Read our COVID-19 research and news.

The Invisible Graffiti of Angkor Wat
Bjørn Christian Tørrissen; (inset) Antiquity Publications Ltd.

3D Imaging Reveals How Paintings Were Made

There’s more to a painting than meets the eye. Under the surface of a sun-dappled landscape or a scrumptious still life lie dozens of meticulously applied layers of paint, forming a complex 3D structure that is all but invisible to viewers. Now, an imaging technique borrowed from biomedical research promises to let art historians and conservators peer into the depths of paintings without damaging them, providing new insights into how these works were made.

“Right now, if an art conservator wants to understand the three-dimensional layering structure of a painting, they almost certainly take a scalpel to it,” removing tiny core samples to study its stratigraphy, says Warren Warren, a chemist and biomedical engineer at Duke University in Durham, North Carolina. He spends most of his time developing laser systems used to image human tissue. But when he visited an exhibit on detecting art forgeries in London’s National Gallery a few years ago, he began wondering what art historians and conservators could learn about artwork if they had access to the state-of-the-art imaging technologies like the ones in his lab.

One method Warren works on is called pump-probe microscopy, which uses carefully timed pulses of laser light to electrically excite the molecules in a sample. As the molecules gain and lose energy in reaction to the pulses, they emit signals that serve as identifying “fingerprints” that reveal their chemical makeup. Pump-probe microscopy is especially useful for studying biological pigments like melanin in skin. So Warren wondered: Could it work on other kinds of pigments, too? Like, say, paint?

“We built a laser system that was designed to do a good job of diagnosing skin cancer and then realized that we could use exactly that same laser system to look at Renaissance artwork,” he says. The low-powered laser pulses travel deep into a painting without scattering as conventional light sources do, returning a remarkably clear picture of its subsurface structure as well as chemical fingerprints of the pigments in each layer.

The team initially tested the technique on mock-up paintings made with historically accurate Renaissance pigments, proving that pump-probe microscopy can distinguish between the 3D structures of a purple created by mixing red and blue pigments and a similar shade made by layering red over blue. Then, the researchers turned their laser eye on an actual Renaissance painting: The Crucifixion, painted by Puccio Capanna around 1330. By imaging small sections of the blue robes of the Virgin Mary and one of the flying angels, they revealed that Capanna used very different pigments to create each one, despite their similar colors. Mary’s robe is composed of a thick layer of ground-up lapis lazuli, a deep blue stone that at the time was “more expensive than gold,” Warren says. The blue of the angel’s robe, on the other hand, was created through a complex layering of several less precious pigments, with just a hint of lapis lazuli, the team reports online this week in the Proceedings of the National Academy of Sciences.

“Honestly, for me it was like a glimpse into the future,” says Francesca Casadio, a conservation scientist at the Art Institute of Chicago in Illinois who was not involved in the research. Pump-probe microscopy could be especially useful for identifying places on aging paintings where the pigments have started to decay, she says. That could help conservators fine-tune their efforts to halt such deterioration. “Such a boost in technology is what the art conservation and museum fields need to ensure that unique works of art are and remain protected in the best possible manner,” agrees Koen Janssens, an analytical chemist at the University of Antwerp in Belgium who was not involved in the research.

Warren hopes pump-probe microscopy might also aid in the identification of forgeries. If the 3D structure of brushstrokes varies from artist to artist, for example, it could serve as a kind of signature, helping historians distinguish between the work of a master and an imitator.

Casadio is skeptical, however, that such identification will ever be precise enough to supplant the sophisticated techniques historians and appraisers already use. She emphasizes that it will be quite some time before pump-probe microscopy becomes practical for most museums. Not only does it now take hours to analyze a few square millimeters of a painting, but the work also needs to be done in a lab with the help of trained scientists. Museums need a smaller system they can use themselves, she says. Not to worry, Warren says: Biomedical researchers are already shrinking down pump-probe microscopy systems, and it’s only a matter of time before these new eyes start looking at art.