Strategies for Coagulation Pretreatment of Low-Pressure Membranes

Abstract

To mitigate natural organic matter (NOM)-induced fouling of low-pressure membranes (LPMs) in drinking water applications, coagulation pretreatment is widely applied. However, multiple process configurations exist and simulating full-scale performance at bench-scale remains a significant challenge. Accordingly, the main objectives of this dissertation were to compare the performance of alternative coagulation pretreatment strategies and to evaluate coagulation/flocculation mixing configurations for bench-scale LPM filtration studies. This research work was structured into four phases: (i) comparison of alternative coagulation pretreatment strategies across seasons and membrane types to identify optimal approaches; (ii) systematic investigation of the impact of helically coiled tube flocculator (HCTF) geometry on floc properties; (iii) evaluation of the influence of mixer configuration on bench-scale inline coagulation (C-IN) performance; and (iv) assessment of phased inline coagulation (C-IN-P), where coagulant is added for only a portion of the filtration cycle, as a more sustainable and cost-effective alternative to continuous C-IN. Conventional coagulation-flocculation-sedimentation (CF-SED) has proven effective for LPM pretreatment, although, coagulation-flocculation-dissolved air flotation (CF-DAF) represents a promising high-rate alternative. C-IN, which omits particle separation prior to membrane filtration, has gained popularity due to its reduced footprint and operational simplicity. For two different membrane types, CF-DAF and CF-SED performed equally well, despite improved NOM removal by CF-DAF. In contrast, the performance of C-IN was membrane dependent as it yielded the lowest fouling rate amongst alternatives for an inside-out polyethersulfone (PES) membrane, but the highest fouling rate for an outside-in polyvinylidene fluoride (PVDF) membrane. This finding highlighted the importance of membrane material/configuration for the selection of optimal LPM pretreatment, even when using the same source water. Seasonal differences in C-IN performance appeared to be caused by changes in the organic matrix affecting the coagulated zeta potential (ZP), suggesting that real-time ZP monitoring may serve as a useful optimization tool for C-IN-LPM hybrid systems. HCTFs represent a promising approach for improving continuous flow C-IN at bench-scale due to their strong secondary flow, which promotes effective mixing even for firmly laminar conditions. A replicated full-factorial design was used to evaluate the impact of the coil pitch, coil diameter, and pipe length on the properties of flocs. HCTF geometry was found to exert statistically significant impact on floc size and concentration, while floc structure (i.e., fractal dimension) remained relatively constant across all studied geometries. The system flowrate (i.e., Reynolds number), however, exhibited the greatest impact on HCTF performance, likely the result of a trade-off between the hydrodynamic condition and the hydraulic residence time. Compared to a batch-reactor, the tested HCTF configurations produced smaller sized flocs in greater concentrations, while relative to a commercial inline static mixer (ISM), HCTFs yielded larger sized flocs in lower concentrations. Bench-scale continuous-flow C-IN has proven difficult due to the low-flowrates involved. Accordingly, previous studies have used a wide variety of mixing configurations. The third phase of this study sought to elucidate the impact of mixer type/configuration on bench-scale C-IN performance by comparing three different continuous-flow flocculator types, namely, incidental flocculation in a straight pipe, an ISM, and two HCTFs, with and without the inclusion of rapid mixing. The performance of these continuous-flow approaches was additionally compared with a batch reactor, which have been extensively used in past studies due to their experimental simplicity. C-IN using a batch-reactor (with an extended contact time) yielded the lowest level of membrane fouling. Compared to rapid mixing or flocculation alone, combined rapid mixing and flocculation yielded reduced magnitude LPM fouling. Amongst alternative continuous-flow flocculator types, HCTFs had the lowest fouling rate, further supporting their suitability for bench-scale C-IN studies. These results indicate that LPM fouling is highly dependent on mixer type/configuration; direct comparison across studies using different C-IN approaches (including batch systems) may not be appropriate. The final phase evaluated the performance of C-IN-P using three synthetic waters with varying hydrophobicity. Intermittent addition over 75% of the filtration cycle achieved comparable fouling mitigation and NOM removal to continuous dosing, with only marginal performance losses when the length of coagulant addition was further reduced to 50% of the cycle. The optimal membrane coating index (C-IDX), an operational control tool for C-IN-P defined as the mass of coagulant per unit membrane area, was higher than values reported in the literature, even for the same membrane type and low hydrophobicity water types. This suggests that the optimal C-IDX is uniquely defined for each membrane-water combination. For a given C-IDX value, optimal performance was achieved when the jar-test optimal dose was used rather than different dosing rates applied for extended or reduced durations. Collectively, this work provides a comprehensive evaluation of alternative coagulation pretreatment strategies for LPMs and offers practical insights into the design, optimization, and bench-scale testing of these processes. The findings highlight the importance of membrane-specific optimization as well as the impact of mixer type/configuration on the outcome of bench-scale LPM studies.