EFFECTS OF PARAMETERS ON SUPERCRITICAL-FLUID EXTRACTION OF TRIAZINES FROM SOIL BY USE OF MULTIPLE LINEAR-REGRESSION

The effects of various parameters on supercritical fluid extraction (S FE) of triazines from soil were studied with an experimental design, b ased on multiple linear regression. The SFE was performed using a mult i-extraction unit with simultaneous extraction of eight samples. Diffe rent types of soil samples were spiked with a series of five triazines with different polarities. The developed design was a compromise of v arious objectives like the relative importance of the different parame ters, the total amount of experiments and instrumental limitations. El even series of experiments using different conditions were performed, resulting in a data set of over 200 data. Regression analysis was appl ied to evaluate the data set of each individual triazine component. Fu rthermore, the influence of the different parameters was tested, resul ting in a limitation of the original parameter set as well as a combin ation of some parameters to avoid interactions. The influence of the p ressure on the recovery appeared to be very important, recoveries incr eased with increasing pressures. The influence of the modifier was als o essential, only when it was added directly to the extraction cell, a nd the effect is increasing with component polarity. The effects of th e temperature and extraction time were slightly negative and not signi ficant, whereas a small effect of the type of soil was observed. Two o ther models, combining the whole data set for all triazines, were appl ied resulting in a more pronounced effect of the individual parameters . Multiple linear regression appeared to be a useful tool to study the effects of the many parameters in SFE, in order to reduce the number of experiments, to facilitate the evaluation of data and to distinguis h possible interactions between several parameters.