Cleaning strategies and membrane flux recovery on anti-fouling membranes for pressure retarded osmosis

Author’s Accepted Manuscript Cleaning Strategies and Membrane Flux Recovery on Anti-fouling Membranes for Pressure Retarded Osmosis Xue Li, Tao Cai, Gary Lee Amy, Tai-Shung Chung www.elsevier.com/locate/memsci

PII: DOI: Reference:

S0376-7388(16)31053-5 http://dx.doi.org/10.1016/j.memsci.2016.09.016 MEMSCI14736

To appear in: Journal of Membrane Science Received date: 20 July 2016 Revised date: 10 September 2016 Accepted date: 12 September 2016 Cite this article as: Xue Li, Tao Cai, Gary Lee Amy and Tai-Shung Chung, Cleaning Strategies and Membrane Flux Recovery on Anti-fouling Membranes for Pressure Retarded Osmosis, Journal of Membrane Science, http://dx.doi.org/10.1016/j.memsci.2016.09.016 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Cleaning Strategies and Membrane Flux Recovery on Anti-fouling Membranes for Pressure Retarded Osmosis Xue Lia,b, Tao Caib, Gary Lee Amya, Tai-Shung Chunga*

a

Department of Chemical & Biomolecular Engineering, National University of Singapore, Kent Ridge, Singapore 117585 b College of Chemistry and Molecular Science, Wuhan University, Wuhan, Hubei 430072, P. R. China *

Correspondence to: T. S. Chung. Tel: (65) 6516-6645; Fax: (65) 6779-1936. E-mail: [email protected] Abstract

Study of cleaning procedures to maintain the anti-fouling properties is urgently needed to sustain high osmotic power density of the fouling resistant membranes in the pressure retarded osmosis (PRO) process. Therefore, various cleaning agents were evaluated for a charged hyperbranched polyglycerol grafted thin-film composite (CHPG-TFC) membrane: (i) deionized (DI) water; (ii) a high pH alkaline solution, (iii) a low pH acid solution, and (iv) a chelating solution comprised of ethylenediaminetetraacetate (EDTA). Compared with other cleaning agents, the EDTA cleaning was more effective to restore the original surface of the CHPG-TFC membrane and sustain its osmotic power density using a real wastewater effluent as the feed. The power density was maintained at a stable range of 6.0-6.7 W/m2 in three repeated PRO tests. In comparison, a non-modified PESTFC membranes with EDTA cleaning showed a large fluctuation of power density from 3.6 to 4.8 W/m2. Instrumental analyses were conducted to reveal the physicochemical relationship between cleanings and membrane properties. The investigations confirmed the effectiveness of EDTA cleaning to mitigate the sulfate scaling, silica fouling, and calcium deposition. In summary, the EDTA cleaning imparted good recovery to the antifouling properties of the CHPG-TFC membrane and maintained its resilience to foulants in osmotic power generation by PRO. 1

Keywords Osmotic power generation, Pressure-retarded osmosis, Hyperbranched polyglycerol grafting, Anti-fouling, Chemical cleaning Introduction Demands for renewable clean energy without CO2 emission have brought pressure retarded osmosis (PRO) into global attention because there are tremendous amounts of feed streams for osmotic power generation are available. [1-4] Since the 1970s, attempts have been made to harvest osmotic power by mixing fresh water with salt water through a semi-permeable membrane. [5-10] The spontaneous permeation of water across the membrane expands the volume in the salt water compartment and increases its hydraulic pressure. As a result, one can harvest the osmotic energy from the high-pressure compartment via a hydro turbine or pressure exchanger. If PRO is integrated with seawater reverse osmosis (SWRO) and wastewater recycling plants, the mixing energy between SWRO retentates and wastewater effluents can not only offset the energy consumption for SWRO, but also mitigate the disposal and environmental issues of SWRO brine. [11, 12] Theoretically, the PRO-SWRO-wastewater hybrid system should produce a significant amount of osmotic energy due to their large difference in salinity gradient. However, membrane fouling from the wastewater stream seriously diminishes energy production as a variety of foulants are present in wastewater effluents. [13-15]

Fouling reduces membrane performance and shortens membrane life. [16-26] In PRO operations, various pretreatments of feed solutions have been studied to minimize membrane fouling. [12, 13] The current state-of-the-art PRO membranes are made of 2

porous polyethersulfone (PES) substrates, [12, 27-29] due to their good mechanical strength and easy fabrication, but the PES substrates are highly susceptible to organic and inorganic foulants. [30] On the other hand, fouling resistance in most cases is determined by the surface properties of the PRO membrane. Thus, attempts have been explored to alter their surface properties to diminish PRO fouling without sacrificing their mechanical strength. [31-35]

General strategies to control membrane fouling are either to introduce hydrophilic groups to enhance water-surface interactions or to impart charges to induce electrostatic repulsion. [36-39] The recent developed hyperbranched polyglycerol (CHPG) grafted PRO membrane is one example. [32] Its negatively charged polymer branches provide the membrane surface with a repulsive and hydrated layer. When testing it for osmotic power generation using a concentrated wastewater as the feed, this membrane exhibited good resistances against proteins and bacteria, and maintained ~65% water flux of the initial value after four runs of wastewater testing. DI water was employed to clean the fouled membrane, but its surface functional groups were still partially shielded by the foulants. As a result, its anti-fouling properties gradually deteriorated in further testing. Although the CHPG-grafting method has significantly improved membrane properties to resist fouling, cleaning is crucial to sustain its anti-fouling performance. Therefore, for long-term PRO operation, one must combine both membrane modifications and efficient cleaning methods to ensure sustainable operation. However, there are limited relevant studies [14, 40-43] reported on cleaning procedures and methods.

3

To rejuvenate and maintain the anti-fouling properties of modified membranes for osmotic power generation, the objectives of this study are to determine the effective cleaning methods for anti-fouling membranes using the CHPG-TFC membrane as an example and to reveal the mechanisms behind the chemical cleaning. Four cleaning solutions/agents were selected; namely, deionized water, sodium hydroxide (NaOH), hydrochloric acid (HCl), and ethylenediaminetetraacetate (EDTA) because they represent simple aqueous solutions with near-neutral pH (DI), high pH, low pH, or an added chelating agent, respectively. Their cleaning effects are evaluated by using a real wastewater effluent as the feed in high-pressure PRO operations. With an appropriate combination of anti-fouling effects and cleaning efficiency, the favorable properties of modified PRO membranes may substantially sustain favorable operation. This work may provide meaningful insights for the osmotic power generation using SWRO brine and wastewater effluent as the feed pair.

2. Materials and Methods 2.1 Chemicals and Materials Poly(ether sulfone) (PES) hollow fiber supports were fabricated from Radel® A PES (Solvay Advanced Polymer). Polyamide active layers were prepared by the interfacial reaction between m-phenylenediamine (MPD, Tokyo Chemical Industry, >99%) and 1,3,5-benzenetricarbonyl trichloride (TMC, Sigma-Aldrich, 98%) on the PES supports. Sodium hydroxide (Sigma-Aldrich, ≥97%), hydrochloric acid (J.T.Baker, Avantor, 36.5%-38%), and ethylenediaminetetraacetic acid (EDTA, Sigma-Aldrich, ≥98.0%) were used to prepare the cleaning solutions. Hexane (reagent grade), the solvent of TMC solution, and sodium chloride (NaCl), the salt of synthetic RO brine, were procured from

4

Merck Pte. Ltd., Singapore. All chemicals were used as received.

2.2 Hollow Fiber Membrane Fabrication and Modification The PES hollow fiber supports were fabricated by the non-solvent induced phase inversion from a PES/NMP/PEG (polyethylene glycol)/water mixed solution. [12, 28] The porosity of the PES hollow fiber supports, 68%~71%, was measured from 10 samples. Then a thin and active polyamide layer was prepared on the inner surface of the hollow fiber supports, [28] while CHPG polymers [32] were introduced to their outer surface which faces the wastewater effluent in PRO tests with the aid of polydopamine (self-polymerization from dopamine hydrochloride, Tokyo Chemical Industry, 99%). The synthesis protocol of CHPG is the same as that described in the previous publication [32]. The as-formed CHPG-TFC membranes were rinsed thoroughly and stored in deionized water (DI water, ELGA MicroMEG) before testing.

2.3 Membrane Fouling and Cleaning Protocol in PRO Processes Membrane fouling and cleaning performance were evaluated by a laboratory-scale PRO system described in our previous studies. [28, 44-46] Hollow fiber membranes were loaded in the active-facing-draw mode (i.e., the PRO mode) where the active layer faces the draw solution, a synthetic SWRO brine (0.81 mol/L NaCl at pH 5~6). A hydra cell pump (Minneapolis, MN) was used to recirculate the draw solution through the lumen side of the hollow fiber module, and a peristaltic pump (Cole-Palmer, Vernon Hills, IL) was employed to recirculate the feed solution through the shell side of the modules. A wastewater effluent from a municipal water recycling plant was used as the feed solution, with a BOD value (biochemical oxygen demand) of 7.3 ppm, COD value (chemical 5

oxygen demand) of 190 ppm, and TDS value (total dissolved solids) of 1100 ppm. The major cations and anions have been reported in our previous publication [12]. Calcium ion and sulfate ion dominated among the inorganic foulants. The loaded membranes were first stabilized at a trans-membrane pressure of 15.0 ± 1.0 bar for 0.5 h using DI water and RO brine as the feed pair. An initial permeation flux of 20 L/m2h was measured using DI water feed before the fouling testing. Then the wastewater was introduced to replace the DI water during the fouling tests. Once the accumulative permeate volume reached 50 f2 L/m2, the fouling testing was stopped. After each testing cycle, backwash with the aid of chemical cleaning was conducted immediately at 24 °C for 5 min by applying a pressure of 15.0 ± 1.0 bar to the active layer of hollow fibers; a pressurized DI water was pumped to the module lumen side at 1.2 m/s while a particular cleaning agent (0 bar) was fed to the module shell side at 0.1 m/s, either DI water, HCl solution at pH 2.8 (dosage per membrane area of 11 g/m2), NaOH solution at pH 11.8 (dosage per membrane area of 77 g/m2), or EDTA solution at pH 10.9 (dosage per membrane area of 203 g/m2). At least three testing and cleaning cycles were performed to evaluate the performance and water flux recovery of the membranes.

2.4 Membrane Characterizations The total organic carbon (TOC) in each spent cleaning solution was measured by a TOC analyzer (Analytik Jena AG, multi N/C 3100). To further investigate components in the cleaning solutions, an ultraviolet-visible (UV-Vis) spectrophotometer (Shimadzu) was employed to investigate the UV spectra of cleaning solutions before and after washing.

6

The hydrophilicity of the virgin and fouled membranes was evaluated by a Contact Angle Goniometer (Rame Hart) at 24 °C. At least 20 measurements were tested for each membrane to minimize the experimental error. Fourier transform infrared spectroscopy (FTIR, Bio-Rad FTS 135) was utilized to analyze the chemical structures of the virgin and fouled membranes over the wavenumbers of 600–2500 cm-1 under the attenuated total reflectance (ATR) mode. Chemical functionalities on the surface of fouled membranes were studied by X-ray photoelectron spectroscopy (XPS, a Kratos AXIS UltraDLD spectrometer) with a monochromatized Al Kα X-ray source (1486.71 eV, 5 mA, 15 kV).

The surface charge properties of the virgin and fouled membranes were measured by a SurPASS electrokinetic analyzer (Anton Paar, Austria). A sodium chloride solution (0.01 mol/L, 450 mL) was employed as the electrolyte solution under the neutral pH value. To analyze the zeta potential, ζ, the electrolyte solution was first autotitrated with 0.1 mol/L HCl from the neutral pH to pH 2, followed by an autotitration with 0.1 mol/L NaOH from pH 2 to 12. The surface morphology of the virgin and fouled membranes was examined by field emission scanning electronic microscopy (FESEM, JEOL JSM-6700F), and the topology was measured by atomic force microscopy (AFM, Agilent Technologies, Santa Clara, CA) under the tapping mode. 3. Results and Discussion 3.1 Fouling and Cleaning in PRO Processes The fouling behavior of the CHPG-TFC membrane and the cleaning efficiency of four cleaning agents (i.e., DI water, EDTA, NaOH at pH 11.8, and HCl at pH 2.8) under PRO operation are compared in Figure 1. Run 1 in different cases were performed under 7

identical conditions at a trans-membrane pressure of 15.0 ± 1.0 bar, thus the membrane performance of each Run 1 was almost identical. For the normalized accumulative permeate volume >0 L/m2, the membrane performance deteriorated immediately because CHPG-TFC was subjected to the wastewater effluent: The flux was sharply reduced to less than 90% of the initial value. Over the subsequent testing period, water flux slowly declined to ~75% of the initial flux. Figure 1(a) clearly shows that the anti-fouling effect of CHPG-TFC was gradually weakened when a simple DI water cleaning was employed. Its water flux decreased to less than 75% in Run 2, and subsequently dropped to less than 70% after Run 3. After cleaning by an EDTA solution at pH 10.9, the membrane performance was well recovered (Figure 1(b)). An additional two PRO runs did not further decline the water flux. After Run 3, it remained at ~75%, which was comparable to the value of Run 1. In comparison, the higher pH cleaning (NaOH at pH 11.8) adversely affected the water flux, decreasing to less than 60% after the second run, as shown in Figure 1(c), because high pH value of the cleaning agent may decrease the stability of the PDA coating and impair the structure of the hollow fiber support. Figure 1(d) shows the result of a low pH cleaning (NaCl at pH 2.8), showing insufficient effects to recover the anti-fouling resistance of the CHPG-TFC membrane, thus the water flux gradually reduced to less than 70% of the initial value. Clearly, the EDTA cleaning is more efficient among the four cleaning agents. Although the functional groups of CHPG have a combination of a hydration effect and electrostatic repulsion to foulants, their antifouling effectiveness is compromised by the adsorption of some ions and foulants existing in the wastewater. [12] Since EDTA has strong effects to chelate specific ions (e.g., calcium), scalants (e.g., CaSO4) and natural organic matter (NOM) acids (e.g.,

8

humic/fulvic) can be desorbed from the membrane surface.

The cleaning efficiency of the EDTA solution with the non-modified PES-TFC membrane (i.e., without CHPG grafting) for three PRO runs is compared in Figure 2. Unsurprisingly, a fast flux decline by ~50% of the initial value was observed prior to the EDTA cleaning. A subsequent EDTA cleaning recovered the water flux back to more than 90% of the initial value. However, the water flux inevitably decreased to 50~55% in the following wastewater cycles. The results suggest that a mere EDTA cleaning was not effective enough to maintain the performance of PES-TFC membrane without protection of an anti-fouling layer. Table 1 compares the power density of the non-modified and CHPG modified membranes with EDTA cleanings. The power density of PES-TFC decreased from 8.3 W/m2 to 4.3-4.8 W/m2 (after Run 1), then to 3.6-4.3 W/m2 (after Run 2), eventually to 4.0-4.6 W/m2 (after Run 3). On the other hand, CHPG-TFC showed a more favorable power density with the aid of EDTA cleaning. Its power density remained in the range of 6.0-6.7 W/m2 from Run 1 to Run 3, indicating that the anti-fouling effectiveness of the CHPG grafting was recovered with the aid of EDTA cleaning.

3.2 Characterization of Cleaning Solutions Table 2 compares the total organic carbon (TOC) of different cleaning agents before and after washing the fouled PES-TFC and CHPG-TFC membranes. The DI water washing on the PES-TFC membrane was used as the baseline for comparison. Originally, DI water had no organic components, but it washed 0.40 ± 0.01 mg/L TOC away from the PESTFC surface. Similarly, the DI water washed 0.25 ± 0.02 mg/L TOC away from the CHPG-TFC surface. The low pH solution removed slightly more organic compounds 9

from the CHPG-TFC membrane, it had a TOC of 0.98 ± 0.11 mg/L after washing. In contrast, the high pH cleaning washed away a TOC of 88.11 ± 7.49 mg/L from the membrane, which was higher than that in the feed wastewater (28~32 mg/L at a neutral pH value). This is because the PES substrate and anti-fouling grafting layer may be partially damaged if the cleaning solution has a pH value of 11.8 close to their limitations, leading to a high TOC value. TOC per unit area of membrane surface is compared in Table 2. DI water and low pH cleaning agents only removed 0.013~0.050 mg/cm2 TOC from the membrane surface, while the high pH solution washed away 4.488 mg/cm2 TOC from the membrane. In the case of EDTA cleaning, the TOC obtained from the membrane surface was negligible as compared to that of the bulk solution.

As shown in Figure 3, the UV-Vis spectra of cleaning solutions after washing are consistent with their TOC results. The plain water after cleaning CHPG-TFC and PESTFC membranes had very weak absorption at 190 nm, confirming very few organic and inorganic compounds were removed. Such absorption at 190 nm was amplified in the case of the low pH washing because of the n to σ* transitions of alcohols, ethers, nitrogen containing compounds, and halide compounds. The high pH washing solution had an intensive and sharp peak at 196 nm, probably due to the n to σ* transitions of nitrogen or sulfur containing compounds dissolved from the PES polymer or CHPG polymer. In addition, its broad but weak absorption from 250 to 400 nm may be attributed to the n to π* or π to π* transitions of unsaturated compounds. In the case of the EDTA washing solution, strong absorption was observed from 190 nm to 270 nm. The small redshift

10

compared to its standard UV spectrum [47] may be due to the foulants obtained from the membrane surface.

3.3 Morphology and Topology of Fouled Membranes Figure 4 illustrates the outer surface of CHPG-TFC membranes after being washed. The PES-TFC membrane cleaned with DI water is presented as a reference. In the SEM image shown in Figure 4(a), a thick fouling layer covered a large part of the PES surface. The CHPG-TFC membrane cleaned with DI water was also investigated and showed much less surface coverage because it had the anti-fouling CHPG layer (Figure 4(b)). After the EDTA washing (Figure 4(c)), only very few parts of the membrane surface were covered by foulants. In contrast, many nodules were formed on the CHPG-TFC membrane surface after the high pH cleaning (Figure 4(d)). These nodules may have resulted from the re-precipitation of the PDA after being partially dissolved in the high pH solution. For the low pH cleaned membrane, foulants were still present on the membrane surface to some extent (Figure 4(e)). These results confirm that DI water and low pH cleaning solutions modestly disrupt foulant adhesions, while the EDTA washing is the most effective way to clean the CHPG-TFC membrane.

AFM characterizations were conducted to examine the surface topology of fouled membranes. Figure 5 illustrates the AFM images over a membrane area of 10 μm ×10 μm. The reference PES-TFC membrane with DI water cleaning showed the smoothest membrane surface with a mean roughness of 25.9 ± 10.4 nm. This was due to the large coverage of foulants on membranes that diminished the concave feature of membrane pores. In comparison, a much rougher surface of 68.0 ± 32.6 nm was observed for the 11

CHPG-TFC membrane cleaned by a high pH NaOH solution, which confirmed the structure damage by the strongly alkaline condition. On the other hand, the CHPG-TFC membranes after DI water, EDTA, and low pH cleanings possessed similar roughnesses to the virgin CHPG-TFC membrane as shown in Figure S1, indicating that their fouling layers are not thick enough to affect the topological feature of the CHPG-TFC membrane.

3.4 Surface Properties of Fouled Membranes The effect of cleaning agents on the hydrophilicity of the membrane surface is illustrated in Figure 6. The surface of porous layer which faces wastewater effluent was investigated for each membrane. The virgin PES-TFC membrane possessed a water contact angle of 67.3°. Foulants attached on the porous layer of PES-TFC and made it more hydrophobic (85.2°) even after being washed with DI water, consistent with its SEM image shown in Figure 4(a). In comparison, all porous layers of CHPG-TFC membranes demonstrated a more hydrophilic nature with average values of contact angle in the range of 49-65°. Their hydrophilic nature is a result of the CHPG grafting, resisting significant adhesion of most hydrophobic components. However, a few regions of the membrane surface showed adhesion by foulants, resulting in a significant variation in the contact angle of the fouled CHPG-TFC membrane. Three cleaning solutions, i.e., DI water, EDTA and high pH solutions, effectively reduced the contact angles to ~55°. However, the CHPG-TFC membrane with low pH cleaning showed a higher contact angle value of 64.2 ± 24.0° and a larger deviation than the other three membranes because some parts of the membrane surface were covered by hydrophobic foulants.

12

XPS is a powerful analytical tool to detect surface chemistry for membranes. Figure 7(a) shows wide scan spectra of the porous layer of CHPG-TFC membranes after being washed with different solutions. All membranes had resonance peaks for C 1s, O 1s, and N 1s, attributable to the presence of the CHPG polymer and organic foulants on the membrane surface. Sulfur signals (S 2s and S 2p) were also clearly observed from the wide scan. Among four membranes, the CHPG-TFC membrane with the low pH washing exhibited the highest sulfur peaks. Its high energy resolution spectrum of S 2p (Figure 7(c)) split into two peak components, due to spin-orbit splitting of 2p3/2 and 2p1/2. The binding energy of S 2p3/2 peak was observed to be near the typical range of metal sulfate. This evidence shows that more sulfate salts were left over by using the acid washing. As shown in Figure 7(d), the intensity of S 2p and S 2s peaks in the case of CHPG-TFC with the high pH cleaning decreased as compared to that of CHPG-TFC after low pH washing. The S 2p core-level spectrum had identical signals with that in Figure 7(c) but with a lower peak height, also indicating the presence of sulfate salts on the membrane surface. On the other hand, being washed with the EDTA solution or DI water largely removed sulfate salts, and reduced their resonance intensity. Moreover, the sulfur signal of C-S functional groups of the CHPG grafting layer was evident at the binding energy of ~164 eV. In the case of CHPG-TFC washed by EDTA (Figure 7(e)), such signals were further amplified and able to be identified as doublet peaks of 2p3/2 (C-S) and 2p1/2 (C-S). That means the EDTA cleaning is the most efficient way to remove sulfate depositions, and expose the original CHPG grafting layer.

13

In addition to the sulfur element, different XPS wide scan spectra exhibited different signals for silicon and calcium elements. Silica fouling is a result of the dehydrationpolymerization of silicic acid, which is prevalent in natural water. Its initial monomer, silicic acid, is soluble at low pH value and generally deionized at natural pH value. After long-term operations, silica-based scaling is particularly difficult to remove. As shown in Figure 7(a), the CHPG-TFC washed by a low pH solution showed the most intensive Si 2p and Si 2s peaks, and that washed by DI water had the second most intensive Si peaks, followed by those washed by high pH and EDTA solutions. The significant difference in Si signals is a result of the diversity of silica-based scaling in different chemical states. The chemistry of silica-based scaling is extremely complex since silica may appear in various forms in aqueous solutions, allowing for a great diversity of products. For example, it can be either amorphous or crystalline forms of silica and silicates as well as dissolved or colloidal. The high energy resolution spectra of Si 2p were compared in Figure 7(b). The spectral resolution of spin-orbit components worsened, making 2p3/2 and 2p1/2 peaks overlap with each other. This is the evidence of more amorphous silicon existing than crystalline forms. In addition, polymerization of silicic acid is strongly influenced by pH. Its polymerization rate is very fast at neutral pH values, giving a high Si 2p peak in Figure 7(b). At low pH values of 2-3, silicic acids naturally exist in the form of weak acid and seldom polymerize. However, the occurrence of some fouling components, such as metal sulfate salts and organic compounds, may be correlated with the occurrence of silicic acids, and make the membrane surface more hydrophobic and silicon-rich. At high pH values, complex anhydrous silicate incorporates metals in the copolymerization (e.g., Ca(OH)2). The XPS results are consistent with theory, and a Ca

14

2p peak in the wide scan spectrum is evident in the spectrum of CHPG-TFC washed by the high pH solution, and absent in other cases. In summary, the XPS analyses provide more details on the characterization of foulants adhering to the CHPG membrane surface, and one can clearly distinguish that the EDTA cleaning has not only diminished sulfur deposition and silica fouling, but also eliminated scaling of some divalent cations.

Figure 8 illustrates the zeta potential on the surface of porous layer for different CHPGTFC membranes, i.e., the virgin membrane, fouled membrane, and fouled membrane after EDTA cleaning. For 2
4. Conclusions

15

Four cleaning agents (i.e., DI water, EDTA, high pH, and low pH solutions) have been employed to investigate the fouling behaviors and cleaning efficiency for the charged hyperbranched polyglycerol (CHPG) membrane used in PRO processes. The anti-fouling CHPG layer was chemically grafted onto the outer surface of PES-TFC membranes. In high pressure PRO tests, the EDTA cleaning exhibited effectiveness to maintain the water flux to be ~75% of the initial value in three testing and cleaning cycles, while other cleaning methods had larger flux declines. TOC, UV, SEM, AFM, contact angle, and zeta potential measurements were performed to characterize the cleaning solutions or membranes, suggesting that the surface properties of CHPG-TFC membranes were not substantially altered by fouling although some foulants were observed on the membrane surface. In addition, XPS measurements have verified the successful mitigation of sulfate scaling, silica fouling and some divalent cation deposition by the EDTA cleaning on the CHPG membrane surface. In summary, EDTA cleaning imparts good recovery to the fouled CHPG-TFC membrane and maintains its favorable resilience to foulants for osmotic power generation.

Acknowledgements This research grant is supported by the Singapore National Research Foundation under its Environment & Water Research Programme and administered by PUB, Singapore’s national water agency. It is funded under the projects entitled ĀMembrane Development for Osmotic Power Generation” (1102-IRIS-11-01 and 1102-IRIS-11-02) and NUS Grant No. R-279-000-381-279 and R-279-000-382-279. Special thanks are due to Mr. S. Japip,

16

Ms. Y. Zhang, Mr. Z. L. Cheng, Ms. J. Gao, Ms. W. X. Gai, Mr. T. S. Yang and Mr. C. F. Wan for their useful comments and assistance.

References 1. Thelin, W. R.; Sivertsen, E.; Holt, T.; Brekke, G. Natural organic matterfouling in pressure retarded osmosis. J Membr Sci 438 (2013) 46–56. 2. Achilli, A.; Cath, T. Y.; Childress, A. E. Power generation with pressure retarded osmosis: An experimental and theoretical investigation. J. Membr. Sci. 343 (2009) 42−52. 3. Gerstandt, K.; Peinemann, K. -V.; Skihagen, S. E.; Thorsen, T.; Holt, T. Membrane processes in energy supply for an osmotic power plant. Desalination 224 (2008) 64−70. 4. Han, G.; Zhang, S.; Li, X.; Chung, T.S. Progress in pressure retarded osmosis (PRO) membranes for osmotic power generation. Progr. Polym. Sci. 51 (2015) 1–27. 5. Loeb, S. Production of energy from concentrated brines by pressure-retarded osmosis. I. Preliminary technical and economic correlations. J. Membr. Sci. 1 (1976) 49–63. 6. Wang, Y.-N.; Wei, J.; She, Q.; Pacheco, F.; Tang, C. Y. Microscopic characterization of FO/PRO membranes – A comparative study of CLSM, TEM and SEM. Environ. Sci. Technol., 46 (2012) 9995–10003. 7. Kim, D. I.; Kim, J. W.; Shon, H. K.; Hong, S. Pressure retarded osmosis (PRO) for integrating seawater desalination and wastewater reclamation: Energy consumption and fouling. J. Membr. Sci. 483 (2015) 34−41.

17

8. Bui, N. N. and McCutcheon, J. R. Hydrophilic Nanofibers as new supports for thin film composite membranes for engineered osmosis. Environ. Sci. Technol., 47 (2013) 1761–1769. 9. Sivertsen, E.; Holt, T.; Thelin, W.; Brekke, G. Pressure retarded osmosis efficiency for different hollow fibre membrane module flow configurations. Desalination 312 (2013) 107−123. 10. Chung, T. S.; Li, X.; Ong. R. C.; Ge, Q. C.; Wang, H. L.; Han, G. Emerging forward osmosis (FO) technologies and challenges ahead for clean water and clean energy applications. Curr. Opin. Chem. Eng. 1 (2012) 246−257. 11. Chung, T. S.; Luo, L.; Wan, C. F.; Cui, Y.; Amy, L. G. What is next for forward osmosis (FO) and pressure retarded osmosis (PRO). Sep. Purif. Technol. 156 (2015) 856−860. 12. Wan, C. F. and Chung, T. S. Osmotic power generation by pressure retarded osmosis using seawater brine as the draw solution and wastewater brine as the feed. J. Membr. Sci. 479 (2015) 148−158. 13. Chen, S. C.; Amy, G. L.; Chung, T. S. Membrane fouling and anti-fouling strategies using RO retentate from a municipal water recycling plant as the feed for osmotic power generation. Water Res. 88 (2016) 144−155. 14. Yip, N. Y. and Elimelech, M. Influence of natural organic matter fouling and osmotic backwash on pressure retarded osmosis energy production from natural salinity gradients. Environ. Sci. Technol. 47 (2013) 12607−12616. 15. Liu, X.; Foo, L. X.; Li, Y.; Lee, J. Y.; Cao, B.; Tang, C. Y. Fabrication and characterization of nanocomposite pressure retarded osmosis (PRO) membranes with

18

excellent anti-biofouling property and enhanced water permeability. Desalination 389 (2016) 137−148. 16. Escobar, I. C. and Van der Bruggen, B. Modern Applications in Membrane Science and Technology, ACS Symposium Series; American Chemical Society: Washington, D.C., 2011. 17. Mi, B. X. and Elimelech, M. Organic fouling of forward osmosis membranes: Fouling reversibility and cleaning without chemical reagents. J. Membr. Sci. 348 (2010) 337−345. 18. Laabs, C. N.; Amy, G. L.; Jekel, M. Understanding the size and character of foulingcausing substances from effluent organic matter (EfOM) in low-pressure membrane filtration. Environ. Sci. Technol. 40 (2006) 4495–4499. 19. Gu, Y.; Wang, Y. N.; Wei, J.; Tang, C. Y. Organic fouling of thin-film composite polyamide and cellulose triacetate forward osmosis membranes by oppositely charged macromolecules. Water Res. 47 (2013) 1867−1874. 20. Lu, H. J.; Xue, Z.; Saikaly, P.; Nunes, S. P.; Bluver, T. R. ; Liu, W-T. Membrane biofouling in a wastewater nitrification reactor: Microbial succession from autotrophic colonization to heterotrophic domination. Water Res. 88 (2016) 337−345. 21. Wang, Y.-N. and Tang. C. Y. Fouling of nanofiltration, reverse osmosis, and ultrafiltration membranes by protein mixtures: The role of inter-foulant-species interaction. Environ. Sci. Technol. 45 (2011) 6373–6379. 22. Kim, Y.; Elimelech, M.; Shon, H. K.; Hong, S. Combined organic and colloidal fouling in forward osmosis: Fouling reversibility and the role of applied pressure. J. Membr. Sci. 460 (2014) 206−212.

19

23. Mi, B. X. and Elimelech, M. Gypsum scaling and cleaning in forward osmosis: Measurements and mechanisms. Environ. Sci. Technol. 44 (2010) 2022−2028. 24. Verliefde, A. R. D. ; Cornelissen, E. R.; Heijman, S. G. J.; Petrinic, I.; Luxbacher, T. ;Amy, G. L.; Van der Bruggen, B. ; van Dijk, J. C. Influence of membrane fouling by (pretreated) surface water on rejection of pharmaceutically active compounds (PhACs) by nanofiltration membranes. J. Membr. Sci. 330 (2009) 90−103. 25. Xie, Y. H.; Tayouo, R.; Nunes, S. P. Low fouling polysulfone ultrafiltration membrane via click chemistry. J. Appl. Polym. Sci. 132 (2015) 41549. 26. Gullinkala, T. and Escobar, I. A green membrane functionalization method to decrease natural organic matter fouling. J. Membr. Sci. 360 (2010) 155−164. 27. Chou, S.; Wang, R.; Shi, L.; She, Q.; Tang, C.Y.; Fane, A.G. Thin-film composite hollow fiber membranes for pressure retarded osmosis (PRO) process with high power density. J. Membr. Sci. 389 (2012) 25–33. 28. Sukitpaneenit, P. and Chung, T. S. High Performance thin-film composite forward osmosis hollow fiber membranes with macrovoid-free and highly porous structure for sustainable water production. Environ. Sci. Technol. 46 (2012) 7358–7365. 29. Zhang, S. and Chung, T. S. Minimizing the instant and accumulative effects of salt permeability to sustain ultrahigh osmotic power density. Environ. Sci. Technol. 47 (2013) 10085−10092. 30. Susanto, H. and Ulbricht, M. Photografted thin polymer hydrogel layers on PES ultrafiltration membranes: Characterization, stability, and influence on separation performance. Langmuir 23 (2007) 7818−7830.

20

31. Li, X.; Cai, T.; Chung T. S. Antifouling behavior of hyperbranched polyglycerolgrafted poly(ether sulfone) hollow fiber membranes for osmotic power generation. Environ. Sci. Technol. 48 (2014) 9898−9907. 32. Li X., Cai, T.; Chen, C. Y.; Chung T. S. Negatively charged hyperbranched polyglycerol grafted membranes for osmotic power generation from municipal wastewater. Water Res. 89 (2016) 50–58. 33. Cai, T.; Li, X.; Wan, C. F.; Chung T. S. Zwitterionic polymers grafted poly(ether sulfone) hollow fiber membranes and their antifouling behaviors for osmotic power generation. J. Membr. Sci. 497 (2016) 142–152. 34. Zhang, S.; Zhang, Y.; Chung T. S. Facile preparation of antifouling hollow fiber membranes for sustainable osmotic power generation. ACS Sustain. Chem. Eng. 4 (2016) 1154−1161. 35. Zhang, L.; She, Q.; Wang, R.; Wongchitphimon, S.; Chen, Y.; Fane, A. G. Unique roles of aminosilane in developing anti-fouling thin film composite (TFC) membranes for pressure retarded osmosis (PRO). Desalination 389 (2016) 119−128. 36. Saeki, D.; Tanimoto, T.; Matsuyama, H. Anti-biofouling of polyamide reverse osmosis membranes using phosphorylcholine polymer grafted by surface-initiated atom transfer radical polymerization. Desalination 350 (2014) 21−27. 37. Kang, G. D. and Cao, Y. M. Development of antifouling reverse osmosis membranes for water treatment: A review. Water Res. 46 (2012) 584−600. 38. Hu, M., Zheng, S. X; Mi, B. X. Organic Fouling of graphene oxide membranes and its implications for membrane fouling control in engineered osmosis. Environ. Sci. Technol. 50 (2016) 685−693. 39. Akthakul, A.; Salinaro, R. F.; Mayes, A. M. Antifouling polymer membranes with subnanometer size selectivity. Macromolecules 37 (2004) 7663−7668. 21

40. Sim, J. W.; Nam, S. H.; Koo, J. W.; Choi, Y. J.; Hwang, T. M. Organic fouling and osmotic backwashing in PRO. Desalin. Water Treat. 57 (2016) 10086−10092. 41. Kim, J.; Park, M. J.; Park, M.; Shon, H. K.; Kim, S. H.; Kim, J. H. Influence of colloidal fouling on pressure retarded osmosis. Desalination 389 (2016) 207−214. 42. Kim, D. I.; Kim, J.; Hong, S. Changing membrane orientation in pressure retarded osmosis for sustainable power generation with low fouling. Desalination 389 (2016) 197−206. 43. Nguyen, T. P. N.; Jun, B. M.; Park, H. G.; Han, S. W.; Kim, Y. K.; Lee, H. K.; Kwon, Y. N. Concentration polarization effect and preferable membrane configuration at pressure-retarded osmosis operation. Desalination 389 (2016) 58−67. 44. Li, X. and Chung, T. S. Effects of free volume in thin-film composite membranes on osmotic power generation. AIChE J. 59 (2013) 4749–4761. 45. Han, G.; Zhang, S.; Li, X.; Chung, T. S. High performance thin film composite pressure retarded osmosis (PRO) membranes for renewable salinity-gradient energy generation. J. Membr. Sci. 440 (2013) 108–121. 46. Cheng, Z. L.; Li, X.; Liu, Y. D.; Chung, T. S. Robust outer-selective thin-film composite polyethersulfone hollow fiber membranes with low reverse salt flux for renewable salinity-gradient energy generation. J. Membr. Sci. 506 (2016) 119–129. 47. Thomas, O. and Burgess, C. UV-Visible Spectrophotometry of Water and Wastewater, Techniques and Instrumentation in Analytical Chemistry; Elsevier: Amsterdam, 2007.

22

Figure 1. Fouling behaviors of CHPG-TFC membranes with different cleaning agents. The draw solution initially contained 0.81 mol/L NaCl, and a wastewater effluent was used as the feed solution. The hydraulic pressure difference between feed and draw solutions was 15.0 ± 1.0 bar. The first data point in each run was tested in DI water other than in wastewater. Runs 1, 2, 3 are testing cycles after cleaning. Figure 2. Fouling behaviors of the non-modified PES-TFC membrane using EDTA at pH 10.9 as the cleaning agent. The draw solution initially contained 0.81 mol/L NaCl, and a wastewater effluent was used as the feed solution. The hydraulic pressure difference between feed and draw solutions was 15.0 ± 1.0 bar. The first data point in each run was tested in DI water other than in wastewater. Runs 1, 2, 3 are testing cycles after cleaning. Figure 3. UV-Vis spectra of cleaning agents after membrane washing. 23

Figure 4. SEM detection on outer surfaces of PES-TFC and CHPGTFC hollow fibers after the 2

nd

cleaning. Dashed circle

indicates the film-like foulants. Figure 5. AFM images of outer surfaces of PES-TFC and CHPG-TFC nd

a

hollow fibers after the 2 cleaning. Ra is the mean roughness. Figure 6. Water contact angles of the porous layer of membranes in

different

treatments.

Demonstrated

by

flat-sheet

membranes. Figure 7. XPS spectra of CHPG-TFC membranes: (a) wide scan spectra, (b) Si 2p core-level spectra, and S 2p core-level spectra after being washed by (c) low pH, (d) high pH, (e) EDTA, and (f) DI water solutions. Figure 8. Zeta potential curves of the porous layer of CHPG-TFC membranes.

24

Table 1. Power densities of the non-modified PES-TFC and CHPGTFC membranes using EDTA as the cleaning agent. 2

Membrane

Power density (W/m ) Initial

After Run 1

After Run 2

After Run 3

Non-modified PESTFC

8.3

4.3-4.8

3.6-4.3

4.0-4.6

CHPG-TFC

8.3

6.0-6.6

6.1-6.7

6.0-6.5

Note: 0.81 mol/L NaCl and wastewater were used as the feed pair in PRO tests. The hydraulic pressure difference between feed and draw solutions was 15.0 ± 1.0 bar. Power density after a testing run is the value before cleaning.

25

Table 2. Total organic carbon (TOC) of different cleaning agents. DI, for PESTFC

DI, for CHPGTFC

Low pH, for CHPG -TFC

High pH, for CHPGTFC

Wastewate r

EDTA, for CHPG -TFC

TOC in the cleaning agent before washing (mg/L)

<0.0 1

<0.01

<0.01

<0.01

28~32

>9000

TOC in the cleaning agent after washing (mg/L)

0.40 ±0.0 1

0.25±0.0 2

0.98 ±0.11

88.11±7.4 9

-

>9000

TOC obtained from membran e surface 2 (mg/cm )

0.02 0

0.013

0.050

4.488

-

0.1~0. 3

Note: DI, low pH, high pH and EDTA refer to the cleaning agents being used.

26

Highlights ·

Cleaning procedures for pressure retarded osmosis (PRO) membranes were studied.

·

Charged hyperbranched polyglycerol grafted PRO membranes were examined.

·

The fouling resistant membrane exhibits sustainably high osmotic power density after proper cleaning.

27

Run 3

20

60

80

100

120

140

30

0

20

40

80

100

120

140

30 0

40

60

80

Run 2 100

120

140

Run 3

20

40

60

100

(L/m2)

80

Run 2

120

140

Run 3

(d) Cleaned by HCl at pH 2.8

20

Run 1

0

Run 1

(b) Cleaned by EDTA

Normalized accumulative permeate volume

60

40

50

50

Run 3

60

60

Run 2

70

70

Run 1

80

80

40

90

90

30

100

(c) Cleaned by NaOH at pH 11.8

0

40

100

30

40

40

60

60

Run 2

70

70

Run 1

80

80

50

90

90

50

100

100

(a) Cleaned by DI

Figure 1. Fouling behaviors of CHPG-TFC membranes with different cleaning agents. The draw solution initially contained 0.81 mol/L NaCl, and a wastewater effluent was used as the feed solution. The hydraulic pressure difference between feed and draw solutions was 15.0 f 1.0 bar. The first data point in each run was tested in DI water other than in wastewater. Runs 1, 2, 3 are testing cycles after cleaning.

Normalized water flux (%)

30

40

50

60

70

80

90

100

0

40

60

80

Run 2

100

120

140

Run 3

Normalized accumulative permeate volume (L/m2)

20

Run 1

Figure 2. Fouling behaviors of the non-modified PES-TFC membrane using EDTA at pH 10.9 as the cleaning agent. The draw solution initially contained 0.81 mol/L NaCl, and a wastewater effluent was used as the feed solution. The hydraulic pressure difference between feed and draw solutions was 15.0 f 1.0 bar. The first data point in each run was tested in DI water other than in wastewater. Runs 1, 2, 3 are testing cycles after cleaning.

Normalized water flux (%)

200

250

300

400

450

Wavelength (nm)

350

',IRU3(67)& ',IRU&+3*7)&

/RZS+IRU&+3*7)&

+LJKS+IRU&+3*7)&

('7$IRU&+3*7)&

500

1050 1100

Figure 3. UV-Vis spectra of cleaning agents after membrane washing.

Absorbance

100 nm

1µm

(b) CHPG-TFC cleaned by DI

1µm

100 nm

(c) CHPG-TFC cleaned by EDTA

100 nm

1µm

(d) CHPG-TFC cleaned by high pH

1µm

100 nm

(e) CHPG-TFC cleaned by low pH

Figure 4. SEM detection on outer surfaces of PES-TFC and CHPG-TFC hollow fibers after the 2nd cleaning. Dashed circle indicates the film-like foulants.

1µm

100 nm

(a) PES-TFC cleaned by DI

Ra(nm)=36.5f16.2

(b) CHPG-TFC cleaned by DI

Ra(nm)=34.2f14.1

(c) CHPG-TFC cleaned by EDTA

Ra(nm)=68.0f32.6

(d) CHPG-TFC cleaned by high pH

Ra(nm)=34.9f13.2

(e) CHPG-TFC cleaned by low pH

Figure 5. AFM images of outer surfaces of PES-TFC and CHPG-TFC hollow fibers after the 2nd cleaning. a Ra is the mean roughness.

Raa(nm)=25.9f10.4

(a) PES-TFC cleaned by DI

Virgin PESTFC

Fouled PESTFC

Virgin CHPGTFC

Fouled CHPGTFC

CHPGTFC, DI

CHPG- CHPG- CHPGTFC, TFC, TFC, EDTA Low Lo pH high pH H

Figure 6. Water contact angles of the porous layer of membranes in different treatments. Demonstrated by flat-sheet membranes.

Water contact angle (º)

0

N 1s

O 1s

95

800

By low pH

105

By DI water

By EDTA

By high pH

100

By DI By high pH By EDTA

(b) Si 2p By low pH

Binding Energy (eV)

400

Ca 2p

C 1s

S 2p S 2s

Si 2p Si 2s

(a) Wide scan

168

172

168

168

164

C-S

168

(f) By DI (S 2p)

164

C-S

172

172

172

(e) By EDTA (S 2p)

164

(d) By high pH (S 2p)

164

(c) By low pH (S 2p)

Figure 7. XPS spectra of CHPG-TFC membranes: (a) wide scan spectra, (b) Si 2p core-level spectra, and S 2p core-level spectra after being washed by (c) low pH, (d) high pH, (e) EDTA, and (f) DI water solutions.

Intensity (Arb. units)

-350

-300

-250

-200

-150

-100

-50

0

50

100

150

2

pH 2.8 4

6

pH

8

10

pH 10.9

Virgin CHPG-TFC

Fouled CHPG-TFC

12

pH 11.8

CHPG-TFC washed by EDTA

Figure 8. Zeta potential curves of the porous layer of CHPG-TFC membranes.

Zeta potential (mV)

GRAPHICAL ABSTRACT