Solid State Sciences 12 (2010) 1029e1034

Contents lists available at ScienceDirect

Solid State Sciences

Synthesis and acid catalysis of cellulose-derived carbon-based solid acid

Materials and Structures Laboratory, Tokyo Institute of Technology, 4259-R3-33 Nagatsuta-cho, Midori-ku, Yokohama 226-8503, Japan Kanagawa Academy of Science and Technology, Sakado 3-2-1, Takatsu-ku, Kawasaki 213-0012, Japan c Research Institute of Instrumentation Frontier, National Institute of Advanced Industrial Science and Technology (AIST), Central 5, 1-1-1 Higashi, Tsukuba 305-8565, Japan
b

a r t i c l e i n f o
Article history: Received 2 January 2009 Received in revised form 8 January 2010 Accepted 28 February 2010 Available online 27 March 2010 Keywords: Solid acid catalyst Amorphous carbon Brnsted acid Hydrolysis of b-1,4 glucan

a b s t r a c t
SO3H-bearing amorphous carbon, prepared by partial carbonization of cellulose followed by sulfonation in fuming H2SO4, was applied as a solid catalyst for the acid-catalyzed hydrolysis of b-1,4 glucan, including cellobiose and crystalline cellulose. Structural analyses revealed that the resulting carbon material consists of graphene sheets with 1.5 mmol g1 of SO3H groups, 0.4 mmol g1 of COOH, and 5.6 mmol g1 of phenolic OH groups. The carbon catalyst showed high catalytic activity for the hydrolysis of b-1,4 glycosidic bonds in both cellobiose and crystalline cellulose. Pure crystalline cellulose was not hydrolyzed by conventional strong solid Brnsted acid catalysts such as niobic acid, Naon NR-50, and Amberlyst-15, whereas the carbon catalyst efciently hydrolyzes cellulose into water-soluble saccharides. The catalytic performance of the carbon catalyst is due to the large adsorption capacity for hydrophilic reactants and the adsorption ability of b-1,4 glucan, which is not adsorbed to other solid acids. 2010 Published by Elsevier Masson SAS.

1. Introduction Sulfuric acid is one of the most popular acid catalysts for practical chemical processes and is widely used in the production of industrially important chemicals. Such homogeneous acids generally require special processing in the form of neutralization, which involves energy-inefcient catalyst separation from products and results in the formation of a large amount of sulfate wastes [1e4]. The development of environmentally sustainable chemical processes has stimulated the use of solid acid catalysts, because solid acids can be easily separated from the reaction matrix by simple ltration or decantation for repeated use. Replacing conventional homogeneous acids with heterogeneous counterparts, such as silica-alumina [5], zeolites [6e8], niobic acid [9,10], and strong ion-exchangeable resins [11e13], is a most promising solution to this problem; however, the catalytic activity and stability of these acids are much lower than those of homogeneous acids. Therefore, there has been increasing demand for the development of stable and highly active solid acid catalysts [14e17]. Recently, we have developed novel solid acid materials referred to as carbon-based solid acids that contain nano-sized graphene
* Corresponding author. Materials and Structures Laboratory, Tokyo Institute of Technology, 4259-R3-33 Nagatsuta-cho, Midori-ku, Yokohama 226-8503, Japan. Tel.: 81 45 924 5311; fax: 81 45 924 5381. E-mail address: [email protected] (M. Hara). 1293-2558/$ e see front matter 2010 Published by Elsevier Masson SAS. doi:10.1016/j.solidstatesciences.2010.02.038

sheets (ca. 1.5 nm) with high densities of sulfonic acid groups in the amorphous carbon, which function as efcient and reusable solid acid catalysts for industrially important chemical reactions such as hydration, hydrolysis, and esterication: the catalytic activity of the carbon-based solid acids in these reactions is comparable to that of sulfuric acid [18e22]. These catalysts can be prepared by incomplete carbonization of sulfopolycyclic aromatic hydrocarbons [18] or sulfonation of incompletely carbonized natural organic materials containing D-glucose units [19e22]. These D-glucose-derived carbon-based solid acids can incorporate large amounts of hydrophilic molecules, including water, into the carbon bulk, due to the presence of hydrophilic functional groups such as eSO3H, eCOOH, and phenolic eOH groups on the exible carbon sheets [19e21]. Incorporation of hydrophilic molecules provides good access by reactants to the SO3H groups, which gives rise to high catalytic performance, despite the low surface area (ca. 2 m2 g1) [18e22]. The conversion of cellulose into useful saccharides is an attractive route to the environmentally benign production of chemicals, including ethanol [23,24]. Sulfuric acid can function as an effective catalyst for the hydrolysis of cellulose into water-soluble saccharides. However, this catalyst requires a special processing in the form of neutralization, involving costly and inefcient catalyst separation. This results in large energy consumption [25e27]. Recently, we found that amorphous carbon bearing eSO3H, eCOOH, and phenolic eOH groups, prepared by sulfonation of partially carbonized cellulose, can catalyze the hydrolysis of

1030

S. Suganuma et al. / Solid State Sciences 12 (2010) 1029e1034

cellulose [22]. The particulate catalyst can be readily separated from the aqueous solution containing saccharides, allowing for the repeated reuse without minimum energy consumption. The reaction does not proceed in the tested conventional strong solid Brnsted acid catalysts such as niobic acid, H-mordenite, Naon NR-50 and Amberlyst-15. While the carbon catalyst can function as an active catalyst for the hydrolysis of cellulose, the details of the structure and acid catalysis of the cellulose-derived carbon catalyst have not yet been cleared. In this study, the structure and acid catalysis of the carbon-based solid acid have been studied through structural analyses and the hydrolysis of cellobiose and cellulose. 2. Experimental 2.1. Preparation of the carbon-based solid acid The carbon-based solid acid was prepared by sulfonation of incompletely carbonized cellulose. 500 g of microcrystalline cellulose powder was heated for 1 h at 673 K under N2 ow to obtain amorphous carbon, which was then ground to a powder. The powder (5 g) was then boiled in 150 mL of fuming sulfuric acid (15 wt% SO3) at 353e393 K under N2 ow. After heating for 10 h and then cooling in an ice bath, 1000 mL of distilled water was added to the mixture to form a black precipitate. The precipitate was washed repeatedly in hot distilled water at 373 K until sulfate ions were not detected in the wash water. 2.2. Characterization Structural information for the prepared carbon material was obtained by powder X-ray diffraction (XRD; Ultima IV, RIGAKU), Raman spectroscopy (NRS-3100, JASCO), and 13C cross-polarization (CP) magic angle spinning (MAS) nuclear magnetic resonance (NMR). 13C CP/MAS NMR spectra were measured at room temperature using a Bruker ASX-200 spectrometer at a Lamor frequency of 50.3 MHz. A Bruker MAS probehead was used with a 7 mm zirconia rotor. The spinning rate of the sample was 4.0 kHz. The frequency of the spectra is expressed with respect to pure tetramethylsilane. Glycine was used as a second reference material, with a carbonyl signal set at 176.48 ppm. The Brunauer-Emmet-Teller (BET) surface areas and water vapor adsorption isotherms for the samples were measured using a surface area analyzer (NOVA-4200e, QUANTACHROME) and volumetric absorption equipment (AUTOSORB MP/VP, QUANTACHROME), respectively. The amount of functional groups bonded to the carbon-based solid acid was estimated by elemental analysis (vario MICRO cube, ELEMENTAR) and cation-exchange analysis. According to X-ray photoelectron spectroscopic (XPS) analysis, all S atoms in the carbon material, which possesses SO3H, COOH, and phenolic OH, is expected to be conned to SO3H groups [19e22]. The densities of SO3H groups were thus estimated based on the sulfur content determined from sample compositions obtained by elemental analysis. The total SO3H COOH and SO3H COOH OH contents were estimated from the exchange of Na in aqueous NaCl and NaOH solutions, respectively, to afford the proportions of each functional group [22]. 2.3. Hydrolysis of cellobiose Hydrolysis of cellobiose was carried out at 373 K in cellobiose solution (cellobiose, 0.1 g; water, 10 mol) over 0.2 g of carbon catalyst. Niobic acid (Nb2O5$nH2O, CBMM), Naon NR50 (peruorosulfonated ionomer, Aldrich), and Amberlyst-15 (polystyrene-based cation-exchangeable resin with SO3H, Aldrich) catalysts (0.2 g) were used for comparison. The liquid phase during

reaction was analyzed by liquid chromatography (LC; LC-2000 plus, JASCO). 2.4. Hydrolysis of cellulose The hydrolysis of pure microcrystalline cellulose (Avicel) was typically carried out in a Pyrex reactor (catalyst, 0.300 g; cellulose, 0.025 g; distilled water, 0.700 g) at 373 K. The catalysts tested for hydrolysis of cellulose were sulfuric acid (H2SO4; 96%, KANTO Chemicals), niobic acid, Naon NR50, Amberlyst-15, and the carbon catalyst. After the reaction, an aliquot of the supernatant solution was decanted and analyzed. The yields of glucose and water-soluble b-1,4 glucan were determined by LC, and the enzymatic hydrolysis of water-soluble b-1,4 glucan using cellulase, respectively. The amount of glucose produced in the collected supernatant solution was estimated by LC. 0.2 g of crude cellulase obtained from Trichoderma viride (Wako Pure Chemical industries) was then added to the collected supernatant solution, and the solution was warmed for 48 h at 313 K in order to hydrolyze the water-soluble b-1,4 glucan in the solution into glucose. After the enzymatic hydrolysis, glucose produced by hydrolysis of b-1,4 glucan was also analyzed by LC. Cellulose conversion was obtained using the following equations [22]. Cellulose conversion (%) 100(B C)/A. A: total amount (mol) of glucose monomer in cellulose B: amount (mol) of glucose produced by acid-catalyzed hydrolysis C: total amount (mol) of glucose monomer in water-soluble b-1,4 glucan produced by acid-hydrolyzed hydrolysis. 3. Results and discussion 3.1. Characterization of carbon materials Fig. 1 shows XRD patterns and Raman spectra for the carbonized cellulose before and after sulfonation. The XRD patterns in Fig. 1(A) exhibit a weak and broad C(002) diffraction peak (2q 10e30 ) attributed to amorphous carbon. This indicates that both the carbon precursor and carbon catalyst are composed of aromatic carbon sheets oriented in a considerably random fashion [28]. The Raman spectra (Fig. 1(B)) display two broad signals assigned to the D-band (ca. 1400 cm1, A1g D breathing mode) and G-band (ca. 1590 cm1, E2g G mode). The intensity ratio of the D- and G-bands for these samples is nearly equal, indicating that there is no signicant difference in the average sizes of graphene among these materials. The average size of graphene was estimated to be approximately 1.2 nm [28]. Fig. 2 shows 13C CP/MAS NMR spectra for carbonized cellulose before and after sulfonation. Signals at 130 and 155 ppm with spinning sidebands attributed to polycyclic aromatic carbon atoms and phenolic OH groups are observed in these spectra, indicating that the carbon frameworks of these materials consist mainly of sp2-derived carbon sheets with phenolic OH groups. The broad signal for aliphatic hydrocarbon moieties (15 ppm) observed for carbonized cellulose in Fig. 2(a) disappeared and a weak signal at 180 ppm, assigned to eCOOH groups, appears after sulfonation (Fig. 2(b)). This suggests that the aliphatic hydrocarbons were decomposed and/or carbonized by the strong oxidative and dehydrating properties of fuming H2SO4 when heated. The formation of eCOOH groups in the carbon catalyst can probably be attributed to the partial oxidation of graphene sheets by H2SO4. The peak due to aromatic carbon bonded to the SO3H group (Ar-SO3H, ca. 140 ppm) is not observed in the spectrum (Fig. 2(b)), because broad peaks due to aromatic carbon atoms

S. Suganuma et al. / Solid State Sciences 12 (2010) 1029e1034

1031

Intensity / a.u.

(b) (b) (a)

(a)
0 10 20 30 40 2 / degree 50

Fig. 1. XRD patterns (A) and Raman spectra (B) for carbonized cellulose (a) before and (b) after sulfonation.

(130 ppm) and phenolic OH groups (155 ppm) completely obscure the peak. The introduction of sulfonic acid groups was directly conrmed by XPS analysis. The XPS spectra for carbon materials after sulfonation exhibit a single S2p peak attributable to SO3H groups. Fig. 3 shows S2p XPS spectra for carbonized cellulose before and after sulfonation. Although no signal is observed in the spectrum for carbonized cellulose (Fig. 3(a)), a single peak at 168 eV assigned to the SO3H groups is clearly observed in the spectrum for sulfonated carbon, indicating that sulfonation of amorphous carbon was successfully achieved in fuming H2SO4 to form sulfonic acid groups, and all sulfur in the carbon catalyst is contained in the SO3H groups. From the elemental analysis and cation-exchange experiments, the amounts of SO3H, COOH, and phenolic OH groups bonded to the graphene were estimated to be 1.5, 0.4, and 5.4 mmol g1, respectively [22]. The proposed carbon structure is schematically illustrated in Fig. 4. The prepared material is amorphous carbon consisting of SO3He, COOHe, and phenolic OHbearing graphene sheets in a considerably random fashion. The carbon catalyst possesses a high density of almost neutral phenolic OH groups in addition to Brnsted acid sites (SO3H and COOH),

which is distinct from conventional solid acids that have single functional groups, such as sulfonated resins. One of the important features of the carbon catalyst is that it can adsorb large amounts of hydrophilic molecules. A carbon catalyst prepared from D-glucose was reported to have a high density of effective Brnsted acid sites and exhibited high catalytic performance for acid-catalyzed reactions, despite the low surface area (2 m2 g1) [19e21]. Fig. 5 shows the H2O vapor adsorptionedesorption isotherm of the carbon catalyst at 298 K. Although the carbon catalyst prepared from cellulose also has a low surface area (2e3 m2 g1), the amount of adsorbed H2O at 1.5 kPa is much smaller than that at saturated water vapor pressure (ca. 3 kPa) exceeds 0.01 mol g1. Assuming that the adsorption cross sectional area of H2O is 0.125 nm2, the effective surface area of the carbon catalyst is estimated to exceed over 560 m2 g1, which means that the carbon catalyst can incorporate large amounts of water into the carbon bulk, due to the high density of hydrophilic functional groups bound to the exible nanographene sheets. This provides good access by reactants in solution to the SO3H groups in the

168 eV
130 ppm

155 ppm

180 ppm 15 ppm * * (a) 300 250 200 150 / ppm 100 50 0 * * (b)

(b)

(a)

175

170 165 Binding energy / eV

160

Fig. 2. 13C CP/MAS NMR spectra for carbonized cellulose (a) before and (b) after sulfonation. * denotes spinning sidebands.

Fig. 3. S2p XPS spectra for carbonized cellulose (a) before and (b) after sulfonation.

1032

S. Suganuma et al. / Solid State Sciences 12 (2010) 1029e1034

20
H SO 3

OH

HO
CO OH

SO3H (1.5 mmol g-1) COOH (0.4 mmol g-1) OH (5.4 mmol g-1) Surface area: 2 m2 g-1

(a) 15 D-glucose yield / % (b) 10 (c)

Fig. 4. Proposed structure for carbon-based solid acid prepared from cellulose.

carbon material, resulting in high catalytic performance, despite the low surface area [22]. 3.2. Hydrolysis of cellobiose by solid acid catalysts Fig. 6 shows time courses for D-glucose formation from cellobiose over the carbon catalyst. For comparison, the results for Naon resin (NR50), Amberlyst-15, and niobic acid are included. Naon NR50 and Amberlyst-15 are polymer-based strong solid acids with high densities of SO3H, and very high activity for a variety of reactions [11e13]. Niobic acid is a typical inorganic oxide strong solid acid that is widely used in industrial acid-catalyzed reactions [9,10]. Although the catalytic activity of niobic acid is moderate, the SO3Hbearing solid acid catalysts exhibit high catalytic activity for this reaction. The catalytic activity for the carbon catalyst is approximately twice as high as those of Amberlyst-15 and Naon NR-50. Therefore, the difference in the catalytic properties of sulfonic acidbearing catalysts for this reaction can be adequately explained as not being due to acid amount or acid strength. Catalysis over the carbon catalyst with SO3H, COOH, and phenolic OH was thus further investigated by examining the adsorption of cellobiose on each of the solid acid catalysts used in this study [29]. 0.1 g of each solid acid was added to 2 cm3 (cellobiose, 3 mol) of aqueous cellobiose solution (1.5 mmol L1) and then stirred at room temperature. Cellobiose was not hydrolyzed into glucose at room temperature under the experimental conditions, and the amount of adsorbed cellobiose was estimated by LC of the supernatant solution. The conventional solid acids, such as Naon, and Amberlyst-15, did not adsorb cellobiose in water even after stirring for 24 h, whereas the carbon catalyst adsorbed 24% of the cellobiose in just 3 h. Cellobiose adsorption reached a plateau at 5 h, with adsorption of 26% (0.8 mol) of the

(d)

0 0 1 2 4 3 Time / h 5 6 7

Fig. 6. Time courses of D-glucose formation from cellobiose over various solid acid catalysts at 373 K: (a) carbon-based solid acid, (b) Naon NR50, (c) Amberlyst-15, and (d) niobic acid. Reaction conditions: 0.2 g catalyst, 0.1 g cellobiose, and 10 mL water.

cellobiose dissolved in the solution, which suggests that the carbon material has a stronger afnity for b-1,4 glucan than the other solid acids. It should be pointed out that the carbon catalyst cannot adsorb D-glucose at all, suggesting that the glycosidic bond in b-1,4 glucan participates in adsorption on the carbon catalyst. The ability to adsorb b-1,4 glucan may be attributed to the phenolic OH groups bonded to the graphene sheets. In the cellobiose adsorption experiment for the COOH-containing (Amberlite IRC-50 (Aldrich)) and SO3H-containing (Amberlyst-15) resins, cellobiose was not adsorbed to the resins, indicating that these functional groups do not adsorb cellobiose. With regard to the stronger afnity for b-1,4 glucan, the large surface area of the carbon catalyst in a water medium can enhance the incorporation of cellobiose, providing good access by reactants to the SO3H groups in the carbon material and resulting in high catalytic activity. The results for the carbon catalyst reusability experiment are shown in Fig. 7. Judging from statistics, the experimental error in glucose yield was estimated to be 3% yield [30]. As a result, there is no signicant difference in activity among the catalyst reuse

600 500

(b)

Volume [cm3 / g]

400 300 200

(a)
100 0 0

0.2

0.4

0.6

0.8

Relative pressure [p / p0]

S. Suganuma et al. / Solid State Sciences 12 (2010) 1029e1034

1033

100 80 60 40 20 0 0 (b)

(a)

(c), (d), (e)

2 4 Reaction time / h

Fig. 8. Time courses of cellulose conversion by hydrolysis (0.300 g catalyst, 0.025 g cellulose, 0.700 g water, and reaction temperature of 373 K) for pure crystalline cellulose using (a) carbon material, (b) sulfuric acid, (c) niobic acid, (d) Naon NR-50, and (e) Amberlyst-15.

experiments. In addition, no decrease in activity and SO3H density is observed even after 4 reuses for the hydrolysis of cellobiose by the carbon material [29]. These results indicate that SO3H groups are tightly bonded to the graphene sheets, and thus the carbon catalyst can function as a stable and highly active solid acid for the hydrolysis of b-1,4 glycosidic bond. 3.3. Hydrolysis of cellulose by solid acid catalysts The performance of various acid catalysts and sulfuric acid for the hydrolysis of pure microcrystalline cellulose powder was examined at 373 K; the time courses of cellulose conversion are shown in Fig. 8. Cellulose could not be hydrolyzed into glucose or water-soluble b-1,4 glucan using conventional solid acid catalysts such as niobic acid, Amberlyst-15, and Naon NR-50. However, the carbon catalyst exhibited remarkable performance for the hydrolysis reaction, with catalytic activity comparable to that of

sulfuric acid. All cellulose was converted into water-soluble saccharides over the carbon catalyst within 6 h, and the catalyst could be readily separated by decantation from the solution containing dissolved saccharides after the reaction. Fig. 9 shows the product distribution for the reaction, as determined by LC. While the conventional solid strong Brnsted acids of niobic acid, Naon NR-50, and Amberlyst-15 did not produce any water-soluble saccharides, the carbon catalyst hydrolyzed cellulose into watersoluble oligosaccharides and D-glucose, as did sulfuric acid. The water-soluble oligosaccharides produced in this reaction were analyzed by matrix-assisted laser desorption/ionization-time-ofight-mass spectrometry (MALDI-TOF-MS) of the reaction solution after 3 h in the presence of the carbon catalyst. Large signals appeared at m/z 162 intervals, which is the mass number of the glucose monomer (e(eOeC6H10O4e)ne) in b-1,4 glucan, indicating that the water-soluble hydrolysis product from cellulose was b-1,4 glucan (results not shown). The b-1,4 glucan component was estimated to be C6H11O5e(eOeC6H10O4e)2w10eOeC6H11O5 by MALDI-TOF-MS. The results indicate that the hydrolysis of cellulose on the carbon material proceeds as well as that on sulfuric acid: H attacks hydrogen and b-1,4 glycosidic bonds in solid crystalline cellulose to form water-soluble b-1,4 glucan, followed by hydrolysis of the b-1,4 glycosidic bonds in b-1,4 glucan to form glucose. Therefore, efcient conversion of cellulose into glucose using a solid acid catalyst requires strong interaction between the solid acid and b-1,4 glucan (solid cellulose and water-soluble b-1,4 glucan), because the Brnsted acid sites of the solid acid cannot approach the cellulose surface without such interaction. This is distinct from hydrolysis in the homogeneous sulfuric acid system. The SO3H-bearing carbon material, which is capable of adsorbing b-1,4 glucan, can thus be expected to have much higher activity for hydrolytic catalysis than the other solid acids examined in this study. 4. Conclusion The sulfonation of carbon material prepared by low temperature carbonization of cellulose resulted in a highly active, stable solid acid catalyst as a replacement for sulfuric acid. The carbon catalyst, consisting of polycyclic aromatic carbon with SO3H, COOH, and phenolic OH, was demonstrated to function as a highly active catalyst for the direct hydrolysis of b-1,4 glucan in solid cellulose, despite the low surface area of the catalyst. The high catalytic activity for the reaction can be attributed to the ability to adsorb b-1,4 glucan and the large effective surface area of the catalyst in water. Acknowledgement This work was supported by the New Energy and Industrial Technology Development Organization (NEDO, 04A32502), the Research and Development in a New Interdisciplinary Field Based on Nanotechnology and Materials Science programs of the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan, and a Grant-in-Aid for Scientic Research (18206081) from the Japan Society for the Promotion of Science (JSPS). References
[1] [2] [3] [4] [5] [6] [7] P.T. Anastas, M.M. Kirchhhoff, Acc. Chem. Res. 35 (2002) 686. J.M. DeSimone, Science 297 (2002) 799. B. Harton, Nature 400 (1999) 797. P.T. Anastas, J.B. Zimmermann, Environ. Sci. Technol. 37 (2003) 94A. J.H.D. Boer, W.J. Visseren, Catal. Rev. 5 (1972) 55. S. Namba, N. Hosonuma, T. Yashima, J. Catal 72 (1981) 16. H. Ishida, Catal. Surv. Jpn. 1 (1997) 241.

Fig. 9. Product distributions for the hydrolysis of pure crystalline cellulose over various solid acids and sulfuric acid at 373 K. Reaction condition: 0.300 g catalyst, 0.025 g cellulose, 0.7 g water, and reaction time of 3 h.

Cellulose conversion / %

1034 [8] [9] [10] [11] [12] [13] [14] [15] [16] [17]

S. Suganuma et al. / Solid State Sciences 12 (2010) 1029e1034 [21] A. Takagaki, M. Toda, M. Okamura, J.N. Kondo, S. Hayashi, K. Domen, M. Hara, Catal. Today 116 (2006) 157. [22] S. Suganuma, K. Nakajima, M. Kitano, D. Yamaguchi, H. Kato, S. Hayashi, M. Hara, J. Am. Chem. Soc. 130 (2008) 12787. [23] A.J. Ragauskas, C.K. Williams, B.H. Davison, G. Britovsek, J. Cairney, C.A. Eckert, W.J. Frederick, J.P. Hallett, D.J. Leak, C.L. Liotta, J.R. Mielenz, R. Murphy, R. Templer, T. Tschaplinski, Science 311 (2006) 484. [24] B. Hahn-Haegerdal, M. Galbe, M.F. Gorwa-Grauslund, G. Liden, G. Zacchi, Trends Biotechnol. 24 (2006) 549. [25] E.C. Sherrard, F.W. Kressman, Ind. Eng. Chem. 37 (1945) 5. [26] W.L. Faith, Ind. Eng. Chem. 37 (1945) 9. [27] E.E. Harris, E. Beglinger, Ind. Eng. Chem. 38 (1946) 890. [28] A.C. Ferrari, J. Robertson, Phys. Rev. B 61 (2000) 14095. [29] M. Kitano, D. Yamaguchi, S. Suganuma, K. Nakajima, H. Kato, S. Hayashi, M. Hara, Langmuir 25 (2009) 5068. [30] D. Yamaguchi, M. Kitano, S. Suganuma, K. Nakajima, H. Kato, M. Hara, J. Phys. Chem. C 113 (2009) 3181.

K. Eguchi, T. Tokiai, H. Arai, Appl. Catal 34 (1987) 275. I. Nowak, M. Ziolek, Chem. Rev. 99 (1999) 3603. K. Tanabe, S. Okazaki, Appl. Catal 133 (1995) 191. G.A. Olah, P.S. Lyer, G.K.S. Prakasn, Synthesis (1986) 513. M.M. Sharma, React. Polym. (1995) 1. A. Chakrabati, M.M. Sharma, React. Polym. (1993) 1. J.H. Clark, Acc. Chem. Res. 35 (2002) 791. M. Misono, C.R. Acad. Sci. Ser. IIc 3 (2000) 471. T. Okuhara, Chem. Rev. 102 (2002) 3641. K. Smith, G.A. El-Hiti, A. Gamal, A.J. Jayne, K. Butters, Org. Biomol. Chem. 1 (2003) 1560. [18] M. Hara, T. Yoshida, A. Takagaki, T. Takata, J.N. Kondo, K. Domen, S. Hayashi, Angew. Chem. Int. Ed. 43 (2004) 2955. [19] M. Toda, A. Takagaki, M. Okamura, J.N. Kondo, S. Hayashi, K. Domen, M. Hara, Nature 438 (2005) 178. [20] M. Okamura, A. Takagaki, M. Toda, J.N. Kondo, T. Tatsumi, K. Domen, M. Hara, S. Hayashi, Chem. Mater. 18 (2006) 3039.