Co-decorated carbon nanotubes as a promoter of Co–Mo–K oxide catalyst for synthesis of higher alcohols from syngas

Available online at www.sciencedirect.com

Applied Catalysis A: General 340 (2008) 87–97 www.elsevier.com/locate/apcata

Co-decorated carbon nanotubes as a promoter of Co–Mo–K oxide catalyst for synthesis of higher alcohols from syngas Xiao-Man Wu, Yan-Yan Guo, Jin-Mei Zhou, Guo-Dong Lin, Xin Dong, Hong-Bin Zhang * Department of Chemistry, College of Chemistry and Chemical Engineering, State Key Laboratory of Physical Chemistry for Solid Surfaces, Xiamen University, Xiamen 361005, China Received 14 June 2007; received in revised form 27 September 2007; accepted 28 September 2007 Available online 9 February 2008

Abstract A metal cobalt-decorated multi-walled carbon nanotube (MWCNT)-promoted Co–Mo–K oxide-based catalyst was developed, with excellent performance for the selective formation of C2–9-alcohols from syngas. Under reaction condition of 5.0 MPa and 593 K, the space–time-yield of C2–9-alcohols reached 628 mg/(g h) over the Co1Mo1K0.05–12%(4.2% Co/MWCNT) catalyst. The addition of a minor amount of the Co-decorated MWCNTs into the Co1Mo1K0.05 host catalyst caused little change in the apparent activation energy for the higher alcohol synthesis (HAS), but led to an increase of surface concentration of the two kinds of catalytically active species, CoO(OH)/Co3O4 and Mo4+, both closely associated with the alcohol generation. Addition of 5% CO2 in the feed gas at properly elevated reaction temperature (593 K) could further enhance the surface concentration of those active Mo and Co species. An excellent adsorption performance of the Co-decorated MWCNTs as promoter for H2 would be conducive to generating a surface micro-environment with a high concentration of H-adspecies on the functioning catalyst, thus increasing the rate of surface hydrogenation reactions in the HAS. In addition, high concentration of H-adspecies on the catalyst would, through synergistic action with the CO2 in the feed gas, greatly inhibit the water–gas-shift side-reaction. All these factors contribute to an increase in the yield of alcohols. # 2008 Elsevier B.V. All rights reserved. Keywords: Co-decorated carbon nanotubes; Co–Mo–K catalyst; Higher alcohol synthesis

1. Introduction The higher alcohols (C2+-alcohols), together with methanol and dimethyl ether (DME), have been considered as the most important species among coal-based clean synthetic fuels and chemical feedstocks. The higher alcohols have been confirmed to be a better and cleaner automobile fuel. They feature high octane numbers, and lower emissions of NOx, ozone, CO and aromatic vapors [1]. Recently, use of methyl tert-butyl ether (MTBE) has been prohibited in some countries or regions as additive of oil-based fuel due to the new legal requirements in environment protection; this change has greatly renewed interest in hydrogenation of syngas to the C2+-oxygenates as gasoline blends. Higher alcohol syntheses (HAS) on modified methanol catalysts [2,3], modified Fischer–Tropsch catalysts [4,5], alkali-promoted MoS2 catalysts [6,7] and catalysts containing Mo, Group VIII metals and alkali [8–11], have been

* Corresponding author. Tel.: +86 592 2184591; fax: +86 592 2086116. E-mail address: [email protected] (H.-B. Zhang). 0926-860X/$ – see front matter # 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.apcata.2007.09.051

extensively studied since 1980s. Progress in this field has considerably contributed to the growing understanding of the nature of those catalytic reaction systems. Nevertheless, the existing technology of HAS is still on a small scale. The singlepass-conversion of the feed syngas and selectivity to C2+-alcohols were both relatively low. Most systems produce methanol (over alkali-promoted MoS2 catalysts) or hydrocarbons (over modified Fischer–Tropsch catalysts) as the main product instead of C2+-alcohols [12–14]. Development of catalysts with high efficiency and selectivity for C2+-alcohols has been one of the key objectives for research and development efforts. Multi-walled carbon-nanotubes (MWCNTs) [15], as novel nanocarbon supports or promoters of catalysts, have drawn increasing attention recently [16–18]. The catalytic applications range from selective hydrogenation [19], hydrofomylation [20], ammonia synthesis [21], Fischer–Tropsch synthesis [22], methanol and higher alcohol synthesis [23–25], to selective dehydrogenation [26], selective oxidation [27] and electrocatalysis [28]. The catalytic studies conducted so far on MWCNT-based systems have shown promising results in terms

88

X.-M. Wu et al. / Applied Catalysis A: General 340 (2008) 87–97

of activity and selectivity. From a chemical catalysis point of view, the excellent performance of MWCNTs in adsorption– activation of hydrogen and in promoting spillover of adsorbed H-species is very attractive, in addition to its high-mechanical strength, nanosize channels, sp2-C constructed surfaces, graphite-like tube-walls and high-thermal/electrical conductivity; modification of some transition-metals to MWCNTs might further improve their performance for adsorbing and activating hydrogen, as well as promoting H-adspecies spillover [29]. In studies of preparation of MWCNT-based functionalized materials, a most straightforward route by melting the elements on MWCNT surfaces was earlier reported by Dujardin et al. [30]. However, research efforts by Ebbesen and coworkers [31] showed that only liquids with low surface tensions could wet MWCNT surfaces. Ang et al. [32] reported a so-called ‘‘twostep sensitization–activation method’’, where the surfaces of oxidized MWCNTs were activated by the introduction of catalytic nuclei, and then the activated MWCNTs catalyze metal deposition specifically onto their surfaces upon immersion in electroless plating baths. However, there might be a small concentration of some heterogeneous metal element such as Sn or Pd remaining in the sensitization–activation process in metal-decorated MWCNTs. Figlarz et al. [33–35] developed a so-called ‘‘polyol method’’ for synthesis of fine metallic powders of Co, Cu, Ni, Pb and Ag, in which the metal precursor(s) are suspended or dissolved in a polyol such as ethylene glycol. The resultant glycol–metal precursor mixture is then heated to reflux and the metallic moieties precipitated out of solution. This can be a viable catalyst-free method for the deposition of fine metallic powders. Metal-decorated MWCNTs could also be prepared by metal vapor deposition. Nevertheless, the dispersion of metal or metal compounds on MWCNTs by chemical means is still largely based on conventional catalyst preparation techniques, such as wet impregnation followed by chemical reduction. In the present work, several metal cobalt-decorated MWCNTs, abbreviated as x% Co/MWCNT (where x% represents mass percentage), were prepared by an intermittent microwave irradiation-assisted polyol reduction–deposition method, and characterized through transmission electron microscope (TEM), scanning electron microscope (SEM)/ energy dispersive spectrum (EDS), XRD and H2-TPD (temperature-programmed desorption) measurements. Using the x% Co/MWCNTas promoter, we prepared several x% Co/MWCNTpromoted Co–Mo–K oxide-based catalysts by the combined coprecipitation and impregnation method. The catalysts displayed higher catalytic activity and selectivity for HAS from syngas, as compared to the MWCNT-free counterpart and to a reference catalyst promoted by the simple MWCNTs. 2. Experimental 2.1. Preparation of MWCNTs The MWCNTs were prepared from catalytic decomposition of CH4 by the method reported previously [36]. The prepared

MWCNTs were a herringbone-type of multi-walled carbon nanotubes, with the outer diameters of 10–50 nm, inner diameters of 3–7 nm, and N2-BET surface area of 140 m2/ g. Such nanotubes were constructed by a superposition of many graphene layer facets, which were tilted at a certain angle with respect to the axis of the central hollow nanofibre, as if a number of cones were placed one on top of the other [37]. The freshly prepared MWCNTs were purified with boiling concentrated nitric acid, followed by rinsing with deionized water, then drying at 383 K under dry N2 atmosphere. Openend MWCNTs with hydrophilic surface were thus obtained. In the purified MWCNTs, the contents of the total carbon and the graphitized carbon were 99.5% and 90% (mass percentage), as evidenced by elemental analysis and O2-TPO (temperature-programmed oxidation) measurements, respectively. Tests of H2-temperature-programmed hydrogenation (TPH) showed that the temperature needed for initiating the hydrogenation reaction of the MWCNTs with H2 was 773 K, indicating that this type of MWCNTs was stable in H2atmosphere at the reaction temperatures for the HAS [25]. 2.2. Preparation of Co-decorated MWCNTs Each x% Co/MWCNT was prepared according to the following procedure in reference to the synthesis parameters used by Kurihara et al. [38]. In brief, 0.66 g of Co(Ac)2 (of AR grade) was put into 50 mL ethylene glycol (of AR grade) in a 100-mL beaker, followed by agitating till the Co precursor salt was completely dissolved. 2.11 g of the purified MWCNTs were put into the above solution under agitating by ultrasonicating. The beaker was placed in the center of a household microwave oven (2450 MHz, 720 W). After microwave-heating for 2 min, the suspending mixture was agitated for 10 s, then put into the microwave oven and heated for 1 min. After another round of the agitating/microwave-heating process, the resulting suspension was filtered. The precipitate was then rinsed several times with deionized water, and finally dried at 373 K. A Co-decorated MWCNTs material was thus obtained. Element analysis showed that the cobalt content in the prepared material was at 4.2% (mass percentage). By changing the loading amounts of MWCNTs and Co(Ac)2, one could prepare a series of x% Co/MWCNT with varying Co contents. 2.3. Preparation of Co/MWCNT-promoted Co–Mo–K oxide catalysts A series of x% Co/MWCNT-promoted Co–Mo–K oxidebased catalysts, denoted as CoiMojKk–y%(x% Co/MWCNT) (where y% represented mass percentage), were prepared by the combined co-precipitation and impregnation method. Two aqueous solutions containing calculated amounts of Co(NO3)2
6H2O and (NH4)6Mo7O24
4H2O (all of AR grade), respectively, were simultaneously added dropwise under vigorous stirring into a Pyrex flask containing a calculated amount of the x% Co/MWCNT at constant temperature of 353 K. The addition was adjusted to maintain a constant pH of 5. The precipitate was continuously stirred for 4 h at 353 K,

X.-M. Wu et al. / Applied Catalysis A: General 340 (2008) 87–97

followed by cooling down to room temperature, aging overnight, and then filtering. The filtered cake (precipitate) was repeatedly rinsed with deionized water until the filtrate became neutral in pH. The obtained solid was dried at 383 K for 5 h and calcined at 773 K for 6 h in N2 atmosphere, followed by cooling down to room temperature and impregnating with potassium carbonate aqueous solution containing a calculated amount of K by the conventional incipient wetness method. The solid was then dried at 383 K and calcined at 673 K under N2 atmosphere for 4 h, yielding the oxide precursor of (x% Co/ MWCNT)-promoted Co–Mo–K catalysts. The MWCNT-free host catalyst and the simple MWCNT-promoted counterpart, used as references, were prepared in a similar way. All samples of catalyst precursor were pressed, crushed, and sieved to a size of 40–80 mesh for the activity evaluation. 2.4. Catalyst evaluation Performance of the catalysts for the HAS from syngas was evaluated in a fixed-bed continuous-flow reactor and gas chromatograph (GC) combination system. Prior to the reaction, the sample of the oxide precursor of catalyst was prereduced in situ under H2 stream at 0.1 MPa and 2400 mL/ (g h). The reduction temperature was programmed to rise from room temperature to 623 K, to maintain that temperature for 12 h, and then to decrease to the desired temperature for the catalyst test. The HAS reaction was conducted at a stationary state under reaction conditions of 523–603 K, 2.0– 5.0 MPa, V(H2):V(CO):V(N2) = 60:30:10 or V(H2):V(CO):V(CO2):V(N2) = 60:30:5:5. Exit gas from the reactor was immediately brought down to atmospheric pressure and transported, while its temperature was maintained at 403 K, to the sampling valve of GC (Model GC-950 by Shanghai Haixin GC Instruments, Inc.), which was equipped with dual detectors (TCD and FID) and dual columns filled with carbon molecular sieve (TDX-01) and Porapak Q-S (USA), respectively, for online analysis. The former column (2.0-m length) was used for the analysis of N2 (as internal standard), CO and CO2, and the latter (2.0-m length) for C1–9-hydrocarbons, C1– 9-alcohols and other oxygenates (mainly DME and acetaldehyde in a small amount). CO hydrogenation–conversion (namely the CO-conversion after deducting the contribution of water–gas-shift (WGS) side-reaction, simplified as X(CO) thereafter) was determined through an internal standard, and the carbon-based selectivity for the carbon-containing products, including alcohols, hydrocarbons, and other oxygenates (simplified as S(alc.), S(HC), etc. thereafter) was calculated by an internal normalization method. 2.5. Characterization of the MWCNT-based promoter and catalysts Transmission electron microscope and scanning electron microscope observations, as well as energy dispersive spectrum observations were performed with Technai F30 and LEO-1530 electron microscopes, respectively. XRD measurements were carried out on an X’Pert PRO X-ray Diffractometer (PANa-

89

lytical) with Cu Ka (l = 0.15406 nm) radiation. A continuous scan mode was used to collect 2u data from 108 to 908. The voltage and current were 40 kV and 30 mA, respectively. X-ray photoelectron spectroscopy (XPS) measurements were done on a Quantum 2000 Scanning ESCA Microprobe instrument with Al Ka radiation (15 kV, 25 W, hn = 1486.6 eV) under ultrahigh vacuum (10 7 Pa), calibrated internally by the carbon deposit C(1s) (Eb = 284.6 eV). Specific surface area (SSA) was determined by N2 adsorption using a Micromeritics Tristar3000 (Carlo Erba) system. The comparative investigations of adsorption of H2 on the 4.2% Co/MWCNT, simple MWCNTs, as well as on the Co–Mo–K catalysts promoted by these promoter materials, were conducted by using H2-temperature-programmed desorption method on a home-made adsorption–desorption/ reaction system. About 50 mg of sample was used for each test. Prior to TPD measurements, each sample was treated in situ in the TPD equipment by H2 gaseous (of 99.999% purity) at 623 K for 12 h and then flushed by an Ar (of 99.999% purity) stream at 623 K for 30 min to clean its surface, followed by cooling down to 433 K, switching to a H2 (of 99.999% purity) stream for hydrogen adsorption at 433 K for 60 min and subsequently at room temperature for 120 min, and then flushing by the Ar stream at room temperature till the stable baseline of GC appeared. The rate of temperature increase was 5 K/min. Change of hydrogen signal was monitored using an online GC (Shimadzu GC-8A) with a TC detector. 3. Results and discussion 3.1. Characterization of the prepared x% Co/MWCNT promoter material The morphology, metal-particle size distribution and surface element composition of the prepared 4.2% Co/MWCNT nanocomposites were observed with TEM (FEI-F30) and SEM/EDX (LEO-1530) instruments. The results (Fig. 1) showed that Co nanoparticles were fairly uniform in shape and size and were well distributed on the MWCNT surface. The Coparticle diameters were estimated to be below 10 nm. EDS analysis demonstrated that carbon, oxygen and cobalt were the only three elements at the surface of 4.2% Co/MWCNT, with atomic percentage of 87.82%, 10.41% and 1.77%, respectively. The corresponding mass percentages were 79.54%, 12.56% and 7.90%, respectively. The surface oxygen most probably originated from the pre-oxidation/carboxylation treatment of the MWCNTs by the concentrated nitric acid. Fig. 2 displays the XRD patterns of the purified MWCNTs and the Co-decorated MWCNTs with varying Co-decorating amounts. The observed XRD peaks at 2u = 26.18, 43.18 and 53.58 for these systems were due to the diffraction of (0 0 2), (1 0 0) and (0 0 4) faces of the MWCNTs, respectively [37,39]. The weak but distinguishable features at 2u = 34.08, 38.68 and 60.18 appeared only in the XRD patterns of the Co-decorated MWCNTs. They could be attributed to the contributions from crystallite phase of metallic cobalt. Using the well-known

90

X.-M. Wu et al. / Applied Catalysis A: General 340 (2008) 87–97

Fig. 2. XRD patterns of: (a) MWCNTs; (b) 3.2% Co/MWCNT; (c) 3.8% Co/ MWCNT; (d) 4.2% Co/MWCNT; (e) 4.6% Co/MWCNT; (f) 5.1% Co/ MWCNT.

3.2. Performance of Co–Mo–K oxide-based catalyst promoted by x% Co/MWCNT for HAS

Fig. 1. TEM (a) and SEM (b) images and energy dispersive spectrum (EDS) (c) of 4.2% Co/MWCNT.

Scherrer’s equation, we estimated the particle size of metallic cobalt to be 8.2 nm for the 4.2% Co/MWCNT, which is very close to that observed from the above TEM image. For the samples with the Co-decorating amount at 3.2%, 3.8%, 4.6% and 5.1%, these values were 6.2, 7.6, 9.8 and 11.1 nm, successively, showing an upward trend of the particle size of metallic cobalt with increasing Co-decorating amounts.

Our previous investigation [40] showed that, among a series of MWCNT-promoted oxide CoiMojKk catalysts, the catalyst with the composition of Co1Mo1K0.05–12%MWCNT displayed the highest catalytic activity for HAS from syngas, and that the presence of CO2 in the feed synthesis gas suppressed the activity of the catalysts. In the present work, with the composition of CoiMojKk host and the additive level of promoter remaining unchanged and using the Co-decorated MWCNTs in place of the simple MWCNTs as the promoter, the effect of modification of cobalt to MWCNT promoter on performance of the corresponding catalyst for HAS was investigated. The results indicated that the observed carbon-containing products included CO-hydrogenation products, C1–9-alcohols (total alc.) and C1–9-hydrocarbons (total HC), as well as CO2 coming from the WGS side-reaction. The Co1Mo1K0.05–12%(4.2% Co/MWCNT) catalyst displayed the best catalytic performance. Under the reaction conditions of 2.0 MPa, 563 K, V(H2)/V(CO)/V(N2) = 60/30/10, GHSV = 2400 mL/(h g), the X(CO) reached 28.2%, with the corresponding total alcohol yield (i.e., the product of X(CO) and S(total alc.), simplified as Y(total alc.) thereafter) reaching 25.4% over this catalyst; over the other four catalysts promoted by the x% Co/MWCNT (x% = 3.2%, 3.8%, 4.6% and 5.1%, respectively), the X(CO) was 16.6%, 22.1%, 24.9% and 23.2%, successively, with the corresponding Y(total alc.) values being 15.0%, 19.9%, 21.9% and 19.7%, respectively. In order to evaluate the performance of the catalysts under the working condition with higher extent of reaction, we conducted HAS reactions from syngas at higher pressure (5.0 MPa) and GHSV (8000 mL/(h g)). Fig. 3 shows the results of the comparative assay of HAS reaction activity over the 4.2% Co/MWCNT-promoted catalyst and over the simple MWCNT-promoted counterpart. Similar reaction–chemical behavior was observed on both catalysts. With the reaction temperature raised progressively from 523 K, X(CO) increased

X.-M. Wu et al. / Applied Catalysis A: General 340 (2008) 87–97

Fig. 3. Reactivity of HAS over catalysts: (a) Co1Mo1K0.05–12%(4.2% Co/ MWCNT); (b) Co1Mo1K0.05–12%MWCNT; reaction conditions: 5.0 MPa, V(H2)/V(CO)/V(N2) = 60/30/10, GHSV = 8000 mL/(h g), 18 h.

monotonously, while the decline of S(total alc.) was slow at temperatures 563 K, but accelerated at temperatures >563 K. Y(total alc.) approached a maximum value when the reaction temperature went up to 573 K. In order to obtain high yield of alcohols and low consumption of feed gas, we took 563 K as the optimal operating temperature. Under the reaction conditions of 563 K, X(CO) and S(total alc.) reached 21.1% and 85.0% over the Co1Mo1K0.05–12%(4.2% Co/ MWCNT) catalyst, with the corresponding Y(total alc.) and space–time-yield (STP) being 17.9% and 331.1 mg/(h g), while these values over the Co1Mo1K0.05–12%MWCNT catalyst were 18.7%, 85.0%, 15.9% and 285.9 mg/(h g), successively. Fig. 4 shows the product distribution of HAS over the (4.2% Co/MWCNT)-promoted Co1Mo1K0.05 catalyst. The products were mixtures consisting of linear primary alcohols and hydrocarbons (as well as small amounts of acetaldehyde and DME), but the carbon-number distribution of the alcohols and hydrocarbons did not follow the classical Anderson–Schulz– Flory rule. The S(C2–9-alc.) reached 80.7 C%. The maximum component of the alcohol-product distribution shifted to

Fig. 4. Product distribution of HAS over Co1Mo1K0.05–12%(4.2% Co/ MWCNT) catalyst; reaction conditions: 5.0 MPa, 563 K, V(H2)/V(CO)/ V(N2) = 60/30/10, GHSV = 8000 mL/(h g), 18 h.

91

Fig. 5. Effect of CO2 content in the feed gas on the reactivity of HAS over Co1Mo1K0.05–12%(4.2% Co/MWCNT) catalyst; reaction condition: 5.0 MPa, V(H2)/V(CO)/V(CO2)/V(N2) = 60/30/x/(10 x), GHSV = 10,000 mL/(h g), 18 h.

C7-alcohol (HepOH) from C8-alcohol (OcOH) for the simple MWCNT-promoted counterpart [40]. The S(C7-alc.) reached 17.4 C%. Unlike the cases of the HAS systems reported previously [7,12,13] and also including the simple MWCNT-promoted Co1Mo1K0.05 catalyst developed in our group [40], the presence

Fig. 6. Reaction activity (a) and product distribution (at 593 K) (b) of HAS over Co1Mo1K0.05–12%(4.2% Co/MWCNT) catalyst fed with CO2-containing feed gas V(H2)/V(CO)/V(CO2)/V(N2) = 60/30/5/5, at 5.0 MPa and GHSV = 10,000 mL/(h g), 18 h.

92

X.-M. Wu et al. / Applied Catalysis A: General 340 (2008) 87–97

of a proper amount of CO2 in the feed gas over the (4.2% Co/ MWCNT)-promoted Co1Mo1K0.05 catalyst under relatively high-reaction temperature (583 K) was found to be highly beneficial, instead of harmful, to hydrogenation–conversion of CO and selective formation of C2–9-alcohols. Such results are shown in Fig. 5. Under the optimal reaction temperature (593 K) and on the stream of feed syngas with the optimal CO2 content (5%), the Y(total alc.) reached 26.9%, which was 1.31 times as high as the corresponding value (20.5%) on the stream of CO2-free feed gas. Fig. 6 shows the reactivity of HAS over the Co1Mo1K0.05– 12%(4.2% Co/MWCNT) catalyst fed with the CO2-containing feed gas. Increasing the reaction temperature from 563 to 593 K while simultaneously adding an appropriate amount (5%) of CO2 to the feed gas led to an increase by 63% in X(CO) (34.5% vs. 21.1%) and a marked improvement in Y(total alc.) (26.9% vs. 17.9%) (Fig. 6(a) vs. Fig. 3); at the same time, the maximum component of the alcohol-product distribution shifted to C5alcohol (PeOH) from C7-alcohol (HepOH) (Fig. 6(b) vs. Fig. 4). The S(C5-alc.) reached 15.5 C%. Fig. 7 shows the operation stability of the Co1Mo1K0.05– 12%(4.2% Co/MWCNT) catalyst for HAS lasting 200 h under the two types of reaction conditions with the CO2-free

Fig. 7. Operation stability of HAS lasting 200 h over Co1Mo1K0.05–12%(4.2% Co/MWCNT) catalyst under reaction conditions: (a) 5.0 MPa, 563 K, V(H2)/ V(CO)/V(N2) = 60/30/10 and GHSV = 8000 mL/(h g); (b) 5.0 MPa, 593 K, V(H2)/V(CO)/V(CO2)/V(N2) = 60/30/5/5 and GHSV = 10,000 mL/(h g).

or CO2-containing feed gas. After about 24 h of ‘‘running in’’ stage of the reaction, the catalyst attained a stable operating state, with no obvious deactivation observed after the reaction for 200 h. Table 1 lists the selectivity and STY of HAS over the (4.2% Co/MWCNT)-promoted Co–Mo–K catalyst and the related reference systems. Under the reaction conditions of 563 K and CO2-free feed gas, the STY of the C2–9-alcohols over the Co1Mo1K0.05–12%(4.2% Co/MWCNT) catalyst reached 294.3 mg/(h g). This value was 1.54 and 1.09 times that (191.3 and 269.0 mg/(h g)) of the MWCNT-free host catalyst and the simple MWCNT-promoted counterpart, respectively, under the same reaction conditions. Moreover, under the reaction conditions of 593 K and 5% CO2 in feed gas, the STY of the C2–9-alcohols over the Co1Mo1K0.05–12%(4.2% Co/ MWCNT) catalyst reached 628.3 mg/(h g). This value was 3.28 and 2.33 times that (191 and 269 mg/(g h)) of the MWCNT-free host catalyst and the simple MWCNT-promoted counterpart, respectively, under the respective optimal reaction conditions. In the products of HAS under the above two types of reaction conditions, the C2–9-alc./MeOH ratio was 18.7 (C-based selectivity ratio) for the former (Fig. 4) and 44.8 for the latter (Fig. 6(b)). Such ratios mean that there is a good prospect for such mixed alcohol products to be used as gasoline additives. Table 1 also lists the result of ‘‘blank’’ catalytic measurement with 4.2% Co/MWCNT, from which it can be seen that the Co-decorated MWCNTs displayed fairly high activity for hydrogenation–conversion of CO under the reaction conditions for the HAS, with C1–8-hydrocarbons (fed with the CO2-free feed gas) or C1–6-hydrocarbons (fed with the CO2-containing feed gas) as the predominant product. Such results are consistent with the result reported by Bezemer et al. [41], suggesting that the Co-decorated MWCNTs material may be a good Fischer–Tropsch synthesis catalyst, yet in the present work when used as additive, it has made little direct contribution to the hydrogenation–conversion of syngas, most probably due to its low content and to being covered with the Co–Mo host components. The apparent activation energy (Ea) of HAS reaction was measured under the reaction conditions with mass transfer limitation ruled out, and the results are shown in Fig. 8. Over the three catalysts of Co1Mo1K0.05–12%(4.2% Co/MWCNT), Co1Mo1K0.05–12%MWCNT and Co1Mo1K0.05, the observed Ea of HAS reaction was 92.7, 93.5 and 95.3 kJ/mol, respectively, under the reaction conditions of 543–583 K and CO2-free feed gas, and 61.1, 61.0 and 62.7 kJ/mol, respectively, under the reaction conditions of 573–613 K and 5% CO2 in feed gas. The Ea values within the each contrast group were fairly close to each other, indicating that appropriate incorporation of a minor amount of either simple MWCNTs or Co-decorated MWCNTs into the Co1Mo1K0.05 host catalyst did not cause an marked change in the Ea for the HAS reaction. We can conclude that the addition of a minor amount of the MWCNTs or the Co-decorated MWCNTs to the Co1Mo1K0.05 did not alter the major reaction pathway of hydrogenation–conversion of synthesis gas.

X.-M. Wu et al. / Applied Catalysis A: General 340 (2008) 87–97

93

Table 1 Selectivity and STY of HAS over the (4.2% Co/MWCNT)-promoted Co–Mo–K catalyst and the related reference systems Catalyst

Co1Mo1K0.05–12%(4.2% Co/MWCNT)

4.2% Co/MWCNT a b

Selectivity of hydrogenation products (C%)

STY (mg/(h g))

ROH + HC

Total alc.

C2–9 alc.

Total HC

Total alc.

a

a

a

34.5

Co1Mo1K0.05–12% MWCNT [40] Co1Mo1K0.05 [40] Sample for blank catalytic test

Conversion of CO (%) to

a

CO2 22.5

a

77.9

76.2

22.1

655.2

a

C2–9-alc. 628.3

Total HC

a

161.9a

21.1b 18.1b

11.4b 8.7 b

85.0b 85.0b

80.7b 82.6b

14.6b 14.2b

331.1b 285.9b

294.3b 269.0b

88.2b 42.9b

13.0b

8.9 b

82.5b

81.6b

16.3b

196.7b

191.3b

35.0b

Conversion of CO (%) to

Selectivity of hydrogenation products (C%)

ROH + HC

CO2

Total alc.

Total HC

MeOH

EtOH

PrOH

BuOH

C1–6-HC

C7–8-HC

16.8a 15.0b

6.2 a 4.9 b

20.5a 19.5b

79.5a 80.5b

10.5a 9.0 b

7.5 a 5.3 b

2.5 a 2.7 b

–a 2.5b

79.5a 48.5b

–a 32.0b

Reaction conditions: 5.0 MPa, 593 K, V(H2)/V(CO)/V(CO2)/V(N2) = 60/30/5/5 and GHSV = 10,000 mL/(h g). Reaction conditions: 5.0 MPa, 563 K, V(H2)/V(CO)/V(N2) = 60/30/10 and GHSV = 8000 mL/(h g).

3.3. Post-analysis of the tested catalysts by XRD and XPS The XRD post-analysis of the tested catalysts showed that in the position and shape of XRD features, there were few marked differences among the XRD patterns taken on the three catalysts tested for HAS for 18 h under the reaction condition: (1) CO2-free feed gas at 563 K or (2) CO2-containing feed gas at 593 K, except the feature at 2u = 26.18 for the MWCNTcontaining systems, which was due to the diffraction of (0 0 2) plane of graphite-like tube-wall of the MWCNTs [39] (see Fig. 9). In these tested catalysts, the Co–Mo components existed mainly in the forms of CoO and MoO3 (with the observed XRD features at 2u = 37.08/43.18/62.08 and 35.88, respectively) [42], and the presence of crystallite phases of CoMoO4, K2MoO4 and K4MoO5 (with the weak, even ambiguous, XRD features at 2u = 28.58/32.58, 25.98 and 14.08, respectively) could not be excluded [42]. Moreover, the content of metal Co0x phase was under XRD-detection limit,

even for the system that included metal Co-decorated MWCNTs. XPS measurements revealed that certain differences existed among the above catalysts in the valence-states or microenvironments of their surface Co and Mo species. Fig. 10(A) shows the Co(2p)-XPS spectra of the tested catalysts. The observed Co(2p3/2) and Co(2p1/2) XPS peaks appeared at 781.0 and 797 eV, respectively, with their area–intensity ratio of I(781.0) to I(797.0) being approximately 2. These values are characteristics of Co-species with positive oxidation valence. With reference to Ref. [43] and through computerfitting, it could be found that each of those Co(2p)-XPS spectra involved the contribution from three kinds of surface Con+species (2  n  3) with different micro-environments: CoO(OH)/Co3O4, CoMoOx(3  x < 4) and CoMoO4 in varying degrees (see Fig. 10(A) and Table 2). The molar percentage of Con+-species in the form of CoO(OH)/Co3O4 in the total Co amount at the surface of tested Co1Mo1K0.05–12%(4.2% Co/

Fig. 8. Arrhenius plots of HAS over the catalysts: (a) Co1Mo1K0.05–12%(4.2% Co/MWCNT); (b) Co1Mo1K0.05–12%MWCNT; (c) Co1Mo1K0.05; taken under the reaction conditions: (1) 2.0 MPa, 543–583 K, V(H2)/V(CO)/V(N2) = 60/30/ 10, GHSV = 10,000 mL/(h g) or (2) 2.0 MPa, 573–613 K, V(H2)/V(CO)/ V(CO2)/V(N2) = 60/30/5/5, GHSV = 20,000 mL/(h g), 18 h.

Fig. 9. XRD patterns of the tested catalysts: (a) Co1Mo1K0.05–12%(4.2% Co/ MWCNT); (b) Co1Mo1K0.05–12%MWCNT; (c) Co1Mo1K0.05. Reaction condition: (1) fed with CO2-free feed gas at 563 K and/or (2) fed with CO2containing feed gas at 593 K, 18 h.

94

X.-M. Wu et al. / Applied Catalysis A: General 340 (2008) 87–97

Fig. 10. XPS spectra of Co(2p) (A) and Mo(3d) (B) of the tested catalysts: (a) Co1Mo1K0.05–12%(4.2% Co/MWCNT); (b) Co1Mo1K0.05–12%MWCNT; (c) Co1Mo1K0.05. Reaction condition: (1) fed with CO2-free feed gas at 563 K and/ or (2) fed with CO2-containing feed gas at 593 K, 18 h.

MWCNTs) catalyst fed with CO2-free feed gas at 563 K reached 38.1 mol%, being 1.16 and 1.33 times the corresponding value (32.9 and 28.6 mol%) of the simple MWCNTpromoted counterpart and the MWCNT-free host system. Properly raising the reaction temperature (from 563 to 593 K) while simultaneously doping an appropriate amount (5%) of

CO2 to the feed gas led to a further increase of concentration of the surface CoO(OH)/Co3O4 species to 39.9 mol%. It followed that the ratio of molar percentage of Con+-species in the form of CoO(OH)/Co3O4 at the surface of the three tested catalysts under the reaction conditions (1) and/or (2) was: (a 2)/ (a 1)/(b 1)/(c 1) = 39.9/38.1/32.9/28.6. This sequence was in line with the observed sequence of HAS reaction activity over those catalysts. This result shows that there was some correlation between such a kind of surface Con+-species (CoO(OH)/Co3O4) and the formation of higher alcohols. The higher concentration of the surface CoO(OH) species was conducive to increasing the probability of chain termination to form oxygenated products. Mo(3d)-XPS spectra of the tested catalysts are shown in Fig. 10(B). Slight differences in the position and shape as well as relative intensity of the features associated with the surface Mo-species existed among the three catalysts. Using Ref. [44] and assuming Eb(Mo 3d3/2)–Eb(Mo 3d5/2) = 3.1 eV (B.E.) and I(Mo 3d5/2)/I(Mo 3d3/2) (intensity ratio of the peaks) = 1.5 for each kind of the Mon+ species in the same valence-state, analysis and fitting of these Mo(3d)-XPS spectra were carried out. The results (Fig. 10(B) and Table 2) showed that, under the reactive atmosphere of HAS, most of the surface Mo6+ species were reduced to lower valence: the major portion to Mo5+ and the minor one to Mo4+. This was analogous to the case of coexistence of Mo4+ with Mo5+ (major) and Mo6+ (minor) in the related systems [45]. The obtained ratio of molar percentage of Mo4+-species in the total Mo amount at the surface of the three tested catalysts under the reaction condition (1) and/or (2) was: (a 2)/(a 1)/(b 1)/(c 1) = 23.9/17.4/17.1/16.2. This sequence was again in line with the sequence of HAS reactivity over those catalyst samples. Catalysts containing Mo, Group VIII metals and alkali have been extensively investigated by Fujimoto and Oba [8], Inoue et al. [9], Tatsumi et al. [10] and Sun et al. [11]. They observed that the activity of the catalysts was mainly associated with the presence of Mo, and that lower oxidation state of Mo was more active in CO hydrogenation [46,47], and that cobalt was a necessary promoter for high STY of higher alcohol on reduced K–Mo-based catalysts. The results of the current XPS study show that CoO(OH)/Co3O4 and Mo4+ were the two kinds of catalytically active species closely associated with the alcohol generation, and that one of the additional effects of Co-decorated

Table 2 XPS binding energy and relative content of the Co and Mo species with different valence-states or micro-environments at the surface of the tested catalysts Catalyst

Co1Mo1K0.05–12%(4.2% Co/MWCNT) Co1Mo1K0.05–12%MWCNT Co1Mo1K0.05

Reaction condition

Relative content (mol%) CoO (OH)/Co3O4 with Con+-2p3/2  780.0 eV (B.E.)

CoMoOx (3  x < 4) with Con+-2p3/2  781.0 eV (B.E.)

CoMoO4 with Con+-2p3/2  782.5 eV (B.E.)

Mo4+ with Mo4+-3d5/2  229.8 eV (B.E.)

Mo5+ with Mo5+-3d5/2  230.9 eV (B.E.)

Mo6+ with Mo6+-3d5/2  233.8 eV (B.E.)

(2)

39.9

38.9

21.2

23.9

56.5

19.6

(1) (1) (1)

38.1 32.9 28.6

40.0 44.3 49.4

21.9 22.8 22.0

17.4 17.1 16.2

59.0 56.1 52.7

23.6 26.8 31.1

Reaction conditions: (1) 5.0 MPa, 563 K, V(H2)/V(CO)/V(N2) = 60/30/10 and GHSV = 8000 mL/(h g); (2) 5.0 MPa, 593 K, V(H2)/V(CO)/V(CO2)/V(N2) = 60/30/5/ 5 and GHSV = 10,000 mL/(h g).

X.-M. Wu et al. / Applied Catalysis A: General 340 (2008) 87–97

MWCNTs was in promoting the increase of surface concentration of the two kinds of catalytically active species. The addition of 5% CO2 in the feed gas at properly elevated reaction temperature (593 K) was found to be able to further enhance the concentration of those active Mo and Co species. The roles that CO2 played may also include preventing the deep reduction of the active Con+ and Mon+ on the catalyst surface, and maintaining their activities towards yielding more alcohols and less alkanes (see Fig. 3 vs. Fig. 6(A) and (B)). This scenario was similar to the Cu–ZnO–Al2O3 catalyst used in methanol synthesis, where a certain amount of CO2 in the feed gas could prevent the deep reduction of Cun+ to Cu0 [48].

Table 3 Relative area–intensity of H2-TPD profiles of the prereduced catalysts and the related materials Sample

Relative area–intensitya (293–723 K)

Contrast group (1)

MWCNTs 4.2% Co/MWCNT

67 100

Contrast group (2)

Mo1Co1K0.05 Mo1Co1K0.05– 12%MWCNT Mo1Co1K0.05–12% (4.2% Co/MWCNT)

a

3.4. H2-TPD test of hydrogen-prereduced catalysts and nature of promoter action by MWCNT-based additives Recently, there has been increasing interest in the use of nanostructured carbon materials (especially CNTs, carbon nanofibers and mechanically milled graphite) as hydrogen sorbents. Ishikawa et al. [49] demonstrated that graphitized carbon black surfaces were capable of rapidly equilibrating H2/ D2 mixtures. A dissociation rate of 2.5  1017 molecules/ (s (m2-ASA)) was measured at ambient temperatures and pressures. The active surface area (ASA) was described in terms of atoms located at edge positions on the graphite basal plane and was determined from the amount of oxygen able to chemisorb at these sites, regardless of the nature of the carbon material under investigation. Our previous H2-TPD investigation [50] showed that hydrogen adsorption on the MWCNTs can occur at ambient temperatures and pressures, and that the desorbed product was almost exclusively H2 at temperatures lower than 723 K, while it included CH4, C2H4 and C2H2, in addition to H2, at temperatures of 773 K and above. This implies that H2 adsorption on the MWCNTs may be in two forms: associative (molecular state) and dissociative (atomic state), as had been evidenced in our Raman spectroscopic study of H2/MWCNTs adsorption system [39]. Fig. 11(a) and (b) showed the TPD profiles of H2 adsorbed at 433 K followed by cooling down to room temperature on the

Fig. 11. TPD profiles of H2 adsorption on: (a) MWCNTs; (b) 4.2% Co/ MWCNT; (c) Co1Mo1K0.05; (d) Co1Mo1K0.05–12%MWCNT; (e) Co1Mo1K0.05–12%(4.2% Co/MWCNT).

95

42 68 100

With the area–intensity of the strongest profile in each group as 100.

MWCNTs and 4.2% Co/MWCNTs, respectively. The decoration of metallic cobalt to the MWCNTs resulted in a significant increase in its capacity of adsorbing hydrogen, especially the dissociatively adsorbing hydrogen H(a). Considering that the hydrogenation of some surface carbon by H-adspecies (which would lead to consumption of part of H-adspecies and formation of C1–2-hydrocarbons) could occur at temperatures of 773 K and above [39,50], we estimated the relative area– intensity of these H2-TPD profiles in the region of 293–723 K. The obtained ratio was: SMWCNTs/S4.2% Co/MWCNT = 67/ 100 (see Table 3), indicating that the decoration of the metallic cobalt to the MWCNTs led to a 49% increase in its hydrogenadsorbing capacity in the temperature range of 293–723 K. Fig. 11(c)–(e) shows the H2-TPD profiles of H2 adsorbed at 433 K followed by cooling down to room temperature on the H2-prereduced catalysts: Co1Mo1K0.05, Co1Mo1K0.05– 12%MWCNT and Co1Mo1K0.05–12%(4.2% Co/MWCNT), respectively. Overall, each H2-TPD profile contained a lower-temperature peak (peak I, 293–523 K) and a highertemperature peak (peak II, 523–773 K). Peak I resulted from the desorption of weakly adsorbed H-species, most probably molecularly adsorbed hydrogen, and peak II was attributed to the desorption of strongly adsorbed H-species, perhaps dissociatively chemisorbed hydrogen. The relative area– intensity ratio of these H2-TPD profiles was estimated to be S(c)/S(d)/S(e) = 42/68/100 (see Fig. 11(c)–(e) and Table 3). This sequence was in line with the observed sequence of HAS reaction activity over those catalysts. Based on the above results, we concluded that a considerably greater amount of reversibly adsorbed hydrogen-species would be present on the functioning catalyst promoted by Co-decorated MWCNTs, which would generate a surface micro-environment with higher stationary-state concentration of hydrogen-adspecies. Those active H-adspecies could be easily transferred to CoiMojKk active sites via the MWCNT-promoted hydrogen spillover, thus increasing the rate of the surface hydrogenation reactions. This was very similar to the cases in synthesis of methanol and higher alcohols from syngas over the MWCNT-promoted Cu–ZnO– Al2O3 [23,24] and Co–Cu [25] catalysts, respectively. Moreover, the high stationary-state concentration of Hadspecies at the surface of catalyst would considerably

96

X.-M. Wu et al. / Applied Catalysis A: General 340 (2008) 87–97

inhibit WGS side-reactions, which would also contribute to an increase in the main product yields. 4. Conclusion (1) The Co-decorated MWCNTs could serve as an excellent catalyst-promoter. The Co-decorated MWCNT-promoted Co–Mo–K catalyst achieved highly effective and selective formation of C2+-alcohols from syngas. (2) Incorporation of a proper amount of the Co-decorated MWCNTs into the Co–Mo–K catalyst caused little change in Ea for HAS reaction, but led to an increase of surface concentration of the two kinds of catalytically active species (CoO(OH)/Co3O4) and Mo4+, both closely associated with the alcohol generation. Raising the reaction temperature appropriately and doping a certain amount of CO2 in the feed gas could further enhance the surface concentrations of those active Mo and Co species. The doped CO2 also played roles in preventing the deep reduction of the active Co and Mo species. (3) The decoration of the metallic cobalt to the MWCNTs led to a considerable increase in its hydrogen-adsorbing capacity, which would be in favor of generating a surface microenvironment with higher stationary-state concentration of hydrogen-adspecies at the surface of the functioning catalyst, thus increasing the rate of the surface hydrogenation reactions. High concentration of hydrogen-adspecies on the catalyst surface would, through synergistic action with the CO2 in the feed gas, also inhibit the WGS sidereactions. For better understanding of the nature of the promoting action by Co-decorated MWCNTs, further studies, especially in situ characterization of reaction intermediates under the HAS reaction conditions, would be highly desired. Acknowledgments The authors are grateful for the financial supports from National Natural Science Foundation (Project Nos. 20473063 & 20590364) and National Basic Research (‘‘973’’) Project (Project No. 2005CB221403) of China. References [1] R.R. Chianelli, J.E. Lyons, G.A. Mills, Catal. Today 22 (1994) 361–396. [2] K.J. Smith, R.B. Anderson, J. Catal. 85 (1984) 428–436. [3] P. Courty, D. Durand, E. Freund, A. Sugier, J. Mol. Catal. 17 (1982) 241– 254. [4] A. Razzaghi, J.P. Hindermann, A. Kiennemann, Appl. Catal. 13 (1984) 193–210. [5] A. Kiennemann, S. Boujana, C. Diagne, P. Chaumette, Stud. Surf. Sci. Catal. 75 (1993) 1479–1492. [6] J.G. Santiesteban, C.E. Bogdan, R.G. Herman, K. Klier, in: M.J. Phillips, M. Ternan (Eds.), Proceedings of the Ninth International Congress on Catalysis, vol. 2, 1988, pp. 561–568. [7] R.G. Herman, Stud. Surf. Sci. Catal. 64 (1991) 266–349. [8] K. Fujimoto, T. Oba, Appl. Catal. 13 (1985) 289–293.

[9] M. Inoue, T. Miyake, Y. Takegami, T. Inui, Appl. Catal. 29 (1987) 285– 294. [10] T. Tatsumi, A. Muramatsu, T. Fukunaga, H. Tominaga, in: M.J. Phillips, M. Ternan (Eds.), Proceedings of the Ninth International Congress on Catalysis, vol. 2, 1988, pp. 618–625. [11] Z.H. Sun, J. Bao, Y.L. Fu, G.Z. Bian, Chin. J. Catal. 24 (2003) 826–830. [12] P. Forzatti, E. Tronconi, I. Pasquon, Catal. Rev. Sci. Eng. 33 (1991) 109– 168. [13] A.B. Stiles, F. Chen, J.B. Harrison, X. Hu, D.A. Storm, H.X. Yang, Ind. Eng. Chem. Res. 30 (1991) 811–821. [14] J.C. Slaa, J.G. van Ommen, J.R.H. Ross, Catal. Today 15 (1992) 129–148. [15] S. Iijima, Nature 354 (1991) 56–58. [16] K.P. de Jong, J.W. Geus, Catal. Rev. Sci. Eng. 42 (2000) 481–510. [17] P. Serp, M. Corrias, P. Kalck, Appl. Catal. A: Gen. 253 (2003) 337–358. [18] H.-B. Zhang, G.-D. Lin, Y.-Z. Yuan, Curr. Topics Catal. 4 (2005) 1–21. [19] J.M. Planeix, N. Coustel, B. Coq, V. Brotons, P.S. Kumbhar, R. Dutartre, P. Geneste, P. Bernier, P.M. Ajiayan, J. Am. Chem. Soc. 116 (1994) 7935– 7936. [20] Y. Zhang, H.-B. Zhang, G.-D. Lin, P. Chen, Y.-Z. Yuan, K.-R. Tsai, Appl. Catal. A: Gen. 187 (1999) 213–224. [21] H.-B. Chen, J.-D. Lin, Y. Cai, X.-Y. Wang, J. Yi, J. Wang, G. Wei, Y.-Z. Lin, D.-W. Liao, Appl. Surf. Sci. 180 (2001) 328–335. [22] E. van Steen, F.F. Prinsloo, Catal. Today 71 (2002) 327–334. [23] H.-B. Zhang, X. Dong, G.-D. Lin, Y.-Z. Yuan, P. Zhang, K.-R. Tsai, in: C.J. Liu, R.G. Mallinson, M. Aresta (Eds.), Proceedings of the ACS Symposium Series 852, American Chemical Society, Washington, DC, (2003), pp. 195–209. [24] X. Dong, H.-B. Zhang, G.-D. Lin, Y.-Z. Yuan, K.-R. Tsai, Catal. Lett. 85 (2003) 237–246. [25] H.-B. Zhang, X. Dong, G.-D. Lin, X.-L. Liang, H.-Y. Li, Chem. Commun. (2005) 5094–5096. [26] Z.-J. Liu, Z.-D. Xu, Z.-Y. Yuan, D.-Y. Lu, W.-X. Chen, W.-Z. Zhou, Catal. Lett. 72 (2001) 203–206. [27] C.-B. Li, W.-X. Pan, W.-K. Wong, J.-L. Li, X.-Q. Qiu, X.-P. Chen, J. Mol. Catal. A: Chem. 193 (2003) 71–75. [28] W. Li, C. Liang, J. Qiu, W. Zhou, H. Han, Z. Wei, G. Sun, Q. Xin, Carbon 40 (2002) 791–794. [29] X.-M. Ma, G.-D. Lin, H.-B. Zhang, Catal. Lett. 111 (2006) 141–151. [30] E. Dujardin, T.W. Ebbesen, H. Hiura, K. Tanigaki, Science 265 (1994) 1850–1852. [31] T.W. Ebbesen, J. Phys. Chem. Solids 57 (1996) 951–955. [32] L.M. Ang, T.S.A. Hor, G.Q. Xu, C.H. Tung, S.P. Zhao, J.L.S. Wang, Carbon 38 (2000) 363–372. [33] M. Figlarz, F. Fievet, J.P. Lagier, French Patent 82 21483; Europe Patent 011 3281; U.S. Patent 4 539 041; Finland Patent 74416. [34] F. Fievet, J.P. Lagier, B. Blin, B. Beaudoin, M. Figlarz, Solid State Ionics 32/33 (Pt 1) (1989) 198–205. [35] C. Ducamp-Sanguesa, R. Herrera-Urbina, M. Figlarz, Solid State Ionics 63/65 (1993) 25–30. [36] P. Chen, H.-B. Zhang, G.-D. Lin, Q. Hong, K.-R. Tsai, Carbon 35 (1997) 1495–1501. [37] P. Chen, H.-B. Zhang, G.-D. Lin, K.-R. Tsai, Chem. J. Chin. Univ. 19 (1998) 765–769. [38] L.K. Kurihara, G.M. Chow, P.E. Schoen, NanoStruct. Mater. 5 (1995) 607–613. [39] H.-B. Zhang, G.-D. Lin, Z.-H. Zhou, X. Dong, T. Chen, Carbon 40 (2002) 2429–2436. [40] H.-B. Zhang, X. Dong, G.-D. Lin, X.-L. Liang, Y.-Y. Guo, P. Zhang, in: Proceedings of the 232nd ACS National Meetings, Paper No. 244 (Oral Presentation), Fuel Chemistry Division, San Francisco, 2006, preprints. [41] G.L. Bezemer, J.H. Bitter, H.P.C.E. Kuipers, H. Oosterbeek, J.E. Holewijn, X.D. Xu, F. Kapteijn, A.J. van Dillen, K.P. de Jong, J. Am. Chem. Soc. 128 (2006) 3956–3964. [42] XRD Data Bank Attached to X’Pert PRO X-ray Diffractometer, PANalytical, The Netherlands. [43] J.F. Moulder, W.F. Stickle, P.E. Sobol, K.D. Bomben, Handbook of X-ray Photoelectron Spectroscopy—A Reference Book of Standard Spectra for

X.-M. Wu et al. / Applied Catalysis A: General 340 (2008) 87–97 Identification and Interpretation of XPS Data, Physical Electronics Inc., Eden Prairie, MN, USA, 1995. [44] J. Abart, E. Delgado, G. Ertl, H. Jeziorowshi, H. Kno¨zinger, N. Thiele, X.Z. Wang, E. Taglauer, Appl. Catal. 2 (1982) 155–176. [45] A.M. Venezia, Catal. Today 77 (2003) 359–370. [46] A. Muramatsu, T. Tatsumi, H. Tominaga, J. Phys. Chem. 96 (1992) 1334– 1340.

97

[47] M. Saito, R.B. Anderson, J. Catal. 63 (1980) 438–446. [48] K. Klier, Adv. Catal. 31 (1982) 243–313. [49] Y. Ishikawa, L.G. Austin, D.E. Brown, P.L. Walker Jr., in: P.L. Walker, Jr., P.A. Thrower (Eds.), American Carbon Society, vol. 12, Marcel Dekker, New York, 1975, p. 39. [50] Z.-H. Zhou, X.-M. Wu, Y. Wang, G.-D. Lin, H.-B. Zhang, Acta Phys. Chim. Sin. 18 (2002) 692–698.