| dc.description.abstract | Sold phase micro extraction (SPME) has grown wide popularity in recently years in analyzing volatile or semi-volatile organic compounds in aqueous samples due to its ease of use, adequate sensitivity, solvent free in extraction, etc. The most attractive advantage with SPME over other means of extraction lies in the fact that no solvent is needed in the extraction procedure, the polymer phase coated on a thin fused silica fiber is employed as the extraction media to absorb organic analytes in aqueous samples. Subsequently, the fiber is inserted in the GC injection port to rapidly desorb the analytes onto the column. Assuming the volume of the polymer phase is negligible compared to that of the aqueous phase, the sensitivity of this technique is proportional to the volume of the polymer phase. Nevetheless, the geometry of regular GC injector forbids any forms of sample introduction other than needle type of insertion, as which the commercial SPME device is configured. To further lower the detection limit of SPME the injection port has to be redesign to accommodate a polymer phase of significant larger volume than the commercial SPME fiber.
In this study, a thick film of polydimethylsiloxane (PDMS) polymer was coated on a glass rod of 5 mm O.D. x 5 cm length. A renovated thermal desorption device was designed and constructed to accommodate the thick PDMS rod. Furthermore, to avoid peak broadening in chromatography, a microtrap packed with chemical sorbents was employed to focus analytes desorbed from the large volume of PDMS, which resulted in comparable peak width but considerable peak height than the commercial SPME. Based on the analysis of standards containing 50 ?g/L of benzene, toluene, ethylbenzene, xylene and various true underground water samples, the self prepared large volume PDMS and accompanying device is several tens fold to 2 order of magnitude in sensitivity enhancement.
We further employed the advantage of large volume PDMS extraction technique to develop a unique calibration method for BTEX based on the rapid exponential decrease in concentration from repeated extraction of a standard mixture. A linear relationship of log FID response versus extraction number was determined to be linear, which is consistent with the partition theory of extraction. Although the calibration was based on one standard mixture of unknown concentrations, the statistic risk arising from one-point calibration was minimized by the linear fitting of the numerous points generated by the multiple extraction procedure. | en_US |