Ni-modified Mo2C catalysts for methane dry reforming

Applied Catalysis A: General 431–432 (2012) 164–170

Contents lists available at SciVerse ScienceDirect

Applied Catalysis A: General journal homepage: www.elsevier.com/locate/apcata

Ni-modified Mo2 C catalysts for methane dry reforming Chuan Shi a,b,∗, Anjie Zhang a,b, Xiaosong Li b, Shaohua Zhang a,b, Aimin Zhu b, Yufei Ma a,b, Chaktong Au c,∗∗ a b c

Key Laboratory of Industrial Ecology and Environmental Engineering, Dalian University of Technology, Dalian, China Laboratory of Plasma Physical Chemistry, Dalian University of Technology, Dalian, China Chemistry Department, Hong Kong Baptist University, Hong Kong, China

a r t i c l e

i n f o

Article history: Received 30 August 2011 Received in revised form 29 March 2012 Accepted 27 April 2012 Available online 4 May 2012 Keywords: Methane Carbon dioxide Molybdenum carbide Nickel Dry reforming

a b s t r a c t Dry reforming of methane with CO2 (DRM) was studied over Ni–Mo2 C catalysts with Ni/Mo molar ratios of 1/3, 1/2, and 1/1 (denoted as Ni–Mo2 C (1/3), Ni–Mo2 C (1/2), and Ni–Mo2 C (1/1), respectively) aiming to investigate the catalytic roles of Ni and the carbide. The results of XRD and XPS characterizations indicated that the carbonization process was promoted by the presence of Ni. The CH4 -TPR and CO2 -TPO over the fresh samples proved that CH4 dissociation was greatly enhanced by Ni. The Ni–Mo2 C (1/2) catalyst showed the best catalytic activity and stability for CH4 /CO2 (1/1) dry reforming. Above 80% of CH4 and CO2 conversions were maintained at 800 ◦ C during a test run of 20 h at W/F = 0.3 g s cm−3 . Characterizations of the spent samples revealed that the deactivation of Ni–Mo2 C (1/1) was due to coke formation whereas that of Ni–Mo2 C (1/3) was due to bulk oxidation of Mo2 C into MoO2 . Only at a Ni/Mo molar ratio of 1/2, a catalytic oxidation–reduction cycle could be established. It was suggested that Ni–Mo2 C was a typical bi-functional catalyst. In CH4 /CO2 dry reforming, the dissociation of CH4 was catalyzed by Ni, while the activation of CO2 took place on Mo2 C. By regulating the molar ratio of Ni and Mo2 C, a catalytic redox cycle could be established. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Catalytic dry reforming of methane with CO2 (DRM) has become an interesting alternative for syngas production due to the fact that the greenhouse gases CO2 and CH4 can be utilized through the reaction. The DRM process produces syngas with a H2 /CO ratio of 1 that is suitable for Fischer–Tropsch (F–T) and methanol synthesis [1–3]. The DRM reaction has been performed over a wide range of catalysts, including noble as well as non-noble metals [4,5]. Among them, the nickel-based catalysts are preferred due to the inherent availability and low cost of nickel. However, the major drawback is coking, in particular when the reaction was conducted over nickelbased catalysts [6–9]. It is hence highly desirable to develop a new catalyst that is inexpensive, active and stable for the DRM reaction. In recent years, transition metal carbides have attracted much attention, owing to the fact that the metal carbides show catalytic properties similar to those of noble metals in a variety of

∗ Corresponding author at: Key Laboratory of Industrial Ecology and Environmental Engineering (MOE), Dalian University of Technology, Dalian, China. Tel.: +86 411 84986083. ∗∗ Corresponding author at: Department of Chemistry, Hong Kong, China. Tel.: +852 34117067. E-mail addresses: [email protected] (C. Shi), [email protected] (C. Au). 0926-860X/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.apcata.2012.04.035

reactions such as methane reforming, hydrogenation reaction and hydrocarbon isomerization [10–12]. Previous studies indicated that unsupported ␤-Mo2 C was catalytically active for the DRM process, showing high thermal stability and good resistance toward coking [13,14]. According to Claridge et al., stable activity (72 h) at 850 ◦ C could be achieved in CH4 /CO2 reforming over ␤-Mo2 C at high pressure (8 bar) [15]. Cubic ␣-MoC1−x was also tested for DRM, although the conversion levels over ␣-MoC1−x were higher than those over ␤-Mo2 C, the ␣-MoC1−x phase was transformed to ␤Mo2 C phase during the reaction [16]. On the other hand, bimetallic carbides were used as catalysts for DRM [17–19]. Stable catalytic activity (80 h) was observed by Shao et al. over bimetallic Co–W carbides at 850 ◦ C and 3.4 bar [17]. Despite the positive results and high thermal stability of these low-cost materials, the carbide catalysts deactivated rapidly due to oxidation by CO2 at atmospheric pressure as pointed out by Thomson [14]. In our previous studies, we demonstrated that the Ni modified Mo2 C catalyst performed well in CH4 /CO2 reforming [20,21]. The advantage of Ni–Mo2 C over hexagonal Mo2 C or other traditional carbide catalysts is its stable performance at atmospheric pressure. Herein, we report the results of our recent study on the catalytic performance of Ni–Mo2 C, with attention given to the synergistic effect of Ni and Mo2 C. The different catalytic behaviors of Ni–Mo2 C catalysts with different Ni/Mo ratios allowed us to discriminate the catalytic roles of Ni and carbide for DRM.

C. Shi et al. / Applied Catalysis A: General 431–432 (2012) 164–170

165

2. Experimental

Mo2C

2.1. Catalyst preparation We prepared the NiMoOx precursor by stirring an aqueous solution of (NH4 )6 Mo7 O24 ·4H2 O and Ni(NiO3 )2 ·6H2 O (Ni/Mo molar ratio = 1/3, 1/2, and 1/3) at 80 ◦ C for 4 h. The as-obtained NiMoOx was filtered out and dried at 110 ◦ C for 12 h and calcined at 550 ◦ C for 4 h. Then NiMoOx was carburized in CH4 /H2 (20 vol% CH4 ) to Ni–Mo2 C following a series of temperature-programmed processes: temperature was raised from room temperature (RT) to 300 ◦ C in a span of 1 h, then from 300 ◦ C to 700 ◦ C at a rate of 1 ◦ C/min, and subsequently kept at 700 ◦ C for 2 h. The as-obtained material was cooled down to RT in flowing CH4 /H2 and passivated in a mixture of 1%O2 /Ar for 12 h. In a similar manner, MoO3 powder purchased from Tianjin Kermel Chemical Reagent was carburized as well. The Ni–Mo2 C bimetallic carbide catalysts with Ni/Mo molar ratios of 1/3, 1/2, and 1/1 are denoted hereinafter as Ni–Mo2 C (1/3), Ni–Mo2 C (1/2), and Ni–Mo2 C (1/1), respectively.

Intensity (a. u.)

Ni

Ni/Mo=1/1

(d)

Ni/Mo=1/2

(c)

Ni/Mo=1/3

(b)

Mo2C 20

30

(a) 40

50

60

70

80

2 Theta(º) Fig. 1. XRD patterns of Mo2 C and Ni–Mo2 C catalysts.

2.2. Catalyst characterization X-ray powder diffraction (XRD) analysis was conducted using a XRD-6000 (Shimadzu, Kyoto, Japan) equipment with Cu K␣ radiation ( = 0.1542 nm), operating at 40 kV and 30 mA and at a scanning rate of 2◦ /min; phase identification was achieved through comparison of XRD patterns to those of “Joint Committee on Powder Diffraction Standards (JCPDS)”. X-ray photoelectron spectroscopy (XPS) analysis was done with a Leybold Max 200 spectrometer using AlKa radiation as the photon source, generated at 15 kV and 20 mA. The pass energy was set at 192 eV for the survey scan and at 48 eV for the narrow scan.

BET surface area and average pore diameter of the catalystsorbent were measured through nitrogen adsorption at liquid-nitrogen temperature (77 K) by a surface area analyzer (NOVA4200). The samples were digested with mixed acids, and their chemical compositions were determined by inductively coupled plasma-atomic emission spectroscopy (ICP-AES). Carbon dioxide temperature-programmed oxidation (CO2 -TPO) studies were performed using a mass spectrometer (OmniStarTM Pfeiffer Vacuum, Germany) interfaced with a computer. With the sample (0.1 g) securely placed in a quartz tubular reactor.

Fig. 2. Deconvolution of Mo3d spectra of Mo2 C and Ni–Mo2 C catalysts.

C. Shi et al. / Applied Catalysis A: General 431–432 (2012) 164–170 1000

ICP-AES Dpore (nm)

Mo weight (%)

Ni weight (%)

Ni/Mo molar ratio

6.8 27.1 9.4 9.6

1.38 3.76 4.23 1.46

– 78.26 60.65 55.57

– 14.98 18.53 36.38

– 0.322 0.514 1.065

8.00E-008

800

600

m/z=44

400

4.00E-008

Temperature (oC)

Ni/Mo=1/3 Fresh

856 oC

Mo2 C Ni–Mo2 C (1/3) Ni–Mo2 C (1/2) Ni–Mo2 C (1/1)

BET SBET (m2 /g)

Intensity (a.u.)

Sample

1.20E-007

950 oC

Table 1 Physical and chemical properties of the as-prepared samples.

675 oC

166

200

m/z=28 0.00E+000

0

20

40

60

80

100

0

Time (min) 9.00E-008

6.00E-008

800

600

m/z=44 400

3.00E-008

2.3. Catalyst evaluation

200

m/z=28

Catalytic performance of the Ni–Mo2 C catalysts with various Ni/Mo ratios was evaluated. The reaction was carried out in a fixedbed micro-reactor at atmospheric pressure. Prior to the reaction, the catalyst was pre-reduced with pure H2 at 500 ◦ C for 2 h. CH4 and CO2 (CH4 :CO2 = 1:1) was allowed to pass through the catalyst bed at a flow rate 30 mL/min (W/F = 0.3 g s cm−3 ). The relative error of our experiment data is less than 3%.

0.00E+000

0

20

40

60

80

100

0

Time (min) 1000

950 oC

800

Fig. 1 shows the XRD patterns of Ni–Mo2 C (1/3), Ni–Mo2 C (1/2), and Ni–Mo2 C (1/1) catalysts. The peaks at 2 of 34.8◦ , 38.4◦ , 39.8◦ , 52.5◦ , 61.9◦ , 69.9◦ , 74.9◦ and 76.1◦ can be attributed to the presence of hexagonal Mo2 C. No peaks corresponding to metallic molybdenum and Ni–Mo alloy were observed over the Ni–Mo2 C samples. Notably, specific peaks of metallic nickel (2 = 44.4◦ and 51.8◦ ) were detected, disclosing the presence of metal species (Ni) on Mo2 C surface. In addition to that, it was observed that with the increase of Ni/Mo molar ratio, there was an increase of Ni peak intensity. As calculated by Scherrer equation based on the Ni diffraction peak at 44.4◦ , the Ni particle size were 17, 19, and 20 nm for Ni–Mo2 C (1/3), Ni–Mo2 C (1/2), and Ni–Mo2 C (1/1), respectively. The BET specific surface areas of Ni–Mo2 C catalysts of different Ni/Mo ratios are presented in Table 1. The BET specific surface area of the Ni–Mo2 C (1/3) sample is 27.1 m2 /g, and the surface areas of Ni–Mo2 C (1/2) and Ni–Mo2 C (1/1) samples are 9.4 and 8.6 m2 /g, respectively. It is clear that with rise of Ni content, there is decrease in BET specific surface areas across the Ni–Mo2 C samples. According to ICP results, the real Ni/Mo ratios are similar to the nominal ones. The Mo 3d spectra of the Ni–Mo2 C catalysts measured by XPS are shown in Fig. 2. The doublet peaks should have a splitting of ∼3.2 eV and a Mo 3d5/2 to Mo 3d3/2 ratio of 3:2. By means of deconvolution, the distribution of molybdenum oxidation states was estimated. As shown in Fig. 2A, there are three molybdenum species. The one with Mo 3d5/2 binding energy (BE) of 228.1 eV is attributed to Mo2+ species involved in Mo C bonding. The other two with Mo 3d5/2

m/z=44

6.00E-008

600

854 oC

Intensity (a.u.)

3. Results and discussion

400 3.00E-008

Temperature ( oC)

689 oC

Ni/Mo=1/1 Fresh

9.00E-008

3.1. Physicochemical properties of the catalysts

Temperature ( oC)

854 oC 950 oC

680 oC

1000

Ni/Mo=1/2 Fresh

Intensity

Temperature-programmed oxidation using CO2 (CO2 -TPO) was carried out by introducing 10%CO2 /Ar with a total flow rate of 100 mL/min into the system while the sample temperature was raised from RT to a desired temperature at a rate of 10 ◦ C/min. The signal intensity of CO (m/e = 28) and CO2 (m/e = 44) were monitored. Methane temperature-programmed surface reduction (CH4 TPSR) studies were performed using a mass spectrometer (OmniStarTM Pfeiffer Vacuum, Germany) interfaced with a computer. With the sample (0.1 g) securely placed in a quartz tubular reactor, CH4 -TPR profile was obtained by heating the sample from RT to 950 ◦ C at a rate of 10 ◦ C/min in a flow of 10% CH4 /Ar (100 mL/min). The signal intensity of CH4 (m/e = 15), H2 (m/e = 2), H2 O (m/e = 18), CO (m/e = 28), and CO2 (m/e = 44) were monitored.

200

m/z=28 0.00E+000

0

20

40

60

80

100

0

Time (min) Fig. 3. CO2 -TPO profiles of Ni–Mo2 C catalysts with various Ni/Mo ratios.

binding energy of 230.7 and 232.0 eV were identified as Mo5+ and Mo6+ , respectively. No peak corresponding to metallic Mo (Mo 3d5/2 BE at 227.3 eV) was detected. It is clear that although molybdenum oxides identified as Mo5+ and Mo6+ were present, a portion of Mo species is in the form of carbides. The detection of molybdenum oxides should be attributed to surface oxides formed during passivation [22]. However, one cannot exclude the possibility that some of the oxides were due to incomplete carburization; in XRD analysis, these oxides could not be detected because they are minute in particle size [23]. In the case of Ni–Mo2 C (1/3), the characteristic peaks of Mo2+ species is much smaller compared to those of Mo5+ and Mo6+ , indicating the lower ratio of Mo2 C to total Mo species (Fig. 2B). In Fig. 2C and D, the peaks ascribed to Mo2+ become much stronger than that ascribed to Mo5+ and Mo6+ , especially in the case of Ni–Mo2 C (1/1). By curve-fitting the Mo 3d profiles, the

C. Shi et al. / Applied Catalysis A: General 431–432 (2012) 164–170 Table 2 Mo 3d5/2 binding energies and Mo2+ /(Mo2+ + Mo5+ + Mo6+ ) ratios of the as-prepared samples. Catalysts

Mo2 C Ni–Mo2 C (1/3) Ni–Mo2 C (1/2) Ni–Mo2 C (1/1)

Mo2+ /(Mo2+ + Mo5+ + Mo6+ )

Mo 3d5/2 (eV) Mo2+ (carbide)

Mo5+ (oxide)

Mo6+ (oxide)

228.1 227.8 228.3 228.4

230.7 230.5 230.7 230.7

232.0 232.2 232.6 232.5

28.6 13.3 27.3 54.2

ratio of surface molybdenum species to total molybdenum species, denoted as Mo2+ /(Mo2+ + Mo5+ + Mo6+ ), were obtained and summarized in Table 2. When the Ni/Mo molar ratio was increased from 1/3 to 1/1, the value of Mo2+ /(Mo2+ + Mo5+ + Mo6+ ) increased from 0.13 to 0.54, indicating that the increasing amount of molybdenum carbides on the surface. This is in accordance with the report that the presence of nickel promotes the caburization process [24]. In addition, the results of XPS analysis disclosed features of nickel interaction with the molybdenum carbide species. Nickel addition may cause a change in the electronic band structure of Mo species. It is interesting to know that with increase of Ni/Mo molar ratio from 1/3 to 1/1, the Mo2+ 3d peaks shift to higher binding energy. The results indicate there is electron transfer from Ni to Mo2 C. 3.2. Surface reactions of CH4 and CO2 over Ni–Mo2 C We studied the surface reaction of CO2 and CH4 over the Nimodified Mo2 C catalysts. Shown in Fig. 3 are the CO2 -TPO profiles of fresh samples. In the case of Ni–Mo2 C (1/3), two peaks are observed, one at 675 ◦ C and the other at 856 ◦ C. In our previous studies, we demonstrated by XRD measurements at fixed temperatures that

167

Table 3 Curve fitting results of CO2 consumption profiles during CO2 -TPO (Fig. 3) over the fresh samples. Sample

Peak area of surface carbon

Peak area of carbides

Ratio of surface carbon to the carbide

Ni–Mo2 C (1/3) Ni–Mo2 C (1/2) Ni–Mo2 C (1/1)

7.00E−7 6.87E−7 6.91E−7

1.16E−6 7.76E−7 5.95E−7

0.60 0.88 1.16

the former CO2 consumption peak should be due to the oxidation of surface carbon while the latter to the bulk oxidation of Mo2 C [20]. Over the samples of Ni–Mo2 C (1/1) and Ni–Mo2 C (1/2), it is observed that the temperature for bulk oxidation of Ni–Mo2 C samples remained unchanged, while the oxidation of surface coke needs higher temperature over samples of higher Ni/Mo ratio. By curve-fitting of the two CO2 consumption peaks, we calculated the ratio of surface carbon to carbides (Table 3). It is obvious that the ratio of surface carbon increased with increasing Ni/Mo ratio from 1/3 to 1/1, indicating that the dissociation of CH4 was enhanced by Ni, resulting in higher surface coke formation during the carbonthermal reduction process. Fig. 4 displays the results of CH4 -TPR obtained over the Ni–Mo2 C samples in a gas stream of 10% CH4 /Ar. The consumption peaks of CH4 observed at lower temperatures (<500 ◦ C) accompanied with the formation of H2 O, CO, CO2 and H2 should be a result of CH4 interaction with the oxygen species that were left behind due to incomplete carburization [20]. CH4 consumption at higher temperature, which was accompanied by obvious formation is ascribed to CH4 dissociation on the catalysts. In the case of Ni–Mo2 C (1/1), the CH4 dissociation reached its maximum at 930 ◦ C. While over the sample with lower Ni/Mo ratios, CH4 dissociation did not reach its maximum till 950 ◦ C. In Fig. 4, we compared the CH4 consumption profiles (m/e = 15) across the three samples. It is clear that the degree of CH4 dissociation over higher Ni/Mo ratio samples is much

Fig. 4. CH4 -TPR profiles of Ni–Mo2 C catalysts with various Ni/Mo ratios.

168

C. Shi et al. / Applied Catalysis A: General 431–432 (2012) 164–170

Fig. 5. Lifetime study of Ni–Mo2 C catalysts in CH4 /CO2 dry reforming.

deeper than that over the samples with lower Ni content. Based on the comparison result, it is obvious that CH4 dissociation was promoted by the presence of Ni. Summarizing the studies for the fresh samples, it is apparent that Ni particle size increased with rise of Ni/Mo ratio across the Ni–Mo2 C samples. The increasing Ni content and Ni particle size promotes the carburization process as indicated by higher ratios of surface molybdenum carbides to total molybdenum species. The enhancement of CH4 dissociation by Ni was also revealed by CO2 TPO and CH4 -TPR. The concentration of free carbon increased with increasing Ni/Mo ratio, indicating the dissociation of CH4 was promoted by Ni. Moreover, from CH4 -TPR over the Ni–Mo2 C samples, it is clear that the temperature for dissociation was lowered with higher Ni content samples. All these results of characterization confirm that the promotion effect of Ni on CH4 dissociation.

3.3. Catalytic performance The performance variation of Ni–Mo2 C (1/3), Ni–Mo2 C (1/2), and Ni–Mo2 C (1/1) with time on stream are shown in Fig. 5. It is obvious that both catalyst activity and stability are dependent on Ni/Mo molar ratio. The Ni–Mo2 C (1/2) catalyst shows high conversion and stability over a run of 20 h. On the other hand, rapid decline of performance was observed over Ni–Mo2 C (1/3) and Ni–Mo2 C (1/1). It is worth pointing out that there is no DRM activity over pure Mo2 C under similar reaction conditions, and it is clear that there is synergistic effect between nickel and molybdenum carbide that facilitates the catalysis of Ni–Mo2 C (1/2).

3.4. Characterizations of used catalysts Oxidation and coking are usually the main reasons for deactivation of carbide and nickel-based catalysts, respectively. To understand the causes of the catalyst deactivation, CO2 -TPO and XRD investigation were performed over the used catalysts. Displayed in Fig. 6A and B are the CO2 -TPO profile and XRD pattern of a used Ni–Mo2 C (1/3) sample. The former shows CO2 consumption starting at around 730 ◦ C while the latter indicates that there is oxidation of Mo2 C to MoO2 during exposure to carbon dioxide and methane. In other words, the cause of Ni–Mo2 C (1/3) deactivation in DRM process was the bulk oxidation of Mo2 C by CO2 . In the case of Ni–Mo2 C (1/1), there are two CO formation peaks, rising at 550 ◦ C and 730 ◦ C as shown in Fig. 7A. The former should be due to the oxidation of surface carbon while the latter to the bulk oxidation of Mo2 C. The XRD pattern of used Ni–Mo2 C (1/1) (Fig. 7B) confirms that there is no bulk oxidation of Ni–Mo2 C (1/1) during the DRM process, suggesting that the reason of deactivation for Ni–Mo2 C (1/1) was coking rather than oxidation. When explaining the effect of Ni/Mo ratio, we must take into account the fact of carbon deposition on nickel and bulk Mo2 C oxidation which are the main reasons for deactivation of Ni catalysts and carbide, respectively. The Ni–Mo2 C (1/3) catalyst was oxidized to MoO2 after exposure to methane and carbon dioxide within 5 h. Meanwhile, coking was not observed over Ni–Mo2 C (1/3). The results indicate that the oxidation of Mo2 C with oxygen species originated from the dissociation of CO2 is the main reason for Ni–Mo2 C (1/3) deactivation. We deduce that over Ni–Mo2 C (1/3), the rate of methane dissociation was lower than that of CO2 dissociation. In such a situation, the oxidized Mo species cannot

C. Shi et al. / Applied Catalysis A: General 431–432 (2012) 164–170

m/e=44

m/e=28

200

300

400

m/e=28

500

600

700

800

900

200

300

400

Temperature (ºC)

B

500

600

700

800

Temperature(ºC)

B

Ni/Mo=1/3 Used

MoO2

Ni/Mo=1/1 Used

Intensity (a. u.)

Ni

Intensity (a. u.)

730 oC

730 oC

Intensity (a. u.)

m/e=44

Ni/Mo=1/1 Used

550 oC

A

CO-TPO 2

Ni/Mo=1/3 Used

Intensity (a. u.)

A

169

Mo2C

Ni Mo2C

Ni-MoC 2

20

25

30

35

40

45

50

55

60

65

70

2 Theta (º)

20

25

30

35

40

45

50

55

60

65

70

2 Theta (º)

Fig. 6. (A) CO2 -TPO profile and (B) XRD pattern of the used Ni–Mo2 C (1/3). Fig. 7. (A) CO2 -TPO profile and (B) XRD pattern of the used Ni–Mo2 C (1/1).

be recarburized by CH4 and an “oxidation–recarburization” cycle cannot be established. In contrast, bulk oxidation of Mo2 C was not observed over the Ni–Mo2 C (1/1) catalyst (Fig. 7B). The result indicated that the cause for Ni–Mo2 C (1/1) deactivation is completely different from that of Ni–Mo2 C (1/3) deactivation. For Ni–Mo2 C (1/1), the main deactivation reason is carbon deposition. It is known that coking is common over nickel-based catalysts in DRM processes. At a Ni/Mo molar ratio of 1/1, the rate of carbon deposition is high and surface carbon

cannot be timely removed by the oxygen species that are originated from CO2 dissociation. The net result is carbon deposition that causes the ultimate deactivation of Ni–Mo2 C (1/1). It is only when at a proper Ni/Mo molar ratio that an “oxidation–recarburization” cycle can be established and a stable DRM activity can be maintained as demonstrated in the case of Ni–Mo2 C (1/2) (Scheme 1).

4. Conclusion

Scheme 1. Catalytic oxidation–recarburization cycle over Ni–Mo2 C catalyst.

Ni-modified Mo2 C catalysts were studied for CO2 reforming of methane at atmospheric pressure. It was found that the presence of Ni could promote the carburization of Mo2 C and dissociation of CH4 . Through the optimization of Ni/Mo molar ratio, stable catalytic performance could be achieved at a Ni/Mo molar ratio of 1/2. Characterization of the spent samples revealed that the deactivation of Ni–Mo2 C (1/1) was due to coke formation whereas that of Ni–Mo2 C (1/3) was due to Mo2 C bulk oxidation. Only at a Ni/Mo molar ratio of 1/2, the catalytic oxidation–reduction cycle could be established. Based on the results, it is clear that Ni–Mo2 C catalyst is a typical bi-functional catalyst for CH4 /CO2 dry reforming. The dissociation of CH4 is catalyzed by Ni, while the activation of CO2 takes place on Mo2 C. By regulating the molar ratio of Ni and Mo2 C, there is a matching of CH4 dissociation and CO2 activation rates. Thus, a catalytic redox cycle is established and the deactivation due to carbon accumulation or Mo2 C oxidation could be avoided.

170

C. Shi et al. / Applied Catalysis A: General 431–432 (2012) 164–170

Acknowledgments The work was supported by the National Natural Science Foundation of China (Nos. 20573014 and 21073024), Natural Science Foundation of Liaoning Province (No. 201102034) and by the Program for New Century Excellent Talents in University (NCET07-0136), as well as by the Fundamental Research Funds for the Central Universities (No. DUT12LK23). References [1] M.S. Fan, A.Z. Abdullah, S. Bhatia, ChemCatChem 1 (2009) 192–208. [2] J.R.H. Ross, A.N.J. van Keulen, M.E.S. Hegarty, K. Seshan, Catal. Today 30 (1996) 193–199. [3] M.A. Vannice, Catal. Rev. 14 (1976) 153–191. [4] H.Y. Wang, C.T. Au, Appl. Catal. A: Gen. 155 (1997) 239–252. [5] F. Frusteri, F. Arena, G. Calogero, T. Torre, A. Parmaliana, Catal. Commun. 2 (2001) 49–56. [6] M.C.J. Bradford, M.A. Vannice, Appl. Catal. A: Gen. 142 (1996) 73–96. [7] S. Tang, L. Ji, J. Lin, H.C. Zeng, K.L. Tan, K. Li, J. Catal. 194 (2000) 424–430. [8] K. Takanabe, K. Nagaoka, K.I. Aika, Catal. Lett. 102 (2005) 153–157. [9] N. Laosiripojana, W. Sutthisripok, S. Assabumrungrat, Chem. Eng. J. (Lausanne) 112 (2005) 13–22.

[10] R.B. Levy, M. Boudart, Science 181 (1973) 547–549. [11] H.H. Hwu, J.G. Chen, Chem. Rev. (Washington, DC, USA) 105 (2004) 185–212. [12] Z. Yao, A. Zhu, J. Chen, X. Wang, C.T. Au, C. Shi, J. Solid State Chem. 180 (2007) 2635–2640. [13] S. Naito, M. Tsuji, T. Miyao, Catal. Today 77 (2002) 161–165. [14] D.C. LaMont, W.J. Thomson, Chem. Eng. Sci. 60 (2005) 3553–3559. [15] J.B. Claridge, A.P.E. York, A.J. Brungs, C.M. Alvarez, J. Sloan, S.C. Tsang, M.L.H. Green, J. Catal. 180 (1998) 85–100. [16] M. Tsuji, T. Miyao, S. Naito, Catal. Lett. 69 (2000) 195–198. [17] H. Shao, E.L. Kugler, W. Ma, D.B. Dadyburjor, Ind. Eng. Chem. Res. 44 (2005) 4914–4921. [18] T. Hirose, Y. Ozawa, M. Nagai, Chin. J. Catal. 32 (2011) 771–776. [19] J.M. Cheng, W. Huang, Fuel Process. Technol. 91 (2010) 185–193. [20] J. Guo, A.J. Zhang, A.M. Zhu, Y. Xu, C.T. Au, C. Shi, Advances in CO2 Conversion and Utilization, American Chemical Society, 2010, pp. 181–196. [21] A. Zhang, A. Zhu, B. Chen, S. Zhang, C. Au, C. Shi, Catal. Commun. 12 (2011) 803–807. [22] J. Patt, D.J. Moon, C. Phillips, L. Thompson, Catal. Lett. 65 (2000) 193–195. [23] N.M. Schweitzer, J.A. Schaidle, O.K. Ezekoye, X.Q. Pan, S. Linic, L.T. Thompson, J. Am. Chem. Soc. 133 (2011) 2378–2381. [24] N. Ji, T. Zhang, M. Zheng, A. Wang, H. Wang, X. Wang, J.G. Chen, Angew. Chem. Int. Ed. 47 (2008) 8510–8513.