Benchmarking of candidate detectors for multiresidue analysis of pesticides by comprehensive two-dimensional gas chromatography

Abstract

The present study discusses the relevance, performance and complementarities of flame photometric detector in phosphorus (FPD/P) and sulfur (FPD/S) modes, micro electron capture detector (μECD), nitrogen phosphorus detector (NPD), flame ionization detector (FID) and time-of-flight mass spectrometer (TOF/MS) for the comprehensive two-dimensional gas chromatography (GC×GC) analysis of pesticides. A mix of 41 pesticides including organophosphorus pesticides, synthetic pyrethroids and fungicides was investigated in order to benchmark GC×GC systems in terms of linearity (R2), limits of detection (LOD), and peak shape measures (widths and asymmetries). A mixture of pesticides which contained the heteroatoms phosphorus, sulfur, nitrogen and one or several halogens, was used to acquire a comparative data set to monitor relative detector performances. GC×GC datasets were systematically compared to their GC counterpart acquired with an optimized one-dimensional GC configuration. Compared with FID, considered the most appropriate detector in terms of suitability for GC×GC, the element-selective detector FPD/P and μECD best met the peak widths (0.13–0.27s for FPD/P; 0.22–0.26s for μECD) and tailing factors (0.99–1.66 for FPD/P; 1.32–1.52 for μECD); NPD exhibited similar peak widths (0.23–0.30s), but exceeded those of the above detectors for tailing factors (1.97–2.13). These three detectors had improved detection limits of 3–7 times and 4–20 times lower LODs in GC×GC mode compared with FID and TOF–MS, respectively. In contrast FPD/S had poor peak shape (tailing factor 3.36–5.12) and much lower sensitivity (10–20 fold lower compared to FPD/P). In general, element-selective detectors with favorable detection metrics can be considered viable alternatives for pesticide determination using GC×GC in complex matrices. The controversial issue of sensitivity enhancement in GC×GC was considered for optimized GC and GC×GC operation. For all detectors, we found no significant LOD enhancement in GC×GC.