The Mode of Formation of Thecotrichite, a Widespread Calcium Acetate Chloride Nitrate

The Mode of Formation of Thecotrichite, a Widespread Calcium Acetate Chloride Nitrate Efflorescence

Lorraine T. Gibson, Brian G. Cooksey, David Littlejohn, Kirsten Linnow, Michael Steiger and Norman H.Tennent

The widespread occurrence of thecotrichite, in the museum environment is explained theoretically by construction and examination of its phase diagram. Thecotrichite formation was simulated in the laboratory to identify the key factors involved in its production. This efflorescence occurs on porous limestone or calcareous artefacts such as pottery, stored in wooden cabinets that generate acetic acid vapour. Salt production depends on the moisture content of the object and the concentration of acetic acid in its surroundings. Furthermore, for thecotrichite to form the artefact must contain soluble chloride and nitrate salts.

INTRODUCTION

Porous materials such as stone or brick absorb moisture at times of high humidity [1]. With fluctuations in the ambient relative humidity (RH), soluble salts may be drawn in liquid form to the surface of the object by capillary forces and evaporate to form white crystalline deposits. These deposits, termed 'efflorescence', occur in a variety of crystalline habits, generally related to the nature of the salt, most commonly prisms and acicular, whisker-like crystals [2, 3]. Whisker growth may occur in any type of crystal, irrespective of whether the structural form is predicted to be prismatic or cubic [3]. These efflorescent salts, most often inherent in the material of the objects, can arise from a variety of sources [1, 2]: the burial environment of the object; absorption of salt-laden groundwater into building materials such as stone, brick and tiles; or from previous cleaning or conservation treatments. They are mainly chlorides, nitrates, sulphates and carbonates of the alkali and alkaline earth metals [2].

In contrast, the particular type of efflorescence studied here involves the formation of crystalline surface deposits by reaction of the material of the object, and impurities within it, with acetic acid pollutant vapour. The formation of salt efflorescence on porous calcareous artefacts, as a result of attack by organic acids, has been observed for many years. Efflorescence on shell collections, known as Byne's disease, was first documented in 1896 [4] and investigated more fully m 1899 [5]. It took a further three decades, however, before it was recognized that acetic acid vapour, emitted from oak wooden drawers was responsible for the formation of calcium acetate salts on the surface of the shells [6]. Since then, a number of studies have related the presence of organic acid vapours, caused by the materials used to construct museum cabinets, to pollution-induced deterioration (see, for example [7—10]). In some cases these crystalline salts are novel chemical compounds whose only known occurrence is as efflorescence on museum artefacts [8-10]. A wide range of porous materials (for example fossils, pottery and limestone items) stored in wooden cabinets are adversely affected by the production of acetate efflorescence. One of the most ubiquitous acetate efflorescences, designated calclacite, Ca(CH3COO)Cl·5H2O [11], has been frequently observed and reported, notably on a range of Greek and Roman objects where much of the pottery had been previously treated with hydrochloric acid [12].

Another unusual form of acetate efflorescence has also been found to occur widely on a number of museum objects [9, 10] and is the subject of the studies reported here. It was first referred to by FitzHugh and Gettens [10J in an article that cites Van Tassel's description of this efflorescence as 'fibrous crystals showing polysyn-thetic twinning, negative elongation, monoclinic or tri-clinic symmetry, oblique extinction about 20°, with refractive indices d = 1.504 and a = 1.492'. Chemical analysis identified calcium, acetate and chloride ions in the salt, but full chemical characterization was not achieved at this time and it was designated as efflorescence X [10]. It "was not until a subsequent investigation using ion Chromatographie analysis that efflorescence X was identified as Ca3(CH3COO) 3Cl(NO3) 2·7H2O [13]. Interestingly, the composition of the compound, which we have named 'thecotrichite' — a hairy compound from a storage cabinet, from the Greek thêkê (chest) and triches (hair) — was shown to be consistent on 11 limestone or ceramic items from different collections around the world [14]. Comprehensive analytical, infrared and X-ray diffraction data have been reported [13].

The widespread occurrence of thecotrichite has since become evident. Samples from numerous collections in Europe and North America have been identified. In particular, many samples previously examined by X-ray diffraction at the Natural History Museum, London, but never identified are now known to be thecotrichite [15]. As thecotrichite is now believed to be as ubiquitous as calclacite in the museum environment, a full investigation into the mode of formation and stability of this complex efflorescent salt was undertaken. It had been proposed that, for thecotrichite formation to occur, a calcareous porous material must be contaminated with chloride and nitrate salts and exposed to acetic acid vapour [13]. The acid vapour is responsible for dissolution of calcium carbonate from the matrix of the material, and the resultant solution containing chloride, nitrate, acetate and calcium ions then forms thecotrichite. This study was initiated in an attempt to confirm this supposition. Theoretical studies were also undertaken to elucidate the phase diagram of thecotrichite.

METHODSAND MATERIALS

Impregnation of limestone blocks with chloride and nitrate salts

Oolitic limestone blocks (8x5x1 cm) were repeatedly steeped in distilled water to extract any soluble salts from the body of the limestone. Ion Chromatographie analyses of the water confirmed the absence of ions in the washings after five extractions over seven days. The blocks were then air dried and weighed. To estimate the porosity of the limestone, the maximum mass of water held by each block of limestone was determined. All blocks were found to be similar; each block absorbed approximately 5% of water by weight of the limestone. After drying, four limestone blocks were impregnated with a solution containing low (1:1) or high (1:2) molar ratios of calcium chloride and calcium nitrate, with either low or high weights of impregnation (Table 1). Six sets of similarly impregnated limestone blocks were prepared. Using these conditions, calcium ions from both the salt solutions and the dissolution of calcium carbonate will contribute to the production of thecotrichite. It should be noted that the addition of calcium salts to a limestone block would not eliminate the corrosion effect (acetic acid reacting with calcium carbonate on the surface of the object) observed in real situations. Moreover, the use of calcium salts enabled saturation of the solution with calcium in order to produce thecotrichite in a reasonable time frame.

Table 1 Preparation of impregnated limestone blocks in an attempt to synthesize thecotrichite.

Generation of acetic acid environments

Two different volumes of glacial acetic acid were added to saturated solutions of sodium chloride (75% RH) or 40% v/v calcium chloride (33% RH). The resulting acidic salt solutions were placed at the bottom of glass desiccators to create environments contaminated with 150 or 10 mg·m-3, equivalent to 61 or 4 parts per million (ppm), acetic acid at relative humidities of 75 or 33% RH (Table 2). These are the theoretical relative humidities; the actual RH was not measured with

Table 2 Creation of four environments contaminated with acetic acid vapour.

hygrometers or dataloggers due to the high corrosiveness of the environments. Changes to the final vapour pressure of the salt solutions, caused by the addition of small quantities of acetic acid, were neglected as the intention was to provide high or low humidity environments as opposed to those with exactly 75 or 33% RH. Although wooden cabinets in the museum environment have been found to give rise to concentrations generally below 2.45 mg·m-3 (1 ppm) [9], higher concentrations were used in this study to permit crystal formation in a relatively short period of time (a few months). Over time, as the acetic acid vapour reacts with the calcareous materials inside the desiccator, the acid concentration will decrease due to exhaustion of the acetic acid in the salt solution. Despite this, solutions were not replenished during the experiment as it is thought that vapour phase concentrations might also decrease in real situations.

Determination of the experimental acetic acid vapour concentration

The vapour phase concentrations, generated in the glass desiccators, were measured accurately using passive diffusion tube samplers as described elsewhere [16], with one modification to the method of tube preparation. Instead of glycerol, ethylene glycol dimethyl ether (Fluka, UK) was used as a wetting agent in the 1 M potassium hydroxide trapping solution. This provides a more stable solution for the trapping of acetic acid vapour.

Ion chromatography

Samples of the efflorescent salts were analysed using ion chromatography. A known weight of the sample was dissolved m 10 mL of distilled water. Prior to injection into the ion Chromatograph, all solutions were filtered using inorganic membrane filters (Anotop 10 C, 0.2 mm pore size, 10 mm diameter, Whatman). Salt solutions were injected in quadruplicate and the mean results are presented. The calibrant ion solutions were prepared by diluting 1000 mg-mL-1 stock solutions of each analyte.

A Dionex 4000i ion Chromatograph was used with an AG4A guard and AS4 columns for analysis of acetate ions, an AG5 and AS5 set of columns for analysis of chloride and nitrate ions, and CS12 guard and separator columns for cations. The eluents for anion determination, at a flow rate of 2 mL per minute, were 6 mM disodium tetraborate decahydrate (Merck, UK) or a solution of 0.75 mM sodium carbonate and 2.2 mM sodium hydrogencarbonate (Dionex HiPerSolv vial, Merck, Poole). For cations, an 18 mM solution of methane sulphonic acid (Fluka, UK) eluent was used at 1 mL per minute. All eluents were purged with helium for 20 minutes prior to use. Analytes were detected after chemical suppression by a Dionex CDII conductivity cell set to a detection range of 30 μS.

The relative standard deviation (R.SD) values of signals obtained by replicate injections were normally less than 2%. Recoveries of analytes, measured by analysing solutions of efflorescence spiked with 5, 5, 4 or 8 mg-mL-1 of calcium, acetate, chloride or nitrate ions, respectively, ranged from 96 to 104%.

Calculation of stoichiometric formulae from ion Chromatographie analyses

A full explanation of the method used to determine the stoichiometry of efflorescent salts by ion chromatography has been published elsewhere [13]. The calculation of the stoichiometric ratios of anions (nia) and cations (nic) can be summarized by the following equations:

where mia and mic are the molar masses of the ith anion and cation, zia and zic are the charge on the ith anion and cation, Σviazia and Σviczic represent the theoretical total number of anion and cation equivalencies, and EeTa and EeTc represent the total experimental number of anion and cation equivalencies, respectively, in a solution as determined by ion Chromatographic analysis. The concentrations of sodium, potassium and calcium ions were measured in every solution by ion chromatography. However, in each case the concentration of calcium ions was > 99%, hence for simplicity the molar ratio of the minor ions, sodium and potassium, are not reported here.

RESULTS AND DISCUSSION

Production of thecotrichite using impregnated limestone blocks in acetic-acid-contaminated environments

Using the controlled acetic acid environments, attempts were made to produce thecotrichite on blocks of oolitic limestone. Four acetic-acid-contaminated environments of approximately 150 or 10 mg·m-3, with either a low (33%) or high (75%) RH were prepared (Table 2). Six sets of four blocks were impregnated with chloride and nitrate as outlined in Table 1 and placed in desiccators A to F, where A to D contained acetic acid as indicated in Table 2. Desiccators E and F were control desiccators with no acetic acid, but a RH of 75% or 33%, respectively.

The limestone blocks were monitored visually. After only three weeks, a small amount of crystals was clearly visible on one of the limestone blocks stored in desiccator A, which had an acetic acid environment of approximately 150 mg·m-3 at 75% RH (see Table 2). After 18 months, all four limestone blocks in this environment had produced crystals (Figure 1), which grew up to approximately 2 cm in length (Figure 2). The highest amount of crystals was produced on the block impregnated with a 1:1 molar ratio of calcium chloride: calcium nitrate and a weight of impregnation of 0.5% w/w. Different crystal habits were observed on the blocks, despite the relatively stable RH over the period of the experiment. Acicular and fibrous crystals, and compact encrustations were observed on the four limestone blocks in desiccator A.

Sub-samples of the salts were carefully removed under a microscope, weighed and dissolved in distilled water to provide salt solutions for ion Chromatographic analysis (fibrous crystals and encrustations were removed separately from block 4). The stoichiometries of the salts were determined as described previously. The results (Table 3) show that, despite differences in the molar ratio of chloride and nitrate salts, and the weights of impregnation, all of the synthetic salt samples had a composition similar to thecotrichite found in the museum environment. Interestingly, block 4 had a higher concentration of nitrate and chloride (compared to acetate), which may suggest the presence of the double salt CaCl(NO3)·2H2O in combination with thecotrichite. However, a solution of this composition is extremely hygroscopic and it is expected that no solids would precipitate at humidities greater than 21% RH. Therefore it is suggested that the salt was contaminated with a small amount of the mother liquor. At the end of the experiment, after 18 months, salt formation was not observed on any of the similarly impregnated limestone blocks stored in desiccators B, C or D, nor was salt formation observed on any of the limestone samples held in the control environments in desiccators E or F.

The above experiments suggest that the combination of high RH and high acetic acid concentration has a profound efFect on the rate oí thecotrichite formation on contaminated limestones. Where the acetic acid concentration was high, but the RH low (desiccator B) no efflorescence formation was observed. Equally, when the acetic acid concentration was low, regardless of high or low humidity conditions (desiccators C and D, respectively), salts were not observed on the surface of the limestone, despite having been impregnated with calcium chloride and calcium nitrate. When the acid concentration and RH were high enough to produce efflorescence, the stoichiometric compositions of the salt produced were similar, at least for the 1:1 and 1:2

Figure 1 The four limestone blocks that were placed in desiccator A for a period of 18 months. Figure 2 Close-up of crystals growing on the side of one of the limestone blocks.

Table 3 Stoichiometric formulae of synthetic samples of thecotrichite grown in dessicator A.

molar ratios at different weights of impregnation studied here.

Analysis of objects that have promoted thecotrichite formation

The above experiments demonstrate that thecotrichite is a stable salt and will commonly crystallize on limestone or porous materials such as pottery containing calcium carbonate when environmental conditions permit. However, in practice, it is often the case that a number of such materials are held in the same contaminated storage or display environment, but are affected to very different extents. Many instances have occurred where a range of shells [4—6, 8], birds' eggs [7] and earthenware objects [9] have been damaged by salt formation, whereas neighbouring objects have remained in a pristine condition. Obviously, in addition to the surrounding atmospheric environment, the nature of the material must also be considered.

Material was carefully removed from two previously studied objects that supported thecotrichite formation; an Egyptian limestone relief and a Dutch tile [14]. Limestone scrapings were obtained for a series of depths up to 20 mm from the edge of the limestone relief and up to 7 mm from the back of the Dutch tile. The samples were weighed, submersed in water and the extracted solutions were analysed using ion chromato-graphy. The results for both artefacts (Figures 3a—3c) indicate a decrease in all three ions inwards from the surface. The compact nature of the fired clay body of the tile enabled a greater number of discreet scrapings to be taken, thereby demonstrating the sharp decrease in anion concentration from the surface to a depth of about 7 mm. For the limestone relief, it was impossible to obtain a sample of the friable surface layer that was not intermingled with crystals of efflorescence and so these results are not reported in Figure 3. These observations suggest an external source of acetate and that chloride and nitrate ions exist deep into the body of the artefact.

Figure 3 Concentration profiles of (a) acetate ions, (b) nitrate ions and (c) chloride ions extracted from limestone scrapings of an Egyptian limestone and a Dutch tile.

Enrichment of chloride and nitrate close to the surface of the tiles may be explained by capillary transport and evaporation, consistent with their removal from relatively damp walls to a more controlled museum environment. Further evidence was obtained by monitoring the environment surrounding the Egyptian limestone relief. Passive samplers were deployed in the wooden cabinets for 14 days. The concentration of acetic acid measured was approximately 11 mg·m-3, but less than 0.005 mg·m-3 of nitrogen dioxide was measured. It is also interesting to note that formaldehyde and formic acid were present at appreciable concentrations of 0.4 mg·m-3 and 0.3 mg·m-3, respectively, yet no formate was present in the efflorescent salt.

A sample of the fired clay matrix removed from the Dutch tile was also examined by X-ray diffraction analysis. A second tile was also sampled for comparison. This tile was located in the same area as the first but no thecotrichite efflorescence had formed. Although both tiles contained significant quantities of quartz, only the tile that supported thecotrichite formation contained calcite. These observations suggest that more highly fired (and consequently more vitreous) wares, which do not contain calcite, are resistant to formation of acetate efflorescence. Therefore to induce thecotrichite formation porous materials not only need to be contaminated with soluble salts prior to acid exposure, but must also be comprised of or contain calcium carbonate.