Color prediction from first principle quantum chemistry computations: a case of alizarin dissolved in methanol

The electronic spectrum of alizarin (AZ) in methanol solution was measured and used as reference data for color prediction. The visible part of the spectrum was modelled by different DFT functionals within the TD-DFT framework. The results of a broad range of functionals applied for theoretical spectrum prediction were compared against experimental data by a direct color comparison. The tristimulus model of color expressed in terms of CIE XYZ and CIE Lab parameters was applied both to experimental and predicted spectra. It was found that the HSE03 method along with the 6-31G(d,p) basis set provides the most accurate color prediction, much better than other commonly used functionals such as B3LYP, CAM-B3LYP or PBE0. Besides, the influence of potential errors, introduced by theoretical predictions on color estimation, was examined for different wavelengths. The obtained results showed that color prediction is significantly dependent on the type of basis set and functional applied. The proposed methodology provides a simple, straightforward and more reliable way of theoretical protocols validation than just a comparison of experimental and estimated values of maximum absorbance wavelength.