Engineered Slippery Surface to Mitigate Gypsum Scaling in

Article pubs.acs.org/est

Cite This: Environ. Sci. Technol. 2018, 52, 14362−14370

Engineered Slippery Surface to Mitigate Gypsum Scaling in Membrane Distillation for Treatment of Hypersaline Industrial Wastewaters Vasiliki Karanikola,† Chanhee Boo,† Julianne Rolf, and Menachem Elimelech* Department of Chemical and Environmental Engineering, Yale University, New Haven, Connecticut 06520-8286, United States

Environ. Sci. Technol. 2018.52:14362-14370. Downloaded from pubs.acs.org by WESTERN SYDNEY UNIV on 01/13/19. For personal use only.

S Supporting Information *

ABSTRACT: Membrane distillation (MD) is an emerging thermal desalination process, which can potentially treat high salinity industrial wastewaters, such as shale gas produced water and power plant blowdown. The performance of MD systems is hampered by inorganic scaling, particularly when treating hypersaline industrial wastewaters with high-scaling potential. In this study, we developed a scaling-resistant MD membrane with an engineered “slippery” surface for desalination of high-salinity industrial wastewaters at high water recovery. A polyvinylidene fluoride (PVDF) membrane was grafted with silica nanoparticles, followed by coating with fluoroalkylsilane to lower the membrane surface energy. Contact angle measurements revealed the “slippery” nature of the modified PVDF membrane. We evaluated the desalination performance of the surface-engineered PVDF membrane in direct contact membrane distillation using a synthetic wastewater with high gypsum scaling potential as well as a brine from a power plant blowdown. Results show that gypsum scaling is substantially delayed on the developed slippery surface. Compared to the pristine PVDF membrane, the modified PVDF membranes exhibited a stable MD performance with reduced scaling potential, demonstrating its potential to achieve high water recovery in treatment of high-salinity industrial wastewaters. We conclude with a discussion of the mechanism for gypsum scaling inhibition by the engineered slippery surface.

■

cooling water.5 In addition, many power plants are designed to use reclaimed municipal wastewater effluent or surface water as cooling process water, which contain diverse inorganic and organic matter as well as microorganisms.6 The concentrated cooling water contains high concentrations of total dissolved solids (TDS) and chemical additives. Hence, wastewater from a power plant is challenging to treat by conventional water treatment practices. Membrane distillation (MD) is a thermal separation process driven by a vapor gradient resulting from a temperature gradient created across a hydrophobic microporous membrane.7 Theoretically, MD can achieve complete salt rejection, owing to its phase-change based desalination mechanism. Furthermore, performance of MD is only slightly sensitive to the feed salinity, because the vapor pressure of feed solution does not change much with salt concentration.8 Nonetheless, MD is energetically intensive as any thermal separation process.9 Energy-efficient MD operation is possible if abundant waste or low-grade heat is readily available.10 The waste heat produced from a power plant cooling tower can be exploited to drive the MD process. With its effective incorporation into the thermoelectric power plant scheme, MD can provide

INTRODUCTION Thermoelectric power plants use a substantial amount of water, which accounts for approximately 40% of freshwater withdrawals in the United States.1−3 There is a critical need to impose stringent discharge regulations on power plants to reduce their freshwater consumption, wastewater discharge, and environmental impact. The U.S. Environmental Protection Agency (EPA) established a regulatory limit of wastewater discharge from power plants to the environment to be reduced by ∼57 billion gallons per year.4 Brine discharge from thermoelectric power plants is strictly prohibited in regions with limited freshwater supplies. In such conditions, the facilities need to achieve zero liquid discharge (ZLD) for waste (brine) streams. Sustainable management of wastewaters from steam electric power generation is of critical importance at the water-energy nexus. Steam-driven electric power plants require water in two processes: one as the means to generate steam to drive a turbine and one as the cooling water to remove heat from the condensed steam and the system components;5 the latter accounts for the majority of the water consumption. During cooling tower operation, the process water is continuously concentrated up to eight times due to evaporation, environmental losses, and infrastructure leakage,6 which causes scaling and corrosion of the system. To prevent scaling and corrosion, many chemicals, including biocides, biodispersants, corrosion inhibitors, pH adjusters, and antiscalants are added to the © 2018 American Chemical Society

Received: Revised: Accepted: Published: 14362

August 28, 2018 November 6, 2018 November 14, 2018 November 14, 2018 DOI: 10.1021/acs.est.8b04836 Environ. Sci. Technol. 2018, 52, 14362−14370

Article

Environmental Science & Technology

0.45 μm and an average thickness of 125 μm (HVHP, Millipore).20 The first step was to immerse the PVDF substrate into a 7.5 M NaOH solution at ∼70 °C for 3 h to functionalize the surface with hydroxyl groups. The alkaline-treated PVDF membrane was then rinsed with DI water and oven-dried at 80 °C for 1 h. The dried membranes were subsequently placed in 1% v/v APTES in ethanol under continuous stirring for 1 h. APTES covalently binds with the hydroxyl groups to produce amine terminal groups, rendering the membrane surface positively charged. Then, the amine-functionalized PVDF substrate was immersed in an aqueous SiNP suspension for 1 h under gentle mixing. SiNPs are negatively charged and bound electrostatically to the positively charged PVDF membrane surface. The aqueous SiNP suspension was prepared by adding 1 wt % SiNPs in acetate buffer with an ionic strength of ∼1 mM. The pH of the SiNP suspension was adjusted to 4 to promote effective electrostatic attraction between the negatively charged SiNPs and the positively charged amine terminal groups on the PVDF substrate. Before use, the SiNP suspension was bath sonicated for 30 min to minimize particle aggregation. Finally, the SiNP-coated PVDF substrate was functionalized with fluoroalkylsilane via covalent bonding (i.e., (heptadecafluorotetrahydrodecyl)trichlorosilane, hereafter denoted as 17-FAS) to lower the membrane surface energy via vapor phase silanization for 12 h under vacuum and heating at 70 °C. Membrane Characterization. Surface morphology of the pristine and modified PVDF membranes was examined by scanning electron microscopy (SEM, Hitachi SU-70). Before imaging, membrane samples were sputter-coated with a 4 nm iridium layer (BTT-IV, Denton Vacuum, LLC, Moorestown, NJ). The SEM images were obtained to confirm SiNP coating after surface modification and to observe scalants formed on the membrane surface after MD scaling experiments. Scaling of MD membranes on the surface and inside the pores was further analyzed by elemental mapping using SEM equipped with an energy dispersive X-ray spectroscopy (EDS, Bruker XFlash 5060FQ Annular detector). The scaled membrane samples were sputter-coated with a 4 nm iridium layer (BTTIV, Denton Vacuum, LLC, Moorestown, NJ) to eliminate charging during EDS analysis. We note that EDS elemental analysis is not affected by the iridium coating as it has binding energy far from the elements expected on our samples. The elemental composition of the pristine and surface modified PVDF membranes was analyzed by X-ray photoelectron spectroscopy (XPS, PHI VersaProbe II, Physical Electronics Inc., MN). Membrane samples were oven-dried at 70 °C overnight prior to performing XPS. The XPS spectra were collected using monochromatic 1486.7 eV Al Kα X-ray source with a 0.47 eV system resolution. The energy scale was calibrated using Cu 2p3/2 (932.67 eV) and Au 4f7/2 (84.00 eV) peaks on a clean copper plate and a clean gold foil, respectively. Static contact angle and dynamic surface wettability of the PVDF membranes with hydrophobic, hydrophilic, and slippery surface were measured by a contact angle goniometer (OneAttension, Biolin scientific instrument) using the sessile drop method. Static water contact angles were measured by placing a 5-μL droplet on the membrane surface. The shape of water droplet was photographed for 30 s using a digital camera. The digital images obtained were analyzed by a postprocessing software (OneAttension software), and the data were averaged. Dynamic surface wettability was tested by dropping a 10-μL

sustainable process water management by achieving ZLD for waste streams.11 Membrane scaling is of critical concern when treatment of high-salinity industrial wastewater at high water recovery is desired, such as desalination of power plant blowdown.12,13 At higher MD system recovery, TDS concentrations in the feed increase to supersaturated conditions.14 Nucleation of the crystals in the bulk solution and subsequent crystal deposition on the membrane surface, as well as heterogeneous nucleation of the crystals on the membrane surface, take place under such supersaturated conditions.15,16 The scaling mechanisms in MD are closely related to the membrane surface properties, feed solution chemistry, and the system hydrodynamic conditions.15,17 Scaling leads to pore wetting and to a subsequent decreasee in the desalination performance of MD membranes.18 In addition, the scaling layer aggravates temperature and concentration polarization by reducing the effectiveness of heat and mass transfer near the membrane surface.19 Recent studies have shown that engineering the surface with special wettability can enhance MD membrane performance.20−22 Polyvinylidene fluoride (PVDF) is one of the commercially available hydrophobic membranes widely used for MD applications.23−26 However, in many cases, conventional PVDF membranes are susceptible to scaling.27−29 The nonadhesive surface nature adapted from a lotus leaf can potentially be leveraged to fabricate antiscaling MD membranes.28,30−32 Coating a hydrophobic substrate with ultralow interfacial energy material can create a surface with unique properties, including a significantly high water contact angle, low contact angle hysteresis, and nonadhesive slipperiness. Such unique surface properties are expected to reduce scaling potential of MD membranes by limiting energetically favorable regions for crystal growth.33−36 In this study, we demonstrate the potential application of MD membranes with superhydrophobic, slippery surface properties in treatment of high-salinity industrial wastewaters at high water recovery. MD membranes with a slippery surface were prepared via facile and scalable surface engineering of hydrophobic PVDF membranes. Water flux and salt rejection performance of the slippery membrane were evaluated and compared to membranes with hydrophobic and hydrophilic surfaces in direct contact membrane distillation, using supersaturated gypsum feedwater. We further evaluated MD performance of the slippery-surface membrane in the desalination of power plant blowdown wastewater. On the basis of the observed results, we discuss possible mechanisms for reduced scaling potential of the engineered slippery surface.

■

MATERIALS AND METHODS Materials and Chemicals. ACS grade sodium hydroxide (NaOH, J.T. Baker), (3-Aminopropyl)triethoxysilane (99%, APTES, Sigma-Aldrich), anhydrous ethanol (100%, Decon Laboratories, Inc., PA), silica nanoparticles (SiNPs, Ludox SM, avg. diameter 8 nm, 30%, Sigma−Aldrich), and (heptadecafluoro-1,1,2,2-tetrahydrodecyl)trichlorosilane (FDTS, Gelest Inc., PA) were used for the surface modification of the polyvinylidene fluoride substrate. Supersaturated gypsum (CaSO4·2H2O) solution was prepared by mixing calcium chloride dihydrate (CaCl2·2H2O, Sigma-Aldrich) and sodium sulfate (Na2SO4, Sigma-Aldrich) in deionized (DI) water. Surface Modification of PVDF Membrane. A four-step protocol was used to modify a flat sheet polyvinylidene fluoride (PVDF) hydrophobic membrane with a nominal pore size of 14363

DOI: 10.1021/acs.est.8b04836 Environ. Sci. Technol. 2018, 52, 14362−14370

Article


Figure 1. (a) Schematic illustrating the protocol of engineering a slippery surface of PVDF membrane. (i) The surface of a hydrophobic PVDF membrane is hydrolyzed by alkaline treatment with a 7.5 M NaOH solution. (ii) (3-Aminopropyl)triethoxysilane (APTES) is grafted on the hydrolyzed PVDF membrane surface. (iii) Negatively charged silica nanoparticles (SiNPs) bind electrostatically on the APTES functionalized PVDF membrane. (iv−v) SiNPs are covalently bound with (heptadecafluorotetrahydrodecyl)trichlorosilane (17-FAS) via vapor-phase silanization, creating (vi) a nonadhesive, slippery surface. (b) SEM top-down image of the SiNP-coated PVDF membrane. (c) SEM-EDS cross-section image of the SiNP-coated PVDF membrane, revealing a thin-layer SiNP surface coating. (d) XPS analysis of the hydrophobic PVDF membrane, SiNPcoated PVDF membrane (denoted as hydrophilic membrane), and modified slippery PVDF membrane. (e) Photograph of water droplet on the modified slippery PVDF membrane.

membrane area of 20.0 cm2, was used for testing. The temperatures of feed and permeate solutions were maintained at 60 °C and 20 °C, respectively. Low (8.5 cm/s) and high (17 cm/s) feed cross-flow velocities were employed to investigate the effect of hydrodynamics on the membrane scaling behavior. The water vapor flux, Jw, across the membrane was determined by measuring the increase in permeate weight over time. Concentration of salt in the permeate was measured using a calibrated conductivity meter (Oakton Instruments, Vernon Hills, IL). The electrical conductivity of the permeate solution was monitored to calculate the salt rejection.

water droplet from a 4 cm height and monitoring the movement of the droplet on the surface over time. Synthetic and Blowdown Feed Waters. DI water (1 L) at 60 and 20 °C was used as feed and permeate streams, respectively, for initial 1 h direct contact membrane distillation (DCMD) runs. Calcium chloride (1 M) and sodium sulfate (1 M) stock solutions were prepared in advance and filtered with a 0.45 μm cellulose acetate membrane filter (Corning, Tewksbury, MA). After 1 h stabilization, stock solutions were added to the feed reservoir to obtain a scaling solution comprising 20 mM CaCl2 and 20 mM Na2SO4 without pH adjustment. The gypsum (CaSO4·2H2O) saturation index (Supporting Information, SI) of this solution at 60 °C was calculated using OLI Stream Analyzer (OLI Systems, Inc., Morris Plans, NJ) as 1.1. The gypsum scaling experiments were conducted until the permeate flux nearly reached zero and the conductivity of the permeate begun to increase. Blowdown water from the cooling towers at Redhawk Power Station in Arlington, Arizona, was prefiltered through 11-μm filter (Ashless grade 44, GE Whatman, PA) to remove particulate and suspended organic matter prior to DCMD experiments. The major composition and key properties of the prefiltered blowdown water were analyzed by a third-party environmental laboratory (Environmental Service Laboratories, Inc., PA) (Table S1). The blowdown water was treated with chlorine and stored at 4 °C to prevent biological growth. The DCMD experiments with blowdown water were conducted until the maximum amount of feed solution (initial volume of 1 L) was recovered before the conductivity of the permeate started to increase  an indication of pore wetting. Membrane Distillation Scaling Experiments. We evaluated the scaling behavior of the pristine and surfacemodified PVDF membranes using a laboratory-scale direct contact membrane distillation (DCMD) unit. A custom-made acrylic cell with channel dimensions of 77 mm in length, 26 mm in width, and 3 mm in depth, translating to an effective

■

RESULTS AND DISCUSSION Properties of Slippery Surface Membrane. Figure 1a schematically illustrates the steps to synthesize a slippery surface on a hydrophobic PVDF substrate. The PVDF substrate is first treated with an NaOH (7.5 M) solution at 70 °C to functionalize the surface with hydroxyl groups. During this process, simultaneous defluorination and oxygenation take place on PVDF polymer chains by substitution of fluorine into hydroxyl moieties.37 The NaOH-treated PVDF membrane remained mechanically robust and flexible as shown in Figure S1-A, and no morphological differences were observed in SEM images of the PVDF membranes before and after alkaline treatment (Figures S1-B and -C). The alkaline treated PVDF substrate is then aminosilanized (APTES 1% v/v in ethanol) to render the surface positively charged. APTES binds covalently with the hydroxyl groups on the membrane and serves as a substrate for SiNP deposition onto the membrane surface. The negatively charged SiNPs are electrostatically bound with the amine-functionalized substrate surface. Finally, the PVDF membrane with attached SiNPs is vapor silanized with 17-FAS through a hydrolysis−condensation reaction to produce a surface with ultralow interfacial energy.33 14364


Article


Figure 2. (a) Static contact angle of hydrophobic, hydrophilic, and slippery surface PVDF membranes measured with a 5-μL DI water droplet placed on the sample surface. (b) Dynamic surface wettability of the membranes evaluated by dispensing a 10-μL DI water droplet from a syringe at a 4 cm height and recording the motion of the water droplet with time.

Surface wettability and the nonstick slippery nature of the pristine and surface-modified PVDF membranes were evaluated using a contact angle goniometer with DI water as the testing liquid (Figure 2). The pristine hydrophobic surface membrane showed a relatively high static contact angle of ∼120°. As expected, the hydrophilic surface membrane exhibited the lowest static water contact angle of ∼70°, while the slippery surface membrane showed the highest hydrophobicity with a water contact angle of >150° (Figure 2a). When the pristine hydrophobic membrane was subjected to dynamic surface wettability testing, the water droplet remained in place with no bouncing (Figure 2b-1). This observation indicates that water and the surface do not strongly repel each other despite the inherent surface hydrophobicity. The hydrophilic surface membrane exhibited the lowest tendency to repel water as demonstrated by stable droplet settling on the surface during the dynamic surface wettability test (Figure 2b2). This observation indicates that the attraction between the hydrophilic surface and water is higher than the kinetic energy exerted by the bouncing droplet. In contrast, the slippery surface membrane illustrates immediate bouncing of water droplet upon contact with the engineered slippery surface (Figure 2b-3). This observation suggests that the modification produced a low energy surface where the kinetic energy of the bouncing droplet can overcome the attractive forces between the droplet and membrane surface. Gypsum Scaling Behavior. The gypsum scaling behavior of the hydrophobic, hydrophilic, and slippery surface membranes was studied through DCMD experiments. Permeate flux was measured over time as the membranes were exposed to a supersaturated gypsum feed solution saturation index (SI = 1.1) prepared by mixing 20 mM of CaCl2 and Na2SO4. We assessed the gypsum scaling potential of these membranes by monitoring the flux and electric conductivity of the permeate simultaneously in a batch DCMD mode (i.e., feed solution was concentrated over time). The effect of cross-flow rate was examined by applying two crossflow velocities (i.e., 8.5 and 17 cm/s), both in the laminar flow regime.

Scanning electron microscopy (SEM) and energy dispersive X-ray spectroscopy (EDS) were used to confirm SiNP coating on the PVDF substrate. The top-down view of the SiNPcoated membrane clearly shows a dense layer of the spherical nanoparticles on the PVDF substrate (Figure 1b). The crosssection image of SEM-EDS elemental mapping for the SiNPcoated membrane indicates that SiNP modification creates a several micrometer thick coating layer on the porous substrate (Figure 1c). We also observed SiNP coating inside the substrate pores, consistent with our previous studies.33 The internal pore functionalization is expected to increase the resistance to pore wetting and potentially reduce scaling potential, as we discuss later in this work. XPS spectra for the pristine PVDF membrane (denoted as hydrophobic), SiNP-coated PVDF membrane (denoted as hydrophilic), and SiNP and 17-FAS functionalized PVDF membrane (denoted as slippery) verify the substrate coating with SiNPs and surface fluorination with 17-FAS (Figure 1d). The pristine hydrophobic PVDF membrane exhibits two major peaks at 289 and 285 eV, which are assigned to the binding energies of carbon to fluorine (CF2) and carbon to hydrogen (CH2), respectively.38 The peak intensities for CH2 and CF2 are nearly equal, consistent with the chemistry of PVDF. The hydrophilic surface membrane exhibits binding energy peaks identical to those of the pristine hydrophobic membrane; however, their intensities are lower because the SiNP coating covers the surface elemental composition of the native PVDF. Instead, we observed strong XPS spectra relevant to oxygen (O) and silicon (Si) for the SiNP-coated PVDF membrane (Figure S2). After functionalization with 17-FAS, the slippery membrane exhibits a slightly different spectrum, which displays an energy peak at 293 eV ascribed to the terminal CF3 group of 17-FAS chain. We also observed a shift of the CF2 peak to 291 eV due to the formation of CF2−CF2 after surface 17-FAS functionalization. The obtained XPS spectrum corroborates successful surface modification of the pristine hydrophobic substrate to produce membranes with an engineered slippery surface that provides a high water repellency (Figure 1e). 14365


Article

Environmental Science & Technology Figure 3a illustrates the water flux decline (expressed as normalized permeate flux) and permeate conductivity of the

engineered slippery surface resists gypsum scaling better than those with hydrophobic and hydrophilic surfaces. We further performed the gypsum scaling experiments at a higher feed cross-flow velocity (17 cm/s) to evaluate the effect of hydrodynamics on the scaling behavior of membranes with different degrees of surface wettability (Figure 3b). The high cross-flow velocity reduced the gypsum scaling potential of the hydrophobic and slippery surface membranes but not the hydrophilic surface membrane. The reduced gypsum scaling potential at a higher feed cross-flow velocity is attributed to better mixing, which in turn reduces the buildup of scale near the membrane surface.40 Interestingly, the gypsum scaling potential of the hydrophilic surface membrane is likely independent of the hydrodynamics as this membrane exhibits similar flux decline rate at the two different feed cross-flow velocities. The hydrophilic surface is expected to strongly attract the hydrated gypsum scalants to form a strong hydration layer by increasing hydrogen bonding interactions with water molecules.41 Once the gypsum scalants are deposited on the hydrophilic surface they are harder to remove by shear force (i.e., higher cross-flow) compared with the hydrophobic and slippery surfaces that provide lower interaction with the scalants.42 Thus, cross-flow velocity has less influence on sweeping away the deposited gypsum scalants from the hydrophilic surface membrane. To better understand the effect of surface wettability on gypsum scaling behaviors of MD membranes, we investigated the gypsum scalants formed on the membrane surface at two stages of scaling. The first is identified as the early stage of scaling, which occurs just before the flux started to decline (points b-1 and b-3 in Figure 4a), at approximately 150 mL cumulative permeate volume. The second is identified as the long-term stage which occurs when the permeate flux reached nearly zero and simultaneously the conductivity of the permeate began to rapidly increase (points b-2 and b-4 in Figure 4a). The membranes were taken out of the cell as is and examined under SEM. Figure 4b shows the SEM images of the membrane surface after each scaling experiment. The slippery surface membrane taken at the early scaling stage exhibited a clean surface without discernible nucleation of gypsum crystals (Figure 4b-3), while rosette-like embryo gypsum crystals were found on the hydrophobic surface membrane (Figure 4b-1), a manifestation of heterogeneous surface nucleation.43,44 SEM images of both the hydrophobic and slippery surface membranes obtained after long-term scaling experiments showed that the surfaces were substantially covered with needle-like crystals resulting from bulk deposition of gypsum scalants (Figures 4b-2 and 4b-4).45 The slippery surface membrane, however, achieved a higher water recovery compared to the hydrophobic surface membrane before a significant water flux decline appeared, suggesting a retardation of bulk crystal deposition on this membrane. Such an enhanced resistance to deposition of gypsum crystals further reduced the potential of pore wetting of the slippery surface membrane as demonstrated by the delay of conductivity increase in the permeate (Figure 4a). Desalination Performance with a Blowdown Wastewater. We conducted DCMD experiments with a brine collected from the cooling tower at the Redhawk Power Station (Phoenix, AZ) to evaluate the desalination performance of the pristine hydrophobic and modified slippery surface membranes. Major composition and key properties of the prefiltered blowdown water are given in the SI (Table S1).

Figure 3. Normalized water flux and permeate conductivity of the hydrophobic, hydrophilic, and slippery surface PVDF membranes obtained during MD scaling experiments at (a) low feed cross-flow velocity (8.5 cm/s) and (b) high feed cross-flow velocity (17 cm/s). The feed solution contained 20 mM CaCl2 and 20 mM Na2SO4, with a gypsum (CaSO4·2H2O) saturation index (SI) of 1.1, pH 7.08, and temperature of 60 °C. An initial volume of 1 L was employed for both feed and permeate solutions. Temperature of permeate solution was maintained at 20 °C. The average initial water flux was 18.0, 18.5, and 14.0 L m−2 h−1 at a low feed cross-flow velocity and 21.0, 22.6, and 15.5 L m−2 h−1 at a high feed cross-flow velocity for the hydrophobic, hydrophilic, and the slippery surface membranes, respectively.

hydrophobic, hydrophilic, and slippery surface membranes obtained during MD scaling experiments at a low feed crossflow velocity (8.5 cm/s). The hydrophobic and hydrophilic surface membranes exhibited similar flux decline behavior, with the hydrophilic membrane allowing for a slightly higher water recovery (∼12.5%) before a rapid flux decline. With the sudden drop of permeate flux, we observed a sharp increase of permeate conductivity, which is attributable to the onset of scaling-induced pore wetting.39 The slippery surface membrane outperformed both the hydrophobic and hydrophilic surface membranes by delaying both the decline in permeate water flux and increase of permeate conductivity. This observation clearly indicates that the MD membrane with an 14366


Article


Figure 4. (a) Normalized water flux and permeate conductivity of hydrophobic and slippery surface PVDF membranes obtained during scaling experiments at low feed cross-flow velocity (8.5 cm/s). The feed solution contained 20 mM CaCl2 and 20 mM Na2SO4, with a gypsum (CaSO4· 2H2O) saturation index (SI) of 1.1, pH 7.08, and temperature of 60 °C. The average initial water flux under these conditions was 18.0 and 14.0 L m−2 h−1 for the hydrophobic and the slippery surface membranes, respectively. Long-term scaling experiments were conducted until a significant decline in permeate flux was observed. Independent scaling experiments were conducted until the cumulative permeate volume reached 150 mL (indicated as “Early Terminated”). (b) SEM images of the membrane surface after long-term and early terminated gypsum scaling experiments.

Figure 5. (a) Normalized water flux and permeate conductivity of the hydrophobic and slippery surface PVDF membranes as a function of water recovery obtained during DCMD experiments with an industrial blowdown wastewater. Feed cross-flow velocity of 8.5 cm/s and temperature of 60 °C were employed. The average initial water flux under these conditions was 18.0 and 10.7 L m−2 h−1 for the hydrophobic and the slippery surface membranes, respectively. (b) Images of SEM-EDS elemental mapping for cross-section and top surface of the scaled membranes after DCMD experiments with the blowdown wastewater. Chemical composition of the blowdown wastewater is presented in Table S1.

water recovery due to limitations of the experimental DCMC bench-scale setup. Nonetheless, the results demonstrate that membranes with an engineered slippery surface increase water recovery significantly compared to the conventional hydrophobic surface membrane in treatment of saline power plant blowdown with high-scaling potential. To explain why the slippery surface membrane outperformed the hydrophobic surface membrane, SEM imaging (Figure S3) and SEM-EDS elemental mapping (Figure 5b) were conducted for the cross-section and top surface of the membranes after long-term scaling experiments with the blowdown wastewater. Only representative crystal elements expected to form on the membrane were examined. We also note that the green signal represents fluorine (F), which is the main elemental composition of the PVDF substrate. The SEMEDS analysis showed high intensity of Na and Ca on the surface of the hydrophobic membrane (Figure 5b-2) possibly due to accumulation of Na2SO4 and CaSO4 crystals. We also detected a wide distribution of Na inside the pores of the

Redhawk’s blowdown has a neutral pH (∼7.1) and a high conductivity (23.9 mS/cm) attributed to the high concentrations of various inorganic salts. In particular, relatively high levels of calcium (620 mg/L) and sulfate (6390 mg/L) were measured, implying that the blowdown water has a significant gypsum scaling potential. Flux decline curves expressed as normalized water flux and permeate conductivity as a function of water recovery were obtained from DCMD experiments with the Redhawk’s blowdown as the feed solution (Figure 5a). A rapid increase in permeate conductivity, which indicates a failure of MD desalination capacity by pore wetting, was observed at ∼40% of water recovery for the hydrophobic surface membrane. At this point, the hydrophobic surface membrane experienced ∼80% reduction in water flux compared to the initial performance (Figure 5a). In contrast, no substantial increase of permeate conductivity was observed for the slippery surface membrane until water recovery reached ∼60%. Note that we were not able to continue DCMD experiments beyond this 14367


Article


Figure 6. (a) Gibbs free energy of gypsum crystals (CaSO4·2H2O) as a function of surface porosity and contact angle which corresponds to the energy that the gypsum crystal must overcome to start forming heterogeneous crystallization on a surface (calculated based on classical nucleation theory). (b) A diagram illustrating the effect of surface wettability on gypsum crystal deposition on the MD membrane. θ1, θ2, and θ3 indicate the static water contact angles of the hydrophilic, hydrophobic, and slippery surface membranes measured in Figure 2a.

porosity provides a higher energetic barrier, which translates to lower scaling potential. We previously observed a relatively clean surface with no discernible gypsum crystals on the slippery surface membrane (Figure 4b-3) while the hydrophobic surface membrane showed sparse growth of rosette-like embryo crystals at the earlier stage of scaling experiments (Figure 4b-1). Provided that the reduction of the membrane surface porosity after a thin layer coating with relatively small SiNPs (∼8 nm) is insignificant, the obtained lower heterogeneous scaling potential of the slippery surface membrane compared to the hydrophobic surface membrane is attributed to its higher surface hydrophobicity (or lower surface energy shown in Figure 2a), which develops a larger energetic barrier for surface gypsum crystallization. Given the use of supersaturated gypsum feed solution (SI = 1.1) during DCMD experiments, deposition of scalant crystals on the membrane surface (bulk deposition) is expected to be an important mechanism governing MD membrane scaling behavior.51 Previous studies claim that a rapid flux decline with concomitant increase in permeate conductivity during MD scaling experiments are mainly attributed to bulk crystal deposition on the membrane surface followed by pore wetting.52 Figure 6b illustrates the impact of surface wettability on the contact between the gypsum crystals and the membrane surface. MD membranes with an engineered slippery surface make minimal contact with the scaling solution due to their high surface hydrophobicity, and thus allowing limited gypsum crystal deposition on the surface. The nonstick, slippery surface nature (Figure 2b-3) further reduces the chance of crystal accumulation on these membranes. In contrast, a moderately hydrophilic surface (i.e., water contact angle, θ ≈ 70°) and a conventional hydrophobic surface (i.e., θ ≈ 120°) form much more intimate contact with the scaling solution compared to the slippery surface (i.e., water contact angle, θ ≈ 150°) (Figure 6b), resulting in an increased probability of gypsum crystals deposition or heterogeneous nucleation on the membrane surface. In summary, we demonstrated the fabrication of MD membranes with an engineered slippery surface that provides improved resistance to scaling. The modified slippery surface membrane exhibited a stable MD performance with high water recovery of a supersaturated gypsum scaling solution as well as brine from a power plant blowdown. On the basis of the results obtained, we discussed the mechanisms responsible for the reduced scaling potential of the slippery surface for further

hydrophobic surface membrane, which suggests membrane pore wetting (Figure 5b-1). In contrast, the slippery surface membrane showed less crystal accumulation on the surface (Figure 5b-4) and very little presence of Na and Ca inside the pores (Figure 5b-3) compared to the hydrophobic surface membrane. This observation clearly demonstrates that the slippery surface membrane provides a higher resistance to scaling than the hydrophobic surface membrane, thereby allowing less crystal deposition on the surface as well as inside the pores and, consequently, a lower potential for pore wetting. Mechanism for Reduced Scaling with Slippery Surface. Classical nucleation theory (CNT) explains the influence of surface properties on heterogeneous surface nucleation of gypsum crystal based on the thermodynamic expression that describes the Gibbs free energy required for the formation of spherical particles.46 The Gibbs free energy, ΔG, of the forming phase, assuming spherical nuclei of radius r and molecular volume υ, was calculated by an energy balance between the sum of surface excess Gibbs free energy (positive) and volume excess Gibbs free energy (negative).47 The maximum Gibbs free energy, ΔGhomogeneous * , corresponding to the energy barrier that must be overcome by the nuclei to form a crystal in the bulk solution (also referred to as homogeneous nucleation) can be calculated using the following:48 * ΔG homogeneous =

βγ 3υ2 1 (kT )2 ln 2 S

(1)

where β is the surface geometrical factor (16π/3 for spheres), γ is the precipitate surface energy (here CaSO4), υ is the molecular volume of the CaSO4 crystallizing phase, k is the Boltzmann constant, T is the temperature, and S is the supersaturation index.49 The potential of heterogeneous surface nucleation can be deduced from ΔGhomogeneous * by incorporating membrane properties, including surface hydrophobicity and porosity, using the following equation:50 * ΔG heterogeneous = É3 ÄÅ 2Ñ ÑÉÑÄÅÅ Ñ ÅÅ 1 ÅÅ (2 + cos θ)(1 − cosθ)2 ÑÑÑÅÅÅÅ1 − ε (1 + cosθ) ÑÑÑÑ * ΔG homogeneous ÑÑÅÅ ÅÅ 4 2Ñ ÑÖÅÅÇ ÅÇ (1 − cosθ) ÑÑÑÖ

(2)

where θ is the apparent contact angle between the solution and membrane and ε is the membrane surface porosity. Figure 6a presents the calculated maximum Gibbs free energy of heterogeneous nucleation, ΔGheterogeneous * , indicating that a membrane with higher hydrophobicity and lower surface 14368


Article


(4) Postponement of Certain Compliance Dates for the Effluent Limitations Guidelines and Standards for the Steam Electric Power Generating Point Source Category. In 179th ed.; EPA, Ed. Federal Register: September 18, 2017, 2017; Vol. 82, pp 43494−43500. (5) Koeman-Stein, N.; Creusen, R.; Zijlstra, M.; Groot, C.; van den Broek, W. Membrane distillation of industrial cooling tower blowdown water. Water Resources and Industry 2016, 14, 11−17. (6) Yu, X.; Yang, H.; Lei, H.; Shapiro, A. Experimental evaluation on concentrating cooling tower blowdown water by direct contact membrane distillation. Desalination 2013, 323, 134−141. (7) Khayet, M.; Matsuura, T. Membrane Distillation: Principles and Applications; Elsevier: Oxford, 2011. (8) Lawson, K. W.; Lloyd, D. R. Membrane distillation. J. Membr. Sci. 1997, 124 (1), 1−25. (9) Boo, C.; Elimelech, M. Carbon nanotubes keep up the heat. Nat. Nanotechnol. 2017, 12, 501. (10) Deshmukh, A.; Boo, C.; Karanikola, V.; Lin, S.; Straub, A. P.; Tong, T.; Warsinger, D. M.; Elimelech, M. Membrane distillation at the water-energy nexus: limits, opportunities, and challenges. Energy Environ. Sci. 2018, 11 (5), 1177−1196. (11) Xie, M.; Shon, H. K.; Gray, S. R.; Elimelech, M. Membranebased processes for wastewater nutrient recovery: Technology, challenges, and future direction. Water Res. 2016, 89, 210−221. (12) Kim, J.; Kwon, H.; Lee, S.; Lee, S.; Hong, S. Membrane distillation (MD) integrated with crystallization (MDC) for shale gas produced water (SGPW) treatment. Desalination 2017, 403, 172− 178. (13) Davenport, D. M.; Deshmukh, A.; Werber, J. R.; Elimelech, M. High-Pressure Reverse Osmosis for Energy-Efficient Hypersaline Brine Desalination: Current Status, Design Considerations, and Research Needs. Environ. Sci. Technol. Lett. 2018, 5 (8), 467−475. (14) Kim, J.; Kim, J.; Hong, S. Recovery of water and minerals from shale gas produced water by membrane distillation crystallization. Water Res. 2018, 129, 447−459. (15) Shaffer, D. L.; Tousley, M. E.; Elimelech, M. Influence of polyamide membrane surface chemistry on gypsum scaling behavior. J. Membr. Sci. 2017, 525, 249−256. (16) Tijing, L. D.; Woo, Y. C.; Choi, J.-S.; Lee, S.; Kim, S.-H.; Shon, H. K. Fouling and its control in membrane distillationA review. J. Membr. Sci. 2015, 475, 215−244. (17) Tong, T.; Zhao, S.; Boo, C.; Hashmi, S. M.; Elimelech, M. Relating Silica Scaling in Reverse Osmosis to Membrane Surface Properties. Environ. Sci. Technol. 2017, 51 (8), 4396−4406. (18) Guillen-Burrieza, E.; Thomas, R.; Mansoor, B.; Johnson, D.; Hilal, N.; Arafat, H. Effect of dry-out on the fouling of PVDF and PTFE membranes under conditions simulating intermittent seawater membrane distillation (SWMD). J. Membr. Sci. 2013, 438, 126−139. (19) Gryta, M. Fouling in direct contact membrane distillation process. J. Membr. Sci. 2008, 325 (1), 383−394. (20) Boo, C.; Lee, J.; Elimelech, M. Omniphobic polyvinylidene fluoride (PVDF) membrane for desalination of shale gas produced water by membrane distillation. Environ. Sci. Technol. 2016, 50 (22), 12275−12282. (21) Wang, Z.; Elimelech, M.; Lin, S. Environmental Applications of Interfacial Materials with Special Wettability. Environ. Sci. Technol. 2016, 50 (5), 2132−2150. (22) Ma, M.; Hill, R. M. Superhydrophobic surfaces. Curr. Opin. Colloid Interface Sci. 2006, 11 (4), 193−202. (23) Nejati, S.; Boo, C.; Osuji, C. O.; Elimelech, M. Engineering flat sheet microporous PVDF films for membrane distillation. J. Membr. Sci. 2015, 492, 355−363. (24) Khayet, M.; Matsuura, T. Preparation and Characterization of Polyvinylidene Fluoride Membranes for Membrane Distillation. Ind. Eng. Chem. Res. 2001, 40 (24), 5710−5718. (25) Woo, Y. C.; Kim, Y.; Shim, W.-G.; Tijing, L. D.; Yao, M.; Nghiem, L. D.; Choi, J.-S.; Kim, S.-H.; Shon, H. K. Graphene/PVDF flat-sheet membrane for the treatment of RO brine from coal seam gas produced water by air gap membrane distillation. J. Membr. Sci. 2016, 513, 74−84.

improvement of antiscaling properties. MD membranes that are highly resistant to scaling will be a critical component of a high recovery MD system for sustainable treatment of hypersaline industrial wastewaters with minimal production of waste (brine) streams.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.est.8b04836.

■

Composition and properties of power plant blowdown wastewater after pretreatment through an 11-μm filter (Table S1); mechanical stability and SEM imaging of the PVDF and NaOH-treated PVDF membrane (Figure S1); XPS survey spectra of the hydrophilic, hydrophobic and slippery surface membranes (Figure S2); and crosssection and top view SEM images of scaled membranes taken after an experiment with an industrial blowdown wastewater (Figure S3) (PDF)

AUTHOR INFORMATION

Corresponding Author

*Tel: +1 (203) 432-2789; e-mail: menachem.elimelech@yale. edu. ORCID

Vasiliki Karanikola: 0000-0003-3249-6517 Chanhee Boo: 0000-0003-4595-9963 Julianne Rolf: 0000-0001-6655-2690 Menachem Elimelech: 0000-0003-4186-1563 Author Contributions †

V.K. and C.B. contributed equally to this work.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This research was made possible by the postdoctoral fellowship (to V.K.) provided from the Agnese Nelms Haury Program in Environment and Social Justice and the University of Arizona. We acknowledge the supported received from National Science Foundation (NSF) Graduate Research Fellowship Program (GRFP) to J.R. (Fellowship 2016227750) and the National Science Foundation through the Engineering Research Center for Nanotechnology-Enabled Water Treatment (EEC-1449500) to C.B. Facilities used were supported by the Yale Institute of Nanoscale and Quantum Engineering (YINQE). The author also thanks the assistance of Dr. Min Li (Yale West Campus Materials Characterization Core) with the XPS measurements and SEM imaging. The characterization facilities were supported by the Yale Institute for Yale West Campus Materials Characterization Core (MCC).

■

REFERENCES

(1) Averyt, K.; Fisher, J.; Huber-Lee, A.; Lewis, A.; Macknick, J.; Madden, N.; Rogers, J.; Tellinghuisen, S. Freshwater Use by U.S. Power Plants: Electricity’s Thirst for a Precious Resource; Union of Concerned Scientists: Cambridge, MA, 2011; pp 1−62. (2) Maupin, M. A.; Kenny, J. F.; Hutson, S. S.; Lovelace, J. K.; Barber, N. L.; Linsey, K. S. Estimated Use of Water in the United States in 2010; U.S. Geological Survey: November 5, 2014; pp 1−64. (3) Mekonnen, M. M.; Hoekstra, A. Y. Four billion people facing severe water scarcity. Science advances 2016, 2 (2), e1500323. 14369


Article


(47) Clouet, E. Modeling of nucleation processes. arXiv preprint arXiv:1001.4131 2010. (48) Becker, R.; Döring, W. Kinetische behandlung der keimbildung in übersättigten dämpfen. Ann. Phys. 1935, 416 (8), 719−752. (49) Tröger, J.; Lunkwitz, K.; Bürger, W. Determination of the surface tension of microporous membranes using contact angle measurements. J. Colloid Interface Sci. 1997, 194 (2), 281−286. (50) Curcio, E.; Fontananova, E.; Di Profio, G.; Drioli, E. Influence of the structural properties of poly (vinylidene fluoride) membranes on the heterogeneous nucleation rate of protein crystals. J. Phys. Chem. B 2006, 110 (25), 12438−12445. (51) Di Profio, G.; Curcio, E.; Drioli, E. Supersaturation Control and Heterogeneous Nucleation in Membrane Crystallizers: Facts and Perspectives. Ind. Eng. Chem. Res. 2010, 49 (23), 11878−11889. (52) Gryta, M. Calcium sulphate scaling in membrane distillation process. Chemical Papers 2009, 63 (2), 146−151.

(26) Mavukkandy, M. O.; Bilad, M. R.; Kujawa, J.; Al-Gharabli, S.; Arafat, H. A. On the effect of fumed silica particles on the structure, properties and application of PVDF membranes. Sep. Purif. Technol. 2017, 187, 365−373. (27) Li, X.; García-Payo, M.; Khayet, M.; Wang, M.; Wang, X. Superhydrophobic polysulfone/polydimethylsiloxane electrospun nanofibrous membranes for water desalination by direct contact membrane distillation. J. Membr. Sci. 2017, 542, 308−319. (28) Liu, K.; Tian, Y.; Jiang, L. Bio-inspired superoleophobic and smart materials: design, fabrication, and application. Prog. Mater. Sci. 2013, 58 (4), 503−564. (29) Gryta, M.; Karakulski, K. The application of membrane distillation for the concentration of oil-water emulsions. Desalination 1999, 121 (1), 23−29. (30) Lee, Y.; Park, S. H.; Kim, K. B.; Lee, J. K. Fabrication of hierarchical structures on a polymer surface to mimic natural superhydrophobic surfaces. Adv. Mater. 2007, 19 (17), 2330−2335. (31) Nosonovsky, M.; Bhushan, B. Biomimetic superhydrophobic surfaces: multiscale approach. Nano Lett. 2007, 7 (9), 2633−2637. (32) Wong, T.-S.; Kang, S. H.; Tang, S. K. Y.; Smythe, E. J.; Hatton, B. D.; Grinthal, A.; Aizenberg, J. Bioinspired self-repairing slippery surfaces with pressure-stable omniphobicity. Nature 2011, 477, 443. (33) Lin, S.; Nejati, S.; Boo, C.; Hu, Y.; Osuji, C. O.; Elimelech, M. Omniphobic membrane for robust membrane distillation. Environ. Sci. Technol. Lett. 2014, 1 (11), 443−447. (34) Shillingford, C.; MacCallum, N.; Wong, T.-S.; Kim, P.; Aizenberg, J. Fabrics coated with lubricated nanostructures display robust omniphobicity. Nanotechnology 2014, 25 (1), 014019. (35) Dalvi, V. H.; Rossky, P. J. Molecular origins of fluorocarbon hydrophobicity. Proc. Natl. Acad. Sci. U. S. A. 2010, 107 (31), 13603− 13607. (36) Sunny, S.; Vogel, N.; Howell, C.; Vu, T. L.; Aizenberg, J. Lubricant-Infused Nanoparticulate Coatings Assembled by Layer-byLayer Deposition. Adv. Funct. Mater. 2014, 24 (42), 6658−6667. (37) Ross, G.; Watts, J.; Hill, M.; Morrissey, P. Surface modification of poly (vinylidene fluoride) by alkaline treatment1. The degradation mechanism. Polymer 2000, 41 (5), 1685−1696. (38) Shaulsky, E.; Nejati, S.; Boo, C.; Perreault, F.; Osuji, C. O.; Elimelech, M. Post-fabrication modification of electrospun nanofiber mats with polymer coating for membrane distillation applications. J. Membr. Sci. 2017, 530, 158−165. (39) Warsinger, D. M.; Swaminathan, J.; Guillen-Burrieza, E.; Arafat, H. A.; Lienhard V, J. H. Scaling and fouling in membrane distillation for desalination applications: A review. Desalination 2015, 356, 294− 313. (40) 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 (1), 245−255. (41) Tiraferri, A.; Kang, Y.; Giannelis, E. P.; Elimelech, M. Highly Hydrophilic Thin-Film Composite Forward Osmosis Membranes Functionalized with Surface-Tailored Nanoparticles. ACS Appl. Mater. Interfaces 2012, 4 (9), 5044−5053. (42) Chen, L.; Thérien-Aubin, H.; Wong, M. C. Y.; Hoek, E. M. V.; Ober, C. K. Improved antifouling properties of polymer membranes using a ‘layer-by-layer’ mediated method. J. Mater. Chem. B 2013, 1 (41), 5651−5658. (43) Lee, S.; Lee, C.-H. Effect of operating conditions on CaSO4 scale formation mechanism in nanofiltration for water softening. Water Res. 2000, 34 (15), 3854−3866. (44) Jawor, A.; Hoek, E. M. V. Effects of feed water temperature on inorganic fouling of brackish water RO membranes. Desalination 2009, 235 (1), 44−57. (45) Benecke, J.; Haas, M.; Baur, F.; Ernst, M. Investigating the development and reproducibility of heterogeneous gypsum scaling on reverse osmosis membranes using real-time membrane surface imaging. Desalination 2018, 428, 161−171. (46) Volmer, M.; Weber, Α. Keimbildung in übersättigten Gebilden. Z. Phys. Chem. 1926, 119 (1), 277−301. 14370


Engineered Slippery Surface to Mitigate Gypsum Scaling in

Recommend Documents