Corrosion inhibitor testing in archaeological conservation

Metal objects from archaeological contexts often suffer serious damage by corrosion. Various methods for inhibiting corrosion have been developed, but their effects need to be evaluated. Here new research is described on how treatments to inhibit the corrosion of copper and copper-alloy artefacts may be tested.

Metal objects fr om arch aeological contexts often suff er serious damage by corrosion.Various methods fo r inhibiting corrosion have been developed, but th eir effe cts need to be evaluated.
Here new research is described on how treatments to inhibit th e corrosion of copp er and copper-alloy artefacts may be tested.

C
orrosion inhibitors are used to arrest corrosion on archaeologi cal metal artefacts and stop them from deteriorating further, so that they can be preserved for future generations.Inhibitors are mainly applied as solutions and these alter the mineralized metal surface.The corn pounds not only form a chemical bond with the patina, but can also alter the appearance of an object.They can cause colour changes and partial dis solution ofthe admired patina.This article focuses on the evaluation of inhibitor treat ments for copper and copper alloys, and outlines various ways of testing the damage caused by a new conservation treatment.These treatments aim to preserve artefacts with as little change as possible to the sur viving archaeological information.
Archaeological copper and copper alloys frequently contain actively corroding areas on their surface (Fig. 1).These corroding parts can destroy archaeological evidence contained in the mineralized surface, by a process known as "cyclic chloride corro sion".To prevent this sort of continuous damage occurring, artefacts are often treated with corrosion inhibitors.However, the in hibitor most frequently used (benzotriazole) is not always successful in preventing chlo ride corrosion.
The research outlined here had the aim offinding a corrosion-testing procedure by which corrosion inhibitors might be as sessed without testing them on archaeo logical artefacts, because there is a danger that damage might be inflicted on an arte fact by action of a new unknown com pound.The project also examined colour changes induced by corrosion inhibitors.Many artefacts are highly prized for the colour of their patina, which is known as a noble corrosion surface.The corrosion inhibitors applied react with this surface of the artefact and are adsorbed onto the minerals present.This can cause drastic changes in the appearance of an object.Any new compound chosen as an inhibitor should not alter the appearance of an arte fact.

Corrosion of ancient copper and copper alloys
Ancient copper and copper alloys usually have a stratified corrosion layer covering a metallic core.The innermost layer cover ing the remaining copper is the ruby red cuprous oxide known as cuprite (Cu20).It is in turn covered either with a dark green copper carbonate, malachite (Cu2(0H)2C0 3 ), or with the less commonly found blue cop per carbonate, azurite (Cu 3 (0H}AC0 :1lzl .If harmful chlorides are present in the corro sion layers, pale grey waxy cuprous chlo ride, nantokite (CuCl), will probably cover parts of the remaining metal.Nantokite is unstable when in contact with moist air and will change into basic copper chloride, either green paratacamite (CuCl23Cu(OH)2), bright to darkgreen atacamite (Cu2(0H) 3 Cl), or rarer blue-green botallackite (Cu 2 (0Hk Cl.H20).Bronzes going through this proc-  Conservators have treated artefacts with corrosion inhibitors, to prevent the conver sion of cuprous chloride to basic copper chloride.Only benzotriazole (BTA) has been widely accepted in the stabilization of cop per and copper-alloy artefacts.Madsen pro posed BTA for archaeological conservation based on a vast amount of research done for industrial purposes.
2 Because of its per ceived advantages and satisfactory results, BTA has been the predominant corrosion inhibitor used in the conservation of copper and copper alloys.Since that time Oddy, 3 Sease 4 and others have investigated the properties of BTA, but very little research has been done on other corrosion-inhibit ing compounds in conservation.The most recent corrosion inhibitor suggested for copper-corrosion inhibition in conserva tion is 2-amino-5-mercapto-1 ,3,4-thiadi azole (AMT). 5However, AMT has not been accepted generally for use in archaeologi cal conservation, because of the lack of in formation available on its effects, both in industrial applications and in the field of archaeological conservation.
Generally, conservation research into corrosion prevention aims to compare the effectiveness of different treatments,6 and it usually uses archaeological samples to compare their effectiveness.However, cor rosion testing using archaeological copper and copper alloys has several disadvan tages: • the size and shape of the artefacts are dif ferent Bearing these disadvantages in mind, a corrosion-testing procedure was developed by the author using specially prepared copper coupons, covered in copper-chloride corrosion products, for testing corrosion inhibitors for copper and copper-alloy ar chaeological artefacts Te sting procedure Virtually (99.9 per cent) pure copper cou pons were immersed in cupric-chloride solutions to produce cuprous-chloride cor rosion, the structure of which is similar to the corrosion found directly against the re maining metal of copper artefacts.These coupons in contact with 95 per cent rela tive humidity at ambient temperature pro duced paratacamite (Fig. 2).This is similar to the reaction that occurs on archaeolog ical artefacts containing chloride corrosion.
The corrosion inhibitors in Table 1 were tested with this new corrosion test.

Nitrogen based inhibitors
Formula Abbr. Benzotriazole The inhibitor test found that none of the inhibitors was 100 per cent successful in inhibiting the conversion from nantokite to paratacamite.BTA and MBT prevented corrosion best.The other inhibitors were less effective or completely ineffective.All ofthem induced colour changes on the cor roded surface of the coupons, and, to deter mine the amount of colour changes, further experimental work was undertaken.

Colour changes attributable to cor rosion inhibitors
Some archaeological artefacts are covered in very beautiful green and red corrosion products, considered indicative of the an cient pedigree of the artefact.However, the colour changes induced on these surfaces by corrosion-inhibitor treatments have not been accurately assessed qualitatively or quantitatively.Describing slight alterations in colour poses many difficulties when evaluating corrosion surfaces after a con servation treatment.An attempt was made to qualify and quantify colour changes in duced by corrosion inhibitors for copper and copper-alloy archaeological artefacts, as precisely as possible.
Every observer perceives colour differ ently.A major obstacle encountered when comparing colours is the choice of descrip-ARCHAEOLOGY INTERNATIONAL tive words.This was found to be largely dependent on the observer's colour per ception, and on changes in light source.It is also necessary to take into account the possibility that some people might suffer from a small degree of undetected colour blindness.The drawbacks encountered with a visual assessment of colour changes led to the use of a colour-measuring device (colorimeter), which assigns a numerical value to colour.This avoids individual colour perception, and results in a higher accuracy and reproducibility of colour values.The Minolta Chroma Meter CR-200 expresses colours in precise numerical values and was used for colour matching in the research.The instrument relies on mod ern optoelectronic technology, and com pletely discounts the subjective variations of individual observers.This tool was ideal for measuring the small variations in col our observed on copper minerals treated with corrosion inhibitors.
To evaluate the colour changes induced by corrosion-inhibitor treatments, the inhib itors chosen were individually dissolved in ethanol and exposed to cuprite, malachite and nantokite.These are the three main copper-corrosion products found on archae ological copper and copper alloys.This evaluation indicated that all corrosion in hibitors will change the colour of a miner alized surface containing cuprite, malachite and nantokite.Chloride-free copper and copper alloys will not benefit from treat ment with these inhibitors.They cause col our changes and therefore should not be used.
If the artefact has to be treated because of active copper-chloride corrosion present in the corrosion surface, a drastic colour change of the nantokite can be expected.For example, BTA causes almost double the amount of colour change in the nantokite than AMTa colour change that is very apparent to the naked eye.

Conclusion
The testing procedure developed shows that corrosion inhibitors do not have to be tested on archaeological material.The inhib itors can be quantitatively compared using the technique outlined and the best per forming inhibitor can be selected for further testing for factors such as colour changes.
The evaluation of colour changes led to the recommendation that copper and copper-alloy archaeological artefacts, with a highly prized mineralized surface, should not be treated with these inhibitors be cause of the colour changes induced.If the artefact contains chlorides, it should be stored in a desiccated environment and its condition monitored at regular intervals.However, it must also be borne in mind that a corrosion-inhibitor treatment might be of benefit to the long-term stability of an artefact, and the colour changes might be acceptable.This means that the decision to carry out an inhibitor treatment must be made by looking at each object as an indi vidual case.

0 cmFigure 1
Figure 1 A copper-alloy arrowhead in the AI Ain Museum, Abu Dhabi, Un ited Arab Emirates, covered in copper corrosion .Th e tip of the arrowhead is undergoing active copper-chloride corrosion.
ess are often referred to as suffering from "bronze disease" . 1

•
the type and amount of alloying constit uents of the original metal are often not known and might influence the corro sion inhibitor treatment • the type and amount of corrosion prod ucts present on and in the surface of the artefact vary greatly • the porosity of corrosion products may differ • the inhibitor treatment might damage the patina of the archaeological artefact • the long-term effects attributable to an inhibitor treatment are not known • the corrosion inhibitor treatment is probably irreversible.

Figure 2
Figure 2 Copper coupons containing artificial copper-chloride corrosion.

7
Notes 1. R. B. Faltermeier, Th e evaluation of cor rosion inhibitors for application to copper and copper alloy arch aeological artefacts(PhD thesis, University College London, Institute of Archaeology, 1995).2. H. Madsen, "A preliminary note on the use ofbenzotriazole for stabilising bronze objects", Studies in Conservation 12, 163-6, 1967.3. W. Oddy, "Toxicity of benzotriazole", Studies in Conservation 19, 188-9, 1974.4. C. Sease, "Benzotriazole: a review for con servators ", Studies in Con serva tion 23, 76-85, 1978. 5. M. C. Ganorkar, V. PanditRao,P.Gayathri, T. A Sreenivasa Rao, "A novel method for conservation of copper and copper-based artifacts", Studies in Conservation 33, 97-101, 1988.6. I. D. McLeod, "Conservation of corroded copper and copper alloys: a comparison of new and traditional methods for removing chloride ions", Studies in Conservation 32, 25-40, 1987.7. The author wishes to thank E. Pye and R. Gibbs for their encouragement.con structive dialogue and assistance with the English.This research was supported financially by the Institute of Archaeol ogy, University College London, and by the Novartis Corporation, Base!, Switzer land.