Use of experimental design in the investigation of stir bar sorptive extraction followed by ultra-high-performance liquid chromatography–tandem mass spectrometry for the analysis of explosives in water samples

Abstract A method for the sensitive quantification of trace amounts of organic explosives in water samples was developed by using stir bar sorptive extraction (SBSE) followed by liquid desorption and ultra-high performance liquid chromatography–tandem mass spectrometry (UHPLC–MS/MS). The proposed method was developed and optimized using a statistical design of experiment approach. Use of experimental designs allowed a complete study of 10 factors and 8 analytes including nitro-aromatics, amino-nitro-aromatics and nitric esters. The liquid desorption study was performed using a full factorial experimental design followed by a kinetic study. Four different variables were tested here: the liquid desorption mode (stirring or sonication), the chemical nature of the stir bar (PDMS or PDMS-PEG), the composition of the liquid desorption phase and finally, the volume of solvent used for the liquid desorption. On the other hand, the SBSE extraction study was performed using a Doehlert design. SBSE extraction conditions such as extraction time profiles, sample volume, modifier addition, and acetic acid addition were examined. After optimization of the experimental parameters, sensitivity was improved by a factor 5–30, depending on the compound studied, due to the enrichment factors reached using the SBSE method. Limits of detection were in the ng/L level for all analytes studied. Reproducibility of the extraction with different stir bars was close to the reproducibility of the analytical method (RSD between 4 and 16%). Extractions in various water sample matrices (spring, mineral and underground water) have shown similar enrichment compared to ultrapure water, revealing very low matrix effects.