Structure of Ultrathin CeO2 Films on Pt(111) by Polarization

Article pubs.acs.org/JPCC

Structure of Ultrathin CeO2 Films on Pt(111) by PolarizationDependent X‑ray Absorption Fine Structure P. Luches,*,† F. Pagliuca,†,‡ S. Valeri,‡ and F. Boscherini§,∥ †

Centro S3, Istituto Nanoscienze, Consiglio Nazionale delle Ricerche, Via G. Campi 213/a, I-41125 Modena, Italy Dipartimento di Scienze Fisiche, Informatiche e Matematiche, Via G. Campi 213/a, I-41125 Modena, Italy § Dipartimento di Fisica e Astronomia, Università di Bologna, Viale C. Berti-Pichat 6/2, I-40127 Bologna, Italy ∥ Istituto Officina dei Materiali, Consiglio Nazionale delle Ricerche, Operative Group in Grenoble, c/o ESRF, BP 220, F-38043 Grenoble, France ‡

ABSTRACT: We present a study of the structure of cerium oxide ultrathin films supported on Pt(111), focused on the evolution of the epitaxial strain in films of different thickness. The stoichiometry and oxidation state of the films are determined by X-ray photoemission spectroscopy, and the surface structure, measured by low-energy electron diffraction, is compared with the results obtained by the analysis of X-ray absorption fine structure measurements at the Ce L3 edge, exploiting the polarization dependence of the cross section to probe the in-plane and the out-of-plane atomic correlations. The obtained results allow to establish the epitaxial relation between the cerium oxide film and the Pt substrate and give an accurate evaluation of the cerium oxide layer structure. The 2 ML films have a fluorite structure which is compressed in the (111) plane. The measured compression is compatible with the assumption of a coincidence lattice between overlayer and substrate, in which three CeO2 surface unit cells match four Pt unit cells. The films’ three-dimensional structure is compared with the one expected assuming the bulk-phase elastic constants. The strain is released when the film thickness is increased to 10 ML, and the lattice parameters assume the bulk values.

■

systems, i.e., epitaxial cerium oxide ultrathin films or nanostructures on single crystal metal surfaces, can be of help in elucidating some crucial aspects related to the structure of nanosized materials and their interplay with metals. Several previous works, devoted to the study of morphology and electronic properties of ultrathin cerium oxide films supported on single crystal metal surfaces, showed that it is possible to obtain good quality flat films and supported lowdimensional structures.4,10−19 Cerium dioxide has the fluorite structure, and the (111) surface is the most stable one. In spite of the considerable lattice mismatch with most transition and noble metal (111) surfaces (30−40%), cerium oxide films have been found to grow epitaxially on most of these substrates,

INTRODUCTION Cerium oxide is the subject of much recent research, since it finds application in several different fields including catalysis and energy.1,2 The most remarkable property of this material, on which most of the applications are based, is its ability to store, release, and transport oxygen ions. Thanks to this property, cerium oxide is used in several catalytic applications which often involve also the use of metals either as supported nanoparticles, to get the desired catalytic activity, or as dopants, to increase its oxygen conductivity.3,4 The properties of cerium oxide in the nanostructured formbasically the reducibility have been shown to be optimized compared to the bulk phase.5,6 Several works report a lattice expansion as the nanoparticles size is decreased, which is ascribed to a partial reduction of the cations with remarkable consequences on the properties of the oxide.7−9 The proximity of metal atoms may lead to further structural rearrangements in the material and as a consequence to a different behavior. Studies of model © XXXX American Chemical Society

Received: October 19, 2012 Revised: December 18, 2012

A

dx.doi.org/10.1021/jp310375t | J. Phys. Chem. C XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry C

Article

LEED showed a sharp (1 × 1) pattern. Cerium oxide ultrathin films were grown by reactive evaporation of Ce, using an ebeam evaporator with a rate of ∼0.2 Å/min, measured by a quartz microbalance. The amount of deposited cerium oxide is expressed in nominal monolayers, where one monolayer (ML) is defined as the thickness of an O−Ce−O trilayer in the (111)oriented fluorite structure of cerium dioxide, i.e., 3.12 Å. During the evaporation molecular oxygen was supplied through a nozzle in a background O2 pressure of 1 × 10−7 Torr, and the Pt substrate was kept at room temperature. After the growth the cerium oxide films were annealed at 1040 K in 1 × 10−7 Torr O2 pressure for 15 min, a procedure which has been shown to optimize the film structure, morphology, and stoichiometry.14,32 Two cerium oxide ultrathin films of 2 and 10 ML nominal thickness are investigated in the present study. The films were checked in stoichiometry, thickness, and structure by in-situ XPS and LEED after the preparation. The XPS measurements were performed at normal emission using Al Kα photons. The Ce 3d spectra were fit using the methods described in our previous works.14,32 The samples were kept in a nonreactive atmosphere during their transfer to the synchrotron radiation laboratory. In order to check the stability of the samples, we exposed similar films to air and to N2 atmosphere for times comparable to the ones necessary for their transfer to the synchrotron. After this procedure XPS showed a very small reduction of the sample surface (probably due to surface hydroxylation), and the LEED pattern was very similar (apart from a slightly increased background) before and after the treatment. The quality of the sample surface was also checked after the samples were carried back from the synchrotron, finding similar results. We consider these as proofs that the samples were not relevantly modified before the XAFS measurements. The XAFS measurements at the Ce L3 absorption near-edge (X-ray absorption near-edge spectroscopy XANES) and extended energy region (extended X-ray absorption fine structure, EXAFS) were performed at the GILDA beamline33 of the European Synchrotron Radiation Facility. We exploited the polarization dependence of the XAFS cross section by changing the relative orientation between the sample normal and the polarization of the impinging X-ray beam in order to preferentially probe either the in-plane or the out-of-plane atomic correlations. In particular, we used two experimental configurations: the spectra indicated as PER are measured with the sample in vertical position, with the photon beam at 20° from the sample surface and the electric vector at 20° from sample normal; the spectra indicated as PAR are measured with the sample in horizontal position, with the photon beam at a few degrees from the sample surface and the electric vector in the surface plane.34 The spectra from the films were acquired in the fluorescence yield mode. Besides the two cerium oxide films, a reference sample made of CeO2 powder, i.e., with Ce mostly in the 4+ oxidation state, and a cerium silicate film with Ce in the 3+ oxidation state35 were also measured for comparison. The CeO2 powdered sample was measured in the transmission mode, while the silicate film was measured in the fluorescence yield mode in the PER configuration. Because of the presence of an unavoidable, though small, Cr−K adsorption edge signalprobably originating from diffused light scattering from the chamber wallswe had to limit the measured spectra to 5980 eV photon energy.

although the structural details of the interface and the driving force for the obtained epitaxial relation are still unclear. In a recent study we have shown that cerium oxide films with flat terraces and a good epitaxial quality can be obtained on Pt(111), and we detected an in-plane compressive strain by low-energy electron diffraction (LEED), which is released with increasing film thickness.14 Also, Grinter et al. measured a slight contraction in the surface lattice parameter of ultrathin cerium oxide films on Pt(111),13 but the mechanism responsible for this compression has not been identified. In a recent theoretical work Spiel et al. studied the structure and the electronic properties of the interface between a single cerium oxide layer and the Pt(111) surface, finding that the most stable structure is obtained when three cerium oxide surface unit cells match four substrate cells. The study also gives interesting insight into the different adsorption geometries and the respective density of states.20 A detailed atomic scale description of the CeO2/ Pt(111) interface structure is interesting also in view of understanding the experimentally observed higher reactivity toward CO oxidation observed for the CeOx/Pt(111) system compared to the clean Pt(111) surface.21,22 The epitaxy of cerium oxides on other metal substrates has also been studied. Mašek et al. using reflection high-energy electron diffraction measured an average in-plane 5% compression of the lattice constant at the early stages of the growth of cerium oxide on Cu(111).23 However, for this system the interface has been shown to be quite complex, with the reduction of a full cerium oxide layer and the formation of oxygen vacancies at the interface.24 Scanning tunneling microscopy studies on the same system revealed a strong dependence of the in-plane structure on the growth conditions and on the oxide island orientation.25 An in-plane contraction of the surface lattice parameter can also be deduced from the moiré pattern of cerium oxide films on Ru(0001).12 On the contrary on Rh(111) a slight in-plane expansion has been observed.11 It has to be mentioned also that submonolayer cerium oxide films grown by oxidation of a CePt alloy show an in-plane expansion of the lattice parameter.26 Polarization-dependent X-ray absorption fine structure (XAFS) is an excellent tool to investigate the three-dimensional local atomic structure of ultrathin films.27 It has been successfully used to obtain detailed information about epitaxial metal/oxide interfaces.28−31 The aim of this work is to obtain an accurate picture of the local structure at the interface between cerium oxide ultrathin films and the Pt(111) surface and to discuss the physical mechanisms for the observed epitaxial relation in this system. The results provide a structural basis for the understanding of the physical and chemical properties of the CeO2/Pt(111) system.

■

EXPERIMENTAL SECTION The samples for the present experiment have been grown in a UHV apparatus using two experimental chambers connected in ultrahigh vacuum (UHV): one allows for the growth of the cerium oxide layers, and the second one is equipped with facilities for characterization by Auger electron spectroscopy (AES), X-ray photoelectron spectroscopy (XPS) and LEED. The base pressure in the UHV apparatus is 1 × 10−10 Torr. The Pt(111) substrate was cleaned by repeated cycles of sputtering (1 keV, 1 μA) and annealing (1040 K). After the cleaning procedure the concentration of impurities on the Pt surface was below the XPS and AES detection limits and the B



Article

Initial data processing and background subtraction were performed using ATHENA software.28 Quantitative analysis was performed by nonlinear fitting of the EXAFS signals with the ARTEMIS package36 using theoretical scattering functions simulated using FEFF 8.2.37 Because of the relatively limited energy range available, analysis was limited to the first coordination shell of Ce, consisting of eight oxygen atoms. For all samples, the k and R ranges used for fitting (k2 weight) were 2−8 Å−1 and 1.25−2.35 Å, respectively; to avoid systematic errors also the bulk data were analyzed in this small k range. Many body effects influence the EXAFS spectrum. In fact, the presence of multiple excitations at the L edges of rare earths has been reported,38 and Fonda et al. have carefully taken into account the mixed valence nature of the excited state and the presence of multiple excitations in the analysis of high-quality EXAFS data of polycrystalline CeO2.39 In view of the restricted energy range in our analysis, we neglected the presence of weak multiple excitation features, the amplitude of which is very low.

The right panel of Figure 1 reports the near-edge regions of the Ce L3-edge XAFS spectra for the two films in the PER configuration and for the two reference samples containing Ce4+ and Ce3+, normalized to the edge jump. The two main peaks observed in the XANES of the CeO2 reference sample are also observed in the spectra of the films, and they are a further indication that the films contain Ce formally in the 4+ oxidation state. The many body transitions which give rise to the observed peaks are quite complex.42,43 The XANES spectra of Ce ions in the 3+ oxidation state are characterized by a single peak at ∼5 eV lower photon energy compared to the Ce4+ edge (Figure 1). The LEED patterns of the 2 and 10 ML films, measured using electrons of 80 eV kinetic energy, are shown in Figure 2.

■

RESULTS The XPS spectra of the 2 and 10 ML CeO2 films are reported in the left panel of Figure 1. The spectra have been normalized

Figure 2. LEED patterns of (a) 2 ML and (b) 10 ML cerium oxide films on Pt(111). The electron energy is 80 eV. The Pt- and CeO2related spots are indicated.

The 2 ML film (Figure 2a) shows both the pattern of the Pt substrate (outermost hexagonal pattern) and of the cerium oxide film (inner hexagonal pattern). The Pt spots are visible because at this coverage and in the used preparation conditions a fraction of the substrate surface is left uncovered, as shown by scanning tunneling microscopy measurements published in a previous work by the authors.14 The symmetry of the LEED patterns of the substrate surface and of the film indicate that the cerium oxide films grow with the (111) surface orientation and with the surface symmetry azimuths aligned to the ones of the Pt(111) substrate. The film diffraction spots are quite sharp, indicating a good long-range order. From the difference in the reciprocal lattice vectors corresponding to Pt and to the film, averaged in the three directions, we estimated for the 2 ML film a surface lattice parameter of 3.79 ± 0.03 Å. This value is slightly contracted compared to the bulk value of 3.826 Å. On the LEED pattern of the 10 ML film the substrate spots are not visible anymore because the surface is completely covered by the film in this case. The spots originating from the 10 ML film, though slightly broader than in the 2 ML film pattern, are still quite sharp and indicate a good long-range order also in this case. The surface lattice parameter of the 10 ML film has been evaluated to be 3.82 ± 0.05 Å, expanded with respect to the 2 ML one and comparable to the bulk value. Figure 3 shows the Fourier transforms of the k2-weighted χ(k) and the relative first-shell fits in the PAR and PER configurations, together with the raw χ(k) of the 10 ML sample illustrating a good signal-to-noise ratio. First of all, the bulk CeO2 data were analyzed in order to check the reliability of the theoretical scattering functions and to estimate the many body amplitude reduction factor S02. For bulk data four fitting

Figure 1. Left: Ce 3d XPS spectra of the 2 and 10 ML cerium oxide films (dots) and corresponding fits (solid lines). The spectra have been normalized in intensity for comparison. On the bottom the components and background used to fit the 2 ML spectrum are also shown. Right: Ce L3 edge XANES for the 2 and 10 ML cerium oxide films, for a CeO2 powdered sample (Ce4+ reference), and for a cerium silicate sample (Ce3+ reference).

in intensity to allow for a better comparison of their shape. The Ce3+- and Ce4+-related components and the background used to fit the 2 ML spectrum are also shown. Two doublets have been used for the Ce3+ and three (plus one to account for asymmetry) for Ce4+. They are due to different 4f configuration in the final state. For the details of their assignment and of the fitting procedure, we refer to previous works.40,41 The fitting of the spectra resulted in Ce4+ relative intensity of 91% for the 2 ML and 98% for the 10 ML sample within the XPS probing depth, in agreement with previous results on similar films.14 The lower Ce4+ relative intensity in the 2 ML sample is possibly due to the higher concentration of low-coordinated edge sites, expected to be more easily reducible, or to a larger weight of interface sites where some charge transfer from the oxide to the Pt substrate has been predicted by DFT studies.20 C



Article

good quality long-range order can be grown on the Pt(111) surface: the films have a (111) surface orientation and the surface symmetry azimuths of the films and of the substrate are aligned, as shown by the LEED patterns in Figure 2 and observed in previous works.13−15,17,18 The coincidence lattice between the Pt and CeO2 (111) surface unit cells (aPt = 2.775 Å, aCeO2 = 3.826 Å) which requires the lowest strain is obtained by matching five cerium oxide unit cells and seven Pt unit cells. In this case a 1.5% in plane expansion of the surface lattice parameter is expected. A second coincidence lattice with a low strain can be obtained if three cerium oxide surface unit cells match four Pt surface unit cells. This hypothesis implies a contraction of the cerium oxide lattice of 3.3%. Although the strain in the 5aCeO2 = 7aPt coincidence hypothesis is half than in the 3aCeO2 = 4aPt case, the latter configuration can be favored, since it requires a lower number of surface cell repetitions with a reasonable degree of strain. The contraction of the surface lattice parameter of the cerium oxide films measured by LEED on the 2 ML film may therefore originate from the 3aCeO2 = 4aPt coincidence relation with the substrate. However, the large relative errors connected to the LEED measurement and its sensitivity only to the in-plane lattice parameter hinder further considerations on this point. The presence of patterns reflecting the interface periodicity, e.g., (4 × 4) or (7 × 7) superstructures, are not expected to be observed on the 2 ML sample, since the layer actually forms islands, with a height of 2−3 ML covering a fraction of the substrate surface,14 and the LEED technique is very surface sensitive. The comparison with the EXAFS-measured first-neighbor distances in the two experimental configurations used allows to have more information on the strain of the film and compare the three-dimensional local structure measured with the one expected assuming that the cerium oxide lattice contracts to match the Pt lattice in the 3aCeO2 = 4aPt coincidence hypothesis. Whenever a film is strained in the surface plane, its perpendicular lattice constant varies according to elastic theory. Although when ultrathin films are concerned the materials do not always respond with the bulk-phase elastic constants, an analysis based on the behavior expected for a bulk may help in elucidating the details of the structure of ultrathin films. Since cerium oxide films on Pt expose their (111) surface, we are in the case of a trigonal distortion, i.e., an isotropic distortion in the (111) plane followed by a modification of the lattice parameter in the [111] direction, for which the following relation holds in the case of small strains:44,45

Figure 3. Modulus of the Fourier transform of the k2-weighted Ce L3 edge χ(k) (solid lines) and first-shell fits (dashed lines) for the CeO2 standard sample and for the 10 and 2 ML CeO2 films in the PAR and PER geometry. The inset shows the Ce L3 edge raw χ(k) data for the 10 ML film.

parameters were used: S02, an energy origin shift, a shift of the interatomic distance with respect to the theoretical simulation, and a Debye−Waller factor. A good fitting of the bulk CeO2 data was obtained (top curves in Figure 3) with S02 = 0.82. The data for the two samples in the two orientations were analyzed in a similar fashion, the only difference being that only three fitting parameters were used, S02 being fixed. The best estimates of the interatomic distances and Debye−Waller factors obtained from fitting are reported in Table 1. Since our Table 1. Parameters Resulting from the Fitting of the XAFS Spectra of the 2 and 10 ML Cerium Oxide Filmsa sample

R-factor (%)

CeO2 bulk; XRD reference 10 ML PER 10 ML PAR 2 ML PER 2 ML PAR

0.052 0.055 0.036 0.01

dCe−O (Å) 2.343 2.354 2.342 2.312 2.291

(0.018) (0.017) (0.015) (0.024)

σ2 (10−3 Å2) 13.0 7.6 11.8 9.6

(1.5) (1.4) (1.2) (2.0)

a R-factor, Ce−O first-neighbor distances (dCe−O), and Debye−Waller factor (σ2). The table also reports the Ce−O first-neighbor distance in the bulk phase (from ref 50).

interest is to study the variations of the interatomic distances with thickness and orientation and since the data relative to the samples intrinsically do not have the quality obtainable in the bulk, we have not considered the presence of a superposition of signals due to the mixed valence nature of the excited state of CeO2 as performed by Fonda et al.;39 an attempt to perform this analysis on the sample did not, in fact, give reliable results. The good quality of the fits of the spectra measured on the films indicates that also in the ultrathin limit of a few atomic layers the local structure of cerium oxide is the fluorite one. The fits of the 10 ML spectrum in the PAR and PER configurations show that the first-neighbor distance is almost relaxed to the bulk value (Table 1). The values for the 2 ML film, instead, are contracted both in the PAR and in the PER experimental configurations (Table 1).

⎛ a′ ⎞−γ c′ = c ⎜ ⎟ ⎝a⎠

(1)

where c and c′ are the lattice constants before and after the strain along the [111] direction, a and a′ are the lattice constants before and after the strain in the (111) plane, and γ is given by γ = (C11 + 2(C12 − C44))/(C11 + 2(C12 + C44))

(2)

where C11, C12, and C44 are the elastic constants of the material. The elastic constants for bulk cerium oxide are C11 = 4.03 × 1012 dyn/cm2, C12 = 1.05 × 1012 dyn/cm2, and C44 = 0.60 × 1012 dyn/cm2.46 Using eq 2 yields γ = 0.67 for cerium oxide. Assuming that the cerium oxide film adopts the 3aCeO2 = 4aPt coincidence lattice with the Pt(111) substrate, its in-plane lattice constant after the strain a′ is 3.3% contracted with

■

DISCUSSION Cerium dioxide and Pt have a quite large lattice mismatch of 38%. In spite of this, single crystalline cerium oxide films with a D



Article

respect to its bulk value, i.e., a′ = 0.967a = 3.700 Å, and using (1) with c = 3.124 Å (distance between cerium planes), c′ = 3.195 Å is obtained. We now compare these values, obtained under the hypothesis of 3aCeO2 = 4aPt coincidence and bulk elastic constants, with the Ce−O distances measured by EXAFS in the two measurement geometries. The Ce atoms in the CeO2 fluorite structure are in a cubic coordination with eight O nearest neighbors (Figure 4). Considering the (111) surface

O(1) and Ce−O(2) bond lengths, with weights proportional to the product of the number of O atoms and cos2 θ, θ being the angle between the interatomic vector and the polarization direction. If the electric vector were perfectly aligned to the Ce−O(1) bond, it can easily be calculated that the weights of the Ce−O(1) and Ce−O(2) contributions would be equal. With a 20° inclination of the electric vector from the (111) direction the relative weights of the six Ce−O(2) bonds will not be equal and will depend, in principle, also on the azimuthal angle in the (111) plane; however, considering an isotropic orientation of the local coordination environment around Ce, it can be estimated that the weights of the Ce−O(1) and Ce− O(2) contributions are 58% and 42%, respectively. Again in the hypothesis of a 3aCeO2 = 4aPt coincidence lattice, the two Ce−O(1) interatomic distances are expanded d′Ce−O(1) = 3/4c′ = 2.396 Å. Since the difference between the two distances d′Ce−O(1) and d′Ce−O(2) is very small, the resolution of the measurement is not good enough to resolve them. As a first approximation it is possible to consider the experimental value as weighted average of the two distances,47 obtaining dCe−OPER = 2.348 Å. This value is larger than the experimentally measured values for the 2 ML in the PER configuration dCe−OPER = 2.312 ± 0.015 Å and slightly smaller than the value for the 10 ML film dCe−OPER = 2.354 ± 0.018 Å. The interatomic distances measured experimentally and evaluated under the hypothesis of 3aCeO2 = 4aPt and 5aCeO2 = 7aPt coincidence are compared in Figure 5.

Figure 4. Model of the atomic geometry of the first coordination shell of Ce atoms in the CeO2 flourite-type structure. The Ce atom (gray) has a cubic coordination with two oxygen neighbors in the [111] direction (O(1) atoms) and six oxygen neighbors (O(2) atoms) which have in- and out-of-plane components of the distance from the Ce atom with respect to the (111) plane.

plane, two of the eight O nearest neighbors of a Ce atom are located along the [111] direction (O(1) atoms in Figure 4), while the remaining six have both in- and out-of-plane components of the Ce−O interatomic vector with respect to the (111) plane (O(2) atoms in Figure 4); the Ce−O(2) vectors form an angle of sin−1(1/3) with respect to the (111) plane. In the PAR geometry the electric vector of the X-ray beam is perpendicular to the Ce−O(1) bonds; hence, in this geometry the EXAFS signal probes only the Ce−O(2) interatomic distance. The Ce−O(2) distance can be expressed as d′Ce−O(2) = [(d′Ce−O(2)∥)2 + (d′Ce−O(2)⊥)2]1/2, where d′Ce−O(2)∥ and d′Ce−O(2)⊥ are the components of the Ce−O(2) interatomic vector parallel and perpendicular to the (111) plane. In the hypothesis of 3aCeO2 = 4aPt coincidence between the cerium oxide film and the Pt substrate, the in-plane component d′Ce−O(2)∥ is compressed, while the out-of-plane component is expanded compared to the bulk values. We obtain d′Ce−O(2)∥ = a′/√3 = 2.136 Å and d′Ce−O(2)⊥ = c′/4 = 0.799 Å which give d′Ce−O(2) = 2.280 Å. The obtained value is in good agreement with the value obtained from the fitting of the data in the PAR configuration for the 2 ML film, dCe−OPAR = 2.291 ± 0.024 Å, while the value of the 10 ML film dCe−OPAR = 2.342 ± 0.017 Å is expanded with respect to the 3aCeO2 = 4aPt coincidence hypothesis and relaxed to the bulk value (Table 1). The EXAFS measurements in the PER configuration were performed with the electric vector of the impinging X-ray beam inclined by 20° with respect to the Ce−O(1) interatomic bond. In this geometry the EXAFS signal depends on both the Ce−

Figure 5. Ce−O interactomic distances measured by XAFS with the electric field parallel (dCe−OPAR) and perpendicular (dCe−OPER) to the sample surface for the 2 ML (black triangle) and for the 10 ML (blue triangle) films. The experimental values are compared with the value expected for a bulk (red dot) and with the values expected under the hypotheses of 3aCeO2 = 4aPt (red open circle) and 5aCeO2 = 7aPt (red open square) coincidence assuming the bulk elastic constants.

The results of the measurements allow to exclude the 5aCeO2 = 7aPt coincidence hypothesis. The 2 ML film in fact exhibits an in-plane contraction (2.4% with respect to the bulk), which is compatible with the hypothesis of a 3aCeO2 = 4aPt coincidence with the underlying Pt surface. The same coincidence structure was found to be more stable than the 5aCeO2 = 7aPt one by DFT calculations on a CeO2 single layer on Pt(111).20 The formation of PtO2 islands on the Pt surface during the CeO2 film growth may also be of help in stabilizing the observed inplane epitaxial relation as discussed in ref 14. The out-of-plane interatomic distance for the 2 ML film is compressed with E



Article

comparison of the values obtained by the two techniques shows that the average in-plane strain is, as expected, slightly larger than the surface one. The mechanisms for the release of the epitaxial strain at the CeO2/Pt interface are indeed an interesting issue, which will be the subject of further investigation.

respect to the value expected assuming the bulk elastic constants. This is probably due to the reduced dimensionality of the system, which may lead to a corrugation of the interface and surface layer, which have a geometric structure different from the bulk, with the interface and surface Ce atom missing the bottom and top oxygen nearest neighbor, respectively (O(1) atoms in Figure 4). It has been in fact recognized that the continuum elastic theory can correctly describe the behavior of films which are at least 5−6 layers thick, while thinner films require an atomic scale description.48 A similar out-of-plane contraction of ultrathin oxide films has already been observed and ascribed to deviations from the bulk elastic behavior.28,49 In their DFT study Spiel et al. calculated the relaxed atomic positions for inequivalent Ce, surface, and interface O atoms for a single cerium oxide layer on Pt(111).20 In the most stable adsorption geometry, i.e., the 3:4 coincidence structure with a surface oxygen atom on top of a Pt atom in the coincidence cell, Ce and O atoms have a nonnegligible corrugation. On average, interface O atoms are closer to Ce atoms by 0.1 Å in the vertical direction, while surface O atoms are slightly more distant by 0.01 Å.20 Our experimental data on the 2 ML film cannot be quantitatively be compared to the published DFT results on a 1 ML film. However, it is reasonable to believe that the driving forces which induce an average out-of-plane contraction of the Ce−O interatomic distances for the monolayer will come into play for the experimentally observed out-of-plane contraction of the 2 ML film. Furthermore, the possible presence of defects and of adsorbate species on the surface may also play a role on the structure of ultrathin films. The 10 ML cerium oxide film on the contrary shows a relaxation of both in- and out-of-plane interatomic distances, which assume values compatible with the bulk one. We observe that for a number of substrates, such as Pt(111),13 Rh(111),11 and Ru(0001),12 ultrathin cerium oxide films on metal surfaces show slight modification of the surface lattice parameter compatible with the smallest coincidence lattice which can be obtained with a reasonable strain. In the case of a Cu(111) substrate, on the contrary, in spite of the negligible (0.2%) lattice mismatch expected for the 2:3 coincidence, the cerium oxide films are significantly (3−5%) compressed at the early stages of the growth.23−25 This effect has been ascribed simply to the low dimensionality of the films. The fact that on other substrates the strain is not only compressive, but also expansive (e.g., on Rh(111)11), points toward a non-negligible influence of epitaxial constraints. While values of the in-plane lattice parameter for the 10 ML sample measured by LEED and EXAFS are very well compatible (EXAFS: aCeO2 = 3.819 ± 0.031 Å; LEED aCeO2 = 3.82 ± 0.05 Å), for the 2 ML sample the in-plane lattice parameter measured by the LEED is slightly less contracted than the one measured by EXAFS (EXAFS: aCeO2 = 3.732 ± 0.044 Å; LEED: aCeO2 = 3.79 ± 0.03 Å). Although the relative errors of both measurements are quite large, the difference can be explained by considering that LEED is sensitive mainly to the outermost surface layer, which may be more relaxed than the layers underneath, while EXAFS measures an average of the interatomic distance over all the layers. It has to be remembered in fact that samples of very low thickness grown in the same conditions used for this work actually form islands, with height of 2 ML or more, which cover a fraction of the substrate surface.14 For the nominally 2 ML thick sample the

■

CONCLUSIONS We have performed an EXAFS study of ultrathin cerium oxide films on Pt(111), well controlled in stoichiometry. The measurements demonstrate that, even at reduced dimensionality, the films have a fluorite structure. The influence of the substrate is evident at the early stages of the growth, where it forces the cerium oxide films to adopt an in-plane compression to match the substrate with a 3aCeO2 = 4aPt coincidence lattice. The out-of-plane parameter is less expanded than expected assuming the bulk elastic constants, as often found in the case of ultrathin films. The epitaxial compression is almost completely relaxed at 10 ML. This work, establishing the exact epitaxial relation between the ceria film and the Pt substrate and giving an accurate evaluation of the cerium oxide film structure, provides a basis for studies of the properties, in particular of the reducibility, of the CeO2/Pt system, where Pt is expected to play a non-negligible role.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The work has been supported by the Italian MIUR through the FIRB Project RBAP115AYN “Oxides at the nanoscale: multifunctionality and applications”. We acknowledge the GILDA beamline and the European Synchrotron Radiation Facility for provision of beamtime and excellent support for the experiment. Support of this work by the COST Action CM1104 is gratefully acknowledged. S. Fugattini and I. Valenti participated in the experiment, and their help is gratefully acknowledged.

■

REFERENCES

(1) Trovarelli, A.; de Leitenburg, C.; Boaro, M.; Dolcetti, G. Catal. Today 1999, 50, 353−367. (2) Park, S.; Vohs, J. M.; Gorte, R. J. Nature 2000, 404, 265−267. (3) Fu, Q.; Saltsburg, H.; Flytzani-Stephanopoulos, M. Science 2003, 301, 935−938. (4) Rodriguez, J. A.; Ma, S.; Liu, P.; Hrbek, J.; Evans, J.; Pérez, M. Science 2007, 318, 1757−1760. (5) Carrettin, S.; Concepcion, P.; Corma, A.; Lopez Nieto, J. M.; Puntes, V. F. Angew. Chem., Int. Ed. 2004, 43, 2538−2540. (6) Migani, A.; Vayssilov, G. N.; Bromley, S. T.; Illas, F.; Neyman, K. M. Chem. Commun. 2010, 46, 5936−5939. (7) Tsunekawa, S.; Ishikawa, K.; Li, Z.-Q.; Kawazoe, Y.; Kasuya, A. Phys. Rev. Lett. 2000, 85, 3440−3443. (8) Tsunekawa, S.; Wang, J.-T.; Kawazoe, Y. J. Alloys Compd. 2006, 408−412, 1145−1148. (9) Zhang, F.; Chan, S.-W.; Spanier, J. E.; Apak, E.; Jin, Q.; Robinson, R. D.; Herman, I. P. Appl. Phys. Lett. 2002, 80, 127−129. (10) Castellarin-Cudia, C.; Surnev, S.; Schneider, G.; Podlucky, R.; Ramsey, M. G.; Netzer, F. Surf. Sci. 2004, 554, L120−L126. (11) Eck, S.; Castellarin-Cudia, C.; Surnev, S.; Ramsey, M. G.; Netzer, F. P. Surf. Sci. 2002, 520, 173−185.

F



Article

(44) Landau, L. D.; Lifshitz, E. M. Course of Theoretical Physics, Theory of Elasticity; Pergamon Press: Oxford, 1986; Vol. 7. (45) Marcus, P. M.; Jona, F. Surf. Rev. Lett. 1994, 1, 15−21. (46) Nakajima, A.; Yoshihara, A.; Ishigame, M. Phys. Rev. B 1994, 50, 13297−13307. (47) Tormen, M.; De Salvador, D.; Natali, M.; Drigo, A.; Romanato, F.; Rossetto, G.; Boscherini, F.; Mobilio, S. J. Appl. Phys. 1999, 86, 2533−2539. (48) Marcus, P. M. Surf. Sci. 1996, 366, 219−227. (49) Groppo, E.; Prestipino, C.; Lamberti, C.; Luches, P.; Giovanardi, C.; Boscherini, F. J. Phys. Chem. B 2003, 107, 4597−4606. (50) Trovarelli, A. Catalysis by Ceria and Related Materials; Imperial College Press: London, 2002.

(12) Lu, J.-L.; Gao, H.-J.; Shaikutdinov, S.; Freund, H.-J. Surf. Sci. 2006, 600, 5004−5010. (13) Grinter, D. C.; Ithnin, R.; Pang, C. L.; Thornton, G. J. Phys. Chem. C 2010, 114, 17036−17041. (14) Luches, P.; Pagliuca, F.; Valeri, S. J. Phys. Chem C 2011, 115, 10718−10726. (15) Hardacre, C.; Roe, G. M.; Lambert, R. M. Surf. Sci. 1995, 326, 1−10. (16) Mullins, D. R.; Zhang, K. Z. Surf. Sci. 2002, 513, 163−173. (17) Wilson, E. L.; Grau-Crespo, R.; Pang, C. L.; Cabailh, G.; Chen, Q.; Purton, J. A.; Catolw, C. R. A.; Brown, W. A.; de Leeuw, N. H.; Thornton, G. J. Phys. Chem. C 2008, 112, 10918−10922. (18) Schierbaum, K.-D. Surf. Sci. 1998, 399, 29−38. (19) Staudt, T.; Lykhach, Y.; Hammer, L.; Schneider, M. A.; Matolín, V.; Libuda, J. Surf. Sci. 2009, 603, 3382−3388. (20) Spiel, C.; Blaha, P.; Suchorski, Y.; Schwartz, K.; Rupprechter, G. Phys. Rev. B 2011, 84, 045412. (21) Suchorski, Y.; Wrobel, R.; Becker, S.; Strzelczyk, B.; Drachsel, W.; Weiss, H. Surf. Sci. 2007, 601, 4843−4848. (22) Suchorski, Y.; Wrobel, R.; Becker, S.; Weiss, H. J. Phys. Chem. C 2008, 112, 20012−20017. (23) Mašek, K.; Beran, J.; Matolín, V. Appl. Surf. Sci. 2012, 259, 34− 38. (24) Szabová, L.; Stetsovych, O.; Dvořaḱ , F.; Farnesi Camellone, M.; Fabris, S.; Mysliveček, J.; Matolín, V. J. Phys. Chem. C 2012, 116, 6677−6684. (25) Dvorǎk,́ F.; Stetsovych, O.; Steger, M.; Cherradi, E.; Matolıńová, I.; Tsud, N.; Škoda, M.; Skala,́ T.; Myslivecěk, J.; Matolıń, V. J. Phys. Chem. C 2011, 115, 7496−7503. (26) Breinlich, C.; Essen, J. M.; Barletta, E.; Wandelt, K. Thin Solid Films 2011, 519, 3752−3755. (27) Boscherini, F. XAFS in the study of semiconductor heterostructures and nanostructures. In Characterization of Semiconductor Heterostructures and Nanostructures; Lamberti, C., Ed.; Elsevier: Amsterdam, 2008. (28) Lamberti, C.; Groppo, E.; Prestipino, C.; Casassa, S.; Ferrari, A. M.; Pisani, C.; Giovanardi, C.; Luches, P.; Valeri, S.; Boscherini, F. Phys. Rev. Lett. 2003, 91, 046101. (29) Luches, P.; Bellini, V.; Colonna, S.; Di Giustino, L.; Manghi, F.; Valeri, S.; Boscherini, F. Phys. Rev. Lett. 2006, 96, 106106. (30) Malvestuto, M.; Carboni, R.; Boscherini, F.; D’Acapito, F.; Spiga, S.; Fanciulli, M.; Dimoulas, A.; Vellianitis, G.; Mavrou, G. Phys. Rev. B 2005, 71, 075318. (31) Malvestuto, M.; Scarel, G.; Wiemer, C.; Fanciulli, M.; D’Acapito, F.; Boscherini, F. Nucl. Instrum. Methods Phys. Res., Sect. B 2006, 246, 90−95. (32) Luches, P.; Pagliuca, F.; Valeri, S.; Illas, F.; Preda, G.; Pacchioni, G. J. Phys. Chem. C 2012, 116, 1122−1132. (33) Pascarelli, S.; Boscherini, F.; D’Acaptito, F.; Hrdy, J.; Meneghini, C.; Mobilio, S. J. Synchrotron Radiat. 1996, 3, 147−155. (34) Maurizio, C.; Rovezzi, M.; Bardelli, F.; Pais, H. G.; D’Acapito, F. Rev. Sci. Instrum. 2009, 80, 063904. (35) Pagliuca, F.; Luches, P.; Valeri, S. Surf. Sci. 2013, 607, 164. (36) Ravel, B.; Newville, M. J. Synchrotron Radiat. 2005, 12, 537− 541. (37) Ankudinov, A.; Ravel, B.; Rehr, J. J.; Conradson, S. Phys. Rev. B 1998, 58, 7565−7576. (38) Solera, J. A.; Garcia, J.; Proietti, M. G. Phys. Rev. B 1995, 51, 2678. (39) Fonda, E.; Andreatta, D.; Colavita, P. E.; Vlaic, G. J. Synchrotron Radiat. 1999, 6, 34−42. (40) Skála, T.; Sutara, F.; Prince, K. C.; Matolín, V. J. Electron Spectrosc. Relat. Phenom. 2009, 169, 20. (41) Skála, T.; Sutara, F.; Skoda, M.; Prince, K. C.; Matolín, V. J. Phys: Condens. Matter 2009, 21, 055005. (42) Bianconi, A.; Marcelli, A.; Dexpert, H.; Karnatak, R.; Kotani, A.; Jo, T.; Petiau, J. Phys. Rev. B 1987, 35, 806−812. (43) Soldatov, A. V.; Ivanchenko, T. S.; Della Longa, S.; Kotani, A.; Iwamoto, Y.; Bianconi, A. Phys. Rev. B 1994, 50, 5074−5080. G


Structure of Ultrathin CeO2 Films on Pt(111) by Polarization

Recommend Documents