The non-destructive analysis of works of art and more specifically the paintings with the aim of a non-ambiguous
identification of their components and the understanding of the techniques of the artists still remains a challenge. The
aim of our research is to elaborate a purely optical way for this identification, based on the exclusive use of the intrinsic
characteristic optical parameters of the components, instead the derived parameters presently commonly used, depending
on several other parameters (morphology, environment...).
The approach we propose is based on the resolution of the RTE using the 4-Flux approximation, combined with the Mie
theory, allowing the identification of the pigments via the spectrum of their complex optical index entered into the model
via a database. The key point of this approach is the index data bank. We report in this communication one the method's
crucial steps: the determination of the intrinsic optical index of pigments under the form of grains of micrometric size.
This step is far from trivial and presents many difficulties that are not completely solved. This is one of the reasons why
a more rigorous analysis of the paintings has not been up to now developed.
We illustrate this problem with a red pigment: vermillion randomly dispersed at low concentration in a transparent
polymer. The morphology of the sample is well characterized (thickness, concentration, size and dispersion of the
pigments, surface roughness) as well as the index of the matrix. We use the same approach and model as presented
above, applied this time to the calculation of the complex index of the pigments. The model is supposed to account for
the diffuse flux and the specular flux, both measured on our samples, by spectrophotometry with an integrating sphere in
the visible spectral range 400-800 nm. This resolution allows determining independently the coefficients of scattering
and absorption of the pigment, which are finally related to the complex index of refraction through Mie's Theory.