UCLA

The objective and focus of this study is to fully understand trace organic pollutant transport through NF/RO membranes. An extension of the classical solution-diffusion model had been developed that relates transport through NF/RO membranes directly to membrane structure descriptors (i.e., effective barrier layer pore size, porosity and thickness, etc.). In general, model predictions agreed well with experimental data suggesting the model captures the phenomenological behavior of commercial NF/RO membranes for separations relevant to modern water treatment objectives. The model also provides new mechanistic insights about the "effective structure" of NF/RO composite membranes and how trace organic solutes are rejected. These results suggest it is possible and important to fine-tune the surface energy of membrane and membrane structure (pore size, porosity, thickness) to achieve high membrane selectivity for certain solute.

The effects of feed solution ionic strength, pH and divalent cation content on NF/RO membrane structure and performance were elucidated experimentally and fitted with the newly developed model. Generally, water permeabilities of all three membranes decreased with ionic strength and divalent cation content, but increased with pH. For RO membranes, neutral solute rejection decreased with pH and divalent cation content, but increased with ionic strength and the salt rejection remained independent with water chemistry except for very low pH of 3; for a NF membrane, solute rejection was more sensitive to water chemistry and neutral solute rejection decreased with ionic strength, pH, but increased with divalent cation content. Ultimately, these new insights may be useful in selection of already commercial or design of new NF/RO membranes for removal of chemicals of emerging concern in water treatment.

Four different organic solute removals by six different commercial NF/RO membranes in laboratory re-created groundwater matrix were experimentally determined. SWRO membranes exhibited excellent removal efficiency (> 90%) for both NDMA and 1,4-dioxane in groundwater, while NF membranes showed inefficient separation. Correlation studies suggested that both size exclusion and thermodynamic partitioning play important roles in trace organics removal and a partition coefficient, which combines both steric effects and solute-membrane interactions, can be employed to predict organic solute rejection by NF/RO membranes.