Experimental problems in Raman spectroscopy applied to pigment identification in mixtures

Raman spectroscopy provides useful information to detect and identify pictorial materials in artworks, although some problems are involved when the identification of individual pigments in mixtures is treated. With the hypothesis of the Principle of superposition, the mixture spectrum should be the direct sum of each pondered individual spectrum. In this work, we will show several mixtures where it can be noticed that the mixture spectrum is not qualitatively proportional to the sum of pondered individual spectra. Also there were some cases where the bands of one of the pigments could not be detected in the mixture spectrum. This non-linear behaviour could be attributed to specific proprieties of each material that are revealed when they interact with each other. We conjecture that, for instance, the different reflectances or the wavelength of the laser source could be determinant factors of the obtained results. In this paper an experimental method has been designed in order to characterize the quantitative behaviour of the Raman bands corresponding to each pigment in a mixture. Adequate coefficients are defined and calculated to facilitate the study of the spectral contribution of the different components of a mixture.