Controlling reduction degree of graphene oxide

0 downloads 0 Views 906KB Size Report
Apr 30, 2018 - graphene oxide membranes for improved water permeance, ... in-plane oxygen functional groups can hinder the water transport in graphene.
Accepted Manuscript Article Controlling reduction degree of graphene oxide membranes for improved water permeance Qing Zhang, Xitang Qian, Khalid Hussain Thebo, Hui-Ming Cheng, Wencai Ren PII: DOI: Reference:

S2095-9273(18)30236-6 https://doi.org/10.1016/j.scib.2018.05.015 SCIB 406

To appear in:

Science Bulletin

Received Date: Revised Date: Accepted Date:

23 April 2018 29 April 2018 30 April 2018

Please cite this article as: Q. Zhang, X. Qian, K.H. Thebo, H-M. Cheng, W. Ren, Controlling reduction degree of graphene oxide membranes for improved water permeance, Science Bulletin (2018), doi: https://doi.org/10.1016/ j.scib.2018.05.015

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 proof before it is published in its final 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.

Article Received 23 April 2018 Received in revised form 29 April 2018 Accepted 30 April 2018

Controlling reduction degree of graphene oxide membranes for improved water permeance Qing Zhang,1,2 Xitang Qian,1,2 Khalid Hussain Thebo,1,3 Hui-Ming Cheng,1,2,4 Wencai Ren1,2,* 1

Shenyang National Laboratory for Materials Science, Institute of Metal Research,

Chinese Academy of Sciences, Shenyang 110016, China 2

School of Materials Science and Engineering, University of Science and Technology

of China, Shenyang 110016, China 3

University of Chinese Academy of Sciences, Beijing 100049, China

4

Tsinghua-Berkeley Shenzhen Institute (TBSI), Tsinghua University, Shenzhen

518055, China. Email: [email protected]

Abstract Tailoring the pore structure and surface chemistry of graphene-based laminates is essentially important for their applications as separation membranes. Usually, pure graphene oxide (GO) and completely reduced GO (rGO) membranes suffer from low water permeance because of the lack of pristine graphitic sp2 domains and very small interlayer spacing, respectively. In this work, we studied the influence of reduction degree on the structure and separation performance of rGO membranes. It was found that weak reduction retains the good dispersion and hydrophilicity of GO nanosheets. More importantly, it increases the number of pristine graphitic sp2 domains in rGO

nanosheets while keeping the large interlayer spacing of the GO membranes in most regions at the same time. The resultant membranes show a high water permeance of 56.3 L m−2 h−1 bar−1, which is about 4 times and over 104 times larger than those of the GO and completely reduced rGO membranes, respectively, and high rejection over 95% for various dyes. Furthermore, they show better structure stability and more superior separation performance than GO membranes in acid and alkali environments.

Keywords: graphene oxide, reduction, separation membrane, water permeance

1. Introduction With the increasing demand for clean water, advanced membrane filtration technology has been extensively developed to remove heavy metal ions [1-3], organic dyes [4, 5], and micropollutants [6] from waste water. In the membrane filtration system, membranes serve as thin barriers between feed and permeate liquids for preferential

mass-transport

to

separate,

concentrate,

or

fractionate

under

pressure-driven conditions. As for nanofiltration membranes, permeability and selectivity are two important factors to evaluate the separation performance of the membranes, both of which strongly depend on the microstructures and chemical properties [7] of membranes such as the thickness, pore structure, hydrophilicity and functional groups. Recently, graphene oxide (GO) membranes have emerged as a new type of potential filtration membrane for efficient water purification [8-12], especially with the rapid development of mass production techniques [13-15]. The pure GO membranes have appropriate interlayer spacing in the range of 0.7 to 1.0 nm, which allows the sieving of smaller water molecules from mixed solution. In GO membranes, it was suggested that the pristine sp2 graphene domains provide fast low-friction channels for water flow and the oxygen functional groups clustered together act as spacers to provide appropriate spaces for water transport [16-18]. However, the

in-plane oxygen functional groups can hinder the water transport in graphene nanochannels by hydrogen bond interaction [19, 20]. Therefore, pure GO membranes usually show low water permeance due to the presence of a great number of in-plane oxygen functional groups. In addition, the negatively charged GO nanochannels are easily damaged in aqueous solutions due to hydration and electrostatic repulsion of the ionized carboxyl groups under hydration condition [21, 22]. Therefore, for practical filtration applications, GO membranes need to be further modified in order to maintain their structural stability and improve the water permeance without sacrificing the rejection. Recently, various chemical modification methods such as small molecule cross-linking [23-27] and macromolecule or nanomaterial intercalation [28-30] have been developed to improve the stability and water permeance. However, the rejection is usually decreased due to the enlarged nanochannels [26, 28, 31]. Moreover, the nanoscale intercalation agents such as multi-walled carbon nanotubes [29] and polymers [32, 33] made it difficult to obtain uniform nanochannels. It is well known that reduction can eliminate the oxygen containing groups and recover the π-conjugation in GO, which provide a possible way to minimize the hindrance of functional groups in nanochannels. However, direct reduction of GO membranes using thermal annealing [16] or chemical treatment [34] usually gives rise to graphitization, and the significantly improved π-π interaction leads to serious aggregation of nanosheets and capillaries collapse. The resulting small interlayer spacing of ~0.36 nm blocks all substances in the solutions including water molecules. In addition, the complete elimination of oxygen functional groups reduction makes the

membranes

hydrophobic,

which

limit

their

practical

applications

in

pressure-driven water purification. Here, we demonstrate the controlled reduction of GO nanosheets by hydrothermal reduction method. It is found that the different reduction degrees result in different microstructures, chemical characteristics, and separation performance of the assembled membranes. The membranes made by weakly reduced GO nanosheets show a high water permeance of 56.3 L m−2 h−1 bar−1, which is about 4 times and over

104 times larger than those of the GO and completely reduced rGO membranes, respectively, and high rejection over 95% for various dyes. Furthermore, they show better structure stability and more superior separation performance than GO membranes in acid and alkali environments.

2. Materials and methods 2.1 GO synthesis GO dispersion was prepared following the modified Hummers method [35]. All chemicals were purchased from Alfa Aesar or Sinopharm Chemical Reagent Co. and used without further purification. Briefly, 5 g of graphite was first dispersed in 120 mL H2SO4 with ice-cold bath and magnetic stirring. 15 g of KMnO4 was then slowly added while keeping the bath temperature under 10 °C. After homogeneous mixing to form dark-green-color suspension, the temperature was increased to 35 °C and kept for 2 h, and then further to 50 °C for 2 h. Then, 150 mL distilled water was carefully dropped into the mixture, over a period of 30 min, to realize the intercalation of water molecules into graphite oxide. Two hours later, 500 mL iced distilled water was added to stop the oxidation process. To reduce the residual permanganate and manganese dioxide to soluble manganese sulfate, 10 mL of H2O2 was added while keeping stirring for 30 min. The obtained yellow graphite oxide suspension was subjected to centrifugation several times in 3% HCl and distilled water at 5,000 r min−1 to remove the acid and residual manganese ions, and then transferred into the dialysis tube for several days to further purify. Finally, certain amount of graphite oxide slurry was diluted and ultra-sonicated in deionized water to obtain GO dispersion with a concentration of 1 mg mL−1. 2.2 Hydrothermal reduction of GO Sixty mL diluted GO dispersion (0.1 mg mL−1) was put into 100 mL Teflon stainless steel autoclaves, and then heated to 120, 150 and 180 °C for 12 h,

respectively, for hydrothermal reduction. The rGO samples obtained were named as 120-HTrGO, 150-HTrGO and 180-HTrGO, respectively. 2.3 Membrane fabrication GO and HTrGO membranes were prepared by vacuum filtration of 25 mL GO and HTrGO aqueous dispersion (5 mg L−1), respectively, on polyether sulfone (PES) microfiltration membrane. After that, the membranes were dried in air at room temperature to maintain their regular layered structures. The PES membranes act as supports to enhance the mechanical properties of ultrathin GO-based membranes. In these PES supported membranes, the mass loading of GO/HTrGO was 100 mg m−2. To characterize the structure of the membranes, we also fabricated thick free-standing membranes by soaking the PES supported membranes in water for several minutes to separate the PES substrates, in which the mass loading of GO/HTrGO was 2.4 g m−2. 2.4 Materials characterization Scanning electron microscopy (SEM, Nova NanoSEM 430, at 10 kV) was used to characterize the structure of GO/HTrGO nanosheets and their assembled membranes. The chemical composition was characterized by X-ray photoelectron spectroscopy (XPS) on ESCALAB250 using Al Kα radiation, and all spectra were calibrated to the binding energy of C=C bonds (284.6 eV). Raman spectra were taken on LabRAM HR800 (532 nm laser, spot size ~1 µm2). X-ray diffractometer (XRD, D-MAX/2400 using Cu Kα radiation) was used to analyze the interlayer spacing of GO/HTrGO membranes. Hydrophilicity of the membranes was analyzed at 298 K using a contact angle dropmeter, and the zeta potential was measured on Malvern Zetasizer Nano ZS90. The dye concentration in feed and permeate solutions were determined by a UV-vis spectrophotometer (Varian Carry 5000). 2.5 Measurements on the water permeance and dye rejection of the membranes The water permeance and dye rejection of the membranes were measured on a

vacuum-assisted dead-end filtration device with an effective area of 10.75 cm2 at room temperature, and the applied pressure was 0.95 bar (1 bar = 0.1 MPa). The contention of feed dye solution was 10 mg L−1. The water permeance J (L m−2 h−1 bar−1) and rejection R (%) were calculated according to Eqs. (1) and (2), respectively. ,

(1) ,

(2)

where V (L) is the volume of permeated water, A (m2) is the effective membrane area, Δt (h) is the permeate time, P (bar) is the applied pressure, and Cp and Cf are the concentration of permeate and feed solution, respectively.

3. Results and discussion As shown in Fig. 1a, the GO nanosheets we used mostly have a lateral size of 1‒2 μm. For hydrothermal reduction [36], under the high temperature and internal pressure conditions, supercritical water acts as reducing agent and offers a green chemistry alternative to other chemical agents such as hydrazine hydrate or hydroiodic acid. Figure 1b, c shows the GO dispersion in water before and after hydrothermal treatment. It can be found that after hydrothermal treatment at 120 ºC for 12 h, the yellow GO dispersion becomes black color, indicating that the GO nanosheets are reduced. More importantly, it is worth noting that the rGO nanosheets are homogenously dispersed in water without any visible particles even after one week, which is essentially important for the fabrication of uniform membranes. In contrast, the GO dispersions after reduction by other reducing agents such as hydrazine hydrate [37] usually show severe agglomeration of rGO nanosheets. The zeta potential is an important parameter to evaluate the electrical properties of interfacial layers in dispersion. In general, a zeta potential between 30 and 40 mV corresponds to appropriate stability while higher than 40 mV is required to get high

stability for colloidal suspensions [38]. We then measured the zeta potential of 120-HTrGO, 150-HTrGO and 180-HTrGO dispersions, which give the values of -43.1, -35.6 and -33.9 mV, respectively (Fig. 1d). The high zeta potential is responsible for the homogeneous dispersion of rGO nanosheets observed in Fig. 1c.

Fig. 1. Characterization of the GO and HTrGO dispersions. (a) SEM image of GO nanosheets. Photos of

GO dispersion (b) and

120-HTrGO dispersion (c) after 1

week without sonication. Zeta potential (d) and UV-vis absorption spectra (e) of GO, 120-HTrGO, 150-HTrGO and 180-HTrGO dispersions.

Figure 1e shows the UV-vis absorption spectra of GO and HTrGO dispersion. GO dispersion shows a strong absorption peak in the range of 230 to 270 nm, corresponding to π→π* transitions of aromatic C–C bonds, and a shoulder peak at 301 nm, corresponding to the n→π* transitions of C=O bonds. After hydrothermal treatment, the shoulder peak disappears, indicating the removal of C=O bonds. Moreover, the strong absorption peak upshifts with increasing the hydrothermal treatment temperature, indicating the gradual restoration of conjugate system [39]. We then analyzed the chemical composition and bonding states of GO and HTrGO nanosheets by XPS. The C/O atomic ratio of GO nanosheets is ~2.39, and the corresponding values of 120-HTrGO, 150HTrGO, and 180HTrGO nanosheets are 4.32, 5.92, and 6.04, respectively. Figure 2a-d show the fitted C 1s spectra of GO and HTrGO membranes. It can be found that the intensity of C−O bonds (286.6 eV, hydroxyl and epoxy groups) and C=O bonds (288.1 eV, carbonyl groups) decreases with increasing the hydrothermal treatment temperature. These results indicate that all the HTrGO samples are partially reduced and the reduction degree can be easily controlled. The residual oxygen functional groups endow HTrGO dispersion a high zeta potential and good dispersion as shown above. Moreover, with increasing the reduction temperature, the full width at half maximum (FWHM) of C 1s peak decreases from 1.75 to 1.41, 1.36 and 1.34, respectively. The narrower FWHM confirms that the amount of in-plane oxygen content is reduced and the more ordered

sp2 carbon structures are generated [40].

Fig. 2. Characterization of GO and HTrGO nanosheets. High-resolution C 1s XPS spectra of GO (a), 120-HTrGO (b), 150-HTrGO (c), and 180-HTrGO (d) nanosheets. (e) Raman spectra of GO, 120-HTrGO, 150-HTrGO, and 180-HTrGO nanosheets.

Raman spectroscopy is a highly sensitive technique to identify the microstructure of carbon-based materials. The G peak is due to bond stretching of sp2 carbon pairs in both rings and chains, and the D peak is due to the breathing mode of aromatic sp2 carbon rings and requires defects for its activation [41]. As shown in Fig. 2e, compared to the GO samples, the intensity ratio of D band to G band (ID/IG) first increases and then decreases when increasing the reduction degree. The increase in ID/IG of 120-HTrGO indicates the increase of the number of isolated pristine sp2 domains with the removal of oxygen functional groups [15]. Combined with the XPS data, the subsequent decrease in ID/IG of 150-HTrGO and 180-HTrGO suggest that more sp2 domains are generated and some of them joint together to form big sp2 domains when more oxygen functional groups are removed [36]. We then used XRD to study the structure of GO and HTrGO membranes (Fig. 3a). GO membranes shows a sharp XRD peak at 11.5°, corresponding to an interlayer spacing of 0.76 nm [17, 23]. Interestingly, 120-HTrGO membrane shows a broad peak at 24.3º in addition to the strong peak at 11.7°. This suggests that some sp2 domains are stacked together with the increase in the number of isolated sp2 domains [42]. As the sp2 domains expands with further reduction, the relative intensity of the 24.3º XRD peak increases, indicating more stacking occurs. Although the GO nanosheets are not completely reduced at 180 ºC, the 180-HTrGO membranes show only one strong XRD peak at 24.3º, which is similar to the completely reduced GO membranes [34]. We further used SEM to characterize the cross-sectional structure of the different membranes fabricated by the same mass loading of GO/HTrGO nanosheets. It can be seen that the pure GO membranes show densely stacked and well aligned layered structure with a thickness of 1.36 μm (Fig. 3b). In contrast, the 120-HTrGO

membrane shows irregular laminated structure with varying interlayer spacing (Fig. 3a). Interestingly, it shows a similar thickness (1.25 μm) with GO membranes (Fig. 3c) although some sp2 domains are restacked based on XRD data. This indicates that there are still enough residual oxygen functional groups as spacer to ensure large interlayer spacing in most regions. As the reduction temperature is increased to 150 and 180 °C, the rGO nanosheets becomes strongly restacked and the thickness of the membranes are greatly reduced to 1.16 ad 0.84 m, respectively (Fig. 3d,e).

Fig. 3. Characterization of GO and HTrGO membranes. (a) XRD patterns of GO, 120-HTrGO, 150-HTrGO, and 180-HTrGO membranes. Typical cross-sectional SEM images of GO (b), 120-HTrGO (c), 150-HTrGO (d), and 180-HTrGO (e) membranes that were fabricated with the same mass loading of GO/HTrGO nanosheets.

All the above results indicate that (1) weak reduction increases the number of pristine sp2 domains in rGO nanosheets while keeping the large interlayer spacing of the GO membranes in most regions at the same time, and (2) moderate reduction results in the expansion of sp2 domains and consequently severe stacking of rGO nanosheets although there are still many oxygen functional groups. As reported previously, strong reduction by traditional chemical [34] or thermal treatments [43] can further improve the ordering and stacking of rGO nansheets, evidenced by a sharp XRD peak at round 26º, because of the complete elimination of oxygen functional groups and recovery of π-conjugation. It is well accepted that, in GO membranes, the oxidized domains act as spacer to provide relatively large interlayer distance to accommodate water molecules and the pristine graphitic sp2 domains facilitate rapid water permeability by nearly frictionless flow [16]. Therefore, it is expected that 120-HTrGO membranes should have greatly improved water permeance compared to pure GO membranes, moderately reduced 150-HTrGO and 180-HTrGO membranes, and strongly reduced rGO membranes [34], because they have large interlayer spacing and more pristine sp2 domains simultaneously. At the same time, the similar interlayer spacing with GO membranes ensures the excellent separation performance of

120-HTrGO membranes.

For the system of osmotically moderated aqueous separation and organic solvent nanofiltration, the wettability of membrane surface with the solvents plays an important role in the permeability and fouling resistant property. In general, a hydrophilic surface is favorable for water permeation and antifouling performance [3]. We then measured the water contact angles of GO and HTrGO membranes to estimate the surface hydrophilicity. As shown in Fig. 4a, the GO membrane shows hydrophilic surface with an average water contact angle of about 46°. The hydrophilicity of the membranes decreases with increasing the reduction degree. However, the 120-HTrGO, 150-HTrGO and 180-HTrGO membranes are still hydrophilic, which show contact angle of 76°, 87° and 88°, respectively.

Fig. 4. Wettability and separation performance of GO and HTrGO membranes. (a) Photos of a water droplet on GO, 120-HTrGO, 150-HTrGO and 180-HTrGO membrane surface at room temperature. The contact angle (CA) was given in each case. (b) The water permeance of GO, 120-HTrGO, 150-HTrGO and 180-HTrGO membranes in pure water and methylene blue (MB) solution. (c) The water permeance and rejection rate of 120-HTrGO membrane in different dye solutions. (d) The water permeance and dye rejection rate of 120-HTrGO membrane fabricated with different mass loading of HTrGO nanosheets. (e) The UV-Vis spectra of EB filtered through the 120-HTrGO membrane fabricated with a mass loading of HTrGO nanosheets of 100 mg m–2.

We then performed dead-end pressure filtration measurements to evaluate the water permeance of GO and HT-rGO membranes. To rule out the influence of membrane thickness, all the membranes were fabricated with the same mass loading of GO/HTrGO nanosheets (100 mg m−2). As shown in Fig. 4b, the GO membrane shows a water permeance of 15.5 L m−2 h−1 bar−1, which is similar to those reported previously [42]. As expected, the 120-HTrGO membrane shows a greatly improved

water permeance as large as 56.3 L m−2 h−1 bar−1 due to the large interlayer spacing and more frictionless water flow channels formed by pristine sp2 domains. However, the 150-HTrGO and 180-HTrGO membranes are almost impermeable to water with a water permeance of ~1 and ~0.01 L m−2 h−1 bar−1, respectively, because of the greatly narrowed interlayer spacing. To characterize the membrane separation performance, we tested the filtration of 120-HTrGO membranes to various dyes with different charges and molecule size, including MB (negatively charged, 10 mg L–1), Rhodamine B (RhB, negatively charged, 10 mg L–1), evens blue (EB, positively charged, 10 mg L–1) and eriochrome black T (EBT, negatively charged, 10 mg L–1). The concentration of dyes in the feed and permeate solutions were monitored by UV-vis analysis. The results show that the water permeance of the 120-HTrGO membranes in dye solution is only a bit lower than that in pure water, with a minimum value of ~40 L m−2 h−1 bar−1 (Fig. 4c). Importantly, such membranes show very good rejection performance for all the dye molecules investigated. The rejection rate is 97%, 95.3%, 98%, and 99.5%, respectively, for EB, MB, RhB, and EBT. Considering the charges and molecule size of these dyes, we suggest that the steric hindrance effect plays the most important role in the excellent dye rejection performance of our membranes [29, 42]. The thickness-dependent separation performance of the 120-HTrGO membranes was further studied. As shown in Fig. 4d, the water permeance increases from 20.1 to 76.0 L m−2 h−1 bar –1 when the mass loading of HTrGO nanosheets is decreased from 150 to 75 mg m–2, corresponding to the thickness from ~110 to ~50 nm. This is because of the shortening of the pathway for liquid transport [31, 42, 44, 45]. However, the thinnest 120-HTrGO membrane shows poor dye rejection of about 36%. Therefore, the 68-nm-thick 120-HTrGO membranes with the mass loading of 100 mg m–2 show optimal water permeance and dye rejection at the same time. Figure 4e clearly shows its excellent rejection performance. Considering the different operational environments of nanofiltration membranes, we further tested the structure stability as well as water permeance and dye rejection rate of GO and 120-HTrGO membranes under different pH environments. As shown

in Fig. 5a, the GO membrane changes from brown color in pure water to bright yellow in acid solution and dark black in alkali solution, indicating the structure changes. In contrast, the 120-HTrGO membrane shows very good stability in different solutions, with no obvious color and structure change being observed. As a result, GO membrane shows a dramatic decrease in water permeance in both acid (3.2 L m−2 h−1 bar−1) and alkali solutions (5.4 L m−2 h−1 bar−1) compared to that in pure water (13.4 L m−2 h−1 bar−1), as shown in Fig. 5b. In acid solution, protonation of carboxylic acid in GO nanosheets can narrow the regular nanochannel of GO membranes by weakening the interlayer electrostatic repulsion forces [46]. In alkali solution, with the diffusion of Na+ into GO membrane nanochalles, both the cation–π interaction [24] and ionic screening effects [46] can enhance the interlayer interactions of GO membranes, leading to the decrease in interlayer spacing. Therefore, we suggest that the narrowing of nanochannels of GO membranes in both acid and alkali solution is responsible for the great decrease in water permeance. Because of the same reason, the rejection of GO membranes to RhB is slightly increased from 90% to 93% and 95% (Fig. 5b). In sharp contrast, the 120-HTrGO membrane only shows a slight decrease in both water permeance and rejection in both acid and alkali solutions (Fig. 5b). It still remains a water permeance over 47 L m−2 h−1 bar−1 and rejection over 95% to RhB, which are about 10 times higher than and comparable to those of GO membranes, respectively. The good structure stability and high separation performance under different environments open up the possibility of 120-HTrGO membranes for a broad range of nanofiltration applications.

Fig. 5. Stability and separation performance of GO and HTrGO membranes in various aqueous solutions. (a) Photos of GO (left) and 120-HTrGO (right) membranes in aqueous solutions with different pH for one week. (b) The water permeance and rejection to RhB molecules of GO and 120-HTrGO membranes in aqueous solutions with different pH values.

4. Conclusion In conclusion, we demonstrate the controlled mild reduction of GO nanosheets by hydrothermal reaction and study the influence of reduction degree on the structure and separation performance of rGO membranes. It was found that weak reduction retains the good dispersion and hydrophilicity of GO nanosheets. More importantly, it increases the number of pristine graphitic sp2 domains in rGO nanosheets while keeping the large interlayer spacing of GO membranes in most regions at the same time. The resultant membranes show a high water permeance of 56.3 L m−2 h−1 bar−1, which is about 4 times and over 104 times larger than those of the GO and completely reduced rGO membranes, respectively, and high rejection over 95% for various dyes. Furthermore, they show better structure stability and more superior separation performance than GO membranes in acid and alkali environments. This work gives more understanding on the influence of reduction on the structure of GO membranes and provides a new strategy for designing high-performance GO-based separation membranes and more evidence on their mass transport and rejection mechanism. Compared with other two-dimenstion (2D) materials-based membranes, such as transition metal dichalcogenides and MXenes based membranes, the hydrothermal reduced GO membranes are more flexible on the chemical states and structure, which are important to meet different applications in filtration industry.

Conflicts of Interest The authors declare that they have no conflict of interest.

Acknowledgments This work is supported by the National Key Research and Development Program of

China (2016YFA0200101), the National Natural Science Foundation of China (51325205, 51290273, and 51521091), and Chinese Academy of Sciences (KGZD-EW-303-1, KGZDEW-T06, 174321KYSB20160011, and XDPB06).

References [1] Cohen-Tanugi D, Grossman JC. Water desalination across nanoporous graphene. Nano Lett, 2012, 12: 3602-3608 [2] Jiang Z, Karan S, Livingston AG. Water transport through ultrathin polyamide nanofilms used for reverse osmosis. Adv Mater, 2018, 30: 1705973 [3] Zhang RN, Liu YN, He MR, et al. Antifouling membranes for sustainable water purification: Strategies and mechanisms. Chem Soc Rev, 2016, 45: 5888-5924 [4] Marchetti P, Solomon MFJ, Szekely G, et al. Molecular separation with organic solvent nanofiltration: A critical review. Chem Rev, 2014, 114: 10735-10806 [5] Yao X, Guo L, Chen X, et al. Filtration-based synthesis of micelle-derived composite membranes for high-flux ultrafiltration. ACS Appl Mater Interfaces, 2015, 7: 6974-6981 [6] Mi B. Materials science. Graphene oxide membranes for ionic and molecular sieving. Science, 2014, 343: 740-742 [7] Park HB, Kamcev J, Robeson LM, et al. Maximizing the right stuff: the trade-off between membrane permeability and selectivity. Science, 2017, 356: 1138-1148 [8] Boukhvalov DW, Katsnelson MI, Son YW. Origin of anomalous water permeation through graphene oxide membrane. Nano Lett, 2013, 13: 3930-3935 [9] Sun P, Zhu M, Wang K, et al. Selective ion penetration of graphene oxide membranes. ACS Nano, 2013, 7: 428-437 [10] Amadei CA, Vecitis CD. How to increase the signal-to-noise ratio of graphene oxide membrane research. J Phys Chem Lett, 2016, 7: 3791-3797 [11] Deng J, You Y, Bustamante H, et al. Mechanism of water transport in graphene oxide laminates. Chem Sci, 2017, 8: 1701-1704 [12] Jiao S, Xu Z. Non-continuum intercalated water diffusion explains fast permeation through graphene oxide membranes. ACS Nano, 2017, 11: 11152 [13] Chua CK, Pumera M. Chemical reduction of graphene oxide: A synthetic chemistry viewpoint. Chem Soc Rev, 2014, 43: 291-312 [14] Ren W, Cheng HM. The global growth of graphene. Nat Nanotechnol, 2014, 9: 726 [15] Pei S, Wei Q, Huang K, et al. Green synthesis of graphene oxide by seconds timescale water electrolytic oxidation. Nat Commun, 2018, 9: 145 [16] Nair RR, Wu HA, Jayaram PN, et al. Unimpeded permeation of water through helium-leak-tight graphene-based membranes. Science, 2012, 335: 442-444 [17] Joshi RK, Carbone P, Wang FC, et al. Precise and ultrafast molecular sieving through graphene oxide membranes. Science, 2014, 343: 752-754

[18] Abraham J, Vasu KS, Williams CD, et al. Tunable sieving of ions using graphene oxide membranes. Nat Nanotechnol, 2017, 12: 546 [19] Chong JY, Wang B, Li K. Water transport through graphene oxide membranes: the roles of driving forces. Chem Commun, 2018, 54: 2554-2557 [20] Wang Z, Tu Q, Zheng S, et al. Understanding aqueous stability and filtration capability of mos2 membranes. Nano Lett, 2017, 17: 7289-7298 [21] Zheng S, Tu Q, Urban JJ, et al. Swelling of graphene oxide membranes in aqueous solution: Characterization of interlayer spacing and insight into water transport mechanisms. ACS Nano, 2017, 11: 6440-6450 [22] Yeh CN, Raidongia K, Shao JJ, et al. On the origin of the stability of graphene oxide membranes in water. Nat Chem, 2015, 7: 166-170 [23] Hung WS, Tsou CH, De Guzman M, et al. Cross-linking with diamine monomers to prepare composite graphene oxide-framework membranes with varying d-spacing. Chem Mater, 2014, 26: 2983-2990 [24] Chen L, Shi G, Shen J, et al. Ion sieving in graphene oxide membranes via cationic control of interlayer spacing. Nature, 2017, 550: 380-383 [25] Yu W, Yu T, Graham N. Development of a stable cation modified graphene oxide membrane for water treatment. 2D Mater, 2017, 4: 045006 [26] Hu M, Mi B. Enabling graphene oxide nanosheets as water separation membranes. Environ Sci Technol, 2013, 47: 3715-3723 [27] Thebo KH, Qian X, Zhang Q, et al. Highly stable graphene-oxide-based membranes with superior permeability. Nat Commun, 2018, 9: 1486 [28] Huang H, Song Z, Wei N, et al. Ultrafast viscous water flow through nanostrand-channelled graphene oxide membranes. Nat Commun, 2013, 4: 2979 [29] Goh K, Jiang W, Karahan HE, et al. All-carbon nanoarchitectures as high-performance separation membranes with superior stability. Adv Funct Mater, 2015, 25: 7348-7359 [30] Han Y, Jiang Y, Gao C. High-flux graphene oxide nanofiltration membrane intercalated by carbon nanotubes. ACS Appl Mater Interfaces, 2015, 7: 8147-8155 [31] Gao SJ, Qin H, Liu P, et al. Swcnt-intercalated go ultrathin films for ultrafast separation of molecules. J Mater Chem A, 2015, 3: 6649-6654 [32] Nam YT, Choi J, Kang KM, et al. Enhanced stability of laminated graphene oxide membranes for nanofiltration via interstitial amide bonding. ACS Appl Mater Interfaces, 2016, 8: 27376-27382 [33] Song X, Zhou Q, Zhang T, et al. Pressure-assisted preparation of graphene oxide quantum dot-incorporated reverse osmosis membranes: antifouling and chlorine resistance potentials. J Mater Chem A, 2016, 4: 16896-16905 [34] Su Y, Kravets VG, Wong SL, et al. Impermeable barrier films and protective coatings based on reduced graphene oxide. Nat Commun, 2014, 5: 4843 [35] Chen J, Yao BW, Li C, et al. An improved hummers method for eco-friendly synthesis of graphene oxide. Carbon, 2013, 64: 225-229 [36] Zhou Y, Bao Q, Tang LAL, et al. Hydrothermal dehydration for the “green” reduction of exfoliated graphene oxide to graphene and demonstration of tunable

optical limiting properties. Chem Mater, 2009, 21: 2950-2956 [37] Nguyen ST, Ruoff RS, Stankovich S, et al. Synthesis of graphene-based nanosheets via chemical reduction of exfoliated graphite oxide. Carbon, 2007, 45: 1558-1565 [38] Krishnamoorthy K, Veerapandian M, Yun K, et al. The chemical and structural analysis of graphene oxide with different degrees of oxidation. Carbon, 2013, 53: 38-49 [39] Zhang J, Yang H, Shen G, et al. Reduction of graphene oxide vial-ascorbic acid. Chem Commun, 2010, 46: 1112-1114 [40] Voiry D, Yang J, Kupferberg J, et al. High-quality graphene via microwave reduction of solution-exfoliated graphene oxide. Science, 2016, 353: 1413-1416 [41] Wu ZS, Ren W, Gao L, et al. Synthesis of graphene sheets with high electrical conductivity and good thermal stability by hydrogen arc discharge exfoliation. ACS Nano, 2009, 3: 411-417 [42] Han Y, Xu Z, Gao C. Ultrathin graphene nanofiltration membrane for water purification. Adv Funct Mater, 2013, 23: 3693-3700 [43] Ghosh T, Biswas C, Oh J, et al. Solution-processed graphite membrane from reassembled graphene oxide. Chem Mater, 2012, 24: 594-599 [44] Karan S, Samitsu S, Peng X, et al. Ultrafast viscous permeation of organic solvents through diamond-like carbon nanosheets. Science, 2012, 335: 444-447 [45] Kim SJ, Kim DW, Cho KM, et al. Ultrathin graphene oxide membranes on freestanding carbon nanotube supports for enhanced selective permeation in organic solvents. Sci Rep, 2018, 8: 1959 [46] Huang H, Mao Y, Ying Y, et al. Salt concentration, ph and pressure controlled separation of small molecules through lamellar graphene oxide membranes.Chem Commun, 2013, 49: 5963-5965

Qing Zhang received her B.S. degree from Jilin University in 2015. Now she is a master student at the Institute of Metal Research, Chinese Academy of Science, and University of Science and Technology of China. Her current research interests include design and synthesis of graphene-based membranes for molecular separation.

Wencai Ren is a professor in materials science at the Institute of Metal Research, Chinese Academy of Sciences, and University of Science and Technology of China. His research interests mainly focus on the synthesis of graphene and other two-dimensional materials and their applications in energy storage, composites, separation membranes and optoelectronics. He has received several awards including the 2rd Class of the National Natural Science Award (2017, 2006) and the National Science Fund for Distinguished Young Scholars (2013).