The panelists swirled and smelled each sample for about 15 s, then began to rate the intensity of each attribute. The following parameters were analysed: overall perception of quality (1 = faulty, 3 = acceptable, Screening Library 6 = outstanding), body (1 = light, 3 = medium, 5 = full), alcohol level (1 = low, 3 = medium, 5 = high), flavour length (1 = short, 3 = medium, 5 = long), flavour intensity (1 = low, 3 = medium, 5 = pronounced), and approximate wine age (no scale). The intensity of each sensory parameter was measured with

the structured scales applied to Levels 3 and 4 of the Wine and Spirits Education Trust (2009), a worldwide wine school that prepares professionals to taste all types of wines and spirits. These scales are use by professionals worldwide to evaluate wine quality. Thus, the sensory results obtained by the WSET method seems to be the closest approach to the consumer’s perception. All panelists were fully trained and had more than five years of experience in evaluating all types of wine using these scales, which means that the sensory evaluation of red wines with the WSET scales was routine for all the assessors. For high-performance liquid chromatography coupled with a diode array and fluorescence detection (HPLC–DAD–FL), Agilent Technologies 1200 series equipment containing

a quaternary pump, a 20 μL injection loop, and a UV detector was used. The HPLC was controlled buy BIBW2992 by a PC running HP Chem Station Software system. Stock solutions of all standards were prepared in methanol/water, and the calibration

curves were obtained from triplicate injections of at least five concentrations. For all standard curves, correlation coefficients (r) were above 0.990. For the HPLC analysis, polyphenols were identified by comparing their retention times with those of pure standards. Flavanols (catechin, epicatechin), hydroxybenzoic acids (gallic acid and vanillic acid), and hydroxycinnamic acids (caffeic acid, p-coumaric acid, and ferulic acid) were measured in triplicate with a Luna Phenomenex C18 column and a guard column kept at 30 °C with 200 × 4.6 mm i.d. 5 μm particle size. The mobile phases consisted of acetonitrile, acetic acid, and water, where the gradient elution conditions were as follows: 0 min (5% acetic Adenosine triphosphate acid: 15% methanol; 80% water), 5 min (5% acetic acid: 20% methanol; 75% water), and 40 min (5% acetic acid: 45% methanol; 50% water). The flow rate was 0.2 mL min−1, and the injection volume was 20 μL ( López et al., 2001). Before analysis, wines were filtered with a 0.45 μm filter with a PTFE membrane (Millipore, São Paulo, Brazil). The programmable variable wavelength UV–Vis detector system allows detecting at different wavelengths; so λ = 271 nm for gallic acid, λ = 279 nm for (+)-catechin and (−)-epicathechin, λ = 296 nm for vanillic and ferulic acids, λ = 308 nm for p-coumaric acid, and λ = 325 nm for caffeic acid.