Growth of Gold on a Pinwheel TiO∼1.2 Encapsulation Film Prepared

Nov 24, 2014 - encapsulated by an ordered hexagonal pinwheel TiO∼1.2 ultrathin ... existence of hexagonal pinwheel (or wagon-wheel) patterns of...
0 downloads 0 Views 9MB Size
Article pubs.acs.org/Langmuir

Growth of Gold on a Pinwheel TiO∼1.2 Encapsulation Film Prepared on Rhodium Nanocrystallites R. Gubó,† L. Ó vári,‡ Z. Kónya,†,‡ and A. Berkó*,‡ †

Department of Applied and Environmental Chemistry, University of Szeged, 6720 Szeged, Rerrich Béla tér 1, Hungary MTA-SZTE Reaction Kinetics and Surface Chemistry Research Group, 6720 Szeged, Rerrich Béla tér 1, Hungary



ABSTRACT: Rh nanoparticles of 50−100 nm diameter and 20-40 atomic layer thickness with a (111) flat top facet parallel to the support surface were grown on a TiO2(110) surface via physical vapor deposition (PVD) at room temperature (RT) followed by annealing at 1050 K. These nanoparticles were completely encapsulated by an ordered hexagonal pinwheel TiO∼1.2 ultrathin oxide (w-TiO-UTO) film. STM, XPS, and low energy ion scattering (LEIS) methods were used to characterize the postdeposition of gold and the effects of annealing on the Au/w-TiO-UTO/Rh-particle system. The adlayer exhibits 3D growth and Rh−Au bond formation at 500 K. The 3D Au nanoparticles of 2−3 nm diameter and ∼1 nm height are partially covered by TiOx species at RT and sinter via an Ostwald-ripening in the range of 500−800 K. The adparticles are gradually getting free of TiOx decoration, and at around 900 K they exhibit a double layer height with 2D character. Two different arrangements were found for these Au particles: (i) a compressed Au(111)-(1 × 1) and (ii) a reconstructed Au(111)-(2 × 1), both of them pseudomorphic with the Rh lattice underneath. Above 900 K, the thickness of these 2D particles tends to become a single layer, while they spread out and form a continuous gold layer on the Rh nanoparticles. This behavior indicates a thermally activated replacement of the w-TiO-UTO film by an Au ultrathin layer. The gold layer is stable up to 1000 K, where extended 1D interfaces are formed between gold and w-TiO-UTO layers. Pd,17−19 and Rh.20,21 Moreover, similar pinwheel patterns were also observed for other UTO layers supported on different metals: VOx/Rh(111),22−24 CrO x/Pt(111). 25 The main common characteristics of these layers are the followings: (i) they consist of a bilayer of oxygen and metal ions; (ii) the metal ions are directly bonded to the support metal surface; (iii) the hexagonally arranged UTO metal ions form a lattice of ∼0.32 nm lattice constant, which is completed by a hexagonal supercell of 1.5−3.0 nm; (iv) this supercell is the result of the lattice misfit between the support and the UTO layer combined with an inhomogeneous distribution of the oxygen ions resulting in the pinwheel pattern; (v) in some cases the missing metal ions at the corner points of the supercell or the strong lateral lattice distortion result in the formation of periodic nucleation centers for admetal atoms (template effect). Due to these interesting properties, the study of postdeposition of metals on the UTO films has also received remarkable attention.1−3,7−9,21,26,27 Although in most cases a metal− UTO−metal sandwich structure was formed, the exchange process between the UTO film and the postdeposited metal was also detected for several material compositions: Rh/TiOUTO/Rh(111)21,27 and Co/VO-UTO/Rh(111).5,22 Naturally,

1. INTRODUCTION Recent atomic level studies on the fabrication and application of ultrathin oxide (UTO) films revealed a wide range of facilities for tailoring and control of electronic and morphologic properties of oxide−metal interfaces and metal−oxide−metal sandwiches.1−7 Connected to the extended research on the special properties of graphene and other self-supporting 2D materials, which became a highly explored field in itself, the breakthrough in the field of two-dimensional oxides and their applications is just in front of us.8,9 Although several different (less or more sophisticated) methods were elaborated to produce UTO films, it turned out that the final atomic structure of the interfaces is mainly determined by the specific materials involved in the formation and the thermodynamic properties of the system. For example, quite similar nanoscale patterns were found for TiOx ultrathin layers grown on a Pt(111) single crystal by oxidative deposition of titanium10−13 and on Pt nanoparticles supported and decorated by self-limiting TiOx films formed via a decoration/SMSI (strong metal−support interaction) process on a TiO2(110) surface.14,15 By accumulation of the structural and electronic data obtained on specific UTO layers, it was revealed that these systems also have quite general common characteristics. For example, among strongly reduced phases of TiOx films grown on different metals, the existence of hexagonal pinwheel (or wagon-wheel) patterns of only slight variation is quite a general feature: Au,16 Pt,11,13 © 2014 American Chemical Society

Received: September 22, 2014 Revised: November 12, 2014 Published: November 24, 2014 14545

dx.doi.org/10.1021/la503756c | Langmuir 2014, 30, 14545−14554

Langmuir

Article

conductive sample in order to avoid charging effects during spectroscopy and microscopy measurements. Rh and Au were deposited by a commercial 4-pocket PVD source (Oxford Applied Research) using high-purity (99.95%) Rh and Au. The coverage of Rh and Au is expressed in equivalent monolayers (ML), defined as the surface concentration of Rh(111) and Au(111), respectively (for Rh 1 ML ∼1.60 × 1015 cm−2, while for Au 1 ML ∼1.39 × 1015 cm−2). Special attention was paid to the cross-calibration of metal coverage in the two chambers by AES uptake-curves recorded at 300 K with the same primary energy and with constant pass energy in both cases. In the XPS-LEIS chamber the coverage for both metals was checked also by XPS and a quartz crystal microbalance (QCM), and the estimated values agreed with a precision of 10%. In the STM chamber the volume determined in the STM records of the particles formed on a clean TiO2(110) surface was also used for an in situ estimation of the coverage.

this latter process is temperature dependent, as demonstrated clearly by TDS and AES measurements for the Rh/TiO-UTO/ Rh(111) system.28 Assumingly, the exchange process is preferred for the admetals, which tend to be encapsulated by UTO film. In contrary to rhodium, gold was not found to be encapsulated on TiO2 support;15,29,30 consequently, it seemed to be interesting to investigate the Au/w-TiO-UTO/Rh(111) system and to compare its properties to those found for the Rh/w-TiO-UTO/Rh(111) system.21 In the present work, postdeposition/thermal effects of Au on Rh crystallites supported on the TiO2(110) surface and encapsulated by an ordered TiO∼1.2 pinwheel structure (wTiO-UTO) are investigated by scanning tunneling microscopy (STM), X-ray photoelectron spectroscopy (XPS), and low energy ion scattering spectroscopy (LEIS) methods.

3. RESULTS 3.1. STM Measurements. The actual surface morphology and oxidation state of a TiO2(110) single crystal can sensitively be tuned by Ar+ sputtering and annealing cycles.15 This property is also determined by the oxidation state of the subsurface layers (prior treatments of the probe), which is revealed by the apparent color and transparency of the sample. The experiments presented here were carried out on a bulkterminated TiO2(110)-(1 × 1) surface exhibiting defect features (dot-like and 1D like stripe nanostructures) of low concentration.32 The deposition of ∼1 ML of Rh at room temperature (RT) followed by annealing at higher temperatures for a few minutes results in separated Rh nanoparticles of round or slightly elongated shape, as was described in detail in our former studies.20,33 For an enhanced amount of Rh (∼10 ML) exposed at RT and annealed at 1050 K, a totally different morphology was obtained where Rh particles exhibit a net of stripe-like nanoparticles (Figure 1). It is important to note that

2. EXPERIMENTAL SECTION The experiments were carried out in two separate ultrahigh vacuum (UHV) systems evacuated down to 5 × 10−8 Pa. The first one was equipped with a commercial room temperature (RT) scanning tunneling microscope (WA-Technology), a cylindrical mirror analyzer with a central electron gun (Staib-DESA-100), and a quadrupole mass spectrometer (Balzers-PRISMA). Note that, in this work, Augerelectron spectroscopy (AES) and mass spectrometry (MS) techniques were mainly applied for checking the surface/ad-metal cleanness and the gas phase composition. The second UHV system was equipped with facilities for X-ray photoelectron spectroscopy (XPS) and low energy ion scattering spectroscopy (LEIS). STM images of 256 × 256 pixels were generally recorded in constant current mode at a bias of +1.5 V on the sample and a tunneling current of 0.1 nA. Pt−Ir tips were applied in the course of the experiments. The conditioning of the tip was a relatively simple procedure: several seconds at +3 V and 10 nA. The constant current (cc) morphology images are shown in top-view representation, where brighter areas correspond to higher Z values. Images of enhanced lateral resolution were recorded in constant height (ch) mode, where the local-dependent variation of tunneling current (no feedback) served for construction of the image. In the case of some high resolution images, a gentle FFT treatment was also applied. The X-Y-Z calibration of the STM images was performed by measuring the characteristic morphological parameters of the TiO2(110)-(1 × 1) support (lateral unit cell: 0.296 nm × 0. 650 nm, step height: 0.297 nm). In the other chamber the same (Leybold Heraeus) hemispherical electron energy analyzer was used for XPS and LEIS. An Al Kα X-ray source was applied for XPS, performed with constant pass energy. The binding energy scale was calibrated against the 4f7/2 peak of a thick Au layer, fixed at 84.0 eV. The takeoff angle was 16° with respect to the surface normal. For LEIS, He ions of 800 eV kinetic energy were applied at a low ion flux (0.03 μA/cm2), using a SPECS IQE 12/38 ion source. The incident and detection angles were 50° (with respect to the surface normal), while the scattering angle was 95°. Note that LEIS performed with noble gas ions supplies information almost exclusively about the outermost atomic layer.31 A one sided epi-polished rutile TiO2(110) single crystal of 5 × 5 × 1 mm3 was directly fixed to a Ta filament by an oxide adhesive (Aremco 571), and it was mounted on a transferable sample holder. The probe was indirectly annealed by the current flowing through the Ta filament. The temperature of the probe was measured with a chromel−alumel (K-type) thermocouple stuck to the side of the sample by the same oxide adhesive. The sample cleaning was started by a gradual increase of the temperature up to 1050 K, and it was continued by Ar+ bombardment at 1000 K with stepwisely decreased ion energies of 2, 1.5, and 1 keV and an average current density of 4-6 μA/cm2 for 1−2 h. This procedure not only resulted in the purification of the sample and getting rid of Ca and K contamination but also caused some reduction in the subsurface layer, resulting in a more

Figure 1. (a) STM image (200 × 200 nm2) recorded after the deposition of ∼7 ML of Rh at 320 K and annealing at 1050 K for 10 min. (b and c) High resolution STM images of 20 × 10 nm2 recorded (b) on TiO2 (110) terraces (dark regions of image (a)) in constant current mode and (c) on the top facet of Rh stripes (bright regions) in constant height mode. (d) Line profiles along the traces indicated in image (a). (e) A characteristic STM image (200 × 200 nm2) recorded after the deposition of ∼0.2 ML of Au onto the surface characterized in image (a). 14546

dx.doi.org/10.1021/la503756c | Langmuir 2014, 30, 14545−14554

Langmuir

Article

a lower annealing temperature (max. 950 K) results in the preservation of the continuity of Rh films with a thickness >30 ML.21 In Figure 1(a) a constant current (cc) STM image of 200 × 200 nm2 is shown for a TiO2(110) surface exposed to 7 ML of Rh at RT and annealed subsequently at 1050 K for 10 min. The bright area is the Rh-covered region, the dark area is the Rh-free TiO2(110) surface, as was checked by recording STM images of high lateral resolution in the two different regions (Figure 1(b) and (c)). The cc-image of 20 × 10 nm2 (b) recorded on a bare TiO2(110) region shows bright rows running periodically in the [001] crystallographic orientation of the TiO2(110) support and separated by a distance of 0.65 nm. This morphology is characteristic for bulk a terminated support surface which is decorated by several added bright 1D and dotlike 0D defects, as mentioned above. The constant height chimage of 20 × 10 nm2 (c) recorded on the bright region exhibits a well ordered hexagonal pinwheel structure characteristic for the TiO∼1.2 decoration layer (w-TiO-UTO film) described in detail in our previous works.20,21 It is important to note that, in spite of the more or less perfect cation distribution in a quasi-hexagonal lattice, the exact position of the Ti ions exhibits a large (10−20%) distortion from the positions on an ideal hexagonal lattice.21 The line profiles in Figure 1(d) depict the actual height along the white traces indicated in the large scale image (a). It can be seen that the average height of Rh stripes is approximately 4−5 nm (15-20 atomic layers). The top facet of these nanostructures is atomically flat and essentially parallel to the (110) plane of the support oxide. It can clearly be seen in Figure 1(a) that the edges of the Rh nanostructure follow several distinguished crystallographic orientations of the support, such as the close packed [001] (indicated by an arrow) and two other directions of [01̅1] and [012]. The application of the growth method described above allows obtaining a configuration where both the Rh(111) top facets covered by the w-TiO-UTO layer and the metal-free TiO2(110) surface are easily accessible by the STM tip. Both regions contain sufficiently large (∼20 × 20 nm2) atomically flat terraces, enabling to study the effect of the subsequent metal-deposition at the atomic level. The main advantage of this configuration is that the changes can concurrently be followed on both kinds of regions. As an example, in Figure 1(e) we present a large scale STM image recorded after Au exposure of ∼0.2 ML at room temperature followed by annealing at 900 K. It can clearly be seen that Au nanoparticles with unlike size distribution appear on the two different regions. 3.1.1. Au Postdeposition: Monolayer Coverage. In the subsequent experiments, the thermal stability of Au (∼1 ML) deposited at 500 K on the surface characterized in Figure 1(a) and the effect of annealing at different temperatures (700 K, 800 K, 900 K, 950 K) for 10 min was followed by recording STM cc-images. Characteristic areas of 20 × 20 nm2 are shown for both TiO2(110) terraces and the Rh-supported w-TiOUTO film in Figure 2(a)−(e) and (f)−(j), respectively. The height profiles along the traces indicated in the corresponding images are collected at the bottom of Figure 2. After gold deposition at 500 K, a significant difference was found for the particle distribution in the two cases. For TiO2(110) terraces (a), the average height of the particles (∼0.8 nm) is approximately double that appearing on the w-TiO-UTO film (f). The average diameter of the Au nanoparticles was also higher in the former case (3-4 nm), while the particle density was 2−3 times higher in the latter case (∼5 × 1012 cm−2). This means that the nucleation probability is higher on w-TiO-UTO

Figure 2. STM images (20 × 20 nm2) taken (a−e) on the Rh-free TiO2 (110) terraces and (f−j) on encapsulated Rh top facets (a, f) after the deposition of ∼1 ML Au at 500 K and 10 min of annealing cycles at (b, g) 700 K, (c, h) 800 K, (d, (i) 900 K, and (e, j) 950 K. Adherent line profiles along the traces indicated in the images (a−j) are presented in the bottom panel.

than on TiO2(110). The annealing at 700 and 800 K results in a remarkable increment (150-200%) of the height of the particles in both cases, and this process is also accompanied by some sintering of the particles resulting in a rise of the average particle diameter (Figure 2(b) and (c), (g) and (h); see also the height profiles). The enhancement of particle height is especially significant for TiO2(110) regions where the particle concentration decreases below 1 × 1012 cm−2 (Figure 2(c)). The concurrent monitoring of the w-TiO-UTO film shows that the top facet of the adparticles becomes more flat and the average particle height of ∼1 nm indicates a thickness of 3-4 atomic layers (Figure 2(h)). Thermal treatments at higher temperatures (900 K, 950 K) induce significantly different 14547

dx.doi.org/10.1021/la503756c | Langmuir 2014, 30, 14545−14554

Langmuir

Article

changes in the two regions. Concerning TiO2(110) terraces, the sintering process accelerates and results in the formation of large 3D particles with an average height of 2−3 nm (approximately 10 atomic layers) and a diameter of 5−6 nm (Figure 2(d) and (e)). Note that image (e) depicts a particle of a size typical for all TiO2(110) regions, but around these single particles there are extended empty terrace regions. Concerning the particles on the w-TiO-UTO surface, they tend to spread out and to transform clearly into 2D-like nanoparticles consisting only of a double or single monolayer (Figure 2(i) and (j)). In the latter case the average apparent particle height is only approximately 0.1 nm (this low value is explained below). In this temperature range (900−950 K) a gradual and a significant decrease of the initial amount of gold (∼1 ML) can also be observed on the top facets: ∼0.4 ML at 900 K and ∼0.2 ML at 950 K. It is important to note that the thermal treatments at even higher temperatures (1000 and 1050 K, not shown here) result in a gradual disappearance of the adparticles, which is most probably due to the desorption of gold from the surface. Note that the initial amount of Au on the TiO2(110) terraces is stable up to 950 K: a significant drop can be detected only above 1050 K. As was mentioned above, an extremely low height (around 0.1 nm) was measured for the adparticles on the w-TiO-UTO regions recorded after the annealing at 950 K (Figure 2(j)). For a better understanding of this feature, a thin 2D particle of this type is shown with cc- and ch-imaging in Figure 3(a) and (b), respectively. The line profile measured along X1 on the ccimage exhibits a height of 0.13(±5%) nm (Figure 3(a)). The ch-image of higher lateral resolution shows that the pinwheel structure around the adparticle is almost perfect, while the top terrace of the 2D adparticle is certainly an atomically flat Au(111) surface (Figure 3(b)). This statement is supported by LEIS measurements presented below. It can clearly be seen that there is no decoration layer on top of the Au adparticle, and the ch-image of 3 × 3 nm2 indicates a more or less uniform hexagonal arrangement of the atomic sites (inset, Figure 3(b)). Surprisingly, the lattice constant determined from this image is 0.26(±3%) nm, which fits well to the Rh−Rh crystallographic distance of 0.269 nm (for Au this value is 0.289 nm). Since LEIS results, presented below, suggest that these adparticles mainly consist of Au, it means that these Au clusters are grown in a pseudomorphic way on the Rh(111) surfaces. A precise check of the inserted image in Figure 3(b) suggests the presence of foreign atoms of a few percentage exhibiting a different contrast (lower brightness) indicated by circles. This second component in the Au layer probably consists of Ti atoms alloyed in gold. This assumption is also supported by LEIS measurements presented below and by our previous study on Au/Rh bilayers supported on TiO2(110).34 Note that we used typically a bias of +1.5 V and a tunneling current of 0.1 nA for recording cc-images, but we tested also the variation of the bias on the detectable step height,as is shown in Figure 3(c). It is evident that in our case the step height depends only slightly on the bias applied for imaging and shows a shallow minima of ∼0.14 nm at +1.5 V. This latter value fits well to that measured in Figure 3(a). The thickness of the w-TiO-UTO film itself was determined from images where a discontinuity of the encapsulation layer was found before gold deposition (Figure 3(d) and (e)). This region was imaged in both cc- and chmodes, as is shown in Figure 3(d) and (e). We note that the discontinuity of the encapsulated TiOx ultrathin film is also possible to be purposely created by the tunneling tip.21 The line

Figure 3. Two-dimensional Au gold particle formed on the top facet of an encapsulated Rh stripe after the deposition of ∼1 ML Au at 500 K and annealing at 950 K: STM imaging (a) in cc mode and (b) in ch mode (image size 10 × 10 nm2). The inset in (b) shows a ch-image (3 × 3 nm2) taken on the Au particle. (c) The apparent height of the Au nanoplate in image (a) as a function of the bias used for imaging. A region of a disrupted hw-TiO-UTOF layer (see the text) imaged (d) in cc mode and (e) in ch mode (image size 20 × 20 nm2). The line profiles (X1, X2) along the traces indicated in images (a) and (d) are shown in the middle panel.

profile recorded along X2 indicated in Figure 3(d) is also shown in this figure. It can be seen that the bottom of the dark region is atomically flat and the depth is 0.08(±5%) nm. The ch-image in Figure 3(e) shows clearly that the area around the pit exhibits a defected pinwheel structure. If we take the sum of the height of the adparticles (∼0.14 nm and the depth of the pit (∼0.08 nm) in the disrupted encapsulation layer, the result is ∼0.22 nm, which belongs to the typical value of average atomic layer−layer distance in metals. This fact suggests that the single layer Au particles bond immediately to the Rh(111) top facet of the crystallites, while they are circumvented by a w-TiO-UTO film of smaller apparent thickness. This clear structural setup can be seen by STM only for the surfaces annealed at higher temperatures. The question arises whether the bonding of Au to the Rh layer takes place immediately after the deposition of Au at 500 K or the exchange between Au and w-TiO-UTO is activated only at higher temperatures. It cannot be answered without the evaluation of LEIS and XPS spectroscopy measurements presented below. 3.1.2. Au Postdeposition: Multilayer Coverage. As we have seen in Figure 2 for ∼1 ML coverage, annealing at 900 K results in a significant loss (60%) of the initial amount of gold detectable after the deposition at 500 K on the top facets of the encapsulated Rh nanocrystallites. Moreover, this thermal treatment resulted in 1−2 atomic layer thick adparticles 14548

dx.doi.org/10.1021/la503756c | Langmuir 2014, 30, 14545−14554

Langmuir

Article

3.1.3. Reconstruction of Au Layers. As described above, the 2D Au adparticles formed after annealing at/above 900 K exhibited characteristically a compressed pseudomorph (1 × 1) arrangement on Rh(111) facets and their atomic structure revealed the incorporation of foreign atoms (probably Ti atoms) of a few percentage in their lattice (Figure 3(b)). Beside this simple case, the Au nanoplates also showed a more complex atomic structure, as is depicted in Figure 5. The

which are free of the encapsulation TiO∼1.2 layer (Figure 3). In order to affirm the inferences drawn from the STM results obtained at monolayer initial coverage of Au, we performed similar investigations at higher (∼2−3 ML) initial coverages. The annealing at 900 K (2 min) leads also in this case to a significant loss of Au detectable on the surface; the actual coverage is below 1 ML (Figure 4). The cc-image in Figure

Figure 5. STM images recorded on top of stripe-like Rh nanoparticles showing (2 × 1) reconstructed 2D Au nanoparticles of one or two layers formed after annealing at 950 K: (a) cc-image of 50 × 50 nm2; (b) ch-image of 40 × 40 nm2; (c) interface between w-TiO-UTO and a reconstructed Au monolayer imaged in constant height (ch) mode (image size: 15 × 15 nm2); (d) ball model of a pseudomorph (2 × 1) Au reconstructed monolayer on Rh(111) with different phases.

Figure 4. 2D Au particle of double layers and a single layer formed on the top facet of an encapsulated Rh stripe as an effect of the deposition of ∼2 ML of Au at 500 K and 2 min of annealing (a, b) at 900 K and (c, e) at 950 K: STM imaging (a) in cc mode and (b) in ch mode (image size 20 × 20 nm2). (c) An extended layer and islands of Au on top of a Rh nanostripe imaged in a region of 40 × 40 nm. (d) Height profile recorded along the trace indicated in image (c). (e) STM image (100 × 50 nm2) recorded on the region showing a stripe-like Rh nanoparticle covered partially by an extended Au layer.

constant current image of 50 × 50 nm2 in Figure 5(a) clearly exhibits larger Au 2D particles with wavy terraces in both the first and the second layers. At first glance, this structure is reminiscent of the herringbone feature characteristic of extended Au(111) single crystal surfaces.35 However, the constant height images recorded in similar regions lead to a different conclusion (Figure 5(b, c)). It can be clearly seen that the structure consists of domains exhibiting parallel rows with a characteristic distance of 0.47(±3%) nm (Figure 5 c). Moreover, this row structure breaks periodically after a distance of approximately 5−6 nm and shifts in most cases by a half period of the rows or is rotated by 120 grad (right top of the image). Considering the Rh−Rh distance of 0.269 nm, the row structure can be identified as a 2 × 1 reconstructed Au layer, as is drawn in Figure 5(d). Note that the 1D interface between the w-TiO-UTO and the Au1 layer is quite sharp. Taking into account that the second Au layer (Au2) has just the same structural appearance as the Au1 layer, it seems reasonable to assume that the first Au layer is unreconstructed (1 × 1) under the reconstructed Au2 particle. It is important to note that the unreconstructed and pseudomorphic (1 × 1) layer is twice as dense as the (2 × 1) reconstructed layer. Accordingly, the latter one is a more relaxed epilayer of Au on the Rh(111) support. Nevertheless, in the course of our experiments on this system, we detected mainly unreconstructed (1 × 1) Au particles. We may speculate on the effect of embedding Ti atoms, which make the unreconstructed Au layers more relaxed, contributing to their stabilization in this dense form.

4(a) shows an extended, monolayer thick, hexagonal particle covered partially by a second layer. The ch-image recorded in the same region reveals clearly that the interparticle area is covered by a pinwheel w-TiO-UTO structure, while both the lower and the upper facets of the 2D nanocrystallites are free of the TiO∼1.2 encapsulation layer (Figure 4(b)). As an effect of further annealing at 950 K (2 min), the extension of single layer adcrystallites is enhanced due to the dissolution/diffusion of the second layer (Figure 4(c−e)). If the amount of surface Au is sufficiently high, the outspreading monolayer crystallites could cover the entire surface of the Rh crystallites. Figure 4(e) shows a large Rh crystallite of which the upper facet indicates two separate regions: the right-upper half of it is completely covered by gold while its left-lower half indicates a small Au crystallite surrounded by the w-TiO-UTO film. By imaging a similar top facet configuration with higher resolution (Figure 4(c)), it is clearly seen that the left side of the region is completely covered by a layer which has just the same height (∼0.14 nm) as that of the individual Au crystallite, as shown by the z-profile in Figure 4(d). The change of height level indicates also that the height at the atomic step between the lower and upper level facets of 2D Au crystallites is ∼0.22 nm, which corresponds to the characteristic interlayer distance within a bulk metal (for Au it should be 0.23 nm). 14549

dx.doi.org/10.1021/la503756c | Langmuir 2014, 30, 14545−14554

Langmuir

Article

3.2. XPS and LEIS Spectroscopy Measurements. In the case of lateral averaging techniques such as X-ray photoelectron and low energy ion scattering spectroscopy, the dual character of the Rh/TiO2(110) system used for STM studies above, which consisted of Rh stripes and TiO2(110) patches, would have hindered a clear interpretation of spectra due to the mixing of spectral features derived from the different regions. Accordingly, the measurements described here were performed on a TiO2(110) surface covered completely by a continuous Rh film. As has been shown in our previous work, a Rh layer of ∼30 ML (or thicker) deposited at 300 K retains its continuity during annealing up to ∼950 K and exhibits a closed and ordered encapsulation layer of pinwheel TiO∼1.2 structure.21,34 Figure 6 shows the characteristic XPS spectra recorded for the

curves (iv and v). The most significant changes in the Ti 2p peak shape induced by the annealing are the appearance of a component at 458.9 eV. The former energy position (454.4 eV) is characteristic for a strongly reduced Ti (almost metallic) state, in contrast to the latter position, which indicates the formation of Ti4+ ions, suggesting the complete oxidation of a part of the surface Ti atoms. As was discussed in an our previous work, the strongly reduced Ti states can be identified with Ti atoms alloyed in the 2D Au clusters, while the appearance of the Ti4+ state probably can be explained by the formation of TiO2 nanoclusters.34 Since the 37 ML thick Rh film keeps its continuity during an annealing at 930 K,21,34 the contribution of the TiO2 substrate to the Ti4+ component can be excluded. However, the diffusion of bulk oxygen from the TiO2 crystal to the surface through the continuous rhodium film can proceed at this temperature.34 The alloying of Ti in the Au clusters is also supported by the STM measurements presented above in this work. The inset in Figure 6 shows the ratio of Ti 2p and Au 4f peak areas as a function of the annealing temperature. It can be seen that there is a slow gradual decrease of this ratio in the temperature range of 500− 850 K followed by a significant drop of this value at higher temperatures. This behavior indicates a covering or a replacement of the w-TiO-UTO encapsulation film by Au. Considering the STM results above, namely that the Au 2D clusters bond immediately to the top Rh layer of the Rh nanocrystallites and transform from multilayer to single layer particles of more extended diameter, we can conclude that gold replaces the encapsulation TiO∼1.2 film. In Figure 7 the LEIS results are collected for three different initial Au coverages. In the case of ∼1 ML Au deposition on the pinwheel TiO∼1.2 encapsulation layer formed on Rh multilayers at 930 K, Figure 7(a) and (b) show LEIS spectra and the LEIS signal intensities for Au, Ti, and O. It can be seen that, before Au deposition, the LEIS spectrum shows only Ti and O signals (no measurable Rh signal), which is in good harmony with the STM results that the Rh multilayer is completely encapsulated with an ordered w-TiO-UTO film (bottom spectrum of Figure 7(a)). As an effect of the deposition of Au (∼1 ML) at 500 K, the Ti and O signal intensities attenuate by 10−15% and a new peak characteristic of Au appears (Figure 7(a), (b)). The subsequent annealing cycles (5 min) of the sample up to 650 K do not cause any significant change of the Au, Ti, and O signals; however, above this temperature the Au signal exhibits a gradual and pronounced increment up to 930 K. Concurrently, the Ti and O signals show a moderate decrease above 750 K (Figure 7(b)). Dosing higher amounts (2.5 ML, 4 ML) of Au at 500 K, the annealing cycles result in a qualitatively similar behavior of the characteristic LEIS signal intensities (Figure 7(c), (d)). A small difference is the slight increase of the Ti and O signals in the range of 500−750 K. Comparing the three cases, the following main features can be revealed: (i) the drop in Ti and O LEIS intensities and the enhancement of the Au signal intensity due to the deposition of gold are approximately proportional to its amount; (ii) regarding the drop of the Ti and O signal intensities (in the range of 10−25%), the screening effect of the deposited gold is rather weak, which means that the Au does not form a complete layer on top of the w-TiO-UTO film even after the deposition of 4 ML of gold; (iii) the larger the amount of gold deposited, the higher the temperature at which a significant enhancement of the Au LEIS signal sets in during annealing. Taking into account the STM results presented above, the

Figure 6. Ti 2p region of the XPS spectra recorded on (i) the clean TiO2(110) surface and (ii) the w-TiO-UTO encapsulation layer formed on 37 ML of Rh by 10 min of annealing at 930 K, followed by (iii) Au deposition (2.5 ML) at 500 K and subsequent annealing for 10 min at (iv) 850 and (v) 930 K. Inset: The area ratio of the Ti 2p and Au 4f regions as a function of temperature.

Ti 2p region after different sample treatments. As a reference, curve (i) shows the Ti 2p region registered on the clean TiO2(110) just before the Rh deposition. The main Ti 2p3/2 peak at 458.9 eV is characteristic of the nominal Ti4+ oxidation state of a stoichiometric sample. The Ti 2p doublet was not detectable after the deposition of 37 ML of Rh at 300 K, because the titanium XPS signal of the titania substrate was completely screened by the continuous Rh overlayer (not shown). The Ti 2p3/2 peak appears at a completely different position (455.4 eV) after a subsequent annealing at 930 K for 10 min, as is shown by the spectrum (ii). The binding energy of the latter case indicates a dominating oxidation state of Ti2+, which is characteristic for the pinwheel TiO∼1.2 encapsulation layer.21 The asymmetry of the peak toward higher binding energies may be originated from those Ti sites of the encapsulation layer which are coordinated to more O ions.10 The subsequent deposition of Au (2.5 ML) at 500 K results only in small changes in the Ti 2p peak shape (spectrum (iii)). Annealing at 850 K leads to the appearance of a barely visible shoulder at 454.4 eV, which persists up to 930 K, as shown by 14550

dx.doi.org/10.1021/la503756c | Langmuir 2014, 30, 14545−14554

Langmuir

Article

Figure 7. (a) LEIS spectrum (enc) of the w-TiO-UTO encapsulated surface (37 ML of Rh deposition at 300 K on TiO2(110) followed by 5 min of annealing at 930 K) and further LEIS spectra collected after subsequent deposition of 1 ML of Au at 500 K and after annealing at different temperatures. (b) LEIS intensities obtained during the measurement presented in (a). The first points in (b) correspond to the encapsulated surface, before the deposition of Au. The horizontal green arrow in (b) marks the Au intensity obtained after the deposition of 1 ML of Au on TiO2(110) at 500 K and serves for comparison (see the text). In (c) and (d), the LEIS intensities are displayed for identical measurements with higher initial Au coverages (2.5 ML, 4 ML). In the latter cases the intensities of Au are attenuated by 10.

Figure 8. Scheme of the processes occurring for Au deposition on wTiO-UTO films formed on the top facets of Rh nanoparticles and for thermal activation of the system.

particles with decreasing layer thickness occurred at higher temperatures (Figures 2 and 8).

4. DISCUSSION 4.1. Analysis of Surface Diffusion and Crystallization of Au on the w-TIO-UTO Film. The formation of epitaxial ultrathin oxide layers on top of metal nanoparticles supported on reducible oxides is quite a general phenomenon.29 It was shown in our previous works that, in the case of the Rh/ TiO2(110) system, the growth of the pinwheel TiO ∼1.2 decoration film on the (111) top facets of Rh nanoparticles is well reproducible, providing excellent facility for further studies.20,21 The impinging and stabilization of a metal atom on a Rh(111) facet decorated by a w-TiO-UTO film is a rather complex process, as was deduced from LEIS measurements for Rh21 and in this work for Au (Figure 7). Not all impinging gold atoms are stabilized on top of the oxide layer, but already at relatively low temperatures (500 K) a significant part of them diffuses to the underlying rhodium sublayer. Remember that the w-Ti-UTO lattice is not an ideal hexagonal structure but exhibits strong distortions,21 which provide more open sites for forming Au−Rh bonds. This configuration provides very active nucleation centers, supporting also further exchange between the adatoms and the decoration film. The high complexity of

following processes can be identified (Figure 8): the formation of 3D gold nanoparticles and a definite exchange between a wTiO-UTO film and incoming Au atoms already at low temperatures (500 K). To justify this conclusion, we note that the Au LEIS intensity obtained after the deposition of 1 ML of Au on w-TiO-UTO at 500 K (Figure 7(b)) is only onethird of the LEIS intensity observed after dosing the same amount of Au at the same temperature on a clean TiO2(110) surface. This latter intensity level is indicated by a green arrow in Figure 7(b). The surprisingly small intensity obtained on wTiO-UTO can be originated either by an unusually large gold cluster height on w-TiO-UTO or by the migration of a part of the Au atoms below the TiO∼1.2 layer. Since according to STM measurements the height of the gold clusters stabilized on the encapsulated Rh stripes was actually smaller than the height of those formed on the TiO2(110) patches (Figure 2), it is safe to conclude that the diffusion of Au below the encapsulation layer proceeds already during the deposition at 500 K. As an effect of annealing, an Ostwald ripening was also found at moderate temperatures and a gradual 3D → 2D transformation of the Au 14551

dx.doi.org/10.1021/la503756c | Langmuir 2014, 30, 14545−14554

Langmuir

Article

the energetic balance between the Au monolayer and the encapsulation w-TiO-UTO film adsorbed on a Rh(111) facet is a very sensitive factor, and it may have a decisive role on the properties of titania supported bimetallic Rh−Au catalysts under working conditions. In our previous studies, it was clearly demonstrated that Rh−Au core-shell nanoparticles can be easily formed on titania surfaces at relatively low temperatures, at around 500 K.42 It was also shown that, for sufficiently high Au coverages, the Au cover layer hinders the activation of the TiOx decoration process up to 950 K.34 Our attention in this work focused on the effect of Au on the behavior of a compact decoration layer. The combined STM and LEIS measurements clearly indicated that the Au dispersion is rather high when gold is deposited on top of the encapsulation layer. This implies the formation of a more active gold than in the case of the same amount of gold deposited on a TiO2(110) surface, since smaller gold clusters are generally more active catalytically. In the temperature range of 500−900 K the main tendency is the gradual ordering of the Rh−Au interface area by pushing out the w-TiO-UTO layer. Naturally, the sintering of the particles leads also to the increase of the particle size, but surprisingly, this process is accompanied by a better wetting of the Au particles which form plates of 1-2 atomic layers. This ordering process due to the decrease of surface free energy results in undecorated 2D gold plates. If the coverage of gold is sufficient high, the gold occupies completely the top facet of the Rh nanoparticles. The Ti and O atoms of the oxide film replaced by gold diffuse probably to the perimeter side of the Rh nanostripes, the region which is difficult to check by STM, and finally dissolve into the TiO2 crystal. The annealing treatment generates also a secondary process; namely, the defective sites of the w-TiO-UTO film are healed out, and this more or less perfect encapsulation layer promotes the detachment of diffusing Au atoms and, in this way, accelerates the desorption of gold at around 1050 K.

these seeds results in a quasi 3D growth mechanism. The high concentration of nucleation centers refers clearly to a low activation energy of the exchange. This observation is completely supported by the finding that the exchange can be activated below room temperature.5,28 Comparing the nucleation probability on the clean TiO2(110) terraces and on the top facets covered by w-TiO-UTO, we estimated a higher value by a factor of 2−3 for the latter case. The large variation of the diffusion ability of gold on different Ptsupported TiO-UTO films was recently analyzed in detail both experimentally and theoretically.36 The key factor in the present case is probably the penetration ability of Au atoms to the support metal surface or the site exchange between the admetal and the UTO film. When raising the temperature, the Ostwaldripening process is activated and the Au−Rh interface will be stabilized and the defects in the w-TiO-UTO film will be healed out. Concurrently, another process will also be activated: the Au adparticles form larger and flatter top facets exhibiting no decoration, while their layer thickness decreases to 2 monolayers. The increased wetting of Rh by Au can be understood considering that the surface free energy of gold (∼1.6 J m−2) is much smaller than that of rhodium (∼2.6 J m−2). Note that the surface free energies of TiOx surfaces are in the range of 1.8−2.1 J m−2, which is very close to that of Au.15 The removal of the TiOx layer from Au is in harmony with the low tendency of gold for encapsulation, which in turn can be rationalized with the low surface free energy of gold and probably also by the small interface energy between TiOx and Au (the relatively weak interaction of Ti and Au). Note that, according to previous calculations on alloys, the formation energy per Ti atom for Au4Ti is ∼ −1.6 eV, while for Rh2Ti it is ∼ −2.3 eV, indicating that Au−Ti bonding is weaker than Rh− Ti bonding.37 This configuration on our surface is stable in the temperature range of 900−1000 K. The fine analysis of the XPS curves indicated also several further features. The Au particles contain metallic Ti atoms of a few percentage as a result of alloying between Au and Ti.38,39 The extension of this alloying process as well as the appearance of a second reconstructed Au phase (Figure 5) was rather uncertain, and it was probably caused by the variation in the reduction state of the bulk support or, in other words, the concentration of diffusing Ti atoms/ions in the surface/subsurface region. Mixing of Rh with Au cannot be completely excluded,40 but it is probably negligible in our case on the basis of our LEIS results. The last morphological change of the Au nanoparticles bonded immediately to the Rh(111) layer and circumvented by the wTiO-UTO film is the disappearance of the second layer of gold at around 1000 K, which can be explained by both diffusion of Au atoms from the Au nanoparticles to the TiO2(110) terraces and desorbing of Au atoms immediately from the decorated Rh(111) surface. The measurements also suggest that if the gold concentration is sufficient high, the formation of a more stable complete Au cover layer takes place on the Rh particles (Figure 4). However, the desorption of gold from this layer appears at much lower temperatures (1050−1100 K) than from the Rh(111) single crystal surface (1315 K).41 This behavior can be explained by the activated surface diffusion of gold at the Au/w-TiO-UTO one-dimensional interface and desorption of Au from the w-TiO-UTO film where the binding energy of gold is much smaller. 4.2. Competitive Wetting of Surface UTO and Au Layers on the Rh(111) Facet and Its Importance in Catalyst Preparation. As is described in the paragraph above,

5. CONCLUSIONS The presented STM-XPS-LEIS study contributes to the deeper understanding of the atomic scale structure and processes of core-shell Au−Rh bimetallic catalysts supported on reducible oxides and serves as a first-hand model study for the material systems showing similar properties. In these catalysts the formation of a reduced oxide decoration phase on the oxidesupported metal particles, typical for Rh/TiO2 surfaces, will significantly be modified by the gold-shell layer. It was unambiguously demonstrated in this work that Au atoms deposited on the pinwheel TiO1.2 structure replace the ultrathin oxide layer and undecorated 2D gold islands grow immediately on the rhodium nanoparticles. Moreover, extended 1D interfaces are formed between the gold cover and the remaining TiOx decoration atomic layers, providing an excellent platform for further catalytic model studies.



AUTHOR INFORMATION

Corresponding Author

*Phone: +36 62 544 646. Fax: +36 62 420 678 aberko@chem. u-szeged.hu. Notes

The authors declare no competing financial interest. 14552

dx.doi.org/10.1021/la503756c | Langmuir 2014, 30, 14545−14554

Langmuir



Article

Studies of the Strong Metal−Support Interaction: Surface Structure Identified by STM on Pd Nanoparticles on TiO2(110). J. Catal. 2005, 234, 172−181. (20) Majzik, Z.; Balázs, N.; Berkó, A. Ordered SMSI Decoration Layer on Rh Nanoparticles Grown on TiO2(110) Surface. J. Phys. Chem. C 2011, 115, 9535−9544. (21) Berkó, A.; Gubó, R.; Ó vári, L.; Bugyi, L.; Szenti, I.; Kónya, Z. Interaction of Rh with Rh nanoparticles encapsulated by ordered ultrathin TiO1+x film on TiO2(110) surface. Langmuir 2013, 29 (51), 15868−15877. (22) Parteder, G.; Allegretti, F.; Surnev, S.; Netzer, F. P. Growth of Cobalt on a VO(111) Surface: Template, Surfactant or Encapsulant Role of the Oxide Nanolayer? Surf. Sci. 2008, 602, 2666−2674. (23) Schoiswohl, J.; Sock, M.; Eck, S.; Surnev, S.; Ramsey, M. G.; Netzer, F. P.; Kresse, G. Atomic level growth study of vanadium oxide nanostructures on Rh(111). Phys. Rev. B 2004, 69, 155403−155413. (24) Schoiswohl, J.; Surnev, S.; Sock, M.; Eck, S.; Ramsey, M. G.; Netzer, F. P.; Kresse, G. Reduction of vanadium-oxide monolayer structures. Phys. Rev. B 2005, 71, 165437−165444. (25) Zhang, L. P.; vanEk, J.; Diebold, U. Spatial Self-organization of a Nanoscale Structure on the Pt(111) Surface. Phys. Rev. B 1999, 59 (8), 5837−5846. (26) Bäumer, M.; Freund, H.-J. Metal deposits on well-ordered oxide films. Prog. Surf. Sci. 1999, 61, 127−198. (27) Bugyi, L.; Ó vári, L.; Kónya, Z. The Formation and Stability of Rh Nanostructures on TiO2(110) Surface and TiOx Encapsulation layers. Appl. Surf. Sci. 2013, 280, 60−66. (28) Bugyi, L.; Szenti, I.; Kónya, Z. Promotion and Inhibition Effects of TiOx species on Rh Inverse Model Catalysts. Appl. Surf. Sci. 2014, 313, 432−439. (29) Fu, Q.; Wagner, T. Interaction of nanostructured metal overlayers with oxide surfaces. Surf. Sci. Rep. 2007, 62, 431−498. (30) Labich, S.; Taglauer, E.; Knö zinger, H. Metal-support interactions on rhodium model catalysts. Top. Catal. 2001, 14, 153− 161. (31) Brongersma, H. H.; Draxler, M.; de Ridder, M.; Bauer, P. Surface Composition Analysis by Low-Energy Ion Scattering. Surf. Sci. Rep. 2007, 62, 63−109. (32) Berkó, A.; Magony, A.; Szökő , J. Characterization of Mo Deposited on a TiO2 (110) Surface by Scanning Tunneling Microscopy and Auger Electron Spectroscopy. Langmuir 2005, 21, 4562−4570. (33) Berkó, A.; Ménesi, G.; Solymosi, F. STM study of rhodium deposition on the TiO2(110)-(1 × 2) surface. Surf. Sci. 1997, 372, 202−210. (34) Ó vári, L.; Berkó, A.; Gubó, R.; Rácz, Á .; Kónya, Z. The effect of a Au Cover Layer on the Encapsulation of Rh by TiOx on TiO2(110). J. Phys. Chem. C 2014, 118 (23), 12340−12352. (35) Barth, J. V.; Brune, H.; Ertl, G.; Behm, R. J. Scanning tunneling microscopy observations on the reconstructed Au(111) surface: Atomic structure, long-range superstructure, rotational domains, and surface defects. Phys. Rev. B 1990, 42, 9307−9318. (36) Sedona, F.; Sambi, M.; Artiglia, L.; Rizzi, G. A.; Vittadini, A.; Fortunelli, A.; Granozzi, G. Mobility of Au on TiOx substrates with different Stoichiometry and defectivity. J. Phys. Chem. C 2008, 112, 3187−3190. (37) Curtarolo, S.; Morgan, D.; Ceder, G. Accuracy of Ab Initio Methods in Predicting the Crystal Structures of Metals: A Review of 80 Binary Alloys. Comp. Coupl. Phase Diag. Thermochem. 2005, 29, 163−211. (38) Tsud, N.; Šutara, F.; Matolínová, I.; Veltruská, K.; Dudr, V.; Cháb, V.; Prince, K. C.; Matolín, V. Structure and electronic properties of gold adsorbed on Ti(0001). Appl. Surf. Sci. 2006, 252, 5428−5431. (39) Martinez, W. E.; Gregori, G.; Mates, T. Titanium diffusion in gold thin films. Thin Solid Films 2010, 518, 2585−2591. (40) Chantry, R. L.; Atanasov, I.; Siriwatcharapiboon, W.; Khanal, B. P.; Zubarev, E. R.; Horswell, S. L.; Johnston, R. L.; Li, Z. Y. An atomistic view of the interfatial structures of AuRh and AuPd nanorods. Nanoscale 2013, 5, 7452−7457.

ACKNOWLEDGMENTS This work was supported by the Hungarian Scientific Research Fund (OTKA) through K81660 project. The authors gratefully acknowledge the support of the financial sources of TÁ MOP4.2.2/A-11/1/KONV-2012-0047/Hungary and EU-COST Actions CM1104 & MP0903.



REFERENCES

(1) Gavioli, L.; Cavaliere, E.; Agnoli, S.; Barcaro, G.; Fortunelli, A.; Granozzi, G. Template-assisted Assembly of Transition Metal Nanoparticles on Oxide Ultrathin Films. Prog. Surf. Sci. 2011, 86, 59−81. (2) Freund, H.-J.; Pacchioni, G. Oxide Ultra-thin Films on Metals: New Materials for the Design of Supported Metal Catalysts. Chem. Soc. Rev. 2008, 37, 2224−2242. (3) Nilius, N. Properties of Oxide Thin Films and their Adsorption Behavior Studied by Scanning Tunneling Microscopy and Conductance Spectroscopy. Surf. Sci. Rep. 2009, 64, 595−659. (4) Pacchioni, G.; Freund, H.-J. Electron Transfer at Oxide Surfaces; The MgO Paradigm: from Defects to Ultrathin Films. Chem. Rev. 2013, 113, 4035−4072. (5) Surnev, S.; Fortunelli, A.; Netzer, F. P. Structure-Property Relationship and Chemical Aspects of Oxide-Metal Hybrid Nanostructures. Chem. Rev. 2013, 113, 4314−4372. (6) Wu, Q.-H.; Fortunelli, A.; Granozzi, G. Preparation, characterisation and structure of Ti and Al ultrathin oxide films on Metals. Int. Rev. Phys. Chem. 2009, 28, 517−576. (7) Chen, M. S.; Luo, K.; Kumar, D.; Wallace, W. T.; Yi, C.-W.; Gath, K. K.; Goodman, D. The structure of ordered Au films on TiOx. Surf. Sci. 2007, 601, 632−637. (8) Pacchioni, G. Two-Dimensional Oxides: Multifunctional Materials for Advanced Technologies. Chem.Eur. J. 2012, 18, 10144−10158. (9) Chen, M.; Goodman, D. W. Catalytically active gold on ordered titania supports. Chem. Soc. Rev. 2008, 37, 1860−1870. (10) Barcaro, G.; Agnoli, S.; Sedona, F.; Rizzi, G. A.; Fortunelli, A.; Granozzi, G. Structure of Reduced Ultrathin TiOx Polar Films on Pt(111). J. Phys. Chem. C 2009, 113 (14), 5721−5729. (11) Barcaro, G.; Cavaliere, E.; Artiglia, L.; Sementa, L.; Gavioli, L.; Granozzi, G.; Fortunelli, A. Building Principles and Structural Motifs in TiOx Ultrathin Films on a (111) Substrate. J. Chem. Phys. 2012, 116, 13302−13306. (12) Sedona, F.; Rizzi, G. A.; Agnoli, S.; Llabrés i Xamena, F. X.; Papageorgiou, A.; Ostermann, D.; Sambi, M.; Finetti, P.; Schierbaum, K.; Granozzi, G. Ultrathin TiOx Films on Pt(111): A LEED, XPS, and STM Investigation. J. Phys. Chem. B 2005, 109, 24411−24426. (13) Sedona, F.; Agnoli, S.; Granozzi, G. Ultrathin Wagon-Wheel-like TiOx Phases on Pt(111): a Combined Low-Energy Electron Diffraction and Scanning Tunneling Microscopy Investigation. J. Phys. Chem. B 2006, 110 (31), 15359−15367. (14) Dulub, O.; Hebenstreit, W.; Diebold, U. Imaging Cluster Surfaces with Atomic Resolution: The Strong Metal-Support Interaction State of Pt Supported on TiO2(110). Phys. Rev. Lett. 2000, 84, 3646−3649. (15) Diebold, U. The surface Science of Titanium Dioxide. Surf. Sci. Rep. 2003, 48, 53−229. (16) Wu, C.; Marshall, S. J.; Castell, M. R. Surface Structures of Ultrathin TiOx Films on Au(111). Phys. Chem. C 2011, 115, 8643− 8652. (17) Bennett, R. A.; Pang, C. P.; Perkins, N.; Smith, R. D.; Morrall, P.; Kvon, R. I.; Bowker, M. Surface Structures in the SMSI State: Pd on (1 × 2) Reconstructed TiO2(110). J. Phys. Chem. B 2002, 106, 4688−4 696. (18) Bennett, R. A.; McCavish, R. D. Non-stoichiometric Oxide Surfaces and Ultra-thin Films: Characterisation of TiO2. Top. Catal. 2005, 36, 11−19. (19) Bowker, M.; Stone, P.; Morrall, P.; Smith, R.; Bennett, R.; Perkins, N.; Kvon, R.; Pang, C.; Fourre, E.; Hall, M. Model Catalyst 14553

dx.doi.org/10.1021/la503756c | Langmuir 2014, 30, 14545−14554

Langmuir

Article

(41) Rodriguez, J. A.; Kuhn, M.; Hrbek, J. Repulsive Interactions between Au and S on Mo(110) and Rh(111): An Experimental and Theoretical Study. J. Phys. Chem. 1996, 100, 3799−3808. (42) Ó vári, L.; Berkó, A.; Balázs, N.; Majzik, Z.; Kiss, J. Formation of Rh-Au Core-shell Nanoparticles on TiO2(110) Surface Studied by STM and LEIS. Langmuir 2010, 26 (3), 2167−2175.

14554

dx.doi.org/10.1021/la503756c | Langmuir 2014, 30, 14545−14554