Determination of chemical warfare agents in water samples by solid phase microextraction and GC-FID

Abstract

The Norwegian Defence Research Establishment (FFI) is presently developing a mobile laboratory for the identification of chemical warfare agents (CWA) and related compounds. The laboratory will be used in the field, and it is therefore advantageous that the analytical methods give a high priority to low solvent consumption and minimal need for sample preparation. Solid phase microextraction (SPME) meets these requirements, since the technique is based on automatic extraction and concentration of the analyte from the sample without the use of solvents. The aim of this study was thus to develop a fast screening method for selected CWA in water by the use of SPME coupled to a gas chromatograph (GC) with a flame ionisation detector (FID). Parameters such as fiber selection, extraction time, desorption temperature and desorption time were investigated by using spiked distilled and de-ionised water samples. Thereafter, the developed method for SPME was compared with the existing analytical procedure, which is based on liquid-liquid extraction (LLE). The work clearly shows that a 50/30 μm divinylbenzene/carboxen/polydimethylsiloxane (DVB/CAR/PDMS) fiber is the best choice for the extraction of sarin (GB), soman (GD), sulphur mustard (HD) and cyclohexyl sarin (GF) from distilled and de-ionised water. Optimal conditions for SPME of the selected CWA were obtained when the fiber was immersed for 30 min in water saturated with sodium chloride, and subsequently desorbed for 2 min at 250 ºC in the inlet of the GC. Freshly spiked water samples gave method limit of detection (MLOD) in the range from 0.2 to 2 μg/L, which is 14 - 42 times better than by LLE. Interestingly, the greatest improvement in the MLOD was in the case of GF, which has not been determined by SPME in previous reported work. The presently obtained linear ranges were in the interval from 0.54 ng/mL to 4.9 μg/mL, depending on the analyte, with squared regression coefficients from 0.96 to 0.997. The precisions (%) measured as the relative standard deviations at the concentrations of 0.01 μg/mL (n = 7) and 1 μg/mL (n = 6) CWA were 2 - 8% and 4 - 10%, respectively. The developed method was also applied successfully for determination of CWA in spiked natural water samples. The developed SPME method is less time-consuming than the standard LLE technique, as the total time for the sample preparation and the analysis is approximately one hour compared to 4 - 5 hours for the LLE technique. In addition, the developed method does not involve the use of carcinogenic solvents. Therefore, the developed SPME method is a significant contribution towards a faster and a more user-friendly determination of CWA, which is suitable for a mobile laboratory in the field.