Why are the trees blue?

Hello Everyone!

Over the past few weeks I have been working towards perfecting a procedure to simulate surface enhanced Raman spectroscopy of an organic yellow art sample.  Many organic yellow dyes produce beautiful SERS spectra with the traditional procedure, involving the coating of a minute sample with centrifuged silver nanoparticles and shooting a laser beam directly onto the particle.  However, lake pigments involve the mordanting of a dye with an inorganic binding medium, making it difficult for the silver nanoparticles to interact properly with the surface of the dye molecules and produce accurate spectra.

Lakes were frequently used in historic paintings for their luminous hues, despite their known status as “fugitive” or rapidly fading colorants.  The suspected presence of these lake pigments can be deduced through logical observation.  Assuming that these colonial portraits were naturalistic and that the artists weren’t getting too expressive with their color choices, trees and other plants should be painted green.  However, many paintings show foliage with a distinctly blue hue, suggesting that the artist mixed a colorfast inorganic blue colorant with a yellow lake pigment to create a vibrant, life-like green color.  Over time, the yellow has faded from exposure to light and the elements, while the mineral blue remains unscathed.  My job is to confirm these suspicions, and determine exactly what was in that ghost of a green plant.

Therefore, in order to visualize the spectra characteristic of the components of these lake dyes, I tried various methods of extracting the organic dyes from the inorganic binder.  The pigments available to me in lab were stil de grain lake, reseda lake, and a lake combining the two.  Stil de grain represents a more processed form of a dye made from ground buckthorn berries.  After acquiring these ground berries, my goal was to collect identical spectra from the lake pigment and its respective dye.

After consulting the literature, I performed many trials combining different amounts of hydrochloric acid, methanol, and ultra-pure water to extract the dyes from these lake pigments.  Although stil de grain reacted favorably with a strongly acidic solution, it was too harsh for the reseda lake.  I eventually settled on a solution that was one part hydrochloric acid and three parts methanol.

With a scalpel, I sampled a tiny amount of the pigment and mixed it carefully with the solution of hydrochloric acid and methanol, allowing it to evaporate on a glass slide for several hours.  I then coated the spot with silver nanoparticles, allowing the surface interaction to occur.  While the spot was still wet, I collected SERS spectra.

This procedure was strong enough to extract all three lakes, but gentle enough not to overwhelm the more sensitive reseda lake.

Unfortunately, I have so far been unable to develop a procedure to identify gamboge dye.  Gamboge, a brilliant organic yellow composed of about 20% water soluble gum and 80% water insoluble resin, was frequently used in historic paintings.  This resinous quality makes it very difficult to work with, as it is soluble in a solution of ethanol and water.  Even with these solvents, gamboge has not yielded the expected Raman spectrum.

The next step towards simulating the use of these methods on an actual art sample is to make paints with linseed oil.  I have also been fortunate enough to visit the paintings conservation lab at the DeWitt Wallace Collections and Conservation Building in Colonial Williamsburg.  Next time I will discuss mixing oil paints and taking a scalpel to an 18th century portrait!