Graphene oxide membrane for liquid phase organic molecular separation

CARBON

7 7 ( 2 0 1 4 ) 9 3 3 –9 3 8

Available at www.sciencedirect.com

ScienceDirect journal homepage: www.elsevier.com/locate/carbon

Graphene oxide membrane for liquid phase organic molecular separation Renlong Liu a, Girish Arabale a, Jinseon Kim a, Ke Sun b, Yongwoon Lee a, Changkook Ryu a, Changgu Lee a,b,* a

School of Mechanical Engineering, Sungkyunkwan University, 2066, Seobu-ro, Jangan-gu, Suwon, Gyeonggi, Republic of Korea Sungkyunkwan Advanced Institute of Nanotechnology, Sungkyunkwan University, 2066, Seobu-ro, Jangan-gu, Suwon, Gyeonggi, Republic of Korea b

A R T I C L E I N F O

A B S T R A C T

Article history:

The selective permeation of organic solvents and water through graphene oxide (GO) mem-

Received 28 February 2014

branes has been demonstrated. Water was found to permeate through GO membranes fas-

Accepted 3 June 2014

ter than various alcohols. The permeation rates of ethanol, 1-propanol and 2-propanol (IPA)

Available online 13 June 2014

are about 80 times lower than that of water. Taking advantage of the differences in the permeation rates, we separated water from the alcohols and obtained alcohols with high purity. For ethanol and 1-propanol, binary solutions of the alcohol and water were filtered efficiently to produce alcohols with concentration of about 97%. However, the selectivity of the filtration of methanol is significantly lower than those of the other alcohols. To understand the mechanism we followed the structural changes in the GO membranes by X-ray diffraction analysis. From the X-ray diffraction results we speculate that the selectivity of the permeation of water and alcohols is closely related to the molecular sizes of the solvents and their polarity. In order to demonstrate the potential applications of this process for the selective removal of water from aqueous organic mixtures, we performed the separation of water from a bio-oil containing 73% of water. The majority of the water was filtered out resulting in a higher purity bio-oil. Ó 2014 Elsevier Ltd. All rights reserved.

1.

Introduction

The carbon atoms in graphene are so tightly bonded that no gas molecules can permeate through the one atom thick membrane of graphene [1–4]. If atomic defects are generated in the membrane through oxidation or chemical reaction methods, some gas or liquid molecules can pass through it depending on the sizes of the defects and the molecules, which suggests that graphene membranes can be used as molecular filters [5–8]. However, single layer graphene membranes with defects are mechanically too weak to be * Corresponding author. E-mail address: [email protected] (C. Lee). http://dx.doi.org/10.1016/j.carbon.2014.06.007 0008-6223/Ó 2014 Elsevier Ltd. All rights reserved.

used in large volume molecular separation. In contrast, graphene oxide (GO) paper with a stack structure [9] has good chemical stability, excellent mechanical stiffness and strength [10–12], and can easily be produced with large sizes for macroscale applications [13,14]. Functional groups in the GO structure open gaps between the stacked platelets that are approximately 1 nm wide [15]. These gaps in GO paper can act as nanopores for molecular transport [16], and, depending on the gap size or functional groups in the GO membrane, can permit the selective passage of gas or liquid molecules through the pores [17–19]. Thus GO membranes

934

CARBON

7 7 ( 2 0 1 4 ) 9 3 3 –9 3 8

can be used as molecular sieves. One of the interesting studies which has been done recently reported that GO membrane does not permit the permeation of helium gas molecules whereas water molecules can pass through it easily due to the hydrophilic characteristics of the material [17]. Although that experiment demonstrated the possibility of the selective filtration of gas molecules using GO membranes, their sieving characteristics with respect to liquid phase molecules have not been investigated systematically. Chemical feed stock, solvents, and the products of reactions in the liquid phase are often complex mixtures of organic solvents and water [20]. The separation of these mixtures remains a difficult task for industry, and is mostly performed by energy-consuming distillation techniques [21,22] sometimes combined with the use of entrainers [23,24]. Very recently, Tang et al. [25] and Talyzin et al. [26] demonstrated the selective permeation of water and ethanol through GO membranes. In this context GO membranes can offer many advantages in liquid phase separations because of their inherent porous structures that creates pathways for the selective passage of different molecules resulting in the separation on the basis of their permeation rates. The immediate advantages are higher selectivity, lower energy consumption, moderate cost to performance ratio, and modular design. In this study, we assessed the effectiveness of graphene oxide membranes in the liquid phase separation of mixtures of alcohol and water. First, we measured the permeation rates of water and various alcohols with varying carbon numbers, then performed filtering experiments on water and alcohol mixtures in order to observe the selective permeation of the liquid molecules and to investigate the mechanism behind the selectivity. The permeation rate of water was found to be higher than those of the alcohols and we speculate that this is due to the differences in their molecular sizes and polarity. We also tested the filtration of a mixture of many kinds of organic molecules [27] and water and showed that it is feasible to use GO as a versatile liquid phase filter membrane.

2.

Experimental procedure

GO were prepared with modified Hummer’s method and characterized by using various methods (Figs. S1–S5) [28–30]. Graphene oxide membranes were prepared by filtering diluted graphene oxide dispersions through a PTFE filter with 0.5 lm pore size [31,32]. The thickness of the membranes was approximately 5 lm for all experiments (Fig. S6). These membranes were all prepared from GO solutions of the same concentration and quantity, and thus their thicknesses were almost identical. A series of liquid phase permeation experiments were conducted with a modified vacuum filtering device and the GO membranes, as shown in the schematic diagram in Fig. S7. Although freestanding GO membranes possess sufficient mechanical strength for the applications tested in our experiments, we used GO membranes on top of PTFE substrates to extend their life time and check their reusability. The permeation rates of methanol/ethanol/ propanol and water through the PTFE filter were observed to be five orders of magnitudes higher than those through the

GO membranes, so the rate of permeation through the GO/PTFE stack filter structure is determined by the GO membranes. The permeation rates were calculated by following the changes in the density as well as weight loss of the water/alcohol mixtures over a time of one day.

3.

Results and discussion

We varied the alcohol concentration in water from 0 to 100 wt.% in order to investigate the filtration selectivity. Here, alcohol concentrations of 0 wt.% and 100 wt.% correspond to pure water and pure alcohol. The permeation rates of the pure alcohols were found to be much lower than that of water, as shown in Fig. 1. The permeation rates of pure water and the pure alcohols are as follows; water: methanol: ethanol: IPA: 1-propanol = 1:0.19:0.017:0.015:0.012 (see Fig. S8 for clear comparison between alcohols). The methanol permeation rate is the highest of those of the alcohols. As the concentration of alcohol in the solutions of water and ethanol/isopropanol/1-propanol decreases, the permeation rate of the alcohol increases (Fig. 1(b)–(d)). We speculate that water assists the permeation of the alcohols because these alcohols dissolve in water. Pure ethanol/isopropanol/1propanol molecules pass through the nanochannels in GO paper quite slowly compared to water, as mentioned above. However, when they are mixed with water, water molecules adhered to these alcohol molecules through van der Waals bonds or hydrogen bonds can drag the alcohol molecules through the nanogaps between the GO flakes, which usually have a size of approximately 1 nm, because of the high affinity between water and the functional groups in GO. If the solubility of the alcohol in water is lower, this mechanism is likely to be less effective. Hence we also tested mixtures of water and alcohols with large molecular sizes, i.e., butanol and pentanol. These alcohols can be mixed with water to concentrations of approximately 7.6 wt.% (1-butanol) and 2.2 wt.% (1-pentanol) in water at 20 °C [33]. We added the alcohols to the solutions up to the highest possible concentrations in water and the filtration results of the two alcohols are shown in Fig. S9 along with other alcohols. In the filtering test, since the permeation rate of water is much higher than these alcohols, some of the alcohol is separated from the solution as water permeates and we observed the formation and level change of boundary of water and the undissolved alcohol. The permeation of the alcohols with higher molecular weight is almost completely blocked by the GO film and the permeation rate 1 or 2 orders lower than other alcohols, so the unfiltered solutions are almost pure alcohol (i.e., the mass and density of the unfiltered liquid are the same as those of the alcohol previously added into the binary mixture). These results show that the solubility of the alcohol in water is important for GO paper permeation. The trend in the permeation rates for the components of the solutions of methanol and water is opposite to those for the other three alcohol mixtures with water (Fig. 1(a)). As the concentration of methanol in the mixtures increases, the methanol permeation rate increases. These results indicate that water prevents the permeation of methanol through the GO membrane.

12

water methanol

12

water ethanol

10

2

10 8

-6

6 4 2 0

0

25

50

75

100

8 6 4 2 0

0

25

12

water isopropanol

-6

2

10 8 6 4 2 0

0

25

50

75

50

75

100

Ethanol content (wt%)

100

Permeation rate (10-6 mm×g/cm2×s×bar)

Methanol content (wt%) Permeation rate (10 mm×g/cm ×s×bar)

935

7 7 (2 0 1 4) 9 3 3–93 8

Permeation rate (10 mm×g/cm ×s×bar)

Permeation rate (10-6 mm×g/cm2×s×bar)

CARBON

12

water 1-propanol

10 8 6 4 2 0

0

25

50

75

100

1-propanol content (wt%)

IPA content (wt%)

Fig. 1 – Rates of permeation of binary solutions of water and (a) methanol (b) ethanol (c) isopropanol (d) 1-propanol. The concentration error is the standard deviation of 20 measured data. (A color version of this figure can be viewed online.)

Table 1 – Solvents and their physical properties. Solvent Water Methanol Ethanol 1-propanol 2-propanol (IPA) 1-butanol 1-pentanol

Kinetic diameter ˚ @298 K) [35] (A

Relative polarity [34]

2.6 3.6 4.5 4.7 4.7 5.0 6.7

1.000 0.762 0.654 0.617 0.546 0.586 0.568

measurements of the variations in concentration (or density) over time, and second, X-ray diffraction measurements of the gaps between the GO flakes in water or alcohol solvent. First, the variation with time in the concentration of the alcohol in the water–alcohol (50–50%) binary mixtures is different for methanol (Fig. 2). We monitored the concentrations of the three alcohols, methanol, ethanol, and 1-propanol, for 7 days while filtering the binary mixtures. After 5 days, 97 wt.% of 1-propanol and ethanol were obtained from their

100 95 90

Concentration (wt%)

Although the molecular size of methanol is greater than that of water, it is less than those of the other alcohols (refer to Table 1 [34,35]). The effective molecular size of methanol is close to that of water, so methanol molecules might compete with water to pass through the nanochannels in the GO membrane. The gap size between GO flakes is approximately 8– ˚ , so it should be possible for most alcohol molecules to 10 A pass through the GO nanochannels. However, higher energies will be required for larger molecules to penetrate the channels, thus resulting in lower penetration rates. However, molecular size might not be the only factor determining permeation through the GO membrane. Polarity can also affect the permeation rate because there are various kinds of functional groups on the surfaces of the GO flakes that will affect the adsorption of polar liquid molecules (refer to Table 1). In order to investigate the permeation behaviors of the alcohols further, we carried out two other experiments: first,

85 80 75 70 65 60

methanol ethanol 1-propanol

55 50 0

1

2

3

4

5

6

7

Time (d) Fig. 2 – The variations with time in the concentrations of the alcohol-water binary solutions. (A color version of this figure can be viewed online.)

936

CARBON

7 7 ( 2 0 1 4 ) 9 3 3 –9 3 8

50 wt.% alcohol/water mixtures, which demonstrates the effective separation of water from these alcohols. In contrast, the dehydration of the methanol/water mixture progressed much more slowly, not saturating even after 7 days. This result again indicates that methanol is competing with water to permeate the membrane. Second, we measured the spacings between the GO flakes soaked in pure solvents by using X-ray diffraction (XRD). Samples of graphene oxide paper were loaded into pure liquid solvents (water, methanol, ethanol, 1-propanol, 1-butanol, and 1-pentanol) and examined with X-ray diffraction (XRD) at ambient temperatures, as shown in Fig. S10. The XRD graphs in Fig. 3 are for GO paper soaked in these solvents for 2 h and the graph of ‘ambient’ is from GO in ambient humidity without being soaked in a solvent and is for comparison with the wet GO membrane samples. Immersing the GO samples in liquid solvents results in a significant shift in the (0 0 1) reflection, which corresponds to an increase in the interlayer distance. It can also be seen that as the molecular weight of the alcohol molecule increases, the expansion of the corresponding soaked GO membrane increases. The spacing between the GO sheets in 1-pentanol-soaked GO is the largest ˚ , which is 2.4 times larger than of the samples, up to 19.9 A

that of the dry GO sample. Based on spacing and molecular diameter, and taking into account the observation that the electronic clouds around graphene sheets extend over a dis˚ [17,36], at least 2 layers of solvent molecules tance of a  3.5 A occupy the GO interlayer space so it is not a single layer of water or of alcohol molecules that is intercalated into one nanocapillary. Note that although XRD data were obtained at the atmospheric, same pressure condition (1 bar) across the GO were kept during all the filtration tests indicating the d-spacing ratio between the soaked conditions of alcohols should be maintained as long as the pressure conditions are kept the same. Thus the XRD data explain the penetration behavior of the liquid molecules shown in Fig. 1. When the number of carbon atoms increases, the volume of the alcohol molecule increases (the kinetic diameters of the solvent molecules are shown in Table 1). Hence, the GO interlayer distance is expected to be higher when intercalated with larger molecules, as shown in Fig. 3. However, this expansion requires energy to overcome the van der Waals forces between the GO layers [37]. Larger molecules such as pentanol, butanol, propanol, and ethanol have to overcome larger energy barriers, which prevents them from penetrating into the GO membrane. Since the interlayer distance of

Fig. 3 – XRD measurement result. (a) XRD patterns of GO membranes soaked in various pure alcohol solvents and water (b) analysis of XRD patterns. In (b), ‘d-spacing’ indicates the interlayer distance. The d-spacings were calculated by Bragg’s law: k = 2dsin(h) where k is the wavelength of the X-ray beam (0.154 nm), d is the distance between the adjacent GO flakes, h is the diffraction angle of (0 0 1) peak. (A color version of this figure can be viewed online.)

CARBON

7 7 (2 0 1 4) 9 3 3–93 8

˚ , which is quite close to the spacmethanol-soaked GO is 14 A ˚ , methanol molecules occupy a ing in water-soaked GO, 13.5 A volume comparable to that of water when intercalated with water into GO interlayers. Thus the permeation rate of methanol is not as low as that of other alcohols, which helps explain why methanol exhibits the poorest selectivity with respect to water among all the alcohol solvents we tested. Note also that the FWHM (full width at half maximum) of the (0 0 1) peaks of the solvent/GO samples decreases with increases in the molecular size, as shown in Fig. 3. Usually a XRD data gives the information of structure of the crystal. If the width of a peak is broad, we can infer that the crystal is not well ordered. In Fig. 3, the peak of water in GO is the broadest, which indicates that this system undergoes the largest decrease in GO platelet stack order. When water is introduced into a GO membrane, the swollen GO structure becomes semi-ordered in contrast to the well-ordered structure of dry GO before the introduction of water [38]. However, it seems that the alcohol-soaked GO membranes somehow retain well-ordered laminated structures. The FWHM of soaked GO appears to be closely related with polarity of the solvents (Table 1) [34]. The heavy decoration of graphene oxide sheets with polar oxygen-containing functional groups (hydroxyl, carbonyl, carboxyl; see the FTIR spectroscopy results in Fig. S2), increases the strengths of GO interactions with polar solvents [39]. In the case of water, GO stack structure seems to become less structured when intercalated. The poor stack order of water-soaked GO will permit water to pass the through it faster than other solvents lowering the penetration energy barrier. Therefore, in addition to the effects of molecular size, the permeation rate of a liquid through GO also appears to depend on the interaction between polar molecules and functional groups on the GO sheets. In other words, intermolecular forces between the liquid and the GO sheets drive solvent molecules to diffuse into the GO nanocapillaries and to some extent determine the diffusion rate. Higher polarity molecules are expected to diffuse faster through GO nanocapillaries under the same concentration gradient. Hence, for linear chained alcohols, molecular size and polarity synergistically affect the permeation performance. These two factors cooperate to affect permeation rather than competition. For branch chained alcohols, for example, 2-propanol, which possesses smaller molecular size than linear chained 1-butanol and 1-pentanol, permeates almost 10 times faster than them even though it has smaller polarity. We believe that in this case, molecular size is the dominating factor. In addition to our experiments with alcohols, we also carried out a bio-oil filtering test. The oil was produced by the pyrolysis of a biomass, Geodae-Uksae [27], which is a variety of Miscanthus sacchariflorus discovered in Korea. The pyrolysis process produces biochar, bio-oil, and bio-gas; we used the oil portion to test the GO membrane’s liquid separation of water. Biomass is becoming an important renewable energy source for the reduction of greenhouse gas emissions. However, the liquid phase bio-oil extracted from biomass contains a high proportion of water, which degrades its quality as a fuel. Hence it is of high importance to filter

937

water from the oil for increased heating value and improved ignition properties during combustion. The bio-oil is a mixture of more than 30 organic solvents (Table S1) and water. The as-made bio-oil contains 75.62 wt.% of water. With the same method as used for the alcohol mixtures, the bio-oil solution was filtered to remove water for 6 days. Elementary analysis and mass spectrometry show that this filtration process reduced the moisture content in the bio-oil from 73.62 wt.% to 22.96 wt.%. By modifying the functional groups and the thickness of the GO membrane, the water filtration efficiency can be increased even further. This bio-oil filtering test shows that GO membranes can be used to improve the quality of bio-oil at low cost and with a simple set-up.

4.

Conclusion

In conclusion, we have measured the rates of permeation of water and alcohols through a GO membrane and performed the filtration of alcohol and water mixtures. Water has a smaller molecular size and high polarity, so its permeation rate and filtration selectivity are higher than those of the alcohols. Methanol exhibits different behavior from the other alcohols. The permeation rate of pure methanol is higher than those of the other alcohols and its permeation rate increases with decreasing methanol concentration in water. The molecular sizes of the alcohols and our measurements with the XRD method of the spacings between GO flakes soaked in the solvents suggest that the small size of methanol makes it compete with water molecule to penetrate through the GO membrane unlike other alcohols with bigger molecular sizes. Further, the XRD results for the water-soaked GO membrane show that water passes through the GO membrane much faster than the other molecules because of its small size and high polarity. In order to demonstrate the GO membrane’s capability to separate water from complicated solutions, we attempted to use the GO membrane to remove the water from a liquid phase bio-oil containing more than 70% water. The GO membrane was found to filter out the water from the bio-oil and reduced the water content down to approximately 20%. This GO membrane is expected to be useful for removing water in diverse chemical purification processes, and in water purification processes.

Acknowledgments This study was supported by the Basic Science Research Program (2011-0014209, 2009-0083540) and the Global Frontier Research Center for Advanced Soft Electronics (20110031630) through a National Research Foundation of Korea Grant funded by the Korean government Ministry of Science, ICT and Future Planning.

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.carbon. 2014.06.007.

938

CARBON

7 7 ( 2 0 1 4 ) 9 3 3 –9 3 8

R E F E R E N C E S

[1] Leenaerts O, Partoens B, Peeters FM. Graphene: a perfect nanoballoon. Appl Phys Lett 2008:93. [2] Bunch JS, Verbridge SS, Alden JS, van der Zande AM, Parpia JM, Craighead HG, et al. Impermeable atomic membranes from graphene sheets. Nano Lett 2008;8:2458–62. [3] Geim AK, Novoselov KS. The rise of graphene. Nat Mater 2007;6:183–91. [4] Lee C, Wei XD, Kysar JW, Hone J. Measurement of the elastic properties and intrinsic strength of monolayer graphene. Science 2008;321:385–8. [5] Stankovich S, Dikin DA, Dommett GHB, Kohlhaas KM, Zimney EJ, Stach EA, et al. Graphene-based composite materials. Nature 2006;442:282–6. [6] Robinson JT, Zalalutdinov M, Baldwin JW, Snow ES, Wei ZQ, Sheehan P, et al. Wafer-scale reduced graphene oxide films for nanomechanical devices. Nano Lett 2008;8:3441–5. [7] Jiang DE, Cooper VR, Dai S. Porous graphene as the ultimate membrane for gas separation. Nano Lett 2009;9:4019–24. [8] Blankenburg S, Bieri M, Fasel R, Mullen K, Pignedoli CA, Passerone D. Porous graphene as an atmospheric nanofilter. Small 2010;6:2266–71. [9] Chen CM, Yang QH, Yang YG, Lv W, Wen YF, Hou PX, et al. Self-assembled free-standing graphite oxide membrane. Adv Mater 2009;21:3007–11 [3541-3541]. [10] Chen SS, Brown L, Levendorf M, Cai WW, Ju SY, Edgeworth J, et al. Oxidation resistance of graphene-coated Cu and Cu/Ni alloy. ACS Nano 2011;5:1321–7. [11] Dikin DA, Stankovich S, Zimney EJ, Piner RD, Dommett GHB, Evmenenko G, et al. Preparation and characterization of graphene oxide paper. Nature 2007;448:457–60. [12] Chen H, Muller MB, Gilmore KJ, Wallace G, Li D. Mechanically strong, electrically conductive, and biocompatible graphene paper. Adv Mater 2008;20:3557–61. [13] Zhu YW, Murali S, Cai WW, Li XS, Suk JW, Potts JR, et al. Graphene and graphene oxide: synthesis, properties, and applications. Adv Mater 2010;22:3906–24 [5226-5226]. [14] Dreyer DR, Park S, Bielawski CW, Ruoff RS. The chemistry of graphene oxide. Chem Soc Rev 2010;39:228–40. [15] Eda G, Chhowalla M. Chemically derived graphene oxide: towards large-area thin-film electronics and optoelectronics. Adv Mater 2010;22:2392–415. [16] Paul DR. Creating new types of carbon-based membranes. Science 2012;335:413–4. [17] Nair RR, Wu HA, Jayaram PN, Grigorieva IV, Geim AK. Unimpeded permeation of water through helium-leak-tight graphene-based membranes. Science 2012;335:442–4. [18] Kim HW, Yoon HW, Yoon SM, Yoo BM, Ahn BK, Cho YH, et al. Selective gas transport through few-layered graphene and graphene oxide membranes. Science 2013;342:91–5. [19] Li H, Song ZN, Zhang XJ, Huang Y, Li SG, Mao YT, et al. Ultrathin, molecular-sieving graphene oxide membranes for selective hydrogen separation. Science 2013;342:95–8. [20] Bowen TC, Noble RD, Falconer JL. Fundamentals and applications of pervaporation through zeolite membranes. J Membr Sci 2004;245:1–33.

[21] Lai ZP, Bonilla G, Diaz I, Nery JG, Sujaoti K, Amat MA, et al. Microstructural optimization of a zeolite membrane for organic vapor separation. Science 2003;300:456–60. [22] Ulbricht M. Advanced functional polymer membranes. Polymer 2006;47:2217–62. [23] Buchaly C, Kreis P, Gorak A. Hybrid separation processescombination of reactive distillation with membrane separation. Chem Eng Process 2007;46:790–9. [24] Hommerich U, Rautenbach R. Design and optimization of combined pervaporation/distillation processes for the production of MTBE. J Membr Sci 1998;146:53–64. [25] Tang YP, Paul DR, Chung TS. Free-standing graphene oxide thin films assembled by a pressurized ultrafiltration method for dehydration of ethanol. J Membr Sci 2014;458:199–208. [26] Talyzin AV, Hausmaninger T, You SJ, Szabo T. The structure of graphene oxide membranes in liquid water, ethanol and water-ethanol mixtures. Nanoscale 2014;6:272–81. [27] Lee Y, Eum PRB, Ryu C, Park YK, Jung JH, Hyun S. Characteristics of biochar produced from slow pyrolysis of Geodae-Uksae 1. Bioresource Technol 2013;130:345–50. [28] Hummers Jr WS, Offeman RE. Preparation of graphitic oxide. J Am Chem Soc 1958;80:1339. [29] Chen CN, Yu CW, Fu W. Synthesis of graphite oxide (GO)/Cu2O nanocomposite and its catalytic performance under the ultrasound. Nano 2013:8. [30] Allahbakhsh A, Sharif F, Mazinani S. The Influence of oxygen-containing functional groups on the surface behavior and roughness characteristics of graphene oxide. Nano 2013:8. [31] Xu YX, Bai H, Lu GW, Li C, Shi GQ. Flexible graphene films via the filtration of water-soluble noncovalent functionalized graphene sheets. J Am Chem Soc 2008;130:5856–7. [32] Li D, Muller MB, Gilje S, Kaner RB, Wallace GG. Processable aqueous dispersions of graphene nanosheets. Nat Nanotechnol 2008;3:101–5. [33] Marsden CM, Mann S. Solvents guide. 2nd ed. Cleaver-Hume; 1963. p. 633. [34] Linstrom PJ, Mallard WG. The NIST chemistry WebBook: a chemical data resource on the internet. J Chem Eng Data 2001;46:1059–63. [35] Wu HH, Gong QH, Olson DH, Li J. Commensurate adsorption of hydrocarbons and alcohols in microporous metal organic frameworks. Chem Rev 2012;112:836–68. [36] Guo F, Silverberg G, Bowers S, Kim SP, Datta D, Shenoy V, et al. Graphene-based environmental barriers. Environ Sci Technol 2012;46:7717–24. [37] McAllister MJ, Li JL, Adamson DH, Schniepp HC, Abdala AA, Liu J, et al. Single sheet functionalized graphene by oxidation and thermal expansion of graphite. Chem Mater 2007;19:4396–404. [38] Putz KW, Compton OC, Segar C, An Z, Nguyen ST, Brinson LC. Evolution of order during vacuum-assisted self-assembly of graphene oxide paper and associated polymer nanocomposites. ACS Nano 2011;5:6601–9. [39] Paredes JI, Villar-Rodil S, Martinez-Alonso A, Tascon JMD. Graphene oxide dispersions in organic solvents. Langmuir 2008;24:10560–4.