X-ray photoelectron spectroscopy, x-ray absorption spectroscopy, and x-ray diffraction characterization of CuO – TiO 2 – CeO 2 catalyst system M. S. P. Francisco, P. A. P. Nascente, V. R. Mastelaro, and A. O. Florentino Citation: Journal of Vacuum Science & Technology A 19, 1150 (2001); doi: 10.1116/1.1345911 View online: http://dx.doi.org/10.1116/1.1345911 View Table of Contents: http://scitation.aip.org/content/avs/journal/jvsta/19/4?ver=pdfcov Published by the AVS: Science & Technology of Materials, Interfaces, and Processing Redistribution subject to AVS license or copyright; see http://scitation.aip.org/termsconditions. Download to IP: 186.217.234.225 On: Tue, 14 Jan 2014 15:32:40 http://scitation.aip.org/content/avs/journal/jvsta?ver=pdfcov http://oasc12039.247realmedia.com/RealMedia/ads/click_lx.ads/test.int.aip.org/adtest/L23/1291604448/x01/AIP/Hiden_JVACovAd_1640x440Banner_12_10and12_17_2013/1640x440_-_23874-BANNER-AD-1640-x-440px_-_USA.jpg/7744715775302b784f4d774142526b39?x http://scitation.aip.org/search?value1=M.+S.+P.+Francisco&option1=author http://scitation.aip.org/search?value1=P.+A.+P.+Nascente&option1=author http://scitation.aip.org/search?value1=V.+R.+Mastelaro&option1=author http://scitation.aip.org/search?value1=A.+O.+Florentino&option1=author http://scitation.aip.org/content/avs/journal/jvsta?ver=pdfcov http://dx.doi.org/10.1116/1.1345911 http://scitation.aip.org/content/avs/journal/jvsta/19/4?ver=pdfcov http://scitation.aip.org/content/avs?ver=pdfcov Redist X-ray photoelectron spectroscopy, x-ray absorption spectroscopy, and x-ray diffraction characterization of CuO– TiO 2 – CeO2 catalyst system M. S. P. Francisco Departamento de Fı´sica e Ciência dos Materiais, Instituto de Fı´sica de Sa˜o Carlos, Universidade de Sa˜o Paulo, 13560-970 Sa˜o Carlos, SP, Brazil P. A. P. Nascentea) Centro de Caracterizac¸ão e Desenvolvimento de Materiais, Departamento de Engenharia de Materiais, Universidade Federal de Sa˜o Carlos, 13565-905 Sa˜o Carlos, SP, Brazil V. R. Mastelaro Departamento de Fı´sica e Ciência dos Materiais, Instituto de Fı´sica de Sa˜o Carlos, Universidade de Sa˜o Paulo, 13560-970 Sa˜o Carlos, SP, Brazil A. O. Florentino Departamento de Quı´mica, Instituto de Biocieˆncias, Universidade Estadual Paulista, 18618-000 Botucatu, SP, Brazil ~Received 13 September 2000; accepted 11 December 2000! X-ray photoelectron spectroscopy~XPS!, x-ray diffraction ~XRD!, and x-ray absorption spectroscopy~XAS! techniques have been applied to characterize the surface composition and structure of a series of CuO–TiO2–CeO2 catalysts. For a small loading of cerium, ceria was mainly dispersed on the titania surface and a minor amount of CeO2 crystallite appeared. At higher loading of cerium, the CeO2 phase increased and the atomic Ce/Ti ratio values were smaller than the nominal composition, as a consequence of cerium agglomeration. This result suggests that only a fraction of cerium can be spread on the titania surface. For titanium-based mixed oxide, we observed that cerium is found as Ce31 uniquely on the surface. The atomic Cu/~Ce1Ti! ratio values showed no influence from cerium concentration on the dispersion of copper, although the copper on the surface was shown to be dependent on the cerium species. For samples with a high amount of cerium, XPS analysis indicated the raise of second titanium species due cerium with spin-orbit components at higher binding energies than those presented by Ti41 in a tetragonal structure. The structural results obtained by XAS are consistent with those obtained by XRD and XPS. ©2001 American Vacuum Society.@DOI: 10.1116/1.1345911# ly th id i du er r de tiv he be a io g n iv ata- ort. ss tive ata- ed use s d to rs- of ied on- ia- ed ace fter . e tros- nd py ma I. INTRODUCTION Methanol oxidation to methyl formate is an industrial important reaction because formate is a raw material in production of formic acid, dimethyl formamide, acetic ac formamide, and cyanhydric acid. The oxidative catalyze the most used process, which permits us to obtain a pro with higher commercial value from molecules with low aggregate value.1,2 Among the active catalytic systems fo methanol dehydrogenation, alkaline metal oxide, zinc oxi and copper-containing catalysts are employed.1,3,4 Copper- supported catalysts show high catalytic activity and selec ity. A possible explanation for this behavior comes from t fact that metallic oxides, which change the oxidation num from one unit are good oxidation catalysts.4 Recently, copper supported on titanium oxides has demonstrated a high c lytic performance for steam reforming and dehydrogenat of methanol.5,6 According to the literature, there are some disadvanta when TiO2 is used as support for CuO, such as thermal a mechanical instability, sintering of the support, and act phase and titania phase transformation.1,7 a!Author to whom correspondence should be addressed; electronic nascente@power.ufscar.br 1150 J. Vac. Sci. Technol. A 19 „4…, Jul ÕAug 2001 0734-2101 Õ200 ribution subject to AVS license or copyright; see http://scitation.aip.org/term e , s ct , - r ta- n es d e The stabilization of the active phase and supported c lysts depends firmly on the characteristics of the supp Cerium oxide is known to improve resistance to thermal lo of the supported catalyst surface area, to stabilize the ac phase in a fine dispersed state, and also to improve the c lytic activity.7 The change in oxidation state is associat with the reversible removal and addition of oxygen beca of its low redox potential.8 This behavior probably explain the improvement described above when cerium is adde the catalytic system. In the alumina modified with ceria La son and Andersson9 evidenced an increase of combustion CO until CuO particles are formed. They have also stud ethanol oxidation in stationary applications on catalysts c sisting of copper oxide supported on titania and cer modified titania.7 Their results showed that ceria enhanc the activity of the copper species and stabilized the surf area of the titania support. We propose to investigate the catalysts structure a loading cerium and copper oxides to the TiO2 anatase phase X-ray diffraction ~XRD! was employed to determine th crystalline phases and grain sizes. X-ray absorption spec copy~XAS! allowed us to probe the short range order arou Ce, Cu, and Ti atoms. X-ray photoelectron spectrosco il: 11501Õ19„4…Õ1150Õ8Õ$18.00 ©2001 American Vacuum Society sconditions. Download to IP: 186.217.234.225 On: Tue, 14 Jan 2014 15:32:40 th th in a b re lc nd n- e id t o in fo he s in U ° s er t a l a a 72 de F ry io e s. th dg - - te ion ow ted a. ed g ak- nd ted t he m - - xed ana- uce ned ies sed n itiv- 1 re O, op- u- rved , d at he - ted her 1151 Francisco et al. : XPS, XAS, and XRD characterization 1151 Redist ~XPS! was used to investigate the chemical changes of elements present at the surface of the catalysts. II. EXPERIMENTAL PROCEDURES Titania-based mixed supports were prepared using sol–gel technique. Commercial~NH4!2Ce~NO3!6 powder ~Vetec! was dissolved in nitric acid aqueous solution~1.5 mol/l!. Then, Cu~NO3!2.3H2O powder ~Vetec! was added, heated until 353 K and kept at this temperature for 30 m we denoted the resulted solution A. A second solution B w prepared with tetraisopropyl orthotitanate~C12H28O4TiO, Merck! which was dissolved in isopropyl alcohol~mole rate51!. The solution A was added to solution B and su mitted to a 50 W ultrasonic vibration for 2 min. The mixtu rested for 24 h in a saturated atmosphere of isopropyl a hol. Finally, the resulting gel was dried at 383 K for 16 h a than calcined at 723 K for 16 h in air. Solution A was co trolled under acid conditions (pH 1 – 2) in order to guarante a better salt solubilization and a rapid first step alkox group protonation when added to solution B. The amoun water added to solution A was calculated in spite of be suitable for cerium salt solubilization and in excess alkoxide hydrolysis during its polymerization process. T prepared powder CuO–CeO2/TiO2 samples are denoted a Ce0.91O1.91Cu0.09, Ti0.91O1.91Cu0.09, Ce0.09Ti0.82O1.91Cu0.09, Ce0.27Ti0.64O1.91Cu0.09, and Ce0.41Ti0.41O1.82Cu0.18, and ex- press the amount of each component in atomic units. The x-ray powder diffraction patterns were obtained us an automatic Rigaku Rotaflex diffractometer model R 200B with CuKa radiation~40 kV/40 mA, 1.5405 Å! and a nickel monochromator filter. The scanning range was 20 60° ~2u! with a step size of 0.02° and a step time of 3.0 The average crystallite sizes were determined by the Sch equation. The broadening of half-height peak widths due slit sizes and x-ray source characteristics was calculated was equal to 0.16°~2u!. To determine the instrumenta broadening, a reference x-ray pattern from powder Si w recorded. The silicon powder with grain size around 5mm was used. The identification of crystalline phases was complished by comparison with JCPDS file Nos. 21-12 43-1002, and 41-254 for titania, ceria, and copper oxi respectively. The XAS measurements were carried out at the XA beamline of the Brazilian Synchrotron Light Laborato ~LNLS!. The experiments were performed in transmiss mode. The incident and transmitted photon intensities w determined by means of standard ion chamber detector double crystal Si~111! was used as a monochromator and energy steps were 0.8 eV for x-ray absorption near-e spectroscopy~XANES! and 2.0 eV for extended x-ray ab sorption fine spectroscopy~EXAFS! regions. The measure ments were performed by analyzing TiK edge, CeL III edge, and CuK edge at room temperature. The EXAFS analysis was carried out by a microcompu using a program set written by Michalowicz,10 according to the recommended procedures described by the Internat Workshop on Standards and Criteria in XAFS.11 After re- JVST A - Vacuum, Surfaces, and Films ribution subject to AVS license or copyright; see http://scitation.aip.org/term e e , s - o- e f g r g – . rer o nd s c- , , S n re A e e r al moval and normalization of atomic absorption, thek3x(k) weighted EXAFS oscillation was Fourier transformed toR distance space in the range of 3.2–12 Å21 for Cu and 2.3– 9.8 Å21 for Ce. In each case, the Kaiser apodization wind with t52.5 was used. The energy threshold was selec arbitrarily at the inflexion point of the absorption spectr The contribution of the first coordination shell was extract by a back Fourier transform inR space and then fitted usin experimental phase and amplitude functions. A quantitative analysis of the Ce–O shell was made t ing a CeO2 compound as reference~RCe–O52.34 Å and NCe–O58!. To model the Cu–O bond, the phase shift a backscattering amplitude for the Cu–O pair were extrac from a CuO compound settingN54 andR51.96 Å. In all the fittings, the number of freeNpar parameters was kep smaller than the number of independentNind points, defined asNind52DRDk/p, whereDR is the width of theR-space filter windows andDk is the actual interval of the fitting in the k space.11 For the fitting at the Cu and Ce edges, t errors were estimated to be approximately equal to60.01 Å in distance and65% in coordination numbers. The XPS analysis was performed in ultrahigh vacuu ~low 1027 Pa range! by using a KRATOS XSAM HS spec trometer, with a MgKa (hn51253.6 eV) x-ray source op erated at 15 kV and 15 mA. The powder samples were fi on a steel holder by a double side adhesive tape, and lyzed as received. An electron flood gun was used to red charge effects. The high-resolution spectra were obtai with analyzer pass energy of 20 eV. The binding energ were referenced to an adventitious carbon 1s line set at 284.8 eV. For curve fitting, Gaussian line shapes were u for C 1s, O 1s, and N 1s, and a mixed Gaussian/Lorentzia function for Cu 2p, Ti 2p, and Ce 3d. The Shirley back- ground and a least-square routine were applied. The sens ity factors for quantitative analysis were referenced to Fs 51.0. III. RESULTS AND DISCUSSION A. X-ray diffraction XRD patterns of the samples calcined at 723 K a shown in Fig. 1. Despite the larger amount of copper, Cu and Cu2O phases were not observed. Following the devel ment of XRD patterns, it is possible to notice a great infl ence of loading cerium on the Ti0.91O1.91Cu0.09 sample in which only a very defined anatase phase could be obse ~25.2°/2u!. After ceria addition to the Ti0.91O1.91Cu0.09 (Ce0.09Ti0.82O1.91Cu0.09 sample!, the diffractogram changed depicting weak broad peaks whose main one was centere the ~101! anatase position, and the major peak of t CeO2–cerianite phase was barely visible~28.2°–28.3°/2u!. After addition of 27% and 45% of cerium (Ce0.27Ti0.64O1.91Cu0.09 and Ce0.45Ti0.46O1.91Cu0.09 samples!, the peaks related to the CeO2 phase become more pro nounced and easily indexed, and only the main peak rela to the TiO2 phase could be observed~;25.2°/2u!. The observation that bulk CuO is segregated at hig copper contents in Cu–Ce~La!–O catalysts was done by Liu sconditions. Download to IP: 186.217.234.225 On: Tue, 14 Jan 2014 15:32:40 ox es r- ivi g, al ec ce g de th ys itio e u e pp u ll iz y m o ne o as t e ows all and as of e in he- ted f he ame pic om to iO f tita- g t e ing e 1152 Francisco et al. : XPS, XAS, and XRD characterization 1152 Redist and Flytzani-Stephanopoulos.12 They concluded that only a small amount of copper is necessary to promote the CO dation. The excess of copper forms bulk CuO particl which contributes little to the catalyst activity. Coppe containing phases were not detected in Cu/TiO2 catalysts with lower copper loading as a consequence of the sensit and size limits of the XRD technique.13 Larsson and Andersson7,9 have studied CuOx– CeO2–Al2O3 and CuOx– CeO2–TiO2 catalyst systems. For low copper loadin in the first system they identified Cu atoms as a copper minate surface phase, and in the second system they det no peak from the CuO phase. They observed the presen bulk CuO crystallites in both systems only for high loadin They concluded that the distribution of copper species pends not only on the copper oxide loading but also on ceria loading. Our XRD results revealed the presence of only two cr tallographic phases: TiO2 anatase and CeO2 cerianite. The amount of both phases depends on the support compos but it is independent of the copper content. The absenc Cu phases in our samples is in agreement with the res recently published.7,9 As the amount of copper increased, w can state that copper atoms are well dispersed on the su structure. The titania grain size was calculated by the Scherrer eq tion for each sample. We observed a trend toward sma TiO2 crystallites as cerium was added: the titania grain s presented by Ti0.91O1.91Cu0.09 was equal to 12.3 nm and b the sample with the highest amount of ceriu (Ce0.45Ti0.46O1.91Cu0.09 sample! 2.4 nm. Furthermore, due t the fact that any significant peak shift of the crystalli phases mentioned above was observed, the formation solid solution can be discarded. B. XAS measurements 1. Ti K-edge: XANES spectra The TiK-edge spectrum for titania at the anatase ph obtained as a reference is very similar to that presented in FIG. 1. XRD patterns:~a! Ti0.91O1.91Cu0.09, ~b! Ce0.09Ti0.82O1.91Cu0.09, ~c! Ce0.27Ti0.64O1.91Cu0.09, ~d! Ce0.41Ti0.41O1.82Cu0.18, calcined at 723 K. TiO2–anatase phase~* ! and CeO2–ceranite phase~j!. J. Vac. Sci. Technol. A, Vol. 19, No. 4, Jul ÕAug 2001 ribution subject to AVS license or copyright; see http://scitation.aip.org/term i- , ty u- ted of . - e - n, of lts ort a- er e f a e he literature.14,15 All catalysts seemed to have titanium at th anatase phase, as it was identified by XRD. Figure 2 sh the pre-edge region of the Ti absorption edge. For samples at least three pre-edge peaks were presented some of them had their intensities strongly modulated cerium concentration is increased: the intensities A3(4971.9 eV! andA1(4968.8 eV! peaks decreased and th A2 peak~4970.9 eV! became clear, appearing to dominate the Ce0.41Ti0.41O1.82Cu0.18 sample. The intensity of theB peak~4974.6 eV! did not change significantly. In a previous study of the TiK pre-edge, Chenet al.14,15 investigated titanium dioxide nanoparticles~19 Å! which had their TiK-edge spectra compared to the spectrum of octa dral Ti sites in bulk anatase~500 Å!. The XANES data re- vealed the rise of titanium in the surface in penta-coordina sites as a result of distortions of Ti sites in smaller TiO2 nanoparticles. Chenet al.14 distinguished the coexistence o the surface layer from bulk TiO2 by measuring XAS spectra of particles with many different sizes and observing t changes in the spectral features. As the particle size bec smaller, more Ti atoms were on the surface in an anisotro environment, which caused distortions around Ti atoms fr an octahedral TiO6 unit.14 Fargeset al.16 associated titanium with @4#Ti, @5#Ti, and @6#Ti coordination to the following peak positions: 4969.5, 4970.5, and 4971.5 eV. Our results indicated that due to the addition of cerium the support, the number of Ti atoms in an octahedral T2 structure decreased as a consequence of the decrease o nium particle size. Moreover, the intensity of theA3 peak, which is assigned to@6#Ti coordination decreased, reinforcin the previous statement. 2. Ce L III-edge results: XANES ÕEXAFS spectra Figure 3 shows the CeL III XANES spectra of the catalys samples together with CeO2 reference spectrum in which th cerium atom is tetravalent, surrounded by oxygen ions fill the tetrahedral interstitial sites. The two major peaks~de- noteda andb! of CeO2 spectrum are in agreement with th FIG. 2. Ti K-edge XANES spectra of:~a! TiO2–anatase phase,~b! Ti0.91O1.91Cu0.09, ~c! Ce0.09Ti0.82O1.91Cu0.09, ~d! Ce0.27Ti0.64O1.91Cu0.09, and ~e! Ce0.41Ti0.41O1.82Cu0.18. sconditions. Download to IP: 186.217.234.225 On: Tue, 14 Jan 2014 15:32:40 te rg fo he m s r he O F - a- a to o h a lys 6 le pr th ok . ig. the e ere ive ll. g a all . A the om- ller the sec- ted ta- his at- 1153 Francisco et al. : XPS, XAS, and XRD characterization 1153 Redist literature results.17 These two components were also detec in the catalyst samples with an intensity ratio and ene position close to that found in the CeO2 sample. Figure 4 presents a comparison of EXAFS spectra CeO2 and the catalysts. The EXAFS spectra of t Ce0.27Ti0.64O1.91Cu0.09 and Ce0.41Ti0.41O1.82Cu0.18 catalysts are comparable to the CeO2 spectrum, whereas the spectru of the Ce0.09Ti0.82O1.91Cu0.09 sample is rather different. Thi difference can also be observed in the Fourier transfo spectra displayed in Fig. 4. For the samples with a hig amount of cerium, the second peak~between 3 and 5 Å! was observed, although its intensity was lower than in the Ce2. This peak is related to Ce–Ce and Ce–O bonds~second neighbors!. The phase and amplitude used to simulate the EXA spectra were obtained from the CeO2 reference and the re sults are presented in Table I. Both Ce0.27Ti0.64O1.91Cu0.09 and Ce0.41Ti0.41O1.82Cu0.18 samples showed a first coordin tion shell similar to that of CeO2. A better simulation was obtained for the Ce0.09Ti0.82O1.91Cu0.09 sample when two dis- tances for the first coordination shell were considered, though the average number of first neighbors was equal and the average Ce–O distance was 2.32 Å. 3. Cu K-edge results: XANES ÕEXAFS spectra The CuK-edge XANES spectra were collected for tw copper patterns, CuO~Cu21 in a square planar symmetry! and Cu2O ~Cu11 coordinated by two oxygen atoms!, and for the catalysts. The XANES spectra are shown in Fig. 5. T CuO and Cu2O XANES spectra presented edge energies 8984.3 and 8980.5 eV, respectively, whereas the cata presented a transition positioned at approximately 898 eV. These transitions were ascribed to a dipole-enabled e tronic transition of 1s→4p.18 Moreover, theA peak~9016.0 eV! disappeared and theB peak~8997.9 eV! broadened when cerium was added to the support. Thus, the catalysts sented quite different XANES spectra when compared to CuO and Cu2O reference compounds. FIG. 3. CeL III -edge XANES spectra of:~a! CeO2 reference, ~b! Ce0.91O1.91Cu0.09, ~c! Ce0.27Ti0.64O1.91Cu0.09, ~d! Ce0.09Ti0.82O1.91Cu0.09, and ~e! Ce0.41Ti0.41O1.82Cu0.18. JVST A - Vacuum, Surfaces, and Films ribution subject to AVS license or copyright; see http://scitation.aip.org/term d y r m r S l- 8 e t ts .4 c- e- e The Cu-K edge EXAFS spectra presented in Fig. 6 lo very similar, but they differ from the CuO EXAFS spectrum The Fourier transform EXAFS spectra are presented in F 6. The position of the main peak, which corresponds to first Cu–O coordination shell, is similar to that of all th catalysts and CuO. Backscattering phase and amplitude w extracted from the CuO compound to provide quantitat structural information of the first Cu–O coordination she The simulation results are presented in Table I, showin coordination number approximately equal to 4.0 for samples and a Cu–O mean bond length at around 1.95 Å decrease in the intensity of the first peak was observed in Fourier transform spectra of the catalyst samples when c pared to the CuO sample as a result of the Debye–Wa factor increasing. Regarding more distant shells, it can be observed that Fourier transform spectra of all catalyst samples show a ond peak located between 2.2 and 3 Å, which is rela mainly to the Cu–Cu first neighbors. This peak in the ca lysts is less intense than that of the CuO compound. T difference is probably due to the presence of Ti and Ce oms in the structure of the catalyst samples. FIG. 4. ~Top! CeL III -edge EXAFS spectra of:~a! CeO2 phase, ~b! Ce0.09Ti0.82O1.91Cu0.09, ~c! Ce0.27Ti0.64O1.91Cu0.09, and ~d! Ce0.41Ti0.41O1.82Cu0.18. ~Bottom! Fourier transform of the CeL III -edge EXAFS spectra of: ~a! CeO2 phase, ~b! Ce0.09Ti0.82O1.91Cu0.09, ~c! Ce0.27Ti0.64O1.91Cu0.09, and~d! Ce0.41Ti0.41O1.82Cu0.18. sconditions. Download to IP: 186.217.234.225 On: Tue, 14 Jan 2014 15:32:40 1154 Francisco et al. : XPS, XAS, and XRD characterization 1154 J. Vac. Sci. Techno Redistribution subject to AV TABLE I. Simulation results for CeL III -edge and CuK-edge EXAFS spectra. Sample CeL III edge CuK edge N/Ce–O ~60.5! R ~Å! ~60.02! s ~Å! ~60.01! N/Cu–O ~60.4! R ~Å! ~60.01! s ~Å! ~60.01! CuO — — — 4.0 1.96 — Ti0.91O1.91Cu0.09 — — — 4.2 1.95 0.04 CeO2 8.0 2.34 — — — — Ce0.09Ti0.82O1.91Cu0.09 6.2 2.25 0.07 3.8 1.95 0.05 2.2 2.40 0.13 ¯ ¯ ¯ Ce0.27Ti0.64O1.91Cu0.09 7.9 2.32 0.06 — — — Ce0.41Ti0.41O1.91Cu0.18 7.8 2.31 0.06 4.1 1.96 0.06 n in a p d ts e or e l y - rb g r of be- r m- C. XPS analysis Information about changes in the oxidation number a structure from copper, cerium, and titanium due to load cerium were obtained by XPS analysis. Binding energy v ues of the main peaks in the XPS of references and sam are summarized in Table II. Figure 7 shows the Ti 2pXPS spectra of samples an TiO2–anatase. Luet al.19 reported the Ti 2p3/2 peaks at 458.8, 457.5, and 455.1 eV for TiO2, Ti2O3, and TiO, re- spectively. The Ti 2p doublet of TiO2–anatase presented i spin-orbit components at 459.0 eV(2p3/2) and 464.6 eV(2p1/2) with full width at half maximum and rela- tive intensity~2:1! in good agreement with results from th literature.19 We deconvoluted the 2p3/2 and 2p1/2 peaks of the Ti0.91O1.91Cu0.09 sample by single components each, c responding to Ti41 in a tetragonal structure. Otherwise, w fitted the 2p spectrum of the Ce0.09Ti0.82O1.91Cu0.09 sample by two curves at each peak. We assigned the Ti 2p3/2 peak at lower binding energy~459.2 eV! to titanium in a tetragona structure and the Ti 2p3/2 peak at higher binding energ ~460.8 eV! to species at the same oxidation number 41. Analyzing the intensities from 2p3/2 peaks of the two tita- nium species in the Ce0.09Ti0.82O1.91Cu0.09 sample, the tetrag onal structure is predominant on the surface. The spin-o components in the spectra of Ce0.27Ti0.64O1.91Cu0.09 and FIG. 5. CuK-edge XANES spectra of:~a! Cu2O, ~b! CuO, ~c! Ti0.91O1.91Cu0.09, ~d! Ce0.09Ti0.82O1.91Cu0.09, ~e! Ce0.27Ti0.64O1.91Cu0.09, and ~f! Ce0.41Ti0.41O1.82Cu0.18. l. A, Vol. 19, No. 4, Jul ÕAug 2001 S license or copyright; see http://scitation.aip.org/term d g l- les - it Ce0.27Ti0.64O1.91Cu0.09 samples were also better fitted takin into account the presence of two species. The numbe distorted species from tetragonal structure on the surface came higher with cerium addition. Similar results to ours were obtained fo IrO2– TiO2– CeO2 electrocatalysts.20 The Ti 2p spectra be- came larger with shifts to higher energy values for ceriu FIG. 6. ~Top! CuK-edge EXAFS spectra:~a! CuO, ~b! Ti0.91O1.91Cu0.09, ~c! Ce0.09Ti0.82O1.91Cu0.09, and ~d! Ce0.41Ti0.41O1.82Cu0.18; ~Bottom! Fourier transform of CuK-edge EXAFS spectra of:~a! CuO, ~b! Ti0.91O1.91Cu0.09, ~c! Ce0.09Ti0.82O1.91Cu0.09, and~d! Ce0.41Ti0.41O1.82Cu0.18. sconditions. Download to IP: 186.217.234.225 On: Tue, 14 Jan 2014 15:32:40 s and 1155 Francisco et al. : XPS, XAS, and XRD characterization 1155 JVST A - Vacuum, Redistribution subject to AV TABLE II. XPS binding energy values; XPS atomic ratios as a function of Ce content for the reference samples;u-/I Total and (u81v8)/I Total ratios were calculated from Ce 3d peaks of reference and samples. Sample B.E. eV Cu 2p3/2 Ti 2p3/2 Ce 3d5/2 O 1s OH2 Olattice CeO2 — — 882.7 531.7 529.4 TiO2 — 459.0 — 531.7 530.1 Ce0.91O1.91Cu0.09 933.7 — 882.3 531.4 529.4 Ti0.91O1.91Cu0.09 933.7 459.0 — 531.4 529.4 Ce0.09Ti0.82O1.91Cu0.09 933.6 459.2/460.8 883.2 532.1 530.5 Ce0.27Ti0.64O1.91Cu0.09 933.7 429.3/460.6 882.6 531.5 530.0 Ce0.45Ti0.46O1.91Cu0.09 933.8 458.7/460.5 882.2 531.6 529.9 CuO 933.9 — — 531.8 529.9 Cu2O 932.1 — — 531.9 530.2 Atomic ratio Ce/Ti Cu/(Ti1Ce) Cu/Ce Cu/Ti u-/I Total (u81v8)/I Total CeO2 0.11 0.14 Ce0.91O1.91Cu0.09 — 0.17 0.17 — ¯ ¯ Ti0.91O1.91Cu0.09 — 0.14 — 0.14 ¯ ¯ Ce0.09Ti0.82O1.91Cu0.09 0.10 0.17 1.91 0.19 0.07 0.27 Ce0.27Ti0.64O1.91Cu0.09 0.33 0.15 0.60 0.20 0.08 0.20 Ce0.45Ti0.46O1.91Cu0.09 0.64 0.15 0.41 0.26 0.09 0.22 o ia th a ed te s en e ly e - nd a nin be he at- k ne ne. - sts per dis- . the the s a II. ent 1%, ms is in- rium O e he er, en- eria rich samples, suggesting that this element is present in m than one species. Strong interaction between TiO2 and SiO2 in mixed oxides was reported to present higher Ti 2p3/2 bind- ing energy than the value for pure TiO2. 21 This upward shift was explained by the increase in the interatomic potent due to the decrease of the coordination number of Ti and shortening of the Ti–O bond. The XPS result indicating th the presence of Ti41 in the tetragonal structure decreas with cerium addition on the surface titanium is corrobora with our TiK-edge XANES and grain size analyses. Figure 7 shows the Ce 3dXP spectra, where six peak corresponding to three pairs of spin-orbit doublets were id tified and ascribed to Ce~IV! and the other two peaks wer ascribed to Ce~III !. The spin-orbit doublets are common denotedu and v and represent the 3d5/2 and 3d3/2 states, respectively, in the range 880–920 eV.22 Thev andv9 peaks have been attributed to a mix of 3d94 f 2(O 2p4) and 3d94 f 1(O 2p5)Ce(IV) final states, and v- peak to 3d94 f 0(O 2p6)Ce(IV) final state.7 The v8 peak corresponds to the 3d94 f 1(O 2p6)Ce(III ) final state.7 The series ofu structures from the 3d3/2 level can be explained in the sam way. The ratio between theu- intensity and the total inten sity of Ce 3d lines depends on the Ce41/Ce31 ratio. To- gether withu- peak behavior, Ce41 reduction causes thev8 andu8 increase.7,22 Those ratios calculated for reference a samples are displayed in Table II, where the trends tow smalleru-/I Total ratio and higher (u81v8)/I Total ratio with cerium addition may be observed. Then, samples contai cerium seem to have part of this element as Ce31. Although the possibility of Ce31 presence in the catalysts cannot ruled out by XAS analysis, the existence of Ce31 must be limited only to the surface. The Cu 2pXPS results are presented in Table II. T Surfaces, and Films S license or copyright; see http://scitation.aip.org/term re ls e t d - rd g Cu 2pXP spectrum of CuO was easily identified by the s ellite at about 940–945 eV and the 2p3/2 main peak at 934.2 eV.1 The Cu 2pXP spectra for both Ti0.91O1.91Cu0.09 and Ce0.91O1.91Cu0.09 references were fitted by only one pea centered at 933.7 eV. The satellite for the former o ~;944.4 eV! appeared more prominent than for the last o For the catalyst spectra, the Cu 2p3/2 components were posi tioned very close to each other at 933.8 eV. The cataly presented a weak satellite, probably due to their low cop concentration. All catalysts had their copper species as persed Cu21 since no copper peak was identified by XRD XP spectra were also obtained in order to analyze surface composition changes caused by loading cerium to support. The surface ratios and nominal composition a function of Ce content of samples are presented in Table The atomic Ce/Ti ratio values increased with cerium cont on the support. The Ce0.09Ti0.82O1.91Cu0.09, Ce0.27Ti0.64O1.91Cu0.09, and Ce0.45Ti0.46O1.91Cu0.09 samples presented their respective atomic Ce/Ti ratios as 9%, 2 and 36% lower than their nominal compositions. It see that for low cerium concentration most of this element simply dispersed on the matrix. As cerium concentration creases, the dispersed species originated the second ce species. The XRD data indicated an evolution of the Ce2 phase: the Ce0.09Ti0.82O1.91Cu0.09 sample presented only th major diffraction peak related to this phase, while t Ce0.27Ti0.64O1.91Cu0.09 and Ce0.45Ti0.46O1.91Cu0.09 samples showed a very defined ceria diffraction pattern. Moreov XAS data showed that cerium in the Ce0.09Ti0.82O1.91Cu0.09 sample had its Ce–O distance distorted from the CeO2 ref- erence and from other samples with higher cerium conc tration. So, our results revealed that a small amount of c sconditions. Download to IP: 186.217.234.225 On: Tue, 14 Jan 2014 15:32:40 . of i po r l- qu t . ed o tio tiv c A nd sion rium on- port PS all xed even the due ted in- the area igher e- S is on- Pq at tal. in r- . D. . B J. 1156 Francisco et al. : XPS, XAS, and XRD characterization 1156 Redist could be presented in a dispersed form on the support~invis- ible by XRD!. An excess of CeO2 caused its crystallization Other investigators7 have reported that a limited amount cerium can be confined to the titania in order to stabilize and the excess is segregated on the CeO2 phase. Comparing the atomic Cu/Ce ratio of the Ce0.91O1.91Cu0.09 sample with the Ti/Ce atomic ratio of the Ti0.91O1.91Cu0.09 sample, the former was 41% higher than its nominal com sition, while the later was only 29%. Thus, copper is mo exposed in pure ceria than in titania matrix. From the atom Cu/~Ti1Ce! ratio value of the samples with cerium, we ca culated an average raise from the nominal composition e to 41%. This result suggests that copper dispersion on mixed support is independent from cerium concentration The atomic Cu/Ce ratio values from all ceria-modifi titania samples presented the same raise from nominal c position ~an average value equal to 48%!, while the Cu/Ti values suffered influence from cerium addition. These ra decreased with cerium loading compared to their respec nominal composition. A Ti-surface enrichment probably o curred due to titania size decrease as the results from X FIG. 7. ~Top! Ti 2p XP spectra of: ~a! Ce0.45Ti0.46O1.91Cu0.09, ~b! Ce0.27Ti0.64O1.91Cu0.09, ~c! Ce0.09Ti0.82O1.91Cu0.09, ~d! Ti0.91O1.91Cu0.09, and ~e! TiO2 . ~Bottom! Ce 3d XP spectra of:~a! Ce0.45Ti0.46O1.91Cu0.09, ~b! Ce0.27Ti0.64O1.91Cu0.09, ~c! Ce0.09Ti0.82O1.91Cu0.09, ~d! Ce0.91O1.91Cu0.09, and ~e! CeO2. J. Vac. Sci. Technol. A, Vol. 19, No. 4, Jul ÕAug 2001 ribution subject to AVS license or copyright; see http://scitation.aip.org/term t, - e ic al he m- s e - S and XRD indicated, in agreement with the titanium seco species observed by XPS analysis. The copper disper must be more connected to ceria since we observed a ce segregation with cerium addition, and a higher copper c tent on the ceria surface than on the titania support. IV. CONCLUSIONS The amount of observed phases depends on the sup composition, but it is independent of copper content. X and XAS analyses suggest that CuO is in the form of sm crystallized grains well dispersed on the surface of the mi support structure, as we observed no CuO bulk phases, for high copper content. According to this statement, all samples presented a distorted local order of copper shell to the presence of Ti and Ce. XPS surface analysis indica that copper dispersion on the mixed support suffered no fluence of cerium concentration, but it is dependent on cerium species. The trend toward smaller TiO2 crystallites with cerium addition was evidenced by XRD measurements, surface decrease, changes in TiK post-edge region in the XANES spectra, and the appearance of a second species with h Ti 2p3/2 binding energy in the XP spectra. This trend is r lated to a decrease of the Ti coordination number. A higher cerium local distortion was observed in XA analysis for low cerium concentration, implying that CeO2 is mainly in a dispersed form on the titanium surface. Th analysis did not evidence cerium distortion for samples c taining a higher amount of CeO2. ACKNOWLEDGMENTS This work has been supported from FAPESP and CN Brazilian agencies. This research was partially performed LNLS, Brazil. 1A. Guerrero-Ruiz, I. Rodriguez-Ramos, and J. L. G. Fierro, Appl. Ca 72, 119 ~1991!. 2J. M. Tatiboue¨t, Appl. Catal., A148, 213 ~1997!. 3L. Domokos, T. Katonaand, and A´ . Molnàr, Catal. Lett. 40, 215 ~1996!. 4G. C. Bond,Catalysis by Metal, 1st ed.~Academic, London, 1962!. 5N. Takezawa and N. Iwasa, Catal. Today36, 45 ~1997!. 6J. P. Breen and J. R. H. Ross, Catal. Today51, 521 ~1999!. 7P.-O. Larsson and A. Andersson, J. Catal.179, 72 ~1998!. 8H. Inaba and H. Tagawa, Solid State Ionics83, 1 ~1996!. 9P.-O. Larsson and A. Andersson, Appl. Catal. B24, 175 ~2000!. 10A. Michalowicz, in Logiciels pour la Chimie~Société Française de Chimie, Paris, 1991!, p. 102. 11Report on the International Workshops on Standards and Criteria XAFS, in X-Ray Absorption Fine Structure: Proceedings of the VI Inte national Conference on X-Ray Absorption Fine Structures, edited by S. S. Hasnain~Ellis Horwood, New York, 1991!, p. 752. 12W. Liu and M. Flytzani-Stephanopoulos, J. Catal.153, 304 ~1995!. 13G. Balducci, J. Kaspar, P. Fornasiero, M. Graziani, M. S. Islam, and J Gale, J. Phys. Chem. B101, 1750~1997!. 14L. X. Chen, T. Rajh, Z. Wang, and M. C. Thurnauer, J. Phys. Chem 101, 10688~1997!. 15L. X. Chen, T. Raij, W. Jager, J. Nedelijkovic, and M. C. Thurnauer, Synchrotron Radiat.6, 445 ~1999!. sconditions. Download to IP: 186.217.234.225 On: Tue, 14 Jan 2014 15:32:40 k, S c., 1157 Francisco et al. : XPS, XAS, and XRD characterization 1157 Redist 16F. Farges, J. Non-Cryst. Solids244, 25 ~1999!. 17P. Ghigna, G. Spinolo, M. Scavini, A. Tamburini, and A. V. Chadwic Physica C253, 147 ~1995!. 18Y. Okamoto, T. Kubota, H. Gotoh, Y. Ohto, H. Aritani, T. Tanaka, and Yoshida, J. Chem. Soc., Faraday Trans.94, 3743~1998!. 19L. Lu, S. L. Bernasek, and J. Schwartz, Surf. Sci.458, 80 ~2000!. JVST A - Vacuum, Surfaces, and Films ribution subject to AVS license or copyright; see http://scitation.aip.org/term . 20V. A. Alves, L. A. Silva, S. C. Castro, and F. C. Boodts, J. Chem. So Faraday Trans.94, 711 ~1998!. 21A. Y. Stakheev, E. S. Shapiro, and J. Apijok, J. Phys. Chem.97, 5668 ~1993!. 22D. R. Mullins, S. H. Overbury, and D. R. Huntley, Surf. Sci.409, 307 ~1998!. sconditions. Download to IP: 186.217.234.225 On: Tue, 14 Jan 2014 15:32:40