Integrating Tunable Anion Exchange with Reverse ... - ACS Publications

Apr 3, 2015 - Department of Civil and Environmental Engineering, Lehigh University, ... recovery of the brackish water RO plant in El Paso, TX from...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/est

Integrating Tunable Anion Exchange with Reverse Osmosis for Enhanced Recovery During Inland Brackish Water Desalination Ryan C. Smith and Arup K. SenGupta* Department of Civil and Environmental Engineering, Lehigh University, Bethlehem, Pennsylvania 18015, United States S Supporting Information *

ABSTRACT: For inland brackish water desalination by reverse osmosis or RO, concentrate or reject disposal poses a major challenge. However, enhanced recovery and consequent reduction in the reject volume using RO processes is limited by the solubility of ions present in the feedwater. One of the most common and stubborn precipitate formed during desalination is calcium sulfate. Reducing or eliminating the presence of sulfate would allow the process to operate at higher recoveries without threat to membrane scaling. In this research, this goal is accomplished by using an appropriate mixture of self-regenerating anion exchange resins that selectively remove and replace sulfate by chloride prior to the RO unit. Most importantly, the mixed bed of anion exchange resins is self-regenerated with the reject brine from the RO process, thus requiring no addition of external chemicals. The current work demonstrates the reversibility of the hybrid ion exchange and RO (HIX-RO) process with 80% recovery for a brackish water composition representative of groundwater in San Joaquin Valley in California containing approximately 5200 mg/L of total dissolved solids or TDS. Consequently, the reject volume can be reduced by 50% without the threat of sulfate scaling and use of antiscaling chemicals can be eliminated altogether. By appropriately designing or tuning the mixed bed of anion exchange resins, the process can be extended to nearly any composition of brackish water for enhanced recovery and consequent reduction in the reject volume.

■

membrane scaling and fouling.5−11 The phenomenon of concentration polarization further aggravates the situation because the concentrations of precipitating electrolytes tend to be highest at the membrane−water interface.12 While the prevention of CaCO3 scaling is amenable through acid dosing, the same approach is not feasible for sulfate precipitation. Addition of antiscaling chemicals, that are normally salts of poly(acrylic acid) or polyphosphonic acid, are routinely administered into the feedwater to inhibit scaling of sulfate salts.13 One growing concern is that the deep well injection of reject is finding greater resistance and less acceptability when externally added chemicals are present. Ideally, high recovery is desirable without any externally added antiscaling chemicals. Since scaling and fouling caused by CaSO4 is more pervasive than other sparingly soluble inorganic salts, it is the primary focus of the study. However, the methodology proposed is amenable to application for other sparingly soluble salts including calcium phosphate or barium sulfate.

INTRODUCTION The application of inland brackish water desalination by reverse osmosis (RO) membrane processes is growing steadily around the world including the U.S.1 Besides high quality permeate, RO processes concurrently produce concentrated waste brine or reject that must be discarded. Unlike seawater desalination, inland plants are often too far from the coast to make seawater disposal of concentrate economically feasible. Consequently, disposal methods like deep well injection, evaporation ponds, or discharge to surface water are used.2 The Kay Bailey Hutchison Desalination Plant in El Paso, TX is currently the largest inland brackish water desalination plant in the world with 27.5 MGD treatment capacity. The concentrate or the reject from the plant is pumped 22 miles away from the plant for deep well injection.3 The costs associated with concentrate disposal from inland brackish water desalination processes often constitute more than 50% of the total operating cost.4 Reducing the volume of the concentrate produced would result in significant savings in disposal costs, but this can only be accomplished by operating the desalination plants at higher recoveries. Increasing the recovery of the RO process, even by a small percentage, would result in a large reduction in the volume of concentrate produced. For example, increasing the recovery of the brackish water RO plant in El Paso, TX from current 80−83% to 90% would result in nearly 50% decrease in the volume of concentrate produced. However, the recovery of brackish water desalination is limited by the solubility of commonly formed precipitates like CaSO4, CaCO3, or BaSO4 which, at high recoveries, tend to precipitate resulting in © XXXX American Chemical Society

■

CONCEPTUALIZED ENHANCED RO RECOVERY WITHOUT ANTI-SCALANT Limitations on Recovery Due to CaSO4 Precipitation. Due to the difficulty in controlling CaSO4 scaling, the Received: November 6, 2014 Revised: March 5, 2015 Accepted: April 3, 2015

A

DOI: 10.1021/es505439p Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

Should sulfate concentration in the feed be reduced by 80% and replaced by equivalent amount of chloride, percentage recovery can exceed well over 80% while the SI value for CaSO4 still remains under unity. Obviously, the challenge remains in sustaining such sulfate removal process without externally added chemicals. During the last five decades, many attempts have been made to reduce Ca2+ concentrations at the RO membrane through use of zeolite or polymeric cation exchange resins or using waste RO brine as an ion exchange regenerant.15−27 However, there is little variation in the types of cation exchange resins available. On the contrary, polymeric anion exchange resins have significantly more number of composition variables, namely, matrix, basicity, and size of the alkyl group in the functionality, to influence and alter divalent-monovalent anion selectivity.28−31 Therefore, anion exchange resins are wellsuited for the HIX-RO process due to the large selection of commercially available anion exchange resins. Concept of Hybrid Reversible Anion Exchange Process. Conceptually, a reversible anion exchange process, as illustrated in Figure 2, allows optimal sulfate-chloride exchange nearly eliminating any chance of sulfate supersaturation. Two trains of anion-exchange columns along with the RO membrane system constitute the core of the following process steps: Step 1: Incoming brackish water is passed through the first anion exchange column (no. 1) in chloride form and sulfate is removed by the following reaction:

desalination process recovery is kept below a critical point where sulfate scaling is thermodynamically favorable. This point is determined by calculating the dimensionless saturation index (SI) of calcium sulfate over a range of recoveries. SI is calculated from eq 1 below: SI =

{Ca 2 +}{SO24 −} K sp 2+

(1)

{SO42−}

Where {Ca } and indicate the activity of calcium and sulfate in the concentrate, and Ksp is the solubility product of CaSO4. When SI exceeds unity, the precipitation of CaSO4 is thermodynamically favorable. Figure 1 illustrates the gradual build-up and eventual supersaturation of CaSO4 with an increase in recovery for

2(R 4N+)Cl− + SO24 − → (R 4N+)2 SO24 − + 2Cl−

(2)

Overbar denotes the solid ion exchanger phase while R4N+ denotes anion exchanging functional group. Note that in order to achieve preferred removal of sulfate over chloride, the sulfate-chloride separation factor or αS/Cl needs to be greater than unity. Step 2: Now the effluent from the first anion exchange column, with a significant reduction in sulfate, is desalinated by RO at higher recovery without any threat to scaling. Step 3: The concentrated reject from RO, rich in chloride, is now used to regenerate the previously exhausted column (no. 2) to bring it back in chloride form:

Figure 1. Change in Saturation Index of CaSO4 with increasing recovery for a feed corresponding to San Joaquin Valley groundwater.

RO processes; the brackish water composition corresponds to that of San Joaquin Valley in California whose composition is shown in the inset of Figure 1.11 Saturation index values at each recovery were determined using OLI Stream Analyzer software, and detailed calculations are given in the Supporting Information Section S1.14 Note that at 50% recovery, the SI for CaSO4 is nearly equal to unity.

(R 4N+)2 SO24 − + 2Cl− → 2(R 4N+)Cl− + SO24 −

(3)

Figure 2. Depiction of the reversible anion exchange desalination process. B

DOI: 10.1021/es505439p Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology The composition of brine exiting from the column would be equivalent to the brine exiting the RO process if no ion exchange component were present. In order for this step to be thermodynamically favorable, αS/Cl is now required to be less than unity, that is, the same anion exchange resin would have to prefer chloride over sulfate. Note that Column no. 1 and Column no. 2 switch during the process from brackish water influent to RO reject and their sulfate/chloride selectivity is reversed, thus needing no external regenerant under proposed process conditions. Scientific Challenge and Design of Tunable Mixed Anion Exchanger. Brackish water undergoing RO treatment may vary in total electrolyte concentration (i.e., meq/L or mg/ L) widely depending on the location and/or source. In order to sustain the proposed process with higher permeate recovery but without introduction of any external regenerant, the sulfate− chloride separation factor must alter back and forth between the forward and regeneration cycle. The key scientific challenge lies in tailoring an anion exchange resin that satisfies this requirement. Previous studies have demonstrated that divalentmonovalent anion selectivity, namely, SO42−/Cl−, HPO42−/Cl−, SO42−/NO3−, and SO42−/HCrO4−, of a synthetic polymeric anion exchanger is influenced by polymer matrix, basicity, and steric property of the functional group.28−31 For any given brackish water composition and percentage permeate recovery of RO, it is possible to attain such a tunable anion exchanger, that is, αS/Cl > 1 during step 1 of the process and αS/Cl < 1 during step 3 through use of a mixture of anion exchange resins of different compositions. Assuming ideality, the equilibrium constant, K, for an ion exchange reaction is known as the selectivity coefficient, and for the reaction between sulfate and chloride, KS/Cl can be presented as KS/Cl =

yS* =

Q* = Q A ΦA + Q B(1 − ΦA )

Figure 3. Variation in α*S/Cl with mixing ratio of polystyrene and polyacrylic resins (The shaded region shows the region of reversibility for sulfate/chloride selectivity with 80% recovery).

(4)

resins for both feedwater and the reject. The following observations are noteworthy: when (i) fraction of polyacrylic resin is equal to one, that is, no polystyrene resin is present, * is greater than one for both the feedwater and the reject, αS/Cl and (ii) fraction is zero, that is, only polystyrene resin is present, α*S/Cl is less than one for both feed and the reject. However, for a mixed resin with polyacrylic fraction varying between 0.2 and 0.7, αS/Cl * switches from greater than one from the feed to less than one for reject corresponding to 80% recovery. The effect of both matrix and change in total electrolyte concentration (CT), also known as electroselectivity effect, contributes to the selectivity reversal between sulfate and chloride.32 The objective of the paper is to demonstrate through a series of laboratory experiments that for a representative brackish water composition corresponding to San Joaquin Valley brackish groundwater, higher recovery for RO processes is attainable through use of an appropriate 50/50 mixed bed of anion exchange resins while keeping the SI value of CaSO4 consistently below unity. Key specific goals are to validate that the process is reversible, sustainable, and no external addition of regenerant chemicals is necessary.

yS xCl xSyCl

(5)

Considering a mixture of two different resins “A” and “B” with masses mA and mB, the ratio of resin A, ΦA, is ΦA =

mA = 1 − ΦB mA + mB

(6)

And the overall column separation factor, α*S/Cl, may be calculated by * = αS/Cl

yS*xCl xSy* Cl

(9)

The Supporting Information, Section S2, provides details stepwise calculation of effective separation factor values for two resins with different mixing ratios at varying total electrolyte concentrations or CT. Figure S2, in the Supporting Information, gives a chart depicting how changing certain resin properties can also be a method of increasing or decreasing αS/Cl. Considering San Joaquin Valley brackish water with 80% RO permeate recovery and a mixture of two different strong base anion exchange resins with polystyrene and polyacrylic matrix, Figure 3 presents variation of αS/Cl * with mixing ratios of the two

Where x is the fraction of sulfate or chloride in solution, y is the corresponding fraction on the resin, CT is the total equivalent concentration of anions in meq/L, and Q is the resin capacity in meq/g; subscripts S and Cl are sulfate and chloride, respectively. The value of KS/Cl, however, does not provide insight as to whether sulfate is preferred over chloride or vice versa. Instead, the separation factor, αS/Cl, does and may be calculated from αS/Cl =

(8)

And the effective anion exchange capacity, Q*, is

2 yS xCl CT

xSyCl2 Q

ySA Q A ΦA + ySB Q B(1 − ΦA ) Q*

■

MATERIALS AND METHODS Resin Isotherms. Throughout the study, two strong base anion exchange resins with polystyrene and polyacrylic

(7)

Where C

DOI: 10.1021/es505439p Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

Figure 4. (A) Experimentally determined sulfate-chloride isotherms at 25 ± 2 °C for pure polyacrylic and polystyrene resins at 80 meq/L (B) Variation in α*S/Cl with mixing of resins (C) Experimental isotherm plots for ΦA = 0.5 at 80 meq/L and 400 meq/L.

at the exit of the concentrate side of the RO unit. Feedwater osmotic pressure was measured at 3.58 atm which closely matched the predicted value of 3.62 using OLI Analyzer.14 Typical transmembrane pressure during runs was 44−46 atm. Flow rate and flux data are given in the Supporting Information Section S3. During the last step of HIX-RO, the concentrate solution from RO was passed upflow through the ion exchange column with a 12 min EBCT. SEM-EDX analysis was performed using a Philips XL-30 environmental scanning electron microscope. Resin beads were sliced in half using a razor blade dipped in liquid nitrogen, mounted onto pegs using double sided carbon tape, and sputter coated with iridium using an Electron Microscopy Sciences high vacuum sputter coater (model EMS575X).

matricies were used. Individual resin properties are located in the Supporting Information Table S4. Before each isotherm, resins were placed in a glass column and a chloride solution was fed through the column until the effluent concentration was equal to the influent concentration. Resins were then washed with deionized water and left to airdry at room temperature. After drying, varying masses of resin were placed in separate Nalgene bottles and mixed with equal volumes of sulfate solution prepared from ACS grade sodium sulfate. Bottles were closed, sealed using Parafilm, and placed on a rotary mixer for at least 24 h. Upon reaching equilibrium, the concentration of sulfate and chloride was measured. All isotherms were carried out at 25 ± 2 °C. Chloride was analyzed by titration with silver nitrate in the presence of a potassium chromate indicator, and sulfate was analyzed by precipitation with barium chloride according to Standard Methods.33 HIX-RO Runs. Ten cycles of hybrid ion exchange-reverse osmosis (HIX-RO) were performed using either pure polystyrene resin or mixed polystyrene and polyacrylic resins. Before use in HIX-RO runs, all resin was washed with deionized water and packed in a custom-made 1L clear PVC column. Glass wool was placed at the top and bottom of the column to prevent loss of resin during operation. Resin was put in chloride form by passing a dilute solution of sodium chloride until effluent chloride concentration was the same as the influent. The column was washed again with deionized water until no chloride was detected in the effluent. No additional external regenerant or washing of the column was performed. An Accumet Conductivity Meter (model no. AP75) was used to measure conductivity, and osmotic pressure was measured using a Wescor Vapro5520 vapor pressure osmometer. For one cycle of HIX-RO, 20L of feed solution was passed down flow through the prepared ion exchange column using a peristaltic pump with an empty bed contact time (EBCT) of 6 min. Feed solution was prepared from ACS grade chemicals and deionized water. Effluent solution from the ion exchange column was collected and desalinated by reverse osmosis using the setup shown in Supporting Information Figure S3. Solution was fed by a stainless steel piston pump attached to a 1.5 hp electric motor. A prefilter (GE SmartWater GXWH20F) was placed before the RO membrane to prevent damage from any particulate matter that may be present in the feedwater. RO was performed using a spirally wound Dow Filmtec SW30−2540 RO membrane. Solution in the feed tank was kept at 20 °C ± 2 °C by submersing a cooling coil into the tank. Recovery of the RO process was adjusted to 80% by opening or closing a valve

■

RESULTS Isotherms for Parent and 50/50 Mixed Polyacrylic/ Polyacrylic Resins. An analysis for anion exchange resin properties in the previous section revealed that a 50/50 mixture of anion exchange resins with polyacrylic and polystyrene matrix with quaternary ammonium functional groups will allow reversal of sulfate/chloride selectivity for 80% recovery of brackish water equivalent to San Joaquin Valley water. Figure 4A shows sulfate-chloride isotherms at 80 meq/L total electrolyte concentration for two strong base resins with different matricies: polyacrylic and polystyrene. Note that while the sulfate-chloride separation factor, αS/Cl, is greater than unity for the polyacrylic matrix (i.e., the isotherm is above the diagonal), the same is less than unity for the polystyrene matrix, all other conditions remaining identical. Figure 4B shows how the α*S/Cl value of the mixed resin increases with an increase in the fraction of polyacrylic resin, ΦA; the solid line presents the predicted separation factor values assuming exchanger-phase ideality. Using the experimental data for 50/50 mixture of polyacrylic and polystyrene anion exchange resin (i.e., ΦA = 0.5), sulfate-chloride isotherm plots were developed for total electrolyte concentrations of both 80 meq/L and 400 meq/L, as illustrated in Figure 4C. Note that at 80 meq/L, corresponding to feed brackish water concentration, average sulfate/chloride separation factor is consistently greater than unity (average α*S/Cl = 1.6) while at 400 meq/L, corresponding * is to the reject concentration at 80% recovery, average αS/Cl * = 0.62). Neither a polystyrene lower than unity (average αS/Cl nor polyacrylic matrix alone can attain such favorable selectivity reversal between the feedwater and the reject. HIX-RO Runs with Mixed and Single Anion Exchange Resins. HIX-RO runs were conducted for two systems under D

DOI: 10.1021/es505439p Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology identical conditions excepting that while one system had a 50/ 50 mixture of two anion exchange resins with polyacrylic and polystyrene matrix, the other had only a strong-base anion exchange resin with polystyrene matrix. Ten consecutive cycles of HIX-RO runs were performed using the sulfate and chloride concentrations 60 meq/L Cl− and 20 meq/L SO42− identical to that in San Joaquin Valley water. One cycle consists of passage of feedwater through the first anion exchange column and then to the RO unit and finally exiting through the second column. Figure 5A depicts the sulfate concentrations in the feed and at the RO inlet after the mixed-bed anion exchange unit. Note

Figure 6. SI values for mixed bed and pure polystyrene bed.

In contrast, for the HIX-RO system using the anion exchange resin with polystyrene matrix alone, SI values remained consistently well above 1 for the entire period. Thus, the precipitation potential is very high and the addition of antiscaling agent is essential for the protection of the RO membrane. For comparison, the SI values were also computed for 80% recovery of the feedwater through RO without anion exchange and superimposed in Figure 6. Note that resin with polystyrene matrix had a minimal impact in reducing CaSO4 scaling propensity. Sulfate mass balance was carried out for both systems from cycle 5−10 to confirm that the steady states were fully attained and the sulfate concentration at RO inlet is unlikely to change for continued operation. Supporting Information, Figure S4A and B, demonstrate that the total milliequivalents of sulfate entering the treatment system remains practically equal to the sulfate leaving after HIX-RO process, confirming the attainment of the steady states for both operations. Evaluation of In-Column Preciptiation. The initial goal of the research was to prevent precipitation of CaSO4. Sulfate removed from the brackish feedwater during the forward pass prevents supersaturation and potential scaling of CaSO4 onto the RO membrane. During the reverse pass, sulfate is desorbed by RO reject and sulfate concentration in the exiting stream is significantly greater than that in the feed. Although this stream of concentrate goes only to waste and is never in direct contact with RO membrane, in-column precipitation is thermodynamically feasible because the saturation index is greater than unity. Following the procedure outlined in Supporting Information Section S4, HIX-RO runs were carried out using a small-scale ion exchange column using the same feedwater and mixed polystyrene and polyacrylic resins. Synthetic RO concentrate simulating the brine produced at 80% recovery was used for regeneration and was identical to that during HIX-RO runs with mixed resins. Three consecutive cycles were run and no precipitation was observed within the mixed-bed anion exchange column during the 10 min contact time. However, after nearly 120 min of settling, precipitates were formed in the tubes of the sample collector, as shown in Figure 7A. For more quantitative information, samples were collected from the effluent during the third cycle immediately after exiting and after visible precipitation had occurred. For both cases, calcium and sulfate were analyzed and the saturation index was calculated using OLI.14 Figure 7B provides the SI values plotted against bed volumes of reject regenerant passed for the two sets of samples. The absence of precipitation inside an ion exchange column under supersaturating conditions has been the subject of

Figure 5. (A) Sulfate concentration at inlet of RO for mixed-bed anion exchange (B) Sulfate concentration at inlet of RO for polystyrene anion exchange column.

that within four cycles, the HIX-RO system attained steady state and the sulfate concentration is consistently 90% lower than in the feed even under 80% recovery. Consequently, the propensity of scale formation at the membrane-water interface by sulfate salts is greatly reduced even in the absence of antiscaling agent. In contrast to Figure 5A, and B presents the results of 10 cycles using only the polystyrene anion exchange resin. It is quite apparent that the sulfate concentration at the RO inlet did not drop significantly and thus did not diminish scale forming potential of sulfate salts by calcium or other alkaline-earth metal cations. Effect on Saturation Index. The primary purpose of sulfate removal through reversible anion exchange is to eliminate the possibility of CaSO4 precipitation onto the RO membrane surface, that is, to maintain its SI value less than unity without warranting regeneration. Figure 6 shows experimentally determined SI values of the concentrates for both HIX-RO runs for complete ten cycles. Note that for the mixed bed of anion exchange resins with polyacrylic and polystyrene matrix, SI values are consistently well below unity. E

DOI: 10.1021/es505439p Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

the process, the anion exchange resins are continuously regenerated without requiring any external regenerant. Thus, the process is sustainable with no introduction of chemicals whatsoever. One salient finding of the investigation is that the use of only strong-base anion exchange resin with polystyrene matrix, as illustrated in Figure 6, completely failed to control precipitation of CaSO4, that is, SI was significantly greater than unity. Thus, it is not any anion exchange process but an appropriate mixture of anion exchange resins that are tunable to attain the reversibility of sulfate-chloride selectivity between the feedwater and the RO reject that forms the heart of the process. According to information in the open literature, the successful use of mixed anion exchange resins to prevent CaSO4 precipitation during RO process has not been reported to date. The approach is not just limited to sulfate fouling only. Attempts are underway to apply RO to recover and reuse treated municipal wastewater that often contains 2−5 mg/L of phosphate as P. Both calcium phosphate, pKsp Ca3(PO4)2 = 24.0, and calcium sulfate, pKsp = 4.59, are highly insoluble species and both are likely to precipitate and foul the RO membrane especially at high recoveries.35 For typical brackish groundwater, phosphate exists predominantly as a divalent anion, HPO42−, and hence, the same approach of using a mixed bed of anion exchange resins to attain a sustainable operation with higher permeate recovery can be achieved. Also, ligand exchangers with high phosphate selectivity may help attain the desired reversibility.36,37 Removal of silica is also a significant issue in brackish water desalination. However, at neutral pH, silica exists as nonionized hydrated silicic acid, H4SiO4 which is not amenable to removal by conventional ion exchange treatment. For high pressure steam generators in large thermal power stations, the HERO Process,38 is used to maximize silica removal in conjunction with ion exchange to avoid fouling of steam turbines. In cases where the feedwater has high bicarbonate or alkalinity, acid dosing may be prohibitively expensive. For these cases, excess bicarbonate may be removed by a stoichiometric amount of carbon dioxide or waste acid using weak acid cation exchange resins fibers.39 Anti-Scaling Chemicals and Deep Well Injection. Nearly every brackish water RO desalination plant resorts to dosing of antiscaling chemicals to sequester calcium in order to avoid sulfate precipitation and membrane fouling. Antiscaling chemicals are mostly phosphorus based salts and/or acrylic acid. For inland brackish water desalination, deep well injection is the most widely used methodology for disposal of the RO reject stream including that for the Kay Bailey Hutchison Plant in El Paso, Texas.3,40,41 Previous studies have shown that antiscaling chemicals promote the formation of biofilms in RO systems resulting in increased biofouling.42−44 Phosphorus based inhibitors, due to their known potential as a limiting nutrient responsible for eutrophication, are being currently avoided and gradually replaced by acrylic and polyacrylic acids.45 However, acrylic acids are biodegradable and in the presence of high sulfate concentration and naturally present sulfate-reducing bacteria in the reject may lead to odorous hydrogen sulfide or H2S. Retrofitting an existing RO plant with tunable and reversible mixed bed of anion exchange resins, may provide a holistic solution that will eliminate use of antiscaling chemicals altogether without sacrificing permeate recovery.

Figure 7. (A) Observations of the formation of CaSO4 precipitates in the reject solution after 10 and 120 min following exit from the column; (B) Calculated CaSO4 SI values immediately after exiting and after 24 h (C) SEM-EDX analysis of anion exchange resin beads for detection of sulfate, chloride, sodium, and calcium.

previous studies and is known as ion exchange induced supersaturation or IXISS.34 In order to further confirm the absence of precipitation inside the gel phase or pores of individual ion exchange beads, a few spherical resin beads were collected, sliced, and EDX characterization was performed. Figure 7C shows complete absence of calcium confirming no CaSO4 precipitation within the resin.

■

DISCUSSION Tunability of Anion Exchange Resins. Results of this research demonstrate that the propensity of CaSO4 precipitation on membrane surface can be eliminated by intelligently choosing a mixture of anion exchange resins. From an application viewpoint, this finding is particularly significant because such anion exchange resins are already commercially available. For nearly any brackish water TDS concentration and percentage recovery from RO, anion exchange resins of different matrix and functional groups with varying sulfate and chloride selectivity may attain the foregoing goal. During F

DOI: 10.1021/es505439p Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

■

(12) Antony, A.; Low, J. H.; Gray, S.; Childress, A. E.; Le-Clech, P.; Leslie, G. Scale formation and control in high pressure membrane water treatment systems: A review. J. Membr. Sci. 2011, 383, 1−16. (13) Halevy, S.; Korin, E.; Gilron, J.; Box, P. O. Kinetics of gypsum precipitation for designing interstage crystallizers for concentrate in high recovery reverse osmosis. Ind. Eng. Chem. Res. 2013, 52, 14647− 14657. (14) OLI Stream Analyzer, Version 9.0; OLI Systems, Inc.: Cedar Knolls, NJ, 2013. (15) Klein, G.; Villena-Blanco, M.; Vermeulen, T. Ion-exchange equilibrium data in the design of a cyclic sea water softening process. Ind. Eng. Chem. Process Des. Dev. 1964, 3 (3), 280−287. (16) Klein, G.; Cherney, S.; Ruddicks, E. L.; Vermeulen, T. Calcium removal from sea water by fixed-bed ion exchange. Desalination 1968, 4 (2), 158−166. (17) Hoek, C. V.; Kaakinen, J. W.; Haugseth, L. A. Ion exchange pretreatment using desalting plant concentrate for regeneration. Desalination 1976, 19 (1−3), 471−479. (18) Kaakinen, J. W.; Eisenhauer, R. J.; van Hoek, C. High recovery in the Yuma desalting plant. Desalination 1977, 23 (1−3), 357−366. (19) Wilf, M.; Konstantin, M.; Chencinsky, A. Evaluation of an ion exchange system regenerated with seawater for the increase of product recovery of reverse osmosis brackish water plant. Desalination 1980, 34 (3), 189−197. (20) Barba, D.; Brandani, V.; Foscolo, P. U. A method based on equilibrium theory for a correct choice of a cationic resin in sea water softening. Desalination 1983, 48 (2), 133−146. (21) Vermeulen, T.; Tleimat, B. W.; Klein, G. Ion-exchange pretreatment for scale prevention in desalting systems. Desalination 1983, 47 (1−3), 149−159. (22) Shain, P.; Klein, G.; Vermeulen, T. A mathematical model of the cyclic operation of desalination-feedwater softening by ion-exchange with fluidized-bed regeneration. Desalination 1988, 69 (2), 135−146. (23) Muraviev, D.; Khamizov, R.Kh.; Tikhonov, N. A.; Morales, J. G. Clean (“green”) ion-exchange technologies. 4. High-Ca-selectivity ionexchange material for self-sustaining decalcification of mineralized waters process. Ind. Eng. Chem. Res. 2004, 43, 1868−1874. (24) Tokmachev, M. G.; Tikhonov, N. A.; Khamizov, R.K. Investigation of cyclic self-sustaining ion exchange process for softening water solutions on the basis of mathematical modeling. React. Funct. Polym. 2008, 68, 1245−1252. (25) Venkatesan, A.; Wankat, P. C. Desalination of the Colorado River water: A hybrid approach. Desalination 2012, 286, 176−186. (26) Abdulgader, H. Al; Kochkodan, V.; Hilal, N. Hybrid ion exchangePressure driven membrane processes in water treatment: A review. Sep. Purif. Technol. 2013, 116, 253−264. (27) Hanegbi, Y.; Mansdorf, Y. Status of Palmachim Desalination Project. In Program and Papers of 8th Annual Conference, Proceedings of the Innovations and Applications of Sea-Water and Marginal Water Desalination, Technion Institute of Technology, Haifa, Israel, 2006; Semiat, R.; Hasson, D., Eds.; Israel Desalination Society: Israel, 2006; Abstract 1.2, pp 11−16. (28) Clifford, D.; Weber, W. J., Jr. The determinants of divalent/ monovalent selectivity in anion exchangers. React. Polym. 1983, 1, 77− 89. (29) Sarkar, S.; SenGupta, A. K. A new hybrid ion exchangenanofiltration (HIX-NF) separation process for energy-efficient desalination: Process concept and laboratory evaluation. J. Membr. Sci. 2008, 324, 76−84. (30) SenGupta, A. K.; Roy, T.; Jessen, D. Modified Anion Exchange resin for improved chromate selectivity and increased efficiency of regeneration. React. Polym. 1988, 9, 293−299. (31) Guter, G. A. Removal of nitrates from contaminated water supplies using a tributyl amine strong base anion exchange resin. U.S. Patent, No. 4,479,877, Oct. 30, 1984. (32) Li, P.; SenGupta, A. K. Genesis of selectivity and reversibility for sorption of synthetic aromatic anions onto polymeric sorbents. Environ. Sci. Technol. 1998, 32, 3756−3766.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information provides details regarding the construction of Figure 1 (Section S1), stepwise calculation of mixing ratio for ion exchange resins (Section S2), individual resin properties (Table S4), controlling αS/Cl by changing resin properties (Figure S2), characterization of RO system (Section S3), setup for RO (Figure S3), mass balance on sulfate for both HIX-RO runs (Figure S4A and S4B), breakthrough curve for 50/50 resin mixture with no regeneration (Figure S5), and the detailed procedure for the in-column precipitation study (Section S5). This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Phone: (610) 758-3534; e-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This research received funding from the National Science Foundation Accelerating Innovation Research (NSF-AIR #1311758) Grant and the Pennsylvania Infrastructure Technology Alliance (PITA). A few insightful comments from reviewers helped refine a few premises of the paper including inclusion of a few relevant references.

■

REFERENCES

(1) Thompson, J.; Rahardianto, A.; Gu, H.; Uchymiak, M.; Bartman, A.; Hedrick, M.; Lara, D.; Cooper, J.; Faria, J.; Christofides, P. D.; Cohen, Y. Rapid field assessment of RO desalination of brackish agricultural drainage water. Water Res. 2013, 47, 2649−2660. (2) Xu, P.; Cath, T. Y.; Robertson, A. P.; Reinhard, M.; Leckie, J. O.; Drewes, J. E. Critical review of desalination concentrate management, treatment and beneficial use. Environ. Eng. Sci. 2013, 30, 502−514. (3) Hutchison, W. R. Deep-well injection of desalination concentrate in El Paso, Texas. Southwest Hydrology, March/April 2008, 28−30. (4) National Research Council. Desalination: A National Perspective; The National Academies Press: Washington, D.C., 2008. (5) Shih, W.; Rahardianto, A.; Lee, R.; Cohen, Y. Morphometric characterization of calcium sulfate dehydrate (gypsum) scale on reverse osmosis membranes. J. Membr. Sci. 2005, 252 (1−2), 253−263. (6) Gabelich, C. J.; Williams, M. D.; Rahardianto, A.; Franklin, J. C.; Cohen, Y. High-recovery reverse osmosis desalination using intermediate chemical demineralization. J. Membr. Sci. 2007, 301 (1−2), 131−141. (7) Tzotzi, Ch.; Pahiadaki, T.; Yiantsios, S. G.; Karabelas, A. J.; Andritsos, N. A study of CaCO3 scale formation and inhibition in RO and NF membrane processes. J. Membr. Sci. 2007, 296 (1−2), 171− 184. (8) Sheikholeslami, R. Composite scale formation and assessment by the theoretical Scaling Potential Index (SPI) proposed previously for a single salt. Desalination 2011, 278, 259−267. (9) Waly, T.; Kennedy, M. D.; Witkamp, G.; Amy, G.; Schippers, J. C. The role of inorganic ions in the calcium carbonate scaling of seawater reverse osmosis systems. Desalination 2012, 284, 279−287. (10) Hasson, D.; Drak, A.; Semiat, R. Inception of CaSO4 scaling on RO membranes at various water recovery levels. Desalination 2001, 139, 73−81. (11) Rahardianto, A.; Shih, W.; Lee, R.; Cohen, Y. Diagnostic characterization of gypsum scale formation and control in RO membrane desalination of brackish water. J. Membr. Sci. 2006, 279 (1−2), 655−668. G

DOI: 10.1021/es505439p Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology (33) Standard Methods for the Examination of Water and Wastewater, 18th ed.; American Public Health Association: Washington, DC, 1992. (34) Muraviev, D. N.; Khamizov, R. Ion-exchange isothermal supersaturation: Concept, problems, and applications. In Ion Exchange and Solvent Extraction; SenGupta, A. K., Marcus, Y., Eds.; Marcel Dekker, Inc.: New York, 2004; Vol. 16, pp 119−210. (35) Snoeyink, V. L.; Jenkins, D. Water Chemistry; John Wiley & Sons: New York, 1980. (36) Blaney, L.; Cinar, S.; Sengupta, a. Hybrid anion exchanger for trace phosphate removal from water and wastewater. Water Res. 2007, 41, 1603−1613. (37) Zhu, Y.; SenGupta, A. K. Sorption enhancement of some hydrophilic organic solutes through polymeric ligand exchange. Environ. Sci. Technol. 1992, 26 (10), 1990−1998. (38) Mukhopadhyay, D. Method and apparatus for high efficiency reverse osmosis operation. U.S. Patent, No. 6,537,456, Mar. 25, 2003. (39) Greenleaf, J. E.; SenGupta, A. K. Environmentally benign hardness removal using ion-exchange fibers and snowmelt. Environ. Sci. Technol. 2006, 40, 370−376. (40) Greenlee, L. F.; Testa, F.; Lawler, D. F.; Freeman, B. D.; Moulin, P. Effect of antiscalant degradation on salt precipitation and solid/ liquid separation of RO concentrate. J. Membr. Sci. 2011, 366, 48−61. (41) Smith, R. C. Integrating tunable anion exchange with reverse osmosis for enhanced recovery during inland brackish water desalination. Ph.D. Dissertation, Lehigh University, Bethlehem, PA, 2015. (42) Vrouwenvelder, J. S.; Manolarakis, S. A.; Veenendaal, H. R.; van der Kooij, D. Biofouling potential of chemicals used for scale control in RO and NF membranes. Desalination 2000, 132 (1−3), 1−10. (43) Vrouwenvelder, J. S.; Manolarakis, S. A.; van der Hoek, J. P.; van Paassen, J. A. M.; van der Meer, W. G. J.; van Agtmaal, J. M. C.; Prummel, H. D. M.; Kruithof, J. C.; van Loosdrecht, M. C. M. Quantitative biofouling diagnosis in full scale nanofiltration and reverse osmosis installations. Water Res. 2008, 42 (19), 4856−4868. (44) Sweity, A.; Oren, Y.; Ronen, Z.; Herzberg, M. The influence of antiscalants on biofouling of RO membranes in seawater desalination. Water Res. 2013, 47, 3389−3398. (45) Yuchi, A.; Gotoh, Y.; Itoh, S. Potentiometry of effective concentration of polyacrylate as scale inhibitor. Anal. Chim. Acta 2007, 594, 199−203.

H

DOI: 10.1021/es505439p Environ. Sci. Technol. XXXX, XXX, XXX−XXX