ABSTRACT
Reynolds-Averaged Navier-Stokes (RANS) approach with the k-ε closure model is employed for the first time to simulate direct Solid-Phase Micro-Extraction (SPME) computationally. Simulations are performed by using COMSOL Multiphysics in order to examine methods to decrease the extraction time. Experiments are also conducted to support data obtained from the numerical framework. Di-n-Butyl Phthalate (DNBP) and etched steel wire are chosen as the analyte and the adsorbent, respectively. Stirring rate, fiber's location, stirrer magnet's size, and the method of sample rotation are examined to decrease the extraction time. In addition, the effects of adding a baffle to the vial and implementing a periodic function for stirring the sample are studied. Increasing the stirring rate from 700 to 3000 rpm reduces the extraction time from 50 to 36 minutes. Present results suggest that changing the fiber's location and implementing a periodic stirring pattern significantly decrease the extraction time of SPME. The optimum fiber's location is found at 7.5 mm far from the vial's center, which leads to a 51% decrease in the extraction time. Furthermore, using a periodic stirring pattern results in a 60% decrease in the extraction time compared to the traditional uniform stirring pattern. To examine the robustness and reliability of the presented framework, two additional analytes, namely pyrene and Dioctyl phthalate (DOP), are chosen with one additional fiber, Al2O3. In this regard, the additional materials are used in the framework to investigate the effects of fiber's location and stirring pattern. By optimizing the fiber's location, the extraction time is reduced by 36% and 40% for SPME of pyrene and DOP, respectively. On the other hand, by using a periodic stirring pattern, the extraction time is reduced by 42% and 46% for SPME of pyrene and DOP, respectively.
