Redox behaviour and catalytic properties of Ce0.5Zr0.5O2-supported palladium catalysts

Redox behaviour and catalytic properties of Ce0.5Zr0.5O2-supported palladium catalysts

Applied Catalysis A: General 189 (1999) 15–21 Redox behaviour and catalytic properties of Ce0.5 Zr0.5 O2 -supported palladium catalysts Meng-Fei Luo ...

226KB Sizes 0 Downloads 8 Views

Applied Catalysis A: General 189 (1999) 15–21

Redox behaviour and catalytic properties of Ce0.5 Zr0.5 O2 -supported palladium catalysts Meng-Fei Luo ∗ , Xiao-Ming Zheng Institute of Catalysis, Zhejiang University, Xixi Campus, Hangzhou 310028, PR China Received 9 March 1999; received in revised form 4 June 1999; accepted 7 June 1999

Abstract Ce0.5 Zr0.5 O2 -supported palladium catalysts were prepared and used for CO and methane oxidation. The redox behaviour of these catalysts was investigated by using H2 -TPR and CO-TPR. The shape and temperature of TPR peaks depend on the nature of reducing agent. Two peaks (␣1 and ␥) were observed when H2 was used as reducing agent for the fresh catalyst, while three peaks (␣0 , ␤0 and ␥0 ) were observed when CO was used as a reducing agent. The addition of a noble metal effectively promotes the reduction of Ce0.5 Zr0.5 O2 support, which is attributed to hydrogen spillover. After re-oxidation treatments, the PdO reduction process splits into two peaks (␣1 and ␣2 ). On the basis of the catalytic activity and CO-TPR results, the activity of CO oxidation is related to the finely dispersed PdO, while the activity of methane oxidation is related to the large PdO particles. ©1999 Elsevier Science B.V. All rights reserved. Keywords: Redox behaviour; Ce0.5 Zr0.5 O2 ; Palladium; CO oxidation; Methane oxidation

1. Introduction Volatile organic compounds (VOCs) and CO are emitted from an increasing number of industrial processes [1,2]. Catalytic combustion technology has become an effective means of removing toxic odours, VOCs and CO, as the energy consumption of catalytic combustion is less than in traditional thermal methods. Precious metals, such as platinum and palladium, are well-known complete oxidation catalysts with high activity and stability, and are widely used for control of exhaust gas and automotive exhaust [3,4]. Since methane is the most refractory hydrocarbon, it often

∗ Corresponding author. Tel./fax: +86-571-827-3283 E-mail address: [email protected] (M.-F. Luo)

used as the model hydrocarbon compound for activity tests [5]. The search for new catalysts is continuing in order to design complete oxidation catalysts with improved activity. The fluorite-type oxides, such CeO2 and ZrO2 , have been widely investigated owing to the broad range of applications in different fields. It has been reported that the ceria–zirconia mixed oxides show enhanced thermal [6], redox [7–11] and catalytic properties [12,13] compared to pure CeO2 . However, little study on Ce0.5 Zr0.5 O2 mixed oxide as a support for palladium catalyst for CO and methane oxidation has been reported. The present work is concerned with redox behaviour of PdO/Ce0.5 Zr0.5 O2 catalysts and with the oxidation activity for CO and methane. Our aim is to aid the development of PdO/Ce0.5 Zr0.5 O2 catalyst for CO and methane oxidation.

0926-860X/99/$ – see front matter ©1999 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 6 - 8 6 0 X ( 9 9 ) 0 0 2 3 0 - 6

16

M.-F. Luo, X.-M. Zheng / Applied Catalysis A: General 189 (1999) 15–21

2. Experimental 2.1. Preparation and crystal structure characterization of Ce0.5 Zr0.5 O2 solid solution The Ce0.5 Zr0.5 O2 solid solution was prepared by evaporating an aqueous solution of the mixed metal nitrates containing an equivalent amount of citric acid to obtain a gel, followed by decomposition at 950◦ C for 4 h. Its BET surface area is 11 m2 g−1 . XRD data for Rietveld analysis were obtained on a Rigaku D/max-III B powder diffractometer, using CuK␣ radiation and a power of 40 kV × 30 mA. The intensity data were collected at 25◦ C over a 2θ range of 10–130◦ with a step interval of 0.02◦ and a counting time of 8 s per step. BaF2 , which was annealed at 500◦ C for 2 h in an air atmosphere, was used as a non-intrinsic broadening sample to measure instrument function and to extract the micro-structure date [14]. The micro-structure parameters of samples were determined by means of Rietveld analysis combined with Fourier analysis for broadened peaks. The phenomenological relationship, describing the trend of profile width and shape as a function of diffraction angle as popular Rietveld refinement routines, was replaced by fitting parameters of crystallite size, shape and micro-strain. And fitting parameters were transformed to the parameters of a pseudo-Voigt (pV) function by inverting the Warren–Averbach procedure for a single peak. Then the experimental date was fitted. And some constraint was applied according to the symmetry properties of the crystal [15]. Rietveld refinements were performed on a 586 PC computer with ls1 software [16]. Fig. 1 shows the Rietveld refinement patterns for Ce0.5 Zr0.5 O2 . This indicates that both the cubic and tetragonal phases are observed in Ce0.5 Zr0.5 O2 solid solution. The XRD pattern observed for the sample was in agreement with the powder X-ray diffraction profile calculated. Table 1 lists the phase composition and lattice parameters. 2.2. Preparation of catalysts The PdO/Ce0.5 Zr0.5 O2 catalyst was prepared by the conventional wet impregnation method using an aque-

Fig. 1. Rietveld refinement patterns of Ce0.5 Zr0.5 O2 . The observed data are indicated by dots, the calculated data by the solid line overlaying them. The short vertical lines mark the positions of possible Bragg reflections (above: Tetragonal; below: Cubic), and the lower curve shows the difference between the observed and calculated powder diffraction patterns.

Table 1 The phase composition and the lattice parameters of Ce0.5 Zr0.5 O2 solid solution Phase composition

Cell parameters (nm)

Cubic (31.2%) Tetragonal (68.8%)

0.5328(1) 0.3726(1)

a

b

c 0.5295(1)

ous solution of H2 PdCl4 , dried overnight in an oven at 120◦ C and then was heated in air at 650◦ C for 4 h. The loading of Pd is 0.05, 0.10, 0.25, 0.75, 2.0 and 5.0%, respectively. These catalysts are denoted as PdO/Ce0.5 Zr0.5 O2 (X%). 2.3. Activity measurements Catalytic activity measurements were carried out in a fixed bed reactor (i. d. 6 mm) using a 150 mg of catalyst of 20–60 mesh size. The total gas flow rate was set at 80 ml min−1 . For CO oxidation, the gas consisted of 2.4% CO and 1.2% O2 in N2 . For methane oxidation, the gas consisted of 2.8% CH4 and 8% O2 in N2 . The catalysts were directly exposed to 80 ml min−1 of reaction gas while the reactor temperature stabilized without any pre-treatment. The products were analyzed by

M.-F. Luo, X.-M. Zheng / Applied Catalysis A: General 189 (1999) 15–21

gas chromatography with Molecular Sieves 13X and Porapak Q columns, both operating at 50◦ C.

17

reduction was measured by a TCD. The effluent CO2 during CO-TPR was adsorbed with a 5A molecular sieve.

2.4. Measurement of Pd dispersion The dispersion of palladium was calculated using a CO : Pd stoichiometry factor of 1, and assuming that all CO was adsorbed on the exposed Pd atoms. Adsorption of CO on Pd was measured by a pulse method. After the catalyst was reduced with hydrogen at 100◦ C, 0.25 ml of CO was injected with a syringe repeatedly until no adsorption was observed. The amount of CO adsorbed was calculated from the difference between amounts of injection and effluent CO. The amount of CO adsorbed on Pd atoms was calculated by the difference between the amount on the catalyst and that on Ce0.5 Zr0.5 O2 support.

3. Results and discussion 3.1. Catalytic activity for CO and methane oxidation Figs. 2 and 3 show the PdO/Ce0.5 Zr0.5 O2 catalyst oxidation, respectively. It PdO/Ce0.5 Zr0.5 O2 (0.25%)

light-off for CO can be is most

curves of the and methane seen that the active among

2.5. H2 -TPR The reduction properties of PdO/Ce0.5 Zr0.5 O2 catalysts were measured by means of the temperatureprogrammed reduction (TPR), which was described in our previous paper [17]. The weight of the sample is 25 mg in this work. The heating rate is 20◦ C min−1 . The purpose of the re-oxidation treatment is to reveal the oxidation properties of the reduced sample. Before re-oxidation, the sample was reduced in a flow of H2 –N2 (5 : 95) mixed gas at a heating rate of 20◦ C min−1 up to 950◦ C. Subsequently the temperature of the sample was decreased to a desired temperature (i.e. 30, 100, 200 or 500◦ C) and the gas flow was simultaneously switched to oxygen for 0.5 h. Then the temperature of the sample was decreased to 20◦ C. Finally, TPR of re-oxidation sample was carried out. The amount of H2 uptake in TPR was estimated from the integrated peak areas by comparison with those obtained by using pure CuO as a standard.

Fig. 2. Light-off curve of PdO/Ce0.5 Zr0.5 O2 catalysts for CO oxidation.

2.6. CO-TPR CO-TPR measurement was made in a flow system. A 50 mg catalyst was pre-treated in air at 200◦ C and placed in a TPR cell at 20◦ C through which CO–He (10 : 90) mixed gas flowed. The temperature of the sample was programmed to rise at a constant rate of 10◦ C min−1 . The amount of CO uptake during the

Fig. 3. Light-off curve of PdO/Ce0.5 Zr0.5 O2 catalysts for methane oxidation.

18

M.-F. Luo, X.-M. Zheng / Applied Catalysis A: General 189 (1999) 15–21

decrease of the catalytic activity for CO oxidation. However, for methane oxidation, the catalytic activity obviously increases with an increase in palladium loading. With low Pd loading, such as 0.05, 0.10 and 0.25%, the activity of catalysts is less than that of pure PdO. We think that finely dispersed PdO (small particles) is not active, and that large PdO particles are the active sites for methane oxidation. This result is consistent with the PdO–CeO2 catalyst results in our previous paper [18], and with the SiO2 , Al2 O3 and SiO2 –Al2 O3 -supported palladium catalyst results for methane oxidation reported by Muto et al. [19]. Fig. 4. Effect of palladium loading in PdO/Ce0.5 Zr0.5 O2 catalyst on activity of CO and methane oxidation.

Table 2 Dispersion of palladium on PdO/Ce0.5 Zr0.5 O2 catalysts Catalyst PdO/Ce05 Zr0.5 O2 PdO/Ce05 Zr0.5 O2 PdO/Ce05 Zr0.5 O2 PdO/Ce05 Zr0.5 O2 PdO/Ce05 Zr0.5 O2 PdO/Ce05 Zr0.5 O2

Dispersion (%) (0.05%) (0.10%) (0.25%) (0.75%) (2.0%) (5.0%)

57.8 41.9 25.5 14.5 9.1 5.0

these catalysts for CO oxidation; However, the PdO/Ce0.5 Zr0.5 O2 (5%) is most active for methane oxidation. Fig. 4 shows the relationship between T20 (the temperature for 20% conversion) and Pd loading for both CO and methane. The activity of PdO/Ce0.5 Zr0.5 O2 (0.75%) for methane is near that of pure PdO (Fig. 3). However, the activity of the PdO/Ce0.5 Zr0.5 O2 catalyst is much higher than the activities of both pure PdO and Ce0.5 Zr0.5 O2 support for CO oxidation. Only a very small amount of palladium is needed to form an active site for CO oxidation. A maximum in activity enhancement is observed at a Pd concentration of 0.25%. Table 2 lists the dispersion of Pd on PdO/Ce0.5 Zr0.5 O2 catalysts with various Pd loading catalysts. As shown in Table 2, the dispersion of Pd decreases with increase in the amount of Pd loading. On the basis of the activity and dispersion results, we conclude that finely dispersed PdO (small particles) is the active site. And the excess Pd forms the larger PdO particles, which may cover a part of the active sites and result in the

3.2. Redox behaviour of PdO/Ce0.5 Zr0.5 O2 catalysts Fig. 5 shows the H2 -TPR profiles for fresh PdO/Ce0.5 Zr0.5 O2 catalysts with various Pd loadings. The reduction profile of the Ce0.5 Zr0.5 O2 solid solution is characterized by a single peak at about 640◦ C. Two peaks with maxima at 60 and 170◦ C are found in all PdO/Ce0.5 Zr0.5 O2 catalysts. These peaks are designated by α 1 and γ in Fig. 5. The H2 consumption of the α 1 and γ peak is also listed in Table 3. The H2 consumption of the α 1 peak strongly depends on Pd loading, and its peak temperature shifts to a

Fig. 5. H2 -TPR profiles of fresh PdO/Ce0.5 Zr0.5 O2 catalysts with various Pd loadings.

M.-F. Luo, X.-M. Zheng / Applied Catalysis A: General 189 (1999) 15–21

19

Table 3 The H2 Consumption of TPR peaks over PdO/Ce0.5 Zr0.5 O2 catalysts H2 Consumption (mmol g−1 )

Catalyst

Ce05 Zr0.5 O2 PdO/Ce05 Zr0.5 O2 PdO/Ce05 Zr0.5 O2 PdO/Ce05 Zr0.5 O2 PdO/Ce05 Zr0.5 O2 PdO/Ce05 Zr0.5 O2

(0.05%) (0.25%) (0.75%) (2.0%) (5.0%)

␣ peak

␤ peak

– 0.024 0.057 0.105 0.254 0.460

0.538 0.540 0.540 0.541 0.540 0.542

lower temperature with increasing Pd loading, but H2 consumption and the position of the γ peak remain unchanged with increasing Pd loading. As the α 1 peak strongly depends on Pd loading, the α 1 peak is attributed to the reduction of the PdO precursor. H2 consumption for this peak is higher than that expected for theoretical PdO reduction. For example, consumption of 0.105 mmol g−1 was measured for the ␣1 peak in the PdO/Ce0.5 Zr0.5 O2 (0.75%); this is obviously higher than 0.06 mmol g−1 expected for theoretical PdO reduction. The difference decreases with increasing Pd loading. This difference may be attributed to concurrent reduction of surface Ce4+ , which occurs at this lower temperature [8]. The γ peak contributes to the reduction of the Ce0.5 Zr0.5 O2 support. In the presence of the metal, the peak at 640◦ C is shifted to 160◦ C. This indicates that the addition of the noble metal effectively promotes the reduction of the Ce0.5 Zr0.5 O2 support. This is attributed to the ability of the supported metal (Rh, Pt or Pd) to activate H2 and then to spill it over onto the support [8]. In the absence of the metal, H2 activation is difficult and may become the rate determining step. Fig. 6 shows the effect of re-oxidation temperatures on TPR profiles of reduced PdO/Ce0.5 Zr0.5 O2 (0.75%) catalyst. The TPR of the fresh catalyst is also shown in Fig. 6 for comparison. After re-oxidation treatment, a new peak at 80◦ C is observed (henceforth, this new peak will be indicated as the ␣2 peak). The re-oxidation treatments do not effect ␥ peak of the catalyst. As the temperature of the re-oxidation treatment decreases, the ␣2 peak becomes clear, while the intensity of ␣1 peak decreases. However, the ␣1 peak disappears after re-oxidation at 100◦ C. The total H2 consumption (␣1 + ␣2 ) of the re-oxidation sample decreases slightly

Fig. 6. H2 -TPR profiles of fresh PdO/Ce0.5 Zr0.5 O2 (0.75%) catalyst and of re-oxidized catalysts at different temperatures.

Fig. 7. H2 -TPR profiles of (a) fresh PdO/Ce0.5 Zr0.5 O2 (0.75%) catalyst, of (b, c, and d) recycled catalyst, respectively, 1, 2 and 3 times.

with decreasing re-oxidation temperatures. Fig. 7 shows the effect of recycled time on the TPR profile of PdO/Ce0.5 Zr0.5 O2 (0.75%) catalyst re-oxidized at

20

M.-F. Luo, X.-M. Zheng / Applied Catalysis A: General 189 (1999) 15–21

500◦ C. The recycled times do not affect the ␥ peak of the catalyst. However, on increasing the recycled time from 1 to 3, the intensity of the ␣2 peak increases, while the ␣1 peak decreases. On further increasing the reduction–oxidation cycle times, the TPR profile reaches the final aspect shown in trace (D) in Fig. 7, which features two peaks, but the overall H2 consumption of α 1 and α 2 peaks is almost constant. From the re-oxidation experiment, it can be seen that the splitting of the PdO reduction process into two peaks heavily depends on the condition of re-oxidation. On the basis of experiment results and our previous work [17], we suppose that there are two reasons. One is non-uniform size distribution of PdO; the other is that part of PdO moves into the lattice of Ce0.5 Zr0.5 O2 solid solution after reduction–oxidation. On the basis of the H2 -TPR result, we do not differentiate directly the finely dispersed PdO from the crystal phase PdO over all fresh PdO/Ce0.5 Zr0.5 O2 catalysts, because only one reduction peak is observed. Numerous studies have reported that the reduction temperature of metal oxide by H2 can be lowered by the addition of transition metals, with Pt and Pd being the most efficient metals [20]. For example, surface hydrogen atoms formed on Pt, Pd, Cu and Ni can increase the rate of reduction of CuO, NiO and Co3 O4 . In the case of hydrogen reduction, the effect of Pt and Pd can be attributed to hydrogen spillover from Pt particles, leading to activated hydrogen atoms, which can reduce metal oxides more easily. We suppose that crystal phase PdO is first reduced to form Pd atoms at lower temperatures and then H2 is dissociated over these Pd atoms and the hydrogen atoms spillover to the surface of finely dispersed PdO. Thus the reduction temperature of the finely dispersed PdO by H2 can be decreased. In other words, the metal palladium formed is thought to dissociate hydrogen so that the reduction is autocatalytic [21]. Therefore, only one reduction peak is observed in Fig. 5 and the peak temperature shifts to a lower temperature with increasing Pd loading.

3.3. CO-TPR In order to differentiate the finely dispersed PdO from bulk-like crystalline PdO, we have investigated CO-TPR of PdO/Ce0.5 Zr0.5 O2 catalysts, since the

Fig. 8. CO-TPR profiles of fresh PdO/Ce0.5 Zr0.5 O2 catalysts with various Pd loadings.

spillover of carbon monoxide occurs with more difficulty than that of hydrogen. The CO-TPR profiles for these PdO/Ce0.5 Zr0.5 O2 catalysts are shown in Fig. 8. Compared with Fig. 5, it can be seen that the shape and temperature of TPR peak depends on the nature of the reducing agent. The reduction of profile of the support is characterized by a single peak at about 510◦ C. It is very interesting that when carbon monoxide is used as reduction agent, three peaks (␣0 , ␤0 and ␥0 ) are observed, while two peaks appear when H2 was used as a reducing agent for the fresh catalyst. From Fig. 8, one can see that temperatures of both ␣0 and ␤0 peaks shift to slightly lower temperature with an increase of Pd loading. However, the ␥0 peak obviously shifts to lower temperature. The intensity of ␤0 peak depends strongly on Pd loading, while the intensities of ␣0 and ␥0 peaks are almost independent of Pd loading from 0.1 to 5%. We suppose that the ␣0 peak is due to the finely dispersed PdO and that this is the active phase on PdO/Ce0.5 Zr0.5 O2 catalysts for CO oxidation. As the ␣0 peak is very small, the amount of the fine dispersed PdO is very small. So we think this is another reason for the lack of any H2 -TPR peak of finely dispersed PdO in the fresh catalyst. We suggest that the ␤0 peak can be attributed to the large PdO particles, and the ␥0 peak can be

M.-F. Luo, X.-M. Zheng / Applied Catalysis A: General 189 (1999) 15–21

attributed to the reduction of the support. As the peak is very weak, the amount of dispersion PdO is lower. From Fig. 8, it can be seen that the curves of the CO-TPR pattern usually cannot return to the base line at high temperature, indicating that the Boudouard reaction, i.e. 2CO → C + CO2 , may occur under high temperature [22,23]. On the basis of the results of catalytic activity and CO-TPR, we found that the CO oxidation activity order of PdO/Ce0.5 Zr0.5 O2 is consistent with the ␣0 peak (in Fig. 8). However, the order of methane oxidation activity is consistent with the intensity of ␤0 peak (in Fig. 8) except for Pd loading of 5%. This also indicated that the CO oxidation activity is related to Pd hydroxide and/or the finely dispersed PdO, while the methane oxidation activity is related to the bulk-like crystalline PdO. However, if the content of the PdO surpasses 2%, its further enhancement cannot cause an increase of the activity.

Acknowledgements This research project was supported by the Zhejiang Provincial Natural Science Foundation of China. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10]

4. Conclusions PdO/Ce0.5 Zr0.5 O2 catalysts of different Pd loading have been prepared using impregnation methods. The reduction–oxidation behaviour of the PdO/Ce0.5 Zr0.5 O2 (0.75%) catalyst was investigated by H2 -TPR: two peaks can be observed on a fresh catalyst. ␣1 peak is attributed to the reduction of PdO; ␤ peak is ascribed to the reduction of Ce0.5 Zr0.5 O2 support. The presence of Pd improves the reduction of Ce0.5 Zr0.5 O2 support. The reduced catalyst is easily oxidized. After re-oxidation treatment, the PdO reduction process splits into two peaks. For CO oxidation, a maximum in activity enhancement is observed at a Pd concentration of 0.25%. For methane oxidation, the catalytic activity obviously increases with palladium loading. On the basis of catalytic and CO-TPR results, the CO oxidation activity is related to Pd hydroxide or/and dispersed PdO, while the methane oxidation activity is related to the bulk-like crystalline PdO.

21

[11] [12] [13]

[14] [15] [16] [17] [18] [19] [20] [21] [22] [23]

J.J. Spivey, Ind. Eng. Chem. Res. 26 (1987) 2165. J.N. Armor, Appl. Catal. B 1 (1992) 221. Y.F.Y. Yao, J. Catal. 87 (1984) 152. I. Matasuura, Y. Hashimoto, O. Takayasu, Appl. Catal. 74 (1991) 273. W. Liu, M. Flytzani-Stephanopoulos, J. Catal. 153 (1995) 304. M. Pijolat, M. Prin, M. Soustelle, O. Touret, P. Nortier, J. Chem. Soc. Faraday Trans. 91 (1995) 3941. A. Trovarelli, F. Zamar, J. Llorca, C. de Leitenburg, G. Dolcetti, J.T. Kiss, J. Catal. 169 (1997) 490. P. Formasiero, R. Di Monte, G. Ranga Rao, J. Kaspar, A. Trovarelli, M. Graziani, J. Catal. 151 (1995) 168. P. Formasiero, J. Kaspar, M. Graziani, J. Catal. 167 (1997) 576. P. Formasiero, G. Balducci, R. Di Monte, J. Kaspar, V. Sergo, G. Gubitosa, A. Ferrero, M. Graziani, J. Catal. 164 (1996) 173. P. Vidmar, P. Fornasiero, J. Kaspar, G. Gubitosa, M. Graziani, J. Catal. 171 (1997) 160. F. Zamar, A. Trovarelli, C. de Leitenburg, G. Dolcetti, J. Chem. Soc. Chem. Commun., 1996, 965. C. de Leitenburg, A. Trovarelli, F. Zamar, S. Maschio, G. Dolcetti, J. Llorca, J. Chem. Soc. Chem. Commun., 1995, 2181. D. Louêr, J.I. Langford, J. Appl. Cryst. 21 (1988) 430. L. Lutterotti, R. Scardi, J. Appl. Cryst. 23 (1990) 246. L. Lutterotti, R. Scardi, J. Appl. Cryst. 25 (1990) 459. M.-F. Luo, Z.-Y. Hou, X.-X. Yuan, X.-M. Zheng, Catal. Lett. 50 (1998) 205. M.-F. Luo, X.-M. Zheng, J. Chem. Research (S), (1998) 274. K.-J. Mnto, N. Katada, M. Niwa, Appl Catal A 134 (1996) 303. W.C. Conner Jr., J.L. Falconer, Chem. Rev. 95 (1995) 759. N.W. Hurst, S.J. Gentry, A. Jones, B.D. Mcnicol, Catal. Rev.-Sci. Eng. 24 (1992) 233. W.-.P. Dow, T. Huang, J. Catal. 147 (1994) 322. C. Serre, F. Garin, G. Belot, G. Maire, J. Catal. 141 (1993) 1.