Botanical origin characterisation of tannins using infrared spectroscopy

Different approaches to analysing the botanical origin of tannins have been proposed in the last fifteen years, but are generally time consuming and require the use of advanced instrumentation. This study aims to suggest an effective, easy, rapid and cheap method based on the acquisition of FT-IR spectra of 3g/L hydroalcoholic tannin solutions, overcoming possible disadvantages due to sample or particle size inhomogeneity. 114 commercial powder tannins from 7 different botanical sources (oak, chestnut, gall, quebracho, tea, grape skin and grape seed) were collected and the FT-IR spectra were acquired in the region 926-5011cm^-1. Partial Least Squares regression, Discriminant Analysis and Artificial Neural Networks were applied to FT-IR spectra to investigate the possibility of differentiating the 7 botanical origins. The best results were obtained using Discriminant Analysis, with 95% correct re-classification, and 97% grouping of grape skin and seed in a single source.