Study of the precision in the purge-and-trap–gas chromatography–mass spectrometry analysis of volatile compounds in honey

Data precision in the analysis by purge-and-trap coupled on-line to gas chromatography–mass spectrometry (P&T-GC–MS) of honey volatiles has been studied by statistical analysis. The contribution of non-random factors to dispersion of quantitative results was proven by comparing several statistical parameters (correlation coefficients, principal component analysis (PCA) eigenvalues and loadings) from both experimental and simulated data. PCA was also useful for grouping volatiles with similar dispersion behaviour; these groups being generally related to compounds with common properties or structural features. The use of area ratios improves data precision for compounds within the same group. Results from this study could be used for a better selection of internal standards in quantitative analysis of volatiles by P&T-GC–MS.