- Email: [email protected]
SiO2 catalysts
Applied Catalysis A: General 320 (2007) 91–97 www.elsevier.com/locate/apcata
The role of additives in the catalytic reduction of NO by CO over Pd-In/SiO2 and Pd-Pb/SiO2 catalysts Takashi Hirano, Yoshiyuki Ozawa, Takayuki Sekido, Takashi Ogino, Toshihiro Miyao, Shuichi Naito * Department of Material and Life Chemistry, Kanagawa University, 3-27-1, Rokkakubashi, Kanagawaku, Yokohama 221-8686, Japan Received 29 November 2006; received in revised form 18 December 2006; accepted 21 December 2006 Available online 3 January 2007
Abstract Addition effect of In and Pb on the NO-CO reaction over SiO2 supported Pd catalysts was studied, using a closed gas circulation system as well as in situ infrared, XRD and XPS spectroscopies. Formation of intermetallic compounds was observed in the cases of both Pd-In/SiO2 and Pd-Pb/ SiO2 catalysts. These compounds caused a drastic enhancement of the reaction rate of N2O formation. Infrared analyses revealed the weakening of CO adsorption on Pd metal by the formation of intermetallic compounds, which is the main reason of the enhancement of reaction rate. Over Pd/ SiO2, the reaction may proceed via redox mechanism through NO dissociation to form the oxidized surface, followed by its reduction with CO. By adding In or Pb, a new reaction pathway is opened at the lower temperature region via NO dimer-like intermediates. The role of additives is the stabilization of these unique intermediates on the slightly oxidized intermetallic compound surfaces. # 2007 Elsevier B.V. All rights reserved. Keywords: NO-CO reaction; Pd/SiO2 catalyst; Addition effect; Pd-In; Pd-Pb
1. Introduction In the catalytic conversion of NO by CO, rhodium is well known to be the most active of all precious metals. For the oxidation of CO, platinum is generally accepted to be the most efficient catalyst [1]. Palladium is the third component in the so called three-way Rh-Pt-Pd automotive catalysts. In the past decade, many studies have been performed to replace Rh in automotive catalysts in view of its high cost and scare resources. Among them, catalytic properties of Pd have received increasing attention because of its potentiality for the substitution of Rh in the conventional three-way Rh/Pt catalyst [2–5]. However, the Pd-only catalyst has a poor NO reduction efficiency compared to Rh, especially in an oxygen rich environment [6,7]. However, it could be a good basis for a new type of bimetallic system if its selectivity for NO reduction could be increased by the presence of a second metal such as Ag and Cr [8].
* Corresponding author. Tel.: +81 45 481 5661; fax: +81 45 491 7915. E-mail address: [email protected] (S. Naito). 0926-860X/$ – see front matter # 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.apcata.2006.12.015
Attempts to improve the catalytic performance of Pd have also been made by placing it in contact with various base metal oxides and rare earth oxides. This was successfully achieved with Pd promoted by cobalt oxide [8] or with Pd-Mn [9] or PdCr [10] couples, or with Pd-VOx/Al2O3 catalysts [11], which have shown light-off temperatures of conversion lower by 80 to 100 K compared with that of pure Pd. In the case of Pd-Mn, the bimetallic cluster of Pd3Mn was reported to be the active species for this reaction and adsorbed NO with a long N-O bond and quasi-parallel to the surface was a precursor for the dissociation. In the case of Pd-VOx, the reduction of VOx species was enhanced by a close contact with Pd; partially reduced vanadia decorated the noble metal particles and enhanced NO decomposition. Valden et al. presented an interesting paper using Pd/La2O3/ Al2O3 catalysts in which the reactivity between NO and CO and their adsorption was studied by TPD and IR [12]. The latter technique indicates different behavior, since NO was adsorbed molecularly on the support and lanthanum oxide, while NO dissociation followed by the formation of N2 and N2O was seen on all Pd catalysts. Pd catalysts supported on Al2O3, La2O3Al2O3 and LaAlO3 were prepared: these contain different amounts of Pd. CO adsorbed molecularly on all Pd catalysts
92
T. Hirano et al. / Applied Catalysis A: General 320 (2007) 91–97
forming linear and bridged CO species. The presence of PdO was found to decrease the CO binding energy on Al2O3supported Pd catalysts. The destabilization of CO bonding was even further enhanced on Pd catalysts supported on LaAlO3 and La2O3-Al2O3. In contrast to the negligible interaction between CO and supports, NO was observed to adsorb molecularly on the supports. It was reported that the addition of MoO3 to Pd/Al2O3 catalysts improved the NO activity with high selectivity to N2 in the presence of a small excess amount of oxygen [13]. The addition of 8 wt% molybdenum covered the surface of alumina and led to the disappearance of surface hydroxyls [14,15]. The MoO3 addition also induced modifications on the metallic phase, which causes the dilution or blocking of Pd surface by molybdenum oxide species. The catalytic behavior could be explained by a bi-functional mechanism. After adsorption and dissociation of NO on Mo4+, the dissociated oxygen is transferred to the Pd. The CO adsorbed on the Pd surface reacts with oxygen. A similar effect was observed in the cases of Pd-WOx/Al2O3 catalysts [16]. In the present study, we investigated the effect of adding In and Pb to silica supported Pd in the reduction of NO by CO: we found that these additives exhibit a large acceleration effect. Especially in the cases of Pd-In/SiO2 and Pd-Pb/SiO2, an intermetallic compound formed between Pd-In is extraordinary active for NO-CO reaction, and N2O and CO2 are formed steadily even at room temperature. We applied in situ infrared, XPS and XRD analyses to elucidate the role of additives in the tremendous enhancement effect in NO-CO reactions. 2. Experimental Silica supported Pd, Pd-In, and Pd-Pb (5 wt% Pd) catalysts were prepared by a conventional co-impregnation method using SiO2 (Aerosil 300) and (NH4)2PdCl4, InCl34H2O and PbCl4 (Wako chemicals) as precursors. The molar ratio of Pd and additive was 1:1. The catalyst (0.2 g) was reduced by hydrogen at 733 K for 10 h to form intermetallic compounds. The NOCO reaction was carried out in a closed gas circulation system under a 1:1 ratio of NO and CO (4 kPa each). The composition of the gas phase during the reaction was followed by TCD gas chromatography (Shimazu GC-17A). For infrared spectroscopic experiments, the catalyst was pressed into a 20 mm diameter disk and put into an infrared cell, which was connected to a closed gas circulation system. After the pretreatment mentioned above, the adsorbed species were measured by IR spectrometry (JEOL, FT-IR, Diamond 20). A transmission electron microscope (JEM2010, JEOL) with an acceleration voltage of 2000 kV and LaB6 cathode was applied for the observation of the images of supported catalysts. A X-ray photoelectron spectroscope (JPS-9010, JEOL) with a Mg Ka X-ray source (10 kV, 10 mA) was applied for the analysis of the electronic states of supported catalysts. Samples were prepared by molding in thin disk shapes. Each was introduced into the preparation chamber. After reduction by H2 at 573 K, each sample was transferred into analysis chamber without exposure to air. The catalysts were also characterized
by XRD (Rigaku, RAD-gX) and H2 and CO chemisorption method (Beckman Coulter, Omnisorp 100CX). 3. Results and discussion 3.1. Characterization of catalysts Fig. 1 shows the XRD patterns of Pd/SiO2, Pd-In/SiO2 and Pd-Pb/SiO2 catalysts after H2 reduction at 733 K for 10 h. In the case of Pd/SiO2 the diffraction pattern typical for FCC metal was observed as shown in A (2u = 40.0, 46.5, 68.7 and 82.0). When In was added, four characteristic peaks were shifted as shown in B (2u = 39.2, 56.8, 70.9, 84.2), which can be assigned to Pd0.48 In0.52 intermetallic compound. In the case of Pd-Pb/ SiO2, new XRD peaks emerged at 2u = 39.0, 40.5, 57.5, 65.0 and 71.0 by higher temperature reduction, which can be assigned to Pd3Pb2 intermetallic compound, as shown in Fig. 1C. Fig. 2 shows the TEM images of these catalysts. The particle sizes of Pd metal without additives were rather big (more than 10 nm) after 733 K reduction, but these became very small (2–3 nm) by the addition of In, forming Pd0.48 In0.52 intermetallic compounds. In the case of Pd-Pb/SiO2, binary distribution of Pd particles were observed; the particles were around 5–6 and 10–15 nm in diameters and both contained Pd and Pb, as determined by EDS analysis. Figs. 3 and 4 represent the XPS spectra of Pd3d, In3d and Pb4f transitions in Pd/SiO2, Pd-In/SiO2 and Pd-Pb/SiO2 catalysts after the reduction at 573 K in the in situ cell (spectrum (a) in each figure). In the case of Pd/SiO2, the binding energy of Pd3d5/2 was close to the zero valent Pd metal value in the literature (335.0 eV), whereas by adding In or Pb it shifted 0.8 eV higher binding energy side (335.6 eV) even after the reduction at higher temperatures. On the other hand, the binding energy of In3d5/2 in Pd-In/SiO2 was 0.1 eV lower than the zero valent In (443.7 eV) value, suggesting the formation of Pd-In intermetallic compound as shown in Fig. 4A(a). In the
Fig. 1. XRD patterns of Pd/SiO2, Pd-In/SiO2 and Pd-Pb/SiO2 catalysts after reduction at 733 K: (A) Pd/SiO2, (B) Pd-In/SiO2 and (C) Pd-Pb/SiO2.
T. Hirano et al. / Applied Catalysis A: General 320 (2007) 91–97
93
Fig. 2. TEM images of various catalysts: (A) Pd/SiO2, (B) Pd-In/SiO2 and (C) Pd-Pb/SiO2.
case of Pd-Pb/SiO2 catalyst, two Pb4f7/2 transitions were observed, at 137.1 and 139.1 eV, which can be assigned to zero valent and positively charged Pb species (Fig. 4B(a)). As shown in Fig. 1C, Pd3Pb2 intermetallic compound must be formed together with the oxidation of excess Pb to PbO2. 3.2. Kinetic study of NO-CO reaction
Fig. 3. XPS spectra of Pd3d transition in: (A) Pd/SiO2, (B) Pd-In/SiO2 and (C) Pd-Pb/SiO2 after various treatments in the in situ cell. (a) H2 reduction at 373 K, (b) NO-CO reaction at 373 K, (c) NO-CO reaction at 473 K and (d) NO-CO reaction at 573 K.
Fig. 4. XPS spectra of In3d and Pb4f transitions in: (A) Pd/SiO2, (B) Pd-In/ SiO2 and (C) Pd-Pb/SiO2 after various treatments in the in situ cell. (a) H2 reduction at 373 K, (b) NO-CO reaction at 373 K, (c) NO-CO reaction at 473 K and (d) NO-CO reaction at 573 K.
Fig. 5 summarizes the temperature dependences of the rates of (N2O + N2) formation in NO-CO reactions over Pd/SiO2, PdIn/SiO2 and Pd-Pb/SiO2 catalysts. As shown in A, the reaction started only at around 500–600 K over Pd/SiO2, and the main product was N2O with a small amount of N2. When In or Pb was added to this system, the reaction temperature was lowered more than 200 K and only N2O and CO2 were formed even at room temperature. Above 373 K, a small amount of N2 started to be formed together with N2O, which became the main product above 473 K. As shown in Fig. 5B and C, the dependence of the N2O + N2 formation rates upon the temperatures was very unique over Pd-In/SiO2 and Pd-Pb/ SiO2 catalysts. First of all, it increased linearly up to 473 K,
Fig. 5. Temperature dependences of the formation rates of N2O + N2 in NOCO reaction over: (A) Pd/SiO2, (B) Pd-In/SiO2 and (C) Pd-Pb/SiO2. Solid line(r): raising the reaction temperature, broken line(l): lowering the reaction temperature.
94
T. Hirano et al. / Applied Catalysis A: General 320 (2007) 91–97
showing a maximum at around 473–500 K, then decreased gradually up to 573 K, showing a minimum at around 573– 600 K, and then increased again. To see whether this behavior is reversible with temperatures or not, we have compared the temperature-increasing reaction with the temperature-decreasing reaction. After the reaction at 673 K, the temperature was lowered reversibly as shown by the dotted lines in Fig. 5. In the case of Pd/SiO2 almost the same temperature profile was obtained in A. On the other hand, over Pd-In/SiO2 the reaction rate was decreased seriously by the deactivation process at higher temperature reaction, however, it showed a similar temperature dependence to that in the case of temperatureincreasing reaction. Over Pd-Pb/SiO2, although the maximum point shifted to higher temperatures, it still showed a similar behavior to that in the case of Pd-In catalyst. Fig. 6 represents the comparison of Arrhenius plots of (N2O + N2) formation rates over Pd/SiO2 together with (N2O + N2) formation rates between the temperature ranges at 373–473 K (designated as LTR) and at 573–673 K (HTR) over Pd-In/SiO2 catalysts. From the slopes of the Arrhenius plots, the activation energies were estimated to be as follows; 70 kJ/mol for (N2O + N2) formation rate in the case of Pd/SiO2, 32 kJ/mol for the LTR over Pd-In/SiO2, and 66 kJ/mol for the HTR over Pd-In/SiO2 catalysts. These results suggest that the reaction mechanism of HTR over Pd-In/SiO2 may be similar to that over Pd/SiO2. On the other hand, at lower temperature region a new reaction pathway for NO-CO reaction may be emerging due to the addition of In or Pb. The appearance of the maximum in the reaction rate may be explained by the stability of the reaction intermediates of this new pathway or the stability of the active surface structures with temperatures. To study whether there are any structural changes between these temperature ranges, in situ XPS measurements were carried out, as summarized in Figs. 3B(b–d) and C(b–d), and 4A(b–d) and B(b–d). As mentioned already in the previous section, the binding energy of Pd3d5/2 of Pd-In and Pd-Pb catalysts shifted 0.8 eV to the higher binding energy side (335.6 eV) from the zero valent Pd metal value in the literature (335.0 eV) even after the reduction at 300 K. The binding energy of In3d5/2 in Pd-In/SiO2, however, was 0.1 eV lower than the zero valent In (443.7 eV) value, suggesting the
Fig. 6. Arrhenius plots of the formation rates of N2O + N2 in the low temperature region and in the high temperature region over Pd/SiO2 and Pd-In/SiO2 catalysts.
formation of Pd-In intermetallic compound. After the NO-CO reaction over Pd-In/SiO2 at 373 K, the Pd3d5/2 peak did not change much; however, the binding energy of In3d5/2 shifted 0.6 eV to the higher side, indicating the oxidation of surface In by NO. A similar situation was observed in the case of Pd-Pb catalysts: the Pb4f5/2 peak at 137.1 eVassignable to zerovalent Pb disappeared completely after NO-CO reaction at 373 K and only the PbO2 peak at 139.1 eV remained. When the reaction temperature was raised up to 573 K where the minimum reaction rates were observed, all of these XPS peaks are shifted to a lower binding energy side to a certain extent, indicating that HTR reaction takes place over somewhat reduced surfaces. To investigate the structural changes of the intermetallic catalyst during catalytic oxidation–reduction cycle, we oxidized freshly reduced Pd-In/SiO2 by NO at various temperatures, followed by the reduction of the oxidized catalyst with CO. Fig. 7A shows the amount of products obtained when only NO was exposed over the reduced Pd-In/ SiO2 surface for 4 h at various temperatures. Until 450 K the main product was N2O with a trace amount of N2, which may correspond to the oxidation of surface intermetallic compound to form oxidized Pd and In species. Above 450 K, N2 was formed linearly with temperatures corresponding to the oxidation of a bulk intermetallic compound forming PdO and In2O3 in the bulk. Fig. 7B represents the extent of reduced oxygen (reduction%) by CO for 4 h at each temperature, after the oxidation of freshly reduced Pd-In/SiO2 catalysts at the same temperature. Until 450 K, only 30% of the oxygen was removed from the oxidized catalyst, indicating that only surface Pd atoms were reduced, leaving In2O3 species on the surface. After that the reduction percent increased linearly with reduction temperatures up to 500 K and then again stayed constant value (about 80%) until 623 K. At this stage, bulk PdO may be reduced until Pd and In until the stoichiometry of In2O. At the last stage, all the oxygen was reduced by CO to form Pd0.48In0.52 intermetallic compound. It is interesting to note that the transient region between LTR and HTR reactions
Fig. 7. (A) Temperature dependence of the amount of products formed during NO decomposition over freshly reduced Pd-In/SiO2. (B) Temperature dependence of the reduction extents (%) by CO over the oxidized surface of Pd-In/ SiO2.
T. Hirano et al. / Applied Catalysis A: General 320 (2007) 91–97
Fig. 8. In situ XRD measurements of the Pd-In/SiO2 during the oxidation and reduction processes. (a) R.T., (b) H2 reduction at 733 K for 90 min, (c) H2 reduction at 733 K for 150 min, (d) NO oxidation at 733 K for 5 min, (e) NO oxidation at 733 K for 60 min, (f) CO reduction at 300 K for 30 min, (g) CO reduction at 400 K for 30 min, (h) CO reduction at 500 K for 30 min, (i) CO reduction at 600 K for 30 min and (j) CO reduction at 700 K for 30 min.
corresponds to the second stage where the redox cycle is operating with In(III) to In(I) as well as Pd(II) to Pd(0) species. The above mentioned oxidation–reduction processes were followed by in situ XRD spectroscopy as summarized in Fig. 8. Diffraction patterns (a–c) exhibit typical fcc peaks of Pd0.48In0.52 intermetallic compound, which oxidized completely by NO oxidation at 733 K for 1 h (e). Several peaks that emerged can be assigned to PdO (2u = 31.88, 45.68, 56.48) and In2O3 (2u = 30.68, 35.58, 51.08, 60.78), respectively. After that, the catalyst was reduced by CO at elevated temperatures, as shown in (f–j). At around 500–600 K, some new peaks emerged at 2u = 39.58, 67.58 and 82.08 which may be assigned to Pd metal and InO species, accompanied with the decrease of PdO and In2O3 species. Finally, original Pd0.48In0.52 intermetallic compound was obtained by CO reduction at 700 K.
95
Fig. 9. FT-IR spectra of adsorbed CO followed by TPD over: (A) Pd/SiO2, (B) Pd-In/SiO2 and (C) Pd-Pb/SiO2. (A): (a) CO adsorption at R.T., (b) evacuation at R.T., (c) 373 K, (d) 473 K and (e) 573 K. (B): (a) CO adsorption at R.T., (b) evacuation at R.T., (c) 323 K, (d) 373 K, (e) 423 K and (f) 473 K. (C): (a) CO adsorption at R.T. and (b) evacuation at R.T.
2040 cm 1 was very weak, probably because of the inhibition of CO adsorption by excess Pb oxide surface layers. Fig. 10 demonstrates the adsorbed NO over Pd/SiO2, Pd-In/ SiO2 and Pd-Pb/SiO2 catalysts, respectively. Two peaks at around 1725 and 1645 cm 1 can be assigned to linearly adsorbed NO(a) and bridged NO(a) species over Pd/SiO2; these peaks shifted to 1760 and 1680 cm 1 by the addition of In over Pd-In/SiO2. Desorption experiments after adsorption again indicated the weakening of NO adsorption strength by the addition of In. On the other hand, in the case of NO adsorption over Pd-Pb/SiO2, new adsorption peaks emerged at 1995 and 1890 cm 1 instead of the 1725 and 1645 cm 1 bands over Pd/ SiO2. These new peaks may be assigned to the positively charged NO(a) species suggesting some electronic interaction between Pb and NO molecules. Fig. 11 represents the FT-IR spectra of adsorbed species over Pd/SiO2 after NO-CO reaction at various temperatures,
3.3. Infrared spectroscopic study Fig. 9 demonstrates the FT-IR spectra of adsorbed CO over Pd/SiO2, Pd-In/SiO2 and Pd-Pb/SiO2 catalysts at room temperature, followed by TPD of adsorbed CO. Spectrum (a) in Fig. 9A shows the adsorbed CO on Pd/SiO2, where two peaks can be assigned to linear (2080 cm 1) and bridged (1984 cm 1) CO(a) species. On the other hand, over Pd-In/SiO2 the bridged CO(a) disappeared completely and only linearly adsorbed CO (2080 cm 1) was observed. This spectral change may be explained by the ensemble effect of In atoms on the surface of a Pd-In intermetallic compound. TPD experiments after adsorption revealed that over Pd-In/SiO2 linearly adsorbed CO disappeared at around 373–423 K, where most of the bridged CO(a) still remained over Pd/SiO2. From these results, we can conclude that one of the roles of added In would be the weakening of the adsorption strength of CO, which may enhance the reaction rate. This situation was pronounced even more in the case of PdPb/SiO2, as shown in Fig. 9C. The linearly adsorbed CO band at
Fig. 10. FT-IR spectra of adsorbed NO followed by TPD over (A) Pd/SiO2, (B) Pd-In/SiO2 and (C) Pd-Pb/SiO2. (A) (a) NO adsorption at R.T., (b) evacuation at R.T., (c) 323 K, (d) 373 K, (e) 423 K and (f) 473 K. (B) (a) NO adsorption at R.T. and (b) evacuation at R.T. and (C) (a) CO adsorption at R.T. and (b) evacuation at R.T.
96
T. Hirano et al. / Applied Catalysis A: General 320 (2007) 91–97
Fig. 11. FT-IR spectra of adsorbed species after NO-CO reaction for 2 h over Pd/SiO2 at (A) 373 K, (B) 473 K and (C) 573 K, followed by CO TPR from room temperature. (a) NO-CO reaction, (b) CO TPR at R.T., (c) 373 K, (d) 473 K and (e) 573 K.
Fig. 13. FT-IR spectra of adsorbed species after NO-CO reaction for 2 h over Pd-Pb/SiO2 at: (A) 373 K, (B) 473 K and (C) 573 K, followed by CO TPR from room temperature. (a) NO-CO reaction, (b) CO TPR at R.T., (c) 373 K, (d) 473 K and (e) 573 K.
followed by CO TPR of the adsorbed species from room temperature. As shown in Fig. 11A(a), after the NO-CO reaction at 373 K for 2 h, both CO(a) and NO(a) were observed at the positions similar to those for the independent adsorption (see Figs. 9A(a) and 10A(a)). After the evacuations of the gas phase, only CO was introduced at room temperature and the CO TPR experiment was carried out. A significant increase of bridged CO(a) band at 1984 cm 1 was observed, with the decrease of NO(a) bands at 1740 and 1671 cm 1. Similar spectra changes were observed in the cases of higher temperature reactions, although no adsorbed species were observed after 573 K reaction as shown in Fig. 11C(a). In the case of NO-CO reaction over Pd-In/SiO2 at 373 K, the situation was rather different from the independent adsorption experiment (Fig. 9(b)) and the intensities of bridged CO(a) and NO (a) species at 1973 and 1620 cm 1 were increased considerably compared to the linearly adsorbed ones, as shown in Fig. 12A(a). These results may suggest the decreasing of the ensemble effect of In on the surface of Pd-In intermetallic
compound, probably due to the partial oxidation of In. When the reduction temperature was raised, the intensity of the bridged peak decreased considerably with the increase of the linear peak, suggesting the reformation of the intermetallic compounds. Similar spectra changes were observed in the cases of higher temperature reactions, although no adsorbed species was observed after 573 K reaction as shown in Fig. 12C(a). Fig. 13 shows the species adsorbed during NO-CO reactions at various temperatures over Pd-Pb/SiO2. Similar to the individual adsorption, we could not observe any adsorbed CO during NO-CO reaction; instead, two characteristic bands were observed at 1870 and 2160 cm 1, which can be assigned to adsorbed NO and NCO species.
Fig. 12. FT-IR spectra of adsorbed species after NO-CO reaction for 2 h over Pd-In/SiO2 at: (A) 373 K, (B) 473 K and (C) 573 K, followed by CO TPR from room temperature. (a) NO-CO reaction (b) CO TPR at R.T. (c) 373 K (d) 473 K and (e) 573 K.
3.4. Reaction mechanism Two reaction mechanisms have been proposed for the N2O and N2 formations in NO-CO reaction. The first one involves the unimolecular dissociative adsorption of NO to form N(a) and O(a), followed by a rapid removal of O(a) by CO to form CO2 [17]. N2O is formed by the reaction of N(a) and NO(a), whereas N2 is formed by the recombination of two N(a). The second mechanism involves a direct bimolecular reaction between NO and CO to form N(a) and CO2 [18]. N2O and N2 may be formed by the same reaction scheme as mentioned above. To distinguish these two mechanisms, we introduced only NO onto freshly reduced Pd/SiO2 (618 K) and Pd-In/SiO2 (333 K), and we compared the (N2O + N2) formation rate with that in NO-CO reaction. In the case of Pd/SiO2, the initial rates of both processes were almost the same, indicating that the NOCO reaction involves an NO dissociation process. It is well known that the dissociative adsorption of NO on Pd is strongly structure-sensitive [5] and requires higher temperatures as 500 K [18], which is consistent with the present study. On the other hand, over Pd-In/SiO2 the latter rate was much faster than the former one, indicating that NO-CO reaction proceeds without NO dissociation. Accordingly our result may be explained by the bimolecular mechanism between NO and
T. Hirano et al. / Applied Catalysis A: General 320 (2007) 91–97
CO mentioned above, but reported activation energies for this process are much larger (80–120 kJ/mol) than that obtained in this study (32 kJ/mol). Another characteristic point of Pd-In/ SiO2 is that N2O formation from NO and CO is faster than N2 formation from N2O and CO with a smaller activation energy, which may suggest the involvement of a completely different reaction mechanism, such as N2O formation from NO dimer and CO by the assistance of In. As mentioned already, over PdIn/SiO2 and Pd-Pb/SiO2 catalysts the reaction rate exhibited a maximum and minimum at around 500 and 600K, respectively (see Fig. 5), and the activation energy was much smaller (32 kJ/ mol) in the LTR region compared to that in the HTR region(70 kJ/mol). In situ XPS as well as FTIR investigations suggest that in the former temperature region bimolecular or NO dimer-like intermediates may be predominant, whereas in the latter region unimolecular dissociative mechanism may be important. The role of added In and Pb is considered the stabilization of this unique reaction intermediates to open a facile reaction pathway for lower temperature NO reduction. 4. Conclusions (1) Formation of intermetallic compounds was observed in the cases of both Pd-In/SiO2 and Pd-Pb/SiO2 catalysts; such compounds caused the drastic enhancement of the reaction rate of N2O formation in NO-CO reaction. (2) The infrared analyses revealed the weakening of adsorption strength of CO on Pd metal by the formation of intermetallic compounds, which is one of the main reasons of the enhancement of reaction rate. (3) Over Pd/SiO2 catalyst, the reaction may proceed via redox mechanism through NO dissociation to form the oxidized surface, followed by its reduction with CO.
97
(4) By adding In or Pb, a new reaction pathway was opened at lower temperature region via NO dimer-like intermediate. The role of additives is the stabilization of these unique intermediates on the slightly oxidized intermetallic compound surfaces.
References [1] K.C. Taylor, Catal. Rev. Sci. Eng. 35 (1993) 457. [2] A. El Hamdaoui, G. Bergeret, J. Massardier, M. Primet, A. Renouprez, J. Catal. 148 (1994) 47. [3] C.T. Williams, A.A. Tolia, H.Y.H. Chan, C.G. Takoudis, M.J. Weaver, J. Catal. 163 (1996) 63. [4] M. Marwood, C.G. Vayenas, J. Catal. 170 (1997) 275. [5] D.R. Rainer, S.M. Vesecky, M. Koranne, W.S. Oh, D.W. Goodman, J. Catal. 167 (1997) 234. [6] H. Muraki, K. Yokota, Y. Fujitani, Appl. Catal. 48 (1989) 93. [7] S. Subramanian, R.J. Kudla, C.R. Peters, M.S. Chattha, Catal. Lett. 16 (1992) 323. [8] Y.J. Mergler, J. Hoebink, B.E. Nieuvenhuys, J. Catal. 167 (1997) 305. [9] J.F. Trillat, J. Massadier, B. Moraweck, H. Praliaud, A.J. Renouprez, J. Catal. 167 (1997) 103. [10] A. El Hamdaoui, G. Bergeret, J. Massardier, M. Primet, A.J. Renouprez, J. Catal. 148 (1994) 47. [11] C. Neyertz, M.A. Volpe, C. Gigola, Catal. Today 57 (3–4) (2000) 255. [12] M. Valden, L.R. Keiski, N. Xiang, J. Pere, J. Aaltonen, M. Pessa, T. Manula, A. Savimaki, A. Lahti, M. Harkonen, J. Catal. 161 (1996) 614. [13] F.B. Noronha, M.A.S. Baldanza, M. Schmal, J. Catal. 188 (1999) 270– 280. [14] M. Schmal, M.A.S. Baldanza, A. Vannice, J. Catal. 185 (1999) 138– 151. [15] M.A.S. Baldanza, L.F. De Mello, A. Vannace, F.B. Noronha, M. Schmal, J. Catal. 192 (1) (2000) 64. [16] M.A. Sica, E.C. Gigola, E. Carios, Appl. Catal. 239 (1–2) (2003) 121. [17] G. Xi, J. Bao, S. Shao, S. Li, J. Vac. Sci. Technol. A10 (1992) 2351. [18] A. Obuchi, S. Naito, T. Onishi, K. Tamaru, Surf. Sci. 112 (1982) 235.
The role of additives in the catalytic reduction of NO by CO over Pd-In/SiO2 and Pd-Pb/SiO2 catalysts Takashi Hirano, Yoshiyuki Ozawa, Takayuki Sekido, Takashi Ogino, Toshihiro Miyao, Shuichi Naito * Department of Material and Life Chemistry, Kanagawa University, 3-27-1, Rokkakubashi, Kanagawaku, Yokohama 221-8686, Japan Received 29 November 2006; received in revised form 18 December 2006; accepted 21 December 2006 Available online 3 January 2007
Abstract Addition effect of In and Pb on the NO-CO reaction over SiO2 supported Pd catalysts was studied, using a closed gas circulation system as well as in situ infrared, XRD and XPS spectroscopies. Formation of intermetallic compounds was observed in the cases of both Pd-In/SiO2 and Pd-Pb/ SiO2 catalysts. These compounds caused a drastic enhancement of the reaction rate of N2O formation. Infrared analyses revealed the weakening of CO adsorption on Pd metal by the formation of intermetallic compounds, which is the main reason of the enhancement of reaction rate. Over Pd/ SiO2, the reaction may proceed via redox mechanism through NO dissociation to form the oxidized surface, followed by its reduction with CO. By adding In or Pb, a new reaction pathway is opened at the lower temperature region via NO dimer-like intermediates. The role of additives is the stabilization of these unique intermediates on the slightly oxidized intermetallic compound surfaces. # 2007 Elsevier B.V. All rights reserved. Keywords: NO-CO reaction; Pd/SiO2 catalyst; Addition effect; Pd-In; Pd-Pb
1. Introduction In the catalytic conversion of NO by CO, rhodium is well known to be the most active of all precious metals. For the oxidation of CO, platinum is generally accepted to be the most efficient catalyst [1]. Palladium is the third component in the so called three-way Rh-Pt-Pd automotive catalysts. In the past decade, many studies have been performed to replace Rh in automotive catalysts in view of its high cost and scare resources. Among them, catalytic properties of Pd have received increasing attention because of its potentiality for the substitution of Rh in the conventional three-way Rh/Pt catalyst [2–5]. However, the Pd-only catalyst has a poor NO reduction efficiency compared to Rh, especially in an oxygen rich environment [6,7]. However, it could be a good basis for a new type of bimetallic system if its selectivity for NO reduction could be increased by the presence of a second metal such as Ag and Cr [8].
* Corresponding author. Tel.: +81 45 481 5661; fax: +81 45 491 7915. E-mail address: [email protected] (S. Naito). 0926-860X/$ – see front matter # 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.apcata.2006.12.015
Attempts to improve the catalytic performance of Pd have also been made by placing it in contact with various base metal oxides and rare earth oxides. This was successfully achieved with Pd promoted by cobalt oxide [8] or with Pd-Mn [9] or PdCr [10] couples, or with Pd-VOx/Al2O3 catalysts [11], which have shown light-off temperatures of conversion lower by 80 to 100 K compared with that of pure Pd. In the case of Pd-Mn, the bimetallic cluster of Pd3Mn was reported to be the active species for this reaction and adsorbed NO with a long N-O bond and quasi-parallel to the surface was a precursor for the dissociation. In the case of Pd-VOx, the reduction of VOx species was enhanced by a close contact with Pd; partially reduced vanadia decorated the noble metal particles and enhanced NO decomposition. Valden et al. presented an interesting paper using Pd/La2O3/ Al2O3 catalysts in which the reactivity between NO and CO and their adsorption was studied by TPD and IR [12]. The latter technique indicates different behavior, since NO was adsorbed molecularly on the support and lanthanum oxide, while NO dissociation followed by the formation of N2 and N2O was seen on all Pd catalysts. Pd catalysts supported on Al2O3, La2O3Al2O3 and LaAlO3 were prepared: these contain different amounts of Pd. CO adsorbed molecularly on all Pd catalysts
92
T. Hirano et al. / Applied Catalysis A: General 320 (2007) 91–97
forming linear and bridged CO species. The presence of PdO was found to decrease the CO binding energy on Al2O3supported Pd catalysts. The destabilization of CO bonding was even further enhanced on Pd catalysts supported on LaAlO3 and La2O3-Al2O3. In contrast to the negligible interaction between CO and supports, NO was observed to adsorb molecularly on the supports. It was reported that the addition of MoO3 to Pd/Al2O3 catalysts improved the NO activity with high selectivity to N2 in the presence of a small excess amount of oxygen [13]. The addition of 8 wt% molybdenum covered the surface of alumina and led to the disappearance of surface hydroxyls [14,15]. The MoO3 addition also induced modifications on the metallic phase, which causes the dilution or blocking of Pd surface by molybdenum oxide species. The catalytic behavior could be explained by a bi-functional mechanism. After adsorption and dissociation of NO on Mo4+, the dissociated oxygen is transferred to the Pd. The CO adsorbed on the Pd surface reacts with oxygen. A similar effect was observed in the cases of Pd-WOx/Al2O3 catalysts [16]. In the present study, we investigated the effect of adding In and Pb to silica supported Pd in the reduction of NO by CO: we found that these additives exhibit a large acceleration effect. Especially in the cases of Pd-In/SiO2 and Pd-Pb/SiO2, an intermetallic compound formed between Pd-In is extraordinary active for NO-CO reaction, and N2O and CO2 are formed steadily even at room temperature. We applied in situ infrared, XPS and XRD analyses to elucidate the role of additives in the tremendous enhancement effect in NO-CO reactions. 2. Experimental Silica supported Pd, Pd-In, and Pd-Pb (5 wt% Pd) catalysts were prepared by a conventional co-impregnation method using SiO2 (Aerosil 300) and (NH4)2PdCl4, InCl34H2O and PbCl4 (Wako chemicals) as precursors. The molar ratio of Pd and additive was 1:1. The catalyst (0.2 g) was reduced by hydrogen at 733 K for 10 h to form intermetallic compounds. The NOCO reaction was carried out in a closed gas circulation system under a 1:1 ratio of NO and CO (4 kPa each). The composition of the gas phase during the reaction was followed by TCD gas chromatography (Shimazu GC-17A). For infrared spectroscopic experiments, the catalyst was pressed into a 20 mm diameter disk and put into an infrared cell, which was connected to a closed gas circulation system. After the pretreatment mentioned above, the adsorbed species were measured by IR spectrometry (JEOL, FT-IR, Diamond 20). A transmission electron microscope (JEM2010, JEOL) with an acceleration voltage of 2000 kV and LaB6 cathode was applied for the observation of the images of supported catalysts. A X-ray photoelectron spectroscope (JPS-9010, JEOL) with a Mg Ka X-ray source (10 kV, 10 mA) was applied for the analysis of the electronic states of supported catalysts. Samples were prepared by molding in thin disk shapes. Each was introduced into the preparation chamber. After reduction by H2 at 573 K, each sample was transferred into analysis chamber without exposure to air. The catalysts were also characterized
by XRD (Rigaku, RAD-gX) and H2 and CO chemisorption method (Beckman Coulter, Omnisorp 100CX). 3. Results and discussion 3.1. Characterization of catalysts Fig. 1 shows the XRD patterns of Pd/SiO2, Pd-In/SiO2 and Pd-Pb/SiO2 catalysts after H2 reduction at 733 K for 10 h. In the case of Pd/SiO2 the diffraction pattern typical for FCC metal was observed as shown in A (2u = 40.0, 46.5, 68.7 and 82.0). When In was added, four characteristic peaks were shifted as shown in B (2u = 39.2, 56.8, 70.9, 84.2), which can be assigned to Pd0.48 In0.52 intermetallic compound. In the case of Pd-Pb/ SiO2, new XRD peaks emerged at 2u = 39.0, 40.5, 57.5, 65.0 and 71.0 by higher temperature reduction, which can be assigned to Pd3Pb2 intermetallic compound, as shown in Fig. 1C. Fig. 2 shows the TEM images of these catalysts. The particle sizes of Pd metal without additives were rather big (more than 10 nm) after 733 K reduction, but these became very small (2–3 nm) by the addition of In, forming Pd0.48 In0.52 intermetallic compounds. In the case of Pd-Pb/SiO2, binary distribution of Pd particles were observed; the particles were around 5–6 and 10–15 nm in diameters and both contained Pd and Pb, as determined by EDS analysis. Figs. 3 and 4 represent the XPS spectra of Pd3d, In3d and Pb4f transitions in Pd/SiO2, Pd-In/SiO2 and Pd-Pb/SiO2 catalysts after the reduction at 573 K in the in situ cell (spectrum (a) in each figure). In the case of Pd/SiO2, the binding energy of Pd3d5/2 was close to the zero valent Pd metal value in the literature (335.0 eV), whereas by adding In or Pb it shifted 0.8 eV higher binding energy side (335.6 eV) even after the reduction at higher temperatures. On the other hand, the binding energy of In3d5/2 in Pd-In/SiO2 was 0.1 eV lower than the zero valent In (443.7 eV) value, suggesting the formation of Pd-In intermetallic compound as shown in Fig. 4A(a). In the
Fig. 1. XRD patterns of Pd/SiO2, Pd-In/SiO2 and Pd-Pb/SiO2 catalysts after reduction at 733 K: (A) Pd/SiO2, (B) Pd-In/SiO2 and (C) Pd-Pb/SiO2.
T. Hirano et al. / Applied Catalysis A: General 320 (2007) 91–97
93
Fig. 2. TEM images of various catalysts: (A) Pd/SiO2, (B) Pd-In/SiO2 and (C) Pd-Pb/SiO2.
case of Pd-Pb/SiO2 catalyst, two Pb4f7/2 transitions were observed, at 137.1 and 139.1 eV, which can be assigned to zero valent and positively charged Pb species (Fig. 4B(a)). As shown in Fig. 1C, Pd3Pb2 intermetallic compound must be formed together with the oxidation of excess Pb to PbO2. 3.2. Kinetic study of NO-CO reaction
Fig. 3. XPS spectra of Pd3d transition in: (A) Pd/SiO2, (B) Pd-In/SiO2 and (C) Pd-Pb/SiO2 after various treatments in the in situ cell. (a) H2 reduction at 373 K, (b) NO-CO reaction at 373 K, (c) NO-CO reaction at 473 K and (d) NO-CO reaction at 573 K.
Fig. 4. XPS spectra of In3d and Pb4f transitions in: (A) Pd/SiO2, (B) Pd-In/ SiO2 and (C) Pd-Pb/SiO2 after various treatments in the in situ cell. (a) H2 reduction at 373 K, (b) NO-CO reaction at 373 K, (c) NO-CO reaction at 473 K and (d) NO-CO reaction at 573 K.
Fig. 5 summarizes the temperature dependences of the rates of (N2O + N2) formation in NO-CO reactions over Pd/SiO2, PdIn/SiO2 and Pd-Pb/SiO2 catalysts. As shown in A, the reaction started only at around 500–600 K over Pd/SiO2, and the main product was N2O with a small amount of N2. When In or Pb was added to this system, the reaction temperature was lowered more than 200 K and only N2O and CO2 were formed even at room temperature. Above 373 K, a small amount of N2 started to be formed together with N2O, which became the main product above 473 K. As shown in Fig. 5B and C, the dependence of the N2O + N2 formation rates upon the temperatures was very unique over Pd-In/SiO2 and Pd-Pb/ SiO2 catalysts. First of all, it increased linearly up to 473 K,
Fig. 5. Temperature dependences of the formation rates of N2O + N2 in NOCO reaction over: (A) Pd/SiO2, (B) Pd-In/SiO2 and (C) Pd-Pb/SiO2. Solid line(r): raising the reaction temperature, broken line(l): lowering the reaction temperature.
94
T. Hirano et al. / Applied Catalysis A: General 320 (2007) 91–97
showing a maximum at around 473–500 K, then decreased gradually up to 573 K, showing a minimum at around 573– 600 K, and then increased again. To see whether this behavior is reversible with temperatures or not, we have compared the temperature-increasing reaction with the temperature-decreasing reaction. After the reaction at 673 K, the temperature was lowered reversibly as shown by the dotted lines in Fig. 5. In the case of Pd/SiO2 almost the same temperature profile was obtained in A. On the other hand, over Pd-In/SiO2 the reaction rate was decreased seriously by the deactivation process at higher temperature reaction, however, it showed a similar temperature dependence to that in the case of temperatureincreasing reaction. Over Pd-Pb/SiO2, although the maximum point shifted to higher temperatures, it still showed a similar behavior to that in the case of Pd-In catalyst. Fig. 6 represents the comparison of Arrhenius plots of (N2O + N2) formation rates over Pd/SiO2 together with (N2O + N2) formation rates between the temperature ranges at 373–473 K (designated as LTR) and at 573–673 K (HTR) over Pd-In/SiO2 catalysts. From the slopes of the Arrhenius plots, the activation energies were estimated to be as follows; 70 kJ/mol for (N2O + N2) formation rate in the case of Pd/SiO2, 32 kJ/mol for the LTR over Pd-In/SiO2, and 66 kJ/mol for the HTR over Pd-In/SiO2 catalysts. These results suggest that the reaction mechanism of HTR over Pd-In/SiO2 may be similar to that over Pd/SiO2. On the other hand, at lower temperature region a new reaction pathway for NO-CO reaction may be emerging due to the addition of In or Pb. The appearance of the maximum in the reaction rate may be explained by the stability of the reaction intermediates of this new pathway or the stability of the active surface structures with temperatures. To study whether there are any structural changes between these temperature ranges, in situ XPS measurements were carried out, as summarized in Figs. 3B(b–d) and C(b–d), and 4A(b–d) and B(b–d). As mentioned already in the previous section, the binding energy of Pd3d5/2 of Pd-In and Pd-Pb catalysts shifted 0.8 eV to the higher binding energy side (335.6 eV) from the zero valent Pd metal value in the literature (335.0 eV) even after the reduction at 300 K. The binding energy of In3d5/2 in Pd-In/SiO2, however, was 0.1 eV lower than the zero valent In (443.7 eV) value, suggesting the
Fig. 6. Arrhenius plots of the formation rates of N2O + N2 in the low temperature region and in the high temperature region over Pd/SiO2 and Pd-In/SiO2 catalysts.
formation of Pd-In intermetallic compound. After the NO-CO reaction over Pd-In/SiO2 at 373 K, the Pd3d5/2 peak did not change much; however, the binding energy of In3d5/2 shifted 0.6 eV to the higher side, indicating the oxidation of surface In by NO. A similar situation was observed in the case of Pd-Pb catalysts: the Pb4f5/2 peak at 137.1 eVassignable to zerovalent Pb disappeared completely after NO-CO reaction at 373 K and only the PbO2 peak at 139.1 eV remained. When the reaction temperature was raised up to 573 K where the minimum reaction rates were observed, all of these XPS peaks are shifted to a lower binding energy side to a certain extent, indicating that HTR reaction takes place over somewhat reduced surfaces. To investigate the structural changes of the intermetallic catalyst during catalytic oxidation–reduction cycle, we oxidized freshly reduced Pd-In/SiO2 by NO at various temperatures, followed by the reduction of the oxidized catalyst with CO. Fig. 7A shows the amount of products obtained when only NO was exposed over the reduced Pd-In/ SiO2 surface for 4 h at various temperatures. Until 450 K the main product was N2O with a trace amount of N2, which may correspond to the oxidation of surface intermetallic compound to form oxidized Pd and In species. Above 450 K, N2 was formed linearly with temperatures corresponding to the oxidation of a bulk intermetallic compound forming PdO and In2O3 in the bulk. Fig. 7B represents the extent of reduced oxygen (reduction%) by CO for 4 h at each temperature, after the oxidation of freshly reduced Pd-In/SiO2 catalysts at the same temperature. Until 450 K, only 30% of the oxygen was removed from the oxidized catalyst, indicating that only surface Pd atoms were reduced, leaving In2O3 species on the surface. After that the reduction percent increased linearly with reduction temperatures up to 500 K and then again stayed constant value (about 80%) until 623 K. At this stage, bulk PdO may be reduced until Pd and In until the stoichiometry of In2O. At the last stage, all the oxygen was reduced by CO to form Pd0.48In0.52 intermetallic compound. It is interesting to note that the transient region between LTR and HTR reactions
Fig. 7. (A) Temperature dependence of the amount of products formed during NO decomposition over freshly reduced Pd-In/SiO2. (B) Temperature dependence of the reduction extents (%) by CO over the oxidized surface of Pd-In/ SiO2.
T. Hirano et al. / Applied Catalysis A: General 320 (2007) 91–97
Fig. 8. In situ XRD measurements of the Pd-In/SiO2 during the oxidation and reduction processes. (a) R.T., (b) H2 reduction at 733 K for 90 min, (c) H2 reduction at 733 K for 150 min, (d) NO oxidation at 733 K for 5 min, (e) NO oxidation at 733 K for 60 min, (f) CO reduction at 300 K for 30 min, (g) CO reduction at 400 K for 30 min, (h) CO reduction at 500 K for 30 min, (i) CO reduction at 600 K for 30 min and (j) CO reduction at 700 K for 30 min.
corresponds to the second stage where the redox cycle is operating with In(III) to In(I) as well as Pd(II) to Pd(0) species. The above mentioned oxidation–reduction processes were followed by in situ XRD spectroscopy as summarized in Fig. 8. Diffraction patterns (a–c) exhibit typical fcc peaks of Pd0.48In0.52 intermetallic compound, which oxidized completely by NO oxidation at 733 K for 1 h (e). Several peaks that emerged can be assigned to PdO (2u = 31.88, 45.68, 56.48) and In2O3 (2u = 30.68, 35.58, 51.08, 60.78), respectively. After that, the catalyst was reduced by CO at elevated temperatures, as shown in (f–j). At around 500–600 K, some new peaks emerged at 2u = 39.58, 67.58 and 82.08 which may be assigned to Pd metal and InO species, accompanied with the decrease of PdO and In2O3 species. Finally, original Pd0.48In0.52 intermetallic compound was obtained by CO reduction at 700 K.
95
Fig. 9. FT-IR spectra of adsorbed CO followed by TPD over: (A) Pd/SiO2, (B) Pd-In/SiO2 and (C) Pd-Pb/SiO2. (A): (a) CO adsorption at R.T., (b) evacuation at R.T., (c) 373 K, (d) 473 K and (e) 573 K. (B): (a) CO adsorption at R.T., (b) evacuation at R.T., (c) 323 K, (d) 373 K, (e) 423 K and (f) 473 K. (C): (a) CO adsorption at R.T. and (b) evacuation at R.T.
2040 cm 1 was very weak, probably because of the inhibition of CO adsorption by excess Pb oxide surface layers. Fig. 10 demonstrates the adsorbed NO over Pd/SiO2, Pd-In/ SiO2 and Pd-Pb/SiO2 catalysts, respectively. Two peaks at around 1725 and 1645 cm 1 can be assigned to linearly adsorbed NO(a) and bridged NO(a) species over Pd/SiO2; these peaks shifted to 1760 and 1680 cm 1 by the addition of In over Pd-In/SiO2. Desorption experiments after adsorption again indicated the weakening of NO adsorption strength by the addition of In. On the other hand, in the case of NO adsorption over Pd-Pb/SiO2, new adsorption peaks emerged at 1995 and 1890 cm 1 instead of the 1725 and 1645 cm 1 bands over Pd/ SiO2. These new peaks may be assigned to the positively charged NO(a) species suggesting some electronic interaction between Pb and NO molecules. Fig. 11 represents the FT-IR spectra of adsorbed species over Pd/SiO2 after NO-CO reaction at various temperatures,
3.3. Infrared spectroscopic study Fig. 9 demonstrates the FT-IR spectra of adsorbed CO over Pd/SiO2, Pd-In/SiO2 and Pd-Pb/SiO2 catalysts at room temperature, followed by TPD of adsorbed CO. Spectrum (a) in Fig. 9A shows the adsorbed CO on Pd/SiO2, where two peaks can be assigned to linear (2080 cm 1) and bridged (1984 cm 1) CO(a) species. On the other hand, over Pd-In/SiO2 the bridged CO(a) disappeared completely and only linearly adsorbed CO (2080 cm 1) was observed. This spectral change may be explained by the ensemble effect of In atoms on the surface of a Pd-In intermetallic compound. TPD experiments after adsorption revealed that over Pd-In/SiO2 linearly adsorbed CO disappeared at around 373–423 K, where most of the bridged CO(a) still remained over Pd/SiO2. From these results, we can conclude that one of the roles of added In would be the weakening of the adsorption strength of CO, which may enhance the reaction rate. This situation was pronounced even more in the case of PdPb/SiO2, as shown in Fig. 9C. The linearly adsorbed CO band at
Fig. 10. FT-IR spectra of adsorbed NO followed by TPD over (A) Pd/SiO2, (B) Pd-In/SiO2 and (C) Pd-Pb/SiO2. (A) (a) NO adsorption at R.T., (b) evacuation at R.T., (c) 323 K, (d) 373 K, (e) 423 K and (f) 473 K. (B) (a) NO adsorption at R.T. and (b) evacuation at R.T. and (C) (a) CO adsorption at R.T. and (b) evacuation at R.T.
96
T. Hirano et al. / Applied Catalysis A: General 320 (2007) 91–97
Fig. 11. FT-IR spectra of adsorbed species after NO-CO reaction for 2 h over Pd/SiO2 at (A) 373 K, (B) 473 K and (C) 573 K, followed by CO TPR from room temperature. (a) NO-CO reaction, (b) CO TPR at R.T., (c) 373 K, (d) 473 K and (e) 573 K.
Fig. 13. FT-IR spectra of adsorbed species after NO-CO reaction for 2 h over Pd-Pb/SiO2 at: (A) 373 K, (B) 473 K and (C) 573 K, followed by CO TPR from room temperature. (a) NO-CO reaction, (b) CO TPR at R.T., (c) 373 K, (d) 473 K and (e) 573 K.
followed by CO TPR of the adsorbed species from room temperature. As shown in Fig. 11A(a), after the NO-CO reaction at 373 K for 2 h, both CO(a) and NO(a) were observed at the positions similar to those for the independent adsorption (see Figs. 9A(a) and 10A(a)). After the evacuations of the gas phase, only CO was introduced at room temperature and the CO TPR experiment was carried out. A significant increase of bridged CO(a) band at 1984 cm 1 was observed, with the decrease of NO(a) bands at 1740 and 1671 cm 1. Similar spectra changes were observed in the cases of higher temperature reactions, although no adsorbed species were observed after 573 K reaction as shown in Fig. 11C(a). In the case of NO-CO reaction over Pd-In/SiO2 at 373 K, the situation was rather different from the independent adsorption experiment (Fig. 9(b)) and the intensities of bridged CO(a) and NO (a) species at 1973 and 1620 cm 1 were increased considerably compared to the linearly adsorbed ones, as shown in Fig. 12A(a). These results may suggest the decreasing of the ensemble effect of In on the surface of Pd-In intermetallic
compound, probably due to the partial oxidation of In. When the reduction temperature was raised, the intensity of the bridged peak decreased considerably with the increase of the linear peak, suggesting the reformation of the intermetallic compounds. Similar spectra changes were observed in the cases of higher temperature reactions, although no adsorbed species was observed after 573 K reaction as shown in Fig. 12C(a). Fig. 13 shows the species adsorbed during NO-CO reactions at various temperatures over Pd-Pb/SiO2. Similar to the individual adsorption, we could not observe any adsorbed CO during NO-CO reaction; instead, two characteristic bands were observed at 1870 and 2160 cm 1, which can be assigned to adsorbed NO and NCO species.
Fig. 12. FT-IR spectra of adsorbed species after NO-CO reaction for 2 h over Pd-In/SiO2 at: (A) 373 K, (B) 473 K and (C) 573 K, followed by CO TPR from room temperature. (a) NO-CO reaction (b) CO TPR at R.T. (c) 373 K (d) 473 K and (e) 573 K.
3.4. Reaction mechanism Two reaction mechanisms have been proposed for the N2O and N2 formations in NO-CO reaction. The first one involves the unimolecular dissociative adsorption of NO to form N(a) and O(a), followed by a rapid removal of O(a) by CO to form CO2 [17]. N2O is formed by the reaction of N(a) and NO(a), whereas N2 is formed by the recombination of two N(a). The second mechanism involves a direct bimolecular reaction between NO and CO to form N(a) and CO2 [18]. N2O and N2 may be formed by the same reaction scheme as mentioned above. To distinguish these two mechanisms, we introduced only NO onto freshly reduced Pd/SiO2 (618 K) and Pd-In/SiO2 (333 K), and we compared the (N2O + N2) formation rate with that in NO-CO reaction. In the case of Pd/SiO2, the initial rates of both processes were almost the same, indicating that the NOCO reaction involves an NO dissociation process. It is well known that the dissociative adsorption of NO on Pd is strongly structure-sensitive [5] and requires higher temperatures as 500 K [18], which is consistent with the present study. On the other hand, over Pd-In/SiO2 the latter rate was much faster than the former one, indicating that NO-CO reaction proceeds without NO dissociation. Accordingly our result may be explained by the bimolecular mechanism between NO and
T. Hirano et al. / Applied Catalysis A: General 320 (2007) 91–97
CO mentioned above, but reported activation energies for this process are much larger (80–120 kJ/mol) than that obtained in this study (32 kJ/mol). Another characteristic point of Pd-In/ SiO2 is that N2O formation from NO and CO is faster than N2 formation from N2O and CO with a smaller activation energy, which may suggest the involvement of a completely different reaction mechanism, such as N2O formation from NO dimer and CO by the assistance of In. As mentioned already, over PdIn/SiO2 and Pd-Pb/SiO2 catalysts the reaction rate exhibited a maximum and minimum at around 500 and 600K, respectively (see Fig. 5), and the activation energy was much smaller (32 kJ/ mol) in the LTR region compared to that in the HTR region(70 kJ/mol). In situ XPS as well as FTIR investigations suggest that in the former temperature region bimolecular or NO dimer-like intermediates may be predominant, whereas in the latter region unimolecular dissociative mechanism may be important. The role of added In and Pb is considered the stabilization of this unique reaction intermediates to open a facile reaction pathway for lower temperature NO reduction. 4. Conclusions (1) Formation of intermetallic compounds was observed in the cases of both Pd-In/SiO2 and Pd-Pb/SiO2 catalysts; such compounds caused the drastic enhancement of the reaction rate of N2O formation in NO-CO reaction. (2) The infrared analyses revealed the weakening of adsorption strength of CO on Pd metal by the formation of intermetallic compounds, which is one of the main reasons of the enhancement of reaction rate. (3) Over Pd/SiO2 catalyst, the reaction may proceed via redox mechanism through NO dissociation to form the oxidized surface, followed by its reduction with CO.
97
(4) By adding In or Pb, a new reaction pathway was opened at lower temperature region via NO dimer-like intermediate. The role of additives is the stabilization of these unique intermediates on the slightly oxidized intermetallic compound surfaces.
References [1] K.C. Taylor, Catal. Rev. Sci. Eng. 35 (1993) 457. [2] A. El Hamdaoui, G. Bergeret, J. Massardier, M. Primet, A. Renouprez, J. Catal. 148 (1994) 47. [3] C.T. Williams, A.A. Tolia, H.Y.H. Chan, C.G. Takoudis, M.J. Weaver, J. Catal. 163 (1996) 63. [4] M. Marwood, C.G. Vayenas, J. Catal. 170 (1997) 275. [5] D.R. Rainer, S.M. Vesecky, M. Koranne, W.S. Oh, D.W. Goodman, J. Catal. 167 (1997) 234. [6] H. Muraki, K. Yokota, Y. Fujitani, Appl. Catal. 48 (1989) 93. [7] S. Subramanian, R.J. Kudla, C.R. Peters, M.S. Chattha, Catal. Lett. 16 (1992) 323. [8] Y.J. Mergler, J. Hoebink, B.E. Nieuvenhuys, J. Catal. 167 (1997) 305. [9] J.F. Trillat, J. Massadier, B. Moraweck, H. Praliaud, A.J. Renouprez, J. Catal. 167 (1997) 103. [10] A. El Hamdaoui, G. Bergeret, J. Massardier, M. Primet, A.J. Renouprez, J. Catal. 148 (1994) 47. [11] C. Neyertz, M.A. Volpe, C. Gigola, Catal. Today 57 (3–4) (2000) 255. [12] M. Valden, L.R. Keiski, N. Xiang, J. Pere, J. Aaltonen, M. Pessa, T. Manula, A. Savimaki, A. Lahti, M. Harkonen, J. Catal. 161 (1996) 614. [13] F.B. Noronha, M.A.S. Baldanza, M. Schmal, J. Catal. 188 (1999) 270– 280. [14] M. Schmal, M.A.S. Baldanza, A. Vannice, J. Catal. 185 (1999) 138– 151. [15] M.A.S. Baldanza, L.F. De Mello, A. Vannace, F.B. Noronha, M. Schmal, J. Catal. 192 (1) (2000) 64. [16] M.A. Sica, E.C. Gigola, E. Carios, Appl. Catal. 239 (1–2) (2003) 121. [17] G. Xi, J. Bao, S. Shao, S. Li, J. Vac. Sci. Technol. A10 (1992) 2351. [18] A. Obuchi, S. Naito, T. Onishi, K. Tamaru, Surf. Sci. 112 (1982) 235.