ATR versus Diffuse FTIR spectroscopy for Varnished paintings

This is an FTIR spectrometer, which we use to analyze pigments, binders, and varnishes.
One of the main problems when working with real paintings is that they are usually varnished.
This raises a key question: can FTIR identify pigments when a varnish layer is present?

To investigate this, I prepared a color chart with different pigments. For example, here we have artificial ultramarine: the upper part is coated with dammar varnish, while the lower part is unvarnished.

We will test two FTIR sampling modes:

• Diffuse Reflectance

• ATR (Attenuated Total Reflectance) with a diamond crystal

The main difference between these two probes is their depth of penetration.
ATR probes only a few microns (typically 2–5 µm), meaning it is sensitive almost exclusively to the surface layer. Diffuse reflectance, on the other hand, penetrates hundreds of microns, sampling the varnish, paint layer, ground, and even the support.

ATR Measurements

Let’s start with Prussian blue and use the ATR probe.

First, we measure unvarnished paint. The spectrum clearly shows the characteristic absorption band of Prussian blue around 2075–2100 cm⁻¹.

When we repeat the measurement on Prussian blue covered with dammar varnish, this diagnostic band completely disappears. Instead, the spectrum is dominated by varnish features.

The same behavior is observed with lead white.
The strong diagnostic bands of lead white are clearly visible in the unvarnished sample, but once a dammar layer is applied, these features vanish entirely. The resulting spectrum closely matches that of dammar alone.

This confirms that ATR only probes the varnish layer when a coating is present, making it unsuitable for identifying underlying pigments in varnished paintings.

Diffuse Reflectance Measurements

Now let’s examine the same samples using diffuse reflectance FTIR, which probes much deeper into the stratigraphy.

With lead white, even diffuse reflectance struggles. The varnish and the support strongly contribute to the spectrum, and the diagnostic bands of lead white are either suppressed or masked. In this case, pigment identification remains unreliable.

However, the situation changes with Prussian blue.
Prussian blue has an extremely strong absorption band around 2100 cm⁻¹. Even when covered by dammar varnish, this band remains visible in diffuse reflectance spectra, although its shape is altered due to complex scattering and absorption effects.

Conclusions

• ATR FTIR is highly surface-sensitive and effectively detects only the varnish when a coating is present.

• Diffuse reflectance FTIR samples multiple layers and produces complex spectra influenced by varnish, paint, ground, and support.

• For most pigments, varnish masks diagnostic features even in diffuse reflectance.

• Only very strong absorbers, such as Prussian blue, can still be identified beneath varnish layers.

In summary, FTIR has severe limitations for pigment identification in varnished paintings, and only a small subset of pigments can be reliably detected under these conditions.