Humic Acid Fouling Mitigation by Antiscalant in ... - ACS Publications

Corresponding author e-mail: [email protected]; phone: 0086-21-54748942; fax: 0086-21-54748942., †. Shanghai Jiao Tong University. , ‡. Chinese Aca...
3 downloads 3 Views 1MB Size
Download PDF
Environ. Sci. Technol. 2010, 44, 5153–5158

Humic Acid Fouling Mitigation by Antiscalant in Reverse Osmosis System Q I N G F E N G Y A N G , * ,† Y A N G Q I A O L I U , ‡ AND YAJUAN LI† School of Environmental Science and Engineering, Shanghai Jiao Tong University, Shanghai 200240, China, and State Key Laboratory of High Performance Ceramics and Superfine Microstructures, Shanghai Institute of Ceramics, Chinese Academy of Sciences, Shanghai, 200050, China

Received February 16, 2010. Revised manuscript received May 24, 2010. Accepted May 25, 2010.

The mitigation of humic acid (HA) fouling on reverse osmosis (RO) membrane was investigated by using an environmentally benign antiscalant, polyaspartic acid (PASP). In the presence of PASP, HA fouling decreased with increasing Ca2+ concentration. This positive effect of Ca2+ was not due to the electrostatic repulsion as measured by zeta potential, but probably due to the formation of a stable water-soluble complex HA-Ca-PASP through Ca2+ bridging. Fouling inhibition efficiency φ increased with increasing PASP concentration, but overdosing could lead to an adverse effect. At higher feedwater pH, HA fouling was alleviated and the φ was slightly improved. HA fouling increased when increasing initial permeate flux and decreasing cross-flow velocity, but the fouling behaviors became less susceptible to the two hydrodynamic parameters in the presence of PASP. The φ was hardly affected by the initial permeate flux and the cross-flow velocity. HA fouling decreased with decreasing feed temperature in the presence of PASP, likely owing to the improved stability of HA-Ca-PASP at lower temperature. The implication of this paper is that it presented an attractive and feasible approach for organic fouling control in RO system by dosing antiscalant.

Introduction Reduction of fouling by natural organic matter (NOM), such as humic acid (HA), is one of the major challenges in membrane filtration (1). Various physical and chemical influence factors have been studied on membrane fouling by HA (2–7). Severe flux reduction may occur at low pH, high ionic strength, high calcium concentration, low feedwater temperature, high permeation rate, and low cross-flow rate (2–8). Flux decline at low pH and high ionic strength is primarily due to reduced electrostatic repulsion between HA molecules and that between HA and membrane surfaces. On the other hand, hydrodynamic conditions affect fouling mainly through increased drag force, reduced shear, or reduced turbulence. Specific cake resistance of HA fouling layer increases at low temperature, leading to an increase in flux decline (9). Among chemical factors, the presence of calcium ions has a marked effect on HA fouling. Calcium, which is a major * Corresponding author e-mail: [email protected]; phone: 008621-54748942; fax: 0086-21-54748942. † Shanghai Jiao Tong University. ‡ Chinese Academy of Sciences. 10.1021/es100513e

 2010 American Chemical Society

Published on Web 06/07/2010

divalent cation in natural and waste waters, has been known to interact and form complexes with carboxylic functional groups of organic molecules. Its presence was found to aggravate organic fouling phenomenon supposedly via charge neutralization, complexation, and formation of calcium bridges (2–6, 10–12). Selective removal of calcium ions via pretreatment was suggested to reduce the adverse effect of calcium (13, 14). Besides the pretreatment of feedwater and the control of operation parameters, membrane surface modification (15–18) is another way to mitigate organic fouling, which involves material synthesis. The modified membranes demonstrated more hydrophilic and smoother surfaces, showing significantly lower fouling rates (15, 16). Antiscalants, such as polyelectrolytes and organophosphorus compounds, are widely used now and known to be very effective in controlling inorganic scales such as CaCO3 and CaSO4 formed on heat transfer and RO membrane surfaces (19, 20). These compounds inhibit scale nucleation and crystallization processes. Antiscalant dosing is an attractive option because it has the combinative advantages of operational and capital cost reduction, environmental acceptability, and safety when compared to the alternative technologies. However, there are seldom studies on the mitigation of organic fouling on RO membrane surface by using antiscalant. One available report shows that certain commercially available antiscalants and dispersants increased the rate of membrane fouling by HA (21). Some manufacturers declared their proprietary polymeric antiscalant could control organic fouling in membrane systems, but no further actual application report is available. The objectives of this study were (a) to find an environmentally benign antiscalant to mitigate HA fouling on RO membrane surface effectively and (b) to reveal fouling mitigation mechanisms by the antiscalant. A systematic investigation was carried out for the effect of antiscalant polyaspartic acid (PASP) on RO membrane fouling by HA. Bench-scale fouling tests were performed under various chemical and physical conditions in the presence of PASP. The mechanisms of fouling mitigation by antiscalant were studied by zeta potential measurements, particle size measurements, and various permeation flux decline curves.

Materials and Methods Organic Foulant of Humic Acid. Aldrich humic acid (SigmaAldrich, H16752, technical grade, St. Louis, MO) was pretreated following the method described in ref 6 to remove fulvic, metal, and ash contents. Stock solutions of 2 g/L were prepared from the freeze-dried HA by using distilled (DI) water, stored in the dark at 4 °C, and discarded after 3 months. Feed solutions for fouling experiments were prepared from the stock solution. RO Membrane. A polyamide thin-film composite RO membrane of TW30 was used, which was provided by AMFOR Inc. According to the catalogue specifications, TW30 is a high flow, moderate chlorine rejection (98%) membrane, operating at ultra low pressures. The maximum applied pressure and maximum working temperature for the membrane are 21 bar and 45 °C, respectively. The membrane has an operational pH range of 2-11, and was supplied and stored as dry coupon. Solution Chemistry. Unless otherwise specified, all reagents and chemicals were analytical grade. Sodium chloride, calcium chloride, disodium ethylenediaminetetraacetate (Na2-EDTA), sodium hydroxide, and hydrochloric VOL. 44, NO. 13, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5153

acid were purchased from Sinopharm Chemical Reagent Co., Ltd. DI water was used to prepare fouling test solutions. A constant HA concentration of 20 mg/L was used in all experiments involving organic foulant. The concentration was confirmed by measuring the total organic carbon (TOC) of the feed solution with a TOC analyzer (Shimadzu Corporation, Kyoto, Japan). For all fouling tests, the total solution ionic strength, I, was maintained at 10 mM by adjusting NaCl concentration. PASP is a nontoxic, highly biodegradable, and watersoluble antiscalant (22). The compound consists of polymerized R- and β-aspartyl residues, each containing a carboxylic functional group that can combine with H+ or metal ions to form different H-PASP or metal-PASP species at different pHs. PASP is fully deprotonated at pH 7 (23). PASP (sodium salt), provided by Lebond Chemicals, China, was adopted. According to the product specification, it is a brown clear liquid, and has a solid content of 40%, density of 1.23 g/cm3, pH of 9.5 ( 1.0, and molecular weight of 4000-9000. In this study, the concentrations of PASP were given in terms of the as-received brown clear liquid, not in terms of its active material (20). The antiscalant concentrations used for the test solutions were 2-50 mg/L, which corresponds to 0.064-1.6 mg/L equivalent TOC (A sample solution of 800 mg/L PASP was analyzed and the corresponding TOC was 25.78 mg/L). Zetal Potential and Particle Size Measurements. Zeta potential of the membrane surface was determined by streaming potential measurements conducted with an electrokinetic analyzer (Brookhaven Instruments Corp.). Zetal potential values of foulants were measured with a zeta potential analyzer (Brookhaven Instruments Corp.). Zeta cells were washed three times with DI water and two times with sample prior to measurement. Reported values were the average of five replicates. Particle sizes of HA in the absence and presence of PASP were analyzed with a ZetaPlus particle size analyzer (Brookhaven Instruments Corp.) under the same solution conditions used in the fouling experiments. Membrane Fouling Tests. Fouling experiments were carried out in a bench-scale cross-flow RO membrane test system (see Figure S1, Supporting Information). The system contained a test cell at applied pressures up to 15 bar. The test cell was made from two Plexiglas plates with a 2-cmthick quartz glass window fixed to the top plate. Double “Orings” were used to provide a leak-proof seal. Two metallic plates sandwiched the top and bottom Plexiglas plates. Fourteen stainless steel threaded bolts with threaded nuts compressed the metal plates, which provided extra stability for the double O-ring seal. The dimensions of the cross-flow channel were 2 mm height, 9.5 cm width, and 14.6 cm length. A membrane sample was placed face up over the top of the permeate spacer that lay over the permeate collection port in the bottom plate. Membrane coupons (14.6 × 9.5 cm) were thoroughly rinsed with DI water, and soaked in DI water for 24 h, before being loaded into the membrane testing cell. A feed spacer 0.787 mm thick was also placed in the test cell. Solution temperature, Tb, was maintained constant (at 20 or 30 °C) during the fouling runs by a recirculating heater/ chiller (DC0506, Yueping Instruments Corp.) with a stainless steel coil submerged in the feed reservoir. The feed solution was fed to the test cell by a diaphragm pump (GMB78-24, Qian Engineering Inc.). The cross-flow velocity was controlled by a bypass valve, and the trans-membrane pressure was regulated by a backpressure regulator. To avoid concentration of organics in the system, both the concentrate and the permeate were recirculated to the feed reservoir. Feed solution flow rate and membrane permeate flux were measured by rotameters. The concentration of free Ca2+ was measured by an ionometer (PXS-270, LeiCi Instruments Corp.) with a calcium ionic electrode. Total calcium was 5154

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 13, 2010

monitored by EDTA titration. Conductivity, pH, and turbidity measurements were also made during fouling experiments to monitor salt rejection, and to ensure that the chemical conditions were constant throughout tests. Membrane fouling experiments were conducted in three stepsscompacting with DI water, conditioning with foulantfree electrolyte solution, and fouling following the method described in ref 5. The membrane was first compacted with DI water in the membrane test cell for 12 h. Then, the membrane was stabilized and equilibrated for an additional 6 h with foulant-free electrolyte solution having solution chemistry identical to that used for the subsequent fouling run. After attaining a stable permeate flux, antiscalant was added to the feed solution for the experiments to test fouling inhibition effect. Then, the initial flux, J0, and the cross-flow velocity, ub, were adjusted to 12-18 µm/s and 0.21-0.3 m/s, respectively. After a 2-min sufficient mixing of antiscalant with feedwater through agitation, organic fouling was then initiated by adding HA stock solution to the feedwater to achieve a 20 mg/L concentration of HA foulant. Permeate flux was continuously monitored for the next 10 h. During fouling tests, the applied pressure was kept constant. All fouling experiments were conducted at least twice. The variations between flux data of replicate experiments were found to be 1-2.5%.

Results and Discussion Inhibition Effect of Antiscalant on HA Fouling. To investigate the inhibition effect of PASP, 5 mg/L PASP was dosed to the feed solution for the fouling experiments conducted at the same initial permeate flux of 13.6 µm/s. As comparison, a strong chelating agent EDTA of 1 mM was also used in the fouling inhibition test. The running results are shown in Figure 1. It can be seen that the HA fouling on RO membrane surface could be mitigated by dosing 5 mg/L PASP. To discuss the inhibition effect better, an index of inhibition efficiency, φ, was adopted. The values of the normalized flux, J/J0, at the end of 10-h fouling experiments were used to calculate φ. The choice of a 10-h fouling period is arbitrary. Although the use of a different fouling period would lead to a change of the absolute φ, it would not affect the findings on the effects of calcium ion concentration, antiscalant concentration, feed solution pH, hydrodynamic conditions, and feed solution temperature on fouling phenomena, which will be described below (19). The φ is defined as φ)

(J/J0)inhibition - (J/J0)blank × 100% 1 - (J/J0)blank

(1)

where (J/J0)blank and (J/J0)inhibition are the normalized flux values in the absence and presence of antiscalant, respectively. The value of φ is in the range of 0-100% when (J/J0)inhibition changes from (J/J0)blank to 1. The permeate flux decline at the end of experiment and the calculated φ for Figure 1a,b are demonstrated in Figure 1c. In the absence of PASP, the permeate flux decline increases from 14.5% to 16.2% and then to 24.4% as Ca2+ concentration increases from 0 to 1 mM and then to 2 mM. Such fouling trend is consistent with that in other studies (2–6). The aggravated fouling phenomenon in the presence of Ca2+ can be explained by charge neutralization, complexation, and formation of calcium bridges. However, the situation is reversed when PASP is present. As Ca2+ concentration increases from 0 to 1 mM and then to 2 mM in the presence of 5 mg/L PASP, the flux decline at the end of experiment is reduced from 12% to 10.5% and then to 8.8%, while the corresponding φ also increases from 17% to 35% then to 64%. This result demonstrates that in the presence of PASP, the presence of Ca2+ within a certain concentration range could improve the fouling inhibition ability. The

FIGURE 1. Effect of PASP on HA fouling behavior: (a) with 1-2 mM Ca2+, (b) with 0-1 mM Ca2+, (c) permeate flux decline at the end of experiment and inhibition efficiency of HA fouling; φ was calculated by using J/J0 value at the end of experiment. Other experimental conditions: initial permeate flux J0 ) 13.6 µm/s, solution temperature Tb ) 30 °C, solution pH ) 7.0, total ionic strength I ) 10 mM, cross-flow velocity ub ) 21 cm/s. phenomenon was also observed by addition of 1 mM Ca2+ to individual effluent organic matter (EfOM) fractions (24). They attributed this observation to the possibilities of different composition of the real organic components, and the formation of organic aggregates easily carried away by shear or changes in hydrophobicity of organic foulants. Fouling reduction was also found with fatty acids (octanoic acids) owing to the decreased organic hydrophobicity after calcium addition (25). The addition of 1 mM EDTA also presents fouling inhibition effect. The φ value for EDTA is 54%, which is lower than that of 64% with 5 mg/L PASP under identical conditions. Therefore, it is more cost affordable using antiscalant PASP than EDTA. It is well-known that HA can form HA-Ca complex with Ca2+, and therefore effectively reduces the negative charge of HA. Owing to the reduced interchain electrostatic repulsion by sequestering Ca2+, HA molecules form a small, coiled conformation, and subsequently a more compact fouling layer (26). Bridging between HA macromolecules mediated by calcium complexation may also contribute to the formation of a dense fouling layer (2). For PASP molecule, both carboxylic acid (-COOH) and imino (-NH-) groups are

contained, which can provide lone pair electrons and form PASP-Ca complex with Ca2+ (23). Through the formation of such water-soluble complex, inorganic scales can be prevented by dosing trace amounts (several mg/L) of PASP (22). Calcium has the ability to bridge several HA molecules together (12), while one PASP molecule can sequester several calcium ions according to scale inhibition mechanisms (22, 27). Therefore, when PASP is added to solution containing HA and Ca2+, an intermolecular complex of HA-Ca-PASP may form via Ca2+ ion bridging. Due to the excellent water solubility of PASP (sodium salt) in feedwater at pH 7, such intermolecular complex may have a strong interaction with water molecule, i.e., higher water solubility and does not easily deposit on RO membrane surface compared to HACa complex. In the absence of Ca2+, HA fouling could also be mitigated by PASP, as demonstrated in Figure 1b. However, the φ value of 17% is lower than 35% and 64% in the presence of 1 and 2 mM Ca2+, respectively. The interaction between PASP and HA may be via neutral functional groups such as -NH in PASP molecules (28). Owing to the lack of Ca2+ ion bridging, such interaction may be weak. At higher Ca2+ concentration, more HA molecules might be bridged by Ca2+ through possibly formed HA-Ca-PASP. This may be the reason HA fouling was reduced more efficiently when Ca2+ concentration increased, as demonstrated in Figure 1c. Complexation of Calcium Ions by HA and PASP. To test the complexation ability of HA and PASP, two experiments with different addition sequences for HA and PASP were carried out by measuring free Ca2+ concentration in fouling solution. In the first experiment, the following operating steps were adopted for preparing a fouling solution in a 1000-mL beaker: addition of 1 mM CaCl2, 5 mg/L PASP, and 20 mg/L HA, and then adjusting solution pH from 7 to 11. In the second experiment, the operating steps were changed to the following: addition of 1 mM CaCl2, 20 mg/L HA and 5 mg/L PASP, and then adjusting solution pH from 7 to 11. The time interval was 20 min between each step, and strong solution stirring was achieved with a magnetic stirrer. The results are shown in Figure 2a. It can be seen that 20 mg/L HA and 5 mg/L PASP has similar ability to sequester Ca2+. Furthermore, no difference is present for the different addition sequences of HA and PASP. The decline of free Ca2+ concentration is 10%, 17%, and 26% at each operating step. Although PASP could reduce Ca2+ concentration by complexation, this action might be not the main mechanism for the mitigation of HA fouling, since an increase in Ca2+ concentration could increase φ, as illustrated in Figure 1. Monitoring of free Ca2+ and total calcium during fouling tests in both the absence and presence of PASP indicated that the free Ca2+ concentration hardly decreased, but total calcium decreased. This fact demonstrates that the fouling mitigation is realized by reducing the deposition of complexed calcium, likely by HA-Ca-PASP action. Therefore, HA fouling mitigation mechanisms between EDTA and antiscalant PASP may be different. EDTA realizes HA fouling reduction by removing free and HA-complexed calcium ions (2) whereas PASP does so probably by forming stable water-soluble complex of HA-Ca-PASP. As solution pH increases from 7 to 11, no deprotonation occurs for PASP since PASP is fully deprotonated at pH 7 (23), but more functional groups of HA become deprotonated with increasing pH (5, 6). Thus, the increased number of negatively charged functional groups of HA could offer more binding sites to Ca2+, leading to a decrease in free Ca2+ concentration. Zetal Potential and Particle Size Analyses. Zeta potential of the membrane surface measured was -20.6 mV under conditions of pH 7, temperature 30 °C, and 10 mM NaCl. However, further zeta potential measurements indicated VOL. 44, NO. 13, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5155

FIGURE 2. Fouling soution tests: (a) complexation of calcium ions by HA and PASP, (b) zeta potential of test solutions.

FIGURE 3. Effect of PASP concentration: (a) normalized flux versus time, (b) inhibition efficiency versus PASP concentration, φ was calculated by using J/J0 value at the end of experiment. Other experimental conditions: Tb ) 30 °C, pH ) 7.0, I ) 10 mM, ub ) 21 cm/s. that drastic modifications of the initial physicochemical properties of the membrane surface occurred when the membrane surface contacted the fouling solutions. Thus, foulant-deposited-foulant interactions replaced the initial foulant-membrane interactions for the subsequent fouling processes. Since membrane fouling happened immediately when HA was introduced to the feedwater in the present study, foulant-deposited-foulant interactions had dominant role for the whole fouling processes. Zeta potential of solution systems used for fouling experiments in Figure 1b was measured, and the result is illustrated in Figure 2b. The dosage of 5 mg/L PASP slightly increases the zeta potential in both the absence (HA compared to HA+PASP) and presence (HA+Ca2+ compared to HA+Ca2++PASP) of 1 mM Ca2+. On the other hand, the presence of 1 mM Ca2+ greatly reduces the zeta potential in both the absence (HA compared to HA+Ca2+) and presence (HA+PASP compared to HA+Ca2++PASP) of 5 mg/L PASP. As shown in Figure 1b, the least flux decline occurs for the solution system HA+Ca2++PASP. The severity of flux decline for the four solution systems increases in the following order: HA+Ca2++PASP, HA+PASP, HA, and HA+Ca2+. The corresponding zeta potential values are -32.6, -52.5, -50.1, and -28.4 mV, respectively. It can be seen that the solution system of HA+Ca2++PASP with lower negative charge (zeta potential -32.6 mV) has less flux decline than that of HA+PASP with the highest negative charge (zeta potential -52.5 mV). This implies that in the presence of antiscalant, the reduction of HA fouling by increasing Ca2+ is not due to the electrostatic repulsion but may be due to the formation of water-soluble complex HA-Ca-PASP. In another study (16), membrane fouling by sodium alginate (NaAlg) was enhanced by addition of bovine serum albumin (BSA). Particle size analysis of the foulant solutions by dynamic light scattering revealed that interaction be5156

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 13, 2010

tween BSA and NaAlg led to formation of large aggregates. The mean foulant particle diameter increased from 523 to 742 nm when BSA was added to the NaAlg solution. In this study, the mean foulant particle diameter measured by ZetaPlus particle analyzer decreased from 2636 to 483 nm when 5 mg/L PASP was added to the HA+Ca2+ solution. Although the measured foulant sizes may significantly deviate from the actual sizes as explained by the literature (16), the result indicates that PASP could reduce the aggregation of HA-Ca complex. The fouling mitigation effect by PASP might mainly be attributed to the stability of the formed watersoluble complex rather than the increased negative charge (only 4.2 mV negative potential increase from HA+Ca2+ to HA+Ca2++PASP systems). In the complex structure of HA-Ca, Ca may be seen as an active site, through which the complex interacts with bulk and deposited HA, ultimately aggravating membrane fouling. PASP can probably block the active site by complexation and thus mitigate the membrane fouling. Effect of PASP Concentration. The effect of PASP concentration is presented in Figure 3. The result shows that the fouling φ increases proportionally with PASP concentration from 2 to 10 mg/L, and reaches 91% efficiency at 10 mg/L PASP. This indicates that more HA-Ca complexes have been stabilized by the increased PASP molecules. However, further increasing PASP concentration to 50 mg/L results in a great reduction of φ to 35%. Therefore, PASP would lead to an adverse effect when overdosing. This may be explained by the decreased water solubility of HA-CaPASP at high antiscalant concentration (29). Hence, there is an optimal concentration for antiscalant to achieve a better fouling mitigation effect. Effect of Feed Solution pH. The effect of feed solution pH on HA fouling was studied with two different pH values of 7 and 10 (see Figure S2, Supporting Information). The

FIGURE 5. Effect of feed solution temperature. φ was calculated by using J/J0 value at the end of experiment. Other experimental conditions: pH ) 7.0, I ) 10 mM, ub ) 21 cm/s.

FIGURE 4. Effect of hydrodynamic conditions: (a) effect of initial permeate flux, (b) permeate flux decline versus initial flux, φ was calculated by using J/J0 value from (a) at the same cumulative permeate volume of 5.51 L, (c) effect of cross-flow velocity, φ was calculated by using J/J0 value at the end of experiment. Other experimental conditions: Tb ) 30 °C, pH ) 7.0, I ) 10 mM. result demonstrates that fouling is alleviated at higher pH in both the absence and presence of PASP. Moreover, the φ of 54.7% at pH 10 is slightly higher than that of 50.7% at pH 7. This behavior may be explained by the increased electrostatic repulsion due to the charge increase of humic macromolecules and the membrane surface at higher solution pH. Effect of Initial Permeate Flux. The effect of initial permeate flux on HA fouling is presented in Figure 4a,b. The flux decline behavior is plotted in the way of normalized flux vs cumulative permeate volume, rather than time. This is more appropriate when comparing fouling runs performed at different initial fluxes (2, 3). All fouling experiments were performed at the same solution temperature 30 °C. Under this condition, a greater initial flux was achieved by using a higher trans-membrane pressure. In the absence of PASP, a greater flux decline is observed for a higher initial flux, which is consistent with results in the literature (2, 3, 6). This fouling behavior is attributed to the increased permeation drag overcoming the electrostatic repulsion between HA and membrane initially and that between HA and deposited-HA subsequently, increased concentration polarization, and more compact fouling layer at higher applied pressures when the initial flux increases (2, 3, 6, 30). In the presence of 5 mg/L PASP, a greater flux decline is also observed for a higher initial flux, as illustrated in Figure

4a. However, the fouling behaviors become less susceptible to initial flux in the presence of PASP, which can be seen from the slope of the regression lines, k, as shown in Figure 4b. The four data points of permeate flux decline in Figure 4b were taken from fouling curves in Figure 4a at the same cumulative permeate volume of 5.51 L. The k value is 1.07 in the absence of PASP, whereas it decreases to 0.30 in the presence of PASP. The values of φ, calculated by using the four data points, are 50.7% and 52.7% for the experiments performed at initial flux of 12 and 18 µm/s, respectively. Therefore, φ is hardly affected by the variation of initial flux. Effect of Cross-Flow Velocity. The effect of cross-flow velocity on HA fouling is illustrated in Figure 4c. All fouling experiments were performed at the same initial permeate flux of 12 µm/s. Two cross-flow velocities of 21 and 30 cm/s were adopted, and the corresponding Reynolds numbers Re are 245 and 342, respectively (31). The results clearly show that HA fouling is lessened by increasing the cross-flow velocity regardless of the presence of antiscalant. The effect of cross-flow velocity on HA fouling is attributed to the increase in the shear rate and the resulting reduction in HA and calcium concentrations at the membrane surface (3). The fouling behaviors with PASP become less susceptible to cross-flow velocity compared to those without PASP. This can be known from the slope of the regression lines, k (see Figure S3, Supporting Information). The k value is -0.39 in the absence of PASP, whereas it changes to -0.16 in the presence of PASP. The values of φ are 50.6% and 48.2% for the experiments performed at the cross-flow velocity of 21 and 30 cm/s, respectively. This indicates that φ is almost not affected by the cross-flow velocity. Effect of Feed Solution Temperature. The effect of feed solution temperature on organic fouling of RO membrane was less studied in the literature. The factor was investigated in this paper at temperatures of 20 and 30 °C. The applied pressure was adjusted to maintain a constant initial permeate flux 12 µm/s at the two different temperatures. Higher feed temperature results in lower permeate flux decline in the absence of PASP, as shown in Figure 5. The result is in agreement with another study (9), and can be explained by a reduced specific cake resistance as feed temperature increases. The lower applied pressure at higher feed temperature would result in less physical compaction of the HA deposits, which could lead to lower specific cake resistance. However, in the presence of antiscalant, the effect of feed temperature was found to be opposite that in the absence of antiscalant. As shown in Figure 5, the permeate flux decline increases when the feed temperature increases in the presence of 5 mg/L PASP. In the presence of PASP, the operation condition of higher applied pressure at lower feed temperature is similar to that in the absence of PASP, but VOL. 44, NO. 13, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5157

lower feed temperature does not result in higher permeate flux decline. This fact indicates that the key difference in the HA cake layer at different temperatures in the presence of PASP may be the accumulation rate rather than the specific resistance. The above result demonstrates that the modified solution chemistry by antiscalant has greater influence on RO membrane fouling than feed temperature. The values of φ are 64.2% and 56.7% for the experiments performed at the feed temperature of 20 and 30 °C, respectively. The result demonstrates that the φ decreases with an increase in feed temperature and HA was more efficiently stabilized by PASP at lower temperature. This may be due to higher stability constant for HA-Ca-PASP at lower temperature, which would cause a lower deposited mass of HA foulant on RO membrane surface. The implication of this paper is that it presents an attractive and feasible approach for organic fouling control in RO system by dosing antiscalant. The organic fouling may be mitigated under appropriate conditions such as optimal antiscalant dosage, solution pH, calcium concentration, permeate flux, cross-flow velocity, and bulk temperature.

(10) (11) (12) (13)

(14) (15) (16)

(17)

Acknowledgments This work was financially supported by the National Natural Science Foundation of China (20306015, 20676077), Shanghai Education Committee (09ZZ18), Science and Technology Commission of Shanghai Municipality (08QA14073), and the Shanghai Institute of Ceramics (SCX200709).

(18)

(19)

Supporting Information Available Schematic diagram of the bench-scale RO membrane test system (Figure S1); effect of feed solution pH (Figure S2); final permeate flux decline versus cross-flow velocity (Figure S3). This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Tang, C. Y.; Kwon, Y.-N.; Leckie, J. O. Characterization of humic acid fouled reverse osmosis and nanofiltration membranes by transmission electron microscopy and streaming potential measurements. Environ. Sci. Technol. 2007, 41, 942–949. (2) Hong, S.; Elimelech, M. Chemical and physical aspects of natural organic matter (NOM) fouling of nanofiltration membranes. J. Membr. Sci. 1997, 132, 159–181. (3) Seidel, A.; Elimelech, M. Coupling between chemical and physical interactions in natural organic matter (NOM) fouling of nanofiltration membranes: implications for fouling control. J. Membr. Sci. 2002, 203, 245–255. (4) Zularisam, A. W.; Ismail, A. F.; Salim, R. Behaviours of natural organic matter in membrane filtration for surface water treatment s a review. Desalination 2006, 194, 211–231. (5) Lee, S.; Emelimelech, M. Relating organic fouling of reverse osmosis membranes to intermolecular adhesion forces. Environ. Sci. Technol. 2006, 40, 980–987. (6) Tang, C. Y.; Kwon, Y.-N.; Leckie, J. O. Fouling of reverse osmosis and nanofiltration membranes by humic acidsEffects of solution composition and hydrodynamic conditions. J. Membr. Sci. 2007, 290, 86–94. (7) Violleau, D.; Essis-Tome, H.; Habarou, H.; Croue´, J. P.; Pontie´, M. Fouling studies of a polyamide nanofiltration membrane by selected natural organic matter: an analytical approach. Desalination 2005, 173, 223–238. (8) Bowen, W. R.; Mongruel, A.; Williams, P. M. Prediction of the rate of cross-flow membrane ultrafiltration: A colloidal interaction approach. Chem. Eng. Sci. 1996, 51, 4321–4333. (9) Jin, X.; Jawor, A.; Kim, S.; Hoek, E. M. V. Effects of feed water temperature on separation performance and organic fouling of

5158

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 13, 2010

(20) (21) (22) (23) (24)

(25) (26) (27) (28) (29) (30) (31)

brackish water RO membranes. Desalination 2009, 239, 346– 359. Lee, S.; Ang, W. S.; Elimelech, M. Fouling of reverse osmosis membranes by hydrophilic organic matter: implications for water reuse. Desalination 2006, 187, 313–321. Ang, W. S.; Elimelech, M. Protein (BSA) fouling of reverse osmosis membranes: Implications for wastewater reclamation. J. Membr. Sci. 2007, 296, 83–92. Stevenson, F. J. Humus Chemistry; John Wiley and Sons: New York, 1982. Emongor, V. E.; Khonga, E. B.; Ramolemana, G. M.; Marumo, K.; Machacha, S.; Motsamai, T. Suitability of treated secondary sewage effluent for irrigation of horticultural crops in Botswana. J. Appl. Sci. 2005, 5, 451–454. Jin, X.; Huang, X. F.; Hoek, E. M. V. Role of specific ion interactions in seawater RO membrane fouling by alginic acid. Environ. Sci. Technol. 2009, 43, 3580–3587. Gilron, J.; Belfer, S.; Va¨isa¨nen, P.; Nystro¨m, M. Effects of surface modification on antifouling and performance properties of reverse osmosis membranes. Desalination 2001, 140, 167–179. Li, Q. L.; Xu, Z. H.; Pinnau, I. Fouling of reverse osmosis membranes by biopolymers in wastewater secondary effluent: Role of membrane surface properties and initial permeate flux. J. Membr. Sci. 2007, 290, 173–181. Nguyen, Q. T.; Glinel, K.; Pontie´, M.; Ping, Z. H. Immobilization of bio-macromolecules onto membranes via an adsorbed nanolayer: An insight into the mechanism. J. Membr. Sci. 2004, 232, 123–132. Lasquellec, V.; Marcq, J.; Sbaı¨, M.; Filmon, R.; Nguyen, T. Q.; De Nardi, F.; Pontie´, M. An innovative approach in the range to antibiofouling processes by LBL deposition of a cationic polyelectrolyte on the AN69 dialysis anionic membrane. J. Water Sci. 2007, 20, 175–183. Yang, Q. F.; Liu, Y. Q.; Gu, A. Z.; Ding, J.; Shen, Z. Q. Investigation of calcium carbonate scaling inhibition and scale morphology by AFM. J. Colloid Interface Sci. 2001, 240, 608–621. Hasson, D.; Drak, A.; Semiat, R. Induction times induced in an RO system by antiscalants delaying CaSO4 precipitation. Desalination 2003, 157, 193–207. Suratta, W. B.; Andrews, D. R.; Pujals, V. J.; Richards, S. A. Design considerations for major membrane treatment facility for groundwater. Desalination 2000, 131, 37–46. Ross, R. J.; Low, K. C.; Shannon, J. E. Polyaspartate scale inhibitorssbiodegradable alternatives to polyacry1ates. Mater. Perform. 1997, 36, 53–57. Wu, Y. T.; Grant, C. Effect of chelation chemistry of sodium polyaspartate on the dissolution of calcite. Langmuir 2002, 18, 6813–6820. Zhao, Y.; Song, L. F.; Ong, S. L. Fouling of RO membranes by effluent organic matter (EfOM): Relating major components of EfOM to their characteristic fouling behaviors. J. Membr. Sci. 2010, 349, 75–82. Ang, W. S.; Elimelech, M. Fatty acid fouling of reverse osmosis membranes: Implications for wastewater reclamation. Water Res. 2008, 42, 4393–4403. Li, Q. L.; Elimelech, M. Organic fouling and chemical cleaning of nanofiltration membranes: Measurements and mechanisms. Environ. Sci. Technol. 2004, 38, 4683–4693. Martinod, A.; Neville, A.; Euvrad, M.; Sorbie, K. Electrodeposition of a calcareous layer: Effects of green inhibitors. Chem. Eng. Sci. 2009, 64, 2413–2421. Wang, X. Y.; Lee, B. I.; Mann, L. Dispersion of barium titanate with polyaspartic acid in aqueous media. Colloid Surf., A 2002, 202, 71–80. Gill, J. A novel inhibitor for scale control in water desalination. Desalination 1999, 124, 43–50. Tang, C. Y.; Leckie, J. O. Membrane independent limiting flux for RO and NF membranes fouled by humic acid. Environ. Sci. Technol. 2007, 41, 4767–4773. Yang, Q. F.; Drak, A.; Hasson, D.; Semiat, R. RO module RTD analyses based on directly processing conductivity signals. J. Membr. Sci. 2007, 306, 355–364.

ES100513E