Subscriber access provided by UNIV OF LOUISIANA
C: Surfaces, Interfaces, Porous Materials, and Catalysis
Reducibility of Ag- and Cu-modified Ultrathin Epitaxial Cerium Oxide Films Gabriele Gasperi, Luca Brugnoli, Alfonso Pedone, Maria Cristina Menziani, Sergio Valeri, and Paola Luches J. Phys. Chem. C, Just Accepted Manuscript • Publication Date (Web): 06 May 2019 Downloaded from http://pubs.acs.org on May 6, 2019
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
Reducibility of Ag- and Cu-modified Ultrathin Epitaxial Cerium Oxide Films Gabriele Gasperi1,2, Luca Brugnoli3, Alfonso Pedone3, Maria Cristina Menziani3, Sergio Valeri1,2, Paola Luches2* 1
Dipartimento FIM, Università degli Studi di Modena e Reggio Emilia, Via G. Campi 213/a, Modena, Italy 2
S3, Istituto Nanoscienze, Consiglio Nazionale delle Ricerche, Via G. Campi 213/a, Modena, Italy
3
Dipartimento di Scienze Chimiche e Geologiche, Università degli Studi di Modena e Reggio Emilia, Via Campi 103, 41125 Modena, Italy
Abstract The functionality of cerium oxide, and in particular its reactivity, can be significantly altered by the addition of diluted cationic species with different electronic properties as compared to cerium. We investigate the modifications induced by Ag and Cu as modifier cations in cerium oxide ultrathin epitaxial films. The reducibility is assessed by following the modifications of the oxidation state of surface Ce ions by x-ray photoemission spectroscopy, during thermal treatments in ultra-high vacuum and oxygen partial pressure. A significantly higher reducibility is observed in Ag- and Cu-modified films as compared to pure CeO2 films of the same thickness. The thermal stability of the cation modifier concentration and the changes of the surface structure with the reducing treatments are also discussed. The modifications induced in the material are explained by comparison with density functional theory calculations, which indicate that the oxygen vacancy formation energy is significantly modified by the addition of Ag or Cu in the cerium oxide matrix. The obtained results are of help in view of a rational design of catalysts with optimized performance. 1. Introduction Cerium oxide based materials are very important for catalytic applications, largely due to the possibility for the oxide to release and take up oxygen.1-3 This property is in turn linked to the ability of Ce cations to easily and reversibly switch between the 4+ and 3+ oxidation states, accommodating the electrons left after oxygen release in empty Ce 4f levels.4-5 Moreover, cerium oxide can host a high density of oxygen vacancies without a significant deformation of its fluorite-type lattice. The properties of oxides can be modified by introducing small concentrations of modifier cations, which may serve as electron donors or acceptors.6-7 The redox activity of pure ceria has shown an important enhancement when part of the cations are substituted by other metallic ions.8 Indeed, the properties of cation modified oxides in general depend on the specific cation used as modifier (size, valence, etc), and on its concentration.7, 9-10 Catalysts based on modified CeO2 have been shown to have an O ion mobility and a chemical activity larger than the pure oxide.11-15 Fundamental studies based on model systems and theoretical models have largely contributed to an atomic scale understanding of the effects of oxygen vacancies and dopants in cerium oxide.3, 9, 15-24 Doping in ceria has been shown to induce a relevant increase of ionic conductivity.25-26 A dopant atom in general induces a local distortion in the crystal structure, which typically reflects in a decrease of the oxygen vacancy formation energy in its neighbourhood.27 Moreover, if the valence state of the dopant is different from the 4+ valence of Ce ions in CeO2, the observed decrease of oxygen vacancy formation energy is also induced by the modification of the electronic structure of the material.21, 28 In order to obtain active cerium oxide based catalysts the oxygen vacancy formation energy should not be too small, nor 1 ACS Paragon Plus Environment
The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 2 of 26
negative, otherwise the formed vacancies would be stable and they would not contribute to the dynamic oxygen storage capacity, on which the catalytic properties of the material are based. Moreover, dynamic changes of modified ceria-based materials during operation have to be considered in order to properly describe and optimize the material functionalities. As an example, we mention the Pt-modified ceria electrodes for PEM fuel cells developed by Fiala et al.29-30 Such materials have been shown to undergo important modifications with the diluted Pt species being in atomically dispersed ionic configurations under oxidizing conditions and to partially merge into metallic Pt particles of subnanometric size under reducing conditions.31 The abundance of the materials involved in the applications is indeed also a relevant issue, which requires huge efforts towards the development of materials for catalysis with an ultra-low concentration of platinum group metals (PGM), or which do not contain PGM at all. In this context, the present work focuses on an investigation of the reducibility of ceria epitaxial films, modified with Ag and Cu, less critical materials as compared to Pt. We interpret the evolution of the electronic properties during controlled reducing and oxidizing cycles with the help of results obtained by density functional theory (DFT).
2. Experimental and computational details Cerium oxide films were grown and characterized in an ultra-high vacuum (UHV) apparatus, equipped with facilities for reactive molecular beam epitaxy growth and in-situ characterization by X-ray photoemission spectroscopy (XPS), Auger electron spectroscopy (AES), and low energy electron diffraction (LEED). The substrate used was a Pt(111) single crystal, prepared by repeated cycles of Ar+ sputtering (1 keV, 1 μA) and annealing (770 °C) until the level of contaminants was below the sensitivity of AES and the LEED pattern showed sharp spots. Cerium oxide films of controlled thickness and epitaxial quality were grown by depositing metallic cerium, using an e-beam evaporator, in an oxygen background pressure of 1×10-7 Torr with the Pt substrate at room temperature (RT), following a wellestablished procedure.32 To obtain modified films the metal modifier, either Cu or Ag, was evaporated from effusion cells simultaneously with Ce, after having properly tuned the evaporation rates through a quartz crystal microbalance. The operating temperatures of the Ag, Cu and Ce evaporation cells were set to obtain a concentration for Ag and Cu of the order of 10 at%. The resulting atomic concentration of modifier cations was estimated by XPS after film growth to be 13 at% (see Supporting Information). Post growth heating in oxygen background pressure (PO2= 1×10-7 Torr) was performed to optimize stoichiometry, morphology and epitaxial quality of cerium oxide films. A heating temperature of 500°C, relevantly lower as compared to previous studies,32-34 was used in order to limit the mobility of the modifier cations within the lattice. To estimate the reducibility of the oxide films, thermal cycles in UHV and oxygen background pressure were performed following the procedures used in previous studies.33, 35 The oxide composition and the Ce ions oxidation state before and after the different thermal treatments were measured in-situ by XPS, using Al K photons and a hemispherical electron analyser. The photoelectrons were detected both at normal and at grazing emission (0 and 65 from sample normal) to change the sensitivity to the uppermost layers. The concentration of Ce3+ ions was estimated by fitting the Ce 3d peak, following a well-established procedure,36-37 with three doublets related to Ce4+ ions and two related to Ce3+ ions as done in previous works.32-33 The the binding energy of Ag 3d 2 ACS Paragon Plus Environment
Page 3 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
peaks was estimated by fitting with a Voigt-shaped doublet. The binding energy alignment of the spectra of the different samples was checked using the binding energy of Pt 4f peaks, estimated by a fitting the spectra with a Doniach-Sunjich shaped doublet. Unless otherwise stated, the XPS spectra shown and used to evaluate the concentration of the different species are acquired at normal emission. The evolution of the surface structure with the different treatments was assessed by a four-grids rear view LEED using a LaB6 filament and a primary beam energy of 80 eV. The patterns were acquired using a cooled CCD camera and subtracted by a background image acquired with the fluorescent screen at ground potential. All the calculations were performed employing the PBE038 exchange correlation functional implemented in the Crystal program (CRYSTAL17) 39, using a computational protocol similar to that used in ref.40 to study of O vacancies on CeO2 (111). All-electrons Gaussian Type Orbitals (GTOs) Basis Sets (BSs) were employed for describing O41 and Cu42, while for the heavier Ce and Ag atoms two small core Effective Core Potentials43-44 were used to describe the core electrons and GTOs BSs were used for the valence electrons45-46. The 4s24p64d105s25p66s25d14f1 and 5s25p64d10 valence electrons were treated explicitly for Ce and Ag, respectively. With these settings the computed bulk lattice constant for Ceria is 5.402 Å, in good agreement with the experimental value.47 The CeO2(111) surface was modelled employing a p(2x2) supercell formed by 6 atomic layers (OCe-O-O-Ce-O), for a total of 24 atoms. The surface energy for CeO2(111) converges to the value of 0.86 J/m2 with 9 atomic layers, however with 6 atomic layer we obtained a sufficiently close value of 0.85 J/m2. The modification with Ag and Cu was modelled as substitutional point defects of the Ce in the second atomic layer. Other than the pristine structures for pure and modified CeO2 (111), we modelled these systems with one and two O vacancies. The O vacancies were realized by removing the nuclear and electronic charge from the selected O atoms, but leaving a dummy atom (ghost) with the O GTO basis set in place. This allows for a better description of the electron density in the vacancy. The removed O belongs to the first coordination sphere of the substituted Ce. The structural relaxations involve all the atomic layers, only the cell dimensions were fixed to the optimized bulk values. To evaluate the oxidation state of Ce, Ag and Cu we computed the magnetization on the metal atoms and on the coordinating O from the spin population evaluated from the Mulliken’s analysis. The reciprocal space was sampled through a regular sublattice with a Monkhorst-Pack shrinking factor of 6 in the periodic directions x and y, corresponding to 20 points in the irreducible Brillouin Zone (IBZ). Numerical DFT integration was performed considering 75 radial points and 974 angular points of the default integration grid. The Coulomb and the exchange infinite lattice series were truncated with threshold values of 8, 8, 8, 8, and 20 for ITOL1, ITOL2, ITOL3, ITOL4 and ITOL5, respectively. Further details can be found on the manual of the code.48 A tolerance of 10-7 a.u. was used for the convergence of the total energy in the self-consistent field procedure. The structures were relaxed up to simultaneously reach the default convergence threshold for the maximum and root-mean square atomic forces and displacements (4.5 × 10-4, 3.0 × 10-4, 1.8 × 10-3, and 1.2 × 10-3 a.u respectively). All the calculations were spin-polarized.
3 ACS Paragon Plus Environment
The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 4 of 26
3. Results 3.1
Experimental results
Three cerium oxide films with the same nominal thickness of 2 ML grown on a Pt(111) single crystal were investigated. The three samples differ only by the presence of modifier cations, one of them being a pure CeO2 film, the second one containing 13 at% of Ag, and the third one 13 at% of Cu. Each sample was reduced by two heating steps at 250C and 500C for 15 min in UHV, while the reoxidation was performed by a single heating step at 500C for 15 min in an oxygen background pressure of PO2= 1×10-7 Torr. The heating/cooling rate was 180C/min. After each step the samples were cooled down to RT and measured by XPS and LEED. Films with subnanometric thickness were chosen in order to induce a relevant Ce3+ concentration by thermal treatments in vacuum up to maximum temperatures of 500C, based on the results of previous works.33, 35 3.1.1 XPS Results Figure 1 (a) shows the Ce 3d XPS spectra of the three samples after each heating step. While the spectrum of the pure sample only shows mild modifications during the cycles, the spectra of the cation modified samples show a significant evolution after the two reduction steps, with a relevant increase of the intensity of the main Ce3+-related features at binding energies around 886 eV and 905 eV. The shape of the spectra after reoxidation are similar in the three samples and comparable to the initial spectra, showing that the reduction is fully reversible in all cases.
Figure 1. a) Ce 3d XPS spectra of the pure (blue), Ag- (grey) and Cu- (brown) modified cerium oxide films acquired at RT after each step of the reducing and oxidizing cycle. b) Evolution of the Ce3+ concentration with reducing/oxidizing steps, obtained from the fits of the Ce 3d XPS spectra.
Quantitative information on the evolution of the Ce3+concentration was obtained by the fitting the Ce 3d XPS spectra using Ce3+- and Ce4+-related components (see Supporting Information for details). 4 ACS Paragon Plus Environment
Page 5 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
Figure 2. a) Ag 3d XPS spectra of the Ag- modified cerium oxide film and b) Cu 2p XPS spectra of the Cumodified cerium oxide film acquired at RT after each step of the reducing and oxidizing cycle.
The results are shown in Figure 1 (b). Before the heating cycle the three films have a stoichiometry close to CeO2, with the Cu-modified sample having a slightly higher Ce3+ concentration (8%) compared to the other two films. While the Ce3+ concentration in the pure sample increases only to values below 10% with the reducing treatment, both the Ag- and the Cu- modified films reach a Ce3+ concentration around 26-28% at 250C and around 37-40% at 500C. This demonstrates that the Ag and Cu cations as modifiers induce a significantly higher reducibility in cerium oxide. The overall change in Ce3+ concentration induced by the full reducing treatment is smaller in the Cu- modified sample (cCe3+=37%-8%=29%) than in the Ag- modified sample (cCe3+=40%-1%=39%). The final thermal treatment in oxygen decreases the Ce3+ concentration to below 5% in the three samples. XPS allowed to obtain information also on the modifier cation oxidation state within the cerium oxide matrix and on the stability of the material with the reducing/oxidizing treatments. A correct assignment of the oxidation state of the low-concentration cations based on XPS lineshapes is in general non trivial, being the signals rather weak and often superimposed to an intense background. Moreover, in some cases the comparison with reference data acquired on bulk materials may be misleading, due to the very different chemical environments.49 Nevertheless, some indications can be obtained. Figure 2 shows the Ag 3d and the Cu 2p XPS spectra acquired on the Ag- and Cu- modified samples, respectively. The Ag 3d spectra (Figure 2 a) appear symmetric and rather narrow. The binding energy of the center of the Ag 3d5/2 peak is 367.5 eV. It has to be pointed out that the Ag 3d binding energy shows an anomalous negative shift in Ag oxides with respect to Ag in metallic state, due to extra‐atomic relaxation effects.50 The shifts are - 0.3 eV and - 0.8 eV for Ag1+ and Ag2+, respectively.51 The binding energy of Ag 3d in the Ag- modified cerium oxide sample in Figure 2a appears shifted by - 0.3 eV compared to a metallic Ag sample measured with the same equipment in the same conditions,52 in agreement with the hypothesis of a dominant +1 oxidation state for the Ag ions in the cerium oxide matrix. The Cu 2p spectra of the Cu- modified cerium oxide sample, shown in Figure 2b, are also rather sharp and they show the presence of very weak satellites between the two Cu 2p1/2 and Cu 2p3/2 peaks, suggesting a dominant +1 oxidation state for Cu ions in the cerium oxide 5 ACS Paragon Plus Environment
The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 6 of 26
matrix.53 Both Ag and Cu therefore seem to adopt their most stable oxidation states when introduced as modifiers in cerium oxide. The Ag 3d spectra in Figure 2a do not show marked modifications in shape, energy or intensity with the reducing/oxidizing thermal treatments, indicating a good stability of the Ag- modified samples. On the contrary, the Cu 2p spectra in Figure 2b show a significant decrease of intensity after the
Figure 3. Ag (grey) and Cu (brown) concentration in the two cation-modified cerium oxide films after sample growth, after the first preparation treatment in oxygen and after the reducing/oxidizing steps. The solid and dashed lines indicate the modifier cation concentration measured at normal and grazing emission, respectively. The numbers are obtained from the ratio of the areas of the Ag 3d or Cu 2p and the area of Ce 3d XPS lines (see Supporting Information for details).
reducing treatment at 500C. The modifications of the Ag and Cu concentration in the two samples after the treatments (Figure 3) are evaluated from the ratio between the intensity of the XPS signal from modifier cations (Ag 3d or Cu 2p) and from Ce (Ce 3d) in the oxide matrix (see also Supporting Information). In particular, Figure 3 reports the modifier cation concentration measured after the growth at RT (as-grown sample), after the initial thermal treatment in oxygen background pressure, which optimizes the structure (as-prepared sample), and after each of the reducing and oxidizing steps. The two cation modified samples both have an initial modifier cation concentration of 12-13 at%. If evaluated from the XPS data acquired at normal emission the Ag concentration does not show significant modifications after the initial oxidation step in oxygen, after reduction and re-oxidation treatments. If the data measured at grazing emission are considered, a moderate increase of Ag concentration can be detected after the first and after the second oxidation step (Figure 3). This trend can be ascribed to a mild tendency for Ag ions to segregate towards the surface layers with the oxidizing treatment. On the contrary, the Cu concentration shows a drastic decrease after the first oxidation treatment, it is not modified after reduction at 250C, while it shows a further mild decrease after UHV heating at 500°C. This may indicate either a partial migration of Cu ions to deeper layers or the tendency for the ions to reach the surface and evaporate from the sample. The latter hypothesis is considered rather unlikely, since it should involve some degree of surface enrichment before surface depletion, while the Cu concentration measured at grazing emission was systematically lower than the one measured at normal emission (Figure 3). The migration into deeper layers is considered more likely and it may possibly be favoured by some non-negligible tendency to alloy with the Pt substrate. 6 ACS Paragon Plus Environment
Page 7 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
3.1.2 LEED results LEED was used to have information on the evolution of the surface structure in the three samples with the reducing treatment. The LEED patterns acquired before and after reduction are shown in Figure 4. The surface structure after the preparation is the same for the three systems, with the spots corresponding to the Pt(111) surface arranged in a six-fold symmetry pattern, and the spots corresponding to the cerium oxide film surface arranged with a smaller periodicity in reciprocal space
Figure 4. LEED patterns (E=80 eV) of (a) pure, (b) Ag-modified and (c) Cu-modified cerium oxide epitaxial films acquired after the deposition and post-annealing in O2 at 500°C and (d-f) of the same samples after the final reduction step at 500°C in UHV.
compared to Pt, i.e. with the expected (1.37×1.37) periodicity with respect to the Pt surface cell. The patterns are compatible with an epitaxial (111) orientation and a good crystal quality of the films. The comparable quality and symmetry of the LEED patterns of pure and modified films allows to demonstrate the very similar crystal quality of the three different films and the absence of additional ordered phases induced by the modifier cations. In addition, the LEED patterns acquired on the three different samples after reduction are qualitatively similar and they show the appearance of satellites compatible with a (7×7) periodicity with respect to the underlying Pt. It is noteworthy that, in spite of the different Ce3+ concentrations induced by the reduction treatment in the pure and cationmodified films, the (7×7) surface periodicity is common to the three samples. The same surface periodicity was already observed in similar conditions and ascribed to the stabilization of a metastable phase, possibly mediated by the epitaxy with the Pt substrate.54 The 5:7 coincidence supercell, obtained by the superposition of 5 cerium oxide surface unit cells on 7 Pt surface unit cell has a very small (
