Determination of major and minor elements in Mexican red wines by microwave-induced plasma optical emission spectrometry, evaluating different calibration methods and exploring potential of the obtained data in the assessment of wine provenance

Abstract

In this work, the determination of major and trace elements in Mexican wines was carried out by microwave-induced plasma optical emission spectrometry (MIP OES), setting sample dilution as the pretreatment procedure. The elements were grouped according with the required dilution fold: 1:100 for K; 1:50 for Na, Mg, Ca; 1:10 for Rb, Sr, Mn, Fe, Al and 1:4 for Cu, Cr. Due to troublesome background noise, different for each sample, matrix-matching with individual wines was necessary and three quantification methods were explored: classic standard additions (SA), standard dilution analysis (SDA) and multi-energy calibration (MEC). In order to avoid extrapolation in assessing the actual concentration of internal standard at each dilution point and to allow for longer integration time during signal acquisition, SDA was modified (nominated MSDA) and was recommended for the determination of minor elements whereas for major elements the suitable method was MEC. The results obtained by MSDA and MEC were in good agreement with those provided by SA; however, background correction was not required and both methods offer more simplicity as compared to SA. The concentration ranges found in the Mexican wines were as follows: 866–4896 mg L−1 K, 32.7–326 mg L−1 Na, 56.6–164 mg L−1 Mg, 63.9–121 mg L−1 Ca, 0.56–1.90 mg L−1 Sr, 0.35–6.31 mg L−1 Rb, 0.21–2.11 mg L−1 Fe, 0.69–1.72 mg L−1 Mn, 0.28–0.78 mg L−1 Al, 0.10–0.42 mg L−1 Cu, and Cr was not detected. Principal component analysis (PCA) allowed for separation of wines according with the winery regions: Valle de Guadalupe, Guanajuato, Aguascalientes. Furthermore, PCA revealed association of Rb, K, Ca, Sr with wines produced in Guanajuato and association of Na, Cu and Mn with those produced in Valle de Guadalupe; wines from Aguascalientes contained relatively high concentration of Fe and Al, although tight clustering of variables was not observed in this case. To establish reliable discrimination model, more samples from these and possibly from other winery regions in Mexico need to be analyzed.