Fabrication of reduced graphene oxide membranes for highly efficient water desalination

Abstract

Membrane desalination technologies have become important energy-efficient means to ensure freshwater resources all over the world. The well-defined nanochannels and high water permeability of graphene-based membranes give them properties of purification and desalination. In this work, reduced graphene oxide membranes are fabricated using dopamine followed by vacuum filtration. The resultant membranes allow faster permeation of water compared with pristine graphene oxide membranes, but a higher retention rate of solutes. The increase of interaction between functional groups of reduced graphene oxide and the ions or water molecules was responsible for these excellent performances, which made graphene-based membranes promising materials used in desalination and water treatment.

---

Introduction

In recent decades, the shortage of freshwater resources caused by population growth and water pollution has become more and more significant around the world, which forces us to propose some new advanced water treatment technologies that can provide an alternative supply in a low-cost, highly efficient and environmentally friendly way.^1,2 Membrane-based desalination techniques, mainly reverse osmosis (RO), are currently considered as more environment-friendly and energy-efficient in comparison with adsorption or thermal desalination methods. Currently, membranes made from polymers, zeolite and silicon are applied in industry.³ However, these technologies suffer from low desalination capacity or high capital costs.⁴

Recently, considerable interest has been aroused by graphene-based nanomaterials for their potential exploitation in water desalination and purification, which depends on their unique properties including distinctive structure characteristics, great mechanical strength and negligible thickness.^5–13 Several simulation studies have indicated that nanoporous graphene (NPG), introducing nanoscale pores with various sizes, shapes and functionalities into the matrix, exhibited high water flow rates coupled with high salt rejection capacity.^14–16 Nevertheless, these results are mainly based on a single layer of graphene sheet which is difficult to assemble on a large scale in actual industry.

Alternatively, graphene oxide (GO), which is synthesized using the improved Hummers' method,¹⁷ is a better candidate for potentially being used in practical applications due to its characteristics of low cost and ease of production on a large scale.^18–23 Typically, a large number of GO sheets with polar oxygen-containing functional groups on the basal plane and the edges can be stacked layer-by-layer to form a laminate structure.^7,18,24 Within the laminate, a network of nanocapillaries is formed by the connected interlayer space and edges of GO sheets. Water molecules can appear ultrafast flow through the nanocapillaries while other liquids, vapors, and gases are completely restricted.^7,8,25,26 Based on the molecular sieving effect and chemical interactions due to the narrow magnitude of nanocapillaries and the existence of numerous functionalities, GO membranes can afford excellent selectivity toward various ions.^10,11

However, GO membranes performed poorly with regard to water desalination due to rapid ion transport.^9,11 Notably, Mi has proposed that for applications of GO membranes in water desalination, the GO interlayer spacing within the laminates has to be reduced to less than 0.7 nm to sieve the hydrated Na⁺ ions (hydrated radii: 0.36 nm) from water.²⁷ Nevertheless, Geim et al. demonstrated that with the reduction of the GO membranes, the permeation of water is significantly weakened due to the decrease of hydrophilic groups, even resulting in impermeable barrier films in some cases.^7,28 Fortunately, Liu et al. recently reported that after treating GO laminates with hydroiodic (HI) acid, the resulting reduced graphene oxide (RGO) membranes have high rejection toward salts, but allow rapid transport of water in the forward osmosis filtration process.²⁹ Sun et al. reported that by intercalating monolayer titania (TO) nanosheets into the GO laminates, followed by mild ultraviolet (UV) reduction, as-prepared RGO/TO hybrid membranes possess excellent ion rejection while the high water permeation of the pristine GO/TO hybrids can be preserved.¹² Moreover, Liu et al. reported recently a facile and environmentally friendly method to fabricate RGO membranes, or called as poly-dopamine (PDA) coated RGO membranes (PDA–RGO), within which there are still abundant hydrophilic functionalities of PDA such as catechol groups and amine groups, resulting that the as-prepared films exhibited super-hydrophilic properties.³⁰ This work provided a new idea for us to design GO-based membranes capable of rejecting ions while preserving high water permeations.

In our investigation, RGO membranes are fabricated using dopamine followed by vacuum filtration. The resultant membranes are found to allow faster permeation of water but a higher retention rate of solutes compared with GO membranes. These exceptional properties may provide a forward outlook for graphene-based materials in applications of industrial desalination and water treatment.

Experimental procedure

Preparation of GO & PDA–RGO membranes

GO was synthesized using the improved Hummers' method starting from natural flake graphite (325 mesh), which was exposed to a mixture of concentrated sulfuric acid, phosphoric acid and potassium permanganate for oxidation.¹⁷ Typically, a mixture of concentrated H₂SO₄/H₃PO₄ (180 : 20 mL, AR) was added into graphite flakes (1.5 g) under stirring in an ice bath. KMnO₄ (9.0 g, AR) was added slowly to keep the temperature of suspension lower than 20 °C. The reaction was then heated to 50 °C and stirred for 3 h. Successively, the system was transferred to 85 °C and vigorously stirred for about 0.5 h. Then, 250 mL ice-water was added slowly and followed by a slow addition of ∼5 mL H₂O₂ (30%, AR), turning the color of the solution from dark brown to bright yellow. After settled overnight and supernatant liquid removed, the remaining solid material was then washed and subsequently centrifuged (5000 rpm for 10 min) by turns until the supernatant was neutral (pH = 6.0–7.0). The last solid was vacuum-dried overnight at 45 °C and then grinded, obtaining graphite oxide flakes.

The prepared graphite oxide flakes were exfoliated in water by sonication (125 W for 10 min) to obtain monolayer or multilayer graphene oxide nanosheets (GO-Ns) suspension. GO membranes were formed by vacuum filtration of a certain volume of the obtained suspension on a commercial mixed cellulose (MC) filter (diameter: 50 mm, pore size: 0.22 μm, thickness: 100 μm). Equal GO dispersion was diluted to 0.2 mg mL⁻¹ and then a solution of dopamine hydrochloride with Tris buffer (pH ≈ 8.5) was added under magnetic stirring at room temperature to obtain RGO suspension, with which a PDA–RGO film was fabricated by vacuum filtration.

Characterizations

The morphologies of membranes were examined by Scanning Electron Microscope (SEM, SIGMA, Carl Zeiss). The interlayer spacing of lamellar films was determined by X-ray diffraction (XRD, X′ Pert Pro, Panalytical) in a range of 5° ≤ 2θ ≤ 50°. The characterization of X-ray photoelectron spectroscopy was carried out with an X-ray photoelectron spectrometer (XPS, ESCALAB 250Xi, Thermo Fisher Scientific) to determine the surface chemical composition. Fourier transform infrared spectra were separately recorded in a FTIR spectrophotometer (FTIR 5700, Nicolet) over the range of 4000–500 cm⁻¹ to identify the functional groups.

Permeation experiments

The ion diffusion experiments were carried out using a handmade apparatus, as shown in Fig. 1(a and b), which consists of two compartments separated by the studied GO (or PDA–RGO) membranes with the MC membrane underneath (GO/MC or PDA–RGO/MC) in the middle. The membranes were sandwiched between two copper foils (30 mm in diameter) with an aperture of 5 mm in diameter in the center. The copper foils were clamped between rubber rings. In a typical experiment, 80 mL of a 0.1 mol L⁻¹ saline solution and deionized water was poured into the source and drain vessels, respectively, at the same speed. Mild magnetic stirrings were applied to both compartments to avoid possible concentration gradients near the membranes. During the experimental process, the electrical conductivities of both source and drain solutions were measured by a conductivity meter (LEICI, DDSJ-308A). GO membranes were substituted with cellulose tape sandwiched between the copper foils with all the other steps remaining the same, resulting that any permeation could not be detected, which proves that the device does not have any possible cracks or leaks.

Fig. 1 (a, b) Photographs of ion diffusion experiment apparatus with two vessels. (c, d) Experimental setups for pressure-driven water flow.

Even if GO membranes allow unimpeded permeation of water molecules, the magnitude of the penetration rate is very low, as reported in the literatures.^10,12 In order to increase the flow rates, water flow tests were performed by sealing GO (or PDA–RGO) membranes with rubber rings in a dead-end device pressurized with N₂ flow, as illustrated in Fig. 1(c and d). In a typical experiment, 50 mL of a certain feed solution was injected into the chamber from Orifice I with Orifice II opened for balancing the internal pressure. Then Orifice I was blocked by a stopper and N₂ flow was introduced through Orifice II for applying pressure. The feed solution was allowed to percolate through the studied membrane under a pressure of 0.05 MPa. The corresponding permeate solution was collected with time for measuring the fluxes.

Results and discussion

GO flakes were prepared via improved Hummers' method combined starting from natural graphite. Fig. 2(a) shows a typical photograph of a 0.2 mg mL⁻¹ GO aqueous suspension. The RGO suspension was produced by mixing GO dispersion and a solution of dopamine-hydrochloride with magnetic stirring for 12 h at room temperature, turning into a black liquid, presented in Fig. 2(b). The GO (or PDA–RGO) membranes, as schematically shown in Fig. 2(a and b), were subsequently constructed by a method of vacuum suction.

Fig. 2 (a) Photographs of GO suspension and membrane. (b) Photographs of RGO suspension and membrane. (c) SEM characterization of the cross-section of GO membranes. (d) XRD spectra of GO and RGO. (e, f) XPS spectra of GO and RGO. (g, h) C 1s XPS spectra of GO and RGO. (i) FTIR spectra of GO and RGO.

The cross-section of GO membranes is shown in the Fig. 2(c), revealing a laminated structure with thickness about 1 μm. XRD characterizations of the lamellar structure were further investigated, as shown in Fig. 2(d), resulting that the diffraction peak was located at 2θ = 9.35°, which corresponds to an average interlayer spacing of 9.45 Å in dry state. While the RGO showed amorphous structure, which resulted in the decrease of actual nanocapillaries. The composition changes of GO before and after reduction were examined by XPS and was presented in Fig. 2(e and f), revealing that the oxygen content of GO increased with reduction. The peak associated with C–C became dominant in the C 1s spectrum of PDA–RGO. Notably, the significant decrease in intensities of C–O and C=O species and the appearance of the C–N peak component indicated that GO had been successfully reduced and doped, as shown in Fig. 2(g and h). The incorporation of PDA into the RGO membrane was further confirmed by FTIR spectroscopy, presented in Fig. 2(i). It's worth noting that peak at 1510 cm⁻¹ were assigned to the amide N–H shearing vibration and peak due to the C=O stretching modes at 1729 cm⁻¹ of GO were not detected in the spectrum of PDA–RGO, suggesting that GO was convincingly reduced and doped.

The ion permeation experiments were conducted using a homemade apparatus. Notably, the MC filters underneath were retained because the substrates could provide a mechanical strength enhancement for the GO (or PDA–RGO) membranes. To compare the permeation rates of different salts through the membranes, relative conductivities (defined as the conductivity of the filtrate divided by that of a 0.1 mol L⁻¹ initial feed solution) were considered. In the first control experiment, a bare MC film was installed in the device with all the other steps remaining the same to subtract its effect, appearing that various salts can quickly pass through it without significant differences in speed, as shown in Fig. 3(a), which indicated that the filters have no selectivity. In the next experiments, the effects of salts permeation through GO (or PDA–RGO) membranes were studied respectively under the same concentration gradient (0.1 mol L⁻¹). The relative conductivities of drains in the case of GO/MC are plotted as a function of time in Fig. 3(b), showing that the process of ion transfer was obviously weakened, and chloride salts of different cations had a certain selectivity compared with the bare MC film. The permeation rate of potassium and magnesium was significantly higher (∼2.5 times) than that of sodium and calcium. When the GO was reduced by dopamine and formed PDA–RGO membranes, the various salts were effectively blocked, as shown in Fig. 3(c). On the other hand, water is much easier to flow through the PDA–RGO membranes relative to initial GO films, as illustrated in Fig. 3(d). The rejection coefficient of salts in the case of PDA–RGO/MC is about 15 times more than that of bare MC, and 4–5 times than that of GO/MC. Meanwhile, the water permeation of PDA–RGO/MC can reach about 12 times more than GO/MC.

Fig. 3 (a) Salt permeations through bare microfilters (MC). (b) Salt permeations through GO/MC. (c) Salt permeations through PDA–RGO/MC. (d) Fluxes through membranes in pressurized condition.

The mechanism for the effective desalination was discussed. According to the previous reports, we know that molecular sieving effect and the delicate chemical interactions between different small ions and functional groups were responsible for the permeation of metal ions.^10,13 Larger solutes with hydrated radii >0.45 nm can be blocked by the physical size effect of the nanocapillaries within GO membranes, based only on which the size-fitted species can enter in and further permeate through.¹⁰ The negative charges of GO membranes originating from the ionization of the attached oxygen functionalities (e.g., carboxylates) give rise to electrostatic repulsion toward anions and electrostatic attraction toward cations in solution.¹¹ The diverse interactions with different strengths between different ions and the different regions lead to the diversity toward various species in solutions when permeating through, resulting the phenomenon in Fig. 3(b).

In our RGO membranes, the significant decrease of oxygen-containing groups, synchronously reducing their interactions with solute particles, and the sharp diminution of nanocapillaries with the formation of amorphous structure weakened the diffusion of ions. After the preparation procedure, the cross-linking polymeric dopamine and residues tris(hydroxymethyl) aminomethane molecules between GO flakes within the resultant film restrict the movement of metal ions in the nanochannels, but maybe can adsorb more water molecules into the membranes and facilitate water transmembrane permeations because of their hydrophily.²⁸ On the other hand, with reduction, the sp² matrix increased relatively, where the water molecules can slip through with little resistance.^7,31,32 Additionally, with the oxygen-containing groups decreasing, the hydrogen-bonding between the water and carboxyl decreased sharply, which contributed the faster water flow.

Conclusions

In summary, reduced graphene oxide membranes are fabricated using dopamine followed by vacuum filtration. The resultant membranes allow faster permeation of water compared with bare micro-filters and graphene oxide membranes, but a higher retention rate of solutes. Concretely, the salt rejection can be increased by 5 times compared with the pristine graphene oxide membranes, whereas the water permeation can reach 12 times. The results presented in this study indicate that our membranes are a promising candidate in future applications for water desalination and wastewater purification.

Conflict of interest

The authors declare no conflict of interest.

Acknowledgements

The authors are grateful for financial support from the National Natural Science Foundation of China (No. 50906064) and the Fundamental Research Funds for the Central Universities (No. 2042016kf0023).

References

1. M. A. Shannon, P. W. Bohn, M. Elimelech, J. G. Georgiadis, B. J. Mariñas and A. M. Mayes, Nature, 2008, 452, 301.
2. M. Elimelech and W. A. Phillip, Science, 2011, 333, 712.
3. P. Bernardo, E. Drioli and G. Golemme, Ind. Eng. Chem. Res., 2009, 48, 4638.
4. K. P. Lee, T. C. Arnot and D. Mattia, J. Membr. Sci., 2011, 370, 1.
5. A. K. Geim and K. S. Novoselov, Nat. Mater., 2007, 6, 183.
6. X. Zhang, Y. Liu, C. C. Sun, H. F. Ji, W. F. Zhao, S. D. Suna and C. S. Zhao, RSC Adv., 2015, 5, 100651.
7. R. R. Nair, H. A. Wu, P. N. Jayaram, I. V. Grigorieva and A. K. Geim, Science, 2012, 335, 442.
8. Y. Han, Z. Xu and C. Gao, Adv. Funct. Mater., 2013, 23, 3693.
9. M. Hu and B. Mi, Environ. Sci. Technol., 2013, 47, 3715.
10. R. K. Joshi, P. Carbone, F. C. Wang, V. G. Kravets, Y. Su, I. V. Grigorieva, H. A. Wu, A. K. Geim and R. R. Nair, Science, 2014, 343, 752.
11. P. Z. Sun, M. Zhu, K. L. Wang, M. L. Zhong, J. Q. Wei, D. H. Wu, Z. P. Xu and H. W. Zhu, ACS Nano, 2013, 7, 428.
12. P. Z. Sun, Q. Chen, X. D. Li, H. Liu, K. L. Wang, M. L. Zhong, J. Q. Wei, D. H. Wu, R. Z. Ma, T. Sasaki and H. W. Zhu, NPG Asia Mater., 2015, 7, e162.
13. S. Gahlot, P. P. Sharma, H. Gupta, V. Kulshrestha and P. K. Jha, RSC Adv., 2014, 4, 24662.
14. S. P. Koenig, L. Wang, J. Pellegrino and J. S. Bunch, Nat. Nanotechnol., 2012, 7, 728.
15. K. Sint, B. Wang and P. Král, J. Am. Chem. Soc., 2008, 130, 16448.
16. D. Cohen-Tanugi and J. C. Grossman, Nano Lett., 2012, 12, 3602.
17. D. C. Marcano, D. V. Kosynkin, J. M. Berlin, A. Sinitskii, Z. Z. Sun, A. Slesarev, L. B. Alemany, W. Lu and J. M. Tour, ACS Nano, 2010, 4, 4806.
18. D. A. Dikin, S. Stankovich, E. J. Zimney, R. D. Piner, G. H. B. Dommett, G. Evmenenko, S. T. Nguyen and R. S. Ruoff, Nature, 2007, 448, 457.
19. G. Eda and M. Chhowalla, Adv. Mater., 2010, 22, 2392.
20. G. Eda, G. Fanchini and M. Chhowalla, Nat. Nanotechnol., 2008, 3, 270.
21. X. Huang, X. Qi, F. Boey and H. Zhang, Chem. Soc. Rev., 2011, 41, 666.
22. X. Huang, Z. Yin, S. Wu, X. Qi, Q. He, Q. Zhang, Q. Yan, F. Boey and H. Zhang, Small, 2011, 7, 1876.
23. C. Tan, X. Huang and H. Zhang, Mater. Today, 2013, 16, 29.
24. D. R. Dreyer, S. Park, C. W. Bielawski and R. S. Ruoff, Chem. Soc. Rev., 2010, 39, 228.
25. D. W. Boukhvalov, M. I. Katsnelson and Y. W. Son, Nano Lett., 2013, 13, 3930.
26. G. Algara-Siller, O. Lehtinen, F. C. Wang, R. R. Nair, U. Kaiser, H. A. Wu, A. K. Geim and I. V. Grigorieva, Nature, 2015, 519, 443.
27. B. Mi, Science, 2014, 343, 740.
28. Y. Su, V. G. Kravets, S. L. Wong, J. Waters, A. K. Geim and R. R. Nair, Nat. Commun., 2014, 5, 4843.
29. H. Liu, H. Wang and X. Zhang, Adv. Mater., 2015, 27, 249.
30. N. Liu, M. Zhang, W. F. Zhang, Y. Z. Cao, Y. N. Chen, X. Lin, L. X. Xu, C. Li, L. Feng and Y. Wei, J. Mater. Chem. A, 2015, 3, 20113.
31. E. Mamontov, C. J. Burnham, S.-H. Chen, A. P. Moravsky, C.-K. Loong, N. R. de Souza and A. I. Kolesnikov, J. Chem. Phys., 2006, 124, 194703.
32. V. Henk, S. C. Melissa and J. Li, Small, 2007, 3, 1996.

---

Footnote

† These authors contributed equally to this work.

---