AVS2013 Session SS+AS-TuM: Synthesis, Structure and Characterization of Oxides

In this talk, I will summarize surface science studies conducted at the iNANO center addressing a prototypical model oxide system – the rutile TiO₂(110)–(1 × 1) surface. The identification of point defects on the terraces such as oxygen vacancies, hydroxyl groups, and near surface defect such as Ti interstitials will be discussed. Using high-resolution scanning tunnelling microscopy (STM) several examples for reaction of molecules with oxygen vacancies will be presented [1]. These reactions were directly imaged in so-called STM movies. Whereas the reactions on the terraces can be imaged directly, a combination of static STM results with temperature-programmed desorption (TPD) measurements was used to study the reactions at step edges, with the focus particularly on the <1-11> step edges [2,3]. Both, for the reaction on the terraces and at the step edges, water and ethanol were used as probe molecules [1-3]. Accompanying density functional theory (DFT) calculations support the assignments made in the STM studies. In addition, I will address the complex oxygen–TiO₂(110) interaction [4]. Specifically, the role of bulk defects in the oxygen chemistry on reduced rutile TiO₂(110) is highlighted. Finally, I will show that these surface science studies help to improve our understanding of TiO₂ as photo-catalysts. Specifically, the photo-reaction of ethanol on differently prepared TiO₂(110) surfaces will be compared.

[1]J. Ø. Hansen, P. Huo, U. Martinez, E. Lira, Y.Y. Wei, R. Streber, E. Lægsgaard, B. Hammer, S. Wendt, F. Besenbacher, Phys. Rev. Lett. 107, 136102 (2011).

[2] U. Martinez, J. Ø. Hansen, E. Lira, H. H. Kristoffersen, P. Huo, R. Bechstein, E. Lægsgaard, F. Besenbacher, B. Hammer, S. Wendt, Phys. Rev. Lett. 109, 155501 (2012).

[3] H. H. Kristoffersen, J. Ø. Hansen, U. Martinez, Y.Y. Wei, J. Mathiessen, R. Streber, R. Bechstein, E. Lægsgaard, F. Besenbacher, B. Hammer, S. Wendt, Phys. Rev. Lett. 101, 146101 (2013).

[4] S. Wendt, P. T. Sprunger, E. Lira, G. K. H. Madsen, Z. Li, J. Ø. Hansen, J. Matthiesen, A. Blekinge-Rasmussen, E. Lægsgaard, B. Hammer, F. Besenbacher, Science 320, 1755 (2008).

BaTiO₃ is a ferroelectric material that finds applications in a range of traditional and emerging devices. Understanding the surface is prerequisite to controlling interactions during device processing, potential for surface mediated device properties and surface chemical reactions.

Here we present scanning tunneling microscopy (STM) and scanning tunneling spectroscopy (STS) of atomically resolved reconstructions on BaTiO₃ (001) surfaces, √ 5 x √ 5 R26.6^{\circ}$, √ 2 x √2 , $c(4 x 4)$, and √ 13 x √13 R33.7^{\circ}$. We experimentally show the coexistence of two reconstructions √ 2 x √ 2 and $c(4 x 4)$ while DFT calculations indicate that $c(4 x 4)$ is a thermodynamically favorable phase, the √ 2 x √ 2 persists due to a more kinetically favorable path of formation. The structures of the boundaries between the two reconstructions is determined. In the case of the √ 5 x√ 5 R26.6^{\circ}$ surface a comparison of electronic structure determined by STS with theoretical calculations shows that the surface is Ti terminated. In this and other reconstructions Ti-ad atoms create filled and empty states that are imaged in STM. We will briefly discuss a complete BaTiO₃ (001) phase diagram and the influence to Ti adatoms in the choice of step edge direction in the case of the various surface reconstructions mentioned above.

There is widespread interest in discovering materials that can effectively harvest sunlight in the visible region of the solar spectrum in order to drive chemical processes on surfaces. Hematite (α-Fe₂O₃) has received renewed interest recently as the active photoanode in photoelectrochemical water splitting to store solar energy as H₂ [1]. Hematite is very abundant, has a narrow bandgap of 2.1 eV, and is stable in aqueous and oxidizing environments. However, it has extremely poor electron and hole mobilities, which results in short hole diffusion lengths and ultrafast recombination of photogenerated e^- - h⁺ pairs. Additionally, it does not efficiently absorb photons in the 500-600 nm range where the solar spectrum is most intense. The utility of hematite could therefore be significantly increased by red shifting its band gap to harvest more of the solar spectrum.

Eskolaite (α-Cr₂O₃) has a band gap of 3.3 eV, but alloying Cr and Fe to make α-(Fe_1-xCr_x)₂O₃ solid solutions as epitaxial films has been found to result in a reduction in gap down to as low as ~1.7 eV at x = ~0.5 [2,3]. We examine the optical absorption in α-(Fe_1-xCr_x)₂O₃ thin films grown by oxygen-plasma-assisted molecular beam epitaxy (OPA-MBE) on α-Al₂O₃(0001) substrates and explain the observed excitations, and the nature of the band gap dependence on x, through first principles calculations. Photoconductivity (PC) measurements show that the onset of photocurrent decreases by nearly 0.5 eV when x is increased from 0 to ~0.4. However, for x > ~0.6, the films are not sufficiently conductive for PC to be measured.

Karelianate (α-V₂O₃) is weakly metallic at room temperature, so α-(Fe_1-xV_x)₂O₃ films may overcome the conductivity challenges of the α-(Fe_1-xCr_x)₂O₃ system. Substitutional V impurities have also recently been predicted to lower the band gap of α-Fe₂O₃ further than that of Cr-doped α-Fe₂O₃ [4]. Heteroepitaxial thin films of α-(Fe_1-xV_x)₂O₃ were deposited on α-Al₂O₃(0001) substrates by OPA-MBE. Film quality was monitored in situ by reflection high energy electron diffraction (RHEED). In situ x-ray photoemission spectroscopy (XPS) was utilized to characterize the charge states of the cations. Film crystallinity and lattice parameters were determined ex situ by high resolution x-ray diffraction (HRXRD). The effects of the doping concentration on the conductivity and optical properties will be discussed.

[1] K. Sivula et al., Chemsuschem 4, 432 (2011).

[2] H. Mashiko, et al., Appl. Phys. Lett. 99, 241904 (2011).

[3] S.E. Chamberlin et al., J. Phys. Chem. Lett., submitted (2013).

Low-dimensional model systems are of fundamental importance to elucidate emergent phenomena in condensed matter physics and chemistry. Here we present 1-dimensional oxide-metal hybrid model systems, Co-oxide and Mn-oxide nanowires coupled to a Pd surface, which have been fabricated by the decoration of steps of a vicinal Pd(100) surface. We have studied the growth of Co-oxide and Mn-oxide nanostructures on a Pd(1 1 23) surface using LEED and STM, and have probed their electronic properties by STS in a low-T (5K) STM. At low oxide coverages and suitable kinetic conditions, the Pd step edges are decorated by oxide nanostripes, while the Pd(100)-like terraces are covered by a (2x2) layer of chemisorbed oxygen (as a result of the reactive metal oxide deposition procedure). However, there are significant structural differences between Co- and Mn-oxide nanowires: whereas Mn-oxide decorates the Pd steps by monoatomic MnO_x lines in a pseudomorphic (1x1) arrangement, Co-oxide grows in form of 2-4 atom-wide CoO_x nanostripes, with a strained hexagonal structure partially embedded into the outer terrace areas. In addition, rod-like structures attached to the CoO_x step edges are periodically observed, which have been identified by STS as rows consisting of 4±1 Pd atoms that have been ejected from the Pd steps. This indicates that Co-Pd atom exchange at the step edges is a dominant feature during oxide step decoration and growth. The absence of such structures on the MnO_x covered Pd(1 1 23) surface demonstrates that Mn-Pd atom exchange is much less pronounced, thus revealing a significant difference in growth behavior between MnO_x and CoO_x nanostructures and in their interactions with Pd surfaces.

Work supported by the ERC Advanced Grant SEPON.

The detailed structure and electronic properties of the surfaces of the corundum-type oxides α -Al₂O₃, α -V₂O₃, α -Cr₂O₃, and α -Fe₂O₃ continue to be of significant interest due to their importance in catalysis, photocatalysis, and environmental chemistry. The surface structure, and thus the properties of the surface, can be dramatically altered with different processing or treatment conditions. Recently, Henderson¹ has shown that the Cr-terminated α -Cr₂O₃(0001)surface can be fully passivated by exposure to activated oxygen from an electron cyclotron resonance (ECR) plasma source, blocking both the dissociative and the molecular H₂O adsorption channels. We use x-ray photoelectron spectroscopy (XPS) and x-ray photoelectron diffraction (XPD) to investigate the surface structure of α -Cr₂O₃(0001) epitaxial thin films before and after exposure to activated oxygen from an ECR plasma source. The films are deposited on α-Al₂O₃(0001) substrates by oxygen-plasma-assisted molecular beam epitaxy. When cooled or annealed in vacuum, strong evidence for a Cr-Cr-O₃- termination is obtained by comparing the Cr³⁺ XPD angular scan to single scattering simulations. However, after plasma exposure, a high binding energy feature is observed in the Cr 2p XPS spectrum that possesses an ordered structure distinct from the underlying Cr³⁺ of Cr₂O₃, which remains Cr-Cr-O₃-like. Investigation of this new surface structure with simulations of various candidate structures tentatively rules out CrO₂-like configurations. The high binding energy feature likely arises from a higher oxidation state of Cr, although a quantitative charge state assignment is not straightforward. One possibility is the oxidation of the surface layer of Cr to Cr^6-δ with a double chromyl structure (O=Cr=O).

1. M. A. Henderson, Surf. Sci.604, 1502 (2010).

The rare earth oxides (REOs) have potentially versatile applications in heterogeneous catalysis. Catalysis of reactions like dehydrogenation and selective oxidation of organic compounds, methane conversion to syntheses gas or applications in three way catalysis may be mentioned as examples. The versatility of the REOs is based on the ease of valency changes of rare earth metals (change of f-electron configuration) as compared to, e.g., that of d-block elements. In particular, the possible dioxides in the REO series (ceria, praseodymia and terbia) are highly interesting materials for oxidation catalysis due to their ability to easily give off oxygen and switch to different oxidation states. As lattice oxygen participates directly in the molecule-surface reactions (cf. the Mars-van Krevelen mechanism), the selectivity of REO catalysts in oxidation chemistry depends strongly on their oxygen storage/release capabilities. The oxide of samarium for instance, which only forms the sesquioxide Sm₂O₃, seems to be the most effective REO catalyst for partial (selective) oxidation of methane.

In the recent years considerable progress has been made in the study of ceria thin films and oxide structures on a variety of transition metal substrates. Surface Science studies of REO films other than ceria are however fairly limited. In this context we will present two approaches to study other complex REOs like samaria, praseodymia and also terbia in ultra high vacuum (UHV). The presentation concerns a more classical approach for the preparation of samaria and terbia on a Pt(111) single crystal by reactive physical vapor deposition. Praseodymia thin films on the other hand were prepared on a Si(111) substrate by molecular beam epitaxy and new routes for adjusting the oxidation state of those films will be presented. By means of plasma treatment, e.g., we achieved to prepare even the dioxide of praseodymium (PrO₂), a complex oxide which has not been available for UHV studies up to now. In combination with temperature-programmed desorption of oxygen a range of different oxidation states from PrO₂ to Pr₂O₃ can be adjusted and made accessible for model studies. This concept has been expanded to ceria on Si(111) as well, making it possible to even stabilize the surface of the iota-phase (Ce₇O₁₂), which contains ordered oxygen vacancies, for UHV model studies.

While it is a common concept in microelectronics and adjacent fields to dope and mix

oxides to tune their properties this concept has not yet entered the world of surface

catalytic studies of thin films to a large extend. The talk will close with an outlook and first results on the mixture of REOs for UHV model studies.

We present theoretical and experimental results for the surface core-level shifts, SCLS, of MgO(100). Since MgO is an ideal ionic insulator, [1] one would expect shifts to higher binding energy, BE, for a surface Mg²⁺ cation and shifts to lower BE for a surface anion; furthermore, the magnitudes of the shifts of opposite sign should be comparable. This expectation follows from the electrostatics of the different coordination of the bulk and surface atoms. However, we have found that while the Mg(2p) core-level shift is ~+1 eV, as expected, the O(1s) SCLS is almost zero. This unexpected difference between the cation and anion SCLS has been found for a large range of theoretical models for MgO, which include extended embedded cluster models and periodic calculations. The experiments, which also show the different SCLS, were carried out for MgO films on Mo supports. While the electron density of the bulk Mg and O ions and the surface Mg ions have the dominantly spherical shape expected for closed shell species, the distortion of the surface O anions from spherical symmetry is the effect that leads to the different chemical shifts. Thus the SCLS provide another manifestation of the polarizability of the O anions that is associated with the catalytic activity of thin oxide films; see, for example, Ref. [2]. However, in this case the polarization is of the electronic charge density rather than of the geometric structure of the oxide. We analyze the nature of the polarization of the surface O anions and show how this leads to the anomalously small SCLS for these anions. We also consider the broadening of the XPS lines and we relate the Full Width at Half Maximum of the core-levels to the final state geometric modification of bond distances between Mg and O. [3]

1. C. Sousa, T. Minerva, G. Pacchioni, P. S. Bagus, and F. Parmigiani, J. Electron Spectrosc. Relat. Phenom. 63, 189 (1993).

2. H. J. Freund, Chem. Eur. J. 16, 9384 (2010).

3. C. J. Nelin, P. S. Bagus, M. A. Brown, M. Sterrer, and H.-J. Freund, Angew. Chem. Int. Ed. 50, 10174 (2011).