This Technology Allows Researchers to View Centuries-Old Paintings in Unprecedented Detail

Madonna in Preghiera" by the workshop of Giovanni Battista Salvi da Sassoferrato. The painting was studied using a terahertz reflectometry technique. | Alexandre Locquet

By re-purposing technology used for oil exploration and geological surveys, scientists and historians are now able to identify the individual layers of paint that go into transforming a blank canvas into a work of art.

A transatlantic team of scientists and art historians has developed a new technique for scanning artwork from hundreds of years ago. Using high-speed scanners and complex signal processing methods, the system provides detailed information on how old masterpieces were prepared, painted, and preserved.

For art historians, the new technique could be a big deal, as it allows researchers to peer through layers of pigment and paint to determine the history of a work of art. In an odd twist, the new technique uses existing technology that was developed for geological surveys and petroleum exploration.

Researchers at the Georgia Institute of Technology, working with academic, museum, and commercial partners in France, adapted the technology and published details on their work in the journal Scientific Reports.

Technically known as time-domain terahertz reflectometry imaging — terahertz imaging for short — the technique works by pulsing electromagnetic waves through a painting that is mounted on a special gantry. As the EM waves pass through successive layers of paint and canvas, portions of the beam are reflected back to the source.

The scanner’s computer analyzes the data in real time, using a signal processing technique known as sparsity-based time-domain deconvolution. (Researchers have yet to coin a shorter term for that one.) The bottom line is that by separating the signals reflected by each layer, the scanner can construct a three-dimensional map of the painting itself.

This image shows reflected raw terahertz signals measured across the painting “Madonna in Preghiera” by the workshop of Giovanni Battista Salvi da Sassoferrato. |David Citrin

Because each layer has its own terahertz “fingerprint,” historians can determine the materials the artist originally used and how they were incorporated into the artwork.

David Citrin, co-author of the paper and a professor in the Georgia Tech School of Electrical and Computer Engineering, said the level of detail generated by the system is unprecedented, thanks to the incredible scale and speed of the energy pulses themselves.

“We are using very short pulses — about one picosecond long,” Citrin told Seeker. “That means if we stacked pulses back to back, there would be one trillion of them in one second.”

By working at this scale, the signal processing end of the system can accurately distinguish between layers of paint just 20 microns thick. For comparison, consider that an average human hair is around 100 microns thick.

“It’s like sending a ping into something and then listening to the echoes as the ping bounces off the boundaries between different layers,” Citrin said. “By accounting for the time between echoes, and knowing a bit about the materials, we can determine the thickness of the material between the various echoes.”

This image shows reflected raw terahertz signals measured across the painting “Madonna in Preghiera” by the workshop of Giovanni Battista Salvi da Sassoferrato. David Citrin

To demonstrate the new technique, researchers studied the painting “Madonna in Preghiera” from the workshop of 17th century Italian baroque painter Giovanni Battista Salvi da Sassoferrato. That’s where the museum collaborators come in. The painting was borrowed from the Musée de la Cour d’Or, Metz Métropole in France.

Using the new terahertz scanning method, researchers generated a three-dimensional model of the painting, separating out the various layers, such as the canvas surface, the pictorial layer itself, and the imprimatura — the initial stain of neutral color used by painters to improve color quality. Analysis of the imaging data also revealed a previously unknown restoration of the varnish layer.

“An interesting thing about the project is that it involves people at Georgia Tech’s campus in Atlanta and at a French government laboratory located at Georgia Tech Lorraine — our European campus in Metz, France,” Citrin said. “In addition, we had important input from our collaborator Marcello Melis of Profilocolore, a multispectral imaging company in Rome, Italy.”

“The equipment and the idea of terahertz imaging was developed for various general purposes,” Citrin said. “The fact that such commercial systems are now available means we could focus on the measurement and not on tweaking the system.”

The result of this unique collaboration among public and private groups, in several different countries, is that art historians worldwide now have a powerful new research tool. The terahertz imaging technique can be used on its own, or to supplement conventional art analysis techniques such as X-rays, nuclear magnetic resonance imaging, and optical imaging, Citrin said.

“I think it’s important for the general public to see how science and engineering can be used to provide important information in the humanities.”