The Electronic Structure of NiO by means of Electron Spectroscopies and theoretical calculations

DEPARTAMENTO DE FÍSICA APLICADA I FACULTAD DE CIENCIAS CONFERENCIA The Electronic Structure of NiO by means of Electron Spectroscopies and theoretical calculations Keywords: Photoemission, X-ray absorption, NiO/oxide interface, Local density approach calculations Leonardo Soriano de Arpe Catedrático de Física Aplicada de la Universidad Autónoma de Madrid Jueves 2 DE NOVIEMBRE A LAS 12:00 h AULA JACQUES-LOUIS LIONS (M2), MÓDULO DE MATEMÁTICAS FACULTAD DE CIENCIAS, UNIVERSIDAD DE MÁLAGA Organiza: Dpto. de Física Aplicada I y Vicerrectorado de Investigación y Transferencia Surface effects in the Ni 2px-ray photoemission spectra of NiO L. Soriano, I. Preda, A. Gutiérrez, and S. Palacín Departamento de Física Aplicada, Universidad Autónoma de Madrid, Cantoblanco 28049 Madrid, Spain M. Abbate Departamento de Física, Universidade Federal do Paraná, Caixa Postal 19091, 81531-990 Curitiba PR, Brazil A. Vollmer BESSY, Albert Einstein Strasse 15, D-12489 Berlin, Germany 共Received 26 March 2007; published 28 June 2007兲 We report experimental and theoretical evidence of surface effects in the Ni 2px-ray photoemission spectra 共XPS兲of NiO. The Ni 2p3/2 surface-enhanced XPS of a NiO sample show a relative enhancement of the intensity of the known satellite at 1.5 eV higher binding energy from the main line, indicating a considerable surface contribution of this satellite. The results are discussed in terms of bulk-octahedral and surfacepyramidal Ni symmetries. Other contributions, like nonlocal screening effects, cannot be neglected. DOI: 10.1103/PhysRevB.75.233417 PACS number共s兲: 73.20.At, 72.80.Ga, 79.60.⫺i I. INTRODUCTION Nickel oxide is one of the most investigated transition metal oxides, whose electronic structure has been controversial for years. Not only has the nature of the band gap been a matter of discussion, but also the electronic structure of their surfaces, defects, etc. Nowadays, it is accepted that NiO is a 3d8charge transfer oxide and the ground state is a mixture of 3d8,3d9Lគ, and 3d10Lគ 2configurations.1–3This gives rise to a complex line shape of the experimental spectra when using spectroscopic techniques, in particular x-ray photoemission spectroscopy 共XPS兲. The Ni 2pphotoemission peak line shape of NiO shows a multipeaked structure which has been widely discussed in the literature. Nowadays, it has been shown that cluster calculations of the Ni 2pphotoemission spectrum for a NiO6cluster give rise to three different peaks, according to the above description of the electronic structure. However, the very-well-known additional shoulder separated by 1.5 eV at higher binding energies from the main line is absent in those calculations.4This satellite appears only when the calculations are extended to a larger Ni7O36 cluster. As a consequence of that, the satellite was called a nonlocal screening satellite and was explained as a screening process due to oxygen atoms belonging to the outer NiO6clusters.4 In this work we present additional experimental evidence and a theoretical model that show that the origin of the satellite is not only due to nonlocal effects, but also to surface effects, being its main contribution that surface effects are separated from the nonlocal screening model. These experimental results are theoretically interpreted by cluster model calculations in terms of the reduced symmetry at the surface 共see Table I兲. The reason for using single NiO5and NiO6 clusters in our calculation is to highlight the surface contribution of the 1.5-eV satellite, even without including the nonlocal screening effect from adjacent clusters. In this picture, the main line 共cគ3d9Lគ兲of the XPS spectra of surface Ni ions shifts 1.0 eV towards higher binding energies, thus significantly contributing to the satellite intensity. These conclusions are completely consistent with recent results showing a splitting of the Ni egstates at the NiO surface5and suggest a new interpretation of the Ni 2pXPS spectra, especially those concerning nanostructured NiO, where surface effects are dominant due to the large surface-to-volume ratio. The so-called nonlocal satellite in the Ni 2pXPS spectra has been conventionally assigned to the presence of Ni3+ species at the NiO surface.6In general, it has been observed that the intensity of this satellite increases with the creation of defects. For instance, Uhlenbrock et al.7concluded that Ar bombardment of a freshly cleaved NiO 共100兲surface produces the formation of Ni3+ species and the simultaneous reduction to Ni0when analyzing the intensity of the satellite with respect to the main line. However, it is well known that Ar bombardment produces oxygen vacancies due to preferential sputtering of oxygen, leading to the reduction of Ni atoms to metallic Ni.8Other interesting systems to study the effect of defects in NiO are highly defective 3–5-nm NiO nanoparticles, whose Ni 2pXPS spectra show a clear enhancement of the relative intensity of the satellite with respect to the spectrum of a NiO single crystal.9This effect was related to the good catalytic properties of the nanoparticles and explained as due to the presence of Ni3+ defects. TABLE I. Parameters of the cluster model calculation. ⌬, charge transfer; U, electron repulsion; pd ␴ ,p-dhybridization; Q,B,C, Racah parameters; 10Dq, crystal field. Cluster parameters U7.5 ⌬4.0 pd ␴ 1.5 Q9.0 Multiplet parameters B0.13 C0.58 10Dq 0.1 PHYSICAL REVIEW B 75, 233417 共2007兲 1098-0121/2007/75共23兲/233417共4兲©2007 The American Physical Society233417-1 However, in a posterior x-ray absorption spectroscopy 共XAS兲study of these nanoparticles,10 Ni atoms were unambiguously characterized as high-spin Ni2+ species. This conclusion together with the results obtained in previous studies of hole-doped LixNi1−xO共Refs. 11 and 12兲make doubtful the existence of Ni3+ defects at the NiO surface. Other authors found a dependence of the intensity of the satellite on the emission angle of a NiO single crystal.13,14 They concluded that the satellite could be associated with the surface Ni atoms but no appropriate theoretical model was proposed to explain the experiment. Other interpretations appearing in the literature assign the satellite to cគ3d10Lគ 2共Ref. 15兲states or cគ3d9multiplets.16 Therefore, the origin of the satellite at 1.5 eV higher binding energies with respect to the main line in the Ni 2pXPS spectra in NiO remains unclear. As mentioned above, the interpretation of the satellite as a nonlocal process is the most accepted theory at present. This model is supported by experimental results in the systems: NixMg1−xO共Ref. 17兲and one epitaxial NiO monolayer grown on MgO single crystal.18 In these systems, the nonlocal satellite is absent as corresponds to systems where, in spite of Ni atoms are octahedrally surrounded by oxygen atoms, no second neighboring oxygen atoms exist. The nonlocal screening model suggests that the satellite structure is intrinsic for bulk NiO. However, when the calculations are restricted to a Ni6O30 cluster, unexpectedly the intensity of the satellite increases. The authors conclude that this could be due to a different coordination of the cluster with the core hole—i.e., fivefold to sixfold.4In turn, the model does not explain the absence of the satellite in the Ni 2pXPS spectrum of La2NiO4as corresponds to a purely octahedral Ni compound.14 Other recent results for nanoscopic NiO systems, where the large surface-to-volume ratio permits the detection of surface effects, concern NiO nanoparticles and planar NiO nanoislands grown on highly oriented pyrolitic graphite 共HOPG兲.5In these systems, the O 1sXAS spectra show a splitting of the Ni egstates at the surface of the nanostructures. This splitting has been explained as arising from the pyramidally coordinated Ni atoms at the surface. Similar surface states have also been experimentally observed19,20 and calculated for a NiO 共100兲surface using local density approach 共LDA兲calculations including electron repulsion.21 The effect of the large surface-to-volume ratio on the Ni 2pXPS spectra has been studied in NiO nanoparticles, giving an increase of the satellite intensity.22 It seems reasonable to think that the above splitting of the Ni 3dlevels at the surface could also be reflected in the Ni 2pXPS spectra. To investigate this possibility, we have performed two simple experimental XPS measurements of a bulk NiO sample. In XPS there are two main techniques to make the spectrum more sensitive to the surface by minimizing the inelastic mean free path23 共␭兲: using a grazing takeoff angle and reducing the kinetic energy of the photoelectrons to approach the minimum of the universal curve. We present below the Ni 2pXPS spectra of a NiO sample measured at different takeoff emission angles and at different photon energies. II. EXPERIMENTAL DETAILS A NiO thin film 共200 Å兲was grown in situ by reactive evaporation of Ni following conventional methods on HOPG.5The measurements at different takeoff angles were performed with a CLAM4 electron analyzer in our laboratory whereas the measurements at different photon energies were performed at the PM4 beamline of the Synchrotron BESSY 共Berlin兲using a Phoibos analyzer. The energy resolution was approximately 0.8 eV in both cases and the angular resolution was about ±3°. III. RESULTS The Ni 2p3/2 XPS spectra of the NiO thin film taken with Al K ␣ radiation at various emission angles 共 ␪ 兲are shown in Fig. 1. The spectrum taken at ␪ =0° 关Fig. 1共a兲兴agrees with that of a NiO single crystal.4The spectrum taken at ␪ =80° 关Fig. 1共b兲兴shows a clear enhancement of the intensity of the satellite at 856.0 eV with respect to that measured at ␪ =0°. Band dispersion and matrix element effects are important for valence levels at low energies, like angle-resolved photoemission spectroscopy 共ARPES兲, but are not expected to significantly influence the core-level spectra at higher energies 共XPS兲. Therefore, the results seem to indicate that the satellite contains a considerable surface contribution. Thus, the first experiment is qualitatively consistent with the interpretation made in this paper. In Fig. 2, the Ni 2p3/2 XPS spectra of the NiO thin film taken at different photon energies are depicted. The kinetic energy of the photoelectrons in the conventional spectrum taken with Al K ␣ radiation 关Fig. 2共a兲兴is about 630 eV, whereas in the spectrum taken in the synchrotron at 1000 eV photon energy 关Fig. 2共b兲兴is about 145 eV, thus giving values FIG. 1. Experimental XPS spectra of a NiO thin film measured with Al anode at 共a兲0° and 共b兲85° emission angle 共 ␪ 兲. The inset shows the geometry of the measurements. BRIEF REPORTS PHYSICAL REVIEW B 75, 233417 共2007兲 233417-2 for the inelastic mean free path of 12.66 Å and 4.24 Å, respectively.20 The spectrum taken with synchrotron light at 1000 eV shows a large increase of the relative satellite intensity. The changes of the cross section with energy affect all the component of the core-level spectra and cannot account for the change of only one component as observed in the experimental data. Therefore, these results are again consistent with the interpretation that the satellite at 856.0 eV has a large surface contribution. However, detailed XPS quantitative analysis of both peaks indicates that the intensity of the satellite is too high to be explained as only due to the surface contribution. This seems to indicate that other processes, like nonlocal screening, could also contribute to the satellite intensity. In order to give an interpretation to the experimental data, we have performed cluster model calculations to theoretically reproduce the Ni 2pcore-level spectra. The clusters were NiO6in octahedral symmetry 共bulk兲and NiO5in pyramidal symmetry 共surface兲. The ground state was expanded in the 3d8,3d9Lគ, and 3d10Lគ 2configurations, where Lគdenotes a symmetry adapted O 2phole.24 The final state was expanded in the cគ3d8,cគ3d9Lគ, and cគ3d10Lគ 2configurations, where cគhere denotes a hole in Ni 2pcore level. The Ni 2pphotoelectron spectra were then calculated using the sudden approximation.18 The parameters of the model were the charge transfer energy ⌬=4.0 eV, the Mott-Hubbard repulsion U=7.5 eV, the core-hole potential Q=9.0 eV, and the p-dtransfer integral T ␴ =2.6 eV.25 These values give the best agreement with the experiment and are in agreement with previous estimates. The multiplet splitting was given in terms of the Raccah parameters B=0.13 eV and C=0.58 eV, as well as the crystal field parameter 10Dq=0.10 eV. The transfer of the x2-y2 level involved T ␴ in both octahedral and pyramidal symmetries. The transfer of the z2level involved also T ␴ in octahedral symmetry, but only 2 3T ␴ in pyramidal symmetry due to the lack of the apical O.26 Figure 3shows the results of the calculation in octahedral 共a兲and pyramidal 共c兲symmetries. For the bulk NiO6cluster, the calculation presents the standard cគ3d9Lគ,cគ3d10Lគ 2, and cគ3d8peaks. For the surface NiO5cluster, the calculations show these peaks but the energy spread is smaller 共the additional peak in the spectrum is due to the different hybridization of the x2-y2and z2levels兲. The overall energy separation between the peaks is roughly given by ⌬E2=共Q−⌬兲2+4Teff 2. The smaller energy spread above is attributed to the reduced Teff in pyramidal symmetry. The cគ3d9Lគpeak at the surface is shifted 1.0 eV toward lower energies with respect to the bulk. These results suggest that the main line of the Ni 2pXPS spectra would come from the bulk, whereas the nonlocal structure would contain a considerable surface contribution. In fact, the mixture of 60% bulk and 40% surface 共although nonlocal effects are also expected to contribute to the satellite intensity兲components in Fig. 3共b兲resembles the experimental Ni 2pXPS. Other effects would also place a significant surface contribution at the nonlocal structure. For instance, the changes in the Madelung potential would affect the charge-transfer energy ⌬and would also reduce the binding energy of the surface contribution by about 0.5 eV.24 The differences in the hybridization were the key to interpret the O 1sXAS spectra of NiO nanostructures.5It is difficult to tell which of these effects dominates in the case of the Ni 2pXPS specFIG. 2. Experimental XPS spectra of a NiO thin film measured at 共a兲1486.8 eV and 共b兲1000 eV photon energies. The measurements were performed at normal emission angle 共 ␪ =0兲.FIG. 3. Results of the cluster calculations performed in 共a兲octahedral symmetry, 共b兲a combination of a 60% octahedral and 40% pyramidal symmetries, and 共c兲pyramidal symmetry. BRIEF REPORTS PHYSICAL REVIEW B 75, 233417 共2007兲 233417-3 trum. A more realistic calculation considering all these effects would be needed to solve this question. IV. CONCLUSIONS In summary, we have studied the line shape of the Ni 2p photoemission spectra by means of experimental XPS spectra and theoretical cluster calculations. Surface enhanced XPS spectra show that the satellite has an important contribution from the surface Ni atoms. Cluster calculations performed in octahedral and pyramidal symmetry show that for surface Ni atoms 共pyramidal兲the main line of the spectrum is shifted by 1.0 eV towards higher binding energies, thus significantly contributing to the intensity of the satellite at 856.0 eV. However, the intensity of the satellite is much higher than that corresponding to a surface layer. This indicates that other effects like nonlocal screening have also to be taken into account. This interpretation of the Ni 2pXPS line shape opens a revision of the Ni 2pXPS spectra of NiO with a large surface-to-volume ratio. ACKNOWLEDGMENTS We would like to thank G. A. Sawatzky for useful suggestions, encouragement, and comments. We want to thank the MECD of Spain for support under Contract No. FIS2006-06240, Comunidad de Madrid/UAM under Contract No. CCG06-UAM/MAT-0227, and the European Union through Contract No. R II 3.CT-2004-506008. We also thank the staff of BESSY for technical support. 1A. Fujimori, F. Minami, and S. Sugano, Phys. Rev. B 29, 5225 共1984兲. 2G. A. Sawatzky and J. W. Allen, Phys. Rev. Lett. 53, 2339 共1984兲. 3S. Hüfner, Solid State Commun. 52, 793 共1984兲. 4M. A. van Veenendaal and G. A. Sawatzky, Phys. Rev. Lett. 70, 2459 共1993兲. 5L. Soriano, A. Gutiérrez, I. Preda, S. Palacín, J. M. Sanz, M. Abbate, J. F. Trigo, A. Vollmer, and P. R. Bressler, Phys. Rev. B 74, 193402 共2006兲. 6M. J. Tomellini, J. Electron Spectrosc. Relat. Phenom. 58,75 共1992兲. 7St. Uhlenbrock, Chr. Scharfschwerdt, M. Neumann, G. Illing, and H.-J. Freund, J. Phys.: Condens. Matter 4, 7973 共1992兲. 8J. M. McKay and V. E. Henrich, Phys. Rev. B 32, 6764 共1985兲. 9A. R. González-Elipe, J. P. Holgado, R. Alvarez, and G. Munuera, J. Phys. Chem. 96, 3080 共1992兲. 10L. Soriano, M. Abbate, J. Vogel, J. C. Fuggle, A. Fernández, A. R. González-Elipe, M. Sacchi, and J. M. Sanz, Chem. Phys. Lett. 208, 460 共1993兲. 11 P. Kuiper, G. Kruizinga, J. Ghijsen, G. A. Sawatzky, and H. Verweij, Phys. Rev. Lett. 62, 221 共1989兲. 12M. Abbate, F. M. F. de Groot, J. C. Fuggle, A. Fujimori, Y. Tokura, Y. Fujishima, O. Strebel, M. Domke, G. Kaindl, J. van Elp, B. T. Thole, G. A. Sawatzky, M. Sacchi, and N. Tsuda, Phys. Rev. B 44, 5419 共1991兲. 13F. Parmigiani, in Cluster Models for Surface and Bulk Phenomena, edited by G. Pacchioni et al. 共Plenum Press, New York, 1992兲, Vol. 1, p. 475. 14L. Sangaletti, L. E. Depero, and F. Parmigiani, Solid State Commun. 103, 421 共1997兲. 15M. Oku, H. Tokuda, and K. Hirokawa, J. Electron Spectrosc. Relat. Phenom. 53, 201 共1991兲. 16K. S. Kim and R. E. Davis, J. Electron Spectrosc. Relat. Phenom. 1, 251 共1972兲. 17M. Atanasov and D. Reinen, J. Electron Spectrosc. Relat. Phenom. 86, 185 共1997兲. 18D. Alders, F. C. Voogt, T. Hibma, and G. A. Sawatzky, Phys. Rev. B54, 7716 共1996兲. 19A. Freitag, V. Staemmler, D. Cappus, C. A. Ventrice, Jr., K. Al Shamery, H. Kuhlenbeck, and H.-J. Freund, Chem. Phys. Lett. 210,10共1993兲. 20B. Fromme, M. Möller, Th. Anschütz, C. Bethke, and E. Kisker, Phys. Rev. Lett. 77, 1548 共1996兲. 21S. L. Dudarev, A. I. Liechtenstein, M. R. Castell, G. A. D. Briggs, and A. P. Sutton, Phys. Rev. B 56, 4900 共1997兲. 22V. Biju and M. Abdul Khadar, J. Nanopart. Res. 4, 247 共2002兲. 23S. Tanuma, C. J. Powell and D. R. Penn, Surf. Interface Anal. 37, 1共2005兲. 24M. D. Towler, N. M. Harrison, and M. I. McCarthy, Phys. Rev. B 52, 5375 共1995兲. 25G. van der Laan, C. Westra, C. Haas, and G. A. Sawatzky, Phys. Rev. B 23, 4369 共1981兲. 26G. Zampieri, F. Prado, A. Caneiro, J. Briatico, M. T. Causa, M. Tovar, B. Alascio, M. Abbate, and E. Morikawa, Phys. Rev. B 58, 3755 共1998兲. BRIEF REPORTS PHYSICAL REVIEW B 75, 233417 共2007兲 233417-4