Comparison of gas chromatography–pulsed flame photometric detection–mass spectrometry, automated mass spectral deconvolution and identification system and gas chromatography–tandem mass spectrometry as tools for trace level detection and identification

The complexity of a matrix is in many cases the major limiting factor in the detection and identification of trace level analytes. In this work, the ability to detect and identify trace level of pesticides in complex matrices was studied and compared in three, relatively new methods: (a) GC–PFPD–MS where simultaneous PFPD (pulsed flame photometric detection) and MS analysis is performed. The PFPD indicates the exact chromatographic time of suspected peaks for their MS identification and provides elemental information; (b) automatic GC–MS data analysis using the AMDIS (“Automated Mass Spectral Deconvolution and Identification System”) software by the National Institute of Standards and Technology; (c) GC–MS–MS analysis. A pesticide mixture (MX-5), containing diazinon, methyl parathion, ethyl parathion, methyl trithion and ethion was spiked, in descending levels from 1 ppm to 10 ppb, into soil and sage (spice) extracts and the detection level and identification quality were evaluated in each experiment. PFPD–MS and AMDIS exhibited similar performance, both superior to standard GC–MS, revealing and identifying compounds that did not exhibit an observable GC peak (either buried under the chromatographic background baseline or co-eluting with other interfering GC peaks). GC–MS–MS featured improved detection limits (lower by a factor of 6–8) compared to AMDIS and PFPD–MS. The GC–PFPD–MS–MS combination was found useful in several cases, where no reconstructed ion chromatogram MS–MS peaks existed, but an MS–MS spectrum could still be extracted at the elution time indicated by PFPD. The level of identification and confirmation with MS–MS was inferior to that of the other two techniques. In comparison with the soil matrix, detection limits obtained with the loaded sage matrix were poorer by similar factors for all the techniques studied (factors of 5.8, >6.5 and 4.0 for AMDIS, PFPD–MS and MS–MS, respectively). Based on the above results, the paper discusses the trade-offs between detectivity and identification level with the compared three techniques as well as other more traditional techniques and approaches.