Fabrication and characterization of antifouling carbon nanotube/polyethersulfone ultrafiltration membranes

Abstract

In this study, carbon nanotubes coated with poly(sulfobetaine methacrylate) (SBMA@CNT) particles were synthesized via a precipitation polymerization method. The SBMA@CNT particles were used as a novel kind of modifier to fabricate polyethersulfone (PES) ultrafiltration membranes by a non-solvent induced phase separation method (NIPS). During the membrane formation process, the surface segregation phenomenon of the SBMA@CNT particles was found, which was demonstrated by an energy dispersive spectrometer (EDS) mapping and X-ray photoelectron spectroscopy (XPS). This phenomenon was ascribed to the self-organization of hydrophilic SBMA@CNT particles spontaneously on the interface of membrane/water in the formation of membranes. In the ultrafiltration of a bovine serum albumin (BSA) feed solution, the best-performing membrane was found to effectively reduce protein adhesion. Its irreversible and reversible flux declines were remarkably decreased and the flux recovery was as high as 98.9%.

---

1. Introduction

Membrane technology is widely applied in different fields such as food processing, environmental protection, wastewater treatment and protein separation. However, membrane fouling leads to a remarkable decline of separation efficiency.¹ The accumulation and adsorption of proteins, natural organic matter (NOM) and microorganisms on the membrane surfaces and pores are the main reasons for membrane fouling. Therefore, much attention has been paid in academia and industry to fabricate antifouling membranes with high flux, good separation properties and low energy cost.^2,3

Antifouling ability of ultrafiltration membranes can be remarkably improved through surface modification. Surface coating, surface grafting and surface segregation have been investigated for improving the hydrophilicity of membranes, which could reduce the interaction force between foulants and membrane surfaces.^4–6 With the purpose of construction hydrophilic membrane surfaces, polyzwitterions have drawn attention in recent years. Polyzwitterions possess strong hydration capability and wonderful anti-adhesion properties for protein and bacteria.⁷ Generally speaking, polyzwitterions have both positive and negative charged functional groups on the same repeating unit.^8–10 The electrostatic interaction of zwitterionic functional groups induces a strong hydration layer. This layer acts as a buffer for foulant deposition. Liu synthesized a new aromatic zwitterionic polyimide copolymer to fabricate an antifouling ultrafiltration membrane.⁹ Xiang fabricated antifouling membranes via in situ cross-linked polymerization of poly(sulfobetaine methacrylate) (SBMA) with polyethersulfone (PES). The recovery ratios of the membranes in protein filtration were nearly 100%.¹⁰ In our research group, the antifouling properties of PVC membranes were improved by functionalizing the PVC with zwitterionic mononers.¹¹ The zwitterionic PVC membranes exhibited better permeability and stronger fouling resistance ability than the unmodified PVC membranes.

Carbon nanotubes (CNTs) have been applied in some of separation fields such as nanofiltration, ultrafiltration, and pervaporation.^12–15 Incorporating CNTs into membranes improved mechanical property and permeation.¹⁶ Those membranes performed higher flux, less membrane fouling, higher thermal stability and lower energy requirement than conventional polymeric membranes.¹⁷ Zwitterion–carbon nanotube has been synthesized to fabricate ion-responsive membranes.¹⁸ In our group, Liu synthesized SBMA@CNT particles to fabricate gas separation membranes.¹⁹ The CO₂ separation property enhanced due to the SBMA@CNT particles.

In this work, a composite membrane was designed by blending SBMA@CNT particles into the PES membrane in order to enhance antifouling property. SBMA is a typical polyzwitterion. Also, eight water molecules could be strongly bonded on each sulfobetaine unit while more water molecules are loosely adsorbed simultaneously.²⁰ CNT was used as the carrier for SBMA. The SBMA@CNT particles were prepared by precipitation polymerization method.^19,21 The modified ultrafiltration membranes exhibited significantly enhanced antifouling property. The chemical structure and morphology of SBMA@CNT particles were confirmed by Fourier transform infrared spectroscopy (FTIR) and transmission electron microscopy (TEM). The morphology and chemical structure of the modified ultrafiltration membranes were studied with field emission electron microscope (FESEM) and X-ray photoelectron spectroscopy (XPS). The mechanical properties of the membranes were analyzed by a material-testing machine. The interaction force between the protein and membrane surfaces was measured by atomic force microscope (AFM). Meanwhile, the antifouling abilities of the modified PES membranes were evaluated through ultrafiltration experiments.

2. Experimental section

2.1 Materials

Polyethersulfone (PES, 6020P, M_w 59 000, flake form) was obtained from BASF Co. (Germany). Before using, it was dried at 120 °C for 12 h. Polyethylene glycol (PEG) with an average molecular weight of 2000, methacrylic acid (MAA), 2,2′-azoisobutyronitrile (AIBN), N,N-dimethylformamide (DMF) and ethanol were supplied by Guangfu Fine Chemical Research Institute (Tianjin, China). 3-(Methacryloxy)propyltrimethoxysilane (MPS) and ethylene glycol dimethacrylate (EGDMA) were purchased from Alfa Aesar China Co., Ltd. Hydroxyl modified multiwalled CNT were obtained from Nanjing XFNANO Materials Tech Co., Ltd. Zwitterionic monomer, N-(3-sulfopropyl)-N-(methacryloxyethyl)-N,N-dimethylammonium betaine (SBMA) was provided by Sigma-Aldrich (St. Louis, MO). Acetonitrile was provided by Tianjin Kewei Co., Ltd. Bovine serum albumin (BSA), commercially analytical grade, was obtained from the local reagent corporation. In the experiments, the water was the deionized water.

2.2 Synthesis of the SBMA@CNT particles

As shown in Scheme 1, CNTs and MPS were mixed in ethanol at 50 °C for 48 h to insure grafting CNTs with C=C bonds as previous studies.¹⁹ Then, the polymerization of SBMA was carried out to modify CNTs via the precipitation polymerization method. The synthesis was carried out as follow: 0.08 g of CNTs were dispersed in a mixed solution of acetonitrile/water (3 : 1 vol%). After ultrasonication for 1 h, 0.80 g of SBMA, 0.04 g of AIBN, 0.20 mL of MAA and 0.80 mL of EGDMA were added in the solution, respectively, and heated to boiling for 80 min. After polymerization, the products were collected, washed with ethanol, and dried at 40 °C in a vacuum oven for 24 h.

Scheme 1 Schematic illustration for the synthesis process of the SBMA@CNT particles.

2.3 Preparation of the membranes

PES and SBMA@CNT particles blend ultrafiltration membranes were prepared through the non-solvent induced phase inversion (NIPS) method, using DMF as solvent and water as the non-solvent coagulation bath.²² The casting solution contained PES, SBMA@CNT, pore-forming agent PEG2000 and DMF. SBMA@CNT particles were dispersed in DMF by an ultrasonic cell disruption machine. After dispersing, the solutions were continued stirring for about 6 h and heating at 60 °C. And then, the solutions left for 8 h to make sure to degas completely. After cooled to room temperature, the casting solutions were cast on a glass plate using a steel knife and put into a deionized water coagulation bath for several minutes. The thin polymeric membranes were peeled off from the glass. All membranes were stored in deionized water in order to eliminate residual solvent. The modified membrane was named as PES/SBMA@CNT(x), where x was the particle content (wt%) in a polymer matrix.

2.4 Membrane characterization

A transmission electron microscopy (TEM, Tecnai G2 30 STWIN) was applied to observe the morphology of the SBMA@CNT particles. Fourier transform infrared spectrometer (FTIR Nicolet FTIR6700 equipped with horizontal attenuated total reflectance accessory) was used to analyze the chemical structure of the particles.

The cross-section and membrane surfaces morphologies of membranes were observed by field emission electron microscope (FESEM, Nova NanoSEM 430, FEI Co., USA). The membrane samples freeze-dried with vacuum freeze dryer (FD-1C-50, Boyikang Co., China) were broken in liquid nitrogen. Then they were sputtered with gold for producing electric conductivity before FESEM measurement. The element distribution of the modified membranes was studied by energy dispersive spectrometer (EDS) mapping during FESEM observation. The surface chemical composition of membranes was analyzed by X-ray photoelectron spectroscopy (XPS, Perkin Elmer Phi 1600 ESCA system) using Mg Kα (1254 eV) as the radiation source. Survey scans were in the range of 0–1100 eV for taken-off angle of 90°. The mechanical properties of membranes were analyzed by a material-testing machine (AXM-350-10KN, Testometric Co., UK). Before testing, membranes were dried and cut into 50 mm × 10 mm (length × width) rectangle strips. Atomic force microscopy (AFM, Multimode 3, Bruker Co.) was applied to measure the interaction force between membrane surfaces and a AFM tip.²³ The AFM tip was modified with model foulant BSA.

2.5 Ultrafiltration experiments and antifouling properties evaluation

Flux decline resistant and flux recovery properties were evaluated by a dead-end stirred cell system (model 8200, Millipore Co.) connected with a nitrogen gas cylinder. The system had an effective area of 28.7 cm². This cell could take 200 mL liquid. The whole filtration process was conducted at a stirring speed of 200 rpm. Meanwhile, the nitrogen gas was used as pressure. Before testing the performance of membranes, membranes were compacted at a high pressure of 0.15 MPa for 30 min aiming to obtain a steady flux. After compacting process, the pressure was adjusted to the operating pressure of 0.10 MPa. The pure water flux J_w1 (L m⁻² h⁻¹) was measured at this condition and it was calculated by the following equation:where V (L) is the volume of permeated water, A (m²) is the effective membrane area and Δt (h) is the permeation time. After the pure water experiment, the cell was refilled with protein solution (BSA, 1.0 mg mL⁻¹ in phosphate buffered solution (PBS), pH = 7.0). The flux of feed solution was recorded as J_p (L m⁻² h⁻¹) according to the water quality penetrating the membranes at the 0.10 MPa pressure. Membrane rejection efficiency was recorded as R. R was calculated according to the following equation:where C_p and C_f were the concentrations of BSA in penetrating fluid and feed solution, respectively. A UV-spectrophotometer (UV-2800, Hitachi Co., Japan) was applied to measure the BSA concentrations at wavelength of 278 nm. After the protein ultrafiltration process, the membranes were washed with deionized water for about 20 min. After that the water flux of the cleaned membranes, J_w2 (L m⁻² h⁻¹), was measured again. The total flux decline ratio (R_t), reversible flux decline ratio (R_r), irreversible flux decline ratio (R_ir) and flux recovery ratio (FRR) were calculated to analyze the antifouling capacities of the membranes. These were defined as reported in our previous work:²⁴

Furthermore, the sum of R_ir and R_r was R_t. Hydraulic cleaning of the membrane could reduce reversible BSA adsorption on the membrane surface. However, hydraulic cleaning could not reduce the irreversible fouling. Also, the excellent antifouling membranes have the high value of FRR and the low value of R_t.

3. Results and discussion

3.1 Characterization of SBMA@CNT particles

The morphology and chemical structure of SBMA@CNT particles were investigated by TEM and FTIR-ATR. The TEM images of CNTs and SBMA@CNT particles were shown in Fig. 1. Compared with the CNTs TEM image, it could be clearly seen that CNTs were fully coated with the SBMA layer. The interface between CNT and SBMA could be observed clearly, indicating that SBMA was coated on the outermost shell of CNTs. In addition, the modified CNT had less aggregation. FTIR-ATR spectrum of SBMA@CNT particles and CNTs were shown in Fig. 2. In the CNTs spectrum, the peaks at 1728 cm⁻¹ and 1163 cm⁻¹ were assigned to the C=O and the C–OH stretching vibrations of the carboxylic groups, respectively. After modification with SBMA, the intensity of the peaks enhanced. More importantly, the presence of the SBMA was evidenced by the –SO₃ stretch at 1036 cm⁻¹.^25,26 The FTIR-ATR spectra analysis evidenced that SBMA@CNT particles were successfully synthesized via a precipitation polymerization method.

Fig. 1 TEM image of CNTs and SBMA@CNT particles.

Fig. 2 FTIR spectrum of SBMA@CNT and CNTs.

3.2 Characterization of the membranes

The ultrafiltration membranes were fabricated via NIPS using PES as bulk material, PEG2000 as pore-forming agent, SBMA@CNT particles as modifiers, and water as coagulation bath agent, the composition of the casting solution was given in Table 1. The casting solutions still remained excellent membrane-forming ability after adding SBMA@CNT particles. To make out the influence of SBMA@CNT particles for membrane structure and morphology, the cross-section and surfaces of the membranes morphologies were observed by FESEM. As shown in Fig. 3, the unmodified PES membrane revealed a typical asymmetric structure, which possessed of a thin dense top-layer and porous finger-like structures. The structures of the PES/SBMA@CNT membranes consisted of the similar finger-like structures and dense top-layers. Also, it was clear that all the membrane surfaces exhibited smooth and flat surface morphologies (Fig. S1†). There was no appreciable structural and morphological change after blending the SBMA@CNT particles.

Table 1 The casting solutions for PES and PES/SBMA@CNT ultrafiltration membranes

Membrane	Composition of casting solution (g)
	PES (g)	PEG (g)	SBMA@CNT (g)	DMF (g)
Unmodified PES	3.2	3.2	0	13.600
PES/SBMA@CNT(0.50)	3.2	3.2	0.016	13.584
PES/SBMA@CNT(0.75)	3.2	3.2	0.024	13.576
PES/SBMA@CNT(1.00)	3.2	3.2	0.032	13.568
PES/SBMA@CNT(1.25)	3.2	3.2	0.040	13.560
PES/SBMA@CNT(1.50)	3.2	3.2	0.048	13.552

---

Fig. 3 The cross-section FESEM morphologies of PES, PES/SBMA@CNT(0.50), PES/SBMA@CNT(1.00), and PES/SBMA@CNT(1.50) membranes.

The element distribution of PES/SBMA@CNT(1.25) membrane across the thickness was characterized by EDS mapping analysis. Element mapping of the region in Fig. 4(a) was shown in Fig. 4(b–e), for carbon, nitrogen, oxygen and sulfur elements, respectively. Nitrogen was the characteristic element of SBMA. The appearance of nitrogen element on the membrane surface was attributed to the segregation of SBMA@CNT particles, which was used as a kind of surface modifier to improve the antifouling ability. The element of nitrogen appeared no apparent agglomeration in the Fig. 4(c). In addition, the proportion of nitrogen on the membrane surface was larger than other parts of the membrane in the Fig. 4(c), which indicated that there was a large coverage of SBMA@CNT particles on the surface of PES/SBMA@CNT(1.25) membrane. Element mapping of nitrogen confirmed that the SBMA@CNT particles successfully segregated into the membrane surface in a degree.

Fig. 4 Cross-section FESEM images of (a) PES/SBMA@CNT(1.25) and EDS mapping analysis of PES/SBMA@CNT(1.25) membrane (b–e).

The surface elements of PES/SBMA@CNT were also investigated by XPS analysis. Four characteristic XPS signals for C 1s (∼285 eV), O 1s (∼532 eV), N 1s (∼401 eV) and S 2p (∼167.5 eV) were observed in the wide-scan XPS spectra of the PES/SBMA@CNT(0.50), PES/SBMA@CNT(1.00) and PES/SBMA@CNT(1.50) membranes in Fig. 5. The elemental atomic percentages on membrane surfaces were shown in Table 2. There was no nitrogen element in the PES membrane. However, there was nitrogen element derived from SBMA in the PES/SBMA@CNT membranes due to adding the SBMA@CNT particles. The XPS results clearly confirmed the existence of SBMA@CNT particles on the membrane surfaces. In addition, the elemental atomic percentages of nitrogen were much larger than the actual contents of nitrogen in the casting solutions. This point confirmed that the surface segregation of SBMA@CNT particles assuredly happened. The nitrogen elemental atomic percentages increased with the increase of the SBMA@CNT particles.

Fig. 5 XPS wide-scan of PES/SBMA@CNT membranes.

Table 2 The XPS elemental atomic percentages of PES, PES/SBMA@CNT(0.50), PES/SBMA@CNT(1.00) and PES/SBMA@CNT(1.50) membrane surfaces

Membrane	Membrane surface composition [at%]
	C	O	S	N
Unmodified PES	74.6	19.1	6.3	—
PES/SBMA@CNT(0.50)	75.1	19.5	4.2	1.2
PES/SBMA@CNT(1.00)	73.9	20.0	4.1	2.1
PES/SBMA@CNT(1.50)	70.5	23.9	2.2	3.4

---

The surface segregation during the phase inversion was reported in the preparation of membranes.^27,28 Zhao synthesized F127-b-PDMS amphiphilic copolymers. The hydrophilic segments segregated freely on water/membrane interface because of its hydrophilicity.²⁹ Similarly, in the phase inversion process of the PES/SBMA@CNT membranes, as shown in Scheme 2, the SBMA@CNT particles could undergo self-organization at the interface of membrane/coagulation bath due to the superior intrinsic hydrophilicity of SBMA@CNT particles. This self-organization behavior provided a surface with large coverage of hydrophilic SBMA@CNT particles.

Scheme 2 Schematic illustration of the migration behavior of SBMA@CNT particles during the coagulation step in the phase inversion process.

The interaction force between the membrane surfaces and foulants was measured by AFM with the BSA-immobilized tip.²³ Fig. 6 showed the adhesion force between the membrane surface and the tip. The interaction force between the unmodified PES membrane surface and the tip was about 16 nN because of the hydrophobicity of PES. However, the interaction force between PES/SBMA@CNT(1.00) membrane surface and the tip was 10 nN. The SBMA@CNT particles segregated on the membrane surface during membrane formation. The zwitterionic SBMA groups leaded to the strong hydration layer due to electrostatic interactions on the membrane surfaces, which reduced the interaction of BSA with the membrane surfaces.

Fig. 6 Force–extension curves recorded with a BSA-immobilized tip against (a) PES, (b) PES/SBMA@CNT(1.00) membranes.

The mechanical properties of PES/SBMA@CNT membranes were studied by a material-testing machine. The mechanical properties data were listed in Table 3. Generally speaking, there was no apparent change of the strength, plasticity and elasticity of the modified membranes after the incorporation of the SBMA@CNT particles. The interconnected structure between SBMA@CNT particles and PES matrix might slightly enhance the mechanical properties.

Table 3 Mechanical properties of PES, PES/SBMA@CNT(0.50), PES/SBMA@CNT(1.00), PES/SBMA@CNT(1.50) membranes

Membrane	Young's modulus (MPa)	Maximum elongation (%)	Maximum strength (MPa)
Unmodified PES	55.76	15.64	3.99
PES/SBMA@CNT(0.50)	64.65	15.37	3.95
PES/SBMA@CNT(1.00)	60.22	28.58	3.92
PES/SBMA@CNT(1.50)	64.75	16.83	3.97

---

3.3 Antifouling property of membranes

Hydrodynamic conditions, membrane properties and feed characteristics affected the membrane filtration behavior, especially membrane fouling.^30,31 Electrostatic interactions, hydrophobic interaction and other factors often caused pore blocking, adsorption and cake formation in the BSA filtration.³² The permeation fluxes of the SBMA@CNT membranes during the filtration of BSA solution were described in the Fig. 7. After the BSA filtration, the membranes were cleaned by water in order to clean the reversible BSA adsorption on membrane surfaces. It was noted that the fluxes of membranes were recovered in a degree. R_r revealed the fouling that could recover after hydraulic cleaning. R_ir revealed the fouling that could not recover by hydraulic cleaning.

Fig. 7 (a) Time-dependent water permeation flux during BSA filtration of PES, PES/SBMA@CNT(0.50), PES/SBMA@CNT(0.75), PES/SBMA@CNT(1.00), PES/SBMA@CNT(1.25), and PES/SBMA@CNT(1.50) membranes, and (b) a summary of the corresponding FRR, R_t, R_r, and R_ir values.

In the fouling experiments, the BSA rejection rates were almost 100% for all the PES/SBMA@CNT membranes. Flux values of the PES/SBMA@CNT membranes during BSA ultrafiltration were used to analyze the antifouling properties. The unmodified PES membrane had the severer flux decline than the PES/SBMA@CNT membranes. The corresponding R_t value of the unmodified PES membrane was as high as about 62.8%. The R_r value was only 20.6%, and R_ir value was about 42.2%. And the FRR value was just 79.4%. The PES/SBMA@CNT membranes had better antifouling properties than the unmodified PES membrane. R_t values of the PES/SBMA@CNT membranes were 51.1, 45.0, 37.7, 43.5 and 57.2%, respectively, corresponding to the dosage of SBMA@CNT particles from 0.5 to 1.5 wt% in the casting solution. These R_t values were all smaller than the value of unmodified PES membrane. Also, the FRR values were 84.6, 91.9, 98.9, 92.1 and 84.5%, respectively. Similarly, these FRR values were larger than that of the unmodified one. In general, the hydrophilic membranes have hydration layers formed via hydrogen bonds, electrostatic interactions, which shows a lower flux decline and better antifouling properties.³³ The irreversible fouling mainly resulted from the adhesion and irreversible adsorption of BSA on the membrane surfaces. On the contrary, the reversible fouling resulted from the weak interaction and deposition of BSA on the membrane surfaces.³⁴ The improved antifouling properties of the modified PES membranes were attributed to the hydration layer formed by surface segregation of the hydrophilic SBMA@CNT particles. In addition, the hydration layer acted as a protective screen to minimize the chance for BSA directly attachment. With increasing the dosage of the SBMA@CNT particles, FRR values of the PES/SBMA@CNT membranes gradually increased and reached a peak of 98.9% for PES/SBMA@CNT(1.00) membrane. However, with the dosage of SBMA@CNT particles further increased, FRR values began to decrease. FRR value of PES/SBMA@CNT(1.50) membrane was decreased to 84.5%, the probable reason was pore blocking during protein ultrafiltration process.

4. Conclusions

In this study, the antifouling ultrafiltration membranes were designed and prepared by blending SBMA@CNT particles into PES membranes. EDS mapping and XPS analysis confirmed that the surface segregation of SBMA@CNT particles on the membrane surface really happened. AFM results exhibited that the incorporation of SBMA@CNT particles could reduce the interaction force between foulants and membrane surfaces. PES/SBMA@CNT(1.00) membrane performed the strongest pollution resistance among the modified membranes, whose irreversible and reversible flux declines were remarkably decreased and the flux recovery was as high as 98.9%. The enhanced antifouling properties could be ascribed to the existence of multifunctional polyzwitterion groups SBMA@CNT particles on the membrane surfaces.

Acknowledgements

This research is supported by Tianjin Natural Science Foundation (No. 13JCYBJC20500, 14JCZDJC37400), and National Science Fund for Distinguished Young Scholars (21125627).

References

1. L. Shao, Z. X. Wang, Y. L. Zhang, Z. X. Jiang and Y. Y. Liu, J. Membr. Sci., 2014, 461, 10–21.
2. S. Azari and L. Zou, Desalination, 2013, 324, 79–86.
3. H. Meng, Q. Cheng and C. Li, Appl. Surf. Sci., 2014, 303, 399–405.
4. S. Azari and L. Zou, J. Membr. Sci., 2012, 401–402, 68–75.
5. X. Zhao, Y. Su, H. Dai, Y. Li, R. Zhang and Z. Jiang, J. Mater. Chem. A, 2015, 3, 3325–3331.
6. Y. Zhu, F. Zhang, D. Wang, X. F. Pei, W. Zhang and J. Jin, J. Mater. Chem. A, 2013, 1, 5758.
7. Q. Shao and S. Jiang, Adv. Mater, 2015, 27, 15–26.
8. G. Liu, L. Zhang, S. Mao, S. Rohani, C. Ching and J. Lu, Sep. Purif. Technol., 2015, 152, 55–63.
9. Y. Liu, C. Ma, S. Wang, H. Guo, B. Zhang, L. Zhang, K. Gu and J. Gu, RSC Adv., 2015, 5, 21316–21325.
10. T. Xiang, C.-D. Luo, R. Wang, Z.-Y. Han, S.-D. Sun and C.-S. Zhao, J. Membr. Sci., 2015, 476, 234–242.
11. J. Zhu, Y. Su, X. Zhao, Y. Li, J. Zhao, X. Fan and Z. Jiang, Ind. Eng. Chem. Res., 2014, 53, 14046–14055.
12. S. Wang, S. Liang, P. Liang, X. Zhang, J. Sun, S. Wu and X. Huang, J. Membr. Sci., 2015, 491, 37–44.
13. H. J. Kim, K. Choi, Y. Baek, D.-G. Kim, J. Shim, J. Yoon and J.-C. Lee, ACS Appl. Mater. Interfaces, 2014, 6, 2819–2829.
14. B. Gao, Z. Jiang, C. Zhao, H. Gomaa and F. Pan, J. Membr. Sci., 2015, 492, 230–241.
15. J. Zhu, M. Tian, J. Hou, J. Wang, J. Lin, Y. Zhang, J. Liu and B. Van der Bruggen, J. Mater. Chem. A, 2016, 4, 1980–1990.
16. Y. Feng, K. Wang, C. H. J. Davies and H. Wang, Nanomaterials, 2015, 5, 1366–1378.
17. E.-S. Kim, G. Hwang, M. Gamal El-Din and Y. Liu, J. Membr. Sci., 2012, 394–395, 37–48.
18. T.-Y. Liu, H.-G. Yuan, Q. Li, Y.-H. Tang, Q. Zhang, W. Qian, B. V. D. Bruggen and X. Wang, ACS Nano, 2015, 9, 7488–7496.
19. Y. Liu, D. Peng, G. He, S. Wang, Y. Li, H. Wu and Z. Jiang, ACS Appl. Mater. Interfaces, 2014, 6, 13051–13060.
20. P. Bengani, Y. Kou and A. Asatekin, J. Membr. Sci., 2015, 493, 755–765.
21. G. He, Z. Li, Y. Li, Z. Li, H. Wu, X. Yang and Z. Jiang, ACS Appl. Mater. Interfaces, 2014, 6, 5362–5366.
22. E. Celik, H. Park, H. Choi and H. Choi, Water Res., 2011, 45, 274–282.
23. Y. Liu, Y. Su, X. Zhao, Y. Li, R. Zhang and Z. Jiang, J. Membr. Sci., 2015, 486, 195–206.
24. Y. Li, Y. Su, X. Zhao, R. Zhang, J. Zhao, X. Fan and Z. Jiang, J. Membr. Sci., 2014, 455, 15–23.
25. P. Kasák, Z. Kroneková, I. Krupa and I. Lacík, Polymer, 2011, 52, 3011–3020.
26. Y.-C. Chiang, Y. Chang, A. Higuchi, W.-Y. Chen and R.-C. Ruaan, J. Membr. Sci., 2009, 339, 151–159.
27. M. A. Shannon, P. W. Bohn, M. Elimelech, J. G. Georgiadis, B. J. Marinas and A. M. Mayes, Nature, 2008, 452, 301–310.
28. Y.-q. Wang, T. Wang, Y.-l. Su, F.-b. Peng, H. Wu and Z.-y. Jiang, Langmuir, 2005, 21, 11856–11862.
29. X. Zhao, Y. Su, Y. Li, R. Zhang, J. Zhao and Z. Jiang, J. Membr. Sci., 2014, 450, 111–123.
30. J.-y. Tian, M. Ernst, F. Cui and M. Jekel, J. Membr. Sci., 2013, 446, 1–9.
31. F. Xiao, P. Xiao, W. Zhang and D. Wang, J. Membr. Sci., 2013, 447, 144–152.
32. M.-J. Corbaton-Baguena, S. Alvarez-Blanco and M.-C. Vincent-Vela, Desalination, 2015, 360, 87–96.
33. C. M. Pang, P. Hong, H. Guo and W.-T. Liu, Environ. Sci. Technol., 2005, 39, 7541–7550.
34. X. Fan, Y. Su, X. Zhao, Y. Li, R. Zhang, T. Ma, Y. Liu and Z. Jiang, J. Membr. Sci., 2016, 499, 56–64.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra02991d

---