Scientific Art Authentication: What Tests Can and Can't Prove

The logic of anachronism

Scientific analysis in art authentication rests on a simple asymmetry. A test can show that a work contains something that did not exist when the work was supposedly made, and that finding can be decisive. A test cannot show that a particular person made the work. At best, it can show that the materials and methods are consistent with the claimed artist and date.

This is the logic of anachronism. A pigment introduced after an artist's death, a panel cut from a tree that was still growing when the painting was supposedly finished, or a modern synthetic binder in a work attributed to an earlier century each rules out the attribution, or at least demands an explanation such as later restoration. The absence of any anachronism, by contrast, is not proof. It only means the tests found nothing inconsistent.

Understanding that asymmetry is the most useful thing a collector can take into a conversation with a laboratory. It shapes what to ask for, and how to read the report that comes back.

Pigment and materials analysis

Materials analysis identifies what a work is made of: pigments, fillers, binding media, grounds, varnishes and supports. Some techniques, such as X-ray fluorescence, can be used without taking a sample and report the chemical elements present at a point on the surface. Others, such as Raman or infrared spectroscopy and the study of paint cross-sections under a microscope, can identify specific compounds and layer structures, but some of them require a tiny sample taken by a conservator.

The value of these results depends on how well the materials found can be dated. Some pigments have well-documented introduction dates. Titanium white, for example, only came into widespread use as an artists' pigment in the twentieth century, and many synthetic organic pigments are later still. Finding such a material in the original paint of a work attributed to an earlier period is strong evidence against the attribution.

The caveat matters. Many works have been restored, and modern materials in a retouched area say nothing about the original paint. A careful analyst establishes which layers are original before drawing conclusions.

Dating the support: dendrochronology and radiocarbon

Dendrochronology dates wooden panels by matching the pattern of annual growth rings to reference chronologies for the region the timber came from. It dates the latest ring present, from which an earliest likely felling date can be estimated. It does not date the painting: timber was seasoned and stored before use, panels were sometimes reused, and the method only works for certain species and regions with good reference data.

Radiocarbon dating measures the decay of carbon in organic materials such as canvas, wood, paper and some binding media. For older materials, calibration produces a date range rather than a single year. For more recent works, atmospheric nuclear testing in the mid-twentieth century left a distinctive signal that can help show whether organic material was formed before or after that period.

Both methods share a limitation. They date the material, not the act of painting. An old canvas or panel can be acquired and painted on much later, and forgers have long done exactly that. An old support is necessary for an old attribution, but it is not sufficient.

Imaging: X-radiography, infrared reflectography and UV

Imaging techniques look beneath or across the surface without taking samples. X-radiography records how strongly different materials absorb X-rays. Lead-based white paint shows clearly, so X-rays can reveal changes to the composition, an earlier painting beneath the visible one, the weave of the canvas and the construction of a panel or stretcher.

Infrared reflectography penetrates many paint layers and shows carbon-based underdrawing, where it exists. Comparing the character of an underdrawing, and the changes made between drawing and final paint, with those in securely attributed works can be informative about working method. Ultraviolet examination shows the fluorescence of varnishes and the presence of later retouching on the surface.

These images rarely settle a question on their own. What they provide is evidence about how a work was made, which specialists can compare with what is known of the artist's practice. That comparison depends on good reference material from works whose attribution is secure.

What scientific tests can't prove

The limits are as important as the capabilities, and a report that glosses over them should be read with caution. Laboratories themselves are usually careful to describe results as consistent or inconsistent with a date or attribution, and that language should be taken at face value.

Commissioning analysis sensibly

The most useful analysis starts with a specific question. Whether a painting is genuine is not something a laboratory can answer. Whether the white paint in the original layers is consistent with the claimed date, or whether the panel's timber could have been available when the work was supposedly made, are questions it can address.

Non-invasive methods are usually tried first, and sampling is considered only where it would answer something that imaging and surface analysis cannot. Results should be interpreted by specialists who understand both the science and the art history, and read together with the provenance file and the view of the recognised authority for the artist. A laboratory report filed alongside that evidence, rather than presented as a verdict on its own, will carry more weight with anyone who later relies on it.