Boron Removal in Seawater Desalination by Reverse Osmosis Membranes — the Impacts of Operating Conditions

摘要

Production of drinking water through seawater desalination using reverse osmosis (RO) membranes is becoming increasingly attractive especially in coastal areas with limited freshwater sources. However, one challenge in such conventional desalination RO plants is the difficulty of meeting boron standards in product waters. Therefore, most of the current desalination plants employ additional treatment steps including pH adjustment of feedwater, dilution of RO permeate with other sources, ion exchange post-treatment of RO permeate, and/or double-pass staging for permeate. All these further treatment options increase the cost of desalination. Although membrane manufacturers have been developing modified RO membranes with enhanced boron removal capacities such membranes still should be improved from operational flux and pressure perspectives. The main objective of this work was to determine the impacts of operational conditions (membrane pressure, cross-flow velocity and flux) and water chemistry on boron rejections using two commercial RO membranes specified for enhanced boron removal (TorayTM UTC-80-AB and FilmtecTM SW3OHR). A lab-scale cross-flow flat-sheet configuration test unit (SEPA CF II, Osmonics) was used for all RO experiments. Seawater samples were collected from the Mediterranean Sea, Alanya-Ktzilot shores, south Turkey. For all experiments, mass balance closures were between 91 and 107%, suggesting relatively low loss of boron on membrane surfaces during 14 h of operation. Boron rejections were relatively constant (a maximum change of ±3%) during the 14 h of operation period for all experiments, suggesting that steady state dynamic membrane conditions were immediately achieved within couple hours. Boron rejections obtained with Toray and FilmTec membranes at pH of original seawater (8.2) and at other various operating conditions ranged between 85 and 92%, resulting in permeate boron concentrations of about 0.2-0.9 mg/L. On the other hand, for both membranes, much higher boron removals were achieved at a pH of 10.5 (>98%), resulting in permeate boron concentrations less than 0.1 mg/L. The charged boron species are expected to be dominant at pH values >9.24 (pKa of boric acid) compared to the neutral boric acid. Therefore, as expected, both membranes exhibited higher boron rejections at a pH of 10.5. Salt rejections (as measured by conductivity) were generally 97-99% at both pH values. Boron rejections were independent of feed water boron concentrations up to 6.6 mg/L. For each membrane type, permeate fluxes at constant pressure were generally lower at pH of 10.5. The ranges of permeate fluxes measured in all experimental conditions were 11-15, 13-17 and 19-21 L/m~2-h for 600, 700 and 800 psi (41, 48 and 55 bar) pressures, respectively, after an operation period of 14 h. For all experimental conditions, permeate fluxes gradually decreased during the 14 h operation although a leveling off was observed after 12 h. At constant membrane pressure of 800 psi and pH of 8.2, feed flowrate thus the cross-flow velocity (0.9 and 0.5 m/s) did not exert any significant impact on boron rejection.