AN INVESTIGATION OF THE REMOVABILITY OF NATURALLY AGED SYNTHETIC PICTURE VARNISHES
SUZANNE QUILLEN LOMAX, & SARAH L. FISHER
THE REMOVABILITY of two aging methacrylate varnishes, poly(isoamyl methacrylate), known as 27H, and various poly(n-butyl methacrylate)s, has been a concern at the National Gallery of Art due to the relatively large number of paintings covered with these coatings and the knowledge that they cross-link in accelerated aging experiments, rendering them insoluble (Feller et al. 1981).
The traditional natural resin varnishes, such as dammar resin and gum mastic, form brittle, yellowed films as they deteriorate (Feller 1958; Constable 1979; Thomson 1957; LaFontaine 1979). It has been noted that photochemical autoxidation is the method by which these natural resins undergo degradation (de la Rie 1987b, 1988). In general, the oxidation proceeds to a degree at which time toluene-acetone mixtures (40:60 v:v) are required for varnish removal (Feller and Curran 1975). On the other hand, some synthetic varnishes are suspect under the principle of reversibility (AIC 1985), requires that materials used in a conservation treatment, such as a picture varnish, should be removable without damage to the object (Gettens 1935; Feller 1960, 1959, 1957b). Some also seem to violate the principle of having the appropriate aesthetic qualities (de la Rie 1987a; De Witte et al. 1981).
All natural and synthetic varnishes tend to be less soluble in solvents of low polarity over time due to oxidation. However, synthetic varnishes may also become less soluble due to the phenomenon of cross-linking. Most synthetic varnishes are made up of long-chain polymeric structures consisting of molecules of hundreds of repeating units. If these long chains become chemically linked, the polymer eventually becomes insoluble in the solvents in which it was originally soluble. When the number of cross-links is relatively low, the polymer can be removed by swelling it by the proper solvent, followed by mechanical removal. If the degree of cross-linking has progressed to an advanced state, this method of removal is no longer possible. The cross-linking reaction is activated by ultraviolet radiation (Morimoto and Suzuki 1972), but it has also been shown to proceed in the presence of visible light (Feller et al. 1981; Feller 1967). It has been observed, on the basis of artificial aging studies, that the methacrylate polymers most sensitive to cross-linking have the alkyl side chains of isoamyl, isobutyl, 2-methylbutyl, and n-butyl (Feller et al. 1981; Feller et al. 1985). Side chains possessing tertiary hydrogen atoms exhibit greater reactivity in the cross-linking process, as these hydrogen atoms are more readily abstracted by the free radical species involved in the process.
The two varnishes examined in this study are 27H (poly(isoamyl methacrylate)) and poly(n-butyl methacrylate). According to the conservation records, 27H was reported to have been applied to 92 paintings in the period 1952–57 at the National Gallery of Art. Most of these are “old master” paintings that at least since 1952 have been exposed to light under normal exhibition conditions. In this paper, the term “old master” refers to works by great European artists of the 16th to 18th centuries. The conservation records also show that a poly(n-butyl methacrylate) varnish was applied to 184 additional paintings now in the National Gallery of Art collection. This varnish was applied between 10 and 40 years ago either by National Gallery of Art conservators or by private conservators in the case of paintings not then in the gallery collection.
The varnish 27H was developed in the early 1950s as part of a Mellon Institute research project at the National Gallery of Art (Feller and Raynolds 1953). The resin was never commercially manufactured but was prepared in batches and distributed to selected conservators. It was observed in 1955 that 27H did not completely depolymerize on heating, a situation attributed to cross-linking of the polymer (Feller 1955a). (An excellent review of the cross-linking studies of 27H and other methacrylate polymers can be found in Feller et al. 1985, appendix A.) Samples were exposed to accelerated aging conditions (National Accelerated Fading Unit, Type XV, Corex filter) and were subsequently found to be insoluble in mild solvents such as hexane, benzene, or toluene (Feller 1955b). UV absorbers were added (Feller 1956), among them Uvinul 400 and Uvinul D-49, but their use was discontinued because they led to yellowing of the varnish, and the original formulation of 27H was no longer produced after 1957 (Feller 1957b).
Poly(n-butyl methacrylate) has been used as a picture varnish since the 1930s. It is known under a wide variety of trade names, including Lucite 44 (DuPont), Elvacite 2044 (DuPont), Acryloid F-10 (Rohm and Haas), Acryloid B-66 (Rohm and Haas, a terpolymer of poly(n-butyl methacrylate) and an ethyl methacrylate/methyl acrylate copolymer) (Feller 1971–72), M varnish (Ralph Mayer), Soluvar (Permanent Pigments) (Feller 1973–74), and Synvar (Weber). Feller et al. (1981) has suggested that there is an induction time of 11 years before insoluble matter begins to form in commercial n-butyl and isobutyl polymers on well-illuminated museum walls. Testing revealed that n-butyl and isobutyl methacrylate polymers develop 50–80% insoluble matter after exposure to 9 million footcandle hours of radiation from daylight-type fluorescent lamps. This exposure is approximately equal to 50 years of exposure to daylight under an average illumination level of 50 footcandles. Although ultraviolet filters (Plexiglas UF3) were installed at the National Gallery of Art in 1963, thereby extending the suggested lifetimes of these coatings, concerns remained about the removability of the methacrylate varnishes.
In an attempt to understand the aging behavior of these coatings, removability studies were carried out on paintings believed to be varnished with them. It should be pointed out that these tests were undertaken only to determine qualitatively the removability of the varnish. Neither the degree of cross-linking nor the amount of insoluble material was determined, but rather the solvent required to remove the varnish. It is obvious that some insoluble material can be removed by the mechanical action of rubbing (Feller and Curran 1975).