R&D Reports, Publications and Presentations
Title: Establishing Correlations Between Membrane Fouling and Water Composition
Publication Date: March 2005
Abstract:
The deterioration of RO membrane performance is determined by complex interactions resulting from process operating conditions, membrane properties, and source water quality. Because fouling necessitates application of greater pressure to maintain product flow, it requires the expenditure of more energy to produce the same amount of water, and thus can dramatically diminish process efficiency and cost efficiency in a water processing (purification) system. Analytical methods and equipment were developed to correlate membrane fouling with feed water quality parameters under controlled operating conditions. A self-contained, portable bench-scale RO test unit was fabricated and installed at locations with source waters including surface water, lime-clarified and microfiltered secondary treated wastewater, tertiary treated wastewater, and agricultural drainage water. Four commercially available membranes were tested: three polyamide membranes and one cellulose acetate membrane. Membranes were operated in a single pass configuration, each at a constant flux of 10 gfd and 4 percent recovery. Water quality analyses included general minerals, microbial analysis by heterotrophic plate counts and epifluorescence microscopy. At the end of each test, membranes were autopsied to determine the nature of the biological material accumulated on the membrane surface (protein, carbohydrate, microbiological analysis by heterotrophic plate counts and epifluorescence microscopy, community analysis by 16S rRNA). Collected data were compiled in a database. Chemical, physical, and biological characteristics of the source waters, membranes and fouling layers were measured and used to develop relationships between source water quality, membrane composition, and membrane fouling (loss of water flux and rejection) using artificial neural net analysis. Using the generated models it was possible to explain variations in data and their affect on specific water flux and percent rejection.
Article Download: CEC RO Fouling Database Project Final Report (11.37 MB)
Project Funding:
Funding provided by the California Energy Commission through the Desalination Research and Innovation Partnership (DRIP), managed by the Metropolitan Water District (MWD) of Southern California.
Title: Sensitivity of RO Membranes Materials to Chlorine
Publication Date: March 2005
Abstract:
The treatment of municipal wastewater using advanced membrane processes such as reverse osmosis (RO) has gained in popularity due to the development of the polyamide (PA) membrane. This membrane is attractive because of the high water quality produced and the relatively low pressures required for operation. Despite its many advantages, the PA membrane is prone to biological fouling, especially when treating municipal wastewater. Disinfectants such as chlorine are typically used in wastewater treatment plants to maintain stable levels of bacterial activity. This can pose a problem when operating PA membranes since they are susceptible to oxidation and degradation from exposure to free chlorine and combined chlorine. The objective of this project was to evaluate the sensitivity of PA membranes materials to chlorine using RO test cells designed to simulate operations of manufactured membrane elements.
A series of four commercial RO membranes were evaluated using a small RO test cell designed to simulate operations of a spiral RO membrane element. Membranes selected represent those used at the various member agencies of the Desalination Research and Innovation Partnership and include the BW-30 (Dow FilmTec), ESPA-2 and LFC-1 (Hydranautics), and TFC-HR (Koch Membrane Systems). Membrane performance (defined as water transport and solute (salt) rejection) was evaluated upon operations with free chlorine (as hypochlorite (OCl-) and hypochlorous acid (HOCl)) and combined chlorine (as monochloramines (NH2Cl)). Combined chlorine testing was conducted in the presence and absence of iron (as Fe+2) to confirm whether or not a transition metal can influence the rate of membrane oxidation. Results indicated that all PA membranes, regardless of manufacturer, were susceptible to oxidative degradation in varying degrees depending on the chlorine species: HOCl > OCl-> NH2Cl+iron > NH2Cl. RO membrane life and replacement may not be dictated solely by the incidence of irreversible fouling. Rather, consideration of PA membrane degradation resulting from long-term exposure to chlorine is warranted.
Article Download: CEC Oxidation Final Report 2005
Project Funding:
Funding provided by the California Energy Commission through the Desalination Research and Innovation Partnership (DRIP), managed by the Metropolitan Water District (MWD) of Southern California.
Title: Evaluation of Commercially Available High-Voltage Capacitance (HVC) and Magnetic Water Treatment (MWT) Devices for their Effectiveness in Limiting RO and MF Fouling to Enhance Overall
Membrane Performance
Publication Date: March 2005
Abstract:
The treatment of municipal wastewater using advanced membrane processes such as reverse osmosis (RO) and microfiltration (MF) are subject to biological, colloidal and precipitative fouling. Resulting operations and maintenance costs associated with increased energy requirements, chemical cleaning frequencies and membrane replacement challenge the economic feasibility of membrane treatment. Two, nontraditional, electro-technologies were tested as a means of additional pretreatment prior to MF and RO in an effort to limit the incidence of membrane biological fouling. These technologies consisted of a high-voltage, capacitor-based system (HVC) and a magnetic water treatment (MWT) device. These technologies were selected for testing because
they could: (1) provide continuous cleaning in place; (2) ease of installation into existing RO and MF piping; (3) low operating cost and (4) minimal maintenance requirements. Identical, single-pass membrane systems, each capable of testing four spiral wound membranes, were designed and constructed for this study. Both systems were operated simultaneously with one system operating as a control and either the HVC or MWT device connected to the other. Each RO membrane was operated at a constant water flux of 12 gallons per ft2-day (gfd) and a recovery of 6% (17.5gfd and 3% recovery for the MF). Membrane performance including water flux and solute (salt) rejection were monitored daily. Results indicate that the HVC was effective in significantly increasing membrane performance in 3 of the 4 RO membranes tested. Operational data obtained from the MWT testing indicated that it was ineffective in improving performance in any of the 4 RO membranes. Evaluations of the HVC and MWT device for enhancement of MF membrane performance indicated that neither was effective.
Article Download: CEC ZetaLSI Final Report 2005
Project Funding:
Funding provided by the California Energy Commission through the Desalination Research and Innovation Partnership (DRIP), managed by the Metropolitan Water District (MWD) of Southern California.
Title: Evaluation of a Rapid Test for Quantifying Assimilable Organic Carbon (AOC) in Membrane Desalination Feedwaters
Publication Date: April 2005
Abstract:
This study was designed to evaluate new methods for rapid determination of assimilable organic carbon (AOC). The AOC represents a fraction of the total organic carbon (TOC) in water that bacteria can use for growth and other metabolic processes. High levels of AOC are associated with rapid biofilm formation, loss of membrane performance and poor water quality. Currently there is no rapid test available that can be used to monitor the levels of AOC. A Standard Method to determine the AOC concentrations is available, but requires 5-14 days to complete. A rapid bioassay made available by Checklight was evaluated for its speed, sensitivity, accuracy and complexity to perform the assay. A number of defined tests were run using the Checklight bioassay and it was determined that the sensitivity ranged from 3-100ppb using a mixed carbon solution. Whereas, the sensitivity range using simple and complex carbon compounds ranged from 5-100ppb. Light produced by Vibrio fischeri cells used in the assay was greater with glucose and fructose-carbon than with glycerol, sodium acetate and the mixed carbon solution. OCWD secondary municipal wastewater was tested using the Checklight bioassay and the values were compared to Standard Methods that uses different organisms. The two values were within 20% of each other. The AOC concentrations could not be calculated
when deep well injection water and Santa Clara secondary municipal wastewater were evaluated. A range between 34-55ppb was calculated for Fountain Valley potable drinking water. The Coulter Multizer was used to observe an increase in cell biomass by measuring cell volume. This method was tested to determine if the Coulter Multizer could be used to calculate growth rates. Several inconsistencies, which easily could not be explained, were observed with controlled samples and the OCWD secondary municipal wastewater at different time intervals. The Checklight assay was evaluated under defined and controlled conditions and it was determined that the assay is highly variable due to the physiology of V. fischeri cells. Several modifications are outlined in this study that improved the variability, sensitivity and accuracy of the assay.
Article Download: AOC Study Final Report
Project Funding:
Funding provided by the California Department of Water Resources through the Desalination Research and Innovation Partnership (DRIP), managed by the Metropolitan Water District (MWD) of Southern California.
Title: Role of Microfiltration (MF) Cake Layer Composition and Stability in Desalination Efficiency
Publication Date: April 2005
Abstract:
In recent years, microfiltration (MF) has been shown to be less costly and more effective than conventional lime clarification as a pre-treatment method for reverse osmosis (RO). Orange County Water District (OCWD) plans to utilize an MF/RO process in a 70-mgd advanced water reclamation facility. The feedwater for this large facility is municipal secondary treated wastewater, which possesses significant nutrient loading, and high biological activity. These factors contribute to rapid biofouling of the MF membranes with consequent loss of performance. MF cake formation is associated with increased hydrodynamic resistance (i.e., flux loss) and necessitates frequent chemical cleaning of the MF membranes, thereby driving up operational costs. Bacteria, nanoparticulates and dissolved organic matter (DOM) are the major components of membrane fouling. The primary objective of this project was to elucidate the fouling mechanism of MF membranes and to suggest mitigation strategies based on this mechanism. Experimental MF cakes were formed under controlled laboratory conditions using a 0.22-µm polypropylene membrane material and secondary treated wastewater. Two mechanisms of MF fouling were observed; the classical MF cake formation by microparticulates and fouling by microbial cell residue (such as DOM proteins, carbohydrates and phopholipids) and nanoparticulates. The cake layer (bacterial large particulate layer) is responsible for a relatively small fraction of the observed loss of hydraulic conductivity, and the cake is easily removed by backwashing (air sparging), which forces the cake off the surface. On the other hand, microbial residues such as DOM and nanoparticulates with dimensions smaller than 0.2 µm are responsible for the majority of the observed reduction of hydraulic conductivity and may be more difficult to remove. Adsorbed material may foul MF membranes in the absence of applied transmembrane pressure.
Article Download: DWRP13_MFCake_FinalReport
Project Funding:
Funding provided by the California Department of Water Resources through the Desalination Research and Innovation Partnership (DRIP), managed by the Metropolitan Water District (MWD) of Southern California.
Title: Identification and Evaluation of Unique Chemicals for Optimum Membrane Compatibility and Improved Cleaning Efficiency
Publication Date: April 2005
Abstract:
In an effort to gain a better understanding of the molecular interactions between chemical cleaning agents and reverse osmosis membranes, a group of potential chemical cleaning agents was exposed to eight different reverse osmosis membranes. These membranes included five thin-film composite polyamides (PA), a polyamide-urea (PA-U), a cellulose acetate (CA) and an experimental PA. Each membrane was exposed to 37 individual compounds that included nonionic, cationic, anionic, zwitterionic (neutral compound with both positive and negative charges), chelating and oxidizing chemical compounds. Following exposure to the cleaning agents it was determined that 82% of the compounds strongly associated with the membrane surface. The nonionic surfactant Mega 10 was the least reactive. The Hydranautics LFC3 and FilmTec BW-30 were the most resistant to the chemical exposure in that little chemical changes to these membranes were observed. Poor membrane performance often conflicted with the infrared spectroscopic analysis that indicated no adsorbed compounds on the surface. Limited applications of cleaning chemicals to fouled membranes produced mixed results, affecting water flux and solute flux inconsistently. Quantitative structure activity relationship (QSAR) molecular descriptors were generated by computer for each of the potential chemical-cleaning agents. Artificial neural network models were constructed and the correlation of molecular properties with membrane performance were determined. Three models were constructed (a PA, a CA and a PA-U) that were used to predict a chemical compound’s effect on water flux and solute flux. Charge, polarity and hydrogen bonding properties appeared to be the most influential factors in determining a compound’s effect on water flux and solute rejection.
Article Download: ROClean_Final5_15
Project Funding:
Funding provided by the California Department of Water Resources through the Desalination Research and Innovation Partnership (DRIP), managed by the Metropolitan Water District (MWD) of Southern California.
Title: Effects of Molecular and Environmental Properties on Removal of Pharmaceuticals, Endocrine Disruptors and Disinfection Byproducts by Polyamide RO Membranes
Publication Date: January 2008
Abstract:
Removal by thin film composite polyamide reverse osmosis (RO) membranes of a number of trace-organic compounds and hormones considered of interest by State and Federal regulatory agencies was studied using cross-flow membrane test cells. Compounds in RO feed and product water were analyzed using EPA methods 508 and 524 as well as methods developed by OCWD for hormones and endocrine disrupting compounds (EDCs). An empirical artificial neural network (ANN) model relating log compound removal to compound physicochemical properties and membrane properties was successfully constructed. Log removal was used as a transformation function instead of the traditional expression of percent rejection to provide a more even representation of performance value variations across the range of observed compound removal values. Log P most strongly influenced RO removal, followed by molecular weight and the number of methyl groups on the compound. Membrane roughness was weakly related to compound removal. The rejection behavior of membranes in response to altering feedwater quality parameters (pH and salinity) was also investigated. A surface-response statistical approach in which the range between the minimum and maximum predicted removal of the compound over a range of pH and salinity was used to define influence of these parameters. It was typically observed that variation in compound removal due to these environmental properties was not dramatic; overall, changes in removal over the study range of pH and salinity was less than one log. Highly rejected compounds were not observed to become poorly rejected, and vice versa.
Observations and conclusions were limited by the range of compounds, membranes and environmental variations employed in the study, but nonetheless help expand understanding of principals governing RO rejection of organic substances of public health concern.
Article Download: EPA II _OCWD_ 2008 Final Report EDC & PPCP removal-Bench
Project Funding:
Funding provided by the United States Environmental Protection Agency (USEPA) through the Desalination Research and Innovation Partnership (DRIP), managed by the Metropolitan Water District (MWD) of Southern California.
Title: Rejection of Pharmaceuticals by Reverse Osmosis Membranes: Quantitative Structure Activity Relationship (QSAR) Analysis
Publication Date: October 2004
Abstract:
In this study, 51 radiolabeled surrogate compounds selected from an initial compound list of over 200 organic compounds, mostly of public health concern, were used to construct a series of quantitative structure activity relationship (QSAR) based empirical multivariate models describing the interaction of the compounds with several commercial polyamide (PA) and cellulose acetate (CA) reverse osmosis (RO) membranes. Models were constructed using artificial neural networks (ANNs) based on data obtained from calculated QSAR molecular descriptors and direct measurements of compound-membrane associations. The penetration of molecules through the membranes, the adsorption/absorption of molecules on/in the membranes and the rejection of molecules at the feed/membrane interface were associated with molecular properties that included charge/polarity, structural complexity, hydrogen bonding and hydrophobicity. Percent rejection, calculated from the ANN model predictions, compared favorably with published values. Models developed in this study were capable of predicting the compound-membrane interactions of 57% to 70% of the organics in the initial compound list. In addition to the individual membrane models, a “Universal” PA model was constructed from individual PA membrane performance data capable of predicting the compound-membrane interactions for 76% of the compounds. A gap analysis that could improve model performance was discussed. A fully-atomistic geometry-optimized model of a PA membrane was created and used to study the free diffusion behavior of 1,1,2,2 tetrachloroethylene (PCE) and N-nitrosodimethylabine (NDMA). Predicted PCE diffusion was 4-fold less than NDMA. This result agreed in general with the relationship between PCE and NDMA relative membrane fluxes; however, absolute values were much overestimated compared to laboratory results, although water flux measurements were not. Movement of compounds through the membrane by low-frequency, longer range “jumps” as opposed to local diffusion and underestimation of the solute partition coefficients may account for the discrepancies. A simplified membrane model system using a single PA membrane “pore” to speed investigation of compound-membrane interactions is proposed.
Article Download: NWRI Organics Rejection QSAR Study Final Report
Project Funding:
Funding provided by the National Water Research Institute (NWRI), the West Basin Municipal Water District (WBMWD), the Orange County Sanitation District (OCSD) and the Unites States Bureau of Reclamation (US BuREC).
Title: Development of a Percolation Decay Model to Guide Future Optimization of Surface Water Recharge Basins
Publication Date: October 2007
Abstract:
Reductions in percolation over time due to foulant accumulation are often a significant problem with surface water recharge basins. The ability to predict the impact of foulant accumulation on basin percolation facilitates evaluation of the potential effectiveness of cleaning strategies or water pretreatment schemes in improving percolation efficiency. In this study, both laboratory and field data were used to develop a relatively simple mathematical model capable of describing basin percolation kinetics. This model suggests that initial percolation rate, foulant loading and the interaction between foulant and sediment at or near the sediment/water interface can describe the decay of percolation over time.
Simulating performance of a recharge basin indicates that the interaction of foulant with sediment and reduction of overall foulant loading is perhaps the most effective method of improving recharge basin efficiency.
Article Download: Don Phipps Percolation Decay Model for ISMAR6
Title: Percolation Decay Model for Optimization of Surface Water Recharge Basins
Publication Date: December 2008
Abstract:
Managed Aquifer Recharge (MAR) operations, such as those at the Orange County Water District (OCWD), are challenged to maintain sustainable high percolation rates in surface water recharge basins. At OCWD, water diverted from the Santa Ana River (SAR) is subject to several stages of gravity desilting before entering terminal recharge basins. Despite desilting efforts, deposition of fluvial-transported fine-grained suspended solids (convective transport) at the sediment/water interface forms a thin, clay-like layer. This layer of accumulated fine-grained detrital sediment reduces the capacity of the basin to transmit water to the underlying aquifer, and is the primary contributor to percolation decay. Laboratory studies indicate that the majority (> 90%) of this “fouling layer” consists of inorganic solids less than 62.5 microns in apparent diameter (silt- to clay-sized particles). Removal of suspended solids in this size range by gravity settling is ineffective. Hypothesized is that loss of percolation is primarily a function of the accumulation of detritus. A log-decay model describing instantaneous percolation in terms of deposition of detrital sediments at the sediment/water interface using a single “sediment/foulant interaction coefficient” was developed. Integration of this relationship as a function of time yields percolation decay kinetics closely resembling those observed in full-scale recharge basins. The design of the study is to test this hypothesis by employing flow cells packed with sand and loaded with sediment-laden water as analogs of field-scale basins. Investigation of the effects on percolation decay kinetics and on the sediment/foulant interaction coefficient consisted of two parts: first, variations in concentration and particle size distribution of the suspended solids, and second, variations in particle-size distribution of basin sediments. The goal of this study is to validate the current percolation decay model and generalize it to accommodate a broader range of sediment and suspended solids compositions. A broadly applicable percolation decay model will be of value for rapid evaluation of potential basin performance optimization strategies and provide a tool for cost/benefit analyses for pre-treatment/desilting of recharge waters.
Article Download: 102908_PercDecay_UNESCOPaper
|
|
|
 |
|
|
|