The influence of solute polarizability and molecular volume on the rejection of trace organics in loose nanofiltration membrane processes

Abstract

The removal of trace organic compounds of emerging concern (TrOC) from ground water was evaluated using a split-feed, center exit, nanofiltration (NF) pilot process. Ground water was dosed with varying amounts of bisphenol-A, caffeine, carbamazepine, N,N-diethyl-meta-toluamide, estrone, gemfibrozil, naproxen, sucralose, and sulfamethoxazole between 150 ng/L and 4.5 mg/L, and processed with NF membranes operating at a feed flow rate of 60,636 L/h (267 gpm), a flux rate of 25.6 L m⁻² h⁻¹ (15.1 gsfd), and 85% water recovery. TrOC rejection by the NF process ranged from 68% for caffeine to below detection for gemfibrozil and sucralose. Correlations between rejection and various chemical and physical compound properties were investigated. It was found that TrOC rejection correlated well with polarizability (0.94 R²) and molecular volume (0.94 R²), and to a lesser extent hydrophobicity/hydrophilicity (0.87 R²). However, in this work, molecular weight and log D were not well correlated with solute rejection. Analysis of TrOC rejection data collected from five prior independent loose NF research studies representing 61 different TrOCs were found to correlate well with polarizability (0.71 R²) and molecular volume (0.72 R²), suggesting that polarizability and molecular volume are useful in estimating TrOC removal from fresh ground water using loose NF membrane processes.

Keywords:

• Pharmaceuticals
• Compound of emerging concern
• Pilot plant
• Polarizability
• Endocrine disrupting compound

Acknowledgments

This work was funded by Jupiter Water Utilities (Jupiter, FL) and Kimley-Horn & Associates, Inc. (West Palm Beach, FL) via UCF project 16208114. The authors would like to acknowledge the Town of Jupiter Utilities staff, including David Brown, Amanda Barnes, Paul Jurczak, and Rebecca Wilder, for their assistance and support. The authors would also like to acknowledge engineering consultants who assisted on this project, including Ian Watson (RosTek Associates Inc.) and John E. Potts (Kimley-Horn & Associates, Inc.). Additional thanks are offered to Dr Cherie Yestrebsky and Dr Carolina Franco (UCF Chemistry Department) for their assistance in obtaining a portion of trace organic compound data, and for UCF CECE Department graduate and undergraduate students who assisted in the pilot experiments.