Towards a semiquantitative non invasive characterisation of Tyrian purple dye composition: Convergence of UV-Visible reflectance spectroscopy and fast-high temperature-high performance liquid chromatography with photodiode array detection.

In this paper, partial least square (PLS) regression is innovatively applied for a semi-quantitative non invasive study of the most precious dye of Antiquity: Tyrian purple. This original approach for the study of organic dyes in the cultural heritage field, is based on the correlation of spectrophotometric (UV-Visible) and chromatographic (Fast-HT-HPLC-PDA) data from an extensive set of textiles prepared with different snail species according to historical recipes. A cross-validated PLS model, based on the quantity of 6,6'-dibromoindigotin, displays an excellent correlation factor (R(2)Y = 0.987) between values determined by chromatography and those predicted from reflectance spectra. This indicates that the spectral features of Tyrian purple on textile fibre is strictly related to the amount of this indigoid component whose content may be non invasively predicted from reflectance spectrum. The studied correlation also highlights that, independently of the dyeing method and nature of the textile fibre used, the relative content of 6,6'-dibromindigotin may be used as a parameter to distinguish samples prepared with Hexaplex trunculus L. snails from those prepared with further mollusc species. To validate this model, archaeological textile fragments dating from the Roman period were successfully examined. The results achieved open an absolutely new way in Tyrian purple analysis in cultural heritage by non invasive spectroscopic techniques attesting their convergence with HPLC and giving them a semi-quantitative value.