ARTICLE pubs.acs.org/Langmuir
Zeta Potential of Ion-Conductive Membranes by Streaming Current Measurements He Xie, Tomonori Saito, and Michael A. Hickner* Department of Materials Science and Engineering, The Pennsylvania State University, University Park, Pennsylvania 16802, United States ABSTRACT: Surface charge properties have a significant influence on membrane retention and fouling performance. As a key parameter describing the surface charge of membranes used in aqueous applications, zeta potential measurements on membranes of various types have attracted great attention. During the zeta potential characterization of a series of ion-conductive sulfonated poly(sulfone) membranes, it was found that the measured streaming current varied with the thickness of the sample, which is not predicted by the classical Smoluchowski equation. Moreover, for higher conductivity membranes with an increased concentration of sulfonate groups, the zeta potential tended toward zero. It was determined that the influence of membrane bulk conductance on the measured streaming current must be taken into account in order to correctly interpret the streaming current data for ion-conductive polymers and understand the relationship between membrane chemical composition and zeta potential. Extrapolating the measured streaming current to a membrane thickness of zero has proven to be a feasible method of eliminating the error associated with measuring the zeta potential on ion conductive polymer membranes. A linear resistance model is proposed to account for the observed streaming currents where the electrolyte channel is in parallel with the ion-conductive membranes.
1. INTRODUCTION Many countries in arid regions now rely on desalination of seawater as an important fresh water source. Multistage flash (MSF) distillation and reverse osmosis (RO) processes are the two most commercially important desalination technologies, and by 2009, over 15 000 desalination plants were in operation worldwide, with approximately 50% of those being RO plants.1 RO accounts for more than 70% of the seawater and brackish water desalination market in Europe, and almost all recently installed desalination plants employ RO.2 The recent growth in RO capacity is, in part, a result of improved membranes that have long lifetimes, high salt rejection, and high water flux. However, during the RO and NF processes, various particles in the feed solution, for example, aluminosilicate clays or humic acids, or biological growth may deposit on the membrane surface, and these fouling processes result in a significant reduction of transmembrane permeability.3,4 Periodic cleaning is not able to completely alleviate membrane fouling, and fouling mitigation processes and accommodation of the decreased productivity of the membranes increase the maintenance and operating costs of the water treatment process. Thus, preventing fouling at its earliest stages has become a major research focus. Fouling processes are strongly influenced by the surface charge properties of the membrane in contact with the raw water and the species suspended in the feed solution.58 Also, electrostatic interactions between ions in the water to be desalinated and the membrane surface charge plays an important role in ion separation based on Donnan exclusion.9 Zeta potential, which provides useful information about the surface charge properties of the membrane in solution, is an important parameter in fouling r 2011 American Chemical Society
studies as well as in desalination studies. A number of techniques for measuring the zeta potential have been developed based on one of the three electrokinetic effects, electrophoresis, electroosmosis, and streaming potential, out of which streaming potential, or streaming current, is the most suitable for flat membrane surfaces.10 Streaming potential characterization has been carried out on several commercial nanofiltration (NF) membranes and reverse osmosis (RO) membranes by Elimelech,10 Childress,3,11,12 Peeters,13 and others, to investigate the influence of solution chemistry on membrane surface charge. Additionally, zeta potential has been used to rationalize membrane retention mechanism and fouling properties.13,14 Studies dealing with solution ion type and concentration, solution pH, surfactant, and natural organic matter (NOM) have been performed.3,1012 In most cases, a change of zeta potential is explained by considering the dissociation of surface functional groups or the preferential adsorption of ions and macromolecules from solution. Other studies have focused on how membrane chemical composition influences the zeta potential of the membrane surface where different types of materials have different relationships between zeta potential and pH.15,16 Hinke and Staude have characterized the zeta potentials of chemically modified poly(sulfone) microporous membranes and have shown distinctive differences between poly(sulfone) membranes with neutral Received: December 28, 2010 Revised: March 7, 2011 Published: March 28, 2011 4721
dx.doi.org/10.1021/la105120f | Langmuir 2011, 27, 4721–4727
Langmuir functional groups, such as chloromethyl moieties, methyl esters, and amines. At low pH, membranes with neutral functional groups displayed negative zeta potentials due to the adsorption of anions from the solution, while the membranes with tethered amines displayed positive zeta potentials due to protonation of the amine surface groups.17 Werner and Jacobasch studied a series of sulfonated poly(ethersulfone) hollow fiber membranes with different degrees of sulfonation and observed that, as the degree of sulfonation increased, the zeta potential became less negative.18 Mockel et al. reported a similar trend for carboxylated poly(sulfone) membranes.19 Both of these research groups explained their results on membranes with tethered negative charges by the extraordinary hydrophilicity of the acid groups that increased the thickness of the interfacial swelling layer at the membranes’ surface as the degree of substitution was increased. As the thickness of the swelling layer increased, the shear plane moved toward the solution bulk and a lower zeta potential was obtained on samples with at higher degree of tethered ions. Using sulfonated poly(phenylsulfone) as a model membrane system, this article reports detailed analysis of the measured streaming current in a microfluidic channel with sulfonated membrane walls and subsequent calculation of zeta potential for materials with different ion contents. The aim of this work was to investigate and explain the influence of membrane bulk conductance on zeta potential measurements for samples that display significant ion conductivity such as would occur with a charged membrane, or a membrane that absorbs electrolyte. Also proposed is a more accurate analysis method and linear resistance model to compute the zeta potential of these kinds of ion-conductive materials, which has not been reported previously.
2. BACKGROUND When a solid surface is brought into contact with an aqueous solution, most materials will acquire a surface charge due to the dissociation of surface ionic groups, the preferential adsorption of one type of ion from the aqueous solution, or some other charging mechanism.20 The electrostatic charge on the solid surface will attract counterions and repel co-ions in the solution. The electric double layer (EDL) is a region close to the charged surface in which there is an excess of counterions to neutralize the surface charge.20 In the liquid layer adjacent to the solid surface, termed the Stern layer, ions are strongly attracted to the surface charge and are immobile.21 Further away from the surface in the diffuse layer, ions are less affected by electrostatic interactions near the solid surface and the net charge density gradually decreases to that of the bulk liquid. The diffuse layer can flow under the influence of tangential stress and the shear plane is defined where the diffuse layer “slips” or flows past the charged surface. The electrical potential at the shear plane, which can be measured experimentally, is defined as the zeta potential, ζ. Since the electrical potential at the solid surface cannot be measured directly, zeta potential is often the most valuable characterization of double-layer properties.15 Zeta potential plays important roles in various applications such as microfluidics,2224 colloid chemistry,25,26 and membrane fouling.3 The zeta potential is influenced by surface composition, as well as solution properties such as the nature
ARTICLE
Figure 1. Parallel plate microchannel geometry for streaming current measurements.
of the ions, ionic strength, and pH.3,10 While some phenomena such as electrophoresis27 and colloid vibration current28 have been used to characterize the zeta potential, measurement of the streaming potential in a channel geometry is the most commonly used technique for characterizing the zeta potential of flat surfaces. In the streaming potential technique, a pressure differential, ΔP, is applied between the ends of a channel or capillary containing electrolyte to induce hydrodynamic flow. The ions in the diffuse layer of the EDL are carried in the direction of the flow, resulting in an electrical current defined as the streaming current, Is. The streaming current can be measured when the electrical potential difference between the two ends of the channel is zero. At open circuit, the streaming potential, Es, generated by the hydrodynamic flow of electrolyte is the potential difference between the ends of the channel which induces a flow of ions in the opposite direction of the flow, defined as the conduction current Ic. Streaming potential is measured when the flow reaches a steady state and the conduction current equals the streaming current, making the total electrical current in the channel zero. Figure 1 displays a slit microchannel geometry usually employed for the streaming potential measurement; L is the length of channel, W is the width of the channel, h is the channel height, and dm is the thickness of the membrane. The Smoluchowski equation describes the relationship between zeta potential (ζ) and streaming potential (eq 1) or streaming current (eq 2)15 ζ¼
dEs η L dEs η ¼ λ0 dP 3 εε0 3 WhR dP 3 εε0
ð1Þ
ζ¼
dIs η L dIs η ¼ λ0 R dP 3 εε0 3 Wh dP 3 εε0
ð2Þ
where dEs/dP is the change in streaming potential with pressure, dIs/dP is the change in streaming current versus pressure, R is the resistance of the measuring cell, η is the electrolyte viscosity, ε0 is the vacuum permittivity, and ε is the dielectric constant of the electrolyte. Equation 3 is obtained through integration of the product of the volume charge density, F, and velocity, V, of the electrolyte across the channel. In this case, F can be expressed as the onedimensional variant of the Poisson equation (eq 4)29 Z y ¼ h=2 IS ¼ 2W FðyÞV ðyÞ dy ð3Þ y¼0
FðyÞ ¼ εε0 4722
d2 Ψ dy2
ð4Þ
dx.doi.org/10.1021/la105120f |Langmuir 2011, 27, 4721–4727
Langmuir
ARTICLE
Figure 2. Repeat unit of sulfonated poly(phenylsulfone).
Table 1. Properties of Sulfonated Poly(phenylsulfone) Samples DSNMRa
IECa (mmol g1)
water uptake (%)
PSf-I
0.26
0.6
7
PSf-II
0.46
1.1
7
PSf-III
0.73
1.6
22
PSf-IV
0.93
2.0
27
PSf-V
1.28
2.5
70
Figure 3. Streaming current versus applied pressure for a poly(phenylsulfone) membrane.
a
Degree of sulfonation (DS) determined by 1H NMR using BrukerSpectrospin, 400 MHz in DMSO. Ion exchange capacity (IEC) computed from DS.
where y is the distance from the center plane of the channel and Ψ is the electrical potential at any position y. We can assume that Ψ and dΨ/dy become zero at y = 0 and designate Ψ at y = h/2 as the zeta potential, ζ; then, eq 3 could be rewritten and eq 5 is derived. IS ¼
ΔPεε0 hWζ ηL
ð5Þ
Though the Smoluchowski equation is widely accepted, it does not take into account surface conductance. The Fairbrother-Mastin procedure for eliminating the contribution of surface conductance is based on the assumption that surface conductance is suppressed at sufficiently high electrolyte concentration.30 The conductivity term in eq 1 and eq 2 is replaced by λ0 ¼
λH R H R
ð6Þ
where RH is the resistance of the channel when the cell is filled with a solution of high salt concentration, λH is the conductivity of this solution, and R is the resistance of the channel when filled with the measurement solution. Yaroshchuk and Ribitsch31 have pointed out that for certain materials, such as porous membranes, the conduction current Ic does not necessarily take the same path as the streaming current. If the conduction current circulates through the membrane body, then the membrane wall conductance cannot be eliminated in the simple Fairbrother-Mastin procedure because it is proportional to the conductance in the channel and will not be suppressed. Yaroshchuk and Ribitsch described the streaming current/ potential cell as a “sandwich” composed of the channel and two conducting layers with the total electrical conductance (surface conductance is neglected) being 1 Wh Whm ¼ λ0 þ 2 λm R L L
ð7Þ
where hm is the thickness of the membrane and λm is its conductivity. Substituting eq 7 into eq 1 leads to dEs εε0 ζ ¼ ηλ0 dP
1 λm h m 1þ λ0 h
ð8Þ
In eq 8, the reciprocal of dEs/dP is a linear function of cell height h. If the measurement is carried out at several channel heights, it is possible to perform a linear extrapolation to infinitely large channel heights and a correct value of ζ can be determined. In Yaroshchuk and Ribitsch’s analysis, they assumed that, during the streaming current measurement, there would be no conduction current in the membrane body; thus, the streaming current was not affected by membrane conductance and did not need the correction relating to cell height. In a study of porous membranes by Yaroshchuck and Luxbacker,32 it was observed that the cell resistance and streaming current coefficient (dIs/dP) were linear functions of channel height. Given the ohmic behavior of the streaming channel observed in the previous studies, we set out to test whether the conductance of ioncontaining membranes in parallel with the electrolyte channel could be described in a similar fashion using a parallel resistance model. In our investigations, streaming current was measured and the zeta potential is obtained via eq 2. Our results show that the streaming current measurement is affected by the membrane conductance and the influence of the membrane must be taken into account when trying to interpret the results and obtain information about the zeta potential behavior of materials that are ion conductive.
3. EXPERIMENTAL SECTION Materials. Poly(phenylsulfone) (Radel R-5500, MW 63 000) was kindly supplied through a donation from Solvay Advanced Polymers, LLC (Alphareta, GA). A series of sulfonated poly(phenylsulfone) samples with different ion contents (PSf-I-V) were prepared through sulfonation of the polymers with trimethylsilyl chlorosulfonate (Aldrich, 99%) in tetrachloroethane (Aldrich, >98%) solution.33 The repeat unit for the prepared sulfonated poly(phenylsulfone) is shown in Figure 2 and the properties of the samples studied in this work are summarized in Table 1. Radel and sulfonated poly(phenylsulfone) were dissolved in dimethylformamide (DMF) and then cast on glass plates and dried in a vacuum oven at 80 °C to form flat membranes. By varying the casting solution concentration, membranes with thicknesses ranging from 20 to 4723
dx.doi.org/10.1021/la105120f |Langmuir 2011, 27, 4721–4727
Langmuir
Figure 4. Zeta potential versus pH for sulfonated poly(phenylsulfone) of different IECs.
ARTICLE
Figure 5. Zeta potential at pH 5 versus IEC for the series of sulfonated poly(phenylsulfone) membranes.
120 μm were obtained. After casting, the membranes were treated in a 50 °C bath of 0.5 M H2SO4 for 24 h to fully protonate the sulfonate groups, and the membranes were rinsed with deionized water repeatedly to remove excess acid. The samples were stored in deionized water before measurement. Dilute solutions of sulfonated poly(phenylsulfone) were cast on Kapton films to form very thin layers (less than 5 μm) of sulfonated polymers for zeta potential measurement. Characterization of Zeta Potential. A SurPASS (Anton Paar GmbH, Graz, Austria) electrokinetic analyzer was used to measure the streaming current or streaming potential as a function of the applied channel pressure for the sulfonated poly(phenylsulfone) samples across a range of pH values.34 The slit-type channel was 2 cm in length and 1 cm in width and had a variable cell height that was set to approximately 100 μm for all experiments. Flow was induced in the measurement cell by linearly ramping the differential pressure from 0 to 300 mbar in both directions. Two cycles of pressure ramping in each direction were conducted and the average dIs/dP values for all four cycles were used to compute the zeta potential after the raw data confirmed a linear relationship between the applied pressure and the measured streaming current (Figure 3). A 0.001 M KCl solution was used as the electrolyte and HCl (0.1 M) and NaOH (0.1 M) were used to adjust the pH. Before measurements, membranes were placed under constant flow of 0.001 M KCl for at least 30 min to allow the ion exchange process to reach equilibrium. Samples were also rinsed for 3 min at each pH point before the zeta potential was measured. The zeta potential (ζ) of the membrane was computed using the Smoluchowski equation (eq 2) from the measured average dIs/dP values.
4. RESULTS AND DISCUSSION The computed zeta potentials from streaming current measurements of the sulfonated poly(phenylsulfone) membranes with thicknesses of approximately 2030 μm were negative at high pH and became less negative as pH decreased (Figure 4). This increase in zeta potential with a decrease in pH could be explained by the promoted protonation of sulfonic acid groups in a lower pH environment leading to more contact ion pairs, which did not contribute to the surface charge. A similar trend of less negative zeta potential is observed for neutral surfaces that tend to attract anions. When pH decreases, the absorption of cations is enhanced; thus, a less negative surface charge is acquired. At pH 5, the zeta potentials computed directly from the measured streaming current for each sample became more negative as the ion exchange capacity (IEC) of the membrane increased up to 1.1 mmol g1. When the IEC exceeded 1.1 mmol g1, the computed zeta
Figure 6. Zeta potential versus pH for (a) PSf-II (IEC: 1.1 mmol g1) and (b) PSf-IV (IEC: 2.0 mmol g1) membranes of different thicknesses.
potentials became less negative with increasing IEC (Figure 5). For surfaces without ionizable groups, the adsorption of ions from solution can result in surface charges comparable to those of surfaces with tethered ionizable groups.35 Often, anion adsorption to hydrophobic surfaces (because of the anions minimal hydration shell compared to cations) causes the surface to display a negative zeta potential. The surface adsorption of ions changes with pH, and thus, the computed zeta potentials are a function of solution pH and often solution composition. In this series of experiments on 4724
dx.doi.org/10.1021/la105120f |Langmuir 2011, 27, 4721–4727
Langmuir
ARTICLE
Figure 7. Illustration of the streaming current generated by electrolyte flow through the channel and the back flow current through the membrane bulk.
Figure 8. Measured streaming current as a function of applied pressure dIs/dP at pH 5 versus membrane thickness dm for Radel (Δ), PSf-I (b), PSf-II ()), and PSf-IV (9) membranes.
polymers with fixed ionic charges, the surfaces with low ionizable functional group density displayed more negative zeta potentials than the unfunctionalized polymer surface. It is reasonable to assume that membranes with negatively charged sulfonate groups covalently bound to the polymer should have a comparable or more negative zeta potential than their unsulfonated analogue. At low IECs, the functional group ionization on the surface of the membrane lowered the measured zeta potential compared to the unfunctionalized surface; e.g., in Figure 5, the zeta potential of the unfunctionalized surface was about 40 mV at pH 5, while the zeta potential for a the surface of a membrane with 0.6 mmol g1 was 52 mV and decreased to 72 mV for a membrane with 1.1 mmol g1 sulfonate groups. When the fixed charge density (IEC) of the membranes increased further, the conductance of the membranes had an influence on the measured streaming currents, so the computation of zeta potential directly from the streaming current was not valid, as will be discussed in the following paragraphs. Werner et al. also reported the phenomenon of decreasing zeta potential with increasing IEC in their streaming potential measurements on sulfonated poly(ethersulfone) membranes: the zeta potential for less substituted poly(ethersulfone) membranes was more negative than for materials with higher ionic substitution.18 They explained this trend by hypothesizing that the sulfonic acid groups increased the thickness of the swelling layer at the membranes’ surface and changed the location of the shear plane. Their interpretation does not explain the phenomena observed in the present experiments. First of all, the more negative zeta potential of PSf-II (IEC 1.1 mmol g1) compared to PSf-I (0.6 mmol g1) does not fit their explanation. Second, a correlation between measured zeta potential and membrane thickness was observed (Figure 6). These data cannot be rationalized by the increased thickness of the swelling layer at the
surface of the membranes, since all of the membranes were of the same material with the same membrane/solution interface. Figure 6 shows that the computed zeta potential versus pH curves trend toward zero with increasing thickness of membranes with the same composition. The hydrophilicity of sulfonic acid groups, per Werner et al.’s explanation, is not able to account for the change in zeta potential with membrane thickness, so the effect cannot be just at the surface. From the prospective of membrane conductivity throughout the thickness of the sample, the trend is clearly in accordance with the data presented in Figure 4 where samples with lower resistance (higher thickness or higher ion conductivity) display less negative zeta potential values. On the basis of these observations, instead of membrane hydrophilicity and swelling at the interface, membrane conductivity appears to be an important factor in the streaming current measurement and thus the calculation of zeta potential. In the slit microchannel geometry used, the electrolyte flow sheared away the ions in the diffusion layer of the surface and inducted a streaming current. In the case of a nonconductive channel boundary, the amount of ions driven by the hydrodynamic force to the end of the channel will be measured by drawing the current off through a low impedance electronic circuit, which shortcircuits the return path through the conducting liquid.15 The zeta potential is calculated from the measured streaming current and the dimensions of the cross section of the slit channel. However, with the sulfonated poly(phenylsulfone) membranes in this study and Werner et al.’s sulfonated poly(ethersulfone) membranes, the conductivity of the membrane bulk is comparable to or greater than the conductivity of the electrolyte in the channel. As a result, the ions driven by hydrodynamic force could flow back through the membrane due to the potential difference generated by the streaming current, leading to a smaller value of streaming current measured between the two ends of the channel as shown in Figure 7. In the case of a conductive membrane, the highest resistance path is through the electrolyte solution, so ionic current will naturally flow through the conductive membrane bulk instead of in the electrolyte channel. As shown in eq 9, the original streaming current I is reduced by the backflow current through the membrane Im, resulting in a smaller measured streaming current Is. Accordingly, the calculated zeta potential would become smaller when significant back flow of current occurs through the ion-conductive membrane. IS ¼ I 2Im
ð9Þ
Using this framework of a conductive path through the membrane, the change of zeta potential with IEC or membrane thickness is explicit. As IEC increases, membrane conductivity increases and membrane resistance decreases. Thus, more ions flow back through the membrane bulk, resulting in a smaller measured 4725
dx.doi.org/10.1021/la105120f |Langmuir 2011, 27, 4721–4727
Langmuir
ARTICLE
Table 2. Measured and Corrected Values of dIs/dP and Zeta Potential for Radel, PSf-I, PSf-II, and PSf-IV Membrane of Sample Thicknesses of 2030 μm sample
dIs/dP (mA/mbar)
ζ (mV)
corrected dIs/dP (mA/mbar)
ζc (mV)
Radel PSf-I PSf-II PSf-IV
1.8 107 2.1 107 2.6 107 1.3 107
40.9 51.6 71.8 34.6
2.0 107 2.4 107 3.5 107 4.4 107
49.7 59.6 97.7 116.7
Figure 10. Zeta potential versus pH for thin films, PSf-II (9) and PSfIV (2), on Kapton substrates.
Figure 9. Corrected zeta potentials (ζc) versus membrane IEC values.
streaming current and a smaller computed zeta potential. For membranes of the same IEC polymer, as membrane thickness increases, membrane resistance also decreases, which will decrease the streaming current accordingly. Since the surface properties do not change with membrane thickness, the backflow current, Im, must be a function of membrane thickness to explain the data shown here. Previous work has demonstrated that the streaming channel can be treated as ohmic,32 so in the simplest ohmic model for the membrane resistance, we can use a linear extrapolation of dm to determine if a parallel ohmic resistance description is sufficient for the materials and conditions examined. For the Radel and PSf-I (IEC: 0.6 mmol g1) samples, the membrane conductivities are low, and the conductivity influence on the measured streaming current is small. There is only a small decrease of measured streaming current with increasing membrane thickness for Radel and PSf-I membranes (Figure 8). For PSf-I and PSf-II membranes with thicknesses of 2030 μm, the PSf-I membrane displays a less negative zeta potential than PSfII. This result reflects the lower density of sulfonate groups in PSf-I compared to PSf-II, and their bulk conductivities are small enough not to have a large effect on the streaming current. These two samples containing negatively charged sulfonate groups displayed measured zeta potentials more negative than that of the unsulfonated membrane. The measured dIs/dP values at pH 5 were plotted as a function of sample thickness (dm) and extrapolated to zero membrane thickness (membrane resistance infinite), as shown in Figure 8. Linear regression with respect to dm will eliminate the influence of Im and result in the original streaming current without the influence of membrane conductance. The “corrected” dIs/dP values and the accordingly corrected zeta potential (ζc) values display a different trend (Table 2) than was shown previously in Figure 5. For the corrected results, membranes of higher IEC had
more negative zeta potentials as shown in Figure 9, which is in agreement with the expectation that higher IEC membranes contain more negatively charged functional groups and thus should display a more negative zeta potential. Additional measurements on thin (∼5 μm) sulfonated poly(phenylsulfone) films on Kapton substrates also support the thickness-corrected results obtained on thick films. A more negative zeta potential was computed from streaming current measurements on the thin-film sample with higher IEC, Figure 10. As the membrane became thin enough, the affect of membrane conductivity was minimized and the zeta potential was predominately determined by surface charge density or, by extension, the polymer IEC. These results, along with the data presented in Figure 8, demonstrate that the entire membrane resistance must be taken into account in the streaming current analysis and the membrane resistance may be treated as ohmic. Simply considering the surface hydration or other polymer/ electrolyte interfacial phenomena is not sufficient to capture the data presented here. The deviation between thin film zeta potential and the corrected zeta potential extrapolated to zero membrane thickness still shows the influence of membrane conductance, even though the effect is small in the case of thin films.
5. CONCLUSIONS Accurately characterizing zeta potential has been an important issue for investigating the charging behavior of solid surfaces in contact with aqueous solutions. While it remains difficult to perform accurate measurements on many surfaces, characterizing the zeta potential of membranes is crucial for understanding of the influence surface composition, solution pH, and ionic strength as well as other parameters imposing on the surface charge in a hydrated environment. During the zeta potential measurements of sulfonated poly(phenylsulfone) membranes, we found that membrane bulk conductivity could induce large perturbations of the measured streaming current, and thus the computed zeta potential, for ion-conductive membranes. Measuring membrane zeta potential at different membrane thicknesses and using a parallel ohmic resistance model to eliminate the influence of bulk conductivity has proven to be a feasible route to obtaining results that reflect the membrane composition. 4726
dx.doi.org/10.1021/la105120f |Langmuir 2011, 27, 4721–4727
Langmuir
’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected], Tel: 814-933-2204, FAX: 814865-2917.
ARTICLE
(33) Dyck, A.; Fritsch, D.; Nunes, S. P. J. Appl. Polym. Sci. 2002, 86, 2820–2827. (34) Luxbacher, T. Desalination 2006, 199, 376–377. (35) Zimmermann, R.; Freudenberg, U.; Schweiss, R.; K€uttner, D.; Werner, C. Curr. Opin. Colloid Interface Sci. 2010, 15, 196–202.
’ ACKNOWLEDGMENT The authors acknowledge the support of the U.S. Office of Naval Research through grants N00014-08-1-0730 and N0001410-1-0875. The Penn State Materials Research Institute and Penn State Institutes of Energy and the Environment are acknowledged for infrastructure support. ’ REFERENCES (1) Greenlee, L. F.; Lawler, D. F.; Freeman, B. D.; Marrot, B.; Moulin, P. Water Res. 2009, 43, 2317–2348. (2) Fritzmann, C.; Lowenberg, J.; Wintgens, T.; Melin, T. Desalination 2007, 216, 1–76. (3) Childress, A. E.; Elimelech, M. J. Membr. Sci. 1996, 119, 253–268. (4) Zularisam, a; Ismail, a; Salim, R. Desalination 2006, 194, 211–231. (5) Al-Amoudi, A. S. Desalination. 259, 1-10. (6) Hong, S. K.; Elimelech, M. J. Membr. Sci. 1997, 132, 159–181. (7) Yang, J.; Lee, S.; Lee, E.; Lee, J.; Hong, S. Desalination 2009, 247, 148–161. (8) Zhu, X. H.; Elimelech, M. Environ. Sci. Technol. 1997, 31, 3654–3662. (9) Schaep, J.; Van der Bruggen, B.; Vandecasteele, C.; Wilms, D. Sep. Purif. Technol. 1998, 14, 155–162. (10) Elimelech, M.; Chen, W. H.; Waypa, J. J. Desalination 1994, 95, 269–286. (11) Childress, A. E.; Elimelech, M. Environ. Sci. Technol. 2000, 34, 3710–3716. (12) Childress, A. E.; Deshmukh, S. S. Science 1998, 118, 167–174. (13) Peeters, J. M. M.; Mulder, M. H. V.; Strathmann, H. Colloids Surf., A. 1999, 150, 247–259. (14) Ernst, M.; Bismarck, A.; Springer, J.; Jekel, M. J. Membr. Sci. 2000, 165, 251–259. (15) Hunter, R. J. Zeta Potentials in Colloid Science: Principles and Applications; Academic Press: London, 1988. (16) Guiver, M. D.; Croteau, S.; Hazlett, J. D.; Kutowy, O. Br. Polym. J. 1990, 23, 29–39. (17) Hinke, E.; Staude, E. J. Appl. Polym. Sci. 1991, 42, 2951–2958. (18) Werner, C.; Jacobasch, H. J.; Reichelt, G. Biomater. Sci. 1995, 7, 61–76. (19) Mockel, D.; Staude, E.; Dal-Cin, M.; Darcovich, K.; Guiver, M. J. Membr. Sci. 1998, 145, 211–222. (20) Probstein, R. F. Physicochemical Hydrohynamics; Wiley-Interscience, 1994. (21) Stern, O. Z. Elektrochem. Angew. P. 1924, 30, 508–516. (22) Lin, J.-L.; Lee, K.-H.; Lee, G.-B. J. Micromech. Microeng. 2006, 16, 757–768. (23) Wu, Z.; Li, D. Electrochimi. Acta. 2008, 53, 5827–5835. (24) Erickson, D.; Li, D. Q. Langmuir 2002, 18, 1883–1892. (25) Reiber, H.; K€oller, T.; Palberg, T.; Carrique, F.; Ruiz Reina, E.; Piazza, R. J. Colloid Interface Sci. 2007, 309, 315–22. (26) Jodar-Reyes, A. B.; Ortega-Vinuesa, J. L.; Martín-Rodríguez, A. J. Colloid Interface Sci. 2006, 297, 170–81. (27) Delgado, A. V.; Gonzalez-Caballero, F.; Hunter, R. J.; Koopal, L. K.; Lyklema, J. J. Colloid Interface Sci. 2007, 309, 194–224. (28) Dukhin, A. S.; Ohshima, H.; Shilov, V. N.; Goetz, P. J. Electrophoresis 1999, 3445–3451. (29) Laboratories, B. T.; Hill, M. Chem. Phys. 1961, 35, 1584–1589. (30) Fairbrother, F.; Mastin, H. J. Chem. Soc. 1924, 125, 2319–2330. (31) Yaroshchuk, A.; Ribitsch, V. Langmuir 2002, 18, 2036–2038. (32) Yaroshchuk, A.; Luxbacher, T. Langmuir 2010, 26, 10882–10889. 4727
dx.doi.org/10.1021/la105120f |Langmuir 2011, 27, 4721–4727
