12
Preparation and Characterization of Me 2 O 3 -CeO 2 (Me ) B, Al, Ga, In) Mixed Oxide Catalysts. 2. Preparation by Sol-Gel Method B. Bonnetot, V. Rakic, ‡,§ T. Yuzhakova, § C. Guimon, 4 and A. Auroux* Laboratoire des Multimatériaux et Interfaces, Unité Mixte de Recherche (UMR) 5615, UCB Lyon 1, 69622 Villeurbanne Cedex, France; Faculty of Agriculture, Department of Chemistry, UniVersity of Belgrade, Nemanjina 6 11080 Zemun, Serbia; Institut de Recherches sur la Catalyse et l’EnVironnement de Lyon, UMR 5256, Centre National de la Recherche Scientifique (CNRS) - UniVersité Lyon 1, 2 AV. Einstein, 69626 Villeurbanne Cedex, France; and Institut Pluridisciplinaire de Recherche sur l’EnVironnement et les Matériaux (IPREM) - Equipe de Chimie Physique (ECP), UMR 5254, CNRS - UniVersité de Pau et des Pays de l’Adour, 2 aVenue P. Angot, 64053 Pau Cedex 9, France ReceiVed July 10, 2007. ReVised Manuscript ReceiVed NoVember 21, 2007 The present work focuses on the combination of ceria mixed with another oxide from group III, using the sol–gel method. B 2 O 3 - Al 2 O 3 - Ga 2 O 3 - or In 2 O 3 -CeO 2 mixed oxides have been prepared in order to improve the catalytic properties of these materials. The structural, textural, and surface properties of these catalysts have been fully characterized by means of a variety of techniques (Brunauer-Emmett-Teller, BET; X-ray diffraction analysis, XRD; Raman; scanning electron microscopy, SEM; thermogravimetric analysis, TG; and temperature-programmed reduction/oxidation, TPR-TPO). The highest surface area was achieved for the Al 2 O 3 -CeO 2 mixed oxide. Only the fluorite structure of CeO 2 was observed by XRD for all prepared mixed oxides; the presence of oxygen vacancies was proven by Raman spectroscopy. The acid–base properties were estimated by the adsorption of probe molecules (NH 3 and SO 2 ); two methods, microcalorimetry and X-ray photoelectron spectroscopy (XPS), have been employed for that purpose. All investigated mixed oxides express surface amphoteric character; however, it seems that surface basicity is more pronounced than surface acidity. The surface bacisity, which is mainly of the Brönsted type, has been found as dependent on the character of the group III metal. Red-ox properties were investigated for the In 2 O 3 -CeO 2 sample, the achieved degree of reduction being 34%. TPR/TPO experiments performed up to 830 °C irreversibly changed the morphology and the structure of the samples, leading to the crystallization of In 2 O 3 . 1. Introduction Cerium oxide is one of the most important catalytic materials that can play multiple roles owing to its unique acid–base and red-ox properties. It expresses a high reduc- ibility of Ce 4+ , which is a consequence of the O 2- high mobility inside the ceria fluorite type structure. The ability to cycle easily between reduced (Ce 3+ ) and oxidized (Ce 4+ ) states and, particularly, its oxygen storage capacity are additional important properties of this material which promote its applicability in the domain of catalysis. Under common catalytic conditions, the fluorite structure of ceria is usually preserved; however, its catalytic efficiency may be reduced at elevated temperatures, due to sintering and a loss of surface area. However, despite a vast number of literature reports concerning this topic, the preparation of ceria-containing catalytic materials with sufficiently high specific area is not well-known. Therefore, the improvement of ceria properties is an attractive subject for numerous investigators. Very often, ceria is “stabilized” by its disper- sion on the supports that express high surface and thermal stabilities, such as alumina or silica. Particularly, it is often employed in mixed-oxide formulations or in conjunction with active (noble) metals. Supported CeO 2 - and CeO 2 -based mixed oxides have versatile applicability: cerium oxide is included in materials used in fuel cell processes 1–8 and in oxygen permeation membrane systems; 9,10 additionally, those materials can serve as promoters for industrially important processes such is fluid catalytic cracking. 11 Importantly, the possibility of their application as catalysts for environmental purposes has been tested and proven: these materials are effective catalysts in three-way catalysis, 11–15 in oxidation * Corresponding author e-mail: [email protected]. UCB Lyon 1. University of Belgrade. § CNRS - Université Lyon 1. 4 Université de Pau et des Pays de l’Adour. (1) Shao, Z.; Halle, S. M. Nature 2004, 431, 170. (2) Laosiripojana, N.; Assabumrungrat, S. Appl. Catal., B 2005, 60, 107. (3) Ramirez-Cabrero, E.; Atkinson, A.; Chadwick, D. Appl. Catal., B 2005, 47, 107. (4) Wong, F.-Y.; Wan, B.-Z.; Cheng, S. J. Solid State Electrochem. 2005, 9, 168. (5) Panzero, G.; Madafferi, V.; Candamano, S.; Donato, A.; Frusteri, F.; Antonucci, P. L. J. Power Sources 2004, 135, 177. (6) Sato, K.; Hashido, T.; Yashiro, K.; Yugami, H.; Kawada, T.; Mizusaki, J. J. Ceram. Soc. Jpn. 2005, 113, 562. (7) Kumar, A.; Devi, P. S.; Maiti, H. S. Chem. Mater. 2004, 16, 5562. (8) Shanwen, T.; Jonh, T. S. I. Chem. Mater. 2004, 16, 4116. (9) Levy, C.; Guizard, Ch.; Julbe, A. Sep. Purif. Technol. 2004, 32, 327. (10) Yin, X.; Hong, L.; Liu, Z.-L. Appl. Catal., A 2006, 300, 75. (11) Trovarelli, A.; de Leitenburg, C.; Boaro, M.; Dolcetti, G. Catal. Today 1999, 50, 353. (12) Zhu, T.; Kundakovic, Lj.; Dreher, A.; Flytzani-Stephanopoulos, M. Catal. Today 1999, 50, 381. (13) Solinas, V.; Rombi, E.; Ferino, I.; Cutrufello, M. G.; Colón, G.; Navío, J. A. J. Mol. Catal. A: Chem. 2003, 204–205, 629. 1585 Chem. Mater. 2008, 20, 1585–1596 10.1021/cm071809e CCC: $40.75 2008 American Chemical Society Published on Web 01/19/2008

Preparation and Characterization of Me 2 O 3 −CeO 2 (Me = B, Al, Ga, In) Mixed Oxide Catalysts. 2. Preparation by Sol−Gel Method

  • Upload
    a

  • View
    214

  • Download
    0

Embed Size (px)

Citation preview

Preparation and Characterization of Me2O3-CeO2 (Me ) B, Al, Ga,In) Mixed Oxide Catalysts. 2. Preparation by Sol-Gel Method

B. Bonnetot,† V. Rakic,‡,§ T. Yuzhakova,§ C. Guimon,4 and A. Auroux*,§

Laboratoire des Multimatériaux et Interfaces, Unité Mixte de Recherche (UMR) 5615, UCB Lyon 1,69622 Villeurbanne Cedex, France; Faculty of Agriculture, Department of Chemistry, UniVersity of

Belgrade, Nemanjina 6 11080 Zemun, Serbia; Institut de Recherches sur la Catalyse et l’EnVironnementde Lyon, UMR 5256, Centre National de la Recherche Scientifique (CNRS) - UniVersité Lyon 1, 2 AV.

Einstein, 69626 Villeurbanne Cedex, France; and Institut Pluridisciplinaire de Recherche surl’EnVironnement et les Matériaux (IPREM) - Equipe de Chimie Physique (ECP), UMR 5254, CNRS -

UniVersité de Pau et des Pays de l’Adour, 2 aVenue P. Angot, 64053 Pau Cedex 9, France

ReceiVed July 10, 2007. ReVised Manuscript ReceiVed NoVember 21, 2007

The present work focuses on the combination of ceria mixed with another oxide from group III, usingthe sol–gel method. B2O3- Al2O3- Ga2O3- or In2O3-CeO2 mixed oxides have been prepared in orderto improve the catalytic properties of these materials. The structural, textural, and surface propertiesof thesecatalystshavebeenfullycharacterizedbymeansofavarietyof techniques (Brunauer-Emmett-Teller,BET; X-ray diffraction analysis, XRD; Raman; scanning electron microscopy, SEM; thermogravimetricanalysis, TG; and temperature-programmed reduction/oxidation, TPR-TPO). The highest surface areawas achieved for the Al2O3-CeO2 mixed oxide. Only the fluorite structure of CeO2 was observed byXRD for all prepared mixed oxides; the presence of oxygen vacancies was proven by Raman spectroscopy.The acid–base properties were estimated by the adsorption of probe molecules (NH3 and SO2); twomethods, microcalorimetry and X-ray photoelectron spectroscopy (XPS), have been employed for thatpurpose. All investigated mixed oxides express surface amphoteric character; however, it seems thatsurface basicity is more pronounced than surface acidity. The surface bacisity, which is mainly of theBrönsted type, has been found as dependent on the character of the group III metal. Red-ox propertieswere investigated for the In2O3-CeO2 sample, the achieved degree of reduction being 34%. TPR/TPOexperiments performed up to 830 °C irreversibly changed the morphology and the structure of the samples,leading to the crystallization of In2O3.

1. Introduction

Cerium oxide is one of the most important catalyticmaterials that can play multiple roles owing to its uniqueacid–base and red-ox properties. It expresses a high reduc-ibility of Ce4+, which is a consequence of the O2- highmobility inside the ceria fluorite type structure. The abilityto cycle easily between reduced (Ce3+) and oxidized (Ce4+)states and, particularly, its oxygen storage capacity areadditional important properties of this material whichpromote its applicability in the domain of catalysis.

Under common catalytic conditions, the fluorite structureof ceria is usually preserved; however, its catalytic efficiencymay be reduced at elevated temperatures, due to sinteringand a loss of surface area. However, despite a vast numberof literature reports concerning this topic, the preparation ofceria-containing catalytic materials with sufficiently highspecific area is not well-known. Therefore, the improvementof ceria properties is an attractive subject for numerousinvestigators. Very often, ceria is “stabilized” by its disper-sion on the supports that express high surface and thermalstabilities, such as alumina or silica. Particularly, it is often

employed in mixed-oxide formulations or in conjunction withactive (noble) metals. Supported CeO2- and CeO2-basedmixed oxides have versatile applicability: cerium oxide isincluded in materials used in fuel cell processes1–8 and inoxygen permeation membrane systems;9,10 additionally, thosematerials can serve as promoters for industrially importantprocesses such is fluid catalytic cracking.11 Importantly, thepossibility of their application as catalysts for environmentalpurposes has been tested and proven: these materials areeffective catalysts in three-way catalysis,11–15 in oxidation

* Corresponding author e-mail: [email protected].† UCB Lyon 1.‡ University of Belgrade.§ CNRS - Université Lyon 1.4 Université de Pau et des Pays de l’Adour.

(1) Shao, Z.; Halle, S. M. Nature 2004, 431, 170.(2) Laosiripojana, N.; Assabumrungrat, S. Appl. Catal., B 2005, 60, 107.(3) Ramirez-Cabrero, E.; Atkinson, A.; Chadwick, D. Appl. Catal., B 2005,

47, 107.(4) Wong, F.-Y.; Wan, B.-Z.; Cheng, S. J. Solid State Electrochem. 2005,

9, 168.(5) Panzero, G.; Madafferi, V.; Candamano, S.; Donato, A.; Frusteri, F.;

Antonucci, P. L. J. Power Sources 2004, 135, 177.(6) Sato, K.; Hashido, T.; Yashiro, K.; Yugami, H.; Kawada, T.; Mizusaki,

J. J. Ceram. Soc. Jpn. 2005, 113, 562.(7) Kumar, A.; Devi, P. S.; Maiti, H. S. Chem. Mater. 2004, 16, 5562.(8) Shanwen, T.; Jonh, T. S. I. Chem. Mater. 2004, 16, 4116.(9) Levy, C.; Guizard, Ch.; Julbe, A. Sep. Purif. Technol. 2004, 32, 327.

(10) Yin, X.; Hong, L.; Liu, Z.-L. Appl. Catal., A 2006, 300, 75.(11) Trovarelli, A.; de Leitenburg, C.; Boaro, M.; Dolcetti, G. Catal. Today

1999, 50, 353.(12) Zhu, T.; Kundakovic, Lj.; Dreher, A.; Flytzani-Stephanopoulos, M.

Catal. Today 1999, 50, 381.(13) Solinas, V.; Rombi, E.; Ferino, I.; Cutrufello, M. G.; Colón, G.; Navío,

J. A. J. Mol. Catal. A: Chem. 2003, 204–205, 629.

1585Chem. Mater. 2008, 20, 1585–1596

10.1021/cm071809e CCC: $40.75 2008 American Chemical SocietyPublished on Web 01/19/2008

processes (of hydrocarbons8,11,14,16,17 or in wet oxidationprocesses of organic compounds14,18), in the removal of totalorganic carbon from industrial wastewaters,19 in automotiveexhaust gas conversion,11,15,16 in the water-gas shiftreaction,20–23 and in deNOx catalysis.24–26

In our previous work, which can be considered as a partI of our investigation concerning ceria-based mixed oxides,27

we reported the investigation of textural, structural, red-ox,and acid–base characteristics of formulations containing ceriaand other oxides of elements belonging to group III (boria,alumina, gallia, or india). The oxides of group III elementswere chosen as additives to ceria, knowing their catalyticactivities in various industrially important reactions: purealumina, having acidic character, is widely used as a catalystfor several reactions where it activates hydrogen-hydrogen,carbon-hydrogen, and carbon-carbon bonds.28 Addition-ally, boria-based catalysts can be effectively used in theselective oxidation of hydrocarbons such as ethane,29 whilegallium- or indium-supported oxides are promising catalystsfor combustion reactions and NOx abatement processes, suchas the selective catalytic reduction of NOx by hydrocar-bons.15,30–34 The materials were prepared using coprecipi-tation, a method known to provide good dispersion andhomogeneous distribution of mixed metal oxides in thebulk.12,35,36 Indeed, formulations with satisfactory distribu-tions of two oxides were reached; in addition, it was shownthat the properties of mixed oxides depend on the characterand the amount of the group III element.

The necessity to improve the formulations for applicationsin environmentally important processes imposed the tasksof investigating new formulations of ceria-based materialsand modifying the preparation procedures. The insight intothe versatility of preparation procedures applied for obtainingceria-based materials and reported in the literature revealstheir important influence on textural and other properties ofceria and ceria-based materials.

Among the other known techniques, the sol–gel techniquebased on the hydrolysis of alkoxides is one which has beenalso successfully employed for the synthesis of ceria-basedmaterials.32,37–40 The so-called “mild” conditions of the sol–gelmethod provide reaction systems that can be performed underkinetic control. An important hard point using the sol–gelmethod is a possibility to get rid of the organic intermediatesformed during the synthesis. In addition, slight changes in theexperimental parameters can lead to substantial modificationsof the resulting supramolecular assemblies. The resultingnanostructures, their degree of organization, and their propertiesprove that soft-chemistry-based processes offer innovativestrategies to obtain tailored nanostructured materials.37,40

In this work, we report data concerning the preparationof mixed oxides, based on ceria mixed with the same groupIII oxides as published recently: B2O3-CeO2, Al2O3-CeO2,Ga2O3-CeO2, and In2O3-CeO2, but using a sol–gel routeas the method of preparation. The obtained materials werefully characterized to investigate the properties important forcatalytic application; their structural, textural, red-ox, andacid–base characteristics were investigated using a varietyof appropriate techniques.

2. Experimental Section

2.1. Sample Preparation. As already mentioned, cerium-basedmixed oxides were prepared using the sol–gel method. The synthesisof B2O3-CeO2, Al2O3-CeO2, Ga2O3-CeO2, and In2O3-CeO2

samples were performed in isopropanol (Aldrich, 99+, GC purity).The precursor for cerium oxide was a solution of cerium IVtetramethoxyethoxide in methoxyethanol (18-20% wt/wt), suppliedby Alfa Aeser. Taking the high reactivity of the cerium precursortoward hydrolysis into account, no determination was made to verifythe precursor concentration, which can result in important errorsin the mixed oxide composition. The precursors for group III metaloxides were trimethylborate (Fluka, purum 99%), used to obtainboron oxide; aluminum triisopropoxide Al(OiPr)3 (Strem Chemi-cals), used as an alumina source; and indium(III) acetylacetonate(98%, Strem Chemicals), used to produce indium oxide. In the caseof gallia, two routes were used: one from gallium(III) acetylaceto-nate (99.99%, Strem Chemicals), yielding a low surface area catalyst(∼ 1 m2 g-1), and a second one using gallia(III) isopropoxide (99%,Strem Chemicals), yielding better results (surface area ∼ 40 m2

g-1). Among the obtained, only the gallia-ceria sample, having ahigher surface area, was investigated in this work. The hydrolysiswas performed using a solution of ammonia and water (SDS, d )0.92, 20% NH3) dissolved in isopropanol. Typically, 5 g of the

(14) Damyanova, S.; Bueno, J. M. C. Appl. Catal., A 2003, 253, 135.(15) Ozaki, T.; Masui, T.; Machida., K.-I.; Adachi, G.-Y.; Sakata, T.; Mori,

H. Chem. Mater. 2000, 12, 643.(16) Aneggi, E.; de Leitenburg, C.; Dolcetti, G.; Trovarelli, A. Catal. Today

2006, 114, 40.(17) Zhao, S.; Gorte, R. K. Appl. Catal., A 2004, 277, 129.(18) Garcia, T.; Solsona, B.; Taylor, S. Catal. Lett. 2005, 105, 183.(19) Mishra, V. S.; Mahajani, V. V.; Joshi, J. B. Ind. Eng. Chem. Res.

1995, 34, 2.(20) Weibel, M.; Garin, F.; Bernhardt, P.; Maire, G.; Prigent, M. Stud.

Surf. Sci. Catal. 1991, 71, 195.(21) Zafiris, G. S.; Gorte, R. J. J. Catal. 1993, 139, 561.(22) Jacobs, G.; Khalid, S.; Patterson, P. M.; Sparks, D. E.; Davis, B. H.

Appl. Catal., A 2004, 268, 255.(23) Trovarelli, A. Catal. ReV. Sci. Eng. 1996, 48, 439.(24) Kawi, S.; Tang, Y. P.; Hidajat, K.; Yu, L. E. J. Metastable Nanocryst.

Mater. 2005, 23, 95.(25) Neylon, M. K.; Castagnola, M. J.; Castagnola, N. B.; Marshall, Ch.

L. Catal. Today 2004, 96, 53.(26) Baudin, F.; Da Costa, P.; Thomas, C.; Calvo, S.; Lendresse, Y.;

Schneider, S.; Delacroix, F.; Plassat, G.; Djega-Mariadassou, G. Top.Catal. 2004, 30/31, 97.

(27) Yuzhakova, T.; Rakic, V.; Guimon, C.; Auroux, A. Chem. Mater. 2007,19, 2970.

(28) Knözinger, H.; Ratnasamy, P Catal. ReV. Sci.Eng. 1978, 17, 31.(29) Colorio, G.; Védrine, J. C.; Auroux, A.; Bonnetot, B. Appl. Catal., A

1996, 137, 55.(30) Gervasini, A.; Perdigon-Melon, J. A.; Guimon, C.; Auroux, A. J. Phys.

Chem. B 2006, 110, 240.(31) Postole, G.; Gervasini, A.; Caldararu, M.; Bonnetot, B.; Auroux, A.

Appl. Catal,. A 2007, 325, 227.(32) Zahir; Md, H.; Katayama, S.; Awano, M. Mater. Chem. Phys. 2004,

86, 99.(33) Park, P. W.; Ragle, C. S.; Boyer, C. L.; Lou Balmer, M.; Engelhard,

M.; McCready, D. J. Catal. 2002, 210, 97.(34) Perdigon-Melon, J. A.; Gervasini, A.; Auroux, A. J. Catal. 2005, 234,

421.(35) Plassat, G.; Djega-Mariadassou, G. Top. Catal. 2004, 30/31, 97. Yu,

J. C.; Zhang, L.; Lin, J. Interface Sci. 2003, 260, 240.(36) Lee, J.-S.; Choi, K.-H.; Ryu, B.-K.; Shin, B.-C.; Kim, I.-S. Mater.

Res. Bull. 2004, 39, 2025.

(37) Levy, C.; Guizard, Ch.; Julbe, A. Sep. Purif. Technol. 2004, 32, 327.(38) Labery-Robert, C.; Long, J. W.; Lucas, E. M.; Pettigrew, K. A.; Stroud,

R. M.; Doescher, M. S.; Rolison, D. R. Chem. Mater. 2006, 18, 50.(39) Rane, N.; Zou, H.; Buelna, G.; Lin, J. Y. S. J. Membrane Sci. 2005,

256, 89.(40) Khalil, K. M. S.; Elkabee, L. A.; Murphy, B. Microporous Mesoporous

Mater. 2005, 78, 83.

1586 Chem. Mater., Vol. 20, No. 4, 2008 Bonnetot et al.

ceria precursor (15 mmol of pure cerium(IV) methoxyethoxide)was dissolved in 100 mL of isopropanol in a three-neck balloonunder an argon flow. The required amount of the group III oxideprecursor was added after dissolution in 50 mL of isopropanol undera continuous argon flow. The solution was brought to reflux, anda solution of 3 mL of concentrated ammonia and water dissolvedin 50 mL of isopropanol was added dropwise. The addition tookabout 1 h, and the refluxing was performed overnight. After cooling,the “hard solution” was filtered (high-porosity glass filter), and awaxy gel was recovered. This gel was dried under low pressure (1Pa) for at least 12 h at room temperature (RT). A light brownpowder was recovered and ground before calcination to eliminatethe organic species residues. To determine the suitable temperatureof calcination, thermogravimetry (TG) was performed on the crudesamples.

Pure ceria was obtained by precipitation from a solution of(NH4)2Ce(NO3)6 and dilute ammonia (28 wt %/wt), which wasgradually added dropwise to a solution of cerium salt, with vigorousstirring, until the precipitation was complete (at pH ) 8). Then,the precipitate was filtrated and washed with hot distilled water(0.5 L) and flushed twice in an absolute ethanol solution in orderto facilitate agglomeration, followed by drying overnight in an ovenat 110–120 °C. Subsequently, ceria was calcined in the air at 500°C for 5 h (the temperature applied for calcination has been chosenon the basis of TG measurements).

The samples investigated in this work are denoted as XCe-n,where X denotes B, Al, Ga, or In and n is a number representingthe weight percentage of the group III metal oxide in the respectivesample. The list of prepared catalysts and their legends aresummarized in Table 1.

2.2. Characterization. The methods used for characterizationof the samples are the same as those applied for characterizationof the materials investigated in our previous work, devoted to thesame type of Me2O3-CeO2 mixed oxides but synthesized by acoprecipitation route. Here, these experimental procedures arebriefly summarized, while more details can be seen in ref 27.

The contents of nonmetals (B) or metals (Al, Ga, and In) andmetallic oxides in the prepared mixed oxides were determined usingthe atomic emission spectroscopy-inductively coupled plasma(AES)-ICP technique, using a SpectroflameICP instrument. Theirstructure and morphology were investigated using X-ray diffraction(XRD; the patterns were recorded on a Bruker (Siemens) D5005diffractometer at room temperature using Cu KR, from 3° to 80°2θ in 0.02° steps); surface areas were determined using theBrunauer-Emmett-Teller (BET) method from the low-temperatureadsorption of nitrogen, performed at -196 °C after appropriatepretreatment. In addition, the morphologies of the samples havebeen examined using a scanning electron Jeol 55 CR microscope.

In order to determine the lowest temperature needed for thecalcination of fresh samples, thermogravimetric/differential ther-mogravimetric analysis (TG-dTG) was applied. These measure-ments were performed using “Labsys - TG” Setaram equipment;the crude samples were heated from 25 up to 500 °C with a heatingrate of 5°/min in a flow of air.

The Raman spectra of investigated samples were collected on aDILOR XY spectrometer, equipped with a liquid-nitrogen-cooled

charge-coupled device detector. The excitation was provided bythe 514.5 nm line of an Ar + Kr+ ion laser (Spectra Physics),keeping the sample under a microscope, under ambient atmosphericconditions. The wavenumber values reported from the spectra areaccurate to within 2 cm-1.

The analysis of elemental surface concentrations was performedby means of the X-ray photoelectron spectroscopy (XPS) technique,which was done using a Surface Science Instruments (SSI) 301spectrophotometer using a monochromatic Al KR radiation source,equipment that is fully described elsewhere.41 The XPS analysisdepth was around 5–10 nm. The XP spectra of Al (2p), O (1s), Ce(3d5/2), B (1s), Ga (2p3/2), and In (3d5/2) were recorded in detail.The binding energy of the Ce3d5/2 line (CeO2), chosen as the internalreference, is 882.2 eV for all samples, and it corresponds to Ce4+.42

The red-ox properties of the investigated samples have beenstudied by temperature-programmed reduction/temperature-pro-grammed oxidation (TPR/TPO) experiments. The system is equippedwith a thermal conductivity detector. TPR/TPO experiments wereperformed from RT to 830 °C with heating rate of 5 °C min-1

under a H2 (5%)/Ar flow, or O2(1%)/He flow, respectively. ForInCe-24, TPR/TPO experiments were additionally performed upto 500 °C. The details about the experimental setup, the quantifica-tion of the signal, and the acquisition of the data are given in refs27 and 43.

The acid–base properties were studied using the adsorption ofprobe molecules (NH3 and SO2) at 80 °C by means of twoappropriate methods: adsorption microcalorimetry and XPS.

Adsorption microcalorimetry is a method for describing in detailthe features of surface sites, simultaneously from a quantitative andan energetic viewpoint.44,45 Additionally, XPS investigations werecarried out in order to acquire additional data about the nature ofsurface-active sites. It is well-known that the XPS gives informationon the catalysts’ surface composition and on the distribution andelectronic/oxidation state of the elements in the surface layer.42

Moreover, XPS can be applied in order to estimate the characterof surface-active sites. In this work, ammonia and sulfur dioxidewere chosen to probe the acidity and basicity of the catalysts,respectively.

Microcalorimetric measurements were performed using thecoupled microcalorimetric-volumetric system, which enables thedetermination of heats evolved as a result of probe molecules’adsorption, and a quantitative determination of the amounts ofadsorbed gases. For these investigations, mixed oxides werepretreated in a quartz cell by heating overnight under a vacuum at400 °C with a heating rate of 0.6 °C min-1. Subsequently, theadsorption of a gas probe was carried out isothermically at 80 °Cin a heat flow calorimeter (Setaram C80), coupled with a volumetricline equipped with a Barocel capacitance manometer for pressure

(41) Keranen, J.; Guimon, C.; Iiskola, E.; Auroux, A.; Niinisto, L. J. Phys.Chem. B 2003, 107, 10773.

(42) Galtayries, A.; Sporken, R.; Riga, J.; Blanchard, G.; Caudano, R. J.Electron. Spectrosc. Relat. Phenom. 1998, 88–91, 951.

(43) Jouguet, B.; Gervasini, A.; Auroux, A. Chem. Eng. Technol. 1995,18, 243.

(44) Auroux, A. Top. Catal. 1997, 4, 71.(45) Solinas, V.; Ferino, I. Catal. Today 1998, 41, 179.

Table 1. Chemical Analysis and BET Surface Areas for Investigated Samples

sample labelICP/Me3+

(wt %)ICP/Me2O3

(wt %)atomic ratio

Me/Ce, ICP (bulk)atomic ratio

Me/Ce, XPS (surface)BET surface

area (m2 g-1)

B2O3-CeO2 BCe-36 11.3 36.4 1.96 4.50 77Al2O3-CeO2 AlCe-36 18.9 35.7 1.26 3.40 269Ga2O3-CeO2 GaCe-16 12.1 16.2 0.26 1.00 40In2O3-CeO2 InCe-24 19.8 24.0 0.21 0.80 71CeO2 CeO2 88

1587Chem. Mater., Vol. 20, No. 4, 2008Preparation and Characterization of Me2O3-CeO2

measurements. Successive doses of gas were sent to the sampleuntil a final pressure of 67 Pa was obtained. Two adsorptionisotherms were monitored: after the first one was completed, thesample was evacuated at 80 °C. In that way, physically adsorbedspecies were removed, and a second adsorption run was performedup to the equilibrium pressure of 27 Pa. The difference betweenthe amounts adsorbed in the first and second adsorptions at 27 Parepresents the irreversibly chemisorbed amount (Virr) of a respectivegas, which provides the estimation of the number of strong acidic/basic sites.

For the investigation of NH3 or SO2 adsorptions by XPS, thesamples were activated overnight under helium at 350 °C andsubsequently exposed to one of the probes, at 80 °C. The adsorptionwas followed by desorption over 1 h under a helium atmosphereat the same temperature. The XPS measurements were performedat room temperature, preserving inert conditions during all samplehandling steps (with the help of a glovebox coupled with theintroduction chamber of the spectrometer). NH3 or SO2 adsorptionswere investigated through the recording of N1s or S2p lines,respectively. The binding energy of the Ce (3d5/2) line (CeO2) at882.2 eV was chosen as the internal reference.

3. Results and Discussion

3.1. Morphology and the Structure of the Samples. Thegels recovered after synthesis were dried under low pressure(10-2 mbar) at room temperature and subsequently calcinedin the air. Keeping in mind that a high temperature ofcalcination might affect the morphology of the sample,46 thetemperature suitable for calcination was chosen on the basisof results obtained from the thermogravimetric experiments,as the lowest temperature at which the mass loss wasstabilized. The samples investigated in this work lost about20–22 wt %; this relatively high mass loss could be expected,since only a drying procedure at room temperature wasapplied after a sol–gel procedure. As an example, Figure 1presents TG and dTG profiles recorded during the temper-ature-programmed heating of the In2O3-CeO2 sample con-taining 24% india up to 500 °C. It is evident that thermo-grams are complex. One low-temperature process (in therange of 80–100 °C) which is due to the loss of molecularlyadsorbed water from the surface has been noticed. Oneadditional process was found in the range 200-300 °C; most

probably, it can be attributed to the decomposition and theloss of impurities from the bulk.

Generally, the results of TG-dTG experiments haveshown that the weight loss was stabilized in the range380-450 °C. On the basis of these data, 450 °C was chosenas the temperature of calcination, which was performed over5 h, in the air.

The hydrophilic character of the BCe-36 sample wasstudied: after this sample was calcined at 450 °C, it remainedfor several days at open atmospheric conditions. The weightloss obtained using the TG method afterward stems fromthe physisorbed water, and it was similar to the TG profileof a crude sample, in the same temperature range.

The samples synthesized in this work are listed in Table1. The results of chemical analysis, performed either by theICP or XPS method, and the values of BET surface areas ofcalcined samples, are also presented in Table 1. It can benoticed that the samples obtained by the sol–gel method,investigated in this work, possess importantly differentsurface areas which seem to be dependent on the nature ofthe group III metal oxide, but also on its content. As in thecase of the samples synthesized using coprecipitation, thehighest surface is found for the Al2O3-CeO2 mixed oxide:it is significantly higher in comparison with pure ceria,indicating the high dispersion of Al2O3 in the mixed-oxideformulation. Quite to the contrary, the surface area ofGa2O3-CeO2 is significantly lower than that of pure ceria,while B2O3-CeO2 and In2O3-CeO2 express surface areassimilar to that of CeO2. In this work, we synthesized theGa2O3-CeO2 catalyst which contains the same amount ofgallia as the one prepared by coprecipitation (16 wt %).27

The sample prepared by coprecipitation expressed greatersurface area (117 m2/g), in comparison with that synthesizedhere using the sol–gel route. The obtained results confirmthat the preparation route has an important influence on theachieved mixed oxide and its morphology, as already knownfrom the literature data.

The XRD patterns of various mixed oxides prepared andcalcined at 450 °C showed only the diffraction lines of afluorite structure of CeO2,47 as seen in Figure 2, which

(46) Kung, H. H. Stud. Surf. Sci. Catal. 1989, 45, 121.(47) Reddy, B.; Khan, A.; Yamada, Y.; Kobayashi, T.; Loridant, S.; Volta,

J.-C. J. Phys. Chem. B 2003, 107, 11475.

Figure 1. TG and dTG profiles obtained for the InCe-24 sample.Figure 2. XRD patterns of (1) CeO2, calcined at 500 °C, and (2) BCe-36,(3) AlCe-36, (4) GaCe-16, and (5) InCe-24, calcined at 450 °C.

1588 Chem. Mater., Vol. 20, No. 4, 2008 Bonnetot et al.

presents the XRD patterns of pure ceria and all samplessynthesized using the sol–gel method. The presented dif-fractogramms clearly show that the crystallinities of theprepared samples are lower in comparison with that of pureceria: their intensities decrease with addition of the secondoxide; the lowest intensity values were obtained for alumina-and boria-containing samples. Keeping in mind thatAl2O3-CeO2 and B2O3-CeO2 mixed oxides have the sameweight percents of metallic oxides, it might be inferred thatthe degree of crystallinity depends on both the nature andthe amount of the group III metal oxide. Importantly, extralines that could originate from other compounds or mixedphases between ceria and the group III metal oxide werenot noticed in X-ray diffractogramms. However, it can beseen in Table 1 that surface atomic Me/Ce ratios (found usingthe XPS method) differ significantly from those obtained forbulk samples (found using the ICP method). These resultsindicate that the existence of some surface aggregates of thegroup III metal oxide, not detectable using the XRDtechnique, is possible.

Indeed, Figure 3, which shows the SEM image obtainedfor the InCe-24 sample investigated in this work (after thecalcination done in the air), indicates the possible existenceof such surface species. The appearance of some surfacesegregations of the group III metal oxide can explain the factsthat the BET surface area found for this sample is lowerthan that noticed for pure ceria (see Table 1).

Figures 4a and b present Raman spectra collected for allthe samples synthesized applying a sol–gel route. Additio-anlly, a Raman spectrum of pure CeO2 is also presented inthe same figure. All spectra exhibit main peaks at 270, 465,and 600 cm-1 already assigned to CeO2 phase in theliterature.47 In consistence with XRD measurements, Ramanlines that could be assigned to Me2O3 were not observed.However, the decrease in the intensities of the main peakassigned to ceria (centered at 465 cm-1) can be observed.The following order in the intensity decrease is found for

the samples investigated in this work: CeO2 > InCe-24 >GaCe-16 > BCe-36 > AlCe-36; this order is in agreementwith XRD data, where the decrease in the crystallinities canbe seen with doping by the second metal oxide.

In addition, it is important to notice here that the mainpeak assigned to ceria (at 465 cm-1) was shifted to lowerwavenumbers with the addition of a second metal oxide; theinteraction between oxides could be the explanation for thatshift. As it has been stated already, XRD measurements didnot show any evidence of third compounds forming. Eventhough, this shift in position of the peak at 465 cm-1

indicates that some interaction between oxides happened. Theinsight into the spectra presented in Figure 4 reveals evidencethat the shift of the peak centered at 465 cm-1 differs forthe samples investigated in this work; which is a clearindication that the character of the group III element playsan important role in the final structure of mixed oxides.

Another important indication was observed from Ramanspectra: it is known from the literature that a band recordedat 600 cm-1 is associated with oxygen vacancies in theCeO2-x lattice. Being also related with oxygen mobility, thispresence of oxygen vacancies is important because of thepossible role of these sites in combustion reactions. There-fore, it is important to notice here that, in the case of thesamples prepared by the sol–gel technique, this band becamemore pronounced for india-ceria mixed oxide, while in thecase of boria-ceria and alumina-ceria, the intensity of thispeak was significantly decreased (Figure 4b). Generally, itcan be pointed out that the intensities of all three peaks thatcan be observed in Raman spectra showed the highestdecreases in the case of alumina-ceria.

3.2. Red-Ox Properties. In this work, the investigationof reduction/oxidation properties was performed for theIn2O3-CeO2 sample, using the TPR/TPO method. Withinthe investigated Me2O3, only the indium oxide can be reducedat relatively low temperatures: the complete reduction of bulkindium oxide and the formation of metallic indium occur inone step,48 while in the case of In2O3 loaded on γ-alumina,it has been shown that the complete reduction of pure indiaproceeds in one step with a maximum at 755 °C.33 Thereduction temperature of indium oxide has also been foundto be related to particle size: for In2O3 supported on alumina,two different indium oxide species have been found for theformulations containing more than 2.5 wt % indium. It hasbeen accepted that the smallest particles are reduced at lowtemperatures.33,34 TPR profiles for pure CeO2 have alsoalready been published in the literature;14,18,49–52 the reduc-tion of this material occurs in two temperature regions:between 350 and 500 °C—the peaks found in this temper-ature region have been assigned to the reduction of surfaceCe4+ to Ce3+-and in the region above 600 °C, where thereduction of bulk CeO2 begins. Here, the reduction range

(48) Mihalyi, R. M.; Beyer, H. K.; Mavrodinova, V.; Minchev, Ch.;Neinska, Y. Microporous Mesoporous Mater. 1998, 24, 143.

(49) Piras, A.; Trovarelli, A.; Dolcetti, G. Appl. Catal., B 2000, 28, L77.(50) Damyanova, S.; Perez, C. A.; Schmal, M.; Bueno, J. M. C. Appl.

Catal., A 2002, 234, 271.(51) Gu, X.; Ge, J.; Zhang, H.; Auroux, A.; Shen, J. Thermochim. Acta

2006, 451, 84.(52) Trovarelli, A.; Dolcetti, G.; de Leitenburg, C.; Kašpar, J.; Finetti, P.;

Santoni, A. J. Chem. Soc., Farady Trans. 1992, 88, 1311.

Figure 3. SEM micrograph of InCe-24, previously calcined in the air at450 °C.

1589Chem. Mater., Vol. 20, No. 4, 2008Preparation and Characterization of Me2O3-CeO2

from RT up to 830 °C was chosen for the indium-containingcatalyst since the volatilization of In2O3 takes place at 850°C. The obtained results are in agreement with the findingspreviously stated in the literature.

Figure 5a shows the hydrogen uptakes as a function ofthe temperature for InCe-24, prepared using the sol–gel route,and one obtained for pure ceria, which is performed for the

sake of comparison, although this sample has been synthe-sized by the precipitation route. The TPR profile of pure ceriashows a broad peak at 475 °C, thus proving that the reductionof surface ceria particles from Ce4+ to Ce3+ happened. Thereduction of bulk ceria only begins in the temperature regionapplied for TPR experiments here, as can be seen from theappearance of another incomplete peak which started at600 °C.

In the TPR profile obtained for InCe-24, one well-definedsymmetrical peak with a maximum at 345 °C can beobserved, and a broad unfinished process that starts from500 °C and expresses a maximum at about 700 °C. It isimportant to notice that this profile is different in comparisonwith those obtained in our previous investigation that wasfocused on india-ceria mixed oxides, containing between4.5 and 12 wt % india, obtained by coprecipitation.27 There,the appearance of a low-temperature complex TPR profile(below 300 °C), composed of two overlapped peaks, wasnoticed. The reduction in this low-temperature region hasbeen explained by the existence of well-dispersed india onthe surface of ceria, while the appearance of a complex TPRprofile has been assigned to the reduction of surface indiumparticles having different sizes; the observed changes of theTPR profile with the increase in india content were inter-preted as a result of the changes in the distribution of indiaparticles.

The TPR profile obtained for the sample investigated inthis work, which contains 24 wt % india, confirms theimportance of indium oxide content in the reductionprocess—the profile is centered at a higher temperature (at345 °C), in comparison with the results found for copre-cipitated samples. Importantly, the TPR peak is narrow, withone well-resolved maximum, which can be an indication ofuniform distribution of In2O3 particles. Additionally, thebroad reduction peak formed at 700 °C can be assigned tothe reduction of bulk indium oxide. It is important to noticethat this high-temperature peak is shifted to a lower reduction

Figure 4. Raman spectra of mixed oxide samples and the CeO2 sample: (1) InCe-24, (2)GaCe-16, (3) BCe-36, (4) AlCe-36, and (5) CeO2.

Figure 5. (a) TPR profiles of (1) CeO2 and (2) InCe-24, performed up to830 °C. (b) TPO profiles of (1) CeO2 and (2) InCe-24, performed up to830 °C.

1590 Chem. Mater., Vol. 20, No. 4, 2008 Bonnetot et al.

temperature, in comparison with the result obtained for pureindia,33 thus indicating that the presence of ceria can lowerthe reduction temperature for india particles in the bulk.

The extent of reduction was calculated and compared tothe theoretical H2 consumption of these samples, assumingthat the total reduction of In3+ to metallic In0 requires 3 molof hydrogen per mole of the india phase (In2O3 + 3H2 f2In + 3H2O), while in the case of ceria, the reduction ofCe4+ to Ce3+ requires 1 mol of hydrogen per 2 mol of theceria phase (2CeO2 + H2 f Ce2O3 + H2O). The sampleprepared by the sol–gel method showed a relatively highreduction of india particles (34%).

Figure 5b shows the TPO profiles obtained for pure ceriaand for InCe-24. The bulk ceria sample shows only one peakcentered at 90 °C. As already published in the literature, thispeak can be attributed to the oxidation of the ceria surfacewith the formation of cation vacancies (O2- or O-) or tothe oxidation of Ce2O3 to CeO2. As a result of oxygenconsumption, the InCe-24 sample produced one pronouncedpeak at 345 °C, the same temperature where a maximumfor reduction of india was detected. However, a complexTPO profile was found in the region of higher temperatures:a shoulder of the previous peak with a maximum at 405 °Cand large peak with a maximum centered at 600 °C werefound. Again, it can be noticed that this complex TPO profilefound for the india-ceria sample synthesized by the sol–gelroute is different from that found for india-ceria with 14wt % In2O3 synthesized by coprecipitation, where the mainoxygen consumption was detected in the lower temperatureregion.27 In addition, the appearance of one overlappeddouble peak in the TPO profile (in the region 250-500 °C)can be interpreted by the possible previous sintering ofindium, during the TPR experiment.

The possibility of the existence of some surface segrega-tions of group III oxide has been already noticed from thehigh surface Me/Ce ratios found by XPS. Hence, thedifferences in TPR/TPO profiles that have been noticed, incomparison with already published data obtained for InCe-14,27 could come from higher In2O3 loading and from adifferent distribution of the particles, in the case of InCe-24. Indeed, the morphology of this sample was significantlychanged by the reduction/oxidation cycle: the SEM micro-graph recorded after these experiments (Figure 6) clearlyshows that the forming of spherical particles happened. Mostprobably, these spherical particles could be formed bysintering of the metallic indium species; since india can bereduced more easily than ceria, the process of sintering couldbe also favored by the high surface concentration of indiumat the surface of this sample.

As already mentioned, additional TPR/TPO measurementwas performed up to 500 °C. After both TPR/TPO proce-dures were completed, the structure was checked by XRD.As already shown, only a slight decrease of intensities wasobserved as a result of pure ceria mixing with india (Figure7, patterns 1 and 2). As a result of the TPR/TPO cycleperformed up to 500 °C, some weak intensity diffractionpeaks of In2O3 crystallites appeared (pattern 3).

However, the reduction of InCe-24 up to 830 °C, followedby oxidation up to the same temperature, leads to a better

crystallization of the CeO2 phase (pattern 4, Figure 7), andto the visible appearance of In2O3 crystallites. As a result ofhigh-temperature treatment, oxygen vacancies can be gener-ated, which produces the changes in the CeO2-x fluoritestructure: the shift of the main ceria peak toward lower 2θvalues (2θ ) 28.60–28.70) can be observed, which is inaccordance with already published literature data.47 Hence,high-temperature reduction changes irreversibly the morphol-ogy of the samples, leading to the formation of defects in aCeO2-x fluorite structure. In addition, it is important to noticethat the decreases in the surface area of 40% and 90%happened as a result of TPR/TPO cycles performed up to500 and 830 °C, respectively. It seems that the high amountof india in InCe-24 is a key feature that influences theinstability of this mixed oxide.

From a generalized view on the morphology and red-oxproperties of the samples investigated in this work, it canbe underlined that, evidently, there is an influence of thepreparation route on these features. It is important to pointout here that, although we had planned to prepare samplesthat would have comparable amounts of group III oxides asin ref 27, samples having higher final amounts than expected(except in the case of GaCe-16) have been obtained. Themost probable reason is one difficult step in the preparationprocedure: it is not possible to remove the solvent withoutchanging the composition of the gel, which is caused bydifferent partial solubilities of components into the dispersingmedium. These higher amounts of group III oxides that aremixed with ceria provoked a different dispersion of theirparticles. As already stated, the appearance of aggregateshas been proven by SEM imaging. The existence ofaggregates explains the values of surface areas, which are,except in the case of AlCe-36, lower for all other samplesprepared by the sol–gel technique, in comparison withcorresponding mixed oxides prepared by coprecipitation.Similarly, the red-ox behavior of InCe-24 can be also

Figure 6. SEM micrograph of InCe-24 after the red-ox cycle performedup to 830 °C.

1591Chem. Mater., Vol. 20, No. 4, 2008Preparation and Characterization of Me2O3-CeO2

explained by the same phenomenon: the higher temperaturesthat are necessary for both the reduction and oxidationprocesses, in comparison with InCe-14,27 clearly show thatdispersions of india are not the same in these two samples.

3.3. Surface and Acid/Base Properties. The adsorptionof probe molecules (NH3 and SO2) was used to investigatethe acid–base properties, by means of XPS and microcalo-rimetry. Additionally, those techniques also provide theinformation about the adsorption possibilities of investigatedsamples to capture the pollutants containing ammonia andsulfur dioxide. The information obtained from the investiga-tions can be considered to supplement each other and not asa basis of pure comparison of results, since the principles ofthe experimental techniques (XPS and calorimetric investiga-tions) are not the same. Calorimetric study allows preciseinformation on the concentration of active sites (basic andacidic) present on the catalytic surface, while XPS measure-ment can help to differentiate the types of these sites(Brönsted or Lewis), giving in the same time informationonly on the strong and medium-strength acid sites. Namely,in XPS experiments, titrating of the weak sites, which cannotkeep ammonia molecules under ultrahigh (10-7 Pa) vacuumconditions applied during analyses, is not possible.

3.4. The Analysis of B 1s, Al 2p, In 3d5/2, and Ce3d5/2 XPS Bands. XPS data obtained for investigatedcatalysts are summarized in Table 2. The Me/Ce atomic ratioobtained by XPS (Me/Ce surface) and by ICP (Me/Ce bulk)methods are already given in Table 1. B/Ce, Al/Ce, Ga/Ce,and In/Ce ratios on the surface were determined from theXPS band areas of B1s (Al2p, Ga2p, or In3d5/2) and Ce3d5/2

peaks. Because of the large uncertainty on the XPS atomicpercentage due particularly to the Scofield intensity factors,which are average theoretical values, only the relative atomicratios in each series are considered. The quantitative datafor GaCe-16 are approximate because of the overlapping ofGa2p and the Ce4d bands.

As already explained in the case of mixed oxides samplesobtained by co precipitation,27 the analysis of XP spectra ofCe 3d core-electron levels revealed the existence of CeO2

as well as Ce2O3, although XRD could not detect any visibleCe2O3 lines. The cerium is mainly present in the Ce4+

oxidation state. The presence of Ce(III) can be mainlyattributed to the removal of surface hydroxyl groups andoxygen from the CeO2 surface during exposure of thesamples to the X-rays in the ultra-high vacuum chamber

Figure 7. XRD patterns of CeO2 and InCe-24 pretreated in different ways: (1) CeO2 calcined in air at 500 °C, (2) InCe-24 calcined in air at 450 °C, (3)InCe-24 after TPR/TPO cycles up to 500 °C, and (4) InCe-24 after TPR/TPO cycles up to 830 °C. (*) Crystal particles of In2O3.

Table 2. XPS Data Obtained for the Investigated Samples

BCe-36 AlCe-36 GaCe-16 InCe-24 CeO2

adsorption (80 °C) NH3 SO2 NH3 SO2 NH3b SO2 NH3 SO2 NH3 SO2

Me/Ce 3.9 4.5 3.7 3.4 1.0 0.6 0.8CeIII/CeIV 25/75 20/80 35/65 55/45 25/75 25/75 25/75 25/75 20/80N/(Me+Ce)a 0.003 0.01 0.01 0.008S/(Me+Ce) 0.29 0.4 3.3 0.4 0.44

N1sa

NH2- 397.6 397.2 397.5

(50%) (65%) (40%)Lewis 399.6 400.7 399.2 399.7

(50%) (35%) (60%) (100%)

S2p3/2

L (O2-) 166.7 166.3 166.9 166.7 166.9(5%) (25%) (20%) (35%) (10%)

B (OH-) 168.3 168.3 168.4 168.2 168.3(95%) (75%) (80%) (65%) (90%)

a The values are very approximate because of the very low intensity of the N1s band. b Not done because of the overlap between the N1s band andthe L2M45M45 Auger band.

1592 Chem. Mater., Vol. 20, No. 4, 2008 Bonnetot et al.

(UHV) under mild reduction condition.53 This phenomenonis confirmed by the increase of the intensity of the compo-nents corresponding to Ce(III) with the irradiation time ofthe sample. The AlCe-36 sample showed the highest reduc-ibility of ceria to Ce(III)—up to 35% after NH3 adsorptionand 55% after SO2 adsorption, in comparison with thecorresponding values found between 20 and 25%, for bothNH3 and SO2 adsorptions, in the case of the other samplesinvestigated here. The higher degree of reduction of thissample might be caused by presence of a high amount ofalumina, which is an n-type semiconductor,54 as a conse-quence of electron charge transfer from alumina to ceria.

3.5. The Analysis of S 2p and N 1s XPS. XPS experi-ments have shown that all samples have more basic thanacidic character as the S/(Me + Ce) molar ratios weresignificantly higher than the corresponding N/(Me + Ce).The atomic ratios of S/(Me + Ce) and N/(Me + Ce) weremeasured from the ratios between the intensities of the mainband associated with the adsorbed sulfur dioxide or ammoniamolecules (where S is the S 2p band for SO2 and N is the N1s for NH3), and bands associated with metal oxide (whereMe is Al 2p for Al2O3, B 1s for B2O3, Ga 2p for Ga2O3, In3d5/2 for In2O3, and Ce 3d for CeO2). The quantitative dataand the binding energy of the Ga 2p band are particularlyapproximate because of the overlapping of this peak and theCe 4d band. The same stands for the N/(Ga + Ce) ratioobtained after the adsorption of ammonia, because of theoverlapping between the N 1s band of nitrogen and the Augerband L2M45M45 of gallium.

The quantitative analysis of NH3 adsorption is difficultdue to the very weak concentration of adsorbed ammoniaon all samples and thus the large signal/noise ratio. This cancome from a small concentration of surface acidic sites orfrom a weak strength of these sites (provoking a partialdesorption of NH3 under the UHV conditions of the XPSanalyses). The obtained N/(Me + Ce) ratios are in the range10-2 to 10-3. In addition to NH3-Me species, the formingof NH2

- was detected on the surface of the samples, mainlyAlCe-36, due to the possible dissociative adsorption of anammonia molecule. The dissociative adsorption of NH3

proceeds with involvement of the surface active oxygen: O-

(ad) + NH3 (g)f OH (ad) + NH2-.55 Since the forming of

amide species goes with the participation of basic O- sites,the population of NH2

– might be directly related to thepopulation of oxygen basic sites. Those oxygen species,which are active in the dehydrogenation of ammonia, mightbe active in the H abstraction reaction as well.

From the XPS analysis of ammonia adsorption, it can benoticed that all of the samples investigated in this workexpress very weak acidity, even in the case of boria-ceriamixed oxides. It is important to notice the difference betweenthese results and those obtained for the same types of mixedoxides prepared by coprecipitation.27 There, weak aciditieswere found for alumina-ceria, gallia-ceria, and india-ceria.However, higher values of (N/Me + Ce) ratios were found

for boria-ceria, and a trend of their increase with an increasein boria content was found as well.

The values of binding energies and the percent proportionsof the sulfur species adsorbed on Lewis basic sites(SO2-O2-) or Brönsted basic sites (SO2-OH-) are sum-marized in Table 2. XPS 2p peaks obtained after theadsorption of SO2 at 80 °C on all investigated catalysts areshown in Figure 8. Generally, this peak is decomposed intotwo doublets (2p3/2-2p1/2), as found in the case of mixedoxides prepared by coprecipitation. The relative intensitiesof the doublets seem to be dependent on the character andthe amount of the group III oxide. The first 2p3/2 peakappeared at about 167 eV, and the second peak appeared at169 eV. Like in the case of TiO2,56 the former doublet couldbe assigned to SO2 in interaction with O2- anions, that is,to Lewis basic sites. The latter doublet could correspond tothe interaction of SO2 molecules with hydroxyl (OH-)surface groups (Brönsted sites). The assignment is proposedon the basis of the evolution of the relative intensities of thecomponents with an increase in the activation temperature:

(53) Pamukchieva, V.; Gonbeau, D.; Guimon, M.-F.; Skordeva, E.;Dedryvere, R.; Arsova, D. Physica B (Amsterdam, Neth.) 2006, 371,302.

(54) Szabó, Z. G.; Kalló, D. Contact Catalysis; Akademiai Kiada: Budapest,1976; Vol. I, p 396.

(55) Carley, A. F.; Dollard, L. A.; Norman, P. R.; Pottage, C.; Roberts,M. W. J. Electron. Spectrosc. Relat. Phenom. 1999, 98–99, 223.

(56) Guimon, C.; Gervasini, A.; Auroux, A. J. Phys. Chem. B 2001, 105,10316.

Figure 8. S2p XP spectra of (1) BCe-36, (2) AlCe-36, (3) GaCe-16, and(4) InCe-24 activated at 350 °C, after adsorption of SO2 at 80 °C.

1593Chem. Mater., Vol. 20, No. 4, 2008Preparation and Characterization of Me2O3-CeO2

when the activation temperature is increased from 350 to700 °C, the relative intensity of the first doublet associatedwith the sulfite groups (167 eV for S 2p3/2) increases. Thisevolution can be related to the lesser concentration of OH-

groups after activation at high temperatures (more effectivedehydroxylation of the surface at 700 °C). In the same time,the S/Ce atomic ratio decreases, as an important part ofBrönsted sites dissapears while the concentration of Lewiscenters remains quasi-constant. The relative ratio of B(OH-)and L(O-) bands shows that the main part of the surfacebasicity for all samples originates from surface hydroxylgroups. The highest percent of B(OH-) sites is found onthe boria-ceria surface (95%), while the lowest foundpercentage is 65% for india-ceria.

3.6. Microcalorimetry Results. Due to noticeable dif-ferences in BET surface area values, the comparison of thecalorimetric results were presented in µmol · m-2 and notin µmol · g-1.

Table 3 compiles the data obtained from microcalori-metric measurements of SO2 and NH3 adsorption on allinvestigated samples. The table presents the total amountsof adsorbed gases, the amounts of irreversibly adsorbed(chemisorbed) gases, and the amounts of reversiblyadsorbed (physisorbed) gases. As already mentioned inthe Experimental Section, by subtracting the adsorbedvolume of the secondary isotherm (not presented in thefigure) from that of the primary isotherm at the sameequilibrium pressure (P ) 27 Pa), the amount of irrevers-ibly adsorbed gas was calculated. The vertical parts ofthe isotherms correspond to irreversible adsorption, whilethe horizontal parts can be assigned to reversibleadsorption.44,57

As already published in our previous contribution27 andin the literature, pure ceria is an amphoteric material. It canbe seen from Figures 9 and 10 and from Table 3 that CeO2

adsorbed both NH3 and SO2 with high initial heats ofadsorption, and that the adsorbed amounts of both probeswere important. Importantly, the results obtained in this workshow that the acid–base character of ceria was influencedby the addition of a group III metal oxide.

Figure 9 presents the isotherms obtained for SO2 adsorp-tion on the investigated samples, and the profiles of dif-ferential heats versus surface coverage. It can be concluded

from the results presented in Table 3 and in Figure 9 thatthe type of group III metal oxide has a decisive role in thebasic character of the investigated mixed oxides. A briefinsight into the results presented by Figure 9 and the amountsof adsorbed SO2 (both reversible and irreversible) revealsevidence that the addition of boria and alumina decreasedthe bacisity of pure ceria, while the addition of india andgallia led to, as a consequence, an increase in basic character.Evidently, the lowest and highest adsorbed amounts of SO2

were found for boria- and gallia-containing samples,respectively.

Ceria investigated in this work exhibited strong basicity:the initial heats of SO2 adsorption are very high (>200kJ mol-1). Plateau-like profiles with a slight decrease ofdifferential heat values up to ∼180 kJ mol-1 were foundfor CeO2, InCe-24, and GaCe-16. Evidently, gallia-ceriaand india-ceria mixed oxides possess basic sites whichare of the same strength as CeO2. Thus, it can beconcluded that the addition of In2O3 and Ga2O3 increasedthe number of sites active for SO2 adsorption but did notinfluence their strength. The shape of the reversible partof the isotherm of the GaCe-16 sample is quite differentfrom the others, showing a sudden increase for SO2

pressure higher than 20 Pa. This can be attributed to aspecific interaction of SO2 with the gallia-ceria. It is alsoimportant to notice that the values of differential heats

(57) Dragoi, B.; Gervasini, A.; Dumitriu, E.; Auroux, A. Thermochim. Acta2004, 420, 127.

Table 3. Calorimetric Measurements for SO2 and NH3 Adsorption

SO2 adsorbedamount at 80 °C/µmol m-2

NH3 adsorbedamount at 80 °C, µmol m-2

sample Nta Nreadb Nirrc Nt Nread Nirr

BCe-36 0.74 0.32 0.42 0.81 0.60 0.21AlCe-36 3.29 0.18 3.11 1.47 0.92 0.55GaCe-16 6.6 0.82 5.78InCe-24 5.25 0.15 5.10 1.33 0.75 0.58CeO2 3.73 0.50 3.23 1.88 1.05 0.83

a Total amount of gas (µmol m–2) adsorbed under an equilibriumpressure of 27 Pa. b Physisorbed amount of gas (µmol m-2) determinedunder equilibrium pressure of 27 Pa by readsorption after pumping.c Number of strong sites corresponding to the irreversibly adsorbed gasamount (Nirr), calculated by subtracting the primary (Nt) and secondary(Nread) isotherms (µmol m-2).

Figure 9. The isotherms and the profiles of differential heats obtained as aresult of SO2 adsorption.

1594 Chem. Mater., Vol. 20, No. 4, 2008 Bonnetot et al.

indicate the existence of weaker sites for SO2 adsorptionon boria-ceria and alumina-ceria, in comparison withpure ceria.

All mixed oxides investigated in this work expressed onlyweak acidity. It is clearly evidenced by the adsorptionisotherms and the differential heats profiles of ammoniaadsorption, presented in Figure 10, that pure ceria expressedhigher acidity than any other catalysts investigated here. Itcan be seen from Table 3 that the highest values of reversiblyand irreversibly adsorbed ammonia are found in the case ofCeO2.

It is important to notice that, in our previous investigation,significant acidity was noticed for boria-ceria samples.27

In addition, the increase of acidic character was evidencedas a consequence of increasing the boria content, from 6 to17 wt %. However, even if in this work the boria-ceriasample had a higher amount of boria (36 wt %), this sampleexpressed only insignificant acidity: a small number of weaksites for NH3 adsorption was found (see Table 3 and Figure10). It seems that a high amount of boria in the sample wasa factor which produced an unfavorable distribution of boriaparticles in the mixed oxide; in that way, the sites active inNH3 adsorption were blocked.

It has to be pointed out here that all mixed oxidesinvestigated here express basic character: the results obtainedfrom microcalorimetry are in agreement with the dataobtained from XPS investigations. Let us recall here alsothat the XPS spectra of SO2 adsorption showed mainly thepresence of Brönsted basic sites on the catalytic surface.

Generally, as already stated concerning morphology andred-ox properties, it seems that acid–base properties are also

influenced by the applied preparation method—the acid–basefeatures of investigated mixed oxides are quite different fromthose of the samples obtained by coprecipitation. Thisinfluence is most evident in the case of GaCe-16: even ifthe composition is the same as for the sample obtained bycoprecipitation, it expresses an increased basicity. It is knownthat both ceria and GaCe-x, obtained by coprecipitation,express amphoteric character. The analysis of the resultspresented here reveals that the applied procedure is decisivefor the dispersion/mixing of both group III oxide and ceria.In the case of GaCe-16, it seems that a specific situation isachieved by applying the sol–gel procedure: the contributionof acidic sites is decreased, meaning at the same time thatthe influence of ceria in the mixture is minimized.

4. Conclusions

The properties of B2O3-CeO2, Al2O3-CeO2, Ga2O3-CeO2

,and In2O3-CeO2 mixed oxides investigated in this work havebeen shown to depend on the group III element—its characterand its amount—and on the preparation procedure. High specificsurface area and a good dispersion were reached in the case ofAlCe-36, while it seems that, for the other formulations of mixedoxides investigated here, the high amounts of group III oxidesprovoked a more or less pronounced decrease in surface areasand homogeneity.

For all synthesized mixed oxides, only a fluorite crystalstructure of CeO2 was detected; the crystallinity of ceriadiminishes with the presence of the second group III metaloxide. Importantly, XRD measurements did not show anyevidence of third compounds forming; however, from Ramanspectroscopy measurements, some interactions between twooxide phases have been evidenced for all investigated samples.Raman measurements provided also evidence of oxygen vacan-cies present on the surface of ceria. Importantly, it has beenfound that the concentration of these species is higher for theindia-ceria sample, in comparison with pure ceria, indicatingthat this catalyst could be possibly active in environmentallyimportant combustion reactions. Raman spectroscopy experi-ments have also shown that Al2O3-CeO2, although expressingvery high specific surface area and a good homogeneity, doesnot contain oxygen vacancies.

The TPR/TPO profiles were examined for the InCe-24sample, in relation to the nature of the phases and the extentof reduction. After performing a red-ox cycle up to hightemperatures, the structure and morphology of this samplewere irreversibly changed: crystalline structures of In2O3 andCeO2 were formed by the process of sintering. InCe-24 ismore stable after applying a reduction/oxidation cycle up to500 °C, although the crystallization of In2O3 started even atthis temperature, to a small extent.

The results obtained from the microcalorimetry of ammoniaand sulfur dioxide confirmed the amphoteric character of ceria,which had already been known from the literature. Both XPSand microcalorimetry experiments have shown that theMe2O3-CeO2 samples synthesized by the sol–gel route alsoexpress amphoteric features. However, it was evidenced thatall investigated samples are less acidic than ceria, while thebasic character is diminished for boria-ceria and alumina-ceriaformulations and remained similar to that of ceria for gallia-ceria

Figure 10. The isotherms and the profiles of differential heats obtained asa result of NH3 adsorption.

1595Chem. Mater., Vol. 20, No. 4, 2008Preparation and Characterization of Me2O3-CeO2

and india-ceria. The surface basicity comes mainly from thesites of Brönsted type, while the surface acidity is mainly ofLewis type; these data could be of particular importance forpossible catalytic applications.

Acknowledgment. This article is especially dedicated toBernard Bonnetot who passed away on December 25, 2006. This

work was supported by a NATO grant. Special thanks are addressedto Dr. Stephane Loridant, Institut de Recherches sur la Catalyze etl’Environnement de Lyon, for assistance in Raman measurementsand to the technical services of IRCELYON for valuable contribu-tion in XRD and chemical analysis.

CM071809E

1596 Chem. Mater., Vol. 20, No. 4, 2008 Bonnetot et al.