Professional Documents
Culture Documents
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
123
Barrie D. Dunn
School of Engineering
University of Portsmouth
Portsmouth
UK
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).
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.
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
1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
xiii
xiv Contents
Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 629
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 639
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 655
Introduction
1
It is always impressive to look at the hardware of space fossil fuels and industrial processes trap heat and lead to the
ventures, whether in the form of launch vehicles culminating harmful warming of the planet, can be evaluated by using
with the successful launch and landings of the Space Shuttle, the new Earth resources satellites. These remote sensing
of land-sited test chambers, of satellites being tested, of large spacecraft can also study deforestation, damage to the rain
antenna dishes, or of complex electronic circuitry under high forests of South America, Africa, and Asia, and the effect of
magnication. But this is not the real space capability. The sewage and industrial waste on our oceans and shorelines.
real capability lies in the people, in their technical compe- Meteorology, Earth phenomena observations,
tence, and in their manipulations of the metals and materials space-based satellite navigation systems (GPS) and com-
which have made space communication programmes and munications by satellites are four major areas of more
space science programmes possible. immediate interest where space technology can be applied
Every 24 h an average of 45,000 storms break around the usefully for the service of mankind. GPS saves mankind
world. Meteorology is probably one of the oldest sciences, billions of dollars per year in wasted fuel for cars, aircraft
but it is presently one of the least accurate even though the and wear/tear! The Galileo global navigation satellite system
four elements which enable the weather to be forecast (GNSS) should see 30 satellites in operation by 2019, pro-
(clouds, atmospheric pressure, temperature, and wind) are viding users with horizontal and vertical position measure-
perfectly measurable. The various space programmes ments within a 1-m precision. Telecoms satellites in a
undertaken by the US, the USSR, Europe, and more recently geostationary orbit, 22,236 miles above the Earths surface
Japan, India, China, South Korea and Brazil, are providing enable internet signals to travel from one user to another
considerable amounts of information about the environment. anywhere in the world within 700 ms. These satellites allow
The weather develops in a restricted area consisting mainly us to speak and to see from one country or continent to
of the troposphere, the lowest layer of the atmosphere, which another; to receive the same television picture simultane-
never exceeds 16 km in depthbarely twice the height of ously in Britain, Luxemburg, America, Russia, or New
Mount Everest. But no part of the atmosphere will act Zealand will educate and entertain the people of the world
independently of the other, and there is a continuing need to we may hope that the so-called telecom satellites will help to
answer such questions as, how stable is our climate, or, how abolish the frontiers of misunderstanding and ignorance.
much additional atmospheric or water pollution can be tol- This could be the key to the continuation of our civilization.
erated without drastically altering it? Such knowledge comes The aforementioned attributes to space also support the
in many ways, and no more dramatically than in recent years United Nations goals of enhancing the literacy of everyone
from sensors in satellites orbiting the Earth, as illustrated in on the plant, eradicating poverty and improving health.
Figs. 1.1, 1.2, and 1.3, from observations and photographs Of particular concern to the space industries is the
made by astronauts, and even by correlation with the unpredictability of the Suns weather. It has been predicted
atmospheres of Mars, Venus, and Saturn which have also that after years of relative calm, the years 2015 onwards may
been investigated recently by space probes and landers. see a period of more intense activity such as solar flares,
Wonderful images made by the Hubble Space Telescope coronal mass ejections and magnetic storms. This so-called
have focused on galaxies in an innite universe, but this space weather has been added to the UKs National Risk
orbiting observatory has also assisted in the expansion of Register as neutron-storms risk degrading satellite data and
knowledge of our own planet. The greenhouse effect, in hardware, as well as on-Earth facilities that use sensitive
which pollutants put into the atmosphere by the burning of electronic components such as metal-oxide-semiconductor
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 been realized. Chinese scientists are planning an orbital
project called High Energy Solar Physics Data in Europe, space station where the core module will be launched in
will continue to observe solar flares, explosions on the Sun 2018, followed by two laboratory modules in 2020 and
and the magnetic phenomena that flings radiation and par- 2022. Settlements on the Moon are periodically proposed by
ticles across the Solar System. NASA and other space agencies, but since the discovery of
Permanently manned space stations such as the Rus- water-ice close to one of the lunar poles, this concept may
sian MIR and more substantially, the International Space become a reality. The remarkable Mars exploration projects
Stationexpected to remain operational until 2024have involving orbiting spacecraft and rovers are establishing the
1 Introduction 3
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 processing and these are facilitating the development of
that water exists below the Martian surface, the future highly sophisticated spacecraft subsystems and electronic
presence of manmade colonies on Mars may not be just the devices. As these developments have continued it is perhaps
fantasy of science-ction writers. surprising to note that, in the main, it is the ordinary metals
The advancements in space technology have been made and plastic materials and their related manufacturing or
possible by many specic breakthroughs in materials science, assembly processes that nd their way into the construction
manufacturing processes and novel technological advance- of space hardware. Consequently we have now determined
ments. Novel organic and inorganic materials have been that these ordinary materials produce the greatest prepon-
developed by new methods of synthesis and metallurgical derance of failures during the various stages of spacecraft
4 1 Introduction
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 operating stresses and environmental conditions is of great
limitations is a real requirement during the development of help in the diagnosis of a failure mode.
spacecraft systems in order that materials engineers can There is a grave danger that the line of development of
predict realistic margins of safety. space equipment and instrumentation may be lost if care is
Failures may occur as a result of over-testing, overload- not taken to preserve documentation related to past failures.
ing, or over-pressurization when it can be demonstrated that Much information is contained in in-house laboratory reports
no fault is attached to the material itself. Other failures occur which are often led and forgotten. This book may go a little
because of poor choice of material, shortcomings in design, way to collate a small percentage of the examinations per-
mistreatment during construction, or when the part was not formed by the author in order that similar design or pro-
adequate to withstand a particular fatigue or corrosive duction problems may be minimized in the future. The
environment. Spacecraft failures occur during their fabrica- classical failure modes of fatigue, stress corrosion cracking,
tion, assembly, integration, and environmental testing, and hydrogen embrittlement, and the degradation of polymers by
the generally short period of approximately four years ageing, outgassing radiation, etc. have frequently been
between design and launch necessitates that failure analyses associated with spacecraft hardware. It is expected that
are rapidly performed to identify failure modes and their chemists, metallurgists and other material or manufacturing
causes. This can be achieved only when sufcient informa- engineers, who have metallic failures to contend with, will
tion is available about the history of the failed part, from its be able to draw parallel examples to their work from the
initial composition and heat treatment, through manufac- illustrated case histories included within this book. Many of
turing details to records of all post-manufacture storage and these examples necessitate that some of the material
testing. Fortunately, the high degree of surveillance by the requirements for space flight are understood, as these are
Product and Quality Assurance teams of space hardware frequently directly attributable to particular failure modes.
contractors usually enables proper documentation of most An overview of the specic requirements of spacecraft
details pertinent to a failure. This knowledge of the precise materials is given in Chap. 2, and the role of material
1 Introduction 5
evaluations vis--vis spacecraft product assurance schemes microprobe analysers, high-denition radiographic units,
is described in Chap. 3. scanning-laser acoustic microscopies, infrared spectroscopy,
The case histories have been selected from a large num- etc. as diagnostic tools for failure analysis. These are fre-
ber of examinations involving standard material testing quently available on loan, or can be rented from local uni-
techniques, and they are divided into three characteristic versities or research establishments. It must be emphasized
groupings. Chapter 4 concerns problems encountered during that the majority of these tools give very limited information
certain spacecraft manufacturing phases; Chap. 5 relates to to the inexperienced. Only with sufcient practical experi-
failures which possibly occurred during testing; and Chap. 6 ence or on-the-spot guidance will the investigator be able to
concerns failures which may cause deterioration of electrical piece together information gained from various stages of
interconnections. nondestructive testing, supercial binocular microscope
Chapter 7 deals with the many aspects of whisker growth observations, physical and analytical tests, and
as they might affect practising materials engineers associated metallography.
with structural materials, electronics, and space-related In the scope of this book it is not possible to include a
industries. The case histories presented are limited to elec- description of the equipments selected for particular inves-
trical device failures that have resulted directly from the tigations, nor is there place for detailed accounts of the
growth of whiskers. Corrective actions are proposed for each methods which have been chosen. The traditional guidelines
case, and it will be evident that the published literature does applicable to all engineering failures have been followed by
not hold a solution for every situation. the author who is indebted to his highly professional col-
Chapter 8 is a short overview of the effect that the space leagues as well as access to well-equipped materials labo-
environment has on materials. Here, examples are given ratories and test facilities. As will be discussed, the majority
from the large number of analyses made on materials of organic material and metallurgical investigations are made
returned from Low Earth Orbit (LEO). The effects of out- during hardware production or after equipment level testing.
gassing, temperature, micrometeoroids, and atomic oxygen Some defective items also originate from units or structures
are illustrated from data amassed by Space Shuttle flight which have been installed on-board engineering or quali-
experiments, the Long Duration Exposure Facility, Eureca, cation model spacecraft, so that results and recommenda-
Medet and materials retrieved from the Hubble Space tions from material and failure review boards can be fed
Telescope during its various repair missions. back to project designers and engineers. This procedure will,
The major tool utilized for each investigation is that of we hope, eliminate future problems with flight model
microscopical metallographya technique developed by H. spacecraft. This book deals not only with failure analysis but
C. Sorby in the 1860s for the examination of geological also with the measures which may be taken for failure pre-
samples. However, the number of tools available to materials vention by improving product reliability. Finally, the term
engineers for the examination of polymeric molecules, failure can be construed to have many meanings ranging in
microstructure and the like, has increased enormously over scope from the trivial to the calamitous. Throughout this text
the last decade. It is often necessary to utilize advanced failure is employed as a technical term meaning cessation
instruments such as ESCA and Auger spectrometers, laser of function or usefulness.
Requirements for Spacecraft Materials
2
2.1 General Background capabilities of the assigned launch vehicle which will take
the satellite from the Earths surface and inject it into the
The Space Age began in 1957, with an 83 kg Russian desired orbit. The lighter the satellite, the cheaper will be the
Sputnik satellite bleeping greetings to a surprised world. launch costs. Another major performance factor, reliability,
Since that spectacular beginning, intensive effort has gone can also be purchased if money is preferentially funneled
into the scientic exploration of space, exploration of the into reliability and test programmes rather than launch
Moon and distant planets, manufacturing of materials in vehicles. The important point is that performance factors of
space laboratories, and exploiting orbiting satellites for weight, reliability, and cost are all interrelated. The designer
communication, navigation and observation of the Earth. of an applications satellite will be more willing to pay for the
The early steps have passed into history, and most equip- reliability level that would give him 10 years of operation
ment and instrumentation has been and will continue to be than the designer of a scientic satellite designed to shut off
replaced by lighter and more complex substitutes. The transmission after only one year when the mission objectives
remarkable achievements of the Apollo Lunar Exploration are attained.
Programme two decades ago still tend to overshadow the One of the major aims of the European satellite manu-
unmanned automated satellite flights, and it is not always facturer has been to set up a European communications
realized that spacecraft orbiting above all continents of the programme which will develop and launch long-life satel-
world have already revolutionized global communications, lites. A supporting technology programme has been under-
maritime navigation, and worldwide weather forecasting. taken to develop and qualify most of the critical subsystems
These satellites are now vital links in a global network. They that will enter the design of future operational satellites. An
would not have been economically or technically feasible experimental satellite (Orbital Test SatelliteOTS) was
before the advent of near-Earth space explorations. launched in 1978 to evaluate and test the performance of the
Satellite communications started on a commercial basis various subsystems of future European communication
with the launch of Early Bird in 1965, less than eight years satellite systems. OTS and its launcher are illustrated in
after the launch of the rst Sputnik. This was the rst Figs. 2.1a and 2.2. Its major subsystems under evaluation
satellite to remain stationary over the Earth, and it was able included:
to provide a continuous connection between any two Earth
stations. Until comparatively recently these so-called ap- communicationsto relay information (data and com-
plications satellites were merely assemblies of separately mands) between Earth and satellite and, in concept, to
designed components rather than thoughtfully integrated and from other spacecraft.
systems. Often component interfaces failed to match, power supplyto provide electrical power to all satellite
reducing the overall system performance. subsystems.
These satellites, and to a more limited extent the scien- on-board propulsionto provide thrust for orbit changes,
tic satellites, are now incorporating standardized subsys- station-keeping, and deorbiting.
tems in an attempt to optimize performance factors including Environmental controlto maintain specied tempera-
weight, reliability, and cost. tures, radiation levels, electromagnetic environment, etc.
It seems likely that the spacecraft designer has placed structureto support and maintain satellite conguration
greatest emphasis on mass, as this is usually set by the on the ground, during launch and in orbit.
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
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 11
reinforced plastics (CFRP). Because of their exact alignment light (or one-millionth of an inch), a calculation error caused
requirements some solar panels, dish antennas, and antenna the mirror which was originally launched to have been
platforms are fabricated from CFRP which, because of its fabricated with a curvature that was too shallow with a total
small coefcient of expansion, will retain dimensional centre-to-edge error of about 2 m (or 1/50 thickness of a
accuracy under the changing temperature conditions of an human hair). The result was that light rays hitting the mirror
orbit (160 to +180 C). Launch vehicles, satellites, space edges eventually made focus to a point that was slightly
probes and manned modules are predominantly built by away from where light rays from the centre of the mirror
industrial concerns engaged in aircraft manufacture (e.g. focused: a defect called spherical aberration. The HST,
prime contractors such as Boeing, Airbus, Lockheed delayed three years by the Challenger disaster, was launched
Martin, Alenia, Aerospatiale and Astrium). Because of this, in April 1990. Despite the flawed mirror, which rendered
designers will prefer to choose structural and mechanical many of Hubbles initial observations fuzzy, the new
parts from traditional metal alloys and composites, and will spaceborne telescope quickly demonstrated the advantages
limit manufacturing to joining and nishing technologies of an orbiting platform free from the interference of the
which already exist in their respective plants. When com- Earths atmosphere. After the dramatic December 1993
pared to a mass production industry there is often little repair mission, using astronauts from Space Shuttle
incentive to promote the use of advanced materials and Endeavour to correct the mirror and solar panel (see Sect. 8.
alloys which may improve reliability and be weight-saving 2) problems, Hubble began to demonstrate its full potential
but will suffer the drawback of requiring costly fundamental to peer into the universe.
testing and qualication before being incorporated into space
hardware.
Discussions between customers, prime contractors and 2.2.2 Some Futuristic Ideas
their sub-tier suppliers involve contract requirement negoti-
ations related to Materials and Processes issues and a con- Advanced materials are nding more and more applications
siderable number of reviews will take into account such topics in new designs, and this is particularly true of reinforced
as design, materials selection, and fabrication processes. polymers based on carbon or Kevlar bres, clean materials
These are held throughout the various stages of every space (with low outgassing), and several new types of lightweight
project, from inception on the drawing board until the envi- metal alloys. The microminiature electronic circuits so
ronmental testing and qualication of manufactured hardware important for the relay of enormous volumes of data within a
prior to launch. However, it is not until the actual hardware is fraction of a second are also incorporating new materials
seen that one is struck by the results of cooperation between with unique physical characteristics. Microdevices continue
the many engineering disciplines. It is probable that the to be designed and prototyped (David 1996)today these
introduction of computer-aided design now means that are termed micro-electro-mechanical systems (MEMS).
spacecraft subassemblies and piece-parts are being fabricated Although many MEMS devices have been manufactured, to
to the closest tolerances ever achieved. The optimization of date, the only devices that have flown are accelerometers and
structural weight and the smaller design margins mean that a gyroscopes (de Rooij 2009). In the USA, the JPL Centre for
thorough knowledge of the materials selected for the appli- Space Microelectronics Technology has already produced a
cation must be well established. This is particularly true for micro seismometer having a diameter of 12 mm and a
new, advanced materials, as the small design margins means camera on a chip about the size of a ngernail. These kinds
there is no longer a reserve of strength built into the structure, of advancement will certainly lead to smaller, lighter, and
as was the case for earlier spacecraft, to cover ignorance of less costly spacecraft for the future. Even the so-called nano
design loads or stress intensities. The safety margins required satellite, weighing about 12 kg, is thought to be feasible
of materials are real, but the over-conservative designs orig- due to breakthroughs in small-scale engineering of MEMS.
inating from so-called gloom factors or scatter in materials The cost of launching a satellite into LEO by the Space
properties should be a thing of the past. Shuttle was about 14,000 per kilogram and now, 5000 to
To illustrate the accuracy demanded of modern machin- 12,000 per kilogram when an ELV is selected. Costs to
ing capabilities one can consider the unfortunate situation of place a spacecraft into a geosynchronous transfer orbit
the flawed primary mirror of the Hubble Space Telescope (GTO) are estimated to be between and 20,000 per kilo-
(HST). The prime objective of the HST mission was to gram. Either launch vehicles should become less expensive,
obtain images of astronomical objects in approximately ten or satellites need redesigning to become far smaller and
times sharper detail than that obtained by ground-based lighter so that multiple payloads (or even nanosats) can be
telescopes. The HST 2.4 m mirror was designed to be a launched simultaneously. A proliferation in the number of
precisely calculated hyperboloid. Although the mirror is miniaturised satellites (often referred to as CubeSats) have
actually smooth to a precision of 1/64 the wavelength of been built by companies such as Clyde Space and SSTL
12 2 Requirements for Spacecraft Materials
(now part of Airbus) but probably the majority of those face. Similarly, custom compositions can be 3D printed so
presently in orbit originate from schools and universities. that undesirable compositions (possibly brittle intermetallic
These have low construction costs combined with fewer compounds, magnetic phases or corrodible compounds) in
materials and processes requirements. Many have applica- any binary or ternary phase diagram can be avoided by only
tions beyond those of academic research or technology depositing the useful compositions. The waste-, weight- and
demonstration, and are used for Earth observation and money-saving attributes of AM have already attracted
defense purposes. manufacturers in all elds of advanced technology to
The future will see more advanced manufacturing pro- incorporate AM into the production lines for their cus-
cesses involved with the construction of space hardware tomized parts. These include heat-resistant bosses for turbine
even traditional methods such as casting and forging will cases, large bearing housings, rocket engine injectors,
probably be to closer specication and under more highly landing gear support struts, and numerous spare parts.
inert atmospheres. The autoclave curing of composites will Creative endeavors, like that noted in Fig. 2.3, may even
be done under clean conditions without the use of low enable additive layer manufacturing (ALM) to print a hab-
volatile organic materials and any mold release agents will itable structure on the Moon (Redahan 2014). Figure 2.3a
not contain silicones as they are difcult to remove prior to illustrates how functional habitation modules could be
painting. Friction stir welding and FricRiveting can be brought from Earth; the surface of the thin-walled inflatable
envisaged for joining metals to thermoplastics; and laser structures would then be coated by 3-D printing a powder
materials processing will involve localized, intense heating made entirely of regolith, having a particle size of around
of solid targets and components by laser, to achieve ultrafast, 200 m, onto the thin-wall until a sufciently large wall
novel and economic joining and surface engineering. thickness could be built up to protect human space-workers
It used to be impossible to select a material without a full from radiation and micrometeoroids. Regolith is the name
knowledge of how it might best be processed into a nal given to lunar dust and this local resource has already been
piece-partbut today it seems that 3D printing has opened encountered by humans and analyzed with respect to particle
up a world of endless possibilities for designing and creating size and composition (see Fig. 2.3b, c). Quantitative optical
everything from a complex space mechanism to printed and electron-probe studies by the UK Institute of Geological
chocolates and foods for astronauts! 3D printing, also known Sciences (Simpson 1970) have shown that lunar samples can
as additive manufacturing (AM), produces three- contain ilmenite, pyroxene, chrome-titanium spinel, troilite,
dimensional items by printing them, layer-by-layer from native iron, iron-nickel alloy, and even native copper (as
raw ingredients consisting of powdered plastics, aluminium shown in Fig. 8.1). This concept for constructing a human
alloys, titanium alloys, low expansion alloys and other outpost on the Moon using lunar soil, and ways to monitor
spacecraft materials. The most usual processing method is to the buildings progress from Earth by means of an industrial
introduce metal powder into a laser beama precise depo- CCD camera positioned on the printer, have been described
sition of either sintered or melted powder is directed onto a by Ceccanti (2010) and Colla (2014).
flat table. The laser beam is controlled by CAD programs to Another rapid prototyping process that demonstrates
raster across the sintered metal powder layer so consolidat- great promise has been described by Maxwell et al. (2013).
ing the deposit before another layer is added. The item is Known as Hyperbaric Pressure Laser Chemical Vapour
then built up, layer-by-layer to create a net-shaped part. This Deposition (HP-LCVD), this rapid prototyping process
revolutionary rapid protyping process can now be used to incorporates a mixture of reactive gases into which laser
create nalized spacecraft parts with a 40 % weight saving. beams penetrate for the growth of materials from atomic
Some improvements in the chemical purity of the powders level to large structures by means of thermally- or
used should increase yields to 100 % and, already, a 3D photolytically-induced decomposition of the gaseous pre-
printer is in operation on the International Space station. cursor. Exploitation of this HP-LCVD process, as prescribed
This rst 3D printing in space was performed in December by Dynetics and the NASA Marshall Space Centre (Maxwell
2014 by Butch Wilmore as part of a Zero gravity demon- et al. 2013), may enable 3D rapid prototyping in-space for
strationengineers up-linked a custom-made digital design the fabrication of components, replacement parts and even
le of a ratchet wrench to the 3D printer and produced a tool nuclear thermal propulsion systemsby the use of precursor
measuring 11.4 cm in length. This process will enable fragile gases and raw materials found, often in abundance, within
items to be manufactured in space without the need to our own Solar System.
incorporate robustness and extra weight for surviving the Numerous advanced materials and manufacturing tech-
shocks, vibration and mechanical loads encountered during niques will be individually described in Chap. 4.
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
2.2 Considerations for Materials and Processes 13
Fig. 2.3 a An artists impression of an igloo, built on the Moon by cycling. Concept and illustration courtesy of ESA and architects Foster
means of a 3D printer attached to the arm of the robotic vehicle seen to + Partners. b Apollo 11 astronauts footprint in lunar soil, made up of
the right. Printing powder material is Moon dust (regolith), processed small, dust-like particles of regolith (courtesy NASA photo
into a cell-like structure of high strengththe idea is to initially inflate AS11-40-5877). c Grain size distribution of lunar soil from three
a thin folded dome brought from Earth and protect it with a cellular different sites; about 50 % is greater than 100 m from Heiken et al.
shell using the 3D printerthe pressurised enclosure so sheilds (1974)
astronauts from solar radiation, micrometerites and severe thermal
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
2.2.3 Some Basic Considerations Regarding Rooij 1989a) and is presented in Table 2.1. However, de
Corrosion Prevention Rooij has simplied this Table into Groupings of metallic
alloys and the modied tabulation now appears as shown in
It is necessary to ensure that any newly selected material will Table 2.2. The material may need a high resistance to Stress
retain its functional properties during all stages of the Corrosion Cracking (SCC) before launch, and in such cases
spacecrafts designed life, up to the end of the mission. can be selected from those alloys listed in Table 2.3.
During manufacturing, the material must not degrade The primary and secondary structures will be made from
because of contamination from processing steps such as the light alloys based on aluminium and magnesium, together
release agents used for items moulded from CFRP, or by with titanium and to a very limited extent beryllium. Nickel
cutting oils used in the machining of alloys. Galvanic and alloys are often selected for their high-temperature perfor-
general surface corrosion must be avoided during environ- mance and oxidation resistance; they are often known by
mental testing and ground storage by the correct selection of trade names, rather than by their specication code numbers.
surface nishes such as anodic lms, chemical conversion Commercially pure nickel, easy to form into complex
lms, and paints. When electrical grounding is required, shapes, is used in the construction of spacecraft electronics
only contacts having a compatible coupling of less than where its electrical and magnetic properties are crucial.
0.5 V should be chosen. The static corrosion potential for a Mechanical designers often select Inconel alloys 600 and
large number of metals and alloys has been established (de 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 The new millennium saw a great increase in the size and
from SCC in pure water, at temperatures above 300 C complexity of spacecraft. This has necessitated ground test-
(steam generators for nuclear plants). It may be wiser to use ing facilities to become modernized, physically larger, and
a development of the 625 alloy when high temperatures may more sophisticated for the exposure of space hardware to the
be expected, for instance in propulsion systems. This environments listed in A to C above. The Test Facility within
development involved the additions of molybdenum and space agencies are comprised of high capital investments
niobium to 625 to impart solid solution hardening and the such as Large Space Simulators that reproduce the vacuum,
formation of Ni3Nb, a very effective hardening precipitate. certain radiations and the cryogenic-to-elevated temperatures
This is known as the super alloy Inconel 718 and has encountered by space hardware. Mechanical and acoustic
become the most widely used high temperature nickel alloy. tests simulate the launch environment, the magnetic
2.2 Considerations for Materials and Processes 21
characteristics of spacecraft are evaluated and all tests are and other operator handling errors, etc.). It was seen that the
performed under various classes of cleanroom conditions. reliability of electrical interfaces between equipments, and
As mentioned previously, a development plan for a new the mutual compatibility between the constituent subsys-
spacecraft design will involve a model philosophy where tems, was seriously jeopardized by fundamental oversights
models of the craft will be dynamically testedwithout (e.g. high electrical resistance between gold-plated and
testing the risk of failures is too great. In the early days it was aluminium-nished interfaces due to galvanic corrosion;
not uncommon for space authorities to build four models for migration of silver to produce short circuits; the use of
testing prior to actually building a flight spacecraft. The austenitic steels having work-hardened and therefore slightly
model philosophy will be accounted for in the next para- magnetic surfaces in locations required to be magnetically
graph, but it is emphasised that as the space industry has clean; etc.). Remedial actions were taken by suppliers and
matured, the design margins have become established (for assemblers as a result of failure review boards (FRBs).
structures and electronic systems) so that there is now more Lessons learnt documents (nowadays called internal problem
focus on analysis and less on actual testing. It is now the notication documentsIPNs) were written and circulated.
norm to build only one prototype for testingoften this Finally, a fully assembled qualication model was built and
build and test is completed only six months before work subjected to a comprehensive series of environmental
commences on manufacturing the flight spacecraft. ground tests. These usually included: sine vibration, spin,
It is interesting to remember the aims of the model acoustic noise vibration, centrifuge acceleration, and solar
philosophy for the early (70s and 80s) European simulation. Each of these major test phases was preceded by
telecommunication satellite projects. The structural model an integrated systems test in order to verify that the func-
was subjected to a programme of tests which exceeded the tional behaviour of the satellite was correct. It was not
expected launch environment conditions (each type of uncommon that the qualication model would be recognized
launch vehicle has its own characteristic levels of vibration as a flight spare in case of launcher failure.
and acoustic noise). Typical test congurations are shown in Flight model test programmes were, and continue to be,
Fig. 2.1ac. The enormous amount of energy released during more limited than those used on the qualication model
launch can be witnessed from Fig. 2.2. Weakness in designs (most customers now refer to this as the protoflight model).
may be exposed by this model, such as failures resulting It may be assumed that some subsystem parts that will
from fatigue of welds, struts, electronic box hold-down operate for relatively short periods after launch can accu-
points, and the like. The thermal model was subjected to mulate several hundred hours of test operation before the
solar simulation and thermal balance testing to conrm and actual time of launching. Materials selected for use under the
update previously determined mathematical models. During vacuum conditions of space may therefore have to operate for
these tests, decient designs may promote several material periods under normal atmospheric conditions. This may
failures related to thermal fatigue, overheating, and embrit- create special problems. One is immediately reminded that
tlement of incompatible joining techniques and metal alloys. very thin (tens of ngstroms1) lms of lead, or molybdenum
By simply modifying the paint nish of the spacecraft sur- disulphide, for the lubrication mechanisms, can rapidly oxi-
face, or by the attachment of reflective mirrors, it was found dize under terrestrial conditions and become the cause of
possible to adjust and reduce the local temperature envi- malfunction. This goes to illustrate the need to know the
ronment of each subsystem or equipment and reduce the effects that ground testing may have on delicate surfaces.
chance of thermally induced failures. Workmanship prob- A nal concern of the writer relates to the participation of
lems abounded, with non-conformances relating to open materials experts on spacecraft project review boards. It is
circuits in cable harnesses due to wires separating from paramount that an experienced materials engineer is incor-
crimp barrels and cold soldered joints on circuit boards (such porated into each of the four major reviews during the design
events became less frequent once operator and inspector and construction of individual spacecraft. Whenever ECSS
training schemes were introducedsee Sect. 6.14). The Q-ST-70 is included as a contractual document this become
engineering model ensured that integration and performance a requirementas a minimum, the materials engineer will
could be achieved. A review of the old project and labora- manage the steps taken for the project-approval of declared
tory failure investigation reports was made by the author (at lists for every flight material, mechanical part and their
the time of ling and archiving as a result of which they were related processes (i.e. DML, DMPL and DPL). Tasks should
lost forever!). This revealed that several material problems also include cleanliness and contamination control, the
only came to light during integration, particularly at the testing and validation of new materials, assistance in 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 1 equals 1010 m.
1
22 2 Requirements for Spacecraft Materials
qualication of mechanical parts and the verication of new (b) Sublimation and evaporation
or critical processes.
Spacecraft, such as CubeSats and university flight The minimum altitude for an Earth-orbiting satellite is
experiments may follow less rigorous requirements and a 200 km (125 miles), and, once this altitude has been reached,
reduced M&P programme is suggested in Appendix 8. appreciable changes can be produced in common engineer-
Reviews are usually contractual milestones and essen- ing materials, whether they be metals, plastics, or ceramics.
tially question: will the design, hardware, software, and The vacuum in space is very high, the pressure falling from
operational approach satisfy the mission objectives? The 106 mm Hg at 200 km to less than 1012 mm Hg beyond
review names and their main objectives are as follows: 6500 km. Some polymers will decompose and some metals
will tend to sublimate under vacuum. The rate at which the
PDR (preliminary design review): after evaluation of thermal molecules or atoms leave a surface in vacuum will rise
and/or engineering models, to approve and release the rapidly with an increase in temperature according to the
preliminary design, including materials and processes. equation:
CDR (critical design review): this establishes the nal design
and agrees that flight hardware manufacturing can com- G 5:04 103 PM=T 1=2
mence (all declared materials being approved and pro-
where
duction process, when required, are veried as being
suitable).
G grams of material evaporated or sublimated per square
QR (qualication review): assess that all qualication
centimetre per day
activities on subsystems are completefor certain projects
P Vapour pressure of the evaporating species in torr2
a qualication model is built.
M Molecular weight of the material
LRR (launch and operations readiness review): this checks
T Absolute temperature, K
that all lower-level acceptance reviews have been suc-
cessfully completed, the flight model spacecraft accepted, Temperature has an enormous effect on the amount of
and it is then authorized to be launched. 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:
2.3 The Effect of a Space Environment Cadmium has a vapour pressure of 108 at 70 C
(approximately)
107 at 90 C
106 at 120 C
(a) General 105 at 150 C and
104 at 180 C
The purpose of this section is to provide the reader with an
overview of the salient points concerning the effect of a The relationship between the vapour pressure of a metal
space environment on spacecraft parts and materials. Far (P) and temperature (T) is given by the following equation:
greater discussion with examples will be made of these
effects throughout the remainder of this book. P P/ eE=RT
Each spacecraft material will be required to suffer no
where P = a constant (i.e. vapour pressure at T = ),
vibrational fatigue damage during the launch. In orbit, it
E = heat of evaporation (e.g. joule mole1), and R = gas
will need to survive the space environment (see
constant (8.3 J mol1 K1).
Table 2.4) and, in particular, to possess low-outgassing-
The temperatures for given metallic sublimation rates
under-vacuum properties, whether it be a lubricating
are listed in Table 2.5. The thermal environment in space is
grease or a structural plastic. Radiation and thermal
completely different from thermal conditions on Earth.
cycling must not degrade thermal-control surfaces or
Without an atmosphere, the only means of exchanging
joints in materials possessing different coefcient of
thermal energy is by thermal radiation and conduction.
expansion. The presence of atomic oxygen in low Earth
Certain parts of satellites have been calculated to follow
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 2
The term torr is generally used instead of mm Hg by international
atomic oxygen durability need to be developed. These agreement of several vacuum societies. Torr honours the name of
aspects will be detailed in Chap. 8. Torricelli, who discovered atmospheric pressure in 1643.
2.3 The Effect of a Space Environment 23
Table 2.4 Characteristics of the space environmentorder of magnitude only (Dauphin 1984)
Altitude Pressure (mm Kinetic Gaseous density Composition Ultraviolet Particle radiation
(km) Hg or torr) temperature (particle cm1) radiation (particles cm2 s1)
(K)
Sea 760 300 2.5 1019 78 % N2, Section of solar
level 21 % O2, 1 % A spectrum 0.3
30 10 4 1017 N2, O2, A Absorption zone
6
200 10 1200 1010
N2, O, O2, O +
Full solar
spectrum
800 109 1300 106 O, He, O+, H Full solar
spectrum
6500 1013 103 H+, H, He+ Full solar 104 protons 35 MeV
spectrum 104 electrons 40 keV
22,000 1013 101102 85 % H+, Full solar 108 protons 5 MeV
15 % He2+ spectrum
108 electrons 40 keV
104 electrons 1.6 MeV
thermal excursions from approximately 160 to +180 C. proong of any of the spacecrafts components. Magnesium
The actual temperature attained will differ from one space- parts could pose sublimation problems after long exposure to
craft to another, the major temperature effect arising from the vacuum at temperatures greater than 125 C, and experi-
spacecrafts spin rate. Surfaces of non-spinning satellites ments have shown (Frankel 1969) that magnesium sheets
exposed to direct solar radiation may be unable to dissipate held at 230 C and 1 107 mm Hg for only 168 h (one
thermal energy efciently, and will reach higher tempera- week) became severely pitted and dramatically decreased in
tures than spinning satellites. Temperature variations will static strength properties. Because of their relatively high
also depend upon the amount of albedo radiation and the strength-to-weight ratios, magnesium alloys are often
amount of thermal radiation to space from the spacecraft. employed as structural parts, but it is essential that these
Both active and passive thermal control systems are parts are nished with an adequate plating or chemical
employed on satellites in order to restrict the oscillating conversion coating which will prevent both the corrosion of
temperature extremes. The active systems have made use of the part before launch and the subsequent sublimation
thermostatically controlled heaters. Passive systems involve problem in orbit. Tinlead alloys, such as those employed
the surface absorptance/emittance, /, properties of material for soldering electrical components, have not been seen to
nishes. Solar reflectors have low / ratios, being generally sublimate under spacecraft environments as they are
white paints or clear anodized aluminium. Black paints and restricted for use in areas which are thermally controlled to a
inorganic black anodized aluminium wave / values of maximum of about 80 C. Solder alloys are used for joining
approximately 1. Solar absorbers have an / value greater silver-plated molybdenum interconnector strips between
than 1, and these are generally polished metals since the solar cells on the solar arrays of spinning satellites. As such
emittance values of uncoated metals are very low (<0.l). satellites rotate, the maximum temperature of the arrays does
Examples are described in Sect. 5.5. not degrade the soldered joints. The stationary communica-
The majority of metals do not sublimate at normal tion satellites will not be able to dissipate the absorbed
spacecraft temperatures. However, as can be seen in thermal radiation on the solar arrays efciently, so that
Table 2.5, cadmium and zinc must be excluded from use as welded interconnectors are necessary.
they will readily sublimate and could condense in the form The events of sublimation and evaporation cause the
of thin conductive deposits on electrical insulators, or opa- release of metal atoms which travel and are capable of
que deposits on optical components which may be situated recondensing on cooler surfaces. They are readily ionized
within the satellite or on its external surfaces. All cadmium, and may be contributors to corona and arcing phenomena.
zinc, or tin-plated surfaces, such as the protective nishes on These metallic ions may also cause the complete failure of a
equipment or components including commercial connectors, satellite mission by recondensing between slip rings, causing
must be avoided, because, as will be described in Chap. 7, electrical short circuits, or recondensing on optical surfaces
they are known to grow single-crystal whiskers exceeding causing the loss of a specic wavelength transmission. When
2 cm length in vacuum. Extreme care must be taken to such ions recondense on the spacecrafts highly reflective
ensure that these metals are not used in the corrosion thermal control surfaces the thermal balance can be so
24 2 Requirements for Spacecraft Materials
degraded that severe overheating will promote malfunctions. been developed using this metal, but CNES evaluations have
Cesium is an uncommon metal with a melting point of 28 C determined that cesium can easily contaminate spacecraft
and a boiling point of 671 C but it has been proposed as a surfaces due to its low vapor pressure and re-evaporate from
propellant for eld effect electronic propulsion (FEEP) when surfaces warmer than 30 C to cause further contamination
very low thrust applications are required. Thrusters have and react chemically with polymers or some oxides to
2.3 The Effect of a Space Environment 25
change their optical properties (Tondu 2011). The impracticable to enclose such moving parts within hermet-
BepiColumbo mission dedicated to the study of Mercury is a ically sealed containers, so special lubricants must be found
challenging ESA project where temperatures may reach which do not decompose or sublimate under vacuum. Films
300 C at all external surfaces and metallic sublimation is of of low shear strength, such as molybdenum disulphide or
major concern (as is the outgassing of organic materials). vacuum deposited soft metals (e.g. gold, lead, or silver), are
A reference to Table 2.3 will immediately show that mate- most efcient, particularly when designed to be situated
rials containing zinc, cadmium (always prohibited in space between hard substrates which will support the load and
applications) and lead are unsuitable candidates. Sublimation keep the contact area small.
and cold welding of solar array drive mechanisms will The possibility of contacting metallic surfaces becoming
require strict controls for the bearings, slip-rings and cables cold welded to each other during, for instance, the operation
(Fink et al. 2009a, b). of spacecraft mechanisms or loading of threaded fasteners
For additional examples of the effect of sublimation on under vacuum conditions depends on a number of factors;
spacecraft hardware, the reader may refer to Sect. 5.6. there will be a greater chance of cold welding if
(c) Radiation and particle damage (a) the relative phase diagram indicates that the contacting
metals or alloys form a solid solution with each other,
Organic materials, such as those used for electrical insula- (b) the metals are soft and have the same crystal structure,
tion, may be damaged by ionization due to protons and (c) contact surfaces are clean, or possess easily
electrons from radiation belts, solar emissions, and cosmic damaged/removable oxide lms. Surface oxides are
rays. The Van Allen radiation belt is especially damaging to normally more brittle than metals and are therefore
organic materials and even inorganic materials which make more likely to crack and expose underlying metal if it is
up optical lenses, ceramic insulators, and sensitive electronic too soft to provide a rm support under load, and
components. The metallic materials of most Earth-orbiting (d) contact pressures are higher. The rst step in order to
spacecraft, as well as deep-space probes, are unaffected by minimize the possibility of cold welding is to select
small particle radiations, and are only slightly degraded by those metal combinations known to resist adhesive
erosion from a cloud of meteoric dust which surrounds the wear, as for instance those shown in Fig. 2.4a. The next
Earth and other planets. Larger particle damage was, how- step is to consider the surface nish.
ever, a major problem for the Halleys Comet yby mission
called Giottowhich encountered that comet in 1986. Clean metallic surfaces react with the surrounding
Special armour plating was developed to surround that atmosphere to form oxides, nitrides, or other compounds that
spacecraft, which was impacted by rock debris travelling at are held together by either strong chemical bonds, or weak
hypervelocities of 1070 km/s. These aspects are reviewed van der Waals forces. These surface lms reduce the pos-
in Chap. 8. sibility of metalmetal adhesion that would otherwise occur
on intimate contact; they can be considered as naturally
(d) Friction and wear formed lubricating lms. Under the space environment such
lms are unlikely to be self-healing if they become dis-
One of the major material problem areas for advanced placed. For this reason protective lms of PTFE (Teflon),
spacecraft is that of friction and wear of surfaces which must graphite, and molybdenum disulphide are frequently selec-
rub or slide over each other under conditions of temperature ted to prevent wear and cold-welding.
cycling and high vacuum. This may be encountered in the Suitable non-lubricated pairs of engineering alloys for
operation of hinges, gears, bearings, and electrical contacts sliding wear situations can be selected from engineering
used in a vast number of spacecraft mechanisms. During alloys having very different hardnessestraditional refer-
sliding under normal terrestrial conditions most contacting ence books may be consulted (Brandes 1983; Lansdown and
metallic surfaces are protected by a surface lm of oxide, Price 1986), or for very precise data, the recent work of
oil, grease, or other contaminant which will act as a shear Merstallinger et al. (2009) can be checked for evidence of
layer and prevent binding. Under vacuum conditions such fretting wear and cold welding (it is intended that this ref-
contaminants outgas, and oxides, once disrupted or erence work is maintained on the internet and updated as
removed, are unable to reform. Also, the minute junction new pairs of engineering material pairs become tested).
spots which carry the full load between contacting metallic In general, steels can be coupled with either copper alloys
surfaces will usually have vastly increased friction coef- or steels of a different alloy type and hardness. Copper alloys
cients, and probably high rates of wear, so that many can be coupled with chromium plating, high-chrome steels,
metallic couples will tend to cold-weld. It is usually and tungsten steels. Austenitic stainless steels have a
26 2 Requirements for Spacecraft Materials
Fig. 2.4 a Depicts the theoretical possibility of clean metal surface lubricants) such sticking will be avoided. The very thin oxide lm on
pairs becoming cold-welded upon contact. Choice of a metal to resist austenitic stainless steel is easily ruptured under the sliding conditions
adhesive wear with another specied metal. The blacker the circle, of torquing-up nuts and bolts. High vacuum operation has contributed
the better will be resistance to adhesive wear and cold welding under to the complete seizure by cold-welding of this 316 alloy vacuum
vacuum (Lansdown and Price 1986). Same-to-same metal contacts will chamber support xture. The 45 mm diameter bolt was initially cut to
result in solid solution, sticking and cold welding. b Demonstration of a release the threaded portion which was later cross-sectioned. The
case where same metal contacts have become cold welded (a form of pointer shows the main region of cold-welded asperities
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 standard (Labruyre and Urmston 1995; Doyle and Hubbard
they are both soft and unable to form thick protective 2010; ECSS-E-ST-32-08 2013).
chromium oxides. Alternatively, with very hard substrates It should be noted that all dry lubricants wear, and will
such as chromium plating, surface deformation is so small possess a nite life. However, improved wear characteristics
that the surface chromium oxide lm is never disrupted (see can be achieved by carefully selecting the process of applying
Fig. 2.5). Materials suitable for spacecraft bearing applica- the dry lubricant. Ion-plated lead and sputtered molybdenum
tions are titanium-carbide coated balls located in raceways disulphide are now well proven, having low coefcients of
fabricated from 440C or SAE 52100. A particularly good friction and a long life. Burnished or spray-bonded molyb-
anti-wear surface, as for instance in gears, bushes, and piv- denum disulphide has inferior friction properties. It is
ots, is plasma-nitrided steel (Rowntree and Todd 1988). important to store the dry lubricated spacecraft mechanisms
Thermal spraying is also a novel process for the coating of in a dry inert gas in order to prevent moisture pick-up, and as
spacecraft subsystems enhancing wear resistance and as a neither lead nor molybdenum disulphide perform well in air
thermal barrierhere thermal spraying can deposit both low the number of operations in normal atmosphere should be
and high melting point materials such as polymers and restricted (Rowntree and Todd 1988).
ceramics and metallic layers such as aluminium onto CFRP
substrates such as antennae face-skins (Sturgeon and Dunn (e) Cryogenic temperatures
2006; Saber-Samandari and Berndt 2010). Case histories
related to wear are detailed in Sects. 5.2.7, 5.11, and 5.12. All spacecraft structural metals will undergo changes in
A large number of rules and design recommendations for the properties when cooled from normal ambient temperatures
avoidance of wear and cold welding (for instance at mech- to temperatures in the subzero range encountered during
anism end stops, hold-down and release springs, sliding solar eclipse periods or when voyaging on deep-space mis-
contacts, and ball bearings) have been listed in the form of a sions. This will be an important factor when liquid helium
2.3 The Effect of a Space Environment 27
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.
by eld emission from a liquid metal source that can be the most outer bodies known to orbit the Sun, the RGA
indium. Ion release is usually for a short time, until the power system will be turned on and seven science instru-
spacecraft potential is reduced, or reaches zero. ments activated.
Soyuz rocket. Supplies of equipment and food are trans- utilizing only the liquid or solid materials on-board; at pre-
ported to the ISS using either the Russian Progress space- sent, there is no possibility of using atmospheric oxygen.
craft of the European Automatic Transfer Vehicle. Since its The term propellant is used to denote the two chemical
rst flight in 2008, the ATV has played a vital role in ISS products, the oxidizer and the fuel, contained inside
logistics serving as a cargo carrier, space tug and storage conventional rockets. The fuel is burnt with the oxidizer in
facility. ATV will evolve into the European Service Module order to achieve the enormous amounts of energy needed for
(ESV) designed to support the NASA Orion spacecraft. The liftoff. The propellants can be in either liquid or solid state.
deep-space missions envisaged by NASA will rely on the Modern rockets are powered by cryogenic propellants:
Orion spacecraftit has been flown on a successful test liquids which at atmospheric pressure have boiling points
flight aboard a Delta 4 Heavy booster launched from the below 0 C. Examples are liquid oxygen (183 C) and
Kennedy Space Centre at the end of 2014. liquid hydrogen (253 C).
Propellants for launch vehicles are regarded as materials. The most simple solid propellant is prepared from a
All rockets are propelled into the vacuum of space by mixture of nitrocellulose and nitroglycerinethis is the fuel.
30 2 Requirements for Spacecraft Materials
can be easily rolled and formed into complicated shapes, oxygen. Similarly, any pressure vessel liner materials based
then weldedafter this a suitable heat treatment is made on organic materials need to be compatibility testedthis is
which produces very hard and tough material properties. more difcult, only one resin system, Torlon Al-10, being
Liquid fuels and oxidizers are usually stored in pressure found to be compatible with liquid oxygen (Healy et al.
tanks which may be made of titanium or aluminium alloys, 1995).
as shown in Fig. 2.8. These inner surfaces which make The Zenit ELV of the former Soviet space agency was
contact with the liquid fuels must be tested and found formally announced in 1989. The vehicle is a two-stage
compatible and non-igniting with the liquid. For instance, liquid oxygen and kerosene rocket which, like all CIS
the titanium alloy Ti6A14V is known to be compatible with launchers, is assembled horizontally. Its launcher assembly,
hydrazine, and aluminium alloys compatible with liquid payload integration, and launch preparation phases have
32 2 Requirements for Spacecraft Materials
Fig. 2.14 a Night launch of Space Shuttle. b Illustration of the thermal photographed at the KSC Visitors Centre after retirement; showing
protection systems (TPS) for the Space Shuttle fleet. Key: RCC the triple paned, optical quality windows, the LRSI ceramic tiles and
reinforced carbon/carbon; HRSI and LRSI high- and low-temperature beryllium is associated with the window frames. d Scanning electron
ceramic tiles; FRSI felt reusable surface insulation. c Atlantis image of Shuttle tile (silica bers in unglazed condition) 1000
can hold a six-inch square tile at its edges and not feel any seen to vary in diameter from 1 to 10 m. This material has a
heat when the centre of the tile is glowing red-hot). There are density of 136 kg m3, but more rugged varieties are denser
about 31,000 tiles on each Shuttle, each moulded and cut to and some contain boron, which causes the bres to weld
t the orbitors contours. The basic material for the Shuttle together. The thermal protection surfaces developed from
insulation is a very ne glassy bre of pure silica (SiO2). Shuttle experience are now very robust. The
Figure 2.14a shows the approximate layout of the various alumina-enhanced thermal barrier tiles overcoated with
Shuttle thermal protection systems (TPSs). The reinforced reacting cured glass have been produced that can withstand
carbon-carbon (RCC) regions can reach temperatures of up the impact of 800 kph raindrops (Healy et al. 1995). The tiles
to 1500 C during descent through the atmosphere. The are bonded onto the aluminium skin of the Shuttle with a
reusable felt surface insulation may reach 500 C during silicone adhesive known as RTV 560, which is a somewhat
ascent. The HRSI and LRSI ceramic tiles differ only in their soft material that avoids cracking of the rigid tiles as the
surface coatings (Vaughn 1985). The high-temperature aluminium airframe deflects in flight. Thermal protection is
(HRSI) tiles are coated with a high-emissivity black reac- not needed for lower surface temperatures; for instance,
tion cured borosilicate glass (to re-radiate the heat efciently Inco-617 could be selected for temperatures under 980 C
during re-entry). The low-temperature material is coated and Ti-1100 for temperatures below 700 C (Baumgartner
with a white silica/alumina coating designed to reflect the and Elvin 1995).
suns radiation as the craft is in orbit. Figure 2.14b shows the The retirement of the Shuttle fleet in 2011 has meant that
microstructure of a ceramic tiles interior. Here, the bres are crew members from the ISS can only return to Earth by
36 2 Requirements for Spacecraft Materials
holder is transported into a vacuum chamber having speci- permissible temperature for a few days in order to
ed dimensions. Testing is made under vacuum at 1 Pa, at a remove residual contaminants, process contaminants and
temperature of 120 C, for 24 h. A collector plate is main- handling contaminants.
tained at 25 C and situated at 10 mm above the sample
during the test period. Figure 2.19 illustrates the basic con- Several compilation tables of outgassing data of materials
cept of the outgassing test method. After the samples intended for spacecraft use have existed since the 1960s. Of
temperature/vacuum/time exposure, the sample is again course, much of that data is now obsolete as material man-
weighed. It is then re-conditioned at 55 %RH (in order to ufacturers and suppliers update their products and withdraw
re-absorb any water) and re-weighed. The mass of con- some from the market. Importantly, some manufacturers
densed material on the collector plate is also measured by may have retained their products Trade Name, but changed
accurate weighing. the procedure by which it is made. This makes it important
The following terms should be understood: to check outgassing data on every batch of production
unless safeguards are made to ensure that products conform
collected volatile condensable material (CVCM)quantity to certicates of compliance, or that process identication
of outgassed matter from a test specimen that condenses on a documents (PIDs) are established and audited by the
collector maintained at a specic temperature for a specic procuring agent.
time Websites can be visited to acquire some preliminary data,
NOTE CVCM is expressed as a percentage of the initial unfortunately they are often not kept up-to-date, and include:
specimen mass and is calculated from the condensate
mass determined from the difference in mass of the col- http://outgassing.nasa.gov/
lector plate before and after the test. http://TEERM.nasa.gov/links/html
recovered mass loss (RML)total mass loss of the specimen http://esmat.esa.int/Services/outgassing_data/outgassing_
itself without the absorbed water data.html
NOTE 1 The following equation holds: RML = TML
WVR. Testing for outgassing may be made at centres such as:
NOTE 2 The RML is introduced because water is not Integrity Testing Laboratory, Markham, Ontario,
always seen as a critical contaminant in spacecraft Canada; AAC, Wiener Neustadt, Austria; ESA-Estec,
materials. Noordwijk, the Netherlands; INTA, Spain; TS-Space Sys-
The WVR is the water vapour regained by the sample after tems, UK, and the Goddard Space Flight Centre, Greenbelt,
reconditioning. MA, USA.
total mass loss (TML)total mass loss of material outgassed Outgassing is distinctly different from the offgassing of a
from a specimen that is maintained at a specic constant material. Outgassing, as previously discussed, occurs under
temperature and operating pressure for a specied time a reduced pressure or vacuum. When materials emit volatile
NOTE TML is calculated from the mass of the specimen compounds in an ambient atmosphere this is termed off-
as measured before and after the test and is expressed as a gassing. It can be likened to the smells that are emitted from
percentage of the initial specimen mass. the interior of a new car. In the 1970s the offgassing of art
objects in a museum were investigated and found to be
The following Accept Criteria are generally contractual related to the degradation of both antiquities and their
requirements set by the customers procuring space hardware, packaging materialsa standard Oddy test method was
but it is accepted that many spacecraft units can be baked in devised using gas chromatography-mass spectrometry
order to satisfy these materials requirements: (GCMS) analysis. Many offgassed species were found to
As a minimum, the outgassing screening parameters for a degrade stored antiques, they were based on alcohols, acids,
material selection shall be as follows: amines and siloxanes (Tsukada 2012). In one specic case,
amines were found to volatilise from polyester polyurethane
1. RML < 1.0 %;
foams that were used to cushion and padding within storage
2. CVCM < 0.10 %.
cases. The amines released by these foams had severely
NOTE 1 For materials used in the fabrication of optical damaged Ming Chinese jars by causing thick white
devices, or in their vicinity, the acceptance limits can be efflorescent deposits to form on these ancient pieces. Very
more stringent than those stated below. similar contamination from offgassed foams has resulted in
NOTE 2 It is nowadays becoming standard practice to the corrosion of stored spacecraft electronic components:
bake critical hardware (such as structural parts, harness, leads nished with silver, nickel and also tin-lead have
electronic boxes and thermal blankets) to the highest 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 2 Requirements for Spacecraft Materials
rendered non-solderable, even with activated solder flux. It is radiation induced damage to these bresbut this will
strongly recommended to avoid the following packaging depend on the orbital altitude and proximity to the Van Allen
foams and sheets: polyvinyl chloride, Styrofoam, rubber, belt. The Long Duration Exposure Facility (LDEFsee
paper, wool, silk, nylon and some untested polyurethanes. Fig. 8.12) demonstrated that after 69 months in LEO no
Better long-term storage media, some of which provide for damage ensued except for one cable hit by a micrometeorite.
anti-static discharge, would be glass, stainless steel, poly- Adhesives, coatings and ATOX-resistant paints:
ethylene, Dow Ethafoam 220 or similar, and Azote poly- Araldite AV 138 with hardener epoxy from Huntsman;
olen foams and Zotek polymer foams from Zotefoams. DC93-500 space grade encapsulant from Dow Corn-
Offgassing is associated with the safety of human life and ing; RTV 566 silicone from Momentive Performance
is addressed during the assessment of materials and specic Materials; MAPSIL 213 silicone potting and MAP ATOX
equipment that will be used in a manned environment such 41-8 varnish for atomic oxidation protection, also electri-
as the International Space Station. European standards for cally conductive brown adhesive QS 0225 EA83, thermally
offgassing and toxicity were developed in the 80s during the conductive white adhesive QS 1123 TA66 and black general
construction of the ESA Spacelab in order to avoid causing adhesive QS 1123 CEIT, all from MAP Technology.
harm to both astronauts and to the hardware in these manned Adhesive tapes (for flight or temporary use): Scotchtape
compartments. In fact for such projects, as well as the var- No. 5 from 3 M Company; Eccoshield PST CA from
ious equipment and experiments that are supplied to the ISS, Emerson and Cummings.
it is mandatory that all organic materials are tested for off- Paints (can be epoxy-, silicone- and polyurethane-based,
gassing, toxicity, bacterial and fungal growth as well as with or without pigments such as zinc oxide and titanium
flammability. The relevant standards include ECSS-Q-ST- oxide (white) or carbon black (black)): MAP SG121FD
70-21 and -70-29 and NASA-STD-6001. 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-
2.5.2 Classes of Non-metallic Materials ductive paint; all from MAP Technology; Aeroglaze Z306
black polyurethane with high thermal absorptivity from
The space community has adopted several classes of LORD Corporation.
non-metallic materials. As a brief summary some of the vast Lubricants: Dry MoS2 (alone or within a Teflon or
number of suitable materials are listed hereunder. However, polyimide binder), Fomblin Z25 silicone oil from Solvay
it must be recognized that every material will require to be Plastics; see also organic lubricants listed in Table 5.2.
tested to ensure it is suitable for a given project application. Potting compounds, sealants and foams. It should be
Some of the products listed are contained in the datasheets noted that there is a great variation in the effectiveness of
annexed to ECSS-Q-ST-70-71. sealantsthat is, the time for moisture to permeate various
Optical materials such as organic glasses, oxide ceramics sealant materials. Generally silicones are poor sealants for
and amorphous inorganic glasses (need to consider radiation moisture, epoxies are better followed by fluoro-carbons, but
environment to avoid damage and yellowing): Makrolon for truly hermetic seals only glasses or metals can be con-
(polycarbonate) from Bayer; and Spectrosil fused silica from sidered: Solithane 113300 two part urethane from EV
Optik Heraeus. Zerodur is selected by many companies for Roberts; Stycast 2651-40 dielectric epoxy, and other low
the manufacture of space and ground-based telescopes, it is a outgassing products from Henkel or Emerson and Cuming.
lithium aluminium silicon oxide glass ceramic characterized Reinforced plastic composites (usually carbon, glass,
by evenly distributed nano-crystals within a glass phase Kevlar, Zylon or boron bres, whiskers or chopped bres, in
matrix having an extremely low expansion coefcient. a polymer matrix). These are used for both structural and
Optical bres have come into a class of their own and are electronic applications: Applications for reinforced plastics
more frequently selected for box-to-box optical communi- can be found throughout this book by making an index
cations within spacecraft. Each bre, generally made of silica search. Fasteners such as bolts, screws and pins give ultimate
sometimes containing a germanium dopant, is an optical weight saving choices for applications with moderate loads,
waveguide that can transmit light along its axis. Optical bres with the weight of aluminium and 1/5th the weight of steel
are manufactured as a core inside a cladded layer and an outer they are resistant to harsh chemicals. With other com-
protective coating. They have advantages over copper elec- posite materials they have a matched CTE and superior
trical wires in that they are lighter, chemically inert, do not galvanic corrosion properties while being electrically and
corrode and can be easily joined (see also Sect. 6.14). magnetically transparent. Various fastener products from
Important for space use, they are immune to electromagnetic Click Bond, Inc.
interference, and are electrically non-conductive so avoiding Rubbers and elastomers. Used for damping systems,
short circuits and grounding problems. One drawback is the clamps, seals and bladders for fluids: Viton B-910 from
2.5 Non-metallic Materials 43
DuPont; SIFA-35 and AF-E-332 from PSI Corp. It is 2.5.3 Novel Non-metallics
essential to know the properties of these materials over a
wide range of temperatures as will be highlighted by the (a) Vespel (DuPont) has been utilized more frequently for
explosion, and loss of life, of the space shuttle Challenger spacecraft applications during recent years as inferred
(STS-51). This accident was caused by a sequence of events, by the previous paragraph. It is a polyimide-based
but the engineering cause related to the low temperature plastic and has particularly good space properties,
mechanical properties of the elastomeric O-ring seals that such as a low outgassing under vacuum, even at high
interfaced the joint between the two cylindrical lengths of temperatures. The basic material (SP-1) has a low wear
each booster. The seals were designed to be compliant rate and low coefcient of friction in vacuum and
between the two metallic joint faces in order to prevent hot moisture-free environments. The tribological properties
gas, under huge pressure, escaping during take-off. The of Vespal can be further enhanced by the additions of
STS-51 launch took place during an unusually cold period at 15 % graphite (SP-21), 10 % PTFE (SP-211) or 15 %
KSC which caused the elastomer to shrink and loose elas- molybdenum disulphide (SP-3). The low thermal
ticitythis caused the hot combustion gases to escape, burn expansion of all Vespel variants are between 35
the seal and release enough gaseous fuel to cause melting of (SP-21) and 55 (SP-211) ppm/C, relatively low for
the case and the catastrophic explosion. polymers (when compared to values listed in Appen-
Thermoplastics. Often used for multi-layer insulation dices 1C and 1E). Vespel has been used in a spacecrafts
(MLI), second surface mirrors, electronic wire insulation and liquid helium cryostat, as a valve seat, selected for its
sleeving: a multitude of products such as Kapton H, FEP, stability at cryogenic temperatures (see Fig. 5.71). As an
PTFE are produced by companies such as BASF, ICI and example of its versatile temperature use, Vespel has also
Aventis. been demonstrated as a successful non-metallic inter-
Thermosetting plastics. Usually thermosetting polymers face material against aluminium in the design of the
consisting of base, hardener and catalyst which when mixed electric propulsion pointing hold down and release
have a determined shelf life: products include Araldite mechanism of Bepi, qualied from 40 to +150 C
CT205 as used for printed circuit board lamination; and, under contact loads of between 15 and 60 MPa (Janu
Epikote-828. et al. 2009).
Obsolescence can be a problem with inorganic materials. (b) Graphine. This is the thinnest compound known to
Production lines may have ceased due a materials lack of man and comprises of a thin layer of pure carbon. The
demand. The space market is very small compared to other carbon atoms are bonded together in the form of a
industries and some costly production processes may cause hexagonal, honeycomb lattice. It is a two-dimensional
the manufacture of space-approved products to be unprof- carbon monolayer and being only one atom in thick-
itable. For instance, RT/Duroid 5813 was extensively use in ness, it is the lightest material known to man (1 m2
the space industry for ball bearing cage applications, but weighs about 0.77 mg). Graphine is also the most
manufacture of this precision bearing material stopped in the electrically conductive material available and although
mid-90s. Stocks were exhausted just before a replacement it has the potential to be used for spacecraft commu-
material was foundidentied as PGM-HT (a composite of nication equipment, with vast weight savings, it can
PTFE, glass bre and MoS2). Although thought to be a only be produced in very small sizes and quantities.
direct replacement for Duroid, the PGM-HT was found to Applications include durable display screens, electrical
shrink on returning to room temperature after heating under circuits and solar cells. Graphine is produced by
vacuum. After much more searching and testing at the chemical vapour deposition (CVD) and this process
European Space Tribology Laboratory (Buttery 2011) deposits carbon in the form of multi-crystals having a
another material was found: Sintimid 15M/Tecasint 1391 single-layer sheet of carbon atoms. It is these grain
which is currently considered to be a viable alternative to the boundaries that are thought to affect the larger-scale
recent, US-sourced, Vespel SP-3. Incidentally, the properties of the sheet, such as thermal and electrical
RT/Duroid laminates produced by Rogers Corporation for conductivity. The grain orientations and surface
space electronic printed circuit boards (high-power and topology of graphine has been observed by
microwave) continue to be produced, they consist of PTFE atomic-resolution transmission microscopy (Chuvilin
lled with random glass or ceramic micro-bres. et al. 2014). Possibly graphine could also be used to
44 2 Requirements for Spacecraft Materials
the hairs were thought to be governed by the that are either exothermic (at the lower temperatures) or
ultra-efcient application of van der Waals forces. The endothermic (as the material sublimates). The charred
mimicking of gecko hair properties has been attempted in layer is a good insulator, it has a good emissivity and, as
a large Italian study (Gregoratti et al. 2013). Vertically its surface is eroded in the flow of passing air, further heat
aligned multiwall carbon nanotubes (VA-CNTs) were is dissipated away from the structure. Extensive studies
grown on both rigid and flexible metal substrates using into chemical reaction, ablation mechanism and mor-
lithographic techniques: some gecko adhesive properties phology of the resultant black charred layer include those
were achieved but it seems more questions than answers of Asma et al. (2010). A recent EU-funded project called
are recorded at this early stage in the study (Gregoratti Aerofast has provided promising results for two new
et al. 2013): How the do the gecko properties respond to families of TPS materials fabricated from cork supplied
changes in temperature and humidity? What is their by Norcoat-Liege and Amorium Cork Composite
long-term reliability in specic environments? Can such (ACC). The work (Pinaud et al. 2014) also validated a 3D
a material surface be made robust and reliable? And, charring ablation model taking into account the decom-
most importantly, can they ever be fabricated position of the TPS at various temperatures, its decrease
cost-effectively over large areas? in thickness and required dimensions for various aero
(h) Although not entirely a new material to the space dynamical shapes.
materials engineer, cork is nding more interesting and The protective cover attached to US spacecraft bound
novel applications. Owing to the chemical and physical for Mars is called an Aeroshell. The Aeroshell is
structure of corkthe natural bark growing on the trunks constructed from an aluminium honeycomb, covered
of cork oak trees (quercus suber)this materials has with graphite-epoxy face-skins that are in turn covered
found important uses as ablative material on re-entry with a phenolic honeycomb containing ablative mate-
vehicles. As listed in Table 2.7, cork has a low density rial. The ablator is a unique mixture of specially graded
and low thermal conductivity. Corkcalled natures cork granules, binder and silica glass microspheres. The
foam by NASAis 90 % gas and is essentially a struc- chemical formulation of this mix is designed so that as
ture made up of dead cells, each having a polyhedron of the probe or landing craft enters the Martian atmo-
only 3040 m in diameter (Coelho 2009). Ablative sphere, friction causes the ablator to undergo chemical
materials protect space vehicles from the severe aero- reactions and char. This has the effect of dissipating heat
dynamic heating during launch and during re-entry into which is taken away by the hot wake of gas behind the
the atmosphere at velocities greater than 7 km/s. vehicle. Thermal protection system must be able to
The ablative materials used on the re-entry capsules of resist temperatures which could reach 1600 C during
rst US manned missions, Mercury, Gemini and Apollo, entry into the Martian atmosphere, and yet maintain a
were based on cork. It was used on the solid rocket temperature of less than 125 C inside the probe. The
boosters of the Shuttle and applied to the booster tiles can be formed and individually machined on all six
nose-cone and frustum of Delta launch vehicles. In faces, then bonded to the probes surfaces with silicone
Europe, cork-based ablative materials were applied to blue, ESP 495.
the front and back covers of Beagle-2, are used on (i) A material nding many new applications is expanded
Ariane-5, the Atmospheric re-entry demonstrator and polytetrafluoroethylene (ePTFE). The microstructure
Mars Express (Bouilly et al. 2006, 2013). Illustrations of mimics some cell structures found in nature, as seen in
cork being applied to space hardware are shown in Fig. 2.26. The main usage of ePTFE in space applica-
Figs. 2.24 and 2.25. tions is as a low dielectric constant (1.3) insulation
Cork is re resistant and chars rather than burns. As a material for high-speed digital coax cables. Here,
thermal protection system (TPS) it undergoes several ePTFE provides high signal quality and optimum per-
physiochemical changes each of which are formance in miniaturized systems where conductor
heat-absorbing. Initially it heats up, then the front sizes as small as AWG 42 can be manufactured. As this
exposed surface is thermally degraded (charred), gasses material is inherently inert, thermally resistant and
released during pyrolysis pass though the charred layer hydrophobic, it can also be processed into
which becomes thicker. Gas reaction with the carbona- high-flowrate microporous membranes. Applications
ceous layer creates channels through which gas and will be as lters within manned spacecraft modules for
reaction products are released as ablation vapours. This air and potable water purication systems. ePTFE l-
releases thermal energy into the air flowing past the aments have recently been incorporated into a
structure. The cork composite is formulated so that as it bre-reinforced matrix (such as PEEK) as a reservoir
chars, the interface between the advancing char and the for solid lubricants. The porosity of ePTFE, as seen in
underlying (virgin) material causes chemical reactions Fig. 2.26, can be processed to vary from 5 to 90 %.
2.5 Non-metallic Materials 47
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
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
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
discussed, in future missions assembly may occur on the
surface of our Moon and possibly other planets). So far, little
One named Ikaros recently passed Venus and JAXA work has been specically dedicated to intra-vehicular and
are now considering future sail missions, powered only extravehicular activity (IVA and EVA) joining and cutting
by the Suns light particles, to Jupiter and the asteroid issues in space. This section will revisit some of the ndings
belt. of Dunkerton et al. (2001) and include more recent welding
developments having relevance to space applications. By far
Returning to LCP wire, Vectran is said to be the only the largest contributions made to date, on the development of
commercially available melt-spun LCP bre in the world; welding and brazing technologies for use in space (Dzha-
weight-for-weight this bre is ve times stronger than steel, nibekov et al. 1991; Paton 2003), have been made by Russia:
has a high abrasion resistance, cut resistance and has good
mechanical property stability over a wide range of temper- 1969Vulkan, for electron beam (EB), low-pressure plasma
atures. These bres can be coated with copper, nickel or and consumable electrode welding
silver and is available (Liberator Conductive Fiber) in 1979Isparitel, the depositing of thin coatings by conden-
bundles comprising of 20, 40 and 80 individual laments. sation and evaporation
This is a new product and has been used as a woven braid 1983Isparitel M, deposition of braze metals
material for cables. The new braid material will be of interest 1984development of a versatile hand-held tool for cutting,
to spacecraft harness fabricators due to its substantial weight brazing, welding, heating and coating
savings over copper braid as illustrated in Fig. 2.27. 1988YantarEB coating and welding
1990Universal Hand-held 3 kW manual EB gun used on
Mir Space Station. (Figs. 2.28 and 8.33)
2.6 The Potential for Welding
and Joining in a Space Reference to the numerous metallurgical experiments
Environment performed in the space environment by astronauts and cos-
monauts, for instance on Skylab, Soyuz, Mir and the Salyut
2.6.1 Background Considerations spacecraft, have been compiled by Flom (2005) and by
Paton in his reference book (2003). The majority concerned
Large dimensions and mass mean that structures such as the the melting and joining of metals in space. Flom (2005,
International Space Station (ISS) cannot be transported as 2006) noted that the vacuum of space is higher than a typical
one complete structure into low Earth orbit. They must be Earth-situated vacuum chamber can provide so that electron
launched in sections and assembled in space (and as already beam welding would require very little power and cathode
50 2 Requirements for Spacecraft Materials
2.6.3 Expectations
control, and inspection and testing. Standard metallurgical Lloyds of London or the Underwriters Laboratories). ISO
processes rely on the generation of pre-production test 9000 is an auditing type of standard that requires companies
samples and metallography for the setting up of optimum to document all aspects of their operation. A company
process variables, generation of workmanship standards, and normally begins by documenting all their procedures as
the determination of accept/reject criteria for porosity or standards. They write, or rewrite, their quality manual to
inclusions within materials and joints. Inspection functions, cement their newly established procedures, which should
frequently referred to as quality control, might utilize met- enable them to consistently produce products to the quality
allographic techniques for establishing the cause and possi- levels demanded by their customers. All changes to manu-
ble acceptability of mechanical damage. Heat-treatment facturing and testing procedures must be properly docu-
condition, approximate chemical composition, dimensional mented. In some companies where material processes have
tolerances for machined or electroformed hardware having been performed for many years, the skills of the process are
complex shapes, etc. can each be evaluated or inspected on a often passed down from one operator to another. For
sampling basis by means of metallography. Similarly, instance, the precise times and temperatures for a brazing
organic materials and their processes will be assessed by the operation may not be documented in a controlled procedure
kinds of tests discussed in Sect. 3.2.5. Material traceability is and, instead, the writer has seen a broad range of tempera-
important; this will usually start from the procurement tures and times handwritten on a piece of paper glued to the
source and be controlled by documentation to stores and side of a console above the brazing furnace. Similarly, the
then to the assembly area. Materials with a limited shelf life, cleaning operation may not be documented and different
such as adhesives, paints, potting compounds, and confor- operators are observed to clean printed circuit board
mal coatings should be kept in containers that have been assemblies in a bath of solvent for different times, some only
labeled by QA staff to show product name and code, date of swilling the board around while others use brushes to
manufacture and expiry date. In addition these materials remove contaminants from difcult-to-reach surfaces of the
should be linked to specications that detail how they should board beneath components. And no-one really knows when
be stored (temperature limitations and possibly refrigeration) the solvent is so contaminated that it has to be replaced. In
and this documentation may specify how, or if, measures can these cases the procedures manuals will be difcult to write,
be made to extend their shelf life. as these will set the standards to which all personnel must
Critical materials will probably be those requiring out- work, and there can be many points of conflict for individ-
gassing and flammability testing or stress corrosion testing uals. We do a good job, have done for ten years and no-one
so that they can be validated for a given space project. complained before, It was done for OTS-2 and that satellite
Critical processes are those that are irreversible (product was still working 11 years after its two-years designed life!,
cannot be repaired) and when that process is applied to very etc. These comments come from people who have yet to
expensive and complex assemblies. They may be considered appreciate the concept of QA. Tension can certainly run high
critical if the processed part or sub-assembly cannot be when the rst set of procedures is produced, despite all the
inspected. Processes such as coating, soldering are said to be assurances that it is the starting point and requires feedback
veried if technology samples are evaluated by testing and from staff and operators in rening the wording and content.
found acceptable. Some operators will think it is an administrator in the ofce
As an aside, it is interesting to observe that in European that is telling them how to suck eggs, or just wishing to
industry, many materials engineers, who considered their create paperwork for the sake of it. Again, these people have
careers to be limited by their purely technical role, have preconceived ideas and are reluctant to see what something
turned to QA engineering. It has been remarked that they have new may be able to do for them. Someone has to produce the
a reasonableness with regard to taking into account many written procedure, and it is seldom that this is written by the
ill-dened influences when making a decision, and therefore doers. They are, however, more likely to provide an input
make good QA managersa somewhat misguided inference to the draft when they can see what is intended. QA is not
to materials engineering as being an inexact science! about checking or everyone following one set of rulesit is
All companies contracting to the space industry, whether about effective management which will enable problems to
as manufacturers, suppliers, or test laboratories, are aware of be identied, analysed, and rectied. It does not mean that
the fact that quality is synonymous with survival. The there will never be mistakes, but that they will be minimized
implementation of slogans such as higher productivity, and have less chance to cause schedule and launch delays.
lower costs, on-time deliveries, and increasing product Materials and Failure Analysis laboratories can be audited
quality are strategies considered essential in order to com- and certied to the international management standard, ISO
pete in the international market. One way to help achieve 9001. This is widely used by space manufacturing compa-
these goals is to apply for registration, or seek to be qualied nies for managing the quality of the lab service. However,
according to the ISO Standard 9000 with a registrar (e.g. this standard does not evaluate the technical competence of
3.1 General Product Assurance and the Role of Materials 57
the laboratory or that its test procedures, inspection methods devices, and additional structures, to improve the overall
or calibration are suitably accurate. A more suitable standard system design.
to address these aspects is ISO/IEC 17025; this standard will The reliability of electrical and mechanical components,
also assess the competence of the lab staff, whether the test metallic joints, and structures are calculated from test pro-
methods are valid, equipment is calibrated to national stan- grammes that incorporate the stresses and environments
dards and the suitability of the labs own QA system. expected in spacecraft service. Such programmes rely
Implementation of the system of auditing required by ISO heavily on the results of materials lab work which are
has highlighted and then strengthened some weaknesses in brought together in report forms referred to as Destructive
the smaller space manufacturing companies. It is also an Physical Analysis (DPA) and Failure Mode Analysis
effective tool to build trust and a higher degree of coopera- (FMA). The objectives of the DPA are to catch unapproved
tion between management and the shop floor because it design changes or unacceptable workmanship. Normally, a
keeps all procedures up to date, everything is written and small number of parts are selected at random from a batch of
documented, and employees know what their job is and mechanical or electronic parts and are sent for evaluation.
where their boundaries and responsibilities lie. Another They provide the baseline for the quality of the item which
advantage is that the number of customer visits for quality will then be submitted to both static and dynamic tests (e.g.
audits is reduced when the ISO certication is accepted in vibration, shock, thermal cycling under vacuum) as part of
place of the visit. Machining rms, heat-treatment, plating the reliability test programme. The FMA will subsequently
shops and painting facilities are the kind of companies likely be made to identify the various actual failure modes expe-
to benet as more emphasis is made of in-line process rienced or to pinpoint possible failure modes. Several
controls. This has been found to reduce the scrap rate at nal examples of these kinds of materials examination will be
inspection and thus offer quality products at lower prices. illustrated in subsequent chapters; they relate to the relia-
In the space manufacturing industries, the introduction of bility of spacecraft structures, fasteners, battery cells, elec-
ISO 9000 and its related specications, as well as the stan- tronic components, and the like.
dard space project contractual documents such as the In the case of manned programmes, such as the
European Cooperation for Space Standardization, ECSS now-retired Space Shuttle Orbiters, Spacelab, and the Inter-
series, has helped not only quality but has also ensured that national Space Station (see Fig. 3.1a, b), it is also essential to
state-of-the-art technology has been exploited. In recent evaluate all materials used from the safety viewpoint, and
years this has included computer-aided design and analysis tests are conducted for flammability, toxicity, and odour.
(to model stresses in a system, estimate load-bearing capa- Hazardous metallic materials such as beryllium and mercury
bilities, etc.), manufacturing techniques and tooling (CNC are generally excluded from designs, but with additional
machining cuts steps between design and manufacturing, safety precautions they might be permitted for applications
computer-controlled lathes are accurate, flexible, and fast to such as X-ray windows or charged particle thruster engines.
set up), computer programs (for statistical process control Each of these materials would be tested in its application
and computerization of test and inspection equipmentthis set-up. Another aspect of safety includes the testing and
can measure tighter tolerances), and materials (e.g. rejection of metal alloys found to be susceptible to stress
high-alloy steels with better cleanliness, fewer inclusions, corrosion cracking (this test is described in Sect. 4.5; it
and a lower gas content to enable higher strength levels). 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-
3.1.3 Reliability and Safety trols). The maintenance of reusable spacecraft, like that of
aircraft, has always been considered superior to the mainte-
Reliability activities are directed toward ensuring that the nance of vehicles for other modes of travel. The amount of
design, selection of materials, and choice of fabrication time and effort expended by spacecraft designers, mechanics,
methods for the various spacecraft subsystems are based and inspectors in providing a high level of safety is unsur-
upon sound principles of reliability engineering and analysis. passed by any other industry. Yet, although few in number,
In recent years the concept of reliability has acquired a new there have been occasions where mistakes were made which
dimension which encompasses the generation of numerical resulted in accidents. No matter how small the calculated
predictions of the reliability of a spacecraft system. This is probability is that an adverse event will occur, there is still
made possible by the study of individual components, metal that chance that it will happen. To increase the safety margin
joints, etc. under environmental testing or during real life so against structural failure, spacecraft designers employ
that their life times can be predicted. Single-point failures redundant load paths and redundant systems. Finally, safety
are generally not permitted on spacecraft; they will be also includes the so-called fracture control procedures. Since
replaced by redundant electronic circuits, mechanical all engineering materials contain flaws introduced during
58 3 The Integration of Materials into Product Assurance Schemes
processing of the basic material, or fabrication of piece-parts, investigation. Fracture-control screening is generally per-
it is essential that analyses and tests are performed to establish formed for the complete spacecraft structure. The structural
the limits of detection of these flaws, using nondestructive test items are those parts and assemblies that have been designed
(NDT) methods such as eddy currents, ultrasonics, dye pen- to sustain loads or pressures, provide stiffness and stability, or
etrants, and radiography. Again, metallography is employed provide support or containment. These may be: pressure
to quantify the actual sizes and forms of cracks under NDT vessels, composite structures, joints, and other load-bearing
3.1 General Product Assurance and the Role of Materials 59
parts. ESA space vehicles, and, when implemented, are also applications, mechanical devices, propulsion systems, and
able to satisfy the NASA approval authorities (by means of electronics. As previously stated, the ordinary materials
the standard that denes the project safety programme and its known and well-proven in aircraft applications are most often
safety technical requirements; ECSS-Q-ST-40). 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
3.1.4 Materials and Processes optimized grain size and with unusual surface protection
layers. Many changes can be seen during visits to the fac-
Spacecraft project management will follow the contractual tories and workshops involved with the manufacture of space
requirements formally agreed between customer and usually hardware. Often the remnants of outdated methods are visi-
the prime contractor. Sub-tier contractors will often have those ble, capstan and turret lathes lie idle in corners or are seldom
same requirements flowed down in order to ensure a coherent used by the skilled persons needed to operate them. The
approach. Materials and processes will be well-dened and newest CNC (computer numerical control) turning machines
cover such topics as selection criteria and rules; evaluation, are now in the spotlight with driven tools, automated by
validation and qualication testing; the generation and gantry robot-handling systems and many other newly avail-
maintenance of Declared Materials, Mechanical Parts and able accessories for unmanned operations. These facilities
Processes Lists; and procurement and receiving inspection have been purchased not for high-volume runs, where they
requirements. In Europe, these requirements often follow the can be left permanently set up for one part, but for
ECSS-Q-ST-70 standard, the latest issue should be consulted multi-operations where lathes have great flexibility and can
for details. Full adherence to the ECSS-Q-ST-70 requirements produce the large variety of turned parts required for space-
is preferable, but may be too demanding for some low cost craft. The tool magazines may contain up to 90 different tools
projects such as those carried out by universities where cost is and, by means of interactive programming, even inexperi-
the driving factor rather than highest reliability. Appendix 8 is enced operators are able to soon learn how to programme
proposed for the basic management of low cost Materials and complex work pieces. The small batches of turned compo-
Processes activitiesthis is a rudimentary scheme that may nents having many shapes, as are required for space projects,
need to be enhanced as the project develops. can now be made on a single turning machine where the
Examples of ctitious Declared Materials Lists and metal input is in bar or billet form. In conventional machining
Declared Processes Lists are shown in Appendices 10 and 11. the use of multiple machines has become inefcient for small
The utilization of such DMLs and DPLs are considered batches. Furthermore, when high surface nishes and close
essential by the writerthey follow the hereditary of the tolerances are required for spacecraft mechanisms, grinding
space subsystem and can be immediately recalled in case of can be made with a new generation of tools for instance
failures when material composition will need to be identied. having CBN (cubic boron nitride) inserts. Ideally, the com-
The majority of materials used in spacecraft manufacture plete grinding sequence can be done in one operation to
are selected from commercial metallic alloys, ceramics and reduce time and handling. Swiveling grinding heads are
polymers that can be processed using conventional tech- available for aerospace parts having powerful 40 kW grind-
niques. This is due to four main reasons: ing spindles which offer a speed range of up to 9000 rpm.
A compensating system is available for high-accuracy parts,
Contractors prefer to stick to tried and trusted which negates the effect of thermal expansion. These are
technologies. being used for the manufacture of lightweight spacecraft
The validation of new state of the art materials, or the mirrors. Depending on the alloy being turned or ground, a
verication of new processes for a given space pro- standard coolant or oil can be used. This is supplied under
gramme, usually require much additional effort (and high pressure, and the coolant vapour is collected within the
costs) in terms of generation of quality documents, pro- enclosure of the working area by a vacuum system. It can be
cess controls, reliability test data, etc. said that modern technology has redened the lathe and
The quantities of materials utilized on spacecraft are grinder, with 8760 working hours every year these new,
measured in kilograms and not tens of thousands of tons automatic machines can continued without the need for
as would be preferred by commercial material producers. sleep. Joining techniques have also advanced and whilst, for
The time from the start of a new space project until its instance, the main structural alloy All 2219-T851 for the ESA
launch is about 4 or 5 years, and contracts may be too Spacelab project (Fig. 3.1a) was generally TIG-welded or
late for any new material development. electron-beam-welded in the early 1980s, the same wrought
alloy is being jointed by the recently evaluated Variable
Nevertheless, during the past decade we have seen the Polarity Plasma Arc (VPPA) welding, providing far superior
development of new, advanced materials for use in structural weld quality for the Columbus/Space-Station elements
60 3 The Integration of Materials into Product Assurance Schemes
(Fig. 3.1b) and the Ariane V propellant tanks (see also projects employ dedicated materials engineers for this
Fig. 4.57). It was thought likely that VPPA welding would be function. To be effective, each spacecraft material and
the only suitable process for welding the wide range of alu- process must be documented and then approved by the
miniumlithium alloys that have recently become commer- M&P engineerthis task will be facilitated by the exis-
cially available (Ilyushenko 1993). Conventional welding tence of a specialist support team having access to an
techniques for AlLi alloys produce unacceptable levels of adequately equipped and manned materials test laboratory.
porosity due to the presence of lithium compounds that may Concerning processes, it is probable that the introduction
be hydrides, hydrocarbonates, and hydrated oxides. During of computer-aided design now means that spacecraft
welding, such compounds decompose with the evolution of sub-assembles and piece-parts are being fabricated to the
gas bubbles. They are entrapped by the metal being solidied closest tolerances ever achieved. To illustrate the accu-
and form pores. Nowadays Friction Stir Welding (FSW) has racy demanded of modern machining capabilities one can
become the preferred method for joining AlLi space struc- consider the unfortunate situation of the flawed primary
tures. The longest such weld being 445 inches, by Lockheed mirror of the Hubble Space Telescope (HST). The prime
Martin on the Orion spacecraft (for joining the forward cone objective of the HST mission was to obtain images of
assembly to the aft barrel assembly). Developments in FSW astronomical objects in approximately ten times sharper
at the NASA Marshall Space Center have substantially detail than that obtained by ground-based telescopes. Or,
reduced weld defects so that costly and risky repairs on Al in other words, HST is designed to point on a celestial
Li, superlight spacecraft tanks and structures (Russell 2014). target with a stability measured in tens of nano-radians:
FSW is termed a key technology for primary aerospace roughly equivalent of keeping the HST line-of-sight on a
structures in Europe and this clean, cost effective joining human eyeball at a distance of 320 km. The HST 2.4 m
technology has been transferred to the manufacture of many mirror was designed to be a precisely calculated hyper-
European spacecraft tanks and structural systems (such as boloid. Although the rst mirror flown was actually
AlLi cryogenic LH2 and LO2 tank demonstrators, the upper smooth to a precision of 1/64 the wavelength of light (or
stage of Ariane V and its main motor thrust frame (Brooker one-millionth of an inch), a calculation error had caused it
2001)). FSW is described in more detail in Sect. 4.25. to have been fabricated with a curvature that was too
The following is a generalized account concerning the shallow with a total centre-to-edge error of about 2 m
acceptance of a new alloy product for application in a (1/50 thickness of human hair). Results from the mirror,
spacecraft: during the period 19901993, showed that light rays
hitting the mirror edges eventually focused to a point that
If the new metal alloy, no matter how promising, is was slightly away from where light rays from the centre of
developed by an engineering company, it will not be used the mirror focus: a defect called spherical aberration. The
on a space project unless an individual, or team of mate- shuttle mission to revisit the HST and to install new solar
rials and design engineers related to the project, recognizes arrays and new instruments, and to correct the defective
its potential merits. They need the courage to evaluate it, mirror was made successfully in December 1993.
and use it if the new alloy lives up to its promise. That
alloy, as with all previous metals selected for space use, The space industry places extraordinary demands on the
would have been subjected to a sequence of evaluation properties, performance and cost of materials as these are
tests specically called up by the various technical disci- intertwined with considerations of safety and reliability. In
plines of the projects PA team. The Materials and Pro- this book, the testing and evaluation of materials will mainly
cesses (M&P) engineer reviews the evaluation test results focus on the various phases of selection of individual
and also the conclusions and recommendations given by materials for a particular design based on properties and their
the other PA engineering disciplines. The new materials performance in a particular environment. Cost may be
will only be approved by the M&P engineer once he is important, but generally the best material for the job will be
convinced that the alloy is totally suited for its application selected bearing in mind that the price of procured materials
and that its method of fabrication could not adversely for a complete space project will probably only lie within 1
affect the spacecrafts integrity. Metallography will be and 10 % of overall costs. As will be noted in the chapters
utilized during all the evaluation tests (probably involving relating to failure analyses, it is often the environment that is
stress corrosion tests, accelerated fatigue, testing for ultimately responsible for any particular hardware failure.
hydrogen embrittlement), and this technique will often These environmental factors can be extraordinary difcult to
establish certain test acceptance criteria. comprehend and it is essential that engineers from all dis-
The imposition of an M&P policy is occasionally met with ciplines collaborate to both prevent, and where necessary,
some reluctance by industry, and only the larger space solve problems associated with failures. The complexities
3.1 General Product Assurance and the Role of Materials 61
associated with launcher and spacecraft failures necessitate atmospheric pressure and, during such tests, the nozzle
an interdisciplinary approach to failure analysis where the was found to be subjected to axially tensile forces. The
materials engineer will play an important and active role. failure mode that led to the loss of Flight 157 was a
Two immediate examples come to mind: decrease in pressure around the nozzle as the rocket
gained altitude. That loss in pressure during launch
The Space Shuttle Challenger accident in 1986 was ini- placed the nozzle under gradually increasing axially
tially observed to result from puffs of black smoke compressive forces and this resulted in a deformation of
originating from a leaking lower seal joint on its the upper portion, causing it to buckle. Distortion of the
right-hand side metallic Solid Rocket Booster (SRB). nozzle increased the thermal flux in a localised area until
The seal, actually a complex design incorporating an the nickel-based alloy melted and the cooling system was
elastomeric O-ring, is needed to prevent hot gasses perforated. This launcher was immediately destroyed by
escaping from the highly pressurized rocket motor during a command from the range safety ofcer. The main les-
take-off. Within a minute of launch a flame grew from the son learned from this experience is the fact that while the
leaking pressure seal, the flame impinged on the sur- materials may have been correctly selected, qualication
face of the External Tank causing it to breach and release authorities were unable to reproduce the flight environ-
hydrogen. Total failure occurred within milliseconds ment. The reducing atmospheric pressure could not be
once the adjacent oxygen-containing Intertank was also included as part of a launch simulation.
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 3.1.5 Component Part Selection,
two metallic segments of the failed joint (they has grown and Procurement
due to previous useage of this reusable SRB, but were
still within tolerance); the elastomeric O-ring seals The work of the component engineer is analogous to that of
resilience at an unusually low launch temperature (the the M&P engineer in that electronic components will be
Florida launch site was at 1 C, whereas most launches selected and procured from the types already listed as either
occur at 24 C when the O-ring is ve times more having been tested and qualied for space use, or those
responsive to return from its compressed to its uncom- nonstandard devices that will require qualication by testing.
pressed shape); the possible presence of ice within the The major objective of the component engineer will be to
joint volume (it had been raining before launch and later ensure that electric, electronic and electromechanical parts
tests showed ice formation to be a possibility); stresses (EEE parts) are selected and procured for space projects,
during SRB ignition and ring can slightly open the whether these are launch vehicle, space probes or satellites.
seals tang and clevis and, at the low temperature, this The format for compiling component data as a basis for
gap may not have been lled by the O-ring due to its approval of a component, or EEE part, for a space project is
lower response time to reach the uncompressed dimen- by means of a continually updated and assessed Parts
sion. Elastomers shrink and loose elasticity at low tem- Approval Document. Analogous to the Declared Materials
peratures. The mechanical properties of the O-ring List, components engineers will review each projects
material were known for all temperatures of operation. Declared Components List (DCL), identifying all compo-
However, the importance of the change in seal gap nents needed for the project, their full specications, quali-
dimension, together with the reduced elasticity of the cation data and approval rating. Screening and Lot Approval
O-ring material to ll the gap at the low environmental testing will be specied and particular attention will be given
temperature, were not realised by all concerned in the to radiation verication testing (RVT) with respect to the
launch decision. anticipated dose calculated for the component during its
The other example, in 2002, concerns the failure of the particular in-orbit or exposure life to electromagnetic radia-
Ariane 5 ECA rocket (Flight 157) when it suffered major tion (i.e. solar radiation consists of ions and electrons plus
trouble with its Vulcain 2 main stage engine before going trapped electrons and protons in the van Allen Belts and in the
out of control and exploding over the Atlantic Ocean. South Atlantic Anomaly, see also Sect. 8.2.2). These and
Failure analyses reported by Drogoul (2005) found the other forms of radiation data, including cosmic rays and solar
problem to have been caused by a deformation to the proton events, both able to knock out spacecraft, can be read
engines nozzle (Fig. 2.13 shows the engine and location online from sites such as http://spaceweather.com and http://
of a helical structure of welded Inconel tubes through www.spenvis.oma.be/. Excepting the specic electrical tests,
which hydrogen flows to cool the nozzle). Ground test- most components will be subjected to the same vibrational,
ing, as shown in Fig. 2.13, was of course conducted at shock, and thermal cycling tests as were described in
62 3 The Integration of Materials into Product Assurance Schemes
Sect. 3.2.2. As mentioned, tests may be required involving an together with connectors, hook-up wires and cables and
assessment of electromechanical devices when degradation coating/potting materials. Such boxes are heavy and require
resulting from radiation exposure may be expected (the special handling procedures. Boxes have been known to be
in-orbit levels of electromagnetic, cosmic, and particle radi- constructed from aluminium, magnesium and steel materi-
ation do not affect metals). All classes of components ranging als. The thickness of the walls will need to satisfy
from the relatively simple resistor to the highly complex mechanical design requirements to ensure no damage during
microprocessors and memories will require to be evaluated vibration and other loading. Radiation analysis should also
by a comprehensively equipped laboratory. In addition, be madeoften the wall thickness is selected so that total
advanced test equipment is needed for the characterization of radiation within the box will not exceed 10 kRad(Si).
very large-scale integrated circuits. Typically 2 mm thick aluminium is presently used as the
Space-qualied components are generally covered by norm for shielding electronic componentsas some FPGA
ESA SCC specications. These specications also control packages can be badly affected with less than 100 kRad
the selection of peripheral parts for a spacecraft, such as dosages. Shielding studies are continually performed in an
connectors, wires, cables, and fuses. It is no use sending a attempt to nd a light-weight solution because of the
multi-million pound spacecraft into orbit if, during its launch inherently high mass of traditional aluminium boxes. Some
phase, connector pins break, fall out, or wires fracture. Here ideas are to use CFRP composites that contain carbon
the materials specialists will assist the components engineers nanotubes together with diamond like coatings deposited
in the choice of materials and processes best suited to sur- onto tungsten foils. Aluminium alloy boxes are often surface
vive long-term duration in space. With the ever-increasing treated with Alodine 1200 and later, nally painted black or
complexity of space systems and the demands for a reduc- white with an Aeroglaze or MAP product. Electrostatic
tion in the size and weight of interconnection devices (e.g. precautions are always necessary to avoid ESD and
for the ESA Cluster spacecraft the 50 km of harness was operators/handlers need themselves to be grounded by
made from AWG 28 stranded wires) there has been a need to means of wrist straps and conductive shoes.
make even more stringent materials tests to ensure the Wires and cables used in electrical applications such as
durability and electrical integrity of interconnectors. cable harnesses and wires within electronic boxes are strin-
Micro-D connectors have been developed specially for the gently controlled for space applications. A multitude of
space market. They are small and light, and have higher variants exist and the space-approved types are specied by
density contact congurations than are available in tradi- ECSS documents. These specications dene the require-
tional rectangular connectors. Some have aluminium alloy ments for a particular family of wires or cables. They include
shells with a nal nish of gold plating; these are nonmag- the maximum rating, physical characteristics, design (i.e.
netic and offer electromagnetic and radio frequency shield- shielded or not, AWG wire size, number of strands and
ing (beware of commercial connectors which are totally physical dimensions). Extensive qualication tests (involv-
unsuited for space applications owing to their ing accelerated ageing at 250 C, mechanical tests such as
cadmium-plated steel shell, sublimation, and magnetism abrasion and notch resistance, solderability, possibly
problems). The new designs also possess a silicone elas- flammability, etc.) are performed according to the relevant
tomer compression interfacial seal to provide a moisture and detailed specication in the ESCC 3901/xx series. The
humidity seal between each contact and the contacts and the most-use conductors for larger gauge wires (AWG 1624)
shell (the silicone passes outgassing requirements whereas are silver-coated, soft or annealed oxygen-free high con-
the organic insulants incorporated in commercial connectors ductivity copper. The smaller gauges (AWG 2630) usually
can outgas under vacuum). require silver coated high strength (alloyed) copper strands.
The design and selection of materials for electronic boxes The European space standards require that the silver has a
rest with the industrial companiesand the specic space minimum thickness of 2 m in any one location; this is to
project contractual requirements. Usually the housing is avoid the problem of red plague as will be discussed in
machined and/or assembled from a stress corrosion resistant Sect. 6.8.3. In recent years silver-plated aluminium con-
alloy listed in Table 2.2. Care must be taken to avoid ductors have been selected for power distribution systems on
unwittingly incorporating nickel plating on the surfaces of satellites where weight saving has been essential. Insulation
any electronic box designed for magnetically clean scientic materials are generally, polyimides, cross-linked ETFE, and
spacecrafta costly problem involving the use of electro- expanded-PTFE. For low Earth orbit (LEO) the cable jacket
less nickel is described in the case study presented in must be resistant to atomic oxygen and a dual insulation of
Sect. 4.10.2. Electronic boxes will contain runners, frames polyimide/fluorothermoplastic has been found suitable.
and spaces that locate component-populated printed circuit In recent years, as system integrators attempt to manage
boards. Boxes may only house 3 boards, but some complex higher reliability from their system components together
boxes have been known to house up to fty boards with the need for greater cost-savings, there has been much
3.1 General Product Assurance and the Role of Materials 63
The most usual lifting gear uses steel wire, although some By changing the number and arrangement of wires in each
systems requiring less load-bearing capability favour the use bundle (for instance by making ropes with ne metal bre
of synthetic bre ropes and plastic-covered wires. Synthetic cores, by laying the bundles helically with either right lay or
ropes are made of Kevlarthese now compete with steel left lay, or selecting higher-strength steels for the outer
wire and cable, as they have a lower weight and can be bundles in order to achieve optimum wear resistance) it can
operated using decreased motor and hoist hardware sizes. be seen that the number of possible wire rope combinations
Similar, braided Kevlar is used for tethering small satellites becomes profuse. However, only a small fraction of these
or space experiments to the Space Shuttlethis tether has a constructions are sufcient for industrial purposes in order to
diameter of less than 2.5 mm and can withstand loads of provide strength and stiffness with flexibility in bending.
180 kg (approximately 700 MPa) as the hardware is Expensive alloy steels with their more complex heat treat-
deployed from the Shuttle and back by means of a simple ments are seldom selected for slings and holding devices.
reel (Scala 1996). In comparison, the steel wire strands that However, they are seen in the cleanrooms of spacecraft
are in common use for hoisting applications will have integration halls, as these alloys do not corrode and are less
strengths in the order of 2000 MPa, although ropes made of likely to generate dust or wear particles. In less clean envi-
even higher strengths are available (Shipley 1988). ronments these wires are essentially a plain carbon steel of
Wire rope is constructed from twisted bundles containing carefully controlled composition and a carbon content gen-
strands of small wires having the same cross-sectional area erally close to the eutectoid. The patenting process of
of metal. Whereas a solid wire, for example having a isothermal transformation and work hardening resulting
diameter of 10 mm, is very stiff and will stand relatively little from successive wire drawing is largely responsible for
back-and-forth bending before breaking, the same aggre- achieving the 2000 MPa strengths.
gate sectional area of metal in stranded form will be found to The spacecraft and materials handling systems such as
have gained greatly in flexibility, tensile strength, and life. hand hoists, power hoists, wall and floor cranes, grabs,
66 3 The Integration of Materials into Product Assurance Schemes
slings, and holding devices must be fully maintained to The laboratories, which perform testing and failure
written procedures that also include inspection steps for analyses, are run to give rapid functional support particularly
signs of wire spreading, wear damage and overloading. in the form of reporting to Critical-Design Reviews, Material
Maintenance schedules are attributed to wire ropes which Review Boards, and Failure Review Boards. The items of
will be checked possibly at least once per month for signs of spacecraft hardware that may be submitted for evaluation
marring, stretch, and other deformations. The nature of rope can range in size from thermocompression weldments that
construction provides many parallel load paths, giving it a interconnect 25 m diameter gold wires within miniature
high level of redundancy. The essentially helical nature of electronic circuits, to the massive 4 m diameter forged and
rope construction usually causes the outer bundles or strands heat-treated Spacelab structural rings, the largest to have
to be marginally greater stressed in service than the inner been fabricated in Europe from a high-strength aluminium
strands and these will be preferentially broken in situations alloy.
of overload. Strands in the bundles then slide slightly under The role of metallurgical tests in the somewhat overlap-
surface shear forces to take up the applied load from the ping range of spacecraft project PA activities can be illus-
broken strand. There is no stress concentration to promote trated by the following list:
the fracturing of adjacent strands, and the location of the Quality assurance (QA): Metallography and mechanical
break can be readily found during routine visual inspections. testing for the setting up and control of optimum process
Such inspections will also check for damage from wear, variables (such as those encountered during welding, braz-
crushing, corrosion, nicks, and cuts. Owing to the often ing, heat treatment, electroplating, etc.), and for the gener-
harsh duty cycle of hoisting ropes it is inevitable that some ation of workmanship standards. In-depth testing and
degradation will occur over time. It is the ground-handling qualication programmes for the approval of industrial
managers responsibility, guided by the materials engineer, fabrication lines (such as for double-sided and multilayer
to monitor degradation and remove rope and lifting gear printed-circuit-board production, automatic machine solder-
before it becomes a risk to both the facility operators and the ing, and critical process assessment).
spacecraft hardware. The author is not aware of any space- Reliability and safety: Hardware or unit-level testing
craft handling failures attributable to rope or cable structures, directed toward ensuring that design, selection of materials,
even though these are subjected to high load capacity and choice of fabrication methods are based on sound
combined with flexibilityprobably due to the inherent principles of reliability, engineering, and safety analyses.
redundancy and damage tolerance of the numerous single Prediction of lifetimes, failure-mode analyses, and identi-
woven wires. Other commercial industries have suffered cation of single-point defects. Hardware is subjected to
from catastrophic steel cable failures, most often attributed both static and dynamic tests (e.g. high-temperature ageing,
to corrosion and local fatigue (Piskoty 2009). 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-
3.2 The Materials Laboratory ing of new materials and processes before their being
ascribed to spacecraft production, screening of metals for
3.2.1 Major Objectives of Laboratory specic applications, and the review of certain QA functions
(listed above). Metallography is utilized during the majority
The company materials laboratory can represent over one of evaluation tests (possibly following material or intercon-
million pounds of investment in facilities that enable the nection exposure to corrosion, mechanical loading, acceler-
provision of services to spacecraft projects and associate ated fatigue, testing for hydrogen embrittlement, or a
R&D programmes. It is preferable that the work is con- combination of these environments).
ducted as an integral part of the PA effort described in the Components part selection: This activity is analogous to
previous section, and this will include such topics as the that of M&P in that electronic components are selected from
selection and approval of metallic materials for spacecraft those previously tested and approved for space use and those
structures, the assessment of electronic materials for specic nonstandard devices that need to be qualied by testing.
applications, the qualication of critical spacecraft pro- Metallurgical evaluations complement work performed by
cesses, the testing and approval of novel fabrication and components laboratories in areas related to glass-to-metal
surface-coating techniques, corrosion control, and the relia- sealing, packaging, microwire bonding, solderability, elec-
bility testing of electrical interconnections. tromigration, and the formation of whiskers.
3.2 The Materials Laboratory 67
3.2.2 Facilities and Instrumentation systems are particularly useful for inspecting for cracks,
voids, and solder paste volumes.
3.2.2.1 Macroscopic Examination
Macro-examinations relate to material features that are 3.2.2.2 Microscopic Examination
apparent to the naked eye, although in practice magnica- A microscopic examination is performed at magnications
tions of up to ten times may be used. A number of greater than ten diameters. It is far broader in scope than in
stereo-zoom microscopes and cameras exist to reveal and macroscopic examination, and consequently much addi-
record surface morphologies such as cracks, voids, inclu- tional information can be generated. Both fractography (the
sions, deformation, blistered coatings, de-wetted soldered microscopic examination of the irregular surfaces produced
joints, contaminants, dust particles, and the like. An example when a piece of material is broken) and metallography (the
of the use of macro-examinations is shown in Fig. 3.6; the examination of the microstructure of metals and alloys as
morphology of the titanium propulsion tanks internal weld revealed by cutting, polishing, and etching a sample through
bead was initially macrophotographed before being sec- the plane of interest) are used. Correct observation and
tioned and polished to reveal the weld macrostructure. interpretation of the etched microstructures depend heavily
A good magnier and light source are needed to perform upon the ability of the examiner to interpret what is revealed.
inspections efciently and to incur as little eyestrain as An excellent reference dedicated to the systematic prepara-
possible. Traditional overhead lighting, either fluorescent or tion of a vast range of material surfaces and a comprehensive
incandescent, is not appropriate for most inspection tasks overview of microstructural analysis has been prepared by
because each produces glare that interferes with the viewing Bouseld (1992, 1997). Reagents of microetching metals
of small parts and details. The best light source is adjustable, and alloys are detailed in Appendix 3.
on an articulated arm, so it can be directed to enhance The metallurgical laboratory microscope shown in
contrast and colours. It should be independent of the mag- Fig. 3.7a) is supporting part of a multilayer printed-circuit-
nier lens to reduce glare and reflection from metallic board that has been mounted in a room-temperature-curing
specimens. The magnier having a circular lens works well plastic and then ground, polished, and etched to reveal the
for small objects, but if large numbers of production items, cross-section of a plated-through hole. The area of interest
such as printed-circuit-boards, need to be inspected it is can be viewed through the eyepieces or, as shown, projected
preferable to initially scan each board for defects under a onto a viewing screen. Attachments exist for making digital
rectangular or stadium-shaped lens, which permits more imaging and storage, 35 mm slides or photographs and, now
visibility of the board at one time. Areas of interest can be obsolete, Polaroid instant prints. Other accessories available
identied with detachable plastic markers before the board is for these optical microscopes include an interferometer,
transferred to a stereo-zoom microscope or camera. At the polarized-light illumination, and a micro-hardness tester.
specialist end of inspection, 3-D imaging capabilities are The traceability and recording of all laboratory samples
available. This enables volumes to be measured, and the 3-D such as mechanical test pieces, metallographic mounts, etc.
Fig. 3.6 Girth weld of a hydrazine propulsion tank. These macropho- macrostructure, revealed in cross-section after grinding, polishing,
tographs highlight: a the root bead of the electron beam weld and and etching. Vibration could cause particle detachment, resulting in
associated weld sputter particles, and (b) the Ti6A14 V weld blockage of this spacecraft propellant system
68 3 The Integration of Materials into Product Assurance Schemes
is extremely important. They are needed for laboratory test under investigation (Pohl 1996). With these improvements it
reporting, and may be requested for re-examined during is found that even nonconducting surfaces can be examined,
future yearsparticularly if failures related to evaluated without the need for coatings. Ceramics, plastics, and com-
spacecraft parts are encountered during testing of during the posites can now be examined for minute surface features
long life-times required of modern spacecraft. The departure having a good lateral resolution down to 4 nm. Occasionally
of lab staff and re-hiring of new personnel means that the there will be the need to resolve precipitate particles within the
transfer of knowledge may only be possible if good house- matrix of a metallic sample, grain boundaries, and dislocation
keeping and recording is effective. Examples of good data defects. These can be viewed by means of the Transmission
logging, and sample traceability are illustrated in Fig. 3.7b). Electron Microscope (TEM); either the sample must be
This subject will be further discussed in Sect. 3.3.3. thinned by special electrochemical thinning techniques, or
High-temperature optical microscopy is available at a thin lms of carbon or polymeric material must be used to
number of laboratories. This can be a very useful technique produce a replica of the sample surface and it is the replica that
for the study of diffusion and phase changes, particularly is used for the examination. The TEM has a very high
those that may occur during the temperature excursions resolving power, as illustrated in Fig. 3.9a. More recently, the
experienced by spacecraft materials during exposure to solar Scanning Transmission Electron Microscope (STEM) has
heating or during re-entry into the Earths atmosphere. The been developed for thicker samples, and this instrument can
sample is prepared, usually as a polished section, and placed also produce diffraction patterns, so enabling the characteri-
on the hot stage of the microscope. Some microscopes are zation of crystal structures.
advertised as operating to temperatures of 1500 C; such There are several occasions when the materials laboratory
temperatures are essential for studying refractory metals, but may be requested to examine organic materials and even
in practice the instruments rarely reach such temperatures sample types such as bacteria, algae, spore, fungal growths,
without the need for extensive redesigns. A full run can take and human hair. The latter types may occur as particulate
several hours so that it is advisable to record the optical contaminants inside cleanrooms, and even if present in low
image on video, allowing repeated study of events. Investi- levels, they can be harmful during the production of
gations where the hot-stage microscope has provided useful advanced electronics such as integrated circuits (ICs). They
data have included: accelerated growth of the intermetallics may be present as human-sourced contamination capable of
that form between gold and tin when solder alloy is applied degrading wafers and producing on-die contamination so
to gold-plated terminations; the effect of dehydration on the reducing the reliability of ICs and hybrid packaged elec-
densication and cracking of chemical conversion coatings tronics. As part of the Space Station life support activities,
such as Alodine 1200 on aluminium substrates; viewing the research is progressing to nd alternative sources of oxygen
crack-opening displacement during the thermal fatigue of for the manned modules. One source is the Microbial Eco-
solder joints resulting from z-axis expansion of logical Life Support System Alternative. In this system,
printed-circuit boards (see also Fig. 6.17); and observing the different bacteria and cyanobacteria are grown in a culture
preponderance of adjacent phases in a metallurgical system and used to convert organic waste and carbon dioxide into
to tarnish, oxidize, and become less reflective. food and oxygen. The bacteria are being examined in the
In many situations even greater magnications are useful. authors laboratory, optical microscopy is satisfactory for
This capability is provided by the Scanning Electron Micro- making size and shape determinations, but SEM examination
scope (SEM) shown in Fig. 3.8, which can magnify to 200 is needed to see details (see, for instance, Fig. 3.9). Here, the
000 with good resolution. The great depth of eld of this bacteria are collected on a 0.4 m cellulose membrane lter,
instrument has made it particularly useful for examining washed with water-soluble solvent, and then allowed to dry.
rough fracture surfaces to determine the cause of a component The nal stage is to sputter-coat the lter with a thin layer of
failure. The majority of SEM investigations involve electri- gold and examine using low-energy electrons (about 5 kV) so
cally conductive objects which need no special preparation. as to inflict minimal damage to the organic specimen.
However, when electrically nonconducting parts are exam- Recent experiments (Thiel 2014) have surprisingly
ined in the SEM they can become charged and this causes the revealed that DNA can survive the extreme conditions of
image quality to suffer. It is therefore necessary to coat such space and the high temperatures (1000 C) of re-entry to
objects with a very thin layer of conductive material, either by Earth through the atmosphereraising the possibility of
sputtering or by vapour depositiongold is most often used molecules of extraterrestrial life arriving from space. Plas-
for this purpose. Recently designed SEMs can operate at a low mid DNA, circular strands found naturally in bacteria, was
cathode voltage, down to 1 kV or less, and with renements in formed into 12 samples and applied to the surfaces of a
controlling the equipments beam path, this has the effect that sounding rocket payload and its screw threads. These sam-
the electron beam hardly penetrates the surface of the material ples were examined before and after flight using a scanning
70 3 The Integration of Materials into Product Assurance Schemes
Fig. 3.8 Scanning Electron Microscopes have evolved, but early Duo detector, which can be switched between ultra-high resolution
models, a are still put to good use in laboratories where they have been in-lens secondary electron (SE) imaging and on energy selective
the workhorse in materials investigations since the 1960s. Very backscattered detection (EsB) capable of detecting extremely small
modern versions such as the EVO, b can be used in studies of organic differences in materials composition. For instance, c shows the rear-side
materials and life science samples at nano-scale interaction (at different of a focused ion beam aperture after a few 100 h exposure, 4000.
temperatures, pressures and humidities), they capture outstanding d shows the appearance of the fused near-eutectic tin-lead nish on the
topographical details at low voltages. Such state-of-the-art SEMs and track of a printed circuit boardthe microstructural features are
their sophisticated electron columns, are useful for electronic materials contrast-enhanced by use of the SEM in back-scattered electron mode
work: they enable high resolution imaging up to 0.8 nm (scanning (lead being the white dendrites), X400. bd are courtesy of T. Simpson
transmission mode) and nano-point analysis. Space materials scientists UWO, and Zeiss
can gain maximum information from a sample with the on axis in-lens
electron microscope at the University of Zurich. 35 % of the The AFM gives similar microstructural results to the
DNA retained its biological function. Clearly, these results STM. Also, a needle probe scans over the specimen in a
suggest that even greater sterilization and other safety pre- raster pattern with great precision. Unlike optical micro-
cautions should be taken to prevent carrying terrestrial DNA scopes that use mechanical gears to position the
to other moons and planets. They also suggest that DNA specimen-stage, the AFM uses piezoelectric forces to posi-
from outer space, could enter the atmosphere by means of tion and move the tip over the sample. The precision of these
meteorites. movements is in the order of 0.01 nm. The needle-tip is
New techniques for the microstructural examination of actually mounted onto a soft spring, known as a cantilever.
metallographic specimens are continually being developed. As the tip scans across the surface it may be pushed away
Two recent methods are the Scanning Tunnelling Micro- from very local areas of that surface by a repulsive force.
scope (STM) and the Atomic Force Microscope (AFM). The AFM detects the springs movement by sensing changes
The STM functions by scanning a ne-pointed needle across in laser light that reflects from the cantilever into a photo-
the specimen surface with a distance from the surface of a detector. The photodetector then initiates a piezoelectric
few tenths of a nanometre (nm). A potential is applied force on the cantilever to push the cantilever and needle-tip
between the specimen and the point, and the resultant current back into its original position above the specimen. This
flow is measured. The STM measures the height of the piezoelectric signal is converted, by a computer in the AFM,
specimens surface and can ideally examine grain bound- into a topographical image or picture. The STM and AFM
aries; with multiphase materials, the aim is to remove some are becoming increasingly used in the semiconductor
nanometres less of one phase than of the others during industry because they can detect surface features on semi-
etch-preparation of the polished specimen. conductor devices with a resolution of better than 1 nm.
3.2 The Materials Laboratory 71
was determined to have caused these bonds to fail in the eld specication. This can be done as part of the procuring
and, by Ion Milling, specic attention could be given to companys goods inwards inspection when a certicate of
microstructures and the corrosive ions that degraded the conformance or a chemical analysis record, provided by the
bonds. metal supplier, must be compared against the purchase order
requirements. In some instances when certicates of con-
3.2.2.3 Chemical Analysis 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-
(a) General Analysis in the Metallurgical Laboratory cation or submitted to qualication testing (e.g. expensive
fracture mechanics tests and stress corrosion cracking tests),
An energy-dispersive X-ray analyser (EDAX) attachment to it will be necessary to perform some simple mechanical
the SEM provides instant elemental analysis of materials. tensile testing and chemical composition analysis on samples
The instrument will identify some elements in concentra- cut from the delivered alloy.
tions of only 20 parts per million. Elements from sodium to The method by which the chemical analysis is made must
uranium are readily identied. The method can be used to be well dened and agreed with the laboratory performing
analyse only small areas, but is nondestructive. The results the work. As illustrated by Table 3.2 on page 102, the dif-
are generally qualitative, but a computer program can be ferent families of the new aluminiumlithium alloys have
developed to generate quantitative data by means of standard rather similar compositions. Clearly, the analytical technique
specimens. The primary use of this technique is the exami- must be highly accurate and reproducible in order to dis-
nation of unknown materials, batch comparisons, failure tinguish between these alloys.
analysis, and contamination analysis. Spacecraft microelec- A review of the various methods for chemically analysing
tronic devices, microwelded solar cell interconnections, aluminiumlithium alloys was made with Mr K. David of
electrical slip rings, and optical lenses all need to be MTS Daventry Ltd. (UK). Virtually all chemical analysis of
assembled or processed in highly clean areas. The writer has aluminium alloys is now carried out by instrumental spec-
been engaged in the analysis of ionic and particulate con- trometric methods, the longer, time-consuming classical wet
taminants that have severely degraded the quality of each of methods having been superseded. Atomic emission, atomic
the aforementioned space hardwarendings were that absorption, and X-ray fluorescence methods are most com-
these contaminants, and EDAX analyses determined the monly used nowadays for metallurgical analysis, but there
chemical composition. Dried ngerprints contained sodium, are limitations to each, and the analysis of aluminium
potassium, and chlorine; human hairs contained carbon and lithium alloys does present particular difculties as sum-
calcium; cosmetics contained silicon, sulphur, and titanium; marized below:
and dandruff contained calcium and sulphur. X-Ray fluorescence spectrometry. This technique is lim-
Energy-dispersive spectroscopy provides the main chemical ited by the inability to analyse for elements of low atomic
analysis and raises the awareness level of everyone involved weight. Although modern instruments can now analyse for
in the handling of very small spacecraft parts of the relia- beryllium and magnesium, lithium is too light an element to
bility hazards posed by human-sourced contaminants. To be determined by this method.
enable a more denite identication of contaminants, addi- Atomic absorption spectrometry. This technique has the
tional modern methods are required. A wide range of advantage of being comparatively cheap, but some elements,
modern chemical analytical methods, such as mass spec- e.g. boron, which is found in some aluminiumlithium
troscopy, auger surface analyses, X-ray photo-electron alloys, cannot be determined by this method.
spectroscopy, etc., are generally available at the sites of Atomic emission spectrometry by direct-reading instru-
larger companies, but may be accessible when required at ments. This is probably the most common method for the
local universities and research institutes (for additional analysis of aluminium alloys, in which a solid sample is
information see Sect. 3.2.2.6). For a complete coverage of excited by sparking, and the resulting emission spectrum
the techniques for, and the chemical analysis of small par- analysed. However, instruments need to be calibrated by the
ticle identication, the reader is directed to the work of use of reference materials of known composition, and as yet
McCrone and Deley (1973). there are no suitable aluminium-lithium certied reference
standards commercially available.
(b) Instrumental Spectroscopic Methods (Reference to Al Atomic emission spectrometry by inductively coupled
Li Alloys) plasma excitation. This is almost certainly the only suitable
method for the analysis of aluminiumlithium alloys. The
Whenever metal alloys are procured, it is important to ensure technique involves dissolving samples in acid, and pumping
that the delivered material conforms to its raw material the solutions into a plasma, again interpreting the resulting
3.2 The Materials Laboratory 73
spectrum. Instruments still need to be calibrated with suit- use on welded pressure vessel material having a curved
able standards, but as the technique involves solutions rather prole. A typical report (Thomas et al. 1993) that incorpo-
than solids, these standards can be prepared as cocktails of rated much hardness testing across the weld pool and
pure elemental solutions. These can be accurately made up heat-affected zones of Ti6A14 V concluded that as-welded
to cover a wide range of different elemental concentrations. spacecraft components were able to meet the mechanical
Modern instruments, particularly if coupled with mass property requirements and there was no special reason to go
spectrometry, offer lower and lower detection limits. in for any elaborate post-weld treatments.
The hardness of thin wires, coatings, and the individual
3.2.2.4 Hardness Testing microconstituents of alloys are usually measured on pol-
A number of hardness testers are available for the testing of ished cross-sections by means of a Vickers microhardness
metals, ceramics, rubbers, and plastics. The metals micro- tester (see Sect. 6.3 for examples) or a Knoop microhard-
hardness tester employs a square-based, diamond pyramid ness tester (having a rhomboidal indentor). In some of the
penetrator that can be selectively loaded with deadweights. latest hardness test machines, a means of electronic cali-
The indentation made in the test specimen by this penetrator bration can be provided to maintain the accuracy to
is measured with the aid of a microscope. The Vickers test-block standards. However, these machines can only
hardness number (VPN) is calculated by dividing the load by take measurements on surfaces greater than about 5 m
the area of indentation, but in practice it is usual to measure wide. Many of the current anti-wear surface coatings, such
the diagonal of the square and then consult tables. as the TiC-coated bearing steels discussed in Sect. 5.11,
There are two elds in which hardness measurements are have thicknesses in the order of 2 m. In response to the
particularly important: in establishing standards and in demand to check the hardness of such thin layers, a
maintaining product uniformity (e.g. a simple hardness Nano-hardness tester has been developed.
check can be used to inspect the uniformity of a particular The principle of the Nano-hardness tester is described by
treatment such as a forming operation, a heat-treatment its producer, the Swiss rm CSEM. Nanoindentation mea-
cycle, or a series of surface-hardening processes). surements are performed with a standard Vickers indenter,
Hardness testing is performed at some stage during the mounted on a cantilever spring. The indenter tip and can-
production of the majority of spacecraft metallic parts and tilever are driven to the precise surface location by piezo-
subsystems. It may be carried out as a rough check that the electric actuators. The applied force is measured via the
correct material is being withdrawn from the stockroom, or cantilever beam by a sensorthis operation is fully software
later during fabrication of a part, as a process control. Along drivenas the indenter penetrates the samples surface. The
with the concept of total quality management, it has been microhardness and Youngs Modulus of the sample are
found useful to include the measurement of hardness in automatically recorded by means of the machines software,
statistical process control. Spacecraft parts range in size from and a 3-D topographic inspection of the indentation area,
small bearings and micro-ne gold bonding wires to pres- prior to and after the indent, can be made by scanning atomic
sure vessel weldments and very large booster motor cases. force microscopy. The image and a typical force-penetration
Each of these parts must possess correct thermomechanical curve obtained during the microhardness testing of a thin
treatments and surface processing to achieve the required gold-coated specimen are shown in Fig. 3.10.
hardness. This is as crucial as machining operations for
achieving correct dimensional t. Hardness is related to the 3.2.2.5 Mechanical Testing
tensile strengths of some metals including steel, it can be an The mechanical behaviour of materials is characterized by
indication of their wear properties, it can assess the degree of relations between the stresses and strains imparted to the
temper as a result of heat-treatment, and will conrm whe- material under service conditions. A tensile test machine
ther a material is homogeneous. (Fig. 3.11a) enables standard test specimens to be pulled
Large items will probably be hardness tested and the apart at different strain rates by measured loads. Load cells
results recorded in terms of VPN. A typical investigation is may be changed so that full-scale chart readings of between
to assess the suitability of spacecraft gas bottles, reaction 2 N and 100 kN (this is generally the maximum capacity)
control system tanks (Fig. 3.6), and payload support struc- can be achieved in both tension and compression (see also
tures (Sect. 4.4.4) which are constructed from Ti6A14 V by Fig. 3.11b). This equipment can be used to investigate the
means of electron beam and TIG welding. Test samples will mechanical properties of all materials at temperatures rang-
be made and subjected to various post-weld treatments, such ing from 150 C to +300 C (a custom-built environmental
as annealing or mechanical stress-relief or solution treatment chamber may be suspended from rails attached to the
followed by ageing. Hardness test samples are small and machines xed upper crosshead). The equipment has the
inexpensive to investigate. They can replace the larger ten- ability to make fatigue tests by applying repeated loads of
sile test samples which are both expensive and unsuitable for the same magnitude to a specimen or component.
74 3 The Integration of Materials into Product Assurance Schemes
Extensive thermal cycling of component-assembled curve relating the magnitude of the thermal cycle (equivalent
printed circuit boards will eventually produce failure to the stress range) to the number of cycles necessary to
mechanisms in different locations of the circuit. The results produce the rst failure (Halpin 1985). Figure 3.13 illustrates
of numerous thermal cycling programmes have been anal- the concept of a thermal fatigue curve for component
ysed and it has been found possible to generate a fatigue-life assembled multilayer boards manufactured to military
76 3 The Integration of Materials into Product Assurance Schemes
quality. As part of a spacecraft new technology evaluation a It may be noted that only static SCC test methods have
circuit could be submitted to verication tests following the been described; these are considered to be most suited for
thermal cycling prole shown in Fig. 3.12b. The results can the classication of metal alloys according to their resistance
be placed on the thermal fatigue curve and test circuits that to SCC (see also Sect. 4.6.1). Dynamic SCC tests are fre-
survive are assumed to have fatigue curves no worse than the quently performed by laboratories, and data can be achieved
dashed curve shown on Fig. 3.13. When the new technology within hours rather than the weeks (30 of 80 days) required
is put into spacecraft service the operating stresses are less for the constant load test method. When cylindrical tensile
than those on the dashed fatigue curve and the length of time specimens or precracked specimens are evaluated by vari-
before failure can be expected to exceed the spacecrafts able strain rate techniques the results should only be used to
designed life. The same curves are applied as a rule of thumb compare batches of material within the same family of alloy
by the author during the verication testing of surface- composition. At high strain rates the electrochemical reac-
mounted leadless components. The number of cycles nec- tions associated with SCC do not have sufcient time to
essary to produce the rst leadless component failed occur, and the deformation process will depend on disloca-
soldered joint may be 3750 for a 180 C temperature range tion movements occurring throughout the specimen to cause
and 1250 for a 215 C temperature range (Frear et al. 1993, necking, a high elongation, and, ultimately, transcrystalline
Table 9.1). breakage of the metal grains. As the strain rate is decreased
while the specimen is exposed to the corroding solution, for
3.2.2.7 Corrosion Testing instance a saline solution, the susceptibility of the specimen
Corrosion tests are performed on spacecraft materials in to SCC may or may not increase. This uncertainty depends
accordance with several international specications. Corro- on several factors, but mainly on the ability of the metallic
sion laboratory equipment such as that shown in Fig. 3.15a, b surface to form a protective passive lm. If the lm does
will evaluate a wide range of features, including the suscep- form it will rupture under further strain to form a crack, so
tibility of structural materials and springs to stress-corrosion producing localized anodic dissolution of the metal under
racking, the resistance of anodic and chemical conversion test. If the strain rate is sufciently slow the repassivation
coatings to surface corrosion, the intergranular corrosion of process will have time to ensure a new lm above the crack
weld-heat-affected zones, the protective properties of paints, before dislocation movement causes the emergence of the
and the compatibility of bimetallic contacts. Such electronic next slip step. The ability of the alloy to repassivate is the
devices as component packages, battery cells, and. wires and predominant factor controlling the propagation of the stress
cables should be screened against corrosive environments corrosion crack under dynamic testing; other factors are
that include humidity, the residues of soldering fluxes, and associated with hydrogen-induced cracking. The existence
salt spray. of a different threshold strain rate for the repassivation of
Large-scale stress-corrosion tests necessary for the eval- each metal alloy under test renders the results of dynamic
uation of new spacecraft alloys (such as aluminiumlithium SCC testing difcult to interpret for classication against
alloys and particular-reinforced alloys) are usually made SCC susceptibility. Selection of the most appropriate
under constant-load conditions by alternate immersion method for performing SCC testing must taken into account
cycles of 10 min in a solution of 3.5 % NaCl and 50 min a review of the spacecraft hardware material under consid-
drying in air; typical apparatus is shown in Fig. 3.15c. eration and its applicable in-service environment. The
78 3 The Integration of Materials into Product Assurance Schemes
majority of components will be exposed to static loads, as 3.2.2.8 Fracture Mechanics Testing
the period of launch is small in comparison to their required The concepts and principles of fracture mechanics are rela-
lifetimes, hence the constant load method is to be preferred. tively new, but when the discipline is applied it can offer a
To duplicate the stress corrosion failure of spacecraft system-type approach to the failure prevention of launcher
battery cell terminals and the leads of electronic component casings and pressure vessels. In view of the high cost of
packages, fabricated from Nilo 42 and Nilo K respectively, launching materials, it is increasingly pertinent to optimize
testing is made in the laboratory and under marine atmo- the use of metal alloys and their heat treatment with per-
spheric exposure conditions. Since the failure of both alloys formance. Damage tolerance is generally applied to all
in service had been in situations where plastic strain (from manned spacecraft piece-parts where a failure could cause a
bending and forming) and elastic strain were present, stress catastrophic hazard. In applying the damage tolerance phi-
corrosion tests were carried out by the loop test devised by losophy the designer must assume that crack-like defects
Thompson (1961). The marine test sites and test samples are exist in each safe-life piece of hardware, and that, during
shown in Fig. 3.16. dened intervals such as the mission duration or inspection
A review of the methods for testing ferrous and nickel interval, these cracks will not propagate to cause failure.
alloys and their weldments for a wide range of corrosive This activity is greatly assisted by the establishment of
environments (general, pitting, intergranular, crevice, SCC empirical fracture mechanics tests and fatigue tests on rep-
and galvanic corrosion) has been compiled by Pettersson resentative materials or proof tests on the actual fabricated
(2007). This extensive paper discusses the applications of item. It is important to realize that both fracture and fatigue
stress corrosion standards. The Constant Load Test testing for the generation of crack growth, material tough-
(CLT) which is preferred by the European space community ness, and fatigue properties are influenced by many param-
seems to be the most used test for the oil and gas industries. eters such as the precise alloy composition, heat treatment,
3.2 The Materials Laboratory 79
grain direction, and test environment. Only by extensive specications. The testing is generally performed with the
testing can basic material data be derived for spacecraft following equipment:
fracture mechanics analysis. This is also particular relevant
for composite materials; life predictions for composite Servo hydraulic testing machines with capacity in the
structures are more frequently made because of the increased range 100500 kN (computer-controlled input capabili-
use of these materials in aerospace applications. ties for variable amplitude load spectra to simulate tran-
Fracture toughness tests and crack growth tests are made sient and random service loading).
with specimens that have been precracked with either sur- Facilities for measuring crack growth.
face or internal cracks. The ESA fracture control require- Environmental temperature testing, high pressure or
ments, (ECSS-Q-ST-40, 2008) dene when fracture testing in vacuum, and possibly testing in an aggressive
toughness tests are performed to determine the KIC (plane gas or liquid environment.
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 Fatigue tests are performed to determine the number of
rectangular cross-section test-piece in which a notch is load cycles that a material can withstand without fracture
machined. The notch extends from one edge through half the the tests load the specimen for a denite stress cycle, applied
depth of the specimen, and the specimen width is half the at a dened frequency. For all metals there is a limiting
depth. The specimen is rst subjected to fatigue loading to range of stress below which fracture will not occur for any
cause a real crack to form at the root of the notch, and loaded number of loading cycles, however large. When there is
until the crack has grown to 5 % of the overall notch length. fatigue crack growth it is possible to assess the growth rate
The crack length is monitored with a travelling microscope, by inspecting the polished edge of the specimen (as for
and measured accurately after test. Test samples made from instance was done for producing the fracture toughness
AA2219 forgings are shown in Fig. 3.17. The samples were specimen in Figs. 3.17a and b. Fatigue fracture surfaces are
loaded in tension. The load versus displacement is measured illustrated in future chapters (a classical fatigue failure is
and the loading rate determined as being within specica- shown in Fig. 5.2), but for modern metallic alloys it is the
tion. The KIC values are assessed from tables. The methods microstructure of the alloy that provides for their low fatigue
of testing are becoming well dened in recognized crack growth rate (FCGR). The superior FCGR resistance of
3.2 The Materials Laboratory 81
aluminiumlithium alloys can be traced to their micro- crack length, and number of load cycles), deduced data
structure. This is shown in Figs. 3.17c and d, where the produced from raw data by numerical methods such as curve
fatigue crack has a tortuous crack path resulting from the tting or differentiation, and design data intended for veri-
alloys highly planar slip characteristics and the presence of fying a structural design, such as lower-bound values.
texture. This tortuosity increases the energy required to Industrial companies can access the ESA FRAMES-2 data-
propagate the crack and impedes crack growth. The facets on base as an aid in their selection of structural materials.
the fatigue crack surface have been studied in detail by Wu
et al. (1994); they are separated by 109.5 and this angle of 3.2.2.9 Printed-Circuit-Board and Component
separation coincides with the angle between the favourable Assembly Evaluations
slip plane of aluminiumlithium{III} and the plane that is Printed-circuit-board (PCB) evaluations are made in metal-
parallel to the rolling direction of this plate material. lurgical laboratories during the quality assessment of PCB
The present standards for fracture mechanics testing were manufacturing lines. New PCB laminates and fabrication
originally developed for the testing of homogeneous metallic techniques can be studied for possible use on future space-
alloys. Future modications must be made to the test pro- craft. Equipment exists for the visual inspection of circuits,
cedures when ceramics, metal matrix composites, and peel-strength testing of conductor tracks, and the solder-
metallic weldments are tested. ability testing of terminals, whether they be pads or
Data on the strength of materials have been derived over plated-through holes.
the years by industry and national agencies. A computerized Component assembly on PCBs and wire/cable inter-
materials database is available for compiling this data. connection by manual soldering, automatic soldering,
FRAMES-2 contains raw data taken from measurements microwelding, and crimping are some of the most critical
performed on test specimens (such as force, displacement, spacecraft processes. Standard test equipment exists in the
82 3 The Integration of Materials into Product Assurance Schemes
laboratory to evaluate the material properties and metal- In the most simple case a PCB consists of an
lurgical characteristics of samples produced by these join- epoxy-impregnated breglass-laminated weave with copper
ing techniques: conductor lines bonded to the epoxy laminate surface by
means of a further epoxy adhesive layer (called the top
Mounting, grinding, and polishing equipment, with coat). Electronic components are soldered to copper pads of
ultrasonic cleaning baths for the metallographic prepa- plated-through holes. More modern boards are made of
ration of soldered joints, plated-through holes, etc. polyimidequartz laminate, and these are particularly
Photographic camera for the general photography of well-suited for the assembly of leadless surface-mount
incoming samples. components such as ceramic chip capacitors and resistors,
Solder globule solderability tester for wetting/dewetting leadless ceramic chip carrier packages, and the like. In more
checks on component leads, wires, terminal posts, etc. complex, and more expensive, PCB assemblies, the lami-
Rotary dip machine for solderability and solder-float nated substrate consists of a ceramic dielectric with a
testing of PCB pads and plated-though holes. Some three-dimensional refractory metallized conductor network
effects of this test on poor-quality products are illustrated with pads that have a series of metallized terminations onto
in Fig. 3.18. which electronic components can be soldered. Multichip
Peel test made by using attachment to tensile test modules (MCMs) can be mounted onto these
machine (see Sect. 3.2.2.5). ceramic-laminated structuresthis is essentially the
Thermal cycling systems (see Sect. 3.2.2.6). mounting of integrated circuits onto a ceramic board that has
Laminar flow bench providing a dust-free working space incorporated a wiring circuit with signal layers, power
for the assembly and inspection of spacecraft hardware planes, and ground planes built in. Simply put, the chip
under ideal lighting conditions and a Class-100 envi- carrier becomes the ceramic printed circuit, or multilayer
ronment (US Fed. Std 209a). board. Obviously, these MCMs offer a decreased size and
Stereo-zoom binocular microscope and metallurgical weight and a faster performance (up to 34 times as fast
microscopes for quality inspection of electronic traditional PCBs), and they are presently under consideration
hardware. for qualication to spacecraft test standards.
3.2 The Materials Laboratory 83
Ceramics such as alumina oxide (A12O3), aluminium 3.2.2.10 Conventional Nondestructive Testing
nitride (AIN), silicon (Si), and beryllia (BeO) are a great deal (NDT)
more fragile and susceptible to microcracking than metals A large number of NDT techniques are available to the
and PCB materials. Consequently, all metallographic materials engineer, and spacecraft piece-parts and small
preparations must be made extremely carefully and by using components are frequently inspected for defects such as
methods that distribute compressive stresses evenly across microcracks in brittle materials and porosity in welded
the ceramics surface. Samples should be potted in a hard structures. The defects can be generally categorized as:
plastic and slowly ground down to the plane of interest.
Sawing, even with diamond-coated blades, is unsuitable, as raw material defects (e.g. rolling defects, blow holes in
microcracks having long cleavage paths will be generated castings, seams, segregation of alloying elements),
and it will often be impossible to grind back sufciently to spacecraft processing defects (e.g. brazing and welding
eliminate these cracks. When conductor thicknesses are to be anomalies, incorrect location of encased units, delami-
measured it is essential to over-plate the metallizations with nation, machining defects, misalignment),
a hard layer of nickel or thick copper in order to prevent service defects resulting from overtesting, incorrect
edge-bevelling during the polishing of microsections. The choice of materials, fatigue cracking.
evaluation of ceramic PCBs typically does not require any
special operator training beyond an emphasis on extra care in Conventional NDT techniques require trained and com-
handlinga ceramic board when dropped will very often petent operators and methods for record keeping. A good,
shatter. clear, enlarged photograph or radiograph is worth many
84 3 The Integration of Materials into Product Assurance Schemes
words, and occasionally video tapes may be useful to record Galileo spacecraft (Johnson 1989), when the high-gain
visual examination. The most important methods are: antenna could not be fully deployed. The ribs of the
Visual examinationThis is probably the single most umbrella-shaped antenna were locked for launch, but failed
important method of inspection, and relies on a trained eye, to open. Investigations have shown that fretting during
magnifying devices, good lighting, endoscopes, mirrors, and transport and lift-off caused the ribs to cold weld together in
brescopes. the launch position. Further more recent examples will be
Liquid penetrant testsPenetrants may be simple mix- given in Sect. 5.1.2.
tures of alcohol and water so that, after a suitable penetration The aim of a recently developed tribology test (Mer-
time, the excess liquid is removed and the surface dried. stallinger 2009) is to provide designers and engineers with
Liquid seepage out of crack-like defect may be made more data on adhesion forces that have to be considered in design
apparent by means of dye, fluorescent additives, and thin, of mechanisms, in order to assess, on the one hand, cold
white surface coatings which become stained by the seeping welding on bare metal contacts and, on the other, the per-
liquid. Penetrants can contaminate spacecraft if not com- formance of coatings in preventing adhesion. The test
pletely removed. enables comparison of results for different material pairings,
Ultrasonic testingAll spacecraft metals readily transmit and the collection of data for entry to a database.
ultrasonic vibrations which are reflected or scattered (atten- The impact test facility, for loads up to 100 N, enables the
uated) by flaws or inclusions. Methods include transmission, measurement of adhesion forces under vacuum between
pulsed echo, and acoustic imaging. Composite materials are contact points after cyclic contacts, varying from static con-
more difcult to assess and should be double checked by tacts (long-term compression) to slow closing cyclic contacts
radiography particularly in the case of honeycomb sandwich (static adhesion), and after impacts with impact energies up to
panels. 0.02 J. The impact test facility consists of an ultra high
RadiographyRadiographs can be produced by means vacuum system (108 mbar after bake-out at 130 C) with an
of X-rays and gamma-rays that pass through the item under ion getter pump and an air damping system for vibration-free
investigation to produce a photographic image. X-ray measurements. The contact is made between a ball and a flat
absorption will occur when dense material and refractory disc. These are the two metals under evaluation. The ball is
inclusions are present, giving rise to light areas. Voids, mounted on a pushrod, which is driven by an electromagnet.
porosity, and cracks, being hollow, show up as dark areas. A low-friction loading system enables an accuracy of 1 mN
Organic composite materials may be more difcult to inspect (0.1 g) for the adhesion force measurement, which is made
because the X-ray absorption of resin and graphite bres is directly above the pin in the vacuum chamber using a piezo
similar. Penetrant-enhanced radiography can be used to force transducer. The transducer measures the impact force as
detect disbonds within honeycomb skin lay-ups. well as the adhesion, i.e. the force needed to separate the two
Coin or tap testGross defects can be located by lis- materials.
tening to the ring of a metallic item; flawed parts may emit a Cyclic loading may be done either slowly (static) or by
dull note when impacted with a coin or small hammer. impact (dynamic). By varying the ball radius (typically 2
Disbonds between face skins and honeycomb panels can be 20 mm), the contact pressure can be adjusted to the yield
easily detected by carefully dropping a coin, edgeways, onto strength of most materials. As an option, the contact surfaces
the face skin surface. A dull rather than a harsh response will can be cleaned in situ by glow discharge before the test.
result above areas of disbond. Listening sensitivity can be The fretting facility allows investigations of the influence
improved by means of a stethoscope. of fretting on the tendency to cold welding of materials.
After a certain number of fretting cycles the adhesion force
3.2.2.11 Tribological Test Methods between contacts is measured. The loading mechanism is
Spacecraft mechanisms are generally designed to incorporate similar to that of the impact test device described above. The
a lubricant that does not degrade either during ground testing loading and adhesion forces are measured by the z-direction
in air, or testing and operation under vacuum. Metal to metal of a 3-axis piezo transducer mounted directly below the disc
contacts are composed of dissimilar metals as discussed in under vacuum. The friction force due to the fretting move-
Sect. 2.3(d). A common failure mode seen during the testing ment is measured in the x-direction. The fretting movement
and operation of spacecraft is termed cold welding. Euro- (sine, triangle or square wave) is generated by a piezo
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 tudes up to 100 m. This lateral movement is transduced to
occur under conditions of impact or fretting. These surfaces the pin via a CuBe plate and controlled at the contact by a
may be bare metals, or inorganically or organically coated triangulation sensor.
metals and their alloys. One documented example of a This facility has allowed for the rst time the simulation
failure due to cold welding after fretting occurred on the of high-frequency vibrations resulting from bearings, for
3.2 The Materials Laboratory 85
example, combined with measurements of the adhesion acoustic attenuation and absorption properties of the
force. underlying material.
The surface waves are detected by a scanning laser beam
with a small spot size, which permits an imaging resolution
3.2.3 The Use of New Laboratory Techniques of the order of 10 m. The acoustic image is displayed on a
for NDT TV monitor, where white regions correspond to areas of the
sample with good acoustic transmission properties, and
3.2.3.1 Holographic Interface Tester darker regions denote areas where ultrasonic attenuation is
The metallographic methods employed for the case investi- higher.
gations detailed in this book are based on both optical and The SLAM uses a video-signal processor to integrate the
electron microscopy, and X-ray analysis. In recent years, the signals from laser scans to a number of grey levels, each of
continued demand for better monitoring of quality and which is assigned a different colour. This is demonstrated in
reliability has led to the introduction of more modern Fig. 3.21. Which shows an example of SLAM inspection of
inspection methods and laboratory techniques than those a silicon chip attached to a ceramic carrier by means of a
outlined in the previous section. They can supplement con- solder preform. Although the colour images contain the
ventional procedures by providing remarkably accurate same amount of information as the black and white ones,
nondestructive data. This can be useful for monitoring the they are easier to interpret.
growth of defects and for determining the optimum locations Photographs, termed acoustic-amplitude micrographs,
for making microsections. are taken of the acoustic TV image as seen in Fig. 3.21b. In
The usefulness of holographic NDT inspection will be addition to displaying the acoustic-amplitude distribution
demonstrated in Sect. 5.10 as part of the failure analysis of a throughout the eld of view, the SLAM provides an acoustic
defective spacecraft antenna. Debonded and fractured parts interference mode of operation. Acoustic interferograms
of the CFRP face-skins were uniquely identied by the (Fig. 3.21c) exhibit a series of alternating light and dark
Holographic Interface Tester (as developed by Schliekel- bands. For acoustically homogeneous samples, these bands
mann 1972), and the extent of the damage could be photo- (interference fringes) are arranged parallel to one another
graphically recorded before progressing with the destructive and are equally spaced.
metallographic failure mode assessment. The rst SLAM purchased for a European laboratory was
in 1980. This instrument has provided decades of useful
3.2.3.2 Acoustic Microscopy 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
(a) Scanning-laser acoustic microscopy cells making up a spacecraft solar array. These welds are
made by a process requiring close control of every process
The Scanning Laser Acoustic Microscope (SLAM) produces parameter, particularly for the electrode pressure and current
images of the internal structure of metals, ceramics, and even pulse. Weld quality could not be determined by conventional
biological tissues, and these can be displayed in real time on nondestructive methods including high-denition radiogra-
a television monitor, as shown in Fig. 3.19. The SLAM uses phy, and reference to workmanship standards also failed to
very-high-frequency ultrasound to produce images. provide information about these very small weld zones. This
The basic transducer and detection scheme employed by problem was solved by the SLAM method, which managed
the SLAM is illustrated in Fig. 3.19b. In this case solar-cell to locate and determine the extent of weld zones between
interconnector welds are under evaluation. The samples are solar-cell metallization and interconnectors. The progress of
hand-held and are easily manipulated on a simple specimen thermal fatigue within the welded zones can also be fol-
stage. As shown in the sketch, acoustic waves are trans- lowed. The accuracy and high resolution of the SLAM
mitted into the bottom face of the solar-cell sample by means method have been conrmed by detailed metallography,
of a distilled-water coupling medium. particularly or interconnected samples that had been sub-
Ultrasonic waves will not propagate across an air gap, mitted to 30,000 cycles between 100 C and +100 C as
and are therefore attenuated in areas where there is lack of part of a weld reliability programme for a low-orbiting
bonding. The waves will pass through where solar-cell spacecraft.
interconnector weldment is good. They are then reflected at The SLAM instrument is no longer manufactured having
the samples top surface. The transmitted acoustic energy been replaced by a series of acoustic microscopes without
imparts pressure waves on the top surface of the specimen. scanning laser features. The latest version is shown in
The oscillations have the same frequency as the incident Fig. 3.20 together with images made using two modes of
wave, but the variation in their amplitude depends on the operation. The reflected mode, where pulsed ultrasound has
86 3 The Integration of Materials into Product Assurance Schemes
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
capacitor have been made possible by the recording 50 Fig. 3.22including those described for the SLAM in the
images, each taken at an equal, small increment throughout preceding paragraphs. In some circumstances the technique
the body thickness. As major defects have appeared (the can be used instead of making destructive microsections.
white zones at increments 15 to 18) they were highlighted by Metals, polymers, ceramics, and certain composite materials
the yellow arrows, but also different material densities can be have sufcient acoustic transmissivity to permit imaging.
seen by the changing shades of grey. The C-SAM is Defects that are positioned beneath other discontinuities are
described in the next section. not detectable by C-SAM imaging (Martel and Adams 1996;
Rameriz and Adams 2002).
(b) C-mode scanning acoustic microscope (C-SAM)
3.2.3.3 High-Definition Radiography
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 (a) X-ray radiography
mechanically rasters a small dot of ultrasound over the top
surface of the sample. As the ultrasound is pulsed into the Conventional X-ray units are primarily designed to provide
surface, a reflected wave (echo) is generated at each interface the sharpest of images in the shortest of times. These radi-
within the volume of the sample and returned to the trans- ological examinations are based on the fact that X-rays and
ducer for processing. A high-speed digital signal processing gamma-rays can pass through materials which are optically
unit permits information to be gathered from all depths opaque. The quality or denition of a radiograph will
within the sample. Images can be generated from specic depend on factors that affect the transmission of radiation
depths or through the complete sample thickness. The through the specimen, and these will include material
high-frequency ultrasound (10100 MHz) is very sensitive thickness, composition and homogeneity of the material,
to internal discontinuities in materials. The method has been and intensity of the initial radiation emanating from the
developed for many NDT applicationssee for instance equipments focal spot. Modern equipments will detect the
88 3 The Integration of Materials into Product Assurance Schemes
3.2.3.4 Acoustic Emission to Monitor Damage reinforcement. A rather modern technique of detecting the
Accumulation in Materials acoustic emission (AE) from materials under stress is being
and Welds developed which can provide a direct method for studying
Advanced spacecraft structures and certain mechanisms can mechanical property degradation, interfacial disbonding, and
be constructed from a number of composite materials in the development of microcracks. An overview summarizing
order to achieve increased specic strength when compared the method has been written by Hinton (1995). AE refers to
to conventional monolithic alloys. Composites can also relatively low-amplitude sound waves that are generated by
contribute to other improved characteristics such as better various dynamic energy transfer mechanisms from micro
wear resistance or a specially controlled coefcient of areas in a material. According to ASTM E610-82, the de-
expansion. The mechanical properties of intermetallic com- nition of AE is: Acoustic emissions are transient elastic
posites and metal matrix composites (IMCs and MCCs) as waves generated by the rapid release of energy from energy
well as bre-reinforced plastics (FRPs) depend greatly on sources within a material. In metals these can be the
the integrity of the bond between the matrix and the cracking or decohesion of inclusions and the interaction of
3.2 The Materials Laboratory 91
slip lines, as shown in Fig. 3.26. The formation of twin Many aerospace materials have been evaluated by AE
bands and grain boundary cracking also emit AE signals (an (Hamstad 1986), but because it is not possible to distinguish
extreme case is that of pure, polycrystalline tin which damage modes or failure mechanisms by this technique
releases an easily audible sound, or cry, as it is being alone, the method has received a somewhat lukewarm
deformed). In composites the AE is generated by bre reception by materials engineers. However, AE response
breakage, matrix damage and, as previously stated, the behaviour has been characterized for a welded maraging
delamination and pull-out of reinforcement particles from steel and acceptance criteria determined for the control of a
the matrix. spacecraft pressure chamber and undergoing a proof pres-
The detection of AE is conventionally performed by sure test (Chelladurai et al. 1995). AE has also been used to
placing a sensor either directly, or by means of a couplant, monitor fatigue damage in samples of Inconel 901 and to
onto the surface of the item under investigation. Sensors are predict the onset of crack initiation (Fang and Berkovits
piezoelectric crystals such as Piezo-Zirconate-Titanate 1995). The same authors found that the threshold stress
(PZT), mounted as transducers, and these are extremely intensity factor values determined by AE tests were com-
sensitive to acoustic waves having an amplitude as low as parable to those obtained from conventional fracture
25 pm (Pollack 1989). A variety of waveguides, electronic mechanics tests, but for a much reduced time and cost.
ampliers, and electronic lters may be used before the The major asset of AE is that it is now a sufciently
signal is processed by a secondary amplier and nally mature technique to assist in both the production and the
displayed on an oscilloscope. By using multiple transducers quality controls for composite materials. For instance, under
attached to the surface of a device or structure it is possible proof loading, spacecraft hardware can be AE monitored for
to precisely locate the source of a defect within the volume First Damage Detection, Flaw Detection, and Failure
of the material (this can be considered as analogous to Warning. Also, the emissions that are originating from the
nding the epicentre of an earthquake). relief of residual stresses can be monitored in situ during
either thermal or mechanical processing.
92 3 The Integration of Materials into Product Assurance Schemes
AE was applied by industry for monitoring materials used bond between the CFRP and the aluminium surface which is
in the construction of the Concord supersonic planes in the most critical. Numerous sensors are attached to surfaces both
1960s and is now being used to study microssures in space in the region of the aluminium-to-CFRP bonds and the
materials such as hybrid glasscarbon and Kevlar laminates CRFP panel-to-panel joints. The AE testing is mainly con-
(Perami et al. 1992), aluminium and titanium matrix com- ducted with the SPELDA in compression (to simulate the
posites (Roman et al. 1992), and graphite-epoxy which is compressive forces expected from the payload and fairing
wound onto a liner/mandrel pressure vessel consisting of during launch). Any unacceptable AE can give an early
two aluminium hemispheres that had been tungsten arc warning of failure so that testing is halted before ultimate
welded at the equator (Downs and Hamstad 1995). AE was failure occurs and the problem areas can be pinpointed and,
also applied in a manner similar to that presented in the last after being subjected to a material review board, there is the
reference during the development of the lament-wound possibility of reworking the structure to an acceptable level
MAGE apogee boost motors discussed in Sect. 4.8.3. Here, of condence. The base line criteria for this condence were
the optimization of manufacturing process parameters were gained from AE testing on development and qualication
partly determined by AE responses generated during pres- models which were loaded to failure. The typical failure
surization of these vessels up to both the proof load and their criteria for one-shot launch vehicles are very different to
actual burst pressure. In-process quality controls utilizing those applied to aircraft or space shuttle vehicles with
AE are also applied during the fabrication of the Ariane IV 20-year design lives, and may include points such as: a
launch vehicles various payload fairing (SPELDA) designs. Felicity ratio (FR) of greater than 0.95 measured at 90 % of
Each SPELDA (see Fig. 2.7) takes up the stresses during the loadthe FR being dened as the load at which the
multi-satellite launches and provides the aerodynamically onset of signicant emissions occurs on a subsequent
shaped external protection skin that is released by the det- loading divided by the previous maximum load (see
onation of pyrotechnic devices just prior to the spinning and ASTM E 1316 for details), and no rapid increase of AE with
ejection into space of the individual satellites. The SPEL- stress.
DAs are constructed from thin carbon-bre reinforced plastic Spacecraft materials are often subjected to severe tem-
skins which are laid up and bonded to aluminium perature conditions and gradients of fast heating and cooling.
load-bearing rings. These bonded joints are difcult to This leads to thermal microcracking and delamination due to
control by nondestructive methods such as radiography or the mismatch in thermal expansion coefcients of the vari-
ultrasonics owing to the bre content and the presence of ous constituents of a composite material. Potential materials
small voids, but they are assessed by AE. The test setup for for a reusable aerospace vehicle such as carbon-reinforced
this part of the European launcher is described by Cole and bismaleimide or polyimide thermostable matrix composites
Thomas (1989); the locations of attached sensors are shown have been studied (Favre and Raud 1995) by thermal cycling
in Fig. 3.27. In fact, it was found that the small voids have between +250 and 110 C with low and high (to 15 C per
no effect on the strength of the joint, as it is the interfacial min) cooling rates. Special heat-resistant detectors were
3.2 The Materials Laboratory 93
metal, heat-affected zone (HAZ), and weld bead) will be the the HAZ of the girth welds are hoop tension and longitudinal
summation of the residual stress, and those stresses con- compression. Immediately after welding the hoop stress can
tributed by the weight of the tank and its propellant. reach 375 MPa. This value was obtained from hole-drilling
In the case of 33 mm thick forged AA2219 plate that had strain-gauge techniques described in (b) below. After a
been multipass TIG welded with 2319 ller metal, extremely suitable stress-relief heat-treatment there is a remarkable,
high residual stresses caused the plate to bend, conrming order of magnitude reduction in residual stress down to
that the 280 MPa yield point of the ller metal had been about 37 MPa (Haupt 1996). Neutron diffraction mea-
exceeded (Verderaime and Vaughan 1995). surements for a similar tank design, but produced by a dif-
With thin-walled titanium pressure vessels (yield ferent manufacturer (see (d) below), indicated lower
point = 820 MPa) it is known that in the residual stresses in
3.2 The Materials Laboratory 95
as-welded tensile stresses and a less effective stress-relief temperaturethis is considered to be the stress-free refer-
treatment. ence. X-rays have a low depth of penetration, and only very
The main methods for measuring residual stress, in their slightly subsurface values for residual strain can be made
order of industrial popularity, are described below. with this method. A full description of the X-ray diffraction
method for residual stress measurement, together with many
(b) The hole-drilling strain-gauge method case studies, is given by AMS International (1991).
This is dened by the ASTM method E837 (1994) and (d) Neutron diffraction method
consists of attaching a special three-element strain-gauge
rosette to the surface of the metallic piece under investiga- The principle is similar to that of the previous X-ray method
tion. As the gauge is quite small it can be used on surface except that neutrons have a 1000 times greater penetration
positions in the weld metal, HAZ, or anywhere on the parent depth. Also, absolute strains in materials can usually be
metal. It is usual to initially dress the weld bead in order to calculated from measurements made at any depth from the
achieve a reasonably flat surface for strain-gauge attachment. samples surface. In the case of neutron diffraction the
The three gauge grids are wired and electrically connected to specimen has to be taken to a nuclear reactor. Stresses, for
a static strain indicator. A small, shallow hole is then drilled instance in the various local regions of a weld, can be
through the centre of the rosette in small increments. The mapped by studying how the neutrons are scattered by the
introduction of a hole into the residually stressed metal planes of atoms in the solid weld material. The neutron beam
relaxes the stresses at that location, causing the local strains has a xed wavelength (usually 1.65 ), selected to make
on the surface to change correspondingly. Readings are the scattering angle close to 90. The strains are again
made of the relaxed strains and, by means of special charts, determined by observing the diffraction angles and calcu-
these are converted into values of residual stress (in three lating the lattice spacing.
angular orientations). The accuracy of this method is con- Peaks in the diffracted intensity are observed when the
sidered to be 10 %. The method is widely applied in Bragg condition is satised. This condition is given by the
Europe and the USA. An interesting case study related to the expression
NASA MARS Global Surveyor propellant tanks has been
described by Haupt (1996), who concluded that this was a k 2d sin h
practical method, being cost-effective and the least destruc- where is the neutron wavelength, d is the interplanar
tive technique for gathering residual stress data. spacing, and 2 is the angle between the incident and dif-
fracted beams.
(c) X-ray diffraction method In a state of tension the interplanar spaces are expan-
ded. The strain, , is a dimensionless quantity, the frac-
When external or residual stresses are applied to a structure tional change of d from that in a stress-free reference
made of metallic grains, the crystalline lattice is distorted sample, d0:
and the interatomic distances are changed. As the stress is
increased, so the interatomic spacing will be proportionally d
e 1
increased. Monochromatic X-rays are directed onto the d0
sample under investigation, and these rays are then reflected
from successive parallel planes of atoms in the metals grain Thus, careful determinations of 2 and give direct mea-
structure. The reflected rays are reinforced in one direction, surements of d and the strain.
which is called the diffraction direction. The angle between A three-dimensional map of strains in a series of volume
the incident or reflected beams and the surface of the elements in the weld, the HAZ, and the base metal can be
reflecting planes is measured by either lm or goniometer generated. From this the residual stresses in each of the three
techniques. Braggs law denes the condition for diffraction, principal directions can be calculated. Further information is
and by means of a simple equation the interplanar spacing, contained in Hutchings and Krawitz (1992).
d, can be calculated. In order to evaluate the strain from The technique has been successfully applied to calculate
the interplanar distances or from the lattice parameter, a the residual stresses and strains in several spacecraft appli-
reference sample free of strain is needed. The standard way cations. Titanium alloy propellant tank weldments were
to achieve this is to grind material from the various parts of measured; initially residual stresses approaching 250 MPa
the sample, particularly the weld metal and the parent metal, were recorded, post-weld heat-treatments reducing this to
as these may have different compositions. The powder pro- about 50 MPa. AA 2219 welded plates and AISI 304
duced is thermally treated and slowly cooled to room stainless steel components were also evaluated. These
96 3 The Integration of Materials into Product Assurance Schemes
examples are shown in Figs. 3.29a and b, courtesy of Prof. equilibrium positions. Other methods depend on the material
F. Rustichelli, University of Ancona, Italy. being ferromagnetic, so they may be useful for steels and
The Rustichelli work has been extended in order to nickel alloys. Stress has been shown to affect magnetic
measure the residual stresses in Ti6Al4 V fuel tanks fabri- parameters including a materials permeability, coercivity,
cated for space use by the welding together (TIGtungsten and redistribution of magnetic domainsall changes to
inert gas) of two cold drawn half-cylinders supplied by the these parameters from values made when the material is in a
author (Bruno 2000). Three welded curved samples were stress-free condition can be related to both applied and
investigated, each having different thicknesses. The welded residual stress levels.
6 and 2 mm samples behaved similarly (hoop stresses being
300 MPa in the weld, and 100200 MPa in the parent metal), 3.2.3.6 Modern Techniques for Surface Analysis
whereas the thinnest 1.6 mm sample possessed lower, A great number of surface analytical techniques have been
180 MPa hoop stress in the welded zone, and zero residual developed since the mid-1960s. All rely on the evaluation of
stress in the parent metal. samples under ultra-high vacuum conditions in order to
Very little work has been done on Ti6Al4 V welded eliminate the effects of surface contamination by residual
spacecraft tanks so the investigations were resumed on two gases. The scanning electron microscope (SEM) has become a
further 2 mm thick curved samples and the results compared common laboratory tool, and it has largely replaced the
to the weldment microstructures. These samples were stress transmission electron microscope owing to its ease of opera-
relieved after TIG welding (485 C for 4 h in a 104 torr tion. The characteristic X-rays generated at the surface of a
vacuum). The main result found that this stress relief treat- specimen by the incident electron beam in the SEM can be
ment reduced the weldments hoop stress from 300 MPa utilized to display or analyse the chemical elements present at,
down to about 160 MPa (Bruno 2004). and slightly below, the materials surface. The signals are
Thermal spraying is widely used in industry to apply usually detected by an energy-dispersive X-ray analytical
surface protection to engineering components. This process (EDAX) system. Several operational modes exist, and these
has been considered for space vehicle applications, including have been detailed by Goldstein et al. (1981). The
composite, ceramic and metallic coatings for protection EDAX-system (see also Sect. 3.2.2.3) has an elemental sen-
against heat as encountered during re-entry into the Earths sitivity of approximately 0.5 % if calibrated against good
atmosphere, or entry through the atmospheres surrounding standards, a lateral resolution of 50100 1010 m. This
distant planets and their moons. Such coatings have been method is virtually nondestructive to the specimen surface,
evaluated (Luzin 2010) for residual stresses by means of and under normal operation local heating effects will be
neutron diffraction; the evolution of residual stresses during introduced only if electrically insulating materials are
deposition were established for thick deposits (1030 mm) examined.
down to thin deposits having a thickness of 0.5 mm. Anal- A battery of alternative methods for the surface analysis
yses found stress distributions through the deposits due to of materials is available to the materials engineer; each has
the thermal mismatch that occurred between the substrate its own merits, and although Table 3.1 is introduced as a
materials and the coating. This is complicated by the means for the initial selection of certain analytical tech-
quenching stresses created by the molten particles as they niques, it is stressed that expert advice must be solicited for
rapidly solidify on impact. The interfacial stresses can be any particular application. Fourier transformation infrared
monitored and when high, spalling of the coating from the (FTIR) spectrometers are particularly useful for analysing
substrate can be expected. However, spalling, or detachment organic residues and examples of this technique are given in
may not occur if there is some solid solubility between the Fig. 3.30, and in Fig. 5.76. The FTIR equipment described
coating material and the substrate (without the formation of by Smith (2007) has uncovered many examples where
brittle intermetallics) and here, for metals, the relevant phase organic species have had a detrimental effect on electronic
diagrams need to be consulted. equipment at various stages of their livesthe case studies
are a recommended read and cover: acrylic adhesive and
(e) Less common methods paint residues found to damage ball bearing races in an
electro-optic system and stepper motors; brown residues of
These have been briefly covered by Wedgewood (1994), rosin on printed circuit boards which were identied as rosin
but are unlikely to nd wide applications. An ultrasonic derived from pine tree resin rather than synthetic resin;
method depends on the small changes to the ultrasonic methyl cellosolve transferred during pcb cleaning to pene-
velocity as ultrasound is passed through the material. The trate a connector and cause electrical insulation breakdown
change in velocity is very small, about 10 ppm, and is between the connectors pins; degradation of greases due to
caused by the strained atoms as they are pulled from their long term storage; and, carbonate deposits on electronics
operating in a sea-coast environment.
3.2 The Materials Laboratory 97
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 The primary functions of an analytical chemistry labora-
Test Laboratories tory are to determine thermal, optical and chemical properties
of organic materials. Activities and analyses will include:
These laboratories are essential support facilities for the Optical Spectroscopy: Fourier Transform Infrared Spec-
testing and evaluation of space materials. Most space trometry (FT-IR) for surface analysis, high resolution
agencies and prime contractors will provide laboratory ser- infra-red-imaging, depth proling, reaction kinetics, quality
vices and possess equipment to follow standard test meth- control, quantitative determination of chemical species, a
ods, but the true space labs will have dedicated systems for collection of IR-reference spectra for structural
more exotic tests, those that produce accelerated ageing and assignment/conrmation will be available from databases.
those providing for the simultaneous exposure of samples to Liquid and Gas Chromatographic Techniques: determi-
components of the space environmentso-called syner- nation of the type of monomer in polymers, structure elu-
gistic testing. Analytical laboratories will involve special cidation, quantitative determination, quality control, Mass
testing techniques and instrumentation. These sophisticated Spectroscopy reference spectra will be available from
procedures necessitate that trained and learned operators are databases.
available in order to obtain accurate, optimized and relevant Elemental Analysis: by Atom Adsorption Spectrometer,
results that will be useful to the projects requesting data. X-ray fluorescence and X-ray diffraction.
Numerous reference books can be consulted for detailed Environmental test and screening facilities have been
descriptions and space applications for these techniques described elsewhere in this book, for instance the micro-
(Dauphin 1992; Bosma 2009; Pretsch 2009; Gross 2011; VCM test which provides quantitative gravimetric measure-
Lund 2011; Kleiman 2013; Voigtlander 2015). ment of outgassing and condensable materials was illustrated
3.2 The Materials Laboratory 99
Fig. 3.30 Analyses of human contaminants. Fourier transformation fastener was also submitted to FT-IR analysis following the discovery
infrared (FT-IR) spectrometers are able to provide a rapid analysis of of particle contaminants within its threads. The SEM image is shown in
various organic and polymer contaminants which may be found on the (b) together with the IR spectrum of the white particlesa skin-flake
surfaces of critical spacecraft hardware. This has been made possible by reference spectrum is also shown for comparison. EDAX analysis also
developments in the speed and power of small, affordable computers. conrmed the presence of the elements: S, CI, K, Ca, and Na-detectable
Libraries of standard spectra are stored in the computer memory and ingredients of skin for this method. Further investigations indicated
may be compared against spectra from contaminant species by these to be small scales of dead skin (dandruff) (courtesy of Torbjorn
relatively simple search routines. Two case histories are illustrated Lindblom, Celsius Materialteknik, Karlskoga, Sweden.) a Fingerprint
(a) shows the optical photograph of a ngerprint on a polished steel contamination on polished mirror and analysis. b Skin flakes in bolt
mirror together with the FT-IR spectrum and its analysisthe nger thread and analysis
marks contain some water and typical skin secretions. A threaded steel
in Sect. 2.5. The condensable materials that are collected considered to be 8 km/s the stream of atomic oxygen on the
from the outgassing test can be chemically analyzed by front surface has a charge energy of approximately 5 eV.
comparison to spectra available in databases as shown in The interaction of atomic oxygen causes an optical glow
Fig. 3.31. The dynamic outgassing testing of samples, close to the spacecrafts surface. Few laboratories are able to
material pieces or individual components, can also be mea- produce an atomic oxygen (AO) flux and the majority of
sured by specialized equipmenthere the sample is con- realistic assessments of AO on materials have been made by
tinually monitored for mass loss against time, rstly at 25 the examination of samples returned from space, as will be
C, then by steps of 25 C every 24 h to a temperature of shown in Sect. 8.5. Ground laboratory Atomic Oxygen
125 C. This test can provide data about the long term effect testing generally follows the ASTM E2089 (2014) standard
of the space environment on a material, so supplementing practice. Facilities designed and commissioned for
the more simplistic VCM screening test (Rampini 2009). on-ground AO exposure include ESA-Estec, Noordwijk;
The effect of Atomic Oxygen on materials and parts will Physical Sciences Inc., MA, USA; Integrity Testing Labo-
be reviewed in Sect. 8.5 and have been extensively reported ratory, Markham, Canada; and, SAS Lab., Rome. AO pro-
by Banks (2003). The atomic oxygen existing in the low duced by methods that include the dissociation of molecular
Earth orbital environment has formed by the oxygen using a pulsed laser source. The lab typical AO
photo-disassociation of diatomic oxygen. These oxygen energy is about 5.5 eV and temperature of sample can be
atoms, O+, are highly corrosive with most of the materials controlled between 150 and +150 C in a chamber under
they contact. When a spacecrafts orbital velocity is vacuum of less than 105 mbar.
100 3 The Integration of Materials into Product Assurance Schemes
Fig. 3.31 Example of chemical analysis by infrared spectroscopy. correspond to the database stored spectrum for a commercial epoxy,
a Photograph showing the collected volatile condensable material as it they have no resemblance to the fluorocarbon bre contained in the
has condensed onto the central area of a standard test plate (ECSS sample (which had not outgassed/condensed). ECSS Q-ST-70-05
Q-ST-70-02 denes the outgassing-under-vacuum test). b Sample of an denes the method for detecting organic contamination of surfaces
IR spectroscope reading, showing the condensed species peaks around by IR spectroscopy (courtesy ESA)
3500, 2200 and 1600 cm1 on the horizontal axis. These peaks
3.3.2 Laboratory Records and Reports (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
One of the most important documents generated during readable (the original readings may be required again if
laboratory studies, whether for a PA function or as a basic instrument fluctuations are found at a later date etc.).
research and development activity, is a properly kept labo- (e) The initiator of work project, all reference to material
ratory notebook. The steps taken in the DPA, failure anal- history, environmental test parameters, time of failure,
ysis, or routine investigation need to be accurately recorded etc. should be recorded.
to prevent the loss of ndings, to validate the quality of (f) All preliminary examinations of the item and visual
experimentation, to enable the work to be repeated in an inspection results should be recorded, and special sur-
identical manner by other laboratories, and to prevent face features photographed; then possible nondestruc-
repeating the work at some time later because it was not tive tests performed and their results recorded; nally, a
known that the work had been tried and abandoned. record should be kept of how the item was cut to aid
In the space industry it is common that one of the QA further SEM and metallographic examination (e.g.
tasks is to ensure that laboratory notebooks are kept updated precise orientation of cross-section), the mounting
by technical staff. The records of laboratory activities, media (particularly if under pressure, or with exother-
together with the records of meetings with engineers from mic resin), polishing media, magnication of micro-
other disciplines, will be periodically required for compila- graphs, etc. (this will permit a similarly qualied
tion into a wide range of documents ranging from monthly worker to reconstruct and repeat the work as originally
reports to ofcial technical reports, working papers, papers performed).
for publication in journals or at conferences, and nally as (g) Supporting records should be glued, stapled, or taped
data to be presented at material review boards and failure into the laboratory notebook (particularly
review boards. digitally-stored photomicrographs, graphs, hardness
The materials and processes tasks performed in the space results, etc.).
industry are sometimes of a routine nature, but it is the (h) Adequate references to procedure specications, other
authors experience that certain test failures, or last-minute records, drawings, sketches, negative numbers (this
malfunctions, can be diagnosed only by a crash laboratory will facilitate upgrading of results into the format of an
investigation when operators may be required to work ofcial report) should be made.
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- 3.3.3 Report of Materials Data to Spacecraft
uled, or postponed pending further work or equipment Projects
changes. The accurate and extensive recording of experi-
mental or test data in laboratory notebooks is therefore A book concerning the uses of applied materials science and
essential. Sample traceability has been coved previously in techniques in the spacecraft manufacturing industry would
Sect 3.2.2.2an example of good met lab practice was not be complete without a short account of the role of report
illustrated in Fig. 3.7b). writing. A metallurgists or chemists work is never nalized
The author has found it necessary to apply the following until he/she has described what has been accomplished by
guidelines. When adhered to, they should obviate continuity the all the practical examinations, and this will necessitate
problems that might arise when staff are absent on leave or the writing of a nal ofcial technical report for his/her
sickness: organization. The impact of the report will be reflected by its
contents and the format in which it is presented. It is the
(a) Laboratory notebooks should contain bound, authors experience that materials examinations and studies
pre-numbered pages (to assure management that vital for a spacecraft project are usually performed to establish the
records are not missing because they were not led). quality of a product, to determine a mode of failure, or to
(b) All entries must be clearly written, and in a multi- assess the suitability of a materials or device for space use.
national organization, preferably in the English or Ofcial reports will be based on the detailed entries con-
French language (a notebook is of no use if an entry tained in the laboratory notebook, the contents of which
cannot be read). were discussed in the previous subsection. When novel
(c) The use of permanent ink is preferred (this will also experimental or test techniques were employed, a full
facilitate photocopying). 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 wide assortment of non-conformances. The MRB will
information. As internal reports, they will need to be written frequently draw upon the metallographic ndings of a lab-
for internal reasons, and only a summarized nal document oratory study when considering the dispositioning of articles
may be required. This will be circulated so as to feed back or materials. These dispositions will normally fall into four
information to other departments, to justify design changes, categories, such as:
and for PA archival purposes. Long records of details will
not be needed, and in most cases the recipients will only be suitable for use on spacecraft,
interest in, or have the time to read about, the state of the suitable only after article has been repaired or
engineering problem in the introduction and what has been refurbished,
done to solve it in the conclusion or summary. not acceptable, return to suppliers, and
In any presentation, photographs, sketches, and photo- not acceptable, scrap.
micrographs are extremely helpful. When they are suitably
captioned they can reduce the need for long descriptive The failure review board (FRB) is convened to review
narratives. They will also be useful during oral presentations failures, evaluate failure analyses of the kind illustrated in
when materials or metallographic data will need to be shown this book, and establish corrective action plans. Again,
and discussed at material review boards and failure review specialist reports are discussed, and the recommendations
boards. made as a result of the metallographic studies are usually
Frequently it will be necessary to dispatch material to accepted. Corrective actions might include engineering
chemical laboratories, (e.g. Sect. 3.2) for testing or accurate changes, materials changes, and inspection or test procedure
analyses of the constituent polymers or alloys. The presen- changes.
tation of the major alloy elements may be important to The metallographic presentations at both MRBs and
identify the alloy designation and, from that, the supplier. FBRs are required by QA and reliability disciplines, but they
Aluminiumlithium alloys may soon nd some spacecraft will always be important to the design and production
applications. It can be seen from Table 3.2 that AlLi fully engineers who were responsible for the choice of material
heat-treated sheets can have large density and mechanical and manufacture of the part. Care must be taken to avoid
property variations due to very slight modications in their laying the blame for a problem or failure on any individual
elemental composition. person. The oral presentation or report must be conned to
Materials review board (MRBs) (increasingly termed metallurgical and other technical considerations, and it is
NRB, non-conformance board) consist of members respon- preferable to avoid reference to political areas. This will (or
sible for engineering production. P A, and materials or should) be taken up at a higher level in the organization. It is
components specialists. Such boards meet to critically also advisable to give maximum credit to the production
examine new materials, new fabrication processes, and a 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 103
been well-made (spacecraft parts are often one-off items and often found industrial placements as part of their university
may consist of complex shapes made from unusual, course (Puig 2006). Usually material student trainees will
difcult-to-machine materials). It is important that the nd themselves somewhere between engineering, spacecraft
materials engineer does not take on the reputation of being integration and quality assurance teams. They will spend
an over-critic. Full cooperation between all personnel will be vacations or up to one year receiving such training in prime
needed during future troubleshooting exercises, and these and sub-tier space companies, being paid and supervised so
are hindered when police tactics are used by PA laboratory as to make full benet of the study or managerial function
staff. undertaken. Topics covered can range from wire harness
In summary, keep clear and legible records; never spec- assembly methods to metallurgy, composites, surface treat-
ulate publically as to the potential causes of a failure, par- ments, bonding and testing materials in a simulated space
ticularly if the work has not been completed; never make environment. As observed by Puig, this is a win/win sit-
casual or flippant remarks on any aspect of your work; do uation; often the student obtains enough information and
not criticize colleagues. In the space business if there is an laboratory data to be able to submit his/her thesis, while the
accident, the engineer is keen to discover the cause. It is the company may benet from a well-researched report, better
writers (fortunate) experience that spacecraft accidents do access to local university facilities and an insight as to
not result in disputes, arbitration, and costly litigation whether the student could be a potential candidate for
involving the legal profession. This would have a negative post-graduate full employment.
effect, as lawyers want to determine who was responsible The use of a smaller test-house laboratory generally gives
and this could cause conflict. In other elds, engineering a better customer service, being more competitive and more
disputes and claims are part of business and, here, the personalized. Small laboratories may, at the same time and
reporting of materials data may be unfavourable to the client on account of their size, be in a weaker position than the
who may decide to bury the report and look elsewhere for larger laboratories as they need to concentrate on the certain
another opinion (Allen and Jubb 1991). research or test disciplines that they do best, and a contra-
dictory situation will exist if they attempt to cover them-
selves too widely.
3.3.4 Training of Materials Engineers When complex spacecraft technologies need to be
and Laboratory Staff assessed in a materials laboratory there is a need for both
good technician resources and graduate materials specialists.
Visits to different materials or metallurgical laboratories in So-called generalists may also be recruited to coordinate
Europe have shown some national differences concerning the various work orders or investigative projects; these staff
the deployment of staff. Certainly differences exist which may know less of the ingredients of the different disciplines
varied according to the size of the facility; however, it is felt but more about how they relate together. The skilled tech-
that continental companies, particularly in Germany and nicians must understand the objectives of the laboratory
France, placed a much greater emphasis on machine-shop work, particularly if it is part of a technology assessment or a
crafts, understanding of laboratory equipment, and inter- materials failure analysis, so that they can intervene and alert
mediate skills than is experienced in Britain or Scandinavia. the specialist during an investigation whenever a new, or
In all laboratories there is an amplied role for technicians, unexpected, situation arises. It is particularly important that
and whereas some continental companies will upgrade their materials lab staff understand the full history of the sample
technician staff by sponsorship through technical university they are preparing so that facts are not lost during, for
courses, the converse is also seen, whereby university instance, the specimen cleaning stage, or during
graduate engineers are deployed in technician roles. Many hot-mounting when microstructural evidence can be easily
graduates are too specialized or research-oriented to become destroyed by the introduction of heat and pressure. The
easily socialized into an industrial or commercial labora- integrity of the material sample preparation method is also
tory, and they have difculty coping with the upkeep and receiving more attention, particularly when quality controls
repair of modern laboratory equipment, liaison with service and failure investigations question aspects related to coating
departments such as machine-shop personnel, or even porosity, the particulate congurations within metal matrix
operation of simple lathes for the preparation of specimens. composites, and the presence of microcracks within envi-
When graduates are flexible and do integrate well into a ronmentally tested electronic components and interconnec-
company laboratory it is found that small rms provide for tions. The principles of material investigations, such as
more congenial surroundings, and this promotes a higher metallography, chemical analysis, fracture mechanics test-
level of loyalty and more successful on-the-job training than ing, etc. are taught at Technical Universities and Colleges,
the larger, more anonymous rms with their systematic and an extensive library of video courses has been produced
training programmes. Materials technology students are by the American Society of Metals International so that rms
104 3 The Integration of Materials into Product Assurance Schemes
can run on-site training with the help of acknowledge and sustainability of resources. But also during their work
experts. It is also clear that the staff of test laboratories all which can involve dangerous lab chemicals, hazardous
around the world are making a growing use of the so-called materials at the launch siteconsiderations must be given to
information highway afforded by Internet. These staff, personal welfare, health and safety. It may be necessary to
including those of the author, are nding other materials make a risk assessment before performing dangerous tasks.
engineers and scientists with similar interests. For instance, take care/precautions while sawing and pol-
There is a need for laboratory staff to be retrained every ishing beryllium, and handling vessels that might contain
ve or ten years. The active life of a rst university degree residuals such as hydrazine. Materials and processes engi-
can be as short as ve years according to Professor Goodhew neers hold a privileged and trusted position in society and
of Surrey University, and owing to the rapid pace of tech- therefore should only perform services within their areas of
nological advance a materials graduate may need to attend competence. As previously stated it is essential that their
refresher courses six times during his/her entire career. New hierarchy understand the need to keep up their skills and
material developments are continually occurring, particularly knowledge bases. Evidence gained from lab examinations,
in the elds of carbon bre composites, high-performance studies and failure investigations should be interpreted and
oxide ceramics, high-temperature superconducting ceramics reported honestly, accurately and without bias. These state-
particularly for large passive electronic devices such as ments may seem to be obvious, but under pressure from
RF cavities and antennae, super alloys manufactured by inert project or clients management, engineers and lab staff need
gas automization, and hot isostatic pressing, etc. Engineers to avoid improper influences that may lead to cutting corners
and technicians must also be kept up-to-date with new lab- and making conclusions based on insufcient test or his-
oratory techniques that may be available only for loan or hire torical evidence.
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- 3.4 The Future for Materials Failure
structed cryostats, or the use of radioactive trace techniques Investigations
for the investigation of absorption rates of, for instance,
water, by the surfaces of materials following exposure to 3.4.1 The Larger Company
space simulation environments. New concepts in metallo-
graphic preparation may also need to be established, par- The majority of the larger companies engaged in the man-
ticularly when small, unique samples are presented for ufacture of aircraft structures and components, marine sys-
failure analysis. Here there is no opportunity for the tradi- tems, road vehicles, and the construction of power stations
tional trial and error approach, and a systematic sequence and telecommunications equipment, will have already
of grinding and polishing will require the selection of pol- invested in the capital equipment required for the smooth
ishing cloths, abrasive materials and their size, speed of running of a metallurgical failure laboratory as described in
polishing, and pressure for different groups of difcult Sect. 3.2. Failure analysis is becoming increasingly impor-
materials (Bouseld 1988, 1992). In short, read the litera- tant for a number of reasons. Some failures are very costly,
ture, go to the talks and conferences, and when and where and not only can valuable equipment become destroyed but
necessary augment formal training with comprehensive high-volume product output may be interrupted. Downtime
courses. can cost hundreds of thousands of pounds per day whether it
Finally, it is also important that there should be a strong be associated with an aircraft, power station, or satellite
interaction between the metallurgist, chemist or materials communications link. It is probable that in the future,
engineer and the mechanical, electrical, or aerospace engi- material costs for heavy engineering units will be dispro-
neer responsible for the hardware being evaluated. This will portionately more expensive than they are today so that
enable the problem-oriented topic to be adequately addres- investment in the tools for performing failure investigations
sed with a good mutual understanding of design, product will be increased. In this case the cost of determining a
processing, or service operational environment requirements failure can be justied.
an extremely useful form of on-the-job training. On the other hand, nancial constraints now necessitate
that commercial electronic systems, such as television
receivers, computer hardware, and even land-based data
3.3.5 Ethical Issues relay stations, incorporate the technique of redundant circuits
and throw-away modules. This means that there is seldom a
Professional engineers and laboratory staff, as will be noted requirement for engineers to investigate the mechanisms that
in Sect. 3.5, are expected to pay regard to the environment result in failures. Other factors in this argument include the
3.4 The Future for Materials Failure Investigations 105
growing shortage of skilled maintenance technicians and the peaks and troughs encountered by space hardware pro-
wide availability of mass-produced low-cost electronic cir- curement cycles, Lens part-nance the purchase of state of
cuits: thus the throw-away maintenance philosophy could the art equipment for their partners who, in turn, manufac-
become more and more appealing. ture and deliver piece-parts (Leijtens 2015). Engineering
design, together with electronic and mechanical assembly, is
made in-house. Product verication and space-qualication
3.4.2 The Smaller Company testing (e.g. the complete range of environmental testing) is
performed in external facilities. The tasks of material and
In general, it is the authors view that, within the smaller component analyses, destructive and failure analyses, as well
company (e.g. small, medium enterprisesSMEs), top as reporting, are also out-sourced as a joint collaboration
management is not giving the M&P function all the attention with internationally accredited laboratories. Although unu-
it deserves. One particular issue that may be especially sual, this philosophy of working has been approved by
important in the space industry is the complaint that the various space agencies.
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 3.4.3 Product Liability
PA. Another reason is that frequently top management is
unaware of the need to employ a team of materials or Products made for the domestic consumer have, in the
component parts specialists to perform DPAs and FMAs. It course of their use, proved to be defective, fail, and cause
is not until costly failures occur that consideration is given to injury. It is becoming more common that Product Liability
the need of an on-site materials laboratory for performing Laws are transferring the nancial consequences of a failure
controls on raw materials and processes, and the occasional more directly onto the producer. This will, in turn, cause the
defect or failure analysis work. legal profession to make more use of the failure analysis
The many cases the occurrence of failures has meant laboratory to unravel the clues, and then place responsibility
resorting to lab testing in order to answer the immediate for the failure. Product liability is becoming a serious matter
question, Why?. This is not simply to satisfy curiosity but in the commercial world, and it seems that these matters may
to avoid repeating the same mistake. Certain cases have also become evident in the space industry (Thoma 1981).
demonstrated the forensic engineering nature of the inves- The space environment is ideal for some metallurgical
tigation. Here, by reconstructing the sequence of events processes, such as pressure or diffusion welding under vac-
leading to the material failure it should be possible to uum, but the working risks will be high, so that it is unlikely
establish and report on the underlying cause of the problem. that astronauts will be called to perform routine repairs in
Against this background it is important that top management low Earth orbit. Secondly, spacecraft hardware that is not
in SMEs realise the contribution that laboratory work can situated in very high equatorial orbits may be retrievable by
make to the success of spacecraft projects. If a smaller means of the crew of a space plane such as the former fleet
company cannot afford to employ materials/components of Shuttles or Hermes, and laboratory analyses could then
specialists, or have available materials/components labora- identify those individual manufacturers who produced
tories, other alternatives must be found. It is often experi- defective hardwarepossibly responsible for breakdown of
enced that national research establishments and local a multi-million pound commercial Earth resources satellite
universities possess excellent facilities, and their staff are or space observatory. It is evident that PA will continue to
more than willing to participate in solving space materials play an essential role in future projects, and that product
problems. These experts can give advice for the choice of safety policies will place great importance on laboratory
materials and production methods, and they will perform techniques to ensure a higher product reliability.
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- 3.5 Greener Spacecraft
ded from the ESA Business Incubation Centre and based on
technologies developed for space application within the As a consequence of mans actions, the oceans, air, and land
Dutch Micro Technology development programme). This suffer from ever-increasing exploitation and contamination.
two- to three-man company, operating in the Netherlands, This has led to the implementation of various national and
works together with highly specialized partners (not called international measures, laws, and regulations. These range
suppliers). Rather than investing themselves in costly from requirements that operators continuously monitor par-
machines which could remain idle during the well-known ticulate emissions from metallurgical industries such as
106 3 The Integration of Materials into Product Assurance Schemes
foundries during metal production, to the partial or total competition between international companies. At the present
banning of hazardous materials and ozone-depleting chem- time there is an attempt to curb massive exploitation of our
icals. Until recently, large-scale monitoring was possible environment by the worlds large manufacturing industries.
only by local, on-site detectors such as those situation close It is therefore very important that the space industry should
to chimney stacks. Aircraft are also used to routinely check also monitor its own operations and ensure that it is not
for oil pollution in coastal areas, air pollution above indus- contributing unnecessarily to the range of lethal fumes and
trial sites, or monitor for any increase in the toxic chemicals toxic wastes discharged into the surroundings, and that it too
that seep from landll sites into waterways and the soil. With utilizes the Earths material resources in the most sustainable
the advent of Earth observation satellites, extensive and way. Most engineers also consider themselves to be envi-
repetitive surveys have become possible and in some ronmentalistswe care about the environments of both
instances these can be more effective and cheaper than Earth and space, and the need to protect them for the future
routine airborne surveys. Radar sensors overcome obscuring of our children. It is up to engineers to provide the tech-
clouds and haze by using the microwave spectral bandwidth nology for the solutions to environmental problems. The
for collecting surface information independent of weather German word for engineer is ingenieur, which comes from
and viewing conditions. Often optical satellite imagery is the Latin word meaning ingenious. Landll or incineration
sufcient to show the required information (see, as exam- disposal of scrap, byproducts, ground station gantries (see
ples, the images presented in Fig. 3.32). Fig. 3.33) and unused spacecraft hardware is optimal.
The continuity of satellite observations will be ensured by Environmental regulations may eventually lead to a ban on
the second generation of remote sensing satellites being both so recycling, which has the potential to offer revenues
designed, fabricated, and launched. These include the long beyond the space projects end of life, is certainly the
European ERS-2, Envisat and Sentinal spacecraft which best option. The recycling of electronic items will be further
have demonstrated highly sensitive detectors for global discussed below, and an assessment of the metals, organic
ozone monitoring, the US Upper Atmospheric Research compounds and oxides available to be recycled or extracted
Satellite, Landsat, and NOAA, as well as a multitude of from waste spacecraft electronic boxes will be given in
instruments on-board the International Space station. Sect. 3.2.2.
The constant developments in materials and processes However, the arguments and balances which determine
ensure that all forms of spacecraft become lighter and more the selection of materials and processes when environmental
accurate. Advances in their various technologies are main- factors are introduced to the equations are complex. For
tained owing, to a large extent, to the existence of worldwide instance, it is widely known that weight-saving is a critical
Fig. 3.32 a. SAR multitemporal image made by ERS of an astronomers light pollution should be made an offence. The scourge
85 85 km area around Rome. The technique requires several satellite of light pollution may be one of the most detrimental spin-offs of our
passes over the same area and reveals pollution and oil-spills in the sea, technological age; excluding younger urban generations from the
wave patterns, some topographical relief, and changes in soil moisture experience of a truly dark sky and the stars. (Picture courtesy
(courtesy of ESA.) b. A satellite optical photograph of Europe lit up at International Dark-Sky Association, Tucson, Arizona)
night, evidence of the wasteful use of energy resourcesto some
3.5 Greener Spacecraft 107
Fig. 3.33 ESAs launch complex ELA-2 as seen in August 2011. It is building for the vehicle and integration of its satellite payload. The
extensively damaged by wide-spread corrosion of its painted mobile gantry was rolled back to its parked location prior to the
zinc-plated-steel support structure, its roof structure and internal launchers lift-off. One month after this picture was taken, ELA-2 was
platforms. This facility launched 119 space missions of the Ariane 2, brought down by a controlled explosion and dismantled. The metal
3 and 4 series over the period from 1986 to 2003. It is split into two structure and electronic contents were stripped down, compressed and
zones, the launcher pad itself (to the left of picture) and the mobile returned by ship to Europe for recycling (photograph, Dunn)
gantry, seen here 950 m to the right, which provided a protective
factor in the selection of satellite materials, particularly when manufacturing industry is forever changing because of reg-
the cost of taking one kilogram into a low Earth orbit is very ulations and rules, many emanating from the EU. Most often
approximately 2000. A kilogram saved by selecting observed are the restriction, or compliance, of materials
beryllium in place of an aluminium alloy for a structural part resulting from REACH (pertaining to restriction and
might seem attractive. However, the extra cost of imple- authorization of all dangerous chemicals) and RoHS (ban-
menting the safety requirements which control the toxic dust ning dangerous substances that are used in electronic prod-
and scrap generated during machining of beryllium and ucts). The interaction between REACH and RoHS is
disposing of this waste in an environmentally friendly way complex but complementary, and both can be related to
may mitigate against this advanced material. This is despite WEEE (waste electrical and electronic equipment). These
the fact that when weight-saving factors are translated into regulations impact the space industry, but are beyond the
overall launch vehicle weight, there should have been a scope of this book. Already the author is aware of several
saving of fuel or an increase in the number of payloads, large organizations and companies in the space industry who
which would have in turn meant less pollution from the are making studies to nd alternative substances for an
propellants being red during launch. arbitrary table of which might be called, The most haz-
High-quality workmanship and a good standard of living ardous materials to be presently used in the manufacture and
both at home and at work have been achieved in the modern launch of spacecraft. This table contains the following
world, and to some extent this will be maintained by the materials:
implementation of internationally agreed specications, such
as ISO 9000. This document has brought into focus the need Ozone-depleting chemicals (ODCs) such as the
for effective working standards and defect-free products. The chlorofluorocarbons (CFCs) used most extensively as
document also aids in the quest for greener products. The solvents in the cleaning processes for metals and
108 3 The Integration of Materials into Product Assurance Schemes
electronic assemblies. The most common such chemical cadmium-containing silver solders can pose a serious
which was used in the space industry and has now been health hazard to those making these joining operations,
almost totally replaced is trichlorotrifluoroethane, also even when fume extraction is provided. Alternatives are
known as CFC-113. This substance was also used in the available, such as zinc plating and cadmium-free joining
simulation of more hazardous liquids, the fuel hydrazine alloys. Note, both cadmium and zinc are prohibited from
and its oxidizer nitrogen tetroxide, during the ground items exposed to the space environment owing to prob-
testing of propellant tanks and associated reaction-control lems associated with sublimation and whisker growth, as
equipment. See also Sect. 4.21.1. described elsewhere in this book.
Chromate primers which passivate metallic surfaces Lead-bearing solders have been banned from use in
including aluminium and magnesium alloys and silver by domestic plumbing to prevent ingestion of lead from
simply immersing the parts in a chromate solution to drinking water. Lead and lead-containing substances are
form a coating of chromate on the base metal. These very toxic and pending legislation and U.S. Environ-
conversion coatings are widely specied as very mental Agency regulations threaten to ban the use of lead
cost-effective means of enhancing the corrosion protec- in the solders used widely in the electronics and space
tion of spacecraft surfaces (see Sect. 4.5.3) and increas- industry. The main reason for this is the belief by some
ing the adhesive strength of paints or adhesive joints to politicians that by discarding domestic electrical goods
structural surfaces (see Sect. 5.10). However, inhalation into landll sites there will be an ecological problem in
of the chromate ion is recognized as potentially car- years to come due to the seepage of lead-bearing com-
cinogenic, leading to serious health and safety concerns pounds into our water supplies. The long historical use of
about its use in the workplace. Unless properly handled, tinlead solder, reaching back to artisans in Mesopotamia
chromates can cause ulceration and environmental con- in 4000 BC, means that there is an extensive database of
tamination. As yet, despite extensive testing of new solder properties. At the present time both tinlead and
products (e.g. Alodine 5700, Nabutan STI/310 and Iridite indiumlead solders are the preferred solders for space-
NCP), no effective, low-toxicity alternatives appear to be craft electronic assembly. Alternatives, having alloy
commercially available (Pereira 2008). The most com- constituents that are essentially nontoxic, with a good
mon chemical conversion coating (CCC) currently used oxidation resistance, and an ability to wet and form
for the improvement of corrosion resistance on alu- metallurgical bonds to the wide range of metals used in
minium alloys is Alodine 1200, manufactured by Henkel. the electronics industry, are still being evaluated. Early
This is a chromium (V1)-based CCC. Due to the envi- works found none of the lead-free alloys to possess the
ronmental impact of hexavalent chromium this product is overall good properties and low cost of eutectic tin-lead
likely to be forbidden for space-use in future years. An alloy (Akinade et al. 1995; Jacobson and Humpston
important property of Alodine 1200 is that has a low 1995). Little progress has been made and it is certain that
electrical resistance and can comply with spacecraft the reliability of lead-free products will depend on the
requirements for electrical bonding (to ensure the craft component package type and circuit board material and
has proper electrical grounding, an equi-potential design. Certainly, due to the higher melting temperature
between sub-systems and avoidance of corona dis- and higher strength of lead-free solder alloys versus
charge under vacuum). A resistance of less than 0.0025 eutectic tin-lead, the reliability of reworked and repaired
ohms is possible for 3xxx, 4xxx, 5xxx and 6xxx-series electronic circuits will be severely reduced (Ma 2012; Ma
alloys is achievable with Alodine 1200 coatings. The 2013; Konoza 2012). Compared to tin-lead, the lead-free
high-strength alloys, 2xxx and 7xxx with the same processing windows are reduced when assembly and
coating, have electrical resistances below 1.0 and often rework are considered, there is less margin for error. Pcb
below 0.1 ohms which is suitable for removable contacts assemblies that would have been acceptable with Sn-Pb
(for instance, between an electronic housing and a process parameters are at risk of delaminating and failing
structure when both are Alodined). during operation when lead-free solders are employed
The ground equipment associated with launch sites and (Reid 2007). For these, and many other reasons, ESA has
tracking stations should avoid the use of cadmium as a advocated the application of Lead Free Control Plans for
corrosion protection nish on mild and low-alloy steels the space industry (Dunn 2012). One might question the
owing to the highly toxic nature of cadmium and its impact that the electronics industry has had on the
by-products. Even the pickling of cadmium-plated parts worldwide usage of lead? In the USA it is estimated that
can cause signicant quantities of cadmium to enter the only 0.6 % of all lead consumption is in the electronics
effluent system. A number of platers have already been industry, whereas 80 % is used by the car battery industry
penalized for discharging cadmium. Similarly, (Craig 1994). These values are also probably reflected in
silver-cadmium brazing alloys and the traditional European industrial consumption. 100,000 km of lead
3.5 Greener Spacecraft 109
sheet was used in the UK during 2010, mostly, for house prototype thrusters are being tested in Italy using
roofs (Anon 2011). The real conundrum is that the same monopropellant H2O2 and bi-propellant H2O2 with
persons who are attempting to ban lead in electronics also ethane. Another material, that is said to have the potential
want to reduce the use of fossil fuels and cut down on the to replace both AP and hydrazine is ammonium dini-
number of petrol-driven vehicles by replacing them with tramide (ADN), NH4N(NO2)2. Swedish tests on a solid
battery-driven electric cars (based on lead acid AND propellant and an AND-based liquid monopropel-
batteries!). lant have been successfully tested (Larsson 2011).
Lead in brass and bearing materials, can be thought of Another, so-called green propellant is Hydroxyl
as an essential ingredient. It is normally added as Ammonium Nitrate fuel/oxidizer mix, also is known as
between 2.5 and 3.5 %wt. The lead is present as discrete AF-M315E. This was developed at the Air Force
islands within the copper-zinc matrix, usually at grain Research Laboratory at Edwards Air Force Base in
boundaries. It is trapped within the brass dual grain California and is proposed to replace hydrazine in
microstructure, being insoluble in both alpha and beta thruster motor applications (Hawkins 2010).
grains. Lead ensures less tool wear during machining as Brominated flame retardants, (BRFs) and less fre-
it acts as a solid lubricant, it provides for a more uniform quently chlorinated flame retardants are added to ther-
surface nish compared to lead-free brass, and piece moplastics, electronic packaging and printed circuit
parts can be machined to better dimensional tolerances board materials to achieve a reduced or zero possibility
with less cutting force, and chip size (Pantazopoulos of re. These substances do have a damaging effect on
2008). health as they release dioxins when disposed of by
Volatile organic compounds (VOCs), such as the sol- incineration. They may be prohibited, but in removing
vents that are mixed with paints, are harmful to operators BRFs from electronics, there can be a signicant negative
and must be extracted from the work-area. They are also impact on human health as the risk of re may cause
environmentally hazardous, so cannot be simply emitted injury or death.
into the outdoor atmosphere. Regulations in European Particulate emissions, particularly during the machining
and other countries mean that all VOC-emitting plants and grinding of toxic spacecraft materials, must be con-
must be registered and monitored for environmental tinuously monitored to satisfy environmental legislation.
pollution by a number of authorities. These VOC sol- The monitors for particle emission may utilize the prin-
vents have little productive use once the paint has been ciple of particle impingement (triboelectricity) whereby
deposited on the part. The space industry utilizes many the collision of a particle with a metallic surface (here the
different paint products (for both corrosion protection and sensors probe) causes a charge transfer to occur between
thermal control). The answers are to use abatement the sensor and the probe. This charge generates a signal
equipment together with higher solid-content paints, with which is recorded in the sensor electronics. A data logger
up to 75 % solids. can store and display the dust particle numbers collected
Green propellants, how they can be identied and uti- daily or weekly, and this information will satisfy the
lized as launcher and thruster motor fuel, is one of the Environmental Protection Act in the UK. It is especially
greatest challenges for propulsion engineers. We are necessary to monitor the dust emitted during the
presently almost totally dependent on the family of machining and shot blasting of materials such as beryl-
hydrazine-based liquid fuels, and ammonium lium and beryllia, which are toxic in the nely divided
perchlorate-based oxidizer (AP) for solid fuels, both of form; such particles are shown in Figs. 5.15 and 5.41.
which are highly toxic and carcinogenetic. Legislation Advanced materials are slowly becoming accepted into
presently under discussion in the EU, and a common the commercial manufacturing industry, but there are
desire in the space business, points to the need to steer some unlikely factors which might limit this trend. One is
away from these substances. If hydrazine is replaced, the desire that it should be possible to recycle all of these
then its oxidizing agent dinitrogen tetroxide will proba- materials. Although this argument must be taken into
bly also be abandoned. Apart from the toxicity of this account, the most important reason for material selection
liquid propellant combination, both agents have good, is its performance during service life. In the case of the
well proven, long-term storage capabilities on ground expensive beryllium alloys referred to above, it is clear
and in space; in operation they have a very high specic that scrap material or unused spacecraft parts can be
impulse; and once ignited combustion cannot fail. recycled. But this involves the careful handling, storage,
Alternatively, hydrogen peroxide (H2O2) is a green oxi- and traceability of scrap which can then be sold back to
dizer (Meaker 2006) and has been used with kerosene for the metal producer. More problems have been experi-
the main propulsion of the UK designed and manufac- enced with the aluminiumlithium alloys earmarked for
tured Black Arrow launcher seen in Fig. 3.34. Presently aircraft structures where there can be extreme wastage.
110 3 The Integration of Materials into Product Assurance Schemes
Between 80 and 93 % is lost during the milling of parts contaminated with lithium, and even small amounts
for the Tornado project. These aluminiumlithium alloys (30 ppm according to Cleave and Morton 1989) will give
are 10 % lighter and 10 % stiffer than current aluminium unacceptable levels of porosity in conventional alu-
alloys; however, they are also between 3 and 4 times the minium alloy castings.
cost of the alloys they are designed to replace, so it is The electrodeposition of metals such as gold, silver, and
clear that their recyclability is an important factor during copper is made from cyanide baths for purposes such as
material selection. Various approaches are being carried electrical contacts, conductors, and solar reflectors. This
out by the manufacturers of these alloys, but it is now calls for extreme care when operators are making cyanide
clear that great care must be taken to avoid the mixing of additions to the tanks. Management must be sure that
scrap AlLi with other aluminium alloys, as this would spent plating solutions and rinse water solutions are fed
mean that the secondary metal chain would become into an approved waste treatment process. Several
3.5 Greener Spacecraft 111
companies have decided to attempt the change to 3.6.2 Elemental Distribution for Spacecraft
noncyanide-plating baths. Only copper plating was suc- Electronic Box
cessfully switched to a noncyanide, alkaline electrolyte
for both rack and barrel installations. Some obstacles and
Background
difculties may be encountered which are overcome by
Some years ago European space scientists proposed and
changing to a different pre-plate chemistry, and more
designed a spacecraft with the scientic objective of per-
frequent in-house bath analyses are needed until a com-
forming a complete global survey of the Moons physical
fort level is established for the consistency and reliability
and chemical properties. This complex programme, which
of the product. It has been found that the noncyanide
was later abandoned, relied on a low attitude lunar orbiter in
copper gives better plating coverage in
combination with a relay satellite. The project was intended
low-current-density areas than the cyanide copper, but
to supplement the wealth of data already obtained from the
overall it has a 1015 % slower deposition rate.
six landing sites provided by the US NASA Apollo missions
As will be seen in Sect. 3.6, it may be more cost effective
(see also Sect. 8.1.2). Spectrometers and multi-spectral
to retain the use of some banned substances and utilize
stereo-imaging would produce a map of the elemental
mechanical processing in the recycling of waste elec-
abundances over the Moons surface and also map its min-
tronic equipment.
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
3.6 The Potential for Recycling Electronic map would be of great interestthis should indicate the
Waste concentrations of these elements to an accuracy of at least
20 %, across the entire lunar crust and so establish the lateral
3.6.1 General and vertical heterogeneity of the Moons surface.
Chemical analyses would be made using two instruments:
The prospect of recycling electronic waste materials became
intriguing when the Institute of Materials, Minerals and X-ray spectrometer
Mining declared in 2012 that the urban gold mine will be up The experiment observes x-rays emitted by the lunar surface
to 50 times more lucrative in the coming years because of as a result of solar x-ray excitation.
the 320 tons of gold, 7500 t of silver and immeasurably more
tons of tin going annually into products such as computers, Gamma-ray spectrometer
televisions and phones. Until now only 15 % of these Chemical mapping to a depth of about 30 cm from the
materials are recycled and recovered, but this will become a Moons surface might be achieved by observing the
lucrative business for urban minerswww.gesi.org. gamma-rays produced from the excitation of nuclei by cos-
In particular, gold and the rare earth metals, often the mic rays (mainly secondary neutrons). In polar regions, H2O
subject of political bartering, should be recycled because was thought detectable down to a concentration of 0.4 %.
although their concentration is low, it is still higher than in It had been proposed that both spectrometers should be
mineral ores. Electronic equipment can be crushed, shred- mounted on booms in order to minimize the effect of
ded, magnetically separated and by means of hydro- and spacecraft-produced background emissions that would
extraction metallurgy, these valuable metal elements can be reduce the accuracy of the lunar surface analyses.
recovered; far better than diverting potentially toxic mate- Under ideal circumstances the spacecrafts own back-
rials to land-ll sites. Mechanical processing of pcb assem- ground emissions could be characterized by operating the
blies allows for the recovery of 80 % of the constituent two spectrometers during their flight towards the Moon.
metals, particularly copper which represents 75 % of the A knowledge of this elemental distribution could be con-
metallic fraction (Veit 2002). Pyrolysis by conventional and sidered as a perturbation factor which, when computer
microwave heating can be used to separate the majority of modelled, would be subtracted from the crafts own analysis
organic compounds from the inorganic materials, and the of the lunar surface. Studies of the Moons gamma radiation
resultant gasses can be used to fuel the further pyrolysis of have been made by Russian Lunas spacecraft, the US
printed circuit board assemblies (Hall 2007; Sun 2012). The Apollos 15 and 16 as well as numerous other craftthe
isolated metals can then be recycled following traditional gamma-ray spectrometers and X-ray fluorescence spec-
routes for scrap metals. 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 3.4 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) Some chemical analysis data could be established from
and Schrunk (1999). the certicates of conformance issued by the material sup-
Special attention would have been paid to the materials pliers. Here full analysis records were available for the actual
selected for the construction of the spacecraft in order to aluminium billet used to make the electronic housing. Sim-
reduce this chemical background. For instance, the struc- ilarly, analyses were available for purchased batches of
ture could be made entirely from aluminium, beryllium or aluminium foil and stainless steel screws.
titanium. It was recognized that a major part of the space- The remaining items, such as the external paint, potting
crafts mass comprises of electronic boxes, each loaded with foam and the various electronic components were individ-
high-density printed circuit board assemblies. ually crushed, pulverized and their material constituents
The space scientists requested that a computer model be measured by various methods of chemical analysis. Special
made pertaining to the global chemical (elemental) content efforts were made with the soldered joints. Here the
of a typical spacecraft electronic box. copper-plated through holes were mechanically removed
from several multi-layer circuits and submitted for analysis.
Method and Results Full chemical analyses have been attempted. Fluorine,
A discarded flight-spare unit from a successful space project chlorine, bromine, iodine, sulphur, phosphorus, etc., have
was used for this work (Dunn and Nicholas 1998). Initially been detected by emission spectroscopy, X-ray fluorescence
the box was opened, the printed circuit board assemblies spectroscopy and activation analysis. Gas chromatography
were de-mounted from their supports, potting foam was and infrared absorption spectroscopy have been used to
removed and the electronic components were de-soldered. separate and determine the volatile organic components.
114 3 The Integration of Materials into Product Assurance Schemes
Non-volatile components have been identied by both ionic is somewhat limited by the fact that different analytical
and non-ionic chromatography on ion exchange columns. techniques were used to establish the weight percent of the
The metallic elements were determined by Spark Source constituent materials. The identication of organic and
Mass Spectroscopy (SSMS) of the pulverized powders. intermetallic compounds has not been recorded. The main
Further analysis was made by dissolving some of the oxides discovered from analysis are listed in Table 3.4.
metallic particles in mixed acids and measuring the metal ion
content of the solution. The recording of these results might be useful for future
An attempt has been made to compile all the chemical scientic projects. They may also be of interest to organi-
analysis results and this is shown, based on elemental dis- zations involved in the recycling of electronic hardware.
tribution, in Table 3.3. The accuracy of the tabulated results
Spacecraft ManufacturingFailure Prevention
and the Application of Material Analysis 4
and Metallography
4.1 Sources of Failure was set by one of the top dynamics engineers at 10 gRMS.
Although this was based on good judgement and previous
Potential sources of failure are frequently identied during experience, the equipment manager sought out a transducer
the manufacturing and assembly stages of spacecraft pro- expert for corroboration. The channel setting was raised
duction. They may related to one or more of the following: from 10 gRMS to 15 gRMS by the transducer expert (who was
conservative by nature). The qualication plan was then
unsuitable aspects of design, given to the project management and another group of
incorrect material selection, measurement experts, who added their margins to the test. It
human error (e.g. machine tool operator becomes is not surprising that, after further stops along the organi-
careless), zational chain, the test ended up at 50 gRMS.
materials imperfections (e.g. internal porosity or harmful The relationship between human issues and engineering
inclusions), failures has been addressed by Le May and Deckker (2009)
defects introduced during reworking, repair or modi- and researching routes to failure-avoidance found that
cation of existing structures, human causes were responsible for 80 % of accidents,
corrosion damage during storage. so-called unintentional mistakes. These issues can be caused
by: organizational change, where staff may change their role
Whether materials are custom-made for the space indus- in a company without having the necessary re-training;
try or bought off-the-shelf from general commercial markets, fatigue and loss of alertness; failure to maintain equipment
there will be a need for careful planning and the use of a and perform calibration tasks; and, failure to write and/or
wide range of quality control techniques. These will monitor comply with manufacturing procedures or process identi-
whether fabrication processes have been under control and cation documents.
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 4.2 Drawings and Workmanship
both destructive techniques, such as mechanical testing, as
well as nondestructive investigations like radiography, dye 4.2.1 Design and Manufacturing Drawings
penetrant testing, and ultrasonic testing. Each of the potential
material failure sources listed above can be evaluated by The design of spacecraft is based on the practices developed
metallographic inspection either by conventional optical initially for the aircraft industry. The main goal of the design
microscopy or by more sophisticated procedures such as engineers is to minimize the structural mass without jeop-
scanning electron microscopy, and scanning laser acoustic ardizing the integrity of the spacecraft structure during the
microscopy. phases of test, launch, zero-gravity, and possibly re-entry.
Failures are occasionally caused by human error and Almost equally important are the designs of the many sub-
may relate to the overtesting of perfectly spaceworthy systems for communication, power, attitude control, and the
equipment. This can be illustrated by the following anecdote like, which must also survive extensive testing and launch
that was brought to the authors attention. Several electronic loads. The design of the individual piece-parts that are used
equipment units were to be qualication tested and had been to construct the nal subsystems may undergo several
heavily instrumented with vibration and shock accelerome- changes from the initial concept of the spacecraft design.
ters. Early in the programme the vibration range for the test Depending on the model philosophy, these changes might be
needed following the testing of a thermal model, engineering letters that refer, sequentially, tothe part family (e.g.
model, or qualication model. As discussed in Sect. 2.2, the electronic housing, support strut, bearing, thruster
freezing of the design should occur at the time of the critical motor, or electrical drawing);the spacecraft job
design review, but some changes do occasionally happen number;the assembly area location;and the draw-
after this time and will need to be qualied prior to the ing serial number and issue number.
launch readiness review. An overview of various spacecraft (b) The accurately drawn picture of the part with all its
structure designs is given by Turner (1995), and is outside views, dimensions, and surface nishes. This must be
the scope of this book, but, as will be illustrated in the fully understood by all operators both in the parent
following chapters, initial designs are sometimes totally company and those located at sub-contractors sites.
inadequate owing to the complex interactions between Dimensions are usually in millimetres. Tolerances,
materials, their surface nishes, and their exposure to sim- where necessary, need to be specied for radii (changes
ulated space environments. in section can lead to stress concentrations), surface
An engineering drawing will be assigned to each single roughness, and interlocking parts. Abbreviations
part. It is important that these drawings are closely controlled should only be used if they are universally accepted.
(under conguration control) and that they have a standard- (c) A list of materials needed to make the part must be
ized format. The large numbers of spacecraft manufacturing shown on the drawing.
contractors, combined with the potential for numerous (d) The manufacturing processes must be listed in the
changes to dimensional detail, material nish, etc., necessi- sequence of production. This may commence with saw
tate that drawings originating from one design ofce are also cut, and continue through dip braze, cleaning,
fully understood by designers and production persons, not nondestructive inspection, stress relieve, shot
only in other companies, but in other countries and on dif- blast, chemical conversion coat, paint-colour, and
ferent continents. Drawings will be prepared to contain an bake. Each process identied for flight hardware
identication number, an issue date, and details that cover the should be accompanied by a process identication
complete manufacturing picture of the product. It is probable document reference number and its issue number. The
that each engineering drawing will be approved by the per- issue number is important. It must reflect the actual
sons responsible for design, selection of materials, produc- process document and process parameters utilized for
tion, and product assurance. All materials and processes will the production of qualication hardware.
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 4.2.2 Workmanship Standards
spacecraft systems in Appendices 10 and 11. This approval
activity occurs both at the preliminary design review and, Bitter experience has shown that things do not work if a
again, after modications, at the critical design review. These certain number of manufacturing precautions and controls
reviews will consider factors such as suitability of materials are not exercised. This is particularly true for automated
for structural use (mechanical properties, fracture mechanics, processes such as resistance spot welding or the making of
resistance to stress corrosion cracking), operation under mechanical joints such as crimped electrical connections.
vacuum (sublimation and outgassing), and suitability for Here all process variables must be evaluated, and, once
production (solderability of electronic materials, weldabil- optimized process schedules have been attained, they must
ity). The processes identied on the engineering drawings, be preset and maintained during production runs. Unfortu-
and later tabulated on the projects DPL, will also be scru- nately, the quality of most metallurgical joints is difcult to
tinized to see if they have been suitably veried or qualied assess by visual inspection alone. In these instances it is
and whether the process documentation, which ensures important to produce so called technology samples during
reproducibility, has also been approved. preproduction runs, and these will be inspected and then
The engineering drawings are therefore of great impor- subjected to destructive tests such as mechanical pull and
tance during design, production, approval reviews, and, metallography. Duplicate samples can also be used to gen-
should failures occur, during the failure investigation. erate workmanship standards or visual aids to illustrate the
A summary of the useful data that must be recorded on the quality characteristics of joints by reference to various
drawings is as follows: machine settings (e.g. pressure, time, current).
Workmanship (Dunn 2008) can be dened as, the skill
(a) Drawing number and revision number (contained in a with which something was made or done. It is the product
box at the lower right-hand corner). The number will of effort or endeavour and relates very much to the skills of
usually be a long string of individual numbers and 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 during production, in-line quality control samples will be
shop floor assembling spacecraft hardware. Such skills are taken at regular intervals and subjected to rapid metallo-
acquired by training, and it is the physical skills, such as the graphic preparation for comparison against these standards.
handling of a blow-torch or soldering iron, which may result This will ensure that the variable settings on automated
in either excellent workmanship, or shoddy workman- process equipment have not altered since they were initially
ship. See also Sect. 6.14.2 concerning training schools. set up, thereby giving added condence that the overall
A set of photomicrographs, photographs and sketched quality levels of flight equipments will remain consistently
illustrations may be used as workmanship standards and, as high.
illustrated in Figs. 4.1, 4.2, 4.3 and 4.4, they can dene Optical microscopy is also widely used to examine pol-
process acceptance and rejection criteria. When required ished metallographic samples for inclusion content. The
Fig. 4.3 Workmanship standards exist for a wide range of welding unacceptable weld bead width and the degree of permitted meander.
technologies and specic standards need to be consulted that relate to These illustrations are reprinted from ASME BPE-2014 (pp. 132 and
the metal alloy, the design and the welding equipment. Spacecraft 135), by kind permission of The American Society of Mechanical
weldments can be visually controlled to ECSS Q-ST-70-39 and this Engineers. All rights reserved. No further copies can be made without
standard also contains workmanship guides for accept/rejection of written permission from ASME. a Acceptable, b misalignment
surface oxidation on post-welded titanium and stainless steel parts. The (mismatch), c O.D. concavity, d I.D. concavity (suckback), e lack of
illustrations shown here dene acceptable and unacceptable weld penetation, f convexity. A Acceptable weld bead, B excessive weld
proles for tube welds. They depict what might be observed on the bead width variation, C excessive weld bead meander. General note
polished and etched surfaces of microsections made either during weld Applies only to nonproduct contact surfaces and only if weld on
parameter set up, or during in-line quality control sampling. The lower product contact surface cannot be inspected
set of illustrations depict visual inspection criteria for acceptable and
simplest method is to survey the as-polished sample and metallurgical structure that exists in the joints. It should be
compare the number and form of inclusions with standard noted, however, that satisfactory surface appearance does
charts. This method is employed for the assessment of not guarantee a sound metallurgical bond. Solder defects are
thin-walled pressurized devices which might otherwise leak usually classied as major, minor, or cosmetic. Major
if an inclusion were to occupy the wall thickness. The defects could reduce the operational life of the electronic
assessment of size, shape, and distribution of inclusions has hardware and must always be repaired or reworked (see
also been successful in the selection of material for space- Sect. 6.12.3). Minor defects, such as borderline solder vol-
craft pyrotechnic devices as will be discussed in Sect. 5.4. umes (either excessive or insufcient), generally do not
Finally, workmanship standards exist for all the designs cause failuresfor high-reliability electronic circuits they
of solder joints employed for spacecraft electronics. They are would be reworked. Cosmetic defects do not affect the
used to determine the quality of solder joints through visual performance or operating life of the product (e.g. component
criteria, as there is a degree of correlation between surface tilting, small folds on the solder llet); they do given an
appearance and shape of the solder llet with the undesirable appearance and indicate that the production
120 4 Spacecraft ManufacturingFailure Prevention
(b)
122 4 Spacecraft ManufacturingFailure Prevention
assembly process needs to be better controlled before real convolutions were damaged, and that in these regions the
problems occur. Solder contact angles less than 90 (positive section thickness of the tubing was greatly reduced and
wetting) are acceptable for solder joints; those greater than hence mechanically weakened.
90 (negative wetting) are unacceptable. The goal of per- Complete beryllium copper removal had occurred at the
forming visual inspection against documented workmanship pinhole site. The detail of one of the defective areas (etched,
standards is to maintain process control and minimize Fig. 4.5c) shows the damaged internal wall of the tube to
rework. have slightly collapsed. This indicates that a compressive
Workmanship standards related to Area Grid Arrays force had been applied to the tube surface. Reagents for
(AGAs) are reviewed in Sect. 6.15.10. AGAs are particularly microetching metals and alloys are detailed in Appendix 3.
difcult to assess because the terminations, all present on the Also, the microstructure of the undamaged tube is shown
underside of these packages, are often impossible to visually to consist of non-deformed equiaxed grains. In the damaged
inspect and some newer devices have more than 1000 region, the grains have been slightly deformed; they contain
interconnections. X-ray systems have been evaluated a high density of parallel clip bands and mechanical twins
(Wickham et al. 1999) and many of the resulting which change direction from grain to grain. Such features
X-radiographs were incorporated into the ECSS standard for are characteristic of mechanical damage that must have
surface mount technology (ECSS Q-ST-70-38). Similar occurred subsequent to the precipitation treatment because,
workmanship standards have been published by the Institute at 300 C, they would have been annealed out.
for Printed Circuits (IPC 610D and IPC 7095) and the Joint At this stage, it was possible to rule out the initial
Electron Devices Engineering Council (JEDEC-217). The judgement that penetration had been caused by forming or
major defects addressed in these standards include: corrosion. A visit to the manufacturers shop-floor revealed
misalignments, missing solder balls or columns, presence of that an inexperienced soldering operator had assembled this
voids in the solder, non-wetted pads on the pcb, and the particular batch of waveguides. Many of his llets had
presence of small unwanted solder balls created during the contained more solder than was normal. By means of pulling
soldering process that might eventually cause short circuits. a ne cord, loaded with abrasive grit, around the llets, he
Such NDT inspection is tedious and some advancements in was able to remove the excess solder and produce a clean
the automated X-ray inspection of AGAs are available for bright llet. In doing so, this action had also produced wear
this work in order to reduce variability and decrease the time and excessive thinning of the waveguide tube. The charac-
needed for inspection (Said 2012). 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
4.3 Mechanical Damage Revealed treatment of the tube.
by Microstructure
Figure 4.5a shows a flexible waveguide from a batch which, 4.4 Hydrogen Embrittlement
at nal inspection, was rejected owing to the occurrence of
small pinholes adjacent to the tube-to-flange soldered joints. 4.4.1 Interaction of Metal with Hydrogen
These defects were visible only at 10 magnication. The
tube material was solution-treated beryllium copper and the Since the early days of alloy development there has been the
pinholes were limited to the rst two tube convolutions. challenge of hydrogen. Owing to their smaller diameter,
At the start of the investigation, two causes for the defect hydrogen atoms can be easily absorbed into metal lattices,
had been proposed by the waveguide manufacturer: rst, the particularly those with large interstices like Al, Pd, Ir, Mo,
tubes had been supplied with drawing defects which had Ta, W, Zr, and Ti, and with difculty into metals with small
been opened out by the heat dissipated during the soldering interstices such as Cu, Ag, Au, Ba, and Cr. The reaction of
operation, and, second, the strong soldering flux required to hydrogen with metals may result in the formation of
produce smooth and bright solder llets had chemically hydrides, molecular hydrogen, and gaseous products of the
corroded the tube. Another opinion was that the soft reaction between hydrogen and inclusions or alloy phases
solution-treated alloy had been penetrated by sharply cor- held within the metal matrix. In all these cases hydrogen has
nered tooling jigs used to bend the tube into its spacecraft a harmful effect on the mechanical properties of the metal
conguration. After forming, the assembly had been and causes embrittlement or brittle fracture (Raymond 1988;
precipitation-hardened at 315 C for 2 h and then soldered to Robertson and Teter 1996).
the flange at 240 C. Several theories have been developed to explain the
The defective region was microsectioned and polished interactions between hydrogen and metals. They presume a
(see Fig. 4.5b). It can be seen that at least two of the generation of very high internal pressures caused by the
4.4 Hydrogen Embrittlement 123
recombination of atomic interstitial hydrogen to form the pickling and plating of aluminium waveguides and steel
hydrogen gas, a trapping of dislocations and hence a springs; and the brazing of commercial copper component
reduction in mechanical deformation by the increased solu- parts in a hydrogen furnace. The following subsections
bility of hydrogen in the strained zones of the material, a represent typical investigations associated with the hydrogen
reduction in the bonding forces between metal atoms in the embrittlement of spacecraft equipment.
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 4.4.2 Hydrogen Embrittlement of Spring Steel
numerous examples exist where spacecraft hardware, after
successful trials, will suddenly exhibit hydrogen embrittle- Several small springs which had been installed for 2 days
ment. Examples are the methanol embrittlement of titanium, within the equipment of an engineering model satellite were
the hydrazine embrittlement of certain stainless steels, and heard to snap into two halves before equipment level testing.
the hydrogen embrittlement of battery cells. The metallurgical analysis found the failed springs to have
Laboratory failure investigations have highlighted the fact been fabricated from a standard 0.4 % carbon low-alloy
that atomic hydrogen may be introduced into spacecraft spring steel which had been hardened and tempered, des-
piece-parts during their production by the forging of end caled, and nally plated with, rst, copper and then nickel.
domes and barrel sections of titanium propellant tanks; the The spring fracture surfaces were examined in a SEM
heat treatment of steel mechanical parts and titanium bolts; (Fig. 4.6a), and seen to consist of delaminated bands sepa-
the welding of apogee boost motors made of titanium alloy; rated by dimpled rupture regions. The fracture path direction
ities, so enabling unrestricted rotation and low signal loss. A comprehensive programme of tensile testing, bend, and
The rotors were then withdrawn from their cavities, so that sustained load tests were made on welded material from the
small permanent magnets could be adhesively bonded to their flight batch. Chemical analyses were made together with
bases. The heat cycle used to cure the epoxy adhesive was extensive metallographic inspections for the presence of
seen to produce blisters in the plated layers. Hydrogen atoms titanium hydrides in order to acquire an acceptable con-
absorbed during plating were thought to have recombined to dence level in the flight unit.
form molecular hydrogen at the plating-to-substrate interface. The investigation results showed that the weld tensile
As shown in Fig. 4.7 the disbonded plating is raised locally to strengths lay between the values expected for pure titanium
form a surface blister. The blisters prevented reinsertion of and Ti6A14V, and the welds showed a good ductility, far
each rotor into its matched cavity as it had blocked the 2 m greater than that experienced in the Saturn rocket tank fail-
air gap. Future batches of rotors were hydrogen relieved by ure. The sustained load test in this programme (at
careful baking treatments involving slow heating cycles to 220 N/mm2 for 14 days) was designed to induce delayed
200 C for 4 h between each plating operation. This redis- degradation of the weld sample due to additional migration
tributed or released any absorbed hydrogen to give a of hydrogen and secondary precipitation of hydrides, as was
blister-free surface. observed in the aforementioned helium tank failure. No
delayed embrittlement was encountered, and no titanium
hydrides were associated with the fracture path. Also, the
4.4.4 Examination for Titanium Hydride chemical analysis showed that interstitial elements were
Precipitates lower than required by the raw material specications.
The entire spectrum of possible microstructures, indicated
A manned spacecraft payload support structure has been by the phase diagram inserts (see Fig. 5.18), has been
constructed of Ti6A14V alloy struts. Fitting and tubular observed on the polished and etched specimens. These are
parts are machined from wrought alloy bar, whereas conical shown and detailed in Fig. 4.8ad. The captions describe the
parts are made as alloy castings. These parts have been detailed photomicrographs. Titanium hydrides were
tungsten-inert-gas (TIG)-welded. Unalloyed titanium-ller observed in small isolated clusters set in the weld bead face
metal rod was inadvertently used in making each weld. run, but not with the weld root run. No hydrides were pre-
These 0.045 in. diameter (1.14 mm) rods had been supplied sent in the HAZ or parent Ti6A14V material. No hydrides
with a certicate of analysis purporting the material to be Ti were associated with any of the fracture paths seen by
6A14V. Spectrographic analyses made after samples and metallography of the mechanical test specimen.
flight equipment had been built showed that the user ller It should be noted that titanium and its alloys are more
rods were in fact pure titanium. Interstitial elements (i.e. difcult to prepare by metallography than steels. Of partic-
hydrogen, nitrogen, and oxygen contents) were not deter- ular importance when investigating for hydrides, is that the
mined. Following these welding activities all remaining mounting material is a low-viscosity epoxy resin that cures
noncompliant welding rods were removed from store and are slowly without much exothermic reaction, so that the tita-
no longer available. nium sample is not heated by more than 10 C. Some
The use of unalloyed titanium ller metal in the welding acrylic, fast curing resins can reach very high temperatures
of titanium Ti6A14V and other titanium alloys is common which, like hot-mounted specimens in a press, can reach up
practice in the aerospace industry. However, the materials to 150 C and these temperatures can modify the morphol-
and process must be extremely well controlled in order to ogy of any precipitated TiH within sections microstructure.
prevent embrittlement of the nal weldment from interstitial A good procedure for the specimen preparation of titanium
elements (i.e. hydrogen, oxygen, nitrogen). The solid solu- and its alloys is given by Vander Voort (2008).
bility of hydrogen in titanium alloys, with the exception of The macrographs of each weld show the root weld to
beta-alloys, is very low at room temperature and this results have become particularly well alloyed with the parent metal
in the precipitation of a hydride phase (Robertson and Teter forging. The presence of beta-stabilizing vanadium in this
1996). The hydrides are brittle and have a detrimental effect portion of the weld will absorb interstitial elements to a far
on the mechanical properties. As a consequence of titanium greater extent than pure alpha titanium. The lack of titanium
embrittlement problems encountered by NASA [most nota- hydrides within the rst properly alloyed weld deposits is
bly the failure of a helium pressure vessel which led to the explained by the presence of vanadium. As the strength of
destruction of a Saturn rocket during a test sequence (Wil- the tested weld sample was at least six-fold better than the
liams et al. 1979)], the welding of commercially pure tita- expected maximum service stress, a decision was made to fly
nium to alloyed titanium is now prohibited. the payload support structure.
126 4 Spacecraft ManufacturingFailure Prevention
The particular case-history from the European Spacelab material review board exercise which led to the decision to
project resulted from a breakdown in raw materials control use as is was extremely costly. Alternatively, any decision
by the relevant Materials Quality Engineer. The welding rods to reject the nished hardware would have led to even greater
were incorrectly identied by the vendor as having the costs and some delay to the launch date. Welding wire mis-
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- caused failure of the main propulsion engine and damage to
pany accepted the vendors paperwork and did not bother to its test stand when improper welding wire was used to weld
forward a sample of welding rod to the Materials Acceptance together the steer horn hydrogen lineinstead of Inconel
laboratory for chemical spectrographic analysis. Had this 718, a weaker weld wire, Inconel 600, was used, and here the
been done, the welding rods would have been seen to have mistake was also blamed on a supplier.
the wrong composition (unalloyed titanium) and they would As concluding remarks, titanium has a reputation for
have been rejected and returned to the vendor. Instead, the being a difcult material to weld. This is largely due to
rods were used to weld flight equipment. Fortunately, this unfamiliarity with the material and the necessity to take
equipment was found to be metallurgically acceptable but the anti-contamination precautions. The surfaces to be joined
4.4 Hydrogen Embrittlement 127
need thorough cleaning, even ngerprint greases can lead silvercorrect shielding, satisfactory
to weld porosity. The shielding gases (usually argon or light straw and dark strawslight contamination, but
helium, or a 75 %Ar25 %He mixture of the two) must acceptable
have a very low dew point (below 50 C) in order to dark blueheavier contamination, may be acceptable
prevent rstly the pick-up of hydrogen by the liquid weld depending on service
metal, and later the formation of titanium hydrides. Rubber light blueheavy contamination, unlikely to be acceptable
hose for transportation of the shielding gases is to be grey-blue, grey and white (loose deposit)very heavy
avoided, as this can release moisturevinyl plastic is contamination, unacceptable.
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 4.4.5 Embrittlement of Copper
generate interface colours ranging from light straw to dark
blue and grey. A rule of thumb classication for the Tough-pitch copper can, if insufciently deoxidized, contain
acceptability of titanium welds is given by Ellis and Gittos up to 1 % cuprous oxide distributed throughout the copper
(1995) for the colours of the weld zone. Workmanship matrix as small discrete included particles. Such material
standards that compile acceptable and reject surface was machined to form electronic component piece-parts,
colourations are now illustrated in the latest, 2015, ECSS including items for use within travelling wave tubes that
welding standard, Q-70-39. The following is a guide: were subsequently furnace heat-treated and brazed with
Fig. 4.9 The brazeability of copper is generally excellent, but formation of high-pressure steam. A sample of oxygen-free high
difculties may occur during furnace brazing in hydrogen atmospheres conductivity copper was simultaneously heated in the same atmosphere
that have a reducing potential. Oxygen-containing or tough-pitch to produce the fully annealed grain structure shown in (c, d). Note It is
phosphorous deoxidized coppers are subject to embrittlement when difcult to obtain a scratch-free micrograph with the tough-pitch copper
heated above 450 C. a, b show the effect of heating tough pitch copper as the porosity traps abrasive with mechanical polishing and chemical
to 850 C in a hydrogen atmosphere. Grain boundary embrittlement and polishing pits the pores and enlarges them. Courtesy of Mr T. Pinder
internal cavitation are caused by the reaction of copper oxide and the
128 4 Spacecraft ManufacturingFailure Prevention
silvercopper eutectic ller metal. Hydrogen-reducing have to be designed to withstand strongly corrosive envi-
atmosphere furnaces were employed. At temperatures ronments during its useful life of possibly 40 years in orbit.
above approximately 450 C absorbed hydrogen reacts with Such equipment includes battery units, heat pipes and sys-
the inclusions to form steam in the microstructure: tems, and pressure vessels containing liquid fuels. Excepting
emergency conditions, space hardware on-board proposed
Cu2 O + 2H ! 2Cu + H2 O steam) spaceplanes, will not be exposed to uncontrolled terrestrial
environments. However, pre-launch phases and ground
Owing to this reaction, local high pressures cause the
storage periods for single shot rockets and reusable space-
ssuring of the copper grain boundaries, leading to the
craft may be long and it has been necessary to develop and
complete break-up of the part by intergranular failure.
implement a plan which ensures that all structural or
Only oxygen-free copper grades can be considered for load-bearing metals are adequately protected from general
such piece-parts, particularly if brazing, welding, or surface corrosion and stress corrosion cracking. Manned
heat-treatment is to be performed in a reducing atmosphere. space modules such as those of the International Space
A routine test may be performed to assess the suitability of Station will also be subject to humidity and condensation on
copper grades. It is recommended in ASTM B379, and it internal cold surfaces. Assessments of corrosion protection
involves annealing off-cuts from the copper batch at 800875 nishes are generally made by accelerated testing, using
C for 20 min in atmospheres containing at least 15 % hydro- saline solutions (e.g. 3.5 % NaCl) for structural hardware,
gen. Typical results are shown in Fig. 4.9ad. The test speci- and temperature/humidity exposure for electronics. Specic
men must then pass a subsequent bend test without cracking. cases and test environments will be described in other Sec-
Brazing procedures are discussed further in Sect. 6.10. tions of the book.
For electrical connections, it is highly desirable to select
metallic combinations from the approved couples detailed in
4.4.6 Future Developments Tables 2.1 or 2.2. However, it is also important to note that
many of the less noble metals shown in this tabulation
The design of hypersonic vehicles fuelled by hydrogen will require additional protection from general surface corrosion
be exposed to a variety of hydrogen environments. Pressure in the form of platings, conversion coatings, anodic lms,
vessels and fuel management devices will be exposed to paints, etc., and additional polarization testing is often nec-
liquid hydrogen at cryogenic temperatures. Motors and essary. Metals never considered for use in structural systems
exhaust nozzles will experience high hydrogen temperatures of a spacecraft are widely used in the fabrication of its
at great pressures. At temperatures above about 650 C most electronic equipment. Some of the rarer metals are found in
structural materials will become saturated with atomic transistors and micro-miniature circuits. As an example of
hydrogen, and, as such, these materials can be considered as the potential problems to be encountered from dissimilar
hydrogen-alloyed (Nelson 1988). On cooling, the matrix material corrosion some of the different alloy systems used
may become supersaturated with hydrogen, leading to in the construction of various spacecraft electrical systems
property degradation due to hydride formation or the cre- are indicated in the following list:
ation of molecular hydrogen. Further damage may result
from a reheating cycle when dissociation of metal hydride 1. Iron and steel (ferrous alloys) are used as component
particles causes a localized hydrogen supersaturation. leads, magnetic shields, transformer cores, brackets,
Attention must also be drawn to the fact that hydrogen in racks, and general hardware.
its atomic form will react with graphite at ambient tempera- 2. Aluminium and aluminium alloys are widely used in sys-
tures to form methane gas. Metalmatrix composites con- tems as antennas, structures, chassis, supports, and brackets.
taining graphite and silicon carbide bres may readily react 3. Magnesium alloys are occasionally used as equipment
with hydrogen to form cavities of high-pressure methane gas. housings.
4. Stainless steel is used for mounting racks, brackets, and
4.5 General Corrosion Problems hardware.
5. Copper and copper-based alloys are generally used in
4.5.1 Bimetallic Corrosion-Related Failures systems as contacts, springs, leads, connectors,
printed-circuit-board lands, and wire.
Corrosion has to be considered during all manufacturing and 6. Fibre-reinforced plastics are used in waveguide fabrica-
pre-launch phases of spacecraft. Certain equipment may also tion (graphite bres), circuit boards (glass bres), and
4.5 General Corrosion Problems 129
component support adhesives (metal and oxide-loaded nish to provide excellent corrosion protection to the mag-
epoxies). nesium alloy AZ31 (Yao & Basaran 2013). Another Cr-free
7. Nickel and tinlead plating are used for protective coat- chemical conversion coating for magnesium alloys is the
ings and for solderability purposes. The use of tin in MAGNOLYA treatment (European Patent pending) being
solder is a well-known application; however, tin-plating process-developed by Tecnalia. This coating can be applied
is also common on RF shields. by immersion of parts into the solution, or by brush. It leaves
8. High-purity electrodeposited gold has wide application in a thin coating on the magnesium alloys so far tested
electrical connectors, printed-circuited runs and edge (EV31A, AZ91 and AM60). The coating process involves
connectors, miniature coaxial connectors, semiconduc- pickling and activation pre-treatments, followed by
tors, leads, and contacts. phosphate-permanganate-rare-earth conversion coating and
9. Silver is used normally as a plating material over copper drying. It appears to be suitable for parts with very close
in waveguides, miniature and microminiature circuits, tolerance mating surfaces and may be a candidate for
wires, contacts, high frequency cavities, and RF spacecraft housings, such as the one shown in Fig. 4.10a.
shielding. The bare MAGNOLYA coating on EV31A meets one of the
space industrys requirement standards, ASTM B-117, as
From this listing it is evident that the potential for gal- shown in Fig. 4.10be. The test samples shown in Fig. 4.10
vanic corrosion is inherent in the design of such equipments; (Brusciotti et al. 2014) were exposed to salt spray at 6 off
this will become apparent in later chapters. from the vertical and exposed for 168 h to the salt spray.
Until recently, magnesium alloys with their high Additional salt spray test exposures were made on EV31A
strength-to-weight ratios have been frequently selected for samples with resin and primer coatings applied to both the
the construction of spacecraft structures and the housing of Cr-free coating and its reference Cr6+/Mn coating; the results
electronic systems. General spacecraft requirements now found the MAGNOLYA coated system to be comparable or
stipulate that all units must be electrically grounded to their superior to the presently-used system.
structures, and, at the same time, possess an adequate surface Of the many hard anodizing treatments developed for
corrosion protection system. This has led to the virtual magnesium, the HAE and DOW 17 processes are the most
exclusion of magnesium from present spacecraft design commonly used, and, when present as a thick lm, they offer
(however, selective copper plating, see Sect. 4.28, small excellent abrasion resistance. A useful highly protective
areas of the magnesium skin can be solder-connected with nish for spacecraft (even when exposed to tropical launch
wires to the grounding structure in alternative designs where site environments) was the application of a sealing layer to
the entire remaining surface will be painted). Commercial the anodic lm comprising a thin, clear high-temperature
conversion coatings, such as DOW 7, may be suitably baking epoxy resin, followed by a chromated epoxy primer
conductive but, being thin, they frequently give rise to cor- and epoxy top coat. It would appear that such nishes are
rosion problems (see Fig. 4.10a). Figure 4.10a shows the resistant to temperatures of 200 C (Rendu and Tawil 1988).
effect of slight condensed moisture on an experiment hous- More recently, the plasma oxidation treatment of magnesium
ing constructed of magnesium alloy after equipment quali- alloys, which utilizes micro-discharges to change to surface
cation by thermal cycling. A non-conformance review of the alloy into hard ceramic layers, has become wide-
board considered that the housing had been removed from spread. Here the magnesium is both protected from corro-
the thermal chamber while still very coldleading to water sion and this coating also possesses extremely good wear
condensing onto all surfaces. The white, dust-like magne- properties (Blawert and Srinivasan 2010).
sium oxide corrosion product was also considered to be a Contamination of an aluminium alloy cooling loop by a
contamination hazard. few minute particles of copper swarf (Fig. 4.11) has been
New, high-strength magnesium alloys and metal matrix identied as promoting the failure of this system. Connec-
composites having high strength-to-weight ratios may nd tion of copper, or its alloys, to aluminium is a most unfa-
new applications on launch vehicles. Here, corrosion must vourable galvanic couple, particularly in aqueous
be avoided by the strict control of surface contamination, environments. In this instance, a small number of hemi-
flux inclusions from welding operations, galvanic couples, spherical pinholes, originating from the inner surface of the
and ensuring that surface protection schemes are suited to channel, have been produced by the localized corrosion cell
the application. Chromate conversion coatings will delay the set up between the cathodic copper particle and the alu-
onset of natural surface oxidation for a short time, and it is minium. Away from the pinholes, the aluminium surfaces
preferable that they are coated with resin-based paints. One have not been corroded.
chemical conversion coating free of hexavalent chromium Figure 4.12 details a severe corrosion pit within a
(Cr-free) is based on molybdate; recent work shows this camera-pointing mechanism constructed of aluminium alloy
130 4 Spacecraft ManufacturingFailure Prevention
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 131
Fig. 4.15 Al-2219-T851 (forged rings, no surface protection) follow- solution treatment. These CuAl2 particles have no doubt been broken
ing exposure to a 3.5 % NaCl solution for 30 days. This micrograph up and redistributed during the forging treatment, but they exist in a
clearly highlights pitting corrosion adjacent to the coarse network of displaced network that had originally formed during solidication of
electro-positive intermetallic particles (red-orange, arrowed). Most of the cast metal. The micrograph also shows both Al2(OH)3 (white,
these large inclusions are shown up by polarized light as predominately arrowed) and green CuCl corrosion products seeping from corroded
CuAl2 which, owing to the high copper content of this alloy, was micro-channels. See also Fig. 4.19d for further explanation. 500
unable to dissolved completely into the saturated solid during the magnication (Dunn 1984)
weight) of 2219 generally depresses the electrode potential based on the need for a material having an optimum com-
of aluminium in the cathodic (more noble) direction, as bination of properties including those of mechanical strength,
shown in Table 2.1. fracture toughness, and resistance to general corrosion and
From both a mechanical strength and an optimum corro- stress-corrosion cracking. This material is commercially
sion resistance stand-point it is essential that, during alloy available as sheet and plate, extruded rod, and bar; it can be
fabrication, copper should be fully dissolved into the alu- forged and made available as Alclad sheet and plate.
minium. It is the solid-solution, or homogenizing, treatment In the 1980s, Spacelab (see Fig. 3.1) became the largest
that dissolves the copper, and this is performed in the tem- joint space programme undertaken by Europe and the USA.
perature range 535 5 C. Cooling from this temperature must Since its rst flight in 1983 and 1998 there have been 22
be rapid to prevent the formation of the intermetallic com- Spacelab missions flown by the Space Shuttle, and the
pound CuAl2 at the grain boundaries. If excessive numbers of overall flight performance of Spacelab has been excellent
CuAl2 intermetallics do form at grain boundaries they are (see also Sect. 8.7.3). The welded 2219 module structure has
surrounded by adjoining volumes of alloy depleted in copper, remained leak tight in orbit even though it has been
and this will facilitate corrosion attack as shown in Fig. 4.15. assembled and disassembled many times for experiment
The decisions to select the structural alloy Al-2219 for payload changes. It has outlasted the design requirement for
manned spacecraft such as the Space Shuttle, Space Station, 10 years life (even though the number of missions during
Spacelab, the Columbus Module and Ariane 5 and 6, were this time has not reached the 50 which were also expected).
134 4 Spacecraft ManufacturingFailure Prevention
storage of spacecraft equipment in an uncontrolled sea-coast Al-6061 Good resistance to marine environments. Nature
environment and THERMAL CYCLING UNDER of attack influenced by thermal treatment.
VACUUM, an in-orbit space environment known to degrade Al-7075 Attacked by a marine environment and generally
surface coatings due to a mismatch in the coefcient of less corrosion-resistant than other wrought alloys.
expansion between the coating and substrate. The properties
All specimens were obtained from rolled plate and
of the alloys and their coatings were assessed by measuring
machined to a standard surface roughness before being
infrared reflectivity, solar reflectivity, resistivity, optical
Alodined for three different periods to produce three layer
properties (SEM and photographic recording), outgassing
thicknesses. A typical coating is seen in Fig. 4.18a. The
rates, and infrared absorptivity.
conclusions from this study were as follows:
The general corrosion properties of the unprotected alloys
are as follows:
(a) The surface structure of the chromate conversion
coatings was similar for all alloys.
Al-2024 Poor resistance to marine environments. More
(b) The corrosion properties were found to depend on the
susceptible to corrosive attack than other groups
copper content of the substrate and thickness of the
of wrought-aluminium alloys. Nature of attack
Alodine layer. As the copper content increases, so the
influenced by thermal treatment.
corrosion resistance decreases. The thicker coatings
always resulted in better corrosion protection.
(c) The effect of thermal cycling under vacuum on the
(a) 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.
(b) 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.
(Fluorotitanic acid chemistry from Chemo-Phos), and matrix, it was the coating itself and the matrix that suffered
Iridite NCP (thought to be fluorotitanate chemistry from degradation. Corrosion products rich in aluminium and
MacDermid) together with Alodine 1200S (chromic acid with oxygen were seen to have deposited around the intermetallic,
complex fluorides from Henkel) as the control coating. and their volume increased with time of exposure to the NaCl
Chemical cleaning was also applied by different methods prior corrosive solution. For the very small precipitates (<1 m
to the coating process in an attempt to achieve a clean, diameter) corrosion had penetrated underneath the precipi-
defect-free initial surface. The extensive report (Pereira et al. tate, so releasing it from the samples surface.
2008) concluded that only Alodine 1200S survived thermal The workhorse, high strength alloy Al-2219, in the T6
cycling and that the other coatings had a tendency to spall-off and T8 heat treated condition has been selected for Spacelab,
(producing contaminant particulates). The electrical resistivity the Columbus module of the International Space Station,
values of the other samples were very similar, even after Arianes 5, 6 and many other US and European space pro-
thermal cycling. Only Alodine 1200S survived the salt spray jects. Its corrosion properties, discussed within this Chapter,
testing on the high strength alloys. It was recommended that are of particular importance due to the airborne salt particles
space projects should continue with this well-proven product, that will deposit onto rockets and other space hardware
and to continue the search for an alternative. stored at sea-coast launch sites. Al-2219 has a matrix of
The four chemical conversion coatings are likely to be aluminium containing a Cu content of up to 6.8 %, Fe 0.3 %,
suitable for most non-critical applications when applied to Mn 0.4 %, and Ti 0.1 %. Our analyses have shown that the
the lower-strength alloys such as Al-1100, Al-3003, Al-5052 undissolved and precipitated second phase intermetallics
and alloys in the Al-6xxx series. These, and the high strength contain mainly AlCu (as CuAl2) with small fractions of Fe
Al-2xxx and Al-6xxx series of alloys will also benet from and Mn (Al-2219 contains a saturated solid solution of
all coatings when they are applied as a base for nal nish copper, designed for maximum precipitation hardening
painting. properties so there will be inevitable undissolved CuAl2
A third programme was designed to investigate the within the matrix microstructure). The cathodic nature of the
manner in which a conventional Cr6+ conversion coating intermetallics with respect to the aluminium matrix will
reacts with a bare aluminium alloy containing undissolved cause preferential oxidation of the less noble matrix as was
CuAl2 inclusions, such as Al-2219, (Grilli et al. 2010). By seen in Fig. 4.15. In the absence of an effective coating,
combining scanning auger microscopy (SAM) and EDX sever pitting corrosion is known to be prevalent at sea-coast
images it was possible to deduce the distribution of the launch site environments. The schematic diagrams in
conversion coating as a function of alloy microstructure. The Fig. 4.19 are thought to represent the corrosion mechanisms
conversion coating is thick and uniform all over the alloy most likely to occur. Progressive corrosion leads to exten-
matrix (and XPS indicated this was a mixture of Cr3+ and sive pitting, or crevice corrosion, and these defects could
Cr6+ species), cracking occurs during, or shortly after the becoming stress raisers. For alloys susceptible to stress
completion of the Alodining process to expose the alu- corrosion cracking this can lead to catastrophic fractures (see
minium substrate (Fig. 4.18b). On the intermetallic inclusion for instance Figs. 4.21 and 4.22).
surfaces the chromium concentration is much lower,
implying that the conversion coating is not able to protect
these more noble areas of the surface. 4.5.4 Cleaning, Passivation, and Plating
A fourth programme involved investigating samples of of Spacecraft Steels
Al-2219 having the titanium based Nabutan STI/310 coating.
SEM and SAM analyses conrmed the TiF based coating The preferred steels for spacecraft construction have been
was also generally uniform across the surface, but as with the tested and found to possess a high resistance to
Alodine samples, the CuAl2 inclusions only supported traces stress-corrosion cracking; these are listed in Table 2.3. This
of a titanium compound. Again, this indicated a reduced listing contains a complete range of steels from the
corrosion protection for the inclusions. Three precipitates low-strength plain carbon family to complex, high-alloy
were selected on the samples surfaceeach had a different duplex steels which have a very high strength if they are
size. The Ti- chemical conversion coated sample was then correctly precipitation-hardened. The plain carbon steels and
exposed to a 3.5 %NaCl solution for progressive exposures the carbon, low-alloy steels all have a low resistance to
of 15, 45 min, 2, 4, and 8 h. At the end of each exposure the corrosion. That is, they produce a loose oxide which breaks
sample was re-entered into the vacuum chamber and exam- off under very little stress, allowing further corrosion to take
ined by SEM, AES and EDX analyses. Out ndings (Grilli place. Such steels have few spacecraft applications but may
et al. 2010) revealed the fact that the coating had deposited be used as shear blades in pyrotechnic cutters, as springs (see
preferentially on the aluminium matrix, but, because of the Sect. 4.4.2) and as component housing. They must always be
cathodic nature of the intermetallics with respect to the plated to prevent the occurrence of corrosion. They may be
138 4 Spacecraft ManufacturingFailure Prevention
forbidden from some scientic experiments owing to their Martensitic high-strength alloys such as 440C are subject
strong magnetic properties. to hydrogen embrittlement and intergranular attack when
Steels that have poor resistance to corrosion (1000 series exposed to acids; they also have a low resistance to SCC.
and 4000 series) are vapour-degreased and pickled (possibly Quality control tests are usually required for the in-line
in hydrochloric acid of strength 15 %, or in a 75 % solution testing of spacecraft items.
of sulphuric acid in water) for a short time in order to remove For plated layers the thickness can be measured
light oxides. The plating is applied after all basis metal destructively on samples by microsectioning. Nondestructive
heat-treatments, machining, welding, brazing, and forming eddy current measurements can be made using equipment
have been performed. An underplating of copper having a calibrated to known standards. Plating adhesion is tested by
thickness of between 5 and 10 m is often initially applied. clamping and bending the sample and inspecting for ruptures
This is usually followed with an overplate of 1020 m of and delamination. To check for embrittlement, samples are
electrodeposited nickel. There is a hazard that hardened and held for 200 h at a load of about 115 % of the yield load, and
tempered cold-worked or cold-straightened parts might crack then examined for cracks or fracture.
during cleaning and plating. Such parts need to be stress- Passivated parts can be checked for continuity of oxide
relieved at 180200 C for three or more hours prior to lm by exposing samples to the copper sulphate test. The
plating. The same heat-treatment must be performed again test solution is made up of 40 g cupric sulphate, which is
after the plating operation (and within 4 h of the plating) to dissolved into a glass beaker containing 900 ml of distilled
provide for hydrogen-embrittlement relief. water to which has been added 100 g of sulphuric acid
Steels that are corrosion-resistant are usually not plated, (having a specic gravity of 1.84). The parts can be
but given a so-called passivation treatment. Passivation is a immersed into the solution or swabbed continually for 6
post-fabrication method of maximizing the inherent corro- 8 min. Visual inspection shall verify that the tested surfaces
sion resistance of stainless steels, but care should be taken to are free of copper deposition. The test solution must be
apply the correct method to the particular steel, as will be removed from useful parts by immersing them into cold
stated below. This treatment forms a protective oxide on the running water, then re-passivate in one of the aforemen-
steel surface in a controlled and repeatable way. Two solu- tioned solutions.
tions based on nitric acid are applied separately to two dif- Several specications exist covering the chemical passi-
ferent steel groupings: vation of stainless steel parts. The favoured ones are: ASTM
A380M, ASTM A967M and for electroplating such parts
Passivation treatment A, made up of 2.5 % (by weight) ASTM B254-92 (all dated 2014).
sodium dichromate and 25 % (by volume) nitric acid, at 48 Attempts to utilize greener passivation processes have
50 C for 20 min, for types been successfulthey are based on the use of citric acid
passivation; increasingly popular with fabricators wishing to
Precipitation hardening: A-286, 15-PH, PH15-7Mo, 17-7, avoid the use of mineral acids or sodium dichromate. Citric
AM-350, AM-355 acid is a bio-based material that releases no toxic fumes
Free machining: 303, 347, 430 during passivation, requires low concentrations and is easily
Ferritic 400: 405, 409, 429 disposable. Parameters for passivation with citric acid for a
Martensitic 400: poor SCC resistance 440C, 410, 414, 416, wide range of stainless and ferritic-FM steels is provided by
420, 431 DeBold and Kosa (2004). This nish has been fully studied
by NASA, including exposures at the KSC Beachside cor-
Passivation treatment B, made up of 2545% (by volume) rosion test facility (see Fig. 3.16) where samples showed no
nitric acid, at 2030 C for 30 min, for types signs of corrosion after 6 months (Lewis et al. 2013).
trade winds. The chloride concentration in precipitation in launch site, may result in erroneous conclusions. It is rec-
the Guiana region was measured at the turn of the twentieth ommended that where corrosion by the atmosphere is an
century to be equivalent to 193 lb/acre/year (216 kg/ha/year, important factor, on-the-spot tests are needed. A simple test,
or 60 mg/m2/day of deposited salt) (Miller 1906). The author such as the samples wire-on-bolt test (Doyle et al. 1963)
could not uncover more recent data! Similar precipitation can be conducted on the site, with test samples located at
can be expected at the KSC. Corrosion rates of carbon steel various heights above the ground. However, facilities such
at KSC are almost four times those further along the coast at as those depicted in Fig. 2.16 will provide far better data,
Daytona Beach and forty times those inland in Pittsburg PA and give reliable predictions as to a materials suitability for
(Calle 2014). the launch site environment.
Wind moving over breaking sea waves will pick up small The various launch sites (see Figs. 2.2 and 2.17) depict
droplets of saltwater. When these droplets and tiny particles several metallic structures such as the Mobile Gantries, and
or airborne salt accumulate on exposed metal surfaces, cor- the lightening masts that give protection the launch vehicle in
rosion can occur, particularly when the atmospheric relative case of tropical storms Pylons and masts are slender and have
humidity is high. Sub-micron-sized salt particles may deposit a certain amount of flexibility under high wind conditions.
within the cleanrooms of spacecraft integration halls, and The structures are usually constructed from mild steel and are
later, when atmospheric conditions of temperature and heavily covered with hot-dipped galvanization having often a
humidity permit, will absorb water and form a highly con- composition of 0.3 % lead, 1.0 % tin and some bismuth, rest
ductive solution. The science of breaking waves and the zinc. Metallurgical considerations have been given to such
release of salt particles has become a science in its own right structures, as their bases are often welded and can suffer from
and Sharkovs book (2007) contains astounding amounts of wind fatigue, resulting in fracture. The study of low cycle
information, mostly gained from Russian Cosmos satellites, fatigue of galvanized lightening masts has been performed at
and interpretation of data by the Russian Academy of Sci- the University of Genoa and calculation procedures are pre-
ence. Small droplets (aerosols) of salt water produced by surf sented for elimination such failure (Solari & Repetto 2014).
tend to be coarse (520 m) and those produced by whitecaps An associated problem with welded galvanized steel is the
are generally smaller (0.53 m). Due to gravity, the large phenomenon of liquid metal embrittlement (LME), and liquid
droplets will deposit close to the shore-line, whereas the small metal induced stress corrosion. These cracks occur when
aerosols can be transported considerable distances (Cole et al. liquid zinc penetrates between the steel grainsit reduces
2004). The Cole paper includes salinity maps of Australia. ductility and causes high localized stresses between the
Here, very few coastal regions reach the 60 mg/m2/day salt grains. Trained and certied welding operators will know of
deposition rates of the Guianese Space Centre. It may be these problems and, particularly if the structure is under stress
remembered that the proposed Cape York Spaceport project during welding, they may wish to try and unload the parts or
was to have transformed the northern tip of Queensland into a take measures to reduce the stress levels. Unfortunately these
high tech, international space centre. According to Coles cracks are not often visible from the surface, which is coated
map, the proposed location appears to have appreciably low with zinc. Dye penetrants or X-ray inspections are not always
airborne salinity levels (approximately 4 mg/m2/day). useful, but eddy current testing can detect LME cracks. New
Plans for the UK to become the home of Europes rst welding wires have been developed to avoid or reduce LME;
spaceport will be developed in 2016 when a detailed technical salt spray testing of such weldments appear to show
specication is due to be published. Interestingly, each pro- promising results (Kodama et al. 2010).
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 4.6 Stress-Corrosion Resistance of Metals
can differ between 2 and 8 as detailed in the Zinc Millennium
Map, published in the excellent design and metallurgical 4.6.1 Stress-Corrosion Cracking
publication: 2014 Engineers & Architects Guide.
Corrosion by electrolytic action can result when two The stress-corrosion cracking (SCC) of metals is initiated at
dissimilar metals are involved. Also, corrosion of a single several sites by localized corrosion, by pit formation, or by
metallic part can occur when, for instance, a saline solution intergranular attack. The influence of stress on the corrosion
reacts chemically with the parts surface. Impingement of rate is not very specic above a particular threshold stress,
salt laden aerosols can of course, provide a conductive and the other principal factors involved in a failure are
electrical path and short-circuit electrical equipment. environment, time, grain orientation, and heat-treatment
Acceptance of the laboratory corrosion tests described in the condition. The stresses that will produce failure by SCC in
previous sections as an exact representation of the corrosion service may be due to the residual stresses in the metal caused
occurring at a specic spacecraft assembly location, or by fabrication (quenching, forging, differential expansion),
140 4 Spacecraft ManufacturingFailure Prevention
by assembly stresses (resulting from improper tolerances classes of structural alloys. Surprisingly little is known about
during t-up, overtorquing, press ts), or a combination of such weldments as, in general, it is the parent metals that are
residual and operating stresses induced by service loads. thoroughly tested (Irving 1992). The Ariane 4 ELV has also
SCC at the launch site (see also Sect. 4.5.5) is a major suffered from SCC (see Sect. 5.19) and, even when alloys
contribution to failures in ground support equipment. At the are listed in Table 2.3, they can fail by SCC when exposed to
NASA Kennedy Spaceflight Centre in Florida, the problem non-saline environments. The auxiliary power units of the
has been highlighted by Majid (1988), who considered that Space Shuttle Columbiawhich supply power needed for
75 % of all SCC failures could have been prevented if those activation of the rudder, speed brake, and elevonsfailed
persons responsible for design had been sufciently aware of owing to stress-corrosion cracking of the Hastelloy B
the problem. In designing piece-parts the following must be injector tubes. Here the environment was exposure under
seriously considered: stress to ammonium hydroxide vapours, resulting in the
leakage of liquid hydrazine (Korb et al. 1985).
1. Tensile properties of metals. Probably the most incongruous failures are those which
2. The materials susceptibility to SCC. are life threatening; for instance, the selection of an AA
3. Fracture toughness. 2014-T6 forging for a naval helicopter main rotor blade
4. Type of corrosive environment. grip. This alloy is known to have a poor resistance to SCC,
yet it was specied on the manufacturers drawingthe part
Testing for stress-corrosion resistance may be undertaken failed by SCC in the short transverse plane (Parker 1991).
by plotting applied stress against specimen life in a particular Whereas the paper by Johnson (1973) lists many NASA
corrosive environment. The shapes of such curves will pre- failures due to the SCC of aluminium alloys, there exists an
sent data on stress-corrosion susceptibility. The sample equally interesting early report (Stanley 1968) that describes
conguration may have the form of uncracked tensile NASA launch vehicle failures due to the SCC of steel parts
test-pieces having a cylindrical or dumb-bell conguration, including 17-4 PH, H-11, and 440C.
or precracked samples that resemble fracture toughness The ideal material is one that suffers from no stress-
test-pieces. corrosion crack propagation below the maximum stress
In testing it is common practice to load specimens for levels that may occur during the service life of the compo-
30 days at a single stress level, that of 75 % of the mean nent. Numerous papers have been devoted to the topic of
yield stress, although there have been some instances where SCC and a good overview is given by Bussu (2009). There
a higher value, up to 90 % of the yield stress, has been appears to be a growing consensus that regardless of the
favoured. The SCC resistance of metals, particularly alloys specic mechanism, hydrogen is involved with the SCC
of aluminium, is always less when tension is applied in a propagation of both ferritic and austenitic stainless steel
direction transverse to the grain orientation. It is least for the (Hehemann 1985), and high-strength aluminium alloys
short transverse direction. The criteria used in the selection (Gruhl 1984; Newman 1990).
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 4.6.2 SCC Evaluation
load method is most favoured (i.e. specimens stretched in
spring-loaded devices) over other methods involving con- An important design goal is to achieve a highly efcient,
stant strain loading (i.e. specimens stretched or bent in jigs to low-weight spacecraft. This is possible only if materials are
a xed position). It is accepted that specimens having short selected from high-strength alloys such as those listed in
life will fail in a similar time when tested by either the Table 2.3 with a high resistance to SCC. In Europe, space-
constant stress or the constant strain methods. However, craft construction, materials, and their weldments are gener-
materials with a high resistance to SCC will give longer lives ally evaluated for SCC by the procedure detailed in ESA
under constant strain owing to relaxation of the stress by specication ECSS-Q-ST-70-36 Determination of the sus-
creep (plastic deformation). Structural material failures in the ceptibility of metals to stress corrosion cracking. The method
early USA space programmes were most commonly caused is particularly applied to all new structural metals and their
by SCC, resulting in much loss of hardware, time, and special heat treatments. It is used whenever metals have been
money (Johnson 1973). Of particular concern were complex welded or unusually processed and is applied as a quality
parts that had been machined from thick sections. Particu- control during production when in-line samples are tested.
larly vulnerable were the regions exposed to short transverse Many of the metals characterized as having a high resistance
loads, because SCC threshold stresses are lowest in this to SCC, as listed in Table 2.3, were evaluated as follows: The
transverse direction. Also, of critical importance is the sus- tests were carried out under alternate immersion conditions in
ceptibility of welds to SCC. This applies to all the main 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 The 0.2 % proof stress was established for each type of
same environment to provide a basis for comparison in sample from an average of three tensile tests. These values
assessing stress-corrosion susceptibility of alloys that survive would establish the stresses at which the stress-corrosion
30 days in the test. The susceptibility is assessed by tensile tests were to be conducted and they provide reference values
tests to compare the residual strengths of the specimens of tensile strength and elongation for comparison with
exposed stressed and unstressed, and by metallographic specimens to be tensile tested after surviving 30 days in the
examination of microsections from stressed and unstressed stress-corrosion test.
specimens to distinguished between stress-corrosion and The stress-corrosion tests were carried out in triplicate,
intergranular corrosion or pitting occurring independently of the specimens being loaded in tension by calibrated springs.
stress. A typical test rig is shown in Fig. 3.15c. 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 the SCC ratings have been determined (i.e. highly resistant
corresponding unstressed sample. After removal from the moderate, or low resistance to SCC) it may be necessary to
salt environment, and at the end of the thirty-day period, all perform fracture mechanics testing. The additional work is a
specimens were washed in warm water, with gentle scrub- requirement for alloys having a moderate or low rating which
bing to remove most of the salt and corrosion products on will be used in any structural application. The ESA test is
them. They were dried in warm air and stored in a desic- similar to a screening test for alloys and weldments, and
cator. The samples were then tensile tested and subjected to indicate that below 75 % of the materials 0.2 % proof stress
metallographic examination in accordance with the there will be no onset of SCC. The fracture mechanics tests
ECSS-Q-ST-70-37 test procedure. involve pre-cracked specimens (see Sect. 3.2.2.8), and it is
The ESA constant-load test overcomes the repro- necessary to obtain a threshold value, at a crack tip, when the
ducibility problems associated with constant-strain tests, stress-intensity factor (known as KISCC) will not cause the
which suffer owing to relaxation of the stressing jig; or after crack to propagate while the sample is being exposed to the
the onset of SCC in small specimens, there is plastic strain liquid test environment. Note that the usual test liquid is a
and creep with an attendant reduction in the initial load. Once 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 4 Spacecraft ManufacturingFailure Prevention
Fig. 4.21 BS L93 alloy plate, stressed to 25 % of PS showing a pit at surface and b detail of the crack tip
4.6 Stress-Corrosion Resistance of Metals 145
Fig. 4.22 a AlLi alloy 8090-T81 3 mm sheet, stressed in the long stresses are introduced into a mechanical connectionexcessive
transverse direction at 75 % PS. Severe SCC causes specimen failure clearance between lugs result in normal stresses parallel to the
after only 3 days alternate immersion. Etched microsection (50). b Al short-transverse (ST) direction of the materialsuch a design must
Li alloy 8090-T8771 plate, stressed in the short transverse direction at not incorporate AlLi 8090-T8771 thick plate, or any other Table III
50 % PS. Failure, and severe SCC is noted after only 9 days alternate alloy per ECSS Q-ST-70-36 (the standard listing alloys as having a
immersion. This alloy is categorized as being Class III per ECSS high, medium and low resistance to SCC), (Bussu and Dunn 2002),
Q-70-37 (the SCC test method standard). Picture shows the as-polished courtesy of ESA
microsection (200). c illustrates a typical example of how sustained
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- seldom, if ever, used for springs. These metals have been
erties, and the temperature limits of these materials (the listed in Table 4.5, and are not considered further.
SRAMA as originator is gratefully thanked for permission to A full evaluation of all other alloys for spring manufacture
publish this information). is given in Tables 4.6, 4.7, 4.8 and 4.9. Where appropriate the
Some of the metals in Table 2.3 are not capable of being relevant UK, USA, and/or DIN specication that applies to
hardened sufciently for springmaking, and others are 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 4 Spacecraft ManufacturingFailure Prevention
together with the metallurgical condition in which the alloy is cleaning after part manufacture (see Sect. 4.21)
used. Many of the alloys have mechanical properties that passivation and control that surface is completely passi-
depend on section size, and in these cases it will be necessary vated (see Sect. 4.5.4)
to refer to the specication for these data. The spring oper- application of oils or greases (see Table 5.2)
ating temperature limits indicated are those known to be safe. Storage of bearings, mechanisms and similar corrosion
The maximum temperature is generally governed by spring sensitive parts in sealed, leak-tight bags containing dry
relaxation properties. Where information is known about an nitrogen.
alloys usage below 0 C, this has been given.
These preventative measures should avoid the damage to
bearings or similar mechanisms. They are based on the lessons
4.6.4 Bearing Materials learned from some diligent failure analysis work (Saltzman
et al. 2009) performed on a solar array drive actuator where the
Bearings for spacecraft mechanisms consist of balls, which root cause was concluded to be: stress corrosion due to
are the rolling elements, and the rings which form the incomplete cleaning, the presence of small amounts of chlorine,
raceways. Where very high radial and thrust loads are sodium, water vapour and very long storage times.
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 4.7 Control of Printed Circuit Boards
bearings, these are based either on high alloy steels, or
occasionally, ceramics are used for special applications (for 4.7.1 Chemical Composition of Tin-Lead
instance the Shuttle and Ariane cryogenic bearings for from Microstructure
pumping liquid hydrogen have silicon nitride rolling ele-
ments). Commercial bearings are often made from AISI It is generally accepted in the space business that
52100, a low alloy steel containing 1.4 % chromium (see printed-circuit boards (PCBs) to be assembled by soldering
Appendix 6 for some European equivalent designations). support eutectic composition tinlead nishes which have
Aerospace steels contain prescribed amounts of been electro-plated and then fused onto their high-purity
high-chromium, carbon and molybdenum (1417 %Cr, 0.6 electrolytic copper conductors.
1.1 %C and 0.75 %Mo). A common corrosion resistant At least eight other pcb nishes are available but it seems
steel is designated 440C; this has been selected for most unlikely that these will nd extensive use for space boards.
space bearings because of excellent performance and a Briefly, these are:
successful heritage. The combination of chromium and
carbon ensure that heat treatment, mechanical deformation HASLhot air solder level, poor shelf live as solderability
(forging, etc.) and induction hardening, provide a mixture of degrades with time (Fig. 4.37).
ne M7C3 and M23C6 carbides (see also Fig. 5.7) in the OSPorganic solderable preservative, not possible to use
steels tempered martensitic microstructure. This composi- test probes without damaging coating and coatings probably
tion provides the surface hardness required of the bearing, not stable under vacuum
usually between 59 and 61 HRC. The chromium gives the ENIGelectroless nickel immersion gold, tarnishes and can
steel some corrosion resistance, at least 12 %Cr is required be susceptible to black pad.
at the surface to form a thin, insulating chromium oxide Immersion tingood for ne pitch packages but may suffer
surface lm. Due to the formation of chromium carbides the from whisker growth.
initially high-chromium available at the surface becomes Nickel palladium with immersion silver or goldtarnishes,
depleted, but the presence of molybdenum avoids the critical or gold tin intermetallic problem.
depletion to below 12 %Cr. Immersion silvertarnishes and may suffer from electro-
Due to the critical surface condition care must be taken chemical migration.
during the handling of products made from 440C. This is not Reflowed tin-leadpresently preferred, good solderability,
an austenitic steel, such as stainless AISI 304 or 316, (nothing solders better than solder), solder slumping may
possessing thick, very adherent chromium oxide nishes. occur but, as will be shown below, many process precautions
The thin chromium oxide lm on 440C can easily become and quality controls are needed.
degraded and render the surface at risk to rusting and even
stress corrosion. In fact 440C is known to be a Table III From experience to date, hot oil fusing of tinlead n-
alloy, susceptible to stress corrosion cracking per ECSS and ishes by means of a two-stage heating cycle (see also
NASA standards. The strict post-manufacture controls Fig. 4.24) has proved to be a controllable process (Dunn
should include: 1980). One supplier of hi-rel PCBs has, however,
4.7 Control of Printed Circuit Boards 149
encountered systematic problems resulting in rough, fused The surface cavities and spherical pores associated with
tin-lead nishes containing a large degree of porosity the rejected boards are considered to result from the
(Fig. 4.23a). A metallurgical investigation was instigated to volatilization of occluded organic materials during the hot
determine the severity of the problem and to recommend a oil fusion process. At some time during solder fusion, before
procedure which would eliminate the observed defects. the alloy had cooled to its solidus temperature, the presence
Microsections (Fig. 4.23b, c) were analysed by EDAX and of tin-rich dendrites has entrapped gas pockets which would
line-scan attachments to a SEM. Wet chemical analysis otherwise have been released at the liquid solder-to-oil
conrmed the presence of slight copper contamination and a interface. A revised plating procedure eliminated all trace
hypereutectic composition of the tin-lead, tin being 69.4 % contaminants and, once fused, provided a microstructure of
whereas the eutectic composition of pure solder is 62.9 %. almost exactly eutectic composition (Fig. 4.23d). No surface
The detailed photomicrographs highlight several large, imperfections were observed on PCBs made to this revised
partially sealed pores in the cross-sectioned coating; the procedure.
normal coppertin intermetallic layer is just visible (A and B It was found to be essential that tinlead fusing trials
of Fig. 4.23b, respectively). The microstructure of the should be performed to establish the optimum fusion
plated-through hole shown in Fig. 4.21c also details severe schedule for both double-sided and multilayer boards in
porosity (A) and a ne laminated eutectic structure (C). This order that they retain a good solderability after long term
phase distribution would be expected in a tinlead alloy storage. Research has shown that the metallographic
containing 69 % tin, the relative volumes of islands and inspection of test coupons from boards can adequately
eutectic being given by the tinlead binary phase diagram. 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 151
Oil fusion was followed by cleaning in isopropyl alcohol Results and discussion
and bake drying for 2 h. The ve submitted boards were noted to possess totally
Five boards were submitted for this investigation. They different topographic features, as can be seen from
had been fused according to the aforementioned procedures Figs. 4.25, 4.26, 4.27, 4.28, 4.29, 4.30, 4.31, 4.32, 4.33 and
and, in addition, three were given a further fusion treatment 4.34. The captions describe features ranging from, lumpy
as a rework operation (210 C until thought re-melted by solder islands with high contact angles and bright flat
the operator) in an attempt to improve their surface visual dewetted regions, to, matt, dark, gritty nishes with large
appearances. All the boards were subsequently rejected at dull islands. Detailed photographs highlight the pads around
incoming inspection by a company engaged in the compo- plated-through holes; they were taken under identical par-
nent assembly of spacecraft circuit boards, as detailed in allel lighting conditions which accentuated the observed
Table 4.10. The reasons for rejection were that each board features. These pad areas represent the most critical board
had an irregular surface appearance. Only two boards were surfaces; they are required to have excellent solderability to
later dispositioned as rejectable by the PCB manufacturer. facilitate component assembly with both nonactivated fluxes
and low soldering iron temperatures.
Method of evaluation Only the illustrations associated with Board no. 1 will be
The submitted boards were evaluated as follows: 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
Visual inspection and photography with oblique lighting. technique and observations. Table 4.11 summarizes every
Samples removed for microsectioning. result of the various thickness measurements and solder-
Samples removed for rotary dip solderability testing of ability tests. All boards, with the exception of Board no. 1,
plated-through holes (ECSS-Q-ST-70-10 requires wet- had adequate tinlead coverage around the corners of each
ting in 3 s and no dewetting after 8 s). plated-through hole.
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.
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. 4.30 Board no. 1 (A), hole corner is covered with 1.5 m
intermetallic and has dewetted
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
conditions the operator was in a position to judge and prevented from flowing around the plated-through hole
compare the ease of making each joint. Table 4.11 compiles corners, and in addition there are signs of severe dewetting
the results of these investigations, and it can be seen that all on the bottom (B) pad surfaces. Severe solderability prob-
board which had been reworked by a second fusing opera- lems will exist when exposed coppertin intermetallics are
tion had potential soldering problems. present at the knees of plated-through holes (Hagge and
Photographs and microsections of the hand-soldered Davis 1985). The formation of coppertin intermetallic does
joints made to Board no. 1 shown in Figs. 4.32, 4.33 and not adversely affect the soldering as long as there is a coating
4.34. As with the solderability test, it is noted that solder is of tinlead to protect the intermetallic from oxidation. At
4.7 Control of Printed Circuit Boards 155
least 1.5 m of coating was recently determined to be critical Destructive evaluations can be made on the test coupons
to ensure solderability (Ray et al. 1995). However, the attached to delivered PCBs; they must include solder-
0.7 m of intermetallic that forms during the hot oil fusion ability tests, pad adhesion tests, and microsectioning.
will continue to grow at a rate of about 1.0 m per year In the non-heat-treated condition, as-plated tinlead is
dependent on storage temperature and humidityso it can present as a duplex structure consisting of almost pure
be seen that all tinlead coatings will have a nite useful life lead grains which are dispersed within a tin-rich matrix
(Craven 1993). It should be noted that in both the rotary dip (Dunn 1977). Each zone will have a theoretical melting
test and the soldering trials the solder was applied from the point of 327 C for lead and 232 C for tin. It is therefore
top side (A) of each board. Joint inspection was made to desirable for the as-plated duplex grain size to be small in
accept/reject criteria of spacecraft soldering specications, order to facilitate solid-state diffusion and interalloying.
and no attempt was made to rework the defective joints. Under these conditions eutectic liquidation will be
Although the hand-soldered joints depicted in Fig. 4.32 do achieved during a shorter oil fusing time (for any given
not meet visual inspection requirements, it can be argued oil bath temperature).
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 4.7.3 Evaluation of Multilayer Board Internal
importance is made of the cosmetic appearance of solder Connections
llets and wetting angles.
Plated-through holes provide the electrically conductive paths
Conclusions from one side of a double-sided PCB to the other side. In the
case of multilayer PCBs the plated-through hole may inter-
The reworking of a high-rel spacecraft PCB by a connect more than 20 layers of conductor patterns. The ther-
second fusing operation is considered unacceptable. The mal expansion coefcient for solder and copper in a multilayer
additional reflow treatment reduces the solderability of board assembly is approximately 1517 m/mm/C. In con-
each through-plated hole and increases the band of trast, the epoxyglass board substrate expands at approxi-
coppertin intermetallic from a thickness of 0.7 m mately 100340 m/mm/C in its Z-axis and is nonlinear in its
(maximum) to over 1.5 m. These thicker layers of temperature increase between room temperature and the sol-
intermetallic produce dewetting of surfaces to be sol- der melting temperature of 183 C (see also Table 6.1). From
dered (e.g. at the corners of plated-through holes and on this it can be seen that during solder fusing, soldering opera-
termination pads). Unacceptable solder joint quality can tions, and thermal cycling under service conditions the
result if an attempt is made to mount components to epoxy-glass expands more than the copper plating within the
such boards. hole. This will cause mechanical stressing at the point where
The unusual surface features of as-fused PCBs are the internal pads connect with the copper plated-through hole.
undesirable and will cause problems during incoming Work hardening and subsequent fatigue cracking can result at
inspection (Bulwith 1986). No soldering problems were these junctions, as can be seen by the examples illustrated in
encountered on non-reworked tinlead nishes, but by Fig. 4.35ad. These illustrations show how inherently defec-
visual inspection alone it is impossible to establish if a tive multilayer boards (from nonapproved sources) can pass
surface feature is a cosmetic or an actual defect. electrical continuity tests on receipt at a spacecraft contractors
156 4 Spacecraft ManufacturingFailure Prevention
4.7 Control of Printed Circuit Boards 157
b Fig. 4.35 a Photomicrographs of cracked pth internal connection. temperature or thermal fatigue will result in these cracks growing and
Cross-section made across top part of plated-through hole. Note crack causing electrical open circuits. c Usage of the multilayer board shown
(between rst internal copper conductor and the plated-through copper) (as-received) in (b) has resulted in open circuits. Thermal shock created
extends completely around hole. Detail of crack at the interface during the soldering operation has caused wide cracks at the internal
between internal copper conductor and plated-through copper. Electri- conductor to plated-through hole interfaces (courtesy of ZVE).
cal contact is made at only one small point (arrowed). Additional d Beginning of a corner crack in the electroplating of a
cracking across internal conductor at a lower circuit. b Photomicro- plated-through hole caused by resin smear on the edge of the copper
graphs to illustrate manufacturing defects on multilayer board (8 layers) cladding (courtesy of IFE)
supplied to experimenter from non-approved source. Exposure to high
site. It is important that destructive microsections are made increasing the copper plating thickness; and by locating the
both by the circuit board manufacturer as part of his nal functional pads at the centre of the board thickness (i.e.
inspection and by the customers incoming inspection internal connections at the neutral axis). By optimizing these
department; this is done on the test coupons attached to flight design parameters, impressive improvements have been
quality boards. It is now recognized that this microsectioning demonstrated by Yoder et al. (1993), with no failures when
task must be performed excellently in order to locate the plated-through holes were subjected to 2000 thermal cycles
presence of the very ne microcracks present in Fig. 4.35b. (55 to +100 C) over a sixteen-month period as part of a
Often, grinding and polishing will cover over the microcracks. NASA space environment qualication programme.
Alternatively, when the copper is etched, an uninformed eye Specications for the procurement of multilayer boards
may mistake the dark line between the copper laminate and the rely heavily on metallographic examination of
plated copper as a crack when it could be a series of grain plated-through holes from the test coupons representative of
boundaries. The thermal shock caused during the component the production boards in order to ensure quality and relia-
lead soldering operation has immediately opened the bility. These examinations are made after coupons have been
defective interfaces shown in Fig. 4.35c. Alternatively, ther- exposed to thermal excursions that simulate conditions that
mal cycling at unit level testing can thermally fatigue these the board might be exposed to during initial assembly
connections. Worse still would be the generation of open operations and repair or reworking operations. Separation of
circuits caused by in-orbit thermal fatigue due to component internal connections and the presence of resin smear are
heat dissipation or solar heating. examined for in the as polished condition. The microsec-
Other defects may be shown up during thermal cycling, tion is then lightly etched to detect cracks in the copper
such as the delamination and resin recession referred to in plating or foil, etching conditions, lamination, drilling
Fig. 3.19b. Delamination can be prevented by black-oxidizing parameters, hole cleaning, and etchback processes.
the copper foils before laminating the multilayers in order to The EDAX attachment to the SEM and Auger Electron
increase the available surface area for bonding. Processes Spectroscopy can be used to detect contaminants, especially
such as etching the copper to a minimum depth of 1.25 m, in cases when internal layer separations are observed.
control of the etchant composition, and strict control of the Figure 4.36 shows how microsectioning was used to
black copper oxide thickness can be assessed by standard peel ensure the integrity of multiple component replacement
strength and solder float test methods followed by metallog- activities and the effects of these operations on the quality of
raphy (Jackson 1987). It is essential to avoid delamination if the internal connections within the multilayer board to which
the thermal fatigue of internal connections is to be avoided. they had been assembled. Several components including this
The plated-through copper must also have a very high purity capacitor were assembled and desoldered three times, using
and a suitable grain structure to enable a certain amount of the flight board test coupon. Checks were also made to
plastic deformation during temperature excursions. Trace ensure that the capacitors internal solder (low melting
elements or impurities in the electrodeposit interfere with silver-loaded tinlead alloy) had not melted to produce either
dislocation motion, and very ne grain structures prevent open circuits due to lead separation, or short circuits due to
extensive dislocation movement during plastic deformation, the presence of reflowed solder particles between the elec-
causing the plating to be hard. If dislocations cannot move, the trodes. Figure 4.36b, c conrms the absence of cracked
applied loads caused by thermal expansion mismatches cause platings and a good internal layer connection following the
the plated hole to fracture, resulting in intermittent open cir- thermal stressing. The reliability of inner layer junctions can
cuits. Suitable copper electrodeposits have a low hardness be assessed by submitting samples to thermal shock by
(less than 100 kp/mm2) and an elongation at fracture of greater floating them on the surface of molten solder. This test is
than 10 %. recommended if solder joint reworking or extensive com-
The fatigue life of multilayer boards, when designed to ponent replacement is envisaged. The rework of assembled
withstand thermal cycling, can be markedly improved by: multilayer boards can be a problem, and should be attempted
increasing the diameter of the plated-through holes; by 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 159
reworking activities and their effect on PCB reliability will will be required to assess the effect of vibration fatigue on
be presented in Sect. 6.12. the soldered connections.
A note must be made to readdress epoxy smearing. This Copper is the most common conductor material. The
is the drill smear of epoxy resin which has been transferred preferred type is rolled annealed (RA) copper, as this has the
to the edges of copper foil in the holes of multilayer boards. greatest flexibility and a high resistance to mechanical fati-
It has been seen both to be a uniform coating and to be gue cracking than the cheaper, electrodeposited copper.
present in scattered patches. It is supposed to be totally When using RA copper it is important that electrical
removed in the etchback process of mechanically and/or designers are aware that the copper grain direction should
chemically cleaning the hole prior to through-hole plating. run parallel to the length of the flexing section of the circuit.
The epoxy smear is undesirable because it can electrically The thinner coppers are more flexible. They can be pur-
isolate the conductive internal layers from the plated-through chased with the same nomenclature as is used for the
hole. The epoxy smear can be caused by improper drill thickness of copper on rigid PCB substrates. This is stated in
feed-to-speed ratio, blunt drills, incomplete cure of the B ounces per square foot but can be related to copper thick-
stage epoxy or smear induced by overfeeding the drill into ness; for example, 1 oz/ft2 equals a thickness of 35 m.
the backing plastic sheet. Almost any hole drilled in a Other electrical conductors can be silver, nichrome, or
multilayer board may contain epoxy smears; it is therefore a berylliumcopper alloy.
crucial task of the processor to control for each possible The most common spacecraft dielectric selected for
cause. Smears are impossible to detect without resorting to flexible circuits is polyimide (e.g. Kapton), as this has
microsectioning (see, for instance, Fig. 4.35d) and even then excellent mechanical, electrical, and chemical properties.
it can only be detected if the mount is very carefully pol- A usual thickness is 50 m. Fluorinated ethylene propylene
ished. It is impossible to predict the number and location of (FEP) is selected for spacecraft microwave applications; this
smears, and they may miss detection during the in-line is a thermoplastic with an upper temperature limit of 270 C.
quality control steps associated with the test coupon. However, it is extremely expensive and difcult to process.
An illustrated handbook containing the rationale and Polyesters have high outgassing properties, and whilst they
procedure for the evaluation of multilayer PCBs has been are the most cost-effective commercial materials, they are
written by Jellison (1986). It places great emphasis on the unsuitable for space use.
problems addressed in this section and can be used as a Flexible circuits are often used as the umbilical electrical
practical aid for those in the space industry. The handbook, connection to mechanical devices that operate under elec-
prepared from NASA experience, is certainly still appli- trical power, particularly in the close proximity to hinged
cable today, as it contains photomicrographs of subassemblies. They form part of heating devicesfor
cross-sectioned holes to illustrate defect types, acceptable instance, it is important that the titanium pressurized tanks
and unacceptable conditions, and methods of measurement which contain hydrazine fuel are not permitted to fall in
that are also similar to European PCB requirements temperature when the spacecraft passes into eclipsehere
(ECSS-Q-ST-70-10). nichrome elements are bonded inside the flexible circuit
Finally, attention is made to a collation of ber and par- which, in turn, can be adhesively bonded to the surface of
ticle contamination defects seen during the inspection and the titanium tank. Where rigid multilayer boards (Sect. 4.7.3)
testing of spacecraft pcbs (Heltzel 2014). His presentation are used in conjunction with flexible laminates, the con-
details failure mechanisms caused by electrochemistry, dis- struction is called a flexi/rigid circuit. In this case the
continuities in the laminate and discusses the sources of the flexible polyimide circuit passes into, and is sandwiched
contamination and relates the problem to pcb design. The between, the layers of the multilayer board. The flexi/rigid
included viewgraphs would make excellent Workmanship circuits are used to provide a total package combining a
Standards for pcb inspection and quality tasks. 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
4.7.4 Flexible Circuits may even be in a different plane to that of the mother board.
The flexible circuit should have an oxidation protection
Electrical power lines and data transfer conductors are often layer such as gold-plating on the copper conductor lines.
constructed with flexible circuitry. The range of construc- Tinlead should be present on termination areas where
tions for these spacecraft flexible circuits is as wide as the joining by soldering is envisaged. Fused tinlead is not
variations in materials selected for rigid PCBs. The main recommended, as the hot oil fusion process can cause the
difference is that electronic components are very seldom flexible circuit to delaminate. When the flexible circuit is
assembled onto flexible circuits and, in the rare event that located on the outside of a spacecraft designed to operate in
this is a special design feature, special qualication testing low Earth orbit, it is important to protect the polyimide
160 4 Spacecraft ManufacturingFailure Prevention
4.7.6 Solder Assembly of Component commercial industry. This is particularly true of metalma-
Packages onto Multilayer Boards trix composites which seem to be just emerging from the
with High Heat Capacity research and development stage.
The word composite in composite materials means two
A common problem encountered when hand-soldering or more materials, called constituents, that are combined in
components onto thick, up to 24-layer boards, having a certain conguration to form a new material with specic
back- or ground-planes is that the solder does not flow from desired properties. Usually, the strengthening phases are
the soldering side of the board to the top surface. It will not whiskers or discontinuous bres embedded in a matrix. In
be possible to only use a higher soldering temperature the utilization of composite materials a number of conditions
(contravening the ECSS Q-ST-70-08 maximum of 360 C have to be met. These conditions are related to the properties
for difcult boards) as this can easily cause delamination, of both bre and matrix. Primarily, the bres are expected to
fracture of internal plated hole connections, pad lifting and bear the tensile load without yielding. At the same time the
may exceed the recommended maximum soldering temper- matrix should distribute the load evenly over the bres,
ature listed on a data-sheet. Wave soldering for these board protect their surfaces from oxidation, and diminish any
designs was successful on one project. It should be noted possible crack propagation.
that recent soldering standards have relaxed the visual Although such research has been done on compounds
inspection requirements for such boards75 % hole ll with such as SiO2, Al2O3, and SiC, the main interest to day seems
25 % circumferential llets may now be permitted. to be focused on graphite and boron bre with strengths
Soaking the board in an oven at 125 C and then sol- approaching their theoretical values of 210 GN m2.
dering is possible, but dangerous for the operator. A study at If we turn our attention to the matrix material it will be
one of the ESA training schools (Strachan 2000) attempted clear that although it is not required to bear the load, the
three pre-heating methods: IR lamp; Hot plate with hot air matrix must possess the following minimum properties:
stream; and, the application of a hot air gun.
Results indicated the IR lamp (Martin HB02 Hot Beam) a good bondability to the bre,
gave the best resultsheating being on the component side. a compatible coefcient of expansion to avoid thermal
This allowed the soldering side to reach 100 and 130 C on fatigue problem,
the component side. No damage ensued, and the process was a protective layer against corrosion, and
comfortable and convenient. However, some component suitable elasticity and toughness.
bodies could shadow the joint being soldered.
The hot plate method (Pace HeatWave) gave good results The bondability is necessary to transmit and redistribute
with solder side at 100 C and the component side at 140 C. shear stresses, as a result of loading, to all active bres. It is
Some poor visibility of the soldering area was noticed. also important that there is minimal reaction between matrix
The hot air system (Leister 7A1) gave a very fast heat up and bres, thus avoiding the growth of brittle intermediate
rate which could lead to destructive temperatures being phases, cracks, or internal residual stresses. Unfortunately, it
reached. However, this system was best for rework and is found that structurally perfect, strong interfaces are dif-
visability. cult to form.
The practical problems associated with incorporating
reinforcing whiskers into a metal matrix are shown in
4.8 Control of Composite Materials Fig. 4.38a. These aspects were addressed by Calow (1974)
with particular reference to the composites Ni/Al2O3 for use
4.8.1 MetalMatrix Composites for Space at elevated temperatures and Al/C, Al/SiC, and Cu/C for use
Structures at ambient temperatures. It would seem that the predictions
from this early work are still relevant today in that the future
The eld of composite materials is relatively new, with the for structural composite materials will rest entirely on eco-
greatest expansion in growth occurring during the last dec- nomic factors, dependent on lowering the cost of producing
ade. The impetus in growth has been to achieve high bre materials and also for thermally cycling environments,
strength-to-weight ratios in materials. An ever increasing achieving a more permanent interfacial bond between bre
number of papers are being published on this subject, and it and matrix.
may be quite frustrating for potential users of such materials Much work has been made into investigations for
that, despite the vast amount of work that has been com- replacing the very heavy high-temperature nickel-based
pleted in research and development, there is still a striking superalloys by SiC bre reinforced titanium alloy Ti6A14V.
lack of technically and economically available materials for 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 163
material by condensation from the gas phase (PVD). The properties can be achieved depending on the specic lay-up
coated bres are then bundled, encapsulated, evacuated, and design. For unidirectional plates the room temperature ten-
hot-isostatically pressed at a temperature of about 900 C sile strength can reach 1000 MPa with a tensile modulus of
and a pressure of 1900 bar. The material composite is 300 GPa. For thick plate with 0/45/90 angle plies the
formed by the Ti-matrix flowing into the initial cavities strength is 240 MPa and modulus 120 GPa. These carbon
between the coated bres. The porosity-free composite is magnesium MMCs have a coefcient of expansion ranging
conrmed by metallography as shown in Fig. 4.38b. The from 0 to 4 ppm, dependent on the bre lay-up direction.
shear strength between the bres and the matrix of this Such MMCs are suitable for applications where a low
MMC was determined by Hemptenmacher et al. (1994) us- expansion is essential (e.g. optical bench) but they will
ing a method developed by Grande et al. (1988). Here, require suitable corrosion-resistant surfaces (see Sect. 4.5.1).
individual bres were pushed out of a flat parallel test-piece Large space structure prototypes have been constructed in
using a hardness test indenter. The shear strength is deter- the USA incorporating graphitemagnesium struts and t-
mined by the force required for bre debonding, divided by tings (Paprock 1988); full metal inltration was demon-
the surface area of the cylindrical bre under test. A bre strated by metallography.
with a protective carbon coating has a far lower bond Silicon carbide particles are also used to reinforce mag-
strength to the matrix than a bre with no coating, as seen in nesium and magnesium alloy composites in order to achieve
Fig. 4.38c, d. However, as noted by Hemptenmachter et al., low density and high stiffness-to-weight ratio products for
the shear strength depends also on the temperatures to which applications such as aerospace equipment. These MMC
the composite has been exposed. A special review of the composites may be fabricated by both powder metallurgy
interfacial phenomena in MMCs, how they can be examined, and ingot metallurgy routes (Krishnadev 1993), but it is
and how they affect performance, has been prepared by Feest found that the former route results in material having the best
(1994). Ti6A14V/SiC MMC with 32 % volume fraction SiC properties for the Mg/10 vol.% SiC composites.
bre has a tensile strength of 1450 MPa parallel to the bres Low-expansion materials can also be produced from
and 450 MPa transverse to the bres. aluminium MMCs. A special gas pressure inltration pro-
One of the most promising metalmatrix composites is cess has been developed by AMAGFVA in Austria for the
the newly developed graphite bre/magnesium composite, production of continuous-bre reinforced mouldings for
which offers great potential in terms of high strength and high-strength and thermal stability. These aluminium MMCs
stiffness combined with light weight (Goddard 1984). A new are comparable to those described by Cornec and Clariou
graphite coating system that allows molten magnesium to (1994) and mentioned above for magnesiumMMC. How-
wet the bres has been developed. This means that it is ever, AlMMCs have a greatly improved surface-corrosion
possible to lay up the coated graphite bres in a mould of resistance. A wide variety of bres and aluminium compo-
any desired conguration, and then cast molten magnesium sitions are produced by AMAGFVA. A typical plate pro-
onto them. The result is a metal composite cast to nearly the duct is composed of 60 % high-modulus carbon bres in a
net shape which can be subsequently machined by conven- matrix of 40 % high-purity aluminium (Keramol
tional tooling practice. In addition, as the composite follows 1085/C-HM/60f). This plate material has a tensile strength at
the rule-of-mixtures behaviour, each casting can be tailored room temperature of 360 MPa, with a tensile modulus of
for a particular application. This is demonstrated by con- 95 GPa. The tensile strength and modulus can be increased
sidering an article that requires a good longitudinal strength by designing products having a thinner cross-section (the
and a reasonable transverse strengthin this case the casting so-called bre paradox). This MMC has a coefcient of
could be made up of 30 % bres (by volume) oriented thermal expansion of 0.4 ppm and a density of 2.2 g/cm3.
longitudinally and 10 % oriented transversely. Of particular The revolutionary concept of metal-matrix composites,
interest to spacecraft designers is the fact that this material combined with their complex requirements for structural
possesses zero, or near-zero, thermal expansion over a wide applications (particularly for aircraft and spacecraft), pre-
range of temperatures. Special bres giving more options in sents a formidable uncertainty impeding acceptance by
strength, stiffness, light weight, and endurance against heat design engineers. Some problems are highlighted in
have been developed (Klein 1988). They range in diameter Figs. 4.38 and 4.39. Because of this, designers of today will
from 150 m to 2 nm. prefer to choose structural and mechanical parts from tra-
An alternative method for producing carbonmagnesium ditional metal alloys, and will limit manufacturing to joining
MMCs for lightweight space structures has been developed and nishing technologies which already exist in their
at Aerospatiale (Cornec and Clariou 1994), which is based respective plants. Whenever MMCs are selected it is
on an organic prepreg route (solid densication) or on a advisable to discuss the material requirements face to face
liquid phase inltration route. A wide range of mechanical with the producer.
164 4 Spacecraft ManufacturingFailure Prevention
Fig. 4.39 Micrographs depicting Avco SiC monolaments within an produce a compact MMC plate (by courtesy of the Metals Technology
AA 2014 matrix. This metal matrix composite (MMC) has been centre, Harwell Laboratory, UK). a General optical micrograph
recently identied as a potential advanced spacecraft structural showing the etched AA 2014 matrix structure of the MMC plate.
material; it is produced by multistage processing. A carbon core, made The bres possess a graphite core surrounded by silicon carbide and a
from a polymer precursor, is coated with SiC by chemical vapour thin layer of carbon. b TEM view showing internal structure of SiC at
deposition initially as a carbon-rich deposit and then as a columnar the transition between the inner carbon-rich and the outer stoichiomet-
deposit possessing grains of preferred orientation and a stoichiometric ric SiC regions (at point arrowed on the optical micrograph). c SEM
composition. The SiC monolaments are coated with a carbon coating: view of a tensile fracture face showing lament pull-out (minus the
the laments are then wound onto AA 2014 foil. Additional aluminium carbon layer), and a predominantly ductile matrix failure
is added to the foil by plasma spraying. Hot-pressed layers of these foils
4.8.2 Composite Contact Devices between electrical contacts. Typically, metal will transfer
from the anode to the cathode in a d.c. circuit relay after
The insertion of bres of a high-melting point metal into a several thousands of switching operations. This occurrence
matrix of a highly conductive metal for the purpose of becomes serious when microscopic particles of detached
producing contact materials was rst described by J.K. Ely deposits break loose within spacecraft devices. They can
in a patent of 1942. Fibres were of tungsten or molybdenum cause short-circuits between adjacent conductor paths.
in a matrix of a coppersilver alloy. The bres imparted Typical particles from worn contacts are shown in Fig. 4.40.
mechanical strength, reduced mechanical wear, and pre- Alternative metalmatrix composite materials are being
vented particles of the contact from crumbling away. Similar considered for advanced contacts. The silvernickel com-
contact materials are still being employed even in spacecraft posite and the copperpalladium system appear to be suit-
electronic systems of today. They unfortunately suffer from able; they have been developed and tested in Germany
the phenomenon of metal migration, i.e. silver transfer (Kocker and Stockel 1979) and are noted to possess a high
4.8 Control of Composite Materials 165
apogee-boost motors have already placed seven satellites in cause low bond strength between lay-ups, so that low buckling
their required orbital positions. or twisting forces will easily delaminate this composite
Special surface pretreatments of graphite bres have material. As with all composites, full quality assurance
resulted in exceptionally high bond strengths with epoxy surveillance, including the testing of standard-in-line samples,
matrices. Such composites have been successfully fabricated must be made during all steps of manufacture, and similar
into complex spacecraft antenna face-skins that must survive samples may be submitted to thermal cycling tests (see Fig. 3.
in excess of 10,000 thermal cycles between +100 and 150 C 14b). The ultra-high-modulus bres used in spacecraft have
during a ten-year life in orbit around the Earth. Despite the been found to possess a considerable variation in properties
superior properties of such materials the author has found that from one batch to another, so that particular care is needed in
the human factor in the physical lay-up of such structures can the interpretation of the results of materials tests.
168 4 Spacecraft ManufacturingFailure Prevention
The environmental effects on spacecraft components con- made of either brass (1900 g/m length of waveguide) or
structed of composites have been reviewed (King 1988). aluminium alloys (600 g/m length). The developments made
Atomic oxygen has been seen to cause carbon breepoxy in Germany with thin-walled lightweight (less than 100 g/m
systems to lose weight by surface erosion, whereas Teflon and length) CFRP waveguides metallized on their internal sur-
silicones are almost unaffected. Structures manufactured of faces with either copper, silver, or gold have also led to a
CFRP systems in low Earth orbit will require some form of similar fabrication line for complex components having
protection, as atomic oxygen has been calculated to erode different shapes, such as those needed for choke flange
surfaces to a depth of 3.3 mm during a 30-year period, connections, couplers, and bends (Wagner 1992). For
thickness loss being made up of both bre and matrix degra- straight waveguides, mechanically removable and reusable
dation. Coatings of metals and oxides afford some protection. mandrels are selected. The mandrel material is secret, but it
Metallized CFRP waveguides have proved to be enor- can be machined and polished to a very high accuracy. This
mously successful on both the ERS 1 and the ERS 2 surface is then plated with one or two microns of copper,
spacecraft. This technology is sure to nd increased usage silver, or gold. An adhesion treatment is applied and then
during the next millennium when low attenuation and phase carbon bres are wound around the plated layer. Special
stable power transmission requirements will be more winding directions and layer congurations have been
demanding. Typical commercial spacecraft waveguides are developed to avoid bending, during thermal cycling service
4.8 Control of Composite Materials 169
operation, due to mismatches between the expansion of the and CFRP; torsion mechanical analysis (TMA) to determine
metallization and that of the carbon bres. For complex the glass transition temperature and ensure complete curing
shapes, sacricial mandrels are employed; after curing of the of the laminate; X-ray inspection to determine faults in the
CFRP part they are chemically removed. These parts are metallization; and, ultrasonic C-scan inspection of the CFRP
subjected to many of the quality assurance activities laminate for voids and delaminations. The qualication
described in Chap. 3 of this book. These include tests on testing of these waveguide constructions included more than
samples manufactured in-line with flight parts, such as 10,000 thermal cycles, in the temperature range 70 to
microsectioning (to establish complete densication of the +115 C, without any degradation of peel strength or elec-
laminate and the thickness of the metallization); peel trical attenuation. In orbit they have performed faultlessly for
strength tests to check the adhesion between metallization more than six years.
170 4 Spacecraft ManufacturingFailure Prevention
The CFRP processing advances are very important for constructions of SiC/MAS-L have been fabricated and later
construction of space structures having a high precision. The characterized (Peres et al. 1994). Unitape plies are made by
antenna reflector for the FIRST project will consist of a 3 m winding (solvent, resin, glass powder) slurry-impregnated
diameter parabolic antenna with a focal distance of 1.58 m and SiC yarn onto a mandrel. The plies are stacked into a furnace
a surface accuracy of 7 m RMS (root mean square). It will be and the resin binder is burnt out. Hot-pressing at a high
made totally of CFRP, comprising a sandwich core and two temperature permits the glass ceramic to flow into the inter-
face-skins of high stiffness and low weight (Ehmann 1994). stices between individual bres to form a composite densied
One of the most important factors to understand when up to 98 % of the theoretical density. The microstructure of
employing CFRP is the effects of a humid environment the composite sheets shows a 2-D pattern of seven layers
(moisture) on the dimensional stability, strength, and stiff- where the bre plies are oriented at 0 and 90. A few
ness of the laminated composite. Moisture will diffuse residual pores are visible within the as-manufactured sheets.
through the resin of the CFRP regardless of the resin type. 1000 thermal cycles between 100 and +100 C cause a
The diffusion process is similar for both thermoplastic and small amount of microcracks which run parallel to the bre
thermoset resinmatrix composites (Lucas and Zhou 1993). lay. Thermal shocks from 1000 C into water at 20 C cause
Some dimensional changes can be expected due to swelling microcracks that run at a 45 inclination to the bre lay.
of the resin; this is not detectable in the direction of the Thermal cycling caused a slight reduction in room tempera-
bres, but may be extensive in the width and thickness ture tensile strength of the composite from 179 MPa down to
directions. It is possible to optimize the winding directions 165 MPa. Thermal shock was more dramatic in reducing this
of carbon bres so that the net effect of expansion of the value to 68 MPa. However, in compression and in bending,
resin matrix will have no influence on the nal structure. the strength decreases caused by thermal cycling and thermal
This has, in part, been demonstrated by Hammesfahr et al. shock were not so great, particularly when the mechanical
(1995) in association with the considerations for an X-ray testing was made at elevated temperatures up to 1000 C. It
mirror assembly for the XMM project which was based on was concluded that this new class of material, with its very
carbon-bre technology. Initially the CFRP solution was low coefcient of expansion, should be considered for
preferred for the construction of the 58 Wolter type cylin- high-stability structural applications where thermal cycling
drical mirrors, their supporting structure, and a thermal might be present between 100 and +100 C.
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 4.8.5 CarbonCarbon Composites
presence of moisture.
Carboncarbon composites are employed almost exclusively
in high-temperature applications. Occasionally, very thin
4.8.4 Fibre-Reinforced Glass Ceramics carboncarbon face-skins are selected and bonded onto
honeycomb spacers for lightweight panel applications. The
Glass ceramics are formed in two main processes. Initially, the main uses are associated with propulsion systems and
constituents of the glass are melted then press-formed into re-entry vehicles. During the combustion of the solid chem-
shapes. This is followed by heat treatments that nucleate cer- ical propellants (described in Sect. 2.4) high pressures are
tain phases in the glass microstructure and cause crystalliza- generated together with the liberation of large quantities of
tion. The most common glass ceramic systems are energy in a small volume. The subsequent expansion of these
lithia-alumino-silicate (LAS), zinc-alumino-silicate (ZAS), high-temperature products of combustion through a nozzle
and magnesia-alumino-silicate (MAS). They are used to make converts thermal energy into directed kinetic energy for
radomes, electronic devices and the blanks for mirrors. An spacecraft propulsion. Woven carboncarbon composites
excellent overview has been presented by Hanson and Fernie have been found to be ideal materials for the construction of
(1993). Important characteristics of the glass ceramics are their rocket motor nozzles as they have a low density, a high
ability to form strong ceramic-to-metal bonds, their strength at operational temperatures, and little chemical
tailor-made coefcients of expansion (020 ppm), a high reaction with the exhaust gases during propellant burn time.
strength, and their high softening temperatures (up to 1400 C). Nozzles for the Ariane launch vehicles are shown in Figs. 2.7
A glass ceramic matrix which combines the oxides of and 2.11. A similar external view of a solid propellant apogee
magnesium, aluminium, silicon, and lithium named MAS-L motor is shown in Fig. 4.44. Carboncarbon is also an
has recently been processed. It was selected not only for its exceptionally good material for thermal protection systems
high elastic modulus, low coefcient of expansion, and up to and is selected for the highest temperature regions of re-entry,
1000 C creep strength, but also for its good compatibility vehicles, the nose cones and leading sections of wings (see
and bonding to silicon carbide Nicalon bres. Several Figs. 2.14 and 3.4).
4.8 Control of Composite Materials 171
The fabrication route for carboncarbon composite parts Tight processing parameters must be maintained and the
is complex and involves many variable parameters. For this end-product should contain sufcient offcuts for mechanical
reason it is useful to follow production steps by means of performance tests.
microsectional controls. In-line control samples can be cut A carboncarbon composite consists of two phases of the
from the product at different stages during its production. element carbon, a carbonaceous matrix that is reinforced by
Fig. 4.45 Propellant tanks manufactured from Ti6Al4V; spun domes d Fractograph SEM image showing angle ply lay-up of SiC monol-
are diffusion bonded to SiC bre-reinforced titanium matrix composite aments in an aluminium matrix (this is a high magnication fracture
tubular lengths, a before and, b after burst testing. c Tungsten cored surface of an angle ply 100 m bre in aluminium matrix, an
silicon carbide monolament selected for the tubular lengths showing alternative choice of material to the titanium matrix seen in (c) (Cour-
darker laminar carbon coating interface to the titanium alloy matrix. tesy of TISIC, Farnborough)
172 4 Spacecraft ManufacturingFailure Prevention
carbon (or graphite) bres. The carbon bres may be woven 4.9 Control of Capillary Screens
and physically formed into the required shape of the
piece-part. For low-cost products, where little mechanical Present spacecraft liquid propulsion systems concentrate on
loading will be encountered, nonwoven bres are used. The monopropellant hydrazine for in-orbit propulsion. These
carbon bres are impregnated with a polymeric resin which systems use catalytic and electrothermal decomposition
is then cured. The composite is then subjected to its rst techniques that convert liquid hydrazine into gaseous
carbonization, when it is heat-treated to temperatures hydrogen and ammonia. Short bursts of these gases from
between 500 and 1000 C, and in this stage the polymer small thruster motors enable spacecraft manoeuvres by
matrix is pyrolysed to form the secondary carbon phase, three-axis attitude control. The large titanium vessels that
which is porous and may contain transverse matrix cracking. contain liquid hydrazine fuel onboard satellite presently
Microsectioning of control samples, together with nonde- incorporate a rubber bladder that is pressurized by a gas, so
structive testing of the remaining material, will establish if forcing the fuel to flow to the thruster motors. However, such
defects such as voids and microcracks are in sufcient organic membranes rupture after relatively few expulsion
numbers to render the material as being unsuitable for fur- cycles and can be dissolved to cause contamination of the
ther processing, i.e. workmanship standards would identify hydrazine fuel as described in Sect. 5.8. A new flow system
rejectable situations. This is an important step, as further should make obsolete the perishable bladderit is known as
proprietary processing operations that involve a second the surface-tension tank. See also Sects. 5.8 and 5.20.
heat-treatment can cause the isolated defects to develop into Spherical surface-tension tanks manufactured in Europe
gross delaminations which cannot be densied by further are flying on Frances Spot spacecraft, and have been pro-
processing. 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
4.8.6 Metal Matrix Composites for Spacecraft Propellant Management Device (PMD) which enables
Pressure Vessels gas-free liquid propellants to be fed to the engines (e.g.
thrusters). The PMD comprises an internal assembly of tubes,
Metal matrix composites for structural applications were vanes, and ne-mesh screens which, by capillary action,
described in Sect. 4.8.1. The development of bre reinforced perform the propellant separation and gas-free feeding.
pressure vessels and propellant tanks by TISICS Ltd., Once the PMD is incorporated into the tank, the two
Farnborough started with a xenon gas tank design based on hemispheres are either TIG-welded or EB-welded along their
requirements for the SSTL Tech-Demo-Sat-1 satellite. equatorial girth. Present PMD capillary screens are made of
TISICS is the only commercial plant outside of the US 304L stainless-steel mesh, woven to provide 5 m diameter
producing CVD (chemical vapour deposition) silicon car- interstices. To achieve greater weight savings, the PMDs of
bide monolaments. Pressure vessels were designed and the future will be constructed of titanium alloy (only a few
fabricated with their own SiC bre-reinforced titanium materials, including titanium and stainless steel, are fully
matrix composite in the tubular region. These tanks (see compatible with the hydrazine propellant), and efforts are
Fig. 4.45) were tested and passed against flight release test being made to acquire titanium capillary screens. Recent
schedules. Testing to burst at greater than 4-times the powder metallurgy techniques have produced a Ti6A14V
maximum expected operating pressure resulted in a rupture alloy with a controlled porosity level of around 40 % by using
failure similar to a metallic rather than composite failure a combination of high-temperature and high-argon backll
mode. Quality controls included metallography of in-line pressure. This could nd applications within PMD designs.
test samples and X-ray tomography of the composite region The pores have a cubic morphology (Fig. 4.46a) and can be
and the interface-to-monolithic titanium diffusion bond. made either interconnecting or ne and discrete
Further work is underway to develop thin wall titanium and (non-connecting). Although in their infancy, titanium screens
aluminium composite propellant and pressure tanks for the made from 5 m diameter bres (Fig. 4.46b) are being
new green propellants (Fig. 4.45c, d). Fibre performance evaluated. Titanium wires are initially covered with protec-
and increased production efciency has been achieved for tive alloy sheaths and drawn through dies until a minimum
both the 100 m bres required for thin wall angle ply diameter is reached. The drawn wires are annealed by heat
systems, and the stronger and stiffer 140 m bres for most treatment, bundled together again, and redrawn. These steps
other applications. are repeated until the desired nal bre dimension is
4.9 Control of Capillary Screens 173
Fig. 4.48 SEM view of cracked knee on bent package lead: a general Fig. 4.49 SEM view of de-golded and solder-tinned packaged lead:
appearance; b the base Kovar is exposed beneath the cracked a general appearance; b solder has not covered volume of crack
gold-plated EN
containing 12 % or more phosphorus are said to retain their furnace, is that of laser annealingthis readily achieves
nonmagnetic properties even after heat treatment at 300 C nanocrystallisation of the electroless nickel, without affecting
(Tulsi 1986). Experience has shown that heat treatments at the properties of its substrate material. A recent study into the
200 C, as for hydrogen embrittlement relief, result in an utilisation of lasers for modifying NiP and Ni-W-P elec-
improved corrosion resistance. Heat treatments made over the troless coatings, in an attempt to increase the corrosion pro-
range 250400 C cause a signicant reduction in corrosion tective properties of these coatings, has been published by
resistance due to the precipitation of nickel phosphide. This Liu et al. (2014). Laser heat treatment of a Ni-3.9W-13.4P
causes the EN deposit to harden and can promote microc- (wt%) as-plated amorphous coating causes nanorecrystalli-
racking because of anisotropic dimensional changes. Fig- sation, and the precipitation of the Ni3P phase, this is a stable
ure 4.50b illustrates the crystal structure of the phosphorus compound in a corrosive medium. Corrosion (preferential
nickel layer. Possibly brittle cleavage fracture occurs at the dissolution) of nickel occurs, leading to an enriched phos-
boundaries of these columnar-shaped crystals. The internal phorous surface layer and a surface coated with adsorbed
volume of these crystals appears to contain nano-sized pre- hypophosphite anions which blocks further hydration, or
cipitates of nickel phosphide. Differential thermal expansion corrosion of nickel. For the as-plated, amorphous coating,
between the EN deposit and its substrate will further impair this corrosion protection mechanism is weak. It is only after
deposit integrity. The problem related to black pad will be heat treatment, with the formation of Ni3P, that the best
further discussed in Sect. 6.6.2. corrosion resistance is achieved. Such desirable nishes can
As mentioned above, heat treatment after plating is often be attained (Liu et al. 2014) without damaging underlying
carried out; this ensures a good adhesion of nickel to sub- material, by the laser annealing/surface modication process.
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 4.11 Control of Electroforming Processes
the relatively high temperature of 200300 C. This process
may be damaging for the substrate, it may alter the substrates A wide range of spacecraft piece-parts have been success-
heat-treatment condition, and some parts may be restricted fully fabricated by electro-forming. They include electro-
from entering a furnace due to their large size. A relatively formed copper waveguides, CFRP waveguides supporting
new process which can obviate the need for an oven or silver-plated internal faces, microwave horns, and a
Fig. 4.51 a General view of microwave horn. b Showing electroformed internal surface (dark area of damaged castellations are arrowed)
onto the mandrel. The next step is to apply a thicker back-up using 5-axis numerical control milling machines. The mandrel
metallic layer to strengthen the waveguide. The back-up can was then silver-plated following a proprietary pretreatment.
be solid copper or nickel, and this may later be machined to the This was followed by nickel plating from a nickel sulphamate
correct outer dimensions of the nished waveguide. Complex electrolyte (an alternative plating solution based on nickel
multi-horn feeds for the British Orian satellite (Teare 1994) sulphate/nickel chloride was unsuitable because it produced
could only be manufactured by electroforming, owing to the nickel deposits having a high internal stress level). The
complex shape of the horns and the stringent geometric mandrel was removed by conventional dissolution in sodium
requirements. A one-piece aluminium mandrel was produced hydroxide without damaging either the internal silver or the
4.11 Control of Electroforming Processes 179
ller metal, has been the slower heat-up rates of the thicker under low loading conditions was attributed to the presence
components in the assembly. This can lead to insufcient of zones of incomplete braze penetration, as highlighted in
braze metal flow and the formation of unsatisfactory joints. Fig. 4.55. The manufacture of automobile semi-nished
Recent developments have shown that brazing sheet or foil products by means of brazing and the usage of braze clad-
can be fabricated with a core of 3003 alloy, clad on both ding is fully described in the useful reference by Sicking
sides with AlSi brazing ller metal produced from (2005).
high-purity components. The brazing sheet can be tack Finally, some general precautions must be taken during
welded between castings, and plates made from different the selection of a suitable aluminium alloy for flux bath dip
aluminium alloys. After special cleaning, the assembly is brazing:
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 The solidus temperature of the alloy being joined must be
(possibly only 14 C below the melting point of the casting, above the liquidus temperature of the braze alloy.
so that temperature control is critical). The assembly is Chemical composition limits should not exceed 0.5 %
allowed to cool slightly so that the brazing salt can be flu- copper, 1.5 % silicon, and 1.75 % magnesium (wrought
shed away and then slowly cooled to room temperature. sheet and plate meeting these requirements are: pure
Large parts for the USA Space Shuttle Orbiter have been aluminium 1050A, 1080A, 1200, 1350;
successfully fabricated in this manner (Anon 1981). They non-heat-treatable alloy 3103; and the heat-treatable
are later radiographically inspected and nally heat-treated alloys 6061, 6063, and 6082).
and age-hardened. Non-heat-treatable materials are brazed with Al 7.5 %Si,
It should be noted that a competing process to dip brazing which is fully molten at 615 C; in the case of the
is known as controlled atmosphere brazing (CAB). Only in heat-treatable alloys the solidus of the parent metal
recent years has CAB has become a state of the art tech- determines the choice of braze alloy, and an Al 1013 %
nology; it has been developed for the production of alu- Si braze alloy having a melting range of 595585 C is
minium heat exchangers used in the automobile industry. used (additions of 25 % copper to the Al 1013 %Si
CAB is not known to have been applied to the production of alloys will lower the braze metal solidus temperature to
spacecraft parts as it involves the use of specialized costly approximately 550 C).
equipment, ideal for mass production but not for the few As the brazing temperature is higher than the recrystal-
parts used in the space industry. Additionally, the actual lization temperature of aluminium and its alloys, anneal-
furnace atmosphere is a well-kept secret. It is known that ing takes place during brazing, and much mechanical
successful ller alloys are in the AWS BAlSi series, and that strength is consequently lost; the heat-treatable alloys
a suitable flux, containing potassium fluoroaluminate com- 6061, 6063, and 6082 can be strengthened by quenching
pounds, has been developed by Solvay Fluor GmbH having after brazing followed by natural or articial ageing.
the name Nocolok. This flux is considered to be The sublimation of aluminium alloys and their brazing
non-corrosive before, and as a residue after brazing, being alloys needs to be considered when applied to spacecraft
only active at the brazing temperature whilst in the protec- parts designed to operate under vacuum and at high
tive inert CAB atmosphere (Zhao et al. 2013). Obviously temperatures; here, the reader is directed to Sect. 5.6.1 for
such fluxes would need to be entirely removed for any an assessment of sublimation data.
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 4.13 Considerations for the Assembly
alloys. Figure 4.55 illustrates the importance of accurate of Subsystems by Welding
jigging. Here, brazing sheet has been stamped into a com-
plex shim to interface between the two half-bodies of a 4.13.1 General Welding Methods and Controls
spacecraft electromechanical device that would, when
brazed, incorporate several externally sealed cavities. Non- Automatic welding processes are becoming much more
axial compressive forces applied by an unsuitable jigging widely used by both the larger and the smaller companies
arrangement have caused tilting at one side of the device, engaged in the assembly of spacecraft subsystems. However,
and, although a satisfactory bond has been formed along the manual tungsten inert gas (TIG) arc welding still nds a
lower side of the brazing shim, there has been only partial place owing to its versatility, and novel innovations may be
contact between the upper surfaces. All external braze llets found for applications when the occasional welding of parts
were inspected, assessed to be acceptable, and the device is performed by the contractor. For instance, a range of
was then electroplated with silver. The failure of one device flexible welding enclosures have been designed for
182 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) engineers have access to these data if they are to raise quality
failures occur. The failure of welds by brittle fracture is more and prevent costly failures.
serious in its consequence but rarely has been seen to occur. Many descriptions of the commonly used TIG welding
The fatigue life of a structural part is determined by the process are described throughout this book. Not described
number of stress fluctuations, the magnitude of these fluc- are the precautions that need to be taken to avoid adversely
tuations, and the kind of stress concentration in the part (see affecting the health of the welding personnel. For instance,
Fig. 5.2 for a typical fatigue fracture). The designer has arc welding can give rise to ozone emissions and fumes that
control over the last two and is able to plan to avoid sudden exceed the permitted levels dened by national standards
changes in dimension and, by the choice of weld details, to (Anderson and Wiktorowicz 1995; Sterjovski et al. 2010)
prevent cracks and porosity caused by improper processing. this does produce respiratory problems and can even con-
There are large differences in the fatigue performance of taminate and cause damage to spacecraft electrical hardware.
various welded geometries, and the problems of selecting The eyes must also be completely protected against injury
optimum designs cannot be overstated. Similarly, the from directly viewing bright light emitted from welding arcs.
harmful effect of a last minute, uncontrolled, attachment Some case studies can be presented for the more recently
weld is all too often overlooked. This is illustrated by the developed welding techniques such as electron beam and
failure of a hydrazine thruster chamber (see Fig. 4.14) which laser welding methods, and these will be presented in the
was modied by an electron beam weld that was never following paragraphs.
properly assessed several months prior to the launch of a
telecommunications satellite. Although the modication
overcame one problem, it generated another, which even- 4.13.2 Electron Beam Welding
tually caused the device to fail. The launch was delayed and
the defective thrusters were replaced at the launch site. Our The fusion process that is accomplished by electron beam
knowledge about the influence of weld defects is always welding (EBW) occurs when a beam of electrons is accel-
increasing. It is important that both designers and process erated up to nearly the speed of light and then focused onto
4.13 Considerations for the Assembly of Subsystems by Welding 185
the metallic pieces to be joined. The kinetic energy of the EBWs are considered to be the most favorable welding
electrons is converted into thermal energy, causing the metal techniques for spacecraft piece-parts. Much experience has
to melt and vaporize. The very high power density can be been gained in recent years and unusually combinations of
focused onto a spot having only about 0.25 mm diameter. metallic materials may be successfully welded when other
The depth of penetration of the electron beam is dependent on welding methods have failed. This is true for titanium joined
how well it is focused, its power density, and the type of to beryllium, stainless steel to copper, and deep-drawn alu-
material being welded. Copper can dissipate heat rapidly minium to itself (ingrained lubricating oils were outgassed in
(Fig. 4.14) and has a wider weld pool than titanium (Fig. 3.6) the EBW vacuum chamber, whereas under arc welding the
or aluminium. The greatest advantage of EBW is its ability to occluded oil produced massive porosity in the weld pool).
produce welds that are far deeper and narrower than TIG arc Some care is needed when magnetic materials are being
welds (compare Fig. 3.6 with Fig. 4.8 for similar wall joined, as the electron beam can meander off-centre owing to
thicknesses of Ti6A14V alloy, and Fig. 4.20a with b for magnetic eld effects. Also, because of the high-energy
Al-2219 alloy plate). Also, again illustrated by Fig. 4.20a, b, density of the electron beam, care must be taken to avoid the
the EBW eliminates the need for multiple-pass welds. Weld kind of unbeaded spatter as was shown in Fig. 3.6. The
pools formed by conventional arc welding are not only wide, notoriously difcult combination of titanium to stainless steel
but are associated with a wide heat-affected zone, as shown in could not be joined by EBW, as very brittle irontitanium
Fig. 4.8. A similar EBW can have virtually no heat-affected intermetallics are formed in the weld pool. This interfacing of
zone and this results in minimized distortion, enables welding metals is often designed out of RCS fuel lines although some
to be made in close proximity to sensitive areas such as attempts to use alternative joining methods have been
glass-to-metal seals, and of most importance, the overall attempted (see Sects. 4.13.4, 4.18, and 6.10).
mechanical properties of the joined parts are virtually the There are two important disadvantages associated with
same as the parent materials. Figure 4.59 was a detail from a EBW. The main one is that this process needs a vacuum
very deep EBW, made in a Ti6A14V housing, situated chamberhere, the size of the chamber controls the dimen-
adjacent to an electro-optical sensor. The fragile device was sions of the parts to be joined. The second problem is that
not damaged and, as shown in the microsection, the X-rays are generated at the electron beam-to-workpiece inter-
microstructure of the parent material is not affected by this face and therefore operations such as the lid sealing of elec-
butt-weld. The lack of any heat-affected zone adjacent to the tronic packages or the use of EBW close to electronic boxes
weld pool was conrmed by a microhardness traverse across must be prohibited as this radiation may damage components.
the weld. 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.
Complex adhesive bonding processes that require surface need power levels up to 3 kW during eclipse and up to 10 kW
preparation of the parts to be joined coupled with heat cycles sunlight power. Earth resources satellites may also need
for curing that may damage surrounding areas. 10 kW but, like the HST, they are in LEO where there will be
In some ways the welding of thermoplastics is similar to greater stresses on the arrays and batteries. Manned missions
welding of the light metals. Methods include laser welding, such as Spacelab and the International Space Station are also
friction stir welding and ultrasonic welding. Only like-to-like in LEO and need power levels of between 3 and 50 kW.
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 4.14.2 Welded Solar Arrays
being transparent to the laser beam. Usually the parts are
placed together by a pressure xture so enabling weld widths The current delivered by a solar array is essentially propor-
of between 10 and 100 m. Specialist papers should be tional to the solar energy received by the array. This current
consulted when specifying methods for non-destructive will vary owing to the pointing angle of the array to the Sun
inspection and mechanical testing of these connections (cosine factor) and the distance between the Sun and the
(Troughton et al. 2013). 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
4.14 Control of Power System Weldments transfer from the array is by radiation to deep space during
periods of eclipse. Spacecraft in LEO, such as the Hubble
4.14.1 General Space Telescope (HST), are at this moment surviving about
30,000 thermal cycles between 100 and +100 C during a
The two most important elements of every spacecraft power ve-year period of operation. A more benign orbit may cause
system are the energy sourceusually a solar arrayand the temperature cycling between 80 and +80 C whereas a
energy storagemainly nickelcadmium or nickel-hydrogen severe cycle would be from 150 to +100 C. The HST has
batteries used to provide basic power in eclipse, or peak two solar panel assemblies as shown in Fig. 4.63 (each panel
power in sunlight. The charging and discharging of the is made up of ve power-generating sections).
spacecrafts batteries will be determined by the routine Every cycle between hot and cold results in the array
interruption of the power source as the Earth passes between components expanding and contracting according to their
the orbiting satellite and the Sun. The characteristics of the individual coefcients of thermal expansion (CTE). The sin-
orbit can be drawn into four main groupings: gle strings of solar cells are interconnected by means of flex-
ible foils that have been formed to include a stress-relief bend.
low Earth orbit (LEO): close to the Earth the eclipses last One design of solar array is shown schematically in Fig. 4.64;
about 0.5 h during each 1.5 h orbit here, the solar array has a particularly high electromagnetic
geostationary orbit (GEO): 36,000 km above the Earth, cleanliness due to the presence of grounded cover glasses.
with a 1.2 h eclipse for a 24 h orbit In the past 30 years, monocrystalline silicon solar cells
highly elliptical orbit (HEO): perigee close to the Earth have been used on European space solar generators. The
and apogee far removed, which gives rise to variable cells are connected by silver mesh or silver foil. Early joints
eclipse/orbit ratios between the silver interconnectors and the silver-plated front
deep space orbits: result in very few eclipses, but large or back face of the cell were achieved by soldering. The
variations in solar energy; in this case a radioisotope cyclic temperature range and high upper temperature caused
generator may be the energy source. the microstructure of the solder joints to coarsen and rapidly
fatigue. Welded interconnector-to-cell joints were developed
The most demanding orbit listed above is the LEO, with and accelerated testing together with actual flight experience
its high eclipse-to-sunlight ratio of about 30 %. This means has proved that these joints will withstand 10 years in LEO
that a heavy discharge of power during the 0.5 h eclipse will and 15 years in GEO. The gradual degradation of the welded
need a full battery replenishment in a period of less than 1 h. joints can be followed nondestructively by means of scan-
Other orbits will provide for longer charging cycles, but there ning laser acoustic microscopy (SLAM, described in Sect. 3.
can be quite different power requirements imposed by the 2.3.2) or by destructive metallography. The method is shown
mission. Science missions have various orbits and generally in Fig. 3.15 and the captions of Figs. 4.65, 4.66, 4.67 and
require lower power levels (200 W1.5 kW). The Hubble 4.68 describe actual cases where welding processes were
Space Telescope (HST) array, with a surface area of about assessed for quality and reliability.
130 m2, is providing 4.6 kW of electrical power of 34 V. Large thermal stresses build up on the welded joints
Telecommunication satellites will be in GEO orbits and may 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) of 48,760 BSFR silicon solar cells to provide more than 5 kW of
solar array shortly after retrieval. This shows one of the two HST solar electrical power. During inspection (see also Fig. 8.4), 704 large
array wing assemblies being opened out at the MMS-UK facility where hypervelocity impacts were photographed. 148 impacts had completely
it had originally been fabricated. All outer surfaces were photographed penetrated the arrays thickness but they did not cause functional failure
to assess the effects of 3.6 years operation at 600 km above the Earth. of the array (a total of 3862 impacts was counted on the front side of one
This is the largest roll-out solar generator built to date, employing a total 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. 4.68 Microsection of design. All can indirectly affect the nal weld quality, and
silver-coated molybdenum the following controls are essential:
interconnector weld from the
Hubble Space Telescope (see also
Fig. 4.63) after 30,000 thermal the interconnector material for chemical purity, hardness,
cycles between 100 and +100 thickness, and form;
C. The weld zone has no fused the solar cell for signs of contamination of the silver
nugget, which is commonly
observed between metallized layer (e.g. sulphides and oxides), thickness,
resistance-welded parts, but the and surface roughness; and
evidence of grain growth the repeatability of all welding parameters, such as
(arrows) does indicate that the electrode dimensions, pressure, voltage, and duration of
temperature, time, and pressure of
welding have produced weld pulse.
strain-induced, grain-boundary
migration. There is no sign of Even if these joints are made to the most optimized
thermal fatigue 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.
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 195
4.15 Problems Associated with Residual likely that these corrosive processing environments caused
Stresses in Weldments crack initiation sites which later were to grow by a
stress-corrosion mechanism.
Electron beam (EB) welding is frequently employed during Based on the failure analysis results, subsequent struts were
spacecraft construction. It is particularly suitable for weld- carefully heat-treated in a furnace immediately after the
ments that are bodies of rotation, e.g. pressure vessels and welding operation. The residual welding stresses, which can
tubular assemblies. The welding motion is mechanized, and reach levels as high as the yield strength of the welded struc-
generally narrower weld widthsand therefore reduced ture, were sufciently relieved during this heat cycle. At the
distortion and heat-affected zones (HAZ)are produced. It same time, the solution-treated HAZ was modied to a state of
should be remembered that even EB welds will contain controlled higher strength owing to precipitation-hardening.
certain residual stresses which are caused by expansion No further cracking problems were experienced after the
differentials, shrinkage of the weld metal as it solidies, and adoption of this post-weld heat-treatment procedure.
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 4.16 Electromagnetic Emission from TIG
had been machined from the same high-strength Welding Equipment
precipitation-hardening alloy Al 2618.
Figure 4.57 shows the appearance of this painted cylin- Arc welding equipment is known to produce a high level of
drical weldment; the section shows that adequate weld electromagnetic emission which can cause interference to
penetration has fused and welded the end of the tube to the electrical communications systems. Laws covering the
end-piece without the need for a ller metal. During inte- electromagnetic compatibility of industrial welding equip-
gration, this weldment was observed to fracture under very ment came into force within the European Union in 1996.
low loading conditions. The crack path had propagated This legislation makes it an offence to produce excessive
through the weld region in a manner superimposed on the electromagnetic emission. The emission (mainly radiation)
micrograph. can be produced by both the welding power source and the
As part of this failure investigation, an as-welded, but arc itself (Lucas 1995). Unfortunately it is impossible to
never loaded, strut was microsectioned. This sample was eliminate these emissionsthey should also be considered
lightly etched (Fig. 4.58) to reveal the presence of a ne when welders bodies may be exposed to locally conned
network of cracks at the weld pool-to-parent metal interface. high magnetic eld strengths, such near the welding torch,
Of great importance is the evidence of multipass welding as its cables and power source (Mair 2005). They begin at the
shown by the liquation zones present in the etched weld moment the welding arc is started. The space between the
pool. Some large, secondary intergranular cracks have also welding electrode and the metal workpiece is ionized by
formed in the HAZ. It was later conrmed that cracks did not high-voltage sparks (540 kV at a frequency within the
exist in newly made weldments. In fact, they had initiated range 101000 kHz). Further sparks are generated during
and grown at some stage during the struts 6-month storage welding owing to the presence of surface oxides on the
period. In this case, the reasons for crack initiation and piece-parts being joined; this has an effect similar to light-
growth are believed to be twofold. The strut material is a ning as miniature flashes propagate through the shielding gas
rather complex precipitation-hardening alloy which under- (e.g. argon). In addition, it has been observed that the
goes various submicroscopic changes during fabrication. welding cables connecting the weld head to the power
The alloy adjacent to the weld pool (HAZ) heats up during supply/controller, which may be 10 m away, will act as
the welding operation and undergoes a localized solution antennae, causing the level of radiation to increase.
treatment. During the second, third, or subsequent fusion Many electronic devices are sensitive to electromagnetic
there is extreme thermal expansion and contraction in the radiation. Components such as CMOS packages are partic-
base of the weld. The local stresses may be so great as to ularly vulnerable and are known to have failed owing to
cause yielding and bending of the material (Verderaime and radiation emitted from TIG welding equipment which was
Vaughan 1995). At room temperature, the residual stresses, being used, in close proximity to the electronic circuitry, to
which have been set up within the material as a result of carry out repair welds on both the Voyager and Galileo
weld metal contraction, cause the solution-treated HAZ to spacecraft (Lumsden and Whittlesey 1981). Concerns that
gradually harden because of natural strain ageing. Unfortu- similar damage to electronic devices might occur were
nately, such a condition may become embrittled, particularly expressed during the replacement of hydrazine propulsion
if exposed to a mildly corrosive environment. It should be system tubing (RCS thruster pipes) on board flight spacecraft
noted that, following the welding process, these struts had of the Cluster project. These pure titanium tubes were
been chemically cleaned, alodined, and nally painted. It is thought to be over-etched and had shown signs of weakness
196 4 Spacecraft ManufacturingFailure Prevention
when incorrect xation procedures had caused a length of time there are no examples of any aerospace usage, but
unsupported tube to fracture during ground transportation of owing to their characteristics of light weight, surface sta-
the spacecraft. bility, potentially high operation temperatures, and oxidation
A specialized circumferential in-place pipe welding head, resistance, the titanium aluminides are certainly worth a
developed for aerospace applications, was used to both cut mention as being revolutionary engineering materials.
the defective pipes (3.17 mm diameter, 0.25 mm wall thick- These new materials are not solid solutionable alloys, but
ness) and, later, weld together the replacement pipes. Some are intermetallic compounds having a fairly xed composi-
manual saw cutting of the pipes was also undertaken using tion. There are two classes of titanium aluminide. The rst
nitrogen purging and special contamination controls to ensure has the composition Ti3Al, termed the alpha two alloys with
that slivers of metal were not introduced into the propellant useful structural properties up to 700 C. The second is based
system. Welding was performed by rotating the TIG electrode on TiAl, the so-called gamma titanium aluminides, and
around the joint on a ring gear. The butt joints were welded these are reported to have been developed in the USA to
autogenously with arc voltage and other parameters con- reach strength levels as high as 1000 MPa at room temper-
trolled by a microprocessor. Experiments were made to ature and 500 MPa at 1000 C (Kim 1995). The European
ensure that minimal electromagnetic emission was produced space community appears to favour the development of TiAl
during the weld strike. This work followed the practices with aluminium contents of 48 at.%. As a binary alloy it has a
outlined by Lumsden and Whittlesey (1981) at the Jet moderate elongation to fracture at room temperature, but
Propulsion Laboratory. Here, the use of a whisker arc start inadequate creep strength and poor oxidation resistance at
instead of the usual transient-arc method was noted to reduce elevated temperatures. These disadvantages are avoided
the initial burst of radiation. A 50 m diameter pure titanium when the gamma aluminide is alloyed with approximately
metal whisker is placed between the electrode and workpiece; 2 at.% of chromium, manganese, or vanadium, and/or 24 at.
it is vaporized by the weld current, leaving a plasma to initiate % of niobium, tantalum, molybdenum, or tungsten. More
the arc. The spark gap was reduced to 0.5 mm, to reduce the recent developments also include the deliberate additions of
spark voltage. Low-impedance braid (less than 10 m) was up to 1 at.% of silicon or boron. The initial group of elements
used to ground the spacecraft structure and the welder chassis is introduced to improve further the room temperature duc-
to the facility earth. The welder-to-weld head cable was tility (they enhance mechanical twinning and the movement
completely shielded with additional aluminium foil and of dislocations). The second and third groups improve the
Eccoshield tape wrap and grounded to the welder chassis. high-temperature properties of creep strength and oxidation
Some of the sensitive electronic boxes were removed prior to resistance. It is very important to avoid the pick-up of
the repair welding operation, and others were covered with interstitial impurities such as oxygen, as these severely
aluminium foil and rectangular ferrite pieces to isolate them reduce the room temperature ductility from about 4 to 0.5 %
from the electromagnetic radiation, albeit reduced by the elongation at fracture. At temperatures of between 600 and
whisker-wire start. Prior to spacecraft repair welds being 700 C there is a remarkable increase in ductility, with
made, test specimens were produced and controlled by fracture elongations of up to 85 %. This increase is due to
X-radiography, bend testing, microhardness checks, and dynamic recovery enhanced grain boundary sliding, and the
metallography. No radiation damage was caused during these start of dynamic recrystallization (Clemens and Schretter
operations and the resulting flight weldments are considered 1996). The fabrication routes for forging and rolling
to be of excellent metallurgical quality. gamma-TiAl alloys (mainly Ti48Al2Cr) have been well
researched and the different possibilities of optimizing the
microstructure of these intermetallics by heat treatment to
4.17 Titanium Aluminides produce a variety of structures from near gamma grains, to
for High-Temperature duplex grains, and even ne lamellae in a small grain size,
Applications have been described in a number of papers published by
Plansee Aktiengesellschaft and referenced by Clemens and
Structural metals for spacecraft applications are predomi- Schretter. Sheet material with a small grain size (about 10 m
nantly selected from either aluminium alloys or titanium diameter) has been tested in tension at 950 C and shown to
alloys. The latter are well established as lightweight engi- have elongations of 200 %. Dynamic recrystallization pro-
neering materials and can be developed to withstand tem- cesses that accompany the deformation at these high tem-
peratures in the order of 600 C. A new class of material, the peratures are responsible for achieving superplastic forming
titanium aluminides, is receiving much attention in the 1990s processes where the deformation ratio is in the order of
and it is probable that they will nd applications in the next 100 %. A prototype example of Ti48Al2Cr sheet that has
generation of aero engines, space launcher vehicles with been pack rolled to dimensions up to 480 205 2 mm and
re-entry capabilities, and space platforms. At the present then superplastically deformed and diffusion bonded can be
4.17 Titanium Aluminides for High-Temperature Applications 197
seen in Fig. 4.71. Whereas diffusion bonding of the inter- composition and place of invention (nickeltitanium Naval
metallic gamma-TiAl is well established as a joining process, Ordnance Laboratory). Other, less-expensive shape-memory
more research needs to be made into how it may be joined by alloys based on copper, zinc, and aluminium were developed
conventional methods of welding. It would appear that the in a collaboration between the former Fulmer Research
heat input (and related expansion or contraction) during Institute and the Raychem Corporation. These ternary alloys
welding of all the TiAl intermetallics causes local cracking were manufactured into devices which operated as thermo-
in the colder parent metal where the thermal stresses and static valves for controlling domestic radiators and opening
distortions are difcult to be absorbed by a material with a or closing window arms. More recently, high-temperature
room temperature elongation of only 14 %. Crack-free shape-memory alloys have been developed (Golberg et al.
joints have been made by heating gamma-TiAl to above its 1995). These are based on titaniumpalladium alloys with a
600 C brittle-to-ductile transition temperature just prior to martensitic transformation temperature of 550 C, and these
welding, but such preheating is an impractical process for could nd future use in hot spacecraft structures. The
large piece-parts unless a special welding furnace is available shape-memory alloys possess a mechanical memory. When
(Clements et al. 1996). A stability test has been performed on they are heated to a certain transformation temperature they
a welded panel (500 250 1 mm) supporting superplasti- remember a previous shape and promptly revert back to that
cally formed and laser-welded stiffeners (similar to the shape. Considerable force is exerted as the alloy changes
diffusion-bonded panel shown in Fig. 4.71). The test panel shape, and this can be used to perform useful work, such as
survived thermal and compressive loading to test tempera- the opening and closing of ofce windows, or to provide
tures of 750 C. This material has been proposed for further power for activation and operation of robotic arms on plan-
development as it has the potential to replace superalloys in etary lander vehicles.
the temperature range of 700950 C and therefore could nd Shape-memory alloys (SMAs) undergo a crystalline
applications in hypersonic aerospace technology. 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
4.18 Shape-Memory Alloys for Spacecraft lattice structure. For an understanding of the process of
Devices shape-change the crystallographic forms of the alloy must be
known. Nickeltitanium alloys have the majority of aero-
Several metal alloys exhibit the so-called shape-memory space applications and the behavior of these alloys is fairly
effect. The rst alloy to be exploited for commercial appli- well known. As a demonstration we can consider the man-
cations was one discovered in the early 1960s by W. ufacturing route for a spacecraft antenna that has a form
J. Buehler. It has been named Nitinol, reflecting its similar to a dish-shaped spiders web. The construction
198 4 Spacecraft ManufacturingFailure Prevention
material is Nitinol wire that has been wound into a large coil. deployable struts. Examples are sketched in Figs. 4.72, 4.73
The coil is placed into an oven containing an inert gas. The and 4.74.
temperature is raised to 650 C and the metal undergoes a The SMS alloys have been found suitable for the cou-
phase change to that having the crystal structure of austenite. pling of a variety of tube materials, particularly the
The wire is now cooled to room temperature. As it cools a difcult-to-weld stainless steel to titanium alloy tubes. Fig-
new crystal phase called martensite begins to form at a ure 4.75 shows the effect of heat on previously expanded
precise temperature of 60 C; the wire is completely trans- Cryocon and Cryot coupling rings. Each ring is made of
formed to martensite on reaching 52 C. While the wire is in Nitinol (NiTi), NiTiFe or NiTiNb, and has a 35 % rever-
the martensitic state, conveniently at room temperature, it is sible change in diameter. On reaching the
cut into short lengths and tack welded into the form of the martensite-to-austenite transition temperature, the ring con-
dish-shaped spiders web. The nished antenna can then be tracts and four knife-like protrusions lock mechanically into
crushed, or folded into a small volume (suitable for the the tubes surfaces to form a hermetic seal. The force that the
conned space of a launch vehicle). In order to restore the alloy exerts in reducing its strain is approximately the yield
original shape of the antenna the folded volume is heated. strength of the warm alloy. Since this is about 410 MPa, an
On reaching a temperature, also well dened at 71 C, the enormous force is present around the tubes circumference.
martensitic phase begins to transform into the austenitic These couplings have been used in special situations
phase and this modication of the metals crystal structure is namely on a European space project when a ll and drain
accompanied by dimensional changes in the wire that cause nozzle, located in the forged dome of a Ti6Al4V hydrazine
the individual segments of the web to unfurl. At 77 C the tank, could not be re-welded owing to its very short length.
remaining martensite is completely transformed to austenite The Cryot coupling ring was successful as a last resort-
and the antenna will be seen to have been restored to its repair and was adequately qualied for this unique appli-
original shape. cation. It is important to note, based on lessons learnt, that
The antenna example has been fully demonstrated and, by properly welded fluid connections are inherently more
the same metallurgical phase changes, Nitinol has been reliable than mechanical ttings and weldments should be
employed to produce hinged tubes, deployable masts, and incorporated in fluid propulsion designs whenever possible
4.18 Shape-Memory Alloys for Spacecraft Devices 199
variance only accepted when accompanied by appropriate be employed in reaction control systems. As for the previ-
risk analyses have been assessed (Gilbrech et al. 2005). ously described HST application, release triggers made from
The Solar-Array Drive Mechanism (SADM) for the Nitinol avoid the complex circuitry associated with the safe
Hubble Space Telescope (HST) was built with release trig- keeping of pyrotechnic devices (particularly when there is a
gers made of Nitinol (Fig. 4.76). 60 elements of Nitinol were danger of lightning strikes) and, when activated, they do not
procured and submitted to metallurgical characterization transmit shock loads to the spacecraft. However, it is difcult
(each, having a diameter of 10 mm and a length of 56 mm, to accurately predict the performance of designs involving
produces a force of more than 450 N over a full stroke of shape-memory alloys; basic data are lacking and a path of
6 mm). Once the HST was in orbit the trigger elements were trial and error under the guidance of a metallurgical engineer
successfully activated by electrical-resistance heating to a may have to be followed in order to achieve a reliable
temperature of 115 C (this particular alloy contained a small product. Predictions for service life data may need to be
addition of cobalt in order to slightly raise the transition generated in the laboratory. For instance, if the Nitinol
temperature and avoid the possibility of a premature acti- devices are to be used repeatedly (i.e. about 20 times) it has
vation caused by solar heating). been seen that stresses should not exceed 150 MPa and
Shape-memory alloys have been proposed for a wide strains need to be kept below 3 % so that degradation by
range of future spacecraft applications. Nitinol alloys are permanent setting will be avoided (Luciano and Galet 1995).
nonmagnetic, have a good corrosion resistance, and are They withstand 10 to the power 5 times of deformation with
thought to be compatible with hydrazine fuel so they could a 2 % strain and 10 to the power 7 times of deformation with
200 4 Spacecraft ManufacturingFailure Prevention
a 0.5 % strain according to Japanese sources. A similar deployment mechanisms for protective shrouds, seals for
non-explosive separation device is being developed in South solid rocket booster cases, and other release mechanisms for
Korea for nano-sats (Tak et al. 2009). boosters, rocket nozzle detachment at apogee, and the
A rotating arm that can be powered by short lengths of deployment of drogues and parachutes. They may also give
Nitinol has been built at the NASA Lewis Research Center spacecraft landing on distant planets an improved margin of
and installed on-board the Pathnder mission to Mars. This safety if heavy landings cause damage to Nitinol legs that
is the rst multicycle actuator used in a space application and might be repaired using some form of electrical heating.
has been described by Jenkins and Landis (1995). The arm is Commercial applications include the use of Nitinol in air
connected to a cover glass and when electrically heated it conditioning units, orthodontic wires, eyeglass frames, and
rotates to completely uncover a solar cell detector. When the even brassier frames. A power generation system that
current to the wire is shut off, the wire expands and a flat incorporates Nitinol as an energy conversion element to
spring pulls it back to the rest position and so returns the operate below 100 C has been designed and studied in
cover glass to protect the solar cell. Additional promising Japan (Anon 1995). This study concentrated on absorbing
space applications for shape-memory alloys have been power from low-temperature waste heat. It concluded that a
detailed by McDonald Schetky (1991), and these include piston-type heat engine could be used for every economical
4.18 Shape-Memory Alloys for Spacecraft Devices 201
Fig. 4.75 Use of shape memory alloys (SMA) in high reliability with plastic shrink-tubing. b Photograph of a Cryot connection
fastening applications produced by Raychem. They have been in between tubes having different diameters (after test programme).
service for more than 40 years. Nitinol (NiTi) alloy was later replaced c Micrograph of cross-section through a Cryot coupling made to join
by NiTiFe and NiTiNb alloys as the SMA. a Photograph of a a stainless steel tube to a titanium tube. Shrinkage of the Nitinol alloy
Cryocon; this is an electrical connector for individual wires and part has caused its internal knife edges to slightly deform the
cables. The gold-plated pin is mechanically secured into the left-side thin-walled titanium tube. Hardness values are shown (Dunn unpub-
socket by means of a shrunk Nitinol ring. Stranded wires are crimped lished report). d The extract beneath, courtesy of Raychem Corpora-
or soldered into the gold-plated pin and socket buckets, then covered tion, describes the assembly process
power generation. By replacing solar heating for warm waste varies with temperature and alloy composition. With certain
water, it should be feasible to design a similar heat engine Nitinol alloys, maximum damping occurs around room
for operation in space. temperature when the material is stressed such that it lies
Another property of shape-memory alloys concerns their within the pseudoelastic hysteresis loop. For Nitinol the time
damping capacity. This can be tailored to be very high and it delay of strain with respect to stress results in the dissipation
202 4 Spacecraft ManufacturingFailure Prevention
formed Al-5083 sheet has been selected to produce a pro- The rst European spacecraft propulsion tanks to have
tective cover for the sensitive instruments on board. This been produced by superplastic forming were made in Britain
will shield them from increased temperature and radiation as in 1975. They were designed for three-axis stabilized satel-
the spacecraft flies by Venus, on two occasions, during the lites requiring long-term storage of the monopropellant
rst stage of the mission. The formed Al-5083 cover is hydrazine. Expulsion of the fuel would be by means of an
chemically etched in six places to a thickness of only elastomeric diaphragm (ethylene-propylene-diene- mono-
0.2 mm in order to save weight. mer). The starting material for these tanks is Ti6Al4V sheet.
Fig. 4.79 a External and b internal views of Ti6Al4V superplastically fast fracture in Ti6Al4V [locations are 3 and 5 given on (b)]. d Optical
formed propellant tank (480 mm diameter) after burst test. The micrographs made complementary to SEM images shown in (c). Note
locations for SEM fracture surface and metallographic inspection are necking down of tank wall at origin of burst and presence of secondary
indicated. No corrosive sites are seen on the internal surface after subsurface cracks. At tail end of crack the superplastically formed
two years in contact with hydrazine. c Pairs of SEM fractographs made tank has a thicker wall section, but local necking is also apparent (pairs
at 60 and 660 magnications showing the origin of the burst crack have identical magnications)
and the tail end of the crack. The ductile morphology is typical of a
4.20 Superplastic Forming and Diffusion Bonding of Metals 205
Two hemispheres are made initially by superplastic forming. The burst tank was cut in half just outside the girth weld and
This is done at a temperature just below 1000 C, under a submitted to metallurgical assessments, as shown in
pressure produced by an inert gas that presses and forms the Fig. 4.79ad.
sheet into a hemispherical mould. After cooling, the hemi- At the present time superplastic forming is being per-
spheres are machined at their circumferential edges and formed in Germany to produce hemispheres that are welded
surface etched. Etching is necessary to remove any hard, together as propellant tanks for the Ariane-5 launch vehicle.
embrittling alpha-case that might have formed when the hot The tank design and rationale for selecting this method of
alloy comes in contact with traces of air in the inert gas (see production against conventional forging and machining are
Sect. 5.18.2 for etching method). The hemispheres are then presented in Fig. 4.80, based on the paper by Muller-
TIG welded together at the equatorial (girth) location. As Wiesner et al. (1994). Each launcher incorporates two
part of a qualication test programme, one post-life test propellant tanks. These store the hydrazine fuel that is fed to
culminated with a burst test. The qualication burst stress six 400 N thrusters arranged in two groups of three. The
was quoted to be 612 MPa whereas the actual burst pressure thrusters deliver the forces and torque designed to control
was 706 MPa, providing an acceptable result with a margin the roll and three-axis stability during the ballistic phase of
of 15 %. The fracture started just outside the boss weld in the the launch; additional rings are provided for later accel-
tank shell (coinciding with the thinnest section of tank wall). eration and de-spinning operations. The superplastic
Fig. 4.82 Two examples of creep-resistant structural panels made from discernible. When held at 600 C for 48 h, samples have a hot tensile
high-temperature titanium alloys by superplastic forming and diffusion strength of 600 MPa. At room temperature the tensile strength is above
bonding in a single operation. The photographs show thin test-pieces cut 1000 MPa. After 5 days at 600 C, oxidation is noted to reduce the room
from large panels. In microsection, the diffusion bond lines are not temperature strength to 750 MPa (courtesy of Dornier Space Systems)
4.21 Cleaning of Mechanical Parts destruction of the stratospheric ozone layer, international
agreements (in particular the Montreal Protocol of 1987) and
4.21.1 General Background government regulations should see the phase-out of ODCs in
developed countries by the year 2000. In the early 1990s it
This section concerns general cleaning processes associated became clear that the deadlines set by the Montreal Protocol
with spacecraft mechanical systems. More specic cleaning were inadequate and far greater damage was occurring to the
processes are addressed elsewhere. Section 4.5.4 describes ozone layer than had been earlier predicted. Consequently,
the cleaning and passivation of steels, and Sect. 6.9.7 amendments were later agreed in Copenhagen and then
highlights some means of cleaning electronic systems. Brussels (in December 1992). The European Community
The space industry has previously relied very heavily on [now European Union (EU)] agreement in Brussels ruled
the use of ozone-depleting compounds (ODCs), but with the that the two major ODCs would be phased out as follows:
208 4 Spacecraft ManufacturingFailure Prevention
Chlorofluorocarbons (CFCs) would be cut by 85 % by the on other volatile organic compounds. In an important recent
end of 1993, and 100 % cut by the end of 1994 paper the actions taken by ve of the largest US aerospace
1,1,1-trichloroethane would be cut by 50 % at the end of organizations in order to replace the offending cleaning
1994, and 100 % cut by the end of 1995. agents are described (Patterson and Mykytiuk 1996). More
than 30 other major companies and US government agencies
It would appear from visits to European space companies also detail their cleaning applications, what was used as
during 1996 that the majority (particularly those in Scandi- cleaners and the new, alternative solvent system. It appears
navia) have completely abandoned the use of any ODCs. that 80 % of the industry that was using 1,1,1-trichloroethane
Some companies have stockpiled trichlorotrifluoroethane (TCA) and CFCs has now turned to aqueous cleaning sys-
solvents, in particular CFC-113. These companies were tems. However, many larger companies are hindered from
concerned that where mechanisms had been fabricated prior changing to water-based cleaning because of the problem of
to the 1990s with processes incorporating the use of metallic corrosion and the ability to heat the water sufciently
CFC-113 and then space-qualied, it would be difcult, or to ensure effective cleaning.
impossible to use drop-in non-ODC solvent replacements. The Patterson and Mykytiuk (1996) paper relates how
Fourteen of the main European companies associated large aerospace companies have evaluated their proposed
with the fabrication and assembly of spacecraft mechanisms new cleaners before they progress into pilot plant evaluation.
were questioned about their present preferred cleaning sol- The individual piece-parts of units can consist of a variety of
vents. In this purposely very general survey it was impos- metal alloys, such as nickel, cobalt, stainless steel, titanium,
sible to assess the actual volume or throughput of hardware aluminium, and low-alloy steels, which support contaminants
along the individual cleaning lines referred to by the dif- including machining and tapping oils, hydraulic fluids,
ferent companies. However, the main observations were that drawing oils, adhesive tape residues, and dye-penetrant
until 1990, all companies had utilized CFC-113 as the fluids. A logical sequence of events is to make a preliminary
standard cleaning medium and were in possession of costly review of all cleaning media on the market from a literature
capital equipment designed to vapour degrease, ultrasoni- search, check material safety data, and request information
cally clean, or high-pressure spray clean spacecraft from health and safety organizations. Cleaners should not
piece-parts using ODCs. During the early 1900s a large contain chromate compounds or sulphides. They must have a
number of chemical suppliers were offering non-ODC low volatile organic compound (VOC) content, be noncar-
products, but very little was ever published in the open lit- cinogenic, and have a low toxicity together with an inoffen-
erature by independent authorities to actually quantify the sive odour. Next, samples of the cleaners will be subjected to
suitability of these products. Very little technical cooperation compatibility testing with the metals. Nickel test-pieces will
occurred between these companies, as they are frequently in be exposed to the cleaner to check for sulphur embrittlement.
open competition with each other, and consequently, no All metals will be tested for their resistance to general surface
joint-sponsored activities to nd a common product corrosion and stress-corrosion cracking in the cleaner.
replacement to CFC-113 were ever established. This can be Finally, tests will be made for rinsability of the cleaning
observed from the ndings into the present-day cleaning solution and, separately, checks will be made to ensure that
systems used by these companiesthey are very diversied dried precipitates can be removed.
and are based on the following classes of solvents: 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
Hydrocarbons 3
suppliers, and avoid internal bottlenecks in the company
Halogenated solvents, including 6
(Mertens and Gessford 1995). The following are also rec-
(HCFCs)
ommended: not to wait for drop-in alternatives, as they may
Aqueous and semi-aqueous 13
not exist for any one product line, seek expert advice, pay
Alcohol and acetone 9
attention to details (especially what happens to liquid waste),
Exotics 2 (both plasma systems)
do not expect perfection on the rst run of the new system
The HCFCs, less-damaging to the ozone layer than CFCs, (keep the old system running until the new one has been
are not currently regulated by the EU but are expected to be tested and is operating properly), and ensure that training
phased out by 2015. However, apart from HCFC-141b, it is and maintenance programmes are documented.
known that many HCFCs have poor solvent properties and In the following sections, spacecraft mechanism parts will
they have not all been submitted to toxicological testing, so be considered with respect to their surface cleanliness and
this has also curtailed their use. surface condition. Space systems, instruments, and equip-
The aerospace industry in the USA is keenly aware of the ments include a vast range of items with mechanisms that
phase-out schedule for ODCs and the tightening regulations include:
4.21 Cleaning of Mechanical Parts 209
occurs in high humidity. Whenever possible, emphasis should The importance of performing materials compatibility
be placed on the prevention of particle deposition rather than testing with potential solvents at an early stage in a space
counting on their subsequent removal. Careful handling of programme cannot be overstated. Some companies offer
space equipment is essential, storage must be in control training for solvent cleaning, material compatibility testing
environments (sealed bags, containers, inert atmospheres, and test kits, such as MaxicheckTM for process optimization.
etc.), and the tasks of assembly and integration must be An overall company cleaning plan should be devised by
undertaken in a specially designated cleanroom. 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
4.21.3 Cleaning of Individual Parts identical materials as will be selected for the space hardware.
Where steels will be heat-treated to produce a certain hard-
The most effective cleaning processes are those performed ness, only that heat-treated condition must be tested. If one
on individual piece-parts. Whenever possible it is strongly chemical conversion coating is applied to the end item it is not
recommended that mechanisms are completely disassembled appropriate to test a similar coating as there can be signicant
and that the individual parts having the same chemical differences in the surface compositionthe acceptance of test
composition are placed into different glass containers. It may results by similarity is often inappropriate.
be necessary to use different solvents for the different groups The author has been involved with some compatibility
of materials. Certainly special thought must be given to the testing, the results of which can be used to illustrate some
hardened steels from bearing assemblies, as they may rust; to important ndings. The common solvents utilized by Euro-
the magnesium housings, as they may produce copious pean space mechanisms companies were listed in
volumes of corrosion product, and to all the organic mate- Sect. 4.21.1. A limited number of tests have been made on:
rials, as these can swell and become embrittled by many of
Hydrocarbons
the common solvents. Breaking the mechanism down into
All of the major petrochemical companies provide
individual parts enables the cleaning media to contact the
benzene-based solvents for cleaning. The metals selected for
entire surface, there is no danger of galvanic corrosion, and
spacecraft parts manufacturing are generally compatible with
less likelihood of crevice corrosion. Additionally, the sur-
pure benzene. Some rusting of ferritic steels has been seen,
faces can usually be easily dried and then visually inspected
but this was possibly due to a slight water content in the
for residual contaminants such as drying stains, grease,
hydrocarbon. An exotic cleaning method using distillable
particles, and corrosion products.
hydrocarbons is shown in Fig. 4.85.
The method and equipment used for the cleaning opera-
Some plastics and most synthetic rubbers swell and dis-
tion will need to be assessed by making tests on represen-
solve in benzene.
tative samples. Demonstrations are possible at most
All hydrocarbon solvents will present some re risk.
suppliers of automated equipment. Manual cleaning using
Some are highly flammable (having a flash-point less than
brushes, possibly with ultrasonic enhancement, is frequently
32 C) and are subject to specic controls. Most proprietary
made when the number of parts is few.
formulations on offer have flash-points of around 45 C or
The cleaning process may be performed in different parts
above. All need to be kept in re-resisting stores. Particular
of the factory. Cutting oils may be cleaned from parts in the
care must be taken not to spray flammable solvents onto hot
machine shop. Possibly the approach to cleaning in this area
or electrically live equipments.
may be: detergent spray, water rinse, and alcohol drying.
Special areas may be designated for the ne machining of
Halogenated solvents
beryllium parts due to health and safety regulations. Here,
An enormous number of cleaning products and associated
the beryllium parts may be ultrasonically cleaned in alcohol
trade names exist on the market.
and blow-dried (for this material see also Fig. 5.15 and
All metals need to be assessed. Two of the more common
Sect. 5.7 as particle detection using Millipore lters during
solvents are trichloroethylene and dichloromethane (also
cleaning may be desirable, together with a nal chemical
called methylene chloride). Long-term storage tests
etching of the surface). Parts may be withdrawn from stores,
(30 days) have shown that type 304 stainless steel, alu-
where they may have been packaged in corrosion-resisting
minium alloy 2219, and titanium alloy Ti6A14V are com-
oils and greases. On entering the cleanroom, halogenated
patible with both the liquid and the vapour phases of these
solvents may be specied for cleaning off these oils and
solvents. Low-alloy steel (AISI 4140) corrodes in both liquid
greases before these parts are further processed and assem-
solvents and dichloromethane produces voluminous oxides
bled into a mechanism. Additional storage could be in sealed
on the magnesium alloy MgZE41A, as shown in Fig. 4.83.
containers lled with dry nitrogen gas.
4.21 Cleaning of Mechanical Parts 211
In order to improve quality, sustainability and safety, most units are fully automated and include high pressure spray
aerospace companies have researched alternative cleaning washing, vapour degreasing and nally, drying under vacuum.
methods to the open top degreasers. In the past operators had to Non-metals such as sintered PTFE tape and aluminized
wear breathing apparatus and they were using signicant Teflon were hardly affected by long-term exposure to tri-
volumes of chlorinated solvents which lead to worryingly high chloroethylene or to dichloromethane. However, epoxy resin
emissions of Volatile Organic Compounds (VOCs). Nowa- (Scotchweld EC 2216) sheets and silicone rubbers (e.g.
days cleaning solvents such as trichloroethylene are used in DC93-500) generally show appreciable weight gains and
modern, closed loop systemsthese ensure VOC swelling, together with a reduction in mechanical properties.
emission-free cleaning. Sizes of such degreasing machines Polyurethane sheet (Solithane 113) disintegrated catas-
range from small units, up to about 3 4 m, and they are able to trophically both during and after exposure to both solvents
clean a mass of about 700 kg of aluminium in one cycle. Many (see Fig. 4.84).
These groups of solvents are not normally recommended warm circulating air and achieves a spot-free surface. Ace-
for in situ cleaning and should not be used in this way tone is seldom used because it can slightly attack many
without adequate precautions to control exposure to the materials and, on drying, leaves some stain marks. These low
operator. They are all volatile (although the degree varies) flash-point solvents require to be housed in properly designed
and access to the degreaser and supplies should be controlled equipment; they should be carefully handled and may need to
to prevent unauthorized use. be regulated by health and safety inspectorates.
Pure IPA does not attack any of the metals employed in
Aqueous and semi-aqueous spacecraft applications. Most pure alcohol contains about
Aqueous cleaning systems are frequently used for cleaning 0.1 % water and this water content is sufcient to passivate
mechanisms in Europe. They rely on certain important aluminium and other alloys without producing corrosion
components: the surfactants (such as emulsiers or splitters) problems. When completely free from water, and at its
and their concentration, the water quality (such as whether it boiling point (for IPA this is 82 C), very pure alcohol will
is de-ionized, ltered, or puried by reverse osmosis) and the attack the surface of aluminium.
method of drying (e.g. dry, hot air, or air knives or various Non-metals are generally immune to any room tempera-
methods followed by a nal rinse in alcohol and air drying). ture reaction with IPA. Owing to its low surface tension, it
Various detergents may be added to the water. reduces spotting on optical surfaces. At elevated tempera-
Ferritic and martensitic steels may suffer from surface tures many plastics and rubbers will swell when immersed in
corrosion during cleaning (particularly 440C and 52100 alcohol or acetone.
bearing steels), and as mentioned earlier, the presence of a
strong detergent can activate surfaces and accelerate corro- Exotics
sion. Most other metals also will tarnish (oxidize) and cor- Plasma cleaning technology is occasionally used where
rode in aerated water. All metallic materials must be pristine surfaces are needed prior to surface treatment and
evaluated in the aqueous cleaning line, with detail given to bonding processes. The method produces little or no waste
the duration of the process, its temperature, and the drying products. Special equipment is required. Plasma is produced
method. Part designs should be inspected to ensure than no by applying a high-frequency AC voltage to gases, or a
blind overlapping surfaces exist, such as capillary gaps, mixture of gases, under partial vacuum conditions (below
which could entrap these solvents, so causing corrosion 1 Torr) inside a processing chamber.
during later storage periods. Plasma cleaning can be applied to the majority of metals
Plastics and rubbers are usually unaffected by immersion and alloys whenever organic contaminants need to be
into water. Water can permeate through most polymers, and removed from sensitive surfaces (Ries 1996; Ward 1996;
this must be considered when the polymer is present as a thin Engemann 1996). Enhanced cleaning rates are being eval-
layer on the surface of a corrodible metal part. Baking, or uated; they produce spot-free surfaces and improve the
post-conditioning the part under partial vacuum, is some- bonding properties of metals.
times used to ensure that any absorbed water is removed Precision optical glasses can be plasma cleaned. This
from spacecraft organic materials. technique may also be useful for removing the enamel
An interesting report that documents the actions taken by insulation from the surfaces of ne copper wires prior to
a large engineering company in replacing the cleaning sol- soldering processes, or to plasma etch machine parts made of
vents CFC-113 and trichloroethylene with aqueous solutions materials such as Delrin and Teflon into complex patterns.
has been prepared for the US Environmental Protection Many other exotic cleaning methods are nding appli-
Agency (EPA 1994). These company-cleaning changes have cations as replacements to ODS cleaning technologies.
not only complied with the environmental goals of the Chambers that can be lled with solvent and then pressur-
Montreal Protocol, but they resulted in improved cleaning at ized whilst the part is rotated and subjected to ultrasonic
dramatically reduced costs. All cleaning steps were identi- vibrations have been developed. Figure 4.85 illustrates one
ed and converted under the supervision of a team of per- piece of equipment where the solvent cleaning can be per-
sons involving the company vice-president, the production formed under either pressure or, in the vapour phase, partial
manager, the environmental ofcer, quality personnel, the vacuum. The solvent is then drained from the chamber and
metallurgist and chemist, and the purchasing department. the part then dried under partial vacuum. Another method,
which seems to be nding new uses in the USA, incorpo-
Alcohol and acetone rates supercritical fluids (SCFs). Carbon dioxide is com-
Alcohol (isopropyl alcoholIPA) is a major cleaning sol- monly used at very high pressures (above 73 atmospheres)
vent. It is usually used in conjunction with de-ionized water and at 31 C. Here, the highly compressed gas becomes a
or an azeotrope of cyclohexane. IPA is often used as a rinsing fluid, but it has the same diffusivity of gaseous carbon
agent following aqueous cleaning. It can be easily dried in 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 partial vacuum of about 5 mbars. The solvent is continuously ltered
industry for cleaning mechanical parts. A closed treatment chamber for particles larger than 1 m and regenerated in a vacuum still.
contains the parts. The cleaning fluid can be one of several distillable A modied system is suitable for cleaning PCB assemblies and
hydrocarbons with flash points between 55 and 100 C. The solvent removing rosin flux and solder pastes (see also Sect. 6.9). (By
initially sprays the parts (which can be rotated) at high pressure. They permission of Branson Ultrasonics B.V.)
are then vapour rinsed in a partial vacuum at 80 C and nally dried in a
holes, and capillary gaps for fluid retention and is con- silicone oil (typical for space lubricants), 2 % fluorescent
structed from typical spacecraft alloys. The blind holes have powder (Fluoflux, used to represent fluorescent
a diameter of 1 mm and a depth of 4 mm, and were drilled crack-detection particles), and 4 % oil-soluble analytical
into the steel base plate to represent pockets that could trap tracer marker (Scandium-46 tracer element, in standard
contaminants. One hundred test-pieces were manufactured to solution of 100 g Sc/g). A standard technique was devel-
high dimensional and surface nish tolerances. They were oped to contaminate the individual piece-parts. They were
individually numbered. An articial contaminant was con- assembled within the slurry (homogeneous suspension) of
cocted using a base consisting of cutting oil (Blasocut 200 contaminants by means of the bolt and nut, and then tight-
CF can be emulsied with water) which was mixed with 2 % ened to a dened torque of 8 Nm. The lock wire and lead
4.21 Cleaning of Mechanical Parts 215
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 4 Spacecraft ManufacturingFailure Prevention
The ndings of the round-robin cleaning study were 4.21.5 Maintenance of Cleanliness
many and included the following:
The title of this chapter refers to metallography as it is
all participating companies utilized different cleaning applied to spacecraft production. The numerous sections
methods (25) have promoted metallography as a technique that can use-
from the rst view, all returned test-pieces appeared fully control and maintain a high level of quality during the
visually clean manufacture of spacecraft mechanisms and structures. Sim-
after disassembly, residual contaminants could be clearly ilar aspects, but directed only to spacecraft electrical sys-
seen under UV-light tems, will be detailed in the following chapter. The topic of
the amounts of residual contaminants vary by 34 orders cleaning has also been covered. Once it has been decided
of magnitude (reflecting the great differences of ef- that a particular piece-part must be cleaned (probably
ciency between the cleaning techniques) immediately after a dirty production process) there will be
none of the steel parts or titanium parts had corroded a need to check that the cleaning procedure is adequate for
aqueous alkaline cleaners and also plasma cleaning pro- the part and that all subsequent handling operations are
duced some surface corrosion on the aluminium alloy as performed in a correct manner (possibly using lint-free
seen in Fig. 4.89 (other cleaning methods had no effect) gloves) within the connement of an environmentally con-
the molybdenum disulphide surface lubrication layer on trolled room. Company handling procedures and facilities
the titanium bolt was unaffected by al the cleaning will need to be scrutinized by some form of audit. Fig-
methods except that performed by plasma techniques ures 4.91 and 4.92 show the production and test facilities for
the cleaning process (e.g. equipment design, time of large cryogenic tanks that are installed onto conventional
cleaning, use of pre-cleaning steps, use of ultrasonics, launchers. Special attention is paid to the welding techniques
etc.) is more important than the selected solvent when these large, thin, curved metal shells are processed and
most of the non-ODS cleaning techniques are as good as to the manner by which they are handled and later submitted
the reference CFC-113 methods, particularly when to a combination of mechanical and thermal loads during
ultrasonic agitation is applied (Fig. 4.90). performance testing (Pradier and Dosio 1996).
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 4 Spacecraft ManufacturingFailure Prevention
The gures illustrate facilities that are not classied as Airborne particles are assessed in units of total particle count
cleanrooms. These are locations in the factories where and distribution of particle size. Much must be known about
good housekeeping and a tidy appearance is necessary the dirt: whether it is a soluble or insoluble contaminant.
during spacecraft fabrication and test. Dirty, untidy facil- Ionic contamination lends itself to electrical measurements
ities are inherently detrimental to the end-product. Dirt can these are based on the change in conductivity (or resistivity)
be anything that settles on the part from air, such as dust, of a polar solvent after extraction of the dirt from the parts
lint, or other matter. Numerous examples have been surface. It may be trapped in hard-to-reach places where
shown in this book where particles, ngerprints, oil, and dynamic cleaning methods incorporating the use of ultrasonic
the like, have promoted or caused the failure of expensive cleaning may be required.
parts. The cleanliness of all piece-parts will steadily deteriorate
The maintenance of cleanliness in a spacecraft production after processing. Unless the part or assembly is sealed,
facility will necessitate that different tests are performed contamination build-up continues until the time of launch.
these are dictated by the nature of the contaminating dirt. 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 method was insufciently accurate (requiring at least a
responsibility of the manufacturer to know or study what the 2.4 102 m thick silicone contamination on the surface).
limits are for his own assemblies. 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
4.21.6 Cleaning of Silicone Contamination detecting a layer of only 5 103 mmaking the contact
angle method at least 5 times more sensitive than the FTIR
Section 5.15 will describe individual case studies where method). However, the disadvantage is the substantially
silicones have contaminated space hardware, and offer higher effort needed for surface angle measurements com-
advice about how such contaminants may be removed. pared to FTIR measurements.
Silicone products used for spacecraft range from oils and The second part of the Liedke study was to contaminate
greases to cured products as used for sealing materials, ultra-clean glass slides with a range of common spacecraft
adhesives, paints, or conformal coatings. They can offer silicone products. Each product was applied to an individual
excellent properties, e.g. low vapour pressure and low sur- slide by smearing, using the edge of another slide, so as to
face tension (oil lubricants), high thermal conductivity achieve a very thin coverage. The contaminated test slide
(thermally conductive vacuum greases), chemical inertness, was then assessed for cleanability by exposure to a range
good dielectric properties, and ductility (conformal coatings of commercially available cleaning fluids. An attempt was
and paints). Silicones on the other hand can act as major made to standardise the cleaning methodlens tissue was
sources of contamination if they are incorrectly handled or immersed into the cleaning fluid, the tissue was then stroked
used. Their low surface tension will cause oils to creep over across the contaminated slide ve times. After this cleaning
clean surfaces, spreading contaminations into gaps and other process the slide was dried and submitted to the contact
areas that are almost impossible to clean. Under vacuum, angle measurement test. The following contaminants were
such trapped reservoirs can act as a long-term source for applied to slides; the cleaning agents are also listed:
further contamination. Therefore, a thorough removal before The following silicone contaminants, representing typi-
launch is mandatory. cal species that may come into contact with spacecraft, were
A method for assessing the efciency of cleaning silicone been selected:
contaminated surfaces has been developed based on contact
angle measurements (Liedke 2011). Glass slides deliberately Poly-dimethylsiloxane oil (Dow Corning DC 200)
contaminated with a standard silicon oil solution were used Pentaphenyl-trimethyl-trisiloxane oil (Dow Corning DC
together with FTIR spectroscopy as a reference. This solu- 705)
tion was: 0.5 g DC 200 silicone oil in 49.5 g dichlor- Mould release agent (330 LV Silicone Emulsion from
omethane (Merck, p.a. grade). It was found that the FTIR Basildon Chemicals)
Cured silicone resin (DC 93-500 from Dow Corning) now back-side coated with 5 m thick diamond lms giving
applied by rubbing on substrate better than 400 W/mK conduction at room temperature.
Cured silicone resin (DC 93-500 from Dow Corning) Similar lms, applied to the circuit-side of IC are used to
applied as an extract in tetrachloroethylene provide space electronics with partial radiation hardening.
Thermally conductive vacuum grease (DC 340 from Nano-tubes and nano- wires based on graphine are also
Dow Corning). proposed for thermal transport, based on their inherently
high thermal conductivity (Schelling et al. 2005).
Cleaning agents (most are dangerous substances and Metal-graphite foam and sheets are manufactured commer-
should be only used by trained professionals, some may have cially for heat-sinking. The graphite bre Thermal Strap
REACH restrictions): assemblies touched on in Sect. 2.5c) are convenient,
light-weight and have been supplied for US spacecraft
Methylene chloride CH2Cl2, (from Merck) including the Orion multi-purpose crew vehicle. The
Toluene C6H5CH3, (from Merck) assemblies use GraFlexTM bundles (made from graphite
n-Hexane C6H14 (from Merck) bre with a thermal conductivity of about 800 W/mK).
Fluxclene, (manufactured by Electrolube) Overall, the special thermal link assemblies can be consid-
Silgest SD0001, (distributed by Polymer Systems Tech- ered to have a conductance of up to 5 W/K and are used in
nology Ltd.) near-cryogenic systems, satellite electronic boxes and power
Ortimex Alkaline Cleaner, (distributed by H.C. Stehelin) electronic systems, as seen in Fig. 4.93.
SWAS (Safewash Super), (manufactured by Electrolube) The high conductivity of metal substrates makes them
Vericlean defluxer (from Microcare) attractive for thermal management of electronic devices.
Vertrel CCA (DuPont, distributed by Microcare) However, in many applications it is necessary to provide
Vertrel CMS (DuPont, distributed by Microcare) electrical insulation to the thermal plane to prevent shorts.
Dowclene 1601 (based on 1-Methoxy-2-propanol) from While this is readily accomplished by a polymer dielectric,
(Dow Chemicals) the relatively poor thermal conductivity of even
Dowclene PX-16S (from Dow Corning) ceramic-lled resins represents a signicant barrier in the
thermal path, even when the layers are relatively thin. One
The test results are somewhat subjective because not all solution to this problem is to use a pure ceramic dielectric.
process variants were strictly controlled (time between At least one company has developed a process for con-
contamination and cleaning, room temperature variations verting the surface of aluminium to nano-grained alumina
and cleaning tissue pressure was operator dependant), but (Shashov 2013). Remarkably, this is done using an elec-
the author considers the compilation presented in Table 4.15 trochemical process since the resulting alumina is electri-
should be a good starting point for further evaluations. 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-
4.22 Novel Thermal Management tionally thin layer of dielectric possessing moderate thermal
Materials conductivity (7 W/mK), together with a very low interface
resistance between the aluminium and the ceramic means
Thermal management systems are used to maintain space- that this material has the lowest through-thickness thermal
craft hardware within acceptable temperature ranges. Exist- resistance of insulated metal substrates available commer-
ing passive and active systems that achieve this, are cially at the time of writing. 3-dimensional heat sinks made
described in other Sections. New developments in materials from aluminium with nano-ceramic dielectric and support-
will ensure that heat flow away from hot spots, or heat flow ing copper circuits, as illustrated in Fig. 4.94, are available
towards cold spots (such as mechanisms located in the and should have several space applications. By growing
vicinity of cryogenic tanks) is via passive materials rather alumina on both sides and capping with copper, complete
than heat pipes, radiators and resistance heaters. In the metal core printed circuit boards (MCPCB) can be
absence of an air stream, thermal management within space fabricated.
systems is best made by the incorporation of Diamond has a heat conductivity four times that of copper.
high-conductivity materials. Materials under development, Measurements range from 800 to 2000 W/mK for industrially
or being qualied, include composite materials and produced diamond, and 420 W/mK for copper. The larger the
heat-spreader substrates manufactured by novel processes. diamond grain size, the greater the thermal conductivity.
Copper is the most common spreader, but more efcient Composites of diamond grit embedded in a copper, or cop-
materials will incorporate diamonds, silicon carbide and per alloy matrix have been produced in small volumes for use
aluminium nitride. Electronic circuits based on silicon are as high current heat sinks (power devices producing
4.22 Novel Thermal Management Materials 221
Fig. 4.93 a Graphite bre thermal strap assemblies are light-weight, operate reliably under severe vibration. These straps provide passive,
and b have found a number of advanced thermal management high performance heat transfer in vacuum from hot zones (e.g.
applications within spacecraft. On a thermal conductivity/mass basis travelling wave tubes) to radiator plates. Courtesy of Technology
they are more efcient than copper or aluminium and, c have sufcient Applications Inc.
flexibility to accommodate installation and alignment tolerances; they
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 223
kW cm2). Here, the target was to achieve a flat composite materials are generally limited to copper, aluminium, tita-
sheet having 500 W/mK conductivity that could be utilized in nium and some cermets (which need to be cold sprayed with
the construction of multi-layer printed circuit boards, or a ductile binder). This is an alternative to electroplating and
alternatively, special heat sinks. Complex powder metallurgy hot, thermal spraying. CGDS can be directly applied to the
processes are needed to produce such compositesalthough surfaces of metallic-, ceramic- and organic-nished products.
it should be noted that the cutting tools industry has Space hardware applications will include the deposition of
employed brazed diamond cutting and grinding tools using metallization layers onto electrical components and electrical
active metal silver-copper and nickel-based brazing alloys for grounding points on organic substrates. Specic examples
many years (Tillmann and Boretius 2008). One has to con- include aluminium metallization of polymer composite
sider thermal expansion mismatch between diamond crystal structures and silver deposited as a solid lubricant for use in
size and the particular alloy matrix. Of greatest importance is space. The deposits are free of the oxide lms and porosity
the need for additives that can bond the matrix to the usually seen in more conventionally thermal sprayed coat-
almost-inert diamond facets; here, chromium, boron, zirco- ings, and give improvement in properties such as electrical
nium or titanium (all carbide formers) can be considered. The and thermal conductivity. The lack of heat input into the
microstructure of a metal matrix composite under develop- substrate during coating application makes the CGDS pro-
ment, containing 30 % diamond grit, is seen in Fig. 4.95. cess of interest for depositing metal coatings onto polymer
Consolidation of the powder mix and mechanical treatments substrates such as antennae face skins and structural parts.
are proprietary, but some released data for the European The process is explained in Fig. 4.96.
Project ExtreMat (Schubert et al. 2008) concluded their most The CGDS facility at the Institute of Manufacturing,
promising composite to be a copper alloy-diamond possessing University of Cambridge, together with TWI have success-
a thermal conductivity of 640 W/mK combined with a CTE of fully demonstrated the process for applying aluminium onto
about 11 106 K1 from room temperature to 200 C. antenna face-skin material samples made of PEEK com-
posite (Sturgeon and Dunn 2006). The spraying was per-
formed within an enclosed chamber built on a bespoke CNC
4.23 Cold Sprayed Coatings controlled 3-axis system, with the nozzle assembly attached
to a z-axis actuator and the target substrates xed to the
This process is termed cold gas dynamic spraying (CGDS). translatable base. The chamber protected the surrounding
Only a limited number of materials can be presently cold environment from any loose powder and facilitated the
sprayed onto substrates but this deposition process has recovery of exhaust process gas that was then recycled.
enormous potential for specic spacecraft application. The Helium was chosen to be the principle process gas for this
system because of the greater deposition efciencies and these two temperatures. A total of 500 such cycles were
material integrity that can be achieved compared to nitrogen, undertaken with the samples then removed and examined.
even without gas heating. However, there is a cost penalty in Such thermal cycling is often used for the verication of
using helium because of its high cost per unit volume that is spacecraft electronic components. The samples were then
compounded by the increased flow rates because of the subjected to visual inspection, tape testing to check for
higher gas velocities. Consequently, helium gas recycling is coating delamination and microsectioning for coating mor-
required to make the process commercially viable. phology and thickness. The sample ndings are shown in
After coating, the samples were subjected to environ- Fig. 4.97.
mental tests that consisted of holding in a high humidity air The work showed that it is feasible to deposit a CGDS
atmosphere at elevated temperature followed by thermal aluminium coating of typically 0.5 mm directly onto a chop-
cycling. The exposure to high humidity was made in a ped carbon bre reinforced PEEK composite. The processes
chamber with the humidity level set at 85 %RH and the did required careful selection of nozzle design and spraying
temperature set to 85 C. After 14 days (336 h) the coated conditions. However the range of carbon bre reinforced
samples were removed and examined visually. The same polymer composites that can be coated with metal using
coated samples were then put through a 55 to +100 C CGDS will only be known after more extensive evaluations.
thermal cycle in air, cycling at a rate of 10 C/min between 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.
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 very high local pressures (believed to be several GPa).
deposition rate and thickness limit, applied voltage, elec- Although very high local surface heating appears to make
trolyte type, electrolyte life, phase formation, process speed the processed surface red, or white hot, only surface oxi-
and scalability. The PEO process is also of signicant dation occurs and there are no metallurgical changes to the
interest as it can offer improved surface properties to mag- substrate. Metallography has conrmed that treated compo-
nesium and aluminium alloys and may be a suitable nents only have modied surfaces and their mechanical
replacement for conventional acid based processes such as properties are not changed. Inspection of processed items
DOW and other chemical processes that contain hexavalent show no heat affected zones or dimensional distortion. The
chromium. It has been used as a drop-in replacement for the overall process temperature and that of the component are
anodisation of aluminium alloys where the anodising elec- typically in the range of 1230 C, so modication of any
trolytes are based on chromic acid and sulphuric acid. heat treatment condition is not possible. Sophisticated lab-
The Keronite is an advanced surface technology based oratory examinations of resultant oxide surfaces using
on plasma electrolytic oxidation (PEO). It is an environ- crystallography show these PEO nishes to contain some
mentally friendly and safe process for the preparation of elemental co-deposition from the electrolyte, creating a
corrosion and wear resistant nishes (Shrestha et al. 2003). ceramic layer that consists of both crystalline and amorphous
These nishes comprise of ceramic layers on aluminium phases of alumina.
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 4.24.2 Characteristics of PEO Coatings
of the PEO coating can be changed from grey to black; this
is a factor of the alloys compositions. Generally with high An SEM image of an as-prepared surface of the Keronite
copper content, such as AA2024 or AA2219 the nal colour layer is shown in Fig. 4.99a which depicts an abundance of
is black. The Keronite process employs a low concentration spherical and irregular shaped particles-like structures that
alkaline-neutral solution that is free of heavy metals and were formed during the PEO process. An as-prepared sur-
contains a salt concentration of less than 4 %. The com- face of the hard-anodised layer is shown in Fig. 4.99b which
monly used sulphuric acid hard anodising chemistry is far shows several voids and micro-porosity on the coating sur-
different from that of Keronite and comprises of between 10 face. In addition, the surface clearly shows parallel lines that
and 20 vol. % of acid concentration. Some parameters for form vertical through thickness cracks. A typical cross sec-
the Keronite PEO and sulphuric acid hard anodising pro- tion of the Keronite layer on AA7075 alloy at the outer
cesses are given in Table 4.16. corner/edge is shown in Fig. 4.100a. The Keronite layer has
Coating of aluminium alloys using the Keronite PEO some micro-scale porosity as would be expected from the
process involves the creation of a plasma discharge around a PEO process but the coating can be considered relatively
component immersed in an electrolyte (Fig. 4.98). The dense. A typical cross section of the hard-anodised layer on
mechanism of oxide layer formation is complex, being a AA7075 alloy at the outer corner is shown in Fig. 4.100b
combination of oxide growth with subsequent fusing and which shows the presence of extensive V-shaped columnar
then re-crystallisation of the oxide lm. A small amount of cracks. These cracks are seen through the entire coating
the substrate is simultaneously dissolved into the electrolyte. thickness and extend down to the underlying substrate, and
At the time of Keronite PEO processing it appears to be very in several areas even expose the bare metal substrate. Some
aggressive due to extensive plasma discharges on the surface general characteristics of the Keronite coating on three dif-
of the piece-part; at microscopic level the process creates ferent aluminium alloys are given in Table 4.17.
226 4 Spacecraft ManufacturingFailure Prevention
(a)
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 whereas, the anodised coating displayed extensive cracking
was noted to have very low adhesion as shown in Fig. 4.101. and chipping of the coating surface in a fretting environment.
Actual adhesion values are: Keronite-coated-AA2219 against The differences in the damage mechanisms between the
a 52,100 steel ball (adhesion force 108 mN); and Keronite and anodised coating surfaces are shown in
NiCr-plated-AA7075 against an anodised AA7075 (adhesion Fig. 4.101. While the friction coefcient of Keronite to
force 242 mN). The aluminium alloy 7075 without a coating, Keronite has been reported relatively high about 0.50.6
to itself, resulted in extremely high adhesion (adhesion (Shrestha et al. 2003), this can be substantially reduced to an
7330 mNequivalent to cold-welding) after fretting. The extremely low value of about 0.04 by impregnating the
Keronite-coated-AA7075 surface against itself, displayed porosity (see Fig. 4.102) in the coating with solid lubricants
329 mN. Despite low adhesion values displayed by the e.g. MoS2. The Instituto de Astrofsica de Canarias (IAC) and
Keronite and hard anodised coatings against steel surfaces, Keronite have undertaken a co-operative research on cryo-
the Keronite coating shows no sign of surface damage, genic plain sliding bearings. Keronite + MoS2 composite
4.24 Advanced Plasma Electrolytic Oxidation Treatment for Aluminium, Magnesium and Titanium Alloys 227
coating has been developed as a tribological surface with a process capable of producing a charcoal black nish on most
low sliding friction coefcient, maximum resistance to wear, wrought and some cast aluminium alloys has recently been
minimum debris generation, vacuum compatibility, resis- developed an industrial scale. Unlike the process discussed
tance to thermal shock and thermal conductivity better than above, this new process converts the aluminium alloys into a
plastics. black nish and the resulting surface characteristics and
Keronite PEO coatings on copper-containing aluminium properties are not dependent on the substrate alloy compo-
alloys such as 2219 and 7075 can be prepared black and these sition (Shrestha et al. 2007). A typical SEM image of this new
have been extensively studied as thermal control coatings. black Keronite surface is shown in Fig. 4.102. This coating
Recently there has been growing interest by industries on surface has very low reflectance (and high absorptance) of
black nish coatings on other aluminium alloysa PEO about <0.1 % (Fig. 4.103). This reflectance value is less than
228
a
ASTM B117
Spacecraft ManufacturingFailure Prevention
4.24 Advanced Plasma Electrolytic Oxidation Treatment for Aluminium, Magnesium and Titanium Alloys 229
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. 4.103 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
4.25 Joining by Friction Stir efciency, no ller wire necessary and no gas shielding for
joining aluminium alloys. It can even be carried out under
4.25.1 Friction Stir Welding water.
FSWs have been utilized for the welding of spacecraft
Friction Stir Welding (FSW) was invented as recently as tanksthe welds being around the domes-to-cylindrical
1991 by Wayne Thomas at The Welding Institute parts. The rst space FSW was the Interstage Module of a
(TWI) Ltd., Cambridge and this technique for joining Delta-II launch vehicle successfully flown in 1999 (Kallee
metals has now been incorporated into the manufacture of 2001). It was used for welding the Space Shuttle External
aeroplane and spacecraft structures where high quality, tanks that stored liquid hydrogen and liquid oxygen (alloys
reproducible welds are a necessity. The FSW process is a being AA 2219-T87 and AA 2195-T8M4). It is used for
solid-state joining process capable of fabricating butt or lap welding the US Ares launch vehicle tanks and curved
joints in a wide range of materials and thicknesses (Kallee structural parts. In Europe, Fokker successfully demon-
2001). In aerospace the common alloys are based on alu- strated that FSW can be used to weld Ariane-5s motor
minium as these are light and, because the process does not thrust frame on the cryogenic main stage. This is an
involve the melting and softening of the alloy, the resultant unpressurised conical structure where the majority of parts
joints have little distortion and can maintain the mechanical are made from AA 7075 and are in the form of lap joints
properties of the parent plates. Also, due to the non-melting [Fig. 4.105 (Brooker 2001)].
of the joint, difcult-to-weld metal combinations can often Thales Alenia Space (TAS) in Turin have also designed,
be joined without the formation of damaging voids, brittle manufactured and tested integrally friction stir welded
intermetallics or distorted interfaces. Benets also include aluminium-lithium (AA2195) cryogenic tanks that are
no arcing or generation of fumes, no spatter, good envisaged for both the European Expendable- and the
232 4 Spacecraft ManufacturingFailure Prevention
Fig. 4.106 a AlLi cryogenic tank made by Thales Alenia Space weld pin during FSW of the Orion crew module at Michoud Assembly
Turin where the longitudinal cylinder weld and the dome-to-cylinder Facility in New Orleans; this FSW xture is the largest in the world
welds are by FSW (courtesy ESA-TAS). b Close-up of the rotating 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
Fig. 4.110 Macrostructure of tranverse cross-section made on FSW at the advancing side of the 14 mm diameter steel rotating tool. Here,
between oxygen free copper with high phosphorous content and the weld was made with the lowest heat input (Galvao 2011)
aluminium alloy AA 5083-H111. The harder Cu-DHP plate was placed
4.25.2 Friction Stud Welding electrical contacts, wear surfaces for mechanisms, producing
specic absorptance or reflective nishes for thermal control
As with friction stir welding, friction stud welding can be systems, and for enhancing solderability or brazability prior to
used to join difcult, and sometimes impossible, combina- assembly of parts. The main deposits are hard-chromium,
tions of metals pairs. The stud is selected as the harder of the copper, gold, lead, nickel, silver and tin-lead. The quality of an
two interfaces. Stainless steel studs have been used to join to electroplated layer will be tested, usually on in-line produced
aluminium alloy plates for electrical grounding applications. samples, or occasionally on the flight hardware.
Here a pin-like protrusion on the flat face of the stud is Non-destructive tests may include adhesion (tape testing such
rotated until the softer material heats up and becomes plas- as to ECSS-Q-ST-70-11), surface roughness, and brightness;
ticized. The pin penetrates into the softer plate and when the destructive tests will include thickness and uniformity by
shoulder of the stud contacts the plate a force is applied to metallography, corrosion-resistance, ductility by bending,
weld the surfaces together. There is no fusion at the inter- hardness and testing for porosity. Gold plating should receive
faces but a true metallurgical joint is formed and a high special attention as the purer the gold, the better the electrical
strength connection achieved. Many kinds of friction welds and thermal conductanceboth key material properties for
are based on this principal. For instance, two cylindrical bars spacecraft. However, pure gold has relatively poor wear
can be welded together by rotating one bar onto the end of resistance, so for surfaces that will be mechanically abraded,
the other barthe friction heating produces some material to such as electronic pins and sockets, pcb end-connectors, and
be expelled from the interface after which the relative motion electronic housings, it becomes essential to co-deposit very
is stopped and a higher forging force applied for effec- small quantities of other elements, such as cobalt or nickel
tively welding the interfaces. This joining process was dis- with the gold nish, to enhance wear resistance (and to some
cussed in Sect. 2.6.2 where it is proposed as a joining degree, avoid cold-welding).
technique for in-space assembly of hardware. 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
4.26 Selective Brush Electroplating electroplate selected areas of a piece-part. It is performed
locally, using a D.C. current through two cables. One cable
Electroplating is a well-established process employed for the is clamped to the work-piece (the negative, cathode con-
electrolytic deposition of coatings on spacecraft parts and nection seen in Fig. 4.111) and the anodic cable is plugged
sub-systems. Plating will be selected for both protective into a stylus that comprises of a carbon anode surrounded
(against corrosion) and functional purposes such as for sliding with an absorbent cotton pad, covered with a permeable
4.26 Selective Brush Electroplating 235
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. 4.113 Illustrations showing nickel being applied to components by adhesive tape testing, flatness by laser optical equipment, and
of the ISS by brush plating at NASA MSFC (module, nodes, laboratory surface roughness down to 16 microinches Ra (0.4 m Ra). Nickel
and airlock). During plating of the flight articles in-line samples were plating thickness was measured by microsectioning and beta backscat-
made for destructive testing. The nickel plating quality was determined ter methods (courtesy of NASA and SIFCO)
exposed to air it will immediately oxidise. Various methods the Ti6Al4V sheet, followed by acid copper to 350 m. The
have been studied for the removal of this oxide lm prior to copper was then machined to reduce the copper roughness.
selective nickel plating. Most were unsuccessful, but one The machined copper surface was nally plated with a
successful method involved keeping the titanium under 60 m nickel cap, so demonstrating a useful surface modi-
potential control at all times, the plated area being 100 % cation schedule. This may be of use in different space
covered by the wrapped brush anode and rapidly switching hardware applications where bimetallic couples are to be
from anodic to cathodic conditions as detailed in Table 4.18. reduced to less than 0.5 Vfor instance the nickel to alu-
The resulting anodic etching has caused the titanium to minium EMF difference is 0.45, whereas titanium to alu-
become micro-roughened and there appears to be no oxide minium is 0.63 (see Table 2.1).
lm (see Fig. 4.112). This enables the nickel plating to form Probably one of the most demanding brush plating
a very strong bondit is dense and standard testing on campaigns has been the electroplating of over three hundred
samples has shown the layer to be completely adherent areas on various partially assembled components of the
during bend testing under a cantilever load at 85 % of yield International Space Station. This took place at the NASA
for 24 h. Signicantly, additional testing proved the samples Marshall Spaceght Centre (MSFC) in Huntsville, Alabama
to be negligibly affected by hydrogen embrittlement. Further over a period of one and a half years. Nickel plating was
brush plating tests were made: 35 m nickel was added to carried out by a team of two, an operator and a support
4.26 Selective Brush Electroplating 237
person, working from scaffolding and lift platforms. Great because of the presence of the large intermetallic particles.
care was needed to avoid spillage onto flight hardware. The Most brush nickel-plated areas were typically within hull
primary material of the ISS pressurized modules and nodes cut-outs that accommodate electrical connectors, fluid and
is Al-2219, usually in the T6 or occasionally in the T851 gas line feed-throughs, and ventilation ducting (Fig. 4.113).
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 4.27 Control of Coatings and Bonded
super-saturated with copper for the purpose of precipitation Items by Tape Testing
hardening (see Figs. 4.15 and 5.8). The inhomogenous
macro-surface of machined Al-2219 contains areas of vari- Coatings are used on a wide range of spacecraft surfaces.
able electro-chemical potential owing to the more noble These extend from corrosion protection surfaces to decora-
CuAl2 particles which accelerate the dissolution or corro- tive nishes (for instance, the painted flags and logos on
sion, in ionic solutions, of the surrounding aluminium. By launch vehicles). Adhesively bonded items may be solar cell
analogy, during electro-etch cleaning and electroplating, this covers and labels for part identication. These materials are
potential variation over the surface renders the Al-2219 as applied towards the end of system or sub-system production
being difcult-to-plate. The SIFCO selective plating teams by means of electrodeposition, chemical deposition, dipping,
found it important to keep pre-plating etching and polishing spraying, brush painting or just pressing. Coatings are not
to a minimum during their MSFC campaign, presumably effective unless some bonding occurs with the substrate, by
mechanical attachment, diffusion or chemical/metallurgical More accurate, less subjective tests are selected from
reactions. The essential considerations for cleaning, partic- ASTM D3359 or ECSS-Q-ST-70-13. These measure the peel
ularly the removal of silicones prior to coating, priming and and pull-off strength of coatings and nishes using
painting were extensively described in Sect. 4.21the pressure-sensitive tapes. May be applies to the X knife cut,
maxim failure to prepare is preparing to fail is certainly or to a cross-hatch pattern where multiple preset blades make
applicable to the nishing of spacecraft surfaces! A variety properly spaced and parallel incisions into the coating. Often
of recognized quality control methods can be applied to test no cuts are made and the coating is tested as-manufactured.
the suitability of coatings, both at the end of application, and The tape test used most often for European space projects
after a variety of environmental test exposures which have involves the application of tape (specied range of adhesion
been extensively presented in earlier paragraphs. Most strengths is from 220 to 670 g/cm) with a compressive load
common in commercial industry is the knife test; here, the (5 kg for 60 s) onto the coating sample xed to a top hat
knife blade makes two cuts into the coating down to the support. The tape is then withdrawn and pulled apart from the
substrate with a 3045 angle between the legs to make a sample surface at 0.2 cm/min until separation occurs. The
X. The knife is then used to pull up the coating at the sample may not have been disturbed by the test and can be
centre of the X. This is highly subjective and only used in considered as passed. Samples which have been disturbed
the space industry during the initial setting up of process are investigated further in order to determine the mode of
parameters. failure (see the schematics bd presented in Fig. 4.114).
4.27 Control of Coatings and Bonded Items by Tape Testing 239
The results of the tape test are photographically recorded laser interferometry, and electrostatic (capacitive) sensing.
together with the pull-off load. If results are far lower than The capacitive sensor comprises a hollow cubic molybdenum
expected, this may indicate: wrong selection of coating housing with gold-coated sapphire electrodes mounted in the
material for the intended application and environment of faces. Differential gravitational forces are measured by
testing, poor workmanship (maybe de-greasing before knowing the precise location of the test massesprovided by
coating application), or wrong application procedure. QC responses from the sapphire electrodes and a complex system
tests are made on all Ariane 5 components (Fig. 4.115). 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
4.28 The Application of EB Welding to be completely non-magnetic. The EH is made from sin-
Machine for Reflow Brazing tered pure molybdenumthis is an unusual metal for space
applications as it is very dense and has the very high melting
The LISA Pathnder mission aims to test several technolo- temperature of 2600 C. Molybdenum was selected as it
gies including those designed to establish methods for possess a similar coefcient of expansion to the large sap-
gravitational sensing. This is the rst mission to in-flight test phire crystals designed to be mounted to its surfaces (both are
low frequency gravitational wave detection metrology. An between 5 and 6 106 C1), together with the fact it has a
electrode housing (EH) is in the heart of the Inertial Sensor high magnetic cleanliness, a high thermal stability and can be
Sub-system, as shown in Fig. 4.116. Test masses, consisting machined to the EH designed tolerance of a few microns.
of 1.96 kg cubes made of gold-platinum mono-phase alloy Figure 4.117 shows an electrode housing (EH); this is the
have been chosen because of their extremely low magnetic 3D capacitive sensor containing many sapphire crystals,
susceptibility and high density. The test masses are located in fully gold-plated and in a semi-nished condition. A proto-
the EH and their exact positions are found by high resolution type mechanical model of this EH was made in order to
demonstrate the materials and processes used in its manu- A multidisciplinary team considered several options for
facture. Thermal cycling and vibration were performed on assembling and recovering the EH. Eventually two alterna-
the model, but it failed catastrophically releasing the faces tive assembly processes were selected; both were based on
which had fractured at their interfaces. Failure analyses the use of an electron beam (EB) heating source and the
indicated the poor mechanical strength was likely caused by application of a braze metal (both Au80Sn20 and pure Ti
two inadequacies: the use of a low solder melting alloy to were proposed). The sapphire crystals were already attached
bond the gold-plated surface, and the use of an inadequate to the internal surface housing walls. In order to avoid
soldering oven process. Thin solder preform sheets had been damaging the extremely expensive crystals and their
clamped between the surfaces to be joined. However, due to attachment bonds, the maximum temperature that could be
the high mass of the molybdenum parts, the temperature applied to the piece-parts was limited. Selection of EB
prole of the soldering oven had not incorporated a suf- heating meant that there was the possibility to produce very
ciently long dwell period to enable the entire structure to high spot temperatures adjacent to the braze path.
reach an equilibrium temperature. The solder llets appeared Molybdenum is a poor conductor of heat, and the brazing
compliant on inspection of the external surfaces, but operation is performed within a few seconds under a high
post-vibration, the fracture surfaces consisted of some good vacuum, so that oxidation and damage to the sapphire bonds
metallurgical bonding but many areas of non-wetting and could be excluded. One complication was the fact that some
some zones of AuSn4 intermetallic. A failure review board molybdenum parts were only gold-plated, whereas others
concluded that a new manufacturing route should be inves- had either an intermediate layer of nickel between gold and
tigated for assembly of the EH. 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. 4.119 Type 1 samples. a Visual appearance and microsections to minimal porosity. Needles of the peritectic phase, , are seen
illustrate satisfactory EB brazed joints with Au80Sn20 alloy between throughout this braze. d Nickel is only discernable at higher
Au on Ni on molybdenum. b Gold-tin binary phase diagram detailing magnications, less than 1 m remains on the Mo surface due to
the Au-rich side. The solidication range is from about 520 to 280 C. dissolution into the braze alloy. The lap-shear test joints produced
c Macro-section through the lap-shear joint showing good wetting and strengths in the range 2540 MPa
seen on the (fractured) shear test sample. The Au rich liquid 550 C. However, the obtained lap-shear strengths were
will partially alloy with the titanium interface as shown by 30 MPa, fracture being in the molybdenum and not in the
the phase diagram shown in Fig. 4.120b, but the resulting titanium braze. This metallurgical joining method has been
microstructure can be complex and will contain several successfully used with both palladium and titanium pure
phases that will be dependent on the time and temperature of metal braze ller metals by Chinese researchers (Lin 2013)
the braze cycle. Xu et al. (2013) in their studies found that but using oven instead EB. Figure 4.121 show the appear-
temperatures of brazing can exceed 350 C, but cooling ance of the titanium braze and its macro-section.
should be rapid, and particularly there should be no dwells at Qualication of the Electrode housing materials and
190 C in order to avoid the formation of Au5Sn as this processes were made using the Type 2 braze alloy together
phase was weaker than the ner AuSn/Au5Sn eutectic with a well-understood EB brazing process thermal cycle.
microstructure when bonded to a titanium coated substrate. Operators, trained and certied by the company HTC (High
Type 3 samples, consisting of molybdenum EB brazing Technology Company, Foligno, Italy) performed EB
with pure titanium foils, have been attempted on a limited brazing against a completed process identication speci-
number of samples. These were made as a back-up to sample cation (PID) where all the variable parameters were xed
type 2 (with Au80Sn20 ller). The temperature needed to for these qualication sample production runs (see
obtain this Ti brazing on Mo was far higher than that needed Fig. 4.122). The braze cycle was made rapidly, with a high
for Au80Sn20probably in the region of 1800 C against traverse speed of the electron beam (focused on the Mo top
4.28 The Application of EB Welding Machine for Reflow Brazing 243
Fig. 4.122 Operators producing the EH qualication samples, parts was placed over the samples to protect the sapphire crystals from the
are aligned with the beam and monitored by closed loop TV and electron beam and from any condensed material that could have
microprocessor. Vacuum chamber runs at 106 bar. A tungsten mask sublimated during this high temperature/vacuum process
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 repeated along the 3 axes with a minimum duration of
braze shims. Cooling by conduction would be rapid, thus 120 s. Further tests were made on all axes for 600 s.
avoiding the solid state precipitation of Au5Sn at 190 C, Clearly the EH faces were over-tested with respect to the
and the EB welding vacuum chamber was only opened to LISA pathnder mission equipment vibration requirements
ambient pressure when the items had reached a temperature of: 120 s and 25 g r.m.s. The robustness of the qualica-
when they could be comfortably handled. The qualication tion parts were such that it was not possible to establish a
plan specied that the testing would be performed on EH rst mode of failure. The success of this process develop-
faces BB. All vibration test campaign sessions were suc- ment is important as it involved an innovative joining
cessful. No failures or surface visual degradations were techniqueonly one other reference to a similar process
seen on any of the EB brazed parts. All frequency reso- was established (Lin 2013). With this added condence in
nance test proles were identical, before and after the electron beam reflow-brazing of Au80Sn20 between tita-
vibration runs. The test frequency range was between 20 nium plated molybdenum faces, the conditions for Flight
and 2000 Hz, the level up to 36 g r.m.s. Tests were Model Repair were thus in-hand (Fig. 4.123).
Metallography Applied to Spacecraft Test
Failures 5
5.1 Application of Electron Microscope do not always crack plastically. Likewise, brittle materials
need not display brittle fractures. It is the stress level and
5.1.1 SEM Examination of Fracture Surfaces other loading conditions such as loading rate, duration, and
temperature which determine the appearance of a fracture
The failure mode of spacecraft materials can often be surface.
assessed by detailed examination of the fracture surfaces. Fractures often leave characteristic markings from which
The scanning electron microscope (SEM) is an invaluable the cause of a component failure may be quickly deduced.
tool in such failure investigations because it extends both the However, for accurate identication of failure modes, frac-
depth of eld and the magnications obtainable by con- tography should always be followed by metallographic
ventional light microscopes. Fracture surfaces rarely require examination.
special preparation for SEM inspection. Only when thick The appearance of a fracture surface has always served as
corrosion products or other nonconductive substances are a criterion of the quality or suitability of a material. Brittle
present in the area of examination will a thin layer of gold be fractures are taken to mean fracture without macroscopic
applied to prevent electrical charging effects. Laboratory plastic strain. However, there is invariably local microscopic
personnel should be encouraged to examine as many spec- plastic strain ahead of the tip of a brittle crack propagating in
imens as possible in an uncoated form, as only the uncoated metals. An example of a brittle fracture in a polycrystalline
fracture surfaces can display the true surface structure. If a metal is shown in Fig. 5.1. The top view of this sample
coating is required in order to avoid the build-up of surface shows brittle cracks propagating along particular crystallo-
charge, or to provide a surface layer that produces a higher graphic planes of the grain, termed cleavage planes, to
secondary electron yield than the specimen material, two form a transcrystalline fracture. Brittle cracks can also
options are recommended: follow along a metals grain boundaries, forming an inter-
granular fracture. Plastic or ductile fracture is preceded by
Sputter coating. Here, low-vacuum equipment is used to macroscopic plastic strain due to deformation in the slip
deposit gold or, platinum at a rate of up to 1 nm/s onto planes without crack propagation at either cleavage planes or
the specimen. It must be remembered that sputtered grain boundaries.
metals are deposited in the form of islands, not as a An excellent treatise (Engel and Klingele 1981) fully
continuous coating. To avoid the shadow created in the describes SEM observations of a wide selection of metal
coating by a rough fracture surface it is preferable to coat fractures and damaged surfaces. The book is intended for
the specimen tilted at 45 to the target, and then repeat engineers who wish to locate and identify their own frac-
with the specimen tilted in the opposite direction. tographs with those presented. More recently, fracture surfaces
Carbon deposition. This method involves heating carbon resulting from the premature failure of industrial components
bres to approximately 2000 C under vacuum. Carbon and structures has been compiled by the Italian Welding
will then evaporate and form thin conducting lms on Institute, Genoa, whos metallurgical laboratory is recognized
adjacent areas. Care must be taken to avoid overheating for performing failure analyses (De Marco et al. 2015).
the specimen. As an insight to the usefulness of the SEM in performing
spacecraft materials failure investigations the appearances of
The classication of material fractures is usually made some damaged metals are briefly illustrated in Figs. 5.2, 5.3
with reference to their structure, loading mode, and crack and 5.4. However, no attempt is made to delve deeply into
propagation mechanism. Materials that are said to be plastic detailed metallurgical explanations concerning these modes
(a)
(b)
(c)
transmission work, and this has tended to limit its use in unable to pin, or anchor, dislocations that are moving under
spacecraft failure analysis. The utilization of the TEM is applied stress and this results in an overall reduction of the
unjustiably considered to be solely a university materials alloys mechanical properties.
science tool or a company R&D method for studying the size Other illustrations of the TEMs usefulness as a failure
and distribution of microconstituents in new alloys. Whereas investigation tool are shown in Figs. 4.50b and 5.43, which
the instrument has uniquely assisted in the understanding of show, respectively, precipitation hardening in an electroless
physical metallurgy and, in particular, dislocation move- nickel plating, and an evaluation of dislocations and grain
ment, it can also be applied during the assessment of met- growth with their effect on the embrittlement of beryllium
allurgical failure mechanisms. Figures 5.7 and 5.8 clearly sheet.
demonstrate the elegance of the method and highlight
microstructural deciencies caused by the incorrect pro-
cessing of high-strength structural alloys having vastly dif- 5.2 Fasteners
ferent compositions. Both case histories refer to alloys
which, when freshly quenched, form a supersaturated solid 5.2.1 Spacecraft Fasteners
solution that is soft. Heating to a predetermined temperature
for a specied time causes second phase precipitates to form Fastener systems are the essential elements which may hold
preferentially on certain habit planes. As precipitation takes together both small parts of a device and large subsystems to
place, the alloys become harder, but, if the precipitates are a spacecraft structure. Care must be taken to address
permitted to grow owing to holding the material for too long potential problems associated with nuts, bolts, washers, and
at high temperature, the alloys slowly begin to soften. This rivets, as these items are considered by engineers to be
phenomenon is known is over-ageing. As the initial pre- simple and are often overlooked. Fasteners are normally
cipitates grow in size and change morphology they become installed by torquing the bolt with a wrench so that it
cut threads in both air and water (Kephart and Hayden corrosion is not exposed to tensile stresses). Experience has
1993). The improved properties of rolled threads (for most shown that the fastening of structural parts by rivets is a
metal alloy fasteners) are due to interactive factors, including reliable process; however, in the light of two recent rivet
the presence of benecial residual compressive microstresses failures the following recommendations have been made
in the critically stressed thread root region, reduction of (Jha et al. 1996):
plastic strains during loading, and the formation of a
favourable microstructure. AA 2024 rivets must be driven within 30 min following
A wide variety of rivet designs are employed in the solution treatment followed by quenching, or must be
aerospace industry to hold together metal plates. The kept refrigerated until needed to prevent age hardening
headless end of the rivet is passed through two holes in the prior to driving
plates and then beaten out or pressed down to mechanically material chemical cleanliness is important, the presence
hold them in contact. Wrought aluminiumcopper alloys of second-phase particles and undissolved CuMgAl2
are most often selected for spacecraft applications for the distributed throughout the matrix causes shearing along
fastening of brackets and panels. There is usually no gal- the plane of maximum stress (resulting in the popping
vanic corrosion problem, as all parts are made of a similar off of rivet heads)
aluminium alloy composition with a primary surface nish a smooth llet radius should be provided at the head to
of chromate conversion coating. Rivet compositions are shank junction of the rivet, to prevent stress concentra-
usually AA 2017 (4 %Cu 0.6 %Mg 2 %Ni Al alloy) and tions should the rivet be subjected to bending moments.
less frequently AA 2024 (4 %Cu 1.6 %Mg Al alloy). Both
alloys are selected because they can be solution treated and Industry is presently at risk due to the rise in counterfeit
then age hardened (precipitation treatment) to enhance their fasteners appearing on the market. These are copies or
mechanical properties. The 2 % nickel addition to AA 2017 substitutes whose material and performance (mechanical
acts as a grain-rening element which stabilizes the rivet so properties) are misrepresented by the vendor or manufac-
delaying room temperature age hardening. The AA 2024 turer. It is essential that bolts, nuts and washers are pur-
rivets are slightly more difcult to process, as after solution chased from authorized dealers and where possible, directly
treatment (holding at 500 C then water quenching) they from the manufacturer. European spacecraft manufacturers
should be stored below 0 C in order to avoid natural generally specify the ECSS-Q-ST-46 when procuring fas-
ageing. Both alloys can be readily cold worked (e.g. riv- teners. This document requires fastener heads to be hot or
eted) in the solution-treated condition. After the rivet has cold forged before heat treating; the head-to-shank llets and
been deformed it can be allowed to age harden, so inducing the threads will be formed by rolling; identication marking
high strength. Neither AA 2017 nor AA 2024 has a high will be by depressed characters that univocally identify the
resistance to stress-corrosion cracking, but by special dis- fastener batch and supplier; and non-destructive inspection
position, they are approved for given spacecraft applications requirements are well dened by reference to ASTM and
as the rivet shanks are not loaded in the short transverse ISO standards. Major aerospace fastener manufacturers
direction (i.e. the grain direction most prone to stress include: Blanc-Aero Industries (France), Linread (UK) and
254 5 Metallography Applied to Spacecraft Test Failures
Fairchild Fasteners (Germany), SPS Technologies and Huck manufactured by forging with hot heading of the sockets
Fasteners Int. (USA). Further quality controls are described a process commonly performed at approximately 1000 C
in Sect. 5.2.9. 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
5.2.2 Fastener Failure Due to Forging Defect insert-to-head ratio. The specication called for these bolts
to be quenched and tempered to a hardness of 315 Brinell
The hexagonal socket head from a steel bolt manufactured to minimum (equivalent to a UTS of 1110 MPa). All bolts had
an aircraft standard, as shown in Fig. 5.9, failed during a hardness of between 363 HB in the head volume and 380
satellite testing. The steel was analysed and found to belong HB within their rolled threads. Only by reinspecting the
to the medium carbon martensitic high chromium family. as-polished section under high magnication could the crack
The failed bolt, together with intact bolts originating from initiation site of the broken fastener be established. This is
what was understood to be the same batch, was subjected to shown in Fig. 5.9b to exist as a very ne seam of oxide
visual inspection and metallography. All had been present along 80 % of the crack path. Several other seams
5.2 Fasteners 255
were observed adjacent to the fracture surface. Full material additional 24 h bake is recommended for the removal of
traceability had not been established for this type of fastener, hydrogen. Although cadmium plating is a restricted nish for
and it was considered that the failed item had originated use in high-vacuum environments, it is commonly employed
from defective stock. The seams were believed to have as a general corrosion protection nish to fasteners for
formed during the socket head forging operation because of ground equipment. The deleterious effect of thick cadmium
an inefcient forging-die design, and possibly the forging electroplating and resulting hydrogen embrittlement of
operation was unduly fast for this design. Retroactive action high-strength steel bolts is well documented, and where this
was to randomly sample additional bolts from all batches is of major concern the cadmium should be deposited by a
that had been utilized on the spacecraft. These were longi- vacuum deposition process. Aluminium coatings are also
tudinally cross-sectioned and metallographically inspected. gaining popularity as the environmental problems associated
No seams or further defects were observed on those samples. with cadmium are becoming recognized.
Finally, as a last precaution, magnetic particle inspection Aluminium can be ion vapour deposited onto fasteners
was applied to all fasteners in an attempt to screen-out any made of steel, titanium alloys, and aluminium alloys. This is
other rogue items from the flight batch. 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
5.2.3 Laps and Surface Irregularities in Threads another metal, and no problem of hydrogen embrittlement.
The aluminium coating is adherent and can be chromate.
Fastener specications will permit the existence of minute This nish of high-purity aluminium can be used for appli-
laps, seams, nicks, or gouges on non-load-bearing surfaces cations which require electrically conductive surfaces, such
of a bolt head. No discontinuity or surface contamination is as electrical bonding and grounding for fasteners requiring a
permitted at the head-to-shank llet or at the root of bolt low EMF potential when coupled to an aluminium alloy
threads or other bearing locations. The presence of such spacecraft structure (see also Table 2.1).
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 5.2.5 Embrittlement of Titanium Alloys
as crevices in the surface of metal that have been closed, but
not welded, by the working of the metal. Typical examples Titanium alloys, including Ti6A14V, are known to become
of rejected spacecraft bolts can be seen in Fig. 5.10. These embrittled when held in contact with certain solid metals.
seams are difcult to detect by nondestructive means, and Cadmium can embrittle such alloys at temperatures as low as
they cause a stress concentration to intensify in an even 150 C, whereas a threshold temperature of 204 C exists for
smaller, more localized, region of the thread. Dye penetrant silver and gold (Stoltz and Stulen 1979). The actual mech-
and magnetic particle inspection will detect defects having a anism of solid metal embrittlement is not understood, but it
0.02 mm wide seam, but smaller faults may pass unnoticed is known that cracking in cadmiumtitanium couples occurs
to cause a reduced fatigue life and premature failures. 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
5.2.4 Hydrogen Embrittlement of Steel the grain boundaries.
Fasteners Other recognized titanium failure modes include those
associated with the stress-corrosion fracture of structural
A common source of failure of bolted assemblies is the parts under service loads caused by their exposure to cleaning
hydrogen embrittlement of high-strength steels (UTS in liquids including methanol and certain of the halogenated
excess of 750 MPa). Failures have occurred in a short time hydrocarbons. Free chloride ions present in methanol have
under applied stresses as low as 40 % of the steels yield promoted the SCC failures of titanium alloys. The most
strength. This form of material degradation has been dis- dramatic were the failures of Apollo tanks containing
cussed in Sect. 4.4. The low ductility of such steel fasteners methanol (Johnston et al. 1967); fortunately, these failures
resulted from the absorption of hydrogen during strong-acid occurred during proof testing, and regulations were intro-
pickling. It is now common practice to bake high- and duced to avoid similar damage to flight tanks. The following
medium-strength bolts after pickling at 180 C for 3 h. Some halogenated hydrocarbons are known to promote cracking in
bolts may support an electroplated layer of gold or silver as a titanium alloys (Brown 1972): carbon tetrachloride, methy-
solid lubricant. The process controls for successful applica- lene chloride, methylene iodide, trichloroethylene, tri-
tion of these metals is very important, and, after plating, an chlorofluoromethane, and octafluorocyclobutane. Certain
256 5 Metallography Applied to Spacecraft Test Failures
fluorinated hydrocarbons, registered under the trade names Each time titanium alloys are selected for a space project
Freon TF (C2Cl3F3) and Freon MF (CCl3F), can also lower they will be reviewed by means of the project declared
the fracture toughness KISCC behaviour of Ti6A14V, materials list. By controlling the cleaning liquids, excluding
although the TF grade appears to cause less susceptibility to certain environments (e.g. N2O4 and cadmium plating) and
stress corrosion than the MF grade. These solvents may no selecting fully tested alloys, it can be stated that
longer be purchased, but some companies maintain stock- high-strength titanium alloys have been remarkably free
piles which may be recycled until alternative liquids have from embrittlement problems. Such controls are rigorously
been qualied as replacements (see also Sect. 4.21). maintained for all spacecraft fastener applications, and the
5.2 Fasteners 257
author is not aware of any titanium alloy fastener failure utilize materials, particularly organic ones, with very low
resulting from either solid metal embrittlement or SCC. 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,
5.2.6 Galvanic Corrosion of Fasteners clean packaging, special monitors, and the like as described
in Sect. 4.21.
When dissimilar metals are fastened together, accelerated Liquids and greasesone aspect which should not be
corrosion can occur at their interfaces owing to the previ- overlooked is the generation of contamination from fastener
ously discussed mechanism of galvanic corrosion. Table 2.1 devices which may have been procured to high-reliability
listed metals in the order of their relative nobility in an aircraft standards. Standard oils and greases applied for
aqueous environment. The list proceeds down to the more corrosion protection and against galling are unsuitable for
active (anodic) metals, and it is these that will corrode as a space use as they quickly volatilize or boil away under
result of current flow from the anode to the less reactive conditions of high vacuum, either directly into space or
(more noble) cathodic metal. The rate of galvanic corrosion possibly to condense onto the spacecraft surfaces. Contam-
will depend on several factors, including: ination from organic sources can also collect on the walls
and glass windows of the huge vacuum chambers used to
The conductivity of the electrolyte (e.g. condensed water expose an actual spacecraft to high vacuum and solar sim-
on an electronic box after its removal from a cold envi- ulation as part of its ground testing phase (see Fig. 5.12).
ronmental test chamber). High conductivity is often Suitable lubricants for spacecraft fasteners are listed in
caused by the presence of chloride ions from nger Table 5.2 and further information may be obtained from
marks, or by residual acids or alkalis from chemical textbooks such as that compiled by Fox et al. (2010).
processing operations (fasteners, washers, and their sur- Solid lubricantsmolybdenum disulphide-coated fas-
rounding areas must be maintained in a clean condition). teners nd extensive applications in spacecraft mechanisms;
The area ratio will determine the rate of galvanic cor- they are resistant to most cleaning solvents, but this layer can
rosion; as the area of the more noble metal (cathode) be removed during plasma cleaning as was shown in Fig. 4.
increases, so the rate of corrosion of the anodic metal will 90. MoS2 coatings are applied to commercial fasteners and
be more raid (care must be taken to ensure that the fas- other sliding contacts by burnishing and resin bonding, but
tener metal is either the same polarity or cathodic to the great care needs to be taken if these are selected for space
surfaces being joined). applications as they are known to have poor adhesion and
wear characteristics that generate large volumes of contam-
Great care must be taken when nonmetallic materials must inant particles. They may be suitable when incorporated in
be joined to metals with metallic fasteners. Table 2.1 shows special space-approved greases (Table 5.2) as the MoS2
carbon (graphite) to be more noble than gold. Extensive powder is bound-up by the grease. However, if space
corrosion of steel fasteners has occurred during laboratory mechanisms engineers request a solid lubricant, then MoS2
tests when composite materials made from carbon bre can be applied, either by RF sputtering, or using a new
reinforced plastic (CFRP) were joined to metallic sheets. technique called high ion current density magnetron sputter
Fasteners made from titanium and nickel-based alloys can be ion plating (Fox et al. 1999). As will be seen in other
used without problems because of their high resistance to sections of this book, MoS2 coatings degrade when in con-
corrosion. As additional protection against galvanic corro- tact with humidity. Space mechanisms that incorporate
sion, screws made of titanium alloys are anodically oxidized, MoS2 perform well in vacuum (but are best shielded from
but the ion vapour deposition of aluminium coatings cannot solar radiation). Long in-orbit lifetimes have seen this nish
be recommended for CFRP structures as the voltage potential to be the most preferred dry lubricant. It is essential that such
CFRP-Al lies at approximately 0.9 V, and the fastener is the mechanisms and fastener nishes are bagged in dry
anodic part to the large surface area of the cathodic nitrogen-lled bags during storage. Operation in humid air
CFRP. The harmful effect of coupling CFRP with a generally causes an increase in friction coefcient, a greatly reduced
corrosion-resistant alloy is illustrated in Fig. 5.11. 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
5.2.7 Contamination and Organic Fastener have far improved mechanical and tribological properties in
Lubrication Systems air environments that MoS2. This coating includes a titanium
content of about 15 % and a 1 m thickness is recommended
The control of contamination has led to increases in the (Teer 2001) for terrestrial applications. TEM studies of this
reliability and precision of spacecraft. Much effort is made to 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 poor adhesion and generation of contaminant particles. More
in a coating with high hardness. research is being conducted into DLCs which may result in
Graphite is a material that is both a good electrical con- an improved wear resistant coatings.
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 5.2.8 Metallic Particle Generation
extreme corrosiveness, graphite lubricants in any form
should not be used on spacecraft equipment. Diamond-like Fastener platings and coatings can prevent galling and cold
coatings (DLCs) are very hard, with high internal stress, and welding of the kind illustrated in Fig. 2.4. Silver is primarily
are also unsuitable for space-use due to their brittle nature, a lubricant for high-temperature operations, and can be
5.2 Fasteners 259
b Fig. 5.13 Metallic particles generated from spacecraft fasteners. or magnetic elds.) b Silver slivers generated from the torque loading
a Thousands of 8 m thick nickel chips had flaked off these M5 steel of steel bolts into silver-plated steel stiff-nuts; these cause electrical
fasteners from a microgravity experiment. (Note, nickel is magnetic and short-circuits and must be removed by vacuum cleaning
these chips could accumulate under zero-g in regions of high electrical
5.2.9 Quality Assurance Controls for Fasteners Sect. 5.2.1. As a brief overview of European fastener stan-
dards, the QA controls must ensure the following aspects:
The preceding subsections have described a wide variety of
individual case histories related to European fastener failures (a) Fastener design, material, thread sizes, and surface
and problems during both fastener manufacture and space- nish must comply with national standards and draw-
craft assembly. NASA has also experienced a number of ings (e.g. LN 29950, Cheese head screws with internal
fastener problems as were compile during a technical serrations, made from titanium alloy).
exchange meeting in the mid-90 (Gamwell 1995). Out of (b) Qualication and acceptance tests acceptance/rejection
240 lots of fasteners inspected at one NASA establishment criteria are well dened (e.g. LN 65072, Fasteners
over a two-year period, 42 lots were rejected (i.e. 18 %) made of titanium alloystechnical specication).
during routine incoming inspection tests. The classication (c) Acceptance tests are performed on all lots of fasteners,
and numbers of these nonconforming fasteners are shown in both at the manufacturers plant (results will be recor-
Table 5.3. As part of an overall space project Product ded on certicates, e.g. following DIN 500493.1B)
Assurance Scheme it is essential that Quality Assurance and at the purchasers plant as part of the in-house
(QA) is applied to the procurement and assembly of all incoming inspection procedure.
fasteners whether they be nuts, bolts, screws, threaded studs, (d) Qualied manufacturers are selected from recognized
or rivets. The minimum requirements for the manufacturing, national lists (based on the successful history of manu-
inspection, and quality control of treated fasteners for ESA facture of aerospace fasteners) and following a quality
spacecraft are dened in ECSS-ST-70-46 (2009) as stated in audit of the manufacturing facilities. Process
262 5 Metallography Applied to Spacecraft Test Failures
The microstructures which result from heat-treating this heat-treated correctly. However, as a result of abnormal
alloy in various manners are shown schematically. On slow burning of the propellant, a hot spot occurred in the case
cooling from the phase, the resultant ambient tempera- which reached a temperature of 800 C. This caused the case
ture microstructure would appear as a mixture of marten- to plastically deform and aggravate the liner problem. The
site () and primary . On fast fooling to ambient microstructure shown in Fig. 5.17b illustrates that the rup-
temperature, the phase is replaced by a totally martensitic ture initiation site must have reached a temperature of
() structure. approximately 1000 C for the ( + ) to transform com-
Phase diagrams are the metallurgists blueprints. For pletely into phase. At this temperature, the case initiated a
any combination of metals, a characteristic form of phase ssure which propagated owing to the cases internal pres-
diagram is obtained. For instance, when a piece of polished sures. After case rupture, the hot spot must have cooled
alloy is viewed under a microscope and then heated or rapidly, so that the phase fully transformed into the
cooled at its equilibrium rate, the observer will notice metastable martensitic structure.
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- 5.4 Effect of Inclusions Within
cated by the phase diagram are frequently considerable. the Microstructure
From experience and consultation of other data, e.g. time of Explosively Deformed
temperaturetransformation diagrams, it is possible for the Material
metallurgist to assess metastable structures which result from
nonequilibrium cooling. Pyrotechnic or electro-explosive devices are used exten-
Examination of the failed ABM case proved that there sively in spacecraft to push or pull loads, to pull switches
were no defects in the case structure and that it had been and pins, and to cut cables, wires, and release mechanisms.
5.4 Effect of Inclusions Within the Microstructure of Explosively Deformed Material 265
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
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 5 Metallography Applied to Spacecraft Test Failures
surface roughness, manufacturing lot number, ageing effects, It is recommended that all paints and coatings are
and a lack of adequate processing controls. assessed in specialized test houses by exposing samples to
The major challenge to the materials engineer is to ensure relevant environments prior to their selection and applica-
that the / ratio for any surface remains stable. Organic tion. These may be for LEO and/or GEO applications.
coatings may be not be expected to withstand the combined Thermo-optical properties will include absorptivity, emis-
forces of ultraviolet and particle radiations, temperature sivity, and / ratio. Handling and mechanical properties
cycling, and high vacuum; these factors can cause noticeable might include adhesive strength, abrasion and cleaning
colour changes to such surfaces. Metallic surfaces can be properties, shedding of contaminant particles and suscepti-
modied by thermal fatigue, corrosion during ground stor- bility to fungal growth. Chemical properties might include
age, sublimation, and contamination, but they are not resistance to radiation and atomic oxygen, flammability,
affected by the particle radiation of space. Unless special outgassing-under-vacuum, off gassing and resistance to UV.
attention is given to manufacturing process controls and the Thermal and electrical properties will include resistance to
choice of nishes, the / stability of metallic nishes may thermal cycling (with consideration to substrate), maximum
be degraded as shown in the following case histories. and minimum operational temperatures, thermal conductiv-
Spacecraft thermal control paints and coatings (some ity and electrical conductivity.
being electrically conductive), usually black or white, but
many having different colours, should be procured from
specialist companies. These include: 5.5.2 Low-Emissivity Surfaces
For paints (mostly used on ESA and NASA spacecraft):
A gold-plated (low emissivity, ) aluminium alloy panel
many paints, MAP, Pamiers, France; supporting the electronic boxes of a satellite transmitter and
many paints, Henkel, Dusseldorf, Germany; receiver was noted to have different optical properties (/
Nextel Black, Acktar Ltd., Kiryat-Gat, Israel, and ratio) before and after testing. The aluminium panel had been
Hohenaspe, Germany; nished with initially copper (6 m), silver (20 m), and
many paints, AZ Technology, AL, USA; nally gold (1030 m). It was suspected that the gold layer
many paints, Lord Corporation, Manchester, UK, and shown in Fig. 5.23 was porous, and that the optical property
NC, USA; change was due to surface silver sulphide formed by the
many paints, HTRI, Technology Research Institute, Illi- presence of trace hydrogen sulphide in the industrial atmo-
nois, USA. sphere of the testing site. The resulting laboratory examina-
tion demonstrated the presence of a porous gold layer by
For coatings and conversion coatings: placing part of the panel into concentrated nitric acid.
Streams of bubbles emerged from the pores, account for by:
Aluminized polyester adhesive foils, RUAG Space
GmbH, Vienna, Austria. Ag 2HNO3 AgNO3 H2 O NO2 " gas
Kepla-Coat (previously Plasmocer), AHC, Berlin, Metallography revealed that the gold had been deposited
Germany; in an arboriferous fashion, and, as seen in Fig. 5.23b, it is
coatings for discrete components and roll/sheet for very porous. The gold-plating procedure was revised by
application, Acktar Ltd., Kiryat-Gat, Israel, and Hohe- increasing the current density of the bath so as to produce
naspe, Germany; more nucleation sites, and alkaline brighteners were added to
Black anodizing, Hegin Metal Finishing BV, Heerde, the the electrolyte to ensure the growth of ner gold grains.
Netherlands; Many plating trials were required before an acceptable
Colinal, (black inorganic metal dye applied with an procedure could be established. It was found benecial to
electric current for UV stability), Ateco, Rheinfelder, continually lter the alkaline gold cyanide electrolyte
Switzerland, and Steiger Galvanotechnique SA, through a 2 m lter core, and oil-free air was used as the
Chatel-St-Denis, Switzerland; principal means of agitation. The electrolysis slowly pro-
Ebonol Black, Jackson Plating, Hounslow, UK; duces unwanted organic contamination by cyanide elec-
white, grey and black highly stable PEO coatings, Ker- tropolymerization, and as this phenomenon can also produce
onite Advanced Surface Technology Ltd., Haverhill, UK. rough deposits the bath was monitored daily by Hull cell
various, Sheldahl Corp., Northeld, MI, USA tests. The nal flight equipment panels were noted to possess
Type II anodizing, with black inorganic metal dyes extremely stable optical properties due to the brightness of
applied with an electric current for UV stability, 2Dye4 the gold nish and its freedom from irregularities such as
Anodizing, Inc., Rohnert Park, CA, USA. surface roughness, voids, pits, and visible contamination.
5.5 Degradation of Passive Thermal Control Systems 269
5.6.1 General
Fig. 5.27 Orion crew module pressure vessel. This structural element primer coating. These welded panels still have to be covered with
is machined from aluminium lithium alloy Al-2195 (same as Space thermal protection tiles (Courtesy NASA)
Shuttle lightweight external tank) and protected with green chromated
useful, as they have a relatively low melting temperature for microbalance system. No sublimation was detected. In a
silver-based alloys and reasonably good metallurgical com- further test run at 392 C for 10 h, the sublimation rate was
patibility with titanium alloys. One Gapasil brazing alloy measured to be less than 2.8 105 g/cm2 per year. These
was utilized to join a complex assembly of interlocking tests conrmed that the braze alloy would be suitable for
titanium and beryllium piece-parts during the manufacture of vacuum applications.
an X-ray detector housing for a scientic satellite. The alloy
selected was Gapasil-9, with a composition of Ag9Ga9Pd, a Tungsten Heater Elements
liquidus temperature of 880 C, and a solidus temperature of Heaters for indirectly heated cathodes used in spacecraft
845 C. Gallium is a rare metal with a melting point of only travelling wave tubes are made of tungsten wire. This wire is
30 C. Remarkably, gallium also does not boil at atmo- very slightly doped with thoria to prevent sagging and
spheric pressure below 2300 C. recrystallization. These tungsten wires are designed for use at
Owing to the extremely stringent cleanliness require- 1200 C under conditions of high vacuum. Some life test tubes
ments assigned to the detector housing it was considered were destructively analysed after 44,092 h operation (a little
advisable to assess the sublimation properties of Gapasil-9. more than ve years). Substantial thinning of the coiled wires
This alloy was considered for vacuum brazing operations at due to sublimation was observed and, interestingly, it was seen
920 C, and during service life, the detector housing parts that the evaporated tungsten atoms had condensed in cooler
would occasionally reach a temperature of 200 C. The test areas of the travelling wave tube heater wire so as to form small
apparatus would not reach 920 C, so tables in the Metals tungsten whiskers (further described in Sect. 7.2.2).
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 5.6.2 Sublimation of and Condensation
silver at the same temperature. The brazing cycle only of Cadmium and Zinc
caused the Gapasil to be liquid for less than two minutes, so
virtually no evaporation or condensation of metallic gallium One of the largest vacuum test chambers in Europe was
was to be expectedthis was conrmed during inspections recently used for the testing of a space structure under
after the brazing operation. The sublimation test was per- conditions of thermal cycling between 100 and +120 C at
formed at 258 C for a duration of 10 h using the in situ low pressure (less than 106 torr). The structure consisted of
5.6 Sublimation of Metals 275
which account for the cadmium contamination inside the A large number of heater devices were tested by thermal
environmental test chamber described above. 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
5.6.3 Heater Sublimation Problem Associated power, this test being performed under a vacuum of
with Thruster Motor 105 mm Hg. It was established that nonleaking heaters sys-
tematically would pass this test, whereas leaking heaters
Telecommunications and meteorology satellites, in geosta- would fail because of open circuits.
tionary Earth orbit, are able to alter their orientation (altitude Failed devices were submitted to high-denition
control) and positioning (orbit control) by means of X-radiography (see Fig. 3.23), metallography, and SEM
propulsion systems based on fuels such as either monopro- inspection. Radiography identied several types of defect as
pellant hydrazine or a bipropellant of nitrogen tetroxide with shown in Fig. 5.30. Metallography revealed cracked brazes
monomethyl hydrazine. Small propellant motors, known as and welds as seen in Fig. 5.31. SEM examination of the
thrusters, usually provide a pulsed operational mode for heater lament coil conrmed that the loss in wire resistance
altitude control and a continuous operational mode for orbit is a result of chromium and nickel sublimation caused by
control. Monopropellant hydrazine thrusters derive their operation of the lament in vacuum. These metal elements
energy from the decomposition of liquid hydrazine into were found to have redeposited throughout the heater inte-
nitrogen, ammonia, and hydrogen. rior, and the coils had thinned to the point of total fracture as
2N2 H2 liquid ! N2 H2 2NH3 highlighted in Fig. 5.32. EDAX analyses of the microsection
made of this lament (Fig. 5.33) and others revealed that
The hydrazine decomposition may be achieved by cat- chromium depletion had progressed in advance of the sub-
alytic and/or thermal means. A strongly exothermic reaction limation of nickel. This can be expected by reference to the
results in the formation of hot gases that can be expanded in plots of sublimation rate against temperature given in
a nozzle to convert their thermal energy into kinetic energy Fig. 5.34.
and thereby produce the thrust necessary to reposition the The welding and brazing techniques used to fabricate
spacecraft. The catalyst employed in thruster motors is these heaters have been changed to eliminate the possibility
iridium, which is dispersed over porous aluminium oxide for the inert argon gas to leak out. Other improvements to
pellets so as to provide a large surface area; see also Sect. 5.8 manufacturing and testing, including leak tests to 108 cm3/s,
concerning the deactivation of this catalyst. The catalyst have been implemented to extend operational life to at least
pellets are located in a cylindrical bed surrounded by three 10 years in space.
specially designed heaters.
After the in-orbit failure of several thruster heaters, and in
view of the fact that similar heaters were being physically 5.6.4 Sublimation of Klystron Cathode-Heaters
integrated into the flight units of follow-on satellites, a
detailed failure investigation was initiated. The in-orbit A large number of cathode heater assemblies were subjected
failures were identied from onboard temperature and cur- to long-term life-testing. They had been made to standard
rent measurements to be due to electrically open circuits. military production procedures but with extra quality
They occurred approximately 13 months after successful assurance surveillance. The device is a scaled-down version
operations following the launch of this satellite. of a typical klystron cathode heater gun; it is designed for the
The heater designs incorporate a 0.025 mm diameter same voltage as a travelling wave tube gun, and its fabri-
Nichrome V coiled element which is welded to a 0.25 cation route was considered as a dry run for the manufacture
diameter platinum lead. This heating coil is supported by of space hardware designated for Earth resources pro-
alumina spacers and is surrounded with 0.040 mm alumina grammes. A number of metal sublimation problems were
powder particles. The heater coils and ceramic parts are identied during the course of the life-testing.
packed into a cylindrical Inconel housing as shown in One failed device was found to have a detached heater
Fig. 5.29. The leadwires are embedded in an Inconel 600 leg; it is compared to an intact leg in Figs. 5.35 and 5.36.
tube containing magnesium oxide for insulation. After The high operational temperatures of these cathode hea-
assembly the heaters are backlled with dry argon to prevent ters (approximately 1185 C) have caused the platinum
oxidation. return leg ribbons to undergo excessive grain growth.
A literature search established that similar heater designs According to the engineering drawing the leg should be
had failed owing to inadequate joining of the lament coil to separated from the alumina ceramic by an air gap of 1.5 mm.
the lead wires, and to a loss of the protective argon gas Grain growth has caused the failed ribbon to weaken and sag
(Schatz et al. 1979). under its own weight, causing it to make contact with the
5.6 Sublimation of Metals 277
Fig. 5.29 Heater investigation. a The heater assemblies were heater subassembly showing positions where housings were cut (by ne
subjected to the Gross Leak Test specied in Mil-883 (Method 1014 turning on a lathe). The delicate lament wires were not disturbed. The
test condition C). Each complete heater was immersed into fluorocar- welded plug ends were removed and the ceramic discs extracted from
bon at 125 C. Leaking heaters were identied by streams of bubbles cylinders. Certain of the brazed seal portions with jumper wires were
issuing from both plug and transition junction (arrowed) demonstrating withdrawn
leakage in excess of 103 atm. cm3/s. b Schematic cross-section of
cylindrical molybdenum heater container. Solid-state diffu- platinum metals are given by Arblaster (1996). Improve-
sion of platinum into the contacting molybdenum (arrowed ments to this cathode heater were made by lowering the
on Fig. 5.35b) was conrmed by microsectioning as a molybdenum cylinders so that they could no longer contact
wear-out mechanism. The platinum ribbon subsequently the platinum ribbon. It is envisaged that greater reliability
thinned, became overheated, and caused an acceleration in would be achieved by replacing the pure platinum ribbon by
sublimation rate from the large-grained region. It is expected one made from dispersion-strengthened platinum [known as
that the platinum ribbon had operated at a temperature of ZGS and developed by Johnson Matthey (Selman et al.
1500 C for approximately one year before its failure. Fig- 1974)]. The following benets can be expected:
ure 5.37 illustrates how sublimation and preferential thin-
ning at the platinum ribbon grain boundary have caused this A higher recrystallization temperature (with less grain
return leg to overheat and fail by fusion. growth during operational life).
From the viewpoint of sublimation (see Table 2.5) it is The sagging, or high-temperature creep, of the return
recommended that parts fabricated from pure platinum and leg would be reduced or eliminated.
intended for long-life applications (e.g. in excess of eight ZGS platinum is easier to form than pure platinum.
years under vacuum) should not be operated at temperatures Higher resistance to shock loads (e.g. launch vibration).
greater than 1100 C. Accurate vapour pressure data for the Less chance of tearing in weld heat-affected zones.
278 5 Metallography Applied to Spacecraft Test Failures
It is emphasized that the direct replacement of the con- tube can be varied by changing the power input to the heater.
ventional materials with ZGS platinum can be made only Test programmes have been completed under conditions of
after a careful investigation has ensured that the redesign high vacuum showing the potential suitability of rhenium for
will not affect the electrical characteristics of the cathode this application. However, certain precautions must be taken
heater. 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
5.6.5 Sublimation of Rhenium rhenium tube while in the hot, powered condition.
Rhenium surfaces rapidly pick up oxygen to form
Recently developed thruster motors incorporate a coiled unstable, low-melting-point oxides. Rhenium in its pure
length of rhenium heater tube. The tube is heated by means state has a melting point of 3170 C. However, like other
of its own electrical resistance such that it reaches temper- refractory metals, rhenium is oxidized both in air and
atures approaching 2000 C. This electrothermal thruster under partial vacuum at elevated temperatures. These
device produces thrust by expanding the gases of decom- resulting oxides, particularly the higher ones, are volatile
posed hydrazine (ammonia, hydrogen, and nitrogen) by and have very low melting points (e.g. Re2O7 melts at
heating as they flow through the tube. The temperature of the only 296 C).
5.6 Sublimation of Metals 279
Metallography indicated that testing of one of the tubes at electrical resistance increases to promote even higher oper-
about 2000 C for 300 h caused a reduction in its wall ational temperatures and a further increase in the material
thickness by 0.04 mm. If it is assumed that sublimation of loss. Intense enlargement of rhenium grains occurs above
pure rhenium accounts for this thinning effect, then this will 2000 C (grain growth). This has resulted in tube failures, as
occur on both internal and external tube surfaces (e.g. a shown by Fig. 5.38, by the degradation of the part by
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 accelerating sublimation rates particularly along exposed
sublimation rate can occur if the rhenium is held at tem- grain boundaries,
peratures in excess of 2050 C. excessively high temperatures of operation,
As the cross-sectional area of tubes decreases, either grain boundary sliding once grains have the same
because of oxidation or because of sublimation, so their diameter as the tube wall thickness.
280 5 Metallography Applied to Spacecraft Test Failures
5.7.1 General
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
5.7.6 Investigation of Microcracked Thin-Foil improve the materials formability. Some minor improve-
Detector Windows ments in ductility were achieved by these laboratory heat
treatments, and on the basis of the results new foils were
Beryllium foil is widely used for windows in spacecraft fabricated following three different routes. The processing
X-ray and gamma-ray detectors because of its low absorp- history and selected laboratory test ndings are presented in
tion of these radiations. When vacuum hot-pressed beryllium Fig. 5.42. The appearance of dislocation networks and
block is cross-rolled into sheet and foil its property values recrystallized grains are shown in Fig. 5.43, and from this
are increased by cold work and the generation of crystallo- transmission electron microscope information it is clear that
graphic anisotropy. However, it must be noted that as rolling rolling produces a characteristic microstructure having a
progressively increases the properties of strength and duc- high dislocation density. Annealing processes at 600 C
tility in the plane of the sheet, the out-of-plane properties are produce a more well-developed subgrain structure and lower
signicantly reduced. This behaviour became evident during dislocation densitiesthe microstructure indicated this
the assembly and ground operation of a low-intensity material to be partially recrystallized, whereas annealing at
gamma radiation detector before integration into a research 650 or 710 C produces a more recrystallized structure
satellite. The detector under investigation incorporated, by which is almost dislocation free.
necessity, a window foil of only 3050 m thickness and Failure in the Type 1 foil from the defective detector is
was required to house (hermetically sealed) a controlled and due to its exceedingly low ductility, the presence of a rough
stable internal pressure of argon gas. The detector was found surface condition, and the existence of dislocation walls and
to fail during operation under vacuum by networks leading to transgranular cleavage. The chemical
composition of this material was within its specied
(a) leakage of argon through the foil, and high-purity limits. Fracture of the 650 C annealed foil was a
(b) cracking during bowing of the foil as a result of dif- mixture of transgranular cleavage and intergranular crack
ferential pressure across the window. propagation. Failure in the fully recrystallized material was
by intergranular fracturethis would suggest that during
Mechanical testing and metallography determined that the annealing weakening of the grain boundaries and/or possible
argon leakage failure resulted from microcracks in the foil strengthening within the grains had occurred. No precipita-
that had been introduced during the cold-rolling operation. tion was observed within the grains which could produce a
The internal argon pressure induced bowing of the foil strengthening effect, and it is likely therefore that the fracture
during operation of the detector under vacuum, causing the mode was due entirely to embrittlement of the grain
microcracks to propagate through the thickness of this boundaries. Grain boundary embrittlement during annealing
35 m thick foil. is common in both structural steels and nickel-based alloys.
Various processing routes were evaluated with the In both cases the effect is due primarily to the segregation of
assistance of the Fulmer Research Institute to determine a phosphorus or sulphur, although other elements also play an
preferred rolling and heat-treatment sequence which might important part. The specic mechanism by which phos-
produce a more isotropic foil that could bend or buckle phorus, or any other element, weakens the intrinsic cohesive
without the formation of microcracks. It may be noted that strength of a boundary is not well understood. Very little, if
beryllium has a hexagonal crystal structure whose basal any, work has been done on grain boundary embrittlement in
plane is both the twin plain and the cleavage plane. Twin- beryllium. However, by analogy with steels or certain nickel
ning is promoted by deformation on the basal plane and can alloys, it is quite possible that the intergranular fracture in
result in the formation of microcracks when twins intersect the 710 C annealed beryllium is a result of grain boundary
grain boundaries (Tetelman and McEvily 1967). The flow embrittlement due to segregation of impurity elements dur-
stresses for basal slip are remarkably low at room tempera- ing the anneal. This should be an important area for further
ture and decrease further with increase in temperature (Beurs investigation since the grain boundary monolayer composi-
et al. 1987). tion can be easily determined by Auger electron spec-
Beryllium foil from the same batch as used for the leaking troscopy, and this, in turn, may influence the mechanical
detector was submitted to chemical analysis, mechanical properties of the annealed material.
testing both normal to, and parallel with, the nal rolling The conclusions drawn from this investigation are as
direction, bend testing, microsectioning, fractography, and follows:
TEM evaluation at the Fulmer Research Institute.
Foils from the initial batch were later annealed for Samples which had been taken from unidirectionally
20 min, some at 650 C and others at 710 C, in an attempt rolled foil stock have been found to have extremely
to both remove any residual stresses caused by rolling and 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 5 Metallography Applied to Spacecraft Test Failures
equi-axed grain structures of a dimension approaching In order to study the compatibility of hydrazine with the
the foil thickness. various component materials used in the construction of the
The cross-rolled samples (foil types 6 and 7) have a reaction control system (RCS) an investigation was initiated
uniformly high strength both parallel and transverse to to determine the gas generation rate of hydrazine versus
the foils long direction. This material has rather low temperature and time by means of a long-term test utilizing a
ductility (better, however, than the cross-rolled batch of space quality thruster motor. The test lasted 12 months, and,
foil used for the leaking detectorfoil type 1). after a throughput of 23 kg of hydrazine, resulted in a sig-
nicant deactivation of the thruster catalyst. This catalyst
The main recommendations were to advise that foil types degradation has been the focus of the following
15 be rejected. Foil type 7 was later demonstrated to be the investigation.
most suitable, and has been successfully incorporated into at The test equipment consisted of a hydrazine tank fabri-
least one series of gamma-ray detector housings. 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
5.7.7 Aluminium-Beryllium Alloys description of the test procedure has been published by Dunn
and Papenburg (1986).
Beryllium is one of the lightest aerospace structural mate- The test was terminated after a signicant thruster deac-
rials, with a density of 1.85 g/cc, while aluminiums density tivation and various materials investigations were initiated.
is approximately 45 % higher at 2.7 g/cc. A family of These included:
low-density alloys is being developed based on aluminium,
with additions of between 10 and 75 % beryllium. They have Opening the thruster at the injector end, then inspecting
been found to combine the high specic stiffness of beryl- for particle contaminants.
lium with the ductility and ease of fabrication of aluminium. Detailed examinations of the used catalyst particles and
The new alloys have been produced by powder and ingot comparative studies with unused particles by visual,
metallurgy methods at the Brush Welman company in scanning electron microscopy (SEM), surface analysis
Cleveland, Ohio. One of the more successful binary alloys (EDAX and XPS), mass spectroscopy, and metallography.
was noted (Haschiguchi et al. 1992) to be Al40 wt%Be. Emptying and opening the propellant tank to sample the
When cast and hot rolled, sheet material of this composition AFE-332 diaphragm material by X-ray and SEM analy-
had interesting mechanical properties: yield strength sis. A full characterization of this elastomer has been
289 MPa, ultimate tensile strength 386 MPa, and elongation published by Martin (1976) and Schreib et al. (1979).
at fracture of 17 %. Analysis of hydrazine samples (that had been drained off
after the various thruster rings) by different methods to
determine the inorganic impurities.
5.8 Deactivation of Catalyst Particles
for Hydrazine Decomposition
different appearance from the new particles, being covered 5.8.3 Mechanism of Particle Deactivation
with a dense, white, almost continuous shell. These shells
vary in thickness from one particle to another. The maxi- A loss in performance of thrusters can be caused by cat-
mum shell thickness is about 80 m (this might be an alyst deactivation. The test programme has identied a
oblique cut through the layer), but it is generally 10 m failure mode related to the formation of a surface shell of
deep. Some particles are further coated with a thin trans- amorphous silica which acts as a barrier layer in preventing
parent glaze of only 24 m thickness. Line scan analyses intimate contact between hydrazine and the iridium
conrmed the particles to be coated with a silicon-rich layer, catalyst.
probably SiO2. Similar ndings were reached by XPS and The silicon contaminant is believed to originate from the
mass spectroscopy. AF-E-332 diaphragm elastomer. Hydrophobic silica is
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).
Fig. 5.47 Polished cross-section through NEW, AS-RECEIVED a dark appearance, and except for some small areas (arrowed) does not
catalyst particles. Note the rounded corners and porous nature of this have the bright, reflective surface characteristic of a metal. c Area B.
material. Many open-mouthed pores provide these pellets with an Showing friable nature of these particles and some bright lamellae
extremely large surface area. a Typical NEW catalyst particle showing within the matrix. d Area C. Bright lamellae are also present in central
locations of detailed micrographs. b Area A. Even at high magnica- region of particle
tions there is no evidence of a free iridium surface layer. The pellet has
The free barium diffuses along the pores of the tungsten The extent of wear-out can be assessed by metallographic
to activate the cathode surface. A comparison between means. The tungsten pellets are removed from the test
Figs. 5.52 and 5.53 shows the long-life cathode to have assembly and mounted in epoxy resin. As the tungsten pores
produced an eroded emission surface. The tungsten grains contain barium, calcium, and aluminium oxides, no lubricant
have ion-etched at various rates dependent upon their crystal can be used as this might otherwise dissolve or dislodge the
orientation to the electron flow direction. They are often oxide particles. A suitable means using a microtome for
striated; their grain boundaries have become clearly visible, making cross-section through these pellets has been found;
and EDAX analysis has shown that the entrapped pores may the mounted pellets can be sliced in 2 m steps with a very
become depleted in the impregnated oxide when accelerating sharp tungsten-carbide knife edge. Microprobe line-scan
factors of current density or temperature are incorrectly analyses are then made, as shown in Fig. 5.54, commencing
selected and far greater than those encountered during ser- at the cathode emitter surface and following a path normal to
vice operation. In fact, when service failures do occur they it. Recently introduced sectioning methods, since this
usually take place at the beginning of the cathode life owing examination was made, might have provided more details
to poor workmanship or material defects and not by from our cathode investigationthey include both the FIB
wear-out of the barium-emitting compound. and argon ion milling methods described in Sect. 3.2.2.2 and
5.9 Cathode Emitter Degradation 293
the observations of Kim (2012)and obviate the sample plastics (CFRPs) were probably rst introduced via the
smearing noticed in Fig. 5.54. manufacture of antennas to provide for an electrically con-
The life-tested porous-tungsten cathode showed no con- ductive, highly stable surface prole, even under severe
centration gradients for barium and aluminium throughout thermal cycling in vacuum. They are now often used in
its volume. This indicates that free barium is still available satellites for structural components between 4 and 450 K as
for diffusion, and wear-out is not imminent. Calcium does reported by Schedler (1988). The design and development of
show a depletion toward the emitter surface due to its a 2 m diameter antenna, with a complex convex beam
inability to diffuse, in the oxide form, along the intercon- reflector, provided a challenging task to European designers
necting pores of the pellet. and manufacturers. The mechanical design problems asso-
Investigations have shown that excessively high test ciated with the launch environment and extreme thermal
temperatures (above 1100 C) can cause the porous tungsten environment in orbit necessitated that the antenna reflector
pellets to become more compact by the continuation of the should be constructed from a sandwich of CFRP face-skins
sintering process. This causes a specic mode of failure (e.g. with an aluminium honeycomb core bonded together by
no emission) owing to the seal-off, by sintering and grain means of lm adhesives. This assembly would satisfy the
growth, of the interconnecting network of pores to prevent critical requirements of low mass, high rigidity, low thermal
the barium activator diffusing from the pellet bulk to the deformation over a wide temperature range, and good
cathode emitter surface as illustrated in Fig. 5.55. reflection of radio frequencies.
A test programme was attempted on a qualication model
antenna (Fig. 5.56) to demonstrate that the design aims had
5.10 Investigation of a Failed Spacecraft been achieved. Vibration testing invoked damage to the
Antenna antenna, and once material failures became visible on the
surfaces the tests were terminated. A request was made that a
Spacecraft continue to make increased use of thermosetting materials evaluation be performed on the failed antenna,
bre-based composites to reduce the weight and increase the particularly in the regions of debonding and fracturing of the
stiffness of structural components. Carbon bre reinforced CFRP. A wide range of nondestructive test methods was
294 5 Metallography Applied to Spacecraft Test Failures
Fig. 5.49 Appearance of RIB-SIDE of used hydrazine tank has been used for many years for elastomeric diaphragm and bladder
diaphragm. This side had been in contact with hydrazine. The reverse tanksthe elastomer is ethylene propylene diene modied (EPDM)
side had been in contact only with nitrogen. Material is AF-E-332 and polymer (Marker is 1 cm)
interference patterns were produced with an even distribu- face-sheet lay-ups (Fig. 5.61) and the form and extent of
tion across the individual surface areas inspected (Fig. 5.60). wetting achieved by the adhesive lm on the aluminium
Debonded areas, not previously apparent from visual honeycomb and face-sheets. The samples were cut from
inspection, were carefully marked on the antenna surface. regions noted by holography to be either mechanically sound
Similar holographic interferometry is performed on a regular or debonded. Areas located adjacent to CFRP delamina-
basis for the NDT inspection of 2 m length cylindrical tions and fractures (Fig. 5.62) or debonds (Fig. 5.63) could
spacecraft motor cases constructed from composite materials have degraded and were also microsectioned (Dunn and
(Barbier and Le Floch 1985). Collins 1978).
Several special chemical and mechanical tests were per- An assessment of the investigation results showed that the
formed on the antenna, but probably the most important hand laid-up antenna face-sheets were not observed to
deductions were made from the results of metallographic contain any major defects. Also, mechanical tests made
examinations. Microsections were taken from the various uniaxially (mode I), or in pure shear (mode II) on tensile
areas of the antenna to establish the integrity of the CFRP samples cut from the face-sheets, produced values of
296 5 Metallography Applied to Spacecraft Test Failures
Fig. 5.53 Appearance of worn out cathode pellet. Many thousands 5.11 The Wear of Ball Bearings
of hours under test at very high current densities has depleted the
barium content of the subsurface Spacecraft mechanisms frequently incorporate
high-precision bearings consisting of an inner race, outer
suitably high strength. However, application of low twisting race, and either balls or rollers manufactured of hardened
or buckling forces (e.g. mixed-mode I/II) appeared to easily and tempered chromium steels such as AISI 440C and AISI
delaminate this CFRP material, separation occurring E52100. The balls are retained and separated from each
between lay-ups of differing orientation. The bond strength other by means of a metallic or nonmetallic cage. Particular
between the aluminium honeycomb and face-sheets was applications on-board spacecraft will encounter environ-
considered inadequate owing to: 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 299
Fig. 5.57 Debond and fracture of CFRP on back face, adjacent to mounting pillar
operation at cryogenic temperatures. For these reasons the had not caused these materials to become stressed beyond
common oil and grease lubrication systems used for terres- their yield stresses; that is, no plastic deformation occurred
trial bearings cannot be used, and either organic (e.g. poly- on a macroscale. Local deformation on a microscale had
tetrafluoroethylene), metallic (e.g. gold and lead thin lms), taken place, however, so that the two metal contact surfaces
and ceramic (e.g. titanium carbide, titanium nitride, and could be pushed sufciently close together under vacuum
silicon nitride) coating materials have to be considered. The and at room temperature to create a minute cold pressure
choice of space-approved synthetic oils is severely limited weld. These microwelds rupture almost instantaneously to
(see Table 5.2), and either temperature limitations or the produce a roughened surface that becomes progressively
close proximity of optical systems may preclude their more prone to further microwelding as the peaks on the race
selection. side become compressed during subsequent contact with the
Operation of bearings under vacuum will cause the ball ball.
bearings to make contact with the adjacent races by rolling Modern bearings now support customized thin coatings
and sliding. Conditions such as long operational life, high such as silicon nitride and silicon carbide on conventional
loading conditions, or poor workmanship (e.g. wrong steel bearing components, and, for European spacecraft,
alignment or high surface roughness) will cause the lubri- bearings coated with titanium carbide (TiC) have widespread
cation lm to either wear or become displaced, so that applications. TiC coatings are also frequently used to avoid
metal-to-metal contacts occur. This can lead to an increased the seizure or galling of highly loaded contacts in hold-down
bearing friction, the risk of cold welding, and ultimately and release mechanisms, sliding pivots, and solar array
seizure of the mechanism. Figure 5.64 illustrates the damage hinges (Briscoe 1982). The procedure for attaining
incurred by cold-welded ball-bearings. Loading conditions 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. 5.61 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 303
Boving and Hintermann (1987). In this case, TiC is applied preferred ball material (better than 440C) to be used with
by chemical-vapour deposition (CVD) onto steels such as steel races for cryoliquid pump systems. One reason is that
vacuum-remelted 440C. Deposition takes place at a rate of Si3N4 balls lower heat generation and improve heat transfer
the order of 1 m/h. Boving and Hintermann have also (Ryan et al. 1995). Tests in Europe by E.S.K. (1995) also
demonstrated a much reduced dry friction coefcient for revealed that Si3N4 balls can operate at very high speed, with
sliding contacts of TiC-coated steel against steel (0.3 both in high load-carrying capacity (even at 800 C) without seizure
dry air and under vacuum) when compared to steel against or cold welding during dry running.
steel (greater than 1, both in dry air and under vacuum). The Metallography is employed as a quality control tool for
TiC CVD coating on 440C steel balls is initially too rough to the measurement of coating continuity and thickness as well
be immediately used, and the balls must be high-precision as a check to ensure a correct microstructure of the substrate
lap-polished before bearing assembly. A large amount of (see, for instance, Fig. 5.65). The TiC-modied surface
data was generated during the post-flight tribological avoids cold-welding at solid-to-solid interfaces between
assessment of the Hubble Space Telescope solar array balls and races in the absence of oils or greases and under
mechanisms. The bearings were noted to be in good con- extreme environments such as vacuum and either elevated or
dition (TiC-coated titanium carbide balls and cryogenic temperatures. The coating has a slight residual
outer-race-riding leadbronze cage) but the outer race, compressive stress on its substrate (Chollet et al. 1985) but
manufactured from the bearing steel 52100, showed some does not crack or spall even when exposed to the stresses
local corrosion spots (some method of surface protection is imposed by thermal cycling between 269 and 30 C.
needed for this steel, or the use of 440C material). As a nal note, it may be important to reconsider the
Titanium nitride coatings are often used to improve the choice of cleaning solvent. An Alert which circulated the
friction and wear properties of unlubricated sliding contacts space industry in 1996 provided mechanical engineers with
in terrestrial applications where the presence of ambient the following information:
humidity is benecial. At low humidities, or under vacuum,
Bearings and other lubricated components used in long-duration
the TiN coating is associated with increased friction and spacecraft have traditionally used ozone-depleting chemicals
wear (Mohrbacher et al. 1995), and this may limit its use- (ODCs) such as freon as a nal cleaning process prior to
fulness for space mechanisms. lubrication. As these chemicals are being phased out, new
Ball and roller bearings made of silicon nitride ceramics cleaning processes have taken their place such as supercritical
carbon dioxide, perfluoromalocarbons, aqueous detergents, etc.
have been proven to provide smooth operation even under These new processes are effective cleaners, but leave the bearing
extreme conditions. They are identied by NASA as the surfaces with a different residual surface chemistry. Some of
304 5 Metallography Applied to Spacecraft Test Failures
5.12.1 General
Fig. 2.4. The wear mechanisms of fasteners and ball bearings spacecraft structure or its sub-system. The specic DC
were also covered, respectively, in Sects. 5.2 and 5.11. resistance between the mechanism and the grounding plane
Great care must be taken during the selection of materials will depend on whether the electronic components are
and their coatings in order to prevent wear and cold welding, mounted within the mechanism i.e. less than 5 m, or if
or adhesion, between piece-parts such as gears, solar array there are no electronic components, only protection against
hinges, bearings, electrical commutators, and the like, which electrostatic discharge is needed 0.1 (see
operate under vacuum. As mentioned in previous sections, ECSS-E-ST-33-01).
the tribological behaviour under space environmental con- Probably the most demanding conditions of sliding con-
ditions differs signicantly from terrestrial conditions. Oxide tacts are those associated with the Infrared Space Observa-
and nitride layers which are formed in the terrestrial envi- tory (ISO), which was launched in 1995 and is operating
ronment on pure metal surfaces, and can be regarded as successfully. The cryostat cover was ejected 10 days after
natural protection layers, are not formed in space. During the launch and all scientic instruments have since been tested
service life of spacecraft mechanisms these surfaces can and have been making observations of various objects in
outgas, wear, fret, and become generally cleaner and this space. One instrument utilizes focusing and shutter mecha-
will increase the probability of cold welding. Also the inti- nisms that have to operate at 4 K. Another incorporates a
mate contact of nonmoving parts (e.g. fasteners) may lead to lter change mechanism with gear drive and shaft-mounted
cold welding in the space environment and this may hinder ball bearings; these also operated at cryogenic temperatures.
the disassembly of orbiting space hardware by astronauts. Here, lubrication is by means of a molybdenum disulphide
This necessitates the use of suitable materials, coatings, lm and the gears are made from titanium alloy with titanium
and/or lubricants, many of which can now be assessed by the carbide surfaces. Another area of challenge is that associated
test method prescribed by Merstallinger et al. (2009). In the with the Space Station Alpha and the Polar Platform projects;
case of the Hubble Space Telescope solar array refurbish- these spacecraft will have the largest arrays yet built in
ment mission, the arrays could be wound up by hand and Europe, with nominal power capacities of around 15 kW. The
removed from the spacecraft by means of turning space station arrays will be extended and retracted at
silver-plated steel bolts. six-monthly intervals for a lifetime of 30 years. The Polar
In this Section, an overview of space tribology will focus Platform arrays will be Sun-following, with a rotation of 6
on mechanisms that operate under conditions where surfaces revs per day. These arrays have led to the design of rotating
may have a tendency to cold weld under sliding wear, under power transfer assemblies with large-diameter titanium car-
impact forces, by surface fretting, and under long-term bide bearings and high-power slip rings.
loading. It may be pertinent, when considering surface n- Several new developments are taking place in the area of
ishes for mechanisms, to point out that all mechanisms need tribological surfaces. Filled and reinforced polymer com-
to be electrically bonded (for grounding purposes) to the 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-
Fig. 5.65 Micrograph of a section through a space-quality TiC-coated ing wear and friction. The most frequently used in the
440C steel ball after lapping and polishing. Maximum surface space industry simulates the behaviour of sliding wear and
roughness is 0.012 m CLA. Vickers microhardness of coating is
35,000 N/mm2. Microstructure is revealed by etching in Marbles
is tested under vacuum conditions by the pin-on-
solution (500) (courtesy of the Swiss Center for Electronics and rotating-disc method (especially DIN Std. 50324). Other
Microtechnology) methods are the pin-on-ring and pin-on-plate. The test
306 5 Metallography Applied to Spacecraft Test Failures
parameters should mirror those seen in the space mecha- made of Densimet (D175K) for the body side, and stainless
nism application and include: contact pressure, sliding steel (Inox Z6CNU17 04) for the mobile bracket, was found
speeds, loads due to thermal distortion and preload varia- to resist cold-welding and this combination of materials was
tions plus application loads, temperatures, etc. These have incorporated into later designs of the black-body mechanism
been detailed by Labruyre and Urmston (1995). The tri- (Fig. 5.67).
bological components, contacting surfaces, and lubricants A test equipment for the evaluation of cold-welding under
utilized by the mechanisms of the Hubble Space Telescope cyclic loading and high vacuum was designed and produced
solar array were microscopically inspected and reported on by the former Austrian Research Centre, Seibersdorf. This
after their return to Earth. The ndings demonstrated that equipment is sited at the Aerospace and Advanced Com-
the present approach to the selection and testing of con- posites (AAC) GmbH site in Wiener-Neustadt, Austria. It is
tacting materials is adequate. There was generally no evi- capable of simulating a wide range of contact parameters
dence that, after approximately 3.5 years of spaceflight including static and impact loading. The equipment has been
operation, modications to materials or lubrication systems upgraded to include the simulation of fretting wear and
would have been benecial to this particular spacecraft resultant cold-welding under vacuum. A standard method
(Anderson et al. 1995). 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
5.12.2 Cold Welding Due to Cyclic, Impact equipment was also shown to successfully reproduce the
Loading sequence of events that led to the cold-welding failure of the
calibration mechanism described above (Merstallinger et al.
One of the early Meteosats suffered problems when its 1995a, b; Merstallinger and Semerad 1996). Material was
black-body calibration mechanism became blocked after cut from the engineering model black-body mechanism and
several years of successful operation. The mechanism would submitted to impact adhesion testing. The impact force was
remain blocked for 7 h during operation in orbit in what was seen to effectively break down the black coating on the
called the bracket-to-end-stop position. As one of several magnesium alloyeven with a low impact energy of
failure modes (e.g. relaxation of returning springs, problem 0.004 Nm. The highest cold-welding adhesive forces
with magnet) cold-welding between bracket and end stop attained during this test were: bare magnesium alloy
was attributed as being the main cause of failure. An engi- (GA3Z1)-to-black-coated alloy (60 mN); after breakdown of
neering model of the calibration mechanism, having the coating Mg-to-Mg (280 mN); and, Densimet-to-Inox
same build, was ground tested under vacuum and after (100 mN). The test conditions for cold-welding were quite
repeated impacts between the bracket and its end stops the similar to the actual in-orbit failure which occurred after
onset of cold-welding was observed after several thousand 8230 activations with an impact energy of 0.0112 Nm, 30 g
operations. This mechanism is shown in Fig. 5.66a. Both the load, that is with a 300 mN force required to unblock the
end stop and the bracket had been constructed from the mechanism. A repeat laboratory test was conducted with a
magnesium alloy GA3Z1 (AZ31) having the composition similar impact energy of 0.02 Nm; this resulted in adhesion
3 %Al, l %Zn, 0.3 %Mn, balance Mg. All surfaces had been forces of up to 280 mN between metallic contacts of GA3Z1
nished with a black chemical conversion coating (black being measured. This gives some condence that the new
mordancage). The on-ground laboratory tests conrmed test equipment can actually simulate working mechanisms
that the repeated impact forces gradually wore down the making impacts under vacuum.
black coating, until it had been completely removed at the Impact testing has also been performed on a modied rig
small points of impact. Microscope inspections revealed that by CSEM in Switzerland. It was found that all tested mate-
once the coating had been removed there was sticking rials were subject to some form of cold-welding and that
between the mating surfaces. Mechanically deformed parti- coatings of molybdenum disulphide or titanium carbide can
cles were noted to have been pulled from one surface and reduce this phenomenon (Maillat et al. 1995). Both the
redeposited during the impacting cycles, as seen in Merstallinger et al. and the Maillat et al. reports indicate that
Fig. 5.56b. This conrmed the cold-welding theory and no similar studies are available in the literature. Both con-
initiated the investigation of a test method suitable for the clude that the cold-welding (adhesion force) measured as a
evaluation of materials and surfaces that were expected to be function of number of impacts can be divided into three
subjected to wear and cold-welding. In parallel, many impact phases. Initially there is no, or negligible, adhesion during the
tests were made using the calibration mechanism shown in rst few thousand impact cycles; then there is gradual
Fig. 5.66, but with modications so that different small increase in adhesion force followed by a nal phase where
metallic inserts could be placed at the bracket impact posi- the adhesion force reaches a constant value. It is postulated
tion and on the end-stop position. A suitable pair of contacts that the constant value is related to some kind of self-healing
5.12 Cold Welding of Mechanisms 307
Fig. 5.66 a Engineering model of the black-body calibration mech- 5.12.3 Cold-Welding Due to Fretting
anism with BeCu springs S and magnet M. C shows the mechanism
bracket in one end stop position where the blocking occurred. This is the form of cold-welding, or galling, occurring on the
b Scanning electron microscope view of flattened particles and some
evidence of cratering. The particle arrowed was analysed by EDAX contacting surfaces of moderately-to-heavily loaded metals
and seen to be composed of the same elements found in the magnesium and alloys which are subject to slight relative movement.
alloy (GA3Z1) together with traces of potassium and chromium from Fretting taking place in mechanisms and between structural
the worn black coating (600) parts when there is a small amount of play between adja-
cent surfaces. Under atmospheric conditions this gives rise
of the surfaces oxide layers due to diffusion of interstitial to galling and the tearing off of tiny metallic particles which
gases from within the metal to its surface, or from the residual soon oxidize and then may worsen the problem as the oxi-
gases within the vacuum chamber (held at 107 torr). dized particles can be very abrasive (e.g. aluminium oxide
More recent results have shown that theoretical predic- and tin oxide). Under vacuum conditions the galling will still
tions are by no means comparable with experimental data occur, but now in the absence of oxygen, the released
308 5 Metallography Applied to Spacecraft Test Failures
particles can themselves become cold-welded to one or both thicker specialized coating to ensure there will be no cold
of the adjacent surfaces. welding under vacuum. Such coatings were evaluated by
A literature review of fretting associated with space means of the AAC test method depicted in Fig. 5.67 and
conditions was made by Vine and Price (1992), clearly little described in the previous Section. Preliminary results of a test
empirical data are available concerning the fretting and wear campaign aiming to nd suitable nishes for titanium and its
rates of spacecraft materials at different temperatures in alloys having a high resistance to cold welding under condi-
vacuum. Cold-welding due to fretting during ground trans- tions of impact and fretting are described in the captions to
portation of the US Galileo spacecraft may have caused Fig. 5.69ad. Trade name coating processes based on ceramic
damage and later a malfunction of its main antenna conversion treatment (plasma electrolytic oxidation treatment)
deployment systemthis became only partially unfurled at a of titanium include: Apicoate, Balinite, Dicronite and Kero-
critical stage during this crafts journey to Jupiter. Damage nite 3000. For further reading the following references are
by fretting has been observed during the ground testing of a recommended: Li and Dong (2010), Jiang and Wang (2010),
latching device, similar to a Marman-clamp, on the Cassini Li and Zhang (2014) and Merstallinger and Dunn (2009).
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 5.13 Defective Black-Anodized Electrical
occurred when the anodized AA 6061 was coupled to Connector
stainless steel 316, causing damage to both steel and the
anodized layer. Following extensive testing a suitable com- Spacecraft electrical connectors consist of the metallic shell
bination of surfaces was found: 60 m thick hard anodized (endbell, coupling nut, and barrel), the insulator and grom-
AA 2219 coupled to a latch clamp ring made of PEEK met materials (possessing low outgassing properties), and
(plastic) coated AA 2219 (Ylikorpi et al. 1995). Sticking the connector pin and socket contacts which are usually
resulting from fretting was also considered and tested for the machined from brass, have a copper and nickel strike, and
holddown and release unit of the Envisat solar arrayvi- are nished with at least 2 m of electrolytic gold plate. The
bration tests conrmed the acceptance of anodized alu- shell parts are made of aluminium alloys or 300-series
minium running against nickel-plated aluminium stainless steels. Aluminium parts are either anodized or
(Zwanenburg 1995). nickel-plated, whereas the steel parts may be passivated or
The European spacecraft Hipparcos Baffle Cover gold-plated. The plating or anodization thicknesses are not
Mechanism failed a ground functional test during open always specied, but connectors must pass the salt spray
close operations. This was found to result from an unac- corrosion tests according to ML-STD-202, method 101, test
ceptably high increase in friction between a titanium alloy condition B. It should be noted that the usual commercial or
bush and a stainless steel spring. These parts were supposed military connectors supporting chromate coating over
to slide over each other, but during vibration testing fretting cadmium-plating cannot be used under high vacuum owing
had occurred. This caused some small surface areas to make to the problem of sublimation as discussed in Sect. 5.6.
contact and become cold-welded. Figure 5.68a shows where The connector shown in Fig. 5.70 possesses a
sticking had taken place, and Fig. 5.68bd highlights in black-anodized nish. The aluminium alloy is 2024, and a
greater detail the area where fretting and cold-welding had 0.2 mm translucent oxide has been built up in a
taken place. Tests showed the 177 PH steel to have a low-temperature sulphuric acid electrolyte to produce a hard,
microhardness of 736 Hv, and the Ti6A14V bush a micro- wear-resistant lm. During anodizing the oxide lm consists
hardness of 385 Hv. The titanium alloy bush was modied, of two layers: a thin inner barrier layer which is composed of
and instead of an anodized nish, the outer surface was compact Al2O3, and a thick 1200 m outer porous hydrated
coated with chemical-vapour-deposited titanium carbide. oxide lm (Gabe 1978). The anodized lm had been dyed
This solution was found to avoid further fretting and the black by a nickel sulphide process (ECSS-ST-70-03) and
mechanism performed perfectly, both during qualication nally sealed in boiling deionized water.
testing and later, when the spacecrafts telescope baffles The connector seen in Fig. 5.70 shows the effect of
were deployed in orbit. thermal cycling followed by a salt spray corrosion test. The
Titanium and its alloys are used extensively in structural body is covered in supercial white corrosion products, and
application and spacecraft mechanisms. Mating surfaces are metallography reveals microcracking in the anodized layer.
now known to easily cold weld because either the natural, or A close inspection of the cross-sectioned layer reveals a
anodic blue oxide nishes are too thin and have poor nonuniform lm growth due to the presence of copper-rich
adhesion under mechanical wear, particularly under vacuum. inclusions distributed throughout the aluminium. The thick
These nishes have good corrosion resistance but, for sepa- oxide lm is very hard and insufciently compliant to sur-
rable mating contact surfaces, it may be essential to build-up a vive the thermal cycling test procedure. As a result of
5.13 Defective Black-Anodized Electrical Connector 309
anodizing trials suitable parameters were established to during inspection and test and, the deleterious consequences
produce a thinner, 35 m oxide lm that did not craze during of their presence within space hardware have been discussed.
100 thermal cycles between 100 and +100 C. Salt spray In this section, several examples will be briefly described
testing produced no evidence of corrosion products, and the where costly spacecraft subsystem failures have been caused
connector was subsequently space-qualied. 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
5.14 Contaminant ParticlesIdentification useful information to supplement the detailed inspection and
of Their Sources microscopy work was gained from the various projects
Declared Materials List. The Declared Materials List
A number of problems associated with contaminant particles (DML) is a compulsory compilation describing each mate-
have been described and illustrated in this book. Until now, rial selected for use on-board the spacecraft. The list is
the particles have been identied in situ, and it is clear that produced by every contractor producing flight hardware and
they originated from processes such as welding (Fig. 3.6), is used as a formal document for the approval of the material
sliding wear (Fig. 4.40), laceration of screw threads and its intended application. The list is formatted so that
(Figs. 5.13 and 5.14), from inserts that had been drilled and each material is identied by commercial name, by chemical
threaded but inadequately cleaned (Fig. 5.15), or particles that composition, by manufacturers specication, by processing
spalled from the surface of work-hardened beryllium during parameters, and its precise use and location must be
thermal cycling (Fig. 5.41). These particles were revealed described. As many hundreds of metallic alloys are used on
310 5 Metallography Applied to Spacecraft Test Failures
Fig. 5.69 Preliminary results of a test campaign aiming to nd suitable at 100 % EL. Under these specic test conditions, the following
nishes for titanium and its alloys having a high resistance to cold coatings have shown very little adhesion: Balinite A (TiN), Balinite B
welding under conditions of impact and fretting. Balinite A (TiN), (TiCN), Dicronite + (WS2 + CrC), Keronite and Ni-PTFE. Fretting
Balinite B (TiCN), Dicronite + (WS2 + CrC), Keronite and Ni-PTFE results are not shown. It is recommended that coatings produced by any
coatings appear to be good contenders. It is important that Process manufacturer should be tested if cold welding under vacuum is
Identication Documents are generated for each coating process to considered an important spacecraft parameter. (from Merstallinger and
ensure all parameters are maintained under strict control. Once Dunn 2009). b FIB milled image through top surface of 6 m thick
approved, there can be no improvements made to the process unless Keronite coating showing its columnar structure beneath a 200 nm
the new ideas are re-assessed by testing. The results shown in a can be thick fused oxide. EDX analyses only revealed Ti, Al and V as can be
traced to the coating morphologies illustrated in bd for each expected of this PEO coating. c FIB milled image of Balinite B coating,
designated coating (courtesy of AAC, Austria). a Experimental cam- top is surface with micro-indents, middle is 2.9 m coating thickness
paign showing results of maximum adhesion force recorded following and lower area is substrate. EDX analyses revealed Ti, Al with some C
standardised impact testing. Note that for impact testing, three contact and N. d FIB milled image of Dicronite coating. The double layer
pressure values are used. The rst 10,000 impact cycles are made at thickness of (WS2 + CCr) is about 10 m
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
Fig. 5.71 Case 1. Difculties were experienced in the qualication of composed of iron, chromium, and nickel, together with a binder having
a liquid helium cryostat valve. Nonuniform leak rates occurred owing a high chlorine content with calcium, potassium, silicon, and sulphur.
to the presence of metallic contamination particles that had become d Show the length of the re-entrant path which, despite extensive fluid
embedded on the Vespal valve seat. The valve was disassembled and a cleaning of the internal surfaces, became a source for particle release
section cut at the valve port (material AISI 304 stainless steel)-to-outlet during the flow of liquid helium. The binder and particles originate
pipe weldment (AISI 316 stainless steel) using a diamond cutting tool. from the cutting pastes used to dress the AISI 304 valve port prior to
Many particles were seen entrapped within a small gap between the welding. The pipe did not release particles, as AISI 316 contains
pipe and valve port, close to the weld. Chemical analyses were made by molybdenum and this element was not present in the particles.
EDAX of the particles: those embedded on the valve seat had the same Corrective actions included a revised cleaning procedure before
composition as those trapped in the gap; they are seen in ac and are electron beam welding
the optical detector, after integration into the housing, was manufacturing routes for an antenna and horn are associated
considered unacceptable. It was thought likely that silicone with, respectively, Figs. 4.514.53, 5.56, 5.57, 5.58, 5.59,
from the pore sites could be redeposited onto a glass lens 5.60, 5.61, 5.62 and 5.63. Where the service region on the
during the thermal vacuum testing that was planned to Earth is irregular in shape, the antennas reflector surface
qualify this instrument. Corrective action was to manufac- may be formed to appropriately cover this shape. The
ture a new housing and exclude silicone lubricants from the manufacturing of the carbon bre/epoxy reflector skin,
instrument. 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
5.15.3 Contamination of Invar Moulding Tool tools for antennae must provide temperature resistance, a
very low coefcient of expansion (CTE) to match the
Spacecraft antennae are usually of the reflector type, such composite, vacuum integrity, thermal conductivity, and
as a paraboloid illuminated by a horn. Examples of 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. 5.75 Case 5. Stainless steel corrosion products in the form of of polyvinyl chloride (PVC) tape that had been inadvertently left
particle growths and curly bres were found on the internal surface of a attached to the bellows as they were stress-relieved in an oven at
spacecraft propulsion systems nonreturn valve. The growths were red approximately 180 C. PVC starts to decompose at about 80 C, and
brown in colour and the bres, as seen in a and arrowed on b, have a consequently evolves hydrogen chloride gas (HC1). This scenario was
diameter of 510 m and a length of 500 m. They were analysed and repeated under laboratory conditions. PVC tape attached to polished
found to be composed of iron, oxygen, and chlorine. The bellow AM350 was heat-treated in an oven at 180 C. The resulting
material is AM350, a semi-austenitic, precipitation-hardening stainless laboratory-grown bres are shown in c and have the same morphology
steel. The same corrosion products were later found on other valves as the contaminated flight parts. It is postulated that only the small
and, more worrying, these products were present on the PTFE valve volumes of delta ferrite, present in the microstructure of AM350, are
poppet seals. As a result of a long investigation it was concluded that corrosively attacked and produce this form of corrosion product
the corrosion products were caused by the exposure of AM350 to the (courtesy of Mr. P.W. Martin, MMS-UK) (Color gure online)
vapours of hydrochloric acid. This acid was released from some lengths
Other, cured solid silicones may be mechanically SILGEST SD0001 from Dunlop Adhesives, Birming-
removed by peeling [this is possible for some conformal ham, UK.
coatings (Dunn and Desplat 1994)].
Two special removal agents are also available: Both agents are extremely hazardous and require great
care in their use. They are skin irritants, the products being
DIGESIL from C&M Research Ltd., Eneld, UK (or flammable and corrosive.
from Merck in other countries), and
316 5 Metallography Applied to Spacecraft Test Failures
5.15 Silicone Contamination 317
b Fig. 5.76 a Topography of black-anodized aluminium housing as seen The contaminant was extracted with dichloro methane, a powerful
by SEM inspection (backscattered electron image). Superimposed is an solvent and submitted to FT-IR analysis. This conrmed the presence
EDAX spectrum clearly identifying the presence of silicon (from a of silicone oil (the ngerprint corresponded to a reference sample of
silicone oil contaminant) within a surface pore. This image reveals silicone oil)see box, top right. b Pinch-offs were contaminated with a
parallel machining marks that had been present on the original thin layer of silicone oil which was presumed to have condensed from
aluminium alloys surface. During the sulphuric acid anodizing process, vapour onto the internal surfaces of these annealed aluminium tubes
a porous lm of hydrated aluminium oxide would have built up. This during an incorrect vacuum pumping operation. (i) Hermetic sealing by
was then dyed by a nickel sulphide process, and then sealed by cold welding had not been successful and the location of minute leaks
immersion into boiling water. The crystallites of alumina are separated were only found after helium leak testing. (ii) Special cleaning steps
by the curved lines; the remaining pores appear as dark spots. The main were introduced and once totally free of contamination, consistent
elements on the surface are: Al and O being alumina; Ni and S from the vacuum-tight joints were achieved. The processed pinch-offs are sharp
nickel sulphide inorganic dye; and Si from the silicone contaminant. and very delicate so that threaded end caps are attached
5.15.5 Contamination of Aluminium Tubes them to deform, flatten together and, like a cold extrusion,
for Vacuum Pinch-Offs the internal surfaces should cold weld together. Constant
pressure, without any lateral movement, nally causes the
Ductile metal tubulations are frequently designed into metal to thin down to a pointat which time it will instantly
spacecraft systems in order to hermetically seal a unit that sever.
may have been pumped down to vacuum, or may later be Pinch-off sealing pressure can be delivered by either
lled with a particular gas or liquid. These tubes are usually pneumatic or hydraulic pumpsit is a delicate procedure
made of half-hard aluminium (such as AA-3003-H14 (see where tube material and outer diameter must be matched to a
Appendices 5 and 6 for heat-treatment and composition). particular tool size. Some parameters need to be optimized
Occasionally, annealed OFHC copper is selected, when and this must be done using sample tubes having precisely
glass-to-metal seals are included in designs for power the same characteristics as the flight-material (i.e. chemical
devices; and high-purity nickel tubing may be used for composition, hardness, grain-size, dimensions and surface
systems holding liquids such as ammonia heat pipes, or cleanliness). Samples will be visually inspected, occasion-
when high unit bake-out temperature will be envisaged ally microsectioned, and might be submitted to gross (bub-
(nickel will have lower outgassing and less oxidation than Al ble) and ne (helium) leak tests. The delicate sharp end of
or Cu when bake-out temperatures exceed 150 C). Pinch-off the pinch must be protected as shown in Fig. 5.76b. It is not
seals avoid the need for heavy valves and they have been recommended to enhance the quality of these cold welded
effective for generations of spacecraft. Before pumping seals by further processing such as melting, welding or
down the system it is essential that the area of tube to be soldering the tube end.
pinched-off is exceptionally clean. Ultrasonic and mechani- For a detailed account of evacuation tube closure, the
cal cleaning is preferred over chemical cleaning. process development report by Zirker (2007) and co-workers
However, in the case of a defective pinch-off (described at the Idaho National Laboratory is recommended reading.
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 5.16 Magnetic Problems
aluminium evacuation tube was connected by weldments to
an electronic housing containing electronic circuits for high The magnetic eld generated by a spacecraft interacts with
power applications. It had been designed as a leak-tight the local eld from the Earth and this results in an external
housing for evacuation, then lling with the gas sulphur coupling force being exerted on the craft. The effect is
hexafluoride (SF6). The chemical stability and high thermal analogous to a compass needle in the Earths magnetic eld
conductivity of SF6 allow it to extinguish arcs by thermal and may torque the spacecraft and alter its attitudeso
transport; its breakdown voltage exceeds that of other gas- requiring the ring of thrusters for station-keeping
ses, having a maximum of about 140 kV at 0.2 MPa pres- manoeuvres. Owing to the high density of ferromagnetic
sure. The concept was that this unit could be tested in the materials they are seldom used for large structural space
laboratory, could be calibrated and space-approved, then applications and are therefore unlikely to affect attitude
later would be expected to operate during the launch phase control. On the other hand, the presence of small-sized
under an ever decreasing ambient pressure, and nally in the magnetic materials on-board a spacecraft can limit the ef-
vacuum of space. ciency of some detectors, such as magnetometers. For this
The tools for pinch-off processes include tungsten carbide reason, care is usually taken to exclude these materials from
inserts that are pressed into the tube outer surfaces, causing critical designs and only select from nonmagnetic materials
318 5 Metallography Applied to Spacecraft Test Failures
or those having paramagnetic properties (i.e. those materials to be carried by all practising metallurgists!) or if they
that, when inside a magnetic eld, behave with feeble interact with a small pocket compass. Alternatively, a more
magnetic properties). sensitive test is to degrease the parts and immerse them for a
Several magnetic materials are used for their specic few minutes in a solution of acidied copper sulphate
properties on-board spacecraft. The largest number are in metallic copper will deposit on those parts of the surface that
electronic devices, with most of the remainder being in are magnetic (i.e. deposition will be on areas possessing a
electronic power apparatus (e.g. motors, transformers choke martensitic or ferritic grain structure).
coils, travelling wave tubes, reed switches, recorder heads, The factors affecting magnetic properties of metals
ferrite-core memories, etc.). Where stray magnetic elds can include their crystal structure and Curie temperature. It is
produce unwanted effects on neighbouring equipments, it is common that as the temperature is decreased, a materials
possible to shield them with sheets or cans fabricated from magnetic properties will increase. Conversely, as the tem-
magnetic metals and ferrite cupssuch shields form closed perature is increased there can be a temperature where the
magnetic paths. A magnetic shield material often used in material becomes nonmagnetic (the Curie temperature is
space applications is Mumetal (ironnickelcopper alloy), dened as that temperature at which all of the metals
which can be wrapped around the magnetic components of electron spins are disoriented by the randomising energy of
travelling wave tubes. thermal vibration, making the net magnetization effectively
Electrolytic and electroless nickel plating can have strong zero). Details of the magnetic properties of metals and their
magnetic properties (see Sect. 4.10) and have given pertur- alloys are contained in specialist books including the edition
bative effects on the instruments of scientic satellites by Cahn and Haasen (1996).
(Delahais 1985). As stated above, the austenitic steels are generally
Many austenitic steels (that are usually stainless and thought to be either nonmagnetic or paramagnetic, and this
nonmagnetic) can become strongly magnetic when they are was also considered true for most of the Inconel alloys (e.g.
heavily work-hardened. The 18-8 type of steels (designated Inconel 718, 625, and 600).
by AISI as 301, 302, 304, 316, and 321) are used throughout Cryogenic launch vehicles, such as Ariane 5, consist of a
spacecraft for applications such as fasteners, struts, stringers, main stage (referred to as the EPC from its title in French,
inserts, and in mechanisms. They have a high resistance to Etage Principal Cryotechnique) that is composed of an
stress-corrosion cracking and need no form of corrosion aluminium tank with two compartments: one containing
protection except possibly a deliberately oxidized passiva- 133 tonnes of liquid oxygen (boiling point at ambient
tion layer. These alloys are nonmagnetic in the annealed pressure is 90 K), the other containing 26 tonnes of liquid
(fully austenitic) condition. However, the alloys become hydrogen (27 K). At the base of the EPC is the Vulcain
magnetic when they are cold-worked owing to the engine assembled mainly from stainless steels and Nimonic
deformation-induced transformation of nonmagnetic alloys possessing a low magnetic permeability.
austenite to magnetic alpha-prime martensite. Small increa- Scientic experiment detectors such as those installed in
ses in the nickel content of these 300-series steels decrease the ISO satellites cryostat operate below the temperature of
the extent of magnetization by cold-working: 301 steel with liquid helium (7 K) at 4 K. Whereas the mechanical prop-
about 7 %Ni becomes strongly magnetic, whereas 304 steel erties of many steels and nickel alloys are known at these
with about 10 %Ni becomes somewhat less magnetic. This cryogenic temperaturesparticularly their ductile-to-brittle
increase in magnetic permeability is often called martensitic fracture characteristicsvery little is known, or has been
stress induced transformation, MSIT. It must also be noted published, concerning the magnetic properties of these
that during high temperature heat treatment, 300-series steels engineering alloys in this temperature regime. It is now
may undergo a microstructural change from austenite to known that most of the austenitic steels and Inconel alloys
martensite if they are slow cooled from the austenitizing change their magnetic states when exposed to sufciently
temperatures. cold temperatures. The precise Curie temperature depends
The most common spacecraft parts to become magnetic on the materials composition. The 300-series of steels can
when they have been cold-worked are cold-rolled or change from the metastable austenitic (paramagnetic) phase
cold-drawn springs. Also, heavily machined or cold-worked to a ferritic or martensitic (ferromagnetic) phase when
(by cold forging) stainless steel fasteners may have ferro- exposed to cryogenic temperatures. (The steel AM-350
magnetic outer surfaces in regions of their shanks and contains 17 %Cr 4 %Ni, and 3 %Mo, and is deliberately
threads. This can lead to problems when they are used for hardened by reducing the temperature to 75 C for at least
the mechanical assembly of delicate electronic instruments. 3 h in order to transform all the austenite to martensite;
A quick check to screen out magnetic items is to see if they subsequent tempering at 520 C for 3 h produces good
are attracted to a small magnet (an essential diagnostic tool strength and ductility.)
5.16 Magnetic Problems 319
At very low temperatures a partial transformation of maximum temperature envelope of 10 C. This would be
austenite to martensite takes place in some of the less alloyed representative of equipment operating on the inside of a
stainless steels such as AISI 301 and 304. It is impossible to spacecraft body where the thermal equilibrium is controlled
quote, or tabulate, precise transformation temperatures (e.g. so that the absorbed energy is balanced by the radiated
Curie temperatures) for the different alloys. This is because energyas was described in earlier chapters.
signicant changes can occur as the composition of the alloy Temperature changes cause materials to expand and
changes within its own permitted specied limits, and these contract in different ways. Between room temperature and
values will also be dependent on the parts thermomechan- that of liquid nitrogen, the total linear contraction of space
ical history. It is therefore recommended that special labo- materials may vary from that of Invar and Pyrex glass
ratory tests are made for magnetic permeability for each (0.05 %) to thermosetting resins (more than 5 %). Problems
critical application. can soon arise when materials are used together without
Whether or not Inconel 718 can be used as a nonmagnetic proper forethought. A simple example might be a stainless
(paramagnetic) material will depend on the temperature, the steel bolt and nut that were used to assemble an aluminium
exact composition, and the limit of magnetic susceptibility housing to a spacecrafts aluminium structure. As the craft
that can be tolerated for a given application (Goldberg et al. was warmed by solar heating, the aluminium parts would
1990). This is generally considered as a nonmagnetic expand far greater than the steel (23 ppm against 16 ppm) and
material and is used in many spacecraft applications. How- the steel bolt would be mechanically stressed so as to cause it
ever, it appears to become strongly magnetic at a tempera- to yield and stretch. On cooling to a low temperature during
ture close to that of the boiling point of liquid hydrogen the eclipse portion of the crafts thermal cycle, the bolt would
(27 K). As mentioned in the Goldberg et al. paper (1990), now be too long and a gap between housing and structure
this characteristic of Inconel caused a hydrogen fuel bleed would be formed. In this case of the housing there is unlikely
valve system to malfunction on the Space Shuttle Discovery. to be a problem, as electrical grounding will be maintained by
The monitor for the valve position was made of Inconel 715, the steel bolt. However, in the case of an aluminium bracket
and on cooling to the temperature of the liquid hydrogen it attached to a thruster motor by means of a steel fastener,
produced erroneous readings; this initiated a metallurgical continuous thermal excursions did work loose the xture.
failure investigation and caused a delay to the launch of Probably one of the most difcult engineering problems
Discovery in 1988. to overcome is the stress-induced deformations that are
Finally, to make matters even worse, the transformation produced by temperature gradients within a cryogenic
of austenitic material into ferromagnetic martensite is propulsion system such as that shown in Figs. 2.112.13 for
accompanied by a volume expansion which can spoil the t the Ariane V launch vehicle. Here, gears rotating and
of accurately machined piece-parts such as shafts and interlocking at 27 K must not be subjected to thermally
flanges. If this is likely to be a problem, the stock material induced compressive forces that can cause friction, wear, or
can be immersed into either liquid nitrogen (77 K) or a seizure.
more appropriate liquied gas and repeatedly
thermal-cycled between the liquid and room temperature
until no signicant dimensional changes occur. This will 5.17.2 Stress-Relaxation by Thermal Gradients
also often have the effect of stabilizing the low-temperature
magnetic properties. 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
5.17 Thermal Stress-Induced Dimensional detector that had been designed to operate inside a large
Changes liquid helium cryostate as seen in Fig. 5.77. The detector
instruments actually operate at a temperature of 2 K and are
5.17.1 General Problems electrically connected through the walls of the cryostat to the
rest of the spacecraft by means of a specially designed wire
The thermal environments that spacecraft parts might be harness. The signal wires of this harness are constructed
subjected to range from cryogenic temperatures to those from single strands of enamelled steel, each having a
where materials are almost white-hot. Clearly, it is the diameter of slightly less than 100 m. Steel was selected in
propulsion systems that witness the greatest thermal excur- order to minimize the heat input to the cold instruments
sions. At the other extreme, an ideal isothermal mass tem- (normally copper wire would be used for all electrical con-
perature is taken to be 25 C, plus or minus 10 C, with a nections, but this has a high thermal conduction). Electrical
320 5 Metallography Applied to Spacecraft Test Failures
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
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)
2
The solar absorptance to solar emittance ration, /, for polished steel
is 0.5/0.13 = 3.88. For aluminized FEP Teflon, the / ratio is 0.1/0.8 = 0.12.
3
5.17 Thermal Stress-Induced Dimensional Changes 323
cylindrical concertina thermal shield, the value for this 5.18.2 Alpha-Case Embrittlement
parameter has been reduced to 0.02 C per second, a
reduction factor of 25. 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
5.18 Defects in Titanium Piece-Parts elevated temperature air exposure causes the oxygen, and to
a lesser extent nitrogen, to stabilize the alpha phase of the
5.18.1 General alloy, so producing a hard and brittle layer, termed the alpha
case. This is often associated with the presence of visible
The main titanium material to have found extensive use in surface contamination consisting of oxide and nitride layers.
spacecraft applications is the alphabeta alloy, Ti6A14V. The hard, oxygen-rich layer needs to be removed from the
A limited volume of pure titanium has been employed, but nished forging as it is likely to adversely affect bend duc-
owing to some manufacturing problems it is the alloys that tility and fatigue strength and lead to difculties during
will be employed for future advanced applications. Some machining and welding. Generally the surface oxides and
have been described in Sects. 4.17 and 4.20. The majority of nitrides as well as all the subsurface alpha case is removed
propulsion and reaction control systems are fabricated from from forgings by rst dislodging any adherent scale by
titanium because the pure metal and most of its alloys have a surface blasting processes and then acid pickling, chemical
good strength-to-weight ratio and are compatible with both milling, and/or machining. As previously inferred, the key
the hydrazine-based fuels and the nitrogen tetroxide oxidants. structural components in the systems for manoeuvring
Fuel lines are made of long lengths of formed and arc-welded spacecraft are the Ti6A14V pressure vessels. Massive alpha
titanium tubes. Also, the nonreturn valves that are situated as inclusions caused the failure of several Apollo-era hardware,
redundant pairs in the pressurant lines between the pressure including the Lunar Module descent propulsion system
regulator and the propellant tanks are constructed from tita- which failed during proof testing (Ecord 1972). Experience
nium parts. The fuel, oxidizer, and high-pressure gas tanks has shown that for unmanned projects the pressure vessel
are generally fabricated from forged, machined, and welded design can be successful with a burst pressure factor of
titanium hemispheres (Fig. 3.6). Some tanks have been made safety of 1.5 provided strict controls are implemented for the
by superplastic forming (see Fig. 4.77), but the majority are critical steps of forging, heat treatment, welding, cleaning,
from forged parts. Spacecraft detector housings, such as for and testing.
the Italian SAX scientic satellite, have been made from Problems arose during the fabrication of one large vessel
formed titanium sheet that was later joined to beryllium for a satellite. The manufacturing steps were to follow a
transition couples by explosive welding or to thin beryllium standard route. Two forged and machined hemispheres, one
windows by special braze alloysboth processes being half with an integral mounting boss, the other with a pene-
described in other chapters of this book. tration port for lling the fuel, were submitted for girth
The Ti6A14V alloy is forged above the beta transus (see welding. The conventional inspection techniques for the
Fig. 5.18) to take advantage of the better plasticity and the hemispheres, both before and after welding, were to involve
lower forging pressures needed at these temperatures (usu- radiography, ultrasonics, and dye penetrant testing. The
ally 870960 C). However, during forging there can be a material was chemically analysed and test-pieces were ten-
build up of temperature due to internal frictional energy and sile tested and metallographically examined using off-cuts
this can affect the nal microstructure and cause the surface from the forged blanks. All results were acceptable. How-
of the alloy to become more prone to contamination. To ever, one of the half-shells supported an unusual small area,
avoid overheating it is advisable to follow the surface about 2 cm2, where the external circumferential machine
temperature of the forged item with an optical pyrometer. In markings were associated with surface microcracks. They
order to achieve a uniform alpha-beta duplex structure it is could be highlighted by the dye penetrant but not by
preferable to cool the parts to just below the beta transus and X-radiography. Owing to this surface anomaly, the part was
produce the nal working operation isothermally using the considered unsuitable for flight use; however, it was welded
superplastic properties of the material. A post-forge heat to a sound hemisphere (this was for welding operator
treatment will also be applied in order to anneal the material. training purposes). Proof pressure tests were made and the
The main causes of failures to titanium piece-parts have surface microcracks readdressed. Some cracks appeared to
been identied as resulting from either the formation of hard, have enlarged but this could not be quantied. The tank was
surface-diffusion layers (alpha case) on forgings, or the for- nally submitted to a series of destructive mechanical and
mation of titanium hydrides within the parts microstructure. metallographic tests to assess the quality of the weld as well
Examples of both are given in the following paragraphs. as account for the microcracks.
324 5 Metallography Applied to Spacecraft Test Failures
Metallography showed the forged hemispheres to have the A particularly difcult case of hydride embrittlement was
correct ne-grain, alphabeta microstructure. In the defective discovered on a spacecraft when a length of titanium tube in
region the surface was seen to have been contaminated to a the propellant fuel line was found to have broken. The tube
depth of about 20 m. Here, the microstructure consisted of a had been fabricated from pure titanium, with an outer diameter
hard, light-grey zone that contained several cracks to a of 3.2 mm and a wall thickness of, nominally, 0.3 mm. The
maximum subsurface depth of 10 m. Clearly, this integration of the fuel lines into the spacecraft had necessitated
air-contaminated surface layer had not been correctly much ground handling and transportation between manufac-
removedthe anomaly was reviewed with the forging sup- turing sites. The nal cause of tube failure was this trans-
plier. It became apparent that alpha-case formation had portation. The length was incorrectly (severely) loaded by a
occurred during the forging operation. The dimensional temporary xation device that held the complicated
design of the dies for these close-to-form precision forgings three-dimensional network of pipes onto a transportation
was incorrect and insufcient material was being left for the frame. Metallurgical examination of the titanium pipe in the
complete removal of the hard, alpha-rich subsurface layer. vicinity of the fracture revealed many problems and deviations
The hard alpha case is difcult to machine and may reach to a from the material specication AMS 4942 (see Fig. 5.84):
depth of 500 m. In this condition the forging can be pickled in
hydrofluoric acid (24 % aqueous solution) and material will be large hydride precipitates lying on slip planes and along
removed at a rate of about 25 m per minute. To minimize grain boundaries
hydrogen pick-up from the hydrofluoric acid bath, the solution a large grain size (ASTM 6 instead of the specied 710)
is usually modied with the addition of about 30 % nitric acid an elongation at fracture of only 18 % (specied as 20
this solution has a reduced layer removal rate (about 10 m per 32 %); importantly, the tube supplier had tested and
minute), but is essential if further problems associated with found this to be 32 % prior to delivery
hydrogen contamination are to be avoided. 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
5.18.3 Titanium Hydride Embrittlement
These ndings, together with additional information from
The problem of hydrogen contamination in titanium alloy the company producing these fuel lines, conrmed that the
weldments was discussed in Sect. 4.4.4, where it was associ- titanium tubes had been degraded by the application of
ated with the wrong choice of welding rod material. The incorrect processing procedures. The damage was irreversible
potential for hydride formation in Ti6Al4V when using unal- and as the tubes could not be reclaimed they were rejected as
loyed welding rods was probably rst recognized in the being unworthy for flight. The causes of the problem were:
Apollo-era as discussed by Ecord (1972) in his summation of
propulsion system failures occurring in that period. Because of this batch of tubes had been kept in an uncontrolled
the small solubility of hydrogen in alpha titanium the dominant storeroom for many years and had developed a thick
embrittlement of pure titanium and the alpha alloys is caused oxide coating
by the precipitation of titanium hydride. The beta phase of this coating had to be removed to ensure a suitable sur-
titanium can absorb larger volumes of hydrogen (up to 0.5 %) face nish for tube-to-tube TIG welding and to avoid
and generally all that happens is a distortion to the crystal lattice contaminating the spacecrafts liquid propellants
of the beta phase. Hence, unless severely exposed to hydrogen, chemical cleaning (pickling) had been too aggressive,
the alphabeta alloys such as Ti6A14V are seldom affected by removing excessive material from the surface and load-
embrittlement, as all the hydrogen contaminant is absorbed by ing the titanium with hydrogen
the beta phase. It would seem from the literature that titanium a further heat treatment had precipitated the hydrogen as
hydride phases can occur, one of which is dependent on the titanium hydride
presence of a tensile stress [and can decompose when later the presence of surface notches and hydride platelets at
exposed to a compressive stress (Nakasa and Liu 1993)]. the tube surface caused the initiation and growth of
Hydrogen can contaminate titanium from many sources: cracks resulting in breakage of the tube (the hydride
water vapour, pickling solutions, electroplating, chemical pol- platelets are the sources of Grifth cracks, which develop
ishing, the presence of surface lms such as grease during heat because of the poor cohesion between the precipitates
treatment operations, etc. Once hydrides nucleate they appear to and the matrix and the difference in their elastic modulus
grow rapidly and deplete the matrix of interstitial hydrogen. and plasticity)
5.18 Defects in Titanium Piece-Parts 325
calculations showed these stresses to have reached a value of heat-treated AA 7020 samples had given extra condence that
190 MPa in the short transverse direction of the formed these new tanks would possess an improved resistance to
plates. This is about 68 % of the AA 7020 materials yield SCC. No further SCC problems were encountered on Ariane
strength (the 0.2 % off-set yield strength = 280 MPa). launchers during subsequent years and, after many successful
Laboratory SCC tests (see also Sect. 4.6) made on AA flights, the water tank design was changedso as to achieve
7020 welded sheet material showed that, even in the long appreciable weight saving, these tanks are now fabricated
transverse direction, this alloy is prone to SCC in damp from CFRP material.
(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 5.20 Compatibility of Liquid and Solid
the welded material as having a low resistance to SCC (i.e. Propellants with Components
according to ECSS-Q-ST-70-37 this is a Class III material and Subsystems
and is unsuitable for spacecraft structural applications).
Corrective actions were to scrap the existing water tanks. An overview of materials considered to be compatible and
New tanks were manufactured (at that time, from the same incompatible with liquid propellants, based on evidence
alloy) with a modied welding technique that resulted in going back to 1965 is offered in the standard
lower residual stresses. Also, the new as-welded tanks were ECSS-E-ST-35-10. This standard covers the complexities
submitted to a nal stress-relieving heat treatment. The results associated with compatibility testing of liquid propellants
of laboratory SCC tests performed on welded and then with materials. It is probably easier to list materials with
5.20 Compatibility of Liquid and Solid Propellants with Components and Subsystems 327
6.1 Material Problems Removable ones are classed as H (resistance less than 0.1 )
or S (resistance less than 1.0 ), these resistance measure-
There are several classications of processes which will ments can be met with chemical conversion coatings such as
metallurgically join parts as utilized in the production of Alodine 1200. The aluminium alloys of the 3-, 4-, 5- and
electronic circuitry. These interconnection techniques are 6000 series can easily meet the H and S goals, but the 2000
usually formed by the application or generation of heat, and, and 7000 series of alloys may have difculties if subjected to
depending on the temperatures attained, are generally any humidity. All aluminium alloys can be plated with
referred to as soldering, brazing, or welding. Where once nickel (magnetic) to achieve bonds having less than
dissimilar metals have been in contact, the joining temper- 0.0025 and these meet another class of bond, classed as R.
atures produce a continuous volume of metal or metal alloy Vibration and g forces encountered during the launch
that is intended to enhance the electrical properties of the phase may produce open circuits as a result of the de-mating
joint and preclude interfacial corrosion. of the contact surfaces of connectors, and the effect of fret-
Interconnections that do not involve metallurgical reac- ting or mechanical fatigue of the interconnection materials
tions or the formation of alloyed interfaces are termed sol- may result in wire breakages. Once the spacecraft is in orbit
derless, mechanical, or physical joints and include the interconnections will be subjected to the thermal excur-
operations such as crimping, wire wrapping, and screw sions described in Sect. 2.3; the extremities of these excur-
fastening. sions will depend on whether there is a spin velocity. For
An incorrectly chosen method for making interconnec- stationary, non-spinning spacecraft, such as the Space
tions on any spacecraft may cause equipment electrical Telescope, interconnectors situated on surfaces exposed to
interfaces to degrade in several phases of their operating the Sun may be subjected to quite high temperatures
lives. Between assembly and launch, it is important that the (c. 150 C) as a result of solar radiation. High temperatures
interconnections should be able to withstand certain mildly may also be expected in the proximity of power sources and
corrosive environments. They must possess chemical resis- where heavy electrical currents have heating effects. In such
tance to the acids secreted from an operators skin and to locations the use of low-melting-point solders would be
cleaning agents and the trace impurities that may be dis- restricted. Special electronic circuits will operate at cryo-
solved in them. Resistance to atmospheric corrosion in genic temperatures, such as those being designed for the ISO
marine and industrial regions is critical when dissimilar spacecraft when electrical interconnections must operate for
metals have a high galvanic incompatibility (see also Sect. 2. several years at liquid helium temperature (269 C). Cables
2 and Tables 2.1 and 2.2). of superconducting NbTi may also be spliced with leadtin
When payloads such as electronic housings (black solder and are expected to possess good mechanical strength
boxes) are mechanically attached to the spacecrafts struc- and a low voltage drop at 4.2 K.
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 6.2 Welded Lead Wire Interconnections
metals Table 2.2 should be consulted) and these can be
painted black/white or anodized to ensure against general Resistance spot welding is an automated process in which all
corrosion. The contacting attachments to the structure are process variables must be preset and maintained constant.
often removable (screw fasteners) but can be permanently This is necessary becauseonce a weld has been made
xed. There are some general rules for the electrical bonds. present, nondestructive methods of inspection are difcult to
made under microprocessor computer control during the Occasionally, welded parts may contain intermetallic
assembly of electrical components into a cordwood module. compounds at the interface between dissimilar metals. These
The monitoring system assess the thermal expansion of the intermetallics are generally of a very brittle nature, and, once
weld nugget and surrounding material during heating and they reach a critical volume, will jeopardize the welds
melting. Welding at a relatively high current level produces resistance to impact or mechanical shock loading. Although
a rapid grain growth across the interfaces, and some liquid is undesirable, it may be impossible to avoid metallic combi-
pushed out from the weld zone itself, culminating in a rapid nations which produce intermetallics, and in these instances
weld electrode contraction. Welding at a low heat setting efforts must be made to restrict their rate of growth by opti-
produces too small a spot weld size and an inadequate weld mizing welding schedules to short times at the lowest pos-
strength under tension-shear. sible temperatures. It may also be necessary to limit
332 6 Failure Analysis of Electrical Interconnections and Recommended Processes
Fig. 6.4 An evaluation of weak weldments between the tantalum wire wire lead and its nickel wire extension on tantalum foil capacitors.
lead and its nickel wire extension on tantalum foil capacitors. a General d Detail of weld zones on (c). Nickel wire has recrystallized, a dendritic
view of capacitor after lead bending. b Fragile lead that fractured during structure containing Ni, Sn, and Ta has been expelled, and a white
rst bend through 90. c Section through butt-welded lead after etching intermetallic phase (Ni3Ta) exists at wire interfaces. e Some welds
in nitric acid. An evaluation of weak weldments between the tantalum possess stress-raisers around their circumference (arrowed)
oxides from the contact surface of the aluminium, so that vary, but one typical example of an open transistor circuit is
strong bonds occur only where the oxide has been removed presented in Fig. 6.5. Many papers have been written to
(Xu et al. 2009) and it is at this intimate interface between explain the characteristics of goldaluminium interconnec-
wire and aluminium that forms a metallurgical bond. tions, and just about as many conflicting conclusions have
During bonding, and even at temperatures within the been reached. The most noteworthy set of papers reviewed
range to which assembled micro-circuits are sometimes by the author were presented as long ago as 1965 at the
exposed, the gold and aluminium undergo mutual solid-state Western Electronics Show and Convention in San Francisco
interdiffusion to form a number of intermetallic compounds. (Wescon 1965). It seems that there is a great danger in
If the bonding temperature/time prole is excessive, or extrapolating the conclusions drawn from one component
operating temperatures of the nished device are too high, type to another, as each can have different conditions which
thick layers of interfacial intermetallic can grow and these promote variations in intermetallic phases, kinetics of
conditions have been observed to promote open circuits and intermetallic formation, effect of silicon, and the strength of
component failures. The mode of failure has been seen to 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 SEM view of the same device shows the plague to have a ne
attached to aluminium metallization on the silicon wafer. The wires are crystalline appearance. a Optical view of transistor, 80. b Detail of
attached by a thermo-compression bonding technique. b Purple Purple purple plague, 200. c SEM reflected mode appearance of
plague (arrowed) is formed at the goldaluminium junction and same bonds, 300
spreads (migrates) along the aluminium at elevated temperatures. c The
gold wires to aluminium metallised pads has been ne-tuned ball bond and a sharp edge in the zone of Au2A1 compound
such that bonding times have been reduced to 210 ms with which surrounds the inner circumference of detachment
a substrate temperature of 175 C. A recent interesting (Figs. 6.6 and 6.7). The intermetallic zone was investigated
investigation (Xu et al. 2010) concluded that during the few by actually microsectioning these minute ball bonds. The
microseconds of the bonding process the initial metallurgical packages of transistors similar to the failed device were
bond occurs at the peripheral regions of the gold ball. The opened, and all conductors were prepared by electroplating
dimension of the contact area increased rapidly from the their surfaces in the laboratory with a protective layer of
circumference and slowed as the bond progressed to the copper. With meticulous care several bonds were micro-
centre of the circular bond area. sectioned and polished. These details are shown in Fig. 6.8.
The electronic device depicted in Fig. 6.5 is typical of a The schematic diagram in Fig. 6.8c details two distinct
component which was known to have a poor reliability. An mechanisms postulated to initiate the failure of these bonds:
investigation was initiated using traditional metallographic
techniques in order to evaluate its most prevalent mode of the void and notch formation due to mass transfer, and
failure. the notch-initiated microcracks along strained interfaces
The author believes that the appearance of purple pla- due to lattice mismatch between the intermetallics and
gue, as shown in Fig. 6.5, is in fact due to the optical effect silicon substrate (particularly during thermal cycling).
created when white light is diffracted from the extremely ne
network of intermetallic crystals surrounding the bond. It is understood that three intermetallics may occur in the
Under SEM examination (Fig. 6.5a) these crystals can be bond region, AuAl2, Au5Al2, and Au4Al, as shown by the
resolved as having a very rough, irregular surface appearance phase diagram reproduced in Fig. 6.8d. The microcracks are
with a composition of AuAl2 (see also caption of Fig. 6.6). thought to separate the gold-rich intermetallic phases, and, to
The investigation demonstrated that associated with the establish if this interface was also brittle, microhardness
purple plague formation, and lying hidden underneath the impressions were carefully placed in this region of the bond.
bond, is another compound of gold which is very brittle and Sets of hardness impressions are shown for one particular
possesses a low adherence to the silicon substrate (as bond in Fig. 6.8e, f. They were noted to induce slight cracks
depicted in Fig. 6.6). This intermetallic (Au5Al2white only at the intermetallicsilicon interface and inter-metallic
plague) has been shown (Bushmire 1977) to increase the gold interface. Incidental information determined the hard-
electrical resistance of bonds signicantly as a function of ness values of the constituents as:
time and temperature in excess of 150 C. The migration, or
Kirkendall mass transfer effect, of gold into the aluminium Silicon 1690 MHN 5 g
metallization away from the goldaluminium junction is Gold-rich intermetallics 385 MHN 5 g
seen to leave microvoids around the periphery of the gold Gold 59 MHN 5 g
6.3 Purple Plague 335
Fig. 6.6 Details of failed transistor thermocompression bond. The a brittle intermetallic compound (b); the insert highlights the Kirkendall
rough crystalline circle of purple plague. a is noted to possess a sharp microvoiding effect associated with the purple plague formation
well-dened inner edge (see insert). Lying underneath the gold bond is
Fig. 6.8 Evaluation of the intermetallic systems associated with sketched schematically in (c). Evaluation of the intermetallic systems
purple plague. a In order that thermocompression bonds could be associated with purple plague. c Schematic diagram of observations
microsectioned they were initially plated with copper in the laboratory. made during inspection of microsectioned bonds. d Phase diagram of
b Vigilant metallography reveals the thermocompression bond and the the goldaluminium system. Diamond pyramid hardness impressions
goldaluminium interaction layer. This sample was lightly stained, and made under different loads in selected regions of thermocompression
microscopical inspection could discern between two slightly different bond. e Shows optical micrograph, and f SEM image of same area
coloured phases within the intermetallicthese observations are (appears as mirror image)
that utilize Focused Ion Beam milling and Ion Beam milling may have become smeared or covered by gold and therefore
as illustrated by Fig. 6.9 (Erdman et al. 2006) and discussed appear to be less extensive as those seen in Fig. 6.9.
in Sect. 3.2.2.2. These new methods provide the same data A failure of this kind can be detected only by in situ tests,
as grinding, polishing and etching specimens, but there is which must be designed to simulate the calculated
no surface damage and evidence of cracking is much clearer thermal/vibrational environment at a particular location
to observe. For instance, the microcracks seen in Fig. 6.8 within, or external to, the spacecraft structure.
6.4 Mechanical Electrical Connections 337
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.
usually employed on AWG 22 and smaller wire sizes, as Quality controls will include visual examination for
there is a danger that mechanical-type strippers can stretch absence of overlapping wires or gaps between adjacent turns
these ne-gauge wires. The heat produced by such stripping and absence of pin deformation; electrical contact resistance;
tools must be regulated to avoid blistering and excessive tensile pull-off strength; an unwrapping test to check that the
melting of the insulation, otherwise a thin lm of insulant wire has not become embrittled; and a gas-tightness test that
could be drawn over the conductor so as to degrade the subjects wrapped connection samples to a specied chemical
subsequent wire-wrapped joint. Large-diameter wires are attack. The sample illustrated in Fig. 6.11 had been exposed
usually stripped by means of precision automatic to an atmosphere of hydrogen sulphide (H2S); the exposed
power-driven strippers. Handheld precision strippers with area of the connection (both wire and terminal) has become
accurately machined and factory-preset cutting heads may be discoloured. The gas-tight area of the connection is exposed
used, with a caution to ensure that correct die openings are when the uninsulated wire is carefully unwrapped from the
used. Operator inspection steps must check that the plated terminal. Microsections of the as-made connection, as shown
nish of the solid conductor is not scored by this operation, in Fig. 6.10, allow further assessment of joint integrity. They
and that the wire is neither stretched nor twisted. show the depth of corner penetration into the wire and
A variety of commercial wrapping tools are available for illustrate how the post corners may have deformed, thereby
interconnecting panels and backplanes constructed of glass indicating the suitability of the post for any subsequent
epoxy PCB material supporting up to several hundred repair operation. Once the electrical reliability of the
wire-wrapping pins. Spacecraft wire-wrapped connections process-controlled wire-wrapping operation becomes estab-
consist of at least six turns of wire around the post for AWG lished, only visual inspections and the tensile pull-off
26 and larger, and seven turns for thinner wires. An addi- strength should be necessary as in-line quality sampling
tional turn of insulated wire around the bottom of the ter- tests. Each nished panel should be electrically inspected
minal post is also required, as this adds a certain amount of immediately after it is produced in order to verify the
mechanical stability to the wire and helps to resist vibration. accuracy of wrap locations and to detect hidden defects.
6.4 Mechanical Electrical Connections 339
cold-weld and stranded wires become very rigid. Nickel between soldering and brazing is to assume that the lower
coatings, being harder, will slide over each other. However the temperature soldered joints are nonpermanent and are of
disadvantages of nickel must be taken into consideration: it is little strength, whereas brazed joints are designed to be both
strongly magnetic, it gives an increased resistance at elevated permanent and strong.
temperatures, and is not suitable for low-voltage applications. Usual purchase requirements for the composition of
space-grade tin 63lead 37 solder are as follows:
Fig. 6.14 Details of microsection shown in Fig. 6.13. a Detail of (arrowed). c Detail of poor plating. d Wire strand-to-strand interface
highly deformed wire strand-to-barrel interface. A very small gap (2.0 showing good interconnection. The strand plating is 3.5 m nickel
0.5 m) exists along this interface. b Possibly point contact at this which is seen to be well adherent and continuous around all strands
strand-to-barrel interface, but it is in region of detached plating examined within the crimp connection
the mounting of electronics parts onto PCBs or by operators soldered interconnections which were discovered during the
making wire harnesses for power supplies. Before the issue course of constructing more than ten different spacecrafts.
of the ECSS specications contractors had followed their Some of these ndings are presented in the following
own in-house soldering specications which, when sections.
reviewed, were found to be extremely scant, while it was Soldered assemblies were subjected to a 100 % visual
apparently assumed that the process was capable of being inspection at up to 7 magnication. In these instances, it
carried out by even the most inexperienced personnel. was essential that workmanship standards, in the form of
The European standards for soldering spacecraft elec- work samples or visual aids, were readily available to clearly
tronics (ECSS-Q-ST-70-08, 70-26, and 70-38) presuppose illustrate the quality characteristics of the soldered connec-
common PCB designs, materials, procedures, operator tions. Acceptable solder connections are generally charac-
training courses, workmanship, and inspection methods. It terized by:
was considered desirable to support this standardizing effort
by initiating a comprehensive solder-joint test programme, (a) clean, smooth, undisturbed surfaces,
which would investigate semi-quantitatively the failure (b) concave llets between solder and joined surfaces,
modes of space-quality joints. In addition, an effort would be (c) complete wetting as evidenced by a low contact angle
made to reveal and understand the failure and problem between the solder and the joined surfaces.
6.5 Soldered Interconnections 343
Unacceptable solder conditions which may be cause for respective assemblies with a low-speed,
rejection are: diamond-impregnated rotary saw. The samples were mounted
in a resin curing at room temperature, sometimes in a vacuum
(a) damages, crushed, cracked, melted, corroded, etc., impregnation apparatus. The cured mounts were hand-ground
surfaces, on abrasive SiC papers. It was essential that the specimen and
(b) improper tinning, papers were frequently cleaned to remove loose particles of
(c) flux residues or other contamination, sintered ceramic and abrasive that might later become
(d) cold joints, embedded in the specimen. To obtain satisfactory nal results
(e) fractured joints, it was found essential to use diamond pastes for polishing.
(f) pits, holes, or voids in the joint which are not attribu- Alumina polishing compounds were found to leave steps on
table to liquid-to-solid solder shrinkage, the sample surface. Initially, diamond paste having a 6 m
(g) excessive or insufcient solder, particle size was used for polishing. A napless cloth was used
(h) splattering of flux or solder on adjacent areas, for the rst stages to remove the unidirectional scratches from
(i) dewetting, etc. the nal SiC paper. After a few minutes the specimens were
transferred to 1 m, then 0.25 m diamond pastes which had
The preparation of soldered joint specimens for metallo- been applied to short-nap cloths. Each polishing stage was
graphic examination presents problems that are not common kept to a minimum time to reduce the formation of a step, or
to the preparation of most other materials. This is because bevel, which would have made examination of the interfaces
low-temperature solder alloys are so soft that they deform even very difcult. Often the microstructure of solder alloys could
under low pressures. Also, these alloys have low recrystal- be distinguished in the specimens as-polished state. The tin
lization temperatures, so that normal Bakelite hot-pressure lead and indiumlead alloys could be etch-polished (i.e.
mounting of samples is not feasible. The microsections that are alternately etching and polishing until a bright, lustrous,
presented in the following sections contain materials of vastly scratch-free surface was obtained).
differing hardness (e.g. indium-to-sintered alumina; tin The following etchants were employed (see also
solder-to-Kovar; and joints containing hard intermetallics set Appendix 3):
in soft matrices), so that special metallographic preparations
were required. Generally, the joints selected for microsec- (a) Diluted ammonium hydroxide with a few drops of
tioning are coated with a low exotherm, reasonably hard resin 30 % hydrogen peroxide, for tinlead solder on copper
to prevent disturbances as they were carefully sawn from their 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 Thermal fatigue cracking is caused by the stresses
acid) for darkening tinlead and enhancing the repeatedly imposed on solder joints during thermal cycling,
appearance of intermetallic compounds. when free thermal expansion is part or entirely restrained.
(c) Mixed acids (10 cm3 HNO3 + 30 cm3 acetic The magnitude of these stresses will vary according to the
acid + 50 cm3 glycerine) to show up the general thermal expansion coefcients of the different materials
structure of tinlead alloys. 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
6.5.3 The Effect of Thermal Fatigue because of poor workmanship, were subjected to three
on Solder-Assembled Leaded phases of extensive thermal cycling. This culminated with
Components cycles between +100 and 100 C. The eutectic tinlead
solder alloy was of high purity and contained less than
Several solder-joint studies undertaken in the late 1960s and 0.05 % trace elements. A full description of the test condi-
early 1970s into the reliability of conventional electronic tions and the materials employed are given elsewhere (Dunn
components assembled onto PCBs by soldering had shown 1979). A sketch of each joint design studies is shown in
conclusively that thermal fatigue was the primary cause of Fig. 6.18. The joints were visually and metallograpahically
joint failure. The mode of thermal fatigue failure was inspected at various stages throughout the thermal cycling.
exacerbated by the fact that the samples were made before As the joints visual appearance deteriorated with
the existence of the relevant space-quality materials and cumulative thermal fatigue, a record was kept of those joint
process specications. In addition it was later shown that the llets which could be described as bearing surface discon-
solder alloy contained impurities which might embrittle the tinuities pertaining to a crack. This rather loose description
soldered joints. The thermal fatigue failures are characterized could be substantiated only by destructive metallographic
by the optical and metallographic pictures shown in examinations. The summarized results are shown visually in
Fig. 6.17. The mode of electrical open circuit failure is Fig. 6.19. A number of photomicrographs are presented in
depicted in Fig. 6.17d. Simulated launch environments, Figs. 6.20 and 6.21. For each gure the caption highlights
including mechanical shock and vibration, were found, on the signicant features. In the absence of brittle intermetallic
the other hand, to have little effect on the appearance and phases, crack propagation is noted to progress
quality of solder joints.
Fig. 6.17 Stud-mounted resistor from thermal fatigue programme cracking of PCB. Intermetallics are just visible along coppersolder
using commercial-quality PCBs and solder alloy (solder contained interface (between arrows). d. Non-approved PCB manufacturers
0.5 % antimony, 0.1 % copper, 63 % tin, and remainder lead). a View products may promote thermal fatigue in solder llets, as seen in (a) to
of circumferential fatigue crack on llet. b Microsection reveals that (c). The mechanical integrity of the PCB is important, particularly at
virtually no plastic deformation occurred on solder llet, 125. c Detail interfaces between laminate-copper and plated through copper (from
from (b) showing that crack propagation in solder initiated by corner Dunn 1979)
intergranularly around the tin phases (white islands) and lower than the values obtained in standard tests involving
transgranularly through the lead phases. higher strain rates.
The relationship between the physical changes induced in The initial thermal fatigue to which each PCB is subject
the various solder joint materials and the characteristics of occurs during immersion in a hot oil bath, which heat-cycles
joint failure resulting from fatigue has been studied. It was the boards in a controlled manner to approximately 205 C
found that a solder joint is unlikely to fail because its ulti- to fuse the tinlead plated circuitry. This has been described
mate tensile strength has been exceeded, but that failure is in Sect. 4.7. This cycle is followed by a further heat input to
most likely to be caused by an accumulation of large cyclic the plated-through hole during component assembly to the
strains, which result from differences in the expansion board, when soldering iron temperatures are generally
coefcients of the constituent materials. The strain rates around 250 C but are permitted by assembly process
imposed by this form of thermal fatigue are very low, and, if specications to reach a maximum of 320 C. Space-quality
the mechanical properties of bulk materials are evaluated at PCBs are not expected to be damaged by these
corresponding strain rates, their strengths are found to be far high-amplitude short-duration thermal cycles, although
346 6 Failure Analysis of Electrical Interconnections and Recommended Processes
Fig. 6.18 Various component boards of lower quality may be degradedas shown in
joints subjected to thermal Fig. 6.16by the formation of cracks in the corners of
fatigue. a Clinched lead.
b Stud-mounted lead. c Off-set plated-through holes.
lap joint. d Lap joint 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
It is probable that this large calculated displacement does For a rst approximation, the dimensional mismatch
not take place fully during the soldering of actual termina- between the PCB and each component lead can be calculated
tions because of the restraining effect of the plated-through for the temperature extremes chosen for the various phases
copper. It is evident, however, that such copper deposits of the programme. Mismatch of the component and the PCB
ought to possess sufcient elongation and ductility to pre- in the directions parallel to the surface of the PCB are not
vent corner cracking problems. considered, in view of the relatively compatible values of
The various forms of soldered joints that were subjected coefcients of expansion between the PCB in the x and
to the different phases of thermal cycling during this pro- y directions and those of the various lead materials. Also, as
gramme are shown in Fig. 6.18. In view of the difference in shown in Fig. 6.18, each joint design will allow for sufcient
coefcient of expansion values between the PCB material stress relief to compensate for such minute lateral
and the component lead materials, each solder llet will have movements.
been stressed to an extent dependent on the actual design and The greatest problem during thermal cycling results from
material composition of each joint. The true stresses within relative motion between the component lead material and the
solder llets are, however, extremely complicated to assess, plated-through hole. This motion is caused by the high
owing to metallurgical effects which take place in the solder values of coefcient of expansion in the thickness, or z axis,
alloy. At different temperatures, the competing effects of of the PCB. The amount of mismatch, dl, is calculated from
strain hardening, time-dependent thermal softening, and the expression
creep will each have a signicance diverging from the actual
mechanical properties of the alloy. d1 aPCB alead t=2DT
Fig. 6.20 Solder joint, after 800 cycles. The component side (a) of this within the joint. There are no defects associated with the clinched
joint shows two circumferential discontinuities related to the hole in the portion of the joint, but the component side (c) is seen to possess real
PCB and to the area of solder adjacent to the component lead. cracks, which have propagated a distance of 400 m alongside the
Metallography has revealed (b) the presence of a medium-sized void 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 joints shows extensive cracking alongside each of the leads f. The
on each of its three soldered leads (a shows lead No. 1; b shows lead detailed micrograph g reveals that these cracks are associated with the
No. 2). Microsection c shows the position of a transipad tp. brittle goldtin intermetallic AuSn2, which results from inadequate gold
Examination of lead No. 2 (d and e) reveals that these cracks have removal from these leads by the operator before component assembly
propagated from the llet surface into the solder at an angle of 45 to onto the PCB. a 20. b 20. c 13. f 300. g 1350 (oil immersion
the Kovar lead material. Inspection of the component side of the solder 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
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 The following observations can be drawn from the ther-
solder as a result of plastic shear will assist in solid-state mal fatigue studies;
diffusion causing this band of material to coarsen as thermal
cycling proceeds. Initially the ne-grained, equiaxed solder (1) The number of solder joint cracks on space-quality
will undergo inter-crystalline sliding at the leadtin inter- soldered joints was smaller and their severity was less
phase boundaries, and this creep-like process has been than in the case of similar studies performed on PCBs
observed to produce superplastic extensions to a few thou- made according to commercial-quality procedures.
sand per cent before fracture occurred (Raman and Reiley (2) 360 C cracks on solder llet identied by visual
1987). Decohesion at the phase boundaries does not occur inspection are often termed failures. This is consid-
unless certain impurities, such as aluminium, gold, tungsten, ered inappropriate unless the visual inspection is sub-
and copper, are present within the grain boundaries. High stantiated by metallographic or electrical inspections.
concentrations of contaminants can contribute to cracking Visual inspection alone can lead to misleading results.
problems during thermal cycling, leading to a reduced Continual plastic deformation of the solder llet during
interconnect conductor area and degraded electrical perfor- thermal cycling causes surface discontinuities that may
mance (OClock et al. 1987). The onset of grain coarsening be described as folds, ridges, orange-peel effects, and
will further concentrate subsequent deformation at the wide-mouthed cracks.
coarsened region, and after several hundred thermal cycles (3) No interconnections were considered to have failed at
the large grain size of the solder will no longer be able to the end of this programme. Offset lap joints, as used for
accommodate the imposed strain by grain boundary sliding, mounting transistor leads, and the lap joints used for
and this might result in the formation of intergranular flat packages suffered no cosmetic or metallurgical
cracks. The 1000 thermal cycles experienced in the authors degradation. Solder cracking on clinched leads was
solder joint study showed no evidence that cracks in the conned to the bend radius of the clinch, and no sur-
coarsened region would eventually lead to ultimate failure. face changes were observed along that portion of lead
This fact is attributed to the presence of stress-relief bends in lying parallel to the PCB conductor. Solder cracks
the component leads and the selection of a very-high-purity were present around the llets of stud joints, but they
eutectic solder alloy containing less than 0.05 % trace were generally wide-mouthed and lacked appreciable
elements. depth.
6.5 Soldered Interconnections 351
(4) Grain boundary sliding of the tinlead solder alloy size of the chip carrier is increased, causing carriers to break
provides for stress relaxation and a creep-mechanism- away from the board. For this reason it is better to use mixed
based possibility on lattice self-diffusion. However, if types of surface-mounted packages, i.e. J-lead components
brittle intermetallic phases, such as goldtin, are pre- and ne-pitch, high-lead-count, gullwing Very Large Scale
sent in the joint, then crack propagation is through both Integration (VLSI) carriers rather than solely leadless
the intermetallic and the lead-rich phase. components.
(5) Hairline cracks, which are potentially dangerous, were Multilayer ceramic capacitors are the most widely used
occasionally observed to run alongside the Kovar or surface-mounted components in spacecraft electronics. They
copper-clad Kovar. They were not associated with are also very brittle and must be solder-attached to the circuit
copper lead materials. by methods that will not cause thermal cracking of the
(6) Non-approved mounting methods, i.e. soldering of ceramic near to the metallized termination area. Preheating
gold, use of poor solder-ability lead materials, etc., will of the part and substrate is a critical factor in avoiding
produce substandard joints which may lead to prema- microcracks in chip capacitors such as that illustrated in
ture electrical failures. 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
6.5.4 Effect of Thermal Fatigue on Leadless 55 and +80 C. Thermal cycling propagates the initial
Components thermal shock crack in the vertical direction and will even-
tually split the capacitor in half.
Spacecraft electronics generally comprise the leaded com- Evaluation programmes have been performed using dif-
ponents evaluated under thermal fatigue in the previous ferent solder alloys and various stand-off heights (i.e. dis-
section. These conventional devices are mounted onto PCBs tance between termination area on underside of component
by inserting their leads or package pins into plated-through and that of the substrate) to achieve suitable connections
holes in the PCB. Flat-packs with leads emerging from the between chip capacitors and a highly thermally mismatched
package sidewalls have been mounted by lap-soldered joints substrate (Teflon PCB in intimate contact with an aluminium
to the board surface. The survival rate of these devices is base plate). 75In25Pb solder was found to be slightly more
high because the formed leads can easily absorb stresses and reliable than 63Sn37Pb, followed by a 2 % silver-loaded
the solder alloy provides a reliable connection to the circuit. tinlead solder. A satisfactory stand-off height was deter-
Microcircuit assembly has gained full acceptance for mined to be 0.2 mm. Laboratory tests, based on mechanical
spacecraft applications; here the components and electronic straining a wide variety of solder lap joints, were made to
packages are surface mounted to the PCB by means of pin-point preferred designs for SMT assembly (Adile and
soldering directly between the component body and the Dunn 1999); they concluded:
circuit termination pad. In mounting these modern leadless
devices the solder is the only load-bearing member, and it large solder llets are more reliable than small llets (all
must absorb stresses as well as provide electrical intercon- joints having same prole),
nection. Whereas ceramic chip capacitors and resistors are a stand-off of 0.2 mm is better than either very small or
commonly selected for spacecraft applications it is the large stand-offs,
ceramic chip carrier and area grid arrays that provide the signicantly longer life-times were seen when fatigue
most benets for space-bound electronics. The chip carrier testing was made in an inert atmosphere (argon), possibly
provides a safe package for sensitive integrated circuits and similar to the observation for testing under vacuum with
offers a greater than 10 to 1 reduction in volume over present no oxidation of the propagating crack and the occurrence
duel-in-line package devices. The low mass of the chip of cold welding.
carrier improves mechanical shock and vibration resistance,
and its high internal component density requires far less The importance of increasing the compliance between a
PCB surface area, provides a lower inductance, and leadless device and its substrate by incorporating a suitable
improves switching times. stand-off can also be seen from Fig. 6.24. Even a small
The stresses in leadless surface-mounted devices must be amount of solder in this gap can considerably increase the
absorbed by the solder. The origin of stress is due either to number of cycles to failure. One specication requires from
mismatch in thermal expansion between board and carrier, 0.1 to 0.4 mm stand-off (ECSS-Q-ST-70-38). Such a gap
or to thermal stressstrain loading on the joints by power will facilitate the cleaning of flux residues from beneath the
dissipation. As expected, this problem is increased as 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 quickly than the ceramic dielectric). b Under thermal cycling the type
interface as a result of thermal shock during soldering (the capacitors of crack seen in (a) can propagate under thermal or power cycling to
metallized terminations and electrodes heat up and expand much more cross the silver electrodes
thermal fatigue. Solder llet heights were found to be more increasing the stand-off height,
reliable when solder extended to a distance of 0.4 mm from increasing the ductility of the solder alloy.
the base of the chip termination. Evaluation of mounted Area
Grid Array packages, using CSAM and electrical resistance The rst solution to avoid thermal mismatch is to select
test monitoring, during thermal cycling also showed that ceramic substrates having the same composition as the
cracks propagate faster in low stand-off solder joints than in ceramic package. A second solution is to utilize advanced
the high stand-off solder joints (Liu and Lu 2003). When PCB materials. Extensive testing (Houlberg 1988) estab-
solder llets have unequal volumes, as shown in Fig. 6.25, lished that the following matched coefcient substrate
the total strain resulting from ceramic-PCB expansion mis- materials were suitable for space use:
match is taken up by the smaller llet.
The CofnManson model has been used by several (A) Epoxy/glass with two CuInvarCu cores. The core
workers to determine the fatigue life of leadless ceramic chip thickness is 0.254 mm and the ratio of CuInvarCu is
carriers soldered onto PCBs, and the reader is recommended to 12.5, 75, 12.5 %.
consult the original paper of Lau and Rice (1985) for an account (B) Polyimide/glass with two CuInvarCu cores as above.
of the model predictions, i.e. fatigue life is increased by: (C) Polyimide with quartz bre reinforcement.
(D) Epoxy with Kevlar bre reinforcement.
reducing the size of chip carrier,
decreasing temperature cycle excursion, The PCB substrates A, B and C, used for early European
decreasing coefcient of expansion mismatch, space projects, are no longer on the commercial market and
6.5 Soldered Interconnections 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
Fig. 6.24 Graphs to illustrate the importance of having an adequate surface-mounting technology is rapidly progressing to
stand-off height, or sandwich of solder, between the metallization of include large ne pitch devices as well as large Area Grid
the component and the multilayer board (MLB) termination pad during Array packages, having both ball and column terminations,
temperature cycling. These plots have been drawn to show the
predicted mean number of temperature cycles-to-failure for leadless
many issues related to reliability, processing, and availability
ceramic chip carriers (0.850-inch square package) assembled by solder of space-approved components remain unresolved.
to three types of MLB substrate. This form of thermal fatigue occurs ECSS-Q-ST-70-38 and the IPC J-STD-001 and its Space
during power cycling, the nominal ceramic chip temperature is 25 C Addendum, have included important surface-mount
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
requirements. Some guidelines have been published con-
of thermal expansion are 5.5 ppm for the chip and the ceramic MLB, cerning the principles involved in the design, assembly- and
21.5 ppm for the epoxyglass MLB, and 8.6 ppm for the copper-clad testing for surface-mounted electronic components [many
invar MLB (Used with permission from Dr. Engelmaier 1995) based on the groundwork of Lea (1988b) and Hinch (1988)],
and some will be reviewed in Sect. 6.15.10.
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 6.5.5 The Effect of Thermal Fatigue
35 and 85 N. Other specialist laminates with matched on Semi-rigid Cable Connections
coefcients to ceramic component packages are
non-woven, short aramid bres embedded in either epoxy or Solder joint cracking problems associated with the construc-
polyimide (trade name Thermount, but this DuPont material tion of semi-rigid coaxial cables have plagued many space-
has been withdrawn from the market although large stocks craft programmes. Some cracks were due to poor handling of
are available with space-approved PCB manufacturers and it the preformed cables during integration onto the spacecraft,
is possible a license to formulate this material may be whereas others resulted from thermal cycling when the PTFE
transferred to another producer). A new PCB technology dielectric contained within these cables expanded and tended
utilizes a carbon bre pseudo-fabric weave, embedded in to protrude out from the cable ends. This protrusion can cause
resin and sandwiched by copper foilreferred to as carbon the soldered interconnections to become stressed, and suc-
core laminate with the trademark name, Stablecore. It is cessive thermal excursions may lead to solder joint cracking at
claimed to be superior to the aforementioned substrates one end of the cable and the formation of air gaps between the
(A) to (D) and Thermount. The benets of Stablecore PTFE and adjacent mating surfaces at the opposite end. The
include: ability to tailor CTE between 2.5 and 12 ppm; a problem is compounded when the conductors and mating
lower z-axis expansion than Thermount; is a good thermal parts possess gold plating, as this nish will form a layer of
dissipater due to carbon weave and can reduce (weight) brittle goldtin intermetallic when incorrectly soldered. These
amount of copper in ground planes; and, has extremely low factors have been addressed by Fletcher (1978), and it is now
moisture absorption. realized that strict process controls must be adhered to if the
At the present time (2014) more than 75 European degradation of cable interconnections is to be avoided.
companies and institutions are utilizing surface-mounting Additional examples are given in Sect. 6.15.7, and Fig. 6.77
technology on spacecraft. Soldering methods vary from hand reveals 5 mm of extruded PTFE!
soldering, vapour phase, and infrared soldering with solder Cross-sections made through defective joints are shown
pastes, to the reflow melting of specially constructed solder and explained in Fig. 6.27. It is thought that deep drawing of
columns around which are wound ne coiled wire that the copper cable sheath has created minute surface striations
354 6 Failure Analysis of Electrical Interconnections and Recommended Processes
Fig. 6.27 Section through connector-to-coax assemblies showing mismatch between the coaxial cable assembly and the unit. Assembly
cracked solder joints resulting from PTFE extrusion and resultant force will be forced into position, so stressing the solder joints. d Typical
on solder. See also Fig. 6.77. a Suitable connector with non-captive nut appearance of cracked joint due to either bending stress or PTFE Teflon
and good solder workmanship. b Non-approved captive nut; during extrusion (after thermal cycling of non-preconditioned assembly).
torqueing, stressing of solder joints can occur if inadequate allowance e Macrosection through cracked solder and revealing the extent of
has been made for torqueing distance (arrowed). Detail shows location PTFE dielectric protrusion from semi-rigid cable following thermal
of circlip that permits rotation of the coupling nut, but not retraction as cycling
photographed in (a). c Poor lay-up, when there is a dimensional
on the internal wallsthese cause ratcheting of the PTFE Selection of at least one cable-end connector with a
towards one direction as it expands/contracts during each noncaptive centre contact and coupling nut (to avoid
heating/cooling cycle. Correct processing of flight cables stressing during integration).
should follow the requirements of ECSS-Q-ST-70-18, when Final soldering using 96Sn4Ag solder alloy to ensure
the following critical steps must be adhered to: adequate joint strength.
6.5 Soldered Interconnections 357
(e)
Coax tube
Connector body
Solder joint
Preconditioning the formed cable by heat treatments (at provide surfaces which are better suited for soldering. Two
least three complete thermal cycles between 45 and types of metallic nish are employed for this purpose. They can
+75 C) with trimming of any protruding dielectric. be loosely categorized into surfaces designed to be soldered to
Degolding of all surface areas intended to be joined by and those that should be soldered through.
soldering.
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
lead. Such coatings are usually very adherent as they are surface. Oxidation or tarnish of surface zinc then prevents
applied by hot mechanical working operations. They usually solder spread, which cannot be promoted even in the pres-
give no problem during the thermal shock of soldering or ence of an activated liquid soldering flux. By specifying the
later during the mechanical loadings of service environ- presence of a barrier layer of 23 m of copper or
ments. However, if the coating is not well attached to the nickel-plating, to separate the zinc containing brass from the
substrate material by a sound metallurgical joint, electrical nal nish, the problem of poor solderability will be
failures may result from detached bonds, resulting in the removed. Solderability data recorded from brass samples
form of open circuit as highlighted in Fig. 6.28. with and without a barrier layer have decisively illustrated the
Thick coatings may also be applied to machined electronic value of using a barrier layer when poor solderability could
parts for other reasons. An example of a brass turret terminal cause a loss in joint integrity during service (Scott 1987).
is shown by the series of photographs presented in Fig. 6.29. The condition known as black pad has been considered
The solderability of tinned parts which have been fabricated in Sect. 4.10 and a transmission electron micrograph is
from brass is severely reduced by long storage. Zinc is known included in Fig. 4.50b showing the presence of small nickel
to diffuse from the substrate and onto the components free phosphide precipitates distributed within an electroless
6.6 Problems Associated with Coatings for Soldering Applications 359
must be taken to avoid thermally shocking these terminations mechanical or thermomechanical forces (i.e. during thermal
but activated fluxes must occasionally be used for oxidized cycling when components have a different coefcient of
parts that were procured many years before they are used. It is expansion to the substrate). The addition of only 0.2 wt%
recommended that such devices/packages are stored in gold suppresses this mechanism, and with 1 wt% of gold,
nitrogen-lled bags or that de-golding and pre-tinning are voids were seen to grow and promote cracking, leading to
performed at the earliest opportunity. early creep failures.
Gold removal from soldered gold-plated dipoles was Solder-plating with a tinlead alloy is probably the most
omitted during the assembly and electrical integration of an trouble-free coating which can be soldered through. It is
antenna system for a meteorological satellite. The antenna generally fused to eliminate porosity. Palladium is also a
was submitted to thermal cycling between 25 and +125 C useful coating material, but is very expensive and nds little
as part of its qualication testing. As a result of gold usage within even the highest-cost electronic system selected
embrittlement, many of the soldered joints were actually for spacecraft. Solder coatings have been presented previ-
heard to fracture by snapping within only a few days of the ously, as for example in Sect. 4.7, but it must be remembered
systems test completion. Metallurgical evaluation of these that even these coatings may become nonsolderable and
electrically open-circuit joints revealed the presence of a embrittled if they are subjected to high temperatures. This is
high volume of brittle AuSn4 needles that had formed during due to the formation and growth of two intermetallic layers,
the soldering operation (i.e. gold dissolution into the liquid Cu3Sn and Cu6Sn5, both of which are hard and brittle:
solder followed by precipitation of the needles during
cooling to room temperature). Figure 6.31 clearly shows this Cu3 Sn 400 Microhardness Number and 5 g load
fracture surface to consist of a continuous network of AuSn4. Cu6 Sn5 550 MHN 5 g
Experiments by Hannech and Hall (1992) duplicated the Copper 95 MHN 5 g
solid-state diffusion of gold (Fig. 6.31b, c) that occurs during Solder 38 MHN 5 g
thermal cycling to 125 C. Clearly, once the brittle network
of intermetallic reached a critical thickness, it fractured. Under normal spacecraft conditions these intermetallics
Each joint had been slightly loaded in tension by the nor- do not give cause for concern, and even thermal cycling will
mally acceptable tolerances of antenna positioning. not cause them to grow to any appreciable thickness (in
A far more obnoxious if not less costly failure due to gold Fig. 6.20 they are just visible). Their growth becomes a real
embrittlement involved the intermittent electrical failure of a problem only when high temperatures (80120 C) are
discrete transistor located on a ceramic platform. Excessive encountered for long periods, because, when stressed, they
gold dissolution from both a gold-plated Kovar pin and the are prone to the form of cleavage fracture depicted in
thick-gold metallization on the ceramic resulted in gold sat- Fig. 6.33. A further consideration is that of the effect of
uration of the molten tinlead solder joint. The as-solidied elevated temperatures encountered either in a burn-in
interconnecting network of intermetallics became slightly treatment before the actual service of an electronic compo-
stressed during a thermal cycling test, and produced a criss nent or during actual assembly. This may not only degrade
cross of ne cracks; these are shown in Fig. 6.32. solderability owing to oxidation of the surfaces to be sol-
Gold has a high solubility in liquid solder and is very dered, but the growth of intermetallic layers which are
quickly leached from coated articles during the soldering composed of tin and copper can make the surface become
operation. When the concentration of gold in solidied lead-rich and this will also produce a solderability problem
solder reaches 10 %, the joint is hard, brittle, and on (Billot et al. 1982). A systematic investigation has been
stressing the solder volume has essentially zero elongation. undertaken to assess the growth rates of compound layers on
Even 2 or 3 % gold in a joint markedly reduces the super- various metal bases plated with both tin and its alloys
plasticity of an uncontaminated solder joint. The presence of (Warwick and Muckett 1983; Hagge and Davis 1984). After
brittle needles of goldtin intermetallic in the joints seen in an initial fast rate of growth, the intermetallic compounds
Figs. 6.31 and 6.32 will obviously degrade the quality of the were generally seen to grow in a uniform manner as a
interconnection, but in many cases defective joints have function of time. Graphical presentations for most combi-
shown failure along the solder-to-gold interface when there nations of materials and solder alloys show that these growth
were no needle-like intermetallics within the solder volume. rates obey the Arrhenius law.
In this instance, it is the distribution of the intermetallic layer A coating of fused tinlead eutectic composition alloy is
(shown schematically in Fig. 6.31c) that is important and not strongly recommended for use in high-reliability electronic
the absolute gold concentration. Work by Kramer et al. applications whenever the temperature constraints permit
(1994) have shown that creep of pure, eutectic solder is (i.e. extended periods below 80 C). This nish could also
dominated by grain boundary sliding. This is the mechanism replace tin, cadmium, and zinc coatings which, as will be
by which solder relaxes when joints are strained by seen in Chap. 7, have been specically excluded from
362 6 Failure Analysis of Electrical Interconnections and Recommended Processes
(b)
Solder
fillet
(c) (d)
6.6 Problems Associated with Coatings for Soldering Applications 363
(a) (c)
200 m
(b)
20 m
100 m
Fig. 6.32 ac Illustrate a severely gold-embrittled solder joint. The pin and metallized pattern on ceramic results from the n network of
general photomicrograph a shows the effect of specimen bevelling due cracks b which propagate longitudinally through the acicular goldtin
to the extreme hardness variations between alumina ceramic, the Kovar intermetallics, and c along the solder-to-pin interface
pin, and soft solder alloy. Intermittent electrical open-circuit between
spacecraft hardware as they can support stress-induced necessary to provide an excellent surface conductivity
whisker growth. Also, as has been discussed, the latter two combined with the absence of surface oxide formation. In
metals can sublimate in the space environment. 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
6.7 The Use of Indium Solder Alloys be masked and selectively plated with metallic nishes
compatible with tinlead soldering. When selective plating
The number of integrated circuits and PCB substrates is not possible because of circuit restraints it is important that
required for microwave purposes is growing rapidly. Con- tinlead soldering to gold is avoided; as shown in Sect. 6.6,
ventional spacecraft PCBs utilize tinlead nished circuits gold is rapidly dissolved into solder with the formation of
with suitable spaces to interconnect the passive and active extremely brittle intermetallic compounds. The critical
components. They are, however, less desirable for composition of gold in solder for the formation of brittle
high-frequency applications owing to unavoidable signal joints appears to be about 0.6 % (McKay 1983) and may be
loss and distortion, and gold-plated conductor material is as low as 0.2 % (Kramer et al. 1994).
364 6 Failure Analysis of Electrical Interconnections and Recommended Processes
(b)
10 m
(c)
5 m
The most common indium-lead alloy contains 50 wt% of has a good wettability to gold with mildly activated flux and
indium, it has a liquidus temperature of 210 C and a solidus this is most suited for soldering to gold platings that have a
of about 185 C. On cooling from the liquid phase, lead-rich thickness of between 1 and 10 m. Soldering techniques are
dendrites form and these can give solder joint llets a generally the same as for tinlead solders, using hot-plates,
somewhat dimpled or frosty appearance. The 50InPb alloy soldering irons and solder in ribbon or as a paste.
6.7 The Use of Indium Solder Alloys 365
(a)
(b)
Fig. 6.34 Metallographic evaluation of indiumlead soldered to circuit. d Hardness impressions in the AuIn2 layer on (e) to indicate its
various gold interfaces after thermal cycling (100 cycles +100 to ductility, i.e. no radial cracking as would be the case with goldtin
100 C) and 12.8 years room temperature storage. a Overall view of intermetallics. e Edge of solder llet to highlight 2.5 m thick gold
microsectioned chip capacitor. All circuit tracks are gold plated, as is layer. This produces an intermetallic having a thickness of 14.0 m and
the aluminium support plate. b Detail of circuit-board-to-aluminium some free gold is still present (arrowed). f A gold-plated beryllium
support plate soldered connection. Aluminium is plated with an copper alloy ribbon interconnector was also soldered to the Teflon
intermediate layer of copper, and the nal gold layer has transformed PCB. There has been complete reaction with the gold on the
entirely to AuIn2. The gold-plated copper PCB ground plane is also interconnection side of the joint, and some free gold exists on the
converted to AuIn2. c Detail of lower corner of chip termination. Note PCB side. Although the IMC is extremely thick, there is no evidence
that excessive pressure during solder assembly has deformed copper that these joints have degraded
It has been shown that indium-lead solders form unusually A method for metallographic preparation and etching of
ductile intermetallic compounds with gold. These solders do indium-to-gold soldered joints has been detailed by Millares
nd use within electronic hybrid circuits for attachment of and Pieraggi (1992). An example of such an indium-soldered
small chip capacitors and resistors to ceramic and PCB connection after extensive thermal cycling and 12.8 years of
substrates (Yost 1976). The quality of such joints is fre- room temperature storage is seen in Fig. 6.34. It can be seen
quently assessed by the metallographic examination of test that the reaction layer between gold and indium does produce
samples made in parallel with flight production hardware. a layer of intermetallic, but this layer does not have an
366 6 Failure Analysis of Electrical Interconnections and Recommended Processes
(c)
(d)
acicular form, is relatively soft, and therefore causes no be AuIn2. Similar indium-solder interconnections were
embrittlement. The hardness of the main intermetallic is: investigated by Hauer (1977), who found that joints made to
miniature passive ceramic chip components have no
AuIn2 280 MHN 5 g decrease in strength after 10 thermal shocks from 193 to
It is interesting to note that a 2.5 m plated gold layer will +70 C and 100 quick temperature cycles from 55 to +85
transform by solid diffusion to form a 14 m reaction C after a high-temperature storage of 10,000 h at 85 C.
intermetallic layer. Electron probe microanalyses have con- They can be considered suitable for most spacecraft
rmed the intermetallic composition shown in Fig. 6.34 to applications.
6.7 The Use of Indium Solder Alloys 367
(e)
(f)
Some caution must be observed when indium-lead sol- The fatigue and shear strength properties of soldered joints
ders are selected for high-reliability applications. These have been studied by Wild (1975) and commented on by
alloys were noted (Goldmann et al. 1977; Banks et al. 1990) Jackson (1973). Between 55 and +125 C they established
to be more susceptible to corrosion than the SnPb alloys that 50 %In50 %Pb provided a one-hundredfold increase in
when submitted to high humidity and high temperatures. fatigue life over eutectic SnPb. The excellent fatigue lives of
Whereas SnPb is stable at 85 C and 85 %RH, corrosion indium solders were also conrmed by Hashimoto et al.
products containing mainly In(OH)3 were detected on the (1991), who investigated VLSI package flip-chip-soldered
InPb alloy. The presence of chlorides appears to stimulate joints thermal cycled between liquid nitrogen and room
the reaction by the formation of hygroscopic salts (e.g. temperature. The packages were nished with Au/Pt/Ti and
InCl3), but it would appear from the work of Kessowsky Au/Pt/Cr metallizations and the In40 Pb solder alloy was
et al. (1978) that a highly acidic environment, possessing a seen to be twice as resilient to thermal fatigue as Sn37 Pb
pH of less than 4, is a prerequisite for corrosion, and from solder. The indiumlead alloy was found particularly
their work indiumlead soldered to gold has proved quite well-suited for hybrid assembly because it absorbed gold at a
reliable from the point of view of joint strength, intermetallic slower rate than tinlead. A recent study conrmed that the
formation, and effect of long-time exposure. 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 369
is far less than the growth rate of other intermetallics com- can avoid reflowing or deforming the initial soldered
monly found in soldered joints made with SnPb, SnSb, connections.
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 6.8 Wires and Cables
zones having a similar appearance to those observed by
Maciolek et al. (1978) that formed during 2000 h at a storage 6.8.1 Selection of Plated Finish on Copper
temperature of 70 C. However, once ageing is performed at Conductors
temperatures in excess of 125 C the zones become
nonuniform and possess large columnar grains and large There is no wire conductor material that can be recom-
voids. Voiding and an unstable intermetallic interface asso- mended as the panacea for all potential corrosion, process-
ciated with the intermetallic Au9In4 has been observed after ing, and application problems. The suitability of various
ageing at 150 C (Yost et al. 1976). Experience has shown metallic nishes on copper wires for spacecraft applications
that InPb solder is suitable for the electrical and mechanical has been reviewed, and the choice nally limited to tin,
interconnection of gold-plated articles with the important silver, and nickel. However, owing to its low melting point,
proviso that operational temperatures do not exceed 70 or tin is generally not considered for the insulation materials
80 C. The mechanical test data presented in Fig. 6.35 that require high-temperature processing, such as PTFE,
support the view that under these conditions InPb alloys are FEP, Teflon, and Tefzel, and Kapton.
far superior to those alloys based on tinlead. (Incidentally, commercial insulation materials cannot be
The 50InPb alloy has been selected for the assembly of used in space or vacuum applications. The most common
components onto circuit board intended to operate at cryo- are: polyvinyl chloride (PVC); this has a high outgassing, is
genic temperatures. This alloy losses its strength properties considered a toxic material, and when heated releases
on cooling, but remains soft and ductile, even down to liquid hydrochloric acid which can surface-corrode and
helium temperature. When compared to eutectic tinlead, it stress-corrode metals. Silicone rubber outgasses excessively
is assumed that the greater stress relaxation in indium-lead under vacuum. Teflon, even though widely used, tends to
soldered joints during thermal cycling within cryogenic cold-flow, and wire strands have been seen to creep through
regimes will provide those circuits with a higher reliability. this insulation in the locations of sharply bent wire har-
The mechanical properties of metallic and organic electronic nesses; it also has a lower temperature rating than Tefzel.)
materials have been measured down to 4.2 K (Fink et al.
Tin-plated nishes
2008) and will be reviewed in Sect. 6.15.13. Two of the
Tin-plated conductors have been successfully employed on
common solders can be compared:
most spacecraft, but there remains some doubt as to their
long-term reliability, particularly after long-term ageing. The
50InPb remains soft (E = 14 GPa) and ductile (A = 20 %) at
reduction in solderability of tin-plated copper will be dis-
4.2 K but has low strength (Rm = 76 MPa) and low proof
cussed in Sect. 6.8.2, but this is also accompanied by a
stress (RP0.2 = 33 MPa)
reduction in flexibility and an increase in electrical resistivity
63SnPb is stiff (E = 43 GPa) and brittle (A = 0.3 %) at 4.2 K
owing to the progressive growth of tincopper intermetallic
but has a higher strength (Rm = 136 MPa) and higher proof
compounds. The effects of tincopper intermetallics have
stress (RP0.2 = 116 MPa)
been studied by McCune et al. (1970), and the problems that
could develop during the in-service refurbishment or modi-
Pure indium and indium solders are also used for the
cation to a reusable space vehicle might include:
sealing of glass and beryllium windows within the units of
satellite sun sensors and small-particle detectors such as
the increase in voltage drop through the intermetallic
X-ray and gamma-ray counters.
layers present within crimped electrical contacts (see also
Summarizing, indium-based solders have melting points
Sect. 6.4.3, p. 341),
ranging from 93 to 314 C, and in general they will wet and
difculty to wet with solder, and
flow onto soldering surfaces in the same way as tinlead
simple handling during repair conditions could result in
solders. The wide range of melting temperatures enables step
wire fracturing.
soldering processes to attach wire leads in close proximity to
each other by means of the progressive use of lower A review of the literature, discussions with cable spe-
melting-point alloys. The lower soldering iron temperatures 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 371
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 Fig. 6.36 General view of Philips Globule solderability test equip-
2.0 m minimum coating, and nished wires from approved ment. A60/40 solder globule (weight 100 mg). Bstripped wire in
manufacturers can be subjected periodically to accelerated position for test
tests designed to evaluate their susceptibility to red plague
corrosion (Dunn et al. 1984). Tin-plated conductors have the Tin-plated copper AWG 24
lowest cost, but owing to the low melting temperature of tin Silver-plated copper AWG 24
these conductors can be insulated only with low melting
point or irradiated insulations. The electrical conductivity of Ageing tests
tin-coated conductors is not as good as silver-coated con-
ductors owing to the lower conductivity of tin and the for- Steam ageing The samples were suspended above boiling
mation of brittle intermetallic tincopper compounds. distilled water in a clean, glassware condensator column
Initially the solderability of tin-plated conductors is good, for either 1 or 24 h (MIL STD-202E Method 208C).
but experience and literature sources show that this deteri- Dry-heat ageing The samples were suspended for 16 h
orates with shelf life (Comiskey 1979). within an air-circulating laboratory oven held at 155 2
C (B.S. 2011, test B).
To avoid potential corrosion problems and the possibility of
corona effects when small amounts of ionic flux residues Solderability testing
outgas from beneath wire insulation, it is recommended that
each soldering assembly operation is performed with (non- Philips Globule method Performed to BS 2011, part 2T,
activated) pure rosin flux. This necessitates that all soldering with the appropriate solder pellet, at 235 C. Both wire
surfaces must have an excellent solderability if the resulting and solder pellet were rst coated with a nonactivated
solder wetting angles are to remain low. flux solution. Wires were lowered to split the solder
The following subsections describe the simple ageing and globule in two, and the time was taken for the halves of
solderability testing of lengths of tin-plated and silver-plated the globule to flow around the wire and unite above it (as
conductors. Both conductors had been fabricated to military a measure of its solderability). This apparatus is shown in
specication requirements, insulated with the same radiation Fig. 6.36.
crosslinked extruded modied fluoropolymer by the same Dip test Performed with Tri Moore solderability tester, to
manufacturer, and both had a shelf-life since manufacture of Mil Std 202E method 208C, with fluxing in nonactive
less than one year. rosin flux, then draining off flux until it becomes tacky.
Wire samples After skimming the solder bath individual samples were
Each sample consisted of an 8 cm length of wire that had immersed for 5 s in the solder held at 230 C. Each
been mechanically stripped to reveal 3 cm of bared con- dip-coated wire conductor was then thoroughly clean of
ductor strands. Cutting, stripping, and storage operations flux and the 2 cm dipped portion examined by stereo-
were performed in a laminar flow clean air bench. The zoom microscope at 10 magnication. According to Mil
samples were handled with clean lint-free gloves, and the Spec, accept/reject criteria include a 95 % coverage by
bared conductors were not touched. Five samples were new solder coating and pinholes or voids not exceeding
applied to each test variable. The supplied wires were: 5 % of total area.
372 6 Failure Analysis of Electrical Interconnections and Recommended Processes
temperatures of the order of 150 C. The micrographs in entirely transformed to Cu3Sn5 and/or Cu3Sn. The micro-
Fig. 6.39 indicate that the wire strands were hot tinned, as the graphs shown in Figs. 6.39 and 6.40 clearly show that the
reaction layer between copper and tin has an appreciable dark-grey bands of Cu3Sn exist on signicant areas of the
thickness (about 0.5 m of Cu3Sn). It is expected that the free external conductor strand surfaces. Where intermetallic
tin existing on the surface of the as-received strands will compounds are exposed they become oxidized. It is probably
facilitate solder wetting. The ageing processes are known to not tin oxide, but this oxidized intermetallic compound results
increase the intermetallic compound thickness by an Arrhe- in poor solderability (Davy and Skold 1985; Warwick and
nius rate of growth until the free tin external surface has been Muckett 1983).
374 6 Failure Analysis of Electrical Interconnections and Recommended Processes
6.8.3 Red Plague Corrosion and 6.42), is the cuprous oxide (possibly with some black
of Silver-Plated Copper, cupric oxide) corrosion product that forms when a galvanic
and Plagues on Other Plated cell is formed between copper and silver. Green plague is
Stranded Wires the reaction product of some resin-based solder fluxes with
copper oxide; it has been found on tin-plated and
Two potential corrosion problems were recognized to be silver-plated conductors, and is considered to be an unde-
associated with the use of silver-plated copper conductors in sirable cosmetic defect.
the 1960s. Because of their characteristic corrosion products, In recent years, the term white plague (or fluorine
they are referred to throughout the electronics industry as attack) has been coined to account for the white,
red plague and green plague. Red plague (see Figs. 6.41 non-solderable deposit occasionally found on the outer
376 6 Failure Analysis of Electrical Interconnections and Recommended Processes
strands of nickel-plated copper conductors. White plague 0.003 %) had a signicant amount of red plague, and that
is only found on fluoropolymer-insulated electrical wires and no malfunctions could be attributed to this phenomenon.
cables having nickel plated conductors. It is thought to occur Several investigative activities were undertaken in Europe
during the manufacturing of wires and cables. In-coming (Dunn et al. 1984) in the mid-1970s. These included labo-
inspection of nickel-plated stranded wires should include ratory evaluations, a European red plague survey, and
visual checks for this condition, if found it can be construed eventually the auditing of wire manufacturers premises. It
as reject condition. During the high-temperature extrusion of was concluded that silver-plated copper conductors would
fluorocarbon resin onto stranded wires some gaseous mate- be suitable for general spacecraft usage, but it was essential
rials outgas, these include carbonyl-difluoride (COF2), an that the platings be defect-free and of uniform thickness (at
extremely reactive compound which, in the presence of least 2 m, and preferably 4 m). It was considered that red
water and the nickel surface forms an adherent, porous white plague could be avoided if the wire manufacturer and the
deposit. [The recent usage of this term should not be con- user companies implemented certain mandatory controls.
fused with the same name, attributed by component engi- The causes and the effect of red plague can be sum-
neers during the wire bonding in microelectronics, see marized as follows:
Sect. 6.3here gold micro-wires bonded to aluminium form
two intermetallics: Au5Al2 (white plague) and AuAl2 (purple (1) Imperfections ranging from porosity to deep scrapes pen-
plague)]. etrate the silver plating and reveal the copper wire of a wire
Summaries of the mechanism by which red plague occurs strand. Such damage may result from the use of defective
are given by (Dunn 1984) and Cooke (2010). wiredrawing dies, wear action between strands during
A number of reports including the work by Anthony and stranding and braiding operations, or when conductor
Brown (1965) were written in the 1960s, which established strands are wound and rewound from reels to bobbins.
that red plague originates at breaks in the silver plating of (2) Moisture may penetrate beneath the insulation of wires
copper wire strands in the presence of moisture and oxygen. and cables during the wire fabrication process, or
Although some authors considered the problem to be per- during insulation testing, or from the humidity in air.
sistent, uncontrolled, and unsolved, others presented solu- (3) The presence of moisture in the region of exposed
tions, thought likely to avoid the problem. A NEMA (1972) copper can produce an electrolytic cell, the copper
investigation reported that out of a total USA production of tending to corrode and protect its more noble silver
18 to 20 109 ft of silver-plated copper wire, only 105 ft (i.e. plating. The following reactions were proposed:
6.8 Wires and Cables 377
Fig. 6.41 a Characteristic red corrosion products red plague on silver determined by the sodium polysulde test (Dunn not published).
plated copper strands of AWG 26 single wire (after removal of c Despite the controls placed on procurement of silver plated wires it is
insulation), and b, microsection through one strand showing galvanic essential to avoid damaging the plating as revealed in this illustration of
corrosion between small (arrowed) defect in silver plate. All corrosion flight equipment (Cooke 2010). The insert shows stranded, silver-coated
products appear as Cu2O, conrmed by having a bright orange copper wire with varying stages of red plague, damaged plating and
colouration under polarized light. This wire conforms to ASTM B exposed copper (photo credit NASA Johnson Space Center)
298-07 requiring not less than 1 m silver average, and no porosity as
378 6 Failure Analysis of Electrical Interconnections and Recommended Processes
At the anode (i.e. exposed copper): 4Cu 4OH As the water dries out, the cuprous oxide crystals,
2Cu2 O 2H2 O 4e which have a characteristic red appearance, precipitate
At the cathode (i.e. silver plating): O2 4H out onto the silver plating.
4e 2H2 O (5) Excessive corrosion can impair the wires fatigue life
The sum reaction is: 4Cu O2 4H 4OH and electrical conductivity.
2Cu2 O 4H2 O Auger (AES) surface analyses were made on the
Note that the water is not consumed, and in the presence of silver-plated strands from 51 different wires and cables samples,
oxygen the galvanic copper corrosion can proceed indenitely. each originating from European space companies stock rooms
(Dunn 1984). The same samples were submitted to the accel-
(4) Copper hydroxides are soluble in water and will con- erated corrosion test. All surfaces supported varying amounts of
taminate the water that surrounds the stranded wires. chlorine (up to 8 at.%) and fluorine (to 4.8 at.%). The influence
6.8 Wires and Cables 379
of Cl and F on the corrosion rate is not understood; the nitrogen with active desiccant should have an unlimited
chloride reaction with silver-plate will form AgCl, which is shelf-life).
insoluble in water and can be expected to coat the surface with a control the environment by water-vapour-proof
protective lm. Fluorine however, forms silver fluorides, AgF, packaging/bagging during shipping and storage, use
AgF2 and Ag2F and these are all highly soluble in water. This humidity indicators in bags, assembly area shall be
contaminant has evolved from the fluoropolymer-insulation environmentally controlled, aqueous solvents shall not be
surrounding these wire strands and its presence is expected to used for cleaning.
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 6.8.4 Manganin Wire
the formation of red plague is still uncertain and continues to be
studied (Kauffmann and Wolf 2013). Manganin is an interesting alloy. The French physicist Emile
Although the incidence of red plague corrosion of Hilaire Amagat (18411915) worked on high ranges of
silver-plated conductors in service is extremely rare, it is pressureone of his earliest inventions was the
emphasized that for high-reliability space applications, such self-tightening bolt which allowed pressures to be devel-
as those involving manned vehicles, it is necessary to control oped and maintained far in excess of those reached before. He
all steps in the design, handling, and processing of wires and studied pressure-measuring devices and in 1911 reported the
cables constructed with such conductors. Red plagued wires design of an electrical resistance gauge which enabled him to
were found on the ESA Spacelab after service [see Sect. 8.7. measure pressures up to 1270 MPa. The secret is now known
3 and the publication by Dunn and Stanyon (1997)] and to be based on one of the rst uses of the alloy Manganin. The
more recent damage, resulting in red plague was observed on electrical resistance of Manganin was found to be a linear
other space equipment (see Fig. 6.41c). function of pressure up to 1180 MPaand by extrapolation,
It is recommended that the basic hot-rolled and redrawn pressure measurements could be made up to 2010 MPa.
copper rods are submitted to verication tests such as Manganin is a coppermanganesenickel alloy possessing
chemical composition, electrical resistivity, and absence of a high electrical resistance and a low temperature coefcient
surface and internal defects before silver plating (Pops 1985). of resistance (0.00000 at 25 C). The original alloy has a
After drawing of the silver-plated rods it is important that all composition of 84Cu 12Mn 4Ni, but this specication for-
stranding and braiding machines must be controlled by wire mulation has been slightly modied in recent years so that the
and cable manufacturers to ensure that they do not damage temperature coefcient could be adapted to suit equipment
the continuity of the 2 m silver plating. If this is an area of operating at elevated temperatures. Manganin wires have
concern, it would be expedient to produce strands to a greater found many applications on board spacecraft in the con-
nominal silver thickness (possibly 34 m). Humidity con- struction of precision resistors and in the electrical measuring
trol for cable reels (spools) can be achieved by packaging in a and control units for space experiments. The Tyco company
sealed transparent plastic bag (polyvinyl chloride is usually produce a NASA flight qualied (2015) shielded ribbon cable
prohibited). Desiccant or dry-nitrogen lling should be used consisting of 24 twisted pairs of 40 AWG manganin wire.
in the packaging. An easy means for indicating that the These shielded light-weight interconnections are designed to
humidity content is below 50 %RH should be provided, operate between 273 C and +200 C, mainly for spacecraft
capable of being read without opening the package. Finally, Cryogenic Telescope instruments. The wires can be soldered
an accelerated corrosion test can be used as a screening or welded into circuits, but any bending operations will cause
method for silver-plated wires, particularly for very critical the alloy to work-harden; this produces good spring proper-
applications, such as the test proposed by Anthony and ties but a marked reduction in electrical stability. However,
Brown (1965) and adapted in the European specication any cold work can be removed by an annealing heat treat-
ECSS-Q-ST-70-20 Test procedure to determine the suscep- ment, in an inert atmosphere, at 300 C for 2 h. The
tibility of plated copper wire/cable to red plague corrosion. microstructure of properly heat-treated Manganin possesses
The NASA Red Plague Control Plan (JSC 64647) and equiaxed grains where the addition elements are in stable
NASA STD-6012 (2012) include many of the remarks made solid solution. All processing steps associated with the
earlier, mitigation assembly of Manganin wire must be carefully performed, as
the alloy is susceptible to selective oxidation and galvanic
only use qualied suppliers, ensure traceability and cer- corrosion. These effects can result in manganese-depleted
tied test reports, recommend shelf life of 10 years. regions along the wire causing a greatly increased tempera-
(Although, fully traceable wires/cables stored in dry ture coefcient of resistance.
380 6 Failure Analysis of Electrical Interconnections and Recommended Processes
6.8.5 High-Voltage Wires, Cables, been degassed prior to pouring into the connector shell and
and Connections might have contained small bubbles that caused the partial
discharge activity, or that the wires were not clean and flux
Spacecraft power systems have continuously increased out- residues might have reduced the dielectric properties of the
put from a few watts in the 1960s to the several kilowatts compound. In discussions with experts, it would appear that a
demanded by the present space-based activities. As the well-designed high-voltage connector does not require pot-
power levels increase, so the transmission voltages will have ting for its high-voltage withstand capability. With an open
to be increased in order to decrease currents and so minimize construction, space vacuum is the best high-voltage insula-
the conductors weight and resistance losses. For instance, tion, and is the lowest weight insulation! (Wetzer and
some travelling wave tube ampliers (TWTAs) are now Wouters 1993, 1994; Haq Qureshi and Dayton 1995).
operating at 15 kV. High-voltage connections where corona suppression is
The International Space Station requires even higher necessary will require special designs. For instance, all
electrical loads. Here the photovoltaic electric power gen- aspects of soldered joints should be covered by smooth llets,
eration solar arrays supply about 83 kW. The wires and free of discontinuities or severe changes in contour, i.e. sharp
cables for electrical transmission and distribution systems edges and points must be avoided (ECSS-Q-ST-70-08).
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- 6.8.6 Cold Welding of Stranded Wires
tant, they must also be able to operate under high vacuum. and Cables
From a vacuum point of view, a high-voltage cable is a
poorly vented component. It may have a length of many Concerns were raised when a band cable for an optical
meters. In manned spacecraft the problem might be how to terminal was under development. Some difculties were
prevent condensation of water vapour onto the conductors noticed after mechanical testing during which the cable had
when there are changes to the cabin temperature (note the been rotating through 200 in a vacuum environment. The
red plague corrosion phenomena of Sect. 6.8.3) but, when cable became stiff and the bending was not smooth. Cold
launched into a vacuum, there needs to be sufcient time for welding of the individual silver-plated strands within insu-
air to completely vent from the cable and for the cable lated wires of the cable was suspected and a detailed labo-
materials to outgas before switching on any power supplies. ratory test initiated (Fig. 6.43). Anomalies associated with
If such practice is not applied, there can be partial discharges spacecraft Moving Mechanical Assemblies have been
and corona effects that will damage the cable. addressed by Pirnack (2012) and suggestions have been
To ensure proper venting it is particularly important that made how to avoid this problem. However, if magnetic
wire and cable ends are not blocked by potting compounds. materials can be accepted on-board spacecraft, it is the wri-
These have given problems on certain telecommunication ters opinion that for a given design, nickel-nished strands
satellites when wires were soldered onto the pins of will be far more resilient to cold-welding than silver nishes.
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 6.9 Problems Associated
was itself lled with sulphur hexafluoride gas, the most with Soldering Fluxes
common gas used as a dielectric (having a very low break-
down strength, good heat transfer properties, and chemical 6.9.1 Purpose of a Flux
stability). The vacuum side of the connector consisted of ve
high-voltage, stranded wires, each partially stripped of The importance of a flux in soldering operations is usually
insulation and soldered to their respective connector pins. underestimated. For conventional soldering the use of a flux
Solder flux had been cleaned from the joints, with the pos- is imperative, and when correctly formulated it will:
sible ingress of alcohol and water beneath the insulation
jacket. The opposite ends of the wires were also soldered to (a) form the initial heat bridge between the heat source and
connectors, then clean and protected with heat-shrinkable the part;
sleeves. The TWTA connector was then back-potted with a (b) react with, and either modify or remove, surface oxides,
hard compound that essentially caused the internal volume of passivation lms, and contamination lms;
the wires to become airtight. This caused the inside pressure (c) eliminate air from the liquid solder front because, rst, it
of the wires to be uncontrolled and, during early operations in wets, and, second, it spreads by surface tension forces.
orbit, is thought to have been the cause of partial electrical
discharges, or flashovers at this side of the connector. Other In general the space-quality specications limit the choice
suggestions for this problem were that the potting had not of material nishes which may be interconnected to those
6.9 Problems Associated with Soldering Fluxes 381
having an excellent solderability, so that highly activated years (Parker 1977). Essentially, this product consists of an
fluxes are not needed during the actual assembly process. outer length of plastic tubing that has been extruded to a
Fluxes of high activity, which are potentially more dangerous desired nal diameter. The tube is then irradiated to crosslink
from a corrosion viewpoint, are permitted during the initial the plastics molecular chains, and this will enable it to
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. when heated. The tube is next warmed and mechanically
After pretinning, the flux residues must be thoroughly cleaned widened to an enlarged diameter. A preformed ring of solder
from the component lead surface to preclude time-dependent containing a resin-based flux is inserted into the plastic tube.
corrosive attack. Similarly, the nonactivated rosin-type fluxes Two small plastic sealing rings are also pressed into the ends
which are permitted for component assembly and wire harness of the tube. Wires, which have been stripped of their insu-
manufacture must be removed by special cleaning procedures lation, can be inserted into the tube with the intention of
immediately after the soldered joint has been made. Proper making lap-joints. By heating the tube with hot air, or an
cleaning is considered essential in order to avoid spacecraft infrared gun, the sealing rings melt and flow around the
problems associated with the effects of flux residue contami- insulated wire, the tube shrinks, and after the flux has flowed
nation, such as outgassing, corona effects, and both over the wire strand surfaces the solder alloy melts and flows
stress-corrosion cracking and galvanic corrosion. These to interconnect the wire strands. On cooling, the solder joint
effects will be overviewed in the following sections. solidies and the flux and its residues are entrapped within
the outer sleeve.
Visual inspection of this kind of device after soldering
6.9.2 Heat-Shrinkable Sleeves Containing is difcult because of the opaque nature of the sleeve
Solder Preforms material and masking by the retained flux. For this reason it
is essential to assess the quality of these interconnections by
A special product designed for joining wires and cables by making metallographic examinations on a periodic basis.
spliced joints has been commercially available for several 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 An attempt was made to make soldered interconnections
suitable for commercial and aerospace applications. These with the same device but with the incorporation of solid
devices have been thoroughly evaluated for spacecraft solder preforms which did not support or contain any kind of
applications (Premat 1977; Dunn 1978). It was found that flux. From visual inspections the resulting joints were ini-
the outer sleeve is only a partial hermetic seal, and that the tially believed to be of reasonable quality. It was not until the
outgassing of flux from each assembly (under vacuum con- sleeves had been removed, and microsections made, that the
ditions) will represent between 30 and 60 % of the initial majority of joints were seen to be totally unacceptable owing
weight of the flux. Some additional flux outgassing data are to nonwetting. Figure 6.44 shows how the solder alloy has
given in Table 6.4. The problems associated with outgassing been pressed around the silver-plated wire braid (normally
and condensation of volatile materials were outlined in this nish possesses an excellent solderability), but no
Chap. 2. In this wire-joining application, under space con- metallurgical joint has been achieved. These devices have
ditions, gas pockets under the wire insulation and outer been prohibited for European space projects
sleeve can give rise to the additional danger of corona, (ECSS-Q-ST-70-08) but are permitted by NASA for
especially when chlorides in the vaporized flux ionize and attaching grounding wires to shields and for wire splicing
have a lower dielectric constant than the insulant material. (NASA-STD-8739.4 (2011).
Fig. 6.44 To avoid the inclusion of soldering flux within these pressed around the braid but no metallurgical bonding had taken place
devices, an attempt was made to employ a fluxless solder preform. (arrowed). b Enlargement of area arrowed in (a). c Detail reveals an
Visual inspection of the resulting joints had given acceptable results, open electrical connection and lack of alloy formation between the tin
a Metallographic inspections showed that the molten solder was lead eutectic solder and the silver-plated braid
6.9 Problems Associated with Soldering Fluxes 383
6.9.3 Stress Corrosion of Component Lead Both the scratched areas and the porous gold plating are
Material capable of initiating and sustaining SCC growth by diffusion
of chloride ions from the residual flux to the crack tip. The
Notwithstanding the tight soldering process controls applied electrochemical mechanism of SCC in alloy steels is accel-
by space projects, several problems associated with the erated by the application of anodic currents, and a similar
corrosive nature of liquid-soldering fluxes and their residues situation is thought to exist in the case of these gold-plated
have cost much wasted time and effort. They include the Kovar leads, anodic currents being set up between the por-
failure of various gold-plated Kovar leads on flat-packaged ous plating and the less noble Kovar alloy. It is to be noted
components. These leads had been formed and then prepared that the corrosion products occupy a larger volume than the
for gold removal prior to soldering. Degolding and pretin- Kovar from which they are formed, and these products will
ning were performed by dipping the leads into solder baths tend to increase the SCC propagation rate by creating a
fluxed with a mildly activated rosin. After a few months of wedging action and additional stress concentrations at the
storage, the leads from several batches of flat packages were crack tip.
observed to fracture completely when exposed to light The fracture of Kovar alloy as a result of exposure to
handling operations. Metallographic investigations, as SCC environmental conditions has not been widely reported
shown in Fig. 6.45, strongly indicated that the failure (Weirick 1975). Effective protection of Kovar component
mechanism of these leads was one of stress-corrosion leads from SCC has been achieved by rstly chemically
cracking of the Kovar alloy due to the combined effect of removing the work-damaged (from lead-stamping opera-
residual stresses in the lead material after the forming tions) surface of the lead before thin gold plating. The thin
operation and the presence of a thin surface lm of sup- plate is designed to protect the lead from oxidation during
posedly noncorrosive flux residue. Other failures have component manufacture, but is nally removed by
included the fracture of mounted transistor leads after dip-coating with a ductile pore-free nish of eutectic tin
equipment level vibration testing; fractography revealed that lead solder, and any gold plate remaining adjacent to the
stress corrosion had initiated a crack in the lead material, and component-to-lead glass-to-metal seal is additionally pro-
this had later propagated by a fatigue mechanism. tected by a silicone varnish (Elking and Hughes 1969).
An extensive test programme was initiated to assess the Another method to minimize the SCC failure of leads is
corrosive effects of soldering fluxes then used by the to plate the Kovar with 12.5 m of nickel before gold
majority of contractors engaged in the manufacture of plating. However, even this process will not overcome
electronic hardware for European spacecraft. Seventy-one chloride-assisted SCC if the nickel mechanically cracked
samples of flux were received and tested (Dunn and Chan- during either component manufacture or subsequent
dler 1980) by the use of standard and ad hoc corrosion tests lead-forming operations (see also Sect. 4.10).
and the determination of flux halide content and pH values. The programme of flux evaluation was able to make
The corrosion tests were performed in a warm, damp several conclusions and recommendations which could help
atmosphere for 56 days and after deliberate contamination increase the reliability of not only spacecraft, but also
by the fluxes of samples of electronic materials. commercial electronic systems. These included the need to
Probably the most revealing results to have been pro- solder with only nonactivated flux; the need for flux
duced by this evaluation programme are those which indi- removal; and the use of stress-relief bends in leads between
cated the low stress-corrosion cracking resistance of part termination and component body.
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 6.9.4 Flux-Corrosion of Silver-Plated
under deliberately scratched regions and from below what is Stranded Wires
now considered to be the porous gold plating.
All samples of components were subjected to metallo- In a continuation of the work presented in the preceding
graphic inspection of their longitudinally cross-sectioned subsection an attempt was also made to assess the corrosive
leads. The test conguration and some typical photomicro- effects of many soldering fluxes on silver-plated copper wire.
graphs are presented in Fig. 6.46. The extensive degree of This was considered necessary because corrosion products
general and stress-corrosion cracking (SCC) shown by these had been observed on satellite wire harnesses. These corrosion
longitudinal sections is quite surprising because close visual products were of a bright-green and grey-brown nature, and
inspection of the lead surfaces does not always indicate any were not thought to be related to the phenomenon of red
sign of surface corrosion. 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 To enhance any mode of wire degradation the lengths of
such discoloration results from surface corrosion and is only a wires which were submitted to the test programme were
cosmetic defect. However, many spacecraft electronic boxes scratched so as to expose the underlying copper core. Each
and cable harnesses have long storage lives before launch. wire was covered by a controlled amount of the standard
Once in orbit, it is difcult to assess the long-term effect of this commercial soldering flux under investigation and then
corrosion mechanism on electronic conductors within pres- submitted to a long-term damp heat exposure. A majority of
surized systems which have been designed to operate for 10 to the fluxes caused extensive surface discoloration of the wires
40 years (viz. the International Space Station). after the total 52-day test period. Metallographic
6.9 Problems Associated with Soldering Fluxes 385
examination showed that some of the strongly activated than in seawater! Rosin-based fluxes are extremely difcult
fluxes did produce a limited amount of copper wasting. The to analyse, and they can include a vast range of additives
worst example is presented in Fig. 6.47. Here, 15 m of known only to the flux manufacturers. However, a typi-
copper have been removed from the 76 m diameter cal simple formulation (Weirick 1975) contains: 85 % abi-
strandbut it should be noted that this flux contained etic acid, 12 % pimaric acid and 3 % other natural
62.5 %CI, which is approximately 280 times more chlorine compounds.
386 6 Failure Analysis of Electrical Interconnections and Recommended Processes
form of tarnish existing on the workpiece surface, and the In recent years, it has been found cost-effective to prepare
joints should be designed so that the flux can be both surfaces for soldering with a dual metallic nish, particularly
properly introduced to the joint area and then totally when long storage periods are envisaged before the nal
removed after soldering. The various methods of soldering assembly of spacecraft parts. In these cases, otherwise
which transmit heat energy to the workpiece (by means of difcult-to-solder surfaces are plated with copper or nickel,
hand-held irons, torches, resistance soldering, oven or fur- and subsequently coated with a eutectic tinlead composi-
nace soldering, wave-soldering, etc.) may necessitate special tion which can be fused in hot oil to form a 412 m thick
fluxes which avoid factors such as the evolution of corrosive pore-free protective layer possessing excellent solderability
fumes during soldering; low flash-point constituents; irri- even after many years of exposure to industrial atmo-
tating or dermatitis-producing liquids and fumes. spheres. The choice of highly solderable nishes precludes
When torch-soldering is used, a more active flux may be the need for strongly activated fluxes which are often dif-
necessary to deal with the extra tarnish formed during cult to remove in post-soldering cleaning operations and
heating by the flame. Table 6.3 may be used as a guide in the may later lead to extensive corrosion of the less noble metals
selection of flux types for the soldering of common metals. It in any joint design.
is observed that the less noble metals listed in this tabulation The chemical composition of strong fluxes may be based
require either a more active fluxing system or a more sol- on aqueous solutions of halides, such as ammonium and zinc
derable nish, such as copper plating. chlorides or orthophosphoric acid; these are widely employed
6.9 Problems Associated with Soldering Fluxes 389
in general engineering. For use on electronic assemblies, and was retrieved by the shuttle Endeavour. During this
rosin-based fluxesoften in the form of alcoholic solutions mission, all objectives were achieved but several problems
are chosen. A range of rosin types are commonly used, and anomalies were observed during the course of the mis-
including nonactivated halide-free, organically activated, and sion (see also Chap. 8). One concerned the failure of Gyro
halide-activated rosins. A multitude of commercially available No. 3examination of this mechanism when it returned to
fluxes are on the market. These can include a vast range of Earth identied the failure to be due to an open phase of the
additives known only to the flux manufacturers. 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
6.9.6 Control of Galvanic Corrosion others were abnormally discoloured. Metallurgical analyses
found excessive amounts of flux residues to be adhering to
As a general design rule, it is desirable to avoid joining the wires and SEM images identied that acidic chlorides had
together metals which have greatly separated positions on caused corrosion to all wires and resulted in the
the galvanic scale (see also Sect. 2.2). Table 6.3 include data failure/rupture of one. This is a clear lesson that flux residues
of the relative activity of metals in relation to a standard must be carefully removed after soldering.
reference electrode based on a seawater environment. Many Flux residues must always be removed from systems
authorities recommend that, to avoid corrosion couples, exposed to vacuum, as it has been shown that rosin flux
metallic combinations which will be subjected to uncon- consists of substances nearly all of which have appreciable
trolled terrestrial environments do not exceed a potential volatility at typical equipment operating temperatures (Sin-
difference of 0.25 V. When the working environment is clair et al. 1985). In some instances it may be possible to use
controlled, e.g. in the case of electrical units operating under a new ultrasonic soldering technique which does not require
spacecraft assembly cleanroom conditions, compatible the presence of flux. The method is not suitable for
corrosion-free couples may be separated by potential dif- through-hole or surface mount assembly of components to
ferences not exceeding 0.5 V. PCBs where it is the capillary flow of solder (preceded by
It is fortunate that tinlead, the most common soldering the flow of flux) that is critical in forming good connections.
alloy (Group 10 of Table 6.3), is generally compatible with Also, some authorities consider that ultrasonic vibrations can
most solderable metals, with the important exclusion of gold. damage the internal microwire bonds within active compo-
The chart showing the galvanic series in Table 2.1 is useful nent packages. However, ultrasonic soldering, whereby
for a rst approximation in selecting materials for both ultrasonic energy is introduced into a bath containing liquid
solderability and corrosion control, but for many applica- solder, will have many applications, as oxides are mechan-
tions it may be too simplistic because it does not provide ically removed from a submerged workpiece by the forces of
information about corrosion rates or changes in surface imploding cavitation bubbles. Component leads can be
chemistry which may pacify or accelerate corrosion at effectively degolded and pre-tinned. Wire strands can be
bimetallic interfaces. 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
6.9.7 Cleaning of Flux-Contaminated molten solder temperature can also be lower for ultrasonic
Surfaces soldering than when a flux is used and this causes less dross
formation on the solder bath. The technique may also be
After soldering, the flux, dross inhibitors, solder resists, and useful (Saxty 1995) for regenerating the solderability of
any oils must be promptly removed from the joint in a difcult items such as terminal pins, wire tags, and com-
manner which does not change the hardware. A clean sol- ponent leads that have aged and would normally de-wet if
vent is recommended for the removal of all residues even soldered by conventional ECSS-Q-ST-70-08 techniques.
though the flux manufacturer may have classied the par- A method adopted by NASA and ESA for selecting
ticular flux, thinner, oil, etc. as being noncorrosive. If flux materials is the micro-VCM test. Samples of the RMA and RA
residues are to be left on soldered surfaces in pressurized fluxes were submitted to the volatile condensable material test.
electronic boxes, it is essential to perform environmental Initially they were spread onto aluminium foil and cured at
tests to ensure that such residues do not cause surface cor- room temperature for eight weeks and then, according to the
rosion or stress-corrosion cracking, thereby lowering the test procedure, the weight losses of their volatile condensable
integrity of the part. Flux contamination existing beneath products were evaluated after a thermal vacuum exposure of
surface mount devices is discussed in Sect. 6.15.9. the samples for 24 h at 125 C and 106 mm Hg.
The spacecraft EURECA was launched from the space The results given in Table 6.4 clearly indicate that these
shuttle Atlantis in July 1992. It remained in orbit for 286 days 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 after the use of deionized water. However, with
Sample Type TML/RML CVCM increased miniaturisation water-based processes cannot
(%) (%) penetrate into extremely low stand-off recesses.
Kenco RMA, rosin based 48.2/48.0 12.0
365 containing 0.01 % The Montreal Protocol established under the aegis of the
chloride United Nations is now, in effect, to contain and reverse the
Multicore RA, rosin based 21.2/21.0 5.0 depletion of atmospheric ozone which, if not halted, is likely
PC 21A containing 0.23 %
to have a catastrophic effect on the Earths environment.
chloride
Chemicals such as the chlorofluorocarbons (CFCs) are most
TML Total mass loss
RML Recovered mass loss
extensively used as solvents in the electronics industry, but it
CVCM Collected volatile condensable material is now recognized that their release into the atmosphere is a
Note Micro-VCM test per ECSS-Q-ST-70-02, A thermal vacuum test primary cause of the phenomenon of ozone depletion. Space
for the screening of space materials industries are no longer engaged in the production and use of
From Dunn and Bergendahl (1987)
the CFCs including the item CFC-113, also known as tri-
chlorotrifluoroethane, CC12F-CC1F2.
spacecraft owing to contamination and other phenomena Modern methods for the cleaning of mechanical parts
such as corrosion and corona. They may also cause complete were addressed in Sect. 4.21. Some manufacturers of elec-
failure of a satellite mission by recondensing between slip tronic equipments have considered the no clean and
rings, contact points (causing electrical open circuits), and low-residue solder flux routes after abandoning the CFC-113
on optical or thermal control surfaces. solvent. Certainly some excellent products exist for com-
The solvent cleaning of flux-contaminated surfaces may be mercial and some military applications, as was evident from
performed by immersion cleaning, sometimes with ultrasonic a Soldering Task Force test programme and results (Iman
agitation (not for electronics which support small integrated et al. 1995). However, the space community still insists on
circuits, as the miniature wire bonds can be degraded). Solder aesthetically perfect assembled PCBs for proper conformed
joints can be more readily inspected after flux residues are coating adhesion, low outgassing, and no possibility for the
removed. In those cases where inorganic or organic acid migration of tinlead as seen in Figs. 6.30 and 6.49. The
fluxes have been employed, it is especially critical to remove cleaning machine market appears buoyant as now there are
residual flux immediately after soldering and even before the different machines on sale for each variety of solvent (an
workpiece has cooled to room temperature. The nal rinse example of a successful machine is seen in Fig. 4.85).
baths may contain clean solvent or deionized water; they can Periodic testing for cleanlinessgenerally to check for
be controlled by electrical resistivity measurements made at the absence of ionic residuesof pcb assemblies after nal
predetermined intervals based on the number and size of cleaning (and prior to conformal coating or further integra-
articles being cleaned over a given period. Solvents should be tion of the assembly into electronic housings) is required by
properly labelled, and those showing evidence of contami- all space standards, and if these tests fail, the entire batch of
nation or decomposition should not be used. assemblies will need to be re-cleaned and re-tested.
The following solvents are acceptable for the cleaning of Ionic residue testing is based on the solvent extraction
electronic equipment provided that cleaning takes place soon method; a solution of isopropyl alcohol and deionized water
after the soldering operation (e.g. before the flux and resi- in the ratio 75/25 (volume/volume) is puried to a low
dues have had time to polymerize or age). electrical conductivity (i.e. the base line) by pumping it
through an ion-exchange column in a closed-loop; the board
(a) Ethyl alcohol, 99.5 or 95 % pure by volume. to be tested is introduced to the solution; ionic residues are
(b) Isopropyl alcohol, 99 % pure. washed off the board leading to an increase in conductivity
(c) Any mixture of the above. (or decrease in resistivity) that is a measure of the amount of
(d) Terpine. ionic residues dissolved in the solution. The levels of con-
(e) Any commercial product proven to be equal to or tamination are expressed as, micrograms/cm2 sodium
better than the solvent CFC-113 regarding cleanliness chloride (NaCl) equivalent ionic or ionisable flux residue.
and reliability issues. This is a long established test method dating back to the
(f) Hydrofluorether, methoxynonafluorbutan [C4F9-O-CH3] 1970s and many similar instruments for conduction the test
a clear, colourless and low-odour liquid has found some are available on the market. They are seen in all component
acceptance in the aerospace and electronics industries. assembly lines and include: the Contaminometer, Ion
(g) Deionized water at 40 C maximum may be used for Chaser, Ionograph and Omega Meter. An ESA report found
certain fluxes. Items shall be thoroughly dried directly models of these instruments to be suitable for the
6.9 Problems Associated with Soldering Fluxes 391
determination of ionic contamination levels, but the amount (b) PCB assemblies shall not be immersed in cleaning
of contamination indicated by each instrument differed solvents for more than 30 min for each cleaning oper-
slightly according to the type of flux under investigation ation (as long immersion times can promote galvanic
(Bergendahl and Dunn 1984). Equivalence factors for each corrosion between adjacent metallic surfaces).
instrument were proposed. The maximum limit of surface
contamination required by IPC and ECSS standards is: less ECSS-Q-ST-70-38 called for a stand-off of between 0.1
than 1.56 micrograms/cm2 NaCl equivalent. Other bodies and 0.4 mm to aid cleaning and minimize failure due to
require lower limits such as 1.3 or 1.0 g/cm2, and such thermal cycling.
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 6.9.8 Flux Residues, Their Ingress
some important ndings: during testing the solution can pick into Top-Coat of PCB Surfaces,
up some ionic contaminants such as CO2 and other gases and Bake Out After Cleaning
present in the surrounding air (e.g. forming carbonic acid) so
it is necessary to keep a lid on the instruments sample The preceding section described how the space industries
chamber; surface insulation resistance (SIR) testing is suit- have been searching for alternative solvent cleaners. The
able for assessing whether residues (ionic and non-ionic) recent use of surface-mounted components in spacecraft
might compromise circuit performance from an systems requires additional considerations, as the majority of
electro-chemical standpoint; FTIR and ion chromatography soldering techniques necessitate the application of active
are effective at measuring the presence of ionic salts but they fluxes to ensure that adequate solder wetting is achieved and
are not capable of determining whether their presence will the minute balls of solder within solder paste do not scatter
compromise circuit reliability. as they melt (ECSS-Q-ST-70-38).
A nal topic which had not been adequately addressed Several studies have been made to assess the suitability of
prior to the study by Zou et al. (1999) was: what duration of test equipment for the detection of ionic contamination on
time, and under what conditions of temperature, can both conventional and surface-mounted component assem-
assembled (and repaired/modied) PCBs be left before they bled PCBs (Bergendahl and Dunn 1984; Dunn and Ber-
are submitted to a cleaning process? This study was made gendahl 1987; Tegehall and Dunn 2006). They have found
on bare boards, typically assembled boards with that the distribution of flux and flux residues under indi-
through-hole components, and boards with surface mounted vidual components may be impossible to remove (or detect),
components. Flux types were: pure rosin, mildly activated even when utilizing CFC solvents, unless adequate stand-off
flux (cored solder wire) and solder paste fluxall were heights are specied. At the present time it appears that the
supplied by a preferred space product manufacturer. Results CFC-substitute solvents selected by ESAs contractors can-
showed the fluxes difcult to clean after 48 h storage at room not adequately remove the various classes of fluxes
temperature, by simple immersion into IPA. It was recom- employed during electronic assembly. Within each class of
mended that cleaning be performed within 48 h, preferably flux, the composition may vary considerably. Certain prod-
within 24 h. Heating the boards to 80 and 120 C did not ucts with known trade names have been chemically analysed
have an adverse effect on the recommended period before and found to vary considerably in composition and acidity
cleaning. The choice of cleaning fluid proved interesting. from one batch to another; and even these fluxes can contain
A proprietary solvent cleaner out-performed other solvents contaminant particles (Dunn and Chandler 1980; Nasta et al.
by a factor of two, for simple immersion cleaning. This is 1990). The 1980 study was remarkable because flux sup-
particularly important for surface mount components since pliers agreed that in-coming materials were not adequately
brushing and scrubbing had little influence. For plated hole analysed prior to them being formulated into deliverable
components brushing and scrubbing was effective. Cleaning liquid flux products. More recently the author visited a
efciency increased to the fourth power of the stand-off manufacturer of liquid flux, and it was clear that the natural
height, but the cleaning of SM components presents more rosin was sourced from both Portugal and Chinano
problems than through-hole components. chemical analyses were made on the incoming batches
These test results were considered by a panel for the (being produced from pine tree sap, they are certain to
ECSS-Q-ST-70-08 revision who adopted the requirements possess different chemistries and acidity). These crystalline
for: solids were then chemically modied by thermal treatment,
addition of solvents and other chemicals that gave consistent
(a) Flux and residue shall be removed within a maximum viscosity and halogen-activation.
period of 8 h after soldering operations.
392 6 Failure Analysis of Electrical Interconnections and Recommended Processes
Fig. 6.50 Actual ECM failures from operating electronics. From top growth progressing over dielectric and, leaded flat package showing
left, SEM image of ECM (similar to Fig. 6.33, but here a component ECM over conformal coating. The electrical leakage paths are clearly
lead is soldered to a nonfunctional pad). Ceramic chip with dendritic visible (Ring 2010)
patterns onto test boards to be soldered according to com- from these fluxes, as examined in this investigation, was
pany process and cleaning standards (IPC-TM-650 2000). very low (0.13 g/cm2 eq. NaCl). These, and the organic
The ndings from the test programme are depicted in acid fluxes, are unsuitable for space assemblies [poor
Fig. 6.51 and can be summarized as follows: properties were also found by Sohn and Ray (1995)].
The application of conformal coatings to PCBs has been
The surface structure of the epoxy top-coat present on both shown to improve their SIR in air, especially at high
sides of FR-4 PCB laminates makes them very difcult to humidities. This does not necessarily mean that the reliability
clean. Boards procured for space use should be thoroughly will be enhanced. If there are hygroscopic contaminants
cleaned prior to component assembly. Subsequent pro- beneath a conformal coating, the presence of the coating may
cesses such as soldering must avoid the use of chemicals or have very detrimental effects (Schweigart 2007).
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. 6.9.9 Conductive Anodic Filament
Residues from nonactivated or mildly activated rosin (CAF) Formation and Particulate
fluxes have been seen to increase the SIR of PCBs Contamination
operating under humid or dry conditions. However, there
may be severe corrosion of the metallic conductors. Electrochemical migration (ECM) has been described in the
Residues from low-solid fluxes based on carboxylic acids Sects. 6.6.3 and 6.9.8 and it must be noted that those
decrease the SIR. The level of contamination originating examples related to Mechanism 1, the growth of dendrites
6.9 Problems Associated with Soldering Fluxes 395
Fig. 6.51 Photographs (ac) showing appearance of some test boards for bare boards and all coated boards). d Compilation of results from
(comb patterns) after the environmental tests; to illustrate the effect of environmental tests: exposure to temperature/humidity then vacuum,
flux residues on PCBs (bare and with various spacecraft conformal with details shown on graph of insulation resistance against time in
coatings). a Elsold F-SW26. Low ionic content No corrosion, no SIR hours. Registered surface insulation resistance (SIRMohm) values for
problems, clear residue. (No SIR problem with any of the coated conformally coated test boards (all coating types gave similar results,
boards). b Kester 1585. Medium to high ionic content. Conductor track proving that water molecules are able to diffuse quickly through the
shows severe corrosion. Clear residues on PCB. No SIR problems. (No silicone coatings). The insert lists the ionic contamination levels on the
SIR problem with any of the coated samples). c Alpha 850, High ionic boards after poor cleaning (gures quoted are average values for three
content, plus conformal coating type CV-1140. No visual residues, but test boards soldered using the same flux) (from Tegehall and Dunn
severe leadtin migration. Very low SIR. (Dendrites and short circuits 1992)
on the surface of a PCB, resulting from lead ions going into thickness of a PCB. CAF testing to the IPC method dened
solution at the anode (one solder-coated track) and plating in TM-650 helps to determine the reliability of a PCB
out at the cathode. These growth have different morpholo- laminate material or a nished product. Incidents of CAF
gies, ranging from fern- or tree-like formations (Fig. 6.30) to failure are increasing because conductor spacing and overall
the nodular deposits in Figs. 6.49 and 6.50. Mechanism 2 part sizes are becoming smaller and smaller, the greater need
concerns the growth of conducting anodic laments for higher performance and an increased density of com-
(CAF) as shown in Fig. 6.52. CAF forms within the 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
follow glass bres (often surrounding individual bres), weave of the glass so that it cannot be seen in a horizontal or
cracks in the laminate or weaknesses in the resin bond. The vertical cross-section. Often the deposit is hygroscopic and
deposited lament is usually copper. Often CAF is not dis- washes away during sectioning. If poor PCB manufacturing
covered until electrical failure has occurred resulting in some processes are suspected, boards can be submitted to acceler-
localized burning, resin vaporization and breaking of glass ated testing when CAF may appear around anodic through
bres. It is a difcult laboratory exercise to make failure hole vias after 140 h at 100 V during climatic chamber
analysis of CAF: the lament is often very thin, it follows the 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 breglass such as that contained within PCB laminates.
CAF (Heltzel 2014) caused the re-visiting of publications by Exposed bres can penetrate into the skin and cause irrita-
Rogers et al. (1999) and Karavakis and Bertling (2004), and tions. Fibres released into the working area can circulate in
recent laboratory investigations appear to conrm the fol- the air and lead to itching of persons exposed skin.
lowing reactions: Wounds to the body and eyes can be inflicted by the wrong
use of cutting tools such as scalpels and even lead clippers.
Anode: Operators are trained to select lead clippers having a sharp, 90
cutting edge which is placed over the lead such that the per-
Cusolid ! Cunaq ne and
2H2 O ! O2 gas 4Haq 4e pendicular cutting sides are facing the component body
(ECSS-Q-ST-70-08). When hard component leads, such as
Cathode: those made from Kovar or hard-drawn ironnickel alloys, are
cut there is a very large release in energy (shock) which is
2H2 O 2e ! H2 gas 2OH
evident from the speed and distance of the projected small
Cunaq ne ! Cusolid length of lead wire. If the clippers are used the wrong way
round, the released energy can easily damage the component
Follow-on evaluations of defective multilayer printed circuit
or shatter its glass-to-metal seals. However, when correctly
boards, due to short circuits or deceased insulation resistance,
performed the operator must avoid damage to his own or his
have identied particulate contamination to exist between
colleagues eyes from the small projectile, and similarly this
internal conductors. Contamination, such as small bres, is a
small length of conductor must not be allowed to land in the
known problem on prepreg and copper clad laminates and is
midst of flight hardware. Irritant dermatitis can also be caused
possibly caused by air-borne particles as a result of poor
by the wrist straps worn by operators in order to avoid elec-
handling and storage during PCB fabrication. Accelerated
trostatic discharges. These straps often contain a stainless steel
testing (85 C, 75 %RH, with 100 VDC electrical bias) has
or nickel-plated earth connection that is also in contact with the
shown copper to migrate along bres bridging between internal
operators wrists. It has been considered that 10 % of all
conductor tracks (Mc Brien and Heltzel 2013). The
females have an allergy to nickel, particularly when it is lea-
contaminant bres were observed using an IR camera and
ched out from an armband with perspiration. This may be
SEM as part of the failure analysis. Long bres had been coated
compounded by the presence of small amounts of liquid sol-
with electrochemically migrated copper during the test period,
dering flux which are known to contain acids, amines, sol-
much like the CAF mechanism described previously. This
vents, and other mildly toxic substances. A study of skin
caused leakage currents and test failures. The detailed
disorders among hand solderers in the electronics industry
investigation concluded that existing cleanliness requirements
found that work-related dermatitis of the arms or forearms was
for laminate materials should be made tighter, particularly
commonplace and that the use of gloves did lower the
when laminates are being processed for the space industry.
prevalence of such disorders (Koh et al. 1990, 1994).
Colophony fumes are released during soldering operations
6.9.10 Potential Health Hazards and can have a mild chronic obstructive effect on the
in the Electronic Assembly Area breathing of female solderers (Lee et al. 1994). They are a
product from the solder flux, and are listed by the British
Spacecraft electronics are generally made as a low-series Health and Safety Executive as one of the six substances
production and it is common for individual circuit boards to responsible for most cases of occupational asthma (Evans
have unique designs and layouts which do not warrant the 1995). This is characterized by coughing, shortness of breath,
extra cost of conducting the trials necessary for setting up and wheezing, and chest pain. The solder flux fumes contain a
optimizing the many parameters associated with machine wide range of toxic compounds, including hydrochloric acid,
soldering operations. For this reason companies will employ acetone, hydrocarbons, and carbon monoxide. It is important
a skilled team of personnel who, dressed in their clean-room to utilize fume extraction units when space companies are
protective clothing, will assemble semiconductor devices and performing solder-assembly work. These extractors may be
PCBs using processes that involve the cutting of component sited locally on each operators work bench or may be
leads, the application of fluxes based on colophony or ami- positioned above a soldering machine such as is used for
nes, soldering with very hot irons, performing repairs, and wave-soldering. An important recent study has found that
mixing and applying staking compounds and conformal hand soldering operators in the electronics industry, using
coatings. This work, so-called hand-, or manual-soldering, is soldering irons and a solder of 60 %SnPb, are not being
labour intensive and can expose the operators to various exposed to a lead hazard (Jalbert et al. 1994). Care should be
working hazards including the risk of acquiring dermatitis. taken to avoid a strong air flow as this can affect the quality of
Physical wounds must also be avoided when cutting 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 alumina ceramics with molybdenummanganese-sintered
the joint being brazed. It is recommended that simple cal- metal powder and braze in a hydrogen atmosphere with Ag
culations are made utilizing the coefcients of expansion Cu or AuNi ller metals (see, for instance, Fig. 4.12).
presented in Table 6.7. These calculations should be fol- Routes for brazing alumina to Kovar, as used for battery cell
lowed up by trial operations using representative test sam- joining and parts of a travelling wave tube, are shown
ples. For most brazing alloys, effective capillary action schematically in Fig. 6.53.
occurs when the braze path gap at the brazing temperature is
maintained within the range 0.0250200 mm. 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
6.10.2 Brazeability of Materials and Braze
before brazing). Bronzes and coppernickel alloys must be
Alloy Compositions
in the stress-free condition prior to brazing. Oxygen-bearing
coppers may not be brazed in high hydrogen content
It is not possible to detail the optimum heating methods and
atmospheres owing to severe embrittling effects as illustrated
ller metals for the wide range of base metals and combi-
in Fig. 4.8. Alloys containing elements with high vapour
nations of base metals. Literature searches and experimen-
pressures are unsuitable for vacuum brazing.
tation must be performed before the brazing of uncommon
materials. It is important to know if the surfaces of the
Magnesium alloys
metallic piece-parts being brazed will oxidise or be reduced
These are usually limited to dip-brazing using
in a particular brazing atmosphere. The Ellingham diagrams
magnesium-based ller metals (BMg-1, BMg-2). Many
presented in Appendix 7 can be consulted to determine:
magnesium alloys cannot be brazed, so specialist advice is
always necessary.
(a) the relative ease of reducing a given metallic oxide to
metal
Nickel alloys
(b) the ratio of carbon monoxide to carbon dioxide able to
Nickel and some nickel-rich alloys are subject to intergran-
reduce the oxide to metal
ular penetration by certain silver-based ller metals, partic-
(c) the ratio of hydrogen to water vapour that will be able
ularly when brazed in a state of stress. Penetration may be
to reduce the oxide to metal at a given temperature.
prevented if PdAgCu alloys are used in dry hydrogen
without the use of a flux. The special nickelchromium
Limited details for various parent materials that have been
aluminium high-strength alloys require special fluxes to
successfully brazed for spacecraft applications are as follows:
remove their very stable oxide lms. A nickel-to-copper
Aluminium alloys joint using silver-copper eutectic is shown in Fig. 6.54.
Brazing is possible for those alloys whose melting point is
above that of the available ller alloy. Clad brazing sheet Refractory metals (W, Mo, Ta, and Nb)
normally requires no extra ller metal. All the All these metals readily form oxides in contact with air, and
non-heat-treatable alloys can be brazed. The most suitable embrittlement can occur when these oxides are exposed to
heat-treatable alloys for brazing are AA 1100, AA 3003, AA nitrogen- and hydrogen-containing atmospheres. Tungsten
6061, AA 6063, and AA 6951; see also Sect. 4.12. The alloys and molybdenum should not be heated above their recrys-
AA 2011, AA 2014, AA 2017, AA 2024, and AA 7075 are tallization temperatures (1300 and 1150 C respectively);
very difcult to braze owing to their low melting points. their surfaces must be thoroughly deoxidized, and brazing is
best accomplished in vacuum or under dry hydrogen, using
Beryllium alloys PdAgCu or AgCu ller metals. Tantalum and niobium
These alloys are highly reactive and extremely toxic; they should be brazed in argon, helium, or vacuum, using special
can be brazed with AISi and Ag alloys, in pure dry alloys based on Ni, Zr, Au, or Ti (specialist advice is
hydrogen or in vacuum. recommended).
costly CuAgZnCd ller metals are particularly suitable annealed condition. Also carbide precipitation in the heated
for furnace brazing. Graphite powder mixed with an area lowers their corrosion resistance (if high-temperature
appropriate carrier has recently been found to provide a good braze alloys are chosen these carbides can be redissolved by
bond between low-carbon steelsbrazing is in situ at 1200 heat treatment after brazing). The use of low-carbon alloys
C for 5 min under vacuumthe brazing liquid carburizes such as 304 L will minimize this condition. The stabilized
both sides of the joint, and this is followed by rapid diffusion alloys 321 and 347, however, do not undergo carbide pre-
of carbon into the steel. cipitation, and should be used wherever possible.
Stainless steel Austenitic nonhardenable stainless steels Ferritic nonhardenable stainless steels are best brazed
are subject to stress-corrosion cracking in the presence of with AgCu ller metals, and when exposed to water or high
molten ller material, and best results are achieved when humidity special braze alloys are required to prevent crevice
brazing is performed only when these steels are in their corrosion. Martensitic stainless steels requiring heat
6.10 Problems Associated with Brazing 403
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 piece-parts is maintained and often improved by controlled
atmospheres, particularly when they reduce the surface
One advantage in the use of controlled atmospheres in oxides. This will facilitate inspection steps and any nished
brazing is the elimination of potentially dangerous fluxes operations. The most common types of atmosphere are
and their residues which may become embedded within the dissociated ammonia, combusted town gas, hydrogen, inert
solidied ller metal. The ability of the atmosphere to reduce gases, and partial vacuum.
or totally prevent the formation of surfaces oxidation can be The operation of a high-temperature brazing furnace is
achieved under FURNACE BRAZING, but such atmo- shown, together with some spacecraft piece-parts, in
spheres are also benecial when used in conjunction with Fig. 6.56. The enclosed brazing chamber is superior to the
other heating methods such as INDUCTION and use of continuous furnaces, where there is often insufcient
RESISTANCE BRAZING. The surface quality of spacecraft oxidation protection. Stainless steel and nickel alloys require
location, and form of braze alloy), details of the brazing 2. Quality of joint
equipment (e.g. atmospheres, temperature, times with pos- 2:1 Contour
sible preheating and cooling cycles), approved methods for The contour of an outside ller joint shall be of a
flux removal, and nally inspection. uniform radius with a minimum amount of excess
Acceptable brazed joints are not ware generally charac- braze or flash over the adjacent surfaces.
terized by: 2:2 External defects
2:2:1 Surface porosity (see Figs. 6.57 and 6.58)
(a) clean, smooth surfaces with the absence of external External voids are not permitted.
porosity, Incomplete fusion is not permitted.
(b) full penetration of the braze path, Single porosity: no single pore shall exceed
(c) complete wetting between the ller metal and the 0.5 mm.
jointed surfaces without excessive melting or erosion of Concentrated porosity: accumulated total of
the base metals. diameters shall not exceed 1.0 mm in
diameter in any 10.0 mm of braze.
The visual criteria for rejection will depend on the criti- 2:2:2 Cracks
cality of the joint, and may include: Cracks are not acceptable regardless of
location.
(a) excessive braze metal, 2:2:3 Blisters
(b) inadequate contour of braze llets, Overheating of the base metal resulting in
(c) flux residue or other contamination, blistering on the surface of the base metal or
(d) presence of surface porosity and blisters, any plated layers shall be cause for rejection.
(e) evidence of overheating by penetration of the base metal 2:2:4 Excess braze metal
by the ller metal, or under heating as seen by presence Braze ller metal which may be in excess of
of unmelted braze alloy or inadequate penetration. that required of the joint is acceptable pro-
vided that it does not interfere with the
Other inspection test methods may include radiography, functioning of the completed assembly.
ultrasonics, and pressure leak testing (possibly helium leak 2:2:5 Undercutting
testing for ne leaks and dye penetration for gross leaks). At Melting or erosion of the base metal on the
the commencement of an unfamiliar brazing job it will be outside surface adjacent to the brazed joint
desirable to make destructive mechanical and metallographic shall be limited to a maximum depth of 5 %
tests on preproduction samples. These tests may include the of the stock thickness and a maximum of
proof loading of assemblies by subjecting them to loads 15 %, cumulative, of the braze llet length.
slightly in excess of those required in service, and corrosion 2:2:6 Unmelted braze alloy
and stress-corrosion tests, which simulate the intended The presence of unmelted brazing alloy in a
working environment of the part (e.g. battery cell cap joint is unacceptable and shall be cause for
interconnections). For high-reliability jobs random samples rejection.
should be selected from each lot and subjected to in-line or 2:2:7 Residual flux
end inspections which may include nondestructive and/or No residual flux shall be permitted on the
destructive tests. Dimensional inspections should also be surface of a brazed joint or base metal.
made for compliance with the relevant drawings. 2:2:8 Foreign material
The brazed joint shall be free of all foreign
material.
6.10.7 Inspection Criteria for Brazed 2:3 Penetration and internal voids
Aluminium Alloy Braze ller metal must have full penetration from
Waveguide-to-Flange Joints the external surface to the internal surface. The
criteria for internal voids (i.e. at the end of the braze
1. General path) shall be a maximum void length of 5 mm for
These inspection criteria relate to interconnections made individual voids. The total unbrazed area, including
by any method of brazing, including torch, furnace, trapped flux, scattered porosity, and voids, shall not
induction, resistance, and dip. It is based on the exceed 20 % of the internal length of the internal
requirements of MIL-B-007883C. interface area.
6.10 Problems Associated with Brazing 407
perform investigations and report to the project teams within conform with the relevant workmanship standard, or it may
a short period of time. Such activities were discussed in be an assembly that is judged to be insufciently clean and in
Sect. 3.3. need of a further cleaning cycle)
The problem may be related to the materials being used, Repairoperations performed on a nonconforming item to
the quality of the components, the suitability and repro- place it in a usable and acceptable condition according to an
ducibility of the process, and on the dexterity of the operator authorized repair procedure/standard. Repair is distinguished
performing the soldering-related task. Similarly, the from rework (repair may consist of a component change with
inspector who flags up the problem must himself have ini- all the associated solder removal and replacement proce-
tially decided if the situation could be rectied by rework, dures, or it may involve the xing down of a lifted pad)
repair, or modication. Here, it is important to distinguish Modicationthe revision of interconnections on a PCB
between these three activities as each will have different such as interrupting conductors and/or adding components
consequences. The presence of specially trained and certied and wire connections.
operators and inspectors will be essential if the correct
decisions are to be made (this responsibility will be expan- A maximum of three reworks are permitted on any one
ded on in Sect. 6.14, related to skills training schools). The soldered joint without the need for an MRB. Repairs require
following denitions originate from the European soldering some form of authorization from an MRB, as they might be
specications ECSS-Q-ST-70-08 and 70-28: proposed in order to repair some damage that occurred
during the assembly work, or, at the time of testing space-
Reworkthe reprocessing of a nonconforming item, to flight equipmentseach repair can affect the long-term
make it conform to drawing, specication, or contract (for a reliability of the hardware. Modications may result from
soldered connection this could be a solder llet containing a late changes to design when there is insufcient time to order
void that can be re-heated with a soldering iron to make it 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 6.12.3 Effect of Rework Electronic
repairs and modications are permitted to be performed on Components
any one circuit board by the contractor without the need for
authorization from the customer, but this will depend on the Visual defects on soldered llets are mainly caused by a
specic terms of the contract. The effects of such practices poor solderability of the component leads, or metallized
on the materials and individual parts being assembled will terminations, and the effect of high heat-sink by the pack-
now be addressed. 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
6.12.2 Cosmetics of Solder Fillets trade-off between accepting a cosmetically inferior joint and
requesting rework is that there can be damage to the com-
Quickly frozen solder, in the presence of flux to prevent ponent during the rework, or touch-up operation.
oxidation, produces a solder joint that is smooth, bright and Problems that have been caused during the reworking of
having an excellent appearance. Slowly frozen solder, due to spacecraft assemblies included the following:
different heat-sinking effects of surrounding conductor
materials, can look dull and lumpy with some evidence of Damage to adjacent joints and plastic encapsulants by the
dendrites from the early solidication of the lead phase in the soldering iron when the boards have a high package
solder. Multilayer circuit boards commonly have up to 14 density.
layers and may be assembled with side-brazed duel in-line Ceramic capacitors such as the types, CKR 05 and CKR
packages. Such connections will have a high heat capacity 06, have copper leads and internal lead-to-silver metal-
not only will they take longer to heat, they will also take lization soldered interconnections (see Fig. 4.36). These
longer to cool to room temperature and will have solder internal joints have melted when excessive heat or time is
llets with a different appearance to those of other smaller involved when reworking the PCB soldered joint. This
components. The high heat capacity connections may also causes the internal solder to make short-circuits between
have imperfect solder llets on the component side of the the silvered ceramic capacitance plates or sometimes to
board. Relaxations to the original issue of the PSS-01-708 produce open-circuits between lead and metallization;
specication now permit (ECSS-Q-ST-70-08) only a partial such defects cannot be visually inspected owing to the
llet on this side, as long as there is evidence of some solder components mouldable plastic encapsulant and are only
flow-through and proper wetting angles. exposed by equipment functional tests. Although this
The presence of small voids and blow-holes in solder problem was evaluated many years ago (Dunn et al.
llets may also be cosmetic problems. There is no evidence 1983), it continues to resurface on present-day projects.
that they will be associated with electrical service faults. Cracking of glass-to-metal seals in dual-in-line packages
They are usually caused by outgassing of the boards with the ingress of solder flux into the package due to
plated-through hole or by the entrapment of solder flux. reworking the solder llets from the component-side of
Corrosion due to flux entrapment has not been a problem, as the PCB. Here, the permitted minimum soldering distance
space assemblies used nonactivated fluxing products. Also, from the seal is given by the component specication as
during thermal cycling, the voids are through to act as 1.5 mm; this requirement was breached by the rework.
stress-relief points, as they are spherical in shape and would
contribute by crack blunting. Voids do not promote crack For information related to the assembly, inspection and
propagation during extensive thermal cycling as is evident reworking of Area Grid Arrays (e.g. BGAs and CGAs) the
from also Fig. 6.60a). These were the ndings from many reader is referred to Sect. 6.15.10.
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 6.12.4 Effect of Rework on Plated-Through
accept the cosmetic defect rather than apply mechanical or Holes
thermal damage during rework.
In the majority of cases, solder joints possessing minor Thick multilayer boards (MLBs) are most likely to become
defects should be left as-is without being submitted to degraded by the continuous thermal shocks that their internal
rework. Rework can damage the component, the PCB copper tracks-to-plated-through holes can experience during
(particularly if it is a thick multilayer board) and can make each application of a hot soldering iron. Microsections of
the solder alloy harder, and therefore less compliant with the these connections were shown in Fig. 4.35b.
substrate and component during thermal cycling. These A detailed investigation into the effect of repeated rework
points will be readdressed in the following sections. on the reliability of spacecraft MLBs has been performed by
6.12 Effects of Rework and Repair on Soldered Interconnections 411
Fig. 6.60 Illustrations to highlight the importance of knowledge, skills wetting of solder to component lead and plated through hole (pth) is
training is essential for understanding the ramications of rework. perfect. d This is a cosmetic defectrework to the solder llet by
a Appearance of soldered joint after very severe thermal cycling (1000 re-melting the solder will cause thermal stresses loading the delicate
cycles between 100 and +100 C). Thermal fatigue cracks appear at internal connections. If ground planes were present, they would have a
the llet but are not associated with the large internal voidthis heat-sinking effect and may cause the operator to apply excessive heat
indicates that large voids have no influence on reliability, electrical or pressure from the soldering iron, leading to damage of the chemical
conductivity being guaranteed in solder volume opposite void. b The bond between laminate copper and plated through copper. Microsection
pit appears as a blow-hole resulting from flux entrapment, it may satisfy in (d) has been etched to reveal interfaces and there are no debonds
the requirements of ECSS-Q-ST-70-08 if the bottom of the pit can be within pth
observed under high magnication. c Electrical connection is good,
412 6 Failure Analysis of Electrical Interconnections and Recommended Processes
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
with insufcient pressure there can be occurrences of no followed the manufacturers recommendationsall were
electrical conduction. In the case of conductive adhesives it mixed then degassed under vacuum. Two conventional
is very important to monitor the electrical resistance of joints surface mount devices were selected: 208QFPs with SnPb
continuously during environmental testingparticularly nish and 0.2 mm pitched copper leads, and, C2225 chip
during thermal cycling. Conductive adhesives and eutectic capacitors with SnPb plated metallisations. Stencil printing
tinlead solders were used during the reliability testing of could only be optimized for 8175A and E2116. Exposure to
flipchip interconnection technology. It was found that 40 C/9395 %RH for 504 h, followed by shear testing,
during temperature cycling, failures were detected at both showed all samples to decrease in strength. Storage under
hot and cold temperature regions, whereas the soldered vacuum did not degrade these joints. Powdery tin oxides
connections were successful (Boustedt and Hedemalm 1995; were detected within joints to the SnPb-nished PCBs,
Liu et al. 1996). Keusseyan and Dilday (1993) also found acting as a weak interface during exposure to hot/wet
that, although mechanically good, most of the conductive environments. Weak oxide layers within the bonds made to
adhesives they tested showed an increase in electrical the porous gold flash also produced weak interfaces. Elec-
resistivity during thermal cycling. The increase in electrical trical monitoring during humidity exposures produced open
resistivity of these joints was correlated to the material and circuits of the QFPs bonds, but not to the soldered joints, as
surface properties of the ller metal powders and to the seen in Fig. 6.68b. Under thermal cycling ECA joint resis-
deformation properties of the organic carrier. tance displayed considerable data scatter (Fig. 6.68c), there
A metallographic section through a typical surface-mounted were no deviations in resistance for the soldered intercon-
component joint made with conductive adhesive is seen in nections. The thermally cycled ECAs did not display the
Fig. 6.68. The electrical conductivity depends on factors such extensive oxidation/corrosion observed in humidity exposed
as the size, shape, and surface oxidation of the silver particles. specimen. Our conclusions were:
Their asperities must touch and must not be affected by
deformation or expansion of the organic matrix. ECAs cannot be used as drop-in replacements for con-
European spacecraft PCBs are only component-assembled ventional soldering (however, it is emphasized that fused
with soldering technologiesi.e. metallurgically joined tinlead is not a suitable nish for stencil designs, device
connections. Electrically conductive adhesives are not per- geometry or device co-planarity),
mitted for this use, but they have found a limited number of ECA joints possess electrical resistance at least an order
applications during the assembly of very small chip compo- of magnitude greater than soldered joints,
nents within hybrid packages. Hybrid technology processes dispensing and assembly for ne pitch devices
are strictly controlled/qualied for space-use and the her- (208QFPs) was more difcult than C2225 devices due to
metically sealed packages contain a dry gas. This condition their accurate requirements,
obviates some other problems associated with conductive Ablebond 8175A performed better than Epotek E2116
adhesives, such as the possibility of silver migration, oxida- under the environmental exposure programme, and
tion of the metallic particles, and the poor thermal conduction ECAs were not adversely effected by vacuum exposure.
of the adhesive (Harris 1995). Under conditions of vacuum
the heat dissipation from an active component is via its sol- Clearly, the ECA technology and processing methods
dered interconnection; silver-loaded epoxy has an inferior have advanced since these tests were conducted. It is noted
thermal conductivity and this must be borne in mind during that NASA have experienced silver-lled ECAs to be unre-
the selection of processes for spacecraft assemblies. liable on some space projects due to changes in conductivity
The question, is it possible to consider ECAs as drop-in with time, temperature and environment (Mielke 2002).
replacements for solder pastes? has been raised many times Newer materials and nishes will need to be evaluated, in a
when last-minute modications were identied for assem- similar manner to the aforementioned testing, before they can
bled PCBs. Also, this process was suggested should repairs be applied on current and future space programmes.
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 6.14 Training and Certification
suitability of ECA processing and its effect on standard
components and material nishes used for space projects 6.14.1 General
(Thomas and Dunn 1999). PCBs were FR4 having two
conventional track/termination nishes: fused tinlead and The most valuable resource available to space manufacturing
gold-flashed nickel plated copper. Three silver-lled epoxies and fabrication industries is trained people. Management
were selected as being space-approved materials (Ablebond personnel, scientists, and engineers will usually possess good
8175A, Epotek E2116 and EG8050HC) and processing 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 c The plotted graphs show Quad Flat Packs with SnPb coated leads
small crack in the adhesive between the epoxy and the chip attached to pcb tracks having either SnPb or Au/Ni nishes: bonding
metallization is termed pull back by the adhesive supplier who materials are SnPb solder alloy and two ECAs: b the effect of storage in
claimed that there is no risk of propagation during thermal cycling. humidity on joint resistance, and c the effect of storage then thermal
Chip metallization is platinumpalladiumsilver. The carrier is gold cycling between 55 and +100 C in vacuum on resistance (Thomas
thick-lm paste on alumina (courtesy of Jrgen Svenssen, SES). b and and Dunn 1999)
operators and inspectors involved in the actual manufacture of demonstrate, by means of practical and written examinations,
spacecraft hardware are also trained to high levels of com- that attendees have the manual skills and sufcient knowledge
petence and technical skill. Most equipment procurement of detailed requirements to perform operations that include
specications require that persons engaged in the operation the different types of welding, soldering, crimping, etc.
and inspection of so-called critical processes must possess Diplomas in welding are granted to successful candidates,
certicates indicating that they have been suitably trained. at both engineer and operator levels, by the national welding
Skills training courses are held at recognized institutes institutes of all major countries. Such facilities are not
throughout Europe and the USA. Final tests must widely available for the teaching and skills-training of the
6.14 Training and Certification 417
assembly processes necessary to companies building and Polymeric application on electronic assemblies, no ECSS
inspecting spacecraft hardware. This is because the space standard; NASA-STD 8739.1
market is relatively small and, although there is a need to
re-certify personnel every second year, the total number of The NASA standards marked * have been transferred to
potential students is not high. the requirements of the Space Applications Electronic
Hardware Addendum to IPC J-STD-001E Requirements for
soldered electrical and electronic assembliessince 5th
6.14.2 Certification for Electronic Assembly September, 2012.
Techniques Six European Centres perform training in a variety of
languages:
The training and certication of personnel involved in European
spacecraft electronic assembly techniques is an important DenmarkHytek, Aalborg
activity that reflects the changing requirements and technologies FranceInstitut de Soudure, Paris
of the space industry. Training courses have been organized for GermanyZVE and IFE, both located in Oberpfaffenhofen
mixed groups of inspectors and operators from more than 30 ItalyItalian Institute of Welding, Genoa, and
countries since 1973; the initial European courses were insti- UKUniversity of Portsmouth,
gated following an Information Exchange Programme orga-
nized by NASA and ESRO (now ESA) when engineers attended Details of particular courses can be found on the websites
a wide range of practical training courses at TRW, in California, of these schools. Training programmes follow the contrac-
in preparation for joint NASA-ESA projects such as Spacelab tual requirements of ESA and other agencies quality
(the follow-up to Skylab) (Dunn 1973). Those rst courses assurance and processes specications. The training pro-
followed European Space Agency hand-soldering requirements grammes provided by these six institutes are very similar but
for conventional components onto PCBs. A scheme for the are held in different languages. They meet ESA-Approved
training and certication of operators, inspectors and course Skills Training Schools document ESA STR-258 (Novem-
instructors (within industry) has been established (Dunn 2008). ber, 2008). A common level of training has been agreed for
The aims of the European training are to: Category 2 (Inspectors) and Category 3 (Operators) and
all schools have similar course formats. The end-of-course
Develop a pool of personnel trained in the best practice practical and theoretical examinations are continually
for electronic assembly harmonized. It is common that the listed schools also
Develop a network of organisations capable of providing provide courses organized by the IPCseveral of the
training to appropriate standards, and European schools teachers are also Master Instructors per
To improve the quality of electronic assemblies used in ANSI/IPC-A-610.
space projects. Corresponding courses to the NASA STD process
requirements have been held since the 1960s at:
Courses are held to cover the following European and
(equivalent) NASA process standards (see also Appendix 9 NASA Jet Propulsion Laboratory, Pasadina, California, and
for titles) as well as a range of IPC courses: NASA Goddard Space Flight Center, Greenbelt, Maryland.
Machine, wave soldering, ECSS Q-ST-70-07: NASA-STD A Training and Certication Board now supervises the
detail in 8739.3 (Dec 1997) approved schools, reviews the training programmes, class-
Manual soldering, ECSS Q-ST-70-08; NASA-STD 8739.3 room facilities, and lesson plans. Numbers of attendees to
(change 2, 2011) the European schools courses now exceed 1000 per year
RF semi-rigid soldering coax cables, ECSS Q-ST-70-18; and all individuals will have a range of backgrounds. They
NASA-STD 8739.3 and 8739.4 include Ph.D.-holding scientists producing space experi-
Crimping, ECSS Q-ST-70-26; NASA-STD 8739.4 (change ments; engineers aiming for a career in space design and
6, 2011)* manufacturing; technicians producing electronics for satel-
Repair and modication, ECSS Q-ST-70-28; NASA-STD lites, probes and launch vehicles such as Ariane 5 and Vega;
no equivalent* and persons from other industries wishing to achieve a
Wire wrapping, ECSS Q-ST-70-30; NASA-STD 8739.4 higher standard of workmanship (Sampson 2007). Of par-
(change 6, 2011) ticular importance are the training courses for Category 1
Surface-mount technology, ECSS Q-ST-70-38; personnel (i.e. Instructors and Examiners for any of the
NASA-STD 8739.2 (cancelled in 2011)* ECSS processes certications). This course will involve
Optical Fibres, ECSS Q-ST-70-51; NASA-STD 8739.5 either two or three weeks attendance and has been stan-
(change 2, 2011) dardized to:
418 6 Failure Analysis of Electrical Interconnections and Recommended Processes
Week 1: skills training in soldering at operator level (can hardware that highlights excellent quality products. Simi-
be waived if personnel have a current Category 2 or 3 larly, actual assemblies should be available to be used as
certicate for soldering per ECSS-Q-ST-70-08); workmanship standardsthese will represent conditions
Week 2: training for remaining solder-related processes where trainees can distinguish between acceptable and
(wave soldering, RF cable interconnection, repair of rejectable hardware during inspection sessions. Discussions
spacecraft electronics, and the assembly of can also refer to photographs of workmanship boarder-line
surface-mounted devices, following the respective pro- cases such as the example given in Fig. 6.60.
cess specications ECSS-Q-ST-70-07, 70-18, 70-28 and Each soldering or crimping process must be considered as
70-38). Advice is also provided concerning the selection part of a closed-loop, where inspection results can be fed
of materials for space applications, effects of vacuum and back for possible design or materials changes. Note that
thermal cycling on electrical systems, suitable cleaning some changes to the materials or processes must be agreed
methods/controls, and preferred conformal coatings; by the customer, as process qualication status may become
Week 3: solderless connections including crimping invalidated by such changes, even if they are thought to
(70-26), wire wrapping (70-30), and a review of the main improve the product. The writer recalls how an engineer
Quality Assurance requirements. Much emphasis is given authorized a slight change to the manufacturing route of a
to course presentation and teaching methods during particular insulated cable that was found to be difcult to
this nal week, together with a mandatory 20-min strip. The change was made to the automatic line where
demonstration of teaching abilities per person. spools were feeding stranded silver-plated copper conductor
wire into a machine which applied the Tefzel insulating
It is clear that the Category 1 courses have beneted from jacket. Simply, just before entering the insulating machine,
the considerable amount of knowledge accumulated by the the wire was made to pass over a sponge that had been
participants. Some attendees were experienced operators soaked in silicone oil. The end result was a wire jacket with
transferred from production jobs to training, but had little or excellent stripability. Solderability problems in industry
no experience of, or skills in, training. Some belonged to alerted quality personnel about the situation. The process
quality circles. Others had experience either in training or in change caused the wire product to become noncompliant
related activities within or outside the space business, but little with its original space qualied process identication
or no practical operating skills. On completing the training, all documentation. Unfortunately, wires from this contaminated
will have gained some knowledge of the metallurgical factors batch of production had been installed on the harnesses of
that control the reliability of these electrical connectionsthe three spacecraft where assembly work had utilized crimping
causes of good or poor solder wetting, flux reactions, galvanic processes. Laboratory investigations found the silicone oil to
corrosion, intermetallic growth, solid-state diffusion and cold have a low purity and a very high outgassing-under-vacuum
welding at mechanical interfaces, and the like. Attendees to value. Costly reworking operations were instigated to
these courses have beneted from the extensive practical and remove all the affected wires and cables from these space-
theoretical knowledge of both the instructors as well as other craft. Inspectors are permitted to request authorized rework,
trainees from different companies and institutes. Such inter- repair and modicationscompilations of approved meth-
actions will certainly be of use when they return to their own ods exist in some standards (e.g. ECSS-Q-ST-70-28).
companies and need to introduce new interconnection tech- However, they must know the underlying cause of the
nologies, address workmanship issues or organize their own problemthis may be related to poor-quality circuits,
in-house training courses. 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
6.14.3 Understanding Process-Induced the contour reveals the quality of wetting, and wetting is the
Failures and the Importance most important physical attribute of the connection in
of Workshops determining its strength and reliability. The question will be:
to leave as is, or to rework? Solder llets that have a
A brief mention is now made to the importance of under- non-ideal texture and lustre will not affect reliability. To
standing how design, materials, and processes affect the rework cosmetically inferior connections can cause severe
quality of the product. Skills training schools should possess damage to circuits (as was discussed in Sect. 6.12), and the
physical examples of failed equipment, obtained from correct answers to such questions will be one of the benets
industry or electrical appliance repair stores, which can be of good inspector training.
used to illustrate case histories and be available to pass Electrostatic discharge (ESD) is an important topic to
round the classroom as subject matter for discussion pur- cover during training. There are dedicated courses for ESD,
poses. A bonus is if these defective circuit board assemblies, but as electronic components are becoming smaller and
flux-corroded packages, etc. can be compared to similar real faster than before, they are becoming increasingly
6.14 Training and Certification 419
susceptible to damage by ESD. Grounded wrist bands must specications prepared by ESA, NASA, and IPC (1996).
be worn by all persons handling all space PCB assemblies. These have, in the main, been fully approved by the agencies
Individual ultra-sensitive devices will need special precau- procuring such electronic hardware. The processes are for-
tions, workstations must be grounded and a wide range of mally approved by means of the space projects Declared
other precautions should be undertaken as listed in Agency Process List (DPL), and all that needs to be ensured are that
standards such as the excellent NASA JSC technical stan- the written procedures are followed and that the assembly
dard 66552, dated 1st May 2013. It is good practice to have operators and inspectors have been suitably trained and
static shielding bags available for training purposesthese certied (as was described in the preceding section). Qual-
are often made of polyethylene to which an antistatic com- ication of hardware will be made on the spacecraft sub-
pound has been added prior to bag extrusion. system or unit, by exposing electronic boxes and the like to
As a means of keeping the training schools instructors the range of environmental testing dened by each particular
and the space community up-to-date concerning failure spacecraft contract.
modes, new materials, novel processes and the like, the Processes for SMT solder attachments are not so well
author has established an annual workshop together with the understood as those for leaded components. Until recently
University of Portsmouth entitled: Electronic Materials and they were considered as nonevaluated processes where no
Processes for Space (EMPS). The venue for these workshops information existed concerning their use in a space envi-
rotates from one school to another in order to allow each ronment and for which no useful data concerning reliability
school to present the ndings of their research, promote the and performance could be given to the project in question.
schools facilities and address the needs of STEM activities. The most important difference is that the SMT processes do
The presentations can be seen and downloaded from the not always provide for sufcient stress-relief between the
EMPS website since its inception in 2010. The schools also materials being interconnected and therefore accelerated
publish the results of their own work in a language appro- reliability testing should always be required to determine the
priate for that countryfor example, the extensive work of actual failure modes/mechanisms and provide data for which
Moliterni (2010, 2012) is based on the experience of the the space hardwares service life can be calculated. It has
Italian school, gained from performing courses, from already been shown that future spacecraft circuits will utilize
research, from a close contact with industry and solving large numbers of leaded and leadless surface-mounted
problems in its failure analysis laboratory. devices (SMDs), as these offer a greater opportunity for
higher-density electronics with improved electrical perfor-
mance and associated weight savings. Documents such as
6.15 Verification of Surface-Mount the ECSS-Q-ST-70-38 and IPC J-STD 001 (with space
Technology and Prevalent addendum) contain the main design rules that should be
Failure Mechanisms followed if spacecraft SMD assemblies are to survive the
environments of storage, launch, and operation. It is usual
6.15.1 Verification Testing that any electronics manufacturing company proposing to
build spacecraft equipment that incorporates SMT will be
An introduction to the design concepts and manufacturing required to demonstrate the reliability of the SMT by means
procedures that affect the reliability of surface-mount tech- of verication testing on test samples. If the technology
nology (SMT) have been described in Sect. 6.5. involves completely novel processes the procuring agency
Surface-mounted components have caused problems when will normally request so-called technology samples which
they were rst assembled onto spacecraft circuits. Because a will be assessed by that agency for any major problems
failed electrical connection can impair the performance, or associated with materials selection, wrong designs, or
destroy a major part or function of a spacecraft, this tech- incompatible metallurgical processing. When these tech-
nology has been identied by some approving authorities as nology samples are found to be acceptable, the new SMT
a so-called critical process. The main steps that are usually process line might be subjected to a systematic and inde-
required prior to the approval of SMT for spacecraft, and pendent examinationdened by ISO 8402 as an audit
probably other hi-rel applications, will now be covered. (here the company quality procedures will be checked for
Examples of common failure mechanisms will be presented suitability to general product assurance requirements levied
in order to illustrate the importance of understanding the on the contractor by the procuring agencywhich will
broad range of environmental stresses that affect SMT include an on-site review of process documentation, design
reliability. documentation, parts storage control, training and certica-
Soldering processes related to the assembly of conven- tion procedures, suitability of processing equipment, clean-
tional leaded components onto through-hole PCBs follow ing methods and their controls, whether all materials
well-established design rules which are covered by 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 Sweep speed: 1 octave per minute
stage it will be important to establish that a Process Identi- Duration: 1 cycle from 10 to 2000 to 10 Hz
cation Document (PID) exists within the companythis is
a summary document that lists the range of electronic parts Random vibration
associated with the SMT, lists the processing equipment, and
in particular lists the materials and all of the process pro- Frequency range: 202000 Hz (CPS) at 15 g (RMS)
cedures associated with the SMT line. Power spectral density: 0.1 g21 Hz
It is important that the contents of any SMT verication Duration: 10 min per axis
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 It is emphasized that companies and institutes should
agreed, together with a complete description of the test remain in close contact with the qualication authority (or
environments and a denition of what constitutes a test customer). Technical standards/requirements can be modi-
failure. A verication sample PCB assembly is seen in ed, or drastically changed as new modes of failure are
Fig. 6.69a, the assembly is conformal coated and has some identied. Materials may become obsolete or project envi-
electrical monitoring of parts. A failure criterion may be the ronmental test requirements may be made more or less
rst interruption of electrical continuity even when it is of demanding. In Europe, the ECSS Q-ST-70-38 requirements
very short duration. Only the continuous monitoring of are continually assessed by a working group, so it is clear
electrical properties during environmental testing will give that latest issues of such documents must be obtained or
an accurate and repeatable indication of when test failures freely downloaded from the internet.
occur. This is because short electrical interruptions are often A laboratory report containing all visual inspection
noticed when the test circuit is operating at the test tem- ndings, electrical test results, and micrographs will be
perature extremes. The failure mechanisms and their rates of compiled, but this may be quite voluminous and it is
occurrence should correspond closely to the failure mecha- therefore useful to also provide the customer with a sum-
nisms and thresholds that are expected to be seen at the mary table of the SMDs veried to the test specication.
actual end of life of an operational spacecraft. All veri- This table should include all important information,
cation test samples must be manufactured according to the including:
PID. These component-assembled circuit board samples
should include all the design congurations, materials, sol- the address of the manufacturing line
dering fluxes, and the cleaning methods which are envisaged the identication of the test specication
for future spacecraft hardware. If staking compounds, ther- full reference to the PID
mally conductive adhesives, and conformal coatings are to assembly process, e.g. vapour phase, hand soldering,
be applied to flight hardware, it will be essential that these etc.
same materials are present on the verication samples. Each solder alloy, e.g. 63Sn, 62Sn, 50 indium lead, etc.
of these materials will have different physical properties and conformal coating type, e.g. no coating, Urelane
might contribute to a particular failure mechanism. 5750LV, etc.
Thermal cycling will generally follow a specication each part should be listed according to SMD class, type
such as that proled in Fig. 3.12b and SMT verication will number, body dimensions (in mm) and identify any that
require that test samples survive at least 500 cycles (for were staked
soldered connections that cannot be readily inspected, such the PCB substrate should be well dened together with
as area grid arrays, the number of cycles may be 1500). its coefcient of expansion value and glass transition
Limits for the permitted extent of solder joint cracking, or temperature (see also Appendix 1 for specic values)
electrical failure are given in ECSS Q-70-38. Vibration state the full reference of the company verication report.
testing is usually performed in two axes, one parallel to the
board and the other perpendicular to the board. Vibration The author is particularly indebted to the companies who
levels related to the spacecraft requirements and the char- agreed that summaries of different SMT failure mechanisms
acteristics of the intended launch vehicle. A typical set of could be recounted in the following sections. These failures
vibration test levels would be: were experienced during verication testing. Each condition
for failure is different. It is the authors opinion that it is
Sine vibration
unwise to predict the service life for any surface-mounted
device from formulations based on either the CofnManson
Vibration amplitude: (peak to peak) 1070 Hz at 1.5 mm
equation [i.e. where the logarithm of the life is linearly
Frequency range: (702000 Hz (CPS) at 15 g
proportional to the fatigue ductility exponent of the solder
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms 421
alloysee, for instance, Lau and Rice (1985)], or the vari- spacecraft SMT can unfortunately involve imprecise mate-
ous Engelmaier expressions. These mathematical solder rial processes. Because of the wide number of variables it is
fatigue models should be used with great caution because unreasonable to try to predict a solder joints structural
they rely on precise physical and mechanical data associated integrity and reliability from models based on assumptions.
with each of the materials making up the soldered inter- Practical thermal cycling experiments and verication tests
connection. As will be seen in the following sections, 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 whenever possible, rigid multilayer PCB designs should
incorrect if the following examples of inconsistencies exist be employed; thin PCBs need to be secured at short
within the hardware: distances to stiff brackets in order to avoid resonance,
low natural frequencies and high amplitudes
residual stresses in the laminate position heavy parts close to the edges of PCBs and
the solder alloy becomes hardened by leaching from xation points
adjacent copper or silver surfaces, for indium solders use staking compounds and conformal coatings to damp
leaching may be from gold plate the board and increase mechanical support.
intermetallic layers at solder interfaces
variable stand-off heights (i.e. distance between the PCB It is emphasized that particular attention must be given
pad surface and the component termination that is lled whenever engineers contemplate using automatic electrical
with solder; see particularly Fig. 6.24) test equipments. Here, computer software manages some
presence of different organic materials beneath the electrical function checks on the assembled board by
component body, such as staking compounds, thermally directing long, spring-loaded metal probes onto predeter-
conductive adhesives, and conformal coatings (note that mined pad terminations. It is intended that the sharp probes
the thermal expansion coefcients and glass transition should descend onto each selected pad in close proximity to
points of these organic compounds vary from batch to the soldered SMD llets but without impacting the device.
batch) Close inspection of automatically tested assemblies conrms
for leaded SMDs, a variation in the Youngs modulus that when the components are exactly positioned onto their
and yield strength of the lead material, and residual termination pads, a series of penetration marks is present on
stresses from coplanarity problems or hot bar soldering the pad. Ideally, these impressions should be made into the
general variability in the solder joint such as voids, llet tinlead coating well away from the soldered device. In
geometries (especially with large solder llets on one practice, the SMDs show some misalignment on their pads
side of a component and small llets on the opposite and this, coupled with the inaccuracies of the test equipment,
lengthas seen in Fig. 6.25), and misalignment on ter- causes indentations to be made either onto the sloping side
mination pads. 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
6.15.2 Failure Under Mechanical Overloading major problem here is that the high loading force that is
applied to the probe tips will damage many of the mounted
Mechanical shock is a possible accelerator of failure with components. This was recently witnessed on several flight
SMT. It may occur when an assembled PCB is accidentally boards which had been automatically tested without proper
dropped onto a hard surface or in operation if the circuit is authorization (i.e. this inspection method was not part of the
situated close to a pyrotechnic actuator device. Certainly the approved Process Identication Document). Failed devices
ceramic packages will be prone to such accidental damage were microsectioned and clearly showed that brittle cracks,
but only slightly more so than conventional leaded compo- which passed throughout the complete thickness of the
nents. The failure of SMDs due to mechanical shock has not component, had been initiated at the point where the probe
been encountered by the author. However, a common source had made contact with the metallized surface.
of failure has been the detachment of SMDs from their PCBs
during the vibration test environment which formed part of
verication testing. 6.15.3 Failures Due to Board Flatness
SMDs with ne-pitched leads such as quad flat packages Problems
(some down to 500 m pitch) are the most frequent to suffer
from vibration-induced failures. The leads and interconnects The required nish for the majority of spacecraft PCBs is
flex, become work-hardened, and nally fracture. Other tinlead on the external copper conductor patterns. The tin
parts seen to dislodge or become damaged during vibration lead has a neareutectic composition and is initially elec-
testing are leadless chip carrier packages and ceramic chip troplated onto the copper and later fused by either immersing
components such as that seen in Fig. 6.69b, c. the complete circuit board into hot oil or melting the elec-
Recommendations for avoiding mechanical overloading troplated layer by infrared heating (see also Sect. 4.7). Most
of SMDs are: board distortion problems occur during this heat cycle.
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms 423
The base laminate is heated above its glass transition tem- solder fusing; any thermal treatments should be gradual and
perature (Tg Cthe temperature in C when the material involve dwell times so that the board materials reach thermal
changes from a glassy state to a plastic state) and the epoxy equilibrium; and, distorted boards intended for SMT appli-
or polyimide matrix becomes soft, causing a relaxation of cations must not be corrected by pressing at elevated tem-
the support medium which is usually woven glass bre or, peratures, as the warp will reappear when the board is
occasionally, woven quartz bre. reheated above its Tg C.
Another form of board distortion occurs when the board The assembly process engineers should also avoid ther-
is suddenly heated when an assembly is passed over a solder mal shocking the board during oven bake-outs to remove
wave during machine soldering. In all instances it is humidity prior to soldering or during the actual soldering
important to pre-heat to avoid thermal shock. In a similar operation; flux cleaning should occur only after the assembly
manner, immersion of a hot circuit board into a cold cleaning has cooled to ambient temperature. Ductile solder llets of
fluid can also cause distortion. The soldered components will soft tinlead or indiumlead composition, together with a
hold the assembly in a distorted condition. suitable stand-off height, are to be preferred.
Distortion is often referred to as warp or twist, and the Inspectors and operators should check for warp and
usual requirement when space-quality boards are purchased twist. Suspect assembled boards should be documented and
is that their flatness should be within a tolerance of 1 %. subjected to a company Material Review Board before they
Board suppliers will seldom offer a guaranteed flatness of are mounted onto supporting structures.
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 6.15.4 Failure Due to Co-planarity Problems
overall distortions, as was seen in the Fig. 3.18.
If SMDs are mounted onto distorted boards they will be The three most commonly used solder joint congurations
subjected to severe mechanical stress each time the assembly for SMDs are those known as leadless (Figs. 6.25 and 6.26),
is flexed or vibrated. This is particularly a problem when gull winged (Fig. 6.18d), and J-leaded (Fig. 6.71). The
warped assemblies are straightened during insertion of the gull-wing components have the disadvantage that they
board into a cardguide or when it is screw mounted onto occupy more board space than the other congurations. Their
stiffeners and positioned in an electronic box. The failure protruding leads can also become distorted during trans-
analysis of several cracked chip capacitors determined the portation and handling. Lead coplanarity is the term given to
cause of failure as the straightening of a warped multilayer all adjacent leads along the sides of a component lying in the
board assemblyone component is illustrated in Fig. 6.70. same plane. Noncoplanarity is the distance between the
Unfortunately, these types of failure are difcult to observe, highest and lowest leads when the package rests on a per-
as the cracks are extremely narrow. fectly flat surface. When some leads have become acciden-
PCB distortion can be minimized by taking account of tally bent their noncoplanarity can result in soldered joints
some simple rules: having widely different solder llets. In order to achieve
The circuit design should ensure that the conductor pat- joints having a similar visual appearance, the assembler may
terns on either side of the board are evenly distributed in be tempted to either press down the package so that all the
terms of density and direction; SMDs should be evenly leads are spread to make contact with their PCB terminations
distributed over the board surface and high package densities during hand-soldering. Alternatively, the hot bar method of
in any one region should be avoided; with multilayer boards reflow soldering may be usedhere a resistance-heated
the neutral axis (i.e. location of zero strain during manu- metal bar is applied simultaneously to the gull-wing leads of
facture, assembly, and test) should be at the centre of the the SMD. It has been used for components such as standard
boards thickness and this can only be accomplished if the flat packs, ne pitch quad flatpacks (QFPs), small-outline
copper foil thickness and the ground and power planes are integrated circuits (SOICs), and small-outline transistors
equally distributed on both sides of the neutral axis. Poly- (SOTs). High levels of residual stresses are generated, both in
imide base material is generally superior to epoxy resin as a the component lead and in the solder alloy llet, whenever
PCB matrix for SMT. the leads are forced against the PCB termination pads and
The PCB fabricator should procure only the highest held there during solidication of the solder.
quality of laminate, ensuring that it possesses no residual Solder joint stress-rupture failures caused by the residual
stresses; the boards should not be articially restrained stresses in strong Kovar gull wing leads that were forced to
during fabrication processes such as etching, plating, and lie flat against the PCB during the soldering account for
424 6 Failure Analysis of Electrical Interconnections and Recommended Processes
many qualication test failures. During thermal cycling, or leads may produce very small solder llets. As J-leads are
during high-temperature storage, the elasticity within the very difcult to inspect, these defects can go undetected.
lead causes it to move and become displaced out of the Electrical tests and metallography have clearly proven
solder. At slightly elevated temperatures the solder alloy has that noncoplanar leads that contain residual stresses are more
very little strength and will creep to nally release the lead, prone to solder joint failure, and after fewer thermal cycles,
so causing the joint to become electrically open-circuit. when they are embedded in a hard conformal coating such as
The problem caused by non-coplanarity is to a far lesser epoxy resin or a polyurethane compounds. This is probably
extent seen with J-leaded components where the leads have caused by the high expansion of these coatings which causes
become misaligned. Here, the damaged leads are usually the component package to be displaced in the z-direction
displaced upwards and out of plane. During soldering these (i.e. the same direction as the lead movement).
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms 425
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
having an identical COE. Figure 6.73 shows that this is not 6.15.5.2 Presence of Thermally Conductive
always the case and very localized COE mismatches within Adhesives as Staking Compound
the interconnection itself must be considered, such as the Excessive heat generated by power dissipation in compo-
differences between solder and the adjacent ceramic. nents can threaten the reliability of spacecraft electronic
Recent advancements in the production of new materials assemblies and in such cases a high heat extraction is nee-
which can be tailored to have a range of well-dened COEs ded. In vacuum, forced air cooling is not possible. Therefore,
have started to revolutionize the microelectronics, power if the heat dissipation through the component lead termina-
semiconductor, microwave and laser diode packaging tions and by radiation is not large enough, it will be nec-
industries. Reference has already been made in previous essary to apply thermally conductive adhesives (staking
sections to COE-compensated PCBs (see also the typical compounds) between the package and PCB. It is probable
values in Appendix 1)these PCB materials help to reduce that Area Grid Arrays (see Sect. 6.15.10) will not require
the strains that cause thermal fatigue failure of soldered extra conductive sandwiched between these packages and
joints. Similar advancements have led to new metalmatrix board due to the presence of highly conductive balls and
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms 427
Fig. 6.74 Evaluation of a commercial metal-matrix composite (MMC) microconstituents of the MMC after slight etching of the polished
electronic package based on silicon carbide particle reinforcement and section. The angular SiC particles range in size from 4 to 200 m
an aluminium alloy matrix. a Shows an X-ray micrograph of the 50 mm (10 m average). The particulate volume fraction is approximately
by 25 mm package and its lid showing no evidence of porosity, particle 60 %. The matrix is composed of aluminium, magnesium, silicon, and a
density variation across the parts, or cracking. Hermeticity was trace of silver together with the coarse A and C phases (analysed by
conrmed by helium-leak testing, b shows the pseudo-isometric ED AX to be compounds of MgSi and without the presence of
microstructure of the as-polished MMC in the planes longitudinal undesirable aluminium carbides). This MMC package is one third of
(L), short transverse (ST), and transverse (T) to the longest package the weight of a similarly designed Kovar packaged with the same COE.
dimension. These micrographs conrm that this MMC is fully dense, In addition, the thermal conductivity of this MMC material is far
contains an even distribution of the various SiC particle sizes, and superior to that of Kovar (100 vs. 17 W/mK) (micrographs courtesy of
indicates isotropic physical propertiesl. c Shows details of the various Mr. G. Ramusat, ESTEC, Noordwijk)
columns. However, ne pitch leaded packages have minimal of certain types of material based predominantly on
thermal conductive paths and often need additional heat outgassing-under-vacuum data (Tegehall and Dunn 2001b).
sinking. Thermally conductive adhesives, and similar com- Three products were initially selected: Epo-tek 930,
pounds used for staking (to provide more resilience during CV-2946 and Multimat M-4030LD. The latter product
vibration) will have different thermal expansion properties proved impossible to form into the kind of mounds neces-
than the adjacent materials of component and PCB. This sary to meet ECSS requirements and was abandoned. The
may cause stresses in the joints during thermal cycling. The test vehicles were leaded ceramic quad flat packs (CQFPs)
hardness of the adhesive is also important since a harder with 144 I/Os soldered onto polyimide-glass PCBs.
adhesive is more prone to transfer the stress. The CQFP lead-frames were alloy 42, with solder-plated
A review of the market for thermally conductive adhesives gull leads having a 0.65 mm pitch. In cross-section the
supplied to the aerospace industries produced 35 products that samples appeared as in Fig. 6.75.
were thermally conductive but not electrically conductive. Epo-tek 930 (two part epoxy system with boron nitride
23 products were both thermally and electrically conductive, particles as ller) was stated in the trade literature to have a
15 were encapsulants and 4 lm products were available. thermal conductivity of 4.2 W/mK, (also conrmed by lab-
These have been compiled into a report that justies selection oratory testing).
428 6 Failure Analysis of Electrical Interconnections and Recommended Processes
CV-2946 is a two part silicone material, also containing components were solder mounted (63Sn solder) without any
boron nitride, was stated in the trade literature to have a form of stress-relief onto electrolytic copper pads. The board
thermal conductivity of 3.8 W/mK, but laboratory tested to material is a form of PTFE (polytetrafluoroethylene) that is
be only 1.7 W/mK. mechanically attached to an aluminium base plate. Extensive
Outgassing under vacuum, per ECSS Q-70-02 gave: cracking was seen after 500 thermal cycles. The cause of
these cracks is thought to result from a high COE mismatch
Epo-tek 930 TML 0.01 %; RML 0.01 %; CVCM 0.00 % between the ceramicKovar component materials (5 ppm/
CV-2946 TML 0.36 %; RML 0.08 %; CVCM 0.00 % C) and the PTFE-aluminium substrate (about 22 ppm/C).
This mismatch has stressed the copper conductor such that it
Coefcient of expansion tested to ASTM E-831-93 gave: became work-hardened and tore at the point where there is
an abrupt change in section between the pad and track as
Epo-tek 930 26.3 ppm/C (50 to +50 C) and 90.93 ppm/ seen in Fig. 6.76.
C (100150 C) Circuits were later manufactured which passed all veri-
CV-2946 209.7 ppm/C for 50150 C cation tests. These supported three modications: use of the
soft 50 % indium50 % lead alloy; a slight stand-off between
Hardness (Shore A) Epo-tek 930 = 98, CV-2946 = 75. the ribbon lead and the copper pad to allow for stress-relief;
All assembled PCBs were verication tested to ECSS and a conductor track having a slightly wider dimension of
Q-70-38 (thermal cycling then vibration. 0.5 mm.
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 6.15.7 Failure of RF Cables Connected
be suggested for space-use. by SMT
Fig. 6.75 Macro-section to show stand-off, form of bent leads and the The ne lines and close conductor spacing utilized in
adhesive (white) which had been introduced through a hole drilled into spacecraft SMT may necessitate that conformal coatings are
the PCB at the centre of the QFP always applied to the nal, cleaned PCB assemblies.
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms 429
The main purposes of conformal coatings are to provide need support during exposure to vibration and acoustic
protection against airborne contaminants and humidity dur- noise) and it is almost impossible to remove for circuit repair
ing ground-handling operations and, with the exception of without resorting to mechanical abrasion.
polyxylene, to ensure that the components are mechanically The relative differences in the expansion and contraction
stable during exposure to shock and vibration under launch of the PCBs, the component parts, and the incorrect selection
conditions. In service coatings will prevent shorting due to of a conformal coating applied to these assemblies have been
conductive contaminants floating and resting between con- seen to produce catastrophic failures to both operational
ductors in zero gravity and they mitigate against tin whisker spacecraft in orbit and flight-quality electronics that were
growths. Conformal coatings can be applied by the immer- undergoing qualication tests. Actual failures are seldom
sion of PCB assemblies into uncured coating, but brushing disclosed in the scientic literature. However, a detailed
and spraying are the more usual processes as they have more analysis of the cause of an open-circuit solder joint discov-
chance of avoiding the lling of stress-relief bends and can ered on the Magellan spacecrafts radar digital unit was
achieve thinner coverages (weight-saving). Polyxylene made by the Jet Propulsion Laboratory, which led to the
(Trade name Parylene) is applied to assemblies by a vacuum successful repair of many pieces of affected hardware on that
processit has advantages being very thin, impervious to vehicle as well as the Galileo spacecraft. The problem was
moisture, low outgassing, and covers all surfaces; however, caused by the presence of a high coefcient of expansion
it provides no mechanical support to mounted devices (some coating (Solithane 113) located in the large gap (1.3 mm)
430 6 Failure Analysis of Electrical Interconnections and Recommended Processes
between Dual Inline Packages (DIPs) and the PCB surface component than when the component lies flush against the
(Ross 1989; Lee 1990). Similar problems with space hard- PCBs surface. They also provide for a sandwich of soft,
ware have been traced to damage caused by conformal ductile solder between the terminations which is able to creep
coatings during thermal cycling (ASTROS star tracker, when there is a mismatch in coefcient of expansion (COE).
Magellan data formatter, etc. reported in JPL Lessons This creep of the solder alloy gives some stress-relief
Number 275). when the component is subjected to environmentally
Previous sections (e.g. 6.54) have demonstrated the pref- imposed thermal-cycling conditions or when the component
erence of having suitably high stand-off distances between dissipates thermal energy under spacecraft operational
the solder terminations on surface-mounted devices (SMDs) conditions.
and PCBs. These stand-offs facilitate cleaning operations, as Several evaluation programmes have been undertaken to
flux residues can be removed more easily from beneath the 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)
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
7
1
1
cracked joint, also showing physical construction of the multilayer
7 = difcult
3
6
6
5
7
l = cheap
5
1
SCOTCHCAST
CONATHANE
URALANE
CV 1144-0
have been forbidden for the nal cleaning of PCBs and, per
112
280
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms 433
6.15.10 Area Grid Array (AGA) Packaging The preferred non-destructive inspection methods are
those using perpendicular and angled systems and X-ray
6.15.10.1 General transmission X-ray laminography. The principal of these
Integrated circuit packages for space electronics were, his- methods are shown in Fig. 6.84 and, with the inclusion of
torically, based on leaded devices designed with leads for tear-drops in pad design, this can facilitate inspection for
insertion into the plated through holes of PCBs. Only good solder flow and wetting, or rejectable items as seen in
so-called through-hole dual in line packages (DILs) now Fig. 6.85a, b. Clearly hair line cracks caused by the move-
remain in usagethese were followed by industries use of ment of this package during machine soldering will not be
pin grid arrays (PGAs) but such through-hole mounted large seen by radiography.
packages were not popular for the space industry. Nowadays, A treatise on BGA assemblies, optimization of process-
the vast majority of packages are based on surface mounted ing, materials and X-ray inspection has been issued by the
technology (SMT). They include devices, discussed earlier in IPC (2004) and is still relevant today. Of great practical use
this book, such as castellated chip carriers, quad flat packages is a wide-ranging booklet compiled by Bernard (2006) that
(QFPs) with L and gull-winged leads (recongured from covers briefly the salient theoretical, then practical issues
through-hole to surface mounting), and J leaded devices related to the X-ray inspection of SMT assemblies, with
based on ceramic or plastic leaded chip carriers. great detail being given to the interpretation of
Area grid arrays (AGAs) for space-use are all leadless. X-radiographs made of defective AGA soldered intercon-
The interconnection between AGA device and the solder nections. The requirements of SMT soldered joints are pre-
pads of the PCB is always by means of the solder which scribed in ECSS Q-ST-70-38 and workmanship standards
provides the electrical circuit connections as well as the are shown together with accept/reject criteria for wetting,
mechanical connection. The forms of the solder are either voids in solder balls, acceptable tilt angles for CGAs and the
balls or columns having a non-eutectic composition. Ball size and distribution of contaminant solder particles ejected
Grid Arrays (BGAs may, however, contain some eutectic from solder paste during machine soldering [Fig. 6.86,
SnPb solder at their interface with the PCB pad. Column (Wickham et al. 1999)]. Dr Bernards booklet would seem to
Grid Arrays (CGAs) are always composed of high temper- go hand in hand with the ECSS document, particularly when
ature solder, usually 90Pb10Sn, and can be reinforced with a the interpretation of X-ray images becomes taxing.
copper spiral ribbon as seen in Fig. 6.72. New developments have enabled high denition 2D
Circuit designs are becoming denser and hence SMT digital radiographs to be rendered by computerized tomog-
packages have evolved so that leaded devices with peripheral raphy (CT) to produce partial computerized scans. Once a
interconnections are now ne (and ultra-ne) pitch. AGAs have localized are is digitally separated, it can be closely exam-
developed from 250 to over 1000 interconnections (I/Os); often ined and rendered to become a 3D image (as illustrated in
with the same coverage area on a PCB as a 208 pin QFP. Fig. 6.87)here a space saving package on package
The assembly of AGAs (BGA and CGA packages) onto (POP) interconnection layer is manipulated to reveal two
multilayer PCBs involves steps such as: baking of PCBs and defective (incomplete) joints that had produced electrically
any moisture sensitive packages; the deposition of solder paste intermittent open circuits (Bernard and Golubovic 2012).
to the PCB pads using stencil printing and occasionally syr- This method has been developed to produce an automatic
inge deposition; inspection of paste deposits; AGA placement void detection system for AGA interconnections (Said et al.
by machine; solder reflow in IR or convection ovens; cleaning 2012) said to be applicable to all void types in solders
by solvents; further inspection (see next sections); minimal (planar, micro, pinhole, Kirkendahl and shrinkage). Void
rework; and nally bagging or storage in nitrogen lled con- percentages were also measurable by an operator from 3D
tainers until it is time to populate the board with additional rendered images where solder balls were seen as a series of
SMT devices. Both BGAs and CGAs are very delicate and 2D cross sectional images. Additional new work has been
must be handled with caution in order that the small, soft reported by NASA JPL (Ghaffarian 2013) for the 2D and 3D
interconnections are not bent or contaminated (see Fig. 6.82). inspection of CGA1752 packages after thermal cycling. JPL
results indicated a better denition of packaging and
6.15.10.2 Workmanship and Inspection assembly by cross-sectional viewing, although the 3D fea-
AGAs are not suitable for visual inspection although some tures were somewhat limited (less magnication) when
evidence of ball and column joint quality can be ascertained compared to the 2D X-ray versions. Defects in solder col-
from a close inspection of the rst lines of peripheral umns (micro-cracks, voids, tilting) could not be detected by
interconnections as shown in Fig. 6.83. Optical equipment the 3D method due to insufcient magnication.
and binocular microscopes can be used; the Ersascope has Certain electronic components are susceptible to damage
been specially designed for these packages. by ionizing radiation. During X-ray inspection the sample is
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms 435
soldered to polyimide-glassno failures after 1000 (g) Very small cracks were seen beneath the BGA and
cycles. CGA mounting pads (in the laminate material of the
(d) The use of underlls had no positive or negative effects PCB). A later study (Tegehall and Dunn 2003) showed
for the BGAs. these cracks to initiate in the low temperature regime of
(e) Underll beneath AGAs mounted on polyimide-aramid the thermal cycling, this might have been benecial in
caused cracking of the jointseven with the more that fatigue lives were extended due to the hinging
resilient CGAs, as seen in Fig. 6.89. effect of the copper pad.
(f) The volume of solder paste printed onto the PCB solder (h) A suitable destructive dye penetrant test method was
pads is critical for the fatigue life of CBGAs. A de- introduced for quantifying the extent of solder fatigue
crease of the paste volume to 0.068 mm3, about 10 % cracks within the soldered balls and columns.
less than the minimum solder paste volume recom-
mended by the device manufacturer, caused a decrease The failure of AGA balls and columns during thermal or
in fatigue life of 2040 %. other environmental test exposure can be monitored by
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms 439
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)
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 power supply units and other power distribution circuits) that
pins and posts, difcult accessibility with a soldering iron, as operate between 400 and 900 V. It was aimed to shed light
well as the inability to build up molten solder around the on whether protrusion-free solder llets are technically
joint because the materials being joined have both an required for spacecraft units operating under a low partial
excellent solderability and a high thermal conductivity. pressure. Samples were made by a certied operator using
Many hours have been spent in attempting to rework such space-approved materials having low outgassing-under-
soldered joints in order to cover all wire strands and avoid vacuum properties. They were tested at operational poten-
every protrusion. A detailed evaluation (Materassi et al. tials ranging between 400 and 900 V, with different elec-
2000) was made which concentrated on typical soldered trode spacing (insulation gaps) and under various partial
joints found on space-quality printed circuit board assem- pressures. Two distinct batches of specimens were produced
blies and wire bundle terminations (as used in electronic 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
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
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
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 signicant: when Napoleons men froze in the Russian
positive terminal to be rather smooth, but containing craters and a few winter when tin pest disintegrated the buttons on their
small nodules trench-coats; and, when Scott attempted to reach the South
Pole but was hampered by leaking kerosene cans due to tin
congurations with sharp solder llets were found to pest of their soldered seams. Tin pest is prevented from
have a Paschen curve minimum value of 550 V. This is occurring by adding small amounts of antimony and bismuth
only 50 V lower than the 600 V minimum detected for to pure tin when ingots are stock-piled by governments. Any
round, smooth joints. allotropic changes would render the element almost worth-
Two predominant factors in determining whether a pair less, and these elements, Sb and Bi, are now typically found
of joints will be prone to corona inception discharge are as traces in solder alloys. Tin pest is manifest in the images
the pad-to-pad gap and the condition of the pad edges seen in Figs. 6.101 and 6.102.
(e.g., copper pad edges that are clean, free of solder-flux Although tin pest is a potentially an insidious risk to the
residues, coated with a smooth layer of fused tinlead are reliability of electronics it does not appear to have caused
to be preferred). Figure 6.100a, c reveal that a small any reported failures in the space community. However,
portion of the right-hand pad had not been covered with laboratory studies have established that cast tin alloys can
solder: this should be a reject featurebut is only visible also suffer from tin pest if held for long periods (years) at
under a magnication of greater than *20 times. temperatures of 18 C, with even greater occurrence at
6.15 Verification of Surface-Mount Technology and Prevalent Failure Mechanisms 449
40 C (Plumbridge 2011). It has been stated that tin can reviews. The tin whisker study (Dunn 1987) included the
be converted back to its metallic state if heated above 13 C, subjection of more than 100 pure tin whiskers to cryogenic
but without regaining its structural integrity (Kostic 2011), temperatures. They were held in liquid nitrogen for periods
but this is not fully substantiated. Pure tin is prohibited from from 0.5 to 168 h, after which the samples were rapidly
the designs of European and US spacecraft (except in the transferred to an SEM. Despite many repeated and careful
case of fused-and-drawn tin coatings on stranded wires, examinations of the cryogenically exposed tin whiskers,
which are immune to transformation and whisker growth) so none were seen to have acquired a changed surface
tin pest is unlikely to have been the subject of any hardware morphology.
450 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
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
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
2.19E06
1.05E07
5.97E08
5.51E09
Sn96Ag4
Elec. resistivity ( cm)
resistance) of Sn60Pb40
bulk
1.75E06
3.30E08
1.75E06
2.69E07
1.34E08
2.68E07
4.18E08
2.86E09
[4]
Literature
Literature
Std. Dev.
Std. Dev.
Std. Dev.
Resource
Resource
Resource
Mean
Mean
LHe
RT
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
7.1 The Problem of Whisker Growth Since in-service electrical failures continue to result from
the current-carrying capabilities of whiskers, this topic is
The ultimate idea in distributed communications might be the repeatedly discussed between specialist materials engineers.
development of a wristwatch telephone system which pro- Commercial considerations are known by the author to have
vides for instant electronic communication between two prevented electronics failures analysis reports related to
locations anywhere on Earth. This concept of a miniature whisker growths from being made unrestrictedly available.
personal terminal is forwarded neither as science ction, nor Of particular signicance is the authors belief that little
as a prophecy of future technology. It illustrates the current work is being undertaken to assess the overall problems
rapid development of numerous communications techniques related to whisker growth and to investigate ways by which
that could make such capabilities as personal portable ter- it can be curtailed. The writer has encountered several
minals both feasible and economically viable. Two techno- instances of whisker growth that have seriously jeopardized
logical developments in particularcommunications an area of his concern, that of the reliability of spacecraft
satellites and package switchingare the largest contribut- hardware. In two instances, between 1998 and 2000, they
ing factors in achieving highly responsive, user-oriented involved the failure of geostationary communications satel-
telecommunications networks. These were the words the lites; caused by the growth of tin whiskers on relays, or
author used to introduce the problems associated with whisker switches, causing processors to be disabled by
growth in 1987we are now long past this milestone as today short-circuiting whiskers. The majority of modern compo-
miniaturization has enabled us to purchase wrist watch phones nents and assembled electrical circuits can be subjected to
with web browsers, GPS-maps and, although uncomfortable screening techniques that very effectively reduce infant
to view for long periods, TV and on-line videos! mortalities. Unfortunately, problems associated with whis-
Metallic lms are used as electrical conductors in very ker growths are not linked to such early failures; they occur
many microelectronic devices. The thickness of these lms during the useful life of electronics and are associated with
generally ranges from a few hundred angstroms to 25 m. physical metallurgical processes that lead to either catas-
The extraordinary development of the integrated circuit has trophic failures, or at best, a long-term wear-out mechanism.
gone hand-in-hand with an increase in reliability; however, The introduction of the concept of communication by
the speed of evolution itself brings about greater reliability wristwatch terminals was intended to emphasize the
risks due to the reduction in dimensions of the devices, but it reduction in dimensions of electronic devices. This scaling
seems that the design and construct the wrist watch-sized down since 1960 resulted in integrated circuits doubling in
portable terminals possess a totally acceptable mean time complexity every yearthe so-called Moores Law. The
between failures (MTBFs). The mechanism of electronic exponential progress of the integrated circuit can be fol-
device failure is the physical process leading to that failure. lowed; the average number of components on a silicon chip
Following decades of research, we now understood more being 5 in 1960, 100 in 1970, 106 in 1984, and 2 108 in
about electrical failure mechanisms related to the growth of 1990. This was possible due to the high yields, larger wafers
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 manufacturers to keep the cost per area down. It seems
packages of electronic components. The term whisker is that now, in the 2010s, this trend is coming to an end
used to describe bre-like crystals, usually in the form of because of the physical challenge lithography faces: the
monocrystalline laments, these are often highly conductive impossibility to scale down any further. Nowadays, com-
and are known to produce electronic failure by short-circuit. ponent manufacturers, stimulated by strong competition in
the market, are forced to offer devices not yet assessed, and specialist at the Centre for Manufacturing Technology of the
some of the initial problems are thus transferred to the user. Nederlandse Philips Bedrijven B.V. (1984), The progres-
As the designed lifetime of spacecraft missions has increased sive miniaturization in the electronic engineering elds is
over the last decade from 3 to more than 15 years (partic- being increasingly harassed by whiskers, which may cause
ularly for weather and communications satellites), it is clear flashovers, short-circuits, audible noise, etc. In this sphere
that state-of-the-art electronics with failure mechanisms that more than anywhere else, prevention is better than cure.
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 7.2 Analysis of Failures Due to Whisker
long lives: Pioneer 10 was 30 years, and Voyagers 1 and 2 Growth
continue to provide data after 38 years. Launched in 1977
these craft have explored Jupiter, Saturn, Uranus and Nep- 7.2.1 Molybdenum Whiskers on Metallized
tune (in 1989)in 2012 Voyager 1 made the historic entry Miniature Circuits
into interstellar space. As of 2015, Voyagers 1 and 2 con-
tinue to transmit signals to Earth as they monitor conditions An integrated circuit device was received for failure analysis
outside of our Solar Systemmore than 5 instruments are after the discovery that it contained some short-circuit paths
still working and are expected to continue transmitting data that could not be discerned by nondestructive
until 2025 when there may not be sufcient power to support high-resolution X-radiography. Tests were also carried out to
the operations. check for package hermeticity before package opening. In
Fibrous crystals, that we now call whiskers, have been this test the entire package was subjected to a helium pres-
observed and have interested scientists since long before the sure of several atmospheres for 15 min. No leakage was
advent of the electronics industry. A few naturally occurring apparent when using a helium sniffer mass spectrometer.
brous crystals were known four centuries ago, and an early The package was then drilled open in a specially designed
example, recorded by Robert Boyle in his famous work The vessel connected to the mass spectrometer. The smallest
Skeptical Chymist (published in 1661), was the observation amount of water or gaseous contamination contained within
that protuberances of silver on certain oxides increased in the cavity of sealed packages can be determined by this
length from one day to the next. analytical method. The internal atmosphere of the failed
It was not, however, until the 1940s that monocrystalline device was found to contain a large amount of water vapour.
whiskers were found to grow on tin-, zinc-, and The integrated circuit (IC) package was next de-lidded by
cadmium-plated materials and cause electrical failures. mechanically grinding the top until the edge weld had been
Problematic electronic circuits were rectied by maintenance penetratedthe top was then peeled back to reveal the chip,
staff who were employed to change components, make circuit, and microbonds.
repairs to burnt-out circuits, and also clean away the bundles The circuit metallic lm conductors had been prepared
of ne metallic whiskers which often appeared only as during manufacture by the chemical vapour deposition of
accumulated piles of dust. successively titanium, molybdenum, and gold multilayers
It is possible, in principle, to classify whiskers into two onto the silicon semiconductor chip. The total conducting
categories which differ primarily in terms of their growth lm thickness (referred to as metallization) was about 1 m
mechanisms. The rst category includes whiskers that thick and 20 m wide.
crystallize onto a substrate from the vapour phase, the liquid A close SEM examination of the circuit revealed a den-
phase, or the solid phase, with growth occurring at the tip. dritic type of electrochemical migration between conductors
The second category contains whiskers that have grown that resulted in the formation of numerous needle-like
spontaneously from the solid phase, with growth occurring whiskers. These are shown in Fig. 7.2. Such growths are not
at the base of the whiskers. A disquieting number of to be confused with electromigration, which occurs within
whisker-related problems stemming from each category of thin-lm conductors at current densities in excess of
growth continue to be identied during the assembly and test 106 A/cm2. An energy dispersive X-ray analyser (EDAX)
of spacecraft systems. Some examples are presented in the attachment to the SEM showed the whiskers to be composed
following sections. They may relate to development parts of molybdenum with traces of gold and chlorine. The TiMo
such as that shown in Fig. 7.1 which possesses a large layer was intended to serve as a barrier to gold dissolution
number of spectacular growths, or to flight model transistors into the silicon substrate and as an agent for bonding the
with a single whisker that has grown to produce a glassication to the metal stripe. During electrical testing the
short-circuit between its tin-plated lead and the metallic circuit had operated under d.c. conditions, occasionally with
capsule of the transistor. Clearly a greater vigilance must be reversed bias. Failure was evident when the large leakage
applied during the selection of materials for these parts. To current through the short-circuiting molybdenum whiskers
quote from Mr R.J. Klein Wassink, a former materials became sufciently high to cause the microcircuit to fuse.
7.2 Analysis of Failures Due to Whisker Growth 463
The IC device had received a 168 h (1 week) burn-in at However, migrating gold-resistive short-circuits were dis-
125 C, but was noted to fail only after several thousand cussed by Shumka and Pietry (1975) with co-workers in the
hours testing at room temperature (during the mid-1970s. They claimed that gold dendrites form by elec-
space-qualication of this new technology component). The trodeposition from a gold halide complex solution. The rate
presence of trace amounts of Cl, detected as a surface layer of deposition from the solution was limited by diffusion.
on the whisker, suggests that contamination was present In the case of the molybdenum whiskers, three stages of
within this device before it washermetically sealed by the whisker propagation are suggested: dissolution of molybde-
fabricator. Halogens in the presence of water and electrical num by a halide solution, then an incubation period for the
bias are known to be electrochemically active. formation of active sites at the molybdenum surface (either at
Occurrences of molybdenum whisker failures have not pores in the gold or at the track edges), followed by growth
previously been reported as far as the author is aware. across the electrode gap. Figure 7.2b reveals Mo whiskers to
464 7 Whisker Growths
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. 2. Use contamination-monitoring equipment to check for
Normally growth would be expected from a cathodic to an ionic residues (see footnote 1).
anodic stripe; in the case of the failed IC, the current had 3. Cover the device circuit with a layer of electron beam
operated in both directions, resulting in the formation and evaporated quartz (taking care to mask over the surface
growth of whiskers away from adjacent stripes. areas required for microwire bonding); this manufactur-
Possible actions that are suggested to alleviate the growth ing step is already undertaken by the majority of
of molybdenum whiskers are: microcircuit fabricators.
b Fig. 7.3 Observations and model of tungsten whisker growth on showed substantial thinning caused by high-temperature vacuum
non-potted TWT heater coil. a X-radiograph of failed cathode/heater exposure. e Model of tungsten whisker. The smooth sides act as a
assembly (10). b and c Optical microscope views of tungsten whisker catchment area for tungsten atoms; they diffuse to the active tip
growths. d Detail of whisker growth on part of tungsten surface that
7.2.2 Tungsten Whisker Growth Within electrical short-circuiting may ensue. They also act as
Travelling Wave Tubes sources for corona discharge which would cause power
surges within travelling wave tubes, causing them to switch
A critical device used in a high-power wide-bandwidth com- off. Another failure mechanism would be for the whisker to
munication satellite is the travelling wave tube (TWT). This become negatively charged, break off, and be transported by
vacuum device consists of an electron gun (cathode and the high internal currents within the travelling wave tube.
anode) producing an electron beam, a slow-wave structure in Under zero gravity the whisker could create extensive fluc-
the form of a helix surrounded by magnets used for beam tuations at the cathode emitter surface.
focusing, and a beam collector. In operation, the radio fre- Modications to the cathode heater coils were success-
quency signal to be amplied is impressed on the helix at the fully incorporated to eliminate the possibility of whiskers
input end of the tube and takes the form of a wave travelling on nucleation. The main design change was to use a tungsten
the helix. Interaction between this wave and the electron beam heater coil that was coated with a thin layer of alumina paste,
results in an amplied signal at the output end of the tube. For and, once inserted into the heater body, was back-lled with
operation of this device, high voltages provided by a suitably a potting of densely packed ne alumina particles. By so
designed power supply must be impressed on the anode, embedding the heater, a uniform temperature gradient was
cathode, and collectors. Isolation of these TWT elements is achieved along its length, and free tungsten surfaces were
achieved by the use of ceramic insulators. Dispenser cathode not exposed to the vacuum environment.
assemblies with circular emitting surfaces and containing A review of the literature indicates that this potential mode
coiled tungsten heater wires were submitted for metallurgical of failure has not previously been reported. Tungsten whiskers
examination. One sample had failed after 44,092 h life testing deliberately grown from the vapour phase have been studied
because of tungsten heater lament irregularities. by Fumio (1974) and found, by transmission electron micro-
The failure analysis photomicrographs (Fig. 7.3) of the scopy, to have increased in length by the screw-dislocation
defective heater coil show that the effective cross-sectional method of crystal growth. It is known that tungsten whiskers
area of the tungsten heater wire has been reduced by subli- have excellent electrical and mechanical properties and can be
mation of the tungsten metal at the high service temperatures used for eld emitters. With this objective several researchers
(about 1100 C) during the extensive life of this vacuum tube. have attempted to grow these whiskers using different growth
Close examination of the heater coil through a binocular mechanisms. A recent article documents the successful pro-
microscope showed the coil surface to be covered with small duction of tungsten whiskers by a vapour phase deposition
whiskers as shown in Fig. 7.3b, c. The whiskers were analysed method, seeded by a nickel catalyst (Ma et al. 2012). Growth
in the SEM and found to be of the same composition as the was along the <100> direction. A complex growth process
wire, i.e. tungsten. The whisker detailed in Fig. 7.3c has a was postulated: water vapour containing WO2(OH)2 was
length of 46 m and a diameter of 7 m. A typical whisker was initially produced by tungsten oxidation, this was decomposed
closely examined in the SEM (Fig. 7.3d); it has a straight to produce a liquid tungsten fluid. Tungsten atoms then
cylindrical form with no appreciable change in diameter along nucleate at a site where nickel is present in sufcient con-
its length, nor at its tip or base. centration within this liquid phase; homo-epitaxial growth
At lower magnications it was seen that the whiskers follows as tungsten atoms deposit at the nucleation site on
grew in random directions with respect to the coil surface. It a <100> plane. Growth follows from the tip of the nucleation
is postulated that these whiskers grow by the condensation site to form a continuously growing whisker in a very similar
of tungsten emitted from a local sublimating hot spot on the manner to that proposed in Fig. 7.3e but in a liquid phase
tungsten heater coil. A temperature gradient of only a few environment rather than the vacuum environment of the
degrees causes the tungsten vapour to condense without cathode heater narrative.
chemical change at a cooler growth surface. The mecha-
nisms of growth is probably similar to that reported by 7.2.3 Metal Oxide Whisker Precipitation
Nabarro and Jackson (1958) of mercury whiskers produced in Glass Seals
in vacuum at 63 C, and shown schematically in Fig. 7.3e.
The presence of whiskers, like those of tungsten, is most The electrical leakage of high-quality integrated circuit
undesirable in vacuum tubes operating at pressures below packages has been investigated. Leakage currents were
104 mm Hg since, once they have grown long enough,
experienced at locations in the glass-to-metal seal region
7.2 Analysis of Failures Due to Whisker Growth 467
Fig. 7.4 Metal oxide precipitation in glass seal. a External view of dark-eld illumination. The oxide-rich glass volumes are highlighted.
electrically leaky IC package. b Transverse polished section through d Excessive glass-rich oxide whiskers (arrowed)
lead-to-package glass window seal. c As (b) but viewed under
between the IC packages and certain of their leads. A pho- Decarburizing is performed by ring the Kovar pieces in wet
tograph detailing these glass seal windows is shown in hydrogen at near 1000 C for about 30 min. The oxidation
Fig. 7.4a. step is the second step-during which a thin layer of metal
To discuss the failure mechanism it is necessary to briefly oxide is built up on the metal leads and component bodies.
outline the main steps in the manufacture of the IC packages. Suppliers of Kovar today recommend the following heat
The package and lead items are formed from Kovar, an iron treatment for glass-to-metal sealing: oxidationheat treat in
nickelcobalt alloy developed at Westinghouse in the 1940s an electric air furnace to 850900 C until parts are cherry
to provide the required match with the expansion properties red (dull red heat). The length of heat cycle is approximately
of a variety of soft lead and borosilicate types of glasses. 3 min, but due to differences in humidity and furnaces, the
The IC package is fabricated by assembling a base plate, a proper cycle needs to be varied. Then reduce heat approxi-
lead frame, and a ring or wall frame together, using a simple mately 10 C per minute. When parts are cooled, oxide will
jigging device. Ideally, the leads should be centrally posi- be formed. The oxide may appear from light grey to black in
tioned inside the open windows formed by the slots in the colour. Black is normally considered to be over oxidation
wall frame and the base plate. Inspection of the photomi- and is not desirable for a good glass to metal seal.
crographs in Fig. 7.4 reveals that this correct alignment was This is a difcult step to perform as both under-oxidation
not achieved in the defective devices. and over-oxidation must be controlled. Too little oxide has
The success of the matched-glass sealing process is been seen to cause lack of bonding (hermetically leaking),
totally dependent on three heat-treatment stages: decarbur- and too thick an oxide causes damage to the lead and also
ization, preoxidation, and sealing (Pask 1948). results in leakage (McCormick and Zakraysek 1979). The
468 7 Whisker Growths
third step by which glass flows and bonds to the metal or the H2, N2, and H2O atmosphere is now possible (Yext et al.
metal oxide is complex, and not entirely understood (Borom 1983). More recently it has been established that glasses can
and Pask 1966). The oxide layer may be composed of a be supersaturated with iron ions (a constituent of Kovar) and
surface scale layer and an underlying layer of intergranular on cooling from the glass melt, the spontaneous crystalliza-
oxide. The scale layer dissolves into the dried tion of iron oxides may occur (Somiya et al. 2013).
powdered-glass (frit) coating once the glass becomes molten. Magnetite is a conductor of electricity: about 2.5 104
Frequently, this sealing is performed in a H2N2 furnace ( m)1 according to Tsuda (2000) and, even present as
atmosphere at 1000 C for 6 min. The red pieces are minute crystals in a glass seal, is certain have been the cause
removed from the heat and allowed to cool in air without any of the observed leakage currents.
further annealing. The external dimensions of the IC package shown in
The composition of the glass frit employed for sealing the Fig. 7.4 are approximately 6 4 1 mm. The glass window
defective ICs was not revealed by the fabricator. Initially it seals are rather miniature, with a measured surface area of
was thought that a soft lead glass had been used and that the 0.3 mm2. However, even these dimensions restrict the
hydrogen atmosphere during sealing reduced some of the package density of components on present-day printed cir-
lead oxide to its elemental (and electrically conductive) cuit boards (PCBs). Component packaging technology is at
form. Failure analyses employing ESCA have identied this moment being revolutionized by the development of
such a failure mode where Pb2O3 and Pb were detected on vapour-phase or condensation soldering. Present-day com-
the surfaces of electrically leaking lead glass seals (Kelly ponent packages with leads that are designed for hand or
et al. 1984). wave-soldering into the plated-through holes of PCBs are
The defective packages were mounted in plastic and were being replaced by components of at most half their size.
metallographically prepared by standard techniques. Fig- With the introduction of even smaller glass window seals for
ure 7.4b shows a typical lead seal window. The adherence of the new generation of component packages it is essential to
glass to metal can be attributed to both ionic bonds and the have strict control over the scale and internal oxidation
physical roughness of the metal window. The photomicro- layers during lead-sealing operations. The new processing
graph shows that the oxide scale has fully dissolved into the renements, together with improved jigging to ensure the
glass. Also, the intergranular oxide, which formed at the proper alignment of leads within the IC, or dual-in-line
Kovar grain boundaries, appears to provide for a good package windows, should prevent the growth of oxide
chemical/mechanical bonding of glass to metal. The extent whiskers that could result in leakage currents.
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 7.2.4 Integrated Circuit Failure Modes Due
close proximity to their wall frames, as for instance shown in to ElectromigrationAluminium
Fig. 7.4d. In this case, during sealing, the volume of glass Whisker Growth and Solder Joint
has dissolved all the metal scale from these closely adjacent Voiding
surfaces. On cooling from 1000 C, this volume has become
supersaturated with metal oxide, and a glass-rich oxide Whisker growth from aluminium thin-lm semiconductors
phase has nucleated at the intergranular oxide locations on has been noted under several sets of conditions. Blech et al.
the metal surface. This oxide phase has grown in the form of (1975) reported whiskers after high-temperature (350 C)
whiskers. Qualitative EDAX analysis of the whiskers was annealing. These whiskers developed and grew, so relieving
difcult owing to the dielectric properties of the glass. No the compressive stress generated in the aluminium by the
lead was present, and only iron was detected within the thermal mismatch between aluminium and the underlying
oxide whisker. silicon dioxide. They were also seen to grow when the
It is assumed that the observed whiskers are composed aluminium lm was sandwiched between a silicon dioxide
primarily of Fe3O4 (magnetite) which has a cubic structure, substrate and a thin cover of titanium nitride. The growth
is known to occur in whisker form, and is magnetic. These mechanism was observed directly by transmission electron
laments (occasionally present as platelets) have a far lower microscopy, and whisker nucleation was seen to occur at
electrical resistivity than glass, and, as can be seen in grain boundary junctions, single grains, or hillocks. Alu-
Fig. 7.4d, it is likely that they connect the lead to the IC minium whiskers have grown by a similar mechanisms to
package wall (a distance of only 10 m) thereby causing the cause extensive damage to a number of types of thin-lm
measured leakage current. An industrial study focused on the metal-oxide semiconductors. Extraordinary SEM photomi-
preoxidation of Kovar as the most critical step in the matched crographs are reported by Turner and Parsons (1982); they
sealing process. A tight control over the depth of scale and show long (200 m) whiskers growing from silicon-doped
intergranular oxide, before sealing, by the precise blending of aluminium metallization pad areas adjacent to gold
7.2 Analysis of Failures Due to Whisker Growth 469
thermocompression bonded wires. Some whiskers were long thick, and between 7 and 10 m wide. Manufacturing pro-
enough to short to the bond wires from adjacent pads, some cesses ensured a small aluminium grain size, and the con-
shorted to conductor tracks, and some were even found to be ductor stripes were nished with a thin glass passivation
shorted to other whiskers growing from adjacent pads. layer. The results of theoretical studies predicted that at
The present author has evaluated the failure of spacecraft current densities of 1.2 106 A/cm2 the device conductors
semiconductor devices resulting from the growth of alu- should withstand several hundred hours before failure. The
minium whiskers. The communications industry is com- prescribed test conditions were noted to result in device
mitted to the development and use of Very Large Scale mean time before failure (MTBF) of approximately 50 h.
Integrated (VLSI) devices that are being designed to incor- SEM examinations revealed the classical modes of electro-
porate submicron-sized circuit components. Futuristic migration failure illustrated in Fig. 7.5:
user-oriented telecommunications networks and new forms
of package switching were postulated in Sect. 7.1. Indeed, (a) Nucleation and growth of voids
we are now seeing the fabrication of microelectronic devices
with conductor spacings and conductor widths of only a few The photomicrographs shown in Fig. 7.6 detail the damage
thousand angstroms. As the state of the art in integrated caused to the aluminium metallization due to electromigra-
circuit design and fabrication approaches VLSI, the ability of tion. Mass flow has caused local mass depletion (holes
metallic interconnections to resist solid-phase whisker appear where the mass flow leaving these areas is larger than
growths becomes an increasingly important consideration. the mass flow entering). Under the high current stressing
The mechanism of electromigration has been reviewed by there is a transport of ions along grain boundaries. For alu-
Richards and Footner (1984), Ross and Evetts (1987), minium this is generally in the direction of electron flow.
Hummel (1994), Niehof (1995) and Morris (1996). It is Figure 7.6 shows voids to have formed at what appear to be
assumed that aluminium metal is transported by momentum ne grain boundaries; they are present only in regions
exchange between thermally activated aluminium ions and toward the cathodic direction of the conductor. A motion
conducting electrons. The aluminium ions travel in the lm study has been made of void formation in an SEM while
direction of electron flow and form hillocks and internally an aluminium conductor was actively undergoing electro-
stressed regions, while voids in the lm of thin conductor migration (Thomas and Calabrese 1983); the most striking
areas form in the opposite areashown schematically in observation made from the lm was that voids were actually
Fig. 7.5. migrating toward the cathode. The voids were dynamic,
Electromigration has been observed during the evaluation continuously changing shape as aluminium disappeared
of a novel semiconductor device. For commercial reasons from the leading edge and relled at the trailing edge. It
only the areas of the circuit directly related to this failure seems likely that as electrons and aluminium ions (in grain
mechanism will be included. The device was designed for boundaries) flow toward the anode there is a flow of
long operational life under normal ambient conditions, but vacancies in the opposite direction, producing local vacancy
accelerated testing was made at a temperature of 150 C and supersaturations. Preferred void nucleation sites appear from
under slightly greater electrical loads than would be expec- Fig. 7.6 to be at grain boundaries both within the conductor
ted in service. Conductors were of pure aluminium, 1.2 m and at the edges of the conductor. Open-circuit failure is the
result of extensive void accumulations.
Fig. 7.6 a and b Void formation in aluminium metallization 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
(c) Aluminium whisker growth
Whisker growth is quite unpredictable. Long forms may The reliability of VLSI as well as other thin-lm semi-
develop at conductor edges, and they may have very low conductor devices may be adversely affected by increased
projection angles as illustrated by Fig. 7.8. The longest current densities in conductor tracks. A detailed account of
whiskers are seen to grow in regions where there is a lack of electromigration-induced failure calculations was published
voids; they exist at the anodic side/site of a conductor. The by Ghate (1982). The objective of the Ghate paper was to
accumulation of material at silica-to-silicon steps such as in dene some current density guidelines. Ghate emphasized
emitter metallizations, or as shown in Fig. 7.9, can also that to ensure the reliability of VSLI devices, current densities
create hillock and whisker growths. 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 471
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 integrated circuits (Hummel 1994; Feng et al. 2013). The
Ghate suggests that the device depicted in Figs. 7.67.9 had following four conductor parameters are identied: current
been overloaded. It is recommended that the best current density; geometric dimension; material properties; and, the
density for these devices should be revised downward to temperature of the metallic track. Feng describes, using the
2 105 A/cm2. However, after three decades of research on Monte Carlo sampling method, how each parameter can affect
electromigration a real breakthrough to completely eliminate the susceptibility of the others for physically causing the
failures has not been found and it is necessary to analyse the conditions for electromigration. It would appear that temper-
four main parameters involved with the electromigration ature has the least influence on electromigration and for an IC,
phenomenon in order to signicantly improve the reliability of 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
Fig. 7.9 a Some whiskers have a pyramid-shape, longitudinal
within the solder (mass diffusion and coalescence of
striations, and bases with a diameter of the order of 4 m. b They vacancies) caused by electromigration and this lead to
grow at SiO2-to-Si steps owing to electromigration increased ohmic resistance and failure.
472 7 Whisker Growths
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 473
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
Fig. 7.14 The occurrence of a whisker changing its shaft diameter. tin plate to account for the volume of the whisker. It is
Note the uniform longitudinal striations
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).
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 products although for space projects the ECSS and NASA
laminated with copper foil. Etching away the copper in requirements specically forbid this nish (Dunn 2012). It is
unmasked regions leaves behind the copper circuit which the special thermal cycling environment of the spacecraft
will be drilled, to facilitate plated-through holes which that makes this option extremely dangerous, as it can pro-
electrically interconnect both faces of the board. The mote the growth of tin whiskers from sites within pure tin
plated-through hole always contains a thick layer of copper plated-through holes. Perpendicular to the laminate, the glass
followed by a nal nish of tinlead plating that is eventu- weave does not restrict the resin expansion, and in this
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
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
Each temperature excursion produces a transverse stria- generate higher compressive stresses than, for instance
tion on the whiskers. copper-leaded interconnectors (17 106 K1) which, hav-
These striations are far more discernible when cycling is ing a similar COE to tin, do not generate thermal-cycling
in air. induced whisker growth (only Cu6Sn5-induced growth, as
The mechanism for growth is suppressed after about will be described in paragraph 7.4.3).
1500 cycles. Calculations have been made to evaluate whether the
Pre-tinning in a solder bath effectively prevents whisker levels of stress produced during the thermal cycling of
growth. ceramic (Al2O3) capacitors and resistors will exceed the
elastic limits of the outer tin-plated end-terminations (Moser
From these ndings it would appear that electronic units 2002). It was found that all the commonly used thermal
containing tin-plated parts will be more susceptible to cycling tests (see also Sect. 3.2.2.6) far exceed the elastic
whisker induced short circuits, or arcing, than their limit of tin. Per the ECSS Q-ST-70-08, the cooling step from
ground-based counterparts. 100 to 55 C was calculated to produce a theoretical stress
The Nemotos observed reduction in whisker growth rate of 160 MPa, whereas the actual elastic limit of pure tin is
at about 1500 thermal cycles is very similar to that seen by around 14 MPa. The potential in-orbit thermal cycles seen
Ditees et al. (2003) for growths measured on tin-plated alloy by operational spacecraft are often between 5 and 35 C
42 (i.e. FeNi42 CTE 4.3 106 K1; far lower than Sn CTE and even this typical cycle is expected to cause the elastic
23.0 106 K1). limit of tin to be slightly exceeded.
Dittes considered that under thermal cycling, tin whisker Some X-ray diffraction (XRD) residual stress measure-
length may have a parabolic relationship to the number of ments were instigated by the writer as part of an extensive
cycles. Indeed, the graphical display of results showing failure investigation involving the growth of tin whiskers on
whisker length against time presented in the Nemoto and passive component terminations. Interestingly, as-received
Dittes papers are almost identical. components that had been stored at room temperature for
The CTE of alumina ceramic chips is between 6 and 8 years and had not grown whiskers were found to possess
7 106 K1. These large CTE mismatches are expected to residual stresses in their tin-plating of about 7 MPa (i.e.
478 7 Whisker Growths
Fig. 7.22 SEM observations made on thermally cycled tin transverse striations encircling each whisker, they are more pronounced
plate-on-nickel-on-copper metallised ceramic chip capacitors. a shows in air c than under vacuum d. The distance between the striations
whiskers after 1040 cycles in air [max length 44 m, average 29 m], appear slightly greater when cycling is under vacuum. By kind
b after 1040 cycles under vacuum, [max length 67 m, average 43 m]. permission of Norio Nemoto
Thermal cycles can be followed by noting the distance between the
7.3 Tin Whisker Growths 479
side (in the region of the brazed joint) to have grown a high for the severely deformed tin seen in Fig. 7.30. The external
density eld of whiskers having lengths of up to 1.0 mm. work applied by the indenter is then equated to the increase
These observation will certainly highlight the long-term in surface energy as whiskers are grown. The indentation test
risk that tin-plated crimp terminations pose. Both is said to also replicate the micro-stresses induced by the
short-circuit and arcing potential failure modes must be volumetric growth (with time) of Cu6Sn4 intermetallic and
mitigated for, possibly by either selecting fused tin crimp its influence on whisker growth.
barrels, or by ensuring that whisker growths are contained
within protective shrink-sleeving.
One unknown is the effect that crimping will have on the 7.3.4 The Nucleation, Growth and Mechanism
external sides of this often-used crimped barrel. The of Growth of Tin Whiskers
crimping process is discussed in Sect. 6.4.3. Clearly large Results from a C-Ring Test
localized compressive forces are applied by the crimping Programme
tool in order to deform the barrel and compress/cold-weld
the enclosed bare strands of wire. Severe compression and The results of a comprehensive laboratory study of tin
deformation of the external tin-plated surfaces is akin to the whisker growth have related their ubiquity to particular types
forces applied to tin-plate when it is tested for of electro-plated tin and characteristics of the substrate
micro-hardness. Shibutani (2009) is one of several material (Dunn 1987). Forty C-ring samples, supporting
researchers who have attempted to identify whisker-prone different tin-platings and substrate variants, were testedthe
surfaces by implanting a hardness impression into the programmes 32nd year results are fully described by Ash-
tin-plating (see Fig. 7.24c). Shibutani introduces a worth and Dunn (2015). One C-ring sample and a sketch
creep-based model to account for whisker growth. The high showing its physical geometry is seen in Fig. 7.25. Com-
pressure exerted during the hardness test is said to cause tin pressive stresses were calculated and applied to the inner
atoms to diffuse away from the compressed volume and form surface of each the C-ring by torqueing the threaded links.
extruded whiskers; this mechanism has probably accounted The C-ring substrates were machined from brass and steel.
482 7 Whisker Growths
They were over-plated with 3 varieties of tin (normal; high (d) Plated tin, deliberately contaminated with occluded
current density; and, from a contaminated bath containing organic particles, on copper gave an unexpected result:
organic particles), half of the samples supported a barrier one sample only nucleated whiskers during the nal
layer of copper between the substrate and each tin plating. 12 years of storage, and these whiskers grew over the
One sample, supporting each material variant, was placed in following 20 years to lengths of 1.8 mm.
an inert gas oven so that the outer tin plated layer could be (e) Nodular eruptions are frequently observed to precede
fused at 250 C (pure tin melts at 232 C). those whiskers having a short incubation period. Most
The main ndings recorded at the 32-year test period whiskers have parallel sides, but they may contain
examination are: sharp kinks and can vary in diameter from around 6 m
down to only 6 nm. Whiskers are seen to possess
(a) Compressive stresses applied to tin plate did not numerous morphologies: from 1 to 20 m in diameter,
accelerate whisker growth rates. Whiskers were seen to some are straight but many have numerous directions
grow over all C-ring surfaces. of growth. Their density appear to vary from about 0.1
(b) Incubation periods before whisker growth are shown to to 200/mm2. Whiskers were noted to change their ori-
be strongly dependent on the substrate material. entation during growth, and by rotating became
Tin-plated brass has a short (several days) nucleation ensnarled with other growths (see Fig. 7.26).
period and produces whiskers with growth rates of (f) The results of the longest whiskers as a function of
8 m per day, whereas tin-plated steel takes several storage time are shown in Fig. 7.27. None of the fused
months to exhibit short whiskers. tin-platings nucleated or grew whiskers during the
(c) Samples with a copper barrier layer saw normal tin 32 years storage period (i.e. fused tin on: brass, steel,
plating to nucleate whiskers within 5 months and these copper-plated-brass and -steel).
grew to lengths between 1 and 4.5 mm.
7.3 Tin Whisker Growths 483
(g) Metallography revealed that, despite the relatively large A ve-stage model for whisker growth has been pro-
volume of tin material being ejected in the form of posed, based on the premise that local micro-stresses can be
nodules and long whiskers, there was no evidence of eliminated by the ejection of tin atoms from the plated layer
depressions or local subsidence in the plated layer. at sites of rotating screw dislocations. It is also postulated
Intermetallic growth after 32 years storage is shown, that a long-range atom-transport mechanism (diffusion)
for 4 samples, in Fig. 7.28. These phases have been sustains these growths (these stages of incubation and
identied, for all samples and are described in Ash- growth are shown schematically in Fig. 7.29).
worth and Dunn (2015).
484 7 Whisker Growths
Fig. 7.28 Back-scattered electron images showing the extent of identied as Cu6Sn4no other phases are present. c The fused 5 m Sn
intermetallic formation for normal Sn deposits on steel after storage at deposit has alloyed with steel to form FeSn2. d Fused 5 m Sn deposit
room temperature for 32 years: a 5 m Sn deposit, there is evidence of with 3 m Cu interlayer; here both Cu3Sn and Cu6Sn4 are present at the
a SnFe intermetallic at the plating-substrate interface. b 5 m Sn copper interface along with some voiding; islands of Cu6Sn4 within the
deposit with a 3 m Cu interlayer, here the intermetallic phase has been tin (Ashworth and Dunn 2015)
7.3 Tin Whisker Growths 485
Fig. 7.29 Schematic diagrams to illustrate tin whiskers growing on rotate and become displaced outwards as a whisker; intermetallic
tin-plated brass substrate (top left to bottom right): nodular eruptions growth, Cu6Sn5, causes additional compressive micro-stresses in the tin
possessing no preferred crystallographic texture are pressed from as it grows at the expense of Sn and Cu; as the intermetallic is less
individual grains of tin by localized residual stress from plating; dense than the metals it forms from, there will be volumetric increases
Frank-Read sources having suitable orientation generate edge disloca- and continuous stress-relief as more Sn is released into the whisker; Sn
tions; edge dislocation is rotated to a screw segment and Sn atoms grain growth (or recrystallization) at ambient temperature also creates
attach themselves to the edge dislocation causing screw segment to additional micro-stress as smaller grains are annihilated
7.3.5 Some Properties of Tin Whiskers Such broken whisker dust can bridge the spac-
ings between tracks of un-coated PCBs with the
The following sub-sections are mainly derived from Dunn risk of causing cross-talk or short-circuit failures.
(1987 and 1989) with additional data compiled from col- (ii) The theoretical natural frequency of tin whiskers
leagues and recent citations. depends on the individual whiskers length and
radius. For instance, one of the largest whiskers
(a) Attachment to plated tin seen in the C-ring experiment (Sect. 7.3.4) having a
(i) Tin whiskers can be physically dislodged by length of 4.0 mm and a diameter of 4 m can be
mechanical force. The sliding spring-loaded con- calculated to possess a natural frequency of 36 fn
nector nger has mated and de-mated several times (Hz). An average whisker may be one having a
to the PCB track causing whiskers to nucleate and length of 2.0 mm and a diameter of 2 m, so
grow in an elastically stressed region (up to 200 m having a fn of 71 Hz. A short whisker with a length
from the connectors edge). Additional mating of 200 m and a diameter of 4 m has a fn of
cycles have caused the plastically deformed mound 14,260 Hz. In fact the complete range of fn is
of tin to advance into the whisker-growth region expected to vary from 36 to 14,260 Hz.
causing them to detach and create whisker debris.
486 7 Whisker Growths
The vibration spectrum encountered by electronic pack- Samples of whiskers growing on electronic tin-plated
ages during, for instance, a spacecraft launch is severely items were exposed to the predominant frequency ranges
limited by deliberately isolating a unit or the damping effects and magnitudes expected to be encountered during elec-
of metal structure, bolts and packages, so that the very high tronic box ground handling, and during launch when much
whisker resonance frequencies will never be encountered. additional vibration is derived from acoustic noise (i.e. the
Only the range from 102000 Hz need be considered. sound eld developed by the extreme turbulence of the jet
7.3 Tin Whisker Growths 487
exhaust downstream from the launchers motor). The sam- actuation. New samples were subjected to 50 shocks
ples were bonded to a thick steel plate bolted to the shaker at a displayed level of 2060 g. Surprisingly, as for
head of a 7 ton Ling vibration tester. High magnication the vibrated samples, no whiskers were noted to
photographs and sketches recorded the number and location break of become deformed. This indicates that nei-
of whiskers at the start of testing. Low level testing was ther their yield of ultimate strength had been
initially performed, followed by inspections when no dam- reachedpresumably due to the very low inertia of
age to whiskers was recorded. The samples were then sub- each whisker.
jected to two sweeps in the 102000 Hz range of
frequencies, at a speed of 2 oct/min. These frequencies were The ndings of (Hashemzadeh 2005) also showed no
split into three groups of accelerations: damage to whiskers shocked for 6 ms and accelerations of
500 and 1000 m/s2.
1016.6 Hz 38 mm peek-to-peek
16.678.1 Hz 200 cm/s (b) Mechanical properties
78.12000 Hz 100 g
Further laboratory investigations have provided data related
No whiskers were seen to break or become permanently to the mechanical properties of tin whiskers. The whisker
deformed during sinusoidal sweeping through their natural samples had grown directly from the surfaces of a tin-plated
frequencies. mild steel electronic housing seen in Fig. 7.12. They were
Recent tests conrm these ndings (Hashemzadeh 2005): extracted from the housing surface using micro-mechanical
exposures culminating with the hardest vibration test for tools and placed onto glass slides. Individual whiskers were
electronics applied by the automotive industry (2000 h at the attached to slides using epoxy or low temperature melting
highest frequency 2000 Hz) failed to damage tin whiskers. indium solder. Cantilever beam loading and tensile loading
was performed using microne gold wire as weights.The
(iii) Long-term vibration at dened frequencies was made mechanical properties of tin whiskers were found to be
with the same samples exposed to ii). The samples were approximately:
tested on a Dunegan Model 4501 laboratory system at a 0
series of pre-selected frequencies: 50, 100, 200, and Young s modulus E 885 GPa
250 Hz, each at 6 g and for a duration of 60 s. Ultimate tensile strength UTS 0:8 kgf=mm2 8 MPa
Undoubtedly, whiskers having the natural frequencies
of this test series existed on the samples under test, but The observed scatter in some results is likely to be due to
again by meticulous microscopical inspection no evi- variations in each whiskers unit-cell parameter (different for
dence of detachment or damage was found. It was each of the ve crystallographic growth directions of tin
considered that the fatigue lives of the whiskers was whiskers) and to difculties in measuring the absolute
never exceeded (endurance limit stresses not reached, cross-sectional area of whiskers supporting flutes (with
number of fatigue cycles did not reach end of life, and rosette-shaped sections).
the whiskers are so homogenous and free from surface
discontinuities that no surface stress concentrations (c) Electrical properties
were available for fracture initiation.
(iv) Ultrasonic vibration in liquidthe same samples were Electrical failures attributable to tin whiskers include: heat
nally placed in a metal basket and lowered into a pacemakers, F-15 aircraft radar system, US missile elec-
commercial ultrasonic cleaning bath lled with iso- tronic assemblies, air-to-air missile systems, Patriot missile
propyl alcohol. The bath was operating at nominal systems, more than ve US commercial communications
frequency in the range 2025 kHz (above the frequency satellites, a nuclear reactor shut-down, rocket motor failure,
of human audibility). The cleaning cycle was main- and, commercial motor car failures due to tin whisker shorts
tained for 2 min. Visual inspection of the samples leading to mortalities and the recall of tens of thousands of
showed that all the whiskers on exposed flat surfaces vehicles (Sood et al. 2011; George and Pecht 2013). Sur-
had been removed. Whiskers in holes or re-entrant prisingly, these problem notications emanate from the
spaces appeared generally intact. US where there seems to be far greater transparency when
(v) The effect of mechanical shocksuch shock loading eld failures occur. Other, anecdotal accounts of whisker
may occur when electronics are exposed to rough failures abound, they are not documented in the open liter-
handling during transportation, accidental dropping ature due to commercial pressures. The number of failures
and at the moment of spacecraft pyrotechnic and various modes of whisker-induced failure necessitates
that all characteristic properties of whiskers are understood.
488 7 Whisker Growths
Mitigation against the risk of failure can then be attempted Whiskers can be attracted by electric elds and the
(see section related to creation of a lead-free control plan). degree of attraction has been found to be a function of the
The most comprehensive compilation of world-wide tin electric eld strength, the angle at which the whisker grows
whisker literature references, illustrations and reports, can be from a substrate and the dimensions of the whisker. This
found on the NASA website: http://nepp.nasa.gov/whisker/. electrostatic attraction will increase the likelihood of shorts
The electrical properties of tin whiskers were found (Dunn occurring between whiskers and conducting surfaces at
1988) to be dependent on their actual diameters. different electric potential (Hunt 2012).
Room-temperature resistivity can vary from 11 to 60 108 Tin whiskers can short-circuit under vacuum to create
. Actual current flow through 3 m diameter whiskers can tin-plasmas which have been seen to carry tens to hundreds
reach 32 mA. Hashemzadeh (2005) using similar test meth- of amperes and blow spacecraft fuses and destroy relays
ods found whiskers with a diameter of 5.5 m to also reach a (Mason and Eng 2007). Paschen curves showing the dis-
critical current of 32 mA before fusing. Current-carrying tance (between whisker and another electrical conductor)
capacities appear to be around 26,000 A/mm2 (Hunt 2012). multiplied by the pressure of the environment against volt-
The resistivity of tin whiskers were found to be lower than for age will predict the minimum voltage below which plasmas
pure -tin. Signicant damage can be caused to space sys- will be initiated. However, these investigators found by
tems when whiskers bridge electrical paths as illustrated by experimentation that sustained plasmas can form in vacuum
Fig. 7.31 and its caption. at a power supply DC voltage of only 4 V. Under ambient
High-voltage discharge can emanate from the sides and conditions the same tin-plasma can form at 28 V. (Paschen
tips of tin whiskers situated on electrically charged surfaces. curves were discussed in Sect. 6.15.4 with respect to corona
Spark discharges (i.e. corona) and nonluminous point dis- discharge between soldered joints on PCBs). Figure 7.33
charges from whiskers have been demonstrated in a scan- shows whiskers to be making short circuits between the
ning electron microscope using photographic methods bright-tin plated nickel-iron terminal pins of a relay. Relays
(Fig. 7.32). 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.
Fig. 7.32 The tip of this tin whisker supports a ne point from which shown toward the end of the 5 min display (d). No damage (corrosion
sparks discharged for more than 5 min. The emissions, all at 20 kV, or melting of the tip) was observed at the end of this event; a 10 s.
have the appearance of discrete white particles or a ne spray as b 30 s. c 40 s. d 200 s
related to the shape of the surface orice through which they between 20 and 30 nm, both from FIB sectioning and, as
have grown. Occasionally, growths have very irregular shown in Fig. 7.34, from XPS analysis. Newly grown
growth directions and odd changes in cross-sectional area. whiskers, only 2 weeks old, were found by the same mea-
Many appear to grow from nodules and some have suring techniques to support a dual oxide layer of just 2 nm
pointed-needle like shapes. As seen in Fig. 7.26, whisker (0.5 nm). It is postulated that long whiskers growing in a
growth is accompanied by a rotation of the shaft about a line short space of time will have a greater potential to cause
drawn normal from the nucleation point on the tin-plated short-circuits than those with a low growth rate to reach a
surface. This rotation, or gyratory movement can cause similar length. The thicker oxides on older whiskers will
whiskers with low projection angles to become ensnarled require greater breakdown voltages or more mechanical
with other whiskers and these may be have a greater force to create the conditions for shorting.
potential for causing electrical short-circuits.
Whiskers surfaces are now known to be covered with tin (e) Structure and growth directions
oxide (Hillman and Chumbley 2006). A dual oxide lm,
consisting of stannous and stannic oxides have been found White -tin has a tetragonal crystal structure with lattice
on both new and old whiskers (Dunn and Mozdzen 2014). constants of a = 5.831 and c = 3.182 at 26 C.
Whiskers removed from 48-year old electronic hardware Thinned whiskers subjected to transmission electron
could be bent so that their brittle oxide lms could be microscopy were found to be relatively transparent to the
measured in situ, these were found to have a thickness of electron beam of 200 keV; the ndings indicated the absence
490 7 Whisker Growths
of grain boundaries within the whiskers, no dislocations or For tin plated on brass: [001], [100] and [111]
stacking faults, no second phases (it should be noted that the
kinked regions were not explored). More than 100 whiskers were subjected to cryogenic
X-ray diffraction using the Rotating Crystal Method temperatures by immersion into both gaseous and liquid
showed growth directions to be: nitrogen (196 C), for times between 0.5 and 168 h. Tin
pest, or the transformation to grey -tin, was not observed
For tin-plated copper on steel: [101] and [100] (see also Sect. 6.15.12).
7.4 Precautions to Avoid General Whisker Growths 491
7.4 Precautions to Avoid General between a whisker and the housing. Redesign of the cath-
Whisker Growths ode heater introduced a complete back-potting that encased
the heater within small particles of alumina. This restricted
The case studies that have been described in detail in the the sublimation of tungsten by limiting the free surface area
preceding sections report on the effect that surface growths of the heater coil, and also enabled heat to be dissipated from
of whiskers have on modern telecommunications equipment. local hot spots by thermal conduction.
These investigations were made by standard laboratory
failure analysis techniques and have identied several 3. Human error
potential sources of whisker-induced failures that can
severely jeopardize the reliability of spacecraft systems. Human error can occur when an operator becomes careless
They can be related to one or more of the following: and causes a microcircuit to be electrically overloaded. One
result may be to cause the phenomenon of electromigration
1. Corrosion damage during storage and application and generate whiskers. Alternatively, the reason for
electromigration-related whisker growth can be wrong
Entrapped contamination can lead to corrosion and electro- design by the engineer responsible for the microcircuit. The
chemical migration within electronic devices. Electronic problem is highlighted by Figs. 7.57.9. Electromigration
circuits must remain humidity-free and should be hermeti- occurs because of the migration of metal atoms along a
cally sealed in metallic packages to avoid the condensation conductor under the influence of a large current. In most
of moisture and ingress of contaminants. electronic circuits electromigration manifests itself as mate-
The time-related short-circuit failures that are so clearly rial is transported along grain boundaries by a diffusion
shown in Fig. 7.2 were caused by the formation of dendritic mechanism. This occurs at high current densities (i.e. in
whiskers resulting from an electrochemical process. Such excess of 2 105 A/cm2) and at relatively low temperatures
failures require the presence of three elements: (much lower than the melting point of the conductor). As
atoms tend to migrate to positive ends of the conductor,
Reactive chemicals such as halogens (the trace amounts vacancies will travel in the opposite direction giving rise to
of chlorine detected in both good and failed integrated hillocks or whisker growths and voiding where the vacancies
circuit devices indicate that it is difcult to remove all accumulate.
traces of these chemicals from the microcircuit). Human error can also be considered to account for the
A layer of water to form an electrolyte between neigh- glass-rich oxide whiskers revealed within the insulation glass
boring stripes (formed by condensation when the tem- windows of integrated circuit flat-packs. Figure 7.4 clearly
perature drops below the dew point). shows that during manufacturing the assembly operator has
A bias for the electrochemical reaction to take place. misplaced the lead and wall frames of this device (this may
also be caused by a jigging problem or poor jig design). Had
Special precautions have to be taken to avoid this elec- the pre-oxidized parts been separated by a greater volume of
trochemical migration form of whisker growth. Cleanliness sealing glass it is certain that this volume would not have
testing has been proposed, using equipment that can detect become supersaturated with metal oxide; and, on cooling to
ionic residues. Alternative measures such as lling packages room temperature, whiskers would not have formed.
with dry nitrogen and adding a nal glassication layer (e.g.
sputtered silica) to the circuit are two alternative 4. Incorrect material selection
possibilities.
It may be arguable that each of the previous mechanisms of
2. Unsuitable aspects of design failure could have been prevented if the relevant design
engineer had chosen other materials for these applications.
The tungsten heater coil (Fig. 7.3) developed hot spots that However, the case histories that refer to the growth of tin
caused tungsten to sublimate. Released tungsten atoms whiskers (Sect. 7.3) are identied as being true cases of
passed through the vapour phase (in this case the evacuated incorrect selection of materials, and it is clearly seen that
travelling wave tube) and condensed at the whisker growth pure, as-plated tin will nucleate and grow whiskers in a
sites that were established by the presence of a temperature spontaneous manner regardless of environment. Figure 7.35
gradient and the existence of a suitable nucleation point. The shows a herring bone analysis of all the factors leading to tin
whisker growth process was accelerated by the inadequate whisker formation. It is a complicated subject as so many
design of the cathode heater piece-part and led to this device processing scenarios will promote such growths and not all
failure by spark emission and a possible short-circuit mitigation steps will be completely successful (e.g. the
492 7 Whisker Growths
encasement of tin-plated surfaces with conformal coatings circuits. Enormous quantities of whiskers have also been
may not be sufcient, as tin whiskers have been seen to seen to grow on household appliances. They are seldom
penetrate such coatings (McDowell 1993; Dunn 2012; reported in the literature, although accounts have been given
Wickham et al. 2014). A more detailed discussion of miti- of whiskers forming on the zinc-plated panels of refrigera-
gation will be covered in Sect. 7.5. tors and ovens (Garner et al. 1976), and one French paper
As a nal remark to the case studies one can take cog- (Bacquias 1982), which makes very interesting reading,
nizance of whisker growth situations that do not give rise to describes tin whiskers that were found growing from the
such calamitous effects as those afflicting complex electronic decorative plated nish on souvenirs stored in shop
7.4 Precautions to Avoid General Whisker Growths 493
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 7 Whisker Growths
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 The exemptions will be listed below, but it must be rec-
Plans ognized that lead-acid batteries account for more than 80 %
of lead consumption and these are exempt from EU regu-
7.5.1 General lations. Huge quantities of lead are used in the construction
industry with few regulations, whereas only 0.5 % of mar-
Humans have used lead since at least 6500 BC for the keted lead is incorporated into electronics. The concern is
ducting of water and for ornaments and jewelry. It is only in that the recycling of electronics by primitive means in third
recent years that the harmful effects of lead have been rec- world countries will harm persons due to the inhalation of
ognized and, as lead has been eliminated from paints and lead, but this has been disputed bearing in mind that lead
petroleum, there has been a dramatic reduction in boils at the very high temperature of 1750 C. Studies such
lead-related illnesses. It is considered by many engineers that as those by the University of Stuttgart, (http://leadfree.ipc.
the sweeping changes introduced by the European Parlia- org/les/RoHS_15.pdf) have indicated the replacement
ment and Council in 2006 will have a negative effect on the lead-free alloys for electronic use (i.e. tin-silver-copper) can
reliability of electrical systems being manufactured for both be more damaging, to both the environment and humans,
the commercial and professional markets. The EU directives than the lead-containing solder alloys. Also, as discussed in
concern WEEE (Waste Electrical and Electronic Equip- Sect. 3.6, there is much to gain from the recycling of elec-
ment) and RoHS (the Restriction of the use of certain tronic waste and the extraction of valuable metallic elements.
Hazardous Substances in electrical and electronic equip- Exemptions to the EU directives (RoHS) that ban the use
ment). These directives now specically forbid the tradi- of lead-containing alloys or glasses in electronics include
tional use of lead in the composition of component surfaces certain industries related directly to military, aerospace,
and solder alloys used to assemble electronic circuits. transportation and space.
7.5 The Creation of Lead-Free Control Plans 495
At the time of writing the following wording is cited: elevated humidity testing with high magnication inspec-
tions, before and after exposure, and the recording of
P7- TA-PROV(2010)0431 whisker growths in terms of length and density. These
A7-0196/2010 methods are not considered suitable for the acceptance of
space hardware for the following reason: only whiskers
Restriction of the use of certain hazardous substances in having a length greater than 40 m are considered poten-
electrical and electronic equipment tially harmful and shorter whiskers are accepted. The
mechanism for whisker growth is not yet understood and the
4. This Directive does not apply to: impact of very long storage and operational lives for
spacecraft parts cannot be determined from these short-term
(a) equipment which is necessary for the protection of the tests. As noted in Sect. 7.3.4, some substrates take many
essential interests of the security of Member States, years before whiskers nucleate and growin one instance
including arms, munitions and war material intended an incubation time of 12 years was necessary before whis-
for specically military purposes; kers emerged and grew to lengths exceeding 1 mm. It is
(b) equipment designed to be sent into space; considered expedient to totally avoid the use of pure tin and,
(c) equipment which is specically designed and to be as suggested by other authorities, alternative nishes such as
installed as part of another type of equipment that is nickel-palladium, nickel-gold, NiPdAu and even bare copper
excluded or does not fall within the scope of this can be suitable for most applications (George and Pecht
Directive, which can full its function only if it is part 2013; Morilla et al. 2015).
of that equipment, and which can be replaced only by
the same specically designed equipment;
(d) large-scale stationary industrial tools; 7.5.2 Methods for Reprocessing Pure Tin
(e) large-scale xed installations; Terminations
(f) means of transport for persons or goods, excluding
electric two-wheel vehicles which are not The most effective methods for the avoidance of tin whiskers
type-approved are those that completely replace pure tin nishes.
It is essential that the re-processing methods do not
A guidelines document suggesting the contents for a degrade the component or piece-part due to thermal shock,
company lead free control plan has been published (Dunn chemical attack caused by aggressive solder fluxes, or
2012) based on the known problems associated with lead-free chemical attack due to stripping and etching solutions. The
technology. Commercial off-the-shelf (COTS) components following reprocessing methods are suggested:
are discussed as well as methods for checking if pure tin is
present during goods-inwards inspection of spacecraft parts. It (a) Method for chemical stripping and replacing tin with
is proposed that swabs and sticks that contain chemicals which tin-lead by solder dipping
turn red when more than 10 % lead is present are a cheap initial
control. Other more accurate methods are included, but care This method must be performed by a specialist company and it
should be taken when using energy dispersive spectroscopy as is essential that in depth studies are performed in order to
false readings have been observed: when only large tin-rich or ensure that no damage has occurred at glass-to-metal seals and
lead-rich grains are analysed rather than an area containing that there is no ingress of chemical liquids into the component,
mixed grains, or when thin pure tin layers are penetrated and either by cracking, dissolution of seal material or by ingress
readings include elements of the substrate. through plastic materials that may be incompatible with the
Two standards have been established with the objective chemicals. Care must be taken to ensure traceability of the
of predicting whether newly tin-plated electronic parts will devices as chemicals can remove marking inks. This method
grow whiskersthey contain quality control type require- has been demonstrated to be suitable for a wide range of
ments and necessitate that samples are subjected to accel- components, including those with glass-to-metal seals as pure
erated environmental testing. The standards JESD22-A121A tin plating can be removed right up to the seal as seen in
and IEC 60068-2-82 include temperature cycling and 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 497
The leads can then be tinned by dipping using standard tin-lead. When possible this is the most effective and
solder-pot or -wave methods. The cleaning of any adhering cost-efcient method of re-processing. It may be problematic
aggressive solder flux is an important part of this process. when solder wetting is difcult and time/temperature of
Components that have been successfully re-processed by immersion is critical to avoid the presence of residual whis-
this method have been listed. Similarly, a list exists showing kers as shown in Fig. 7.38. It is likely that this method will not
components that were damaged by this method. Barrier be permitted on components having a glass-to-metal seal
layers or substrate lead and termination metals may be because these seals might be damaged by thermal shocks if the
exposed adjacent to seals, but these new surfaces are component lead is submerged completely up to the seal.
acceptable for space programme use. However, a form of verication testing has approved this
method when it was shown, as for Method a), that no com-
(b) Method for solder dipping and the reapplication of ponent damaged ensued. This process must be carefully
solder onto tin-nished surfaces controlled and processing parameters need to be documented,
these will include: solder pot temperature, composition of the
Solder dipping of tin-plated surfaces, when performed cor- solder pot alloy, pre-heating rates, immersion depth, cooling
rectly with tin-lead alloys, will eliminate the formation of rates, electrostatic discharge, nal cleaning to remove residual
whiskers as all of the pure tin will have been replaced by solder flux. Robotic solder dipping is performed at specialised
Fig. 7.38 Components were supplied with leads having a pure tin effect of the package, and the presence of a thin (20 nm) tenacious
nish; during storage they developed whiskers; these gures show the oxide on the pure tin whiskers, there has been insufcient heat to
leads after tin-lead reprocessing (pre-tinned by dipping to completely dissolve the tin whiskers into the solder; oxide thicknesses
ECSS-Q-ST-70-08C (2009)); micro-section and EDX analyses con- on whiskers have been measured by Mozdzen (2012); some theories
rmed the presence of pure tin (whisker) coated with tin-lead; pure tin propose that if any pure tin plating exists below the dipped solder layer,
melts at 232 C and eutectic SnPb at 183 C; due to the heat-sinking this will reactivate whisker growth
498 7 Whisker Growths
companies for leaded and leadless component types. These SnPb immersion: Chemical stripping followed by partial
same companies (Walmsley 2015 and Tyler 2015) are also SnPb dip; SnPb dip with standoff (i.e. gap between seal and
capable of reprocessing a very wide range of area grid arrays, solder meniscus)
replacing lead free interconnections with both tin-lead balls
and columns. Tables of suitable and unsuitable component (b) At the PCB assembly level
types have been compiled by the companies who perform the
reprocessing. Physical Barrier (e.g. Mylar foil, space-approved solder
sleeve)
(c) Over-plating pure Tin nishes with pure Lead Pottings: Hard potting; soft potting or encapsulation
Soldering variations: Soldered with SnPb solder with
This method has been used for small chip components. After qualication /verication of coverage (with or without sam-
plating lead onto the pure tin surfaces the device is heated to pling); soldered with SnPb without verication on bottom
promote complete inter-diffusion. The author is not aware termination or small leadless chip parts; SnPb with 13 % Pb;
that this process has been applied by European companies. 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
7.5.3 Mitigation Approaches (atomic layer deposition) and vapour deposition (such as
Parylene, polyparaxylylene)
(a) Component terminations
Tin variations: hot dipped tin; large grain size or actually a Note: The use of conformal coating does not eliminate the
matte tin; thick tin >15 microns; immersion tin potential for whisker failure entirely. For most coatings, tin
Underplating with: nickel or silver whiskers have been shown to eventually grow through thin
Although all tin nishes have some risk of whiskers, coating regions and potentially short any other whiskers that
some nishes have greater risk than others. Selection of a have grown similarly (see Fig. 7.39). In addition, for some
matte tin with a nickel underplate (>2 m) or a Pb-free tin piece part types (e.g. through-hole mounting leaded piece
alloy with nickel underplate (>2 m) are considered lower parts, area array packages), it is difcult to entirely encap-
risk nishes than bright tin or tin nishes over copper. Other sulate the individual interconnects so there remains a direct
nishes include SAC nish or dip, SnAg nish; SnBi nish path for shorting due to whisker growth.
Heat treatment: above fusing temperature (best); below
fusing temperature
Fig. 7.39 (Left) A growing tin whisker is attempting to break through conformal coating. This can result in an electrical short circuit.
a silicone-type of conformal coating; and (right) a tin whisker that is (Electron micrographs courtesy of C. Hunt, the National Physics
growing from an adjacent surface, is attempting to penetrate a 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 reproducible results even when material variants are kept
risk under control. However, it is important that engineers
working in the space, and other sectors requiring high reli-
Many studies have been performed and they provide limited ability products, are able to make an informed judgement,
quantication of when and how failure conditions might for instance during the course of a nonconformance review.
occur as a result of whisker growth. The basic mechanism Table 7.2 can be consulted as a guide to specic risks. An
for whisker nucleation and growth is to date not fully attempt has been made to put a numerical value to each risk
understood. Experiments do not always provide for 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 A comparison between the sum of the positive values
and quantify the effectiveness of mitigation methods for assigned to the cumulative risks and the sum of the
either preventing whisker growth or limiting the risk of negative values assigned to any mitigation methods used
whiskers causing failure; here Table 7.3 can be consulted, (or mitigating design features) is intended to indicate the
and again methods and associated values can be cumulative. acceptability of a given situation.
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.
Assessment of Post-flight Materials
8
8.1 General is just over 450 metric tons (tonnes). Awe-inspiring, as ISS
was built by astronauts and cosmonauts, 216 have lived
8.1.1 Hardware Return from Space there, bringing each part of the construction from Earth by
Shuttles and various other cargo-vehicles (2015 data).
The low Earth orbit (LEO) environment is dened as that The space missions referred to above have all involved
region of space between 200 and 1000 km (124 and the launch, exposure to space (in particular the LEO envi-
621 miles) above the Earth. It was reasonably well charac- ronment), and the nal return of some form of space hard-
terized in the 1950s, during the prelude to the race into ware to Earth. Once the mission tasks had been completed
space, by means of sounding rockets equipped with moni- there was little reason to spend extra effort into performing
toring devices for recording air pressure, temperature, and post-space flight analyses of the materials used to construct
gaseous composition. Many of the data obtained from those the spacecraft. Some post-flight observations have been
early space flights are suitably accurate for todays calcula- documented, such as those related to impacts by microme-
tions and have been discussed in Chap. 2 and compiled into teorites during the mission, the degradation of thermal
Table 2.3. In just one decade, the 1960s, human beings control surfaces due to discoloration from exposure to
began to extend their physical boundaries by initially ven- ultraviolet radiation and, of most importance, the effectivity
turing into LEO and then progressing to explore the of the re-entry protective shield. However, the external
mountains and valleys of the Moon. The rst human to orbit surfaces of those spacecraft were constructed from materials
the Earth was cosmonaut Yuri Gagarin in the Vostok 1 that were little affected by the space environment, such as
spacecraft on the 12th April 1961. Since then, manned anodized aluminium alloys, titanium alloys, passivated
spaceflight included the Mercury (Atlas ELV) missions, the stainless steels, and fused silica glass. The re-entry shields
Gemini (Titan ELV) missions, and the Apollo (Saturn ELV) were either of the ablative type that burnt and shed material
missions that gave 12 human beings the opportunity to walk during their hypersonic velocity passage into the oxidizing
on an extraterrestrial surface, with the rst manned landed on atmosphere or, more recently, of the lightweight
the Moon in 1969 with Apollo 11, and the nal silicon-based ceramic tile-types which have a good resis-
Moon-landing in 1972 with Apollo 17. The 1970s saw the tance to high-temperature oxidation, air turbulence, and
launch of several Salyut space stations and three Skylabs, shock during descent.
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 8.1.2 Raw Materials from the Moon
space, retrieval of spacecraft, and utilization of Spacelab
which remained attached to the Shuttles cargo bay). At the The number of documented post-flight examinations of
time of writing, it is the International Space Station spacecraft materials are probably now equal to the several
(ISS) that continues to be the most newsworthy space hundred papers in which the results of Moon-rock analyses
activity. This is the biggest construction to fly in space and were published since 1967. Although the scientic yields
can be readily seen from the ground as it completes 16 orbits from the observations of the astronauts on early Apollo
per day. Huge solar panels, totaling approximately one acre missions were of great importance, they were later over-
are attached to the central truss of ISSthe dimensions are shadowed by the investigations of land-based scientists who
equivalent to the length of a football eld, the habitable studied the lunar samples. These studies included a deter-
volume is as large as a six-bedroom house and the total mass mination of the abundances of all the elements, their
isotopes, and the nature of their coexistence in the various comparison, the rocks retrieved from the lunar highlands by
phases making up the lunar material. The lunar samples later Apollo spacecraft were relatively enriched in oxides of
brought back by the US (manned Apollo) and Soviet (Luna aluminium and calcium (with low Mg and Fe contents).
automated soil collection and return to Earth) space pro- The use of the Moon and lunar materials has often been
grammes revealed that the Moons surface is rich in non- suggested by scientists as the stepping-stone into an indus-
volatile, low atomic mass elements. Lunar rock and soil trialization of space. Certainly there exists a wealth of
samples (termed regolith) were distributed to laboratories minerals, a strategic location, and a reduced gravitational
worldwide for analysis. One particularly interesting rock was force. It should be possible to manufacture a range of
prepared by standard metallography and is shown in structural materials on the lunar surface; oxygen in the form
Fig. 8.1. This micrograph is typical of the Apollo 11 regolith of oxide is present, but three of the most vital elements were
in that it is basaltic and contains a large amount of ilmenite. thought missing: carbon, nitrogen, and hydrogen, and
The rock also contains phases of metallic iron and a small without these there can be no indigenous life-support sys-
amount of metallic copper. Iron-nickel alloy, native iron and tems. In recent years scientists have suggested that water ice
native copper have also been identied during quantitative and other frozen gasses (or volatiles) may be trapped on the
optical and electron-probe studies of lunar samples at the perpetually shaded polar regions of the Moon. If long-term
Institute of Geological Sciences, London (Simpson and support is to be given to human activities on the Moon or
Bowie 1970). The location of metallic iron has been stated to Mars it seems essential that a base will have to be built on
bear a relationship to the lunar magnetic properties and many the Moons surface. Both NASA (Mueller 2014, 2015) and
possible sources of this metal have been postulated (Adler ESA (Ceccanti 2010) have considered the requirements for a
1988). It would appear that the existence of native copper, lunar base to house people and offer protection from mete-
present as a small 4 m diameter sphere in Fig. 8.1, is orites, solar- and gamma-radiation and high temperature
uncommon because chemical analyses of core samples, fluctuationsseveral attempts on Earth to construct dwell-
collected at any of the other eight Apollo sampling sites, ings from lunar regolith simulant have been successful (see
showed that this metal is generally only present as a trace also Sect. 8.2.2). The ultimate success of a permanent lunar
element in the order of 10 ppm. Oxygen is particularly base is expected to depend on how man uses the in situ
abundant in the form of oxides of silicon, titanium, iron, and resources. If it is possible to manufacture propellants on the
magnesium at the lunar mare site visited by Apollo 11. By Moon it will relieve astronauts of the need to transport
8.1 General 503
return propellants. The lunar surface appears to contain 45 The purpose of the following sections to this chapter are
weight percent oxygen, most in the form of silicates and to present some of the ndings from recent experiments and
other mixed metal oxides. With the use of solar, nuclear observations of returned hardware. As well as recording
power and focused mirrors, it has been proposed (Hepp some predictable modes of degradation, data has also iden-
1999) that regolith can be heated and modied to release tied unusual phenomena likely to either enhance or reduce
oxygen and precipitate metals and alloys of aluminium, the durability of spacecraft mechanisms, structures, and their
silicon, iron and titanium. These metallic powders would materials. Some end-of-life properties have been established
then be burnt in a combustion chamber and the simultaneous and these are of great importance to design engineers. Only
release of hot gasses gas through a nozzle would provide a those references considered to be of major importance will
form of propulsion! 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
8.1.3 Recent Investigations Using Retrieved 1993, Anon 1994, Guyenne 1994, Gerlach 1995). More
Materials recently, some data has been published concerning the
MEDET results (Tighe 2010). The following sections will
The IIS was gradually constructed over a period of 12 years. also describe property changes to materials brought about by
This started in 1998 when the rst module, Zarya, built by their exposure to LEO environments. The main effects are
Russia, was launched in 1998. This permanent space station due to each materials exposure to vacuum, radiation, tem-
in LEO has been continuously occupied since 2000. The perature cycling, micrometeoroids and space debris, atomic
present architecture is seen in Fig. 3.1b. Plans are to prepare oxygen, the re-entry environment, and operation within a
the station for commercial use and two international docking manned volume. The space environment effects are illus-
adaptors will be added to the structure and utilize the station trated by the block diagrams in each section. There can be a
until at least 2020. Alongside the operation of ISS has been a strong synergistic influence between these effects as inferred
continual series of materials tests, such as the MISSE by the diagrams (Dauphin et al. 1991). This information will
experiments, where samples of a wide range of be of interest to those engineers responsible for scrutinizing
material-types and components are exposed to the harsh the materials selection process for future missions, some of
environment of space for long periods of time. The data that which are destined for relatively low orbiting altitudes
has been generated from tests performed on materials that (*350 km) in high densities of atomic oxygen.
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 8.2 Space Environmental Effects
many space symposia. Some data have been generated from from Vacuum and Radiation
ground tests in which attempts are made to replicate the LEO
environment in laboratory chambers. The majority of the 8.2.1 Organic Materials and Lubricants
materials exposure data has originated from space missions
such as Salyut-6 (5-year exposures), Salyut-7 (9-year Screening tests are continually performed during the selec-
exposures), Mir station (experiments of 3-year duration), tion of organic materials for space applications. Approxi-
the Long Duration Exposure Facility (LDEF) launched and mately 100,000 test results are available from NASA and
retrieved by Space Shuttles (STS flights) after nearly six ESA sources based on the test method specied in ASTM
years in LEO, and the European Retrievable Carrier (EUR- E-595. The provision is that materials, when tested in vac-
ECA), similarly launched and collected after nearly one year uum at 125 C for 24 h, shall not have a recovered mass loss
in orbit. Short-term materials exposures were made on (RML) greater than 1 %, or a collected volatile condensed
STS-5, STS-8, and STS-46 (for instance, the 58 h Evaluation material (CVCM) content of more than 0.1 %. This, and
of Oxygen Interactions with Materials III (EOIM-III) flight other test methods for organic materials, have been described
experiments). Retrieval of the Solar Max spacecraft provided in Sect. 2.5. Some examples were given for lubricants in
some data, as did the post-flight investigations performed on Table 5.2. If these requirements are neglected there is a
the Hubble Space Telescope solar array that was retrieved danger that spacecraft will become contaminated by the
after about four years in LEO (during the rst HST Servicing outgassing species (the effect of outgassing/condensation
Mission in December 1993; the second service mission took can be seen in Fig. 5.12). Vacuum exists everywhere in
place over a 10-day period in 1997). The most recent test space except in pressurized modules. Vacuum may change
exposures include the European MEDET rig and the previ- from 1014 torr in free space to 104 or 103 torr in poorly
ously stated MISSE project; all are currently being examined vented spacecraft subsystems. The chain of dangers that
in a 1500 sample test programme. result from outgassing products is depicted in Fig. 8.2; gas
504 8 Assessment of Post-flight Materials
clouds can produce corona and arcing, or may condense onto removing molecular contamination which would otherwise
surfaces and so modify them to produce thermal and elec- have migrated to contaminate sensitive spacecraft surfaces
trical problems. when in orbit. Cleaned hardware is best maintained under
A major nding from the recent post-flight materials vacuum or, at least, sealed in appropriate bagging material,
experiments concerns the use of silicone products that were preferably under an ultrapure gas such as nitrogen. A com-
previously considered to have been suitably screened by the parison of identically produced samples that were either
outgassing test. Another, concerns the 24 h vacuum bake at pre-cleaned by the standard practice, or exposed to the
125 C that has become a standard practice in history for vacuum of space has been performed. The results found that
cleaning polished mirrors and other hardware, so benignly the ground-based material conditioning process 125 C
8.2 Space Environmental Effects from Vacuum and Radiation 505
vacuum bake for 24 h is equivalent to holding an article in a spacecraft with the exception of its application as an insu-
space environment at an ambient temperature of 20 C and a lation material for wires and cables.
pressure of 106 torr for a period of one year. There are two Further inspections of the returned HST solar array
provisosthat this practice does not hold true for either revealed several surfaces to have been contaminated by the
silicone products or Teflon FEP materials. Recent tests with condensation of outgassing products. Figure 8.3 is included,
approved silicone products caused outgassing and contami- as it highlights a condensation layer having a variable thick-
nation in LEO. Even very thin layers of contaminant (about ness and is clearer to photograph than surfaces that have been
400 ), together with the presence of ultraviolet (UV) radi- uniformly discoloured. Silicone contamination has been
ation and atomic oxygen, caused signicant degradation of identied on the HST samples, by analysis of collected
thermal control surfaces (Babel et al. 1994). In a similar spectra, to be the main contaminant along with dicyandiamide,
manner, Teflon FEP can degrade with the formation of a which is an outgassing produce from CFRP materials. All
volatile product that is mainly perfluorocarbon, but can items manufactured from Delrin material, namely the primary
include chemically active species containing fluorine, and deployment cable and the boom end cotton reels, changed
these have been seen to chemically etch the surfaces of from a white colour to yellow-brown on their sun-pointing
polished metals and glass (van Eesbeek et al. 1994). Teflon sides. All surfaces covered with the atomic oxygen protective
FEP originating from samples mounted on LDEF for coatings (CV1144-1, DC93500, and RTV S691) became
5.8 years, was compared to samples of the same materials darkened in colour from straw to dark brownthe more
exposed to LEO on the Hubble Space Telescope (HST) for exposure to the Suns radiation the darker the colour (see
both 3.6 and 8.25 years (periods for the service missions). Fig. 8.4). They are expected to cause an increase in tempera-
All AO (atomic oxygen) exposed samples became ture at these locations.
mechanically embrittled and had lower strength properties None of the traditional lubricants based on
when compared to reference ground-stored sheet (Fink vacuum-distilled silicone oils and greases were seen to
2006). The HST solar array bellows and the multilayer degrade during exposure to service life in space. Some were
insulation (MLI) items were severely damaged and cracked, noted to have crept from bearing races onto adjoining
being worse for the longer exposure. Samples taken from surfaces. Rotating surfaces that had been lubricated with thin
LDEF material exposed to the RAM direction (AO: lead coatings survived their designed lives.
8.17 1021 atoms/cm2) became more damaged than samples Concern was raised when three critical pump motors
in the WAKE direction (AO: 9.32 104 atoms/cm2). In failed in close succession following ve months of suc-
general Teflon FEP is not used on the external surfaces of cessful operation in LEO. On returning to Earth, the motors
506 8 Assessment of Post-flight Materials
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, space. However, once the water content of the brushes had
and armatures. Large quantities of brush debris were noted diminished sufciently by outgassing under vacuum, the
associated with the full wear-out of the graphite brush friction and wear rates would be markedly increased, so
lengths. The small stubs of brush material that remained producing these early pump motor failures. Some alternative
were analysed and found to contain only carbon (graphite) brush materials have been described in Sect. 4.8.2, but it is
and copper with no additional elements. The absence of any essential to perform any qualication tests in vacuum rather
so-called adjuvant materials within the microstructure of than dry nitrogen.
these flight brushes were extremely important. It is known More recently with respect to tribology, tungsten carbide
that adjuvants such as molybdenum disulphide (MoS2) or alloyed with MoSx thin lms (deposited by magnetron
barium fluoride are essential components for brushes need- sputtering) has been prepared and tested under vacuum and
ing to operate under vacuum. Investigations found that these LEO. Ball-on-disc samples were evaluated on the ESA Tri-
motors had been qualication tested prior to launch in lab- bolab (installed in the European Technology Exposure
oratory air. This air contained moisture, an essential ingre- Facility xed to the Columbus module of ISS). These tests
dient for graphite brushes that enables them to run for (Brizuela 2009) have indicated WC + MoSx to be a novel,
extensive periods with very little friction and wear. The highly suitable material for operation in LEO and comparable
flight brushes had absorbed sufcient amounts of moisture to conventional MoS2 lms. Very low extraterrestrial friction
so that they behaved correctly when operated for a short levels of 0.04 were reached during more than 1 million wear
period in dry nitrogen, and correctly when launched into cycles on the ball-on-disc set-up and loading at 0.75 GPa.
8.2 Space Environmental Effects from Vacuum and Radiation 507
Table 8.1 Components of the natural space radiation environment whilst black paints tended to bleach. The white paints with an
Solar-particle events inorganic base and a silicate binder were found to be more
Occur sporadically; not predictable stable than any of the organic-based paints. Spacecraft
Energetic protons and alpha particles
adhesives, thermosetting resins, and lubricants are unaffected
by the low dosage of radiation from space, and these are
Solar-flare events may last for four or more days
usually well protected by metallic foils or are present within
Dose very dependent on orbital altitude and amount of spacecraft
electronic black boxes. Some inorganic glasses are quickly
shielding material
damaged by radiation, and even as little as 103 rad can cause a
Presently under study by the SOHO spacecraft
loss of transparency to some wavelength ranges (Dauphin
Galactic cosmic rays 1993). Plastic glasses and optically transparent silicone
Typically 85 % protons, 13 % alpha particles, 2 % heavier nuclei coatings are also damaged by a general yellowing during
Integrated yearly fluence: exposure to solar radiation as seen in Fig. 8.4.
1 108 protons/cm2 (approximately) Microelectronic semiconductors are also damaged by
Integrated yearly radiation dose: solar particle radiation. These are usually split into two
410 rads (approximately) categories: cumulative effects due to prolonged exposure to
radiation, and single-event effects caused by transient high
Geomagnetically trapped radiation
levels of cosmic or high energy proton exposure that can
Primary electrons and protons
lead to worst-case damage to spacecraft electronics.
Radiation dose depends on orbital altitude For instance, the thin gate oxide in metal-oxide-
Manned flights below 300 km altitude avoid Van Allen belts semiconductor devices (e.g. MOSFETs) is degraded to
For solar particle radiation see Table 2.4 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
8.2.2 Radiation Effects becomes greater. The ionizing radiation creates electron/hole
pairs and charging of the oxide to produce leakage current
The term radiation (Table 8.1) covers a wide range of perturbations. Irradiation will also cause deep-level defects
emissions of energy from the Sun in the form of electro- in the bulk silicon lattice. This can cause gain degradation of
magnetic waves (the range passes from X-rays to UV and an npn or pnp transistor, or increased leakage current in a
visible light, to radiowaves) and particles which are trapped diode. Experience has shown that LEO spacecraft have not
in the Van Allen belts (protons and electrons) that surround often suffered from electronic component damage due to
parts of the Earth. A second region that traps highly charged radiation. However, damage does occur when space vehicles
particles band is known as the South Atlantic Anomaly, its travel through the heart of radiation belts as would be the
location is seen on Fig. 8.5, and this presents a signicant case for spacecraft in low Earth polar orbit, or prolonged
hazard to spacecraft in low altitude, low inclination orbits periods in geostationary orbit. Here, radiation hardening is
(Mikaelian 2001). Other particles originate from deep space, required by means of altering the processing steps during
such as cosmic rays, and reach the Earth from all directions, device manufacture or providing countermeasures such as
with high energy and great penetrative power. Radiation at physical shields or foils around the packaged device.
the levels existing in space (generally about 107 rads per year) Detailed handbooks have been compiled by Homes-Siedle
do not modify the bulk properties of metals. Even the largest and Adams (1993) and Mikaelian (2001) to cover the effects
doses of radiation that occur in the Van Allen belt have no of radiation environments on spacecraft electronic equip-
effects on metals. It is only the combination of contamination ment, their components, and materials. It is noted that in
layers together with radiation (see Fig. 8.2b) that can severely recent years researchers have been searching for novels ways
change the surface properties of metals and this may lead to to avoid radiation damage, either integrating functions into a
an increase in solar absorptance and a reduced hemispherical single chip or ASIC for better performance, or by enhanced
emittance. Materials such as coatings and paints are pro- radiation shielding with tantalum and tungsten spot shields.
gressively damaged by solar radiation becoming embrittled These metallic elements are about six times the density of
by both particle and UV radiation (Dauphin 1993). White aluminium and they allow thinner shields to be made, thus
paints were modied to have more absorptive properties only incurring a small weight penalty (Maurer 2008).
508 8 Assessment of Post-flight Materials
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 on-ground studies of the problem it was concluded that the
failure was due to overheating of one of the SADEs
As already alluded to in the previous sections, there have been assembled printed circuit boards (PCBs). A replacement
no instances of degradation to metallic materials due to the model was built in which two power transistor devices were
combined effects of vacuum and radiation even when returned physically removed from the circuit and mounted via
to Earth from the longest exposures to LEO. This is because heat-sink caps onto the internal metallic structure of the
the selection process avoided the use of metals known to SADE (Haines et al. 1995). This modication resulted in a
produce problems such as sublimation (e.g. for cadmium and device junction-to-structure thermal resistance of 30 C per
zinc), whisker growths (tin and zinc), or loss of hermeticity for watt and ensured that even under the worst case conditions,
pressurized items, such as battery cells, hydrazine fuel tanks, the maximum transistor junction temperature would be less
and the like. The problem with silver was related to atomic than 85 C.
oxygen, which will be discussed in Sect. 8.5. The SADE exchange activities on the rst HST servicing
One problem that did come to light, which resulted from a mission were performed at night and under poor lighting
poor electrical design, was the major failure of one of the conditions by astronauts Story Musgrave and Jeff Hoffman.
HST solar array drive electronic (SADE) systems after three Light from Musgraves helmet light was used to illuminate
years of satisfactory operation. At the moment the solar array the SADE system and during one of the most difcult EVA
position data was lost, the drive electronics was automati- (extra-vehicular activity) tasks of the mission, the defective
cally switched off. A redundant unit was powered up soon SADE was removed and replaced (Haines et al. 1995).
after this event for the subsequent eight-month period up to Electrical tests and visual inspections were performed on
the date of the HST service mission. As a consequence from the defective PCB after it had been returned to Earth.
8.2 Space Environmental Effects from Vacuum and Radiation 509
Following removal of the equipments cover, a few minute Little space will be devoted to temperature effects and
spheres of solder alloy were seen attached to the inside thermal cycling in this section, as much has been written in
surface of the lid. Subsequent inspections revealed much this book to cover the topics of sublimation, thermal fatigue,
discoloration of the Solithane conformal coating in the and the importance of matching adjoining materials so that
vicinity of the two previously suspected power transistors. they will have similar coefcients of expansion. The very
Electrical tests were made. The electronic unit had a below gradual thermal fatigue degradation of the HST solar array
nominal power consumption, solar array position data were cell-to-interconnector weldments (only made visible by
not available, and no smoke was seen to emanate from the SLAM investigations) were described in Sect. 4.14. Distor-
suspected fault area. Further disassembly was performed and tion of the HST solar array due to differential heating and
photographs made of the charred and overheated power expansion rates were described in Sect. 5.17, but apart from
transistors (see Fig. 8.6a). Inspection of the diodes adjacent this major defect, the remaining materials compatibility on
to the charred site showed that they had also overheated as the array seem to be good and the complete spacecraft will
their solder joints had remelted and both the conformal be flightworthy until its successor, the James Webb Tele-
coating and the epoxy top-coat of the circuit board had been scope will be launched and become operational in 2017.
completely evaporated away (Figs. 8.6, 8.7 and 8.8) to One of the most severely thermal-cycled solder joints to
expose a region of glass-bre weave (Adams 1994). have been returned from space is shown in Figs. 8.10. This
The cause of the SADE failure is thought to have resulted joint was one of several that existed on the HST arrays
from the inadequate thermal management of the defective so-called tag board. All survived 21,000 thermal cycles. The
PCB assembly. Testing on the ground, in air, had not caused solder llet has seen some stress due to the large thermal
any pre-flight failures because the area surrounding the two expansion coefcient of the FR4-type PCB in its z-direction
power transistors could be kept below a critical temperature (see also Chap. 6). Thermal fatigue is induced by 2.1 m
by means of heat lost by thermal conduction through the board mismatch between the boards glass-bre weave/epoxy
and heat lost by thermal convection through the laboratory air. matrix and the brass terminal pin. This has exceeded the
Under spacecraft conditions, the operating environment of yield strength of the solder alloy during thermal cycling,
vacuum excluded convection as a means of cooling and, after causing the observed orange peel effect on the surface of
a period of high-temperature operation, the transistor solder these solder llets (Fig. 8.10a), but when microsectioned, it
joints melted (above 183 C), so producing the nal failure. is clear that the integrity of the interconnections has
By modifying the circuit such that both power transistors were remained intact (Fig. 8.10b, c).
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. 8.4 Micrometeoroids and Debris
8.4.1 General
8.3 Temperature Cycling
Micrometeoroids and cosmic dust are natural, hypervelocity
The effects of temperature on spacecraft materials are particles that originate from comets, collisions between
depicted in Fig. 8.9a. The main input is from the Sun, asteroids, and interstellar material. All sizes of these particles
sometimes reflected by a planet, and on output (cooling) by are of importance to spacecraft, but those most likely to
emission to the immense sink of deep black space. Tem- cause damage, owing to their statistical distribution, range in
perature excursions are also generated by the operational diameter from 1 m to 5 mm. Before the space age began,
mode of the spacecraft, from high-power electrical systems, this was the only particle population surrounding the Earth
and by the release of energy (exothermic reactions) during and passing through her atmosphere. Since the 1950s, debris
the consumption of liquid fuels. Local thermal cycling is from mans activities in space has accumulated across the
caused by the expenditure of electrical power as devices skies. Space debris includes satellites and rocket upper
such as travelling wave tubes are operated and then switched stages that have completed their mission, together with
off (power cycling). The main overall thermal cycling of a thousands of smaller particles generated by explosions in
spacecraft is experienced as it repeatedly passes from sun- space. Since the launch of Sputnik in 1957, more than 4600
light into eclipse during the course of an orbit. The solar launches have placed some 6000 satellites into orbit. Nearly
array of the HST is subjected to 6000 thermal cycles per year 2200 remain in orbit, of which 450 are still functional.
between the temperatures 113 and +95 C. This means 80 % of space objects are uncontrolled debris.
510 8 Assessment of Post-flight Materials
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
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 8 Assessment of Post-flight Materials
8.4.2 Debris Emanating from Catalytic Bed reaction is strongly exothermic and the hot reaction product
Thruster Motors gasses expand and are forced through a nozzle to produce
thrust, usually in the form of specic, high energy impulses
Almost all unmanned spacecraft continue to incorporate of greater than 10 N. The thruster design, an engineering
monopropellant hydrazine thrusters for attitude control and drawing, and the appearance of catalyst bed particles are seen
trajectory corrections. Certain older thruster designs use a in Figs. 5.445.48. A screen is incorporated into the thruster
catalyst bed consisting of iridium-coated grains of alumina adjacent to the nozzle in order to reduce the number of par-
having a diameter of between 200 and 1200 m. Small grains ticles exiting through the nozzle. However, small particles are
provide a high surface area of the iridium catalyst for known to be expelled through this screen as the mesh size
decomposing anhydrous hydrazine (N2H4) into innocuous cannot be so small as to limit the flow of gases and create
products consisting of nitrogen, ammonia and hydrogen. This 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 thickness of 1.65 mm. c Only at a magnication of 100 can the
attached to the tag board on HST following approximately 21,000 shallow thermal fatigue cracks and surface roughening be detailed, but
thermal cycles between 100 and +90 C. Some circumferential cracks the majority of the joints circumference shows no cracking (de Rooij
are visible and a little orange peel effect (12 magnication). 1995)
b Microsection through the solder joint shown in a; the PCB has a
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
(b) Photoelectric yield data for the vitreous-carbon sensors count over two slides are seen in Fig. 8.11b. The max-
proved they were not affected by thruster heating or imum size of particles to be released is 20 m, the
contamination. majority had a diameter of under 5.4 m.
(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- 8.4.3 Returned Hardware
mina particlesmany thousands were entrapped on the
slides as black spherical particles and these slides were Micrometeorites and the <20-micron-sized particulate con-
subjected to image analysis (Quantimet 720B) and tamination ejected from thruster motors, as documented in
electron microscopy. The particle size and distribution Sect. 8.4.2, are dangerous but do not pose the same threat as
8.4 Micrometeoroids and Debris 515
space hardware debris. They are smaller in size and have a and thermal systems. 68 months after launch, the spacecraft
lower density, being composed of a material similar to sand. was recovered and returned to Earth. At least 34,000 impacts
Debris that originates from spacecraft is generally made of were recorded on all the space-exposed surfaces; some are
heavy metallic materials and may consist of steel bolts, illustrated in Figs. 8.13 and 8.14. Particles were noted to
copper wires, released aluminium lens caps, titanium fuel have pierced aluminium screens 0.8 mm thick but that
tanks, and the like. A great effort is now being made by the multilayer materials of teflonized woven glass were less
various world space agencies to control the amount of debris damaged (Durin et al. 1993). Solar cells with various cell
in space. As mentioned earlier in this chapter, many exper- covers showed that polymer cell covers provide very little
iments have been devised to evaluate the effects of the low protection. When the crater impact diameters were about
Earth orbit environment on spacecraft. Some of the ndings 100 m, 24 % degradation in short-circuit current was
will be summarized here, particularly with regard to impacts observedwith smaller impact sizes, no measurable chan-
of debris and meteoroids on the Eureca spacecraft returned ges in efciency were observed. If the cell itself was dam-
from space and the solar array of the Hubble Space Tele- aged, the loss in current was proportional to the size of the
scope (see also Fig. 8.4). Information on more than 20,000 damaged area (Gruenbaum and Dursch 1993). Similar
catalogued space objects is assessed by engineers as a basis impact holes were seen on the returned HST array (Fig. 8.4)
for risk assessment in the planning of future spacecraft and detailed in Fig. 8.15, but here they accounted for an
orbits. Similarly, a model for the micrometeoroid environ- insignicant loss of power. The Eureca spacecraft was
ment surrounding the Earth is available and continues to be retrieved in 1993 after almost one year in LEO. With a total
upgraded as new data are collected. Unfortunately, the active area of 140 m2 the observed impacts included 73 penetra-
removal of debris from orbit using laser evaporation or tions of the outer layer of the thermal blankets which cov-
debris sweeper mechanisms is in its infancy and may be ered most parts of the spacecrafts external body, and more
economically impossible. The focus must therefore be on than 2000 impacts on the front face of the solar arrays. The
preventative measures such as the maneuvering of spacecraft size of recorded impact features ranges from about 30 m to
and rockets out of heavily trafcked orbits by means of their 6.5 mm (Drolshagen et al. 1996). One impact on Eurecas
thruster motors, preferably into orbits that will carry these painted scuff plate is shown in Fig. 8.16. The ESABASE
parts back to the Earths atmosphere. Also, the practice of meteoroid and debris tool was used to predict particle flux
generating pieces of metal coming from the in-orbit frag- for the Eureca mission (for details of the model see Berthoud
mentation of pyrotechnic cutters, clamp bands, springs, and 1994). Figure 8.17 shows that there is a good comparison
other release mechanisms is now curtailed. One upper stage between Eureca measured and modelled fluxes and LDEF
of an Ariane IV launch vehicle exploded in a Sun syn- measured flux in terms of particle diameter.
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- 8.4.4 Protection Shields
gent general policies concerning launcher debris have been
made subsequent to this explosion (covering both Ariane 5 The growing amount of man-made objects in LEO may also
and 6) which drastically reduce the risk of creating debris. In require the protection of sensitive parts of unmanned
2009 a defunct Russian satellite collided with and destroyed spacecraft such as the electronic boxes for Earth observation
a US commercial satellite and this collision alone introduced satellites. In fact Canadas Radarsat incorporates modica-
another 2000 pieces of trackable debris to the catalogued tions designed to reduce the crippling effects of impacts.
space objects. In an incomprehensible 2007 test, China used Shielding blankets cover fuel lines, wiring bundles, and
a missile to blow-up an old satellite and this created a further electronic boxesthese add 17 kg to the satellites mass but
3000 pieces of debris. are said to increase the odds of survival of 87 % (David
The consequence of interaction between hypervelocity 1995). This satellite orbits the Earth at 800 km where there is
particles and spacecraft surface is rather obvious, but item- a high level of space debris, including tiny paint chips,
ized in Fig. 8.9b. The Space Shuttle windows had to be fragments from spent rocket stages, and fluids that leaked
regularly replaced because of particle impacts. The surfaces from a retired nuclear-powered spacecraft. Shields with
of other spacecraft that have been retrieved from LEO and multi-layer designs are termed Whittle shields as they had
brought back to Earth have been found peppered with craters been initially proposed by Frank Whittle in 1947. The
and pin-holes caused by particle impacts. The NASA LDEF development of the International Space Station (ISS), now
spacecraft, shown in Fig. 8.12, was deployed in an almost with a nal-build cross-section of about 11,000 m2,
circular orbit by STS-41C at an altitude of 477 km. One of prompted NASA and associated space agencies to pay more
its main purposes was to gather in situ data of particles in attention than ever to debris hazards during early design
LEO by means of experiments covering materials, coatings, stages (Chenard 1990). The ISS has been designed to
516 8 Assessment of Post-flight Materials
withstand impactors of up to 1 cm in size by the incorpo- been, shielded against using passive protection techniques.
ration of three kinds of Whittle shields (an aluminium alloy The Giotto spacecraft, seen in Fig. 4.44, achieved protection
bumper-plate to break up and melt a particle on impact; an by means of a Whipple Shield consisting of layers of alu-
outer bumper plate with an underlying blanket of Nextel minium, Nextal and Kevlar. According to studies, shields
ceramic cloth and Kevlar fabric spaced away from the can provide reasonable protection against fragments of less
module pressure shell; and, a multilayer shield consisting of than 1520 mm diameter, with velocities of 1012 km/s.
multiple layers of fabric and metal panels to protect the most These represent 78 % of the debris population. European
critical parts of ISS). studies favour shields of aluminium and ceramic cloth
Since only larger space objects can be catalogued and bumpers. Many possible combinations of protection-shield
tracked, it is these that have been avoided by evasive materials have been investigated by ground simulations and
manoeuversthe uncatalogued smaller debris can, and have one is illustrated in Fig. 8.18.
8.4 Micrometeoroids and Debris 517
The present activities are focusing on advanced mea- vessel pressure below the critical value, catastrophic failure
surement techniques for hypervelocity impact testing. Little can be minimized, thus reducing the risk of secondary
research has been published on the effects of such impacts damage to equipment and personnel.
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 8.5 Effect of Atomic Oxygen on Materials
constructed from aluminium alloy 6061-T62, with a burst
pressure rating of 15.3 MPa. It was indicated from their A satellite in geosynchronous orbit and operating at a height
results that the vessels internal pressure has an effect on the of 22 000 km will experience about 90 thermal cycles per
severity of impact damage. A critical pressure could be year, and at this altitude only vacuum radiation and thermal
identied above which catastrophic failure usually occurred flux will contribute to environmental wear-out mechanisms.
(for those vessels it was about 3.45 MPa). By maintaining A spacecraft in a Low Earth Orbit (LEO) will, in contrast,
518 8 Assessment of Post-flight Materials
experience up to 16 thermal cycles per day, and at an altitude after only 40 h of exposure. This is evident on Kapton
of between 200 and 700 km by far the greatest problem is samples observed after flight (Fig. 8.20), and high recession
degradation and erosion of externally exposed materials by rates have been measured by Pippin (1995).
atomic oxygen (see Table 2.3 and Fig. 8.19). Organic paint binders such as methyl silicone or poly-
The effects of atomic oxygen were not considered until urethane react with atomic oxygen to form volatile
engineers began to notice measurable, visible material loss carbon-based molecules which are removed from the paint
on Space Shuttle missions after only tens of hours in LEO. surface (Zimcik 1987; Jaggers 1993).
Oxidation and erosion were not a signicant problem on the Lubricant lms of molybdenum disulphide were exposed to
relatively short-term Space Shuttle missions, but similar atomic oxygen during some LEO flights. This lubricant is used
rates of material degradation were considered unacceptable on a variety of spacecraft mechanisms such as
on the 3040-year ISS programme. The hazard of radiation release/deployment devices and precision bearings. Post-flight
by neutral atomic oxygen arises from the photo-dissociation examinations indicate that the MoS2 can oxidize to MoO3,
of upper atmosphere molecular oxygen. This atomic oxygen which is an inferior lubricant with a higher friction coefcient.
impacts spacecraft surfaces in the ram direction with an If mechanisms are continually used in orbit they should be
energy of about 45 eV, owing to the relative velocity of the shielded from atomic oxygen to avoid accelerated wear.
orbiting spacecraft. The high velocity of the Shuttle (ap- Metals have been observed to react less than polymers to
proximately 8 km/s) produced a flux of atomic oxygen on a atomic oxygen attack, although important exceptions are
ram-facing surface in the order of 1015 atoms/cm2. This flux silver, osmium, and copper as shown in Table 8.2. Metal
has been found to be strongly chemically reactive with many oxides (SiO2, indium tin oxide, A12O3, etc.) are nonreactive
of the thermal control lms and paints usually used on the because they are already in their highest oxidation state.
external surfaces of conventional spacecraft. In particular, Carbonepoxy composite materials have shown a loss of
many organic polymer materials were severely degraded surface resin, and the carbon bres themselves have been
8.5 Effect of Atomic Oxygen on Materials 519
seen to be attacked and reduced to porous ridges with little Fig. 4.63). The appearance of silver interconnectors attacked
apparent residual strength or stiffness. by atomic oxygen is shown in Fig. 8.21. Progressive
The high capture efciency for atomic oxygen of silver degradation of these electrical connections is attributed to
has created a major problem for the manufacturers of solar the formation of AgO which, on reaching a certain thickness,
cell interconnectors, as these are generally made of either becomes non-adherent and detaches as flakes during thermal
thin silver foil or silver-coated molybdenum foil. Degrada- cycling (de Rooij 1985, 1989, 1995). Recent detailed mod-
tion of these interconnection materials severely jeopardizes elling by de Rooij (2010) uses two transport mechanisms to
the reliability of solar arrays designed or LEO. A combina- account for thee transport and oxidation of silver through
tion of atomic oxygen attack and excessive thermal cycling pores in the built-up oxide layernamely, gas-flow and
(e.g. 30,000 thermal cycles for the Hubble Space Telescope Fickian diffusion. Both mechanisms result in the parabolic
during a 5-year period, and 60,000 thermal cycles during the growth of oxide.
10-year life of Space Station) is known to result in a pre- Protective measures that either slow down or prevent the
mature loss of electrical power generation caused by the attack of atomic oxygen have been proposed by Dauphin
thinning and eventual failure of the solar cell interconnector (1987). They rely on the use of vacuum-deposited protective
itself or failure of the cell-to-interconnector weldment (see coatings known to be resistant to atomic oxygen, such as:
520 8 Assessment of Post-flight Materials
Metallic layers of platinum and aluminium. the atomic oxygen atmosphere to form silicates and silica
Stable oxide layers such as Al2O3, SiO2, or indium tin both protective compounds that are resistant to further
oxide. erosion.
Silicone layers that are possibly protective owing to the It is clear that a great deal of ground-based research is
formation of a thin, self-healing layer of silicon oxide. developing theoretical models to account for material
Fluorinated layers. degradation and protection methods to meet the challenge of
atomic oxygen. Ground-based atomic oxygen simulation
Many articles returned from LEO have been seen to facilities have been designed and commissioned (at ESTEC,
support a light-brown contamination layer. Analyses have PSI, DERTS, AODTS, PPPL, and UTIAS). The ground tests
shown this to be a deposited layer containing silicones on metals and polymers are giving results comparable to
(condensed volatile materials from outgassing sources); this some LEO exposures, but as often seen in nature, it is the
layer may only be a few angstroms thick; it then oxidizes in complex synergistic reactionsbetween atomic oxygen, UV
8.5 Effect of Atomic Oxygen on Materials 523
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 that have a good resistance to high-temperature oxidation, air
are almost impossible to reproduce in the laboratory. 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
8.6 Decelerators and Heat Shield Materials be used for one descent to the Earths surface. The Gemini
and Apollo TPSs were manufactured from an ultralight
8.6.1 General Examples ablator called ULD 100 by the McDonnell Douglas company
in the USA. Beagle 2 and launch vehicles such as Atlas V,
An overview of the various groups of materials that have Arianes 5 and 6, Vega and Falcon employed cork, as seen in
been selected as the outer skins of thermal protection systems Figs. 2.24 and 2.25.
(TPSs) were discussed in Sects. 2.4 and 2.5. Reusable launch The effects of re-entry into the Earths atmosphere on
vehicles and manned capsules that return to Earth have heat materials are shown schematically in Fig. 8.19b. A similar
shields made either of lightweight silicon-based ceramic tiles chain of reactions can be expected for the TPSs attached to
8.6 Decelerators and Heat Shield Materials 525
Fig. 8.21 Degradation of silver mesh solar cell interconnectors due to microcracks, primary and secondary oxide layers depict the growth and
atomic oxygen (Courtesy of A. de Rooij.): a Silver oxide flakes after spalling observed on (a). c A recent observation made by de Rooij
atomic oxygen attack during STS flight. Cross-sections revealed the (2010) and D. Adams, is that the oxidized silver surface consists of
silver interconnector to have thinned from an original thickness of 34 to nano-flakes and pores
29.5 m. b Flaking of oxide layer due to internal stresses and
unmanned space probes that are designed to descent onto the braking is required as the probe makes contact with the planet
surfaces of those planets and moons possessing their own or moons atmosphere at a supersonic speed. On reaching a
gaseous atmospheres. Space probe missions can usually be subsonic speed the decelerator structure is parted from the
divided into four phases: launch from the Earth, coast through probe by the ring of a pyrotechnic device. The probe descent
space for long periods of time, entry into the planet or moons module consists of a TPS (heat shield) that protects the outer
atmosphere, and descent onto its surface. The design of these envelope of the probe, its equipment, and antennae. When the
probes usually consists of two major components: The decelerator is jettisoned, the descent module may be further
high-speed deceleration system and the descent module. The slowed down by the deployment of a pilot parachute which
high-speed deceleration system is often deployable owing to drags out the main parachute. The probe must possess
the volume restrictions of most launch vehicles. It must be energy-absorbing properties to cushion it as contact is made
made of a material having a high heat resistance, low mass, and with the planet or moons surface.
adequate strength at elevated temperatures. This decelerator is The Cassini-Huygens mission is a joint NASA and ESA
used only in the entry phase of the mission when aerodynamic project to Saturn and its major moon, Titan. The Cassini
526 8 Assessment of Post-flight Materials
spacecraft was launched in 1997 in an orbiter (Cassini) and generators were selected as they remain unmatched for power
probe (Huygens) conguration. After encounters with Jupiter output for missions to the outer solar systemat the time of
and a portion of the asteroid belt, it reached Saturn in the year writing, this spacecraft continues to relay incredible images of
2003. On completing its rst orbit of the planet, the spacecraft Saturn and its many moons (Fig. 8.25).
released the Huygens probe in the direction of Titan. Eleven The actual material that was selected for the
days later the Huygens probe entered Titans dense atmo- high-temperature thermal protection of the Huygens probe is a
sphere (expected to be composed of nitrogen and methane). low-density ablator called AQ 60 I (a trademark of
Deceleration will be by means of an aerodynamic braking Aerospatial, France). This material is in full industrial pro-
structure. The design of the Huygens probe is drawn in duction, being used for the protection of strategic military
Fig. 8.22. The expected entry and descent scenario is illus- missiles. It is in fact a felt, made up of short silica bres rein-
trated in Fig. 8.23 and in situ images are included in Fig. 8.24. forced by an impregnation of phenolic resin. AQ 60 I has a nal
NASA attempted to use solar arrays to power Cassini but density of 0.3 and a total porosity of 84 %. It is not reusable
these would not have been possible as the spacecraft would because part of the material becomes pyrolysed between 400
have become too massive. Radioisotope thermoelectric and 1000 C. After pyrolysis, this heat-shield material is still
8.6 Decelerators and Heat Shield Materials 527
Fig. 8.23 a The Huygens probe (ESA) after separation from the and the heat shield separated. c Huygens descended through the clouds
mothership, the Cassini Saturn Orbiter (NASA), is depicted in these of Titan, analyzing its atmosphere and transmitting data for 2 h 27 min,
artists impressions heading towards Titan. b Huygens: after entry in after landing it transmitted signals via Cassini for 72 min before its
Titans atmosphere on 14th January 2005, the parachute was deployed batteries drained (Courtesy or ESA)
self-supporting and becomes an efcient thermal insulator temperature and pressure, atmospheric composition (by mass
which then behaves like all the silicacarbon systems, being spectroscopy and gas chromatography), cloud particle sizes
ablated at temperatures in the region of 2000 C. and composition, lightning detection, and, for the last phase
During descent, the various sets of instruments on-board of the descent, surface imaging was successful as evidenced
the Huygens probe were activated in order to measure 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
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 531
cross-sections. Sharp notches were machined into the Be oxidation and heat-resistant coatings, have potential TPS
from the underside; the samples were then pulled in tension applications. As previously described, some very suitable
and this resulted in brittle fractures at right angles to the materials have been developed and are in use today. However,
anodized surface. Figure 8.28 shows that the nominally thin others are being evaluated for future applications such as those
(5 m)-anodized layer is in fact 2 m in thickness. Similarly, needed for future space transportation vehicles and space-
Fig. 8.29 shows that the thick-anodized layer is also of planes where surfaces will reach 1600 C. Heat shields such as
inadequate thickness, being only 8 m instead of the nominal ceramic tiles may be attached to these spacecrafts structural
15 m. After thermal cycling there is good adherence materials in order to provide for protection during descent
between oxide and substrate, but more microcracks in the phases of a mission but, for future designs, the outer skin
thicker oxide layer. This holds true following thermal cycling structure will also need to be also resistant to this environment.
plus the 600 C nitrogen exposures. After an additional The following materials are presently identied for structural
1000 C exposure, plus salt spray testing for 8 days, there are applications in the various high-temperature ranges:
extensive signs of oxide breakdown and corrosion of the
Be-substrate and this has occurred for both oxide thicknesses.
10001600 C Carboncarbon composites, also with
D. Conclusion active cooling; these include carbon felts
(e.g. Calcarb, a rigid felt developed in
The thin (actually 2 m)-anodized layer shows a better Ireland; Carbone Lorraine, a flexible felt
mechanical performance regarding cracks and detachment, made in France, and Aerospatials AQ
but the thicker layer (actually 8 m) is better suited in view 61)limitations apply owing to oxidiz-
of a high emittance. For both layers the emittance of the ing properties, but in space these mate-
anodized surface decreases below the required = 0.7 after rials have excellent outgassing
the 600 and 1000 C nitrogen exposures. characteristics
7001000 C Carbonsilicon carbide composites, also
with coatings, and compacted silica
8.6.3 Alternative Heat Shield Materials powder (e.g. Microtherm, manufactured
in the UK by the Micropore Company)
An almost innite number of composite materials, metallic 400700 C Titanium aluminides and
alloys, ceramics, and glasses, together with their numerous high-temperature titanium alloys such as
532 8 Assessment of Post-flight Materials
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 industrialized. The disadvantages of this system are that the
coatings thermoviscosity of the coating does not permit for very high
less than 400 C Conventional metal alloys, oxide-disper- mechanical clamping stresses at xation points, or for
sion-strengthened aluminium alloys, extremely high pressures of gas flow. The internal oxidation
carbon bre reinforced plastics, etc. protection system illustrated in Fig. 8.30 has been developed
and tested. Initially, the C/SiC material with an outer coating
Two concepts for the oxidation protection of ceramic of chemical vapour deposited SiC was tested in static oxi-
matrix composites are shown in Fig. 8.30. The carbon bres dation tests (Lamouroux et al. 1994)that was without the
are encapsulated in a matrix of silicon carbide that is intro- internal oxidation protection. The results were encouraging,
duced by processes involving liquid phase inltration and/or but more recently a great improvement has been achieved
chemical vapour inltration. when the matrix composition was modied to contain an
For external oxidation protection, the multicomponent internal oxidation protection formulation. This formulation
oxidation protection (or thermoviscous external oxidation consists of a glass sealing agent (having a good wetting to
coating) was developed in Europe for the Hermes space- the carbon bres and the silicon carbide matrix, and a low
plane. It has an effective operational temperature range from gas pressure under vacuum at high temperatures), an oxygen
500 to 1600 C, a proven repairability, and utilizes relatively getter, and refractory particles.
inexpensive materials and processes that can be readily
8.6 Decelerators and Heat Shield Materials 533
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 Wherever possible the designer will use graphite- and
ceramic-based fasteners for these xations, but screws, bolts,
Adhesive compounds are available to secure heat shield and nuts made of these materials are very difcult to man-
elements onto spacecraft outer structures. The Shuttle tiles ufacture and procure. Also, they can only be used once and
were described in Sect. 2.4 and the fact that on re-entry the need to be drilled free unless in the particular application it is
outer skin of this vehicle is forced to bend owing to aero- possible to apply a suitable lubricant. The ceramic matrix
dynamic pressure and, in part, to the differential heating rates material described at the end of Sect. 8.6.3, and illustrated in
and enormously high temperatures seen by the tiles them- Fig. 8.30, is being developed for fastener applications. Bolts
selves. Silicone adhesives such as RTV 560 are used to bond and nuts (size KM 8) have been fabricated from C/SiC
the ceramic tiles onto the aluminium skin structure, as this material with coatings of SiC. Test programmes utilized
adhesive can flex and absorb most dimensional mismatches. these fasteners (they were coated with the additional external
In many applications, high-temperature fasteners will be oxidation protection layer seen in Fig. 8.30) to mechanically
required for the attachment and disassembly of spacecraft hold boreholed atmospheric re-entry demonstrator plates for
parts including those associated with rigid external insula- long periods at 1600 C in air. Under these conditions the
tion (hot outer shell elements), ceramic heat shields, hot fasteners survived a maximum load of 9.75 kN. Similar bolts
metallic structures, and items connected to propulsion are now being fabricated with enhanced internal oxidation
systems. protection, as drawn in Fig. 8.31.
534 8 Assessment of Post-flight Materials
Metallic fasteners may be more suitable than the graphite osmium is seriously toxic. The remaining twelve metals can
or ceramic fasteners for applications where there are specic be placed into three classes, described as the refractory
demands for ductility, thermal shock resistance, repeatable metals (hafnium, niobium, tantalum, chromium, molybde-
joining, or electrical and thermal conductivity. This may be num, and tungsten), rhenium from the 7th group in the
important during the integration of heat exchangers, Periodic Table, and the platinum group metals (platinum,
propulsion system parts, and heat shields. The most common ruthenium, rhodium, palladium, and iridium). Oxidation
high-strength spacecraft fastener materials are listed in resistance is the most important characteristic which differ-
Table 5.1 and described in Sect. 5.4, but for general space- entiates the three classes. Most refractory metals react
craft applications even the most advanced alloy (Waspaloy) strongly with oxygen at elevated temperatures and then
has a maximum service temperature of only 800 C. become severely embrittled at room temperaturehowever,
When the Periodic Table is consulted, it is noticed that chromium, molybdenum, and tungsten have volatile oxides
there are 15 metallic elements with sufciently high melting and they generally lose section thickness. Rhenium is similar
points (over 1700 C) for the very-high-temperature appli- to, but slightly more favourable than, tungsten. The platinum
cations required by the space industry (Lupton 1990). Some group of metals is usually highly resistant to oxidation but is
have been referred to in earlier chapters with respect to their costly, and has a low strength at temperatures greater than
use in electron tubes. Zirconium and vanadium have to be 1100 C. These strengths can be increased by alloying, and
excluded owing to poor high-temperature strength, and articial hardening is possible by means of adding a ne
8.6 Decelerators and Heat Shield Materials 535
dispersion of a thermodynamically stable oxide phase (e.g. handling or local stressing by the edges, corners, and
zirconium oxide) to alloys based on platinum and rhodium recesses of materials being joined. Although the oxidation
(Selman et al. 1974; Lupton 1990). resistance of assembled Nb-752 fasteners (coated) has been
Metallic fasteners for applications in the region of 1600 successfully demonstrated, it is noted that shear forces
C have been produced from the refractory metals, particu- between screw and nut must be minimized, and this requires
larly those based on niobium (sometimes called columbium). special assembly techniques.
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. 8.7 Manned Compartments
Si20Cr20Fe and SiCrTi), aluminides (e.g. A140Si and
AlSiTi), and noble metals such as platinum and iridium. 8.7.1 General Conditions
An example of one fastener design is shown in Fig. 8.32.
Coating thicknesses may range from 20 to 200 m, Human beings quest to live in space began in 1961, with the
depending on the application, and may be applied by rst orbital flight of Soviet cosmonaut Yuri Gagarin in the
physical vapour deposition, plasma spraying, pack cemen- Vostok 1 spacecraft. Later in the 1960s, the US Mercury and
tation, and slurry deposition. Loss of ductility resulting in Gemini programmes developed manned spaceflight, and
catastrophic failure may occur if local flaws exist in the these were followed by the Apollo programme that culmi-
oxidation-protection coating. Processing techniques and nated with several landings on the Moon. Skylab was the
nondestructive testing for these types of fastener have been rst American space station; it was launched in 1973 and
rened up to a point where coating defects can be elimi- visited by three sets of astronauts for periods of up to
nated. However, it is important to apply stringent assembly 84 days. Soviet cosmonauts stayed in space for even longer
rules that avoid damage to the coatings from either careless periods during the Salyut-6 and Mir space station missions
536 8 Assessment of Post-flight Materials
(longer than one year). This clearly demonstrated that with given to the proper selection of materials and processes
supplies of air, food, and water, people can function effec- during the design and fabrication stages of these spacecraft.
tively in space for long periods in order to perform useful Some considerations are familiar to those engineers who
experiments, observations, and services for those of us who have designed or lived in conned volumes, such as sub-
must remain on Earth. In the former Space Shuttle era, the marines! Past experience has been accumulated by the var-
human role has taken advantage of the lessons learnt by the ious space agencies and this has formed the basis for
early manned flights into space. Although not designed for establishing acceptable requirements. The majority of the
permanent presence in space, it was regrettable that the metallic materials and the manufacturing processes that were
Skylab space station was unceremoniously dumped into the selected for fabricating both manned and unmanned space-
Indian Ocean (and parts of Australia) because delays at the craft have been touched on between the covers of this book.
beginning of the Shuttle programme prevented a refurbish- There are though, specic problems that might occur within
ment mission. However, the Russian Mir space station, manned spacecraftthe main ones being highlighted in
launched in 1986, did prove to be an extremely durable Fig. 8.34and these can only be avoided by additional
vehicle owing to its design innovations which include a requirements. Chapters 2 and 3 detailed the main product
ve-port docking adapter onto which can be attached addi- assurance requirements set by the space agencies to cover
tional laboratory modules (see Fig. 8.33). This legend general spacecraft materials and processes. An excellent
proved Russia to be a leader in space technology as Mir overview of the technical issues related to materials safety
endured 15 years in orbit, three times its planned lifetime. for manned space have been documented by Pedley (2009).
By the time the International Space Station Alpha (Fig. 3.1) For manned vehicles, some of the additional safety
was launched and operational around 2003, Mir was still requirements relate to:
demonstrating its usefulness by some forms of low-cost
extended missions. The European-built Spacelab has flown Structural design providing ultimate factors of safety
on many missions inside Space Shuttle cargo bays since equal to or greater than 1.40; and, primary structures and
1983. After several refurbishments and recongurations with support bracketry precludes failure by the propagation of
new equipment and experiments, detailed inspection of the pre-existing flaws based on fracture control; special
interior structure of the fourteen-year-old Spacelab (see design burst pressure levels for pressure vessels
Fig. 3.2) shows no sign of degradation. Metallic materials being resistant to cracking due to
The success of the existing space station and laboratory stress corrosion (i.e. select from the alloys listed in
projects has been, in part, due to the importance that was Table 2.2)
8.7 Manned Compartments 537
A metallurgical investigation was performed on the video It was concluded that this commercial video camera con-
camera six weeks after the failure event. A summary of the tains design weaknesses that permit the intrusion of liquids.
ndings is as follows: Also, the quality of workmanship fell short of that normally
required for space hardware (ECSS-Q-ST-70-38); some line
the black plastic video camera case was not hermetically spacings were extremely narrow, the residues of soldering
sealed and there were many places where liquid could flux had not been removed, and the circuit had not been
penetrate adequately conformally coated. High humidity and conden-
close inspection showed that not all of the assembled sation of water onto all internal surfaces, including electronic
components were covered with a very thin layer of housings, within manned volumes is always a potential
conformal coating; some of the noncoated regions of the problem. This is due to the fact that life-support systems may
assembly had corroded (as shown on Fig. 8.35) not always be able to remove humidity quickly enough from
the residues from liquid solder flux were present adjacent the habitat module or laboratory module, particularly at times
to many solder joints when there is a large temperature difference between the
one typical region that supported corrosion products was walls, floors, and working space within the galleys.
photographed at a high magnication (as shown in It is interesting to note that the ISS includes a Mainte-
Fig. 8.36 in the colour section), the voluminous products nance Work Area consisting of a portable foldable work-
having characteristic colours: red, blue-green and white bench and tabletop. Astronauts can unfold it anywhere and
the distinct corrosion products were carefully removed clamp it to a slotted trackthese tracks are located on most
from the board and chemically analysed by of the floor-to-ceiling racks inside station. Many tests have
energy-dispersive X-ray analysis. been made related to soldering operations on ISS while
the results are given in Figs. 8.37, 8.38 and 8.39, white orbiting the Earth. These were made during a dedicated work
deposits being mainly lead oxide, bluegreen deposits programme in order to check if solder joint repair is feasible
containing copper chloride, and red deposits containing inside the manned modules (Flin 2005). The results are
cuprous oxide (chlorides are expected to originate from fascinating, and can be summarized as follows:
solder flux, as the AFFF liquid and original French tap
water showed no white precipitation when mixed with a when rosin-cored solder is applied to the heated tip of a
solution of silver nitratea very sensitive test for soldering iron the solder becomes molten and clings to
chlorides) the tip, the solder liquid then begins to spin round and
dark, electrically conductive traces (similar to the elec- round, getting progressively faster.
trochemical migration illustrated in Figs. 6.30, 6.49 and In low-gravity surface tension forces are greater than on
6.50) were seen beneath some of the corrosion products, Earth so soldering processes are more difcult to perform.
indicating that the camera failure mechanism resulted Splatter from liquid solder forms drops held together by
from the migration of tin between the very small track surface tension, these float through the air in space and
spacing when the circuit was damp. can be hazardous.
8.7 Manned Compartments 541
Low gravity eliminates convection and cooling is Smoke created during soldering does not easily disperse,
initiated at the molten solder surfaces, so entrapping possibly leaving the operator without a good view of the
gasses and solder flux and this leads to poor quality workpiece.
joints.
Debris from cutting and soldering neds to be captured Clearly soldering and conducting repairs in space is
and disposed of using the stations vacuum cleaner. challenging and it may be needed to select conductive
542 8 Assessment of Post-flight Materials
adhesives for repair work, despite their inferior properties programme. The hardware was manufactured in Europe after
when compared to traditional soldering (see Sect. 6.13). 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
8.7.3 Inspection of Spacelab Post-flight the Kennedy Space Centre. The inspections were made by
Hardware the writer in the presence of long-term NASA personnel who
had been associated with the day-to-day use and operation of
Spacelab has been one of ESAs most important projects and the hardware. The full report and an extensive description of
was a major contribution to the US Space Shuttle 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 8 Assessment of Post-flight Materials
have been published (Dunn and Stanyon 1997). This (a) Corrosion: of the Alodined weld and its heat affected
end-of-service inspection took place 16 years after Space- zone, degradation of anodised surfaces and condensation
labs integration (Fig. 8.40) and one year after the nal in cold areas.
in-orbit flight (Fig. 8.41). Spacelab flew on 22 Shuttle mis- (b) Painted surfaces: peeling, flaking, degradation of optical
sions between 1983 and 1996. Various flight congurations properties, atomic oxygen effects.
were qualied and flown with the pressurised, habitable (c) Thermal protection: thermal blanket degradation, wear,
laboratory module. The other components were: the Space- tarnishing, etc.
lab Pallet, a U-shaped un-pressurised platform onto which (d) Interface problems: wear, friction on trunnions and
were mounted experiments and telescopes having a direct mounting points, fasteners, wear of electrical connec-
exposure to space; an Instrument Pointing System attached tors, fluid disconnect problems, threaded connections,
to the Pallets for precise pointing and observation of the accidental damage (dings) needing added protection.
Earth and our planetary system; and, a Tunnel connecting (e) Limited life items/physical degradation: seals and asso-
the laboratory to the crew compartment of the Shuttle. ciated leak rates, relief valves (positive and negative
The main objective of the inspections was to identify pressure), brush motor start-up after long periods of
M&P problems which had been experienced and during non-activity (sticktion), harness attachment points and
Spacelabs life, assess their importance and feed this infor- insulation damage, condensate heat exchanger degrada-
mation to the engineers engaged in the selection of materials tion, evidence of microbial contamination.
and processes for future European manned space missions,
such as the Columbus module of ISS. The following gures illustrate some of problems listed
Some problems were identied during the inspection abovetheir captions describe in more detail the topo-
visit, but all-in-all, the post-flight hardware was considered graphical evidence (Figs. 8.42, 8.43, 8.44, 8.45, 8.46, 8.47,
to be in an excellent condition and generally it was the 8.48, 8.49, 8.50, 8.51, 8.52, 8.53, 8.54, 8.55 and 8.56).
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:
8.7 Manned Compartments 545
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. 8.53 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. 8.56 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)
(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
A wide variety of techniques may be used for the identi- Each chemical must be stored and handled according to the
cation of grain structures, phases, and other constituents in manufacturers recommendations. All chemicals are poten-
metals and alloys. Metallographers are often able to predict tially dangerous and it is assumed that the person mixing,
the chemical composition and processing history of a pouring, or etching is thoroughly familiar with their use. If
metallic sample by selectively etching its polished surface there is any uncertainty about their use, toxicity, or means of
and comparing the microstructure to those of reference disposal, the users Chemical and Safety Department should
samples in conjunction with published phase diagrams. be contacted.
The following chemical reagents (etch compositions) are The concentrations of acids are given in terms of specic
recommended by the author for the etching of metals and gravity (s.g.), or as a percentage (%) of the fully concen-
alloys commonly encountered during the metallurgical trated value.
assessment of electrical and structural spacecraft materials.
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
(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
T611 Solution and precipitation heat-treated, low T7352 Solution heat-treated and stress-relieved by
residual stresses compression to produce a permanent set of not
Material may, after quenching from the solution less than 1 % nor more than 5 %, and
heat-treatment temperature, receive minor precipitation heat-treated. The method and
straightening in an amount necessary to meet direction of compression shall be as agreed upon
tolerances specied on the drawing by purchaser and vendor
T62 Solution heat-treated and then articially aged T736 See T74
by the user T73651 Solution heat-treated and stress-relieved by
T651 Solution heat-treated and stress-relieved by stretching to produce a nominal set of 2 % but
stretching to produce a permanent set of 2 % no less than 1.5 % nor more than 3 %, and
nominal but not less than 1.5 % nor more than precipitation heat-treated
3 %, and articially aged. Product shall receive T736511 Solution heat-treated and stress-relieved by
no further straightening after stretching stretching to produce a nominal permanent set
T6510 Solution heat-treated and stress-relieved by of 1.5 % but not less than 1 % nor more than
stretching to produce a permanent set of 1.5 % 3 %, and precipitation heat-treated. Material
nominal, but not less than 1 % nor more than may receive minor straightening after stretching,
3 % and articially aged. Material shall receive to meet required dimensional tolerances
no further straightening after stretching T73652 Solution heat-treated and stress-relieved by
T6511 Solution heat-treated and stress-relieved by compression to produce a permanent set of not
stretching to produce a permanent set of 1.5 % less than 1 % nor more than 5 %, and
nominal, but not less than 1 % nor more than precipitation heat-treated. The method and
3 %, and articially aged. Material may receive direction of compression shall be as agreed upon
minor straightening after stretching by purchaser and vendor
T652 Solution heat-treated and stress-relieved by T74 (Previously T73 or T736.) Solution heat-treated
compression to produce a permanent set of 1.5 and articially aged to resist stress-corrosion
5 %, and precipitation heat-treated. During cracking
compression, primary focus shall be applied in T7451 Solution heat-treated and stress-relieved by
the axial direction and on individual rings stretching to produce a nominal permanent set
approximately nal dimensions of 2 % but not less than 1.5 % nor more than
T66 Solution and precipitation heat-treated 3 %, and precipitation heat-treated. Plate shall
T7 Solution heat-treated and overaged/stabilized receive no further straightening operations after
T71 Solution and precipitation heat-treated (castings) stretching
T72 Solution heat-treated and then articially over- T7452 Solution heat-treated and stress-relieved by
aged by the user compressing to produce a permanent set of 1
T73 See T74 5 % and overaged
T7351 Solution heat-treated and stress-relieved by T76 Solution heat-treated and articially aged suf-
stretching to produce a nominal permanent set cient to produce improved resistance to
of 2 % but not less than 1.5 % nor more than exfoliation
3 %, and precipitation heat-treated. Plate shall T761 Solution heat-treated and precipitation
receive no further straightening operations after heat-treated. The T7 tempers require closer
stretching control on ageing practice variables such as
T73510 Solution heat-treated and stress-relieved by time, temperature, heating-up rates, etc
stretching to produce nominal permanent set of T7651 Solution heat-treated and stress-relieved by
1.5 % but not less than 1 % nor more than 3 %, stretching to produce a permanent set of 2 %
and precipitation heat-treated. Material shall nominal but not less than 1.5 % nor more than
receive no further straightening after stretching 3 %, and articially aged sufcient to produce
T7311 Solution heat-treated and stress-relieved by improved resistance to exfoliation and
stretching to produce nominal permanent set of stress-corrosion cracking. Plate shall receive no
1.5 %, but not less than 1 % nor more than 3 %, further straightening after stretching
and precipitation heat-treated. Material may T76511 Solution heat-treated and stress-relieved by
receive minor straightening, after stretching, an stretching to produce a nominal permanent set
amount necessary to meet required dimensional of 1.5 %, but not less than 1 % nor more than
tolerances 3 %, and precipitation treated. Material may
570 Appendix 5: Aluminium Alloy Temper Designations
receive minor straightening, after stretching, of of 1.5 % but not less than 1 % nor more than
an amount necessary to meet required dimen- 3 %, and precipitation heat-treated
sional tolerances T852 Solution heat-treated and stress-relieved by
T77 Solution-treated and stabilized compression, to produce a permanent set of 1
T8 Solution heat-treated, cold-worked, and then 5 %, and precipitation heat-treated
articially aged T86 Solution heat-treated, cold-worked by a reduc-
T81 Solution heat-treated, cold-worked by the flat- tion of approximately 6 %, and then articially
tening operation, and then articially aged aged
T851 Solution heat-treated and stress-relieved by T861 Solution heat-treated, cold-reduced approxi-
stretching to produce a permanent set of 2 % mately 6 % in thickness, and precipitation
nominal but not less than 1.5 % nor more than heat-treated
3 %, and articially aged. Plate shall receive no T9 Solution heat-treated, articially aged, and then
further straightening after stretching cold-worked
T8511 Solution heat-treated and stress-relieved by T10 Cooled from an elevated temperature-shaping
stretching to produce a nominal permanent set process, cold-worked, and then articially aged
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
(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
(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
(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
(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 other materials. However, it will be some years before this
process is completed and fully implemented.
European EN specications for metal alloys are currently The European designation system for steels is set out in
being generated and adopted. These will progressively the specication EN10027 and in the ECISS information
supersede the various national standards for steels, as with circular DD214:1993 ECISS/IC10:1992. The EN designa-
tions for steels will consist of three parts:
Appendix 6: Metal Alloy Comparison Tables 587
European Standard number, e.g. EN 10083-1. In principal, EN steel numbers can be inferred for steels
Steel Name (grade)Symbolic letters and numbers by using the existing Werkstoff numbers, however there is
expressing the application and principal characteristics, no guarantee that such numbers have currently been either
e.g. 2C50 (C50E). agreed or adopted by any particular country. New speci-
Steel Numbera 5 digit designation based on the cations, once adopted, will be issued by each national
existing German Werkstoff (Wk.) number, with a further standards organisation, and any existing, competing speci-
two digits held in reserve, e.g. 1.1206. cations will be withdrawn.
The implementation of the new standards recognises
The examples given above make up the complete EN three levels of equivalence between EN designations and
designation: EN 10083-1 2C50 (C50E) 1.1206; See existing national grades of steel: Close Equivalent, Near
table 11029. Within the alloy tables, the Steel Name part Equivalent and Approximate Equivalent. The new EN
has been underlined (e.g. EN 10083-1 2C50 (C50E) 1.1206) designations included in the above tables are all near
to indicate the separate the parts of the designation. equivalents of BS970-1 steel grades.
Alloy equivalentNickel based alloys
588
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
(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
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
(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
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
(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 Chemical composition specications for wrought alu-
Aluminium Alloys minium alloys are now contained in a single CEN spec-
ication: EN 573Aluminium and aluminium alloys
European EN specications for metal alloys are currently Chemical composition and form of wrought products.
being generated and adopted. These will progressively Temper designations for wrought aluminium alloys are
supersede the various national standards for aluminium also contained in a single CEN specication: EN 515
alloys, as with other materials. However, it will be some Aluminium and aluminium alloysWrought products
years before this process is completed and fully imple- temper designations.
mented. EN designations specic to particular alloys, forms In addition EN 485 now contains conditions for delivery,
and conditions have been included in the above alloy tables properties and tolerances for wrought aluminium alloy
(many of these are still at the provisional stage). However, products.
there are several EN specications which cover the basic
characteristics of aluminium alloys: Similar specications will be issued in the future to cover
cast aluminium alloys.
(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
(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
(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 and Fe3O4 are more stable than NiO, and that the oxide
temperature. The network is based on work by Ellingham Fe2O4 are more stable than NiO, and that the oxide Fe2O3 is
and Richardson in the 1940s and 1950s and is still very the least stable of all.
useful today (see for instance Ellingham, H.J.T. (1944) Reduced pressures will enhance most reducing reactions
J. Soc. Chem. Ind., 63, 125133). The diagram indicates at as will increasing the ratio of hydrogen or carbon monoxide
any temperature what is thermodynamically possible for present in a gaseous mixturethese effects can also be
metallurgical reactions that are based on either thermal estimated from intersection points. The techniques are
reduction processes or thermal oxidation processes. described in books such as Metallurgical Thermochemistry,
At the point of intersection of any two curves the standard 6th Edition, (1993) by O. Kubaschewski, C. Alcock and P.
free energy for the chemical reaction is zero and this is the J. Spencer, (Pergamon, Oxford) and Free Energy of For-
point of equilibrium. The diagram can be used to show mation of Binary Compounds (1971), by T.B. Reed (MIT
approximately whether reduction or oxidation will take Press, London).
place. It is important to note that the Ellingham diagrams relate
As an example, below a temperature of about 1600 C, to the equilibrium conditionsthey take no account of the
pure magnesium is expected to reduce aluminium oxide. At kinetics of any oxidation or reduction reaction.
temperatures higher than 1600 C, when the curve for For alloys, it is worth noting that the most easily oxidized
magnesium oxide is above that of aluminium oxide, one constituent should be considered when consulting the dia-
would expect that pure aluminium would reduce magnesium gram. It is usual to discount elements that have a concen-
oxide to metallic magnesium. The approximate temperatures tration less than about 1 % as they are not able to form
when gases such as carbon monoxide might reduce metal continuous surface lms.
oxides can, in a similar way, be deduced from the various A recent, short but very useful articleEllingham dia-
points of graph intersections. grams, their Use and Misusetakes account of several case
It is possible to determine the partial pressure of oxygen examples where diagrams can be used for trouble shooting
in equilibrium at a given temperature with the couples Fe/ during processes such as heat treatment, brazing and bright
FeO, Fe/Fe3O4, Ni/NiO, and the like. The diagrams shows annealing (Stratton 2013).
that up to very high temperatures (>1500 C) the oxides FeO
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
Responsible for Materials and Process, Plus Mechanical Parts (eg Materials Manager):
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
Declared Material List Declared Process List Declared Mech. Parts List
greases and oils to not outgas and contaminate optical Microsectioning of welded and crimped joints*
systems, that rotating parts in vacuum do not cold-weld Tensile testing of welds and crimps*
etc.). NDT occasionally for welds*
The DPL should also be reviewed and possibly com- Outgassing tests to ECSS standard (can be done at ded-
piled by the M and P specialist to ascertain if each icated European labs), needed for a few organic materials
process can be considered non- critical or critical (glues, paints, etc.) in vicinity of optical systems (line of
based on previous usage, reliability, inspect ability, sight to consider condensation). Some laboratories
possibility to re-work in case of human or material offering testing for outgassing are mentioned on page 40.
errors (Table 7.A.1).
*Notecould use local university or industry
Appendix 9: Materials and Processes Standards
Related to Space (Released by ECSS, JAXA and
NASA) as of 2015
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.
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 ammability, etc. The Material Group numbers have been
to be approved by the end customer. They are compiled by detailed in Appendix 8.
the Prime Contractor. The contents of the DMLs will depend on the end cus-
Critical materials are tested and Validated. Dependent tomers requirements, these may follow ECSS-Q-ST-70 or
upon the criticality of the spacecraft, tests may include: can be relaxed in view of the specic contractual require-
outgassing-under-vacuum, stress corrosion testing, ments. Further description is given in Sect. 4.2.1.
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)
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 Annealing A generic term denoting a treatment, consisting
beta to the lower-temperature allotrope alpha phase. It may of heating to, and holding at, a suitable temperature followed
have a needlelike appearance in a photomicrograph and may by cooling at a suitable rate. It is used primarily to soften
have needle, lenticular, or attened bar morphology in three metallic materials, but also to simultaneously produce
dimensions. Its typical aspect ratio is about 10:1 desired changes in other properties or in microstructure. The
purpose of such changes may be, but is not conned to:
Activation The changing of a passive surface of a metal to a
improvement of machinability, facilitation of cold work;
chemically active state. (Contrast with passivation.)
improvement of mechanical or electrical properties, and/or
Age hardening Hardening by ageing, usually after rapid increase in stability of dimensions. When the term is used
cooling or cold working. (See also ageing.) without qualication, full annealing is implied. When
applied only for the relief of stress, the process is properly
Ageing A change in the properties of certain metals and
called stress relieving or stress-relief annealing
alloys that occurs with time at ambient or moderately ele-
vated temperatures after working or a heat treatment (natural Assembly A functional subdivision of a component, con-
or articial ageing) or after a cold-working operation (strain sisting of parts or subassemblies that perform functions
ageing). The change in properties is often, but not always, necessary for the operation of the component as a whole.
due to a phase change (precipitation), but it never involves a Examples: regulator assembly, power amplier assembly,
change in chemical composition of the metal or alloy gyro assembly, etc
Alloy A substance having metallic properties and being Assurance All the planned and systematic activities
composed of two or more chemical elements of which at implemented, and demonstrated as needed, to provide ade-
least one is a metal quate condence that an entity will full its requirements
Alphabeta structure A microstructure containing and Axial lead Lead wire extending from a component or
as the principal phases at a specic temperature module body along its longitudinal axis
Alpha case The oxygen-, nitrogen-, or carbon-enriched, - AWG American wire gauge
stabilized surface resulting from elevated temperature
Basketweave Alpha platelets with or without interleaved
exposure
platelets that occur in colonies in a Widmansttten structure
Analysis The determination of the essential qualities, per-
Batch That quantity of material that was subjected to unit
formance, and limitations of an item by cognitive or com-
chemical processing or physical mixing, or both, designed to
putational methods
produce a product of substantially uniform characteristics
Anomaly Any deviation from the expected situation
Bifurcated (split) terminal A terminal with a slot or split
Assurance All the planned and systematic activities opening in which conductors are placed before soldering
implemented, and demonstrated as needed, to provide ade-
Billet (1) A solid, semi-nished round or square product
quate condence that an entity will full its requirements
that has been hot-worked by forging, rolling, or extrusion;
usually smaller than a bloom. (2) A general term for wrought Chemical vapour deposition (CVD) The precipitation of a
starting stock used in making forgings or extrusions metal from a gaseous compound onto a solid or particulate
substrate
Blister Undesirable rounded elevation of the surface of a
polymer, whose boundaries may be more or less sharply Coarse grains Grains larger than normal for the particular
dened wrought metal or alloy or of a size that produces a surface
roughening known as orange peel or alligator skin in
Body-centred cubic lattice structure A unit cell which
wrought alloys
consists of atoms arranged at cube corners with one atom at
the centre of the cube Cold ow Movement of insulation (e.g. Teon) caused by
pressure
Brazeability The capacity of a metal to be brazed under the
fabrication conditions imposed into a specic suitably Cold solder connection A solder connection exhibiting
designed structure and to perform satisfactorily in the poor wetting and a greyish, porous appearance due to
intended service insufcient heat, inadequate cleaning before to soldering, or
excessive impurities in the solder
Brazing A group of processes that join solid materials
together by heating them to a suitable temperature and by Cold working Deforming metal plastically under condi-
using a ller metal having a liquidus above about 450 C tions of temperature and strain rate that induce strain hard-
(840 F) and below the solidus of the base materials. The ening. Usually, but not necessarily, conducted at room
ller metal is distributed between the closely tted surfaces temperature. (Contrast with hot working.)
of the joint by capillary attraction
Cold-worked structure A microstructure resulting from
Bridging A build-up of solder or conformal coating plastic deformation of a metal or alloy below its recrystal-
between parts, components leads, and/or base substrate lization temperature
forming an elevated path (see Fillet)
Colophony A natural resin obtained as the residue after
Brittle Permitting little or no plastic (permanent) deforma- removal of turpentine from the oleoresin of the pine tree,
tion prior to fracture consisting mainly of abietic acid and related resin acids, the
remainder being resin acid esters
Brittle fracture Separation of a solid accompanied by little
or no macroscopic plastic deformation. Typically, brittle Component A functional subdivision of a system, generally
fracture occurs by rapid crack propagation with less a self-contained combination of assemblies performing a
expenditure of energy than for ductile fracture function necessary for the systems operation. Examples:
power supply, transmitter, gyro package, etc
Cable Two or more insulated conductors, solid or stranded,
of equal length, contained in a common covering; or two or Conductor A lead, solid or stranded, or printed wiring path
more insulated conductors, of equal length, twisted or serving as an electrical connection
moulded together without common covering; or one insu-
Conguration Functional and physical characteristics of a
lated conductor with a metallic covering shield or outer
product as dened in technical documents and achieved in
conductor (shielded cable or coaxial cable)
the product (ISO 10007:1995)
Cast or casting Top fabricate an item by pouring molten
Conformal coating A thin electrically nonconductive pro-
metal into a shaped cavity and permitting the metal to
tective coating that conforms to the conguration of the
solidify. A cast can relate to the item or may be a synonym
covered assembly
for heat, that is an identiable chemistry lot
Contact angle The angle enclosed between half-planes,
Catalyst A substance that changes the rate of a chemical
tangent to a liquid surface and a solidliquid interface at their
reaction without undergoing permanent change in its com-
intersection. In particular, the contact angle of liquid solder in
position; a substance that markedly speeds up the cure of a
contact with a solid metal surface. An approximate value for
compound when added in minor quantity as compared to the
this may be determined by shadow projection or other means,
amount of primary reactants
by measuring after the solder has solidied. Note that the
Certication The act of verifying and documenting that contact angle is always the angle inside the liquid
personnel have completed required training and have
Contaminant An impurity or foreign substance present in a
demonstrated specied prociency and have met other
material that affects one or more properties of the material.
specied requirements
Glossary 631
A contaminant may be either ionic or nonionic. An ionic, or the set of information describing the essential characteristics
polar, compound forms free ions when dissolved in water, of a product
making the water a more conductive path. A nonionic sub-
Dewetting The condition in a soldered area in which the
stance does not form free ions, nor increases the waters
liquid solder has not adhered intimately, but has receded,
conductivity. Ionic contaminants are usually processing
characterized by an abrupt boundary between solder and
residue such as ux activators, nger prints, and etching or
conductor, or solder and terminal/termination area leaving
plating salts
irregularly shaped mounds of solder separated by areas
Contractor Supplier in a contractual situation (ISO covered with a thin-solder lm
8402:1994)
Diffusion welding (DFW) A high-temperature, solid-state
Corrective action Action taken to eliminate the causes of welding process that permanently joins faying surfaces by
an existing nonconformity, defect, or other undesirable sit- the simultaneous application of pressure and heat. The pro-
uation in order to prevent recurrence (ISO 8402: 1994) cess does not involve macroscopic deformation, melting, nor
relative motion of parts. A solid ller metal (diffusion aid)
Corrosion The deterioration of a metal by a chemical or
may or may not be inserted between the faying surfaces
electrochemical reaction with its environment
Disturbed solder joint Unsatisfactory connection resulting
Corrosion fatigue Cracking produced by the combined
from relative motion between the conductor and termination
action of repeated or uctuating stress and a corrosive
during solidication of the solder
environment at lower stress levels or fewer cycles than
would be required in the absence of a corrosive environment Dross Oxide and other contaminants that form on the sur-
face of molten solder
Creep Time-dependent strain occurring under stress. The
creep strain occurring at a diminishing rate is called primary, Ductility The ability of a material to deform plastically
or transient, creep; that occurring at a minimum and almost before fracturing. Measured by elongation or reduction of
constant rate, secondary, or steady-rate creep; that occurring area in a tension test, by height of cupping in a cupping test,
at an accelerating rate, tertiary creep or by the radius or angle of bend in a bend test. (Contrast
with brittleness; see also plastic deformation.)
Crevice corrosion A type of concentration cell corrosion;
corrosion caused by the concentration or depletion of dis- Electron beam welding (EBW) A welding process which
solved salts, metal ions, oxygen, or other gases, and such, in produces coalescence of metals with the heat obtained from
crevices or pockets remote from the principal uid stream, a concentrated beam composed primarily of high-velocity
with a resultant building up of differential cells that ulti- electrons impinging upon the surfaces to be joined
mately cause deep pitting. Localized corrosion of a metal
Elongated grain A grain with one principal axis signi-
surface at, or immediately adjacent to, an area that is shiel-
cantly longer than either of the other two
ded from full exposure to the environment because of close
proximity between the metal and the surface of another Elongation A term used in mechanical testing to describe
material the amount of extension of a test-piece when stressed. In
tensile testing, the increase in the gauge length, measured
Critical item Any item that introduces risk which could be
after fracture of the specimen within the gauge length, usually
unacceptable to the project and requires specic attention or
expressed as a percentage of the original gauge length
control in addition to that given to items not so categorized
Embrittlement The severe loss of ductility and/or tough-
Cure A chemical reaction that hardens and changes the
ness of a material, usually a metal or alloy
physical properties of a material
Encapsulating compound An electrically nonconductive
Deformation A change in the form of a body due to stress,
compound used to completely enclose and ll in voids
thermal change, change in moisture, or other causes. Mea-
between electrical components or parts
sured in units of length
Environment Conditions in which an item exists or is
Delamination A separation between plies within a base
operated
material or any planar separation within a multilayer printed
circuit board (PCB) Equiaxed structure A polygonal or spheroidal
microstructural feature having approximately equal dimen-
Design (1) Set of information which denes the essential
sions in all directions. In alphabeta titanium alloys, such a
characteristics of a product. (2) The process used to generate
632 Glossary
term commonly refers to a microstructure in which most of Flux A chemically active compound which, when heated,
the minority phase appears spheroidal removes minor surface oxidation
Equilibrium A dynamic condition of physical, chemical, Foil Among many denitions is: a at-rolled product
mechanical, or atomic balance, where the condition appears 0.127 mm (0.005 in.), or less, in thickness, regardless of
to be one of rest rather than change width. (Any at-rolled product thicker than this dimension is
not considered foil.) Only thickness, not width, is a factor in
Equipment An item designed and built to accomplish a
determining foil
specied purpose, which can be disassembled and retain its
capabilities after reassembly Forging (1) Plastically deforming metal, usually hot, into
desired shapes with compressive force, with or without dies.
Eutectic alloy An alloy of two or more metals that has one
(2) Reshaping a billet or ingot by hammering. (3) The process
distinct melting point. Eutectic solder is a tinlead alloy
of placing a powder in a container, removing the air from the
containing 63 %Sn and 37 %Pb which melts at 183 C
container, and sealing it. This is followed by conventional
Excessive solder joint Unsatisfactory solder connection forging of the powder and container to the desired shape
wherein the solder obscures the conguration of the
Fracture (1) The irregular surface produced when a piece
connection
of metal is broken. (2) To separate a metal or alloy into two
Face-centred cubic lattice structure A unit cell which or more pieces by an applied load
consists of atoms arranged at cube corners with one atom at
Friction welding A solid-state process in which materials
the centre of each cube face
are welded by the heat obtained from rubbing together sur-
Failure The termination of the ability of an item to perform faces that are held against each other under pressure
a required function
Glass meniscus The glass llet of a lead seal which occurs
Fatigue The phenomenon leading to fracture under repeated where an external lead leaves the package body
or uctuating stresses having a maximum value less than the
Grounding The connection of two or more areas to the
tensile strength of the material. Fatigue fractures are pro-
same potential difference. The space environment can result
gressive, beginning as minute cracks that grow under the
in high potential voltage and destructive arcing, most sub-
action of the uctuating stress
systems are grounded to the structure, but the grounding of
Fatigue failure Failure that occurs when a specimen electronic units is complex and can use a oating ground
undergoing fatigue completely fractures into two parts, or concept
has softened, or been otherwise signicantly reduced in
Ground segment The hardware and software required to
stiffness by thermal heating or cracking. Fatigue failure
launch and control a space vehicle, usually the launcher
generally occurs at loads which, if applied statically, would
produce little perceptible effect. Fatigue failures are pro- Hardness A measure of the resistance of a material to
gressive, beginning as minute cracks that grow under the surface indentation or abrasion; may be thought of as a
action of the uctuating stress function of the stress required to produce some specied
type of surface deformation. There is no absolute scale for
Fault mode The observable effect of the mechanism
hardness; therefore, to express hardness quantitatively, each
through which the failure occurs, e.g. short, open, fracture,
type of test has its own scale of arbitrarily dened hardness.
excessive wear
Indentation hardness may be measured by Brinell, Knoop,
Fillet A smooth concave build-up of material between two Rockwell, Scleroscope, and Vickers hardness tests
surfaces, e.g. a llet of solder between a component lead and
HAZ See heat-affected zone
a solder pad or terminal, or a llet of conformal coating
material between a component and printed circuit board Heat-affected zone That portion of the base metal which
(PCB) has not been melted, but whose mechanical properties or
microstructure have been altered by the heat of welding,
Filler metal The metal to be added in making a welding,
brazing, soldering, or cutting
brazed, or soldered joint
Heat treatment Heating and cooling a solid metal or alloy
Flake Powder of an essentially two-dimensional nature
in such a way as to obtain desired conditions or properties.
Flatpack A part with two straight rows of leads that are Heating for the sole purpose of hot working is excluded from
parallel to the part body the meaning of this denition
Glossary 633
Hermetic seal A seal which protects an enclosed circuit metal, or metal alloy lattice. Common examples are oxygen,
from corrosion by preventing the entry of such contaminants nitrogen, hydrogen, and carbon
as water vapour
Interstitial solid solution A type of solid solution that
HIP See hot isostatic pressing sometimes forms in alloy systems having two elements of
widely different atomic sizes. Elements of small atomic size,
Hot isostatic pressing (1) A process for simultaneously
such as carbon, hydrogen, oxygen, and nitrogen, often dis-
heating and forming a powder metallurgy compact in which
solved in solid metals to form this solid solution. The space
metal powder, contained in a sealed exible mould, is sub-
lattice is similar to that of the pure metal, and the atoms of
jected to equal pressure from all directions at a temperature
carbon, hydrogen, oxygen, and nitrogen occupy the spaces
high enough for sintering to take place. (2) A process similar
or interstices between the metal atoms
to the one explained in (1), but applied to castings in order to
close internal porosity Item Anything which can be individually described and
considered
Hot working Deforming metal plastically at such a tem-
perature and strain rate that recrystallization takes place Lap joint Joining or fusing of two overlapping metal sur-
simultaneously with the deformation, thus avoiding any faces with solder without use of any other mechanical
strain hardening attachment or support
HV See Vickers hardness number Liquidus In a constitution or equilibrium diagram, the locus
of points representing the temperatures at which the various
Hydride phase For instance, the phase TiHx formed in
compositions in the system begin to freeze on cooling or
titanium when the hydrogen content exceeds the solubility
nish melting on heating. (See also solidus.)
limit, generally locally due to some special circumstance
Longitudinal direction Usually, the direction parallel to
Hydrogen embrittlement A condition of low ductility in
the direction of working in wrought alloys or the direction of
metals resulting from the absorption of hydrogen
crystal growth in directionally solidied or single-crystal
Immersion cleaning Cleaning where the work is immersed cast alloys. Commonly, it corresponds to the direction par-
in a liquid solution. Impurities, Undesirable elements or allel to the direction of maximum elongation in a worked
compounds in a material material. (See also normal direction and transverse
direction.)
Inclusion A particle of foreign material in a metallic matrix.
The particle is usually a compound (such as an oxide, sul- Machinability The relative ease of machining a metal
phide, or silicate), but may be of any substance that is for-
Machining Removing material from a metal part, usually
eign to (and essentially insoluble in) the matrix
using a cutting tool, and usually using a power-driven
Ingot A casting of simple shape, suitable for hot working or machine
remelting
Macrostructure The structure of metals as revealed by
Inspection An activity such as measuring, examining, macroscopic examination of a specimen. The examination
testing, or gauging one or more characteristics of an entity may be carried out using an as-polished or a polished and
and comparing the results with specied requirements in etched specimen
order to establish whether conformity is achieved for each
Martensite (1) The alpha product resulting from cooling
characteristic (ISO 8402:1994)
from the beta phase region at rates too high to permit
Insufcient solder connection A solder connection char- transformation by nucleation and growth. Martensite is sat-
acterized by incomplete coverage of one or more of the urated with beta stabilizer. Also called martensitic alpha. (2)
metal surfaces being joined or by incomplete solder llets A generic term for micro-structures formed by diffusionless
phase transformation in which the parent and product phases
Integrated-circuit component An individually packaged
have a specic crystallographic relationship. Martensite is
functional circuit formed by depositing an active or passive
characterized by an acicular pattern in the microstructure in
electronic element on to a substrate
ferrous and nonferrous alloys. The amount of high-temper-
Intermetallic compound A phase in an alloy system hav- ature phase that transforms to martensite upon cooling
ing a restricted solid solubility range. Nearly all are brittle depends to a large extent on the lowest temperature attained,
and of stoichiometric composition there being a distinct starting temperature (Ms) and a tem-
Interstitial element An element with a relatively small perature at which the transformation is essentially complete
atom which can assume a position in the interstices of a (Mf), which is the martensite nish temperature
634 Glossary
Maintainability The ability of an item under given condi- Overageing Ageing under conditions of time and temper-
tions of use, to be retained in, or restored to, a state in which ature greater than those required to obtain maximum change
it can perform a required function, when maintenance is in a certain property. (See ageing.)
performed under given conditions and using stated proce-
Overheated joint An unsatisfactory solder joint, charac-
dures and resources (IEC 50:1992)
terized by rough solder surface; dull, chalky, grainy, porous
Material A raw or semi-nished substance or compound, of or pitted
specied characteristics, which is processed to form a part of
Oxidation (1) A reaction in which there is an increase in
a nished product
valence resulting from a loss of electrons. (Contrast with
Matrix The constituent which forms the continuous or reduction.) (2) A corrosion reaction in which the corroded
dominant phase of a two-phase microstructure metal forms an oxide; usually applied to reaction with a gas
containing elemental oxygen, such as air
Measling/measles A condition existing in the base laminate
of printed-circuit board in the form of discrete white spots or Part An element of a component, assembly, or subassembly
crosses below the surface of the base laminate, reecting a that is not normally subjected to further subdivision or dis-
separation of bres in the glass cloth at the weave intersec- assembly without destruction of designed use
tion. During soldering, may be caused by excessive heat,
Passivation The changing of a chemically active surface of
mechanical stresses, or chemical attack
a metal to a much less reactive state
Mechanical part A piece of hardware which is not elec-
Performance Those generally quantied aspects of an item
trical, electronic, or electromechanical, and which performs
observed or measured from its operation or function
a simple (elementary) function or part of a function in such a
way that it can be evaluated as a whole against expected Pickling Removal of the oxide lm on a casting by a
performance requirements and cannot be disassembled chemical process; pickling is sometimes used solely to show
without destroying this capability up defects
Mechanical properties The properties of a material that Physical properties Properties of a metal or alloy that are
reveal its elastic and inelastic (plastic) behaviour when force relatively insensitive to structure and can be measured without
is applied, thereby indicating its suitability for mechanical the application of force; for example, density, electrical con-
(load-bearing) applications. Examples are elongation, fatigue ductivity, coefcient of thermal expansion, magnetic perme-
limit, hardness, modulus of elasticity, tensile strength, and ability, heat capacity, and lattice parameter. Does not include
yield strength chemical reactivity. (Compare with mechanical properties.)
Melting point The temperature at which a pure metal, Plastic deformation The permanent (inelastic) distortion of
compound, or eutectic changes from solid to liquid; the metals under applied stresses
temperature at which the liquid and the solid are in
Plate A at-rolled metal product of some minimum thick-
equilibrium
ness and widthat times less than 610 mm (24 in). (It is
Mission The specic task, duty, or function dened to be relatively thick when compared with sheet.)
accomplished by a system
Plated-through hole A plated-through hole is one formed
Modulus of elasticity A measure of rigidity or stiffness of a by a deposition of metal on the inside surface of a through
metal; the ratio of stress, below the proportional limit, to the hole. Also known as a supported hole. The conguration is
corresponding strain used to provide additional mechanical strength to the sol-
dered termination or to provide an electrical interconnection
Nonconformance Nonfullment of a specied requirement
on a multilayer PCB
(ISO 8402:1994denition of Nonconformity)
Potting compound An electrically nonconductive com-
Nonwetting A condition whereby a surface has contacted
pound used to partially encapsulate or for a ller between
molten solder, but the solder has not adhered to all of the
parts, conductors, or assemblies
surface; basis metal remains exposed
ppm Parts per million
ODS Oxide dispersion strengthening
Precipitation (1) Separation of a new phase from solid or
Outgassing The release of volatile parts from a substance
liquid solution, usually with changing conditions of time,
when placed in a vacuum environment
temperature, and stress. (2) The removing of a metal from a
Glossary 635
solution caused by the addition of a reagent by displacement. Quality assurance All the planned and systematic activities
(3) The removal of a metal from a gas by displacement implemented within the quality system and demonstrated as
needed, to provide adequate condence that an entity will
Precipitation hardening Hardening caused by the precip-
full requirements for quality
itation of a constituent from a supersaturated solid solution
Quality control Operational techniques and activities that
Preform An initially pressed compact to be subjected to
are used to full requirements for quality
repressing or forging
Quenching Rapid cooling. When applicable, the fol-
Preheat An early stage in the sintering procedure when, in a
lowing more specic terms should be used: direct
continuous furnace, lubricant or binder burnoff occurs
quenching, fog quenching, hot quenching, interrupted
without atmosphere protection prior to actual sintering in the
quenching, selective quenching, spray quenching, and
protective atmosphere of the high heat chamber
time quenching
Pressure vessel A container which stores pressurized uids
Ram direction he side that points in the direction of the
and: (a) contains stored energy of 19,310 joules or greater,
spacecrafts motion. In low earth orbit, it is the side
based on the adiabatic expansion of a perfect gas; or (b)
impacting/ramming into the rareed atmosphere that con-
contains a gas or liquid which may result in a hazardous
tains reactive atomic oxygen (AO, or ATOX)
event if released (c) will experience a design limit pressure
greater than 0.69 MPa. Recrystallization (1) Formation of new, strain-free grain
structure from the structure existing in cold-worked metal.
Printed circuit board (PCB) A product resulting from the
(2) A change from one crystal structure to another, such as
process of selectively etching unwanted copper from one or
that occurring upon heating or cooling through critical
both surfaces of a copper-clad insulating substrate to form a
temperature
desired circuity pattern which is subsequently solder- or
gold-plated. Holes may, or may not, be drilled in the board, Reducing atmosphere A chemically active protective
depending on the intended technique for attaching atmosphere which at elevated temperature will reduce metal
components oxides to their metallic state. (Reducing atmosphere is a
relative term and such an atmosphere may be reducing to
Procedure Specied way to perform an activity
one oxide but not to another oxide.)
Process Set of interrelated resources and activities which
Reduction in area (1) Commonly, the difference, expressed
transform inputs into outputs
as a percentage of original area, between the original cross-
Product The result of activities or processes and may sectional area of a tensile test specimen and the minimum
include service, hardware, processed materials, software, or cross-sectional area measured after complete separation. (2)
a combination thereof The difference, expressed as a percentage of original area,
between original cross-sectional area and that after straining
Product assurance A discipline devoted to the study,
the specimen
planning, and implementation of activities intended to
ensure that the design, controls, methods, and techniques in Reliability The probability that an item can perform a
a project result in a satisfactory level of quality in a product required function under given conditions for a given time
interval
Project A unique set of coordinated activities, with denite
starting and nishing points, undertaken by an individual or Rem Remainder
organization to meet specic objectives within dened
Repair Operations performed on a nonconforming article to
schedule, cost, and performance parameters (BS 6079)
place it in usable condition. Repair is distinguished from
Qualication The process of determining that the product, as rework in that alternate processes rather reprocessing are
designed, is capable of meeting all its specied performance employed
requirements in its operational environment with margins
Residual stress Stress remaining in a structure or member
appropriate for the technologies used and the intended
as a result of thermal or mechanical treatment or both.
application
Stress arises in fusion welding primarily because the weld
Quality The totality of characteristics of an entity that bear metal contracts on cooling from the solidus to room
on its ability to satisfy stated and implied needs temperature
636 Glossary
Resistance brazing Brazing by resistance heating, the joint compositions nish freezing on cooling or begin to melt on
being part of the electrical circuit heating. (See also liquidus.)
Rework The reprocessing of an article or material that will Solution heat treatment A heat treatment in which an alloy
make it conform to drawings, specications, and contract is heated to a suitable temperature, held at that temperature
long enough to cause one or more constituents to enter into
Risk A quantitative measure of the magnitude of a potential
solid solution, then cooled rapidly enough to hold these
loss and the probability of incurring that loss
constituents in solution
Safety A state in which the risk of harm (to persons) or
Specication Document stating requirements
damage is limited to an acceptable level
Spot welding Welding of lapped parts in which fusion is
Scaling (1) Forming a thick layer of oxidation products on
conned to a relatively small circular area. It is generally
metals at high temperature. (2) Depositing water-insoluble
resistance welding, but may also be gas tungsten-arc, gas
constituents on a metal surface, as in cooling tubes and water
metal-arc, or submerged-arc welding
boilers
Stress The intensity of the internally distributed forces or
SCC Stress-corrosion cracking
components of forces that resist a change in the volume or
Sheet A at-rolled metal product of some maximum shape of a material that is or has been subjected to external
thickness and minimum width arbitrarily dependent on the forces. Stress is expressed in force per unit area and is cal-
type of metal. It is thinner than plate and has a width-to- culated on the basis of the original dimensions of the cross-
thickness ratio greater than about 50 section of the specimen. Stress can be either direct (tension
or compression) or shear. Usually expressed in megapascals
Shield A metallic sheath surrounding one or more wires,
(MPa)
cables, cable assemblies, or a combination of wires and
cables that is used to prevent or reduce the transmission of Stress-corrosion cracking Failure of metals by cracking
electromagnetic energy to or from the enclosed conductors. under combined action of corrosion and stress, residual or
The shield also includes an insulating jacket that may cover applied. In brazing, the term applies to the cracking of
the metallic sheath stressed base metal due to the presence of a liquid ller metal
Solder A nonferrous, fusible metallic alloy used to join Stress relief Related to electronic assemblies: the formed
metallic surfaces portion of a conductor that provides sufcient length to
minimize stress between terminations
Solderability The property of a surface that allows it to be
wetted by a molten solder Stress-relief heat treatment Uniform heating of a structure
or a portion thereof to a sufcient temperature to relieve the
Solder connection An electrical/mechanical connection
major portion of the residual stresses, followed by uniform
that employs solder for the joining of two or more metal
cooling
surfaces
Stress relieving Heating to a suitable temperature, holding
Solder mask Coating material used to mask or protect
long enough to reduce residual stress, and then cooling slowly
selected areas of a pattern from the action of an etchant,
enough to minimize the development of new residual stresses
solder, or plating
Striation A fatigue fracture feature often observed in elec-
Solder spike A cone-shaped peak or sharp point of solder
tron micrographs that indicates the position of the crack front
usually formed by the premature cooling and solidication
after each succeeding cycle of stress. The distance between
of solder on removal of the heat source
striations indicates the advance of the crack front across that
Solidication The change in state from liquid to solid on crystal during one stress cycle, and a line normal to the
cooling through the melting temperature or melting range striation indicates the direction of local crack propagation
Solid solution A solid crystalline phase containing two or Substrate The layer of metal underlying a coating,
more chemical species in concentrations that may vary regardless of whether the layer is base metal
between limits imposed by phase equilibrium
Subsystem Set of interdependent elements constituted to
Solid solution strengthening A mechanism for strength- achieve a given objective by performing a specied function,
ening the alloy by dissolved elements in solid solution but which does not, on its own, satisfy the customers need
Solidus In a constitution or equilibrium diagram, the locus Supplier An organization that provides a product to the
of points representing the temperatures at which various customer
Glossary 637
Surface hardening A generic term covering several pro- Verication Conrmation by examination and provision of
cesses applicable to a suitable ferrous alloy that produces, by objective evidence that specied requirements have been
quench hardening only, a surface layer that is harder or more fullled
wear resistant than the core
Vickers hardness number (HV) A number related to the
Surface mounting The electrical connection of components applied load and the surface area of the permanent impres-
to the surface of a conductive pattern that does not utilize sion made by a square-based pyramidal diamond indenter
part holes having included face angles of 136
Technology Readiness Level (TRL) Levels 19 used to Viscosity A measure of the resistance of a uid to ow
dene the maturity of a technical concept, from basic prin-
Waiver Written authorization to use or release a product
ciple to mission proven
which does not conform to the specied requirements
Temper In nonferrous alloys the hardness and strength
Wake The side opposite to the Ram side of a spacecraft. It
produced by mechanical or thermal treatment, or both, and
faces away from the spacecrafts motion
characterized by a certain structure, mechanical properties,
or reduction in area during cold working Wave soldering A process wherein printed circuit boards
are brought in contact with the surface of continuously
Tensile strength In tensile testing, the ratio of maximum
owing and circulating solder
load to original cross-sectional area. Also called ultimate
strength Weave exposure A surface condition of a printed-circuit-
board laminate in which the unbroken woven-glass cloth is
Test A formal process of exercising or putting to trial a
not uniformly covered by resin
system or item by manual or automatic means to identify
differences between specied, expected, and actual results Weldability A specic or relative measure of the ability of
a material to be welded under a given set of conditions.
Thermal shunt A device with good heat-dissipation char-
Implicit in this denition is the ability of the completed
acteristics used to conduct heat away from an article being
weldment to full all service designed into the part
soldered
Wetting Flow and adhesion of a liquid to a solid surface,
Tinning The coating of a surface with a uniform layer of
characterized by smooth, even edges, and a low dihedral
solder
angle
Traceability Ability to trace the history, application, or
Wicking A ow of molten solder or cleaning solution by
location of an entity by means of recorded identications
capillary action. Occurs when joining stranded wire; solder
Transverse direction Literally across. Usually signifying is drawn within the strands, but may not be visible on outer
a direction or plane perpendicular to the direction of work- surface of strands. Wicking may also occur within the stress-
ing. In rolled plate or sheet, the direction across the width is relief bend of a component lead
often called long transverse, and the direction through the
Widmansttten structure A structure characterized by a
thickness, short transverse
geometrical pattern resulting from the formation of a new
Ultimate strength The maximum stress (tensile, compres- phase along certain crystallographic planes of the parent
sive, or shear) a material can sustain without fracture, solid solution. The orientation of the lattice in the new
determined by dividing maximum load by the original cross- phase is related crystallographically to the orientation of
sectional area of the specimen. Also known as nominal the lattice in the parent phase. The structure was origi-
strength or maximum strength nally observed in meteorites, but is readily produced in
many alloys, such as titanium, by appropriate heat
Unaided eye Normal Snellen 20/20 vision, including eye
treatment
glasses required to correct defective vision to 20/20 equiv-
alent. Does not include microscopes, eye loupes, or any Wire A thin, exible, continuous length of metal, usually of
other magnifying device circular cross-section, and usually produced by drawing
through a die
Vacuum melting Melting in a vacuum to prevent contam-
ination from air, as well as to remove gases already dis- Workmanship The physical characteristics relating to the
solved in the metal; the solidication may also be carried out level of quality introduced by the manufacturing and
in a vacuum or at low pressure assembly activities
638 Glossary
Wrought A metal or alloy which has been deformed plas- exhibit a yield point. If there is a decrease in stress after
tically one or more times and which exhibits little or no yielding, a distinction may be made between upper and
evidence of cast structure lower yield points
Yield point The rst stress in a material, usually less than Yield strength The stress at which a material exhibits a
the maximum attainable stress, at which an increase in strain specied deviation from proportionality of stress and strain.
occurs without an increase in stress. Only certain metals An offset of 0.2 % is used for many metals
References
Adams, D. M. (1994). Failure investigation into HST solar array drive ASTM D3359. (2010). Standard test methods for measuring adhesion
electronics. ESTEC Metallurgy Report No. 2038 (condential). by tape test.
Adile, E., & Dunn, B. D. (1999). Effecto della geometria del raccordo ASTM E595. (2007). Standard test method for total mass loss and
di lega saldante sul comportamento a fatica meccanica di saldature CVCM from outgassing in a vacuum environment.
di component elettronici a montaggio superciale. Rivista Italiana ASTM E835-94A. (1994). Standard method for determining residual
della Saldatura,4, 401412. stresses by the hole-drilling strain gage method.
Adler, I. (1988). The analysis of extraterrestrial materials. In Chemical ASTM E2089. (2014). Standard practices for ground laboratory atomic
analysis (Vol. 81). New York: Wiley. oxygen interaction evaluation of materials for space applications.
Akinade, K., et al. (1995). Lead-free solder pastes evaluation. Soldering Athimoolam, M., & Moorthy, T. V. (2012). Polymer nanocomposite
and SMT Journal,20, 5053. materials and shape memory applicationsA review. Procedia
Allen, J. S., & Jubb, J. E. M. (1991). Dispute resolutionThe Engineering,38, 33993408.
increasing role of the expert. Welding and Metal Fabrication Autumn, K., et al. (2000). Adhesive force of a single gecko foot-hair.
Journal, 8184. Nature,405, 681685.
Ambat, R., et al. (2009). Solder ux residues and electrochemical Babel, H. W., et al. (1994). Understanding and controlling the
migration failures of electronic devices. In Proceedings of Eurocorr degradation of silicone on exposed spacecraft surfaces. In Proceed-
2009 (pp. 114), paper 8141, Nice, France, September 2009. ings of the 6th International Symposium on Materials in Space
AMS 4942C. (1984). Titanium tubing, seamless. Society of Automo- Environment, ESTEC (pp. 1926). ESA SP-368.
tive Engineers. Bacquias, G. (1982). La formation des whiskers en bain detain acide
AMS International. (1991). Practical applications of residual stress brillant sur un support en maillechort. Oberche-Surface,23,
technology. In Conference Proceedings, Indianapolis, USA, May 194196.
1517. Ballinger, I. A., & Sims, D. (2003). Development of an EPDM
Anderson, P., & Wiktorowicz, R. (1995). Ozone emissions during arc elastomeric material for use in hydrazine propulsion systems. In
welding. Welding and Metal Fabrication Journal, 385388. 39th AIAA Propulsion Conference, Huntsville, AL, July 21. AIAA
Anderson, M. J., et al. (1995). A post-ight tribological assessment of 2003-4611.
the Hubble Space Telescope solar array mechanism. In Proceedings Banks, B. (2003). Atomic oxygen effects on spacecraft materials,
of the 6th European Space Mechanisms and Tribology Symposium NASA TM-2003-212484.
(pp. 203210). Zurich, October 46 (ESA SP-304). Banks, D. R., et al. (1990). SnPbIn solder corrosion with water
Anon. (1981). Keys to improved brazing technology. Industrial soluble ux. In ASM 3rd International Electronic Materials and
Heating Journal, 31. Processing Conference (pp. 359364). San Francisco, CA, August
Anon. (1994). EURECA, the European retrievable carrier, ESA WPP- 2023.
069. Banks, S. (1995). Reow soldering to gold. In Electronic packaging
Anon. (1995). Feasibility study of power generation system with shape and production (pp. 6972).
memory alloys. New Energy Development Organisation, Tokyo, Barber, A. H., Lu, D., & Pugno, N. M. (2015). Extreme strength
NEDO-P-9324 (in Japanese). observed in limpet teeth. Journal of the Royal Society Interface,12,
Anon. (2011). Easy leadIndustry and innovations. Materials World: 20141326.
45. Barbier, P., & Le Floch, C. (1985). HolographyThe NDT of
ANS (American National Standard). (1996). Requirements for soldered composite structures. In Proceedings of the Third European
electrical and electronic assemblies, J-STD-001B. Symposium on Spacecraft Materials in Space Environment,
Anthony, P. L., & Brown, O. M. (1965). Red plague corrosion. Noordwijk, ESA SP-232 (pp. 161165).
Materials Protection,4(3), 818. Barbier, C., et al. (1992). ATOXThe ESTEC atomic oxygen
Arblaster, J. W. (1996). Thermodynamic properties of the platinum simulation facility. Preparing for the Future,2(2), 1214.
metals. Platinum Metals Review,40, 6263. Baumgartner, R. I., & Elvin, J. D. (1995) Lifting bodyAn innovative
Ashworth, M., & Dunn, B. D. (2015). An investigation into tin whisker concept (pp. 953531). AIAA publication.
growth over a 32-year period (to be published). Beal, J. D., et al. (2006). Forming of titanium and titanium alloys (Vol.
Asma, C. O., et al. (2010). Infrared thermography measurements on 14B, pp. 656669). ASM Handbook.
ablative thermal protection systems for interplanetary space Bergendahl, C. G., & Dunn, B. D. (1984). Evaluation of test equipment
vehicles, QIRT Symposium, paper 007, Quebec, Canada, July 27 for the detection of contamination on electronic circuits. ESA STM-
30. 234.
Bernard, D. (2007). A practical guide to X-ray inspection criteria and Brooker, M. J., et al. (2001). Applying FSW to the Ariane 5 main motor
common defect analysis. Nordson-Dage Publication. thrust frame. ESA SP-468 (pp. 507511).
Bernard, D., & Blish, R. C. (2005). Considerations for minimizing Brown, B. F. (1972). Stress-corrosion cracking in high-strength steels,
radiation doses to components during X-ray inspection. In Pro- and in titanium and aluminium alloys. U.S: Naval Research
ceedings ofthe 7th Electronic Packaging Technology Conference Laboratory.
(EPIC), Singapore, December 79. Brown, B. F. (1977). Stress corrosion cracking control measures, NBS
Bernard, D., & Golubovic, D. (2012). 3D board level X-ray inspection Monograph 156, Ch. 5 (pp. 3541).
via limited angle computer tomography. In SMTA International Bruno, G., Dunn, B. D., et al. (2000). Neutron diffraction measure-
Conference, Orlando, FL, USA, October 1428. ments for the determination of residual stress in Ti6Al4V welded
Berthoud, L. (1994). Micro-impacts on Eureca solar panels. In 6th plates. Material Science Forum,347349, 674690.
International Symposium on Materials in a Space Environment, Bruno, G., & Dunn, B. D. (2004). Surface and bulk residual stress in
Noordwijk, 239248 (ESA SP 368). Ti6Al4V welded aerospace tanks. Transactions of the ASME,126,
Bester, M. H. (1968). Metallurgical aspects of soldering to gold. 284292.
Internepcon (pp. 211231). Brusciotti, F., et al. (2014). MAGNOLYA: Advanced environmentally
Beurs, J., et al. (1987). TEM-in situ deformation of beryllium single friendly chemical surface treatments for cast magnesium helicopter
crystals. Acta Metallurgical,35(9), 22772289. transmission alloys preservation. In Eurocorr 2014, Poster, Pisa,
Billot, M., et al. (1982). Ensuring solderability in electronic compo- Italy, September 812.
nents for space applications. Tin and Its Uses,131, 13. Bruzzone, P. (1987). Electrical properties of low MP alloys at 4.2 K.
Black, J. R. (1969). Electromigration failure modes. Proceedings of the Cryogenics,27, 433436.
IEEE,57, 15871594. Buckley, D. H. (1981). Surface effects in adhesion, friction, wear and
Blackman, B. R. K., et al. (2012). The fracture behaviour of adhesively- lubrication. Amsterdam: Elsevier.
bonded composite joints: Effects of rate of test and mode of loading. Buis, A., & Schijve, J. (1992). SCC of AlLi 2090-T83 in articial
International Journal of Solids and Structures,49(13), 14341452. seawater. Corrosion,48(11), 898909.
Blawert, C., & Srinivasan, P. B. (2010). Plasma electrolytic oxidation Bullettti, A., Dunn, B. D., et al. (2012). Investigation of resistivity
treatment of magnesium alloys. In H. Dong (Ed.), Surface Engi- variation of PCB laminates due to aging. IEEE Transactions on
neering of Light Alloys. Cambridge, UK: Woodhead Publishing Ltd. Components, Packaging, and Manufacturing Technology,2, 2001
Blech, I. A., et al. (1975). Whisker growth in aluminium thin lms. 2012.
Journal of Crystal Growth,32, 161169. Bulwith, R. A. (1986). The rough of grainy solder phenomenon.
Borom, M. C., & Pask, J. A. (1966). Role of adherence oxides in the Printed Circuit Assembly,1(1), 1423.
development of chemical bonding in glass seals. Journal of the Burnette, T., & Koschmeider, T. (2003). Solder joint failure analysis
American Ceramic Society,49, 1. Dye penetrant technique. Advanced Packaging, 12, 2527.
Bosma, S. J., et al. (1978). Evaluation of grounding systems. ESA STM Burrell, M. C., & Keane, J. J. (1988). Characterization of copper/
204, ESA STM 212 and ESA STM 213. enamel interfacial reactions during aging. Surface and Interface
Bosma, J. (2009). Powerhouse for Quality. ESA Bulletin, May 311. Analysis,11, 487496.
Bouilly, J.-M., et al. (2006). Ablative TPS for entry in Mars Bushmire, D. W. (1977). Resistance increases in gold aluminium
atmosphere. In 4th International Planetary Probe Workshop, interconnects with time and temperature. IEEE Transactions on
Pasadena, CA, USA, June 2730. Parts, Hybrids and Packaging,PHP-13(2), 152156.
Bouilly, J.-M., et al. (2013). Aerofast: Development of innovative Bussu, G. (2009). Stress-corrosion cracking (Ch 12, materials and
thermal protections. In 7th European Workshop on TPS and Hot safety). In G. E. Musgrave (Ed.), Safety Design for Space Systems.
Structures, Noordwijk, The Netherlands, April 810. Oxford, UK: Elsevier Ltd.
Bouseld, B. (1988). A systematic approach to sample preparation. Bussu, G., & Dunn, B. D. (2002). ESA approach to the prevention of
Metals and Materials,4(12), 758761. stress corrosion cracking in spacecraft hardware. In Joint ESA-
Bouseld, B. (1992). Surface preparation and microscopy of materials. NASA Space Flight Safety Conference. ESA-Estec, Noordwijk, The
Chichester, UK: Wiley. Netherlands.
Bouseld, B. (1997). The metallography of deformation by material Buttery, M., et al. (2011). Modern self-lubricating composites for space
removal. Praktische Metallographie,34(1), 222. applications: PGM-HT and Sintimid 15 M. In 14th European Space
Boustedt, K., & Hedemalm, P. (1995). The Nordic ip-chip project. In Mechanisms and Tribology Symposium (ESMATS 2011). Con-
Proceedings of the 1995 International Flip Chip, Ball grid, TAB and stance, Germany, September 2830.
advanced packaging Symposium (pp. 294299), San Jose, Ca., USA. Cable, N. (1988). Use of pyrotechnics on spacecraft. ESA Bulletin,54,
Boving, H. J., & Hintermann, H. E. (1987). Properties and perfor- 6671.
mances of chemical-vapour-deposited TiC-coated ball-bearing Cahn, R. W., & Haasen, P. (1996). Physical metallurgy. Netherlands:
components. Thin Solid Films,153, 253266. Elsevier Science B.V. (see Chapter 29).
Bradley, E. (2007). Lead-free electronics: iNEMI projects lead to Calle, L. M. (2014). NASAs corrosion lab at KSC: Anticipating,
successful manufacturing, pub. Hoboken, USA: Wiley. managing, and preventing corrosion. In International Workshop on
Brandes, E. A. (1992). Smithells metals reference book (7th ed.). Environmental and Alternate Energy. Kennedy Space Centre,
Oxford, UK: Butterworth-Heinemann. October 2124.
Brewer, D. H. (1970). Solders for thick gold plate. Welding Research Calow, C. A. (1974). Are metal matrix composites viable? Practical
Sup (pp. 465470). metallic composites 3. Spring Meeting, Institution of Metallurgists.
Briscoe, H. M. (1982). Proceedings of the international symposium on Campbell, J. E. (1980). Alloys for structural applications at subzero
spacecraft materials in space environment (pp. 2734). ESA SP-178. temperatures. Metals Handbook (9th ed., Vol. 3, pp. 721772). ASM.
Brizuela, M., et al. (2009). Tribolab: An experiment on space tribology, Capineri, L., Dunn, B. D., et al. (2003). Partial discharge testing of
in orbit data at the IIS. In Proceedings of the 13th European Space solder llets on PCBs in a partial vacuum: new experimental results.
Mechanisms and Tribology Symposium (ESMATS 2009). Vienna, IEEE Transactions on Electronics Packaging Manufacturing,26,
Austria, September 2325. 294305.
References 641
Case, R. K., et al. (1984). BerylliumThe lightweight contender. Coope, T. S., et al. (2014). Repeated self-healing of microvascular
Machine Design, 56, 5861. carbon bre reinforced polymer composites. Smart Materials and
Ceccanti, F., et al. (2010). 3-D printing technology for a Moon outpost Structures, 23(11), 115002.
exploiting lunar soil. 61st International Astronautical Congress Corocher, G., et al. (2009). Evaluation of different test techniques to
(pp. 19). Prague, CZ (IAC-10-D3.3.5). assess the effect of environmental exposure on the assembly of
Chafn, R. J. (1981). Migration of silver from silver-loaded polyimide MCGA packages for printed circuit boards. 11th ISMSE Interna-
adhesive chip bonds. In IEEE Transactions on C.H.M. Technology tional Symposium, Aix-en-Provence, France, September 1518.
(Vol. 4, No. 2, pp. 214216). June. Cornec, L., & Clariou, J.-P. (1994). Carbonmagnesium MMC for
Chaillot, A., et al. (2014). ENEPIG nish: An alternative solution for light-weight aerospace structures. In Proceedings of the Interna-
printed circuit boards. In Electronic Materials and Processes tional Symposium on Advanced Materials, ESTEC, Noordwijk
Workshop (EMPS-5). Estec, Noordwijk, The Netherlands, May (pp. 563567). ESA WPP-070.
2024. Craig, B. (1994). Lead-free solder: An update. Electronic Production,
Charrier, I., Dunn, B. D., et al. (2003). Moisture pick-up and minimum 2125.
drying conditions of epoxy and polyimide-glass printed circuit Craven, E. (1993). SolderabilityThe effect on quality manufacturing.
boards. In Brasage/Soldering, International Conference. Brest, Electronic Production, 3335.
France, October 810. Damskey, R. G. (1988). A satellite structureManagers look at
Chelladurai, T., et al. (1995). Acoustic emission response of 18 % Ni beryllium. The Beryllium Users Symposium,Europe.
maraging steel weldment with inserted cracks of varying depth to Danford, M. D. (1994). The corrosion protection of several aluminium
thickness ratio. Materials Evaluation,53(6), 742746. alloys by chromic acid and sulphuric acid anodizing. NASA
Chen, Y., et al. (2014). A comprehensive investigation of the galvanic Technical Paper 3490.
corrosion induced Ag-Al bond degradation in microelectronic de Salazar, Gomez, et al. (1988). Diffusion bonding of aluminium
packaging using argon ion milling, SEM, dual beam FIB-SEM, alloys using silver coatings. Practice Metals,25, 532542.
STEM-EDS and TOF-SIMS. 40th ISTFA, Houston, TX, USA, de Rooij, A. (1985). The degradation of metal surfaces by atomic
November 913. oxygen. In Proceedings of 3rd European Symposium on Spacecraft
Chenard, S. (1990). Space: Dirty and dangerous? Space Markets,5, Materials in Space Environment (p. 99). ESA SP-232, October.
253256. de Rooij, A. (1989). Bimetallic compatible couples. ESA Jnl.,13, 199
Choi, S. J., et al. (1996). A Nitinol-based solar array deployment 209.
mechanism. In 30th Aerospace Mechanisms Symposium, Hampton, de Rooij, A. (1989b). An approach to long term prediction of the
Va. (103118). NASA Conf. Pub. No. 3328. May. atomic oxygen effect on materials (to be published).
Chollet, L., et al. (1985). Journal of Mater Energy System,6(4), 293299. de Rooij, A. (1995a). Atomic oxygen effects on HST-SA1 metals. In
Chudnovsky, B. H. (2002). Degradation of power contacts in industrial Hubble Space Telescope Solar Array Workshop (pp. 341353),
atmospheres: silver corrosion and whiskers. In Proceedings of the Noordwijk, May.
48th IEEE Holm Conference on Electrical Contacts, Orlando, USA. de Rooij, A. (1995b). Space exposure of HST-SA1 tagboards. Hubble
Chuvilin, A., et al. (2014). Using TEM to simulate and understand the Space Telescope Solar Array Workshop, ESTEC (pp. 449462).
formation of defects in graphine. In Microscopy and Analysis, ESA WPP-77.
November/December 710. de Rooij, A., & Collins, D. C. (1995). SLAM investigations on HST-
Cleave, J. F., & Morton, J. R. (1989). The effects of impurity levels of SA1 solar cell welds. In Hubble Space Telescope Workshop
lithium on aluminium casting alloys. The Foundryman,82, 549592. Results from Post-ight Investigations, ESTEC, Noordwijk, May
Clemens, H., & Schretter, P. (1996). Microstructure and properties of 3031. ESA WPP-77.
Gamma-TiAl based alloys. Practical Metallogarphy Journal,33(1), de Rooij, A. (2010). Exposure of silver to atomic oxygen. In
1735. EUROCORR 2010, Moscow, Russia, September 1317.
Clements, H., et al. (1996). Intermetallic gamma titanium aluminides De Rooij, N., et al. (2009). MEMS for Space. IEEE Transducers
Potential candidates for spacecraft structures. ESA Conference on Conference (pp. 1724). Denver, CO, USA, June 2125.
Spacecraft Structures, Materials and Mechanical Testing, Noord- Dauphin, J. (1984). Selection and control of materials for space
wijk, March 2729. applications. ESA Journal,8(1), 53.
Coelho, A. (2009). Comeback for cork. Materials World, 17, 24. Dauphin, J. (1986). Space materials in Europe. In Proceedings of the
Cohen, A. (1995). Copper and copper alloy wires, cables and SAMPE Conference (pp. 276290), Las Vegas.
conductors. ASTM Standardization News, 2327. Dauphin, J. (1987). In J. de Bossu (ed.), Atomic oxygenA low orbit
Cole, L. S., et al. (2004). Some recent trends in corrosion science and plague. Looking ahead for materials and processes (pp. 345367).
their application to conservation. In Proceedings of Metal (National Amsterdam: Elsevier Science Publishers B.V.
Museum of Australia, Canberra), October 216. Dauphin, J. (1992). Touchstone for success: The materials and
Cole, P. T., & Thomas, N. (1989). Ariane IV Spelda-Flight proving by processes laboratory, ESA BR-52.
acoustic emission. In Proceedings of AECM-3 (pp. 144150). Paris, Dauphin, J. (1993). Radiation effects in materials. Radiation Physics
July 1721. and Chemistry,43, 4756.
Colla, V., et al. (2014). Monitoring concepts for a 3-D printer applied to Dauphin, J., Dunn, B. D., Judd, M. D., & Levadou, F. (1991). Materials
build a human outpost on the Moon. In UKSim-AMSS 16th for space application. Metals and Materials, 422430.
International Conference on Computer Modelling and Simulation David, L. (1995). Radarsat built to resist debris bumps. Space News.
(pp. 205210). David, L. (1996). Incredible shrinking spacecraft. Aerospace America,
Comiskey, J. M. (1979). Materials that make wire for spacecraft use. In 2024.
Aerospace Corporation,Proceedings of Symposium on Wire Appli- Davy, J. G., & Skold, R. (1985). Computer-aided solderability testing
cations in Spacecraft, December. for receiving inspection. Circuit World, 3441.
Cooke, R. W. (2010). Red plague control plan (RPCP), the control of DeBold, T. A., & Kosa, T. (2004). Pasivating stainless stell parts.
cuprous/cupric oxide corrosion. http://ntrs.nasa.gov/archive/nasa/- NASA Technical Briefs, 5152.
casi.ntrs.nasa.gov/20100009723.pdf. Degisher, H. P., & Simancik, F. (1995). Recyclable foamed aluminium
as an alternative to composites. In H. Warlimont (Ed.),
642 References
Environmental Aspects in Materials Research (pp. 137140). DGM Dunn, B. D. (1982a). Evaluation of various solar cell-to-interconnector
Informations gesellschaft. welds by means of scanning laser acoustic microscopy and
Delahais, M. L. L. (1985). Materials and processes in scientic metallography. ESA STM-225, August.
spacecraft. In Proceedings of the 3rd European Symposium in Dunn, B. D. (1982b). Evaluation of various solar cell to interconnector
Spacecraft Materials in Space Environment. ESA SP-232, October welds by SLAM and metallography. ESA STM-225.
57. Dunn, B. D. (1984). The corrosion properties of Spacelab structural
Deloo, P., et al. (1995). HST Solar Array 1 Dynamic Performance. HST alloy aluminium 2219-T851, ESA STR-212.
Solar Array Workshop (pp. 89100). ESTEC, ESA WPP-77, May Dunn, B. D. (1987). A laboratory study of tin whisker growth. ESA
3031. STR-223.
De Marco, M., et al. (2015a). Introduzione ai modi di frattura ed Dunn, B. D. (1988). Mechanical and electrical characteristics of tin
allanalisi fracttograca. Parte 1. Rev. It. Della Saldatura, 1, 95106. whiskers. ESA Journal,11 and 12, 117.
De Marco, M., et al. (2015b). Introduzione ai modi di frattura ed Dunn, B. D. (2001). Innovative Materials and Processes. Space
allanalisi fracttograca. Parte 2. Rev. It. Della Saldatura, 2, 229242. Technology Innovation Workshop (pp. 21134). Copenhagen.
Dittes, M., et al. (2003). The effect of temperature cycling on tin ESA WPP-197, September 67.
whisker formation. In Proceedings of the Electronics Packaging Dunn, B. D. (1993). Metallurgical evaluation of steam aged LCCC
Technology Conference (pp. 183188). devices following solderability testing. Soldering and SMT,15,
Didziulis, S. V. (1994). Detergent cleaning of bearings for spacecraft 2632.
applications. In Non-Ozone Depleting Chemical Cleaning and Dunn, B. D. (2008). Workmanship standards and their application on
Lubrication of Space System Mechanical Components for Multi- ESA projects. Soldering and SMT Journal,20(4), 3744.
Year Operations, EPA Workshop, Denver, Co, September 2627. Dunn, B. D. (2008). ESA-approved skills training schoolsElectronic
Dohle, G. R., et al. (1996). A new bonding technique for microwave assembly techniques, ESA-STR-258, November.
devices. IEEE Transactions on Components Hybrids and Manu- Dunn, B. D. (2012). Guidelines for a creating a lead-free control plan.
facturing Technology,19(1), 5763. ESA STM-281, October.
Dorsey, J. T., et al. (2012). An efcient and versatile means for Dunn, B. D., & Bergendahl, C. G. (1987). An evaluation of the solvent
assembling and manufacturing systems in space. In AIAA Space extraction method for the detection of ionic contamination on
Conference and Exposition, Pasadena, CA, USA, September 1113. substrates supporting large surface mounted devices. Brazing and
Downs, K. S., & Hamstad, M. A. (1995). High sensor density A.E. Soldering Journal, 2027, 32.
monitoring of graphite/epoxy pressure vessels. In The American Dunn, B. D., & Chandler, C. (1980). The corrosive effect of soldering
Society for Nondestructive Testing, Proceedings of AECM-4 uxes and handling on some electronic materials. Welding
(pp. 1322). Sundsvall, July 1014. Journal,59, 289307.
Doyle, D. P., et al. (1963). A rapid method for determining the Dunn, B. D., & Collins, D. S. (1978). Materials investigation for a
corrosivity of the atmosphere at any location. Nature,200(4912), failed spacecraft antenna. ESA Journal,2, 223235.
11671168. Dunn, B. D., & Desplat, P. (1994). Evaluation of conformal coatings
Doyle, E., & Hubbard, P. (2010). Innovation in nitriding. International for future spacecraft applications. ESA SP-1173, August.
Heat Treatment and Surface Engineering,4(1), 3339. Dunn, B. D., & Nicholas, D. (1998). Elemental distribution of parts and
Dricot, F., et al. (1994). Programme for the evaluation of materials used materials making up an electronic box. Circuit World,24(2), 2123.
in low Earth orbit and manned spacecrafts. In 6th International Dunn, B. D., & Mozdzen, G. (2014). Tin oxide coverage on tin whisker
Symposium on Materials in a Space Environment (pp. 293297). surfaces, measurements and implications for electronic circuits.
Noordwijk, (ESA SP-368). Soldering and SMT Jnl.,26(3), 139146.
Drolshagen, G. (1994). Eureca meteoroid/debris analysis. In Eureca Dunn, B. D., & Papenburg, H. (1986). Material investigation into the
Technical Report (pp. 507512). ESA WPP-069. deactivation of catalyst particles for hydrazine decomposition. In
Drolshagen, G., et al. (1996). Optical survey of micrometeoroid and Proceedings of the Fifteenth Inernational Symposium on Space
space debris impact features on Eureca. Planetary Space Science,44 Technology and Science, Tokyo.
(4), 317340. Dunn, B. D., & Stanyon, R. (1997). An Inspection of Spacelab
Drogoul, S. (2005). From the failure to the successThe return to Hardware. ESA STR-241, September.
ight of the Ariane 5 ECA launcher. In 56th International Dunn, B. D., & Steinz, J. A. (1974). Contamination and heating effects
Astronautical Congress, Session D2.2, IAC-05-D2.2.08, Fukuoka, of a hydrazine thruster intended for use on-board the scientic
Japan. geostationary satellite. ESRO (ESA) TM-152, April.
Dunkerton, S., Dunn, B. D., & Fernie, W. B. (2001). Welding and Dunn, B. D., et al. (1983). Evaluation of the temperatures attained by
joining in a space environment. Revista Italiana della Saldatura, electronic components during various manual soldering operations.
Anno LIII,5, 615627. Brazing and Soldering Journal,5, 4553.
Dunn, B. D. (1973). European team visit TRW Systems to learn Dunn, B. D., et al. (1984). Corrosion of silver plated copper conductors.
techniques for the future. TRW Sentinel (p. 6), November 9. ESA Journal,8, 307335.
Dunn, B. D. (1975). The properties of near-eutectic tin/lead solder Durin, C., et al. (1993). System results from FRECOPA. In 3rd Post
alloys tested between +70 and 60 C and the use of such alloys in Retrieval Conference (pp. 11531165). Williamsburg, November
spacecraft electronics, ESA TM-162, September. 812, NASA Pub. 3275 part 3.
Dunn, B. D. (1977). Metallurgical evaluation of tinlead nishes on Dzhanibekov, V. A., et al. (1991). Welding equipment for space
PCBs. Transactions of the Institute of Metal Finishing,55, 8692. applications. In Proceedings of the Welding in Space and the
Dunn, B. D. (1978). Examination of new types of solder-splices. ESA Construction of Space Vehicles by Welding, sponsored by
TM-210, September. American Welding Society, USA and E O Paton Welding Institute,
Dunn, B. D. (1979). The resistance of space quality solder joints to Ukraine, September 1991.
thermal fatigue. Circuit World part 1 5(4), 1117; part 2 6(1), 1627. Ecord. G. M. (1972). Apollo Experience reportPressure vessels.
Dunn, B. D. (1980). The Fusing of tinlead plating on high quality NASA TN-D-6975, September.
printed circuit boards. Transactions of the Institute of Metal ECSS-Q-ST-70. (2014). Materials, mechanical parts and process
Finishing,58, 2628. Space product assurance requirements.
References 643
Ehmann, D. (1994). Space structures of high precision. Dornier Post,1, Felton, L. E., et al. (1991). Cu-Sn precipitates in solder joints. Scripta
1517. Metallurgica et Materialia,25, 23292333.
Elking, M. J., & Hughes, H. E. (1969). Prevention of stress corrosion Feng, W., et al. (2013). Electro-migration model parameters sensitivity
failure in semiconductor leads. Bell Telephone Report in: Physics of analysis based on Monte Carlo method. Chemical Engineering
Failure in Electronics, 5, 447495. Transactions,33, 10931098.
Ellis, B. N. (1996). The correlation between short- and long-term SIR Fernie, J. A., & Sturgeon, A. (1992), Joining ceramic materials. Metals
testing. Circuit World,22(2), 4950. and Materials, 212217
Ellis, M. B. D. (1996). Fusion welding of aluminium-lithium alloys. Fink, M., & Dunn, B. D., et al. (2006). Characterisation of Teon FEP
Welding and Metal Fabrication Journal. 5560. (Hubble ST and LDEF) following long term exposure to LEO. In 11th
Ellis, M. B. D., & Gittos, M. F. (1995). Tungsten inert gas welding International Symposium on ISMSE, Collioure, France, June 1923.
of titanium and its alloys. Welding and Metal Fabrication Journal, Fink, M., Dunn, B. D., et al. (2008). Measurement of mechanical
912. properties of electronic materials down to 4.2K. Cryogenics
Ely, J. K. (1942). Method of making electrical contact members. U.S. Journal,48, 497510.
patent 2295338 of 8 September. Fink, M., Semerad, E., & Dunn, B. D., et al. (2009). Electrical
Engel, L., & Klingele, H. (1981). An Atlas of Metal Damage. London: resistivity of solders at low temperatures. In 11th International
Carl Hanser Verlag and Wolfe Publishing Ltd. Symposium on ISMSE, Aix-en-Provence, France, September 1516.
Engelmaier, W. (1995). Fatigue life of leadless chip carrier solder joints Fink, M., et al. (2009). In Proceedings of the 13th European Space
during power cycling. In Solder Joint Reliability Workshop Notes, Mechanisms and Tribology Symposium, Vienna, September 2325.
Engelmaier Associates Inc., Mendham, NJ, USA (see also IEEE Fletcher, P. (1978). Problems encountered with the use of semi-rigid
Transactions Components, Hybrids and Manufacturing Technol- co-axial cable in spacecraft electronics, ESA-STM-209, September.
ogy6(3), 232237, 1983). Flin, E. D. (2005). Soldering in space holds surprises, Aerospace
Engemann, J. (1996). Chancen und moglichkeiten der plasmatechnolo- America, 2425.
gie: eine bestandsaufnahme. Vakuum in Forschung und Praxis,1, Flom, Y. (2005). Brazing in spaceEnabling the next frontier. In
2530. TheAWS 35th International Brazing and Soldering Symposium,
EPAUS Environmental Protection Agency. (1994). Replacing solvent Dallas, TX, USA, April 26.
cleaning with aqueous cleaning, EPA-600/R-94-131. Flom, Y. (2006). Electron beam brazing of titanium for construction in
Erdman, N., Campbell, R., & Asahina, S. (2006). Argon beam cross space, NASA GSFC. In Proceedings of the 3rd International Brazing
sectioning. Advance Materials and Processes, 3335. and Soldering Conference, San Antonio, TX, USA, April 2426.
ECSS-E-ST-32-08. (2013). Space engineering: Materials, draft, Flury, W. (Eds.). (1993). First European conference on space debris.
December. ESA SD-01.
ECSS-Q-ST selected ECSS standards related to materials and processes Foster, C. L., et al. (1995). The solar array-induced disturbance of the
can be seen listed in Appendix 9. HST pointing system. NASA Technical Paper3556, May.
ESA PSS-01-401. (1994). EAS fracture control requirements. Fox, V. C., et al. (1999). The structure of tribologically improved
ESA PSS-01-701. (1994). Data for selection of space materials. Issue 1, MoS2Metal composite coatings and their industrial applications.
rev. 3. Surface and Coatings Technology,116119, 492497.
ESA PSS-01-703 A method for black anodising aluminium with Fox, R. M., et al. (2010). Chemistry and technology of lubricants, Ch.
inorganic dyes. 13, Liquid lubricants for space applications. Heidelberg: Springer
ESA PSS-01-708. The manual soldering of high-reliability electronics. Publishing.
ESA PSS-01-721. Application and testing of Cuvertin 306, a black Frankel, H. E. (1969). Effect of vacuum on materials, ESRO TN-77.
paint for space use. Frear, D. R., & Vianco, P. T. (1994). Intermetallic growth rate and
ESA PSS-01-730. (1981). The wire-wrapping of high-reliability mechanical behaviour of low and high melting temperature solder
electrical connections. alloys. Metallurgical and Materials Transactions A,25, 15091523.
ESA PSS-01-736. (1981). Material selection for controlling stress- Frear, D. R., et al. (1993). The mechanics of solder alloy interconnects.
corrosion cracking of ESA spacecraft and associated equipment, May. New York: Van Nostrand Reinhold.
ESA PSS-01-737. (1982). Test method for determination of the stress- Gabe, D. R. (1978). Principles of metal surface treatment and
corrosion susceptibility of spacecraft structural alloys. protection. Oxford: Pergamon Press.
ESA PSS-01-738 (Issue 1). (1991). High-reliability soldering for Galleo, K. (1995). Assembly with conductive adhesives. IPC Technical
surface-mount and mixed technology printed. Review, 2028.
ESA PSS-01-746. (1993). General requirements for threaded fasteners. Galvao, I., et al. (2011). Formation and distribution of brittle structures
E. S. K. (1995). Engineered ceramics. Paper from Elektroschmelzwerk in FSW of aluminium and copper: Inuence of process parameters.
Kempten GmbH, Munich, Germany. Science and Technology of Welding and Joining,16(8), 681689.
Evans, R. (1995). Fume extraction. Electronic Production, 1416. Gamwell, W. R. (1995). NASA-wide fastener technical interchange
Eyre, T. S. (1991). Friction and wear control in industry. Metals and meeting. NASA Conference Publication 3308, July.
Materials, 143148. Garner, H. R., et al. (1976). Whisker growth from decorative zinc deposits.
Fang, D., & Berkovits, A. (1995). Fatigue design model based on In Proceedings of the 9th World Congress on Metal Finishing.
damage mechanisms revealed by AE. Trans ASME,117, 200208. Garrison, A., et al. (1995). How much rework is too much. Soldering
Favre, J.-P., & Raud, C. (1995). Acoustic emission to monitor crack and SMT Journal,19, 4851.
accumulation in crossply CFRP under mechanical loading or thermal George, E., & Pecht, M. (2013). Tin whisker analysis of an automobile
cycling. In The American Society for Nondestructive Testing, engine control unit, microelectronics reliability. http://dx.doi.org/
Proceedings of AECM-4 (pp. 3342). Sundsvall, July 1014. 10.1016/j.microrel.2013.07.134.
Feest, A. (1994). Interfacial phenomena in metal-matrix composites. Gerlach, L. (1995). Co-ordinator. In Proceedings of the Hubble Space
Composites Journal,25(2), 7586. Telescope Solar Array Workshop, ESA WPP-077.
Fellas, C. (1982). Anti-static coat for solar arrays. In Proceedings of the Ghate, P. B. (1982). Electromigration-induced failures. In IEEE 20th
3rd European Symposium on Photovoltaic Generators in Space, Annual Proceedings, Reliability Physics (pp. 292299). San Diego,
Bath, UK (p. 305). MarchApril.
644 References
Ghaffarian, R. (2013). 3D X-ray CT for BGA/CGA workmanship defect Hagge, J. K., & Davis, G. J. (1984). Ageing, solder thickness, and
detection. NASA JPL Publication 12-21 12/12. solder alloy effects on circuit board solderability. IPC Technique
Ghidini, T., et al. (2008). Stress-corrosion cracking assessment in Review,25(9), 1323.
spacecraft propulsion systems, Technology Evaluation for Envi- Hagge, J. K., & Davis, G. J. (1985). Ageing, solder thickness and solder
ronmental Risk Mitigation (TEERM) Workshop, San Diego, CA, effects on circuit board solderability. Circuit World,11(3), 815.
USA, November 1820. Haines, J. E., et al. (1995). HST solar array drive electronics: In-ight
Gilbrech, J., et al. (2005). Cloud-aerosol LIDAR and IR Pathnder history and servicing mission activities. In HST Solar Array
satellite observation (CALIPSO) spacecraftIndependent technical Workshop, ESTEC, (ESA WPP-77) (pp. 111118).
assessment, NASA/TM-2005-213231/Version 2.0, March. Hall, W., & Williams, P. (2007). Processing waste pc boards for
Gobbi, S. L., & Zhang, L. (1994). Laser welding for lightweight material recovery. Circuit World,33, 4350.
structures. In International Symposium on Advanced Materials for Halpin, J. C. (1985). Proceedings of the 31st Annual Technical Meeting
Lightweight Structures, ESA-ESTEC (pp. 107113). 2224 March. of the Meeting of the Institute of Environmental Sciences (pp. 206
ESA WPP-070. 218). Las Vegas, April 30May 2.
Goddard, D. M. (1984). Report on graphite/magnesium castings. Metal Hamacher, H., et al. (1995). Surface erosion in low Earth orbit. In
Progress, 4952. ATOXSymposium on Scientic Results on the German Spacelab
Golberg, D., et al. (1995). High-temperature shape memory effect. Mission. D-2, Norderney: WPF publication.
Materials Letters,22, 241248. Hammesfahr, A. E., et al. (1995). An X-ray mirror assembly based on
Goldberg, I. B., et al. (1990). Magnetic susceptibility of Inconel alloys carbon bre technology. Space Technology,14(6), 391397.
at cryogenic temperature. Advances in Cryogenic Engineering Hamstad, M. A. (1986). A review: Acoustic emission, a tool for
(Materials) Journal,36, 755761. composite-material studies. Experimental Mechanics, March 713.
Goldmann, L. S., et al. (1977). Leadindium for controlled-collapse Hannech, E. B., & Hall, C. R. (1992). Diffusion controlled reactions in
chip joining. In Electronic Components Conference (pp. 2529). gold/solder system. Materials Science and Technology,8, 817824.
Goldstein, J. I., et al. (1981). Scanning electron microscopy and X-ray Hansen, M. (1958). Constitution of binary alloys. New York: McGraw-
microanalysis. New York: Plenum Press. Hill.
Gonnet, L. (1995). Towards an ergonomic design of welding stations. Haq Qureshi, A., & Dayton, J. A. (1995). Insulation requirements of
Rivista Italiana della Saldatura,2, 177183. high-voltage power systems in future spacecraft. NASA Technical
Goodhew, P. J. (1988). Modular MSc degrees: A way of keeping up Paper No.3520, June.
with materials and techniques. Metals and Materials,4(2), 112. Harris, P. G., et al. (1992). Stripping and replacement of damaged
Grady, M. M., & Wright, I. P. (1996). Space invaders. European solderable coatings. Soldering and SMT Journal,12, 2126.
Microscopy and Analysis Journal 42, June 1314. Harris, P. G. (1995). Conductive adhesives a critical review of progress
Grande, D. H., et al. (1988). Fibrematrix bond strength studies. to date. Soldering and SMT Journal,20, 1921.
Journal of Materials Science,23, 311328. Haschiguchi, D., et al. (1992). Aluminiumberyllium alloys for
Grant, L. A. (1983) Beryllium sheet materials to satellite structure. In aerospace applications. In Proceedings of the International Sympo-
24th Annual AIAA SDM Conference, Lake Tahoe, May. sium Advanced Materials for Lightweight Structures (pp. 165169).
Greenberg, M. A., et al. (2013). Inhomogenieties of the interface ESTEC, March 2527, ESA SP-336.
produced by explosive welding. Physics of Metals and Metallog- Hashemzadeh, M. N. (2005). Study of tin whisker growth and their
raphy,113(2), 176189. mechanical and electrical properties (PhD thesis, Linkopings
Gregoratti, L., et al. (2013). Large gecko mimetic tapes as new joining Universitty, Stockholm, 2005), September.
technology. In Proceedings of the 12th International Symposium on Hashimoto, K., et al. (1991). Flip chip interconnection technology for
Materials in Space Environment, Noordwijk, The Netherlands GLSI packages operated in liquid nitrogen. IEICE Transactions,
(ESA SP-705). E74, 23622368.
Grey, M. W. (1989). Inner layer or post cracking on multilayer printed Hassan, J., et al. (2014). Recent advancements in focused ion beam
circuit boards. Circuit World,15(2), 2229. applications: A review. International Journal of Recent Scientic
Grilli, R., Dunn, B. D., et al. (2010). Locallised corrosion of Research,5, 123127.
Aluminium 2219 alloy exposed to a 3.5 %NaCl solution. Corrosion Hauer, H. (1977). Properties of soft soldered joints on gold plated thin
Science,52, 28552866. lm conductors. In International Conference on Thin Film Tech-
Grilli, R., Dunn, B. D., et al. (2010). Localized corrosion on 2219 nology (Vol. 60, pp. 185190). NTG-Facher.
aluminium alloy coated with a titanium based conversion coating. Haupt, C. W. (1996). Residual stress measurement of titanium weld
Surface and Interface Analysis Journal,42(67), 610615. certication rings. In 32nd AIAA Joint Propulsion Conference
Grinberg, N. M. (1982). The effect of a vacuum on fatigue crack (p. 10). Lake Buena Vista, FL, paper AIAAA-96-2750, July 13.
growth. International Journal of Fatigue, April 8395. Hawkins, T.W., et al. (2010). Reduced toxicity, high performance
Griner, C. S. (1968). Redeposition of vacuum outgassed products. monopropellant at the US air force laboratory. In 4th International
NASA TM X-53 801, MSFC, November. Association for the Space Safety Conference, Huntsville, AL, May.
Gross, J. H. (2011). Mass spectroscopy. Heidelberg: Springer. Healy, T. J., et al. (1995). Key technologies being developed for RLV
Gruenbaum, P., & Dursch, H. (1993). Space environment effects on (pp. 19953608). AIAA paper.
solar cells: LDEF and other ight test. In 3rd Post Retrieval Hegin, B. V. (n.d). Electroplating and black anodisation of aluminium.
Conference (pp. 11671177). Williamsburg, November 812, Galvanische Industrie, Heerde, Holland (trade leaet).
NASA Publication 3275 part 3. Hehemann, P. F. (1985). Stress corrosion cracking of stainless steels.
Gruhl, W. (1984). Stress corrosion cracking of high strength aluminium Metallurgy Transactions A,16A, 19091923.
alloys. Z. Metallkde,75, 819826. Heiken, G. H., et al. (1974). Lunar deposits of possible pyroclastic
Gurklis, J. A. (1972). Non-traditional machining of beryllium. Metals origin. Geochimica Cosmochimica Acta,38, 17031718.
and Ceramics Information Center, Battelle, MCIC, January 3. Hemptenmacher, J., et al. (1994). Characteristics of SiCFibre/Ti6A14
Guyenne, T.-D. (1994). Compiler. In 6th International Symposium on VMatrix composites by macro-indentation tests. Practical
Materials in a Space Environment, ESA SP-368. Metals,31(3), 110119.
References 645
Hendrich, CH., et al. (2014). Inuence of water content in an ADN- Jagt, van der, et al. (1991). Anti-wear/corrosion treatment for austenitic
based liquid monopropellant on performance characteristics. In steel: The Hardcor process. Materials & Design,12(1), 4146.
Space propulsion, May 1922. Germany: Cologne. IPC-7095A. (2004). Design and assembly process implementation for
Hepp, A. F., et al. (1999). Metal alloy propellants produced from lunar BGAs, October.
resources. In Symposium Materials in SpaceScience, Technology IPC-D-279. (1996). Design guidelines for reliable surface mount
and Exploration (pp. 3947). Boston, MA, USA, 29th November technology printed board assemblies, July.
2nd December. IPC-SM-780. (1987). Advanced packaging and interconnection of
Heltzel, S. (2014). Latent short circuits failure in high-rel pcbs caused electronic components.
by lack of cleanliness of pcb processes and base materials, 5th IPC-SM-782. (1992). Surface mount land patterns (conguration and
EMPS Workshop, Estec, Noordwijk, The Netherlands, May 22. design rules).
Hinton, Y. L. (1995). Acoustic emission. ASTM Standardization News, IPC-TM-650. (2000). Electrochemical migration resistance test.
3639. Irving, B. (1992). SCC: Weldings No. 1 nemesis. Welding Journal,
Hipolite, W. (2015). NASA 3D prints copper rocket engine part. 3D 3740.
Printing, April 22. Isakowitz, S. J. (1995). International reference guide to space launch
Heycock, C. T., & Neville, F. H. (1900). Philosophical Transactions of systems (2nd ed.). Waldorf, MD, USA: AIAA publication.
the Royal Society, 194A: 201 Jackson, B. C. (1987). Oxide coatings for bonding multilayer PCBs.
Hillman, D. D., & Chumbley, L. S. (2006). Characterization of tin Transactions IMF,65(8), 89.
oxidation products using sequential electrochemical reduction Jackson, C. R. (1973). Leadindium solder alloy wont embrittle in
analysis (SERA). Soldering and SMT Journal,8(3), 3141. gold. Circuit Manufacturing, 4041.
Hinch, S. W. (1988). Hand book of surface mount technology. Harlow, Jacobson, D. M., & Humpston, G. (1992). Diffusion soldering.
England: Longman Scientic and Technical publishers. Soldering and SMT Journal,10, 2732.
Hirnyj, S., & Indacochea, J. E. (2008). Phase transformations in Jacobson, D., & Humpston, G. (1995). Depressing the melting point of
Ag70.5Cu26.5Ti3 ller alloy during brazing processes. Chemistry solders and brazes by eutectic alloys. GEC Journal of Research,12
Metals Alloys,1, 323332. (2), 112121.
Holt, T. (1995). New applications in high power laser welding. Welding Jacobson, D. M., & Sangha, S. (1996). Novel applications of diffusion
and Metal Fabrication Journal, 230234. soldering. Soldering and SMT Journal,23, 1215.
Homes-Siedle, A., & Adams, L. (1993). Handbook of radiation effects. Jaggers, C. H. (1993). Exposure of LDEF materials to A.O. In 3rd
Oxford, UK: Oxford University Press. Post-Retrieval Symposium (pp. 931941). Williamsburg, November
Hornbogen, E., & Schemme, K. (1990). Prospects of superlight metals 812.
and their laser surface alloying with ceramic phases. In Proceedings Jalbert, M., et al. (1994). Risque saturnin des oprations de
of the First Thermal Structures Conference (pp. 99109). Char- microsoudure en lectronique. Archives des maladies profession-
lottesville, VA, November 1416. nelles et de mdecine du travail,55(8), 589594.
Houlberg, K. (1988). Printed circuit boards for surface mounted Janu, P., et al. (2009). Electric propulsion pointing hold down and
devices. Unpublished report. release mechanism test results. In Proceedings of the 13th European
Hu, S. J., et al. (1995). Gold wire weakening in the thermosonic Space Mechanisms and Tribology Symposium (p. 8)., Vienna,
bonding of the rst bond. IEEE Transactions, Components, Austria, September 2325
Hybrids, Manufacturing Technology,18(1), 228234. Jaworske, D. A. (2002). Solar selective coating for power application,
Huang, C.-Y., et al. (2011). Material characterization and failure NASA/TM-2002-211333.
analysis for microelectronics assembly processes. Available from: JAXA-QTS selected JAXA standards related to materials and processes
http://www.intechopen.com /books/wide-spectra-of-quality-con- can be seen listed in Appendix 9.
trol/material-characterization-and-failureanalysis-for-microelectron- Jellison, J. (1986). Evaluation of multilayer printed wiring boards by
ics-assembly-processes. metallographic techniques. NASA Reference Publication 1161.
Hummel, R. E. (1994). Electromigration and related IC failure Jenkins, P. P., & Landis, G. A. (1995). A rotating arm using shape-
mechanisms. International Materials Review,39, 97113. memory alloy. In 29th Aerospace Mechanism Symposium, League
Hunt, C. (2009). Implementation and reliability issues with lead-free City, Texas (pp. 167171). May 1719.
solders. In SMTA International Conference and AIMS Workshop, JESD22-A121A. (2008). Measuring whisker growth on tin and tin
San Diego, CA, USA, October 48. alloy surface nishes, reafrmed, May 2014.
Hunt, C. (2011). Reliability isues with lead-free, EMPS-2 Workshop, Jha, A. K., et al. (1996). Failure of AA 2024 aluminium alloy rivets.
Institut Du Soudure, Paris, France, May 4. Praktische Metallographie,33(5), 264272.
Hunt, C., Dunn, B. D., et al. (2012). Increased shorting in tin whiskers due Jiang, B. L., & Wang, Y. M. (2010). Plasma electronic treatment of
to electric elds and contact pressure. In 6th International Symposium aluminium and titanium alloys. In H. Dong (Ed.), Surface engineering
on Tin Whiskers, Loughborough, UK, November 2728. of light alloys. Cambridge, UK: Woodhead Publishing Ltd.
Hutchings, M., & Krawitz, A. D. (1992). Measurements of residual and Johnson, G. W. (1979). How metallography and SEM analysis can be
applied stress by neutron diffraction. In NATO ASI series E: Applied used in quality control for microelectronics. In Metallography as a
Science (Vol. 216). Dordrecht, The Netherlands: Kluwer. Quality Control Tool, Symposium of A.S.M. (pp. 261278). July 89.
Hutchinson, B., et al. (2004). Whisker growth from tin coatings, IM- Johnson, M.R. (1989). The Galileo High Gain Antenna Deployment
2004-708. Stockholm: Swedish Institute for Metal Research. Anomaly. California Institute of Technology, Jet Propulsion Lab-
Hwang, J. S. (2014). Tin whiskers: capsulization. SMT Magazine, 1629. oratory (JPL), Pasadena, CA.
Ilyushenko, R. V. (1993). Weldability of commercial aluminium Johnson, R.E., et al. (1966). A case history of titanium stress corrosion
lithium alloys. Aluminium-Verlag,69(4), 364370. in nitrogen tetroxide. American Society Metals
Iman, R. L., et al. (1995). Evaluation of low-residue soldering for military Johnson, R. E. (1973). Apollo experience reportThe problem of stress
and commercial applicatons: Sandia National Labs Report, June. corrosion cracking. NASA TN D-7111, March.
IEC 60068-2-82. (2007). Environmental testingWhisker test methods Johnston, R. L., et al. (1967). Stress corrosion cracking of Ti6Al4 V
for electronic and electric components. alloy in methanol. NASA TN D-3868, February.
646 References
Kallee, S. W., Nicholes, D., & Thomas W. (2001). Friction stir welding Korb, L. J., et al. (1985). A metallurgical analysis of the failure of two
Invention, innovations and applications. In INALCO 8th Inter- APUs of the spacecraft Columbia. In Space Technology Confer-
national Conference on Joints in Aluminium, Munich, Germany, ence Proceedings, Anaheim, USA, September 2325, 5.15.34.
March 2830. Kostic, A. D. (2011). Lead-free electronics reliabilityAnd update,
Karavakis, K., & Bertling, S. (2004). Conductive anodic lament GEOINT Development Ofce, Aerospace Corporation, August.
(CAF). MEPTEC Report, Quarter 4, 2427. Kleiman, J., Tagawa, M., & Kimoto, Y. (2013). Protection of materials
Kauffmann, B., & Wolf, D. (2013). Development of low uoride and structures from the space environment. Heidelberg: Springer.
evolution uoropolymer wiring materials for aerospace systems. In Kramer, P. A., et al. (1994). The effect of low gold concentrations on
Aerospace Electrical Interconnect Symposium, Tampa, FL, USA, the creep of eutectic tin-lead joints. Metallurgical and Materials
October. Transactions A,25, 12491257.
Kelly, M. A., et al. (1984). Using ESCA in failure analysisSome Krishnadev, M. R. (1993). The structure and properties of magnesium
recent developments. In Proceedings of ISTFA (pp. 3539). Los matrix composites. JOM,45(8), 5254.
Angeles. Krumbein, S. J. (1988). Metallic electromigration phenomena. IEEE
Kephart, A. R., & Hayden, S. Z. (1993). Benets of thread rolling Transactions Components, Hybrids, Manufacturing Technology,
process to the stress corrosion cracking and fatigue resistance of CHMT-11, 515.
high strength fasteners. In 6th International Symposium on Envi- Kuchler, G. (1995). Assessment of exible substrate integrity. In
ronmental Degradation of Materials in Nuclear Power Systems Hubble Space Telescope Solar-Array Workshop, ESTEC, Noord-
(pp. 18331847). San Diego, CA, August 35. wijk (pp. 253256). ESA Pubn WPP-77, May 1995.
Kessowsky, R., et al. (1978). Corrosion of indium based solders. In Kugel, U., Verain, M., & Zarraga, F. M. (1995). Post-ight investi-
16th Annual Proceedings, IEEE Reliability Physics (pp. 200205). gations of the Eureca Mechanisms. In Proceedings of Sixth
Keusseyan, R. L., & Dilday, J. L. (1993). Electric contact phenomena European Space Mechanisms Symposium (pp. 195201). Zurich,
in conductive adhesive connections. In Proceedings of International (ESA SP-374).
Surface Mount Conference and Exposure (pp. 567571). San Jose, Labruyre, G., & Urmston, P. (1995). ESA mechanisms requirements.
Surface Mount International. In Proceedings of 6th European Space Mechanism and Tribology
Ki, Young-Won. (1995). Gamma titanium aluminides: their status and Symposium, Zurich, (ESA SP-374), October 46
future. JOM,47(7), 3941. Lamouroux, F., et al. (1994). Kinetics and mechanisms of oxidation of
Kim, K.-H., et al. (2011). Observation of high resolution microstruc- 2-D woven C/SiC composites: Experimental approach. Journal of
tures in thermal sprayed coatings and single deposited splats using American Ceramic Society,77(8), 20492057.
ion beam milling. Materials Transactions,52(3), 439446. Lansdown, A. R., & Price, A. L. (1986). Materials to resist wear.
King, D. B., et al. (1976). Effects of inclusions on mechanical Oxford: Pergamon Press.
properties of Be. AFML-TR-76-33 Brush Wellman Inc., April. Laporte-Weywada, H. J, et al. (1993). CNES policy about launcher
King, R. L. (1988). Review, opportunities for advanced composites in space debris mitigationAriane V application. In Proceedings of
spacecraft. Advanced Composite Bulletin,2(1), 410. the First European Conference on Space Debris (pp. 585589).
Kingsley, N. (2008). Liquid crystal polymer: Enabling next generation Darmstadt, Germany, April 57, ESA SD-01.
conformal and multilayer electronics. Microwave Journal, 51, 188 Larsson, A., & Wingborg, N. (2011). Green propellants based on
200. ammonium dinitramide (ADN). www.intechopen.com/download/
Klein, A. J. (1988). Specialty reinforcing bres. Advanced Compos- pdf/13473.
ites,3(3), 3244. Lau, J., & Rice, D. (1985). Solder joint fatigue in SMT: State of the art.
Kleinteich, T., et al. (2015). Whats inside a shy suction cup? Solid State Technology, 91104.
Microscopy and Analysis,17, S8S10. Le May, I., & Deckker, E. (2009). Reducing the risk of failure by better
Klein Wassink, R. J. (1984). Soldering in electronics. Ayr, Scotland: training and education. Engineering Failure Analysis,16, 11531162.
Electrochemical Publications Ltd. Lea, C. (1988). Restrictions on the use of chlorouorocarbons for
Klinkrad, H. (2006). Space debris: Models and risk analysis. Heidel- cleaning soldered PCBs. Circuit World,14(4), 412.
berg, Germany: Springer-Praxis Publishing. Lea, C. (1988). A scientic guide to surface mount technology. Ayr,
Kocker, H. H., & Stockel, D. (1979). Material transfer of composite Scotland: Electrochemical Publications.
contact materials. In IEEE Transactions on Components, Hybrids Lea, C. (1991). Evidence that visual inspection criteria for soldered
and Manufaturing Technoloby (pp. 1519). CHMT-2. joints are no indication of reliability. Soldering and SMT Journal,9,
Kodama, S., et al. (2010). Development of stainless steel welding wire for 1924.
galvanized steel sheet, Rivista Italiana della Saldatura4: 479486. Lee, H. S., et al. (1994). Symptoms in peak expiratory ow rate among
Koh, D., et al. (1990). Dermatological hazards in the electronics female solderers. American Journal of Industrial Medicine,26,
industry. Contact Dermatitis,22, 17. 613619.
Koh, D., et al. (1994). Skin disorders among hand solders in the Lee, S.-B., et al. (2006). Electrochemical migration characteristics of
electronics industry. Occupational Medicine,44(1), 2428. eutectic SnPb solder alloy in printed circuit board. Thin Solid
Kohman, G. T., et al. (1955). Silver migration problem encountered in a Films,504, 294297.
telephone exchange. Bell Systems Technology Journal,34, 1115 Lee, Shen, Y. (1990). Thermomechanical properties of polymeric
1147. materials and related stresses. SAMPE Quarterly, 4851.
Konoza, A., et al. (2012). Analysis of the electronic assembly repair Leitao, C. M. A. (2013). Inuence of base material plastic properties
process for lead-free parts under combined loading conditions. and process parameters on friction stir weldability (Dissertation for
IEEE Transactions Components Packaging Manufacturing Tech- PhD, Department of Mechanical Engineering, University of Coim-
nology,2(9), 15581567. bra, Portugal, 2013).
Koons, H. C., et al. (1980). Spacecraft chargingResults from the Lestrat, D., & Prel, Y. (1994). Evolution et choix technologiques pour
SCATHA satellite. Astronautics and Aeronautics,18(11), 4447. les reservoirs des lanceurs Ariane. In Proceedings of the Interna-
Korb, G., & Dunn, B. D. (1996). Assessment of cleaning methods for space tional Symposium on Advanced Materials for Lightweight Struc-
mechanisms. ESTEC Metallurgy Report No. 2261 (condential). tures, ESTEC (pp. 441453). Noordwijk, March (ESA WPP-070).
References 647
Levine, A. S. (Eds.). (1993). Third post-retrieval symposium. NASA Maciolek, R. B., et al. (1978). Thermal aging characteristics of InPb
Conference Publication 3275, Part 3. solder bonds to gold. Honeywell Report.
Lewis, P. L., Kolody, M., & Curran, J. (2013). Alternative to nitric acid McCormick, J., & Zakraysek, L. (1979). A metallographic test for
passivation of stainless steel alloys. NASA Report, casi.nts.02809. glass-to-metal seals. In 17th Annual Proceedings of Reliability
Leyva, I. A., et al. (2011). Space Mission and Design Analysis, Physics, IEEE (pp. 4450). San Francisco, April.
Chap. 18Propulsion Systems, Air Force Research Lab., Edwards McCormack, S., & Meschter, S. (2009). Probabilistic assessment of
AFB., CA., USA. Report No. AFRL-RZ-ED-BK-2011-057. component lead-to-lead tin whisker bridging. In International Confer-
Li, H., & Zhang, J. (2014). Preparation of microarc oxidation ceramic ence on Soldering and Reliability, Toronto, Canada, May 2022.
layer on Ti6Al4V titanium alloy and its oxidation resistance McCrone, W. C., & Delly, J. C. (1973). The particle atlas (2nd ed.,
property. International Heat Treatment and Surface Engineering,8 Vol. 13). Michigan, USA: Ann Arbor Science Publishers.
(2), 5154. McCune, T. B., et al. (1970). Heat ageing evaluation of common coated
Li, X., & Dong, H. (2010). Ceramic conversion treatment of titanium- copper conductors. In 19th International Wire and Cable Sympo-
based alloys. In H. Dong (Ed.), Surface engineering of light alloys. sium, Atlantic City, N.J., December 24.
Cambridge, UK: Woodhead Publishing Ltd. McDonald, G. E. (1974). Renement in black chrome for use as a solar
Liedke, V., & Dunn, B. D., et al. (2008). Cleaning of silicone selective coating. NASA TMX-3136, December.
contaminations: development of test methods and assessment of McDonnald Schetky, L. (1991). Shape memory alloy applications in
cleaning efciency. In 11th ISMSE Symposium, Aix-en-Provence, space systems. Materials and Design,12(1), 2932.
France, September. McDowell, M. E. (1993). Tin whiskersA case study. In 1993 IEEE
Liejtens, J. (2015). Will the future of hirel components be pure SCOTS Aerospace Applications Conference (14th) (pp. 207215). Steam-
or only SCOTTish, IAA-B10-1203. In Deutsches Zentrum fur Luft- board Springs, CO, USA, January 31February 5.
und Raumfahrt 10th Symposium, Berlin, Germany, April 2023. McKay, C. A. (1983). Causes and effects of solder contamination, part
Lin, C.-C., et al. (2012). High temperature brazing of molybdenum. 2, Electronics: 4144.
Advanced Material Research Journal,586, 6973. Maillat, M., et al. (1995). Adhesion of materials by impact contacts in
Liu, H., et al. (2014). Laser nanocrystallisation and corrosion behavior vacuum. In Proceedings of 6th European Space Mechanisms
of electroless Ni-W-P coating with high phosphorous content. Conference, (pp. 431436). Zurich, Switzerland, (ESA SP-374).
Transactions IMF,92(4), 212217. Mair, P. (2005). Assessment of EMF and biological effects in arc
Liu, J., et al. (1996). Development of ip-chip joining technology. welding applications. In International Institute of Welding Manu-
Circuit World,22(2), 1924. facturing (pp. 110). IIW Document XII-1848-05.
Liu, X., & Lu, G.-Q. (2003). Effects of solder joint shape and height on Majid, W. A. (1988). Controlling SCC in mechanism components of
thermal fatigue lifetime. IEEE Transactions on Components and ground support equipment. In 22nd Aerospace Mechanisms Sym-
Packaging Technologies26(2): 445465. posium (pp. 163174). Hampton Virginia, May 46 (NASA
Lobley, G. R. (1990). Soldering problems on printed circuit boards. Conference Publication 2506).
Practice Metals,27, 627631. Maliutina, I. N., et al. (2014). Structure and microhardness of Cu-Ta
Lucas, J. P., & Zhou, J. (1993) The effects of sorbed moisture on resin joints produced by explosive welding. Rivista Italiana della
matrix composites. J.O.M., December issue, 3740. Saldatura5: 839848.
Lucas, W. (1995). Electromagnetic emission from TIG welding Malla, R. B., & Ghoshall, A. (1994). Thermally induced vibrations of
systems. Welding and Metal Fabrication,63(7), 261265. structures in space. Progress in Astronautics and Aeronautics,168,
Luciano, G., & Galet, G. (1995). Space mechanism actuated by a shape 6889.
memory alloy component. In Proceedings of the 6th European Marchaise, M., & Glodowski, K. A. (1991). Real-time microfocus
Space Mechanical and Tribology Symposium (pp. 221225). radiography for electronic failure analysis. Materials Evaluation:
Zurich, October 46. ESA SP-374. 14811485.
Lumsden, J. B., & Allen, A. T. (1988). The stress corrosion cracking Marrocco, T., & Harvey, D. (2012). Cold StandardThe cold spray
behaviour of AlLi alloy 8090. Corrosion Science-Nace,44(8), 527 process. Materials World: 3031.
532. Martell, S., & Adams, T. (1996). Virtual cross-sectioning of materials.
Lumsden, J. M., & Whittlesey, A. (1981). Automatic ARC welding of Materials World,3(7), 323324.
propulsion system tubing in close proximity to sensitive electronic Martin, J. W. (1976). A new laboratory procedure for determining
devices. In AIAA/SAE/ASME 17th Joint Propulsion Conference, compatibility of positive expulsion elastomers with liquid propel-
Colorado Springs, July, (AIAA-81-1569). lants. In SymposiumCompatibility of Plastics (pp. 16). Indian
Lund, A. (2011). Principals and applications of ESR Spectroscopy. Head, Md., USA, paper III-E, April.
Heidelberg: Springer. Mason, M., & Eng, G. (2007). Understanding tin plasmas: a new
Lupton, D. F. (1990). Heraeus GmbH, private communication. approach to tin whisker plasma risk assessment. In IEEE IRPS,
Luzin, V., et al. (2010). Use of neutron diffraction for stress CALCE, University of Maryland, USA, April.
measurements in thin and thick thermal sprayed coatings. Interna- Materassi, M., Dunn, B. D., & Capinari, L. (2000). The inuence of
tional Heat Treatment and Surface Engineering,4(1), 1724. solder llet geometry on the occurance of corona discharge during
Ma, H., et al. (2012). Effects of rework on the accelerated thermal operation between 400 and 900 V in partial vacuum. IEEE
cycling and shock performance of lead-free BGAs. IEEE Transac- Transactions on Electronics Packaging Manufacturing,23, 104115.
tions on Components Packaging Manufacturing Technology,2(11), Matloff, G. L. (2013). The speed limit for graphine interstellar solar
18241830. proton sails. Journal of the British Interplanetary Society,66,
Ma, Y., et al. (2012). A large-scale fabrication of ower-lie sub- 377380.
micrometer-sized tungsten whiskers via metal catalysis. Nanoscale Maurer, R. H., et al. (2008). Harsh environments: Space radiation
Research Letters,7(325), 17. environment, effects, and mitigation. John Hopkins APL Technical
Ma, H., & Lee, T.-K. (2013). Effects of board design variations on the Digest,28(1), 1729.
reliability of lead-free solder joints. IEEE Transactions on Compo- Mayo, M. J., & Nix, W. D. (1989). Direct observation of superplastic
nents Packaging Manufacturing Technology,3(1), 7178. ow mechanism in torsion. Acta Met,37(4), 11211134.
648 References
Maxwell, J. L., et al. (2013). High-temperature nano-composites for Mueller, R. P., et al. (2015). Additive construction using basalt regolith
nuclear thermal propulsion and in-space fabrication by HP-LCVD. nes, NASA KSC, casi.ntrs.nasa.gov/20150000305.pdf [cited
JBIS,66, 328333. access June 2015].
Mc Brien, C. M., & Heltzel, S. (2013). Insulation resistance of dielectric Muller-Wiesner, A., et al. (1994). Materials and manufacturing
materials under environmental testing. In IPC APEX EXPO Confer- technologies for the tank and thruster components of the Ariane 5
ence Proceedings, San Diego, Ca, USA, February 1921. SCA-system. In Ariane V structures and technologies, Cepadues-
Meaker, A. J., et al. (2006). Hydrogen peroxideFrom bridesmaid to Editions. Toulouse: CNES publication.
bride. In 3rd ESA International Conference on Green Propellants Munns, I. (1995). Adhesive bond inspection using NDT. Materials
for Space Propulsion, Poitier, France, September. World,3(11), 527529.
Medgyes, B., et al. (2011). In situ optical inspection of electrochemical Munson, T. (2006). White residues: are they or arent they? Circuit
migration during THB tests. The Journal of Materials Science: Assembly, page unknown, January.
Materials in Electronics,22, 694700. Nabarro, F. R., & Jackson, P. J. (1958). Growth of crystal whiskers.
Meltsner, K. J. (1995). Understanding the INTERNET: A guide for Growth and perfection of crystals (pp. 11101). New York: Wiley.
materials scientists and engineers. JOM,47(4), 913. Naisbitt, G. (2015). How clean is clean to be functionally good? EMPS-
Merstallinger, A., & Semerad, E. (1996). Test method to evaluate cold- 6 Workshop, DLR, Oberpfaffenhofen, Germany, April 1516.
welding under static or impact loading. Unpublished ESTEC Nakasa, K., & Liu, J. (1993). The effect of stress on hydride formation
Metallurgy Report 2222 in Ti6A14V alloy. In Titanium (Vol. 92, 1st ed., pp. 619627).
Merstallinger, A., & Semerad, E. (1996). Test method to evaluate the TMS publication.
sublimation of metals by weight loss. Unpublished ESTEC Metal- NASA NAS 5300.4 (3J). (1996a). Workmanship standard for staking
lurgy Report 2242. and conformal coating of printed wiring boards and electronic
Merstallinger, A., et al. (1995). Study on adhesion/cold-welding under assemblies, May.
cyclic load and high vacuum. In Proceedings of the 6th European NASA NAS 5300.4 (3M). (1996b). Workmanship standard for surface
Space Mechanisms Conference (pp. 293298). Zurich, Switzerland, mount technology, May.
(ESA SP-374). NASA-STD selected NASA standards related to materials and
Merstallinger, A., et al. (1995). Wire bend test. Unpublished report. processes can be seen listed in Appendix 9.
Merstallinger, A., & Dunn, B. D., et al. (2009). Assessment of cold Nasta, M. A., et al. (1990). Contaminant particles in electronic solder
welding between separable contact surfaces due to impact and uxes. Soldering and SMT Journal,4, 67.
fretting under vacuum, ESA STM-279, and, (2015) http://esamul- Nelson, H. G. (1988). Hydrogen and advanced aerospace materials.
timedia.esa.int/multimedia/publications/STM-279/. SAMPE Quarterly,20(1), 2023.
Mertens, J., & Gessford, D. (1995). Implementation of alternative NEMA. (1972). High temperature insulated wire. Report on Copper
cleaning solvents. Precision Cleaning Journal,3(2), 4550. Corrosion, NEMA Standards Publication, No. HP 2-1968 (reaf-
Mielke, E. W. (2002). Silver lled conductive epoxies, NASA Materials rmed by NEMA).
Engineering Branch, TIP No. 079, rev. October. Nemoto, N. (2007). Evaluation of tin whisker growth in vacuum
Mikaelian, T. (2001). Spacecraft charging and hazards to electronics thermal cycling condition. In 3rd International Symposium on Tin
in space, York University publication, May. Whiskers, Lyngby, Deenmark, June 2324.
MIL-C-13924 Coating, oxide, black, for ferrous metals. Newman, R. C. (1990). Stress corrosion cracking: 19651990. British
MIL-Handbook-697A. (1974). Titanium and titanium alloys, June. Corrosion Journal,25, 259269.
Millares, M., & Peraggi, B. (1992). Metallography of goldindium Niefhof, J. (1995). An empirical model for electromigration. Solid-State
couples and electronic components soldered with indium alloys. Electronics,38, 18171827.
Praktische Metallographie,29, 555563. Noller, E. W. (1996). Unique application of plastics and composite
Miller, N. H. (1906). The amounts of nitrogen as ammonia and nitric materials in the design of magnetometer instruments. In 41st
acid, and of chlorine in the rainwater collected at Rothamsted. International SAMPE Symposium and Exhibition (Vol. 1, pp. 841
Journal of Agricultural Sciences,1, 280303. 846). Anaheim, Ca, March 2428.
Minges, M. L. (1989). Electronics materials handbook (Vol. 1). Metals Nordwall, B. D. (1986). Air force links radar problems to growths of tin
Park, OH, USA: ASM International. whiskers. Aviation Week and Space Technology30: 6569.
Mohrbacher, H., et al. (1995). The inuence of humidity on the fretting Norwood, L. B. (1985). How beryllium proved successful on the space
behaviour of PVD TiN coatings. Wear,180, 4352. shuttle orbiter. Journal of Spacecraft and Rockets,22, 560566.
Moliterni, L. (2010). Il saldatore manual utilizzato nella saldatura in Novikov, I., & Grushko, O. (1995). Hot cracking of Al-Cu-Li alloys.
elettronica. Rivista Italianadella Saldatura, Lugili/Agosto 4: 455459. Materials Science and Technology,11, 926932.
Moliterni, L. (2012). Lafdabilita degli assemblaggi elettronici. Rivista Nuez, D., & Tan, P. (2014). Failure localization of intermittent short
Italianadella Saldatura, Maggio/Giugno 3: 351362. circuit failures caused by vertical CAF formation. In ISTFA
Morilla, D. J., et al. (2015). Tin whisker growth on surface mount Conference Proceedings, Houston, TX, USA, November 913.
devices: A study using standardized test conditions. In EMPS-6, Nurse, M. (1996). Caution remains on aluminiumlithium alloys. MBM
Wessling, Germany, April 1516. (pp. 4245). February.
Morris, J. W. (1996). The metallurgical control of electromigration. OClock, G. D., et al. (1987). Pb-Sn microstructure: potential reliability
JOM,48, 4346. indicator for interconnects. IEEE Transactions Components,
Moser, M. (2002). Calculation of stresses produced in the tin-plated Hybrids, Manufacturing Technology,10, 8288.
terminations of resistors during thermal cycling. Unpublished Estec Okuyama, F. (1974). Vapor-grown tungsten whiskers induced by
MGI report 3496. vacuum discharges. Journal of Applied Physics,45(10), 42394241.
Motz, J. M. (1988). MicrosegregationsAn easily unnoticed inuenc- Pantazopoulos, G., & Vazdirvanidis, A. (2008). Characterization of the
ing variable in the structural description of cast materials. Practice microstructural aspects of machinable Alfa Beta phase brass
Metals,25, 285293. (pp. 1316). Microscopy and Analysis, September.
Mueller, R. P., et al. (2014). Space mining and planetary surface Paprocki, S. J. (1988). Metal matrix composites for space applications.
construction. In ASCE Earth and Space Conference, St. Louis, In Composite Technology, 5th Annual Conference on Materials
Missouri, USA, October 2729. Technology (pp. 3343). Carbondale, IL, April 1415.
References 649
Parker, D. L. (1977). Special devices fabricated in heat shrinkable Piskoty, G., et al. (2009). Structural failures of rope-based systems.
plastics. In Proceedings of the Technical Program, Internepcon77 Journal of Engineering Failure Analysis,16, 19291939.
(pp. 133136). Brighton UK, October 1820. Plumbridge, W. J. (2011). Tin pest in lead-containing solders.
Parker, D. S. (1991). Analysis of a failed helicopter main rotor blade Soldering and SMT Journal,22(1), 5657.
grip. In International Conference on Failure Analysis (pp. 141 Poe, R. F., & Ruckner, M. A. (1993). Evaluation of pressurized vessels
145). Montreal, Canada, July 811 (published 1992 by ASM). following hypervelocity particle impact. In Proceedings of the First
Parquet, D. T., & Boggs, D. W. (1995). Alternatives to HASL. In European Conference on Space Debris (pp. 441446). Darmstadt,
Electronic Packaging and Production (pp. 3842). August. Germany, April 57, (ESA SD-01).
Pask, J. A. (1948). New techniques in glass-to-metal sealing. Pohl, D. (1996). Advances in scanning electron microscopy. Praktische
Proceedings of the IRE, Waves and Electrons Section,36, 286289. Metallographie,33(5), 235245.
Paton, B. E. (2003). Space: Technologies, Materials, Structures. In Polat, A., et al. (2008). Properties of alumina coating formed by
Welding and allied processes (Vol. 2). Ukraine: E. O. Paton Electric microarc oxidation technique on 6061 aluminium alloy. Praktische
Welding Institute. Published in English by Taylor and Francis Inc., Metallographie,12, 594606.
New York, USA. Pollack, A. A. (1989). Acoustic emission inspection. In Metals
Patterson, R. F., & Mykytiuk, P. D. (1996). Solid-state, vapour phase, Handbook (Vol. 17, 9th ed., pp. 278294). ASM International.
and aqueous cleaners for aerospace manufacturing operations. Pops, H. (1985). Copper rod requirements for magnetic wire. In
Sampe Journal,32(3), 4049. Proceedings of the 17th Electrical/Electronics Insulation Confer-
Pearce, R. (1987). SuperplasticityAn overview (pp. 11 to 124). ence (pp. 329337). Boston, USA, September 30October 3.
AGARD Lecture Series No. 154, August. Pradier, A. L., & Dosio, D. (1996). Technology development in
Pecht, M. G., et al. (1999). Moisture ingress into organic laminates. structures for reusable launch vehicles. In Reaching for the Skies
IEEE Transactions on Components and Packaging Technology,22, (Vol. 15, p. 7). ESA Publication.
104110. Premat, G. (1977). Astudy of the outgassing of solder splices. ESA TM-
Pedley, M. D. (2009). Materials safety, Ch. 12 in book safety design for 158.
space systems. Amsterdam, The Netherlands: Elsevier Publishing. Pretsch, E. (2009). Structure determination of organic compounds.
Peel, C. J., & McDarmaid, D. S. (1989). The present status of the Heidelberg: Springer.
development and application of Al-Li alloys 8090 and 8091. Prymak, J. D., & Bergenthal, J. (1995). Capacitance monitoring while
Aerospace,16(5), 1823. ex testing. IEEE Transactions on CPMT, Part A,18(1), 180186.
Peng, X., et al. (2014). Conductivity improvement of silver akes lled Puig, C. (2006). When space adheres to bondingAn interview.
ECAs via introducing silver-graphine nanocomposites. Journal of Qualite Espace,43, 5763. Mars.
Materials Science: Materials in Electronics,25, 11491155. Puippe, J-Cl. (1999). Surface treatments on aluminium for space
Perami, R., et al. (1992). Applications of AE to the study of applications. Galvanotechnique,90(10), 30033009.
microssure damage to composites used in the aeronautics and Puttlitz, J. K. (1990). Preparation, structure and fracture modes. IEEE
space industries. In The American Society for Nondestructive Transactions Parts, Hybrids, Packaging,13, 648655.
Testing, Proceedings of AECM-4, (pp. 99108), Seattle, July 2731. Quinn, G. D. (2008). Fractography of brittle materials: Analysis
Pereira, A., Pimenta, G., & Dunn, B. D. (2008). Assessment of chemical of fractures in ceramics and glasses. Microscopy and Analysis:
conversion coatings for the protection of aluminium alloys, (A 2124.
comparison of Alodine 1200 with Chromium Free Conversion Raiser, G. F., & Amir, D. (2005). Solder joint reliability improvement
Coatings), ESA-STM-276, February. using cold ball pull metrology. In Proceedings of InterPACK05,
Peres, P., Semerad, E., & Dunn, B. D. (1994). Material characteristics San Francisco, CA, USA, July 1722.
of SiC/MAS-L bre re-inforced glass composites. In Proceedings of Raman, V., & Reiley, T. (1987). Cavitation and cracking of
6th International Symposium on Materials in Space Environmental superplastic PbSn eutectic during high-temperature fatigue. Jour-
(pp. 309314). Noordwijk, September 1923, 1994, (ESA SP-368). nal of Materials Science Letters,6, 549551.
Perez, F., & Eymard, P. (1985). Increasing use of composite structures Rameriz, P., & Adams, T. (2002). Acoustic imaging and screening of
on Ariane launchers (pp. 239244). ESA SP-243. ceramic chip capacitors. Passive Component Industry: 1620.
Pergola, P., et al. (2011). Low-thrust missions for expanding foam space Rampini, R., et al. (2009). Dynamic outgassing test: Revisited. In 11th
debris removal. In IEPC-2011-126, 32nd International Electric ISMSE Symposium, Aix-en-Provence, September 1518, Poster.
Propulsion Conference, Wiesbaden, Germany, September 1115. Ratchev, P., et al. (2005). A study of brittle to ductile fracture transitin
Peters, S. T., & Wesling, N. (1968). Corrosion of silver plated copper temperatures in bulk lead-free solders, EMPC, Brugge, Belgium,
conductors. In SAMPE 13th National Symposium. June 1215.
Pettersson, C., et al. (2007). Corrosion testing of welds, a review of Ray, U., et al. (1995). Inuence of temperature and humidity on the
methods. Rivista Italiana Saldatura,4, 529553. wettabiality of immersion tin coated pwbs. IEEE Transactions on
Pettit, R. B., & Sowell, R. R. (1976). Solar absorptance and emittance CPM Technology Part A,18(1), 153161.
properties of several solar coatings. Journal of Vacuum Science and Raymond, L. (1988). Hydrogen embrittlement: Prevention and control.
Technology, 13(2). No: ASTM Special Technical Publication. 962.
Pinaud, G., et al. (2014). Aerofast: Development of cork TPS material Redahan, E. (2014). 3D printingMaterial wise to grain sizeA review
and a 3D comparative thermal/ablation analysis of an Apollo and a of the 3D printing and additive manufacturing industrial applications
biconic sled shape for an aerocapture mission. http://solarsystem.- summit (London 2013). In Materials World, January 1012.
nasa.gov/docs/7_aerofast_paper.pdf (2015). Reed, R. P., & Clark, A. F. (1983). Materials at low temperature.
Pippin, H. G. (1995). Analysis of silverized teon thermal control ASM.
material ow on LDEF. NASA Contractor Report 4663, July. Reid, P. (2009). PWB reliabilityThe next step. Circuit World,33(4),
Pippin, H. G., & Bourassa, R. J. (1995). Effects of space exposure on 5159.
metals on LDEF. NASA Contractor Report 4662, June. Rendu, M., & Tawil, D. (1988). Improved protection of magnesium
Pirnnack, S. (2012). Lessons learned to avoid coax cable failure in alloys against synthetic aviation lubricants at elevated temperatures,
moving mechanical mechanisms. In Proceedings of the 41st Aero- paper 880869. In 24th Annual Aerospace/Airline Plating and Metal
space Mechanisms Symposium, NASA JPL, Pasadena, CA, USA. Finishing Forum, Phoenix, April.
650 References
Reuther, J. (2010). Orion thermal protection system, advanced Schatz, et al. (1979). Development and ight experience of the Voyager
development project. In 7th International Planetary Probe Work- propulsion system. In AIAA/SAE 15th Joint Propulsion Conference
shop, NASA Ames Research Centre, June 16. (pp. 113), Las Vegas, June.
Richards, B. P., & Footner, P. K. (1990). Failure mechanisms in active Schedler, A. (1988). Fibre composites in satellites. Cryogenics,28,
surface mount components. GEC Journal of Research,7(3), 146156. 220223.
Ries, W. (1996). The combined plasma cleaning facility. Product Scheel, W. (2004). Electronics assembly handbook (2nd ed.). Port Erin,
Finishing,49(1), 68. I.O.M, UK: Electrochemical Publications.
Roman, I., et al. (1992). Interfacial shear properties and acoustic Schelling, P., et al. (2005) Managing heat for electronics. Materials
emission behaviour of model aluminium and titanium matrix Today: 3035.
composites. In The American Society for Nondestructive Testing, Schemme, K., & Wittkamp, I. (1993). Microstructure development and
Proceedings of AecM-4 (pp. 109114), Seattle, July 2731. examination of MgLi alloys. Praktische Metallographie,30(7),
Richards, B. P., & Footner, P. K. (1984). Morphological aspects of 334343.
electromigration. GEC Journal of Research,2(3), 157168. Schlesinger, M. (2010). Electroless deposition of nickel. In M.
Ring, K. (2010). Protection of electronic assemblies in condensing Schlesinger & M. Paunovic (Eds.), Modern electroplating. Chich-
atmospheres. EMPS-1, Portsmouth, UK, February 17. ester, UK: Wiley.
Roach, T. A. (1984). Aerospace high performance fasteners resist stress Schliekelmann, R. J. (1972). Holographic interference as a means for
corrosion cracking. Material Performance,239, 4245. quality determination of adhesive bonded metal joints. In 8th
Robertson, I. M., & Teter, D. (1996). Microscopic studies on hydrogen Congress of the International Council of the Aeronautical Sciences,
effects on mechanical properties. JOM,48(11), 5560. Paper No. 72-06.
Rogers, K., et al. (1999). Conductive lament formation in a pcb. Schmidt, E. W. (1984). Hydrazine and its derivatives. New York:
Circuit World,25, 69. Wiley.
Ross, C. A., & Evetts, J. E. (1987). A model for electromigration. Schoenbeck, J., & Dunn, B. D. (2004). Evaluation of a process for the
Scripta Metals,21, 10771082. repair of area array and other surface mounted packages, ESA
Ross, R. G. (1989). Magellan/Galileo solder joint failure analysis and STM-272, September.
recommendations. NASA-CR-186294, September. Schoenthaler, D. (1984). Accelerated ageing for solderability evalua-
Rothschild, B. F. (1981). Electroplating of solderable coatings. Metal tions. IPS-TR-464.
Progress: 2529. Schoyer, H. F. R. (1996). ESAs new solid propellant based on
Rowntree, R. A., & Todd, M. J. (1988). Tribology for spacecraft. hydrazinium nitroformate. Preparing for the Future,6(1), 67.
Chartered Mechanical Engineer Journal: 2830. Schreib, R., et al. (1979). Flow anomalies in small hydrazine thrusters.
Rubinstein, M. (1978). High speed electroplating of gold. Gold In AIAA 15th Joint Propulsion Conference (pp. 17), Las Vegas,
Bulletin,11(1), 38. June.
Russell, C., et al. (2014). Welding technology takes ight with NASA. Schrunk, D., et al. (1999). The MoonResources, future development
Welding Journal,93(5), 3844. and colonization. Chichester, UK: Wiley-Praxis.
Rustichelli, F., & Dunn, B. D. (1997). Assessment of residual stresses Schubert, Th, et al. (2008). Interfacial design of Cu-based composites
in AA 2219 weldments by neutron diffraction. To be published in prepared by powder metallurgy for heat sink applications. Materials
Materials Science and Engineering. Science and Engineering: A,475, 3944.
Ryan, R. S., et al. (1995). Development problems and their solutions Schuerch, H. U. (1968). Certain properties and applications of Nitinol
for the Space Shuttle main engine. NASA Technical Paper 3533, NASA CR-1232.
May. Schweigart, H. (2007). Conformal coating issues: When reliability goes
Saber-Samandari, S., & Berndt, C. C. (2010). IFTHSE Global 21: Heat astray. Global SMT and Packaging, 3.
treatment and surface engineering in the twenty-rst century. Scott, B. C. (1987). Zinc diffusion in tin coatings on brass. Transac-
International Heat Treatment and Surface Engineering,4(1), 713. tions on IMF,65, 9098.
Said, A. F., et al. (2012). Automated void detection in solder balls in Scott, P. G. (1985). An investigation of alodine surface nish on three
the presence of vias and other artifacts. IEEE Transactions on CPM aluminium alloys. EWP No. 1435 (Unpublished).
Technology,2(11), 18901901. Sechi, Y., et al. (2009). Composition dependence of titanium in Ag-Cu-
Saltzman, M., et al. (2009). Problems encountered during the Ti alloy braze metals for laser brazing of boron nitride ceramics and
recertication of a solar array drive actuator. In 13th Space cemented carbides. Material Transactions,50(6), 12941299.
Mechanisms and tribology Symposium, Vienna, Austria, September Seghesio, J. L. (1986). Heurs et malheurs de la corrosionle tore deau
2324. dAriane. CNES Quality Espace,16(5), 2932.
Sampson, B. (2007). Soldering onTraining focus. In Professional Selman, G. L., et al. (1974). Dispersion strengthened platinum.
Engineering (p. 48), April. Platinum Metals Review,18(2), 213.
Sandera, J. (1996). Investigation into the mode of degradation of Semerad, E., & Dunn, B. D. (1990). Evaluation of black anodised
crimped joints due to corrosive effects. Unpublished ESTEC beryllium (Unpublished report).
Metallurgy Report 2256. Semprimoschnig, C., & Eesbeek, M. (2007). Materials research at Estec
Saunders, C. G. (1988). Ion techniques in surface engineering. Metals focus on self-healing materials for space applications. In STFC
and Materials,4(11), 678682. Materials Workshop, London, October 10.
Saxer, W. (1977). Moeglichkeiten funktioneller Edelmetallueberzuege Sharkov, E. A. (2008). Breaking Ocean Waves - geometry, structure
anhand aktueller Beispiele aus der Praxis. Galvanotechnik, 68(11) and remote sensing. Heidelberg, Germany: Springer-Praxis.
(Black inorganic coating developed by Werner Fleuhmann AG, Sharma, A. K. (1989). Gold plating on aluminium alloys for space
Dueberdorf, Switzerland). applications. Transactions of the Institute of Metal Finishing,67,
Saxty, P. (1995). Ultrasonic soldering in the electronics industry. 8788.
Metallurgica: 287. Shashkov, P. (2013). Aluminium nanoceramic substrates for thermal
Scala, E. P. (1996). A brief history of composites in the US. JOM,48 management in electronics. In Electronic Materials and Proceed-
(2), 4548. ings For Space Workshop, (EMPS-4), Aalborg, Denmark.
References 651
Sheldahl Inc. (2004). Thermal control material and metallised lms. Spedding, V., & Hargreaves, P. (1995). Materials on line. Materials
General Trade Literature and Specications, and The Little Red World,3(10), 483484.
Book. Sriveeraraghavan, S., et al. (1995). Chemical removal of solder and
Shibutani, T. (2009). Effect of grain size on pressure-induced tin solder electroplates. Metal Finishing: 3436.
whisker formation. In 3rd International Tin Whisker Symposium, Stanley, J. K. (1968). Solutions to some stress corrosion cracking
Lyngby, Denmark, June 2324. problems in aerospace situations. In Proceedings of 1st Interna-
Shiganov, N. V. (1973). Some special features of arc welding with a tional Aerospace and Marine Corrosion Technical Seminar,
hollow electrode in vacuum. Welding Production, 20(9), 3033 NACE, LA, USA, July.
Shipley, E. A. (1988). Developments in ultra high tensile wire ropes. In Stoltz, R. E., & Stulen, R. H. (1979). Slid metal embrittlement of Ti
Proceedings of Conference on the Hoisting of Men, Materials and 6A16V2Sn by cadmium, silver and gold. Corrosion-NACE,35
Minerals (pp. 12951307), Toronto. (4), 165169.
Shrestha, S., Merstallinger, A., Sickert D., & and Dunn, B.D. (2003). Strachan, W. (2000). Processes for manual solder assembly of high
Some preliminary evaluations of black coating on aluminium heat capacity boards (Unpublished).
AA2219 alloy produced by plasma electrolytic oxidation (PEO) Stratton, P. (2013). Ellingham diagramsTheir use and misuse.
process for space applications. In Proceedings of 9th International International Heat Treatment and Surface Engineering Journal,7
Symposium on Materials in a Space Environment (pp. 5765), The (2), 7073.
Netherlands, June 16-20. Sterjovski, Z., et al. (2010). The effect of voltage and metal transfer
Shrestha, S., et al. (2007). Recent developments in black nish coatings mode on particulate fume size. Rivista Italiana della Saldatura,6,
on aluminium by Keronite plasma electrolytic oxidation. Proceed- 765780.
ings of the Global Power Train Congress, Engine, Transmission Sturgeon, A., & Dunn, B. D. (2006). Cold sprayed coatings for polymer
and Propulsion, Berlin,3942, 207215. composite structures. In Proceedings of 10th ISMSE Conference,
Shrestha, S., & Dunn, B. D. (2010). Plasma electrolytic oxidation and Collier, France, June 1923 (ESA SP-616).
anodizing of aluminium alloys for spacecraft applications. In H. Suchentruck, R. (1986). Electroforming of complex parts. Transactions
Dong (Ed.), Surface engineering of light alloys. Cambridge, UK: IMF,64, 1923.
Woodhead Publishing Ltd. Sun, J., et al. (2012). Kinetic study of the pyrolysis of waste pcbs
Shrestha, S., et al. (2015). Silk corrosion and stress corrosion cracking subject to conventional and microwave heating. Energies,5, 3295
behavior of Plasma electrolytic oxidized Al-SiC metal matrix 3306.
composite, to be published. Surkov, Y. (1997). Exploration of terrestrial planets from spacecraft
Shumka, A., & Pietry, R. (1975). Migrated-gold resistive shorts in (2nd ed.). Chichester, UK: Wiley-Praxis.
microcircuits. In Proceedings of 13th IEEE International Reliability Suslov, D. I., et al. (2010). Convective and lm cooled nozzle
Physics Conference (pp. 9398), Las Vegas. extension for a high pressure rocket subscale combustion chamber,
Sicking, R. (2005). Brazing of aluminium heat exchangers. Welding AIAA 2010-1150. In 48th AIAA Aerospace Sciences Meeting
and Cutting Journal,4(3), 150159. Including New Horizons Forum, Orlando, FL, USA, January 47.
Silverman, E. M. (1995). Space environmental effects on spacecraft: Tak, W, et al. (2009). SMA actuator based satellite separation device
LEO materials guide. NASA Contractor Report 4661 Part 1, using plastic deformation. In 13th European Space Mechnisms and
August. Tribology SymposiumESMATS, Vienna, September 2325 (ESA
Simpson, P. R., & Bowie, S. H. U. (1970). Quantitative optical and SP-670).
electron-probe studies of the opaque phases. Science,167, 619621. Tamai, T. (1995). Effects of silicone vapour and humidity on contact
Sinclair, J. D., et al. (1985). Volatile components in rosin ux: reliability. In Proceedings of IEEE Holm Conference (pp. 252259).
characterisation and reliability effects. In 35th IEEE Electronic Tamai, T., & Aramata, M. (1993). Safe levels of silicone contamination
Components Conference (pp. 144150), Washington DC, May 2022. for electrical contact. In Proceedings of IEEE Holm Conference on
Singh, J., et al. (2005). Welding in Space. Advanced Materials and Electrical Contact Phenomena (pp. 269273), England
Processes Journal: 5257. Teare, M. (1994). Manufacture of a multi-horn feed by electroforming.
Smith, C. A. (2007). Chemical characterization of materials in In International Symposium on Advanced Materials for Lightweight
electronic systems using infrared spectroscopy. Circuit World,33 Structures, ESA-ESTEC (pp. 305309), March 2224. ESA WPP-
(3), 3847. 070.
Sogame, E., et al. (1990). Recent developments in the hydrazine related Teer, D. G. (2001). New solid lubricant coatings. Wear,251, 1068
propulsion technology in NASDA. In Proceedings of 17th Inter- 1074.
national Symposium on Space Technology and Science (Vol. 1, Tegehall, P.-E. (1991). The impact of crevices beneath smds on the
pp. 107112), Tokyo. cleaning efciency. Soldering and SMT Journal,8, 4652.
Sohn, J. E., & Ray, U. (1995). Weak organic acids and surface Tegehall, P.-E., & Dunn, B. D. (1992). Inuence of ux residues and
insulation resistance. Circuit World,21(4), 2225. conformal coating on the SIR properties of spacecraft pcbs. ESA
Solari, G., & Repetto, M. P. (2014). IL comportamento delle structure Journal,16(3), 255274.
metalliche saldate nei confronti delliazione del vento. Rivista Tegehall, P.-E., & Dunn, B. D. (2001). Assessment of the reliability of
Italianadella Saldatura: 851859. solder joints to ball and column grid array packages for space
Soli, L., & Dunn, B. D., et al. (2015). An electron beam brazing reow applications, ESA-STM-266.
technique. Giorate Nazionali di Saldatura, GNS8, Genoa, Italy, Tegehall, P.-E., & Dunn, B. D. (2001a). Evaluation of thermally
May 2829. conductive adhesives as staking compounds during the assembly of
Somiya, I., et al. (2013). Products with special electrical and magnetic spacecraft electronics. ESA STM-265, August.
properties, Ch. 7. In Handbook of advanced ceramics: Materials, Tegehall, P.-E., & Dunn, B. D. (2003). Impact of cracking beneath
applications, processing and properties. Amsterdam, The Nether- solder pads in PCB laminate on reliability of solder joints made to
lands: Elsevier Inc. ceramic ball grid array packages. ESA-STM-267.
Sood, B., et al. (2011). Whisker analysis of electronic throttle controls. Tegehall, P.-E., & Dunn, B. D. (2005). Impact of reworking ceramic
Circuit World,37(3), 49. AGAs on the integrity of PCB laminates. ESA STM-273, June.
652 References
Tegehall, P.-E., & Dunn, B. D. (2006) Evaluation of cleanliness test Tyler, D. (2015) Reliability of ball grid arrays converted from lead-free
methods for spacecraft PCB assemblies. ESA-STM-275. to tin-lead by robotic hot solder dipping, EMPS-6, Wessling,
Tegehall, P.-E., & Wetter, W. (2015). Impact of laminate cracks under Germany, April 1516.
solder pads on the fatigue lives of BGA solder joints, Microelec- Vail, J. R., et al. (2011). Polytetrauoroethylene ber reinforced
tronics Reliability, article in press. polyetheretherketone composites. Wear,270, 737741.
Thiel, C., et al. (1993). Functional activity of plasmid DNA after entry Vander Voort, G. F. (2008) Titaniumspecimen preparation,
into the atmosphere of Earth investigated by new biomarker Advanced Material and Processes, February 2527.
stability assay for ballistic spaceight experiments. PLOS One, van Eesbeek, M., et al. (1994) Degradation of Teon FEP due to VUV
November 26. and atomic oxygen exposure. In Proceedings of 6th International
Thiel, M., et al. (2003). The Rosetta Lander Anchoring System. In Symposium on Matls in Space Environment, ESTEC (pp. 165173).
Proceedings of the 10th European Space Mechanisms and Tribol- ESA SP-368.
ogy Symposium (pp. 239246), San Sebastian, Spain, September Vancso, J, et al. (2009) Whats new in atomic force microscopes for
2426. polymers. Microscopy and Analysis, May 511.
Thoma, W. (1981). Product liability. In 2nd ESA Product Assurance Vaughn, R. L. (1985). Space shuttleA triumph in manufacturing.
Symposium (pp. 3133), ESA SP-163. Dearborn, USA: Society of Manufacturing Engineers.
Thomas, G., et al. (1993). The effect of pre- and post-weld heat VEGA. (2006) Users Manual, Issue 3 rev.0, and ESA-BR-257 (2007).
treatments on the mechanical properties of EB welded Ti6A14V. Veit, H. M., et al. (2002). Using mechanical processing in recycling
Journal of Material Science,28, 48924899. printed wiring boards. JOM, June 4550.
Thomas, G., & Dunn, B. D. (1999). Assessment of electrically Verbraak, C. A., TNO Metaalinstituut. Delft: Private Communication.
conductive adhesives (ECAs) for joining surface mount devices to Verderaime, V., & Vaughan, R. (1995) Aluminium U-groove weld
printed circuit boards. In 2nd IEEE Symposium on Polymeric enhancement based on experimental stress analysis. NASA Techni-
Electronic Packaging (PEP 99) (pp. 913), Gothenburg, Sweden, cal Paper3581.
October 2428. Verdin, D., & Duck, M. J. (1985). Conductive coatings to minimise the
Thomas, R. W., & Calabrese, D. W. (1983). Phenomenological electrostatic charging of Kapton. 3rd European Symposium on
observations on electromigration. In 21st Annual Proceedings in Spacecraft Materials in Space Environment (pp. 125131). ESA
Reliability Physics (pp. 19), Phoenix, Arizona, April 57. SP-232.
Thompson, D. H. (1961). A simple stress corrosion cracking test for Vine, M. K., & Price, W. B. (1992) A literature review of fretting in
copper alloys. Materials Research and Standards,1(2), 108111. launch, re-entry and space environments. ESTEC Contractor Report
Thomson, P., et al. (1995). The use of shape memory alloys for passive from ESTL (TM-106).
structural damping. Smart. Materials and Structures,4, 3642. Vitt, B. (1987). Black-cobalt coating for solar collectors. Philips
Tighe, A. et al (2010) Overview of results from the materials exposure Technical Review,43, 244252.
and degradation experiment (MEDET) after 18 months in orbit on Voigtlander, B. (2015). Scanning Probe Microscopy. Heidelberg:
the ISS, ESA paper. Springer.
Tillmann, W., & Boretius, H. (2008). Properties of nickel-based joints Volkert, C., & Minor, A. M. (2007). Focused ion beam microscopy and
between diamond and steel for diamond grinding tools. Welding micromachining. MRS Bull,32, 389395.
and Cutting,7(4), 228234. Vranish, J. M., & Gorevan, S. (1995). Basic space payload fastener.
Toensmeier, P. A. (2009). Shape memory polymers reshape product 29th Aerospace Mechanisms Symposium, NASA JSC (pp. 152157).
design. In Plastics engineering, April 2. NASA Conf. Publication 3293.
Towner, J. M., et al. (1984). Grain growth in Al alloy conductors as a Wagner, R. (1992). Metallized CFRP technology. Germany: Private
result of rapid annealing. Applied Physics Letters,44, 198199. Communication, Dornier Space Systems.
Tondu, T., et al. (2011). Cesium droplet evaporation as contamination Walmsley, M. (2015). Ruggedised solder joint technology for leadless
route by cesium eld effect electric propulsion. Journal of semiconductor packages. EMPS-6, Wessling, Germany, April 15
Spacecraft and Rockets,48(3), 513519. 16.
Troughton, M., et al. (2013). State of the art in welding thermoplastics Ward, P. (1994). Effects of bearing wash procedures in performance.
and in assessing thermoplastic welded joints. Rivista Italiana della Non-Ozone depleting Chemical Cleaning EPA Workshop, Denver,
Saldatura,6, 887895. Co, September 2627.
Tsuda, N. (2000). Electronic conduction in oxides (p. 243). Heidelberg, Ward, P. P. (1996). Plasma cleaning techniques and future applications
Germany: Springer. in environmentally conscious manufacturing. SAMPE Journal,32
Tsukada, M., et al. (2012). A new strategy for assessing off-gassing (1), 5154.
from museum materials. AIC News,37(1), 17. Warwick, M. E., & Muckett, S. J. (1983). Observations on the growth
Tulsi, S. S. (1986). Properties of electroless nickel. Transactions of intermetallics on tin-coated substrates. Circuit World,9(4), 511.
IMF,64, 7376. Watson, J. A., et al. (2002) A history of astronaut construction for large
Turlach, G. (1985). Improving fatigue strength of aerospace metal space structures at NASA LRC. In IEEE International Space
fasteners by surface work hardening. Surface Engineering,1(1), Station Conference. Nyact, NY, USA, March 916.
1722. Webster, D. (1994). Aluminiumlithium, the next generation.
Turner, I., Dunn, B. D., & Barrnes, C. (2013). A study into the re- Advanced Materials and Processes,145(5), 1824.
processing of pure tin termination nishes into tin-lead. Soldering Wedgewood, A. (1994). Measuring residual stress. Materials World,3
and SMT Journal,25(4), 218228. (1), 57.
Turner, R. F. (1995). Chapter 9. In P. Fortescue & J. Stark (Eds.), Weirick, L. J. (1975). A metallurgical analysis of stress corrosion
Spacecraft systems engineering (2nd ed.). Chichester: Wiley. cracking of Kovar package leads. Solid State Technology,18(3), 25
Turner, T. E., & Parsons, R. D. (1982). A new failure mechanism: A1 30.
Si bond pad whisker growth. IEEE Transaction on Components, Wescon. (1965). Session 16B Papers (Western Electronic Show and
Hybrids and Manufacturing Technology,CHMT-5, 431435. Convention, California, 1965), July.
References 653
Wetzer, J. M., & Wouters, P. (1993). The effect of insulator charging on Xu, H., et al. (2013). Wafer-level SLID bonding for MEMS encap-
breakdown and conditioning. IEEE Transactions,EI 28(4), 681691. sulation. Advances in Manufacturing,1, 226235.
Wetzer, J. M., & Wouters, P. (1994). High-voltage design of spacecraft Yao, W., & Basaran, C. (2014). Damage mechanics of electromigration
vacuum components. Paper from the High-voltage and EMC and thermomigration in lead-free solder alloys under alternating
Group, Eindhoven University of Technology, The Netherlands current: An experimental study. International Journal of Damage
Whalen, M. V. (1988). The compatibility of dispersion-strengthened Mechanics,23(2), 203221.
platinum with propellants. Platinum Metals Rev,32(1), 210. Yao, Y., Zhou, Y., & He, L. (2013). Corrosion behavior of molybdate
Wickham, M., Hunt., Adams, D., & Dunn, B. D. (1999). An conversion coating on AZ31 magnesium alloy in NaCl solution.
investigation into ball grid array inspection techniques, ESA Anti-Corrosion Methods and Materials,60(6), 301311.
STM-261. Yext, W. F., et al. (1983). Improved glass-to-metal sealing through
Wickham, M. et al. (2014). Whisker mitigation measurements using a furnace atmosphere control. IEEE Transactions on Components,
test vehicle based on a printed circuit assembly. 8th International Hybrids and Manufacturing Technology CHMT-6 (pp. 455459),
Symposium on Tin Whiskers. Raleigh, NC, USA, October 29. December
Wild, R. N. (1975) Some fatigue properties of solders and solder joints. Ylikorpi, T., et al. (1995). Development of an actuator for CAPS on
INTER-NEPCON, Brighton, UK. IBM report No. 74Z000481, spacecraft Cassini. In Proceedings of 6th European Space Mecha-
October. nisms Conference (pp. 97101). Zurich, Switzerland, (ESA SP-374).
Williams, D. N, et al. (1970). Hydrogen segregation in Ti6A14V Yoder, D. D., et al. (1993). Technical article on plated through hole
weldments made with unalloyed titanium ller material. Welding fatigue life, Part 2. IPC Review,34(5), 2027.
Journal Research Supplement, May 207212. Yost, F. G. (1976). Indium soldering on gold. In Proceedings of
Wingborg, N., de Flon, J., Johnson, C., & Whitlow, W. (2008) Green International Microelectronics Symposium (p. 61).
Propellants Based on ADN. Space Propulsion 2008. Heraklion, Yost, F. G., et al. (1976). Layer growth in Au-Pb/In solder joints.
Crete, Greece. ESA, 3AF, SNPE, May 58, 2008. Metallurgical Transactions A,7A, 11411156.
Winkler, W. (1975). Conductive coating: problem of electrostatic Zhang, W., & Jin, H. (2009). Study of tungsten metallization surface
cleanliness. Acta Astronautica,2(78), 745754. states for multilayer ceramics. In IEEE Proceedings of ICEPT-HDP
Wolfenden, R. T. (1996). Problems with solvent vapours in arc welding 09 International Conference. Beijing, China, August 1013.
fabrication environments. Welding and Metal Fabrication,64(4), Zhao, H., et al. (2013). Controlled atmosphere brazing of aluminium
136139. heat exchangers. Welding Journal, February 4446.
Wolff, R. H. (1963). Corrosion of plated metals after heating in Zimcik, D. G. (1987). Effect of long-term exposure to LEO space
vacuum. Plating Journal, 905910, October. environment on spacecraft materials. Canadian Aeronautics and
Wright, C. (1977). The effect of solid-state reactions upon solder lap Space Journal,33(1), 410.
shear strength. IEEE PHP,13(3), 202207. Zirker, L. R., et al. (2007). Weld tests conducted by the Idaho National
Wu, X. J., et al. (1994). The orientation dependence of fatigue crack Laboratory, INL/EXT-07-12289.
growth in 8090 A1Li plate. Metallurgical and Materials Trans- Zou, L., Dunn, B. D., et al. (1999). An evaluation of the effect of ageing
actions A,25, 575588. on the cleanability of solder ux residues. Soldering and SMT
Xu, H., et al. (2009). A re-examination of the mechanism of Journal,11, 2735.
thermosonic copper ball bonding on aluminium metallised pads. Zwanenburg, R. (1995) The qualication test programme of the Envisat
Scripta Meterialia,61(2), 165168. solar array mechanism. In Proceedings of 6th European Space
Xu, H., et al. (2010). The role of bonding duration in wire bond Mechanisms Conference (pp. 6572), Zurich, Switzerland. ESA
formation: a study of footprints of thermosonic gold wire on SP-374..
aluminium pad. Microelectronics International,27(1), 1116.
Index
Note: Bold page numbers are used for main references. An italic number refers to an illustration on the page.
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
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
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
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