Selective oxidation of alcohols by molecular oxygen is an attractive and environmentally friendly reaction route in the synthesis of fine chemicals and intermediates [1-3]. Recently, gold nanoparticles supported on carbonaceous materials have received considerable attention since they exhibit high selectivity and stability for the oxidation of primary alcohol [2-8]. The catalytic performance of Au/C catalysts depends largely on the size and the shape of the Au nanoparticles and the nature of the supports [9-11]. The surface chemistry of different carbon materials is a key factor to their catalytic performance [5, 12-14]. It has been proposed that the amount and the type of surface O-containing functional group have a strong effect on the catalytic activity of Au/C catalysts in the selective alcohol oxidation [13-15]. Furthermore, the crystalline nature of the carbon supports was also reported to play an important role in determining the Au-C interaction and high crystallinity of carbon supports corresponds to high activity and selectivity of Au/C catalysts in the selective oxidation of glycerol .
Compared to other carbon materials such as activated carbon and carbon nanotubes, graphite oxide (GO) contains abundant oxygen functional groups such as carboxyl, carbonxyl and hydroxyl groups on the surface, which makes it a good candidate as support for Au nanoparticles in the application of selective oxidation of alcohols [17-21]. In our previous study, three different carbon materials, including reduced graphene oxide (rGO), graphite (GC), and active carbon (AC), have been investigated as the supports for Au nanoparticles and the catalytic performance in the aerobic oxidation of benzyl alcohol (BA) has been compared . It was found that Au/rGO catalyst shows a much higher activity than Au/AC and Au/GC in the liquid phase aerobic oxidation of benzyl alcohol. The superior catalytic activity of Au/rGO was suggested to be related to the presence of surface O-containing functional groups and moderate graphitic character of rGO supports . However, it is difficult to compare directly the support effect by using different carbon materials since they differ largely in the strcucture including volume size distribution, surface area, surface functional groups, and electronic structure etc. To obtain a better understanding of the support effcet, especially the effect of surface functional group and graphitic structure, in the present study, GO supports were firstly modified by thermal and chemical treatments, followed by the immobilization of Au naoparticles with similar average size. Then the prepared Au/GO catalysts were characterized and examined in the aerobic oxidation of benzyl alcohol. Finally, the support effect was discussed on the basis of structure analysis and catalytic activities results.II. EXPERIMENTS
All the chemical agents used in the present study were purchased from Sinopharm Chemical Reagent Co. Ltd. if not mentioned specially. GO was prepared from expandable graphite (Qingdao Jin Ri Lai Graphite Co. Ltd.) by pressurized oxidation [22, 23]. Typically, the graphite (0.6 g), potassium permanganate (KMnO4) (3 g) and sodium nitrate (NaNO3) (0.6 g) were put into a Teflon reactor, and then sulphuric acid (98%, 30 mL) was added. As soon as the sulphuric acid was added, the reactor and stainless steel autoclave were covered and fastened down. The autoclave was kept at 0 ℃ for 2 h and then heated at 100 ℃ in an oven for 2 h. The obtained mud was diluted with 300 mL water. With vigorous stirring, H2O2 (30%) was dipped into the suspension until the slurry turned golden yellow. The suspension was washed with HCl (5%) and deionized water until the pH reached 7, and then the precipitate was dispersed in water (150 mL) by sonication for 2 h. Finally, the resulted mixture was dried at 60 ℃ for 24 h. The as-obtained GO was calcinated at 200, 400, 700 ℃ for 1 h in N2 and is marked as GO200, GO400, and GO700, respectively. GO200 was treated with 0.1 mol/L NaOH at room temperature for 1 h (denoted as GO200-NaOH), or with 0.1 mol/L HNO3 at boiling temperature for 1 h (denoted as GO200-HNO3). Also, GO200 was reduced by NaBH4 at 100 ℃ for 4 h to get rGO [24, 25].
The supported Au/GO catalysts were prepared using a sol-immobilization method . First, 52 mL of HAuCl4 aqueous solution (about 3×10-4 mol/L) was mixed with 0.32 mL of PVA solution (0.56wt%) with stirring. Next, 0.75 mL of 0.1 mol/L NaBH4 solution was added dropwise under vigorous stirring condition. After 30 min, the solution was acidized by diluted HCl solution to pH≈2. Then GO was added to the obtained solution, the amount of which was calculated to have a final Au loading of 1wt%. After stirring for 2 h, the slurry was filtered and washed with deionized water and alcohol until no Cl- ion was detected. The solid was then dried in vacuum at room temperature for 12 h. All the synthesized Au/GO catalysts were calcinated at 200 ℃ for 3 h in flowing air to remove PVA protecting agents before the activity measurements.
GO samples and the prepared Au/GO catalysts were characterized by various analytical techniques. Powder X-ray diffraction patterns (XRD) were recorded on a MXPAHF X'Pert PRO diffractometer using nickle-filtered Cu Kα radiation source (λ=0.15418 nm). Transmission electron microscopy (TEM) images were obtained on JEM-2100F with electron acceleration voltage of 200 kV. X-ray photoelectron spectroscopy (XPS) measurements were performed on an ESCA-LAB 250 high-performance electron spectrometer using monochromatized Al Kα as the excitation source. All the binding energies were calibrated by the C1s peak of adventitious carbon (BE=284.8 eV). Raman spectroscopy was obtained using a LABRAM-HR Confocal Laser Raman Spectroscopy with excitation by a 514.5 nm argon laser line in the back-scattering configuration. In temperature programmed desorption-mass spectroscopy (TPD-MS) measurements, 50 mg of Au/GO catalysts were put in U-shape quartz tube and heated to 800 ℃ in Ar carrier gas at a heating rate of 10 ℃/min. Both CO and CO2 mass signals were monitored by QMS analyzer (QIC20, Hiden).
The catalytic reaction was carried out in a 50 mL stainless steal autoclave. Typically, 0.5 mL BA, 20 mL deionized water, and 75 mg Au/GO catalysts were mixed with 5 mL Na2CO3-NaHCO3 buffer solution (pH=9). Before reaction, the autoclave was purged five times with O2 and then the pressure was kept at 3 atm during the reaction. The reaction temperature is 100 ℃ with stirring rate of 1200 r/min. After reaction for 8 h, the reaction mixture was cooled down to room temperature. The reaction products were extracted with ethyl acetate and then analyzed by a FID-equipped GC.III. RESULTS AND DISCUSSION A. Characterization of catalysts
Figure 1 displays the TEM images of Au/GO200, Au/GO400, Au/GO700, Au/GO200-NaOH, Au/GO200-HNO3, Au/rGO catalysts, and the corresponding size distributions of Au particles. The TEM images show that Au nanoparticles disperse randomly on different GO supports and all the supported Au nanoparticles show a similar size distribution with an average size of ~4.0 nm. It suggests that GO supports with various surface treatments have no significant influences on the size distribution of the supported Au nanoparticles during Au immobilization and subsequent calcination.
Figure 2(a) shows XRD patterns of all the prepared Au/GO catalysts. All patterns exhibit a broad diffraction peak at 24.8°, a characteristic structure of reduced graphene oxide, which is assigned to the presence of `graphitic' carbon domains [17, 27-29]. It indicates that the GO layers restack to roughly the initial graphite spacing with the removal of the functional groups when the GO supports are modified by thermal and chemical treatments [17, 27-29]. On the other hand, the similar XRD patterns also suggest that GO structures are almost kept for all the Au/GO catalysts after various treatments. For all the Au/GO samples, the diffraction peaks at 2θ of 38.2° and 44.5° correspond to the Au (111) and (200) crystal face, respectively, clearly suggesting the presence of metallic Au nanoparticles. The chemical state of supported Au colloids was also identified by XPS measurements, and the results are shown in Fig. 2(b). The binding energies of Au4f7/2 were observed at 84.1 eV for all the prepared Au/GO catalysts, confirming the supported Au nanoparticles in the metallic form.
A varity of previous investigations have provn that TPD is an efficient method to analyze surface properties of carbon supports [6, 13, 30, 31]. By monitoring the CO and CO2 thermal desorption signals, it is possible to determine the type and amount of O-containing functional groups present on carbon materials. For example, it was reported that CO2 desorption alone at low temperatures (150-450 ℃) is assigend to the decomposition of carboxylic acids while that at high temperatures (600-800 ℃) to lactones, carboxylic anhydrides originate from CO and CO2 in the temperature range of 400-650 ℃, dissociation of the groups such as phenols (600-800 ℃), ethers (700 ℃), and carbonyls and quinones (750-1000 ℃) release CO exclusively in TPD measurements [6, 13, 30, 31]. Moreover, the peak intensity of released CO and CO2 desorption is closely related to the amount of surface oxygen species present on the surface. The TPD results of various modified GO supports are shown in Fig. 3. For GO200, the predominant CO2 desortpion below 400 ℃ is ascribed to the decomposition of carboxylic acid while the simultaneous CO and CO2 desortpion between 450 and 750 ℃ is related to the dissociation of carboxylic anhydrides. Meanwhile, CO desorption alone above 750 ℃ also suggests the contribution from phenols, ethers, carbonyls and quinones on the surface. For GO400, the desortpion signal below 400 ℃ is negligible while CO and CO2 desorption at higher temperature is very close to that of GO200. It suggests that carboxylic acid species is removed significatntly from the surface while other surface functional groups are almost kept unchanged for GO400 compared to GO200. In the case of GO700, almost no desorption siganls are observed in TPD sepctra except CO desorption above 750 ℃. It means that the surface species are mostly phenols, carbonyls and quinones after heating GO supports at 700 ℃. After GO200 was treated by NaOH, the CO2 desorption is geneally suppressed, which suggests that the amount of carboxylic acid and carboxylic anhydrides is reduced by NaOH treatment . In the case of GO200-HNO3, the simultaneous desorption of CO and CO2 are enhanced in the temperature range of 450-600 ℃, suggesting more carboxylic anhydrides present on the surface. Moreover, the increased intensity of CO desorption peak between 600 and 700 ℃ also implies a more contribution from phenols species. When GO sample was reduced by NaBH4 (rGO), TPD spectra show that CO desorption between 400 and 800 ℃ is clearly enhanced while CO2 desorption in the same temperature range decreases in intensity. Therefore, it indicates that the coverages of surface groups such as phenols/ethers are relatively increased at the expense of those of surface groups such as carbxylic acids, lactones and carboxylic anhydrides, in consistence with the surface reduction process .
The type and amount of O-containing surface functional groups on Au/GO samples after various treatments were also examined by XPS measurements and the resultant O1s spectra are shown in Fig. 4. According to the assignment established in previous literatures [12, 31], the broad O1s peak was deconvoluted into four peaks: C=O in carbonyls and quinones (BE≈531.1 eV), -CO-OR in esters and anhydrides (BE≈532.3 eV), -CO-OR in esters, anhydrides, ethers and oxygen atoms in phenols (BE≈533.3 eV) and -CO-OH in carboxylic groups (BE≈534.2 eV). After fitting results of O1s spectra, the relative concentrations of oxygen functional groups are listed in Table I. Similar to the TPD results, XPS spectra also suggest that the amount of oxygen functional groups decreases with an increase in the heating temperature. The total oxygen concentration is estimated to be 18.86%, 17.07%, and 10.35% for GO200, GO400, and GO700, respectively. In comparsion with GO200, GO400 exhibits a much less amount of caboxylic acid species while other surface groups remian almost unchanged, in agreement with the TPD results. For GO700, the surface oxygen functional groups are mostly carbonyls and quinones (O1s at 531.1 eV), lactones (O1s at 532.3 eV), and ethers and phenols (O1s at 533.3 eV). When the surface was treated by NaOH, the surface concentration of carboxylic acid decreases accordingly and the contribution from O1s feature at 532.3 eV is relativly enhanced, which is assigned to lactones in combination with TPD results. In the case of GO200-HNO3, the amount of oxygen groups increases slightly; especially the coverages of carboxylic anhydrides (O 1s at 532.3 eV) and phenols/ethers (O1s at 532.3 eV) increase relatively, which agrees well with TPD results. Furtermore, reduction by NaBH4 induces an apparent decreases in the abundance of carbonyls (O1s at 531.1 eV) and carboxylic acids (O1s at 534.2 eV) as well as an increase in the surface concentration of ethers (O1s at 532.3 eV), in consistence with TPD results.
Raman spectroscopy has been often employed to obtain the structural information of GO-derived materials. Figure 5 shows the Raman spectra of various Au/GO supported catalysts. For all the Au/GO samples, two bands are detected at ~1350 and ~1580 cm-1, commonly denoted as D and G-band . The former one is related to the disorder carbon structure induced by lattice defects and the latter is associated to well-ordered structure [16, 28]. And the intensity ratio of D-band to G-band (ID/IG) is used to mesure the quaility of the graphitic character [16, 28]. Usually the ID/IG ratio increases with the extent of the disorder for GO materials, decreasing to zero for completely defect-free graphite . However, for thermally treated Au/GO samples, the ID/IG ratio is estimated to be 1.07 for Au/GO200, 1.09 for Au/GO400, and 1.12 for Au/GO700, respectively. In the case of Au/rGO, the ID/IG ratio remains almost constant with a value of 1.08 compared to that of GO200. The above results are very suprising since thermal/chemical reduction of GO usually leads to a decreased ID/IG ratio with removal of oxygen functional groups and restoration of the aromaticity of the graphene lattice [24, 35]. Similar observations were also reported by Paredes and Stankovich et al. [35, 36]. To account for the unexpected changes of the ID/IG ratio for thermally/chemically reduced GO materials, it was proposed that GO had a significant distoration of the aromatic rings with a certain amorphous character and during reduction the distoration was removed and the carbon lattice returned to a graphitic but highly defected state [35, 36]. In addition, the ID/IG ratio increases to 1.26 and 1.10, respectively, when GO200 supports were treated by HNO3 and NaOH. The increased ID/IG ratio for GO200-HNO3 is due to an introduction of structural disorder in the carbon lattice by the oxidation of GO materials, as indicated by the TPD and XPS results .B. Catalytic test
The activities of various 1.0wt%Au/GO catalysts including Au/GO200, Au/GO400, Au/GO700, Au/GO200-NaOH, Au/GO200-HNO3 to Au/rGO in the selective oxidation of BA were carried out, the conversions of BA were 73%, 69%, 64%, 65%, 56%, and 90%. Blank experiments show that all the studied GO supports without Au loading exhibit a negligible BA conversion (＜2%). And all Au/GO catalysts show high benzaldehyde selectivity (≥98%) (not shown). Among the studied Au/GO catalysts, Au/RGO has the highest conversion of BA (90%). For thermally treated Au/GO catalysts, the activity decreases with an increase in the heating temperature of GO supports. When GO was pretreated by NaOH, the BA conversion decreases slightly to 65%. In comparision, the pretreatment of HNO3 suppresses the actvity of Au/GO200 significantly with BA conversion decreasing from 73% to 56%.
It has been shown that the catalytic performance of Au nanoparticles supported on carbon-based materials in the aerobic oxidation of alcohols is closely related to several factors including the chemical state of Au, the size of Au nanoparticles, the support structure, and the presence of surface functional groups on the supports [11, 13-16]. In the present work, Au nanoparticles were synthesized by a sol-gel method and then were immobilized on modified GO supports with similar metal loading (1%). TEM and XPS results indicate that both the size and the electonic structure of Au nanoparticles are very similar. Therefore, the effect of the properties of Au nanoparticles on the catalytic activity can be neglected in the present work. The observed different activities of Au/GO catalysts are mostly dependent on the nature of the modified GO supports such as the crystal structure and the surface oxygen functional groups. Firstly, the BA conversion decreases slightly from 73% for Au/GO200 to 69% for Au/GO400. Compared to Au/GO200, Au/GO400 catalyst has a similar structural disorder (as indicated by the ID/IG ratios) but exhibits a much lower concentration of carboxylic acid while the relative concentrations of other oxygen functional groups are very similar based on TPD and XPS results. It clearly indicates that the amount of surface carboxylic acid is not a critical factor, similar to previous investigations [13, 14]. Secondly, the activity and Raman results show that Au/GO700 has a similar BA coversion and similar ID/IG ratio to Au/GO200-NaOH. Meanwhile, XPS results indicate that the amount of oxygen functional group on Au/GO700 is only about 56% of that on Au/GO200-NaOH. It implies that the structural disorder of GO supports, instead of the amount of O-contaning groups, is the determining factor to the activity. Such observation is further confirmed by comparing the catalytic activity tendency with that of the ID/IG ratios of various Au/GO catalysts. The BA conversion increases with decreasing the ID/IG ratio except Au/rGO. For example, Au/GO-HNO3 has a higher ID/IG ratio of 1.26 (therefore more disorder structure) but shows a lower BA conversion (56%). The critical factor of the order structure of carbon-based supports was also reported in Ref.. For Au/C catalyst in glycerol oxidation, it is proposed that carbon materials with the higher crystallinity can result in a greater number of graphite edges exposed in an orderly manner, ensuring a strong anchoring of small metal particles . Therefore, it promotes Au-C interaction and faciliate the proton abstraction from glycerol . The superior catalytic performance of Au/rGO is still not clear. It has a similar ID/IG ratio to Au/GO200 but shows a much higher BA conversion (~90%) (Fig. 5). One possible explanation is that the electronic conductivity of GO supports is increased during the reduction process . A recent study suggests that the delocalized $\pi$-elecron in ordered carbon support can lead to an enhancement of electronic mobility and therefore promote both the adsorption and regeneration of hydroxide ions, which are the key intermediates in alcohol oxidation on Au/C catalysts [13, 38].IV. CONCLUSION
In the present work, various Au/GO catalysts were prepared by immobilizing gold nanoparticles with similar average sizes on various modified GO supports. The thermal and chemical treatments have a large influence on the presence of surface oxygenated groups and the crystalline structure of GO supports. A strong support effect was observed on the catalytic activity of Au/GO catalysts for aerobac oxidation of benzyl alcohol. Compared to the amount and the type of surface oxygen functional groups, it is proposed that the ordered structure of GO supports plays a more important role in determining the catalytic performance of the prepared Au/GO catalysts.V. ACKNOWLEDGMENTS
This work was supported by the Natural Science Foundation of Anhui Province (No.11040606M39) and the Fundamental Research Funds for the Central Universities.
|||A. Sheldon R, W. C. E. Arends I, J. T. Brink G, and Dijksman A, Acc. Chem. Res. 35 , 774 (2002). DOI:10.1021/ar010075n|
|||Mallat and A. Baiker T, Chem. Rev. 104 , 3037 (2004). DOI:10.1021/cr0200116|
|||D. Pina C, Falletta E, and Rossi M, Chem. Soc. Rev. 41 , 350 (2012). DOI:10.1039/C1CS15089H|
|||Prati and M. Rossi L, J. Catal. 176 , 552 (1998). DOI:10.1006/jcat.1998.2078|
|||Rodríguez-reinoso F, Carbon 36 , 159 (1998). DOI:10.1016/S0008-6223(97)00173-5|
|||L. Figueiredo J, F. R. Pereira M, M. A. Freitas M, and J. M. Órfão J, Carbon 37 , 1379 (1999). DOI:10.1016/S0008-6223(98)00333-9|
|||Besson and P. Gallezot M, Catal. Today 57 , 127 (2000). DOI:10.1016/S0920-5861(99)00315-6|
|||D. Pina C, Falletta E, Prati L, and Rossi M, Chem. Soc. Rev. 37 , 2077 (2008). DOI:10.1039/b707319b|
|||Bianchi C, Porta F, Prati L, and Rossi M, Top. Catal. 13 , 231 (2000). DOI:10.1023/A:1009065812889|
|||Meenakshisundaram S, Nowicka E, J. Miedziak P, L. Brett G, L. Jenkins R, Dimitratos N, H. Taylor S, W. Knight D, Bethell D, and J. Hutchings G, Faraday Discuss. 145 , 341 (2010). DOI:10.1039/B908172K|
|||Fang W, Chen J, Zhang Q, Deng W, and Wang Y, Chem. Eur J17 , 1247 (2011).|
|||L. Figueiredo and M. F. R. Pereira J, Catal. Today 150 , 2 (2010). DOI:10.1016/j.cattod.2009.04.010|
|||G. Rodrigues E, F. R. Pereira M, Chen X, J. Delgado J, and J. M. Órfão J, J. Catal. 281 , 119 (2011). DOI:10.1016/j.jcat.2011.04.008|
|||G. Rodrigues E, J. Delgado J, Chen X, F. R. Pereira M, and J. M. Órfão J, Ind. Eng. Chem. Res. 51 , 15884 (2012). DOI:10.1021/ie302159m|
|||Zhu J, A. C. Carabineiro S, Shan D, L. Faria J, Zhu Y, and L. Figueiredo J, J. Catal. 274 , 207 (2010). DOI:10.1016/j.jcat.2010.06.018|
|||Gil S, Muñoz L, Sánchez-Silva L, Romero A, and L. Valverde J, Chem. Eng. J. 172 , 418 (2011). DOI:10.1016/j.cej.2011.05.108|
|||R. Dreyer D, Park S, W. Bielawski C, and S. Ruoff R, Chem. Soc. Rev. 39 , 228 (2010). DOI:10.1039/B917103G|
|||Yu X, Huo Y, Yang J, Chang S, Ma Y, and Huang W, Appl. Surf. Sci. 280 , 450 (2013). DOI:10.1016/j.apsusc.2013.05.008|
|||M. Julkapli and S. Bagheri N, Int. J. Hydrogen Energy 40 , 948 (2015). DOI:10.1016/j.ijhydene.2014.10.129|
|||Rostamnia S, Doustkhah E, Karimi Z, Amini S, and Luque R, ChemCatChem 7 , 1678 (2015). DOI:10.1002/cctc.v7.11|
|||Zahed and H. Hosseini-Monfared B, Appl. Surf. Sci. 328 , 536 (2015). DOI:10.1016/j.apsusc.2014.12.078|
|||S. Hummers and R. E. Offeman W, J. Am. Chem. Soc. 80 , 1339 (1958). DOI:10.1021/ja01539a017|
|||Bao C, Song L, Xing W, Yuan B, A. Wilkie C, Huang J, Guo Y, and Hu Y, J. Mater. Chem. 22 , 6088 (2012). DOI:10.1039/c2jm16203b|
|||J. Shin H, K. Kim K, Benayad A, M. Yoon S, K. Park H, S. Jung I, H. Jin M, K. Jeong H, M. Kim J, Y. Choi J, and H. Lee Y, Adv. Func. Mater. 19 , 1987 (2009). DOI:10.1002/adfm.v19:12|
|||L. Li X, Y. Zhang G, D. Bai X, M. Sun X, R. Wang X, Wang E, and J. Dai H, Nat. Nanotechnol. 3 , 538 (2008). DOI:10.1038/nnano.2008.210|
|||Pritchard J, Kesavan L, Piccinini M, He Q, Tiruvalam R, Dimitratos N, A. Lopez-Sanchez J, F. Carley A, K. Edwards J, J. Kiely C, and J. Hutchings G, Langmuir 26 , 16568 (2010). DOI:10.1021/la101597q|
|||M. Scheuermann G, Rumi L, Steurer P, Bannwarth W, and Mülhaupt R, J. Am. Chem. Soc. 131 , 8262 (2009). DOI:10.1021/ja901105a|
|||Nie R, Wang J, Wang L, Qin Y, Chen P, and Hou Z, Carbon 50 , 586 (2012). DOI:10.1016/j.carbon.2011.09.017|
|||J. McAllister M, L. Li J, H. Adamson D, C. Schniepp H, A. Abdala A, Liu J, Herrera-Alonso M, L. Milius D, Car R, K. Prud'homme R, and A. Aksay I, Chem. Mater. 19 , 4396 (2007). DOI:10.1021/cm0630800|
|||L. Figueiredo J, F. R. Pereira M, M. A. Freitas M, and J. M. Órfão J, Ind. Eng. Chem. Res. 46 , 4110 (2007). DOI:10.1021/ie061071v|
|||Brender P, Gadiou R, C. Rietsch J, Fioux P, Dentzer J, Ponche A, and Vix-Guterl C, Anal. Chem. 84 , 2147 (2012). DOI:10.1021/ac102244b|
|||Fan X, Peng W, Li Y, Li X, Wang S, Zhang G, and Zhang F, Adv. Mater. 20 , 4490 (2008). DOI:10.1002/adma.v20:23|
|||Gao W, B. Alemany L, Ci L, and M. Ajayan P, Nat. Chem. 1 , 403 (2009). DOI:10.1038/nchem.281|
|||C. Ferrari and J. Robertson A, Phys. Rev ,B61 , 14095 (2000).|
|||I. Paredes J, Villar-Rodil S, Solís-Fernández P, Martínez-Alonso A, and M. D. Tascón J, Langmuir 25 , 5957 (2009). DOI:10.1021/la804216z|
|||Stankovich S, A. Dikin D, D. Piner R, A. Kohlhaas K, Kleinhammes A, Jia Y, Wu Y, T. Nguyen S, and S. Ruoff R, Carbon 45 , 1558 (2007). DOI:10.1016/j.carbon.2007.02.034|
|||Laref S, Delbecq F, and Loffreda D, J. Catal. 265 , 35 (2009). DOI:10.1016/j.jcat.2009.04.010|
|||N. Zope B, D. Hibbitts D, Neurock M, and J. Davis R, Science 330 , 74 (2010). DOI:10.1126/science.1195055|