From electron microscopes to X-rays, high-tech tools expose low-tech art forgery
Icilio Federico Joni (1866-1946) was a masterful Italian artist of medieval paintings. Trained in the techniques of Middle Age artisans, Joni used his skills to legitimately restore 14th-century works. Then, he started forging them.
“As to beauty — they are lovely! But there are too many of them,” art historian Mary Berenson said of the fake medieval works of art cropping up in the early 1900s. (Nevertheless, her husband, Bernard, seemed to continue passing them off as genuine even after figuring out what Joni was up to.)
Madonna and Child, a painting believed to be by 14th-century painter Duccio di Buoninsegna but later suspected of being a forgery. Click/drag/scroll to zoom. Indianapolis Museum of Art
In 1951, donors gifted the Indianapolis Museum of Art with Madonna and Child, a 14th-century painting they’d purchased a few years earlier from an Italian art firm.
The gilded painting, depicting Mary in a blue robe holding a steady-gazed baby Jesus, hung in the museum until the 1990s. That’s when a curator, who’d recently attended a lecture about a prominent forger, noticed literal cracks in the painting’s facade.
After the painting had spent a few decades in storage, Greg Smith, the museum’s senior conservation scientist, decided to take a closer look at the suspected forgery.
“In our laboratory we have about one-and-a-half million dollars worth of scientific equipment and imaging systems, and so we threw all of that artwork in kind of a crescendoed approach,” he told Digital Trends. Glennis Rayermann, then a Ph.D student at the University of Washington, assisted with the project, attempting to figure out if the painting was indeed a fake and if Joni was responsible.
Mind the beeswax
In 2006, Kim Muir and Narayan Khandekar published a paper in the Journal of the American Institute for Conservation about the techniques they’d used to analyze a Joni painting. While the 20th-century artist was a restorer, anything he touched was suspect. As the researchers noted, the curator Gianni Mazzoni once said, “There was a time, around 1930, when the notoriety of Joni had grown to such a point that it conferred an air of uncertainty onto every gold ground painting that came from Sienese and Florentine antique shops.”
Dr. Gregory Smith, senior conservation scientist at the Indianapolis Museum of Art, uses a microscope to search for signs of forgery on a 17th century painting. Indianapolis Museum of Art
Using Muir and Khandekar’s paper as a template, Rayermann outlined a plan to analyze the Indianapolis museum painting from the ground up. “You do as much as you can at the beginning non-invasively, non-destructively,” she explained. That meant photography, X-rays, infrared, ultraviolet lights, and microscopes.
With the X-rays, she was “looking for the bones of the painting,” she said. Painted on wood panels, the Madonna and Child is made of two planks held together by four wooden joints. Other elements revealed by the X-ray, like the grain of the wood, the shape of the nails, and the placement of the fasteners, all appeared consistent with medieval manufacture. It’s not easy to assign those results to the “it’s not a fake” column, though. Joni learned these methods and would often use very old wood to make his forgeries look authentic.
“I have a couple inklings that what we’re seeing …. isn’t from the Middle Ages.”
One non-invasive method of aging wood is dendrochronology, like counting a tree’s rings. But the Joni painting is coated in beeswax, obscuring the telltale lines. It’s a centuries-old practice artists used to keep the wood from warping. If only one side of the wood is painted, the back tends to swell over time, Smith explained.
“This could have been a trick that Joni did in order to try and prevent you from seeing the back of the wood, something that might give away the fact that this isn’t ancient,” Smith said, “Or he could have been doing it intentionally to try and mimic the technique.”
The X-ray revealed details beyond just the painting’s construction. Rayermann describes the painting like a layer cake, and from the bottom up it’s stacked like this: the wood panels, the ground layer, the preparatory layer, the paint layers, and the surface. The x-ray didn’t skip over all of them and go straight to the wood. It showed the ghostly outlines of mother and child on the preparatory layer. The faces have no distinguishable features.
How the Indianapolis Museum of Art documents an painting layer by layer.
Typically in the Middle Ages, and even into Rembrandt’s time, artists would use lead white on this underlayer to paint in noses, brows, upper lips, and other areas to make highlights and give the final image more dimension. They show up brightly in x-rays but are missing from Madonna and Child.
The pigments artists used in the 13th and 14th centuries were limited, and peachy flesh tones often have a greenish cast. The painters used a green earth pigment for shading and adding depth to faces. With special lamps, Rayermann captured infrared images from the painting and was able to determine what pigments were present.
Cameras and x-rays could only take the analysis so far, and some techniques required an actual sample of the painting.
The Madonna and child’s faces have a verdigris cast under normal light, but those areas look blush pink in false color infrared images. Photoshopped versions of the infrared images then show how a pigment absorbs and reflects the light. Green earth would appear a dark brown, almost black. Nothing like that shows in the Indianapolis museum’s painting. Chrome oxide green, cadmium green, viridian, and cobalt titanate green look pink or red under in false colored images.
“All of those are the modern, synthetic pigments so they didn’t exist prior to around 1700,” said Rayermann. “From even just the initial data, I have a couple inklings that what we’re seeing in the painting isn’t from the Middle Ages.”
After visiting Joni’s workshop in 1899, Mary Berenson described him in her diary as “a rakish-looking man of 30, very free and easy — a good fellow.” According to her, the forgeries were a communal affair, created by “a rollicking band of young men,” each with their own task: drawing, caking dirt on the painting, and crafting the frame. A group of children and their large dog watched over the paintings as they dried in the sun. They viewed themselves as tradesmen, not artists, Berenson explained, with no reason to hide what they were doing.
Cameras and x-rays could only take the analysis so far, and some techniques required an actual sample of the painting. To keep the damage to a minimum, Rayermann took a tiny slice — on the scale of millimeters — out of an area next to a crack. There are a few tools scientists can use to make miniature holes, including eye surgeons’ scalpels or a chemically etched tungsten needle. To lift the sample out of the painting, she used a single strand of deer fur. The sliver ended up being less than the width of a human hair.
A before (left) and after image of the sample used in running tests on the Madonna and Child painting. At less than a milimeter in size, the sample would be smaller than the tip of a pencil. Indianapolis Museum of Art
Despite the minute sample size, Rayermann extracted a lot of information. She uses the analogy of a layer cake. If the modern pigments detected on the infrared image were merely the icing atop other types of paint, then they could’ve been the work of a genuine restorer trying to salvage a damaged 14th century artwork, for example. To get a sense of what each layer’s makeup, Rayermann encased the sample in plastic resin, cured it, and polished a new, flat edge.
As you may remember from art class, tempera paints favored by medieval artists was a mixture of pigments and egg yolk. It dried quickly, created a vibrant hue, and stuck fast to the surface. “My impression is tempera is a pain in the butt to work with,” said Rayermann. For that reason, oil painting began spreading in the 15th century, because it was more versatile and allowed artists to layer colors on top of each other. Using a variety of instruments, Rayermann wanted to see the Madonna and Child’s paint binding material. If it was egg, that would be more consistent with a medieval work.
Dr. Greg Miller and Victor Chen running tests on a painting in the Conservation Science Laboratory of the Indianapolis Museum of Art. Indianapolis Museum of Art
First, using Fourier-transform infrared spectroscopy, she found the binding media’s “molecular fingerprint,” which is based on which wavelengths of infrared light the substance absorbs. There are libraries of known fingerprints, and Rayermann matched the paint sample with aged egg yolk. But the fingerprint on its own isn’t enough to convict. Maybe the egg yolk had an accomplice.
Mass spectrometry bombards a sample with electrons, destroying it but allowing the researcher to see the types and proportion of molecules present. The ratios of the fatty acids in the painting slice didn’t correlate with pure egg yolk. (To avoid further damaging the painting, Rayermann decided not to take another sample to test the proteins as well.)
In a CSI episode, the rate at which one of the molecules vaporized would let a lab tech calculate the age of the paint, but in real life, a variety of factors can mess with this calculation. Pigments can speed up the reaction, as can heating the painting — which is certainly something Joni did to age his forgeries.
To get help identify the the type of compounds used in paint colors, conservation scientists will test samples using a method called Fourier-transform Raman spectroscopy. This machine will then identify the molecular fingerprint of a specific type of material/paint.
To investigate the egg yolk substance further, Rayermann applied a protein-binding stain to the sample to see where it fluoresced. As you’d expect with tempera, which contains protein, the paint layer lit up. When a lipid stain was applied, the same area flouresced again. It reinforced the mass spec results, that the binding media wasn’t just egg but also contained an oil, other than what’s naturally found in eggs.
“All these different little pieces were kind of coming together,” said Rayermann. At the bottom of the sample, the lipid stain also lit up the preparatory layer. It could have been a restorer’s work, but it’s unlikely a painter from the middle ages would have added an oil-based paint to that under portion of the painting.
In 1932, Joni published a memoir in which he detailed some of his forgery techniques. “It’s basically his tips and tricks for how to make an old painting,” said Rayermann. “It’s pretty great.”
A microscopic crack that appears in the Madonna and Child painting. The crack contains clues to the kinds of compounds used during its creation. If the painting contains trace amounts of compounds from dyes that didn’t come about until the 1700s, for example, it would be a clear giveaway that it isn’t a 14th century painting. Indianapolis Museum of Art
One that he didn’t mind divulging was a method of mimicking craquelure, the naturally occurring cracks that form in tempera and oil paintings as the layers respond to changes in temperature and humidity. After etching in the cracks, Joni would roll up the painting to create them.
Eventually, his fake craquelure became too recognizable, and he had to find a new way of inducing them.
In a letter Joni wrote shortly before his death, he explained a bit of his new process of spraying and drying his forgeries to make the cracks. But he had also figured out that he needed to create cracks in the gesso or ground layer, too, otherwise the paint’s craquelure looked artificial.
“It’s not just modern touches added on top of medieval paint.”
On the Madonna and Child painting, there are two sizes of cracks, ones that are larger and can be measured in millimeters that go to the ground layer and smaller ones, on the magnitude of microns, that are surface-level. In his letter, Rayermann notes, Joni uses two different Italian words, depending on which type of crack he’s describing.
Because Joni was fracturing multiple layers on his paintings, the gesso would occasionally peek through the ruptured paint. This was a dead giveaway that he fixed by covering the white gesso with umber.
An electron microscope allows the viewer to see even tiny samples in extreme detail. When the electron beam is set to high enough energy, the sample emits x-rays that can reveal what elements are present. Part of sample where there was a crack contained manganese and iron, consistent with a pigment made of iron oxide and manganese oxide — otherwise known as umber.
Icilio Federico Joni (1866-1946), was a 19th century painter who specialized creating forgeries of Renaissance paintings.
There were other elements in the crack as well. There was phosphorous. When Rayermann compared the result with the reference library, it matched with a pigment called bone black.
“You could imagine that a painting that’s hundreds of years old, if it was displayed by a fireplace then soot might collect in the cracks,” she said, “but the fact that it’s bone black — it’s a manmade pigment, and so the fact that it’s present in the crack means somebody intentionally put that black pigment in there.” Without a little accumulated dust, a forger’s cracks would appear too neat and new.
Using the same electron microscopy on the paint layers also surfaced an interesting element. “In three very diverse regions of the painting — an area of blue, an area of the flesh, and an area of the red robe — I’m seeing titanium top to bottom throughout the paint layer,” said Rayermann
That matches with the false color images, where pigments show up as raspberry hue. The sample’s molecular fingerprints match with anatase titanium white, viridian, cobalt violet light, and synthetic ultramarine pigments, none of which were created until after 1800. The white pigment wasn’t widely available until 1918. “That was the most damning of them,” said Rayermann. “It’s not just modern touches added on top of medieval paint. The whole thing is modern.”
“I would guess there’s not a museum out there that doesn’t have a fake in its collection,” said Smith. The stack of evidence suggesting the Indianapolis Museum of Art’s Madonna and Child is the work of Joni is compelling, but that doesn’t mean it’s doomed to stay in storage forever.
“All of the science and technology which goes into unraveling or uncovering one of these fakes can be fascinating to our public and can really highlight what the authentic artwork is,” he said. “So imagine an exhibition where the Joni painting is hanging next to a Duccio or Jacopo painting that is an original, and you can highlight how one is different from the other.”
Joni once said he could easily teach others how to make a painting look old, but he’d never been able to pass on his gift for painting “with the spirit and the soul” of the artist he was copying. Having been recognized for his talent, the forger has had his works exhibited in Siena, Italy. After one exhibition, art historian Joseph Connors visited a villa and spotted what looked like a Renaissance painting. “Grandfather thought it was Botticelli,” the owner told him. “Father suspected it was a fake. Now we’re hoping it’s Joni.”
- Talking Tech with Elon Musk!
- Trends with Benefits Weekly Tech Podcast
- Trends with Benefits Weekly Tech Podcast
- Trends with Benefits Weekly Tech Podcast
- Trends with Benefits Weekly Tech Podcast