JAIC 1997, Volume 36, Number 3, Article 3 (pp. 207 to 230)
JAIC online
Journal of the American Institute for Conservation
JAIC 1997, Volume 36, Number 3, Article 3 (pp. 207 to 230)




For the paint or textile industry seeking to determine the lightfastness of a material, testing can be done to measure slight color changes in carefully standardized materials, thus avoiding complicated fading behaviors and the dependence on factors such as the depth of shade. Conservators face the exact opposite of this idealized situation, in which the materials in a collection will inevitably vary widely in their color and in their prior fading history. Even knowing the pigments in a paint, one can probably only estimate to within a few grades the actual lightfastness of any given application (in other words, one may only be able to estimate the fading rate to about a factor of 10). With such a wide range of possible light sensitivity, it is exceedingly difficult to make any but an informed guess as to whether extensive fading will result from exhibition under specified lighting conditions. This is a particularly troublesome situation when the lighting conditions cannot be altered substantially and the decision is essentially whether or not to exhibit an object.

In response to this situation some institutions have considered implementing programs to carefully track the color of exhibited objects in order to assess the actual lightfastness of the materials and to better judge the consequences of exhibition. Aside from the technical challenge of making precise color measurements periodically at long intervals, such efforts often force some selection of objects or colored materials in order to make such a monitoring program practical. With this in mind it is useful to consider whether one can choose those types of paint applications that might be most at risk from light damage and thus be sensible targets of a color monitoring program.

The experiments described in this article indicate that for a given colorant, glazes having intermediate reflectance minima will tend to display the largest color changes upon fading. For glazes that are already only partially absorbing across the spectrum, almost any property (if measured precisely) can reliably track the progress of fading. Color differences (ΔE); changes in lightness (ΔL∗), tristimulus values, or Munsell chroma or value; or changes in reflectance at the absorption peak or across the spectrum will all be sensitive and reliable indicators of color stability.

For the high chroma glazes that reflect very little of the light at the absorbed wavelengths, on the other hand, many of the usual measures to indicate color changes may not be sensitive to the colorant fading. Colorant loss in these glazes does not result in the conventional appearance changes associated with “fading” but instead results in hue shifts or chroma increases. These changes will be much less distinct than the value increase or chroma decrease in fading glazes having intermediate reflectance minima. The quantities that change for dark glazes (such as the high chroma) may be difficult to measure precisely, or they may be affected by the details of the colorant reaction (such as the hue shift). The best monitor for the colorant loss in these highly absorbing glazes over light-colored substrates may be the measurement of changes in the entire reflectance spectrum. The relative precision of the spectral measurements should allow a good determination of changes in the profile that reflect the progress of the fading. This profile may also be the best indicator for when the reflectance of the absorbed wavelengths is about to rise above nearly zero (i.e., when the rapid second stage of fading is imminent). Another possibility to track the course of fading is to monitor the reflectance at a wavelength on the side of the absorption peak, where the reflectance is greater than zero. Here the reflectance will increase before changes can be observed at the peak absorption wavelengths. In fact, it was only by measuring the reflectance at these off-peak wavelengths that the chemical kinetics of the colorant losses could be followed in the highly absorbing glazes of this study.

While the glazes that are on the verge of the rapid second stage of fading do not display the greatest light sensitivity, it can be argued that they are most at risk of suffering objectionable light-induced damage. These glazes still have a high chroma and low value (i.e., they retain the appearance that probably motivated the choice of a glaze application in the first place). Further fading will put them at risk of entering the more rapid fading stage, when chroma will decrease and value will increase. Identifying these glazes (by their just barely saturated absorption peaks) and tracking the course of their fading and the approach of the rapid second stage of fading should be considered priorities for color monitoring programs. Again, the profile of the entire reflectance spectrum will be most useful for judging how saturated the absorption peak is.

Finally, it is tempting to interpret the results of this study as justification for assuming a greater stability for old, previously faded glazes, but it would be premature to use this study alone as the basis for such an assumption. The glazes tested here did in fact show lower sensitivities after prior fading, even as those glazes entered into the more rapid second stage of fading (figs. 11 and 12), perhaps through protection by the fading reaction products or through having already lost the most fugitive pigment components. Without much more study, however, it is not possible to assess how much more stable faded glazes of other colorant formulations may be, particularly those traditional organic colorants whose composition, and thus stability, may depend on the detailed chemistry of a plant or animal source. Further, it is impossible to judge accurately from the appearance of a glaze how much fading it may have experienced in the past, and in this study this characteristic was a determinant of the light sensitivity. While it would thus be imprudent to assume a greater durability for previously faded glazes and to exhibit them at higher light levels or for longer periods, it may still be sensible to make such glazes lower priority targets for color monitoring programs. Fresh, unfaded glazes, which retain a high chroma and may be on the verge of the rapid second stage of color change, may be more at risk and make more reasonable subjects for color monitoring.

Copyright 1997 American Institute for Conservation of Historic and Artistic Works