Search

Warm lime softening is commonly used to reduce hardness, silica, and a small fraction of organic matter from steam-assisted gravity drainage (SAGD) produced water through the addition of lime, soda ash, MgO, coagulant and flocculant. We report a systematic study on the impact of solution chemistry on the epichlorohydrin-dimethylamine coagulant demand for the treatment of synthetic SAGD produced water. Concentrations of magnesium, calcium, sodium bicarbonate, clay (mimicking suspended solids), sodium metasilicate (representing silica), and humic acid (mimicking dissolved organic matter) were varied to study their impact on coagulant demand. The impact of the concentration of lime, soda ash, and MgO on coagulant demand was also studied. Within the studied concentration range, the coagulant dose increased linearly with increasing concentration of humic acid (Ycoagulant = 29 + 0.703XHA) and silica (Ycoagulant = 52 + 0.537Xsilica), and increased slightly with increasing concentration of lime and soda ash, but remained almost unchanged with increasing concentration of dissolved hardness, clay, or MgO. The observations were correlated to the understanding of the electrokinetic properties of CaCO3 and Mg(OH)2 particles in lime softening. The findings provide insights for evaluating onsite coagulant dose and optimizing the process.

Significant effort has been made to measure and understand the surface charge of CaCO3 and Mg(OH)2 particles. Many laboratory experiments have been reported on zeta potential of natural and prepared CaCO3 and a few have also been published for Mg(OH)2, however, there are very few reported measurements of zeta potential of CaCO3 and Mg(OH)2 particles at conditions relevant to lime softening, despite lime softening being a common and established process for water treatment. The present study aims to understand the interactions and electrokinetic properties of these two particles in lime softening. Effects of various experimental parameters such as pH, temperature, aging, inorganic carbon (CO32−/HCO3−), and divalent cations (Ca2+/Mg2+) on the electrokinetic properties (i.e. zeta potential) of CaCO3 and Mg(OH)2 particles were individually studied. The interactions between humic acid (mimicking natural organic matter), silicate (representing silica), clay (mimicking suspended solids) and CaCO3/Mg(OH)2 particles were studied, as well as the interactions between CaCO3 and Mg(OH)2. Thermodynamic modeling was used to predict precipitates as a function of solution chemistry and assist with data interpretation. The results provide considerable insight into factors that are of importance to lime softening.

This study used computer simulation to select appropriate process designs to make coal seam gas (CSG) associated water comply with beneficial reuse regulations. The hypothesis was that computational simulation may predict the unit operations required for treatment of CSG associated water over a wide range of salinities. AqMB software facilitated complete analysis of water treatment from initial storage pond to final pH and sodium adsorption ratio (SAR) adjustment. Three samples of CSG associated water ranging from 3650 to 22123 mg/L total dissolved solids content were evaluated. Common to each process design was a settling pond, coarse & fine filtration, softening, anti-scalant addition, and reverse osmosis. When the water hardness was very hard, then a lime softening system may be required. Both microfiltration and ultrafiltration were considered, and in both cases the level of suspended solids was reduced to close to zero. Sodium exchanged WAC resin removed the majority of alkaline earth ions and inhibited downstream scale formation. The presence of substantial alkalinity was addressed by use of acid exchanged WAC resin in combination with an air stripper. However, process economics indicated this option was not favourable. A two-stage Reverse Osmosis system not only recovered from ca. 74–79 % water with all CSG associated water types but also the water quality was compliant with beneficial reuse regulations. The computer predictions mirrored actual data from operating industry sites.

One of the most direct means for human exposure to nanoparticles (NPs) released into the environment is drinking water. Therefore, it is critical to understand the occurrence and fate of NPs in drinking water systems. The objectives of this study were to develop rapid and reliable analytical methods and apply them to investigate the fate and transportation of NPs during drinking water treatments. Rapid single particle ICP-MS (SP-ICP-MS) methods were developed to characterize and quantify titanium-containing, titanium dioxide, silver, and gold NP concentration, size, size distribution, and dissolved metal element concentration in surface water and treated drinking water. The effectiveness of conventional drinking water treatments (including lime softening, alum coagulation, filtration, and disinfection) to remove NPs from surface water was evaluated using six-gang stirrer jar test simulations. The selected NPs were nearly completely (97 ± 3%) removed after lime softening and alum coagulation/activated carbon adsorption treatments. Additionally, source and drinking waters from three large drinking water treatment facilities utilizing similar treatments with the simulation test were collected and analyzed by the SP-ICP-MS methods. Ti-containing particles and dissolved Ti were present in the river water samples, but Ag and Au were not present. Treatments used at each drinking water treatment facility effectively removed over 93% of the Ti-containing particles and dissolved Ti from the source water.

Phosphorus (P), calcium (Ca) and natural organic matter (NOM) naturally occur in all aquatic ecosystems. However, excessive P loads can cause eutrophic or hyper-eutrophic conditions in these waters. As a result, P regulation is important for these impaired aquatic systems, and Ca-P co-precipitation is a vital mechanism of natural P removal in many alkaline systems, such as the Florida Everglades. The interaction of P, Ca, and NOM is also an important factor in lime softening and corrosion control, both critical processes of drinking water treatment. Determining the role of NOM in Ca-P co-precipitation is important for identifying mechanisms that may limit P removal in both natural and engineered systems. The main goal of this research is to assess the role of NOM in inhibiting Ca and P co-precipitation by: (1) measuring how Ca, NOM, and P concentrations affect NOM's potential inhibition of co-precipitation; (2) determining the effect of pH; and (3) evaluating the precipitated solids. Results showed that Ca-P co-precipitation occurs at pH 9.5 in the presence of high natural organic matter (NOM) (≈30 mg L(-1)). The supersaturation of calcite overcomes the inhibitory effect of NOM seen at lower pH values. Higher initial P concentrations lead to both higher P precipitation rates and densities of P on the calcite surface. The maximum surface density of co-precipitated P on the precipitated calcite surface increases with increasing NOM levels, suggesting that NOM does prevent the co-precipitation of Ca and P.

An experimental study was carried out to demonstrate and quantify the feasibility of antiscalant (AS) removal from brackish water RO concentrate of high gypsum scaling propensity via lime treatment prior to seeded gypsum precipitation. Based on studies with model solutions, it was shown that sufficient AS removal (up to ∼90%) from RO concentrate is feasible via a lime treatment step (at a dose significantly lower than that required for conventional lime softening) to enable effective subsequent seeded gypsum precipitation. This two-step chemically-enhanced seeded precipitation (CESP) treatment of primary RO concentrate is suitable as an intermediate concentrate demineralization (ICD) stage for high recovery desalting employing secondary RO desalination. Analysis of gypsum precipitation and lime treatment kinetic data suggests that, after adequate CaCO 3 precipitation has been induced for effective AS scavenging, CaSO 4 desupersaturation can be achieved via seeded gypsum precipitation without retardation due to seed poisoning by AS. Also, the lime dose required to prevent seed poisoning during subsequent gypsum desupersaturation via seeded gypsum precipitation can be adequately assessed with a precipitation kinetics model that considers AS seed poisoning based on a Langmuir adsorption isotherm. The degree of AS removal after lime treatment increased linearly with the logarithm of the single lime dose additions. Staged lime dosing (i.e., multiple lime additions), however, removed a higher degree of AS relative to an equivalent single lime dose addition since a higher driving force for CaCO 3 precipitation could be maintained over the course of the lime treatment period.

Dissolved organic nitrogen (DON) derived from wastewater effluent can participate in reactions that lead to formation of nitrogenous chlorination by-products, membrane fouling, eutrophication, and nitrification issues, so management of DON is important for both wastewater reuse applications and nutrient-sensitive watersheds that receive discharges from treated wastewater. This study documents DON occurrence in full-scale water/wastewater (W/WW) treatment plant effluents and assesses the removal of wastewater-derived DON by several processes (biodegradation, coagulation, softening, and powdered activated carbon [PAC] adsorption) used for advanced treatment in wastewater reuse applications. After varying levels of wastewater treatment, the dominant aqueous nitrogenous species shifts from ammonia to nitrate after aerobic processes and nitrate to DON in tertiary treatment effluents. The fraction of DON in total dissolved nitrogen (TDN) accounts for at most 52% in tertiary treated effluents (median=13%) and 54% in surface waters impacted by upstream wastewater discharges (median=31%). The 5-day biodegradability/bioavailability of DON (39%) was higher, on average, than that of dissolved organic carbon (DOC, 26%); however, upon chlorination, the DON removal (3%) decreased significantly. Alum coagulation (with ≥8 mg/L alum per mg/L DOC) and lime softening (with pH 11.3-11.5) removed

Advances in the application of ion exchange membrane processes for material separation provide new approaches for improving the sustainability of industrial operations. In this work, we report on the application of Donnan dialysis to recover metal cations from lime softening sludge, a byproduct of drinking water production. A cation exchange membrane, which preferentially allows the transport of cations, was used in the Donnan cell. Slurries containing lime softening sludge, gypsum, or calcite were placed on the feed side of the cell and an acidic solution was placed on the sweep side. A concentration gradient between the two sides of the cell extracts metal cations from the feed in exchange for hydrogen ions in the sweep side solution. The influence of acid type, starting pH, and time were evaluated. Although the limited solubility of calcite inhibited its study, significant calcium ion recovery was measured from the feed solutions with gypsum. The results were consistent with theoretical expectations and demonstrated that the pH of the sweep side solution controlled the extent and rate of cation recovery. For the treatment of the lime softening sludge, over 99% of the hydrogen ions were exchanged in the recovery of up to 20% of the calcium and 50% of the magnesium from the lime softening sludge. Over time, depletion of the sweep-side hydrogen ions resulted in the preferential recovery of magnesium from the softening sludge. Our results suggest the process exhibits promise in treating the sludge by recovering metal cations for reuse.

Process design for a 15 mgd (2370 m 3 /h) reverse osmosis (RO) plant in El Paso, Texas, in the desert area of the US southwest presents many challenges. High recoveries in the pilot plant were limited by silica and barium sulfate scaling. Following testing at 85–90% recovery using a silica polymerization inhibitor and antiscalant, the disposal of a projected 2.1 mgd (332 m 3 /h) of an unstable supersaturated RO concentrate presents a significant problem. While deepwell and evaporative disposal alternatives are being considered, a pilot program towards a zero liquid discharge process to recover more water and perhaps useful by-products has progressed with very encouraging results. We report here our initial results, attaining a further recovery of 70% of water, and the production from a lime-softening step of sludge with high silica content, and promise as a useful road-base and embankment materials additive.

In order to improve the removal efficiency of phosphate and mitigate membrane fouling, lime softening as a pretreatment for ceramic membrane microfiltration of phosphorus-containing wastewater was investigated. The effects of pretreatment on filtration flux, particle size, permeate quality, membrane fouling characteristic, and the recovery of membrane permeate flux after cleaning were studied. At the lime dosage of 680mg·L−1, the removal of phosphate increased from 11.0% without pretreatment to 99.7%, and the permeate flux was about 60% greater than that when lime was not used. The permeate obtained from softening and microfiltration met the National Discharge Standard and could be reused under very safe conditions. Pretreatment increased particle size, which mitigated membrane fouling by reducing the extent of pore plugging and forming a more porous cake on the membrane surface. The result was validated by ESEM micrographs. With the same cleaning method, the recuperation of permeate flux is approximately 18% greater with lime pretreatment than without pretreatment, which may be due to the differences in morphology of the deposited cake and the degree of irreversible fouling.

In this research, means to increase the removal of natural organic matter (NOM) during drinking water treatment were investigated. Specifically, the modification of calcium carbonate surfaces by cationic polyelectrolytes was tested as a way to improve humic acid adsorption to floc surfaces during lime softening. Results demonstrated that under high pH conditions (pH 9.5), coating the particles with high charge density cationic polyacrylamide (CPAM) significantly increased the adsorption of humic acid, while coating the particles with low charge density CPAM had no positive effect. The amount of humic acid adsorbed also depended on the molecular weight of CPAM, with high molecular weight coatings being more effective at enhancing adsorption than low molecular weight coatings. Electrophoretic mobility measurements demonstrated that the cationic polymer reduced repulsive electrostatic interactions between humic acid and the particle surface. Under low pH conditions (pH 7.5), adsorption of polyacrylamide was low, and coating the particles had less effect on the amount of humic acid adsorbed. The results presented here suggest that the modification of calcium carbonate floc surfaces using cationic polymers with high charge density may enhance natural organic matter removal during drinking water treatment.

In a previous paper the CAPS softening process is suggested as a pretreatment step for RO: in addition to removal of calcium it is an effective filtration and may thus replace part of the customary pretreatment. In the present study the removal of silica and heavy metals in addition to water softening was investigated. Precipitation with both sodium hydroxide and with lime was studied. With lime precipitation in CAPS, it was found that magnesium is necessary for silica removal, if no other additives such as Al and Zn are used. With conventional methods of precipitation, the well-known reagents for silica removal, aluminium salts, lead to very slow precipitation or difficult filtration. In contrast, aluminium chloride used as an additive in CAPS can remove a major part of the silica without interfering with filtration, if the precipitation conditions are chosen correctly; optimal conditions were in keeping with the regime of softening plants. It was found that zinc chloride, which is environmentally more acceptable than the aluminium salt, could also be used for silica removal, but in larger amounts than aluminium chloride. With these two salts, a substantial fraction of the silica can be removed, but the precipitation of silica results in decreased calcium removal. To achieve simultaneous softening and silica removal, it is necessary to add carbonate. Heavy metals are coprecipitated with the calcium carbonate.

Reverse osmosis (RO) and physical/chemical technology were evaluated for treatment of neutralized spent acid effluent (seepage) containing high concentration levels of TDS (7500 mg/l), Ca (400 mg/l), CrVI (42 mg/l), nitrate-nitrogen (827 mg/l), ammonia-nitrogen (33 mg/l), fluoride (13 mg/l), phenolics (45 mg/l) and COD (620 mg/l). The calcium concentration level in the seepage could be reduced from approximately 400 to 5 mg/l with soda ash softening. Initial permeate flux (feed and bleed system, 85% water recovery) was 278 l/m2·d. Permeate flux, however, dropped rapidly in the beginning of the run and then remained approximately constant to the end of the run. However, chemical cleaning of the membranes was necessary to maintain flux. The TDS of the RO feed could be reduced in one case from 34,253 to 1560 mg/l (95.5% removal) at 85% water recovery. Nitrate and ammonia nitrogen were reduced from 2691 and 103 mg/l to 414 (84.6% removal) and 15 mg/l (85.3% removal), respectively. ChromiumVI and fluoride were reduced from 183 and 90 mg/l to 0.38 (99.8% removal) and 2.8 mg/l (96.9% removal), respectively. COD removals varied between 60 and 80%. No phenolics, however, could be removed from the feed (approximately 32 mg/l) with the cellulose acetate RO membranes. Phenolics, however, could be effectively removed (

With recognizing that particles represent transport vehicles for undesirable chemical contaminants, and potentially disease causing microbial pathogens, the removal of particle materials becomes important to protect the public health. However, due to the complicated filtration mechanisms and interactions between the polymer and particles, the selection of polymers currently remains empirical. This study was initiated to investigate the feasibility of the use of polymers as filter aids for water containing lime softening particles; explore the roles of polymer characteristics in the filtration process; and examine major factors affecting the performance of polymer aided filtration. Seven representative polymers with different molecular weights and charges were tested using a declining rate filter pilot plant. The results show that at a starting flow rate of 15 m/h, the removal of turbidity particles could be significantly improved by using polymers as a filter aid. It was also found that the impacts of initial filter ripening could be substantially reduced and no turbidity breakthrough was observed after 3 days of operation. However, the use of polymer might significantly increase the filtration headloss, especially for the polymers with high molecular weight. To produce a high quality filtrate while ensuring the acceptable filtration productivity, low or moderately low molecular weight polymers are recommended. For a low molecular weight polymer, its optimum mixing intensity and polymer dose were found around 700 s−1 and 0.01 mg/l, respectively. On the basis of the results and particle properties, it is believed that interparticle bridging is the dominant mechanism underlying the interactions between polymers and lime softening particles.

In 1988 the City of Hollywood, Florida prepared a 10-year water supply and treatment plan as a condition for renewal of its permit for consumptive use of the Biscayne Aquifer as raw supply to the City's 37.5 mgd lime softening plant. The three-phase plan which was adopted involves staged replacement of obsolescent softeners and filters with RO membranes, using both the Biscayne and the deeper, better contained, but brackish Floridan Aquifer as raw water supplies. Operating and blending protocols had to be developed which would ensure at all times during and after project implementation: Satisfaction of public water demand; water quality in compliance, present and future; Eastern Biscayne Aquifer withdrawal at or below current permitted amounts, with eventual curtailment; and residuals and concentrate disposal effected in an environmentally acceptable manner. These operating constraints, plus the unusual requirements for blending control make this project an important and instructive prototype for our changing water utility industry.

Increasingly, nanofiltration is being used to soften water in Florida — particularly for applications where the feed water has high color and dissolved organics (which are precursors to regulated disinfectant byproducts). The slightly higher costs for membrane softening than lime softening, are more than justified for many utilities that desire the generally superior product water quality. These utilities anticipate the ability to meet future drinking water standards, even those not yet identified, without the need to modify the process or add additional processes required for lime softening.

In polyelectrolyte-enhanced ultrafiltration (PEUF), a water-soluble anionic polyelectrolyte (in this study sodium polystyrene sulfonate or PSS) is added to hard water. The calcium and magnesium bind to the polymer which has a high enough molecular weight to be rejected by an ultrafiltration membrane. The permeate is softened water. Economically, the PSS needs to be recovered from the retentate for reuse. Three methods of recovery developed in this study were addition of NaCl, Na2CO3 or HCl to PEUF to regenerate PSS. Of the three PEUF processes considered, NaCl/PEUF as compared to Na2CO3/PEUF and HCl/PEUF provided the best scheme for the water softening process. PEUF is shown in this study to be competitive with lime softening at low flow rates. The PEUF process is more expensive than ion exchange for a stream containing only hardness ions. However, PEUF becomes nearly comparable with ion exchange for a stream containing hardness ions as well as bacteria, viruses and pyrogen. The cost comparisons are based on fully continuous operations and include treatment of waste streams from each process.

Experiments were carried out for pretreatment of feed water for RO by adsorption of colloidal matter on Mg(OH) 2 (or hydrated MgO) formed as well as comparative experiments for brackish water pretreatment by: (1) coagulation and flocculation; (2) lime softening, coagulation and flocculation; (3) lime softening and adsorption on the Mg(OH) 2 formed; (4) lime-alkaline softening and adsorption on the Mg(OH) 2 formed; (5) dolomitic lime softening and adsorption with hydrated MgO formed. It was established that colloidal matter in brackish lake water is sharply reduced only by adsorption on the Mg(OH) 2 (or hydrated MgO) formed in sufficient quantity and the pretreated feed water meets the requirements of the spiral wound and hollow fibre membranes. Optimum results for pretreatment of the tested water by adsorption on the Mg(OH) 2 formed and maximum rate of precipitation of the CaCO 3 and Mg(OH) 2 are achieved when the ratio CaCO 3 :Mg(OH) 2 is 3:2. It was found that technologically and economically the most advantageous method for brackish lake water pretreatment is dolomitic lime softening, where dolomite dosage of 450 mg/l and a reaction time of more than 1 h is needed in order to reach heterogeneous equilibrium.

In 1976, a Florida-based equipment manufacturer conceived the idea of using modified reverse osmosis membrane instead of lime softening to treat waters with low total dissolved solids. The growth of the application was slow at first, but in 1989, this blossomed to over 125 mgd of facilities planned, designed or being built in Florida. Membrane softening (MS) is a desalination process since it removes dissolved inorganics as well as organics from the feed water and the product water complies with present SDWA regulations. MS is particularly cost effective in treating ground water in Florida which is high in hardness and natural organics such as humic and fulvic acids. The MS process is similar to conventional reverse osmosis (RO) except three stages are employed for 90% recovery, operation is at 90 to 120 psi with a booster pump on the third stage; and pretreatment is greatly simplified. Capital costs for MS facilities are less than $0.90 per USgpd capacity (over 30 USmgd capacity) and O&M costs are in the order of $0.40 per 1,000 US gallons. Utilizing blending of waters treated by MS and other techniques further enhances the application of MS. Desalination by MS of waters below 1,000 ppm TDS extends the economical treatability range of desalination for drinking water supplies. It is a technology whose time has come.

The molecular weight distributions (MWDs) of several commercially prepared humic and fulvic acids and organic matter present in natural waters were analyzed by gel chromatography. The responses of these substances to treatment by alum coagulation, lime softening, and activated carbon adsorption were also analyzed, as were their trihalomethane formation potentials before and after each treatment. The treatability characteristics and trihalomethane formation properties of the different organic sub- stances were then related to their respective MWDs. Differences in the MWDs were found to effect differences in the behavior of lumped parameter measures of organic matter (such as TOC) with respect to the several treatment operations. The treatments in turn were observed to alter the MWDs of the organic substances as well as their phenomenological behavior in subsequent process operations.

The 96 million gallon per day Yuma desalting plant is part of a complex project intended to solve a long standing water quality problem between the United States and Mexico. Designed to treat up to 129,000 acre-feet of 3200 mg/l irrigation return flows, the plant will deliver 283 mg/l product water to maintain a 115 mg/l differential between water delivered to U.S. users and to Mexico . Design of the plant was preceded by an extensive research program to determine pretreatment requirements, membrane performance, plant capacity and equipment and instrumentation performance. Two pretreatment systems were tested. Partial lime softening followed by multi-media filtration was chosen for the final design. Nine membrane systems (both reverse osmosis and electrodialysis) were tested. Two spiral wound reverse osmosis systems were chosen as most efficient from physical and economic viewpoint, and contracts for site preparation and construction of the intakes and sedimentation basin is scheduled for award summer 1979 with construction of the main plant and installation of desalting equipment scheduled for award summer of 1980. Based on January 1979 prices, total cost of the desalting complex is approximately $190 million (including interest during construction). The investment cost per daily gallon of installed capacity is approximately $2. per 1000 gallons of desalted water (capital amortization plus operation and maintenance) is $0.77.

The removal of organic contaminants by lime softening and molecular characteristics influencing removal were experimentally and theoretically examined. Coprecipitation appeared to be the major removal mechanism for organic substances in low concentrations, and only those substances able to adsorb onto CaCO3 were able to be removed by lime softening. All of the removable anionic compounds were able to coordinate with Ca2+ to form relatively insoluble complexes. Fatty acids, which form relatively soluble complexes with Ca2+, and hydrophobic chemicals unable to coordinate with Ca2+ were not removed. Molecular characteristics influencing removal included molecular charge, functionality, degree of polymerization, molecular geometry and affinity for water. In general, softening is effective in removing polymeric electrolytes possessing acidic oxygen-containing residues, such as carboxyl, phenol, enol, phosphoryl, phosphonyl, sulfuryl and sulfonyl groups. Simple monomeric molecules are not expected to be effectively removed unless they polymerize or possess phosphorus-containing functional groups such as phosphoryl and phosphonyl.

The City of Ft. Myers currently line softens up to 10 MGD (1600 m3/hr) of raw water from a well field in a process including aeration, lime softening, clarification, recarbonation, chlorination, and media filtration. The water, as presently treated, exceeds the regulatory limit for THM by a factor of four (400 ug/L versus regulatory limit of 100 ug/L). Due to a need to expand capacity and comply with THM regulations, the City investigated process alternatives, and found that membrane separation of organics (THM precursor) and hardness, more reliably meets existing requirements and proposed future THM limits, at no additional cost compared to more conventional chemical processes. Subsequent to the treatment study, membrane pilot testing was conducted, a membrane supplier was selected, and proof testing was successfully completed during the summer of 1988. This paper describes the project background, and membrane proof test objectives, procedures, equipment, and results.

The properties and the conditions of formation of a metastable hydrated form of calcium carbonate (CaCO36H2O) are described. This compound often crystallises when Essex river waters are softened by the lime/soda process. Formation of the hydrated salt is favoured by low water temperatures and by high orthophosphate levels. Calcium carbonate hexahydrate forms pure well-defined crystals which are considerably more soluble in water than calcite, the usual product of lime/soda softening. The efficiency of the softening process is very much reduced when the hydrated compound crystallises, and increased quantities of chemicals are required both for softening and for subsequent stabilisation of the softened water. Considerable success has been achieved on the practical scale in reducing the incidence of formation of the hexahydrate by precipitating phosphate before softening using ferric salts and by addition of pre-formed calcite nuclei in the form of commercial Whiting; this has led to substantial economies in chemical consumption.

Organic and inorganic fouling problems associated with reverse osmosis treatment of municipal wastewater were studied at the Advanced Wastewater Treatment facilities at Hoover Dam. This study indicated that the severity of organic fouling can be reduced by improving the quality of the RO feedwater; namely, by removal of food sources for bacterial growth including BOD, TOC, and ammonia. This, besides the normal secondary treatment, required lime softening and filtration. Lime softening and filtration alone, without secondary treatment, were not found to be effective. Chlorination is necessary to discourage bacterial and fungal growth. Where dechlorination is required to protect the membrane, in the case of polyamide membranes for example, sodium bisulfite was found to be superior to sodium thiosulfate. Use of ultraviolet radiation showed promise as an alternate to chlorination for RO feedwater disinfection. Inorganic fouling can be controlled by pretreatment processes including softening and removal of iron and manganese, and by reducing water recovery.

The 363 400 m 3 /d capacity Yuma Desalting Plant has been designed to operate using membrane desalting at about 70 percent product water recovery with the remaining 30 percent being wasted reject brine. An engineering feasibility study was undertaken to identify the best methods of achieving higher recoveries and thus of reducing the brine volume. Results indicate that 90 percent recovery can be attained most economically by increasing the membrane area by roughly 50 percent through additional desalting equipment and by better pretreatment removal of calcium and silica through use of partial lime softening- clarification followed by additional ion-exchange softening. Reject desalting brine would regenerate the ion exchangers. A test program is underway intended to develop this approach and to provide the required design data should it be decided later to alter the Yuma Desalting Plant for high recovery.