Optimization and Modeling of Glyphosate Removal by Nanofiltration at a Pilot Scale, Using Response Surface Methodology

Abstract

The removal of glyphosate by nanofiltration of contaminated water with a glyphosate commercial formulation at a pilot scale was studied. The combined effect of glyphosate concentration in feed [Gly], pH and the transmembrane pressure (TMP) at 20 °C was investigated and optimized for the first time using Response Surface Methodology. The optimum values of these factors were 160 mg/L, 10 and 4 bar respectively. A rejection of glyphosate of 99.6% was estimated and verified under these optimal conditions. Glyphosate remaining in permeate was below the limit established by the U.S. EPA (0.7 mg/L). The acute toxicity tests with fish in permeate showed that the rest of the toxic components of the glyphosate formulation were also removed. The high rejections of glyphosate despite its molecular weight below the molecular weight cut-off of the membrane were related to the combined effect of Donnan Exclusion and Dielectric Exclusion. The adjusted model was adequate with an R2 = 0.96. The linear and quadratic effects of pH and [Gly] factors were statistically significant (pvalue <0.05), as well as the antagonistic interaction between the two factors. The pH was the factor with major effect on rejection, followed by [Gly], the TMP effects were not relevant from the practical point of view.