[Headspace solid phase micro-extraction method optimization for residual solvent analysis].

Abstract

A systematic headspace SPME method optimization is described for the residual solvent analysis using a polidimethyl-siloxane/divinylbenzene (PDMS/DVB) fiber. The first step was the chromatographic system optimization in which a narrow bore capillary column with a thick film (1 micron) of stationary phase, low starting column temperature (30 degrees C) and a narrow bore injector liner (2 mm I.D.) were used. It was found that for the investigated components a desorption temperature of minimum 150 degrees C should be used. The second step was the extraction optimization. The optimum equilibration time for all components was 30 min. It was found that the sample headspace volume has an important effect on method sensitivity and precision. At low headspace volumes (less than 1/3 of vial volume) sensitivity improves but at the same time precision worsen. The optimum headspace volume was found to be 4.6 ml. The total organic content does also have an important effect on method sensitivity and precision. At low organic content sensitivity increases but precision drops significantly. Over 1% organic content in the sample the system becomes unstable due to stationary phase swelling by the organic components. The optimum range for total organic content was found to be between 0.1% and 1%. The added NaCl quantity does increase the extraction yield. The optimum salt quantity to be added was 2 g NaCl in 5 ml sample. The last step was the desorption optimization. The influence of injector temperature and injection depth on desorption were investigated and it was found that it does not have an important effect on desorption yield. The optimum desorption temperature was 220 degrees C and the optimum injection depth was 2 cm into the injector. Among the investigated fibers the best detection limits and chromatographic behavior were given by the PDMS/DVB fiber. The measured detection limits were between 10 and 100 pg/ml and the RSD data were between 1-3%.