in the order of 600 C. A new class of material, the
titanium aluminides, is receiving much attention in the 1990s
and it is probable that they will nd applications in the next
generation of aero engines, space launcher vehicles with
re-entry capabilities, and space platforms. At the present
4 Spacecraft ManufacturingFailure Prevention
time there are no examples of any aerospace usage, but
owing to their characteristics of light weight, surface sta-
bility, potentially high operation temperatures, and oxidation
resistance, the titanium aluminides are certainly worth a
mention as being revolutionary engineering materials.
These new materials are not solid solutionable alloys, but
are intermetallic compounds having a fairly xed composi-
tion. There are two classes of titanium aluminide. The rst
has the composition Ti3Al, termed the alpha two alloys with
useful structural properties up to 700 C. The second is based
on TiAl, the so-called gamma titanium aluminides, and
these are reported to have been developed in the USA to
reach strength levels as high as 1000 MPa at room temper-
ature and 500 MPa at 1000 C (Kim 1995). The European
space community appears to favour the development of TiAl
with aluminium contents of 48 at.%. As a binary alloy it has a
moderate elongation to fracture at room temperature, but
inadequate creep strength and poor oxidation resistance at
elevated temperatures. These disadvantages are avoided
when the gamma aluminide is alloyed with approximately
2 at.% of chromium, manganese, or vanadium, and/or 24 at.
% of niobium, tantalum, molybdenum, or tungsten. More
recent developments also include the deliberate additions of
up to 1 at.% of silicon or boron. The initial group of elements
is introduced to improve further the room temperature duc-
tility (they enhance mechanical twinning and the movement
of dislocations). The second and third groups improve the
high-temperature properties of creep strength and oxidation
resistance. It is very important to avoid the pick-up of
interstitial impurities such as oxygen, as these severely
reduce the room temperature ductility from about 4 to 0.5 %
elongation at fracture. At temperatures of between 600 and
700 C there is a remarkable increase in ductility, with
fracture elongations of up to 85 %. This increase is due to
dynamic recovery enhanced grain boundary sliding, and the
start of dynamic recrystallization (Clemens and Schretter
1996). The fabrication routes for forging and rolling
gamma-TiAl alloys (mainly Ti48Al2Cr) have been well
researched and the different possibilities of optimizing the
microstructure of these intermetallics by heat treatment to
produce a variety of structures from near gamma grains, to
duplex grains, and even ne lamellae in a small grain size,
have been described in a number of papers published by
Plansee Aktiengesellschaft and referenced by Clemens and
Schretter. Sheet material with a small grain size (about 10 m
diameter) has been tested in tension at 950 C and shown to
have elongations of 200 %. Dynamic recrystallization pro-
cesses that accompany the deformation at these high tem-
peratures are responsible for achieving superplastic forming
processes where the deformation ratio is in the order of
100 %. A prototype example of Ti48Al2Cr sheet that has
been pack rolled to dimensions up to 480 205 2 mm and
then superplastically deformed and diffusion bonded can be
4.17 Titanium Aluminides for High-Temperature Applications
Fig. 4.71 Gamma-titanium
aluminide sheet, diffusion bonded
to superplastically deformed ribs
of the same intermetallic
compound, as a demonstration
piece for an airframe structure
(courtesy of Plansee GmbH)
seen in Fig. 4.71. Whereas diffusion bonding of the inter-
metallic gamma-TiAl is well established as a joining process,
more research needs to be made into how it may be joined by
conventional methods of welding. It would appear that the
heat input (and related expansion or contraction) during
welding of all the TiAl intermetallics causes local cracking
in the colder parent metal where the thermal stresses and
distortions are difcult to be absorbed by a material with a
room temperature elongation of only 14 %. Crack-free
joints have been made by heating gamma-TiAl to above its
600 C brittle-to-ductile transition temperature just prior to
welding, but such preheating is an impractical process for
large piece-parts unless a special welding furnace is available
(Clements et al. 1996). A stability test has been performed on
a welded panel (500 250 1 mm) supporting superplasti-
cally formed and laser-welded stiffeners (similar to the
diffusion-bonded panel shown in Fig. 4.71). The test panel
survived thermal and compressive loading to test tempera-
tures of 750 C. This material has been proposed for further
development as it has the potential to replace superalloys in
the temperature range of 700950 C and therefore could nd
applications in hypersonic aerospace technology.
4.18 Shape-Memory Alloys for Spacecraft
Devices
Several metal alloys exhibit the so-called shape-memory
effect. The rst alloy to be exploited for commercial appli-
cations was one discovered in the early 1960s by W.
J. Buehler. It has been named Nitinol, reflecting its
197
composition and place of invention (nickeltitanium Naval
Ordnance Laboratory). Other, less-expensive shape-memory
alloys based on copper, zinc, and aluminium were developed
in a collaboration between the former Fulmer Research
Institute and the Raychem Corporation. These ternary alloys
were manufactured into devices which operated as thermo-
static valves for controlling domestic radiators and opening
or closing window arms. More recently, high-temperature
shape-memory alloys have been developed (Golberg et al.
1995). These are based on titaniumpalladium alloys with a
martensitic transformation temperature of 550 C, and these
could nd future use in hot spacecraft structures. The
shape-memory alloys possess a mechanical memory. When
they are heated to a certain transformation temperature they
remember a previous shape and promptly revert back to that
shape. Considerable force is exerted as the alloy changes
shape, and this can be used to perform useful work, such as
the opening and closing of ofce windows, or to provide
power for activation and operation of robotic arms on plan-
etary lander vehicles.
Shape-memory alloys (SMAs) undergo a crystalline
phase change when they are heated. This change in phase
alters the alloy lattice-constant, resulting in a change in
dimension. Upon cooling, the alloy reverts to its original
lattice structure. For an understanding of the process of
shape-change the crystallographic forms of the alloy must be
known. Nickeltitanium alloys have the majority of aero-
space applications and the behavior of these alloys is fairly
well known. As a demonstration we can consider the man-
ufacturing route for a spacecraft antenna that has a form
similar to a dish-shaped spiders web. The construction
198
Fig. 4.72 Hinged-tube
application [from Schuerch
(1968)]
material is Nitinol wire that has been wound into a large coil.
The coil is placed into an oven containing an inert gas. The
temperature is raised to 650 C and the metal undergoes a
phase change to that having the crystal structure of austenite.
The wire is now cooled to room temperature. As it cools a
new crystal phase called martensite begins to form at a
precise temperature of 60 C; the wire is completely trans-
formed to martensite on reaching 52 C. While the wire is in
the martensitic state, conveniently at room temperature, it is
cut into short lengths and tack welded into the form of the
dish-shaped spiders web. The nished antenna can then be
crushed, or folded into a small volume (suitable for the
conned space of a launch vehicle). In order to restore the
original shape of the antenna the folded volume is heated.
On reaching a temperature, also well dened at 71 C, the
martensitic phase begins to transform into the austenitic
phase and this modication of the metals crystal structure is
accompanied by dimensional changes in the wire that cause
the individual segments of the web to unfurl. At 77 C the
remaining martensite is completely transformed to austenite
and the antenna will be seen to have been restored to its
original shape.
The antenna example has been fully demonstrated and, by
the same metallurgical phase changes, Nitinol has been
employed to produce hinged tubes, deployable masts, and
4 Spacecraft ManufacturingFailure Prevention
deployable struts. Examples are sketched in Figs. 4.72, 4.73
and 4.74.
The SMS alloys have been found suitable for the cou-
pling of a variety of tube materials, particularly the
difcult-to-weld stainless steel to titanium alloy tubes. Fig-
ure 4.75 shows the effect of heat on previously expanded
Cryocon and Cryot coupling rings. Each ring is made of
Nitinol (NiTi), NiTiFe or NiTiNb, and has a 35 % rever-
sible change in diameter. On reaching the
martensite-to-austenite transition temperature, the ring con-
tracts and four knife-like protrusions lock mechanically into
the tubes surfaces to form a hermetic seal. The force that the
alloy exerts in reducing its strain is approximately the yield
strength of the warm alloy. Since this is about 410 MPa, an
enormous force is present around the tubes circumference.
These couplings have been used in special situations
namely on a European space project when a ll and drain
nozzle, located in the forged dome of a Ti6Al4V hydrazine
tank, could not be re-welded owing to its very short length.
The Cryot coupling ring was successful as a last resort-
repair and was adequately qualied for this unique appli-
cation. It is important to note, based on lessons learnt, that
properly welded fluid connections are inherently more
reliable than mechanical ttings and weldments should be
incorporated in fluid propulsion designs whenever possible
4.18 Shape-Memory Alloys for Spacecraft Devices
Fig. 4.73 Nitinol-deployed
lattice mast
variance only accepted when accompanied by appropriate
risk analyses have been assessed (Gilbrech et al. 2005).
The Solar-Array Drive Mechanism (SADM) for the
Hubble Space Telescope (HST) was built with release trig-
gers made of Nitinol (Fig. 4.76). 60 elements of Nitinol were
procured and submitted to metallurgical characterization
(each, having a diameter of 10 mm and a length of 56 mm,
produces a force of more than 450 N over a full stroke of
6 mm). Once the HST was in orbit the trigger elements were
successfully activated by electrical-resistance heating to a
temperature of 115 C (this particular alloy contained a small
addition of cobalt in order to slightly raise the transition
temperature and avoid the possibility of a premature acti-
vation caused by solar heating).
Shape-memory alloys have been proposed for a wide
range of future spacecraft applications. Nitinol alloys are
nonmagnetic, have a good corrosion resistance, and are
thought to be compatible with hydrazine fuel so they could
199
be employed in reaction control systems. As for the previ-
ously described HST application, release triggers made from
Nitinol avoid the complex circuitry associated with the safe
keeping of pyrotechnic devices (particularly when there is a
danger of lightning strikes) and, when activated, they do not
transmit shock loads to the spacecraft. However, it is difcult
to accurately predict the performance of designs involving
shape-memory alloys; basic data are lacking and a path of
trial and error under the guidance of a metallurgical engineer
may have to be followed in order to achieve a reliable
product. Predictions for service life data may need to be
generated in the laboratory. For instance, if the Nitinol
devices are to be used repeatedly (i.e. about 20 times) it has
been seen that stresses should not exceed 150 MPa and
strains need to be kept below 3 % so that degradation by
permanent setting will be avoided (Luciano and Galet 1995).
They withstand 10 to the power 5 times of deformation with
a 2 % strain and 10 to the power 7 times of deformation with
200
Fig. 4.74 Deployable nitinol
strut
a 0.5 % strain according to Japanese sources. A similar
non-explosive separation device is being developed in South
Korea for nano-sats (Tak et al. 2009).
A rotating arm that can be powered by short lengths of
Nitinol has been built at the NASA Lewis Research Center
and installed on-board the Pathnder mission to Mars. This
is the rst multicycle actuator used in a space application and
has been described by Jenkins and Landis (1995). The arm is
connected to a cover glass and when electrically heated it
rotates to completely uncover a solar cell detector. When the
current to the wire is shut off, the wire expands and a flat
spring pulls it back to the rest position and so returns the
cover glass to protect the solar cell. Additional promising
space applications for shape-memory alloys have been
detailed by McDonald Schetky (1991), and these include
4 Spacecraft ManufacturingFailure Prevention
deployment mechanisms for protective shrouds, seals for
solid rocket booster cases, and other release mechanisms for
boosters, rocket nozzle detachment at apogee, and the
deployment of drogues and parachutes. They may also give
spacecraft landing on distant planets an improved margin of
safety if heavy landings cause damage to Nitinol legs that
might be repaired using some form of electrical heating.
Commercial applications include the use of Nitinol in air
conditioning units, orthodontic wires, eyeglass frames, and
even brassier frames. A power generation system that
incorporates Nitinol as an energy conversion element to
operate below 100 C has been designed and studied in
Japan (Anon 1995). This study concentrated on absorbing
power from low-temperature waste heat. It concluded that a
piston-type heat engine could be used for every economical
4.18 Shape-Memory Alloys for Spacecraft Devices
Fig. 4.75 Use of shape memory alloys (SMA) in high reliability
fastening applications produced by Raychem. They have been in
service for more than 40 years. Nitinol (NiTi) alloy was later replaced
by NiTiFe and NiTiNb alloys as the SMA. a Photograph of a
Cryocon; this is an electrical connector for individual wires and
cables. The gold-plated pin is mechanically secured into the left-side
socket by means of a shrunk Nitinol ring. Stranded wires are crimped
or soldered into the gold-plated pin and socket buckets, then covered
power generation. By replacing solar heating for warm waste
water, it should be feasible to design a similar heat engine
for operation in space.
Another property of shape-memory alloys concerns their
damping capacity. This can be tailored to be very high and it
201
with plastic shrink-tubing. b Photograph of a Cryot connection
between tubes having different diameters (after test programme).
c Micrograph of cross-section through a Cryot coupling made to join
a stainless steel tube to a titanium tube. Shrinkage of the Nitinol alloy
part has caused its internal knife edges to slightly deform the
thin-walled titanium tube. Hardness values are shown (Dunn unpub-
lished report). d The extract beneath, courtesy of Raychem Corpora-
tion, describes the assembly process
varies with temperature and alloy composition. With certain
Nitinol alloys, maximum damping occurs around room
temperature when the material is stressed such that it lies
within the pseudoelastic hysteresis loop. For Nitinol the time
delay of strain with respect to stress results in the dissipation
202
Fig. 4.76 a Orthogonal view of the HST SADM off-load device; the
memory alloy elements (MAE) are mounted in the bent form and
enclosed in tubular heaters. b When heated to the austenite transfor-
mation temperature, the elements (arrowed) become straight and lift the
latch
of energythis is due to the mechanism by which the
applied stress causes the material to undergo a phase change
(i.e. the stress-induced martensitic transformation, or pseu-
doelasticity). Damping devices constructed from these alloys
may be important for large space structures that support
4 Spacecraft ManufacturingFailure Prevention
delicate instruments and telescopes needing stringent
pointing accuracies and vibration control (Thomson et al.
1995). Also, the international Space Station Alpha may
require large vibration-free truss-like structures for space
manufacturing and crystal growth.
4.19 Foamed Aluminium for Damping
Purposes
In the previous section reference was made to the need for
materials having good damping properties. Structural
damping materials based on foamed aluminium and its
alloys are commercially available for use in aeroplanes,
hotels, and industrial buildings where there is a need for
sound absorption. The same foamed aluminium products
seem attractive for space applications, as they combine
properties such as low cost, ultra lightness, good mechanical
properties for structures, high impact energy absorption,
high noise control efciency for instance during launch
operations, good electromagnetic shielding, isotropy, and
easy methods for fabrication. In addition, these materials
could be considered for manned spacecraft applications, as
they are non-flammable and not toxic.
The foamed aluminium materials have been produced by
two basic preparation methods. The rst uses a foaming
compound, such as titanium hydride, which is mixed into the
molten aluminium. The compound decomposes in the melt
and releases large volumes of hydrogen-containing bubbles.
An alternative method is to heat a precompacted mixture of
powdered aluminium alloy with granules of a foaming
compound. The compaction can be heated within a hollow
mould that has a shape and dimensions similar to those of
the end item. As the temperature reaches the melting point of
the aluminium matrix, the compound decomposes and
expands the preform into a porous cellular solid. Such foams
can even be produced in situ between sheets of any metal
that has a higher melting point to produce a strong sandwich
structure. Production and property investigations into this
new branch of materials have been reported by Degisher and
Simancik (1995).
The apparent density of these foamed aluminium prod-
ucts varies from 0.53 to around 0.8 g/cm3. The actual value
depends on whether the base metal is pure aluminium or an
alloy such as Al-6061, AlSil2, or AlMg5. A very new
material called Alulight has been produced with open
porosity; it has a density as low as 0.48 g/cm3. This product
is under evaluation for space use. The open pores allow gas
to escape during ascent into space (see Fig. 4.77). The sound
absorption coefcient of the foam has been measured in
impedance tubes with diameters of 99 and 29 mm, according
4.19 Foamed Aluminium for Damping Purposes 203

4.20 Superplastic Forming and Diffusion

Bonding of Metals

4.20.1 Forming of Propellant Tanks

Superplasticity is a rare ability of the solid, crystalline metal

Fig. 4.77 Cross-sectional appearance of aluminium alloy having an
open porosity and density of 0.48 g/cm3
to the test method specied in DIN 52 215 (incidence angle
0). A comparison of the sound absorption coefcients of
foamed aluminium with open cells with that of bulk alu-
minium, polyurethane foam, and glass bre mat is given in
Fig. 4.78. It appears as though sound waves enter the open
pores of the material, the acoustic waves are then partially
reflected by the walls of the pores, and this causes them to be
neutralized by interference. This sound absorption has been
found by experiments to be more effective for ne porosity
and increases as the degree of pore interconnection
increases.
to flow with the characteristics of hot polymers or glasses.
Only a few metals and alloys behave in this manner. Small
tensile, compression, or torsional forces produce ductility
well in excess of conventional behaviour, and without the
effect of localized necking. The phenomenon is specically
associated with eutectic or eutectoid alloys having a very
ne grain size. Commercial processes have been developed
which utilize the superplastic properties of metals similar to
the way that polymers are fabricated into useful items. This
is by blow moulding and vacuum forming the metal at an
elevated temperature. The current superplastic alloy main-
stays of the space industry are the titanium alloys and
specically Ti6A14V (see also Sect. 5.3). Aluminium alloys
are also being superformed by stretching and forming sheet
material when heated within their plastic range. Primary
structures for spacecraft applications require high strength, a
high fracture toughness, and high resistance to stress cor-
rosion. These properties can be met with Superformed
Al-7475. For less demanding applications the alloys Al-2005
and Al-5083 can be superplastically formed into equipment
covers and electronic housings. The Cassini spacecraft is due
to be launched in 1998. The main objectives of this space-
craft and the Huygens probe it supports are to measure the
temperatures and chemical composition of the atmospheres
of Saturn and Titan (its largest moon). Superplastically

Fig. 4.78 Sound absorption

coefcient of different foamed
aluminium of 0.48 g/cm3 with
open cell structure at different
levels of adepth of the air
cavity behind the sample
(determined according to DIN 52
215) A: a = 0 mm, B: a = 20 mm,
C: a = 40 mm, compared with D:
bulk aluminium: a = 0 mm, E:
PU-foam: a = 0 mm, F: glass bre
mat: a = 0 mm (courtesy of Prof.
Degischer)
204
formed Al-5083 sheet has been selected to produce a pro-
tective cover for the sensitive instruments on board. This
will shield them from increased temperature and radiation as
the spacecraft flies by Venus, on two occasions, during the
rst stage of the mission. The formed Al-5083 cover is
chemically etched in six places to a thickness of only
0.2 mm in order to save weight.
Fig. 4.79 a External and b internal views of Ti6Al4V superplastically
formed propellant tank (480 mm diameter) after burst test. The
locations for SEM fracture surface and metallographic inspection are
indicated. No corrosive sites are seen on the internal surface after
two years in contact with hydrazine. c Pairs of SEM fractographs made
at 60 and 660 magnications showing the origin of the burst crack
and the tail end of the crack. The ductile morphology is typical of a
4 Spacecraft ManufacturingFailure Prevention
The rst European spacecraft propulsion tanks to have
been produced by superplastic forming were made in Britain
in 1975. They were designed for three-axis stabilized satel-
lites requiring long-term storage of the monopropellant
hydrazine. Expulsion of the fuel would be by means of an
elastomeric diaphragm (ethylene-propylene-diene- mono-
mer). The starting material for these tanks is Ti6Al4V sheet.
fast fracture in Ti6Al4V [locations are 3 and 5 given on (b)]. d Optical
micrographs made complementary to SEM images shown in (c). Note
necking down of tank wall at origin of burst and presence of secondary
subsurface cracks. At tail end of crack the superplastically formed
tank has a thicker wall section, but local necking is also apparent (pairs
have identical magnications)
4.20 Superplastic Forming and Diffusion Bonding of Metals
Two hemispheres are made initially by superplastic forming.
This is done at a temperature just below 1000 C, under a
pressure produced by an inert gas that presses and forms the
sheet into a hemispherical mould. After cooling, the hemi-
spheres are machined at their circumferential edges and
surface etched. Etching is necessary to remove any hard,
embrittling alpha-case that might have formed when the hot
alloy comes in contact with traces of air in the inert gas (see
Sect. 5.18.2 for etching method). The hemispheres are then
TIG welded together at the equatorial (girth) location. As
part of a qualication test programme, one post-life test
culminated with a burst test. The qualication burst stress
was quoted to be 612 MPa whereas the actual burst pressure
was 706 MPa, providing an acceptable result with a margin
of 15 %. The fracture started just outside the boss weld in the
tank shell (coinciding with the thinnest section of tank wall).
Fig. 4.80 Design of Ariane-5
hydrazine propellant tank
showing hemispheres and the
location of welds. The rationale
for selecting superplastic forming
(SPF) is given in the insert. The
initial condition of the Ti6Al4V
sheet is ne grain, solution
treated, and annealed (STA)
(courtesy of DASA)
205
The burst tank was cut in half just outside the girth weld and
submitted to metallurgical assessments, as shown in
Fig. 4.79ad.
At the present time superplastic forming is being per-
formed in Germany to produce hemispheres that are welded
together as propellant tanks for the Ariane-5 launch vehicle.
The tank design and rationale for selecting this method of
production against conventional forging and machining are
presented in Fig. 4.80, based on the paper by Muller-
Wiesner et al. (1994). Each launcher incorporates two
propellant tanks. These store the hydrazine fuel that is fed to
six 400 N thrusters arranged in two groups of three. The
thrusters deliver the forces and torque designed to control
the roll and three-axis stability during the ballistic phase of
the launch; additional rings are provided for later accel-
eration and de-spinning operations. The superplastic
206 4 Spacecraft ManufacturingFailure Prevention

forming procedure is similar to that described above;

however, prior to forming, the plates are machined into a
pre-contoured blanks. This enables the required nal
hemisphere dimensions to be achieved in one high-
temperature forming step. The process of superplastic
forming relies on exact, repeatable forming temperatures
and soak times, the full recommended processes being
detailed in an ASM Handbook (Beal et al. 2006). Arianes
propellant tanks have a wall variation from 0.9 mm in the
undisturbed shell area to 1.3 mm in the pole area and
1.6 mm at the equator. Experience has shown that the pre-
cision in wall thickness realized during hemisphere forming
is better than 0.1 mm.

4.20.2 Diffusion Bonding

Diffusion bonding may result when two clean surfaces are
pressed into close contact while held at an elevated tem-
perature. There is no macroscopic deformation of the sur-
faces, as is seen during cold welding, and often the original
interface between the surfaces disappears completely.
Industrial joining processes are becoming more reliant on
diffusion bonding and, providing the mating surfaces can be
kept free from oxides, a very wide range of metals and alloys
can be successfully joined. Protection from contamination
during joining can be achieved by providing a vacuum
environment or by using an inert or reducing gas shield. The
process is usually slow and relies on the diffusion of metal
atoms and vacancies.
Intermediate coatings can be applied as bonding agents,
and these are essential if the parent metals being joined are
metallurgically incompatible. Diffusion bonding is usually
carried out at 0.50.8 of the melt temperature. Silver is a
particularly good intermediate coating and can be applied by
electroplating or as a clean foil. Silver-plated aluminium
alloys, Al-5 Mg and Al-13Si with 45 m of silver, have
been successfully joined by heating parts to a temperature of,
respectively, 425 and 500 C, and applying a pressure of
0.6 MPa for approximately 2 h (Gomez de Salazar et al.
1988). In the space industry, high-conductivity copper has
been suitably diffusion bonded to electroformed nickel by
means of a coppertin intermediate layer. A microsection of
this connection is shown in Fig. 4.81. There is no visible
Fig. 4.81 Diffusion-bonded butt joint between copper and nickel with
a bronze alloy intermediate coating on the nickel (125 magnication)
interdiffusion between the joined surfaces, but joints have
been tested and found to break at between 220 and 250 MPa
with a 100 % joint efciency.
4.20.3 Superplastic Forming and Diffusion
Bonding in One Operation
Stiffened skin sections have many spacecraft applications.
They may be used in primary structures or supports for
equipment and space station walkways. An airframe
structure made of titanium aluminide was shown in
Fig. 4.71. Here, the ribs were superplastically formed and
in a later operation, using different tooling, the parts were
diffusion bonded together. Forming and bonding are now
possible in one operation, and have been demonstrated for
the fabrication of hollow box structures with internal
stiffeners. Figure 4.82 shows a close view of two stiff
structural panels. These designs were made from two sheets
of advanced creep-resistant titanium alloys. The panel
dimensions measure approximately 50 cm by 50 cm and
have been successfully made using Ti1100, IMI 829, and
IMI 834. The IMI 834 panels can be used at operational
temperatures of 600C and are envisaged for use in future
spaceplanes.
4.21 Cleaning of Mechanical Parts
Fig. 4.82 Two examples of creep-resistant structural panels made from
high-temperature titanium alloys by superplastic forming and diffusion
bonding in a single operation. The photographs show thin test-pieces cut
from large panels. In microsection, the diffusion bond lines are not
4.21 Cleaning of Mechanical Parts
4.21.1 General Background
This section concerns general cleaning processes associated
with spacecraft mechanical systems. More specic cleaning
processes are addressed elsewhere. Section 4.5.4 describes
the cleaning and passivation of steels, and Sect. 6.9.7
highlights some means of cleaning electronic systems.
The space industry has previously relied very heavily on
the use of ozone-depleting compounds (ODCs), but with the
207
discernible. When held at 600 C for 48 h, samples have a hot tensile
strength of 600 MPa. At room temperature the tensile strength is above
1000 MPa. After 5 days at 600 C, oxidation is noted to reduce the room
temperature strength to 750 MPa (courtesy of Dornier Space Systems)
destruction of the stratospheric ozone layer, international
agreements (in particular the Montreal Protocol of 1987) and
government regulations should see the phase-out of ODCs in
developed countries by the year 2000. In the early 1990s it
became clear that the deadlines set by the Montreal Protocol
were inadequate and far greater damage was occurring to the
ozone layer than had been earlier predicted. Consequently,
amendments were later agreed in Copenhagen and then
Brussels (in December 1992). The European Community
[now European Union (EU)] agreement in Brussels ruled
that the two major ODCs would be phased out as follows:
208
Chlorofluorocarbons (CFCs) would be cut by 85 % by the
end of 1993, and 100 % cut by the end of 1994
1,1,1-trichloroethane would be cut by 50 % at the end of
1994, and 100 % cut by the end of 1995.
It would appear from visits to European space companies
during 1996 that the majority (particularly those in Scandi-
navia) have completely abandoned the use of any ODCs.
Some companies have stockpiled trichlorotrifluoroethane
solvents, in particular CFC-113. These companies were
concerned that where mechanisms had been fabricated prior
to the 1990s with processes incorporating the use of
CFC-113 and then space-qualied, it would be difcult, or
impossible to use drop-in non-ODC solvent replacements.
Fourteen of the main European companies associated
with the fabrication and assembly of spacecraft mechanisms
were questioned about their present preferred cleaning sol-
vents. In this purposely very general survey it was impos-
sible to assess the actual volume or throughput of hardware
along the individual cleaning lines referred to by the dif-
ferent companies. However, the main observations were that
until 1990, all companies had utilized CFC-113 as the
standard cleaning medium and were in possession of costly
capital equipment designed to vapour degrease, ultrasoni-
cally clean, or high-pressure spray clean spacecraft
piece-parts using ODCs. During the early 1900s a large
number of chemical suppliers were offering non-ODC
products, but very little was ever published in the open lit-
erature by independent authorities to actually quantify the
suitability of these products. Very little technical cooperation
occurred between these companies, as they are frequently in
open competition with each other, and consequently, no
joint-sponsored activities to nd a common product
replacement to CFC-113 were ever established. This can be
observed from the ndings into the present-day cleaning
systems used by these companiesthey are very diversied
and are based on the following classes of solvents:
Hydrocarbons 3
Halogenated solvents, including 6
(HCFCs)
Aqueous and semi-aqueous 13
Alcohol and acetone 9
Exotics 2 (both plasma systems)
The HCFCs, less-damaging to the ozone layer than CFCs,
are not currently regulated by the EU but are expected to be
phased out by 2015. However, apart from HCFC-141b, it is
known that many HCFCs have poor solvent properties and
they have not all been submitted to toxicological testing, so
this has also curtailed their use.
The aerospace industry in the USA is keenly aware of the
phase-out schedule for ODCs and the tightening regulations
4 Spacecraft ManufacturingFailure Prevention
on other volatile organic compounds. In an important recent
paper the actions taken by ve of the largest US aerospace
organizations in order to replace the offending cleaning
agents are described (Patterson and Mykytiuk 1996). More
than 30 other major companies and US government agencies
also detail their cleaning applications, what was used as
cleaners and the new, alternative solvent system. It appears
that 80 % of the industry that was using 1,1,1-trichloroethane
(TCA) and CFCs has now turned to aqueous cleaning sys-
tems. However, many larger companies are hindered from
changing to water-based cleaning because of the problem of
metallic corrosion and the ability to heat the water sufciently
to ensure effective cleaning.
The Patterson and Mykytiuk (1996) paper relates how
large aerospace companies have evaluated their proposed
new cleaners before they progress into pilot plant evaluation.
The individual piece-parts of units can consist of a variety of
metal alloys, such as nickel, cobalt, stainless steel, titanium,
aluminium, and low-alloy steels, which support contaminants
including machining and tapping oils, hydraulic fluids,
drawing oils, adhesive tape residues, and dye-penetrant
fluids. A logical sequence of events is to make a preliminary
review of all cleaning media on the market from a literature
search, check material safety data, and request information
from health and safety organizations. Cleaners should not
contain chromate compounds or sulphides. They must have a
low volatile organic compound (VOC) content, be noncar-
cinogenic, and have a low toxicity together with an inoffen-
sive odour. Next, samples of the cleaners will be subjected to
compatibility testing with the metals. Nickel test-pieces will
be exposed to the cleaner to check for sulphur embrittlement.
All metals will be tested for their resistance to general surface
corrosion and stress-corrosion cracking in the cleaner.
Finally, tests will be made for rinsability of the cleaning
solution and, separately, checks will be made to ensure that
dried precipitates can be removed.
In order to implement new cleaning technologies it is
important that company management appoint one person to
oversee the operation, liaise between solvent and equipment
suppliers, and avoid internal bottlenecks in the company
(Mertens and Gessford 1995). The following are also rec-
ommended: not to wait for drop-in alternatives, as they may
not exist for any one product line, seek expert advice, pay
attention to details (especially what happens to liquid waste),
do not expect perfection on the rst run of the new system
(keep the old system running until the new one has been
tested and is operating properly), and ensure that training
and maintenance programmes are documented.
In the following sections, spacecraft mechanism parts will
be considered with respect to their surface cleanliness and
surface condition. Space systems, instruments, and equip-
ments include a vast range of items with mechanisms that
include:
4.21 Cleaning of Mechanical Parts
One-shot mechanisms
Hinged booms
Solar array primary deployment mechanisms
Antenna deployment mechanisms
Rotating mechanisms
Momentum wheels, gyroscopes, reaction wheels
Scanners, gimbals, solar array drive mechanisms
Antenna pointing mechanisms
Sensors (e.g. Sun sensor on a solar array)
Mechanical components
Electric motors, commutators
Slip rings
Gears and bearings
Tape recorder transport units
Fasteners associated with the above mechanisms.
Once cleaned, many of the above-listed mechanisms will
require to be suitably lubricated (see also Table 5.2 and
Sect. 5.11). Some of the lubricating oils with low outgassing
properties are also ODCs (e.g. fluorocarbons) and might be
replaced by lubricants listed in Table 5.2 or a recently
approved synthetic hydrocarbon, Pennzane-X2000. There is
a strong interdependence between lubricant life, surface
chemistry, and tribological factors. This underlines the
desirability for performing materials investigations and
life-tests whenever new materials or cleaning procedures are
introduced. Initial investigations need to identify promising
cleaning candidates, and this should be followed up by
conducting detailed evaluations on typical mechanisms, such
as bearing assemblies, where cleaning efciencies can be
checked in the laboratory against existing cleaning solutions.
The microscope, as well as advanced techniques for surface
analysis (see Sect. 3.2.3.6), must be employed to ensure
materials compatibility with any metals, plastics, and elas-
tomers. Some examples will be illustrated in Sect. 4.21.3.
The nal tasks are to decide whether pre-cleaning and
ultrasonic assistance are necessary, and how drying should
be optimized for the various forms of hardware under
consideration.
4.21.2 Metallic Surfaces
A mechanism surface never consists of a regular plane of
metal atoms. It will always support a thin layer (ranging
from monatomic to several microns in thickness) that forms
as a result of the metals contact with its environment.
Where the environment is wet, the metal surface may
become hydrated and this facilitates ionization. Under dry
conditions there are metal/gas or metal/vapour reactions
involving oxygen, halogens, hydrogen sulphide, sulphur
vapour, etc. to form oxidation, scaling, and tarnishing.
Should the formed layer be volatile, or discontinuous, further
209
reactions may be possible. Complex surface contaminant
layers may grow once the initial lm is exposed to industrial
atmospheres, chemical fumes, chemicals such as salts, and
human perspiration. The surface may be specially hardened.
This is achieved by heat-treatment methods such as
case-hardening, nitriding, carbo-nitriding, cyaniding, car-
burizing, or induction hardening. Alternatively, mechanical
methods can be used such as peening, and cold working by
drawing or rolling. All these hardened surfaces may contain,
or support, contaminants such as oxides, partially adherent
particles, absorbed gases, a region of mechanical damage,
porosity, residues from quenching oils, lubricants, salts, or
microscopic corrosion sites.
Neither liquid solvents nor vapour phase cleaners will be
able to entirely remove the vast numbers of contaminants
mentioned above. Certainly the CFC products, used for
cleaning before the restrictions on ODCs, were never
achieving total removal of all contaminants from the sur-
faces of space mechanisms. Recent studies have shown that
some of the non-ODCs are more aggressive than the CFCs
they are replacing. Aggressive detergent cleaners can strip
away additive surface pre-treatments and oxide lms, and
this factor has produced early failures during wear tests
(Didziulis 1994). What is needed must be a reproducible
level of cleanliness. The CFC-113 is known to leave a car-
bonaceous residue on ball races after cleaning. A similar
layer has been noted after mild detergent cleaning so that
these cleaners might be considered to be interchangeable.
Another report (Ward 1994) conrms that surface chemistry
can be detrimentally changed by aggressive cleaning meth-
ods. 440C steel bearing races with silicon nitride balls with
no retainer (the ceramic balls rub against each other with no
separation) when cleaned with the old ODC system ran
satisfactorily. When cleaned in a new ODC-free solvent, the
bearings torque level rose appreciably and the ceramic balls
were covered in scratches. Indications were that the ODC
system left behind a submicron surface layer lm of tramp
oil, and the new solvent cleaner was more effective than
expected.
As a nal consideration, it has been shown that small
contaminant particles on a dry surface are primarily held in
place by strong van der Waals forces. When the surface of an
article is immersed into a liquid, such as a solvent, the total
force of adhesion is signicantly reduced and the particle will
be expected to be release from the surface. Ultrasonics and
other mechanical agitation can remove the particles far from
the surface. However, if the concentration of particles in the
liquid is high (i.e. the cleaning bath solvent is in need of
renewal) some will remain within droplets on the surface when
the article is withdrawn from the cleaning bath. The residual
particles will be held on the surface by predominantly capil-
lary forces as a liquid bridge is formed between the particles
and surface. This same phenomenon of capillary action also
210
occurs in high humidity. Whenever possible, emphasis should
be placed on the prevention of particle deposition rather than
counting on their subsequent removal. Careful handling of
space equipment is essential, storage must be in control
environments (sealed bags, containers, inert atmospheres,
etc.), and the tasks of assembly and integration must be
undertaken in a specially designated cleanroom.
4.21.3 Cleaning of Individual Parts
The most effective cleaning processes are those performed
on individual piece-parts. Whenever possible it is strongly
recommended that mechanisms are completely disassembled
and that the individual parts having the same chemical
composition are placed into different glass containers. It may
be necessary to use different solvents for the different groups
of materials. Certainly special thought must be given to the
hardened steels from bearing assemblies, as they may rust; to
the magnesium housings, as they may produce copious
volumes of corrosion product, and to all the organic mate-
rials, as these can swell and become embrittled by many of
the common solvents. Breaking the mechanism down into
individual parts enables the cleaning media to contact the
entire surface, there is no danger of galvanic corrosion, and
less likelihood of crevice corrosion. Additionally, the sur-
faces can usually be easily dried and then visually inspected
for residual contaminants such as drying stains, grease,
particles, and corrosion products.
The method and equipment used for the cleaning opera-
tion will need to be assessed by making tests on represen-
tative samples. Demonstrations are possible at most
suppliers of automated equipment. Manual cleaning using
brushes, possibly with ultrasonic enhancement, is frequently
made when the number of parts is few.
The cleaning process may be performed in different parts
of the factory. Cutting oils may be cleaned from parts in the
machine shop. Possibly the approach to cleaning in this area
may be: detergent spray, water rinse, and alcohol drying.
Special areas may be designated for the ne machining of
beryllium parts due to health and safety regulations. Here,
the beryllium parts may be ultrasonically cleaned in alcohol
and blow-dried (for this material see also Fig. 5.15 and
Sect. 5.7 as particle detection using Millipore lters during
cleaning may be desirable, together with a nal chemical
etching of the surface). Parts may be withdrawn from stores,
where they may have been packaged in corrosion-resisting
oils and greases. On entering the cleanroom, halogenated
solvents may be specied for cleaning off these oils and
greases before these parts are further processed and assem-
bled into a mechanism. Additional storage could be in sealed
containers lled with dry nitrogen gas.
4 Spacecraft ManufacturingFailure Prevention
The importance of performing materials compatibility
testing with potential solvents at an early stage in a space
programme cannot be overstated. Some companies offer
training for solvent cleaning, material compatibility testing
and test kits, such as MaxicheckTM for process optimization.
An overall company cleaning plan should be devised by
personnel from manufacturing, Quality Assurance, and the
materials laboratory. The unique nature of any one companys
product necessitates actual tests to be performed in-house on
identical materials as will be selected for the space hardware.
Where steels will be heat-treated to produce a certain hard-
ness, only that heat-treated condition must be tested. If one
chemical conversion coating is applied to the end item it is not
appropriate to test a similar coating as there can be signicant
differences in the surface compositionthe acceptance of test
results by similarity is often inappropriate.
The author has been involved with some compatibility
testing, the results of which can be used to illustrate some
important ndings. The common solvents utilized by Euro-
pean space mechanisms companies were listed in
Sect. 4.21.1. A limited number of tests have been made on:
Hydrocarbons
All of the major petrochemical companies provide
benzene-based solvents for cleaning. The metals selected for
spacecraft parts manufacturing are generally compatible with
pure benzene. Some rusting of ferritic steels has been seen,
but this was possibly due to a slight water content in the
hydrocarbon. An exotic cleaning method using distillable
hydrocarbons is shown in Fig. 4.85.
Some plastics and most synthetic rubbers swell and dis-
solve in benzene.
All hydrocarbon solvents will present some re risk.
Some are highly flammable (having a flash-point less than
32 C) and are subject to specic controls. Most proprietary
formulations on offer have flash-points of around 45 C or
above. All need to be kept in re-resisting stores. Particular
care must be taken not to spray flammable solvents onto hot
or electrically live equipments.
Halogenated solvents
An enormous number of cleaning products and associated
trade names exist on the market.
All metals need to be assessed. Two of the more common
solvents are trichloroethylene and dichloromethane (also
called methylene chloride). Long-term storage tests
(30 days) have shown that type 304 stainless steel, alu-
minium alloy 2219, and titanium alloy Ti6A14V are com-
patible with both the liquid and the vapour phases of these
solvents. Low-alloy steel (AISI 4140) corrodes in both liquid
solvents and dichloromethane produces voluminous oxides
on the magnesium alloy MgZE41A, as shown in Fig. 4.83.
4.21 Cleaning of Mechanical Parts 211

Fig. 4.83 Results of

compatibility testing MgZE41A
alloy with dichloromethane
(methylene chloride) showing
voluminous oxide growths
(courtesy of Mr. D. Bagley)

In order to improve quality, sustainability and safety, most
aerospace companies have researched alternative cleaning
methods to the open top degreasers. In the past operators had to
wear breathing apparatus and they were using signicant
volumes of chlorinated solvents which lead to worryingly high
emissions of Volatile Organic Compounds (VOCs). Nowa-
days cleaning solvents such as trichloroethylene are used in
modern, closed loop systemsthese ensure VOC
emission-free cleaning. Sizes of such degreasing machines
range from small units, up to about 3 4 m, and they are able to
clean a mass of about 700 kg of aluminium in one cycle. Many
units are fully automated and include high pressure spray
washing, vapour degreasing and nally, drying under vacuum.
Non-metals such as sintered PTFE tape and aluminized
Teflon were hardly affected by long-term exposure to tri-
chloroethylene or to dichloromethane. However, epoxy resin
(Scotchweld EC 2216) sheets and silicone rubbers (e.g.
DC93-500) generally show appreciable weight gains and
swelling, together with a reduction in mechanical properties.
Polyurethane sheet (Solithane 113) disintegrated catas-
trophically both during and after exposure to both solvents
(see Fig. 4.84).

Fig. 4.84 Polyurethane samples

are totally embrittled after
exposure to dichloromethane
(methylene chloride) (courtesy of
Mr. D. Bagley)
212
These groups of solvents are not normally recommended
for in situ cleaning and should not be used in this way
without adequate precautions to control exposure to the
operator. They are all volatile (although the degree varies)
and access to the degreaser and supplies should be controlled
to prevent unauthorized use.
Aqueous and semi-aqueous
Aqueous cleaning systems are frequently used for cleaning
mechanisms in Europe. They rely on certain important
components: the surfactants (such as emulsiers or splitters)
and their concentration, the water quality (such as whether it
is de-ionized, ltered, or puried by reverse osmosis) and the
method of drying (e.g. dry, hot air, or air knives or various
methods followed by a nal rinse in alcohol and air drying).
Various detergents may be added to the water.
Ferritic and martensitic steels may suffer from surface
corrosion during cleaning (particularly 440C and 52100
bearing steels), and as mentioned earlier, the presence of a
strong detergent can activate surfaces and accelerate corro-
sion. Most other metals also will tarnish (oxidize) and cor-
rode in aerated water. All metallic materials must be
evaluated in the aqueous cleaning line, with detail given to
the duration of the process, its temperature, and the drying
method. Part designs should be inspected to ensure than no
blind overlapping surfaces exist, such as capillary gaps,
which could entrap these solvents, so causing corrosion
during later storage periods.
Plastics and rubbers are usually unaffected by immersion
into water. Water can permeate through most polymers, and
this must be considered when the polymer is present as a thin
layer on the surface of a corrodible metal part. Baking, or
post-conditioning the part under partial vacuum, is some-
times used to ensure that any absorbed water is removed
from spacecraft organic materials.
An interesting report that documents the actions taken by
a large engineering company in replacing the cleaning sol-
vents CFC-113 and trichloroethylene with aqueous solutions
has been prepared for the US Environmental Protection
Agency (EPA 1994). These company-cleaning changes have
not only complied with the environmental goals of the
Montreal Protocol, but they resulted in improved cleaning at
dramatically reduced costs. All cleaning steps were identi-
ed and converted under the supervision of a team of per-
sons involving the company vice-president, the production
manager, the environmental ofcer, quality personnel, the
metallurgist and chemist, and the purchasing department.
Alcohol and acetone
Alcohol (isopropyl alcoholIPA) is a major cleaning sol-
vent. It is usually used in conjunction with de-ionized water
or an azeotrope of cyclohexane. IPA is often used as a rinsing
agent following aqueous cleaning. It can be easily dried in
4 Spacecraft ManufacturingFailure Prevention
warm circulating air and achieves a spot-free surface. Ace-
tone is seldom used because it can slightly attack many
materials and, on drying, leaves some stain marks. These low
flash-point solvents require to be housed in properly designed
equipment; they should be carefully handled and may need to
be regulated by health and safety inspectorates.
Pure IPA does not attack any of the metals employed in
spacecraft applications. Most pure alcohol contains about
0.1 % water and this water content is sufcient to passivate
aluminium and other alloys without producing corrosion
problems. When completely free from water, and at its
boiling point (for IPA this is 82 C), very pure alcohol will
attack the surface of aluminium.
Non-metals are generally immune to any room tempera-
ture reaction with IPA. Owing to its low surface tension, it
reduces spotting on optical surfaces. At elevated tempera-
tures many plastics and rubbers will swell when immersed in
alcohol or acetone.
Exotics
Plasma cleaning technology is occasionally used where
pristine surfaces are needed prior to surface treatment and
bonding processes. The method produces little or no waste
products. Special equipment is required. Plasma is produced
by applying a high-frequency AC voltage to gases, or a
mixture of gases, under partial vacuum conditions (below
1 Torr) inside a processing chamber.
Plasma cleaning can be applied to the majority of metals
and alloys whenever organic contaminants need to be
removed from sensitive surfaces (Ries 1996; Ward 1996;
Engemann 1996). Enhanced cleaning rates are being eval-
uated; they produce spot-free surfaces and improve the
bonding properties of metals.
Precision optical glasses can be plasma cleaned. This
technique may also be useful for removing the enamel
insulation from the surfaces of ne copper wires prior to
soldering processes, or to plasma etch machine parts made of
materials such as Delrin and Teflon into complex patterns.
Many other exotic cleaning methods are nding appli-
cations as replacements to ODS cleaning technologies.
Chambers that can be lled with solvent and then pressur-
ized whilst the part is rotated and subjected to ultrasonic
vibrations have been developed. Figure 4.85 illustrates one
piece of equipment where the solvent cleaning can be per-
formed under either pressure or, in the vapour phase, partial
vacuum. The solvent is then drained from the chamber and
the part then dried under partial vacuum. Another method,
which seems to be nding new uses in the USA, incorpo-
rates supercritical fluids (SCFs). Carbon dioxide is com-
monly used at very high pressures (above 73 atmospheres)
and at 31 C. Here, the highly compressed gas becomes a
fluid, but it has the same diffusivity of gaseous carbon
dioxide. The liqueed gas is very effective for the cleaning
4.21 Cleaning of Mechanical Parts 213

Fig. 4.85 Sketch illustrating a novel method used by the space
industry for cleaning mechanical parts. A closed treatment chamber
contains the parts. The cleaning fluid can be one of several distillable
hydrocarbons with flash points between 55 and 100 C. The solvent
initially sprays the parts (which can be rotated) at high pressure. They
are then vapour rinsed in a partial vacuum at 80 C and nally dried in a
partial vacuum of about 5 mbars. The solvent is continuously ltered
for particles larger than 1 m and regenerated in a vacuum still.
A modied system is suitable for cleaning PCB assemblies and
removing rosin flux and solder pastes (see also Sect. 6.9). (By
permission of Branson Ultrasonics B.V.)

of metals contaminated with oils and greases. It is probably

incompatible with many organic materials. The liquid can be
sprayed onto the parts, called snow cleaning, and will reach
into blind holes.

4.21.4 Cleaning of Metallurgically Joined

Assemblies

Several space mechanisms cannot be disassembled into their

individual piece-parts. This may be where the parts have
been welded together in order to ensure that rotating surfaces
are separated by a xed, unalterable distance. Other dis-
similar metals may be joined by welding, brazing, and (in-
correctly) soldering for structural purposes or to produce
hermetically sealed devices. From experience, soldering with
soft alloys does not reliably full these requirements.
Whenever assemblies are cleaned, care should be taken to
ensure that the entire unit is compatible with the cleaning
process and the selected solvent. In a recent assessment of
cleaning procedures used by European spacecraft contrac-
tors, a standard test-piece was designed to represent a worst
case mechanical device fabricated from different metallic
materials (Korb and Dunn 1996). The test-piece is shown in
Fig. 4.86. It was designed to create conditions for galvanic
corrosion, surface corrosion, galling of screw threads, blind Fig. 4.86 Assembly of dissimilar metals used to assess the efciency
of cleaning processes
214
Fig. 4.87 Photograph of one
disassembled test-piece, after
cleaning, photographed under
UV-light (marking is 5)
holes, and capillary gaps for fluid retention and is con-
structed from typical spacecraft alloys. The blind holes have
a diameter of 1 mm and a depth of 4 mm, and were drilled
into the steel base plate to represent pockets that could trap
contaminants. One hundred test-pieces were manufactured to
high dimensional and surface nish tolerances. They were
individually numbered. An articial contaminant was con-
cocted using a base consisting of cutting oil (Blasocut 200
CF can be emulsied with water) which was mixed with 2 %
4 Spacecraft ManufacturingFailure Prevention
silicone oil (typical for space lubricants), 2 % fluorescent
powder (Fluoflux, used to represent fluorescent
crack-detection particles), and 4 % oil-soluble analytical
tracer marker (Scandium-46 tracer element, in standard
solution of 100 g Sc/g). A standard technique was devel-
oped to contaminate the individual piece-parts. They were
assembled within the slurry (homogeneous suspension) of
contaminants by means of the bolt and nut, and then tight-
ened to a dened torque of 8 Nm. The lock wire and lead
4.21 Cleaning of Mechanical Parts 215

Fig. 4.88 Photograph showing

different xylene
solutions/suspensions from
after-cleaning

seal were applied to ensure that no disassembly was carried Table 4.14 Residual contaminant values (powder and oil) on cleaned
out when the test-pieces were later cleaned by the different test specimen
participating companies. Special containers were used to Residual contaminant analysis
package and ship the contaminated test-pieces to the differ- Cleaning technique Powder (mg) Oil (mg)
ent companies. New containers were supplied for returning A1 0.16 4.8
test-pieces. On return to the test laboratory a wide range of A2 0.62 50.7
inspections were made. These included visual inspections at
B 1.80 100
magnication for damage, particles, and surface corrosion.
C 0.07 6.7
The lead seals were broken and the nuts were opened with a
torque wrench, noting the various torque values. The D 0.95 21.1
fluorescent powder could be inspected and photographed E 0.05 7.4
under UV-light in order to assess the amount of residual F 0.14 8.1
powder contaminants within the blind holes and at contact- G1 0.48 40.7
ing flat surfaces (see, for instance, Fig. 4.87). The particles G2 2.44 273
and the oil (containing the scandium tracer) were collected H 0.31 2.2
from all surfaces and blind holes of the disassembled parts
I 0.05 2.2
by immersing them in xylene contained in Teflon beakers.
J 0.30 16.7
Complete cleaning was established using mechanical and
ultrasonic agitation. The amount of residual powder could K 0.80 67.8
again be assessed as it formed a suspension in the xylene L 1.73 61.5
giving rise to colorations that differed owing to powder M1 2.63 114.8
concentration (see Fig. 4.88). The amount of organic con- M2 2.66 219.3
taminant (oils) contained in the xylene was determined by N 0.13 4.4
analysis of the scandium-46 tracer. Here, the samples were O 0.04 0.7
freeze dried and irradiated in the core of the ASTRA nuclear
P 0.03 3
research reactor at Seibersdorf, Austria. The irradiation of
Q 1.67 56.3
each sample consisted of 20 h under a neutron flux of 3.10 to
the 13 n/cm2 s. After 4 weeks of delay the samples were R 0.18 6.3
measured with a germaniumlithium detector and multi- S 1.50 40.7
channel analyser enabling an accurate assessment of the T 0.80 104.1
residual oil contaminant. The results of the residual powder U1 0.09 6.7
and oil contaminants found during the study are shown in U2 0.04 4.8
Table 4.14.
216
The ndings of the round-robin cleaning study were
many and included the following:
all participating companies utilized different cleaning
methods (25)
from the rst view, all returned test-pieces appeared
visually clean
after disassembly, residual contaminants could be clearly
seen under UV-light
the amounts of residual contaminants vary by 34 orders
of magnitude (reflecting the great differences of ef-
ciency between the cleaning techniques)
none of the steel parts or titanium parts had corroded
aqueous alkaline cleaners and also plasma cleaning pro-
duced some surface corrosion on the aluminium alloy as
seen in Fig. 4.89 (other cleaning methods had no effect)
the molybdenum disulphide surface lubrication layer on
the titanium bolt was unaffected by al the cleaning
methods except that performed by plasma techniques
the cleaning process (e.g. equipment design, time of
cleaning, use of pre-cleaning steps, use of ultrasonics,
etc.) is more important than the selected solvent
most of the non-ODS cleaning techniques are as good as
the reference CFC-113 methods, particularly when
ultrasonic agitation is applied (Fig. 4.90).
Fig. 4.89 Assembly Nr.
14 (Code C) showing a detail of
the surface-corrosion residue
present after aqueous cleaning on
the aluminium part (mainly
aluminium hydroxide)
4 Spacecraft ManufacturingFailure Prevention
4.21.5 Maintenance of Cleanliness
The title of this chapter refers to metallography as it is
applied to spacecraft production. The numerous sections
have promoted metallography as a technique that can use-
fully control and maintain a high level of quality during the
manufacture of spacecraft mechanisms and structures. Sim-
ilar aspects, but directed only to spacecraft electrical sys-
tems, will be detailed in the following chapter. The topic of
cleaning has also been covered. Once it has been decided
that a particular piece-part must be cleaned (probably
immediately after a dirty production process) there will be
a need to check that the cleaning procedure is adequate for
the part and that all subsequent handling operations are
performed in a correct manner (possibly using lint-free
gloves) within the connement of an environmentally con-
trolled room. Company handling procedures and facilities
will need to be scrutinized by some form of audit. Fig-
ures 4.91 and 4.92 show the production and test facilities for
large cryogenic tanks that are installed onto conventional
launchers. Special attention is paid to the welding techniques
when these large, thin, curved metal shells are processed and
to the manner by which they are handled and later submitted
to a combination of mechanical and thermal loads during
performance testing (Pradier and Dosio 1996).
4.21 Cleaning of Mechanical Parts 217

Fig. 4.90 Assembly Nr. 20 (Code D) showing a portion of the titanium fastener thread. The MOS2 layer is removed in large areas. It only remains
where it has been protected by the nut

Fig. 4.91 a Spacecraft assembly area for welding large, thin-walled shells of aluminium alloy AA 2219. b One welding jig being set up for VPPA
welding of an Ariane liquid propellant booster tank (courtesy of Alenia Spazio Turin)
218
Fig. 4.92 Thermomechanical
test facility (courtesy DLR)
The gures illustrate facilities that are not classied as
cleanrooms. These are locations in the factories where
good housekeeping and a tidy appearance is necessary
during spacecraft fabrication and test. Dirty, untidy facil-
ities are inherently detrimental to the end-product. Dirt can
be anything that settles on the part from air, such as dust,
lint, or other matter. Numerous examples have been
shown in this book where particles, ngerprints, oil, and
the like, have promoted or caused the failure of expensive
parts.
The maintenance of cleanliness in a spacecraft production
facility will necessitate that different tests are performed
these are dictated by the nature of the contaminating dirt.
4 Spacecraft ManufacturingFailure Prevention
Airborne particles are assessed in units of total particle count
and distribution of particle size. Much must be known about
the dirt: whether it is a soluble or insoluble contaminant.
Ionic contamination lends itself to electrical measurements
these are based on the change in conductivity (or resistivity)
of a polar solvent after extraction of the dirt from the parts
surface. It may be trapped in hard-to-reach places where
dynamic cleaning methods incorporating the use of ultrasonic
cleaning may be required.
The cleanliness of all piece-parts will steadily deteriorate
after processing. Unless the part or assembly is sealed,
contamination build-up continues until the time of launch.
Gross contamination is obviously a hazard. The parts and
4.21 Cleaning of Mechanical Parts 219

mechanisms may be immune to small dirt levels but it is the
responsibility of the manufacturer to know or study what the
limits are for his own assemblies.
4.21.6 Cleaning of Silicone Contamination
Section 5.15 will describe individual case studies where
silicones have contaminated space hardware, and offer
advice about how such contaminants may be removed.
Silicone products used for spacecraft range from oils and
greases to cured products as used for sealing materials,
adhesives, paints, or conformal coatings. They can offer
excellent properties, e.g. low vapour pressure and low sur-
face tension (oil lubricants), high thermal conductivity
(thermally conductive vacuum greases), chemical inertness,
good dielectric properties, and ductility (conformal coatings
and paints). Silicones on the other hand can act as major
sources of contamination if they are incorrectly handled or
used. Their low surface tension will cause oils to creep over
clean surfaces, spreading contaminations into gaps and other
areas that are almost impossible to clean. Under vacuum,
such trapped reservoirs can act as a long-term source for
further contamination. Therefore, a thorough removal before
launch is mandatory.
A method for assessing the efciency of cleaning silicone
contaminated surfaces has been developed based on contact
angle measurements (Liedke 2011). Glass slides deliberately
contaminated with a standard silicon oil solution were used
together with FTIR spectroscopy as a reference. This solu-
tion was: 0.5 g DC 200 silicone oil in 49.5 g dichlor-
omethane (Merck, p.a. grade). It was found that the FTIR
method was insufciently accurate (requiring at least a
2.4 102 m thick silicone contamination on the surface).
The study found that the surface angle measurement most
appropriate for (semi-)quantitative detection of very thin
surface layers (the sensitivity seems sufcient for safely
detecting a layer of only 5 103 mmaking the contact
angle method at least 5 times more sensitive than the FTIR
method). However, the disadvantage is the substantially
higher effort needed for surface angle measurements com-
pared to FTIR measurements.
The second part of the Liedke study was to contaminate
ultra-clean glass slides with a range of common spacecraft
silicone products. Each product was applied to an individual
slide by smearing, using the edge of another slide, so as to
achieve a very thin coverage. The contaminated test slide
was then assessed for cleanability by exposure to a range
of commercially available cleaning fluids. An attempt was
made to standardise the cleaning methodlens tissue was
immersed into the cleaning fluid, the tissue was then stroked
across the contaminated slide ve times. After this cleaning
process the slide was dried and submitted to the contact
angle measurement test. The following contaminants were
applied to slides; the cleaning agents are also listed:
The following silicone contaminants, representing typi-
cal species that may come into contact with spacecraft, were
been selected:
Poly-dimethylsiloxane oil (Dow Corning DC 200)
Pentaphenyl-trimethyl-trisiloxane oil (Dow Corning DC
705)
Mould release agent (330 LV Silicone Emulsion from
Basildon Chemicals)

Table 4.15 Summary of test results

Designation 330 LV DC 200 DC 705 DC 93-500 rub DC 93-500ext DC 340
Methylene chloride ++ o +
Toluene ++ o + + +
n-Hexane +
Fluxclene ++ ++ ++ ++
Silgest SD0001 o + + + + +
Ortimex + o o
Safewash Super + o + + +
VeriClean o + + o +
Vertrel CCA + +
Vertrel CMS + o
CH2Cl2/Toluene 1:1 + o
CH2Cl2/n-Hexane 1:1 ++ o + + +
Dowclene 1601 + o o
Dowclene PX-16S o
++ means good cleaning, + reasonable, O remaining contamination, poor, and means unsuitable (it will be essential to make further tests on
replicate samples of contaminated hardware because solvents can damage a wide range of organic substrate materials)
220
Cured silicone resin (DC 93-500 from Dow Corning)
applied by rubbing on substrate
Cured silicone resin (DC 93-500 from Dow Corning)
applied as an extract in tetrachloroethylene
Thermally conductive vacuum grease (DC 340 from
Dow Corning).
Cleaning agents (most are dangerous substances and
should be only used by trained professionals, some may have
REACH restrictions):
Methylene chloride CH2Cl2, (from Merck)
Toluene C6H5CH3, (from Merck)
n-Hexane C6H14 (from Merck)
Fluxclene, (manufactured by Electrolube)
Silgest SD0001, (distributed by Polymer Systems Tech-
nology Ltd.)
Ortimex Alkaline Cleaner, (distributed by H.C. Stehelin)
SWAS (Safewash Super), (manufactured by Electrolube)
Vericlean defluxer (from Microcare)
Vertrel CCA (DuPont, distributed by Microcare)
Vertrel CMS (DuPont, distributed by Microcare)
Dowclene 1601 (based on 1-Methoxy-2-propanol) from
(Dow Chemicals)
Dowclene PX-16S (from Dow Corning)
The test results are somewhat subjective because not all
process variants were strictly controlled (time between
contamination and cleaning, room temperature variations
and cleaning tissue pressure was operator dependant), but
the author considers the compilation presented in Table 4.15
should be a good starting point for further evaluations.
4.22 Novel Thermal Management
Materials
Thermal management systems are used to maintain space-
craft hardware within acceptable temperature ranges. Exist-
ing passive and active systems that achieve this, are
described in other Sections. New developments in materials
will ensure that heat flow away from hot spots, or heat flow
towards cold spots (such as mechanisms located in the
vicinity of cryogenic tanks) is via passive materials rather
than heat pipes, radiators and resistance heaters. In the
absence of an air stream, thermal management within space
systems is best made by the incorporation of
high-conductivity materials. Materials under development,
or being qualied, include composite materials and
heat-spreader substrates manufactured by novel processes.
Copper is the most common spreader, but more efcient
materials will incorporate diamonds, silicon carbide and
aluminium nitride. Electronic circuits based on silicon are
4 Spacecraft ManufacturingFailure Prevention
now back-side coated with 5 m thick diamond lms giving
better than 400 W/mK conduction at room temperature.
Similar lms, applied to the circuit-side of IC are used to
provide space electronics with partial radiation hardening.
Nano-tubes and nano- wires based on graphine are also
proposed for thermal transport, based on their inherently
high thermal conductivity (Schelling et al. 2005).
Metal-graphite foam and sheets are manufactured commer-
cially for heat-sinking. The graphite bre Thermal Strap
assemblies touched on in Sect. 2.5c) are convenient,
light-weight and have been supplied for US spacecraft
including the Orion multi-purpose crew vehicle. The
assemblies use GraFlexTM bundles (made from graphite
bre with a thermal conductivity of about 800 W/mK).
Overall, the special thermal link assemblies can be consid-
ered to have a conductance of up to 5 W/K and are used in
near-cryogenic systems, satellite electronic boxes and power
electronic systems, as seen in Fig. 4.93.
The high conductivity of metal substrates makes them
attractive for thermal management of electronic devices.
However, in many applications it is necessary to provide
electrical insulation to the thermal plane to prevent shorts.
While this is readily accomplished by a polymer dielectric,
the relatively poor thermal conductivity of even
ceramic-lled resins represents a signicant barrier in the
thermal path, even when the layers are relatively thin. One
solution to this problem is to use a pure ceramic dielectric.
At least one company has developed a process for con-
verting the surface of aluminium to nano-grained alumina
(Shashov 2013). Remarkably, this is done using an elec-
trochemical process since the resulting alumina is electri-
cally insulating, yet the coatings produced can be tens of
microns in thickness. The breakdown voltage of the grown
coating exceeds 60 V/m. The combination of an excep-
tionally thin layer of dielectric possessing moderate thermal
conductivity (7 W/mK), together with a very low interface
resistance between the aluminium and the ceramic means
that this material has the lowest through-thickness thermal
resistance of insulated metal substrates available commer-
cially at the time of writing. 3-dimensional heat sinks made
from aluminium with nano-ceramic dielectric and support-
ing copper circuits, as illustrated in Fig. 4.94, are available
and should have several space applications. By growing
alumina on both sides and capping with copper, complete
metal core printed circuit boards (MCPCB) can be
fabricated.
Diamond has a heat conductivity four times that of copper.
Measurements range from 800 to 2000 W/mK for industrially
produced diamond, and 420 W/mK for copper. The larger the
diamond grain size, the greater the thermal conductivity.
Composites of diamond grit embedded in a copper, or cop-
per alloy matrix have been produced in small volumes for use
as high current heat sinks (power devices producing
4.22 Novel Thermal Management Materials
Fig. 4.93 a Graphite bre thermal strap assemblies are light-weight,
and b have found a number of advanced thermal management
applications within spacecraft. On a thermal conductivity/mass basis
they are more efcient than copper or aluminium and, c have sufcient
flexibility to accommodate installation and alignment tolerances; they
221
operate reliably under severe vibration. These straps provide passive,
high performance heat transfer in vacuum from hot zones (e.g.
travelling wave tubes) to radiator plates. Courtesy of Technology
Applications Inc.
222 4 Spacecraft ManufacturingFailure Prevention

Fig. 4.94 Nano-ceramic technology. The ceramic alumina layer is and can be grown to a pre-determined thickness between 5 and 50 m.
grown from the aluminium heat-sink substrate, it has a high thermal Metal tracks can be applied directly onto the ceramic surface. Courtesy
conductivity (to 7.2 W/mK) and dielectric strength (to 110 KVmm1) of Cambridge Nanotherm Ltd.

Fig. 4.95 Diamond grit metal matrix composite. The diamond crystals Photomicrograph shows boron nitride (grey lower face), the
are 3060 m in diameter, matrix is an alloy of CuAg with additions copper-rich phase has preferentially etched, and the diamonds appear
of Sn and Ti to promote interfacial bonding. The consolidation was in partial contact within a plane close to the upper surface of the sheet
made on a former using hexagonal boron nitride release agent.
4.22 Novel Thermal Management Materials
kW cm2). Here, the target was to achieve a flat composite
sheet having 500 W/mK conductivity that could be utilized in
the construction of multi-layer printed circuit boards, or
alternatively, special heat sinks. Complex powder metallurgy
processes are needed to produce such compositesalthough
it should be noted that the cutting tools industry has
employed brazed diamond cutting and grinding tools using
active metal silver-copper and nickel-based brazing alloys for
many years (Tillmann and Boretius 2008). One has to con-
sider thermal expansion mismatch between diamond crystal
size and the particular alloy matrix. Of greatest importance is
the need for additives that can bond the matrix to the
almost-inert diamond facets; here, chromium, boron, zirco-
nium or titanium (all carbide formers) can be considered. The
microstructure of a metal matrix composite under develop-
ment, containing 30 % diamond grit, is seen in Fig. 4.95.
Consolidation of the powder mix and mechanical treatments
are proprietary, but some released data for the European
Project ExtreMat (Schubert et al. 2008) concluded their most
promising composite to be a copper alloy-diamond possessing
a thermal conductivity of 640 W/mK combined with a CTE of
about 11 106 K1 from room temperature to 200 C. antenna face-skin material samples made of PEEK com-
4.23 Cold Sprayed Coatings
This process is termed cold gas dynamic spraying (CGDS).
Only a limited number of materials can be presently cold
sprayed onto substrates but this deposition process has
enormous potential for specic spacecraft application. The
Fig. 4.96 Schematic of cold
spray system. A high-pressure
process gas, whether it be He or
N2 is accelerated through a de
Laval nozzle and directed onto a
substrate. Simultaneously, a
minor gas stream is fed through a
powder hopper, and the
gas-powder stream injected
axially into the main gas flow a
short distance from the nozzle
throat. The powder, in this case
aluminium, is deposited directly
onto a PEEK composite sample
located 40 mm below the nozzle
exit
223
materials are generally limited to copper, aluminium, tita-
nium and some cermets (which need to be cold sprayed with
a ductile binder). This is an alternative to electroplating and
hot, thermal spraying. CGDS can be directly applied to the
surfaces of metallic-, ceramic- and organic-nished products.
Space hardware applications will include the deposition of
metallization layers onto electrical components and electrical
grounding points on organic substrates. Specic examples
include aluminium metallization of polymer composite
structures and silver deposited as a solid lubricant for use in
space. The deposits are free of the oxide lms and porosity
usually seen in more conventionally thermal sprayed coat-
ings, and give improvement in properties such as electrical
and thermal conductivity. The lack of heat input into the
substrate during coating application makes the CGDS pro-
cess of interest for depositing metal coatings onto polymer
substrates such as antennae face skins and structural parts.
The process is explained in Fig. 4.96.
The CGDS facility at the Institute of Manufacturing,
University of Cambridge, together with TWI have success-
fully demonstrated the process for applying aluminium onto
posite (Sturgeon and Dunn 2006). The spraying was per-
formed within an enclosed chamber built on a bespoke CNC
controlled 3-axis system, with the nozzle assembly attached
to a z-axis actuator and the target substrates xed to the
translatable base. The chamber protected the surrounding
environment from any loose powder and facilitated the
recovery of exhaust process gas that was then recycled.
Helium was chosen to be the principle process gas for this
224
system because of the greater deposition efciencies and
material integrity that can be achieved compared to nitrogen,
even without gas heating. However, there is a cost penalty in
using helium because of its high cost per unit volume that is
compounded by the increased flow rates because of the
higher gas velocities. Consequently, helium gas recycling is
required to make the process commercially viable.
After coating, the samples were subjected to environ-
mental tests that consisted of holding in a high humidity air
atmosphere at elevated temperature followed by thermal
cycling. The exposure to high humidity was made in a
chamber with the humidity level set at 85 %RH and the
temperature set to 85 C. After 14 days (336 h) the coated
samples were removed and examined visually. The same
coated samples were then put through a 55 to +100 C
thermal cycle in air, cycling at a rate of 10 C/min between
Fig. 4.97 a CGDS coating of aluminium on carbon bre reinforced
PEEK composite, C3 after high humidity and thermal cycle exposure
(surfaces highlighted with side illumination). b Cross-section through
C3 after high humidity, thermal cycling and tape testing (low
particulate separation). Spraying conditions were later modied to
achieve a more uniform coating
4 Spacecraft ManufacturingFailure Prevention
these two temperatures. A total of 500 such cycles were
undertaken with the samples then removed and examined.
Such thermal cycling is often used for the verication of
spacecraft electronic components. The samples were then
subjected to visual inspection, tape testing to check for
coating delamination and microsectioning for coating mor-
phology and thickness. The sample ndings are shown in
Fig. 4.97.
The work showed that it is feasible to deposit a CGDS
aluminium coating of typically 0.5 mm directly onto a chop-
ped carbon bre reinforced PEEK composite. The processes
did required careful selection of nozzle design and spraying
conditions. However the range of carbon bre reinforced
polymer composites that can be coated with metal using
CGDS will only be known after more extensive evaluations.
In cold spraying the powders remain below their melting
point and are not oxidised. Microsectioned coatings will
reveal a mixture of mechanically interlocked and metallur-
gically bonded particlesthe kinetic energy of the acceler-
ated particles creates interfacial deformation and a limited
amount of local heating upon impact with the substrate so
effecting a bonded interface. CGDS is now nding many
applications in the aerospace industry for the repair of gear
boxes, propellers and landing gear (Marrocco and Harvey
2012). Recently cold sprayed particles include titanium, tan-
talum, stainless steels and nickel alloys with future possibil-
ities for nanophase, intermetallic and amorphous materials.
4.24 Advanced Plasma Electrolytic
Oxidation Treatment
for Aluminium, Magnesium
and Titanium Alloys
4.24.1 General Process
Plasma electrolytic oxidation (PEO) and micro-arc oxidation
(Shrestha et al. 2007; Polat 2008) are relatively new envi-
ronmentally safe electrolytic coating processes, applicable to
light metals and their alloys and represents a rapidly devel-
oping sector in surface engineering. These processes
involves the use of far higher voltages than in anodising, the
electrolyte usually consists of low concentration alkaline
solutions and all variants of this process are considered to be
environmentally friendly. Each process results in the for-
mation of a ceramic layer that offers protection to the base
alloy in terms of corrosion and wear; they also offer other
functional characteristics including thermo-optical, dielec-
tric, thermal barrier, low friction coefcient and also can be
used as pre-treatment for topcoat paints and other
metal/ceramics to create composite coatings. A number of
similar processes are being offered by research laboratories
and commercial companies around the world (Shrestha and
4.24 Advanced Plasma Electrolytic Oxidation Treatment for Aluminium, Magnesium and Titanium Alloys 225

Table 4.16 Typical process parameters during Keronite PEO and hard anodizing of Al-7075 alloy
Coating Aluminium Pre-treatment Electrolyte Total salt Typical Nominal Coating Voltage Process Coating Surface
process alloy content (%) pH thickness rate (V) temp formation appearance
(m) (m/min) (C) method
Keronite 7075 Degrease only Proprietary <4 712 1560 14 200 1230 Plasma Grey to
(AC) alkaline free of 900 oxidation charcoal
Cr, V or other black
heavy metals
Hard 7075 Degrease and Sulphuric acid 1020 <3 <60 0.81 4550 10 to 0 Oxidation Bluish
Anodising chemical clean H2SO4 without black
(DC) in alkaline plasma
solution

Dunn 2010). However, they each differ in terms of coating
deposition rate and thickness limit, applied voltage, elec-
trolyte type, electrolyte life, phase formation, process speed
and scalability. The PEO process is also of signicant
interest as it can offer improved surface properties to mag-
nesium and aluminium alloys and may be a suitable
replacement for conventional acid based processes such as
DOW and other chemical processes that contain hexavalent
chromium. It has been used as a drop-in replacement for the
anodisation of aluminium alloys where the anodising elec-
trolytes are based on chromic acid and sulphuric acid.
The Keronite is an advanced surface technology based
on plasma electrolytic oxidation (PEO). It is an environ-
mentally friendly and safe process for the preparation of
corrosion and wear resistant nishes (Shrestha et al. 2003).
These nishes comprise of ceramic layers on aluminium
alloysthey have seen increasing usage over recent years in
the aerospace and defence industries. Finished products
include those used for optical instruments because the colour
of the PEO coating can be changed from grey to black; this
is a factor of the alloys compositions. Generally with high
copper content, such as AA2024 or AA2219 the nal colour
is black. The Keronite process employs a low concentration
alkaline-neutral solution that is free of heavy metals and
contains a salt concentration of less than 4 %. The com-
monly used sulphuric acid hard anodising chemistry is far
different from that of Keronite and comprises of between 10
and 20 vol. % of acid concentration. Some parameters for
the Keronite PEO and sulphuric acid hard anodising pro-
cesses are given in Table 4.16.
Coating of aluminium alloys using the Keronite PEO
process involves the creation of a plasma discharge around a
component immersed in an electrolyte (Fig. 4.98). The
mechanism of oxide layer formation is complex, being a
combination of oxide growth with subsequent fusing and
then re-crystallisation of the oxide lm. A small amount of
the substrate is simultaneously dissolved into the electrolyte.
At the time of Keronite PEO processing it appears to be very
aggressive due to extensive plasma discharges on the surface
of the piece-part; at microscopic level the process creates
very high local pressures (believed to be several GPa).
Although very high local surface heating appears to make
the processed surface red, or white hot, only surface oxi-
dation occurs and there are no metallurgical changes to the
substrate. Metallography has conrmed that treated compo-
nents only have modied surfaces and their mechanical
properties are not changed. Inspection of processed items
show no heat affected zones or dimensional distortion. The
overall process temperature and that of the component are
typically in the range of 1230 C, so modication of any
heat treatment condition is not possible. Sophisticated lab-
oratory examinations of resultant oxide surfaces using
crystallography show these PEO nishes to contain some
elemental co-deposition from the electrolyte, creating a
ceramic layer that consists of both crystalline and amorphous
phases of alumina.
4.24.2 Characteristics of PEO Coatings
An SEM image of an as-prepared surface of the Keronite
layer is shown in Fig. 4.99a which depicts an abundance of
spherical and irregular shaped particles-like structures that
were formed during the PEO process. An as-prepared sur-
face of the hard-anodised layer is shown in Fig. 4.99b which
shows several voids and micro-porosity on the coating sur-
face. In addition, the surface clearly shows parallel lines that
form vertical through thickness cracks. A typical cross sec-
tion of the Keronite layer on AA7075 alloy at the outer
corner/edge is shown in Fig. 4.100a. The Keronite layer has
some micro-scale porosity as would be expected from the
PEO process but the coating can be considered relatively
dense. A typical cross section of the hard-anodised layer on
AA7075 alloy at the outer corner is shown in Fig. 4.100b
which shows the presence of extensive V-shaped columnar
cracks. These cracks are seen through the entire coating
thickness and extend down to the underlying substrate, and
in several areas even expose the bare metal substrate. Some
general characteristics of the Keronite coating on three dif-
ferent aluminium alloys are given in Table 4.17.
226 4 Spacecraft ManufacturingFailure Prevention

(a)

(b) Modulated electrical voltage

Electrolyte
Electrode
Sample

Spark
discharge

Cooler
Cooler Filter
Filter

Fig. 4.98 a Microprocessor controlled Keronite unit (1000 l tank, 160 kW) for processing aluminium alloys, and b schemetic of process (courtesy
of Keronite Ltd.)

The Keronite ceramic surface after fretting tests in vacuum
was noted to have very low adhesion as shown in Fig. 4.101.
Actual adhesion values are: Keronite-coated-AA2219 against
a 52,100 steel ball (adhesion force 108 mN); and
NiCr-plated-AA7075 against an anodised AA7075 (adhesion
force 242 mN). The aluminium alloy 7075 without a coating,
to itself, resulted in extremely high adhesion (adhesion
7330 mNequivalent to cold-welding) after fretting. The
Keronite-coated-AA7075 surface against itself, displayed
329 mN. Despite low adhesion values displayed by the
Keronite and hard anodised coatings against steel surfaces,
the Keronite coating shows no sign of surface damage,
whereas, the anodised coating displayed extensive cracking
and chipping of the coating surface in a fretting environment.
The differences in the damage mechanisms between the
Keronite and anodised coating surfaces are shown in
Fig. 4.101. While the friction coefcient of Keronite to
Keronite has been reported relatively high about 0.50.6
(Shrestha et al. 2003), this can be substantially reduced to an
extremely low value of about 0.04 by impregnating the
porosity (see Fig. 4.102) in the coating with solid lubricants
e.g. MoS2. The Instituto de Astrofsica de Canarias (IAC) and
Keronite have undertaken a co-operative research on cryo-
genic plain sliding bearings. Keronite + MoS2 composite
4.24 Advanced Plasma Electrolytic Oxidation Treatment for Aluminium, Magnesium and Titanium Alloys 227

Fig. 4.99 Scanning electron

images showing a sharp
edge/corner of the coating on
Al-7075: a good and uniform
corner retention with Keronite
layer; and, b wide-opening cracks
with hard anodized layer

Fig. 4.100 Optical micrographs

of a Keronite coating showing a
uniform coating coverage and
good edge retention; and b hard
anodized coating, showing
extensive V-shaped,
through-thickness cracks between
columnar oxide growth extending
to the substrate (at corner of
piece-part)

coating has been developed as a tribological surface with a
low sliding friction coefcient, maximum resistance to wear,
minimum debris generation, vacuum compatibility, resis-
tance to thermal shock and thermal conductivity better than
plastics.
Keronite PEO coatings on copper-containing aluminium
alloys such as 2219 and 7075 can be prepared black and these
have been extensively studied as thermal control coatings.
Recently there has been growing interest by industries on
black nish coatings on other aluminium alloysa PEO
process capable of producing a charcoal black nish on most
wrought and some cast aluminium alloys has recently been
developed an industrial scale. Unlike the process discussed
above, this new process converts the aluminium alloys into a
black nish and the resulting surface characteristics and
properties are not dependent on the substrate alloy compo-
sition (Shrestha et al. 2007). A typical SEM image of this new
black Keronite surface is shown in Fig. 4.102. This coating
surface has very low reflectance (and high absorptance) of
about <0.1 % (Fig. 4.103). This reflectance value is less than
 228

Table 4.17 Characteristics of Keronite PEO coatings on aluminium alloys

a
Coating type Alloy Colour Roughness Porosity Phase Hardness Corrosion Solar Thermal Reflectance UV Vacuum TSR 196
(Ra, m) (%) composition (HV) resistance, absorptance emittance (%) resistance outgassing to +100 C
(h) (S) (n) (TML%)
Keronite 2219 Grey to 11.6 <5 Crystalline 1200 >360 0.880.89 0.710.75 Pass Yes
black Type 1 charcoal alumina 1550
black
Keronite 7075 Grey to <1 <2 Crystalline 1600 >360 0.810.83 0.700.72 Pass <0.1 Yes
black Type 1 charcoal alumina 1650
black
Keronite new 6082 charcoal 11.5 3050 Amorphous 5001200 >1000 0.94 0.84 0.1 Pass Yes
black Type 2 black alumina
TSR thermal shock resistance; UV resistancegrey scale rating 4
4

a
ASTM B117
Spacecraft ManufacturingFailure Prevention
4.24 Advanced Plasma Electrolytic Oxidation Treatment for Aluminium, Magnesium and Titanium Alloys
Fig. 4.101 Surface after fretting test of a Keronite coating (with no visual damage) and b anodized coating showing extensive surface cracks
(Shrestha et al. 2003)
Fig. 4.102 Surface of a new black Keronite on AA6060 alloy showing
surface morphology with crater-like porosity
reported reflectance values for super black NiP coating and
signicantly less than that for black sulphuric acid anodised
coating and black paints.
4.24.3 Applications
(a) Keronite coatings have been extensively evaluated for
spacecraft material applications as thermal control
coatings, and as bore seals (against Torlon and Vespel
SP3 pistons) for cryogenic bearing applications. Vari-
ous space and aerospace applications have been
achieved with Keronite coated parts, including ne and
coarse sun sensors, a large trunion and a lens barrel for
an optical unit.
(b) CNES, ESA, the University of Southampton and
ONERA have participated in a cooperative effort to
229
develop a test-bed called the Material Exposure and
Degradation Experiment (MEDET) that was launched
in February 2008 and operated in-orbit for 19 months.
Keronite-coated thermal control micro-calorimeters
were mounted on the MEDET flight hardware
(Fig. 4.104) that were located on the external payload
facility of ESAs Columbus Laboratory on the ISS.
They flew in the ram-facing direction. Measurements
on return to Earth showed the Keronite samples to have
suffered no degradation and maintained a solar
absorptance, s, of 0.84.
(c) Aluminium metal matrix composites (MMC) such as
AMC640xa (6061/40SiC) offer low density, low ther-
mal expansion, high modulus and high strength char-
acteristics and it has been demonstrated to have good
resistance to stress corrosion cracking. However, fur-
ther improvement to this alloys resistance to corrosion
and wear was sought in order for this alloy to be
considered for wider space applications such as struc-
tures to be exposed to hot, saline and moisture condi-
tions as those existing at launch sites (Kourou, French
Guiana and KSC, Florida). Surface treatments by
conventional acid based anodising and Cr/heavy metal
based electrolytic processes do not offer reliable and
sufcient corrosion protection to the MMC alloys.
Keronite plasma electrolytic oxidation (PEO) has been
demonstrated in a recent project to offer improved
corrosion and wear resistance to this alloy containing
40 vol.% SiC (Shrestha et al. 2015).
(d) Keronite has also been applied to titanium alloys in
order to prevent galling, to provide for dielectric sur-
face properties and help protect titanium parts from
aggressive galvanic corrosion. Some hold-down points
on the James Webb Telescope have been so treated.
230 4 Spacecraft ManufacturingFailure Prevention

Fig. 4.103 a Reflectance from the surface of a Keronite black sample, to be made of lightweight material, able to survive extreme thermal
measured at 25 C with a xed angle of incidence of 30. b Fine sun cycling and radiation environment. Stray illumination from planetary
sensor experimental baffle, machined from aluminium matrix material, albedo and reflections from the spacecraft itself (e.g. solar panels) must
Keronite-coated. This is a basic element of the attitude and orbit be minimised by the sensors baffles which will have a low reflectance
control system of virtually all satellites. Fine sun sensors are often over the entire solar spectrum (see also Sect. 5.5.3) (Courtesy of Hill
based on CCD or CMOS activated pixel sensor technology. They need Media Ltd. and Keronite)
4.25 Joining by Friction Stir 231

Fig. 4.104 a Photograph of (a)

MEDET instruments prior to (b)
launch, b in-flight (arrow)
attached to the Columbus Module
and c a close-up image taken in
March 2008 as part of EVA. The
micro-calorimeter materials were
exposed in the RAM direction for
19 monthsaccording to
pre-flight absorptance
measurements are: SSM 0.084,
PSBN 0.122, RSF 0.115, Y100
0.150, SG121 0.189, ITO/Kapton
0.335, PCBE 0.267, RSR 0.237,
MAPATOX 0.361, Upilex 0.387,
Keronite (grey sample second
from left on top row) 0.844, black
body 1 0.991 (Tighe 2010, photo
MEDET
credits ESA and NASA)
(c)

4.25 Joining by Friction Stir
4.25.1 Friction Stir Welding
Friction Stir Welding (FSW) was invented as recently as
1991 by Wayne Thomas at The Welding Institute
(TWI) Ltd., Cambridge and this technique for joining
metals has now been incorporated into the manufacture of
aeroplane and spacecraft structures where high quality,
reproducible welds are a necessity. The FSW process is a
solid-state joining process capable of fabricating butt or lap
joints in a wide range of materials and thicknesses (Kallee
2001). In aerospace the common alloys are based on alu-
minium as these are light and, because the process does not
involve the melting and softening of the alloy, the resultant
joints have little distortion and can maintain the mechanical
properties of the parent plates. Also, due to the non-melting
of the joint, difcult-to-weld metal combinations can often
be joined without the formation of damaging voids, brittle
intermetallics or distorted interfaces. Benets also include
no arcing or generation of fumes, no spatter, good
efciency, no ller wire necessary and no gas shielding for
joining aluminium alloys. It can even be carried out under
water.
FSWs have been utilized for the welding of spacecraft
tanksthe welds being around the domes-to-cylindrical
parts. The rst space FSW was the Interstage Module of a
Delta-II launch vehicle successfully flown in 1999 (Kallee
2001). It was used for welding the Space Shuttle External
tanks that stored liquid hydrogen and liquid oxygen (alloys
being AA 2219-T87 and AA 2195-T8M4). It is used for
welding the US Ares launch vehicle tanks and curved
structural parts. In Europe, Fokker successfully demon-
strated that FSW can be used to weld Ariane-5s motor
thrust frame on the cryogenic main stage. This is an
unpressurised conical structure where the majority of parts
are made from AA 7075 and are in the form of lap joints
[Fig. 4.105 (Brooker 2001)].
Thales Alenia Space (TAS) in Turin have also designed,
manufactured and tested integrally friction stir welded
aluminium-lithium (AA2195) cryogenic tanks that are
envisaged for both the European Expendable- and the
232
Fig. 4.105 a Assembly of Ariane 5 main motor thrust frame from 12
integrally machined flat panels and b Fokkers concept for the new
cone sub-assembly jig, which is multifunctional, i.e. for both Hi-lok
TM riveting and FSW (Brooker 2001)
Fig. 4.106 a AlLi cryogenic tank made by Thales Alenia Space
Turin where the longitudinal cylinder weld and the dome-to-cylinder
welds are by FSW (courtesy ESA-TAS). b Close-up of the rotating
4 Spacecraft ManufacturingFailure Prevention
Reusable-Launch Vehicles as seen in Fig. 4.106. Large
launcher tank structures have also been made by FSW at
MT-Aerospace.
The open literature contains numerous references to
FSW. A particularly concise thesis has been presented by
Leitao (2013), at the University of Coimbra, Portugal. This
work compares the weldability in FSW of different alu-
minium alloys against welding parameters and provides data
about the joining mechanism, metallurgical and mechanical
transformations and the influence of plate thickness. It also
provides information about recent papers published for dis-
similar FSW materials (Fig. 4.107).
The main parameters affecting the FSW process are
shown schematically in Fig. 4.108. The tool dimensions
regarding pin and shoulder are noted as well as the concavity
angle and pin length. The dynamic parameters are the plunge
depth, tilt angle, rotational speed, traverse speed and
downward force. During FSW the flow of material around
the tool is not symmetrical and there are signicant differ-
ences between advancing and retreating sides, seen as AS
and RS on Fig. 4.109. The thermo-mechanically affected
zone, TMAZ, is where the tool has plastically deformed the
material and caused a fully recrystallized volume (the nug-
get) and a partially or non-recrystallised zone. The heat
affected zone, HAZ, has experienced a thermal cycle that has
modied its microstructure without any plastic deformation
occurring.
Figure 4.110 shows copper welded by FSW to aluminium
(Galvao 2011). Both sides are in plate form, having a
weld pin during FSW of the Orion crew module at Michoud Assembly
Facility in New Orleans; this FSW xture is the largest in the world
with a 6.7 m diameter turntable (courtesy NASA/LM)
4.25 Joining by Friction Stir 233

Fig. 4.107 Material pairs already assessed and presented in the literature for dissimilar FSW. Citations are given in Leitao (2013)

Fig. 4.108 Schematic of the FSW static and dynamic parameters of the rotating weld pin

weldments due to the rapid formation of the intermetallics

Fig. 4.109 Macrostructure of a polished and etched FSW (two
aluminium alloy plates), superimposed with the different microstruc-
tural zones. Retreating side, RS on right and advancing side, AS on left
(Leitao 2013)
thickness of 1 mm. Under normal fusion welding this
combination of materials would result in porous, very brittle
CuAl2 and Cu9Al4. In fact this combination of metals falls
into the category of very-difcult-to join. The
copper-aluminium intermetallics are not present in the
microstructure which shows limited thermo-mechanical
mixing and some solid state diffusion, but intermetallic
precipitation is absent under the optical microscope. This
joint is somewhat similar to a diffusion bonded interface. It
should be stated that the FSW seen in Fig. 4.110 was formed
with the lowest heat input and that other similar FSWs, with
higher heat inputs, possessed some CuAl2 and Cu9Al4
intermetallic phases that would have contributed to a lower
mechanical strength and lower ductility.
234
Fig. 4.110 Macrostructure of tranverse cross-section made on FSW
between oxygen free copper with high phosphorous content and
aluminium alloy AA 5083-H111. The harder Cu-DHP plate was placed
4.25.2 Friction Stud Welding
As with friction stir welding, friction stud welding can be
used to join difcult, and sometimes impossible, combina-
tions of metals pairs. The stud is selected as the harder of the
two interfaces. Stainless steel studs have been used to join to
aluminium alloy plates for electrical grounding applications.
Here a pin-like protrusion on the flat face of the stud is
rotated until the softer material heats up and becomes plas-
ticized. The pin penetrates into the softer plate and when the
shoulder of the stud contacts the plate a force is applied to
weld the surfaces together. There is no fusion at the inter-
faces but a true metallurgical joint is formed and a high
strength connection achieved. Many kinds of friction welds
are based on this principal. For instance, two cylindrical bars
can be welded together by rotating one bar onto the end of
the other barthe friction heating produces some material to
be expelled from the interface after which the relative motion
is stopped and a higher forging force applied for effec-
tively welding the interfaces. This joining process was dis-
cussed in Sect. 2.6.2 where it is proposed as a joining
technique for in-space assembly of hardware.
4.26 Selective Brush Electroplating
Electroplating is a well-established process employed for the
electrolytic deposition of coatings on spacecraft parts and
sub-systems. Plating will be selected for both protective
(against corrosion) and functional purposes such as for sliding
4 Spacecraft ManufacturingFailure Prevention
at the advancing side of the 14 mm diameter steel rotating tool. Here,
the weld was made with the lowest heat input (Galvao 2011)
electrical contacts, wear surfaces for mechanisms, producing
specic absorptance or reflective nishes for thermal control
systems, and for enhancing solderability or brazability prior to
assembly of parts. The main deposits are hard-chromium,
copper, gold, lead, nickel, silver and tin-lead. The quality of an
electroplated layer will be tested, usually on in-line produced
samples, or occasionally on the flight hardware.
Non-destructive tests may include adhesion (tape testing such
as to ECSS-Q-ST-70-11), surface roughness, and brightness;
destructive tests will include thickness and uniformity by
metallography, corrosion-resistance, ductility by bending,
hardness and testing for porosity. Gold plating should receive
special attention as the purer the gold, the better the electrical
and thermal conductanceboth key material properties for
spacecraft. However, pure gold has relatively poor wear
resistance, so for surfaces that will be mechanically abraded,
such as electronic pins and sockets, pcb end-connectors, and
electronic housings, it becomes essential to co-deposit very
small quantities of other elements, such as cobalt or nickel
with the gold nish, to enhance wear resistance (and to some
degree, avoid cold-welding).
The selective brush plating process can incorporate all the
principals, objectives and quality controls of the conven-
tional plating baths described above. Brush plating is used to
electroplate selected areas of a piece-part. It is performed
locally, using a D.C. current through two cables. One cable
is clamped to the work-piece (the negative, cathode con-
nection seen in Fig. 4.111) and the anodic cable is plugged
into a stylus that comprises of a carbon anode surrounded
with an absorbent cotton pad, covered with a permeable
4.26 Selective Brush Electroplating 235

uniform heat transfer between mating components. It has

Fig. 4.111 Schematic drawing to show the principals of selective
brush plating (Rubinstein 1978)
sheath acting as a brush. The plating solution is pumped,
or by immersion contained in the cotton pad, so that this
brush can be applied to the area requiring plating. The
process is known in the aircraft industries and found useful
for the repair of damaged surfaces. It is applied in the space
industry for the brush plating of high-value parts in order to
provide an adherent deposit that will give good corrosion
protection, reduce electrical contact resistance and allow
been applied by the author for the successful deposition of
copper onto electronic housings constructed from magne-
sium and aluminium alloys. The overall housings can then
be protected with non-conductive space-approved paint, but
electrical grounding of each housing is made by soldering
one end of an insulated copper wire to the copper patch, the
other end being crimped or soldered to the grounding system
of the spacecraft.
Signicant cost savings can be made by using selective
plating as the process avoids disassembly of integrated
equipment, or where the part may be too large to immerse
into a plating tank. Rubinstein (1978) reports on NASAs use
of brush plating for the repair of gold-plated edge contacts
on pcbs. It can be used for the localized plating on ultra-high
strength steel with minimum hydrogen embrittlement.
Brush plated coatings have also been applied to what are
generally considered very difcult-to-plate titanium alloys,
including the topical spacecraft material, Ti6Al4V. Titanium
is intrinsically very reactive and when the cleaned surface is

Table 4.18 Brush plating process for producing nickel coatings on Ti6Al4V
Step Operation Material Conditions
R
1 Abrade Scotch-Brite Wet with etch/activate solution
2 Etch Etch/activate 14 V anodic clean, 10 s
3 Activate Solution 48 V cathodic, 1 min
4 Strike plate Acid nickel 818 V cathodic, 0.078A * h/cm2
Courtesy of S. Clouser

Fig. 4.112 a Cross-section

photomicrograph of nickel
deposit on titanium 6Al-4V.
Titanium surface etch/activated
by procedure in Table 4.18.
b Scanning electron microscope
image of the interface between
Ti6Al4V (dark upper left) and
nickel strike plate (courtesy of S.
Clouser, SIFCO applied surface
concepts)
236
Fig. 4.113 Illustrations showing nickel being applied to components
of the ISS by brush plating at NASA MSFC (module, nodes, laboratory
and airlock). During plating of the flight articles in-line samples were
made for destructive testing. The nickel plating quality was determined
exposed to air it will immediately oxidise. Various methods
have been studied for the removal of this oxide lm prior to
selective nickel plating. Most were unsuccessful, but one
successful method involved keeping the titanium under
potential control at all times, the plated area being 100 %
covered by the wrapped brush anode and rapidly switching
from anodic to cathodic conditions as detailed in Table 4.18.
The resulting anodic etching has caused the titanium to
become micro-roughened and there appears to be no oxide
lm (see Fig. 4.112). This enables the nickel plating to form
a very strong bondit is dense and standard testing on
samples has shown the layer to be completely adherent
during bend testing under a cantilever load at 85 % of yield
for 24 h. Signicantly, additional testing proved the samples
to be negligibly affected by hydrogen embrittlement. Further
brush plating tests were made: 35 m nickel was added to
4 Spacecraft ManufacturingFailure Prevention
by adhesive tape testing, flatness by laser optical equipment, and
surface roughness down to 16 microinches Ra (0.4 m Ra). Nickel
plating thickness was measured by microsectioning and beta backscat-
ter methods (courtesy of NASA and SIFCO)
the Ti6Al4V sheet, followed by acid copper to 350 m. The
copper was then machined to reduce the copper roughness.
The machined copper surface was nally plated with a
60 m nickel cap, so demonstrating a useful surface modi-
cation schedule. This may be of use in different space
hardware applications where bimetallic couples are to be
reduced to less than 0.5 Vfor instance the nickel to alu-
minium EMF difference is 0.45, whereas titanium to alu-
minium is 0.63 (see Table 2.1).
Probably one of the most demanding brush plating
campaigns has been the electroplating of over three hundred
areas on various partially assembled components of the
International Space Station. This took place at the NASA
Marshall Spaceght Centre (MSFC) in Huntsville, Alabama
over a period of one and a half years. Nickel plating was
carried out by a team of two, an operator and a support
4.26 Selective Brush Electroplating
person, working from scaffolding and lift platforms. Great
care was needed to avoid spillage onto flight hardware. The
primary material of the ISS pressurized modules and nodes
is Al-2219, usually in the T6 or occasionally in the T851
condition (see Appendix 5). Section 4.5.2 has accounted for
the microstructure of Al-2219 which contains large inter-
metallic CuAl2 particles set in a matrix of aluminium
super-saturated with copper for the purpose of precipitation
hardening (see Figs. 4.15 and 5.8). The inhomogenous
macro-surface of machined Al-2219 contains areas of vari-
able electro-chemical potential owing to the more noble
CuAl2 particles which accelerate the dissolution or corro-
sion, in ionic solutions, of the surrounding aluminium. By
analogy, during electro-etch cleaning and electroplating, this
potential variation over the surface renders the Al-2219 as
being difcult-to-plate. The SIFCO selective plating teams
found it important to keep pre-plating etching and polishing
to a minimum during their MSFC campaign, presumably
Fig. 4.114 Denition of fracture
locations in coatings following
tape testthese depict the
weakest layer in the coating
system (from ECSS-Q-ST-70-13)
237
because of the presence of the large intermetallic particles.
Most brush nickel-plated areas were typically within hull
cut-outs that accommodate electrical connectors, fluid and
gas line feed-throughs, and ventilation ducting (Fig. 4.113).
4.27 Control of Coatings and Bonded
Items by Tape Testing
Coatings are used on a wide range of spacecraft surfaces.
These extend from corrosion protection surfaces to decora-
tive nishes (for instance, the painted flags and logos on
launch vehicles). Adhesively bonded items may be solar cell
covers and labels for part identication. These materials are
applied towards the end of system or sub-system production
by means of electrodeposition, chemical deposition, dipping,
spraying, brush painting or just pressing. Coatings are not
effective unless some bonding occurs with the substrate, by
238 4 Spacecraft ManufacturingFailure Prevention

Fig. 4.115 Examples of many

paints coated onto a variety of
substratesall will require to be
tested for peel and pull-off
strength (adhesion) by means of
pressure sensitive tapes: a White
anti-static paints are used on
multiple sub-systems of all launch
vehicles. Ariane 5 systems where
white paint is applied are shown.
The cold, white coatings have an
/ ratio <1 and, on the launch
pad help to minimize heating of
the cryogenic stage. The
anti-static coatings avoid
electrical surface discharge due to
air friction during the launch
phase. Conductive coatings have
an electrical resistance of between
105 and 109 /square. b Thermal
control coatings may have to
resist solar flux (e, p+) and rays
(UV, gamma), heat (170 to
130 C), vacuum and atomic
oxygen (LEO) for more than
15 years. The photo shows a large
optical bafflestellar diurnal
sensor of the PILOT experiment
to be flown under a stratospheric
balloon (illustrations are courtesy
of MAP)

mechanical attachment, diffusion or chemical/metallurgical
reactions. The essential considerations for cleaning, partic-
ularly the removal of silicones prior to coating, priming and
painting were extensively described in Sect. 4.21the
maxim failure to prepare is preparing to fail is certainly
applicable to the nishing of spacecraft surfaces! A variety
of recognized quality control methods can be applied to test
the suitability of coatings, both at the end of application, and
after a variety of environmental test exposures which have
been extensively presented in earlier paragraphs. Most
common in commercial industry is the knife test; here, the
knife blade makes two cuts into the coating down to the
substrate with a 3045 angle between the legs to make a
X. The knife is then used to pull up the coating at the
centre of the X. This is highly subjective and only used in
the space industry during the initial setting up of process
parameters.
More accurate, less subjective tests are selected from
ASTM D3359 or ECSS-Q-ST-70-13. These measure the peel
and pull-off strength of coatings and nishes using
pressure-sensitive tapes. May be applies to the X knife cut,
or to a cross-hatch pattern where multiple preset blades make
properly spaced and parallel incisions into the coating. Often
no cuts are made and the coating is tested as-manufactured.
The tape test used most often for European space projects
involves the application of tape (specied range of adhesion
strengths is from 220 to 670 g/cm) with a compressive load
(5 kg for 60 s) onto the coating sample xed to a top hat
support. The tape is then withdrawn and pulled apart from the
sample surface at 0.2 cm/min until separation occurs. The
sample may not have been disturbed by the test and can be
considered as passed. Samples which have been disturbed
are investigated further in order to determine the mode of
failure (see the schematics bd presented in Fig. 4.114).
4.27 Control of Coatings and Bonded Items by Tape Testing
The results of the tape test are photographically recorded
together with the pull-off load. If results are far lower than
expected, this may indicate: wrong selection of coating
material for the intended application and environment of
testing, poor workmanship (maybe de-greasing before
coating application), or wrong application procedure. QC
tests are made on all Ariane 5 components (Fig. 4.115).
4.28 The Application of EB Welding
Machine for Reflow Brazing
The LISA Pathnder mission aims to test several technolo-
gies including those designed to establish methods for
gravitational sensing. This is the rst mission to in-flight test
low frequency gravitational wave detection metrology. An
electrode housing (EH) is in the heart of the Inertial Sensor
Sub-system, as shown in Fig. 4.116. Test masses, consisting
of 1.96 kg cubes made of gold-platinum mono-phase alloy
have been chosen because of their extremely low magnetic
susceptibility and high density. The test masses are located in
the EH and their exact positions are found by high resolution
Fig. 4.116 a Schematic diagram
of LISA pathnders inertial
sensor head. The electrode
housing and test mass are
produced by Thales Alenia Space
they operate within a titanium
vacuum chamber at less than
105 Pa
239
laser interferometry, and electrostatic (capacitive) sensing.
The capacitive sensor comprises a hollow cubic molybdenum
housing with gold-coated sapphire electrodes mounted in the
faces. Differential gravitational forces are measured by
knowing the precise location of the test massesprovided by
responses from the sapphire electrodes and a complex system
of bre optics. In order to prevent damage to this delicate
set-up during launch, a locking device and release mecha-
nisms are employedall construction materials are designed
to be completely non-magnetic. The EH is made from sin-
tered pure molybdenumthis is an unusual metal for space
applications as it is very dense and has the very high melting
temperature of 2600 C. Molybdenum was selected as it
possess a similar coefcient of expansion to the large sap-
phire crystals designed to be mounted to its surfaces (both are
between 5 and 6 106 C1), together with the fact it has a
high magnetic cleanliness, a high thermal stability and can be
machined to the EH designed tolerance of a few microns.
Figure 4.117 shows an electrode housing (EH); this is the
3D capacitive sensor containing many sapphire crystals,
fully gold-plated and in a semi-nished condition. A proto-
type mechanical model of this EH was made in order to
240
Fig. 4.117 a Overall view of an
electrode housing, and b removal
of one of the molybdenum faces
reveals the gold-plated internal
surfaces and the size and location
of the many sapphire crystals
demonstrate the materials and processes used in its manu-
facture. Thermal cycling and vibration were performed on
the model, but it failed catastrophically releasing the faces
which had fractured at their interfaces. Failure analyses
indicated the poor mechanical strength was likely caused by
two inadequacies: the use of a low solder melting alloy to
bond the gold-plated surface, and the use of an inadequate
soldering oven process. Thin solder preform sheets had been
clamped between the surfaces to be joined. However, due to
the high mass of the molybdenum parts, the temperature
prole of the soldering oven had not incorporated a suf-
ciently long dwell period to enable the entire structure to
reach an equilibrium temperature. The solder llets appeared
compliant on inspection of the external surfaces, but
post-vibration, the fracture surfaces consisted of some good
metallurgical bonding but many areas of non-wetting and
some zones of AuSn4 intermetallic. A failure review board
concluded that a new manufacturing route should be inves-
tigated for assembly of the EH.
4 Spacecraft ManufacturingFailure Prevention
A multidisciplinary team considered several options for
assembling and recovering the EH. Eventually two alterna-
tive assembly processes were selected; both were based on
the use of an electron beam (EB) heating source and the
application of a braze metal (both Au80Sn20 and pure Ti
were proposed). The sapphire crystals were already attached
to the internal surface housing walls. In order to avoid
damaging the extremely expensive crystals and their
attachment bonds, the maximum temperature that could be
applied to the piece-parts was limited. Selection of EB
heating meant that there was the possibility to produce very
high spot temperatures adjacent to the braze path.
Molybdenum is a poor conductor of heat, and the brazing
operation is performed within a few seconds under a high
vacuum, so that oxidation and damage to the sapphire bonds
could be excluded. One complication was the fact that some
molybdenum parts were only gold-plated, whereas others
had either an intermediate layer of nickel between gold and
the Mo substrate or an intermediate layer of titanium.
4.28 The Application of EB Welding Machine for Reflow Brazing 241

Samples for evaluating the EB process were therefore made EB was focused onto a spot between 1 and 3 mm from the
having the following congurations: braze shims, melting of the shims being effected by ther-
mal conductivitywhen the beam was focused onto the
1. MoNiAu to Au80Sn20 ller to AuNiMo (these braze alloy itself, there was severe disruption to the molten
sample were produced in a short time frame, the incor- alloy, poor wetting and no capillary flow into the braze
poration of nickel will not be used on flight equipment as path.
it is magnetic) The full recovery plane test results have been recently
2. MoTiAu to Au80Sn20 ller to AuTiMo (samples published (Soli and Dunn 2015). Processing parameters have
with titanium represent actual flight materials for the EH) been optimized for each of the braze alloy and molybdenum
3. MoTiAu to Titanium ller to AuTiMo nishes listed as 13. The appearance of these samples and
their microstructures are presented in Figs. 4.119, 4.120 and
The recovery plan for the EH was divided into three 4.121 and the main comments and tensile strength results are
activities: Feasibility, Qualication and Flight model compiled into the captions.
repair. Sample type 2 was considered to be the main contender
Feasibility included a wide range of activities such as an for progress into the Qualication phase based on sample
attempt to electron beam weld Mo to Mo using a Mo ller results and the metallurgical attributes of this system. Tita-
sheet. This was unsuccessful, despite varying parameters nium has a complete solid solubility in Mo, there are no
that included: EB current, voltage, weld velocity, number intermetallic reactions during brazing and hence the
and thickness of shims to be melted, and changes to cooling Ti-to-Mo interface will be very strong. Considering the
rates. One sample has been described in the caption to Au80Sn20 ller, with optimum heat input, the shims will all
Fig. 4.118. fuse together, the melted braze alloy then dissolves the gold
These results suggested that progress should be directed plating and then wets to what remains of the Ti. This pro-
to the production of EB brazed samples using the sample duced a brazed joint that, when mechanically tested, had the
types 13 listed above. EB brazed joints were produced most fracture surface in the braze metal (centre). Fracture is
using a ne focused electron beam directed at 45 into rough with peaks of braze alloy. Poor brazing, with insuf-
corner of the samples; a range of parameter variants, each cient heat input provides no bond between Ti and Mo. This
test attempting to ne-tune the process to achieve accept- is the weakest interface and this is where the fracture can
able joints from a perspective of visual inspection, tensile occurthe fracture surface is flat and grey in colour, repli-
strength and microstructural soundness. Importantly, the cating the detached Ti to Mo interface. Some flat areas were

Fig. 4.118 Section of a visu-

ally acceptable EB
weld/braze joint using a pure
Mo foil clamped between two
Mo plates. The electron beam
was perpendicularly ne
focused onto the centre-line
braze path. The resulting
inter-granular cracking, seen
on this polished and etched
microsection, conrmed litera-
ture data that sintered molyb-
denum is an extremely difcult
metal to weld. Liquid weld pool
metal releases gas bubbles and
on solidifying at 2600 C and
on cooling to ambient temper-
ature there is extensive shrink-
age cracking
242
Fig. 4.119 Type 1 samples. a Visual appearance and microsections to
illustrate satisfactory EB brazed joints with Au80Sn20 alloy between
Au on Ni on molybdenum. b Gold-tin binary phase diagram detailing
the Au-rich side. The solidication range is from about 520 to 280 C.
c Macro-section through the lap-shear joint showing good wetting and
seen on the (fractured) shear test sample. The Au rich liquid
will partially alloy with the titanium interface as shown by
the phase diagram shown in Fig. 4.120b, but the resulting
microstructure can be complex and will contain several
phases that will be dependent on the time and temperature of
the braze cycle. Xu et al. (2013) in their studies found that
temperatures of brazing can exceed 350 C, but cooling
should be rapid, and particularly there should be no dwells at
190 C in order to avoid the formation of Au5Sn as this
phase was weaker than the ner AuSn/Au5Sn eutectic
microstructure when bonded to a titanium coated substrate.
Type 3 samples, consisting of molybdenum EB brazing
with pure titanium foils, have been attempted on a limited
number of samples. These were made as a back-up to sample
type 2 (with Au80Sn20 ller). The temperature needed to
obtain this Ti brazing on Mo was far higher than that needed
for Au80Sn20probably in the region of 1800 C against
4 Spacecraft ManufacturingFailure Prevention
minimal porosity. Needles of the peritectic phase, , are seen
throughout this braze. d Nickel is only discernable at higher
magnications, less than 1 m remains on the Mo surface due to
dissolution into the braze alloy. The lap-shear test joints produced
strengths in the range 2540 MPa
550 C. However, the obtained lap-shear strengths were
30 MPa, fracture being in the molybdenum and not in the
titanium braze. This metallurgical joining method has been
successfully used with both palladium and titanium pure
metal braze ller metals by Chinese researchers (Lin 2013)
but using oven instead EB. Figure 4.121 show the appear-
ance of the titanium braze and its macro-section.
Qualication of the Electrode housing materials and
processes were made using the Type 2 braze alloy together
with a well-understood EB brazing process thermal cycle.
Operators, trained and certied by the company HTC (High
Technology Company, Foligno, Italy) performed EB
brazing against a completed process identication speci-
cation (PID) where all the variable parameters were xed
for these qualication sample production runs (see
Fig. 4.122). The braze cycle was made rapidly, with a high
traverse speed of the electron beam (focused on the Mo top
4.28 The Application of EB Welding Machine for Reflow Brazing 243

Fig. 4.120 Type 2 samples:

Au80Sn20 alloy between Au on
Ti on molybdenum. a Four shims
of Au80Sn20 foil were used to
produce the excellent llet
appearance. b the AuTi phase
diagram indicates that, at the
brazing temperature, little Ti will
dissolve into the alloy. Although
the llet shows excellent melting
and wetting of the braze metal,
there is some indication that the
temperature was not sufcient to
fully melt the entire length of
these shims. It is probable that the
maximum temperature of brazing
was well under 350 C (and the
cooling cycle was rapid to
ambient temperature, thus
avoiding long dwells at 190 C
when the AuSn phase becomes
unstable and brittle Au5Sn might
form). The lap-shear samples all
produced values above 30 MPa,
being on average stronger than
the Type 1 samples
244 4 Spacecraft ManufacturingFailure Prevention

Fig. 4.121 Type 3 samples.

a EB brazed joint, Mo to titanium
ller to Mo. The molybdenum
titanium phase diagram seen in
(b) is simple, showing complete
mutual solid solubility of both
elements. Inter-diffusion between
the Ti braze and the Mo substrate
will be rapid and will have a high
strength. The intermediate very
thin layers of Au will be rapidly
dissolved in the TiMo alloy. No
intermetallic compounds will be
present in the main braze path.
The high strength has been
conrmed by lap-shear testing to
be always far greater than
30 MPa, fracture being in the
sintered molybdenum
4.28 The Application of EB Welding Machine for Reflow Brazing
Fig. 4.122 Operators producing the EH qualication samples, parts
are aligned with the beam and monitored by closed loop TV and
microprocessor. Vacuum chamber runs at 106 bar. A tungsten mask
Fig. 4.123 Environmental testing of EH faces was totally successful and as such, the EB process is considered qualied
surface) at a given distance from the interface containing
braze shims. Cooling by conduction would be rapid, thus
avoiding the solid state precipitation of Au5Sn at 190 C,
and the EB welding vacuum chamber was only opened to
ambient pressure when the items had reached a temperature
when they could be comfortably handled. The qualication
plan specied that the testing would be performed on EH
faces BB. All vibration test campaign sessions were suc-
cessful. No failures or surface visual degradations were
seen on any of the EB brazed parts. All frequency reso-
nance test proles were identical, before and after the
vibration runs. The test frequency range was between 20
and 2000 Hz, the level up to 36 g r.m.s. Tests were
245
was placed over the samples to protect the sapphire crystals from the
electron beam and from any condensed material that could have
sublimated during this high temperature/vacuum process
repeated along the 3 axes with a minimum duration of
120 s. Further tests were made on all axes for 600 s.
Clearly the EH faces were over-tested with respect to the
LISA pathnder mission equipment vibration requirements
of: 120 s and 25 g r.m.s. The robustness of the qualica-
tion parts were such that it was not possible to establish a
rst mode of failure. The success of this process develop-
ment is important as it involved an innovative joining
techniqueonly one other reference to a similar process
was established (Lin 2013). With this added condence in
electron beam reflow-brazing of Au80Sn20 between tita-
nium plated molybdenum faces, the conditions for Flight
Model Repair were thus in-hand (Fig. 4.123).
 Metallography Applied to Spacecraft Test
Failures
5.1 Application of Electron Microscope
5.1.1 SEM Examination of Fracture Surfaces
The failure mode of spacecraft materials can often be
assessed by detailed examination of the fracture surfaces.
The scanning electron microscope (SEM) is an invaluable
tool in such failure investigations because it extends both the
depth of eld and the magnications obtainable by con-
ventional light microscopes. Fracture surfaces rarely require
special preparation for SEM inspection. Only when thick
corrosion products or other nonconductive substances are
present in the area of examination will a thin layer of gold be
applied to prevent electrical charging effects. Laboratory
personnel should be encouraged to examine as many spec-
imens as possible in an uncoated form, as only the uncoated
fracture surfaces can display the true surface structure. If a
coating is required in order to avoid the build-up of surface
charge, or to provide a surface layer that produces a higher
secondary electron yield than the specimen material, two
options are recommended:
Sputter coating. Here, low-vacuum equipment is used to
deposit gold or, platinum at a rate of up to 1 nm/s onto
the specimen. It must be remembered that sputtered
metals are deposited in the form of islands, not as a
continuous coating. To avoid the shadow created in the
coating by a rough fracture surface it is preferable to coat
the specimen tilted at 45 to the target, and then repeat
with the specimen tilted in the opposite direction.
Carbon deposition. This method involves heating carbon
bres to approximately 2000 C under vacuum. Carbon
will then evaporate and form thin conducting lms on
adjacent areas. Care must be taken to avoid overheating
the specimen.
The classication of material fractures is usually made
with reference to their structure, loading mode, and crack
propagation mechanism. Materials that are said to be plastic
Springer International Publishing Switzerland 2016
B.D. Dunn, Materials and Processes, Springer Praxis Books,
DOI 10.1007/978-3-319-23362-8_5
5
do not always crack plastically. Likewise, brittle materials
need not display brittle fractures. It is the stress level and
other loading conditions such as loading rate, duration, and
temperature which determine the appearance of a fracture
surface.
Fractures often leave characteristic markings from which
the cause of a component failure may be quickly deduced.
However, for accurate identication of failure modes, frac-
tography should always be followed by metallographic
examination.
The appearance of a fracture surface has always served as
a criterion of the quality or suitability of a material. Brittle
fractures are taken to mean fracture without macroscopic
plastic strain. However, there is invariably local microscopic
plastic strain ahead of the tip of a brittle crack propagating in
metals. An example of a brittle fracture in a polycrystalline
metal is shown in Fig. 5.1. The top view of this sample
shows brittle cracks propagating along particular crystallo-
graphic planes of the grain, termed cleavage planes, to
form a transcrystalline fracture. Brittle cracks can also
follow along a metals grain boundaries, forming an inter-
granular fracture. Plastic or ductile fracture is preceded by
macroscopic plastic strain due to deformation in the slip
planes without crack propagation at either cleavage planes or
grain boundaries.
An excellent treatise (Engel and Klingele 1981) fully
describes SEM observations of a wide selection of metal
fractures and damaged surfaces. The book is intended for
engineers who wish to locate and identify their own frac-
tographs with those presented. More recently, fracture surfaces
resulting from the premature failure of industrial components
and structures has been compiled by the Italian Welding
Institute, Genoa, whos metallurgical laboratory is recognized
for performing failure analyses (De Marco et al. 2015).
As an insight to the usefulness of the SEM in performing
spacecraft materials failure investigations the appearances of
some damaged metals are briefly illustrated in Figs. 5.2, 5.3
and 5.4. However, no attempt is made to delve deeply into
detailed metallurgical explanations concerning these modes
247
248
Fig. 5.1 Brittle fracture of polycrystalline zinc at 77 K after uniaxial
tensile loading. a Top view of flat specimen after tensile failure.
Cleavage cracks follow slip lines yielding a cleavage transcrystalline
fracture. There is evidence of microscopic twinning and local plastic
deformation (optical microscope). b Fracture surface of grain high-
lighted in (a). A characteristic property of brittle fracture is that these
features are reminiscent of river patterns (optical microscope). The
fracture has started in a cleavage plane (center-left containing parallel
striations) and progressed across the wider river markings towards the
right-hand side, in the direction of the ner river patterns (Dunn,
unpublished report)
5 Metallography Applied to Spacecraft Test Failures
Fig. 5.2 SEM photograph of fatigue striations on the fracture surface
of a failed satellite strut (Al 6061-T651) after vibration testing. Crack
propagation is in direction of arrows
of failurethat can be made by consulting either the Engel &
Klingele publication, recent editions of the ASM Handbooks
(volume 11, failure analysis and prevention; and volume 19,
fatigue and fracture) as well as the Italian language failure
analyses compilations by De Marco et al. (2015).
Vibrational fatigue failure is represented in Fig. 5.2. This
fractograph details the fatigue striations as seen in the SEM
on the failed surface of a strut, part of a satellites structure,
after vibration testing. The fracture path is perpendicular to
the stress axis, and each of the parallel striations has resulted
from a single stress cycle. The density of the striation
spacings and their denition also provide information about
the loading conditions.
Figure 5.3a highlights the appearance of typical stress-
corrosion cracks found in a high-strength aluminium alloy
test specimen which was stressed to 75 % of the materials
yield point and then exposed to a saline environment for
80 days. Some corrosion pitting was observed adjacent to
this specimens surface. A stress concentration had devel-
oped at the root of the deepest pit and promoted
stress-corrosion cracking (SCC) along this alloys rather
large grain boundaries. Figure 5.3b shows the characteristic,
very small dimple shapes at the centre of the specimen. They
result from tensile overloading and the ductile failure of the
alloy after 80 days of environmental testing.
The aluminiumlithium alloy AA 8090 shown in Fig. 5.4
possesses a microstructure consisting of large, elongated
original grains that have been pinned by a network of ne
5.1 Application of Electron Microscope
Fig. 5.3 SEM views of fractured AA 2219-T851 stress-corrosion
specimen after 80 days exposure to a saline environment. a Intergran-
ular cracking attributed to stress corrosion. b Ductile dimple fracture
following tensile overloading of aluminium plate
intermetallic particles. A subsequent heat treatment has
produced a substructure of subgrains. Stress-corrosion
cracking and tensile fracture had occurred within 9 days of a
standard 30-day alternate immersion test. Unlike the AA
2219 alloy shown in Fig. 5.3, which is covered in corrosion
products and pitting, the AA 8090 alloy supports very little
indication of surface corrosion and no pitting; both
stress-corrosion cracking and tensile fracture occurred
mainly along the original grain boundaries with secondary
cracks at sub-grain boundaries. Microsections made in the
vicinity of the main stress corrosion crack show very little
pitting or corrosion products (Fig. 5.4c illustrates these
249
(a)
(b)
(c)
0.4mm
Fig. 5.4 SEM views of AA 8090-T8771 plate following the
stress-corrosion failure of tensile specimen after 9 days exposure to
a saline environment (see also Table 4.4). a Bright areas are the steep
sides on the fracture surface caused by SCC. The propagating cracks
have generally followed the large, original grain boundaries, some
follow the subgrain boundaries of this alloy to produce intergranular
cracks supporting small facets. b Shows examination in the same area
as a, but by backscattered electron imaging which highlights a
network of comminuted acicular intermetallics along the original grain
boundaries. The fracture surfaces a and b support no corrosion
products. Similarly, pits and corrosion are not visible on the
specimens surfaces (c)
250 5 Metallography Applied to Spacecraft Test Failures

insidious hair-line defectsthey are very difcult to nd by

standard non-destructive tests when inspecting flight
hardware).
A Belleville spring is a convexconcave washer that
stores energy when flattened. The stacking arrangement of
mated pairs of these dish-shaped high-strength steel washers
is shown in Fig. 5.5. The illustrated stacks of springs were to
have flown as a vibration-absorbing system but, after an
actual flight hardware failure in the USA, it was deemed
necessary to assess whether these similar steel parts would
be susceptible to the same stress-corrosion mode of failure.
Sets of Belleville springs were loaded with
flight-representative preload and installation stresses. Two
groups of springs were exposed to what should have been a
30-day environmental test. The rst group of samples was
submitted to a standard alternate immersion test using a
one-hour cycle that includes a 10-min immersion period in
an aqueous solution of 3.5 % sodium chloride, followed by a
50-min period out of the solution during which the samples
are allowed to dry. Rather than surviving the 30-day
acceptance test, several of the dished washers failed by
cracking after only 6 h. The second group of samples was
exposed to a warm, humid environment (60 C and 95 %
RH). These samples failed after only 72 h. The fracture
surfaces are shown in Fig. 5.6. They illustrate classical
intercrystalline cracking resulting from stress corrosion that
has propagated rapidly along prior-austenite grain bound-
aries. Some evidence of micropores and other fracture sur-
face facets is also detailed. A review of the high-strength
spring steel washers found them to have been fabricated
from an alloy having only a moderate resistance to
Fig. 5.6 SEM views of Belleview spring fracture surface. a Arrows
stress-corrosion cracking. Replacement washers were indicate grain boundaries and micropores. b Showing the ductile
hairlines and crows-feet markings (courtesy of A. de Rooij)

fabricated from an alloy represented in Table 4.6; later, they

were tested and found suitable for this application.

5.1.2 TEM Examination of Metallic Failures

The SEM is, primarily, a tool for studying surface topog-

Fig. 5.5 One Belleville spring assembly during SCC testing. The
second photo is 90 turned relative to the rst. The number of cracks is
17. One dished washer lost a part (arrowed)
raphy with the advantage of a depth of focus which is about
ten times that of the optical microscope for a given magni-
cation. The transmission electron microscope (TEM) has
about ten times more resolving power, and it can provide
even greater magnication than the SEM, as was illustrated
in Fig. 3.9. The metal sample must be either subjected to a
thinning process to ensure electron transmission for the
direct observation of the interior, or, when a surface requires
laboratory examination, replica techniques must be
employed (see also Sect. 3.2.2.2).
By far the most difcult part of TEM operation is that of
producing and handling thin enough specimens for
5.1 Application of Electron Microscope 251

transmission work, and this has tended to limit its use in
spacecraft failure analysis. The utilization of the TEM is
unjustiably considered to be solely a university materials
science tool or a company R&D method for studying the size
and distribution of microconstituents in new alloys. Whereas
the instrument has uniquely assisted in the understanding of
physical metallurgy and, in particular, dislocation move-
ment, it can also be applied during the assessment of met-
allurgical failure mechanisms. Figures 5.7 and 5.8 clearly
demonstrate the elegance of the method and highlight
microstructural deciencies caused by the incorrect pro-
cessing of high-strength structural alloys having vastly dif-
ferent compositions. Both case histories refer to alloys
which, when freshly quenched, form a supersaturated solid
solution that is soft. Heating to a predetermined temperature
for a specied time causes second phase precipitates to form
preferentially on certain habit planes. As precipitation takes
place, the alloys become harder, but, if the precipitates are
permitted to grow owing to holding the material for too long
at high temperature, the alloys slowly begin to soften. This
phenomenon is known is over-ageing. As the initial pre-
cipitates grow in size and change morphology they become
unable to pin, or anchor, dislocations that are moving under
applied stress and this results in an overall reduction of the
alloys mechanical properties.
Other illustrations of the TEMs usefulness as a failure
investigation tool are shown in Figs. 4.50b and 5.43, which
show, respectively, precipitation hardening in an electroless
nickel plating, and an evaluation of dislocations and grain
growth with their effect on the embrittlement of beryllium
sheet.
5.2 Fasteners
5.2.1 Spacecraft Fasteners
Fastener systems are the essential elements which may hold
together both small parts of a device and large subsystems to
a spacecraft structure. Care must be taken to address
potential problems associated with nuts, bolts, washers, and
rivets, as these items are considered by engineers to be
simple and are often overlooked. Fasteners are normally
installed by torquing the bolt with a wrench so that it

Fig. 5.7 Carbon replicas of a

12 % chromium steel alloyed with
Mo and V for high strength and
good creep and oxidation
properties. Precipitation
hardening produces an optimized
microstructure of rod-shaped
precipitates (see a and b transition
carbides, mainly Cr7C3) giving
enhanced physical properties.
Sections of the same piece-part
were exposed to excessively high
temperature 750 C/2 h. The
transition carbides have converted
to massive spherically shaped
particles (c and d show
equilibrium carbides of M23C6),
accompanied by an increase in
ductility and corresponding
decrease in hardness, proof stress,
and UTS. Note, these carbon
extraction replicas were prepared
from lightly etched polished
specimens (Dunn unpublished
work)
252
Fig. 5.8 a Transmission electron micrographs of Al-6 wt%Cu plate,
solution treated, stretched, and aged. Precipitation hardening during
ageing gives particles on [100] planes. This is consistent with the Al
2219 plate treatment according to the T851 condition (175 C/18 h).
TIG-welding of the plate causes a reduction in hardness from 140 DPN
in the parent plate, to only 90 DPN in the over-aged
weld-to-heat-affected zone (HAZ). Examination of the foils from the
HAZ shows b less precipitates and more evidence of CuA12
precipitation adjacent to the parent plate and c total over-ageing
adjacent to the weld pool. The strengthening effect of the phase is lost
as it is converted to produce coarse CuA12 particles (Dunn 1984)
5 Metallography Applied to Spacecraft Test Failures
becomes seated and this is followed by the locking
torque which will prevent the joint becoming loose under
conditions of vibration and thermal fatigue. Torque wren-
ches are tools which need regular calibration since their
usage causes them to loose their accuracy and setting points;
calibration is often made by certied contractors at periodic
intervals dened by the Quality Department of the assem-
bling company. Fastener failures have occurred because of
mechanical overloading, corrosion, and a variety of manu-
facturing defects, so that full quality controls are required
both during fastener fabrication and at the time of spacecraft
assembly and integration. The selection of materials and
their nishes must be suited to the expected service envi-
ronment. This may vary from cryogenic temperatures to
temperatures in excess of the melting point of standard
alloys. Spacecraft fasteners will be required to reasonably
match the thermal expansion coefcient of the structure, to
possess a resistance to stress corrosion, and to be resistant to
hydrogen embrittlement. In most applications there must be
galvanic compatibility with the parts being joined, as
detailed in Table 2.1, and electrical bonding requirements or
grounding of surfaces will eliminate the use of organic paints
or sealants for corrosion protection. Electrical bonding may
require a maximum d.c. resistance of 2.5 m across joint
surfaces, so that, in the area of the fastener, surfaces must be
made clean (paint and anodization lms being removed) and
subsequently coated with either a metal layer or a chemical
conversion lm.
The majority of threaded fasteners selected for spacecraft
application possess a high strength coupled with an inherent
resistance to both surface and stress corrosion. The most
widely used high-strength materials are listed in Table 5.1.
All have an appreciable resistance to SCC, and, as such,
were listed among the alloys in Table 2.3.
The most usual sequence of processing used in the
manufacture of fasteners is to initially solution heat treat the
hot rolled bar or wire stock, then cold reduce to size, warm
forge, precipitation heat treat, roll thread, and nish surfaces.
The thread-rolling work-hardening process after heat treating
is used to build high residual compressive stresses at the
notch roots and thus overcome any fracture toughness
problems that might otherwise exist. It should be noted that
it is also benecial to work-harden the nonthreaded zone of
the bolt in a separate operation, otherwise the fatigue
strength of the nonthreaded zone is less than the threaded
part (Turlach 1985). For high-strength Inconel fasteners
exposed to very severe SCC and fatigue environments, it
may be better to roll the threads at room temperature rather
than warm roll at 400 C. However, in all cases, rolled
threads have improved SCC and fatigue lives over those of
5.2 Fasteners 253

Table 5.1 Mechanical properties of some corrosion-resistant fastener materials

Material Typical chemical Room temp. tensile properties Remarks
composition (see also UTS 0.2 % Elongation Reduction
Appendix 6) Mpa PS UTS (%) of area %
(ratio)
15-5PH 15Cr 5.5Ni 4.5Cu bal.Fe 1250 0.9 6 15
Ti6A14V 6A14V balance Ti 1100 0.95 15 45
a
286 15Cr 1.2Mo 25Ni 960 0.70 14 20 (Cold work and ageing produce UTS of
1380 MPa). Max. service temp. 670 C
Waspaloy 21Cr 5Mo 15Co 3.3Ti 1250 0.70 12 18 Max. service temp. 800 C
2Fe bal.Ni
a
Inconel 718 21Cr 3.3Mo 1Co 55Ni 1250 0.80 13 20 (Cold work and ageing produce UTS of
bal.Fe 1517 MPa). Max. service temp. 650 C
Multiphase 20Cr 10Mo 35Co 35Ni 1800 0.90 10 50 (Cold work and ageing produce UTS of
MP35 N 1790 MPa)
a
May be lubricated with 510 m silveralways install washer to avoid silver contact with aluminium or titanium structure

cut threads in both air and water (Kephart and Hayden
1993). The improved properties of rolled threads (for most
metal alloy fasteners) are due to interactive factors, including
the presence of benecial residual compressive microstresses
in the critically stressed thread root region, reduction of
plastic strains during loading, and the formation of a
favourable microstructure.
A wide variety of rivet designs are employed in the
aerospace industry to hold together metal plates. The
headless end of the rivet is passed through two holes in the
plates and then beaten out or pressed down to mechanically
hold them in contact. Wrought aluminiumcopper alloys
are most often selected for spacecraft applications for the
fastening of brackets and panels. There is usually no gal-
vanic corrosion problem, as all parts are made of a similar
aluminium alloy composition with a primary surface nish
of chromate conversion coating. Rivet compositions are
usually AA 2017 (4 %Cu 0.6 %Mg 2 %Ni Al alloy) and
less frequently AA 2024 (4 %Cu 1.6 %Mg Al alloy). Both
alloys are selected because they can be solution treated and
then age hardened (precipitation treatment) to enhance their
mechanical properties. The 2 % nickel addition to AA 2017
acts as a grain-rening element which stabilizes the rivet so
delaying room temperature age hardening. The AA 2024
rivets are slightly more difcult to process, as after solution
treatment (holding at 500 C then water quenching) they
should be stored below 0 C in order to avoid natural
ageing. Both alloys can be readily cold worked (e.g. riv-
eted) in the solution-treated condition. After the rivet has
been deformed it can be allowed to age harden, so inducing
high strength. Neither AA 2017 nor AA 2024 has a high
resistance to stress-corrosion cracking, but by special dis-
position, they are approved for given spacecraft applications
as the rivet shanks are not loaded in the short transverse
direction (i.e. the grain direction most prone to stress
corrosion is not exposed to tensile stresses). Experience has
shown that the fastening of structural parts by rivets is a
reliable process; however, in the light of two recent rivet
failures the following recommendations have been made
(Jha et al. 1996):
AA 2024 rivets must be driven within 30 min following
solution treatment followed by quenching, or must be
kept refrigerated until needed to prevent age hardening
prior to driving
material chemical cleanliness is important, the presence
of second-phase particles and undissolved CuMgAl2
distributed throughout the matrix causes shearing along
the plane of maximum stress (resulting in the popping
off of rivet heads)
a smooth llet radius should be provided at the head to
shank junction of the rivet, to prevent stress concentra-
tions should the rivet be subjected to bending moments.
Industry is presently at risk due to the rise in counterfeit
fasteners appearing on the market. These are copies or
substitutes whose material and performance (mechanical
properties) are misrepresented by the vendor or manufac-
turer. It is essential that bolts, nuts and washers are pur-
chased from authorized dealers and where possible, directly
from the manufacturer. European spacecraft manufacturers
generally specify the ECSS-Q-ST-46 when procuring fas-
teners. This document requires fastener heads to be hot or
cold forged before heat treating; the head-to-shank llets and
the threads will be formed by rolling; identication marking
will be by depressed characters that univocally identify the
fastener batch and supplier; and non-destructive inspection
requirements are well dened by reference to ASTM and
ISO standards. Major aerospace fastener manufacturers
include: Blanc-Aero Industries (France), Linread (UK) and
254 5 Metallography Applied to Spacecraft Test Failures

Fig. 5.9 As-polished

cross-section of forged 4520 steel
fastener after fracture during
spacecraft vibration testing. The
general view a highlights location
of fracture, and the detail b shows
the presence of a thin oxide layer
along crack path. The steel is
noted to contain very slight
internal oxidation in this location.
Excessive deformation during
hot-heading of this batch of
fasteners produced incipient
seams which became oxidized.
Crack propagation from this
stress-raising defect caused shear
fracture of head during test

Fairchild Fasteners (Germany), SPS Technologies and Huck
Fasteners Int. (USA). Further quality controls are described
in Sect. 5.2.9.
5.2.2 Fastener Failure Due to Forging Defect
The hexagonal socket head from a steel bolt manufactured to
an aircraft standard, as shown in Fig. 5.9, failed during
satellite testing. The steel was analysed and found to belong
to the medium carbon martensitic high chromium family.
The failed bolt, together with intact bolts originating from
what was understood to be the same batch, was subjected to
visual inspection and metallography. All had been
manufactured by forging with hot heading of the sockets
a process commonly performed at approximately 1000 C
when the material is in the austenitic phase. Figure 5.9a
shows the failed bolt to have undergone a large volume
displacement during forging; intact fasteners had a far lower
insert-to-head ratio. The specication called for these bolts
to be quenched and tempered to a hardness of 315 Brinell
minimum (equivalent to a UTS of 1110 MPa). All bolts had
a hardness of between 363 HB in the head volume and 380
HB within their rolled threads. Only by reinspecting the
as-polished section under high magnication could the crack
initiation site of the broken fastener be established. This is
shown in Fig. 5.9b to exist as a very ne seam of oxide
present along 80 % of the crack path. Several other seams
5.2 Fasteners
were observed adjacent to the fracture surface. Full material
traceability had not been established for this type of fastener,
and it was considered that the failed item had originated
from defective stock. The seams were believed to have
formed during the socket head forging operation because of
an inefcient forging-die design, and possibly the forging
operation was unduly fast for this design. Retroactive action
was to randomly sample additional bolts from all batches
that had been utilized on the spacecraft. These were longi-
tudinally cross-sectioned and metallographically inspected.
No seams or further defects were observed on those samples.
Finally, as a last precaution, magnetic particle inspection
was applied to all fasteners in an attempt to screen-out any
other rogue items from the flight batch.
5.2.3 Laps and Surface Irregularities in Threads
Fastener specications will permit the existence of minute
laps, seams, nicks, or gouges on non-load-bearing surfaces
of a bolt head. No discontinuity or surface contamination is
permitted at the head-to-shank llet or at the root of bolt
threads or other bearing locations. The presence of such
defects, and particularly seams in the threaded portion of
bolts, is rather common, and strict quality controls are
required by both manufacturer and user. Seams are dened
as crevices in the surface of metal that have been closed, but
not welded, by the working of the metal. Typical examples
of rejected spacecraft bolts can be seen in Fig. 5.10. These
seams are difcult to detect by nondestructive means, and
they cause a stress concentration to intensify in an even
smaller, more localized, region of the thread. Dye penetrant
and magnetic particle inspection will detect defects having a
0.02 mm wide seam, but smaller faults may pass unnoticed
to cause a reduced fatigue life and premature failures.
5.2.4 Hydrogen Embrittlement of Steel
Fasteners
A common source of failure of bolted assemblies is the
hydrogen embrittlement of high-strength steels (UTS in
excess of 750 MPa). Failures have occurred in a short time
under applied stresses as low as 40 % of the steels yield
strength. This form of material degradation has been dis-
cussed in Sect. 4.4. The low ductility of such steel fasteners
resulted from the absorption of hydrogen during strong-acid
pickling. It is now common practice to bake high- and
medium-strength bolts after pickling at 180 C for 3 h. Some
bolts may support an electroplated layer of gold or silver as a
solid lubricant. The process controls for successful applica-
tion of these metals is very important, and, after plating, an
255
additional 24 h bake is recommended for the removal of
hydrogen. Although cadmium plating is a restricted nish for
use in high-vacuum environments, it is commonly employed
as a general corrosion protection nish to fasteners for
ground equipment. The deleterious effect of thick cadmium
electroplating and resulting hydrogen embrittlement of
high-strength steel bolts is well documented, and where this
is of major concern the cadmium should be deposited by a
vacuum deposition process. Aluminium coatings are also
gaining popularity as the environmental problems associated
with cadmium are becoming recognized.
Aluminium can be ion vapour deposited onto fasteners
made of steel, titanium alloys, and aluminium alloys. This is
a particularly useful nish as it can completely cover the
roots of threads, recesses, and sharp corners on the fasteners.
There is no need for a preliminary coating or plating of
another metal, and no problem of hydrogen embrittlement.
The aluminium coating is adherent and can be chromate.
This nish of high-purity aluminium can be used for appli-
cations which require electrically conductive surfaces, such
as electrical bonding and grounding for fasteners requiring a
low EMF potential when coupled to an aluminium alloy
spacecraft structure (see also Table 2.1).
5.2.5 Embrittlement of Titanium Alloys
Titanium alloys, including Ti6A14V, are known to become
embrittled when held in contact with certain solid metals.
Cadmium can embrittle such alloys at temperatures as low as
150 C, whereas a threshold temperature of 204 C exists for
silver and gold (Stoltz and Stulen 1979). The actual mech-
anism of solid metal embrittlement is not understood, but it
is known that cracking in cadmiumtitanium couples occurs
more rapidly in vacuum than in air. Solid metal embrittle-
ment, like liquid metal embrittlement, results in intergranular
fractures, indicating a reduction in the cohesive strength of
the grain boundaries.
Other recognized titanium failure modes include those
associated with the stress-corrosion fracture of structural
parts under service loads caused by their exposure to cleaning
liquids including methanol and certain of the halogenated
hydrocarbons. Free chloride ions present in methanol have
promoted the SCC failures of titanium alloys. The most
dramatic were the failures of Apollo tanks containing
methanol (Johnston et al. 1967); fortunately, these failures
occurred during proof testing, and regulations were intro-
duced to avoid similar damage to flight tanks. The following
halogenated hydrocarbons are known to promote cracking in
titanium alloys (Brown 1972): carbon tetrachloride, methy-
lene chloride, methylene iodide, trichloroethylene, tri-
chlorofluoromethane, and octafluorocyclobutane. Certain
256 5 Metallography Applied to Spacecraft Test Failures

Fig. 5.10 Seams introduced

during rolling were extensive
enough to cause bolt
deterioration, but they passed the
initial nondestructive tests and
were revealed only during the
metallurgical examination of
samples. a General views of two
Ti6A14V bolts. b Detail of thread
form. Note defects extend inside
the thread centre-line diameter
(cause for rejection). c Detail of
defects around thread peaks

fluorinated hydrocarbons, registered under the trade names
Freon TF (C2Cl3F3) and Freon MF (CCl3F), can also lower
the fracture toughness KISCC behaviour of Ti6A14V,
although the TF grade appears to cause less susceptibility to
stress corrosion than the MF grade. These solvents may no
longer be purchased, but some companies maintain stock-
piles which may be recycled until alternative liquids have
been qualied as replacements (see also Sect. 4.21).
Each time titanium alloys are selected for a space project
they will be reviewed by means of the project declared
materials list. By controlling the cleaning liquids, excluding
certain environments (e.g. N2O4 and cadmium plating) and
selecting fully tested alloys, it can be stated that
high-strength titanium alloys have been remarkably free
from embrittlement problems. Such controls are rigorously
maintained for all spacecraft fastener applications, and the
5.2 Fasteners
author is not aware of any titanium alloy fastener failure
resulting from either solid metal embrittlement or SCC.
5.2.6 Galvanic Corrosion of Fasteners
When dissimilar metals are fastened together, accelerated
corrosion can occur at their interfaces owing to the previ-
ously discussed mechanism of galvanic corrosion. Table 2.1
listed metals in the order of their relative nobility in an
aqueous environment. The list proceeds down to the more
active (anodic) metals, and it is these that will corrode as a
result of current flow from the anode to the less reactive
(more noble) cathodic metal. The rate of galvanic corrosion
will depend on several factors, including:
The conductivity of the electrolyte (e.g. condensed water
on an electronic box after its removal from a cold envi-
ronmental test chamber). High conductivity is often
caused by the presence of chloride ions from nger
marks, or by residual acids or alkalis from chemical
processing operations (fasteners, washers, and their sur-
rounding areas must be maintained in a clean condition).
The area ratio will determine the rate of galvanic cor-
rosion; as the area of the more noble metal (cathode)
increases, so the rate of corrosion of the anodic metal will
be more raid (care must be taken to ensure that the fas-
tener metal is either the same polarity or cathodic to the
surfaces being joined).
Great care must be taken when nonmetallic materials must
be joined to metals with metallic fasteners. Table 2.1 shows
carbon (graphite) to be more noble than gold. Extensive
corrosion of steel fasteners has occurred during laboratory
tests when composite materials made from carbon bre
reinforced plastic (CFRP) were joined to metallic sheets.
Fasteners made from titanium and nickel-based alloys can be
used without problems because of their high resistance to
corrosion. As additional protection against galvanic corro-
sion, screws made of titanium alloys are anodically oxidized,
but the ion vapour deposition of aluminium coatings cannot
be recommended for CFRP structures as the voltage potential
CFRP-Al lies at approximately 0.9 V, and the fastener is the
anodic part to the large surface area of the cathodic
CFRP. The harmful effect of coupling CFRP with a generally
corrosion-resistant alloy is illustrated in Fig. 5.11.
5.2.7 Contamination and Organic Fastener
Lubrication Systems
The control of contamination has led to increases in the
reliability and precision of spacecraft. Much effort is made to
257
utilize materials, particularly organic ones, with very low
outgassing properties. It is also necessary to ensure particle
cleanliness by employing a contamination control procedure
involving the use of clean rooms, laminar airflow facilities,
clean packaging, special monitors, and the like as described
in Sect. 4.21.
Liquids and greasesone aspect which should not be
overlooked is the generation of contamination from fastener
devices which may have been procured to high-reliability
aircraft standards. Standard oils and greases applied for
corrosion protection and against galling are unsuitable for
space use as they quickly volatilize or boil away under
conditions of high vacuum, either directly into space or
possibly to condense onto the spacecraft surfaces. Contam-
ination from organic sources can also collect on the walls
and glass windows of the huge vacuum chambers used to
expose an actual spacecraft to high vacuum and solar sim-
ulation as part of its ground testing phase (see Fig. 5.12).
Suitable lubricants for spacecraft fasteners are listed in
Table 5.2 and further information may be obtained from
textbooks such as that compiled by Fox et al. (2010).
Solid lubricantsmolybdenum disulphide-coated fas-
teners nd extensive applications in spacecraft mechanisms;
they are resistant to most cleaning solvents, but this layer can
be removed during plasma cleaning as was shown in Fig. 4.
90. MoS2 coatings are applied to commercial fasteners and
other sliding contacts by burnishing and resin bonding, but
great care needs to be taken if these are selected for space
applications as they are known to have poor adhesion and
wear characteristics that generate large volumes of contam-
inant particles. They may be suitable when incorporated in
special space-approved greases (Table 5.2) as the MoS2
powder is bound-up by the grease. However, if space
mechanisms engineers request a solid lubricant, then MoS2
can be applied, either by RF sputtering, or using a new
technique called high ion current density magnetron sputter
ion plating (Fox et al. 1999). As will be seen in other
sections of this book, MoS2 coatings degrade when in con-
tact with humidity. Space mechanisms that incorporate
MoS2 perform well in vacuum (but are best shielded from
solar radiation). Long in-orbit lifetimes have seen this nish
to be the most preferred dry lubricant. It is essential that such
mechanisms and fastener nishes are bagged in dry
nitrogen-lled bags during storage. Operation in humid air
causes an increase in friction coefcient, a greatly reduced
lifetime and widespread generation of contaminant debris.
A relatively new coating, based on a co-deposit of titanium
with MoS2 (registered and patented as MoSTTM) is known to
have far improved mechanical and tribological properties in
air environments that MoS2. This coating includes a titanium
content of about 15 % and a 1 m thickness is recommended
(Teer 2001) for terrestrial applications. TEM studies of this
coating indicate that titanium is present in solid solution
258 5 Metallography Applied to Spacecraft Test Failures

Table 5.2 Organic lubricants suitable for use under high vacuum
Trade name Chemical composition Type of product Typical outgassing Manufacturer
TML RML CVCM
SILCO-30 Silicone Thermally conductive grease 0.070 0.040 0.010 Cos, France
0.080 0.060 0.010
0.070 0.050 0.010
0.100 0.100 0.030
0.040 0.040 0.000
FOMBLIN-Z25 Perfluoralkylether Oil 0.010 0.010 0.000 Montedison, Italy
0.010 0.010 0.000
0.060 0.060 0.000
0.050 0.040 0.020
0.060 0.060 0.010
FOMBLIN Perfluoropolyether Grease 0.140 0.120 0.030 Montedison, Italy
Y-VAC 0.670 0.630 0.130
DC-C61103 Silicone Grease 0.060 **** 0.020 Dow Corning, USA
0.170 **** 0.010
0.170 **** 0.010
GRA-D85BIS Molybdenum Lubricant 0.000 **** 0.100 Graphoil, France
disulphide 0.800 **** 0.100
0.460 **** 0.000
Molykote-X106 Molybdenum Lubricant (should be oven baked 0.300 0.180 0.010 Molykote, USA
disulphide for best CVCM) 0.490 0.140 0.000
0.740 0.520 0.040
0.450 0.170 0.000
Apiezon-L Hydrocarbon Grease lubricant 0.150 0.120 0.020 Shell, The Netherlands
0.280 0.260 0.040
Teflon, Delrin, Polytetrafluoroethylene Mouldable thermoplastic solid 0.230 **** 0.000 Du Pont, USA
Halon, etc. (PTFE) lubricant 0.420 **** 0.010 Hoechst, Germany,
etc.
0.400 **** 0.000
0.410 **** 0.010
Braycote Polymer with MoS2 Grease 0.120 0.110 0.010 Bray Oil, USA
3L38MS
Tiolube 460 Molybdenum Dry lubricant 0.280 0.050 0.000 Triodize Inc., USA
disulphide 0.100 **** 0.060
TML Total mass loss. RML Recovered mass loss. CVCM Collected volatile condensed material. **** Not measured

within the MoS2 lattice; this causes lattice strain and results
in a coating with high hardness.
Graphite is a material that is both a good electrical con-
ductor and a good lubricant. However, graphite acts elec-
trochemically like a noble metal and, as stated previously, is
cathodic (corrosive) to all structural metals. Because of its
extreme corrosiveness, graphite lubricants in any form
should not be used on spacecraft equipment. Diamond-like
coatings (DLCs) are very hard, with high internal stress, and
are also unsuitable for space-use due to their brittle nature,
poor adhesion and generation of contaminant particles. More
research is being conducted into DLCs which may result in
an improved wear resistant coatings.
5.2.8 Metallic Particle Generation
Fastener platings and coatings can prevent galling and cold
welding of the kind illustrated in Fig. 2.4. Silver is primarily
a lubricant for high-temperature operations, and can be
5.2 Fasteners
Fig. 5.11 Photograph shows results of 8-day salt spray tests. Carbon
bre composite sheet was sandwiched between titanium and aluminium
6061 alloy. Only the aluminium has suffered extensive galvanic
corrosion (almost total wasting of 0.5 mm thickness)
Fig. 5.12 Condensed organic contamination on the high-vacuum side
of a thick glass window after a satellite solar simulation test
applied to high-strength steels (see Table 5.1) but not to any
of the titanium alloy fasteners. Soft metals such as lead,
which can be easily sheared, are also applied by ion plating
to fasteners not expected to support heavy loads. As previ-
ously mentioned, vapour-deposited aluminium is now
replacing cadmium platings. The ion plating of fasteners
with titanium nitride (TiN) is also beginning to take place,
particularly on titanium and on stainless steel fasteners to
reduce pick-up and sticking (Saunders 1988). Metallic par-
ticle contamination has, however, been noted to originate
from fastener platings and coatings as shown in Figs. 5.13
and 5.14. Great care must be taken to ensure the total
removal of metal slivers, particularly when electrical devices
259
are being mounted. A visual inspection of the astronomical
facilities for the detection of new X-ray sources mounted on
board a scientic satellite revealed the presence of a ne
metallic dust. Closer examination showed that this contam-
ination had been generated by the scoring action of lock-nuts
applied to the copper-plated threads of antivibration mounts.
Laceration of the mount threads resulted in extremely ne
flakes of plating becoming detached, as highlighted in
Fig. 5.14. These flakes contained bands of the mild steel
thread material as shown in the enlarged cross-section (d) of
the copper sliver. Under spacecraft conditions of
zero-gravity and strong electrical elds, these magnetic
particles would be free to float and cause a variety of
equipment malfunctions. There was insufcient time before
launch to redesign this item, so an alternative solution was to
apply a thin coating of a viscous, space-qualied silicone
lubricant to the threads before assembly of the lock-nuts.
The silicone grease was seen to agglomerate all the gener-
ated flakes; however, the design of this locking assembly
remained defective.
The advantages of beryllium (see also Sect. 5.7) have
been exploited for the fabrication of a wide range of
spacecraft mechanism and structures. In recent years the
assembly operations for beryllium structures have success-
fully incorporated bolts, screws, and blind bolt fasteners.
The discovery of minute particles of beryllium within and
adjacent to the drilled and tapped holes of a bearing housing
manufactured from hot-pressed beryllium has given great
concern, particularly as this contamination was detected on
cleaned flight unit parts. There seemed every possibility that
the small particles of beryllium shown in Fig. 5.15 could be
released from the vicinity of every drilled hole, and that such
particles would interfere with the smooth running of the
bearing parts. In retrospect, it is clear that special design
criteria and manufacturing requirements must always be
adhered to during the manufacture of beryllium piece-parts.
This is particularly important for the production drilling and
fastener assembly spacecraft structures. Certain prerequisites
will include:
Fastener holes must incorporate clearance for the fastener
head-shank radius.
Surface defects and partially released grains introduced
by drilling and tapping must be removed by chemical
etching (e.g. 20 % nitric acid by volume, 1 % sulphuric
acid by volume, 79 % deionized water for 34 min at a
bath temperature of 35 C).
The female screw-threads must be designed to take
account of a chemical etching rate of 12 m per minute.
Interference ts between fasteners and hole are not per-
mitted; they can give rise to further microcracking and
particle detachment.
260 5 Metallography Applied to Spacecraft Test Failures
 5.2 Fasteners
b Fig. 5.13 Metallic particles generated from spacecraft fasteners.
a Thousands of 8 m thick nickel chips had flaked off these M5 steel
fasteners from a microgravity experiment. (Note, nickel is magnetic and
these chips could accumulate under zero-g in regions of high electrical
Fig. 5.14 Damage by the
scoring action of lock-nuts.
a Copper slivers on damaged
thread. b Copper slivers retained
in silicone coating on thread.
c and d Microsection to reveal
copper-plated mild steel thread;
the detached particle
contamination is of copper
containing a band of mild steel
5.2.9 Quality Assurance Controls for Fasteners
The preceding subsections have described a wide variety of
individual case histories related to European fastener failures
and problems during both fastener manufacture and space-
craft assembly. NASA has also experienced a number of
fastener problems as were compile during a technical
exchange meeting in the mid-90 (Gamwell 1995). Out of
240 lots of fasteners inspected at one NASA establishment
over a two-year period, 42 lots were rejected (i.e. 18 %)
during routine incoming inspection tests. The classication
and numbers of these nonconforming fasteners are shown in
Table 5.3. As part of an overall space project Product
Assurance Scheme it is essential that Quality Assurance
(QA) is applied to the procurement and assembly of all
fasteners whether they be nuts, bolts, screws, threaded studs,
or rivets. The minimum requirements for the manufacturing,
inspection, and quality control of treated fasteners for ESA
spacecraft are dened in ECSS-ST-70-46 (2009) as stated in
261
or magnetic elds.) b Silver slivers generated from the torque loading
of steel bolts into silver-plated steel stiff-nuts; these cause electrical
short-circuits and must be removed by vacuum cleaning
Sect. 5.2.1. As a brief overview of European fastener stan-
dards, the QA controls must ensure the following aspects:
(a) Fastener design, material, thread sizes, and surface
nish must comply with national standards and draw-
ings (e.g. LN 29950, Cheese head screws with internal
serrations, made from titanium alloy).
(b) Qualication and acceptance tests acceptance/rejection
criteria are well dened (e.g. LN 65072, Fasteners
made of titanium alloystechnical specication).
(c) Acceptance tests are performed on all lots of fasteners,
both at the manufacturers plant (results will be recor-
ded on certicates, e.g. following DIN 500493.1B)
and at the purchasers plant as part of the in-house
incoming inspection procedure.
(d) Qualied manufacturers are selected from recognized
national lists (based on the successful history of manu-
facture of aerospace fasteners) and following a quality
audit of the manufacturing facilities. Process
262 5 Metallography Applied to Spacecraft Test Failures

identication documents (PIDs) must exist, and no

changes to the design, material composition, sequence of
thermomechanical treatments, nishes, etc. can take
place without the product being requalied to national
standards or the purchasers own space product
specication.
(e) Incoming inspection procedures for space quality fas-
teners must be available at each user company, in the
form of a written specication that includes: a sampling
plan (e.g. bolts selected at random (AQL) from each
lot), full details of the test methods, and details of
acceptance and rejection criteria.
(f) Nonconformances are reported by means of a project
material review board.
(g) Purchase orders for the procurement of spacecraft fas-
teners should be reviewed and signed off by the com-
pany quality assurance department. Care must be taken
to avoid the purchase of counterfeit fasteners by
procuring directly from manufacturers or known and
trusted stockists.
(h) All fastener materials used for spacecraft have to be
listed and reviewed for acceptability by means of the
project Declared Materials List.
(i) High-performance fasteners should be controlled in the
following manner (Roach 1984): material cleanliness to
national specications for metal composition; all criti-
cal surfaces of the fastener, e.g. bodies and underhead
llet, must be ground; threads rolled after heat treat-
ment for fatigue resistance; underhead llets cold
worked for fatigue resistance. Controls are generally on
samples by microsection and metallography, with
Fig. 5.15 Minute particles of beryllium released from drilled and 100 % magnetic particle, or fluorescent penetrant
threaded inserts on a bearing housing (and collected after ultrasonic inspection for cracks, laps, etc. The nal fastener ver-
cleaning on a Millipore lter) ication may include extensive property determinations
such as fatigue life, tensile strength, shear strength, and
Table 5.3 Classication of nonconforming fasteners rejected over a notch sensitivity, etc.
two-year period by one NASA establishment (Gamwell 1995)
Nonconformance No. of lots
Dimensional (threads) 8
Wrong material or nish 6 5.3 Thermal History from Microstructure
Mismarked or no marking 5
Halfway through the trial ring of an apogee boost motor
Lot integrity 4
(ABM), the outer case material was observed to breach and
Cracked 4
fail catastrophically. An anomaly was known to exist in the
Workmanship 4 ablative case-liner which separated the propellant from the
Dimensional (nonthread) 2 Ti6A14V case material, but it was considered important to
Running torque 2 perform a metallurgical failure analysis to ensure that no
Missing features 2 local defects had been present within the titanium alloy case
Damaged 2 before rupture.
Fastener cut to length 1 Two metals samples were carefully cut from the case
wall: one adjacent to the failure initiation site and the other
Nutplate inserts fall out 1
from a region through which the crack had propagated. The
Failed tensile 1
fracture surface at the origin of the fracture was
5.3 Thermal History from Microstructure 263

Fig. 5.16 Origin of ABM case fracture (material Ti6A14V)

photographed (arrowed in Fig. 5.16), and both samples were

then mounted, polished, and lightly etched in a solution of
HF, HNO3, and water to reveal the materials microstructure.
The case-wall thickness is clearly seen to have necked
down before failure. Such extreme ductility of this particular
alloy will occur only within the temperature range 800
1000 C, when it is known to undergo superplastic defor-
mation. Ti6A14V is highly superplastic at about 950 C; it
does not grain-coarsen or cavitate, and is the current
superplastic mainstay of the aerospace industry (Pearce
1987). The most signicant observation made during this
examination was the existence of two markedly different
microstructures within the failed case material. Well away
from the rupture initiation site, the microstructure,
Fig. 5.17a, is seen to consist of the typical room temperature
stable ( + ) structure. However, in the vicinity of the
rupture, a wholly martensitic () structure exists, Fig. 5.17b.
The ternary phase diagram of this alloy dictates that it is
necessary for the Ti6A14V composition to be heated to a
temperature between 910 and 1000 C for the ( + )
structure to transform completely into the single
Fig. 5.17 Etched cross-sections reveal local microstructure variations.
high-temperature phase. a Typical microstructure of the case, 2-phase ( + ), possessing small
In the partial phase diagram of this alloy system shown grain size. b Microstructure at origin of fracture is large grain and fully
in Fig. 5.18, the fully region is denoted by location A, martensitic (see also partial phase diagram in Fig. 5.18)
264
Fig. 5.18 Partial phase diagram
of the Ti6A14V system and
schematic representation of
microstructures resulting from
fabrication of Ti6A14V at
various temperatures (from
Mil-Handbook-697A 1974)
The microstructures which result from heat-treating this
alloy in various manners are shown schematically. On slow
cooling from the phase, the resultant ambient tempera-
ture microstructure would appear as a mixture of marten-
site () and primary . On fast fooling to ambient
temperature, the phase is replaced by a totally martensitic
() structure.
Phase diagrams are the metallurgists blueprints. For
any combination of metals, a characteristic form of phase
diagram is obtained. For instance, when a piece of polished
alloy is viewed under a microscope and then heated or
cooled at its equilibrium rate, the observer will notice
microstructural changes which may include phase modi-
cations corresponding to the alloys phase diagrams. In
practice, alloys are rarely cooled at equilibrium rate, and, as
shown in Fig. 5.17, divergences from the conditions indi-
cated by the phase diagram are frequently considerable.
From experience and consultation of other data, e.g. time
temperaturetransformation diagrams, it is possible for the
metallurgist to assess metastable structures which result from
nonequilibrium cooling.
Examination of the failed ABM case proved that there
were no defects in the case structure and that it had been
5 Metallography Applied to Spacecraft Test Failures
heat-treated correctly. However, as a result of abnormal
burning of the propellant, a hot spot occurred in the case
which reached a temperature of 800 C. This caused the case
to plastically deform and aggravate the liner problem. The
microstructure shown in Fig. 5.17b illustrates that the rup-
ture initiation site must have reached a temperature of
approximately 1000 C for the ( + ) to transform com-
pletely into phase. At this temperature, the case initiated a
ssure which propagated owing to the cases internal pres-
sures. After case rupture, the hot spot must have cooled
rapidly, so that the phase fully transformed into the
metastable martensitic structure.
5.4 Effect of Inclusions Within
the Microstructure
of Explosively Deformed
Material
Pyrotechnic or electro-explosive devices are used exten-
sively in spacecraft to push or pull loads, to pull switches
and pins, and to cut cables, wires, and release mechanisms.
5.4 Effect of Inclusions Within the Microstructure of Explosively Deformed Material
Fig. 5.19 Pyrotechnic cutter found cracked (arrowed) and leaking
after ring
Various forms of such explosive-actuated devices are
available. They consist generally of a cylindrical body
containing a pyrotechnic explosive charge and an igniter
which can be red by lead wires. The high-pressure gases
which are developed inside the cylinders as a result of ring
cause a piston or bellow to move forward within a few
milliseconds. As piston actuators they can move loads of up
to 750 N over distances of approximately 15 mm, they can
sever structural metal up to 6 mm thick, and cut wires and
bolts used as launch restraints, commonly up to 6 mm
diameter and supporting 1 tonne in tension (Cable 1988).
Figure 5.19 shows the end of a red satellite cable cutter
cylinder; the round imprint has been caused by a piston rod
attached to a cutter blade (their relative positions have been
superimposed, but see also Fig. 3.25, the neutron radiograph
of a similar device). It is highly desirable to prevent spent
explosive from contaminating the spacecraft or its environ-
ment by ensuring that these devices retain their hermeticity
after ring. The Inconel alloy cylinder shown in Fig. 5.19, as
with many prototypes, developed cracks (arrowed) during
ring, and this resulted in the escape of a considerable
quantity of contaminants.
The metallographically prepared section shown in
Fig. 5.20 reveals a high propensity of microporosity
extending through the thickness of the cylinder wall.
Additional examinations and testing determined the mode
of cylinder failure. The cylinders were found to have been
machined from commercial purity Inconel 600 in bar form.
This NiCrFe alloy exists as a stable solid solution. The
rather high carbon, silicon and sulphur content which is
265
Fig. 5.20 A transverse section highlights severe porosity
permitted by commercial specications can give rise to
inclusions of chromium carbide and nickel sulphide. Lon-
gitudinal sections made from bar stock showed such inclu-
sions to be numerous and as having generally a high
length-to-width ratio. Closer examination of the leaking
cartridge revealed that each microvoid in the structure was
associated with an inclusion (see Fig. 5.21a). These voids
have formed by the pressure differences caused when the
shock wave of the explosion passed over the inclusion and
created microcracking, separation, and cavitation at the
inclusion-to-matrix interface (Verbraak, private communi-
cation). It is thought likely that cracks have formed after
ring owing to the high gas pressures within the cylinder.
These cracks are seen to propagate through the Inconel in a
ductile transgranular manner, linking the microvoids and so
forming a continuous leak path.
Rigorous quality controls involving inclusion counting by
metallographic techniques were imposed to ensure that only
high-purity clean alloy was procured for the manufacture
of additional cylinders. These were successfully tested, and
their hermeticity and freedom from microvoiding is attrib-
uted to the almost total absence of inclusions within the new
stock material.
Recent developments have made it possible to make
very accurate quality controls on the surfaces of micro-
sectioned materials such as those used for pyrotechnic
devices. A computer-backed electron beam microanalyser
can be used to store large amounts of information con-
cerning so-called element distribution proles (EDP).
Whereas pyrotechnic cylinder materials are now assessed
by chemical analysis coupled with metallography to indi-
cate the distribution of second phases and inclusions, it is
likely that in the future many materials selected for critical
266 5 Metallography Applied to Spacecraft Test Failures

Fig. 5.21 a The detailed

examination shows inclusions
(arrowed) surrounded by
microvoidsthe major crack has
propagated in a ductile manner as
evidenced by the deformed
Inconel grains. b The element
distribution prole showing the
content relief of silicon
(inclusions) from a polished
cross-section of alloy

applications will be quality controlled by EDP facilities 5.5 Degradation of Passive Thermal
(Motz 1988). EDPs can provide the starting points for Control Systems
predicting the properties of explosively deformed materials,
particularly when microsegregation through the grain 5.5.1 General Background
structure of a metal alloy causes solid solution strength-
ening to localized regions of the lattice. An example of an The intense emission of light and heat from far ultraviolet
EDP is illustrated in Fig. 5.21b. to far infrared is the most obvious of all solar features,
5.5 Degradation of Passive Thermal Control Systems 267

and gold on aluminium can reach values approaching 10.

A general schematic for and values of surfaces is given in
Fig. 5.22. Black anodized aluminium (inorganic dye to
prevent degradation under vacuum as prescribed in ESA
PSS-01-713) and a black paint such as Cuvertin 306 which
can be applied according to ESA PSS-01-721 are known as
flat absorbers and possess / ratios very close to 1. Values
of / for spacecraft materials under solar radiation are given
in Table 5.4, but casual application of these data has to be
avoided, as these values are averages over a large number of
samples. Variations may be caused by surface contaminants,

Table 5.4 / for a selection of surfaces and nishes (see also test
method ECSS-Q-ST-70-09C, 2008)
Surface Absorptance Emittance /

Gold/Kapton 0.3 0.03 10.0
Gold plate on Al 7075 0.3 0.03 10.0
Polished beryllium 0.4 0.05 8.0
Gold on aluminium 0.26 0.03 6.5
Polished silver 0.07 0.02 3.5
Polished stainless steel 0.5 0.13 3.85
Polished aluminium 0.25 0.1 2.5
Polished copper 0.28 0.13 2.2
(oxidized)
Fig. 5.22 Generalisation of and for common spacecraft surfaces Vapour-blasted stainless 0.6 0.33 1.8
(from Gilmore, Stuckey and Fongthermal surface nishes, NASA
steel
document)
Epoxy black paint 0.95 0.85 1.12
Black-anodized 0.94 0.85 1.11
particularly above the Earths atmosphere. Thermal control aluminium
is a major factor affecting spacecraft reliability, and it is Acrylic black paint 0.97 0.91 1.07
essential for the satisfactory operation of all types of
Silicone white paint 0.19 0.88 0.22
instruments. This is also one of the most serious of space
After 3 years u.v. 0.39 0.88 0.44
materials problems. The methods by which control of tem- irradiation
perature might be achieved can be classied as either active
Silicate white paint 0.14 0.94 0.15
or passive. Active systems include electrical, mechanical,
After 3 years u.v. 0.27 0.94 0.29
and chemical mechanisms such as heat pipes based on liquid
irradiation
ammonia to generate and distribute heat energy. Passive
Silicon solar cell, bare 0.82 0.64 1.3
techniques are static in nature and include the surface n-
Silicon solar cell, silica 0.82 0.81 1.0
ishes chosen for structural materials, optical solar reflectors,
cover
protective foils such as aluminized Mylar and Kapton, or
Silicon solar cell, silica 0.78 0.81 0.96
second surface mirrors. In the vacuum of space all exchange cover, blue lter
of heat between the spacecraft and space will be via radia-
Silicon solar cell, silica 0.7 0.81 0.86
tion. As already discussed in Sect. 2.3, the radiation char- cover, red lter
acteristics of surfaces are dependent on the ratio between
Kapton (5 mil)/aluminium 0.48 0.81 0.6
their solar absorptivity and emissivity properties (/ ratio).
In2O3/kapton/aluminium 0.4 0.71 0.56
If stable nishes are made available the temperature control
problem is markedly reduced and mathematical models can Quartz fabric/tape 0.19 0.6 0.3
be generated to optimize designs. OSR (quartz mirror) 0.08 0.81 0.1
silvered teflon
Solar reflectors are approximated by white paints (/ less
than 0.2) containing oxide pigments. Solar absorbers with an FEP (5 mil)/silver 0.11 0.78 0.14
/ ratio greater than 1 include most bare metallic nishes FEP (2 mil)/silver 0.05 0.62 0.08
268
surface roughness, manufacturing lot number, ageing effects,
and a lack of adequate processing controls.
The major challenge to the materials engineer is to ensure
that the / ratio for any surface remains stable. Organic
coatings may be not be expected to withstand the combined
forces of ultraviolet and particle radiations, temperature
cycling, and high vacuum; these factors can cause noticeable
colour changes to such surfaces. Metallic surfaces can be
modied by thermal fatigue, corrosion during ground stor-
age, sublimation, and contamination, but they are not
affected by the particle radiation of space. Unless special
attention is given to manufacturing process controls and the
choice of nishes, the / stability of metallic nishes may
be degraded as shown in the following case histories.
Spacecraft thermal control paints and coatings (some
being electrically conductive), usually black or white, but
many having different colours, should be procured from
specialist companies. These include:
For paints (mostly used on ESA and NASA spacecraft):
many paints, MAP, Pamiers, France;
many paints, Henkel, Dusseldorf, Germany;
Nextel Black, Acktar Ltd., Kiryat-Gat, Israel, and
Hohenaspe, Germany;
many paints, AZ Technology, AL, USA;
many paints, Lord Corporation, Manchester, UK, and
NC, USA;
many paints, HTRI, Technology Research Institute, Illi-
nois, USA.
For coatings and conversion coatings:
Aluminized polyester adhesive foils, RUAG Space
GmbH, Vienna, Austria.
Kepla-Coat (previously Plasmocer), AHC, Berlin,
Germany;
coatings for discrete components and roll/sheet for
application, Acktar Ltd., Kiryat-Gat, Israel, and Hohe-
naspe, Germany;
Black anodizing, Hegin Metal Finishing BV, Heerde, the
Netherlands;
Colinal, (black inorganic metal dye applied with an
electric current for UV stability), Ateco, Rheinfelder,
Switzerland, and Steiger Galvanotechnique SA,
Chatel-St-Denis, Switzerland;
Ebonol Black, Jackson Plating, Hounslow, UK;
white, grey and black highly stable PEO coatings, Ker-
onite Advanced Surface Technology Ltd., Haverhill, UK.
various, Sheldahl Corp., Northeld, MI, USA
Type II anodizing, with black inorganic metal dyes
applied with an electric current for UV stability, 2Dye4
Anodizing, Inc., Rohnert Park, CA, USA.
5 Metallography Applied to Spacecraft Test Failures
It is recommended that all paints and coatings are
assessed in specialized test houses by exposing samples to
relevant environments prior to their selection and applica-
tion. These may be for LEO and/or GEO applications.
Thermo-optical properties will include absorptivity, emis-
sivity, and / ratio. Handling and mechanical properties
might include adhesive strength, abrasion and cleaning
properties, shedding of contaminant particles and suscepti-
bility to fungal growth. Chemical properties might include
resistance to radiation and atomic oxygen, flammability,
outgassing-under-vacuum, off gassing and resistance to UV.
Thermal and electrical properties will include resistance to
thermal cycling (with consideration to substrate), maximum
and minimum operational temperatures, thermal conductiv-
ity and electrical conductivity.
5.5.2 Low-Emissivity Surfaces
A gold-plated (low emissivity, ) aluminium alloy panel
supporting the electronic boxes of a satellite transmitter and
receiver was noted to have different optical properties (/
ratio) before and after testing. The aluminium panel had been
nished with initially copper (6 m), silver (20 m), and
nally gold (1030 m). It was suspected that the gold layer
shown in Fig. 5.23 was porous, and that the optical property
change was due to surface silver sulphide formed by the
presence of trace hydrogen sulphide in the industrial atmo-
sphere of the testing site. The resulting laboratory examina-
tion demonstrated the presence of a porous gold layer by
placing part of the panel into concentrated nitric acid.
Streams of bubbles emerged from the pores, account for by:
Ag 2HNO3 AgNO3 H2 O NO2 " gas
Metallography revealed that the gold had been deposited
in an arboriferous fashion, and, as seen in Fig. 5.23b, it is
very porous. The gold-plating procedure was revised by
increasing the current density of the bath so as to produce
more nucleation sites, and alkaline brighteners were added to
the electrolyte to ensure the growth of ner gold grains.
Many plating trials were required before an acceptable
procedure could be established. It was found benecial to
continually lter the alkaline gold cyanide electrolyte
through a 2 m lter core, and oil-free air was used as the
principal means of agitation. The electrolysis slowly pro-
duces unwanted organic contamination by cyanide elec-
tropolymerization, and as this phenomenon can also produce
rough deposits the bath was monitored daily by Hull cell
tests. The nal flight equipment panels were noted to possess
extremely stable optical properties due to the brightness of
the gold nish and its freedom from irregularities such as
surface roughness, voids, pits, and visible contamination.
5.5 Degradation of Passive Thermal Control Systems
Fig. 5.23 a SEM micrograph showing large-grained, severely porous
gold plating. A marked change in the reflectivity of this totally
unacceptable nish resulted from the migration of silver sulphide
tarnish which emanated from an intermediate silver layer which, as seen
on the photomicrograph (b), is partially exposed to the atmosphere
In a separate, unrelated investigation, Sharma (1989) also
investigated the problem of silver-based gold plating on
spacecraft aluminium panelsagain, porous gold was noted
to turn the coating black owing to the reaction with atmo-
spheric sulphides, resulting in the degradation of optical
properties. As a consequence, Sharma developed and pub-
lished a sequence for zincating, electroless nickel plating,
gold striking, and gold plating on aluminium. He produced a
gold nish with a very low infrared emissivity (0.020.03)
which can withstand both very low (196 C) and very high
(150 C) temperatures, together with no changes in optical
properties after exposure to humidity, thermal cycling, and
thermal shock in liquid nitrogen.
269
5.5.3 High-Absorption Surfaces
Certain surface areas of spacecraft require a high absorp-
tance () over the solar spectrum. This is particularly true for
those surfaces which will be periodically shaded from the
Sun, to protect these regions from cooling down to extre-
mely low temperatures; these surfaces should, in addition,
have a low emittance () to reduce thermal radiative heat
loss. A black-nickel plating has been extensively reported
and shown to have good solar selective properties (Pettit and
Sowell 1976). It was incorporated into the design of certain
spacecraft equipments. Although the optical properties of
these equipment surfaces were initially very stable, subse-
quent environmental testing which incorporated thermal
cycling between 100 and +120 C caused degradation.
Metallography revealed that the nickel was in fact very
brittle, and because of the difference in expansion coef-
cients between this plated surface and its aluminium sub-
strate, a network of cracks soon formed (see Fig. 5.24), and
further extensive thermal cycling caused the nickel to spall
and flake off. Alternative black nishes to metals have been
subsequently approved after the successful completion of
environmental testing. All have desirable solar selective
properties and can be applied to all common spacecraft
structural metal surfaces. They include:
Black chromiuma durable decorative nish which is
extensively used in the electroplating industry and pre-
viously assessed by NASA (McDonald 1974).
Black cobalt coatingsa possible candidate, but unless
all cobalt is converted to oxide the residual cobalt-rich
phase will be ferromagnetic (Vitt 1987).
Ebonol blackcopper plating a substrate and then
electrolytically black-oxidizing it in a proprietary
solution.
Copper sulphide black inorganic coatingchemically
deposited with a brous texture having particularly sharp
microcrystalline surface developed for satellite mass
spectrometer (Saxer 1977).
Black-anodized aluminiumwhich utilizes an inorganic
dye; pure aluminium can be deposited on various sub-
strates and later black anodized (Hegin B.V., the
Netherlands).
Colinal 3100developed by Alusuisse and applied by
Steiger S.A., Switzerland, for the ISO spacecraft, ano-
dized AA 6061 is black-coloured with stannous sulphate.
Black oxide coatings on ferrous metalsalkaline and
fused salt oxidizing processes are specied in
MIL-C-13924.
A solar collector developed for space power applications
has been developed by NASA Glen (Jaworske 2002). High
270
Fig. 5.24 Topography of black nickel plating, a as-received and
b after a satellite thermal cycling test. The crazing and subsequent
spalling of this brittle nish results from the differential expansion
coefcient between nickel and its aluminium substrate (13.3 106 and
25.3 106 per C, respectively)
solar absorptance is needed in order to collect as much
incident solar radiation as possible, and low infrared emit-
tance is needed to keep radiant energy losses to a minimum.
The collector surface coating is sputter coated with a cermet
layer composed of molecular islands of metal (aluminium)
embedded in alumina. This coating produced a solar
absorptance () of 0.797 and an infrared emittance () of
0.131, yielding an / ratio of 6.
Colinal 3100 coated on Al-6061 has been extensively
tested (Puippe 1999) in a wide range of environments and
found to have remained rather stable after an equivalent
6.5 years exposure to a low Earth polar orbit. The initial
solar absorptance () was 0.91 and infrared emittance ()
was 0.85. These values changed to 0.91 and 0.85 respec-
tively after the exposure to the polar orbit environment; a
5 Metallography Applied to Spacecraft Test Failures
decrease in / ratio of only 3 %. Similarly, exposure to a
simulated atomic oxygen test (7.2 1021 atoms/cm2),
equivalent to a LEO exposure of 1.4 years, produces negli-
gible change in / ratio. Puippe (1999) indicates that his test
results show Colinal 3100 (anodized then pigmented by
electrolytic tin) to be consistently better than the anodized
and black inorganically dyed (nickel or cobalt sulphide
pigmentation) formulation given in ESA PSS-01-703 [now
superseded by ECSS-Q-ST-70-03, which in the 2008 issue
C, states that black anodizing should not be used on alu-
minium alloys containing high amounts of copper (5 %),
zinc (6 %) or silicon (5 %)].
5.5.4 Rigid Optical Solar Reflectors
Optical solar reflectors (OSR) used for thermal control are
nding extensive applications on research satellites. They
are usually made of thin quartz or cerium doped glass with a
rear side layer of vacuum-deposited highly reflective silver
protected by Inconel which can be bonded onto the space-
craft structure. These small thermal control mirrors, maxi-
mum area 80 mm2, are bonded to satellite panels with
high-purity silicone adhesives (e.g. CV 2566 is used on
Envisat and the Hubble Space Telescope). In the mid-1970s
it was realized that nonconducting spacecraft surfaces can
charge up to potentials of several kilovolts, and satellite
electrostatic cleanliness is presently a new and fashionable
space discipline (Winkler 1975; Fellas 1982). The charging
mechanism in geostationary orbit is complex and is gener-
ated by the low-energy electron environment at this altitude.
Charging events occur in the outer magnetosphere and dis-
charges have cause the anomalous behaviour of more than
twenty spacecraft and at least one catastrophic failure
(Koons et al. 1980). Antennas would also change gain or
orientation spontaneously and inexplicably. Some satellites
such as Intelsat III and IV and MARECS A would suddenly
start spinning because of spurious signals and would have to
be commanded to despin. It has therefore become necessary
to apply conductive coatings to as many nonconducting
satellite surfaces as possible. Both solar cell covers and
optical solar reflectors (OSRs) are therefore often coated
with a conductive but transparently thin layer of indium
oxide. These conducting surfaces have been developed to be
uniform and stable to all mechanical, thermal, and envi-
ronmental influences of the spacecraft, and show a minimum
resistance of approximately 40 k/m2. Some scientic
satellites have incorporated four small contacting spots on
each OSR to electrically connect rows of OSRs. Each spot is
built up from metallized layers of titanium, palladium, and
silver, and each contact spot is interconnected to an adjacent
OSR by means of a welded silver wire. The multimetal layer
5.5 Degradation of Passive Thermal Control Systems
Fig. 5.25 The complementary fracture surface of a circular metallized
(TiPdAg) contact pad which had lifted off a quartz optical solar
reflector. Fracture, by thermal fatigue, originated in location X. This is
shown in a as the rough-cut edge of the quartz, which acts as a stress
raiser. It was recommended that the quartz edge should be
chamfered-off. The vacuum-deposited metal pad b lifted, with a crack
front seen from the beach marks to be initially symmetrical with
respect to the origin of fracture X. Residual stresses, coupled to those
induced by thermal cycling, account for the direction changes of crack
propagation. The pad nally was attached only in the location A.
Optical fractograph
was designed to ensure electrical continuity with the thin
indium oxide conductive coating and to provide a weldable
surface which once joined to a silver wire would not degrade
by formation of brittle intermetallics, corrosion, or
solid-state diffusion. The various metallic layers are applied
by vacuum deposition at 250 C so that on cooling it is
probable that the quartz, which has a coefcient of expansion
only one third of that of silver, will contain residual tensile
stresses at room temperature. This thermal expansion mis-
match did cause extensive thermal fatigue failures, particu-
larly after thermal cycling as shown in Fig. 5.25. Once the
various metals were applied in exceedingly thin layers (i.e.
1000 1 Pd and 10 m Ag) it appeared that the differences
1 (angstrom unit) = 1010 m.
1
271
in expansion rates could be taken up by plastic deformation
of the pad materials. A full qualication programme has
demonstrated that a 100 % conductively coated spacecraft
which is really electrostatically clean can be achieved.
Recommended reading about fractography of brittle mate-
rials, particularly glasses, are the works of Quinn (2008).
5.5.5 Flexible Second Surface Mirrors
Special thermal blankets are fabricated from Kapton or
Mylar lms that have internal layers that are metallized on
both sides by vacuum-deposited pure aluminium. These
blankets are widely used both for thermal control and as a
means of avoiding electrostatic charging. The surface pos-
sesses a low solar absorptance and a low absorptance and a
low emittance, and can be considered as a total reflector.
Surface modications to Kapton lms have been devel-
oped by Sheldahl Inc. (2004, etc.); they have a very thin
duplex layer of indiumtin oxide deposited on one side of
the Kapton and an aluminized rear surface. The
thermo-optical properties of such lms have been measured
to be = 0.39, = 0.77, providing an / ratio of 0.5 (i.e. for
Kapton lms having a thickness of 50 m). This value is far
better than those thermo-optical properties of other second
surface mirrors, coupled with the fact that the surface
resistivity is very low and therefore suitable for the avoid-
ance of electrostatic charging (Verdin and Duck 1985).
The rigid OSRs described in Sect. 5.5.4 consisted of
individual units (20 40 mm) which were each grounded or
interconnected by four metallized conductive tabs on each
corner of the indium oxide surface. This method provided
for a high reliability and ensured that ESA research satellites
such as GEOS, ISEE-B and SOHO remained electrostati-
cally clean during their lifetimes. Venus Express had OSRs
mounted on the rear of the solar array panels and between
the solar cells on the front of the panels.
This type of panel is, however, very expensive, and
requires considerable labour, owing to the mounting and
interconnection techniques. For these reasons the cheaper,
flexible, second surface mirrors based on indium
oxide-coated polymer are preferred for large communica-
tions satellites; they can also be attached to such spacecraft
in the form of multilayer insulation blanket. Unfortunately
it is extremely difcult to ground such metallized lms.
Welding by electrical resistance pressure techniques proved
unsuccessful owing to the fragile nature of the coatings,
and many of the electrically conductive adhesives were
also found unacceptable owing to an increase in joint
resistivity with time that was probably related to a galvanic
corrosion mechanism. A novel method of grounding has
been developed with a conductive silicone adhesive based
on RTV 566 mixed with CHO-bond 1029B. The method
272 5 Metallography Applied to Spacecraft Test Failures

uses a special heated probe, similar to a soldering iron but

with a flat end face. The probe is held by a ceramic grip,
and the hot end face can be applied, under various loads, to
effect an elevated temperature curing of the adhesive. The
bond lines generally have a thickness of only 0.1 mm, and
the curing takes place at 160 C under a load of 200 g. The
electrical grounding of aluminized Kapton tape and
indiumtin oxide-coated Kapton flexible second surface
mirrors has been space-qualied where this technique is
used (Bosma et al. 1978, etc.).

5.6 Sublimation of Metals

5.6.1 General

Current space hardware requires structural materials capable

of offering satisfactory mechanical properties in the range
150 to +150 C. This has led to aluminium alloys, titanium
alloys, and carbon bre reinforced epoxy resin composites
becoming the standard spacecraft materials, much as they
have in aircraft construction. Several spacecraft systems will
experience higher temperatures, such as the structural parts
associated with an apogee boost motor, propulsion units, and
electronic assemblies including heaters and cathodes
employed within travelling wave tubes. It is not always
possible to predict the exact temperatures expected of these
items, but they can lie in the range from 200 to 1400 C, and
under high-vacuum conditions it is important to consider the
effect of material loss by vaporization on properties such as
mechanical strength or electrical conductivity over the parts
operational lifetime. The process by which a solid substance
passes into the vapour state from the solid state without
passing through a liquid phase is known as sublimation. This
can occur when a metal is close to its melting point and has a
vapour pressure not far removed from the external pressure.
By decreasing the pressure, all metals will sublimate, and the
rate of metal loss will depend on the temperature and on the
ameliorating effects of coatings, alloying agents, etc. (Fran-
kel 1969). Individual sublimation rates for metals were listed
in Table 2.5, and curves relating the vapour pressure of a
metal to a specic temperature range are given in Fig. 5.26
and in chemistry handbooks. Fig. 5.26 a Vapour pressure of a metal as a function of temperature
(after Frankel 1969). b Loss of the metals (Cd, Zn, Mg, and Al) as a
As an example to be considered further in Sect. 7.2.2, it
function of temperature (after Frankel 1969)
will be noted from Fig. 5.26 that at 2800 C tungsten
requires a pressure of 105 mm Hg (or 0.00133 Pa) for it to
vaporize. Spacecraft materials will not be generally con- connectors (Griner 1968) and why magnesium can become
cerned with these very high melting point metals (two fur- troublesome at temperatures greater than 200 C.
ther important exceptions are, however, discussed in the The evaporation rates of the more volatile metals are
following case histories), but the vapour pressure curves do shown graphically in Fig. 5.26b. It is difcult in practice to
illustrate why cadmium or zinc plating is not a good nish make actual measurements of the evaporation rates of
for the corrosion protection of components such as electrical volatile materials, and it is more common to describe the
5.6 Sublimation of Metals
redeposition of metallic elements by means of a quartz
crystal microbalance (QCM). This device has a sensitivity of
109 g/cm2. According to Griner (1968), cadmium began to
redeposit at 70 C and produced a lm thickness of up to
130 . She found that, under high vacuum, rapid vapor-
ization occurred before the sample reached 100 C, well
below the 321 C melting point of cadmium. Other work has
found that cadmium and zinc coatings sublimed only
between temperatures of 150 and 260 C, and that chromate
lms did not act as a retarding barrier to the sublimation
rates (Wolff 1963).
Several of the modern high-strength, lightweight alloys
and metalmatrix composites that could nd applications in
the next generation of space vehicles are composed of what
appear, from the tables and graphs, to be elements having
high sublimation rates (i.e. magnesium, lithium, zinc, and
aluminium). Sublimation of these materials, as already
mentioned, can cause them to lose weight, lose mechanical
strength, and increase in electrical resistivity. On the other
hand, the recondensation of metallic lms can cause
immense problems to electronic and optical systems. The
sublimation rates of the pure metals can be roughly assessed
from the graphs of vapour pressure versus temperature.
When they are covered with porous coatings (i.e. chemical
conversion lms), or plated layers where solid-state diffusion
can take place, it is almost impossible to assess the risk of
sublimation. Similarly, the sublimation rates of alloys are not
well understood. They probably depend on the composition
of the phases that make up the alloys microstructure, the
activity of the grain boundaries, and the occurrence of nat-
ural barrier layers that might hinder evaporation.
A method to evaluate the sublimation of metals by a
weight loss test has recently been developed. It is based on a
test for the outgassing of organic materials (ESA PSS
01-702) and assesses the mass loss of metals and alloys
under thermal vacuum at temperatures up to 700 C, mea-
sured on-line with a vacuum microbalance (Merstallinger
and Semerad 1996). The initial part of the test establishes the
dynamic sublimation rate; here, the temperature is slowly
increased until a mass loss of 1 mg is achieved, and then
slowly decreased. A static sublimation rate is then deter-
mined as a function of time at constant temperature. During
the static part of the test, individual samples are retained
under vacuum (less than 105 torr) and quickly reheated to
the desired temperature which is held until a constant sub-
limation rate is achieved. The static test is performed at two
temperatures (at least) and the sublimation rates calculated as
a function of temperature. Also the activation energy for
sublimation is calculated. Condensation rates can also be
measured. Some preliminary results from the sublimation
tests can be given as follows:
273
Aluminiumlithium alloys
The binary alloy (Al3Li), ternary alloy (Al3Li1Cu), and
a commercial alloy (AA 8090 obtained both as Ospray
material and in the as-rolled condition) have been tested. All
show similar sublimation and condensation characteristics.
Initially during the dynamic test, there is a mass increase due
to the formation of a surface layer of lithium oxide and then
measured sublimation parts. It appears as though this oxide
layer acts as a barrier, but once broken down, sublimation
can occur at lower temperatures. During the rst temperature
cycle, layer growth was observed between 350 and 400 C,
but during this rst dynamic test, condensation of some
sublimated material commenced at 300 C. During the
second and third runs (static test) the onset of sublimation
was at a lower temperature than the preceding run. The
measured sublimation rate for each of the aluminium
lithium alloys was in the order of 1 102 g/cm2 per year at
400 C, and 5 g/cm2 per year at 600 C (Met. Report 2241).
These sublimation rates occur at high temperatures and that
are not exceeded in-flight for space hardware. However,
even if higher temperatures are reached, these alloys are
encased by means of anodic coatings that act as protective
barrier layers as seen in Fig. 5.27.
Magnesiumlithium alloys
These alloys have the lowest density of the light alloys
(between 1.3 and 1.5 g/cm3). They have been characterized
by Hornbogen and Schemme (1990) and Schemme and
Wittkamp (1993), and would appear to possess sufcient
strength for certain spacecraft applications (e.g. secondary
structures). However, as was seen in Table 2.5, both mag-
nesium and lithium possess rather high rates of sublimation.
Microbalance tests were performed under vacuum on sam-
ples of the alloys Mg5Li (alpha phase), Mg8Li (alpha-beta
phase), and Mg10Li (virtually fully beta phase). As was the
case for the AlLi alloys, a barrier surface layer formed in
the MgLi samples as they were heated in vacuum. On
reaching a temperature between 350 and 400 C the barrier
breaks down and sublimation rates become very high. Each
alloy sample was cooled and then maintained at a set tem-
perature whilst mass loss measurements were made. At
about 300 C the sublimation rate for each of these alloys
was found to be 1 g/cm2 per year. At 400 C the sublimation
rate increased to more than 10 g/cm2 per year (this value lies
between the sublimation rate for pure magnesium and that
for pure lithium).
Gapasil brazing alloy
Gapasil is a trade name of the US Wesgo company and
covers a range of brazing alloys based on silver with addi-
tions of gallium and palladium. These alloys are particularly
274
Fig. 5.27 Orion crew module pressure vessel. This structural element
is machined from aluminium lithium alloy Al-2195 (same as Space
Shuttle lightweight external tank) and protected with green chromated
useful, as they have a relatively low melting temperature for
silver-based alloys and reasonably good metallurgical com-
patibility with titanium alloys. One Gapasil brazing alloy
was utilized to join a complex assembly of interlocking
titanium and beryllium piece-parts during the manufacture of
an X-ray detector housing for a scientic satellite. The alloy
selected was Gapasil-9, with a composition of Ag9Ga9Pd, a
liquidus temperature of 880 C, and a solidus temperature of
845 C. Gallium is a rare metal with a melting point of only
30 C. Remarkably, gallium also does not boil at atmo-
spheric pressure below 2300 C.
Owing to the extremely stringent cleanliness require-
ments assigned to the detector housing it was considered
advisable to assess the sublimation properties of Gapasil-9.
This alloy was considered for vacuum brazing operations at
920 C, and during service life, the detector housing parts
would occasionally reach a temperature of 200 C. The test
apparatus would not reach 920 C, so tables in the Metals
Handbook were consultedthey showed that at 920 C,
pure gallium would sublimate at 1 g/cm2 per year. This was
approximately the same rate as that for the sublimation of
silver at the same temperature. The brazing cycle only
caused the Gapasil to be liquid for less than two minutes, so
virtually no evaporation or condensation of metallic gallium
was to be expectedthis was conrmed during inspections
after the brazing operation. The sublimation test was per-
formed at 258 C for a duration of 10 h using the in situ
5 Metallography Applied to Spacecraft Test Failures
primer coating. These welded panels still have to be covered with
thermal protection tiles (Courtesy NASA)
microbalance system. No sublimation was detected. In a
further test run at 392 C for 10 h, the sublimation rate was
measured to be less than 2.8 105 g/cm2 per year. These
tests conrmed that the braze alloy would be suitable for
vacuum applications.
Tungsten Heater Elements
Heaters for indirectly heated cathodes used in spacecraft
travelling wave tubes are made of tungsten wire. This wire is
very slightly doped with thoria to prevent sagging and
recrystallization. These tungsten wires are designed for use at
1200 C under conditions of high vacuum. Some life test tubes
were destructively analysed after 44,092 h operation (a little
more than ve years). Substantial thinning of the coiled wires
due to sublimation was observed and, interestingly, it was seen
that the evaporated tungsten atoms had condensed in cooler
areas of the travelling wave tube heater wire so as to form small
tungsten whiskers (further described in Sect. 7.2.2).
5.6.2 Sublimation of and Condensation
of Cadmium and Zinc
One of the largest vacuum test chambers in Europe was
recently used for the testing of a space structure under
conditions of thermal cycling between 100 and +120 C at
low pressure (less than 106 torr). The structure consisted of
5.6 Sublimation of Metals 275

many metallic Invar struts onto which experiment housings

had been attached. The Invar parts were nished with a black
paint. After extensive thermal vacuum testing the chamber
was opened and the structure removed. It was at this time
that the test engineers observed that the internal walls of the
test chamber had changed colour. The chamber walls are
actually shrouds, nished with a coating of Chemglaze Z306
black paint. These surfaces had changed from black to grey.
In a similar way, the black Electrodag 501 surface of
infrared detection equipment also installed in the test
chamber was also covered in a grey deposit. A Materials
Review Board team was formed to investigate the deposit,
the source of the deposit, to develop a method to remove the
deposit, and to assess the effect of the cleaning method on
the Chemglaze Z306 black paint.
Material was removed from the black chamber shrouds and
analysed by scanning electron microscopy and EDAX. The
general morphology of the paint is seen in Fig. 5.28athe
black paint nodules have been contaminated by a myriad of
ne columnar deposits. An EDAX analysis of the general
contaminated surface is seen in Fig. 5.28b and when the
background composition of the paint was subtracted from this
X-ray spectrum it became obvious that the cause of the con-
tamination was an unknown source of metallic cadmium and
zinc. The offending source was later found to be the Invar
structurethis had been electroplated with both cadmium and
zinc, as corrosion protection layers, and then black-painted.
The experiment manager had not declared the presence of
cadmium or zinc on his materials list; this was to be an
extremely costly omission.
A method for the removal of both cadmium and zinc from Fig. 5.28 a SEM view of condensed acicular metallic contaminants on
the shroud and infrared (IR) detector surfaces was developed surface of painted shroud at 500. b EDAX analysis in small grain
and documented. The contaminant metals could be success- region of shroud, identifying the presence of cadmium, zinc, sulphur,
chromium, chlorine, and sodium on its surface (courtesy of Dr A. de
fully dissolved in a 10 % solution of ammonium nitrate Rooij)
without any damage to the black-painted surfaces. Applica-
tion of the solution was by swabbing and the entire chamber
was refurbished (owing to the toxic nature of cadmium and spacecraft or ground-handling applications that involve
cadmium oxide, this work was completed in the presence of a exposure to vacuum.
safety ofcer). The cleaned chamber was then baked out at A series of experiments involving the laboratory subli-
100 C for 24 h. Witness plates were placed in the chamber mation test was performed to quantify the sublimation rates
and a shroud, cooled by liquid nitrogen, was positioned for cadmium. Various samples of cadmium coatings were
opposite to what had been the most severely contaminated evaluated. Interestingly, the presence of a chromate con-
wall. No cadmium or zinc could be detected and these metals version coating on cadmium-plated steel (as is applied, for
were judged to have been completely removed. However, it instance, to commercial electrical connector shells as a
required much additional effort in order to remove the traces corrosion protection layer), or the presence of naturally
of ammonium nitrate solution from the chamber walls. occurring oxide lms on aged cadmium surfaces, did not
This experience with the effect of cadmium and zinc reduce the rate of cadmium sublimation. At 120 C a sub-
sublimation/condensation clearly reinforces the strict rule limation rate was found to be 0.3 g/cm2 per hour. The
that these metals must continue to be forbidden from all ndings of these tests have provided semi-quantitative data
276
which account for the cadmium contamination inside the
environmental test chamber described above.
5.6.3 Heater Sublimation Problem Associated
with Thruster Motor
Telecommunications and meteorology satellites, in geosta-
tionary Earth orbit, are able to alter their orientation (altitude
control) and positioning (orbit control) by means of
propulsion systems based on fuels such as either monopro-
pellant hydrazine or a bipropellant of nitrogen tetroxide with
monomethyl hydrazine. Small propellant motors, known as
thrusters, usually provide a pulsed operational mode for
altitude control and a continuous operational mode for orbit
control. Monopropellant hydrazine thrusters derive their
energy from the decomposition of liquid hydrazine into
nitrogen, ammonia, and hydrogen.
2N2 H2 liquid ! N2 H2 2NH3 highlighted in Fig. 5.32. EDAX analyses of the microsection
The hydrazine decomposition may be achieved by cat-
alytic and/or thermal means. A strongly exothermic reaction
results in the formation of hot gases that can be expanded in
a nozzle to convert their thermal energy into kinetic energy
and thereby produce the thrust necessary to reposition the
spacecraft. The catalyst employed in thruster motors is
iridium, which is dispersed over porous aluminium oxide
pellets so as to provide a large surface area; see also Sect. 5.8
concerning the deactivation of this catalyst. The catalyst
pellets are located in a cylindrical bed surrounded by three
specially designed heaters.
After the in-orbit failure of several thruster heaters, and in
view of the fact that similar heaters were being physically
integrated into the flight units of follow-on satellites, a
detailed failure investigation was initiated. The in-orbit
failures were identied from onboard temperature and cur-
rent measurements to be due to electrically open circuits.
They occurred approximately 13 months after successful
operations following the launch of this satellite.
The heater designs incorporate a 0.025 mm diameter
Nichrome V coiled element which is welded to a 0.25
diameter platinum lead. This heating coil is supported by
alumina spacers and is surrounded with 0.040 mm alumina
powder particles. The heater coils and ceramic parts are
packed into a cylindrical Inconel housing as shown in
Fig. 5.29. The leadwires are embedded in an Inconel 600
tube containing magnesium oxide for insulation. After
assembly the heaters are backlled with dry argon to prevent
oxidation.
A literature search established that similar heater designs
had failed owing to inadequate joining of the lament coil to
the lead wires, and to a loss of the protective argon gas
(Schatz et al. 1979).
5 Metallography Applied to Spacecraft Test Failures
A large number of heater devices were tested by thermal
vacuum and leak testing. Those devices showing leaks
(Fig. 5.29) were then submitted to a simulated life test: 900 C
heater body temperature, heater operation for 96 h at 22 W/in2
power, this test being performed under a vacuum of
105 mm Hg. It was established that nonleaking heaters sys-
tematically would pass this test, whereas leaking heaters
would fail because of open circuits.
Failed devices were submitted to high-denition
X-radiography (see Fig. 3.23), metallography, and SEM
inspection. Radiography identied several types of defect as
shown in Fig. 5.30. Metallography revealed cracked brazes
and welds as seen in Fig. 5.31. SEM examination of the
heater lament coil conrmed that the loss in wire resistance
is a result of chromium and nickel sublimation caused by
operation of the lament in vacuum. These metal elements
were found to have redeposited throughout the heater inte-
rior, and the coils had thinned to the point of total fracture as
made of this lament (Fig. 5.33) and others revealed that
chromium depletion had progressed in advance of the sub-
limation of nickel. This can be expected by reference to the
plots of sublimation rate against temperature given in
Fig. 5.34.
The welding and brazing techniques used to fabricate
these heaters have been changed to eliminate the possibility
for the inert argon gas to leak out. Other improvements to
manufacturing and testing, including leak tests to 108 cm3/s,
have been implemented to extend operational life to at least
10 years in space.
5.6.4 Sublimation of Klystron Cathode-Heaters
A large number of cathode heater assemblies were subjected
to long-term life-testing. They had been made to standard
military production procedures but with extra quality
assurance surveillance. The device is a scaled-down version
of a typical klystron cathode heater gun; it is designed for the
same voltage as a travelling wave tube gun, and its fabri-
cation route was considered as a dry run for the manufacture
of space hardware designated for Earth resources pro-
grammes. A number of metal sublimation problems were
identied during the course of the life-testing.
One failed device was found to have a detached heater
leg; it is compared to an intact leg in Figs. 5.35 and 5.36.
The high operational temperatures of these cathode hea-
ters (approximately 1185 C) have caused the platinum
return leg ribbons to undergo excessive grain growth.
According to the engineering drawing the leg should be
separated from the alumina ceramic by an air gap of 1.5 mm.
Grain growth has caused the failed ribbon to weaken and sag
under its own weight, causing it to make contact with the
5.6 Sublimation of Metals
Fig. 5.29 Heater investigation. a The heater assemblies were
subjected to the Gross Leak Test specied in Mil-883 (Method 1014
test condition C). Each complete heater was immersed into fluorocar-
bon at 125 C. Leaking heaters were identied by streams of bubbles
issuing from both plug and transition junction (arrowed) demonstrating
leakage in excess of 103 atm. cm3/s. b Schematic cross-section of
cylindrical molybdenum heater container. Solid-state diffu-
sion of platinum into the contacting molybdenum (arrowed
on Fig. 5.35b) was conrmed by microsectioning as a
wear-out mechanism. The platinum ribbon subsequently
thinned, became overheated, and caused an acceleration in
sublimation rate from the large-grained region. It is expected
that the platinum ribbon had operated at a temperature of
1500 C for approximately one year before its failure. Fig-
ure 5.37 illustrates how sublimation and preferential thin-
ning at the platinum ribbon grain boundary have caused this
return leg to overheat and fail by fusion.
From the viewpoint of sublimation (see Table 2.5) it is
recommended that parts fabricated from pure platinum and
intended for long-life applications (e.g. in excess of eight
years under vacuum) should not be operated at temperatures
greater than 1100 C. Accurate vapour pressure data for the
277
heater subassembly showing positions where housings were cut (by ne
turning on a lathe). The delicate lament wires were not disturbed. The
welded plug ends were removed and the ceramic discs extracted from
cylinders. Certain of the brazed seal portions with jumper wires were
withdrawn
platinum metals are given by Arblaster (1996). Improve-
ments to this cathode heater were made by lowering the
molybdenum cylinders so that they could no longer contact
the platinum ribbon. It is envisaged that greater reliability
would be achieved by replacing the pure platinum ribbon by
one made from dispersion-strengthened platinum [known as
ZGS and developed by Johnson Matthey (Selman et al.
1974)]. The following benets can be expected:
A higher recrystallization temperature (with less grain
growth during operational life).
The sagging, or high-temperature creep, of the return
leg would be reduced or eliminated.
ZGS platinum is easier to form than pure platinum.
Higher resistance to shock loads (e.g. launch vibration).
Less chance of tearing in weld heat-affected zones.
278
Fig. 5.30 X-radiographs
showing a cracks in Inconel
heater housing, and b severe
thinning of Nichrome heater coil
on device with incorrectly welded
end plug
It is emphasized that the direct replacement of the con-
ventional materials with ZGS platinum can be made only
after a careful investigation has ensured that the redesign
will not affect the electrical characteristics of the cathode
heater.
5.6.5 Sublimation of Rhenium
Recently developed thruster motors incorporate a coiled
length of rhenium heater tube. The tube is heated by means
of its own electrical resistance such that it reaches temper-
atures approaching 2000 C. This electrothermal thruster
device produces thrust by expanding the gases of decom-
posed hydrazine (ammonia, hydrogen, and nitrogen) by
heating as they flow through the tube. The temperature of the
5 Metallography Applied to Spacecraft Test Failures
tube can be varied by changing the power input to the heater.
Test programmes have been completed under conditions of
high vacuum showing the potential suitability of rhenium for
this application. However, certain precautions must be taken
to prevent material deterioration of the rhenium tube of the
kind illustrated in Fig. 5.38.
It was found essential to avoid any oxidation of the
rhenium tube while in the hot, powered condition.
Rhenium surfaces rapidly pick up oxygen to form
unstable, low-melting-point oxides. Rhenium in its pure
state has a melting point of 3170 C. However, like other
refractory metals, rhenium is oxidized both in air and
under partial vacuum at elevated temperatures. These
resulting oxides, particularly the higher ones, are volatile
and have very low melting points (e.g. Re2O7 melts at
only 296 C).
5.6 Sublimation of Metals 279

Fig. 5.31 Identication of leak

paths allowing argon to escape
from thruster heaters. a Brazed
end seals showing inadequate
volume and penetration of braze
alloy between Nichrome sheath
and space ring (inset shows
cracked braze).
b Photomicrograph of welded end
plug. Sample has been etched to
differentiate between plug
material and Inconel housing
(section is slightly away from
heater central axis). Note the very
shallow depth (0.25 mm) of weld
penetration and leak path from
inside of heater, alongside of plug
to atmosphere via the microcracks
(arrowed) (63)

Metallography indicated that testing of one of the tubes at
about 2000 C for 300 h caused a reduction in its wall
thickness by 0.04 mm. If it is assumed that sublimation of
pure rhenium accounts for this thinning effect, then this will
occur on both internal and external tube surfaces (e.g. a
sublimation rate of 0.02 mm per 300 h). This is equivalent to
6 102 cm/year. A review of Table 2.5 shows that this
sublimation rate can occur if the rhenium is held at tem-
peratures in excess of 2050 C.
As the cross-sectional area of tubes decreases, either
because of oxidation or because of sublimation, so their
electrical resistance increases to promote even higher oper-
ational temperatures and a further increase in the material
loss. Intense enlargement of rhenium grains occurs above
2000 C (grain growth). This has resulted in tube failures, as
shown by Fig. 5.38, by the degradation of the part by
accelerating sublimation rates particularly along exposed
grain boundaries,
excessively high temperatures of operation,
grain boundary sliding once grains have the same
diameter as the tube wall thickness.
280
Fig. 5.32 Appearance of lament coils after heater life testing under
vacuum. Leaking devices a showed excessive sublimation, whereas
those heaters that retained their internal argon gas b remained intact
Fig. 5.33 Microsection made through thinned Nichrome lament
element shown in Fig. 5.32a. Note excessive grain growth and internal
cavitation. EDAX analysis revealed remaining wire composition to be
mainly nickel with only traces of chromium (1000)
5 Metallography Applied to Spacecraft Test Failures
Fig. 5.34 Plots of sublimation rate against temperature for Nichrome
wire (80Ni20 Cr)
It is probable that rhenium tubes are suitable only for
long-life electrothermal thruster applications when exposed
to decomposed hydrazine (internal flow) by ensuring an
external environment of ultra-high vacuum and by lowering
the operating temperature to well below 2000 C. Alterna-
tive materials, including platinumyttria and platinumzir-
conia, have also been evaluated for a similar thruster
application. They were found to be generally compatible, on
the basis of mass loss, with the multipropellant capability of
resistojets for space station auxiliary propulsion units
(Whalen 1988).
5.7 Beryllium for Spacecraft Applications
5.7.1 General
Beryllium is a metal with a better combination of strength
and stiffness properties than most other materials. It has a
particularly good stiffness-to-weight ratio, high specic heat,
excellent thermal conductivity, and a good resistance to
fatigue cracking, and it possesses the ability to maintain
useful properties at both high and low temperatures. It would
appear that beryllium has found most usage in USA space-
craft applications. This is due to a European consideration
that the metal is both brittle and difcult to fabricate, as well
as being fairly toxic. These may be misunderstood factors, as
5.7 Beryllium for Spacecraft Applications
Fig. 5.35 a Intact return leg device is not in contact with
molybdenum. b Failed return leg device has contacted and diffused
into molybdenum. Note (i) Excessive platinum grain growth in region
above ceramic. (ii) Intergranular fracture on (b). (iii) Roughening and
wasting away of platinum along arrowed region on (b)
experience in production methods by some advanced aero-
space companies has led to the metal being used in a broad
range of applications from satellite structures to optical
platforms (Case et al. 1984; Damskey 1988; Grant 1983).
Beryllium is almost entirely compacted by the powder
metallurgy route in order to achieve the ne-grained (5
15 m) microstructures necessary for successful machining
and suitable mechanical properties. Modern manufacturing
techniques now produce blanks by hot isostatic processing;
281
they have a near-net-shape and considerably lower the cost
of fabricating complex parts.
The major factors affecting physical and mechanical
properties are grain size, grain orientation, and purity, par-
ticularly with respect to oxygen (O), silicon (Si), iron (Fe),
and aluminium (Al). The highest strength, greatest ductility,
and greatest resistance to crack propagation occur when the
grain size is ne and equi-axed.
5.7.2 Health and Safety
Beryllium machining and metal working require several
steps to control the inhalation of dust and fumes which
otherwise could result in acute or chronic berylliosis. Ingress
of the metal into the skin must also be avoided, or skin
reactions can occur. National laws usually stipulate that
personnel are monitored to ensure that they do not exceed
certain limits. If sufcient care is taken with the control of
processes involving beryllium metal, or its oxide, beryllia,
the risks can be very low. The most important control is to
ensure that individual, local exhaust ventilation is able to
collect all the dust and fumes when beryllium is being
machined. Contamination outside the site is prevented by
careful ltration of exhausted air, decontamination of all
metal and oxide surfaces, control of effluent entering the
drains, and control of all persons employed on the site.
The recommended limit quoted by the British Health and
Safety Executive is stated to be 0.002 mg of beryllium or its
oxide per cubic metre of air (8 h TWA value). This means
that all incoming components should be checked for loose
dust before the batch is accepted by the supplier.
Great care must be practised by the material engineer
concerned in the production, handling, or metallographic
preparation of test specimens made from beryllium or beryllia
components. Special containers must be kept for the collec-
tion of al waste material, which should be suitably sealed and
labelled and later disposed of by the company Health and
Safety Ofcer. It is self-evident that eating, drinking, and
smoking should be prohibited in the work areas.
Beryllia is nding increased usage in the electronics and
electrical industries because of its special properties of being
an electrical insulator whilst being a highly effective con-
ductor of heat. This means that assemblies may incorporate
electrical components such as frames for modules and
packages, or substrates for hybrid circuits and power devices
made from beryllia. It has been found that beryllia, owing to
its matched thermal expansion coefcient with ceramic
packaged components, is more suitable than copper as a heat
spreader. Also, the thermal resistance and heat capacity of
beryllia and copper are essentially the same. For these rea-
sons the cutting, grinding, and polishing operations of
282 5 Metallography Applied to Spacecraft Test Failures

Fig. 5.36 SEM examination of

intact leg shows some
sublimation. a 100. b Transition
from large equi-axed grains to
small grains. Platinum thinning
by sublimation occurs
predominantly at grain
boundaries. These are effectively
cracks (arrowed). c Detail of
grain boundary-to-free surface
illustrating metal removal
5.7 Beryllium for Spacecraft Applications 283

Fig. 5.37 Detail of the fracture of the failed leg seen in Fig. 5.35b.
Sublimation and thinning at grain boundaries networks have caused
these interfaces to fuse (arrowed) and become detached

standard destructive physical analyses can be hazardous

unless full details of all the component assembly materials
are known and special precautions are taken.

5.7.3 Integrity of Machined Beryllium

Metallic components that are required to possess inherent

properties of high strength and resistance to fatigue must be
machined, or surface treated, to the correct shape, dimen-
sions, and surface nish. Beryllium is particularly sensitive
to surface damage classied as resulting from microstruc-
tural changes, cracking, and residual stress. It is well known
that beryllium, because of its hexagonal close-packed atomic
structure, is susceptible to twinning of the microstructure to
varying depths below the surface. When twins are present
they can intersect with neighbouring grains to create
microcracks which grow along preferred crystallographic
planes. Twins are produced to various subsurface depths by
the different machining techniques employed. They may also
form after a period subsequent to the machining operation
when the residual microstresses are relieved by local
deformations that create twins.
The results of an investigation into the effects of milling Fig. 5.38 Example of failed rhenium tube from electrothermal thruster
on subsurface structures are shown in Fig. 5.39. This work device after very-high-temperature operation under vacuum. a SEM
fractograph of failed tube showing absence of rhenium deformation at
and the following text relating to the photomicrographs are fracture surface. Cleavage of the large grain structure is mainly by
reproduced by courtesy of W.C. Heraeus GmbH. Directly intercrystalline crack propagation. b Detail from external vacuum
beneath the surface, a heavily twinned zone was observed exposed surface arrowed on (a). Surface appears strongly etched by
extending to a depth of between 40 m for a surface milled grain boundary depressions and wavy markings. c Optical micrograph
of cross-section made in region of (b). Surface was nickel-plated to
with heavy cuts (0.5 mm cut) and 15 m for a ne-milled retain edge of the hard rhodium surface during microsectioning. Note
surface (0.025 mm cut). In each case individual twins were greatest sublimation occurs at grain boundaries, and certain grain
seen reaching to about three times this depth. Examples of orientations are more susceptible to sublimation
284 5 Metallography Applied to Spacecraft Test Failures

the twinned microstructure beneath the coarse- and

ne-milled surface are shown in Fig. 5.39a, b.
There are essentially two ways of removing the surface
damage and residual stresses caused by machining. As
shown in Fig. 5.39c, the twins can be removed by heat
treatment. A typical heat treatment consists of heating the
machined beryllium component to 790 C under vacuum at a
maximum rate of 300 C per hour, holding for 1 h, and
slowly cooling to room temperature at a rate of less than 50
C per hour. An alternative method for removing the surface
damage is to actually remove the surface zone by etching.
This treatment will also lead to a readjustment of the residual
stress level and equalize its distribution throughout the bulk
of the material, in certain instances equivalent to a thermal
stress-relief treatment. A range of etchants has been rec-
ommended for this purpose, and the choice of a particular
solution depends usually on the specic requirements (e.g.
high or low etching rate, matt or bright surfacesee
Sect. 5.7.5). A well-tried production route for machined
beryllium components consists of rough machining, heat
treatment, ne machining, and a nal etching treatment with
a removal of more than 0.1 mm from the surface. Fig-
ure 5.39d shows the subsurface zone of beryllium thus
treated, and Fig. 5.40 shows a beryllium component for an
interferometer which was produced in this way. As a further
measure to prevent distortion in service, the component is
subjected to a nal stabilization treatment in which it is
thermally cycled between extremes of temperature slightly in
excess of the temperature range anticipated in service.

5.7.4 Thermal Cycling on Work-Hardened

Beryllium

Attention has been focused on a problematic component for

a series of spacecraft apogee-pointing mechanisms and solar
array drive mechanisms. Close examination of these
as-delivered beryllium piece-parts showed their surfaces to
occasionally support wavy lines of small pits possibly
caused by the sintering and hot-pressing operation that
achieved a near-net-shape before machining to size.
Adhesive tape was applied to the outer conical surface of
the parts to cover both the wavy lines and the machined and Fig. 5.39 Subsurface microstructure of milled and stress-relief-treated
polished surface regions. The tapes were slowly pulled from beryllium (type HP 20). a Coarse-milled, 0.5 mm cut. b Fine-milled,
the surfaces and carefully inspected under stereo zoom 0.025 mm cut. c Coarse milled and heat treated 1 h at 790 C. d Coarse
microscopes for released beryllium particles. Many particles milled, heat treated, ne milled, and etched
were found to be adhering to the tape (see Fig. 5.41).
Each part was then cleaned to a standard practice by The beryllium part was next bagged and placed in an air
swabbing and immersion into an industrial solvent (1,1,1 circulating oven and exposed to ve slow thermal cycles
trichlorotrifluoroethane). Cleaning continued until the between 30 and +70 C. A further tape test over the suspect
flushing solvent was found, by ltration over Millipore l- area resulted in the release of a very large number of
ters, to contain no particles. The tape test was repeated, and beryllium particles from along the entire wavy line length. In
only one particle was seen to be released. a nal attempt to remove partially adherent particles the part
5.7 Beryllium for Spacecraft Applications 285

Fig. 5.40 Interferometer bed for

Jupiter space probe, produced
according to procedure in
Fig. 5.39d

was subjected to ultrasonic cleaning in a fresh solvent for

periods of three minutes. Each cleaning step was monitored
by the Millipore ltration method, and particle size analysis
was made of the collected debris. Once the number of
detached particles had been reduced to zero the surface was
considered clean. The part was submitted to a further ve
thermal cycles between 30 and +70 C. No particles were
released after a nal tape test.
The conclusions to be drawn from this case history are:

The standard cleaning method (immersion and swabbing)

for beryllium as-received parts is inadequate for the
removal of partially adherent beryllium particles.
Thermal cycling can cause the detachment of many
loosely adherent particles from non-etch, incompletely
cleaned piece-parts. (It is assumed that as beryllium is
anisotropic in nature, microscopic movements between
neighbouring grains having either a different degree of
work-hardening from machining, or a different grain
orientation, may tend to dislodge each other, particularly
in the vicinity of the wavy line defect.)
Whereas ultrasonic cleaning is found to be an improved
method of removing surface particles it is recommended that
all structural spacecraft parts made of beryllium are chem-
ically etched in order to remove surface layers of mechanical
damage (see the following section). Etching will remove the
subsurface twinned layer which could otherwise initiate
microcracks and cause the premature failure of structural
parts and the generation of metallic debris.
Incidentally, it is a requirement that all beryllium com-
ponents for Space Shuttle applications shall be chemically
etched, and in-process control specimens must be prepared
to ensure, by metallography, that all twinning has been
removed (Norwood 1985).
Fig. 5.41 Detail of one mound of released particles. Close inspection
shows them to be composed of needle-like slivers (all highly reflective
and thus metallic)
5.7.5 General Etching Solutions for Beryllium
The following solutions have been used successfully for the
removal of work-hardened layers from the surface of beryl-
lium. Trials are needed to determine the precise rates of
dissolution, as these will change as a function of temperature.
Solution A 2 % nitric acid, 2 % sulphuric acid, 2 %
hydrofluoric acid, balance distilled water (in-
formation from King et al. 1976).
Solution B 15 % ammonium bifluoride in water with
20 grams per litre beryllium.
Solution C 20 % nitric acid, 1 % hydrofluoric acid in water,
with 10 grams per litre beryllium (information
from Gurklis 1972).
286
5.7.6 Investigation of Microcracked Thin-Foil
Detector Windows
Beryllium foil is widely used for windows in spacecraft
X-ray and gamma-ray detectors because of its low absorp-
tion of these radiations. When vacuum hot-pressed beryllium
block is cross-rolled into sheet and foil its property values
are increased by cold work and the generation of crystallo-
graphic anisotropy. However, it must be noted that as rolling
progressively increases the properties of strength and duc-
tility in the plane of the sheet, the out-of-plane properties are
signicantly reduced. This behaviour became evident during
the assembly and ground operation of a low-intensity
gamma radiation detector before integration into a research
satellite. The detector under investigation incorporated, by
necessity, a window foil of only 3050 m thickness and
was required to house (hermetically sealed) a controlled and
stable internal pressure of argon gas. The detector was found
to fail during operation under vacuum by
(a) leakage of argon through the foil, and
(b) cracking during bowing of the foil as a result of dif-
ferential pressure across the window.
Mechanical testing and metallography determined that the
argon leakage failure resulted from microcracks in the foil
that had been introduced during the cold-rolling operation.
The internal argon pressure induced bowing of the foil
during operation of the detector under vacuum, causing the
microcracks to propagate through the thickness of this
35 m thick foil.
Various processing routes were evaluated with the
assistance of the Fulmer Research Institute to determine a
preferred rolling and heat-treatment sequence which might
produce a more isotropic foil that could bend or buckle
without the formation of microcracks. It may be noted that
beryllium has a hexagonal crystal structure whose basal
plane is both the twin plain and the cleavage plane. Twin-
ning is promoted by deformation on the basal plane and can
result in the formation of microcracks when twins intersect
grain boundaries (Tetelman and McEvily 1967). The flow
stresses for basal slip are remarkably low at room tempera-
ture and decrease further with increase in temperature (Beurs
et al. 1987).
Beryllium foil from the same batch as used for the leaking
detector was submitted to chemical analysis, mechanical
testing both normal to, and parallel with, the nal rolling
direction, bend testing, microsectioning, fractography, and
TEM evaluation at the Fulmer Research Institute.
Foils from the initial batch were later annealed for
20 min, some at 650 C and others at 710 C, in an attempt
to both remove any residual stresses caused by rolling and
5 Metallography Applied to Spacecraft Test Failures
improve the materials formability. Some minor improve-
ments in ductility were achieved by these laboratory heat
treatments, and on the basis of the results new foils were
fabricated following three different routes. The processing
history and selected laboratory test ndings are presented in
Fig. 5.42. The appearance of dislocation networks and
recrystallized grains are shown in Fig. 5.43, and from this
transmission electron microscope information it is clear that
rolling produces a characteristic microstructure having a
high dislocation density. Annealing processes at 600 C
produce a more well-developed subgrain structure and lower
dislocation densitiesthe microstructure indicated this
material to be partially recrystallized, whereas annealing at
650 or 710 C produces a more recrystallized structure
which is almost dislocation free.
Failure in the Type 1 foil from the defective detector is
due to its exceedingly low ductility, the presence of a rough
surface condition, and the existence of dislocation walls and
networks leading to transgranular cleavage. The chemical
composition of this material was within its specied
high-purity limits. Fracture of the 650 C annealed foil was a
mixture of transgranular cleavage and intergranular crack
propagation. Failure in the fully recrystallized material was
by intergranular fracturethis would suggest that during
annealing weakening of the grain boundaries and/or possible
strengthening within the grains had occurred. No precipita-
tion was observed within the grains which could produce a
strengthening effect, and it is likely therefore that the fracture
mode was due entirely to embrittlement of the grain
boundaries. Grain boundary embrittlement during annealing
is common in both structural steels and nickel-based alloys.
In both cases the effect is due primarily to the segregation of
phosphorus or sulphur, although other elements also play an
important part. The specic mechanism by which phos-
phorus, or any other element, weakens the intrinsic cohesive
strength of a boundary is not well understood. Very little, if
any, work has been done on grain boundary embrittlement in
beryllium. However, by analogy with steels or certain nickel
alloys, it is quite possible that the intergranular fracture in
the 710 C annealed beryllium is a result of grain boundary
embrittlement due to segregation of impurity elements dur-
ing the anneal. This should be an important area for further
investigation since the grain boundary monolayer composi-
tion can be easily determined by Auger electron spec-
troscopy, and this, in turn, may influence the mechanical
properties of the annealed material.
The conclusions drawn from this investigation are as
follows:
Samples which had been taken from unidirectionally
rolled foil stock have been found to have extremely
anisotropic mechanical properties and fully recrystallized
5.7 Beryllium for Spacecraft Applications 287

Fig. 5.42 Summary of processing steps for beryllium foil with resulting mechanical properties and grain structures
288
equi-axed grain structures of a dimension approaching
the foil thickness.
The cross-rolled samples (foil types 6 and 7) have a
uniformly high strength both parallel and transverse to
the foils long direction. This material has rather low
ductility (better, however, than the cross-rolled batch of
foil used for the leaking detectorfoil type 1).
The main recommendations were to advise that foil types
15 be rejected. Foil type 7 was later demonstrated to be the
most suitable, and has been successfully incorporated into at
least one series of gamma-ray detector housings.
5.7.7 Aluminium-Beryllium Alloys
Beryllium is one of the lightest aerospace structural mate-
rials, with a density of 1.85 g/cc, while aluminiums density
is approximately 45 % higher at 2.7 g/cc. A family of
low-density alloys is being developed based on aluminium,
with additions of between 10 and 75 % beryllium. They have
been found to combine the high specic stiffness of beryl-
lium with the ductility and ease of fabrication of aluminium.
The new alloys have been produced by powder and ingot
metallurgy methods at the Brush Welman company in
Cleveland, Ohio. One of the more successful binary alloys
was noted (Haschiguchi et al. 1992) to be Al40 wt%Be.
When cast and hot rolled, sheet material of this composition
had interesting mechanical properties: yield strength
289 MPa, ultimate tensile strength 386 MPa, and elongation
at fracture of 17 %.
5.8 Deactivation of Catalyst Particles
for Hydrazine Decomposition
5.8.1 Testing Procedure
The inflight experience of a telecommunications satellite
showed that when 0.5 N thrusters are operated in the
pulse-mode after a shut-off period of several days, the
impulse bit from the rst few pulses may be very low and
then progressively increase to the nominal design value.
Since the roll control loop of this spacecraft has been
designed to operate from the rst pulse onwards, it was
necessary to investigate the cause of this thruster
underperformance.
5 Metallography Applied to Spacecraft Test Failures
In order to study the compatibility of hydrazine with the
various component materials used in the construction of the
reaction control system (RCS) an investigation was initiated
to determine the gas generation rate of hydrazine versus
temperature and time by means of a long-term test utilizing a
space quality thruster motor. The test lasted 12 months, and,
after a throughput of 23 kg of hydrazine, resulted in a sig-
nicant deactivation of the thruster catalyst. This catalyst
degradation has been the focus of the following
investigation.
The test equipment consisted of a hydrazine tank fabri-
cated in Ti6A14V incorporating the expulsion diaphragm
elastomer type AF-E-332. The tank would be stored for the
duration of the test in a water bath held at 40 C. A full
description of the test procedure has been published by Dunn
and Papenburg (1986).
The test was terminated after a signicant thruster deac-
tivation and various materials investigations were initiated.
These included:
Opening the thruster at the injector end, then inspecting
for particle contaminants.
Detailed examinations of the used catalyst particles and
comparative studies with unused particles by visual,
scanning electron microscopy (SEM), surface analysis
(EDAX and XPS), mass spectroscopy, and metallography.
Emptying and opening the propellant tank to sample the
AFE-332 diaphragm material by X-ray and SEM analy-
sis. A full characterization of this elastomer has been
published by Martin (1976) and Schreib et al. (1979).
Analysis of hydrazine samples (that had been drained off
after the various thruster rings) by different methods to
determine the inorganic impurities.
5.8.2 Material Investigation
(a) Destructive analysis of thruster
The thruster drawing is shown in Fig. 5.44. The unit was cut
at the injector and nozzle ends in order to inspect for signs of
obvious blockages. No contamination was observed at the
injector tube end, and a mass flow measurement with
deionized water, according to a standard procedure, gave
results that were identical to those produced before the start
of the long-term test. This clearly proves that the injector
5.8 Deactivation of Catalyst Particles for Hydrazine Decomposition
Fig. 5.43 Transmission electron micrographs of beryllium foils.
a Typical recrystallized grain structure after both 650 and 710 C
annealing treatment (e.g. foil types 2 and 3). b Low-temperature
cross-rolled foil showing high dislocation density and well developed
networks (foil type 6). c Partially recrystallized grain structure after foil
type 6 is annealed at 600 C for 30 min (foil type 7)
289
tube was not blocked. A visual inspection of the catalyst
chamber screens, made from Haynes 25 Alloy, showed them
to have turned from grey to a whitish colour, but again there
was no sign of blocking.
(b) Characteristics of catalyst bed
Nature of catalyst. The catalyst particles are identied as
Shell 405 developed and manufactured by the Shell Devel-
opment Company. This catalyst is considered to be the most
successful for the controlled catalytic decomposition of
monopropellants such as hydrazine. Shell 405 catalyst is
prepared on carrier particles of preattrited Reynolds
RA-alumina which after special processing acquires a sur-
face composition of 32 % iridium (Schmidt 1984).
Visual inspection of catalyst grains. The nozzle-end
screen of the test thruster was removed and the catalyst
grains were carefully extracted. They were classied into
three zones (nozzle, middle, and injector), and these grains
were then compared with the visual appearance and were
within the same size range (0.51.5 mm diameter). They
were jet black with no further colorations. Visual inspection
of the used particles at a magnication of 20 showed them
to have changed colour, from the original black through grey
and grey/white to white.
(c) Morphology and composition of new and used catalyst
particles
SEM inspection. The new particles are shown in Fig. 5.45.
They have a very large surface area due to the facets and
crystallites that are present. Only the particle corners appear
almost polished and rounded-off. The new particle surfaces are
highly electrically conductive, and from the EDAX analysis
are composed of iridium and presumably aluminium oxide.
Particles extracted from the catalyst bed of the test
thruster were viewed in the SEM. The contamination layer is
present as a white amorphous cover on used grains. This is
shown in Fig. 5.46 to be composed of mainly silicon.
Metallographic results. Probably the most interesting
ndings are those given by the mounted, cross-sectioned,
and polished particles. The NEW catalyst particles are seen
in Fig. 5.47. These sections are dull, and not highly reflec-
tive. There is no evidence of any free metallic iridium on the
surfaces of these particles. Some small regions within the
grain indicate the presence of a metal in a ne lamellar form.
Line scan analyses identied an inhomogeneous distribution
of iridium through the particles.
Micrographs of a typical contaminated particle are
included in Fig. 5.48. The used grains have a markedly
290 5 Metallography Applied to Spacecraft Test Failures

Fig. 5.44 Engineering drawing

of 0.5 N thruster

different appearance from the new particles, being covered
with a dense, white, almost continuous shell. These shells
vary in thickness from one particle to another. The maxi-
mum shell thickness is about 80 m (this might be an
oblique cut through the layer), but it is generally 10 m
deep. Some particles are further coated with a thin trans-
parent glaze of only 24 m thickness. Line scan analyses
conrmed the particles to be coated with a silicon-rich layer,
probably SiO2. Similar ndings were reached by XPS and
mass spectroscopy.
5.8.3 Mechanism of Particle Deactivation
A loss in performance of thrusters can be caused by cat-
alyst deactivation. The test programme has identied a
failure mode related to the formation of a surface shell of
amorphous silica which acts as a barrier layer in preventing
intimate contact between hydrazine and the iridium
catalyst.
The silicon contaminant is believed to originate from the
AF-E-332 diaphragm elastomer. Hydrophobic silica is

(d) Diaphragm material ndings

The rubber diaphragm from the test (Fig. 5.49) was cut into
samples for optical and scanning electron microscopy.
The macrostructure of the cross-sectioned diaphragm
material is strongly inhomogeneous. Flow lines are clearly
seen, and on one photograph (Fig. 5.50) there appears to be
very slight separation along one flow line situated at the
ridge of the moulded rib.
The diaphragm AF-E-332 contains 17.6 % by weight of
Aerosil R-972, a silane-treated silicon dioxide, 2.9 % zinc
oxide, and 2.9 % calcium oxide. It is inhomogeneous mixing
of these additives to the elastomer (required for strengthen-
ing and wetting purposes) which gives rise to the flow lines.
Analysis of the diaphragm was made in many locations
on the same sample (both hydrazine-exposed and
nitrogen-exposed sides). Quantitative microprobe analysis
results show a slight inhomogeneous distribution of those
elements detected. The metallic elements were present as
0.6 % calcium, 0.9 % zinc, and 7.5 % silicon (average
weight per cent results).

(e) Chemical analysis of the hydrazine

Chemical analysis by the ICP method (Inductive Coupled

Plasma) has been performed on all hydrazine samples taken
from the test. A linear increase in silicon content was Fig. 5.45 Appearance of as-received new catalyst particles showing
observed against storage time, as depicted by Fig. 5.51. a rounded-off corners, and b the highly faceted surface
5.8 Deactivation of Catalyst Particles for Hydrazine Decomposition
b subsequent detailed gures. b SEM photo showing a transition between
Fig. 5.46 Typical grain from catalyst bed of tested thruster. a The
surface was noted in the SEM to be semiconducting; this makes
photography difcult. Superimposed letters relate to the positions of
291
a rough crystalline region and what appears toe an amorphous coating
region having rather a smooth overall nish. Side views indicated
coating to range in thickness between 0.5 and 1.5 m. c White
amorphous layer appears to seal the underlying surface. EDAX surface
analysis identied the presence of silicon and aluminium
probably the only ller additive to be chemically
attacked/leached from the diaphragm, as neither zinc nor
calcium oxides (the remaining llers in the rubber) have
been found on used catalyst grain surfaces.
The silicon concentration measured in the hydrazine
increased from 0.1 to 2.1 ppm during the 12-month test period.
Increases are also likely to occur when similar diaphragms
have been selected for operational spacecraft reaction control
systems. The failure investigation (Dunn and Papenburg
1986) concluded that leaching of silica from the diaphragm
material caused it to be incompatible with hydrazine and this
had caused deactivation of the thruster catalyst.
Alternative diaphragm materials such as ethylene propy-
lene (Sogame et al. 1990; Ballinger and Sims 2003) may be
more compatible with hydrazine, but it should also be pos-
sible to spray a thin, inert layer of silica-free elastomer onto
the AF-E-332 diaphragm surface intended to make contact
with hydrazine.
5.9 Cathode Emitter Degradation
The manufacture of travelling wave tubes for spacecraft
requires that a high reliability is attained in order to satisfy
life requirements of up to 8 years. Life-testing of new
cathode emitters is generally performed by using accelerat-
ing factors such as an increase in operating current density
and higher temperatures of operation.
Fabrication of barium-impregnated cathodes includes the
following steps:
(a) Sintering of pure tungsten powder under pressure to
form a porous pellet.
(b) Filling the tungsten pores with barium calcium alumi-
nate by melting a mixture of 4 mol BaO, 2 mol CaO,
and 1 mol Al2O3 at 2000 K. This compound is allowed
to soak into the porous tungsten under an inert
atmosphere.
During activation of the cathode, its emitting surface is
covered with barium metal, this being the most effective
emitting element (lowest work function within the cathode).
This free barium is produced by the following reaction:
6BaO)3 Al2 O3 W ! Ba3 WO6 6BaO)2 Al2 O3 3Ba
292
Fig. 5.47 Polished cross-section through NEW, AS-RECEIVED
catalyst particles. Note the rounded corners and porous nature of this
material. Many open-mouthed pores provide these pellets with an
extremely large surface area. a Typical NEW catalyst particle showing
locations of detailed micrographs. b Area A. Even at high magnica-
tions there is no evidence of a free iridium surface layer. The pellet has
The free barium diffuses along the pores of the tungsten
to activate the cathode surface. A comparison between
Figs. 5.52 and 5.53 shows the long-life cathode to have
produced an eroded emission surface. The tungsten grains
have ion-etched at various rates dependent upon their crystal
orientation to the electron flow direction. They are often
striated; their grain boundaries have become clearly visible,
and EDAX analysis has shown that the entrapped pores may
become depleted in the impregnated oxide when accelerating
factors of current density or temperature are incorrectly
selected and far greater than those encountered during ser-
vice operation. In fact, when service failures do occur they
usually take place at the beginning of the cathode life owing
to poor workmanship or material defects and not by
wear-out of the barium-emitting compound.
5 Metallography Applied to Spacecraft Test Failures
a dark appearance, and except for some small areas (arrowed) does not
have the bright, reflective surface characteristic of a metal. c Area B.
Showing friable nature of these particles and some bright lamellae
within the matrix. d Area C. Bright lamellae are also present in central
region of particle
The extent of wear-out can be assessed by metallographic
means. The tungsten pellets are removed from the test
assembly and mounted in epoxy resin. As the tungsten pores
contain barium, calcium, and aluminium oxides, no lubricant
can be used as this might otherwise dissolve or dislodge the
oxide particles. A suitable means using a microtome for
making cross-section through these pellets has been found;
the mounted pellets can be sliced in 2 m steps with a very
sharp tungsten-carbide knife edge. Microprobe line-scan
analyses are then made, as shown in Fig. 5.54, commencing
at the cathode emitter surface and following a path normal to
it. Recently introduced sectioning methods, since this
examination was made, might have provided more details
from our cathode investigationthey include both the FIB
and argon ion milling methods described in Sect. 3.2.2.2 and
5.9 Cathode Emitter Degradation 293

Fig. 5.48 Polished cross-section

through a contaminated tested
catalyst particle. a Cross-section
through a contaminated particle
showing locations of further
detailed photomicrographs. Note
that the outer shell appears white
this is assumed to be a
contamination layer. b Area AB.
The outer shell supports a thin
(4 m) glaze that is just visible
between arrows. c Area C.
Further evidence of the thin
surface glaze present on the
white shell of contamination.
d Area D. Photomicrograph taken
in the central region of the
contaminated particle by means
of polarized light. This region has
two phases (light and grey)
indicating two different
compositions

the observations of Kim (2012)and obviate the sample
smearing noticed in Fig. 5.54.
The life-tested porous-tungsten cathode showed no con-
centration gradients for barium and aluminium throughout
its volume. This indicates that free barium is still available
for diffusion, and wear-out is not imminent. Calcium does
show a depletion toward the emitter surface due to its
inability to diffuse, in the oxide form, along the intercon-
necting pores of the pellet.
Investigations have shown that excessively high test
temperatures (above 1100 C) can cause the porous tungsten
pellets to become more compact by the continuation of the
sintering process. This causes a specic mode of failure (e.g.
no emission) owing to the seal-off, by sintering and grain
growth, of the interconnecting network of pores to prevent
the barium activator diffusing from the pellet bulk to the
cathode emitter surface as illustrated in Fig. 5.55.
5.10 Investigation of a Failed Spacecraft
Antenna
Spacecraft continue to make increased use of thermosetting
bre-based composites to reduce the weight and increase the
stiffness of structural components. Carbon bre reinforced
plastics (CFRPs) were probably rst introduced via the
manufacture of antennas to provide for an electrically con-
ductive, highly stable surface prole, even under severe
thermal cycling in vacuum. They are now often used in
satellites for structural components between 4 and 450 K as
reported by Schedler (1988). The design and development of
a 2 m diameter antenna, with a complex convex beam
reflector, provided a challenging task to European designers
and manufacturers. The mechanical design problems asso-
ciated with the launch environment and extreme thermal
environment in orbit necessitated that the antenna reflector
should be constructed from a sandwich of CFRP face-skins
with an aluminium honeycomb core bonded together by
means of lm adhesives. This assembly would satisfy the
critical requirements of low mass, high rigidity, low thermal
deformation over a wide temperature range, and good
reflection of radio frequencies.
A test programme was attempted on a qualication model
antenna (Fig. 5.56) to demonstrate that the design aims had
been achieved. Vibration testing invoked damage to the
antenna, and once material failures became visible on the
surfaces the tests were terminated. A request was made that a
materials evaluation be performed on the failed antenna,
particularly in the regions of debonding and fracturing of the
CFRP. A wide range of nondestructive test methods was
294
Fig. 5.49 Appearance of RIB-SIDE of used hydrazine tank
diaphragm. This side had been in contact with hydrazine. The reverse
side had been in contact only with nitrogen. Material is AF-E-332 and
Fig. 5.50 Scalpel blade cuts through the diaphragm show its
inhomogeneous nature. Flow lines into the rib volume are clearly
seen. Separation along the flow lines was observed (arrowed)
considered for this work. Many are able to detect voids, areas
of porosity and cracking within the adhesive layers, (Munns
1995) but it was the inability of existing nondestructive
5 Metallography Applied to Spacecraft Test Failures
has been used for many years for elastomeric diaphragm and bladder
tanksthe elastomer is ethylene propylene diene modied (EPDM)
polymer (Marker is 1 cm)
methods to detect poor adhesive strength with any reliability
which inhibited this antenna failure investigation. Hologra-
phy was both an industrial technique and a research tool at
the time of the investigation, and provided excellent nonde-
structive information. Metallography complemented other
techniques and could be used to quantify the failure mode.
The more obvious defects were photographed (see
Figs. 5.57, 5.58 and 5.59) and then both sides of the antenna
were subjected to holographic nondestructive test inspection
employing a Fokker-VFW holographic interface bond tester.
This method requires no physical contact between the test
equipment and the antenna surface, the face-skins being
heated gradually with hot air to provide thermal distortion
and local displacements. Laser light (spatially and longitu-
dinally coherent and highly monochromatic) supplied at a
laser gun power of 1 W was reflected off the failed antenna
surfaces under both hot and cold conditions and recorded on
a photographic plate as a hologram. Good holograms and
5.10 Investigation of a Failed Spacecraft Antenna
Fig. 5.51 Plot of hydrazine
contamination levels (inorganic)
as a function of storage time.
Samples were decanted from test
propellant tank held at 40 C
interference patterns were produced with an even distribu-
tion across the individual surface areas inspected (Fig. 5.60).
Debonded areas, not previously apparent from visual
inspection, were carefully marked on the antenna surface.
Similar holographic interferometry is performed on a regular
basis for the NDT inspection of 2 m length cylindrical
spacecraft motor cases constructed from composite materials
(Barbier and Le Floch 1985).
Several special chemical and mechanical tests were per-
formed on the antenna, but probably the most important
deductions were made from the results of metallographic
examinations. Microsections were taken from the various
areas of the antenna to establish the integrity of the CFRP
295
face-sheet lay-ups (Fig. 5.61) and the form and extent of
wetting achieved by the adhesive lm on the aluminium
honeycomb and face-sheets. The samples were cut from
regions noted by holography to be either mechanically sound
or debonded. Areas located adjacent to CFRP delamina-
tions and fractures (Fig. 5.62) or debonds (Fig. 5.63) could
have degraded and were also microsectioned (Dunn and
Collins 1978).
An assessment of the investigation results showed that the
hand laid-up antenna face-sheets were not observed to
contain any major defects. Also, mechanical tests made
uniaxially (mode I), or in pure shear (mode II) on tensile
samples cut from the face-sheets, produced values of
296
Fig. 5.52 Appearance of unused cathode emitter surface
Fig. 5.53 Appearance of worn out cathode pellet. Many thousands
of hours under test at very high current densities has depleted the
barium content of the subsurface
suitably high strength. However, application of low twisting
or buckling forces (e.g. mixed-mode I/II) appeared to easily
delaminate this CFRP material, separation occurring
between lay-ups of differing orientation. The bond strength
between the aluminium honeycomb and face-sheets was
considered inadequate owing to:
5 Metallography Applied to Spacecraft Test Failures
only partial wetting of the adhesive lm to aluminium,
and
unsuitable contact between aluminium and face-sheets
with occasional areas showing aluminium to have hardly
submerged in the adhesive lm.
Sandwich or flatwise tensile tests showed the rear
face-skin to possess the weakest bond to the aluminium
honeycomb.
It was concluded that vibration testing had initially
caused debonding of the face-skins from the honeycomb
structure (due to the poor bonding process). Once debonded,
the skins are likely to have been subjected to buckling forces
which are believed initially to have produced delamination
within the CFRP and later actual fractures (as shown in
Figs. 5.57 and 5.62). This is consistent (Blackman et al.
2012) with mixed mode I/II fractures, occurring when the
skins oscillate under vibration once they have detached from
the aluminium honeycomb, to result in interlaminar fracture
of the composite face-skins.
It was nally recommended that an attempt should be
made to improve the strength of the adhesive bond between
antenna honeycomb and CFRP face-skins, by:
degreasing, then surface treatment of the aluminium
honeycomb with a chemical conversion coating such as
Iridite or Alodine 1200, followed by adequate cleaning,
and either
application of a thicker layer of FM 24-02 adhesive lm,
or
use of a primer and a complete change of adhesive sys-
tem (e.g. Ciba Geigy BSL 312).
All materials and bonding developments led to process
changes that were nally incorporated into the build of a
further qualication model antenna. These changes in mate-
rials and procedures were monitored by standard testing of
in-line samples. In addition, the full quality assurance
surveillance made during all steps of manufacture has con-
tributed to the assembly and launch of a flight-worthy antenna.
5.11 The Wear of Ball Bearings
Spacecraft mechanisms frequently incorporate
high-precision bearings consisting of an inner race, outer
race, and either balls or rollers manufactured of hardened
and tempered chromium steels such as AISI 440C and AISI
E52100. The balls are retained and separated from each
other by means of a metallic or nonmetallic cage. Particular
applications on-board spacecraft will encounter environ-
mental conditions of high vacuum, and may necessitate
5.11 The Wear of Ball Bearings 297

Fig. 5.54 Line-scans taken along path XX on microtome section of life-tested cathode. a Cathode cross-section line represents a depth of 1 mm
from the emitter surface. b Barium line-scan. c Calcium line-scan. d Aluminium line-scan
298 5 Metallography Applied to Spacecraft Test Failures

Fig. 5.55 Cathode emitter degraded by additional sintering of porous b Detail showing reduced number of pores toward emitter surface.
tungsten during very-high-temperature operation (above 1100 C for c Detail of porosity network
25,000 h). a Cathode pellet after polishing of microtomed cross-section.

Fig. 5.56 View of 2 m diameter L-band dish antenna after test. Note reflector cut-out positions (arrowed)
5.11 The Wear of Ball Bearings
Fig. 5.57 Debond and fracture of CFRP on back face, adjacent to mounting pillar
operation at cryogenic temperatures. For these reasons the
common oil and grease lubrication systems used for terres-
trial bearings cannot be used, and either organic (e.g. poly-
tetrafluoroethylene), metallic (e.g. gold and lead thin lms),
and ceramic (e.g. titanium carbide, titanium nitride, and
silicon nitride) coating materials have to be considered. The
choice of space-approved synthetic oils is severely limited
(see Table 5.2), and either temperature limitations or the
close proximity of optical systems may preclude their
selection.
Operation of bearings under vacuum will cause the ball
bearings to make contact with the adjacent races by rolling
and sliding. Conditions such as long operational life, high
loading conditions, or poor workmanship (e.g. wrong
alignment or high surface roughness) will cause the lubri-
cation lm to either wear or become displaced, so that
metal-to-metal contacts occur. This can lead to an increased
bearing friction, the risk of cold welding, and ultimately
seizure of the mechanism. Figure 5.64 illustrates the damage
incurred by cold-welded ball-bearings. Loading conditions
299
had not caused these materials to become stressed beyond
their yield stresses; that is, no plastic deformation occurred
on a macroscale. Local deformation on a microscale had
taken place, however, so that the two metal contact surfaces
could be pushed sufciently close together under vacuum
and at room temperature to create a minute cold pressure
weld. These microwelds rupture almost instantaneously to
produce a roughened surface that becomes progressively
more prone to further microwelding as the peaks on the race
side become compressed during subsequent contact with the
ball.
Modern bearings now support customized thin coatings
such as silicon nitride and silicon carbide on conventional
steel bearing components, and, for European spacecraft,
bearings coated with titanium carbide (TiC) have widespread
applications. TiC coatings are also frequently used to avoid
the seizure or galling of highly loaded contacts in hold-down
and release mechanisms, sliding pivots, and solar array
hinges (Briscoe 1982). The procedure for attaining
space-quality TiC coatings on steel has been detailed by
300 5 Metallography Applied to Spacecraft Test Failures

Fig. 5.58 Typical debond area adjacent to mounting pillar, back face

Fig. 5.59 Cracks in CFRP on back face

5.11 The Wear of Ball Bearings 301

Fig. 5.60 Holographic

interference inspection of failed
antenna. Secondary fringes
denote various areas of CFRP
internal fracture and debonds.
a Underside face-skin; detail of
debond photographed in
Fig. 5.58. Damage extends further
than was expected from the visual
inspection. b Underside face-skin
viewed from edge. Note debonds
on edge skin close to cut-out
(arrowed). c Underside face-skin
showing severely weakened fault
extending between small cut-out
and the three screw-down points
302 5 Metallography Applied to Spacecraft Test Failures

Fig. 5.61 Metallography can be readily applied for the quality b microvoids in resin between toes; c change in bre density. Note A
inspection of the antenna face-skin lay-ups. In this case four toes of toe is a sheet containing millions of parallel-lying carbon bres, held
carbon bre were hand-laid at 90, 180, 180, and 90. Certain together in partially cured resin
discrepancies were seen for the failed antenna: a changes in thickness;
5.11 The Wear of Ball Bearings
Fig. 5.62 Scanning electron
micrographs: Fracture surface
from debond area arrowed in
Fig. 5.57. a (50) shows
delamination to have occurred
between lay-ups. b (500) shows
carbon bres to have broken; they
are supported by resin which
appears to have pulverized.
c (2000) and d (5000) show
detail of bre ends and resin
Boving and Hintermann (1987). In this case, TiC is applied
by chemical-vapour deposition (CVD) onto steels such as
vacuum-remelted 440C. Deposition takes place at a rate of
the order of 1 m/h. Boving and Hintermann have also
demonstrated a much reduced dry friction coefcient for
sliding contacts of TiC-coated steel against steel (0.3 both in
dry air and under vacuum) when compared to steel against
steel (greater than 1, both in dry air and under vacuum). The
TiC CVD coating on 440C steel balls is initially too rough to
be immediately used, and the balls must be high-precision
lap-polished before bearing assembly. A large amount of
data was generated during the post-flight tribological
assessment of the Hubble Space Telescope solar array
mechanisms. The bearings were noted to be in good con-
dition (TiC-coated titanium carbide balls and
outer-race-riding leadbronze cage) but the outer race,
manufactured from the bearing steel 52100, showed some
local corrosion spots (some method of surface protection is
needed for this steel, or the use of 440C material).
Titanium nitride coatings are often used to improve the
friction and wear properties of unlubricated sliding contacts
in terrestrial applications where the presence of ambient
humidity is benecial. At low humidities, or under vacuum,
the TiN coating is associated with increased friction and
wear (Mohrbacher et al. 1995), and this may limit its use-
fulness for space mechanisms.
Ball and roller bearings made of silicon nitride ceramics
have been proven to provide smooth operation even under
extreme conditions. They are identied by NASA as the
303
preferred ball material (better than 440C) to be used with
steel races for cryoliquid pump systems. One reason is that
Si3N4 balls lower heat generation and improve heat transfer
(Ryan et al. 1995). Tests in Europe by E.S.K. (1995) also
revealed that Si3N4 balls can operate at very high speed, with
high load-carrying capacity (even at 800 C) without seizure
or cold welding during dry running.
Metallography is employed as a quality control tool for
the measurement of coating continuity and thickness as well
as a check to ensure a correct microstructure of the substrate
(see, for instance, Fig. 5.65). The TiC-modied surface
avoids cold-welding at solid-to-solid interfaces between
balls and races in the absence of oils or greases and under
extreme environments such as vacuum and either elevated or
cryogenic temperatures. The coating has a slight residual
compressive stress on its substrate (Chollet et al. 1985) but
does not crack or spall even when exposed to the stresses
imposed by thermal cycling between 269 and 30 C.
As a nal note, it may be important to reconsider the
choice of cleaning solvent. An Alert which circulated the
space industry in 1996 provided mechanical engineers with
the following information:
Bearings and other lubricated components used in long-duration
spacecraft have traditionally used ozone-depleting chemicals
(ODCs) such as freon as a nal cleaning process prior to
lubrication. As these chemicals are being phased out, new
cleaning processes have taken their place such as supercritical
carbon dioxide, perfluoromalocarbons, aqueous detergents, etc.
These new processes are effective cleaners, but leave the bearing
surfaces with a different residual surface chemistry. Some of
304 5 Metallography Applied to Spacecraft Test Failures

b Fig. 5.63 Sample carefully machined from defective areas of the

antenna (as located in Fig. 5.60) were submitted to metallography.
Some inherent defects are observed from the microsections: a Good
appearance of adhesive ller which bonds aluminium honeycomb to
face-skin. b Aluminium was not in contact with face-skin and had not
been adequately wetted by adhesive (inadequate strength caused failure
of this bond during testing). c Uneven, low-strength bond

these surface chemistries have shown enhanced reactivity with

the applied lubricants (particularly with perfluoropolyethers) in
static and dynamic tests. This reactivity can lead to decreased
bearing lifetimes and since these represent single-point failures,
reduced spacecraft lifetimes.

Additional information concerning replacement cleaning

solvents is given in Sect. 4.21.1, and the recent overview by
Patterson and Mykytiuk (1996) provides important data as to
how US companies have tackled this subject.

5.12 Cold Welding of Mechanisms

5.12.1 General

Methods of resisting adhesive wear and cold welding have

been discussed in Sect. 2.3 and illustrated by means of a
chart showing which metals have the greatest chance to
cold-weld (Fig. 2.3) with an example of a cold welded large
45 mm diameter stainless steel nut and bolt being shown in

Fig. 5.64 An example of a failed ball-bearing after operation under

vacuum. Inadequate lubrication had been applied to the uncoated 440C
steel ball
5.12 Cold Welding of Mechanisms
Fig. 2.4. The wear mechanisms of fasteners and ball bearings
were also covered, respectively, in Sects. 5.2 and 5.11.
Great care must be taken during the selection of materials
and their coatings in order to prevent wear and cold welding,
or adhesion, between piece-parts such as gears, solar array
hinges, bearings, electrical commutators, and the like, which
operate under vacuum. As mentioned in previous sections,
the tribological behaviour under space environmental con-
ditions differs signicantly from terrestrial conditions. Oxide
and nitride layers which are formed in the terrestrial envi-
ronment on pure metal surfaces, and can be regarded as
natural protection layers, are not formed in space. During the
service life of spacecraft mechanisms these surfaces can
outgas, wear, fret, and become generally cleaner and this
will increase the probability of cold welding. Also the inti-
mate contact of nonmoving parts (e.g. fasteners) may lead to
cold welding in the space environment and this may hinder
the disassembly of orbiting space hardware by astronauts.
This necessitates the use of suitable materials, coatings,
and/or lubricants, many of which can now be assessed by the
test method prescribed by Merstallinger et al. (2009). In the
case of the Hubble Space Telescope solar array refurbish-
ment mission, the arrays could be wound up by hand and
removed from the spacecraft by means of turning
silver-plated steel bolts.
In this Section, an overview of space tribology will focus
on mechanisms that operate under conditions where surfaces
may have a tendency to cold weld under sliding wear, under
impact forces, by surface fretting, and under long-term
loading. It may be pertinent, when considering surface n-
ishes for mechanisms, to point out that all mechanisms need
to be electrically bonded (for grounding purposes) to the
Fig. 5.65 Micrograph of a section through a space-quality TiC-coated
440C steel ball after lapping and polishing. Maximum surface
roughness is 0.012 m CLA. Vickers microhardness of coating is
35,000 N/mm2. Microstructure is revealed by etching in Marbles
solution (500) (courtesy of the Swiss Center for Electronics and
Microtechnology)
305
spacecraft structure or its sub-system. The specic DC
resistance between the mechanism and the grounding plane
will depend on whether the electronic components are
mounted within the mechanism i.e. less than 5 m, or if
there are no electronic components, only protection against
electrostatic discharge is needed 0.1 (see
ECSS-E-ST-33-01).
Probably the most demanding conditions of sliding con-
tacts are those associated with the Infrared Space Observa-
tory (ISO), which was launched in 1995 and is operating
successfully. The cryostat cover was ejected 10 days after
launch and all scientic instruments have since been tested
and have been making observations of various objects in
space. One instrument utilizes focusing and shutter mecha-
nisms that have to operate at 4 K. Another incorporates a
lter change mechanism with gear drive and shaft-mounted
ball bearings; these also operated at cryogenic temperatures.
Here, lubrication is by means of a molybdenum disulphide
lm and the gears are made from titanium alloy with titanium
carbide surfaces. Another area of challenge is that associated
with the Space Station Alpha and the Polar Platform projects;
these spacecraft will have the largest arrays yet built in
Europe, with nominal power capacities of around 15 kW. The
space station arrays will be extended and retracted at
six-monthly intervals for a lifetime of 30 years. The Polar
Platform arrays will be Sun-following, with a rotation of 6
revs per day. These arrays have led to the design of rotating
power transfer assemblies with large-diameter titanium car-
bide bearings and high-power slip rings.
Several new developments are taking place in the area of
tribological surfaces. Filled and reinforced polymer com-
posites are nding space applications, a polyimide (trade
name Vespol) was selected for the Giotto mission cameras
scanning mirror system gear wheels. Even slightly reactive
materials, such as PTFE and polydimethyl siloxane, can be
bonded to strong metallic substrates using surface etching
and reduce surface friction with a nonstick nish. Ceramics
have received extensive exposure over the last few years and
specially engineered surface treatments are available for
most metals (Eyre 1991). Vapour phase deposition, ion
implantation, and laser treatments are being researched.
Whereas the main bearing steels are liable to corrode, a new
anti-wear treatment can now be applied to the softer auste-
nitic stainless steels such as AISI 304 and 316. The nish
has a surface hardness of up to 1200 Hv and a gradual
decrease in hardness from the surface to the metal core
eliminates the risk of delamination (van der Jagt et al. 1991).
There are many well established methods for determin-
ing wear and friction. The most frequently used in the
space industry simulates the behaviour of sliding wear and
is tested under vacuum conditions by the pin-on-
rotating-disc method (especially DIN Std. 50324). Other
methods are the pin-on-ring and pin-on-plate. The test
306
parameters should mirror those seen in the space mecha-
nism application and include: contact pressure, sliding
speeds, loads due to thermal distortion and preload varia-
tions plus application loads, temperatures, etc. These have
been detailed by Labruyre and Urmston (1995). The tri-
bological components, contacting surfaces, and lubricants
utilized by the mechanisms of the Hubble Space Telescope
solar array were microscopically inspected and reported on
after their return to Earth. The ndings demonstrated that
the present approach to the selection and testing of con-
tacting materials is adequate. There was generally no evi-
dence that, after approximately 3.5 years of spaceflight
operation, modications to materials or lubrication systems
would have been benecial to this particular spacecraft
(Anderson et al. 1995).
5.12.2 Cold Welding Due to Cyclic, Impact
Loading
One of the early Meteosats suffered problems when its
black-body calibration mechanism became blocked after
several years of successful operation. The mechanism would
remain blocked for 7 h during operation in orbit in what was
called the bracket-to-end-stop position. As one of several
failure modes (e.g. relaxation of returning springs, problem
with magnet) cold-welding between bracket and end stop
was attributed as being the main cause of failure. An engi-
neering model of the calibration mechanism, having the
same build, was ground tested under vacuum and after
repeated impacts between the bracket and its end stops the
onset of cold-welding was observed after several thousand
operations. This mechanism is shown in Fig. 5.66a. Both the
end stop and the bracket had been constructed from the
magnesium alloy GA3Z1 (AZ31) having the composition
3 %Al, l %Zn, 0.3 %Mn, balance Mg. All surfaces had been
nished with a black chemical conversion coating (black
mordancage). The on-ground laboratory tests conrmed
that the repeated impact forces gradually wore down the
black coating, until it had been completely removed at the
small points of impact. Microscope inspections revealed that
once the coating had been removed there was sticking
between the mating surfaces. Mechanically deformed parti-
cles were noted to have been pulled from one surface and
redeposited during the impacting cycles, as seen in
Fig. 5.56b. This conrmed the cold-welding theory and
initiated the investigation of a test method suitable for the
evaluation of materials and surfaces that were expected to be
subjected to wear and cold-welding. In parallel, many impact
tests were made using the calibration mechanism shown in
Fig. 5.66, but with modications so that different small
metallic inserts could be placed at the bracket impact posi-
tion and on the end-stop position. A suitable pair of contacts
5 Metallography Applied to Spacecraft Test Failures
made of Densimet (D175K) for the body side, and stainless
steel (Inox Z6CNU17 04) for the mobile bracket, was found
to resist cold-welding and this combination of materials was
incorporated into later designs of the black-body mechanism
(Fig. 5.67).
A test equipment for the evaluation of cold-welding under
cyclic loading and high vacuum was designed and produced
by the former Austrian Research Centre, Seibersdorf. This
equipment is sited at the Aerospace and Advanced Com-
posites (AAC) GmbH site in Wiener-Neustadt, Austria. It is
capable of simulating a wide range of contact parameters
including static and impact loading. The equipment has been
upgraded to include the simulation of fretting wear and
resultant cold-welding under vacuum. A standard method
has been developed after performing many cold-welding
trials between space mechanism materials the details of
which are reported by Merstallinger et al. (2009). This test
equipment was also shown to successfully reproduce the
sequence of events that led to the cold-welding failure of the
calibration mechanism described above (Merstallinger et al.
1995a, b; Merstallinger and Semerad 1996). Material was
cut from the engineering model black-body mechanism and
submitted to impact adhesion testing. The impact force was
seen to effectively break down the black coating on the
magnesium alloyeven with a low impact energy of
0.004 Nm. The highest cold-welding adhesive forces
attained during this test were: bare magnesium alloy
(GA3Z1)-to-black-coated alloy (60 mN); after breakdown of
coating Mg-to-Mg (280 mN); and, Densimet-to-Inox
(100 mN). The test conditions for cold-welding were quite
similar to the actual in-orbit failure which occurred after
8230 activations with an impact energy of 0.0112 Nm, 30 g
load, that is with a 300 mN force required to unblock the
mechanism. A repeat laboratory test was conducted with a
similar impact energy of 0.02 Nm; this resulted in adhesion
forces of up to 280 mN between metallic contacts of GA3Z1
being measured. This gives some condence that the new
test equipment can actually simulate working mechanisms
making impacts under vacuum.
Impact testing has also been performed on a modied rig
by CSEM in Switzerland. It was found that all tested mate-
rials were subject to some form of cold-welding and that
coatings of molybdenum disulphide or titanium carbide can
reduce this phenomenon (Maillat et al. 1995). Both the
Merstallinger et al. and the Maillat et al. reports indicate that
no similar studies are available in the literature. Both con-
clude that the cold-welding (adhesion force) measured as a
function of number of impacts can be divided into three
phases. Initially there is no, or negligible, adhesion during the
rst few thousand impact cycles; then there is gradual
increase in adhesion force followed by a nal phase where
the adhesion force reaches a constant value. It is postulated
that the constant value is related to some kind of self-healing
5.12 Cold Welding of Mechanisms
Fig. 5.66 a Engineering model of the black-body calibration mech-
anism with BeCu springs S and magnet M. C shows the mechanism
bracket in one end stop position where the blocking occurred.
b Scanning electron microscope view of flattened particles and some
evidence of cratering. The particle arrowed was analysed by EDAX
and seen to be composed of the same elements found in the magnesium
alloy (GA3Z1) together with traces of potassium and chromium from
the worn black coating (600)
of the surfaces oxide layers due to diffusion of interstitial
gases from within the metal to its surface, or from the residual
gases within the vacuum chamber (held at 107 torr).
More recent results have shown that theoretical predic-
tions are by no means comparable with experimental data
307
Fig. 5.67 Simplied schematic of the impact/fretting test equipment.
If fretting is selected, it can only be applied during closed contact and
stopped before unloading. Lateral and vertical forces (load and
adhesion) are measured with the 3-axis force transducer
(Merstallinger and Dunn 2009). The main reason for this is
that the adhesion force is driven by the real contact area
which cannot be predicted (Hertzian theory would predict a
nominal contact area, neglecting surface roughness and
surface contaminationthe latter is unpredictable but is
actually the main contributor). A wide range of material
combinations, some in contact with themselves has been
included in Tables (particularly the Cold weld database
summary tables for fretting and impact contained in
Annex C of the Merstallinger and Dunn report). This work is
continuing, but includes couples: Al-2219, Al-7075, AISI
52100, SS17-7PH, AISI 316L, 440C, Ti6Al4V, Densimet,
Inconel 718 and non-metal Vespel SP3. The metals were
usually coated with space-typical nishes such as Keronite,
hard anodisation, MoS2, Braycote, Balinite and CrN. Those
Tables indicate whether there is no, or negligible adhesion
(where the adhesive force is less than 200 mN) and such
couples might be considered acceptable for long-duration
space use. Other results suggest additional testing is needed,
and two further categories recommend that such couples
should not be used as cold welding is a distinct possibility.
5.12.3 Cold-Welding Due to Fretting
This is the form of cold-welding, or galling, occurring on the
contacting surfaces of moderately-to-heavily loaded metals
and alloys which are subject to slight relative movement.
Fretting taking place in mechanisms and between structural
parts when there is a small amount of play between adja-
cent surfaces. Under atmospheric conditions this gives rise
to galling and the tearing off of tiny metallic particles which
soon oxidize and then may worsen the problem as the oxi-
dized particles can be very abrasive (e.g. aluminium oxide
and tin oxide). Under vacuum conditions the galling will still
occur, but now in the absence of oxygen, the released
308
particles can themselves become cold-welded to one or both
of the adjacent surfaces.
A literature review of fretting associated with space
conditions was made by Vine and Price (1992), clearly little
empirical data are available concerning the fretting and wear
rates of spacecraft materials at different temperatures in
vacuum. Cold-welding due to fretting during ground trans-
portation of the US Galileo spacecraft may have caused
damage and later a malfunction of its main antenna
deployment systemthis became only partially unfurled at a
critical stage during this crafts journey to Jupiter. Damage
by fretting has been observed during the ground testing of a
latching device, similar to a Marman-clamp, on the Cassini
projects plasma spectrometer. Vibration tests caused the
hard anodized nish on AA 6061 to break off when coupled
to Alodine-treated AA 6061. In a repeat test fretting again
occurred when the anodized AA 6061 was coupled to
stainless steel 316, causing damage to both steel and the
anodized layer. Following extensive testing a suitable com-
bination of surfaces was found: 60 m thick hard anodized
AA 2219 coupled to a latch clamp ring made of PEEK
(plastic) coated AA 2219 (Ylikorpi et al. 1995). Sticking
resulting from fretting was also considered and tested for the
holddown and release unit of the Envisat solar arrayvi-
bration tests conrmed the acceptance of anodized alu-
minium running against nickel-plated aluminium
(Zwanenburg 1995).
The European spacecraft Hipparcos Baffle Cover
Mechanism failed a ground functional test during open
close operations. This was found to result from an unac-
ceptably high increase in friction between a titanium alloy
bush and a stainless steel spring. These parts were supposed
to slide over each other, but during vibration testing fretting
had occurred. This caused some small surface areas to make
contact and become cold-welded. Figure 5.68a shows where
sticking had taken place, and Fig. 5.68bd highlights in
greater detail the area where fretting and cold-welding had
taken place. Tests showed the 177 PH steel to have a
microhardness of 736 Hv, and the Ti6A14V bush a micro-
hardness of 385 Hv. The titanium alloy bush was modied,
and instead of an anodized nish, the outer surface was
coated with chemical-vapour-deposited titanium carbide.
This solution was found to avoid further fretting and the
mechanism performed perfectly, both during qualication
testing and later, when the spacecrafts telescope baffles
were deployed in orbit.
Titanium and its alloys are used extensively in structural
application and spacecraft mechanisms. Mating surfaces are
now known to easily cold weld because either the natural, or
anodic blue oxide nishes are too thin and have poor
adhesion under mechanical wear, particularly under vacuum.
These nishes have good corrosion resistance but, for sepa-
rable mating contact surfaces, it may be essential to build-up a
5 Metallography Applied to Spacecraft Test Failures
thicker specialized coating to ensure there will be no cold
welding under vacuum. Such coatings were evaluated by
means of the AAC test method depicted in Fig. 5.67 and
described in the previous Section. Preliminary results of a test
campaign aiming to nd suitable nishes for titanium and its
alloys having a high resistance to cold welding under condi-
tions of impact and fretting are described in the captions to
Fig. 5.69ad. Trade name coating processes based on ceramic
conversion treatment (plasma electrolytic oxidation treatment)
of titanium include: Apicoate, Balinite, Dicronite and Kero-
nite 3000. For further reading the following references are
recommended: Li and Dong (2010), Jiang and Wang (2010),
Li and Zhang (2014) and Merstallinger and Dunn (2009).
5.13 Defective Black-Anodized Electrical
Connector
Spacecraft electrical connectors consist of the metallic shell
(endbell, coupling nut, and barrel), the insulator and grom-
met materials (possessing low outgassing properties), and
the connector pin and socket contacts which are usually
machined from brass, have a copper and nickel strike, and
are nished with at least 2 m of electrolytic gold plate. The
shell parts are made of aluminium alloys or 300-series
stainless steels. Aluminium parts are either anodized or
nickel-plated, whereas the steel parts may be passivated or
gold-plated. The plating or anodization thicknesses are not
always specied, but connectors must pass the salt spray
corrosion tests according to ML-STD-202, method 101, test
condition B. It should be noted that the usual commercial or
military connectors supporting chromate coating over
cadmium-plating cannot be used under high vacuum owing
to the problem of sublimation as discussed in Sect. 5.6.
The connector shown in Fig. 5.70 possesses a
black-anodized nish. The aluminium alloy is 2024, and a
0.2 mm translucent oxide has been built up in a
low-temperature sulphuric acid electrolyte to produce a hard,
wear-resistant lm. During anodizing the oxide lm consists
of two layers: a thin inner barrier layer which is composed of
compact Al2O3, and a thick 1200 m outer porous hydrated
oxide lm (Gabe 1978). The anodized lm had been dyed
black by a nickel sulphide process (ECSS-ST-70-03) and
nally sealed in boiling deionized water.
The connector seen in Fig. 5.70 shows the effect of
thermal cycling followed by a salt spray corrosion test. The
body is covered in supercial white corrosion products, and
metallography reveals microcracking in the anodized layer.
A close inspection of the cross-sectioned layer reveals a
nonuniform lm growth due to the presence of copper-rich
inclusions distributed throughout the aluminium. The thick
oxide lm is very hard and insufciently compliant to sur-
vive the thermal cycling test procedure. As a result of
5.13 Defective Black-Anodized Electrical Connector 309

Fig. 5.68 a Part of a Baffle

Cover Mechanism following
vibration testing and disassembly.
The arrow highlights a wear
track. Fretting has completely
removed a blue anodized
protection layer from the
Ti6A14V bush. This surface
became stuck to the
corresponding internal surface of
the coil spring. Optical
photographs of the contact
surfaces where fretting occurred.
b Shows the removed passivated
layer on the 17-7 PH steel spring.
c Shows the worn surface on the
anodized Ti6A14V bush (both at
60 magnication). Material from
the steel coil spring has
transferred to the titanium alloy
surface as shown in d. This
cold-welded particle has the same
composition as the steel (from
EDAX analysis); it is securely
bonded to the titanium bush

anodizing trials suitable parameters were established to
produce a thinner, 35 m oxide lm that did not craze during
100 thermal cycles between 100 and +100 C. Salt spray
testing produced no evidence of corrosion products, and the
connector was subsequently space-qualied.
5.14 Contaminant ParticlesIdentification
of Their Sources
A number of problems associated with contaminant particles
have been described and illustrated in this book. Until now,
the particles have been identied in situ, and it is clear that
they originated from processes such as welding (Fig. 3.6),
sliding wear (Fig. 4.40), laceration of screw threads
(Figs. 5.13 and 5.14), from inserts that had been drilled and
threaded but inadequately cleaned (Fig. 5.15), or particles that
spalled from the surface of work-hardened beryllium during
thermal cycling (Fig. 5.41). These particles were revealed
during inspection and test and, the deleterious consequences
of their presence within space hardware have been discussed.
In this section, several examples will be briefly described
where costly spacecraft subsystem failures have been caused
by metallic particles originating from locations far removed
from the site of failure. Some degree of detective work was
associated with these failure analyses, but probably the most
useful information to supplement the detailed inspection and
microscopy work was gained from the various projects
Declared Materials List. The Declared Materials List
(DML) is a compulsory compilation describing each mate-
rial selected for use on-board the spacecraft. The list is
produced by every contractor producing flight hardware and
is used as a formal document for the approval of the material
and its intended application. The list is formatted so that
each material is identied by commercial name, by chemical
composition, by manufacturers specication, by processing
parameters, and its precise use and location must be
described. As many hundreds of metallic alloys are used on
310
Fig. 5.69 Preliminary results of a test campaign aiming to nd suitable
nishes for titanium and its alloys having a high resistance to cold
welding under conditions of impact and fretting. Balinite A (TiN),
Balinite B (TiCN), Dicronite + (WS2 + CrC), Keronite and Ni-PTFE
coatings appear to be good contenders. It is important that Process
Identication Documents are generated for each coating process to
ensure all parameters are maintained under strict control. Once
approved, there can be no improvements made to the process unless
the new ideas are re-assessed by testing. The results shown in a can be
traced to the coating morphologies illustrated in bd for each
designated coating (courtesy of AAC, Austria). a Experimental cam-
paign showing results of maximum adhesion force recorded following
standardised impact testing. Note that for impact testing, three contact
pressure values are used. The rst 10,000 impact cycles are made at
40 % of the substrates elastic limit, i.e. 40 %EL in this gure. Then,
the next 5000 impact cycles at 60 % EL, and nally 5000 impact cycles
each spacecraft it is possible, by means of performing some
chemical analysis, metallography, hardness testing, and
consulting the DML, to actually pinpoint the precise location
from where particles may have originated.
The examples illustrated in Figs. 5.71, 5.72, 5.73, 5.74 and
5.75 show how particles have caused particular failures, the
morphology of the particles, and how their origin was determined.
5 Metallography Applied to Spacecraft Test Failures
at 100 % EL. Under these specic test conditions, the following
coatings have shown very little adhesion: Balinite A (TiN), Balinite B
(TiCN), Dicronite + (WS2 + CrC), Keronite and Ni-PTFE. Fretting
results are not shown. It is recommended that coatings produced by any
manufacturer should be tested if cold welding under vacuum is
considered an important spacecraft parameter. (from Merstallinger and
Dunn 2009). b FIB milled image through top surface of 6 m thick
Keronite coating showing its columnar structure beneath a 200 nm
thick fused oxide. EDX analyses only revealed Ti, Al and V as can be
expected of this PEO coating. c FIB milled image of Balinite B coating,
top is surface with micro-indents, middle is 2.9 m coating thickness
and lower area is substrate. EDX analyses revealed Ti, Al with some C
and N. d FIB milled image of Dicronite coating. The double layer
thickness of (WS2 + CCr) is about 10 m
5.15 Silicone Contamination
5.15.1 General
Silicone products are used extensively during the manufac-
ture of spacecraft. More than 50 are identied on some
projects Declared Materials Lists. All products selected for
5.15 Silicone Contamination
Fig. 5.70 Evaluation of launch vehicle connector. a General view of
black-anodized nish on cable connector exhibiting corrosion products.
b Thermal cracking of anodized lm as shown in microsection. c The
presence of inclusions masks the anodization growth front
flight hardware must be carefully screened; for applications
in vacuum certain outgassing tests will be necessary. For
manned spacecraft these silicone products will be assessed
for off-gassing, toxicity, and flammability.
Silicone oils and greases that have been puried by
vacuum-distillation are used as lubricants (see Table 5.2).
Some are chosen for their thermal conductivity properties.
The surface tension properties of silicone liquids are very
low, about 1525 dynes/cm. This is between those for pet-
roleum and ether, so one can expect these products to readily
flow into conned spaces and to creep over clean surfaces.
Some silicone products become polymerized in air. Others
are two-part silicones that must be mixed and cured. Solid
silicones are used as sealing materials, adhesives, paints, and
as conformal coatings, such as those applied to PCB
assemblies. These materials have many useful characteris-
tics, such as good dielectric properties, chemically rather
inert, a good ductility over a wide range of cryogenic and
elevated temperatures, and they are hydrophobic.
311
Unfortunately, several of the uniquely useful character-
istics of silicones can cause them to be major sources of
contamination if they are incorrectly handled. Many indus-
trial processes can be severely affected by the presence of
minute amounts of silicones, and the space industry has
suffered several failures owing to this form of contamination.
This had led several European companies (mainly in Scan-
dinavia) to prohibit all silicone products from their sites.
When they are used, it is essential that operators are trained
in the handling of silicones by either the supplier or the
companys own health and safety personnel. The relative
risks of contamination are usually made available on the
material suppliers data sheet.
The writer is aware of several costly problems associated
with the problem of silicone contamination in the space
industry. Three will be described more fully in the following
sections. Finally, some recommendations will be offered as
to how such contaminants may be removed. Briefly, some
other space hardware problems have included:
extensive contamination on the floor of a space labora-
tory due to inadvertent spillage of silicone lubricating oil
solderability problems with both PCBs and electronic
components due to silicone handborne contamination
originating from cosmetics and skin moisturising creams
used by operators
solder repair problems due to the presence of silicone
residues on the circuit board
the degradation, then nal electrical open-circuit failure,
of spacecraft slip rings operating in a vacuum chamber
caused by the condensation of silicone vapour from an
incorrectly cured silicone sealant (this is a well docu-
mented mode of failureit has been shown that an
atmosphere containing only 7 ppm silicone vapour can
cause the electrical failure of micro-relays and other
commutators during controlled contamination tests
(Tamai and Aramata 1993; Tamai 1995).
5.15.2 Contamination of Black-Anodized Finish
The surface of an optical detector housing, specially nished
with a space-approved black-anodized layer in order to
absorb stray light, became contaminated with silicone oil
thought to have crept from a baffle hinge. Attempts to clean
the surface of this instrument were not successful. Despite
submitting small dismantled parts to a number of cleaning
solvents, the surface appeared to regenerate a sheen caused
by the weeping of oil from micropores in the anodized layer.
The problem is clearly illustrated in Fig. 5.76a and described
in its caption. The risk of causing a further contamination to
312
Fig. 5.71 Case 1. Difculties were experienced in the qualication of
a liquid helium cryostat valve. Nonuniform leak rates occurred owing
to the presence of metallic contamination particles that had become
embedded on the Vespal valve seat. The valve was disassembled and a
section cut at the valve port (material AISI 304 stainless steel)-to-outlet
pipe weldment (AISI 316 stainless steel) using a diamond cutting tool.
Many particles were seen entrapped within a small gap between the
pipe and valve port, close to the weld. Chemical analyses were made by
EDAX of the particles: those embedded on the valve seat had the same
composition as those trapped in the gap; they are seen in ac and are
the optical detector, after integration into the housing, was
considered unacceptable. It was thought likely that silicone
from the pore sites could be redeposited onto a glass lens
during the thermal vacuum testing that was planned to
qualify this instrument. Corrective action was to manufac-
ture a new housing and exclude silicone lubricants from the
instrument.
5.15.3 Contamination of Invar Moulding Tool
Spacecraft antennae are usually of the reflector type, such
as a paraboloid illuminated by a horn. Examples of
5 Metallography Applied to Spacecraft Test Failures
composed of iron, chromium, and nickel, together with a binder having
a high chlorine content with calcium, potassium, silicon, and sulphur.
d Show the length of the re-entrant path which, despite extensive fluid
cleaning of the internal surfaces, became a source for particle release
during the flow of liquid helium. The binder and particles originate
from the cutting pastes used to dress the AISI 304 valve port prior to
welding. The pipe did not release particles, as AISI 316 contains
molybdenum and this element was not present in the particles.
Corrective actions included a revised cleaning procedure before
electron beam welding
manufacturing routes for an antenna and horn are associated
with, respectively, Figs. 4.514.53, 5.56, 5.57, 5.58, 5.59,
5.60, 5.61, 5.62 and 5.63. Where the service region on the
Earth is irregular in shape, the antennas reflector surface
may be formed to appropriately cover this shape. The
manufacturing of the carbon bre/epoxy reflector skin,
suitably bonded to its aluminium honeycomb (described in
Sect. 5.10), involves moulding this composite multilayer on
tools that are then placed inside an autoclave. The moulding
tools for antennae must provide temperature resistance, a
very low coefcient of expansion (CTE) to match the
composite, vacuum integrity, thermal conductivity, and
machinability.
5.15 Silicone Contamination 313

Fig. 5.72 Case 2. Fine, gold dust-like particles were noticed within a
plastic bag used to store connector bodies. Short-circuits could easily
ensue if similar contamination should exist in the vicinity of spacecraft
electronic circuits. The dust was traced and found to originate from the
xation tabs used to mechanically join the half-shells of these
50-contact connector bodies, as shown in (a). A more detailed SEM
inspection of the bent tabs in (b) revealed that plated layers were
spalling from the brass substrate. Metallography c showed the brass to
have been nished with a thin flash of copper (0.5 m) followed by
2 m of gold. Also, each tab was seen to have been stretched beyond
the tensile strength of the plating, causing the coppergold to flake off
at the interface between the brass and copper flash. The defective shells
were noted to have plating thicknesses far below those specied.
Thicker plated layers on the tab bends from a different batch of
connector shells did not crack and spall

Fig. 5.73 Case 3. Contamination was noticed on the surface of a

workstation situated inside a clean-room, where electronic component
assembly work was taking place. A selection of particles was taken for
laboratory analysis. a and b show the morphology of the particles. They
are composed of pure cadmium and, being electrically conductive,
could cause short-circuits if transferred to printed circuit assemblies.
The particles were found to originate from the cadmium-plated surface
of a steel housing designed to contain a plastic wall socket for mains
power supply to the clean bench. During installation the housing had
been mechanically damaged. Scoring had released these cadmium
slivers. Proper clean-room practice had not been followed
314
Fig. 5.74 Case 4. A launch vehicle release gear mechanism (hydraulic
jack) was dismantled following its poor performance. Inspection of the
parts showed the presence of a metallic fragment resting on the internal
surface of the hydraulic oil inlet orice. The fragment was passed to the
metallurgical laboratory for examination. The fractured piece was
measured and found to be 17 mm in the longest direction. A photograph
of an important fracture surface is seen in a. This shows the failed part to
have an intergranular fracture similar to stress corrosion, b highlights
how the fracture had initiated at the root of a heavy machine mark. The
piece was carefully cut in two, one half for metallography, the other for
chemical analysis by atomic absorption spectrophotometry. The frag-
ment was composed of aluminium together with 4.5 % copper and 0.7 %
lead. The presence of lead is very unusual in aluminium alloys and this
was conclusive evidence that pinpointed the fragment to have originated
from an internal part that had been machined from the free-machining
aluminium alloy AA 2011 (nominal composition being 56 %Cu, 0.2
0.6 %Pb, rest Al). This alloys is known to be susceptible to stress
corrosion cracking (photographs courtesy of Mr P.W. Martin, MMS-UK)
The nickeliron alloy, Invar, plays a crucial role as a
moulding tool material. This alloy was discovered in 1896. It
has the composition 36 % nickel, remainder iron. Invar
possesses the lowest CTE of all engineering metal alloys in
the range from room temperature to 230 C (1.3 ppm per
5 Metallography Applied to Spacecraft Test Failures
C); this is very similar to the CTE of antenna reflector skins
and is most suitable for this moulding operation.
Unfortunately, one of the most difcult stages in building
an antenna is detaching the reflector skin from the Invar
mould after the thermal-vacuum curing treatment carried out
in the autoclave. Sticking and delamination of the carbon
bre/epoxy toes can damage the skin. Proprietary parting
compounds are used to dress the mould surface and prevent
stickingthese are usually successful. However, the for-
mulation of this substance (usually applied by spraying) is
not well understood.
It appears as though one batch of parting compound was
manufactured so as to contain a silicone compound. This
may have been introduced by the manufacturer of the
compound, or it may have been introduced by an engineer as
a good idea. One disadvantage of manufacturers trade
names, or product numbers, is that there may be little control
over the products chemical composition. In the space
manufacturing industry it is essential to request that a cer-
ticate of conformance is delivered together with the pur-
chased material, and this must contain a reference to
chemical composition. The unusual parting compound cer-
tainly caused no difculty when one set of antennae was
being built; they could be easily removed from the mould
face. However, it was impossible to nd a painting process
that resulted in an adherent painted nish. Tests made on the
painted surfaces of specimens (utilizing the cut-out material
from positions similar to those arrowed on Fig. 5.56) proved
that silicone compounds were present on the reflector skin.
They had caused the paint to blister and this resulted in the
rejection of several reflectors.
5.15.4 Removal of Silicone Polymers
It is generally difcult to completely remove liquid silicones
from metallic surfaces.
In the case of uncured liquid it is advisable to begin
cleaning before any polymerization takes place. This takes
place quickly when the liquid is in contact with air. White
spirit is quite effective; it has a high flash point and is not
very volatile. Experts advise that wiping with a cloth soaked
in isopropyl alcohol (IPA) gives the best results. As IPA is
flammable, benzyl alcohol is an alternative but is slightly
toxic.
Fluorosilicones, being a separate family of materials,
require the use of polar solvents. Here, methyl isobutyl
ketone or methyl ethyl ketone is recommended.
Methylene chloride may be suitable to remove cured low
molecular silicones.
5.15 Silicone Contamination
Fig. 5.75 Case 5. Stainless steel corrosion products in the form of
particle growths and curly bres were found on the internal surface of a
spacecraft propulsion systems nonreturn valve. The growths were red
brown in colour and the bres, as seen in a and arrowed on b, have a
diameter of 510 m and a length of 500 m. They were analysed and
found to be composed of iron, oxygen, and chlorine. The bellow
material is AM350, a semi-austenitic, precipitation-hardening stainless
steel. The same corrosion products were later found on other valves
and, more worrying, these products were present on the PTFE valve
poppet seals. As a result of a long investigation it was concluded that
the corrosion products were caused by the exposure of AM350 to the
vapours of hydrochloric acid. This acid was released from some lengths
Other, cured solid silicones may be mechanically
removed by peeling [this is possible for some conformal
coatings (Dunn and Desplat 1994)].
Two special removal agents are also available:
DIGESIL from C&M Research Ltd., Eneld, UK (or
from Merck in other countries), and
315
of polyvinyl chloride (PVC) tape that had been inadvertently left
attached to the bellows as they were stress-relieved in an oven at
approximately 180 C. PVC starts to decompose at about 80 C, and
consequently evolves hydrogen chloride gas (HC1). This scenario was
repeated under laboratory conditions. PVC tape attached to polished
AM350 was heat-treated in an oven at 180 C. The resulting
laboratory-grown bres are shown in c and have the same morphology
as the contaminated flight parts. It is postulated that only the small
volumes of delta ferrite, present in the microstructure of AM350, are
corrosively attacked and produce this form of corrosion product
(courtesy of Mr. P.W. Martin, MMS-UK) (Color gure online)
SILGEST SD0001 from Dunlop Adhesives, Birming-
ham, UK.
Both agents are extremely hazardous and require great
care in their use. They are skin irritants, the products being
flammable and corrosive.
316 5 Metallography Applied to Spacecraft Test Failures
 5.15 Silicone Contamination
b Fig. 5.76 a Topography of black-anodized aluminium housing as seen
by SEM inspection (backscattered electron image). Superimposed is an
EDAX spectrum clearly identifying the presence of silicon (from a
silicone oil contaminant) within a surface pore. This image reveals
parallel machining marks that had been present on the original
aluminium alloys surface. During the sulphuric acid anodizing process,
a porous lm of hydrated aluminium oxide would have built up. This
was then dyed by a nickel sulphide process, and then sealed by
immersion into boiling water. The crystallites of alumina are separated
by the curved lines; the remaining pores appear as dark spots. The main
elements on the surface are: Al and O being alumina; Ni and S from the
nickel sulphide inorganic dye; and Si from the silicone contaminant.
5.15.5 Contamination of Aluminium Tubes
for Vacuum Pinch-Offs
Ductile metal tubulations are frequently designed into
spacecraft systems in order to hermetically seal a unit that
may have been pumped down to vacuum, or may later be
lled with a particular gas or liquid. These tubes are usually
made of half-hard aluminium (such as AA-3003-H14 (see
Appendices 5 and 6 for heat-treatment and composition).
Occasionally, annealed OFHC copper is selected, when
glass-to-metal seals are included in designs for power
devices; and high-purity nickel tubing may be used for
systems holding liquids such as ammonia heat pipes, or
when high unit bake-out temperature will be envisaged
(nickel will have lower outgassing and less oxidation than Al
or Cu when bake-out temperatures exceed 150 C). Pinch-off
seals avoid the need for heavy valves and they have been
effective for generations of spacecraft. Before pumping
down the system it is essential that the area of tube to be
pinched-off is exceptionally clean. Ultrasonic and mechani-
cal cleaning is preferred over chemical cleaning.
However, in the case of a defective pinch-off (described
in the caption of Fig. 5.76b) it was necessary to use a pro-
prietary liquid to remove silicone contaminants from this
aluminium tube in order to achieve a hermetic seal. The
aluminium evacuation tube was connected by weldments to
an electronic housing containing electronic circuits for high
power applications. It had been designed as a leak-tight
housing for evacuation, then lling with the gas sulphur
hexafluoride (SF6). The chemical stability and high thermal
conductivity of SF6 allow it to extinguish arcs by thermal
transport; its breakdown voltage exceeds that of other gas-
ses, having a maximum of about 140 kV at 0.2 MPa pres-
sure. The concept was that this unit could be tested in the
laboratory, could be calibrated and space-approved, then
later would be expected to operate during the launch phase
under an ever decreasing ambient pressure, and nally in the
vacuum of space.
The tools for pinch-off processes include tungsten carbide
inserts that are pressed into the tube outer surfaces, causing
317
The contaminant was extracted with dichloro methane, a powerful
solvent and submitted to FT-IR analysis. This conrmed the presence
of silicone oil (the ngerprint corresponded to a reference sample of
silicone oil)see box, top right. b Pinch-offs were contaminated with a
thin layer of silicone oil which was presumed to have condensed from
vapour onto the internal surfaces of these annealed aluminium tubes
during an incorrect vacuum pumping operation. (i) Hermetic sealing by
cold welding had not been successful and the location of minute leaks
were only found after helium leak testing. (ii) Special cleaning steps
were introduced and once totally free of contamination, consistent
vacuum-tight joints were achieved. The processed pinch-offs are sharp
and very delicate so that threaded end caps are attached
them to deform, flatten together and, like a cold extrusion,
the internal surfaces should cold weld together. Constant
pressure, without any lateral movement, nally causes the
metal to thin down to a pointat which time it will instantly
sever.
Pinch-off sealing pressure can be delivered by either
pneumatic or hydraulic pumpsit is a delicate procedure
where tube material and outer diameter must be matched to a
particular tool size. Some parameters need to be optimized
and this must be done using sample tubes having precisely
the same characteristics as the flight-material (i.e. chemical
composition, hardness, grain-size, dimensions and surface
cleanliness). Samples will be visually inspected, occasion-
ally microsectioned, and might be submitted to gross (bub-
ble) and ne (helium) leak tests. The delicate sharp end of
the pinch must be protected as shown in Fig. 5.76b. It is not
recommended to enhance the quality of these cold welded
seals by further processing such as melting, welding or
soldering the tube end.
For a detailed account of evacuation tube closure, the
process development report by Zirker (2007) and co-workers
at the Idaho National Laboratory is recommended reading.
5.16 Magnetic Problems
The magnetic eld generated by a spacecraft interacts with
the local eld from the Earth and this results in an external
coupling force being exerted on the craft. The effect is
analogous to a compass needle in the Earths magnetic eld
and may torque the spacecraft and alter its attitudeso
requiring the ring of thrusters for station-keeping
manoeuvres. Owing to the high density of ferromagnetic
materials they are seldom used for large structural space
applications and are therefore unlikely to affect attitude
control. On the other hand, the presence of small-sized
magnetic materials on-board a spacecraft can limit the ef-
ciency of some detectors, such as magnetometers. For this
reason, care is usually taken to exclude these materials from
critical designs and only select from nonmagnetic materials
318
or those having paramagnetic properties (i.e. those materials
that, when inside a magnetic eld, behave with feeble
magnetic properties).
Several magnetic materials are used for their specic
properties on-board spacecraft. The largest number are in
electronic devices, with most of the remainder being in
electronic power apparatus (e.g. motors, transformers choke
coils, travelling wave tubes, reed switches, recorder heads,
ferrite-core memories, etc.). Where stray magnetic elds can
produce unwanted effects on neighbouring equipments, it is
possible to shield them with sheets or cans fabricated from
magnetic metals and ferrite cupssuch shields form closed
magnetic paths. A magnetic shield material often used in
space applications is Mumetal (ironnickelcopper alloy),
which can be wrapped around the magnetic components of
travelling wave tubes.
Electrolytic and electroless nickel plating can have strong
magnetic properties (see Sect. 4.10) and have given pertur-
bative effects on the instruments of scientic satellites
(Delahais 1985).
Many austenitic steels (that are usually stainless and
nonmagnetic) can become strongly magnetic when they are
heavily work-hardened. The 18-8 type of steels (designated
by AISI as 301, 302, 304, 316, and 321) are used throughout
spacecraft for applications such as fasteners, struts, stringers,
inserts, and in mechanisms. They have a high resistance to
stress-corrosion cracking and need no form of corrosion
protection except possibly a deliberately oxidized passiva-
tion layer. These alloys are nonmagnetic in the annealed
(fully austenitic) condition. However, the alloys become
magnetic when they are cold-worked owing to the
deformation-induced transformation of nonmagnetic
austenite to magnetic alpha-prime martensite. Small increa-
ses in the nickel content of these 300-series steels decrease
the extent of magnetization by cold-working: 301 steel with
about 7 %Ni becomes strongly magnetic, whereas 304 steel
with about 10 %Ni becomes somewhat less magnetic. This
increase in magnetic permeability is often called martensitic
stress induced transformation, MSIT. It must also be noted
that during high temperature heat treatment, 300-series steels
may undergo a microstructural change from austenite to
martensite if they are slow cooled from the austenitizing
temperatures.
The most common spacecraft parts to become magnetic
when they have been cold-worked are cold-rolled or
cold-drawn springs. Also, heavily machined or cold-worked
(by cold forging) stainless steel fasteners may have ferro-
magnetic outer surfaces in regions of their shanks and
threads. This can lead to problems when they are used for
the mechanical assembly of delicate electronic instruments.
A quick check to screen out magnetic items is to see if they
are attracted to a small magnet (an essential diagnostic tool
5 Metallography Applied to Spacecraft Test Failures
to be carried by all practising metallurgists!) or if they
interact with a small pocket compass. Alternatively, a more
sensitive test is to degrease the parts and immerse them for a
few minutes in a solution of acidied copper sulphate
metallic copper will deposit on those parts of the surface that
are magnetic (i.e. deposition will be on areas possessing a
martensitic or ferritic grain structure).
The factors affecting magnetic properties of metals
include their crystal structure and Curie temperature. It is
common that as the temperature is decreased, a materials
magnetic properties will increase. Conversely, as the tem-
perature is increased there can be a temperature where the
material becomes nonmagnetic (the Curie temperature is
dened as that temperature at which all of the metals
electron spins are disoriented by the randomising energy of
thermal vibration, making the net magnetization effectively
zero). Details of the magnetic properties of metals and their
alloys are contained in specialist books including the edition
by Cahn and Haasen (1996).
As stated above, the austenitic steels are generally
thought to be either nonmagnetic or paramagnetic, and this
was also considered true for most of the Inconel alloys (e.g.
Inconel 718, 625, and 600).
Cryogenic launch vehicles, such as Ariane 5, consist of a
main stage (referred to as the EPC from its title in French,
Etage Principal Cryotechnique) that is composed of an
aluminium tank with two compartments: one containing
133 tonnes of liquid oxygen (boiling point at ambient
pressure is 90 K), the other containing 26 tonnes of liquid
hydrogen (27 K). At the base of the EPC is the Vulcain
engine assembled mainly from stainless steels and Nimonic
alloys possessing a low magnetic permeability.
Scientic experiment detectors such as those installed in
the ISO satellites cryostat operate below the temperature of
liquid helium (7 K) at 4 K. Whereas the mechanical prop-
erties of many steels and nickel alloys are known at these
cryogenic temperaturesparticularly their ductile-to-brittle
fracture characteristicsvery little is known, or has been
published, concerning the magnetic properties of these
engineering alloys in this temperature regime. It is now
known that most of the austenitic steels and Inconel alloys
change their magnetic states when exposed to sufciently
cold temperatures. The precise Curie temperature depends
on the materials composition. The 300-series of steels can
change from the metastable austenitic (paramagnetic) phase
to a ferritic or martensitic (ferromagnetic) phase when
exposed to cryogenic temperatures. (The steel AM-350
contains 17 %Cr 4 %Ni, and 3 %Mo, and is deliberately
hardened by reducing the temperature to 75 C for at least
3 h in order to transform all the austenite to martensite;
subsequent tempering at 520 C for 3 h produces good
strength and ductility.)
5.16 Magnetic Problems
At very low temperatures a partial transformation of
austenite to martensite takes place in some of the less alloyed
stainless steels such as AISI 301 and 304. It is impossible to
quote, or tabulate, precise transformation temperatures (e.g.
Curie temperatures) for the different alloys. This is because
signicant changes can occur as the composition of the alloy
changes within its own permitted specied limits, and these
values will also be dependent on the parts thermomechan-
ical history. It is therefore recommended that special labo-
ratory tests are made for magnetic permeability for each
critical application.
Whether or not Inconel 718 can be used as a nonmagnetic
(paramagnetic) material will depend on the temperature, the
exact composition, and the limit of magnetic susceptibility
that can be tolerated for a given application (Goldberg et al.
1990). This is generally considered as a nonmagnetic
material and is used in many spacecraft applications. How-
ever, it appears to become strongly magnetic at a tempera-
ture close to that of the boiling point of liquid hydrogen
(27 K). As mentioned in the Goldberg et al. paper (1990),
this characteristic of Inconel caused a hydrogen fuel bleed
valve system to malfunction on the Space Shuttle Discovery.
The monitor for the valve position was made of Inconel 715,
and on cooling to the temperature of the liquid hydrogen it
produced erroneous readings; this initiated a metallurgical
failure investigation and caused a delay to the launch of
Discovery in 1988.
Finally, to make matters even worse, the transformation
of austenitic material into ferromagnetic martensite is
accompanied by a volume expansion which can spoil the t
of accurately machined piece-parts such as shafts and
flanges. If this is likely to be a problem, the stock material
can be immersed into either liquid nitrogen (77 K) or a
more appropriate liquied gas and repeatedly
thermal-cycled between the liquid and room temperature
until no signicant dimensional changes occur. This will
also often have the effect of stabilizing the low-temperature
magnetic properties.
5.17 Thermal Stress-Induced Dimensional
Changes
5.17.1 General Problems
The thermal environments that spacecraft parts might be
subjected to range from cryogenic temperatures to those
where materials are almost white-hot. Clearly, it is the
propulsion systems that witness the greatest thermal excur-
sions. At the other extreme, an ideal isothermal mass tem-
perature is taken to be 25 C, plus or minus 10 C, with a
319
maximum temperature envelope of 10 C. This would be
representative of equipment operating on the inside of a
spacecraft body where the thermal equilibrium is controlled
so that the absorbed energy is balanced by the radiated
energyas was described in earlier chapters.
Temperature changes cause materials to expand and
contract in different ways. Between room temperature and
that of liquid nitrogen, the total linear contraction of space
materials may vary from that of Invar and Pyrex glass
(0.05 %) to thermosetting resins (more than 5 %). Problems
can soon arise when materials are used together without
proper forethought. A simple example might be a stainless
steel bolt and nut that were used to assemble an aluminium
housing to a spacecrafts aluminium structure. As the craft
was warmed by solar heating, the aluminium parts would
expand far greater than the steel (23 ppm against 16 ppm) and
the steel bolt would be mechanically stressed so as to cause it
to yield and stretch. On cooling to a low temperature during
the eclipse portion of the crafts thermal cycle, the bolt would
now be too long and a gap between housing and structure
would be formed. In this case of the housing there is unlikely
to be a problem, as electrical grounding will be maintained by
the steel bolt. However, in the case of an aluminium bracket
attached to a thruster motor by means of a steel fastener,
continuous thermal excursions did work loose the xture.
Probably one of the most difcult engineering problems
to overcome is the stress-induced deformations that are
produced by temperature gradients within a cryogenic
propulsion system such as that shown in Figs. 2.112.13 for
the Ariane V launch vehicle. Here, gears rotating and
interlocking at 27 K must not be subjected to thermally
induced compressive forces that can cause friction, wear, or
seizure.
5.17.2 Stress-Relaxation by Thermal Gradients
The negative effect caused by the relaxation of stresses
during heating is now illustrated by reference to some fail-
ures associated with the construction of a large spacecraft
detector that had been designed to operate inside a large
liquid helium cryostate as seen in Fig. 5.77. The detector
instruments actually operate at a temperature of 2 K and are
electrically connected through the walls of the cryostat to the
rest of the spacecraft by means of a specially designed wire
harness. The signal wires of this harness are constructed
from single strands of enamelled steel, each having a
diameter of slightly less than 100 m. Steel was selected in
order to minimize the heat input to the cold instruments
(normally copper wire would be used for all electrical con-
nections, but this has a high thermal conduction). Electrical
320 5 Metallography Applied to Spacecraft Test Failures

signals to and from the instruments pass through the walls of

the cryostat by means of circular plug-like connectors con-
taining glass-to-metal pin seals. The processes involved for
the interconnection of the steel wires to the harness con-
nector are as follows. Initially the termination ends of the
thin steel wires are stripped of enamel and then crimped into
a nickel tube. This crimp is a temporary joint which only
serves to hold the wire and tube together so that they can
then be welded. Next, the tube and its slightly protruding
wire are manually welded by means of a small oxyacetylene
torch. The crimped portion of the wire and half of the fused
weld ball are shown in Fig. 5.78. The remaining electrical
connection processes are to insert the welded nickel tube into
a standard copper-alloy connector pin, and to apply solder.
This pin forms part of a circular (male) connector that can,
on completion of the harness, be coupled to the (female)
connector shell in the cryostat wall.
Electrical discontinuities were experienced during opera-
tion of the harness under cryogenic conditions. A metallurgical
investigation traced the problem to the steel wire-to-nickel
tube welding operation. The location of the open circuit is seen
from Fig. 5.79 to be at the interface between steel wire and
weld pool. The end of the steel wire from this interface is
shown in Fig. 5.80anot shown is the element analysis of the
fracture surface where mapping revealed local regions of pure
nickel on the broken surface. A deliberately broken wire
having a good weld can be compared in Fig. 5.80b; the
fracture surface is completely different.
Complementary tests proved that the open-circuit failures Fig. 5.77 The Infrared Space Observatory cryostat prior to thermal
were not caused by low-temperature embrittlement or cor- vacuum testing. This spacecraft was launched by Ariane in November
1995 and provided astronomers world-wide with a facility of unprece-
rosion. The initial cause of the failure was a defective weld; dented sensitivity for a detailed exploration of the Universe at infrared
this was proved to be a random problem associated with the wavelengths. Its 60 cm diameter telescope was cooled by superfluid
manner by which the operator performed the manual welding liquid helium to temperatures of 24 K. It made over 30,000 observations,
operation. The nickel tube was sometimes preheated by the covering all elds of astronomy, literally from comets to cosmology
welding torch for a fraction of a second before the steel wire.
This caused the tube to expand and relax its crimping effect
on the steel wire. Welding and alloying between steel and
nickel did occur (as seen in Fig. 5.79) but the weight of the
wire harness pulled at the weld zone, causing it to part at the
arrows point. On cooling from the welding temperature, the
broken wires were then again held in electrical connection by
the contracting nickel tube. The frustrating phase of this
failure was witnessed during some ground tests after lling
the cryostat with liquid helium. The harness wires contracted
on cooling from room temperature to 4 K, causing all the Fig. 5.78 A good steel-to-nickel tube connection. This microsection is
connections to be subjected to thermally induced stress. lightly etched and shows the position of the crimped nickel tube which
Wires possessing the weak, defective welds then became mechanically holds the steel wire during welding. The perfect welded
zone is on right-hand side
loose and were detected by electrical open-circuit.
5.17 Thermal Stress-Induced Dimensional Changes 321

Fig. 5.79 Cross-section through failed joint. The nickel tube is seen
together with the weld ballthese are soldered inside the copper-alloy
connector pin (solder has etched black and is not visible). The steel wire
break point is arrowed

Part of the remedial action was to check the integrity of

each accessible connection after applying a load to indi-
vidual wires in situ equal to 20 % of the ultimate strength of
a good weld. Those wires that came loose were replaced by
connections made to a revised, more reproducible procedure.
No problems due to this mode of failure were experienced
after launch of this very successful scientic spacecraft.

5.17.3 Thermally Induced Vibrations

The performance and behaviour of structures in space are

greatly influenced by the extreme temperature environments
that exist on the spacecraft. Thermally induced oscillations
and vibrations have been witnessed on many satellites, and
these were surveyed by Malla and Ghoshall (1994). Large
orbiting structures experience substantial change as their
orbit passes from the Sun side of Earth into the Earths
shadow, and out again. The Hubble Space Telescope
(HST) cycles between +105 and 105 C. At its 600 km
altitude, only 90 min are necessary to complete each circular
orbit. The transition from night to day takes the HST
approximately 10 s to complete, and vice versa. As for all
spacecraft, great care was taken to minimize the solar ther-
mal influence on the HST design and structural materials.
For example, the thin solar array and booms will heat and
cool much faster than the heavier spacecraft structures as
they pass in and out of sunlight. The transient mismatch in
thermal expansion/contraction between the thinner parts and
the thicker parts is accounted for in design by the incorpo-
ration of pre-tension spring devices and foam-like cushions.
After its launch in 1990, the problem of spherical aber-
ration of its imaging optics was discovered, followed by
Fig. 5.80 a Fracture surface of failed connector wire. b Good
connector wire was tensile tested; it broke when loaded to 900 g, the
fracture surface is ductile, with much wire elongation. Some areas on
(a) were analysed to be pure nickel, whereas the remaining fracture
surfaces supported mainly iron with traces of nickel and chromium
another problem: a pointing jitter induced by thermally
induced bending of its solar array. The vibrations occurred
when the arrays passed in and out of sunlight. The jitter was
caused by thermo-elastic bending of the HSTs bi-stems.
This deformation was later found to be caused by the
cumulative effect from a thermal gradient that builds up
across the bi-stems diameter together with the forces caused
by differential expansion between the bi-stem and the ther-
mal blanket. A photograph illustrating the bent array was
taken from the Shuttle at the start of the HSTs rst Ser-
vicing Mission in 1993 (see Fig. 5.81). Drawings showing
the location and operation of the bi-stems are shown
322
Fig. 5.81 The Hubble Space Telescope captured into the Space
Shuttle Endeavours cargo bay at the start of the First Servicing
Mission in December 1993. The array with the buckled boom (right
side) could not be retrieved. That on the left was returned to Earth
(courtesy of NASA)
schematically in Figs. 5.82 and 5.83. Calculations have
concluded that the main problem was caused by the bi-stem.
As it passed from eclipse and into the Suns rays, the outer
sleeve (steel spring2) of the bi-stem directly facing the solar
flux became hot (up to about 140 C). The inner steel spring
part of the bi-stem, shown as the insert in Fig. 5.83, warms
far slower than the outer part, as it is not directly exposed to
the Sun and because the thermal conduction path between
the touching surfaces of these steel springs is very inefcient.
The temperature difference between the two steel springs on
each bi-stem, caused by this differential solar heating, was
calculated to be approaching 30 C. This caused a different
expansion rate between the outer and inner springs. Not only
did they cause the 6 m array to bend, but there was some
slippage between the mating steel surfaces and, owing to the
effect of surface friction, this caused the jitter-like vibrations.
In essence, the orbital transition disturbances that occurred
prior to the rst Service Mission were caused by the build-up
and release of thermal gradients in bi-stem booms (Foster
et al. 1995; Deloo et al. 1995).
Corrective actions were to redesign part of the solar array.
The bi-stem deployment booms were covered with cylin-
drical thermal shields that easily deploy and retract with the
booms. This was constructed from circular pieces, cut by a
2
The solar absorptance to solar emittance ration, /, for polished steel
is 0.5/0.13 = 3.88.
5 Metallography Applied to Spacecraft Test Failures
Fig. 5.82 HST solar array deployment/retraction system (from Foster
et al. 1995)
Fig. 5.83 The solar array boom actuator drive chain with inserts to
illustrate the bi-stem deployment and, in section, how the two lengths
of spring steel curl together and effect the deployment of each bi-stem
laser beam from aluminized FEP Teflon,3 which appeared
like a bellows construction around each bi-stem. These
shields have been calculated to reduce the thermal gradient
between the bi-stem springs from almost 30 C, to less than
3 C. The stickslip behaviour of the stowage drum was also
altered and a new tension device was attached to the
spreader bar. These modications were made to the
replacement solar arrays which were attached to the HST in
December 1993. Flight data since that refurbishment mission
show that the new solar arrays are performing well and the
jitter and vibrations caused by thermal distortions have been
eliminated.
The rate of change in temperature for the bi-stems steel
spring facing the Sun, as it passed into and out of eclipse,
was 0.5 C per second on the defective array. The replace-
ment HST solar array has been assessed in orbit and cal-
culations show that with each new bi-stem protected by the
For aluminized FEP Teflon, the / ratio is 0.1/0.8 = 0.12.
3
5.17 Thermal Stress-Induced Dimensional Changes
cylindrical concertina thermal shield, the value for this
parameter has been reduced to 0.02 C per second, a
reduction factor of 25.
5.18 Defects in Titanium Piece-Parts
5.18.1 General
The main titanium material to have found extensive use in
spacecraft applications is the alphabeta alloy, Ti6A14V.
A limited volume of pure titanium has been employed, but
owing to some manufacturing problems it is the alloys that
will be employed for future advanced applications. Some
have been described in Sects. 4.17 and 4.20. The majority of
propulsion and reaction control systems are fabricated from
titanium because the pure metal and most of its alloys have a
good strength-to-weight ratio and are compatible with both
the hydrazine-based fuels and the nitrogen tetroxide oxidants.
Fuel lines are made of long lengths of formed and arc-welded
titanium tubes. Also, the nonreturn valves that are situated as
redundant pairs in the pressurant lines between the pressure
regulator and the propellant tanks are constructed from tita-
nium parts. The fuel, oxidizer, and high-pressure gas tanks
are generally fabricated from forged, machined, and welded
titanium hemispheres (Fig. 3.6). Some tanks have been made
by superplastic forming (see Fig. 4.77), but the majority are
from forged parts. Spacecraft detector housings, such as for
the Italian SAX scientic satellite, have been made from
formed titanium sheet that was later joined to beryllium
transition couples by explosive welding or to thin beryllium
windows by special braze alloysboth processes being
described in other chapters of this book.
The Ti6A14V alloy is forged above the beta transus (see
Fig. 5.18) to take advantage of the better plasticity and the
lower forging pressures needed at these temperatures (usu-
ally 870960 C). However, during forging there can be a
build up of temperature due to internal frictional energy and
this can affect the nal microstructure and cause the surface
of the alloy to become more prone to contamination. To
avoid overheating it is advisable to follow the surface
temperature of the forged item with an optical pyrometer. In
order to achieve a uniform alpha-beta duplex structure it is
preferable to cool the parts to just below the beta transus and
produce the nal working operation isothermally using the
superplastic properties of the material. A post-forge heat
treatment will also be applied in order to anneal the material.
The main causes of failures to titanium piece-parts have
been identied as resulting from either the formation of hard,
surface-diffusion layers (alpha case) on forgings, or the for-
mation of titanium hydrides within the parts microstructure.
Examples of both are given in the following paragraphs.
323
5.18.2 Alpha-Case Embrittlement
The high solid solubility of oxygen in titanium results in the
inward diffusion of oxygen from the surface during hot
working processes such as forging and tube forming. The
elevated temperature air exposure causes the oxygen, and to
a lesser extent nitrogen, to stabilize the alpha phase of the
alloy, so producing a hard and brittle layer, termed the alpha
case. This is often associated with the presence of visible
surface contamination consisting of oxide and nitride layers.
The hard, oxygen-rich layer needs to be removed from the
nished forging as it is likely to adversely affect bend duc-
tility and fatigue strength and lead to difculties during
machining and welding. Generally the surface oxides and
nitrides as well as all the subsurface alpha case is removed
from forgings by rst dislodging any adherent scale by
surface blasting processes and then acid pickling, chemical
milling, and/or machining. As previously inferred, the key
structural components in the systems for manoeuvring
spacecraft are the Ti6A14V pressure vessels. Massive alpha
inclusions caused the failure of several Apollo-era hardware,
including the Lunar Module descent propulsion system
which failed during proof testing (Ecord 1972). Experience
has shown that for unmanned projects the pressure vessel
design can be successful with a burst pressure factor of
safety of 1.5 provided strict controls are implemented for the
critical steps of forging, heat treatment, welding, cleaning,
and testing.
Problems arose during the fabrication of one large vessel
for a satellite. The manufacturing steps were to follow a
standard route. Two forged and machined hemispheres, one
half with an integral mounting boss, the other with a pene-
tration port for lling the fuel, were submitted for girth
welding. The conventional inspection techniques for the
hemispheres, both before and after welding, were to involve
radiography, ultrasonics, and dye penetrant testing. The
material was chemically analysed and test-pieces were ten-
sile tested and metallographically examined using off-cuts
from the forged blanks. All results were acceptable. How-
ever, one of the half-shells supported an unusual small area,
about 2 cm2, where the external circumferential machine
markings were associated with surface microcracks. They
could be highlighted by the dye penetrant but not by
X-radiography. Owing to this surface anomaly, the part was
considered unsuitable for flight use; however, it was welded
to a sound hemisphere (this was for welding operator
training purposes). Proof pressure tests were made and the
surface microcracks readdressed. Some cracks appeared to
have enlarged but this could not be quantied. The tank was
nally submitted to a series of destructive mechanical and
metallographic tests to assess the quality of the weld as well
as account for the microcracks.
324
Metallography showed the forged hemispheres to have the
correct ne-grain, alphabeta microstructure. In the defective
region the surface was seen to have been contaminated to a
depth of about 20 m. Here, the microstructure consisted of a
hard, light-grey zone that contained several cracks to a
maximum subsurface depth of 10 m. Clearly, this
air-contaminated surface layer had not been correctly
removedthe anomaly was reviewed with the forging sup-
plier. It became apparent that alpha-case formation had
occurred during the forging operation. The dimensional
design of the dies for these close-to-form precision forgings
was incorrect and insufcient material was being left for the
complete removal of the hard, alpha-rich subsurface layer.
The hard alpha case is difcult to machine and may reach to a
depth of 500 m. In this condition the forging can be pickled in
hydrofluoric acid (24 % aqueous solution) and material will be
removed at a rate of about 25 m per minute. To minimize
hydrogen pick-up from the hydrofluoric acid bath, the solution
is usually modied with the addition of about 30 % nitric acid
this solution has a reduced layer removal rate (about 10 m per
minute), but is essential if further problems associated with
hydrogen contamination are to be avoided.
5.18.3 Titanium Hydride Embrittlement
The problem of hydrogen contamination in titanium alloy
weldments was discussed in Sect. 4.4.4, where it was associ-
ated with the wrong choice of welding rod material. The
potential for hydride formation in Ti6Al4V when using unal-
loyed welding rods was probably rst recognized in the
Apollo-era as discussed by Ecord (1972) in his summation of
propulsion system failures occurring in that period. Because of
the small solubility of hydrogen in alpha titanium the dominant
embrittlement of pure titanium and the alpha alloys is caused
by the precipitation of titanium hydride. The beta phase of
titanium can absorb larger volumes of hydrogen (up to 0.5 %)
and generally all that happens is a distortion to the crystal lattice
of the beta phase. Hence, unless severely exposed to hydrogen,
the alphabeta alloys such as Ti6A14V are seldom affected by
embrittlement, as all the hydrogen contaminant is absorbed by
the beta phase. It would seem from the literature that titanium
hydride phases can occur, one of which is dependent on the
presence of a tensile stress [and can decompose when later
exposed to a compressive stress (Nakasa and Liu 1993)].
Hydrogen can contaminate titanium from many sources:
water vapour, pickling solutions, electroplating, chemical pol-
ishing, the presence of surface lms such as grease during heat
treatment operations, etc. Once hydrides nucleate they appear to
grow rapidly and deplete the matrix of interstitial hydrogen.
5 Metallography Applied to Spacecraft Test Failures
A particularly difcult case of hydride embrittlement was
discovered on a spacecraft when a length of titanium tube in
the propellant fuel line was found to have broken. The tube
had been fabricated from pure titanium, with an outer diameter
of 3.2 mm and a wall thickness of, nominally, 0.3 mm. The
integration of the fuel lines into the spacecraft had necessitated
much ground handling and transportation between manufac-
turing sites. The nal cause of tube failure was this trans-
portation. The length was incorrectly (severely) loaded by a
temporary xation device that held the complicated
three-dimensional network of pipes onto a transportation
frame. Metallurgical examination of the titanium pipe in the
vicinity of the fracture revealed many problems and deviations
from the material specication AMS 4942 (see Fig. 5.84):
large hydride precipitates lying on slip planes and along
grain boundaries
a large grain size (ASTM 6 instead of the specied 710)
an elongation at fracture of only 18 % (specied as 20
32 %); importantly, the tube supplier had tested and
found this to be 32 % prior to delivery
failed flarability test (indicating poor ductility)
wall thickness revealed by the microsection (Fig. 5.84)
was only 250 m instead of the required 300 microns
the internal and external tube surfaces were extremely rough
These ndings, together with additional information from
the company producing these fuel lines, conrmed that the
titanium tubes had been degraded by the application of
incorrect processing procedures. The damage was irreversible
and as the tubes could not be reclaimed they were rejected as
being unworthy for flight. The causes of the problem were:
this batch of tubes had been kept in an uncontrolled
storeroom for many years and had developed a thick
oxide coating
this coating had to be removed to ensure a suitable sur-
face nish for tube-to-tube TIG welding and to avoid
contaminating the spacecrafts liquid propellants
chemical cleaning (pickling) had been too aggressive,
removing excessive material from the surface and load-
ing the titanium with hydrogen
a further heat treatment had precipitated the hydrogen as
titanium hydride
the presence of surface notches and hydride platelets at
the tube surface caused the initiation and growth of
cracks resulting in breakage of the tube (the hydride
platelets are the sources of Grifth cracks, which develop
because of the poor cohesion between the precipitates
and the matrix and the difference in their elastic modulus
and plasticity)
5.18 Defects in Titanium Piece-Parts
Fig. 5.84 Microsection made through the wall thickness of a titanium
tube that had become embrittled with titanium hydrides. a Both the
internal and external tube surfaces are very rough owing to chemical
etching. b Detail from central part of wall thickness highlighting
locations of hydride formation. Polished sample was very lightly etched
to reveal grain boundaries
The complete tube network of the reaction control system
was replaced in situ. A description of the welding technique
used to integrate the replacement pipes on the spacecraft is
given in Sect. 4.16.
5.19 Leaking Water Tank on Launcher
The majority of space launch vehicles appear to have suffered
from problems related to stress-corrosion cracking (SCC). An
interesting early account of the SCC failures associated with the
Apollo programme was prepared by Johnson (1973). This
report includes many examples of metallic materials that had
been selected for structural applications without due regard to
the potential problem of SCC. The metallurgical investigations
into the Apollo and Saturn failed parts and welded joints often
conrmed that cracks had propagated through the
325
microstructure of materials having either a low or a medium
resistance to SCC (i.e. not those preferred alloys that have been
listed in Table 2.3). Several launchers consist of a liquid-fuelled
core vehicle with two or more solid-rocket booster motors
strapped to it by means of special joints. These joints, and the
bolts that mechanically fasten them, have a notorious record of
SCC problemseven to the extent that a booster motor has
actually fallen off a vehicle when it was still on the launch pad.
Any water that precipitates on launchers must be assumed to be
contaminated with ionic matter and have the possibility to
promote corrosion. Section 4.5.5 discusses how the majority of
this comes from the submicron particles of salt that are picked
up by winds passing over the ocean.
Three early Ariane launchers were forced into long
launch delays owing to the discovery of leaks in their tor-
oidal water tanks. The water tank (tore deau) is situated on
the second stage of this launch vehicle, as shown in
Fig. 5.85. The precise location of the water tank can also be
seen from the overall view of this rocket, together with an
identication of its major structural alloys and composite
materials, which was presented in Fig. 2.7. The toroidal
shape of the water tank enables good use to be made of the
space between the motor and fuel feed lines but its circular
shape makes its integration and xation into the vehicle
rather difcult. The tank is made of welded aluminium alloy
AZ5G (equivalent to AA 7020, the main constituent
elements being 1.2 % Mg, 5 % Zn, k and 0.2 % of Ti + Zr).
The same alloy has been used in several other locations on
this vehicle, but for the new generation of Ariane 5 launchers
it has been superseded by AA 2219.
AA 7020 possesses a good strength-to-weight ratio, is
easily weldable, but without suitable heat treatment is now
known to suffer from a low resistance to SCC. This was
certainly made clear from the failure investigation performed
on the leaking tanks. The initial water leaks were discovered
to flow from cracks in the water tank integrated on-board a
launcher at the Guianese Space Centre in Kourou (Seghesio
1986). Further cracks were then found during the special
inspection of the follow-on tanks that had been fabricated for
two subsequent launchers. The prime leak site of the tank was
submitted to metallographic examination. A microsection of
the crack, see Fig. 5.85c, showed it to have formed close to a
weld bead and to have passed completely through the wall
thickness of the AA 7020 alloy. The external surface of the
tank showed very little obvious sign of general surface cor-
rosion, but inside the metal volume, very thin hair-width
cracks passed intergranularly through its grain structurethe
typical morphology of an SCC mode of failure.
The failure investigation revealed that during its fabri-
cation, welding had introduced some very large residual
stresses into the water tanksome measurements and
326 5 Metallography Applied to Spacecraft Test Failures

Fig. 5.85 a Schematic drawing

to identify the location of the tore
deau (water tank)this cutaway
is part of Fig. 2.7. The actual
toroid-shaped tank, as integrated
within the launch vehicle, is
shown in (b). c Illustrates part of
the stress-corrosion crack in the
tank wall (reproduced by
permission from CNES)

calculations showed these stresses to have reached a value of
190 MPa in the short transverse direction of the formed
plates. This is about 68 % of the AA 7020 materials yield
strength (the 0.2 % off-set yield strength = 280 MPa).
Laboratory SCC tests (see also Sect. 4.6) made on AA
7020 welded sheet material showed that, even in the long
transverse direction, this alloy is prone to SCC in damp
(80 %RH) atmospheres. In saline solutions, the standard
SCC test on welded joints produced failures when the wel-
ded samples were stressed to 160 MPa. These results classify
the welded material as having a low resistance to SCC (i.e.
according to ECSS-Q-ST-70-37 this is a Class III material
and is unsuitable for spacecraft structural applications).
Corrective actions were to scrap the existing water tanks.
New tanks were manufactured (at that time, from the same
alloy) with a modied welding technique that resulted in
lower residual stresses. Also, the new as-welded tanks were
submitted to a nal stress-relieving heat treatment. The results
of laboratory SCC tests performed on welded and then
heat-treated AA 7020 samples had given extra condence that
these new tanks would possess an improved resistance to
SCC. No further SCC problems were encountered on Ariane
launchers during subsequent years and, after many successful
flights, the water tank design was changedso as to achieve
appreciable weight saving, these tanks are now fabricated
from CFRP material.
5.20 Compatibility of Liquid and Solid
Propellants with Components
and Subsystems
An overview of materials considered to be compatible and
incompatible with liquid propellants, based on evidence
going back to 1965 is offered in the standard
ECSS-E-ST-35-10. This standard covers the complexities
associated with compatibility testing of liquid propellants
with materials. It is probably easier to list materials with
5.20 Compatibility of Liquid and Solid Propellants with Components and Subsystems 327

known incompatibilities as these have been categorized from

the results of specic laboratory test methods, from the
failure of items under storage conditions, as well as from
failures occurring during component qualication and even
the catastrophic failure flight equipment. The tests required
for assuming liquids to be compatible are well dened in the
standard, but real knowledge of status can only be built up
over years of usage. For instance, nitrogen tetroxide (N2O4),
the liquid widely used as an oxidizer in spacecraft propul-
sion units, has always been considered compatible with
pressure vessels manufactured the alloy Ti6Al4V. A tank
failure, later determined to have been caused by stress cor-
rosion (Fig. 5.86), is an interesting case. It was determined
that the N2O4 was in fact too pure and this alone had caused Fig. 5.86 Pressure vessel made of Ti6Al4V which failed due to SCC
the tank to fracture (Johnson et al. 1966; Ecord 1972). By because of being lled with too high purity N2O4 [photo courtesy of
trial and error the formulation of the oxidizer was found to Dr. L.J. Korb, Rockwell Int. Space Division, from Brown (1997) and
be compatible with the tank walls if it contained between 0.6 Ghidini et al. (2008)]
and 1.0 % NO. This is now noted in the ECSS standard: the
higher the NO-content, and the lower the water content, the The upper stages of most launch vehicles are propelled by
better the compatibility with metals. liquid propellants, namely mono-methyl hydrazine
The incorrect application of materials can also occur due (MMH) and nitrogen tetroxide. Similarly, most satellites rely
to inadequate traceability, and the mixing of small items on the MMH/N2O4 combination of propellant and oxidizer
released from the stock room. This was a particularly because, despite their high toxicity and poor environmental
vexing problem during the manufacture of large tanks impact, this combination has a long history of usage, they
containing liquid propellant capillary management systems are compatible with materials which are readily available.
(CPSs) destined for use in a series of communications Also, they ignite spontaneously when they come in contact
satellites. The CPSs contain hollow vanes having one side with each other (i.e. they are hypergolic). Green fuels or
exposed to the tanks hydrazine or oxidizer by means of a propellants are so named because they should be easier to
very ne mesh (see Fig. 5.87). The mesh acts as a capillary handle and when combusted, the reaction products transform
pump enabling fuel to be pumped, by means of Ti6Al4V into non-toxic gasses.
tubing, to the spacecraft thrusters. Unfortunately the wash- The most common green propellant is based on the
ers of the 304 stainless steel screws uniting the vanes to the cryogenic combination of liquid oxygen and liquid hydrogen
internal walls of the Ti6Al4V tanks were composed of (LOX/LH2) as presently used for the main engines of
cadmium-plated carbon steel rather than the specied 304
stainless steel. Cadmium is known to be incompatible with
both liquids N2H4 and N2O4 and causes them to decompose
and release, respectively, gaseous ammonia and nitrogen
dioxide. Such gases can degrade the capillary action of the
CPSs and are highly undesirable. This small traceability
error was to cause a vast nancial cost to the project as
each nished tank needed to be disassembled, relieved of
the cadmium items, re-assembled and re-welded. Numerous
other materials are known to have short, or long term
incompatibility with hydrazine-based liquids, hydrocarbons,
Fig. 5.87 Fine mesh is contained in the windows of this hollow vane,
alcohols (possibly used for cleaning propellant hardware),
delicately laser-welded into an arc-shape so that many such vanes could
N2O4 and hydrogen peroxidefor these materials the be interlocked within the fuel tank as part of its diaphragm-free liquid
reader is advised to consult the latest issue of propellant capillary management system. The mesh acts as a capillary
ECSS-E-ST-35-10. pump to force fuel through the vanes, the tank manifold, and into the
titanium alloy fuel lines (windows are 2 mm wide)
328
Fig. 5.88 Extensive corrosion occurring on the external surfaces of a
nitrogen tetroxide storage tank welded from ASTM 321 stainless steel.
This tank was 12 years old and kept in an area shielded from rain, but
exposed to humidity and launch site contaminants such as hydrochloric
acid (HCl) and aluminium particulates. HCl is classed as a reducing
acid and lacks the oxidizing agents that stainless steels need to maintain
their passivated corrosion protection surface. 321 is similar to 304
stainless steel, but contains additions of titanium and niobium to
desensitise the material from carbide precipitation during welding. The
aggressive environment has however promoted a form of weld decay
and extensive pitting corrosion
launchers such as Ariane 5. Here the resultant plume is
composed of water vapour. However, the LOX/LH2 pro-
pellant has not previously been considered for scientic
missions because of the need to prevent the cryogenic liq-
uids from evaporation. This may now be possible for future
missions as cryogenic refrigeration is becoming more
mature. LOX/LH2 should be feasible for the propulsion of
future satellites and space probes, and (as already used for
boosters and launchers) hydrocarbons such as kerosene,
ethanol, methane, propane, etc. combined with LOX can be
expected to propel long-duration space missions.
Several new monopropellants, considered to be far
greener than hydrazine are under investigation. These
include:
HAN NH3OHNO3 (hydroxylammonium nitrate), also
known as AF-M315E,
5 Metallography Applied to Spacecraft Test Failures
HNF N2H5C(NO2)3 (hydrazinium nitroformate),
ADN NH4N(NO2)2 (ammonium di-nitramide), and
FLP-103S, a blend of NH4N(NO2)2, water, methanol
and ammonia, manufactured by ECAPS of Sweden which
decomposes into N2, H2 and CO/CO2. (The designation
FLP has been changed from LMP in order not to be
confused with certain HAN formulations). A newer
development from ECAPS is FLP-106, having a lower
vapour pressure than FLP-103S, this latest formulation also
based on ADN contains 23.9 % water and 11.5 % hydro-
carbon fuel. The water is added to the propellant in order to
reduce the combustion temperature of ADN as very high
operating temperatures would have required all parts of the
thrusters to be manufactured from alloys based on iridium
and rhodium (Hendrich et al. 2014).
The compatibility testing required by ECSS-E-ST-35-10 is
now being performed with the above-listed green propellants.
As was done in former years, these tests will encompass
pressurized tanks, piping, valves, lters, and heat exchangers
and compatibility with engine and thruster surfaces.
Booster motors generally rely on sold propellants based
on the oxidiser ammonium perchlorate (AP), NH4ClO4 and
aluminium powder distributed within a polymer binder
matrix. It should be noted that on burning AP forms huge
amounts of hydrochloric acid. The acidic exhaust plume,
plus some residual aluminium powder, is not only detri-
mental to health, but creates the conditions for accelerated
corrosion of the launch site, its ground station and infras-
tructure as shown in Fig. 5.88. Wingborg et al. (2008) cal-
culated an Ariane 5 launch to generate 270 tons of
concentrated hydrochloric acid per launch. These corrosive
additions are compounding the already corrosive natural
environment imposed by tropical seacoast weather.
In a Swedish development to reduce our dependence on
AP-Al powder solid fuel, a new solid propellant has been
formulated based on ADN. The ADN is formed into ne
particulates by spraying and later mixed with a binder called
GAP (polyglycidylazideC3H5N3O). This greener solid
propellant has been red successfully (Larsson & Wingborg
2011).
 Failure Analysis of Electrical Interconnections
and Recommended Processes
6.1 Material Problems
There are several classications of processes which will
metallurgically join parts as utilized in the production of
electronic circuitry. These interconnection techniques are
usually formed by the application or generation of heat, and,
depending on the temperatures attained, are generally
referred to as soldering, brazing, or welding. Where once
dissimilar metals have been in contact, the joining temper-
atures produce a continuous volume of metal or metal alloy
that is intended to enhance the electrical properties of the
joint and preclude interfacial corrosion.
Interconnections that do not involve metallurgical reac-
tions or the formation of alloyed interfaces are termed sol-
derless, mechanical, or physical joints and include
operations such as crimping, wire wrapping, and screw
fastening.
An incorrectly chosen method for making interconnec-
tions on any spacecraft may cause equipment electrical
interfaces to degrade in several phases of their operating
lives. Between assembly and launch, it is important that the
interconnections should be able to withstand certain mildly
corrosive environments. They must possess chemical resis-
tance to the acids secreted from an operators skin and to
cleaning agents and the trace impurities that may be dis-
solved in them. Resistance to atmospheric corrosion in
marine and industrial regions is critical when dissimilar
metals have a high galvanic incompatibility (see also Sect. 2.
2 and Tables 2.1 and 2.2).
When payloads such as electronic housings (black
boxes) are mechanically attached to the spacecrafts struc-
ture it is important that considerations are given to general
corrosion protection and surface electrical grounding. Tra-
ditional black boxes are made of aluminium alloys (for other
metals Table 2.2 should be consulted) and these can be
painted black/white or anodized to ensure against general
corrosion. The contacting attachments to the structure are
often removable (screw fasteners) but can be permanently
xed. There are some general rules for the electrical bonds.
Springer International Publishing Switzerland 2016
B.D. Dunn, Materials and Processes, Springer Praxis Books,
DOI 10.1007/978-3-319-23362-8_6
6
Removable ones are classed as H (resistance less than 0.1 )
or S (resistance less than 1.0 ), these resistance measure-
ments can be met with chemical conversion coatings such as
Alodine 1200. The aluminium alloys of the 3-, 4-, 5- and
6000 series can easily meet the H and S goals, but the 2000
and 7000 series of alloys may have difculties if subjected to
any humidity. All aluminium alloys can be plated with
nickel (magnetic) to achieve bonds having less than
0.0025 and these meet another class of bond, classed as R.
Vibration and g forces encountered during the launch
phase may produce open circuits as a result of the de-mating
of the contact surfaces of connectors, and the effect of fret-
ting or mechanical fatigue of the interconnection materials
may result in wire breakages. Once the spacecraft is in orbit
the interconnections will be subjected to the thermal excur-
sions described in Sect. 2.3; the extremities of these excur-
sions will depend on whether there is a spin velocity. For
stationary, non-spinning spacecraft, such as the Space
Telescope, interconnectors situated on surfaces exposed to
the Sun may be subjected to quite high temperatures
(c. 150 C) as a result of solar radiation. High temperatures
may also be expected in the proximity of power sources and
where heavy electrical currents have heating effects. In such
locations the use of low-melting-point solders would be
restricted. Special electronic circuits will operate at cryo-
genic temperatures, such as those being designed for the ISO
spacecraft when electrical interconnections must operate for
several years at liquid helium temperature (269 C). Cables
of superconducting NbTi may also be spliced with leadtin
solder and are expected to possess good mechanical strength
and a low voltage drop at 4.2 K.
6.2 Welded Lead Wire Interconnections
Resistance spot welding is an automated process in which all
process variables must be preset and maintained constant.
This is necessary becauseonce a weld has been made
present, nondestructive methods of inspection are difcult to
329
330 6 Failure Analysis of Electrical Interconnections and Recommended Processes

apply and are not completely satisfactory. It is customary,

therefore, to establish welding schedules by experiment and
to maintain the best possible control of process variables by
systematically subjecting test samples to destructive tests
such as mechanical strength and, as described in Sect. 4.2,
routine metallography inspection for comparison against
workmanship standards [as shown, for instance, in Fig. 6.1
(see colour section)].
One form of resistance weld requiring extensive metal-
lographic control is shown in Fig. 6.1. Here, a solder-coated
copper wire has been resistance butt welded to the Kovar
end-cap of an electronic component. This form of resistance
pressure welding causes the asperities on the Kovar (iron
nickelcobalt alloy) surface to deform on a microscale
without actually melting. The current and energy expended
in forcing together these surfaces bring the weld region into
a plastic condition. The thin layers of copper and nickel
oxides which are present as tarnish on these surfaces are
expelled into the weld nugget, and the weld contact area is
increased in size by plastic deformation of the copper wire.
Solid-state diffusion across the clean copper-to-Kovar
interface subsequently provides for a strong metallurgical
bond. Unfortunately, the heat generated by the welding
process, as witnessed by the extent of recrystallization within
the copper wire, causes a further metallurgical reaction at the
wire plating interface. The tin content of the solder-plating
reacts with the copper to form the brittle intermetallic,
Cu6Sn5. Microcracks in the newly formed intermetallic
surface surrounding the copper lead wire (as shown in
Fig. 6.1a) act as stress raisers, and have been seen to be the
origin of fatigue cracks which caused the complete failure of
component leads during the vibration testing of satellite
electronic equipments.
Electronic components are often found to be made with
both nickel and silver leads. Resistance spot welds between
nickel and silver have frequently been attempted by the
manufacturers of spacecraft equipmentusually with
unsuccessful results. An example of such a weld is shown in
Fig. 6.2. Although the visual aspects of the as-made welds
and the microsection have the appearance of a good weld, it
should be noted that such welded metal combinations have
very low mechanical strength. This is because nickel and
silver are totally immiscible in both the solid and the liquid
states (Hansen 1958), being analogous to a mixture of oil
and water. Equipment failures highlighted the total unsuit-
ability of this joining technique when used within dynamic
systems. Repairs to these systems were affected by initially
pretinning such leads and then soldering them together. Fig. 6.1 Electrical resistance weld between electronic component
The most suitable and easily controllable welds are made body (Kovar) and its solder-plated lead wire. a Weld zones 125. Heat
from welding produces (1) tincopper intermetallic which cracks to
by joining like materials. Nickel-to-nickel welded electronic form stress raisers, and (2) recrystallization of copper. b Weld nugget
circuits are particularly reliable as grain growth will take 300. This photomicrograph details (3) laboratory copper-plating for
place across surface interfaces during the weld pressure/heat good edge preparation, (4) remaining solder alloy, (5) the intermetallic
cycle as illustrated in Fig. 6.3. These microweldments were Cu6Sn5 and (6) washed-off oxides of CuO and NiO
6.2 Welded Lead Wire Interconnections 331

Fig. 6.2 a Nickel ribbon: arrow

points to silver nugget. b Silver
wire. c Photomicrograph of
weld. Optical photographs
showing the fracture surfaces of a
nickel ribbon (a) -to-silver
(b) electrical resistance weld. For
most welding combinations the
presence of a nugget would
indicate a sound joint. The
microstructure (c) shows that
deformation has provided some
mechanical strength at this
interface, but there has been no
metallurgical reaction between
the nickel and silver

made under microprocessor computer control during the
assembly of electrical components into a cordwood module.
The monitoring system assess the thermal expansion of the
weld nugget and surrounding material during heating and
melting. Welding at a relatively high current level produces
a rapid grain growth across the interfaces, and some liquid is
pushed out from the weld zone itself, culminating in a rapid
weld electrode contraction. Welding at a low heat setting
produces too small a spot weld size and an inadequate weld
strength under tension-shear.
Occasionally, welded parts may contain intermetallic
compounds at the interface between dissimilar metals. These
intermetallics are generally of a very brittle nature, and, once
they reach a critical volume, will jeopardize the welds
resistance to impact or mechanical shock loading. Although
undesirable, it may be impossible to avoid metallic combi-
nations which produce intermetallics, and in these instances
efforts must be made to restrict their rate of growth by opti-
mizing welding schedules to short times at the lowest pos-
sible temperatures. It may also be necessary to limit
332 6 Failure Analysis of Electrical Interconnections and Recommended Processes
Fig. 6.3 Cross-sections of nickel-to-nickel electrical resistance
microwelds possessing, respectively, high and low mechanical
strengths. a 0.015 inch diameter wire welded to itself. Note grain
growth across interface and presence of weld nugget expulsion.
b 0.015 inch diameter wire welded to 0.007 0.015 inch ribbon at a
low heat setting
in-service temperatures to prevent or at least control the
reaction and the growth of these intermetallics. An example
of one particularly troublesome system which has received
much attention is illustrated in Fig. 6.4. Incoming inspection
of these tantalum foil capacitors revealed some components
with broken leads, and lead-forming operations before
assembly onto printed circuit boards (PCBs) resulted in fur-
ther lead breakages. Bend testing resulted in brittle fracture,
and mechanical testing produced very wide differences in
mechanical strength for welded lead wires having a similar
visual appearance. Failure was always noted to occur at the
interface between the tantalum lead wire and a hard inter-
metallic phase that had formed during the welding process.
Welds with a thick layer of Ni3Ta intermetallic, together with
stress raisers such as cracks and voids around the weld
circumference, possessed the lowest strength. Rather than
optimize the butt-welding process to induce less heat and
consequently a lower volume of intermetallic, it was rec-
ommended that lap welds would be better suited for this
application. These components were found later to be unac-
ceptable for integration into spacecraft electrical systems.
6.3 Purple Plague
The phase diagram for the goldaluminium system was
determined completely in 1900 by Heycock and Neville, but
it was not until the mid-1960s that the electronics industry
became especially aware of the signicance of the gold
aluminium intermetallics that form during the
high-temperature wire bonding of microelectronicsand the
problems associated with purple plague and white pla-
gue. Purple plague is a name the industry has applied to the
region of intermetallics, a mixture of Au2Al and AuAl2,
which frequently surrounds the gold bonds on transistors and
integrated circuits, as shown in Fig. 6.5. The existence of
these purple-coloured small crystals surrounding gold ther-
mocompression bonds frequently throws into doubt the
reliability of certain microelectronic components and their
suitability for use on-board spacecraft. White plague is also
created by an intermetallic, Au5Al2, and as it has a low
electric conductivity, its formation can increase the resis-
tivity of the interconnection.
Gold wires, usually 25 m or smaller in diameter, are
attached to the aluminium metallized conductors of silicon
chips by a thermalcompression bonding technique which
employs the simultaneous application of force and heat. Two
techniques, ball bonding and wedge bonding, are used to
attach the leads, ball bonding being more reliable in most
applications. This microwire bonding provides the electrical
connection between the contact pads of an integrated circuit
die and the leads on its package. It is an extremely important
process, both for the reliability of the device and for nancial
reasonsat this stage in the manufacture of the device, 95 %
of the end cost has already been invested, so that scrap due to
poor processing must be avoided. The ball bonding process
has been developed during recent years and can now be
achieved by means of an electric flame-off technique. The
gold wire is melted to form a ball at its end by means of a
single power supply so that the welding electrode has a
positive charge and the wire a negative charge. During the
ball formation, both the wire and the ball are surrounded by a
plasma of ionized air. The plasma protects the liquid gold ball
and, by optimizing the electrode current and welding (spark)
time as the gold ball is mechanically pressed onto the devi-
ces aluminium contact pad, it is possible to achieve excellent
interconnections (Hu et al. 1995). Studies have shown that
ultrasonic vibration during this bonding process will remove
6.3 Purple Plague
Fig. 6.4 An evaluation of weak weldments between the tantalum wire
lead and its nickel wire extension on tantalum foil capacitors. a General
view of capacitor after lead bending. b Fragile lead that fractured during
rst bend through 90. c Section through butt-welded lead after etching
in nitric acid. An evaluation of weak weldments between the tantalum
oxides from the contact surface of the aluminium, so that
strong bonds occur only where the oxide has been removed
(Xu et al. 2009) and it is at this intimate interface between
wire and aluminium that forms a metallurgical bond.
During bonding, and even at temperatures within the
range to which assembled micro-circuits are sometimes
exposed, the gold and aluminium undergo mutual solid-state
interdiffusion to form a number of intermetallic compounds.
If the bonding temperature/time prole is excessive, or
operating temperatures of the nished device are too high,
thick layers of interfacial intermetallic can grow and these
conditions have been observed to promote open circuits and
component failures. The mode of failure has been seen to
333
wire lead and its nickel wire extension on tantalum foil capacitors.
d Detail of weld zones on (c). Nickel wire has recrystallized, a dendritic
structure containing Ni, Sn, and Ta has been expelled, and a white
intermetallic phase (Ni3Ta) exists at wire interfaces. e Some welds
possess stress-raisers around their circumference (arrowed)
vary, but one typical example of an open transistor circuit is
presented in Fig. 6.5. Many papers have been written to
explain the characteristics of goldaluminium interconnec-
tions, and just about as many conflicting conclusions have
been reached. The most noteworthy set of papers reviewed
by the author were presented as long ago as 1965 at the
Western Electronics Show and Convention in San Francisco
(Wescon 1965). It seems that there is a great danger in
extrapolating the conclusions drawn from one component
type to another, as each can have different conditions which
promote variations in intermetallic phases, kinetics of
intermetallic formation, effect of silicon, and the strength of
the bond. In recent years the process of joining thermosonic
334 6 Failure Analysis of Electrical Interconnections and Recommended Processes
Fig. 6.5 a Silicon transistor (type 2N3680) where gold lead wires are
attached to aluminium metallization on the silicon wafer. The wires are
attached by a thermo-compression bonding technique. b Purple
plague (arrowed) is formed at the goldaluminium junction and
spreads (migrates) along the aluminium at elevated temperatures. c The
gold wires to aluminium metallised pads has been ne-tuned
such that bonding times have been reduced to 210 ms with
a substrate temperature of 175 C. A recent interesting
investigation (Xu et al. 2010) concluded that during the few
microseconds of the bonding process the initial metallurgical
bond occurs at the peripheral regions of the gold ball. The
dimension of the contact area increased rapidly from the
circumference and slowed as the bond progressed to the
centre of the circular bond area.
The electronic device depicted in Fig. 6.5 is typical of a
component which was known to have a poor reliability. An
investigation was initiated using traditional metallographic
techniques in order to evaluate its most prevalent mode of
failure.
The author believes that the appearance of purple pla-
gue, as shown in Fig. 6.5, is in fact due to the optical effect
created when white light is diffracted from the extremely ne
network of intermetallic crystals surrounding the bond.
Under SEM examination (Fig. 6.5a) these crystals can be
resolved as having a very rough, irregular surface appearance
with a composition of AuAl2 (see also caption of Fig. 6.6).
The investigation demonstrated that associated with the
purple plague formation, and lying hidden underneath the
bond, is another compound of gold which is very brittle and
possesses a low adherence to the silicon substrate (as
depicted in Fig. 6.6). This intermetallic (Au5Al2white
plague) has been shown (Bushmire 1977) to increase the
electrical resistance of bonds signicantly as a function of
time and temperature in excess of 150 C. The migration, or
Kirkendall mass transfer effect, of gold into the aluminium
metallization away from the goldaluminium junction is
seen to leave microvoids around the periphery of the gold
SEM view of the same device shows the plague to have a ne
crystalline appearance. a Optical view of transistor, 80. b Detail of
Purple purple plague, 200. c SEM reflected mode appearance of
same bonds, 300
ball bond and a sharp edge in the zone of Au2A1 compound
which surrounds the inner circumference of detachment
(Figs. 6.6 and 6.7). The intermetallic zone was investigated
by actually microsectioning these minute ball bonds. The
packages of transistors similar to the failed device were
opened, and all conductors were prepared by electroplating
their surfaces in the laboratory with a protective layer of
copper. With meticulous care several bonds were micro-
sectioned and polished. These details are shown in Fig. 6.8.
The schematic diagram in Fig. 6.8c details two distinct
mechanisms postulated to initiate the failure of these bonds:
the void and notch formation due to mass transfer, and
the notch-initiated microcracks along strained interfaces
due to lattice mismatch between the intermetallics and
silicon substrate (particularly during thermal cycling).
It is understood that three intermetallics may occur in the
bond region, AuAl2, Au5Al2, and Au4Al, as shown by the
phase diagram reproduced in Fig. 6.8d. The microcracks are
thought to separate the gold-rich intermetallic phases, and, to
establish if this interface was also brittle, microhardness
impressions were carefully placed in this region of the bond.
Sets of hardness impressions are shown for one particular
bond in Fig. 6.8e, f. They were noted to induce slight cracks
only at the intermetallicsilicon interface and inter-metallic
gold interface. Incidental information determined the hard-
ness values of the constituents as:
Silicon 1690 MHN 5 g
Gold-rich intermetallics 385 MHN 5 g
Gold 59 MHN 5 g
6.3 Purple Plague
Fig. 6.6 Details of failed transistor thermocompression bond. The
rough crystalline circle of purple plague. a is noted to possess a sharp
well-dened inner edge (see insert). Lying underneath the gold bond is
Fig. 6.7 Optical view of aluminium metallization and purple plague
surrounding the detached bond (from Fig. 6.6). This device was
subjected to EDAX Analysis. The count rates obtained by using a small
areal raster identied the irregular surface of the purple plague area
A to be Au2Al. The region B compared identically with the standard for
SiO2
335
a brittle intermetallic compound (b); the insert highlights the Kirkendall
microvoiding effect associated with the purple plague formation
To avoid space component bond failures due to inter-
metallic embrittlement, Kirkendall voiding, and intermetallic
resistance at the semi-conductor chip, an effort has been
made to replace gold wire thermocompression bonding with
ultrasonic aluminium wire bonding. This technique results in
an aluminiumaluminium monometallic bond on the chips
metallized pattern.
The majority of todays spacecraft transistors and inte-
grated circuits utilize this more recent bonding technique,
and millions of performance hours with this monometallic
bond indicate no signicant problem with the chip metal-
lization interconnection. Occasionally it is found that par-
ticular components are available only with goldaluminium
chip bonds, and in these cases it is essential that their
interconnections are not exposed to high humidities, or
thermal shock during solder assembly operations (Richards
and Footner 1990).
The procedures for microsectioning the majority of
microelectronic components are rather well established.
They have been reviewed with respect to quality control by
Johnson (1979). Although the majority of laboratories today
still employ traditional metallographic preparation tech-
niques, specialist labs have invested in new modern methods
336 6 Failure Analysis of Electrical Interconnections and Recommended Processes
Fig. 6.8 Evaluation of the intermetallic systems associated with
purple plague. a In order that thermocompression bonds could be
microsectioned they were initially plated with copper in the laboratory.
b Vigilant metallography reveals the thermocompression bond and the
goldaluminium interaction layer. This sample was lightly stained, and
microscopical inspection could discern between two slightly different
coloured phases within the intermetallicthese observations are
that utilize Focused Ion Beam milling and Ion Beam milling
as illustrated by Fig. 6.9 (Erdman et al. 2006) and discussed
in Sect. 3.2.2.2. These new methods provide the same data
as grinding, polishing and etching specimens, but there is
no surface damage and evidence of cracking is much clearer
to observe. For instance, the microcracks seen in Fig. 6.8
sketched schematically in (c). Evaluation of the intermetallic systems
associated with purple plague. c Schematic diagram of observations
made during inspection of microsectioned bonds. d Phase diagram of
the goldaluminium system. Diamond pyramid hardness impressions
made under different loads in selected regions of thermocompression
bond. e Shows optical micrograph, and f SEM image of same area
(appears as mirror image)
may have become smeared or covered by gold and therefore
appear to be less extensive as those seen in Fig. 6.9.
A failure of this kind can be detected only by in situ tests,
which must be designed to simulate the calculated
thermal/vibrational environment at a particular location
within, or external to, the spacecraft structure.
6.4 Mechanical Electrical Connections 337

Fig. 6.9 Modern means of

cross-sectioning can avoid the
need for the laborious
metallographic preparation noted
in the work captured in Fig. 6.8.
A cross section polisher from
JEOL Ltd. uses of a broad argon
ion beam, eliminating problems
associated with conventional
microsectioning. Here, the
instrument has sectioned though
an aluminium-to-gold wire bond.
The result has a very high quality
and, despite the differences in
hardness between Si, Al, and Au,
reveals void formation,
micro-cracking and grain
structure (Backscattered electron
image, courtesy of JEOL)

6.4 Mechanical Electrical Connections retaining part at the opposite temperature extreme may
become negligible, while the dimensional differences may
6.4.1 General not even provide for an interference t.

The denition of these mechanical joints was given in

Sect. 6.1, and it is imperative that each type of connection,
whether it be a crimp joint, wire wrapping, or spring clip, be
subjected to a full qualication test before its incorporation
on a spacecraft. The selection of the contact materials is
critical, and it is necessary that their chemical compositions,
with acceptable tolerances, should be specied, and that
details of their relevant processing and thermal treatment
histories should be given. These joints often rely on inherent
springiness in the deformed materials to maintain the elec-
trical interfaces in contact (i.e. the deformed crimp barrel, or
the stretched wire wrapped around a terminal post). Some
materials, such as correctly heat-treated beryllium copper,
are perfectly acceptable for certain mechanical joints.
Materials that either are too brittle or, on the contrary, have a
very low yield point, will be completely unsuitable.
Often, dissimilar metals are in contact. In such cases,
thermal cycling is likely to produce a marked reduction in
contact efciency. Continual cycling expansion and con-
traction at the metal interface may cause one part slightly to
exceed its material elastic limit at one of the temperature
extremes. As a result, the impinging forces imposed by the
6.4.2 Wire-Wrapped Connections
Conventional wire-wrapped connections consist of a helix of
silver-plated solid copper wire that is wound, by means of an
electrically driven wrapping tool, around a terminal post to
produce a connection that is both mechanically and electri-
cally stable. The terminal post is either square or rectangular,
and is generally made of a nickel alloy, cold-worked brass,
or copper2 % beryllium that has been over-plated with
copper then gold. This form of solderless wrapped connec-
tion can be mechanically and electrically stable only if the
materials are correctly selected and the wire-wrapping pro-
cess is well controlled. A typical wire-wrapped connection is
illustrated in Fig. 6.10 together with some examples of the
quality controls that must be made to ensure suitability for
hi-rel applications.
All tools employed in the wire-wrapping process must be
well controlled by regular calibration. As a wide range of
wire diameters can be used for interconnection it is impor-
tant that the insulation stripper is well suited to the insulation
diameter and the insulation material. Thermal strippers are
338 6 Failure Analysis of Electrical Interconnections and Recommended Processes

Fig. 6.10 Metallurgical

inspection of wire-wrapped joints
(courtesy of the Departement
Fiabilit-Technologie, CNES,
Toulouse). a Typical appearance
of wire-wrapped joints made with
silver-plated single-strand wires
having different insulation
materials. b Cross-section
through wrapped joint after
environmental testing. The copper
is etched, and the micrograph
shows good contact between the
gold-plated terminal post and the
silver-plated wrapped conductor
wire

usually employed on AWG 22 and smaller wire sizes, as
there is a danger that mechanical-type strippers can stretch
these ne-gauge wires. The heat produced by such stripping
tools must be regulated to avoid blistering and excessive
melting of the insulation, otherwise a thin lm of insulant
could be drawn over the conductor so as to degrade the
subsequent wire-wrapped joint. Large-diameter wires are
usually stripped by means of precision automatic
power-driven strippers. Handheld precision strippers with
accurately machined and factory-preset cutting heads may be
used, with a caution to ensure that correct die openings are
used. Operator inspection steps must check that the plated
nish of the solid conductor is not scored by this operation,
and that the wire is neither stretched nor twisted.
A variety of commercial wrapping tools are available for
interconnecting panels and backplanes constructed of glass
epoxy PCB material supporting up to several hundred
wire-wrapping pins. Spacecraft wire-wrapped connections
consist of at least six turns of wire around the post for AWG
26 and larger, and seven turns for thinner wires. An addi-
tional turn of insulated wire around the bottom of the ter-
minal post is also required, as this adds a certain amount of
mechanical stability to the wire and helps to resist vibration.
Quality controls will include visual examination for
absence of overlapping wires or gaps between adjacent turns
and absence of pin deformation; electrical contact resistance;
tensile pull-off strength; an unwrapping test to check that the
wire has not become embrittled; and a gas-tightness test that
subjects wrapped connection samples to a specied chemical
attack. The sample illustrated in Fig. 6.11 had been exposed
to an atmosphere of hydrogen sulphide (H2S); the exposed
area of the connection (both wire and terminal) has become
discoloured. The gas-tight area of the connection is exposed
when the uninsulated wire is carefully unwrapped from the
terminal. Microsections of the as-made connection, as shown
in Fig. 6.10, allow further assessment of joint integrity. They
show the depth of corner penetration into the wire and
illustrate how the post corners may have deformed, thereby
indicating the suitability of the post for any subsequent
repair operation. Once the electrical reliability of the
process-controlled wire-wrapping operation becomes estab-
lished, only visual inspections and the tensile pull-off
strength should be necessary as in-line quality sampling
tests. Each nished panel should be electrically inspected
immediately after it is produced in order to verify the
accuracy of wrap locations and to detect hidden defects.
6.4 Mechanical Electrical Connections
Fig. 6.11 Demonstration of gas-tightness after exposure of intercon-
nection to hydrogen sulphide then unwrapping silver-plated wire.
Cold-welded points remain bright, whereas all other surfaces have
turned black
Particularly useful workmanship drawings and wire-wrap
photographs, together with electrical and mechanical
accept/reject criteria, are contained in the European speci-
cation ECSS-Q-ST-70-30, 2013, Wire wrapping of high-
reliability electrical connections.
6.4.3 Crimped Joints
The factors governing a good crimped joint are controlled to
a large extent by the tools and materials used. Wire stripping
has been mentioned in the context of wire wrapping, and
once stripped of insulation the bare wire must be inserted
into the connector pin until it shows in the inspection hole.
Crimping tools must be calibrated for the terminal size and
type, number of strands, and plating of the wire. Quality
requirements have been prescribed in ECSS-Q-ST-70-26.
339
One type of crimped cable lug which terminated
large-diameter wires for the Spacelab electrical harness was
investigated after the discovery that slight pushing, pulling,
and twisting actions on the crimped connection caused a vast
variation in electrical resistivity.
The problem crimps were subjected to metallographic
inspection and the resulting photomicrographs are presented
in Figs. 6.12, 6.13 and 6.14. The captions to these illustra-
tions describe the as-received condition of these devices.
Without doubt, physical changes would have occurred in
flight, owing to stress relaxation from vibration and thermal
cycling forces, so as to produce a highly dangerous condi-
tion. Further mechanical tests were made which proved that
the process for this crimp interconnection had not been
optimized (taking into account tensile strength, cross-section
area reduction, joint electrical conductivity, and microsec-
tion as prescribed in the governing process specication).
A quick comparison between Fig. 6.12 and a typical work-
manship standard as depicted in Fig. 4.4 shows these sam-
ples to be undercrimped.
A short programme of work was undertaken to optimize
the crimping process for this combination of wire size and
crimp barrels. The process procedure followed the operations
and quality assurance provisions of the relevant specication,
and resulted in design value crimps, optimized as plotted in
Fig. 6.15, with satisfactory cross-sectional appearances.
Silver-plated copper wire strands are used for most
applications when crimped interconnections are envisaged.
The use of nickel- and tin-plated wire in crimp-type con-
nections is not recommended. The reason for this is that
these nishes can present electrical problems. When tin plate
is connected to the gold plating inside the crimp barrel there
is a potential difference between these metals of 0.63 V
(from Table 2.1 it is seen that the EMF of gold is +0.15 V,
that of tin is 0.48 V). Long-term ageing tests, as well as
experiments where soldered joints were made to the opposite
end of a crimped tin-plated wire (with the possibility that
solder flux vapour had contaminated the crimp end), showed
conclusively that tin-plated crimped wires produced a con-
siderable increase in electrical resistance (Sandera 1996).
Tests made at the same time, and under identical environ-
mental conditions, but with silver-plated copper wire,
showed stable results and no increase in electrical resistance.
In the case of the silver-plated wire, the silver has an EMF of
0.03, resulting in a potential difference, when coupled to
gold, of only 0.18 V. An additional factor which may cause
the tingold couple to degrade is that they diffuse together to
form a brittle intermetallic of AuSn4, whereas silver and gold
are completely soluble in each other.
Nickel-plated copper wire is occasionally recommended
for large-diameter wires which are designed to be subjected to
much flexing (i.e. when the harness is joined to a rotating
mechanism). In such applications, silver coatings can
340 6 Failure Analysis of Electrical Interconnections and Recommended Processes

Fig. 6.12 Composite photograph

of 0 AWG crimp assembly. This
transverse section is taken
through volume of maximum
deformation. Although crimp
barrel is evenly deformed, its
appearance is as an undercrimp,
the voids between the wire
strands occupying more than
10 % of the wire volume.
A detailed examination of the
boxed area (arrowed) was made
to assess gas-tightness. From the
macrophoto this barrel area
appears to have the greatest
mechanical interface with
individual strands of the wire.
This American wire gauge
(AWG) 0 stranded wire has a
diameter of approximately 9 mm
and has the largest gauge, known
to the author in 2015, to have
flown on a European spacecraft.
Crimping of these large diameter
stranded wires will normally
require special disposition as it is
mechanically impossible to
achieve the strand-deformations
required by certain standards (see
also Fig. 4.4 on p. 120)

cold-weld and stranded wires become very rigid. Nickel
coatings, being harder, will slide over each other. However the
disadvantages of nickel must be taken into consideration: it is
strongly magnetic, it gives an increased resistance at elevated
temperatures, and is not suitable for low-voltage applications.
between soldering and brazing is to assume that the lower
temperature soldered joints are nonpermanent and are of
little strength, whereas brazed joints are designed to be both
permanent and strong.
Usual purchase requirements for the composition of
space-grade tin 63lead 37 solder are as follows:

6.5 Soldered Interconnections tin 63  0:5 % by mass;

lead remainder
6.5.1 Introduction to Soldering antimony \0:05 %
copper \0:05 %
Soldering and brazing are heterogeneous joining processes
bismuth \0:10 %
in which metals are joined together by a dissimilar, lower
melting point metal. The American Welding Society denes iron \0:02 %
brazing as a process in which the joining metal has a melting arsenic \0:01 %
point above 427 C (800 F). This temperature may be used aluminium zinc cadmium \0:002 %
to distinguish between brazing and soldering, although, in total of other impurities \0:08 %
practice, most solder metals are completely molten at tem-
peratures below 300 C. The majority of soldering and In practice, soldered joints cannot be reworked inde-
brazing metals are alloys based on binary or ternary eutectic nitely because there is always a reaction which creates a thin
systems and designedpossibly by means of addition ele- metallurgical bond between the solder and the metals being
mentsto have a good capillarity at the joining tempera- jointed. Repeated soldering or long soldering times increase
tures, and characteristics suitable to meet specic the volume of the reaction zone, and this has invariably a
environmental conditions. Another way to differentiate deleterious effect on the nal properties of the
6.5 Soldered Interconnections
Fig. 6.13 Composite photo showing detail of wire-to-barrel interface
in volume of greatest deformation (see box, Fig. 6.12). The internal
barrel plating (probably 4 m thick silver, over 0.5 m copper,
substrate 70/30 brass) has detached from the brass barrel in many
places. Letters indicate locations of subsequent photomicrographs
shown in Fig. 6.14 (lightly etched in ammonium peroxide)
interconnection, particularly when intermetallic compounds
are formed. Tin and its alloys form intermetallics with most
base materials during solder-dipping and soldering pro-
cesses. These intermetallics then continue to grow at a
greatly reduced rate by time and temperature-dependent
solid-state reactions during storage and operational life of a
joint. Under certain conditions, the tin component of the
solder may be totally consumed as intermetallic compound.
The higher temperature brazing operations form much
341
broader reaction zones of modied composition, and,
although properly chosen braze alloys will not form brittle
intermetallics, the modied microstructure will elevate the
melting temperature to possibly that of the work-piece liq-
uidus and thus preclude the possibility of rework. Brazing is
distinguished from welding by the fact that the brazed joint
is formed mainly by capillary action between the surfaces
being joined, whereas the welded joint starts by melting and
fusing the base metal surfaces together, usually with the aid
of a welding ller metal.
The basic European requirements for the hand-soldering
of components to PCBs are set out in ECSS Q-70-08. In the
US and other countries it is usual to apply the IPC J-Std.
001 Space addendum document (Appendix 9 lists titles for
ECSS, NASA and JAXA standards). Such standards cover
soldering facilitiessuch as the modern assembly room
shown in Fig. 6.16the tools, materials selection, mounting
designs, soldering parameters, cleaning, inspection and
means to verify solder-assemblies for spacecraft applica-
tions. Related standards refer to soldering of semi-rigid
cables (ECSS Q-70-18), repair and modication of assem-
blies (Q-70-28), surface mount technology (Q-70-38) and
these practices, together with training and certication of
personnel will be discussed further in subsequent sections.
For a comprehensive, practically oriented work covering
the whole area of mounting and soldering of electronics, the
reader is directed to the latest available edition of Scheel
(2004). This continually revised textbook includes sections
on design for manufacture, materials properties and selection
and a compendium on mounting for today and for the
futures very different components.
6.5.2 Inspection of Soldered Joints
To date, it is evident that soldering with eutectic tinlead
solder alloy has proved to be the most suitable method of
forming electrical interconnections designed to operate
below about 80 C for long periods (spacecraft lifetime), or
125 C for short periods (as during burn-in procedures).
However, this technique is clearly unsuitable in instances
where the maximum operating temperature approaches or
exceeds the melting point of the solder alloy as may be
expected at connections made to high-heat dissipating units
like travelling wave tubes, or exposure to the external sur-
faces of spacecraft like solar array wiring and connections
between solar cells. Today the use of lead-free solder alloys,
such as Sn96.5Ag3.0Cu0.5, that comply with RoHS and
WEEE directives are in wide use throughout commercial
industries, but because of reliability concerns the space,
military and medical industries have exemptions that permit
the continued use of SnPb alloys (see also Dunn 2012).
Space-quality soldering specications are followed during
342 6 Failure Analysis of Electrical Interconnections and Recommended Processes
Fig. 6.14 Details of microsection shown in Fig. 6.13. a Detail of
highly deformed wire strand-to-barrel interface. A very small gap (2.0
0.5 m) exists along this interface. b Possibly point contact at this
strand-to-barrel interface, but it is in region of detached plating
the mounting of electronics parts onto PCBs or by operators
making wire harnesses for power supplies. Before the issue
of the ECSS specications contractors had followed their
own in-house soldering specications which, when
reviewed, were found to be extremely scant, while it was
apparently assumed that the process was capable of being
carried out by even the most inexperienced personnel.
The European standards for soldering spacecraft elec-
tronics (ECSS-Q-ST-70-08, 70-26, and 70-38) presuppose
common PCB designs, materials, procedures, operator
training courses, workmanship, and inspection methods. It
was considered desirable to support this standardizing effort
by initiating a comprehensive solder-joint test programme,
which would investigate semi-quantitatively the failure
modes of space-quality joints. In addition, an effort would be
made to reveal and understand the failure and problem
(arrowed). c Detail of poor plating. d Wire strand-to-strand interface
showing good interconnection. The strand plating is 3.5 m nickel
which is seen to be well adherent and continuous around all strands
examined within the crimp connection
soldered interconnections which were discovered during the
course of constructing more than ten different spacecrafts.
Some of these ndings are presented in the following
sections.
Soldered assemblies were subjected to a 100 % visual
inspection at up to 7 magnication. In these instances, it
was essential that workmanship standards, in the form of
work samples or visual aids, were readily available to clearly
illustrate the quality characteristics of the soldered connec-
tions. Acceptable solder connections are generally charac-
terized by:
(a) clean, smooth, undisturbed surfaces,
(b) concave llets between solder and joined surfaces,
(c) complete wetting as evidenced by a low contact angle
between the solder and the joined surfaces.
6.5 Soldered Interconnections
Fig. 6.15 Metallography is an
essential tool for assessing the
quality of crimped joints. Typical
plots are presented showing the
variation in crimp-termination
characteristics with increasing
indentation depth
Unacceptable solder conditions which may be cause for
rejection are:
(a) damages, crushed, cracked, melted, corroded, etc.,
surfaces,
(b) improper tinning,
(c) flux residues or other contamination,
(d) cold joints,
(e) fractured joints,
(f) pits, holes, or voids in the joint which are not attribu-
table to liquid-to-solid solder shrinkage,
(g) excessive or insufcient solder,
(h) splattering of flux or solder on adjacent areas,
(i) dewetting, etc.
The preparation of soldered joint specimens for metallo-
graphic examination presents problems that are not common
to the preparation of most other materials. This is because
low-temperature solder alloys are so soft that they deform even
under low pressures. Also, these alloys have low recrystal-
lization temperatures, so that normal Bakelite hot-pressure
mounting of samples is not feasible. The microsections that are
presented in the following sections contain materials of vastly
differing hardness (e.g. indium-to-sintered alumina; tin
solder-to-Kovar; and joints containing hard intermetallics set
in soft matrices), so that special metallographic preparations
were required. Generally, the joints selected for microsec-
tioning are coated with a low exotherm, reasonably hard resin
to prevent disturbances as they were carefully sawn from their
343
respective assemblies with a low-speed,
diamond-impregnated rotary saw. The samples were mounted
in a resin curing at room temperature, sometimes in a vacuum
impregnation apparatus. The cured mounts were hand-ground
on abrasive SiC papers. It was essential that the specimen and
papers were frequently cleaned to remove loose particles of
sintered ceramic and abrasive that might later become
embedded in the specimen. To obtain satisfactory nal results
it was found essential to use diamond pastes for polishing.
Alumina polishing compounds were found to leave steps on
the sample surface. Initially, diamond paste having a 6 m
particle size was used for polishing. A napless cloth was used
for the rst stages to remove the unidirectional scratches from
the nal SiC paper. After a few minutes the specimens were
transferred to 1 m, then 0.25 m diamond pastes which had
been applied to short-nap cloths. Each polishing stage was
kept to a minimum time to reduce the formation of a step, or
bevel, which would have made examination of the interfaces
very difcult. Often the microstructure of solder alloys could
be distinguished in the specimens as-polished state. The tin
lead and indiumlead alloys could be etch-polished (i.e.
alternately etching and polishing until a bright, lustrous,
scratch-free surface was obtained).
The following etchants were employed (see also
Appendix 3):
(a) Diluted ammonium hydroxide with a few drops of
30 % hydrogen peroxide, for tinlead solder on copper
leads or copper-plated circuit boards.
344 6 Failure Analysis of Electrical Interconnections and Recommended Processes
(b) 2 % Nital (an alcoholic solution containing 2 % nitric
acid) for darkening tinlead and enhancing the
appearance of intermetallic compounds.
(c) Mixed acids (10 cm3 HNO3 + 30 cm3 acetic
acid + 50 cm3 glycerine) to show up the general
structure of tinlead alloys.
6.5.3 The Effect of Thermal Fatigue
on Solder-Assembled Leaded
Components
Several solder-joint studies undertaken in the late 1960s and
early 1970s into the reliability of conventional electronic
components assembled onto PCBs by soldering had shown
conclusively that thermal fatigue was the primary cause of
joint failure. The mode of thermal fatigue failure was
exacerbated by the fact that the samples were made before
the existence of the relevant space-quality materials and
process specications. In addition it was later shown that the
solder alloy contained impurities which might embrittle the
soldered joints. The thermal fatigue failures are characterized
by the optical and metallographic pictures shown in
Fig. 6.17. The mode of electrical open circuit failure is
depicted in Fig. 6.17d. Simulated launch environments,
including mechanical shock and vibration, were found, on
the other hand, to have little effect on the appearance and
quality of solder joints.
Fig. 6.16 Solder assembly
facility for spacecraft electronics.
The operators in this area perform
manual soldering in a room
continually monitored for
airborne particles, cleanliness,
temperature and humidity. Fresh
air is supplied to the room
through a ltering system and
there is a positive air pressure
difference with respect to adjacent
rooms (courtesy of Spur Electron)
Thermal fatigue cracking is caused by the stresses
repeatedly imposed on solder joints during thermal cycling,
when free thermal expansion is part or entirely restrained.
The magnitude of these stresses will vary according to the
thermal expansion coefcients of the different materials
which make up, or are adjacent to, the solder joint. In an
attempt to evaluate the effect of thermal fatigue on spacecraft
electronics, ten space-quality component-assembles PCBs,
together with two PCBs that were rejected upon inspection
because of poor workmanship, were subjected to three
phases of extensive thermal cycling. This culminated with
cycles between +100 and 100 C. The eutectic tinlead
solder alloy was of high purity and contained less than
0.05 % trace elements. A full description of the test condi-
tions and the materials employed are given elsewhere (Dunn
1979). A sketch of each joint design studies is shown in
Fig. 6.18. The joints were visually and metallograpahically
inspected at various stages throughout the thermal cycling.
As the joints visual appearance deteriorated with
cumulative thermal fatigue, a record was kept of those joint
llets which could be described as bearing surface discon-
tinuities pertaining to a crack. This rather loose description
could be substantiated only by destructive metallographic
examinations. The summarized results are shown visually in
Fig. 6.19. A number of photomicrographs are presented in
Figs. 6.20 and 6.21. For each gure the caption highlights
the signicant features. In the absence of brittle intermetallic
phases, crack propagation is noted to progress
6.5 Soldered Interconnections
Fig. 6.17 Stud-mounted resistor from thermal fatigue programme
using commercial-quality PCBs and solder alloy (solder contained
0.5 % antimony, 0.1 % copper, 63 % tin, and remainder lead). a View
of circumferential fatigue crack on llet. b Microsection reveals that
virtually no plastic deformation occurred on solder llet, 125. c Detail
from (b) showing that crack propagation in solder initiated by corner
intergranularly around the tin phases (white islands) and
transgranularly through the lead phases.
The relationship between the physical changes induced in
the various solder joint materials and the characteristics of
joint failure resulting from fatigue has been studied. It was
found that a solder joint is unlikely to fail because its ulti-
mate tensile strength has been exceeded, but that failure is
most likely to be caused by an accumulation of large cyclic
strains, which result from differences in the expansion
coefcients of the constituent materials. The strain rates
imposed by this form of thermal fatigue are very low, and, if
the mechanical properties of bulk materials are evaluated at
corresponding strain rates, their strengths are found to be far
345
cracking of PCB. Intermetallics are just visible along coppersolder
interface (between arrows). d. Non-approved PCB manufacturers
products may promote thermal fatigue in solder llets, as seen in (a) to
(c). The mechanical integrity of the PCB is important, particularly at
interfaces between laminate-copper and plated through copper (from
Dunn 1979)
lower than the values obtained in standard tests involving
higher strain rates.
The initial thermal fatigue to which each PCB is subject
occurs during immersion in a hot oil bath, which heat-cycles
the boards in a controlled manner to approximately 205 C
to fuse the tinlead plated circuitry. This has been described
in Sect. 4.7. This cycle is followed by a further heat input to
the plated-through hole during component assembly to the
board, when soldering iron temperatures are generally
around 250 C but are permitted by assembly process
specications to reach a maximum of 320 C. Space-quality
PCBs are not expected to be damaged by these
high-amplitude short-duration thermal cycles, although
346
Fig. 6.18 Various component
joints subjected to thermal
fatigue. a Clinched lead.
b Stud-mounted lead. c Off-set
lap joint. d Lap joint
Fig. 6.19 Percentage of llets on
space-quality boards with surface
discontinuities after thermal
cycling
6 Failure Analysis of Electrical Interconnections and Recommended Processes
boards of lower quality may be degradedas shown in
Fig. 6.16by the formation of cracks in the corners of
plated-through holes.
The coefcient of expansion of the type G10 PCB
material under test was measured by dilatometry in both the
boards longitudinal directions (x and y) and in the direction
perpendicular to the laminate (z). The results which are
compiled in Table 6.1 reveal thatat the higher temperature
the board will expand in the z direction by an amount that
is at least an order of magnitude greater than that in the x or
y direction. This difference is due to the layers of glass-bre
weave distributed throughout the PCB laminate, which are
effective in restricting expansion of the resin in the x and
y axes, so that the dimensional changes of the resin volume
are directed mainly into the z axis. A simple calculation will
indicate the magnitude of the displacement (dl) within each
plated-through hole during a hand-soldering operation. The
1
thickness of the PCB laminate (t) is 16 inch (1.59 mm), and,
in view of the maximum permitted soldering temperature
(T) of 320, then
dl t : az DT
where t is the change in temperature and z is the thermal
coefcient of expansion of the PCB in the z axis.
Substituting the measured values to be found in
Table 6.1,
dl 1:5932:580  20 95100  80 280125  100
340320  125  106 
0:123 mm
6.5 Soldered Interconnections
It is probable that this large calculated displacement does
not take place fully during the soldering of actual termina-
tions because of the restraining effect of the plated-through
copper. It is evident, however, that such copper deposits
ought to possess sufcient elongation and ductility to pre-
vent corner cracking problems.
The various forms of soldered joints that were subjected
to the different phases of thermal cycling during this pro-
gramme are shown in Fig. 6.18. In view of the difference in
coefcient of expansion values between the PCB material
and the component lead materials, each solder llet will have
been stressed to an extent dependent on the actual design and
material composition of each joint. The true stresses within
solder llets are, however, extremely complicated to assess,
owing to metallurgical effects which take place in the solder
alloy. At different temperatures, the competing effects of
strain hardening, time-dependent thermal softening, and
creep will each have a signicance diverging from the actual
mechanical properties of the alloy.
Fig. 6.20 Solder joint, after 800 cycles. The component side (a) of this
joint shows two circumferential discontinuities related to the hole in the
PCB and to the area of solder adjacent to the component lead.
Metallography has revealed (b) the presence of a medium-sized void
347
For a rst approximation, the dimensional mismatch
between the PCB and each component lead can be calculated
for the temperature extremes chosen for the various phases
of the programme. Mismatch of the component and the PCB
in the directions parallel to the surface of the PCB are not
considered, in view of the relatively compatible values of
coefcients of expansion between the PCB in the x and
y directions and those of the various lead materials. Also, as
shown in Fig. 6.18, each joint design will allow for sufcient
stress relief to compensate for such minute lateral
movements.
The greatest problem during thermal cycling results from
relative motion between the component lead material and the
plated-through hole. This motion is caused by the high
values of coefcient of expansion in the thickness, or z axis,
of the PCB. The amount of mismatch, dl, is calculated from
the expression
d1 aPCB  alead t=2DT
within the joint. There are no defects associated with the clinched
portion of the joint, but the component side (c) is seen to possess real
cracks, which have propagated a distance of 400 m alongside the
copper-clad Kovar lead material (d). a, b 15. c 125. d 300
348 6 Failure Analysis of Electrical Interconnections and Recommended Processes

Fig. 6.21 This stud-mounted transistor shows circumferential cracks
on each of its three soldered leads (a shows lead No. 1; b shows lead
No. 2). Microsection c shows the position of a transipad tp.
Examination of lead No. 2 (d and e) reveals that these cracks have
propagated from the llet surface into the solder at an angle of 45 to
the Kovar lead material. Inspection of the component side of the solder
joints shows extensive cracking alongside each of the leads f. The
detailed micrograph g reveals that these cracks are associated with the
brittle goldtin intermetallic AuSn2, which results from inadequate gold
removal from these leads by the operator before component assembly
onto the PCB. a 20. b 20. c 13. f 300. g 1350 (oil immersion
lens)
6.5 Soldered Interconnections 349

Table 6.1 Coefcients of expansion for various materials making up Phase 1 Phase 3
soldered joints
(20, +80 C) (100, +100 C)
Material Temperature Coefcient of expansion
range (C) (ppm C) Copper leads 1 m 3 m

High-purity 63Sn 100 to 0 17.5 Kovar leads 2 m 5 m

(63 %Sn, 37 %Pb) 0 to +100 14.7
0 to +50 16.1 The various forms of joints utilizing plated-through holes
100 to +100 15.9
have been soldered to termination pads having a similar
PCBs surface area. The gap between the component leads and the
Glass/epoxy type 100 to 0 6.6 walls of the PCB holes has also been standardized at
G10, FR4 0 to +50 5.1
between 0.25 and 0.50 mm. If these conditions are assumed,
(Tg = 125 C)
in x and y directions +50 to +100 11.0 thenas a rst approximationthe maximum stress that
in z direction 100 to 0 29.5 may be expected to built up inside each solder llet as a
0 to +80 32.5 result of the different thermal expansions can be calculated
+ 80 to +100 95.0
from the formula
+100 to +125 280.0
+125 to 150 340.0
E  dl
Fibreglass/Teflon r
t
(Tg = 75 C)
in x and y directions 100 to +70 20.0 where E is the engineering modulus of the solder
in z direction + 70 to +100 55.0
100 to +100 260.0
(31,600 MN/m2).
In the least severe situation of testing (i.e. copper com-
Glass/polyimide (Tg = 250 C)
ponent leads during Phase 1) the calculated resultant stress
in x and y directions 100 to +150 15.0
will be approximately 40 MN/m2. In the most severe case
in z direction 100 to + 150 50.0
(i.e. Kovar leads during Phase 3) forces of approximately
Polyimide/quartz (Tg = 250 C)
200 MN/m2 may be envisaged. Stress values in the solder
in x and y directions 100 to +150 7.0
surrounding Dumet leads will approach those in the case of
in z direction 100 to +150 34.0
Kovar.
Conformal coatings
The yield point of this solder alloy varies from 22 MN/m2
Solithane 113 with 100 to 10 52.8 at 60 C to 57 MN/m2 at 70 C (Dunn 1979). This indicates
300 catalyst (100:73) + 25 to +100 107.0
that, during Phase 1, all the solder llets made to copper leads
Component leads
that are soldered into plated-through holes will marginally
Kovar 0 to +200 5.0 exceed their yield stress. Plastic deformation of the llets is
Dumet 0 to +200 6.0
Copper 0 to +200 17.0
guaranteed during the Phase 3 thermal cycling. High-purity
solder undergoes superplastic deformation when exposed to
Associated package materials
tensile or torsional forces. Studies have shown that shear on
Quartz, fused silica 0 to +200 0.5
alternate planes in the thickness of this solder can push sur-
Alumina 0 to +200 8.0
Beryllia 0 to +200 49 face grains to the extreme surface of the material where they
Aluminium (6061 0 to +200 23.6 no longer participate in the bulk deformation process. These
T-6) have been termed floating grains (Mayo and Nix 1989). The
See also Appendices 1 and 2 initial surface rumpling observed on thermally fatigued sol-
Note Batch variations exist which account for the slight differences in der llets is assumed to result from these floating grains; their
measured values
presence is a precursor to crack initiation.
Thermal fatigue cracks are initiated at sites of plastic strain
where PCB and lead are the coefcients of linear expansion on the surface of the solder joint llet. Throughout this pro-
measured (see Table 6.1) for the PCB in the z axis and the gramme, and as evidenced by the cracks shown in Figs. 6.20
component lead, respectively. t/2 is the half thickness of the and 6.21, crack propagation is predominantly through the
PCB (i.e. 0.795 mm), and T represents the temperature dark lead-rich beta phase. Microstructural changes were
cycle extremes for each phase of the programme. observed in the solder alloy during temperature cycling with
The following mismatches, dl, between PCB and lead the formation of coarse bands of enlarged tin-rich and
materials during the thermal excursion have been calculated lead-rich grains. They appear in regions of the joint that
to occur within each solder llet: encountered the highest shear strains (as shown by Fig. 6.22).
350 6 Failure Analysis of Electrical Interconnections and Recommended Processes
Fig. 6.22 Typical grain coarsening in tinlead microstructure along a band encountering the highest shear strains (between arrows)
It is assumed that dislocations and vacancies generated in the
solder as a result of plastic shear will assist in solid-state
diffusion causing this band of material to coarsen as thermal
cycling proceeds. Initially the ne-grained, equiaxed solder
will undergo inter-crystalline sliding at the leadtin inter-
phase boundaries, and this creep-like process has been
observed to produce superplastic extensions to a few thou-
sand per cent before fracture occurred (Raman and Reiley
1987). Decohesion at the phase boundaries does not occur
unless certain impurities, such as aluminium, gold, tungsten,
and copper, are present within the grain boundaries. High
concentrations of contaminants can contribute to cracking
problems during thermal cycling, leading to a reduced
interconnect conductor area and degraded electrical perfor-
mance (OClock et al. 1987). The onset of grain coarsening
will further concentrate subsequent deformation at the
coarsened region, and after several hundred thermal cycles
the large grain size of the solder will no longer be able to
accommodate the imposed strain by grain boundary sliding,
and this might result in the formation of intergranular
cracks. The 1000 thermal cycles experienced in the authors
solder joint study showed no evidence that cracks in the
coarsened region would eventually lead to ultimate failure.
This fact is attributed to the presence of stress-relief bends in
the component leads and the selection of a very-high-purity
eutectic solder alloy containing less than 0.05 % trace
elements.
The following observations can be drawn from the ther-
mal fatigue studies;
(1) The number of solder joint cracks on space-quality
soldered joints was smaller and their severity was less
than in the case of similar studies performed on PCBs
made according to commercial-quality procedures.
(2) 360 C cracks on solder llet identied by visual
inspection are often termed failures. This is consid-
ered inappropriate unless the visual inspection is sub-
stantiated by metallographic or electrical inspections.
Visual inspection alone can lead to misleading results.
Continual plastic deformation of the solder llet during
thermal cycling causes surface discontinuities that may
be described as folds, ridges, orange-peel effects, and
wide-mouthed cracks.
(3) No interconnections were considered to have failed at
the end of this programme. Offset lap joints, as used for
mounting transistor leads, and the lap joints used for
flat packages suffered no cosmetic or metallurgical
degradation. Solder cracking on clinched leads was
conned to the bend radius of the clinch, and no sur-
face changes were observed along that portion of lead
lying parallel to the PCB conductor. Solder cracks
were present around the llets of stud joints, but they
were generally wide-mouthed and lacked appreciable
depth.
6.5 Soldered Interconnections
(4) Grain boundary sliding of the tinlead solder alloy
provides for stress relaxation and a creep-mechanism-
based possibility on lattice self-diffusion. However, if
brittle intermetallic phases, such as goldtin, are pre-
sent in the joint, then crack propagation is through both
the intermetallic and the lead-rich phase.
(5) Hairline cracks, which are potentially dangerous, were
occasionally observed to run alongside the Kovar or
copper-clad Kovar. They were not associated with
copper lead materials.
(6) Non-approved mounting methods, i.e. soldering of
gold, use of poor solder-ability lead materials, etc., will
produce substandard joints which may lead to prema-
ture electrical failures.
6.5.4 Effect of Thermal Fatigue on Leadless
Components
Spacecraft electronics generally comprise the leaded com-
ponents evaluated under thermal fatigue in the previous
section. These conventional devices are mounted onto PCBs
by inserting their leads or package pins into plated-through
holes in the PCB. Flat-packs with leads emerging from the
package sidewalls have been mounted by lap-soldered joints
to the board surface. The survival rate of these devices is
high because the formed leads can easily absorb stresses and
the solder alloy provides a reliable connection to the circuit.
Microcircuit assembly has gained full acceptance for
spacecraft applications; here the components and electronic
packages are surface mounted to the PCB by means of
soldering directly between the component body and the
circuit termination pad. In mounting these modern leadless
devices the solder is the only load-bearing member, and it
must absorb stresses as well as provide electrical intercon-
nection. Whereas ceramic chip capacitors and resistors are
commonly selected for spacecraft applications it is the
ceramic chip carrier and area grid arrays that provide the
most benets for space-bound electronics. The chip carrier
provides a safe package for sensitive integrated circuits and
offers a greater than 10 to 1 reduction in volume over present
duel-in-line package devices. The low mass of the chip
carrier improves mechanical shock and vibration resistance,
and its high internal component density requires far less
PCB surface area, provides a lower inductance, and
improves switching times.
The stresses in leadless surface-mounted devices must be
absorbed by the solder. The origin of stress is due either to
mismatch in thermal expansion between board and carrier,
or to thermal stressstrain loading on the joints by power
dissipation. As expected, this problem is increased as the
351
size of the chip carrier is increased, causing carriers to break
away from the board. For this reason it is better to use mixed
types of surface-mounted packages, i.e. J-lead components
and ne-pitch, high-lead-count, gullwing Very Large Scale
Integration (VLSI) carriers rather than solely leadless
components.
Multilayer ceramic capacitors are the most widely used
surface-mounted components in spacecraft electronics. They
are also very brittle and must be solder-attached to the circuit
by methods that will not cause thermal cracking of the
ceramic near to the metallized termination area. Preheating
of the part and substrate is a critical factor in avoiding
microcracks in chip capacitors such as that illustrated in
Fig. 6.23a. This form of damage is particularly insidious, as
in the case of the electrically failed capacitor shown in
Fig. 6.23b, where failure occurred when a spacecraft quali-
cation unit was submitted to 200 thermal cycles between
55 and +80 C. Thermal cycling propagates the initial
thermal shock crack in the vertical direction and will even-
tually split the capacitor in half.
Evaluation programmes have been performed using dif-
ferent solder alloys and various stand-off heights (i.e. dis-
tance between termination area on underside of component
and that of the substrate) to achieve suitable connections
between chip capacitors and a highly thermally mismatched
substrate (Teflon PCB in intimate contact with an aluminium
base plate). 75In25Pb solder was found to be slightly more
reliable than 63Sn37Pb, followed by a 2 % silver-loaded
tinlead solder. A satisfactory stand-off height was deter-
mined to be 0.2 mm. Laboratory tests, based on mechanical
straining a wide variety of solder lap joints, were made to
pin-point preferred designs for SMT assembly (Adile and
Dunn 1999); they concluded:
large solder llets are more reliable than small llets (all
joints having same prole),
a stand-off of 0.2 mm is better than either very small or
large stand-offs,
signicantly longer life-times were seen when fatigue
testing was made in an inert atmosphere (argon), possibly
similar to the observation for testing under vacuum with
no oxidation of the propagating crack and the occurrence
of cold welding.
The importance of increasing the compliance between a
leadless device and its substrate by incorporating a suitable
stand-off can also be seen from Fig. 6.24. Even a small
amount of solder in this gap can considerably increase the
number of cycles to failure. One specication requires from
0.1 to 0.4 mm stand-off (ECSS-Q-ST-70-38). Such a gap
will facilitate the cleaning of flux residues from beneath the
component package as well as increasing its resistance to
352 6 Failure Analysis of Electrical Interconnections and Recommended Processes
Fig. 6.23 a Microcrack starts at or just under the ceramictermination
interface as a result of thermal shock during soldering (the capacitors
metallized terminations and electrodes heat up and expand much more
thermal fatigue. Solder llet heights were found to be more
reliable when solder extended to a distance of 0.4 mm from
the base of the chip termination. Evaluation of mounted Area
Grid Array packages, using CSAM and electrical resistance
test monitoring, during thermal cycling also showed that
cracks propagate faster in low stand-off solder joints than in
the high stand-off solder joints (Liu and Lu 2003). When
solder llets have unequal volumes, as shown in Fig. 6.25,
the total strain resulting from ceramic-PCB expansion mis-
match is taken up by the smaller llet.
The CofnManson model has been used by several
workers to determine the fatigue life of leadless ceramic chip
carriers soldered onto PCBs, and the reader is recommended to
consult the original paper of Lau and Rice (1985) for an account
of the model predictions, i.e. fatigue life is increased by:
reducing the size of chip carrier,
decreasing temperature cycle excursion,
decreasing coefcient of expansion mismatch,
quickly than the ceramic dielectric). b Under thermal cycling the type
of crack seen in (a) can propagate under thermal or power cycling to
cross the silver electrodes
increasing the stand-off height,
increasing the ductility of the solder alloy.
The rst solution to avoid thermal mismatch is to select
ceramic substrates having the same composition as the
ceramic package. A second solution is to utilize advanced
PCB materials. Extensive testing (Houlberg 1988) estab-
lished that the following matched coefcient substrate
materials were suitable for space use:
(A) Epoxy/glass with two CuInvarCu cores. The core
thickness is 0.254 mm and the ratio of CuInvarCu is
12.5, 75, 12.5 %.
(B) Polyimide/glass with two CuInvarCu cores as above.
(C) Polyimide with quartz bre reinforcement.
(D) Epoxy with Kevlar bre reinforcement.
The PCB substrates A, B and C, used for early European
space projects, are no longer on the commercial market and
6.5 Soldered Interconnections
Fig. 6.24 Graphs to illustrate the importance of having an adequate
stand-off height, or sandwich of solder, between the metallization of
the component and the multilayer board (MLB) termination pad during
temperature cycling. These plots have been drawn to show the
predicted mean number of temperature cycles-to-failure for leadless
ceramic chip carriers (0.850-inch square package) assembled by solder
to three types of MLB substrate. This form of thermal fatigue occurs
during power cycling, the nominal ceramic chip temperature is 25 C
but this cycles to 50 C and back to 25 C during operation. The MLB
is assumed to reach a maximum temperature of 45 C. The coefcients
of thermal expansion are 5.5 ppm for the chip and the ceramic MLB,
21.5 ppm for the epoxyglass MLB, and 8.6 ppm for the copper-clad
invar MLB (Used with permission from Dr. Engelmaier 1995)
todays electronics mainly rely on epoxy/glass, FR4 and
polyimide/glass technologies. These non-matched lami-
nates include HTg-epoxy 370HR and rigid polyimide Arlon
35 and 85 N. Other specialist laminates with matched
coefcients to ceramic component packages are
non-woven, short aramid bres embedded in either epoxy or
polyimide (trade name Thermount, but this DuPont material
has been withdrawn from the market although large stocks
are available with space-approved PCB manufacturers and it
is possible a license to formulate this material may be
transferred to another producer). A new PCB technology
utilizes a carbon bre pseudo-fabric weave, embedded in
resin and sandwiched by copper foilreferred to as carbon
core laminate with the trademark name, Stablecore. It is
claimed to be superior to the aforementioned substrates
(A) to (D) and Thermount. The benets of Stablecore
include: ability to tailor CTE between 2.5 and 12 ppm; a
lower z-axis expansion than Thermount; is a good thermal
dissipater due to carbon weave and can reduce (weight)
amount of copper in ground planes; and, has extremely low
moisture absorption.
At the present time (2014) more than 75 European
companies and institutions are utilizing surface-mounting
technology on spacecraft. Soldering methods vary from hand
soldering, vapour phase, and infrared soldering with solder
pastes, to the reflow melting of specially constructed solder
columns around which are wound ne coiled wire that
353
becomes the lead wire once the solder has fused (to be
described in Sect. 6.15.5). Tinlead and indiumlead solders
are used together with the coefcient of expansion com-
pensated boards listed (A) to (D) above. Components are
generally leadless, but occasionally companies will them-
selves attach thin wires to packages to achieve some form of
stress relief on noncompensated assemblies. Qualication
programmes, as detailed in Sect. 6.15, that include thermal
cycling and vibration tests should be performed on all new
technologies. Photographs of typical leadless chip carrier
solder joints following testing are shown in Fig. 6.26 toge-
ther with details of the test plan. It is clear that because
surface-mounting technology is rapidly progressing to
include large ne pitch devices as well as large Area Grid
Array packages, having both ball and column terminations,
many issues related to reliability, processing, and availability
of space-approved components remain unresolved.
ECSS-Q-ST-70-38 and the IPC J-STD-001 and its Space
Addendum, have included important surface-mount
requirements. Some guidelines have been published con-
cerning the principles involved in the design, assembly- and
testing for surface-mounted electronic components [many
based on the groundwork of Lea (1988b) and Hinch (1988)],
and some will be reviewed in Sect. 6.15.10.
6.5.5 The Effect of Thermal Fatigue
on Semi-rigid Cable Connections
Solder joint cracking problems associated with the construc-
tion of semi-rigid coaxial cables have plagued many space-
craft programmes. Some cracks were due to poor handling of
the preformed cables during integration onto the spacecraft,
whereas others resulted from thermal cycling when the PTFE
dielectric contained within these cables expanded and tended
to protrude out from the cable ends. This protrusion can cause
the soldered interconnections to become stressed, and suc-
cessive thermal excursions may lead to solder joint cracking at
one end of the cable and the formation of air gaps between the
PTFE and adjacent mating surfaces at the opposite end. The
problem is compounded when the conductors and mating
parts possess gold plating, as this nish will form a layer of
brittle goldtin intermetallic when incorrectly soldered. These
factors have been addressed by Fletcher (1978), and it is now
realized that strict process controls must be adhered to if the
degradation of cable interconnections is to be avoided.
Additional examples are given in Sect. 6.15.7, and Fig. 6.77
reveals 5 mm of extruded PTFE!
Cross-sections made through defective joints are shown
and explained in Fig. 6.27. It is thought that deep drawing of
the copper cable sheath has created minute surface striations
354 6 Failure Analysis of Electrical Interconnections and Recommended Processes

Fig. 6.25 Unequal (asymmetric)

solder llets on small chip
capacitor; after 200 thermal
cycles between 55 and +100 C
the small llet has degraded.
a The small llet shows some
solder cracking beneath the
component (arrowed) and a large
fold in the llet. b Micrograph
showing that the maximum
critical resolved shear stress has
caused tinlead to undergo severe
deformation, and this volume has
subsequently recrystallized.
Crack has propagated to
X (starting underneath the
component). The wide-mouthed
crack in the llet. Y, does not
extend into the solder and appears
more like a fold owing to
extensive plastic movement.
c Owing to the very pure
composition of this eutectic
solder the crack to X appears to be
self-sealing (possibly by
cold-welding)
6.5 Soldered Interconnections 355

Fig. 6.26 50In50Pb solder joints

interconnecting 16-, 24-, 40-, and
64-pin leadless ceramic chip
carriers to PtAu metallized
ceramic multilayer thick lm
substrates. The appearance of the
soldered joints seen in (af) is
typical of both thermal-shocked
and power-cycled substrates. The
microsection (f) conrms no
electrical degradation. Note the
massive recrystallized grain
structure. a Optical view of
LCCC after environmental testing
[see Test Plan (g)]. b Optimum
llet size. c Small llet size.
d Detail of surface rumpling on
(b). e Surface cracking on (c).
f To be found in the colour
illustration section. g Test plan.
(f) Microsection through
mid-plane of llet seen in (c) after
polishing and etching to reveal
InPb grain size. The cracks
detailed in (c) are surface steps
where adjacent grains have been
strained to different extents. No
internal cracks are observed
356 6 Failure Analysis of Electrical Interconnections and Recommended Processes
Fig. 6.27 Section through connector-to-coax assemblies showing
cracked solder joints resulting from PTFE extrusion and resultant force
on solder. See also Fig. 6.77. a Suitable connector with non-captive nut
and good solder workmanship. b Non-approved captive nut; during
torqueing, stressing of solder joints can occur if inadequate allowance
has been made for torqueing distance (arrowed). Detail shows location
of circlip that permits rotation of the coupling nut, but not retraction as
photographed in (a). c Poor lay-up, when there is a dimensional
on the internal wallsthese cause ratcheting of the PTFE
towards one direction as it expands/contracts during each
heating/cooling cycle. Correct processing of flight cables
should follow the requirements of ECSS-Q-ST-70-18, when
the following critical steps must be adhered to:
mismatch between the coaxial cable assembly and the unit. Assembly
will be forced into position, so stressing the solder joints. d Typical
appearance of cracked joint due to either bending stress or PTFE Teflon
extrusion (after thermal cycling of non-preconditioned assembly).
e Macrosection through cracked solder and revealing the extent of
PTFE dielectric protrusion from semi-rigid cable following thermal
cycling
Selection of at least one cable-end connector with a
noncaptive centre contact and coupling nut (to avoid
stressing during integration).
Final soldering using 96Sn4Ag solder alloy to ensure
adequate joint strength.
6.5 Soldered Interconnections 357

(e)
Coax tube
Connector body
Solder joint

PTFE dielectric, protruding Teflon plug

Fig. 6.27 (continued)

Preconditioning the formed cable by heat treatments (at
least three complete thermal cycles between 45 and
+75 C) with trimming of any protruding dielectric.
Degolding of all surface areas intended to be joined by
soldering.
provide surfaces which are better suited for soldering. Two
types of metallic nish are employed for this purpose. They can
be loosely categorized into surfaces designed to be soldered to
and those that should be soldered through.

6.6.2 Surfaces that Can Be Soldered To

6.6 Problems Associated with Coatings These coatings should not appreciably dissolve during the
for Soldering Applications soldering operation. An example of this type has already
been seen in Sect. 6.5.3, and Fig. 6.20 detailed the copper
6.6.1 The Need for Coatings cladding of an otherwise unsolderable Kovar1 component

Many spacecraft electronic devices and assemblies employ

nonmetallics and/or metals that are either impossible to wet or
not easily wetted by tinlead solders. In these cases coatings are
applied by cladding, vacuum deposition, and electroplating to
1
Kovar is a special alloy designed to have a similar coefcient of
expansion to that of glass. It can be oxidized and used to form strong
glass-to-metal seals. Its composition is 2830 % nickel, 1719 %
cobalt, remainder iron.
358 6 Failure Analysis of Electrical Interconnections and Recommended Processes

Fig. 6.28 These micrographs

detail a the fracture of an
electrical connection on a
spacecraft solar array. The special
solder alloy has wetted both the
wire strand and the silver-plated
copper conductor, but failure
b occurred at the
copper-to-silver-plating interface
(arrowed), a 30. b 1000

lead. Such coatings are usually very adherent as they are
applied by hot mechanical working operations. They usually
give no problem during the thermal shock of soldering or
later during the mechanical loadings of service environ-
ments. However, if the coating is not well attached to the
substrate material by a sound metallurgical joint, electrical
failures may result from detached bonds, resulting in the
form of open circuit as highlighted in Fig. 6.28.
Thick coatings may also be applied to machined electronic
parts for other reasons. An example of a brass turret terminal
is shown by the series of photographs presented in Fig. 6.29.
The solderability of tinned parts which have been fabricated
from brass is severely reduced by long storage. Zinc is known
to diffuse from the substrate and onto the components free
surface. Oxidation or tarnish of surface zinc then prevents
solder spread, which cannot be promoted even in the pres-
ence of an activated liquid soldering flux. By specifying the
presence of a barrier layer of 23 m of copper or
nickel-plating, to separate the zinc containing brass from the
nal nish, the problem of poor solderability will be
removed. Solderability data recorded from brass samples
with and without a barrier layer have decisively illustrated the
value of using a barrier layer when poor solderability could
cause a loss in joint integrity during service (Scott 1987).
The condition known as black pad has been considered
in Sect. 4.10 and a transmission electron micrograph is
included in Fig. 4.50b showing the presence of small nickel
phosphide precipitates distributed within an electroless
6.6 Problems Associated with Coatings for Soldering Applications
40 m
Fig. 6.29 Microsections showing details of tin-plated brass turret
terminal pins after attempts to solder them onto PCBs. (a, recently
manufactured item.) Solder has wetted the new terminal pin. Pins from
the old stock were found to be nonsolderable, even with active fluxes
and high soldering temperatures. The high solder contact angles shown
in b (item from old stock) result from the inability of the solder to wet a
thin lm of zinc oxide present on the surface of these items. The X-ray
analysis (EDAX) reveals only zinc (large peaks), tin, and lead, with no
copper, thus proving the diffusion of zinc through this plated nish
nickel layer. The PCBs used in commercial electronic
equipment are often nished with electroless nickel
immersion gold (ENIG). ENIG is frequently accompanied
with a specic, latent defect if the electroless nickel bath is
not properly controlled within dened chemical composi-
tions. The symptom of black pad has been described by
Huang et al. (2011) and many others in the open literature,
without total agreement as to the underlying cause of the
problem. It is seen more often during the operational use of
area grid arrays, specically ball grid arrays, when shear
stresses applied to the solder interconnections causes a
fracture to propagate along the solder-to-nickel interface.
Fractured joints (on the nickel side) appear as black circles
359
having a brittle morphologyhence black pad. As com-
mercial industry tends to specify ENIG as a solution to
going lead-free it is clear that space industries should not
take the risk that black pad could bring to a high-rel unit and
this aspect is highlighted in an ESA guideline document
(Dunn 2012). Probably the best, brief reasoning for black
pad has been made by Gerard OBrian, President of ST and
S Testing in the US: Black pad occurs during the immer-
sion gold process and not during assembly which is only
responsible for revealing the defect. The primary factors are
a very aggressive gold reaction with the electroless nickel
whereby there is hyper-corrosion with the electroless nickel
surface that leaves an elevated phosphorous layer along the
interface of the gold to nickel. Phosphorous is not solderable
and once the gold is dissolved during the soldering process
the underlying non-wettable phosphorous-rich surface can
be exposed (after joint fracture).
6.6.3 Surfaces that Can Be Soldered
Through
These are frequently applied to form a protective layer on
delicate electronic circuitry, to prevent oxidation of the
substrate and to preserve the solderability of the underlying
layer. Gold, silver, palladium, tin, and solder are most
commonly used for this purpose, but under certain condi-
tions each may have considerable drawbacks.
Silver has been known to migrate under certain condi-
tions, and many documented failures due to this phe-
nomenon have been reported since the 1930s (Kohman et al.
1955). More recently, short-circuit failure under high
humidity is termed electrochemical migration (ECM). The
topic has been extensively reviewed by Krumbein (1988).
Silver migration has been seen to cause short-circuits
between the tracks of silver-nished PCBs and between
the sockets of silver-plated terminals. More recently the
author has been confronted with components which have
failed by diffusion of silver from silver-loaded epoxy
materials, and similar failure analysis results have been
reported by Chafn (1981). The environmental conditions
always required a standing d.c. potential, a given time under
potential, high relative humidity, possibly the ability of the
substrate to absorb moisture, and occasionally high tem-
perature. Silver migration is an electrochemical process
(ECM) whereby silver is removed ionically from its initial
location. Dielectric breakdown results when the ionic silver
has been reduced to metal and deposited in an amount and
form such that it provides a metallic conducting path where
none should be. The sites of galvanic dissolution can be
observed by metallographic techniques, but the
short-circuiting paths are generally so thin as to be located
only by surface analytical techniques such as ESCA.
360 6 Failure Analysis of Electrical Interconnections and Recommended Processes
Fig. 6.30 Tinlead electrochemical migration (EMC) under the
influence of an electrical (d.c) potential in combination with damp
conditions. These conductive dendrites shows a ferny morphology and
caused short circuit of conductors on a contaminated FR4 PCB (photo
courtesy of Mr. Wilhelm Maier, IFE, Oberpfaffenhofen)
The ECM of other metals such as tinlead (Fig. 6.30) is
much slower than silver but, because components are
becoming more miniaturized and PCBs have a ner pitch,
these failures are becoming more prevalent. Additionally,
silver will react with the slightest trace of hydrogen sulphide
in the air to form a brown or black patina which is essentially
silver sulphide. With the exception of silver-plated
copper-stranded wire, which is specially stored and pro-
cessed, the use of silver coatings for spacecraft electronic
circuit applications should be specically forbidden so as to
avoid these problems. Due to the numerous failures caused
by silver, NASA has forbidden this plating for use on PCBs,
bus bars, mechanical electrical contacts and special precau-
tions are requested when silver plated wires and cables are
selected for space use (NASA STD-651 2012).
Gold, with its properties of high resistance to tarnish and
corrosion, its low electrical resistivity, and contact resis-
tance, is a highly attractive nish for such spacecraft elec-
tronic components as connectors, terminations, and printed
circuits. However, in addition to its cost, gold has the severe
disadvantage of producing soldered joints of very low
mechanical strength. Joints made to thick gold-plated sur-
faces lose their electrical continuity under both vibration and
thermal cycling conditions. Many papers have covered the
effect of gold on the embrittlement of soldered joints (Bester
1968; Brewer 1970; Rothschild 1981), but even today the
problems associated with gold embrittlement are considered
rather controversial.
The temperatures and times required to cause wetting and
flow of eutectic solder onto a gold surface promote some, or
total, dissolution of gold into the molten alloy. The solid
solubility of gold in both tin and lead is probably less than
0.6 %, so that when the joint cools, gold is precipitated in the
form of a hard brittle intermetallic compound, AuSn4, which,
from metallographic observations, has an acicular structure.
Problems arise when certain authorities recommend that a
maximum of only 1.5 m thick gold platings be soldered.
Even such thin nishes can cause joints to fail due to gold
embrittlement, as has been witnessed by the partly degolded
lead shown in Fig. 6.21. It may be possible that extremely
thin platings of gold, such as less than 0.5 m thickness, will
dissolve totally in the tinlead solder without detrimental
effects, but such nishes are always porous and will cause
oxidation of the substrate during even short storage periods.
With thicker gold nishes it may be possible for some
industries to accept the presence of goldtin intermetallics
between solder and gold-plated surfaces by using materials
with matched expansion coefcients to eliminate thermally
induced stresses (Wright 1977), but this is totally unsuitable for
survival of vibrational loading during the launch of any
spacecraft. Quality standards stipulate that space hardware
shall not include joints made directly to gold plate, and the
recommended method for removing such coatings is to
immerse the part into a bath of pure tinlead solder. After gold
dissolution, the leads should be pre-tinned in a second solder
bath. As well as a double tinning process a dynamic solder
wave may be used for gold removal prior to mounting com-
ponents onto metallised circuits. The removal of gold should be
assessed on a sampling basis by metallographic examination.
An important note contained in the IPC J-Std-001F
(2015) states Gold embrittled solder connections can occur
regardless of gold thickness when solder volume is low or
the soldering process dwell time is not sufcient to allow the
gold to dissolve throughout the entire solder joint.
Many high reliable microelectronic components and
packages utilize ceramics because of their stability under
vacuum, mechanical characteristics and dimensional stability.
These ceramic devices are terminated, initially with a metal-
lization that bonds chemically to the ceramic, and is nished
with metallic layer that will eventually be solder-attached to
PCBs. Much has been reported about the MoMn metalliza-
tion system in the literature (see also Figs. 4.13 and 6.53) but
little is known of the multilayer ceramic tungsten metalliza-
tion applied for decades to packages such as leadless ceramic
chip carriers (LCCC). These metallisations are plated with
nickel, then gold. It is intended that the gold nish is applied to
maintain a good solderability of the nickel underlayer. Two
effects have been observed. One; mechanical and thermal
shocks can break the fragile tungsten-to-ceramic bonds and
this has been reviewed by Zhang and Jin (2009). Two; the poor
solderability of the nickel layer due to porosity in the gold, as
studied by Dunn (1993). Even today, the same problems exist.
De-golding by immersion in solder pots is essential for space
productsbut this has been found difcult owing to porosity
in the thin gold on nickel castellations of some CLCCCs. Care
6.6 Problems Associated with Coatings for Soldering Applications
must be taken to avoid thermally shocking these terminations
but activated fluxes must occasionally be used for oxidized
parts that were procured many years before they are used. It is
recommended that such devices/packages are stored in
nitrogen-lled bags or that de-golding and pre-tinning are
performed at the earliest opportunity.
Gold removal from soldered gold-plated dipoles was
omitted during the assembly and electrical integration of an
antenna system for a meteorological satellite. The antenna
was submitted to thermal cycling between 25 and +125 C
as part of its qualication testing. As a result of gold
embrittlement, many of the soldered joints were actually
heard to fracture by snapping within only a few days of the
systems test completion. Metallurgical evaluation of these
electrically open-circuit joints revealed the presence of a
high volume of brittle AuSn4 needles that had formed during
the soldering operation (i.e. gold dissolution into the liquid
solder followed by precipitation of the needles during
cooling to room temperature). Figure 6.31 clearly shows this
fracture surface to consist of a continuous network of AuSn4.
Experiments by Hannech and Hall (1992) duplicated the
solid-state diffusion of gold (Fig. 6.31b, c) that occurs during
thermal cycling to 125 C. Clearly, once the brittle network
of intermetallic reached a critical thickness, it fractured.
Each joint had been slightly loaded in tension by the nor-
mally acceptable tolerances of antenna positioning.
A far more obnoxious if not less costly failure due to gold
embrittlement involved the intermittent electrical failure of a
discrete transistor located on a ceramic platform. Excessive
gold dissolution from both a gold-plated Kovar pin and the
thick-gold metallization on the ceramic resulted in gold sat-
uration of the molten tinlead solder joint. The as-solidied
interconnecting network of intermetallics became slightly
stressed during a thermal cycling test, and produced a criss
cross of ne cracks; these are shown in Fig. 6.32.
Gold has a high solubility in liquid solder and is very
quickly leached from coated articles during the soldering
operation. When the concentration of gold in solidied
solder reaches 10 %, the joint is hard, brittle, and on
stressing the solder volume has essentially zero elongation.
Even 2 or 3 % gold in a joint markedly reduces the super-
plasticity of an uncontaminated solder joint. The presence of
brittle needles of goldtin intermetallic in the joints seen in
Figs. 6.31 and 6.32 will obviously degrade the quality of the
interconnection, but in many cases defective joints have
shown failure along the solder-to-gold interface when there
were no needle-like intermetallics within the solder volume.
In this instance, it is the distribution of the intermetallic layer
(shown schematically in Fig. 6.31c) that is important and not
the absolute gold concentration. Work by Kramer et al.
(1994) have shown that creep of pure, eutectic solder is
dominated by grain boundary sliding. This is the mechanism
by which solder relaxes when joints are strained by
361
mechanical or thermomechanical forces (i.e. during thermal
cycling when components have a different coefcient of
expansion to the substrate). The addition of only 0.2 wt%
gold suppresses this mechanism, and with 1 wt% of gold,
voids were seen to grow and promote cracking, leading to
early creep failures.
Solder-plating with a tinlead alloy is probably the most
trouble-free coating which can be soldered through. It is
generally fused to eliminate porosity. Palladium is also a
useful coating material, but is very expensive and nds little
usage within even the highest-cost electronic system selected
for spacecraft. Solder coatings have been presented previ-
ously, as for example in Sect. 4.7, but it must be remembered
that even these coatings may become nonsolderable and
embrittled if they are subjected to high temperatures. This is
due to the formation and growth of two intermetallic layers,
Cu3Sn and Cu6Sn5, both of which are hard and brittle:
Cu3 Sn 400 Microhardness Number and 5 g load
Cu6 Sn5 550 MHN 5 g
Copper 95 MHN 5 g
Solder 38 MHN 5 g
Under normal spacecraft conditions these intermetallics
do not give cause for concern, and even thermal cycling will
not cause them to grow to any appreciable thickness (in
Fig. 6.20 they are just visible). Their growth becomes a real
problem only when high temperatures (80120 C) are
encountered for long periods, because, when stressed, they
are prone to the form of cleavage fracture depicted in
Fig. 6.33. A further consideration is that of the effect of
elevated temperatures encountered either in a burn-in
treatment before the actual service of an electronic compo-
nent or during actual assembly. This may not only degrade
solderability owing to oxidation of the surfaces to be sol-
dered, but the growth of intermetallic layers which are
composed of tin and copper can make the surface become
lead-rich and this will also produce a solderability problem
(Billot et al. 1982). A systematic investigation has been
undertaken to assess the growth rates of compound layers on
various metal bases plated with both tin and its alloys
(Warwick and Muckett 1983; Hagge and Davis 1984). After
an initial fast rate of growth, the intermetallic compounds
were generally seen to grow in a uniform manner as a
function of time. Graphical presentations for most combi-
nations of materials and solder alloys show that these growth
rates obey the Arrhenius law.
A coating of fused tinlead eutectic composition alloy is
strongly recommended for use in high-reliability electronic
applications whenever the temperature constraints permit
(i.e. extended periods below 80 C). This nish could also
replace tin, cadmium, and zinc coatings which, as will be
seen in Chap. 7, have been specically excluded from
362 6 Failure Analysis of Electrical Interconnections and Recommended Processes

Fig. 6.31 a Fracture of (a)

spacecraft dipole connection due
to gold embrittlement of solder
alloy. b Semi-rigid cable solder
joint fracture due to gold
embrittlement, as-polished and
(right) etched. c Thickness of
intermetallic layers as a function
of ageing time from experiment
samples held at 125 and 80 C.
d Schematic representation of
(a and b) showing diffusional
flow during reaction between tin
from solder and gold that has
diffused to the interface between
intermetallic and solder [c and
d are courtesy of Hannech and
Hall (1992)]. See also Fig. 6.35
concerning strength reduction
versus time at 80 C

(b)

Connector body Gold

intermetallics

Solder
fillet

(c) (d)
6.6 Problems Associated with Coatings for Soldering Applications
(a)
200 m
(b)
100 m
Fig. 6.32 ac Illustrate a severely gold-embrittled solder joint. The
general photomicrograph a shows the effect of specimen bevelling due
to the extreme hardness variations between alumina ceramic, the Kovar
pin, and soft solder alloy. Intermittent electrical open-circuit between
spacecraft hardware as they can support stress-induced
whisker growth. Also, as has been discussed, the latter two
metals can sublimate in the space environment.
6.7 The Use of Indium Solder Alloys
The number of integrated circuits and PCB substrates
required for microwave purposes is growing rapidly. Con-
ventional spacecraft PCBs utilize tinlead nished circuits
with suitable spaces to interconnect the passive and active
components. They are, however, less desirable for
high-frequency applications owing to unavoidable signal
loss and distortion, and gold-plated conductor material is
363
(c)
20 m
pin and metallized pattern on ceramic results from the n network of
cracks b which propagate longitudinally through the acicular goldtin
intermetallics, and c along the solder-to-pin interface
necessary to provide an excellent surface conductivity
combined with the absence of surface oxide formation. In
some cases heavy gold metallization is plated onto con-
ductor tracks, and the termination pads attached to these
tracks, which will later be soldered with tinlead alloys, can
be masked and selectively plated with metallic nishes
compatible with tinlead soldering. When selective plating
is not possible because of circuit restraints it is important that
tinlead soldering to gold is avoided; as shown in Sect. 6.6,
gold is rapidly dissolved into solder with the formation of
extremely brittle intermetallic compounds. The critical
composition of gold in solder for the formation of brittle
joints appears to be about 0.6 % (McKay 1983) and may be
as low as 0.2 % (Kramer et al. 1994).
364 6 Failure Analysis of Electrical Interconnections and Recommended Processes

Fig. 6.33 Solder copper wire (a)

joint failure after long-term
thermal cycling. a Metallography
reveals the existence of two bands
of intermetallic. The Solder lead-rich phase
environmental testing promoted a
continuous network of brittle
cracks to propagate within the
Cu6Sn5. The wires could be
withdrawn from the joint, and,
Cu6 Sn5 intermetallic
with an SEM, their surfaces were
seen to be covered in containing microcracks
cleavage-fractured Cu6Sn5
crystals. These cleavage facets (b,
c) are seen to contain spherical
pockmarks which denote the
positions of impurity inclusions Cu3 Sn intermetallic
trapped within the microstructure.
(Note the Roman centurion!)
Copper wire

(b)

10 m

(c)

5 m

The most common indium-lead alloy contains 50 wt% of
indium, it has a liquidus temperature of 210 C and a solidus
of about 185 C. On cooling from the liquid phase, lead-rich
dendrites form and these can give solder joint llets a
somewhat dimpled or frosty appearance. The 50InPb alloy
has a good wettability to gold with mildly activated flux and
this is most suited for soldering to gold platings that have a
thickness of between 1 and 10 m. Soldering techniques are
generally the same as for tinlead solders, using hot-plates,
soldering irons and solder in ribbon or as a paste.
6.7 The Use of Indium Solder Alloys
(a)
(b)
Fig. 6.34 Metallographic evaluation of indiumlead soldered to
various gold interfaces after thermal cycling (100 cycles +100 to
100 C) and 12.8 years room temperature storage. a Overall view of
microsectioned chip capacitor. All circuit tracks are gold plated, as is
the aluminium support plate. b Detail of circuit-board-to-aluminium
support plate soldered connection. Aluminium is plated with an
intermediate layer of copper, and the nal gold layer has transformed
entirely to AuIn2. The gold-plated copper PCB ground plane is also
converted to AuIn2. c Detail of lower corner of chip termination. Note
that excessive pressure during solder assembly has deformed copper
It has been shown that indium-lead solders form unusually
ductile intermetallic compounds with gold. These solders do
nd use within electronic hybrid circuits for attachment of
small chip capacitors and resistors to ceramic and PCB
substrates (Yost 1976). The quality of such joints is fre-
quently assessed by the metallographic examination of test
samples made in parallel with flight production hardware.
365
circuit. d Hardness impressions in the AuIn2 layer on (e) to indicate its
ductility, i.e. no radial cracking as would be the case with goldtin
intermetallics. e Edge of solder llet to highlight 2.5 m thick gold
layer. This produces an intermetallic having a thickness of 14.0 m and
some free gold is still present (arrowed). f A gold-plated beryllium
copper alloy ribbon interconnector was also soldered to the Teflon
PCB. There has been complete reaction with the gold on the
interconnection side of the joint, and some free gold exists on the
PCB side. Although the IMC is extremely thick, there is no evidence
that these joints have degraded
A method for metallographic preparation and etching of
indium-to-gold soldered joints has been detailed by Millares
and Pieraggi (1992). An example of such an indium-soldered
connection after extensive thermal cycling and 12.8 years of
room temperature storage is seen in Fig. 6.34. It can be seen
that the reaction layer between gold and indium does produce
a layer of intermetallic, but this layer does not have an
366 6 Failure Analysis of Electrical Interconnections and Recommended Processes
(c)
(d)
Fig. 6.34 (continued)
acicular form, is relatively soft, and therefore causes no
embrittlement. The hardness of the main intermetallic is:
AuIn2 280 MHN 5 g
It is interesting to note that a 2.5 m plated gold layer will
transform by solid diffusion to form a 14 m reaction
intermetallic layer. Electron probe microanalyses have con-
rmed the intermetallic composition shown in Fig. 6.34 to
be AuIn2. Similar indium-solder interconnections were
investigated by Hauer (1977), who found that joints made to
miniature passive ceramic chip components have no
decrease in strength after 10 thermal shocks from 193 to
+70 C and 100 quick temperature cycles from 55 to +85
C after a high-temperature storage of 10,000 h at 85 C.
They can be considered suitable for most spacecraft
applications.
6.7 The Use of Indium Solder Alloys
(e)
(f)
Fig. 6.34 (continued)
Some caution must be observed when indium-lead sol-
ders are selected for high-reliability applications. These
alloys were noted (Goldmann et al. 1977; Banks et al. 1990)
to be more susceptible to corrosion than the SnPb alloys
when submitted to high humidity and high temperatures.
Whereas SnPb is stable at 85 C and 85 %RH, corrosion
products containing mainly In(OH)3 were detected on the
InPb alloy. The presence of chlorides appears to stimulate
the reaction by the formation of hygroscopic salts (e.g.
InCl3), but it would appear from the work of Kessowsky
et al. (1978) that a highly acidic environment, possessing a
pH of less than 4, is a prerequisite for corrosion, and from
their work indiumlead soldered to gold has proved quite
reliable from the point of view of joint strength, intermetallic
formation, and effect of long-time exposure.
367
The fatigue and shear strength properties of soldered joints
have been studied by Wild (1975) and commented on by
Jackson (1973). Between 55 and +125 C they established
that 50 %In50 %Pb provided a one-hundredfold increase in
fatigue life over eutectic SnPb. The excellent fatigue lives of
indium solders were also conrmed by Hashimoto et al.
(1991), who investigated VLSI package flip-chip-soldered
joints thermal cycled between liquid nitrogen and room
temperature. The packages were nished with Au/Pt/Ti and
Au/Pt/Cr metallizations and the In40 Pb solder alloy was
seen to be twice as resilient to thermal fatigue as Sn37 Pb
solder. The indiumlead alloy was found particularly
well-suited for hybrid assembly because it absorbed gold at a
slower rate than tinlead. A recent study conrmed that the
goldindium intermetallic growth rate related to InPb solder
368 6 Failure Analysis of Electrical Interconnections and Recommended Processes

Fig. 6.35 Peel strength versus ageing time at 80 C for various tinlead and indiumlead alloys soldered to gold. a Schematic cross-section of
peel test specimen after ageing. b 63Sn37Pb solder. c 1Sn99Pb solder. d 5Sn95Pb solder. e 50In50Pb solder. f 5In2.5Ag92.5Pb solder
6.7 The Use of Indium Solder Alloys
is far less than the growth rate of other intermetallics com-
monly found in soldered joints made with SnPb, SnSb,
SnAg, and SnIn (Frear and Vianco 1994).
The diffusion couples aged for nearly 13 years at room
temperature (Fig. 6.34) showed uniform and dense reaction
zones having a similar appearance to those observed by
Maciolek et al. (1978) that formed during 2000 h at a storage
temperature of 70 C. However, once ageing is performed at
temperatures in excess of 125 C the zones become
nonuniform and possess large columnar grains and large
voids. Voiding and an unstable intermetallic interface asso-
ciated with the intermetallic Au9In4 has been observed after
ageing at 150 C (Yost et al. 1976). Experience has shown
that InPb solder is suitable for the electrical and mechanical
interconnection of gold-plated articles with the important
proviso that operational temperatures do not exceed 70 or
80 C. The mechanical test data presented in Fig. 6.35
support the view that under these conditions InPb alloys are
far superior to those alloys based on tinlead.
The 50InPb alloy has been selected for the assembly of
components onto circuit board intended to operate at cryo-
genic temperatures. This alloy losses its strength properties
on cooling, but remains soft and ductile, even down to liquid
helium temperature. When compared to eutectic tinlead, it
is assumed that the greater stress relaxation in indium-lead
soldered joints during thermal cycling within cryogenic
regimes will provide those circuits with a higher reliability.
The mechanical properties of metallic and organic electronic
materials have been measured down to 4.2 K (Fink et al.
2008) and will be reviewed in Sect. 6.15.13. Two of the
common solders can be compared:
50InPb remains soft (E = 14 GPa) and ductile (A = 20 %) at
4.2 K but has low strength (Rm = 76 MPa) and low proof
stress (RP0.2 = 33 MPa)
63SnPb is stiff (E = 43 GPa) and brittle (A = 0.3 %) at 4.2 K
but has a higher strength (Rm = 136 MPa) and higher proof
stress (RP0.2 = 116 MPa)
Pure indium and indium solders are also used for the
sealing of glass and beryllium windows within the units of
satellite sun sensors and small-particle detectors such as
X-ray and gamma-ray counters.
Summarizing, indium-based solders have melting points
ranging from 93 to 314 C, and in general they will wet and
flow onto soldering surfaces in the same way as tinlead
solders. The wide range of melting temperatures enables step
soldering processes to attach wire leads in close proximity to
each other by means of the progressive use of lower
melting-point alloys. The lower soldering iron temperatures
369
can avoid reflowing or deforming the initial soldered
connections.
6.8 Wires and Cables
6.8.1 Selection of Plated Finish on Copper
Conductors
There is no wire conductor material that can be recom-
mended as the panacea for all potential corrosion, process-
ing, and application problems. The suitability of various
metallic nishes on copper wires for spacecraft applications
has been reviewed, and the choice nally limited to tin,
silver, and nickel. However, owing to its low melting point,
tin is generally not considered for the insulation materials
that require high-temperature processing, such as PTFE,
FEP, Teflon, and Tefzel, and Kapton.
(Incidentally, commercial insulation materials cannot be
used in space or vacuum applications. The most common
are: polyvinyl chloride (PVC); this has a high outgassing, is
considered a toxic material, and when heated releases
hydrochloric acid which can surface-corrode and
stress-corrode metals. Silicone rubber outgasses excessively
under vacuum. Teflon, even though widely used, tends to
cold-flow, and wire strands have been seen to creep through
this insulation in the locations of sharply bent wire har-
nesses; it also has a lower temperature rating than Tefzel.)
Tin-plated nishes
Tin-plated conductors have been successfully employed on
most spacecraft, but there remains some doubt as to their
long-term reliability, particularly after long-term ageing. The
reduction in solderability of tin-plated copper will be dis-
cussed in Sect. 6.8.2, but this is also accompanied by a
reduction in flexibility and an increase in electrical resistivity
owing to the progressive growth of tincopper intermetallic
compounds. The effects of tincopper intermetallics have
been studied by McCune et al. (1970), and the problems that
could develop during the in-service refurbishment or modi-
cation to a reusable space vehicle might include:
the increase in voltage drop through the intermetallic
layers present within crimped electrical contacts (see also
Sect. 6.4.3, p. 341),
difculty to wet with solder, and
simple handling during repair conditions could result in
wire fracturing.
A review of the literature, discussions with cable spe-
cialists, and the results of metallurgical test programmes
370 6 Failure Analysis of Electrical Interconnections and Recommended Processes

indicated many advantages and disadvantages when choos- other nishes. Contact resistance is affected by the
ing silver or nickel nishes. cleanliness of the surface. Ohmic resistance may be
erratic for nickel-platings (silver-plated copper con-
Silver nishes (more than 2 m thick)
nections exhibit negligible ohmic variations between
Advantages
similar joints).
(c) Nickel-plated conductors exhibit 710 % more DC
(a) Exceptional electrical conductivity (essential for certain
resistance with substantially greater power (I2R) loss.
RF applications), so that smaller conductors may be
(d) Nickel has an unsuitable magnetic moment for certain
employed than if nickel-plated (i.e. possible weight
spacecraft experiments.
saving).
(b) Good resistance to high temperature (200 C).
Enamel nishes
(c) Excellent solderability whether new or after many
Enamel-coated copper is used as a single-strand wire for the
years exposure to uncontrolled environments.
production of spacecraft coils and wire-wound magnets.
(d) Excellent crimpability, and suitable for wire wrapping
Fine magnet wire is coated with a thin, flexible enamel lm.
without cracking.
It is also used for the interconnection of tinlead-coated PCB
(e) Experience with this wire nish is very good in Europe
termination pads by the process known as Emal Technol-
and the USA (conrmed by spacecraft contractors).
ogy [trade name of MATRA (now Airbus Industries),
(f) Least losses from skin effects for RF transmissions.
France]. This process relies on electrical-resistance-welding
electrodes being placed onto the enamel wire surface. The
Disadvantages
application of pressure and current serves to burn off a local
area of enamel, and simultaneously the now-molten tinlead
(a) Poor plating and ingress of water into the wires causes
nish on the PCB is able to wet the exposed copper strand
a potential red-plague hazard (see Sect. 6.8.3). The
and produce a small soldered joint.
optimum silver-plating thickness is 2 m.
(b) The silver surface may react with substances such as
When used in spacecraft electrical devices the enamel
glycol to produce a potential flammability hazard (re-
wire must be capable of withstanding high breakdown
cent studies have also shown this reaction to occur with
voltages. Temperature ratings range from 105 to 220 C,
nickel and tin conductors).
depending on enamel composition. The most common
(c) Slightly higher cost than other coatings.
enamels are based on polyvinyl acetates, polyesters, and
(d) The larger gauge wires (AWG 0 and 4) cannot be
epoxy resins (Cohen 1995).
used with this nish as cold-welding at points of con-
Enamel insulators are known to lose their electrical sta-
tact between the wire strands may increase wire
bility during ageing. Electrical breakdown may occur
rigidity.
between the contact points of degraded enamel-coated wires.
The degradation of a given enamel system is dependent on
Nickel nishes
the thermal ageing temperature, time, and the presence of
Advantages
oxygen. The reactions occurring at the copper-to-enamel
interface have been characterized (Burrell and Keane 1988),
(a) No known potential conductor corrosion risks.
and sensitive surface analyses have shown that two separate
(b) Very high temperature rating (260 C).
ageing mechanism occur. The rst reaction results in the
(c) Minimal storage controls needed.
formation of copper oxide, Cu2O, on the copper wire surface
(d) Good flexibility after long-term ageing at 200 C.
owing to its reaction with oxygen from the air. The second
reaction involves the migration of copper ions into the
Disadvantages
enamel. If the enamel layer is sufciently thin, copper oxide
will grow on the top of the coating surface.
(a) Impossible to solder with the nonactivated fluxes. With
Ageing tests have been made with copper wires that sup-
activated fluxes, it has a variable solderability, depen-
ported approximately 5 m of enamel coating. Twisted wire
dent on the type of nickel-plating and thickness of
pairs were noted to fail by short-circuit after 5 months storage
nickel oxide. Active fluxes are undesirable for soldering
at 240 C. The prime cause of failure was the presence of
hook-up wires, as they are drawn under the insulation
conductive copper ions throughout the enamel coating.
by capillary action of the strands and can cause corro-
Reliable spacecraft enamel-coated wires must be tested and
sion, outgassing, and corona problems under vacuum.
found to be stable under a high-vacuum environment. It is
(b) Crimpability is fairly good, but more frequent adjust-
important that local hot spots do not exist at wire crossover points
ments have to be made to the crimping tool than with
situated on PCBs if long-term electrical shorts are to be avoided.
6.8 Wires and Cables
6.8.2 Effect of Ageing on the Solderability
of Tin-Plated and Silver-Plated
Wires
Introduction
The most common wire strands selected for spacecraft are
drawn and annealed copper that is plated with either silver or
tin. Both silver-plated and tin-plated wires have been suc-
cessfully interconnected within spacecraft subsystems by
soldering and crimping processes. However, precautions
must be taken to ensure a high quality of connection.
Silver-plated conductors have an excellent conductivity,
crimpability, and solderability, but they can suffer from red
plague corrosion unless certain manufacturing steps are
taken to avoid damage to the plated nish. A preferred
requirement for silver-plated conductors is to prescribe a
2.0 m minimum coating, and nished wires from approved
manufacturers can be subjected periodically to accelerated
tests designed to evaluate their susceptibility to red plague
corrosion (Dunn et al. 1984). Tin-plated conductors have the
lowest cost, but owing to the low melting temperature of tin
these conductors can be insulated only with low melting
point or irradiated insulations. The electrical conductivity of
tin-coated conductors is not as good as silver-coated con-
ductors owing to the lower conductivity of tin and the for-
mation of brittle intermetallic tincopper compounds.
Initially the solderability of tin-plated conductors is good,
but experience and literature sources show that this deteri-
orates with shelf life (Comiskey 1979).
To avoid potential corrosion problems and the possibility of
corona effects when small amounts of ionic flux residues
outgas from beneath wire insulation, it is recommended that
each soldering assembly operation is performed with (non-
activated) pure rosin flux. This necessitates that all soldering
surfaces must have an excellent solderability if the resulting
solder wetting angles are to remain low.
The following subsections describe the simple ageing and
solderability testing of lengths of tin-plated and silver-plated
conductors. Both conductors had been fabricated to military
specication requirements, insulated with the same radiation
crosslinked extruded modied fluoropolymer by the same
manufacturer, and both had a shelf-life since manufacture of
less than one year.
Wire samples
Each sample consisted of an 8 cm length of wire that had
been mechanically stripped to reveal 3 cm of bared con-
ductor strands. Cutting, stripping, and storage operations
were performed in a laminar flow clean air bench. The
samples were handled with clean lint-free gloves, and the
bared conductors were not touched. Five samples were
applied to each test variable. The supplied wires were:
371
Fig. 6.36 General view of Philips Globule solderability test equip-
ment. A60/40 solder globule (weight 100 mg). Bstripped wire in
position for test
Tin-plated copper AWG 24
Silver-plated copper AWG 24
Ageing tests
Steam ageing The samples were suspended above boiling
distilled water in a clean, glassware condensator column
for either 1 or 24 h (MIL STD-202E Method 208C).
Dry-heat ageing The samples were suspended for 16 h
within an air-circulating laboratory oven held at 155 2
C (B.S. 2011, test B).
Solderability testing
Philips Globule method Performed to BS 2011, part 2T,
with the appropriate solder pellet, at 235 C. Both wire
and solder pellet were rst coated with a nonactivated
flux solution. Wires were lowered to split the solder
globule in two, and the time was taken for the halves of
the globule to flow around the wire and unite above it (as
a measure of its solderability). This apparatus is shown in
Fig. 6.36.
Dip test Performed with Tri Moore solderability tester, to
Mil Std 202E method 208C, with fluxing in nonactive
rosin flux, then draining off flux until it becomes tacky.
After skimming the solder bath individual samples were
immersed for 5 s in the solder held at 230 C. Each
dip-coated wire conductor was then thoroughly clean of
flux and the 2 cm dipped portion examined by stereo-
zoom microscope at 10 magnication. According to Mil
Spec, accept/reject criteria include a 95 % coverage by
new solder coating and pinholes or voids not exceeding
5 % of total area.
372
Table 6.2 Results of ageing test
Sample and ageing description Evaluation of solderability (test methods) (average of 5 readings)
Each specimen has a length of 8 cm Philips Globule
and is stripped of 3 cm of in solation. Method BS 2011 part 2T
There are 5 specimens for each test (nonactive flux, solder at
conducted 230 C). Time to wet
(1) Tin-plated strands AWG 24
(a) As-received 1.78 s
(b) 1 h steam ageing 13.10 s
(c) 24 h steam ageing No wetting
(d) 16 h dry heat at 155 C No wetting
(2) Silver-plated strands AWG 24
(a) As-received 1.40 s
(b) 1 h steam ageing 2.36 s
(c) 24 h steam ageing 3.34 s
(d) 16 h dry heat at 155 C 10.40 s
See also Figs. 6.37 and 6.38
Pretinning of wires Using a hand-held soldering iron
maintained at 280 C, according to ECSS-Q-ST-70-08,
with Sn 63 cored solder wire containing nonactivated flux.
Scanning electron microscopy (SEM)
To supplement the visual inspection observations, samples
from the solderability tests were further examined in the
SEM.
Microsection
Microsections of the as-received and certain aged wire
conductors were made in order to take measurements of the
various plating thicknesses. Transverse sections were
mounted in a cold-setting resin, ground, and polished to a
0.25 m diamond nish and nally etched.
Results
The solderability test results are given in Table 6.2.
A number of typical SEM views of the post-solderability test
solder coverages on the tin- and silver-coated wire strands
are illustrated in Figs. 6.37 and 6.38. Microsections revealed
that the tin-coating thickness varied from local spots of
almost zero to 1.8 m. The silver plating showed a minimum
thickness of 0.8 m and an average of 1.7 m. After dry heat
ageing only the tin plating was modied (as shown in
Figs. 6.39 and 6.40) mainly into Cu3Sn intermetallic.
Some additional observations were:
the silver-coated conductors were easier to strip than the
tin-coated conductors.
after ageing the exposed tin-coated conductor surfaces
were noted to be slightly duller in appearance than when
newly stripped (particularly after the dry-heat environment
which appears to be the most severe ageing treatment).
6 Failure Analysis of Electrical Interconnections and Recommended Processes
Dip test to MIL std 202E Pretinning of wires to
Method 208C (nonactive flux, ECSS-Q-ST-70-08 with soldering
solder at 230 C). Solder iron held at 280 C (solder63 %
coverage in percent Sn, nonactive flux)
100 3 (easy)
7595 2/1 (moderate)
530 1/0 (difcult)
0 0 (impossible)
100 3 (easy)
100 3 (easy)
100 3 (easy)
100 3/2 (moderate)
Discussion
After exposure to the various accelerated test environments
each of the tin-plated specimens failed to pass either the
globule test or the dip test. By contrast, the silver-plated
conductors were not affected by the steam-aging environ-
ments, and the dry-heat exposure severely affected only the
globule test results. It would appear from the results shown
in Table 6.2 that trained operators would nd no difculty in
performing any of the soldering operations described in
hand-soldering specications when working with the
as-received wires (both tin- and silver-plated wires were less
than one year old). The steam-ageing environment was
chosen to be identical to that described in the USA military
specication for assessing wire solderability. Whereas the
silver-plated conductors were not affected by this environ-
ment (as shown in Fig. 6.37), the tin surfaces did degrade,
and the poor solder wetting illustrated in Fig. 6.38 would,
under wire harnessing conditions, lead to unsatisfactory
solder joints. According to a detailed study by the IPC the
effect of a years storage can be reasonably well duplicated
by exposing tinned copper surfaces to 20 or 24 h of steam
ageing (Schoenthaler 1984). Clearly, in many situations
spacecraft wire harnesses would be more difcult to
rework/repair if tin-plated conductors are chosen instead of
silver nishes. It should also be noted that the present
specications for solderability testing of as-manufactured
wire do not require any ageing to be performed before the
solderability testing, so that the effects of long shelf-lives on
solderability cannot be fully guaranteed.
The maximum use temperature as per the relevant mili-
tary specications is 150 C for tin-plated conductors and
200 C for silver-plated conductors. The results of the
dry-heat ageing tests demonstrate that it would be impossible
to perform soldering rework or repair operations to wires that
might have been exposed for short periods to elevated
6.8 Wires and Cables 373

Fig. 6.37 Silver-plated wire

strands after solderability testing
(dip-test method), a as-received
condition; b after 1 h steam
ageing; c after 24 h steam ageing;
d after 16 h at 155 C. All give
approximately 100 % wetting

temperatures of the order of 150 C. The micrographs in
Fig. 6.39 indicate that the wire strands were hot tinned, as the
reaction layer between copper and tin has an appreciable
thickness (about 0.5 m of Cu3Sn). It is expected that the free
tin existing on the surface of the as-received strands will
facilitate solder wetting. The ageing processes are known to
increase the intermetallic compound thickness by an Arrhe-
nius rate of growth until the free tin external surface has been
entirely transformed to Cu3Sn5 and/or Cu3Sn. The micro-
graphs shown in Figs. 6.39 and 6.40 clearly show that the
dark-grey bands of Cu3Sn exist on signicant areas of the
external conductor strand surfaces. Where intermetallic
compounds are exposed they become oxidized. It is probably
not tin oxide, but this oxidized intermetallic compound results
in poor solderability (Davy and Skold 1985; Warwick and
Muckett 1983).
374 6 Failure Analysis of Electrical Interconnections and Recommended Processes

Fig. 6.38 Tin-plated wire

strands after solderability testing
(dip-test method), a as-received
condition (100 %); b after 1 h
steam ageing (7595 %); c after
24 h steam ageing (530 %);
d after 16 h at 155 C (0 %)

Conclusions (B) The reduced solderability of tin-plated conductors is

(A) The solderability of tin-plated copper strands is severely
degraded by the exposure of such conductors to standard
accelerated ageing tests. Difculties may be expected
during the soldering of aged tin-plated conductors with
cored eutectic tinlead solder containing nonactivated
flux. Similar solderability problems were not experi-
enced when silver-plated copper strands were tested.
considered to result from the natural growth of an
intermetallic layer which consumes the free tin and,
when oxidized, is almost impossible to solder.
(C) It is recommended that spacecraft wires and cables that
are expected to be subjected to long shelf-lives, or
rework/repair operations after exposure to elevated
temperatures, should be fabricated from silver-plated
copper conductors.
6.8 Wires and Cables 375

Fig. 6.39 Micrographs of

tin-plated strands. Typical
locations on microsectioned
wires: a as-received; b after 16 h
at 155 C (dry-heat exposure).
The tin plating varies from 0 to
1.8 m, but is generally about
1.0 m thick. It is possible that
some free tin exists on the
as-received samples, but
approximately 10 % of thickness
is composed of Cu3Sn (dark-grey
band). After ageing, most of the
layer has converted to Cu3Sn
(polished sections were lightly
etched in ammonium peroxide
and examined under oil
immersion lens)

6.8.3 Red Plague Corrosion
of Silver-Plated Copper,
and Plagues on Other Plated
Stranded Wires
Two potential corrosion problems were recognized to be
associated with the use of silver-plated copper conductors in
the 1960s. Because of their characteristic corrosion products,
they are referred to throughout the electronics industry as
red plague and green plague. Red plague (see Figs. 6.41 non-solderable deposit occasionally found on the outer
and 6.42), is the cuprous oxide (possibly with some black
cupric oxide) corrosion product that forms when a galvanic
cell is formed between copper and silver. Green plague is
the reaction product of some resin-based solder fluxes with
copper oxide; it has been found on tin-plated and
silver-plated conductors, and is considered to be an unde-
sirable cosmetic defect.
In recent years, the term white plague (or fluorine
attack) has been coined to account for the white,
376 6 Failure Analysis of Electrical Interconnections and Recommended Processes
Fig. 6.40 Micrograph detailing
some strand drawing defects.
Because of the abrasive nature of
tin oxide the strand dimensional
tolerances are degraded and
subsurface longitudinal ssures
are partly closed embedding
islands of tin. The dry-heat
exposure has converted most tin
to Cu3Sn intermetallic
strands of nickel-plated copper conductors. White plague
is only found on fluoropolymer-insulated electrical wires and
cables having nickel plated conductors. It is thought to occur
during the manufacturing of wires and cables. In-coming
inspection of nickel-plated stranded wires should include
visual checks for this condition, if found it can be construed
as reject condition. During the high-temperature extrusion of
fluorocarbon resin onto stranded wires some gaseous mate-
rials outgas, these include carbonyl-difluoride (COF2), an
extremely reactive compound which, in the presence of
water and the nickel surface forms an adherent, porous white
deposit. [The recent usage of this term should not be con-
fused with the same name, attributed by component engi-
neers during the wire bonding in microelectronics, see
Sect. 6.3here gold micro-wires bonded to aluminium form
two intermetallics: Au5Al2 (white plague) and AuAl2 (purple
plague)].
Summaries of the mechanism by which red plague occurs
are given by (Dunn 1984) and Cooke (2010).
A number of reports including the work by Anthony and
Brown (1965) were written in the 1960s, which established
that red plague originates at breaks in the silver plating of
copper wire strands in the presence of moisture and oxygen.
Although some authors considered the problem to be per-
sistent, uncontrolled, and unsolved, others presented solu-
tions, thought likely to avoid the problem. A NEMA (1972)
investigation reported that out of a total USA production of
18 to 20 109 ft of silver-plated copper wire, only 105 ft (i.e.
0.003 %) had a signicant amount of red plague, and that
no malfunctions could be attributed to this phenomenon.
Several investigative activities were undertaken in Europe
(Dunn et al. 1984) in the mid-1970s. These included labo-
ratory evaluations, a European red plague survey, and
eventually the auditing of wire manufacturers premises. It
was concluded that silver-plated copper conductors would
be suitable for general spacecraft usage, but it was essential
that the platings be defect-free and of uniform thickness (at
least 2 m, and preferably 4 m). It was considered that red
plague could be avoided if the wire manufacturer and the
user companies implemented certain mandatory controls.
The causes and the effect of red plague can be sum-
marized as follows:
(1) Imperfections ranging from porosity to deep scrapes pen-
etrate the silver plating and reveal the copper wire of a wire
strand. Such damage may result from the use of defective
wiredrawing dies, wear action between strands during
stranding and braiding operations, or when conductor
strands are wound and rewound from reels to bobbins.
(2) Moisture may penetrate beneath the insulation of wires
and cables during the wire fabrication process, or
during insulation testing, or from the humidity in air.
(3) The presence of moisture in the region of exposed
copper can produce an electrolytic cell, the copper
tending to corrode and protect its more noble silver
plating. The following reactions were proposed:
6.8 Wires and Cables
Fig. 6.41 a Characteristic red corrosion products red plague on silver
plated copper strands of AWG 26 single wire (after removal of
insulation), and b, microsection through one strand showing galvanic
corrosion between small (arrowed) defect in silver plate. All corrosion
products appear as Cu2O, conrmed by having a bright orange
colouration under polarized light. This wire conforms to ASTM B
298-07 requiring not less than 1 m silver average, and no porosity as
377
determined by the sodium polysulde test (Dunn not published).
c Despite the controls placed on procurement of silver plated wires it is
essential to avoid damaging the plating as revealed in this illustration of
flight equipment (Cooke 2010). The insert shows stranded, silver-coated
copper wire with varying stages of red plague, damaged plating and
exposed copper (photo credit NASA Johnson Space Center)
378 6 Failure Analysis of Electrical Interconnections and Recommended Processes

Fig. 6.42 a Red plague

corrosion associated with
defective silver-plated copper
braiding within spacecraft (4 mm
diameter). b Section through X-X,
the superimposed numbers
referring to corroded strands.
Three of the 143 strands are
corroded to some extent (i.e.
2.1 %). c Detail of strand no. 1.
Adjacent strands were
uncorroded, but support crystals
of cuprous oxide, d one strand in
the centre of the braid has been
totally transformed to a silver
tube. Cuprous oxide (visible as
bright red under polarized light) is
located within the silver tube and
within all the interstrand spaces.
e Tunnelling corrosion along
strand no. 3

At the anode (i.e. exposed copper): 4Cu 4OH
2Cu2 O 2H2 O 4e
At the cathode (i.e. silver plating): O2 4H
4e 2H2 O
The sum reaction is: 4Cu O2 4H 4OH
2Cu2 O 4H2 O
Note that the water is not consumed, and in the presence of
oxygen the galvanic copper corrosion can proceed indenitely.
(4) Copper hydroxides are soluble in water and will con-
taminate the water that surrounds the stranded wires.
As the water dries out, the cuprous oxide crystals,
which have a characteristic red appearance, precipitate
out onto the silver plating.
(5) Excessive corrosion can impair the wires fatigue life
and electrical conductivity.
Auger (AES) surface analyses were made on the
silver-plated strands from 51 different wires and cables samples,
each originating from European space companies stock rooms
(Dunn 1984). The same samples were submitted to the accel-
erated corrosion test. All surfaces supported varying amounts of
chlorine (up to 8 at.%) and fluorine (to 4.8 at.%). The influence
6.8 Wires and Cables
of Cl and F on the corrosion rate is not understood; the
chloride reaction with silver-plate will form AgCl, which is
insoluble in water and can be expected to coat the surface with a
protective lm. Fluorine however, forms silver fluorides, AgF,
AgF2 and Ag2F and these are all highly soluble in water. This
contaminant has evolved from the fluoropolymer-insulation
surrounding these wire strands and its presence is expected to
accelerate corrosion. Due to variations in the quality and
thickness of their silver coatings it was difcult to establish a
relationship between halide surface contamination and the
preponderance for red plague. The involvement of fluoride in
the formation of red plague is still uncertain and continues to be
studied (Kauffmann and Wolf 2013).
Although the incidence of red plague corrosion of
silver-plated conductors in service is extremely rare, it is
emphasized that for high-reliability space applications, such
as those involving manned vehicles, it is necessary to control
all steps in the design, handling, and processing of wires and
cables constructed with such conductors. Red plagued wires
were found on the ESA Spacelab after service [see Sect. 8.7.
3 and the publication by Dunn and Stanyon (1997)] and
more recent damage, resulting in red plague was observed on
other space equipment (see Fig. 6.41c).
It is recommended that the basic hot-rolled and redrawn
copper rods are submitted to verication tests such as
chemical composition, electrical resistivity, and absence of
surface and internal defects before silver plating (Pops 1985).
After drawing of the silver-plated rods it is important that all
stranding and braiding machines must be controlled by wire
and cable manufacturers to ensure that they do not damage
the continuity of the 2 m silver plating. If this is an area of
concern, it would be expedient to produce strands to a greater
nominal silver thickness (possibly 34 m). Humidity con-
trol for cable reels (spools) can be achieved by packaging in a
sealed transparent plastic bag (polyvinyl chloride is usually
prohibited). Desiccant or dry-nitrogen lling should be used
in the packaging. An easy means for indicating that the
humidity content is below 50 %RH should be provided,
capable of being read without opening the package. Finally,
an accelerated corrosion test can be used as a screening
method for silver-plated wires, particularly for very critical
applications, such as the test proposed by Anthony and
Brown (1965) and adapted in the European specication
ECSS-Q-ST-70-20 Test procedure to determine the suscep-
tibility of plated copper wire/cable to red plague corrosion.
The NASA Red Plague Control Plan (JSC 64647) and
NASA STD-6012 (2012) include many of the remarks made
earlier, mitigation
only use qualied suppliers, ensure traceability and cer-
tied test reports, recommend shelf life of 10 years.
(Although, fully traceable wires/cables stored in dry
379
nitrogen with active desiccant should have an unlimited
shelf-life).
control the environment by water-vapour-proof
packaging/bagging during shipping and storage, use
humidity indicators in bags, assembly area shall be
environmentally controlled, aqueous solvents shall not be
used for cleaning.
6.8.4 Manganin Wire
Manganin is an interesting alloy. The French physicist Emile
Hilaire Amagat (18411915) worked on high ranges of
pressureone of his earliest inventions was the
self-tightening bolt which allowed pressures to be devel-
oped and maintained far in excess of those reached before. He
studied pressure-measuring devices and in 1911 reported the
design of an electrical resistance gauge which enabled him to
measure pressures up to 1270 MPa. The secret is now known
to be based on one of the rst uses of the alloy Manganin. The
electrical resistance of Manganin was found to be a linear
function of pressure up to 1180 MPaand by extrapolation,
pressure measurements could be made up to 2010 MPa.
Manganin is a coppermanganesenickel alloy possessing
a high electrical resistance and a low temperature coefcient
of resistance (0.00000 at 25 C). The original alloy has a
composition of 84Cu 12Mn 4Ni, but this specication for-
mulation has been slightly modied in recent years so that the
temperature coefcient could be adapted to suit equipment
operating at elevated temperatures. Manganin wires have
found many applications on board spacecraft in the con-
struction of precision resistors and in the electrical measuring
and control units for space experiments. The Tyco company
produce a NASA flight qualied (2015) shielded ribbon cable
consisting of 24 twisted pairs of 40 AWG manganin wire.
These shielded light-weight interconnections are designed to
operate between 273 C and +200 C, mainly for spacecraft
Cryogenic Telescope instruments. The wires can be soldered
or welded into circuits, but any bending operations will cause
the alloy to work-harden; this produces good spring proper-
ties but a marked reduction in electrical stability. However,
any cold work can be removed by an annealing heat treat-
ment, in an inert atmosphere, at 300 C for 2 h. The
microstructure of properly heat-treated Manganin possesses
equiaxed grains where the addition elements are in stable
solid solution. All processing steps associated with the
assembly of Manganin wire must be carefully performed, as
the alloy is susceptible to selective oxidation and galvanic
corrosion. These effects can result in manganese-depleted
regions along the wire causing a greatly increased tempera-
ture coefcient of resistance.
380 6 Failure Analysis of Electrical Interconnections and Recommended Processes
6.8.5 High-Voltage Wires, Cables,
and Connections
Spacecraft power systems have continuously increased out-
put from a few watts in the 1960s to the several kilowatts
demanded by the present space-based activities. As the
power levels increase, so the transmission voltages will have
to be increased in order to decrease currents and so minimize
the conductors weight and resistance losses. For instance,
some travelling wave tube ampliers (TWTAs) are now
operating at 15 kV.
The International Space Station requires even higher
electrical loads. Here the photovoltaic electric power gen-
eration solar arrays supply about 83 kW. The wires and
cables for electrical transmission and distribution systems
must be shielded both physically and electromagnetically to
prevent coupling of high-voltage pulses and transients into
low-voltage circuitry on-board the spacecraft. Most impor-
tant, they must also be able to operate under high vacuum.
From a vacuum point of view, a high-voltage cable is a
poorly vented component. It may have a length of many
meters. In manned spacecraft the problem might be how to
prevent condensation of water vapour onto the conductors
when there are changes to the cabin temperature (note the
red plague corrosion phenomena of Sect. 6.8.3) but, when
launched into a vacuum, there needs to be sufcient time for
air to completely vent from the cable and for the cable
materials to outgas before switching on any power supplies.
If such practice is not applied, there can be partial discharges
and corona effects that will damage the cable.
To ensure proper venting it is particularly important that
wire and cable ends are not blocked by potting compounds.
These have given problems on certain telecommunication
satellites when wires were soldered onto the pins of
high-voltage connectors. The connectors consisted of Kovar
pins set into glass-to-metal seals that had been welded into
the housing of a TWTA. The hermetically sealed housing
was itself lled with sulphur hexafluoride gas, the most
common gas used as a dielectric (having a very low break-
down strength, good heat transfer properties, and chemical
stability). The vacuum side of the connector consisted of ve
high-voltage, stranded wires, each partially stripped of
insulation and soldered to their respective connector pins.
Solder flux had been cleaned from the joints, with the pos-
sible ingress of alcohol and water beneath the insulation
jacket. The opposite ends of the wires were also soldered to
connectors, then clean and protected with heat-shrinkable
sleeves. The TWTA connector was then back-potted with a
hard compound that essentially caused the internal volume of
the wires to become airtight. This caused the inside pressure
of the wires to be uncontrolled and, during early operations in
orbit, is thought to have been the cause of partial electrical
discharges, or flashovers at this side of the connector. Other
suggestions for this problem were that the potting had not
been degassed prior to pouring into the connector shell and
might have contained small bubbles that caused the partial
discharge activity, or that the wires were not clean and flux
residues might have reduced the dielectric properties of the
compound. In discussions with experts, it would appear that a
well-designed high-voltage connector does not require pot-
ting for its high-voltage withstand capability. With an open
construction, space vacuum is the best high-voltage insula-
tion, and is the lowest weight insulation! (Wetzer and
Wouters 1993, 1994; Haq Qureshi and Dayton 1995).
High-voltage connections where corona suppression is
necessary will require special designs. For instance, all
aspects of soldered joints should be covered by smooth llets,
free of discontinuities or severe changes in contour, i.e. sharp
edges and points must be avoided (ECSS-Q-ST-70-08).
6.8.6 Cold Welding of Stranded Wires
and Cables
Concerns were raised when a band cable for an optical
terminal was under development. Some difculties were
noticed after mechanical testing during which the cable had
been rotating through 200 in a vacuum environment. The
cable became stiff and the bending was not smooth. Cold
welding of the individual silver-plated strands within insu-
lated wires of the cable was suspected and a detailed labo-
ratory test initiated (Fig. 6.43). Anomalies associated with
spacecraft Moving Mechanical Assemblies have been
addressed by Pirnack (2012) and suggestions have been
made how to avoid this problem. However, if magnetic
materials can be accepted on-board spacecraft, it is the wri-
ters opinion that for a given design, nickel-nished strands
will be far more resilient to cold-welding than silver nishes.
6.9 Problems Associated
with Soldering Fluxes
6.9.1 Purpose of a Flux
The importance of a flux in soldering operations is usually
underestimated. For conventional soldering the use of a flux
is imperative, and when correctly formulated it will:
(a) form the initial heat bridge between the heat source and
the part;
(b) react with, and either modify or remove, surface oxides,
passivation lms, and contamination lms;
(c) eliminate air from the liquid solder front because, rst, it
wets, and, second, it spreads by surface tension forces.
In general the space-quality specications limit the choice
of material nishes which may be interconnected to those
6.9 Problems Associated with Soldering Fluxes
Fig. 6.43 SEM images of single
strands of wire (Alloy 135, silver
plated): a before wire 200 bend
testing shows some small contact
marks. b After mechanical bend
testing, 15,000 cycles between
100 and +100 at ambient
temperature, under vacuum,
distinct cold welding marks are
visible and some strands remain
cold welded. c When pulled apart,
the silver plating detaches from
the stands and EDX-distribution
reveals exposed copper (d)
having an excellent solderability, so that highly activated
fluxes are not needed during the actual assembly process.
Fluxes of high activity, which are potentially more dangerous
from a corrosion viewpoint, are permitted during the initial
pretinning of difcult component leadsparticularly in the behave like an elastomer, and particularly it will not melt
case of nickel-based alloysto achieve better solderability.
After pretinning, the flux residues must be thoroughly cleaned
from the component lead surface to preclude time-dependent
corrosive attack. Similarly, the nonactivated rosin-type fluxes
which are permitted for component assembly and wire harness
manufacture must be removed by special cleaning procedures
immediately after the soldered joint has been made. Proper
cleaning is considered essential in order to avoid spacecraft
problems associated with the effects of flux residue contami-
nation, such as outgassing, corona effects, and both
stress-corrosion cracking and galvanic corrosion. These
effects will be overviewed in the following sections.
6.9.2 Heat-Shrinkable Sleeves Containing
Solder Preforms
A special product designed for joining wires and cables by
spliced joints has been commercially available for several
381
years (Parker 1977). Essentially, this product consists of an
outer length of plastic tubing that has been extruded to a
desired nal diameter. The tube is then irradiated to crosslink
the plastics molecular chains, and this will enable it to
when heated. The tube is next warmed and mechanically
widened to an enlarged diameter. A preformed ring of solder
containing a resin-based flux is inserted into the plastic tube.
Two small plastic sealing rings are also pressed into the ends
of the tube. Wires, which have been stripped of their insu-
lation, can be inserted into the tube with the intention of
making lap-joints. By heating the tube with hot air, or an
infrared gun, the sealing rings melt and flow around the
insulated wire, the tube shrinks, and after the flux has flowed
over the wire strand surfaces the solder alloy melts and flows
to interconnect the wire strands. On cooling, the solder joint
solidies and the flux and its residues are entrapped within
the outer sleeve.
Visual inspection of this kind of device after soldering
is difcult because of the opaque nature of the sleeve
material and masking by the retained flux. For this reason it
is essential to assess the quality of these interconnections by
making metallographic examinations on a periodic basis.
The acceptability of the joint is generally based on wetting
382 6 Failure Analysis of Electrical Interconnections and Recommended Processes
angles and flow of solder, and the method of assembly is
suitable for commercial and aerospace applications. These
devices have been thoroughly evaluated for spacecraft
applications (Premat 1977; Dunn 1978). It was found that
the outer sleeve is only a partial hermetic seal, and that the
outgassing of flux from each assembly (under vacuum con-
ditions) will represent between 30 and 60 % of the initial
weight of the flux. Some additional flux outgassing data are
given in Table 6.4. The problems associated with outgassing
and condensation of volatile materials were outlined in
Chap. 2. In this wire-joining application, under space con-
ditions, gas pockets under the wire insulation and outer
sleeve can give rise to the additional danger of corona,
especially when chlorides in the vaporized flux ionize and
have a lower dielectric constant than the insulant material.
Fig. 6.44 To avoid the inclusion of soldering flux within these
devices, an attempt was made to employ a fluxless solder preform.
Visual inspection of the resulting joints had given acceptable results,
a Metallographic inspections showed that the molten solder was
An attempt was made to make soldered interconnections
with the same device but with the incorporation of solid
solder preforms which did not support or contain any kind of
flux. From visual inspections the resulting joints were ini-
tially believed to be of reasonable quality. It was not until the
sleeves had been removed, and microsections made, that the
majority of joints were seen to be totally unacceptable owing
to nonwetting. Figure 6.44 shows how the solder alloy has
been pressed around the silver-plated wire braid (normally
this nish possesses an excellent solderability), but no
metallurgical joint has been achieved. These devices have
been prohibited for European space projects
(ECSS-Q-ST-70-08) but are permitted by NASA for
attaching grounding wires to shields and for wire splicing
(NASA-STD-8739.4 (2011).
pressed around the braid but no metallurgical bonding had taken place
(arrowed). b Enlargement of area arrowed in (a). c Detail reveals an
open electrical connection and lack of alloy formation between the tin
lead eutectic solder and the silver-plated braid
6.9 Problems Associated with Soldering Fluxes
6.9.3 Stress Corrosion of Component Lead
Material
Notwithstanding the tight soldering process controls applied
by space projects, several problems associated with the
corrosive nature of liquid-soldering fluxes and their residues
have cost much wasted time and effort. They include the
failure of various gold-plated Kovar leads on flat-packaged
components. These leads had been formed and then prepared
for gold removal prior to soldering. Degolding and pretin-
ning were performed by dipping the leads into solder baths
fluxed with a mildly activated rosin. After a few months of
storage, the leads from several batches of flat packages were
observed to fracture completely when exposed to light
handling operations. Metallographic investigations, as
shown in Fig. 6.45, strongly indicated that the failure
mechanism of these leads was one of stress-corrosion
cracking of the Kovar alloy due to the combined effect of
residual stresses in the lead material after the forming
operation and the presence of a thin surface lm of sup-
posedly noncorrosive flux residue. Other failures have
included the fracture of mounted transistor leads after
equipment level vibration testing; fractography revealed that
stress corrosion had initiated a crack in the lead material, and
this had later propagated by a fatigue mechanism.
An extensive test programme was initiated to assess the
corrosive effects of soldering fluxes then used by the
majority of contractors engaged in the manufacture of
electronic hardware for European spacecraft. Seventy-one
samples of flux were received and tested (Dunn and Chan-
dler 1980) by the use of standard and ad hoc corrosion tests
and the determination of flux halide content and pH values.
The corrosion tests were performed in a warm, damp
atmosphere for 56 days and after deliberate contamination
by the fluxes of samples of electronic materials.
Probably the most revealing results to have been pro-
duced by this evaluation programme are those which indi-
cated the low stress-corrosion cracking resistance of
gold-plated Kovar leads when stressed close to this mate-
rials yield point and then subjected to the soldering fluxes
under evaluation. Cracks were seen to initiate both from
under deliberately scratched regions and from below what is
now considered to be the porous gold plating.
All samples of components were subjected to metallo-
graphic inspection of their longitudinally cross-sectioned
leads. The test conguration and some typical photomicro-
graphs are presented in Fig. 6.46. The extensive degree of
general and stress-corrosion cracking (SCC) shown by these
longitudinal sections is quite surprising because close visual
inspection of the lead surfaces does not always indicate any
sign of surface corrosion.
383
Both the scratched areas and the porous gold plating are
capable of initiating and sustaining SCC growth by diffusion
of chloride ions from the residual flux to the crack tip. The
electrochemical mechanism of SCC in alloy steels is accel-
erated by the application of anodic currents, and a similar
situation is thought to exist in the case of these gold-plated
Kovar leads, anodic currents being set up between the por-
ous plating and the less noble Kovar alloy. It is to be noted
that the corrosion products occupy a larger volume than the
Kovar from which they are formed, and these products will
tend to increase the SCC propagation rate by creating a
wedging action and additional stress concentrations at the
crack tip.
The fracture of Kovar alloy as a result of exposure to
SCC environmental conditions has not been widely reported
(Weirick 1975). Effective protection of Kovar component
leads from SCC has been achieved by rstly chemically
removing the work-damaged (from lead-stamping opera-
tions) surface of the lead before thin gold plating. The thin
plate is designed to protect the lead from oxidation during
component manufacture, but is nally removed by
dip-coating with a ductile pore-free nish of eutectic tin
lead solder, and any gold plate remaining adjacent to the
component-to-lead glass-to-metal seal is additionally pro-
tected by a silicone varnish (Elking and Hughes 1969).
Another method to minimize the SCC failure of leads is
to plate the Kovar with 12.5 m of nickel before gold
plating. However, even this process will not overcome
chloride-assisted SCC if the nickel mechanically cracked
during either component manufacture or subsequent
lead-forming operations (see also Sect. 4.10).
The programme of flux evaluation was able to make
several conclusions and recommendations which could help
increase the reliability of not only spacecraft, but also
commercial electronic systems. These included the need to
solder with only nonactivated flux; the need for flux
removal; and the use of stress-relief bends in leads between
part termination and component body.
6.9.4 Flux-Corrosion of Silver-Plated
Stranded Wires
In a continuation of the work presented in the preceding
subsection an attempt was also made to assess the corrosive
effects of many soldering fluxes on silver-plated copper wire.
This was considered necessary because corrosion products
had been observed on satellite wire harnesses. These corrosion
products were of a bright-green and grey-brown nature, and
were not thought to be related to the phenomenon of red
plague which was discussed in Sect. 6.8.3. Metallographic
examination of many lengths of wire taken from the non-
conforming harness did not reveal reduced conductor
384 6 Failure Analysis of Electrical Interconnections and Recommended Processes

Fig. 6.45 Metallographic inspection of partially failed component corrosion product can be discerned as the grey volume (arrowed). It is
lead, a SEM photograph of part fractured electronic component lead. present within the crack path and tends to lift blister, (d) see colour
Note the absence of any plastic deformation. All features indicate that illustration section following p. 370. d Same section as shown in (c),
the Kovar-lead has failed by a stress-corrosion cracking mechanism. but after etching in mixed acids. The photomicrographs was taken with
b Detail of the crack and slightly lifted blister. The blister is surrounded an incidentlight interference contrast microscope. Such colour
by a white circle of soldering flux residue, identied by EDAX to enhancement can facilitate observations. The etching process has
contain the element chlorine, c this as-polished longitudinal micro- removed all corrosion products from beneath blister. The component
section of the component lead passes through line X-X on (a). The lead is gold plated and crack propagation is transgranular

cross-sectional areas; it was assumed that in the short term
such discoloration results from surface corrosion and is only a
cosmetic defect. However, many spacecraft electronic boxes
and cable harnesses have long storage lives before launch.
Once in orbit, it is difcult to assess the long-term effect of this
corrosion mechanism on electronic conductors within pres-
surized systems which have been designed to operate for 10 to
40 years (viz. the International Space Station).
To enhance any mode of wire degradation the lengths of
wires which were submitted to the test programme were
scratched so as to expose the underlying copper core. Each
wire was covered by a controlled amount of the standard
commercial soldering flux under investigation and then
submitted to a long-term damp heat exposure. A majority of
the fluxes caused extensive surface discoloration of the wires
after the total 52-day test period. Metallographic
6.9 Problems Associated with Soldering Fluxes 385

Fig. 6.46 Effect of stress in the

presence of soldering fluxes on
the reliability of component leads
(from Dunn and Chandler 1980).
a Test conguration for possible
corrosion of Kovar component
leads under constant deformation
(S = Location of standard
scratch). The inside width of the
containers was accurately
machined to produce a constant
tensile strain on the lead top
surfaces approaching that of the
Kovar materials yield strength.
Metallography had revealed that
these components leads were in
the fully annealed condition.
Mechanical testing of individual
leads provided the following
results: 29.2 kg/mm2 yield
strength; 43.4 kg/mm2 ultimate
tensile strength; 14 104 kg/mm2
Youngs modulus; 43 %
elongation at fracture; and
150 VPN microhardness.
b Extremely severe form of
stress-corrosion cracking through
the lead thickness. Here a single
crack propagates from a porous
location in the gold plating. The
commercial soldering flux which
had been deliberately applied to
this sample contained 0.78 %
halides. c The strongest flux to be
tested in this programme
contained 62.5 % halides and
caused catastrophic
stress-corrosion of lead within a
few days of application. The
section reveals highly branched
cracks emanating from the
standard scratch as well as several
other locations

examination showed that some of the strongly activated
fluxes did produce a limited amount of copper wasting. The
worst example is presented in Fig. 6.47. Here, 15 m of
copper have been removed from the 76 m diameter
strandbut it should be noted that this flux contained
62.5 %CI, which is approximately 280 times more chlorine
than in seawater! Rosin-based fluxes are extremely difcult
to analyse, and they can include a vast range of additives
known only to the flux manufacturers. However, a typi-
cal simple formulation (Weirick 1975) contains: 85 % abi-
etic acid, 12 % pimaric acid and 3 % other natural
compounds.
386 6 Failure Analysis of Electrical Interconnections and Recommended Processes

Fig. 6.47 The strongest and

most hygroscopic flux to be tested
produced severe galvanic
corrosion of the copper
conductors from these stranded
wires a. The SEM photograph
b of one strand shows where the
silver plating was scraped off
before testing; a small area of
copper (arrowed x) can still be
seen within the remaining silver
shell. On microsectioning c the
shell becomes crushed, but
extensive wasting of copper is
observed, a Optical photograph,
b SEM photograph,
c Photomicrograph after etching
in ammonium peroxide

85 % abietic acid 6.9.5 Selection of a Soldering Flux

12 % pimaric acid
3 % others
The green coloration seen after test exposures, and by
inference on the flight hardware, is believed to result from
the reaction product between abietic acid and copper oxide.
This is copper abiate, known to the electronics industry as
green plague, which can normally be safely washed away
during post-soldering cleaning (Peters and Wesling 1968).
As a precaution against galvanic corrosion only the pore-free
silver-plated wires described in Sect. 6.8.2 should be chosen,
and activated fluxes must be either totally forbidden or fully
assessed before used.
or a Solderable Finish
Solderability is the property of a workpiece surface which
permits it to form a good bond with a specic liquid solder
alloy at a specied temperature, and, usually, in the presence
of a liquid flux. The extent of wetting will depend on the
chemical nature of the surface being soldered and will be
facilitated by the presence of a flux in order to satisfy the
physical and chemical reactions that accompany solder flow
described in Sect. 6.9.1.
Fluxes may be applied as an external liquid, a coating, or
a predetermined volume incorporated in the core of a strand
of solder wire. The type of flux must be chosen to suit the
6.9 Problems Associated with Soldering Fluxes 387

Table 6.3 Recommended flux types for engineering metals

Group Metallurgical category EMF between a Flux type and possible protective nish with precautions
no. calomel electrode
and seawater (V)
1 Gold, solid or plated; goldplatinum alloys; +0.15 Both gold and platinum have excellent solderability.
wrought platinum Soldering to gold and its alloys should be avoided because
goldtin intermetallics embrittle joints. Gold platings can
be removed by solder-dipping (to dissolve gold), the
pre-tin in second solder pot. Rosin, nonactivated flux
adequate. Alternatively, use indiumlead solder alloys if
gold cannot be removed. Bright, hard gold platings may
be difcult to wet owing to certain alloying elements or
organic additives from plating solution (in that case, use
activated rosin flux)
2 Rhodium +0.05 Not easyinorganic flux
3 Silver, solid or plated on copper; high-silver 0 Easily soldered with mildly activated flux or, when free of
alloys surface sulphide, nonactivated rosin is preferable. If
chloride-contaminated (from plating bath), may need
abrasion. Use silver-loaded solder when joining to thin
silver plate. The silver-saturated liquid reduces danger of
scavenging. Silver-plated parts are not recommended for
electrical circuits owing to the problem of silver migration
and subsequent short-circuits
4(a) Nickel, solid or plated; Monel 0.15 Difcult to solder. Use inorganic or organic acid for
pretinning with solder and nonactivated rosin after
pretinning
4(b) Titanium 0.15 Impossible to solder. Can be copper plated
5(a) Copper, solid or plated; high-copper bronzes 0.20 With red oxide tarnish, can be soldered with mildly
activated rosin. Black oxide removable only with
activated rosin, organic acids, or zinc ammonium chloride
solutions (e.g. for radiator plates)
5(b) Coppernickel alloys; Nichrome alloys; 0.20 Very difcult to solder with inorganic acid fluxes, zinc
austenitic, high corrosion-resistant steels; chloride, and ammonium chloride solutions. Some
Nilo-K, Kovar, Monel, etc. proprietary brands are available. Can be nickel plated, but
deposit must be nonporous or substrate will become
oxidized by plating salts. All ionic matter, incl. handling
contamination, must be thoroughly and immediately
removed because this group is susceptible to stress
corrosion cracking
6 Commercial bronze 0.25 If thin tarnish lm quite easy to solder with mildly
activated resin. May be copper or silver plated, but ensure
good plating adhesion
7 Commercial brasses (6OCu40Zn, and 0.30 Difcult to solder. Use activated rosin if tarnish is thin.
70Cu30Zn) Impossible to soldereven with inorganic flux-if oxidation
is visible owing to surface lm of zinc oxide. Barrier
plating of more than 3 m nickel or copper recommended
for preserving solderability during shelf life (should
prevent zinc diffusion to surface). 5 m barrier necessary
if highly leaded brass
8 18 % chromium-type corrosion-resisting 0.35 See group 5(b)
steel
9 Tin-plated metals 0.45 New coatings are easily soldered with nonactivated rosin
flux. Activation of rosin depends on extent of tin oxide.
Fused tin preferred, as less porous. Pure tin coatings not
recommended for electronic applications owing to risk of
whisker growth. Should exceed 1 m thickness as
otherwise will react with copper and completely convert
to intermetallic, which is extremely difcult to solder
(continued)
388 6 Failure Analysis of Electrical Interconnections and Recommended Processes

Table 6.3 (continued)

Group Metallurgical category EMF between a Flux type and possible protective nish with precautions
no. calomel electrode
and seawater (V)
10 Tinlead, solid, plated, or fused 0.50 Most suitable nish for easy soldering with nonactivated
rosin flux. Porous platings need activated flux. Should
exceed 1 m thickness, but lead slows down intermetallic
formation. Good shelf life for fused coatings on copper
wire and printed circuits
11 Lead, solid or plated; high-lead alloys 0.55 Mildly activated flux required to penetrate surface oxides.
High dissolution of lead in tinlead solder produces joints
with extremely low shear strength
12 Aluminiumcopper alloys (e.g. Duralumin 0.60 Impossible with tinlead alloys. Generally, welding or
and most of AA2XXX Series) dip-brazing is more suitable. Plate with zincate and
copper. The low-melting solders are preferred to avoid
thermals stressing plating
13 Iron, wrought, grey, or Armco; plain carbon 0.70 Solderability depends on oxide thickness. Clean, pickled
and low-alloy steels surfaces can be easily soldered with mild nor activated
rosin flux. Passivated or phosphated steels require special
fluxes
14 Aluminium and most alloys other than in 0.75 Can be soldered, but with special alloys (e.g. tinzinc or
group 12 (e.g. AA1XXX, 3XXX and 5XXX cadmiumzinc alloys). May be friction-soldered without
series) flux, or, when aluminium oxide is removed, by
ultrasound. Some proprietary fluxes are available.
Corrosion in aluminium-soldered joints is particularly
troublesome; a waterproof coating is essential
15(a) Aluminium cast alloys other than 0.80 As for group 14, but choice of solder and flux requires a
silicon-type specialist advice
15(b) Cadmium platings 0.80 Easy to solder with mildly or fully activated rosin fluxes.
If passivated by chromate lm, they are very difcult and
require ammonium chloride-type flux. Fumes are toxic
16 Hot-dipped zinc plate 1.05 As for group 15(b)
17 Zinc, wrought; zinc-based casting alloys; 1.10 Generally, very difcult to solder owing to oxidation. Use
zinc plate inorganic acid or special proprietary flux. Water-tight
protection necessary to avoid corrosion
18 Magnesium and magnesium-based alloys, 1.60 Not recommended because of poor strength and corrosion
cast or wrought
Note For galvanic compatibility, the maximum potential difference between soldered metals should not exceed 0.5 V

form of tarnish existing on the workpiece surface, and the
joints should be designed so that the flux can be both
properly introduced to the joint area and then totally
removed after soldering. The various methods of soldering
which transmit heat energy to the workpiece (by means of
hand-held irons, torches, resistance soldering, oven or fur-
nace soldering, wave-soldering, etc.) may necessitate special
fluxes which avoid factors such as the evolution of corrosive
fumes during soldering; low flash-point constituents; irri-
tating or dermatitis-producing liquids and fumes.
When torch-soldering is used, a more active flux may be
necessary to deal with the extra tarnish formed during
heating by the flame. Table 6.3 may be used as a guide in the
selection of flux types for the soldering of common metals. It
is observed that the less noble metals listed in this tabulation
require either a more active fluxing system or a more sol-
derable nish, such as copper plating.
In recent years, it has been found cost-effective to prepare
surfaces for soldering with a dual metallic nish, particularly
when long storage periods are envisaged before the nal
assembly of spacecraft parts. In these cases, otherwise
difcult-to-solder surfaces are plated with copper or nickel,
and subsequently coated with a eutectic tinlead composi-
tion which can be fused in hot oil to form a 412 m thick
pore-free protective layer possessing excellent solderability
even after many years of exposure to industrial atmo-
spheres. The choice of highly solderable nishes precludes
the need for strongly activated fluxes which are often dif-
cult to remove in post-soldering cleaning operations and
may later lead to extensive corrosion of the less noble metals
in any joint design.
The chemical composition of strong fluxes may be based
on aqueous solutions of halides, such as ammonium and zinc
chlorides or orthophosphoric acid; these are widely employed
6.9 Problems Associated with Soldering Fluxes
in general engineering. For use on electronic assemblies,
rosin-based fluxesoften in the form of alcoholic solutions
are chosen. A range of rosin types are commonly used,
including nonactivated halide-free, organically activated, and
halide-activated rosins. A multitude of commercially available
fluxes are on the market. These can include a vast range of
additives known only to the flux manufacturers.
6.9.6 Control of Galvanic Corrosion
As a general design rule, it is desirable to avoid joining
together metals which have greatly separated positions on
the galvanic scale (see also Sect. 2.2). Table 6.3 include data
of the relative activity of metals in relation to a standard
reference electrode based on a seawater environment. Many
authorities recommend that, to avoid corrosion couples,
metallic combinations which will be subjected to uncon-
trolled terrestrial environments do not exceed a potential
difference of 0.25 V. When the working environment is
controlled, e.g. in the case of electrical units operating under
spacecraft assembly cleanroom conditions, compatible
corrosion-free couples may be separated by potential dif-
ferences not exceeding 0.5 V.
It is fortunate that tinlead, the most common soldering
alloy (Group 10 of Table 6.3), is generally compatible with
most solderable metals, with the important exclusion of gold.
The chart showing the galvanic series in Table 2.1 is useful
for a rst approximation in selecting materials for both
solderability and corrosion control, but for many applica-
tions it may be too simplistic because it does not provide
information about corrosion rates or changes in surface
chemistry which may pacify or accelerate corrosion at
bimetallic interfaces.
6.9.7 Cleaning of Flux-Contaminated
Surfaces
After soldering, the flux, dross inhibitors, solder resists, and
any oils must be promptly removed from the joint in a
manner which does not change the hardware. A clean sol-
vent is recommended for the removal of all residues even
though the flux manufacturer may have classied the par-
ticular flux, thinner, oil, etc. as being noncorrosive. If flux
residues are to be left on soldered surfaces in pressurized
electronic boxes, it is essential to perform environmental
tests to ensure that such residues do not cause surface cor-
rosion or stress-corrosion cracking, thereby lowering the
integrity of the part. Flux contamination existing beneath
surface mount devices is discussed in Sect. 6.15.9.
The spacecraft EURECA was launched from the space
shuttle Atlantis in July 1992. It remained in orbit for 286 days
389
and was retrieved by the shuttle Endeavour. During this
mission, all objectives were achieved but several problems
and anomalies were observed during the course of the mis-
sion (see also Chap. 8). One concerned the failure of Gyro
No. 3examination of this mechanism when it returned to
Earth identied the failure to be due to an open phase of the
gyros wheel motor (Kugel et al. 1995). The harness to this
motor is made of very thin tinned copper flexible wiring.
Inspection showed one wire to have completely broken and
others were abnormally discoloured. Metallurgical analyses
found excessive amounts of flux residues to be adhering to
the wires and SEM images identied that acidic chlorides had
caused corrosion to all wires and resulted in the
failure/rupture of one. This is a clear lesson that flux residues
must be carefully removed after soldering.
Flux residues must always be removed from systems
exposed to vacuum, as it has been shown that rosin flux
consists of substances nearly all of which have appreciable
volatility at typical equipment operating temperatures (Sin-
clair et al. 1985). In some instances it may be possible to use
a new ultrasonic soldering technique which does not require
the presence of flux. The method is not suitable for
through-hole or surface mount assembly of components to
PCBs where it is the capillary flow of solder (preceded by
the flow of flux) that is critical in forming good connections.
Also, some authorities consider that ultrasonic vibrations can
damage the internal microwire bonds within active compo-
nent packages. However, ultrasonic soldering, whereby
ultrasonic energy is introduced into a bath containing liquid
solder, will have many applications, as oxides are mechan-
ically removed from a submerged workpiece by the forces of
imploding cavitation bubbles. Component leads can be
effectively degolded and pre-tinned. Wire strands can be
pre-tinned and soldered without the need for fluxes that
could be wicked under the wire insulation (so no heating
sinking of the wire is needed to prevent wicking). The
molten solder temperature can also be lower for ultrasonic
soldering than when a flux is used and this causes less dross
formation on the solder bath. The technique may also be
useful (Saxty 1995) for regenerating the solderability of
difcult items such as terminal pins, wire tags, and com-
ponent leads that have aged and would normally de-wet if
soldered by conventional ECSS-Q-ST-70-08 techniques.
A method adopted by NASA and ESA for selecting
materials is the micro-VCM test. Samples of the RMA and RA
fluxes were submitted to the volatile condensable material test.
Initially they were spread onto aluminium foil and cured at
room temperature for eight weeks and then, according to the
test procedure, the weight losses of their volatile condensable
products were evaluated after a thermal vacuum exposure of
the samples for 24 h at 125 C and 106 mm Hg.
The results given in Table 6.4 clearly indicate that these
flux residues are hazardous and can cause degradation of
390 6 Failure Analysis of Electrical Interconnections and Recommended Processes
Table 6.4 Outgassing data for flux residues
Sample Type TML/RML CVCM
(%) (%)
Kenco RMA, rosin based 48.2/48.0 12.0
365 containing 0.01 %
chloride
Multicore RA, rosin based 21.2/21.0 5.0
PC 21A containing 0.23 %
chloride
TML Total mass loss
RML Recovered mass loss
CVCM Collected volatile condensable material
Note Micro-VCM test per ECSS-Q-ST-70-02, A thermal vacuum test
for the screening of space materials
From Dunn and Bergendahl (1987)
spacecraft owing to contamination and other phenomena
such as corrosion and corona. They may also cause complete
failure of a satellite mission by recondensing between slip
rings, contact points (causing electrical open circuits), and
on optical or thermal control surfaces.
The solvent cleaning of flux-contaminated surfaces may be
performed by immersion cleaning, sometimes with ultrasonic
agitation (not for electronics which support small integrated
circuits, as the miniature wire bonds can be degraded). Solder
joints can be more readily inspected after flux residues are
removed. In those cases where inorganic or organic acid
fluxes have been employed, it is especially critical to remove
residual flux immediately after soldering and even before the
workpiece has cooled to room temperature. The nal rinse
baths may contain clean solvent or deionized water; they can
be controlled by electrical resistivity measurements made at
predetermined intervals based on the number and size of
articles being cleaned over a given period. Solvents should be
properly labelled, and those showing evidence of contami-
nation or decomposition should not be used.
The following solvents are acceptable for the cleaning of
electronic equipment provided that cleaning takes place soon
after the soldering operation (e.g. before the flux and resi-
dues have had time to polymerize or age).
(a) Ethyl alcohol, 99.5 or 95 % pure by volume.
(b) Isopropyl alcohol, 99 % pure.
(c) Any mixture of the above.
(d) Terpine.
(e) Any commercial product proven to be equal to or
better than the solvent CFC-113 regarding cleanliness
and reliability issues.
(f) Hydrofluorether, methoxynonafluorbutan [C4F9-O-CH3]
a clear, colourless and low-odour liquid has found some
acceptance in the aerospace and electronics industries.
(g) Deionized water at 40 C maximum may be used for
certain fluxes. Items shall be thoroughly dried directly
after the use of deionized water. However, with
increased miniaturisation water-based processes cannot
penetrate into extremely low stand-off recesses.
The Montreal Protocol established under the aegis of the
United Nations is now, in effect, to contain and reverse the
depletion of atmospheric ozone which, if not halted, is likely
to have a catastrophic effect on the Earths environment.
Chemicals such as the chlorofluorocarbons (CFCs) are most
extensively used as solvents in the electronics industry, but it
is now recognized that their release into the atmosphere is a
primary cause of the phenomenon of ozone depletion. Space
industries are no longer engaged in the production and use of
the CFCs including the item CFC-113, also known as tri-
chlorotrifluoroethane, CC12F-CC1F2.
Modern methods for the cleaning of mechanical parts
were addressed in Sect. 4.21. Some manufacturers of elec-
tronic equipments have considered the no clean and
low-residue solder flux routes after abandoning the CFC-113
solvent. Certainly some excellent products exist for com-
mercial and some military applications, as was evident from
a Soldering Task Force test programme and results (Iman
et al. 1995). However, the space community still insists on
aesthetically perfect assembled PCBs for proper conformed
coating adhesion, low outgassing, and no possibility for the
migration of tinlead as seen in Figs. 6.30 and 6.49. The
cleaning machine market appears buoyant as now there are
different machines on sale for each variety of solvent (an
example of a successful machine is seen in Fig. 4.85).
Periodic testing for cleanlinessgenerally to check for
the absence of ionic residuesof pcb assemblies after nal
cleaning (and prior to conformal coating or further integra-
tion of the assembly into electronic housings) is required by
all space standards, and if these tests fail, the entire batch of
assemblies will need to be re-cleaned and re-tested.
Ionic residue testing is based on the solvent extraction
method; a solution of isopropyl alcohol and deionized water
in the ratio 75/25 (volume/volume) is puried to a low
electrical conductivity (i.e. the base line) by pumping it
through an ion-exchange column in a closed-loop; the board
to be tested is introduced to the solution; ionic residues are
washed off the board leading to an increase in conductivity
(or decrease in resistivity) that is a measure of the amount of
ionic residues dissolved in the solution. The levels of con-
tamination are expressed as, micrograms/cm2 sodium
chloride (NaCl) equivalent ionic or ionisable flux residue.
This is a long established test method dating back to the
1970s and many similar instruments for conduction the test
are available on the market. They are seen in all component
assembly lines and include: the Contaminometer, Ion
Chaser, Ionograph and Omega Meter. An ESA report found
models of these instruments to be suitable for the
6.9 Problems Associated with Soldering Fluxes
determination of ionic contamination levels, but the amount
of contamination indicated by each instrument differed
slightly according to the type of flux under investigation
(Bergendahl and Dunn 1984). Equivalence factors for each
instrument were proposed. The maximum limit of surface
contamination required by IPC and ECSS standards is: less
than 1.56 micrograms/cm2 NaCl equivalent. Other bodies
require lower limits such as 1.3 or 1.0 g/cm2, and such
lower limits seem to be more appropriate for modern space
pcb assemblies due to their higher complexity, miniaturisa-
tion and smaller track spacings. Ionic contamination testing
has been evaluated and reviewed by Naisbitt (2015) with
some important ndings: during testing the solution can pick
up some ionic contaminants such as CO2 and other gases
present in the surrounding air (e.g. forming carbonic acid) so
it is necessary to keep a lid on the instruments sample
chamber; surface insulation resistance (SIR) testing is suit-
able for assessing whether residues (ionic and non-ionic)
might compromise circuit performance from an
electro-chemical standpoint; FTIR and ion chromatography
are effective at measuring the presence of ionic salts but they
are not capable of determining whether their presence will
compromise circuit reliability.
A nal topic which had not been adequately addressed
prior to the study by Zou et al. (1999) was: what duration of
time, and under what conditions of temperature, can
assembled (and repaired/modied) PCBs be left before they
are submitted to a cleaning process? This study was made
on bare boards, typically assembled boards with
through-hole components, and boards with surface mounted
components. Flux types were: pure rosin, mildly activated
flux (cored solder wire) and solder paste fluxall were
supplied by a preferred space product manufacturer. Results
showed the fluxes difcult to clean after 48 h storage at room
temperature, by simple immersion into IPA. It was recom-
mended that cleaning be performed within 48 h, preferably
within 24 h. Heating the boards to 80 and 120 C did not
have an adverse effect on the recommended period before
cleaning. The choice of cleaning fluid proved interesting.
A proprietary solvent cleaner out-performed other solvents
by a factor of two, for simple immersion cleaning. This is
particularly important for surface mount components since
brushing and scrubbing had little influence. For plated hole
components brushing and scrubbing was effective. Cleaning
efciency increased to the fourth power of the stand-off
height, but the cleaning of SM components presents more
problems than through-hole components.
These test results were considered by a panel for the
ECSS-Q-ST-70-08 revision who adopted the requirements
for:
(a) Flux and residue shall be removed within a maximum
period of 8 h after soldering operations.
391
(b) PCB assemblies shall not be immersed in cleaning
solvents for more than 30 min for each cleaning oper-
ation (as long immersion times can promote galvanic
corrosion between adjacent metallic surfaces).
ECSS-Q-ST-70-38 called for a stand-off of between 0.1
and 0.4 mm to aid cleaning and minimize failure due to
thermal cycling.
6.9.8 Flux Residues, Their Ingress
into Top-Coat of PCB Surfaces,
and Bake Out After Cleaning
The preceding section described how the space industries
have been searching for alternative solvent cleaners. The
recent use of surface-mounted components in spacecraft
systems requires additional considerations, as the majority of
soldering techniques necessitate the application of active
fluxes to ensure that adequate solder wetting is achieved and
the minute balls of solder within solder paste do not scatter
as they melt (ECSS-Q-ST-70-38).
Several studies have been made to assess the suitability of
test equipment for the detection of ionic contamination on
both conventional and surface-mounted component assem-
bled PCBs (Bergendahl and Dunn 1984; Dunn and Ber-
gendahl 1987; Tegehall and Dunn 2006). They have found
that the distribution of flux and flux residues under indi-
vidual components may be impossible to remove (or detect),
even when utilizing CFC solvents, unless adequate stand-off
heights are specied. At the present time it appears that the
CFC-substitute solvents selected by ESAs contractors can-
not adequately remove the various classes of fluxes
employed during electronic assembly. Within each class of
flux, the composition may vary considerably. Certain prod-
ucts with known trade names have been chemically analysed
and found to vary considerably in composition and acidity
from one batch to another; and even these fluxes can contain
contaminant particles (Dunn and Chandler 1980; Nasta et al.
1990). The 1980 study was remarkable because flux sup-
pliers agreed that in-coming materials were not adequately
analysed prior to them being formulated into deliverable
liquid flux products. More recently the author visited a
manufacturer of liquid flux, and it was clear that the natural
rosin was sourced from both Portugal and Chinano
chemical analyses were made on the incoming batches
(being produced from pine tree sap, they are certain to
possess different chemistries and acidity). These crystalline
solids were then chemically modied by thermal treatment,
addition of solvents and other chemicals that gave consistent
viscosity and halogen-activation.
392 6 Failure Analysis of Electrical Interconnections and Recommended Processes
Fig. 6.48 Scanning electron micrograph showing the surface of an
FR-4 PCB board laminate
Another complication related to cleaning flux from
assembled PCBs is certainly caused by the variability of
each PCBs topography. Under high magnication, epoxy
laminates possess a sponge-like surface containing microp-
orosity and interconnecting channels. An example is shown
in Fig. 6.48; this clearly illustrates how flux and other con-
taminants can be easily picked up by the board. Once in
place they are extremely difcult to remove and, in course of
time, may diffuse to the board surface and subsequently
affect properties such as surface insulation resistance (SIR).
The effect of various contaminants on current leakage were
reviewed by Tegehall and Dunn (2006) and it is clear that,
for all inorganic contaminants, there is a critical relative
humidity when the contaminant substance will cause leakage
currents on a PCB (as demonstrated using copper comb
patterns). Remarkable examples where ionic matter has
contributed to electrochemical migration (EMC) and a low
SIR are seen in Figs. 6.30, 6.49 and 6.50.
Concerning the dielectric properties of space boards: three
materials were evaluated, as-received, after exposure to
humidity of 90 %RH for 20 days, and nally after baking in air
125 C for 4 h [as part of a programme that included the ageing
of PCB laminates during exposure to ultraviolet radiation, and
to electron beam radiation (Bulletti and Dunn 2012)]. A short
resume of the humidity exposure results showed:
Surface Resistivity, SR,
Thermount: as received, 6.7 1016, after 90 %RH, 1.8 1015
Epoxy: as received, 1.0 1016, after 90 %RH, 3.5 1015
Polyimide: as received, 1.7 1016, after 90 %RH, 3.5 1015
Volume Resistivity, VR, cm
Thermount: as received, 1.3 1016, after 90 %RH, 1.5 1015
Epoxy: as received, 2.6 1016, after 90 %RH, 3.6 1015
Polyimide: as received, 3.3 1016, after 90 %RH, 8.2 1015
The nal bake restored both SR and VR values similar to the
as-received values.
The Thermount laminate appears to have a slightly
greater sensitivity to moisture than the epoxy- and
polyimide-glass woven laminates. This is might result from
Thermounts more porous surface (type is Arlon 85NT, a
unique combination of short aramid reinforcement bres in a
matrix of polyimide resin). However, it is clear that from an
electrical viewpoint, the baking of all assembled boards is
essential:
(a) after cleaning, or
(b) when exposed to a damp environment (e.g. if conden-
sation is present on electronic boxes withdrawn from
thermal test without given time to reach room
temperature).
Moisture ingress into, and bake-out from the organic
laminates used for modern hi-rel electronic assemblies was
also studied by Pecht et al. (1999). Their work included the
space-materials FR-4 epoxy, high temperature FR-4
(HT) and polyimide (PI). Diffusion coefcients were calcu-
lated based on moisture absorption and desorption as a
function of the square root of time. FR-4 and HTFR-4
boards were similar (0.73.3 cm2/s for 5085 C, 5085 %
RH) whereas polyimide diffusion coefcients ranged from
1.2 to 8.2 cm2/s for 5085 C, 5085 %RH. For real PCBs
the presence of copper will form barriers to moisture
movement, but in practice polyimide boards may be
expected to bake out faster than those fabricated from FR-4.
Drying to remove water absorbed by PCBs is essential
before initiating the soldering processes (i.e. to avoid
delamination of the board and damage to plated through
holes should steam be formed within the laminate during
the 200300 C soldering cycle).
A study into optimizing the bake-out times and temper-
atures for PCBs procured from ESA-approved manufactur-
ing lines (ECSS-Q-ST-70-11) gave consistent results for
bare boards (Charrier et al. 2003). Each supplied board was
washed clean to a standard, precise procedure in a mixture of
75 % isopropyl alcohol and 25 % distilled water. The
polyimide (PI) laminated boards were quickest to bake out at
temperatures of 105 and 120 C. The FR-4 epoxy boards
were signicantly slower to bake out. As might be expected,
moisture pick-up after baking showed the PI boards to gain
weight far quicker than the FR-4 boards. These results are in
agreement with those of Pecht (1999). The results indicated
that a reasonable bake out procedure before soldering would
be:
(a) Demoisturisation in an oven held between 90 and 120
C (but also take note of the board manufacturers
recommendations).
6.9 Problems Associated with Soldering Fluxes
Fig. 6.49 Electrochemical migration. a Terminal pin following wave
soldering to PCB and operation under 10 V in damp (not moisture)
conditions. The conductor spacing is only 0.15 mm and this illustrates
the danger of permitting, under waiver, values of less than 0.30 mm.
b Details the leadtin migration and accounts for a low SIR
(b) The process time shall be a minimum of 4 h (but note
that long baking periods at higher temperatures can
reduce the solderability of PCB terminations).
(c) A vacuum bake method can be selected for tempera-
tures less than 90 C.
(d) Demoisturising after 8 h in ambient conditions.
(e) Storage in dry nitrogen lled bags is recommended.
These parameters for bake out were introduced into
ECSS-Q-ST-70-08.
In situ optical observations and photographic recordings
have followed the initiation and growth of dendrites during
electrochemical migration (ECM). The work (Medgyes et al.
2011) is important because the tests were made under ther-
mal humidity biasthis is very similar to a spacecraft test
environment when electronic boxes are submitted to testing
393
after a thermal vacuum test. The electronic circuits can take a
long time to warm to ambient temperature if the vacuum
chamber is opened after a deep cooling cycle. Condensation
has occurred under these conditions, resulting in short cir-
cuits of the kind seen in Figs. 6.49, 6.50 and 6.51.
Care needs to be taken to differentiate between Electro-
chemical Migration (ECM) and Conductive Anodic Fila-
ment (CAF) formation. According to the IPC:
Electrochemical Migration (ECM) The growth of con-
ductive metal laments across or through a dielectric mate-
rial in the presence of moisture and under the influence of
voltage bias.
Conductive Anodic Filament (CAF) Formation The
growth of metallic conductive salt laments by means of an
electrochemical migration process involving the transport of
conductive chemistries across a non-metallic substrate under
the influence of an applied electric eld, thus producing
Conductive Anodic Filaments (see next section).
ECM is mainly associated with PCB surfaces. However,
it can also occur on the surfaces of electronic components
and conformal coatings (see Fig. 6.50). The miniaturization
of components and distance between conductors has
increased the propensity of EMCfor flip chip ICs,
miniaturization has amounted to about 90 % during the past
10 years. This closer spacing d causes an increase in
electric eld, E, which makes local corrosion cells more easy
to form under humid conditions, i.e. E = V/d. Ambat et al.
(2009) found single ceramic capacitors to fail due to severe
EMC because of the presence of dust and flux residues on
the ceramic surfaces (between the capacitors metallised
terminations).
Although not an ESA requirement, most spacecraft
electronic circuits are coated with a thin, electrically non-
conductive protective coating known as a conformal coat-
ing. Such coatings are intended to minimize the degradation
of assemblies by humidity, handling, and other contamina-
tion. Of particular importance is that the coating should help
to protect against a decrease in SIR when the assembled
circuit board is either under test or in its nal working
environment.
An investigation to assess the effect of residual fluxes on
spacecraft PCBs that were only partially cleaned and then
conformally coated was performed by subjecting such
boards to electrical tests both in air and under vacuum
(Tegehall and Dunn 1992). The residual contaminants from
the seven different flux-types (listed in Table 6.5) used in the
space industry were assessed by SIR. This is the only easy
way of forecasting a loss of reliability (Ellis 1996) and it is
strongly recommended that, for space applications, the
propensity for ECM is assessed by incorporating test
394 6 Failure Analysis of Electrical Interconnections and Recommended Processes
Fig. 6.50 Actual ECM failures from operating electronics. From top
left, SEM image of ECM (similar to Fig. 6.33, but here a component
lead is soldered to a nonfunctional pad). Ceramic chip with dendritic
patterns onto test boards to be soldered according to com-
pany process and cleaning standards (IPC-TM-650 2000).
The ndings from the test programme are depicted in
Fig. 6.51 and can be summarized as follows:
The surface structure of the epoxy top-coat present on both
sides of FR-4 PCB laminates makes them very difcult to
clean. Boards procured for space use should be thoroughly
cleaned prior to component assembly. Subsequent pro-
cesses such as soldering must avoid the use of chemicals or
soldering fluxes known to be hygroscopic in nature.
Some contaminants will inevitably penetrate into the
epoxy top-coat, lling pores and channels so as to affect
the properties of the epoxy material.
Residues from nonactivated or mildly activated rosin
fluxes have been seen to increase the SIR of PCBs
operating under humid or dry conditions. However, there
may be severe corrosion of the metallic conductors.
Residues from low-solid fluxes based on carboxylic acids
decrease the SIR. The level of contamination originating
growth progressing over dielectric and, leaded flat package showing
ECM over conformal coating. The electrical leakage paths are clearly
visible (Ring 2010)
from these fluxes, as examined in this investigation, was
very low (0.13 g/cm2 eq. NaCl). These, and the organic
acid fluxes, are unsuitable for space assemblies [poor
properties were also found by Sohn and Ray (1995)].
The application of conformal coatings to PCBs has been
shown to improve their SIR in air, especially at high
humidities. This does not necessarily mean that the reliability
will be enhanced. If there are hygroscopic contaminants
beneath a conformal coating, the presence of the coating may
have very detrimental effects (Schweigart 2007).
6.9.9 Conductive Anodic Filament
(CAF) Formation and Particulate
Contamination
Electrochemical migration (ECM) has been described in the
Sects. 6.6.3 and 6.9.8 and it must be noted that those
examples related to Mechanism 1, the growth of dendrites
6.9 Problems Associated with Soldering Fluxes
Fig. 6.51 Photographs (ac) showing appearance of some test boards
(comb patterns) after the environmental tests; to illustrate the effect of
flux residues on PCBs (bare and with various spacecraft conformal
coatings). a Elsold F-SW26. Low ionic content No corrosion, no SIR
problems, clear residue. (No SIR problem with any of the coated
boards). b Kester 1585. Medium to high ionic content. Conductor track
shows severe corrosion. Clear residues on PCB. No SIR problems. (No
SIR problem with any of the coated samples). c Alpha 850, High ionic
content, plus conformal coating type CV-1140. No visual residues, but
severe leadtin migration. Very low SIR. (Dendrites and short circuits
on the surface of a PCB, resulting from lead ions going into
solution at the anode (one solder-coated track) and plating
out at the cathode. These growth have different morpholo-
gies, ranging from fern- or tree-like formations (Fig. 6.30) to
the nodular deposits in Figs. 6.49 and 6.50. Mechanism 2
concerns the growth of conducting anodic laments
(CAF) as shown in Fig. 6.52. CAF forms within the
395
for bare boards and all coated boards). d Compilation of results from
environmental tests: exposure to temperature/humidity then vacuum,
with details shown on graph of insulation resistance against time in
hours. Registered surface insulation resistance (SIRMohm) values for
conformally coated test boards (all coating types gave similar results,
proving that water molecules are able to diffuse quickly through the
silicone coatings). The insert lists the ionic contamination levels on the
boards after poor cleaning (gures quoted are average values for three
test boards soldered using the same flux) (from Tegehall and Dunn
1992)
thickness of a PCB. CAF testing to the IPC method dened
in TM-650 helps to determine the reliability of a PCB
laminate material or a nished product. Incidents of CAF
failure are increasing because conductor spacing and overall
part sizes are becoming smaller and smaller, the greater need
for higher performance and an increased density of com-
ponents. Hence, the necessity for this test is increasing.
396 6 Failure Analysis of Electrical Interconnections and Recommended Processes

Table 6.5 Properties of fluxes used to contaminate test boards Work performed at the National Physics Laboratory, UK and
Flux Type a
Halide content pH % the Trace Laboratories in the US has provided industry with
(%) solid a greater insight of CAF failure modes.
Kester 145 R None 5.1 25.2 Reduced conductor spacing, small diameter vias and
Alpha 611 RMA 0.03 5.0 34.5 multiple plated through holes (some lled) on multiple levels
within a multilayer board are prime bridging paths for the
Elsold RMA 0.08 6.0 13.7
formation and growth of CAF. Whereas CAF short circuits
Multicore RA <0.5 4.2 17.4
PC29-17 typically form laterally between plated through holes, it can
occur vertically (Nuez and Tan 2014) between internal tracks;
Kester 1585 RA 0.75 4.8 36.4
from the remarkably detailed work of these analysts such
Interflux 2005 LS None 4.5 2.0
failures can easily be missed due to the physical characteristics
Alpha 85033 OA 5.0 1.0 30.8 of CAF, intermittent nature and recoverability of the short
RMA rosin mildly activated circuit. CAF can also be promoted by poor hole drilling that
RA rosin activated
results in cracks or delamination between the laminates glass
LA low solids, water-soluble
OA organic acids, water-soluble, only used in the space industry for bre bundles and the resin matrix. CAF can occur from
pre-tinning difcult to solder component leads/terminations hole-to-hole, hole-to-track, internal track-to-track, and from
a
R rosin pure nonactivated ground plane to track. The appearance of CAF is different from
dendritic growths in that these are lamentary growths that

Fig. 6.52 Compilation of optical

and SEM images to illustration
the appearance of CAF.
a Recorded the contamination
between PCB plated through
holes in FR-4 laminate as
determined by EDX, after testing
(Lee et al. 2006). b This optical
image of a microsection (made in
the horizontal plane, in the middle
of a PCB) reveals CAF growth
between two plated through holes
that had previously caused
shorting within a printed circuit
assembly (IPC-9691A: User
Guide for the IPC-TM-650,
Method 2.6.25), courtesy of
NTS-Trace Labs Inc. c Is a high
magnication micrograph
showing debonded glass bres
adjacent to a drilled hole; copper
plating solution has penetrated
and deposited copper between
these bres (Courtesy Hytek).
d Shows sections taken transverse
to the bresit reveals CAF as
copper deposition (white) at the
bre-resin interface; EDX
spectrum on a thick deposit
conrms pure Cu (Courtesy NPL)
6.9 Problems Associated with Soldering Fluxes
Fig. 6.52 (continued)
follow glass bres (often surrounding individual bres),
cracks in the laminate or weaknesses in the resin bond. The
deposited lament is usually copper. Often CAF is not dis-
covered until electrical failure has occurred resulting in some
localized burning, resin vaporization and breaking of glass
bres. It is a difcult laboratory exercise to make failure
analysis of CAF: the lament is often very thin, it follows the
397
weave of the glass so that it cannot be seen in a horizontal or
vertical cross-section. Often the deposit is hygroscopic and
washes away during sectioning. If poor PCB manufacturing
processes are suspected, boards can be submitted to acceler-
ated testing when CAF may appear around anodic through
hole vias after 140 h at 100 V during climatic chamber
exposure at 85 C and 85 %RH (Lee et al. 2006).
398 6 Failure Analysis of Electrical Interconnections and Recommended Processes
A recent spacecraft power supply failure resulting from
CAF (Heltzel 2014) caused the re-visiting of publications by
Rogers et al. (1999) and Karavakis and Bertling (2004), and
recent laboratory investigations appear to conrm the fol-
lowing reactions:
Anode:
Cusolid ! Cunaq ne and
2H2 O ! O2 gas 4Haq 4e
Cathode:
2H2 O 2e ! H2 gas 2OH
Cunaq ne ! Cusolid
Follow-on evaluations of defective multilayer printed circuit
boards, due to short circuits or deceased insulation resistance,
have identied particulate contamination to exist between
internal conductors. Contamination, such as small bres, is a
known problem on prepreg and copper clad laminates and is
possibly caused by air-borne particles as a result of poor
handling and storage during PCB fabrication. Accelerated
testing (85 C, 75 %RH, with 100 VDC electrical bias) has
shown copper to migrate along bres bridging between internal
conductor tracks (Mc Brien and Heltzel 2013). The
contaminant bres were observed using an IR camera and
SEM as part of the failure analysis. Long bres had been coated
with electrochemically migrated copper during the test period,
much like the CAF mechanism described previously. This
caused leakage currents and test failures. The detailed
investigation concluded that existing cleanliness requirements
for laminate materials should be made tighter, particularly
when laminates are being processed for the space industry.
6.9.10 Potential Health Hazards
in the Electronic Assembly Area
Spacecraft electronics are generally made as a low-series
production and it is common for individual circuit boards to
have unique designs and layouts which do not warrant the
extra cost of conducting the trials necessary for setting up and
optimizing the many parameters associated with machine
soldering operations. For this reason companies will employ
a skilled team of personnel who, dressed in their clean-room
protective clothing, will assemble semiconductor devices and
PCBs using processes that involve the cutting of component
leads, the application of fluxes based on colophony or ami-
nes, soldering with very hot irons, performing repairs, and
mixing and applying staking compounds and conformal
coatings. This work, so-called hand-, or manual-soldering, is
labour intensive and can expose the operators to various
working hazards including the risk of acquiring dermatitis.
Physical wounds must also be avoided when cutting
breglass such as that contained within PCB laminates.
Exposed bres can penetrate into the skin and cause irrita-
tions. Fibres released into the working area can circulate in
the air and lead to itching of persons exposed skin.
Wounds to the body and eyes can be inflicted by the wrong
use of cutting tools such as scalpels and even lead clippers.
Operators are trained to select lead clippers having a sharp, 90
cutting edge which is placed over the lead such that the per-
pendicular cutting sides are facing the component body
(ECSS-Q-ST-70-08). When hard component leads, such as
those made from Kovar or hard-drawn ironnickel alloys, are
cut there is a very large release in energy (shock) which is
evident from the speed and distance of the projected small
length of lead wire. If the clippers are used the wrong way
round, the released energy can easily damage the component
or shatter its glass-to-metal seals. However, when correctly
performed the operator must avoid damage to his own or his
colleagues eyes from the small projectile, and similarly this
small length of conductor must not be allowed to land in the
midst of flight hardware. Irritant dermatitis can also be caused
by the wrist straps worn by operators in order to avoid elec-
trostatic discharges. These straps often contain a stainless steel
or nickel-plated earth connection that is also in contact with the
operators wrists. It has been considered that 10 % of all
females have an allergy to nickel, particularly when it is lea-
ched out from an armband with perspiration. This may be
compounded by the presence of small amounts of liquid sol-
dering flux which are known to contain acids, amines, sol-
vents, and other mildly toxic substances. A study of skin
disorders among hand solderers in the electronics industry
found that work-related dermatitis of the arms or forearms was
commonplace and that the use of gloves did lower the
prevalence of such disorders (Koh et al. 1990, 1994).
Colophony fumes are released during soldering operations
and can have a mild chronic obstructive effect on the
breathing of female solderers (Lee et al. 1994). They are a
product from the solder flux, and are listed by the British
Health and Safety Executive as one of the six substances
responsible for most cases of occupational asthma (Evans
1995). This is characterized by coughing, shortness of breath,
wheezing, and chest pain. The solder flux fumes contain a
wide range of toxic compounds, including hydrochloric acid,
acetone, hydrocarbons, and carbon monoxide. It is important
to utilize fume extraction units when space companies are
performing solder-assembly work. These extractors may be
sited locally on each operators work bench or may be
positioned above a soldering machine such as is used for
wave-soldering. An important recent study has found that
hand soldering operators in the electronics industry, using
soldering irons and a solder of 60 %SnPb, are not being
exposed to a lead hazard (Jalbert et al. 1994). Care should be
taken to avoid a strong air flow as this can affect the quality of
the soldered connection, causing the liquid solder to freeze
6.9 Problems Associated with Soldering Fluxes 399

too rapidly on one side of the joint. High-volume units for ller metal is placed between the overlapping surfaces,
displacing large quantities of air are also available but they pressure is applied but very little melting will occur at the
should not produce excessive noise (should be less than process temperature (see Table 6.6 and also Sect. 6.11). The
50 dB) as this can adversely affect operator performance dip brazing process was reviewed in Sect. 4.12.
when precision work is carried out. A short, concise recom- Brazing can be considered for the manufacture of elec-
mended read is the Safe Soldering Work Practices document tronic assemblies as a process for electrical interconnection.
issued by the Lawrence Berkeley National Laboratory in It is particularly suitable when one or more of the following
Dec., 2010 and available on-line. spacecraft design requirements should be met:
Good lighting is also a prerequisite for quality soldering.
ECSS-Q-ST-70-08 requires a lighting intensity of 1080 lux Joints will be loaded in shear.
on the work surface, and at least 90 % of the work area Joining of thin sections to heavy sections.
should be shadowless. Good lighting will increase produc- Joining of sections too thin for welding.
tivity and reduce the need for rework (see also Sect. 6.12). Hermetically sealed assemblies.
Poor lighting contributes to accidents and reduces the ef- Low distortion with a minimum residual stress distribution.
ciency or effectiveness with which people work. It also Joining dissimilar metal combinations which cannot be
reduces the contrast of objects, which is particularly welded.
important for soldering and inspection operations. Even a Producing complex permanent assemblies with multiple
reduction in lighting of 20 % seriously reduces visibilityat joints which would be inaccessible to welding (for
normal lighting levels this is equivalent to reducing the instance waveguide networks).
illumination at the working surface by a factor of more than
10. Lamps must give a suitable colour quality. High-pressure Table 6.6 Metal combinations used for diffusion brazing/soldering
sodium lamps have the longest life and highest efciency, Substrate Filler Process Remelt
but some metals such as copper and gold-plated parts are metal temp. (C) temp. (C)
difcult to work with under this lighting. Aluminium Cu 550 650
Other problems connected to the soldering environment alloys
are similar to those more widely researched by the Interna- Aluminium AlllSi 550 650
tional Institute of Welding, particularly by Gonnet (1995), alloys 3Cu
who has discussed in detail the ergonomic aspects of Aluminium Al10Si 550 650
working stations, the effect of fumes, the projection of liquid alloys 4Ge
metal spatters, and the neurophysiological aspects (nervous Aluminium ZnlCu 525 650
stress) due to the close attention required by individuals for alloys
watching and producing high-quality interconnections. This Cobalt alloys Ni4B 1175 1475
may be made worse during periods of insufcient work Copper alloys Au 1000 1050
(monotony) or during the periods of excessive work when Nickel alloys Ni4B 1175 1450
the delivery dates for space hardware are nearby. Nickel alloys Ni12P 1100 1450
Silver alloys Ag30Cu 825 950
Steel (including Fe12Cr 1050 1400
6.10 Problems Associated with Brazing stainless) 4B
Titanium alloys Cu50Ni 975 1700
6.10.1 Design Considerations
Titanium alloys Ag15Cu 700 1700
15Zn
Modern usage of the term brazing excludes processes Copper Sn 500 >900
which do not rely on capillary flow of the ller metal metallizations
between the closely overlapping surfaces of the metals being Gold Sn 450 >600
joined. Braze alloy ller metals melt at temperatures above metallizations
427 C (800 F) and are formulated to give the required Gold Hg 50 >310
fluidity which will enable them to flow into small joint gaps metallizations
by capillary attraction. Because wetting can take place only Silver Sn 250 >480
on clean metal surfaces it is essential that thin oxide lms, metallizations
tarnish, and the like are removed from base metals by the Silver In 175 >660
self fluxing action of the ller metal, or the addition of a metallizations
fluxing agent, or the use of inert gases, reducing atmo- All compositions are given in wt%
spheres, or vacuum. In the case of diffusion brazing the Based on the paper by Jacobson and Humpston (1992)
400 6 Failure Analysis of Electrical Interconnections and Recommended Processes
Great care must be taken to assess the actual clearance of
the joint being brazed. It is recommended that simple cal-
culations are made utilizing the coefcients of expansion
presented in Table 6.7. These calculations should be fol-
lowed up by trial operations using representative test sam-
ples. For most brazing alloys, effective capillary action
occurs when the braze path gap at the brazing temperature is
maintained within the range 0.0250200 mm.
6.10.2 Brazeability of Materials and Braze
Alloy Compositions
It is not possible to detail the optimum heating methods and
ller metals for the wide range of base metals and combi-
nations of base metals. Literature searches and experimen-
tation must be performed before the brazing of uncommon
materials. It is important to know if the surfaces of the
metallic piece-parts being brazed will oxidise or be reduced
in a particular brazing atmosphere. The Ellingham diagrams
presented in Appendix 7 can be consulted to determine:
(a) the relative ease of reducing a given metallic oxide to
metal
(b) the ratio of carbon monoxide to carbon dioxide able to
reduce the oxide to metal
(c) the ratio of hydrogen to water vapour that will be able
to reduce the oxide to metal at a given temperature.
Limited details for various parent materials that have been
successfully brazed for spacecraft applications are as follows:
Aluminium alloys
Brazing is possible for those alloys whose melting point is
above that of the available ller alloy. Clad brazing sheet
normally requires no extra ller metal. All the
non-heat-treatable alloys can be brazed. The most suitable
heat-treatable alloys for brazing are AA 1100, AA 3003, AA
6061, AA 6063, and AA 6951; see also Sect. 4.12. The alloys
AA 2011, AA 2014, AA 2017, AA 2024, and AA 7075 are
very difcult to braze owing to their low melting points.
Beryllium alloys
These alloys are highly reactive and extremely toxic; they
can be brazed with AISi and Ag alloys, in pure dry
hydrogen or in vacuum.
Ceramic materials
These materials are inherently difcult to wet unless reactive
braze metals based on Ti, TiVCr, AgCuTi or CuTi by most heating methods. Filler metal choice is usually
systems are employed. Titanium is efcient when present in
quantities greater than 1.25 wt% (Sechi et al. 2009). The
most common process used industrially is to metallize
alumina ceramics with molybdenummanganese-sintered
metal powder and braze in a hydrogen atmosphere with Ag
Cu or AuNi ller metals (see, for instance, Fig. 4.12).
Routes for brazing alumina to Kovar, as used for battery cell
joining and parts of a travelling wave tube, are shown
schematically in Fig. 6.53.
Copper-base alloys
All heating methods can be applied using silver-base or
copper-base ller metals. Berylliumcopper is brazeable
only with silver-based ller metals (should be heat-treated
before brazing). Bronzes and coppernickel alloys must be
in the stress-free condition prior to brazing. Oxygen-bearing
coppers may not be brazed in high hydrogen content
atmospheres owing to severe embrittling effects as illustrated
in Fig. 4.8. Alloys containing elements with high vapour
pressures are unsuitable for vacuum brazing.
Magnesium alloys
These are usually limited to dip-brazing using
magnesium-based ller metals (BMg-1, BMg-2). Many
magnesium alloys cannot be brazed, so specialist advice is
always necessary.
Nickel alloys
Nickel and some nickel-rich alloys are subject to intergran-
ular penetration by certain silver-based ller metals, partic-
ularly when brazed in a state of stress. Penetration may be
prevented if PdAgCu alloys are used in dry hydrogen
without the use of a flux. The special nickelchromium
aluminium high-strength alloys require special fluxes to
remove their very stable oxide lms. A nickel-to-copper
joint using silver-copper eutectic is shown in Fig. 6.54.
Refractory metals (W, Mo, Ta, and Nb)
All these metals readily form oxides in contact with air, and
embrittlement can occur when these oxides are exposed to
nitrogen- and hydrogen-containing atmospheres. Tungsten
and molybdenum should not be heated above their recrys-
tallization temperatures (1300 and 1150 C respectively);
their surfaces must be thoroughly deoxidized, and brazing is
best accomplished in vacuum or under dry hydrogen, using
PdAgCu or AgCu ller metals. Tantalum and niobium
should be brazed in argon, helium, or vacuum, using special
alloys based on Ni, Zr, Au, or Ti (specialist advice is
recommended).
Steels
Carbon and low-alloy steels All can be successfully brazed
dependent upon temperature usage of the brazed part. The
gap-lling Ni brazing alloys are preferred over high-Cu
alloys because of less critical t-up requirements. The more
6.10 Problems Associated with Brazing 401

Table 6.7 Comparisons of materials: coefcient of thermal expansiona

Material High Low Material High Low
Zinc and its alloysc 3.5 1.9 Alloy steels (cast)d 1.5 1.4
c c
Lead and its alloys 2.9 2.6 Age hardenable stainless steels 1.5 1.0
Magnesium alloysb 2.8 2.5 Goldc 1.4
Aluminium and its alloysc 2.5 2.1 High-temperature steelsd 1.4 1.1
Tin and its alloys b
2.3 Ultrahigh-strength steelsd 1.4 1.0
c c
Tin and aluminium brasses 2.1 1.8 Malleable irons 1.3 1.1
Plain and leaded brassesc 2.1 1.8 Titanium carbide cermetd 1.3 0.8
Silver b
2.0 Wrought irons c
1.3
CrNiFe superalloysd 1.9 1.7 Titanium and its alloysd 1.3 0.9
Heat-resistant alloys (cast) d
1.9 1.1 Cobalt d
1.2
Nodular or ductile irons (cast)d 1.9 1.2 Martensitic stainless steelsc 1.2 1.0
Stainless steels (cast) d
1.9 1.1 Nitriding steels d
1.2
Tin bronzes (cast)c 1.8 1.8 Palladiumc 1.2
Austenitic stainless steels c
1.8 1.6 Beryllium 1.1
Phosphor silicon bronzesc 1.8 1.7 Chromium carbide cermetc 1.1 1.0
Coppers c
1.8 Thorium b
1.1
Nickel-base superalloysd 1.8 1.4 Ferritic stainless steelsc 1.1
c c
Aluminium bronzes (cast) 1.7 1.6 Gray irons (cast) 1.1 0.3
Cobaltbase superalloysd 1.7 1.2 Beryllium carbided 1.0
Beryllium copperc 1.7 Low-expansion nickel alloysc 1.0 0.8
Cupro-nickels and nickel silvers c
1.7 1.6 Beryllia and thoriae 0.9
Nickel and its alloysd 1.7 1.2 Alumina cermetsd 0.9
CrNiCoFe superalloys d
1.6 1.4 Molybdenum disilicidee 0.9
Alloy steels d
1.5 1.1 Ruthenium b
0.9
Carbon free-cutting steelsd 1.5 1.5 Platinumc 0.9
Vanadium b
0.9 Osmium and tantalum b
0.6
Rhodiumb 0.8 Zirconium and its alloysb 0.6 0.55
Tantalum carbide d
0.8 Hafnium b
0.6
Boron nitrided 0.8 Zirconiae 0.6
Niobium and its alloys 0.7 0.68 Molybdenum and its alloys 0.6 0.5
Titanium carbided 0.7 Silicon carbidee 0.4 0.39
Steatitec 0.7 0.6 Tungstenb 0.4
Tungsten carbide cermet e
0.7 0.4 Electrical ceramicsc 0.4
Iridium b
0.7 Zircon c
0.3 0.02
Alumina ceramicse 0.7 0.6 Boron carbidee 0.3
Zirconium carbided 0.7 Carbon and graphitec 0.3 0.2
5 1
Units are 10 C
a
Values represent high and low sides of range of typical values
b
Value at room temperature only
c
Value for a temperature range between room temperature and 100390 C
d
Value for a temperature range between room temperature and 540980 C
e
Value for a temperature range between room temperature and 12051580 C
Reprinted from Materials Selector, Reinhold Publishing Co., Penton/IPC
402 6 Failure Analysis of Electrical Interconnections and Recommended Processes

Fig. 6.53 Two routes for joining

metals to ceramics. In the active
metal process, titanium hydride
can be painted onto the alumina
ceramic and brazing is then
performed with AgCu eutectic
foil, or the active metal can be
present as an alloy with the Ag
Cu. The longest established active
metal braze alloys are 65Ag
33.5Cul.5Ti and 68.8Ag
26.7Cu4.5Ti. From Fernie and
Sturgeon (1992) and 70.5Ag
26.5Cu3Ti (Hirnyj 2008). See
also Fig. 4.12

Fig. 6.54 Characterization of a

nickel-to-copper brazed joint for a
spacecraft application.
Microstructure of silver-copper
eutectic braze alloy showing
some penetration of copper and
copper-rich dendrites attached to
the nickel interface

costly CuAgZnCd ller metals are particularly suitable
for furnace brazing. Graphite powder mixed with an
appropriate carrier has recently been found to provide a good
bond between low-carbon steelsbrazing is in situ at 1200
C for 5 min under vacuumthe brazing liquid carburizes
both sides of the joint, and this is followed by rapid diffusion
of carbon into the steel.
Stainless steel Austenitic nonhardenable stainless steels
are subject to stress-corrosion cracking in the presence of
molten ller material, and best results are achieved when
brazing is performed only when these steels are in their
annealed condition. Also carbide precipitation in the heated
area lowers their corrosion resistance (if high-temperature
braze alloys are chosen these carbides can be redissolved by
heat treatment after brazing). The use of low-carbon alloys
such as 304 L will minimize this condition. The stabilized
alloys 321 and 347, however, do not undergo carbide pre-
cipitation, and should be used wherever possible.
Ferritic nonhardenable stainless steels are best brazed
with AgCu ller metals, and when exposed to water or high
humidity special braze alloys are required to prevent crevice
corrosion. Martensitic stainless steels requiring heat
6.10 Problems Associated with Brazing 403

Fig. 6.55 Characterization of a

nickel-to-titanium alloy brazed
joint. Microstructure of gallium
palladiumsilver braze alloy
(white phase is gallium-rich). The
titanium alloy is Ti6AI4V. The
various microhardness values
(Hv) of the layers are: 1 = 437,
2 = 805, 3 = 476, 4 = 1073,
5 = 770, 6 = 138. Despite the hard
region 4, mechanical tests showed
these intermetallic phases did not
embrittle the joint (preparation by
David Adams)

treatment must have a compatible brazing thermal cycle. Fused borax (disodium tetraborate), borates, and
Precipitation-hardening stainless steels also require special fluoroborates for general brazing at higher than 750 C
braze alloys and thermal proles tailored for each individual depending on composition.
parent alloy composition The furnace atmosphere for braz-
ing stainless steels is described in Sect. 6.10.4. They are supplied in paint, paste, and powder form, and
may already be present on certain ller metal preforms.
Titanium alloys
Ideally, they should be easy to mix. They should be easy to
Titanium is an extremely active material and very difcult to
apply to the workpiece and, once molten, need to adhere to
braze; it readily oxidizes and produces brittle intermetallic
vertical surfaces. Fluxes must be selected to be liquid and
compounds with the common braze alloys. Vacuum furnace
fully active at a temperature at least 50 C below the solidus
brazing at <105 torr and inert gas (argon) brazing have been
of the chosen ller alloy and remain active to temperatures at
successfully performed on several scientic spacecraft
least 50 C higher than the liquidus of the alloy. Braze alloy
detector housings, including a large complex system for the
manufacturers will advise on fluxes suitable for penetrating
Italian SAX project, using the braze alloy Gapasil (based
the capillary gaps between various combinations of metals,
on the alloy galliumpalladiumsilver). Microstructures are
and their actual composition may be influenced by the
seen in Fig. 6.55. Incidentally, despite the low melting point
method of heating employed.
of gallium, this braze alloy was tested for sublimation under
Except for the nonactivated boron-based fluxes and sil-
vacuum and found to be completely stable. The great afnity
icofluoride types, all flux residues are hygroscopic and
of titanium for oxygen, hydrogen, and nitrogen means that
potentially corrosive to most parent metals. Usually, flux
the workpiece must be completely cooled to 80 C before the
residues left from the active halide-type fluxes are water
furnace is opened to the atmosphere.
soluble and can be removed by immersion in warm water.
Other residues may be removed only by immersion in hot
10 % sulphuric acid or hot 10 % caustic soda, according to
6.10.3 Brazing Fluxes and Their Removal
the braze alloy and parent metal compositions. Very tena-
cious flux residues may be removed only by mechanical
Brazing fluxes are specially formulated to displace slight
treatments (abrasion, brushing, vapour blasting, etc.). In all
surface contamination but not heavy oxides and greases,
cases flux residues are easier to remove if the workpiece has
which must be eliminated by precleaning operations. Non-
not been subjected to overheating or prolonged heating.
gaseous fluxes are usually proprietary mixtures based on:
There appear to be no neutralizes for flux residues, so to
avoid corrosion, outgassing under vacuum, or general
Active halides, such as chlorides and fluorides, for
spacecraft contamination, they should be completely
low-temperature brazing.
removed.
404 6 Failure Analysis of Electrical Interconnections and Recommended Processes

6.10.4 Atmospheres for Brazing
One advantage in the use of controlled atmospheres in
brazing is the elimination of potentially dangerous fluxes
and their residues which may become embedded within the
solidied ller metal. The ability of the atmosphere to reduce
or totally prevent the formation of surfaces oxidation can be
achieved under FURNACE BRAZING, but such atmo-
spheres are also benecial when used in conjunction with
other heating methods such as INDUCTION and
RESISTANCE BRAZING. The surface quality of spacecraft
piece-parts is maintained and often improved by controlled
atmospheres, particularly when they reduce the surface
oxides. This will facilitate inspection steps and any nished
operations. The most common types of atmosphere are
dissociated ammonia, combusted town gas, hydrogen, inert
gases, and partial vacuum.
The operation of a high-temperature brazing furnace is
shown, together with some spacecraft piece-parts, in
Fig. 6.56. The enclosed brazing chamber is superior to the
use of continuous furnaces, where there is often insufcient
oxidation protection. Stainless steel and nickel alloys require

Fig. 6.56 Small furnace for

high-temperature brazing to
1100 C in a controlled hydrogen
atmosphere. The piece-parts are
arranged inside a 55 1 chamber
containing the reducing gas.
Heating is by radiation and
conduction from the internal wall
of the chamber. a Shows the
red-hot chamber being withdrawn
from the induction heating
furnace. b Shows the resulting
ceramic-to-metal brazed parts
(and their method of jigging) for
the Italian Electron Gun Tethered
Satellite (courtesy of Laben-Proel
Technology)
6.10 Problems Associated with Brazing
the maximum protection afforded by the pure hydrogen
atmosphere of the LabenProel furnace illustrated. Stainless
steel is brazed in the temperature range 9301100 C using a
nickel-brazing alloy. The critical element in stainless steel
with regard to the brazing atmosphere requirement is chro-
mium. Even at a temperature of 1000 C, with 100 %
hydrogen, a dewpoint lower than 37 C for the hydrogen is
required to prevent oxidation of chromium.
6.10.5 Safety Precautions
The brazing of spacecraft parts and materials entails certain
hazards, and it is essential that operators are made fully
aware of the following safety precautions in order to avoid
accidents.
General
Protective clothing, gloves, aprons, and shoes must be worn
by brazers to avoid dangers from hot and liquid metals.
Goggles or spectacles are particularly necessary.
Cadmium
Many common braze alloys contain cadmium as a major
constituent. These are generally prohibited for space use (see
Sect. 2.3) but this restriction may be waived if the item is
intended for a pressurized system. It is important to avoid the
inhalation of cadmium fumes by the use of local exhaust
ventilation. When emissions are excessive, for instance if the
alloy is heated directly by a torch flame or if it is overheated
beyond its melting range, then individual air-supplied res-
pirators are recommended. Air monitoring is often manda-
tory within larger companies.
Fluorides
Compounds containing fluorine such as alkali metal
fluoroborates which are used with silver-based and alu-
miniumsiliconbased ller metals have toxic effects when
inhaled as fumes or when digested orally. Dip brazing with
these fluxes should be performed in baths tted with adequate
exhaust units. Hands should be washed before meals, with
particular attention paid to scrubbing under the ngernails.
Barrier hand creams help prevent irritation, but any cuts or
scratches must be covered immediately. If fluorine com-
pounds are swallowed, a doctor must be called, and the patient
should drink plenty of milk mixed with calcium carbonate.
Beryllium
Beryllium may be present within magnesium ller metals.
The presence of beryllium in any form requires adequate
ventilation, and brazing should commence only when a
health and safety ofcer is satised that there will be no
dangers to the operator or his surroundings.
405
Zinc
Zinc emissions may result from metal platings (galvanized)
on ground-handling equipment or spacecraft pressurized
units. Although the health hazard is not so great as for the
previously mentioned elements, it is essential to provide
suitable exhaust ventilation.
Heating systems
Assemblies to be dip brazed must be completely dry to avoid
explosions. Torch brazing must be performed well away
from inflammable materials. Hands should not be inserted in
or placed closed to the work coil during induction brazing
rings, bracelets, and watches will heat up very rapidly.
Furnace brazing will often use controlled atmospheres which
burn off with colourless flamesplace a piece of asbestos
string in the frame to produce a yellow glow. Mixtures of
many furnace atmospheres with air are extremely explosive
purge all air before heating or igniting.
All personnel working in brazing areas should be familiar
with rst-aid treatment, particularly the treatment of burns
and acid burns. First-aid equipment and foot or elbow
operated high-volume, low-pressure water showers should
always be available close to the brazing shop.
6.10.6 Produce Assurance Applied to Brazing
Operations
Brazed joints are used in a great many spacecraft structural
and electrical applications, each having different functional
requirements. Mechanical strength, hermeticity, and surface
appearance will have varying degrees of importance for each
designed joint. Inadequacies, which can result in failures of
brazed joints, are often attributable to differences in the
brazeability of the parent metals. Regardless of the method
of heating (furnace, inductance, dip, etc.) it is usual to nd
an increase in defective joints as brazeability decreases.
Mechanized brazing may in several applications be more
reliable than manual brazing, but in all instances trained
personnel and 100 % inspection are recommended. Detailed
specications related to the inspection of brazed joints may
be difcult to locate. It is important that inspection criteria
are agreed upon before space hardware is submitted to the
brazing facility. A simple set of inspection requirements
drawn up by the writer for the acceptance of brazed
waveguide flanges is given in Sect. 6.10.7; it might be useful
as a guideline for similar piece-parts.
As in joining by soldering, it is good practice to ensure
that each aspect of the process is well-documented. Process
documents should include: the identication of ller metals
and fluxes for each combination of parent metals, the method
of preparing the joints (cleaning, deburring, clearances,
406 6 Failure Analysis of Electrical Interconnections and Recommended Processes
location, and form of braze alloy), details of the brazing
equipment (e.g. atmospheres, temperature, times with pos-
sible preheating and cooling cycles), approved methods for
flux removal, and nally inspection.
Acceptable brazed joints are not ware generally charac-
terized by:
(a) clean, smooth surfaces with the absence of external
porosity,
(b) full penetration of the braze path,
(c) complete wetting between the ller metal and the
jointed surfaces without excessive melting or erosion of
the base metals.
The visual criteria for rejection will depend on the criti-
cality of the joint, and may include:
(a) excessive braze metal,
(b) inadequate contour of braze llets,
(c) flux residue or other contamination,
(d) presence of surface porosity and blisters,
(e) evidence of overheating by penetration of the base metal
by the ller metal, or under heating as seen by presence
of unmelted braze alloy or inadequate penetration.
Other inspection test methods may include radiography,
ultrasonics, and pressure leak testing (possibly helium leak
testing for ne leaks and dye penetration for gross leaks). At
the commencement of an unfamiliar brazing job it will be
desirable to make destructive mechanical and metallographic
tests on preproduction samples. These tests may include the
proof loading of assemblies by subjecting them to loads
slightly in excess of those required in service, and corrosion
and stress-corrosion tests, which simulate the intended
working environment of the part (e.g. battery cell cap
interconnections). For high-reliability jobs random samples
should be selected from each lot and subjected to in-line or
end inspections which may include nondestructive and/or
destructive tests. Dimensional inspections should also be
made for compliance with the relevant drawings.
6.10.7 Inspection Criteria for Brazed
Aluminium Alloy
Waveguide-to-Flange Joints
1. General
These inspection criteria relate to interconnections made
by any method of brazing, including torch, furnace,
induction, resistance, and dip. It is based on the
requirements of MIL-B-007883C.
2. Quality of joint
2:1 Contour
The contour of an outside ller joint shall be of a
uniform radius with a minimum amount of excess
braze or flash over the adjacent surfaces.
2:2 External defects
2:2:1 Surface porosity (see Figs. 6.57 and 6.58)
External voids are not permitted.
Incomplete fusion is not permitted.
Single porosity: no single pore shall exceed
0.5 mm.
Concentrated porosity: accumulated total of
diameters shall not exceed 1.0 mm in
diameter in any 10.0 mm of braze.
2:2:2 Cracks
Cracks are not acceptable regardless of
location.
2:2:3 Blisters
Overheating of the base metal resulting in
blistering on the surface of the base metal or
any plated layers shall be cause for rejection.
2:2:4 Excess braze metal
Braze ller metal which may be in excess of
that required of the joint is acceptable pro-
vided that it does not interfere with the
functioning of the completed assembly.
2:2:5 Undercutting
Melting or erosion of the base metal on the
outside surface adjacent to the brazed joint
shall be limited to a maximum depth of 5 %
of the stock thickness and a maximum of
15 %, cumulative, of the braze llet length.
2:2:6 Unmelted braze alloy
The presence of unmelted brazing alloy in a
joint is unacceptable and shall be cause for
rejection.
2:2:7 Residual flux
No residual flux shall be permitted on the
surface of a brazed joint or base metal.
2:2:8 Foreign material
The brazed joint shall be free of all foreign
material.
2:3 Penetration and internal voids
Braze ller metal must have full penetration from
the external surface to the internal surface. The
criteria for internal voids (i.e. at the end of the braze
path) shall be a maximum void length of 5 mm for
individual voids. The total unbrazed area, including
trapped flux, scattered porosity, and voids, shall not
exceed 20 % of the internal length of the internal
interface area.
6.10 Problems Associated with Brazing 407

Fig. 6.57 Sketch illustrating

some inspection criteria for
brazed joints (external llet and
microsection). The criteria and
sketches were made and used by
the author for this waveguide
application due to the absence of
other suitable standards

Fig. 6.58 Sketch illustrating

inspection criteria for end of braze
path (waveguideflange assembly
bottom view)
408 6 Failure Analysis of Electrical Interconnections and Recommended Processes
3. Inspection
3:1 Visual inspection
All indicated flaws shall be checked visually under
ten power magnication for verication of the
defect.
3:2 Penetrant inspection
Fluorescent or visible dye penetrant is not permitted
for spacecraft flight hardware.
3:3 Radiography
Radiography shall be performed when required by
the customer to verify conformance to 2.3.
3:4 Flux removal test
A suitable test, such as the absence of a typical
chloride precipitate in a 5 % aqueous solution of
silver nitrate on the cleaned and rinsed part, may be
used to determine that flux has been adequately
removed.
3:5 Metallography (see also Fig. 6.58)
Microsections shall be performed as part of a veri-
cation test to ensure full penetration of the braze
metal. The presence of voids within the braze path
is permitted to a maximum of 20 % of the
cross-sectional area of the braze path (as seen on the
microsection).
6.11 Diffusion Soldering/Brazing
Diffusion characteristics between different metals can be
used to an advantage for the joining of metal surfaces.
Previous sections of this book have discussed the negative
effects of cold welding and wear; some combinations of pure
metals that have a high solid solubility in each other were
shown in Fig. 2.3. A form of diffusion bonding was pre-
sented in Fig. 4.55, when a foil of one alloy is sandwiched
between two other similar alloys and subjected to a high
temperature and some pressure exerted by jigs. However,
true diffusion/soldering is performed at the lowest possible
temperature and pressure is deliberately applied in order to
break down trapped oxide lms and enhance the diffusion
mechanism. Semiconductor chip bonds to the internal sur-
faces of metal packages have been made for many years by
this kind of process. The packages are usually made of
gold-plated Kovar and the silicon chip is rub-pressed either
to the pure gold or to a thin (25 m) goldsilicon alloy wafer
so that melting will occur at about 400 C. During the dif-
fusion process, the low melting component of the alloy is
replaced by an alloy that becomes progressively higher in
melting point. This process is governed by the
phase-diagram of the particular alloy system. Semiconduc-
tors are now being bonded with very thin (below 5 m)
goldtin layers; these can be bonded with a peak temperature
below 280 C within a few minutes and without a scrubbing
action when the processing atmosphere excludes oxygen
(Dohle et al. 1996).
Low-melting diffusion soldering systems such as those
listed in Table 6.6 (on p. 401) have been researched by
Jacobson and Humpston (1992) and Jacobson and Sangha
(1996). Many should nd useful applications during the
manufacture of spacecraft electronics where special tem-
perature limitations might be overcome by the insertion of
low-melting metals and alloys between parts to be joined.
A schematic representation of diffusion soldering is shown
in Fig. 6.59. A particularly useful system for joining
silver-plated copper parts together could, for instance, use an
indium interlayer where the processing temperature would
be only 175 C. The silver plating of other metals and alloys
is also possible, followed by diffusion soldering with
indium. Silver-to-tin interlayer bonding can be made at 250
C as shown in Table 6.6. The lowest-temperature bond
achieved by Jacobson and Humpston was with a processing
temperature of 150 C, by using silver-plated substrates
coated with thin, duplex layers of tin-on-indium. (Indiumtin
eutectic has a melting point of 117 C.) These systems have
an advantage because when metallic couples possessing very
different coefcients of expansion need to be joined, the high
residual stresses caused by thermal mismatches of expansion
during normal high-temperature brazing, can either be
avoided or at least be minimized. Particular space hardware
applications might be associated with difcult component
packaging assemblies that involve joining Invar to copper,
Kovar to aluminium, or aluminium to aluminium MMCs
reinforced with particles of SiC.
6.12 Effects of Rework and Repair
on Soldered Interconnections
6.12.1 General
Rework of soldered joints and the possible need to repair
electronic assemblies are some of the most controversial
aspects of spacecraft product assurance. The decision mak-
ing might involve a large number of personnel from different
departments (e.g. persons from the company manufacturing,
inspection, quality assurance, and materials and processes
departments, as well as project management and marketing
departments). Material review boards (MRBs) will be
organized, which may request the attendance of the space-
craft prime contractor, certain sub-tier contractors, their
various experts, and the nal customer. Obviously, time and
cost will be at a premium, but it is essential that the
underlying problem is understood from a technical view-
point before any corrective actions are performed. It is at this
point that the metallurgical laboratory may be requested to
6.12 Effects of Rework and Repair on Soldered Interconnections
Fig. 6.59 Schematic
representation of the diffusion
soldering process (from Jacobson
and Sangha 1996). Reproduced
with permission from
GEC-Marconi Materials
Technology Ltd, Borehamwood,
UK
perform investigations and report to the project teams within
a short period of time. Such activities were discussed in
Sect. 3.3.
The problem may be related to the materials being used,
the quality of the components, the suitability and repro-
ducibility of the process, and on the dexterity of the operator
performing the soldering-related task. Similarly, the
inspector who flags up the problem must himself have ini-
tially decided if the situation could be rectied by rework,
repair, or modication. Here, it is important to distinguish
between these three activities as each will have different
consequences. The presence of specially trained and certied
operators and inspectors will be essential if the correct
decisions are to be made (this responsibility will be expan-
ded on in Sect. 6.14, related to skills training schools). The
following denitions originate from the European soldering
specications ECSS-Q-ST-70-08 and 70-28:
Reworkthe reprocessing of a nonconforming item, to
make it conform to drawing, specication, or contract (for a
soldered connection this could be a solder llet containing a
void that can be re-heated with a soldering iron to make it
409
conform with the relevant workmanship standard, or it may
be an assembly that is judged to be insufciently clean and in
need of a further cleaning cycle)
Repairoperations performed on a nonconforming item to
place it in a usable and acceptable condition according to an
authorized repair procedure/standard. Repair is distinguished
from rework (repair may consist of a component change with
all the associated solder removal and replacement proce-
dures, or it may involve the xing down of a lifted pad)
Modicationthe revision of interconnections on a PCB
such as interrupting conductors and/or adding components
and wire connections.
A maximum of three reworks are permitted on any one
soldered joint without the need for an MRB. Repairs require
some form of authorization from an MRB, as they might be
proposed in order to repair some damage that occurred
during the assembly work, or, at the time of testing space-
flight equipmentseach repair can affect the long-term
reliability of the hardware. Modications may result from
late changes to design when there is insufcient time to order
new PCBs and these will need to be addressed by an MRB to
410 6 Failure Analysis of Electrical Interconnections and Recommended Processes
ensure that they are feasible. Usually a certain number of
repairs and modications are permitted to be performed on
any one circuit board by the contractor without the need for
authorization from the customer, but this will depend on the
specic terms of the contract. The effects of such practices
on the materials and individual parts being assembled will
now be addressed.
6.12.2 Cosmetics of Solder Fillets
Quickly frozen solder, in the presence of flux to prevent
oxidation, produces a solder joint that is smooth, bright and
having an excellent appearance. Slowly frozen solder, due to
different heat-sinking effects of surrounding conductor
materials, can look dull and lumpy with some evidence of
dendrites from the early solidication of the lead phase in the
solder. Multilayer circuit boards commonly have up to 14
layers and may be assembled with side-brazed duel in-line
packages. Such connections will have a high heat capacity
not only will they take longer to heat, they will also take
longer to cool to room temperature and will have solder
llets with a different appearance to those of other smaller
components. The high heat capacity connections may also
have imperfect solder llets on the component side of the
board. Relaxations to the original issue of the PSS-01-708
specication now permit (ECSS-Q-ST-70-08) only a partial
llet on this side, as long as there is evidence of some solder
flow-through and proper wetting angles.
The presence of small voids and blow-holes in solder
llets may also be cosmetic problems. There is no evidence
that they will be associated with electrical service faults.
They are usually caused by outgassing of the boards
plated-through hole or by the entrapment of solder flux.
Corrosion due to flux entrapment has not been a problem, as
space assemblies used nonactivated fluxing products. Also,
during thermal cycling, the voids are through to act as
stress-relief points, as they are spherical in shape and would
contribute by crack blunting. Voids do not promote crack
propagation during extensive thermal cycling as is evident
from also Fig. 6.60a). These were the ndings from many
thermal cycling programmes performed by the writer and
have been conrmed by others (Lea 1991). Similarly, for
observed blow-holes as shown in Fig. 6.60bd, it is safer to
accept the cosmetic defect rather than apply mechanical or
thermal damage during rework.
In the majority of cases, solder joints possessing minor
defects should be left as-is without being submitted to
rework. Rework can damage the component, the PCB
(particularly if it is a thick multilayer board) and can make
the solder alloy harder, and therefore less compliant with the
substrate and component during thermal cycling. These
points will be readdressed in the following sections.
6.12.3 Effect of Rework Electronic
Components
Visual defects on soldered llets are mainly caused by a
poor solderability of the component leads, or metallized
terminations, and the effect of high heat-sink by the pack-
aging material. In these instances, when deviations from the
workmanship standard are not great, it is advisable for
inspectors to accept the joint without requesting rework. The
trade-off between accepting a cosmetically inferior joint and
requesting rework is that there can be damage to the com-
ponent during the rework, or touch-up operation.
Problems that have been caused during the reworking of
spacecraft assemblies included the following:
Damage to adjacent joints and plastic encapsulants by the
soldering iron when the boards have a high package
density.
Ceramic capacitors such as the types, CKR 05 and CKR
06, have copper leads and internal lead-to-silver metal-
lization soldered interconnections (see Fig. 4.36). These
internal joints have melted when excessive heat or time is
involved when reworking the PCB soldered joint. This
causes the internal solder to make short-circuits between
the silvered ceramic capacitance plates or sometimes to
produce open-circuits between lead and metallization;
such defects cannot be visually inspected owing to the
components mouldable plastic encapsulant and are only
exposed by equipment functional tests. Although this
problem was evaluated many years ago (Dunn et al.
1983), it continues to resurface on present-day projects.
Cracking of glass-to-metal seals in dual-in-line packages
with the ingress of solder flux into the package due to
reworking the solder llets from the component-side of
the PCB. Here, the permitted minimum soldering distance
from the seal is given by the component specication as
1.5 mm; this requirement was breached by the rework.
For information related to the assembly, inspection and
reworking of Area Grid Arrays (e.g. BGAs and CGAs) the
reader is referred to Sect. 6.15.10.
6.12.4 Effect of Rework on Plated-Through
Holes
Thick multilayer boards (MLBs) are most likely to become
degraded by the continuous thermal shocks that their internal
copper tracks-to-plated-through holes can experience during
each application of a hot soldering iron. Microsections of
these connections were shown in Fig. 4.35b.
A detailed investigation into the effect of repeated rework
on the reliability of spacecraft MLBs has been performed by
6.12 Effects of Rework and Repair on Soldered Interconnections
Fig. 6.60 Illustrations to highlight the importance of knowledge, skills
training is essential for understanding the ramications of rework.
a Appearance of soldered joint after very severe thermal cycling (1000
cycles between 100 and +100 C). Thermal fatigue cracks appear at
the llet but are not associated with the large internal voidthis
indicates that large voids have no influence on reliability, electrical
conductivity being guaranteed in solder volume opposite void. b The
pit appears as a blow-hole resulting from flux entrapment, it may satisfy
the requirements of ECSS-Q-ST-70-08 if the bottom of the pit can be
observed under high magnication. c Electrical connection is good,
411
wetting of solder to component lead and plated through hole (pth) is
perfect. d This is a cosmetic defectrework to the solder llet by
re-melting the solder will cause thermal stresses loading the delicate
internal connections. If ground planes were present, they would have a
heat-sinking effect and may cause the operator to apply excessive heat
or pressure from the soldering iron, leading to damage of the chemical
bond between laminate copper and plated through copper. Microsection
in (d) has been etched to reveal interfaces and there are no debonds
within pth
412 6 Failure Analysis of Electrical Interconnections and Recommended Processes

the NASA Goddard Space Flight Center (Garrison et al.

1995). The ndings conrmed that three rework operations
were the maximum number that could be allowed on flight
boards. There was evidence of degradation during the fth
operation. Finite element analyses were made of various
board designs and material compositions. For 2.24 mm
thick, glassepoxy, 10-layer boards, type FR-4, when sol-
dered at 320 C, there is a thermally induced stress at the rst
layer (closest to the applied soldering iron) of 110.9 MPa.
Boards having the same construction, but with only four
layers, showed a maximum stress of 94.4 MPa. Clearly, with
such high stresses acting on the mechanical copper inter-
faces, it will be necessary to only procure MLBs from
approved, high-quality manufacturers.
The exact mechanisms that account for interlayer crack-
ing and internal connection cracking during spacecraft solder
Fig. 6.61 Rework and repair operations cause the solder to be heated
joint reworking operations have been reviewed by Grey
and reheated. At the high temperatures, close to the 183 C melting
(1989). Some recommendations are made to lessen this point of eutectic solder, the tin content of the alloy combines with a
problem, but it is again concluded that the rework of MLBs copper conductor to form an intermetallic, Cu3Sn and Cu6Sn5. The lead
should be attempted only as a last resort and then only under phase is not reactive and coalesces to form the small Chinese script
dendrites seen here. The main part of the solder was etched away from
carefully controlled conditions.
the surface of this sample before the topography was photographed.
(Magnication approx. 200.) Note, on exposure to air, the intermetal-
lic oxidizes and is nonwetted by liquid solder
6.12.5 Effect of Rework on Composition
of Joint

Eutectic tinlead solder is preferred for general soldering,

and particularly for the assembly of leaded and leadless
surface-mounted devices. This is because it can creep and
take up the thermal expansion mismatches between com-
ponent and PCB during thermal cycling. Rework of these
joints causes copper from the conductors to react with, and
to be dissolved into, the liquid solder during the time of the
rework operation.
Both the terminal pad and the component lead can
become heated, so that in some areas, close to where there is
a heat input, all the tin in the solder reacts to form inter-
metallics with these copper conductors. Lead is not reactive
and is left as a lead-rich phase, often in dendritic form, on
these partially exposed surfaces. An example is shown in Fig. 6.62 Microsection through surface-mounted device, copper leads,
Fig. 6.61. The intermetallic layer is silver-coloured and quite with solder present in the stress-relief bend. Standard reflow soldering
reflective when compared to the lead-rich phase or the
copper. Intermetallics in the coppertin system are known to bend lled completely with ductile tinlead, or an only
have a very poor solderability when they become oxidized partially lled bent lead connected to a PCB with a hard,
on exposure to air. Rework or repair in these regions may noncompliant lead-free solder.
result in a much reduced wettability of solder to conductor Transmission electron microscopy has been performed on
which, in turn, can result in repaired joints having a some- the strengthening coppertin precipitates and whiskers pre-
what, nonwetted appearance (see also Figs. 4.32 to 4.34). sent in solder (Felton et al. 1991). It is contended that these
The effect of rework on the composition of the solder is very small particles in the solder matrix can be present to a
illustrated by the colour micrographs shown in Figs. 6.62, size of only 5 nm. At this size they will be able to impede
6.63, 6.64 and 6.65. These microsections were made from deformation of the solder by pinning dislocation motion via
samples submitted to a solder rework investigation. It is the Orowon mechanism. The precise intermetallic phase that
unclear which is the lesser of the two evils: a stress-relief is seen in Fig. 6.41 is considered by Felton to be Cu6S5.
6.12 Effects of Rework and Repair on Soldered Interconnections
Fig. 6.63 Detail from joint shown in Fig. 6.62, showing Cu6Sn5
intermelallic at copper-to-solder interfaces and almost no copper in the
solder matrix
Fig. 6.64 As Fig. 6.63 but after 30 s at 260 C during vapour phase
rework simulation. Some coppertin precipitates in matrix
Fig. 6.65 As Fig. 6.64 but after 90 s at 260 C during vapour phase
rework simulation. Many small coppertin precipitates are seen
throughout the solder matrix. Spherical voids are gas pockets. Courtesy
of Mr. Karl Ring, ZVE
413
6.12.6 Recuperation of Unsolderable PCBs
and Component Leads
Incorrect processing methods, such as overheating and
exposure to surface contaminants, can severely impair the
solderability of an otherwise high-quality circuit board. The
same can be said for certain components where the leads
might have become oxidized or converted into coppertin
intermetallic by deleteriously long burn-in times at high
temperatures. Specially made MLBs can be extremely
expensive. The same can be said for unique integrated cir-
cuits and their packages which may have been produced in
low numbers for the space market. The replacement of both
items would normally involve long lead times and high
procurement costs. As launch deadlines are approached there
will be the, hopefully seldom, occasion when products that
would otherwise be rejected because of poor solderability
may have to be utilized. It is possible to recuperate defective
devices by stripping and replacing their damaged solderable
coatings with pristine ones. Bared copper and nickel alloy
leads can be re-tinned by dipping in a solder bath. The ter-
mination areas on PCBs can be selectively stripped, and if
used within a few days can be easily soldered without the
need for strong fluxes (see Figs. 6.66 and 6.67). Chemical
solutions have been developed which can successfully
accomplish this, removing surface oxides, unreacted tin and
lead, and intermetallic compounds (Harris et al. 1992;
Sriveeraraghavan et al. 1995). The stripping procedure
developed by PMD of Coventry, UK (using a proprietary
tinlead stripping solution based on nitric acid with ferric
ions as oxidants, inhibitors, and stabilizersmarketed as
PMD-936) has been tested on space components by Harris
and his co-workers. The method was recently applied to
components for space projects in difcult operational
circumstances.
6.13 Electrical Conductive Adhesives
Electrical conductive adhesives (ECAs) are most usually
silver-lled thermosetting epoxy resins that are marketed as
thixotropic pastes. They are used commercially for the
assembly of surface-mounted devices and are said to be
benecial because of the lower temperatures that are needed
when compared to tinlead soldering. Also, there is no need
to clean, no involvement of lead or other toxic materials, and
the processing parameters may be wider and easier to control
than is the case for conventional soldering. The temperature
of processing is low and there are many savings to be made
owing to the reduced number of processing steps. Other
types of adhesive have been developed and may be selected
from thermoplastic polyimides and silicones. Naturally, the
polymeric matrices of all conductive adhesives have
414 6 Failure Analysis of Electrical Interconnections and Recommended Processes
Fig. 6.66 As-received PCB showing: a termination pad of a
plated-through hole; b intermetallic ring around hole corner is exposed
to atmosphere; c microsection across region shown in (b). a is SEM
image, b and c are optical images and these samples were electrolyt-
ically plated with cobalt to aid edge retention
excellent dielectric properties and it is only the touching of
the loaded metal particles that produce electrical conductive
paths within the adhesive and to the circuit board or to the
metallized surface of the electronic component. Conductive
adhesives are still in an embryonic stage of development but
were stated to have the possibility of becoming the rst
Fig. 6.67 Termination pads from PCB shown in Fig. 6.66, following
chemical stripping. Note that all the solder and intermetallic layers have
been successfully removed and if stored correctly the PCB can be soldered
with nonactivated flux
choice in the electronic component assembly area (Galleo
1995). ECAs have limitations such as, low electrical and
thermal conductivity, migration of the ller metal under high
voltage and poor impact strength. Studies continue; new
materials are being considered for ECA applications and
recent conductivity improvements have been realized by the
addition of silver-graphine nanocomposites to silver
flake-loaded epoxy matrix adhesives (Peng et al. 2014).
Many other investigators have noticed large problems
when processing conductive adhesives; because they are
difcult to inspect, all hi-rel applications need to be checked
by functional testing. The in situ reworking of adhesively
bonded connections is very difcult, if not impossible, to
achieve on highly populated boards. Bonding conditions
such as pressure, temperature, and time play an important
role in the nal joint quality. Liu et al. (1996) at the IVF
research laboratory found that if improper process conditions
are used, electrical resistivity will increase signicantly.
When the pressure applied to the joining parts is too great
there is a deformation of the board or pad metallization and
6.13 Electrical Conductive Adhesives
with insufcient pressure there can be occurrences of no
electrical conduction. In the case of conductive adhesives it
is very important to monitor the electrical resistance of joints
continuously during environmental testingparticularly
during thermal cycling. Conductive adhesives and eutectic
tinlead solders were used during the reliability testing of
flipchip interconnection technology. It was found that
during temperature cycling, failures were detected at both
hot and cold temperature regions, whereas the soldered
connections were successful (Boustedt and Hedemalm 1995;
Liu et al. 1996). Keusseyan and Dilday (1993) also found
that, although mechanically good, most of the conductive
adhesives they tested showed an increase in electrical
resistivity during thermal cycling. The increase in electrical
resistivity of these joints was correlated to the material and
surface properties of the ller metal powders and to the
deformation properties of the organic carrier.
A metallographic section through a typical surface-mounted
component joint made with conductive adhesive is seen in
Fig. 6.68. The electrical conductivity depends on factors such
as the size, shape, and surface oxidation of the silver particles.
Their asperities must touch and must not be affected by
deformation or expansion of the organic matrix.
European spacecraft PCBs are only component-assembled
with soldering technologiesi.e. metallurgically joined
connections. Electrically conductive adhesives are not per-
mitted for this use, but they have found a limited number of
applications during the assembly of very small chip compo-
nents within hybrid packages. Hybrid technology processes
are strictly controlled/qualied for space-use and the her-
metically sealed packages contain a dry gas. This condition
obviates some other problems associated with conductive
adhesives, such as the possibility of silver migration, oxida-
tion of the metallic particles, and the poor thermal conduction
of the adhesive (Harris 1995). Under conditions of vacuum
the heat dissipation from an active component is via its sol-
dered interconnection; silver-loaded epoxy has an inferior
thermal conductivity and this must be borne in mind during
the selection of processes for spacecraft assemblies.
The question, is it possible to consider ECAs as drop-in
replacements for solder pastes? has been raised many times
when last-minute modications were identied for assem-
bled PCBs. Also, this process was suggested should repairs
be necessary to electronic hardware aboard manned space-
craft, such as ISS, where hot soldering with an iron might be
hazardous. A short programme was devised to check the
suitability of ECA processing and its effect on standard
components and material nishes used for space projects
(Thomas and Dunn 1999). PCBs were FR4 having two
conventional track/termination nishes: fused tinlead and
gold-flashed nickel plated copper. Three silver-lled epoxies
were selected as being space-approved materials (Ablebond
8175A, Epotek E2116 and EG8050HC) and processing
415
followed the manufacturers recommendationsall were
mixed then degassed under vacuum. Two conventional
surface mount devices were selected: 208QFPs with SnPb
nish and 0.2 mm pitched copper leads, and, C2225 chip
capacitors with SnPb plated metallisations. Stencil printing
could only be optimized for 8175A and E2116. Exposure to
40 C/9395 %RH for 504 h, followed by shear testing,
showed all samples to decrease in strength. Storage under
vacuum did not degrade these joints. Powdery tin oxides
were detected within joints to the SnPb-nished PCBs,
acting as a weak interface during exposure to hot/wet
environments. Weak oxide layers within the bonds made to
the porous gold flash also produced weak interfaces. Elec-
trical monitoring during humidity exposures produced open
circuits of the QFPs bonds, but not to the soldered joints, as
seen in Fig. 6.68b. Under thermal cycling ECA joint resis-
tance displayed considerable data scatter (Fig. 6.68c), there
were no deviations in resistance for the soldered intercon-
nections. The thermally cycled ECAs did not display the
extensive oxidation/corrosion observed in humidity exposed
specimen. Our conclusions were:
ECAs cannot be used as drop-in replacements for con-
ventional soldering (however, it is emphasized that fused
tinlead is not a suitable nish for stencil designs, device
geometry or device co-planarity),
ECA joints possess electrical resistance at least an order
of magnitude greater than soldered joints,
dispensing and assembly for ne pitch devices
(208QFPs) was more difcult than C2225 devices due to
their accurate requirements,
Ablebond 8175A performed better than Epotek E2116
under the environmental exposure programme, and
ECAs were not adversely effected by vacuum exposure.
Clearly, the ECA technology and processing methods
have advanced since these tests were conducted. It is noted
that NASA have experienced silver-lled ECAs to be unre-
liable on some space projects due to changes in conductivity
with time, temperature and environment (Mielke 2002).
Newer materials and nishes will need to be evaluated, in a
similar manner to the aforementioned testing, before they can
be applied on current and future space programmes.
6.14 Training and Certification
6.14.1 General
The most valuable resource available to space manufacturing
and fabrication industries is trained people. Management
personnel, scientists, and engineers will usually possess good
academic qualications, but it is equally important that the
416 6 Failure Analysis of Electrical Interconnections and Recommended Processes
Fig. 6.68 a Silver-loaded epoxy used for mounting of chip parts. The
small crack in the adhesive between the epoxy and the chip
metallization is termed pull back by the adhesive supplier who
claimed that there is no risk of propagation during thermal cycling.
Chip metallization is platinumpalladiumsilver. The carrier is gold
thick-lm paste on alumina (courtesy of Jrgen Svenssen, SES). b and
operators and inspectors involved in the actual manufacture of
spacecraft hardware are also trained to high levels of com-
petence and technical skill. Most equipment procurement
specications require that persons engaged in the operation
and inspection of so-called critical processes must possess
certicates indicating that they have been suitably trained.
Skills training courses are held at recognized institutes
throughout Europe and the USA. Final tests must
c The plotted graphs show Quad Flat Packs with SnPb coated leads
attached to pcb tracks having either SnPb or Au/Ni nishes: bonding
materials are SnPb solder alloy and two ECAs: b the effect of storage in
humidity on joint resistance, and c the effect of storage then thermal
cycling between 55 and +100 C in vacuum on resistance (Thomas
and Dunn 1999)
demonstrate, by means of practical and written examinations,
that attendees have the manual skills and sufcient knowledge
of detailed requirements to perform operations that include
the different types of welding, soldering, crimping, etc.
Diplomas in welding are granted to successful candidates,
at both engineer and operator levels, by the national welding
institutes of all major countries. Such facilities are not
widely available for the teaching and skills-training of the
6.14 Training and Certification
assembly processes necessary to companies building and
inspecting spacecraft hardware. This is because the space
market is relatively small and, although there is a need to
re-certify personnel every second year, the total number of
potential students is not high.
6.14.2 Certification for Electronic Assembly
Techniques
The training and certication of personnel involved in European
spacecraft electronic assembly techniques is an important
activity that reflects the changing requirements and technologies
of the space industry. Training courses have been organized for
mixed groups of inspectors and operators from more than 30
countries since 1973; the initial European courses were insti-
gated following an Information Exchange Programme orga-
nized by NASA and ESRO (now ESA) when engineers attended
a wide range of practical training courses at TRW, in California,
in preparation for joint NASA-ESA projects such as Spacelab
(the follow-up to Skylab) (Dunn 1973). Those rst courses
followed European Space Agency hand-soldering requirements
for conventional components onto PCBs. A scheme for the
training and certication of operators, inspectors and course
instructors (within industry) has been established (Dunn 2008).
The aims of the European training are to:
Develop a pool of personnel trained in the best practice
for electronic assembly
Develop a network of organisations capable of providing
training to appropriate standards, and
To improve the quality of electronic assemblies used in
space projects.
Courses are held to cover the following European and
(equivalent) NASA process standards (see also Appendix 9
for titles) as well as a range of IPC courses:
Machine, wave soldering, ECSS Q-ST-70-07: NASA-STD
detail in 8739.3 (Dec 1997)
Manual soldering, ECSS Q-ST-70-08; NASA-STD 8739.3
(change 2, 2011)
RF semi-rigid soldering coax cables, ECSS Q-ST-70-18;
NASA-STD 8739.3 and 8739.4
Crimping, ECSS Q-ST-70-26; NASA-STD 8739.4 (change
6, 2011)*
Repair and modication, ECSS Q-ST-70-28; NASA-STD
no equivalent*
Wire wrapping, ECSS Q-ST-70-30; NASA-STD 8739.4
(change 6, 2011)
Surface-mount technology, ECSS Q-ST-70-38;
NASA-STD 8739.2 (cancelled in 2011)*
Optical Fibres, ECSS Q-ST-70-51; NASA-STD 8739.5
(change 2, 2011)
417
Polymeric application on electronic assemblies, no ECSS
standard; NASA-STD 8739.1
The NASA standards marked * have been transferred to
the requirements of the Space Applications Electronic
Hardware Addendum to IPC J-STD-001E Requirements for
soldered electrical and electronic assembliessince 5th
September, 2012.
Six European Centres perform training in a variety of
languages:
DenmarkHytek, Aalborg
FranceInstitut de Soudure, Paris
GermanyZVE and IFE, both located in Oberpfaffenhofen
ItalyItalian Institute of Welding, Genoa, and
UKUniversity of Portsmouth,
Details of particular courses can be found on the websites
of these schools. Training programmes follow the contrac-
tual requirements of ESA and other agencies quality
assurance and processes specications. The training pro-
grammes provided by these six institutes are very similar but
are held in different languages. They meet ESA-Approved
Skills Training Schools document ESA STR-258 (Novem-
ber, 2008). A common level of training has been agreed for
Category 2 (Inspectors) and Category 3 (Operators) and
all schools have similar course formats. The end-of-course
practical and theoretical examinations are continually
harmonized. It is common that the listed schools also
provide courses organized by the IPCseveral of the
European schools teachers are also Master Instructors per
ANSI/IPC-A-610.
Corresponding courses to the NASA STD process
requirements have been held since the 1960s at:
NASA Jet Propulsion Laboratory, Pasadina, California, and
NASA Goddard Space Flight Center, Greenbelt, Maryland.
A Training and Certication Board now supervises the
approved schools, reviews the training programmes, class-
room facilities, and lesson plans. Numbers of attendees to
the European schools courses now exceed 1000 per year
and all individuals will have a range of backgrounds. They
include Ph.D.-holding scientists producing space experi-
ments; engineers aiming for a career in space design and
manufacturing; technicians producing electronics for satel-
lites, probes and launch vehicles such as Ariane 5 and Vega;
and persons from other industries wishing to achieve a
higher standard of workmanship (Sampson 2007). Of par-
ticular importance are the training courses for Category 1
personnel (i.e. Instructors and Examiners for any of the
ECSS processes certications). This course will involve
either two or three weeks attendance and has been stan-
dardized to:
418 6 Failure Analysis of Electrical Interconnections and Recommended Processes
Week 1: skills training in soldering at operator level (can
be waived if personnel have a current Category 2 or 3
certicate for soldering per ECSS-Q-ST-70-08);
Week 2: training for remaining solder-related processes
(wave soldering, RF cable interconnection, repair of
spacecraft electronics, and the assembly of
surface-mounted devices, following the respective pro-
cess specications ECSS-Q-ST-70-07, 70-18, 70-28 and
70-38). Advice is also provided concerning the selection
of materials for space applications, effects of vacuum and
thermal cycling on electrical systems, suitable cleaning
methods/controls, and preferred conformal coatings;
Week 3: solderless connections including crimping
(70-26), wire wrapping (70-30), and a review of the main
Quality Assurance requirements. Much emphasis is given
to course presentation and teaching methods during
this nal week, together with a mandatory 20-min
demonstration of teaching abilities per person.
It is clear that the Category 1 courses have beneted from
the considerable amount of knowledge accumulated by the
participants. Some attendees were experienced operators
transferred from production jobs to training, but had little or
no experience of, or skills in, training. Some belonged to
quality circles. Others had experience either in training or in
related activities within or outside the space business, but little
or no practical operating skills. On completing the training, all
will have gained some knowledge of the metallurgical factors
that control the reliability of these electrical connectionsthe
causes of good or poor solder wetting, flux reactions, galvanic
corrosion, intermetallic growth, solid-state diffusion and cold
welding at mechanical interfaces, and the like. Attendees to
these courses have beneted from the extensive practical and
theoretical knowledge of both the instructors as well as other
trainees from different companies and institutes. Such inter-
actions will certainly be of use when they return to their own
companies and need to introduce new interconnection tech-
nologies, address workmanship issues or organize their own
in-house training courses.
6.14.3 Understanding Process-Induced
Failures and the Importance
of Workshops
A brief mention is now made to the importance of under-
standing how design, materials, and processes affect the
quality of the product. Skills training schools should possess
physical examples of failed equipment, obtained from
industry or electrical appliance repair stores, which can be
used to illustrate case histories and be available to pass
round the classroom as subject matter for discussion pur-
poses. A bonus is if these defective circuit board assemblies,
flux-corroded packages, etc. can be compared to similar real
hardware that highlights excellent quality products. Simi-
larly, actual assemblies should be available to be used as
workmanship standardsthese will represent conditions
where trainees can distinguish between acceptable and
rejectable hardware during inspection sessions. Discussions
can also refer to photographs of workmanship boarder-line
cases such as the example given in Fig. 6.60.
Each soldering or crimping process must be considered as
part of a closed-loop, where inspection results can be fed
back for possible design or materials changes. Note that
some changes to the materials or processes must be agreed
by the customer, as process qualication status may become
invalidated by such changes, even if they are thought to
improve the product. The writer recalls how an engineer
authorized a slight change to the manufacturing route of a
particular insulated cable that was found to be difcult to
strip. The change was made to the automatic line where
spools were feeding stranded silver-plated copper conductor
wire into a machine which applied the Tefzel insulating
jacket. Simply, just before entering the insulating machine,
the wire was made to pass over a sponge that had been
soaked in silicone oil. The end result was a wire jacket with
excellent stripability. Solderability problems in industry
alerted quality personnel about the situation. The process
change caused the wire product to become noncompliant
with its original space qualied process identication
documentation. Unfortunately, wires from this contaminated
batch of production had been installed on the harnesses of
three spacecraft where assembly work had utilized crimping
processes. Laboratory investigations found the silicone oil to
have a low purity and a very high outgassing-under-vacuum
value. Costly reworking operations were instigated to
remove all the affected wires and cables from these space-
craft. Inspectors are permitted to request authorized rework,
repair and modicationscompilations of approved meth-
ods exist in some standards (e.g. ECSS-Q-ST-70-28).
However, they must know the underlying cause of the
problemthis may be related to poor-quality circuits,
inadequate solderability, contamination, or result from
operator error and bad workmanship. The problem might be
found in the contour of a solder llet. This is important, as
the contour reveals the quality of wetting, and wetting is the
most important physical attribute of the connection in
determining its strength and reliability. The question will be:
to leave as is, or to rework? Solder llets that have a
non-ideal texture and lustre will not affect reliability. To
rework cosmetically inferior connections can cause severe
damage to circuits (as was discussed in Sect. 6.12), and the
correct answers to such questions will be one of the benets
of good inspector training.
Electrostatic discharge (ESD) is an important topic to
cover during training. There are dedicated courses for ESD,
but as electronic components are becoming smaller and
faster than before, they are becoming increasingly
6.14 Training and Certification
susceptible to damage by ESD. Grounded wrist bands must
be worn by all persons handling all space PCB assemblies.
Individual ultra-sensitive devices will need special precau-
tions, workstations must be grounded and a wide range of
other precautions should be undertaken as listed in Agency
standards such as the excellent NASA JSC technical stan-
dard 66552, dated 1st May 2013. It is good practice to have
static shielding bags available for training purposesthese
are often made of polyethylene to which an antistatic com-
pound has been added prior to bag extrusion.
As a means of keeping the training schools instructors
and the space community up-to-date concerning failure
modes, new materials, novel processes and the like, the
author has established an annual workshop together with the
University of Portsmouth entitled: Electronic Materials and
Processes for Space (EMPS). The venue for these workshops
rotates from one school to another in order to allow each
school to present the ndings of their research, promote the
schools facilities and address the needs of STEM activities.
The presentations can be seen and downloaded from the
EMPS website since its inception in 2010. The schools also
publish the results of their own work in a language appro-
priate for that countryfor example, the extensive work of
Moliterni (2010, 2012) is based on the experience of the
Italian school, gained from performing courses, from
research, from a close contact with industry and solving
problems in its failure analysis laboratory.
6.15 Verification of Surface-Mount
Technology and Prevalent
Failure Mechanisms
6.15.1 Verification Testing
An introduction to the design concepts and manufacturing
procedures that affect the reliability of surface-mount tech-
nology (SMT) have been described in Sect. 6.5.
Surface-mounted components have caused problems when
they were rst assembled onto spacecraft circuits. Because a
failed electrical connection can impair the performance, or
destroy a major part or function of a spacecraft, this tech-
nology has been identied by some approving authorities as
a so-called critical process. The main steps that are usually
required prior to the approval of SMT for spacecraft, and
probably other hi-rel applications, will now be covered.
Examples of common failure mechanisms will be presented
in order to illustrate the importance of understanding the
broad range of environmental stresses that affect SMT
reliability.
Soldering processes related to the assembly of conven-
tional leaded components onto through-hole PCBs follow
well-established design rules which are covered by
419
specications prepared by ESA, NASA, and IPC (1996).
These have, in the main, been fully approved by the agencies
procuring such electronic hardware. The processes are for-
mally approved by means of the space projects Declared
Process List (DPL), and all that needs to be ensured are that
the written procedures are followed and that the assembly
operators and inspectors have been suitably trained and
certied (as was described in the preceding section). Qual-
ication of hardware will be made on the spacecraft sub-
system or unit, by exposing electronic boxes and the like to
the range of environmental testing dened by each particular
spacecraft contract.
Processes for SMT solder attachments are not so well
understood as those for leaded components. Until recently
they were considered as nonevaluated processes where no
information existed concerning their use in a space envi-
ronment and for which no useful data concerning reliability
and performance could be given to the project in question.
The most important difference is that the SMT processes do
not always provide for sufcient stress-relief between the
materials being interconnected and therefore accelerated
reliability testing should always be required to determine the
actual failure modes/mechanisms and provide data for which
the space hardwares service life can be calculated. It has
already been shown that future spacecraft circuits will utilize
large numbers of leaded and leadless surface-mounted
devices (SMDs), as these offer a greater opportunity for
higher-density electronics with improved electrical perfor-
mance and associated weight savings. Documents such as
the ECSS-Q-ST-70-38 and IPC J-STD 001 (with space
addendum) contain the main design rules that should be
followed if spacecraft SMD assemblies are to survive the
environments of storage, launch, and operation. It is usual
that any electronics manufacturing company proposing to
build spacecraft equipment that incorporates SMT will be
required to demonstrate the reliability of the SMT by means
of verication testing on test samples. If the technology
involves completely novel processes the procuring agency
will normally request so-called technology samples which
will be assessed by that agency for any major problems
associated with materials selection, wrong designs, or
incompatible metallurgical processing. When these tech-
nology samples are found to be acceptable, the new SMT
process line might be subjected to a systematic and inde-
pendent examinationdened by ISO 8402 as an audit
(here the company quality procedures will be checked for
suitability to general product assurance requirements levied
on the contractor by the procuring agencywhich will
include an on-site review of process documentation, design
documentation, parts storage control, training and certica-
tion procedures, suitability of processing equipment, clean-
ing methods and their controls, whether all materials
including PCB types, solder alloys, staking compounds,
420 6 Failure Analysis of Electrical Interconnections and Recommended Processes
conformal coatings are suitable for space use, etc.). At this
stage it will be important to establish that a Process Identi-
cation Document (PID) exists within the companythis is
a summary document that lists the range of electronic parts
associated with the SMT, lists the processing equipment, and
in particular lists the materials and all of the process pro-
cedures associated with the SMT line.
It is important that the contents of any SMT verication
programme is established by the manufacturing company
and approved by the customer prior to the commencement of
testing. Here, the number of verication samples will be
agreed, together with a complete description of the test
environments and a denition of what constitutes a test
failure. A verication sample PCB assembly is seen in
Fig. 6.69a, the assembly is conformal coated and has some
electrical monitoring of parts. A failure criterion may be the
rst interruption of electrical continuity even when it is of
very short duration. Only the continuous monitoring of
electrical properties during environmental testing will give
an accurate and repeatable indication of when test failures
occur. This is because short electrical interruptions are often
noticed when the test circuit is operating at the test tem-
perature extremes. The failure mechanisms and their rates of
occurrence should correspond closely to the failure mecha-
nisms and thresholds that are expected to be seen at the
actual end of life of an operational spacecraft. All veri-
cation test samples must be manufactured according to the
PID. These component-assembled circuit board samples
should include all the design congurations, materials, sol-
dering fluxes, and the cleaning methods which are envisaged
for future spacecraft hardware. If staking compounds, ther-
mally conductive adhesives, and conformal coatings are to
be applied to flight hardware, it will be essential that these
same materials are present on the verication samples. Each
of these materials will have different physical properties and
might contribute to a particular failure mechanism.
Thermal cycling will generally follow a specication
such as that proled in Fig. 3.12b and SMT verication will
require that test samples survive at least 500 cycles (for
soldered connections that cannot be readily inspected, such
as area grid arrays, the number of cycles may be 1500).
Limits for the permitted extent of solder joint cracking, or
electrical failure are given in ECSS Q-70-38. Vibration
testing is usually performed in two axes, one parallel to the
board and the other perpendicular to the board. Vibration
levels related to the spacecraft requirements and the char-
acteristics of the intended launch vehicle. A typical set of
vibration test levels would be:
Sine vibration
Vibration amplitude: (peak to peak) 1070 Hz at 1.5 mm
Frequency range: (702000 Hz (CPS) at 15 g
Sweep speed: 1 octave per minute
Duration: 1 cycle from 10 to 2000 to 10 Hz
Random vibration
Frequency range: 202000 Hz (CPS) at 15 g (RMS)
Power spectral density: 0.1 g21 Hz
Duration: 10 min per axis
It is emphasized that companies and institutes should
remain in close contact with the qualication authority (or
customer). Technical standards/requirements can be modi-
ed, or drastically changed as new modes of failure are
identied. Materials may become obsolete or project envi-
ronmental test requirements may be made more or less
demanding. In Europe, the ECSS Q-ST-70-38 requirements
are continually assessed by a working group, so it is clear
that latest issues of such documents must be obtained or
freely downloaded from the internet.
A laboratory report containing all visual inspection
ndings, electrical test results, and micrographs will be
compiled, but this may be quite voluminous and it is
therefore useful to also provide the customer with a sum-
mary table of the SMDs veried to the test specication.
This table should include all important information,
including:
the address of the manufacturing line
the identication of the test specication
full reference to the PID
assembly process, e.g. vapour phase, hand soldering,
etc.
solder alloy, e.g. 63Sn, 62Sn, 50 indium lead, etc.
conformal coating type, e.g. no coating, Urelane
5750LV, etc.
each part should be listed according to SMD class, type
number, body dimensions (in mm) and identify any that
were staked
the PCB substrate should be well dened together with
its coefcient of expansion value and glass transition
temperature (see also Appendix 1 for specic values)
state the full reference of the company verication report.
The author is particularly indebted to the companies who
agreed that summaries of different SMT failure mechanisms
could be recounted in the following sections. These failures
were experienced during verication testing. Each condition
for failure is different. It is the authors opinion that it is
unwise to predict the service life for any surface-mounted
device from formulations based on either the CofnManson
equation [i.e. where the logarithm of the life is linearly
proportional to the fatigue ductility exponent of the solder
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms 421

Fig. 6.69 a Verication PCB

with assembled SMT components
after undergoing several tests
such as thermal cycling, vibration
and rework simulation. Post-test
visual inspections and
microsections are made to
evaluate the quality of the solder
joints. b External and
c microsectional views of a
cracked ceramic chip capacitor
due to vibration. Note the
wide-mouthed crack and the
mechanical displacement between
the silver leaves. The large pore
within the solder llet may have
contributed to the damage, but the
main factor was an inadequate
xation of the thin (1.6 mm)
double-sided PCB to its dummy
housing

alloysee, for instance, Lau and Rice (1985)], or the vari-
ous Engelmaier expressions. These mathematical solder
fatigue models should be used with great caution because
they rely on precise physical and mechanical data associated
with each of the materials making up the soldered inter-
connection. As will be seen in the following sections,
spacecraft SMT can unfortunately involve imprecise mate-
rial processes. Because of the wide number of variables it is
unreasonable to try to predict a solder joints structural
integrity and reliability from models based on assumptions.
Practical thermal cycling experiments and verication tests
on representative samples should always be recommended.
422 6 Failure Analysis of Electrical Interconnections and Recommended Processes
Calculated predictions for SMT thermal fatigue life will be
incorrect if the following examples of inconsistencies exist
within the hardware:
residual stresses in the laminate
the solder alloy becomes hardened by leaching from
adjacent copper or silver surfaces, for indium solders
leaching may be from gold plate
intermetallic layers at solder interfaces
variable stand-off heights (i.e. distance between the PCB
pad surface and the component termination that is lled
with solder; see particularly Fig. 6.24)
presence of different organic materials beneath the
component body, such as staking compounds, thermally
conductive adhesives, and conformal coatings (note that
the thermal expansion coefcients and glass transition
points of these organic compounds vary from batch to
batch)
for leaded SMDs, a variation in the Youngs modulus
and yield strength of the lead material, and residual
stresses from coplanarity problems or hot bar soldering
general variability in the solder joint such as voids, llet
geometries (especially with large solder llets on one
side of a component and small llets on the opposite
lengthas seen in Fig. 6.25), and misalignment on ter-
mination pads.
6.15.2 Failure Under Mechanical Overloading
Mechanical shock is a possible accelerator of failure with
SMT. It may occur when an assembled PCB is accidentally
dropped onto a hard surface or in operation if the circuit is
situated close to a pyrotechnic actuator device. Certainly the
ceramic packages will be prone to such accidental damage
but only slightly more so than conventional leaded compo-
nents. The failure of SMDs due to mechanical shock has not
been encountered by the author. However, a common source
of failure has been the detachment of SMDs from their PCBs
during the vibration test environment which formed part of
verication testing.
SMDs with ne-pitched leads such as quad flat packages
(some down to 500 m pitch) are the most frequent to suffer
from vibration-induced failures. The leads and interconnects
flex, become work-hardened, and nally fracture. Other
parts seen to dislodge or become damaged during vibration
testing are leadless chip carrier packages and ceramic chip
components such as that seen in Fig. 6.69b, c.
Recommendations for avoiding mechanical overloading
of SMDs are:
whenever possible, rigid multilayer PCB designs should
be employed; thin PCBs need to be secured at short
distances to stiff brackets in order to avoid resonance,
low natural frequencies and high amplitudes
position heavy parts close to the edges of PCBs and
xation points
use staking compounds and conformal coatings to damp
the board and increase mechanical support.
It is emphasized that particular attention must be given
whenever engineers contemplate using automatic electrical
test equipments. Here, computer software manages some
electrical function checks on the assembled board by
directing long, spring-loaded metal probes onto predeter-
mined pad terminations. It is intended that the sharp probes
should descend onto each selected pad in close proximity to
the soldered SMD llets but without impacting the device.
Close inspection of automatically tested assemblies conrms
that when the components are exactly positioned onto their
termination pads, a series of penetration marks is present on
the pad. Ideally, these impressions should be made into the
tinlead coating well away from the soldered device. In
practice, the SMDs show some misalignment on their pads
and this, coupled with the inaccuracies of the test equipment,
causes indentations to be made either onto the sloping side
of a solder llet or directly onto the metallization area of the
ceramic component. Probe marks on solder llets are clearly
visible. However, these indentations are hardly discernible
when they occur on the components hard metallization. The
major problem here is that the high loading force that is
applied to the probe tips will damage many of the mounted
components. This was recently witnessed on several flight
boards which had been automatically tested without proper
authorization (i.e. this inspection method was not part of the
approved Process Identication Document). Failed devices
were microsectioned and clearly showed that brittle cracks,
which passed throughout the complete thickness of the
component, had been initiated at the point where the probe
had made contact with the metallized surface.
6.15.3 Failures Due to Board Flatness
Problems
The required nish for the majority of spacecraft PCBs is
tinlead on the external copper conductor patterns. The tin
lead has a neareutectic composition and is initially elec-
troplated onto the copper and later fused by either immersing
the complete circuit board into hot oil or melting the elec-
troplated layer by infrared heating (see also Sect. 4.7). Most
board distortion problems occur during this heat cycle.
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms
The base laminate is heated above its glass transition tem-
perature (Tg Cthe temperature in C when the material
changes from a glassy state to a plastic state) and the epoxy
or polyimide matrix becomes soft, causing a relaxation of
the support medium which is usually woven glass bre or,
occasionally, woven quartz bre.
Another form of board distortion occurs when the board
is suddenly heated when an assembly is passed over a solder
wave during machine soldering. In all instances it is
important to pre-heat to avoid thermal shock. In a similar
manner, immersion of a hot circuit board into a cold cleaning
fluid can also cause distortion. The soldered components will
hold the assembly in a distorted condition.
Distortion is often referred to as warp or twist, and the
usual requirement when space-quality boards are purchased
is that their flatness should be within a tolerance of 1 %.
Board suppliers will seldom offer a guaranteed flatness of
better than 1 % because of the variability in both matrix
material behaviour and the effect of circuit board design.
Delaminations within multilayer boards may also promote
overall distortions, as was seen in the Fig. 3.18.
If SMDs are mounted onto distorted boards they will be
subjected to severe mechanical stress each time the assembly
is flexed or vibrated. This is particularly a problem when
warped assemblies are straightened during insertion of the
board into a cardguide or when it is screw mounted onto
stiffeners and positioned in an electronic box. The failure
analysis of several cracked chip capacitors determined the
cause of failure as the straightening of a warped multilayer
board assemblyone component is illustrated in Fig. 6.70.
Unfortunately, these types of failure are difcult to observe,
as the cracks are extremely narrow.
PCB distortion can be minimized by taking account of
some simple rules:
The circuit design should ensure that the conductor pat-
terns on either side of the board are evenly distributed in
terms of density and direction; SMDs should be evenly
distributed over the board surface and high package densities
in any one region should be avoided; with multilayer boards
the neutral axis (i.e. location of zero strain during manu-
facture, assembly, and test) should be at the centre of the
boards thickness and this can only be accomplished if the
copper foil thickness and the ground and power planes are
equally distributed on both sides of the neutral axis. Poly-
imide base material is generally superior to epoxy resin as a
PCB matrix for SMT.
The PCB fabricator should procure only the highest
quality of laminate, ensuring that it possesses no residual
stresses; the boards should not be articially restrained
during fabrication processes such as etching, plating, and
423
solder fusing; any thermal treatments should be gradual and
involve dwell times so that the board materials reach thermal
equilibrium; and, distorted boards intended for SMT appli-
cations must not be corrected by pressing at elevated tem-
peratures, as the warp will reappear when the board is
reheated above its Tg C.
The assembly process engineers should also avoid ther-
mal shocking the board during oven bake-outs to remove
humidity prior to soldering or during the actual soldering
operation; flux cleaning should occur only after the assembly
has cooled to ambient temperature. Ductile solder llets of
soft tinlead or indiumlead composition, together with a
suitable stand-off height, are to be preferred.
Inspectors and operators should check for warp and
twist. Suspect assembled boards should be documented and
subjected to a company Material Review Board before they
are mounted onto supporting structures.
6.15.4 Failure Due to Co-planarity Problems
The three most commonly used solder joint congurations
for SMDs are those known as leadless (Figs. 6.25 and 6.26),
gull winged (Fig. 6.18d), and J-leaded (Fig. 6.71). The
gull-wing components have the disadvantage that they
occupy more board space than the other congurations. Their
protruding leads can also become distorted during trans-
portation and handling. Lead coplanarity is the term given to
all adjacent leads along the sides of a component lying in the
same plane. Noncoplanarity is the distance between the
highest and lowest leads when the package rests on a per-
fectly flat surface. When some leads have become acciden-
tally bent their noncoplanarity can result in soldered joints
having widely different solder llets. In order to achieve
joints having a similar visual appearance, the assembler may
be tempted to either press down the package so that all the
leads are spread to make contact with their PCB terminations
during hand-soldering. Alternatively, the hot bar method of
reflow soldering may be usedhere a resistance-heated
metal bar is applied simultaneously to the gull-wing leads of
the SMD. It has been used for components such as standard
flat packs, ne pitch quad flatpacks (QFPs), small-outline
integrated circuits (SOICs), and small-outline transistors
(SOTs). High levels of residual stresses are generated, both in
the component lead and in the solder alloy llet, whenever
the leads are forced against the PCB termination pads and
held there during solidication of the solder.
Solder joint stress-rupture failures caused by the residual
stresses in strong Kovar gull wing leads that were forced to
lie flat against the PCB during the soldering account for
424 6 Failure Analysis of Electrical Interconnections and Recommended Processes

Fig. 6.70 a and b Machine

soldering of this large-sized
capacitor to a multilayer board
(glassepoxy) had caused the
board to warp by about 2 %. The
nonleaded components became
mechanically stressed when the
board was mounted to a stiffener.
The capacitor termination is a
nonpreferred type, as the silver
palladium metallization is not
covered with a nickel barrier
layer. This has caused silver to be
leached from the metallization
during machine soldering. The
microstructure of the solder llet
shows it to be saturated with
silver precipitates. This causes the
solder to be far harder and less
ductile than would be the case if it
had flowed onto a protective
nickel barrier. The component is
also hard-mounted to the PCB
termination. In summary, there
are not stress-relieving
mechanisms available to avoid
component cracking and only a
small degree of board
straightening will damage these
leadless devices

many qualication test failures. During thermal cycling, or
during high-temperature storage, the elasticity within the
lead causes it to move and become displaced out of the
solder. At slightly elevated temperatures the solder alloy has
very little strength and will creep to nally release the lead,
so causing the joint to become electrically open-circuit.
The problem caused by non-coplanarity is to a far lesser
extent seen with J-leaded components where the leads have
become misaligned. Here, the damaged leads are usually
displaced upwards and out of plane. During soldering these
leads may produce very small solder llets. As J-leads are
very difcult to inspect, these defects can go undetected.
Electrical tests and metallography have clearly proven
that noncoplanar leads that contain residual stresses are more
prone to solder joint failure, and after fewer thermal cycles,
when they are embedded in a hard conformal coating such as
epoxy resin or a polyurethane compounds. This is probably
caused by the high expansion of these coatings which causes
the component package to be displaced in the z-direction
(i.e. the same direction as the lead movement).
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms
Fig. 6.71 Verication testing has revealed these Alloy 42 J-leads to
have a good solderability, but because of the high stiffness of this
material, the solder joints failed before 500 thermal cycles. The rst
joints to fail by thermal fatigue are those at the corners (farthest away
from the neutral axis)
6.15.5 Solder Joint Failure Due to Thermal
Mismatch Between SMD
and Substrate
6.15.5.1 Leaded and Chip SMT Devices
This is the most common failure mechanism associated with
surface-mounted devices. The ceramic body of the compo-
nents may have a coefcient of expansion (COE) between 5
and 7 ppm/C. The COE of a glassepoxy PCB at ambient
temperature can range between 11 and 17 ppm/C. When the
component possesses flexible leads they should take up the
COE mismatch between component and PCB materials
during thermal cycling, but this is by no means certain and
verication testing has revealed some problem leads such as
that shown in Fig. 6.71. The SMDs with short stiff leads, or
425
only leadless metallized terminations, require an intercon-
nection system that has a COE closely matched to their
ceramic bodies. COE matching may be achieved in different
ways, as described in Sect. 6.5.
Successful techniques include the use of Invar or
molybdenum plates within the PCB lamination, as these low
COE metals will constrain the natural high expansion of the
glass breepoxy types of laminate. Board laminates con-
taining quartz or other low-expansion bres based on ara-
mides, set in a polyimide matrix may also be purchasedthe
so-called compensated boards are generally expensive, as
they are only produced in small quantities for a specialized
market. Also, the low-expansion additions are relatively hard
and cause manufacturing difculties such as a rapid tool
wear-out during hole-drilling operations. As an alternative to
changing the base laminate, one can select to produce tall
solder jointsas previously described in Sect. 6.15.1
these having a large standoff height, and it is the solder
volume itself that will creep and take up any thermal mis-
match. Specially prepared pre-forms for leadless chip carri-
ers may also be purchased, such as that illustrated and
described in the caption to Fig. 6.72a. Ceramic column grid
array (CCGA) packages having up to 1500 interconnections
(I/Os) are well designed to resist thermal fatigue, particularly
when they are reinforced by the Six-Sigma columns
(Fig. 6.72b). For details of Area Grid Array (BGA and
CGA) packages see Sect. 6.15.10.
Matching the COEs between component body material
and the PCB substrate is no guarantee of freedom from the
problem of thermal fatigue. Some electronic components are
their own heat source and become very warm (e.g. power
transistors and resistors, which can dissipate several watts of
thermal energy). This is particularly troublesome for space-
craft electronics, as convection cooling is not possible and
all heat dissipation from the component must be via heat
conduction through the soldered joints or along paths made
by placing thermally conductive adhesives beneath these
devices. These adhesives can also promote thermal fatigue in
the solder interconnection, as they have a higher COE than
the solder alloys and create stress in the vertical axis. The
use of such adhesives can only be permitted if they were
incorporated into the verication test samples described in
Sect. 6.15.1. Thermally conductive ceramic packages have
recently emerged on the market, such as aluminium nitride
(AIN), silicon carbide (SiC), and beryllia (BeO). These
materials exhibit improvements in thermal conductivity of at
least ve to ten times that of alumina. They have somewhat
lower COEs than conventional alumina packages, so again
care must be taken to avoid the thermal fatigue of any sol-
dered joints.
As a note of caution, thick-lm hybrids might be con-
sidered to be ideally matched because they consist of cera-
mic (alumina) components soldered onto alumina substrates
426 6 Failure Analysis of Electrical Interconnections and Recommended Processes
Fig. 6.72 a 1995 Technology. View showing Raychem solder
columns that have been machine-soldered to interconnect between
the castillations of the leadless chip carrier and the glassepoxy PCB.
The photograph was taken after verication testing. 500 thermal cycles
(55/+100 C) have caused the solder columns to tilt. Each column
contains a copper coil which can distort to take up COE mismatches
without failure. b 2014 Technology. 6-Sigma CCGA 625 packages
assembled to ENEPIG coated PCB by machine soldering. All parts
passed a cumulative qualication test flow: vibrations, shocks and
thermal cycles up to 1500 cycles (per ECSS Q-70-38) without any
electrical failures. Microsections after testing revealed no defects
(Chaillot et al. 2014). Column grid arrays having Six Sigmas
copper-reinforced columns (originally developed by Raychem), are
said to be 600 times more reliable than standard solder balls for area
grid array attachments
having an identical COE. Figure 6.73 shows that this is not
always the case and very localized COE mismatches within
the interconnection itself must be considered, such as the
differences between solder and the adjacent ceramic.
Recent advancements in the production of new materials
which can be tailored to have a range of well-dened COEs
have started to revolutionize the microelectronics, power
semiconductor, microwave and laser diode packaging
industries. Reference has already been made in previous
sections to COE-compensated PCBs (see also the typical
values in Appendix 1)these PCB materials help to reduce
the strains that cause thermal fatigue failure of soldered
joints. Similar advancements have led to new metalmatrix
Fig. 6.73 Part of a thick-lm hybrid circuit after thermal cycling. The
ceramic capacitor is soldered with indiumlead alloy to gold conductors
red onto a glassalumina dielectric. The very localized COE mismatch
between the gold-saturated indiumlead solder alloy (32 ppm/C) and
the alumina component (6 ppm/C) has caused sufcient strain to crack
and detach its metallization layer
composite (MMC) materials for large-size component
packages having dimensions from 10 mm to more than
100 mm. These MMCs are tailored to match the COEs of the
PCB substrate, whether it be made from glassepoxy
material or even alumina. COE matching is achieved by the
addition of reinforcements, which are mainly carbon bres,
boron bres, or silicon carbide particles, to the matrix of
commercial alloys based on aluminium. The common metals
presently selected for microwave and hybrid microelectronic
packages are made from Kovar (COE = 5 ppm/C), Alloy 42
(COE 4.4 ppm), aluminium (23 ppm), and copper (17 ppm).
Low-expansion MMC packages based on silicon carbide
reinforced aluminium are now available with COEs ranging
from 6 to 14 ppm/C. The low-density aluminium matrix
gives these MMCs a 3:1 weight advantage over Kovar and
Alloy 42. A typical microstructure of such a package
material is shown in Fig. 6.74 and described in that caption.
6.15.5.2 Presence of Thermally Conductive
Adhesives as Staking Compound
Excessive heat generated by power dissipation in compo-
nents can threaten the reliability of spacecraft electronic
assemblies and in such cases a high heat extraction is nee-
ded. In vacuum, forced air cooling is not possible. Therefore,
if the heat dissipation through the component lead termina-
tions and by radiation is not large enough, it will be nec-
essary to apply thermally conductive adhesives (staking
compounds) between the package and PCB. It is probable
that Area Grid Arrays (see Sect. 6.15.10) will not require
extra conductive sandwiched between these packages and
board due to the presence of highly conductive balls and
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms
Fig. 6.74 Evaluation of a commercial metal-matrix composite (MMC)
electronic package based on silicon carbide particle reinforcement and
an aluminium alloy matrix. a Shows an X-ray micrograph of the 50 mm
by 25 mm package and its lid showing no evidence of porosity, particle
density variation across the parts, or cracking. Hermeticity was
conrmed by helium-leak testing, b shows the pseudo-isometric
microstructure of the as-polished MMC in the planes longitudinal
(L), short transverse (ST), and transverse (T) to the longest package
dimension. These micrographs conrm that this MMC is fully dense,
contains an even distribution of the various SiC particle sizes, and
indicates isotropic physical propertiesl. c Shows details of the various
columns. However, ne pitch leaded packages have minimal
thermal conductive paths and often need additional heat
sinking. Thermally conductive adhesives, and similar com-
pounds used for staking (to provide more resilience during
vibration) will have different thermal expansion properties
than the adjacent materials of component and PCB. This
may cause stresses in the joints during thermal cycling. The
hardness of the adhesive is also important since a harder
adhesive is more prone to transfer the stress.
A review of the market for thermally conductive adhesives
supplied to the aerospace industries produced 35 products that
were thermally conductive but not electrically conductive.
23 products were both thermally and electrically conductive,
15 were encapsulants and 4 lm products were available.
These have been compiled into a report that justies selection
427
microconstituents of the MMC after slight etching of the polished
section. The angular SiC particles range in size from 4 to 200 m
(10 m average). The particulate volume fraction is approximately
60 %. The matrix is composed of aluminium, magnesium, silicon, and a
trace of silver together with the coarse A and C phases (analysed by
ED AX to be compounds of MgSi and without the presence of
undesirable aluminium carbides). This MMC package is one third of
the weight of a similarly designed Kovar packaged with the same COE.
In addition, the thermal conductivity of this MMC material is far
superior to that of Kovar (100 vs. 17 W/mK) (micrographs courtesy of
Mr. G. Ramusat, ESTEC, Noordwijk)
of certain types of material based predominantly on
outgassing-under-vacuum data (Tegehall and Dunn 2001b).
Three products were initially selected: Epo-tek 930,
CV-2946 and Multimat M-4030LD. The latter product
proved impossible to form into the kind of mounds neces-
sary to meet ECSS requirements and was abandoned. The
test vehicles were leaded ceramic quad flat packs (CQFPs)
with 144 I/Os soldered onto polyimide-glass PCBs.
The CQFP lead-frames were alloy 42, with solder-plated
gull leads having a 0.65 mm pitch. In cross-section the
samples appeared as in Fig. 6.75.
Epo-tek 930 (two part epoxy system with boron nitride
particles as ller) was stated in the trade literature to have a
thermal conductivity of 4.2 W/mK, (also conrmed by lab-
oratory testing).
428
CV-2946 is a two part silicone material, also containing
boron nitride, was stated in the trade literature to have a
thermal conductivity of 3.8 W/mK, but laboratory tested to
be only 1.7 W/mK.
Outgassing under vacuum, per ECSS Q-70-02 gave:
Epo-tek 930 TML 0.01 %; RML 0.01 %; CVCM 0.00 %
CV-2946 TML 0.36 %; RML 0.08 %; CVCM 0.00 %
Coefcient of expansion tested to ASTM E-831-93 gave:
Epo-tek 930 26.3 ppm/C (50 to +50 C) and 90.93 ppm/
C (100150 C)
CV-2946 209.7 ppm/C for 50150 C
Hardness (Shore A) Epo-tek 930 = 98, CV-2946 = 75.
All assembled PCBs were verication tested to ECSS
Q-70-38 (thermal cycling then vibration.
All samples were noted to pass these tests with no
adhesive-generated damage inflicted to the soldered joints
(following visual and microsectional examination). Both can
be suggested for space-use.
6.15.6 Conductor Track Failure Due
to Thermal Mismatch
Very narrow conductor tracks having a width of only
0.3 mm were seen to crack and generate electrical
open-circuits during the thermal cycling tests that formed
part of verication testing for SMD transistors (type
NE41607) and diodes (type HP 5082-3141). These
Fig. 6.75 Macro-section to show stand-off, form of bent leads and the
adhesive (white) which had been introduced through a hole drilled into
the PCB at the centre of the QFP
6 Failure Analysis of Electrical Interconnections and Recommended Processes
components were solder mounted (63Sn solder) without any
form of stress-relief onto electrolytic copper pads. The board
material is a form of PTFE (polytetrafluoroethylene) that is
mechanically attached to an aluminium base plate. Extensive
cracking was seen after 500 thermal cycles. The cause of
these cracks is thought to result from a high COE mismatch
between the ceramicKovar component materials (5 ppm/
C) and the PTFE-aluminium substrate (about 22 ppm/C).
This mismatch has stressed the copper conductor such that it
became work-hardened and tore at the point where there is
an abrupt change in section between the pad and track as
seen in Fig. 6.76.
Circuits were later manufactured which passed all veri-
cation tests. These supported three modications: use of the
soft 50 % indium50 % lead alloy; a slight stand-off between
the ribbon lead and the copper pad to allow for stress-relief;
and a conductor track having a slightly wider dimension of
0.5 mm.
6.15.7 Failure of RF Cables Connected
by SMT
The reliability of a spaceborne radio-frequency (RF) system,
such as employed for S-band communications, gave cause
for concern when soldered interconnections made to a SMT
verication sample failed during exposure to thermal
cycling. The substrate was PTFE bonded to an aluminium
support. Part of one SMT verication test board is seen in
Fig. 6.77. Whilst all the components survived the test
environment, the semi-rigid RF cables failed because of cold
flow and thermal expansion of the PTFE dielectric that is
present surrounding the cables copper centre conductor and
within the copper external tubing. The failure mechanism is
well known and has been discussed in Sect. 6.5.5 together
with the metallographic illustrations of Fig. 6.27. Remedial
actions were to precondition the cables by thermally cycling
them for several cycles between 45 and +100 C until there
was no further cold flow of the PTFE dielectric. The
extruded cable was mechanically cut flush with the copper
sheath and discarded before forming the centre conductor
and soldering it to the rectangular pads on the circuit board.
Coaxial cable-to-stripline junctions soldered after this pre-
conditioning heat-treatment no longer gave anomalous
results during further verication tests.
6.15.8 SMT Solder Joint Failure Due
to Conformal Coatings
The ne lines and close conductor spacing utilized in
spacecraft SMT may necessitate that conformal coatings are
always applied to the nal, cleaned PCB assemblies.
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms
Fig. 6.76 Thermal fatigue
cracking of copper conductor
located between parts having a far
lower coefcient of expansion
The main purposes of conformal coatings are to provide
protection against airborne contaminants and humidity dur-
ing ground-handling operations and, with the exception of
polyxylene, to ensure that the components are mechanically
stable during exposure to shock and vibration under launch
conditions. In service coatings will prevent shorting due to
conductive contaminants floating and resting between con-
ductors in zero gravity and they mitigate against tin whisker
growths. Conformal coatings can be applied by the immer-
sion of PCB assemblies into uncured coating, but brushing
and spraying are the more usual processes as they have more
chance of avoiding the lling of stress-relief bends and can
achieve thinner coverages (weight-saving). Polyxylene
(Trade name Parylene) is applied to assemblies by a vacuum
processit has advantages being very thin, impervious to
moisture, low outgassing, and covers all surfaces; however,
it provides no mechanical support to mounted devices (some
429
need support during exposure to vibration and acoustic
noise) and it is almost impossible to remove for circuit repair
without resorting to mechanical abrasion.
The relative differences in the expansion and contraction
of the PCBs, the component parts, and the incorrect selection
of a conformal coating applied to these assemblies have been
seen to produce catastrophic failures to both operational
spacecraft in orbit and flight-quality electronics that were
undergoing qualication tests. Actual failures are seldom
disclosed in the scientic literature. However, a detailed
analysis of the cause of an open-circuit solder joint discov-
ered on the Magellan spacecrafts radar digital unit was
made by the Jet Propulsion Laboratory, which led to the
successful repair of many pieces of affected hardware on that
vehicle as well as the Galileo spacecraft. The problem was
caused by the presence of a high coefcient of expansion
coating (Solithane 113) located in the large gap (1.3 mm)
430 6 Failure Analysis of Electrical Interconnections and Recommended Processes

Fig. 6.77 a General view of

SMT verication sample
highlighting b the fractured solder
lap joints associated with the
central conductors of RF cables
that suffered from extensive cold
flow during thermal cycling

between Dual Inline Packages (DIPs) and the PCB surface
(Ross 1989; Lee 1990). Similar problems with space hard-
ware have been traced to damage caused by conformal
coatings during thermal cycling (ASTROS star tracker,
Magellan data formatter, etc. reported in JPL Lessons
Number 275).
Previous sections (e.g. 6.54) have demonstrated the pref-
erence of having suitably high stand-off distances between
the solder terminations on surface-mounted devices (SMDs)
and PCBs. These stand-offs facilitate cleaning operations, as
flux residues can be removed more easily from beneath the
component than when the component lies flush against the
PCBs surface. They also provide for a sandwich of soft,
ductile solder between the terminations which is able to creep
when there is a mismatch in coefcient of expansion (COE).
This creep of the solder alloy gives some stress-relief
when the component is subjected to environmentally
imposed thermal-cycling conditions or when the component
dissipates thermal energy under spacecraft operational
conditions.
Several evaluation programmes have been undertaken to
explore the concern that high COE coating materials in the
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms 431

Fig. 6.78 Leadless chip carrier coated with Scotchcast after 1000
thermal cycles. Electrical open circuits occurred at the 400th cycle.
Overall view shows some detachment of Scotchcast from upper side
face of ceramic package and some delamination from PCB surface
(arrows). The castellations numbered 26 support failed joints but,
even with the best inspection optics, the llets of these soldered joints
cannot be seen to contain cracks which account for the electrical open
circuits (from Dunn and Desplat 1994)

stand-off volume beneath SMDs might induce damaging

stresses (Lee 1990; Dunn and Desplat 1994). It was con-
cluded that epoxy coatings such as Scotchcast 280 and the
polyurethanes Solithane 113 and Conathane EN11 should be Fig. 6.79 Optical micrographs made parallel to a row of solder joint
restricted from SMT applications owing to the problems castellations (shown as the cutting section, X-X in Fig. 6.78). Scotchcast
associated with COE mismatch and solder joint cracking. after 1000 cycles. All castellations contain cracked soldered joints
The series of captions to Figs. 6.78, 6.79 and 6.80 account
for the electrical open circuits experienced on some coated candidate for SMT. In contrast, the glass transition temper-
leadless chip carrier packages. The coatings expanded suf- ature of Solithane 113 and Humiseal 2874 as tested by Lee
ciently at the elevated temperatures to thermally fatigue (1990) is 7 and +5 C respectively. The Tg temperatures of
these solder joints and cause open circuits. At room tem- these coating materials are near to room temperature; there-
perature and below, the contraction of the coatings caused fore they will impose dramatic stresses on neighbouring
the cracked surfaces of the joints to become pressed solder joints because each temperature cycle will pass
togetherelectrical continuity was restored as in the beha- through their Tgs. This is illustrated by the following list of
viour of a thermal switch. selected material properties:
Coatings that can be recommended for spacecraft SMT
were determined from a number of criteria, including the ease Coating type Tg (C) COE (ppm/C)
in which they can be repaired. A summary of results is shown Below Tg Above Te
in Table 6.8. The more suitable coatings where the silicones Uralane 5753LV 65 91 218
CV 11400 and MAPSIL 213; these have a low glass tran-
Solithane 113-300 7 98 247
sition temperature (Tg lower than 100 C) which is below
Humiseal-2B74 +5 88 225
the normal range of temperature-cycling tests. These silicone
materials are also rather soft and can be considered as com-
pliant whenever COE mismatches are encountered (i.e. they Brief mention is given to the fact that hard coatings with
impose low stresses on soldered joints). The polyurethane large COEs can also cause small leadless ceramic compo-
Uralane 5750LV also has a low glass transition temperature nents such as capacitors and resistors to crack during thermal
(Tg being 65 C) and this coating also appears as a suitable cycling. One obvious failure is seen in Fig. 6.81.
432 6 Failure Analysis of Electrical Interconnections and Recommended Processes

Total

22
33
36

42
43

45

51
spacecraft

l = good
Manned

4 = fail
1
1
4

1
4

4
Outgassing

l = good
7 = fail
1
6
4

7
1

4
Resistance to thermal cycling environment
Microwire
bonds

1
2
4

3
5

7
Table 6.8 Summary table of results (from Dunn and Desplat 1994) with listing of conformal coatings according to an order of merit

components

7 = poor
l = best
Leaded

1
1
4

1
5

6
Surface
mount

1
1
1

1
1

7
7 = poor
l = good
Life test

1
4
1

5
7

1
Humidity

7 = poor
l = good

1
3
6

3
7

5
Fig. 6.80 Microsection made progressively in plane YY across the
solder llets shown in Fig. 6.78. Every joint was electrically open
resistance

7 = poor
l = good
Solvent

circuit, but because these defects do not emerge as wide-mouthed

cracks they are not visible through the Scotchcast coating during
inspection at high magnication. a General transverse view through
5
6
1

7
1

1
cracked joint, also showing physical construction of the multilayer
7 = difcult

polyimide board with copper-coated molybdenum inner layer. The

Scotchcast surrounds the package (arrows) and lls the stand-off
l = easy
Repair

volume. b Detail of the irregular crack path. Zones of maximum shear

strain have an enlarged tinlead grain structure. This is castellation no.
1
2
5

3
6

3 (64) (from Dunn and Desplat 1994)

7 = complicated
l = simple

6.15.9 SMT Problems Related to Flux

Process

and White Residues

2
1
3

6
5

7
l = cheap

It is essential that companies perform cleaning efciency

7 = dear

programmes whenever there is a change to their cleaning

Cost

process. This has been particularly relevant in recent years

7
6
3

5
1

when most aerospace companies were required to stop using

SYLGARD 184

SCOTCHCAST
CONATHANE

ozone-depleting substances (ODS) including chemicals such

SOLITHANE
MAPSIL 213
Coating type

URALANE
CV 1144-0

as halogenated chlorofluorocarbons (CFCs) like Freon. More

5750LV

recently other ODS cleaning liquids like 1.1.1-trichloroethane

EN11

have been forbidden for the nal cleaning of PCBs and, per
112

280
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms
Fig. 6.81 Thermal cycling between 55 and +100 C caused this hard
polyurethane coating to crack the ceramic component during verica-
tion testing. From 7 to +100 C the respective coefcients of
expansion for the solder alloy and the coating are approximately 22 and
247 ppm/C
ECSS Q-ST-70-08, only ethyl alcohol, isopropyl alcohol and
mixtures with water seem to be permitted in Europe.
The most common causes for flux residues remaining on
SMD assemblies after cleaning with organic solvents are:
immersion in contaminated solvent
inadequate immersion time and/or temperature
rosin polymerization (ageing of flux, which should be
removed as soon as the assembly cools to room
temperature)
poor racking in vapour degreaser
formation of white residues.
European space projects are often confronted with the
scenario of what to do when white residues are seen on
assembled, cleaned and inspected boards. This happens dur-
ing so-called mandatory inspection points (MIPs) attended by
contractor and customer. An interesting study at IVF (Tege-
hall 1991) utilized the materials and processes recommended
in ECSS Q-ST-70-08 and -38. Surface mount devices, sol-
dered with 63SnPb paste and RMA flux (F-SW 26DIN), was
followed by cleaning with isopropanol by a variety of meth-
ods (spray with ICOM 4000, spray gun and ultrasonics)
white residues were found underneath and beside the
majority of componentsthey were not observed when the
same devices were vapour cleaned in CFC-113 (except for a
few spots on one sample). The study was designed to provide
evidence of the effect of stand-off height on the cleaning
efciency of PCBs mounted with LCCC-68 and PLCC-68
components (50 and 100 mil pitch and stand-off heights
433
ranging from 120 to 870 m). It was concluded that it is
impossible to clean flux from devices below a 100 mil pitch,
or having a stand-off below 240 m. The effects, long term, of
these residual white spots was not resolved.
White residues have a variety of compositions and
cause great problems when cleaned boards are inspected
prior to conformal coating operations. These are deposits
composed of metal salts and polymerized flux which form
during the soldering operation. The metal salts are natural
products from the interaction between metal oxides on the
surfaces of parts being joined and the liquid flux. These
white oxides have been chemically analyses in the
authors laboratory and found to contain tin, lead and traces
of copper and nickel. They are deposited within the flux
along with other residues during the high-temperature sol-
dering operation. If the cleaning solvent used to remove the
flux residues is not very polar, it may leave behind salts,
resulting in a powdery-white residue.
Consulting chemistry handbooks we can nd:
Lead bromide, PbB2 is a white powder, slightly soluble in
hot water; insoluble in alcohol.
Lead carbonate, 2PbCO3 is a white amorphous powder;
insoluble in water, poisonous.
Lead hydroxide, Pb(OH)2 is a white powder, slightly soluble
in water, poisonous.
Stannic oxide, SnO2 is a white powder, insoluble in water.
Stannous oxide, SnO is unstable in air reverting to SnO.
White residues cannot be judged solely by appearances
and type of assembly process used, but instead need to be
investigated to determine the makeup and potential effects
on reliabilitya comment by Munson (2006), an expert on
these residues. Investigation will involve ion chromatogra-
phy with localized extraction of the white residue concerned.
A follow-up investigation with surface insulation resistance
(SIR) testing will establish the long term reliability of a
company soldering technology, flux system and cleaning
process. This testing will provide concrete results and is well
worth the cost. It should answer the subjective conundrum
that a defect is not always a defect.
Present spacecraft manufacturing specications require
that the effectiveness of the cleaning procedure is monitored
for ionic contamination. This is important, because as
described elsewhere in this book, ionic residues can cause
surface insulation resistance problems, dendritic growths,
and the fracture, by stress-corrosion cracking, of leaded
SMDs. However, the problem caused by white residues
will remainonly, as stated above, a detailed laboratory
investigation will help decide if these residual spots are a
cosmetic defect of a long term reliability risk.
434 6 Failure Analysis of Electrical Interconnections and Recommended Processes
6.15.10 Area Grid Array (AGA) Packaging
6.15.10.1 General
Integrated circuit packages for space electronics were, his-
torically, based on leaded devices designed with leads for
insertion into the plated through holes of PCBs. Only
so-called through-hole dual in line packages (DILs) now
remain in usagethese were followed by industries use of
pin grid arrays (PGAs) but such through-hole mounted large
packages were not popular for the space industry. Nowadays,
the vast majority of packages are based on surface mounted
technology (SMT). They include devices, discussed earlier in
this book, such as castellated chip carriers, quad flat packages
(QFPs) with L and gull-winged leads (recongured from
through-hole to surface mounting), and J leaded devices
based on ceramic or plastic leaded chip carriers.
Area grid arrays (AGAs) for space-use are all leadless.
The interconnection between AGA device and the solder
pads of the PCB is always by means of the solder which
provides the electrical circuit connections as well as the
mechanical connection. The forms of the solder are either
balls or columns having a non-eutectic composition. Ball
Grid Arrays (BGAs may, however, contain some eutectic
SnPb solder at their interface with the PCB pad. Column
Grid Arrays (CGAs) are always composed of high temper-
ature solder, usually 90Pb10Sn, and can be reinforced with a
copper spiral ribbon as seen in Fig. 6.72.
Circuit designs are becoming denser and hence SMT
packages have evolved so that leaded devices with peripheral
interconnections are now ne (and ultra-ne) pitch. AGAs have
developed from 250 to over 1000 interconnections (I/Os); often
with the same coverage area on a PCB as a 208 pin QFP.
The assembly of AGAs (BGA and CGA packages) onto
multilayer PCBs involves steps such as: baking of PCBs and
any moisture sensitive packages; the deposition of solder paste
to the PCB pads using stencil printing and occasionally syr-
inge deposition; inspection of paste deposits; AGA placement
by machine; solder reflow in IR or convection ovens; cleaning
by solvents; further inspection (see next sections); minimal
rework; and nally bagging or storage in nitrogen lled con-
tainers until it is time to populate the board with additional
SMT devices. Both BGAs and CGAs are very delicate and
must be handled with caution in order that the small, soft
interconnections are not bent or contaminated (see Fig. 6.82).
6.15.10.2 Workmanship and Inspection
AGAs are not suitable for visual inspection although some
evidence of ball and column joint quality can be ascertained
from a close inspection of the rst lines of peripheral
interconnections as shown in Fig. 6.83. Optical equipment
and binocular microscopes can be used; the Ersascope has
been specially designed for these packages.
The preferred non-destructive inspection methods are
those using perpendicular and angled systems and X-ray
transmission X-ray laminography. The principal of these
methods are shown in Fig. 6.84 and, with the inclusion of
tear-drops in pad design, this can facilitate inspection for
good solder flow and wetting, or rejectable items as seen in
Fig. 6.85a, b. Clearly hair line cracks caused by the move-
ment of this package during machine soldering will not be
seen by radiography.
A treatise on BGA assemblies, optimization of process-
ing, materials and X-ray inspection has been issued by the
IPC (2004) and is still relevant today. Of great practical use
is a wide-ranging booklet compiled by Bernard (2006) that
covers briefly the salient theoretical, then practical issues
related to the X-ray inspection of SMT assemblies, with
great detail being given to the interpretation of
X-radiographs made of defective AGA soldered intercon-
nections. The requirements of SMT soldered joints are pre-
scribed in ECSS Q-ST-70-38 and workmanship standards
are shown together with accept/reject criteria for wetting,
voids in solder balls, acceptable tilt angles for CGAs and the
size and distribution of contaminant solder particles ejected
from solder paste during machine soldering [Fig. 6.86,
(Wickham et al. 1999)]. Dr Bernards booklet would seem to
go hand in hand with the ECSS document, particularly when
the interpretation of X-ray images becomes taxing.
New developments have enabled high denition 2D
digital radiographs to be rendered by computerized tomog-
raphy (CT) to produce partial computerized scans. Once a
localized are is digitally separated, it can be closely exam-
ined and rendered to become a 3D image (as illustrated in
Fig. 6.87)here a space saving package on package
(POP) interconnection layer is manipulated to reveal two
defective (incomplete) joints that had produced electrically
intermittent open circuits (Bernard and Golubovic 2012).
This method has been developed to produce an automatic
void detection system for AGA interconnections (Said et al.
2012) said to be applicable to all void types in solders
(planar, micro, pinhole, Kirkendahl and shrinkage). Void
percentages were also measurable by an operator from 3D
rendered images where solder balls were seen as a series of
2D cross sectional images. Additional new work has been
reported by NASA JPL (Ghaffarian 2013) for the 2D and 3D
inspection of CGA1752 packages after thermal cycling. JPL
results indicated a better denition of packaging and
assembly by cross-sectional viewing, although the 3D fea-
tures were somewhat limited (less magnication) when
compared to the 2D X-ray versions. Defects in solder col-
umns (micro-cracks, voids, tilting) could not be detected by
the 3D method due to insufcient magnication.
Certain electronic components are susceptible to damage
by ionizing radiation. During X-ray inspection the sample is
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms 435

Fig. 6.82 a Manipulation of

AGAs such as this CGA package
must be made using a vacuum pen
to avoid damage to columns.
Following component placement
onto stenciled solder paste pads
the boards are ready for vapour
phase reflow. b Modern vapour
phase machines are simple to
operate once a suitable
temperature/conveyer speed pro-
le has been established on
dummy boards having similar
design/component loading as
flight boards. The machine illus-
trated incorporates a vacuum sys-
tem in the vapour chamber to
enable lower processing tempera-
tures and bar code application
provides for excellent traceability
and documentation. c Large carri-
ers are used for high throughput,
the vacuum applied when the
solder is still liquid is effective
in eliminating voids and improve
wetting as seen through the obser-
vation window of the process
chamber (courtesy of Spur Elec-
tron and IBL Technologies LLC)
436 6 Failure Analysis of Electrical Interconnections and Recommended Processes

bathed in ionizing radiation of high energy photons. Clearly

Fig. 6.83 Visual inspections can focus on peripheral joints to reveal:
a non-flow of solder paste onto solder balls of BGA, and b the
maximum tilt permitted in ECSS Q-70-38 for an assembled CGA
column
the user must ensure that the critical threshold, known for
most devices, is not exceeded. This topic is complex and
should be addressed with component experts. Devices being
X-rayed during routine inspection stages of AGA assembly,
may be given exposures higher than expected because sus-
pected defects in joints will be examined at higher magni-
cation, and possibly for multiple times when other persons
become involved. If there is cause for concern, the steps
needed to minimize damage have been recommended by
Bernard and Blish (2005). These include: increasing distance
between sample and X-ray tube focal point; use additional
lters; minimize time of inspection; check the actual dose
given to the samples; and, use automated systems and record
images.
An example of X-ray CT inspection applied to the
Pb10Sn columns attached to ceramic grid arrays is shown in
Fig. 6.88. When micro-voids are suspected to be present
within either balls or columns attached to AGAs, it is nec-
essary to assess their size and distribution throughout the
interconnection. It was considered likely that only very large
voids would affect the reliability of these devices. By
recording a sufciently large number of digital images
through a multitude of planes it is also possible to produce a
video of these inspections. Playback of the video can reveal
voids and other microstructural features during progression
from the top of the column attachment to its base. It is
interesting to note that by post-processing CT data, such as
that used to produce the images in Fig. 6.88, it is possible to
generate a physical 3D printed model. These models can be
printed as large-scale replicas of interconnections and used
for operator and inspector training purposes. Incidentally,
the technique of exporting CT segmented data, as VRML

Fig. 6.84 X-ray inspection,

shown here for a BGA on a
simple PCB (with no tracks) to
illustrate the usefulness of designs
where tear-drops are incorporated
into pads. The stenciled paste is
seen printed in (a). The X-ray
schematic in (b) shows how
tear-drops can aid inspection for
reflow. c Shows actual
radiographs where one ball solder
llet has not flown and wetted to
the tear-drop padindicating a
potential open circuit
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms
Fig. 6.85 Metallography conrms a only partial reflow of solder paste
in Fig. 6.83a, and b non-wetting of ball to pad and into tear-drop.
However, it is clear that hair-line cracks of the kind seen in the
microsection of c will not be revealed by radiography and here a
destructive dye penetrant method may be useful (Wickham et al. 1999)
les, to a rapid prototyping machine (3D printer) has been
used to create biomechanical models of skeletal and mus-
cular structuresuseful teaching aids for biology and
medical students (Kleinteich et al. 2015).
437
Fig. 6.86 CGA workmanship X-radiograph to highlight bent columns
(pins) and solder splash solder balls having diameter down to 0.1 mm.
Such resolution is required for the inspection of space hardware.
Calibration of the X-ray equipment can be made by attaching a clear
adhesive tape, loaded with solder splash balls down to 0.1 mm
diameter, to the surface of a flight assembled multilayer board.
Resolution to observe a 0.2 mm ball is recommended
6.15.10.3 Effect of Environmental Testing
and Inspection Methods
for AGAs
AGAs are a desirable alternative to other surface mount
devices such as quad flat packs as discussed in previous
sections, they result in a tremendous saving in board real
estate. Early studies aimed at introducing AGAs into space
systems (Tegehall and Dunn 2001a) produced results valid
today, and they have been used as a baseline for the
development of ECSS Q-ST-70-38 process verication
requirements. Resistance to thermal cycling between 55
and +100 C was paramount and tests made on ceramic
BGAs and ceramic CGAs (all 625 I/Os) mounted to a variety
of PCB laminate materials were conducted to ECSS
Q-ST-70-38, followed by vibration tests to the same stan-
dard. Study variants included: the use of different, known
volumes of deposited solder paste; different PCB substrate
materials; the use of underll; and, the application of rework
methods. A summary of results is as follows:
(a) BGAs fail before 500 cycles (most passed 200 cycles)
on polyimide-glass substrates.
(b) The low in-plane CTE values for polyimide-aramid
bre boards improved the fatigue life of BGAs and
CGAs. The BGAs did have cracks after 1000 cycles,
limited to the corner joints. All CGAs were acceptable.
(c) Substitution of the balls on the packages for columns
improved fatigue life considerably on packages
438 6 Failure Analysis of Electrical Interconnections and Recommended Processes

Fig. 6.87 A fast, partial

computer tomographic scan of a
POP device was initially obtained
as a virtual cross-section through
a plane containing the
interconnectionssome joints
appeared to be abnormal.
A detailed scan was made in order
to produce a a virtual cross
section in the vertical plane and
then, b a virtual cross section in
the horizontal plane. c A 3D
rendered view of the intermittent
joints is presented from the CT
model (courtesy of Nordson
DAGE)

soldered to polyimide-glassno failures after 1000
cycles.
(d) The use of underlls had no positive or negative effects
for the BGAs.
(e) Underll beneath AGAs mounted on polyimide-aramid
caused cracking of the jointseven with the more
resilient CGAs, as seen in Fig. 6.89.
(f) The volume of solder paste printed onto the PCB solder
pads is critical for the fatigue life of CBGAs. A de-
crease of the paste volume to 0.068 mm3, about 10 %
less than the minimum solder paste volume recom-
mended by the device manufacturer, caused a decrease
in fatigue life of 2040 %.
(g) Very small cracks were seen beneath the BGA and
CGA mounting pads (in the laminate material of the
PCB). A later study (Tegehall and Dunn 2003) showed
these cracks to initiate in the low temperature regime of
the thermal cycling, this might have been benecial in
that fatigue lives were extended due to the hinging
effect of the copper pad.
(h) A suitable destructive dye penetrant test method was
introduced for quantifying the extent of solder fatigue
cracks within the soldered balls and columns.
The failure of AGA balls and columns during thermal or
other environmental test exposure can be monitored by
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms
Fig. 6.88 X-ray CT inspection of solder columns, a at the
PCB-to-column interface, and b at the package-to-column interface
(images courtesy of JAXA)
electrical continuity if the packages contain internal
daisy-chained circuitry. However, the usual technique is to
perform delicate metallographysomething that can only be
made by highly trained technicians. Green or yellow lumi-
nescent compounds are added to the mounting medium to
better highlight cracks and other defects, as seen in
Fig. 6.89b. The most highly stressed interconnections are
usually those along the outer periphery, or corners of a tested
device, so that visual observations are possible. Figure 6.90
illustrates how visual and metallographic examinations can
439
be complimentary whereas electrical continuity testing is not
always cut and dry because fractured joints give intermittent
connection. During temperature cycling breaks can recon-
nect and act as a switch as COE mismatches widen and pull
back opposing fracture surfaces.
The present specied methods for establishing the extent
of damage to the interconnecting balls and columns of
AGAs at the end of an environmental test programme are by
a destructive dye penetrant procedure (ECSS Q-70-38), to be
reviewed in the next section, or metallography. The pass/fail
criteria is that the maximum crack extending into a single
interconnection is 25 % of the original cross-section of the
ball or column. An alternative method based on C-mode
scanning acoustic microscopy (C-SAM) has been evaluated
in an attempt to overcome the problems associated with the
existing specied methods: dye-penetrant testing followed
by tensile testing to reveal the extent of cracking can lead to
erroneous readings due to ductile (necking) fracture of the
solder; and, metallography, as previously mentioned is a
delicate operation that is time-consuming (particularly if
serial microsections are made through joints in order to
obtain more than one 2D view of the interconnection). The
ndings of this study (Corocher et al. 2009) have shown the
C-SAM to give a reliable measure of the structural damage
in the columns of multilayer ceramic grid arrays (MCGAs)
as illustrated by Fig. 6.91. It is about 4 % more accurate than
serial microsectioning and considerably better than if using
dye penetrants. It is also a quicker inspection tool, but fur-
ther work is needed to develop a standard calibration device
for this C-SAM application.
The dye penetrant test method prescribed in ECSS
Q-70-38 for BGA and CGA packages is detailed in Annex A
of ESA STM-266 (Tegehall and Dunn 2001a). Essentially, a
dam is created around the component to be examined using
modelling wax. The dam is lled with dye (Steel Red from
Dykem proved suitable, being quick to penetrate then dry,
also IFWB-C2 and EOSINA are easy to use and give good
results), penetration of dye into any micro-cracks can be
aided by placing the PCB assembly into an evacuation
chamber and pulling vacuum down to 100 mbar. This will
remove any air trapped in the cracks. The upper side of the
component under investigation is roughened to help the
gluing of a steel cylinder to the surface. The PCB is attached
to an aluminium plate and held by the grips, or an attach-
ment, to a tensile test machine. The steel cylinder is loaded
until the package detaches from the PCB. Severely
degraded/fatigued interconnections will break from the
board and examined under a binocular microscope as shown
in Fig. 6.92. This method is called dye and pry by
Motorola (Burnette and Koschmeider 2003) and is used to
check for cold solder joints, fractured joints (as Fig. 6.92)
and the presence of black pad (when affected electroless
Ni/Au solder pads will appear grey-black after breaking).
440 6 Failure Analysis of Electrical Interconnections and Recommended Processes

Fig. 6.89 Cross-section of corner column of a ceramic CGA with z-direction have caused catastrophic cracking at b, the
epoxy underll (mounted on poyimide-aramid) after 500 thermal cycles column-to-metallized package connection (dark-grey undercoat is just
55 and +100 C. a The PCB pad is intact but stresses in the visible surrounded by green mounting medium)

Fig. 6.90 Column grid array

interconnections following
extensive thermal cycling (until
rst failures) showing a the
composition of the columns,
b ductile distortion and break of
unsupported column, and
c copper ribbon column fracture
at PCB pad. All failures are
caused by material COE
mismatch during thermal cycling.
The conditions of testing
probably far exceed the end of life
of representative satellite, but it is
important to understand the mode
and location of material damage
likely to cause electrical failures
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms
Fig. 6.91 Top image shows a defective solder column as represented
in a C-SAM. The blue circle indicates the columns circumference. The
red colour is the dense defect-free volume within this column and the
green-yellow area depicts the defective perimeter (e.g. crack front)
clearly visible in the microsection (From Corocher et al. 2009)
The method is somewhat improved by increasing the pull
speedto better discriminate among higher strength parts,
the measurement of pull strength and pulling at an angle to
mimic real life scenarios (Raiser and Amir 2005). The dye
penetrant test is applicable to mass produced circuits
(Fig. 6.93) and, as shown by Tegehall and Wetter (2015) to
evaluate the fatigue lives of lead-free solder BGAs.
6.15.10.4 Rework and Repair of AGAs
Rework and repair are more complicated for AGAs than for
conventional packages. Since touch-up of individual solder
joints is not possible, AGAs must always be removed even if
only a single solder joint, ball or column falls outside of the
ECSS Q-ST-70-38 workmanship requirements. Rework
441
systems are commonly used, by certied operators, for the
removal of reject AGAs, they are based on IR and hot gas
heating with real-time, non-contact temperature sensing and
controls. Precision pick, alignment and placing accompanies
modern equipment which has process observation cameras,
monitors and PCB assembly cooling. Importantly, the ECSS
requirements mandate that if SMTs are to be mounted, it is
equally necessary for them to be removable, and replaceable
with alternative, new devices. This reflects reality in the space
business when damage may occur during assembly, when
devices are found to be part of a lot that is designated defective,
or because late design modications are made to the circuit.
A study, aimed to be a step towards writing a draft pro-
cedure for the removal and replacement of area array
packages (for inclusion into ECSS Q-70-28), was com-
pleted with industrial participation. A wide range of SMT
devices was submitted to an evaluation programme for the
repair and visual/NDT inspection of leadless packages using
micro-focus X-radiography. Thermal proles were estab-
lished (all different) for: Ceramic land grid arrays (CLGA)
mounted to Solder Column Interposers; CCGA-625s; Kyo-
cera D-BGA-472 with dimpled ball grid arrays; ceramic
SMD 0.5; SMD2s; Leadless ceramic chip carriers
(LCCC-20s); and, LCCC-44s. All were successfully moun-
ted, reworked and all passed a verication programme to
ECSS Q-70-38 (Schoenbeck and Dunn 2004). This work
was followed up with the objective of establishing the
number of reworks that could be made to difcult to repair
packages. Space-qualied AGAs were reworked using space
materials and qualied PCB multilayer boards (Tegehall and
Dunn 2005). CBGA and CCGA packages with up to 625
I/Os were reliably reworked using semi-automated work
stations and pre-determined temperature proles. They could
be reworked at least ve times without affecting the integrity
of the PCB laminate or solder lands (substantiated by visual
and metallographic inspection). That is two more reflow
cycles than the maximum allowable repair to any of 25 cm2
of PCB surface area according to ECSS Q-ST-70-28.
The repair of flight PCB assemblies by European prime
and sub-tier contractors has been free of major problems
when applying veried repair procedures. However, cer-
tain non-silicone conformal coatings and all underll mate-
rials have caused major non-conformances. As discussed in
the section related to conformal coatings, the silicone based
coatings are often to be preferred as they are easy to apply,
are less likely to cause thermal fatigue of interconnections
during extended thermal cycling and they facilitate
repair-after-coating due to their flexibility and moderately
low adhesion to PCB laminates.
442 6 Failure Analysis of Electrical Interconnections and Recommended Processes

Fig. 6.92 Ceramic BGA

packages with 625 I/Os having
been solder-assembles to a PCB
and verication tested (ECSS
Q-70-38, 55 and +100 C, 500
cycles). Mechanical detachment
after soaking in dye penetrant
plus drying reveals: a solder balls
remaining on the PCB after
removal of device A, b extent of
cracking on the component
sides of device A (left) and
device B (right). c Shows
close-up optical views of
fractured corner joints seen on
A and B. It can be seen that
maximum stresses have occurred
at the edges and corners of the
ceramic packages, causing
micro-fatigue cracks (NP on
gure) to propagate towards the
centre of the packages (Tegehall
and Dunn 2001a)

6.15.11 High Voltage Interconnections spacecraft achieve signal amplication by means of travel-
and Influence of Geometry ling wave tubes (TWTs)these operate at several kVs. As
(Workmanship) on Corona the power levels increase, so the transmission voltages will
Discharge have to be increased in order to decrease currents and so
minimise the conductors weight and resistance losses.
Background
Workmanship standards related to the solder-assembly of
The power conditioning and distribution subsystems of all
high voltage equipment are often included in the specica-
spacecraft include the management of the solar arrays, the
tions as artists impressions drawn to depict requirements for
harnesses and all of the electrical equipment. Numerous
smooth joints (ECSS Q-ST-70-08, NASA-STD-8739.3see
equipment operate at high voltage, particularly nowadays
Fig. 6.94). Words include: for high-voltage terminations,
when telecommunication satellites have power requirements
where coronal suppression is necessaryall joints shall be
of 20 kV and higher. Electrical thrusters for space need high
covered in smooth llets, free of discontinuities (no sharp
voltage and associated high power supplies. Power ranges
points, angles, etc.). For a variety of reasons operators may
from a few Watts for eld emission thrusters to many
have great difculty in achieving such objectives. These can
Kilowatts for gridded of plasma thrusters. Solar arrays have
include: the local poor solderability of cut ends of stranded
a power capability typically ranging from 6 to 25 kW. Some
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms 443

Fig. 6.93 Dye penetrant media can have various colours and are more visible under UV illumination. A set of crack sizes has been established
(courtesy of Huang 2011)

Fig. 6.94 NASA (2002): acceptableall sharp edges are completely covered, unacceptablepartial solder cover. ECSS (2009): acceptable
smooth convex llets, no discontinuities or severe change in contour and no sharp points or edges

wire or component leads, the large size of certain terminal
pins and posts, difcult accessibility with a soldering iron, as
well as the inability to build up molten solder around the
joint because the materials being joined have both an
excellent solderability and a high thermal conductivity.
Many hours have been spent in attempting to rework such
soldered joints in order to cover all wire strands and avoid
every protrusion. A detailed evaluation (Materassi et al.
2000) was made which concentrated on typical soldered
joints found on space-quality printed circuit board assem-
blies and wire bundle terminations (as used in electronic
power supply units and other power distribution circuits) that
operate between 400 and 900 V. It was aimed to shed light
on whether protrusion-free solder llets are technically
required for spacecraft units operating under a low partial
pressure. Samples were made by a certied operator using
space-approved materials having low outgassing-under-
vacuum properties. They were tested at operational poten-
tials ranging between 400 and 900 V, with different elec-
trode spacing (insulation gaps) and under various partial
pressures. Two distinct batches of specimens were produced
by hand soldering: the rst with ideal round llets, the
444 6 Failure Analysis of Electrical Interconnections and Recommended Processes

second with standard llets that exhibited protruding wires PCBprinted circuit board
and sharp contours. Comparison of the test results obtained CDCorona discharge
for both batches was performed.
Two standard specimen designs were selected for this
Test description
work as shown in Fig. 6.95. The rst consists of stranded
When the voltage stress between two electrical interconnec-
space quality wires soldered into the plated through holes of
tion terminals exceeds a critical value it is observed that
PCBs, while the other consists of stranded wires soldered to
luminous discharges are generated due to some form of gas
the pins of a commercial electrical feed-through for vacuum
ionisation. These discharges, which precede the complete
applications.
spark breakdown of the gap, are known as corona discharges.
The soldered interconnections are seen in Fig. 6.95they
They can be continuous, or in the form of individual voltage
were produced by a trained and certied operator. Flux
pulses at the terminal surface. Under the test set-up (Mat-
removal from the samples was paramount, they were also
erassi et al. 2000; Capineri et al. 2003) the voltage pulses
baked to ensure no outgassing products remained.
were detected by a measuring instrument that was coupled to
A schematic diagram of the test facility is seen in Fig. 6.96.
typical spacecraft terminals and calibrated to respond to the
The actual facility is photographed in Fig. 6.97 during
voltage pulses in terms of the charge transferred to the ter-
set-up, and during testing. The typical blue-white coloured
minals. This is identied as the apparent charge; it is
coronal discharge is clearly visible. The CGIV is the voltage
generally not equal to the total charge that occurs locally at
that was measured to indicate the possible effects on dis-
the site of the corona discharge, but it is directly related to the
charge inception caused by the different shapes of soldered
discharge as described in the test standard.
terminationsit is the lowest voltage at which sustained
luminous Corona Discharge occurs as the test voltage is
The list of acronyms that will be used by experimenters is
gradually increased from zero to the test range (400900 V).
as follows:
Values of CGEV, which is the highest voltage at which
sustained CDs are extinguished if the test voltage is grad-
CGIVCorona glow inception voltage
ually decreased after corona is obtained, were also deter-
CGEVCorona glow extinction voltage
mined, as well as the discharge rate at CGIV, and the glow

Fig. 6.95 Visual appearance of

interconnections submitted to
testing. a PCB specimen with
rounded llets (so-called HV
joints); b with sharp, protruding
llets (so-called stud-joints).
c Electrical feedthrough with
rounded llets, d Electrical
feedthrough with binding strand
and sharp llets
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms 445

Fig. 6.96 Vacuum/baking

unitfunctional scheme

Fig. 6.97 View of a PCB test

specimen inside the glass test
chamber, and (colour) under test
showing view of corona glow
discharges around PCB specimen
446 6 Failure Analysis of Electrical Interconnections and Recommended Processes

discharge current. Special calibration procedure is performed a test voltage of 1000 V, this is above the threshold for the
by injecting a known charge into the measuring circuit and onset of corona discharge (for this specimen the CGIV was
observing its response. 780 V) in order to magnify the discharge effects. The dis-
charge was in nitrogen at a pressure of 15 mbar. The topology
Test results
of these solder joints was photographed and examined by
Printed circuit board round and stud joints: The discharge
scanning electron microscopy. The pair of facing joints was
inception voltage for each specimen was recorded. The
then mounted in a room temperature curing plastic and
results of the various measurements of CGIV for the PCB
microsectioned. The polished cross-section was examined by
specimens having 7.5, 5.0, and 2.5 mm electrode spacing
light and scanning electron microscopy. The scanning elec-
were tabulated. Five readings were made at each chamber
tron microscope (SEM) images of the two round solder llets
pressure, and each pressure was selected to be different in
were inspected. The negative electrode supports surface
order to produce a wide range of pressure-distance products.
nodules that were not present at the onset of the testing. The
In a similar manner, the discharge extinction voltages,
positive electrode is smooth with few further irregularities. It
CGEV, were recorded. It was seen that for each environ-
is interesting to note that at this high magnication, a small
mental test pressure, even taking into account the standard
fraction the copper pad of this terminations may not be
deviation of the mean, there is no obvious case for proving
completely wetted. The strange surface morphologies of
that the rounded solder llet is likely to possess a higher
small nodules, protruding from the solder surface to a height
inception voltage than that of a sharp solder llet. Fig-
of 30 m, and lifted craters having a diameter reaching
ure 6.98 shows the pressure-distance product (pd) depen-
approximately 100 m, are detailed in Fig. 6.99. The pol-
dence of the discharge inception voltage, CGIV, for the two
ished microsection made across these round joints is shown
different solder llet shapes; this is repeated for the three
in Fig. 6.100. The actual gap between the joints is shown to
different gap lengths. The Paschen-like curves have been
be 2.7 mm, this is the shortest distance between the facing
evaluated in the region around the inception voltages
copper pad terminations. It is the mid-point of the plated
(pressure range from about 0.120.0 mbar).
through hole-to-hole distance that is 5.0 mm. Very detailed
views of the non-wetted pad edge and the protruding nodules
Wires Soldered to Electrical Feedthroughs: Commercial
are shown in Fig. 6.100c, d. The nodules were chemically
electrical feedthroughs, whose design is typical for spacecraft
analysed by EDAX and found to be lead-rich, whereas the
high voltage equipment (electronic power supply, TWT
bulk solder is tin-rich (no other elements were detected). The
amplier, etc.) consisting of two pins and ceramic-to-metal
melting point of tin is approximately 100 C lower than that
brazed seals for vacuum applications, have been used for the
of lead. The presence of lead-rich phases in the protruding
second set of tests. Most of the pin lengths are masked by
nodules gives support to the hypothesis that it is the lower
ceramic tubes, except for the exposed 5 mm which was used for
melting temperature tin-rich phase that is being removed by
soldering. A ceramic spacer was placed between the pins in
sputtering from the surface.
order to achieve a precise pin separation of 7.9 mm. All spec-
imens were hand-soldered in the same manner. Fixed lengths of
Conclusions
AWG 22 wires were stripped and pre-tinned as for the rst set of
Paschen-like curves have been drawn for all specimen
tests. After soldering, the insulation clearance of the wires was
congurations (instead of breakdown voltage the inception
approximately 3 mm. Lap joints with rounded llets were made
and extinction voltages for corona discharge as a function of
to represent typical high voltage connections. Lap joints that
the product of test environment pressure and electrode
were reinforced by windings of thin stranded wire and then
spacing has been considered).
soldered, represent joints with sharp-cornered llets. All spec-
imens possessed an excellent solderability and locally, between
For the printed circuit board terminations:
conductors, the solder llets have low contact angles. The
rounded llets and sharp llets are illustrated, respectively, in
The results for the joints containing a large protrusion
Fig. 6.95. Paschen-like curves have also been constructed,
(i.e., the normal stud joints) are similar to those of the
based on tabulated results, and are seen in Fig. 6.98.
specially made, rounded, HV joints. The distance
Metallography between the perpendicular, wire-end protrusions is far
The long-term effect of corona discharge on the solder joints greater than the insulation gap between pads (e.g.,
and materials associated with the PCB specimen with roun- compare Fig. 6.95a, b with Fig. 6.100).
ded solder llets, 5 mm apart, was investigated by micro- With slightly rened measurements (Capineri et al. 2003)
scopy. One such specimen was energized for a total of 30 h at the corona inception voltage of the solder joint
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms 447

Fig. 6.98 Paschen-like curves:

a discharge inception voltage as a
function of pressure-distance
product for PCB solder joints
with a 7.5 mm spacing, b 5 mm
spacing, and c 2.5 mm spacing.
d discharge inception voltages as
a function of pressure-distance
product for feedthroughs
(Materassi et al. 2000)
448 6 Failure Analysis of Electrical Interconnections and Recommended Processes
Fig. 6.99 Close SEM inspection of the pair of round soldered joints
showed a the negative terminal to support surface nodules, and b the
positive terminal to be rather smooth, but containing craters and a few
small nodules
congurations with sharp solder llets were found to
have a Paschen curve minimum value of 550 V. This is
only 50 V lower than the 600 V minimum detected for
round, smooth joints.
Two predominant factors in determining whether a pair
of joints will be prone to corona inception discharge are
the pad-to-pad gap and the condition of the pad edges
(e.g., copper pad edges that are clean, free of solder-flux
residues, coated with a smooth layer of fused tinlead are
to be preferred). Figure 6.100a, c reveal that a small
portion of the right-hand pad had not been covered with
solder: this should be a reject featurebut is only visible
under a magnication of greater than *20 times.
For the electrical feedthrough terminations:
The use of a binding strand for securing wires to the
feedthrough lead, which leads to a sharp pointed surface,
seems to have no influence on the inception voltage values.
Large solder llets are to be preferred as these are more
likely to cover the pad edges and be free from any surface
steps in this critical location.
Protruding wires, as present on stud-joints, need not be
reworked unless their ends are located close to other
conducting surfaces such as adjacently stacked circuits,
housing walls or mechanical supports.
By limiting rework to only those solder llets likely to be
susceptible to corona discharge (based on the above) it
will be possible to avoid risking the integrity of heat
sensitive parts of the electronic circuitry.
6.15.12 Tin Pest
Tin pest (or tin plague) is the term given to the allotropic
transformation of pure tin from the normal tetragonal, shiny
white metal to a cubic, grey powdery semiconductor
(Brandes 1992). The equilibrium transition temperature is
13.2 C, but much lower temperatures are needed to make
this crystallographic change in structure occur. The actual
transformation temperature and rate of reaction from one
state to the other, depends on the purity of the tinand as
performed by NPLon whether the white tin can be seeded
with a compound (Hunt 2009), or another metal such as
germanium. Tin pest has become more important since the
move to lead-free electronics. Historically may have been
signicant: when Napoleons men froze in the Russian
winter when tin pest disintegrated the buttons on their
trench-coats; and, when Scott attempted to reach the South
Pole but was hampered by leaking kerosene cans due to tin
pest of their soldered seams. Tin pest is prevented from
occurring by adding small amounts of antimony and bismuth
to pure tin when ingots are stock-piled by governments. Any
allotropic changes would render the element almost worth-
less, and these elements, Sb and Bi, are now typically found
as traces in solder alloys. Tin pest is manifest in the images
seen in Figs. 6.101 and 6.102.
Although tin pest is a potentially an insidious risk to the
reliability of electronics it does not appear to have caused
any reported failures in the space community. However,
laboratory studies have established that cast tin alloys can
also suffer from tin pest if held for long periods (years) at
temperatures of 18 C, with even greater occurrence at
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms 449

Fig. 6.100 Microsection

through the pair of PCB terminal
joints shown in Fig. 6.95a. The
gap between these terminal pads
is measured to be 2.7 mm. b and
c Shows a detailed view of the
facing pad edges seen in (a).
b Shows some roughness on the
negative terminal llet, here the
edge of the copper pad is well
covered with solder. The slightly
non-wetted corner of the pad on
the opposing positive termination
is highlighted in (c), and gives
rise to a surface discontinuity.
d High magnication SEM
secondary electron image of a
protruding nodule on the negative
terminal shown in Fig. 6.99a.
EDAX analysis identied this to
be lead-rich

40 C (Plumbridge 2011). It has been stated that tin can
be converted back to its metallic state if heated above 13 C,
but without regaining its structural integrity (Kostic 2011),
but this is not fully substantiated. Pure tin is prohibited from
the designs of European and US spacecraft (except in the
case of fused-and-drawn tin coatings on stranded wires,
which are immune to transformation and whisker growth) so
tin pest is unlikely to have been the subject of any hardware
reviews. The tin whisker study (Dunn 1987) included the
subjection of more than 100 pure tin whiskers to cryogenic
temperatures. They were held in liquid nitrogen for periods
from 0.5 to 168 h, after which the samples were rapidly
transferred to an SEM. Despite many repeated and careful
examinations of the cryogenically exposed tin whiskers,
none were seen to have acquired a changed surface
morphology.
450 6 Failure Analysis of Electrical Interconnections and Recommended Processes

Fig. 6.101 Time lapse

photographs taken in the same
area of a pure tin sample held at
35 C. In order to accelerate the
transition from shiny white tin
to grey powdery tin the
sample was seeded (the trigger for
this transformation is uncertain).
Courtesy of Hunt, NPL (2011).
Grey tin is *27 % less dense
than white tin, hence its eruptive
appearance in these photos

Fig. 6.102 Sn0.5Cu after ageing

at 18 C, and cross-section of a
sample after 1.5 years storage
(Courtesy Kariya, Williams, and
Plumbridge. The Open
University, UK)
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms 451

Fig. 6.103 Arrangement for

tensile testing at 4.2 K in LHe
featuring the 4 column testing
machine, the LHecryostat
containing the specimen (black
cylindrical object)

6.15.13 Mechanical and Electrical Properties 6.15.13.2 Mechanical Properties

of Electronic Materials The mechanical properties of near-eutectic tinlead alloys
at Temperatures Down to 4.2 K were determined at temperatures between 60 and +70 C in
order to estimate which might be better for spacecraft PCB
6.15.13.1 General assembly joints subjected to thermal cycling (Dunn 1975). It
The majority of state-of-the-art cosmic ray detectors require was found that silver, copper and, to a lesser extent, anti-
cooling for optimum performance. The same statement goes mony would: improve the mechanical strength of purely
for both ground and space-borne cameras and spectrographs. eutectic tinlead, 63Sn37Pb, composition; decrease ductil-
Sensitive photodetectors rely on cooling to reduce noise and ity; increase strain-hardening effects; and raise the
dark current. At temperatures above absolute zero, 0 K, all stress-relaxation temperature. The pure 63Sn37Pb was rec-
matter emits temperature-dependent electromagnetic waves. ommended due to its lowest viscosity during soldering of
By cooling detectors down to liquid helium temperature,
4.2 K, the entire range of infrared wavelengths can be
assessed, particularly the infrared emissions from astro-
nomical objects. The IRAS (Infrared Astronomical Satellite)
was the rst spacecraft to be launched with a dewar (lled
with 127 gallons of LHe and 62 detectors). Because the
entire telescope operated at 4.2 K it emitted virtually no
infrared radiation itself, hence there was no interference with
the observations during its 21 day mission. Many other
satellites have similarly been launched with infrared tele-
scopes maintained at colder temperatures than the objects in
space that they observe, one of the most successful being the
Infrared Space Observatory which took data for 2.5 years
(see Sect. 5.17.2 and Fig. 5.78). Operating temperatures of
spacecraft components can also encounter very cold
regimes, during solar eclipse periods or when voyaging on
deep space missions. Fig. 6.104 Stress versus strain diagram as recorded using a PC
452 6 Failure Analysis of Electrical Interconnections and Recommended Processes

Fig. 6.105 Compilation of results for Elastic Modulus E (GPa) and Elongation at fracture (A%), at room temperature, 77 and 4.2 K
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms 453

Fig. 6.106 Compilation of results for Ultimate Tensile Strength and Proof Strength, at room temperature, 77 and 4.2 K
454 6 Failure Analysis of Electrical Interconnections and Recommended Processes

Fig. 6.107 Fracture surfaces for: 50In50Pb (left, optical) the only ductile solder at all temperatures down to 4.2 K, and Solithane 113
(SEM) which tears at room temperature, but has a brittle glassy fracture at cryogenic temperatures (below Tg)
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms 455

plated through holes, minimizing thermal shock, and pro- Table 6.9 Test matrix listing material samples for electrical resistance
viding for optimum stress relief during thermal cycling. testing at room and cryogenic temperature, and those samples selected
for microstructure evaluation by focused ion beam and SEM
In recent years the property requirements for solder alloys
have been under more scrutiny and data was requires for Material Resistance Microstructure
RT LN2 LHe from FIB
various solders and materials associated with electronic
circuitsfor operation at cryogenic temperatures. The test OFHC Cu 3 3 3
programme results are summarized from the published paper 63Sn37Pb cast 3 3 3
(Fink et al. 2008). Specically, the tensile properties of the 63Sn37Pb wire 3 3 3
following materials were characterised at room temperature,
62Sn36Pb2Ag cast 3 3 3
at liquid nitrogen and at liquid helium temperature:
60Sn40Pb cast 3 3 3 x
60Sn40Pb wire 3 3 3 x
Solder alloys: 63Sn37Pb, 62Sn36Pb2Ag, 60Sn40Pb,
96Sn4Ag, 50In50Pb, 70Pb30In, 96.8Pb1.5Ag1.7Sn, 96Sn4Ag cast 3 3 3 x
96.5Sn3Ag0.5Cu. 96Sn4Ag wire 3 3 3 x
Printed circuit board (multilayer) laminates: polyimide glass 96.8Pb1.5Ag1.7Sn 3 3 3
bre, epoxy glass bre, and Thermount. cast
96.5Sn3.5Ag0.5Cu 3 3 3
cast

Fig. 6.108 FIB cuts through:

a and b Sn96Ag4 cast,
quasi-polygonal grains, 13 m
grain size with transgranular
silver-rich precipitates.
c Sn96Ag4 wire, quasi-polygonal
grains, 17 m grain size with
mainly grain boundary
precipitates, d Sn60Pb40 cast,
oval grains, 517 m grain size,
and e Sn60Pb40 wire, equiaxed
Sn- and Pb-rich grains, up to
10 m
456 6 Failure Analysis of Electrical Interconnections and Recommended Processes

Conformal coatings: Arathane 5750, Sylgard 184, Scotch-

cast 280, Solithane 113, CV-1144-0, Mapsil 213, Conathane
EN4/EN11.
and, oxygen free high conductivity copper (OFE Cu).

Tensile samples were produced by casting as in the pre-

vious study (Dunn 1975) and machiningall were
non-destructively examined and any with defects such as
porosity larger than the minimum detectable size (0.4 mm)
were rejected.
Tests were performed with a constant cross head speed of
1 mm/min and a tensile preload of 1020 N. The exten-
someter was calibrated and zeroed at preload level prior of
each test. The mechanical properties calculated after com-
pletion of each test were based on the recorded signals of the
load cell, the crosshead position sensor and the exten-
someter. Measurements at room temperature were per-
formed with a video extensometer. Tensile tests at
77 K (LN2 temperature) were performed by means of a
liquid nitrogen bath mounted in the tensile test machine.
Measurements at cryogenic temperature were made with a
special low temperature extensometer. Tensile properties at
4.2 K (LHe temperature) were measured by means of a LHe
cryostat mounted in a mechanical testing machine
(Fig. 6.103). By means of this experimental setup tensile
Fig. 6.109 Circuit diagram. R in the circuit diagram is a calibrated
measurements with a maximum load of 200 kN at cryogenic
high-precision measuring resistance. I = 5 A. The solder is maintained
temperatures down to 4.2 K can be performed. A special top at room temperature or either in a bath of LN2 (77 K) or in a LHe
loading mechanism enables a quick change of the samples. cryostat (4.2 K)
Therefore, no heating up of the cryostat between the tests is
warranted and so, the cooling down time of specimen is less
than 2 h (big samples, like PCBs) and less than 1 h (smaller
samples, like OFE copper, solders and conformal coatings).
At least 3 tensile specimens of each material and each
temperature were tested. During each tensile test a stress vs
strain diagram is plotted, from which the tensile properties
were evaluated. Figure 6.104 shows a typical one.
The mechanical properties of all tested materials are
recorded graphically in Figs. 6.105 and 6.106 (averages of
triplicate tests). For an extended analysis of the results, and
the complete tabulations of mechanical test date the original
paper (Fink et al. 2008) should be consulted. The solders
generally show an increase in UTS, the Youngs Modulus
and the stress at the limit of proportionality R0.2 with
decrease in temperature. For the near-eutectic solder alloys
these results are in agreement with the initial study. The Fig. 6.110 Temperature dependence of specic resistance (mean
63Sn37Pb has the greatest ductility at ambient temperature resistance)
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms 457

Sn96Ag4

Sn96.5Ag3Cu0.5
wire
cast

1.06E05
2.27E07

2.21E06
1.89E07

9.04E08
1.46E08
bulk LHe

Pb96.8Ag1.5Sn1.7

LN2
2.14E05
1.00E06

5.58E06
4.02E07
bulk

SC

RT
Sn62Pb36Ag2

1.26E05
5.50E07

2.74E06
3.68E08

1E-8 1E-7 1E-6 1E-5

El. Resistance [Ohm cm]
bulk

SC

Fig. 6.111 Temperature dependence of specic resistance (mean
resistance) of Sn96Ag4
Sn63Pb37

1.92E05
2.50E07

3.94E06
2.57E07
wire

SC

SC
[4]

Sn63Pb37

1.20E05
5.63E07
1.45E05

2.71E06
1.73E07
Table 6.10 Mean values of specic resistance at RT, LN2 and LHe temperature for all materials tested

bulk

SC

SC
[5]

[4]


Sn60Pb40

1.61E05
6.43E07

4.09E06
1.15E06
wire

SC

SC
[4]

Sn60Pb40

1.48E05
1.02E06
1.49E05

3.66E06
3.99E07
bulk

SC

SC
[4]

[4]


1.29E05
1.97E06

2.31E06
6.63E07

5.97E08
1.33E09
Sn96Ag4
wire

Fig. 6.112 Temperature dependence of specic resistance (mean

1.07E05
9.88E07

2.19E06
1.05E07

5.97E08
5.51E09
Sn96Ag4
Elec. resistivity ( cm)

resistance) of Sn60Pb40
bulk

and when silver is added (62Sn36Pb2Ag) there is a slight

OFHC Cu

1.75E06
3.30E08
1.75E06

2.69E07
1.34E08
2.68E07

4.18E08
2.86E09

increase in strength. Interestingly, the 50In50Pb solder

bulk

remains extremely soft with a very low Youngs Modulus

[4]

[4]

and high ductility compared to the other solders. As seen in

Fig. 6.107, the 50In50Pb solder remains ductile at 4.2 K.
Literature

Literature

Literature
Std. Dev.

Std. Dev.

Std. Dev.
Resource

Resource

Resource

A programme to study the brittle-to-ductile fracture

Mean

Mean

Mean

transition temperatures in bulk lead-free solders and eutectic

tinlead has been conducted using the dynamic Charpy
test method (Ratchev et al. 2005). No transition was seen in
LN2

LHe
RT

the SnPb alloy, but distinct transitions were seen for:

458 6 Failure Analysis of Electrical Interconnections and Recommended Processes

Sn0.5Cu alloys at around 125 C; Ag-containing solders in

the range 78 to 45 C. The highest transition was with
Sn5%Ag measuring 45 C, which is ductile only above
30 C. Shock loading (such as pyrotechnic shock on a
spacecraft during launch or solar array deployment) could
cause a catastrophic brittle fracture in lead-free soldered
joints. It would certainly seem better to remain with the tin
lead alloys for low temperature applications. For operation in
liquid helium the alloy 50In50Pb may be advantageous but
no reference to its electrical properties can be found (and,
unfortunately, it was not included in the test matrix shown in
Table 6.9).
The conformal coatings all show an increase in strength
and proof stress at cryogenic temperature, and they are
brittle, exhibiting glassy fractures (Fig. 6.107). The PCB
laminates are not so brittle due to their composite structure.
For pure OFE-copper, at cryogenic temperatures, the proof
stress, elasticity and ductility are all superior to the other
tested materials.
The data collated above will be useful inputs to Finite
Element (FE) Analyses for the simulation of reliability tests
on actual solder joints during thermal loading. Solder joint
geometry signicantly influences the reliability of the
interconnection, together with the rigidity of the PCB lam-
inate, the degree of coefcient of expansion mismatch
between solder and laminate, and the overall mechanical
properties of the included materials. FE analysis, particularly
made to simulate indium-based solders operating at cryo-
genic temperatures, can now be modelled in order to opti-
mize solder joint designs and the effect of harsh thermal
changes.
6.15.13.3 Electrical Resistivity
Data concerning the electrical properties of solder alloys is
scant. Electrical resistivity values for about 50 low melting
point alloys were measured and reported for the 4.2 K
environment, mainly for SnPb and indium alloys (Bruzzone
1987) but only a few values were disclosed. A more
extensive study was considered necessary and this has been
reported (Fink et al. 2009). Here, the resistance of the fol-
lowing solder alloys were measured at room temperature, at
liquid nitrogen and at liquid helium temperature:
63Sn37Pb, 62Sn36Pb2Ag, 60Sn40Pb, 96Sn4Ag,
96.8Pb1.5Ag1.7Sn, 96.5Sn3.5Ag0.5Cu.
Fig. 6.113 Temperature dependence of specic resistance (mean
resistance) of Sn63Pb37
In addition, resistance at the same temperatures, of OFHC
Cu was also measured for comparison with values found in
literature. Resistance measurements (see circuit in
Fig. 6.109) were carried out by a four-wire method by means
of a Keithley 2182A Nanovoltmeter: the ends of the samples
were mechanically pressed into the current leads, the voltage
taps were screwed on the sample holder. For each alloy at
least 3 samples were measured at room temperature, at liquid
nitrogen temperature and at liquid helium temperature. From
each of the samples at least 5 I-V values were measured.
Measurements at 77 K took place in a liquid nitrogen
bath, measurements at 4.2 K took place in a liquid helium
cryostat.
The influence of the grain size of the solders on its
electrical resistivity was also investigated. These solders
were supplied as lengths of wire or as castings as listed in
Table 6.9. The grain morphology of four samples (as-cast,
and annealed-wire forms) are seen in Figs. 6.108.
The mean values of specic resistance for each material at
RT, LN2 and LHe temperature are depicted graphically in
Fig. 6.110; they are listed in Table 6.10. Figures 6.111,
6.112 and 6.113 show a comparison of the specic resistance
at room temperature, 77 and 4.2 K for the castings and wires
of Sn96Ag4, Sn60Pb40 and Sn63Pb37. These alloys are all
superconducting at 4.2 K, independent of their form.
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms 459

The measured values shown in Table 6.10 are very It seems important to note that the resistance of solder
similar to those alloys which have been listed in www. alloys may depend on its grain structure, which is different if
matweb.com and www.nist.gov. cast or in wire form, as shown by the microstructures and
Two alloys not measured or reported until now are: Figs. 6.111, 6.112 and 6.113.
Sn62Pb36Ag2 and Pb96.8Ag1.5Sn1.7. These are both
superconducting at 4.2 K.
 Whisker Growths
7.1 The Problem of Whisker Growth
The ultimate idea in distributed communications might be the
development of a wristwatch telephone system which pro-
vides for instant electronic communication between two
locations anywhere on Earth. This concept of a miniature
personal terminal is forwarded neither as science ction, nor
as a prophecy of future technology. It illustrates the current
rapid development of numerous communications techniques
that could make such capabilities as personal portable ter-
minals both feasible and economically viable. Two techno-
logical developments in particularcommunications
satellites and package switchingare the largest contribut-
ing factors in achieving highly responsive, user-oriented
telecommunications networks. These were the words the
author used to introduce the problems associated with whisker
growth in 1987we are now long past this milestone as today
miniaturization has enabled us to purchase wrist watch phones
with web browsers, GPS-maps and, although uncomfortable
to view for long periods, TV and on-line videos!
Metallic lms are used as electrical conductors in very
many microelectronic devices. The thickness of these lms
generally ranges from a few hundred angstroms to 25 m.
The extraordinary development of the integrated circuit has
gone hand-in-hand with an increase in reliability; however,
the speed of evolution itself brings about greater reliability
risks due to the reduction in dimensions of the devices, but it
seems that the design and construct the wrist watch-sized
portable terminals possess a totally acceptable mean time
between failures (MTBFs). The mechanism of electronic
device failure is the physical process leading to that failure.
Following decades of research, we now understood more
about electrical failure mechanisms related to the growth of
whiskers on the metallic lm integrated circuit and printed and rising equipment productivity which enabled chip
circuit conductors, or on the surface nishes applied to the
packages of electronic components. The term whisker is
used to describe bre-like crystals, usually in the form of
monocrystalline laments, these are often highly conductive
and are known to produce electronic failure by short-circuit.
Springer International Publishing Switzerland 2016
B.D. Dunn, Materials and Processes, Springer Praxis Books,
DOI 10.1007/978-3-319-23362-8_7
7
Since in-service electrical failures continue to result from
the current-carrying capabilities of whiskers, this topic is
repeatedly discussed between specialist materials engineers.
Commercial considerations are known by the author to have
prevented electronics failures analysis reports related to
whisker growths from being made unrestrictedly available.
Of particular signicance is the authors belief that little
work is being undertaken to assess the overall problems
related to whisker growth and to investigate ways by which
it can be curtailed. The writer has encountered several
instances of whisker growth that have seriously jeopardized
an area of his concern, that of the reliability of spacecraft
hardware. In two instances, between 1998 and 2000, they
involved the failure of geostationary communications satel-
lites; caused by the growth of tin whiskers on relays, or
switches, causing processors to be disabled by
short-circuiting whiskers. The majority of modern compo-
nents and assembled electrical circuits can be subjected to
screening techniques that very effectively reduce infant
mortalities. Unfortunately, problems associated with whis-
ker growths are not linked to such early failures; they occur
during the useful life of electronics and are associated with
physical metallurgical processes that lead to either catas-
trophic failures, or at best, a long-term wear-out mechanism.
The introduction of the concept of communication by
wristwatch terminals was intended to emphasize the
reduction in dimensions of electronic devices. This scaling
down since 1960 resulted in integrated circuits doubling in
complexity every yearthe so-called Moores Law. The
exponential progress of the integrated circuit can be fol-
lowed; the average number of components on a silicon chip
being 5 in 1960, 100 in 1970, 106 in 1984, and 2 108 in
1990. This was possible due to the high yields, larger wafers
manufacturers to keep the cost per area down. It seems
that now, in the 2010s, this trend is coming to an end
because of the physical challenge lithography faces: the
impossibility to scale down any further. Nowadays, com-
ponent manufacturers, stimulated by strong competition in
461
462
the market, are forced to offer devices not yet assessed, and
some of the initial problems are thus transferred to the user.
As the designed lifetime of spacecraft missions has increased
over the last decade from 3 to more than 15 years (partic-
ularly for weather and communications satellites), it is clear
that state-of-the-art electronics with failure mechanisms that
are known to include whisker growths across small con-
ductor spacings must be excluded from circuits intended to
possess a high reliability. Some spacecraft have incredibly
long lives: Pioneer 10 was 30 years, and Voyagers 1 and 2
continue to provide data after 38 years. Launched in 1977
these craft have explored Jupiter, Saturn, Uranus and Nep-
tune (in 1989)in 2012 Voyager 1 made the historic entry
into interstellar space. As of 2015, Voyagers 1 and 2 con-
tinue to transmit signals to Earth as they monitor conditions
outside of our Solar Systemmore than 5 instruments are
still working and are expected to continue transmitting data
until 2025 when there may not be sufcient power to support
the operations.
Fibrous crystals, that we now call whiskers, have been
observed and have interested scientists since long before the
advent of the electronics industry. A few naturally occurring
brous crystals were known four centuries ago, and an early
example, recorded by Robert Boyle in his famous work The
Skeptical Chymist (published in 1661), was the observation
that protuberances of silver on certain oxides increased in
length from one day to the next.
It was not, however, until the 1940s that monocrystalline
whiskers were found to grow on tin-, zinc-, and
cadmium-plated materials and cause electrical failures.
Problematic electronic circuits were rectied by maintenance
staff who were employed to change components, make
repairs to burnt-out circuits, and also clean away the bundles
of ne metallic whiskers which often appeared only as
accumulated piles of dust.
It is possible, in principle, to classify whiskers into two
categories which differ primarily in terms of their growth
mechanisms. The rst category includes whiskers that
crystallize onto a substrate from the vapour phase, the liquid
phase, or the solid phase, with growth occurring at the tip.
The second category contains whiskers that have grown
spontaneously from the solid phase, with growth occurring
at the base of the whiskers. A disquieting number of
whisker-related problems stemming from each category of
growth continue to be identied during the assembly and test
of spacecraft systems. Some examples are presented in the
following sections. They may relate to development parts
such as that shown in Fig. 7.1 which possesses a large
number of spectacular growths, or to flight model transistors
with a single whisker that has grown to produce a
short-circuit between its tin-plated lead and the metallic
capsule of the transistor. Clearly a greater vigilance must be
applied during the selection of materials for these parts. To
quote from Mr R.J. Klein Wassink, a former materials
7 Whisker Growths
specialist at the Centre for Manufacturing Technology of the
Nederlandse Philips Bedrijven B.V. (1984), The progres-
sive miniaturization in the electronic engineering elds is
being increasingly harassed by whiskers, which may cause
flashovers, short-circuits, audible noise, etc. In this sphere
more than anywhere else, prevention is better than cure.
7.2 Analysis of Failures Due to Whisker
Growth
7.2.1 Molybdenum Whiskers on Metallized
Miniature Circuits
An integrated circuit device was received for failure analysis
after the discovery that it contained some short-circuit paths
that could not be discerned by nondestructive
high-resolution X-radiography. Tests were also carried out to
check for package hermeticity before package opening. In
this test the entire package was subjected to a helium pres-
sure of several atmospheres for 15 min. No leakage was
apparent when using a helium sniffer mass spectrometer.
The package was then drilled open in a specially designed
vessel connected to the mass spectrometer. The smallest
amount of water or gaseous contamination contained within
the cavity of sealed packages can be determined by this
analytical method. The internal atmosphere of the failed
device was found to contain a large amount of water vapour.
The integrated circuit (IC) package was next de-lidded by
mechanically grinding the top until the edge weld had been
penetratedthe top was then peeled back to reveal the chip,
circuit, and microbonds.
The circuit metallic lm conductors had been prepared
during manufacture by the chemical vapour deposition of
successively titanium, molybdenum, and gold multilayers
onto the silicon semiconductor chip. The total conducting
lm thickness (referred to as metallization) was about 1 m
thick and 20 m wide.
A close SEM examination of the circuit revealed a den-
dritic type of electrochemical migration between conductors
that resulted in the formation of numerous needle-like
whiskers. These are shown in Fig. 7.2. Such growths are not
to be confused with electromigration, which occurs within
thin-lm conductors at current densities in excess of
106 A/cm2. An energy dispersive X-ray analyser (EDAX)
attachment to the SEM showed the whiskers to be composed
of molybdenum with traces of gold and chlorine. The TiMo
layer was intended to serve as a barrier to gold dissolution
into the silicon substrate and as an agent for bonding the
glassication to the metal stripe. During electrical testing the
circuit had operated under d.c. conditions, occasionally with
reversed bias. Failure was evident when the large leakage
current through the short-circuiting molybdenum whiskers
became sufciently high to cause the microcircuit to fuse.
7.2 Analysis of Failures Due to Whisker Growth 463

Fig. 7.1 SEM views of a silicon

diode chip bonded inside a
spacecraft development
flat-package by means of AuSi
eutectic. a Chip top surface
supports a 300 300 10 m
silver plate onto which is
nailhead-bonded a 38 m
diameter gold wire. Whiskers
growths extend to the wedge
bond seen in (b) (courtesy of The
Research Centre S.E.L. AG.,
Stuttgart). It is probable that this
failure mode resulted from the
release of hydrogen sulphide from
packaging material surrounding
the device. These whisker
formations are highly conductive
and, in a detailed study by
Chudnovsky (2002) can be highly
hazardous when present in
electrical apparatus as they grow
to lengths of 68 cm (please turn
the above gures through 90)

The IC device had received a 168 h (1 week) burn-in at
125 C, but was noted to fail only after several thousand
hours testing at room temperature (during the
space-qualication of this new technology component). The
presence of trace amounts of Cl, detected as a surface layer
on the whisker, suggests that contamination was present
within this device before it washermetically sealed by the
fabricator. Halogens in the presence of water and electrical
bias are known to be electrochemically active.
Occurrences of molybdenum whisker failures have not
previously been reported as far as the author is aware.
However, migrating gold-resistive short-circuits were dis-
cussed by Shumka and Pietry (1975) with co-workers in the
mid-1970s. They claimed that gold dendrites form by elec-
trodeposition from a gold halide complex solution. The rate
of deposition from the solution was limited by diffusion.
In the case of the molybdenum whiskers, three stages of
whisker propagation are suggested: dissolution of molybde-
num by a halide solution, then an incubation period for the
formation of active sites at the molybdenum surface (either at
pores in the gold or at the track edges), followed by growth
across the electrode gap. Figure 7.2b reveals Mo whiskers to
464
Fig. 7.2 Short-circuits caused by molybdenum whiskers growing from the metallization pattern on an integrated circuit
have initiated at pores in the gold surface metallization.
Normally growth would be expected from a cathodic to an
anodic stripe; in the case of the failed IC, the current had
operated in both directions, resulting in the formation and
growth of whiskers away from adjacent stripes.
Possible actions that are suggested to alleviate the growth
of molybdenum whiskers are:
1. Eliminate the water content of the package (in a closed
system, temperature fluctuations can result in moisture
condensation).1
7 Whisker Growths
2. Use contamination-monitoring equipment to check for
ionic residues (see footnote 1).
3. Cover the device circuit with a layer of electron beam
evaporated quartz (taking care to mask over the surface
areas required for microwire bonding); this manufactur-
ing step is already undertaken by the majority of
microcircuit fabricators.
1
Analyses of the contamination and residual gases in electronic
packages were performed in 1996 by Oneida Recherche Services,
Sophia Antipolis, France.
7.2 Analysis of Failures Due to Whisker Growth 465
 466
b Fig. 7.3 Observations and model of tungsten whisker growth on
non-potted TWT heater coil. a X-radiograph of failed cathode/heater
assembly (10). b and c Optical microscope views of tungsten whisker
growths. d Detail of whisker growth on part of tungsten surface that
7.2.2 Tungsten Whisker Growth Within
Travelling Wave Tubes
A critical device used in a high-power wide-bandwidth com-
munication satellite is the travelling wave tube (TWT). This
vacuum device consists of an electron gun (cathode and
anode) producing an electron beam, a slow-wave structure in
the form of a helix surrounded by magnets used for beam
focusing, and a beam collector. In operation, the radio fre-
quency signal to be amplied is impressed on the helix at the
input end of the tube and takes the form of a wave travelling on
the helix. Interaction between this wave and the electron beam
results in an amplied signal at the output end of the tube. For
operation of this device, high voltages provided by a suitably
designed power supply must be impressed on the anode,
cathode, and collectors. Isolation of these TWT elements is
achieved by the use of ceramic insulators. Dispenser cathode
assemblies with circular emitting surfaces and containing
coiled tungsten heater wires were submitted for metallurgical
examination. One sample had failed after 44,092 h life testing
because of tungsten heater lament irregularities.
The failure analysis photomicrographs (Fig. 7.3) of the
defective heater coil show that the effective cross-sectional
area of the tungsten heater wire has been reduced by subli-
mation of the tungsten metal at the high service temperatures
(about 1100 C) during the extensive life of this vacuum tube.
Close examination of the heater coil through a binocular
microscope showed the coil surface to be covered with small
whiskers as shown in Fig. 7.3b, c. The whiskers were analysed
in the SEM and found to be of the same composition as the
wire, i.e. tungsten. The whisker detailed in Fig. 7.3c has a
length of 46 m and a diameter of 7 m. A typical whisker was
closely examined in the SEM (Fig. 7.3d); it has a straight
cylindrical form with no appreciable change in diameter along
its length, nor at its tip or base.
At lower magnications it was seen that the whiskers
grew in random directions with respect to the coil surface. It
is postulated that these whiskers grow by the condensation
of tungsten emitted from a local sublimating hot spot on the
tungsten heater coil. A temperature gradient of only a few
degrees causes the tungsten vapour to condense without
chemical change at a cooler growth surface. The mecha-
nisms of growth is probably similar to that reported by
Nabarro and Jackson (1958) of mercury whiskers produced
in vacuum at 63 C, and shown schematically in Fig. 7.3e.
The presence of whiskers, like those of tungsten, is most
undesirable in vacuum tubes operating at pressures below
104 mm Hg since, once they have grown long enough,
7 Whisker Growths
showed substantial thinning caused by high-temperature vacuum
exposure. e Model of tungsten whisker. The smooth sides act as a
catchment area for tungsten atoms; they diffuse to the active tip
electrical short-circuiting may ensue. They also act as
sources for corona discharge which would cause power
surges within travelling wave tubes, causing them to switch
off. Another failure mechanism would be for the whisker to
become negatively charged, break off, and be transported by
the high internal currents within the travelling wave tube.
Under zero gravity the whisker could create extensive fluc-
tuations at the cathode emitter surface.
Modications to the cathode heater coils were success-
fully incorporated to eliminate the possibility of whiskers
nucleation. The main design change was to use a tungsten
heater coil that was coated with a thin layer of alumina paste,
and, once inserted into the heater body, was back-lled with
a potting of densely packed ne alumina particles. By so
embedding the heater, a uniform temperature gradient was
achieved along its length, and free tungsten surfaces were
not exposed to the vacuum environment.
A review of the literature indicates that this potential mode
of failure has not previously been reported. Tungsten whiskers
deliberately grown from the vapour phase have been studied
by Fumio (1974) and found, by transmission electron micro-
scopy, to have increased in length by the screw-dislocation
method of crystal growth. It is known that tungsten whiskers
have excellent electrical and mechanical properties and can be
used for eld emitters. With this objective several researchers
have attempted to grow these whiskers using different growth
mechanisms. A recent article documents the successful pro-
duction of tungsten whiskers by a vapour phase deposition
method, seeded by a nickel catalyst (Ma et al. 2012). Growth
was along the <100> direction. A complex growth process
was postulated: water vapour containing WO2(OH)2 was
initially produced by tungsten oxidation, this was decomposed
to produce a liquid tungsten fluid. Tungsten atoms then
nucleate at a site where nickel is present in sufcient con-
centration within this liquid phase; homo-epitaxial growth
follows as tungsten atoms deposit at the nucleation site on
a <100> plane. Growth follows from the tip of the nucleation
site to form a continuously growing whisker in a very similar
manner to that proposed in Fig. 7.3e but in a liquid phase
environment rather than the vacuum environment of the
cathode heater narrative.
7.2.3 Metal Oxide Whisker Precipitation
in Glass Seals
The electrical leakage of high-quality integrated circuit
packages has been investigated. Leakage currents were
experienced at locations in the glass-to-metal seal region
7.2 Analysis of Failures Due to Whisker Growth
Fig. 7.4 Metal oxide precipitation in glass seal. a External view of
electrically leaky IC package. b Transverse polished section through
lead-to-package glass window seal. c As (b) but viewed under
between the IC packages and certain of their leads. A pho-
tograph detailing these glass seal windows is shown in
Fig. 7.4a.
To discuss the failure mechanism it is necessary to briefly
outline the main steps in the manufacture of the IC packages.
The package and lead items are formed from Kovar, an iron
nickelcobalt alloy developed at Westinghouse in the 1940s
to provide the required match with the expansion properties
of a variety of soft lead and borosilicate types of glasses.
The IC package is fabricated by assembling a base plate, a
lead frame, and a ring or wall frame together, using a simple
jigging device. Ideally, the leads should be centrally posi-
tioned inside the open windows formed by the slots in the
wall frame and the base plate. Inspection of the photomi-
crographs in Fig. 7.4 reveals that this correct alignment was
not achieved in the defective devices.
The success of the matched-glass sealing process is
totally dependent on three heat-treatment stages: decarbur-
ization, preoxidation, and sealing (Pask 1948).
467
dark-eld illumination. The oxide-rich glass volumes are highlighted.
d Excessive glass-rich oxide whiskers (arrowed)
Decarburizing is performed by ring the Kovar pieces in wet
hydrogen at near 1000 C for about 30 min. The oxidation
step is the second step-during which a thin layer of metal
oxide is built up on the metal leads and component bodies.
Suppliers of Kovar today recommend the following heat
treatment for glass-to-metal sealing: oxidationheat treat in
an electric air furnace to 850900 C until parts are cherry
red (dull red heat). The length of heat cycle is approximately
3 min, but due to differences in humidity and furnaces, the
proper cycle needs to be varied. Then reduce heat approxi-
mately 10 C per minute. When parts are cooled, oxide will
be formed. The oxide may appear from light grey to black in
colour. Black is normally considered to be over oxidation
and is not desirable for a good glass to metal seal.
This is a difcult step to perform as both under-oxidation
and over-oxidation must be controlled. Too little oxide has
been seen to cause lack of bonding (hermetically leaking),
and too thick an oxide causes damage to the lead and also
results in leakage (McCormick and Zakraysek 1979). The
468
third step by which glass flows and bonds to the metal or
metal oxide is complex, and not entirely understood (Borom
and Pask 1966). The oxide layer may be composed of a
surface scale layer and an underlying layer of intergranular
oxide. The scale layer dissolves into the dried
powdered-glass (frit) coating once the glass becomes molten.
Frequently, this sealing is performed in a H2N2 furnace
atmosphere at 1000 C for 6 min. The red pieces are
removed from the heat and allowed to cool in air without any
further annealing.
The composition of the glass frit employed for sealing the
defective ICs was not revealed by the fabricator. Initially it
was thought that a soft lead glass had been used and that the
hydrogen atmosphere during sealing reduced some of the
lead oxide to its elemental (and electrically conductive)
form. Failure analyses employing ESCA have identied
such a failure mode where Pb2O3 and Pb were detected on
the surfaces of electrically leaking lead glass seals (Kelly
et al. 1984).
The defective packages were mounted in plastic and were
metallographically prepared by standard techniques. Fig-
ure 7.4b shows a typical lead seal window. The adherence of
glass to metal can be attributed to both ionic bonds and the
physical roughness of the metal window. The photomicro-
graph shows that the oxide scale has fully dissolved into the
glass. Also, the intergranular oxide, which formed at the
Kovar grain boundaries, appears to provide for a good
chemical/mechanical bonding of glass to metal. The extent
of oxide dissolution, and present of intergranular oxide, can
be clearly seen under dark-eld illumination in Fig. 7.4c.
Further examination showed some leads to be in very
close proximity to their wall frames, as for instance shown in
Fig. 7.4d. In this case, during sealing, the volume of glass
has dissolved all the metal scale from these closely adjacent
surfaces. On cooling from 1000 C, this volume has become
supersaturated with metal oxide, and a glass-rich oxide
phase has nucleated at the intergranular oxide locations on
the metal surface. This oxide phase has grown in the form of
whiskers. Qualitative EDAX analysis of the whiskers was
difcult owing to the dielectric properties of the glass. No
lead was present, and only iron was detected within the
oxide whisker.
It is assumed that the observed whiskers are composed
primarily of Fe3O4 (magnetite) which has a cubic structure,
is known to occur in whisker form, and is magnetic. These
laments (occasionally present as platelets) have a far lower
electrical resistivity than glass, and, as can be seen in
Fig. 7.4d, it is likely that they connect the lead to the IC
package wall (a distance of only 10 m) thereby causing the
measured leakage current. An industrial study focused on the
preoxidation of Kovar as the most critical step in the matched
sealing process. A tight control over the depth of scale and
intergranular oxide, before sealing, by the precise blending of
7 Whisker Growths
the H2, N2, and H2O atmosphere is now possible (Yext et al.
1983). More recently it has been established that glasses can
be supersaturated with iron ions (a constituent of Kovar) and
on cooling from the glass melt, the spontaneous crystalliza-
tion of iron oxides may occur (Somiya et al. 2013).
Magnetite is a conductor of electricity: about 2.5 104
( m)1 according to Tsuda (2000) and, even present as
minute crystals in a glass seal, is certain have been the cause
of the observed leakage currents.
The external dimensions of the IC package shown in
Fig. 7.4 are approximately 6 4 1 mm. The glass window
seals are rather miniature, with a measured surface area of
0.3 mm2. However, even these dimensions restrict the
package density of components on present-day printed cir-
cuit boards (PCBs). Component packaging technology is at
this moment being revolutionized by the development of
vapour-phase or condensation soldering. Present-day com-
ponent packages with leads that are designed for hand or
wave-soldering into the plated-through holes of PCBs are
being replaced by components of at most half their size.
With the introduction of even smaller glass window seals for
the new generation of component packages it is essential to
have strict control over the scale and internal oxidation
layers during lead-sealing operations. The new processing
renements, together with improved jigging to ensure the
proper alignment of leads within the IC, or dual-in-line
package windows, should prevent the growth of oxide
whiskers that could result in leakage currents.
7.2.4 Integrated Circuit Failure Modes Due
to ElectromigrationAluminium
Whisker Growth and Solder Joint
Voiding
Whisker growth from aluminium thin-lm semiconductors
has been noted under several sets of conditions. Blech et al.
(1975) reported whiskers after high-temperature (350 C)
annealing. These whiskers developed and grew, so relieving
the compressive stress generated in the aluminium by the
thermal mismatch between aluminium and the underlying
silicon dioxide. They were also seen to grow when the
aluminium lm was sandwiched between a silicon dioxide
substrate and a thin cover of titanium nitride. The growth
mechanism was observed directly by transmission electron
microscopy, and whisker nucleation was seen to occur at
grain boundary junctions, single grains, or hillocks. Alu-
minium whiskers have grown by a similar mechanisms to
cause extensive damage to a number of types of thin-lm
metal-oxide semiconductors. Extraordinary SEM photomi-
crographs are reported by Turner and Parsons (1982); they
show long (200 m) whiskers growing from silicon-doped
aluminium metallization pad areas adjacent to gold
7.2 Analysis of Failures Due to Whisker Growth
thermocompression bonded wires. Some whiskers were long
enough to short to the bond wires from adjacent pads, some
shorted to conductor tracks, and some were even found to be
shorted to other whiskers growing from adjacent pads.
The present author has evaluated the failure of spacecraft
semiconductor devices resulting from the growth of alu-
minium whiskers. The communications industry is com-
mitted to the development and use of Very Large Scale
Integrated (VLSI) devices that are being designed to incor-
porate submicron-sized circuit components. Futuristic
user-oriented telecommunications networks and new forms
of package switching were postulated in Sect. 7.1. Indeed,
we are now seeing the fabrication of microelectronic devices
with conductor spacings and conductor widths of only a few
thousand angstroms. As the state of the art in integrated
circuit design and fabrication approaches VLSI, the ability of
metallic interconnections to resist solid-phase whisker
growths becomes an increasingly important consideration.
The mechanism of electromigration has been reviewed by
Richards and Footner (1984), Ross and Evetts (1987),
Hummel (1994), Niehof (1995) and Morris (1996). It is
assumed that aluminium metal is transported by momentum
exchange between thermally activated aluminium ions and
conducting electrons. The aluminium ions travel in the
direction of electron flow and form hillocks and internally
stressed regions, while voids in the lm of thin conductor
areas form in the opposite areashown schematically in
Fig. 7.5.
Electromigration has been observed during the evaluation
of a novel semiconductor device. For commercial reasons
only the areas of the circuit directly related to this failure
mechanism will be included. The device was designed for
long operational life under normal ambient conditions, but
accelerated testing was made at a temperature of 150 C and
under slightly greater electrical loads than would be expec-
ted in service. Conductors were of pure aluminium, 1.2 m
Fig. 7.5 Schematic drawing of electromigration due to high current
density and associated temperature rise
469
thick, and between 7 and 10 m wide. Manufacturing pro-
cesses ensured a small aluminium grain size, and the con-
ductor stripes were nished with a thin glass passivation
layer. The results of theoretical studies predicted that at
current densities of 1.2 106 A/cm2 the device conductors
should withstand several hundred hours before failure. The
prescribed test conditions were noted to result in device
mean time before failure (MTBF) of approximately 50 h.
SEM examinations revealed the classical modes of electro-
migration failure illustrated in Fig. 7.5:
(a) Nucleation and growth of voids
The photomicrographs shown in Fig. 7.6 detail the damage
caused to the aluminium metallization due to electromigra-
tion. Mass flow has caused local mass depletion (holes
appear where the mass flow leaving these areas is larger than
the mass flow entering). Under the high current stressing
there is a transport of ions along grain boundaries. For alu-
minium this is generally in the direction of electron flow.
Figure 7.6 shows voids to have formed at what appear to be
ne grain boundaries; they are present only in regions
toward the cathodic direction of the conductor. A motion
lm study has been made of void formation in an SEM while
an aluminium conductor was actively undergoing electro-
migration (Thomas and Calabrese 1983); the most striking
observation made from the lm was that voids were actually
migrating toward the cathode. The voids were dynamic,
continuously changing shape as aluminium disappeared
from the leading edge and relled at the trailing edge. It
seems likely that as electrons and aluminium ions (in grain
boundaries) flow toward the anode there is a flow of
vacancies in the opposite direction, producing local vacancy
supersaturations. Preferred void nucleation sites appear from
Fig. 7.6 to be at grain boundaries both within the conductor
and at the edges of the conductor. Open-circuit failure is the
result of extensive void accumulations.
(b) Growth of hillocks and nucleation of whiskers
The formation of hillocks is clearly seen in Fig. 7.7a. It
would appear that such surface swellings can erupt through
the glass passivation layer as curly whiskers (Fig. 7.7b) and
may result in catastrophic short-circuit failure (NB: it is
evident from this gure that conductor masking problems
existed during device fabrication, as aluminium globules are
present within the track spacings). It is believed that the
presence of an SiO2 coating, which retards aluminium sur-
face and grain boundary diffusion, has prevented device
failure at this position by acting as a protective insulating
barrier layer (Black 1969).
470
Fig. 7.6 a and b Void formation in aluminium metallization
(c) Aluminium whisker growth
Whisker growth is quite unpredictable. Long forms may
develop at conductor edges, and they may have very low
projection angles as illustrated by Fig. 7.8. The longest
whiskers are seen to grow in regions where there is a lack of
voids; they exist at the anodic side/site of a conductor. The
accumulation of material at silica-to-silicon steps such as in
emitter metallizations, or as shown in Fig. 7.9, can also
create hillock and whisker growths.
7 Whisker Growths
Fig. 7.7 a Growth of hillocks by electromigration. b Whisker growth
from a crack in the passivation layer may result in a short-circuit
The reliability of VLSI as well as other thin-lm semi-
conductor devices may be adversely affected by increased
current densities in conductor tracks. A detailed account of
electromigration-induced failure calculations was published
by Ghate (1982). The objective of the Ghate paper was to
dene some current density guidelines. Ghate emphasized
that to ensure the reliability of VSLI devices, current densities
should not exceed 2 105 A/cm2, and failure rate predictions
must be based on a conservative approach to the use of modal
7.2 Analysis of Failures Due to Whisker Growth
Fig. 7.8 Whiskers may grow in a direction parallel to the device surface (a), or (b) normal to the surface, with a diameter of 2 m and a length of
70 m. Growth can initiate at the track edge in a location where Al contacts Si as in (c)
parameters. The electromigration test procedure proposed by
Ghate suggests that the device depicted in Figs. 7.67.9 had
been overloaded. It is recommended that the best current
density for these devices should be revised downward to
2 105 A/cm2. However, after three decades of research on
electromigration a real breakthrough to completely eliminate
failures has not been found and it is necessary to analyse the
four main parameters involved with the electromigration
phenomenon in order to signicantly improve the reliability of
Fig. 7.9 a Some whiskers have a pyramid-shape, longitudinal
striations, and bases with a diameter of the order of 4 m. b They
grow at SiO2-to-Si steps owing to electromigration
471
integrated circuits (Hummel 1994; Feng et al. 2013). The
following four conductor parameters are identied: current
density; geometric dimension; material properties; and, the
temperature of the metallic track. Feng describes, using the
Monte Carlo sampling method, how each parameter can affect
the susceptibility of the others for physically causing the
conditions for electromigration. It would appear that temper-
ature has the least influence on electromigration and for an IC,
the mean time for failure is reduced mainly by increasing
current density (seen also for solder alloy failure in Fig. 7.10).
Special techniques must be developed to minimize the
susceptibility of aluminium conductors to electromigration.
Small additions of elements such as copper and magnesium
will control and stabilize the aluminium microstructure, and
heat treatments (e.g. Heatpulse) have been promoted as
having a benecial effect on electromigration (Towner et al.
1984; Morris 1996).
Few articles could be found in the literature concerning
the electromigration (and thermo-migration) of solder alloys.
The latest (Yao and Basaran 2014) compiled several theories
in a review and has included experimental test methods
involving a flip chip test vehicle package. This paper con-
centrates on flip chip ball grid array connections made to a
PCB, where the ball pitch is 270 m. Up to around two
hundred hours of alternating current (AC) loads were applied
at various (high) current densities (between 2.1 105 A/cm2
and 1.76 106 A/cm2) and mean time to failures (MTTFs)
were measured by monitoring the evolution of resistance
with timesome results are seen in Fig. 7.10. SEM
inspection of the failed solder balls showed void growth
within the solder (mass diffusion and coalescence of
vacancies) caused by electromigration and this lead to
increased ohmic resistance and failure.
472 7 Whisker Growths

Fig. 7.10 Current density

dependence of MTTF of solder
joints subjected to 4 MHz AC
stressing at 70 C (from Yao and
Basaran 2014)

7.3 Tin Whisker Growths nucleated within 1 mm of the solder llet peripheries, and
none was present on the tinlead solder surfaces.
7.3.1 Tin Whisker Growth on a Plated Steel
Housing Adhesion test
Several small areas of the tin-plated steel were very carefully
removed from the box wall with a jewelers saw. A con-
The failed unit
trolled adhesion test, employing pressure-sensitive tapes and
One commercial-quality electronic housing containing a
a tensile testing machine, was applied to both surfaces of
failed tuner circuit was submitted for examination. The
these cut samples. The external tin plating was noted to be
donor stated that this unit was approximately ten years old.
rmly attached to the steel substrate, whereas the internal
The steel housing had a wall thickness of 1 mm and was
whisker-supporting tin surface was detached from the sub-
nished with a flash of copper and an electroplated layer of
strate with average loads of 160 g per cm width of tape. The
matt tin. Initial observations revealed an abundance of
underside of the separated tin plate is seen in Fig. 7.11b. The
whiskers growing on the inside surface of the box whereas
longitudinal markings replicate the surface nish of the steel
the external surface supported no whiskers. The unit had
plate. The crazing effect reflects the brittle nature of the
been stored under ambient conditions through its life.
plated deposit.
Visual inspection
Scanning Electron Microscope (SEM) examination
The housing was initially inspected with a binocular
The great depth of focus provided by the SEM facilitated
microscope at magnications up to 10. Certain areas were
topographical examination and measurement of the whisker
then viewed with a Reichert projection microscope at 2
growths (Figs. 7.127.16 and Table 7.1). Frequently the
magnications. At a magnication of 2, the internal walls
whiskers have a uniform direction of growth. Whisker
of the box were noted to have a furry appearance. Bright
diameters vary from 1 to 4 m, and their surfaces are
flashes of reflected light were transmitted from random spots
characterized by the presence of longitudinal striations,
on the boxs surface as its orientation was changed slightly.
which run parallel to the direction of growth. Occasional
At slightly higher magnications, the whisker colonies
whiskers, as in Fig. 7.13, are noted to have very irregular
became clearly visible. Figure 7.11a shows whiskers to have
growth directions. The changes in growth direction, or kinks,
grown sufciently long to have collided with an exposed
are usually associated with an angular displacement of about
inductor coil, clearly highlighting the possibility of electrical
45. The variation in whisker diameter is similar to the
short-circuiting. Several whiskers were measured with a
average size of the nodular-like grain pattern seen on the
graticule attachment to the light microscope and were found
plating surface in Figs. 7.13 and 7.16. The incipient whisker
to have lengths exceeding 2 mm. The internal surfaces of the
seen in Fig. 7.13 has a rounded end, quite different from the
housing had been directly soldered in several positions. It is
stepped end of that in Fig. 7.15, which has the appearance
interesting to note that no whiskers were observed to have
of a rotating growth step. The change in diameter of a
7.3 Tin Whisker Growths
Fig. 7.11 a. Whisker growth clearly visible adjacent to inductor and
coil at 6 magnication, and, b Underside of thin tin-plating after
detachment from steel plate following adhesion test
whiskers shaft, as in Fig. 7.14, appears from its surface
morphology to be due to the presence of an edge dislocation.
By bending the sample across a saw cut that penetrated
90 % of the steel housing wall thickness it was possible to
detach a portion of the internal-plated tin layers. Likewise,
ne flakes of the outer-plated layers were released, peeling
being accomplished by means of hand-held tweezers. These
detached flakes were examined in the SEM and later by
metallography. Figure 7.17 details one flake and a region
that has torn immediately adjacent to the root of a whisker. It
is important to note the absence of local subsidence in the
plating layer and that the whisker appears to have its origins
just below the plating surface. The X-ray energy dispersive
analyser (EDAX) attachment to the SEM was used to make
point analyses of the plating and at several positions along
the length of individual whiskers. The resultant characteristic
473
Fig. 7.12 Scanning electron micrograph of whiskers showing uniform
diameters along their lengths
Fig. 7.13 Detail of a whisker with a very irregular growth direction,
and an incipient whisker
X-ray peaks revealed the presence of only tin in the whisker;
no other elements present in excess of about 50 ppm were
observed. As expected, copper was present within the plated
layer, having been initially deposited as a flash on the steel
surface.
Metallographic examination
A small part of the steel container was carefully removed
with a jewelers saw, mounted transversely in a
room-temperature curing resin, and then ground and
474
Fig. 7.14 The occurrence of a whisker changing its shaft diameter.
Note the uniform longitudinal striations
Fig. 7.15 Apparent fracture surface of a large-diameter (approx. 4 m)
whisker. The thin nearside whisker has a diameter of approx. 1.5 m
polished to a 1 m diamond-paste nish. The mount was
lightly etched, then viewed with a Reichert projection
microscope. The inside tin-plated steel surface is seen in
cross-section in Fig. 7.18. The section reveals that a thin
7 Whisker Growths
Fig. 7.16 Base of whisker has roughly the same diameter as the
nodules on the tin plating. Although the volume of the tin whisker is
quite considerable, there are no signs of local subsidence on plating
surface
strike of copper had been applied to the steel before the nal
tin-plating. The flakes of plating that had been detached for
the SEM work were also microsectioned. Figure 7.19 reveals
the cross-section of a whisker emanating from a grain in the
tin-plating; its diameter is less than that of the grain. This
gure again exemplies the absence of local thinning of the
tin plate to account for the volume of the whisker. It is
apparent that the detached layer of plating fractured within a
thin network of Cu6Sn5 intermetallic grains. Such inter-
metallics have growth rates that are time- and
temperature-dependent, and their presence within the subject
specimen is consistent with the age of this electronic
housing.
An interesting observation can be drawn from Table 7.1;
tin-plate having a thickness greater than 1820 m could be
a mitigation factor likely to avoid the nucleation and growth
of tin whiskers. Data from industry indicates that thicker tin
plating shows a lower propensity for whisker growth
(Bradley 2007), this has been postulated by other researchers
(Hwang 2014), certainly the majority of spontaneous whis-
ker growths have occurred on thinly plated tin (below
10 m).
7.3.2 Tin Whisker Growth on PCB and Other
Electronic Materials During
Thermal Cycling
High failure rates in electronic boxes may be caused by large
thermal stresses in the soldered joints (thermal fatigue) of
electronic components mounted on PCBs, as discussed in
Sect. 6.5.3. The greatest problems that were encountered
resulted from the high value of the coefcient of expansion
in the thickness, or z axis, of the PCB. Spacecraft PCBs are
7.3 Tin Whisker Growths 475

Table 7.1 Plated coatings and whisker distribution on steel electronic housingfrom direct measurements
Steel box Plating thickness (m) Tin whisker features
Copper strike Tin nish Length (m) Total population density (per mm2)
Max Average
Inside surface <1 810 2000 500 1800
External surface 12 1820 none none 0

constructed from glass breepoxy composite sheets that are
laminated with copper foil. Etching away the copper in
unmasked regions leaves behind the copper circuit which
will be drilled, to facilitate plated-through holes which
electrically interconnect both faces of the board. The
plated-through hole always contains a thick layer of copper
followed by a nal nish of tinlead plating that is eventu-
ally fused to form a long-life, highly solderable surface
(Sect. 4.7). More recently, as a result of the RoHS directive,
some manufacturers have decided to use pure tin nishes on
their PCBs. This is standard practice for certain commercial
products although for space projects the ECSS and NASA
requirements specically forbid this nish (Dunn 2012). It is
the special thermal cycling environment of the spacecraft
that makes this option extremely dangerous, as it can pro-
mote the growth of tin whiskers from sites within pure tin
plated-through holes. Perpendicular to the laminate, the glass
weave does not restrict the resin expansion, and in this

Fig. 7.18 Cross-section of surface supporting whisker growth (see

also Table 7.1)

Fig. 7.17 Examination of the detached flake (a) reveals that whisker Fig. 7.19 Tin grains are revealed by light etching in a 0.2 Nital
growth has initiated from just beneath the tin-plated surface (b) EDX mixture. Further staining in ammonium peroxide indicated the presence
analysis of point 1 revealed a trace of copper in the tin, whereas points of a thin, blocky network of coppertin intermetallic along the detached
2 and 3 identied pure tin surface
476 7 Whisker Growths

z-direction the coefcient of expansion is an order of mag-

nitude greater than along the width or length of the board
(see Table 6.1). The whiskers presented in Fig. 7.20 were
observed to be growing on type G10 PCBs that had a
thickness of 1.5 mm. When heated and then cooled through
the temperature range 20160 C these boards were seen to
increase and then decrease in thickness by between 35 and
45 m. The great mismatch between tin and epoxy resin
during thermal cycling causes the plated-through holes to be
subjected to tensile and compressive forces. It is postulated
that the compressive cycles are responsible for the progres-
sive extrusion of needle-like whiskers. Observations of
whiskers during their growth showed that the morphology of
their tips remained unchanged, and it is therefore deduced
that growth occurs by the accretion of fresh material at their
bases (Fig. 7.21). Electronic circuit sizes are rapidly
shrinking, while their capacity and capability continue to
advance. The separation distance between tracks on a
spacecraft PCB may be as little as 0.4 mm, and components
can be mounted directly above plated-through holes so that
the possibility of a short-circuit being caused by whiskers
growths becomes very real. Tin whiskers grow to a length of
2 mm and are strong enough to penetrate a layer of con-
formal coating (McDowell 1993).
Test programmes performed by JAXA on lead-free chip
capacitors were designed to replicate the space environment
(Nemoto 2007). These capacitors were procured from two
manufacturers, their metallisations consisted of pure tin
plating on a nickel barrier layer on copper. Half of the
samples from each manufacturer were solder dipped by
machine to remove the pure tin. All samples were thermal
cycled, but again, half the numbers from each variant were
thermal cycled in air, the rest under a vacuum of 104 Pa.
Thermal cycles for both air and vacuum environments were
+125 to 40 C, dwell times being 60 min and one complete
cycle time of 130 min.
The results showed that samples, from both manufactur-
ers when pre-tinned in a solder bath, produced no indication
of whisker nucleation.
All of the as-received samples grew whiskers when
thermal cycled (see Fig. 7.22), actual results being measured
as follows: in air: 420 cycles [52 m max, average 30 m],
1040 cycles [44 m max, average 29 m] and after 2000
cycles no further growth.
Under vacuum: 420 cycles [29 m max, average 17 m],
1040 cycles [67 m max, average 43 m] and after 2000
cycles no further growth.
To summarize Nemotos ndings:
Fig. 7.20 Micrographs of typical tin whiskers growing within the
electroplated-through-holes of a PCB. One 1.2 mm diameter hole
The rate of tin whisker growth under vacuum is greater (a) was carefully bisected with a diamond saw to facilitate SEM
than when thermal cycled in air. examinations (bc)
7.3 Tin Whisker Growths
Fig. 7.21 Microsection of
plated-through hole has passed
through pyramid base (arrowed).
Etching shows tin plate to vary in
thickness from 7 m in hole to
13 m on pad. The intermetallic
is approximately 1.5 m thick.
a Etched in ammonium peroxide
to reveal base copper and plated
copper. b As (a) etched in 2 %
Nital to reveal tinCu6Sn5
thickness
Each temperature excursion produces a transverse stria-
tion on the whiskers.
These striations are far more discernible when cycling is
in air.
The mechanism for growth is suppressed after about
1500 cycles.
Pre-tinning in a solder bath effectively prevents whisker
growth.
From these ndings it would appear that electronic units
containing tin-plated parts will be more susceptible to
whisker induced short circuits, or arcing, than their
ground-based counterparts.
The Nemotos observed reduction in whisker growth rate
at about 1500 thermal cycles is very similar to that seen by
Ditees et al. (2003) for growths measured on tin-plated alloy
42 (i.e. FeNi42 CTE 4.3 106 K1; far lower than Sn CTE
23.0 106 K1).
Dittes considered that under thermal cycling, tin whisker
length may have a parabolic relationship to the number of
cycles. Indeed, the graphical display of results showing
whisker length against time presented in the Nemoto and
Dittes papers are almost identical.
The CTE of alumina ceramic chips is between 6 and
7 106 K1. These large CTE mismatches are expected to
477
generate higher compressive stresses than, for instance
copper-leaded interconnectors (17 106 K1) which, hav-
ing a similar COE to tin, do not generate thermal-cycling
induced whisker growth (only Cu6Sn5-induced growth, as
will be described in paragraph 7.4.3).
Calculations have been made to evaluate whether the
levels of stress produced during the thermal cycling of
ceramic (Al2O3) capacitors and resistors will exceed the
elastic limits of the outer tin-plated end-terminations (Moser
2002). It was found that all the commonly used thermal
cycling tests (see also Sect. 3.2.2.6) far exceed the elastic
limit of tin. Per the ECSS Q-ST-70-08, the cooling step from
100 to 55 C was calculated to produce a theoretical stress
of 160 MPa, whereas the actual elastic limit of pure tin is
around 14 MPa. The potential in-orbit thermal cycles seen
by operational spacecraft are often between 5 and 35 C
and even this typical cycle is expected to cause the elastic
limit of tin to be slightly exceeded.
Some X-ray diffraction (XRD) residual stress measure-
ments were instigated by the writer as part of an extensive
failure investigation involving the growth of tin whiskers on
passive component terminations. Interestingly, as-received
components that had been stored at room temperature for
8 years and had not grown whiskers were found to possess
residual stresses in their tin-plating of about 7 MPa (i.e.
478
Fig. 7.22 SEM observations made on thermally cycled tin
plate-on-nickel-on-copper metallised ceramic chip capacitors. a shows
whiskers after 1040 cycles in air [max length 44 m, average 29 m],
b after 1040 cycles under vacuum, [max length 67 m, average 43 m].
Thermal cycles can be followed by noting the distance between the
7 Whisker Growths
transverse striations encircling each whisker, they are more pronounced
in air c than under vacuum d. The distance between the striations
appear slightly greater when cycling is under vacuum. By kind
permission of Norio Nemoto
7.3 Tin Whisker Growths 479

compressive stresses below the elastic limit of tin). How-

ever, the same components that had been subjected to 500
cycles between 55 and +100 C were found to have grown
whiskers up to 50 m in length; and, these samples were
found by XRD to possess compressive residual stresses in
the longitudinal direction of -34 MPa 4 MPa. The XRD
measurement was made following the cooling part of the
cycle (100 C to ambient) when the tin can be expected to be
in compression.

7.3.3 Tin Whisker Growth on Crimp

Termination Devices

Many of the small devices used for spacecraft electrical

interconnections are selected from materials processed by
conventional techniques and purchased from commercial
suppliers. They include terminal posts for PCBs, grounding
points, and wire terminals and lugs for crimping operations.
These small parts are frequently machined from either copper
or brass and simply tin-plated to achieve a reasonable sol-
derability and protection from surface corrosion. A surpris-
ingly large number of devices have been observed to support
potentially hazardous growths of tin whiskers. The problem is
illustrated by standard crimp terminals, purchased
off-the-shelf in their thousands from the largest European
manufacturer of these devices and later distributed to space-
craft construction companies. Two of the defective devices are
shown in Fig. 7.23a. Despite the fact that operators had fol-
lowed the many quality control steps to perfect subsequent
crimped joints, neither operators nor inspectors had examined
this device closely enough to expose the presence of whiskers.
In this case they were discovered by chance during a
non-routine test. A metallurgical report was led in 1978, and
the same devices have since been periodically re-examined.
A random elds of view showing the whiskers growth in
1978 is detailed in Figs. 7.23b. The whiskers occurred
around the rim of the cylindrical crimp barrel; they vary in
diameter from 2 to 13 m. Whiskers growing on the inside Fig. 7.23 a. General view of crimp terminal lugs manufactured in
barrel surface were seen from silhouettes to have lengths of January 1978. The brazed lengths of barrels are arrowed. b General
view of whiskersample age since manufacture is six months.
0.15 mm. At that time the devices were considered to have Location of this whisker is at rim of barrel
had a shelf life of six months.
The 1983 inspection noted whiskers to have grown on all
surfaces of the device. The longest whiskers had now
The barrel portion of the crimp device is known, from
attained a length of 0.7 mm.
micro-sectioning samples from the same batch as those in
An inspection in 1985, seven years after the date of
storage, to have been manufactured from copper sheet,
manufacture, reveals that the longest whiskers appear to still
stamped-out, formed, and then brazed to make the continu-
have maximum length of 0.7 mm. However, the number of
ous cylinder.
whiskers having a length within the range 0.30.5 mm has
The latest inspection of the same crimp barrels, stored at
increased (see Fig. 7.24). Although the majority of whiskers
ambient temperature in a desiccator at 40 % RH, in 2015
have large length-to-diameter ratios, there are many elon-
(37 years after they were manufactured), shows the internal
gated nodules and distorted short growths that burgeon from
barrel plating to support 1.2 mm whiskers, and the external
the nodular tin plate.
480 7 Whisker Growths

Fig. 7.24 a and b. General

views of whiskers after seven
years of growth. Although the
maximum whisker length remains
at 0.7 mm there is a larger
population within the 0.30.5 mm
range. [Latest measurements by
optical microscopy in 2015, show
these growths to have reached at
least 1.2 mm, however, additional
growths now appear on the
external face of the barrelthese
are about 1 mm in length and
have a high density].
c Appearance of a diamond
hardness impression made to a
tin-plated copper surface having a
thickness similar to the crimp
barrels. Short whiskers can be
seen erupting from the
compressed tin-plate, some
distance from the impression
(Shibutani 2009). See also
Fig. 7.30
7.3 Tin Whisker Growths
Fig. 7.24 (continued)
side (in the region of the brazed joint) to have grown a high
density eld of whiskers having lengths of up to 1.0 mm.
These observation will certainly highlight the long-term
risk that tin-plated crimp terminations pose. Both
short-circuit and arcing potential failure modes must be
mitigated for, possibly by either selecting fused tin crimp
barrels, or by ensuring that whisker growths are contained
within protective shrink-sleeving.
One unknown is the effect that crimping will have on the
external sides of this often-used crimped barrel. The
crimping process is discussed in Sect. 6.4.3. Clearly large
localized compressive forces are applied by the crimping
tool in order to deform the barrel and compress/cold-weld
the enclosed bare strands of wire. Severe compression and
deformation of the external tin-plated surfaces is akin to the
forces applied to tin-plate when it is tested for
micro-hardness. Shibutani (2009) is one of several
researchers who have attempted to identify whisker-prone
surfaces by implanting a hardness impression into the
tin-plating (see Fig. 7.24c). Shibutani introduces a
creep-based model to account for whisker growth. The high
pressure exerted during the hardness test is said to cause tin
atoms to diffuse away from the compressed volume and form
extruded whiskers; this mechanism has probably accounted
481
for the severely deformed tin seen in Fig. 7.30. The external
work applied by the indenter is then equated to the increase
in surface energy as whiskers are grown. The indentation test
is said to also replicate the micro-stresses induced by the
volumetric growth (with time) of Cu6Sn4 intermetallic and
its influence on whisker growth.
7.3.4 The Nucleation, Growth and Mechanism
of Growth of Tin Whiskers
Results from a C-Ring Test
Programme
The results of a comprehensive laboratory study of tin
whisker growth have related their ubiquity to particular types
of electro-plated tin and characteristics of the substrate
material (Dunn 1987). Forty C-ring samples, supporting
different tin-platings and substrate variants, were testedthe
programmes 32nd year results are fully described by Ash-
worth and Dunn (2015). One C-ring sample and a sketch
showing its physical geometry is seen in Fig. 7.25. Com-
pressive stresses were calculated and applied to the inner
surface of each the C-ring by torqueing the threaded links.
The C-ring substrates were machined from brass and steel.
482
Fig. 7.25 a Overall dimensions
for C-ring specimens,
W = width = 25 mm,
h = thickness = 2 mm, a = outside
radius = 12.5 mm, b = inside
radius = 10.5 mm. b Optical
photograph of C-ring. Some long
whisker growths can be discerned
on both the inner and outer
surfaces
They were over-plated with 3 varieties of tin (normal; high
current density; and, from a contaminated bath containing
organic particles), half of the samples supported a barrier
layer of copper between the substrate and each tin plating.
One sample, supporting each material variant, was placed in
an inert gas oven so that the outer tin plated layer could be
fused at 250 C (pure tin melts at 232 C).
The main ndings recorded at the 32-year test period
examination are:
(a) Compressive stresses applied to tin plate did not
accelerate whisker growth rates. Whiskers were seen to
grow over all C-ring surfaces.
(b) Incubation periods before whisker growth are shown to
be strongly dependent on the substrate material.
Tin-plated brass has a short (several days) nucleation
period and produces whiskers with growth rates of
8 m per day, whereas tin-plated steel takes several
months to exhibit short whiskers.
(c) Samples with a copper barrier layer saw normal tin
plating to nucleate whiskers within 5 months and these
grew to lengths between 1 and 4.5 mm.
7 Whisker Growths
(d) Plated tin, deliberately contaminated with occluded
organic particles, on copper gave an unexpected result:
one sample only nucleated whiskers during the nal
12 years of storage, and these whiskers grew over the
following 20 years to lengths of 1.8 mm.
(e) Nodular eruptions are frequently observed to precede
those whiskers having a short incubation period. Most
whiskers have parallel sides, but they may contain
sharp kinks and can vary in diameter from around 6 m
down to only 6 nm. Whiskers are seen to possess
numerous morphologies: from 1 to 20 m in diameter,
some are straight but many have numerous directions
of growth. Their density appear to vary from about 0.1
to 200/mm2. Whiskers were noted to change their ori-
entation during growth, and by rotating became
ensnarled with other growths (see Fig. 7.26).
(f) The results of the longest whiskers as a function of
storage time are shown in Fig. 7.27. None of the fused
tin-platings nucleated or grew whiskers during the
32 years storage period (i.e. fused tin on: brass, steel,
copper-plated-brass and -steel).
7.3 Tin Whisker Growths 483

Fig. 7.26 Attention is drawn to

the whiskers a and b. The
white arrows show where they
emerge from a tin-plated copper
surface. These SEM images were
taken (A) 57 days and
(B) 181 days after tin plating.
During this 124 day period the
whiskers increase in length are
125 and 110 m respectively for
a and b (approximate growth
rate of 1 m per day). During
growth whiskers have continually
rotated in the directions shown by
the black arrows on (A).
(C) shows the detail of a whisker
having a length of 700 m.
Rotation has caused it to become
ensnarled with other whiskers.
Relatively large latent energy is
contained in this bowed volume

(g) Metallography revealed that, despite the relatively large
volume of tin material being ejected in the form of
nodules and long whiskers, there was no evidence of
depressions or local subsidence in the plated layer.
Intermetallic growth after 32 years storage is shown,
for 4 samples, in Fig. 7.28. These phases have been
identied, for all samples and are described in Ash-
worth and Dunn (2015).
A ve-stage model for whisker growth has been pro-
posed, based on the premise that local micro-stresses can be
eliminated by the ejection of tin atoms from the plated layer
at sites of rotating screw dislocations. It is also postulated
that a long-range atom-transport mechanism (diffusion)
sustains these growths (these stages of incubation and
growth are shown schematically in Fig. 7.29).
484 7 Whisker Growths

Fig. 7.27 Graph comparing the

length of the longest whisker as a
function of storage time at room
temperature for unstressed normal
tin deposits on brass and on steel
with and without a copper
interlayer present. For the
complete set of results see
Ashworth and Dunn (2015).
None of the fused tin samples
grew whiskers

Fig. 7.28 Back-scattered electron images showing the extent of
intermetallic formation for normal Sn deposits on steel after storage at
room temperature for 32 years: a 5 m Sn deposit, there is evidence of
a SnFe intermetallic at the plating-substrate interface. b 5 m Sn
deposit with a 3 m Cu interlayer, here the intermetallic phase has been
identied as Cu6Sn4no other phases are present. c The fused 5 m Sn
deposit has alloyed with steel to form FeSn2. d Fused 5 m Sn deposit
with 3 m Cu interlayer; here both Cu3Sn and Cu6Sn4 are present at the
copper interface along with some voiding; islands of Cu6Sn4 within the
tin (Ashworth and Dunn 2015)
7.3 Tin Whisker Growths
Fig. 7.29 Schematic diagrams to illustrate tin whiskers growing on
tin-plated brass substrate (top left to bottom right): nodular eruptions
possessing no preferred crystallographic texture are pressed from
individual grains of tin by localized residual stress from plating;
Frank-Read sources having suitable orientation generate edge disloca-
tions; edge dislocation is rotated to a screw segment and Sn atoms
attach themselves to the edge dislocation causing screw segment to
7.3.5 Some Properties of Tin Whiskers
The following sub-sections are mainly derived from Dunn
(1987 and 1989) with additional data compiled from col-
leagues and recent citations.
(a) Attachment to plated tin
(i) Tin whiskers can be physically dislodged by
mechanical force. The sliding spring-loaded con-
nector nger has mated and de-mated several times
to the PCB track causing whiskers to nucleate and
grow in an elastically stressed region (up to 200 m
from the connectors edge). Additional mating
cycles have caused the plastically deformed mound
of tin to advance into the whisker-growth region
causing them to detach and create whisker debris.
485
rotate and become displaced outwards as a whisker; intermetallic
growth, Cu6Sn5, causes additional compressive micro-stresses in the tin
as it grows at the expense of Sn and Cu; as the intermetallic is less
dense than the metals it forms from, there will be volumetric increases
and continuous stress-relief as more Sn is released into the whisker; Sn
grain growth (or recrystallization) at ambient temperature also creates
additional micro-stress as smaller grains are annihilated
Such broken whisker dust can bridge the spac-
ings between tracks of un-coated PCBs with the
risk of causing cross-talk or short-circuit failures.
(ii) The theoretical natural frequency of tin whiskers
depends on the individual whiskers length and
radius. For instance, one of the largest whiskers
seen in the C-ring experiment (Sect. 7.3.4) having a
length of 4.0 mm and a diameter of 4 m can be
calculated to possess a natural frequency of 36 fn
(Hz). An average whisker may be one having a
length of 2.0 mm and a diameter of 2 m, so
having a fn of 71 Hz. A short whisker with a length
of 200 m and a diameter of 4 m has a fn of
14,260 Hz. In fact the complete range of fn is
expected to vary from 36 to 14,260 Hz.
486 7 Whisker Growths

Fig. 7.30 Connector contact to

tin-plated PCB track revealing the
location of: a An electrical short
circuit caused by several >1 mm
tin whiskers that originated from
the region of indentation and
severe plastic deformation caused
by a sliding contact. b Small
nodules and relatively long
(*100 m) whiskers have also
grown in a region, away from, but
surrounding the deformed tin.
This eld failure had seen a
one-year life and had been
subjected to thermal cycling and
vibration (courtesy of Karl Ring,
ZVE)

The vibration spectrum encountered by electronic pack-
ages during, for instance, a spacecraft launch is severely
limited by deliberately isolating a unit or the damping effects
of metal structure, bolts and packages, so that the very high
whisker resonance frequencies will never be encountered.
Only the range from 102000 Hz need be considered.
Samples of whiskers growing on electronic tin-plated
items were exposed to the predominant frequency ranges
and magnitudes expected to be encountered during elec-
tronic box ground handling, and during launch when much
additional vibration is derived from acoustic noise (i.e. the
sound eld developed by the extreme turbulence of the jet
7.3 Tin Whisker Growths
exhaust downstream from the launchers motor). The sam-
ples were bonded to a thick steel plate bolted to the shaker
head of a 7 ton Ling vibration tester. High magnication
photographs and sketches recorded the number and location
of whiskers at the start of testing. Low level testing was
initially performed, followed by inspections when no dam-
age to whiskers was recorded. The samples were then sub-
jected to two sweeps in the 102000 Hz range of
frequencies, at a speed of 2 oct/min. These frequencies were
split into three groups of accelerations:
1016.6 Hz 38 mm peek-to-peek
16.678.1 Hz 200 cm/s
78.12000 Hz 100 g
No whiskers were seen to break or become permanently
deformed during sinusoidal sweeping through their natural
frequencies.
Recent tests conrm these ndings (Hashemzadeh 2005):
exposures culminating with the hardest vibration test for
electronics applied by the automotive industry (2000 h at the
highest frequency 2000 Hz) failed to damage tin whiskers.
(iii) Long-term vibration at dened frequencies was made
with the same samples exposed to ii). The samples were
tested on a Dunegan Model 4501 laboratory system at a
series of pre-selected frequencies: 50, 100, 200, and
250 Hz, each at 6 g and for a duration of 60 s.
Undoubtedly, whiskers having the natural frequencies
of this test series existed on the samples under test, but
again by meticulous microscopical inspection no evi-
dence of detachment or damage was found. It was
considered that the fatigue lives of the whiskers was
never exceeded (endurance limit stresses not reached,
number of fatigue cycles did not reach end of life, and
the whiskers are so homogenous and free from surface
discontinuities that no surface stress concentrations
were available for fracture initiation.
(iv) Ultrasonic vibration in liquidthe same samples were
nally placed in a metal basket and lowered into a
commercial ultrasonic cleaning bath lled with iso-
propyl alcohol. The bath was operating at nominal
frequency in the range 2025 kHz (above the frequency
of human audibility). The cleaning cycle was main-
tained for 2 min. Visual inspection of the samples
showed that all the whiskers on exposed flat surfaces
had been removed. Whiskers in holes or re-entrant
spaces appeared generally intact.
(v) The effect of mechanical shocksuch shock loading
may occur when electronics are exposed to rough
handling during transportation, accidental dropping
and at the moment of spacecraft pyrotechnic
487
actuation. New samples were subjected to 50 shocks
at a displayed level of 2060 g. Surprisingly, as for
the vibrated samples, no whiskers were noted to
break of become deformed. This indicates that nei-
ther their yield of ultimate strength had been
reachedpresumably due to the very low inertia of
each whisker.
The ndings of (Hashemzadeh 2005) also showed no
damage to whiskers shocked for 6 ms and accelerations of
500 and 1000 m/s2.
(b) Mechanical properties
Further laboratory investigations have provided data related
to the mechanical properties of tin whiskers. The whisker
samples had grown directly from the surfaces of a tin-plated
mild steel electronic housing seen in Fig. 7.12. They were
extracted from the housing surface using micro-mechanical
tools and placed onto glass slides. Individual whiskers were
attached to slides using epoxy or low temperature melting
indium solder. Cantilever beam loading and tensile loading
was performed using microne gold wire as weights.The
mechanical properties of tin whiskers were found to be
approximately:
0
Young s modulus E 885 GPa
Ultimate tensile strength UTS 0:8 kgf=mm2 8 MPa
The observed scatter in some results is likely to be due to
variations in each whiskers unit-cell parameter (different for
each of the ve crystallographic growth directions of tin
whiskers) and to difculties in measuring the absolute
cross-sectional area of whiskers supporting flutes (with
rosette-shaped sections).
(c) Electrical properties
Electrical failures attributable to tin whiskers include: heat
pacemakers, F-15 aircraft radar system, US missile elec-
tronic assemblies, air-to-air missile systems, Patriot missile
systems, more than ve US commercial communications
satellites, a nuclear reactor shut-down, rocket motor failure,
and, commercial motor car failures due to tin whisker shorts
leading to mortalities and the recall of tens of thousands of
vehicles (Sood et al. 2011; George and Pecht 2013). Sur-
prisingly, these problem notications emanate from the
US where there seems to be far greater transparency when
eld failures occur. Other, anecdotal accounts of whisker
failures abound, they are not documented in the open liter-
ature due to commercial pressures. The number of failures
and various modes of whisker-induced failure necessitates
that all characteristic properties of whiskers are understood.
488
Mitigation against the risk of failure can then be attempted
(see section related to creation of a lead-free control plan).
The most comprehensive compilation of world-wide tin
whisker literature references, illustrations and reports, can be
found on the NASA website: http://nepp.nasa.gov/whisker/.
The electrical properties of tin whiskers were found (Dunn
1988) to be dependent on their actual diameters.
Room-temperature resistivity can vary from 11 to 60 108
. Actual current flow through 3 m diameter whiskers can
reach 32 mA. Hashemzadeh (2005) using similar test meth-
ods found whiskers with a diameter of 5.5 m to also reach a
critical current of 32 mA before fusing. Current-carrying
capacities appear to be around 26,000 A/mm2 (Hunt 2012).
The resistivity of tin whiskers were found to be lower than for
pure -tin. Signicant damage can be caused to space sys-
tems when whiskers bridge electrical paths as illustrated by
Fig. 7.31 and its caption.
High-voltage discharge can emanate from the sides and
tips of tin whiskers situated on electrically charged surfaces.
Spark discharges (i.e. corona) and nonluminous point dis-
charges from whiskers have been demonstrated in a scan-
ning electron microscope using photographic methods
(Fig. 7.32).
Fig. 7.31 Graph to illustrate the effect of whisker diameter on the
possibility to cause a short circuit (from plots of mA versus mV for four
whiskers)the relationship is linear until heating effects cause whisker
to burn-out. Measurements made on actual whiskers (Dunn 1988).
Zone C will cause the maximum encumberence to spacecraft electron-
ics during ground testing, and possible total loss if occuring during
in-orbit operation
7 Whisker Growths
Whiskers can be attracted by electric elds and the
degree of attraction has been found to be a function of the
electric eld strength, the angle at which the whisker grows
from a substrate and the dimensions of the whisker. This
electrostatic attraction will increase the likelihood of shorts
occurring between whiskers and conducting surfaces at
different electric potential (Hunt 2012).
Tin whiskers can short-circuit under vacuum to create
tin-plasmas which have been seen to carry tens to hundreds
of amperes and blow spacecraft fuses and destroy relays
(Mason and Eng 2007). Paschen curves showing the dis-
tance (between whisker and another electrical conductor)
multiplied by the pressure of the environment against volt-
age will predict the minimum voltage below which plasmas
will be initiated. However, these investigators found by
experimentation that sustained plasmas can form in vacuum
at a power supply DC voltage of only 4 V. Under ambient
conditions the same tin-plasma can form at 28 V. (Paschen
curves were discussed in Sect. 6.15.4 with respect to corona
discharge between soldered joints on PCBs). Figure 7.33
shows whiskers to be making short circuits between the
bright-tin plated nickel-iron terminal pins of a relay. Relays
from the same lot were procured for an ESA scientic space
project in the late 1990s, assembled into its processor, but
then dismounted a few months before launch date. The
retrospective demounting followed information, provided by
an alert system, that the same relays had caused four Galaxy
communication spacecraft to be knocked out of service due
to the tin-plasma induced mode of failure. The defective
relays had been installed into on-board computers, known as
spacecraft control processors, to control the flow of elec-
tricity. Once they stop working the satellites cannot maintain
their orientation in space, so preventing their antennae from
pointing down to Earth. It is probable that whiskers only
nucleated after the launch of the communication spacecraft,
they are likely to have grown whilst in orbitthis is only the
authors supposition based on the fact that all the operators
and inspectors involved with the spacecraft build are trained
and certied. Component assembly operations are performed
under strict conditions of cleanliness and involve numerous
visual inspection steps at high magnication. At this point in
time dust, wire debris and whisker growths are likely to have
been identied and cause for flight unit rejection.
(d) Physical appearance
The surface morphologies of tin whiskers can be summa-
rized as follows. They range in diameter from about 6 nm to
7 m (although short blocky growths have diameters in the
order of 20 m). They are usually characterised by the
presence of striations parallel to their growth direction. Some
high magnications indicate that the striations may be
7.3 Tin Whisker Growths
Fig. 7.32 The tip of this tin whisker supports a ne point from which
sparks discharged for more than 5 min. The emissions, all at 20 kV,
have the appearance of discrete white particles or a ne spray as
related to the shape of the surface orice through which they
have grown. Occasionally, growths have very irregular
growth directions and odd changes in cross-sectional area.
Many appear to grow from nodules and some have
pointed-needle like shapes. As seen in Fig. 7.26, whisker
growth is accompanied by a rotation of the shaft about a line
drawn normal from the nucleation point on the tin-plated
surface. This rotation, or gyratory movement can cause
whiskers with low projection angles to become ensnarled
with other whiskers and these may be have a greater
potential for causing electrical short-circuits.
Whiskers surfaces are now known to be covered with tin
oxide (Hillman and Chumbley 2006). A dual oxide lm,
consisting of stannous and stannic oxides have been found
on both new and old whiskers (Dunn and Mozdzen 2014).
Whiskers removed from 48-year old electronic hardware
could be bent so that their brittle oxide lms could be
measured in situ, these were found to have a thickness of
489
shown toward the end of the 5 min display (d). No damage (corrosion
or melting of the tip) was observed at the end of this event; a 10 s.
b 30 s. c 40 s. d 200 s
between 20 and 30 nm, both from FIB sectioning and, as
shown in Fig. 7.34, from XPS analysis. Newly grown
whiskers, only 2 weeks old, were found by the same mea-
suring techniques to support a dual oxide layer of just 2 nm
(0.5 nm). It is postulated that long whiskers growing in a
short space of time will have a greater potential to cause
short-circuits than those with a low growth rate to reach a
similar length. The thicker oxides on older whiskers will
require greater breakdown voltages or more mechanical
force to create the conditions for shorting.
(e) Structure and growth directions
White -tin has a tetragonal crystal structure with lattice
constants of a = 5.831 and c = 3.182 at 26 C.
Thinned whiskers subjected to transmission electron
microscopy were found to be relatively transparent to the
electron beam of 200 keV; the ndings indicated the absence
490 7 Whisker Growths

Fig. 7.33 Optical photographs

showing tin whiskers growing
and making shorts between the
bright tin electroplated terminal
pins on relays built into several
satellites. This relay remained in a
bonded components store, but
was from the same lot as those
that caused the failure of at least
four multi-million dollar
communications satellites

of grain boundaries within the whiskers, no dislocations or
stacking faults, no second phases (it should be noted that the
kinked regions were not explored).
X-ray diffraction using the Rotating Crystal Method
showed growth directions to be:
For tin-plated copper on steel: [101] and [100]
For tin plated on brass: [001], [100] and [111]
More than 100 whiskers were subjected to cryogenic
temperatures by immersion into both gaseous and liquid
nitrogen (196 C), for times between 0.5 and 168 h. Tin
pest, or the transformation to grey -tin, was not observed
(see also Sect. 6.15.12).
7.4 Precautions to Avoid General Whisker Growths
7.4 Precautions to Avoid General
Whisker Growths
The case studies that have been described in detail in the
preceding sections report on the effect that surface growths
of whiskers have on modern telecommunications equipment.
These investigations were made by standard laboratory
failure analysis techniques and have identied several
potential sources of whisker-induced failures that can
severely jeopardize the reliability of spacecraft systems.
They can be related to one or more of the following:
1. Corrosion damage during storage and application
Entrapped contamination can lead to corrosion and electro-
chemical migration within electronic devices. Electronic
circuits must remain humidity-free and should be hermeti-
cally sealed in metallic packages to avoid the condensation
of moisture and ingress of contaminants.
The time-related short-circuit failures that are so clearly
shown in Fig. 7.2 were caused by the formation of dendritic
whiskers resulting from an electrochemical process. Such
failures require the presence of three elements:
Reactive chemicals such as halogens (the trace amounts
of chlorine detected in both good and failed integrated
circuit devices indicate that it is difcult to remove all
traces of these chemicals from the microcircuit).
A layer of water to form an electrolyte between neigh-
boring stripes (formed by condensation when the tem-
perature drops below the dew point).
A bias for the electrochemical reaction to take place.
Special precautions have to be taken to avoid this elec-
trochemical migration form of whisker growth. Cleanliness
testing has been proposed, using equipment that can detect
ionic residues. Alternative measures such as lling packages
with dry nitrogen and adding a nal glassication layer (e.g.
sputtered silica) to the circuit are two alternative
possibilities.
2. Unsuitable aspects of design
The tungsten heater coil (Fig. 7.3) developed hot spots that
caused tungsten to sublimate. Released tungsten atoms
passed through the vapour phase (in this case the evacuated
travelling wave tube) and condensed at the whisker growth
sites that were established by the presence of a temperature
gradient and the existence of a suitable nucleation point. The
whisker growth process was accelerated by the inadequate
design of the cathode heater piece-part and led to this device
failure by spark emission and a possible short-circuit
491
between a whisker and the housing. Redesign of the cath-
ode heater introduced a complete back-potting that encased
the heater within small particles of alumina. This restricted
the sublimation of tungsten by limiting the free surface area
of the heater coil, and also enabled heat to be dissipated from
local hot spots by thermal conduction.
3. Human error
Human error can occur when an operator becomes careless
and causes a microcircuit to be electrically overloaded. One
result may be to cause the phenomenon of electromigration
and generate whiskers. Alternatively, the reason for
electromigration-related whisker growth can be wrong
design by the engineer responsible for the microcircuit. The
problem is highlighted by Figs. 7.57.9. Electromigration
occurs because of the migration of metal atoms along a
conductor under the influence of a large current. In most
electronic circuits electromigration manifests itself as mate-
rial is transported along grain boundaries by a diffusion
mechanism. This occurs at high current densities (i.e. in
excess of 2 105 A/cm2) and at relatively low temperatures
(much lower than the melting point of the conductor). As
atoms tend to migrate to positive ends of the conductor,
vacancies will travel in the opposite direction giving rise to
hillocks or whisker growths and voiding where the vacancies
accumulate.
Human error can also be considered to account for the
glass-rich oxide whiskers revealed within the insulation glass
windows of integrated circuit flat-packs. Figure 7.4 clearly
shows that during manufacturing the assembly operator has
misplaced the lead and wall frames of this device (this may
also be caused by a jigging problem or poor jig design). Had
the pre-oxidized parts been separated by a greater volume of
sealing glass it is certain that this volume would not have
become supersaturated with metal oxide; and, on cooling to
room temperature, whiskers would not have formed.
4. Incorrect material selection
It may be arguable that each of the previous mechanisms of
failure could have been prevented if the relevant design
engineer had chosen other materials for these applications.
However, the case histories that refer to the growth of tin
whiskers (Sect. 7.3) are identied as being true cases of
incorrect selection of materials, and it is clearly seen that
pure, as-plated tin will nucleate and grow whiskers in a
spontaneous manner regardless of environment. Figure 7.35
shows a herring bone analysis of all the factors leading to tin
whisker formation. It is a complicated subject as so many
processing scenarios will promote such growths and not all
mitigation steps will be completely successful (e.g. the
492 7 Whisker Growths

Fig. 7.34 Here, a

fourty-eight-year old whisker is
detailed after being bent, ac
showing cracks in the oxide lm.
The XPS analysis incorporated
sputter proling with Ar ions until
a clear metallic tin peak could be
detected. The tin oxide thickness
for this whisker was found to be
18 2 nm based on a knowledge
of the sputter rate and time to
reach bare tin (Dunn and
Mozdzen 2014)

encasement of tin-plated surfaces with conformal coatings
may not be sufcient, as tin whiskers have been seen to
penetrate such coatings (McDowell 1993; Dunn 2012;
Wickham et al. 2014). A more detailed discussion of miti-
gation will be covered in Sect. 7.5.
As a nal remark to the case studies one can take cog-
nizance of whisker growth situations that do not give rise to
such calamitous effects as those afflicting complex electronic
circuits. Enormous quantities of whiskers have also been
seen to grow on household appliances. They are seldom
reported in the literature, although accounts have been given
of whiskers forming on the zinc-plated panels of refrigera-
tors and ovens (Garner et al. 1976), and one French paper
(Bacquias 1982), which makes very interesting reading,
describes tin whiskers that were found growing from the
decorative plated nish on souvenirs stored in shop
7.4 Precautions to Avoid General Whisker Growths 493

Fig. 7.34 (continued)

windows! These defects may be of only a cosmetic nature, whisker growth are becoming understood they must be
but they are also frustrating and cause nancial losses due to passed on to engineers and craftsmen responsible for a very
the return of aesthetically rejectable goods. As the causes of wide range of manufactured goods.
494
Fig. 7.35 An overview of the main factors contributing to tin whiskers formation and some means to mitigate against growth. Courtesy of
CALCE/University of Maryland
7.5 The Creation of Lead-Free Control
Plans
7.5.1 General
Humans have used lead since at least 6500 BC for the
ducting of water and for ornaments and jewelry. It is only in
recent years that the harmful effects of lead have been rec-
ognized and, as lead has been eliminated from paints and
petroleum, there has been a dramatic reduction in
lead-related illnesses. It is considered by many engineers that
the sweeping changes introduced by the European Parlia-
ment and Council in 2006 will have a negative effect on the
reliability of electrical systems being manufactured for both
the commercial and professional markets. The EU directives
concern WEEE (Waste Electrical and Electronic Equip-
ment) and RoHS (the Restriction of the use of certain
Hazardous Substances in electrical and electronic equip-
ment). These directives now specically forbid the tradi-
tional use of lead in the composition of component surfaces
and solder alloys used to assemble electronic circuits.
7 Whisker Growths
The exemptions will be listed below, but it must be rec-
ognized that lead-acid batteries account for more than 80 %
of lead consumption and these are exempt from EU regu-
lations. Huge quantities of lead are used in the construction
industry with few regulations, whereas only 0.5 % of mar-
keted lead is incorporated into electronics. The concern is
that the recycling of electronics by primitive means in third
world countries will harm persons due to the inhalation of
lead, but this has been disputed bearing in mind that lead
boils at the very high temperature of 1750 C. Studies such
as those by the University of Stuttgart, (http://leadfree.ipc.
org/les/RoHS_15.pdf) have indicated the replacement
lead-free alloys for electronic use (i.e. tin-silver-copper) can
be more damaging, to both the environment and humans,
than the lead-containing solder alloys. Also, as discussed in
Sect. 3.6, there is much to gain from the recycling of elec-
tronic waste and the extraction of valuable metallic elements.
Exemptions to the EU directives (RoHS) that ban the use
of lead-containing alloys or glasses in electronics include
certain industries related directly to military, aerospace,
transportation and space.
7.5 The Creation of Lead-Free Control Plans
At the time of writing the following wording is cited:
P7- TA-PROV(2010)0431
A7-0196/2010
Restriction of the use of certain hazardous substances in
electrical and electronic equipment
4. This Directive does not apply to:
(a) equipment which is necessary for the protection of the
essential interests of the security of Member States,
including arms, munitions and war material intended
for specically military purposes;
(b) equipment designed to be sent into space;
(c) equipment which is specically designed and to be
installed as part of another type of equipment that is
excluded or does not fall within the scope of this
Directive, which can full its function only if it is part
of that equipment, and which can be replaced only by
the same specically designed equipment;
(d) large-scale stationary industrial tools;
(e) large-scale xed installations;
(f) means of transport for persons or goods, excluding
electric two-wheel vehicles which are not
type-approved
A guidelines document suggesting the contents for a
company lead free control plan has been published (Dunn
2012) based on the known problems associated with lead-free
technology. Commercial off-the-shelf (COTS) components
are discussed as well as methods for checking if pure tin is
present during goods-inwards inspection of spacecraft parts. It
is proposed that swabs and sticks that contain chemicals which
turn red when more than 10 % lead is present are a cheap initial
control. Other more accurate methods are included, but care
should be taken when using energy dispersive spectroscopy as
false readings have been observed: when only large tin-rich or
lead-rich grains are analysed rather than an area containing
mixed grains, or when thin pure tin layers are penetrated and
readings include elements of the substrate.
Two standards have been established with the objective
of predicting whether newly tin-plated electronic parts will
grow whiskersthey contain quality control type require-
ments and necessitate that samples are subjected to accel-
erated environmental testing. The standards JESD22-A121A
and IEC 60068-2-82 include temperature cycling and
495
elevated humidity testing with high magnication inspec-
tions, before and after exposure, and the recording of
whisker growths in terms of length and density. These
methods are not considered suitable for the acceptance of
space hardware for the following reason: only whiskers
having a length greater than 40 m are considered poten-
tially harmful and shorter whiskers are accepted. The
mechanism for whisker growth is not yet understood and the
impact of very long storage and operational lives for
spacecraft parts cannot be determined from these short-term
tests. As noted in Sect. 7.3.4, some substrates take many
years before whiskers nucleate and growin one instance
an incubation time of 12 years was necessary before whis-
kers emerged and grew to lengths exceeding 1 mm. It is
considered expedient to totally avoid the use of pure tin and,
as suggested by other authorities, alternative nishes such as
nickel-palladium, nickel-gold, NiPdAu and even bare copper
can be suitable for most applications (George and Pecht
2013; Morilla et al. 2015).
7.5.2 Methods for Reprocessing Pure Tin
Terminations
The most effective methods for the avoidance of tin whiskers
are those that completely replace pure tin nishes.
It is essential that the re-processing methods do not
degrade the component or piece-part due to thermal shock,
chemical attack caused by aggressive solder fluxes, or
chemical attack due to stripping and etching solutions. The
following reprocessing methods are suggested:
(a) Method for chemical stripping and replacing tin with
tin-lead by solder dipping
This method must be performed by a specialist company and it
is essential that in depth studies are performed in order to
ensure that no damage has occurred at glass-to-metal seals and
that there is no ingress of chemical liquids into the component,
either by cracking, dissolution of seal material or by ingress
through plastic materials that may be incompatible with the
chemicals. Care must be taken to ensure traceability of the
devices as chemicals can remove marking inks. This method
has been demonstrated to be suitable for a wide range of
components, including those with glass-to-metal seals as pure
tin plating can be removed right up to the seal as seen in
Figs. 7.36 and 7.37 (Turner et al. 2013).
496 7 Whisker Growths

Fig. 7.36 Acceptable workmanship detailing bare copper after chemical stripping and solder coverage to within 1 mm of glass seal, showing
stripped and pre-tinned view and SEM image of sectioned device

Fig. 7.37 Example of part markings on large TO can transistor, a as-received, and after two stripping parameters where markings are either
b slightly, or c totally removed by chemical stripping d partial immersion in the chemical stripper to the base did not affect the markings
7.5 The Creation of Lead-Free Control Plans
The leads can then be tinned by dipping using standard
solder-pot or -wave methods. The cleaning of any adhering
aggressive solder flux is an important part of this process.
Components that have been successfully re-processed by
this method have been listed. Similarly, a list exists showing
components that were damaged by this method. Barrier
layers or substrate lead and termination metals may be
exposed adjacent to seals, but these new surfaces are
acceptable for space programme use.
(b) Method for solder dipping and the reapplication of
solder onto tin-nished surfaces
Solder dipping of tin-plated surfaces, when performed cor-
rectly with tin-lead alloys, will eliminate the formation of
whiskers as all of the pure tin will have been replaced by
Fig. 7.38 Components were supplied with leads having a pure tin
nish; during storage they developed whiskers; these gures show the
leads after tin-lead reprocessing (pre-tinned by dipping to
ECSS-Q-ST-70-08C (2009)); micro-section and EDX analyses con-
rmed the presence of pure tin (whisker) coated with tin-lead; pure tin
melts at 232 C and eutectic SnPb at 183 C; due to the heat-sinking
497
tin-lead. When possible this is the most effective and
cost-efcient method of re-processing. It may be problematic
when solder wetting is difcult and time/temperature of
immersion is critical to avoid the presence of residual whis-
kers as shown in Fig. 7.38. It is likely that this method will not
be permitted on components having a glass-to-metal seal
because these seals might be damaged by thermal shocks if the
component lead is submerged completely up to the seal.
However, a form of verication testing has approved this
method when it was shown, as for Method a), that no com-
ponent damaged ensued. This process must be carefully
controlled and processing parameters need to be documented,
these will include: solder pot temperature, composition of the
solder pot alloy, pre-heating rates, immersion depth, cooling
rates, electrostatic discharge, nal cleaning to remove residual
solder flux. Robotic solder dipping is performed at specialised
effect of the package, and the presence of a thin (20 nm) tenacious
oxide on the pure tin whiskers, there has been insufcient heat to
completely dissolve the tin whiskers into the solder; oxide thicknesses
on whiskers have been measured by Mozdzen (2012); some theories
propose that if any pure tin plating exists below the dipped solder layer,
this will reactivate whisker growth
498
companies for leaded and leadless component types. These
same companies (Walmsley 2015 and Tyler 2015) are also
capable of reprocessing a very wide range of area grid arrays,
replacing lead free interconnections with both tin-lead balls
and columns. Tables of suitable and unsuitable component
types have been compiled by the companies who perform the
reprocessing.
(c) Over-plating pure Tin nishes with pure Lead
This method has been used for small chip components. After
plating lead onto the pure tin surfaces the device is heated to
promote complete inter-diffusion. The author is not aware
that this process has been applied by European companies.
7.5.3 Mitigation Approaches
(a) Component terminations
Tin variations: hot dipped tin; large grain size or actually a
matte tin; thick tin >15 microns; immersion tin
Underplating with: nickel or silver
Although all tin nishes have some risk of whiskers,
some nishes have greater risk than others. Selection of a
matte tin with a nickel underplate (>2 m) or a Pb-free tin
alloy with nickel underplate (>2 m) are considered lower
risk nishes than bright tin or tin nishes over copper. Other
nishes include SAC nish or dip, SnAg nish; SnBi nish
Heat treatment: above fusing temperature (best); below
fusing temperature
Fig. 7.39 (Left) A growing tin whisker is attempting to break through
a silicone-type of conformal coating; and (right) a tin whisker that is
growing from an adjacent surface, is attempting to penetrate a
7 Whisker Growths
SnPb immersion: Chemical stripping followed by partial
SnPb dip; SnPb dip with standoff (i.e. gap between seal and
solder meniscus)
(b) At the PCB assembly level
Physical Barrier (e.g. Mylar foil, space-approved solder
sleeve)
Pottings: Hard potting; soft potting or encapsulation
Soldering variations: Soldered with SnPb solder with
qualication /verication of coverage (with or without sam-
pling); soldered with SnPb without verication on bottom
termination or small leadless chip parts; SnPb with 13 % Pb;
soldered with SnPb without verication on large leadless chip
Conformal Coatings: Epoxy; polyurethane; silicone
(possibly not very effective as seen in Fig. 7.39); ALD
(atomic layer deposition) and vapour deposition (such as
Parylene, polyparaxylylene)
Note: The use of conformal coating does not eliminate the
potential for whisker failure entirely. For most coatings, tin
whiskers have been shown to eventually grow through thin
coating regions and potentially short any other whiskers that
have grown similarly (see Fig. 7.39). In addition, for some
piece part types (e.g. through-hole mounting leaded piece
parts, area array packages), it is difcult to entirely encap-
sulate the individual interconnects so there remains a direct
path for shorting due to whisker growth.
conformal coating. This can result in an electrical short circuit.
(Electron micrographs courtesy of C. Hunt, the National Physics
Laboratory)
7.5 The Creation of Lead-Free Control Plans 499

Fig. 7.40 An example method of assessing the risk of whisker bridging based on a knowledge of whisker growth rate and storage (life) times
(McCormack and Meschter 2009, Ashworth and Dunn 2015)

(c) An empirical method of assessing mitigation against
risk
Many studies have been performed and they provide limited
quantication of when and how failure conditions might
occur as a result of whisker growth. The basic mechanism
for whisker nucleation and growth is to date not fully
understood. Experiments do not always provide for
reproducible results even when material variants are kept
under control. However, it is important that engineers
working in the space, and other sectors requiring high reli-
ability products, are able to make an informed judgement,
for instance during the course of a nonconformance review.
Table 7.2 can be consulted as a guide to specic risks. An
attempt has been made to put a numerical value to each risk
factor; these risks can be cumulative.

Table 7.2 Plated coatings and product designs susceptibility to whisker induced failure (i.e. conditions likely to initiate whisker growth
mechanisms and their relative propensity to cause electrical shorts and metal vapour arc failures due to whisker bridging)
Risk of failure Specic cause likely to produce failure (+ve)
+8 Bright electroplated pure tin (due to organic brighteners) or matt tin plated onto brass substrate
+5 Bright tin plated on any substrate except nickel or silver
+4 Bright tin plated on nickel or silver (barrier layer or substrate)
+3 Matt tin plated on any substrate except brass
+3 Residual stresses in electroplated tin (no annealing)
+3 Compressive stresses resulting from torqueing a nut or screw (hold-down clamp, etc.) and severe mechanical
deformation as in Fig. 7.30
+3 Coefcient of expansion mismatch between tin and substrate of more than 10 ppm
+1 Bending or stretching of plating, or scratches due to poor handling
+1 Temperature cycling where temperature differences are above 20 C
0 Humidity, vibration or shock, operation under vacuum
See Fig. 7.40 (+4 to Decrease of component package lead separation distance such as ne pitch devices, or distance between termination
+0) and tracks (printed circuit board, electrically conductive housing, etc.)
See Fig. 7.40 (+4 to Life-time of equipment increases to 15 or more years
+0)
Note present-day tin whisker testing methods are unable to predict whether a tin or tin alloy nish will grow whiskers. The Risk of Failure values
are based on the authors published and unpublished work, activities as member of the European Pb-free Electronics Risk Management technical
committee, and participation in the Tin Whisker Symposia (20072015). The values are the opinion of the author
500 7 Whisker Growths

Table 7.3 Mitigators likely to avoid tin whisker bridging; these specic methods, singularly, or in combination, can be selected to provide the
appropriate level of mitigation in order to manage the Risk Factors proposed in Table 7.2
Mitigation factor Specic method of mitigation (ve) against failure
10 (a) Hot solder dip with tin-lead to component body (i.e. all of the lead surfaces have to be in contact with the solder
bath), or complete surface of xture
(b) Chemical strip and solder dip with tin-lead (e.g. Figs. 7.36 and 7.37)
10 Physical barrier such as a plastic shrink sleeve (space qualied) which, when shrunk, will become totally sealed so as to
avoid the release of loose whiskers
8 (a) Hot dip with tin-lead to approved distance from component seal
(b) Fuse electroplated pure tin in hot oil or reducing atmosphere
6 Epoxy potting or complete conformal coating (corners to >5 m)may affect component reliability under thermal
cycling
6 Atomic layer deposition coating and certain vapour deposition coatings
4 Polyurethane conformal coating (corners to >5 m)may affect component reliability under thermal cycling
3 Silicone conformal coatingmay affect component reliability under thermal cycling (see also Fig. 7.39)
3 Annealing electroplated tin at 150200 C for 1 h
See Fig. 7.40 (4 Increase component package lead separation distance, or distance between termination and tracks (printed circuit board,
to 0) electrically conductive housing, etc.)
3 to 0 Generally, thick (>15 m) electroplated tin is thought to be less susceptible to tin whisker growth than thin
electroplating, or immersion tin. Over-plating tin with a lead surface nish is unlikely to prevent whisker growth unless
this bi-layer is fused
Notes see those of Table 7.2

Similarly, a limited amount of data is available to identify
and quantify the effectiveness of mitigation methods for
either preventing whisker growth or limiting the risk of
whiskers causing failure; here Table 7.3 can be consulted,
and again methods and associated values can be cumulative.
Both tables have been updated since the original tables
were compiled in the authors previous publication (Dunn
2012) to be in line with more recent study results.
A comparison between the sum of the positive values
assigned to the cumulative risks and the sum of the
negative values assigned to any mitigation methods used
(or mitigating design features) is intended to indicate the
acceptability of a given situation.
 Assessment of Post-flight Materials
8.1 General
8.1.1 Hardware Return from Space
The low Earth orbit (LEO) environment is dened as that
region of space between 200 and 1000 km (124 and
621 miles) above the Earth. It was reasonably well charac-
terized in the 1950s, during the prelude to the race into
space, by means of sounding rockets equipped with moni-
toring devices for recording air pressure, temperature, and
gaseous composition. Many of the data obtained from those
early space flights are suitably accurate for todays calcula-
tions and have been discussed in Chap. 2 and compiled into
Table 2.3. In just one decade, the 1960s, human beings
began to extend their physical boundaries by initially ven-
turing into LEO and then progressing to explore the
mountains and valleys of the Moon. The rst human to orbit
the Earth was cosmonaut Yuri Gagarin in the Vostok 1
spacecraft on the 12th April 1961. Since then, manned
spaceflight included the Mercury (Atlas ELV) missions, the
Gemini (Titan ELV) missions, and the Apollo (Saturn ELV)
missions that gave 12 human beings the opportunity to walk
on an extraterrestrial surface, with the rst manned landed on
the Moon in 1969 with Apollo 11, and the nal
Moon-landing in 1972 with Apollo 17. The 1970s saw the
launch of several Salyut space stations and three Skylabs,
followed in the 1980s and 90s with the Mir station, the
Kvant module, and the many Space Shuttle launches (for the
placement of satellites into orbit, ejection of probes into
space, retrieval of spacecraft, and utilization of Spacelab
which remained attached to the Shuttles cargo bay). At the
time of writing, it is the International Space Station
(ISS) that continues to be the most newsworthy space
activity. This is the biggest construction to fly in space and
can be readily seen from the ground as it completes 16 orbits
per day. Huge solar panels, totaling approximately one acre
are attached to the central truss of ISSthe dimensions are
equivalent to the length of a football eld, the habitable
volume is as large as a six-bedroom house and the total mass
Springer International Publishing Switzerland 2016
B.D. Dunn, Materials and Processes, Springer Praxis Books,
DOI 10.1007/978-3-319-23362-8_8
8
is just over 450 metric tons (tonnes). Awe-inspiring, as ISS
was built by astronauts and cosmonauts, 216 have lived
there, bringing each part of the construction from Earth by
Shuttles and various other cargo-vehicles (2015 data).
The space missions referred to above have all involved
the launch, exposure to space (in particular the LEO envi-
ronment), and the nal return of some form of space hard-
ware to Earth. Once the mission tasks had been completed
there was little reason to spend extra effort into performing
post-space flight analyses of the materials used to construct
the spacecraft. Some post-flight observations have been
documented, such as those related to impacts by microme-
teorites during the mission, the degradation of thermal
control surfaces due to discoloration from exposure to
ultraviolet radiation and, of most importance, the effectivity
of the re-entry protective shield. However, the external
surfaces of those spacecraft were constructed from materials
that were little affected by the space environment, such as
anodized aluminium alloys, titanium alloys, passivated
stainless steels, and fused silica glass. The re-entry shields
were either of the ablative type that burnt and shed material
during their hypersonic velocity passage into the oxidizing
atmosphere or, more recently, of the lightweight
silicon-based ceramic tile-types which have a good resis-
tance to high-temperature oxidation, air turbulence, and
shock during descent.
8.1.2 Raw Materials from the Moon
The number of documented post-flight examinations of
spacecraft materials are probably now equal to the several
hundred papers in which the results of Moon-rock analyses
were published since 1967. Although the scientic yields
from the observations of the astronauts on early Apollo
missions were of great importance, they were later over-
shadowed by the investigations of land-based scientists who
studied the lunar samples. These studies included a deter-
mination of the abundances of all the elements, their
501
502
Fig. 8.1 Micrograph of Moon
rock returned by US Apollo 11
astronauts, July 1969
isotopes, and the nature of their coexistence in the various
phases making up the lunar material. The lunar samples
brought back by the US (manned Apollo) and Soviet (Luna
automated soil collection and return to Earth) space pro-
grammes revealed that the Moons surface is rich in non-
volatile, low atomic mass elements. Lunar rock and soil
samples (termed regolith) were distributed to laboratories
worldwide for analysis. One particularly interesting rock was
prepared by standard metallography and is shown in
Fig. 8.1. This micrograph is typical of the Apollo 11 regolith
in that it is basaltic and contains a large amount of ilmenite.
The rock also contains phases of metallic iron and a small
amount of metallic copper. Iron-nickel alloy, native iron and
native copper have also been identied during quantitative
optical and electron-probe studies of lunar samples at the
Institute of Geological Sciences, London (Simpson and
Bowie 1970). The location of metallic iron has been stated to
bear a relationship to the lunar magnetic properties and many
possible sources of this metal have been postulated (Adler
1988). It would appear that the existence of native copper,
present as a small 4 m diameter sphere in Fig. 8.1, is
uncommon because chemical analyses of core samples,
collected at any of the other eight Apollo sampling sites,
showed that this metal is generally only present as a trace
element in the order of 10 ppm. Oxygen is particularly
abundant in the form of oxides of silicon, titanium, iron, and
magnesium at the lunar mare site visited by Apollo 11. By
8 Assessment of Post-flight Materials
comparison, the rocks retrieved from the lunar highlands by
later Apollo spacecraft were relatively enriched in oxides of
aluminium and calcium (with low Mg and Fe contents).
The use of the Moon and lunar materials has often been
suggested by scientists as the stepping-stone into an indus-
trialization of space. Certainly there exists a wealth of
minerals, a strategic location, and a reduced gravitational
force. It should be possible to manufacture a range of
structural materials on the lunar surface; oxygen in the form
of oxide is present, but three of the most vital elements were
thought missing: carbon, nitrogen, and hydrogen, and
without these there can be no indigenous life-support sys-
tems. In recent years scientists have suggested that water ice
and other frozen gasses (or volatiles) may be trapped on the
perpetually shaded polar regions of the Moon. If long-term
support is to be given to human activities on the Moon or
Mars it seems essential that a base will have to be built on
the Moons surface. Both NASA (Mueller 2014, 2015) and
ESA (Ceccanti 2010) have considered the requirements for a
lunar base to house people and offer protection from mete-
orites, solar- and gamma-radiation and high temperature
fluctuationsseveral attempts on Earth to construct dwell-
ings from lunar regolith simulant have been successful (see
also Sect. 8.2.2). The ultimate success of a permanent lunar
base is expected to depend on how man uses the in situ
resources. If it is possible to manufacture propellants on the
Moon it will relieve astronauts of the need to transport
8.1 General
return propellants. The lunar surface appears to contain 45
weight percent oxygen, most in the form of silicates and
other mixed metal oxides. With the use of solar, nuclear
power and focused mirrors, it has been proposed (Hepp
1999) that regolith can be heated and modied to release
oxygen and precipitate metals and alloys of aluminium,
silicon, iron and titanium. These metallic powders would
then be burnt in a combustion chamber and the simultaneous
release of hot gasses gas through a nozzle would provide a
form of propulsion!
8.1.3 Recent Investigations Using Retrieved
Materials
The IIS was gradually constructed over a period of 12 years.
This started in 1998 when the rst module, Zarya, built by
Russia, was launched in 1998. This permanent space station
in LEO has been continuously occupied since 2000. The
present architecture is seen in Fig. 3.1b. Plans are to prepare
the station for commercial use and two international docking
adaptors will be added to the structure and utilize the station
until at least 2020. Alongside the operation of ISS has been a
continual series of materials tests, such as the MISSE
experiments, where samples of a wide range of
material-types and components are exposed to the harsh
environment of space for long periods of time. The data that
has been generated from tests performed on materials that
were retrieved from long-term space satellite experiments,
from the return of functioning space hardware, and from
relatively short-term Space Shuttle flights is the subject of
many space symposia. Some data have been generated from
ground tests in which attempts are made to replicate the LEO
environment in laboratory chambers. The majority of the
materials exposure data has originated from space missions
such as Salyut-6 (5-year exposures), Salyut-7 (9-year
exposures), Mir station (experiments of 3-year duration),
the Long Duration Exposure Facility (LDEF) launched and
retrieved by Space Shuttles (STS flights) after nearly six
years in LEO, and the European Retrievable Carrier (EUR-
ECA), similarly launched and collected after nearly one year
in orbit. Short-term materials exposures were made on
STS-5, STS-8, and STS-46 (for instance, the 58 h Evaluation
of Oxygen Interactions with Materials III (EOIM-III) flight
experiments). Retrieval of the Solar Max spacecraft provided
some data, as did the post-flight investigations performed on
the Hubble Space Telescope solar array that was retrieved
after about four years in LEO (during the rst HST Servicing
Mission in December 1993; the second service mission took
place over a 10-day period in 1997). The most recent test
exposures include the European MEDET rig and the previ-
ously stated MISSE project; all are currently being examined
in a 1500 sample test programme.
503
The purpose of the following sections to this chapter are
to present some of the ndings from recent experiments and
observations of returned hardware. As well as recording
some predictable modes of degradation, data has also iden-
tied unusual phenomena likely to either enhance or reduce
the durability of spacecraft mechanisms, structures, and their
materials. Some end-of-life properties have been established
and these are of great importance to design engineers. Only
those references considered to be of major importance will
be cited by the author. The remaining information either
originates from his own work, or has been recorded at early
specialist meetings on space materials (Levine 1993, Flury
1993, Anon 1994, Guyenne 1994, Gerlach 1995). More
recently, some data has been published concerning the
MEDET results (Tighe 2010). The following sections will
also describe property changes to materials brought about by
their exposure to LEO environments. The main effects are
due to each materials exposure to vacuum, radiation, tem-
perature cycling, micrometeoroids and space debris, atomic
oxygen, the re-entry environment, and operation within a
manned volume. The space environment effects are illus-
trated by the block diagrams in each section. There can be a
strong synergistic influence between these effects as inferred
by the diagrams (Dauphin et al. 1991). This information will
be of interest to those engineers responsible for scrutinizing
the materials selection process for future missions, some of
which are destined for relatively low orbiting altitudes
(*350 km) in high densities of atomic oxygen.
8.2 Space Environmental Effects
from Vacuum and Radiation
8.2.1 Organic Materials and Lubricants
Screening tests are continually performed during the selec-
tion of organic materials for space applications. Approxi-
mately 100,000 test results are available from NASA and
ESA sources based on the test method specied in ASTM
E-595. The provision is that materials, when tested in vac-
uum at 125 C for 24 h, shall not have a recovered mass loss
(RML) greater than 1 %, or a collected volatile condensed
material (CVCM) content of more than 0.1 %. This, and
other test methods for organic materials, have been described
in Sect. 2.5. Some examples were given for lubricants in
Table 5.2. If these requirements are neglected there is a
danger that spacecraft will become contaminated by the
outgassing species (the effect of outgassing/condensation
can be seen in Fig. 5.12). Vacuum exists everywhere in
space except in pressurized modules. Vacuum may change
from 1014 torr in free space to 104 or 103 torr in poorly
vented spacecraft subsystems. The chain of dangers that
result from outgassing products is depicted in Fig. 8.2; gas
504
Fig. 8.2 Space environment
effects: a vacuum; b radiation
clouds can produce corona and arcing, or may condense onto
surfaces and so modify them to produce thermal and elec-
trical problems.
A major nding from the recent post-flight materials
experiments concerns the use of silicone products that were
previously considered to have been suitably screened by the
outgassing test. Another, concerns the 24 h vacuum bake at
125 C that has become a standard practice in history for
cleaning polished mirrors and other hardware, so benignly
8 Assessment of Post-flight Materials
removing molecular contamination which would otherwise
have migrated to contaminate sensitive spacecraft surfaces
when in orbit. Cleaned hardware is best maintained under
vacuum or, at least, sealed in appropriate bagging material,
preferably under an ultrapure gas such as nitrogen. A com-
parison of identically produced samples that were either
pre-cleaned by the standard practice, or exposed to the
vacuum of space has been performed. The results found that
the ground-based material conditioning process 125 C
8.2 Space Environmental Effects from Vacuum and Radiation
Fig. 8.3 Thermal cover showing
fringe pattern of condensed
contaminant on exposed Kapton
tape (returned from Hubble Space
Telescope)
vacuum bake for 24 h is equivalent to holding an article in a
space environment at an ambient temperature of 20 C and a
pressure of 106 torr for a period of one year. There are two
provisosthat this practice does not hold true for either
silicone products or Teflon FEP materials. Recent tests with
approved silicone products caused outgassing and contami-
nation in LEO. Even very thin layers of contaminant (about
400 ), together with the presence of ultraviolet (UV) radi-
ation and atomic oxygen, caused signicant degradation of
thermal control surfaces (Babel et al. 1994). In a similar
manner, Teflon FEP can degrade with the formation of a
volatile product that is mainly perfluorocarbon, but can
include chemically active species containing fluorine, and
these have been seen to chemically etch the surfaces of
polished metals and glass (van Eesbeek et al. 1994). Teflon
FEP originating from samples mounted on LDEF for
5.8 years, was compared to samples of the same materials
exposed to LEO on the Hubble Space Telescope (HST) for
both 3.6 and 8.25 years (periods for the service missions).
All AO (atomic oxygen) exposed samples became
mechanically embrittled and had lower strength properties
when compared to reference ground-stored sheet (Fink
2006). The HST solar array bellows and the multilayer
insulation (MLI) items were severely damaged and cracked,
being worse for the longer exposure. Samples taken from
LDEF material exposed to the RAM direction (AO:
8.17 1021 atoms/cm2) became more damaged than samples
in the WAKE direction (AO: 9.32 104 atoms/cm2). In
general Teflon FEP is not used on the external surfaces of
505
spacecraft with the exception of its application as an insu-
lation material for wires and cables.
Further inspections of the returned HST solar array
revealed several surfaces to have been contaminated by the
condensation of outgassing products. Figure 8.3 is included,
as it highlights a condensation layer having a variable thick-
ness and is clearer to photograph than surfaces that have been
uniformly discoloured. Silicone contamination has been
identied on the HST samples, by analysis of collected
spectra, to be the main contaminant along with dicyandiamide,
which is an outgassing produce from CFRP materials. All
items manufactured from Delrin material, namely the primary
deployment cable and the boom end cotton reels, changed
from a white colour to yellow-brown on their sun-pointing
sides. All surfaces covered with the atomic oxygen protective
coatings (CV1144-1, DC93500, and RTV S691) became
darkened in colour from straw to dark brownthe more
exposure to the Suns radiation the darker the colour (see
Fig. 8.4). They are expected to cause an increase in tempera-
ture at these locations.
None of the traditional lubricants based on
vacuum-distilled silicone oils and greases were seen to
degrade during exposure to service life in space. Some were
noted to have crept from bearing races onto adjoining
surfaces. Rotating surfaces that had been lubricated with thin
lead coatings survived their designed lives.
Concern was raised when three critical pump motors
failed in close succession following ve months of suc-
cessful operation in LEO. On returning to Earth, the motors
506
Fig. 8.4 Part of the HST solar array showing the ATOX protection silicone coatings to have become darkened after exposure to the Suns
radiation. One micrometeoroid impact is also visible (arrowed) (see also Fig. 4.63)
were stripped down to reveal their brushes, commutators,
and armatures. Large quantities of brush debris were noted
associated with the full wear-out of the graphite brush
lengths. The small stubs of brush material that remained
were analysed and found to contain only carbon (graphite)
and copper with no additional elements. The absence of any
so-called adjuvant materials within the microstructure of
these flight brushes were extremely important. It is known
that adjuvants such as molybdenum disulphide (MoS2) or
barium fluoride are essential components for brushes need-
ing to operate under vacuum. Investigations found that these
motors had been qualication tested prior to launch in lab-
oratory air. This air contained moisture, an essential ingre-
dient for graphite brushes that enables them to run for
extensive periods with very little friction and wear. The
flight brushes had absorbed sufcient amounts of moisture
so that they behaved correctly when operated for a short
period in dry nitrogen, and correctly when launched into
8 Assessment of Post-flight Materials
space. However, once the water content of the brushes had
diminished sufciently by outgassing under vacuum, the
friction and wear rates would be markedly increased, so
producing these early pump motor failures. Some alternative
brush materials have been described in Sect. 4.8.2, but it is
essential to perform any qualication tests in vacuum rather
than dry nitrogen.
More recently with respect to tribology, tungsten carbide
alloyed with MoSx thin lms (deposited by magnetron
sputtering) has been prepared and tested under vacuum and
LEO. Ball-on-disc samples were evaluated on the ESA Tri-
bolab (installed in the European Technology Exposure
Facility xed to the Columbus module of ISS). These tests
(Brizuela 2009) have indicated WC + MoSx to be a novel,
highly suitable material for operation in LEO and comparable
to conventional MoS2 lms. Very low extraterrestrial friction
levels of 0.04 were reached during more than 1 million wear
cycles on the ball-on-disc set-up and loading at 0.75 GPa.
8.2 Space Environmental Effects from Vacuum and Radiation
Table 8.1 Components of the natural space radiation environment
Solar-particle events
Occur sporadically; not predictable
Energetic protons and alpha particles
Solar-flare events may last for four or more days
Dose very dependent on orbital altitude and amount of spacecraft
shielding material
Presently under study by the SOHO spacecraft
Galactic cosmic rays
Typically 85 % protons, 13 % alpha particles, 2 % heavier nuclei
Integrated yearly fluence:
1 108 protons/cm2 (approximately)
Integrated yearly radiation dose:
410 rads (approximately)
Geomagnetically trapped radiation
Primary electrons and protons
Radiation dose depends on orbital altitude
Manned flights below 300 km altitude avoid Van Allen belts
For solar particle radiation see Table 2.4
8.2.2 Radiation Effects
The term radiation (Table 8.1) covers a wide range of
emissions of energy from the Sun in the form of electro-
magnetic waves (the range passes from X-rays to UV and
visible light, to radiowaves) and particles which are trapped
in the Van Allen belts (protons and electrons) that surround
parts of the Earth. A second region that traps highly charged
particles band is known as the South Atlantic Anomaly, its
location is seen on Fig. 8.5, and this presents a signicant
hazard to spacecraft in low altitude, low inclination orbits
(Mikaelian 2001). Other particles originate from deep space,
such as cosmic rays, and reach the Earth from all directions,
with high energy and great penetrative power. Radiation at
the levels existing in space (generally about 107 rads per year)
do not modify the bulk properties of metals. Even the largest
doses of radiation that occur in the Van Allen belt have no
effects on metals. It is only the combination of contamination
layers together with radiation (see Fig. 8.2b) that can severely
change the surface properties of metals and this may lead to
an increase in solar absorptance and a reduced hemispherical
emittance. Materials such as coatings and paints are pro-
gressively damaged by solar radiation becoming embrittled
by both particle and UV radiation (Dauphin 1993). White
paints were modied to have more absorptive properties
507
whilst black paints tended to bleach. The white paints with an
inorganic base and a silicate binder were found to be more
stable than any of the organic-based paints. Spacecraft
adhesives, thermosetting resins, and lubricants are unaffected
by the low dosage of radiation from space, and these are
usually well protected by metallic foils or are present within
electronic black boxes. Some inorganic glasses are quickly
damaged by radiation, and even as little as 103 rad can cause a
loss of transparency to some wavelength ranges (Dauphin
1993). Plastic glasses and optically transparent silicone
coatings are also damaged by a general yellowing during
exposure to solar radiation as seen in Fig. 8.4.
Microelectronic semiconductors are also damaged by
solar particle radiation. These are usually split into two
categories: cumulative effects due to prolonged exposure to
radiation, and single-event effects caused by transient high
levels of cosmic or high energy proton exposure that can
lead to worst-case damage to spacecraft electronics.
For instance, the thin gate oxide in metal-oxide-
semiconductor devices (e.g. MOSFETs) is degraded to
cause leakage currents. As the line width of conductors on
such devices becomes smaller, so the sensitivity of the
various semiconductor SiSiO2 interfaces to radiation
becomes greater. The ionizing radiation creates electron/hole
pairs and charging of the oxide to produce leakage current
perturbations. Irradiation will also cause deep-level defects
in the bulk silicon lattice. This can cause gain degradation of
an npn or pnp transistor, or increased leakage current in a
diode. Experience has shown that LEO spacecraft have not
often suffered from electronic component damage due to
radiation. However, damage does occur when space vehicles
travel through the heart of radiation belts as would be the
case for spacecraft in low Earth polar orbit, or prolonged
periods in geostationary orbit. Here, radiation hardening is
required by means of altering the processing steps during
device manufacture or providing countermeasures such as
physical shields or foils around the packaged device.
Detailed handbooks have been compiled by Homes-Siedle
and Adams (1993) and Mikaelian (2001) to cover the effects
of radiation environments on spacecraft electronic equip-
ment, their components, and materials. It is noted that in
recent years researchers have been searching for novels ways
to avoid radiation damage, either integrating functions into a
single chip or ASIC for better performance, or by enhanced
radiation shielding with tantalum and tungsten spot shields.
These metallic elements are about six times the density of
aluminium and they allow thinner shields to be made, thus
only incurring a small weight penalty (Maurer 2008).
508
Fig. 8.5 Tristian da Cunha is located in the middle of the South Atlantic Anomaly, which is also where the Earths magnetic eld shows an
anomalous dip. Courtesy of the Danish National Space Centre
8.2.3 Effects of Vacuum on Metals
As already alluded to in the previous sections, there have been
no instances of degradation to metallic materials due to the
combined effects of vacuum and radiation even when returned
to Earth from the longest exposures to LEO. This is because
the selection process avoided the use of metals known to
produce problems such as sublimation (e.g. for cadmium and
zinc), whisker growths (tin and zinc), or loss of hermeticity for
pressurized items, such as battery cells, hydrazine fuel tanks,
and the like. The problem with silver was related to atomic
oxygen, which will be discussed in Sect. 8.5.
One problem that did come to light, which resulted from a
poor electrical design, was the major failure of one of the
HST solar array drive electronic (SADE) systems after three
years of satisfactory operation. At the moment the solar array
position data was lost, the drive electronics was automati-
cally switched off. A redundant unit was powered up soon
after this event for the subsequent eight-month period up to
the date of the HST service mission. As a consequence from
8 Assessment of Post-flight Materials
on-ground studies of the problem it was concluded that the
failure was due to overheating of one of the SADEs
assembled printed circuit boards (PCBs). A replacement
model was built in which two power transistor devices were
physically removed from the circuit and mounted via
heat-sink caps onto the internal metallic structure of the
SADE (Haines et al. 1995). This modication resulted in a
device junction-to-structure thermal resistance of 30 C per
watt and ensured that even under the worst case conditions,
the maximum transistor junction temperature would be less
than 85 C.
The SADE exchange activities on the rst HST servicing
mission were performed at night and under poor lighting
conditions by astronauts Story Musgrave and Jeff Hoffman.
Light from Musgraves helmet light was used to illuminate
the SADE system and during one of the most difcult EVA
(extra-vehicular activity) tasks of the mission, the defective
SADE was removed and replaced (Haines et al. 1995).
Electrical tests and visual inspections were performed on
the defective PCB after it had been returned to Earth.
8.2 Space Environmental Effects from Vacuum and Radiation
Following removal of the equipments cover, a few minute
spheres of solder alloy were seen attached to the inside
surface of the lid. Subsequent inspections revealed much
discoloration of the Solithane conformal coating in the
vicinity of the two previously suspected power transistors.
Electrical tests were made. The electronic unit had a below
nominal power consumption, solar array position data were
not available, and no smoke was seen to emanate from the
suspected fault area. Further disassembly was performed and
photographs made of the charred and overheated power
transistors (see Fig. 8.6a). Inspection of the diodes adjacent
to the charred site showed that they had also overheated as
their solder joints had remelted and both the conformal
coating and the epoxy top-coat of the circuit board had been
completely evaporated away (Figs. 8.6, 8.7 and 8.8) to
expose a region of glass-bre weave (Adams 1994).
The cause of the SADE failure is thought to have resulted
from the inadequate thermal management of the defective
PCB assembly. Testing on the ground, in air, had not caused
any pre-flight failures because the area surrounding the two
power transistors could be kept below a critical temperature
by means of heat lost by thermal conduction through the board
and heat lost by thermal convection through the laboratory air.
Under spacecraft conditions, the operating environment of
vacuum excluded convection as a means of cooling and, after
a period of high-temperature operation, the transistor solder
joints melted (above 183 C), so producing the nal failure.
By modifying the circuit such that both power transistors were
physically attached to the metallic frame of the box, additional
conductive heat loss was made possible and, to date, the
present HST SADE unit is operating perfectly.
8.3 Temperature Cycling
The effects of temperature on spacecraft materials are
depicted in Fig. 8.9a. The main input is from the Sun,
sometimes reflected by a planet, and on output (cooling) by
emission to the immense sink of deep black space. Tem-
perature excursions are also generated by the operational
mode of the spacecraft, from high-power electrical systems,
and by the release of energy (exothermic reactions) during
the consumption of liquid fuels. Local thermal cycling is
caused by the expenditure of electrical power as devices
such as travelling wave tubes are operated and then switched
off (power cycling). The main overall thermal cycling of a
spacecraft is experienced as it repeatedly passes from sun-
light into eclipse during the course of an orbit. The solar
array of the HST is subjected to 6000 thermal cycles per year
between the temperatures 113 and +95 C.
509
Little space will be devoted to temperature effects and
thermal cycling in this section, as much has been written in
this book to cover the topics of sublimation, thermal fatigue,
and the importance of matching adjoining materials so that
they will have similar coefcients of expansion. The very
gradual thermal fatigue degradation of the HST solar array
cell-to-interconnector weldments (only made visible by
SLAM investigations) were described in Sect. 4.14. Distor-
tion of the HST solar array due to differential heating and
expansion rates were described in Sect. 5.17, but apart from
this major defect, the remaining materials compatibility on
the array seem to be good and the complete spacecraft will
be flightworthy until its successor, the James Webb Tele-
scope will be launched and become operational in 2017.
One of the most severely thermal-cycled solder joints to
have been returned from space is shown in Figs. 8.10. This
joint was one of several that existed on the HST arrays
so-called tag board. All survived 21,000 thermal cycles. The
solder llet has seen some stress due to the large thermal
expansion coefcient of the FR4-type PCB in its z-direction
(see also Chap. 6). Thermal fatigue is induced by 2.1 m
mismatch between the boards glass-bre weave/epoxy
matrix and the brass terminal pin. This has exceeded the
yield strength of the solder alloy during thermal cycling,
causing the observed orange peel effect on the surface of
these solder llets (Fig. 8.10a), but when microsectioned, it
is clear that the integrity of the interconnections has
remained intact (Fig. 8.10b, c).
8.4 Micrometeoroids and Debris
8.4.1 General
Micrometeoroids and cosmic dust are natural, hypervelocity
particles that originate from comets, collisions between
asteroids, and interstellar material. All sizes of these particles
are of importance to spacecraft, but those most likely to
cause damage, owing to their statistical distribution, range in
diameter from 1 m to 5 mm. Before the space age began,
this was the only particle population surrounding the Earth
and passing through her atmosphere. Since the 1950s, debris
from mans activities in space has accumulated across the
skies. Space debris includes satellites and rocket upper
stages that have completed their mission, together with
thousands of smaller particles generated by explosions in
space. Since the launch of Sputnik in 1957, more than 4600
launches have placed some 6000 satellites into orbit. Nearly
2200 remain in orbit, of which 450 are still functional.
This means 80 % of space objects are uncontrolled debris.
510 8 Assessment of Post-flight Materials

Fig. 8.6 Visual appearance of

post-flight Hubble Space
Telescopes SADE failed
electronic circuit board. a View
showing the two transistors, in the
background, detailing the
discoloration of the Solithane
coating on the component bodies.
b View showing a joint on a
diode, detailing the melting of the
coating causing it to draw away
from the joint. Other areas of this
pcb assembly are seen in Figs. 8.7
and 8.8

Most of the debris exists in low orbits and is removed by the atmospheric drag and consequently may remain in orbit for
natural cleaning process of being slowed down by atmo- thousands of years. Numerous debris removal concepts have
spheric drag and then burning up by the effects of air fric- been proposed, from electromagnetic ones to surrounding
tion. Debris above 1000 km is only marginally subjected to the object in foam and relying on the enlarged area-to-mass
8.4 Micrometeoroids and Debris 511

Fig. 8.7 View showing a solder

sphere attached to a resistor
adjacent to one of the diodes

Fig. 8.8 Detail of the damage to

the board substrate, some solder
splash between the pad
terminations, and the granular
appearance of the solder

ratio to increase natural atmospheric drag and thus enhance complete treatise on space debris, models and risk analysis
deorbiting and burn out (Pergola 2011). The following the reader may consult the works compiled by Klinkrad
sections include data recorded by the writer, but for a (2006).
512
Fig. 8.9 Space environmental
effects: a temperature;
b micrometeoroids and debris
8.4.2 Debris Emanating from Catalytic Bed
Thruster Motors
Almost all unmanned spacecraft continue to incorporate
monopropellant hydrazine thrusters for attitude control and
trajectory corrections. Certain older thruster designs use a
catalyst bed consisting of iridium-coated grains of alumina
having a diameter of between 200 and 1200 m. Small grains
provide a high surface area of the iridium catalyst for
decomposing anhydrous hydrazine (N2H4) into innocuous
products consisting of nitrogen, ammonia and hydrogen. This
8 Assessment of Post-flight Materials
reaction is strongly exothermic and the hot reaction product
gasses expand and are forced through a nozzle to produce
thrust, usually in the form of specic, high energy impulses
of greater than 10 N. The thruster design, an engineering
drawing, and the appearance of catalyst bed particles are seen
in Figs. 5.445.48. A screen is incorporated into the thruster
adjacent to the nozzle in order to reduce the number of par-
ticles exiting through the nozzle. However, small particles are
known to be expelled through this screen as the mesh size
cannot be so small as to limit the flow of gases and create
high internal pressures. It has been reported that up to 30 % of
8.4 Micrometeoroids and Debris 513

Fig. 8.10 a Detailed view of the soldering side of a terminal pin
attached to the tag board on HST following approximately 21,000
thermal cycles between 100 and +90 C. Some circumferential cracks
are visible and a little orange peel effect (12 magnication).
b Microsection through the solder joint shown in a; the PCB has a
thickness of 1.65 mm. c Only at a magnication of 100 can the
shallow thermal fatigue cracks and surface roughening be detailed, but
the majority of the joints circumference shows no cracking (de Rooij
1995)

these particles have been lost during repeated rings. Such nozzle at right angles to the line of thrust. Temperature
contaminant particles continue cause a degree of debris to the excursions were measured with high accuracy, rapid
space environment and there is an environmentally friendly response (10 ms) thermocouples attached to the GEOS
trend for spacecraft to employ other types of small propulsion hardware. Oil contamination sensors were also distributed
systems based on either electric propulsion using xenon inert around the chamberthese established that there was no
gas as the propellant, or solid green propellants as were contamination caused by the pump down motors of the
described in Sect. 3.5. An extensive overview of propulsion vacuum chamber. The 2N thruster had a nozzle-area
systems and their applications is given by Leyva (2011). expansion ratio of 50, and a nozzle-exit lip angle of 15.
An early experimental set up was used (Dunn and Steinz A diagram showing the motor exhaust plume and sample
1974) to assess both the effects of contamination and heating positions is seen in Fig. 8.11a.
from a 2N hydrazine monopropellant attitude control A brief summary of the results from the test programme
thruster on the performance of conductively-coated solar are as follows:
cells, and very sensitive vitreous-carbon sensors attached to
long booms planned for use on the scientic satellite GEOS. (a) The heating rate and temperature of the sample surfaces
A large vacuum chamber, 3 m long and 2 m in diameter, was exposed to steady state and pulse-mode rings have
used to determine the angular distribution of catalyst parti- temperature gradients dependent upon the nozzle to
cles to be ejected from the thruster during operation. Simple sample distance. After 1000 s, the sample surfaces
measurements were possible by placing glass plates coated reached their maximum, or equilibrium temperatures
with low-outgassing silicone grease in front of the thruster (between 180 and 220 C).
514 8 Assessment of Post-flight Materials

Fig. 8.11 a Schematic diagram

of motor exhaust plume and
position of samples within the
vacuum chamber. b Particle size
and distribution count over
grease-coated slide following
thruster motor ring in pulsed
mode (100 ms on, 900 ms off).
The x-axes show number of
particles per mm2 as well as
actual number of particles
counted. y-axis records the
distance from nozzle centre up to
a circular diameter of 6 cm. The
left-hand graph is for at test
duration of 100 s, on the right is
graph for a 1000 s, with slide at
10 cm from nozzle

(b) Photoelectric yield data for the vitreous-carbon sensors
proved they were not affected by thruster heating or
contamination.
(c) The conductively-coated solar cell cover slips showed
no signicant change in surface resistivity and no
damage from the exhaust plume.
(d) The grease coated slides did collect iridium-coated alu-
mina particlesmany thousands were entrapped on the
slides as black spherical particles and these slides were
subjected to image analysis (Quantimet 720B) and
electron microscopy. The particle size and distribution
count over two slides are seen in Fig. 8.11b. The max-
imum size of particles to be released is 20 m, the
majority had a diameter of under 5.4 m.
8.4.3 Returned Hardware
Micrometeorites and the <20-micron-sized particulate con-
tamination ejected from thruster motors, as documented in
Sect. 8.4.2, are dangerous but do not pose the same threat as
8.4 Micrometeoroids and Debris
space hardware debris. They are smaller in size and have a
lower density, being composed of a material similar to sand.
Debris that originates from spacecraft is generally made of
heavy metallic materials and may consist of steel bolts,
copper wires, released aluminium lens caps, titanium fuel
tanks, and the like. A great effort is now being made by the
various world space agencies to control the amount of debris
in space. As mentioned earlier in this chapter, many exper-
iments have been devised to evaluate the effects of the low
Earth orbit environment on spacecraft. Some of the ndings
will be summarized here, particularly with regard to impacts
of debris and meteoroids on the Eureca spacecraft returned
from space and the solar array of the Hubble Space Tele-
scope (see also Fig. 8.4). Information on more than 20,000
catalogued space objects is assessed by engineers as a basis
for risk assessment in the planning of future spacecraft
orbits. Similarly, a model for the micrometeoroid environ-
ment surrounding the Earth is available and continues to be
upgraded as new data are collected. Unfortunately, the active
removal of debris from orbit using laser evaporation or
debris sweeper mechanisms is in its infancy and may be
economically impossible. The focus must therefore be on
preventative measures such as the maneuvering of spacecraft
and rockets out of heavily trafcked orbits by means of their
thruster motors, preferably into orbits that will carry these
parts back to the Earths atmosphere. Also, the practice of
generating pieces of metal coming from the in-orbit frag-
mentation of pyrotechnic cutters, clamp bands, springs, and
other release mechanisms is now curtailed. One upper stage
of an Ariane IV launch vehicle exploded in a Sun syn-
chronous orbit in 1986. This alone produced 499 catalogued
pieces of debris, among which 61 were still in orbit seven
years later (Laporte-Weywada et al. 1993) and some strin-
gent general policies concerning launcher debris have been
made subsequent to this explosion (covering both Ariane 5
and 6) which drastically reduce the risk of creating debris. In
2009 a defunct Russian satellite collided with and destroyed
a US commercial satellite and this collision alone introduced
another 2000 pieces of trackable debris to the catalogued
space objects. In an incomprehensible 2007 test, China used
a missile to blow-up an old satellite and this created a further
3000 pieces of debris.
The consequence of interaction between hypervelocity
particles and spacecraft surface is rather obvious, but item-
ized in Fig. 8.9b. The Space Shuttle windows had to be
regularly replaced because of particle impacts. The surfaces
of other spacecraft that have been retrieved from LEO and
brought back to Earth have been found peppered with craters
and pin-holes caused by particle impacts. The NASA LDEF
spacecraft, shown in Fig. 8.12, was deployed in an almost
circular orbit by STS-41C at an altitude of 477 km. One of
its main purposes was to gather in situ data of particles in
LEO by means of experiments covering materials, coatings,
515
and thermal systems. 68 months after launch, the spacecraft
was recovered and returned to Earth. At least 34,000 impacts
were recorded on all the space-exposed surfaces; some are
illustrated in Figs. 8.13 and 8.14. Particles were noted to
have pierced aluminium screens 0.8 mm thick but that
multilayer materials of teflonized woven glass were less
damaged (Durin et al. 1993). Solar cells with various cell
covers showed that polymer cell covers provide very little
protection. When the crater impact diameters were about
100 m, 24 % degradation in short-circuit current was
observedwith smaller impact sizes, no measurable chan-
ges in efciency were observed. If the cell itself was dam-
aged, the loss in current was proportional to the size of the
damaged area (Gruenbaum and Dursch 1993). Similar
impact holes were seen on the returned HST array (Fig. 8.4)
and detailed in Fig. 8.15, but here they accounted for an
insignicant loss of power. The Eureca spacecraft was
retrieved in 1993 after almost one year in LEO. With a total
area of 140 m2 the observed impacts included 73 penetra-
tions of the outer layer of the thermal blankets which cov-
ered most parts of the spacecrafts external body, and more
than 2000 impacts on the front face of the solar arrays. The
size of recorded impact features ranges from about 30 m to
6.5 mm (Drolshagen et al. 1996). One impact on Eurecas
painted scuff plate is shown in Fig. 8.16. The ESABASE
meteoroid and debris tool was used to predict particle flux
for the Eureca mission (for details of the model see Berthoud
1994). Figure 8.17 shows that there is a good comparison
between Eureca measured and modelled fluxes and LDEF
measured flux in terms of particle diameter.
8.4.4 Protection Shields
The growing amount of man-made objects in LEO may also
require the protection of sensitive parts of unmanned
spacecraft such as the electronic boxes for Earth observation
satellites. In fact Canadas Radarsat incorporates modica-
tions designed to reduce the crippling effects of impacts.
Shielding blankets cover fuel lines, wiring bundles, and
electronic boxesthese add 17 kg to the satellites mass but
are said to increase the odds of survival of 87 % (David
1995). This satellite orbits the Earth at 800 km where there is
a high level of space debris, including tiny paint chips,
fragments from spent rocket stages, and fluids that leaked
from a retired nuclear-powered spacecraft. Shields with
multi-layer designs are termed Whittle shields as they had
been initially proposed by Frank Whittle in 1947. The
development of the International Space Station (ISS), now
with a nal-build cross-section of about 11,000 m2,
prompted NASA and associated space agencies to pay more
attention than ever to debris hazards during early design
stages (Chenard 1990). The ISS has been designed to
516 8 Assessment of Post-flight Materials

Fig. 8.12 The long duration

exposure facility (LDEF):
a during experiment integration;
b spacecraft in orbit. LDEF was
deployed from the orbiters cargo
bay during STS 41-C mission in
April 1984. The 10-tonne, 9.1 m
long, 4.3 m diameter, and
12-sided space platform was later
retrieved in January 1990.
LDEFs 86 experiment trays
provided immense space
environmental effects data

withstand impactors of up to 1 cm in size by the incorpo-
ration of three kinds of Whittle shields (an aluminium alloy
bumper-plate to break up and melt a particle on impact; an
outer bumper plate with an underlying blanket of Nextel
ceramic cloth and Kevlar fabric spaced away from the
module pressure shell; and, a multilayer shield consisting of
multiple layers of fabric and metal panels to protect the most
critical parts of ISS).
Since only larger space objects can be catalogued and
tracked, it is these that have been avoided by evasive
manoeuversthe uncatalogued smaller debris can, and have
been, shielded against using passive protection techniques.
The Giotto spacecraft, seen in Fig. 4.44, achieved protection
by means of a Whipple Shield consisting of layers of alu-
minium, Nextal and Kevlar. According to studies, shields
can provide reasonable protection against fragments of less
than 1520 mm diameter, with velocities of 1012 km/s.
These represent 78 % of the debris population. European
studies favour shields of aluminium and ceramic cloth
bumpers. Many possible combinations of protection-shield
materials have been investigated by ground simulations and
one is illustrated in Fig. 8.18.
8.4 Micrometeoroids and Debris
Fig. 8.13 Part of the LDEF
surface covered with impact
records
The present activities are focusing on advanced mea-
surement techniques for hypervelocity impact testing. Little
research has been published on the effects of such impacts
on space hardware, and there is almost a total absence of
data related to fracture critical equipment such as pressure
vessels. Poe and Ruckner (1993) have tested pressure vessels
constructed from aluminium alloy 6061-T62, with a burst
pressure rating of 15.3 MPa. It was indicated from their
results that the vessels internal pressure has an effect on the
severity of impact damage. A critical pressure could be
identied above which catastrophic failure usually occurred
(for those vessels it was about 3.45 MPa). By maintaining
517
vessel pressure below the critical value, catastrophic failure
can be minimized, thus reducing the risk of secondary
damage to equipment and personnel.
8.5 Effect of Atomic Oxygen on Materials
A satellite in geosynchronous orbit and operating at a height
of 22 000 km will experience about 90 thermal cycles per
year, and at this altitude only vacuum radiation and thermal
flux will contribute to environmental wear-out mechanisms.
A spacecraft in a Low Earth Orbit (LEO) will, in contrast,
518
Fig. 8.14 Detail of micrometeoroid impact on solar cell
experience up to 16 thermal cycles per day, and at an altitude
of between 200 and 700 km by far the greatest problem is
degradation and erosion of externally exposed materials by
atomic oxygen (see Table 2.3 and Fig. 8.19).
The effects of atomic oxygen were not considered until
engineers began to notice measurable, visible material loss
on Space Shuttle missions after only tens of hours in LEO.
Oxidation and erosion were not a signicant problem on the
relatively short-term Space Shuttle missions, but similar
rates of material degradation were considered unacceptable
on the 3040-year ISS programme. The hazard of radiation
by neutral atomic oxygen arises from the photo-dissociation
of upper atmosphere molecular oxygen. This atomic oxygen
impacts spacecraft surfaces in the ram direction with an
energy of about 45 eV, owing to the relative velocity of the
orbiting spacecraft. The high velocity of the Shuttle (ap-
proximately 8 km/s) produced a flux of atomic oxygen on a
ram-facing surface in the order of 1015 atoms/cm2. This flux
has been found to be strongly chemically reactive with many
of the thermal control lms and paints usually used on the
external surfaces of conventional spacecraft. In particular,
many organic polymer materials were severely degraded
8 Assessment of Post-flight Materials
after only 40 h of exposure. This is evident on Kapton
samples observed after flight (Fig. 8.20), and high recession
rates have been measured by Pippin (1995).
Organic paint binders such as methyl silicone or poly-
urethane react with atomic oxygen to form volatile
carbon-based molecules which are removed from the paint
surface (Zimcik 1987; Jaggers 1993).
Lubricant lms of molybdenum disulphide were exposed to
atomic oxygen during some LEO flights. This lubricant is used
on a variety of spacecraft mechanisms such as
release/deployment devices and precision bearings. Post-flight
examinations indicate that the MoS2 can oxidize to MoO3,
which is an inferior lubricant with a higher friction coefcient.
If mechanisms are continually used in orbit they should be
shielded from atomic oxygen to avoid accelerated wear.
Metals have been observed to react less than polymers to
atomic oxygen attack, although important exceptions are
silver, osmium, and copper as shown in Table 8.2. Metal
oxides (SiO2, indium tin oxide, A12O3, etc.) are nonreactive
because they are already in their highest oxidation state.
Carbonepoxy composite materials have shown a loss of
surface resin, and the carbon bres themselves have been
8.5 Effect of Atomic Oxygen on Materials
Fig. 8.15 Rear-side of HST
solar array illustrating a typical
2 mm diameter hole where
complete penetration of the array
has occurred (through cover
glass, solar cell, and insulation
layers)
seen to be attacked and reduced to porous ridges with little
apparent residual strength or stiffness.
The high capture efciency for atomic oxygen of silver
has created a major problem for the manufacturers of solar
cell interconnectors, as these are generally made of either
thin silver foil or silver-coated molybdenum foil. Degrada-
tion of these interconnection materials severely jeopardizes
the reliability of solar arrays designed or LEO. A combina-
tion of atomic oxygen attack and excessive thermal cycling
(e.g. 30,000 thermal cycles for the Hubble Space Telescope
during a 5-year period, and 60,000 thermal cycles during the
10-year life of Space Station) is known to result in a pre-
mature loss of electrical power generation caused by the
thinning and eventual failure of the solar cell interconnector
itself or failure of the cell-to-interconnector weldment (see
519
Fig. 4.63). The appearance of silver interconnectors attacked
by atomic oxygen is shown in Fig. 8.21. Progressive
degradation of these electrical connections is attributed to
the formation of AgO which, on reaching a certain thickness,
becomes non-adherent and detaches as flakes during thermal
cycling (de Rooij 1985, 1989, 1995). Recent detailed mod-
elling by de Rooij (2010) uses two transport mechanisms to
account for thee transport and oxidation of silver through
pores in the built-up oxide layernamely, gas-flow and
Fickian diffusion. Both mechanisms result in the parabolic
growth of oxide.
Protective measures that either slow down or prevent the
attack of atomic oxygen have been proposed by Dauphin
(1987). They rely on the use of vacuum-deposited protective
coatings known to be resistant to atomic oxygen, such as:
520 8 Assessment of Post-flight Materials

Fig. 8.16 Impact feature on a

painted steel scuff plate. The
crater centre is about 0.5 mm in
diameter. The circular patch
around it where the paint has been
removed by the impact is about
3 mm wide (Drolshagen 1994)

Fig. 8.17 Comparison between

Eureca (EURECAmeas) and
LDEF leading and trailing
measured fluxes (LDEFleadmeas
and LDEFtrailmeas) and the
results of the ESABASE
modelled flux (EURECAtot)
8.5 Effect of Atomic Oxygen on Materials 521

Fig. 8.18 a A simulation of the

wall of the ISS Columbus
Attached Laboratory after being
hit (from the right) by a 10 mm
aluminium projectile at a velocity
of 5.5 km/s. From right to left, the
layers are: the shield or rst
bumper, the second bumper,
multilayer insulation (before the
test, it resembles many layers of
foil but has been shredded by the
impact), the wall of the
spacecraft, and several witness
plates used only in testing,
b Details the damage on the face
of the second plate caused by a
cloud of aluminium plasma and
debris
522
Fig. 8.19 Space environment
effects: a atomic oxygen;
b re-entry
Metallic layers of platinum and aluminium.
Stable oxide layers such as Al2O3, SiO2, or indium tin
oxide.
Silicone layers that are possibly protective owing to the
formation of a thin, self-healing layer of silicon oxide.
Fluorinated layers.
Many articles returned from LEO have been seen to
support a light-brown contamination layer. Analyses have
shown this to be a deposited layer containing silicones
(condensed volatile materials from outgassing sources); this
layer may only be a few angstroms thick; it then oxidizes in
8 Assessment of Post-flight Materials
the atomic oxygen atmosphere to form silicates and silica
both protective compounds that are resistant to further
erosion.
It is clear that a great deal of ground-based research is
developing theoretical models to account for material
degradation and protection methods to meet the challenge of
atomic oxygen. Ground-based atomic oxygen simulation
facilities have been designed and commissioned (at ESTEC,
PSI, DERTS, AODTS, PPPL, and UTIAS). The ground tests
on metals and polymers are giving results comparable to
some LEO exposures, but as often seen in nature, it is the
complex synergistic reactionsbetween atomic oxygen, UV
8.5 Effect of Atomic Oxygen on Materials 523

Fig. 8.20 Effect of atomic

oxygen on Kapton H surface
before and after STS 5 flight.
Degradation is by mass loss and
surface texturing (Micrographs
courtesy of A. de Rooij)
524 8 Assessment of Post-flight Materials

Table 8.2 Reaction of atomic oxygen with selected metals (data from LEO exposure)
Metal (protective Reaction efciency Re 1024 cm3/ Oxide Comment Collated
system) atom type reference
Aluminium 0 A12O3 Self-protecting S
Brass, bronze 0 P
Chromium (100 ) Part eroded Optical properties affected S
Copper (bulk) 0 Cu2O Self-protecting SP
Copper (1000 ) 0.006 Cu2O Optical properties affected S
Gold 0 Nonreactive S
Gold on Ag or Mo 0 Protective until silver diffuses through HR
gold
Iridium 0.0007 S
Lead 0 S
Magnesium ? Slight attack SP
Molybdenum (1000 ) 0.006 MOO3 Slight attack S
Molybdenum (bulk) 0 S
Monel 0 P
Nichrome (100 ) 0 Cr2O3 Self-protecting S
Nickel lm 0 NiO S
Niobium lm 0 S
Osmium >0.026 Heavily attacked S
Palladium ? Optical properties affected S
Platinum lm 0 Nonreactive S
Silver 10.5 AgO Heavily attacked SH
Silver on Mo AgO Heavily attacked H
Solder (SnPb) 0 No visual change H
Stainless steel 0 No visual change P
Tantalum 0 Ta2O5 S
Titanium Slight P
Tungsten 0 S
REF: S Silverman (1995)
H Hamacher et al. (1995)
P Pippin and Bourassa (1995)
R de Rooij (1995)
Volume of material lost 3


Re Total no: of incident O atoms cm atom

exposure, thermal cycling, contamination layers, etc.that
are almost impossible to reproduce in the laboratory.
8.6 Decelerators and Heat Shield Materials
8.6.1 General Examples
An overview of the various groups of materials that have
been selected as the outer skins of thermal protection systems
(TPSs) were discussed in Sects. 2.4 and 2.5. Reusable launch
vehicles and manned capsules that return to Earth have heat
shields made either of lightweight silicon-based ceramic tiles
that have a good resistance to high-temperature oxidation, air
turbulence, and shock (See Figs. 2.14 and 2.15), or from
ablative materials, such as the epoxy resin-based shields on
the Mercury, Gemini, and Apollo capsules, that could only
be used for one descent to the Earths surface. The Gemini
and Apollo TPSs were manufactured from an ultralight
ablator called ULD 100 by the McDonnell Douglas company
in the USA. Beagle 2 and launch vehicles such as Atlas V,
Arianes 5 and 6, Vega and Falcon employed cork, as seen in
Figs. 2.24 and 2.25.
The effects of re-entry into the Earths atmosphere on
materials are shown schematically in Fig. 8.19b. A similar
chain of reactions can be expected for the TPSs attached to
8.6 Decelerators and Heat Shield Materials
Fig. 8.21 Degradation of silver mesh solar cell interconnectors due to
atomic oxygen (Courtesy of A. de Rooij.): a Silver oxide flakes after
atomic oxygen attack during STS flight. Cross-sections revealed the
silver interconnector to have thinned from an original thickness of 34 to
29.5 m. b Flaking of oxide layer due to internal stresses and
unmanned space probes that are designed to descent onto the
surfaces of those planets and moons possessing their own
gaseous atmospheres. Space probe missions can usually be
divided into four phases: launch from the Earth, coast through
space for long periods of time, entry into the planet or moons
atmosphere, and descent onto its surface. The design of these
probes usually consists of two major components: The
high-speed deceleration system and the descent module. The
high-speed deceleration system is often deployable owing to
the volume restrictions of most launch vehicles. It must be
made of a material having a high heat resistance, low mass, and
adequate strength at elevated temperatures. This decelerator is
used only in the entry phase of the mission when aerodynamic
525
microcracks, primary and secondary oxide layers depict the growth and
spalling observed on (a). c A recent observation made by de Rooij
(2010) and D. Adams, is that the oxidized silver surface consists of
nano-flakes and pores
braking is required as the probe makes contact with the planet
or moons atmosphere at a supersonic speed. On reaching a
subsonic speed the decelerator structure is parted from the
probe by the ring of a pyrotechnic device. The probe descent
module consists of a TPS (heat shield) that protects the outer
envelope of the probe, its equipment, and antennae. When the
decelerator is jettisoned, the descent module may be further
slowed down by the deployment of a pilot parachute which
drags out the main parachute. The probe must possess
energy-absorbing properties to cushion it as contact is made
with the planet or moons surface.
The Cassini-Huygens mission is a joint NASA and ESA
project to Saturn and its major moon, Titan. The Cassini
526
Fig. 8.22 A preliminary design
for the Huygens Titan probe. This
exploded view shows a carbon
carbon decelerator and several
beryllium thermal protection
components
spacecraft was launched in 1997 in an orbiter (Cassini) and
probe (Huygens) conguration. After encounters with Jupiter
and a portion of the asteroid belt, it reached Saturn in the year
2003. On completing its rst orbit of the planet, the spacecraft
released the Huygens probe in the direction of Titan. Eleven
days later the Huygens probe entered Titans dense atmo-
sphere (expected to be composed of nitrogen and methane).
Deceleration will be by means of an aerodynamic braking
structure. The design of the Huygens probe is drawn in
Fig. 8.22. The expected entry and descent scenario is illus-
trated in Fig. 8.23 and in situ images are included in Fig. 8.24.
NASA attempted to use solar arrays to power Cassini but
these would not have been possible as the spacecraft would
have become too massive. Radioisotope thermoelectric
8 Assessment of Post-flight Materials
generators were selected as they remain unmatched for power
output for missions to the outer solar systemat the time of
writing, this spacecraft continues to relay incredible images of
Saturn and its many moons (Fig. 8.25).
The actual material that was selected for the
high-temperature thermal protection of the Huygens probe is a
low-density ablator called AQ 60 I (a trademark of
Aerospatial, France). This material is in full industrial pro-
duction, being used for the protection of strategic military
missiles. It is in fact a felt, made up of short silica bres rein-
forced by an impregnation of phenolic resin. AQ 60 I has a nal
density of 0.3 and a total porosity of 84 %. It is not reusable
because part of the material becomes pyrolysed between 400
and 1000 C. After pyrolysis, this heat-shield material is still
8.6 Decelerators and Heat Shield Materials
Fig. 8.23 a The Huygens probe (ESA) after separation from the
mothership, the Cassini Saturn Orbiter (NASA), is depicted in these
artists impressions heading towards Titan. b Huygens: after entry in
Titans atmosphere on 14th January 2005, the parachute was deployed
self-supporting and becomes an efcient thermal insulator
which then behaves like all the silicacarbon systems, being
ablated at temperatures in the region of 2000 C.
During descent, the various sets of instruments on-board
the Huygens probe were activated in order to measure
527
and the heat shield separated. c Huygens descended through the clouds
of Titan, analyzing its atmosphere and transmitting data for 2 h 27 min,
after landing it transmitted signals via Cassini for 72 min before its
batteries drained (Courtesy or ESA)
temperature and pressure, atmospheric composition (by mass
spectroscopy and gas chromatography), cloud particle sizes
and composition, lightning detection, and, for the last phase
of the descent, surface imaging was successful as evidenced
by the Fig. 8.24.
528 8 Assessment of Post-flight Materials

Fig. 8.24 Actual in situ images of Titans surface from Huygens. cobbles (1015 cm in size) are composed of hydrocarbons and water
Bright highland terrains are visible with deep drainage networks ice. (Credit ESA/NASA/JPL/University of Arizona)
indicating erosion by methane rainfall. It is assumed the rounded

Fig. 8.25 NASAs Cassini

spacecraft made its nal close
approach to Saturns large,
irregularly shaped sponge-like
moon Hyperion, taking this image
on 31 May 2015

8.6.2 Beryllium as a Heat Shield high stiffness and strength, and the highest heat capacity
of all metals. Beryllium plate and sheet can be easily
The virtues of beryllium, together with some spacecraft machined, then milled to exacting nal part tolerances
applications, have been reviewed in Sect. 5.7. In TPS by chemical milling, electrochemical machining (in
applications, beryllium can be particularly useful, as it is NaCl and NaNO3 electrolytes), and electric discharge
one of the lightest structural metals known, it has a machining.
8.6 Decelerators and Heat Shield Materials 529

Although beryllium can be chromate conversion coated high-temperature exposure, 600 C in nitrogen (8 h)
with Alodine (as for aluminium alloys) and also anodized to high-temperature exposure, 1000 C in nitrogen (8 h)
increase its corrosion protection properties, little was known salt spray corrosion exposure (5 % NaCl in water for
concerning its suitability as a heat shield material. However, 192 h at 35 C).
it was known that the natural colour of anodized beryllium is
black, and this has the added advantage of increasing emis- The physical and thermo-optical properties were evalu-
sivity (particularly useful for reducing the temperature ated after each of the environmental exposures, and then
excursion of a heat shield during descent) and depressing recorded.
light reflectivity in optical systems. Because of the absence of
experience in the anodization of beryllium in the European C. Results
space industries, a short programme was devised to establish
a processing route and expose samples to a series of envi- (a) Visual aspects
ronmental tests which might simulate some of the rigours of a
heat shield similar to that required by the Huygens probe. To The changes in appearance of the Be-samples due to the
reduce the temperature excursion during descent through the environmental exposures were identical for the thin- and
mainly nitrogenous atmosphere of Titan, a very high surface thick-anodized layers. The thin-anodized samples are shown
emissivity of approximately = 0.7 is desirable. in Fig. 8.26. The original Be-oxide layer appears dark black
The following paragraphs summarize some tests and with the rolling texture of the Be sheet just visible. After
results from this unpublished programme of work (Semerad thermal cycling there is no macroscopic or colour change
and Dunn 1990). (only the thick-anodized nish shows some microscopic
hairline cracks). After the additional 600 C nitrogen exposure
A. Test material there is a pronounced colour change to grey/ochre, but the
oxide layer still shows no macroscopic defects (the micro-
Flat-rolled beryllium sheet, type S-200, with a thickness of scopic hairline cracks are still seen on the thick-anodized
2 mm. Produced by the hot rolling of vacuum hot-pressed samples). After exposure to 1000 C in nitrogen, the surface
block (minimum room temperature properties being becomes rugged. White pustules appear, their number and
344 MPa yield strength, 482 MPa ultimate tensile strength, height being greater for the thick-anodized samples. The col-
and a minimum elongation of 10 %). our reverts to a darker grey and the rolling texture disappears.
Be-samples of size 20 20 2 mm were machined and The salt spray test did not degrade the as-received or
subsequently pre-cleaned in a proprietary solution containing thermally cycled samples. Corrosive products were seen on
carbonates, phosphates, borates, and inhibitors with a pH of all samples exposed to either 600 or 1000 C; this corrosive
9.6. The samples (anode) were held by titanium clamps and the attack occurred only at the microcracks in the anodized layer.
cathode was made of 3 mm diameter aluminium wire.
Anode-to-cathode distance was 20 mm. The anodizing (b) Surface analyses
process was performed in a solution of chromic acid
(100 gCrO3/l) with a current of 16 A/dm2. Magnetic stirring at Infrared spectroscopy showed a low IR-reflectance for the
500 rpm assisted the process. Two anodized-layer thicknesses as-received and thermal cycled samples. After 600 and
were expected, nominally 5 and 15 m, using anodizing times 1000 C exposure in nitrogen, the surfaces appear to have
of, respectively, 10 and 60 min. These are referred to as thin very slightly transformed to nitrate. This was conrmed by
and thick anodized layers in the results. Electrolyte tempera- XPS analysis. X-ray diffraction measurements, with CuK-
ture was 13 C. Samples were spray rinsed in de-ionized water radiation, identied only Be and BeO crystallographic phases
and nally dried in hot air at 120 C. on all the as-received samples and environmentally exposed
samples. However, only the 1000 C nitrogen-exposed
B. Test programme samples contained traces of Be3N2.

Multiple samples were made to assess the initial physical (c) Thermo-optical properties
and thermo-optical properties of the as-received samples.
Groups of samples were then exposed to environmental The solar absorption for the thin(thick)-anodized layers
extremes: decreases from 0.95 (0.96) to 0.65 (0.69) after the 600 C
nitrogen exposure, and increases again to 0.8 (0.73) after the
thermal cycling under vacuum, 200 cycles from 150 to 1000 C nitrogen exposure. Thermal cycling and salt spray
+100 C, per ECSS-Q-ST-70-04 corrosion have only very minor effects. The observed
530 8 Assessment of Post-flight Materials

Fig. 8.26 Change in appearance of the Be-samples after different environmental tests (Be-oxide process 1:5 m)

decrease of solar absorptance in the course of the visual spectral range, by an increased transparency in the
high-temperature nitrogen exposure lowers the heating by infrared, so that part of the radiance is reflected/emitted by
Saturn/Titan and solar radiance, but is considered much less the Be-substrate surface.
important than the observed changes in emissivity. The total
hemispherical emittance (average of three samples) is shown (d) SEM observation of the environmentally tested samples
in Fig. 8.27. The observed decrease in the emittance after the
600 C exposure can be understood either as a change of the The environmentally exposed samples were carefully han-
intrinsic properties of the Be-oxide layer, or in analogy to the dled and mechanically broken to enable examination of their
8.6 Decelerators and Heat Shield Materials
Fig. 8.27 Total hemispherical
emittance, n, of the anodized
Be-samples in sequence of the
environmental tests
cross-sections. Sharp notches were machined into the Be
from the underside; the samples were then pulled in tension
and this resulted in brittle fractures at right angles to the
anodized surface. Figure 8.28 shows that the nominally thin
(5 m)-anodized layer is in fact 2 m in thickness. Similarly,
Fig. 8.29 shows that the thick-anodized layer is also of
inadequate thickness, being only 8 m instead of the nominal
15 m. After thermal cycling there is good adherence
between oxide and substrate, but more microcracks in the
thicker oxide layer. This holds true following thermal cycling
plus the 600 C nitrogen exposures. After an additional
1000 C exposure, plus salt spray testing for 8 days, there are
extensive signs of oxide breakdown and corrosion of the
Be-substrate and this has occurred for both oxide thicknesses.
D. Conclusion
The thin (actually 2 m)-anodized layer shows a better
mechanical performance regarding cracks and detachment,
but the thicker layer (actually 8 m) is better suited in view
of a high emittance. For both layers the emittance of the
anodized surface decreases below the required = 0.7 after
the 600 and 1000 C nitrogen exposures.
8.6.3 Alternative Heat Shield Materials
An almost innite number of composite materials, metallic
alloys, ceramics, and glasses, together with their numerous
531
oxidation and heat-resistant coatings, have potential TPS
applications. As previously described, some very suitable
materials have been developed and are in use today. However,
others are being evaluated for future applications such as those
needed for future space transportation vehicles and space-
planes where surfaces will reach 1600 C. Heat shields such as
ceramic tiles may be attached to these spacecrafts structural
materials in order to provide for protection during descent
phases of a mission but, for future designs, the outer skin
structure will also need to be also resistant to this environment.
The following materials are presently identied for structural
applications in the various high-temperature ranges:
10001600 C Carboncarbon composites, also with
active cooling; these include carbon felts
(e.g. Calcarb, a rigid felt developed in
Ireland; Carbone Lorraine, a flexible felt
made in France, and Aerospatials AQ
61)limitations apply owing to oxidiz-
ing properties, but in space these mate-
rials have excellent outgassing
characteristics
7001000 C Carbonsilicon carbide composites, also
with coatings, and compacted silica
powder (e.g. Microtherm, manufactured
in the UK by the Micropore Company)
400700 C Titanium aluminides and
high-temperature titanium alloys such as
532
Fig. 8.28 Anodized beryllium, thin (2 m) coating, showing fractured sections after: a 200 thermal cycles (150 to +100 C), b thermal cycling
plus 600 C nitrogen exposure/8 h, and c all exposures, plus 1000 C nitrogen/8 h and salt spray/192 h
IMI 834 and Ti 1100 with special
coatings
less than 400 C Conventional metal alloys, oxide-disper-
sion-strengthened aluminium alloys,
carbon bre reinforced plastics, etc.
Two concepts for the oxidation protection of ceramic
matrix composites are shown in Fig. 8.30. The carbon bres
are encapsulated in a matrix of silicon carbide that is intro-
duced by processes involving liquid phase inltration and/or
chemical vapour inltration.
For external oxidation protection, the multicomponent
oxidation protection (or thermoviscous external oxidation
coating) was developed in Europe for the Hermes space-
plane. It has an effective operational temperature range from
500 to 1600 C, a proven repairability, and utilizes relatively
inexpensive materials and processes that can be readily
8 Assessment of Post-flight Materials
industrialized. The disadvantages of this system are that the
thermoviscosity of the coating does not permit for very high
mechanical clamping stresses at xation points, or for
extremely high pressures of gas flow. The internal oxidation
protection system illustrated in Fig. 8.30 has been developed
and tested. Initially, the C/SiC material with an outer coating
of chemical vapour deposited SiC was tested in static oxi-
dation tests (Lamouroux et al. 1994)that was without the
internal oxidation protection. The results were encouraging,
but more recently a great improvement has been achieved
when the matrix composition was modied to contain an
internal oxidation protection formulation. This formulation
consists of a glass sealing agent (having a good wetting to
the carbon bres and the silicon carbide matrix, and a low
gas pressure under vacuum at high temperatures), an oxygen
getter, and refractory particles.
8.6 Decelerators and Heat Shield Materials
Fig. 8.29 Anodized beryllium, thick (810 m) coating, showing fractured sections after: a 200 thermal cycles (150 to +100 C), b thermal
cycling plus 600 C nitrogen exposure/8 h, and c all exposures, plus 1000 C nitrogen/8 h and salt spray/192 h
8.6.4 High-Temperature Fasteners
Adhesive compounds are available to secure heat shield
elements onto spacecraft outer structures. The Shuttle tiles
were described in Sect. 2.4 and the fact that on re-entry the
outer skin of this vehicle is forced to bend owing to aero-
dynamic pressure and, in part, to the differential heating rates
and enormously high temperatures seen by the tiles them-
selves. Silicone adhesives such as RTV 560 are used to bond
the ceramic tiles onto the aluminium skin structure, as this
adhesive can flex and absorb most dimensional mismatches.
In many applications, high-temperature fasteners will be
required for the attachment and disassembly of spacecraft
parts including those associated with rigid external insula-
tion (hot outer shell elements), ceramic heat shields, hot
metallic structures, and items connected to propulsion
systems.
533
Wherever possible the designer will use graphite- and
ceramic-based fasteners for these xations, but screws, bolts,
and nuts made of these materials are very difcult to man-
ufacture and procure. Also, they can only be used once and
need to be drilled free unless in the particular application it is
possible to apply a suitable lubricant. The ceramic matrix
material described at the end of Sect. 8.6.3, and illustrated in
Fig. 8.30, is being developed for fastener applications. Bolts
and nuts (size KM 8) have been fabricated from C/SiC
material with coatings of SiC. Test programmes utilized
these fasteners (they were coated with the additional external
oxidation protection layer seen in Fig. 8.30) to mechanically
hold boreholed atmospheric re-entry demonstrator plates for
long periods at 1600 C in air. Under these conditions the
fasteners survived a maximum load of 9.75 kN. Similar bolts
are now being fabricated with enhanced internal oxidation
protection, as drawn in Fig. 8.31.
534
Fig. 8.30 Methods for
protecting ceramic matrix
composites from oxidation
(Courtesy of MAN Technologie,
Germany)
Metallic fasteners may be more suitable than the graphite
or ceramic fasteners for applications where there are specic
demands for ductility, thermal shock resistance, repeatable
joining, or electrical and thermal conductivity. This may be
important during the integration of heat exchangers,
propulsion system parts, and heat shields. The most common
high-strength spacecraft fastener materials are listed in
Table 5.1 and described in Sect. 5.4, but for general space-
craft applications even the most advanced alloy (Waspaloy)
has a maximum service temperature of only 800 C.
When the Periodic Table is consulted, it is noticed that
there are 15 metallic elements with sufciently high melting
points (over 1700 C) for the very-high-temperature appli-
cations required by the space industry (Lupton 1990). Some
have been referred to in earlier chapters with respect to their
use in electron tubes. Zirconium and vanadium have to be
excluded owing to poor high-temperature strength, and
8 Assessment of Post-flight Materials
osmium is seriously toxic. The remaining twelve metals can
be placed into three classes, described as the refractory
metals (hafnium, niobium, tantalum, chromium, molybde-
num, and tungsten), rhenium from the 7th group in the
Periodic Table, and the platinum group metals (platinum,
ruthenium, rhodium, palladium, and iridium). Oxidation
resistance is the most important characteristic which differ-
entiates the three classes. Most refractory metals react
strongly with oxygen at elevated temperatures and then
become severely embrittled at room temperaturehowever,
chromium, molybdenum, and tungsten have volatile oxides
and they generally lose section thickness. Rhenium is similar
to, but slightly more favourable than, tungsten. The platinum
group of metals is usually highly resistant to oxidation but is
costly, and has a low strength at temperatures greater than
1100 C. These strengths can be increased by alloying, and
articial hardening is possible by means of adding a ne
8.6 Decelerators and Heat Shield Materials
Fig. 8.31 Concept for the
enhanced oxidation protection of
ceramic matrix composite
fasteners (Courtesy of MAN
Technologie, Germany)
dispersion of a thermodynamically stable oxide phase (e.g.
zirconium oxide) to alloys based on platinum and rhodium
(Selman et al. 1974; Lupton 1990).
Metallic fasteners for applications in the region of 1600
C have been produced from the refractory metals, particu-
larly those based on niobium (sometimes called columbium).
Data exist for alloys Nb-752, Mo41Re, and Ta10 W.
Oxidation-protection coatings are essential for these alloys
and several have been developed including silicides (e.g.
Si20Cr20Fe and SiCrTi), aluminides (e.g. A140Si and
AlSiTi), and noble metals such as platinum and iridium.
An example of one fastener design is shown in Fig. 8.32.
Coating thicknesses may range from 20 to 200 m,
depending on the application, and may be applied by
physical vapour deposition, plasma spraying, pack cemen-
tation, and slurry deposition. Loss of ductility resulting in
catastrophic failure may occur if local flaws exist in the
oxidation-protection coating. Processing techniques and
nondestructive testing for these types of fastener have been
rened up to a point where coating defects can be elimi-
nated. However, it is important to apply stringent assembly
rules that avoid damage to the coatings from either careless
535
handling or local stressing by the edges, corners, and
recesses of materials being joined. Although the oxidation
resistance of assembled Nb-752 fasteners (coated) has been
successfully demonstrated, it is noted that shear forces
between screw and nut must be minimized, and this requires
special assembly techniques.
8.7 Manned Compartments
8.7.1 General Conditions
Human beings quest to live in space began in 1961, with the
rst orbital flight of Soviet cosmonaut Yuri Gagarin in the
Vostok 1 spacecraft. Later in the 1960s, the US Mercury and
Gemini programmes developed manned spaceflight, and
these were followed by the Apollo programme that culmi-
nated with several landings on the Moon. Skylab was the
rst American space station; it was launched in 1973 and
visited by three sets of astronauts for periods of up to
84 days. Soviet cosmonauts stayed in space for even longer
periods during the Salyut-6 and Mir space station missions
536 8 Assessment of Post-flight Materials

Fig. 8.32 Silicide-coated

fasteners developed for assembly
of the Thermal Protection System
of re-entry vehicles. These
oxidation-protected fasteners are
machined from niobium alloy
752. After machining (1) the parts
are coated with SiCrFe and
annealed (2). The arrow at (3)
shows the pre-oxidized
deliverable nish. These fasteners
have been tested at elevated
temperatures and do not degrade
in mechanical properties
following exposure in air to
1150 C for more than 500 h.
They also survive temperature
cycling to 1300 C for 40 h
(Courtesy of Plansee
Aktiengesellschaft)

(longer than one year). This clearly demonstrated that with
supplies of air, food, and water, people can function effec-
tively in space for long periods in order to perform useful
experiments, observations, and services for those of us who
must remain on Earth. In the former Space Shuttle era, the
human role has taken advantage of the lessons learnt by the
early manned flights into space. Although not designed for
permanent presence in space, it was regrettable that the
Skylab space station was unceremoniously dumped into the
Indian Ocean (and parts of Australia) because delays at the
beginning of the Shuttle programme prevented a refurbish-
ment mission. However, the Russian Mir space station,
launched in 1986, did prove to be an extremely durable
vehicle owing to its design innovations which include a
ve-port docking adapter onto which can be attached addi-
tional laboratory modules (see Fig. 8.33). This legend
proved Russia to be a leader in space technology as Mir
endured 15 years in orbit, three times its planned lifetime.
By the time the International Space Station Alpha (Fig. 3.1)
was launched and operational around 2003, Mir was still
demonstrating its usefulness by some forms of low-cost
extended missions. The European-built Spacelab has flown
on many missions inside Space Shuttle cargo bays since
1983. After several refurbishments and recongurations with
new equipment and experiments, detailed inspection of the
interior structure of the fourteen-year-old Spacelab (see
Fig. 3.2) shows no sign of degradation.
The success of the existing space station and laboratory
projects has been, in part, due to the importance that was
given to the proper selection of materials and processes
during the design and fabrication stages of these spacecraft.
Some considerations are familiar to those engineers who
have designed or lived in conned volumes, such as sub-
marines! Past experience has been accumulated by the var-
ious space agencies and this has formed the basis for
establishing acceptable requirements. The majority of the
metallic materials and the manufacturing processes that were
selected for fabricating both manned and unmanned space-
craft have been touched on between the covers of this book.
There are though, specic problems that might occur within
manned spacecraftthe main ones being highlighted in
Fig. 8.34and these can only be avoided by additional
requirements. Chapters 2 and 3 detailed the main product
assurance requirements set by the space agencies to cover
general spacecraft materials and processes. An excellent
overview of the technical issues related to materials safety
for manned space have been documented by Pedley (2009).
For manned vehicles, some of the additional safety
requirements relate to:
Structural design providing ultimate factors of safety
equal to or greater than 1.40; and, primary structures and
support bracketry precludes failure by the propagation of
pre-existing flaws based on fracture control; special
design burst pressure levels for pressure vessels
Metallic materials being resistant to cracking due to
stress corrosion (i.e. select from the alloys listed in
Table 2.2)
8.7 Manned Compartments 537

Fig. 8.33 a Photograph of the

Russian Space Agencys Mir
space station docked to the Space
Shuttle Atlantis in June 1995.
b Mir station taken at the time of
the EuroMir mission in 1996 (the
ESA astronaut Thomas Reiter
became the rst non-Russian
astronaut to spend 180 days
on-board)
538
Fig. 8.34 Space environmental
effects: manned volumes
Propellant compatibility
Oxygen compatibility
Avoidance of hazardous materials, including radioactive
metals, beryllium (toxicity problems if scraped), and
mercury (as this liquid metal, when spilt, can penetrate
the aluminium alloy walls of the vehicle causing leakage
of the atmosphere)
Nonmetallic materials being controlled for flammability,
toxicity, and odour (they will therefore have low off-
gassing properties). They need to resist moisture and
fungal growth.
From a European viewpoint, the metallic materials that
were selected for projects associated with manned spacecraft
requirements (i.e. Spacelab and its on-board experimental
apparatus, for Eureca, Hubble Space Telescope, etc.) have
functioned correctly without any problems related to safety
and reliability (Dricot et al. 1994). A short list of Preferred
Materials that was compiled for the Columbus project in
1994 enabling a check to be made concerning the number
and extent of materials tests that were, at that time, still
outstanding has been reproduced in Table 8.3.
8 Assessment of Post-flight Materials
8.7.2 Solder Assembly Defects
Occasionally, equipment associated with Spacelab experi-
ments has failed during flight. When it can be demonstrated
that the potential failure of experiments will not be
life-threatening or a danger to the space mission, some of the
specic materials and processes requirements for manned
projects may be waived. This can achieve cost reductions for
universities, and the like. It can also cause the early cur-
tailment of an interesting in-orbit investigation.
One recent equipment failure involved a video camera that
was being used to record a fluid dynamics experiment. The
liquid under evaluation in zero gravity was made up of 95 %
French tap water mixed with 5 % Ajax washing solution (one
of several candidates being considered as an aqueous
lm-forming foam (AFFF), for the suppression of res in the
Columbus module). An astronaut was lming the morpho-
logical changes that occur under zero-g, when small droplets
of water and foam were injected into an orbit space labora-
tory. Soon after commencing the experiment, the video
camera suddenly stopped functioning. A few small globules
of water had made contact and intruded into the camera.
8.7 Manned Compartments 539

Table 8.3 First list of preferred materials for short-term evaluation

Sample Material identication F1 Tox Odor Mbio Tdec Fow Corr SCC Arc T Outg Tcyc UV Atox FL21
0101 AA 1100 N/A N/A N/A T N/A N/A A A N/A N/A N/A N/A N/A N/A
0102 AA 2024-T81 (bar, rod) N/A N/A N/A T N/A N/A A A N/A N/A N/A N/A E N/A
0103 3.3214-T6 (6061-T6) N/A N/A N/A T N/A N/A A A N/A N/A N/A N/A E N/A
0104 3.4364-T7351 (7075-T73) N/A N/A N/A T N/A N/A A A N/A N/A N/A N/A E N/A
0105 AA 2219-T81 N/A N/A N/A T N/A N/A A A N/A N/A N/A N/A E N/A
0302 INCONEL 718-PH N/A N/A N/A T N/A N/A A A N/A N/A N/A N/A E N/A
0401 3.7164.1 N/A N/A N/A T N/A N/A A A N/A N/A N/A N/A E N/A
0402 3.7164.7 N/A N/A N/A N/A N/A N/A A A N/A N/A N/A N/A N/A N/A
0602 AISI 304L Cond. A N/A N/A N/A T N/A N/A A A N/A N/A N/A N/A E N/A
0603 AISI 316 Cond. A N/A N/A N/A T N/A N/A A A N/A N/A N/A N/A N/A N/A
0604 AISI 347 Cond. A N/A N/A N/A T N/A N/A A A N/A N/A N/A N/A N/A N/A
0606 1.4944.9 (A 286 SH) N/A N/A N/A T N/A N/A A A N/A N/A N/A N/A E N/A
0607 17-7 PH-CH 900 N/A N/A N/A T N/A N/A A A N/A N/A N/A N/A N/A N/A
1001 ARALDIT AV138/HV998 A A A T N/A N/A N/A N/A N/A A N/A N/A E N/A
1002 ECCOBOND 57C/9 T T T T T N/A N/A N/A N/A A E N/A E N/A
1003 RTV-S 691 T T T T T N/A N/A N/A N/A A A A E A
1004 SCOTCH-WELD 2216 B/A A A A T T N/A N/A N/A N/A A E N/A E N/A
1005 PARYLENE C A A A T T N/A N/A N/A N/A A N/A N/A N/A N/A
1101 G410810 N/A A A N/A N/A N/A N/A N/A N/A A E E E A
1102 SCOTCH Y-966 T A A N/A N/A N/A N/A N/A N/A A A N/A E A
1103 TESAMETAL 4500 Silver T T T T N/A N/A N/A N/A N/A A E N/A E N/A
1104 MYSTIC 7367 T T T T T N/A N/A N/A N/A A E E E A
1201 CHEMGLAZE Z306 A A A N/A N/A N/A N/A N/A N/A A A E E N/A
1202 ELECTRODAG +501 A N/A N/A N/A N/A N/A N/A N/A N/A A A A E A
1203 PYROLAC PSG 120 FD A N/A N/A N/A N/A N/A N/A N/A N/A A A E E A
1204 ALEXIT 40623 104 G T A A T T N/A N/A N/A N/A A N/A N/A N/A N/A
1301 BRAYCO 815 Z A A A T T N/A N/A N/A N/A A N/A N/A N/A N/A
1302 BRAYCOAT 601 A A A T T N/A N/A N/A N/A A N/A N/A N/A N/A
1303 BRAYCOAT 602 N/A N/A N/A T T N/A N/A N/A N/A N/A N/A N/A N/A N/A
1304 TIO-LUBE 460 A A A T N/A N/A N/A N/A N/A A E N/A E N/A
1401 ECCOSIL 4952/50 A A A T T N/A N/A N/A N/A A N/A N/A N/A A
1402 ECCOFOAM FPH/12-2 H A A A T T N/A N/A N/A N/A A N/A N/A N/A N/A
1501 CHO-THERM 1671 A A T T T N/A N/A N/A N/A A E N/A N/A N/A
1503 MXB-6001/1581 T T T T N/A N/A N/A N/A N/A T E E E N/A
1504 BETA CLOTH X 389-7 A A A N/A N/A N/A N/A N/A N/A A E E E A
1505 GUDE SPACE 18 DPTH A A A T T N/A N/A N/A N/A A N/A N/A N/A N/A
1506 DURA VER-E-Cu 104 (FR4) A T T T T N/A N/A N/A N/A A N/A N/A N/A N/A
1507 FLEXAFIT SG 2212 A A A T N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A
1601 VITON B A A A T T N/A N/A N/A N/A A E E E N/A
1602 5 SL 1617 T T T T N/A N/A N/A N/A N/A A N/A N/A N/A N/A
1701 DACRON B 4 A A A A N/A N/A N/A N/A N/A N/A A E N/A E N/A
1702 PTFE 9900/0100 A A N/A T T N/A N/A N/A N/A A E E E A
1703 THERMOFIT KYNAR A A A T T N/A N/A N/A N/A A N/A N/A N/A A
1704 INSULITITE HFT-A Blue T A T T T N/A N/A N/A N/A A N/A N/A N/A A
1705 TY RAP TYZ 28 M A A A T T N/A N/A N/A N/A A N/A N/A N/A N/A
1901 Wire type: 1871 T T T T T T N/A N/A T A N/A (A) N/A (A)
1902 Wire type: 3901/002 T T T T T T N/A N/A T A N/A (A) N/A (A)
1903 Wire type: 3901/1 T T T T T T N/A N/A T A N/A (A) N/A (A)
1904 Wire type: 3901/2 T T T T T T N/A N/A T A N/A (A) N/A (A)
1908 Wire type: SPA 2110 T T T T T T N/A N/A T A N/A (A) N/A (A)
(continued)
540 8 Assessment of Post-flight Materials

Table 8.3 (continued)

1909 Wire type: SPB 2110 T T T T T T N/A N/A T A N/A (A) N/A (A)
1910 Wire type: SPC 2110 T T T T T T N/A N/A T A N/A (A) N/A (A)
1911 Wire type: MTV-BTV T T T T T T N/A N/A T A N/A (A) N/A (A)
1912 Coax Cable: 50 CIS T T T T T T N/A N/A N/A A N/A (A) N/A (A)
1913 Coax Cable: R 59 T T T T T T N/A N/A N/A T N/A (A) N/A N/A
1914 Wire type: 1872 T T T T T T N/A N/A T A N/A (A) N/A N/A
1915 Wire type: 3901/001 T T T T T T N/A N/A T T N/A (A) N/A N/A
Tests Test houses Ratings
Fl: Flammability 24.5 % O2 SCC: Stress-corrosion C: DCN T: To be tested
cracking (CERTISM)
Tox: Toxicity ArcT: Arc-tracking D: DERTS N/A: Not applicable
Odor: Odour Outg: Outgassing E: DASA-RIT A: Test data available
Mbio: Microbial resistance Tcyc: Thermal cycling S: SINTEF (A): Tested per ESA space components spec.
Tdec: Thennal decomposition UV: Ultraviolet radiation E: Evaluation
Fow: Flammability of wires Atox: Atomic oxygen
Corr: Corrosion F121: Flammability 21 % O2

A metallurgical investigation was performed on the video
camera six weeks after the failure event. A summary of the
ndings is as follows:
the black plastic video camera case was not hermetically
sealed and there were many places where liquid could
penetrate
close inspection showed that not all of the assembled
components were covered with a very thin layer of
conformal coating; some of the noncoated regions of the
assembly had corroded (as shown on Fig. 8.35)
the residues from liquid solder flux were present adjacent
to many solder joints
one typical region that supported corrosion products was
photographed at a high magnication (as shown in
Fig. 8.36 in the colour section), the voluminous products
having characteristic colours: red, blue-green and white
the distinct corrosion products were carefully removed
from the board and chemically analysed by
energy-dispersive X-ray analysis.
the results are given in Figs. 8.37, 8.38 and 8.39, white
deposits being mainly lead oxide, bluegreen deposits
containing copper chloride, and red deposits containing
cuprous oxide (chlorides are expected to originate from
solder flux, as the AFFF liquid and original French tap
water showed no white precipitation when mixed with a
solution of silver nitratea very sensitive test for
chlorides)
dark, electrically conductive traces (similar to the elec-
trochemical migration illustrated in Figs. 6.30, 6.49 and
6.50) were seen beneath some of the corrosion products,
indicating that the camera failure mechanism resulted
from the migration of tin between the very small track
spacing when the circuit was damp.
It was concluded that this commercial video camera con-
tains design weaknesses that permit the intrusion of liquids.
Also, the quality of workmanship fell short of that normally
required for space hardware (ECSS-Q-ST-70-38); some line
spacings were extremely narrow, the residues of soldering
flux had not been removed, and the circuit had not been
adequately conformally coated. High humidity and conden-
sation of water onto all internal surfaces, including electronic
housings, within manned volumes is always a potential
problem. This is due to the fact that life-support systems may
not always be able to remove humidity quickly enough from
the habitat module or laboratory module, particularly at times
when there is a large temperature difference between the
walls, floors, and working space within the galleys.
It is interesting to note that the ISS includes a Mainte-
nance Work Area consisting of a portable foldable work-
bench and tabletop. Astronauts can unfold it anywhere and
clamp it to a slotted trackthese tracks are located on most
of the floor-to-ceiling racks inside station. Many tests have
been made related to soldering operations on ISS while
orbiting the Earth. These were made during a dedicated work
programme in order to check if solder joint repair is feasible
inside the manned modules (Flin 2005). The results are
fascinating, and can be summarized as follows:
when rosin-cored solder is applied to the heated tip of a
soldering iron the solder becomes molten and clings to
the tip, the solder liquid then begins to spin round and
round, getting progressively faster.
In low-gravity surface tension forces are greater than on
Earth so soldering processes are more difcult to perform.
Splatter from liquid solder forms drops held together by
surface tension, these float through the air in space and
can be hazardous.
8.7 Manned Compartments 541

Fig. 8.35 Video camera

electronics after disassembly. The
miniature electronic components
are surface mounted onto the
facing PCB. Localized regions of
corrosion products are present on
solder joints and tracks (arrowed)

Fig. 8.36 Colour

photomicrograph of corrosion
products seen on failed video
camera electronic circuit
(approximately 50
magnication)

Low gravity eliminates convection and cooling is
initiated at the molten solder surfaces, so entrapping
gasses and solder flux and this leads to poor quality
joints.
Debris from cutting and soldering neds to be captured
and disposed of using the stations vacuum cleaner.
Smoke created during soldering does not easily disperse,
possibly leaving the operator without a good view of the
workpiece.
Clearly soldering and conducting repairs in space is
challenging and it may be needed to select conductive
542 8 Assessment of Post-flight Materials

Fig. 8.37 The majority of

corrosion product on the circuit
board is white. EDAX analysis
shows this to be mainly lead
(oxide) with tin (oxide)
originating from attack of the
solder joints. Compare with the
colour photograph shown in the
colour section (Fig. 8.36)

Fig. 8.38 EDAX analysis of

green product. This contains
mainly copper with associated
chlorine, calcium, and lead. This
is similar to green plague, the
reaction product of solder flux
with copper oxide (solder flux
generally contains chlorine)

adhesives for repair work, despite their inferior properties
when compared to traditional soldering (see Sect. 6.13).
8.7.3 Inspection of Spacelab Post-flight
Hardware
Spacelab has been one of ESAs most important projects and
was a major contribution to the US Space Shuttle
programme. The hardware was manufactured in Europe after
long development activities in the elds of materials and
processes (M&P). This nal section describes the ndings of
a detailed inspection of the various Spacelab-1 hardware at
the Kennedy Space Centre. The inspections were made by
the writer in the presence of long-term NASA personnel who
had been associated with the day-to-day use and operation of
the hardware. The full report and an extensive description of
the objectives, methods and ndings from this investigation
8.7 Manned Compartments 543

Fig. 8.39 EDAX analysis of red product showing much copper and lead. This is similar to red plague, the corrosion product of copper to form
red cuprous oxide

Fig. 8.40 General view of the Spacelab processing and integration area in the O and C building at KSC during 1982
544
Fig. 8.41 Spacelab-1 located in
the Shuttle Cargo Bay in-orbit
have been published (Dunn and Stanyon 1997). This
end-of-service inspection took place 16 years after Space-
labs integration (Fig. 8.40) and one year after the nal
in-orbit flight (Fig. 8.41). Spacelab flew on 22 Shuttle mis-
sions between 1983 and 1996. Various flight congurations
were qualied and flown with the pressurised, habitable
laboratory module. The other components were: the Space-
lab Pallet, a U-shaped un-pressurised platform onto which
were mounted experiments and telescopes having a direct
exposure to space; an Instrument Pointing System attached
to the Pallets for precise pointing and observation of the
Earth and our planetary system; and, a Tunnel connecting
the laboratory to the crew compartment of the Shuttle.
The main objective of the inspections was to identify
M&P problems which had been experienced and during
Spacelabs life, assess their importance and feed this infor-
mation to the engineers engaged in the selection of materials
and processes for future European manned space missions,
such as the Columbus module of ISS.
Some problems were identied during the inspection
visit, but all-in-all, the post-flight hardware was considered
to be in an excellent condition and generally it was the
materials known as having a nite shelf- or service-life
which had degraded. Documented problems (from
non-conformance reports, material review boards, etc.) were
made the subject for detailed inspection, they comprised of:
8 Assessment of Post-flight Materials
(a) Corrosion: of the Alodined weld and its heat affected
zone, degradation of anodised surfaces and condensation
in cold areas.
(b) Painted surfaces: peeling, flaking, degradation of optical
properties, atomic oxygen effects.
(c) Thermal protection: thermal blanket degradation, wear,
tarnishing, etc.
(d) Interface problems: wear, friction on trunnions and
mounting points, fasteners, wear of electrical connec-
tors, fluid disconnect problems, threaded connections,
accidental damage (dings) needing added protection.
(e) Limited life items/physical degradation: seals and asso-
ciated leak rates, relief valves (positive and negative
pressure), brush motor start-up after long periods of
non-activity (sticktion), harness attachment points and
insulation damage, condensate heat exchanger degrada-
tion, evidence of microbial contamination.
The following gures illustrate some of problems listed
abovetheir captions describe in more detail the topo-
graphical evidence (Figs. 8.42, 8.43, 8.44, 8.45, 8.46, 8.47,
8.48, 8.49, 8.50, 8.51, 8.52, 8.53, 8.54, 8.55 and 8.56).
8.7 Manned Compartments 545

Fig. 8.42 The waffle structures

and internal walls consist of
anodised AA2219-T851, the TIG
welded plates are ground,
polished and brush Alodined. The
brush Alodined weldments
a show some discolouration, but
after inspection of all accessible
welds, and at higher
magnication, only one suspect
area b was identied and later
NDT found this to be tarnish. No
areas of corrosion or staining
were observed
546 8 Assessment of Post-flight Materials

Fig. 8.43 a General view of heat

exchanger and the right-hand
region of corrosion. b The
corroded aluminium and brown
corrosion product appears to
result from flux residue
8.7 Manned Compartments 547

Fig. 8.44 Detail, showing some

small areas of Chemglaze white
paint removed from rounded
corners, indicating the need to
check the suitability of the primer.
All other areas coated with
Chemglaze (also external and on
Pallet) were in excellent condition

Fig. 8.45 Solar blankets are

undamaged, there is slight
darkening of the cloth
surrounding ESA logo but none
of the colours have altered or
become bleached due to the space
environmental exposure
(radiation and atomic oxygen)
548 8 Assessment of Post-flight Materials

Fig. 8.46 Delamination of the

goldized-Kapton thermal blanket
is seen in this area and an
indication of staining. Only this
region, subjected to a small bend
radius was damaged. Nomex
cloth is now used and the NASA
logos painted directly to this
material have proven to be
resilient to colour degradation in
space

Fig. 8.47 These plastic

attachment devices for holding
together the blankets have a
tendency to pop-off and the
plastic ring, shown here, can
uncontrollably spring away from
the retaining grove of the tubular
post (insert) during flight
(creating debris)
8.7 Manned Compartments 549

Fig. 8.48 Chromium-plated main module trunnion (approximate aluminium-rich coating on a mounting point; the Al-to-Cr interfacial
diameter 80 mm) removed after flight and ground handling. An couple is not a recommended combination as surfaces can cold-weld
extensive deposit of aluminium alloy is seen to adhere to this bearing under the combined effects of vacuum and pressure (see especially
surface. It was determined to have been transferred from an Figs. 2.4a and 2.5

Fig. 8.49 The ducting/bellows

interface providing cooling to the
racks created a problem because
the bellows material is very rigid.
A more flexible material would
make the connection and
clamp-securing operations in orbit
less difcult and reduce the
amount of damage and shred
material. Scufng and wear
causes contaminant particles to be
released
550 8 Assessment of Post-flight Materials

Fig. 8.50 It is probable that

these flight harness materials were
manufactured 15 years before this
nal inspection. Damage to the
Kapton insulation around a cable
attachment point reveals
silver-plated copper strands of the
shielding. Also, in the cable
bundle, the outer layer of the
wrapped Kapton has crazed and
become embrittled, with some
very small flakes being shed.
Wires from the same batch/spools
kept for the same period in a
controlled store showed no
degradation. The ISS baseline for
wire insulation is Teflon-coated
Kapton, which seems less
susceptible to handling damage
than bare Kapton
8.7 Manned Compartments 551

Fig. 8.51 Several wire bundles

containing spliced wire joints
made with Solder Sleeves. These
wire interconnections would not
be permitted for vacuum
applications due to the entrapped
solder flux and its residues (seen
as a yellow-brown layer beneath
the sleeving)
552 8 Assessment of Post-flight Materials

Fig. 8.52 Top photograph show

a new on-board tool holder made
from Solamide 301 foam for the
ISS. It is expected to have a
longer life than the Pyrel foam
seen for the Spacelab tool holder.
After 6 years of use the Pyrel has
completely degraded. When
compressed by slight handling
during this inspection, it
pulverises to the powder seen in
the lower photograph. Pyrel foam
(non-flammable, no off-gassing
and non-toxic) was used for
sound insulation and packaging in
toolboxes, work boxes and
Spacelab drawers
8.7 Manned Compartments 553

Fig. 8.53 The switches in the left photograph are inadequately protected against accidental contact by an astronaut working under micro-gravity
conditions (see also Fig. 3.1). Covers are necessary, on the right-hand rack location protection is given by bars that protrude above the switches
themselves

Fig. 8.54 General views of de-bonded Velcro tape. These defects were frequently seen. They result from an inadequate amount of adhesivea
continuous lm might be preferable
554 8 Assessment of Post-flight Materials

Fig. 8.55 a This post-flight

hold-down button for the external
multilayer insulation is severely
attacked by atomic oxygen, the
white bands show location of
lacing tape (masking the buttons
surface). b SEM image at 7500
showing the typical cone-shaped
features of ATOX attack on a
polymer, some inorganic
additives are left as white
non-conducting residues on the
surface
8.7 Manned Compartments 555

Fig. 8.56 Another hold down button from Spacelab; here a micrometeoroid hit is displayed. This remarkably small impact crater has a diameter of
about 180 m
 Appendix 1: Coefcient of (Linear) Thermal
Expansion for Selected Materials (COE or CTE)

Coefcient of (linear) thermal expansion, , for selected (continued)

materials (COE or CTE) (units are 106 C1 (i.e. ppm/C)) Indiumlead 33.0
Lead (95 %) tin solder 28.0
Tinlead solder 60/40 25.0
A. Pure metals
Magnesium, AZ31B 26.0
Aluminium 25
Ni-clad Molybdenum 56
Chromium 6
Steel, 1020 12.0
Cobalt 12
Stainless steel (18-8) 17.0
Copper 17
Tungsten/copper (90/10) 6.5
Gold 14
Aluminium MMC with SiC particles 614
Iron 12 (8050 % reinforcement)
Lead 29 C. Insulators and substrate materials (for electronic systems)a
Magnesium 25 E glass 5.5
Molybdenum 5 S glass 2.6
Nickel 13 Glassceramic >3.0
Platinum 9 Silicon 2.6
Silver 19 Diamond 0.9
Tantalum 7 Aluminium nitride 4.5
Tin 20 Silicon nitride 3.7
Titanium 9 Quartz, fused silica 0.5
Tungsten 5 Kevlar 49 5
Zinc 35 Beryllia 69
B. Alloys and MMCs Cubic boron nitride
Alloy 42 4.4 xy 3.7
Aluminium (40 % silicon) 13.5 z 7.2
Aluminium, AA 6061 23.6 E glass/epoxy
Aluminium, AA 3003 23.2 xy 1417
Aluminium, AA 2017 22.9 z 80280
Boron aluminium (20 %) 12.7 E glass/polyimide
Brass 18.0 xy 1216
Copper/invar/copper 20/60/20 thick 5.8 z 4080
Copper/molybdenum/copper 20/60/20 7.0 E glass/PTFE
thick
xy 24
Graphite/aluminium 46
z 260
Invar 36 1.6
Kevlar/epoxy
Invar 42 4.5
xy 57
Inconel 600 13.0
z 70
Kovar (FeNiCo) 5.0
Kevlar/polyimide
(continued)
(continued)

Springer International Publishing Switzerland 2016 557

B.D. Dunn, Materials and Processes, Springer Praxis Books,
DOI 10.1007/978-3-319-23362-8
558 Appendix 1: Coefficient of (Linear) Thermal Expansion for Selected Materials (COE or CTE)

(continued) (continued)
xy 3.46.7 Si3N4 (-phase) 2.9
z 83 Si3N4 (-phase) 2.3
Quartz/polyimide Spinel (MgAl2O4) 7.6
xy 58 Sodalimesilicate glass 9.2 (used in lightbulbs)
z 68.4 Borosilicate glass 4.6 (used with Kovar)
Quartz/bismaleimide Silica (96 % pure) 0.8
xy, 35 % resin 6.2 Silica (99.9 % pure) 0.55
z 35 % resin 41 Zerodur Class 2 0.1
Alumina (90 %) TF substrate 7.0 Zerdur Class 0 Extreme 0.007
Alumina (ceramic chip carrier) 5.97.4 E. Polymers (unorientated)
Epoxy (70 % silica) plastic packaging 2023 Polyethylene 100200
Mulite co-red 4.2 Polypropylene 58100
Gallium arsenide 5.7 Polystyrene 6080
Silicon carbide 3.6 Polytetrauoroethylene 100
Carbon bre 60 %epoxy 1.1 Polycarbonate 66
D. Other ceramics Nylon (6/6) 80
A12O3 6.58.8 Cellulose acetate 80160
BeO 9 Polymethylmethacrylate 5090
MgO 13.5 Epoxy 4590
SiC 4.8 Phenolformaldehyde 6080
Silicon 2.6 Silicones 2040
(continued) a
For temperature range 55 to +100 C
 Appendix 2: Properties of Printed Circuit Laminates

Material Thermal Mechanical

Conductivity CTE CTE Max. use Glass Tensile Yield Elongation
W/M-K X, Y Dir. Z. Dir Temperature transition strength strength %
ppm/C ppm// C Temperature C MPa MPa
C
Polymer composites
Polyimide glass 0.35 1216 4060 215280 250260 345
Epoxy glassa 0.160.2 1418 180 130160 125135 276
Modied epoxyb 1416 140150
PTFEe glass, 0.10.26 20 230260
non-woven
PTFEe glass, 419837 1025 248 3852
woven
Epoxy aramid 0.12 68 66 125 68103
Epoxy quartz 613 62 125
Polyimide aramid 0.28 58 83 250
Polyimide quartz 0.35 612 35 188250 207
Epoxycordierite 0.91.3 3.33.8
Modied epoxy 5.55.6 100 137
aramid
PTFEe quartz 7.59.4 88 19d
Polyimide 4.311.8 4550 260315 67
Metal composites
Cu/Invar/Cu 1518c 5.35.5 16 N/A 310414 170270 36
(20/60/20)
Cu/Invar/Cu 14c 4.4 N/A 380480 240340
(12.5/75/12.5)
Cu/Mo/Cu 90174 2.6 N/A
Ni/Mo/Ni 129.8c 5.26 5.26 N/A 621 552 50
Published with permission from the IPC, 2215 Sanders Road, Northbrook, Illinois, USA. (Table from IPC-D-279 Design Guidelines for Reliable
Surface Mount Technology Printed Circuit Board Assemblies, July 1996)
a
FR-4, G-10
b
Polyfunctional FR-4
c
Z-direction
d
Polymorphic p
e
PTFE=Polytetrauoroethylene

Springer International Publishing Switzerland 2016 559

B.D. Dunn, Materials and Processes, Springer Praxis Books,
DOI 10.1007/978-3-319-23362-8
 Appendix 3: Reagents for Microetching
Metals and Alloys
A wide variety of techniques may be used for the identi-
cation of grain structures, phases, and other constituents in
metals and alloys. Metallographers are often able to predict
the chemical composition and processing history of a
metallic sample by selectively etching its polished surface
and comparing the microstructure to those of reference
samples in conjunction with published phase diagrams.
The following chemical reagents (etch compositions) are
recommended by the author for the etching of metals and
alloys commonly encountered during the metallurgical
assessment of electrical and structural spacecraft materials.
Springer International Publishing Switzerland 2016
B.D. Dunn, Materials and Processes, Springer Praxis Books,
DOI 10.1007/978-3-319-23362-8
Each chemical must be stored and handled according to the
manufacturers recommendations. All chemicals are poten-
tially dangerous and it is assumed that the person mixing,
pouring, or etching is thoroughly familiar with their use. If
there is any uncertainty about their use, toxicity, or means of
disposal, the users Chemical and Safety Department should
be contacted.
The concentrations of acids are given in terms of specic
gravity (s.g.), or as a percentage (%) of the fully concen-
trated value.
561
No. Reagent composition Remarks
562

Metal
Aluminium and its alloys
1 Hydrouoric acid (40 %) 0.5 ml 15 s immersion is recommended. Particles of all common microconstituents are outlined. Colour indications
Hydrochloric acid (1.19) 1.5 ml
Nitric acid (1.4) 2.5 ml
Water 95.5 ml Mg2Si and CaSi2 Blue to brown
(AlFeSi) and (AlFeMn) Darkened
(AlCuFe) Light brown
(Kellers etch) MgZn2, NiAl3, (AlCuFeMn), Al2Cu Mg and Al6CuMg Brown to black
(AlCuFe) and (AlCuMn) Blackened
Al3Mg2 Heavily outlined and pitted
The colours of other constituents are little altered. Not good for high Si alloys
Desmut in 50 % nitric acid if necessary
2 Sodium hydroxide 1g Specimens are etched by swabbing for 10 s. All usual constituents are heavily outlined, except for Al3Mg2
Water 99 ml (which may be lightly outlined) and (AlCrFe) which is both unattacked and uncoloured. Colour indications
FeAl3 and NiAl3 Slightly darkened
(AlCuMg) Light brown
(AlFeSi) Dull brown
(AlFeSi) Rough and attacked; slightly darkened
MnAl6 and (AlFeMn) Coloured brown to blue (uneven attack)
MnAl4 Tends to be darkened
The colours of other constituents are only slightly altered
Beryllium
3 Hydrouoric acid (40 %) 10 ml Etch by immersion for 1030 s to outline grain boundaries and microconstituents
Ethyl alcohol 90 ml
4 Water 95 ml Be alloys may be etched in the reagent (115 s)
Sulphuric acid (1.84) 5 ml
Note: during the preparation of beryllium samples, do not breathe dust, as this is extremely toxic. Cutting operations must be done under controlled conditions, preferably in a glove box. See text for the removal of
mechanical twins. Polishing cloth will be contaminated with beryllia and needs to be disposed of according to local health and safety requirements. Metallographers should wear rubber gloves and avoid contact with
etchants
Chromium
5 Hydrochloric acid (concentrated) Shows striations in electrodeposits
Copper, copper alloys, brass, bronze, etc.
6 Ammonium hydroxide 50 ml Used for copper, many copper-rich alloys
Water 50 ml Gives a grain boundary etch, and also tends to darken the solid solution, leaving the solid solution lighter. The
Hydrogen peroxide (30 vol.) 20 ml hydrogen peroxide content may be varied. Less is required the lower the copper content
7 Ferric chloride, various strengths and compositions Used as a general reagent for copper, brass, bronze, nickelsilver, aluminiumbronze, and other copper-rich alloys. It
To 100 parts of water are added darkens the constituent in brasses and gives grain contrast following ammoniacal or chromic acid etches. The most
suitable composition should be found by trial and error in specic cases. This reagent generally emphasizes scratches
Hydrochloric acid (1.19) Ferric chloride (g)
in imperfectly prepared specimens, and tends to roughen the surface. For sensitive work it is frequently a great
20 1 advantage to replace the water in the reagent by a 50:50 wateralcohol mixture or by pure alcohol
10 5
50 5
Appendix 3: Reagents for Microetching Metals and Alloys

(continued)
(continued)
No. Reagent composition Remarks
8 Potassium dichromate 2g Used for copper, and copper alloys with beryllium, manganese, and silicon. Also suitable for nickelsilver, bronzes,
Water 100 ml and chromiumcopper alloys. This reagent should be followed by a ferric chloride etch to give added contrast
Sodium chloride (saturated) 4 ml
Sulphuric acid (s.g. 1.84) 8 ml
Gold
9 Hydrochloric acid (concentrated) 60 ml Use under a hood, immerse for a few seconds
Nitric acid (concentrated) (Aqua Regia) 40 ml
10 Potassium cyanide, 10 % in water 10 ml Used for gold and its alloys. A fresh solution, warmed if necessary, must be used for each operation. The etching time
Ammonium persulphate, 10 % in water 10 ml varies from 0.5 to 3 min. The attack may be speeded up by the addition of 2 % of potassium iodide, but this is liable
to give staining effects
11 Tincture of iodine, 50 % solution in aqueous potassium iodide Used for gold alloys. With silver-gold alloys a silver iodide lm may form. This may be removed by immersion in
potassium cyanide solution
Indium and indium alloys
12 Hydrochloric acid (1.19) 20 ml General etchant, use by immersion for a few seconds. [For very detailed studies of indiumgold reaction layers, see
Picric acid 4g Millares and Peraggi (1992)]
Ethyl alcohol 400 ml
Appendix 3: Reagents for Microetching Metals and Alloys

Iron and steel

13 Nitric acid (1.40) 1.55 ml to 100 ml Ferrite g.b.s in low-carbon steels. Darkens pearlite and gives contrast with ferrite or cementite network. Etches
Ethanol (Nital etch) martensite and its decomposition products in many steels. Better than Picral for low-alloy steels and for ferritic grain
boundaries 530 s depending on steel
14 Picric acid 1g Attacks prior austenite boundaries
Hydrochloric acid (Vilellas reagent) 5 ml Good for ferritic steels
15 Nitric acid (1.40) 10 ml Immerse up to 30 s. Good for high-chromium steels, austenitic stainless steel, etc.
Hydrochloric acid (1.19) 20 ml Do not keep, discard when yellow, if left this reagent can explode
Glycerol 30 ml
16 Ferric chloride 2g In tool steels attacks ferrite and martensite, outlines carbides, leaves austenite unattacked
Hydrochloric acid (1.19) 5 ml
Water (Kallings reagent) 30 ml Immerse 15 min
Molybdenum
17 (a) Potass, hydroxide Water 10 g to 100 ml Mix equal amounts of (a) and (b) as needed Grain boundary etch
(b) Pot. ferricyanide Water 10 g to 100 ml
18 Ammonia (0.88) 50 ml Boil for up to 10 min General etch
Hydrogen perioxide (3 %) 50 ml
Water 50 ml
Nickel and its alloys
19 Nitric acid (1.40) 10 ml Pure nickel, and nickelchromium alloys. Grain boundaries etched
Hydrochloric acid (1.19) 20 ml
Glycerol 30 ml
20 Nitric acid (1.40) 10 ml Used for pure nickel, cupro-nickel, Monel metal, and nickelsilver
Acetic acid 10 ml
Acetone 10 ml
(continued)
563
(continued)
564

No. Reagent composition Remarks

21 Hydrochloric acid (1.19) 300 ml Stand in fume cupboard for 24 h; use diluted 50:50 with water. Good for Nilo K etc. Can be kept as stock solution
Nitric acid (1.40) 100 ml
Ferric chloride 25 g
Cupric chloride 25 g
(Green special etch, also known as Pinders etch)
Platinum group of metals (e.g. Pt, Pd, Rh, Ru, Ir, Os)
Use Aqua Regiasee Etchant No. 9 (may need to be warmed) Grain contrasting
22 Potasssium ferricyanide 3.5 g Several minutes immersion Most alloys, for general etching of grain boundaries
Sodium hydroxide 1g
Water 150 ml
Silver
23 Ammonium hydroxide 50 ml Recommended for silver, silvernickel, and silverpalladium alloys. Also useful for the examination of silver-
Hydrogen peroxide (3 %) 1030 ml soldered joints
24 Sulphuric acid (10 % in water) to which a few crystals of chromic acid CrO3 This reagent reveals the grain structure of silver and silver rich-alloys
have been added (2 g)
Tin and its alloys
25 Nitric acid (1.40) 2 % in alcohol Micro etch with high contrast, which blackens lead and darkens tin to light brown after prolonged immersion
26 Silver nitrate 5g Micro etch recommended for lead-rich alloys. Darkens primary and eutectic lead and produces a very high grain
Water 100 ml contrast
Titanium and its alloys
27 Hydrouoric acid (40 %) 13 ml 310 s Most useful general etch, especially for Ti6A14V alloy
Plastic mounts must be thoroughly washed to remove all hydrouoric acid as redisudal acid will etch and damage
glass microscope lens
Nitric acid (1.40) 26 ml
Water (Krolls reagent) to 100 ml
28 Potassium hydroxide (40 %) 10 ml 320 s Useful for / alloys, is attacked or stained. unattacked
Hydrogen peroxide (30 %) 5 ml
Water (can be varied to suit alloy) 20 ml
Marinol blue50 % Benzalconium chloride 1015 ml Stain etch to show alpha-stabilized layer
solution
29 Glycerol 40 ml The specimen must be dry, and swab etching gives best control; time of etching varies; continue until specimen turns
Methylated spirit 40 ml light-brown
Hydrouoric acid (40%) 510 ml
Tungsten
30 Sodium hydroxide, 10% in water 10 ml This reagent is used cold and, on immersion of the specimen for approximately 10 s, develops grain boundaries
Potassium ferricyanide, 10% in water 10 ml (Murakamis reagent)
31 Hydrogen peroxide, 3 % in water This reagent develops grain boundaries, but only after some 3090 s in the boiling reagent
Appendix 3: Reagents for Microetching Metals and Alloys
 Appendix 4: Conversion Table for Mechanical
Properties

Conversion table for mechanical properties (N/mm2(MPa) to hbar, (continued)

tonf/in2, lbf/in2and kgf/mm2)
N/mm2 hbar tonf/in.2 lbf/in.2 kgf/mm2
N/mm2 hbar tonf/in.2 lbf/in.2 kgf/mm2 180 18 11.7 26,100 18.4
5 0.5 0.3 700 0.5 185 18.5 12.0 26,800 18.9
10 1 0.6 1500 1.0 190 19 12.3 27,600 19.4
15 1.5 1.0 2200 1.5 195 19.5 12.6 28,300 19.9
20 2 1.3 2900 2.0 200 20 12.9 29,000 20.4
25 2.5 1.6 3600 2.5 205 20.5 13.3 29,700 20.9
30 3 1.9 4400 3.1 210 21 13.6 30,500 21.4
35 3.5 2.3 5100 3.6 215 21.5 13.9 31,200 21.9
40 4 2.6 5800 4.1 220 22 14.2 31,900 22.4
45 4.5 2.9 6500 4.6 225 22.5 14.6 32,600 22.9
50 5 3.2 7300 5.1 230 23 14.9 33,400 23.5
55 5.5 3.6 8000 5.6 235 23.5 15.2 34,100 24.0
60 6 3.9 8700 6.1 240 24 15.5 34,800 24.5
65 6.5 4.2 9400 6.6 245 24.5 15.9 35,500 25.0
70 7 4.5 10,200 7.1 250 25 16.2 36,300 25.5
75 7.5 4.9 10,900 7.6 255 25.5 16.5 37,000 26.0
80 8 5.2 11,600 8.2 260 26 16.8 37,700 26.5
85 8.5 5.5 12,300 8.7 265 26.5 17.2 38,400 27.0
90 9 5.8 13,100 9.2 270 27 17.5 39,200 27.5
95 9.5 6.2 13,800 9.7 275 27.5 17.8 39,900 28.0
100 10 6.5 14,500 10.2 280 28 18.1 40,600 28.6
105 10.5 6.8 15,200 10.7 285 28.5 18.5 41,300 29.1
110 11 7.1 16,000 11.2 290 29 18.8 42,100 29.6
115 11.5 7.4 16,700 11.7 295 29.5 19.1 42,800 30.1
120 12 7.8 17,400 12.2 300 30 19.4 43,500 30.6
125 12.5 8.1 18,100 12.7 305 30.5 19.7 44,200 31.1
130 13 8.4 18,900 13.3 310 31 20.1 45,000 31.6
135 13.5 8.7 19,600 13.8 315 31.5 20.4 45,700 32.1
140 14 9.1 20,300 14.3 320 32 20.7 46,400 32.6
145 14.5 9.4 21,000 14.8 325 32.5 21.0 47,100 33.1
150 15 9.7 21,800 15.3 330 33 21.4 47,900 33.7
155 15.5 10.0 22,500 15.8 335 33.5 21.7 48,600 34.2
160 16 10.4 23,200 16.3 340 34 22.0 49,300 34.7
165 16.5 10.7 23,900 16.8 345 34.5 22.3 50,000 35.2
170 17 11.0 24,700 17.3 350 35 22.7 50,800 35.7
175 17.5 11.3 25,400 17.8 355 35.5 23.0 51,500 36.2
(continued) (continued)

Springer International Publishing Switzerland 2016 565

B.D. Dunn, Materials and Processes, Springer Praxis Books,
DOI 10.1007/978-3-319-23362-8
566 Appendix 4: Conversion Table for Mechanical Properties

(continued) (continued)
2 2 2 2
N/mm hbar tonf/in. lbf/in. kgf/mm N/mm2 hbar tonf/in.2 lbf/in.2 kgf/mm2
360 36 23.3 52,200 36.7 485 48.5 31.4 70,300 49.5
365 36.5 23.6 52,900 37.2 490 49 31.7 71,100 50.0
370 37 24.0 53,700 37.7 495 49.5 32.1 71,800 50.5
375 37.5 24.3 54,400 38.2 500 50 32.4 72,500 51.0
380 38 24.6 55,100 38.7 505 50.5 32.7 73,200 51.5
385 38.5 24.9 55,800 39.3 510 51 33.0 74,000 52.0
390 39 25.3 56,600 39.8 515 51.5 33.3 74,700 52.5
395 39.5 25.6 57,300 40.3 520 52 33.7 75,400 53.0
400 40 25.9 58,000 40.8 525 52.5 34.0 76,100 53.5
405 40.5 26.2 58,700 41.3 530 53 34.3 76,900 54.0
410 41 26.5 59,500 41.8 535 53.5 34.6 77,600 54.6
415 41.5 26.9 60,200 42.3 540 54 35.0 78,300 55.1
420 42 27.2 60,900 42.8 545 54.5 35.3 79,000 55.6
425 42.5 27.5 61,600 43.3 550 55 35.6 79,800 56.1
430 43 27.8 62,400 43.8 555 55.5 35.9 80,500 56.6
435 43.5 28.2 63,100 44.4 560 56 36.3 81,200 57.1
440 44 28.5 63,800 44.9 565 56.5 36.6 81,900 57.6
445 44.5 28.8 64,500 45.4 570 57 36.9 82,700 58.1
450 45 29.1 65,300 45.9 575 57.5 37.2 83,400 58.6
455 45.5 29.5 66,000 46.4 580 58 37.6 84,100 59.1
460 46 29.8 66,700 46.9 585 58.5 37.9 84,800 59.7
465 46.5 30.1 67,400 47.4 590 59 38.2 85,600 60.2
470 47 30.4 68,200 47.9 595 59.5 38.5 86,300 60.7
475 47.5 30.8 68,900 48.4 600 60 38.8 87,000 61.2
480 48 31.1 69,600 48.9
(continued)
 Appendix 5: Aluminium Alloy Temper Designations

The compositions and temper conditions of aluminium 9 the minimum ultimate tensile strength exceeds that
alloys are designated by the Aluminium Association (AA) of the fully hard by at least 10 MPa.
and are recognized internationally. Further information can The second digit indicates the degree of hardening and
be obtained from the AA. is a number from 1 to 9.
The Basic Temper Designations are as follows: (c) Third digit
A third digit may be used to denote a further charac-
F As fabricated
teristic or variation.
O Annealed
H Strained hardened
Spacecraft aluminium alloys are generally subdivisions
T Thermally treated to produce stable tempers other
of the T temper, as shown below
than F, O, H
A more precise description of the T tempers is as follows,
but whenever possible the original material specication
The Subdivisions of H temper are as follows: should be consulted, as some deviations exist:
T1 Cooled from an elevated temperature-shaping
(a) First digit
process and naturally aged to a substantially
H1 strain hardened only
stable condition
H2 strain hardened and partially annealed
T2 Cooled from an elevated temperature-shaping
H3 strain hardened and stabilized
process, cold-worked, and naturally aged to a
(b) Second digit
substantially stable condition
1 1/8 hard
T3 Solution heat-treated, cold-worked, and natu-
2 quarter hard
rally aged to a substantially stable condition
3 3/8 hard
T351 Solution heat-treated and stress-relieved by
4 half hard
stretching to produce a permanent set of 2 %
5 5/8 hard
nominal but not less than 1.5 % nor more than
6 three quarters hard
3 %. Product shall receive no further straight-
7 7/8 hard
ening after stretching
8 fully hard

Springer International Publishing Switzerland 2016 567

B.D. Dunn, Materials and Processes, Springer Praxis Books,
DOI 10.1007/978-3-319-23362-8
568
T3510 Solution heat-treated and stress-relieved by
stretching to produce a nominal permanent set
of 1.5 %, but not less than 1 % nor more than
3 %. Extrusions shall receive no straightening
after stretching
T3511 Solution heat-treated and stress-relieved by
stretching to produce a nominal permanent set
of 11.5 %, but not less than 1 % nor more than
3 %. Extrusions may receive minor straighten-
ing, after stretching, of an amount necessary to
meet the tolerances
T352 Solution heat-treated and stress-relieved by
compression to produce a permanent set of 1.5
5 %. During compression, primary focus shall
be applied in the axial direction
T36 Solution heat-treated and cold-worked by
reduction of approximately 6 %
T361 Solution heat-treated and cold-reduced approx-
imately 6 % in thickness
T4 Solution heat-treated and naturally aged to a
substantially stable condition
T42 Material purchased in any temper and subse-
quently solution heat-treated and naturally aged
to a substantially stable condition by the user
Appendix 5: Aluminium Alloy Temper Designations
T451 Rolled or cold-nished, stress-relieved by
stretching to produce a nominal permanent set of
1.5 % but not less than 1 % nor more than 3 %
Product shall receive no further operations after
stretching unless specically authorized by
purchaser
T4510 Solution heat-treated and stress-relieved by
stretching to produce a permanent set of 1.5 %
nominal, but not less than 1 % nor more than
3 %. Material shall receive no further straight-
ening after stretching
T4511 Solution heat-treated and stress-relieved by
stretching to produce a permanent set of 1.5 %
nominal, but not less than 1 % nor more than
3 %. Material may receive minor straightening
after stretching
T5 Cooled from an elevated temperature-shaping
process and then articially aged (wrought
products). Stress-relieved (castings)
T51 Precipitation heat-treated (castings)
T6 Solution and precipitation heat-treated
T61 Solution heat-treated and precipitation
heat-treated. Quenching from the solution tem-
perature shall be into water at 8085 C
Appendix 5: Aluminium Alloy Temper Designations
T611 Solution and precipitation heat-treated, low
residual stresses
Material may, after quenching from the solution
heat-treatment temperature, receive minor
straightening in an amount necessary to meet
tolerances specied on the drawing
T62 Solution heat-treated and then articially aged
by the user
T651 Solution heat-treated and stress-relieved by
stretching to produce a permanent set of 2 %
nominal but not less than 1.5 % nor more than
3 %, and articially aged. Product shall receive
no further straightening after stretching
T6510 Solution heat-treated and stress-relieved by
stretching to produce a permanent set of 1.5 %
nominal, but not less than 1 % nor more than
3 % and articially aged. Material shall receive
no further straightening after stretching
T6511 Solution heat-treated and stress-relieved by
stretching to produce a permanent set of 1.5 %
nominal, but not less than 1 % nor more than
3 %, and articially aged. Material may receive
minor straightening after stretching
T652 Solution heat-treated and stress-relieved by
compression to produce a permanent set of 1.5
5 %, and precipitation heat-treated. During
compression, primary focus shall be applied in
the axial direction and on individual rings
approximately nal dimensions
T66 Solution and precipitation heat-treated
T7 Solution heat-treated and overaged/stabilized
T71 Solution and precipitation heat-treated (castings)
T72 Solution heat-treated and then articially over-
aged by the user
T73 See T74
T7351 Solution heat-treated and stress-relieved by
stretching to produce a nominal permanent set
of 2 % but not less than 1.5 % nor more than
3 %, and precipitation heat-treated. Plate shall
receive no further straightening operations after
stretching
T73510 Solution heat-treated and stress-relieved by
stretching to produce nominal permanent set of
1.5 % but not less than 1 % nor more than 3 %,
and precipitation heat-treated. Material shall
receive no further straightening after stretching
T7311 Solution heat-treated and stress-relieved by
stretching to produce nominal permanent set of
1.5 %, but not less than 1 % nor more than 3 %,
and precipitation heat-treated. Material may
receive minor straightening, after stretching, an
amount necessary to meet required dimensional
tolerances
569
T7352 Solution heat-treated and stress-relieved by
compression to produce a permanent set of not
less than 1 % nor more than 5 %, and
precipitation heat-treated. The method and
direction of compression shall be as agreed upon
by purchaser and vendor
T736 See T74
T73651 Solution heat-treated and stress-relieved by
stretching to produce a nominal set of 2 % but
no less than 1.5 % nor more than 3 %, and
precipitation heat-treated
T736511 Solution heat-treated and stress-relieved by
stretching to produce a nominal permanent set
of 1.5 % but not less than 1 % nor more than
3 %, and precipitation heat-treated. Material
may receive minor straightening after stretching,
to meet required dimensional tolerances
T73652 Solution heat-treated and stress-relieved by
compression to produce a permanent set of not
less than 1 % nor more than 5 %, and
precipitation heat-treated. The method and
direction of compression shall be as agreed upon
by purchaser and vendor
T74 (Previously T73 or T736.) Solution heat-treated
and articially aged to resist stress-corrosion
cracking
T7451 Solution heat-treated and stress-relieved by
stretching to produce a nominal permanent set
of 2 % but not less than 1.5 % nor more than
3 %, and precipitation heat-treated. Plate shall
receive no further straightening operations after
stretching
T7452 Solution heat-treated and stress-relieved by
compressing to produce a permanent set of 1
5 % and overaged
T76 Solution heat-treated and articially aged suf-
cient to produce improved resistance to
exfoliation
T761 Solution heat-treated and precipitation
heat-treated. The T7 tempers require closer
control on ageing practice variables such as
time, temperature, heating-up rates, etc
T7651 Solution heat-treated and stress-relieved by
stretching to produce a permanent set of 2 %
nominal but not less than 1.5 % nor more than
3 %, and articially aged sufcient to produce
improved resistance to exfoliation and
stress-corrosion cracking. Plate shall receive no
further straightening after stretching
T76511 Solution heat-treated and stress-relieved by
stretching to produce a nominal permanent set
of 1.5 %, but not less than 1 % nor more than
3 %, and precipitation treated. Material may
570
receive minor straightening, after stretching, of
an amount necessary to meet required dimen-
sional tolerances
T77 Solution-treated and stabilized
T8 Solution heat-treated, cold-worked, and then
articially aged
T81 Solution heat-treated, cold-worked by the flat-
tening operation, and then articially aged
T851 Solution heat-treated and stress-relieved by
stretching to produce a permanent set of 2 %
nominal but not less than 1.5 % nor more than
3 %, and articially aged. Plate shall receive no
further straightening after stretching
T8511 Solution heat-treated and stress-relieved by
stretching to produce a nominal permanent set
Appendix 5: Aluminium Alloy Temper Designations
of 1.5 % but not less than 1 % nor more than
3 %, and precipitation heat-treated
T852 Solution heat-treated and stress-relieved by
compression, to produce a permanent set of 1
5 %, and precipitation heat-treated
T86 Solution heat-treated, cold-worked by a reduc-
tion of approximately 6 %, and then articially
aged
T861 Solution heat-treated, cold-reduced approxi-
mately 6 % in thickness, and precipitation
heat-treated
T9 Solution heat-treated, articially aged, and then
cold-worked
T10 Cooled from an elevated temperature-shaping
process, cold-worked, and then articially aged
 Appendix 6: Metal Alloy Comparison Tables

Introduction It is for the user to decide from the composition whether

the related specications are sufciently similar to permit the
Marshall Space Flight Center document MSFC-SPEC-522A, British, French or German alloy to be regarded as a suitable
entitled Design criteria for controlling stress corrosion substitute for the American alloy. In all critical applications
cracking, contains a list of alloys. Each alloy bears a ve- the individual alloy specications must be consulted as
digit classication number, which is made up as follows: precise compositions, tolerance in composition, the present
The rst digit denotes the class: of trace elements, etc. can be important. No attempt has been
made to relate materials on the basis of form, but reference to
1. Steels the specication numbers will give guidance in this matter.
2. Nickel alloys In the case of the steels and the aluminium alloys, the list
3. Aluminium alloys must be used in conjunction with the notes appended at the
4. Copper alloys end of the relevant section.
5. Titanium alloys Following the list, there is an index that enables the ve-
6. Magnesium alloys digit table numbers to be related to the unied numbering
7. Miscellaneous system (UNS) of the Society of Automotive Engineers
(SAE) and the American Society for Testing Materials
The second digit denotes the subclass. (ASTM). The UNS numbers are prexed by letter symbols
The last three digits are a serial number within the that have the following meaning:
subclass.
A Aluminium and aluminium alloys
The list contained in the following pages is based on the
C Copper and copper alloys
list described above, which forms as it were the framework.
G AISI and SAE carbon steels
Into this framework have been intercalated those British,
J Cast steels (except tool steels)
French and German alloy specications that most closely
K Miscellaneous nonferrous metals and alloys
correspond to their American counterparts. If an American
N Nickel and nickel alloys
specication is not followed by such a European specica-
R Reactive and refractory metals and alloys
tion, that means that diligent searching has failed to reveal an
S Heat- and corrosion-resistant (stainless) steels
equivalent.
T Tool steels, wrought and cast

Springer International Publishing Switzerland 2016 571

B.D. Dunn, Materials and Processes, Springer Praxis Books,
DOI 10.1007/978-3-319-23362-8
572 Appendix 6: Metal Alloy Comparison Tables

Alloy equivalentssteels
Composition
Country Designation C Si Mn P S Other
Carbon steels
11001 USA AISI/SAE 1005 (UNS G10050) <.06 *1 <.35 <.04 <.05
UK BS970 015A03 <.06 *1 <.3 <.05 <.05
F AFNOR FD5 .04.07 .2.4 .02 .025
F AFNOR FD4 .04.07 <.1 .2.4 .025 .03
G DIN 17140 D6-2 Wk. 1.0314 <.06 *2 <.4 <.04 <.04 N < .007 *3
G DIN 17140 D5-1 Wk. 1.0312 <.06 *2 <.4 <.05 <.05
11002 USA AISI/SAE 1006 (UNS G10060) <.08 *1 .25.4 <.04 <.05
UK BS970 030A04 <.08 *1 .2.4 <.05 <.05
G DIN 17140 D7-1 Wk. 1.0311 <.08 *2 <.45 <.06 <.05
G DIN 17140 D8-2 Wk. 1.0313 <.08 *2 <.45 <.045 <.04 N < .007 *3
11003 USA AISI/SAE 1008 (UNS G10080) <.1 *1 .3.5 <.04 <.05
UK BS970 040A04 <.08 *1 .3.5 <.05 <.05
F AFNOR FdTu4 <.09 .25.5 N < .006
F AFNOR FdTu2 <.08 .35.6 N < .014 Mn/S > 10
F AFNOR FdTu10 <.1 .25.5 N < .01
F AFNOR FdTu11 <.1 <.1 .25.5
F AFNOR Fd2 .04.1 .2.45 N < .007
F AFNOR Fd12 .04.1 .2.5 N < .007
G Ust4, US14 Wk. 1.0336 <.09 *2 .25.5 <.03 <.03 N < .007
G DIN 1623; 1624; 5512; st2, st12 Wk. <.1 *2 .2.45 <.033 <.035 N < .007 *3
1.0330
G DIN 1623. B11; 001624; st3, st13 Wk. <.1 *2 .2.4 <.025 <.023 N < .007
1.0333
G DIN 17115; Ust35-2; Wk. 1.0207 .06.14 *2 .4.6 <.04 <.04 N < .007 *3Also
AISI 1010
G DIN 17115; Rst35-2; Wk. 1.0208 .06.12 .0325 .4.6 <.04 <.04 Also AISI 1010
G DIN 17111; UQst 36-2; Wk. 1.0204 .08.13 *2 1.25.45 <.04 <.04 N < .007 *3Also
AISI 1010
G DIN 17111; Rst 36-2; Wk. 1.0205 <.13 <.4 25.5 <.05 <.5 N < .007 *3-Also AISI
1010
11004 USA AISI/SAE 1010 (UNS G10100) .08.13 *1 .3.6 <.04 <.05
UK BS970 040A10 .08.13 *1 .3.5 <.05 <.05
UK BS970 045A10 .08.13 *1 .3.6 <.05 <.05
F AFNOR Xc9 .06.12 .05.3 .3.5
G DIN 17210; 1652 Ck10 Wk. 1.1121 .07.13 .15.35 .3.6 <.035 <.035
G DIN 17210; 1652 c9 Wk. 1.0301 .07.13 .15.33 .3.6 <.045 <.043
11005 USA AISI/SAE 1011 (UNS G10110) .08.13 *1 .6.9 <.04 <.05
UK BS970 060A10 .08.13 *1 .5.7 <.05 <.05
11006 USA AISI/SAE 1012 (UNS G10120) .1.15 .3.6 <.04 <.05
UK BS970 040A12 .1.15 *1 .3.5 <.05 <.05
F AFNOR XC 12 .1.16 .05.3 .3.5
G DIN 17210; 1652 Ck10Wk. 1.1121 .07.13 .15.35 .3.6 <.035 <.035
G DIN 17210; 1652 c9 Wk. 1.0301 .07.13 .15.35 .3.6 <.045 <.045
11007 USA AISI/SAE 1015 (UNS G10150) .13.18 *1 .3.6 <.04 <.05
UK BS970 040A15 .13.18 *1 .3.5 <.05 <.05
UK BS970 050A15 .13.18 *1 .4.6 <.05 <.05
F AFNOR XC12 .1.16 .05.3 .3.5
G DIN 17210; 1652 Ck15 Wk. 1.1141 .12.18 .15.35 .3.6 <.035 <.035
11008 USA AISI/SAE 1016 (UNS G10160) .1318 *1 .6.9 <.04 <.05
UK BS970 080A15 .13.18 *1 .7.9 <.05 <.05
G DIN 17111; Rst 44.2 Wk. 1.0419 <.18 <.45 <.8 <.05 <.05 N < .007
11009 USA AISI/SAE 1017 (UNS G10170) .15.2 *1 .3.6 <.04 <.05
UK BS970 040A17 .15.2 *1 .3.5 <.05 <.05
F AFNOR XC18 .16.22 <.25 .4.65
G DIN 17210; 1652 Ck15 Wk. 1.1141 .12.18 .15.35 .3.6 <.035 <.035
(continued)
Appendix 6: Metal Alloy Comparison Tables 573

(continued)
Composition
Country Designation C Si Mn P S Other
11010 USA AISI/SAE 1018 (UNS G10180) .15.2 *1 .6.9 <.04 <.05
UK BS970 080A17 .15.2 *1 .7.9 <.05 <.05
G DIN 17172; st43.7 Wk. 1.0484 <.22 <.4 .51.1 <.04 <.045
11011 USA AISI/SAE 1019 (UNS G10190) .15.2 *1 .71.0 <.04 <.05
UK BS970 080A17 .15.2 *1 .7.9 <.05 <.05
G DIN 17172; St43.7 VVk.1.0484 <.22 <.4 .51.1 <.04 <.045
11012 USA AISI/SAE 1020 (UNS G10200) .18.23 *1 3.6 <.04 <.05
UK BS970 040A20 .18.23 *1 .3.5 <.05 <.05
F AFNOR CC20 .15.25 .1.4 .4.7 <.04 <.04
G DIN 17200; 17242; 17243; 1652. C22 .18 .25 .15.35 .3.6 <.045 <.045
Wk. 1.0402
11013 USA AISI/SAE 1021 (UNS G10210) .18.23 *1 .6.9 <.04 <.05
UK BS970 080A20 .18.23 *1 .7.9 <.05 <.05
G DIN 17172; St47.7 Wk. 1.0409 <.22 .2.45 .71.3 <.04 <.035 N < .009
11014 USA AISI/SAE 1022 (UNS G10220) .18.23 *1 .71.0 <.04 <.05
UK BS970 080A20 .18.23 *1 .7.9 <.05 <.05
G DIN 17172; St47.7 Wk. 1.0409 <.22 .2.45 .71.3 <.04 <.035
11015 USA AISI/SAE 1023 (UNS G10230) .2.25 *1 .3.6 <.04 <.05
UK BS970 040A22 .2.25 *1 .3.5 <.05 <.05
F AFNOR XC18S .15.22 <.25 .4.65 <.04 <.035
G DIN 17200; 1652. Ck22 Wk. 1.1151 .18.25 .15.35 .36 <.035 <.035 Cr < .5
11016 USA AISI/SAE 1025 (UNS G10250) .22.28 *1 .3.6 <.04 <.05
UK BS970 060A25 .23.28 *1 .5.7 <.05 <.05
F AFNOR XC25 .23.29 .1.4 .4.7
G Ck25 Wk. 1.1158 .22.29 .15.4 .4.7 <.035 <.035
11017 USA AISI/SAE 1026 (UNS G10260) .22.28 *1 .6.9 <.04 <.05
UK BS970 080A25 .23.28 *1 .7.9 <.05 <.05
11018 USA AISI/SAE 1029 (UNS G10290) .25.31 *1 .6.9 <.04 <.05
UK BS970 080A27 .25.30 *1 .7.9 <.05 <.05
11019 USA AISI/SAE 1030 (UNS G10300) .28.34 *1 .6.9 <.04 <.05
UK BS970 080A30 .28.33 *1 .7.6 <.05 <.05
11020 USA AISI/SAE 1035 (UNS G10350) .32.38 *1 .6.9 <.04 <.05
UK BS970 080A35 .33.38 *1 .7.9 <.05 <.05
F AFNOR XC32 .3.35 .1.4 .5.8
G DIN 17200; 17240; 0017242. Ck35 .32.39 .15.35 .5.8 <.035 <.035
Wk. 1.1181
11021 USA AISI/SAE 1037 (UNS G10370) .32.38 *1 .71.0 <.04 <.05
UK BS970 080A35 .33.38 *1 .7.9 <.05 <.05
F AFNOR XC35 .32.38 .1.4 .5.8 <.04 <.035
G DIN 17200; 17240; 0017242 Ck35 .32.39 .15.35 .5.8 <.035 <.035
Wk. 1.1181
11022 USA AISI/SAE 1038 (UNS G10380) .35.42 *1 .6.9 <.04 <.05
UK BS970 080A37 .35.4 *1 .7.9 <.05 <.05
F AFNOR XC38 .35.4 .1.4 .5.8 <.035 <.035
G Ck38 Wk. 1.1176 .35.4 .35.5 .5.7 <.035 <.035 N < .007, *3
11023 USA AISI/SAE 1039 (UNS G10390) .37.44 *1 .71.0 <.04 <.05
UK BS970 080A40 .38.43 *1 .7.9 <.05 <.05
F AFNOR XC42 .4.45 .1.4 .5.8 <.035 <.035
G Ck42A1 Wk. 1.1190 .39.44 .25.4 .75.9 <.035 <.035 N < .007 *3
11024 USA AISI/SAE 1040 (UNS G10400) .37.44 *1 .6.9 <.04 <.05
UK B 970 060A40 .38.43 *1 .5.7 <.05 <.05
G Ck40 Wk. 1.1186 .37.43 .15.35 .5.8 <.035 <.035 N < .007 *3
11025 USA AISI/SAE 1042 (UNS G10420) .4.47 *1 .6.9 <.04 <.05
UK BS970 060A42 .4.45 *1 .5.7 <.05 <.05
F AFNOR XC42 .4.45 .1.4 .5.8 <.035 <.035
G Ck42A1 Wk. 1.1190 .39.44 .25.4 .75.9 <.035 <.035 N < .007 *3
(continued)
574 Appendix 6: Metal Alloy Comparison Tables

(continued)
Composition
Country Designation C Si Mn P S Other
11026 USA AISI/SAE 1043 (UNS G10430) .4.47 *1 .71.0 <.04 <.05
UK BS970 080A42 .4.45 *1 .7.9 <.05 <.05
G DIN 17200; 17242; 1652. C45 Wk. .42.45 .15.35 .5.8 <.045 <.045
1.0503
11027 USA AISI/SAE 1045 (UNS G10450) .43.50 *1 .6.90 <.04 <.05
UK BS970 080M46 .42.50 *1 .61.0 <.05 <.05
F AFNOR XC45 .42.48 .1.35 .5.8 <.035 <.035
G DIN 17200; 1652; 0017242 Ck45 Wk. .42.5 .15.35 .5.8 <.035 <.035 N < .007 *3
1.1191
G DIN 17200 Cm45 Wk. 1.1201 .42.5 .15.35 .5.8 <.035 .020.035
CEN EN 10083-1 2C45 (C45E) 1.1191 Near equivalent to
BS970 070M55
11028 USA AISI/SAE 1046 (UNS G10460) .43.5 *1 .71.0 <.04 <.04
UK BS970 080A47 .45.5 *1 .7.9 <.05 <.05
11029 USA AISI/SAE 1049 (UNS G10490) .46.53 *1 .6.9 <.04 <.05
UK BS970 080M50 .45.55 *1 .61.0 <.05 <.05
F AFNOR XC50 .46.52 .15.35 .5.8 <.035 <.035
G CK50 Wk. 1.1206 .47.55 .15.35 .6.9 <.035 <.035
CEN EN 10083-1 2C50 (C50E) 1.1206 Near equivalent to
BS970 080M50
11030 USA AISI/SAE 1050 (UNS G10500) (*4) .48.55 *1 .6.9 <.04 <.05
UK BS970 080A52 .5.55 *1 .7.9 <.05 <.05
11031 USA AISI/SAE 1053 (UNS G10530) .48.55 *1 .71.0 <.04 <.05
UK BS970 080A52 .5.55 *1 .7.9 <.05 <.05
11032 USA AISI/SAE 1055 (UNS G10550) (*4) .5.6 *1 .6.9 <.04 <.05
UK BS970 070M55 .5.6 *1 .5.9 <.05 <.05
F AFNOR XC55 .52.6 .1.4 .5.8 <.035 <.035
G DIN 17200 Cm55 Wk. 1.1209 .52.6 .15.35 .6.9 <.035 .020.035
CEN EN 10083-1 2C55 (C55E) 1.1203 Near equivalent to
BS970 070M55
11033 USA AISI/SAE 1060 (UNS G10600) (*4) .55.65 *1 .6.9 <.04 <.05
UK BS970 080A57 55.6 .1.4 .7.9 <.05 <.05
F AFNOR XC60 .57.65 .15.35 .4.7 <.035 <.035
G DIN 17200 Cm60 Wk. 1.1223 .57.65 .15.35 .6.9 <.035 .02.035
G DIN 17200; 1652; 0017222 Ck60 Wk. .57.65 .15.35 .6.9 <.035 <.035
1.1221
G DIN 17200; 1652; 17222 C60 Wk. .57.65 .15.35 .6.9 <.045 <.045
1.0601
11034 USA AISI/SAE 1064 (UNS G10640) (*4) .6.7 *1 .5.8 <.04 <.05
UK BS970 060A62 .6.65 .1.4 .5.7 <.05 <.05
UK BS970 060A67 .65.7 .1.4 .5.7 <.05 <.05
F AFNOR XC65 .6.69 .1.4 .5.8 <.035 <.055
G DIN 17223 Federstahldraht FD (VD) .6.7 <.25 .5.9 <.03 <.03 (<.02)
Wk. 1.1230
G Ck65 .65 .3 .75 <.035 <.035
11035 USA AISI/SAE 1065 (UNS G10650) (*4) .6.7 *1 .6.9 <.04 <.05
UK BS970 080A62 .6.65 .1.4 .7.9 <.05 <.05
UK BS970 080A67 .65.1 .1.4 .7.9 <.05 <.05
11036 USA AISI/SAE 1069 (UNS G10690) .66.75 *1 .4.7 <.04 <.05
UK BS970 060A72 .7.75 .1.4 .5.7 <.05 <.05
F AFNOR XC68 .65.73 .15.35 .4.7 <.035 <.035
G DIN 0017222 Ck67 Wk. 1.1231 .65.72 15.35 .6.9 <.035 <.035
11037 USA AISI/SAE 1070 (UNS G10700) .65.75 *1 .6.9 <.04 <.05
UK BS970 080A72 .7.75 .1.4 .7.9 <.05 <.05
G DIN001 7222 Ck67 Wk. 1.1 231 .65.72 .15.35 .6.9 <.035 <.035
(continued)
Appendix 6: Metal Alloy Comparison Tables 575

(continued)
Composition
Country Designation C Si Mn P S Other
11038 USA AISI/SAE 1074 (UNS G10740) .7.8 *1 .5.8 <.04 <.05
UK BS970 070A78 .75.82 .1.4 .6.8 <.05 <.05
F AFNOR XC75 .7.8 .15.3 .4.7 <.035 <.035
G DIN0017222 Ck75Wk. 1.1248 .7.8 .15.35 .6.8 <.035 <.033 N < .007 *3
11039 USA AISI/SAE 1075 (UNS G10750) .7.8 *1 .4.7 <.04 <.05
UK BS970 060A78 .75.82 .1.4 .5.7 <.05 <.05
F AFNOR XC75 .7.8 .15.3 .4.7 <.035 <.035
G Wk. 1.1246 .7.77 <.2 .4.6 <.025 <.025
11040 USA AISI/SAE 1078 (UNS G10780) .72.85 *1 .3.6 <.04 <.05
UK BS970 060A78 .75.82 .1.4 .5.7 <.05 <.05
F AFNOR XC75 .7.8 .15.3 .4.7 <.035 <.035
G Wk. 1.1246 .7.77 <.2 4.6 <.025 <.025
11041 USA AISI/SAE 1080 (UNS G10800) .75.88 *1 .6.9 <.04 <.05
UK BS970 080A83 .7.9 .1.4 .7.9 <.05 <.05
F AFNOR XC80 .75.85 .1.4 .5.8 <.035 <.035 Cr < 2
G DIN 0017222 Ck75 .7.8 .15.35 .6.8 <.035 <.035 N < .007 *3
G Ck80 .8 .35 .75 <.035 <.035
11042 USA AISI/SAE 1084 (UNS G10840) .8.93 *1 .6.9 <.04 <.05
UK BS970 080A86 .83.9 .1.4 .7.9 <.05 <.05
F AFNOR XC85 .8 .2.4 .4.7 <.03 <.025
G DIN 0017222.Ck85 Wk. 1.1269 .8.9 .15.35 .45.65 <.035 <.035 N < .007 *3
11043 USA AISI/SAE 1085 (UNS G10850) .8.93 *1 .7-1.0 <.04 <.05
UK BS970 080A86 .83.9 .1.4 .7.9 <.05 <
G 90Mn4 Wk. 1.1273 .85.95 .25.5 .91.1 <.035 <.035
11044 USA AISI/SAE 1086 (UNS G10860) .8.93 *1 .3.5 <.04 <.05
UK BS970 050A86 .83.9 .1.4 .4.6 <.05 <.05
F AFNOR XC90 .85.95 .15.3 .3.5 <.03 <.025
G Mk83 Wk. 1.1262 .8.84 .1.25 .35.55 <.03 <.03 N < .007 *3
G Mk82 Wk. 1.1261 .8.84 .1.25 .25.45 <.025 <.025
11045 USA AISI/SAE 1090 (UNS G10900) .85.98 *1 .6.9 <.04 <.05
UK BS970 060A96 .931.0 .1.4 .5.7 <.05 <.05
11046 USA AISI/SAE 1095 (UNS G10950) .91.03 *1 .3.5 <.04 <.05
UK BS970 060A99 .951.05 .1.4 .5.7 <.05 <.05
F AFNOR Xc90 .951.05 .15.3 .2.45 <.03 <.025
G Mk97 .91.05 .15.25 .3.5 .045.055 .060.070 N < .007 *3
Higher manganese steels
12001 USA AISI 1513 (UNS G15130) .1.16 *1 1.11.4 <.04 <.05
F AFNOR 12M5 .1.15 <.4 .91.4 <.04 <.035
12002 USA AISI 1518 (UNS G15180) .15.21 *1 1.11.4 <.04 <.05
UK BS970 120M19 .15.23 *1 1.01.4 <.05 <.05
F AFNOR 20M5 .16.22 .1.4 1.11.4 <.035 <.035
G 20Mn6 Wk. 1.1169 .17.23 .3.6 1.31.6 <.035 <.035
12003 USA AISI 1522 (UNS G15220) .18.24 *1 1.11.4 <.04 <.05
UK BS970 120M19 .15.23 *1 1.01.4 <.05 <.05
F AFNOR 20M5 .16.22 .1.4 1.11.4 <.035 <.035
F AFNOR 18M5 .16.22 .1.4 1.11.5 <.04 .18.23
G 20Mn6 Wk. 1.1169 .17.23 .3.6 1.31.6 <.035 <.035
12004 USA AISI 1524 SAE 1024 (UNS G15240) .19.25 *1 1.35 <.04 <.05
1.65
UK BS970 150M19 .15.23 *1 1.351.7 <.05 <.05
G 20Mn6 Wk. 1.1168 .17.23 .3.6 1.31.6 <.035 <.035
12005 USA AISI 1525 (UNS G15250) .23.29 *1 .81.1 <.04 <.05
UK BS970 080A25 .23.28 *1 .7.9 <.05 <.05
(continued)
576 Appendix 6: Metal Alloy Comparison Tables

(continued)
Composition
Country Designation C Si Mn P S Other
12006 USA AISI 1526 (UNS G15260) .22.29 *1 1.11.4 <.04 <.05
UK BS970 120M28 .24.32 *1 1.01.4 <.05 <.05
G 9S-24 Mn4 Wk. 1.1136 .20.28 .3.6 .91.2 <.035 <.035
12007 USA AISI 1527 SAE 1027 (UNS G15270) .22.29 *1 1.21.5 <.04 <.05
UK BS970 150M28 .24.32 *1 1.31.7 <.05 <.05
G DIN 17200 Wk. 1.1170 .25.32 .15.4 1.31.65 <.035 <.035
12008 USA AISI 1536 SAE 1036 (UNS G15360) .3.37 *1 1.21.5 <.04 <.05
UK BS970 120M36 .32.4 *1 1.01.4 <.05 <.05
UK BS970 150M36 .32.4 *1 1.31.7 <.05 <.05
F AFNOR 35 M5 .32.38 .1.4 1.11.4 <.035 <.035
G 36Mn5, GS-36Mn5 Wk. 1.1167 .32.4 .15.35 1.21.5 <.035 <.035
12009 USA AISI 1541 SAE 1041 (UNS G15410) .36.44 *1 1.35 <.04 <.05
1.65
F AFNOR 40Mn .36.44 .1.4 1.01.35 <.04 <.035
G 36Mn5, GS-36 Mn5, Wk. 1.1167 .32.4 .15.35 1.21.5 <.035 <.035
12010 USA AISI 1547 SAE 1047 (UNS G15470) .43.51 *1 1.35 <.04 <.05
1.65
F AFNOR 45 M5 .39.48 .1.4 1.21.5 <.04 <.035
12011 USA AISI 1548 SAE 1048 (UNS G15480) .44.52 *1 1.11.4 <.04 <.05
F AFNOR 45 M5 .39.48 .1.4 1.21.5 <.04 <.035
12012 USA AISI 1551 SAE 1051 (UNS G15510) .46.56 *1 .851.15 <.04 <.05
UK BS970 080M50 .45.55 *1 .61.0 <.05 <.05
CEN EN 10083-1 2C50 (C50E) 1.1206 Near equivalent to
BS970 080M50
12013 USA AISI 1552 SAE 1052 (UNS G15520) .47.55 *1 1.21.5 <.04 <.05
F AFNOR 55 M5 .5.6 .1.4 1.21.5 <.05 <.035
12014 USA AISI 1561 SAE 1061 (UNS G15610) .55.65 *1 .751.05 <.04 <.05
UK BS970 080A57 .55.60 *1 .7.9 <.05 <.05
G Ck60 Wk. 1.221 .57.65 .15.35 .6.9 <.035 <.035
12015 USA AISI 1566 SAE 1066 (UNS G15660) .6.71 *1 .851.15 <.04 <.05
UK BS970 080A67 .65.7 *1 .7.9 <.05 <.05
12016 USA AISI 1572 SAE 1072 (UNS G15720) .65.76 *1 1.01.3 <.04 <.05

Free cutting steels

Composition
Country Designation C Si Mn P S Other
13001 USA AISI/SAE 1108 (UNS G11080) .08.13 *1 .5.8 <.04 .08.13
G DIN 17111, U7S10, Wk. 1.0700 <.l *2 .4.7 <.08 .08.12 N < .007
G 10320 Wk. 1.0721 .07.13 .1.4 .5.9 <.06 .15.25
13002 USA AISI/SAE 1109 (UNS G11090) .08.13 *1 .6.9 <.04 .08.13
13003 USA AISI/SAE 1110 (UNS G11100) .08.13 *1 .3.6 <.04 .08.13
The changes in manganese range at this carbon level are not reected in European specications
F AFNOR 12MF (approximate equivalent) .09.15 .1.4 .91.2 <.06 .12.24
13004 USA AISI/SAE 1116 (UNS G11160) .14.2 *1 1.11.4 <.04 .16.23
UK BS970 220M07 <.15 *1 .91.3 <.07 .2.3
G 9S20 Wk. 1.0711 <.13 <.05 .61.2 <.1 .18.25
13005 USA AISI/SAE 1117 (UNS G11170) .14.2 *1 1.01.3 <.04 .08.13
F AFNOR13MF .116 .1.4 .81.1 <.04 .09.13
G 9SMN 28 Wk. 1.0715 <.14 <.05 .91.3 <.1 .24.32
UK BS970 230M07 <.15 *1 .91.3 <.07 .25.35
13006 USA AISI/SAE 1118 (UNS G11180) .14.2 *1 1.31.6 <.04 .0813
13007 USA AISI/SAE 1119 (UNS G11190) .14.2 *1 1.01.3 <.04 .24.33
13008 USA AISI/SAE 1132 (UNS G11320) .27.34 *1 1.351.65 <.04 .08.13
UK BS970 216M28 .24.32 *1 1.11.5 <.06 .12.2
(continued)
Appendix 6: Metal Alloy Comparison Tables 577

(continued)
Composition
Country Designation C Si Mn P S Other
13009 USA AISI/SAE 1137 (UNS G11370) .32.39 *1 1.351.65 <.04 .08.13
UK BS970 225M36 .32.4 <.25 1.31.7 <06 .12.2
F AFNOR 35 M6 .3339 .1.4 1.31.7 <.04 .09.13
G Wk. 1.0726 .3239 .1.4 .5.9 <.06 .15.25
13010 USA AISI/SAE 1139 (UNS G11390) .35.43 *1 1.351.65 <.04 .13.2
13011 USA AISI/SAE 1140 (UNS G11400) .37.44 *1 .71.0 <.04 .08.13
13012 USA AISI/SAE 1141 (UNS G11410) .37.45 *1 1.351.65 <.04 .08.13
UK BS970 212A37 .35.40 *1 1.01.3 <.06 .12.2
13013 USA AISI/SAE 1144 (UNS G11440) .4.48 *1 1.351.65 <.04 .24.33
13014 USA AISI/SAE 1145 (UNS G11450) .42.49 *1 .71.0 <.04 .04.07
13015 USA AISI/SAE 1146 (UNS G11460) .42.49 *1 .71.0 <.04 .0813
UK BS970 212M44 .4.48 *1 1.01.4 <.06 .12.2
UK BS970 225M44 .4.48 *1 1.31.7 <.06 .2.3
F AFNOR 45 MF6 .41 .48 .1 .4 1.31.7 <.04 .24 .32
G 45S20 Wk. 1.0727 .42.5 .1.4 .5.9 <.06 .15.25
13016 USA AISI 12L 13 (UNS G12134) <.13 *1 .71.0 .07.12 .24.33 Pb .15.35
13017 USA AISI/SAE 12L 14 (UNS G12144) <.15 *1 .851.15 .04.09 .26.35 Pb .15.35
F AFNOR 10 Pb2 .05.15 <.3 .3.6 <.04 <.04 Pb .15.30
G 9SMn Pb28 <.14 <.05 .91.3 <.1 .24.32 Pb .15.30
G 10 SPb 20 .0713 .1.4 .5 .9 <.06 .15.25 Pb .15.30
G 9SMn Pb 36 <.15 <.05 1.01.5 <.1 .32.40 Pb .15.30
Note Where USA grades are closely graded they have been grouped together with groups of approximating European specications

Low alloy steels: ManganeseMolybdenum

Composition
Country Designation C Si Mn P S Cr Mo Ni Other
14001 USA AISI/SAE 4012 (UNS .09.14 .2.35 .751.0 <.035 <.04 .15.25
G40120)
G 15Mn Mo53 Wk. 1.5418 (*5) <.2 .3.5 1.11.4 <.04 <.04 .35
14002 USA AISI/SAE 4023 (UNS .2.25 .2.35 .7.9 <.035 <.04 .2-3
G40230)
F AFNOR 18MD4.05 (*5) <.22 .1.4 .91.5 <.035 <.035 <.3 .35.6 V < .04
G 20Mo3 Wk. 1.5416 .16.24 .15.35 .5.8 <.04 <.04 .25.35
14003 USA AISI/SAE 4024 (UNS .2.25 .2.35 .7.9 <.035 .035.05 .2.3
G40240)
F/G See AISI 4023
14004 USA AISI/SAE 4027 (UNS .25.3 .2.35 .7.9 <.035 <.04 .2.3
G40270)
14005 USA AISI/SAE 4028 (UNS .25.3 .2.35 .7.9 <.035 .035.05 .2.3
G40280)
G 15Mo3 Wk. 1.5415 (*5) .12.2 .1.35 .4.8 <.04 <.04 <.3 .2.35
G 22Mo4 Wk. 1.5419 (*5) .18.25 .2.4 .4.7 <.035 <.033 <.3 .34
14006 USA AISI/SAE 4032 (UNS .3.35 .2.35 .7.9 <.035 <.04 .2.3
G40320)
14007 USA AISI/SAE 4037 (UNS .35.40 .2.35 .7.9 <.035 <.04 .23
G40370)
UK BS970 605M30 (*5) .26.34 1.35 1.31.7 <.04 <.05 .22.32
UK BS970 605M36 (*5) .32.4 .1.35 1.31.7 <.04 <.05 .22.32
G GS-35Mn Mo5 Wk. 1.5411 .32.38 .3.5 1.01.4 <.035 <.035 .15.25
14008 USA AISI/SAE 4042 (UNS .4.45 .2.35 .7.9 <.035 <.04 .2.3
G40420)
14009 USA AISI/SAE 4047 (UNS .45.5 .2.35 .7.9 <.035 <.04 .2.3
G40470)
UK BS970 608M38 (*5) .32.4 .1.35 1.31.7 <.04 <.05 .4.55
G GS-40Mn Mo43 (*5) .36.43 .3.5 .91.2 <.035 <.05 .25.35
578 Appendix 6: Metal Alloy Comparison Tables

Low alloy steels: ChromiumMolybdenum

Composition
Country Designation C Si Mn P S Cr Mo Ni Other
14010 USA AISI/SAE 4118 .18.23 .2.35 .7.9 <.035 <.04 .4.6 .08.15
(UNS G41180)
14011 USA AISI/SAE 4130 .28.33 .2.35 .4.6 <.035 <.04 .81.1 .15.25
(UNS G41300)
14012 USA AISI/SAE 4135 .33.38 .2.35 .7.9 <.035 <.04 .81.1 .15.25
(UNS G41350)
F AFNOR 15CD3.5 .14.18 <.35 .3.6 <.04 <.035 .851.15 .15.3
F AFNOR12CD 4 .08.14 .14.4 .5.8 <.04 <.035 .851.15 .15.3
F AFNOR 15CD4.05 <.2 .1.4 .4.85 <.035 <.035 .751.23 .4.6 V < .04
F AFNOR 18CD4(S) *6 .16.22 .1.4 .6.9 <.035 <.035 .851.15 .15.3
F AFNOR 30CD 4 *6 .28.34 .1.4 .6.8 <.035 <.035 .851.15 .15.3
F AFNOR 35CD 4 *6 .33.39 .1.4 .6.9 <.035 <.035 .851.13 .15.3
G DIN 17155; 17175: .1.18 .1.35 .4.7 <.04 <.04 .81.15 .4.65
0017243. 13Cr Mo4.4
Wk. 1.7335
G 15Cr Mo5 .13.17 .15.35 .81.0 <.035 <.035 1.01.3 .2.3
G 20Cr Mo5 .18.23 .15.35 .91.2 <.035 <.035 1.11.4 .2.3
G DIN 17200: 001654. .22.29 .15.4 .5.8 <.035 <.035 .91.2 .15.3
(GS)25Cr Mo4 Wk.
1.7218
G DIN 17200: 001654. .3.37 .15.4 .5.8 <.035 <.033 .91.2 .15.3
(GS)34Cr Mo4 Wk.
1.7220
14013 USA AISI/SAE 4137 .35.4 .2.35 .7.9 <.035 <.04 .81.1 .15.25
(UNS G41370)
14014 USA AISI/SAE 4140 .38.43 .2.35 .751.0 <.035 <.04 .81.1 .15.25
(UNS G41400)
UK BS970 708A37 .35.4 .1.35 .71.0 <.04 <.05 .91.2 .15.25
UK BS970 708M40 .36.44 .1.35 .71.0 <.04 <.05 .91.2 .15.25
F AFNOR 40CD 4 .39.46 .2.50 .5.8 <.03 <.025 .951.3 .15.3
F AFNOR 42CD4 .39.46 .1.4 .6.9 <.035 <.035 .851.15 .15.3
G DIN 17200; 001654. .38.5 .3.5 .5.8 <.035 <.035 .81.2 .2.3
GS42CrMo4Wk.
1.7225
G DIN 17200. .3.37 .15.4 .5.8 <.035 .02.035 .91.2 .15.3
34CrMoS4
CEN EN 10083-1 42CrMo4 1.17225
Near equivalent to BS970 708M40
14015 USA AISI/SAE 4142 .4.45 .2.35 .751.0 <.035 <.04 .81.1 .15.25
(UNS G41420)
14016 USA AISI/SAE 4145 .43.48 .2.35 .751.0 <.035 <.04 .81.1 .15.25
(UNS G41450)
14017 USA AISI/SAE 4147 .45.50 .2.35 .751.0 <.035 <.04 .81.1 .15.25
(UNS G41470)
UK BS970 708 H42 .39.46 .1.35 .651.05 <.04 <.05 .81.25 .15.25
F AFNOR 42CD 4 .39.46 .1.4 .6.9 <.035 .035 .851.15 .15.3
G DIN 17200, 42Cr Mo .38.45 .15.4 .5.8 <.035 .02.035 .91.2 15.3
S4 Wk. 1.7227
14018 USA AISI/SAE 4150 .48.53 .2.35 .751.0 <.035 <.04 .81.1 .15.25
(UNS G41500)
14019 USA AISI/SAE 4161 .56.64 .2.35 .751.0 <.035 <.04 .7.9 .25.35
(UNS G41610)
G DIN 17200. 50Cr .46.54 .15.4 .5.8 <.035 <.035 .91.2 .15.25
Mo4 Wk. 1.7228
G GS-58 Cr Mn Mo443 .54.62 .3.5 .61.2 <.035 <.033 .81.2 .2.3
Wk. 1.7266
Appendix 6: Metal Alloy Comparison Tables 579

Low alloy steels: NickelChromiumMolybdenum

Composition
Country Designation C Si Mn P S Cr Mo Ni Other
14020 USA AISI/SAE 4320 .17.22 .2.35 .45.65 <.035 <.04 .4.6 .2.3 1.652.0
(UNS G43200)
F AFNOR 20 NCD 7 .16.22 .2.33 .45.65 <.03 <.023 .2.6 .2.3 1.652.0 Cu < 35
14021 USA AISI/SAE 4340 .38.43 .2.35 .6.8 <.035 <.04 .7.9 .2.3 1.652.0
(UNS G43400)
G DIN 0017242. 40 Ni Cr Mo .37.44 <.4 .7.9 <.02 <.015 .7.95 .3.4 1.652.0
73 Wk. 1.6562
14022 USA AISI/SAE 4718 .16.21 .7.9 .35.55 .3.4 .91.2
(UNS G47180)
F AFNOR 18NCD4 .16.22 .2.35 .5.8 <.03 <.025 .35.55 .15.3 .91.2 Cu < .35
14023 USA AISI/SAE 4720 .17.22 .2.35 .5.7 <.035 <.04 .35.55 .15.25 .91.2
(UNS G47200)
See AISI 4718 (UNS G47180)
Table No. 14022
14024 USA AISI/SAE 8115 .13.18 .2.35 .7.9 <.035 <.04 .3.5 .08.15 .2.4
(UNS G81150)
UK BS970 805A15 .13.18 1.35 .7.9 <.04 <.05 .4.6 .15.25 .4.7
14025 USA AISI/SAE 8615 .13.18 .2.35 .7.9 <.035 <.04 .4.6 .15.25 .4.7
(UNS G86150)
14026 USA AISI/SAE 8617 .15.2
(UNS G86170)
14027 USA AISI/SAE 8620 .18.23
(UNS G86200)
14028 USA AISI/SAE 8622 .2.25 Other elements as AISI/SAE 8615
(UNS G86220)
14029 USA AISI/SAE 8625 .23.28
(UNS G86250)
14030 USA AISI/SAE 8627 .25. 3
(UNS G86270)
14031 USA AISI/SAE 8630 .28. 33
(UNS G86300)
UK BS970 805A17 .15.2 .1.35 .7.9 <.04 <.05 .4.6 .15.25 .4.7
UK BS970 805A20 .18.23
UK BS970 805A22 .2.25 Other elements as BS970 805A17
UK BS970 805A24 .22.21
F AFNOR15NCD2 .13.18 .1.4 .7.9 <.04 <.035 .4.6 .15.25 .4.7
F AFNOR 20NCD2 .18.23 .1.4 .7.9 <.03 <.025 .4.6 .15.25 .4.7 Cu < .35
F AFNOR 30NCD2 .3.35 .1.4 .7.9 <.04 <.035 .4.6 .15.3 .5.8
G DIN001654. 21 Ni Cr Mo2 .17 .23 .15.4 .6.9 <.035 <.035 .35.65 .15.25 .4.7
Wk. 1.6523
G 21 Ni Cr Mo22 Wk. 1.6543 .18.23 .2.35 .7.90 <.035 <.035 .4.6 .23 .4.7
G 30 Ni Cr Mo22 Wk. 1.6545 .27. .34 15.34 .71.0 <.035 <.035 .4.6 .15.3 .4.7
14032 USA AISI/SAE 8637 .35.4 .2.35 751.0 <.035 <.04 4.6 .15.25 4.7
(UNS G86370)
14033 USA AISI/SAE 8640 (UNS .38.43
G86400)
14034 USA AISI/SAE 8642 (UNS .4.45
G86420)
14035 USA AISI/SAE 8645 (UNS .43.48
G86450) *7
14036 USA AISI/SAE 8650 (UNS .48.53 Other elements as AISI 8637
G86500)
14037 USA AISI/SAE 8655 (UNS .51.59
G86550)
14038 USA AISI/SAE 8660 (UNS .56. 64
G86600)
UK BS970 805A60 .55. 65 .1.35 .71.0 <.04 <.05 .4.6 .15.25 .4.7
F AFNOR 35 NCD2 .32.40 .1.4 .71.0 <.04 <.035 .4.6 .15.3 .4.7
F AFNOR 40 NCD2 .37.40 .1.4 .6.9 <.04 <.035 .4.6 .15.3 .4.7
F AFNOR 40 NCD2TS .38.44 .1.4 .71.0 <.035 <.03 .4.6 .15.3 .4.7
G 40Ni Cr Mo22 Wk. 1.6546 .37.44 .15.34 .71.0 <.035 <.035 .4.6 .15.3 .4.7
580 Appendix 6: Metal Alloy Comparison Tables

Low alloy steels: NickelMolybdenum

Composition
Country Designation C Si Mn P S Cr Mo Ni Other
14039 USA AISI/SAE 4615 (UNS G46150) .13.18 .2.35 .45.65 <.035 <.04 .2.3 1.65.2.0
14040 USA AISI/SAE 4617 (UNS G46170) .15.2
14041 USA AISI/SAE 4620 (UNS G46200) .17.22 Other elements as AISI 4615
UK BS970 665A17 .15.2 .1.35 .45.65 <.04 <.05 <.25 .2.3 1.62.0
UK BS970 665A19 .17.22 Other elements as BS970 665A17
14042 USA AISI 4621 (UNS G46210) .18.23 .2.35 .7.9 <.035 <.04 .2.3 1.652.0
UK BS970 665M20 .17.23 .1.35 .35.75 <.04 <.05 .23 1.52.0
14043 USA AISI/SAE 4626 (UNS G46260) .24.29 .2.35 .45.65 <.035 <.04 .15.25 .71.0
UK BS970 665A22 *5 .2.25 .1.35 .45.65 <.04 <.05 <.25 .2.3 1.62.0
UK BS970 665A24 *5 .22.27 .1.35 .45.65 <.04 <.05 <.25 .2.3 1.62.0

Low alloy steels: Chromium

Composition
Country Designation C Si Mn P S Cr Mo Ni Other
14044 USA AISI/SAE 5115 (UNS G51150) .13.18 .2.35 .7.9 <.035 <.04 .79
14045 USA AISI/SAE 5120 (UNS G51200) .17.22 Other elements as AISI 5115
UK BS970 523A14 *5 .12.17 .1.35 .3.5 <.04 <.05 .35
UK BS970 527A19 .17.22 .1.35 .7.9 <.04 <.05 .79
F AFNOR 18C4 .16.21 .1.4 .6.8 <.04 <.035 .851.15
G DIN 17210:001654. 15Cr3 Wk. 1.7015 .12.18 .15.4 .4.6 <035 <035 .47
G 20Cr MnS33 Wk. 1.7121 .17.23 .2.35 .61.0 <.04 <.02 .61.0
14046 USA AISI/SAE 5130 (UNS G51300) .28.33 .2.35 .7.9 <.035 <.04 .81.1
14047 USA AISI/SAE 5132 (UNS G51320) .30.35 .2.35 .6.8 <.035 <.04 .751.0
UK BS970 530A30 .28.33 .1.35 .6.8 <.04 <.05 .91.2
UK BS970 530A32 .30.35 .135 .6.8 <.04 <.05 .91.2
F AFNOR 28 C4 .253 <.4 .6.9 <.04 <.035 .851.15
F AFNOR 32 C4 .3.35 .1.4 .6.9 <.035 <.035 .851.15
G DIN 17200:001654.34Cr4Wk. 1.17033 .3.37 .15.4 .6.9 <.035 <.035 91.2
14048 USA AISI/SAE 5135 (UNS G51350) .33.38 .2.35 .6.8 <.035 <.04 .81.05
14049 USA AISI/SAE 5140 (UNS G51400) .38.43 .2.35 .7.9 <.035 <.04 .7.9
UK BS970 530A36 .34.39 .1.35 .6.8 <.04 <.05 .91.2
UK BS970 530A40 .38.43 .1.35 .6.8 <.04 <.05 .91.2
F AFNOR 38 C4 .35.4 .1.4 .6.9 <.035 <.035 .851.15
F AFNOR 42 C4 .39.45 .1.4 .6.9 <.035 <.035 .851.15
G DIN 17200:001654. 34Cr4 Wk. 1.7034 .3037 15.4 .6.9 <.035 <.035 .91.2
G DIN 17200:001654. 37Cr4 Wk. 1.7035 .34.41 .15.4 .6.9 <.035 <.035 .91.2
14050 USA AISI/SAE 5145 (UNS G51450) .43.49 .2.35 .7.9 <.035 <.04 .7.9
14051 USA AISI/SAE 5147 (UNS G51470) .46.51 .2.35 .7.95 <.035 <.04 .851.15
F AFNOR 42C4TS .38.44 .1.4 .6.9 <.025 <.03 .851.15 <.3
F AFNOR 45 C4 .41.48 .1.4 .6.9 <.035 <.035 .851.15
14052 USA AISI/SAE 5150 (UNS G51500) .48.53 .2.35 .7.9 <.035 <.04 .7.9
14053 USA AISI/SAE 5155 (UNS G51550) .51.59 .2.35 .7.9 <.035 <.04 .7.9
14054 USA AISI/SAE 5160 (UNS G51600) .56.64 .2.35 .751.0 <.035 <.04 .7.9
UK BS970 526M60 .55.65 .1.35 .5.8 <.04 <.05 .5.8
F AFNOR 50 C4 .46.54 .1.4 .6.9 <.04 <.035 .81.15
14055 USA AISI E51100. SAE 51100 (UNS .981.1 .2.35 .25.45 <.025 <.025 .91.15
G51986)
14056 USA AISI E52100. SAE 52100 (UNS .981.1 .2.35 .25.45 <.025 <.025 1.31.6
G52986)
UK BS970 534A99 .951.1 .1.35 .25.4 <.04 <.05 1.21.6
F AFNOR 100 C6 .951.1 .15.35 .2.4 <.03 <.025 1.351.6
G DIN 0017230:LW. 100Cr6 Wk. 1.3505 951.1 .15.35 .25.4 <.03 <.025 1.351.6
G 100Cr6 Wk. 1.2067 .951.05 .15.35 .25.4 <.035 <.033 1.41.7
Appendix 6: Metal Alloy Comparison Tables 581

Low alloy steels: ChromiumVanadium

Composition
Country Designation C Si Mn P S Cr Mo Ni Other
14057 USA AISI 6118 (UNS G61180) .16.21 .2.35 .5.7 <.035 <.04 .5.7 V .1.15
G 21 CrV4 Wk. 1.7510 .18.24 1.2 .81.0 <.035 <.035 .91.2 V .07. 12
14058 USA AISI/SAE 6150 (UNS G61500) .48.53 .2.35 .7.9 <.035 <.04 .81.1 V > .15
UK BS970 735A50 .46.54 1.35 .6.9 <.04 <.05 .81.1 V > .15
F Y50 CV4 .5 .3 .8 1.0 V .15
G DIN 17200; 17221; 17225 (GS)50Cr .47. 55 .15.4 .71.0 <.035 <.035 .91.2 V. 1.2
V40 Wk. 1.8159
CEN EN 10083-1 51 CrV4 1.8159 Near
equivalent to BS970 735A50

Austenitic stainless steels

Composition
Country Designation C Si Mn P S Cr Mo Ni Other
15001 USA AISI 201 (UNS S20100) <.15 <1.0 5.57.5 <.06 <.03 16.018.0 3.55.5
15002 USA AISI 202 (UNS S20200) <.15 <1.0 7.510.0 <.06 <.03 17.019.0 4.06.0 N < .25
UK BS970 284S16 <.07 <1.0 7.010.0 <06 <.03 16.518.5 4.06.5 N .15.25
G X8 Cr Mn Ni 189Wk. 1.4371 <1 <1.0 7.59.5 <.045 <.03 17.019.0 4.56.5 N .1.2
15003 USA AISI 301 (UNS S30100) <.15 <1.0 <2.0 <.045 <.03 16.018.0 6.08.0
UK BS970 301S21 <.15 .21.0 .52.0 <.045 <.03 16.018.0 6.08.0
F AFNOR Z12CN17-08 .0815 <1.0 <2.0 <.04 <.03 16.018.0 6.58.5
G DIN 17440:0017442. <.07 <1.0 <2.0 <.045 <.03 17.020.0 8.510.0
X5 Cr Ni 18.9 Wk. 1.4301
15004 USA AISI 302 (UNS S30200) <.15 <1.0 <2.0 <.045 <.03 17.019.0 8.010.0
UK BS970 302S25 <12 .21.0 .52.0 <.045 <.03 17.019.0 8.011.0
UK BS970 302S17 <.08 .21.0 .52.0 <.045 <.03 17.019.0 9.011.0
F AFNOR Z10CN 18-09 <.12 <1.0 <2.0 <.04 <.03 17.019.0 8.010.0
F AFNOR Z12CN 18-10 <.15 .2.4 .2.4 <.04 <.03 17.019.0 8.010.0
G X12Cr Ni 18 8 Wk. 1.4300 <.12 <.1.0 <2.0 <.045 <.03 17.019.0 8.510.0
15005 USA AISI 302 B (UNS S30215) <.15 2.03.0 <2.0 <.045 <.03 17.019.0 8.010.0
15006 USA AISI 303 (UNS S30300) <.15 <1.0 <2.0 <.2 >.15 17.019.0 8.010.0
UK BS970 303S21 <.12 .21.0 1.02.0 <.045 .15.3 17.019.0 8.011.0
F AFNOR Z10 CNF 18-09 <12 <1.0 <2.0 <.06 >.15 17.019.0 8.010.0
G DIN 17440:0017442. X12 Cr <.15 <1.0 <2.0 <.045 .15.35 17.019.0 8.010.0
Ni S18 8 Wk. 1.4305
15007 USA AISI 303 SE <.15 <1.0 <2.0 <.2 <.06 17.019.0 8.010.0 Se > .15
(UNS S30323)
UK BS970 303S41 <.12 .21.0 1.02.0 <.045 <.03 17.019.0 8.011.0 Se > .15.3
15008 USA AISI 304 (UNS S30400) <.08 <1.0 <2.0 <.045 <.03 18.020.0 8.010.5
UK BS970 304S15 <.06 .21.0 .52.0 <.045 <.03 17.519.0 8.011.0
UK BS970 304S16 <.06 .21.0 .52.0 <.045 <.03 17.519.0 9.011.0
F AFNOR Z6CN 18-09 <.07 <1.0 <2.0 <.045 <.03 17.019.0 8.011.0
G X5 Cr Ni 18 9 Wk. 1.4301 <.07 <1.0 <2.0 <.045 <.03 17.020.0 8.510.0
15009 USA AISI 304 L (UNS S30403) <.03 <1.0 <2.0 <.045 <.03 18.020.0 8.012.0
UK BS970 304S12 <.03 .21.0 .52.0 <.045 <.03 17.519.0 9.012.0
F AFNOR Z2 CN 18-10 <.03 <1.0 <2.0 <.04 <.03 17.019.0 9.011.0
FG X2 Cr Ni 18 9 Wk. 1.4306 <.03 <1.0 <2.0 <.045 <.03 17.020.0 10.012.5
15010 USA AISI 305 (UNS S30500) <.12 <1.0 <2.0 <.045 <.03 17.019.0 10.513.0
UK BS970 305S19 <.1 .21.0 .52.0 <.045 <.03 17.019.0 11.013.0
F AFNOR Z8 CN 18-12 <.1 <1.0 <2.0 <.04 <.03 17.019.0 11.013.0
G DIN 17445 G-X10 Cr Ni 18 8 <.12 <2.0 <1.5 <.045 <03 17.019.5 18.010.0
Wk. 1.4312
(continued)
582 Appendix 6: Metal Alloy Comparison Tables

(continued)
Composition
Country Designation C Si Mn P S Cr Mo Ni Other
15011 USA AISI 308 (UNS S30800) .08 <1.0 <2.0 <.045 <.03 19.021.0 10.012.0
15012 USA AISI 309 (UNS S30900) <.2 <1.0 <2.0 <.0.45 <.03 22.024.0 12.015.0
UK BS970 309S24 <.15 .21.0 .52.0 <.045 <.03 22.025.0 13.016.0
F AFNOR Z12 CNS 25-13 <.2 1.02.0 <2.0 <.04 <.03 20.023.0 12.014.0
G G-XI5 Cr Ni 25-12 Wk. 1.4830 12 <1.5 <2.0 <.045 <.03 24.026.0 12.014.0
15013 USA AISI 310 (UNS S31000) <.25 <1.5 <2.0 <.045 <.03 24.026.0 19.022.0
UK BS970 310S24 <.15 .21.0 .52.0 <.045 <.03 23.026.0 19.022.0
F AFNOR Z12 CN 25-20 <.15 <1.0 2.0 <.04 <.03 23.026.0 18.021.0
G G-X15 Cr Ni 25-20 Wk. .1.2 <1.5 <2.0 <.045 <.03 24.026.0 19.021.0
1.4840
15014 USA AISI 310S (UNS S31008) <.08 <1.5 <2.0 <.045 <.03 24.026.0 19.022.0
G X5 Cr Ni 25 21 Wk. 1.4335 <.07 <1.0 <2.0 <.045 <.03 19.022.0 19.022.0
15C15 USA AISI 314 (UNS S31400) <.25 1.53.0 <2.0 <.045 <.03 23.026.0 19.022.0
15016 USA AISI 316 (UNS S31600) <.08 <1.0 <2.0 <.045 <.03 16.018.0 2.03.0 10.014.0
UK BS970 315S16 <.07 .21.0 .52.0 <.045 <.03 16.518.5 1.251.75 9.011.0
UK BS970 316S16 <.07 .21.0 .52.0 <.045 <.03 16.518.5 2.253.0 10.0 13.0
F AFNOR Z6 CND 17-11 <.07 <1.0 <2.0 <.04 <.03 16.018.0 2.02.5 10.012.0
G DIN 17440; 17445; 17224. X5 <.07 <1.0 <2.0 <.045 <.03 16.518.5 2.02.5 10.513.5
Cr Ni Mo 1810 Wk. 1.4401
15017 USA AISI 316 L (UNS S31603) <.03 <1.0 <2.0 <.045 <.03 16.018.0 2.03.0 10.014.0
UK BS970 316S12 <.03 .21.0 .52.0 <.045 <.03 16.518.5 2.253.0 11.014.0
F AFNOR Z2 CND 17-12 <.03 <1.0 <2.0 <.04 <.03 16.018.0 2.02.5 11.013.0
G DIN 17440; 17442:001654 X2 <.03 <1.0 <2.0 <.045 <.03 16.518.5 2.02.5 11.014.0
Cr Ni Mo 18-10 Wk. 1.44041
15018 USA AISI 317 (UNS S31700) <.08 <1.0 <2.0 <.045 <.03 18.020.0 3.04.0 11.015.0
UK BS970 317S16 <.06 .21.0 .52.0 <.045 <.03 17.519.5 3.04.0 12.015.0
F AFNOR Z2 CND 19-15 <.03 <1.0 <2.0 <.04 <.03 18.020.0 13.04.0 14.016.0
G DIN 17440 X2 Cr Ni Mo 18-16 <.025 <1.0 <.02 <.025 <.02 17.019.0 3.04.0 15.017.0
Wk. 1.4438
15019 USA AISI 321 (UNS S32100) <.08 <1.0 <2.0 <.045 <.03 17.019.0 9.012.0 Ti > 5 C
UK BS970 321S12 <.08 .21.0 .52.0 <.045 .03 17.019.0 9.012.0 Ti 5 C .7
UK BS970 321S20 <.12 .21.0 .52.0 <.045 <.03 17.019.0 8.011.0 Ti 5 C .9
F AFNOR Z6 CN 18-10 .05.1 <1.0 <2.0 <.03 <.03 16.020.0 8.010.0 Ti
F AFNOR Z6 CNT 18-11 <.08 <1.0 <2.0 <.04 <.03 17.019.0 10.012.0 Ti 5 C .6
G DIN 17440:43720 X10 Cr Ni <.1 <1.0 <2.0 <.045 <.03 17.019.0 9.011.5 Ti > 5 C
Ti 18-9 Wk. 1.4541
15020 USA AISI 347 (UNS S34700) <.08 <1.0 <2.0 <.045 <.03 17.019.0 9.012.0 Nb + Ta > 10 C
UK BS970 347S17 <.08 .21.0 .52.0 <.045 <.03 17.019.0 9.012.0 Nb 10 C 1.0
F AFNOR Z6 CN Nb 18-11 <.08 <1.0 <2.0 <2.0 <.03 17.019.0 10.012.0 Nb + Ta
10 C 1.0
G DIN 17440 X10 Cr Ni Nb 18-9 <.1 <1.0 <2.0 <.045 <.03 17.019.0 9.011.5 Nb > 8 C
Wk. 1.4550
15021 USA AISI 348 (UNS S34800) <.08 <1.0 <2.0 <.045 <.03 17.019.0 9.0 .13.0
Nb + Ta > 10 CTa < .1
Co < .2
15022 USA AISI 384 (UNS S38400) <.08 <1.0 <2.0 <.045 <.03 15.017.0 17.019.0
Appendix 6: Metal Alloy Comparison Tables 583

Ferritic and martensitic stainless steels

Composition
Country Designation C Si Mn P S Cr Mo Ni Other
15023 USA AISI 403 (UNS S40300) <.15 <.5 <1.0 <.04 <.03 11.513.0
UK BS970 403S17 <.08 <.8 <1.0 <.04 <.03 12.014.0 <.5
UK BS970 410S21 .09. <.8 <1.0 <.04 <.03 11.513.5 <1.0
15
F AFNOR Z10 C13 <.12 <1.0 <1.0 <.04 <.03 12.014.0
G X7 Crl4; G-X7 Cr 13 Wk. <.08 <1.0 <1.0 <.045 <03 13.015.0
1.4001
G DIN 17440; 001654 (G-) .08. <1.0 <1.0 <.045 <.03 12.014.0
X10 Cr 13 Wk. 1.4006 12
15024 USA AISI 405 (UNS S40500) <.08 <1.0 <1.0 <.04 <.03 11.514.5 AI .1.3
UK BS970 405S17 <.08 <.8 <1.0 <.04 <.03 12.014.0 <.5 AI .1.3
F AFNOR Z6 CA 13 <.08 <1.0 <1.0 <.04 <.03 11.513.5 <.5 AI .1.3
G DIN 17440 X7 Cr AI 13. Wk. <.08 <1.0 <1.0 <.045 <.03 12.014.0 AI .1.3
1.4002
15025 USA AISI 410 (UNS S41000) <.15 <1.0 <1.0 <.04 <.03 11.513.5
UK BS970 410S21 .09. <.8 <1.0 <.04 <.03 11.513.5 <1.0
15
F AFNOR Z10C-13 <.12 <1.0 <1.0 <.04 <.03 12.014.0
F AFNOR Z12C-13 .08.15 <1.0 .04 <.04 <.03 11.513.5 <.5
G DIN 17440:0017442 X15 .12.17 <1.0 <1.0 <.045 <.03 12.014.0
Cr 13 Wk. 1.4024
15026 USA AISI 414 (UNS S41400) <.15 <1.0 <1.0 <.04 <.03 11.513.5 1.25
2.50
15027 USA AISI 416 (UNS S41600) <.15 <1.0 <1.25 <.06 >.15 12.014.0 <.6
UK BS970 416S21 .09.15 <1.0 <1.5 <.06 .15.3 11.513.5 <.6 <1.0
F AFNOR Z12 CF 13 <.15 <1.0 <1.5 <.06 >.15 12.014.0 <.6 <.5
G X 12 Cr S 13 Wk. 1.4005 <.15 <1.0 <1.0 <.045 .15.25 12.013.0
15028 USA AISI 416 SE (UNS S41623) <.15 <1.0 <1.25 <.06 <.06 12.014.0 Se > .15
UK BS970 416S41 .09. <1.0 <1.5 <.06 <.06 11.513.5 <.6 <1.0 Se
15 .15.35
15029 USA AISI 420 (UNS S42000) >.15 <1.0 <1.0 <.04 <.03 12.014.0
UK BS970 420S29 .14.2 <.8 <1.0 <.04 <.03 11.513.5 >.6 >1.0
UK BS970 420S37 .2.28 <.8 <1.0 <.04 <.03 12.014.0 <1.0
F AFNOR Z20 C13 .15.24 <1.0 <1.0 <.04 <.03 12.014.0 <1.0
G DIN 17440; 17224:0017442 .17.22 <1.0 <1.0 <.045 <.03 12.014.0
X20 Cr 13 Wk. 1.402
15030 USA AISI 420 F (UNS S42020) >.15 <1.0 <1.25 <.06 >.15 12.014.0 <.6
15031 USA AISI 429 (UNS S42900) <.12 <1.0 <1.0 <.04 <.03 14.016.0
15032 USA AISI 430 (UNS S43000) <.12 <1.0 <1.0 <.04 <.03 16.018.0
UK BS970 430S15 <1 <.8 <1.0 <.04 <.03 16.018.0 <.5
F AFNOR Z15 CN 16-02 <.18 .2.4 2.4 <.04 <.03 15.017.0 1.02.0
G DIN 17440; 001654 X8 Cr <.1 <1.0 <1.0 <.045 <.03 15.517.5
17 Wk. 1.4016
15033 USA AISI 430 F (UNS S43020) <.12 <1.0 <1.25 <.06 >.15 16.018.0 <.6
15034 USA AISI 430 F SE (UNS <.12 <1.0 <1.25 <.06 <.06 16.018.0 Se > .15
S43023)
15035 USA AISI 431 (UNS S43100) <.2 <1.0 <1.0 <.04 <.03 15.017.0 1.25
2.5
UK BS970 431S29 .12.2 <.8 <1.0 <.04 <.03 15.018.0 2.03.0
F AFNOR Z15 CN17-03 <.18 .2.4 .2.4 <.04 <.03 15.017.0 1.02.0
G DIN 17440;001654 X22 Cr .15.23 <1.0 <1.0 <.045 <.03 16.018.0 1.52.5
Ni 17 Wk. 1.4057
(continued)
584 Appendix 6: Metal Alloy Comparison Tables

(continued)
Composition
Country Designation C Si Mn P S Cr Mo Ni Other
15036 USA AISI 434 (UNS S43400) <.12 <1.0 <1.0 <.04 <.03 16.018.0 .75
1.25
UK BS970 434S19 <.1 <.8 <1.0 <.04 <.03 16.018.0 .91.3 <.5
F AFNOR Z8CD 17-01 <.1 <1.0 <1.0 <.04 <.03 16.018.0 .91.3 <.5
G DIN 17440 X6 Cr Mo 17 <.07 <1.0 <1.0 <.045 <.03 16.018.0 .91.2
Wk. 1.4113
15037 USA AISI 436 (UNS S43600) <.12 <1.0 <1.0 <.04 <.03 16.018.0 .75 Nb+Ta
1.25 5C.7
15038 USA AISI 440 A (UNS S44002) .6.75 <1.0 <1.0 <.04 <.03 16.018.0 <.75
F AFNOR Z50 CD 14 .5.6 <1.0 <1.0 <.04 <.03 13.015.0 .5.6
G X65 Cr Mo 14 Wk. 1.4109 .6.75 <1.0 <1.0 <.045 <.03 13.015.0 .5.6
G X55 Cr Mo 14 Wk. 1.4110 .5.6 <1.0 <1.0 <.045 <.03 13.015.0 .5.6
15039 USA AISI 440 B (UNS S44003) .75. <1.0 <1.0 <.04 <.03 16.018.0 <.75
95
15040 USA AISI 440 C (UNS S44004) .951.2 <1.0 <1.0 <.04 <.03 16.018.0 <.75
F AFNOR Z100CD17
G DIN 0017230 X105 Cr Mo .951.2 <1.0 <1.0 <.045 <.03 16.018.0 .4.8
17
15041 USA AISI 442 (UNS S44200) <.2 <1.0 <1.0 <.04 <.03 18.023.0
UK BS970 442S19 <.1 <.8 <1.0 <.04 <.03 18.022.0 <.5
15042 USA AISI 446 (UNS S44600) <.2 <1.0 <1.5 <.04 <.03 23.027.0 N < .25
F AFNOR Z10 C24 <.12 <1.5 <1.0 <.04 <.03 23.026.0
G X20 Cr 25 Wk. 1.3810 <.25 .52.0 <.5 24.026.0
G X8 Cr 28 Wk. 1.4083 <.1 <1.0 <1.0 <.045 <.03 27.029.0
15043 USA AISI 501 (UNS S50100) >.1 <1.0 <1.0 <.04 <.03 4.06.0 .4.65
UK BS1504 Grade 1504-625 <.15 <.5 .3.7 <.045 <.045 4.06.0 .45.65 <.4 Cu < 4
G GS-12 Cr Mo 19 5 Wk. 1. .08.15 .3.5 .47 <.035 <.035 4.55.5 .45.55
7363
15044 USA AISI 502 (UNS S50200) <.1 <1.0 <1.0 <.04 <.03 4.06.0 .4.65
UK BS1504 Grade 1504625 <.15 <.5 .3.7 <.045 <.045 4.06.0 .45.65 <.4 Cu < .4
G GS-12 Cr Mo 19 5 Wk. .08.15 .3.5 .4.7 <.035 <.035 4.55.5 .45.55
1.7363

Specially named steels

Composition
Country Designation C Si Mn P S Cr Mo Ni Other
16001 USA Music Wire ASTM .71.0 .1.3 .2.6 <.025 <.03
A228
(UNS K08500)
See AISI 1078; 1086; 1095(UNS G10780; UNS G 10860; UNS G 10950 Table Nos: 11040; 11044: 11046)
16002 USA HY80ASTM A543 <.18 .18.37 <.4 <.02 <.02 1.01.5 .45.60 2.253.25 V < .03
(UNS J42015 (HY80))
*8
USA HY130- <.12 .6.9 .4.7 .3.65 4.755.25 V .05.10
USA HY140Designation no
longer in use
These steels are made in
Europe to the USA
analyses
(continued)
Appendix 6: Metal Alloy Comparison Tables 585

(continued)
Composition
Country Designation C Si Mn P S Cr Mo Ni Other
16003 USA Carpenter 20 Cb Stainless (UNS N08020 (20Cb-3)) Now replaced by
20Cb-3 below
USA Carpenter 20 Cb-3 <06 <1.0 <2.0 <.035 <.035 19.021.0 2.03.0 32.535.0 Cu 3.04.0 Nb
+ Ta 8C < 1.0
16004 USA Allegheny Ludlum .08 .4 1.4 15 1.25 26 Ti 2.15 Al .2
A286 (UNS K66286) .004
USA BoforsA286 .06 15 1.3 25 Ti 2.15 AI.2
.004
F AFNOR Z6NCT25-15
G DIN0017225 X5 Ni Cr Ti <.08 <1.0 1.0 <.03 <.03 13.516.0 1.01.5 24.027.0 Al < .35 Ti 1.9
26 15 Wk. 1.4980 2.0 2.3 B
.003.010V.1
.5
G LN 1.4944.4
EUR AECMA FE PA92HT
16005 USA Allegheny Ludlum .08 .4 1.0 16.5 2.7 4.3 N .1
AM350 (UNS S35000)
16006 USA Allegheny Ludlum .15 .4 1.0 15.5 2.75 4.25 N .1
AM355 (UNS S35500)
16007 USA Carpenter. Custom 455 .03 .25 11.75 9.0 Ti 1.2 Cu 2.2
(UNS S45500) Nb+ Ta 0.3
16008 USA 15-5 PH (UNS S15500) .04 .8 15.0 4.6 Cu 3.3 Nb .27
16009 USA PH 148 Mo .04 .6 15.1 2.2 8.3 AI 1.2 + N
(UNS S14800)
16010 USA PH 157 Mo .07 15 2.2 7 AI 1.1
(UNS S15700)
16011 USA PH 177 (UNS S17700) .07 17 7 AI 1.1
16012 USA SAE H11 Tool Steel .3.4 .81.2 .2.4 4.755.50 1.251.75 V .3.5
(UNS T20811)
USA H11 MOD (UNS K74015)Replaced by UNS T20811
USA Vascojet 1000
UK BS4659 BH11 .32.42 .85 <.4 4.755.25 1.251.75 V .3.5
1.15
F E-40CDV20
G (G-) X38 Cr Mo V 5 1 .36.42 .91.2 .3.5 <.03 <.03 4.85.8 1.4
Wk. 1.2343
G Wk. 1.7784
EUR AECMA FE-PM13S .37.43 .3 4.755.25 1.3 V .5
16013 USA 17/4 PHASTM A 579- <.07 <1.0 <1.0 <.025 <.025 15.517.5 3.05.0 Cu 3.05.0 Nb
AIS1630 (UNS J92200) .1545
Grade 61
16014 USA PH 13-8 Mo (UNS S .03 <.1 12.8 2.2 8.2 AI 1.1
13800)
16015 USA Maraging ASTM 579 <.03 <.1 <.1 <.01 <.01 3.03.5 17.019.0 Ti .15.25 Co
Grade 71 Yield 200 ksi 8.09.0 AI .05
(UNS K92820) 15
Ca .06 Zr .02 B
.003
16016 USA Maraging ASTM 579 <.03 <.1 <.1 <.01 <.01 4.65.2 17.019.0 Ti .3.5 Co
Grade 72 Yield 250 ksi 7.58.3 AI
(UNS K92940) .05.15
Ca .06 Zr .02 B
.003
16017 USA Maraging ASTM 579 <.03 <.1 <.1 <.01 <.01 4.65.2 18.019.0 Ti .5.8 Co
Grade 73 Yield 275 ksi 8.59.5 AI
(UNS K93160) .05.15
Ca .06 Zr .02 B
.003
(continued)
586 Appendix 6: Metal Alloy Comparison Tables

(continued)
Composition
Country Designation C Si Mn P S Cr Mo Ni Other
16018 USA ARMCO 21-6-9 <.08 <1.0 8.0 <.06 <.03 19.021.5 5.57.5 N .15.4
(UNS S21900) 10.0
UK BS970 284S16 <.07 <1.0 7.0 <.06 <.03 16.518.5 4.06.5 N .1525
10.0
G X8 Cr Mn Ni 18 9Wk. <.1 <1.0 7.5 <.045 <.03 17.019.0 4.56.5 N .1.2
1.4371 9.5
16019 USA ALMAR 362 (UNS .03 .2 .3 <.015 <.015 14.5 6.5 Ti .8 Nominal
S36200) comp.
16020 USA Nitronic 33 <.08 <1.0 11.5 <.06 <.03 17.019.0 2.253.75 N .2.4
14.5
USA Nitronic 32 <.1 .5 12 <.06 <.03 18 1.6 N .34
USA Nitronic 60 (UNS <.1 3.54.5 7.0 <.06 <.03 1618 8.09.0 N .08.18
S21800) 9.0
16021 USA Kovar (low expansion <.04 <.2 <.5 <.2 <.2 29 Co 17.0 Fe
alloy) Balance.
Nominal comp.
UK Nilo K 29 Co 17.0 Fe
Balance.
Nominal comp.
F Dilver P0 29 Co 21.8 Fe
Balance.
Nominal comp.
F Dilver P1 29 Co 18.0 Fe
Balance.
Nominal comp.
G Dilaton 29/18 <.05 <.2 <1.0 2830 Co 17.019.0
Wk. 1.3981 Fe Bal.
Nominal comp.
16022 USA Invar .1 .2 .5 36 Nominal comp.
USA Invar 36 36 Nominal comp.
UK Nilo 36 36 Nominal comp.
16023 USA Invar 42 42 Nominal comp.
UK Nilo 42 42 Nominal comp.
Notes for steel tables
<: less than, xy range
>: greater than, x approx
*1: Silicon content depends on whether the steel is rimming, balanced or killed. For killed steel Si < 0.4. For AISI up to but excluding 1015 Si < 0.1. Ranges depend on sheet
making practice
*2: Traces
*3: For electric steel N < 0.012
*4: Spring Steel
*5: Only very approximate equivalent
*6: Can be alloyed with lead
*7: Can have boron content of 0.0005 % minimum. Number then carries a B
*8: Pressure vessel plate steels. Special conditions may be required. Vacuum treatment, special testing, impact testing, nondestructive testing

European CEN Designations for Steels
European EN specications for metal alloys are currently
being generated and adopted. These will progressively
supersede the various national standards for steels, as with
other materials. However, it will be some years before this
process is completed and fully implemented.
The European designation system for steels is set out in
the specication EN10027 and in the ECISS information
circular DD214:1993 ECISS/IC10:1992. The EN designa-
tions for steels will consist of three parts:
Appendix 6: Metal Alloy Comparison Tables
European Standard number, e.g. EN 10083-1.
Steel Name (grade)Symbolic letters and numbers
expressing the application and principal characteristics,
e.g. 2C50 (C50E).
Steel Numbera 5 digit designation based on the
existing German Werkstoff (Wk.) number, with a further
two digits held in reserve, e.g. 1.1206.
The examples given above make up the complete EN
designation: EN 10083-1 2C50 (C50E) 1.1206; See
table 11029. Within the alloy tables, the Steel Name part
has been underlined (e.g. EN 10083-1 2C50 (C50E) 1.1206)
to indicate the separate the parts of the designation.
587
In principal, EN steel numbers can be inferred for steels
by using the existing Werkstoff numbers, however there is
no guarantee that such numbers have currently been either
agreed or adopted by any particular country. New speci-
cations, once adopted, will be issued by each national
standards organisation, and any existing, competing speci-
cations will be withdrawn.
The implementation of the new standards recognises
three levels of equivalence between EN designations and
existing national grades of steel: Close Equivalent, Near
Equivalent and Approximate Equivalent. The new EN
designations included in the above tables are all near
equivalents of BS970-1 steel grades.
Alloy equivalentNickel based alloys
588

Composition (Nibalance unless otherwise stated)

Country Designation C Co Cr Mo V W Al Cu Nb Ta Ti Fe Other
21001 USA/UK Hastelloy C (UNS N10002 <.08 <2.5 14.516.5 1517 <.35 3.04.5 4.07.0 Si < 1.0 Mn < 1.0
G Wk 2.4537 <.02 2.5 15.5 16 0.35 3.7 6 Ni balance. Nominal
comp.
G Wk 2.4602 16 17 4 6 Ni balance. Nominal
comp.
21002 USA/UK Hastelloy X (UNS N06002) .05.15 .52.5 20.523.0 8.010.0 .21.0 1720 Si < 1.0 Mn < 1.0
G Wk. 2.4613
21003 USA/UK Incoloy 800 (UNS N08800) <.1 19.023.0 .15.6 .15.6 Bal. Ni 30.035.0
G X10 NiCrAITi 3220 .07 21 * * Ni 31.0 nominal comp.
UK Wk. 1.4876 BS 3072 NA <.1 19.023.0 .15.6 <.75 .15.6 Bal. Ni + Co 30.035.0
15
21004 USA/UK Incoloy 901 (UNS N09901) <.1 11.014.0 57 <.35 <.5 2.353.1 Bal. Ni 40.040.5 Mn < 1.0
Si < .6 B .01.02
21005 USA/UK Incoloy 903 (UNS N19903) 13.017.0 .31.5 2.43.5 1.01.25 Bal. Ni 36.040.0
21006 USA/UK Incoloy 907 (UNS N19907) 13.0 4.7 1.5 Bal. Ni 38.0 nominal comp.
21007 USA/UK Incoloy 909 (UNS N19909) .1 13.0 4.7 1.5 Bal. Ni 38.0 nominal comp.
21008 USA/UK Inconel 600 (UNS N06600) <.15 14.017.0 <.5 6.010.0 Ni > 72.0
UK BS3072 NA14 <.15 14.017.0 <.5 6.010.0 Ni + Co > 72.0
Mn < 1.0 Si < .5
G Wk. 2.4816 .05 16 10 Nominal comp.
21009 USA/UK Inconel 625 (UNS N06625) <.1 20.023.0 8.010.0 <.4 3.154.15 <.4 <5.0 Ni balance
21010 USA/UK Inconel 718 (UNS N07718) <.08 <1.0 17.021.0 2.83.0 .2.8 4.755.5 .651.15 Bal. Si < .35 Mn < .35 Ni
50.055.0
F/AECMA NI-P100HT
G Wk. 2.4666
CEN EN 2403PR (NI-P100HT Solution treated and precipitation treated precision castingsprovisional spec.)
EN 2404PR (NI-P100HT Solution treated and precipitation treated barsprovisional spec.)
EN 2405PR (NI-P100HT Solution treated and precipitation treated forgingsprovisional spec.)
EN 2407PR (NI-P100HT Solution treated and precipitation treated sheet and strip, a 3 mmprovisional spec.)
EN 2408PR (NI-P100HT Solution treated and precipitation treated plates, a 3 mmprovisional spec.)
EN 2952PR (NI-P100HT Solution treated and cold worked bar for hot upset forging for fasteners, 3 mm d 30 mmprovisional spec.)
EN 2961 PR (NI-P100HT Cold worked and solution treated bar for machining for fasteners, 3 mm d 50 mmprovisional spec.)
EN 3666PR (NI-P100 Cold workedRM 1500 MPabar for machining, 3 mm d 50 mmprovisional spec.)
(continued)
Appendix 6: Metal Alloy Comparison Tables
(continued)
Composition (Nibalance unless otherwise stated)
Country Designation C Co Cr Mo V W Al Cu Nb Ta Ti Fe Other
21011 USA/UK Inconel X750 (UNS <.08 14.017.0 .41.0 .71.2 2.252.75 5.09.0 Si < .5 Mn < 1.0 Ni
N07750) Balance
21012 USA/UK Monel K500 (UNS <.25 2.9 Bal. .35.85 <2.0 Si < .5 Mn < 1.5 Ni
N05500) 63.070.0
G Wk. 2.4360 31.0 2.0 Ni balance. Nominal
comp.
G Wk. 2.4374 3.0 30.0 .8 1.5 Ni balance. Nominal
comp.
G Wk. 2.4375 .25 3.0 30.0 1.0 2.0 Ni balance. Nominal
comp.
Appendix 6: Metal Alloy Comparison Tables

21013 USA/UK NI-SPAN C902 (UNS <.06 4.95.75 .3.8 2.22.75 Bal. Ni 41.043.5 Mn < .8
N09902) Si < 1.0
21014 USA/UK RENE 41 (UNS N07041) <.12 10.012.0 18.020.0 9.010.5 1.41.8 3.03.3 <5.0 Si < .5 Mn < .1 B
.003.010
21015 USA/UK UNITEMP 212 .08 16.0 .15 .5 4.0 Bal. Si .15 Mn .05 B .06 Zr
.05 Ni 25.0
21016 USA/UK WASPALLOY (UNS .03.1 12.015.0 18.020.0 3.55.0 1.21.6 2.753.25 <2.0 B .003.010 Zr .02.12
N07001)
F/AECMA NI-P101HT
CEN EN 2193PR (NI-P101 HT Solution treated and precipitation treated barsprovisional spec.)
EN 2194PR (NI-P101 HT Solution treated and precipitation treated forgingsprovisional spec.)
EN 2195PR (NI-P101HT Solution treated and precipitation treated sheet and strip, a 3 mmprovisional spec.)
EN 2406PR (NI-P101 HT Solution treated and precipitation treated bars for forged bolts, d 25 mmprovisional spec.)
EN 2959PR (NI-P101 HT Solution treated and cold worked bar for hot upset forging for fasteners, 3 mm d 30 mmprovisional spec.)
EN 2960PR (NI-P101 HT Cold worked and solution treated bar for machining for fasteners, 3 mm d 50 mmprovisional spec.)
Notes for Nickel alloys
Trade NamesUsually nominal compositions only available. Some alloys are balance iron, some balance nickel
589
Alloy equivalentsAluminium alloys (wrought)
590

Composition (max. unless otherwise stated) Others

Country Designation Si Fe Cu Mn Mg Cr Zn Ti Each Total Al min. Notes
31001 USA AA 1050 (UNS A91050) .25 .4 .05 .05 .05 .05 .03 .03 99.5 (1050A: Zn < .07)
UK AA 1050A (was BS 1B) .3 .4 .05 .03 .1 99.5 *35
UK BS 5L36 (AA 1050A) .3 .4 .05 .05 .10 0.03 99.5
F NF A-5 (AA 1050A) .3 .4 .05 .05 .03 .03 .1 .05 Rem.
G DIN 1712 AI99.5 (AA 1050A) Wk. 99.5
3.0255
CEN EN 2072 (1050A-H14 sheet and
strip);
EN 2073PR (1050A-H14 tube for structures, 5 mm < d < 100 mmprovisional spec.)
EN 2114PR (1050A-H14 wire for solid rivets, d 10 mmprovisional spec.)
31002 USA AA 1060 (UNS A91060) .25 .35 .05 .03 .03 .05 .03 .03 99.6
31003 USA AA 1100 (UNS A91100) 1.0 (Si+Fe) .05.2 .05 .1 .05 .15 99.0 *31
UK AA 1100 1.0 (Si+Fe) .05.2 .05 .1 .05 .15 99.0 *31
CEN EN 3996PR (1100-H14 sheet and strip, 0.3 mm a 6 mmprovisional spec.)
31004 USA AA 1145 (UNS A91145) .55 (Si+Fe) .05 .05 .03 99.45
UK AA 1145 .55 (Si+Fe) .05 .05 .03 99.45
31005 USA AA 1175 (UNS A91175) .15 (Si+Fe) .1 .03 .02 99.75 *32
UK AA 1080A (was BS 1A) .15 .15 .02 .03 .06 99.8 *33
F NF A8 (AA 1080A) .15 .15 .03 .03 .01 .02 .06 .05 .2 Rem.
G DIN 1712 AI99.7 Wk. 3.0275 .20 .25 .03 .07 .05 .02 .3 Rem.
G DIN 1712 AI99.8 (AA 1080A) Wk. .15 .15 .02 .06 .03 .01 .2 Rem.
3.0285
31006 USA AA 1200 (UNS A91200) 1.0 (Si+Fe) .05 .05 .1 .05 .05 .15 99.0
UK AA 1200 (was BS 1C) 1.0 (Si+Fe) .05 .05 .1 .05 .05 .15 99.0
UK BS 6L16; BS 6L17; BS 4L34 .5 .7 .1 .1 .1 .05 99.0
F NF A-4 .5 .8 .1 .1 .05 .05 .1 .05 Rem.
G DIN 1712 AI99 Wk. 3.0205 .5 .6 .07 .08 .05 .04 1.0 Rem.
31007 USA AA 1230 (UNS A91230) .7 (Si+Fe) .1 .03 .1 .05 99.3
UK AA 1230 .7 (Si+Fe) .1 .03 .1 .05 99.3
31008 USA AA 1235 (UNS A91235) .65 (Si+Fe) .05 .05 99.35
UK AA 1235 .65 (Si+Fe) .05 .05 99.35
31009 USA AA 1345 (UNS A91345) .3 .4 .1 .03 99.45
UK, F, G: See Table 31001 (AA 1050)
31010 USA AA 1350 (UNS A91350) .1 .4 .05 .01 .01 .05 +V .02 .03 .1 99.5 Ga < .03 B < .05
UK AA 1350 (was BS 1 E); BS2897 .1 .4 .05 .01 .01 .05 +V .02 .03 .1 99.5 Ga < .03 B < .05
F A 5L, A 5B 99.5
G DIN 1712 AI99.5 Wk. 3.0255 .3 .4 .05 .07 .05 .03 .5 Rem.
G DIN 1712 AI99.5 Wk. 3.0257
31011 USA 1420 .05 4.56.0 Li 1.92.3 Zr .08.15 Nominal comp.
31012 USA 1430 1.41.8 2.33.0 Li 1.51.9 Zr .08.14 Nominal comp.
31013 USA 1440 1.21.9 .61.1 Li 2.12.6 Zr .1.2 Nominal comp.
31014 USA 1460 2.63.3 .05 Li 2.02.5 Zr < .15 Sc < .14 Nominal comp.
31015 USA AA 2011 (UNS A92011) .4 .7 5.06.0 .3 .05 .15 Rem. *37
UK AA2011 (was BS FC1); BS4300/5; .4 .7 5.06.0 .3 .05 .15 Rem. *37
EN 515; EN 573-3; EN 573-4
F A-U5PbBi 5.5 Rem. Nominal comp.
G DIN 1725 Al Cu Bi Pb. Wk. 3.1655 .4 .7 5.06.0 .3 .05 .15 Rem. *37
31016 USA AA 2014 (UNS A92014) .51.2 .7 3.95.0 .41.2 .2.8 .1 .25 .15 .05 .15 Rem. *310*31
Appendix 6: Metal Alloy Comparison Tables

(continued)
(continued)
Composition (max. unless otherwise stated) Others
Country Designation Si Fe Cu Mn Mg Cr Zn Ti Each Total Al min. Notes
UK AA 2104A (was BS H15); BS .5.9 .5 3.95.0 .41.2 .2.8 .1 .2 +Zr .2 Rem. Ni .2, Pb .05, Sn .05
L102; BS L103;
BS L105; BS L156-L159; BS L163-L168;
BS 2L77;
BS 2L87; BS 2L93; BS 3L63; BS
7L37; DTD 5010A;
DTD 5030A; DTD5040A
F NFA-U4SG .51.2 .7 3.94.9 .41.2 .2.8 .1 .25 .2 Rem.
G DIN 1725 Al Cu Si Mn. Wk. 3.1255 .51.2 .7 3.95.0 .41.2 .2.8 .1 .25 .15 .05 .15 Rem.
CEN EN 2087PR (2014A-T6/T62 clad sheet and stripprovisional spec.)
EN 2088PR (2014A-T4/T42 clad sheet and stripprovisional spec.)
EN 2089 (2014A-T6 sheet and strip)
EN 2100 (2014A-T4511 bar and drawn proles)
Appendix 6: Metal Alloy Comparison Tables

EN 2323PR (2014A-T651 bar 200 mmprovisional spec)

EN 2324PR (2014A-T6 bar and section 150 mmprovisional spec.)
EN 2325PR (2014A-T6 bar 100 mmprovisional spec.)
EN 2384 (2014A-T6511 bar and drawn proles)
EN 2387PR (2014A-T6 tube for structures, 0.6 mm a 12.5 mmprovisional spec.)
EN 2395 (2014A-T4/T42 sheet and strip)
EN 2634PR (2014A-T4511 bars and sections 1.2 mm a/d 200 mm, peripheral coarse grain controlprovisional spec.)
EN 2635PR (2014A-T6511 bars and sections 1.2 mm a/d 150 mm, peripheral coarse grain controlprovisional spec.)
EN 2639PR (2014A-T6 extruded bars and sections 1.2 mm a/d 150 mm, peripheral coarse grain controlprovisional spec.)
EN 2710PR (2014A-T4510 bar and section, 1.2 mm a/d 200 mm, peripheral coarse grain controlprovisional spec.)
EN 2711 PR (2014A-T6510 bar and section, 1.2 mm a/d 150 mm, peripheral coarse grain controlprovisional spec.)
EN 3346PR (2014A-T3 tube for structures, 0.6 mm a 12.5 mmprovisional spec.)
31017 USA AA 2017 (UNS A92017) .2.8 .7 3.54.3 .41.0 .4.8 .1 .25 .15 .05 .15 Rem. *310
UK AA 2017 .2.8 .7 3.54.3 .41.0 .4.8 .1 .25 .15 .05 .15 Rem. *310
F NF A-U49 .3.8 .7 3.54.7 .3.8 .41.0 .1 .25 .2 Rem.
G DIN 1725 Al Cu Mg 1. Wk. 3.1325 .6 .5 3.54.3 .31.0 .41.0 .1 .5 .2 .05 .2 Rem.
CEN EN 2116PR (2017A-H13 wire for solid rivets, d 10 mmprovisional spec.)
EN 2393PR (2017A-T4 drawn tube for structures, 0.6 mm a 12.5 mmprovisional spec.)
EN 2509PR (2017A-T42 drawn tube for structuresprovisional spec.)
EN 2640PR (2017A-T4 extruded bars and sections 1.2 mm a/d 150 mm, peripheral coarse grain controlprovisional spec.)
EN 2655PR (2017A-T42 extruded bars and sections 1.2 mm a/d 150 mm, peripheral coarse grain controlprovisional spec.)
EN 2691 PR (2017A-T3 sheet and strip, 0.4 mm a 6 mmprovisional spec.)
EN 2692PR (2017A-T3 clad sheet and strip, 0.4 mm a 6 mmprovisional spec.)
EN 2705PR (2017A-T44 drawn tube for structures, 0.6 mm a 12.5 mmprovisional spec.)
(continued)
591
(continued)
592

Composition (max. unless otherwise stated) Others

Country Designation Si Fe Cu Mn Mg Cr Zn Ti Each Total Al min. Notes
31018 USA AA 2024 (UNS A92024) .5 .5 3.84.9 .3.9 1.21.8 .1 .25 .15 .05 .15 Rem. *310
UK AA 2024 .5 .5 3.84.9 .3.9 1.21.8 .1 .25 .15 .05 .15 Rem. *310
UK BS 2L97; DTD 5100A; BS .5 .5 3.84.9 .3.9 1.21.8 .1 .2 +Zr .2 Rem. Ni .05, Pb .05, Sn .05
AMD2433 (was 2L98)
F NF A-U4G1 .5 .5 3.84.5 .3.9 1.21.8 .1 .25 .2 Rem.
G DIN 1725 Al Cu Mg 2. Wk. 3.1355 .4 .4 4.04.8 .3.9 1.21.8 .1 .25 .2 .05 .2 Rem.
ISO ISO Al Cu 4 Mg 1
CEN EN 2090PR (2024-T3 clad sheet and strip, 0.4 mm < a < 6 mmprovisional spec.)
EN 2091 PR (2024-T4 clad sheet and strip, 0.4 mm < a < 6 mmprovisional spec.)
EN 2318 (2024-T3511 bar and drawn proles, a > 1.2 mm / d < 150 mm)
EN 2320PR (2024-T3 drawn bar, a 75 mmprovisional spec.)
EN 2321 PR (2024-T4 bar and section, a 150 mmprovisional spec.)
EN 2319PR (2024-T3510 drawn bar, a 75 mmprovisional spec.)
EN 2388PR (2024-T351 tube for structures, 0.6 mm a 12.5 mmprovisional spec.)
EN 2419PR (2024-T351 plate, 6 mm a 80 mmprovisional spec.)
EN 251OPR (2024-T42 drawn tube for structuresprovisional spec.)
EN 2633 (2024-T3511 bar and drawn proles, a > 1.2 mm / d < 150 mm, peripheral coarse grain control)
EN 2638PR (2024-T3 extruded bars and sections, 1.2 mm a/d 150 mm, peripheral coarse grain controlprovisional spec.)
EN 2701 PR (2024-T3 drawn tube, 6 mm d/a 12.5provisional spec.)
EN 2703PR (2024-T42 clad sheet and strip, 0.4 mm a 6 mmprovisional spec.)
EN 2704PR (2024-T3511 drawn bar, a 75 mmprovisional spec.)
EN 2709PR (2024-T3510 bar and section, 1.2 mm a/d 150 mm, peripheral coarse grain controlprovisional spec.)
EN 2806PR (2024-T42 extruded sections, 1.2 mm a 100 mm, peripheral coarse grain controlprovisional spec.)
EN 2814PR (2024-T3511 tube for structures, 0.6 mm a 12.5 mmprovisional spec.)
EN 3347PR (2024-T8511 extruded bars and sections, a/d 150 mm, peripheral coarse grain controlprovisional spec.)
EN 3348PR (2024-T62 plate, 6 mm a/d 50 mmprovisional spec.)
EN 3474PR (2024-T81 sheet and strip, 0.25 mm a/d 6 mmprovisional spec.)
EN 3550PR (2024-T8511 extruded barsand sections, a/d 150 mmprovisional spec.)
EN 3657PR (2024-T3510 drawn bar for machining, d 75 mmprovisional spec.)
EN 3997PR (2024-T3 sheet and strip, 0.4 mm a/d 6 mmprovisional spec.)
EN 3998PR (2024-T42 sheet and strip, 0.4 mm a/d 6 mmprovisional spec.)
EN 4101 PR (2024-T4 sheet and strip, 0.4 mm a/d 6 mmprovisional spec.)
31019 USA AA 2048 (UNS A92048) .15 .2 2.83.8 .2.6 1.21.8 .25
UK AA 2048 .15 .2 2.83.8 .26 1.21.8 .25
31020 USA AA 2090 2.52.75 017.02 Li 2.12.2 Zr .11.12 Nominal comp.
USA 2090 (proprietory) 2.43.0 .25 Li 1.92.6 Zr .08.15 Nominal comp.
UK AA 2090 2.52.75 .017.02 Li 2.12.2 Zr .11.12 Nominal comp.
31021 USA AA 2091 .2 .3 1.82.5 1.11.9 Li 1.72.3 Zr < .1 Nominal comp.
UK AA 2091 .2 .3 1.82.5 1.11.9 Li 1.72.3 Zr < .1 Nominal comp.
31022 USA Weldalite 049 (AA 2095) 4.06.3 .4 Li 1.3 Zr .14 Ag .4 Nominal comp.
31023 USA AA 2124 (UNS A92124) .2 .3 3.84.9 .3.9 1.21.8 .1 .25 .15 .05 .15 Rem. *310
UK AA 2124 .2 .3 3.84.9 .3.9 1.21.8 .1 .25 .15 .05 .15 Rem. *310
CEN EN 2422PR (2124T351 plate, 25 mm a < 120 mmprovisional spec.);
(continued)
Appendix 6: Metal Alloy Comparison Tables
(continued)
Composition (max. unless otherwise stated) Others
Country Designation Si Fe Cu Mn Mg Cr Zn Ti Each Total Al min. Notes
31024 USA AA 2195 3.74.3 .258 Li .81.2 Zr .08.16 Nominal comp.
31025 USA AA 2214 .8 4.5 .6 Nominal comp.
UK AA 2214 .8 4.5 .6 Nominal comp.
CEN EN 2124PR (2214-T651 plate, 6 mm a 140 mm)
EN 2382PR (2214-T6 forgings, 100 mm)
EN 2383PR (2214-T4 forgings, 100 mm)
EN 2485PR (2214-F Extruded or cast forging stock)
EN 2697PR (2214-T6 extruded bar and section, 1.2 a/d 100 mm, peripheral coarse grain controlprovisional spec)
31026 USA AA 2219 (UNS A92219) .2 .3 5.86.8 .2.4 .02 .1 .02.1 .05 .15 Rem. *311
UK AA 2219 .2 .3 5.86.8 .2.4 .02 .1 .02.1 .05 .15 Rem. *311
UK DTD 5004A .25 .35 5.76.5 .2.35 .15 .1 +Zr .2 Rem. *312
F NF A-U6MT .2 .3 5.56.5 .2.3 .05.15 Rem.
Appendix 6: Metal Alloy Comparison Tables

CEN EN 4099PR (2219-T62 clad sheet. and strip, 0.5 mm a 6 mmprovisional spec.)
EN 4100PR (2219-T62 sheet and strip, 0.5 mm a 6 mmprovisional spec)
EN 4102PR (2219-T81 clad sheet. and strip, 0.5 mm a 6 mmprovisional spec.)
31027 USA AA 2419 .015 .18 5.86.8 .2.4 .02 .10 .02.1 .05 .15 Rem. *321
UK AA 2419 .015 .18 5.86.8 .24 .02 .10 .02.1 .05 .15 Rem. *321
31028 USA AA 2618 .1.25 .91.3 1.92.7 1.31.8 .1 .04.1 .05 .15 Rem. *322
UK AA 2618A (was H16) .1.25 .91.4 1.82.7 .25 1.21.8 .15 .04.2 .05 .15 Rem. Ni .81.4
UK BS 1472; Hid RR58 .25 .91.4 1.82.7 .2 1.21.8 .2 .2 Rem. *322
UK DTD717A; 731B; 745A; 5084A; 5014A .25 .91.4 1.82.7 .2 1.21.8 .1 +Zr .2 Rem. Pb, Sn .05, Ni .81.4
F A-U2GN .25 .71.4 1.82.7 .2 1.21.8 .15 Rem. *325
CEN EN 2085PR (2618A-T6 forgings, 150 mm)
EN 2086PR (2618A-T851 [AL-P11-T851] forged bars and labs, 150 mm)
EN 2123PR (2618A-T851 plates, 6 mm a 140 mm)
EN 2256PR (2618A-T852 [AL-P11-T852] forged bars and slabs, 150 mm)
EN 2486PR (2618A-F extruded or cast forging stock)
EN 3552PR (2618A-T6 clad sheet and strip, 0.4 mm a 6 mmprovisional spec.)
EN 3553PR (2618A-T6511 extruded bar and section, 1 2 mm 100 mmprovisional spec.)
31029 USA AA 3003 (UNS A93003) .6 .7 .05.2 1.01.5 .1 .03 .15 Rem. *31
UK AA 3003 (was BS 1470 NS3) .6 .7 .1 .81.5 .1 .2 .2 Rem.
F NF A-MI .6 .7 .2 1.01.5 .05 .1 .5 Rem.
G DIN 1725 Al Mn Wk. 3.0515 .5 .6 .1 .91.4 0.3 .05 .2 .1 .05 .15 Rem.
31030 USA AA 3004 (UNS A93004) .3 .7 .25 1.01.5 .81.3 .25 05 .15 Rem. *31
UK AA 3004 .3 .7 .25 1.01.5 .81.3 .25 05 .15 Rem. *31
F NF A .3 .7 .25 1.01.5 .81.3 .25 .5 Rem.
31031 USA AA 3005 (UNS A93005) .6 .7 .3 1.01.5 .2.6 .1 .25 .1 .05 .15 Rem.
UK AA 3005 .6 .7 .3 1.01.5 .2.6 .1 .25 .1 .05 .15 Rem.
31032 USA AA 3103 (UNS A93103) .5 .7 .1 .91.5 .3 .1 .2 +Zr .1 .05 .15 Rem.
UK AA 3103 .5 .7 .1 .91.5 .3 .1 .2 +Zr .1 .05 .15 Rem.
CEN EN 4004PR (3103-H16 sheet and strip, 0.4 mm a 6 mmprovisional spec.)
31033 USA AA 3105 (UNS A93105) .6 .7 .3 .3.8 .2.8 .2 .4 .1 .05 .15 Rem.
UK AA 3105; ALCAN E4S .6 .7 .3 .3.8 .2.8 .2 .4 .1 .05 .15 Rem.
UK BS N31 (old designation) .25 .7 .25 .41.1 .3.6 .1 .2 .2 Rem.
G DIN 1725 Al Mn 0.5 Mg 0.5 Wk.
3.0505
(continued)
593
(continued)
594

Composition (max. unless otherwise stated) Others

Country Designation Si Fe Cu Mn Mg Cr Zn Ti Each Total Al min. Notes
31034 USA AA 4032 (UNS A94032) 11.013.5 1.0 .51.3 .81.3 .1 .25 .05 .15 Rem. *313
UK AA 4032 11.013.5 1.0 .51.3 .81.3 .1 .25 .05 .15 Rem. *313
UK ALCAN GB38S 10.513.0 .6 .71.3 .2 .81.5 .1 .2 Rem. *314
F NF A-S12UN 10.512.5 .71.3 .81.5 .15 Rem. *314
31035 USA AA 5005 (UNS A95005) .3 .7 .2 .2 .51.1 .1 .25 .05 .15 Rem.
UK AA 5005 .3 .7 .2 .2 .51.1 .1 .25 .05 .15 Rem.
UK BS N41 (old designation) .4 .7 .2 .5 .51.2 .1 .2 .2 Rem.
F NF A-G0.6 .4 .7 .2 .2 .51.1 .1 .2 .5 Rem.
G DIN 1725 Al Mg 1. Wk. 3.3315 .3 .4 .05 .2 .81.2 .1 .2 .1 .05 .15 Rem.
31036 USA AA 5050 (UNS A95050) .4 .7 .2 .1 1.11.8 .1 .25 .05 .15 Rem. *31
UK AA 5050 .4 .7 .2 .1 1.11.8 .1 .25 .05 .15 Rem. *31
F NF A-G1 .4 .7 .2 .7 1.01.8 .1 .25 .05 Rem.
31037 USA AA 5052 (UNS A95052) .45 (Si+Fe) .1 .1 2.22.8 .15.35 .1 .05 .15 Rem. *31
USA AMS 4015E; AMS 4016E; AMS
4017E;
AMS 4069; AMS 4070F; AMS 2.5 .25 Nominal comp.
4071F; AMS 4114B
UK AA 5052 .45 (Si+Fe) .1 .1 2.22.8 .15.35 .1 .05 .15 Rem. *31
ISO ISO Al Mg2
CEN EN 4005PR (5052-O sheet and strip, 0.3 mm a 6 mmprovisional spec.)
31038 USA AA 5056 (UNS A95056) .3 .4 .1 .05.2 4.55.6 .05.2 .1 .05 .15 Rem. *31
UK AA 5056A; BS 3L58 .4 .5 .1 .1.6 4.55.6 .2 .2 .2 .05 .15 Rem. Mn + Cr .1.6
UK BS N6 (old designation) .3 .5 .1 .5 4.55.3 .25 .2 .2 Rem. Mn + Cr .1.5
F A-G5 M (AA 5056A) .4 .4 .05 .55 4.35.5 .3 .2 .1 .05 .15 Rem.
G DIN 1725 Al Mg 5. (AA 5056A) .4 .4 .05 .55 4.35.5 .3 .2 .1 .05 .15 Rem.
Wk. 3.3555
CEN EN 2117PR (5056A-H32 wire for solid rivets, d 10 mmprovisional spec.)
EN 2628PR (5056A-O wire for solid rivets, d 10 mmprovisional spec.)
31039 USA AA 5083 (UNS A95083) .4 .4 .1 .41.0 4.04.9 .05.25 .25 .15 .05 .15 Rem.
UK AA 5083 .4 .4 .1 .41.0 4.04.9 .05.25 .25 .15 .05 .15 Rem.
UK BS N8 .4 .4 .1 .51.0 4.04.9 .25 2 .15 Rem.
F A-G4.5MC .7 4.4 .1 Nominal comp.
G DIN 1725 Al Mg 4.5 Mn Wk. 3.3547 .4 .4 .1 .61.0 4.04.9 .05.25 .2 .1 .05 .15 Rem.
31040 USA AA 5086 (UNS A95086) .4 .5 .1 .2.7 3.54.5 .05.25 .25 .15 .05 .15 Rem.
UK AA 5086 .4 .5 .1 .2.7 3.54.5 .05.25 .25 .15 .05 .15 Rem.
CEN EN 2508PR (5086-H111 drawn tube for sturcturesprovisional spec.)
EN 2693 (5086-H 111 sheet and strip)
EN 2699PR (5086-H111 drawn bar, 6 mm d 50 mmprovisional spec.)
31041 USA AA 5154 (UNS A95154) .45 (Si+Fe) .1 .1 3.13.9 .15.35 .2 .2 .05 .15 Rem. *31
UK AA5154A (was BS N5) .5 .5 .1 .5 3.13.9 .25 .2 .2 .05 .15 Rem. Mn + Cr .1.5 *31
G DIN 1725 Al Mg 3. Wk. 3.3535 .4 .4 .05 .5 2.634 .3 .2 .1 .05 .15 Rem.
31042 USA AA 5251 (UNS A95050) .4 .5 .15 .1.5 1.72.4 .15 .15 .15 .05 .15 Rem.
UK AA 5251; BS 5L44; BS 3L80; BS .4 .5 .15 .1.5 1.72.4 .15 .15 .15 .05 .15 Rem.
3L81
UK BS N4 (old designation) .5 .5 .1 .5 1.72.4 .25 .2 .2 Mn + Cr .5 Rem.
F A-G2 M 2 Nominal comp.
G DIN 1725 Al Mg 2. Mn 0.3 Wk. .3 .4 .05 .3 1.72.4 .3 .2 .1 .05 .15 Rem.
3.3525
Appendix 6: Metal Alloy Comparison Tables

(continued)
(continued)
Composition (max. unless otherwise stated) Others
Country Designation Si Fe Cu Mn Mg Cr Zn Ti Each Total Al min. Notes
31043 USA AA 5252 (UNS A95252) .08 .1 .1 .1 2.22.8 .03 .1 Rem.
UK AA 5252 .08 .1 .1 .1 2.22.8 .03 .1 Rem.
F NF AG-G3 3.0 nom High purity base
31044 USA AA 5254 (UNS A95254) .45 (Si+Fe) .05 .01 3.13.9 .15.35 .2 .05 .05 15 Rem.
UK AA 5154A (was BS N5) .5 .5 .1 .5 3.13.9 .25 .2 .2 .05 .15 Rem. Mn + Cr .1.5 *31
G DIN 1725 Al Mg3. Wk. 3.3535 .4 .4 .05 .5 2.63.4 .3 .2 .1 .05 .15 Rem.
31045 USA AA 5356 (UNS A95356) .5 (Si+Fe) .1 .05.2 4.55.5 .05.2 .1 .06.2 .05 .15 Rem. *31
UK AA 5356 .5 (Si+Fe) .1 .05.2 4.55.5 .05.2 .1 .06.2 .05 .15 Rem. *31
UK AA 5056A; BS 3L58 .4 .5 .1 1.6 4.55.6 .2 .2 .2 .05 .15 Rem. Mn + Cr .1.6
UK BS N6 (old designation) .3 .5 .1 .5 4.55.3 .25 .2 .2 Mn + Cr .1.5 Rem.
F A-G5 M .4 .4 .05 .55 4.35.5 .3 .2 .1 .05 .15 Rem.
G DIN 1725 Al Mg 5. Wk. 3.3555 .4 .4 .05 .55 4.35.5 .3 .2 .1 .05 .15 Rem.
Appendix 6: Metal Alloy Comparison Tables

31046 USA AA 5454 (UNS A95454) .4 (Si+Fe) .1 .51.0 2.43.0 .05.2 .25 .2 .05 .15 Rem.
UK AA 5454 (was BS N51); EN 515; EN .4 (Si+Fe) .1 .51.0 2.43.0 .05.2 .25 .2 .05 .15 Rem.
573-3;
EN 573-4
F A-G2.5MC .7 2.7 .1 Nominal comp.
F A-G3 .4 .5 .1 .1.6 2.63.8 .4 .2 .2 Rem.
G DIN 1725 Al Mg 2.7 Wk. 3.3537 .45 2.7 Nominal comp.
G DIN 1725 Al Mg 3. Wk. 3.3585 .4 .4 .05 .5 2.63.4 .3 .2 .1 .05 .15 Rem.
31047 USA AA 5456 (UNS A95456) .4 (Si+Fe) .1 .51.0 4.75.5 .05.2 .25 .2 .05 .15 Rem.
UK BS N61 .4 (Si+Fe) .1 .61.0 5.05.5 .05.2 .2 .05.2 Rem.
F NF A-G5 .4 .5 .1 .21.0 4.55.5 .4 .2 .2 Rem.
G DIN 1725 Al Mg 5. Wk. 3.3555 .4 .4 .05 .55 4.35.5 .3 .2 .1 .05 .15 Rem.
31048 USA AA 5457 (UNS A95457) .08 .1 .2 .15.45 .81.2 .03 .03 .1 Rem.
UK AA 5457 (UNS A95457) .08 .1 .2 .15.45 .81.2 .03 .03 .1 Rem.
F A9-G1 and see AA 5005 1.0 Nominal comp.
31049 USA AA 5652 (UNS A95652) .4 (Si+Fe) .04 .01 2.22.8 .15.35 .1 .05 .15 Rem. *31
UK AA 5652 .4 (Si+Fe) .04 .01 2.22.8 .15.35 .1 .05 .15 Rem. *31
31050 USA AA 5657 (UNS A95657) .08 .1 .1 .03 .61.0 .03 .02 .05 Rem. *31
UK AA 5657 .08 .1 .1 .03 .61.0 .03 .02 .05 Rem. *31
UK BS BTRS2 (old designation) 1.0 Nominal comp.
31051 USA AA 6003 (UNS A96003) .351.0 .6 .1 .8 .81.5 .35 .2 .1 .05 .15 Rem.
UK AA 6003 .351.0 .6 .1 .8 .81.5 .35 .2 .1 .05 .15 Rem.
F NF ASGM .61.5 .5 .1 .11.0 .61.5 .3 .25 .2 Rem.
G DIN 1725 Al Mg Si 1 .751.3 .5 .1 .41.0 .61.2 .3 .2 .1 .05 .15 Rem.
31052 USA AA 6005 (UNS A96005) .6.9 .35 .1 .1 .4.6 .1 .1 .1 .05 .15 Rem. k
UK AA 6005 .6.9 .35 .1 .1 .4.6 .1 .1 .1 .05 .15 Rem.
31053 USA AA 6053 (UNS A96053) *315 35 .1 1.11.4 .15.35 .1 .05 .15 Rem.
UK AA 6053 *315 .35 .1 1.11.4 .15.35 .1 .05 .15 Rem.
See also AA 6003
(continued)
595
(continued)
596

Composition (max. unless otherwise stated) Others

Country Designation Si Fe Cu Mn Mg Cr Zn Ti Each Total Al min. Notes
31054 USA AA 6061 (UNS A96061) .4.8 .7 .15.4 .15 .81.2 .04.35 .25 .15 .05 .15 Rem.
USA AMS4025D; AMS4026D; AMS numbers for various forms and conditions of AA 6061 alloy
AMS4027E; AMS4043
AMS4053; AMS4079; AMS4080E;
AMS4081A
AMS4082E; AMS4083D;
AMS4115; AMS4116A
AMS4117A; AMS4127B;
AMS4146; AMS4150C; AMS4160;
AMS4161
UK AA 6061; BS L117; BS L118 .4.8 .7 .15.4 .15 .81.2 .04.35 .25 .15 .05 .15 Rem.
UK BS H20 .4.8 .7 .15.4 .2.8 .81.2 * .2 .2 Rem. *Either Mn or Cr .04.35
F A-GSUC .6 .2 1.0 .15 Nominal Camp.
G DIN 1725 Al Mg Si 1 Cu Wk. 3.3211
CEN EN 2391 PR (6061-T4 tube for structures, 0.6 mm a 12.5 mmprovisional spec.)
EN 2392PR (6061-T6 tube for structures, 0.6 mm a 12.5 mmprovisional spec)
EN 2629PR (6061provisional spec.)
EN 2694 (6061-T6/T62 sheet and strip)
EN 2700PR (6061-T6 drawn bar, 6 mm d 75 mm, peripheral coarse grain controlprovisional spec.)
EN 2702PR (6061-T6 extruded bar and section, 1.2 mm a/d 150 mmprovisional spec)
EN 2813PR (6061-T6 tube for hydraulics, 0.6 mm a 12.5 mmprovisional spec.)
EN 3341 PR (6061-T4 sheet and strip, 0.4 mm a 6 mmprovisional spec.)
EN 3342PR (6061-T4 drawn bar and section, 10 mm d 150 mmprovisional spec.)
EN 3702PR (6061-T4 tube for hydraulics, 0.6 mm a 12.5 mmprovisional spec.)
31055 USA AA 6063 (UNS A96063) .2.6 .35 .1 .1 .45.9 .1 .1 .1 .05 .15 Rem.
UK AA 6063; DTD 372B .2.6 .35 .1 .1 .45.9 .1 .1 .1 .05 .15 Rem.
UK H9 (old designation) .3.7 .4 .1 .1 .4.9 .1 .2 .2 Rem.
F NF A-GS (NFA.57.350) .8 .8 Nominal comp.
G DIN 1725 Al Mg Si 0.5 Wk. 3.3206 .35.8 .3 .05 .1 .4.8 .05 .2 .1 .05 .15 Rem.
ISO ISO Al Mg Si
31056 USA AA 6066 (UNS A96066) .91.8 .5 .71.2 .61.1 .81.4 .4 .25 .2 .05 .15 Rem.
UK AA 6066 .91.8 .5 .71.2 .61.1 .81.4 .4 .25 .2 .05 .15 Rem.
UK ALCAN 623; BS 2L84 .81.3 .7 1.02.0 1.0 .51.2 .2 .3 Rem. Ni .2 *316
31057 USA AA 6070 (UNS A96070) 1.01.7 .5 .15.4 .41.0 .51.2 .1 .25 .15 .05 .15 Rem.
UK AA 6070 1.01.7 .5 .15.4 .41.0 .51.2 .1 .25 .15 .05 .15 Rem.
UK ALCAN 623; BS 2L84 .81.3 .7 1.02.0 1.0 .51.2 .2 .3 Rem. Ni .2 *316
31058 USA AA 6081
CEN EN 2695 (6081-T6 sheet and strip);
31059 USA AA 6082 (UNS A96082) .71.3 .5 .1 .41.0 .61.2 .25 .2 .1 .05 .15 Rem.
UK AA 6082 (was H30) .71.3 .5 .1 .41.0 .61.2 .25 .2 .1 .05 .15 Rem.
ISO ISO AISi1 MgMn
CEN EN 2326 (6082-T6 <200 mm bar and drawn proles)
EN 2636 (6082-T6 <200 mm bar and drawn proles, peripheral coarse grain control)
EN 2389PR (6082-T4 tube for structures, 0.6 mm a 12.5 mmprovisional spec.)
EN 2390PR (6082-T6 tube for structures, 0.6 mm a 12.5 mmprovisional spec.)
EN 2420PR (6082-T6 barsprovisional spec.)
EN 2421 PR (6082-T4 wire for rivetsprovisional spec.)
EN 4006PR (6082-T4/T42 sheet and strip, 0.4 mm a 6 mmprovisional spec.)
Appendix 6: Metal Alloy Comparison Tables

EN 4007PR (6082-T6/T62 sheet and strip, 0.4 mm a 6 mmprovisional spec.)

(continued)
(continued)
Composition (max. unless otherwise stated) Others
Country Designation Si Fe Cu Mn Mg Cr Zn Ti Each Total Al min. Notes
31060 USA AA 6101 (UNS A96101) .3.7 .5 .1 .03 .35.8 .03 .1 .03 .1 Rem. B .06
UK AA6101A; BS2898:6101A .3.7 .4 .05 .4.9 .03 .1 Rem.
UK BS 91E .3.7 .5 .04 .4.9 Rem.
G DIN 1725 E-AI Mg Si Wk. 3.2305 .5.6 .1.3 .02 .3.5 .1 .03 .1 Rem. *317
31061 USA AA 6151 (UNS A96151) .61.2 1.0 .35 .2 .45.8 .15.35 .25 .15 .05 .15 Rem.
UK AA 6151 .61.2 1.0 .35 .2 .45.8 .1535 .25 .15 .05 .15 Rem.
See AA 6101
31062 USA AA 6162 (UNS A96162) .4.8 .5 .2 .1 .11.1 .1 .25 .1 .05 .15 Rem.
UK AA 6162 .4.8 .5 .2 .1 .11.1 .1 .25 .1 .05 .15 Rem.
See AA 6101
31063 USA AA 6201 (UNS A96201) .5.9 .5 .1 .03 .6.9 .03 .1 .03 .1 Rem.
UK AA 6201 .5.9 .5 .1 .03 .6.9 .03 .1 .03 .1 Rem.
Appendix 6: Metal Alloy Comparison Tables

See AA 6101
31064 USA AA 6253 (UNS A96253) *315 .5 .1 1.01.5 .15.35 1.62.4 .05 .15 Rem.
UK AA 6253 *315 .5 .1 1.01.5 .15.35 1.62.4 .05 .15 Rem.
31065 USA AA 6262 (UNS A96262) .4.8 .7 .15.4 .15 .81.2 .04.14 .25 .15 .05 .15 Rem. *318
UK AA 6262 .4.8 .7 .15.4 .15 .81.2 .04.14 .25 .15 .05 .15 Rem. *318
31066 USA AA 6351 (UNS A96351) .071.3 .5 .1 .4.8 .4.8 .2 .2 .05 .25 Rem.
UK AA 6351 .071.3 .5 .1 .4.8 .4.8 .2 .2 .05 .25 Rem.
31067 USA AA 6463 (UNS A96463) .2.6 .15 .2 .05 .45.9 .05 .15 Rem.
UK AA 6463; E6; EN 515; EN 5733; .2.6 .15 .2 .05 .45.9 .05 .15 Rem.
EN 573-4
UK BS BTR6 .2.5 .15 .2 .05 .4.8 .05 Rem.
31068 USA AA 6951 (UNS A96951) .2.5 .8 .15.4 .1 .4.8 .2 .05 .15 Rem.
31069 USA AA 7001 (UNS A97001) .35 .4 1.62.6 .2 2.63.4 .18.35 6.88.0 .2 .05 .15 Rem.
UK AA 7001 .35 .4 1.62.6 .2 2.63.4 .18.35 6.88.0 .2 .05 .15 Rem.
F NF A-Z8GU 1.6 2.7 .2 8.0 Nominal comp.
31070 USA AA 7009 (UNS A97009)
UK AA 7009
CEN EN 2093 (7009-T74 Forgings >20 mm and <150 mm)
EN 2094 (7009-T74 Die Forgings >3 mm and <150 mm)
EN 2381 (7009-T7452 Forgings >40 mm and <150 mm)
EN 2385 (7009-T74511 bar and drawn proles)
EN 2487PR (7009-F extruded or casr forging stockprovisional spec.)
EN 2630 (7009-T74511 bar and drawn proles, a 125 mm, peripheral coarse grain control)
EN 2706PR (7009-T736510 bar and section, 1.2 mm a/d 125 mm, peripheral coarse grain controlprovisional spec.)
(continued)
597
(continued)
598

Composition (max. unless otherwise stated) Others

Country Designation Si Fe Cu Mn Mg Cr Zn Ti Each Total Al min. Notes
31071 USA AA 7010 1.7 2.4 6.3 Nominal comp.
UK AA 7010 1.7 2.4 6.3 Nominal comp.
UK DTD 5120 .12 .15 1.52 .1 2.12.6 .05 5.76.7 .05 .15 Rem. *323
UK DTD 5130A .1 .15 1.52 .3 2.22.7 .05 5.76.7 .05 .15 Rem. *323
UK DTD 5636 .12 .15 1.52 .1 2.12.6 .05 5.76.7 .06 .05 .15 Zr .1.16
CEN EN 2681 PR (7010-T736 die forgings, a 150 mmprovisional spec.)
EN 2682PR (7010-T73652 forgings, 50 mm a 150 mmprovisional spec.)
EN 2683PR (7010-T7651 forgings, 80 mm a 160 mmprovisional spec.)
EN 2684PR (7010-T7651 plate, 6 mm a 140 mmprovisional spec.)
EN 2685PR (7010-T7652 forgings, 80 mm a 160 mmprovisional spec.)
EN 2686PR (7010-T73651 hand forgings, 50 mm a 150 mmprovisional spec.)
EN 2687PR (7010-T73651 plate, 6 mm a 150 mmprovisional spec.)
EN 3337PR (7010-T74511 extruded bars and sections a/d 130 mm, peripheral coarse grain controlprovisional spec)
EN 3339PR (7010-T76 die forgings, a 200 mmprovisional spec.)
EN 3343PR (7010-T76511 extruded bars and sections 1 mm a/d 130 mm, peripheral coarse grain controlprovisional spec)
EN 3554PR (7010-T7652 hand forgings, a 200 mmprovisional spec.)
31072 USA AA 7020 .35 .4 .2 .05.5 1.01.4 .1 35 4.05.0 .05 .15 Rem. *324
UK AA 7020; BS4300/14 /15; EN 515; .35 .4 .2 .05.5 1.01.4 .1.35 4.05.0 .05 .15 Rem. *324
EN 573-3; EN 573-4
UK BS H17 .4 .4 .25 .2.7 1.01.5 .25 3.84.8 .1 Rem. Zr .25, Mn + Cr .7
F A-Z5G (NFA.57-702) .3 .8 .15.35 .4 .4.65 .35 4.55.5 .15.25 .05 .15 Rem.
G DIN 1725 Al Zn 4.5 Mg 1 Wk. .5 .5 .1 1.5 11.4 .2 45 .2 .05 .15 Rem.
3.4335
CEN EN 2807PR (7020-T6 extruded sections 1.2 mm a 100 mm, peripheral coarse grain controlprovisional spec.)
31073 USA AA 7039 (UNS A97039) .3 .4 .1 .1.4 2.33.3 .15.25 3.54.5 .1 .05 .15 Rem.
UK AA 7039 .3 .4 .1 .1.4 2.33.3 .15.25 3.54.5 .1 .05 .15 Rem.
31074 USA AA 7049 (UNS A97049) .25 .35 1.21.9 .2 2.02.9 .1.22 7.28.2 .1 .1 .15 Rem.
UK AA 7049 .25 .35 1.21.9 .2 2.02.9 .1.22 7.28.2 .1 .1 .15 Rem.
31075 USA AA 7050 (UNS A97050) .12 .15 2.02.6 .1 1.92.6 .04 5.76.7 .06 .05 .15 Rem. Zr.08.15
UK AA 7050 .12 .15 2.02.6 .1 1.92.6 .04 5.76.7 .06 .05 .15 Rem. Zr.08.15
CEN EN 2688PR (7050-T736 die forgings, a 150 mmprovisional spec.)
EN 2689PR (7050-T73651 plate, 6 mm a 150 mmprovisional spec.)
EN 2690PR (7050-T73652 hand forgings, a 125 mmprovisional spec.)
EN 3334PR (7050-T651 plate, 6 mm a 60 mmprovisional spec)
EN 3338PR (7050-T74511 extruded bars and sections a/d 130 mm, peripheral coarse grain controlprovisional spec.)
EN 3340PR (7050-T76 die forgings, a 200 mmprovisional spec.)
EN 3344PR (7050-T76511 extruded bars and sections a/d 130 mm, peripheral coarse grain controlprovisional spec.)
(continued)
Appendix 6: Metal Alloy Comparison Tables
(continued)
Composition (max. unless otherwise stated) Others
Country Designation Si Fe Cu Mn Mg Cr Zn Ti Each Total Al min. Notes
31076 USA AA 7075 (UNS A97075) .4 .5 1.2-2.0 .3 2.12.9 .18.35 5.16.1 .2 .05 .15 Rem. Zr + Ti .25
UK AA 7075 .4 .5 1.22.0 .3 2.12.9 .18.35 5.16.1 .2 .05 .15 Rem. Zr + Ti .25
UK DTD5074A (now DTD5121) 1.6 2.5 .16 6.2 Nominal comp.
UK DTD5121 (part superceded by .4 .5 1.22.0 .3 2.12.9 .1.25 5.16.4 +Zr .2 Rem. Ni, Pb and Sn < .05
L170); DTD5110
UK BS L160; BS L161; BS L162; BS .4 .5 1.22.0 .3 2.12.9 .18.28 5.16.1 .2 .05 .15 Rem. Zr + Ti .25
L170
F NF A-Z5GU .4 .5 1.22.0 .1.9 2.03.5 .35 5.06.5 .2 Rem.
G DIN 1725 Al Zn Mg Cu 1.5 Wk. .5 .7 1.22.0 .3 2.12.9 .18.35 5.16.1 .2 .05 .15 Rem.
3.4365
CEN EN 2092 (7075-T6/T62 .46 mm sheet and strip)
EN 2126 (7075-T651 680 mm sheet)
EN 2127 (7075-T73511 <100 mm bar and drawn proles)
Appendix 6: Metal Alloy Comparison Tables

EN 2128 (7075-T7351 675 mm drawn bars)

EN 2315PR (7075-T73510/T73511 bars and sections 100 mmprovisional spec.)
EN 2316PR (7075-T73 bars and sections 100 mmprovisional spec.)
EN 2317PR (7075-T73 drawn bars 75 mmprovisional spec.)
EN 2380PR (7075-T73 forgings 125 mmprovisional spec.)
EN 2386PR (7075-T7352 hand forgings 150 mmprovisional spec.)
EN 2394PR (7075-T6511 bars and sections 125 mmprovisional spec.)
EN 2488PR (7075-F extruded or cast forging stockprovisional spec.)
EN 2511 PR (7075-T7351 plate, 6 mm a 100 mmprovisional spec.)
EN 2631 PR (7075-T6511 bars and sections 1.2 mm a/d 125 mm, peripheral coarse grain controlprovisional spec.)
EN 2632 (7075-T73511 <100 mm bar and drawn proles, <100 mm, controlled grain size)
EN 2637PR (7075-T73 extruded bars and sections 1.2 mm a/d 100 mm, peripheral coarse grain controlprovisional spec)
EN 2696 (7075-T6/T62 .4-6 mm sheet and strip)
EN 2698PR (7075-T6510 extruded bar and section, 1.2 mm a/d 100 mmprovisional spec.)
EN 2707PR (7075-T6510 bar and section, 1.2 mm a/d 125 mm, peripheral coarse grain controlprovisional spec.)
EN 2708PR (7075-T73510 bar and section, 1.2 mm a/d 100 mm, peripheral coarse grain controlprovisional spec.)
EN 2804PR (7075-T7651 plate, 6 mm a 25 mmprovisional spec.)
EN 3555PR (7075-T79510 extruded bar and section, 1.2 mm a/d 100 mm, coarse grain controlprovisional spec.)
31077 USA AA 7079 (UNS A97079) .3 .4 .4.8 .1.3 2.9-3.7 .1.25 3.84.8 .1 .05 .15 Rem.
UK AA 7079 .3 .4 .4.8 .1.3 2.93.7 .1.25 3.84.8 .1 .05 .15 Rem.
UK BS H17 .4 .4 .25 2.7 1.01.5 .25 3.84.8 .1 Rem. Zr .25, Mn + Cr .7 *319
G DIN 1725 Al Zn Mg Cu 0.5 Wk. .5 .5 .51.0 .1.4 2.63.6 .1.3 4.35.2 .2 .05 .15 Rem.
3.4345
31078 USA AA 7175 (UNS A97175) .15 .2 1.22.0 .1 2.12.9 .18.28 5.16.1 .1 .05 .15 Rem.
UK AA 7175 .15 .2 1.22.0 .1 2.12.9 .18.28 5.16.1 .1 .05 .15 Rem.
CEN EN 2512PR (7175-T7351 plate, 6 mm 100 mmprovisional spec.);
31079 USA AA 7178 (UNS A97178) .4 .5 1.62.4 .3 2.43.1 .18.35 6.37.3 .2 .05 .15 Rem.
UK AA 7178 .4 .5 1.62.4 .3 2.43.1 .18.35 6.37.3 .2 .05 .15 Rem.
F NF A-Z5GU .4 .5 1.22.0 1.9 2.03.5 .35 .5.06.5 .2 Rem.
31080 USA AA 8090 .2 .3 1.01.6 .61.3 Li 2.12.7 Zr < .16 Nominal comp.
UK AA 8090 .2 .3 1.01.6 .61.3 Li 2.12.7 Zr < .16 Nominal comp.
31081 USA AA 8091 .3 .5 1.82.2 .51.2 Li 2.42.8 Zr < .16 Nominal comp.
UK AA 8091 .3 .5 1.82.2 .51.2 Li 2.42.8 Zr < .16 Nominal comp.
599
600 Appendix 6: Metal Alloy Comparison Tables

Alloy equivalentsAluminium alloys (cast)

Composition (max. unless otherwise stated) Others
Country Designation Si Fe Cu Mn Mg Cr Zn Ti Each Total Al Notes
min.
32001 USA AA 295.0 (UNS .71.5 1.0 4.05.0 .35 .03 .35 .25 .05 .15
A02950)
32002 USA AA B295.0 2.03.0 1.2 4.05.0 .35 .05 .35 .5 .25 .35
UK AA 295.0 .71.5 1.0 4.05.0 .35 .03 .35 .25 .05 .15
UK BS LM 11 .25 .25 4.05.0 .1 .1 .1 .1 .3 Sn .05 Pb .05
UK BS L154; BS L155 1.01.5 .25 3.84.5 .1 .1 .1 .1 .05.25 .05 .15 Sn .05 Pb .05
F NF A-USG7 .3 .35 4.25.0 .1 .15.33 .05 .1 .3 Pb .05
G DIN 1725 Al Cu .18 .2 4.24.9 .05 .15.3 .07 .15.3 .03 .1
4 Ti Mg. Wk. 3.1371
32003 USA AA 319.0 (UNS 5.56.3 1.0 3.04.0 .5 .1 .35 1.0 .25 .5
A03190)
32004 USA AAA319.0 5.56.5 1.0 3.04.0 .3 .1 .35 3.0 .25 .5
UK AA 319.0 5.56.3 1.0 3.04.0 .5 .1 .35 1.0 .25 .5
UK BS LM4 4.06.0 .8 2.04.0 .2.6 .2 .3 .5 .2 Sn .1 Pb .1
UK BS LM22 4.06.0 .7 2.83.8 .3.6 .05 .15 .15 .2 Sn .05 Pb .1
F NF A-S5U; NF A- 5.0 3.2 Nominal comp.
S5U3
G DIN 1725 Al Si 6 5.07.5 1.0 3.05.0 .1.3 .1.3 .3 2.0 .15 .03 .15
Cu 4. Wk. 3.2151
ISO ISO AI-Si5Cu3
32005 USA AA 333.0 (UNS 8.010.0 1.0 3.04.0 .5 05.5 .5 1.0 .25 .5
A03330)
32006 USA AA A333.0 8.010.0 1.0 3.04.0 .5 .05.3 .5 3.0 .25 .5
UK AA 333.0 8.010.0 1.0 3.04.0 .5 05.5 .5 1.0 .25 .5
UK BS LM24 7.59.5 1.3 3.04.0 .5 .1 .5 3.0 .2 Sn .2 Pb .3
F A-S9U3Y4 8.2 3.5 Nominal comp.
F A-S10U4 9.011.0 1.3 3.04.0 .3 .5 .5 .8 .2 Sn .1
Fe + Zn +
Mg + Ni + Sn
2.5
G DIN 1725 Al Si 8 Cu 7.59.5 .8 2.03.5 .2.5 .3 .3 1.2 .15 .05 .15 Sn .1
3. Wk. 3.2161
ISO ISO AI-Si8Cu3Fe
32007 USA AA 355.0 (UNS 4.55.5 .6 1.01.5 .5 .4.6 .25 .35 .25 .05 .15
A03550)
32008 USA AA C355.0 (UNS 4.55.5 .2 1.01.5 .1 .4.6 .1 .2 .05 .15
A33550)
UK BS LM16 4.55.5 .6 1.01.5 .5 .4.6 .25 .1 .2 Sn .1 Pb .1
UK AA 355.0 4.55.5 .6 1.01.5 .5 .4.6 .25 .35 .25 .05 .15
F A-S4UG
G Alloy No. 234 5.06.0 .7 1.01.5 .5 .3.6 .3 .5 .15 .05 .15 Sn .1 Pb .2 *320
ISO ISO Al-Si5Cu1 Mg
32009 USA AA 356.0 (UNS 6.57.5 .6 .25 .35 .2.4 .35 .25 .05 .15
A03560)
32010 USA AA A356.0 6.57.5 .2 .2 .1 .2.4 .1 .2 .05 .15
UK BS LM25 6.57.5 .5 .1 .3 .2.45 .1 .1 .05.2 Sn .05 Pb .1
UK BS 2L99 6.57.5 .2 .1 .1 .2.45 .1 .1 .2 Sn .05 Pb .05
UK BS L173; BS L174 6.57.5 .2 .2 .1 .25.45 .1 .04.25 Be .07 max.
UK AA 356.0 6.57.5 .6 .25 .35 .2.4 .35 .25 .05 .15
F NF A-S7G
G DIN 1725 Al Si 6.57.5 .18 .05 .05 .2.4 .07 .15 .03 .1
7 Mg Wk. 3.2371
ISO ISO AI-Si7 Mg
(continued)
Appendix 6: Metal Alloy Comparison Tables 601

(continued)
Composition (max. unless otherwise stated) Others
Country Designation Si Fe Cu Mn Mg Cr Zn Ti Each Total Al Notes
min.
32011 USA AA 357.0 (UNS 6.57.5 .15 .05 .03 .45.6 .05 .2 .05 .15
A03570)
UK AA 357.0 6.57.5 .15 .05 .03 .45.6 .05 .2 .05 .15
UK BS L169 6.57.5 .2 .1 .1 .5.75 .05 .1 .1.2 .05 .15 Pb .05 Sn .05 Be
.07
F A-S7G 7.0 .3 Nominal comp.
G DIN 1725 Al Si 6.57.5 .18 .05 .05 2.4 .07 .15 .03 .1
7 Mg Wk. 3.2371
32012 USA AA B358.0 (Tens 7.68.6 .3 .2 .2 .4.6 .2 .2 .1.2 .05 .15 Be .1.3
50)
32013 USA AA 359.0 (UNS 8.59.5 .2 .2 .1 .5.7 .1 .2 .05 .15
A03590)
32014 USA AA 380.0 (UNS 7.59.5 2.0 3.04.0 .5 .1 .5 3.0 .5 Sn .35
A03800)
32015 USA AA A380.0 7.59.5 1.3 3.04.0 .5 .1 .5 3.0 .5 Sn .35
UK BS LM24 7.59.5 1.3 3.04.0 .5 .1 .5 3.0 .2 Sn .2 Pb .35
F A-S9U3Y4 8.2 3.5 Nominal comp.
F A-S10U4 9.011.0 1.3 3.04.0 .3 .5 .5 .8 .2 Sn .1 Fe+Zn-
+Mg+Ni+Sn 2.5
G DIN 1725 Al Si 8 Cu 7.59.5 .8 2.03.5 .2.5 .3 .3 1.2 .15 .05 .15 Sn .1
3. Wk. 3.2161
ISO ISO Al-Si8Cu3Fe
32016 USA AA 514.0 (UNS .35 .5 .15 .35 3.54.5 .15 .25 .05 .15
A05140)
UK AA 514.0 .35 .5 . .15 .35 3.54.5 .15 .25 .05 .15
UK BS LM5 .3 .6 .1 .3.7 3.06.0 .1 .1 .2 Sn .05 Pb .05
F NF A-G3T .4 .5 .1 .5 2.53.5 .15 .05 .2 .2 Sn .05 Pb .05
F NF A-G6 .4 4.5 Nominal comp.
G DIN 1725 Al Mg 3 1.3 .6 .3 .6 2.04.0 .3 .2 .05 .15
G DIN 1725 Al Mg 5 1.0 .2 5.0 Nominal comp.
ISO ISO AI-Mg5Si1; ISO
Al-Mg6
32017 USA AA 518.0 (UNS .35 1.8 .25 .35 7.58.5 .15 .15 .25 Sn .15
A05180)
UK AA 518.0 .35 1.8 .25 .35 7.58.5 .15 .15 .25 Sn .15
G DIN 1725 Al Mg 9 .5 .05 .2.5 7.010.0 .1 .15 .05 .15 Nominal comp.
Wk. 3.3292
32018 USA AA 520.0 (UNS .25 .3 .25 .15 9.510.6 .15 .25 .05 .15
A05200)
UK AA 520.0 .25 .3 .25 .15 9.510.6 .15 .25 .05 .15
UK BS LM10; Hid 90 .25 .35 .1 .1 9.511.0 .1 .1 .2 Sn .05 Pb .05
F NF A-G10 .5 1.3 .2 .6 8.511.0 .1 .4 .2 2.0 Sn .1
F NF A-G10Y4 10.2 Nominal comp.
G DIN 1725 Al Mg 10 .3 .3 .05 .3 9.011.0 .1 .15 .05 .15 Sn .05 Pb .05
Wk. 3.3591
32019 USA AA 535.0 (UNS .15 .15 .05 .125 6.27.5 .1 .25 .05 .15 Be .003.007 B
A05350) .002
UK AA 535.0 15 .15 .05 .1.25 6.27.5 .125 .05 .15 Be .003.007 B
.002
UK DTD 5018A .25 .35 .2 1.3 7.47.9 .1 .91.4 .25 Sn .05 Pb .05
G DIN 1725 Al Mg 9 .5 .05 .25 7.010.0 .1 .15 .05 .15 Nominal comp.
Wk. 3.3292
32020 USA AA 707.0 (UNS .2 .8 .2 .4.6 1.82.4 .2.4 4.04.5 .25 .05 .15
A07070)
UK AA 707.0 .2 .8 .2 .4.6 1.82.4 .2.4 4.04.5 .25 .05 .15
(continued)
602 Appendix 6: Metal Alloy Comparison Tables

(continued)
Composition (max. unless otherwise stated) Others
Country Designation Si Fe Cu Mn Mg Cr Zn Ti Each Total Al Notes
min.
32021 USA AAA712.0 (UNS .15 .5 .3565 .05 .6.8 6.07.0 .25 .05 .15
A07120)
32022 USA AAC712.0 .3 .7 .3565 .05 .25.45 6.07.0 .2 .05 .15
1.4
UK AA 712.0 .15 .5 .35.65 .05 .6.8 6.07.0 .25 .05 .15
UK BS LM31; DTD .25 .5 .1 .1 .5.75 .4.6 .1 4.85.7 .05.25 .05 .15 Sn .05 Pb .05
5008B
F A-Z5G .3 .8 .15.35 .4 .4.65 .15.35 .05 4.55.5 .15.25 Sn .05 Pb .05
F NF A-Z5G .6 .5 5.2 Nominal comp.
Notes for wrought and cast aluminium alloys
*31: .0008 max. Be for welding electrodes and ller wire
*32: .03 max. Ga
*33: Cu + Si + Fe + Mn + Zr .2 max
*34: Cu + Si + Fe + Mn + Zr 1.0 max
*35: Cu + Si + Fe + Mn + Zr 0.5 max
*36: .05 B max
*37: .2.6 Bi; .2.6 Pb
*38: .2.7 Pb; .05 max. Sn; .2.7 Bi; .05 max. Sb
*39: .2.6 Pb; .2.6 Bi
*310: .2 max. Zr + Ti
*311: .05.15V; .1.25 Zr
*312: .1 max Ni; .05 max. Sn; .05 max. Pb
*313: .51.3 Ni
*314: .05 max. Sn; .05 max. Pb; .71.3 Ni
*315: 4565 % of Mg
*316: .05 max Sn; .05 max Pb; .05 max Sb
*317: Cr + Mn + Ti + V .03 max
*318: .4.7 Bi; .4.7 Pb
*319: Not closely equivalent, Mg lower than AA 7079
*320: Fe + Mn 1.1 max
*321: .1.25 Zr; .05.15V
*322: .91.2 Ni
*323: .11.17 Zr
*324: Ti + Zr .08.25, Zr .08.2
*325: .81.4 Ni + .25 Zr
Appendix 6: Metal Alloy Comparison Tables 603

European CEN Specifications for
Aluminium Alloys
European EN specications for metal alloys are currently
being generated and adopted. These will progressively
supersede the various national standards for aluminium
alloys, as with other materials. However, it will be some
years before this process is completed and fully imple-
mented. EN designations specic to particular alloys, forms
and conditions have been included in the above alloy tables
(many of these are still at the provisional stage). However,
there are several EN specications which cover the basic
characteristics of aluminium alloys:
Chemical composition specications for wrought alu-
minium alloys are now contained in a single CEN spec-
ication: EN 573Aluminium and aluminium alloys
Chemical composition and form of wrought products.
Temper designations for wrought aluminium alloys are
also contained in a single CEN specication: EN 515
Aluminium and aluminium alloysWrought products
temper designations.
In addition EN 485 now contains conditions for delivery,
properties and tolerances for wrought aluminium alloy
products.
Similar specications will be issued in the future to cover
cast aluminium alloys.

Alloy equivalentsCopper Alloys

Composition
Country Designation Zn Sn Pb Be Fe Al Other
41001 USA CDA 110 (UNS C11000) Cu + Ag > 99.9
UK C101 (CW 003A) <.005 Cu + Ag > 99.9
F NF Cu a1 A53-100 Cu 99.9 oxygen free. Nominal comp.
G DIN 1708 E-Cu57 Wk. Cu 99.95, Ag 0.03, O .005.040 nominal comp.
2.0060
41002 USA CDA 170 (UNS C17000) 1.61.79 Cu > 99.5 Co + Ni > .2 Co + Fe + Ni < .6
UK CB101 (CW101C) 1.71 9 Ni + Co .05.4 Other < .5
G CuBe1.7Wk. 2.1245
41003 USA CDA 172 (UNS C17200) 1.82.0 Cu > 99.5 Co + Ni > .2 Co + Fe + Ni < .6
G Cu Co Be Wk. 2.1285 Co + Fe + Ni < .6
41004 USA CDA194 (UNS C19400) .05.2 <.03 2.12.6 P .015.15 Cu 97.097.8 others < .15
ASTM B465
41005 USA CDA 195 (UNS C19500) <2 .11.0 <.02 1.02.0 <.02 P .01.35 Co .31.3
41006 USA CDA 230 (UNS C23000) Balance <.05 <.03 Cu 84.086.0
ASTM B36: B43: B111:
B134: B135:
B359: B395: B543: B587
UK CZ102 (CW502L) Balance <.l <.1 Cu 84.086.0 other < .4
G DIN 17660 DIN 17670 Balance <.05 <.05 <.05 <.02 Cu 84.086.0 Mn < .05 Ni < .2 Sb < .01 other .05
Wk. 2.0240 total other (except Ni) .3
F NFU-Z15 15 Nominal comp.
41007 USA CDA 260 (UNS C26000) Balance <.07 <.05 Cu 68.571.5
UK CZ106 (CW505L) Balance <.05 <.05 Cu 68.571.5 other < .3
F UZ-30
G DIN 17660 Cu Zu 30 Wk. Balance <.05 <.05 <.05 <.02 Cu 69.071.0 Ni < .2 Sb < .01 other .05 total other
2.0265 (except Ni) .3
41008 USA CDA 353 (UNS C35300) Balance 1.32.3 <1 Cu 59.064.5
UK CZ119 (CW601 N) Balance 1.02.5 Cu 61.064.0
G DIN 17660 CuZu 36 Pb Balance <.1 .72.5 <.2 <.05 Cu 62.0 64.0 Mn < .1 Ni < .3 5b < .01 others total .1.
1.5 Wk. 2.0331 Any other except Ni < .5
41009 USA CDA 422 (UNS C42200) Balance .81.4 <.05 <.05 Cu 86.089.0 P < .35
ASTM B591
41010 USA CDA 443 (UNS C44300) Balance .91.2 <.07 <.06 Cu 70.073.0 As .01.1
ASTM B111: B171:
B359: B395:
B432: B543
(continued)
604 Appendix 6: Metal Alloy Comparison Tables

(continued)
Composition
Country Designation Zn Sn Pb Be Fe Al Other
G DIN 17660 CuZn 28 Sn .91.3 <.07 <.07 Cu 70.072.5 As .02.035 Mn < .1 Ni < .1 P < .01
Wk. 2.0470 As + P < .035 other total < .1
41011 USA CDA 510 (UNS C51000) <.3 4.25.8 <.05 <1 Cu + Sn + P > 99.5 P 03.35
ASTM B100: B103:
BI39: BI59
UK PB102 (CW451 K) 4.56.0 <.02 P .02.4 others < .2
F NF U-E5P
G DIN 17662 CuSn 6 Wk. <.3 5.57.5 <.05 <.1 P .01.4 Ni < .3 other < .2
2.1020

Alloy equivalentsCopper Alloys

Composition
Country Designation Zn Sn Pb Be Fe Al Other
41012 USA CDA 521 (UNS C52100) ASTM <.2 7.09.0 <.05 <1 Cu + Sn + P > 99.5 P 03.35
139:159
UK PB103 (CW452 K) 6.07.5 <.02 P .02.4 others < .2
UK PB104 (CW459 K) 7.59.0 <.02 P .02.4 others < .2
F NF U-E7P
G DIN 17662 Cu Sn 8 Wk. 2.1030 <.3 7.59.0 <.05 <.1 P .01.4 Ni < .3 other together < .2
41013 USA CDA 619 (UNS C61900) <.8 <.6 <.02 3.04.5 8.510.0 Cu + Ag 83.688.5
Cu + Fe + AI > 99.5
ASTM B129: B150: B283
UK CA103 <.4 <1 <.05 Fe + Ni < 4.0 8.810.0 Mn < .5 Mg < .05 total impurities
not Mn < .5
F NF U-A8
G DIN 17665 Cu Al 8 Wk. 2.0920 <.5 <.02 <.5 7.09.0 Mn < .8 Ni < .8 Si < .2 other
total < .3
41014 USA CDA 687 (UNS C68700) Balance <.07 <.06 1.82.5 Cu + Ag 76.079.0 As .02.06
ASTM B111: SB359: SB395:
SB543
UK CZ110 (CW702R) Balance <.07 <.06 1.82.3 Cu 76.078.0 As .02.06
G DIN 17660 Cu Zn 20 Al Wk. 2.0460 Balance <.07 <.07 1.82.3 Cu 76.079.0 As .02.035 Mn < .1
Ni < .1 P < .01 As + P < .035 other
together < .1
41015 USA CDA688 (UNS C68800) ASTM 21.324.1 <.05 <.05 3.03.8 Co .25.55
B592
41016 USA CDA 706 (UNS C70600) <1.0 <.05 1.01.8 Ni 9.011.0 Mn < 1.0
Cu + Ag > 86.5
Cu + Fe + Ni > 99.5
ASTM B111: B122: B151: B171:
B359: B395: B402:B432: B466:
B467: B543: B552
UK CN102 (CW352H) <.01 1.02.0 Ni 10.011.0 Mn .5 1.0 S < .05
total impurities < .3 C < .05
F NF Cu Ni 10 Fe Mn * Ni 10.0 + Fe + Mn, nominal comp.
G DIN 17664 Cu Ni 10 Fe Wk. 2.0872 <.5 <.03 1.01.8 Ni 9.011.0 Mn 0.51.0 S < .05
total other < .1
41017 USA CDA 725 (UNS C72500) <.5 1.82.8 <.05 <.6 Ni 8.510.5 Mn < .2
Cu + Ni + Sn + Co > 99.8
41018 USA CDA 762 (UNS C76200) ASTM Balance <.1 <.25 Ni 11.013.5 Mn < .5 Cu 57.061.0
B122
UK NS104 (CW403 J) Balance <.04 <.25 Ni 11.013.0 Mn .05.3 Cu 60.0-
65.0 total Impurities < .5
G DIN 17663 Cn Ni 12Zn 24 Balance <.2 <.05 <.3 Ni 11.0 13.0 Mn < .5 Cu 63.066.0
other together < .1
(continued)
Appendix 6: Metal Alloy Comparison Tables 605

(continued)
Composition
Country Designation Zn Sn Pb Be Fe Al Other
41019 USA CDA 766 (UNS C76600) FEDQQ- Balance <.1 <.25 Ni 11.013.5 Mn < .5 Cu 55.058.0
C-585
41020 USA CDA 770 (UNS C77000) Balance <.l <.25 Ni 16.519.5 Mn < .5 Cn 53.556.5
ASTM B122: B151: B206
UK NS107 (CW410 J) Balance <.03 <.3 Ni 17.019.0 Mn .05.35 Cu
54.056.0
G DIN 17663 Cu Ni 18 Zn 27 Wk. Balance <.2 <.03 <.3 Ni 17.019.0 Mn < .7 Cu 60.063.0
2.0740
41021 USA CDA 782 (UNS C78200) Balance 1.52.5 <.35 Ni 7.09.0 Mn < .5 Cu 63.067.0
UK NS101 (CW402 J) Balance 1.02.5 <.4 Ni 9.011.0 Mn .2.5 Cu 44.047.0
G DIN 17663 Cu Ni 10Zn 42 Pb Balance <.3 .52.0 <.5 Ni 9.011.0 Mn < .5 Cu 45.048.0
others total < .1

Alloy equivalentsMagnesium Alloys

Composition
Country Designation Al Zn Mn Other
61001 USA M1A ASTM B107, B275; SAE 51, 522, 533 >1.2 Si < .1 Cu < .05 Ni < .01 Ca < .3
(UNS M15100) other < .3
UK 1428:7378
G W3501 DIN 1729 Wk. 3.5200 <.05 <.03 1.22.0 Si < 1 Cu < 05 Fe < .005 others < .1
61002 USA LA141 ASTM B270; MIL SPEC M-46130 <.05 1.5 Cu < .05 Fe < .005 Li 12.015.0
(UNS M14142) Ni < .005 Si .5.6 Na < .005
61003 USA AZ31B ASTM B107, B273; FED QQ-M-31, 2.33.5 .61.4 >.2 Si < .1 Cu < .03 Ni < .005 Fe < .003
M-40, M44, WW-T-825 (UNS M11311) Ca < .04 other < .3
UK BS 3370-MAG-S-111; BS 3373-MAG-E-111; 2.53.5 .61.4 .15.7 Ca < .3 Si < .3 Cu .05 Ni < .005
DTD 742 Fe < .005
G DIN 1729 Mg Al 3 Zn Wk. 3.5312 2.53.5 .51.5 .15.4 Si < .1 Cu < .1 Fe < .003 Ni < .005
Ca < .04 other < .1
61004 USA ZK60A ASTM B91, B107, B275; FED QQ- 4.86.2 Zr > .45 others < .3
M-31, M-40, WW-T-825 (UNS M16600)
USA ZK61A ASTM B403 6.0 Zr .8 nominal comp.
UK BS 3373-MAG-E-161 5.0 Zr .6 nominal comp.
UK DTD 5041A 5.5 Zr .7 nominal comp.
F G-Z 5 Zr <.02 3.55.5 <.15 Zr .41.0 Cu < .03 Si < .01
Fe < .01 N < 001
G Dl N 1729 Mg Zn 6Zr Wk. 3.5161 4.86.2 Zr .45.8 others < .3
61005 USA AZ61A ASTM B91, B107, B275; FED QQ- 5.87.2 .41.5 .15 Si < .1 Cu < .05 Ni < .005 Fe < .005
M-31, M40, WW-T-825 (UNS M11610) other < .3
UK DTD 259A 5.58.5 <1.5 .2.4 Si < .1 Cu < .1 Ni < .005 Fe < .03
F G-A 7 Z1 6.58.5 .51.5 >.12 Si < .3 Cu < .05 Ni < .005 Fe < .007
others < .3
G DIN 1729 Mg AI6Zn Wk. 3.5612 5.57.0 .51.5 .15.4 Si < .1 Cu < .1 Ni < .005 Fe < .03
others < .1
61006 USA AZ80A ASTM B91, B107, B275; FED.QQ- 7.89.2 .2.8 >.12 Si < .1 Cu < .05 Ni < .005 Fe < .005
M-31, M-40 (UNS M11800) other < .3
UK BS 2L121; BS 2L122 7.59.0 .31.0 .15.4 Si < .3 Cu < .15 Ni < .01 Fe < .05
Sn < .1 Cu + Si + Fe + Ni < .4
F G-A 7 Z1 See
AZ61A
G DIN 1729 Mg Al 8 Zn Wk. 3.5812 7.89.2 .2.8 .12.3 Si < .1 Cu < .05 Ni < .003 other < .3
61007 USA LAZ933Ballette Mem. Institute 3.0 3.0 Li 9.0, nominal comp.
61008 USA ZK21A 2.3 Zr .45 nominal comp.
UK BS 3373-MAG-E-151, BS 3374-MAG-P-151, 3.0 Zr .6 nominal comp.
ZW3
61009 UK RZ5Magnesium Elektron 4.0 Zr .7 rare earth 1.2 nominal comp.
606 Appendix 6: Metal Alloy Comparison Tables

Alloy equivalentsMiscellaneous Alloys

Country Designation Composition
71001 USA Beryllium S-100C QMV Grade. Brush BeO 1.2, Be 98.5
Beryllium Corp.
71002 USA Beryllium S-200C QMV Grade. Brush BeO 2.0, Be 98.0
Beryllium Corp.
Beryllium SR-200 QMV Grade. Brush BeO 2.0, Be 98.0 Hot rolled sheet
Beryllium Corp.
71003 USA Beryllium S-300C QMV Grade. Brush BeO 3.0, Be 97.4
Beryllium Corp.
71004 USA MP35N Multiphase (UNS R30035) C < .025, Cr 1921.0, Mo 910.5, Ni 3337, Fe < 1.0, Mn < .15, P < .015, S < .01, Si < .15,
Ti < 1.0, Co balance
71005 USA HS25; L-605; Haynes 25 C .05.15, Cr 19.021.0, Ni 9.011.0, W 14.016.0, Fe < 3.0, Co balance
UK BS 3531/1-4 C .1, Cr 20, Ni 10, W 15, Co balance. Nominal comp.
G Co Cr20 W 15 Ni Wk. 2.4967 C .05.13, Cr 19.021.0, Ni 9.011.0, W 14.016.0, Fe < 3.0, Si < 1.0, Mn 1.02.0, P < .045,
S < .03, Co balance
71006 USA HS188Haynes Alloy (UNS R30188) C .05.15, Cr 2024, Ni 2024, W 1316, Fe < 3.0, Mn < 1.25, La .03.15, Si .2.5, Co
balance.

References for Alloy Equivalents ICS: 49Aronautique et Espace

Work Programme 1995

BSI Catalogue 1995/96. CENEuropean Committee for Standardization, Brussels.
British Standards Institution ISBN: 92-9097-432-X
Iron and Steel Specications, 8th Edition. MIL-HDBK-5 J: Metallic Materials and Elements for
British Iron and Steel Producers Association, December Aerospace Vehicle Structures, Vols. 1 and 2.
1994. Department of Defense, USA. 2003
Unied Numbering System, 4th Edition. Metallic Materials Specication Handbook, 4th Edition
SAE/ASTM 1986 Robert B. Ross
Chapman & Hall, 1992, ISBN: 0-412-36940-0
Stahlschssel, 1977.
Properties of Aluminium and its Alloys Smithells Metals Reference Book, 8th Edition.
Aluminium Federation, 2002 2002, Butterworth-Heinemann
Buyers Guide to progress on European Standards ECSS Q-ST-70-71: Data for the Selection of Space
Aluminium Federation, UK. February 1996 Materials.
AFNOR Catalogue, 1996
ICS: 77Mtallurgie
Appendix 6: Metal Alloy Comparison Tables 607

Alloy equivalentsTitanium Alloys

Composition
Country Designation Al V Cr Other
51001 USA CP Ti (UNS R52250) ASTM B265 C 0.1: H 0.01: Fe 0.2, Commercial
purity Titanium
UK BS TA1: IMI 115: Ti 115: H 0.012: Fe 0.2, Commercial purity
Titanium
F AFNOR T35
G Wk. 3.7024
51002 USA 3AI-2.5V (UNS R56320) 2.53.5 2.03.0 C < .05: H < .013: Fe < .25: N < .02
AMS 4943.4944 ASTM B337 O < .12 Ti Balance
51003 USA 6AI-4V (UNS R56401) 5.56.75 3.54.5 C < .1: H < .015: Fe < .4: N < .05
O < .2 Ti Balance
AMS 4906: 4911: 4934, 4935:
4954:4965:4967
ASTM B265, B348: B367: B381:
AWS A5-16
MIL. SPEC. F83142: T9046: T9047:
T81556: T81915
UK BS TA56: 2TA10: 2TA13: 2TA28: 5.56.75 3.54.5 H < .025: Fe < .3: O + N < .25 Ti
Balance
IMI318: Ti 318A
UK DTD5163: 5173: 5303: 5313: 5323 6.1 4.0 H < .012 nominal comp.
F TA6V 6.0 4.0 Nominal comp.
G Ti Al 6V 4 Wk. 3.7164 6.0 4.0 Nominal comp.
AECMA TI-P63
CEN EN 2517PR (TI-P63 alloy: annealedsheet, strip and plate, a 100 mmprovisional spec.)
EN 2530PR (TI-P63 alloy: annealed900 MPa RM 1160 MPabars, d 100 mmprovisional spec.)
EN 2531 PR (TI-P63 alloy: annealed900 MPa RM 1160 MPaforgings, d 100 mmprovisional spec.)
EN 3310PR (TI-P63 alloy: not heat treatedreference heat treatmentannealedgrade 2 forging stocks, d 360 mm
provisional spec.)
EN 3311 PR (TI-P63 alloy: annealed900 MPa RM 1160 MPabar for machining, d 150 mmprovisional
spec.)
EN 3312PR (TI-P63 alloy: annealed900 MPa RM 1160 MPaforgings, d 150 mmprovisional spec.)
EN 3313PR (TI-P63 alloy: not heat treated- reference heat treatmentsolution treated and agedgrade 2 forging
stocks, d 360 mmprovisional spec.)
EN 3314PR (TI-P63 alloy: solution treated and agedRM > = 1070 MPabar for machining, d 50 mmprovisional
spec.)
EN 3315PR (TI-P63 alloy: solution treated and agedRM > = 1070 MPaforgings, d < = 50 mmprovisional spec.)
EN 3456PR (TI-P63 alloy: annealed920 MPa RM 1180 MPasheet and strip, a 6 mmprovisional spec.)
EN 3457PR (TI-P63 alloy: not heat treatedreference heat treatmentsolution treated and agedgrade 2 forging
stock for fasteners, d 25 mmprovisional spec.)
EN 3458PR (TI-P63 alloy: annealed900 MPa RM 1160 MPabar and wire for machined fasteners, d 25 mm
provisional spec.)
EN 3464PR (TI-P63 alloy: annealed900 MPa RM 1160 MPaplate, 6 mm a 100 mmprovisional spec.)
51004 USA 13V-11Cr-3AI (UNS R58010) 2.54.0 12.514.5 1012 Fe < 0.35: C 0.050.1
AMS4917:4959
AWS A5 MIL SPEC. F-83142: T-
9046: T-9047: T-81588
51005 UK IMI 685 6.0 Zr 5.0: Mo 0.5: Si 0.5, nominal comp.
51006 UK IMI 829 5.5 Sn 3.5: Zr 3.0: Nb 1.0: Mo 0.3: Si
0.3, nominal comp.
608 Appendix 6: Metal Alloy Comparison Tables

UNS No. Table No. (continued)

Carbon steels and alloy steels UNS No. Table No.
G10050 11001 G11170 13005
G10060 11002 G11180 13006
G10080 11003 G11190 13007
G10100 11004 G11320 13008
G10110 11005 G11370 13009
G10120 11006 G11390 13010
G10150 11007 G11400 13011
G10160 11008 G11410 13012
G10170 11009 G11440 13013
G10180 11010 G11450 13014
G10190 11011 G11460 13015
G10200 11012 G12134 13016
G10210 11013 G12144 13017
G10220 11014 G15130 12001
G10230 11015 G15180 12002
G10250 11016 G15220 12003
G10260 11017 G15240 12004
G10290 11018 G15250 12005
G10300 11019 G15260 12006
G10350 11020 G15270 12007
G10370 11021 G15360 12008
G10380 11022 G15410 12009
G10390 11023 G15470 12010
G10400 11024 G15480 12011
G10420 11025 G15510 12012
G10430 11026 G15520 12013
G10450 11027 G15610 12014
G10460 11028 G15660 12015
G10490 11029 G15720 12016
G10500 11030 Low alloy steels
G10530 11031 G40120 14001
G10550 11032 G40230 14002
G10600 11033 G40240 14003
G10640 11034 G40270 14004
G10650 11035 G40280 14005
G10690 11036 G40320 14006
G10700 11037 G40370 14007
G10740 11038 G40420 14008
G10750 11039 G40470 14009
G10780 11040 G41180 14010
G10800 11041 G41300 14011
G10840 11042 G41350 14012
G10850 11043 G41370 14013
G10860 11044 G41400 14014
G10900 11045 G41420 14015
G10950 11046 G41450 14016
G11080 13001 G41470 14017
G11090 13002 G41500 14018
G11100 13003 G41610 14019
G11160 13004 G43200 14020
(continued) (continued)
Appendix 6: Metal Alloy Comparison Tables 609

(continued) (continued)
UNS No. Table No. UNS No. Table No.
G43400 14021 S30900 15012
G46150 14039 S31000 15013
G46170 14040 S31008 15014
G46200 14041 S31400 15015
G46210 14042 S31600 15016
G46260 14043 S31603 15017
G47180 14022 S31700 15018
G47200 14023 S32100 15019
G51150 14044 S34700 15020
G51200 14045 S34800 15021
G51300 14046 S38400 15022
G51320 14047 S40300 15023
G51350 14048 S40500 15024
G51400 14049 S41000 15025
G51450 14050 S41400 15026
G51470 14051 S41600 15027
G51500 14052 S41623 15028
G51550 14053 S42000 15029
G51600 14054 S42020 15030
G51986 14055 S42900 15031
G52986 14056 S43000 15032
G61180 14057 S43020 15033
G61500 14058 S43023 15034
G81150 14024 S43100 15035
G86150 14025 S43400 15036
G86170 14026 S43600 15037
G86200 14027 S44002 15038
G86220 14028 S44003 15039
G86250 14029 S44004 15040
G86270 14030 S44200 15041
G86300 14031 S44600 15042
G86370 14032 S50100 15043
G86400 14033 S50200 15044
G86420 14034 Named steels
G86450 14035 J42015 16002
G86500 14036 J92200 16013
G86550 14037 K08500 16001
G86600 14038 K66286 16004
Stainless steels K92820 16015
S20100 15001 K92940 16016
S20200 15002 K93160 16017
S30100 15003 N08020 16003
S30200 15004 S13800 16014
S30215 15005 S14800 16009
S30300 15006 S15500 16008
S30323 15007 S15700 16010
S30400 15008 S17700 16011
S30403 15009 S21900 16018
S30500 15010 S35000 16005
S30800 15011 S35500 16006
(continued) (continued)
610 Appendix 6: Metal Alloy Comparison Tables

(continued) (continued)
UNS No. Table No. UNS No. Table No.
S36200 16019 A95154 31041
S45500 16007 A95252 31043
T20811 16012 A95254 31044
Nickel alloys A95356 31045
N05500 21012 A95454 31046
N06002 21002 A95456 31047
N06600 21008 A95457 31048
N06625 21009 A95652 31049
N07001 21016 A95657 31050
N07041 21014 A96003 31051
N07718 21010 A96005 31052
N07750 21011 A96053 31053
N08800 21003 A96061 31054
N09901 21004 A96063 31055
N09902 21013 A96066 31056
N10002 21001 A96070 31057
N19903 21005 A96082 31059
N19907 21006 A96101 31060
N19909 21007 A96151 31061
Aluminium alloys (wrought) A96162 31062
A91050 31001 A96201 31063
A91060 31002 A96253 31064
A91100 31003 A96262 31065
A91145 31004 A96351 31066
A91175 31005 A96463 31067
A91200 31006 A96951 31068
A91230 31007 A97001 31069
A91235 31008 A97009 31070
A91345 31009 A97039 31073
A91350 31010 A97049 31074
A92011 31015 A97050 31075
A92014 31016 A97075 31076
A92017 31017 A97079 31077
A92024 31018 A97175 31078
A92048 31019 A97178 31079
A92124 31023 Aluminium alloys (cast)
A92219 31026 A02950 32001
A93003 31029 A03190 32003
A93004 31030 A03330 32005
A93005 31031 A03550 32007
A93103 31032 A03560 32009
A93105 31033 A03570 32011
A94032 31034 A03590 32013
A95005 31035 A03800 32014
A95050 31036 A05140 32016
A95050 31042 A05180 32017
A95052 31037 A05200 32018
A95056 31038 A05350 32019
A95083 31039 A07070 32020
A95086 31040 A07120 32021
(continued) (continued)
Appendix 6: Metal Alloy Comparison Tables 611

(continued) (continued)
UNS No. Table No. UNS No. Table No.
A33550 32008 C76200 41018
Copper alloys C76600 41019
C11000 41001 C77000 41020
C17000 41002 C78200 41021
C17200 41003 Titanium alloys
C19400 41004 R52250 51001
C19500 41005 R56320 51002
C23000 41006 R56401 51003
C26000 41007 R58010 51004
C35300 41008 Magnesium alloys
C42200 41009 M11311 61003
C44300 41010 M11610 61005
C51000 41011 M11800 61006
C52100 41012 M14142 61002
C61900 41013 M15100 61001
C68700 41014 M16600 61004
C68800 41015 Miscellaneous alloys
C70600 41016 R30035 71004
C72500 41017 R30188 71006
(continued)
 Appendix 7: Variation of Standard Free Energy
of Formation of Oxides with Temperature
These are graphs showing how free energy values vary with
temperature. The network is based on work by Ellingham
and Richardson in the 1940s and 1950s and is still very
useful today (see for instance Ellingham, H.J.T. (1944)
J. Soc. Chem. Ind., 63, 125133). The diagram indicates at
any temperature what is thermodynamically possible for
metallurgical reactions that are based on either thermal
reduction processes or thermal oxidation processes.
At the point of intersection of any two curves the standard
free energy for the chemical reaction is zero and this is the
point of equilibrium. The diagram can be used to show
approximately whether reduction or oxidation will take
place.
As an example, below a temperature of about 1600 C,
pure magnesium is expected to reduce aluminium oxide. At
temperatures higher than 1600 C, when the curve for
magnesium oxide is above that of aluminium oxide, one
would expect that pure aluminium would reduce magnesium
oxide to metallic magnesium. The approximate temperatures
when gases such as carbon monoxide might reduce metal
oxides can, in a similar way, be deduced from the various
points of graph intersections.
It is possible to determine the partial pressure of oxygen
in equilibrium at a given temperature with the couples Fe/
FeO, Fe/Fe3O4, Ni/NiO, and the like. The diagrams shows
that up to very high temperatures (>1500 C) the oxides FeO
Springer International Publishing Switzerland 2016
B.D. Dunn, Materials and Processes, Springer Praxis Books,
DOI 10.1007/978-3-319-23362-8
and Fe3O4 are more stable than NiO, and that the oxide
Fe2O4 are more stable than NiO, and that the oxide Fe2O3 is
the least stable of all.
Reduced pressures will enhance most reducing reactions
as will increasing the ratio of hydrogen or carbon monoxide
present in a gaseous mixturethese effects can also be
estimated from intersection points. The techniques are
described in books such as Metallurgical Thermochemistry,
6th Edition, (1993) by O. Kubaschewski, C. Alcock and P.
J. Spencer, (Pergamon, Oxford) and Free Energy of For-
mation of Binary Compounds (1971), by T.B. Reed (MIT
Press, London).
It is important to note that the Ellingham diagrams relate
to the equilibrium conditionsthey take no account of the
kinetics of any oxidation or reduction reaction.
For alloys, it is worth noting that the most easily oxidized
constituent should be considered when consulting the dia-
gram. It is usual to discount elements that have a concen-
tration less than about 1 % as they are not able to form
continuous surface lms.
A recent, short but very useful articleEllingham dia-
grams, their Use and Misusetakes account of several case
examples where diagrams can be used for trouble shooting
during processes such as heat treatment, brazing and bright
annealing (Stratton 2013).
613
614 Appendix 7: Variation of Standard Free Energy of Formation of Oxides with Temperature

Based on a diagram supplied by the British Iron and Steel Research Association
 Appendix 8: Simplied Procedure
for the Management of Materials, Processes
and Mechanical PartsPossible Guidelines
for a Cubesat or Small University Spacecraft

Flow Chart for M,P and MP Lists

Company Structure Exists

Responsible for company Product Assurance:

Responsible for Materials and Process, Plus Mechanical Parts (eg Materials Manager):

Responsible for Design and Verification:

Selection

Evaluation:
Critical
Verification of Processes
Yes Reject
Criticality Evaluation Validation of Materials
Analysis Necessary Qualification of Mechanical Parts New Selection

No
No
Non - Critical
Reject New Test or
Evaluation:
Request for
Verification by Test Results
Accept Deviation
Similarity
Yes
Accept

Accept

Add Material, Process or Mechanical Part to:

DML, DPL, DMPL

Declared Material List Declared Process List Declared Mech. Parts List

Springer International Publishing Switzerland 2016 615

B.D. Dunn, Materials and Processes, Springer Praxis Books,
DOI 10.1007/978-3-319-23362-8
616 Appendix 8: Simplified Procedure for the Management of Materials, Processes and Mechanical Parts

Processing parameter (e.g. heat-treatment for metals

Management of Lists (see Flow Chart)
and mix ratio/curing time/temp for organic)
Use and end location
1. List should be initiated and maintained during the life of
Any test date (reference to reports, literature, etc.)
the project by the Material and Processes (M&P)
Last column is for Approval by M&P specialist and
Manager.
projeet manager who gives nal approval.
2. Inputs mainly from Design Manager (e.g. engineering
(d) All acronyms need to be dened
drawings).
(e) Items with a limited life such as two part adhesives,
3. Lists need to be reviewed at PDR (they will be incom-
paints, etc. must be identied under processing
plete but may identity critical items) and again at CDR.
parameter.
4. SuitabiIity of the items on the 3 Iists should be assessed
by an independent material and process specialist.
5. Technical criteria for the selection of material will
include:
Evaluation, Procurement, Inspection,
(a) Effect of temperature and thermal cycling
Traceability, and Storage
(b) Effect of vacuum (outgassing of organic materials and
sublimation of metaIs such as cadmium (forbidden))
Evaluation
(c) Effect of radiation (not generally an issue except for
Some materials may need to be evaluated for outgassing.
solar cell cover glasses, white paints)
(d) Corrosion and galvanic compatibility (i.e. high Procurement
strength aIuminium alloys need a chemical conver- Some materials have long lead times (versus the project
sion coating that enabIes resistace to surface corro- schedule); these need to be identied before PDR.
sion prior to launch and electrical grounding between Inspection
electronic boxes and structure). Inspection is needed at in-coming to ensure suitability,
check materials are in-life, damage to surface, castings
Note:PDRpreliminary design review (usually an in- may need to be x-rayed etc.
depth assessment, by an interdependent team of discipline
experts and managers, that the design and materials/pro- Traceability
cesses are realistic); CDRcritical design review (to ensure When possible each material and batch should have a
the spacecraft's designed hardware is ready for launch). unique reference number.
Storage
Good controls needed regarding: humidity, cleanliness,
Lay-Out of Lists (see also Appendices 10 refrigeration for certain items, health and safety (regarding
and 11 for examples) toxic materials), ammable materials.

(a) These can be formatted prior to PDR and Materials are

grouped according to Table 7.A.1
(b) All lists need issue status for conguration control
(c) Columns will provideUnique Item number
Material designation (e.g. commercial or recognisable
identication)
International code, e.g. AISI, AA, CDA, etc. (for
instance as listed in Appendix 6)
Manufacturers name
Surface nish
Mechanical Parts and Process Controls
The DPL and DMPL will need to be designed in a similar
manner to the DML.
Exact requirements for DPL and DMPL should be agreed
with the M and P specialist.
The DMPL needs to list all mechanismseach should be
assessed by the M and P specialist (e.g. that lubricating
Appendix 8: Simplified Procedure for the Management of Materials, Processes and Mechanical Parts 617

Table 7.A.1 Material group numbers

Group Description ProcessesControls and Possible
number Assessment for Quality
1 Aluminium and aluminium alloys
2 Copper and copper alloys It would be benecial to review all spacecraft systems and
3 Nickel and nickel alloys sub-systems with a mind to check if suitable controls are
4 Titanium and titanium alloys being made, and whether some minimal verication of
5 Steels unusual processes should be Veried:
6 Stainless steels
7 Filler metals: welding, brazing soldering Structure, mechanical assembly or welding process,
8 Miscellaneous metallic materials simple corrosion protection
9 Optical materials Black boxes, manufacture, painting, grounding
10 Adhesives, coatings, vamishes Harness, crimping to ECSS standard?, wire type (silver
11 Adhesives tapes or tin-plated?)
12 Paints an inks General welding (are materials suitable), bonding,
13 Lubricants painting
14 Potting compounds, sealants, foams Pcb assembly methods, repair, controls (ECSS standard?)
15 Reinforced plastics (including PCBs) Make an attempt to list processes and critical ones may
16 Rubbers and elastomers need laboratory testing on in-line samples.
17 Thermoplastics [e.g non-adhesive tapes and folls
(MLI)]
18 Thermoset plastics (including PCBs)
19 Materials aspects of wires and cables
Need for Access to Space Materials
20 Miscellaneous non-metallic materials, e.g ceramics
Laboratory to Assess Quality
From ECSS Q-ST-70B
and Suitability for Use in Vacuum

greases and oils to not outgas and contaminate optical
systems, that rotating parts in vacuum do not cold-weld
etc.).
The DPL should also be reviewed and possibly com-
piled by the M and P specialist to ascertain if each
process can be considered non- critical or critical
based on previous usage, reliability, inspect ability,
possibility to re-work in case of human or material
errors (Table 7.A.1).
Microsectioning of welded and crimped joints*
Tensile testing of welds and crimps*
NDT occasionally for welds*
Outgassing tests to ECSS standard (can be done at ded-
icated European labs), needed for a few organic materials
(glues, paints, etc.) in vicinity of optical systems (line of
sight to consider condensation). Some laboratories
offering testing for outgassing are mentioned on page 40.
*Notecould use local university or industry
 Appendix 9: Materials and Processes Standards
Related to Space (Released by ECSS, JAXA and
NASA) as of 2015
ECSS-Q-ST-10-04C Critical-item control
ECSS-Q-ST-10-09C Nonconformance control system
ECSS-Q-ST-20C Rev.1 Quality assurance
ECSS-Q-ST-20- Quality and safety assurance for space test
centres
ECSS-Q-ST-20-08C Storage, handling and transportation
of spacecraft hardware
ECSS-Q-ST-20-10C Off-the-shelf items utilization in space
systems
ECSS-Q-ST-30 C Dependability
ECSS-Q-ST-30-02C Failure modes, effects (and criticality)
analysis (FMEA/FMECA)
ECSS-Q-ST-30-09C Availability analysis
ECSS-Q-ST-30-11C Rev.1 DeratingEEE components
ECSS-Q-ST-40C Safety
ECSS-Q-ST-40-02C Hazard analysis
ECSS-Q-ST-40-12C Fault tree analysisAdoption notice
ECSS/IEC 61025
ECSS-Q-ST-60 C Rev.2 Electrical, electronic and elec-
tromechanical (EEE) components
ECSS-Q-ST-60-02C ASIC and FPGA development
ECSS-Q-ST-60-05C Rev.1 Generic procurement require-
ments for hybrids
ECSS-Q-ST-60-12C Design, selection, procurement and
use of die form monolithic microwave integrated circuits
(MMICs)
ECSS-Q-ST-60-13C Commercial electrical, electronic and
electromechanical (EEE) components
ECSS-Q-ST-60-14C Re-ling procedureEEE
components
ECSS-Q-ST-60-15C Radiation hardness assuranceEEE
components
ECSS-Q-ST-70C Rev.1 Materials, mechanical parts and
processes
ECSS-Q-ST-70-01C Cleanliness and contamination control
Springer International Publishing Switzerland 2016
B.D. Dunn, Materials and Processes, Springer Praxis Books,
DOI 10.1007/978-3-319-23362-8
ECSS-Q-ST-70-02C Thermal vacuum outgassing test for
the screening of space materials
ECSS-Q-ST-70-03C Black-anodizing of metals with inor-
ganic dyes
ECSS-Q-ST-70-04C Thermal testing for the evaluation of
space materials, processes, mechanical parts and assemblies
ECSS-Q-ST-70-05C Detection of organic contamination
surfaces by infrared spectroscopy
ECSS-Q-ST-70-06C Particle and UV radiation testing for
space materials
ECSS-Q-ST-70-07C Verication and approval of automatic
machine wave soldering
ECSS-Q-ST-70-08C Manual soldering of high-reliability
electrical connections
ECSS-Q-ST-70-09C Measurements of thermo-optical
properties of thermal control materials
ECSS-Q-ST-70-10C Qualication of printed circuit boards
ECSS-Q-ST-70-11C Procurement of printed circuit boards
ECSS-Q-ST-70-12C Design rules for printed circuit boards
ECSS-Q-ST-70-13C Rev.1 Measurements of the peel and
pull-off strength of coatings and nishes using pressure-
sensitive tapes
ECSS-Q-ST-70-18C Preparation, assembly and mounting
of RF coaxial cables
ECSS-Q-ST-70-20C Determination of the susceptibility of
silver-plated copper wire and cable to "red-plague"
corrosion
ECSS-Q-ST-70-21C Flammability testing for the screening
of space materials
ECSS-Q-ST-70-22C Control of limited shelf-life materials
ECSS-Q-ST-70-26C Crimping of high-reliability electrical
connections
ECSS-Q-ST-70-28C Repair and modication of printed
circuit board assemblies for space use
ECSS-Q-ST-70-29C Determination of offgassing products
from materials and assembled articles to be used in a
manned space vehicle crew compartment
619
620 Appendix 9: Materials and Processes Standards Related to Space (Released by ECSS, JAXA and NASA) as of 2015
ECSS-Q-ST-70-30C Wire wrapping of high-reliability
electrical connections
ECSS-Q-ST-70-31C Application of paints and coatings on
space hardware
ECSS-Q-ST-70-36C Material selection for controlling
stress-corrosion cracking
ECSS-Q-ST-70-37C Determination of the susceptibility of
metals to stress-corrosion cracking
ECSS-Q-ST-70-38C High-reliability soldering for surface-
mount and mixed technology
ECSS-Q-ST-70-45C Mechanical testing of metallic
materials
ECSS-Q-ST-70-46C Rev.1 Requirements for manufacturing
and procurement of threaded fasteners
ECSS-Q-ST-70-50C Particles contamination monitoring for
spacecraft systems and cleanrooms
ECSS-Q-ST-70-53C Materials and hardware compatibility
tests for sterilization processes
ECSS-Q-ST-70-55C Microbial examination of ight hard-
ware and cleanrooms
ECSS-Q-ST-70-56C Vapour phase bioburden reduction for
ight hardware
ECSS-Q-ST-70-57C Dry heat bioburden reduction for ight
hardware
ECSS-Q-ST-70-58C Bioburden control of cleanrooms
ECSS-Q-ST-70-71C Materials, processes and their data
selection
ECSS-Q-ST-80C Software product assurance
ISO 24113: Space systemsSpace debris mitigation
requirements
JAXA-QTS-2120 Wire, Electric, Fluorine Resin/Polyimide
Insulated
JAXA-QTS-2140 Printed Wiring Boards, Rigid-Flexible
NASA-HDBK-5010 FRACTURE CONTROL IMPLE-
MENTATION HANDBOOK FOR PAYLOADS,
EXPERIMENTS, AND SIMILAR HARDWARE
NASA-HDBK-6024 Spacecraft polymers atomic oxygen
durability handbook
NASA-HDBK-6025 Guidelines for the specication and
certication of titanium alloys for NASA ight applications
NASA-HDBK-8719.14 Handbook for limiting orbital
debris
NASA-HDBK-8739.23 NASA complex electronics hand-
book for assurance professionals
NASA-STD-4003 ELECTRICAL BONDING FOR NASA
LAUNCH VEHICLES, SPACECRAFT, PAYLOADS,
AND FLIGHT EQUIPMENT
NASA-STD-5001 structural design and test factors of safety
for spaceight hardware
NASA-STD-5002 Load analyses of spacecraft and payloads
NASA-STD-5005 Standard for the design and fabrication of
ground support equipment
NASA-STD-5006 General fusion welding requirements for
aerospace materials used in ight hardware
NASA-STD-5008 Protective coating of carbon steel, stain-
less steel, and aluminum on launch structures, facilities, and
ground support equipment
NASA-STD-5009 Nondestructive evaluation requirements
for fracture critical metallic components
NASA-STD-5019 Fracture control requirements for space-
ight hardware
NASA-STD-5020 Requirements for threaded fastening
systems in spaceight hardware
NASA-STD-6001 Flammability, offgassing, and compati-
bility requirements and test procedures
NASA-STD-6008 NASA Fastener procurement, receiving
inspection, and storage practices for spaceight hardware
NASA-STD-6012 Corrosion protection for space ight
hardware
NASA-STD-6016 Standard materials and processes
requirements for spacecraft
NASA-STD-8719.14 Process for limiting orbital debris
NASA-STD-8719.9 Standard for lifting devices and
equipment
NASA-STD-8739.1 Workmanship standard for polymeric
application on electronic assemblies
NASA-STD-8739.4 Crimping, interconnecting cables, har-
nesses, and wiring
NASA-STD-8739.5 Fiber optic terminations, cable assem-
blies, and installation
NASA-STD-8739.6 Implementation requirements for nasa
workmanship standards (includes requirements for soldered
electrical and electronic assemblies per IPC J-STD-001
ES)
 Appendix 10: Examples of Declared
Process Lists (DPL)

The Processes selected and used on any spacecraft will need reports are generated: on critical processes are verication
to be approved by the end customer. They are compiled by tested.1
the Prime Contractor. Critical processes are usually eval- Common practice is to tabulate all processes according to
uated by Testing technology samples for which laboratory the following Group Numbers:

Group Description
number
1 Adhesive bonding
2 Composite manufacture
3 Encapsulation/moulding
4 Painting/coating
5 Cleaning
6 Welding/brazing
7 Crimping/stripping/wire wrapping
8 Soldering
9 Surface treatments
10 Plating
11 Machining
12 Forming
13 Heat treatment
14 Special fabrication: processes developed
specically for the programme
15 Marking
16 Miscellaneous processes
17 Inspection procedures

1
Critical processes are often specied by the end customer and will
depend on the service life and operational condition of individual
spacecraft (e.g. manned space vehicles may have more critical
processes). Often the process is considered critical if there are major
difculties or uncertainties in the manufacturing, assembly, inspection
and testing. The process will also be considered critical if it has proven
to be difcult to perform by trained operators, and if it has raised
problems in the past that have not been resolved.

Springer International Publishing Switzerland 2016 621

B.D. Dunn, Materials and Processes, Springer Praxis Books,
DOI 10.1007/978-3-319-23362-8
 622

Appendix 10.A.1
Outerspace company Declared process list DOC.NO.: DPL/PIT/3401/OSC
Spacecraft: PITCAIRNSAT 1 Group 1 adhesive bonding Issue: 3
Subsystem: Date: 10/04/14
Equipment: Page: 16
Item Process Specication R Description/Identication Use Location code User Assoc. C Approval Customer
E code DML R Status C comments and
V Items I approval
T
1.07.00 Preparation of VESPEL and PON H- 8 Bonding LLMKTRA IOLACADM 18.018.00 N PSU: A
PON DELRIN surfaces prior to A011 processes ROSETTA
bond
1.018.00 Bonding by conductive resins HR-012.INT 1 CURE: 15 min at 120 C Shielding of LLMKHPA 10.013.00 N PSU: A
ZTN with EPOTEX H20 E SSPA structure METRO 6
1.019.00 Screw locking with Screw 1 CURE: 48 h at RT or 3 h at LLMKHPA 10.009.00 N PSU: A
ZTN SOLITHANE 113 locking 65 C METRO 6
HR-013/INT
1.021.01 Bonding with SOLITHANE HR-411/INT 1 CURE: 48 h at RT or 3 h at Components KSPA 10.009.00 N PSU: A
ZTN 113 65 C bonding METRO 6
1.025.00 Bonding with ECCOSIL HR-410/INT 2 CURE: 18 h at RT or 3 h at RF Absorber LLMKHPA 10.069.03 N PSU: A
ZTN 4952 65 C bonding 12.012.00 METRO 6
16.006.02
1.026.00 Bonding with EA 9321 PLUS SP 4413 TC LLMCHRM IOLACHRM 10.002.03 A
ON EA 9210 PRIMER SP 4414 APSAPA
1.030.00 Preparation of BS L 312 SP 8841 CURE: 1 Hr at 120 C Inserts and LLMCHRM IOLACHRM 10.015.05 A
ON bonding of M
honeycomb
1.031.00 Bonding with FOAM 410-1 M-O-993 8 CURE: 0.5 h at 120 C + 1 h Splicing STR IOLACADM 10.027.01 N PSU: A
CAT 180 C honeycomb part LLMCADM HELIOS ISO
ARIANE 5
1.094.00 Locking of nuts and screws PIV/20/2014 3 Permenent and non- APSIAPC N A
PIN permanent locking of screws
and nuts
Appendix 10: Examples of Declared Process Lists (DPL)
Appendix 10.A.2
Outerspace company Declared process list DOC. NO.: DPL/PIT/3401/OSC
Spacecraft: PITCAIRNSAT 1 Group 6 welding/brazing Issue: 3
Subsystem: Date: 10/04/14
Equipment: Page: 61
Item Process Specication R Description/ Use Location User Assoc. C Approval Customer
number E identication code code DML R Status C comments
contractor V items I and
T approval
Appendix 10: Examples of Declared Process Lists (DPL)

6.018.00 Vacuum PPC PIC 101 1 Joining metal/or Thruster IPPEIT 7.038.00 C Awaiting O
PNT brazing metallised ceramic 20.051.03 vacuum O
components 3.011.03 hermeticity
6.001.36 testing and
6.013.15 DPA
3.036.00
6.020.00 Filling tube SPECWELD Laser welding Tube on top DOME PWSBTA 3.015.01 N RPT OSC A
TOT welding 491 3.015.03 4448 N
6.021.00 Cell closure M.T.U. TIG Welding Terminal to DOME PWSBTA 3.001.01 N RPT OSC A
TOT welding 11997 3.001.02 4521 N
21.017.01
6.022.00 Brazing with TBD Hard bazing of Cu- IPPEIT 6.008.03 N O
FOB VH 950 wire of 2.0040/anode 7.036.02 N
screw 1.4944
6.023.00 Microwelding FOP Plasmajet IPPEIT 2.001.32 N A
FOB 99-22/13 microwelding of 19.019.03 N
HF-RF-CABLE/Cu
wire
623
 624

Appendix 10.A.3
Outerspace company Declared process list DOC.NO.: DPL/PIT/3401/OSC
Spacecraft: PITCAIRNSAT 1 Group 9 space conversion treatment Issue: 3
Subsystem: Date: 10/04/14
Equipment: Page: 99
Item number Process Specication R Description/identication Use Location code User Assoc. C Approval Customer
contractor identication E code DML R Status C comments and
V Items I approval
T
9.012.00 Anodising of 33448/22/4 A Protection RCPROMD FREPFKIF 4.001.43 N RPT: A
XYZ TA6V titanium FREPFKOF FREPFBOF 21.004.19 XYZ/BOT/FT.236
alloy FREPFKTF RREPRKIF 21.007.30 PSU: TDF
RREPRKPD 21.008.51 8 TVSAT 8
TLC 8
POS Anodising of 33448/22/4 A Protection FREPFKLA FREPFICA 4.001.43 N PSU: ECS INTE-V1 A
TA6V titanium FREPFISC FREPRSIR 21.004.19 GIOTTO ERS-1 N
alloy FREPAIT 21.007.30
21.008.51
9.013.00 Surface passivation 62410/99/1 Passivation of silver Protection RCPROMD FREPFKIF 1.029.03 N RPT: BAB/QT/ A
XYZ FREPFKOF FREPFBOF 3.005.18 81.044
FREPFKIF RREPRKIF 4.007.01 PSU: TVSAT 2
RREPRKPD SPLAB
TLC 8
POS Surface passivation 44291/20/1 Passivation of silver Protection FREPFKLA FREPFICA 3.002.11 N RPT: BAB/QT/ AN
FREPFISC FREPRSIR 1.029.02 81.044
1.029.03 PSU: INTE.V1
1.040.00 ERS-1
1.029.04 EURECA
EUTE.2
9.014.00 Anodising DTD 942 Surface treatment on Ti Surface FREPFCFK 21.004.21 N MARS-2 AN
MOM alloys protection on
xings
9.015.00 Anodising of LIN.914.2. A Support and RREPRICA RREPRKPA 1.017.04 N CLUST-2 AN
LIN aluminium ALLOY rev.1 cooling plate RREPRIKC
9.018.00 Hard anodising 49326/20/1 B Treatment of aluminium Insulation FREPFKLA FREPFICA 1.002.34 N PSU: INTE.V1 AN
POS and aluminium alloy FREPFISC FREPRSIR 1.007.15 GIOTTO
1.029.05 EURECA TLC 2
9.020.00 Chromate 40441/20/2 B Treatment of aluminium Protection FREPFKLA FREPFICA 1.002.35 N RPT: R:
POS conversion coating and aluminium alloy FREPFISC FREPRSIR 1.002.38 XYZ/BOT/FT
1.007.16 PSU: ECS
FREPAIT
1.007.19 INTE.V1
1.029.06 ISS TDF 8
1.040.02
Appendix 10: Examples of Declared Process Lists (DPL)
 Appendix 11: Examples of Declared
Materials Lists (DMLs)
The materials selected and used on any spacecraft will need
to be approved by the end customer. They are compiled by
the Prime Contractor.
Critical materials are tested and Validated. Dependent
upon the criticality of the spacecraft, tests may include:
outgassing-under-vacuum, stress corrosion testing,
Springer International Publishing Switzerland 2016
B.D. Dunn, Materials and Processes, Springer Praxis Books,
DOI 10.1007/978-3-319-23362-8
ammability, etc. The Material Group numbers have been
detailed in Appendix 8.
The contents of the DMLs will depend on the end cus-
tomers requirements, these may follow ECSS-Q-ST-70 or
can be relaxed in view of the specic contractual require-
ments. Further description is given in Sect. 4.2.1.
625
 626

Appendix 11.A.1
Outerspace company Declared material list DOC. NO.: DML/PIT/3466/OSC
Spacecraft: PITCAIRNSAT 1 Group 2 copper and copper alloys Issue: 3
Subsystem: Date: 10/04/14
Equipment: Page: 49
Item Commercial Chemical nature and Procurement Summary of processing Use and location code ENVIR. Size Approval Customers
identication type of product for information parameters code comments
applied condition supplies Use Code R A T Status C and approval
specication
2.021.01 Brass 40Zn Cu 60/Zn 40 QQ-B-626 TIN plating 8AB 00526 Mechanical parts RREPRKIC GS V 3 V1 PSU: BSAT D No pure tin
FOK AAAA FREPFIKC 4 permitted
FREPFKIC
RCPFRGU
2.022.00 Brass 38zn Cu Zn 38, 2Pb 734.662.E Gold plating 4 m Mechanical parts: RREPRKIC GS V 3 V0 PSU: ECS BSAT A
FOK 2Pb VAR. TUN.SCREW Copper underplating bonding stud FREPFIKC 4 EUR
NRS 5 m terminal, screw lock FREPFKIC
VAR. TERMINAL 770-82608-AASG RCPFRGU
NRS
2.022.01 Brass CZ121 Cu58, ZN 38-39, Pb 3-4 Aluminium Silver plating Center conductor FREPFCFK GS V 3 W1 PSU: SPOT A
MOM ROD supplies BS 2874 Solder Low pass lter FREPSDIP 4 N
CDP 7.10E Probe and tube PREPRINF
CDP 7.03
2.022.02 Brass 38Zn Cu Zn 38, 2Pb 734.662.E Sn plating deleted Ag Mechanical parts RREPRKIC GS V 3 V0 PSU: ECS BSAT A Tin no longer
FOK 2Pb VAR. TN.SCREW plating Bonding stud, FREPFIKC 4 EUR N used
NRS 8AB 00526 AAAA terminal, screw lock FREPFKIC
VAR. TERMINAL RCPFRGU
NRS
2.023.00 Copper Cu Sn6 DIN 1756 Pretinning Piece, Parts RREPRKPA GS V 3 W1 PSU: A
LIN CDA 510 ROD GZ.1991.209.1 FREPFKPA 4 TELECOM HISPAS.
INT.V11 EUTELS
ECSS-Q-ST-70-36
2.024.00 Copper AVIOMETAL Electrolytic surface Electrical IPPPSCU GS V 3 W2 PSU: GIOTTO ERS A
SINT 131N22 coating Sn and reowed connections 4 PINSAT M
MIL-T-107-27 1
IP 5009
2.026.00 Copper Cu65 % zn35 % TELKRON Plated with Cu EYELET PCDICU GS V 3 W2 PSU: ISO AN
LIN CDA 260 CRS-EPT-6016 5 m + Sn/Pb PCDPKDU 4
2.027.00 Brass Grommet with Spur Atlanta hardware INSERT MLI grounding IOLAIKCM GS V 3 W3 A
FON MS 20230 grommet plain N
Appendix 11: Examples of Declared Materials Lists (DMLs)
Appendix 11.A.2
Outerspace company Declared material list DOC: NO.: DML/PIT/3466/OSC
Spacecraft: PITCAIRNSAT 1 Group 6 stainless steels Issue: 3
Subsystem: Date: 10/04/14
Equipment: Page: 88
Item Commercial Chemical nature and type of Procurement information Summary of Use and location code ENVIR. Size Approval Customers
identication product for appl. condition supplies specication processing Code comments and
parameters Use Code R A T Status C approval
6.034.00 SS 2343-08 TINGSTAD A/S Lock wire IPPFVV GS E 3 W1 A
FIN F9967.1 4 N
6.035.00 AISI 304L LEE JEVA Passivated FLO restrictor in IPPPSME GS V 3 W1 RPT 3991- A
SMM 09-TSL MIL-S-5002 107 series valve 4 02099-DPQ N
6.036.00 AISI 348 Stainless steel sheet Various Formed Cathode, main IPPEIT GS V 4 SCC I A
CPP PPC 823-170-95911 Spot welded ange NASA NASA- N
Appendix 11: Examples of Declared Materials Lists (DMLs)

PPC 823-170-95981 HDBK-527

10119
6.036.01 AISI 348 Stainless steel ROT Machined Structural parts IPPEIT GS V 4 W1 SCC I A
CPP PPC 8213-170-95921 Spot welded NASA NASA- N
HDBK-527
10119
6.037.00 Nitriding BS S106D Plasma nitride Gear SADM GS V 3 W2 ? O More data required
BET steel 4 h at 555 C 4 N
ESP4443
6.038.00 Maraging Bar DTD 5212 Plasma nitride Pinion SADM GS V 3 W1 ? O More data required
BET steel 4 h at 485 C 4 N
ESP4443
6.039.00 Corrosion Sheet BS S527 Washer SADM GS V 3 W1 A
BET resisting steel 4 N
6.040.00 Stainless BS S80 PASSIVE SADM GS V 3 W2 A
BET steel PS 2089 4 N
6.041.00 AISI 304 X8 Cr Ni 19/10 MULTI SOURCE USE AS IS Ball for sealing IPPFCU GS V 3 W1 PSU: A
XIL AMS-5513/AMS 5639 4 OLYMPUS
HIPPAR
METEO
6.042.00 AISI 1017 Nickel plated Electical stop APSAPA GS V 2 W2 A
ON 5 to 20 m Tooth 4 N
627
 628

Appendix 11.A.3
Outerspace company Declared material list DOC. NO.: DML/PIT/3466/OSC
Spacecraft: PITCAIRNSAT 1 Group 10 adhesive, coatings, varnishes Page: 151
Issue: 3
Subsystem:
Date: 10/04/14
Equipment:
Item Commercial Chemical nature and Procurement Summary of Use and location code ENVIR. Size Approval Customers
identication type of product for information processing Code comments
Appl. condition supplies parameters Use Code R A T Status C and approval
specication
10.014.03 EPO-TEK H 74 EPOX resin A/B Epoxy technol. MIX: 100 pp/ Bonding loops and RREPRIWF GS V 3 W1 PSU: O ECS A
MOM two component CDP 9.39 3 pp supports into housing/ FREPFCFK 4 TTL
electric. conductive CURE: 2 h at bonding ferrites FREPSDIP NASA: NASA RP
100 C TTCSBDN 1124
CDP 7.08
10.015.03 REDUX 312L Epoxy resin lm CIBA GEIGY CURE: Skins to core bonding LLMKTRA GS V 1 W3 RFA: W
CAT 1+D-N-15 90 min at GL 3 RFW/KANT-01/CAS
1+D-N-200 120 C ECSS-Q-ST-70-01
1+D-P-70 TVS 14
10.015.04 REDUX 312UL+ Epoxy resin lm CIBA GEIGY CURE: Skins to core SKDRKANA GS V 1 W2 RFA WM
CAT REDUX 1+D-N-15 60 min at STR GL 3 RFW/KANT-01/CAS
120 C ECSS-Q-ST-70-01
1+D-P-70 TVS 14
10.015.05 REDUX 312L Epoxy resin lm CIBA GEIGY Insert plate LLMCHRM GS V 3 V2 RPT: LT 33 A
ON DSN. 0016 IOLACHRM
10.015.06 REDUX 312/P112 Epoxy adhesive lm CIBA GEIGY CURING: Structural bonding in IOLACADM GS V 3 W3 RPT: BOT REP.002 A
CAT 1+D-N-15E 1 h/120 C sandwich LLMCADM PSU: HELIOS, ISO N
1+D-E-159 1+D-P-70 manufacturing ARIANE6
TVS 14
10.015.07 REDUX 312/5 Epoxy resin lm CIBA CEIGY Bonding UPSGT GS E 3 V2 RPT LT 33 A
ON NT 16101/AQEN ECSS-Q-ST-70-01
NT 16102/AQEN TVS 14
10.015.08 REDUX 312L Epoxy resin lm CIBA CEIGY CURE: Bonding of panels SAW GL V 2 A4 RPT ESA I 668 A
KOF TH5. 917/4/5/6 90 min at 4 PSU: ARA IRAS N
120 C OLYMPUS
TH24. 3006 ULYSSES
TVS 14
10.015.09 REDUX312L/112 Epoxy resin plus CIBA CEIGY CURE: 1 h at Beam APSAPA GS V 2 W2 RFA W
ON primer lm Data sheet/GENES 120 C 4 RFW/GOT-APA/01 N
0021-1184/1179 PSU: EUT
TVS 14
Appendix 11: Examples of Declared Materials Lists (DMLs)
 Glossary
The majority of terms have been compiled from documents issued by ESA, IOM, ISO, NASA and TMS.
Acicular alpha A product of nucleation and growth from
beta to the lower-temperature allotrope alpha phase. It may
have a needlelike appearance in a photomicrograph and may
have needle, lenticular, or attened bar morphology in three
dimensions. Its typical aspect ratio is about 10:1
Activation The changing of a passive surface of a metal to a
chemically active state. (Contrast with passivation.)
Age hardening Hardening by ageing, usually after rapid
cooling or cold working. (See also ageing.)
Ageing A change in the properties of certain metals and
alloys that occurs with time at ambient or moderately ele-
vated temperatures after working or a heat treatment (natural
or articial ageing) or after a cold-working operation (strain
ageing). The change in properties is often, but not always,
due to a phase change (precipitation), but it never involves a
change in chemical composition of the metal or alloy
Alloy A substance having metallic properties and being
composed of two or more chemical elements of which at
least one is a metal
Alphabeta structure A microstructure containing and
as the principal phases at a specic temperature
Alpha case The oxygen-, nitrogen-, or carbon-enriched, -
stabilized surface resulting from elevated temperature
exposure
Analysis The determination of the essential qualities, per-
formance, and limitations of an item by cognitive or com-
putational methods
Anomaly Any deviation from the expected situation
Assurance All the planned and systematic activities
implemented, and demonstrated as needed, to provide ade-
quate condence that an entity will full its requirements
Springer International Publishing Switzerland 2016
B.D. Dunn, Materials and Processes, Springer Praxis Books,
DOI 10.1007/978-3-319-23362-8
Annealing A generic term denoting a treatment, consisting
of heating to, and holding at, a suitable temperature followed
by cooling at a suitable rate. It is used primarily to soften
metallic materials, but also to simultaneously produce
desired changes in other properties or in microstructure. The
purpose of such changes may be, but is not conned to:
improvement of machinability, facilitation of cold work;
improvement of mechanical or electrical properties, and/or
increase in stability of dimensions. When the term is used
without qualication, full annealing is implied. When
applied only for the relief of stress, the process is properly
called stress relieving or stress-relief annealing
Assembly A functional subdivision of a component, con-
sisting of parts or subassemblies that perform functions
necessary for the operation of the component as a whole.
Examples: regulator assembly, power amplier assembly,
gyro assembly, etc
Assurance All the planned and systematic activities
implemented, and demonstrated as needed, to provide ade-
quate condence that an entity will full its requirements
Axial lead Lead wire extending from a component or
module body along its longitudinal axis
AWG American wire gauge
Basketweave Alpha platelets with or without interleaved
platelets that occur in colonies in a Widmansttten structure
Batch That quantity of material that was subjected to unit
chemical processing or physical mixing, or both, designed to
produce a product of substantially uniform characteristics
Bifurcated (split) terminal A terminal with a slot or split
opening in which conductors are placed before soldering
Billet (1) A solid, semi-nished round or square product
that has been hot-worked by forging, rolling, or extrusion;
629
630
usually smaller than a bloom. (2) A general term for wrought
starting stock used in making forgings or extrusions
Blister Undesirable rounded elevation of the surface of a
polymer, whose boundaries may be more or less sharply
dened
Body-centred cubic lattice structure A unit cell which
consists of atoms arranged at cube corners with one atom at
the centre of the cube
Brazeability The capacity of a metal to be brazed under the
fabrication conditions imposed into a specic suitably
designed structure and to perform satisfactorily in the
intended service
Brazing A group of processes that join solid materials
together by heating them to a suitable temperature and by
using a ller metal having a liquidus above about 450 C
(840 F) and below the solidus of the base materials. The
ller metal is distributed between the closely tted surfaces
of the joint by capillary attraction
Bridging A build-up of solder or conformal coating
between parts, components leads, and/or base substrate
forming an elevated path (see Fillet)
Brittle Permitting little or no plastic (permanent) deforma-
tion prior to fracture
Brittle fracture Separation of a solid accompanied by little
or no macroscopic plastic deformation. Typically, brittle
fracture occurs by rapid crack propagation with less
expenditure of energy than for ductile fracture
Cable Two or more insulated conductors, solid or stranded,
of equal length, contained in a common covering; or two or
more insulated conductors, of equal length, twisted or
moulded together without common covering; or one insu-
lated conductor with a metallic covering shield or outer
conductor (shielded cable or coaxial cable)
Cast or casting Top fabricate an item by pouring molten
metal into a shaped cavity and permitting the metal to
solidify. A cast can relate to the item or may be a synonym
for heat, that is an identiable chemistry lot
Catalyst A substance that changes the rate of a chemical
reaction without undergoing permanent change in its com-
position; a substance that markedly speeds up the cure of a
compound when added in minor quantity as compared to the
amount of primary reactants
Certication The act of verifying and documenting that
personnel have completed required training and have
demonstrated specied prociency and have met other
specied requirements
Glossary
Chemical vapour deposition (CVD) The precipitation of a
metal from a gaseous compound onto a solid or particulate
substrate
Coarse grains Grains larger than normal for the particular
wrought metal or alloy or of a size that produces a surface
roughening known as orange peel or alligator skin in
wrought alloys
Cold ow Movement of insulation (e.g. Teon) caused by
pressure
Cold solder connection A solder connection exhibiting
poor wetting and a greyish, porous appearance due to
insufcient heat, inadequate cleaning before to soldering, or
excessive impurities in the solder
Cold working Deforming metal plastically under condi-
tions of temperature and strain rate that induce strain hard-
ening. Usually, but not necessarily, conducted at room
temperature. (Contrast with hot working.)
Cold-worked structure A microstructure resulting from
plastic deformation of a metal or alloy below its recrystal-
lization temperature
Colophony A natural resin obtained as the residue after
removal of turpentine from the oleoresin of the pine tree,
consisting mainly of abietic acid and related resin acids, the
remainder being resin acid esters
Component A functional subdivision of a system, generally
a self-contained combination of assemblies performing a
function necessary for the systems operation. Examples:
power supply, transmitter, gyro package, etc
Conductor A lead, solid or stranded, or printed wiring path
serving as an electrical connection
Conguration Functional and physical characteristics of a
product as dened in technical documents and achieved in
the product (ISO 10007:1995)
Conformal coating A thin electrically nonconductive pro-
tective coating that conforms to the conguration of the
covered assembly
Contact angle The angle enclosed between half-planes,
tangent to a liquid surface and a solidliquid interface at their
intersection. In particular, the contact angle of liquid solder in
contact with a solid metal surface. An approximate value for
this may be determined by shadow projection or other means,
by measuring after the solder has solidied. Note that the
contact angle is always the angle inside the liquid
Contaminant An impurity or foreign substance present in a
material that affects one or more properties of the material.
Glossary
A contaminant may be either ionic or nonionic. An ionic, or
polar, compound forms free ions when dissolved in water,
making the water a more conductive path. A nonionic sub-
stance does not form free ions, nor increases the waters
conductivity. Ionic contaminants are usually processing
residue such as ux activators, nger prints, and etching or
plating salts
Contractor Supplier in a contractual situation (ISO
8402:1994)
Corrective action Action taken to eliminate the causes of
an existing nonconformity, defect, or other undesirable sit-
uation in order to prevent recurrence (ISO 8402: 1994)
Corrosion The deterioration of a metal by a chemical or
electrochemical reaction with its environment
Corrosion fatigue Cracking produced by the combined
action of repeated or uctuating stress and a corrosive
environment at lower stress levels or fewer cycles than
would be required in the absence of a corrosive environment
Creep Time-dependent strain occurring under stress. The
creep strain occurring at a diminishing rate is called primary,
or transient, creep; that occurring at a minimum and almost
constant rate, secondary, or steady-rate creep; that occurring
at an accelerating rate, tertiary creep
Crevice corrosion A type of concentration cell corrosion;
corrosion caused by the concentration or depletion of dis-
solved salts, metal ions, oxygen, or other gases, and such, in
crevices or pockets remote from the principal uid stream,
with a resultant building up of differential cells that ulti-
mately cause deep pitting. Localized corrosion of a metal
surface at, or immediately adjacent to, an area that is shiel-
ded from full exposure to the environment because of close
proximity between the metal and the surface of another
material
Critical item Any item that introduces risk which could be
unacceptable to the project and requires specic attention or
control in addition to that given to items not so categorized
Cure A chemical reaction that hardens and changes the
physical properties of a material
Deformation A change in the form of a body due to stress,
thermal change, change in moisture, or other causes. Mea-
sured in units of length
Delamination A separation between plies within a base
material or any planar separation within a multilayer printed
circuit board (PCB)
Design (1) Set of information which denes the essential
characteristics of a product. (2) The process used to generate
631
the set of information describing the essential characteristics
of a product
Dewetting The condition in a soldered area in which the
liquid solder has not adhered intimately, but has receded,
characterized by an abrupt boundary between solder and
conductor, or solder and terminal/termination area leaving
irregularly shaped mounds of solder separated by areas
covered with a thin-solder lm
Diffusion welding (DFW) A high-temperature, solid-state
welding process that permanently joins faying surfaces by
the simultaneous application of pressure and heat. The pro-
cess does not involve macroscopic deformation, melting, nor
relative motion of parts. A solid ller metal (diffusion aid)
may or may not be inserted between the faying surfaces
Disturbed solder joint Unsatisfactory connection resulting
from relative motion between the conductor and termination
during solidication of the solder
Dross Oxide and other contaminants that form on the sur-
face of molten solder
Ductility The ability of a material to deform plastically
before fracturing. Measured by elongation or reduction of
area in a tension test, by height of cupping in a cupping test,
or by the radius or angle of bend in a bend test. (Contrast
with brittleness; see also plastic deformation.)
Electron beam welding (EBW) A welding process which
produces coalescence of metals with the heat obtained from
a concentrated beam composed primarily of high-velocity
electrons impinging upon the surfaces to be joined
Elongated grain A grain with one principal axis signi-
cantly longer than either of the other two
Elongation A term used in mechanical testing to describe
the amount of extension of a test-piece when stressed. In
tensile testing, the increase in the gauge length, measured
after fracture of the specimen within the gauge length, usually
expressed as a percentage of the original gauge length
Embrittlement The severe loss of ductility and/or tough-
ness of a material, usually a metal or alloy
Encapsulating compound An electrically nonconductive
compound used to completely enclose and ll in voids
between electrical components or parts
Environment Conditions in which an item exists or is
operated
Equiaxed structure A polygonal or spheroidal
microstructural feature having approximately equal dimen-
sions in all directions. In alphabeta titanium alloys, such a
632
term commonly refers to a microstructure in which most of
the minority phase appears spheroidal
Equilibrium A dynamic condition of physical, chemical,
mechanical, or atomic balance, where the condition appears
to be one of rest rather than change
Equipment An item designed and built to accomplish a
specied purpose, which can be disassembled and retain its
capabilities after reassembly
Eutectic alloy An alloy of two or more metals that has one
distinct melting point. Eutectic solder is a tinlead alloy
containing 63 %Sn and 37 %Pb which melts at 183 C
Excessive solder joint Unsatisfactory solder connection
wherein the solder obscures the conguration of the
connection
Face-centred cubic lattice structure A unit cell which
consists of atoms arranged at cube corners with one atom at
the centre of each cube face
Failure The termination of the ability of an item to perform
a required function
Fatigue The phenomenon leading to fracture under repeated
or uctuating stresses having a maximum value less than the
tensile strength of the material. Fatigue fractures are pro-
gressive, beginning as minute cracks that grow under the
action of the uctuating stress
Fatigue failure Failure that occurs when a specimen
undergoing fatigue completely fractures into two parts, or
has softened, or been otherwise signicantly reduced in
stiffness by thermal heating or cracking. Fatigue failure
generally occurs at loads which, if applied statically, would
produce little perceptible effect. Fatigue failures are pro-
gressive, beginning as minute cracks that grow under the
action of the uctuating stress
Fault mode The observable effect of the mechanism
through which the failure occurs, e.g. short, open, fracture,
excessive wear
Fillet A smooth concave build-up of material between two
surfaces, e.g. a llet of solder between a component lead and
a solder pad or terminal, or a llet of conformal coating
material between a component and printed circuit board
(PCB)
Filler metal The metal to be added in making a welding,
brazed, or soldered joint
Flake Powder of an essentially two-dimensional nature
Flatpack A part with two straight rows of leads that are
parallel to the part body
Glossary
Flux A chemically active compound which, when heated,
removes minor surface oxidation
Foil Among many denitions is: a at-rolled product
0.127 mm (0.005 in.), or less, in thickness, regardless of
width. (Any at-rolled product thicker than this dimension is
not considered foil.) Only thickness, not width, is a factor in
determining foil
Forging (1) Plastically deforming metal, usually hot, into
desired shapes with compressive force, with or without dies.
(2) Reshaping a billet or ingot by hammering. (3) The process
of placing a powder in a container, removing the air from the
container, and sealing it. This is followed by conventional
forging of the powder and container to the desired shape
Fracture (1) The irregular surface produced when a piece
of metal is broken. (2) To separate a metal or alloy into two
or more pieces by an applied load
Friction welding A solid-state process in which materials
are welded by the heat obtained from rubbing together sur-
faces that are held against each other under pressure
Glass meniscus The glass llet of a lead seal which occurs
where an external lead leaves the package body
Grounding The connection of two or more areas to the
same potential difference. The space environment can result
in high potential voltage and destructive arcing, most sub-
systems are grounded to the structure, but the grounding of
electronic units is complex and can use a oating ground
concept
Ground segment The hardware and software required to
launch and control a space vehicle, usually the launcher
Hardness A measure of the resistance of a material to
surface indentation or abrasion; may be thought of as a
function of the stress required to produce some specied
type of surface deformation. There is no absolute scale for
hardness; therefore, to express hardness quantitatively, each
type of test has its own scale of arbitrarily dened hardness.
Indentation hardness may be measured by Brinell, Knoop,
Rockwell, Scleroscope, and Vickers hardness tests
HAZ See heat-affected zone
Heat-affected zone That portion of the base metal which
has not been melted, but whose mechanical properties or
microstructure have been altered by the heat of welding,
brazing, soldering, or cutting
Heat treatment Heating and cooling a solid metal or alloy
in such a way as to obtain desired conditions or properties.
Heating for the sole purpose of hot working is excluded from
the meaning of this denition
Glossary
Hermetic seal A seal which protects an enclosed circuit
from corrosion by preventing the entry of such contaminants
as water vapour
HIP See hot isostatic pressing
Hot isostatic pressing (1) A process for simultaneously
heating and forming a powder metallurgy compact in which
metal powder, contained in a sealed exible mould, is sub-
jected to equal pressure from all directions at a temperature
high enough for sintering to take place. (2) A process similar
to the one explained in (1), but applied to castings in order to
close internal porosity
Hot working Deforming metal plastically at such a tem-
perature and strain rate that recrystallization takes place
simultaneously with the deformation, thus avoiding any
strain hardening
HV See Vickers hardness number
Hydride phase For instance, the phase TiHx formed in
titanium when the hydrogen content exceeds the solubility
limit, generally locally due to some special circumstance
Hydrogen embrittlement A condition of low ductility in
metals resulting from the absorption of hydrogen
Immersion cleaning Cleaning where the work is immersed
in a liquid solution. Impurities, Undesirable elements or
compounds in a material
Inclusion A particle of foreign material in a metallic matrix.
The particle is usually a compound (such as an oxide, sul-
phide, or silicate), but may be of any substance that is for-
eign to (and essentially insoluble in) the matrix
Ingot A casting of simple shape, suitable for hot working or
remelting
Inspection An activity such as measuring, examining,
testing, or gauging one or more characteristics of an entity
and comparing the results with specied requirements in
order to establish whether conformity is achieved for each
characteristic (ISO 8402:1994)
Insufcient solder connection A solder connection char-
acterized by incomplete coverage of one or more of the
metal surfaces being joined or by incomplete solder llets
Integrated-circuit component An individually packaged
functional circuit formed by depositing an active or passive
electronic element on to a substrate
Intermetallic compound A phase in an alloy system hav-
ing a restricted solid solubility range. Nearly all are brittle
and of stoichiometric composition
Interstitial element An element with a relatively small
atom which can assume a position in the interstices of a
633
metal, or metal alloy lattice. Common examples are oxygen,
nitrogen, hydrogen, and carbon
Interstitial solid solution A type of solid solution that
sometimes forms in alloy systems having two elements of
widely different atomic sizes. Elements of small atomic size,
such as carbon, hydrogen, oxygen, and nitrogen, often dis-
solved in solid metals to form this solid solution. The space
lattice is similar to that of the pure metal, and the atoms of
carbon, hydrogen, oxygen, and nitrogen occupy the spaces
or interstices between the metal atoms
Item Anything which can be individually described and
considered
Lap joint Joining or fusing of two overlapping metal sur-
faces with solder without use of any other mechanical
attachment or support
Liquidus In a constitution or equilibrium diagram, the locus
of points representing the temperatures at which the various
compositions in the system begin to freeze on cooling or
nish melting on heating. (See also solidus.)
Longitudinal direction Usually, the direction parallel to
the direction of working in wrought alloys or the direction of
crystal growth in directionally solidied or single-crystal
cast alloys. Commonly, it corresponds to the direction par-
allel to the direction of maximum elongation in a worked
material. (See also normal direction and transverse
direction.)
Machinability The relative ease of machining a metal
Machining Removing material from a metal part, usually
using a cutting tool, and usually using a power-driven
machine
Macrostructure The structure of metals as revealed by
macroscopic examination of a specimen. The examination
may be carried out using an as-polished or a polished and
etched specimen
Martensite (1) The alpha product resulting from cooling
from the beta phase region at rates too high to permit
transformation by nucleation and growth. Martensite is sat-
urated with beta stabilizer. Also called martensitic alpha. (2)
A generic term for micro-structures formed by diffusionless
phase transformation in which the parent and product phases
have a specic crystallographic relationship. Martensite is
characterized by an acicular pattern in the microstructure in
ferrous and nonferrous alloys. The amount of high-temper-
ature phase that transforms to martensite upon cooling
depends to a large extent on the lowest temperature attained,
there being a distinct starting temperature (Ms) and a tem-
perature at which the transformation is essentially complete
(Mf), which is the martensite nish temperature
634
Maintainability The ability of an item under given condi-
tions of use, to be retained in, or restored to, a state in which
it can perform a required function, when maintenance is
performed under given conditions and using stated proce-
dures and resources (IEC 50:1992)
Material A raw or semi-nished substance or compound, of
specied characteristics, which is processed to form a part of
a nished product
Matrix The constituent which forms the continuous or
dominant phase of a two-phase microstructure
Measling/measles A condition existing in the base laminate
of printed-circuit board in the form of discrete white spots or
crosses below the surface of the base laminate, reecting a
separation of bres in the glass cloth at the weave intersec-
tion. During soldering, may be caused by excessive heat,
mechanical stresses, or chemical attack
Mechanical part A piece of hardware which is not elec-
trical, electronic, or electromechanical, and which performs
a simple (elementary) function or part of a function in such a
way that it can be evaluated as a whole against expected
performance requirements and cannot be disassembled
without destroying this capability
Mechanical properties The properties of a material that
reveal its elastic and inelastic (plastic) behaviour when force
is applied, thereby indicating its suitability for mechanical
(load-bearing) applications. Examples are elongation, fatigue
limit, hardness, modulus of elasticity, tensile strength, and
yield strength
Melting point The temperature at which a pure metal,
compound, or eutectic changes from solid to liquid; the
temperature at which the liquid and the solid are in
equilibrium
Mission The specic task, duty, or function dened to be
accomplished by a system
Modulus of elasticity A measure of rigidity or stiffness of a
metal; the ratio of stress, below the proportional limit, to the
corresponding strain
Nonconformance Nonfullment of a specied requirement
(ISO 8402:1994denition of Nonconformity)
Nonwetting A condition whereby a surface has contacted
molten solder, but the solder has not adhered to all of the
surface; basis metal remains exposed
ODS Oxide dispersion strengthening
Outgassing The release of volatile parts from a substance
when placed in a vacuum environment
Glossary
Overageing Ageing under conditions of time and temper-
ature greater than those required to obtain maximum change
in a certain property. (See ageing.)
Overheated joint An unsatisfactory solder joint, charac-
terized by rough solder surface; dull, chalky, grainy, porous
or pitted
Oxidation (1) A reaction in which there is an increase in
valence resulting from a loss of electrons. (Contrast with
reduction.) (2) A corrosion reaction in which the corroded
metal forms an oxide; usually applied to reaction with a gas
containing elemental oxygen, such as air
Part An element of a component, assembly, or subassembly
that is not normally subjected to further subdivision or dis-
assembly without destruction of designed use
Passivation The changing of a chemically active surface of
a metal to a much less reactive state
Performance Those generally quantied aspects of an item
observed or measured from its operation or function
Pickling Removal of the oxide lm on a casting by a
chemical process; pickling is sometimes used solely to show
up defects
Physical properties Properties of a metal or alloy that are
relatively insensitive to structure and can be measured without
the application of force; for example, density, electrical con-
ductivity, coefcient of thermal expansion, magnetic perme-
ability, heat capacity, and lattice parameter. Does not include
chemical reactivity. (Compare with mechanical properties.)
Plastic deformation The permanent (inelastic) distortion of
metals under applied stresses
Plate A at-rolled metal product of some minimum thick-
ness and widthat times less than 610 mm (24 in). (It is
relatively thick when compared with sheet.)
Plated-through hole A plated-through hole is one formed
by a deposition of metal on the inside surface of a through
hole. Also known as a supported hole. The conguration is
used to provide additional mechanical strength to the sol-
dered termination or to provide an electrical interconnection
on a multilayer PCB
Potting compound An electrically nonconductive com-
pound used to partially encapsulate or for a ller between
parts, conductors, or assemblies
ppm Parts per million
Precipitation (1) Separation of a new phase from solid or
liquid solution, usually with changing conditions of time,
temperature, and stress. (2) The removing of a metal from a
Glossary
solution caused by the addition of a reagent by displacement.
(3) The removal of a metal from a gas by displacement
Precipitation hardening Hardening caused by the precip-
itation of a constituent from a supersaturated solid solution
Preform An initially pressed compact to be subjected to
repressing or forging
Preheat An early stage in the sintering procedure when, in a
continuous furnace, lubricant or binder burnoff occurs
without atmosphere protection prior to actual sintering in the
protective atmosphere of the high heat chamber
Pressure vessel A container which stores pressurized uids
and: (a) contains stored energy of 19,310 joules or greater,
based on the adiabatic expansion of a perfect gas; or (b)
contains a gas or liquid which may result in a hazardous
event if released (c) will experience a design limit pressure
greater than 0.69 MPa.
Printed circuit board (PCB) A product resulting from the
process of selectively etching unwanted copper from one or
both surfaces of a copper-clad insulating substrate to form a
desired circuity pattern which is subsequently solder- or
gold-plated. Holes may, or may not, be drilled in the board,
depending on the intended technique for attaching
components
Procedure Specied way to perform an activity
Process Set of interrelated resources and activities which
transform inputs into outputs
Product The result of activities or processes and may
include service, hardware, processed materials, software, or
a combination thereof
Product assurance A discipline devoted to the study,
planning, and implementation of activities intended to
ensure that the design, controls, methods, and techniques in
a project result in a satisfactory level of quality in a product
Project A unique set of coordinated activities, with denite
starting and nishing points, undertaken by an individual or
organization to meet specic objectives within dened
schedule, cost, and performance parameters (BS 6079)
Qualication The process of determining that the product, as
designed, is capable of meeting all its specied performance
requirements in its operational environment with margins
appropriate for the technologies used and the intended
application
Quality The totality of characteristics of an entity that bear
on its ability to satisfy stated and implied needs
635
Quality assurance All the planned and systematic activities
implemented within the quality system and demonstrated as
needed, to provide adequate condence that an entity will
full requirements for quality
Quality control Operational techniques and activities that
are used to full requirements for quality
Quenching Rapid cooling. When applicable, the fol-
lowing more specic terms should be used: direct
quenching, fog quenching, hot quenching, interrupted
quenching, selective quenching, spray quenching, and
time quenching
Ram direction he side that points in the direction of the
spacecrafts motion. In low earth orbit, it is the side
impacting/ramming into the rareed atmosphere that con-
tains reactive atomic oxygen (AO, or ATOX)
Recrystallization (1) Formation of new, strain-free grain
structure from the structure existing in cold-worked metal.
(2) A change from one crystal structure to another, such as
that occurring upon heating or cooling through critical
temperature
Reducing atmosphere A chemically active protective
atmosphere which at elevated temperature will reduce metal
oxides to their metallic state. (Reducing atmosphere is a
relative term and such an atmosphere may be reducing to
one oxide but not to another oxide.)
Reduction in area (1) Commonly, the difference, expressed
as a percentage of original area, between the original cross-
sectional area of a tensile test specimen and the minimum
cross-sectional area measured after complete separation. (2)
The difference, expressed as a percentage of original area,
between original cross-sectional area and that after straining
the specimen
Reliability The probability that an item can perform a
required function under given conditions for a given time
interval
Rem Remainder
Repair Operations performed on a nonconforming article to
place it in usable condition. Repair is distinguished from
rework in that alternate processes rather reprocessing are
employed
Residual stress Stress remaining in a structure or member
as a result of thermal or mechanical treatment or both.
Stress arises in fusion welding primarily because the weld
metal contracts on cooling from the solidus to room
temperature
636
Resistance brazing Brazing by resistance heating, the joint
being part of the electrical circuit
Rework The reprocessing of an article or material that will
make it conform to drawings, specications, and contract
Risk A quantitative measure of the magnitude of a potential
loss and the probability of incurring that loss
Safety A state in which the risk of harm (to persons) or
damage is limited to an acceptable level
Scaling (1) Forming a thick layer of oxidation products on
metals at high temperature. (2) Depositing water-insoluble
constituents on a metal surface, as in cooling tubes and water
boilers
SCC Stress-corrosion cracking
Sheet A at-rolled metal product of some maximum
thickness and minimum width arbitrarily dependent on the
type of metal. It is thinner than plate and has a width-to-
thickness ratio greater than about 50
Shield A metallic sheath surrounding one or more wires,
cables, cable assemblies, or a combination of wires and
cables that is used to prevent or reduce the transmission of
electromagnetic energy to or from the enclosed conductors.
The shield also includes an insulating jacket that may cover
the metallic sheath
Solder A nonferrous, fusible metallic alloy used to join
metallic surfaces
Solderability The property of a surface that allows it to be
wetted by a molten solder
Solder connection An electrical/mechanical connection
that employs solder for the joining of two or more metal
surfaces
Solder mask Coating material used to mask or protect
selected areas of a pattern from the action of an etchant,
solder, or plating
Solder spike A cone-shaped peak or sharp point of solder
usually formed by the premature cooling and solidication
of solder on removal of the heat source
Solidication The change in state from liquid to solid on
cooling through the melting temperature or melting range
Solid solution A solid crystalline phase containing two or
more chemical species in concentrations that may vary
between limits imposed by phase equilibrium
Solid solution strengthening A mechanism for strength-
ening the alloy by dissolved elements in solid solution
Solidus In a constitution or equilibrium diagram, the locus
of points representing the temperatures at which various
Glossary
compositions nish freezing on cooling or begin to melt on
heating. (See also liquidus.)
Solution heat treatment A heat treatment in which an alloy
is heated to a suitable temperature, held at that temperature
long enough to cause one or more constituents to enter into
solid solution, then cooled rapidly enough to hold these
constituents in solution
Specication Document stating requirements
Spot welding Welding of lapped parts in which fusion is
conned to a relatively small circular area. It is generally
resistance welding, but may also be gas tungsten-arc, gas
metal-arc, or submerged-arc welding
Stress The intensity of the internally distributed forces or
components of forces that resist a change in the volume or
shape of a material that is or has been subjected to external
forces. Stress is expressed in force per unit area and is cal-
culated on the basis of the original dimensions of the cross-
section of the specimen. Stress can be either direct (tension
or compression) or shear. Usually expressed in megapascals
(MPa)
Stress-corrosion cracking Failure of metals by cracking
under combined action of corrosion and stress, residual or
applied. In brazing, the term applies to the cracking of
stressed base metal due to the presence of a liquid ller metal
Stress relief Related to electronic assemblies: the formed
portion of a conductor that provides sufcient length to
minimize stress between terminations
Stress-relief heat treatment Uniform heating of a structure
or a portion thereof to a sufcient temperature to relieve the
major portion of the residual stresses, followed by uniform
cooling
Stress relieving Heating to a suitable temperature, holding
long enough to reduce residual stress, and then cooling slowly
enough to minimize the development of new residual stresses
Striation A fatigue fracture feature often observed in elec-
tron micrographs that indicates the position of the crack front
after each succeeding cycle of stress. The distance between
striations indicates the advance of the crack front across that
crystal during one stress cycle, and a line normal to the
striation indicates the direction of local crack propagation
Substrate The layer of metal underlying a coating,
regardless of whether the layer is base metal
Subsystem Set of interdependent elements constituted to
achieve a given objective by performing a specied function,
but which does not, on its own, satisfy the customers need
Supplier An organization that provides a product to the
customer
Glossary
Surface hardening A generic term covering several pro-
cesses applicable to a suitable ferrous alloy that produces, by
quench hardening only, a surface layer that is harder or more
wear resistant than the core
Surface mounting The electrical connection of components
to the surface of a conductive pattern that does not utilize
part holes
Technology Readiness Level (TRL) Levels 19 used to
dene the maturity of a technical concept, from basic prin-
ciple to mission proven
Temper In nonferrous alloys the hardness and strength
produced by mechanical or thermal treatment, or both, and
characterized by a certain structure, mechanical properties,
or reduction in area during cold working
Tensile strength In tensile testing, the ratio of maximum
load to original cross-sectional area. Also called ultimate
strength
Test A formal process of exercising or putting to trial a
system or item by manual or automatic means to identify
differences between specied, expected, and actual results
Thermal shunt A device with good heat-dissipation char-
acteristics used to conduct heat away from an article being
soldered
Tinning The coating of a surface with a uniform layer of
solder
Traceability Ability to trace the history, application, or
location of an entity by means of recorded identications
Transverse direction Literally across. Usually signifying
a direction or plane perpendicular to the direction of work-
ing. In rolled plate or sheet, the direction across the width is
often called long transverse, and the direction through the
thickness, short transverse
Ultimate strength The maximum stress (tensile, compres-
sive, or shear) a material can sustain without fracture,
determined by dividing maximum load by the original cross-
sectional area of the specimen. Also known as nominal
strength or maximum strength
Unaided eye Normal Snellen 20/20 vision, including eye
glasses required to correct defective vision to 20/20 equiv-
alent. Does not include microscopes, eye loupes, or any
other magnifying device
Vacuum melting Melting in a vacuum to prevent contam-
ination from air, as well as to remove gases already dis-
solved in the metal; the solidication may also be carried out
in a vacuum or at low pressure
637
Verication Conrmation by examination and provision of
objective evidence that specied requirements have been
fullled
Vickers hardness number (HV) A number related to the
applied load and the surface area of the permanent impres-
sion made by a square-based pyramidal diamond indenter
having included face angles of 136
Viscosity A measure of the resistance of a uid to ow
Waiver Written authorization to use or release a product
which does not conform to the specied requirements
Wake The side opposite to the Ram side of a spacecraft. It
faces away from the spacecrafts motion
Wave soldering A process wherein printed circuit boards
are brought in contact with the surface of continuously
owing and circulating solder
Weave exposure A surface condition of a printed-circuit-
board laminate in which the unbroken woven-glass cloth is
not uniformly covered by resin
Weldability A specic or relative measure of the ability of
a material to be welded under a given set of conditions.
Implicit in this denition is the ability of the completed
weldment to full all service designed into the part
Wetting Flow and adhesion of a liquid to a solid surface,
characterized by smooth, even edges, and a low dihedral
angle
Wicking A ow of molten solder or cleaning solution by
capillary action. Occurs when joining stranded wire; solder
is drawn within the strands, but may not be visible on outer
surface of strands. Wicking may also occur within the stress-
relief bend of a component lead
Widmansttten structure A structure characterized by a
geometrical pattern resulting from the formation of a new
phase along certain crystallographic planes of the parent
solid solution. The orientation of the lattice in the new
phase is related crystallographically to the orientation of
the lattice in the parent phase. The structure was origi-
nally observed in meteorites, but is readily produced in
many alloys, such as titanium, by appropriate heat
treatment
Wire A thin, exible, continuous length of metal, usually of
circular cross-section, and usually produced by drawing
through a die
Workmanship The physical characteristics relating to the
level of quality introduced by the manufacturing and
assembly activities
638
Wrought A metal or alloy which has been deformed plas-
tically one or more times and which exhibits little or no
evidence of cast structure
Yield point The rst stress in a material, usually less than
the maximum attainable stress, at which an increase in strain
occurs without an increase in stress. Only certain metals
Glossary
exhibit a yield point. If there is a decrease in stress after
yielding, a distinction may be made between upper and
lower yield points
Yield strength The stress at which a material exhibits a
specied deviation from proportionality of stress and strain.
An offset of 0.2 % is used for many metals
 References
Adams, D. M. (1994). Failure investigation into HST solar array drive
electronics. ESTEC Metallurgy Report No. 2038 (condential).
Adile, E., & Dunn, B. D. (1999). Effecto della geometria del raccordo
di lega saldante sul comportamento a fatica meccanica di saldature
di component elettronici a montaggio superciale. Rivista Italiana
della Saldatura,4, 401412.
Adler, I. (1988). The analysis of extraterrestrial materials. In Chemical
analysis (Vol. 81). New York: Wiley.
Akinade, K., et al. (1995). Lead-free solder pastes evaluation. Soldering
and SMT Journal,20, 5053.
Allen, J. S., & Jubb, J. E. M. (1991). Dispute resolutionThe
increasing role of the expert. Welding and Metal Fabrication
Journal, 8184.
Ambat, R., et al. (2009). Solder ux residues and electrochemical
migration failures of electronic devices. In Proceedings of Eurocorr
2009 (pp. 114), paper 8141, Nice, France, September 2009.
AMS 4942C. (1984). Titanium tubing, seamless. Society of Automo-
tive Engineers.
AMS International. (1991). Practical applications of residual stress
technology. In Conference Proceedings, Indianapolis, USA, May
1517.
Anderson, P., & Wiktorowicz, R. (1995). Ozone emissions during arc
welding. Welding and Metal Fabrication Journal, 385388.
Anderson, M. J., et al. (1995). A post-ight tribological assessment of
the Hubble Space Telescope solar array mechanism. In Proceedings
of the 6th European Space Mechanisms and Tribology Symposium
(pp. 203210). Zurich, October 46 (ESA SP-304).
Anon. (1981). Keys to improved brazing technology. Industrial
Heating Journal, 31.
Anon. (1994). EURECA, the European retrievable carrier, ESA WPP-
069.
Anon. (1995). Feasibility study of power generation system with shape
memory alloys. New Energy Development Organisation, Tokyo,
NEDO-P-9324 (in Japanese).
Anon. (2011). Easy leadIndustry and innovations. Materials World:
45.
ANS (American National Standard). (1996). Requirements for soldered
electrical and electronic assemblies, J-STD-001B.
Anthony, P. L., & Brown, O. M. (1965). Red plague corrosion.
Materials Protection,4(3), 818.
Arblaster, J. W. (1996). Thermodynamic properties of the platinum
metals. Platinum Metals Review,40, 6263.
Ashworth, M., & Dunn, B. D. (2015). An investigation into tin whisker
growth over a 32-year period (to be published).
Asma, C. O., et al. (2010). Infrared thermography measurements on
ablative thermal protection systems for interplanetary space
vehicles, QIRT Symposium, paper 007, Quebec, Canada, July 27
30.
Springer International Publishing Switzerland 2016
B.D. Dunn, Materials and Processes, Springer Praxis Books,
DOI 10.1007/978-3-319-23362-8
ASTM D3359. (2010). Standard test methods for measuring adhesion
by tape test.
ASTM E595. (2007). Standard test method for total mass loss and
CVCM from outgassing in a vacuum environment.
ASTM E835-94A. (1994). Standard method for determining residual
stresses by the hole-drilling strain gage method.
ASTM E2089. (2014). Standard practices for ground laboratory atomic
oxygen interaction evaluation of materials for space applications.
Athimoolam, M., & Moorthy, T. V. (2012). Polymer nanocomposite
materials and shape memory applicationsA review. Procedia
Engineering,38, 33993408.
Autumn, K., et al. (2000). Adhesive force of a single gecko foot-hair.
Nature,405, 681685.
Babel, H. W., et al. (1994). Understanding and controlling the
degradation of silicone on exposed spacecraft surfaces. In Proceed-
ings of the 6th International Symposium on Materials in Space
Environment, ESTEC (pp. 1926). ESA SP-368.
Bacquias, G. (1982). La formation des whiskers en bain detain acide
brillant sur un support en maillechort. Oberche-Surface,23,
194196.
Ballinger, I. A., & Sims, D. (2003). Development of an EPDM
elastomeric material for use in hydrazine propulsion systems. In
39th AIAA Propulsion Conference, Huntsville, AL, July 21. AIAA
2003-4611.
Banks, B. (2003). Atomic oxygen effects on spacecraft materials,
NASA TM-2003-212484.
Banks, D. R., et al. (1990). SnPbIn solder corrosion with water
soluble ux. In ASM 3rd International Electronic Materials and
Processing Conference (pp. 359364). San Francisco, CA, August
2023.
Banks, S. (1995). Reow soldering to gold. In Electronic packaging
and production (pp. 6972).
Barber, A. H., Lu, D., & Pugno, N. M. (2015). Extreme strength
observed in limpet teeth. Journal of the Royal Society Interface,12,
20141326.
Barbier, P., & Le Floch, C. (1985). HolographyThe NDT of
composite structures. In Proceedings of the Third European
Symposium on Spacecraft Materials in Space Environment,
Noordwijk, ESA SP-232 (pp. 161165).
Barbier, C., et al. (1992). ATOXThe ESTEC atomic oxygen
simulation facility. Preparing for the Future,2(2), 1214.
Baumgartner, R. I., & Elvin, J. D. (1995) Lifting bodyAn innovative
concept (pp. 953531). AIAA publication.
Beal, J. D., et al. (2006). Forming of titanium and titanium alloys (Vol.
14B, pp. 656669). ASM Handbook.
Bergendahl, C. G., & Dunn, B. D. (1984). Evaluation of test equipment
for the detection of contamination on electronic circuits. ESA STM-
234.
639
640
Bernard, D. (2007). A practical guide to X-ray inspection criteria and
common defect analysis. Nordson-Dage Publication.
Bernard, D., & Blish, R. C. (2005). Considerations for minimizing
radiation doses to components during X-ray inspection. In Pro-
ceedings ofthe 7th Electronic Packaging Technology Conference
(EPIC), Singapore, December 79.
Bernard, D., & Golubovic, D. (2012). 3D board level X-ray inspection
via limited angle computer tomography. In SMTA International
Conference, Orlando, FL, USA, October 1428.
Berthoud, L. (1994). Micro-impacts on Eureca solar panels. In 6th
International Symposium on Materials in a Space Environment,
Noordwijk, 239248 (ESA SP 368).
Bester, M. H. (1968). Metallurgical aspects of soldering to gold.
Internepcon (pp. 211231).
Beurs, J., et al. (1987). TEM-in situ deformation of beryllium single
crystals. Acta Metallurgical,35(9), 22772289.
Billot, M., et al. (1982). Ensuring solderability in electronic compo-
nents for space applications. Tin and Its Uses,131, 13.
Black, J. R. (1969). Electromigration failure modes. Proceedings of the
IEEE,57, 15871594.
Blackman, B. R. K., et al. (2012). The fracture behaviour of adhesively-
bonded composite joints: Effects of rate of test and mode of loading.
International Journal of Solids and Structures,49(13), 14341452.
Blawert, C., & Srinivasan, P. B. (2010). Plasma electrolytic oxidation
treatment of magnesium alloys. In H. Dong (Ed.), Surface Engi-
neering of Light Alloys. Cambridge, UK: Woodhead Publishing Ltd.
Blech, I. A., et al. (1975). Whisker growth in aluminium thin lms.
Journal of Crystal Growth,32, 161169.
Borom, M. C., & Pask, J. A. (1966). Role of adherence oxides in the
development of chemical bonding in glass seals. Journal of the
American Ceramic Society,49, 1.
Bosma, S. J., et al. (1978). Evaluation of grounding systems. ESA STM
204, ESA STM 212 and ESA STM 213.
Bosma, J. (2009). Powerhouse for Quality. ESA Bulletin, May 311.
Bouilly, J.-M., et al. (2006). Ablative TPS for entry in Mars
atmosphere. In 4th International Planetary Probe Workshop,
Pasadena, CA, USA, June 2730.
Bouilly, J.-M., et al. (2013). Aerofast: Development of innovative
thermal protections. In 7th European Workshop on TPS and Hot
Structures, Noordwijk, The Netherlands, April 810.
Bouseld, B. (1988). A systematic approach to sample preparation.
Metals and Materials,4(12), 758761.
Bouseld, B. (1992). Surface preparation and microscopy of materials.
Chichester, UK: Wiley.
Bouseld, B. (1997). The metallography of deformation by material
removal. Praktische Metallographie,34(1), 222.
Boustedt, K., & Hedemalm, P. (1995). The Nordic ip-chip project. In
Proceedings of the 1995 International Flip Chip, Ball grid, TAB and
advanced packaging Symposium (pp. 294299), San Jose, Ca., USA.
Boving, H. J., & Hintermann, H. E. (1987). Properties and perfor-
mances of chemical-vapour-deposited TiC-coated ball-bearing
components. Thin Solid Films,153, 253266.
Bradley, E. (2007). Lead-free electronics: iNEMI projects lead to
successful manufacturing, pub. Hoboken, USA: Wiley.
Brandes, E. A. (1992). Smithells metals reference book (7th ed.).
Oxford, UK: Butterworth-Heinemann.
Brewer, D. H. (1970). Solders for thick gold plate. Welding Research
Sup (pp. 465470).
Briscoe, H. M. (1982). Proceedings of the international symposium on
spacecraft materials in space environment (pp. 2734). ESA SP-178.
Brizuela, M., et al. (2009). Tribolab: An experiment on space tribology,
in orbit data at the IIS. In Proceedings of the 13th European Space
Mechanisms and Tribology Symposium (ESMATS 2009). Vienna,
Austria, September 2325.
References
Brooker, M. J., et al. (2001). Applying FSW to the Ariane 5 main motor
thrust frame. ESA SP-468 (pp. 507511).
Brown, B. F. (1972). Stress-corrosion cracking in high-strength steels,
and in titanium and aluminium alloys. U.S: Naval Research
Laboratory.
Brown, B. F. (1977). Stress corrosion cracking control measures, NBS
Monograph 156, Ch. 5 (pp. 3541).
Bruno, G., Dunn, B. D., et al. (2000). Neutron diffraction measure-
ments for the determination of residual stress in Ti6Al4V welded
plates. Material Science Forum,347349, 674690.
Bruno, G., & Dunn, B. D. (2004). Surface and bulk residual stress in
Ti6Al4V welded aerospace tanks. Transactions of the ASME,126,
284292.
Brusciotti, F., et al. (2014). MAGNOLYA: Advanced environmentally
friendly chemical surface treatments for cast magnesium helicopter
transmission alloys preservation. In Eurocorr 2014, Poster, Pisa,
Italy, September 812.
Bruzzone, P. (1987). Electrical properties of low MP alloys at 4.2 K.
Cryogenics,27, 433436.
Buckley, D. H. (1981). Surface effects in adhesion, friction, wear and
lubrication. Amsterdam: Elsevier.
Buis, A., & Schijve, J. (1992). SCC of AlLi 2090-T83 in articial
seawater. Corrosion,48(11), 898909.
Bullettti, A., Dunn, B. D., et al. (2012). Investigation of resistivity
variation of PCB laminates due to aging. IEEE Transactions on
Components, Packaging, and Manufacturing Technology,2, 2001
2012.
Bulwith, R. A. (1986). The rough of grainy solder phenomenon.
Printed Circuit Assembly,1(1), 1423.
Burnette, T., & Koschmeider, T. (2003). Solder joint failure analysis
Dye penetrant technique. Advanced Packaging, 12, 2527.
Burrell, M. C., & Keane, J. J. (1988). Characterization of copper/
enamel interfacial reactions during aging. Surface and Interface
Analysis,11, 487496.
Bushmire, D. W. (1977). Resistance increases in gold aluminium
interconnects with time and temperature. IEEE Transactions on
Parts, Hybrids and Packaging,PHP-13(2), 152156.
Bussu, G. (2009). Stress-corrosion cracking (Ch 12, materials and
safety). In G. E. Musgrave (Ed.), Safety Design for Space Systems.
Oxford, UK: Elsevier Ltd.
Bussu, G., & Dunn, B. D. (2002). ESA approach to the prevention of
stress corrosion cracking in spacecraft hardware. In Joint ESA-
NASA Space Flight Safety Conference. ESA-Estec, Noordwijk, The
Netherlands.
Buttery, M., et al. (2011). Modern self-lubricating composites for space
applications: PGM-HT and Sintimid 15 M. In 14th European Space
Mechanisms and Tribology Symposium (ESMATS 2011). Con-
stance, Germany, September 2830.
Cable, N. (1988). Use of pyrotechnics on spacecraft. ESA Bulletin,54,
6671.
Cahn, R. W., & Haasen, P. (1996). Physical metallurgy. Netherlands:
Elsevier Science B.V. (see Chapter 29).
Calle, L. M. (2014). NASAs corrosion lab at KSC: Anticipating,
managing, and preventing corrosion. In International Workshop on
Environmental and Alternate Energy. Kennedy Space Centre,
October 2124.
Calow, C. A. (1974). Are metal matrix composites viable? Practical
metallic composites 3. Spring Meeting, Institution of Metallurgists.
Campbell, J. E. (1980). Alloys for structural applications at subzero
temperatures. Metals Handbook (9th ed., Vol. 3, pp. 721772). ASM.
Capineri, L., Dunn, B. D., et al. (2003). Partial discharge testing of
solder llets on PCBs in a partial vacuum: new experimental results.
IEEE Transactions on Electronics Packaging Manufacturing,26,
294305.
References
Case, R. K., et al. (1984). BerylliumThe lightweight contender.
Machine Design, 56, 5861.
Ceccanti, F., et al. (2010). 3-D printing technology for a Moon outpost
exploiting lunar soil. 61st International Astronautical Congress
(pp. 19). Prague, CZ (IAC-10-D3.3.5).
Chafn, R. J. (1981). Migration of silver from silver-loaded polyimide
adhesive chip bonds. In IEEE Transactions on C.H.M. Technology
(Vol. 4, No. 2, pp. 214216). June.
Chaillot, A., et al. (2014). ENEPIG nish: An alternative solution for
printed circuit boards. In Electronic Materials and Processes
Workshop (EMPS-5). Estec, Noordwijk, The Netherlands, May
2024.
Charrier, I., Dunn, B. D., et al. (2003). Moisture pick-up and minimum
drying conditions of epoxy and polyimide-glass printed circuit
boards. In Brasage/Soldering, International Conference. Brest,
France, October 810.
Chelladurai, T., et al. (1995). Acoustic emission response of 18 % Ni
maraging steel weldment with inserted cracks of varying depth to
thickness ratio. Materials Evaluation,53(6), 742746.
Chen, Y., et al. (2014). A comprehensive investigation of the galvanic
corrosion induced Ag-Al bond degradation in microelectronic
packaging using argon ion milling, SEM, dual beam FIB-SEM,
STEM-EDS and TOF-SIMS. 40th ISTFA, Houston, TX, USA,
November 913.
Chenard, S. (1990). Space: Dirty and dangerous? Space Markets,5,
253256.
Choi, S. J., et al. (1996). A Nitinol-based solar array deployment
mechanism. In 30th Aerospace Mechanisms Symposium, Hampton,
Va. (103118). NASA Conf. Pub. No. 3328. May.
Chollet, L., et al. (1985). Journal of Mater Energy System,6(4), 293299.
Chudnovsky, B. H. (2002). Degradation of power contacts in industrial
atmospheres: silver corrosion and whiskers. In Proceedings of the
48th IEEE Holm Conference on Electrical Contacts, Orlando, USA.
Chuvilin, A., et al. (2014). Using TEM to simulate and understand the
formation of defects in graphine. In Microscopy and Analysis,
November/December 710.
Cleave, J. F., & Morton, J. R. (1989). The effects of impurity levels of
lithium on aluminium casting alloys. The Foundryman,82, 549592.
Clemens, H., & Schretter, P. (1996). Microstructure and properties of
Gamma-TiAl based alloys. Practical Metallogarphy Journal,33(1),
1735.
Clements, H., et al. (1996). Intermetallic gamma titanium aluminides
Potential candidates for spacecraft structures. ESA Conference on
Spacecraft Structures, Materials and Mechanical Testing, Noord-
wijk, March 2729.
Coelho, A. (2009). Comeback for cork. Materials World, 17, 24.
Cohen, A. (1995). Copper and copper alloy wires, cables and
conductors. ASTM Standardization News, 2327.
Cole, L. S., et al. (2004). Some recent trends in corrosion science and
their application to conservation. In Proceedings of Metal (National
Museum of Australia, Canberra), October 216.
Cole, P. T., & Thomas, N. (1989). Ariane IV Spelda-Flight proving by
acoustic emission. In Proceedings of AECM-3 (pp. 144150). Paris,
July 1721.
Colla, V., et al. (2014). Monitoring concepts for a 3-D printer applied to
build a human outpost on the Moon. In UKSim-AMSS 16th
International Conference on Computer Modelling and Simulation
(pp. 205210).
Comiskey, J. M. (1979). Materials that make wire for spacecraft use. In
Aerospace Corporation,Proceedings of Symposium on Wire Appli-
cations in Spacecraft, December.
Cooke, R. W. (2010). Red plague control plan (RPCP), the control of
cuprous/cupric oxide corrosion. http://ntrs.nasa.gov/archive/nasa/-
casi.ntrs.nasa.gov/20100009723.pdf.
641
Coope, T. S., et al. (2014). Repeated self-healing of microvascular
carbon bre reinforced polymer composites. Smart Materials and
Structures, 23(11), 115002.
Corocher, G., et al. (2009). Evaluation of different test techniques to
assess the effect of environmental exposure on the assembly of
MCGA packages for printed circuit boards. 11th ISMSE Interna-
tional Symposium, Aix-en-Provence, France, September 1518.
Cornec, L., & Clariou, J.-P. (1994). Carbonmagnesium MMC for
light-weight aerospace structures. In Proceedings of the Interna-
tional Symposium on Advanced Materials, ESTEC, Noordwijk
(pp. 563567). ESA WPP-070.
Craig, B. (1994). Lead-free solder: An update. Electronic Production,
2125.
Craven, E. (1993). SolderabilityThe effect on quality manufacturing.
Electronic Production, 3335.
Damskey, R. G. (1988). A satellite structureManagers look at
beryllium. The Beryllium Users Symposium,Europe.
Danford, M. D. (1994). The corrosion protection of several aluminium
alloys by chromic acid and sulphuric acid anodizing. NASA
Technical Paper 3490.
de Salazar, Gomez, et al. (1988). Diffusion bonding of aluminium
alloys using silver coatings. Practice Metals,25, 532542.
de Rooij, A. (1985). The degradation of metal surfaces by atomic
oxygen. In Proceedings of 3rd European Symposium on Spacecraft
Materials in Space Environment (p. 99). ESA SP-232, October.
de Rooij, A. (1989). Bimetallic compatible couples. ESA Jnl.,13, 199
209.
de Rooij, A. (1989b). An approach to long term prediction of the
atomic oxygen effect on materials (to be published).
de Rooij, A. (1995a). Atomic oxygen effects on HST-SA1 metals. In
Hubble Space Telescope Solar Array Workshop (pp. 341353),
Noordwijk, May.
de Rooij, A. (1995b). Space exposure of HST-SA1 tagboards. Hubble
Space Telescope Solar Array Workshop, ESTEC (pp. 449462).
ESA WPP-77.
de Rooij, A., & Collins, D. C. (1995). SLAM investigations on HST-
SA1 solar cell welds. In Hubble Space Telescope Workshop
Results from Post-ight Investigations, ESTEC, Noordwijk, May
3031. ESA WPP-77.
de Rooij, A. (2010). Exposure of silver to atomic oxygen. In
EUROCORR 2010, Moscow, Russia, September 1317.
De Rooij, N., et al. (2009). MEMS for Space. IEEE Transducers
Conference (pp. 1724). Denver, CO, USA, June 2125.
Dauphin, J. (1984). Selection and control of materials for space
applications. ESA Journal,8(1), 53.
Dauphin, J. (1986). Space materials in Europe. In Proceedings of the
SAMPE Conference (pp. 276290), Las Vegas.
Dauphin, J. (1987). In J. de Bossu (ed.), Atomic oxygenA low orbit
plague. Looking ahead for materials and processes (pp. 345367).
Amsterdam: Elsevier Science Publishers B.V.
Dauphin, J. (1992). Touchstone for success: The materials and
processes laboratory, ESA BR-52.
Dauphin, J. (1993). Radiation effects in materials. Radiation Physics
and Chemistry,43, 4756.
Dauphin, J., Dunn, B. D., Judd, M. D., & Levadou, F. (1991). Materials
for space application. Metals and Materials, 422430.
David, L. (1995). Radarsat built to resist debris bumps. Space News.
David, L. (1996). Incredible shrinking spacecraft. Aerospace America,
2024.
Davy, J. G., & Skold, R. (1985). Computer-aided solderability testing
for receiving inspection. Circuit World, 3441.
DeBold, T. A., & Kosa, T. (2004). Pasivating stainless stell parts.
NASA Technical Briefs, 5152.
Degisher, H. P., & Simancik, F. (1995). Recyclable foamed aluminium
as an alternative to composites. In H. Warlimont (Ed.),
642
Environmental Aspects in Materials Research (pp. 137140). DGM
Informations gesellschaft.
Delahais, M. L. L. (1985). Materials and processes in scientic
spacecraft. In Proceedings of the 3rd European Symposium in
Spacecraft Materials in Space Environment. ESA SP-232, October
57.
Deloo, P., et al. (1995). HST Solar Array 1 Dynamic Performance. HST
Solar Array Workshop (pp. 89100). ESTEC, ESA WPP-77, May
3031.
De Marco, M., et al. (2015a). Introduzione ai modi di frattura ed
allanalisi fracttograca. Parte 1. Rev. It. Della Saldatura, 1, 95106.
De Marco, M., et al. (2015b). Introduzione ai modi di frattura ed
allanalisi fracttograca. Parte 2. Rev. It. Della Saldatura, 2, 229242.
Dittes, M., et al. (2003). The effect of temperature cycling on tin
whisker formation. In Proceedings of the Electronics Packaging
Technology Conference (pp. 183188).
Didziulis, S. V. (1994). Detergent cleaning of bearings for spacecraft
applications. In Non-Ozone Depleting Chemical Cleaning and
Lubrication of Space System Mechanical Components for Multi-
Year Operations, EPA Workshop, Denver, Co, September 2627.
Dohle, G. R., et al. (1996). A new bonding technique for microwave
devices. IEEE Transactions on Components Hybrids and Manu-
facturing Technology,19(1), 5763.
Dorsey, J. T., et al. (2012). An efcient and versatile means for
assembling and manufacturing systems in space. In AIAA Space
Conference and Exposition, Pasadena, CA, USA, September 1113.
Downs, K. S., & Hamstad, M. A. (1995). High sensor density A.E.
monitoring of graphite/epoxy pressure vessels. In The American
Society for Nondestructive Testing, Proceedings of AECM-4
(pp. 1322). Sundsvall, July 1014.
Doyle, D. P., et al. (1963). A rapid method for determining the
corrosivity of the atmosphere at any location. Nature,200(4912),
11671168.
Doyle, E., & Hubbard, P. (2010). Innovation in nitriding. International
Heat Treatment and Surface Engineering,4(1), 3339.
Dricot, F., et al. (1994). Programme for the evaluation of materials used
in low Earth orbit and manned spacecrafts. In 6th International
Symposium on Materials in a Space Environment (pp. 293297).
Noordwijk, (ESA SP-368).
Drolshagen, G. (1994). Eureca meteoroid/debris analysis. In Eureca
Technical Report (pp. 507512). ESA WPP-069.
Drolshagen, G., et al. (1996). Optical survey of micrometeoroid and
space debris impact features on Eureca. Planetary Space Science,44
(4), 317340.
Drogoul, S. (2005). From the failure to the successThe return to
ight of the Ariane 5 ECA launcher. In 56th International
Astronautical Congress, Session D2.2, IAC-05-D2.2.08, Fukuoka,
Japan.
Dunkerton, S., Dunn, B. D., & Fernie, W. B. (2001). Welding and
joining in a space environment. Revista Italiana della Saldatura,
Anno LIII,5, 615627.
Dunn, B. D. (1973). European team visit TRW Systems to learn
techniques for the future. TRW Sentinel (p. 6), November 9.
Dunn, B. D. (1975). The properties of near-eutectic tin/lead solder
alloys tested between +70 and 60 C and the use of such alloys in
spacecraft electronics, ESA TM-162, September.
Dunn, B. D. (1977). Metallurgical evaluation of tinlead nishes on
PCBs. Transactions of the Institute of Metal Finishing,55, 8692.
Dunn, B. D. (1978). Examination of new types of solder-splices. ESA
TM-210, September.
Dunn, B. D. (1979). The resistance of space quality solder joints to
thermal fatigue. Circuit World part 1 5(4), 1117; part 2 6(1), 1627.
Dunn, B. D. (1980). The Fusing of tinlead plating on high quality
printed circuit boards. Transactions of the Institute of Metal
Finishing,58, 2628.
References
Dunn, B. D. (1982a). Evaluation of various solar cell-to-interconnector
welds by means of scanning laser acoustic microscopy and
metallography. ESA STM-225, August.
Dunn, B. D. (1982b). Evaluation of various solar cell to interconnector
welds by SLAM and metallography. ESA STM-225.
Dunn, B. D. (1984). The corrosion properties of Spacelab structural
alloy aluminium 2219-T851, ESA STR-212.
Dunn, B. D. (1987). A laboratory study of tin whisker growth. ESA
STR-223.
Dunn, B. D. (1988). Mechanical and electrical characteristics of tin
whiskers. ESA Journal,11 and 12, 117.
Dunn, B. D. (2001). Innovative Materials and Processes. Space
Technology Innovation Workshop (pp. 21134). Copenhagen.
ESA WPP-197, September 67.
Dunn, B. D. (1993). Metallurgical evaluation of steam aged LCCC
devices following solderability testing. Soldering and SMT,15,
2632.
Dunn, B. D. (2008). Workmanship standards and their application on
ESA projects. Soldering and SMT Journal,20(4), 3744.
Dunn, B. D. (2008). ESA-approved skills training schoolsElectronic
assembly techniques, ESA-STR-258, November.
Dunn, B. D. (2012). Guidelines for a creating a lead-free control plan.
ESA STM-281, October.
Dunn, B. D., & Bergendahl, C. G. (1987). An evaluation of the solvent
extraction method for the detection of ionic contamination on
substrates supporting large surface mounted devices. Brazing and
Soldering Journal, 2027, 32.
Dunn, B. D., & Chandler, C. (1980). The corrosive effect of soldering
uxes and handling on some electronic materials. Welding
Journal,59, 289307.
Dunn, B. D., & Collins, D. S. (1978). Materials investigation for a
failed spacecraft antenna. ESA Journal,2, 223235.
Dunn, B. D., & Desplat, P. (1994). Evaluation of conformal coatings
for future spacecraft applications. ESA SP-1173, August.
Dunn, B. D., & Nicholas, D. (1998). Elemental distribution of parts and
materials making up an electronic box. Circuit World,24(2), 2123.
Dunn, B. D., & Mozdzen, G. (2014). Tin oxide coverage on tin whisker
surfaces, measurements and implications for electronic circuits.
Soldering and SMT Jnl.,26(3), 139146.
Dunn, B. D., & Papenburg, H. (1986). Material investigation into the
deactivation of catalyst particles for hydrazine decomposition. In
Proceedings of the Fifteenth Inernational Symposium on Space
Technology and Science, Tokyo.
Dunn, B. D., & Stanyon, R. (1997). An Inspection of Spacelab
Hardware. ESA STR-241, September.
Dunn, B. D., & Steinz, J. A. (1974). Contamination and heating effects
of a hydrazine thruster intended for use on-board the scientic
geostationary satellite. ESRO (ESA) TM-152, April.
Dunn, B. D., et al. (1983). Evaluation of the temperatures attained by
electronic components during various manual soldering operations.
Brazing and Soldering Journal,5, 4553.
Dunn, B. D., et al. (1984). Corrosion of silver plated copper conductors.
ESA Journal,8, 307335.
Durin, C., et al. (1993). System results from FRECOPA. In 3rd Post
Retrieval Conference (pp. 11531165). Williamsburg, November
812, NASA Pub. 3275 part 3.
Dzhanibekov, V. A., et al. (1991). Welding equipment for space
applications. In Proceedings of the Welding in Space and the
Construction of Space Vehicles by Welding, sponsored by
American Welding Society, USA and E O Paton Welding Institute,
Ukraine, September 1991.
Ecord. G. M. (1972). Apollo Experience reportPressure vessels.
NASA TN-D-6975, September.
ECSS-Q-ST-70. (2014). Materials, mechanical parts and process
Space product assurance requirements.
References
Ehmann, D. (1994). Space structures of high precision. Dornier Post,1,
1517.
Elking, M. J., & Hughes, H. E. (1969). Prevention of stress corrosion
failure in semiconductor leads. Bell Telephone Report in: Physics of
Failure in Electronics, 5, 447495.
Ellis, B. N. (1996). The correlation between short- and long-term SIR
testing. Circuit World,22(2), 4950.
Ellis, M. B. D. (1996). Fusion welding of aluminium-lithium alloys.
Welding and Metal Fabrication Journal. 5560.
Ellis, M. B. D., & Gittos, M. F. (1995). Tungsten inert gas welding
of titanium and its alloys. Welding and Metal Fabrication Journal,
912.
Ely, J. K. (1942). Method of making electrical contact members. U.S.
patent 2295338 of 8 September.
Engel, L., & Klingele, H. (1981). An Atlas of Metal Damage. London:
Carl Hanser Verlag and Wolfe Publishing Ltd.
Engelmaier, W. (1995). Fatigue life of leadless chip carrier solder joints
during power cycling. In Solder Joint Reliability Workshop Notes,
Engelmaier Associates Inc., Mendham, NJ, USA (see also IEEE
Transactions Components, Hybrids and Manufacturing Technol-
ogy6(3), 232237, 1983).
Engemann, J. (1996). Chancen und moglichkeiten der plasmatechnolo-
gie: eine bestandsaufnahme. Vakuum in Forschung und Praxis,1,
2530.
EPAUS Environmental Protection Agency. (1994). Replacing solvent
cleaning with aqueous cleaning, EPA-600/R-94-131.
Erdman, N., Campbell, R., & Asahina, S. (2006). Argon beam cross
sectioning. Advance Materials and Processes, 3335.
ECSS-E-ST-32-08. (2013). Space engineering: Materials, draft,
December.
ECSS-Q-ST selected ECSS standards related to materials and processes
can be seen listed in Appendix 9.
ESA PSS-01-401. (1994). EAS fracture control requirements.
ESA PSS-01-701. (1994). Data for selection of space materials. Issue 1,
rev. 3.
ESA PSS-01-703 A method for black anodising aluminium with
inorganic dyes.
ESA PSS-01-708. The manual soldering of high-reliability electronics.
ESA PSS-01-721. Application and testing of Cuvertin 306, a black
paint for space use.
ESA PSS-01-730. (1981). The wire-wrapping of high-reliability
electrical connections.
ESA PSS-01-736. (1981). Material selection for controlling stress-
corrosion cracking of ESA spacecraft and associated equipment, May.
ESA PSS-01-737. (1982). Test method for determination of the stress-
corrosion susceptibility of spacecraft structural alloys.
ESA PSS-01-738 (Issue 1). (1991). High-reliability soldering for
surface-mount and mixed technology printed.
ESA PSS-01-746. (1993). General requirements for threaded fasteners.
E. S. K. (1995). Engineered ceramics. Paper from Elektroschmelzwerk
Kempten GmbH, Munich, Germany.
Evans, R. (1995). Fume extraction. Electronic Production, 1416.
Eyre, T. S. (1991). Friction and wear control in industry. Metals and
Materials, 143148.
Fang, D., & Berkovits, A. (1995). Fatigue design model based on
damage mechanisms revealed by AE. Trans ASME,117, 200208.
Favre, J.-P., & Raud, C. (1995). Acoustic emission to monitor crack
accumulation in crossply CFRP under mechanical loading or thermal
cycling. In The American Society for Nondestructive Testing,
Proceedings of AECM-4 (pp. 3342). Sundsvall, July 1014.
Feest, A. (1994). Interfacial phenomena in metal-matrix composites.
Composites Journal,25(2), 7586.
Fellas, C. (1982). Anti-static coat for solar arrays. In Proceedings of the
3rd European Symposium on Photovoltaic Generators in Space,
Bath, UK (p. 305).
643
Felton, L. E., et al. (1991). Cu-Sn precipitates in solder joints. Scripta
Metallurgica et Materialia,25, 23292333.
Feng, W., et al. (2013). Electro-migration model parameters sensitivity
analysis based on Monte Carlo method. Chemical Engineering
Transactions,33, 10931098.
Fernie, J. A., & Sturgeon, A. (1992), Joining ceramic materials. Metals
and Materials, 212217
Fink, M., & Dunn, B. D., et al. (2006). Characterisation of Teon FEP
(Hubble ST and LDEF) following long term exposure to LEO. In 11th
International Symposium on ISMSE, Collioure, France, June 1923.
Fink, M., Dunn, B. D., et al. (2008). Measurement of mechanical
properties of electronic materials down to 4.2K. Cryogenics
Journal,48, 497510.
Fink, M., Semerad, E., & Dunn, B. D., et al. (2009). Electrical
resistivity of solders at low temperatures. In 11th International
Symposium on ISMSE, Aix-en-Provence, France, September 1516.
Fink, M., et al. (2009). In Proceedings of the 13th European Space
Mechanisms and Tribology Symposium, Vienna, September 2325.
Fletcher, P. (1978). Problems encountered with the use of semi-rigid
co-axial cable in spacecraft electronics, ESA-STM-209, September.
Flin, E. D. (2005). Soldering in space holds surprises, Aerospace
America, 2425.
Flom, Y. (2005). Brazing in spaceEnabling the next frontier. In
TheAWS 35th International Brazing and Soldering Symposium,
Dallas, TX, USA, April 26.
Flom, Y. (2006). Electron beam brazing of titanium for construction in
space, NASA GSFC. In Proceedings of the 3rd International Brazing
and Soldering Conference, San Antonio, TX, USA, April 2426.
Flury, W. (Eds.). (1993). First European conference on space debris.
ESA SD-01.
Foster, C. L., et al. (1995). The solar array-induced disturbance of the
HST pointing system. NASA Technical Paper3556, May.
Fox, V. C., et al. (1999). The structure of tribologically improved
MoS2Metal composite coatings and their industrial applications.
Surface and Coatings Technology,116119, 492497.
Fox, R. M., et al. (2010). Chemistry and technology of lubricants, Ch.
13, Liquid lubricants for space applications. Heidelberg: Springer
Publishing.
Frankel, H. E. (1969). Effect of vacuum on materials, ESRO TN-77.
Frear, D. R., & Vianco, P. T. (1994). Intermetallic growth rate and
mechanical behaviour of low and high melting temperature solder
alloys. Metallurgical and Materials Transactions A,25, 15091523.
Frear, D. R., et al. (1993). The mechanics of solder alloy interconnects.
New York: Van Nostrand Reinhold.
Gabe, D. R. (1978). Principles of metal surface treatment and
protection. Oxford: Pergamon Press.
Galleo, K. (1995). Assembly with conductive adhesives. IPC Technical
Review, 2028.
Galvao, I., et al. (2011). Formation and distribution of brittle structures
in FSW of aluminium and copper: Inuence of process parameters.
Science and Technology of Welding and Joining,16(8), 681689.
Gamwell, W. R. (1995). NASA-wide fastener technical interchange
meeting. NASA Conference Publication 3308, July.
Garner, H. R., et al. (1976). Whisker growth from decorative zinc deposits.
In Proceedings of the 9th World Congress on Metal Finishing.
Garrison, A., et al. (1995). How much rework is too much. Soldering
and SMT Journal,19, 4851.
George, E., & Pecht, M. (2013). Tin whisker analysis of an automobile
engine control unit, microelectronics reliability. http://dx.doi.org/
10.1016/j.microrel.2013.07.134.
Gerlach, L. (1995). Co-ordinator. In Proceedings of the Hubble Space
Telescope Solar Array Workshop, ESA WPP-077.
Ghate, P. B. (1982). Electromigration-induced failures. In IEEE 20th
Annual Proceedings, Reliability Physics (pp. 292299). San Diego,
MarchApril.
644
Ghaffarian, R. (2013). 3D X-ray CT for BGA/CGA workmanship defect
detection. NASA JPL Publication 12-21 12/12.
Ghidini, T., et al. (2008). Stress-corrosion cracking assessment in
spacecraft propulsion systems, Technology Evaluation for Envi-
ronmental Risk Mitigation (TEERM) Workshop, San Diego, CA,
USA, November 1820.
Gilbrech, J., et al. (2005). Cloud-aerosol LIDAR and IR Pathnder
satellite observation (CALIPSO) spacecraftIndependent technical
assessment, NASA/TM-2005-213231/Version 2.0, March.
Gobbi, S. L., & Zhang, L. (1994). Laser welding for lightweight
structures. In International Symposium on Advanced Materials for
Lightweight Structures, ESA-ESTEC (pp. 107113). 2224 March.
ESA WPP-070.
Goddard, D. M. (1984). Report on graphite/magnesium castings. Metal
Progress, 4952.
Golberg, D., et al. (1995). High-temperature shape memory effect.
Materials Letters,22, 241248.
Goldberg, I. B., et al. (1990). Magnetic susceptibility of Inconel alloys
at cryogenic temperature. Advances in Cryogenic Engineering
(Materials) Journal,36, 755761.
Goldmann, L. S., et al. (1977). Leadindium for controlled-collapse
chip joining. In Electronic Components Conference (pp. 2529).
Goldstein, J. I., et al. (1981). Scanning electron microscopy and X-ray
microanalysis. New York: Plenum Press.
Gonnet, L. (1995). Towards an ergonomic design of welding stations.
Rivista Italiana della Saldatura,2, 177183.
Goodhew, P. J. (1988). Modular MSc degrees: A way of keeping up
with materials and techniques. Metals and Materials,4(2), 112.
Grady, M. M., & Wright, I. P. (1996). Space invaders. European
Microscopy and Analysis Journal 42, June 1314.
Grande, D. H., et al. (1988). Fibrematrix bond strength studies.
Journal of Materials Science,23, 311328.
Grant, L. A. (1983) Beryllium sheet materials to satellite structure. In
24th Annual AIAA SDM Conference, Lake Tahoe, May.
Greenberg, M. A., et al. (2013). Inhomogenieties of the interface
produced by explosive welding. Physics of Metals and Metallog-
raphy,113(2), 176189.
Gregoratti, L., et al. (2013). Large gecko mimetic tapes as new joining
technology. In Proceedings of the 12th International Symposium on
Materials in Space Environment, Noordwijk, The Netherlands
(ESA SP-705).
Grey, M. W. (1989). Inner layer or post cracking on multilayer printed
circuit boards. Circuit World,15(2), 2229.
Grilli, R., Dunn, B. D., et al. (2010). Locallised corrosion of
Aluminium 2219 alloy exposed to a 3.5 %NaCl solution. Corrosion
Science,52, 28552866.
Grilli, R., Dunn, B. D., et al. (2010). Localized corrosion on 2219
aluminium alloy coated with a titanium based conversion coating.
Surface and Interface Analysis Journal,42(67), 610615.
Grinberg, N. M. (1982). The effect of a vacuum on fatigue crack
growth. International Journal of Fatigue, April 8395.
Griner, C. S. (1968). Redeposition of vacuum outgassed products.
NASA TM X-53 801, MSFC, November.
Gross, J. H. (2011). Mass spectroscopy. Heidelberg: Springer.
Gruenbaum, P., & Dursch, H. (1993). Space environment effects on
solar cells: LDEF and other ight test. In 3rd Post Retrieval
Conference (pp. 11671177). Williamsburg, November 812,
NASA Publication 3275 part 3.
Gruhl, W. (1984). Stress corrosion cracking of high strength aluminium
alloys. Z. Metallkde,75, 819826.
Gurklis, J. A. (1972). Non-traditional machining of beryllium. Metals
and Ceramics Information Center, Battelle, MCIC, January 3.
Guyenne, T.-D. (1994). Compiler. In 6th International Symposium on
Materials in a Space Environment, ESA SP-368.
References
Hagge, J. K., & Davis, G. J. (1984). Ageing, solder thickness, and
solder alloy effects on circuit board solderability. IPC Technique
Review,25(9), 1323.
Hagge, J. K., & Davis, G. J. (1985). Ageing, solder thickness and solder
effects on circuit board solderability. Circuit World,11(3), 815.
Haines, J. E., et al. (1995). HST solar array drive electronics: In-ight
history and servicing mission activities. In HST Solar Array
Workshop, ESTEC, (ESA WPP-77) (pp. 111118).
Hall, W., & Williams, P. (2007). Processing waste pc boards for
material recovery. Circuit World,33, 4350.
Halpin, J. C. (1985). Proceedings of the 31st Annual Technical Meeting
of the Meeting of the Institute of Environmental Sciences (pp. 206
218). Las Vegas, April 30May 2.
Hamacher, H., et al. (1995). Surface erosion in low Earth orbit. In
ATOXSymposium on Scientic Results on the German Spacelab
Mission. D-2, Norderney: WPF publication.
Hammesfahr, A. E., et al. (1995). An X-ray mirror assembly based on
carbon bre technology. Space Technology,14(6), 391397.
Hamstad, M. A. (1986). A review: Acoustic emission, a tool for
composite-material studies. Experimental Mechanics, March 713.
Hannech, E. B., & Hall, C. R. (1992). Diffusion controlled reactions in
gold/solder system. Materials Science and Technology,8, 817824.
Hansen, M. (1958). Constitution of binary alloys. New York: McGraw-
Hill.
Haq Qureshi, A., & Dayton, J. A. (1995). Insulation requirements of
high-voltage power systems in future spacecraft. NASA Technical
Paper No.3520, June.
Harris, P. G., et al. (1992). Stripping and replacement of damaged
solderable coatings. Soldering and SMT Journal,12, 2126.
Harris, P. G. (1995). Conductive adhesives a critical review of progress
to date. Soldering and SMT Journal,20, 1921.
Haschiguchi, D., et al. (1992). Aluminiumberyllium alloys for
aerospace applications. In Proceedings of the International Sympo-
sium Advanced Materials for Lightweight Structures (pp. 165169).
ESTEC, March 2527, ESA SP-336.
Hashemzadeh, M. N. (2005). Study of tin whisker growth and their
mechanical and electrical properties (PhD thesis, Linkopings
Universitty, Stockholm, 2005), September.
Hashimoto, K., et al. (1991). Flip chip interconnection technology for
GLSI packages operated in liquid nitrogen. IEICE Transactions,
E74, 23622368.
Hassan, J., et al. (2014). Recent advancements in focused ion beam
applications: A review. International Journal of Recent Scientic
Research,5, 123127.
Hauer, H. (1977). Properties of soft soldered joints on gold plated thin
lm conductors. In International Conference on Thin Film Tech-
nology (Vol. 60, pp. 185190). NTG-Facher.
Haupt, C. W. (1996). Residual stress measurement of titanium weld
certication rings. In 32nd AIAA Joint Propulsion Conference
(p. 10). Lake Buena Vista, FL, paper AIAAA-96-2750, July 13.
Hawkins, T.W., et al. (2010). Reduced toxicity, high performance
monopropellant at the US air force laboratory. In 4th International
Association for the Space Safety Conference, Huntsville, AL, May.
Healy, T. J., et al. (1995). Key technologies being developed for RLV
(pp. 19953608). AIAA paper.
Hegin, B. V. (n.d). Electroplating and black anodisation of aluminium.
Galvanische Industrie, Heerde, Holland (trade leaet).
Hehemann, P. F. (1985). Stress corrosion cracking of stainless steels.
Metallurgy Transactions A,16A, 19091923.
Heiken, G. H., et al. (1974). Lunar deposits of possible pyroclastic
origin. Geochimica Cosmochimica Acta,38, 17031718.
Hemptenmacher, J., et al. (1994). Characteristics of SiCFibre/Ti6A14
VMatrix composites by macro-indentation tests. Practical
Metals,31(3), 110119.
References
Hendrich, CH., et al. (2014). Inuence of water content in an ADN-
based liquid monopropellant on performance characteristics. In
Space propulsion, May 1922. Germany: Cologne.
Hepp, A. F., et al. (1999). Metal alloy propellants produced from lunar
resources. In Symposium Materials in SpaceScience, Technology
and Exploration (pp. 3947). Boston, MA, USA, 29th November
2nd December.
Heltzel, S. (2014). Latent short circuits failure in high-rel pcbs caused
by lack of cleanliness of pcb processes and base materials, 5th
EMPS Workshop, Estec, Noordwijk, The Netherlands, May 22.
Hinton, Y. L. (1995). Acoustic emission. ASTM Standardization News,
3639.
Hipolite, W. (2015). NASA 3D prints copper rocket engine part. 3D
Printing, April 22.
Heycock, C. T., & Neville, F. H. (1900). Philosophical Transactions of
the Royal Society, 194A: 201
Hillman, D. D., & Chumbley, L. S. (2006). Characterization of tin
oxidation products using sequential electrochemical reduction
analysis (SERA). Soldering and SMT Journal,8(3), 3141.
Hinch, S. W. (1988). Hand book of surface mount technology. Harlow,
England: Longman Scientic and Technical publishers.
Hirnyj, S., & Indacochea, J. E. (2008). Phase transformations in
Ag70.5Cu26.5Ti3 ller alloy during brazing processes. Chemistry
Metals Alloys,1, 323332.
Holt, T. (1995). New applications in high power laser welding. Welding
and Metal Fabrication Journal, 230234.
Homes-Siedle, A., & Adams, L. (1993). Handbook of radiation effects.
Oxford, UK: Oxford University Press.
Hornbogen, E., & Schemme, K. (1990). Prospects of superlight metals
and their laser surface alloying with ceramic phases. In Proceedings
of the First Thermal Structures Conference (pp. 99109). Char-
lottesville, VA, November 1416.
Houlberg, K. (1988). Printed circuit boards for surface mounted
devices. Unpublished report.
Hu, S. J., et al. (1995). Gold wire weakening in the thermosonic
bonding of the rst bond. IEEE Transactions, Components,
Hybrids, Manufacturing Technology,18(1), 228234.
Huang, C.-Y., et al. (2011). Material characterization and failure
analysis for microelectronics assembly processes. Available from:
http://www.intechopen.com /books/wide-spectra-of-quality-con-
trol/material-characterization-and-failureanalysis-for-microelectron-
ics-assembly-processes.
Hummel, R. E. (1994). Electromigration and related IC failure
mechanisms. International Materials Review,39, 97113.
Hunt, C. (2009). Implementation and reliability issues with lead-free
solders. In SMTA International Conference and AIMS Workshop,
San Diego, CA, USA, October 48.
Hunt, C. (2011). Reliability isues with lead-free, EMPS-2 Workshop,
Institut Du Soudure, Paris, France, May 4.
Hunt, C., Dunn, B. D., et al. (2012). Increased shorting in tin whiskers due
to electric elds and contact pressure. In 6th International Symposium
on Tin Whiskers, Loughborough, UK, November 2728.
Hutchings, M., & Krawitz, A. D. (1992). Measurements of residual and
applied stress by neutron diffraction. In NATO ASI series E: Applied
Science (Vol. 216). Dordrecht, The Netherlands: Kluwer.
Hutchinson, B., et al. (2004). Whisker growth from tin coatings, IM-
2004-708. Stockholm: Swedish Institute for Metal Research.
Hwang, J. S. (2014). Tin whiskers: capsulization. SMT Magazine, 1629.
Ilyushenko, R. V. (1993). Weldability of commercial aluminium
lithium alloys. Aluminium-Verlag,69(4), 364370.
Iman, R. L., et al. (1995). Evaluation of low-residue soldering for military
and commercial applicatons: Sandia National Labs Report, June.
IEC 60068-2-82. (2007). Environmental testingWhisker test methods
for electronic and electric components.
645
Jagt, van der, et al. (1991). Anti-wear/corrosion treatment for austenitic
steel: The Hardcor process. Materials & Design,12(1), 4146.
IPC-7095A. (2004). Design and assembly process implementation for
BGAs, October.
IPC-D-279. (1996). Design guidelines for reliable surface mount
technology printed board assemblies, July.
IPC-SM-780. (1987). Advanced packaging and interconnection of
electronic components.
IPC-SM-782. (1992). Surface mount land patterns (conguration and
design rules).
IPC-TM-650. (2000). Electrochemical migration resistance test.
Irving, B. (1992). SCC: Weldings No. 1 nemesis. Welding Journal,
3740.
Isakowitz, S. J. (1995). International reference guide to space launch
systems (2nd ed.). Waldorf, MD, USA: AIAA publication.
Jackson, B. C. (1987). Oxide coatings for bonding multilayer PCBs.
Transactions IMF,65(8), 89.
Jackson, C. R. (1973). Leadindium solder alloy wont embrittle in
gold. Circuit Manufacturing, 4041.
Jacobson, D. M., & Humpston, G. (1992). Diffusion soldering.
Soldering and SMT Journal,10, 2732.
Jacobson, D., & Humpston, G. (1995). Depressing the melting point of
solders and brazes by eutectic alloys. GEC Journal of Research,12
(2), 112121.
Jacobson, D. M., & Sangha, S. (1996). Novel applications of diffusion
soldering. Soldering and SMT Journal,23, 1215.
Jaggers, C. H. (1993). Exposure of LDEF materials to A.O. In 3rd
Post-Retrieval Symposium (pp. 931941). Williamsburg, November
812.
Jalbert, M., et al. (1994). Risque saturnin des oprations de
microsoudure en lectronique. Archives des maladies profession-
nelles et de mdecine du travail,55(8), 589594.
Janu, P., et al. (2009). Electric propulsion pointing hold down and
release mechanism test results. In Proceedings of the 13th European
Space Mechanisms and Tribology Symposium (p. 8)., Vienna,
Austria, September 2325
Jaworske, D. A. (2002). Solar selective coating for power application,
NASA/TM-2002-211333.
JAXA-QTS selected JAXA standards related to materials and processes
can be seen listed in Appendix 9.
Jellison, J. (1986). Evaluation of multilayer printed wiring boards by
metallographic techniques. NASA Reference Publication 1161.
Jenkins, P. P., & Landis, G. A. (1995). A rotating arm using shape-
memory alloy. In 29th Aerospace Mechanism Symposium, League
City, Texas (pp. 167171). May 1719.
JESD22-A121A. (2008). Measuring whisker growth on tin and tin
alloy surface nishes, reafrmed, May 2014.
Jha, A. K., et al. (1996). Failure of AA 2024 aluminium alloy rivets.
Praktische Metallographie,33(5), 264272.
Jiang, B. L., & Wang, Y. M. (2010). Plasma electronic treatment of
aluminium and titanium alloys. In H. Dong (Ed.), Surface engineering
of light alloys. Cambridge, UK: Woodhead Publishing Ltd.
Johnson, G. W. (1979). How metallography and SEM analysis can be
used in quality control for microelectronics. In Metallography as a
Quality Control Tool, Symposium of A.S.M. (pp. 261278). July 89.
Johnson, M.R. (1989). The Galileo High Gain Antenna Deployment
Anomaly. California Institute of Technology, Jet Propulsion Lab-
oratory (JPL), Pasadena, CA.
Johnson, R.E., et al. (1966). A case history of titanium stress corrosion
in nitrogen tetroxide. American Society Metals
Johnson, R. E. (1973). Apollo experience reportThe problem of stress
corrosion cracking. NASA TN D-7111, March.
Johnston, R. L., et al. (1967). Stress corrosion cracking of Ti6Al4 V
alloy in methanol. NASA TN D-3868, February.
646
Kallee, S. W., Nicholes, D., & Thomas W. (2001). Friction stir welding
Invention, innovations and applications. In INALCO 8th Inter-
national Conference on Joints in Aluminium, Munich, Germany,
March 2830.
Karavakis, K., & Bertling, S. (2004). Conductive anodic lament
(CAF). MEPTEC Report, Quarter 4, 2427.
Kauffmann, B., & Wolf, D. (2013). Development of low uoride
evolution uoropolymer wiring materials for aerospace systems. In
Aerospace Electrical Interconnect Symposium, Tampa, FL, USA,
October.
Kelly, M. A., et al. (1984). Using ESCA in failure analysisSome
recent developments. In Proceedings of ISTFA (pp. 3539). Los
Angeles.
Kephart, A. R., & Hayden, S. Z. (1993). Benets of thread rolling
process to the stress corrosion cracking and fatigue resistance of
high strength fasteners. In 6th International Symposium on Envi-
ronmental Degradation of Materials in Nuclear Power Systems
(pp. 18331847). San Diego, CA, August 35.
Kessowsky, R., et al. (1978). Corrosion of indium based solders. In
16th Annual Proceedings, IEEE Reliability Physics (pp. 200205).
Keusseyan, R. L., & Dilday, J. L. (1993). Electric contact phenomena
in conductive adhesive connections. In Proceedings of International
Surface Mount Conference and Exposure (pp. 567571). San Jose,
Surface Mount International.
Ki, Young-Won. (1995). Gamma titanium aluminides: their status and
future. JOM,47(7), 3941.
Kim, K.-H., et al. (2011). Observation of high resolution microstruc-
tures in thermal sprayed coatings and single deposited splats using
ion beam milling. Materials Transactions,52(3), 439446.
King, D. B., et al. (1976). Effects of inclusions on mechanical
properties of Be. AFML-TR-76-33 Brush Wellman Inc., April.
King, R. L. (1988). Review, opportunities for advanced composites in
spacecraft. Advanced Composite Bulletin,2(1), 410.
Kingsley, N. (2008). Liquid crystal polymer: Enabling next generation
conformal and multilayer electronics. Microwave Journal, 51, 188
200.
Klein, A. J. (1988). Specialty reinforcing bres. Advanced Compos-
ites,3(3), 3244.
Kleinteich, T., et al. (2015). Whats inside a shy suction cup?
Microscopy and Analysis,17, S8S10.
Klein Wassink, R. J. (1984). Soldering in electronics. Ayr, Scotland:
Electrochemical Publications Ltd.
Klinkrad, H. (2006). Space debris: Models and risk analysis. Heidel-
berg, Germany: Springer-Praxis Publishing.
Kocker, H. H., & Stockel, D. (1979). Material transfer of composite
contact materials. In IEEE Transactions on Components, Hybrids
and Manufaturing Technoloby (pp. 1519). CHMT-2.
Kodama, S., et al. (2010). Development of stainless steel welding wire for
galvanized steel sheet, Rivista Italiana della Saldatura4: 479486.
Koh, D., et al. (1990). Dermatological hazards in the electronics
industry. Contact Dermatitis,22, 17.
Koh, D., et al. (1994). Skin disorders among hand solders in the
electronics industry. Occupational Medicine,44(1), 2428.
Kohman, G. T., et al. (1955). Silver migration problem encountered in a
telephone exchange. Bell Systems Technology Journal,34, 1115
1147.
Konoza, A., et al. (2012). Analysis of the electronic assembly repair
process for lead-free parts under combined loading conditions.
IEEE Transactions Components Packaging Manufacturing Tech-
nology,2(9), 15581567.
Koons, H. C., et al. (1980). Spacecraft chargingResults from the
SCATHA satellite. Astronautics and Aeronautics,18(11), 4447.
Korb, G., & Dunn, B. D. (1996). Assessment of cleaning methods for space
mechanisms. ESTEC Metallurgy Report No. 2261 (condential).
References
Korb, L. J., et al. (1985). A metallurgical analysis of the failure of two
APUs of the spacecraft Columbia. In Space Technology Confer-
ence Proceedings, Anaheim, USA, September 2325, 5.15.34.
Kostic, A. D. (2011). Lead-free electronics reliabilityAnd update,
GEOINT Development Ofce, Aerospace Corporation, August.
Kleiman, J., Tagawa, M., & Kimoto, Y. (2013). Protection of materials
and structures from the space environment. Heidelberg: Springer.
Kramer, P. A., et al. (1994). The effect of low gold concentrations on
the creep of eutectic tin-lead joints. Metallurgical and Materials
Transactions A,25, 12491257.
Krishnadev, M. R. (1993). The structure and properties of magnesium
matrix composites. JOM,45(8), 5254.
Krumbein, S. J. (1988). Metallic electromigration phenomena. IEEE
Transactions Components, Hybrids, Manufacturing Technology,
CHMT-11, 515.
Kuchler, G. (1995). Assessment of exible substrate integrity. In
Hubble Space Telescope Solar-Array Workshop, ESTEC, Noord-
wijk (pp. 253256). ESA Pubn WPP-77, May 1995.
Kugel, U., Verain, M., & Zarraga, F. M. (1995). Post-ight investi-
gations of the Eureca Mechanisms. In Proceedings of Sixth
European Space Mechanisms Symposium (pp. 195201). Zurich,
(ESA SP-374).
Labruyre, G., & Urmston, P. (1995). ESA mechanisms requirements.
In Proceedings of 6th European Space Mechanism and Tribology
Symposium, Zurich, (ESA SP-374), October 46
Lamouroux, F., et al. (1994). Kinetics and mechanisms of oxidation of
2-D woven C/SiC composites: Experimental approach. Journal of
American Ceramic Society,77(8), 20492057.
Lansdown, A. R., & Price, A. L. (1986). Materials to resist wear.
Oxford: Pergamon Press.
Laporte-Weywada, H. J, et al. (1993). CNES policy about launcher
space debris mitigationAriane V application. In Proceedings of
the First European Conference on Space Debris (pp. 585589).
Darmstadt, Germany, April 57, ESA SD-01.
Larsson, A., & Wingborg, N. (2011). Green propellants based on
ammonium dinitramide (ADN). www.intechopen.com/download/
pdf/13473.
Lau, J., & Rice, D. (1985). Solder joint fatigue in SMT: State of the art.
Solid State Technology, 91104.
Le May, I., & Deckker, E. (2009). Reducing the risk of failure by better
training and education. Engineering Failure Analysis,16, 11531162.
Lea, C. (1988). Restrictions on the use of chlorouorocarbons for
cleaning soldered PCBs. Circuit World,14(4), 412.
Lea, C. (1988). A scientic guide to surface mount technology. Ayr,
Scotland: Electrochemical Publications.
Lea, C. (1991). Evidence that visual inspection criteria for soldered
joints are no indication of reliability. Soldering and SMT Journal,9,
1924.
Lee, H. S., et al. (1994). Symptoms in peak expiratory ow rate among
female solderers. American Journal of Industrial Medicine,26,
613619.
Lee, S.-B., et al. (2006). Electrochemical migration characteristics of
eutectic SnPb solder alloy in printed circuit board. Thin Solid
Films,504, 294297.
Lee, Shen, Y. (1990). Thermomechanical properties of polymeric
materials and related stresses. SAMPE Quarterly, 4851.
Leitao, C. M. A. (2013). Inuence of base material plastic properties
and process parameters on friction stir weldability (Dissertation for
PhD, Department of Mechanical Engineering, University of Coim-
bra, Portugal, 2013).
Lestrat, D., & Prel, Y. (1994). Evolution et choix technologiques pour
les reservoirs des lanceurs Ariane. In Proceedings of the Interna-
tional Symposium on Advanced Materials for Lightweight Struc-
tures, ESTEC (pp. 441453). Noordwijk, March (ESA WPP-070).
References
Levine, A. S. (Eds.). (1993). Third post-retrieval symposium. NASA
Conference Publication 3275, Part 3.
Lewis, P. L., Kolody, M., & Curran, J. (2013). Alternative to nitric acid
passivation of stainless steel alloys. NASA Report, casi.nts.02809.
Leyva, I. A., et al. (2011). Space Mission and Design Analysis,
Chap. 18Propulsion Systems, Air Force Research Lab., Edwards
AFB., CA., USA. Report No. AFRL-RZ-ED-BK-2011-057.
Li, H., & Zhang, J. (2014). Preparation of microarc oxidation ceramic
layer on Ti6Al4V titanium alloy and its oxidation resistance
property. International Heat Treatment and Surface Engineering,8
(2), 5154.
Li, X., & Dong, H. (2010). Ceramic conversion treatment of titanium-
based alloys. In H. Dong (Ed.), Surface engineering of light alloys.
Cambridge, UK: Woodhead Publishing Ltd.
Liedke, V., & Dunn, B. D., et al. (2008). Cleaning of silicone
contaminations: development of test methods and assessment of
cleaning efciency. In 11th ISMSE Symposium, Aix-en-Provence,
France, September.
Liejtens, J. (2015). Will the future of hirel components be pure SCOTS
or only SCOTTish, IAA-B10-1203. In Deutsches Zentrum fur Luft-
und Raumfahrt 10th Symposium, Berlin, Germany, April 2023.
Lin, C.-C., et al. (2012). High temperature brazing of molybdenum.
Advanced Material Research Journal,586, 6973.
Liu, H., et al. (2014). Laser nanocrystallisation and corrosion behavior
of electroless Ni-W-P coating with high phosphorous content.
Transactions IMF,92(4), 212217.
Liu, J., et al. (1996). Development of ip-chip joining technology.
Circuit World,22(2), 1924.
Liu, X., & Lu, G.-Q. (2003). Effects of solder joint shape and height on
thermal fatigue lifetime. IEEE Transactions on Components and
Packaging Technologies26(2): 445465.
Lobley, G. R. (1990). Soldering problems on printed circuit boards.
Practice Metals,27, 627631.
Lucas, J. P., & Zhou, J. (1993) The effects of sorbed moisture on resin
matrix composites. J.O.M., December issue, 3740.
Lucas, W. (1995). Electromagnetic emission from TIG welding
systems. Welding and Metal Fabrication,63(7), 261265.
Luciano, G., & Galet, G. (1995). Space mechanism actuated by a shape
memory alloy component. In Proceedings of the 6th European
Space Mechanical and Tribology Symposium (pp. 221225).
Zurich, October 46. ESA SP-374.
Lumsden, J. B., & Allen, A. T. (1988). The stress corrosion cracking
behaviour of AlLi alloy 8090. Corrosion Science-Nace,44(8), 527
532.
Lumsden, J. M., & Whittlesey, A. (1981). Automatic ARC welding of
propulsion system tubing in close proximity to sensitive electronic
devices. In AIAA/SAE/ASME 17th Joint Propulsion Conference,
Colorado Springs, July, (AIAA-81-1569).
Lund, A. (2011). Principals and applications of ESR Spectroscopy.
Heidelberg: Springer.
Lupton, D. F. (1990). Heraeus GmbH, private communication.
Luzin, V., et al. (2010). Use of neutron diffraction for stress
measurements in thin and thick thermal sprayed coatings. Interna-
tional Heat Treatment and Surface Engineering,4(1), 1724.
Ma, H., et al. (2012). Effects of rework on the accelerated thermal
cycling and shock performance of lead-free BGAs. IEEE Transac-
tions on Components Packaging Manufacturing Technology,2(11),
18241830.
Ma, Y., et al. (2012). A large-scale fabrication of ower-lie sub-
micrometer-sized tungsten whiskers via metal catalysis. Nanoscale
Research Letters,7(325), 17.
Ma, H., & Lee, T.-K. (2013). Effects of board design variations on the
reliability of lead-free solder joints. IEEE Transactions on Compo-
nents Packaging Manufacturing Technology,3(1), 7178.
647
Maciolek, R. B., et al. (1978). Thermal aging characteristics of InPb
solder bonds to gold. Honeywell Report.
McCormick, J., & Zakraysek, L. (1979). A metallographic test for
glass-to-metal seals. In 17th Annual Proceedings of Reliability
Physics, IEEE (pp. 4450). San Francisco, April.
McCormack, S., & Meschter, S. (2009). Probabilistic assessment of
component lead-to-lead tin whisker bridging. In International Confer-
ence on Soldering and Reliability, Toronto, Canada, May 2022.
McCrone, W. C., & Delly, J. C. (1973). The particle atlas (2nd ed.,
Vol. 13). Michigan, USA: Ann Arbor Science Publishers.
McCune, T. B., et al. (1970). Heat ageing evaluation of common coated
copper conductors. In 19th International Wire and Cable Sympo-
sium, Atlantic City, N.J., December 24.
McDonald, G. E. (1974). Renement in black chrome for use as a solar
selective coating. NASA TMX-3136, December.
McDonnald Schetky, L. (1991). Shape memory alloy applications in
space systems. Materials and Design,12(1), 2932.
McDowell, M. E. (1993). Tin whiskersA case study. In 1993 IEEE
Aerospace Applications Conference (14th) (pp. 207215). Steam-
board Springs, CO, USA, January 31February 5.
McKay, C. A. (1983). Causes and effects of solder contamination, part
2, Electronics: 4144.
Maillat, M., et al. (1995). Adhesion of materials by impact contacts in
vacuum. In Proceedings of 6th European Space Mechanisms
Conference, (pp. 431436). Zurich, Switzerland, (ESA SP-374).
Mair, P. (2005). Assessment of EMF and biological effects in arc
welding applications. In International Institute of Welding Manu-
facturing (pp. 110). IIW Document XII-1848-05.
Majid, W. A. (1988). Controlling SCC in mechanism components of
ground support equipment. In 22nd Aerospace Mechanisms Sym-
posium (pp. 163174). Hampton Virginia, May 46 (NASA
Conference Publication 2506).
Maliutina, I. N., et al. (2014). Structure and microhardness of Cu-Ta
joints produced by explosive welding. Rivista Italiana della
Saldatura5: 839848.
Malla, R. B., & Ghoshall, A. (1994). Thermally induced vibrations of
structures in space. Progress in Astronautics and Aeronautics,168,
6889.
Marchaise, M., & Glodowski, K. A. (1991). Real-time microfocus
radiography for electronic failure analysis. Materials Evaluation:
14811485.
Marrocco, T., & Harvey, D. (2012). Cold StandardThe cold spray
process. Materials World: 3031.
Martell, S., & Adams, T. (1996). Virtual cross-sectioning of materials.
Materials World,3(7), 323324.
Martin, J. W. (1976). A new laboratory procedure for determining
compatibility of positive expulsion elastomers with liquid propel-
lants. In SymposiumCompatibility of Plastics (pp. 16). Indian
Head, Md., USA, paper III-E, April.
Mason, M., & Eng, G. (2007). Understanding tin plasmas: a new
approach to tin whisker plasma risk assessment. In IEEE IRPS,
CALCE, University of Maryland, USA, April.
Materassi, M., Dunn, B. D., & Capinari, L. (2000). The inuence of
solder llet geometry on the occurance of corona discharge during
operation between 400 and 900 V in partial vacuum. IEEE
Transactions on Electronics Packaging Manufacturing,23, 104115.
Matloff, G. L. (2013). The speed limit for graphine interstellar solar
proton sails. Journal of the British Interplanetary Society,66,
377380.
Maurer, R. H., et al. (2008). Harsh environments: Space radiation
environment, effects, and mitigation. John Hopkins APL Technical
Digest,28(1), 1729.
Mayo, M. J., & Nix, W. D. (1989). Direct observation of superplastic
ow mechanism in torsion. Acta Met,37(4), 11211134.
648
Maxwell, J. L., et al. (2013). High-temperature nano-composites for
nuclear thermal propulsion and in-space fabrication by HP-LCVD.
JBIS,66, 328333.
Mc Brien, C. M., & Heltzel, S. (2013). Insulation resistance of dielectric
materials under environmental testing. In IPC APEX EXPO Confer-
ence Proceedings, San Diego, Ca, USA, February 1921.
Meaker, A. J., et al. (2006). Hydrogen peroxideFrom bridesmaid to
bride. In 3rd ESA International Conference on Green Propellants
for Space Propulsion, Poitier, France, September.
Medgyes, B., et al. (2011). In situ optical inspection of electrochemical
migration during THB tests. The Journal of Materials Science:
Materials in Electronics,22, 694700.
Meltsner, K. J. (1995). Understanding the INTERNET: A guide for
materials scientists and engineers. JOM,47(4), 913.
Merstallinger, A., & Semerad, E. (1996). Test method to evaluate cold-
welding under static or impact loading. Unpublished ESTEC
Metallurgy Report 2222
Merstallinger, A., & Semerad, E. (1996). Test method to evaluate the
sublimation of metals by weight loss. Unpublished ESTEC Metal-
lurgy Report 2242.
Merstallinger, A., et al. (1995). Study on adhesion/cold-welding under
cyclic load and high vacuum. In Proceedings of the 6th European
Space Mechanisms Conference (pp. 293298). Zurich, Switzerland,
(ESA SP-374).
Merstallinger, A., et al. (1995). Wire bend test. Unpublished report.
Merstallinger, A., & Dunn, B. D., et al. (2009). Assessment of cold
welding between separable contact surfaces due to impact and
fretting under vacuum, ESA STM-279, and, (2015) http://esamul-
timedia.esa.int/multimedia/publications/STM-279/.
Mertens, J., & Gessford, D. (1995). Implementation of alternative
cleaning solvents. Precision Cleaning Journal,3(2), 4550.
Mielke, E. W. (2002). Silver lled conductive epoxies, NASA Materials
Engineering Branch, TIP No. 079, rev. October.
Mikaelian, T. (2001). Spacecraft charging and hazards to electronics
in space, York University publication, May.
MIL-C-13924 Coating, oxide, black, for ferrous metals.
MIL-Handbook-697A. (1974). Titanium and titanium alloys, June.
Millares, M., & Peraggi, B. (1992). Metallography of goldindium
couples and electronic components soldered with indium alloys.
Praktische Metallographie,29, 555563.
Miller, N. H. (1906). The amounts of nitrogen as ammonia and nitric
acid, and of chlorine in the rainwater collected at Rothamsted.
Journal of Agricultural Sciences,1, 280303.
Minges, M. L. (1989). Electronics materials handbook (Vol. 1). Metals
Park, OH, USA: ASM International.
Mohrbacher, H., et al. (1995). The inuence of humidity on the fretting
behaviour of PVD TiN coatings. Wear,180, 4352.
Moliterni, L. (2010). Il saldatore manual utilizzato nella saldatura in
elettronica. Rivista Italianadella Saldatura, Lugili/Agosto 4: 455459.
Moliterni, L. (2012). Lafdabilita degli assemblaggi elettronici. Rivista
Italianadella Saldatura, Maggio/Giugno 3: 351362.
Morilla, D. J., et al. (2015). Tin whisker growth on surface mount
devices: A study using standardized test conditions. In EMPS-6,
Wessling, Germany, April 1516.
Morris, J. W. (1996). The metallurgical control of electromigration.
JOM,48, 4346.
Moser, M. (2002). Calculation of stresses produced in the tin-plated
terminations of resistors during thermal cycling. Unpublished Estec
MGI report 3496.
Motz, J. M. (1988). MicrosegregationsAn easily unnoticed inuenc-
ing variable in the structural description of cast materials. Practice
Metals,25, 285293.
Mueller, R. P., et al. (2014). Space mining and planetary surface
construction. In ASCE Earth and Space Conference, St. Louis,
Missouri, USA, October 2729.
References
Mueller, R. P., et al. (2015). Additive construction using basalt regolith
nes, NASA KSC, casi.ntrs.nasa.gov/20150000305.pdf [cited
access June 2015].
Muller-Wiesner, A., et al. (1994). Materials and manufacturing
technologies for the tank and thruster components of the Ariane 5
SCA-system. In Ariane V structures and technologies, Cepadues-
Editions. Toulouse: CNES publication.
Munns, I. (1995). Adhesive bond inspection using NDT. Materials
World,3(11), 527529.
Munson, T. (2006). White residues: are they or arent they? Circuit
Assembly, page unknown, January.
Nabarro, F. R., & Jackson, P. J. (1958). Growth of crystal whiskers.
Growth and perfection of crystals (pp. 11101). New York: Wiley.
Naisbitt, G. (2015). How clean is clean to be functionally good? EMPS-
6 Workshop, DLR, Oberpfaffenhofen, Germany, April 1516.
Nakasa, K., & Liu, J. (1993). The effect of stress on hydride formation
in Ti6A14V alloy. In Titanium (Vol. 92, 1st ed., pp. 619627).
TMS publication.
NASA NAS 5300.4 (3J). (1996a). Workmanship standard for staking
and conformal coating of printed wiring boards and electronic
assemblies, May.
NASA NAS 5300.4 (3M). (1996b). Workmanship standard for surface
mount technology, May.
NASA-STD selected NASA standards related to materials and
processes can be seen listed in Appendix 9.
Nasta, M. A., et al. (1990). Contaminant particles in electronic solder
uxes. Soldering and SMT Journal,4, 67.
Nelson, H. G. (1988). Hydrogen and advanced aerospace materials.
SAMPE Quarterly,20(1), 2023.
NEMA. (1972). High temperature insulated wire. Report on Copper
Corrosion, NEMA Standards Publication, No. HP 2-1968 (reaf-
rmed by NEMA).
Nemoto, N. (2007). Evaluation of tin whisker growth in vacuum
thermal cycling condition. In 3rd International Symposium on Tin
Whiskers, Lyngby, Deenmark, June 2324.
Newman, R. C. (1990). Stress corrosion cracking: 19651990. British
Corrosion Journal,25, 259269.
Niefhof, J. (1995). An empirical model for electromigration. Solid-State
Electronics,38, 18171827.
Noller, E. W. (1996). Unique application of plastics and composite
materials in the design of magnetometer instruments. In 41st
International SAMPE Symposium and Exhibition (Vol. 1, pp. 841
846). Anaheim, Ca, March 2428.
Nordwall, B. D. (1986). Air force links radar problems to growths of tin
whiskers. Aviation Week and Space Technology30: 6569.
Norwood, L. B. (1985). How beryllium proved successful on the space
shuttle orbiter. Journal of Spacecraft and Rockets,22, 560566.
Novikov, I., & Grushko, O. (1995). Hot cracking of Al-Cu-Li alloys.
Materials Science and Technology,11, 926932.
Nuez, D., & Tan, P. (2014). Failure localization of intermittent short
circuit failures caused by vertical CAF formation. In ISTFA
Conference Proceedings, Houston, TX, USA, November 913.
Nurse, M. (1996). Caution remains on aluminiumlithium alloys. MBM
(pp. 4245). February.
OClock, G. D., et al. (1987). Pb-Sn microstructure: potential reliability
indicator for interconnects. IEEE Transactions Components,
Hybrids, Manufacturing Technology,10, 8288.
Okuyama, F. (1974). Vapor-grown tungsten whiskers induced by
vacuum discharges. Journal of Applied Physics,45(10), 42394241.
Pantazopoulos, G., & Vazdirvanidis, A. (2008). Characterization of the
microstructural aspects of machinable Alfa Beta phase brass
(pp. 1316). Microscopy and Analysis, September.
Paprocki, S. J. (1988). Metal matrix composites for space applications.
In Composite Technology, 5th Annual Conference on Materials
Technology (pp. 3343). Carbondale, IL, April 1415.
References
Parker, D. L. (1977). Special devices fabricated in heat shrinkable
plastics. In Proceedings of the Technical Program, Internepcon77
(pp. 133136). Brighton UK, October 1820.
Parker, D. S. (1991). Analysis of a failed helicopter main rotor blade
grip. In International Conference on Failure Analysis (pp. 141
145). Montreal, Canada, July 811 (published 1992 by ASM).
Parquet, D. T., & Boggs, D. W. (1995). Alternatives to HASL. In
Electronic Packaging and Production (pp. 3842). August.
Pask, J. A. (1948). New techniques in glass-to-metal sealing.
Proceedings of the IRE, Waves and Electrons Section,36, 286289.
Paton, B. E. (2003). Space: Technologies, Materials, Structures. In
Welding and allied processes (Vol. 2). Ukraine: E. O. Paton Electric
Welding Institute. Published in English by Taylor and Francis Inc.,
New York, USA.
Patterson, R. F., & Mykytiuk, P. D. (1996). Solid-state, vapour phase,
and aqueous cleaners for aerospace manufacturing operations.
Sampe Journal,32(3), 4049.
Pearce, R. (1987). SuperplasticityAn overview (pp. 11 to 124).
AGARD Lecture Series No. 154, August.
Pecht, M. G., et al. (1999). Moisture ingress into organic laminates.
IEEE Transactions on Components and Packaging Technology,22,
104110.
Pedley, M. D. (2009). Materials safety, Ch. 12 in book safety design for
space systems. Amsterdam, The Netherlands: Elsevier Publishing.
Peel, C. J., & McDarmaid, D. S. (1989). The present status of the
development and application of Al-Li alloys 8090 and 8091.
Aerospace,16(5), 1823.
Peng, X., et al. (2014). Conductivity improvement of silver akes lled
ECAs via introducing silver-graphine nanocomposites. Journal of
Materials Science: Materials in Electronics,25, 11491155.
Perami, R., et al. (1992). Applications of AE to the study of
microssure damage to composites used in the aeronautics and
space industries. In The American Society for Nondestructive
Testing, Proceedings of AECM-4, (pp. 99108), Seattle, July 2731.
Pereira, A., Pimenta, G., & Dunn, B. D. (2008). Assessment of chemical
conversion coatings for the protection of aluminium alloys, (A
comparison of Alodine 1200 with Chromium Free Conversion
Coatings), ESA-STM-276, February.
Peres, P., Semerad, E., & Dunn, B. D. (1994). Material characteristics
of SiC/MAS-L bre re-inforced glass composites. In Proceedings of
6th International Symposium on Materials in Space Environmental
(pp. 309314). Noordwijk, September 1923, 1994, (ESA SP-368).
Perez, F., & Eymard, P. (1985). Increasing use of composite structures
on Ariane launchers (pp. 239244). ESA SP-243.
Pergola, P., et al. (2011). Low-thrust missions for expanding foam space
debris removal. In IEPC-2011-126, 32nd International Electric
Propulsion Conference, Wiesbaden, Germany, September 1115.
Peters, S. T., & Wesling, N. (1968). Corrosion of silver plated copper
conductors. In SAMPE 13th National Symposium.
Pettersson, C., et al. (2007). Corrosion testing of welds, a review of
methods. Rivista Italiana Saldatura,4, 529553.
Pettit, R. B., & Sowell, R. R. (1976). Solar absorptance and emittance
properties of several solar coatings. Journal of Vacuum Science and
Technology, 13(2).
Pinaud, G., et al. (2014). Aerofast: Development of cork TPS material
and a 3D comparative thermal/ablation analysis of an Apollo and a
biconic sled shape for an aerocapture mission. http://solarsystem.-
nasa.gov/docs/7_aerofast_paper.pdf (2015).
Pippin, H. G. (1995). Analysis of silverized teon thermal control
material ow on LDEF. NASA Contractor Report 4663, July.
Pippin, H. G., & Bourassa, R. J. (1995). Effects of space exposure on
metals on LDEF. NASA Contractor Report 4662, June.
Pirnnack, S. (2012). Lessons learned to avoid coax cable failure in
moving mechanical mechanisms. In Proceedings of the 41st Aero-
space Mechanisms Symposium, NASA JPL, Pasadena, CA, USA.
649
Piskoty, G., et al. (2009). Structural failures of rope-based systems.
Journal of Engineering Failure Analysis,16, 19291939.
Plumbridge, W. J. (2011). Tin pest in lead-containing solders.
Soldering and SMT Journal,22(1), 5657.
Poe, R. F., & Ruckner, M. A. (1993). Evaluation of pressurized vessels
following hypervelocity particle impact. In Proceedings of the First
European Conference on Space Debris (pp. 441446). Darmstadt,
Germany, April 57, (ESA SD-01).
Pohl, D. (1996). Advances in scanning electron microscopy. Praktische
Metallographie,33(5), 235245.
Polat, A., et al. (2008). Properties of alumina coating formed by
microarc oxidation technique on 6061 aluminium alloy. Praktische
Metallographie,12, 594606.
Pollack, A. A. (1989). Acoustic emission inspection. In Metals
Handbook (Vol. 17, 9th ed., pp. 278294). ASM International.
Pops, H. (1985). Copper rod requirements for magnetic wire. In
Proceedings of the 17th Electrical/Electronics Insulation Confer-
ence (pp. 329337). Boston, USA, September 30October 3.
Pradier, A. L., & Dosio, D. (1996). Technology development in
structures for reusable launch vehicles. In Reaching for the Skies
(Vol. 15, p. 7). ESA Publication.
Premat, G. (1977). Astudy of the outgassing of solder splices. ESA TM-
158.
Pretsch, E. (2009). Structure determination of organic compounds.
Heidelberg: Springer.
Prymak, J. D., & Bergenthal, J. (1995). Capacitance monitoring while
ex testing. IEEE Transactions on CPMT, Part A,18(1), 180186.
Puig, C. (2006). When space adheres to bondingAn interview.
Qualite Espace,43, 5763. Mars.
Puippe, J-Cl. (1999). Surface treatments on aluminium for space
applications. Galvanotechnique,90(10), 30033009.
Puttlitz, J. K. (1990). Preparation, structure and fracture modes. IEEE
Transactions Parts, Hybrids, Packaging,13, 648655.
Quinn, G. D. (2008). Fractography of brittle materials: Analysis
of fractures in ceramics and glasses. Microscopy and Analysis:
2124.
Raiser, G. F., & Amir, D. (2005). Solder joint reliability improvement
using cold ball pull metrology. In Proceedings of InterPACK05,
San Francisco, CA, USA, July 1722.
Raman, V., & Reiley, T. (1987). Cavitation and cracking of
superplastic PbSn eutectic during high-temperature fatigue. Jour-
nal of Materials Science Letters,6, 549551.
Rameriz, P., & Adams, T. (2002). Acoustic imaging and screening of
ceramic chip capacitors. Passive Component Industry: 1620.
Rampini, R., et al. (2009). Dynamic outgassing test: Revisited. In 11th
ISMSE Symposium, Aix-en-Provence, September 1518, Poster.
Ratchev, P., et al. (2005). A study of brittle to ductile fracture transitin
temperatures in bulk lead-free solders, EMPC, Brugge, Belgium,
June 1215.
Ray, U., et al. (1995). Inuence of temperature and humidity on the
wettabiality of immersion tin coated pwbs. IEEE Transactions on
CPM Technology Part A,18(1), 153161.
Raymond, L. (1988). Hydrogen embrittlement: Prevention and control.
No: ASTM Special Technical Publication. 962.
Redahan, E. (2014). 3D printingMaterial wise to grain sizeA review
of the 3D printing and additive manufacturing industrial applications
summit (London 2013). In Materials World, January 1012.
Reed, R. P., & Clark, A. F. (1983). Materials at low temperature.
ASM.
Reid, P. (2009). PWB reliabilityThe next step. Circuit World,33(4),
5159.
Rendu, M., & Tawil, D. (1988). Improved protection of magnesium
alloys against synthetic aviation lubricants at elevated temperatures,
paper 880869. In 24th Annual Aerospace/Airline Plating and Metal
Finishing Forum, Phoenix, April.
650
Reuther, J. (2010). Orion thermal protection system, advanced
development project. In 7th International Planetary Probe Work-
shop, NASA Ames Research Centre, June 16.
Richards, B. P., & Footner, P. K. (1990). Failure mechanisms in active
surface mount components. GEC Journal of Research,7(3), 146156.
Ries, W. (1996). The combined plasma cleaning facility. Product
Finishing,49(1), 68.
Roman, I., et al. (1992). Interfacial shear properties and acoustic
emission behaviour of model aluminium and titanium matrix
composites. In The American Society for Nondestructive Testing,
Proceedings of AecM-4 (pp. 109114), Seattle, July 2731.
Richards, B. P., & Footner, P. K. (1984). Morphological aspects of
electromigration. GEC Journal of Research,2(3), 157168.
Ring, K. (2010). Protection of electronic assemblies in condensing
atmospheres. EMPS-1, Portsmouth, UK, February 17.
Roach, T. A. (1984). Aerospace high performance fasteners resist stress
corrosion cracking. Material Performance,239, 4245.
Robertson, I. M., & Teter, D. (1996). Microscopic studies on hydrogen
effects on mechanical properties. JOM,48(11), 5560.
Rogers, K., et al. (1999). Conductive lament formation in a pcb.
Circuit World,25, 69.
Ross, C. A., & Evetts, J. E. (1987). A model for electromigration.
Scripta Metals,21, 10771082.
Ross, R. G. (1989). Magellan/Galileo solder joint failure analysis and
recommendations. NASA-CR-186294, September.
Rothschild, B. F. (1981). Electroplating of solderable coatings. Metal
Progress: 2529.
Rowntree, R. A., & Todd, M. J. (1988). Tribology for spacecraft.
Chartered Mechanical Engineer Journal: 2830.
Rubinstein, M. (1978). High speed electroplating of gold. Gold
Bulletin,11(1), 38.
Russell, C., et al. (2014). Welding technology takes ight with NASA.
Welding Journal,93(5), 3844.
Rustichelli, F., & Dunn, B. D. (1997). Assessment of residual stresses
in AA 2219 weldments by neutron diffraction. To be published in
Materials Science and Engineering.
Ryan, R. S., et al. (1995). Development problems and their solutions
for the Space Shuttle main engine. NASA Technical Paper 3533,
May.
Saber-Samandari, S., & Berndt, C. C. (2010). IFTHSE Global 21: Heat
treatment and surface engineering in the twenty-rst century.
International Heat Treatment and Surface Engineering,4(1), 713.
Said, A. F., et al. (2012). Automated void detection in solder balls in
the presence of vias and other artifacts. IEEE Transactions on CPM
Technology,2(11), 18901901.
Saltzman, M., et al. (2009). Problems encountered during the
recertication of a solar array drive actuator. In 13th Space
Mechanisms and tribology Symposium, Vienna, Austria, September
2324.
Sampson, B. (2007). Soldering onTraining focus. In Professional
Engineering (p. 48), April.
Sandera, J. (1996). Investigation into the mode of degradation of
crimped joints due to corrosive effects. Unpublished ESTEC
Metallurgy Report 2256.
Saunders, C. G. (1988). Ion techniques in surface engineering. Metals
and Materials,4(11), 678682.
Saxer, W. (1977). Moeglichkeiten funktioneller Edelmetallueberzuege
anhand aktueller Beispiele aus der Praxis. Galvanotechnik, 68(11)
(Black inorganic coating developed by Werner Fleuhmann AG,
Dueberdorf, Switzerland).
Saxty, P. (1995). Ultrasonic soldering in the electronics industry.
Metallurgica: 287.
Scala, E. P. (1996). A brief history of composites in the US. JOM,48
(2), 4548.
References
Schatz, et al. (1979). Development and ight experience of the Voyager
propulsion system. In AIAA/SAE 15th Joint Propulsion Conference
(pp. 113), Las Vegas, June.
Schedler, A. (1988). Fibre composites in satellites. Cryogenics,28,
220223.
Scheel, W. (2004). Electronics assembly handbook (2nd ed.). Port Erin,
I.O.M, UK: Electrochemical Publications.
Schelling, P., et al. (2005) Managing heat for electronics. Materials
Today: 3035.
Schemme, K., & Wittkamp, I. (1993). Microstructure development and
examination of MgLi alloys. Praktische Metallographie,30(7),
334343.
Schlesinger, M. (2010). Electroless deposition of nickel. In M.
Schlesinger & M. Paunovic (Eds.), Modern electroplating. Chich-
ester, UK: Wiley.
Schliekelmann, R. J. (1972). Holographic interference as a means for
quality determination of adhesive bonded metal joints. In 8th
Congress of the International Council of the Aeronautical Sciences,
Paper No. 72-06.
Schmidt, E. W. (1984). Hydrazine and its derivatives. New York:
Wiley.
Schoenbeck, J., & Dunn, B. D. (2004). Evaluation of a process for the
repair of area array and other surface mounted packages, ESA
STM-272, September.
Schoenthaler, D. (1984). Accelerated ageing for solderability evalua-
tions. IPS-TR-464.
Schoyer, H. F. R. (1996). ESAs new solid propellant based on
hydrazinium nitroformate. Preparing for the Future,6(1), 67.
Schreib, R., et al. (1979). Flow anomalies in small hydrazine thrusters.
In AIAA 15th Joint Propulsion Conference (pp. 17), Las Vegas,
June.
Schrunk, D., et al. (1999). The MoonResources, future development
and colonization. Chichester, UK: Wiley-Praxis.
Schubert, Th, et al. (2008). Interfacial design of Cu-based composites
prepared by powder metallurgy for heat sink applications. Materials
Science and Engineering: A,475, 3944.
Schuerch, H. U. (1968). Certain properties and applications of Nitinol
NASA CR-1232.
Schweigart, H. (2007). Conformal coating issues: When reliability goes
astray. Global SMT and Packaging, 3.
Scott, B. C. (1987). Zinc diffusion in tin coatings on brass. Transac-
tions on IMF,65, 9098.
Scott, P. G. (1985). An investigation of alodine surface nish on three
aluminium alloys. EWP No. 1435 (Unpublished).
Sechi, Y., et al. (2009). Composition dependence of titanium in Ag-Cu-
Ti alloy braze metals for laser brazing of boron nitride ceramics and
cemented carbides. Material Transactions,50(6), 12941299.
Seghesio, J. L. (1986). Heurs et malheurs de la corrosionle tore deau
dAriane. CNES Quality Espace,16(5), 2932.
Selman, G. L., et al. (1974). Dispersion strengthened platinum.
Platinum Metals Review,18(2), 213.
Semerad, E., & Dunn, B. D. (1990). Evaluation of black anodised
beryllium (Unpublished report).
Semprimoschnig, C., & Eesbeek, M. (2007). Materials research at Estec
focus on self-healing materials for space applications. In STFC
Materials Workshop, London, October 10.
Sharkov, E. A. (2008). Breaking Ocean Waves - geometry, structure
and remote sensing. Heidelberg, Germany: Springer-Praxis.
Sharma, A. K. (1989). Gold plating on aluminium alloys for space
applications. Transactions of the Institute of Metal Finishing,67,
8788.
Shashkov, P. (2013). Aluminium nanoceramic substrates for thermal
management in electronics. In Electronic Materials and Proceed-
ings For Space Workshop, (EMPS-4), Aalborg, Denmark.
References
Sheldahl Inc. (2004). Thermal control material and metallised lms.
General Trade Literature and Specications, and The Little Red
Book.
Shibutani, T. (2009). Effect of grain size on pressure-induced tin
whisker formation. In 3rd International Tin Whisker Symposium,
Lyngby, Denmark, June 2324.
Shiganov, N. V. (1973). Some special features of arc welding with a
hollow electrode in vacuum. Welding Production, 20(9), 3033
Shipley, E. A. (1988). Developments in ultra high tensile wire ropes. In
Proceedings of Conference on the Hoisting of Men, Materials and
Minerals (pp. 12951307), Toronto.
Shrestha, S., Merstallinger, A., Sickert D., & and Dunn, B.D. (2003).
Some preliminary evaluations of black coating on aluminium
AA2219 alloy produced by plasma electrolytic oxidation (PEO)
process for space applications. In Proceedings of 9th International
Symposium on Materials in a Space Environment (pp. 5765), The
Netherlands, June 16-20.
Shrestha, S., et al. (2007). Recent developments in black nish coatings
on aluminium by Keronite plasma electrolytic oxidation. Proceed-
ings of the Global Power Train Congress, Engine, Transmission
and Propulsion, Berlin,3942, 207215.
Shrestha, S., & Dunn, B. D. (2010). Plasma electrolytic oxidation and
anodizing of aluminium alloys for spacecraft applications. In H.
Dong (Ed.), Surface engineering of light alloys. Cambridge, UK:
Woodhead Publishing Ltd.
Shrestha, S., et al. (2015). Silk corrosion and stress corrosion cracking
behavior of Plasma electrolytic oxidized Al-SiC metal matrix
composite, to be published.
Shumka, A., & Pietry, R. (1975). Migrated-gold resistive shorts in
microcircuits. In Proceedings of 13th IEEE International Reliability
Physics Conference (pp. 9398), Las Vegas.
Sicking, R. (2005). Brazing of aluminium heat exchangers. Welding
and Cutting Journal,4(3), 150159.
Silverman, E. M. (1995). Space environmental effects on spacecraft:
LEO materials guide. NASA Contractor Report 4661 Part 1,
August.
Simpson, P. R., & Bowie, S. H. U. (1970). Quantitative optical and
electron-probe studies of the opaque phases. Science,167, 619621.
Sinclair, J. D., et al. (1985). Volatile components in rosin ux:
characterisation and reliability effects. In 35th IEEE Electronic
Components Conference (pp. 144150), Washington DC, May 2022.
Singh, J., et al. (2005). Welding in Space. Advanced Materials and
Processes Journal: 5257.
Smith, C. A. (2007). Chemical characterization of materials in
electronic systems using infrared spectroscopy. Circuit World,33
(3), 3847.
Sogame, E., et al. (1990). Recent developments in the hydrazine related
propulsion technology in NASDA. In Proceedings of 17th Inter-
national Symposium on Space Technology and Science (Vol. 1,
pp. 107112), Tokyo.
Sohn, J. E., & Ray, U. (1995). Weak organic acids and surface
insulation resistance. Circuit World,21(4), 2225.
Solari, G., & Repetto, M. P. (2014). IL comportamento delle structure
metalliche saldate nei confronti delliazione del vento. Rivista
Italianadella Saldatura: 851859.
Soli, L., & Dunn, B. D., et al. (2015). An electron beam brazing reow
technique. Giorate Nazionali di Saldatura, GNS8, Genoa, Italy,
May 2829.
Somiya, I., et al. (2013). Products with special electrical and magnetic
properties, Ch. 7. In Handbook of advanced ceramics: Materials,
applications, processing and properties. Amsterdam, The Nether-
lands: Elsevier Inc.
Sood, B., et al. (2011). Whisker analysis of electronic throttle controls.
Circuit World,37(3), 49.
651
Spedding, V., & Hargreaves, P. (1995). Materials on line. Materials
World,3(10), 483484.
Sriveeraraghavan, S., et al. (1995). Chemical removal of solder and
solder electroplates. Metal Finishing: 3436.
Stanley, J. K. (1968). Solutions to some stress corrosion cracking
problems in aerospace situations. In Proceedings of 1st Interna-
tional Aerospace and Marine Corrosion Technical Seminar,
NACE, LA, USA, July.
Stoltz, R. E., & Stulen, R. H. (1979). Slid metal embrittlement of Ti
6A16V2Sn by cadmium, silver and gold. Corrosion-NACE,35
(4), 165169.
Strachan, W. (2000). Processes for manual solder assembly of high
heat capacity boards (Unpublished).
Stratton, P. (2013). Ellingham diagramsTheir use and misuse.
International Heat Treatment and Surface Engineering Journal,7
(2), 7073.
Sterjovski, Z., et al. (2010). The effect of voltage and metal transfer
mode on particulate fume size. Rivista Italiana della Saldatura,6,
765780.
Sturgeon, A., & Dunn, B. D. (2006). Cold sprayed coatings for polymer
composite structures. In Proceedings of 10th ISMSE Conference,
Collier, France, June 1923 (ESA SP-616).
Suchentruck, R. (1986). Electroforming of complex parts. Transactions
IMF,64, 1923.
Sun, J., et al. (2012). Kinetic study of the pyrolysis of waste pcbs
subject to conventional and microwave heating. Energies,5, 3295
3306.
Surkov, Y. (1997). Exploration of terrestrial planets from spacecraft
(2nd ed.). Chichester, UK: Wiley-Praxis.
Suslov, D. I., et al. (2010). Convective and lm cooled nozzle
extension for a high pressure rocket subscale combustion chamber,
AIAA 2010-1150. In 48th AIAA Aerospace Sciences Meeting
Including New Horizons Forum, Orlando, FL, USA, January 47.
Tak, W, et al. (2009). SMA actuator based satellite separation device
using plastic deformation. In 13th European Space Mechnisms and
Tribology SymposiumESMATS, Vienna, September 2325 (ESA
SP-670).
Tamai, T. (1995). Effects of silicone vapour and humidity on contact
reliability. In Proceedings of IEEE Holm Conference (pp. 252259).
Tamai, T., & Aramata, M. (1993). Safe levels of silicone contamination
for electrical contact. In Proceedings of IEEE Holm Conference on
Electrical Contact Phenomena (pp. 269273), England
Teare, M. (1994). Manufacture of a multi-horn feed by electroforming.
In International Symposium on Advanced Materials for Lightweight
Structures, ESA-ESTEC (pp. 305309), March 2224. ESA WPP-
070.
Teer, D. G. (2001). New solid lubricant coatings. Wear,251, 1068
1074.
Tegehall, P.-E. (1991). The impact of crevices beneath smds on the
cleaning efciency. Soldering and SMT Journal,8, 4652.
Tegehall, P.-E., & Dunn, B. D. (1992). Inuence of ux residues and
conformal coating on the SIR properties of spacecraft pcbs. ESA
Journal,16(3), 255274.
Tegehall, P.-E., & Dunn, B. D. (2001). Assessment of the reliability of
solder joints to ball and column grid array packages for space
applications, ESA-STM-266.
Tegehall, P.-E., & Dunn, B. D. (2001a). Evaluation of thermally
conductive adhesives as staking compounds during the assembly of
spacecraft electronics. ESA STM-265, August.
Tegehall, P.-E., & Dunn, B. D. (2003). Impact of cracking beneath
solder pads in PCB laminate on reliability of solder joints made to
ceramic ball grid array packages. ESA-STM-267.
Tegehall, P.-E., & Dunn, B. D. (2005). Impact of reworking ceramic
AGAs on the integrity of PCB laminates. ESA STM-273, June.
652
Tegehall, P.-E., & Dunn, B. D. (2006) Evaluation of cleanliness test
methods for spacecraft PCB assemblies. ESA-STM-275.
Tegehall, P.-E., & Wetter, W. (2015). Impact of laminate cracks under
solder pads on the fatigue lives of BGA solder joints, Microelec-
tronics Reliability, article in press.
Thiel, C., et al. (1993). Functional activity of plasmid DNA after entry
into the atmosphere of Earth investigated by new biomarker
stability assay for ballistic spaceight experiments. PLOS One,
November 26.
Thiel, M., et al. (2003). The Rosetta Lander Anchoring System. In
Proceedings of the 10th European Space Mechanisms and Tribol-
ogy Symposium (pp. 239246), San Sebastian, Spain, September
2426.
Thoma, W. (1981). Product liability. In 2nd ESA Product Assurance
Symposium (pp. 3133), ESA SP-163.
Thomas, G., et al. (1993). The effect of pre- and post-weld heat
treatments on the mechanical properties of EB welded Ti6A14V.
Journal of Material Science,28, 48924899.
Thomas, G., & Dunn, B. D. (1999). Assessment of electrically
conductive adhesives (ECAs) for joining surface mount devices to
printed circuit boards. In 2nd IEEE Symposium on Polymeric
Electronic Packaging (PEP 99) (pp. 913), Gothenburg, Sweden,
October 2428.
Thomas, R. W., & Calabrese, D. W. (1983). Phenomenological
observations on electromigration. In 21st Annual Proceedings in
Reliability Physics (pp. 19), Phoenix, Arizona, April 57.
Thompson, D. H. (1961). A simple stress corrosion cracking test for
copper alloys. Materials Research and Standards,1(2), 108111.
Thomson, P., et al. (1995). The use of shape memory alloys for passive
structural damping. Smart. Materials and Structures,4, 3642.
Tighe, A. et al (2010) Overview of results from the materials exposure
and degradation experiment (MEDET) after 18 months in orbit on
the ISS, ESA paper.
Tillmann, W., & Boretius, H. (2008). Properties of nickel-based joints
between diamond and steel for diamond grinding tools. Welding
and Cutting,7(4), 228234.
Toensmeier, P. A. (2009). Shape memory polymers reshape product
design. In Plastics engineering, April 2.
Towner, J. M., et al. (1984). Grain growth in Al alloy conductors as a
result of rapid annealing. Applied Physics Letters,44, 198199.
Tondu, T., et al. (2011). Cesium droplet evaporation as contamination
route by cesium eld effect electric propulsion. Journal of
Spacecraft and Rockets,48(3), 513519.
Troughton, M., et al. (2013). State of the art in welding thermoplastics
and in assessing thermoplastic welded joints. Rivista Italiana della
Saldatura,6, 887895.
Tsuda, N. (2000). Electronic conduction in oxides (p. 243). Heidelberg,
Germany: Springer.
Tsukada, M., et al. (2012). A new strategy for assessing off-gassing
from museum materials. AIC News,37(1), 17.
Tulsi, S. S. (1986). Properties of electroless nickel. Transactions
IMF,64, 7376.
Turlach, G. (1985). Improving fatigue strength of aerospace metal
fasteners by surface work hardening. Surface Engineering,1(1),
1722.
Turner, I., Dunn, B. D., & Barrnes, C. (2013). A study into the re-
processing of pure tin termination nishes into tin-lead. Soldering
and SMT Journal,25(4), 218228.
Turner, R. F. (1995). Chapter 9. In P. Fortescue & J. Stark (Eds.),
Spacecraft systems engineering (2nd ed.). Chichester: Wiley.
Turner, T. E., & Parsons, R. D. (1982). A new failure mechanism: A1
Si bond pad whisker growth. IEEE Transaction on Components,
Hybrids and Manufacturing Technology,CHMT-5, 431435.
References
Tyler, D. (2015) Reliability of ball grid arrays converted from lead-free
to tin-lead by robotic hot solder dipping, EMPS-6, Wessling,
Germany, April 1516.
Vail, J. R., et al. (2011). Polytetrauoroethylene ber reinforced
polyetheretherketone composites. Wear,270, 737741.
Vander Voort, G. F. (2008) Titaniumspecimen preparation,
Advanced Material and Processes, February 2527.
van Eesbeek, M., et al. (1994) Degradation of Teon FEP due to VUV
and atomic oxygen exposure. In Proceedings of 6th International
Symposium on Matls in Space Environment, ESTEC (pp. 165173).
ESA SP-368.
Vancso, J, et al. (2009) Whats new in atomic force microscopes for
polymers. Microscopy and Analysis, May 511.
Vaughn, R. L. (1985). Space shuttleA triumph in manufacturing.
Dearborn, USA: Society of Manufacturing Engineers.
VEGA. (2006) Users Manual, Issue 3 rev.0, and ESA-BR-257 (2007).
Veit, H. M., et al. (2002). Using mechanical processing in recycling
printed wiring boards. JOM, June 4550.
Verbraak, C. A., TNO Metaalinstituut. Delft: Private Communication.
Verderaime, V., & Vaughan, R. (1995) Aluminium U-groove weld
enhancement based on experimental stress analysis. NASA Techni-
cal Paper3581.
Verdin, D., & Duck, M. J. (1985). Conductive coatings to minimise the
electrostatic charging of Kapton. 3rd European Symposium on
Spacecraft Materials in Space Environment (pp. 125131). ESA
SP-232.
Vine, M. K., & Price, W. B. (1992) A literature review of fretting in
launch, re-entry and space environments. ESTEC Contractor Report
from ESTL (TM-106).
Vitt, B. (1987). Black-cobalt coating for solar collectors. Philips
Technical Review,43, 244252.
Voigtlander, B. (2015). Scanning Probe Microscopy. Heidelberg:
Springer.
Volkert, C., & Minor, A. M. (2007). Focused ion beam microscopy and
micromachining. MRS Bull,32, 389395.
Vranish, J. M., & Gorevan, S. (1995). Basic space payload fastener.
29th Aerospace Mechanisms Symposium, NASA JSC (pp. 152157).
NASA Conf. Publication 3293.
Wagner, R. (1992). Metallized CFRP technology. Germany: Private
Communication, Dornier Space Systems.
Walmsley, M. (2015). Ruggedised solder joint technology for leadless
semiconductor packages. EMPS-6, Wessling, Germany, April 15
16.
Ward, P. (1994). Effects of bearing wash procedures in performance.
Non-Ozone depleting Chemical Cleaning EPA Workshop, Denver,
Co, September 2627.
Ward, P. P. (1996). Plasma cleaning techniques and future applications
in environmentally conscious manufacturing. SAMPE Journal,32
(1), 5154.
Warwick, M. E., & Muckett, S. J. (1983). Observations on the growth
of intermetallics on tin-coated substrates. Circuit World,9(4), 511.
Watson, J. A., et al. (2002) A history of astronaut construction for large
space structures at NASA LRC. In IEEE International Space
Station Conference. Nyact, NY, USA, March 916.
Webster, D. (1994). Aluminiumlithium, the next generation.
Advanced Materials and Processes,145(5), 1824.
Wedgewood, A. (1994). Measuring residual stress. Materials World,3
(1), 57.
Weirick, L. J. (1975). A metallurgical analysis of stress corrosion
cracking of Kovar package leads. Solid State Technology,18(3), 25
30.
Wescon. (1965). Session 16B Papers (Western Electronic Show and
Convention, California, 1965), July.
References
Wetzer, J. M., & Wouters, P. (1993). The effect of insulator charging on
breakdown and conditioning. IEEE Transactions,EI 28(4), 681691.
Wetzer, J. M., & Wouters, P. (1994). High-voltage design of spacecraft
vacuum components. Paper from the High-voltage and EMC
Group, Eindhoven University of Technology, The Netherlands
Whalen, M. V. (1988). The compatibility of dispersion-strengthened
platinum with propellants. Platinum Metals Rev,32(1), 210.
Wickham, M., Hunt., Adams, D., & Dunn, B. D. (1999). An
investigation into ball grid array inspection techniques, ESA
STM-261.
Wickham, M. et al. (2014). Whisker mitigation measurements using a
test vehicle based on a printed circuit assembly. 8th International
Symposium on Tin Whiskers. Raleigh, NC, USA, October 29.
Wild, R. N. (1975) Some fatigue properties of solders and solder joints.
INTER-NEPCON, Brighton, UK. IBM report No. 74Z000481,
October.
Williams, D. N, et al. (1970). Hydrogen segregation in Ti6A14V
weldments made with unalloyed titanium ller material. Welding
Journal Research Supplement, May 207212.
Wingborg, N., de Flon, J., Johnson, C., & Whitlow, W. (2008) Green
Propellants Based on ADN. Space Propulsion 2008. Heraklion,
Crete, Greece. ESA, 3AF, SNPE, May 58, 2008.
Winkler, W. (1975). Conductive coating: problem of electrostatic
cleanliness. Acta Astronautica,2(78), 745754.
Wolfenden, R. T. (1996). Problems with solvent vapours in arc welding
fabrication environments. Welding and Metal Fabrication,64(4),
136139.
Wolff, R. H. (1963). Corrosion of plated metals after heating in
vacuum. Plating Journal, 905910, October.
Wright, C. (1977). The effect of solid-state reactions upon solder lap
shear strength. IEEE PHP,13(3), 202207.
Wu, X. J., et al. (1994). The orientation dependence of fatigue crack
growth in 8090 A1Li plate. Metallurgical and Materials Trans-
actions A,25, 575588.
Xu, H., et al. (2009). A re-examination of the mechanism of
thermosonic copper ball bonding on aluminium metallised pads.
Scripta Meterialia,61(2), 165168.
Xu, H., et al. (2010). The role of bonding duration in wire bond
formation: a study of footprints of thermosonic gold wire on
aluminium pad. Microelectronics International,27(1), 1116.
653
Xu, H., et al. (2013). Wafer-level SLID bonding for MEMS encap-
sulation. Advances in Manufacturing,1, 226235.
Yao, W., & Basaran, C. (2014). Damage mechanics of electromigration
and thermomigration in lead-free solder alloys under alternating
current: An experimental study. International Journal of Damage
Mechanics,23(2), 203221.
Yao, Y., Zhou, Y., & He, L. (2013). Corrosion behavior of molybdate
conversion coating on AZ31 magnesium alloy in NaCl solution.
Anti-Corrosion Methods and Materials,60(6), 301311.
Yext, W. F., et al. (1983). Improved glass-to-metal sealing through
furnace atmosphere control. IEEE Transactions on Components,
Hybrids and Manufacturing Technology CHMT-6 (pp. 455459),
December
Ylikorpi, T., et al. (1995). Development of an actuator for CAPS on
spacecraft Cassini. In Proceedings of 6th European Space Mecha-
nisms Conference (pp. 97101). Zurich, Switzerland, (ESA SP-374).
Yoder, D. D., et al. (1993). Technical article on plated through hole
fatigue life, Part 2. IPC Review,34(5), 2027.
Yost, F. G. (1976). Indium soldering on gold. In Proceedings of
International Microelectronics Symposium (p. 61).
Yost, F. G., et al. (1976). Layer growth in Au-Pb/In solder joints.
Metallurgical Transactions A,7A, 11411156.
Zhang, W., & Jin, H. (2009). Study of tungsten metallization surface
states for multilayer ceramics. In IEEE Proceedings of ICEPT-HDP
09 International Conference. Beijing, China, August 1013.
Zhao, H., et al. (2013). Controlled atmosphere brazing of aluminium
heat exchangers. Welding Journal, February 4446.
Zimcik, D. G. (1987). Effect of long-term exposure to LEO space
environment on spacecraft materials. Canadian Aeronautics and
Space Journal,33(1), 410.
Zirker, L. R., et al. (2007). Weld tests conducted by the Idaho National
Laboratory, INL/EXT-07-12289.
Zou, L., Dunn, B. D., et al. (1999). An evaluation of the effect of ageing
on the cleanability of solder ux residues. Soldering and SMT
Journal,11, 2735.
Zwanenburg, R. (1995) The qualication test programme of the Envisat
solar array mechanism. In Proceedings of 6th European Space
Mechanisms Conference (pp. 6572), Zurich, Switzerland. ESA
SP-374..
 Index

Note: Bold page numbers are used for main references. An italic number refers to an illustration on the page.

09 Al-2219 wrought product, 136

0.5 N thruster, 290 Alclad products, 134
2 m diameter L-band dish antenna after test, 298 AlLi alloy plate, 143
3D printing (or Additive manufacturing), 12, 33 AlLi alloys, 103
50In50Pb, 454 AlLi cryogenic tank, 232
50In50Pb solder joints, 355 Alodine, 529, 545
6061, 179 Alodine 1200, 329
96Sn4Ag solder alloy, 356 Alodine 1200 coating on Al-2024-T3, 135
Alodine nishes on common spacecraft aluminium alloys, 134
and for common spacecraft surfaces, 267
A / for a selection of surfaces and nishes, 267
AA 2017, 253 / values, 23
AA 2024-T81 (bar, rod), 539 Alpha-case embrittlement, 323
AA 2219, 33, 51, 103, 217, 225, 325, 545 Alphasat, 8
AA 2219-T81, 539 AlSi brazing ller metal, 181
AA 2618, 183 Alternate immersion test, 250
AA 7020, 325 Alulight, 202
AA 8090-T8771, 249 Aluminides, 535
Ablation, 46 Aluminium alloy, 19
Ablative material, 34, 36, 37, 46, 262, 502, 524. See also Re-entry Aluminium alloy 2219, 98
Ablebond 8175A, 415 Aluminium alloy cooling loop, 129
Accredited laboratories, 105 Aluminium Alloys Designations, 567
Acicular alpha, 126 Aluminium alloy temper designations, 567
Acoustic emission, 89 Aluminium-Beryllium Alloys, 288
Acoustic emission sensors, 94 Aluminium honeycomb to face-skin, 304
Acoustic emission signals, 93 Aluminium-lithium alloy, 36, 83, 273
Acoustic microscope, 88 Aluminium metallization, 469
Acoustic microscopy, 85 Aluminium metal matrix composites, 229
Active metal process, 402 Aluminium on carbon bre reinforced PEEK, 224
Active systems, 23 Aluminium-to-gold wire bond, 337
Additive layer manufacturing, 12 Aluminium whisker growth, 470
Adhesive, 50, 56, 125, 413, 507 Aluminized FEP Teflon, 322
Adhesive compounds, 533 AM-350, 318
Adhesive ller, 304 American Welding Society, 340
Adhesive lm, 295 Ammonium di-nitramide, 328
Adhesive tape, 42, 188, 284 Ammonium perchlorate, 328
Adhesive tape residues, 208 Amorphous silica, 290
Adhesive wear, 26 Analysis of surfaces, 99
Aeroshell, 46 Anodized beryllium, 529, 532
AF-E-332, 43, 290, 294 Anodized lm, 308
AgCu eutectic, 402 Anodizing, 134
Ageing tests, 339 Antenna face-skin, 223
Airborne salt, 139 Apogee-boost motors, 167
AISI 304, 98 AQ 60 I, 526
AISI 440C, 296 Aqueous cleaning systems, 212
AISI E52100, 296 ARALDIT, 539
Al-2195, 274 Araldite AV 138, 42
Al-2219, 132, 133, 137, 237 ArallR, 45

Springer International Publishing Switzerland 2016 655

B.D. Dunn, Materials and Processes, Springer Praxis Books,
DOI 10.1007/978-3-319-23362-8
656 Index

Area grid array (AGA), 122, 426 BETA cloth, 540

Area Grid Array packaging, 434 Bimetallic contacts, 18
Area Grid Array packages, 352 Bimetallic corrosion-related failures, 128
Argon ion milling, 70 Bi-stem deployment, 322
Argweld enclosure, 182 Bi-stem deployment booms, 322
Ariane IV, 31 Black anodized aluminium, 267
Ariane 5, 232, 328 Black anodizing, 270
Ariane 5 systems where white paint is applied, 238 Black arrow, 111
Ariane launchers, 325 Black chromium, 269
Ariane V, 32, 33 Black cobalt, 269
Ariane V half-fairing, 66 Black pad, 176, 358, 359
Asymptotic heating strategy, 150 Black patina, 360
Atmospheres for Brazing, 404 Black-anodized aluminium housing, 317
Atomic force microscope, 70 Black-anodized electrical connector, 308
Atomic oxygen, 168, 505, 522, 525, 540, 547, 554 Black-anodized nish, 311
Atomic oxygen on materials, 517 Black-anodized layer, 311
Atomic oxygen protective coatings, 505 Black-body calibration mechanism, 306, 307
Atomic oxygen testing, 97 Black-nickel plating, 269
Atomic oxygen with selected metals, 524 Blind bolt fasteners, 259
Au2Al, 335 Blow-hole, 160, 411
AuAl2, 334 Board distortion, 423
AU4GN, 30 Borosilicate glass, 35
Au80Sn20 braze alloy, 243 Brass turret terminal, 358
AuIn2, 348, 365, 366 Brass turret terminal pins, 359
AuSn4, 339, 360, 361 Braycoat, 539
Austenitic steels, 318 Brazeability, 400
Automatic electrical test equipments, 422 Braze alloy compositions, 400
Avcoat ablative material, 36 Braze alloy ller metals, 399
AZ31, 129, 306 Brazed joint, 341
AZ5GU, 30 Brazing, 51, 239, 245, 329, 399
Brazing alloy foil, 180
Brazing fluxes and their removal, 403
B Brazing furnace, 404
Bacteria, 68, 72 Brittle fracture, 247, 332
Bacterial and fungal growth, 42 Brittle-to-ductile fracture transition temps. in Pb-free and SnPb solders,
Baffle Cover Mechanism, 309 457
Baffle hinge, 311 Brominated flame retardants, 109
Bake-out times and temperatures for PCBs, 392 Brush alodined weld zones, 59
Balinite, 310 Brush debris, 506
Ball bonding, 332 Brush plating at NASA MSFC, 236
Barium-impregnated cathodes, 291 BS L93 alloy plate, 142
Barrier layer, 358 Burn-in procedures, 341
Beach marks, 271 Butt joint between copper and nickel, 206
Beagle 2, 48 Butt-welding process, 332
Bearing friction, 299
Bearing lifetimes, 304
Bearing materials, 148 C
Bearings coated with titanium carbide, 299 Cable cutter, 28
Bearings under vacuum, 299 Cadmium, 24, 108, 272, 327
Belleview spring fracture surface, 250 Cadmium fumes, 405
Belleville spring, 250 Cadmium plating, 255
Benzene-based solvents for cleaning, 210 Cadmium slivers, 313
Be-oxide layer, 530 CAF growth, 396
BepiColumbo, 25 Calomel electrode, 387
Berylliosis, 281 Capacitor, 352
Beryllium, 280, 283, 405 Capillary action, 341
Beryllium as a heat shield, 528 Capillary attraction, 399
Beryllium as-received parts, 285 Capillary gaps, 214
Beryllium foil, 286 Capillary management systems, 327
Beryllium foil with resulting mechanical properties and grain Capillary pump, 327
structures, 287 Capillary screens, 172
Beryllium machining, 281 Carbon bres, 36, 45, 164, 166, 168, 172, 247, 518, 532
Beryllium S-200C, 20 Carbon bre reinforced polymers, see CFRP
Beryllium sheet, 529 Coefcient of expansion, 557
Beryllium structures, 259 Carbon nanotubes (CNT), 44, 46
Beryllium thermal protection, 526 Carbon nanotube structure, 44
Index 657

Carboncarbon composites, 170, 531 Cold welding of mechanisms, 304

Carboncarbon decelerator, 526 Cold welding of stranded wires, 380
Carbon-epoxy, 36 Colinal, 269
Carbon bre mesh, 44 Collected volatile condensable material, 40
Carbonsilicon carbide composites, 531 Colophony fumes, 398
Cassini, 526, 528 Column grid array interconnections, 440
Catalyst deactivation, 290 Column grid arrays, 426
Catalyst particles for hydrazine decomposition, 288 Combustion chamber, 33
Catalyst particles, 289, 290 Commercial off-the-shelf components (COTS), 63
Catalytic bed thruster motors, 512 Communication satellite systems, 7
Cathode emitter degradation, 291 Company cleaning plan, 210
Cathode emitter degraded by sintering of porous tungsten, 298 Comparison Tables (Alloys), 571
Cathode emitter surface, 293, 296 Compatibility of Liquid and Solid Propellants with Components and
Cavitation at the inclusion-to-matrix interface, 265 Subsystems, 326
Ceramic capacitors, 410 Compatibility testing, 210
Ceramic chip capacitor with internal voiding, 88 Compatible coupling, 17
Ceramic matrix composite fasteners, 535 Component cracking, 424
Ceramic matrix composites, 534 Component part selection, 66
Ceria doped micro-sheets, 28 Component part selection, and procurement, 61
Cesium, 24 Composite contact rivets, 165
CFRP, 11, 26, 36, 91, 166, 293, 505 Computer tomographic scan, 438
CFRP delaminations and fractures, 295 Computerized X-ray tomography, 88
CFRP face-skins with an aluminium honeycomb, 293 CONATHANE, 432
CFRP waveguides, 168 Condensation of outgassing products, 505
Charcoal black nish, 227 Condensation rates, 273
Chemglaze, 547 Condensed cadmium contaminant on surface of painted shroud, 275
CHEMGLAZE, 539 Condensed moisture, 129
Chemglaze Z306 black paint, 275 Condensed organic contamination, 259
Chemical analysis, 71 Conductive adhesives, 425
Chemical conversion, 329 Conductive Anodic Filament (CAF), 393, 394
Chemical conversion coating for magnesium alloys, 129 Conductive coatings, 238
Chemical conversion coatings, 134 Conductive silicone adhesive, 271
Chemical (elemental) content of a typical spacecraft electronic box, 112 Conductor track failure, 428
Chemical stripping, 496 Conformal coating, 393, 428, 432, 540
Chromate conversion coating, 129 Connector bodies, 313
Chromate conversion coating on cadmium-plated steel, 275 Connector-to-coax assemblies, 356
Chromate primers, 107 Contact devices, 164
Chromium and nickel sublimation, 276 Contaminant particles, 309, 549
Circuit design, 423 Contamination, 65
Circumferential in-place pipe welding, 196 Contamination of Invar moulding tool, 312
Cleaning efciency, 391 Controlled atmosphere brazing, 181
Cleaning method, 275 Conversion table for mechanical properties, 565
Cleaning of flux-contaminated surfaces, 389 Co-planarity problems, 423
Cleaning of flux residues, 351 Copper comb patterns, 392
Cleaning of individual parts, 210 Copper-palladium alloy, 165
Cleaning of metallurgically joined assemblies, 212 Copper ribbon column fracture, 440
Cleaning processes associated with spacecraft mechanical systems, 207 Coppersilver eutectic preforms, 132
Cleaning silicone contaminated surfaces, 219 Copper sulphate test for ferrite, 138, 318
Cleanliness, 216, 218 Coppertin intermetallic layer, 149
Clean-room practice, 313 Copper-to-enamel interface, 370
Cleavage cracks, 248 Copper-to-silver-plating interface, 358
Coatings and conversion coatings, 268 Cork, 37, 4648, 524
Coatings for soldering applications, 357 Cork TPS tiles, 48
Coaxial cable assembly, 356 Corona, 27, 390, 442, 445, 504
Coefcient of expansion, 557 Corona discharge, 444
Coefcient of (linear) thermal expansion, 349, 401, 557 Corona effects, 371
Coil spring, 309 Corrosion adjacent to dip brazed llet, 179
Coin or tap test, 83 Corrosion of stored spacecraft electronic components, 40
Cold-drawn springs, 318 Corrosion potential, 17
Cold pressure weld, 299 Corrosion potential of metals, 14
Cold sprayed coatings, 223 Corrosion prevention, 17
Cold-weld, 339, 549 Corrosion product, 541, 546
Cold weld database summary tables, 307 Corrosion testing, 75
Cold-welded particle, 309 Corrosion, 390, 540
Cold welding, 25, 26, 83, 120, 299, 305, 354, 381 Corrosion-resistant fastener materials, 253
Cold welding due to cyclic, impact loading, 306 Cosmetic defect, 119, 411
658 Index

Cosmetics of solder llets, 410 Diamond grit metal matrix composite, 222
Cosmic ray detectors, 451 Diamond pyramid hardness impressions, 336
Counterfeit fasteners, 253 Diaphragm, 38, 288
Cracked barrel, 120 Dichloromethane, 211
Cracked ceramic chip capacitor due to vibration, 421 Dicronite, 310
Cracked leads on a thick-lm carrier package, 174 Dielectric breakdown, 359
Cracking of glass-to-metal seals, 410 Dielectric properties, 392
Crack initiation, 349 Differential heating, 509
Crack propagation rate, 27 Diffusion bonding, 197, 203, 206
Cracks in CFRP, 300 Diffusion brazing, 399
Cracks in Inconel heater housing, 278 Diffusion of zinc, 359
Cranes, 63 Diffusion soldering/brazing, 408
Crimpability, 370 Diffusion soldering process, 409
Crimping tools, 339 Diffusion welding, 105
Crimp joint, 337 DIGESIL, 315
Crimp-termination characteristics, 343 Dimensional stability, 170
Critical design review, 22 Dip brazing of aluminium alloys, 179
Critical processes, 621 Dip brazing, 181, 399, 405
Cross section polisher, 337 Dipole connection, 362
Cross-section through a solar array, 190 Disassembly of orbiting space hardware by astronauts, 305
Cryocon, 198 Disposable mandrels, 177
Cryot, 198, 201 Dissimilar FSW, 233
Cryogenic bearing, 229 DNA, 70
Cryogenic propellants, 29 Dog bone tensile sample, 47
Cryogenic temperatures, 369, 455 DOW 17, 129
Cryogenic temperature materials, 26, 43, 149, 220, 229, 318, 320, 329, Dry-heat ageing, 371
369, 379, 454, 490 Ductile dimple fracture, 249
Cryostats, 27 Ductile intermetallic compounds, 365
Crystallographic planes, 98 Dye penetrant, 442
C-SAM, 439 Dye penetrant testing, 58, 139, 208, 255, 323, 408
Cu2O, 370 Dye penetrant test method, 439
Cu3Sn, 361, 372, 376, 412 Dynamic outgassing testing, 97
Cu6Sn5, 330, 361, 364, 412, 413, 474
CuA12, 131
CuAl2, 133 E
CubeSat, 11, 45, 63, 615 Earths magnetic eld, 508
Curie temperature, 318 EB brazed joints with Au80Sn20, 242
Custom 455, 146 Ebonol black, 269
CV 11400, 431 EB-welded 2219-T851, 141
CV 1144-0, 432 EB welding machine for reflow brazing, 239
ECCOBON, 539
ECCOFOAM, 539
D Eccoshield tape, 196
D6AC, 34 ECM failures, 394
DC93-500 space grade, 42 ECSS-Q-ST-70-02, 38
Debond and fracture of CFRP, 299 ECSS-Q-ST-70-38, 353
Debris, 509, 512 EG8050HC, 415
Declared Materials List (DML), 10, 58, 116, 262, 309, 625 Elastomer type AF-E-332, 288
Declared Process List (DPL), 58, 419, 621 Electrical bonds, 329
Defective solid rocket motor case, 28 Electrical conductive adhesives (ECAs), 413
Defects in titanium piece-parts, 323 Electrical feedthrough, 444
De-golding by immersion, 360 Electrical grounding, 17, 223, 235, 329
Delamination, 84, 303, 392 Electrical interconnections, 329
Delrin, 505 Electrical open-circuit, 320
Delta ferrite, 315 Electrical resistance testing at room and cryogenic temperature, 455
Dendrites, 395 Electrical resistance weld, 330, 331
Densimet, 306 Electrical resistivity, 339, 458
Deployable nitinol strut, 200 Electrochemical migration (ECM), 359, 360, 392, 393, 393, 462, 540
Dermatitis, 398 Electrode housing, 240
Design margins, 11 Electrode housing materials, 242
Destructive physical analysis, 57 Electro-etch cleaning, 237
Dewetted area of pad, 152 Electro-explosive devices, 264
Dewetting of pad, 153 Electroforming processes, 176
Dewetting on areas to be soldered, 150 Electroless nickel deposits, 173
Dewpoint, 405 Electroless nickel plating, 318
Diamond, 220 Electroless nickel plating of aluminium electronic housings, 175
Index 659

Electromagnetic emission, 196 Fibre-reinforced glass ceramics, 170

Electromagnetic emission from TIG welding equipment, 195 Fibre-reinforced plastic composites, 166
Electromigration, 468, 470 Fill and drain nozzle, 198
Electron beam weld, 132, 184 Fingerprint greases, 127
Electron beam welding, 51, 184, 185 Finishes for titanium and its alloys, 310
Electron-beam-welded titanium alloy, 185 First-aid equipment, 405
Electronic box, 62, 88, 110 Five-stage model for whisker growth, 483
Electronic circuitry, 329 Flammability, 42, 540
Electronic housings, 329 Flammability hazard, 370
Electronic package, 427 Flat-packs, 351
Electroplated nickel, 174 Flatwise tensile tests, 296
Electrostatic discharge, 418 Flawed primary mirror, 11, 60
Elemental analysis, 97 Flexible circuits, 159
Ellingham diagrams, 400, 613 Flexible second surface mirrors, 271
Emal Technology, 370 Flexible waveguide, 121
Embrittlement of copper, 127 Flight harness materials, 550
Embrittlement of titanium alloys, 255 Floating grains, 349
Enamel-coated copper, 370 Fluorescent penetrant inspection, 262
Energy conversion element, 200 Fluorides, 405
Energy-dispersive x-ray analyser, 71 Fluorinated ethylene propylene, 159
Engineering drawing, 116 Fluorine attack, 375
Environmental conditions, 20 Flux residue, 384, 391, 546
Epotek E2116, 415 Flux types for engineering metals, 387
Epoxy smearing, 159 Flux-corrosion of silver-plated stranded wires, 383
Epoxy top-coat, 394 Foamed aluminium for damping, 202
ePTFE, 46 Focused ion beam (FIB) microscope, 70
ERS-1 spacecraft, 167 Fourier transformation infrared (FT-IR) spectrometers, 100
Etchants, 561 Four-point bending test, 76
Etching of metals, 561 Fracture at cryogenic temperatures, 454
Etching solutions for beryllium, 285 Fracture locations in coating, 237
Ethical issues, 104 Fracture mechanics testing, 76
Eureca, 520 Fracture surface of a circular metallized (TiPdAg) contact pad, 271
European retrievable carrier (EURECA), 503 Fracture toughness, 83
Eutectic tinlead solder alloy, 341 Fretting, 83
EVA joining/cutting activities, 50 Fretting test, 229
Evaluation of solderability, 372 Friction, 25
Evaporation rates, 272 Friction stir joining and welding, 12, 34, 36, 52, 189, 231, 234
Examination of fracture surfaces, 247 Friction stud welding, 234
Exhaust plume, 328 FTIR, 391
Expendable launch vehicles, 28 FT-IR analysis, 317
Explosively welded transition ring, 187 Fuel lines, 324
Explosive welding, 186 Fusion welding, 182
Futuristic ideas, 11

F
Failed ABM case, 264 G
Failed ball-bearing, 304 Galileo spacecraft, 308
Failed component lead, 384 Galliumpalladiumsilver braze alloy, 403
Failed rhenium tube from electrothermal thruster, 283 Galvanic compatibility, 388
Failed spacecraft antenna, 293 Galvanic copper corrosion, 378
Failed video camera electronic circuit, 541 Galvanic corrosion, 386, 389
Failure investigation, 99 Galvanic corrosion of fasteners, 257
Failure mechanism associated with surface-mounted devices, 425 Gamma-ray detectors, 286
Failure mode analysis, 57 Gamma-rays, 83
Failure of RF cables connected by SMT, 428 Gamma-TiAl alloys, 196
Failure review board (FRB), 21, 102 Gapasil brazing alloy, 273
Failures due to board flatness problems, 422 Gas-tight joint, 120
Faraday shielding, 37 Gas-tightness, 339, 340
Fastener failure due to forging defect, 254 Gas-tightness test, 338
Fastener manufacturers, 253 GCMS, 40
Fastener specications, 255 Gecko biomimetic adhesive tape, 45
Fasteners, 251, 533 Girth weld, 68
Fatigue life, 367 GlareR, 45
Fatigue striations, 248 Glass to metal seal, 467
Fatigue tests, 78 Glass-rich oxide whiskers, 467
Ferromagnetic materials, 317 Glossary, 629
660 Index

Gold, 110 Hillocks, 469, 470

Goldaluminium system, 332 Hi-lok TM fasteners, 34
Gold-embrittled solder joint, 363 Hinged-tube, 198
Gold embrittlement, 360 Hipparcos, 308
Gold embrittlement of solder, 362 Holddown and release unit, 308
Gold in solidied solder, 361 Hold-down button, 554
Gold-plated conductor material, 363 Hold-down points, 229
Gold-plated dipoles, 361 Hole-drilling strain-gauge method, 94
Gold-plated surfaces, 360 Holographic interface bond tester, 294
Gold removal, 348 Holographic interface tester, 85
Gold-rich intermetallic phases, 334 Holographic interference inspection of failed antenna, 301
Gold-tin binary phase diagram, 242 Hot-air-levelled coatings, 160
Gold wires, 332 Hot-dipped galvanization, 139
Grain boundary embrittlement, 127 Hot oil fusing of tinlead, 148
Grain boundary embrittlement in beryllium, 286 Hot plate method, 161
Grain growth, 332 Hot-pressed beryllium, 259
Grain growth and internal cavitation, 280 hot-pressure mounting of samples, 343
Graphine, 43, 44, 220 HST solar array, 506
Graphite bre thermal strap assemblies, 221 Hubble Space Telescope, 11, 50, 60, 190, 321, 322, 505
Graphite lubricants, 258 Human activities on the Moon, 502
Graphitization treatment, 166 Human contaminants, 100
Greases, 311 Human error, 115
Green chromated, 274 Humid environment (moisture), 170
Green CuCl corrosion, 133 Huygens probe, 527
Greener spacecraft, 106 Hybrid packages, 415
Green plague, 375, 386, 542 Hydrazine, 327
Green propellant, 109, 327, 328 Hydrazine contamination levels, 295
Grifth cracks, 324 Hydrazine (N2H4), 30
Ground activities, 20 Hydrazine propulsion tank, 68
Ground handling, 65 Hydrazine tank, 288
Ground-handling facilities, 63 Hydrazine tank diaphragm, 294
Guianese Space Centre, 325 Hydrazinium nitroformate, 328
Hydrogen bake-out, 124
Hydrogen embrittlement, 122, 255
H Hydrogen embrittlement of spring steel, 123
Habitable structure, 12 Hydrogen embrittlement of steel fasteners, 255
Hairline cracks, 184 Hydrogen embrittlement relief, 176
Hard-anodised layer on AA7075 alloy, 225 Hydroxylammonium nitrate, 328
Hard anodizing of Al-7075 alloy, 225 Hygroscopic contaminants, 394
Hard anodizing treatments, 129
Hard chromium, 27
Hardness testing, 73 I
Harness, 319 Identication of leak paths, 279
Health Hazards in the Electronic Assembly Area, 398 Impact crater, 555
Heat Affected Zone (HAZ), 126, 184, 185 Impact feature, 520
Heat-affected zone of laser welds, 186 Impact records, 517
Heat exchanger, 546 Impact test facility, 84
Heat shield, 34 Impact testing, 310
Heat shield materials, 524, 531 Impact/fretting test equipment, 307
Heat transfer in vacuum, 221 Inconel 600, 265
Heater lament, 90 Inconel 718, 20, 27, 126, 147, 253, 319
Heater investigation, 277 INCONEL 718-PH, 539
Heater sublimation problem associated with thruster motor, 276 Inconel alloys, 318
Heat-shrinkable sleeves containing solder preforms, 381 Indium solder alloys, 363
Hemispherical emittance, n, of the anodized, 531 Indiumlead soldered to various gold interfaces, 365
Hermes spaceplane, 532 Indiumtin oxide, 271
Hermetically sealed assemblies, 399 Industrial placements, 103
High-absorption surfaces, 269 Infrared Space Observatory, 27
High-denition radiography, 86 Infrared Space Observatory cryostat, 320
High-denition X-radiography, 276 Infrared spectroscopy, 101
High-performance fasteners, 262 Inorganic glasses, 507
High-precision bearings, 296 Inspection, 434
High-temperature brazing, 404 Inspection criteria for brazed joints, 407
High-temperature fasteners, 533 Inspection criteria, 406
High temperature rating, 370 Insulation materials, 369
High voltage interconnections, 442 Insulation stripper, 337
Index 661

Interference t, 259, 337 Leaking detector, 286, 288

Intergranular cracking attributed to stress corrosion, 249 Leaking heaters, 276, 277
Intergranular cracks, 184 Leaking kerosene cans, 448
Intermetallic compounds, 361 Leaking lead glass seals, 468
Intermetallics, 331 Leaking lower seal joint, 61
International Space Station (ISS), 42, 236, 501, 536 Leaking tanks, 325
Invar, 314, 319 Leak tests, 317
Ion chromatography, 391 Liberator TM, 49
Ionic contamination levels, 391 Lifting gear, 63
IPC J-STD-001, 353 Lightening, 139
Iridium catalyst, 290 Lightning, 37
Ironnickelcopper alloy, 318 Light pollution, 107
ISO 9001, 56 Limited shelf life, 56
Isopropyl alcohol (IPA), 212, 390 Limpet teeth, 45
Limpet tooth material tooth, 47
Liquid and gas chromatographic techniques, 97
J Liquid crystal polymers (LCPs), 48
James Webb Telescope, 229, 509 Liquid helium, 312
J-leads, 424, 425 Liquid helium cryostate, 319
Joining metals to thermoplastics, 12 Liquid helium temperature, 451
Jupiter space probe, 285 Liquid metal embrittlement (LME), 139
Liquid penetrant tests, 82
Liquid phase inltration, 532
K Liquid propellants, 326
Kapton, 518, 523, 550 LISA Pathnder mission, 239
Kapton lms, 271 Lock-nuts, 261
Kapton thermal blanket, 548 Low Earth Orbit (LEO), 501, 518
Keronite coating, 227 Low-emissivity surfaces, 268
Keronite PEO, 225 Low-expansion materials, 163
Kevlar, 63, 516 Low-temperature magnetic properties, 319
Kevlar-49 reinforced plastic, 166, 169 Low-voltage applications, 340
Kirkendall voiding, 335 Lubricants, 42, 503
Kovar, 132, 174, 330, 363, 357, 423, 467 Lubricants suitable for use under high vacuum, 258
Kovar lead material, 173, 348 Lubricating oils with low outgassing properties, 209
Kovar leads, 383 Lubrication, 21
Lunar soil, 13

L
Laboratories, 56 M
Laboratory notebooks, 100 MAPSIL 213, 432
Laboratory records, 100 Macroscopic examination, 67
Laminar flow bench, 81 MAGE apogee boost motors, 90
Laminography, 88 MAGE motor, 169
Lap welds, 332 Magnesium, 20, 24
Large diameter stranded wires, 340 Magnesium alloys, 129
Laser annealing, 176 Magnesiumlithium alloys, 273
Laser beam welding, 185 Magnet, 318
Laser welding, 52 Magnetic cleanliness, 175
Laser-welded spacecraft axle shaft, 186 Magnetic coercivity, 175
Laser-welded Ti6A14V, 186 Magnetic eld generated by a spacecraft, 317
Launch, 20 magnetic moment, 370
Launch and operations readiness review, 22 Magnetic permeability, 318, 319
Launch site exposure and corrosion, 138 Magnetic problems, 317
Launch vehicle connector, 311 Magnetic shield materia, 318
Launch vehicle release gear mechanism, 314 Magnetically clean, 62
LDEF, 503, 516, 517, 520 Magnetometers, 317
Leaching of silica, 291 MAGNOLYA chemical surface treatment, 130
Lead coatings, 505 Mandrel materials for electroforming, 176
Lead-free control plans, 494 Manganin wire, 379
Lead-free solder alloys, 341 Manned compartments, 535
Leadless ceramic chip carriers, 355, 360 Manned spacecraft, 432
Leadless chip carrier, 431 Manned volumes, 538
Leadless surface-mounted devices, 351 Manual tungsten inert gas (TIG) arc welding, 181
Leaking battery cell, 194 Manual welding, 182
Leaking cartridge, 265 Manufacturing processes, 116
Leaking ceramic-to-metal seal, 131 MAPSIL 213, 42, 431
662 Index

Maraging steels, 30 Micro-VCM test equipment, 41

Martensitic high-strength alloys, 138 Microwave horn, 176, 177
Martensitic stress induced transformation, 318 Microwelding, 182
Martensitic transformation, 202 Microweldments, 330
Material group numbers, 617 Migration of silver sulphide tarnish, 269
Material review boards, 408 Mir space station, 52, 537
Materials and processes, 58 MISSE, 503
Materials and processes standards related to space, 619 Model philosophy, 21
Materials engineer, 21 Modern assembly room, 341
Materials laboratory, 64 Modern bearings, 299
Materials making up soldered joints, 349 Modication, 409
Materials review board, 102 Moisture ingress, 392
Mean time before failure, 469 Moisture pick-up, 26
Mechanical electrical connections, 337 Molybdenum, 240, 277, 280, 462
Mechanical fastener tool, 50 Molybdenum disulphide, 25, 305, 506, 518
Mechanical nishes, 134 Molybdenumtitanium phase diagram, 244
Mechanical parts and process controls, 616 Molybdenum whiskers, 462, 464
Mechanical properties MoMn metallization system, 360
AA 8090 and other Al-Li alloys, 102, 143 Mono-methyl hydrazine (MMH) and nitrogen tetroxide, 327
AA 2219 various heat treatments, 102, 141 Monopropellant hydrazine thrusters, 276
beryllium, 287 Monopropellants, 328
BS L93, 142 Moon rock, 502
electronic materials at different temperatures, 368, 451, 458 Moon-rock analyses, 501
fastener materials, 253 Moores law, 461
spring materials, 146 Morphology of wear particles generated from sliding contacts, 165
tin whiskers, 487 MoS2, 217
Mechanical properties conversion table, 565 MoSTTM, 257
Mechanical properties of near-eutectic tinlead alloys, 451 Motor exhaust plume, 514
Mechanical shock, 422 Motor thrust frame, 34, 232
Mechanical testing, 73 Moulding tools, 312
Mechanical-type strippers, 338 Mounting of chip parts, 416
Mechanism and the grounding plane, 305 MP 35N, 147
MEDET instruments, 231 Multilayer board internal connections, 155
MEMS, 11 Multilayer boards with high heat capacity, 161
Metal alloy comparison tables, 571, 611 Multilayer ceramic capacitors, 351
Metal matrix composites for spacecraft pressure vessels, 172 Multi-layer insulation blankets, 28
Metal migration, 164 Multilayer PCB, 84
Metal oxide precipitation in glass seal., 467 Multiphase MP35 N, 253
Metal oxide whisker, 466 Mumetal, 318
Metallic contamination particles, 312 Mutually soluble in the solid state, 165
Metallic fragment, 314 Mylar lms, 271
Metallic particle generation, 258
Metallic particles, 309
Metallographic control, 330 N
Metallographic examination, 343 Nano-ceramic technology, 222
Metallographic presentations, 102 Nanocomposites, 414
Metallography, 97 Nano-grained alumina, 220
Metallurgical joint, 358 Narloy-Z, 32, 33
Metallurgical laboratory practice, 69 Natural rosin, 391
Metallurgical reaction, 331 Neutron diffraction method, 94
Metal-matrix composite, 427, 161 Neutron radiographs, 92
Micro-arc oxidation, 224 Neutron radiography, 86
Microcracked electroless nickel, 173 Nextel ceramic cloth, 516
Microcracked thin-foil detector windows, 286 Ni3P, 176
Microcracking of diode lead surface, 173 Ni3Ta, 332, 333
Microfocus x-radiography, 91 Nichrome heater coil, 278
Micrographs of tin-plated strands, 375 Nickel alloys, 19
Micrometeoroid, 509, 512, 555 Nickel coatings on Ti6Al4V, 235
Micrometeoroid impact, 506, 518 Nickel nishes, 370
Microscopic examination, 67 Nickel phosphide (Ni3P), 174
Microspheres, 36 Nickel phosphide precipitates, 358
Microstructure of milled and stress-relief-treated beryllium, 284 Nickel ribbon, 331
Microstructure of solder alloys, 343 Nickel sulphide process, 308, 317
Microtome, 70, 292 Nickelcadmium battery cell, 117, 131
Microtome section of life-tested cathode, 297 Nickel-clad conductive LCP bres, 49
Micro-VCM test, 97 Nickel-plated copper braid, 49
Index 663

Nickel-to-copper brazed joint, 402 PEEK, 308

Nickel-to-nickel electrical resistance microwelds, 332 PEEK composite, 223
Nickel-to-nickel welded electronic circuits, 330 Peel strength versus ageing time, 368
Nickel-to-phosphorus ratio, 175 Phase diagram, 25
Nickel-to-titanium alloy brazed joint, 403 Philips Globule method, 371
Niobium, 535 Phosphorous-rich surface, 359
Nitinol, 197, 200, 201 Pinch-off for tube sealing, 317
Nitinol-deployed lattice mast, 199 Pinch-off seals, 317
Nitrided stainless steel, 138 Pinhole, 121, 129
Nitrogen tetroxide (N2O4), 30, 327 Pitting and leakage of an aluminium cooling channel, 131
Nomex cloth, 548 Pitting corrosion, 328
Non-captive nut, 356 Plagues, see Green, Purple, Red and White plagues
Non-conformance board, 102 Plasma cleaning technology, 212
Non-conformance reports, 544 Plasma electrolytic oxidation (PEO), 224, 225, 229
Nonconforming fasteners, 262 Plasma electrolytic oxidation treatment, 308
Noncoplanarity, 423 Plated nish on copper conductors, 369
Nondestructive testing, 82 Plated-through hole, 152, 410
Non-metallic materials, 38 Platinum grain growth, 281
Platinum group metals, 534
Platinum ribbon, 277
O Platinum ribbon grain boundary, 277
Oddy test method, 40 Polyacrylonitrile, 166
Odour, 540 Polybutaidene acrylonitrile, 30
Offgassing, 40 Polyglycidylazide, 328
Open circuit, 358, 390, 431 Polyvinyl chloride (PVC), 369
Open-circuit failures, 320 Polyxylene, 429
Optical bres, 42 Porosity during a solderability test, 160
Optical materials, 42 Porosity in weld bead, 118
Optical microscopy, 67 Porous gold plating, 269
Optical properties (/ ratio) before and after testing, 268 Positive air pressure, 344
Optical solar reflectors (OSR), 270 Post-flight inspection, 190
Optical spectroscopy, 97 Post-flight materials, 501
Orbital test satellite (OTS), 7, 28 Post-flight observations, 501
Organic chemistry, 97 Post-flight tribological assessment of the Hubble Space Telescope solar
Organic fastener lubrication systems, 257 array mechanisms, 303
Organic materials, 503 Potting compounds, 42
O-ring, 61 Power cycling, 352
O-ring seals, 43 Power (I2R) loss, 370
Orion crew module pressure vessel, 274 Power system weldments, 189
Orion spacecraft, 36 Precipitation hardening, 138, 251, 252
OTS-2, 9 Precipitation-hardening stainless steels, 403
Outgassing, 38, 540 Preconditioning, 357
Outgassing data for flux residues, 390 Preferred materials for short-term evaluation, 539
Oven bake-outs, 423 Preliminary design review, 22
Over-ageing, 251 Pressure vessel steel, 93
Oxyacetylene gas welding, 182 Primer coating, 274
Ozone-depleting chemicals, 107 Printed circuit board assemblies, 559
Printed circuit boards, 148
Printed-circuit-board (PCB) evaluations, 79
P Problems associated with brazing, 399
Paints, 42 Process documents, 405
Parameters for bake out, 393 Process identication document, 420, 422
Particle radiation, 23 Process identication documentation, 418
Particle size, 72 Produce assurance applied to brazing operations, 405
Particles generated from spacecraft fasteners, 261 Product assurance management, 55
Particles of beryllium, 262 Project review boards, 21
Particulate contamination, 398 Propellant, 29
Parting compound, 314 Propellant Management Device (PMD), 172
Paschen-like curves, 446, 447 Propellant motors, 30
Passivation treatment, 138 Propellant tanks for the Ariane-5 launch vehicle., 205
Passive systems, 23 Propellant tanks manufactured from Ti6Al4V, 171
Passive thermal control systems, 266 Properties of fluxes, 396
Pathnder mission, 200 Properties of printed circuit laminates, 559
Payload-support structures, 167 Properties of tin whiskers, 485
Pcb nishes, 148 Protection shields, 515
PCB laminates properties, 557 PTFE, 25, 43, 49, 175, 188, 211, 258, 305, 310, 315, 353, 356
664 Index

Pull-off strength of coatings, 238 Rosetta, 28

Pulsed laser to repair solar cell interconnection welds, 186 Rotary dip method, 153
Purple Plague, 334336, 332 Rotating weld pin during FSW, 232
Pyrel foam, 552 RTV 560, 35, 533
Pyrex glass, 319 RTV 566 silicone, 42
Pyrotechnic, 92 Rubber diaphragm, 290
Pyrotechnic actuator device, 422 Rubbers, 42
Pyrotechnic cutter, 265

S
Q SADM off-load device, 202
Quad flat packages, 422 Salinity maps, 139
Qualication review, 22 Salt spray corrosion tests, 308
Quality assurance, 55, 66 Salt-spray cabinet, 80
Quality assurance controls for fasteners, 261 Sapphire crystals, 245
Quartz optical solar reflector, 271 SAX spacecraft, 187
Scanning electron microscope (SEM), 68, 71, 247
Scanning laser acoustic microscope (SLAM), 85, 87
R SCC evaluation, 140
Radiation effects, 507 Scotchcast, 431
Radiation testing, 63 SCOTCHCAST, 432
Radiation, 25, 62, 547 SCOTCH-WELD, 539
Radiography, 83 Screen out magnetic items, 318
Random vibration, 420 Sea coast corrosion, 82
Rapid protyping, 12 Sealing of glass and beryllium windows, 369
REACH, 42, 106 SEAMS introduced during rolling, 256
Re2O7, 278 Second phase precipitates, 251
Recovered mass loss, 40 Secondary ion mass spectrometry, 99
Recuperation of unsolderable PCBs and component leads, 413 Selection of materials and processes, 10
Recycling, 106, 108 Selective brush electroplating, 234
Recycling electronic waste, 110 Selective brush plating, 235
Red plague, 370, 375, 376, 377, 378 Self-healing materials, 45
Re-entry, 520, 536 Semi-rigid cables, 341
Re-entry vehicles, 170 Semi-rigid cable solder joint fracture due to gold embrittlement, 362
Reflow of capacitor solder, 158 Semi-rigid RF cables, 428
Refractory metals, 534 Shape-memory alloys, 197
Regolith, 12, 13, 502, 503 Shape memory polymers, 44
Reinforced carbon/carbon, 35 Sharp llets, 444
Release triggers made of Nitinol, 199 Shell 405 catalyst, 289
Reliability and safety, 57, 66 Shielding gases, 127
Removal of silicone polymers, 314 Shock loading, 458
Removal of work-hardened layers, 285 Shock wave, 265
Repair, 409, 432 Short circuit, 313, 360, 395, 486
Repair and modication of assemblies, 341 Shuttle tile, 35, 533
Reprocessing pure tin terminations, 495 Silica glass microspheres, 46
Residual fluxes on spacecraft PCBs, 393 SiC bres embedded in a matrix of Ti6A14V, 162
Residual stress, 93, 98 SiC monolaments in an aluminium matrix, 171
Residual stress measurements, 477 SiC monolaments within an AA 2014 matrix, 164
Residual Stresses in Weldments, 195 SILGEST, 315
Resistance measurements, 329 Silicide-coated fasteners, 536
Resistance pressure welding, 330 Silicon carbide bre reinforced metal matrix composites, 37
Resistance spot welding, 116, 329 Silicon carbide bre reinforced titanium alloy matrix, 39
Resistance to thermal cycling environment, 432 Silicone contaminants, 317
Resistance-welded silvermesh interconnector, 191 Silicone contamination, 219, 310, 505
Resonance, 422 Silicone oil, 311, 418
Reusable tanks, 183 Silicone oil contaminant, 317
Rework, 409 Silicone products, 219, 311, 505
Rework and repair of AGAs, 441 Silver coatings, 360
Rework of soldered joints, 408 Silver nishes, 370
Rework on composition of joint, 412 Silver mesh, 525
Reworking of spacecraft assemblies, 410 Silver migration, 359
Rhenium, 278 Silver wire, 331
River patterns, 248 Silver-coated molybdenum interconnector weld, 193
Rivet compositions, 253 Silver-lled thermosetting epoxy resins, 413
Rocket motor nozzles, 170 Silver-graphine, 414
RoHS, 107, 494 Silver-loaded epoxy, 415
Index 665

Silver-plated conductors, 371 Space Shuttle External tanks, 231

Silver-plated solid copper wire, 337 Space tribology, 305
Silver-plated steel stiff-nuts, 261 Spalling, 270
Silver-plated wire, 339 Specimens made from beryllium, 281
Silver-plated wire strands after solderability testing, 373 Spectroscopic methods, 72
Sine vibration, 420 SPELDA, 94
Skin effects for RF transmissions, 370 Spliced wire joints, 551
Skin-flake, 100 Spot-welding, 179
Skin secretions, 100 Spring clip, 337
Slip rings, 23, 311 Spring materials, 144
SMT solder joint failure due to Conformal Coatings, 428 Stablecore, 353
SMT verication sample, 430 Staking compounds, 422, 426
Sodium polysulde test, 377 Standard free energy of formation of oxides with temperature, 613
Solamide 301 foam, 552 Standard reference electrode, 389
Solar absorbers, 267 Standards for soldering spacecraft electronics, 342
Solar absorptance, 270 Standards related to space, 619
Solar array, 358 Stand-off height, 351, 353, 391
Solar array deployment/retraction system, 322 Steam ageing, 371
Solar-Array Drive Mechanism (SADM), 199 Steel alloys, 19
Solar blankets, 547 Steel alloys generally considered suitable for spring manufacture, 146
Solar reflectors, 267 Steel wires, 320
Solder assembly facility, 344 Steel wire-to-nickel tube welding operatio, 320
Solder column, 426, 439, 441 Stress corrosion, 327
Solder copper wire joint failure, 364 Stress-corrosion crack in the tank wall, 326
Solder dipping, 497 Stress-corrosion cracking (SCC), 19, 139, 325, 384, 540
Solder llet, 352, 443 Stress-corrosion failure, 249
Solder flux, 540 Stress corrosion of component lead material, 383
Solder flux vapour, 339 Stress-corrosion tests, 75
Solder joint repair, 540 Stress-induced deformations, 319
Solder paste, 436 Stress-raising defect, 254
Solder Sleeves, 551 Stress-relaxation by thermal gradients, 319
Solder sphere, 511 Stress-relief bend, 412
Solderability, 155, 371, 386 Structural panels made from high-temperature titanium alloys, 207
Soldered interconnections, 340 Students, 103
Soldering, 329, 340 Sublimation, 24, 272, 283, 508
Soldering fluxes, 380 Sublimation of aluminium alloys, 181
Soldering parameters, 341 Sublimation of and condensation of cadmium and zinc, 274
Solder-plated lead wire, 330 Sublimation of klystron cathode-heaters, 276
Solders, 108 Sublimation of Rhenium, 278
Sold propellants, 328 Sublimation rate against temperature for Nichrome, 280
Solid lubricants, 26 Sublimation tests, 273
Solid solution, 25 Subsurface structures, 283
Solid-state diffusion of platinum, 277 Sulphur hexafluoride (SF6), 317
solid-state interdiffusion, 333 Sun sensor experimental baffle, 230
Solithane, 509, 510 Sun-sensor, 63, 64, 209
SOLITHANE, 432 Superconducting at 4.2 K, 458
Solithane 113, 454 Superconductive SnPb solders, 457, 458
Solvent resistance, 432 Superplastic forming, 196, 203, 206
Sound absorption coefcient of different foamed aluminium, 203 Surface analysis, 96
Sources of Failure, 115 Surface colourations, 127
South Atlantic Anomaly, 508 Surface-corrosion residue, 216
Space-approved greases, 257 Surface-diffusion layers, 323
Spacecraft antennae, 312 Surface electrical grounding, 329
Spacecraft antenna face-skins, 167 Surface insulation resistance (SIR), 392, 395
Spacecraft charging, 27 Surface insulation resistance testing, 391
Spacecraft detectors, 187 Surface mount technology, 341, 419
Spacecraft failures, 61 Surface of an FR-4 PCB board laminate, 392
Space environment, 22, 23 Surface protection treatments for aluminium alloys, 134
Space environment effects, 504 Surface-tension tanks, 172
Spacelab, 27, 42, 339, 501 Sustained stresses, 145
Spacelab post-flight hardware, 542 SYLGARD 184, 432
Spacelab processing and integration, 543
Spacelab-1 located in the shuttle cargo bay, 544
Space launch vehicles, 28 T
Space radiation environment, 507 Tack-weld, 179
Space Shuttle, 319 Tantalum foil capacitors, 333
666 Index

Tantalum lead wire, 332 Tin-plated conductors, 371

Tantalum wire lead, 333 Tin-plated copper, 369
Tape testing (for coatings), 224, 234, 236, 237 Tin-plated wire, 339
Technology Readiness Levels (TRLs), 10 Tin-plated wire strands after solderability testing, 374
Technology samples, 419 Titanium, 20
Teflon FEP, 505 Titanium aluminide, 196, 531
Temperature cycling, 509 Titanium aluminides for high-temperature applications, 196
Temperature dependence of specic resistance, 456, 457 Titanium carbide surfaces, 305
Temperature gradients, 319 Titanium ller metal, 125
Temper conditions of aluminium alloys, 567 Titanium hydride embrittlement, 324
Tensile testing at 4.2 K, 451 Titanium hydride precipitates, 125
Test chamber, 8 Titanium hydrides, 323
Textiles, 38 Titanium MMCs, 37
Thermal control, 270 Titanium nitride, 259, 299
Thermal control paints and coatings, 268 Titanium nitride coatings, 303
Thermal cracking, 311 Toe, 34, 291, 302, 314
Thermal cycling on work-hardened beryllium, 284 Toolboxes, 552
Thermal cycling systems, 74 Toroidal water tanks, 325
Thermal fatigue, 271 Total mass loss, 40
Thermal fatigue cracking, 79 Toxicity, 42, 540
Thermal fatigue cracking of copper conductor, 429 Traceability, 56, 67, 101, 110, 255, 327, 379, 435, 495, 616
Thermal fatigue cracks, 349 Trained operators, 621
Thermal fatigue failures in printed-circuit-board, 78 Training and certication, 341, 415
Thermal fatigue on leadless components, 351 Training, 21, 82, 103, 115, 161, 208, 210, 323
Thermal fatigue on semi-rigid cable connections, 353 Transcrystalline fracture, 247
Thermal fatigue on solder-assembled leaded components, 344 Transistor circuit, 333
Thermal fatigue programme, 345 Transmission electron micrographs of beryllium foils, 289
Thermal history from microstructure, 262 Transmission electron microscope (TEM), 68, 250
Thermal management, 509 Travelling wave tubes, 274, 291, 318, 442, 464
Thermal management materials, 220 Trunnion, 27, 549
Thermal mismatch between SMD and substrate, 425 Tube-to-tube TIG welding, 324
Thermal protection system, 47, 536 Tungsten heater elements, 274
Thermal straps, 44 Tungsten whisker, 464, 466
Thermal strippers, 337 Tungsten-inert-gas (TIG)-welded, 125
Thermalcompression bonding technique, 332 Type I high modulus bres, 166
Thermal-cycled solder joints, 509 Type II high strength bres, 166
Thermally conductive adhesives, 426, 427
Thermally induced bending, 321
Thermally induced vibrations, 321 U
Thermo-compression, 334 Ultrasonic testing, 83
Thermoelectric generators, 28 ultrasonic vibration, 332
Thermomechanical test facility, 218 Ultrasonics and other mechanical agitation, 209
Thermo-optical properties, 529 Ultraviolet (UV) radiation, 505
Thermoplastics, 43 Unequal (asymmetric) solder llets, 354
Thermosetting plastics, 43 University, 10, 22, 63, 103, 617
Thermosetting resins, 319 University Spacecraft, 615
Thermount, 353 URALANE, 432
Thick-lm hybrid, 425, 426 Urban miners, 110
Threaded fasteners, 252
Thread-rolling work-hardening, 252
Throw-away modules, 105 V
Thruster chamber, 132 Vacuum, 22, 40, 49, 135, 258, 272, 380, 508, 551, 617. See also
Ti6A14V system and schematic representation of microstructures, outgassing and sublimation
264 Vacuum test chambers, 274
Ti6A14V, 323, 403 Vapour-deposited aluminium, 259
Ti6Al4V superplastically formed propellant tank, 204 Vapour deposition, 134
TiC coatings on steel, 299 Vapour phase machines, 435
TiC-coated 440C steel ball, 305 Vapour pressure curves, 272
TIG-welded 2219-T851, 141 VectranTM, 49
TIG-welded aluminiumlithium alloy plates, 188 Vega, 36
Tin oxide, 489 Velcro tape, 553
Tin pest, 448 Verication programme, 420
Tin plague, 448 Verication testing, 419
Tin whisker growths, 472 Vespal valve seat, 312
Tin whiskers, 490 Vespel, 28, 43, 229, 307
Tinlead coating procedure, 150 Vibrations caused by thermal distortions, 322
Index 667

Video camera electronics, 541 Welding parameters, 118

Visual criteria, 406 Welding rods, 126
Visual inspections, 421 Weld in nickel alloy pressurized housing., 118
VITON B, 540 Weld nugget, 330, 332
Void formation, 337 Weld penetration, 132
Voids and blow-holes in solder llets, 410 Weld porosity, 127
Volatile organic compounds, 109 Weld proles for tube welds, 119
Volatile oxides, 534 Weld sputter particles, 68
VPPA welding, 217 Weld strike, 196
Vulcain engine, 33, 318 Wettability of solder, 412
Vulcain-2 engine, 34 Whisker bridging, 499
Whisker growths, 461
Whisker nucleation and growth, 499
W Whiskers
Warp or twist, 423 aluminium, 468
Waspaloy, 534 C-ring experiment, 482, 485
Water tank, 326 metal oxide, 466
waveguideflange assembly, 407 mitigation, 498
Waveguides, 176 molybdenum, 462
Waveguide switch, 124 precautions, 491
Waveguide-to-Flange Joints, 406 silver sulphide, 463
Wave soldering, 161 tin, 472490
Wear, 25 tungsten, 464, 465
Wear of ball bearings, 296 White paints (/ less than 0.2), 267
WEEE, 107 White plague, 375
Weld bead width and the degree of permitted meander, 119 White residues, 432, 433
Weld decay, 328 Wire ropes, 63, 64
Welded battery cells, 193 Wires and cables, 62, 369
Welded cryogenic tank, 183 Wire strands, 339
Welded galvanized steel, 139 Wire-to-barrel interface, 341
welded joint, 341 Wire-wrapped joints, 338
Welded lead wire interconnections, 329 Wire wrapping, 337
Welded plate, 98 Wire-wrapping pins, 338
Welded solar arrays, 189 Workmanship, 417, 434
Welding, 329 Workmanship drawings, 339
all methods of welding, 181 Workmanship standards, 116, 119, 342
butt, 206, 332 Workmanship standards for resistance spot welds, 117
cold, see cold welding Workmanship standards related to Area Grid Arrays, 122
diffusion, 105 Worn out cathode pellet, 296
Diffusion, 105
EB, see electron beam welding
electromagnetic emission, 196 X
explosive, 186 X-radiation, 185
friction, see friction stir and stud welding X-ray diffraction, 94
laser, 52, 186 X-ray inspection, 436
manual, 181 X-ray laminography, 434
pulsed laser, 186 X-ray radiography, 86
resistance pressure, 330 X-rays, 83, 185
resistance spot, 116, 329
thermoplastics, 188
Welding and joining in a space environment, 49 Z
Welding in space, 52, 53 Zinc, 24, 272
Welding methods and controls, 181 Zinc diffusion, 359, 387
Welding of aluminiumlithium alloys, 187 Zinc emissions, 405
Welding of apogee boost motors, 123 zinc oxide, 359
Welding of commercially pure titanium, 125 Zinc-plated-steel support structure, 108
Welding of thermoplastics, 188