Study of automated mass spectral deconvolution and identification system (AMDIS) in pesticide residue analysis.

The effects of overlapping levels and concentration ratios of overlapping components, and of scan rates of the mass spectrometer, on the capability of the automated mass spectral deconvolution and identification system (AMDIS) in pesticide residue analysis were studied. To investigate the capability of AMDIS in removing interferences from the overlapping peaks, this system was applied to data files obtained from the gas chromatography/mass spectrometry (GC/MS) analysis of two overlapping (co-eluting) pesticides (beta-HCH and PCNB) in full scan mode. Differences in overlap levels, the concentration ratios of the two overlapping components and the scan rates of the instrument were studied. When the difference in scan number of overlapping compounds was equal to 1 scan, AMDIS incompletely extracted 'purified' mass spectra but as the difference increased to 3 or more scans, complete correct spectra could be extracted. The results also show that when the scan rate was in the range of 0.4-0.90 s/scan and the concentration ratios of the target compound/interference were above 1/5, there were ideal deconvolution results for this approach. To further study the application of AMDIS to pesticide residue analysis, AMDIS was applied to the identification of pesticides spiked in real samples (cabbage and rice). Typical pesticides being evaluated were identified using AMDIS at concentrations >50 ng/g in the extracts.