ARTICLE pubs.acs.org/est
Nanofiltration Membrane Fouling by Oppositely Charged Macromolecules: Investigation on Flux Behavior, Foulant Mass Deposition, and Solute Rejection Yi-Ning Wang and Chuyang Y. Tang* School of Civil and Environmental Engineering, and Singapore Membrane Technology Centre, Nanyang Technological University, Singapore 639798
bS Supporting Information ABSTRACT: Nanofiltration membrane fouling by oppositely charged polysaccharide (alginate) and protein (lysozyme) was systematically studied. It was found that membrane flux decline in the presence of both lysozyme and alginate was much more severe compared to that when there was only lysozyme or alginate in the feed solution. The flux performance for the mixed foulants was only weakly affected by solution pH and calcium concentration. These effects were likely due to the strong electrostatic attraction between the two oppositely charged foulants. Higher initial flux caused increased foulant deposition, more compact foulant layer, and more severe flux decline. The deposited foulant cake layer had a strong tendency to maintain a constant foulant composition that was independent of the membrane initial flux and only weakly dependent on the relative foulant concentration in feed solution. In contrast, solution chemistry (pH and [Ca2+]) had marked effect on the foulant layer composition, likely due to the resulting changes in the foulant foulant interaction. The mixed alginate lysozyme fouling could result in an initial enhancement in salt rejection. However, such initial enhancement was not observed when there was 1 mM calcium present in the feedwater, which may be attributed to the charge neutralization of the foulant layer.
1. INTRODUCTION Fouling of reverse osmosis (RO) and nanofiltration (NF) membranes by colloids and macromolecules has received considerable attention in the past few decades. Prior fouling studies with various feeds have revealed many important factors that affect fouling, including solution chemistry, hydrodynamic conditions, and membrane properties.1 A recent review of the mechanisms and factors controlling fouling of both RO and NF membranes1 pointed out the importance of mass transfer near the membrane surface and colloid membrane (and colloid colloid) interactions that are strongly affected by the feed solution chemistry. Past studies on single macromolecular feeds showed that electrostatic interactions between molecules play an important role, which was inferred from the effect of solution pH and ionic strength on fouling.2 8 Fouling was generally more severe at pH values close to the isoelectric point (IEP) of the macromolecules and at higher ionic strength due to reduced electrostatic repulsion.6 9 Calcium complexation with macromolecules has also been observed to increase the fouling rate markedly.7 11 The feedwater for NF or RO often contains more than one foulant. The handful of existing studies on mixed foulant feeds have shown different dominant effects and fouling mechanisms, including (1) the prefiltering effect, i.e., one foulant acts as a prefilter for another to reduce fouling in porous microfiltration (MF) and ultrafiltration (UF) membranes;12,13 (2) the r 2011 American Chemical Society
synergistic effect that explains the more severe flux decline by the mixed foulants compared to that by any single foulant;6,14,15 (3) the averaging effect where the rate and the extent of flux reduction for the mixture falls between that of the two single foulants;10,16 and (4) adsorption effect where macromolecules coat inorganic particles that alters the surface properties on the latter.17,18 Nevertheless, the effect of solution chemistry and hydrodynamic conditions on RO and NF fouling by mixed macromolecules is yet to be systematically studied.15 Moreover, an understanding of cake formation and ion rejection due to fouling by mixed macromolecules is still lacking. Little work has been performed to understand of the factors controlling the composition of the foulant cake layer. Therefore, the current study had three main objectives: (1) to systematically investigate solution chemistry and hydrodynamic effects on NF membrane fouling by a feed solution containing oppositely charged macromolecules; (2) to correlate mass deposition with flux decline at various conditions; and (3) to compare the change in salt rejection and correlation with cake deposition for different solution chemistries. Received: August 4, 2011 Accepted: September 19, 2011 Revised: September 14, 2011 Published: September 19, 2011 8941
dx.doi.org/10.1021/es202709r | Environ. Sci. Technol. 2011, 45, 8941–8947
Environmental Science & Technology
ARTICLE
2. MATERIALS AND METHODS 2.2. Chemicals. Unless specified otherwise, all reagents and chemicals were of analytical grade with purity over 99%. Ultrapure water with a resistivity of 18.2 MΩ.cm (Millipore Integral 10 Water Purification System) was used to prepare all working solutions. The pH and ionic compositions of the feed solution were adjusted by the addition of sodium chloride, calcium chloride, hydrochloric acid, and sodium hydroxide. Sodium alginate (ALG, Sigma A2158) and lysozyme (LYS, Fluka 62971) were used to represent polysaccharide and protein foulants of opposite charges under the testing conditions. These model foulants were chosen for their ready availability and their wide use in membrane fouling studies.3,6,10 Alginate is negatively charged at neutral pH due to its high content of carboxylic groups, and LYS is positively charged below pH 10.4. The molecular weights of alginate and LYS are ∼12 80 kDa and 14.3 kDa, respectively.3,6,13 Both foulants were received in powder form with purity above 98% and were stored at 4 °C in the dark. Working solutions were freshly prepared prior to each fouling experiment. 2.3. Nanofiltration Membrane. A commercial nanofiltration membrane NF270 (Dow FilmTec) was used in the current study. According to our prior studies,19,20 this NF membrane is a poly(piperazine)-based thin film composite polyamide membrane. Its surface is smooth (root-mean-square roughness ∼9 nm), negatively charged (zeta potential of ∼ 35 mV at pH 7), and highly hydrophilic (contact angle below 30°).19 21 The permeability and the NaCl rejection of the membrane are ∼0.87 L/m2 3 h 3 psi and 50 60%, respectively, when tested at 90 psi using a 10 mM NaCl feedwater.9 2.4. Fouling Experiment. Membrane fouling experiments were performed using a laboratory-scale crossflow filtration setup under constant pressure conditions. A detailed description of the setup can be found elsewhere.9 The fouling test procedure was adapted from Tang and co-workers.9,15,22 Briefly, membrane coupons (active membrane area ∼42 cm2) were first compacted and equilibrated with the background electrolyte solution (of composition identical to that used for the subsequent fouling test except that foulants were not added) for 2 days under pressure. This was to ensure that any subsequent flux decline after foulant addition was not influenced by the mechanical compaction of membrane. At the end of the 2-day equilibration stage, predissolved foulant solutions were added to achieve a desired total foulant concentration of 20 mg/L. The membrane flux immediately before the addition of foulants was recorded as the initial flux. The fouling experiments continued for 4 days. Pressure and crossflow velocity were maintained constant throughout each fouling experiment. The effects of relative foulant concentration, pH, calcium concentration, and initial flux were evaluated by varying one variable at a time while maintaining the rest at constant conditions. Unless specified otherwise, the following reference conditions were applied: • total foulant concentration of 20 mg/L (LYS + ALG for the mixed system); • total ionic strength of 10 mM (adjusted by the addition of NaCl and CaCl2); • crossflow velocity was 9.5 cm/s with diamond patterned spacer used in the feedwater channel; and • feed tank temperature controlled at 20 ( 1 °C. 2.5. Foulant Extraction and Mass Deposition Analysis. Fouled membrane coupons were gently rinsed with Milli-Q
Figure 1. Zeta potential of alginate and lysozyme in 10 mM NaCl as a function of pH. At least two replicate measurements were performed for each solution condition.
water to remove any labile foulants. Membrane samples (area of 1.267 2.534 cm2) were cut from the fouled coupons and used for foulant extraction. Sodium dodecyl sulfate (SDS, 5% 23 and NaOH (0.1%)24 were used for LYS and alginate extraction, respectively. Mild sonication (below 30 °C, 20 min) was employed after one day soaking of the samples in the extraction solution. The extractant of LYS was analyzed using a protein assay kit (Sigma, QuantiPro BCA Assay Kit, 0.5 30 μg/mL) at UV of 562 nm,25 and that of alginate was examined using phenol sulfuric acid method at UV of 485 nm (UV spectrophotometer, UV-1700 Shimadzu).26 2.6. Zeta Potential Measurement of Macromolecules. Zeta potential of LYS and alginate was measured using a Malvern ZetaSizer Nano ZS according to our prior publication.9 Freshly prepared LYS or alginate solutions were used in the measurements. Measurements were conducted in a background electrolyte of 10 mM NaCl, and the solution pH was adjusted by the addition of HCl or NaOH.
3. RESULTS AND DISCUSSION 3.1. Zeta Potential of Foulants. Figure 1 shows the zeta potential results of LYS and alginate. LYS was positively charged at circumneutral pH (∼ 7 mV at pH 7) with an isoelectric point (pHIEP) ∼ 10.4. It was more positively charged at lower pH. Alginate was negatively charged over the entire pH range evaluated (pH 2 9). Its zeta potential value decreased from ∼ 20 mV at pH 2 to ∼ 70 mV at pH 6. The value was nearly constant above pH 6, likely due to the complete deprotonation of carboxylic groups.6 3.2. Flux Behavior and Foulant Mass Deposition. 3.2.1. Effect of Feed Foulant Composition. The effect of feed foulant composition on membrane fouling was investigated by varying the LYS content (0, 5, 30, 50, 70, and 100 wt % of total feed foulant) while keeping the total feed foulant concentration at 20 mg/L. Figure 2a shows the membrane flux performances under an initial flux of 75 L/m2 3 h. Least flux decline was observed over the 4-day fouling tests when the feed contained either alginate alone (0% LYS, ∼ 30% flux decline) or lysozyme 8942
dx.doi.org/10.1021/es202709r |Environ. Sci. Technol. 2011, 45, 8941–8947
Environmental Science & Technology
ARTICLE
Figure 2. Effect of feed foulant composition on flux behavior and foulant mass accumulation: (a) flux performance; (b) foulant mass accumulation after 96-h fouling. Other test conditions: total foulant concentration of 20 mg/L, initial flux of 75 L/m2 3 h, pH 7, ionic strength of 10 mM, crossflow velocity of 9.5 cm/s, and temperature at 20 ( 1 °C.
Figure 3. Effect of solution chemistry on flux behavior and foulant mass accumulation: (a) flux performance; (b) mass accumulation after 96-h fouling. Other test conditions: initial flux of 75 L/m2 3 h, feedwater containing 10 mg/L lysozyme and 10 mg/L alginate, total ionic strength of 10 mM, crossflow velocity of 9.5 cm/s, and temperature at 20 ( 1 °C.
alone (100% LYS, ∼ 20% flux decline). With a small amount of LYS (5%) in the feed, a much more severe flux reduction (∼ 50%) occurred. This was likely due to the increased foulant mass deposition (Figure 2b) as a result of the electrostatic attraction between the oppositely charged LYS and alginate molecules. The most severe fouling occurred at the LYS relative concentration of 30, 50, and 70% with ∼80% flux loss. The fastest rate of initial flux decline was observed for the feed with 70% LYS (inset of Figure 2a). The reason for explaining the effect of foulant relative concentration is discussed in the following text based on the mass deposition analysis. Figure 2b presents the foulant mass and composition of the extracted cake layer. Compared to the single foulant system (0 or 100% LYS), the mixed foulant system had a much greater amount of foulant mass deposition. A small amount of LYS (e.g., 5% LYS) in the mixed foulants increased the total foulant mass deposition to ∼300 μg/cm2, compared to ∼100 μg/cm2 when there was only alginate in the feed. Even more foulant deposition (400 500 μg/cm2) was observed at LYS contents of
30 70%, which correlated well with the more severe flux reduction under those conditions. Interestingly, the deposited mass ratio of LYS to alginate in the foulant layer was nearly constant (mLYS/malg = ∼4) for feeds with 30 70% LYS content. There seemed to be a strong tendency that the relative quantity of LYS on fouled membranes approached a quite constant value (∼ 80% of the total deposited mass) when sufficient concentrations of LYS and alginate were available in the feed solution. Conceptually, there exists a thermodynamically most stable arrangement of foulant cake deposition (corresponding to a minimum free energy state27). To achieve such a minimum energy arrangement, it is reasonable to assume that the oppositely charged LYS and alginate may deposit in a certain configuration to minimize the electrostatic potential energy (e.g., a mosaic arrangement where each molecule is surrounded by adjacent oppositely charged molecules). Therefore, the mass ratio mLYS/malg is presumably determined by the relative charge densities of two foulants, their geometry, and the presence of other additional interactions (e.g., hydrophobic interaction or specific 8943
dx.doi.org/10.1021/es202709r |Environ. Sci. Technol. 2011, 45, 8941–8947
Environmental Science & Technology interactions). This thermodynamic tendency (i.e., constant mLYS/malg) will be preserved regardless of the foulant mass ratio in the feedwater, as long as both LYS and alginate are present in sufficient quantity (LYS 30 70% in this study). The constant mLYS/malg ratio is somewhat analogous to the precipitation (scaling) of inorganic minerals such as CaSO4 (gypsum) where the calcium to sulfate ratio in the precipitation is fixed (at 1:1 in this particular case) irrespective of their ratio in the bulk solution.28 The fouling behavior at lower LYS content in feed is also interesting. Even when there was only 5% LYS in the feedwater, the relative LYS content in the foulant layer was as high as 53%. The corresponding mLYS/malg ratio was ∼1. Although it was still significantly lower than the “thermodynamic constant ratio” of ∼4, this ratio in the foulant layer was an order of magnitude higher than the corresponding ratio in the feed (∼0.053). This observation may suggest that there might be a tendency for the foulant layer mass ratio to approach the thermodynamic constant ratio even at 5% LYS in feed, although this ratio was also limited by the slow deposition of LYS due to its low feed concentration (i.e., due to the kinetic constraint). The same kinetic consideration could also explain why the feeds with 30, 50, and 70% LYS content fouled membrane more severely. It could predict that the fastest initial flux drop shall occur for the feed containing the “desired” foulant proportion for cake formation (i.e., LYS to alginate concentration ratio of ∼4). In this way, the feed containing 70% LYS initiated faster flux decline compared to 50 and 30% LYS content, likely due to the foulants composed of 70% LYS and 30% alginate having the most proximal composition to the desired proportion of cake formation. 3.2.2. Effect of Solution Chemistry. The effect of pH and calcium concentration on flux behavior is shown in Figure 3a for a feed containing 10 mg/L LYS and 10 mg/L alginate. A nearly identical flux behavior was observed at both pH 5 and pH 7, presumably resulting from the similar strong electrostatic attraction between the oppositely charged macromolecules. Unlike the important role of solution pH in single protein fouling,9 it became less crucial during mixture fouling at least for the conditions explored in the current study. This observation is consistent with our prior study that the flux decline for a feed containing a binary protein mixture was only weakly dependent on pH when it was between the IEPs of the two proteins.15 Notwithstanding the identical flux behavior at different solution pHs, the results of foulant mass deposition showed some remarkable differences (Figure 3b). There were more alginate and less LYS deposition at pH 5 compared to pH 7 at the end of fouling tests, and the relative mass of LYS on the fouled membrane reduced from ∼80% at pH 7 to 66% at pH 5. At the reduced pH, LYS was more positively charged whereas alginate was less negatively charged (Figure 1), which may require a lesser amount of LYS to balance the negatively charged alginate. This is consistent with our hypothesis that the mLYS/malg ratio is dependent on the relative charge density of the foulants—an increased LYS to alginate charge ratio leads to a reduced mLYS/malg. Calcium concentration effect was studied by comparing the feeds containing 1 and 0 mM Ca2+ with a total ionic strength of 10 mM at pH 7 (Figure 3a). The presence of 1 mM Ca2+ only led to slightly more severe flux reduction (86% reduction, compared to 80% without Ca2+). It has been well recognized that alginate can form a cross-linked gel layer in the presence of calcium ions as a result of calcium complexation with alginate carboxylic groups.10 When alginate is the sole foulant in the feed, increasing
ARTICLE
Figure 4. Effect of initial flux on flux behavior and foulant mass accumulation: (a) flux performance; (b) mass accumulation after 96-h fouling. Other test conditions: feedwater containing 10 mg/L lysozyme and 10 mg/L alginate, pH 7, Ca2+ concentration of 1 mM, total ionic strength of 10 mM, crossflow velocity of 9.5 cm/s, and temperature at 20 ( 1 °C.
calcium concentration tends to drastically reduce membrane permeate flux.10,29 In contrast, this marked effect of Ca2+ was not observed in the fouling by the LYS alginate mixture. The reduced Ca2+ effect was probably due to the presence of LYS in feed solution, which can cause a strong LYS alginate attraction and form a dense cake layer even if Ca2+ was not present. Alternatively, the positively charged LYS may compete with Ca2+ for the negatively charged sites ( COO groups) in the alginate, leading to a reduced degree of Ca2+ alginate complexation. Introducing Ca2+ into the feed solution only marginally accelerated flux decline at the beginning (within 10 h) upon foulant addition. During such a fouling process, a compound effect of LYS alginate and Ca2+ alginate complexation may take place (noting that Ca2+ had minimal effect on the fouling by LYS alone, see Supporting Information S1). Mass deposition results (Figure 3b) show a significant increase (more than doubled) in the amount of alginate accumulation on the fouled membrane due to the addition of Ca2+, despite the insignificant variation in flux performance. The increased alginate deposition may be attributed to the Ca2+ alginate 8944
dx.doi.org/10.1021/es202709r |Environ. Sci. Technol. 2011, 45, 8941–8947
Environmental Science & Technology
Figure 5. Foulant layer hydraulic resistance and specific cake layer resistance as a function of total foulant mass deposition. Other test conditions: feedwater containing 10 mg/L lysozyme and 10 mg/L alginate, pH 7, Ca2+ concentration of 1 mM, total ionic strength of 10 mM, crossflow velocity of 9.5 cm/s, and temperature at 20 ( 1 °C.
complexation. On the other hand, LYS deposition was reduced, and the relative mass of LYS on membrane was reduced to ∼52%. This further suggested the existence of the compound effect of LYS alginate and Ca2+ alginate complexation, where both LYS and Ca2+ compete for negatively charged COO sites in alginate. The specific Ca2+ alginate interaction and subsequent reduction in the charge density of alginate may also explain the drastically reduced mLYS/malg ratio (∼4 when there was no calcium and ∼1 in the presence of 1 mM Ca2+). Once again, the reduced LYS to alginate mass ratio corresponds to an increased ratio of charge density. 3.2.3. Effect of Initial Flux. The effect of initial flux was evaluated over an initial flux range of 15 120 L/m2 3 h. The feedwater contained 1 mM Ca2+ at pH 7. Much more severe flux reduction was observed at higher initial flux (Figure 4a), which was consistent with the results from our previous fouling studies with single protein9 and binary protein mixture,15 and as well as from many other existing publications.5,22,30 The fluxes started to converge to an identical limiting value although they began with different initial values; this was consistent with our observation in protein fouling and humic acid fouling.7 9,22 The surface-interaction controlled limiting flux model, which states that the mutual action of hydrodynamic drag and foulant depositedfoulant (foulant membrane) interaction determines foulant deposition and the extent of fouling, is appropriate to explain the limiting flux phenomenon in the current study.1,7 A similar effect of initial flux was also observed for the feed without Ca2+ (S2(a), Supporting Information). Foulant mass deposition results (Figure 4b and S2(b)) show an expected trend that more LYS and alginate deposited on membranes fouled under higher initial flux. It is worthwhile to note that the deposited mass of the two foulants varied synchronously such that their mass ratio remained nearly constant (mLYS/malg = ∼1 for the feed containing 1 mM Ca2+ and ∼4 for the feed without Ca2+). This suggests that increasing initial flux (and/or applied pressure) may have negligible effect on the foulant layer composition and that the mLYS/malg ratio may be primarily determined by the foulant foulant interactions but not the hydrodynamic forces.
ARTICLE
Figure 6. Foulant layer composition as a function of feed composition. The feedwater had a total foulant concentration of 20 mg/L, ionic strength of 10 mM, crossflow velocity of 9.5 cm/s, and temperature at 20 ( 1 °C.
As most hydraulic cake resistance was built up at the initial period where the flux dropped significantly (Figure 4a and S2(a)), the large difference in the deposited cake mass under different initial fluxes may develop within the initial fouling period in the current study. Operating at a higher initial flux (or higher applied pressure) had no noticeable benefit on the longer term membrane flux (say, at 96 h), likely due to the densely formed cake layer. Figure 5 shows that the hydraulic resistance of foulant cake layer Rf increased with the total deposited foulant mass mf. In addition, the specific cake layer resistance (i.e., Rf/mf) also increased at higher mf. The increased specific cake layer resistance may be attributed to the cake layer compaction at the higher initial flux (and applied pressure).9 In other words, the higher initial flux resulted in more foulant deposition, a more compact cake layer, and thus greater flux reduction, which agrees well with a previous characterization study on humic acid fouling with RO and NF membranes.31 The various effects (solution chemistry, feed composition, and initial flux) on the foulant layer composition are presented in Figure 6. These can be summarized into the following three aspects: (1) the foulant composition in the feed solution only had weak influence on the relative mass of deposited foulants as long as both foulants were present in the feed in sufficient concentrations (such that the mLYS/malg ratio was not kinetically limited); (2) the initial flux had negligible effect on the foulant composition in the deposited cake layer; and (3) the solution chemistry such as pH and Ca2+ concentration affected the mLYS/malg ratio as a result of the changes in foulant foulant interactions (e.g., the LYS alginate electrostatic attraction and the calcium alginate specific interaction). 3.3. Effect of Fouling on Salt Rejection. Figure 7 presents the normalized salt rejection performances (normalized against the corresponding virgin membrane rejection) as a function of fouling time for different solution compositions. The effect of feed foulant composition is shown in Figure 7a. It is interesting to observe an initial increase in salt rejection upon fouling for all cases. Tang et al.22 attributed this beneficial effect to the preferential deposition of foulants over membrane defects where the localized flux was at maximum. In this way, even a small amount of foulants deposition may significantly improve salt 8945
dx.doi.org/10.1021/es202709r |Environ. Sci. Technol. 2011, 45, 8941–8947
Environmental Science & Technology
ARTICLE
was clearly observable when there was no Ca2+ in the feed solution (at both pH 5 and pH 7). In contrast, such initial enhancement was not observed for the feed containing 1 mM Ca2+, even though (1) the flux decline in the presence of calcium was only slightly more severe than that without calcium, and (2) the foulant cake layer was likely more compact with the presence of calcium (Supporting Information S3). A possible explanation may be that the foulant charge neutralization associated with the calcium alginate complex formation may result in a cake layer with reduced ability of solute retention. Prior studies have demonstrated that rejection of charged solutes can be enhanced by Donnan exclusion.33 Thus, the loss of rejection in the presence of 1 mM Ca2+ might be attributed to the weakening of Donnan exclusion mechanism in addition to the enhanced solute polarization inside the foulant cake layer.
4. IMPLICATIONS The current study investigated the effect of feed foulant composition, solution chemistry, and initial flux on membrane fouling by oppositely charged macromolecular foulants. It was found, for the first time, that the foulant layer composition was primarily governed by the feed solution chemistry and that this composition had little or weak dependence on initial membrane flux or foulant composition in the feed solution. Meanwhile, the flux behavior was regulated by the interplay of membrane flux and foulant foulant (and foulant membrane) interaction. Although the current study was performed for a nanofiltration membrane, the mechanisms revealed in the current study may also apply in other membrane systems, such as reverse osmosis, ultrafiltration, and microfiltration. This study may have significant implications in membrane fouling control, particularly for membrane based wastewater reclamation plants and membrane bioreactors, where fouling by protein/polysaccharide mixture may likely occur. ’ ASSOCIATED CONTENT Figure 7. Salt rejection as a function of fouling duration: (a) effect of feed foulant composition at pH 7 without Ca2+; (b) effect of solution chemistry (i.e, pH and Ca2+ concentration) for a feed containing 10 mg/L lysozyme and 10 mg/L alginate. Other test conditions: total ionic strength of 10 mM, crossflow velocity of 9.5 cm/s, and temperature at 20 ( 1 °C.
rejection by defects sealing.22 Nevertheless, the rejection started to decrease at later stages of fouling probably due to a combination of several factors: (1) the decreased membrane permeate flux which resulted in a reduced dilution factor to the permeate concentration; (2) the cake-enhanced concentration polarization of solutes when the foulant cake layer became more extensive;32 and (3) the significant change of membrane surface charge properties upon fouling. Consistent with the above explanations, the decrease in salt rejection in the later stage of fouling seemed to be more obvious for more severe fouling conditions that were accompanied by greater permeate flux reduction and more foulant mass deposition (e.g., for the feed containing 30 70% LYS). In contrast, when the feed only contained LYS or alginate, relatively stable rejection was maintained, consistent with the milder membrane fouling for these cases. Figure 7b presents the effect of solution chemistry on salt rejection during fouling. An initial salt rejection enhancement
bS
Supporting Information. S1. Effect of ionic composition on flux behavior during LYS fouling; S2. Effect of initial flux on flux behavior and foulant mass accumulation in absence of calcium; S3. Foulant mass deposition and specific cake layer resistance. This material is available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION Corresponding Author
*Tel: +65 6790 5267; fax: +65 6791 0676; e-mail: cytang@ntu. edu.sg; mail: 50 Nanyang Avenue N1-1b-35, School of Civil & Environmental Engineering, Nanyang Technological University, Singapore 639798.
’ ACKNOWLEDGMENT We thank the Environment and Water Industry Programme Office (EWI) under the National Research Foundation of Singapore (Project MEWR C651/06/176) for funding the research work carried out in this manuscript. The comments from Professor William B. Krantz have significantly strengthened this paper. We also thank Dow FilmTec for providing membrane samples. 8946
dx.doi.org/10.1021/es202709r |Environ. Sci. Technol. 2011, 45, 8941–8947
Environmental Science & Technology
’ REFERENCES (1) Tang, C. Y.; Chong, T. H.; Fane, A. G. Colloidal interactions and fouling of NF and RO membranes: A review. Adv. Colloid Interface Sci. 2011, 164, 126–143. (2) Palecek, S. P.; Mochizuki, S.; Zydney, A. L. Effect of ionic environment on BSA filtration and the properties of BSA deposits. Desalination 1993, 90 (1 3), 147–159. (3) Palecek, S. P.; Zydney, A. L. Intermolecular electrostatic interactions and their effect on flux and protein deposition during protein filtration. Biotechnol. Prog. 1994, 10 (2), 207–213. (4) Yuan, W.; Zydney, A. L. Humic acid fouling during ultrafiltration. Environ. Sci. Technol. 2000, 34 (23), 5043–5050. (5) She, Q.; Tang, C. Y.; Wang, Y. N.; Zhang, Z. The role of hydrodynamic conditions and solution chemistry on protein fouling during ultrafiltration. Desalination 2009, 249 (3), 1079–1087. (6) Ang, W. S.; Elimelech, M. Protein (BSA) fouling of reverse osmosis membranes: Implications for wastewater reclamation. J. Membr. Sci. 2007, 296 (1 2), 83–92. (7) Tang, C. Y.; Leckie, J. O. Membrane independent limiting flux for RO and NF membranes fouled by humic acid. Environ. Sci. Technol. 2007, 41 (13), 4767–4773. (8) Tang, C. Y.; Kwon, Y. N.; Leckie, J. O. The role of foulant foulant electrostatic interaction on limiting flux for RO and NF membranes during humic acid fouling - Theoretical basis, experimental evidence, and AFM interaction force measurement. J. Membr. Sci. 2009, 326 (2), 526–532. (9) Wang, Y.; Tang, C. Y. Protein fouling of nanofiltration, reverse osmosis, and ultrafiltration membranes - The role of hydrodynamic conditions, solution chemistry, and membrane properties. J. Membr. Sci. 2011, 376 (1 2), 275–282. (10) Lee, S.; Elimelech, M. Relating organic fouling of reverse osmosis membranes to intermolecular adhesion forces. Environ. Sci. Technol. 2006, 40 (3), 980–987. (11) Li, Q.; Xu, Z.; 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 (1 2), 173–181. (12) Arora, N.; Davis, R. H. Yeast cake layers as secondary membranes in dead-end microfiltration of bovine serum albumin. J. Membr. Sci. 1994, 92 (3), 247–256. (13) Palacio, L.; Ho, C. C.; Pradanos, P.; Hernandez, A.; Zydney, A. L. Fouling with protein mixtures in microfiltration: BSA-lysozyme and BSA-pepsin. J. Membr. Sci. 2003, 222 (1 2), 41–51. (14) Iritani, E.; Mukai, Y.; Murase, T. Separation of binary protein mixtures by ultrafiltration. Filtr. Sep. 1997, 34 (9), 967–973. (15) Wang, Y.; Tang, C. Y. Fouling of nanofiltration, reverse osmosis and ultrafiltration membranes by protein mixtures: The role of interfoulant-species interaction. Environ. Sci. Technol. 2011, 45 (15), 6373–6379. (16) Zazouli, M. A.; Nasseri, S.; Ulbricht, M. Fouling effects of humic and alginic acids in nanofiltration and influence of solution composition. Desalination 2010, 250 (2), 688–692. (17) Lee, S.; Cho, J.; Elimelech, M. Combined influence of natural organic matter (NOM) and colloidal particles on nanofiltration membrane fouling. J. Membr. Sci. 2005, 262 (1 2), 27–41. (18) Contreras, A. E.; Kim, A.; Li, Q. Combined fouling of nanofiltration membranes: Mechanisms and effect of organic matter. J. Membr. Sci. 2009, 327 (1 2), 87–95. (19) Tang, C. Y.; Kwon, Y.-N.; Leckie, J. O. Effect of membrane chemistry and coating layer on physiochemical properties of thin film composite polyamide RO and NF membranes. I. FTIR and XPS characterization of polyamide and coating layer chemistry. Desalination 2009, 242 (1 3), 149–167. (20) Tang, C. Y.; Kwon, Y.-N.; Leckie, J. O. Effect of membrane chemistry and coating layer on physiochemical properties of thin film composite polyamide RO and NF membranes II. Membrane physiochemical properties and their dependence on polyamide and coating layers. Desalination 2009, 242 (1 3), 168–182.
ARTICLE
(21) Tang, C. Y.; Fu, Q. S.; Criddle, C. S.; Leckie, J. O. Effect of flux (transmembrane pressure) and membrane properties on fouling and rejection of reverse osmosis and nanofiltration membranes treating perfluorooctane sulfonate containing wastewater. Environ. Sci. Technol. 2007, 41 (6), 2008–2014. (22) Tang, C. Y.; Kwon, Y.-N.; Leckie, J. O. Fouling of reverse osmosis and nanofiltration membranes by humic acid -- Effects of solution composition and hydrodynamic conditions. J. Membr. Sci. 2007, 290 (1 2), 86–94. (23) Jones, K. L.; O’Melia, C. R. Protein and humic acid adsorption onto hydrophilic membrane surfaces: Effects of pH and ionic strength. J. Membr. Sci. 2000, 165 (1), 31–46. (24) Guo, X.; Chen, X.; Hu, W. Studies on cleaning the polyvinylchloride ultrafiltration membrane fouled by sodium alginate. Environ. Technol. 2009, 30 (5), 431–435. (25) Brown, R. E.; Jarvis, K. L.; Hyland, K. J. Protein measurement using bicinchoninic acid: Elimination of interfering substances. Anal. Biochem. 1989, 180 (1), 136–139. (26) Dubois, M.; Gilles, K. A.; Hamilton, J. K.; Rebers, P. A.; Smith, F. Colorimetric method for determination of sugars and related substances. Anal. Chem. 1956, 28 (3), 350–356. (27) Greiner, W.; Neise, L.; St€ocker, H.; Rischke, D.Thermodynamics and Statistical Mechanics; Springer: New York, 1995 (28) Stumm, W.; Morgan, J., Aquatic Chemistry: Chemical Equilibria and Rates in Natural Waters, 3rd ed.; Wiley-Interscience: New York, 1996; p 1040. (29) Ang, W. S. Optimization of Chemical Cleaning of Organicfouled Reverse Osmosis Membranes. Yale University, 2008. (30) Wu, D.; Howell, J. A.; Field, R. W. Critical flux measurement for model colloids. J. Membr. Sci. 1999, 152 (1), 89–98. (31) 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 (3), 942–949. (32) Hoek, E. M. V.; Elimelech, M. Cake-Enhanced Concentration Polarization: A New Fouling Mechanism for Salt-Rejecting Membranes. Environ. Sci. Technol. 2003, 37 (24), 5581–5588. (33) Schaep, J.; Vandecasteele, C.; Mohammad, A. W.; Bowen, W. R. Analysis of the salt retention of nanofiltration membranes using the Donnan-steric partitioning pore model. Sep. Sci. Technol. 1999, 34 (15), 3009–3030.
8947
dx.doi.org/10.1021/es202709r |Environ. Sci. Technol. 2011, 45, 8941–8947
