Construction and validation of a full-scale model for rejection of organic micropollutants by NF membranes

Abstract

A transport model for the rejection of charged and uncharged organic solutes was derived and validated in this paper. The rejection of 21 pharmaceuticals and 6 pesticides was determined on a pilot plant, operating at 75%, 82.5% and 90% recovery. For the neutral organic solutes, a solute transport model was constructed, based on a convection–diffusion model. The partition coefficient of the solutes in the membrane phase was determined with a steric model, and then compared with a fitted partition coefficient, incorporating solute–membrane affinity. For the charged solutes, the “charge concentration polarisation” concept was used to model the rejection. Both models for charged and uncharged organic solutes were then incorporated into a general model, and the influence of hydrodynamic concentration polarisation was introduced into the model. Afterwards, the general model was translated to a full-scale rejection model, which was then validated. The rejection model seemed to fit/predict the experimentally determined rejection values for neutral, as well as charged organic solutes quite well, at all the different feed water recoveries used in this study. The use of the steric model for the partition coefficient led to an overestimation of rejection values, because solute–membrane affinity was not incorporated.