Effect of oxygen functionalisation on the electrochemical behaviour of multiwall carbon nanotubes for alcohol oxidation reactions

Abstract

Multiwall carbon nanotubes (MWCNTs) have been popularly used as catalyst supports for various electrochemical devices and reactions. In their preparation, surface oxidation by chemical oxidants is often necessary to purify MWNCTs, which also results in the formation of oxygen functional groups. However, the effect of these functionalities on electrochemical behavior of MWCNTs for alcohol oxidations remain largely unknown. In this study, we show that surface oxidation activates MWCNTs for electrochemical oxidation of alcohols in alkaline media (0.1 M KOH). Significantly enhanced catalytic activity in terms of higher current density (j) as well as lower alcohol oxidation onset potentials was observed following controlled oxidations via chemical or electrochemical methods. High-resolution XPS analysis suggests that the surface bound oxygen functionalities e.g. ketonic group (C=O) contribute primarily to the observed increase in catalytic performance. Moreover, to further increase the activity of MWCNTs, hydrothermal treatment was applied to repair the structural damage induced by the harsh oxidation treatment without the sacrifice of oxygen functional groups. Using the hydrothermally treated, surface-oxidized MWCNTs, EtOH undergoes oxidation into acetic acid with ∼99% faradaic efficiency. This study reveals the unique role of oxygen functional groups on MWCNTs towards catalytic alcohol oxidations for possible applications in direct alcohol fuel cells and alcohol sensors.

---

Introduction

Nanostructured carbon materials are being widely used as catalyst support materials in electrochemical energy storage and conversion systems such as direct alcohol fuel cells (DAFCs).¹ Currently noble metal catalysts such as Pt and Pd alloys are popularly used in DAFCs.^1,2 As these active metal catalysts are very costly, efficient utilisation of these materials is important, which often requires preparing nanoparticles of noble metals and immobilizing them onto high surface area carbon supports.³ The strategy is known to be effective and the catalytic activity of a catalyst increases with the increase of reactive surface area and surface-bulk ratio.⁴ Moreover, catalyst support materials strongly influence the size, shape and the dispersion of the nanoparticle catalysts, and play a critical role in ensuring efficient electron transport from the catalyst to the electrode.⁵ Hence, the development of catalyst support materials that have high conductivity, high surface area and high corrosion resistance, and the understanding of their electrochemical behaviors during the catalytic reactions are of great importance.⁶

Nanostructured carbon materials such as carbon black (CB), MWCNTs and graphene are commonly used as catalyst support materials. Among them, MWCNTs are known to be an excellent catalyst support due to its high electrical conductivity, high surface area as well as high corrosion resistance.^7–9 MWCNTs also has unique multilayer structures which can provides sufficient nucleation sites on the outer walls for attachment of metal NPs meanwhile enabling efficient charge transport through the undisturbed inner walls to the substrate electrodes.¹⁰ Nevertheless, to anchor metal nanoparticles (NPs) onto MWCNTs, surface oxidation is an indispensable step to introduce oxygen functional groups and defect sites which can act as nucleation sites for metal NPs.^11,12 Therefore, the presence of oxygen groups on MWCNTs is almost inevitable. Recent studies have shown oxygen functionalities on MWCNTs can improve the carbon material performance for several electrochemical reactions such as OER, HER and ORR.^13–17 However, the role of the oxygen containing groups on the MWCNTs towards the catalytic alcohol oxidation reactions has not been reported to date and remains largely unknown.

In this study we investigate the roles of oxygen groups on MWCNT for electrochemical alcohol oxidations. The intrinsic electrocatalytic activity of surface-oxidized MWCNT towards alcohol oxidation reaction (AOR) was studied in alkaline media by voltammetry. Recent studies have demonstrated that the incorporation of heteroatoms such as N, S, O and B in the graphitic network could alter the electron neutrality of carbon network, offering enhanced electrochemical reactivity.^18–21 Hence, the effects of addition and removal of oxygen functionalities on MWCNTs, were systematically studied through thermal reduction in inert atmosphere and controlled electrochemical oxidation, respectively.

Experimental

Materials

MWCNTs were received from Nanoport Tech., Shenzhen, China, where different treatments were used to produce oxidized MWCNTs with different oxidation degree and oxygen functionalities.

Purification of MWCNTs (pMWCNTs). In order to remove traces of metallic impurities, 0.5 g of MWCNTs were stirred in 100 mL of concentrated HCl for 2 h at 22 °C in 250 mL round-bottomed flask, the resulting mixture was diluted with water and then centrifuged to separate MWCNTs from the solution and washed with deionised water for 6 times to increase the pH of the solution to the neutral pH. Mixture was then dried overnight at 40 °C.

Electrochemically activated pMWCNT (eo-pMWCNT). Electrochemical activation of pMWCNT was performed by anodic polarization of pMWCNTs modified glassy carbon (GC) electrode at 1.76 V (vs. RHE) for 1000 s in 0.1 M KOH.

Mildly oxidized MWCNTs (moMWCNTs). Mild oxidation was performed by stirring 0.1 g of MWCNTs in piranha solution (70 mL of 98% H₂SO₄ and 30 mL of 30% H₂O₂) in 250 mL round bottomed flask for 5 h at 22 °C. The resulting mixture was diluted with deionised water, filtered, washed with water and ethanol and the subjected to overnight drying at 40 °C.

Highly oxidized MWCNTs (hoMWCNTs). 0.1 g of MWCNTs were refluxed at 70 °C for 48 h in 100 mL of 68% HNO₃. The resulting mixture was diluted with copious amount of deionised water and hoMWCNT were collected via vacuum filtration with Nylon filter (0.2 micron pores) which then washed with deionised water and dried under vacuum for 48 h at 22 °C.

Hydrothermal treated moMWCNTs (moMWCNTs-180). In typical preparation, 50 mg of moMWCNT was added into round bottomed flask filled with 35 mL of Milli-Q water and mixture was sonicated for 5 min to disperse the moMWCNT. The suspension was then transferred into Teflon lined 45 mL autoclave and then heated at 180 °C for 18 h.

Thermally annealed MWCNT (pMWCNT₉₀₀). 50 mg of pMWCNT was loaded into annealing alumina boat and treated using horizontal tube furnace were the temperature was set at 900 °C at ramping rate of 5 °C min⁻¹, annealing was performed under the protection of Ar-gas with a flow rate of 2 mL min⁻¹.

Electrochemical measurements

Inks were prepared by mixing 5 mg of MWCNTs samples in a mixture of 0.5 mL of deionised water, 0.5 mL of 95% ethanol and 20 μL of Nafion solution (5 wt%, Sigma-Aldrich). Mixture was homogenized by ultrasonication for 20 min using ElmaSonic. To prepare working electrodes, a known volume of the ink was spread to a polished glassy carbon (GC) electrode (polished with 0.05 μm alumina slurry on micropolishing pad, Buehler, USA) using micropipette tip. All electrochemical measurements were performed using CHI 760 Electrochemical Workstation (CH Instrument, Texas, USA) with standard three-electrodes set-up. An Ag/AgCl (3 M KCl) electrode (CH Instrument, Texas, USA) was used as reference electrode, and potential was calibrated against RHE according to the following equation:

E_RHE = E_Ag/AgCl + 0.197 + (0.059pH)

A typical analysis was performed in an Ar-saturated 0.1 M KOH (pH = 13) solution with known concentration of alcohols. Cyclic voltammetries were collected with a typical scan rate (ν) of 50 mV s⁻¹. The typical catalyst loading on GC electrode (geometric surface area = 0.07 cm²) is 0.2 mg cm⁻².

Physical characterization

Transmission electron microscopy (TEM). Imaging was performed with Phillips CM 200 operated at 200 kV. Samples were prepared by initially preparing the homogeneous ink of the respective MWCNT through ultra-sonication in methanol. The prepared ink was drop casted to holey Cu grid. The modified Cu grid was dried at 40 °C overnight.

Raman spectroscopy. Raman spectroscopy measurements were performed using a Renishaw Raman Microscope (514 nm), (Renishaw Plc., U.K.). Raman sample were prepared by first dispersing MWCNTs samples via 20 min ultrasonication in ethanol producing a black homogeneous ink which was then homogeneously drop-casted onto the surface of clean glass slide and dried in the oven at 100 °C for 3 min.

UV-visible spectroscopy. UV-visible spectroscopic measurements were performed using a Shimadzu UV-2401PC UV-Vis Spectrometer. The scans were performed between 200 nm and 500 nm. Prior to sample measurements, a background scan was taken using 0.1 M KOH. The carboxylic acid control was prepared by mixing glacial acetic acid solution (Ajax) with KOH solution and diluted to achieve a solution with 0.1 M acetic acid and 0.1 M KOH content.

X-ray photoelectron spectroscopy. X-ray photoelectron spectroscopy (XPS) spectra were obtained with a Kratos AXIS imaging XPS microprobe and Auger system.

Gas chromatography

Gas chromatography measurements were carried out using Shimadzu GC-2010 equipped with an AOC-20i autosampler and SGE Solgel-wax premium capillary column. Prior to GC measurements, the KOH in the solution were separated from the volatile compounds by distillation using a vacuum evaporator. The system was connected to a flame ionisation detector (FID). For this measurement the injection and FID temperatures were set at 210 °C and operated under linear velocity mode, other parameters were to the following values: total flow rate was set at 65.8 mL min⁻¹, linear velocity was set at 45.0 cm s⁻¹, column flow at 2.66 mL min⁻¹, purge flow at 3.0 mL min⁻¹ and pressure was set at 100 kPa with an injection volume of 1 μL.

Results and discussions

Raw MWCNTs (rMWCNTs) were first purified from surface bound metal impurities by treating with concentrated HCl solution for 2 h to produce purified MWCNT (denoted as pMWCNT). Following acid treatments, the presence of metal impurities were monitored using transmission electron microscopy (TEM). As shown in Fig. 1A and B, TEM investigations show that after acid washing the surface of MWCNTs is free of metal impurities, which is also confirmed by high resolution X-ray photoelectron spectroscopy (XPS) (Fig. S2†). TEM images also show, as expected, that metal impurities are detected (Fig. 1A and B) inside the MWCNTs. EDS analysis of the impurities (Fig. S1†) confirms the metal impurities are mainly Ni, consistent with the manufacturer report. Nevertheless, all TEM and XPS studies have shown that the Ni impurities are all deeply embedded inside the MWCNTs, and not exposed to the surface. Thus, the oxidized MWCNTs consist of only C and O on the surface, which enables us to study the role of O in catalyzing alcohol oxidations.

Fig. 1 Representative of TEM images for (A) pMWCNT and (B) higher magnification revealing the core-confined inherent metal impurities.

To understand the effect of oxygen functional group created by acid oxidations towards the electrochemical alcohol oxidation reaction (AOR) activity, pMWCNTs with different oxidation degree were prepared using acids such as piranha solution (7 : 3, H₂SO₄ : 30% H₂O₂) for mild oxidation (moMWCNT),^12,21 and hot 63% HNO₃ for high-degree oxidation (hoMWCNT).^12,21 Effect of acid strength on the oxidation degree of MWCNTs can be observed from their XPS spectra displayed in Fig. 2A. As determined by XPS, hoMWCNT has the highest oxygen content of 22.4 at%, while the oxygen content in moMWCNT and pMWCNT are 3.31 at% and 1.47 at%, respectively. As revealed by high resolution XPS (Fig. S3†), the acid treated MWCNTs are mainly comprised of ketonic (C=O) and epoxy (C–O) groups which give peaks at 531.3 eV and 533.3 eV, except for hoMWCNTs which contain mostly epoxy groups.

Fig. 2 (A) XPS of MWCNTs samples with different oxygen functionalisation treatments, indicating the difference in atomic percentage of oxygen content; (B) Raman spectra of all MWCNTs samples, the peaks were fitted with Gaussian–Lorentzian curves.

The effect of acid treatments towards graphitic networks was also closely monitored by Raman spectroscopy (Fig. 2B). The D-band observed at ∼1338 cm⁻¹ corresponds to the deformation and disorder of graphitic arrays such as defects, vacancies, kinks, heptagon–pentagon pairs and heteroatoms, the D′-band at 1608 cm⁻¹ is a double resonance feature induced by disorder, defect and intercalating molecules,^22,23 while the G-band at ∼1572 cm⁻¹ originates from in plane tangential stretching of C–C bonds.¹² The Table S2† shows the comparison of I_G/I_D of MWCNT samples. As shown in the Table S2,† the I_G/I_D values follows the sequence of moMWCNT > pMWCNT > hoMWCNT. The low I_G/I_D of hoMWCNT can be attributed to the extensive graphitic destruction induced by the harsh oxidative environment of 63% HNO₃.¹² The D′-band are more clearly pronounced in all oxidized MWCNT samples which can be ascribed to the strong interaction between oxidized region and the intercalated water molecules.^22,24 This was later confirmed by thermal annealing at 900 °C under inert Ar-atmosphere to deliberately remove oxygen functional groups and intercalated water molecules, resulting in a significant reduction of the D′-band peak (Fig. 2B).

The electrochemical performance of pMWCNT for ethanol (EtOH) oxidation (EOR) was evaluated by cyclic voltammetry. Fig. 3A demonstrates that distinct voltammetric response was detected in 0.1 M KOH electrolyte containing 1 M EtOH. On the anodic scan, the oxidation of EtOH occurs at 1.44 V and gives rise to a major peak with peak potential of 1.58 V. On the reverse cathodic sweep, another oxidation peak is observed at 1.57 V, which is typical for alcohol oxidation process and is attributed to the further oxidation of carbonaceous intermediates that are not fully oxidized during the anodic scan.²⁵ Significantly improved EOR activity was observed for the surface-oxidized MWCNTs relative to pMWCNTs. The onset potential for EOR using moMWCNT was detected at lower potential of 1.43 V, and the oxidation peak current density was about an order of magnitude higher than pMWCNTs.

Fig. 3 (A) The electrochemical activity of pMWCNT, hoMWCNT and moMWCNT (B) the effect of electrochemical activation of pMWCNT. Electrolyte used: 1 M EtOH in 0.1 M KOH, (catalyst loading: 0.2 mg cm⁻²). The background currents were obtained from pMWCNT in 0.1 M KOH.

To confirm the role the oxygen groups on the observed catalytic activity for ethanol oxidation, we deliberately remove oxygen functional groups by thermal annealing the pMWCNTs at 900 °C under inert Ar-atmosphere for 120 min. The XPS confirms that the thermally reduced pMWCNT contains much lower oxygen content of 0.74 at%. Following the thermal reduction process, significantly reduced electrocatalytic current was observed (Fig. S4†). Collectively, the above results suggest the observed catalytic activity is strongly related to the presence of oxygen functional groups on the surface of MWCNTs. On the other hand, hoMWCNT which have significantly higher O content exhibits lower EOR performance relative to moMWCNT. Unlike piranha, oxidation by 63% HNO₃ are known to be very effective for oxygen functionalisation but also induces extensive etching that create significant amount of defects as reflected from the Raman analysis.²¹ Fig. 3A shows that the voltammetric response of hoMWCNT was similar to moMWCNT. However, both of the onset and peak potentials (E_p) of EtOH oxidation were shifted towards higher values by 80 mV and 30 mV, respectively. The higher potential shift was accompanied by decrease of peak current density (j_p) for both 1^st and 2^nd oxidation peaks, which are 1.14 mA cm⁻² and 1.10 mA cm⁻². The shift of E_p, onset potential and the decreased j_p suggest degraded catalytic activity, possibly due to the reduced conductivity of the hoMWCNT as a result of extensive damage induced by harsh oxidation by nitric acid. Moreover, as revealed by XPS in Fig. S3,† unlike the oxidation by piranha solution, the oxidation by nitric acid produces mainly epoxy functional groups, while peaks from ketonic group was not observed.

Another way of oxygen functionalisation was achieved by in situ electrochemical oxidation process which usually takes place through two stages: (i) intercalation of the active ions within MWCNTs induced by electric current; and (ii) electro-oxidation of MWCNTs at high voltages.²⁶ Electrochemical oxidation of pMWCNT was performed in 0.1 M KOH at anodic potential of 1.76 V for 1000 s to produce electrochemically activated pMWCNT (eo-pMWCNT). The electrocatalytic activity of eo-pMWCNT for oxidation of 1 M EtOH in 0.1 M KOH was dramatically enhanced as indicated by increase of j_p by 365% as well as negative shift of the EOR onset potential by 10 mV relative to pMWCNT. In addition to oxygen functionalisation, eo-pMWCNT also has enhanced electrochemical surface area (ECSA). According to Fig. S6B† the double layer capacitance (C_DL) was enhanced to 1.65 F g⁻¹ from an initial value of 1.44 F g⁻¹, the value is directly proportional to the ECSA value.²¹

Furthermore, the Raman spectra shows that electrochemical activation does not induces extensive structural damage as reflected by the preserved I_G/I_D ratio of 1.07 (Fig. 2). The presence of pronounced D′-band also suggests improved interaction of water molecules with eo-pMWCNT. From the XPS in Fig. 1, the electrochemical oxidation treatment leads to enhancement of oxygen content in pMWCNT by an order of magnitude from 1.47 at% to 11.91 at%. This observation correlates well with the significantly enhanced catalytic activity and negative shift of the EOR onset potential of eo-pMWCNT. Furthermore, the epoxy group at 533.3 eV are confirmed to increase following electrochemical oxidation from the initial of 0.83 at% to 4.57 at%. The role of oxygen functionalities thus can be validated by comparing the ratio of intensity (I) of I_(C=O)/I_(C–O) from MWCNT samples with and without electrochemical activation. Table S1† shows that samples having the highest I_(C=O)/I_(C–O) ratio (e.g. moMWCNT with I_(C=O)/I_(C–O) = 0.87), also exhibits highest catalytic performance. The above results suggest oxygen groups, particularly ketonic group (–C=O) are catalytically active for EOR, consistent to the lower activity of hoMWCNT which the surface bound oxygen functional groups were mainly epoxy functionalities. This is caused by the strongly electron withdrawing nature of C=O functionalities which is very likely to induce change in electronic structure of its neighbouring carbons atom causing significant enhancement of electrocatalytic activity.^21,27

One concern related to oxygen functionalisation of MWCNT and other carbon materials is that the potential structural damage created during the oxidative treatments which leads to deterioration of electrochemical properties. Hydrothermal treatment is known to be useful for inducing graphitic re-arrangement useful to repair damages from oxidative treatments without significantly affecting the surface oxygen functionalities.²⁸ By rearranging the disorder created by defect sites, the overall electro-conductivity of the MWCNTs can be improved, thus allowing faster shuttling of electrons transport.²⁹ This hypothesis is consistent with the Raman spectroscopy of the hydrothermally treated moMWCNT (moMWCNT-180) which exhibits the highest I_G/I_D = 1.22. On the other hand, slight reduction of oxygen content was observed, due to removal of less stable oxygen functionalities such as epoxy. Cyclic voltammetry in Fig. 4A reveals that the hydrothermal treatment improves the electrocatalytic activity of moMWCNT. As displayed in Fig. 4A, a significantly higher EtOH oxidation peak of ∼4.00 mA cm⁻² can now be achieved at 1.58 V in contrast to the performance of moMWCNT shown in Fig. 3A.

Fig. 4 (A) CVs for the oxidation of MeOH, EtOH and isopropyl alcohol by electrochemically oxidized moMWCNT-180; (B) chronoamperograms at 1.62 V with EtOH spiking in 0.1 M KOH; (C) Tafel slopes of MWCNTs with different oxidation treatments (electrolyte: 0.1 M KOH containing 1 M EtOH).

The products from EOR catalyzed by moMWCNT-180 were quantified by UV-visible spectroscopy and gas chromatography (GC) following the 3 h bulk-electrolysis at 1.62 V. Fig. S8† shows stable current responses over 3 hours continuous EOR bulk electrolysis. As no gaseous bubble formation is observed on the electrode surface, it is deduced that EOR undergoes a 4 electrons oxidation process from ethanol to form ethanoic acid (CH₃COOH) rather than gaseous products such as CO and CO₂. To prove the hypothesis, we carried out UV-visible spectroscopy and gas chromatography (GC) to characterise the product. In these measurements acetic acid standard solutions were used as the internal standard. UV-visible spectra obtained for the electrolyte (0.1 M KOH with 1 M EtOH) (Fig. S7†) exhibits a significantly enhanced and broadened absorbance with a shoulder absorbance around 216–218 nm. By comparing to the spectra obtained from acetic acid standard solution (1 M acetic acid in 0.1 M KOH), the UV-vis spectra suggests that acetic acid is formed in the electrolyte after electrolysis. The EtOH oxidation product is further quantified by GC.^30,31 In chromatogram, the peak for acetic acid was observed at a retention time of 2.428 min. Using acetic acid standard solutions, a calibration curve of acetic acid concentration vs. area of the peak is established (Fig. S10†). According to the calibration curve, 44.40 ± 6.43 μmol of acetic acid was detected during electrolysis which corresponds to a faradaic efficiency close to unity (∼99%), suggesting that acetic acid was the primary product of the EOR at the moMWCNT.

Finally, it is demonstrated that moMWCNTs-180 are also capable of catalyzing the oxidation of other alcohols such as methanol and isopropyl alcohol (refer to Table S3,† for detailed electrochemical performance comparison). According to the catalytic performance shown by Table S3† and Fig. 4A, moMWCNT-180 exhibits catalytic activities to the alcohols in the order of MeOH > EtOH > isopropyl alcohol, similar to Pt catalyst with greater acidity resulting in higher current density of oxidation peaks.²⁵ In Fig. 4B, the bulk EOR electrolysis at 1.62 V were performed with the treated MWCNTs for comparison purpose. The highest current density was observed with moMWCNT-180, signifying its superior catalytic activity. Fig. 4C shows the comparison of Tafel slopes between moMWCNT-180, moMWCNT, hoMWCNT and pMWCNT which are 51 mV dec⁻¹, 46 mV dec⁻¹, 114 mV dec⁻¹ and 177 mV dec⁻¹ respectively. The significantly lowered Tafel slope value of moMWNT-180 indicates the significantly improved EOR kinetics following the chemical oxidation and hydrothermal treatment.

Conclusions

In conclusion, it was found that following the oxygen functionalization with acids, MWCNTs exhibit enhanced electrocatalytic activity for alcohol oxidations. The mildly oxidized MWCNT by piranha solution treatment exhibits higher alcohol oxidations activity compared to rMWCNT and hoMWCNTs treated by HNO₃, which leads to destruction to MWCNT structures. Electrochemical oxidation treatment of MWCNTs after piranha solution treatment is found to be particularly effective in introducing ketonic groups (–C=O) onto MWCNTs, which are believed to be most active for catalysing the AOR. Although the observed catalytic activity at oxidized MWCNTs for AOR is modest, compared to noble metals-based catalysts such as Pt and Pd, it is revealed that, for the first time, the surface oxidised MWCNTs is not necessarily inert for AOR. This is particularly important for understanding the MWCNTs as catalyst support for metal-based nanoparticles for AOR in alcohol fuel cells and alcohol sensors.

Acknowledgements

The authors thank UNSW Mark Wainwright Analytical Centre for the access of all characterization facilities. The study was financed by an Australian Research Council Discovery Grant (DP160103107).

Notes and references

1. H. Huang and X. Wang, J. Mater. Chem. A, 2014, 2, 6266–6291.
2. C. Bianchini and P. K. Shen, Chem. Rev., 2009, 109, 4183–4206.
3. R. Balgis, G. M. Anilkumar, S. Sago, T. Ogi and K. Okuyama, J. Power Sources, 2012, 203, 26–33.
4. E. Antolini, Appl. Catal., B, 2009, 88, 1–24.
5. J. H. Kim and J.-S. Yu, Phys. Chem. Chem. Phys., 2010, 12, 15301–15308.
6. L. Dai, D. W. Chang, J.-B. Baek and W. Lu, Small, 2012, 8, 1130–1166.
7. W. Li, C. Liang, W. Zhou, J. Qiu, Z. Zhou, G. Sun and Q. Xin, J. Phys. Chem. B, 2003, 107, 6292–6299.
8. X. Lu and C. Zhao, J. Mater. Chem. A, 2013, 1, 12053–12059.
9. Y. Li, W. Zhou, H. Wang, L. Xie, Y. Liang, F. Wei, J.-C. Idrobo, S. J. Pennycook and H. Dai, Nat. Nanotechnol., 2012, 7, 394–400.
10. X. Lu, H. M. Chan, C.-L. Sun, C.-M. Tseng and C. Zhao, J. Mater. Chem. A, 2015, 3, 13371–13376.
11. S. W. Lee, B.-S. Kim, S. Chen, Y. Shao-Horn and P. T. Hammond, J. Am. Chem. Soc., 2009, 131, 671–679.
12. V. Datsyuk, M. Kalyva, K. Papagelis, J. Parthenios, D. Tasis, A. Siokou, I. Kallitsis and C. Galiotis, Carbon, 2008, 46, 833–840.
13. X. Wang, C. Ouyang, S. Dou, D. Liu and S. Wang, RSC Adv., 2015, 5, 41901–41904.
14. N. Cheng, Q. Liu, J. Tian, Y. Xue, A. M. Asiri, H. Jiang, Y. He and X. Sun, Chem. Commun., 2015, 51, 1616–1619.
15. A. L. Mohana Reddy, N. Rajalakshmi and S. Ramaprabhu, Carbon, 2008, 46, 2–11.
16. S. H. Ryan, H. Drew and C. Zhongwei, Nanotechnology, 2010, 21, 165705.
17. J. M. Sieben and M. M. E. Duarte, Int. J. Hydrogen Energy, 2011, 36, 3313–3321.
18. S. Chen, J. Duan, M. Jaroniec and S.-Z. Qiao, Adv. Mater., 2014, 26, 2925–2930.
19. T. Y. Ma, S. Dai, M. Jaroniec and S. Z. Qiao, Angew. Chem., Int. Ed., 2014, 53, 7281–7285.
20. Y. Zhao, R. Nakamura, K. Kamiya, S. Nakanishi and K. Hashimoto, Nat. Commun., 2013, 4, 2390.
21. X. Lu, W.-L. Yim, B. H. R. Suryanto and C. Zhao, J. Am. Chem. Soc., 2015, 137, 2901–2907.
22. S. Hussain, R. Amade, E. Jover and E. Bertran, Nanotechnology, 2012, 23, 385604.
23. S. Osswald, M. Havel and Y. Gogotsi, J. Raman Spectrosc., 2007, 38, 728–736.
24. R. R. Nair, H. A. Wu, P. N. Jayaram, I. V. Grigorieva and A. K. Geim, Science, 2012, 335, 442–444.
25. T. S. Ilana, P. J. Alex, J. Mira and J. Jiří, Catal. Lett., 2013, 143, 777–782.
26. G. Moraitis, Z. Špitalský, F. Ravani, A. Siokou and C. Galiotis, Carbon, 2011, 49, 2702–2708.
27. B. H. R. Suryanto and C. Zhao, Chem. Commun., 2016, 52, 6439–6442.
28. P. V. Kumar, N. M. Bardhan, S. Tongay, J. Wu, A. M. Belcher and J. C. Grossman, Nat. Chem., 2014, 6, 151–158.
29. X. Lu, H. M. Chan, C. L. Sun, C. M. Tseng and C. Zhao, J. Mater. Chem. A, 2015, 3, 13371–13376.
30. E. M. Belgsir, H. Huser, J. M. Leger and C. Lamy, J. Electroanal. Chem. Interfacial Electrochem., 1987, 225, 281–286.
31. C. Lamy, A. Lima, V. LeRhun, F. Delime, C. Coutanceau and J.-M. Léger, J. Power Sources, 2002, 105, 283–296.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra17402g

---