ARTICLES Sintering and Phase Transformation of V-Loaded Anatase Materials Containing Bulk and Surface V Species J. P. Balikdjian, ² A. Davidson,* ,² S. Launay, ‡ H. Eckert, § and M. Che ², | Laboratoire de Re ´ actiVite de Surface, UMR 2609 CNRS, UniVersite ´ Pierre et Marie Curie, Paris, 75252 Cedex 05, France, Laboratoire de Cristallochimie du solide, UniVersite ´ P. et M. Curie, Paris, 75252 Cedex 05, France, Institut fu ¨ r Physikalische Chemie, Schlossplatz 7, D-48149 Munster, Germany, and Institut UniVersitaire de France ReceiVed: February 11, 2000; In Final Form: May 9, 2000 V-loaded anatase materials with V/Ti molar ratios in the range 0.02-0.06 have been prepared by a sol-gel procedure under inert atmosphere. Combining macroscopic (XRD, TEM, and BET) and spectroscopic (XPS, Raman, ESR, and 51 V NMR) techniques, we show that most of the V species are incorporated within the anatase bulk (amount of surface V species below the sensitivity limit of XPS, i.e., lower than 0.5-1 at. %). As revealed by XPS and in situ XRD measurements, upon calcination below 600 °C in flowing oxygen, a migration of V species from the anatase bulk toward its surface, associated with a sintering of anatase particles, is induced. Above 600 °C, V x Ti 1-x O 2 rutile solid solutions are formed. The threshold temperature at which anatase particles transform into rutile is affected by the V/Ti molar ratio. Located near 650 ( 50 °C in our experimental conditions for V-free titania, this temperature regularly increases with V loading and can be located at 750 ( 50 °C for a V/Ti molar ratio of 0.06. Furthermore, on a parallel experiment performed on a mechanical mixture of V-free titania and vanadia (V/Ti molar ratio of 0.05), we have observed that this temperature decreases to 550 ( 50 °C. These observations strongly suggest that the decrease in the threshold temperature at which V-loaded anatase materials are transformed into rutile, which has been previously observed, extensively discussed in the literature and attributed to vanadium, is not actually due to V species incorporated within the anatase bulk. Our results rather suggest that rutile formation is initiated near the surface of anatase particles and that a minimal amount of surface V species is required to decrease the threshold temperature at which rutile can be detected. Introduction Heterogeneous catalysts in which a titania (TiO 2 ) support is associated with surface vanadium oxide are largely used both for oxidation (o-xylene transformation in phthalic anhydride) 1 and reduction (selective catalytic reduction SCR of nitrogen oxides in the presence of ammonia) reactions. 2,3 They exist in two forms, V/TiO 2 (anatase) and V/TiO 2 (rutile). The latter form has received less attention than the former one because catalytic studies have suggested that anatase gives more active and selective catalysts than rutile. 4-6 V/TiO 2 (anatase) catalysts are generally obtained by deposit- ing V species on commercial anatase by impregnation techniques. 7-11 Equilibrium deposition followed by filtration procedures are also often used. 12 The overall V content that can be introduced by the last techniques is largely limited by the surface area of anatase. Segregated vanadia (V 2 O 5 ) crys- tallites are evidenced, once a nominal vanadia monolayer covering the anatase surface is exceeded. 13 Sol-gel chemis- try, 14-16 coprecipitation, 17-19 and laser induced pyrolysis 20 methods have opened the possibility of generating small anatase particles containing highly accessible V V species and therefore supposedly giving high reaction rates per gram of catalyst. Unfortunately, the anatase form of titania is metastable and transforms into the thermodynamically stable rutile form upon calcination. Whatever their origin, it has been found that small anatase particles (specific surface areas typically greater than 100 m 2 /g) transform into rutile more easily than larger particles. For instance, commercial TiO 2 (Tioxide, 122 m 2 /g) is converted at 60% into rutile after a calcination at 620 °C for 100 min, whereas a low area anatase powder (10 m 2 /g) does not convert into rutile at this temperature, even if it is maintained for 1000 min. 10 Furthermore, it has been observed that vanadium is significantly decreasing the threshold temperature at which rutile is formed. For instance, a V-free anatase reference obtained by laser induced pyrolysis of Ti alkoxide has been found to be maintained in its anatase form after a calcination at 530 °C (4 h), whereas parent V 2 O 5 -TiO 2 mixed oxides were partially transformed into rutile after a calcination at a temperature as low as 450 °C. 20 Based mainly on XRD data, it has been suggested that the * Corresponding author. Phone: (1) 44 27 60 04. Fax: (1) 44 27 60 33. E-mail: davidson@ccr.jussieu.fr. ² Laboratoire de Re ´activite de Surface. ‡ Laboratoire de Cristallochimie du Solide. §Institut fu ¨r Physikalische Chemie. | Institut Universitaire de France. 8931 J. Phys. Chem. B 2000, 104, 8931-8939 10.1021/jp000569m CCC: $19.00 © 2000 American Chemical Society Published on Web 08/30/2000