Gerhard Eggert and Britta Schmutzler (Eds.

Archaeological Iron
Conservation Colloquium 2010

Extended Abstracts
 Archaeological Iron
Conservation Colloquium 2010

State Academy of Art and Design Stuttgart

24th to 26th June 2010
INTRODUCTION:
THE MAKING OF A CONFERENCE

GERHARD EGGERT1, BRITTA SCHMUTZLER1

1
State Academy of Art and Design, Am Weißenhof 1, D-53489 Stuttgart, Germany
gerhard.eggert@abk-stuttgart.de; b.schmutzler@abk-stuttgart.de

There's a problem
Guess, when this was written: “As is generally
known, excavated iron objects decay – depen-
ding on their state of deterioration- sooner or
later even in the driest exhibition or storage
rooms despite all precautions.”? No, although
we are today all aware of the problem, that's
not a current description from an
archaeologist. It was written by the Berlin
museum chemist Eduard Krause in 1882, he
already supposed the action of chloride and
Fig. 1: ’Weeping iron’: Who may dry its
found ferrous chloride in 'weeping iron' (Fig.
tears? Photo: I. Wiesner, SABK
1).

But the problem to stabilize iron finds is still with us, washing with water as Krause
recommended does not help very much. The alkaline sulphite treatment described by North
and Pearson was a big step forward for individual finds without organic remains. For dealing
with the tons of iron excavated every year many conservation labs found it too complicated.
While working on a book on iron conservation (Scott and Eggert 2009) it became apparent
that there might be ways to simplify the treatment (no heating, diluted solutions, alternatives
to sulphite to keep the oxygen out). That was start of a research project, generously funded by
the Deutsche Bundesstiftung Umwelt (DBU) who also supports this colloquium to share its
results with the international conservation community. The idea to compare variations was
straight forward: It's not the amount of chloride extracted into the solution, but the chloride
left behind in the object which needs to be measured (Eggert and Schmutzler 2009).
Fortunately, the Romans left so many nails in the soil of Baden-Württemberg that the Landes-
amt für Denkmalpflege (LAD Baden-Württemberg), cooperation partner for the research
project and co-organiser of this conference, was able to provide us with enough samples.
Experimentally, measuring chloride in iron turned out much more complicated than expected
from the literature...
We are not alone
There was not much innovation or systematic research on remedial conservation of iron finds
since North and Pearson (1975). The largest problem in archaeological conservation seemed
to be neglected somewhat. But as soon as we started with our project we heard about many
other initiatives. Cardiff took up earlier work on extraction rates (Watkinson 1996) and
anaerobic desalination. As you can see from the conference programme, London became a hot
spot for research on corrosion (British Museum, English Heritage) and treatment (Museum of
London). In France, the group of Dillmann and Neff applied their highly advanced microana-
lytical tools in the ODéFA project to the effects of conservation treatments. In the US, the use
of subcritical solutions seems to open a totally new route to desalination. In Germany, the
Kulturstiftung des Bundes and the Kulturstiftung der Länder started the KUR-programme for
conservation of mobile heritage objects. One of the projects funded deals with mass treatment
of iron (Archäologische Staatssammlung Bayern). Another one (Landesmuseum für Vor-
geschichte Sachsen-Anhalt in Halle in cooperation with our Academy) tests the use of tetra-
alkyl ammonium hydroxides for desalination. They also report their results here and support
our conference. And there’s so much more…

Let's have a conference

Our idea: If there are so many research and conservation projects going on, the best thing
would be to get all groups together for a conference. Following a series of one day conser-
vation colloquia at the Academy (2005 Metals, 2006 Ceramics, 2007 Bronze, 2008 Glass)
with international participation we developed an international Call for Papers. Spreading the
news was easy: The ICOM-CC WG Metals (see this booklet for their next Metals 2010 confe-
rence in Charleston, South Carolina, Oct. 11-15, 2010) has a specialist group ‘Archaeological
Iron after Excavation’ (AIAE). They joined forces and distributed our call worldwide. The
response was outstanding (simply see the programme!): Some of our own contributions had to
be withdrawn or could only make it as posters to give enough space for our guests. And as
soon as we distributed details for registration among German archaeological conservators our
seats were outsold within few days. That’s why we are confident that our colloquium meets
the interests of the conservation community.
We do hope that the mutual exchange in lectures and discussions at the colloquium will in-
spire future research and will lead to viable solutions for the preservation of our entire excava-
ted ferrous heritage. In 1882 and today, we know the challenge:

‘Rust never sleeps!’ Neither can we.

Acknowledgement
We are grateful to all our sponsors for their support in organizing the conference:
- Staatliche Akademie der Bildenden Künste, Stuttgart
- Landesamt für Denkmalpflege Baden-Württemberg, Esslingen
- Deutsche Bundesstiftung Umwelt, Osnabrück
- Landesamt für Denkmalpflege und Archäologie Sachsen-Anhalt /
Landesmuseum für Vorgeschichte, Halle
- AIAE, sub-WG of ICOM-CC WG ’Metals’

References
Eggert, G., and Schmutzler , B. (2009), ’Lässt sich die Konservierung von Eisenfunden auf Standard
bringen?’, in U. Peltz and O. Zorn (eds) kulturGUTerhalten: Standards in der
Restaurierungswissenschaft und Denkmalpflege, Mainz: v. Zabern, 91-95.
North, N.A., and Pearson, C. (1975), `Alkaline sulfite reduction treatment of marine iron´ in Icom
committee for conservation, 4th Triennial Meeting, Venice, 13-18 October 1975. Preprints, Paris:
International Council of Museums, 75/13/ 3/ 1-14.
Scott, D.A., and Eggert, G. (2009) Iron and Steel in Art: Corrosion, Colorants, Conservation. London:
Archetype Publications.
Watkinson, D. (1996) ‘Chloride extraction from archaeological iron: comparative treatment
efficiencies’, in A. Roy and P. Smith (eds) Archaeological Conservation and its Consequences:
Preprints of the Contributions to the Copenhagen Congress, 26–30 August 1996. London:
International Institute for Conservation, 208–12.
 Extended Abstracts

Session 1:

Iron Conservation Science

PLASMA-REDUCTION, ITS POTENTIAL AND LIMITS IN THE CONSERVATION
OF METALS

KATHARINA SCHMIDT-OTT1

1
Swiss National Museum, Collection Centre, Lindenmoostr.1, CH-8910 Affoltern am Albis,
katharina.schmidt-ott@snm.admin.ch

Introduction
Since their initial application to metal artefacts in 1979, gas plasmas have been the subject of
research in conservation. The application of the plasma method in the conservation of
different metals is investigated to determine its potential and limits. Clarification of the effect
of high frequency excited plasmas on iron, silver and lead objects is covered in the paper.

Method and Materials

In a first step the treatment parameters are improved. Pure hydrogen and previously
established hydrogen-argon mixtures are investigated with optical emission spectroscopy. It is
assumed that by optimising the gas excitation the plasma operation is also optimised.
It is shown that pure hydrogen plasma treatments with relatively low gas flow increase the
radical concentration as compared to hydrogen-argon plasmas consisting of ten parts
hydrogen to one part argon. These results lead to new standard parameters for archaeological
iron objects. The new hydrogen plasmas being at least as efficient as those previously applied
but with reduced object temperatures (about 90 °C) during treatment.

Fig. 1: Iron objects during plasma treatment. The bright light

surrounding the object indicates the presence of atomic hydrogen.
Optimal settings are then assumed to be also valid for the treatment of metal artefacts. A
group of 20 Roman nails is divided into small groups and differently treated. The effects of
hydrogen-argon plasma, pure hydrogen plasma and heat treatment in vaccum are compared.
Metallographic samples from each of the nails are taken both before and after drying and/or
plasma treatment and observed under a scanning electron microscope. It is shown that only
the plasma treatment leads to the desired separation between the original surface layer of the
object and outer corrosion layers. Mechanical removal of outer corrosion layers can thereby
be more easily and precisely carried out. A separation of layers due to dehydration can be
ruled out. The facilitation of mechanical cleaning after plasma reduction is furthermore shown
in a double-blind study.

Fig. 2: SEM of a polished sample of a Roman nail after plasma

treatment. The arrow indicates the separation between outer
corrosion layers and original surface which facilitates the
mechanical cleaning of the objects.

The effect of plasma treatment on the iron corrosion products haematite, magnetite, goethite
and akaganeite is measured by micro-Raman spectroscopy. Measurements on the powdered
samples are carried out before and after the plasma treatment. The reduction of particular
corrosion products in surface layers can be confirmed. The penetration of the plasma into
different corrosion layers is investigated on archaeological artefacts. Corrosion products from
stratigrafically cleaned areas as well as metallographic samples are investigated by means of
Raman spectroscopy both before and after plasma treatment. Chemical reduction during
treatment was confirmed to be active only at the surface layer, the more deeply situated
corrosion layers having not been penetrated.
For the first time, samples are treated with pure argon plasma in order to detect a facilitation
of treatment for subsequent mechanical cleaning. A separation of layers is also observed. This
separation can be related to the high-frequency field and the displacement current in the
sample layers.
As plasma treatment carried out with comparatively low temperatures does not lead to a
removal of chlorides contained in iron objects, the method is combined with desalination in
alkaline sulphite. This combination results in a very good long-term stability of archaeological
artefacts.
Pure hydrogen plasmas are also applied to corroded historical silver artefacts. The effect of
plasma reduction is shown by means of scanning electron microscopy and X-ray fluorescence
analyses on a silver sheet with artificially produced silver corrosion layers. Silver corrosion is
removed without any damage to the surface of the artefacts.
It can be concluded that plasma reduction is a suitable method even for daguerreotypes. The
efficient removal of tarnish layers is further demonstrated on selected historical artefacts. A
combination of plasma reduction and subsequent burnishing retards new corrosion from
forming.
As for the treatment of heavily corroded lead artefacts, the effect of hydrogen and hydrogen-
argon plasmas is investigated. Two samples of the same object are compared. A reduction of
the corrosion layers and changes to surface topography can be shown by means of scanning
electron microscopy and X-ray fluorescence analyses. Penetration into deeper layers is
minimal and comparable to the penetration of iron and silver.

Discussion and Results

The optical spectroscopy of hydrogen plasmas at smaller operation pressure showed that not
only lower treatment temperatures can be reached, but also an increase of radical density is
present as compared to the hydrogen-argon treatment. The influence of the plasma treatment
on frequently occurring iron corrosion products could be made apparent in surface layers.
Investigations of surface regions of archaeological iron samples also showed that a reduction,
for example of hematite to magnetite, did take place. Examination of the stratigraphical
samples showed that the effect of hydrogen plasma in the deeper lying layers, namely the
loosening between layers, is not related to a chemical reduction but to the high frequency
field.
On silver objects hydrogen plasma leads to the reduction of corrosion and therefore the
removal of tarnish layers without any damage to the artefacts.
For more details please see reference below.

References
Schmidt-Ott, K. (2009): Erhaltung von Kulturgütern. Das Plasma in der Metallkonservierung –
Möglichkeiten und Grenzen. Collectio archæologica, Band 7. Zürich: Chronos Verlag.
THE CHLORIDE LEFT BEHIND: (DIS)SOLVING AN ANALYTICAL PROBLEM

BRITTA SCHMUTZLER1, GERHARD EGGERT1

1
State Academy of Art and Design, Am Weißenhof 1, D-70191 Stuttgart, Germany
b.schmutzler@abk-stuttgart.de; gerhard.eggert@abk-stuttgart.de

Introduction
In the research project “Rettung vor dem Rost” systematic research was done to identify the
most efficient desalination method for archaeological iron finds from the soil which
performances well under the restrictions of mass conservation: saving time and money
(funded by DBU, Deutsche Bundesstiftung Umwelt, Osnabrück/ Germany; see the contri-
bution “Simplifying Sulphite Solutions – The DBU-Project `Rettung vor dem Rost´”). During
experimental work it became clear that the analytical procedures for the analysis of desalina-
tion solutions as well as residual chloride content of desalinated objects are not standardised
in any way. In addition, for the applying conservators of alkaline sulphite desalination method
a simple test is needed for controlling the chloride content of desalination solutions.

Qualitative respectively Semi-Quantitative Chloride Determination

For quick and easy chloride analysis of desalination solutions a well-known qualitative
chloride test is recommended: the silver nitrate test (Greiff and Bach 2000). But in case of
sulphite containing solutions like alkaline sulphite it is important to work in strong acidic
conditions, e.g. adjusted with nitric acid. If not, also white silver sulphite precipitates (in
neutral and slightly acidic conditions) which is mistaken for silver chloride. For semi-
quantitative estimation, reference solutions with known chloride content can be used.
Commercial quantitative tests are not recommended, since they are in most cases still
complex in application, need special sample preparation and are in the end not precise enough
for quantitative data (e.g. mercurimetric test basing on two colour reactions as used by
Schmidt-Ott and Oswald 2006).

Quantitative Chloride Determination

In the research project, photometric analysis is used for quantitative chloride measurement of
desalination solutions and desalinated objects as suggested by Wunderlich (Wunderlich
2000). The chemical reaction for chloride determination used is the quantitatively reaction of
Hg(SCN)2 with chloride to HgCl2, and the freed thiocyanate ions (SCN-) react with Fe3+ to
Fe(SCN)3 which is of measurably orange-red colour (Florence and Farrar 1971). In both
cases, preparation of sample solution is necessary.
Determination of chloride-content of desalination solutions
For sample preparation, a solution containing 5 mol/l nitric acid and 10 % hydrogen peroxide
(oxidation of sulphite) is added in ratio 1:1 to the sample (Beaudoin and Bertholon 1994), and
allowed to stand at room temperature for 24 hours in closed bottles. The validity of
measurement was checked with alkaline sulphite solutions containing a known amount of
chloride, which were heated for 3 weeks at 55 °C and pre-treated as described. The error
associated with this pre-treatment lies in a range up to 10 %, which is acceptable considering
the many excessive interfering ions in the analytes.

Validation of Digestion Strategy and Residual Chloride Analysis

In conservation literature, the focus was on the analysis of desalination solutions so far, but
only little attention was paid on the analysis of the residual chloride in desalinated objects.
However, Ellis et al. report total chloride loss in the form of volatile hydrogen chloride,
caused by excessive acid when solving akaganéite powder in sulphuric acid (H2SO4, 20 %
w/w; Ellis et al. 1976). This problem was not considered in conservation science, but has
direct impact on our research – what, if the sample preparation biases the desalination effi-
ciency results by expelling chloride? Therefore, firstly the own analytical procedure was to be
validated, and secondly, checking of published digestion strategies was necessary.
For validation measurements, iron powder was digested in sulphuric acid (20 %) with added
sodium chloride solution. After dissolution, Fe2+ ions were oxidized to Fe3+ ions by hydrogen
peroxide (30 %), and excessive H2O2 was expelled by boiling for 30 minutes. Subsequently,
the sample was prepared as given by the protocol for photometric analysis (Florence and
Farrar 1971), and directly measured. Three validation measurements gave a mean of 1.10
mg/l chloride (standard deviation of ± 0.042) in contrast to the given reference value of 1.0
mg/l chloride. This difference of about 0.1 mg/l can be expressed as error of 10 %, which is
again acceptable for the given reasons.
The protocol of Drews et al. (2004) was reproduced with sulphuric acid to check potential
chloride loss by digestion applying the mentioned standard, protocol, and analytical method.
Measurements show that hot digestion for 48 hours without any covering as suggested by
Drews et al. causes loss of chloride from the iron powder–sodium chloride mixture (Drews et
al. 2004; for detailed problems concerning other digestion strategies and chloride analysis see
Schmutzler and Eggert 2010).

Residual Chloride Content Analysis

For evaluation of the desalination experiments, the desalinated specimens (Roman nails) were
individually cut into small pieces. This allows digestion of two representative, small batches
of the objects for acceleration of the solubilization process, avoiding loss and saving sulphuric
acid. The mean of chloride content is determined in ‰.
In general, working with archaeological objects makes accurate analysis difficult and signi-
ficant differences between desalination strategies are hard to work out. The data quality of all
experiments is expressed in standard deviation and variation coefficient, which expresses in
turn the standard deviation as a percentage of the mean. All in all the variation coefficient of
around 400 residual chloride measurements is in average around 16 %. Taking into conside-
ration that small amounts of chloride ions among excessive interfering ions are to be
measured, this is definitely a satisfying value.

Conclusion
This outline shows how important every step in analytical procedure is. It is also to be kept in
mind that long hot and open digestion causes loss of chloride and, therefore, manipulates the
desalination efficiency data.

Acknowledgements
We would like to thank Dr. J. Peters (H.C. Starck/ Goslar), Dr. J. Opitz (Institute of Organic
Chemistry/ University of Stuttgart) and Dr. M. Völker (fem/ Schwäbisch Gmünd) for helpful
discussions and analyses. We are grateful to Landesamt für Denkmalpflege Baden-Württem-
berg/ Esslingen and Deutsche Bundesstiftung Umwelt DBU/ Osnabrück for their support.

References
Beaudoin, A., and Bertholon, R. (1994), `Le Dosage des ions chlorure dans le sulfite alcalin´, Les
Cahiers Techniques de l´ARAAFU 1 : 39-48.
Drews, M., de Viviés, P., González, N., and Mardikian, P. (2004), `A study of the analysis and
removal of chloride in iron samples from the Hunley´, in Metal 2004, Proceedings of the Int.
Conference on Metals Conservation, Canberra, Australia, 4-8 October 2004, Canberra: National
Museum of Australia, 247-260.
Ellis, J., Giovanoli, R., and Stumm, W. (1976), `Ion exchange properties of β-FeOOH´, Chimia 30(3):
194-197.
Florence, T.M., and Farrar, Y. (1971), `Spectrophotometric determination of chloride at the parts-per-
billion level by the mercury (II) thiocyanate method´, Analytica Chimica Acta, 54: 373-377.
Greiff, S. and Bach, D. (2000), `Eisenkorrosion und Natriumsulfitentsalzung: Theorie und Praxis´,
Arbeitsblätter für Restauratoren, Gruppe 1: 319-339.
Schmidt-Ott, K., and Oswald, N. (2006), `Neues zur Eisenentsalzung mit alkalischem Sulfit´, Beiträge
zur Erhaltung von Kunst- und Kulturgut 2: 126-134.
Schmutzler, B., and Eggert, G. (2010), `Chloride calamities: assessment of residual chloride analysis
and comparison of iron desalination methods´, manuscript submitted for the ICOM-CC Metal
Conference 2010, Charleston, South Carolina
Wunderlich, C., (2000)´Archäosideroprophylakt - Entsalzung von Eisenfunden mit Hydroxylamin´,
Jahresschrift für mitteldeutsche Vorgeschichte 83: 305-316.
FORMATION AND TRANSFORMATION OF AKAGANEITE

DAVID THICKETT1, MARIANNE ODLYHA2

1
English Heritage, Collections Conservation Team, Rangers House, Chesterfield Walk,
London, SE108QX, david.thickett@english-heritage.org.uk
2
Birkbeck College, London, m.odlhya@bbk.ac.uk

Formation of Akaganeite
The formation of akaganeite and its physical damage has been ascribed as the major
mechanism for the deterioration of archaeological iron artefacts. Several RH thresholds were
determined experimentally and presented at the first meeting of the Archaeological Iron Sub
Working Group. Below 16% RH no akaganeite formation was observed. Below 30% the
formation was very slow. Neither of these thresholds appeared to have a strong temperature
dependence. At RHs above 55%, and decreasing with increasing temperature, iron dissolution
becomes the major mechanism. These thresholds are critical to preventive conservation
practice to stop the deterioration of archaeological iron in storage or limit its deterioration on
display (where low RHs are extremely expensive to achieve). Watkinson showed that iron and
iron (II) chloride mixtures with artificially produced akaganeite have a lower threshold of
11% (Watkinson 2002). This is extremely important for storage. The initial experiments
reported at the last meeting had generated a series of akaganeites formed at different RHs;
from 30 to 90% in 10% intervals. These were mixed with iron and iron (II) chloride powders
to form 1:1:1 mixtures which were then exposed to RHs between 10 and 20% at 1% intervals.
The powder mixtures from the lower RH experiments still contained some residual iron and
iron (II) chloride. This was quantified by thermomagnetometry and the amounts modified
accordingly (Thickett and Odlyha 2003). The RHs were generated above glycerol solutions.
After twelve months exposure the powders were analysed with FTIR as KBr disks to quantify
the amount of akaganeite (Thickett 2004). Whilst the threshold RH was lowered to 15% with
the presence of the akaganeites formed at 80 and 90% RH, no effect was observed with the
akaganeites formed at lower RHs. Watkinson’s work used akaganeite precipitated from
acidified iron (III) chloride solution. His conclusions are valid for the system being studied,
i.e. a ship where the akaganeite will have formed at ambient outdoor RHs which are often in
excess of 80%. However most archaeological iron artefacts are stored in dry silica gel from
excavation so are unlikely to have experienced RHs above 50% even with very dilatory
reconditioning of the silica gel. Hence the 16% threshold (in the presence of copper ions) is
appropriate for archaeological iron.

Several volatile organic compounds are reported to accelerate the corrosion of iron or carbon
steel. In a museum storage or display environment the carbonyl compounds; ethanoic and
methanoic acid and methanal are those most commonly encountered (Rance 1956, Grzywacz
1994). Since ethanoic acid is the most damaging and most common, its effects on the
akaganeite formation reaction were investigated. A range of ethanoic acid concentrations
were used to mimic situations frequently encountered;

zero good storage, Stewart boxes (not in wooden cupboards)*

750 good quality showcase, wood products used for base-board and back-board only
and Moistop wrapped or Selabond coated to reduce emissions
1500 showcases with some wood products with no emmision barriers or with Dacrylate
103 coatings
3000 showcases and cupboards made of wood or wood products
6000 poor quality cardboard boxes, paper envelopes,

* Larkin et al reported Stewart boxes failing an accelerated corrosion test with lead, the author
has undertaken over twenty such tests on Stewart boxes purchased over the past fifteen years
without one failing any accelerated corrosion test.

Ethanoic acid atmospheres were generated above the appropriate ethanoic acid solutions
(Merck Analar). These were added to saturated magnesium chloride, magnesium nitrate and
sodium chloride solutions to give RHs of 33.1-33.6, 52-56 and 75.2-75.7%, following the
method developed by Tetrault (Tetrault 1999). Brockerhof et al had reported significant non
linearity in acetic acid concentrations generated above acetic acid and salt mixtures, in
contradiction of Tetrault's findings. The ethanoic acid and salt where also prepared in 250ml
jars and the airborne acid concentration determined for four consecutive 14 day periods with
duplicate diffusion tubes (Gibson et al 1996). This confirmed that the concentrations were
within the 7% instrumental error of the diffusion tube analyses for the two month period. No
systematic decrease in ethanoic acid concentration was detected (t-test at 99% confidence
interval) over the test period. The use of an impermeable barrier on the reaction vessel top
negated the need to change the solutions monthly. The differences between Tetraut's and
Brockerhoff's results are most likely due to different reaction vessels used, particularly in the
permeabilities of the lid materials and sealing of the lids. The ethanoic acid solution
concentration was checked by analysing the solution when the tests were complete with ion
chromatography (Dionex DX200 with ICE column and 18mM borax eluent), confirming very
little loss of ethanoic acid during the tests.

The amount of akaganeite formed was expressed as a percentage of that formed with no
ethanoic acid present. Results are shown below.
 350
compared to clean environment
percentage of akaganeite formed

300 33%
54%
250
75%
75% total
200

150

100
0 1000 2000 3000 4000 5000 6000 7000
ethanoic acid concentration (µg/m3)

All the tests, with the exception of the 75% RH tests at 3000 and 6000ug/m3, produced more
akaganeite. Geothite was obvious in the infra-red spectrum at 3000ug/m3 and 75% and was
the major product at 6000ug. Close examination of the spectrum also revealed some geothite
at 1500 ug/m3. To take this into account, samples were taken from the 75-76% tests and
analysed with TGA to quantify the amounts of akaganeite and geothite present. These results
are shown as 75-76% total above. The formation of goethite at 75% RH and higher acid
concentration is possibly due to the decreased pH during Fe(II)Cl oxidation (Misawa 1974).
An alternative explanation could be initial formation of akaganeite, which is then dissolved by
the lower pH solution and re-precipitates as geothite.

Archaeological iron can be an extremely complex system with many ions and species present.
The experiments have used iron and iron (II) chloride mixtures as model systems. To validate
the results, a survey was undertaken of the 600 or so archaeological iron artefacts on display
throughout English Heritage’s sites. English Heritage has over 50 sites displaying
archaeological iron and the RH in the rooms is highly variable from site to site. Additionally
the showcases used have been built over a ninety year period and are of highly variable
quality. Some can keep RHs below 30%, whilst many can only achieve below 55% with silica
gel and some have no silica gel facilities and experience RHs up to 80%. The survey was
carried out using a criteria anchored methodology (Sully et al 1999). Results are shown
below.
 No ethanoic acid
100% 2 1 1
1 1 1
90% 32
1
80%
percentageof objects

23
10
70% 5
60% 3
50% 97
40% heavy 2
105
30% medium 31 15
20% slight 4
4
10% none
0%
20-30 30-40 40-50 50-60 60-70 70-80
RH ra nge

Ethanoic acid
100% 3
9 7 3
90% 4
12 10 14
80%
percentageof objects

heavy 7
70% 31
medium
60% 25
slight 22
50% none
40% 10
26
30%
43
20% 31 24
10% 4
0%
20-30 30-40 40-50 50-60 60-70 70-80
RH ra nge

As can be seen, the 30% and 55% thresholds appear to hold with real objects.

Conversion of Akaganeite
The conversion of akaganeite to other iron minerals will almost certainly release chloride
back into the object as no other iron mineral contains the same concentration of chloride.
Examination of akaganeite crystals formed, in two situations, revealed very thin yellow layers
on the surface of the crystals. These layers were identified with Germanium ATR FTIR as
Geothite and the lower chloride concentration of the layer compared to the akaganeite
confirmed with SEM-EDX. The conversions had occurred at RHs between 21 and 54% (poor
quality showcases with silica gel changed every six months) and in storage in an
unconditioned store experiencing RHs between 43 and 61% in the cardboard storage boxes.
Both situations had significant concentrations of ethanoic acid. In the cases this originated
from their particleboard construction and in the store from the cardboard boxes.
Stahl has written that there is no release of chloride from akagaleite below 200°C (Stahl
2003). Such isothermal heating experiments are known to underestimate the rates of
transformation for solid state reactions. Misura et al have shown that in aqueous media
akaganeite relatively rapidly transforms into goethite and it is possible that the high
hygroscopicity of the adsorbed chloride on the akaganeite surfaces could provide enough
water for this reaction to occur. Stahl’s work was carried out under nitrogen with a nominal
RH of 0%.
A method has recently been developed to model the kinetics of solid state reactions and
predict the reaction rates under different ambient thermal profiles. Thermogravimetric
analyses were undertaken of both synthetic akaganeite and a relatively pure akaganeite
sample removed from a anglo-saxon coffin clamp. The synthetic sample was prepared from
arial oxidation of ferrous chloride (Aldrych Analar 99.98%) at 60ºC for sixteen hours (Cornell
2003). The samples were coarsely ground between glass slides, and analysed with a Perkin
Elmer TGA7 under 40ml/min of nitrogen at 1,2,5,10, 15 and 20°C/min heating rate. The
reaction rate was fitted allowing both the pre-exponetial factor and activation energy to vary.
The modelling indicated that slow conversion to hematite was likely over a number of years.
Samples of Akaganeite taken from objects over the past twenty five years and analysed
mainly with XRD were reanalysed with Raman. Although XRD is relatively insensitve to
Hematite in Akaganeite each sample was examined under microscopy and the appearance
recorded. Visual examination revealed red spots on many of the earlier samples and Raman
verified that these were Hematite.

References
Brokerhof, A.W., M. van Bommel, (1996) Deterioration of Calcareous Materials by Acetic Acid
Vapour: A Model Study." ICOM Committee for Conservation 2, 769-775.
Cornell, R.M., U. Schwertmann, (2003) The Iron Oxides 2nd Edition, Wiley VCH, New York.
Grzywacz, C.M., N.H. Tennent, (1994) Pollution monitoring in storage and display cabinets,
Preventive Conservation Practice, Theory and Research, A. Roy, P.Smith eds, IIC, 164-170
Larkin, N., N. Blades, (2000) Investigation of volatile organic compounds associated with
polyethylene and polypropylene containers used for conservation storage. The Conservator 24, 42-51.
Misawa, T., K. Hashimoto, (1974) The mechanism of formation of iron oxide and oxyhydroxides in
aqueous solutions at room temperature, Corrosion Science 14, 131-149.
Rance, V., H.G.Cole, (1958) Corrosion of metals from organic acids: a survey, HMSO, London, 17.
Roduit, B. (2000) Computational aspects of kinetic analysis: Part E: The ICTAC Kinetics Project –
numerical techniques and kinetics of solid state processes, Thermochimica Acta 355(1-2), 171-180.
Roduit, B. (2002) Prediction of the progress of solid-state reactions under different temperature
modes, Thermochimica Acta 388(1-2), 377-387.
Tetreault, J., J. Sirois, (1998) Studies of lead corrosion in acetic acid environments, Studies in
Conservation. 43, 17-32.
Thickett, D. (2005) The use of Infra-red and Raman Spectroscopies for Iron Corrosion Products,
Postprints of Sixth Infra-Red Users Group, Florence, Il PratoElsevier, Padua, 86-93.
Thickett, D., M. Odlyha, (2005) Application of thermomagnetry to corrosion studies of archaeological
iron, Journal of Thermal Analysis and Calorimetry, 80 (3), 567-571.
Watkinson, D., M. Lewis, (2004) SS Great Britain iron hull: modelling corrosion to define storage
relative humidity. Proceedings of the international conference on metals conservation, Metals 04, 88-
102
‘UNUSUAL’ CORROSION PRODUCTS OF IRON

QUANYU WANG1

1
Department of Conservation and Scientific Research, The British Museum, Great Russell
Street, London WC1B 3DG, UK qwang@thebritishmuseum.ac.uk

Introduction
Post excavation corrosion of archaeological iron has been widely studied and most of the
research has been focused on akaganeite and iron oxyhydroxides. A wide range of the
corrosion products of iron has recently been summarised by Scott and Eggert (2009). It is the
author’s intention here to report on some of the corrosion products which had previously not
been widely reported. The formation of these “unusual” corrosion products is discussed and
questions are addressed.

Jarosite
A green/yellowish powdery corrosion product was observed on some archaeological iron
objects in storage at the museum as well as on freshly excavated iron objects. The Raman
spectrum of the corrosion products (Figure 1) indicated that it is probably a member of the
jarosite group, i.e. MFe3(SO4)2(OH)6 compounds, where M = K, NH4, Na, etc. It was
identified as ammoniojarosite, NH4Fe3(SO4)2(OH)6 by a combination of X-ray fluorescence
spectroscopy (XRF) and X-ray diffraction (XRD).

Fig. 1: Raman spectrum of the green/yellowish powdery corrosion product of iron identified as
ammoniojarosite

A light yellow powdery material was found on iron knives from the Sutton Hoo 2000
excavation. It was identified as potassium jarosite, KFe3(SO4)2(OH)6, by elemental analysis
using scanning electron microscopy equipped with energy dispersive X-ray spectrometer
(SEM-EDX) and by XRD. It needs to be noted that the Raman spectra of different materials
from the jarosite group have little difference to distinguish between them and the XRD data
for these two members of the group are also very similar, therefore, elemental analysis is
important in the identification of these materials.
How did these sulphates form, from burial or post excavation? These sulphates could have
been formed during burial, because they were found on freshly excavated iron. This does not
exclude the possibility of post-excavation corrosion, because the sulphates were sometimes
present underneath flakes and in cracks. Sulphur arises from decaying organic matter in
burials, from industrial pollutants or from display materials, and the oxidation of sulphides
can take place in air to form sulphates. Ammonia and potassium can come from burials and/or
agricultural fertilisers. These members of the jarosite group are considered to be stable
because of their widespread occurrence in nature (Smith and Lampert 1973).

Iron oxyhydroxides resulted from desalination treatment

Desalination treatment of archaeological iron for the removal of chloride is one of the
approaches taken in the preservation of the objects. A bright yellow corrosion product, which
often has a ‘cotton ball’ morphology, was observed on some of the iron objects treated in
desalination solutions of either sodium hydroxide or alkaline sulphite (Wang 2007a). It was
identified as goethite, α-FeOOH, by Raman spectroscopy. A yellow/brown corrosion product
was also found to have formed on archaeological iron nails during rinsing in deionised water
following desalination in alkaline sulphite solutions. This corrosion product, which has a
loose powdery structure, was identified as lepidocrocite, γ-FeOOH, by Raman spectroscopy.
How were these corrosion products formed? The formation of goethite and lepidocrocite,
instead of akaganéite, could be the result of atmospheric oxidation promoted by residual OH-
from incomplete washing after desalination treatment and may be due to a low chloride ion
concentration. Research carried out by Refait and Genin (1997) suggests that only akaganéite
can form at high Cl- ion levels ([Cl-]/[OH-] ≥ 8) and only goethite forms at low Cl- ion levels
([Cl-]/[OH-] ≤ 6).
Can these corrosion products cause any further deterioration of the objects? Goethite and
lepidocrocite are chemically stable as corrosion products. However, they have been found to
grow perpendicular to the surface on some objects which can cause cracking and flaking on
the surface. Further studies are needed to investigate in what conditions or at which stages
these materials form and how to avoid their formation in the desalination process.

Elongated akaganeite
It has been found from experiments with iron coupons that the corrosion morphology seems
to be affected by relative humidity (RH) levels. Elongated akaganeite was likely to have
formed at RH levels between 50 and 60% as the result of atmospheric oxidation, and weeping
was more likely to occurr at higher RH levels (Wang 2007b). It is the elongated akaganéite
cryastals that often caused the flaking of archeological iron (Fig. 2 & 3). The RH of 54% was
achieved by a saturated Mg(NO3)2.6H2O solution. Comparison of iron coupons exposed,
respectively, to the nitrate and glycerol solution which was used to achieve the same RH level
showed that the nitrate seemed to accelerate the formation and growth of the elongated
akaganeite crystals. This raises the following questions: are other pollutants such as SO2, NO2
harmful to iron objects? Is the formation of elongated cerytalline akaganeite, which is often
observed on archaeological iron, promoted by the presence of nitrate in the enviroment? The
effect of RH levels and pollutants including nitrogen oxides and nitrate on the corrosion
morphology needs further research.

Fig. 2 Elongated akaganeite crystals formed Fig. 3 Flaking of an Anglo-Saxon iron axe
on an iron coupon at 54%RH caused by the formation of elongated
akaganeite

References
Refait, P., and Genin, J.-M.R. (1997) ‘The mechanism of oxidation of ferrous hydroxychloride β-
Fe2(OH)3Cl in aqueous solution: The formation of akaganéite vs. goethite’, Corrosion Science 39 (3):
539–553.
Scott, D.A., and Eggert, G. (2009) Iron and Steel in Art: Corrosion, Colorants, Conservation. London:
Archetype Publications.
Smith, W.L., and Lampert, J.E. (1973) ‘Crystal data for ammoniojarosite, NH4Fe3(SO4)2(OH)6’,
Journal of Applied Crystallography 6: 490–491.
Wang, Q. (2007a) ‘An investigation of deterioration of archaeological iron’, Studies in
Conservation 52 (2): 125-134.
Wang, Q. (2007b) ‘Effects of relative humidity on the corrosion of iron: an experimental view’,
British Museum Technical Research Bulletin 1: 65-73.
METASTABLE IRON SULPHIDES AS CORROSION PRODUCTS OF
ARCHAEOLOGICAL IRON OBJECTS

CELINE REMAZEILLES1, DELPHINE NEFF2, ELODIE GUILMINOT3, SOLENN

REGUER4, PHILIPPE DILLMANN2, PHILIPPE REFAIT1.

1
Laboratoire d’Etude des Matériaux en Milieux Agressifs, Université de La Rochelle, Avenue
Michel Crépeau, F-17042 La Rochelle cedex 01, France.
celine.remazeilles@univ-lr.fr, philippe.refait@univ-lr.fr
2
LAPA/SIS2M UMR3299 CEA/CNRS et IRAMAT UMR5060 CNRS, CEA Saclay, F-
91191 Gif sur Yvette cedex, France
delphine.neff@cea.fr, philippe.dillmann@cea.fr
3
EPCC Arc’Antique, 26 Rue de la Haute Forêt, F-44300 NANTES, France
elodie.guilminot@arcantique.org
4
Synchrotron SOLEIL, L'Orme des Merisiers, BP 48, F-91192 Gif-sur-Yvette cedex, France
solenn.reguer@synchrotron-soleil.fr

Introduction
Archaeological iron objects buried in anoxic conditions can be exposed to microbiologically
influenced corrosion (MIC). In some cases, this form of corrosion can be extremely severe.
The most common process is associated with the development of sulphate reducing bacteria
(SRB) colonies on the metal surface, and requires anoxic conditions. The metabolic activity of
such micro-organisms involves the reduction of sulphates, naturally present in the
environment, into sulphides. Sulphides react with Fe2+ ions, which leads to the iron(II)
sulphide mackinawite (FeS) in the rust layer. Such corrosion processes are often localised and
lead to the formation of broad pits in the metal.
However, the works dealing with MIC were mainly focused on modern steels. In the case of
archaeological ferrous materials, the influence of microorganisms can be suspected similarly
when iron sulphides are detected in the rust layer. Sometimes restorers testify of a specific
odour of “sulphur”, namely hydrogen sulphide (H2S), while MIC was never mentioned as a
possible danger for conservation of archaeological objects. This report is in contradiction with
current studies on MIC, assuming a rapid decay of iron objects when microorganisms
influence the corrosion mechanisms. As exposed in this presentation, iron(II) sulphides like
mackinawite and greigite (Fe3S4) and even SRB have been detected in rust layers of
archaeological iron artefacts although these ones presented no alteration patterns (broad pits)
characteristic of MIC.
Iron(II) sulphides are unstable once exposed to atmosphere. The consequences of their
transformation after excavation have never been clearly demonstrated. So this presentation is
focused on the identification, localisation and evolution of iron(II) sulphides in rust layers of
archaeological iron artefacts. After a short state of art, analysis performed on artefacts
extracted from different anoxic media (bottom of the sea and soil) is presented. For some of
them, namely Roman ingots, dechlorination treatments have been applied (cathodic
polarisation and immersion in a NaOH solution) after which remaining iron/sulphur
containing compounds were detected.

Evolution of iron sulphides in excavated artefacts

Previous studies devoted to the analysis of iron sulphides in archaeological iron artefacts are
scarce. Among all naturally occurring iron/sulphur containing compounds (table 1), only
mackinawite, greigite and framboidal pyrite (FeS2) were mentioned. For instance, these were
identified in rust layers of artefacts extracted from a waterlogged soil (Fiskerton, UK)
(Fell/Ward 1998; Fell/Williams 2004).
General oxidation processes of iron sulphides are widely studied and several mechanisms
deduced from works realised on laboratory systems and synthetic products are proposed in
literature. Parameters like the oxic/anoxic nature of the environment, sulphur supply via
hydrogen sulphide production, temperature, humidity, pH, etc. seem to determine one
pathway or another, leading from mackinawite either to the formation of pyrite or to the
decomposition into sulphur and iron(III) oxyhydroxides. A recent work focused on the
analysis of synthetic mackinawite by Raman spectroscopy demonstrated that this phase can
present three physico-chemical states (nanocrystalline, well crystallised and Fe(III)-
containing mackinawite), each of them being differentiated by a characteristic spectrum
(Bourdoiseau 2008).

Table 1 : Iron/sulphur containing compounds

Mineral Formula Oxidation state
Mackinawite FeS Fe(+II), S(-II)
Troilite FeS Fe(+II), S(-II)
Pyrrhotite Fe1-xS Fe(+II), S(-II)
Greigite Fe3S4 Fe(+II,+III), S(-II)
Smythite Fe9S11 Fe(+II,+III), S(-II) (?)
Pyrite FeS2 Fe(+II), S2(-II)
Marcasite FeS2 Fe(+II), S2(-II)
Sulphur -S8 S(0)

In our study, corrosion layers of 16th century old nails coming from a terrestrial site (Glinet,
France) and of Roman ingots excavated from a wreck off Les Saintes Marie de la Mer
(France) were analysed by combining several methods: scanning electron microscopy (SEM)
coupled with X-ray microanalysis and Raman micro-spectroscopy. In each sample,
mackinawite was detected, and its three physico-chemical forms could be observed. The
corrosion layers were mainly composed of iron(II) compounds such as siderite (FeCO3),
chukanovite (Fe2(OH)2CO3) and -Fe2(OH)3Cl, confirming the anoxic conditions of both
burial environments. Such conditions are favourable for SRB development. As far as the
microbiological activity of the soil of Glinet is concerned, SRB were actually detected in the
rust layers of nails. However as for the nails or the ingots, the sulphur rich zones were located
in the external part of the rust layers, meaning that the microbiologically influenced
phenomenon do not imply the metallic substrate but rather the corrosion products. This result
is not consistent with the classic MIC mechanisms currently admitted. In consequence it is not
obvious that the mechanisms proposed for explaining MIC processes of modern materials can
be transposed to the archaeological samples that we studied. This point will be discussed in
the presentation. Greigite was also identified but its presence is rather explained by the
oxidation of initially present mackinawite during -Raman experiments, which were
performed without protection against oxygen.
In the frame of the French research program ODeFA, which was mainly dedicated to the
understanding of dechlorination mechanisms of corrosion layers during treatments, some
Roman ingots were treated. Cathodic polarisation in a KOH bath and immersion in a 2%
NaOH bath were applied. The analysis of ingots at the end of treatments revealed the presence
of remaining iron/sulphur containing compounds. Orthorhombic elemental sulphur ( -S8)
was identified in large amount in the rust layer of the ingot treated by cathodic polarisation,
while Fe(III)-containing mackinawite was found in the rust layer of the ingot treated in the
NaOH solution. In the first case, the presence of elemental sulphur can be explained as the
result of the oxidation of mackinawite that led simultaneously to goethite. In the second case,
the presence of Fe(III)-containing mackinawite in the middle of a layer mainly composed of
goethite is more difficult to explain.

Consequences for conservation

In most cases the superficial concretion including exogenous compounds is removed during
the treatments. Iron sulphides being located rather in the external part of the rust layer, a large
amount is certainly removed during this stage. However, the example of the ingot treated in a
NaOH solution has shown that when they are located deeper inside the rust layer, they remain.
In this case, they can transform at any moment. In wet organic archaeological materials,
consequences of iron sulphides oxidation are very dangerous (embrittlement and cracking).
For ferrous materials, the impact for conservation is not clearly evaluated yet.

References
Bourdoiseau, J.A. et al. (2008), Corrosion Science, 50, 3247-3255.
Fell, V. and M. Ward (1998) Proceedings of the Interim Meeting of the ICOM-CC Metal WG,
Draguignan, 111-115.
Fell, V. and J. Williams (2004) Proceedings of the Interim Meeting of the ICOM-CC Metal WG,
Canberra, 17-24.
Gerhard Eggert and Britta Schmutzler (Eds.)

Archaeological Iron
Conservation Colloquium 2010

Extended Abstracts
 Archaeological Iron
Conservation Colloquium 2010

State Academy of Art and Design Stuttgart

24th to 26th June 2010
 Extended Abstracts

Session 2:

Iron Conservation Projects

around the World
ABOUT CORROSION AND CONSERVATION PROBLEMS OF IRON ARTEFACTS
COMING FROM ORADEA FORTRESS

OLIMPIA MURESAN1

1
“Tara Crisurilor” Museum, Dacia Boulevard no1-3, Oradea, Romania
olimpia_muresan2010@yahoo.com

Introduction
The today Oradea fortress was built during the XVI-XVII centuries. Ruins of buildings (the
XIII-XVII centuries), artefacts and fragments (discovered during archaeological diggings
1991-2001, 2005-2007, 2009) are a great challenge for studyand conservation nowadays
(Rusu et al. 2002; Muresan 2009). A special attention will be paid to the corroded iron
artefacts.

General considerations about archaeological iron corrosion

“Rust” is associated with iron corrosion which in the language of chemistry is called
oxidation. During the burial period the iron undergoes ionization as a consequence of the
electrochemical phenomena: the anodic process consisting of iron oxidation is balanced by
the depolarization process at the cathodic areas. The depolarization depends on the parameters
of the environment.
It is generally known (Selwyn 2004) that the main phases of rust are oxides and
oxyhydroxides such as: maghemite (γ-Fe2O3) magnetite (Fe3O4), goethite, α-FeOOH), and
lepidocrocite γ-FeO(OH). In addition to the above mentioned corrosion products, other
specific ones may be identified due to the particular soil/environmental
composition/parameters (soil pH, soluble salts, aeration, soil constituents, and synergic
effects).

Iron artefacts from Oradea fortress

The soil of the fortress may be considered near-anoxic and carbonated, because of a stone
pavement from the 18th century that was found at a deeper depth than 20 cm and the other
ruins of ancient buildings present there, too. Therefore, carbonates (Saheb 2007) were
expected to be present. But things are not so simple: by means of ATR-IR [analysis realized at
National Researching Institute for Conservation and Restoration, Bucharest with Bruker
Experimental IR-ATR (Diamond, Helios XPM, 5000-400cm-1, 4cm-1 resolution)], organic
components (aromatic structures and probably oil), kaolinite, quartz, carbonates, and hydrated
water were detected too. The heterogeneity of the soil caused a lot of problems in attributing
the peaks and discriminating the components (for the iron corrosion components, too).
Now, speaking about the iron artefacts coming from Oradea fortress, each digging brought up
hundreds of objects or fragments, with different degrees of mineralization and deformation,
even without possibility of being identified. Their conservation requires solving a lot of
problems. Great battle, few weapons! A strategy may be a useful tool. It is important to
acknowledge: which are the problems to be solved; which are the ideal requirements to
conserve the discovered artefacts and the realistic ways/possibilities to do that. The problems
to be resolved are the following: a lot of metallic artefacts were discovered during each
excavation; the iron ones were corroded and mineralized (most of the artefacts looked
fragile); the artefacts must be stabilized and conserved; no supporting expenses for the
research of the artefacts. To divide the iron artefacts, from the beginnings (by examination
tools), in categories based on their emergency degree and degrees of mineralization seems
time gaining.
Some of badly corroded iron artefacts from Oradea fortress didn’t lost the original shape. It is
the case of a fragment of medieval steel armour that was cremated. Details of the morphology
of the corrosion products were observed by optical microscope (magnification 50x, 100x).
Spectral analysis was performed by ATR-IR on iron compound. Standard spectra were used.
Goethite, lepidocrocite, iron oxide, calcite, were detected by both methods. But the research
must to be continued (complementary and non-destructive tools are necessary (Neff et al.
2004).
Height surface sensitive techniques as IRRAS (infrared reflection absorption spectroscopy,
QCM (quartz crystal microbalance) and TM-AFM (tapping-mode atomic force microscopy)
were used (Kleber et al. 2007) for studying initial atmospheric corrosion of iron, copper and
silver.
Now, topography data are linked to the chemical information on the same sample area. The
Interdisciplinary and Experimental Research Institute from Cluj-Napoca is equipped with a
WITec instrument, an Alpha 300A with Confocal Raman Alpha 300R, that offers the
possibility to acquire non-destructive chemical information with a resolution down to the
optical diffraction limits (200nm). All this sounds great! It makes the tool an ideal one! But
there is a trap here! The surface of the artefact is a rough one and this equipment needs a
smooth surface at a nanoscale level. For the moment, the tests made on an iron artefact
sample, archaeological glass, polished bone have no significance. Then it is necessary to bring
the sample parameters compatible to the equipment!

Final considerations
The conservation of great quantities of excavated iron artefacts needs particular management
strategy. The examination methods are used on every artefact. Complex analysis is done to
representative or particular artefacts. Samples of soil, corrosion products, and fragments of
artefacts may be collected for further analysis. When new scientific tools are to be tested for
studying the artefacts a first stage must be a detailed preliminary analysis.
The analysis techniques at nanoscale (molecular!) level may be used for a better conservation
of artefacts. To “see” a molecular structure is a scientific success or an intrusion!?!
Acknowledgments
The author offers her thanks to the contributions to this paper by Balta Zizi (IR-ATR spectra)
and Dr. Astilean Simion (AFM preliminary test).

References
Derrick, M.R., Stulik, D., Landuy, J.M., (1999) Infrared Spectroscopy in Conservation Science, Getty
Conservation Institute.
Kleber, C., Hilfrich, W., Schreiner, M., (2007) “Surface sensitive analytical methods (TM-AFM,
IRRAS, QCM) for the in-situ investigation of the metal-atmosphere interface during the early stage of
atmospheric corrosion” in Degrigny, Ch., van Langh, R., Joosten, I., Ankersmit, B., Eds. Metal
07:Prerint of the Interm meeting of the ICOM-CC Metal WG Amsterdam, 17-12 sept,.Amsterdam, (2),
42-46.
Muresan O., (2009) CETATEA ORADEA REDIVIVUS (I).Artefacte metalice intre descoperire si
valorificare, Editura Arca.
Neff.D., Reguer, S., Bellot-Gurlet,L., Dillmann,Ph.,Bertholon, R., (2004) Structural characterization
of corrosion products on archaeological iron. An integrated analytical approach to establish corrosion
forms, in Journal of Raman Spectroscopy, 35, 739-745.
Rusu, A.A., and Co., (2002) Cetatea Oradea. Monografie Arheologica (Zona Palatului Episcopal),
Editura Muzeul Tarii Crisurilor Oradea.
Saheb, M., Neff, D., Dillmann, Ph., Matthiesen, H., (2007) “Long term corrosion behaviour of low
car, bon steel in anoxic soils”, in Degrigny, Ch., van Langh, R., Joosten, I., Ankersmit, B., (eds) Metal
07 Preprint of te Interm meeting of the ICOM-CC Metal WG Amsterdam,17-21 sept, Amsterdam, (2),
69-75.
Selwyn, L., (2004) “Overview of archaeological iron: the corrosion problem, key factors affecting
treatment, and gaps in current knowledge” in Ashton, J. and Hallam (eds) Metal 2004 : Proceedings of
interim meeting of the ICOM-CC Metal WG, National Museum of Australia Canberra, 294-306.
THE „CUCAGNA CONSERVATION PROJECT“ - CONSERVATION OF
ARCHAEOLOGICAL FINDS AND CONSERVATION RESEARCH IN FRIAUL

TOBIAS FRIEDRICH1

1
Eythstraße 63, 51103 Köln (Cologne), Germany, tobias.friedrich@mfak.de

Introduction
Subject of the archaeological research project in which my conservation project takes place
forms/constitutes the systematical exploration of the remains of the medieval castle Cucagna
in North-Italy.
Since 2001, every summer for a period of four weeks excavations are taking place at
Cucagna.
The excavation licence is held by Prof. Dr. Sebatian Brather of the Albert-Ludwigs-University
Freiburg while Holger Grönwald M.A. leads the excavation.
The excavations accompany the reconstruction and restauration of the architectural remains of
the castle and are organised by the Istituto per la Ricostruzione del Castello di Chucco-Zucco
(IRCCZ) and conducted during an international seminary (Seminario Estivo Internationale di
Architettura Medioevale). The seminary is led by Dott. Arch. Roberto Raccanello and Dott.
Arch. Katharina von Stietencron.
Initiated in 2007 and carried out since 2008, the excavations and reconstruction are
supplemented by the section „conservation of finds“. This has only been possible by
permission and support of the Soprintendenza Archeologica e per i beni Ambientali,
Architettonici, Artistici e Storici del Friuli Venezia-Giulia.
First and foremost the scope is to conserve the finds. However given the limited period of
time during the annual excavation, the proceeding is forced to be phased. Thus, the primary
aims of the project are first, to determine the status quo of conservation of all finds which
have been excavated since 2001; second, to provide fast intervention of restoration for
important and particulary endangered artefacts and third, to create an infrastructure for a long-
term-treatment for conservation, storage and future display. Considering an amount of 1576
finds out of 240 associations and 3400 single finds out of 343 associations only in 2008 and
2009, the task seems to be enormous, even if only about 5-7% of the listed finds belong to the
group of finds with indication of urgent intervention.
Moreover, there is the point of educating students of conservation sciences and restoration as
supplement to the academical curriculum.
In order to contribute to the already existing academical discourse and future research into
archaeometric analysis of iron and glass finds, this conservation research project will
additionally focus on understanding the complex processes of corrosion of these materials.
The analysis are beeing held comparative, since iron and glass are in need of very different
conditions of their ideal in-situ conservation. Ideally, advantage can be taken of the possibility
of taking soil samples of the surrounding earth of particular finds during excavations.

Building history of the castle Cucagna

The archaeological field research on the area of Cucagna is an example for the historical
research on the evolution of the settlement and ruling system in medieval Friuli, which was
influenced by the interplay of the Holy Roman Empire and central Italy.
While in the 10th century locally interwoven structures of villages with manors and
fortifications dominated, self-sustaining palaces gained importance in the following century.
Therefore, Cucagna is an example since it was founded in 1027 by the Swabian house of
Auersberg, which has been proved by historical documents and archaeological evidence.
Cucagna mainly served to defend the march Verona/Friuli for the German kings and emperors
of the Salier and Staufer, namely Konrad II., Henry III. and Friedrich II. Together with
several other castles, it also served for controlling the trading routes out of Friuli towards the
east.
During the 13th century the subsidiary castle Zucco was added to the fortification system of
Cucagna. Other buildings, belonging to the defence and supply system of Cucagna,
demonstrate its dominant local patronate.
These buildings are the broad outer bailey, the fort of Rodingerio, situated on a higher leveled
ridge, the motte at Borgo Scubla, a tower at Stremiz di Faedis and some mills in the valley of
the nearby river Grivò.
The oldest part of Cucagna is the donjon with a height of 22 m and its enclosing wall. In the
following century the castle was extended by at least two phases of construction. These
phases include two palaces (Palazzo 2a, 2b), a chapel and a cistern inside the inner walls, a
defence-like tower palace outside the walls (Palazzo 3) and another palace (Palazzo 4) as well
as a peel inside the outer bailey. In the middle of the 14th century Cucagna reached its largest
circumference, taking effect of a fortified locality.
After the war between the Savorgnan and the Republic of Venice and the “liberation” or
occupation of Friuli by Venice in 1420, the castle lost its importance. Nevertheless the
extension of the castle continued, i.e. by the demolition of the already destroyed outer bailey
including Palazzo 4 and the construction of a wall between Cucagna and Zucco.
Following its final destruction during the war between Venice and Austria, the castle was
abandoned in 1522.

Archaeometric analysis
Due to the special crest site of the castle there are very different types of soil which not only
were formed as a result of erosion, but also by the influence of the castles population with all
its traces including several destructions.
Amongst the listed buildings, Palazzo 4 is of particular interest for the planned analysis
because the main part of the soil samples are taken from there. As mentioned before, the outer
bailey was deconstructed in the late 14th century when the castle was still in vital use. The
area was then only partly site of later utilization. Thus, the state of soil mostly remained
untouched from further human interaction already before the abandonment of the castle in the
early stages of the 16th century.
Consequently, the preconditions for such a proposition are pretty ideal. For an adequate
outcome of information it is planned to conduct non destructive measurings of the find’s alloy
or matrix and the associated corrosion products as well as the surrounding soil.
For their effect to the corrosion of both materials, the chemical compostion of the soil and the
local climate are considered to be the parameters with most influence. It will be elaborated to
what extend characteristic compositions of the find’s materials effect the corrosion process.
In order to comprehend the mutual reaction between the material of the artefacts and the
surrounding soil, the results of the measurings of the elementary composition of the selected
finds, their corrosion products and the corresponding soil sample will be compared to each
other.
Furthermore, a climate profile will be compiled, showing the development of the special
regional climate at the south-eastern mountainside of the Julian Alpes during the past 600
years. It could provide a basis for estimating the potential of conservation of the soil in
comparison to the possibilities of long term storage under realistic atmospheric conditions.
In fact the theoretical conditions at open air are far better than below the surface, because
corrosion processes mainly base on relative humidity and pH value as well as the rate of
ventilation. Nevertheless, the soil climate could be characterized by major constancy than any
of the atmospheric conditions mentioned above.
The particular horizontal stratigraphy of the phases of constrcution limits the datings of finds
available of analysis: the sampling can only be taken at closed assiciations, i.e. all finds
combined with the early phases of constrcution from 11th to 13th century, which already have
been carried out in the years before 2008, can’t be taken into account. During last year’s
excavations at the area of the outer bailey, only finds from the 14th century were recovered.
Circumstances of excavation allowed the taking of 56 soil samples from which 13 can be
linked with glass finds and 38 with iron finds. Depending on the respective feature, every
sample is linked with one or more disposable find.
In order to reach significant results in regard to the scope of research, every disposable find
has to be measured twice to characterize its alloy/matrix and the corrosion product.
For the characterizition of the chemical composition of glass and metal alloys and their
corrosion products, EDX (energy dispersive X-ray spectroscopy) and NAA (neutron
activation analysis) respectively, are considered to be most suitable. Alternatively XRF (X-ray
fluorescence spectroscopy) could be used for analysis of the iron alloy which admittedly is a
destructive method, since the corroded surface has to be partly grinded down before
measuring.
Corrosion products would be measured additionally with XRD to characterize the relevant
mineral phases. To complete the analysis of the finds, they have to be defined according to
their phenomenology such as colour, transparency, homogenity of the matrix, number of
inclusions and bubbles as well as the state of conservation. The analysis of the soil samples is
more complex. It should take into account organo-chemical and petrographical aspects. Other
parameters like density and content of oxygen would be definable only in-situ. After sampling
these data cannot be acquired anymore. Because of the expected costs of such an attempt,
these parameters have to be disregarded. Furthermore a selection of samples has to be
considered, which however cannot be decided before the end of the estimated end of the
sampling phase of the project in 2011.
In order to gain preliminary expertise of the usability of the measuring method and the actual
significance of the results, several test cycles of analysis should be run.
IRON FROM LONDON’S WATERLOGGED SITES: ASSESSING THE OUTCOMES
OF TREATMENT AND PASSIVE STORAGE

ROSE JOHNSON and HELEN GANIARIS1, LIZ BARHAM and LIZ GOODMAN2

1
Museum of London, London Wall, London EC2Y 5HN, UK,
rjohnson@museumoflondon.org.uk, hganiaris@museumoflondon.org.uk
2
Museum of London Archaeology, London Wall, London EC2Y 5HN, UK,
ebarham@museumoflondon.org.uk, lgoodman@museumoflondon.org.uk

In the 1980s, a number of conservation treatments were used on ‘waterfront’ iron at the
Museum of London and elsewhere. Surveys of condition suggested that the iron objects that
had been treated remained in good condition compared to objects that had not been treated.
However, since 1990, active conservation treatments for iron have been largely discontinued
in the UK. At the Museum of London this is largely due to changes to funding and project
priorities, a minimally interventive approach to treatment and environmental issues regarding
use and disposal of treatment solutions. However, the effect on ironwork of this approach has
not been studied. The original aim of this work was to establish whether untreated iron from
waterlogged contexts from the 1980s and 1990s in good desiccated storage had deteriorated
since excavation compared to treated iron. By assessing several different sites using a
standard statistical method, it is hoped that some conclusions can be made. However as this
work progresses, the limitations of assessments of working archives have become apparent.
This paper will also discuss these limitations.

The material and assessment method

The quality of iron objects from London’s excavations with waterlogged conditions can be
exceptional. Corrosion layers tend to be thin, conforming to the surface, and features such as
plating, pattern welding, maker’s marks, copper alloy, silver and tin inlays survive well. In
contrast, iron from aerobic contexts normally has voluminous corrosion that obscures and
sometimes destroys surfaces and features.

Fig. 1 Saxon blade with brass and copper alloy inlay, typical of the excellent condition of some
iron from London ‘waterfront’ sites
The assessment currently has two parts:
• A reassessment of the condition of a group of ‘waterfront’ iron objects treated in
the 1980s and assessed at regular intervals until 2000.
• Examination of ‘waterfront’ iron objects from sites that have had no active
treatment, relying on passive storage in boxes with silica gel.

Assessments will use the Criterion Anchored Rating Scale (CARS) method to create
statistically valid data. CARS was applied by Suenson-Taylor and Sully (1996) to
archaeological leather and then adapted by Heywood in 2000 for the assessment of iron.
CARS is based on the idea of interval data, that is measurement on a scale of equal intervals,
eg length or weight, rather than ordinal data, which is a physical measurement that can be put
into ranked order or counted.

Work in progress: assessment of untreated iron

At the time of writing of this abstract, the assessment of an untreated group from a 1990s
excavation has been completed using CARS. This group of iron was chosen as it was thought
to have received consistent desiccated storage in good archival packaging throughout its
excavated life, with the intention of comparing it with treated material. The 741 objects from
this one site ranged from pins to knives to styli and a large number of unidentified fragments
or parts of objects. Nearly all the objects are packed in perforated clear polythene bags with
Tyvek labels. The bags are stacked upright in a polythene box with a sealable lid and a bag of
silica gel. The boxes are stored in a dehumidified store designated for metals.
It was found that 684 objects, that is, 92% of those examined had the highest possible CARS
score. CARS does provide for some corrosion in its top score (up to 15% of the object
affected by corrosion is acceptable within this score). As a result, it was decided additionally
to get an indication of the proportion of these that have no corrosion at all. 134 objects were
surveyed; 109 out of the 134 (over 81%) still had no active corrosion.
This demonstrates that most of the objects are in very good condition. This has been achieved,
despite long periods when it was found that the desiccated environment had not been
maintained over the 15 years since these objects were excavated. It had been assumed that we
would be assessing a group of iron that had been in well-maintained silica gel storage for a
valid comparison to treated material. However, when this material was retrieved it was found
that a low RH (below 12%) could only be assumed for approximately half the time it had been
in storage. This was due to periods of examination by finds specialists, illustrators and
photographers, and a lack of staff resources to routinely maintain the gel. For the rest of the
time it will have been stored in an environment of 30-40% RH and within that, short periods
of several months potentially at higher levels when removed from store and not kept in silica
gel.
Work in progress: assessment of treated iron
At the time of writing the group of treated material from the 1980s is about to be assessed. It
will have been over 25 years since this material was treated; the condition of this iron has
been tracked with regular assessments (Keene 1994; Heywood 2000).

Observations to date
With the pressures on budgets in both commercial archaeology and museums, it is essential to
establish the effectiveness of treatment and storage regimes. For excavation archives in use
(eg for assessments, illustration, research), and with limited staff resources, the reality is that
it can be difficult to maintain the desiccated conditions recommended to slow down or halt
corrosion. As a result, we do not have a group of untreated iron that can show the
effectiveness of consistent desiccated storage. Nevertheless, the material examined to date has
survived well despite no treatment and long periods out of desiccation.
We cannot claim the sites chosen to be statistically representative of all the excavated
‘waterfront’ iron in our stores, nor are we likely through this assessment to be able to isolate
factors in the causes of good preservation for our ‘waterfront’ iron such as burial
environment, treatment and length of time in desiccation. Reducing variables is difficult with
‘real’ material. However it does provide a starting point which we would welcome further
researchers’ assistance to better define. Stored material can provide an invaluable resource for
reviews of treatments and storage methods if limitations are factored in.
This body of work is a result of a ‘call to action’ at a recent conference on iron (ICON
Archaeology Group, Archaeological Iron: Reflection and Outlook, London, September 2009).
Our assessments on ‘real’ material are intended to complement the experimental work in
progress at Cardiff University, English Heritage and the British Museum. These studies are
looking at optimum conditions for storage of iron, and the effectiveness of treatment in a
wider body of UK material. It is hoped that the combination of experimental work and
assessments of excavation archives will lead to significant and useful conclusions.

References
Heywood, C. (2000) Ferrous metal treatments undertaken by the Museum of London: observations
and a statistical evaluation, unpublished thesis, Durham University.
Keene, S. (1994) ‘Real-time survival rates for treatments of archaeological iron’, in D. Scott, J.
Podany and B. Considine (eds) Ancient and Historic Metals: Conservation and Scientific Research,
Getty Conservation Institute, 249-264.
Sully, D. and Suenson-Taylor, K. (1996) ‘A condition survey of glycerol treated freeze-dried leather in
long-term storage’, in A. Roy and P. Smith (eds) Archaeological Conservation and its Consequences:
Preprints of the Contributions to the Copenhagen Congress, 26–30 August 1996. London:
International Institute for Conservation, 177-181.
THE KUR-PROJECT “LARGE QUANTITY FINDS IN ARCHAEOLOGICAL
COLLECTIONS”

CRISTINA MAZZOLA1, PETER ALBERT1

1 Archäologische Staatssammlung München, Lerchenfeldstr. 2, D-80538 München, Germany

Cristina.Mazzola@extern.lrz-muenchen.de; Peter.Albert@extern.lrz-muenchen.de

The KUR-project – an introduction

Large quantities of finds – especially those consisting of archaeological iron and waterlogged
wood– are always difficult to handle.
To tackle this challenge many different methods of dealing with these finds have been
developed. The project wants to examine and compare these treatments with regard to
practicability, costs and efficiency.

The three year project is funded by the Kulturstiftung des Bundes (German Federal Cultural
Foundation) and Kulturstiftung der Länder (Cultural Foundation of the Federal States).
Executing institutions are the Archäologische Staatssammlung München (Bavarian State
Archaeological Collection, ASM) and the Römisch Germanisches Zentralmuseum (Roman
Germanic Central Museum, RGZM) in Mainz. It stands under the direction of Prof. Dr.
Rupert Gebhard (ASM) and Prof. Dr. Markus Egg (RGZM).
The materials are divided between the institutions – Waldemar Muskalla and Markus Witt-
köpper work with wood at the RGZM whereas the authors (ASM) focus on iron.

General procedure
The treatments used so far include packing the finds in different kinds of foils, in some cases
adding for example oxygen absorbers or desalination in solutions like alkaline sulphite (see
references). The methods are roughly classified into two categories: Storage and desalination.
The former has two focuses – the methods for initial treatment on site need to be quick and
simple, those for long term storage can involve bigger equipment since this is usually
executed in lab environment. The sample objects are taken from two sites: An early medieval
burial ground in Steinheim near Dillingen, Bavaria, excavated in 1987. The second site is a
settlement located at the Leisenhartfeld in Manching, Bavaria, dug in the first half of 2009.
 An appropriate medium for
testing the efficiency of the
storage methods is a mixture
of iron powder and iron chlo-
ride tetrahydrate (Guggenhei-
mer 2006). Under normal cli-
mate conditions (ca. 25 °C, 40
% rH , 21 % O2) this will turn
into akaganeite after a few
Fig. 1: Powder sample in the beginning and turned into days (fig. 1).
akaganeite. Photo: C. Mazzola, ASM

Storage Desalination All samples are regularly weighed and

Different, preferably Alkaline sulphite- or photographed, objects x-rayed. For the oxygen
transparent foils (PE pure sodium
in several hydroxide-treatment
monitoring a non-invasive oxygen meter is available.
thicknesses, vacuum in different variations The progress of desalination baths is controlled by
bags (PE/PA), (vacuum, normal,
ESCAL, aluminium different titration.
barrier film), temperatures, etc.) For the identification of the different iron minerals x-
Different add-ons Heating in a dry
(oxygen absorber, environment (up to ray diffraction (XRD) and Moessbauer spectroscopy
silicagel, soil) 200 °C) are used.
Storage in a
nitrogen filled case- Gas reduction An overview of the tested methods is listed in table
system 1. Most tests were done under laboratory conditions.
Subcritical treatment
(in Clemson, USA) Some of the objects however have been packed on
Washing in distillied site in Manching by the excavation staff to see how
water in overpressure
the handling of foils and equipment fits in the
Table 1: Overview of tested methods
routine.

First observations
Looking at the storage options it was of little surprise that the PE-bags usually used on
excavations offer only little protection against oxygen. The ESCAL foil holds tight as long as
it is not strained extensively. Household vacuum bags seem to be an acceptable alternative –
at least for a few months.
The conversion into akaganeite – indicated by weight and volume increase and colour change
– seems to start already at oxygen levels as low as 1 %. It showed that with both powder and
object samples the oxygen level drops after a while independently from the presence of an
absorber. Probably the oxygen is used up in the oxidation process.
The tests on site were a mixed success. It was criticised, that the packing methods used were
time consuming and that a ‘packing person’ would need to be budgeted from the start. A time
and material saving alternative would be not to bag every single find in high quality foil.
Instead the finds of the day could be packed in a few bigger bags in the evening. Package on
site however is generally advisable and should be considered when planning further
excavations.
For desalination the suggested bath changing rhythm (3 or more baths à 3-4 weeks; Schmidt-
Ott & Otto 2006) seems to be confirmed by the titration results.
It appeared that the objects from the sodium
hydroxide baths were a bit less brittle after
the treatment than the alkaline sulphite
objects.
Two objects have been sent to Clemson,
USA for subcritical treatment. A moisture
chamber test and analyses in Clemson and
Munich point to a good desalination success.
The main difference to finds which have
been treated in alkaline sulphite solutions is
a thin deep red layer of powdery hematite on
the surface. However there was no
Fig. 2: Desalinated objects – in alkaline distinction in the subsequent exposing of the
sulphite solution (above) and subcritical ‘original surface’.
treatment (below) Photo: C. Mazzola, ASM

Future prospects
Much remains to be done – more desalination variations, e.g. with the in the ASM available
gas reduction machine. Improvement of the packing methods to make them quicker and a
broader research in the foil sector are planned to find cheaper alternatives to ESCAL-foil. At
present the systematic study of storage in nitrogen filled case-system has started.

References
Guggenheimer, S. (2006) Investigation into the potential of low-oxygen and dry/cold storage for
freshly excavated iron artifacts (Thesis). School of the University of Applied Sciences Western
Switzerland, La Chaux-de-Fonds, (unpublished).
Mathias, C., Ramsdale, K. and Nixon, D. (2004) ‘Saving archaeological iron using the Revolutionary
Preservation System’, in J. Ashton & D. Hallam (eds): METAL 04, Proceedings of the International
Conference in Metals Conservation, Canberra, Australia, October 2004. Canberra: National Museum
of Australia, 28 – 42.
Schmidt-Ott, K., Otto, N. (2006) ‚Neues zur Eisenentsalzung mit alkalischem Sulfit‘, VdR Beiträge
zur Erhaltung von Kunst- und Kulturgut 2, 126 – 134.
Selwyn, L. (2004) ‘Overview of archaeological iron: The corrosion problem, key factors affecting
treatment and gaps in current knowledge’, in J. Ashton & D. Hallam (eds): METAL 04,etc. (see
Mathias et al.), 294 – 306.
METALLURGICAL PROPERTIES OF STEEL USED IN A TRADITIONAL
JAPANESE MATCHLOCK GUN

MANAKO TANAKA1 and MASAHIRO KITADA1

1
Graduate School of Conservation for Cultural Property, Tokyo University of the Arts, 12-8
Ueno-Koen, Taito-ku, Tokyo 110-8714, Japan
s1308950@fa.geidai.ac.jp ; kitada.masahiro@palette.plala.or.jp

Introduction
The microstructures, nonmetallic inclusions and mechanical properties of a steel barrel of a
traditional Japanese matchlock gun fabricated in the middle of the Edo period (17th-18th
century) have been investigated. Guns were introduced into Japan in the 16th century by
Europeans. It is said that steel made from domestic iron sand and/or imported steel was used
as the raw material of the barrel. However, the details have not been clarified yet because of
the limited number of investigations. It is considered that the barrel of a Japanese matchlock
gun was made mainly by hot forging, although the detailed fabrication technique of the barrel
is not clear as it was carried out in secret. The purpose of this work is to obtain materials-
science data of the raw materials of the Japanese matchlock gun and to investigate the
fabrication technique of the steel barrel.

Experimental methods
The barrel used for analysis is shown in Fig.1. It is considered to have been made after 1624
from the shape of the sight. Specimens are cut from the center of the barrel. The carbon
concentration is determined by chemical analysis. The metallurgical microstructure and
nonmetallic inclusions of the barrel are observed using an optical microscope, a scanning
electron microscope (SEM) and a transmission electron microscope (TEM). To observe the
metallurgical structure, specimens were etched with 5% nital after being buffed with cerium
oxide. The small-area concentration of a nonmetallic inclusion is obtained by electron
dispersive X-ray analysis (EDS). The mechanical properties are evaluated by tension and
hardness tests. To check the impact of fabrication (forging), a test piece is annealed (700oC x
1h) in inert atmosphere.

Fig. 1: Japanese matchlock gun fabricated in the Edo period. (Property of one of the authors:
M.Kitada.)
Results and discussion
1. Microstructures
The center of the barrel is made of mild steel containing 0.01-0.1 mass% carbon. Cross-
sectional optical micrographs of specimens No.1 and No.2 are shown in Fig.2. Although
specimens No.1 and No.2 both consist of α-Fe, the grain size is variable. The grain diameters
of specimens No.1 and No.2 are 30-200µm and 200-1200µm, respectively. The metallurgical
microstructures of a cross section parallel to the barrel and a surface show the same
characteristic. The following two causes are considered to be responsible for the grain
diameter variation: Firstly, the characteristics of the raw material (steel sheet) of the barrel
itself and, secondly, the effects of the forging process, such as annealing temperature and
annealing time.
A typical transmission electron micrograph of α-Fe is shown in Fig.3. Since many
dislocations are observed in the crystals and pinned by Fe3C particles, the steel was slightly
forged at room temperature after hot forging.

500nm

Fig. 2: Cross-sectional optical micrographs of Fig. 3: Transmission electron micrograph of

steel barrel: (a) specimen No.1 and (b) α-Fe. Dislocations are pinned by Fe3C
specimen No.2. Arrow indicates Neumann particles.
band.

2. Nonmetallic inclusions
Two typical microstructures of the nonmetallic inclusions are shown in Fig.4. Fe, Si, Al, Ca,
K, Mg, P, Na and Ti are detected from nonmetallic inclusions of the barrel. Most Japanese
iron sands contain Ti as ilmenite (FeTiO3), but Japanese iron ore and imported steel do not
contain Ti. Since Ti is detected, it is concluded that iron sand was used to make the steel.
The distributions of nonmetallic inclusions in the cross sections normal to the barrel and
parallel to the barrel and surface are shown in Fig.5. From the distributions of nonmetallic
inclusions in two typical Japanese barrels (Fig.6), it is concluded that the barrel was fabricated
by the joining of two long edges of a steel sheet, which is called Udonbari (in Japanese:
seamed steel pipe).
 Fig. 4: SEM images of two typical microstructures of the nonmetallic inclusions, (a)
FeO + Fayalite + glass matrix and (b) glass single phase.

Fig. 5: Distributions of on-metallic inclusions in (a) cross section normal to the

barrel, (b) cross section parallel to the barrel and (c) surface. Arrow indicates
assumed weld area of steel barrel.

Fig. 6: Schematic illustrations of distribution of nonmetallic inclusions (arrows:

dotted lines) in the steel barrel of a Japanese matchlock gun: (a) Udonbari and (b)
Kazuramaki.

3. Mechanical properties
The maximum Vickers hardness and tensile strength are 132Hv and 366Mpa, respectively.
The values are slightly higher than those of the standard data of annealed mild steel; therefore,
there is a possibility of work hardening by cold forging. Optical micrographs of the specimens
before and after annealing are shown in Fig.7 and Fig.8. Crystal growth was observed after
annealing. The maximum Vickers hardness and tensile strength are decreased by annealing at
118HV and 325MP, respectively. As the grain size and Vickers hardness of the surface
change more markedly than those of the cross section, it is concluded that the steel was forged
at room temperature after hot forging.
 Fig. 7: Cross-sectional optical micrographs of steel barrel: (a) before annealing
and (b) after annealing

Fig. 8: Optical micrographs of surface of steel barrel: (a) before annealing and
(b) after annealing.

Conclusions
The microstructures, nonmetallic inclusions and mechanical properties of the barrel of a
Japanese matchlock gun were investigated. The center of the barrel is made from a mild steel
that contains 0.01-0.1 mass% carbon. Since Ti is detected, it is determined that the iron sand
was used as the raw material of the barrel. The barrel was fabricated by the hot-forge welding
of a long steel sheet. The barrel was forged at room temperature after hot forging.

References
Kimura, S. (1854) Houjutsu kunmo vol.3. Kyoto: Tenshinrou; translated from Van Overstraten, J.P.C.
(1850) HANDLEIDING TOT DE KENNIS DER ARTILLERIE. Holland: Militaire Akademie.
Kitada, M. (2007) ‘Fine Structures of a Japanese Sword Fabricated in the Late Muromachi Era (16th
Century)’, Bulletin of the Faculty of Fine Arts Tokyo National University of Fine Arts and Music 45:
5-56.
Kitada, M. (2007) JFE 21st Century Foundation Study Report:71-80.
Kitada, M. (2009) ‘Microstructures and Mechanical Properties of Japanese Ancient Sword (Straight
Sword) Excavated from Old Mound’, Bulletin of the Faculty of Fine Arts Tokyo National University of
Fine Arts and Music 46: 5-32.
Tanaka, M. and M.Kitada (2009) ‘Microstructure of Japanese Matchlock Gun Fabricated in the Edo
Period’, Journal of the Japan Institute of Metals 73: 778-785.
Tokoro, S. (1964) Hinawaju. Tokyo: Yuzankaku.
IRON AND THE MICROSCOPE

DAVID A. SCOTT1

1
Professor, Department of Art History, UCLA, Los Angeles, California, USA, 90049
dascott@ucla.edu

Ferrous materials have a long history in the crucial stages of development of metallurgical
microscopy. The realization that Anglo-Saxon sword blades from Scandinavian excavations
in the 19th century, were fabricated from pattern-welded iron of different kinds, spurred
investigation and fascination with iron’s past and the achievements of the blacksmith. At the
same time, scientists were making significant advances in trying to prepare iron artefacts to
examine microstructure, and explain the kind of products being produced, as well as
understanding the microstructural reasons for the different behaviour of cast iron, steel and
wrought iron. Observations on the structure of iron began in the 17th century, but the
development of metallurgical microscopy began with the work of Henry Clifton Sorby in the
19th century, which has been reviewed in detail by Smith (1988).
In the Old World, iron was never molten in antiquity, except in small amounts, and was joined
by welding or forging of pieces of iron together. As a result, the microstructure of iron is
variable, due to weld lines, different carbon contents, and slag inclusions, and these
differences are very informative concerning the processes of fabrication which have been
employed to produce different artefacts (Smith 1981, Craddock 2005). Carbon content,
methods of heat-treatment, extent of corrosion and patination can all be investigated with the
microscope, which is part of the reason why conservation measures seek to preserve as much
of the original object as possible for future scientific work. The improvement in conservation
methods means that we can examine the microstructure of finds from excavations without the
fear that they have been altered by treatment as they once might have been. Numerous
conservation measures used to conserve iron in the past have failed to maintain the
microstructural integrity of the material we were supposed to have preserved. Such techniques
as heating iron artefacts to dull red heat, the older plasma reduction processes which used
high temperatures, from 300-450 oC, hydrogen reduction and even electrolytic reduction not
only stripped away all of the corrosion products in the case of the latter but altered the
microstructure of the ferrous material being conserved in the case of the former. Our concern
for metallurgical veracity is a major influence on conservation strategies. This veracity can be
determined by microscopical examination, by x-ray radiography and by new non-destructive
methods such as time-of-flight neutron diffraction analysis, used on a Song Dynasty Chinese
iron coin (Huang et al 2010) to determine phase composition. The different phases of the iron-
carbon system are very sensitive to fabrication techniques and compositional variations in
carbon content.
In this short review of structure, we will examine some typical microstructures of ancient iron
artefacts: wrought iron, low-carbon steel, high carbon-steel, welded structures, quenched
martensite, tempered martensite, grey cast iron and white cast iron. The iron-carbon phase
diagram is an essential backdrop for ferrous metallurgy which will be briefly examined, and
which helps to define the phases of ferrite, austenite, ledeburite, pearlite, cementite and
graphite. Even low contents of carbon, 0.4% , create remarkable changes in the properties of
iron artefacts. Phosphoric iron was also an important product and occasionally nitrogen
created nitrides, hardening the iron produced. Phosphoric iron creates possible contrasts with
low-carbon steel used in pattern welding. Even in the bloomery process, small amounts of
high carbon products, such as grey or white cast iron could be made, although the usual
challenge was to retain or get enough carbon into the iron to create steel (Craddock 1995).
The Chinese and Indian task was to remove enough carbon from cast iron products to produce
steel. The co-fusion of wrought iron and cast iron was one way in which this could be done,
used for one of the Indian crucible steel processes (Bronson 1986), or cast iron could be
decarburized. The Chinese production of cast iron two thousand years before the Old World
could make it in any quantity, resulted in a unique metallurgical development of grey cast
iron, white cast iron and mottled or spheroidal graphite cast iron whose ferrous
microstructures are often the most difficult to decipher in the corroded state. Diagrams which
plot the changes of phases depending on the speed with which steels are quenched, known as
time-temperature-transformation diagrams, are useful in determining the presence of the
phases, bainite, troostite, martensite and other carbides (Rollason 1973). An example of a
Japanese sword blade which shows martensite, troostite and the transition to fine pearlite and
a lower carbon steel core will be discussed (Scott 1991). Hardening steels by quenching and
tempering gave rise to products which were far superior to the copper alloys of the Bronze
Age. Some of these products will be reviewed here, employing the metallurgical microscope
and polished and etched sections of some of the iron artefacts examined in the laboratory.

References
Bronson, B. (1986) The making and selling of wootz, a crucible steel of India. Archeomaterials 1, 13-
51.
Craddock, Paul T. (1995) Early Mining and Metal Production.Edinburgh University Press: Edinburgh.
Huang Wei, Winfried Kockelmann, Evelyn Gordfrey, David A Scott, Wu Xiaohong (2010) Non-
destructive Phase Analysis of Song Dynasty Iron Coins by TOF Neutron Diffraction.
Acta Scientiarum Naturalium Universitatis Pekinensis, Vol. 46, No. 2, 245-250.
Rollason, E. C. (1973) Metallurgy for Engineers. Edward Arnold: London.
Scott, David A. (1991) Metallography and Microstructure of Ancient and Historic Metals. J. Paul
Getty Press: Malibu.
Smith, Cyril Stanley (1981) A search for structure: selected essays on science, art and history. MIT
Press: Cambridge, Massachusetts.
Smith, Cyril Stanley (1988) A History of Metallography. MIT Press: Cambridge, Massachusetts.
Gerhard Eggert and Britta Schmutzler (Eds.)

Archaeological Iron
Conservation Colloquium 2010

Extended Abstracts
 Archaeological Iron
Conservation Colloquium 2010

State Academy of Art and Design Stuttgart

24th to 26th June 2010
 Extended Abstracts

Session 3:

Alkaline Chloride Extraction

EFFICIENCY OF CHLORIDE EXTRACTION WITH ORGANIC AMMONIUM
BASES: THE KUR-PROJECT “CONSERVATION AND PROFESSIONAL
STORAGE OF IRON ARTIFACTS”

CHRISTIAN-HEINRICH WUNDERLICH1, CHARLOTTE KUHN2, VERA DRÖBER3,

GERHARD EGGERT2, THOMAS SCHLEID4

1
State Museum for Prehistory Sachsen-Anhalt, Richard-Wagner-Str. 9, D-06114 Halle
(Saale), chwunderlich@lda.mk.sachsen-anhalt.de
2
State Academy of Art and Design, Am Weißenhof 1, D-70191 Stuttgart
charlotte@charlotte-kuhn.de; gerhard.eggert@abk-stuttgart.de
3
Regierungspräsidium Stuttgart, Referat 84/ Archäologische Denkmalpflege, Berliner Strasse
12, 73728 Esslingen a. N., vera.droeber@rps.bwl.de
4
Institute for Inorganic Chemistry, University of Stuttgart, Pfaffenwaldring 55, D-70569
Stuttgart, schleid@iac.uni-stuttgart.de

Introduction
The Project “Conservation and Professional Storage of Iron Artifacts” has been launched by
the Landesamt für Denkmalpflege Sachsen-Anhalt, Landesmuseum für Vorgeschichte Halle
to ensure the long term conservation of our meaningful collection of archaeological iron
artifacts. As part of the “KUR-Project” it is supported by the german Kulturstiftung des
Bundes and Kulturstiftung der Länder.
The project work includes the practical conservation of our collection of iron artifacts, dating
between the 10th century BC and the 16th century AD. Most of the about 3500 finds were
excavated 50-150 years ago. The collection has been stored at uncontrolled conditions until
now and many finds – restored or untreated - corroded after their excavation. A wide range of
corrosion phenomena, typical for post excavation corrosion, can be observed. Our challenge
now is the conservation of a large amount of iron objects, a well known problem to many
archaeological collections throughout Europe.
The centre of the project´s research activities, which are conducted in cooperation with the
State Academy of Art and Design Stuttgart and the University of Stuttgart, is the evaluation
of a new desalination method, to enable an effective improvement of the corrosion stability of
the iron finds by extracting chloride ions within the corrosion layers, which catalyse ongoing
corrosion processes after excavation.
Desalination of iron artifacts with organic quaternary ammonium hydroxides
Although it is well known, that chloride extraction methods considerably improve the
corrosion stability of treated archaeological iron artifacts (e. g. Keene and Orton 1985),
particularly long treatment times require further ameliorations of established methods.
A desalination treatment in solutions based on organic solvents can be expected to be faster
and more effective than aqueous methods, due to the lower surface tension of such solutions,
resulting in better wetting abilities. Besides, no further corrosion of the artifacts has to be
expected during treatment in waterless solutions (Watkinson 1982).
To shift the currently accepted desalination processes in alkaline, aqueous media to water free
solutions, a hydroxide is needed, which shows a sufficient solubility in organic solvents, and
also forms a sufficiently soluble chloride salt.

Research activities
Within the project solutions of tetra-methyl ammonium hydroxide (TMAH) in methanol
and/or water are tested as an alternative to the common alkali hydroxide based desalination
procedures (e. g. Watkinson 1982, Al Zahrani 1999).
Desalination experiments are conducted using roman nails as test objects. Different
concentrations of TMAH in water, methanol, and mixtures of both are compared to the
commonly used alkaline sulphite solution. Residual chloride in the test objects, as well as
chloride washed into solution, is determined. The behaviour of objects in the tested solutions
is monitored weekly by visual surveys, to observe possible ongoing corrosion processes or
the partial dismantling of corrosion layers during the process.
Additionally, the insoluble, chloride containing corrosion product akaganéite, β-FeO(OH),
has been synthesised and is treated in corresponding desalination solutions. Original and
residual chloride content of the solid is also determined and possible phase transformations
are surveyed using X-ray diffraction.

First results
First desalination experiments on synthetic akaganéite indeed show, that chloride extraction
is basically possible in methanolic, as well as aqueous tetra-alkyl ammonium hydroxide
solutions.

Perspective
Further experiments on akaganéite and archaeological artifacts will be carried out, to allow
quantitative comparison of the efficiency of solutions of varying concentrations and solvent
contents.
Acknowledgements
We are grateful to the Chemisches Labor für Umweltanalytik Halle (CLU GmbH) for the
assistance with the chloride analyses.
We also thank Nicole Ebinger-Rist from the Landesamt für Denkmalpflege Esslingen for
providing test objects.

References
Al Zahrani, A. (1999) Chloride ion removal from archaeological iron and β -FeOOH, PhD
dissertation, University of Wales, Cardiff.
Keene, S. and Orton, C. (1985) Stability of treated archaeological iron: an assessment, Studies in
Conservation 30: 136-142.
Watkinson, D. (1982) An assessment of the lithium hydroxide treatments for archaeological ironwork,
in R. W. Clarke and S. M. Blackshaw (eds) Conservation of iron, Maritime Monographs and reports
of the National Maritime Museum 53: 208-213.
THE USE OF SUBCRITICAL ALKALINE SOLUTIONS FOR THE
STABILIZATION OF ARCHAEOLOGICAL IRON ARTIFACTS

PAUL MARDIKIAN1, NESTOR GONZALEZ1, MICHAEL J. DREWS1,

LIISA NASANEN1

1
Clemson University Conservation Center
Warren Lasch Laboratory, School of Materials Science and Engineering, 1250 Supply Street
Bldg. 255, North Charleston, SC 29405, USA
pmardik@clemson.edu, nestorg@clemson.edu, dmichae@clemson.edu,
lnasane@clemson.edu

New technology for the stabilization of archaeological iron using subcritical fluids has been
tested over the last six years at the Clemson Conservation Center. This technique originated
from the extensive research carried out in the industrial world with sub and supercritical
fluids for waste oxidation, radioactive waste reduction and other chemical reactions and
synthesis.
Water heated under the appropriate pressure above its atmospheric boiling temperature of
100oC and under its critical temperature (Tc) of 374oC and pressure (Pc) of 221 bar is
referred to as subcritical water. Above this temperature and pressure water becomes
supercritical; that is, it exhibits properties between those of a liquid and a gas. For practical
reasons, it was decided to remain within the subcritical region for the purposes of treating
archaeological iron, and define a working temperature as low as possible to avoid alteration
of the metallographic structure of the iron. In order to define the optimal operating
temperature, experiments were conducted within the temperature range of 130 to 230oC with
a pH between 11.6 and 13.1. These temperatures and pH range were selected while taking
into account the practical considerations of treating very large artifacts and the likely
effectiveness of the treatment. Based on these results, it was decided to undertake treatment
of the iron at 180oC at a pH of 13.1, using 0.5% w/w sodium hydroxide (NaOH).
The reasoning behind using an alkaline liquid at subcritical conditions, for the treatment of
chloride-ridden iron artifacts was that the treatment time would be significantly reduced due
to the following reasons:
• The increase in temperature of the solution would result in a significant increase in the
chloride diffusion constants.
• The decrease in the viscosity and density of the treatment solution would improve the
diffusion of hydroxide ions into the corrosion layers and promote a more effective
chloride exchange.
 • The decrease in the surface tension of the solution would improve its wettability and
capacity to penetrate the interstices of the corrosion layers.

A typical treatment using this technique involves the following steps: once the object is
placed inside the pressure vessel, the chamber is filled with a solution of sodium hydroxide
and air is purged from the system. Pressure is maintained at approximately 50 bar and the
solution is gradually heated until it reaches 180oC. The flow of the solution during treatment
is adjusted according to the size of the chamber. The eluent solution is regularly sampled in
order to determine the chloride concentration at any given time of the treatment, as well as to
calculate the total concentration of chlorides.
To date, over 150 experiments have been conducted at the Clemson Conservation Center
using wrought and cast iron archaeological samples from marine and terrestrial sites as well
as synthetic iron oxides. Most of the initial trials were conducted on a relatively small scale
using 40 and 600 ml pressure vessels (Fig.1). The small size of these vessels has restricted the
size of the samples and artifacts that can be treated using this technique. So far, hundreds of
samples and relatively small artifacts have been successfully treated. These include rivets,
nuts and bolts, tools and nails. Recently, a 40-liter reactor was designed and built to enable
the treatment of larger artifacts such as ballast blocks from the H.L. Hunley submarine
(1864), and other large and more complex artifacts and samples from both marine and
terrestrial sites (Fig.2).

Figure 1. Research engineer Nestor Gonzalez operating the first

generation subcritical reactor at the Warren Lasch Conservation
Laboratory.
 Figure 2. Research engineer Nestor Gonzalez and conservator Liisa
Nasanen operating the second generation 40-L subcritical reactor.

The chloride release rates using the subcritical technique were found to be significantly
higher and treatment times much shorter than those observed using electrolytic treatment.
None of the treatments using this technique have exceeded 10 days, compared to over 6
months of treatment using traditional methods on some of the cast iron specimens.
During the first years of experimentation with this technique, chemical digestion of the
specimen was used to determine the total amount of chloride remaining in the metal after
treatment. The residual chloride levels in the treated samples were also found to be at least as
low as chloride levels in those treated by electrolysis. In addition, no damage was observed to
the graphitized layer on the cast iron samples, and the long-term stability for all of the
subcritical treated specimens appears to be very good. Moreover, recent work has shown that
the chlorinated oxyhydroxide phase akaganéite (β-FeOOH), thought to be responsible for the
long-term instability of iron, was completely transformed into more stable corrosion products
such as goethite, magnetite and hematite. These results are particularly encouraging for the
stabilization of terrestrial artifacts where akaganéite is often found.
In these experiments undertaken at the Clemson Conservation Center, the subcritical
technique effectively removed very high levels of entrapped chloride ions from terrestrial and
marine iron objects in very short periods of time. While it is true that continuing research and
development of the subcritical technique will require considerable investment, the potential is
great for reducing the long-term costs associated with lengthy conservation treatments and
expensive environmentally controlled post-excavation storage options. The capacity for this
treatment to significantly assist the conservation community with the mass-treatment and
long-term stabilization of archaeological iron may result in considerable savings both in
terms of treatment time and cost of treatment.
References
de Vivies, P., Cook, D., Drews, M.J., Gonzalez, N.G., Mardikian, P., Memet, J.B., (2007)
‘Transformation of akaganéite in archaeological iron artefacts using subcritical treatment’,
Proceedings of the International Conference on Metals Conservation, Amsterdam, Netherlands, 17-21
Sept 2007, eds. C. Degrigny, R. Van Langh, I. Joosten, B. Ankersmit, Amsterdam, 17-21.
Drews, M. J., de Viviés, P., González, N. G., Mardikian, P., (2004) ‘A study of the analysis and
removal of chloride in samples from the Hunley’, Metal 2004, Proceedings of the International
Conference on Metals Conservation Canberra, Australia, 4-8 October 2004, ed. J. Ashton and D.
Hallam, National Museum of Australia, Camberra, 247-260.
Gonzalez, N.G., Cook, D., deVivies, P., Drews, M.J., Mardikian, P., (2007) ‘The effects of cathodic
polarization, soaking in alkaline solutions and subcritical water on cast iron corrosion products’,
Metal 2007, Proceedings of the ICOM-CC Metal WG Conference, Amsterdam, Netherlands, 17-21
Sept 2007, eds. C. Degrigny, R. Van Langh, I. Joosten, B. Ankersmit, Amsterdam, 32-37.
Mardikian, P., Gonzalez, N. G., Drews, M. J., deVivies, P., (2009) ‘New perspectives regarding the
stabilization of terrestrial and marine archaeological iron’, Iron, Steel & Steamship. Proceedings of
the 2nd Australian Seminar held in Perth, Melbourne and Sydney, 2006, ed. M. McCarthy, Australian
National Centre of Excellence for Maritime Archaeology, Fremantle, 113-118.
SOME NEW ADVANCES IN ALKALINE SULPHITE TREATMENT OF
ARCHAEOLOGICAL IRON

SVETLANA BURSHNEVA1, NATALIA SMIRNOVA2

1
State Hermitage Museum, Dvortsovaya Naberezhnaya, 34, 190000, St-Petersburg, Russia
burshneva@yandex.ru
2
Grabar Art Conservation Center, Vologda brunch office, Zasodimskogo st. 14a, 160000,
Vologda, Russia
natasha-s@bk.ru

Introduction
We assume that condition of most archaeological iron finds can be correlated to one of the
oxidation stages that we describe in our table “Degree of ironwork oxidation”. This table is
based on the analogous one that was proposed by J.M. Cronyn (Cronyn 1990: 183). For
marine ironwork, we propose a similar one for archaeological iron. Our table includes some
data of oxide films and corrosion crusts formation at each oxidation stage, and also a short
description of permissible conservation treatments. According to our table from a
conservation point of view the most problematic are iron objects of the IIIrd and IVth
oxidation stages. At these stages the artifacts have already undergone a considerable
mineralization and their original surface has been hidden by corrosion products. But these
objects still have a sound metal core that continues to corrode and provoke flaking of the
mineral crust. Finally the objects lose their shape and cannot be considered as historical
artifact any more. We examined a number of mineral crusts of different archaeological iron
objects with a scanning microscope. We found that the inner structure of mineral crusts of
buried archaeological iron is uneven and non-regulated, cracked and capillary. Chloride-
containing compounds are concentrated at goethite dominating iron oxyhydrates zones, and at
metal – mineral crust interface. Presence of chloride has not been detected within magnetite
zones. It is evidence of good protective properties of magnetite, but unfortunately it does not
produce a smooth layer.

Problem formulation
The alkaline sulphite method for archaeological iron was proposed by N.A. North and C.
Pearson (1975) for desalination of marine iron. They reported that the final product of the
treatment is mixed iron oxide Fe3O4, magnetite. However N.A. North and C. Pearson did not
describe thermodynamics of the process. Later M. Gilberg and N.J. Seeley (1981: 50-56)
revealed that the magnetite formation is more complicated process than it was described by
previous authors. In 1989 in the Institut de Recherches et de Restoration Archeologiques et
Paleometallurgiques in France 107 Iron age axes were treated with the alkaline sulphite
method. After two years 27 of them showed new corrosion– pustules and craters have
appeared (Beaudoin et al. 1997: 170-177). We decided to analyze once again the sequence of
the chemical transformations that take place during the alkaline sulphite treatment of
archaeological iron. We calculated chemical processes that take place during immersion of iron
objects to alkaline sulphite solution (stage 1) and during washing out after immersion in
distilled water (stage 2).

Experimental
Stage 1. When the alkaline sulphite solution reaches pores of the mineral crust it reacts with
oxygen and takes it up. As a result the solution penetrates deeper into the pores:
2SO32- + O2 = 2SO42- (1)

At the same time Fe3+cations get reduced to Fe2+ cations:

2Fe3+ + SO32- + 6OH- = SO42- + 2Fe(OH)2 + H2O (2)

According to thermodynamic calculations if in the solution both Fe3+ and Fe2+ are present,
there is a high possibility for magnetite formation:
2Fe(OH)2+ + SO32- + Fe(OH)+ + 5OH- = Fe3O4 + SO42+ + 5H2O (3)

This conclusion has been proved with experimental data. The pH of the solution falls from
10.4 to 8.8 but does not reach a critical value lower 7.

Stage 2. During the washing out procedure of the object after alkaline sulphite treatment
concentration of iron (III) hydroxide ions increases, the reaction as follow can take place:
4Fe(OH)2 + O2 + (2n-4)H2O → 2(Fe2O3·nH2O) (4)

The presence of soluble Fe3+ and Fe2+ ions again is favorable for magnetite formation.
Fe(OH)2 + Fe2O3·nH2O → Fe3O4 + (n+1) H2O (5)

According to experimental data:

- the total number of chlorides that were extracted during alkaline sulphite treatment is 10.52
times higher than those extracted during intensive washing in distilled water.
- the mechanism of chloride ions extraction from the object happens due to diffusion and
concentrations smoothening according to osmosis law. During washing out procedure the pH of
the solution decreases. Washing out of the working solution (together with chloride anions)
works only for not deep strata of the mineral crust, because the hydrolysis products (Fe(OH)3)
block pores of the mineral crust when pH is lower than ~ 7. It means that some chloride
anions are stuck within the object. This fact, together with long-term treatment itself, gives
some disadvantages for this method.
To increase extraction of chloride ions and to shorten treatment time of objects we applied an
ultrasonic field to the bath. During our experiment first we used samples to select a suitable
regime, and then applied it to archaeological objects. Treatment time for objects in ultrasonic
bath worked out 1-5 minutes every day during three days, depending object’s size. Washing
out the working solution till pH= 7 took only 5 minutes.

Conclusions
Ultrasonic waves intensify the reduction ability of the solution: the concentration of iron (II)
cations increases, and consequently the output of magnetite from chemical reactions also
increases. Moreover, in the ultrasonic field liquid degasification and colloids dispersion take
place: altogether they favorably affect the penetration ability of the solution and washing
water. The treated objects were tested with a scanning microscope. Chloride anions in the
mineral crust and at the metal – mineral crust interface were not detected. However using
ultrasonic bath in alkaline sulphite treatment has some disadvantages: wrong calculations of
immersion time and cavitation energy can provoke the mineral crust to ruin. Objects that
have fragile and badly adhered mineral crust also cannot be treated in the alkaline sulphite
solution. To protect completeness of objects we propose to use gauze bandages that support
fragile areas of mineral crusts. We fix these bandages with Paraloid B-72 glue. The testing
shows that polymer films created with Paraloid B-72 do not affect seriously the penetration
ability of the alkaline solution, and danger of mineral crust ruining is negligible. Thus we can
conclude that nowadays the alkaline sulphite treatment completed with ultrasonic treatment is
the most effective method for archaeological iron objects stabilization. For the first time this
method has been successfully used in Russia for about ten years for archaeological iron finds
from the Russian North and some regions of Siberia, both just after archeological dig and
after long-term storage. A condition check of these objects did not detect corrosion renewal.

References
Beaudoin A., M-C. Clerice, J. Francoise, J-P. Labbe, M-A. Loeper-Attia and L. Robbiola (1997)
“Corrosion d’objets archeologiques en fer après dechloruration par la metode au sulphite alcalin.
Caracterisation physico-chimique et retraitement electrochimique”, Metal 95: 170-177
Cronyn J.M. (1990) The elements of archaeological conservation. London and NY.
Gilberg M., N.J. Seeley (1981) “The identity of compounds containing chloride ions in marine iron
corrosion products: a critical review”, Studies in Conservation 26: 50-56.
North N.A., Pearson С. (1975) “Alkaline Sulfite Reduction Treatment of Marine Iron”
ICOM-СС, 4th Triennial Meeting. Venice, 13/3: 1-14.
 DEGREE OF IRONWORK OXIDATION
Corrosion products Oxidation process Oxidation products Degree of conservation treatment

environment
appearance
Corrosion

Corrosion
Oxidation

formation
stage

layer
Formation of the primary oxide film that works as a protective layer. Very thin and Black thin layer of FeO possibly through The artifacts do not require any
transparent in the beginning, it gets thicker with the time and turns black or dark Fe2O3 (hematite) turns to Fe3O4 cleaning. Conservation coating is
brown. (magnetite). desirable. Artificial oxidation of the
I
surface can be considered as

Film formation
conservation treatment.
Atmospheric

Reaction of iron with oxygen and water produces rust on the surface upon oxide Iron hydroxide (II) Fe(OH)2 oxidizes to All conservation cleaning methods are
film. Film of rust is uneven and has different shades of yellow and light brown. At iron hydroxide (III) Fe(OH)3, which later permissible: mechanical, chemical,
the same time pitting corrosion, which damages oxide film and irritates metal turns to one of hydrated iron oxides (III) electrochemical and electrolythic.
II
surface, occurs. goethite (α-FeOOH) or lepidocrocite (γ- Conservation coating after cleaning is
FeOOH) required.

Original surface of an artifact is marked by the primary oxide film that was formed Primary crust Chemical cleaning permissible but not
at the stage I of oxidation. Under this primary film the process of mineralization goethite (α-FeOOH) possibly through desirable, especially in case of inlay or
takes place towards the center of the artifact. This primary corrosion crust is dark hematite Fe2O3 slowly turns to magnetite other surface decoration, mechanical
III
brown; its thickness is about 1-2 mm (visible to the naked eye). On the primary Fe3O4. cleaning is required. Stabilizing or
crust another corrosion crust is formed – so-called secondary crust. This secondary desalinating treatment is necessary.
crust is porous, fragile and mixed with organic remains and soil particles. Its color Secondary crust Conservation coating after cleaning is
mainly depends on the type of soil in which the artifact was buried. But shades of goethite (α-FeOOH) + secondary corrosion required.
brown and light brown are predominant. The primary oxide film marks the products, soil particles, etc.
boundary between primary and secondary crusts.
Enlargement of the primary crust provokes the artifact to extend. The crust itself Primary crust Chemical cleaning can provoke the
Crust formation

becomes porous and cracked. The secondary crust also enlarges and hides the hematite Fe2O3 slowly turns to magnetite loss of the original surface of an
Underground

original shape of the object. The loss of the primary corrosion crust at this stage Fe3O4. artifact together with corrosion crust.
IV
means the loss of the artifact itself, because the remains of metal core do not Stabilizing treatment is necessary, but
indicate the original shape. Secondary crust only special nondestructive methods
goethite (α-FeOOH) + secondary corrosion can be applied. After stabilizing
products, soil particles, etc. treatment deep polymer impregnation
is required.
The artifact is totally mineralized and highly extended. The primary corrosion Artifact The artifact needs only cleaning and
crust (i.e. the object itself) is porous and cracked; the cracks are filled with magnetite Fe3O4. recovering of its original shape. After
secondary corrosion products. The secondary corrosion crust totally hides the cleaning deep polymer impregnation
V
shape of object. Instead of metal core – a void filled with friable secondary Secondary crust is necessary because the artifact
products of iron corrosion. The object is stable from corrosion point of view. But goethite (α-FeOOH) + secondary corrosion destruction danger is very high.
the danger of its destruction during storage is very high. products, soil particles, etc.

INDEX
Primary oxide film, original Sound metal Void instead sound metal Primary crust Secondary crust
surface
THE EFFECTIVENESS OF CHLORIDE REMOVAL FROM ARCHAEOLOGICAL
IRON USING ALKALINE DEOXYGENATED DESALINATION TREATMENTS

MELANIE RIMMER1, QUANYU WANG2 and DAVID WATKINSON1

1
Cardiff University, School of History and Archaeology, Humanities Building, Colum Road, Cardiff,
CF10 3EU, United Kingdom. rimmermb@cardiff.ac.uk; watkinson@cardiff.ac.uk
2
The British Museum, Department of Conservation and Scientific Research, Great Russell Street,
London, WC1B 3DG, United Kingdom. qwang@thebritishmuseum.ac.uk

Introduction
Desalination treatments for archaeological iron have not been in general use by British
conservators for many years. The reasons for this include time and resource constraints,
concerns about the effects of treatments on fragile mineral-preserved organic remains, and a
lack of reliable quantitative data on desalination efficiency. Storage in reduced relative
humidity (RH) has been adopted as the primary method of corrosion prevention for
archaeological iron collections. A lack of adequate availability and maintenance of such
storage often leads to RH levels that allow corrosion processes to occur, particularly those
relating to chloride-containing β-FeOOH (akaganéite) which begins to induce iron corrosion
above 12% RH (Watkinson and Lewis 2005). There is a growing worry among conservators
that leaving chloride ions in objects poses significant risks if controlled RH storage fails.

Research design
As part of a collaborative research project with Cardiff University and The British Museum,
desalination treatments were re-examined by determining chloride extraction rate from a
statistically significant number of archaeological samples. Two treatments were selected
based on previous research (Watkinson and Al-Zahrani 2008), which suggested that chloride
extraction of over 95% could be achieved using sodium hydroxide (0.5M NaOH)
deoxygenated either with sulphite ions or nitrogen gas. In the study reported here, both
treatments were modified to use 0.1M NaOH at room temperature; the sulphite ion
concentration was also reduced to 0.05M, based on the method employed at the Swiss
National Museum (Schmidt-Ott and Oswald 2006). Both treatments were tested using a large
number of archaeological iron nails exhibiting varied corrosion morphologies and chloride ion
content, originating from three British sites. The nails were desalinated for up to 96 days at
room temperature, with up to six solution changes. Treatment was terminated when chloride
levels in the treatment solution were less than 10 ppm for two consecutive solutions, although
due to time constraints a small proportion of the objects did not complete the treatment.
Chloride concentration was measured using a specific ion meter. After treatment, the nails
were acid digested at room temperature in covered containers, and their residual chloride ion
content determined.
Chloride extraction efficiency
Neither of the two treatments as tested achieved
the >95% average extraction rate reported by
Al-Zahrani. Both achieved similar results in
terms of the average chloride extraction (c.
75%), and both treatments displayed significant
variability in the extraction (11-99%). Alkaline
sulphite treatment heated to 60°C resulted in
shorter treatment times and slightly improved
overall average extraction rate, although the
extraction range remained large. Although this
makes treatment appear to be an unpredictable
process, in the majority of cases (c. 80% for all
three treatments), over 60% of the total chloride
present could be extracted. Although it is
difficult to determine the precise reasons for
poor extraction rates where these occur, several
factors were identified: bulky corrosion products
and/or adhering soil retarding the diffusion rate,
bound chloride, most likely held within the
Figure 1: Scanning electron microcope
structure of β-FeOOH, or chloride trapped deep image (top) and energy-dispersive
within inaccessible pores and surrounded by spectroscopy map of Cl, showing Cl-
containing corrosion products surrounded
metal (Figure 1).
almost completely by metal.

Statistical analysis of results

Due to the fact that large numbers of objects were treated and digested, statistical analyses of
the results is possible. This shows that there is no significant correlation between the total
chloride extracted during treatment and the residual chloride remaining in the object. A
correlation was found between the final solution concentration and the residual chloride
concentration, indicating that where objects did not reach the specified completion criteria
(<10 ppm solution concentration), residual chloride levels tended to be high. This
demonstrates that some objects require longer treatment times than others, and that
monitoring of the chloride concentration in the desalination solution is important if treatments
are to be completed.
The level of residual chloride which induces significant post-excavation corrosion is an
unknown quantity. North and Pearson (1978) suggest that objects containing <200 ppm may
be considered ‘stable’, while those containing >1000 ppm are at significant risk. Using these
boundaries, approximately 80% of the objects had their chloride levels reduced to less than
1000 ppm by treatment, while 40% of the objects fell below the ‘stable’ 200 ppm boundary. It
must be stressed that these limits do not have any empirical basis; the level of chloride in
objects which may be considered ‘safe’ is not known, and is the subject of further research.

Conclusion
The treatment of large numbers of archaeological samples in alkaline deoxygenated solutions
has resulted in a significant body of data, providing statistically sound information about the
extraction efficiency that can be expected when treating objects. Both nitrogen-deoxygenated
0.1M sodium hydroxide and 0.1M/0.05M alkaline sulphite were capable of extracting
significant quantities of chloride ions from objects in the majority of cases. Although it is not
known how much needs to be extracted from objects to prevent rapid and destructive post-
excavation corrosion, it is considered that any reduction of chloride ion content is beneficial
to an object in terms of reducing its susceptibility to corrosion, particularly where storage RH
exceeds the recommended 12% limit. It is hoped that this new data will lead to the
consideration of desalination treatments as part of iron conservation strategies in the UK.

References
North, N. A., and Pearson, C. (1978) ‘Methods for treating marine iron’, in ICOM Committee for
Conservation 5th Triennial Meeting, Zagreb, 78/23/3. Paris: International Council of Museums, 1-10.
Schmidt-Ott, K., and Oswald, N. (2006) ‘Alkaline Sulfite Desalination: Tips and Tricks’, in VDR
conference handbook, "Archaeological Metal Finds - From Excavation to Exhibition". October 11-
13th 2006, Mannheim, Germany.
Watkinson, D., and Al-Zahrani, A. (2008) ‘Towards quantified assessment of aqueous chloride
extraction methods for archaeological iron: de-oxygenated treatment environments’, The Conservator
31: 75-86.
Watkinson, D., and Lewis, M.R.T. (2005) ‘Desiccated Storage of Chloride-Contaminated
Archaeological Iron Objects’, Studies in Conservation 50: 241-252.
SIMPLIFYING SODIUM SULPHITE SOLUTIONS – THE DBU-PROJECT “RET-
TUNG VOR DEM ROST”

BRITTA SCHMUTZLER1, GERHARD EGGERT1

1
State Academy of Art and Design, Am Weißenhof 1, D-70191 Stuttgart, Germany
b.schmutzler@abk-stuttgart.de; gerhard.eggert@abk-stuttgart.de

Motivation
Especially in the field of archaeological heritage preservation, archaeologists and conservators
are confronted with masses of iron finds and almost all of them suffer from the characteristic
post excavation corrosion induced by chloride ions under uncontrolled storage conditions: the
orange akaganéite needles growing vertically up from an iron object’s metal core after the
surface was flaking off.
To investigate the best possibility of conserving masses of iron finds under restricted
personnel and budget facilities, the research project “Rettung vor dem Rost” was funded by
the DBU (Deutsche Bundesstiftung Umwelt, Osnabrück/Germany). Closely connected to
these questions is the aspect of saving our environment by reducing the needed amount of
chemical agents or energy. Starting point of the project is the Alkaline Sulphite Desalination
Method (North and Pearson 1975). The urgency of the research project was highlighted in
2006 by a German mail survey (Eggert and Schmutzler 2009). In this framework, the effort of
time and budget were identified as main counter argument against the application of alkaline
sulphite desalination. Only 20 % of all interviewed German conservators (nearly 40) applied
the Alkaline Sulphite Method at this time. In this extended abstract, experimental results
regarding the aspect of the simplification of the method and reducing the required budget are
reported.

Aim of research in detail

From a theoretical point of view the Alkaline Sulphite Method should be successful (Cornell
and Giovanoli 1990), and accordingly many comparing experiments show that caustic
solutions are the most efficient desalination solutions (e.g. Rinuy 1979; Rinuy and Schweizer
1982; Watkinson 1996). Initially, sodium sulphite was thought to reduce iron(III)compounds
mainly to magnetite. In fact, no reduction takes place at completely dried out iron finds from
the soil, but the role of sulphite is as an oxygen scavenger in the bath solutions (Gilberg and
Seeley 1987). Actually, the passivating power of the concentrated sodium hydroxide is
enough to stabilize the soaked iron (Hjelm-Hansen et al. 1993). Nevertheless, the de-aeration
function is welcome, since freshly precipitated iron(III)hydroxides are suspected to plug up
the pores of the corrosion product layers which can inhibit in turn the diffusion of the
desalination solution respectively chloride ions.
From this outline it can be deduced that de-aeration of desalination solutions is relevant and
can be reached by chemical strategies - or alternatively by physical strategies. First, Al Zah-
rani tried in 1999 de-aeration by nitrogen atmosphere, but his experimental results based on
10 nails only (Al Zahrani 1999; Watkinson and Al Zahrani 2008). The role of sulphite for
desalination success or more alternatives for chemical de-aeration in addition to nitrogen, e.g.
vacuum and heat are to be considered. Furthermore, iron finds from the soil are not as much
chloride contaminated as finds from the sea, that's why the diminishing of the sodium
hydroxide and the sulphite component are also to be investigated (Stawinoga 1996; Schmidt-
Ott and Oswald 2006). The subsequent aspect is to investigate the potential of omitting
sulphite respectively any de-aeration method. These approaches offer the opportunity to
reduce the costs of active conservation as well as the diminishment of environment pollution
by chemical agents.

Methodology of research
The concept of the research project was to identify the most effective desalination method
under consideration of the above mentioned criteria by comparing experiments, conducted
using archaeological finds. The chloride content of the desalination solutions and the residual
chloride content of the desalinated objects were measured.
Of course, working with archaeological iron finds at this large scale needs a courageous and
generous contributor, which could be won in the Landesamt für Denkmalpflege Baden-
Württemberg. Roman nails from the archaeological site “Köngen” were used as specimens.
The desalination experiments were conducted using around 100 g nails in one litre
desalination solution (alkaline sulphite (each 0.5 mol/l), diluted alkaline sulphite (NaOH 0.1
mol/l, Na2SO3 0.05 mol/l), sodium hydroxide NaOH 0.5 and 0.1 mol/l) at room temperature
(20 °C ± 1 °C); all chemicals used were of high quality purchased for analytical purpose. One
bath ran 60 days, and desalination was finished when little or no chloride could be detected in
the bath solution (around 4 baths).

Results
In case of the Köngen finds, the contamination with chloride was quite low compared to finds
from the sea. Nevertheless, this small amount of chloride is able to stimulate post excavation
corrosion. Hence, as main criterion for an effective desalination, the residual chloride content
is to be considered. The less chloride is inside the metal core and the more chloride was
extracted, the more efficient was the desalination method, and the lower is the risk of post
excavation corrosion.
A clear ranking of the varying desalination solutions can be worked out regarding their
sodium hydroxide concentration: in 75 % of cases the 0.5 molar sodium hydroxide solutions
extracted more chloride and left less inside the objects than desalination solutions with 0.1
mol/l NaOH. In contrast, in case of de-aeration the picture is less clear: the most effective
desalination solutions are de-aerated by vacuum, nitrogen or sulphite (0.5 mol/l). But maybe
this impression is only caused by the cold desalination temperature, and is independent from
de-aeration. Because one of the most efficient solutions is cold sodium hydroxide 0.5 mol/l
without any de-aeration – it simply stood at room temperature in a box with an air-tight lid.
This phenomenon was replicated in a second experiment. In contrast, reproducing this
experiment at 55 °C, a result obtained which is quite contrary to expectations: cold
desalination (with and without de-aeration) is more efficient than warm desalination (only de-
aerated with sulphite, 0.5 mol/l and 0.05 mol/l). The initially planned enhancement of
diffusion of the desalination solution as well as the chloride ions seems not to take place.
Furthermore, the combination of physical de-aeration with sodium sulphite does not improve
the desalination.

Discussion
The efficiency of extraction by treatment in our experiments varies between 79 % and 46 %,
i.e. every object still contained chloride after desalination. However, how important are these
differences in efficiency regarding long-term stability? This question has to be answered by
systematic survey of desalinated objects.

Conclusions
As things are now, it is recommended to use 0.5 molar solutions – but all other approaches to
save money work well: Cold desalination allows working with plastic containers, it saves
energy, and it extracts chloride more efficiently. De-Aeration with sulphite is cheap, with
nitrogen cheaper, but with vacuum it is “for free”. A disadvantage to be mentioned is the
equipment needed for mass treatment under vacuum or nitrogen, since air-tight vacuum
drying ovens are recommended. However, smaller quantities can also be desalinated using
exsiccators and pressure control. Finally, no other chemical agents are needed for disposal of
sulphite, whereby the requirements for environment protection are matched.

Perspectives
The project will end in autumn 2010, so more results will be gained in the next months. It is
planned to ensure the reported preliminary results by analysing finds from a second
archaeological site. Furthermore, it is to be investigated, if the high efficiency of cold de-
aerated desalination solutions is caused by de-aeration or temperature. Results from
experiments with hot caustic solutions are still to be evaluated. At least, it has to be
investigated the relevance of the desalination efficiency rank. Then it can be concluded, if
diluted sodium hydroxide concentration is another way for low-cost mass conservation
treatment. All these other aspects of the research project and the detailed experimental results
will be published in near future.
Acknowledgements
We are grateful to DBU, Deutsche Bundesstiftung Umwelt, Osnabrück/ Germany for funding.
For support with archaeological iron samples and lab space we are deeply indebted to
Landesamt für Denkmalpflege Baden-Württemberg, especially Nicole Ebinger-Rist. Further,
we thank all conservators participating at the survey.

References
Al Zahrani, A.N. (1999), Chloride ion removal from archaeological iron and β-FeOOH, unpublished
PhD Thesis, Universitiy of Wales, Cardiff
Cornell, R.M., and Giovanoli, U. (1990), 'Transformation of akaganeite into goethite and hematite in
alkaline media', Clays and Clay Minerals 38: 469-476.
Eggert, G. and Schmutzler, B. (2009), `Lässt sich die Konservierung von Eisenfunden ‚auf Standard’
bringen?´ In kulturGUTerhalten, ed. U. Peltz and O. Zorn, Mainz: Philip von Zabern: 91-95.
Gilberg, M.R., and Seeley, N.J. (1982), `The alkaline sodium sulphite reduction process for
archaeological iron: a closer look´, Studies in Conservation, 27(2): 180-184.
Hjelm-Hansen, N., Van Lanschot, J., Szalkay, C.D., and Turgoose, S. (1993), `Electrochemical
assessment and monitoring of stabilisation of heavily corroded archaeological iron artefacts`,
Corrosion Science 35: 767-774.
North, N.A., and Pearson, C. (1975), `Alkaline sulfite reduction treatment of marine iron´ in Icom
committee for conservation, 4th Triennial Meeting, Venice, 13-18 October 1975. Preprints, Paris:
International Council of Museums: 75/13/ 3/ 1-14.
Rinuy, A. (1979), `Vergleichende Untersuchungen zur Entsalzung von Eisenfunden´, Arbeitsblätter
für Restauratoren, Gruppe 1: 130-140.
Rinuy, A., and Schweizer, F. (1982), `Entsalzung von Eisenfunden mit Alkalischer Sulfitlösung´,
Arbeitsblätter für Restauratoren, Gruppe 1: 160-174.
Schmidt-Ott, K., and Oswald, N. (2006), `Neues zur Eisenentsalzung mit alkalischem Sulfit´, Beiträge
zur Erhaltung von Kunst- und Kulturgut 2: 126-134.
Stawinoga, G. (1996), `Ein Beitrag zur Eisenentsalzung´, Arbeitsblätter für Restauratoren,
Gruppe 1: 293-294.
Watkinson, D. (1996), Chloride extraction from archaeological iron: comparative treatment
efficiencies`, in Archaeological conservation and its consequences, Preprints of the Contributions to
the Copenhagen Congress, 26-30 August 1996, ed. A. Roy and P. Smith, London: International
Institute for Conservation: 208-212.
Watkinson, D. and A. Al Zahrani, (2008), `Towards quantified assessment of aqueous chloride
extraction methods for archaeological iron: de-oxygenated treatment environments´, The Conservator
31: 75-86.
Gerhard Eggert and Britta Schmutzler (Eds.)

Archaeological Iron
Conservation Colloquium 2010

Extended Abstracts
 Archaeological Iron
Conservation Colloquium 2010

State Academy of Art and Design Stuttgart

24th to 26th June 2010
Extended Abstracts

Session 4

Marine Finds
EFFECT OF DECHLORINATION IN NaOH OF IRON ARCHAEOLOGICAL
ARTEFACTS IMMERSED IN SEA WATER

FLORIAN KERGOURLAY1,4, DELPHINE NEFF1, ELODIE GUILMINOT2, CELINE

REMAZEILLES3, SOLENN REGUER4, PHILIPPE REFAIT3, FRANCOIS MIRAMBET5,
EDDY FOY1, PHILIPPE DILLMANN1

1
LAPA/SIS2M, CEA-CNRS et IRAMAT LMC UMR5060 CNRS, CEA Saclay bat 37,
91191 Gif/Yvette, France, florian.kergourlay@cea.fr and delphine.neff@cea.fr
2
Laboratoire ARC’ANTIQUE, 26, rue de la haute forêt, 44000 Nantes, France,
elodie.guilminot@arcantique.org
3
Laboratoire d’Etudes des Matériaux en Milieux Agressifs (LEMMA), Université de La
Rochelle, Avenue Michel Crépeau, F-17042 La Rochelle cedex 01, France;
celine.remazeilles@univ-lr.fr, philippe.refait@univ-lr.fr
4
Synchrotron SOLEIL, Saint Aubin, France, solenn.reguer@synchrotronsoleil.fr
5
Centre de Recherche et de Restauration des Musées de France (C2RMF), Paris, France

Introduction
The aim of the dechlorination treatments is to remove the chloride and the Cl-containing
phases present inside the corrosion layers. Dechlorination of iron artefacts has been widely
studied in order to understand the physico-chemical mechanisms involved. The final goal is to
optimise the various dechlorination treatments tested by the conservation laboratories. A
general tendency to agree on the fact that dechlorination is more efficient on artefacts freshly
excavated stands in the literature (Drews et al. 2004). Treating an artefact after excavation
stops or reduces the corrosion phenomena. The consequence is a limitation of the activation of
the corrosion processes when artefact is exposed to air through the migration of chloride ions
inside the corrosion layers. The precipitation of akaganeite (β-FeOOH) is often associated to
these phenomenons (Selwyn 2004) specifically on the artefacts stored during long periods.
Our methodology initiated by the Odefa project was to characterise the evolution at a
microscopic scale of the corrosion layers during a dechlorination treatment consisting in
soaking of the artefacts in NaOH 2%. Treatments in basic solutions are the most common in
conservation workshop. Artefacts excavated from the marine site of Les Saintes Maries de la
Mer were chosen to establish the set of samples. Some of them were freshly excavated while
others were excavated and stored without any protection against air two years before applying
the treatment. In a first time the entire ingots were treated and transverse sections were
analysed.
Two complementary analytical protocols have been followed. The first one is based on the
treatment of entire iron ingots in basic medium with or without cathodic polarisation.
Transverse sections of the treated ingots were afterwards analysed. The second consists in
characterising the structural evolution of the phases in situ in solution during the treatment.
For this last protocol the treatment in soda 2%wt has been given priority. Characterisation of
the corrosion products on transverse sections were based on a combination of microbeam
analytical tools: µraman spectroscopy (532 nm, Ø 2 µm), X-ray microdiffraction (µXRD, 17
keV and 30x600 µm²) and Energy Dispersive Spectroscopy coupled to Scanning Electron
Microscope).

Before treatment
The observations highlighted by the Odefa project are briefly reminded in this paragraph. One
freshly excavated ingot stored only nine weeks in tap water has been analysed. Corrosion
products of several millimetres thickness were present and a concretion layer (mainly calcite
or quartz) was partly present on the external surface. The inner part of the corrosion layer was
mainly composed of an iron hydroxychloride (β-Fe2(OH)3Cl). On the external part a thin layer
of magnetite (few 10 µm thickness) almost continuous has been observed. Its presence has
been attributed to the first stage of corrosion. Between the β -Fe2(OH)3Cl and the magnetite
layer a thin layer of akaganeite (few 10 µm thickness) was also observed. Its presence could
be explained by a weak release of chloride during the storage of the ingot in aerated tap water.
This process provoked an oxidation of the superficial part of the β-Fe2(OH)3Cl and induced
the precipitation of akaganeite on the ingot during the immersion in the tap water or during
the drying of the ingot when it was prepared for the analyses. The average Cl content of the
layer is about 18%wt. The corrosion layers of the ingot stored during two years at ambient RH
and temperature were completely different. The corrosion products were mainly composed of
iron oxy-hydroxides. Goethite (α-FeOOH) was the main constituent but akaganeite (β-
FeOOH) was present in veins more or less parallel to metal/corrosion product interface.
Moreover marbles were observed by optical microscope and were correlated to less
crystallised phases as feroxyxhyte (δ-FeOOH)/ferrihydrite (5Fe2O3, 9H2O). The average Cl
content of the layer is about 4%wt.

After treatment on entire ingots

Cathodic polarisation in KOH 1%wt and soaking in NaOH 2%wt have been applied on fresh
and stored ingots. On the freshly excavated ingots chlorides extracted from the Cl containing
phases were not detected anymore. The corrosion products have evolved towards a system
composed of goethite and less crystallised ferrihydrite and/or feroxyhyte. Moreover a layer of
magnetite has been detected at the metal/corrosion product interface. On the two-year stored
ingots goethite and ferrihydrite and/or feroxyhyte were present in the layer but akaganeite is
not completely removed from the external part of the corrosion layer.
In situ dechlorination
To determine how the crystalline phases transform inside the corrosion layer during a
treatment and identify the intermediate phases formed in solution, partial treatment in 2%wt
NaOH on samples cut on a freshly excavated ingot have been conducted. Sample was put in a
cell so that µXRD could be performed on the transverse section allowing the transformation
of the phases inside the corrosion layer to be followed. The samples were carefully selected in
order to present both akaganeite and β-Fe2(OH)3Cl allowing to compare the behaviour of both
phases in the same conditions. Firstly it has been noticed that akaganeite contrary to β-
Fe2(OH)3Cl could not be dechlorinated during treatment lasting at the most 35h. Secondly
after 8h of treatment the dechlorination front is located at about 600 µm of the surface in the
β-Fe2(OH)3Cl zone. Magnetite has been mainly observed in situ in the dechlorinated zone but
Fe(OH)2 was also detected during the dechlorination process. Observations have been
completed using raman microspectroscopy of the sample after the interruption of the
treatment. On the dried sample the dechlorinated zone is composed of a mix of goethite and
magnetite and the dechlorination front have been correlated to the presence of akaganeite,
goethite, magnetite and ferrihydrite. The presence of akaganeite on the dried sample and not
in situ could be explained by the interruption of the treatment that induced its precipitation.
Last in the not dechlorinated zones the β-Fe2(OH)3Cl was still the main phase detected by
raman spectroscopy but some akaganeite was also present under the dechlorination front in
the corrosion layer.

Dechlorination mechanisms
The E-pH equilibrium Pourbaix diagram of iron in concentrated chloride aqueous solution
was recently established (Remazeilles et al. 2009). It shows that an equilibrium exists between
the two ferrous phases β-Fe2(OH)3Cl and Fe(OH)2. This equilibrium is more favourable for
Fe(OH)2 under alkaline conditions. In consequence, immersing objects containing β-
Fe2(OH)3Cl in the rust layer leads to the release of chlorides, according to the chemical
reaction:
β-Fe2(OH)3Cl + OH- → 2 Fe(OH)2 + Cl-
Fe(OH)2 was detected in situ in the reactor. Magnetite was also detected and may result from
the oxidation of Fe(OH)2.

Consequences for conservation

The dechlorination rate of β-Fe2(OH)3Cl in alkaline media proved to be higher than that of
akaganéite. So, for artefacts that remained in anoxic conditions, the dechlorination was more
efficient for freshly excavated objects. In this case, the rust layer more likely contained only
β-Fe2(OH)3Cl, even if a small amount of akaganéite could have been present as the result of a
slight oxidation of the β-Fe2(OH)3Cl layer.
Finally, when Fe(OH)2 forms from β-Fe2(OH)3Cl, the next steps of the treatment will have an
influence on the final corrosion products because this phase is very reactive towards O2. A
step favouring a slow oxidation process of Fe(OH)2 (long treatment, soft drying) should lead
to the formation of magnetite (and possibly goethite), while a step favouring a fast oxidation
process (rinsing in ethanol for instance) should lead to poorly crystallized phases such as δ-
FeOOH.

References
Drews, M., de Viviés, P., González, N., and Mardikian, P. (2004), `A study of the analysis
and removal of chloride in iron samples from the Hunley´, in Metal 2004, Proceedings of the
Int. Conference on Metals Conservation, Canberra, Australia, 4-8 October 2004, Canberra:
National Museum of Australia, 247-260.
Selwyn, L. (2004), `Overview of archaeological iron: the corrosion problem, key factors
affecting treatment, and gaps in current knowledge´, in Metal 2004, Proceedings of the Int.
Conference on Metals Conservation, Canberra, Australia, 4-8 October 2004, Canberra:
National Museum of Australia, 294-306.
Rémazeilles, C., Neff, D., Kergourlay, F., Foy, E., Conforto, E., Guilminot, E., Reguer, S.,
Refait, Ph., and Ph. Dillmann (2009), `Mechanisms of long-term anaerobic corrosion of iron
archaeological artefacts in seawater´, Corrosion Science 51(12): 2932-2941
EVOLUTION OF pH IN THE SOLUTIONS OF DECHLORINATION

CHARLENE PELÉ1, STEPHANE LEMOINE1, ELODIE GUILMINOT1

1
Laboratoire Arc’Antique, 26 rue de la Haute Forêt, 44300 Nantes, France
Elodie.guilminot@arcantique.org

Introduction
Archaeological iron artefacts are usually stabilized by electrolysis or by immersion in alkaline
sulphite. For the two treatments, measurements of the chloride concentration are taken every
week by argentimetry – potentiometry. The treatment solution is changed when the chloride
concentration “stops” increasing. In Arc’Antique, the monitoring of the dechlorination
treatment also includes the pH measurement over 2 years. For the electrolysis, the solution of
KOH (1%w) initially has a pH of 13.3. The pH decreases during the treatment. The value is
approximately 10.5 at the end of each bath. These evolutions of the pH seem to be correlated
with those of chloride concentrations. When the chloride concentration stops increasing, the
pH decreases. For the dechlorination by immersion in alkaline sulphite, with the high
concentrations (NaOH 0.5M + Na2SO3 0.5M), the pH of the solution remains at 13.5 during
all the treatment. But if we use the lower concentrations ((NaOH 0.1M + Na2SO3 0.05M)
concentration advised in the Wang’s paper), the pH decreases during the treatment and it can
reach a value of 10 at the end of the bath. In this presentation, we’ll try correlating the
evolution of the pH with the impacts of dechlorination.

The dechlorination of iron artefacts

The ferrous archaeological artefacts contain chlorinated corrosion products. After excavation,
the chloride corrosion cycle can provoke the rapid destruction of the archaeological object.
For this reason, the objects are currently treated to remove chloride ions. Desalination
methods have been employed for several decades in order to stabilize iron archaeological
artefacts. Electrochemical and immersion treatments are currently used. In literature (North
1978, North 1987, Mardikian 2005, Selwyn 2001), it is mentioned that the immersion
methods extract the chloride by diffusion; and the electrolysis favours the extraction of
chlorides by reducing the CPs. In the electrolysis, the electric field applied to the iron artefacts
also accelerates the chloride removal. The solution of the desalination treatments is chosen so
that the artefact remains in the passive zone of the Pourbaix diagram, it’s an alkaline solution.
For the electrolysis, the objects are polarized at -1.45V/SSE (or -0.8V/NHE) with a stabilised
generator in an alkaline solution (as KOH 1%) maintained at room temperature and stirred
every week. For the immersion in the alkaline sulphite, the objects are immersed in a closed
tank, and the solution is regularly stirred and heated (50°C). The alkaline sulphite solution is
made of soda (NaOH, 0.5M) and sodium sulphite (Na2SO3, 0.5M) (North 1978), but in a
recent paper (Wang 2008), the treatments are also effective in less concentrated solutions
(0.1M NaOH and 0.05M Na2SO3).
During the stabilisation treatments, the main parameters of the treatment solution (tempera-
ture, pH and chloride content) were monitored for this paper. The chloride content was quan-
tified using argentimetry / potentiometry and the pH was measured by a glass electrode (with
corrections of the temperature).

Results and discussion

Electrolysis
During the treatments of the archaeological artefacts, only the chloride rate was measured.
But since a research program on the comparison of the iron stabilization, several parameters
have been monitored: chloride rate, pH, temperature, conductivity. For this study, a Roman
iron ingot (from “Saintes Maries de la Mer”, near Marseille in France) was treated by electro-
lysis in 10L of KOH (1% (w/w)). The results of pH and chloride rate are showed in Fig. 1.
We notice the decrease of pH from 13 to 10.5 in each bath. In the first bath, the pH has re-
mained to 13 for 3 days, whereas it has remained to 13 for 8 days in the third bath. The evolu-
tion of pH wouldn’t depend on time but on the evolution of the chloride rate. The pH decrea-
ses when the chloride rate increases, and the pH stabilizes at 10.5 when the chloride rate rea-
ches its maximum value (whatever the value).

Fig. 1: Evolution of pH and Cl contents during electrolysis of an iron ingot in 10L of KOH
(1%w/w).

During the treatments of large objects (for example, cannons), the bath time is longer and the
volume of solution is bigger. However the same evolutions of pH and chloride rate are noted
(Fig 2.).
Fig. 2: Evolution of pH and Cl contents during electrolysis of 2 cast cannons (1 cannon from
“Breles” and 1 cannon from ”Le Juste” shipwreck (France)) in 3500L of KOH (1%w/w).

In theory, the pH of KOH solution (1 %(w/w) = 0.1782M) is 13.25. With the pH of 10.5, the
concentration of OH- is 3.10-4M. If the pH decreases from 13.25 to 10.5, 0.1779M of OH- are
consumed, or 2.994.10-11M of H3O+ are produced. During the electrolysis, the production of
H3O+ could come from the oxidation of water on the anode (stainless steel grid):
6 H2O → O2 + 4 H3O+ + 4 e-
The reactions of carbonation can also produce H3O+ ions. Carbonation occurs when carbon
dioxide dissolves in water:
CO2 + 2 H2O → H2CO3 + H2O
H2CO3 + H2O → HCO3- + H3O+
HCO3- + H2O → CO32- + H3O+ (pKA=10.3)

The consumption of OH- could be due to the penetration of anions OH- in the corrosion
products of the archaeological object. But normally this phenomenon is limited because the
cathodic polarization of the object supports the anions mobility from the object to the solution
(the anode).
Currently we don’t know which the predominant phenomenon is and how we explain exactly
the pH evolution. However the production of H3O+ ions (or consumption of OH- ions) is
closely dependent on the end of the chloride extraction. Thus it would be possible to use the
pH in the monitoring of electrolysis
Immersion treatments in alkaline sulphite solutions
For a research program on the comparison of the iron stabilization, a Roman iron ingot (from
Saintes Maries de la Mer, near Marseille in France) was also treated by immersion in 10L of
alkaline sulfite (0.5M NaOH and 0.5M Na2SO3) at 50°C. The results of pH and chloride rates
are showed in the Fig. 3. The pH remains constant and it doesn’t depend on the chloride rate.
In theory, the pH of the alkaline sulfite solution is 13.7. The measured pH is between 13.1 and
13.7. The variations could be due to the monitoring of solution level (because of the
evaporation).

Fig. 3: Evolution of pH and Cl contents during immersion treatments of an iron ingot in alkaline
sulphite solutions (0.5M NaOH and 0.5M Na2SO3).

Another Roman iron ingot (from Saintes Maries de la Mer, near Marseille in France) was
treated by immersion in less concentrated alkaline sulfite (0.1M NaOH and 0.05M Na2SO3) at
50°C. The results of pH and chloride rate are showed in Fig. 4. In this solution, the pH
decreases from 13 to 10.5. In theory, the pH of the alkaline sulfite solution (0.1M NaOH) is
13. If the pH decreases from 13 to 10.5, 0.0997M of OH- will be consumed, or 2.99.10-11M of
H3O+ will be produced.
We noticed the same evolution of pH during the monitoring of the archaeological artefacts
(from The Natière site, France) (Fig. 5).
During the immersion in alkaline sulfite, the reactions of carbonation could produce H3O+
ions (cf Electrolysis). The diffusion of Cl- from the object to the solution favours the
penetration of anions OH- in the corrosion products of the archaeological object (Selwyn
2004). If the consumption of OH- is too big, the concentration of OH- will not be sufficient to
protect iron.
Fig. 4: Evolution of pH and Cl contents during immersion treatments of an iron ingot in alkaline
sulphite solutions (0.1M NaOH and 0.05M Na2SO3).

Fig. 5: Evolution of pH and Cl contents during immersion treatments of marine iron artefacts
from The Natière site (France) in alkaline sulphite solutions (0.1M NaOH and 0.05M Na2SO3).
Conclusion
Our measurements showed the pH could change during the dechlorination treatments.
In the case of the immersion treatments, the pH only decreases for the less concentrated
solutions (0.1M NaOH and 0.05M Na2SO3). For the solutions more concentrated (0.5M
NaOH and 0.5M Na2SO3), the pH is maintained around 13.5. Thus, the quantity ions OH-
would be sufficient. On the other hand, in the less concentrated solutions, the pH decreases to
10. Consequently, this solution would risk to not keeping the domain of iron passive. So, the
immersion treatments should be used in higher concentrations (0.5M NaOH and 0.5M
Na2SO3).
As regards the electrolysis, even if the pH decreases (13 to 10.5), the object is always
protected by cathodic polarization. Furthermore, we notice the pH evolution is linked to the
rate of chloride extracted: the pH decreases when the chloride rate reaches its maximum
value. Thus, it will be possible to use the pH as a means of treatments monitoring. The pH
measurement is easiest to use for a routine monitoring of the electrolysis treatments.

References
Mardikian, P., Gonzalez, N., Drews, M. and De Vivies, P. (2005) ‘New perspectives regarding the
stabilization of terrestrial and marine archaeological iron’ Proceedings of Williamsburg conference
The Conservation of Archaeological Materials: Current Trends and Future Directions.
North, N. A. and Pearson, C. (1978) ‘Washing methods for chloride removal from marine iron
artifacts’ Studies in Conservation 23: 174-186.
North, N.A. (1987) ‘Conservation of metals’ in Conservation of Marine Archaeological
Objects, Ed. Colin Pearson, Butterworths, London, 207-232.
Selwyn, L.S., (2004) ‘Overview of archaeological iron: the corrosion problem, key factors affecting
treatment, and gaps in current knowledge’ Proceeding of Metal 2004 – Ed. National Museum of
Australia, Canberra: 294-306.
Wang, Q., Dove, S., Shearman, F., and Smirniou, M. (2008) ‘Evaluation of methods chloride ion
concentration determination treatments and effectiveness of desalination treatments using sodium
hydroxide and alkaline sulphite solutions’ The Conservator 31: 65-74.
CONSERVATION OF IRON ARTIFACTS FROM THE USS MONITOR (1862)

ERIC NORDGREN1

1
The Mariners’ Museum, 100 Museum Drive, Newport News, Virginia, USA
enordgren@marinersmuseum.org

Over 200 tons of metal artifacts have been recovered from the wreck of the USS Monitor
(1862), including the ship’s engine, condenser, Dahlgren shell guns, and the world’s first
ship-mounted rotating gun turret. More than 75 percent of this material is composed of
wrought or cast iron resulting in the need for large scale evaluation and conservation of iron
artifacts at The Mariners’ Museum’s (TMM) Batten Conservation Lab. The approach taken is
a practical one, endeavoring to make optimal use of established treatments and work towards
adopting newer, more efficient options when possible.
Conservation of Monitor iron artifacts begins with maintaining wet storage conditions from
the moment of recovery from the sea until the objects can be safely dehydrated during
conservation treatment. Owing to the tremendous volume of material, some of it of a very
large weight and volume (engine, 30 tons, turret, 120 tons) this is not an insignificant task.
Artifact handling and lifting is a significant challenge due to the size, weight and fragility of
many of the Monitor components, requiring that an artifact handler with extensive rigging
experience is included in the conservation team. Large numbers of artifact tanks are required,
some with volumes up to 300,000 liters. Extensive support services such as plumbing and
electrical and drainage systems are also needed, as well as fresh water for the storage
solutions. In the absence of other complicating factors such as the presence of organic
materials, iron artifacts are stored in 1-2% solutions of sodium hydroxide at a pH of 11-13,
which serves to prevent significant iron corrosion as well as to begin the process of chloride
extraction. Large artifacts have initially been stored in caustic solutions based on tap water for
reasons of economy and availability of water purification facilities. Smaller iron artifacts are
stored in sodium hydroxide solutions based on de-ionized water. As larger facilities for the
production of de-ionized water become available and treatments progess, the large artifacts
such as the engine and turret will also be stored in de-ionized water/sodium hydroxide
solutions, allowing more effective desalination and reducing the possibility of calcium
deposition on the artifacts due to reduced solubility at high pH.
In the case of the gun carriages and turret, current storage is in tap water with the addition of
an impressed current cathodic protection system which polarizes the metal sufficiently to
prevent significant further corrosion without the addition of an electrolyte. This was done to
protect wood elements of the composite gun carriages and to prevent damage to organic
artifacts formerly contained within the turret. Now that excavation of artifacts from the turret
has been completed, future treatment stages for the turret will be conducted in sodium
hydroxide.
The pH, condition and level of the storage solutions are monitored to ensure the best storage
conditions possible. The objects are also monitored for signs of corrosion during storage.
Frequently, significant amounts of chloride salts are extracted into solution from the objects
during wet storage, leading Monitor conservators to term this phase of conservation as ‘active
storage’.
Documentation and cataloguing is continued at this stage, starting with the archaeological
context and continuing with pre-treatment, digital photography, computed radiography,
measurements and descriptions. Radiography is very important in ascertaining the
morphology and level of preservation of the artifact, allowing for careful planning of the
conservation and handling. As the appearance changes and true dimensions are revealed
during conservation, documentation continues throughout the process concluding with after
treatment reports.
The vast majority of marine recovered iron artifacts are covered with a thick layer of
‘concretion’ (Memet 2007) consisting primarily of calcium carbonate, biofouling, and
corrosion products, which must be removed for effective desalination treatment. This is
normally accomplished physically with a combination of hand tools and small pneumatic
chisels. In some cases the Garcia Method of flame deconcretion has been used successfully on
iron artifacts as well (Carpenter 1990). Particular care must be taken when deconcreting
objects made of graphitized cast iron, as these have typically lost nearly all of their
mechanical strength and may suffer loss easily.
It should be noted that disassembly of multi-component iron artifacts can promote effective
release of chlorides from confined spaces, and can also be of great benefit when treating
composite artifacts of iron and other metals or organic materials. Disassembly is done when it
is possible to do so without causing damage to the artifact (Krop and Nordgren 2010).
Once concretion has been removed and disassembly started, desalination proceeds at a more
rapid pace. In some cases this involves passive soaking in 1-2% sodium hydroxide solutions
with chloride monitoring at regular intervals of one week, with the solution being changed
when chloride concentrations in solution reach a plateau. More often, electrolytic reduction in
sodium hydroxide solutions is employed. As well as promoting the release of chlorides, this
functions to reduce iron corrosion products and loosen small fragments of remaining
calcareous concretion. The treatment solution is often circulated with a pump to promote a
rinsing effect and to mix the treatment solution thoroughly before solution samples are drawn
for chloride testing.
Chloride testing during iron desalination treatment is normally accomplished by testing
samples of the treatment solutions with methods such as chloride ion specific electrodes,
potentiometric titration, or ion chromatography. The goal of the iron conservation treatments
is to proceed until levels of chloride below 1ppm (mg/L) are measured for at least 3
consecutive weeks. While it is felt that this gives a good indication of the effectiveness of the
treatment, it is realized that solution analysis does not tell the whole story and in selected
cases destructive sampling followed by digestion has been performed to look at the quantity
of chlorides which may still remain in the metal.
While sodium hydroxide soaking with and without electrolytic reduction are the primary
methods currently used for desalination of monitor artifacts, alternative treatments are being
evaluated. A study comparing these methods with hot and cold alkaline sulfite washing of
Monitor iron artifacts was performed by former TMM employee Laura Reid and will be
presented separately at this Colloquium. Conservators working on the Monitor have followed
with interest the development of subcritical extraction of chlorides (de Vivies, Cook, Drews,
Gonzalez, Mardikian, and Memet 2007) by the team working on the submarine HL Hunley at
the Clemson University Restoration Institute and have submitted a sample of wrought iron
from the Monitor’s turret for subcritical testing.
The use of nitrites for corrosion inhibition, storage and treatment of archaeological iron
artifacts is also being investigated and will be reported in a poster session at the Metal 2010
Conference (Sangouard, Nordgren, and Spohn 2010). The test material, sodium nitrite, has the
potential to prevent corrosion of iron and other metals such as copper alloys at a near neutral
pH, making it an attractive possibility for storage of composite artifacts containing organics or
multiple metal alloys. Testing is also underway to determine how well nitrites may function as
a medium for extracting chlorides.
Following the removal of concretion, corrosion and significant amounts of chlorides, Monitor
iron artifacts are thoroughly rinsed in de-ionized water and dried under controlled conditions.
A coating of 2-10% tannic acid is applied to protect the surface with an iron tannate layer,
followed by coating with clear acrylic coatings. When necessary, components are reassembled
and given supportive mountings to relieve stress on the material. The artifacts are carefully
monitored in storage and on display, in which the climate is controlled to approximately 50-
60% relative humidity (RH) and 20-25 degrees Celsius. It is anticipated that some Monitor
components may require special display environments such as low RH or modified
atmosphere, depending on the extent of desalination treatment that has been achieved.
The general methods and practices described above are used to conserve iron artifacts from
the USS Monitor. Although this list is not exhaustive, and other methods are utilized
depending on certain conditions, these techniques have proved quite effective at stabilizing
artifacts recovered from the shipwreck for long term storage and display.
References
Carpenter, J., (1990) ‘A Review of Physical Methods Used to Remove Concretions from Artefacts,
Mainly Iron, Including Some Recent Ideas’, Bulletin of the Australian Institute for Maritime
Archaeology 14 (1), 25-42.
De Vivies, P., Cook, D., Drews, D., Gonzalez, N., Mardikian, P., and Memet, J.B., (2007)
‘Transformation of akaganeite in archaeological iron artefacts using subcritical treatment’ in:
Degrigny, C., ed, Metal 07: Interim meeting of the ICOM-CC Metal WG, Amsterdam, 17.-21.
September 2007, Amsterdam: ICOM-CC Metal Working Group, 26-30.
Krop, D.S, and Nordgren, E. (2010) ‘Disassembly of USS Monitor’s complex mechanical
components.’, in: Metal 2010: Interim meeting of the ICOM-CC Metals Group, Charleston, South
Carolina, 11-15 October 2010. (forthcoming)
Memet, J.B., (2007) ‘The corrosion of metallic artefacts in seawater: descriptive analysis’, in
Corrosion of Metallic Heritage Artifacts, ed. P. Dillmann, G. Beranger, P. Piccardo, and H.
Matthiesen, Woodhead Publishing, Cambridge, 152-169.
Sangouard, E., Nordgren, E., and Spohn, R., (2010) ‘Evaluation of Sodium Nitrite as a Corrosion
Inhibitor for USS Monitor Artifacts’ poster submission for Metal 2010: Interim meeting of the ICOM-
CC Metals Group, Charleston, South Carolina, 11-15 October 2010. (forthcoming)
RETREATMENT OF ARCHAEOLOGICAL IRONS TEMPORARILY SUBMERGED
IN BRACKISH FLOODWATERS

KENYA BROWN1

1
Maryland Archaeological Conservation Laboratory, 10515 Mackall Road, St. Leonard,
Maryland, USA, kbrown@mdp.state.md.us

Abstract
Irons artifacts excavated at Jamestown, Virginia, USA, (Fig. 1) between 1930 and 1950 led
chemists to coat the artifacts with layers of parrafin wax. Salt from the floodwaters entered
into the abraded wax layers and on the surfaces of the irons causing ferrous oxide corrosion.
Treatment precluded dewaxing in 20 liters of xylene in a solvent still heated at 350°F, 176°C
for 6 to 8 hours; desalination in 1 percent of NaOH and deionized water; application of 5
percent tannic acid mixed with deionized water; and a coating of 10 percent paraloid in a
solution of 50 percent each xylene and ethanol.
Challenges throughout the phases of treatment included wax on the surface of the irons after
the dewaxing process. Residual wax forced conservators on the project to retreat the artifacts
by xylene immersion at room temperature for 24 to 48 hours. Desalination often yielded
weekly volatile chloride readings in treatment batches; the objective being the stabilization of
the level of chlorides in the solution at 10 parts per million (ppm) over a four week period.

Two four week periods of abherrant

Fig. 1: Historic Jamestown, Virginia map and site
guide. Courtesy of the National Park Service.
readings delegated immersion into
plain deionized water instead of the
NaOH solution. The immediate results
revealed the desired level of 10 ppm
between a 1 to 4 week period. Tannic
acid application acted as a stabilizer to
the corrosion resulting in the change of
the orange rusted color of the irons to
their original black appearance. Some
objects, however, were resistant
initially to the first application and
required spot mechanical removal
before a secondary tannic acid
application.
Extended Abstract
In 2003, Jamestown staff members prepared the site storing the object for the onslaught of
Hurricane Isabel in accordance to its emergency plan. Unfortunately, the combination of the
hurricane and its related storm surge brought in waters from the nearby Pitch and Tar Swamp
(which takes in waters from the James River) measuring 5 ft, 152 cm deep at the National
Park Service (NPS)Visitor Center. A significant number of the 900,000 artifacts were
affected. 42,500 of them, including approximately 20,000 iron objects, were sent to the lab as
an agreement between the National Park Service and the MAC lab in 2004.
The iron artifacts were treated with paraffin wax applied as both a consolidant and protective
coating by the Civilian Conservation Corps (CCC) and National Park Service in 1935. The
original wax treatment would remain on the irons until their retreatment at the Maryland
Archaeological Conservation Laboratory (MAC) in St. Leonard, Maryland. Although the wax
from the previous treatment offered some protection, the irons suffered corrosion from the
fissures in the wax that developed over several decades. Consequently, these openings
allowed salt laden waters to enter the surfaces of the objects.
The James River situated near 540 km to the southern portion of the Chesapeake Bay, is a
large watershed stretching from New York to Virginia. Salinity levels in the James river are
affected by the urbanization and farm land use around the bay (Benke, et al., 2005) Runoff
waters originating from the rainfall in these areas create groundwater containing nitrates and
chemical contaminants adversely changing its saline properties. Studies conclude that 48
percent of the nitrates found in groundwater contribute to the contamination of the James
River.
Hurricane Isabel brought 1,190 billion gallons of groundwater between September 19-25,
86.9 billion gallons per day (329.4 billion liters), which is 8 times more than the normal
amount of rainfall to enter the Chesapeake during a 7 day period in September (USGS, 2003).
The James's highest saline counts were recorded at 15 ppm on September 24, 2003 based on
information from a stationed buoy in the river (Virginia’s Department of Environmental
Quality, 2003). In March 2004, samples were taken at Jamestown by NPS staff members
showing readings between 40–88 ppm, a significant increase from the readings taken during
the storm. The Pitch and Tar swamp adjacent to the visitor center measured 157-276 ppm
during the same time period.
Irons submerged for a minimum of 24 hours into floodwaters from the swamp were made
susceptible to these same contaminants and augmented salinity which ultimately led to their
corrosion. Jamestown staff and NPS volunteers rinsed the irons in distilled water and bagged
them in two other holding facilities before the objects were sent to the MAC lab.Upon their
arrival, conservators commenced the first phase of treatment, dewaxing followed by
desalination. A typical batch of irons consisted of 200-300 objects per bin. Chloride readings
taken from these batches concluded a range of 10-15 ppm during the first week immersion.
Alkaline solutions tend to dissolve both organic and inorganic material while producing an
environment where corrosion substances are pliable, allowing an easier removal of Clֿ ions
High temperature in alkaline solutions increases the solubility of iron oxides forcing more of
the corrosion products to dissolve (Selwyn and Argyropoulos, 2005). The lab's relative
humidity (RH) averaged at 65 degrees farenheit, 18 degrees celsius daily, however, there are
periods of increased RH during the summer season.
A majority of the chloride readings were heightened for several months possibly indicating
continued states of corrosion caused by an attraction between Clֿ and Fe +² ions.
The adsorption of Clֿ ions is also contingent on the pH of the solution. If corrosion is
passivated at a high pH, the presence of Fe +² ions are decreased minimizing the attraction
between the two. In acidic conditions, Clֿ ions are adsorbed into iron oxide surfaces of the
object at a retarded rate because the charge on the Fe +² becomes Fe -2, repelling the Clֿ ions.
(Selwyn and Argyropoulos, 2005) In these cases, the results were lower chloride readings.
Because chloride test results fluctuated weekly revealing both high and low readings,
questions regarding the pH of the lab's deionized water became a factor. It was found that the
pH level was measured above 7.0. Furthermore, 1 percent NaOH solutions exhibited ranges
between 12.0 and 18.0pH. 60 percent of the objects in desalination contained chlorides
ranging 30-64 ppm weekly. Irons were removed from these solutions and placed in deionized
water in an effort to deactivate any remaining chlorides. Weekly test results yielded lower
chloride readings.
The initial treatment proposed by lab conservators stipulated a final coating of paraloid. In
order to prevent reactivation of corrosion within a microenvironment, objects in high chloride
solutions did not receive an application of paraloid.
The wax layers that remained on some of the irons prevented tannic acid application from
being effective. About 2 percent of the last 3,000 objects remaining in treatment required
additional xylene immersion at room temperature for 24, 48, and 72 hours. Most of these
objects responded to the second xylene treatment and reacted positively to the tannic acid
application. In other cases where the wax remained on the surfaces, tannic acid applications
were again unsuccessful. It was not advisable to perform a tertiary xylene treatment due to the
fragile and mineralized state of the objects.
References
Argyropoulos, V., and Selwyn L.S. (2005) ‘Removal of Chloride and Iron Ions from Archaeological
Wrought Iron with Sodium Hydroxide and Ethylenediamine Solutions’, Studies in Conservation 50:
81-100.
Benke, A.C., Smock, L.A., and Wright A.B. (2005). Rivers of North America. Burlington,
Massachusetts: Elsevier Academic Press.
Keene, Suzanne. (1984) ‘The Performance of Coatings and Consolidants used for Archaeological
Iron,’ in Adhesives and Consolidants: Contributions to the Paris Congress, 1984. London:
International Institute for Conservation, 104-6.
"U.S. Geological Survey: Impacts of High River Flow and Tropical Storm Isabel on the Chesapeake
Bay and its Watershed." U.S. Geological Survey. 06 October 2003. Web. 22 Jan. 2010.
Virginia Department of Environment Quality (2003). James River Water Quality Data. Richmond,
Virginia: Roger Stewart.
Gerhard Eggert and Britta Schmutzler (Eds.)

Archaeological Iron
Conservation Colloquium 2010

Extended Abstracts
 Archaeological Iron
Conservation Colloquium 2010

State Academy of Art and Design Stuttgart

24th to 26th June 2010
Poster Session
IDENTIFICATION OF ORGANIC REMAINS ON IRON FINDS USING VARIABLE
PRESSURE SCANNING ELECTRON MICROSCOPY (VP - SEM)

ANDREA FISCHER1
1
Staatliche Akademie der Bildenden Künste, Am Weißenhof 1, D-70191 Stuttgart, Germany.
a.fischer@abk-stuttgart.de

The remains of organic material, which have been preserved in the corrosion products, can be
observed on many iron finds. Recent systematic and methodical approaches have increased
the understanding of the original function and use of textiles, leather and other organic mate-
rials. Identification of these materials depends on their condition and state of preservation and
has greatly improved because of the introduction of the SEM in the 1980s.
Using a Zeiss EVO 60, investigations at the State Academy of Art and Design in Stuttgart
have revealed the potential of VP- SEM to study organic remains. Alternative applications of
the microscope were tested on iron finds lifted in a block.
The microscope facilitates the analysis of samples without prior preparation and the sample
chamber can accommodate small archaeological finds such as fibulae, fittings or knifes. It is
possible to study surface details of fragile and brittle materials which had not been recognized
before. The non-destructive observations provides a more thorough insight into organic re-
mains, showing the state of preservation of disintegrated substances and preserved morpho-
logical characteristics, thereby enabling identification of the material.
It must be recorded that the capacity
of the cooling stage is limited as far as
both weight and size are concerned.
Furthermore, using the VP - SEM
accelerates the drying process and
therefore the exposure time for in situ
examinations of waterlogged organic
remains is limited. To some extent,
damp samples can be examined,
because water vapour can be
introduced into the chamber and the
(1)highly degraded fibre, the cuticle is lost
(2) negative cast of fibre Image: A. Fischer, SABK temperature of a Peltier cooling stage
can be controlled. Organic remains
from hydrated soil blocks were lifted in small sections and placed on the cooling stage for ex-
amination. Excellent results have been achieved, distinguishing highly degraded textile and
leather remains used perhaps as belts, bags or wrappings.
THE LABORATORY PROCESSING OF BLOCK – LIFTED FINDS FROM GRAVES

ANDREA FISCHER1; CHRISTINA PEEK2

1
Staatliche Akademie der Bildenden Künste, Am Weißenhof 1, D-70191 Stuttgart, Germany
a.fischer@abk-stuttgart.de
2
Landesamt für Denkmalpflege, Berliner Str. 12, D-73728 Esslingen, Germany,
christina.peek@rps.bwl.de

To reconstruct and gain knowledge about early medieval clothing, archaeological research
relies on the examination of the organic remains preserved in close proximity to metal burial
finds. Lifting the finds at archaeological sites as a block enables the recognition and identifi-
cation of all preserved information, ensuring safe excavation and careful investigative clean-
ing in the conservation laboratory. This „excavation en miniature” is a particular challenge to
conservators and is well suited to the didactical imparting of the basic principles and skills of
archaeological conservation. For this reason, practical studies in processing block-lifted finds
are of significant value in the curriculum of the Objects Conservation Course at the State
Academy of Art and Design Stuttgart.
The cleaning of metal finds with associated organic remains is an irreversible process, deter-
mining whether information is preserved or irrecoverably lost. Only a systematic examination
and treatment, including a
stratigraphical excavation of
all structures and layers,
comprehensive documen-
tation and the identification
of the nature of organic
remains, can provide optimal
basic data for research.
The careful documentation is
of scientific importance,
because most of the organic
remains will be destroyed
Documentation of iron finds with mineral preserved textile
remains Image: K. Bott, SABK after the objects are taken out
of the block.
The process of drawing several plana trains one to recognise delicate and sophisticated
details. Drawing details is learning to see - and learning to see is learning to understand!
THE IRON COLLECTION OF ANCIENT MESSENE: A METHODOLOGICAL
CONSERVATION APPROACH

MARIA GIANNOULAKI1, VASILIKE ARGYROPOULOS1, GEORGE MICHALAKA-

KOS1, THEODOROS PANOU2, ANTIKLEIA MOUNDREA-AGRAFIOTI3, PETROS
THEMELIS4

1
Department of Conservation of Antiquities & Works of Art, TEI of Athens, Ag.Spyridonos
12 210 Egaleo, Athens, Greece
mgiann@teiath.gr; bessie@teiath.gr;
2
Department of Medical Radiologic Technologists, TEI of Athens, Ag.Spyridonos 12 210
Egaleo, Athens, Greece
thpanou@teiath.gr
3
Department of History, Archaeology and Social Anthropology, University of Thessaly,
Argonafton & Philellinon 38 221 Volos, Greece
eagraf@ha.uth.gr
4
Society of Messenian Archaeological Studies, Psaromiligou 33, 10 553Athens, Greece
pthemelis@hotmail.com

In Greece, but also in other Mediterranean countries, in museums that house archaeological
collections next to sites and/or excavation areas, iron finds of every day use may be stored for
many years after excavation without proper conservation or maintenance programs. To
compound this problem, wide temperature and RH fluctuations during the year create further
damages to metal artifacts. Thus, the nature of these collections in terms of typology,
technology and condition is unknown and it is difficult to design an effective conservation
plan to preserve the collection. Furthermore, the current legislation does not easily permit the
transportation of artifacts to a scientific laboratory or sampling for further diagnostic
examination. Consequently, a “re-excavation” and a damage assessment of the collection
should be carried out in-situ.
Under the auspices of the 6th Framework funded project, PROMET, the T.E.I. of Athens has
developed a systematic methodology approach based on statistics for surveying quickly the
technology and condition of large metals collections prior to treatment and possibly many
years from excavation, so as to identify treatment priorities. Such an approach requires also
the application of diagnostic techniques in-situ, in order to identify the value and condition of
each artefact and determine their treatment priority.
A technology survey was performed using the variables that describe the type, such as
morphological characteristics. Burt frequency tables and classification trees were produced
according to the defined variables presenting the characteristic technological types of the
collection. A condition survey was carried out to determine treatment priority levels using the
prediction-answer of the question: Which objects are in urgent need of treatment,
would benefit from treatment, do not require treatment? Variables affecting the
condition (such as burial environment, environment and time after excavation, technological
characteristics, type of corrosion products) and the ‘value’ were used. With the assistance of a
statistical package SPSS12.0 (SPSS 2004), frequency tables were produced showing how all
the probable values of a variable(s) are correlated in the sample. From here, Multinomial
Logistic Regression was chosen to predict the answer to the question.
X-ray Radiography was applied in the entire sample of 549 wrought iron objects, as it is an
essential non-destructive technique for revealing hidden clues as to the methods of
manufacturing, decorative detail, as well as the overall condition of the artifacts (amount of
remaining metal core, presence of cracks, extent of corrosion layers).
The process of the results helped to design an effective conservation program for the
preservation and the protection of the collection, including management after excavation,
storage and treatment and was applied experimentally on a sample of 16 wrought iron
artifacts that retained the original metal core but exhibited clear signs of ‘active’ corrosion
(akaganeite formation, spalling), so as to evaluate the efficiency of the application.
This poster presents the overall results from the application of a systematic methodology
approach for surveying the technology and the condition and from the design and application
of an efficient conservation program for the iron artifacts collection of Ancient Messene.

References
Argyropoulos V., M. Giannoulaki, D. Anglos, P. Pouli , MA. Harith, A. Elhassan, AG. Karydas, Ch.
Zarkadas, V. Kantarelou, A. Arafat, N. Haddad (2007) ‘Developing Innovative Portable Diagnostic
Techniques and Approaches for the Analysis of Metal Artefacts from Museum Collections’ in 7th
European Commission Conference Safeguarded Cultural Heritage. Understanding & Viability for the
Enlarged Europe, Prague, Czech Republic.
Giannoulaki M., V. Argyropoulos, Th. Panou, A. Moundrea-Agrafioti and P. Themelis (2005) ‘The
feasibility of using portable X-Ray radiography for the examination of the technology and the
condition of a metals collection housed in the Museum of Ancient Messene, Greece’ in ART
CONSERVATION-RESTORATION – STUDIES AND PRACTICE, Volume VII, VII Conference of Art
Conservation-Restoration, Students and Graduates, 13-15 October 2005, Torun, Poland, 163-176.
FREEZING CORROSION – A VIABLE STORAGE OPTION?

CHARLOTTE KUHN1, GERHARD EGGERT1

1
State Academy of Art and Design, Am Weißenhof 1, D-70191 Stuttgart, Germany
charlotte@charlotte-kuhn.de; gerhard.eggert@abk-stuttgart.de

Introduction
Several authors assume a good corrosion stability of archaeological iron stored at deep-freeze
temperatures, but systematic studies about the efficiency of this conservation method are
missing. The corrosion of iron in Antarctica (e. g. Buchwald and Clarke 1989), following the
same mechanisms as post excavation corrosion, raises doubts about the stability of iron
artefacts under such conditions.

Experimental and Results

To study the process of post excavation corrosion at deep-freeze temperatures, powder
samples consisting of iron and iron(II)-chloride-tetrahydrate (Watkinson and Lewis 2004) in
similar weight proportions were stored at maximum -20 °C and under common working room
climate as a reference. To observe the transformation of the initial powder samples, FTIR
spectroscopic measurements were carried out at regular intervals. The intensity of the band at
852 cm-1, which is specific for akaganéite, can be referred to for a quantitative evaluation of
the akaganéite content in each measured sample. Therefore, a calibration line was created by
measuring pellets with known amounts of akaganéite (Thickett 2003).
Experimental results show, that significant amounts of akaganéite can form within days under
uncontrolled environmental conditions. Concerning samples stored at deep-freeze
temperatures slight changes in colour as well as minimal changes in the IR spectra also
indicate ongoing chemical reactions after just few days. However, corrosion reactions are
significantly delayed. The deep-freeze storage of archaeological iron finds thus seems to be
maintainable for short periods to minimize active corrosion, but can not be recommended as a
long-term storage solution.

References
Buchwald, Clarke (1989) “Corrosion of Fe-Ni alloys by Cl-containing akaganéite (β-FeOOH): The
antarctic meteorite case”, American Mineralogist 74: 656-667
Watkinson, Lewis (2004) “SS Great Britain iron hull: modelling corrosion to define storage relative
humidity”, Proceedings of Metal 2004, National Museum of Australia, Canberra: 88-103
Thickett (2003) “Analysis of Iron Corrosion Products with Fourier Transform Infra-red and Raman
Spectroscopies”, IRUG 6, Padua: Il Prato: 86–93
METAL 2010: ICOM-CC WG ‘METALS’ INTERIM MEETING IN CHARLESTON,
SOUTH CAROLINA FROM OCT. 11-15, 2010

PAUL MARDIKIAN1
1
Clemson University Conservation Center, Warren Lasch Laboratory, School of Materials
Science and Engineering, 1250 Supply Street Bldg. 255, North Charleston, SC 29405, USA.
pmardik@clemson.edu
ARCHAEOLOGICAL IRON AFTER EXCAVATION (AIAE): A SUB-WORKING
GROUP WITHIN THE ICOM-CC METALS WORKING GROUP

ERIC NORDGREN1

1
Co-ordinator ICOM-CC Metals WG Sub-Working Group Archaological Iron After
Excavation (AIAE); The Mariners’ Museum, 100 Museum Drive, Newport News, Virginia,
USA; enordgren@marinersmuseum.org

Archaeological Iron After Excavation (AIAE) is an international forum for scientists,

conservators, archaeologists and other researchers interested in the investigation and
conservation of iron artifacts after they have been recovered from archaeological contexts.
The objectives of AIAE are:

1. To identify research groups working on any aspects of post excavation changes and
storage conditions of archaeological iron excavated from land or marine sites
2. To compile and maintain an exhaustive and accessible bibliography of these topics
3. To identify research areas that need to be further developed
4. To promote research and disseminate relevant information

AIAE was initiated in 2003 as a sub-working group of the International Council of Museums
Committee for Conservation Metal Working Group (ICOM-CC Metal WG), and encourages
participation by ICOM-CC Members. AIAE is a partner with the State Academy of Art and
Design, Stuttgart, Germany in the 2010 Iron Conservation Colloquium in Stuttgart. For more
information please contact Eric Nordgren, AIAE Sub working group co-ordinator.
CORROSION PROTECTION WITHOUT RED LEAD: A CONTRADICTION IN
TERMS?

MARTINA RAEDEL1, MICHAEL BÜCKER1, MARTIN SABEL1, MANDY REIMANN2,

GEORG HABER2, THOMAS BALDAUF2

1
BAM Federal Institute for Materials Research and Testing, Richard-Willstätter-Strasse 11,
12489 Berlin, Germany, martina.raedel@bam.de
2
Haber & Brandner GmbH, Lichtenfelserstrasse 4, 93057 Regensburg, Germany

When restoring historic iron and cast-iron objects, the selection of a suitable anti-corrosive
protective coating is often a key concern. A drawback of most modern anti-corrosion systems
is that they have been developed for use in industrial applications. As a consequence, these
systems have a monochrome appearance, and short drying times often do not allow for the
artistic embellishment of the object’s surface. Another challenge posed by the conservation of
historic objects is that the surface to be treated typically displays a residual level of rust
(according to ISO surface preparation standard St 2). The corrosion protectant frequently
used in the past, red lead, is an environmental and health hazard.

A two-year research project, titled Korrosionsschutz in der Denkmalpflege (“Corrosion

protection in the preservation of historic monuments”), funded by the German Federation of
Industrial Research Associations (AiF), aims to contribute to the investigation, evaluation and
modification of modern anti-corrosion systems for historic iron and cast-iron objects.
Together with the metal conservation workshop Haber & Brander, BAM will evaluate the
performance of modern anti-corrosion systems in comparison with red lead. A key aim is to
adapt commercial anti-corrosion systems to the requirements associated with the conservation
of historic objects.

At the start of the project, a comparative study of red lead and modern anti-corrosion systems
will be performed. Numerous sample materials will be tested: new and artificially corroded
low-alloy steel; historic cast and wrought iron; as well as materials already treated with red
lead or featuring other residual coatings. After applying various commercially available anti-
corrosion systems, the samples will be subjected to an accelerated aging process in a
controlled-humidity atmosphere, under exposure to UV light and, as the case may be,
corrosive gasses. Using a range of optical and physical methods, the researchers hope to
evaluate the performance of the various anti-corrosion systems. The adaptation of the anti-
corrosion substances as well as the preparation of advisory guidelines for conservation
practice are two additional goals of the project.