Journal of Molecular Catalysis A: Chemical 204–205 (2003) 683–691 Interaction of molecular hydrogen with three-way catalyst model of Pt/Ce0.6Zr0.4 O2 /Al2 O3 type P. Fornasiero a , J. Kaspar a , T. Montini a , M. Graziani a,∗ , V. Dal Santo b , R. Psaro b , S. Recchia c a c Dipartimento di Scienze Chimiche, Via Giorgieri 1, Università di Trieste, 34127 Trieste, Italy b Istituto di Scienze e Tecnologie Molecolari, CNR, Via C. Golgi 19, 20133 Milano, Italy Dipartimento di Scienze CC·FF·MM·Università dell’Insubria, Via Valleggio 11, 22100 Como, Italy Received 13 September 2002; received in revised form 8 January 2003; accepted 20 January 2003 Dedicated to Professor Renato Ugo on the occasion of his 65th birthday Abstract The interaction of H2 with a composite Pt/Ce0.6 Zr0.4 O2 /Al2 O3 system and its redox properties are investigated by means of H2 and O2 chemisorption, temperature programmed reduction (TPR) and desorption techniques. It is shown that the high H2 spillover capabilities of Pt/CeO2 -ZrO2 are maintained even after supporting the system on Al2 O3 . The re-oxidation experiments performed on the reduced moieties show that the stability of Ce(III) species is increased by Al2 O3 , while the presence of Pt favours the re-oxidation. © 2003 Elsevier Science B.V. All rights reserved. Keywords: Ceria-zirconia mixed oxides; Hydrogen spillover; Alumina support; Three-way catalyst; Platinum catalyst 1. Introduction CeO2 and more recently Cex Zr1−x O2 mixed oxides are a key component in three-way catalysts due to their redox Ce4+ /Ce3+ activity responsible for the oxygen storage/release capacity (OSC) [1]. Cex Zr1−x O2 also promote processes related to H2 production such as hydrocarbon reforming and partial oxidation, etc. [2,3]. The favourable effects of Cex Zr1−x O2 are related to both the ability to provide mobile oxygen species and the great affinity of the M/CeO2 (M = metal) systems towards H2 [4]. The adsorption of H2 over NM/CeO2 (NM = noble metal) is reversible at room temperature (RT), leading to generation of Ce3+ species which are ∗ Corresponding author. Fax: +39-040-5583903. E-mail address: graziani@units.it (M. Graziani). re-oxidised by simple evacuation [5]. Increasing the temperature of H2 adsorption leads to the so-called irreversible reduction, i.e. generation of oxygen vacancies and water [5]. While the beneficial effects of Al2 O3 on thermal stability of the CeO2 -ZrO2 system are known [6], H2 adsorption and reactivity with the NM/CeO2 -ZrO2 /Al2 O3 system has received scarce attention, despite their correlation to the catalytic properties [7,8]. 2. Experimental Ce0.6 Zr0.4 O2 (13 wt.%)/␥-Al2 O3 (indicated as CZ60/Al2 O3 ) was prepared using a citrate method [6]. CZ60/Al2 O3 and Al2 O3 were impregnated to incipient wetness with Pt(NH3 )2 (NO3 )2 and calcined at 500 ◦ C 1381-1169/$ – see front matter © 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S1381-1169(03)00352-2 684 P. Fornasiero et al. / Journal of Molecular Catalysis A: Chemical 204–205 (2003) 683–691 for 5 h. Pt content (1.5 and 0.87 wt.%, respectively for Pt/CZ60/Al2 O3 and Pt/Al2 O3 ) was determined by atomic emission spectroscopy (ICP-AES) on a Jobin Yvon (series JY24) instrument after extraction of Pt with aqua regia. Temperature programmed reduction (TPR) was carried out as previously reported using, however, 200 mg of sample to enhance the sensitivity [9]. H2 chemisorption and BET surface area were measured on a Micromeritics ASAP 2000 analyser. The samples (ca. 0.5 g) were first cleaned in flow of O2 (5%)/He for 1 h at 500 ◦ C and then reduced in a flow of H2 (5%)/Ar (25 ml min−1 ) at a heating rate of 10 ◦ C min−1 up to the selected temperature (350, 500 and 800 ◦ C). After 2 h at this temperature, the samples were evacuated at 400 ◦ C for 4 h and cooled under vacuum to the adsorption temperature (−80 or 25 ◦ C). A pressure change of less then 0.01% for 11 consecutive readings taken at 30 s intervals was used as an equi- librium criterion. TPD experiments were conducted on a Micromeritics Pulse Chemisorb 2700 apparatus. Samples were calcined in O2 flow (50 ml min−1 ) at 400 ◦ C for 1 h, purged in Ar flow (50 ml min−1 ) for 1 h at 350 ◦ C, cooled down to −70 ◦ C and then reduced (H2 (8%)/Ar at 15 ml min−1 , heating rate 8 ◦ C min−1 to 500 ◦ C). After cooling down in pure H2 (50 ml min−1 ) to −70 ◦ C, TPD was performed in Ar flow (12 ml min−1 ), heating rate 16 ◦ C min−1 to 500 ◦ C. 3. Results and discussion 3.1. Volumetric H2 chemisorption The results of the H2 chemisorption on the Pt/CZ60/Al2 O3 and Pt/Al2 O3 are reported in Table 1. The comparison of the data obtained at −80 and Table 1 H2 chemisorption and BET surface areas measured over Pt/Ce0.6 Zr0.4 O2 /Al2 O3 and Pt/Al2 O3 Run 1 2 3 4 5 6 7 8 9 10 Sample 1.5% Pt/Ce0.6 Zr0.4 O2 (13%)/Al2 O3 Tox a (◦ C) a BET (m2 g−1 ) H/Ptc −80 ◦ C Adsorbed volumed (time)e (ml g−1 (min)) 25 ◦ C −80 ◦ C 25 ◦ C 0 Torrh 400 Torr 0 Torr 400 Torr – – 350f – 75f 300f 150f 350g 500g 600g 350 500 350 800 – – 350 350 350 350 150 150 150 145 145 145 145 144 145 144 0.57 0.49 0.57 0.16 0.48 0.29 0.27 0.34 0.43 0.39 0.99 0.58 0.77 0.23 1.67 2.61 0.27 0.40 0.49 0.55 0.49 0.42 0.49 0.14 0.41 0.25 0.23 0.29 0.37 0.34 (37 min) (23 min) (35 min) (12 min) (31 min) (223 min) (14 min) (18 min) (19 min) (22 min) 0.92 0.78 0.94 0.20 0.84 0.89 0.75 0.75 0.65 0.66 0.85 0.50 0.66 0.20 1.44 2.25 0.23 0.35 0.42 0.47 (146 min) (57 min) (85 min) (14 min) (198 min) (671 min) (24 min) (23 min) (43 min) (61 min) 1.01 0.66 0.80 0.54 1.61 2.50 0.47 0.55 0.53 0.60 350f – 350f 350 500 350 800 350 186 185 185 171 170 0.68 0.69 0.69 0.54 0.58 0.69 0.69 0.69 0.50 0.54 0.34 0.35 0.35 0.27 0.29 (22 min) (17 min) (23 min) (17 min) (15 min) 0.57 0.60 0.60 0.60 0.33 0.35 0.35 0.35 0.25 0.27 (42 min) (35 min) (37 min) (28 min) (26 min) 0.52 0.40 0.42 0.30 0.31 0.87% Pt/Al2 O3 11 12 13 14 Tred b (◦ C) Tox : temperature at which the sample has been re-oxidised after the previous run for 2 h in flow of 5% O2 in He. Tred : temperature at which the sample has been pre-reduced for 2 h in flow of 5% H2 in He. The sample has been evacuated at 400 ◦ C for 4 h before H2 chemisorption experiments. c H chemisorption experiments carried out at the reported temperature. 2 d Adsorption of H obtained by back extrapolating the linear part of the adsorption isotherm in the region 2–20 Torr (0 Torr) and 2 cumulative adsorption value at 400 Torr. e Time required for reaching the equilibrium at 2 Torr of pressure. f Sample from previous run. g Sample previously reduced at 800 ◦ C. h 1 Torr = 133.332 Pa. b P. Fornasiero et al. / Journal of Molecular Catalysis A: Chemical 204–205 (2003) 683–691 25 ◦ C reveals the presence of H2 spillover effects. Spilling of H2 over the support is an activated process, accordingly by lowering the adsorption temperature this process can be effectively arrested [10]. The time to reach the equilibrium (equilibrium time) is an independent criterion for discriminating the presence/absence of spillover phenomena, since H2 adsorption on the metal was equilibrated within approx. 20 min at an equilibrium pressure of 2 Torr [11]. Most of the equilibrium times at −80 ◦ C were around this value, runs 1 and 3 (Table 1) being somewhat higher (35–37 min), and quite high in run 6 (223 min). Therefore, the contribution of H2 spilt over the support appears minimal in all the experiments conducted at −80 ◦ C, except run 6, where, however a low H/Pt was obtained. In this case, the sample was re-oxidised at 300 ◦ C leading to metal passivation, which can easily account both for the slow H2 uptake and low H/Pt. The equilibrium time at 25 ◦ C largely exceeded 20 min in all the runs except 4, 7 and 8. Accordingly, with the exception of these three runs, significant H2 spillover occurs. Consistently, the equilibrium is always attained within 20–30 min over the Pt/Al2 O3 , where significant spillover should not be expected due to the lack of the acceptor phase. In Pt/Cex Zr1−x O2 systems, H/Pt ratios as high as 12–20 are observed after reduction at moderate temperatures (≈200 ◦ C), that decrease to 1.6–1.7 when the reduction temperature is increased to 300–400 ◦ C [12]. The H/Pt ≈1 observed at 25 ◦ C on freshly reduced Pt/CZ60/Al2 O3 might appear somewhat low given the high surface area of the present material. The presence of exposed Al2 O3 that could have blocked spilling of the adsorbed species cannot be invoked since H2 can spill over the Cex Zr1−x O2 phase at RT even when physically mixed with Pt/Al2 O3 [13]. At RT, spilled H2 is suggested to be adsorbed exclusively at the surface of CeO2 . However, due to the low amount of Ce0.6 Zr0.4 O2 (13 wt.%), we calculate that 36 m2 g−1 of Al2 O3 are covered by Ce0.6 Zr0.4 O2 , if we assume that the (1 0 0) plane of the mixed oxide grows as a monolayer (with cell parameter of 0.5307 nm). H2 chemisorption on platinum can be evaluated by back extrapolating to 0 Torr the linear part of the H2 adsorption isotherm measured at −80 ◦ C to avoid contribution from physisorption. By back extrapolating the linear part of the H2 adsorption isotherm measured at 25 ◦ C it is possible to estimate the amount of hydro- 685 gen chemisorbed both on the metal phase and stored on/in the ceria-zirconia system. Therefore, the amount of spilled H can be calculated by the difference between the amount of H2 adsorbed at 25 and −80 ◦ C, leading to a surface density of 4.1 adsorbed H nm−2 of Ce0.6 Zr0.4 O2 . This value is in line with the values of 3.5 and 6.2 H atoms nm−2 previously found for Pt/Ce0.68 Zr0.32 O2 [8,12]. Evidence for crystalline Ce0.6 Zr0.4 O2 -supported phase could not be detected from the XRD pattern which suggests either the presence of an amorphous phase dispersed on the Al2 O3 surface or an upper limit of 2 nm for the Ce0.6 Zr0.4 O2 particles. By considering a spherical geometry of the particles and a bulk density of 5.58 ml g−1 , a surface density of 4.3 adsorbed H atoms nm−2 is calculated for such a particle size, which is very close to the above value. Both values indicate that extensive H2 spillover does occur in our system and that the presence of Al2 O3 does not depress this property to any appreciable extent. Fig. 1 reports the effects of redox treatments on the chemisorption isotherms. All the isotherms measured at 25 ◦ C feature comparable aspect, i.e. a flat adsorption isotherm for pH2 >100 Torr. The overall chemisorption capability is strongly depressed by increasing the reduction temperature. The aspect of the isotherms measured at −80 ◦ C closely resembles each other, however, no flat zone is observed above 100 Torr, the amount adsorbed still increasing with increasing the H2 pressure. This is an indication that, in addition to the chemisorption process, physisorption of H2 takes place at this temperature. A curious consequence of this phenomenon is that as the reduction temperature is increased up to 800 ◦ C, the H2 adsorption capability decreases more rapidly at 25 ◦ C compared to −80 ◦ C: starting from a situation where the isotherm measured after reduction at 350 ◦ C at 25 ◦ C is above than that obtained at −80 ◦ C, the opposite is observed after reduction at 800 ◦ C. Noticeably, when Pt/CZ60/Al2 O3 is reduced at 800 ◦ C, re-oxidised at 350 ◦ C and then re-reduced at 350 ◦ C, the adsorption capability starts to be recovered partially moving the isotherm obtained at 25 ◦ C above that measured at −80 ◦ C. These large modifications of the adsorption isotherms are related to the presence of the Ce0.6 Zr0.4 O2 phase. Consistently, Pt/Al2 O3 does not feature any unusual behaviour, the adsorption isotherm measured at −80 ◦ C being always above 686 P. Fornasiero et al. / Journal of Molecular Catalysis A: Chemical 204–205 (2003) 683–691 Fig. 1. H2 chemisorption isotherms measured at −80 (䉭) and 25 ◦ C (䊐) over Pt/Ce0.6 Zr0.4 O2 /Al2 O3 subjected to different reducing (R) and oxidation (䊊) treatments at the indicated temperature (◦ C). Fig. 2. H2 chemisorption isotherms measured at −80 (䉭) and 25 ◦ C (䊐) over Pt/Al2 O3 subjected to different reducing (R) and oxidation (䊊) treatments at the indicated temperature (◦ C). P. Fornasiero et al. / Journal of Molecular Catalysis A: Chemical 204–205 (2003) 683–691 that at 25 ◦ C due to occurrence of physisorption (ca. 0.20 ml g−1 ) at −80 ◦ C (Fig. 2). As the reduction temperature is increased from 350 to 800 ◦ C (Table 1), H/Pt measured at −80 ◦ C decreased from 0.57 to 0.16. If a spherical geometry for the Pt particles is assumed, this would account for an increase of average particle size from 2.3 to 7.1 nm, provided that other phenomena do not contribute to the decrease of the extent of H2 chemisorption. However, this is not the case in the NM/CeO2 -containing systems: CeO2 is a reducible oxide and susceptible to the strong metal/support interaction (SMSI) phenomena [14]. However, no evidence for metal decoration effects were observed in Rh/CeO2 after reduction at 500 ◦ C that typically generates the SMSI state (metal decoration effects) in Rh/TiO2 [14]. Consistently, increasing the reduction temperature from 150 to 350 ◦ C, the apparent H/Pt observed over Pt/Ce0.68 Zr0.32 O2 decreased by about 500 and 30%, respectively at 25 and −80 ◦ C without any significant particle size modification and decoration effects as checked by HREM, suggesting a deactivation of the H2 chemisorption due to electronic effects [11]. We believe that both chemical deactivation and loss of spillover capabilities contribute to the above deactivation, as indicated by the comparable H/Pt measured at −80 and 25 ◦ C. This is substantiated by the relatively small decrease of H/Pt experienced by Pt/Al2 O3 and by the fact that when Pt/CZ60/Al2 O3 was re-oxidised even at mild temperatures (75–300 ◦ C), H2 chemisorption was recovered to a significant extent. No Pt re-dispersion is expected to occur below 500 ◦ C. Noticeably, H/Pt = 1.67 is measured at 25 ◦ C after oxidation at 75 ◦ C which increases up to 2.61 at a re-oxidation temperature of 300 ◦ C, indicating recover of spillover capability and favourable effects of the reduction/mild re-oxidation sequence. The equilibrium time is more than three times higher in the sample re-oxidised at 300 ◦ C compared to that at 75 ◦ C. In both cases, the TPR profiles (see below) revealed the presence of reducible PtOx species on the surface. Accordingly, we attribute this phenomenon to a different rate of transport (spillover) of H2 species over the support. This is substantiated by observation of H2 adsorption even at −80 ◦ C, where Pt reduction should be more difficult. We believe that the slower attainment of equilibrium observed increasing the re-oxidation temperature from 75 to 300 ◦ C 687 should be associated with a more difficult reduction of PtOx species passivated by the higher oxidation temperature, leading to a slow rate of spillover. Significantly, almost no hydrogen spillover is observed for Pt/CZ60/Al2 O3 after reduction at 800 ◦ C followed by an oxidation at low temperature and a reduction at 350◦ (runs 7 and 8). For re-oxidation temperatures higher than 500 ◦ C both H/Pt ration and hydrogen spillover significantly increase. As suggested by a referee, the influence of effects like changes in the metal distribution on the catalyst (changing the relative amount of contacts between Pt and ceria-zirconia mixed oxide entities) or encapsulation of active Pt or Ce-Zr mixed oxide entities after the severe redox treatments could be invoked for this behaviour. Moreover, the main beneficial effect of the presence of a solid solution between Ce and Zr is related to avoiding formation of CeAlO3 . However, for high concentration of ceria and after long ageing at high temperature, formation of some CeAlO3 was observed [15]. Re-oxidation of CeAlO3 occurs at high temperature [16] and therefore its presence could be partially responsible for the absence of spillover observed in runs 7 and 8. 3.2. Temperature programmed reduction and oxygen uptake measurements TPR profiles of CZ60/Al2 O3 and Pt/CZ60/Al2 O3 subjected to consecutive redox treatments are reported in Fig. 3. TPR profiles of CeO2 -ZrO2 mixed oxides represent quite complex aspects due to a number of factors: phase non-homogeneity, pre-treatments, etc. [1]. Supporting Cex Zr1−x O2 on Al2 O3 leads to broad reduction features (Fig. 3b) that may appear even at quite low temperatures compared to unsupported mixed oxide [6,17]. Redox ageing (TPR followed by a mild oxidation) shifts the reduction to ca. 300 ◦ C. Importantly, whereas high temperature oxidation moves the reduction peak to ca. 600 ◦ C in the unsupported Cex Zr1−x O2 , presence of Al2 O3 minimises this undesirable deactivation (Fig. 3b). Pt/CZ60/Al2 O3 features quite complex TPR profiles that approximately occur in the following temperature intervals: (i) sub-ambient/ambient reduction, (ii) 100–300 ◦ C, (iii) 300–500 ◦ C, and (iv) at 600–800 ◦ C. Low temperature reduction if favoured by Pt due to its ability to activate and spill H2 [18]. However, the reducibility depends on the nature of the surface 688 P. Fornasiero et al. / Journal of Molecular Catalysis A: Chemical 204–205 (2003) 683–691 Fig. 3. Temperature programmed reduction of Pt/Ce0.6 Zr0.4 O2 /Al2 O3 (thin line) and Ce0.6 Zr0.4 O2 /Al2 O3 (bold line): (a) sub-ambient to ambient temperature; (b) ambient to 800 or 1000 ◦ C. (1) fresh sample, (2) sample from run (1) oxidised at 75 ◦ C, (3) sample from run (2) oxidised at 150 ◦ C, (4) sample from run (3) oxidised at 300 ◦ C, (5) sample from run (4) oxidised at 427 ◦ C, (6) sample from run (5) oxidised at 427 ◦ C, (7) sample from run (6) oxidised at 427 ◦ C, (8) sample from run (7) oxidised at 1000 ◦ C. species generated by the calcination of the Pt precursor: chemisorbed oxygen can react with H2 even at sub-ambient temperature, while PtO and PtO2 are reduced at ca. 50 and 100 ◦ C [19]. Notice that oxidation to give bulk PtO and PtO2 was observed, respectively, at 100 and 300 ◦ C in Pt/Al2 O3 [19]. The TPR profiles of the fresh (calcined 500 ◦ C) and oxidised at 1000 ◦ C samples significantly differ from others as denoted by the absence of significant H2 uptake near RT. This suggests the presence of some bulk Pt oxide— reduced with difficulty—that is generated by the deep oxidation treatment. Also an interaction between the metal particles and oxygen atoms of the CeO2 -ZrO2 phase could have generated hardly reducible Pt-O-Ce species [1], in addition to some surface dehydroxylation, both factors leading to a less efficient reduction process. The intense peak at 180 ◦ C observed in the fresh sample is associated with the reduction of the Pt oxide and Ce0.6 Zr0.4 O2 mixed oxide. Consistent with the high spillover capability of this sample, reduction occurs at a temperature close to that expected for PtO2 reduction. The peak above 600 ◦ C is likely associated with some CeO2 non-incorporated in the sample, as detected previously [6]. Redox ageing consisting of TPR/oxidation (up to 427 ◦ C) improves the reducibility of the system. In fact, the TPR profiles of all the recycled samples are characterised by an intense H2 uptake at RT, in addition to the broad reduction at 100–240 ◦ C. Due to the relatively mild re-oxidation conditions, passivation of the Pt particles by chemisorbed oxygen rather than deep oxidation to form a bulk Pt oxide should have occurred in these oxidation treatments. Such species P. Fornasiero et al. / Journal of Molecular Catalysis A: Chemical 204–205 (2003) 683–691 would be reduced at low temperatures, favouring extensive adsorption of H2 at the Ce0.6 Zr0.4 O2 surface. In agreement with the TPD investigation (see below) we associate the RT peak to reversible Ce(IV) reduction via a spillover mechanism [5] and the peak at 100–240 ◦ C to reduction in the bulk of the mixed oxide. There are some modification of this reduction feature with the pre-treatment, however, no clear trend could be discerned. 3.2.1. Effect of temperature on the re-oxidation process O2 uptake was measured after TPR at different temperatures using a pulse method (Table 2). Re-oxidation of Pt particles in the Pt/CZ60/Al2 O3 cannot be easily distinguished from that of the reduced Ce0.6 Zr0.4 O2−x moiety. By considering the Pt dispersion of 16%, measured in the chemisorption measurements after reduction at 800 ◦ C, passivation of Pt surface with one oxygen per exposed metal atom would require 0.15 ml of O2 g−1 of material. Bulk oxidation of Pt to PtO and PtO2 requires 0.9 and 1.9 ml of O2 g−1 of Pt/CZ60/Al2 O3 , respectively. Re-oxidation of reduced Cex Zr1−x O2 moieties is fast and effective even at RT: about 75% of the oxidation proceeded at RT while re-oxidation was fully accomplished by 100 ◦ C, independently of the oxide composition [20]. 689 Both metal-free and Pt-loaded samples do not show appreciable H2 uptake above 800 ◦ C suggesting comparable degree of reduction of the Ce0.6 Zr0.4 O2 component to be achieved in both systems. Based on this assumption, the excess of the O2 uptake observed in the Pt/CZ60/Al2 O3 compared to CZ60/Al2 O3 (Table 2) can be attributed to Pt-derived “redox reactions”. According to the reduction temperature (800 or 1000 ◦ C), two sets of difference values are obtained: 0.5 and 0.8–0.9 ml O2 g−1 , the latter value suggests oxidation of Pt particles to give bulk-like PtO, PtO2 not being formed under our re-oxidation conditions. When the reduction temperature is increased up to 1000 ◦ C, re-oxidation of Pt particles in the bulk becomes more difficult under the present transient conditions, which could be associated with some Pt sintering induced by the high temperature reduction. Consistently, when the sample oxidised at 75 ◦ C is further oxidised at 427 ◦ C (Table 2), negligible oxidation of Pt particles is observed. Curiously, the difference values slightly decrease as the oxidation temperature is increased from 75 to 427 ◦ C on the sample reduced at 800 ◦ C. As shown by runs 4 and 5 (Table 2), this cannot be attributed to a progressive sintering of Pt particles as these experiments were performed consecutively. Rather, it suggests that the kinetics of re-oxidation of the Ce0.6 Zr0.4 O2−x Table 2 Oxygen uptake measured over Ce0.6 Zr0.4 O2 /Al2 O3 and Pt/Ce0.6 Zr0.4 O2 /Al2 O3 after TPR treatments Runa Treatment Oxygen uptakeb Temperature (◦ C) 1 2 3 4 5 6 7 8 9 TPR TPR TPR TPR TPR TPR TPR TPR 800 800 800 800 800 1000 1000 800 75 150 300 427 427 427 1000 75 427 Amount adsorbed (ml g−1 ) (Ce(III) re-oxidation (%)) CZ60/Al2 O3 Pt/CZ60/Al2 O3 Difference 1.34 1.41 1.40 1.98 1.97 1.63 1.92 1.13 0.56 1.83 2.35 2.29 2.77 2.75 2.10 2.39 1.59 0.63 0.51 0.94 0.89 0.79 0.78 0.47 0.47 0.46 0.07 (52) (54) (54) (76) (76) (63) (74) (43) (21) a Consecutive runs. The fresh sample has been subjected to a cleaning procedure at 500 ◦ C for 1 h under O pulses before the TPR 2 was performed up to indicated temperature. b Assuming a Pt metal dispersion of 16%, the passivation of Pt surface with one oxygen per exposed metal atom would require 0.15 ml of O2 g−1 of Pt (1.5%)/Ce0.6 Zr0.4 O2 (13%)/Al2 O3 . Bulk oxidation of Pt to PtO and PtO2 would require, respectively, 0.9 and 1.9 ml of O2 g−1 . Difference calculated by subtracting the values obtained for Ce0.6 Zr0.4 O2 (13%)/Al2 O3 from those obtained for Pt (1.5%)/Ce0.6 Zr0.4 O2 (13%)/Al2 O3 . 690 P. Fornasiero et al. / Journal of Molecular Catalysis A: Chemical 204–205 (2003) 683–691 moieties might be favoured by the presence of the supported Pt. The highest O2 uptake in the re-oxidation of the reduced Ce0.6 Zr0.4 O2−x /Al2 O3 that corresponds to 76% of Ce(III) being re-oxidised, is fully consistent with the value of 77% previously obtained on unsupported Ce0.6 Zr0.4 O2−x [21]. Only about 70% of the ultimate O2 uptake is achieved by re-oxidation at 75–300 ◦ C, indicating that the presence of Al2 O3 stabilises reduced Ce(III) moieties at the surface as in case of CeO2 -Al2 O3 [1]. CeO2 and CeO2 -ZrO2 mixed oxides tend to form bidimensional patches at the Al2 O3 surface whose redox activity differs from that of the bulk oxide [17]. Redox cycles carried out up to 1000 ◦ C decrease the ability of the system to undergo re-oxidation as denoted by a decrease of 20% of the O2 uptake compared to reduction at 800 ◦ C. Full re-oxidation could be achieved by increasing the oxidation temperature to 1000 ◦ C. This suggests that the deactivation of the redox properties likely occurs through generation of reduced CeAlO3 -like moieties that can be re-oxidised only upon a high temperature oxidation. Presence of some CeAlO3 has been detected by XRD on the redox aged sample subjected to mild oxidation as the last treatment. The relatively high CeO2 content favours phase segregation upon high temperature treatments, CeO2 -rich phase thus generated are expected to react more easily with the Al2 O3 compared to ZrO2 -rich oxides [6]. 3.3. Temperature programmed desorption of H2 Fig. 4 reports the TPD profiles obtained on the investigated samples after pre-treatment in O2 flow at 400 ◦ C for 1 h, reduction at 500 ◦ C and cooling down in pure H2 to −70 ◦ C. TPD of H2 from Pt/CZ60/Al2 O3 features three desorption peaks centred at 50, 190 and 370 ◦ C (Fig. 4). The comparison with the TPD of Pt/Al2 O3 suggests contribution from H2 adsorbed on the metal as being responsible for the H2 desorbed at low temperatures, even though some reverse spillover of H2 presumably contributes to this signal as well. An interaction between Pt and Ce0.6 Z0.4 O2 is in fact suggested by the comparison of the desorption pattern obtained for CZ60/Al2 O3 . It is reasonable to associate the H2 desorption that occurs above 400 ◦ C in CZ60/Al2 O3 (Fig. 4) with the broad peak at 370 ◦ C in Pt/CZ60/Al2 O3 . Pt strongly promotes H2 storage at low temperatures on the support, which is Fig. 4. Temperature programmed desorption of H2 from Pt/Ce0.6 Zr0.4 O2 /Al2 O3 (a), Ce0.6 Zr0.4 O2 /Al2 O3 (b) and Pt/Al2 O3 (c). P. Fornasiero et al. / Journal of Molecular Catalysis A: Chemical 204–205 (2003) 683–691 then reversibly desorbed during the TPD experiment [5]. Spillover phenomenon therefore accounts for the strong increase of the amount H2 chemisorbed on Pt/CZ60/Al2 O3 compared to Pt/Al2 O3 . Interestingly, the desorption feature occurring near to RT suggests that the consumption of H2 found in the TPR profiles might be related to the so-called reversible reduction rather then creation of oxygen vacancies. [2] [3] [4] [5] [6] 4. Conclusions [7] The present results reveal significant effects of Al2 O3 on the hydrogen activation capabilities and the redox properties of the Pt/CeO2 -ZrO2 system: (i) the extent of the hydrogen spillover is not depressed by the presence of Al2 O3 to any appreciable extent; (ii) low temperature reduction in terms of TPR behaviour is promoted, (iii) the re-oxidation behaviour is affected by the presence of Al2 O3 as a support due to stabilisation of reduced Ce(III) moieties, however, the presence of ZrO2 and Pt plays a favourable role in that the re-oxidation process is facilitated. This confirms the ability of ZrO2 to thermally stabilise the OSC property of CeO2 even in these composite systems. [8] [9] [10] [11] [12] [13] [14] [15] Acknowledgements University of Trieste, Fondo Trieste 1999, MURST—PRIN 2000 “Stabilisation under reaction conditions of catalysts based on nano-dispersed metals for use in selective oxidation reactions”, “Catalysis for the reduction of the environmental impact of mobile source emissions”, CNR Agenzia 2000, and Consortium INCA—project “Urban Atmosphere” are gratefully acknowledged for financial support. References [1] J. Kaspar, M. Graziani, P. Fornasiero, Ceria-containing three way catalysts, in: K.A. Gschneidner Jr., L. Eyring (Eds.), [16] [17] [18] [19] [20] [21] 691 Handbook on the Physics and Chemistry of Rare Earths: The Role of Rare Earths in Catalysis, Elsevier, Amsterdam, 2000, Chapter 184, pp. 159–267. T. Takeguchi, S.N. Furukawa, M. Inoue, J. Catal. 202 (2001) 14. X. Wang, R.J. Gorte, Appl. Catal. A: Gen. 224 (2002) 209. K. Otsuka, Y. Wang, M. Nakamura, Appl. Catal. A: Gen. 183 (1999) 317. S. Bernal, J.J. Calvino, G.A. Cifredo, J.M. RodriguezIzquierdo, V. Perrichon, A. Laachir, J. Catal. 137 (1992) 1. R. Di Monte, P. Fornasiero, J. Kaspar, M. Graziani, J.M. Gatica, S. Bernal, A. Gomez Herrero, Chem. Commun. (2000) 2167. A. MartinezArias, M. FernandezGarcia, A.B. Hungria, A. IglesiasJuez, K. Duncan, R. Smith, J.A. Anderson, J.C. Conesa, J. Soria, J. Catal. 204 (2001) 238. N. Hickey, P. Fornasiero, J. Kaspar, J.M. Gatica, S. Bernal, J. Catal. 200 (2001) 181. N. Hickey, P. Fornasiero, J. Kaspar, M. Graziani, G. Blanco, S. Bernal, Chem. Commun. (2000) 357. S. Bernal, J.J. Calvino, G.A. Cifredo, A. Laachir, V. Perrichon, J.M. Herrmann, Langmuir 10 (1994) 717. J.M. Gatica, R.T. Baker, P. Fornasiero, S. Bernal, J. Kaspar, J. Phys. Chem. B 105 (2001) 1191. S. Salasc, V. Perrichon, M. Primet, N. MouaddibMoral, J. Catal. 206 (2002) 82. N. Hickey, P. Fornasiero, R. Di Monte, J. Kaspar, M. Graziani, G. Dolcetti, Catal. Lett. 72 (2001) 45. S. Bernal, F.J. Botana, J.J. Calvino, G.A. Cifredo, J.A. Perez-Omil, Catal. Today 23 (1995) 219. R. Di Monte, P. Fornasiero, J. Kaspar, M. Graziani, Stabilisation of nanostructured CeO2 -ZrO2 solid solutions by addition of Al2 O3 : a suitable way for production of thermally stable oxygen storage/release promoters for three-way catalysts, in: A. Gamba, C. Collela, S. Coluccia (Eds.), Oxide-Based Systems at the Crossroads of Chemistry, Stud. Surf. Sci. Catal., Elsevier, Amsterdam, 2001, pp. 229–236. T. Miki, T. Ogawa, A. Ueno, S. Matsuura, M. Sato, Chem. Lett. (1988) 565. M. FernandezGarcia, A. MartinezArias, A. IglesiasJuez, C. Belver, A.B. Hungria, J.C. Conesa, J. Soria, J. Catal. 194 (2000) 385. P. Fornasiero, R. Di Monte, G. Ranga Rao, J. Kaspar, S. Meriani, A. Trovarelli, M. Graziani, J. Catal. 151 (1995) 168. C.P. Hwang, C.T. Yeh, J. Mol. Catal. A: Chem. 112 (1996) 295. H. Vidal, J. Kaspar, M. Pijolat, G. Colon, S. Bernal, A. Cordon, V. Perrichon, F. Fally, Appl. Catal. B: Environ. 30 (2001) 75. P. Vidmar, P. Fornasiero, J. Kaspar, G. Gubitosa, M. Graziani, J. Catal. 171 (1997) 160.