Science

What light does to paint

Cochineal insects, Dactylopius coccus, the source of carmine red

Ship in a Storm by J. M. W. Turner, a sailing ship under stormy skies

Pieces of gamboge, a yellow tree resin used as a watercolour pigment

Why colours fade, which historic pigments suffered most, and how the ASTM and Blue Wool scales measure a paint's resistance to light.

Light as a chemical event

Most people have seen a poster gone pale on a sunny wall, or a sketchbook that sat on a windowsill and came back washed out. Conservators treat this as damage done by a dose. The exposure a surface receives is the brightness of the light multiplied by the hours it falls there, so a dim room over many years can do what a bright window does in a few weeks. The Canadian Conservation Institute puts it bluntly: there is no true recovery from a faded colour, only replacement.¹

Ultraviolet usually gets the blame, and it does real harm. It drives yellowing, chalking and the weakening of binders and paper. But the CCI's guidance is that visible light is what fades most sensitive colours, and that ultraviolet mainly adds to the fading of colours that already hold up well against ordinary light.¹ Filtering out UV helps, but it does not make a fragile colour safe.

What happens at the molecular level depends on the colourant. Light delivers energy that can drive chemical reactions in the molecules that give a paint its colour, and the product of those reactions may absorb light differently or barely at all. A study of Prussian blue in linseed oil, published in 2011, found evidence that iron ions at the surface of the pigment particles are chemically reduced during light exposure, which is one reason the blue weakens.² In cases like this the colour has not been washed away. The pigment has been changed into something else.

The pigments that suffered

Prussian blue is a useful example because it behaves unexpectedly. On its own it is very lightfast. Mixed with white, particularly lead white or zinc oxide, it fades considerably.² Researchers at the National Gallery in London tested Prussian blues made by different historical methods and found that the way the pigment was manufactured affected its colour change, that an alumina extender sped up fading, and that some of the colour change is reversible.³ So a pale Prussian blue sky in an old painting may be telling you as much about the white it was mixed with as about the blue.

The red lakes have a worse record. Rose madder, made from the root of the madder plant, was the standard transparent red for centuries. In 1868 the German chemists Carl Graebe and Carl Liebermann synthesised alizarin, its main colouring compound, and alizarin crimson became the lab-made replacement. It was more lightfast than the madder it replaced but still carries an ASTM rating of III, and it was eventually overtaken by the quinacridones, developed in 1958 and rated I.⁴ The CCI places alizarin lakes in its medium sensitivity group, with plant-extract lakes, insect reds such as cochineal and many early synthetic dyes like the anilines in the high sensitivity group.¹

Watercolour, laid thinly on white paper, shows the problem most clearly. Many of the bright colours of eighteenth-century watercolour came from plant extracts that faded. Indigo, used in mixtures for grey storm clouds, and yellows such as gamboge, used in greens, could shift the whole balance of a picture as they went. According to Winsor & Newton, William Winsor warned J.M.W. Turner that indigo would fade in daylight, and Turner told him that was his business, not Winsor's.⁵ By the later nineteenth century, longer opening hours and gas lighting in public galleries were exposing watercolours to far more light. Concern about the results led to a report on the action of light on watercolours by the chemist W.J. Russell and the photographic scientist Captain William Abney, presented to Parliament in 1888.⁵⁶

Measuring lightfastness

Two rating systems are in common use. Paint makers following the American standard rate pigments on the ASTM scale, where I means excellent lightfastness, II very good, and III to V progressively worse. The rating comes from exposing paint films to long periods of light and measuring how much the colour changes.⁷

The other is the Blue Wool scale, which began in the textile industry. German standards appeared in 1914 and British textile chemists refined them in the late 1920s. It uses eight strips of wool dyed blue, from 1, the most fugitive, to 8, the most stable. Each step needs roughly twice the light exposure of the one before it to fade by the same amount, so the gap between 1 and 8 is enormous.⁸ A test sample is exposed alongside the strips and rated against whichever one fades to a similar degree.

The CCI's figures give a sense of what these numbers mean in a real room. A colour in the high sensitivity group (Blue Wool 1 to 3) shows a just noticeable fade after somewhere between 1.5 and 20 years at the 50 lux of a carefully lit museum. Under the roughly 5,000 lux near a window or under a study lamp, the same change can take between five days and two months.¹ Almost complete fading is estimated at around 30 to 100 times that.¹ For anyone painting with fugitive colours and hanging the work in a bright room, the timescale is closer to months than generations.