Accounting for matrix effects of pesticide residue liquid chromatography/electrospray ionisation mass spectrometric determination by treatment of background mass spectra with chemometric tools.

Matrix effect (ME) - ionisation suppression or enhancement - in liquid chromatography/electrospray ionisation mass spectrometry (LC/ESI-MS) is caused by matrix components co-eluting with the analytes. ME has a complex and not fully understood nature. ME is also highly variable from sample to sample making it difficult to compensate for. In this work it was studied whether the background ion signals in scanned mass spectra of the LC effluent at the retention time of the analyte offer some insight into the presence and extent of matrix effect. Matrix effects for six pesticides - thiabendazole, carbendazime, methomyl, aldicarb, imazalil and methiocarb - in garlic and onion samples used in the study varied from 1% (suppression 99%) to 127% (enhancement 27%) depending on the pesticide and sample. Also standards in solvent and solvent blanks were included in the study. The ions most strongly varying from sample to sample - and therefore best describing the changes in sample composition and ME - were selected for quantification according to principal component analysis (PCA) for all six pesticides under study. These ions were used to account for ME via partial least-squares (PLS) regression. The calibration set was constructed from 19 samples and standards and the obtained calibration function was validated with seven samples and standards. The average errors from the test set were from 0.05 to 0.27 mg/kg for carbendazim and imazalil, respectively (the respective average pesticide concentrations were 0.22 and 0.88 mg/kg). The PLS results were significantly more accurate compared to the conventional solvent calibration resulting in average errors from 0.07 to 0.69 mg/kg for carbendazime and methiocarb, respectively.