A New Methodology for Accelerated Corrosion Testing

A New Methodology for Accelerated Corrosion Testing

Laurianne Robinet and David Thickett

Accelerated corrosion tests have been used for almost 30 years to exclude corrosive materials from museum display and storage. The method reported by Green and Thickett in 1995 has been developed to allow the evaluation of three metals (silver, copper and lead) simultaneously in a single test. Comparison of a wide range of materials has shown that the differences in results introduced by this modification are less than the differences encountered with replicate tests using the same material employing the previous method. Problems encountered with corrosion in lead control tests have been investigated and the method has been modified to remove this effect. The cleaning method for the test vessels has been modified to exclude the hazardous chemical, chromic acid.

INTRODUCTION

Accelerated corrosion tests have been used in conservation since 1972 [1, 2]. Such tests have much to commend them. They are relatively fast and inexpensive; and when used prior to building or refurbishment work, they are pro-active and exclude unsuitable materials before damage has occurred. They 'model' the showcase/storage situation and do not require knowledge of air exchange rates or the susceptibilities and damage levels of different metals. The tests suffer the drawbacks of all accelerated testing, with uncertainty over the acceleration factors compared to ambient environments and the possibility of encouraging reactions that would not occur under normally encountered conditions. This could lead to materials failing the tests, which if used under the actual display/storage conditions would not cause problems. However, if the object of the testing is to exclude potentially damaging materials, then this is not a serious failing. The intention has always been that the test method should be available to any conservatorand hence not require complex equipment or extensive training.

The test method has developed over the last three decades, with modifications introduced to suit particular institutions or circumstances. This has led to a situation where many minor variations of the test are in use. A survey of UK institutions using the test highlighted the many differences in methods in use [3]. An inter-laboratory comparison between these institutions showed that such small differences could lead to very different results being obtained from the same test material. In the field of analytical chemistry it is well recognized that perfect reproducibility between laboratories does not occur, even with well validated and controlled analytical methods. However, two tests providing different results for the same material can seriously undermine confidence in both the tests themselves and the need for materials testing. Research conducted into the various factors affecting the test method generated a method which balances ease of preparation with reproducibility. Repeating the inter-laboratory comparison, with each participant using the new method, gave very significantly improved reproducibility [4].

The major and fundamental limiting factor on reproducibility for this type of testing is almost certainly the reproducibility of the metal surfaces, which influences their reactivity to the test environment. The aims of the present work were to investigate different ways of running the tests with all three metals in one vessel, to optimize a method for preparation speed and utility, whilst retaining control over those test factors that significantly affect the test results. The test reported here has been evaluated against the method published by Green and Thicken in 1995 [4], using 53 materials, and the procedure is presented in Appendix 1. At the same time, two separate problems identified with the 1995 method — tube cleaning, and corrosion observed on the lead control coupon — were investigated. The results are reported below.

TUBE CLEANING

Cleanliness is a very important part of the test. If the tubes are not cleaned properly, the results of the next test could be adversely affected. The standard cleaning procedure comprises three separate steps, one of which uses chromic acid. Chromic acid is of major concern because it has serious health and safety implications and is damaging to the environment. Two weaker, safer, acid solutions were tested as replacements for chromic acid solution. Decon 90 'Acid Rinse' detergent and dilute nitric acid solution (10% by volume) were used to clean tubes which had contained an unsuitable material, which had corroded lead in a test. Tests with lead, but no material, were then assembled using the cleaned tubes. Lead was used because of its high sensitivity to low pollutant gas concentrations [5]. At the end of the test, no corrosion was observed on any of the lead test coupons. The new cleaning procedure recommended is presented in Appendix 1.

ALTERATION OF THE LEAD CONTROL COUPON

Sometimes, in the standard test, a red oxidation appears on the lead control coupon. Corrosion on the control is problematic, as the batch of tests must be repeated. The effect was not observed on the materials tested concurrently, hence a dirty tube is unlikely to be responsible for the corrosion.

To try to determine the cause of the corrosion, tests were undertaken to investigate various factors such as: different cotton wool; different tissue (used to dry the coupon); different source of lead; coupon not cleaned in acetone (propanone); coupon air-dried after acetonecleaning; coupon in a tube without water; coupon in direct contact with water; and a different container, to avoid contact with the interior surface of the glass. Several of the tests produced white and red corrosion which was identified by micro-Raman spectroscopy and X-ray diffraction as hydrocerrusite (2PbCO3 • Pb(OH)2), litharge (tetragonal PbO), massicot (orthorhombic PbO) and plumbonacrite (6PbCO3 - 3Pb(OH)2- PbO).

The tests set up in a vertical position did not show any corrosion. Some of the previous tests were repeated with the tubes placed vertically, using an aluminium tray with square compartments. At the end of the test no corrosion was observed on any of the lead coupons. The contact created between the coupon and condensation on the glass, by placing the tube at an angle, was responsible for the corrosion on the control coupon. This effect can be avoided by placing the tube vertically in the oven, utilizing an aluminium tray.

THREE COUPONS IN ONE TUBE

The standard method involves setting up a separate test for each metal coupon: silver, copper and lead. In a modified test method [6], the three metal coupons are bent on the edge of a glass beaker, in direct contact with the glass. Unfortunately this arrangement facilitates contact with water condensation which can induce corrosion. Also, because of stress induced in the metal coupon, corrosion appears at the bend, which would not normally be observed.

Placing the three metal coupons in a single tube would greatly simplify' the test, and reduce preparation time. Such a method must avoid contact between the coupons, and also contact with the glass of the tube. The use of a soft stopper in which slots are cut to insert the three different metal coupons (Figure 1) was investigated.

Soft stoppers from different manufacturers were tested for out-gassing, using the standard accelerated corrosion test. Neoprene stoppers were found unsuitable as they tarnish silver and copper. Silicone stoppers produced by VWR and Radleys proved to be inert with respect to the metal coupons used. It was noticed that the part of the silver coupon in the slot of the silicone stopper tarnished sometimes, but the part exposed in the test was not affected.

In order to verify that the modification gives the same results as the standard test, a trial was run on different materials using the standard method and the new method, called the '3 in 1' test. Radleys stoppers were used for all the following experiments.

Figure 1 Metal coupons placed parallel in a soft stopper.

Initially, tests were run on three commonly tested materials: wood and two textiles. Different weights of material (2, 4 and 6 g) were placed in the tubes, to evaluate the effect of the produced gases being split between the coupons, and hence the dose of gas for each coupon being lower than in the standard test. The gases might react preferentially with one or more of the metal coupons. Each standard test was repeated twice on each material and each '3 in 1' test was repeated three times. The whole assembly was placed in the oven at 60°C for 28 days. At the end of the test period, the coupons were assessed according to the standard rating definitions used by the BritishMuseum [7]:

P suitable for use (no corrosion)

T suitable for temporary use, up to six months (slightcorrosion)

U unsuitable for use

Better results were always obtained with 2 g of material. The more material present, the less the coupon corroded. The more material present in the test tube, the less water remained in the small tube.

The repeats of the same test on the same materials gave identical results. Even though the coupons in both tests were affected in the same way and gave the same assessment, their appearance was a little different. In the '3 in 1' test, the corrosion was always slightly less pronounced and mainly concentrated at the bottom. This is probably because the coupons are at a different height, with more distance between the coupon and the fabric than in the standard test; and because they are next to each other, which restricts the mixing of the air in the tube.

Based on the observations from the first trials, the '3 in 1' test was modified. Another batch of tests was run on the textile which gave the greatest difference in the appearance of the coupons between the two tests in the previous trial. Various positions and lengths of coupon were mailed. The arrangement shown in Figure 1 gave results most comparable with the standard test. The appearance of the coupon will never be exactly the sameas in the standard test, because the environment is different. Nevertheless, the two tests produce the same ratings. The presence of more pronounced corrosion in the second trial, compared to the first trial, reveals the importance of the positioning of the coupons in the tube. The '3 in 1' test has the advantage of improving the repeatability of the coupon positioning, a previously unsuspected factor in the test.

In order to decide if the '3 in 1' test could safely replace the standard test, comparison tests were run on 53 different samples commonly encountered in museum use, including textiles, woods, adhesives, paints and paper. The '3 in 1' test was set up according to the procedure described in Appendix 1. The lead coupon was placed in the central position (Figure 1) to avoid any risk of contact with the glass of the tube, which would cause corrosion in the case of water condensation. Lead is by far the most susceptible of the three metals to this effect. The coupon should be assessed at the bottom, the part which reacts the most. However, corrosion sometimes occurred at a higher level on the coupon, in the middle or nearly at the top, generally with lead. The same effect was observed in the corresponding standard test. The vertical stratigraphy of the corrosion had not previously been considered. The '3 in 1' test is superior to the standard test because metal is available near the stopper, unlike the standard test.

With 46 of the materials the results were identical, whatever the materials used or the rating obtained (P, T or U). With seven of the materials, the two tests did not agree for one of the three metals (Table 1). Among the disagreements, the standard test is sometimes more sensitive than the '3 in 1' test and sometimes the opposite is true, so it is not possible to generalize that one test is more sensitive than the other.

The coupons were assessed according to the standard rating described in the test procedure. The percentage of identical results was 91% for silver, 96% for copper and 98% for lead. A statistical comparison between the '3 in 1' test and the standard test method is presented in Appendix 2.

Table 1 Comparison of the level of agreement between the '3 In 7' test and the standard test and for repeats of the standard test

٭Root mean square deviation

CONCLUSION

During this research three independent aspects were examined and improved: tube cleaning, changes in lead control coupons, and developing a way to place the three coupons in one tube.

In the cleaning process, a new procedure has been put in place. Chromic acid is now replaced by either a solution of dilute nitric acid or Decon 90 'Acid Rinse' detergent, both of which are significantly less dangerous to health and non-hazardous towards the environment.

The unusual red corrosion on lead control coupons was explained by the contact of the coupon with the condensed water on the edges of the tube, condensation which was formed during the opening of the oven. To solve the corrosion problem on the control, the position of the tube in the standard test was changed from a leaning to a vertical position by using an aluminium tray.

The utility of the accelerated corrosion test method has been significantly improved. The placing of all three metal coupons in a single tube not only saves time in preparing the tests but also improves the reproducibility of the positioning of the coupons, a factor not previously suspected in the amount of corrosion produced. The new test method gives a very high agreement with results from the previously published method.

ACKNOWLEDGEMENT

The authors would like to thank Susan Bradley from the BritishMuseum for her advice and comments, and Kristie Short for her help in setting up the tests.

APPENDIX 1: '3 IN 1”METHOD PROCEDURE

Equipment required

2 g sample of the material, 50 ml glass tubes, inert silicone stoppers, small glass tubes (0.8 ml), metal foil (Ag, Cu, Pb). Cotton wool, distilled water, high purity acetone, three Petri dishes (one for each metal), watch glass, glass bristle brushes (one for each metal), triangular-end scalpel, tweezers, pair of scissors. Scalpel, tweezers and scissors must be cleaned with acetone before use.

Cleaning procedure

In the case of rust marks remaining on the tube from a previous test, clean with dilute (20%) HC1 for several days. Place the tubes in a dilute nitric acid solution(10% by volume) or Decon 90 'Acid Rinse' solution overnight. Wash the tube in water with Centiclean Super Detergent, using a brush. Put the tube to dry in an oven at 60°C. Pass the tube through a Bunsen burner to eliminate any organic residue.

Preparation of the tubes

Insert 2 g (± 10%) of the material at the bottom of each tube. Place also, at the bottom of the tube, the small tube (0.8 ml) filled with distilled water, closed with a small piece of cotton wool.

Preparation of the metal coupons

For each metal foil (Ag, Cu, Pb), abrade the foil surface very roughly on both sides with a glass brush — the objective is to remove any oxide layer from the metal surface. Cut one coupon of each metal foil to size (10 x 35 mm). Clean the coupons in acetone by placing them in a Petri dish filled with Spectrosol grade acetone and cover the Petri dish with a watch glass. Hold the coupons with tweezers and dry them with a tissue (do not leave the coupon to dry off by itself).

Preparation of the stopper (Figure 1)

Cut three slots, about 5 mm deep and 10 mm across, in the bottom of the silicone stopper with the scalpel. Insert each metal coupon in a slot, placing the lead coupon in the middle (ensure that the coupons are not touching one another). An easy way to insert the coupon is to squeeze the stopper, which will open the slots.

Closing the tube (Figure 2)

Place the coupons in the tube by closing the tube with the stopper which holds the coupons; make sure the tube is perfectly sealed by pushing the stopper, ensuring coupons are not touching the glass of the tube. Place the tube upright in an aluminium tray and place the whole in the oven at 60°C for 28 days.

A control tube is made in exactly the same way, omitting any test material. One control tube is needed for the batch. If any corrosion appears on control coupons, the tests should be repeated.

Notes

In order to avoid any contamination by absorption of harmful gases, such as acetic acid from wood, the stoppers must be stored in an inert container.