Green and Multivariate Approach to Obtain a Fingerprint of Red Wines by HPLC-PAD

Abstract

This work presents the development of a strategy that integrates multivariate statistical analysis and concepts of green chemistry to obtain the chromatographic profile of Brazilian red wines by HPLC-PAD. An initial screening of variables was performed using fractional factorial design (2_IV^7–2) to investigate the main parameters responsible for separating phenolic compounds. It was considered the initial and final % of bioethanol in the organic mobile phase, % acetic acid in aqueous mobile phase, flow rate, running time, type of column chemistry, and column temperature. The number of chromatographic bands was used as empiric response. The most important variables were selected to further optimization by Doehlert factorial design. The optimal condition was as follows: initial % of bioethanol=5%, final % of bioethanol=55%, column temperature= 51.8 °C, % acetic acid in H₂O =0.1%, flow rate = 0.75 mL·min⁻¹, time=30 min, and column chemical composition=Phenyl-Hexyl (X-Select). The optimized method allowed the separation of 24 chromatographic bands with signal/noise (S/N) higher than 100, a response at least 17% higher than observed in the screening step. The method allowed the separation and identification of the main compounds presented in red wines: gallic acid, caffeic acid, syringic acid, p-coumaric acid, resveratrol, kaempferol, catechin, and epicatechin.