Abstract
The feasibility of using a 2(6-1) fractional factorial design to screen the relative importance of six water quality and operational factors in the removal of microcystin-LR (MC-LR) by powdered activated carbon (PAC) was evaluated through jar testing. The factors were: PAC type, PAC dose, total organic carbon (TOC) concentration, turbidity, alum dose, and timing of PAC versus coagulant application. Follow-up tests were performed to examine the interaction of PAC dose and TOC concentrations. All MC-LR analyses were performed by enzyme linked immunosorbent assay (ELISA) and liquid chromatography coupled with tandem mass spectrometry (LC/MS/MS). The top three effect magnitudes were the same by ELISA and LC/MS/MS:PAC dose > PAC type > PAC application time. Correlation coefficients between removals estimated by ELISA and LC/MS/MS were >0.9 (p « .05). With both methods, the effects of PAC type and dose were found to be markedly larger than the other factors. The follow-up tests indicated a greater impact of PAC dose at higher natural organic matter concentrations. Factorial designs are not commonly used to plan drinking water jar test experiments. The results generated in this study were plausible with respect to the existing body of adsorption knowledge, thus helping to demonstrate the feasibility of the factorial approach.