|
|
Forecast of Chemical Aging and Related Color Changes in Paintings
by Boris Zilbergleyt
Abstract
The article describes the potential application of thermodynamic simulation to the problems of chemical aging of painting. Qualitative and numerical results were obtained in a preliminary investigation by applying the method to various mixtures of pigments without and with atmospheric components. The results were compared to historic recommendations on incompatible pigment mixtures with about an 80% match regarding potential color changes in the aged mixtures of pigments. Results for the cadmium yellow-lead white and cadmium lemon-emerald green mixtures are illustrated by pictures, gradually showing color changes related to aging. The method of thermodynamic simulation can be a powerful tool to investigate old paintings, in developing new materials, in conservation and restoration, and to forecast some aspects of the aging of real paintings.
Introduction Color is, obviously, the most important element in a painting but also the most sensitive to degradation factors. From fading to darkening, color changes of paint layers that occur with time can alter the entire appearance and perception of a painting. Painting collections in world museums abound with altered canvases due to internal and external deterioration factors and unfavorable conditions that occurred before they were placed in the controlled museum environment. Although external deterioration agents, such as environmental factors (temperature, humidity, ultraviolet radiation, etc.) and biologic attack are permanently putting works of art at risk, their impact can be controlled and limited in many cases. However, in the long term, the major reason for color change of a painting is the inevitable chemical aging of the paint layers [1]. These changes include alterations in the optical properties of binders, and in the chemical and structural composition of pigments due to chemical interactions between them, assisted by atmospheric species [2]. Chemical deterioration, resulting from these interactions appears to be immanent to the aging process. The aging patterns of paintings have been intensively studied, and seem to be quite clear on a qualitative level [3, 4]. Paintings are composed of a complex of grounds, pigments, organic binders, and varnish and in an ideal case the fresh paint layer contains well-encapsulated pigment particles. The particles are separated from the support (canvas, wood, etc.) by the preparation layer, from each other by the binder, and from the atmosphere by both the binder and varnish layer. As long as the ground, binder and varnish are intact, routine changes of paint layers are extremely slow. But even in the most favorable storage conditions, the upper layer often ages faster than the inner layers. Initially, aging leads to a change in the optical characteristics of the varnish and to the formation of craquelure networks. Moisture and other atmospheric factors sharply accelerate chemical interactions between pigments and other components of the paint layer leading to intensive changes in its chemical and phase composition and to massive changes in the color, brightness and contrast of the painting. What can we do and to what extent can we prevent the chemical changes? Chemical aging of paintings has been extensively studied although a comprehensive understanding has not yet been achieved. We are aware of historic recommendations from old masters on the incompatibility of certain mixtures of inorganic pigments, such as lead white and ultramarine among others, based on the artist’s practical experience [5, 6]. The technical execution of a work of art is another important aspect to take into consideration, but in any case, we need to understand on a quantitative level the aging mechanisms of a painting. Thermodynamic Simulation Method The method of thermodynamic simulation consists of the computer simulation of possible chemical interactions within the paint layer to determine the ultimate chemical and phase equilibrium composition. It allows for numerical calculations of the most probable final composition of the paint layer, resulting from these interactions when all changes are over and the system rests at thermodynamic equilibrium. One can state that in old works of art most of the possible chemical processes, allowed by their pre-museum and museum storage conditions, are either completed or the relevant changes are already well pronounced. Knowledge of the chemical and phase changes in the layer may help to predict optical/color changes and structural damages such as detachment or flaking of the paint layer. The method can be implemented using most of the known simulation software; we used the simulation complex ASTRA-4 [7] and partly HSC Chemistry [8]. Atmospheric pressure (0.1 Pa) and a temperature of 293K (200C) were taken as simulation parameters. In case of pair mixtures of pigments, the mass ratio of main to admixed components varied from 10:1 to 1:1. Taking into account the relatively long lifetime of a painting, atmospheric air as a natural mixture of oxygen and nitrogen was present in the initial compositions up to 10%, with moisture up to 5% of the air mass and some typical pollutants like CO, CO2, sometimes SO2 and H2S up to typical concentrations for urban areas. In cases when not enough thermodynamic data was available to involve some complex pigments of interest into simulation, we used their essential and often the major color carrying fragments. For example, in case of lead white we used the data for lead carbonate only; in this case simulation was carried out with the fragment of the pigment with known data instead of the whole compound. The same situation occurred in case of ultramarine and several others. Though in most cases this was good enough to judge the possible color changes, the results of this work should be considered qualitative for manifesting the method’s ability to predict and describe the chemical aging of paintings. Simulation Results For several chosen groups of pigments we tried chemical interactions within the group between the comprising pigments along with simultaneous interactions of these pigments with atmospheric components; lead white and zinc white with some pigments in presence of atmospheric components; oxide pigments in mixtures with sulfide pigments; special pairs of pigments to check the historic incompatible pigment couples, and some complex mixtures from Rubens palette. Among these mixtures, the incompatible pigments group is one of the most interesting in the light of this work. Corresponding results are shown in Table 1. In some cases, when both dark and white substances were formed together in the mixture, such as FeS, CaS, and Na2CO3 in ultramarine mixtures with ochres and umbers, it was difficult to evaluate if optical changes would happen. Also, gaseous reaction products are not shown in the table. Even with this restriction, in about 80% of the investigated mixtures the predicted possible optical changes have certainly matched the historic alterations. What is remarkable is that the method of thermodynamic simulation also explains possible reasons of the changes in terms of chemical and structure composition thus allowing for visual interpretation. Percentage of the matching Y/N compatibility results in Table 1 is high enough to prove the applicability of the method to the analysis of chemical aging of paint at least on the qualitative level. The quantitative results, even if more difficult to obtain, will definitely offer more information. Table 2 contains numeric simulation results of the incompatible mixture of lead white (represented by lead carbonate) with yellow cadmium (represented by cadmium and zinc sulfides). The abbreviations for colors are B for black, W for white or light colors, and Y for yellow. To account for reduced thermodynamic activity of the lead carbonate (PbCO3) due to its binding into lead white (2PbCO3.Pb(OH)2), its thermodynamic activity coefficient was reduced to 0.6 for simulation. One can see changes in the chemical composition of the mixture due to decay of input basic color carriers, certain structural changes due to the presence of new components, and drastic color changes in the presence of black lead sulfide. Figure 1 shows color visualization of the simulation results, RGB indices were calculated using data for yellow cadmium from [9]. Mixtures’ color was calculated as weighted mean of their components [10]. Figure 2 shows the aged samples of the cadmium lemon mixture with emerald green, compared to the initial colors (upper row); numbers on the swatches show initial moles ratio cadmium/emerald green in the mixture. Interestingly, that cadmium pigment almost totally disappears in all investigated mixtures after aging; its increase in the initial mixture leads to more and more pronounced grey color due to formation of dark, fine dispersed particles of Cu3As and Cu2O along with several white products. Chemical Aging of Some Rubens’ Pigment Mixtures Rubens was one of those rare artists who thoroughly wrote down the major components of the mixtures he used in various purposes, sometimes with a kind of functional names [11], that allows us to experiment with his palette. Qualitative results of the simulation of the mixtures aging in presence of normal air (with typical presence of moisture) are placed in Table 3.
Table 1. Compatibility of various pigments in pair mixtures. N – not compatible, Y – compatible, Leg. - legendary, Sim. - simulated. Black or just dark colored species resulting from the chemical interactions are typed in bold. Conclusion This work presents the results of a computer simulation aging experiment in the field of paintings, where the information was collected from testimonies of old masters and from paintings, where precise initial composition may not be known. The data related to paintings aging and pigments compatibility is transmitted and may vary from one bibliographic source to another, sometimes without clear references. Certain information definitely can be achieved after centuries of natural aging or using accelerated methods, usually run at conditions different than the natural ones, such as elevated temperatures and moisture [1]. To the best of the author’s knowledge, our approach was the first attempt to use an experimental computer simulation method to predict chemical aging. Discussing the results of this work, one should keep in mind that thermodynamic simulation brings about the ultimate results of aging–potential chemical and structural changes providing that all possible interactions in the ground-pigment-binder-varnish system are over, that is the whole system rests in thermodynamic equilibrium. Equilibrium compositions, in their turn, should be considered as the limit towards which the system moves, but not necessarily achieves it in reality. Chemical aging advances up to the latest stages of the painting lifetime but is just one of the possible contributors to the deterioration process. Although the scope of this investigation was restricted to inorganic pigments, the proposed method is applicable to analyze chemical interactions between any kind of substances including organic dyes, binders and grounds. The increasing abilities of computers essentially expand opportunities for simulation and output of the simulation results in such complex materials like paint mixtures. Besides that, the method can be used to investigate chemical behavior not only of a painting but also of any art materials that change by interacting with their environment. Although the method investigates only one important aspect of the aging of paintings, in certain cases its results can help conservators to achieve a better understanding of the aging behavior of compatible and incompatible pigment mixtures. Even though color alterations are generally accepted as part of the authenticity of a painting, this method may help to establish the appropriate conservation methodology. The application of the method of thermodynamic simulation in the field of aging is not easy because only rare pigments are chemically simple enough to have their thermodynamic properties ready in the regular thermodynamic data bases. Calculation or recollection of appropriate experimental information should be the first task prior to the method implementation. The author foresees many objections related to the results of this article. Among them, why most of the species, predicted in the run of that preliminary investigation, were never reported by previous investigators? Well, it could be that the works of art, investigated earlier on, didn’t achieve the extreme stages that could be qualified as equilibrium in the context of this article. Or, what if they were just overlooked because nobody suspected them to be present? Who might be interested in using this method? We believe it can definitely help to investigate possible reasons for the changes that occurred, and how aging will develop. It could be also useful for attribution purposes, solving the so called back task of simulation (common in geochemical simulation of the origination of minerals), that is to find out the initial composition of the paint mixture given the chemical analysis of the aged sample. The manufacturers of art materials could also use the method to evaluate their products in various mixtures and environments. In general, the method could be of a real help in finding proper conditions of the accelerated aging. As concerns to the artists, it’s of a very low probability that Jackson Pollock would ever be willing to hear about any simulation program, but Rubens definitely might be interested to use it. Acknowledgements The original preprint of this article, prepared in the main frame black-and-white computers era, was published in a small amount of copies by the Ministry of Culture of the USSR [12] with kind support of Dr. L. Gorel’chenkova. Recently some results were calculated anew to allow for color illustrations, and the paper in its new form was made available on the Cornell University Library site [13]. Current version contains some amendments as well as special updates for conservation professionals. The author is much obliged to the people of e_conservation magazine, whose attention, energy and help made this publication possible. References 1. R. L. Feller, Accelerated Aging: Photochemical and Thermal Aspects, D. Berland, Ed. The Getty Conservation Institute, Los Angeles (1994) http://www.getty.edu/conservation/publications/pdf_publications/aging.pdf, accessed on 20/10/2008 2. A Study of the Discoloration Products Found in White Lead Paint Films,
http://aic.stanford.edu/sg/bpg/annual/v04/bp04-04.html, accessed on 20/10/2008
3. Margriet van Eikema Hommes, Changing Pictures - Discolouration in 15th to 17th Century Oil Paintings, Archetype (2004) 4. Blakey, R. R. Evaluation of paint durability - natural and accelerated. Progress in Organic Coatings 13:279–96 (1985)
5. Cennini, Cennino, The Craftsman's Handbook http://www.archive.org/details/bookofartofcenni00cennuoft, accessed on 20/10/2008
6. Odnoralov N. Materials in the Visual Arts. Moscow: Enlightment (1983), pp.144 7. ASTRA-4. Modeling of chemical and phase equilibria (manual). Moscow: MGTU (1991), pp.56 8. Outokumpu HSC Chemistry. Finland: Outokumpu Research Oy, 2000, http://www.outotec.com/, accessed on 28/10/2008 9. Web Color Definitions. http://endprod.com/colors/, accessed on 28/10/2008 10. Color. Universal Language and Dictionary of Names, US Department of Commerce (1976) 11. Grenberg Y.I. Technique of painting. Moscow: Visual Arts (1982), pp. 539 12. B. Zilbergleyt, Simulation of the Chemical Aging of Painting with Computers, "Investigation of the Art Remnants, their Materials and Techniques", Moscow, Survey of the Ministry of Culture of the USSR (1989). 13. B. Zilbergleyt, Forecast of the Chemical Aging and Related Color Changes in Painting (2005)
About the author
Boris Zilbergleyt
System Dynamics Research Foundation Contact: sdrf@ameritech.net Boris Zilbergleyt, born in Ukraine, lived mostly in Russia and since 1991, in the USA. He has a bachelor's degree in Computer Science, a master's degree in Metallurgy and a PhD in Chemistry/Physical Chemistry. He worked as engineer and R&D scientist in the fields of Metallurgy, Chemical Engineering and Chemical Thermodynamics. His interests include research in discrete thermodynamics of chemical systems and painting materials. Currently he is affiliated with System Dynamics Research Foundation, Chicago, USA.
▲TOP
To save this article please
Permanent link to this resource: http://www.e-conservationline.com/content/view/701 |
