Impact of flow regime on pressure drop increase and biomass accumulation and morphology in membrane systems

Biomass accumulation and pressure drop development have been studied in membrane fouling simulators at different flow regimes. At linear flow velocities as applied in practice in spiral wound nanofiltration (NF) and reverse osmosis (RO) membranes, voluminous and filamentous biofilm structures developed in the feed spacer channel, causing a significant increase in feed channel pressure drop. Elevated shear by both single phase flow (water) and two phase flow (water with air sparging: bubble flow) caused biofilm filaments and a pressure drop increase. The amount of accumulated biomass was independent of the applied shear, depending on the substrate loading rate (product of substrate concentration and linear flow velocity) only. The biofilm streamers oscillated in the passing water. Bubble flow resulted in a more compact and less filamentous biofilm structure than single phase flow, causing a much lower pressure drop increase. The biofilm grown under low shear conditions was more easy to remove during water flushing compared to a biofilm grown under high shear. To control biofouling, biofilm structure may be adjusted using biofilm morphology engineering combined with biomass removal from membrane elements by periodic reverse flushing using modified feed spacers. Potential long and short term consequences of flow regimes on biofilm development are discussed. Flow regimes manipulate biofilm morphology affecting membrane performance, enabling new approaches to control biofouling.     

Phosphate limitation to control biofouling

Phosphate limitation as a method to control biofouling of spiral wound reverse osmosis (RO) membranes was studied at a full-scale installation fed with extensively pretreated water. The RO installation is characterized by (i) a low feed channel pressure drop increase and (ii) low biomass concentrations in membrane elements at the installation feed side. This installation contrasted sharply with installations fed with less extensively pretreated feed water (and therefore higher phosphate concentrations) experiencing a high-pressure drop increase and high biomass concentrations in lead elements. Membrane fouling simulator (MFS) studies showed that low phosphate concentrations (∼0.3 μg P L⁻¹) in the feed water restricted the pressure drop increase and biomass accumulation, even at high substrate (organic carbon) concentrations. In the MFS under ortho-phosphate limiting conditions, dosing phosphonate based antiscalants caused biofouling while no biofouling was observed when acids or phosphonate-free antiscalants were used. Antiscalant dosage could increase both phosphate and substrate concentrations of the water. Therefore, antiscalant selection may be critical for biofouling control. Since no biofouling was observed at low phosphate concentrations, restricting biomass growth by phosphate limitation may be a feasible approach to control biofouling, even in the presence of high organic carbon levels.     

Effect of conventional chemical treatment on the microbial population in a biofouling layer of reverse osmosis systems

The impact of conventional chemical treatment on initiation and spatiotemporal development of biofilms on reverse osmosis (RO) membranes was investigated in situ using flow cells placed in parallel with the RO system of a full-scale water treatment plant. The flow cells got the same feed (extensively pre-treated fresh surface water) and operational conditions (temperature, pressure and membrane flux) as the full-scale installation. With regular intervals both the full-scale RO membrane modules and the flow cells were cleaned using conventional chemical treatment. For comparison some flow cells were not cleaned. Sampling was done at different time periods of flow cell operation (i.e., 1, 5, 10 and 17 days and 1, 3, 6 and 12 months). The combination of molecular (FISH, DGGE, clone libraries and sequencing) and microscopic (field emission scanning electron, epifluorescence and confocal laser scanning microscopy) techniques made it possible to thoroughly analyze the abundance, composition and 3D architecture of the emerged microbial layers. The results suggest that chemical treatment facilitates initiation and subsequent maturation of biofilm structures on the RO membrane and feed-side spacer surfaces. Biofouling control might be possible only if the cleaning procedures are adapted to effectively remove the (dead) biomass from the RO modules after chemical treatment.     

Short-term adhesion and long-term biofouling testing of polydopamine and poly(ethylene glycol) surface modifications of membranes and feed spacers for biofouling control

Ultrafiltration, nanofiltration membranes and feed spacers were hydrophilized with polydopamine and polydopamine-g-poly(ethylene glycol) surface coatings. The fouling propensity of modified and unmodified membranes was evaluated by short-term batch protein and bacterial adhesion tests. The fouling propensity of modified and unmodified membranes and spacers was evaluated by continuous biofouling experiments in a membrane fouling simulator. The goals of the study were: 1) to determine the effectiveness of polydopamine and polydopamine-g-poly(ethylene glycol) membrane coatings for biofouling control and 2) to compare techniques commonly used in assessment of membrane biofouling propensity with biofouling experiments under practical conditions. Short-term adhesion tests were carried out under static, no-flow conditions for 1 h using bovine serum albumin, a common model globular protein, and Pseudomonas aeruginosa, a common model Gram-negative bacterium. Biofouling tests were performed in a membrane fouling simulator (MFS) for several days under flow conditions similar to those encountered in industrial modules with the autochthonous drinking water population and acetate dosage as organic substrate. Polydopamine- and polydopamine-g-poly(ethylene glycol)-modified membranes showed significantly reduced adhesion of bovine serum albumin and P. aeruginosa in the short-term adhesion tests, but no reduction of biofouling was observed during longer biofouling experiments with modified membranes and spacers. These results demonstrate that short-term batch adhesion experiments using model proteins or bacteria under static conditions are not indicative of biofouling, while continuous biofouling experiments showed that membrane surface modification by polydopamine and polydopamine-g-poly(ethylene glycol) is not effective for biofouling control.     

Biofouling of spiral-wound nanofiltration and reverse osmosis membranes: A feed spacer problem

Biofouling was studied in full-scale and pilot-scale installations, test-rigs and membrane fouling monitors by conventional methods as well as Magnetic Resonance Imaging (MRI). Independent of permeate production, the feed spacer channel pressure drop and biomass concentration increased similarly in a nanofiltration pilot installation. In the presence of a feed spacer the absolute feed channel pressure drop increase caused by biomass accumulation was much higher than when a feed spacer was absent: in both spiral-wound nanofiltration and reverse osmosis systems biofouling is dominantly a feed spacer problem. This conclusion is based on (i) in-situ visual observations of the fouling accumulation, (ii) in-situ non-destructive observations of the fouling accumulation and velocity distribution profiles using MRI, and (iii) differences in pressure drop and biomass development in monitors with and without feed spacer. MRI studies showed that even a restricted biofilm accumulation on the feed channel spacer influenced the velocity distribution profile strongly. Biofouling control should be focused on the development of low fouling feed spacers and hydrodynamic conditions to restrict the impact of biomass accumulation on the feed channel pressure drop increase.     

Impact of ZnO embedded feed spacer on biofilm development in membrane systems

The concept of suppressing biofouling formation using an antibacterial feed spacer was investigated in a bench scale-cross flow system mimicking a spiral wound membrane configuration. An antibacterial composite spacer containing zinc oxide-nanoparticles was constructed by modification of a commercial polypropylene feed spacer using sonochemical deposition. The ability of the modified spacers to repress biofilm development on membranes was evaluated in flow-through cells simulating the flow conditions in commercial spiral wound modules. The experiments were performed at laminar flow (Re = 300) with a 200 kDa molecular weight cut off polysulfone ultrafiltration membrane using Pseudomonas putida S-12 as model biofilm bacteria. The modified spacers reduced permeate flux decrease at least by 50% compared to the unmodified spacers (control). The physical properties of the modified spacer and biofilm development were evaluated using high resolution/energy dispersive spectrometry-scanning electron microscopy, atomic force microscopy and confocal laser scanning microscopy imaging (HRSEM, EDS, AFM and CLSM). HRSEM images depicted significantly less bacteria attached to the membranes exposed to the modified spacer, mainly scattered and in a sporadic monolayer structure. AFM analysis indicated the influence of the modification on the spacer surface including a phase change on the upper surface. Dead-live staining assay by CLSM indicated that most of the bacterial cells attached on the membranes exposed to the modified spacer were dead in contrast to a developed biofilm which was predominant in the control samples.     

Fouling characteristics of NF and RO operated for removal of dissolved matter from groundwater

A pilot study had been performed for about 6 months in order to investigate the removal efficiency of dissolved matter and its fouling potential during nanofiltration (NF) and reverse osmosis (RO) of local groundwater that was pretreated with an ultrafiltration (UF) membrane system. After pilot plant operation, autopsy tests were performed to identify the characteristics of foulants that were attached to the membrane surface. In the autopsy tests, the flux recovery for each specific cleaning scheme (hydraulic washing, acid cleaning, and alkaline cleaning) was also measured using a dead-end filtration cell unit. The washing solution used in each chemical cleaning was analyzed to identify major components of the foulants, and the membrane surface was observed using the scanning electron microscopy (SEM).Among three kinds of membranes tested, one NF and two RO membranes, the NF and RO1 membranes showed a rapid flux decline after 100 days of operation. Especially, the RO1 membrane showed the more serious flux decline than the NF membrane. The RO2 membrane, with the lowest recovery rate, demonstrated a gradual flux decline. The removal efficiency of dissolved inorganic matter (as conductivity) for each NF, RO1 and RO2 membrane was 76.3%, 88.2% and 95.3%, respectively. The removal of dissolved organic matter (as total organic carbon) was found to be about 80% for both NF and RO membranes used in this study. During the membrane autopsy tests, five sections of the fouled membrane were cut along each NF and RO membrane module from the feed inlet side to the concentrate outlet side, the specific flux for each membrane section was measured before and after each cleaning step. As expected, the degree of fouling was intensified along the membrane surface as the feed flow approached the outlet. Based on the analysis results of wash water used in each cleaning step, the major foulants attached to the membrane surface appeared to be Ca bound with inorganic matter and Si bound with organic matter. Fe seemed to be a great contributor to irreversible fouling. The SEM analysis indicated that the organic matter was forming the first fouling layer close to the membrane and that the inorganic matter was layered top of the organic fouling layer in a tetragonal shape. Any evidence of biofouling was not observed in this study because most of microorganisms had been already removed by the UF pretreatment.     

Factors affecting filtration characteristics in membrane-coupled moving bed biofilm reactor

Factors affecting filtration characteristics in submerged hollow fiber membrane were investigated in membrane-coupled moving bed biofilm reactor (M-CMBBR). The trend of membrane biofouling in M-CMBBR was quite different from that in a conventional membrane bioreactor (MBR). The M-CMBBR showed much lower biofouling rate than a conventional MBR. Whereas the membrane biofouling in conventional MBR system is known to be dependent mostly on biochemical effects of mixed liquor (soluble COD, EPS, etc.), the extent of biofouling in M-CMBBR was largely dependent on the potential collision energy of biofilm carriers (media) moving freely and colliding with surfaces of submerged hollow fibers. The collisions between circulating media and hollow fiber membranes gave rise to frictional forces which mitigated the formation of biofilms on the outer surface of hollow fibers. Consequently, the membrane permeability was greatly enhanced. The potential collision energy of moving media was dependent on the media volume fraction as well as the air flow rate. The membrane permeability was found to be proportional to the relative potential collision energy of the biofilm carriers. The frictional effect on the morphology of biofilms formed on the surface of organic membrane under various operating condition was also examined and identified through their visualization with SEM and AFM.     

Do biological-based strategies hold promise to biofouling control in MBRs?

Biofouling in membrane bioreactors (MBRs) remains a primary challenge for their wider application, despite the growing acceptance of MBRs worldwide. Research studies on membrane fouling are extensive in the literature, with more than 200 publications on MBR fouling in the last 3 years; yet, improvements in practice on biofouling control and management have been remarkably slow. Commonly applied cleaning methods are only partially effective and membrane replacement often becomes frequent. The reason for the slow advancement in successful control of biofouling is largely attributed to the complex interactions of involved biological compounds and the lack of representative-for-practice experimental approaches to evaluate potential effective control strategies. Biofouling is driven by microorganisms and their associated extra-cellular polymeric substances (EPS) and microbial products. Microorganisms and their products convene together to form matrices that are commonly treated as a black box in conventional control approaches. Biological-based antifouling strategies seem to be a promising constituent of an effective integrated control approach since they target the essence of biofouling problems. However, biological-based strategies are in their developmental phase and several questions should be addressed to set a roadmap for translating existing and new information into sustainable and effective control techniques. This paper investigates membrane biofouling in MBRs from the microbiological perspective to evaluate the potential of biological-based strategies in offering viable control alternatives. Limitations of available control methods highlight the importance of an integrated anti-fouling approach including biological strategies. Successful development of these strategies requires detailed characterization of microorganisms and EPS through the proper selection of analytical tools and assembly of results. Existing microbiological/EPS studies reveal a number of implications as well as knowledge gaps, warranting future targeted research. Systematic and representative microbiological studies, complementary utilization of molecular and biofilm characterization tools, standardized experimental methods and validation of successful biological-based antifouling strategies for MBR applications are needed. Specifically, in addition, linking these studies to relevant operational conditions in MBRs is an essential step to ultimately develop a better understanding and more effective and directed control strategy for biofouling.     

Application of nanosilver surface modification to RO membrane and spacer for mitigating biofouling in seawater desalination

Biofouling is one the most critical problems in seawater desalination plants and science has not yet found effective ways to control it. Silver compounds and ions are historically recognized for their effective antimicrobial activity. Nanosilver particles have been applied as a biocide in many aspects of disinfection, including healthcare products and water treatment. This study proposes an innovative biofouling control approach by surface modification of the RO membrane and spacer with nanosilver coating. A chemical reduction method was used for directly coating nanosilver particles on the membrane sheet and spacer. The surface-modified membrane and spacer were tested for their antifouling performance in a cross-flow flat-sheet membrane cell, which is a part of a pilot plant in Wukan desalination plant. The silver-coating membranes and spacers, along with an unmodified membrane sheet, were tested in the membrane cell and compared on the basis of their antifouling performance. Permeate flux decline and salt rejection was continuously monitored through the testing period. Meanwhile regrowth of microbial populations on the membrane cell was quantified by a unique microbial counting every three to four days. The results showed that both silver-coated membrane (Ag-cM) with uncoated spacer and silver-coated spacer (Ag-cS) with uncoated membrane perforemed better than the unmodified membrane and spacer (Un-MS), in terms of much slower decrease in permeate flux and TDS rejection. However, the effect of silver-coated spacer on antimicrobial activity was more lasting. In the silver-coated spacer test, there was almost no multiplication of cells detected on the membrane during the whole testing period. Besides, the cells adhering to the membrane seemed to lose their activity quickly. According to the RO performance and microbial growth morphology, the nanosilver coating technology is valuable for use in biofouling control in seawater desalination.     

Combined coagulation-disk filtration process as a pretreatment of ultrafiltration and reverse osmosis membrane for wastewater reclamation: an autopsy study of a pilot plant

The effects of the combined coagulation-disk filtration (CC-DF) process on the fouling characteristics and behavior caused by interactions between effluent organic matter (EfOM) and the membrane surfaces of the ultrafiltration (UF) and reverse osmosis (RO) membranes in a pilot plant for municipal wastewater reclamation (MWR) were investigated. The feed water from secondary effluents was treated by the CC-DF process used as a pretreatment for the UF membrane to mitigate fouling formation and the permeate from the CC-DF process was further filtered by two UF membrane units in parallel arrangement and fed into four RO modules in a series connection. The CC-DF process was not sufficient to mitigate biofouling but the UF membrane was effective in mitigating biofouling on the RO membrane surfaces. Fouling of the UF and RO membranes was dominated by hydrophilic fractions of EfOM (e.g., polysaccharide-like and protein-like substances) and inorganic scaling (e.g., aluminum, calcium and silica). The desorbed UF membrane foulants included more aluminum species and hydrophobic fractions than the desorbed RO membrane foulants, which was presumably due to the residual coagulants and aluminum-humic substance complexes. The significant change in the surface chemistry of the RO membrane (a decrease in surface charge and an increase in contact angle of the fouled RO membranes) induced by the accumulation of hydrophilic EfOM onto the negatively charged RO membrane surface intensified the fouling formation of the fouled RO membrane by hydrophobic interaction between the humic substances of EfOM with relatively high hydrophobicity and the fouled RO membranes with decreased surface charge and increased contract angle.     

Quantitative biofouling diagnosis in full scale nanofiltration and reverse osmosis installations

Biofilm accumulation in nanofiltration and reverse osmosis membrane elements results in a relative increase of normalised pressure drop (ΔNPD). However, an increase in ΔNPD is not exclusively linked to biofouling. In order to quantify biofouling, the biomass parameters adenosine triphosphate (ATP), total cell count and heterotrophic plate count in membrane elements were investigated during membrane autopsies and compared with ΔNPD in test rigs and 15 full scale investigations with different types of feed water. The combination of biomass related parameters ATP and total cell count in membrane elements seem to be suitable parameters for diagnosis of biofouling, whereas plate counts were not appropriate to assess biofouling. The applied ΔNPD measurement was too insensitive for early detection of fouling. Measurements of biological parameters in the water were shown to be not appropriate in quantifying biofouling. Evidently, there is a need for a practical tool, sensitive pressure drop data and systematic research.     

正渗透去除单宁酸的影响因素分析

以单宁酸溶液作为原料液、NaCl溶液作为汲取液,在不同的膜朝向、汲取液和原料液浓度、膜面流速、原料液温度、pH值条件下,研究正渗透膜对单宁酸的去除效果。结果表明,汲取液浓度越高、原料液浓度越低、原料液温度越高、错流流速越大、膜活性层朝向汲取液运行模式均有助于提高膜通量,盐通量与膜通量呈正相关;不同影响因素条件下,正渗透膜对单宁酸的截留率均在97%以上,而在膜活性层朝向原料液模式下运行,对单宁酸的截留率更高。     

超滤膜过滤乳化油废水的试验研究

用聚砜膜制成的中空纤维超滤器，对不同浓度乳化油废水进行过滤。该膜对低浓度（１０ｍｇ／Ｌ左右）乳化油废水过滤时产水量大，极化阻力很小，膜阻力为１１．１３×１０１３（ｍ－１）；对高浓度（２００ｍｇ／Ｌ）乳化油废水过滤时产水量小，膜表面凝胶层形成很快，并得到膜过滤通量与膜两侧压力差之间的函数关系。用表面活性剂清洗后的膜性能与新膜基本相同     

Cleaning strategies for flux recovery of an ultrafiltration membrane fouled by natural organic matter

One of the most common problems encountered in water treatment applications of membranes is fouling. Natural organic matter (NOM) represents a particularly problematic foulant. Membranes may be fouled by relatively hydrophilic and/or hydrophobic NOM components, depending on NOM characteristics, membrane properties, and operating conditions. To maximize flux recovery for an NOM-fouled ultrafiltration membrane (NTR 7410), chemical cleaning and hydraulic rinsing with a relatively high cross-flow velocity were investigated as cleaning strategies. The modification of the membrane surface with either an anionic or a cationic surfactant was also evaluated to minimize membrane fouling and to enhance NOM rejection. Foulants from a hydrophobic NOM source (Orange County ground water (OC-GW)) were cleaned more effectively in terms of permeate flux by acid and caustic cleanings than foulants from a relatively hydrophilic NOM source (Horsetooth surface water (HT-SW)). An anionic surfactant (sodium dodecyl sulfate (SDS)) was not effective as a cleaning agent for foulants from either hydrophobic or hydrophilic NOM sources. High ionic strength cleaning with 0.1 M NaCl was comparatively effective in providing flux recovery for NOM-fouled membranes compared to other chemical cleaning agents. Increased cross-flow velocity and longer cleaning time influenced the efficiency of caustic cleaning, but not high ionic strength cleaning. The membrane was successfully modified only with the cationic surfactant; however, enhanced NOM rejection was accompanied by a significant flux reduction.     

Assessing the effects of bacterial predation on membrane biofouling

Membrane biofouling is one of the major obstacles limiting membrane applications in water treatment. In this study, Bdellovibrio bacteriovorus HD 100, a Gram-negative predatory bacterium, was evaluated as a novel way to mitigate membrane biofouling and its subsequent performance decline. Dead-end microfiltration (MF) tests were carried out on Escherichia coli DH5α and B. bacteriovorus HD 100 co-culture feed solutions. Predation of E. coli was performed at either a low or high multiplicity of infection (MOI), which is defined as the predator to prey cell ratio. The MOIs tested were 2 and 200, and the viability of both the E. coli prey and the predator was monitored over 48 h. The higher MOI (high predator, HP) culture showed a nearly 6-log loss in E. coli number after 24 h when compared to both the control and low MOI (low predator, LP) cultures, whereas the E. coli population within both predated cultures (HP and LP) became nearly identical at 48 h and 4-log lower than that of the control. The unpredated cultures led to significant loss in water flux at 12, 24, and 48 h of culture, but the HP and LP membranes showed less loss of flux by comparison. Analysis of the total membrane resistance showed a similar trend as the flux decline pattern; however, irreversible resistance of the membrane was much higher for the 48 h LP culture compared to the unpredated and HP cultures at 48 h. This increase in irreversible resistance was attributed mainly to E. coli debris, which accumulated in the medium after the predator lysed the prey cells. These results show that pretreatment of wastewater using a suitable concentration of predatory bacteria such as B. bacteriovorus can enhance membrane performance.     

SMBR在次临界通量下的运行特性

利用“通量阶式递增法”测定了两种膜组件的临界通量,在此基础上考察了次临界通量下的运行特性。试验发现,次临界通量操作下的膜污染过程具有明显的两阶段特征,与第一阶段跨膜压差(TMP)平缓直线上升相对应的膜污染机制主要是膜孔堵塞和凝胶层污染,与第二阶段TMP剧烈直线上升相对应的膜污染机制则是颗粒沉积层污染;先清水冲洗再化学清洗的方式能有效恢复膜的过滤能力,其中清水冲洗能有效去除颗粒沉积层污染,而化学清洗则能有效去除膜孔堵塞和凝胶层污染。     

Analysis of microfiltration performance with constant flux processing of secondary effluent

This study involves the microfiltration (MF) of secondary effluent from a sequencing batch reactor processing industrial waste. The MF unit was a hollow fibre module with gas backwash capability, and operated with pumped permeate (controlled flux) and dead-end, crossflow or intermittent feed. The results showed that crossflow had no effect on flux and that intermittent dead-end filtration was less productive than non-intermittent operation. For dead-end filtration the cycle-time between gas backwashes depends very significantly on the imposed flux (varying from about 100 min at 30 L/m² h to about 5 min at 90 L/m² h) and the feed solids content. Optimal operation has to balance operating (energy for backwash) costs and the capital (membrane area) costs. Cost analysis based on capital and energy costs indicates that for lower energy cost the unit needs to be operated at lower imposed flux but to minimise total cost it is necessary to operate the unit above 60 L/m² h imposed flux depending on the maximum transmembrane pressure (TMP) allowed before back washing. Further analysis of TMP profiles showed that membrane resistance increased over time towards a maximum, which tended to increase with imposed flux. This implies more frequent chemical cleaning for high flux operation. Specific cake resistances were deduced from the profiles and indicated cake compression at higher flux and larger maximum TMP. Results of long-term trials are also reported. Water quality analysis shows consistent quality of permeate regardless of operating conditions.     

Fouling indices for low pressure hollow fiber membrane performance assessment

This study evaluated the use of fouling indices to describe low pressure membrane fouling. One critical aspect of this study was the use of a bench-scale hollow fiber membrane system that imitated full-scale operation (constant flux with automatic hydraulic backwash and chemical cleaning). Fouling indices were based on a resistance-in-series model. Two different hollow fiber membrane types (membrane A and B) were tested with water from two water utilities (A and B) and three other natural sources (oligotrophic, algal bloom impacted, and wastewater impaired). The bench-scale testing included use of the same membrane as utilized at Utility B. Most fouling was reversible by hydraulic backwash and chemical cleaning. Specific flux and fouling indices for the bench-scale system were higher than those determined from full-scale data but fouling index ratios were comparable, suggesting a similar fouling nature. At similar organic loading, fouling was specific to water source and membrane type, i.e., no generalization on the impact of water source was possible. Full-scale data were compared with bench-scale data to validate the use of fouling indices. Fouling indices based on a resistance-in-series are useful tools to describe membrane performance data for both raw and pretreated water, for different water sources, and different membrane types.     

Impact of microfiltration treatment of secondary wastewater effluent on biofouling of reverse osmosis membranes

The effects of microfiltration (MF) as pretreatment for reverse osmosis (RO) on biofouling of RO membranes were analyzed with secondary wastewater effluents. MF pretreatment reduced permeate flux decline two- to three-fold, while increasing salt rejection. Additionally, the oxygen uptake rate (OUR) in the biofouling layer of the RO membrane was higher for an RO system that received pretreated secondary wastewater effluent compared to a control RO system that received untreated secondary effluent, likely due to the removal of inert particulate/colloidal matter during MF. A higher cell viability in the RO biofilm was observed close to the membrane surface irrespective of pretreatment, which is consistent with the biofilm-enhanced concentration polarization effect. Bacterial 16S rRNA gene clone library analysis revealed dominant biofilm communities of Proteobacteria and Bacteroidetes under all conditions. The Cramer-von Mises test statistic showed that MF pretreatment did not significantly change the bacterial community structure of RO membrane biofilms, though it affected bacterial community structure of non-membrane-associated biofilms (collected from the feed tank wall). The finding that the biofilm community developed on the RO membrane was not influenced by MF pretreatment may imply that RO membranes select for a conserved biofilm community.