Preparation and Structural Characterization of Sno2 and Geo2 Methanol Steam Reforming Thin ﬁlm Model Catalysts by (HR)TEM

Materials Chemistry and Physics 122 (2010) 623–629

Contents lists available at ScienceDirect

Materials Chemistry and Physics

journal homepage: www.elsevier.com/locate/matchemphys

Preparation and structural characterization of SnO2 and GeO2 methanol steam reforming thin ﬁlm model catalysts by (HR)TEM

Harald Lorenz a, Qian Zhao b, Stuart Turner c, Oleg I. Lebedev c, Gustaaf Van Tendeloo c, Bernhard Klötzer a, Christoph Rameshan a,d, Simon Penner a,∗ a Institute of Physical Chemistry, University of Innsbruck, A-6020 Innsbruck, Austria b School of Materials Science and Engineering, Tianjin University, Tianjin 300072, People’s Republic of China c EMAT, University of Antwerp, B-2020 Antwerp, Belgium d Department of Inorganic Chemistry, Fritz-Haber Institute of the Max-Planck Society, D-14195, Berlin, Germany article info abstract

Article history: Structure, morphology and composition of different tin oxide and germanium oxide thin film catalysts for Received 21 October 2009 the methanol steam reforming (MSR) reaction have been studied by a combination of (high-resolution) Received in revised form transmission electron microscopy, selected area electron diffraction, dark-field imaging and electron 23 December 2009 energy-loss spectroscopy. Deposition of the thin films on NaCl(0 0 1) cleavage faces has been carried out Accepted 23 March 2010 by thermal evaporation of the respective SnO2 and GeO2 powders in varying oxygen partial pressures and at different substrate temperatures. Preparation of tin oxide films in high oxygen pressures (10−1 Pa) Keywords: exclusively resulted in SnO phases, at and above 473 K substrate temperature epitaxial growth of SnO Electron microscopy −3 −2 Electron energy-loss spectroscopy on NaCl(0 0 1) leads to well-ordered films. For lower oxygen partial pressures (10 to 10 Pa), mixtures ␤ SnO of SnO and -Sn are obtained. Well-ordered SnO2 films, as verified by electron diffraction patterns and 5 Thin films energy-loss spectra, are only obtained after post-oxidation of SnO films at temperatures T ≥ 673 K in 10 Pa Catalysis O2. Preparation of GeOx films inevitably results in amorphous films with a composition close to GeO2, which cannot be crystallized by annealing treatments in oxygen or hydrogen at temperatures comparable

to SnO/SnO2. Similarities and differences to neighbouring oxides relevant for selective MSR in the third group of the periodic system (In2O3 and Ga2O3) are also discussed with the aim of cross-correlation in formation of nanomaterials, and ultimately, also catalytic properties. © 2010 Elsevier B.V. All rights reserved.

1. Introduction rheotaxial growth and thermal oxidation, among others [3–7].A concise discussion of different preparation techniques for thin films SnO2 is one of the most extensively studied and best charac- has been outlined by Sberveglieri [8]. Regarding heterogeneous terized oxides thanks to its outstanding chemical and electronic catalysis, the activity and selectivity of SnO2 catalysts can be signif- properties. A comprehensive review of all aspects especially of icantly improved by doping with other metals, such as Cu, Pd, Cr or its surface and materials science has been given by Batzill and Sb, but even pure SnO2 itself is an active oxidation catalyst, espe- Diebold [1]. Recent important applications of SnO2 range from cially in CO oxidation reactions. Since a Mars-van-Krevelen-type solid-state gas sensors (especially for reducing gases such as H2, reaction mechanism is assumed to be prevalent, a central issue in CO or H2S) over transparent conductors to heterogeneous catal- oxidation catalysis of SnO2 refers to its easy (surface) reducibility ysis [1,2]. As these applications strongly depend on the structure [1]. Most important, recent investigations indicate that pure SnO2 and morphology of the SnO2 samples, special attention has been exhibits a pronounced activity and/or selectivity in methanol steam paid to the development of reproducible preparation methods of reforming [9]. The interest to include SnO2 and GeO2 in studies a variety of SnO2 samples, especially ordered thin films suitable on methanol steam reforming is derived from previously stud- 4+ for (high-resolution) transmission electron microscopy HR(TEM) ied oxides such as ZnO [10],Ga2O3 [11] or In2O3 [12]. Since Sn characterization. These methods include sputtering, spray pyrol- and In3+ represent isoelectronic species a potential functional analysis, electron-beam evaporation, chemical vapour deposition or ogy also of their catalytic action may be anticipated. As the same holds for Ge4+,Ga3+ and Zn2+, extension of the studies to oxides of the fourth main group of the periodic system is straightfor- ward. Contrary to SnO2, structural information on GeO2 thin films is ∗ scarce and mainly limited to the preparation of GeO nanomaterials Corresponding author. Tel.: +43 5127055056; fax: +43 5125072925. 2 E-mail address: [email protected] (S. Penner). [13–15].

In order to close the materials gap between surface studies especially on tin oxides, where a considerable amount of data already exists [1] and studies on “real world” powder catalysts, the thin film model concept will be exploited. From the structural point of view, thin films grown on well-ordered substrates are usually epitaxial or at least well-ordered, that is, structurally resembling single crystal studies conducted in UHV. They usually also exhibit a narrow particle size distribution and well-defined particle sizes. Since at the same time, our thin film dedicated micro-reactor setup allows correlation of the catalytic activity/selectivity measured on both thin film and powder samples [11,12], structure–activity correla- tions of both model systems (thin films) and real-world catalysts (powders) can be easily established. The aim of the present con- tribution therefore is – in close correlation to similar studies on Ga2O3 and In2O3 [11,12] – to use NaCl(0 0 1) cleavage faces as substrates to eventually grow single-crystalline tin and germanium oxide films to present an easy and reproducible preparation pathway to different thin film model systems of tin and germanium oxides and subsequently their thorough structural characterization, prior to catalytic testing. Adequate tools to fulfil this task are considered HRTEM, selected area electron diffraction (SAED) and electron energy-loss spectroscopy (EELS). Special attention will Fig. 1. Overview bright-field TEM image of a SnO thin film after deposition at 373 K also be paid to the influence of deposition parameters (O2 back- −1 on NaCl(0 0 1) in 10 Pa O2. The SAED pattern is shown as an inset. ground pressure, substrate temperature) on film stoichiometry and morphology. by freshly cleaving the NaCl(0 0 1) crystals immediately before deposition of the 2. Experimental oxide.

All tin and germanium oxide films were prepared in a high-vacuum chamber operated at a base pressure of 10−4 Pa. Film thicknesses (usually 25 nm) were mea- 3. Results and discussion sured in situ by a quartz crystal microbalance. SnO2 and GeO2 powders (tin(IV) oxide, 99.99%, and germanium(IV) oxide, 99.9999%, both Alfa Aesar) were thermally 3.1. SnO and SnO thin films evaporated from tantalum crucibles onto vacuum-cleaved NaCl(0 0 1) surfaces at 2 −3 varying substrate temperatures (298–603 K) and O2 background pressures (10 to 10−1 Pa). The films were floated and rinsed with distilled water, dried and finally 3.1.1. Preparation of SnO thin film model catalysts mounted on gold grids for electron microscopy. Annealing and oxidative treatments The structure of the SnO films after deposition in 10−1 Pa 5 (10 Pa He or O2 for 1 h at 473 K, respectively) were performed in a circulating batch O2 at low substrate temperatures (373 K) is shown in Fig. 1. reactor. The structure and morphology of these thin films in the as-deposited state Although the film exhibits no pronounced contrast in bright-field and upon oxidative and annealing treatments were monitored by (high-resolution)- transmission electron microscopy (HRTEM). The electron diffraction patterns were TEM images, its regular grain-size structure is easily visible. The externally calibrated with respect to the (1 1 1), (2 0 0) and (2 2 0) Pd reflections of average size of the round-shaped grains is approximately 10 nm. an as-deposited, untreated Pd/Ga2O3 catalyst. Overview TEM imaging and selected The corresponding SAED pattern (shown as inset) represents an area electron diffraction (SAED) were carried out with a ZEISS EM10C microscope, almost amorphous structure and shows a single blurred diffrac- high-resolution imaging, electron energy-loss spectroscopy (EELS), and dark-field imaging were performed using a JEOL 4000EX high-resolution microscope operated tion ring. Raising the substrate temperature to 473 K converts at 400 kV and a JEOL 3000F FASTEM microscope operated at 300 kV and equipped the film into a single-crystalline phase, as revealed by the SAED with a GIF 2000 post-column electron energy-loss spectrometer. The film composi- pattern shown in Fig. 2b. Fourfold split intense diffraction spots tion was checked by energy-dispersive X-ray analysis (EDXS), which only showed at 2.7, 1.9 and 1.3 Å correspond to the SnO (1 1 0), (2 0 0) and peaks due the evaporated elemental thin film constituents (Sn, Ge and O) and the (2 2 0) lattice spacings of the tetragonal SnO phase (lattice con- gold grid (Au). Surface carbon impurities present on the films were removed by Ar- sputtering prior to TEM imaging. No influence of the sputtering on structure and stants a = 3.79 Å and c = 4.84 Å, respectively [dtheor(1 1 0) = 2.686 Å, morphology of the films has been detected. The purity of the substrate was ensured dtheor(2 0 0) = 1.899 Å, dtheor(2 2 0) = 1.343 Å] [16]. As expected for

−1 Fig. 2. Overview bright-field TEM image of a SnO thin film after deposition at 473 K on NaCl(0 0 1) in 10 Pa O2 (a). The SAED pattern (b) evidences the tetragonal SnO crystal structure of the film. H. Lorenz et al. / Materials Chemistry and Physics 122 (2010) 623–629 625

Fig. 4. Dark-field image of the SnO film shown in Fig. 2 obtained using the SnO(2 0 0) diffraction reflection. The bright regions correspond to grains giving rise to the 2 0 0 reflection. Other differently oriented grains remain dark (examples indicated by arrows).

temperature. The high-resolution image in Fig. 5b shows such an area. All reflections in the FFT pattern fit with the tetragonal SnO2 Fig. 3. High-resolution electron micrograph of the SnO thin film shown in Fig. 2. structure imaged along the [0 0 1] zone axis orientation. The inset FFT pattern shows a strong ring corresponding to the 3.0 Å {101} spacing of tetragonal SnO. A single reflection corresponding to the 2.7 Å {110} spacing is The effect of O2 background pressure variation at constant sub- also present. strate temperature is highlighted in Fig. 6. The left panel shows −2 growth of SnO at 473 K in 10 Pa O2, on the right panel the same −3 situation in 10 Pa O2 is outlined. In the TEM image of Fig. 6a the the tetragonal phase, the (1 1 0)/(2 −2 0) spots form an angle of 45◦ former SnO structure formed after deposition at 473 K in 10−1 Pa with the (2 0 0) diffraction spots. The corresponding bright-field O2 is still prevailing, but additional round-shaped grains of about TEM image (Fig. 2a) exhibits much enhanced diffraction contrast 20–30 nm size are also observed. The SAED pattern (inset) still and shows extended dark domains corresponding to differently shows strong diffraction spots of the tetragonal SnO phase, but also oriented SnO grains. These observations are corroborated by addi- additional Debye–Scherrer-type diffraction rings. A detailed analy- tional high-resolution imaging, shown in Fig. 3. The HRTEM image sis reveals the presence of a second metallic ␤-Sn phase, with strong in Fig. 3 reveals a complex arrangement of individual SnO grains reflections at 2.9 and 2.8 Å, corresponding to the ␤-Sn(2 0 0) and ␤ each exhibiting lattice fringes of ∼3.0 Å, corresponding to the SnO -Sn(1 0 1) fringes [dtheor(2 0 0) = 2.915 Å and dtheor(1 0 1) = 2.792 Å] (1 0 1) lattice spacings [dtheor(1 0 1) = 2.989 Å] [16]. The inset fast [18]. It is worth noting, that these sphere-like structure corresponds Fourier transform pattern (i.e. the micro-diffraction pattern) evi- to the first stages of oxide film preparation by rheotaxial growth dences these extended SnO(1 0 1) lattice spacings by an ring at and thermal oxidation [7,8], where metallic tin is deposited to form ∼3.0 Å, in agreement with the SAED patterns obtained from this large droplets and subsequently oxidized to result in an oxide film sample. The FFT pattern also shows some reflections at ∼2.7 Å with a high surface-to-volume ratio. These features are even more spacing, corresponding to the SnO (1 1 0) lattice spacing. Both elec- pronounced if the deposition is carried out with less O2 (Fig. 6b). At −3 tron diffraction and HRTEM therefore unambiguously confirm that 10 Pa O2 background pressure the film structure does not resem- after deposition at 473 K in 0.1 Pa O2 a single, well-ordered SnO ble SnO, but rather shows very large sphere-like aggregates (up to phase is formed. Furthermore, the SAED pattern of Fig. 2b suggests 50 nm in size). The SAED patterns still show SnO, but the intensity a preferential epitaxial growth of SnO on NaCl(0 0 1) in the rela- of the ␤-Sn phase increased considerably. tionship SnO[0 0 1]//NaCl[0 0 1] and SnO[0 1 1]//NaCl[0 1 1]. These The issue of influence of O2 background pressure on phase for- observations agree well with previous studies of Choi et al. or Feng mation has also been addressed by Jimenez et al. upon SnO and et al., where highly oriented tin oxide films (mostly tetragonal SnO2 deposition on MgO and SiO2 [19]. In agreement with the SnO2) were successfully grown on single-crystalline Al2O3 sub- present studies, heating of SnO2 and post-exposition to 0.1 Torr strates [2,17]. From dark-field experiments (shown in Fig. 4)we (∼13 Pa) O2 for 300 s resulted only in SnO stoichiometry. Complete also infer that the contrast variations are due to diffraction con- oxidation to SnO2 was only obtained after exposure to an oxygen trast of differently oriented SnO grains. The dark-field image was plasma at 300 K. In general, SnO thin films are mostly prepared via taken by using the SnO (2 0 0) diffraction spot. Consequently, grains oxidation of Sn layers [20] or spray pyrolysis [21]. However, the giving rise to this diffraction spot appear bright in the dark-field above-outlined preparation method is unique since it yields repro- image. All other differently oriented grains remain dark (see e.g. ducible and well-ordered SnO films at low substrate temperatures. areas indicated by arrows in the lower left corner of the image). A more complex scenario is present if the substrate tempera- 3.1.2. Preparation of SnO2 thin film model catalysts ture is raised to 573 K (Fig. 5). The bright-field TEM image in Fig. 5a The question how to reproducibly prepare SnO2 thin films is shows a complete reconstruction of the SnO structure and extended intimately connected with the ability to carry out deposition of Moire-fringes indicate overlapping lattice of differently oriented tin oxides at oxygen background pressures higher than 10−1 Pa. SnO grains. However, the SAED patterns (not shown) are identi- Usually, e.g. for Ga2O3 or In2O3, an oxygen background pressure of cal to the one shown in Fig. 2b, indicating that most of the film 10−2 Pa is sufficient to fully oxidize the respective thin films during remains SnO. Some small areas of the film have however almost deposition. Deposition of tin oxides is different since both of the fully oxidised to SnO2 as a result of the raising of the substrate oxides, SnO and SnO2, are thermodynamically stable oxides and 626 H. Lorenz et al. / Materials Chemistry and Physics 122 (2010) 623–629

Fig. 7. Overview bright-ﬁeld TEM image of a SnO thin ﬁlm after deposition at 573 K −1 5 on NaCl(0 0 1) in 10 Pa O2 after post-oxidation at 673 K in 10 Pa O2 for 1 h.

−1 deposition in 10 Pa O2 only leads to SnO phases. Higher oxygen background pressures, even if they are needed for SnO2 deposition, are basically excluded since the deposition rate decreases drasti- cally for very high oxygen background pressures due to oxidative inhibition of the formation of volatile gas phase species. Hence, the only preparation pathway to SnO2 thin films, if the deposition is carried out by thermal evaporation of SnO2 is via formation of SnO and adequate post-oxidation procedures. Fig. 7 highlights the structure of a previous SnO film prepared at 473 K in 10−1 Pa oxygen background pressure and post-oxidized 5 at 673 K in 10 Pa O2. The overview TEM image reveals a coarse grain structure similar to the original SnO film and exhibits strong diffraction contrast. However, the corresponding HRTEM images and SAED patterns indicate formation of SnO2. Fig. 8 shows a high- resolution TEM image with extended SnO2(1 1 0) lattice fringes of 3.30 Å spacing, corroborating the SAED analysis. The inset ED ring pattern mainly shows rings corresponding to the SnO2 struc- Fig. 5. Overview bright-field TEM image of a SnO thin film after deposition at 573 K ture (at 3.4, 2.6, 2.4 and 1.8 Å, corresponding to the (1 1 0), (1 0 1), −1 on NaCl(0 0 1) in 10 Pa O2 (a) and a HRTEM image showing a small SnO2 region (2 0 0) and (2 1 1) reflections of the tetragonal SnO2 structure within the SnO film (b). The [0 0 1] zone axis orientation and structure of the SnO 2 (a = 4.73 Å, c = 3.18 Å; dtheor(1 1 0) = 3.347 Å; dtheor(1 0 1) = 2.643 Å; region is evidenced by the inset FFT pattern. dtheor(2 0 0) = 2.369 Å; dtheor(2 1 1) = 1.764 Å] [22]). It is worth noting, that SnO2 appears to be at least partially ordered, as judged from the elongated diffraction spots overlapping with the diffrac-

−2 −3 Fig. 6. Overview bright-ﬁeld TEM image of a SnO thin ﬁlm after deposition at 473 K on NaCl(0 0 1) in 10 Pa O2 (a) and 10 Pa O2 (b). The SAED patterns are shown as insets. H. Lorenz et al. / Materials Chemistry and Physics 122 (2010) 623–629 627

trigonal oxygen coordination in the ﬁrst coordination shell) [23] the electron-energy near-edge structure of the Sn M4,5 and O–K edges is a reliable tool to distinguish between the different types of tin oxides. The top-most spectrum is typical for SnO2, with a wider splitting of the peaks in the region 530–550 eV and the more pronounced multiple scattering resonances at higher energy losses (around 575 eV). In contrast, the middle spectrum shows two peaks of almost equal intensity with a narrower splitting and a slightly less pronounced ELNES around 570–575 eV. These differences are basically due to the different coordination of oxygen in both tin oxides [23]. The bottom spectrum, corresponding to ␤-Sn accord- ing to the SAED patterns, only shows a broad edge typical for a delayed Sn M4,5 edge [23].

3.2. GeO2 thin ﬁlms

The preparation of ordered, single-phase GeO2 thin films is ham- pered by the polymorphism of GeO2, that is, it exists not only a hexagonal ␣-quartz-type and a tetragonal rutile-type structure but also a glassy amorphous phase depending on pressure, temperature and preparation conditions. Concerning the preparation of thin films using NaCl as growth templates, the solubility of the different polymorphs in water is most crucial. Whereas ␣-quartz- type GeO2 is slightly soluble in water, rutile-type GeO2 is insoluble Fig. 8. High-resolution electron micrograph of the thin film shown in Fig. 7. The [24]. Depending on which phase forms, floating the thin films in inset SAED pattern evidences the tetragonal SnO2 crystal structure, a faint diffuse water therefore might pose problems, as was the case for ZnO ring corresponding to the SnO (1 0 1) spacing can also be made out. The arrows [25]. However, as discussed below, we succeeded in preparing indicate examples of widespread defects (stacking faults) present in the film. self-supporting GeO2 films, which are not affected by the floating procedure. Generally, GeO2 samples are almost uniquely prepared tion rings. A single diffuse ring at 3.0 Å is present, arising from a by oxidative treatments of Ge films or powders [18,26]. very small amount of SnO. The arrows in the HRTEM image indi- For GeO2 we followed a similar strategy as for SnO and SnO2 cate the wide spread defective (mainly stacking faults) structure of and tried to assess structure, composition and influence of depo- the SnO2 film. Additional information on the formation of different sition parameters. Fig. 10 shows a bright-field overview TEM −2 tin oxide species is provided by EELS. Fig. 9 shows an overview of image of a GeO2 thin film prepared at 300 K and 10 Pa O2 back- the Sn M4,5 and O–K edges of SnO films after deposition at 473 K in ground pressure. Both the image and the SAED pattern (inset) −1 10 Pa O2 (middle), after post-oxidation at 673 K (top) and after indicate the presence of an amorphous germanium oxide phase. reduction at 573 K in 105 Pa hydrogen (bottom). The interaction of It is worth to note, that neither raising the substrate tempera- the tin oxides with hydrogen is discussed in part B, but for the sake ture to 623 K nor post-oxidation treatments up to 673 K lead to of clarity, the EELS spectrum is included in this figure. Due to the the formation of single-crystalline phases and only low contrast different structural environments in SnO and SnO2 (tetrahedral vs. amorphous structures are obtained. Recent studies have shown that thin films prepared by evaporation of GeO2 are usually sub-

Fig. 9. EELS spectra (Sn M4,5 edge at 485 eV and O–K edge at 532 eV) of differently prepared tin oxide thin films (top: SnO2 corresponding to the thin film shown in Fig. 7; middle: SnO corresponding to the film shown in Fig. 2; bottom: SnO thin film 5 after reduction in H2 at 573 K (10 Pa, 1 h)). The two oxide phases can be discerned by the splitting of the fine structure after 530 eV as well as the elevated peak at Fig. 10. Overview bright-field TEM image of a GeO2 thin film after deposition at −2 550–560 eV in SnO2 (indicated by *) [27]. 300Kin10 Pa O2. The SAED pattern is shown as an inset. 628 H. Lorenz et al. / Materials Chemistry and Physics 122 (2010) 623–629

show a strong dependence of the film structure and morphology on the substrate temperature, although formation of nanomaterials, such as octahedrons or pyramidal-shaped particles as obtained for In2O3, has never been observed. On the contrary, GeO2 films resemble Ga2O3 structures, at least at low substrate temperatures. For both materials, amorphous structures are usually observed, even after oxidative and annealing treatments at temperatures as high as 673 K. As for SnO/SnO2, no nano-architectures have been observed for GeO2 even at high substrate temperatures (623 K). This behaviour might be interpreted in terms of the presence of different growth species during oxide deposition and the resulting altered wetting properties on NaCl(0 0 1). For Ga2O3 and In2O3, these growth species were basically associated with sub-stoichiometric oxides (e.g. Ga2OorIn2O) with a high mobility on NaCl(0 0 1) at high substrate temperatures. Therefore, the formation of the sphere- like aggregates for Ga2O3 and pyramidal-shaped particles for In2O3 is the result of efficient transport processes preceding the oxidation of the volatile sub-stoichiometric oxides and their re-oxidation Fig. 11. EELS spectra (O–K edge at 532 eV and Ge L2,3 at 1217 eV) of differently towards ordered aggregates (apart from crystallographic matching prepared germanium oxide thin films (top: GeO2 corresponding to the thin film 5 of In2O3 and NaCl). For SnO2 we might interpret the growth species shown in Fig. 10; middle: GeO2 film after oxidation in O2 at 673 K (10 Pa, 1 h); 5 as SnO and/or Sn metal, as confirmed by the SAED patterns of films bottom: GeO2 thin film after reduction in H2 at 673 K (10 Pa, 1 h). prepared at different oxygen partial pressures. The highest oxygen partial pressures that can be used in our HV chamber during depo- stoichiometric and can only be crystallized by post-annealing at sition (10−1 Pa) are sufficient to convert Sn metal into SnO, but no very high temperatures [27,28]. Our observations agree well with further oxidation to SnO2 can be achieved. The same probably holds studies of Ardyanin et al. on the structure and photoluminescence for GeO2, but since only amorphous films are obtained, a thorough properties of e-beam evaporated GeOx films. GeO2 was evapo- structural characterization by TEM, apart from EELS, is excluded. −8 ∼ −6 rated in UHV at 10 Torr ( 10 Pa; with pressure increase to It is clear, that the outlined results represent only a first step × −6 ∼ × −4 3 10 Torr ( 3 10 Pa) during deposition) at 373 K on sil- towards the full characterization of these oxide systems also in icon substrates and only amorphous GeOx films were obtained catalytic reactions. Since partly the aim of this study is the correla- (film thickness: 200 nm). Only after annealing at 773 K, disproportion of the results to In2O3 and Ga2O3, a thorough catalytic testing tion of GeOx into crystalline Ge metal and amorphous GeO2 was of these systems in methanol steam reforming has been carried observed [28]. As the method of choice to gain information on out and discussed in detail elsewhere, also in the context of acidic the composition of the amorphous GeO2 films, EELS spectra of the and basic surface sites leading to different activity and selectivity as-deposited thin film, after oxidation at 673 K and after reduc- patterns for SnO, SnO2 and GeO2 [30]. tion in hydrogen at 673 K were collected. This will also ensure In any case, both oxides are promising candidates as supports a direct correlation to the properties of SnOx films. Fig. 11 high- for small Pd catalyst particles, since they can be reproducibly pre- lights a set of O–K edges and Ge L2,3 edges of the films after the pared as free-standing, stable oxide support films, into which the above mentioned treatments. In short, these edges are typical of noble metal particles can easily be embedded. Due to their amor- ␣-quartz-type GeO2 [29] and no substantial differences between phicity, in TEM experiments GeO2 films especially favour the easy the spectra can be detected. Most important, the spectra allow discrimination of the Pd particles within the oxide matrix due to the determination of the film stoichiometry by quantification of enhanced contrast, as it has been, e.g. the case for Pd/Ga2O3 [11]. the oxygen-to-germanium ratios. For the as-deposited film and the Hence, bimetallic single phases of Pd–Sn and Pd–Ge might be oxidized sample, ratios of O:Ge = 66 at.%:34 at.% and 67 at.%:33 at.%, prepared by reduction in hydrogen and catalytically characterized respectively, have been obtained. This indicates that already after in methanol steam reforming, in close correlation to Pd supported deposition the film stoichiometry is close to GeO2. For the reduced on ZnO, Ga2O3 and In2O3. sample, this ratio shifts to O:Ge = 58 at.%:42 at.%, indicating that the overall film stoichiometry changes towards GeO. We therefore assume that the films are close to a GeO2 stoi- Acknowledgements chiometry in the as-deposited state and annealing treatments (also in hydrogen and oxygen) up to 673 K, which are typically used for We thank the FWF (Austrian Science Foundation) for financial catalyst regeneration treatments, are not sufficient to induce a com- support under project P20892-N19. The authors acknowledge sup- plete reduction of GeO2, disproportion of GeOx or the formation of port from the European Union under the Framework 6 program any crystalline phase (metallic or oxidic). under a contract from an Integrated Infrastructure Initiative (Ref- erence 026019 ESTEEM). 4. Conclusions References We successfully demonstrated reproducible preparation path- ways to thin film model catalysts of SnO, SnO2 and GeO2 by simple [1] M. Batzill, U. Diebold, Prog. Surf. Sci. 79 (2005) 47–154. thermal evaporation of SnO2 and GeO2 in different oxygen envi- [2] Y. Choi, S. Hong, Sens. Actuators B 125 (2007) 504–509. ronments and, for SnO , by post-oxidation of the resulting films [3] S. Semancik, R. Cavicchi, Thin Solid Films 206 (1991) 81–87. 2 [4] K. Murakami, K. Nakajima, S. Kaneko, Thin Solid films 515 (2007) 8632–8636. at 673 K. Generally, some differences and similarities to the pre- [5] C.F. Wan, R.D. McGrath, W.F. Keenan, S.N. Frank, J. Electrochem. Soc. 136 (1989) viously studied neighbouring oxides (In2O3,Ga2O3) in the third 1459–1463. group of the periodic system are notable [11,12]. From the struc- [6] Q. Wu, J. Song, J. Kang, Q. Dong, S. Wu, S. Sun, Mater. Lett. 61 (2007) 3679–3684. [7] G. Sberveglieri, G. Faglia, S. Gropelli, P. Nelli, Proc. 6th Int. Conf. Solid-State tural point of view, SnO/SnO2 more closely resembles In2O3 thin Sensors and Actuators, San Francisco, CA, USA, June 23–27, 1991, pp. 165–168. films in terms of crystallinity and epitaxial growth. Both materials [8] G. Sberveglieri, Sens. Actuators B 6 (1992) 239–247. H. Lorenz et al. / Materials Chemistry and Physics 122 (2010) 623–629 629

[9] T. Mori, S. Hoshino, A. Neramittagapong, J. Kubo, Y. Morikawa, Chem. Lett. 13 [21] L. Sodomka, A. Kleprik, Jemna Mech. Opt. 8 (1963) 43–46. (2002) 390–391. [22] W. Baur, Acta Crystallogr. A 9 (1956) 515–520. [10] C. Wöll, Prog. Surf. Sci. 82 (2007) 55–120. [23] M.S. Moreno, R.F. Egerton, J.J. Rehr, P.A. Midgley, Phys. Rev. B 71 (2005) 035103- [11] S. Penner, H. Lorenz, W. Jochum, M. Stöger-Pollach, D. Wang, C. Rameshan, B. 1–035103-6. Klötzer, Appl. Catal. A 358 (2009) 193–202 (and references therein). [24] R.J. Meyer, Gmelins Handbuch der Anorganischen Chemie, 8. Auﬂage, 1958, [12] H. Lorenz, W. Jochum, B. Klötzer, M. Stöger-Pollach, S. Schwarz, K. Pfaller, S. Verlag Chemie, GmbH Weinheim/Bergstrasse. Penner, Appl. Catal. A 347 (2008) 34–42 (and references therein). [25] S. Penner, B. Jenewein, H. Gabasch, B. Klötzer, D. Wang, A. Knop-Gericke, R. [13] C. Jing, J. Hou, Y. Zhang, J. Cryst. Growth 310 (2008) 391–396. Schlögl, K. Hayek, J. Catal. 241 (2006) 14–19. [14] V.V. Atuchin, T.A. Gavrilova, S.A. Gromilov, V.G. Kostrovsky, L.D. Pokrovsky, I.B. [26] Y. Su, X. Liang, S. Li, Y. Chen, Q. Zhou, S. Yin, X. Meng, M. Kong, Mater. Lett. 62 Troitskaia, R.S. Vemuri, G. Carbajal-Franco, C.V. Ramana, Cryst. Growth Des. 9 (2008) 1010–1013. (2008) 1829–1832. [27] M. Ardyanin, H. Rinnert, M. Vergnat, J. Lumin. 129 (2009) 729–733. [15] H. Kim, S. Shim, J. Lee, Appl. Surf. Sci. 253 (2007) 7207–7210. [28] M. Ardyanin, H. Rinnert, X. Devaux, M. Vergnat, Appl. Phys. Lett. 89 (2006) [16] F. Izumi, J. Solid State Chem. 38 (1981) 381–385. 011902-1–011902-3. [17] X. Feng, J. Ma, F. Yang, F. Li, C. Luan, Mater. Lett. 62 (2008) 1809–1811. [29] N. Jiang, J. Spence, J. Non-Cryst. Solids 353 (2007) 2813–2816. [18] J.A. Lee, G.V. Raynor, Proc. Phys. Soc., Lond. 67 (1954) 737–747. [30] Q. Zhao, H. Lorenz, S. Turner, O.I. Lebedev, G. van Tendeloo, C. Rameshan, B. [19] V.M. Jimenez, J.P. Espinos, A.R. Gonzalez-Elipe, Surf. Sci. 366 (1996) 556–563. Klötzer, S. Penner, Appl. Catal. A 375 (2010) 188–195. [20] W.E. Boggs, P.S. Trozzo, G.E. Pelissier, J. Electrochem. Soc. 108 (1961) 124–129.