Membrane ozonation of bromide-containing, high-color natural organic matter (NOM) containing groundwater was performed using single-tube polydimethylsiloxane (PDMS) and multi-tube polytetrafluoroethylene (PTFE) membrane contactors, and compared to batch ozonation. For membrane ozonation, dissolved ozone concentration, water color (VIS₄₃₆), ultraviolet light absorption (UV₂₅₄) and bromate formation were correlated with ozone dose, ozone gas concentration, hydraulic retention time and Hatta number (Ha). NOM color removal of up to 45 % for the single-tube contactor and 17 % for the multi-tube contactor were achieved while containing bromate formation below 10 µg L^-1. Higher color removal using higher ozone doses was associated with high bromate formation i.e. >>10 µg L^-1. In membrane ozonation, low ozone gas concentrations, long hydraulic retention times and high Ha resulted in low dissolved ozone concentrations due to quenching of ozone by NOM. At specific ozone doses of < 0.5 mg O₃/mg DOC and Ha > 1, single-tube ozonation resulted in comparable results to batch ozonation while bromate formation was higher in the single-tube contactor at specific ozone doses > 0.5 mg O₃/mg DOC and Ha < 1. At comparable ozone doses and Ha, bromate formation in the multi-tube contactor was always higher compared to single-tube and batch ozonation. This could be associated with the uneven ozone distribution within the multi-tube contactor. Results show that ozone dose is the major driver for selectivity between bromate formation and NOM color removal in both membrane and batch ozonation. Bromate formation in membrane ozonation may be controlled by adjusting gas concentration, Ha and hydraulic retention time. Membrane module design and process parameters of membrane ozonation reactors significantly affect treatment performance and should be optimized for selective target compound removal over by-product formation.

Jörn Sellmann

Publications