Van Gogh’s darkening legacy

Along with his expansive brush strokes, Vincent van Gogh’s (1853-1890) choice of vibrant and often unrealistic colours to convey mood and emotion were central to his unique style, one which had a powerful influence on the development of modern painting. The new-generation pigments of the 19^th century made it possible for van Gogh to create, for example, the rich yellows used in his celebrated Sunflowers. These striking shades, used in many of his works, contained one of these new pigments, called chrome yellow. Unfortunately, more than 100 years after it left van Gogh’s brush, chrome yellow has in some cases darkened visibly to a less than striking brown, a phenomenon that recently caught the interest of a group of scientists.

An international team led by Koen Janssens of the University of Antwerp, Belgium, believes that chemical changes to chrome yellow (PbCrO₄ · xPbO), brought about by exposure to ultraviolet (UV) light, are responsible for its colour transformation (Monico et al., 2011). The darkening of the pigment in sunlight has been known since its invention. Studies in the 1950s demonstrated that it is caused by the reduction of chromium from Cr(VI) to Cr(III) (see Figure 1, below). Until now, however, the precise mechanism was unknown, and the degradation products were uncharacterised.

Historic paint tubes

To address these unknowns, Janssens’s team began by collecting samples from paint tubes belonging to van Gogh’s contemporary, Flemish painter Rik Wouters (1882-1913). Some tubes contained unmixed chrome yellow paint, whereas others contained paint of a lighter shade of yellow, formed by mixing chrome yellow with a white substance. The researchers artificially aged the samples under UV light, expecting to observe a colour change after several months. To their surprise, in only three weeks, a thin surface layer of the lighter yellow paint had darkened significantly to a chocolate brown. The unmixed samples changed either comparatively little or not at all. “We were amazed,” says Janssens.

Having identified the sample most likely to be undergoing the fatal chemical reaction, the team subjected it to sophisticated analyses based on X-rays. Much of the work was carried out at the European Synchrotron Radiation Facility (ESRF)^w1 in Grenoble, France, where two techniques, XRF and XANES, were used to detect, with extreme sensitivity, the spatial distribution and oxidation state of selected elements in the paint samples (see box).

Analyses revealed that the darkening of the thin surface layer of pigment was linked to a reduction of the chromium in chrome yellow from Cr(VI) to Cr(III); this fits with what has been observed for industrial paints based on lead chromate. In addition, the Cr(III)-containing degradation product was identified for the first time as Cr₂O₃ · 2H₂O, better known as the pigment viridian green. But how can a green pigment’s presence explain the brown colouration observed in the researcher’s experiments? The scientists suspect that the reduced chromium in viridian green is formed during the oxidation of the oil component of the paint. It is this oxidised form of the oil, together with the mixture of green and any remaining yellow pigment, that may be the root of the brown colouration.

Using the X-ray techniques, the researchers were also able to show that the mixed, lighter-coloured paint contained sulphur compounds. They concluded that these compounds were somehow involved in the reduction of chromium, explaining why there was comparatively little darkening in the unmixed paint samples.

Shining the X-ray beam on van Gogh

Having uncovered the chemistry of the reaction in isolated paint samples, the scientists sought to ask whether the darkening of the surface layer of yellow paint in samples taken from two of van Gogh’s paintings, View of Arles with Irises (1888) and Bank of the Seine (1887), could be attributed to the same phenomenon.

XRF spectroscopy was used to map the chemistry of the region encompassing the interface between the dark surface layer and the underlying unaltered yellow layer of paint. XANES spectra were collected at specific points within these regions. The findings mirrored those of the previous experiment: the reduced form of chromium, Cr(III), was found in the darker surface layer, suggesting that its presence here was responsible for the brown colouration. Furthermore, Cr(III) was not distributed uniformly, but occurred in loci that also featured sulphate- and barium-containing compounds.

Chemically, these regions resembled the lighter yellow paint samples from the previous experiment, further supporting the researchers’ conclusion that sulphur compounds were involved in reducing chromium (see equation below). Because of their white colour, van Gogh blended powders containing such compounds with chrome yellow to create the lighter shades that were vital in creating the brightly lit scenes characteristic of a certain period of his life.

One important question remained: how does the supposed trigger for the reaction, UV light, actually work? Quite simply, it supplies the reactants with the energy needed to overcome the activation energy barrier, allowing the reaction to proceed (see Figure 6, below).

What can be done?

Janssens’s team has exposed the chemistry that underlies the darkening of van Gogh’s paintings. But can we use this knowledge to rescue the artist’s work? Ella Hendriks of the Van Gogh Museum^w3 in Amsterdam, the Netherlands, has her doubts: “Ultraviolet light…is already filtered out in modern museums. We display the paintings in a controlled environment to maintain them in the best possible condition.” Part of what constitutes a controlled environment is the maintenance of a low temperature in the museum. As a general rule, an increase of 10 ºC increases the rate of a reaction by a factor of 2-4, and reduction of chromium is no exception to this.

So if both UV levels and temperature are already controlled, what more can be done for van Gogh’s paintings? There is a more radical alternative: rather than slow the degradation process, attempt to reverse it altogether. “Our next experiments are already in the pipeline,” says Janssens. “Obviously, we want to understand which conditions favour the reduction of chromium, and whether there is any hope of reverting pigments to their original state in paintings.”^w4

Although turning back the hands of time in this way would be the supreme solution, Janssens admits that the prospect of reverting the altered pigment to its original colour is at present rather unlikely. Nevertheless, the scientists’ work offers us reassurance that we are doing everything we can to preserve van Gogh’s paintings, and hope that future generations can appreciate what this great artist achieved.

Studying art with a synchrotron

The chemical characterisation of precious works of art can be problematic. It is only possible to take a few very small samples for analysis, and these often consist of a diverse mixture of complex compounds in heterogeneous states of matter. To overcome these challenges, scientists use techniques based on X-rays. The more powerful and precise the X-rays are, the better the quality of the analysis. The most potent X-rays available are produced by a synchrotron source^w2 (see Figure 2, below). In this study, two spectroscopic techniques at ESRF were used on the paint samples: XRF and XANES.

 

XANES

XANES spectroscopy relies on the physics of X-ray absorption. Atoms of a particular element absorb X-rays in a characteristic way. By looking at the X-ray absorption spectrum, which is the pattern of X-ray absorptions of a particular sample (Y axis) against the energy range of the X-rays (X axis), it is therefore possible to identify the sample’s constituent elements. High-resolution X-ray absorption spectra are usually collected in particular energy regions (called XANES) that are close to an absorption edge of an element of interest (see Figures 3, below, and 4). Such detailed spectra can show what oxidation state the element of interest is in. This information was of great interest to the researchers.

 

 

XRF

When they absorb X-rays, atoms enter an unstable excited state. When they then return to a more stable state, they emit secondary X-rays in a process called X-ray fluorescence (see Figure 5). The pattern of X-ray fluorescence (XRF) produced by a particular sample, called the XRF spectrum, can be used to map the distribution of elements across a given area. In contrast, XANES can only be performed on an isolated point in the sample. By combining the information obtained with both XRF and XANES, the authors were able to form a detailed picture of the chemistry of the paint samples.