Pt(111): A Complete Surface


Reactivity of Fe Nanoparticles on TiOx/Pt(111): A Complete Surface...

0 downloads 132 Views 1MB Size

ARTICLE pubs.acs.org/JPCC

Reactivity of Fe Nanoparticles on TiOx/Pt(111): A Complete Surface Science Investigation Luca Artiglia,† Emanuele Cavaliere,‡,§ Alessio Vascon,† Federica Bondino,§ Gian Andrea Rizzi,† Luca Gavioli,*,‡,§ and Gaetano Granozzi† †

Department of Chemical Sciences, University of Padova, via Marzolo 1, I-35131 Padova, Italy Dipartimento di Matematica e Fisica, Universita Cattolica del Sacro Cuore, via dei Musei 41, I-25121 Brescia, Italy § Istituto Officina dei Materiali—CNR, Laboratorio TASC, Area Science Park, Basovizza, Strada Statale 14, Km.163.5 I-34149 Trieste, Italy ‡

ABSTRACT: This work presents a complete surface science investigation of a model system obtained depositing different amounts of Fe on the z0 -TiOx/Pt(111) ultrathin (UT) film, whose structure is known in great detail. The system has been investigated both at room temperature (RT) and after thermal treatments in an ultrahigh vacuum at temperatures in the room temperature to 900 K range. In contrast with standard thermodynamic predictions, we show that Fe nanoparticles (NPs) strongly compete with Ti to bind oxygen, i.e., a redox reaction where Fe oxidizes and Ti is further reduced to an extent proportional to the amount of deposited Fe is observed. The z0 -TiOx UT film is first destabilized by the presence of Fe, but as soon as the temperature is raised, so activating an interdiffusion of Fe into the Pt substrate, a rather ordered UT TiOx film is formed again. However, a new hexagonal (h-TiOx) phase replaces the z0 -TiOx one in the room temperature to 800 K range, which progressively transforms into the most stable z0 -TiOx form at the highest temperature (900 K). At the intermediate temperatures, the system is present in the form of FeOx/TiO2 mixed oxide NPs. This is a paradigmatic example where the nanoscale effects produce unexpected transformations different from those observed in the bulk.

1. INTRODUCTION Iron and iron oxide (FeOx) nanoparticles (NPs) have several applications in fields like biomedicine,1,2 environmental sciences,3 catalysis,4 and microfluidics.5 In many cases, the NPs are dispersed on a suitable oxide support (OS) that can play a relevant role on the overall system behavior (e.g., catalytic activity).4,6 In fact NPs/OS interactions can have an effect on several phenomena such as charge transfer, encapsulation, spillover and reverse spillover, and NP coalescence.6 These effects can be efficiently revealed by a rigorous surface science approach applied to metal/OS model system based on ultrathin (UT) oxide films.7 For noble metals such as Pd and Pt, the NPs anchored to a reducible OS present effects that have been attributed to the strong metalsupport interaction (SMSI).810 For reactive transition metals such as Fe and Co, the system behavior is still a matter of debate, since the actual processes strongly depend on the balance between the oxide surface energies and the metal work function and can lead both to oxidation or encapsulation of the metal NPs.1013 Moreover, the presence of defects and the structure stiffness of the oxide support play a relevant role in the behavior and the stability of the entire system.14,15 In this framework, ultrathin (UT) TiO x films grown on Pt(111) have been extensively investigated during the last 5 years by our group. They are model systems where elementary steps of phenomena of relevance in catalysis can be understood.14 r 2011 American Chemical Society

The overall research strategy passed through different subsequent stages: • Stage 1: Obtaining a set of well-characterized UT TiOx films with a variable stoichiometry and defectivity16,17 • Stage 2: Using such films as template for metal nanoparticle (NP) growth15 • Stage 3: Using such films as a playground where to study dynamical processes, such as cluster mobility and cooperative transformations18 When the UT film growth is carried out by reactive evaporation of Ti in a low oxygen partial pressure followed by annealing in an ultrahigh vacuum (UHV),17 the resulting films are limited to a double layer of Ti and oxygen atoms, where Ti is at the interface with Pt and oxygen forms the outer layer. The UT film itself can be seen as a sort of very thin “barrier” between the metallic substrate and metal NPs that can be evaporated on it, since it can prevent deposited metal to alloy with Pt and at the same time can work as a template to obtain ordered arrays of metal NPs.15 Among the different TiOx UT film, the one labeled as z0 -TiOx (where z stands for zigzag-like because of its peculiar contrast seen by scanning tunneling microscopy, STM, see below) has been thoroughly studied because it represents an effective a Received: April 8, 2011 Revised: June 7, 2011 Published: July 13, 2011 15812

dx.doi.org/10.1021/jp203281g | J. Phys. Chem. C 2011, 115, 15812–15821

The Journal of Physical Chemistry C template for metal NPs growth: in particular it was possible to obtain ordered arrays of Au NPs that reveal an epitaxial alignment with the Pt(111) substrate.19,20 If the deposited metal has a lower redox potential, e.g., iron, the oxide UT film can be chemically reduced upon deposition of the metal. In this case many questions need to be answered. Can these UT TiOx films still work as barrier layer and template? What is the effect of nanodimensionality, both of the UT film and of the deposited metal NPs ? How do the defects of these films influence the final product? As an example of a study pertaining to stage 3 listed above, we report here an extensive surface science experimental investigation on the Fe/z0 -TiOx/Pt(111) model system aiming at answering the above-reported questions. We have been adopting complementary structural and spectroscopic techniques such as synchrotron radiation (SR) X-ray photoelectron spectroscopy (XPS), STM, low-energy electron diffraction (LEED), and thermal programmed desorption (TPD). Here we investigate the structure and the reactivity of the system prepared by depositing different amounts of Fe. It is to be outlined that, according to standard thermodynamics,10 Fe should not be capable of reducing an already reduced TiOx film.21 The reported data represent an example where bulk thermodynamic is not able to predict reactivity in nanosystems. Our investigation provides important information on the morphology of the system, the Fe-TiOx interaction and its thermal evolution and allows us to discuss some phenomena which can be also of relevance in real catalysis applications. This is a full report following a letter14 where a subset of the data reported here was preliminarily communicated.

2. EXPERIMENTAL SECTION The experimental data were collected in three different UHV (base pressure 600 K) also for the Au/z0 TiOx system:20,37 such phase transformation has been described by a shift of 1/2 period of the rectangular unit cell of the z0 -TiOx phase, which leads to a merging of the defective picoholes into a single larger one.39 Such new enlarged defects can be considered as a site where the Fe-NPs remain pinned. A detailed description of such a transformation in the Fe/z0 -TiOx system has been recently obtained on the basis of DFT calculations, and a specific paper will be devoted to such topic.38 It is important to note that the z0 -TiOx f h-TiOx transformation is definitely driven by the presence of the Fe deposit, since the z0 -TiOx phase alone does not show any modification upon UHV annealing in a temperature range up to 970 K. At temperatures higher than 710 K, such h-TiOx arrangement progressively disappears, and that of the z0 -TiOx reappears. This is well described in the high resolution STM image reported in Figure 5g, where the copresence of both the h-TiOx and the z0 -TiOx phases is clearly observed after an annealing at 900 K, since the stripes typical of z0 -TiOx and the hexagons typical of h-TiOx are both well evident. Considering the fact that the nucleation and growth of the Fe-NPs already induces some relevant perturbation of the UT TiOx layer at RT (see section 3.1), the after-annealing STM images and their statistical analysis can be interpreted as follows. In the 410710 K temperature range the UT oxide layer modifies the z0 -TiOx phase organization in favor of an hexagonal ordering, as 15818

dx.doi.org/10.1021/jp203281g |J. Phys. Chem. C 2011, 115, 15812–15821

The Journal of Physical Chemistry C

ARTICLE

Figure 8. (a) XPS data of the Fe/z0 -TiOx systems at different coverages (0.3, 0.5, 0.8 ML) in the spectral region of Ti 2p at RT (blue) and after a UHV annealing at 700 (green) and 900 K (red) for 20 . (b) Difference spectra obtained by subtracting the area-normalized spectra at the different temperatures are reported.

also pointed out by the corresponding rearrangement of the Fe-NPs, that likely remain pinned in the transformed picoholes associated to the h-TiOx arrangement. A concomitant modification of the particle shape from rounded to triangular is also observed. Thence, it is apparent that a relationship exists between the structure of the underlying UT film and the spatial arrangement and shapes of the Fe-NPs. The major differences occur after the annealing at the highest temperature (900 K), where a clear change of all the measured parameters takes place, and corresponds to the partial appearance in the oxide layer of the troughs typical of the z0 -TiOx phase. The recurrence of the typical pattern of the z0 -TiOx phase is to be associated to the intrinsic thermodynamic stability of such a phase and to the progressive migration of the Fe deposit into the substrate at high temperatures. This latter point is well in tune with the XPS data discussed in section 3.2.2. To sharpen our description of the mentioned transformations, we now discuss the coverage dependent behavior of the system annealed at 900 K, by including STM data (parts ac of Figure 7) of the Fe/z0 -TiOx/Pt(111) system at 0.3, 0.5, and 0.8 ML, respectively, taken with the same tipsurface biases and currents after the thermal treatment at 900 K for 20 . The data after such high temperature annealing clearly show the structure of the TiOx UT film. In Figure 7d the histogram of the apparent heights, representing the frequency of each height obtained from the data, is displayed. The zero of the horizontal scale corresponds to the main peak due to the substrate, while the other peaks indicate a preferred height occurring in the image. At 0.3 ML the UT oxide layer shows two different types of reconstructions. Part of the substrate is still organized in an hexagonal symmetry, while the majority of the surface presents the troughs typical of the z0 -TiOx phase. The apparent height of the islands measured with respect to the UT oxide layer is 0.18 nm, as shown by the peak in the black curve of Figure 7d. The distribution of the island area (Figure 7e) suggests that there could be a preferred value corresponding to about 2.6 nm2. At 0.5 ML, the number of the island is increased with respect to 0.3 ML coverage, and the substrate is partly organized in the hexagonal lattice, with some troughs related to the z0 -TiOx phase observed amid the islands. Note that at this coverage the islands show a height distribution with two maxima, at 0.18 and 0.36 nm, as shown by the presence of two peaks in the

red curve of Figure 7d. Figure 7e shows that at this coverage the island area presents preferred value reduced to about 1.5 nm2. At 0.8 ML the substrate is less visible but we might suppose that the order should still follow the hexagonal arrangement with the pitch dictated by NP separation D. The islands height distribution indicates that there is still a bimodal distribution, very similar to that at 0.5 ML Fe coverage. The island area distribution indicates an increased dimension of the islands, but with a much less defined preferred area. These data indicate that there is a relationship between the oxide UT film structure and the amount of deposited Fe, at least up to 0.5 ML. When the coverage is low (0.3 ML), the thermal treatment at high temperature gives rise to a low Fe-NPs density, allowing the substrate to rebuild predominantly the z0 -TiOx reconstruction. At an intermediate coverage (0.5 ML) the islands act as pinning centers that likely do not allow the TiOx to reconstruct in the most stable form (the z0 -TiOx one) but tend to maintain the substrate in the hexagonal arrangement already present at lower temperatures. At high coverage (0.8 ML), the Fe-NPs increase mainly in size, suggesting that only a few more Fe-NP nucleation centers are present within the oxide layer and that the Fe in excess is diffusing on the surface to increase the island size. 3.2.2. Chemical Transformations: XPS Data. The chemical changes as a function of the coverage (0.3, 0.5, and 0.8 ML) and the annealing temperature have also been studied by XPS. In Figure 8 we report the most relevant Ti 2p core levels photoemission data taken at RT and after a UHV annealing at 700 and 900 K for 20 . In a preliminary letter, the raw XPS data of the Fe(0.5 ML)/z0 TiOx system as a function of the temperature were already discussed:14 a rather peculiar behavior as a function of the UHV annealing temperature was described, and a first simplified interpretation was drawn. The main points emerged from the preliminary analysis were: • A thermodynamic and kinetic facile interdiffusion of the Fe into the Pt substrate was put in evidence. The thermodynamic factor was associated to the high tendency of Fe to alloy with Pt.33 Also, the role of the picoholes of the UT oxide film to facilitate the kinetic transfer through the UT film was outlined. 15819

dx.doi.org/10.1021/jp203281g |J. Phys. Chem. C 2011, 115, 15812–15821

The Journal of Physical Chemistry C • After an initial reduction at RT of the z0 -TiOx phase operated by the Fe deposit, at intermediate temperatures a bicomponent material was formed where mixed FeOx/TiO2 NPs are coexisting with a restructured and reduced TiOx UT film. Herein we want to expand the data set to the two further coverages and make a deeper analysis of the XPS data, providing a comprehensive discussion of the whole matter. Let us focus on Ti 2p data reported in Figure 8, where we display the original XPS data at different coverage and temperatures (Figure 8a) together with the difference spectra obtained by subtracting the area-normalized Ti 2p spectra obtained at the different temperatures (Figure 8b). As for the previous case of Fe (Figure 3), the subtraction procedure easily allows to track the spectral changes originated by the annealing. Inspecting the difference spectra between 700 K and RT, it is well apparent for the various Fe coverage that after the UHV annealing from RT to 700 K a large transformation is occurring in the UT oxide film: the lowest BE Ti component, generated at RT by the redox reaction with the deposited Fe and assigned to the PtxTi alloys (see also parts a and c of Figure 3) is strongly reduced, with the concomitant formation of a higher BE peak to be associated with the OTiO stacking sequence. The important point is that such a phenomenon is more evident the higher the Fe coverage is, i.e., the higher the NP density is (see Figure 7), so unambiguously demonstrating that the amount of Fe influences the abundance of the OTiO stacking component, i.e., that FeOx/TiO2 mixed NPs are actually formed. Recent DFT calculations seems to indicate that the formation of a mixed Fe/Ti oxide with approximately stoichiometry FeTiO3 (ilmenite) can be envisioned.40 It is also important to note that the formation of such mixed oxide is mainly based on the retrieval of the Ti alloyed with the substrate, and actually the low BE component almost vanishes at 700 K. However, what is really peculiar is that the stability of such mixed oxide is restricted to an intermediate temperature range. Actually, when the different 900700 K spectrum is analyzed, it is apparent that the higher BE peak associated with the OTiO stacking sequence is completely quenched and that the Ti 2p XPS data (including the Ti 2p fwhm) almost reproduce the ones of the clean z0 -TiOx UT film, in line with the reappearance of the z0 -TiOx motif observed by STM. In addition, the data reported in Figure 8 show that the Ti 2p intensity attenuation resulting from the Fe deposition (proportional to the Fe coverage) at RT is progressively removed as the temperature is raised, indicating that a large part of the deposited Fe has diffused into the Pt support,39 as already pointed out by TPD.14 However, an accurate analysis of the Ti 2p intensity as a function of both the temperature and the Fe coverage shows that at the lowest analyzed Fe coverage (0.3 ML) the intensity of the Ti 2p peaks after thermal treatment at 900 K is lowered by 30% with respect to the same system at RT. This evidence is compatible with an interdiffusion of Ti atoms into the Pt substrate at high temperature. In other words, when a large quantity of Fe is deposited on the TiOx UT film (0.5 or 0.8 ML), the temperature induced diffusion of Ti results to be inhibited. Thence, Ti and Fe are competing not only for the bonding with oxygen (as already pointed out) but also for the interdiffusion into the Pt substrate.

4. CONCLUSIONS By use of a complementary set of surface science tools we have investigated the system obtained when Fe is evaporated on the z0 -TiOx/Pt(111) UT film at 3 different coverage, i.e. 0.3, 0.5, and

ARTICLE

0.8 ML. Since the structure of the z0 -TiOx UT film was known in great detail, such an investigation allowed us to trace with accuracy the transformations occurring in such a model system. The system evolution have been investigated both at RT and after thermal treatment in UHV up to 900 K. At RT Fe-NPs tend to grow with some degree of long-range order because of the already documented templating properties of the z0 -TiOx UT film.15,19,20 However, as a consequence of the high affinity between the Fe and the O of the UT film topmost layer, the templated ordering is less effective than in the case of Au.15 Different from what expected from bulk thermodynamics, the Fe-NPs are oxidized by the reduced underlying TiOx substrate already at RT, as proven by the XPS measurements that clearly show a reduced component (at ca. 455 eV) developing in the Ti 2p spectrum (assigned to a PtxTi alloy), while the spectra corresponding to the Fe 2p present both metallic and oxidized (Fe2+ and Fe3+) components. This effect is related to the nanodimensionality of both the UT film and of the Fe-NPs. When the temperature is raised in UHV, the redox reaction is further boosted, and a composite system is formed where islands of mixed FeOx/TiO2 oxide form at an intermediate temperature range up to 700 K. This is clearly shown by the XPS data where the signature of the OTiO stacking sequence is present. Such an oxidation is assisted by an incipient interdiffusion of Fe into the Pt substrate, with the consequent increase in the O/M(Ti + Fe) ratio. As soon as the Fe interdiffusion is enhanced by temperature raising, a larger amount of oxygen binds back to Ti (according to QMS reading, no appreciable loss of oxygen is observed during the whole process) and some ordering of the UT TiOx film is showing again: a new hexagonal (h-TiOx) phase is observed beside to the z0 -TiOx one in the 700800 K temperature range. During such a process, residual Fe-NPs islands (mostly oxidized) are prevalently pinned at the apex of the hexagons formed by the restructured UT h-TiOx film. At the final examined temperature (900 K), the z0 -TiOx UT film is almost completely re-established indicating its thermodynamic stability.18 As a general outcome of this study we outline that kinetic factors have to be taken into account to explain the observed transformations. In particular, heat-induced mass transport of metal atoms in and out of the substrate bulk are to be considered, as already suggested to clarify coverage dependent phenomena observed during UT film preparation.40 In a sense, we can suggest the picture that the TiOx layer acts as a dynamic nanofilter with respect to the transmission of the deposited Fe into the substrate bulk, which modifies its structure (z0 -TiOx f h-TiOx f z0 -TiOx) depending on the temperature, coverage, and annealing conditions.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected].

’ ACKNOWLEDGMENT This work was supported by the University and Research (MIUR) through the program PRIN 2005 and 2006 and by the University of Padova, through the grant CPDA071781. We thank Alessandro Fortunelli (Pisa) for clarifying discussions on the content of the manuscript. We thank the BACH Beamline staff (Elettra Synchrotron, Trieste) for their technical assistance during the measurements. 15820

dx.doi.org/10.1021/jp203281g |J. Phys. Chem. C 2011, 115, 15812–15821

The Journal of Physical Chemistry C

’ REFERENCES (1) Laurent, S.; Forge, D.; Port, M.; Roch, A.; Robic, C.; Elst, L.-V.; Muller, R.-N. Chem. Rev. 2008, 108, 2064. (2) Gao, J.; Gu, H.; Xu, B. Acc. Chem. Res. 2009, 42, 1097. (3) Zeng, H.; Sun, S. H. Adv. Funct. Mater 2008, 18, 391. (4) Kung, M. C.; Davis, R. J.; Kung, H. H. J. Phys. Chem. C 2007, 111, 11767. (5) Lu, A.-H.; Salabas, E. L.; Schuth, F. Angew. Chem., Int. Ed. 2007, 46, 1222. (6) Freund, H.-J.; Goodman, D.-W. In Handbook of Heterogeneous Catalysis, 2nd ed.; Ertl, G., Kn€ozinger, H., Sch€uth, F.; Weitkamp, J., Eds.; Wiley-VCH Verlagsgesellschaft mbH: Weinheim, 2007; p 1309. (7) Freund, H.-J.; Pacchioni, G. Chem. Soc. Rev. 2008, 37, 2224. (8) Tauster, S.-J.; Fung, S.-C.; Garten, R.-L. J. Am. Chem. Soc. 1978, 100, 170. (9) Tauster, S.-J. Acc. Chem. Res. 1987, 20, 389. (10) Fu, Q.; Wagner, T. Surf. Sci. Rep. 2007, 62, 431. (11) Parteder, G.; Allegretti, F.; Surnev, S.; Netzer, F.-P. Surf. Sci. 2008, 602, 2666. (12) Sun, Y.-N.; Qin, Z.-H.; Lewandowski, M.; Carrasco, E.; Sterrer, M.; Shaikhutdinov, S.; Freund, H.-J. J. Catal. 2009, 266, 359. (13) Lewandowski, M.; Giordano, L.; Goniakowski, J.; Sun, Y.-N.; Qin, Z.-H.; Noguera, C.; Shaikhutdinov, S.; Pacchioni, G.; Freund, H.-J. Angew. Chem., Int. Ed. 2010, 49, 4418. (14) Artiglia, L.; Cavaliere, E.; Rizzi, G. A.; Gavioli, L.; Granozzi, G. J. Phys. Chem. Lett. 2010, 1, 1660. (15) Gavioli, L.; Cavaliere, E.; Agnoli, S.; Barcaro, G.; Fortunelli, A.; Granozzi, G. Prog. Surf. Sci. 2011, 86, 59. (16) Wu, Q. H.; Fortunelli, A.; Granozzi, G. Int. Rev. Phys. Chem. 2009, 28, 517. (17) Sedona, F.; Rizzi, G. A.; Agnoli, S.; Llabres i Xamena, F. X.; Papageorgiou, A.; Ostermann, D.; Sambi, M.; Finetti, P.; Schierbaum, K.; Granozzi, G. J. Phys. Chem. B 2005, 109, 24411. (18) Barcaro, G.; Agnoli, S.; Sedona, F.; Rizzi, G. A.; Fortunelli, A.; Granozzi, G. J. Phys. Chem. C 2009, 113, 5721. (19) Cavaliere, E.; Kholmanov, I.; Gavioli, L.; Sedona, F.; Agnoli, S.; Granozzi, G.; Barcaro, G.; Fortunelli, A. Phys. Chem. Chem. Phys. 2009, 11, 11305. (20) Sedona, F.; Agnoli, S.; Fanetti, M.; Kholmanov, I.; Cavaliere, E.; Gavioli, L.; Granozzi, G. J. Phys. Chem. C 2007, 111, 8024. (21) Sedona, F.; Granozzi, G.; Barcaro, G.; Fortunelli, A. Phys. Rev. B 2008, 77, 115417/1–8. (22) Zangrando, M.; Finazzi, M.; Paolucci, G.; Comelli, G.; Diviacco, B.; Walker, R. P.; Cocco, D.; Parmigiani, F. Rev. Sci. Instrum. 2001, 72, 1313. (23) Barcaro, G.; Fortunelli, A. J. Phys. Chem. A 2009, 113, 14860. (24) Finetti, P.; Sedona, F.; Rizzi, G.-A.; Mick, U.; Sutara, F.; Svec, M.; Matolin, V.; Schierbaum, K.; Granozzi, G. J. Phys. Chem. C 2007, 111, 869. (25) Ringler, S.; Janin, E.; Boutonnet-Kizling, M.; Goethelid, M. Appl. Surf. Sci. 2000, 190, 162. (26) Bhargava, G.; Gouzman, I.; Chun, C. M.; Ramanarayanan, T. A.; Bernasek, S. L. Appl. Surf. Sci. 2007, 253, 4322. (27) Fujii, T.; Alders, D.; Voogt, F. C.; Hibma, T.; Thole, B. T.; Sawatzky, G. A. Surf. Sci. 1996, 366, 579. (28) Fujii, T.; de Groot, F. M. F.; Sawatzky, G. A.; Voogt, F. C.; Hibma, T.; Okada, K. Phys. Rev. B 1999, 59, 3195. (29) Yamashita, T.; Hayes, P. Appl. Surf. Sci. 2008, 254, 2441. (30) Poulin, S.; Franca, R.; Moreau-Belanger, L.; Sacher, E. J. Phys. Chem. C 2010, 114, 10711. (31) Matsumoto, T.; Batzill, M.; Hsieh, S.; Koel, B. Surf. Sci. 2004, 572, 146. (32) Le Moal, S.; Moors, M.; Essen, J. M.; Becker, C.; Wandelt, K. Surf. Sci. 2010, 604, 1637. (33) Jerdev, D.-I.; Koel, B.-E. Surf. Sci. 2002, 513, L391. (34) Cameron, S.-D.; Dwyer, D.-J. Langmuir 1988, 4, 282. (35) Lemire, C.; Meyer, R.; Henrich, V.-E.; Shaikhutdinov, S.; Freund, H. J. Surf. Sci. 2004, 572, 103.

ARTICLE

(36) Sun, Y.-N.; Qin, Z.-H.; Lewandowski, M.; Shaikhutdinov, S.; Freund, H.-J. Surf. Sci. 2009, 603, 3099. (37) Barcaro, G.; Fortunelli, A.; Granozzi, G.; Sedona, F. J. Phys. Chem. C. 2009, 113, 1143. (38) Barcaro G. et al. , in preparation. (39) An experimental proof of Fe diffusion into the Pt substrate has been recently obtained through photoelectron diffraction data on a Fe(0.8 ML)/z0 -TiOx/Pt(111) system. After treatment at 700 K, the Fe 2p signal modulation is that expected for a iron atom in a fcc lattice. (40) Agnoli, S.; Mentes) , T. O.; Ni~ no, M. A.; Locatelli, A.; Granozzi, G. Phys. Chem. Chem. Phys. 2009, 11, 3727.

15821

dx.doi.org/10.1021/jp203281g |J. Phys. Chem. C 2011, 115, 15812–15821