11110
J. Phys. Chem. B 2005, 109, 11110-11118
Structural and Morphological Investigation of Ceria-Promoted Al2O3 under Severe Reducing/Oxidizing Conditions Alessandro Piras,† Sara Colussi,† Alessandro Trovarelli,*,† Valter Sergo,‡ Jordi Llorca,§ Rinaldo Psaro,| and Laura Sordelli| Dipartimento di Scienze e Tecnologie Chimiche, Via del Cotonificio 108, UniVersita` di Udine, 33100 Udine, Italy, CENMAT, Center of Excellence for Nanostructured Materials, UniVersita` di Trieste, Trieste, Italy, Dept. Quimica Inorganica, UniVersitat de Barcelona, Barcelona, Spain, ISTM-CNR and Centro di Eccellenza CIMAINA, Via Golgi 19, 20133 Milano, Italy ReceiVed: December 30, 2004; In Final Form: April 11, 2005
A series of CeO2/Al2O3 samples with different ceria loadings in the range 0-25 wt % (0, 2, 5, 7.5, 15, and 25%) were prepared by incipient wetness and studied using several complementary techniques such as Brunauer-Emmett-Teller (BET), X-ray diffraction (XRD), temperature-programmed reduction (TPR), Raman, high-resolution transmission electron microscopy (HRTEM), and extended X-ray absorption fine structure (EXAFS). The aim of the investigation was to understand the behavior of ceria when deposited on alumina and treated under oxidizing and reducing conditions at high temperature (T g 1273 K). It is shown that ceria can partially stabilize alumina toward the formation of low-surface-area phases up to 1373 K under oxidizing conditions, while enhanced stabilization is observed under reducing conditions, being effective up to 1473 K. A detailed quantitative temperature-programmed reduction (TPR) analysis made at different loadings and calcination temperatures allowed us to identify three characteristic regions where the reduction of small and large ceria crystallites occurs with the formation of CeAlO3 crystallites at high temperature. These are likely responsible for surface-area stabilization. For dispersed ceria samples, reduction takes place almost exclusively at low temperature (1150 K), where the reduction of Ce4+ beyond the thermodynamic value observed for pure ceria24 is driven by interaction with Al2O3. All samples show some features in the low-temperature region due to the presence of small ceria crystallites regardless of loading. For 2CeAl, only one peak is visible in the TPR profile at about 690 K, indicating that in this sample no contributions are observed from the reduction of large crystallites. An increase in ceria loading induces a broadening of the low-temperature peak and a shift of H2 consumption to medium and high temperature. The medium-temperature peak is associated with the reduction of large ceria crystallites; the amount of hydrogen consumed
Piras et al.
Figure 2. Temperature-programmed reduction profiles of samples calcined at 1473 K for 6 h: (a) 2CeAl; (b) 5CeAl; (c) 7.5CeAl; (d) 15CeAl; (e) 25CeAl.
under this peak increases with ceria loading, reaching a maximum with pure CeO2, according to the presence of progressively larger ceria crystallites (see the XRD section). The high-temperature peak, which is not present in the reduction profile of pure ceria, is associated with the reduction of Ce4+ to Ce3+ with the formation of CeAlO3, as evidenced by XRD. The reduction profile of pure ceria is in agreement with previously reported data.23 Due to the low surface area of the sample prepared by decomposition of nitrates, the low-temperature signals are very weak, and the majority of H2 consumption originates from the reduction of large ceria crystallites. Figure 2 shows TPR profiles of samples calcined at high temperature (1473 K). We can always distinguish three different regions of hydrogen consumption at low, medium, and high temperature. However, some distinct features characterize these profiles: (i) The temperature range of hydrogen consumption is significantly shifted to a higher temperature, indicating that calcination at 1473 K affects the dimensions of ceria crystallites, although there is still a fraction of highly reducible CeO2, especially in the low loading samples (100 Å (1473 K), respectively. The Debye-Waller factor is significatively higher in the sample treated at 1273 K, suggesting the presence of a higher structural or morphological disorder when the oxide particles are formed at lower temperatures. The oxidation treatments at 1273 and 1373 K produce coordination numbers lower than the value corresponding to CeO2 in the oxygen first shell. This decrease indicates that part of the Ce atoms have a lower oxygen coordinative sphere than CeIV. In fact, as it occurs in the defective fluoritic structure of CeOx oxides (x ) 1.66-1.83)31 obtained by controlled reduction of CeO2, ordered oxygen vacancy distributions, that is, CeIII ion distributions, have been determined by neutron diffraction, which would correspond to oxygen nearest coordination numbers between 7 and 8. The hypothesis of the presence of a fraction (lower than 10%) of Ce in a lower oxidation state suggested that the TPR data are therefore in agreement with EXAFS results. Following a treatment under reductive atmosphere, a drastic phase transformation is evident in the Fourier transformed spectra in comparison to those after oxidation (Figure 10). There is no indication of any oxygen shell at the Ce-O distance of CeO2 anymore, but three well-resolved peaks are present whose position and integrated area suggest the progressive formation of a CeAlO3 phase (perovskite, a ) 3.820) with increasing reduction temperature. The first peak corresponds to the contribution of the Ce-O and Ce-Al shells at 2.70 and 3.31 Å, respectively; the second and third peaks correspond to the Ce-Ce shells at 3.82 and 5.40 Å. The fit of the sample reduced at 1473 K gives coordination numbers and shell distances in perfect agreement with the crystallographic data of the cerium aluminate up to the fourth shell, suggesting the formation of aggregates with mean diameters of a few nanometers. The reduction at 1373 K produces a similar structure with lower coordination numbers and higher Debye-Waller factors for all the shells. As the O and Al contribution under the first peak could not be resolved, the resulting distance reported in the fit is intermediate between those of the two individual shells. These values suggest that under reduction at 1373 K part of Ce is in the form of crystalline CeAlO3 and part is well dispersed at the alumina surface as an amorphous phase in agreement with Rietveld analysis.
11116 J. Phys. Chem. B, Vol. 109, No. 22, 2005
Piras et al.
TABLE 4: EXAFS Experimental Fits for the 15CeAl Sample Treated for 6 h under Different Conditionsa sample oxidation at 1273 K oxidation at 1373 K oxidation at 1473 K reduction at 1273 K reduction at 1373 K reduction at 1473 K
reduction at 1473 K + oxidation at 1473 K a
shell
N
R (Å)
σDW (Å)
∆E0 (eV)
O Ce O Ce O Ce O Ce O + Al Ce Ce O Al Ce Ce O Ce
7.4 ( 0.2 9.5 ( 0.5 7.7 ( 0.1 10.3 ( 0.3 8.4 ( 0.3 12.2 ( 0.5
2.338 ( 0.003 3.834 ( 0.003 2.355 ( 0.003 3.840 ( 0.003 2.344 ( 0.003 3.837 ( 0.005
0.080 ( 0.003 0.067 ( 0.002 0.078 ( 0.004 0.065 ( 0.002 0.073 ( 0.003 0.064 ( 0.002
-0.2 ( 0.05 1.1 ( 0.5 -0.2 ( 0.05 2.1 ( 0.1 -8.6 ( 0.3 0.4 ( 0.4
7.3 ( 0.1 5.2 ( 0.3 4.2 ( 0.2 11.3 ( 0.2 5.5 ( 0.4 6.1 ( 0.5 5.7 ( 0.5 8.4 ( 0.3 12.0 ( 0.2
3.136 ( 0.003 3.816 ( 0.003 5.405 ( 0.004 2.701 ( 0.003 3.311 ( 0.004 3.795 ( 0.005 5.400 ( 0.005 2.361 ( 0.003 3.841 ( 0.003
0.080 ( 0.004 0.091 ( 0.005 0.071 ( 0.003 0.078 ( 0.003 0.081 ( 0.004 0.093 ( 0.004 0.073 ( 0.005 0.075 ( 0.004 0.064 ( 0.002
14.1 ( 0.05 1.9 ( 0.1 -1.6 ( 0.3 -4.6 ( 0.3 2.5 ( 0.2 1.9 ( 0.4 -1.3 ( 0.5 -0.1 ( 0.04 -4.0 ( 0.4
A mean standard deviation of 5 × 10-3 has been used for the error bar estimation. Fit quality factor χ2ν ) 1-2.
Figure 10. Fourier transformed EXAFS spectra of 15CeAl treated under different conditions at (a) 1273 K, (b) 1373 K, and (c) 1473 K.
The reduction at 1273 K produces a sample whose signal characteristics are similar to those reported above. However, peak intensities were comparable to the background noise, and it was not possible to perform the fitting procedure. The coordination numbers of samples reduced at 1373 K are in agreement with the needle-shaped crystals of CeAlO3 observed in the HRTEM micrographs, as the mean particle diameter suggested by the N values is comparable with the needle thickness. The particles are, however, big with respect to the EXAFS sensitivity. Reoxidation at 1473 K after the reductive treatment at 1473 K completely regenerates the CeO2 phase detected after direct calcination. Discussion Behavior under Oxidative Conditions. Our XRD and Raman data indicate that ceria particles with a well-defined cubic fluorite structure are formed in increasing amounts with increasing cerium loadings. Therefore, in agreement with recent
investigations,13,14 it is shown that the deposition of ceria over Al2O3 forms aggregates of CeO2 with dimensions that are strongly dependent on CeO2 loading and calcination temperatures. Transition between a mainly dispersed and a mainly bulk ceria phase is often located at a loading corresponding to ∼2.7 µmol of CeO2/m2 for samples calcined at 1073 K.11 The breakpoint of our TPR profiles, which shows an increasing contribution for reduction at medium to high temperatures at around 7.5 wt % CeO2 and above, indicates that this point can be considered a boundary between dispersed and bulk phases after calcinations at 1273 K. This corresponds to a loading of ∼2.05 µmol of CeO2/m2 of alumina, a value which is lower than the previous estimates, due to the higher calcination temperatures employed. A shift to values less than 0.51 µmol of CeO2/m2 is observed by increasing calcination temperatures to 1473 K. EXAFS data show that the oxidation treatments at 1273 and 1373 K on the 15% CeO2/Al2O3 sample produce a phase where a fraction of Ce atoms has a lower oxygen coordinative sphere than that of CeIV in CeO2; a behavior that can be attributed to a distribution of oxygen vacancies, corresponding to the presence of CeIII ions. Indirect evidence pointing to the presence of other Ce-containing units derives from quantitative TPR analysis, which indirectly identifies some cerium present in the lower oxidation state, especially in the more disperse samples. This could originate from the presence of very disperse ceria particles, analogous to the 2D Ce entities13 but stabilized in a formally +3 state, like ceria suboxide. Stabilization of ceria suboxide when dealing with aggregates with dimensions of less than a few nanometers has been recently reported in pure nanocrystalline ceria.32,33 Behavior under Reducing Conditions. Before entering into the structural and morphological features of CeAl samples under a reductive environment, let us discuss in detail the reduction profiles collected under TPR conditions. We should briefly review the main features of TPR of pure ceria and extend those conclusions, when possible, to the case of the CeO2/Al2O3 system. The bimodal peak shape of ceria is a consequence of the fact that nanocrystalline and bulk ceria possess different reduction properties,22 and the reduction of nanocrystallites is therefore complete at lower temperature, owing to a lower reduction enthalpy.34 Consequently, smaller ceria crystallites start to be reduced at low temperature (T ∼ 500 K) and can be reduced thermodynamically to a high degree even at these temperatures. Limitation in the degree of reduction is given by the kinetics of the process and by the growth of ceria crystallites.
Ceria-Promoted Al2O3 under Reducing/Oxidizing Conditions The result is that the shape of the first peak is related to the effect of balance between the dynamics of sintering (i.e., surfacearea loss) and the kinetic limitation for the reduction of nanocrystallites. At the same time, upon an increase in temperature, the material undergoes morphological modifications and small crystallites progressively modify to bulk ceria which has different equilibrium reduction properties. The deposition of CeO2 over Al2O3 poses an additional set of problems that must be taken into account to explain the TPR profiles. First, we have a “dispersion” effect; the presence of a high-surfacearea support does stabilize small particles and their growth is inhibited by “steric” interaction with the support in a purely mechanical fashion. The presence of alumina induces also a “chemical” effect due to the possibility of interaction between ceria and alumina to form mixed oxide compounds such as CeAlO3, for example. If we compare the reduction characteristics of ceria in CeAl samples to the theoretical reduction degree of CeO2, as calculated from thermodynamics, we can see that in the whole temperature range examined reduction of CeO2 in CeAl samples exceeds by far the amount dictated by thermodynamics. The majority of thermodynamic data in the literature refer to bulk ceria samples35 and therefore cannot be applied to correctly explain the present TPR profile. For sample 2CeAl and in large part for sample 5CeAl (i.e., samples at low loading), total reduction of ceria occurs in one single peak at low temperature. This is consistent with the reduction of small ceria crystallites according to favorable thermodynamics (for small ceria crystallites, thermodynamics predicts reduction even at room temperature; the smaller the crystallites are, the higher is the reducibility of ceria and its stability in a reduced state36,37). Sintering of ceria particles does not occur in CeAl samples under TPR conditions because (i) before TPR, the sample was already treated at 1273 K and (ii) the presence of the support stabilizes small ceria crystallites when the loading is low. As a consequence, in the 2CeAl and 5CeAl samples, there is no evidence for the second TPR peak at higher temperature (MTP). By increasing the ceria loading, the ability of the support to stabilize ceria in a highly dispersed state is lower as confirmed by XRD, and we observe a progressive increase of the contribution of reduction from the medium-temperature range compared to that at low temperature. The above picture is also in agreement with the observation that samples treated at 1473 K before TPR show a shift of reduction of the entire TPR profile at much higher temperatures, consistent with an increase of the average dimensions of ceria crystallites. In this case, even the 2CeAl sample shows hydrogen consumption at medium and high temperatures. At temperatures higher than 900 K, the interaction between Ce3+ and Al2O3 causes the formation of CeAlO3 at the expense of large ceria crystallites, originating the high-temperature peaks, and in agreement also with bulk thermodynamic data which show that the reaction of formation of CeAlO3 is the most favored at these temperatures.11 The formation of CeAlO3 under reductive conditions can be associated with the stabilization of alumina toward the formation of the R phase with a mechanism similar to that proposed for La2O3 on Al2O3.38,39 It is known that the formation of lanthanum aluminate on the surface of alumina is responsible for its stabilization under an oxidative environment; the presence of La, naturally available as La3+, facilitates this reaction which readily occurs under an oxidative atmosphere. With CeO2/Al2O3, a temperature above 1000 K under reducing conditions favors the formation of bulk CeAlO3. The reason for the thermal stabilization observed under reductive conditions is likely to be related to the geometrical arrangement between CeAlO3 and
J. Phys. Chem. B, Vol. 109, No. 22, 2005 11117 CeO2-x which form microdomains that might act as a barrier for surface diffusion, as suggested by Humbert et al.10 The presence of Ce3+ can also explain the partial stabilization observed under an oxidative environment. Although we do not have direct evidence for the formation of CeAlO3 following calcinations, it appears from either TPR or EXAFS analysis that a small fraction of Ce is in the form of Ce3+ well disperse on Al2O3. Under these conditions, the formation of superficial CeAlO3 by the diffusion of Al3+ ions into the partially reduced ceria lattice is likely to occur.10,14 Conclusions The characterization of ceria deposited on alumina has been for the first time carried out under severe oxidizing and reducing conditions, as required by utilization of supports in the last generation of TWC and high-T catalytic combustion. As a result of the present studies, we have found that most of the structural modifications operated by ceria under oxidative conditions is observed at around 1373 K; this temperature is therefore critical in the study of this system. If treatments are performed at 1273 and 1473 K, the structural and thermal stabilization effects are almost unaffected by the presence of ceria. An unprecedented detailed characterization of the TPR behavior of CeO2/Al2O3 has shown that the presence of Ce3+ plays a crucial role. The formation of Ce3+ occurs under both oxidative and reductive conditions. In the former case, it is related to the presence of very disperse ceria crystallites, which are more susceptible to easy reduction at low temperature due to favorable thermodynamics. A strong interaction between Al2O3 and ceria under reducing conditions results in the formation of crystalline CeAlO3 and in an important stabilization of the textural properties of the CeO2/Al2O3, even at temperatures of 1473 K. The effect is lost upon reoxidation of the material at high temperature. Acknowledgment. MIUR and Regione Friuli Venezia Giulia are acknowledged for financial support. The authors thank the ESRF of Grenoble for the provision of synchrotron radiation under the EEC grant CH-980 and the staff of the BM29 beamline for the technical assistance. References and Notes (1) Catalysis by Ceria and Related Materials; Trovarelli, A., Ed.; Imperial College Press: London, 2002; pp 1-528. (2) Colussi, S.; de Leitenburg, C.; Dolcetti, G.; Trovarelli, A. J. Alloys Compd. 2004, 374, 387. (3) Groppi, G.; Cristiani, C.; Lietti, L.; Ramella, C.; Valentini, M.; Forzatti, P. Catal. Today 1999, 50, 399. (4) Gelin, P.; Primet, M. Appl. Catal., B 2002, 39, 1. (5) Fu, J. C.; Czarnecki, L.; Patellis, C.; Bouney, A.; Whittenberger, W. A. Proc. Air & Waste Manag. Association, 92nd Annual Meeting St. Louis (MI) 1999, 99-217. (6) Oh, Y.-S.; Roh, H.-S.; Jun, K.-W.; Baek, Y.-S. Int. J. Hydrogen Energy 2003, 28, 1387. (7) Roh H.-S.; Jun, K.-W.; Baek, Y.-S.; Park, S.-E. Catal. Lett. 2002, 81, 147. (8) Ozawa, M.; Kimura, M. J. Mater. Sci. Lett. 1990, 9, 291 (9) Morterra, C.; Magnacca, G.; Bolis, V.; Cerrato, G.; Baricco, M.; Giachello, A.; Fucale, M. Stud. Surf. Sci. Catal. 1995, 96, 361. (10) Humbert, S.; Colin, A.; Monceaux, L.; Oudet F.; Courtine, P. Stud. Surf. Sci. Catal. 1995, 96, 829. (11) Shyu, J. Z.; Weber W. H.; Gandhi, H. S. J. Phys. Chem. 1988, 92, 4964. (12) Piras, A.; Trovarelli, A.; Dolcetti, G. Appl. Catal., B 2000, 28, L77. (13) Martinez-Arias, A.; Fernandez-Garcia, M.; Salamanca, L. N.; Valenzuela, R. X.; Conesa, J. C.; Soria, J. J. Phys. Chem. B 2000, 104, 4038. (14) Damyanova, S.; Perez, C. A.; Schmal, M.; Bueno, J. M. C. Appl. Catal., A 2002, 234, 271.
11118 J. Phys. Chem. B, Vol. 109, No. 22, 2005 (15) Riguetto, B. A.; Damyanova, S.; Gouliev, G.; Marques, C. M. P.; Petrov, L.; Bueno, J. M. C. J. Phys. Chem. B 2004, 108, 5349. (16) Morterra, C.; Bolis, V.; Magnacca, G. J. Chem. Soc., Faraday Trans. 1996, 92, 1991. (17) Rohart, E.; Larcher, O.; Deutsch, S.; Hedouin, C.; Aimin, H.; Fajardie, F.; Allain, M.; Macaudiere, P. Top. Catal. 2004, 30/31, 417. (18) Widjaja, H.; Sekizawa, K.; Eguchi, K.; Arai, H. Catal. Today 1997, 35, 197. (19) Forzatti, P.; Groppi, G. Catal. Today 1999, 54, 165. (20) Boaro, M.; Vicario, M.; de Leitenburg, C.; Dolcetti, G.; Trovarelli, A. Catal. Today 2003, 77, 407. (21) Perrichon, V.; Laachir, A.; Abourdanasse, S.; Touret, G.; Blanchard, G. Appl. Catal., A 1995, 69, 129. (22) Giordano, F.; Trovarelli, A.; de Leitenburg, C.; Giona, M. J. Catal. 2000, 193, 273. (23) Trovarelli, A. Catal. ReV.sSci. Eng. 1996, 38, 439. (24) Bevan, D. J. M.; Kordis, J. J. Inorg. Nucl. Chem. 1964, 26, 1508. (25) Bernal, S.; Calvino, J. J.; Gatica, J. M.; Cartes, C. L.; Pintado, J. M. In Catalysis by Ceria and Related Materials; Trovarelli, A., Ed.; Imperial College Press: London, 2002; Chapter 4, p 85. (26) Kepinski, L.; Wolcyrz, M. Catal. Lett. 1992, 15, 329. (27) Graham, G. W.; Weber, W. H.; Peters, C. R.; Usmen, R. J. Catal. 1991, 130, 310.
Piras et al. (28) Kosacki, I.; Suzuki, T.; Anderson, H. U.; Colomban, P. Solid State Ionics 2002, 149, 99. (29) Wen, Q.; Lipkin, D. M.; Clarke, D. R. J. Am. Ceram. Soc. 1998, 81, 3345. (30) Tolpigo, V. K.; Clarke, D. R. Mater. High Temp. 2000, 17, 59. (31) Kummerle, E. A.; Heger, G. J. Solid State Chem. 1999, 147, 485. (32) Tsunekawa, S.; Sivamohan, R.; Ito, S.; Kasuya, A.; Fukuda, T. Nanostruct. Mater. 1999, 11, 141. (33) Wu, L.; Wiesman, H. J.; Moodenbaugh, A. R.: Klie, R. F.; Zhu, Y.; Welch, D. O.; Suenaga, M. Phys. ReV. B 2004, 69, 125415. (34) Hwang, J. H.; Mason, T. O. Z. Phys. Chem. 1998, 207, 21. (35) Trovarelli, A. In Catalysis by Ceria and Related Materials; Trovarelli, A., Ed.; Imperial College Press: London, 2002; Chapter 2, p 15. (36) Zhang, F.; Jin, Q.; Chan, S.-W. J. Appl. Phys. 2004, 95, 4319. (37) Cordatos, H.; Ford, D.; Gorte, R. J. J. Phys. Chem. 1996, 100, 18128. (38) Burtin, P.; Brunelle, J. P.; Pijolat, M.; Soustelle, M. Appl. Catal. 1987, 34, 225. (39) Burtin, P.; Brunelle, J. P.; Pijolat, M.; Soustelle, M. Appl. Catal. 1987, 34, 239.
