Skip to main content
NCBI home page
ACS AuthorChoice logoLink to ACS AuthorChoice
. 2025 Oct 17;147(43):39846–39859. doi: 10.1021/jacs.5c14353

Pt Particles on a Dynamic TiO2 Support in Near-Ambient Conditions–Disentangling Size, Pressure, and Support Effects

Florian Kraushofer †, Matthias Krinninger †, Marina de la Higuera-Domingo †, Lorenz Falling †, Lukas Strauss †, Sebastian Kaiser †, Mohammad Salehi †, Gaurav Anand †, Virginia Pérez Dieste ‡, Monika Blum §, Barbara A J Lechner †,∥,*
PMCID: PMC12576813  PMID: 41105972

Abstract

Platinum particles on reducible oxides are known to form complex and highly dynamic catalyst systems at elevated pressures and temperatures, often adopting active structures that differ from those found at room temperature and under ultrahigh vacuum (UHV). Here, we study the oxidation and structural evolution of subnanometer Pt clusters and nanoparticles supported on rutile TiO2(110) across an oxygen pressure range from UHV to 0.1 mbar, using near-ambient pressure X-ray photoelectron spectroscopy (NAP-XPS), scanning tunneling microscopy (STM) under UHV and NAP conditions, and low-energy ion scattering (LEIS). Our results reveal distinct differences in oxidation behavior and thermal stability between Pt nanoparticles and clusters, which are further modulated by the support stoichiometry and oxygen pressure. Small Pt clusters become oxidized even at room temperature but are susceptible to accelerated sintering in 0.1 mbar O2 at elevated temperatures. In contrast, well-crystallized Pt nanoparticles on near-stoichiometric TiO2 show weaker oxidation. On a reduced, defective TiO2 support, Pt instead quickly becomes deeply buried by new titania layers, which are formed during support reoxidation. This process appears to result primarily from interactions of the support with the gas phase, unlike the classical, self-limited encapsulation that is induced by the strong metal–support interaction (SMSI). Finally, we address the full complexity of real catalysts in a direct side-by-side comparison of the single-crystalline model system with a Pt-loaded TiO2 powder catalyst (P25). We conclude that the stoichiometry of the model supports must be carefully chosen and controlled to accurately reproduce the expected state of powder supports during redox reactions.

graphic file with name ja5c14353_0009.jpg

graphic file with name ja5c14353_0007.jpg

1. Introduction

Oxide-supported platinum particles feature among the most ubiquitous catalytic systems, most prominently in three-way automotive catalysts, where they are key to oxidizing CO and NO x . Understanding and thereby gaining the ability to improve such catalysts is a common goal, and has led to various attempts at elucidating the oxidation state of these Pt particles under working conditions. Historically, many studies have investigated the (111) surface of crystalline platinum, which is expected to be the dominant facet and still serves as a benchmark to this day. − With the advent of near-ambient pressure (NAP) methods, the “pressure gap” between classical ultrahigh vacuum (UHV)-based surface-science studies and the elevated pressures can now often be closed to a significant degree. To extend the understanding gained from single-crystalline Pt(111) surfaces to oxide-supported Pt particles and further to real powder catalysts presents a major challenge, which is in turn described as the “complexity gap”. , While the presence of other facets, edges and kink sites is still somewhat accessible to simple model studies, it is now widely accepted that the influence of oxide supportsin particular reducible oxide supportsmust not be neglected.

Supports can sometimes participate directly in catalytic processes, providing sites for individual reaction steps or supplying oxygen through Mars-van-Krevelen-type mechanisms. , Charge transfer to the support also affects the particles’ reactivity, which is sometimes described as an “electronic metal–support interaction” (EMSI). , Similarly, the so-called “strong metal–support interaction” (SMSI) can change a nanoparticle’s capacity to adsorb molecules from the gas phase, which typically decreases sharply after the catalyst is heated under reducing conditions. , Here, Pt on rutile titania (Pt/TiO2) is a prototypical example, where this SMSI effect is now understood to be due to encapsulation of the particles by a thin, substoichiometric TiO x layer. − This phenomenon has been investigated in depth for metal nanoparticles on many reducible oxides. − In addition to the classical “encapsulation by SMSI”, other related phenomena have also been reported, most importantly de-encapsulation in a mixed reaction atmosphere and a nonclassical SMSI where a thin stoichiometric TiO2 layer overgrows the Pt in oxidizing atmospheres. ,− A further key aspect in oxide-supported Pt catalysts is that small metal clusters (which we define here as particles smaller than 1 nm in size) often exhibit very different properties from larger nanoparticles. As bulk periodicity is lost at small sizes, so is the band structure, yielding a more molecule-like electronic configuration, which can lead to a very different and sometimes significantly higher reactivity than for more bulk-like, several nm large nanoparticles. , Furthermore, the high surface-to-volume ratio and highly undercoordinated surface sites allow clusters to more easily restructure and incorporate oxygen.

In oxygen-rich conditions, a key question concerns the active state of Pt during an ongoing reaction. Here, controversial reports can be found in the literature. On the one hand, it has been proposed that the most active phase for CO oxidation is an O-covered metallic Pt surface, rather than a surface or bulk oxide. ,, On the other hand, exceptionally high CO oxidation activity has been reported for highly oxidized PtO x species. ,, Due to the large number of experimental parameters (sample complexity, pressure range, Pt particle size, and crystallinity, etc.), it is often difficult to directly compare results between the various studies. This has motivated us to investigate the state of supported Pt particles in oxygen by controlling and varying these parameters individually in a systematic, fundamental investigation.

This paper tackles the influence of an oxidative environment on TiO2-supported Pt clusters and nanoparticles. In this context, it is necessary to briefly review how bare rutile TiO2 interacts with gas phase oxygen, which has been investigated in much detail at low pressures, revealing the dominant bulk defects in reduced TiO2–x are titanium interstitials (Tiint), up to a concentration of x ≈4 × 10–4, or one in 1250 unit cells. With sufficiently high near-surface reduction, the most common rutile (110) facet forms a (1 × 2) surface reconstruction. − In our group, we have previously developed preparation recipes to reproducibly obtain single crystals with a well-defined stoichiometry: We distinguish between a “highly reduced” (HR-TiO2) and a near-stoichiometric or “low-reduced” (LR-TiO2) state. The HR-TiO2 samples are bulk-reduced to a level just before the onset of the (1 × 2) surface reconstruction, i.e. they contain the highest concentration of bulk defects that can be obtained without any long-range structural reorganization. When TiO2 is reoxidized, the excess bulk titanium interstitials diffuse to the surface and react with O2, ultimately forming new TiO2 terraces via several intermediate steps. , It has also been shown previously that this oxidative layer growth is accelerated in the vicinity of Pd nanoparticles, and that these particles can be entirely covered by the newly grown TiO2 layers. In a previous study on Pt/TiO2(110), some of the present authors have found that Pt can also become buried when annealed in low oxygen pressures, though more rigorous control of support stoichiometry in the present study reveals that this behavior mainly depends on the availability of excess Ti interstitials.

In this work, we investigate the stability and evolution of Pt on rutile TiO2(110) as a function of oxygen pressure and particle size, using (NAP-)­STM and NAP-XPS, as well as low-energy ion scattering (LEIS) and near-edge X-ray absorption fine structure (NEXAFS) spectroscopies. Strong platinum oxidation is only observed at near-ambient pressures (0.1 mbar) of oxygen, and we find that small metal clusters interact with oxygen much more readily and more strongly than large nanoparticles, forming highly oxidized species. We also systematically control the support stoichiometry, distinguishing LR-TiO2 and HR-TiO2 samples as introduced previously. We show that this parameter is highly relevant to catalyst deactivation, as oxygen will react directly with HR-TiO2–x , leading to the deep, nonclassical burial of the metal particles in the oxide. In contrast, the well-known classical SMSI effect, i.e. encapsulation by a thin and reduced oxide film, is only observed on sufficiently reduced HR-TiO2 supports, , but not on the LR-TiO2 samples. Similarly, we argue that the oxidative deep burial of Pt particles requires a large number of available Tiint. Based on observed and simulated TiO2 growth rates, we deduce that there has to be a rapid exchange of Tiint with the bulk, with typical diffusion paths of several μm below the surface. In this context, we compare the single-crystalline supports to a Pt-loaded TiO2 powder catalyst (P25) and discuss the validity and limits of “model systems”. Nanoscale TiO2 powders grains have a much smaller bulk Tiint reservoir, which means that the same process quickly leads to a full bulk reduction and reoxidation, thus coupling support stoichiometry to the chemical potential of the gas phase. Model single crystal studies allow disentangling these factors to identify how Pt particles are affected by the support stoichiometry and by the gas phase, respectively.

2. Experimental Methods

Single-crystalline TiO2(110) samples were acquired from SurfaceNet GmbH and subjected to cleaning cycles of sputtering (1 keV Ar+ ions) and annealing. HR-TiO2 and near-stoichiometric LR-TiO2 samples were obtained by automated preparation cycles, using the recipes discussed in detail elsewhere. Briefly, this includes a reoxidation step in 5 × 10–6 mbar O2 at 900 K after each sputtering step for both types of sample to avoid long-term drift of the bulk stoichiometry. The highly reduced samples are then annealed at 1100 K in UHV, while the near-stoichiometric samples are instead annealed at 1100 K in 5 × 10–6 mbar O2, then cooled down while keeping the oxygen chemical potential constant by decreasing the pressure simultaneously with the temperature. Pt clusters and nanoparticles were produced on these supports by physical vapor deposition of 0.05–0.2 monolayers (ML) respectively 7 ML Pt (Goodfellow, 99.95%) and sintering for 5 min at 600 K respectively 30 min at 1000–1200 K (1 K/s heating ramps) in UHV. We define 1 ML as one atom per TiO2(110) surface unit cell, i.e. 5.2 × 1014 cm–2. For comparative experiments with a powder sample, we use the same samples as in electron microscopy studies reported in the literature: , TiO2 (Aeroxide P25, Acros Organics) was impregnated with tetraammineplatinum­(II) nitrate (99.995%, Sigma-Aldrich) dissolved in ultrapure Milli-Q water, resulting in a Pt weight loading of 2%. The sample was calcined at 200 °C in static air for 5 h, then heated under He flow (50 mL/min) in a tubular oven to 700 °C for 1 h (heating rate: 10 K/min). For NAP-XPS measurements, the powder was pressed into pellets and mounted like the single-crystal samples. Note that we cannot entirely rule out a substantial systematic error in the temperature readout for the pellet, with the sample being colder than measured, due to limited thermal conductivity.

(NAP-)­STM measurements were conducted in a UHV system consisting of two chambers with a base pressure of <1 × 10–10 mbar. One of the chambers houses an SPM Aarhus 150 NAP instrument (SPECS). STM was performed in constant-current mode, using electrochemically etched tungsten tips. Samples were sputtered (IQE 11, SPECS) and annealed using an electron-beam heater in the preparation chamber. Pt was deposited from an electron-beam evaporator (EBE-1, SPECS), using a quartz-crystal microbalance (OmniVac) to calibrate the deposition rate. Samples were mounted on stainless steel plates, and their temperature was measured during preparation and during experiments by a type K thermocouple pressed to the back of the sample by a spring. O2 gas was acquired from Westfalen AG (grade 5.0). STM images were corrected in ImageJ by row alignment along the slow-scan direction and subsequent plane subtraction. Apparent heights of Pt clusters were determined by evaluating their maximum apparent height with respect to the median height of the supporting TiO2 terrace. Mean apparent heights were averaged over 130–450 clusters, extracted from multiple STM images on different locations on the sample. Clusters at step edges were not included in the analysis, as they cannot be clearly assigned to one terrace.

The TiO2(110) samples for all NAP-XPS, LEIS and NEXAFS experiments were initially also prepared in the NAP-STM setup to obtain the desired support stoichiometry, then transferred in air to the respective facilities. There, they were cleaned by sputtering and annealing. Cleanliness was checked by XPS survey and C 1s spectra. Pt deposition was performed in situ in each setup, using a FOCUS EFM-3 and a SPECS EBE-4 evaporator, respectively.

LEIS and laboratory-based NAP-XPS measurements were performed in-house (base pressure <5 × 10–10 mbar). Pt coverage was controlled using as-deposited XPS spectra and calibrated against samples from the NAP-STM system. Samples were heated from the back using a laser heater (OsTech DioSource, 976 nm). XPS data were acquired with a PHOIBOS 150 NAP hemispherical analyzer (SPECS) with a 300 μm diameter aperture and a monochromated Al Kα X-ray source (μFOCUS 450 with XR-MC, SPECS). The same analyzer and a scannable ion source (IQE 12, SPECS) were used for LEIS (1025 eV He+ ions, 132.5° scattering angle).

Additional NAP-XPS measurements on supported Pt clusters were conducted at beamline 9.3.2 , of the Advanced Light Source (ALS) at the Lawrence Berkeley National Laboratory and NAP-XPS and NEXAFS measurements comparing Pt-loaded P25 powder samples and Pt nanoparticles on TiO2(110) single crystals were performed at the CIRCE beamline at ALBA Synchrotron. Pt deposition rates were estimated based on evaporator flux current, and coverages were later calibrated as discussed in the Supporting Information. All XPS spectra were measured with a beam energy hν of 650 eV and referenced to the Ti 2p peak. NEXAFS was acquired by measuring the total electron yield on the sample, while varying the beam energy. At the ALS, a pyrolytic boron nitride heater was used for temperature control, and the temperature was monitored with a type K thermocouple mounted at the front side of the single crystals. At ALBA, a resistive button heater integrated in the sample holder was used to heat the sample, while the temperature was measured with a type K thermocouple clamped to the front side of the single crystals.

XPS binding energies were referenced to the Ti 2p peak set to 458.5 eV. Peaks were fitted with KolXPD using Shirley backgrounds. In all cases, the Ti 3s peak (which is close to the Pt 4f and can thus be measured with the same probing depth) as well as its plasmon were fitted with Voigt line shapes before Pt was deposited. These peak shapes, as well as the area and position of the plasmon peak relative to the main Ti 3s peak, were then kept constant. Pt 4f peaks were always fitted with a doublet of fixed 4:3 intensity ratio and 3.35 eV energy separation. Pt clusters and the oxidized components of Pt nanoparticles could be well-described by Voigt line shapes, while Doniach-Sunjic line shapes were used to fit the more bulk-like metallic components of Pt nanoparticles. The peak shape was initially left free to obtain the best fit to the as-sintered spectra, then kept fixed in fits for subsequent experimental conditions.

Pt quantification is based on the area ratios between the Pt 4f and Ti 3s peaks, which were always recorded at the same time, thus requiring no normalization. Spectra in the figures below are normalized for clarity, either to the low-energy background or to the Ti 3s peak area (specified in the figure captions). The Pt 4f to Ti 3s peak area ratios are based on numerical integrals of the raw data, as shown in Figures S3 and S6. Integration assumes a background with the start and end points set at the low and high binding energy side of the respective peaks. To minimize the influence of noise on these backgrounds, background support points were determined as the average of a number of neighboring data points. Shirley backgrounds were used when the high-binding-energy background is higher than the low-binding-energy background by more than the noise level (i.e., for most Pt nanoparticle data); otherwise, we default to linear backgrounds. Error bars were obtained by varying the integration ranges, as well as the averaging window for the background support points. The area of the Ti 3s plasmon, which overlaps the Pt 4f peak, is assumed to be a fixed fraction of the main Ti 3s peak; we therefore subtract it from the Pt 4f/Ti 3s area fraction, using the relative areas of the plasmon and main Ti 3s peaks determined before Pt deposition.

3. Results

We have performed comprehensive experiments controlling three independent parameters: particle size, O2 pressure, and support stoichiometry. To distinguish the individual effects, a systematic comparison of experiments is required, where parameters are varied one at a time. Therefore, in the following, we present four sections where the influence of two different pressure regimes on two Pt particle sizes, namely clusters and nanoparticles, are investigated. The influence of the TiO2(110) support stoichiometry is further discussed throughout. Finally, we address the full complexity of real systems by a direct side-by-side comparison of single-crystals and powder Pt/TiO2 samples and discuss the effect of particle crystallinity.

3.1. Pt Clusters in Low Oxygen Pressures

Figure illustrates the simple case of O2 interacting with Pt clusters on reduced rutile TiO2(110). A typical STM image of the HR-TiO2(110) surface after depositing 0.05 ML Pt at room temperature (RT) and sintering at 600 K to form the clusters is shown in Figure a. The Pt clusters (some of which are marked by white arrows) are scattered on the surface, with a distribution of apparent heights around 4.0 ± 0.9 Å. Based on the cluster density analyzed over multiple STM images and the QCM-calibrated amount of deposited Pt, we can estimate that clusters contain on average ca. 9 Pt atoms. Interestingly, some of the clusters appear to be surrounded by a darker region (green arrows), possibly indicating that the surface is slightly etched during the sintering process. Furthermore, after sintering clusters at 600 K, we always find a high density of point defects (blue arrows), which we assign as the Ti2O3-type precursor species of the (1 × 2) reconstruction. − These defects are commonly observed on rutile TiO2(110) surfaces when reducing or reoxidizing at low temperatures. − They can be understood as a metastable feature formed when surface redox processes proceed faster than local structural rearrangement. In this case, the higher concentration of these defects than on the as-prepared surface suggests that the surface becomes slightly more reduced during the UHV sintering step.

1.

1

Open in a new tab

Pt clusters and oxidative layer growth on TiO2. STM images of HR-TiO2(110) (a) after depositing 0.05 ML Pt at RT and sintering at 600 K, and (b) after annealing at 600 K for 5 min in 3 × 10–6 mbar O2. White, green and blue arrows mark some Pt clusters, dark depressions and bright reduced point defects, respectively. Imaging conditions: RT, UHV, (a) U b = 1.9 V, I t = 0.1 nA, and (b) U b = 1.8 V, I t = 0.1 nA.

We then exposed the same sample to 3 × 10–6 mbar O2 and annealed it at 600 K for 5 min (1 K/s heating ramp). The results in Figure b show a distinct roughening of the surface, in analogy to what we have observed previously for the bare support. The Pt clusters are still visible (white arrows), but are embedded in newly grown TiO2 terraces. This terrace growth is a reoxidation of the reduced support via diffusion of bulk Tiint to the surface and reaction with O2. , Our group’s previous NAP-XPS investigation has shown that Pt does not become oxidized under these conditions, and that the new TiO2 layers completely overgrow the Pt clusters in a matter of minutes. The same growth of new TiO2 terraces is not observed on LR-TiO2.

3.2. Pt Clusters in Near-Ambient Oxygen Pressures

The behavior becomes more complex when moving to near-ambient oxygen pressure, as shown in Figure . It now also becomes important to control the support stoichiometry: We have shown previously that the growth of new TiO2 layers becomes primarily limited by the concentration of Tiint in the bulk support, unlike at low pressure, where growth is limited by the availability of gas-phase O2. , Figure a,e show Pt clusters after Pt deposition and 600 K UHV sintering on LR-TiO2 and HR-TiO2, respectively. We note that the clusters in Figure e are already more sintered compared to those in Figure a and more point defects are observed on the surface, caused by significantly longer annealing at the relevant temperatures during preceding STM experiments (not shown).

2.

2

Open in a new tab

Effect of NAP O2 exposure on Pt clusters supported on (a–d) LR-TiO2 and (e–h) HR-TiO2, respectively. STM images of 0.05 ML Pt deposited on (a) LR-TiO2 and (e) HR-TiO2 show Pt clusters after sintering at 600 K. The same samples after (b) 20 min in 0.1 mbar O2 at 600 K, and (f) 30 min in 0.1 mbar O2 at 600 K, then postannealing at 800 K for 20 min in UHV to recover a sufficiently stable surface for STM. Imaging conditions: RT, UHV, (a) U b = 1.7 V, I t = 0.1 nA, (b) U b = 2.0 V, I t = 0.1 nA, (e) U b = 1.6 V, I t = 0.2 nA, and (f) U b = 3.1 V, I t = 0.1 nA. (c,g) NAP-XPS spectra (hν = 650 eV, normalized to Ti 3s peak) of the Pt 4f and Ti 3s region from a separate experiment with equivalent sample preparation, acquired in 0.1 mbar O2 while heating from room temperature to 600 K. Note that the Pt coverage in the experiment in (c,d) is 0.2 ML. (d,h) Area ratios of the integrated Pt 4f and Ti 3s peaks (blue squares) as well as the temperature (red curve) as a function of time. Colored arrows indicate which spectra in (c,g) correspond to the respective data points.

Figure b,f show both samples after exposing them to 0.1 mbar O2 and annealing at 600 K. On LR-TiO2 (Figure b), the modification of the TiO2 support is essentially the same as was found in the absence of Pt, i.e. a variety of bright point defects appear due to oxidation of Tiint. ,, The Pt clusters on LR-TiO2 are still clearly visible, though with decreased density and increased size, suggesting further sintering during the NAP O2 heating. Based on cluster density, the mean number of Pt atoms per cluster almost doubles, from ≈12 to ≈20. Meanwhile, their mean apparent height also increases slightly, from 3.9 ± 0.9 Å to 5.0 ± 1.4 Å. At this point, it should be noted that due to the more inhomogeneous, defect-rich appearance of the TiO2 support, we likely underestimate the apparent heights after O2 exposure, and possibly also under-count small clusters, which are difficult to distinguish from larger defects. Histograms of cluster height distributions before and after NAP O2 treatment for both the 0.05 ML coverage (Figure a,b) and 0.2 ML coverage (Figure S1) are shown in Figure S2, exhibiting a similar distribution for both coverages after the oxygen annealing step.

The changes upon heating in O2 are much more drastic on HR-TiO2 than on LR-TiO2. As with blank TiO2 supports, the surface roughens significantly, so much so that STM imaging directly after NAP O2 treatment was impossible. Figure f was recorded after postannealing the HR-TiO2 sample in UHV at 800 K for 20 min. Despite the postannealing treatment, the roughness is still significantly higher than is ever observed on an as-prepared sample. As discussed in prior work on the bare support, this is due to rapid reoxidation of the surface, where the rate of layer growth exceeds the rate of structural rearrangement into well-ordered (1 × 1) terraces. Notably, no signs of Pt clusters remain after this treatmentit appears as though they are completely overgrown by the support, which we will further corroborate below.

Corresponding NAP-XPS data were acquired under analogous experimental conditions. Note that in these experiments, the deposited amount of Pt was initially only estimated based on evaporator flux and later determined by SESSA , simulations (see Supporting Information for details), giving 0.2 ML Pt on LR-TiO2 and 0.05 ML Pt on HR-TiO2. To exclude a possible coverage effect, we repeated the STM experiment with 0.2 ML Pt on LR-TiO2, shown in Figure S1, which yields comparable results. Selected NAP-XPS spectra of the Pt 4f and Ti 3s region acquired during the temperature ramp are shown in Figure c,g, and the full data sets (including Ti 2p peaks) are shown in Figure S3. Figure d,h show the area ratios of the integrated Pt 4f and Ti 3s peaks as a function of time throughout the experiment, with the applied temperature ramp superimposed in red. For a different visualization, the same data are plotted as a function of temperature in Figure S4.

The difference between LR-TiO2 and HR-TiO2 is again immediately apparent: While the Pt 4f area only decreases slightly over the entire time period on LR-TiO2, the platinum signal falls to zero in a matter of minutes on HR-TiO2 once the temperature reaches ≈500 K. This indicates that the Pt clusters become completely buried on this support, in good agreement with the STM image in Figure f. On LR-TiO2, the long-term behavior also fits well with STM. The slow decrease in Pt 4f signal after reaching 600 K can be understood as continued sintering in NAP O2, as was also observed in STM (Figures b and S1). Interestingly, the Pt 4f to Ti 3s peak area ratio increases from ≈1.3 to ≈1.6 when the sample is heated above ≈400 K, possibly due to a change in the particle shape. Furthermore, a strongly oxidized Pt species is formed under these conditions: The Pt 4f 5/2 component develops a high binding energy side feature at ≈77.9 eV in Figure c (see the light blue, green and dark blue curves).

To illustrate the Pt oxidation on LR-TiO2 more clearly, Figure a,b shows Ti 2p and Pt 4f spectra from the same experiment, acquired before and after the data shown in Figure c. Directly after sintering (black curves in Figure ), the Ti 2p peak exhibits a shoulder at ≈456.5 eV binding energy characteristic for Ti3+, as indicated by the arrow in Figure a. The Pt 4f data (Figure b) can be fitted well by a Voigt-shape doublet (peak intensity ratio 4:3, energy separation 3.35 eV) and a Shirley background. In the following, all Pt 4f peak positions refer to the Pt 4f 7/2 peak, as the Pt 4f 5/2 component is fully dependent. The gray peak at ≈75.7 eV corresponds to the Ti 3s plasmon, which was first fitted without constraints in spectra without Pt to obtain the peak shape, then linked to the Ti 3s peak at ≈62 eV after Pt was deposited.

3.

3

Open in a new tab

In situ measurements of the oxidation of Pt clusters on TiO2. XPS spectra (hν = 650 eV, normalized to Ti 3s peak) for Pt clusters (a,b) on LR-TiO2 and (d) on HR-TiO2 show the Ti 2p, Pt 4f and Ti 3s regions at room temperature in UHV after sintering at 600 K (black), at room temperature in 0.1 mbar O2 (orange), and at room temperature in UHV after heating to 600 K in O2 (green). (c) In situ NAP-STM image of Pt clusters on HR-TiO2, acquired while exposing the sample shown in Figure e to 0.1 mbar O2 at room temperature, before heating to 600 K in O2. Imaging conditions: RT, 0.1 mbar O2, U b = 1.6 V, I t = 0.3 nA.

Introducing 0.1 mbar O2 at RT (orange curves in Figure ) is sufficient to completely quench the Ti3+ signal, and introduces a broad, oxidized Pt 4f component (light blue) at 72.7 eV. The main Pt 4f peak (dark blue) also shifts from 71.1 to 71.5 eV. Here, all energies are referenced to the Ti 2p peak at 458.5 eV, which highlights the relative band alignment of Pt and the support. Interestingly, in the raw data, the main Pt peak position is constant, and all Ti and O peaks instead shift to lower binding energies when introducing oxygen (see Figure S3). This suggests that the observed shift is linked to a change in the support, likely modified band bending in the n-type TiO2 due to transfer of electrons to oxygen, and not just to O2 interacting with the Pt clusters. The same overall trends are seen in the Pt 4f spectra for Pt on HR-TiO2 (Figure d). However, after annealing at 600 K in 0.1 mbar O2 (green curves in Figure ), the behavior on the two samples diverges. The main Pt component (dark blue) on LR-TiO2 shifts by another 0.9 to 72.4 eV, suggesting significant oxidation. The more oxidized component (light blue), which we tentatively assign as Pt2+, is significantly sharper than the one observed at room temperature, and shifted to 74.8 eV. These shifts are the same in the raw data, i.e. the observed Ti and O peak positions are unaffected by O2 annealing. Likewise, no changes are observed in the Ti 2p peak shape. On the HR-TiO2 support, the Pt intensity is almost entirely lost.

An in situ NAP-STM image of Pt on HR-TiO2, acquired at RT while exposing the sample shown in Figure e to 0.1 mbar O2, is shown in Figure c. In the NAP-XPS experiments, this would correspond to the orange data in Figure d. The Pt clusters appear qualitatively unchanged, but the TiO2 support has a much more diffuse appearance, and the abundant point defects seen in Figure e are less apparent, indicating that oxidation begins even at RT.

3.3. Pt Nanoparticles in Low Oxygen Pressures

Having explored the oxidation and burial of Pt clusters, we now turn to more bulk-like Pt nanoparticles. Again, it is instructive to first consider the case of low oxygen pressures. Figure summarizes how annealing HR-TiO2 at low (10–6 mbar) oxygen pressure affects supported Pt nanoparticles. The STM image in Figure a shows Pt nanoparticles obtained by depositing 7 ML Pt and annealing at 1200 K. As reported previously, Pt nanoparticles on HR-TiO2 always exhibit an SMSI overlayer already after the initial sintering step, which at this particle size has a “pinwheel” structure, ,, as seen in the inset to Figure a. Note that there is some similarity between this structure and the one reported for the Pt(111) surface oxide, but we can definitely exclude a Pt oxide at this stage based on the fact that the Pt particles were only annealed in UHV and given there is no spectroscopic evidence for Pt oxidation. The apparent heights of the Pt particles are typically in the range of 1–2 nm, as seen in the light blue line profile in Figure e.

4.

4

Open in a new tab

Effect of low-pressure O2 exposure on Pt nanoparticles supported on HR-TiO2. (a–d) STM images of Pt nanoparticles on HR-TiO2 (a) sintered at 1200 K in UHV, then annealed (b) once and (c) twice for 15 min each at 600 K in 10–6 mbar O2, and (d) postannealed for 20 min at 800 K in UHV. The inset in (a) shows a high-pass-filtered image from the same region, revealing the overlayer structure. The inset in (d) shows a higher-resolution image of the marked area, also high-pass filtered for better visibility. Imaging conditions: RT, UHV, (a) U b = 1.5 V, I t = 3.3 nA, (b) U b = 1.4 V, I t = 0.6 nA, (c) U b = 1.8 V, I t = 1.1 nA, and (d) U b = 2.7 V, I t = 0.6 nA. (e) Apparent height profiles along the horizontal arrows drawn in (a–d). (f) Schematic illustration of the geometry of (left) encapsulated and (right) buried Pt nanoparticles on HR-TiO2(110) and the probing angles for normal and grazing emission XPS. (g,h) XPS spectra (monochromatic Al Kα, normalized to low-binding-energy background) of the Ti 2p, Pt 4f and Ti 3s regions for (blue) pristine TiO2(110) before Pt deposition, (black) 7 ML Pt sintered at 1000 K in UHV, and (green) after 30 min at 600 K in 10–6 mbar O2. Solid lines are at normal emission, dashed lines at 45° emission angle. (i) Area ratios of integrated Pt 4f and Ti 3s peaks (blue squares) as well as the temperature (red curve) as a function of time, acquired while heating in oxygen under normal emission. (j) LEIS (1025 eV He+ ions, 132.5° scattering angle, normalized to full range) acquired in the same experiment as the XPS data in (g–i).

Annealing the sample at 600 K in 10–6 mbar O2 again leads to the growth of new TiO2 layers, as previously seen in Figure . STM images after 15 and 30 min O2 annealing are shown in Figure b,c, respectively. Interestingly, the growth seems to be initially inhibited directly on top of the Pt particles: In an image taken after only 5 min of O2 annealing, the encapsulation layer is still clearly visible, as shown in Figure S5, while the surrounding support is already changing significantly. After 15 min, in contrast, no more protruding particles are observed and instead we see pits in the TiO2 terraces, with flat regions at the bottom (see Figure b and yellow line profile in Figure e). In the discussion below, NAP-XPS will allow us to assign these pits to encapsulated particles.

After 30 min of O2 annealing, pits are still present, but no flat surfaces could be resolved at their bottom anymore in Figure c, suggesting that Pt is becoming fully buried at this point. Postannealing the sample at 800 K for 20 min in UHV also does not recover the particles, as is shown in Figure d. While some regions resemble the pits in Figure b, closer inspection reveals only the TiO2(110)–(1 × 1) structure at their bottom, shown in the inset to Figure d. Figure f shows a rough schematic of this development: Pt is initially encapsulated by a thin, reduced SMSI layer (dashed red), but becomes deeply buried by new, bulk-stoichiometric TiO2 layers when annealed in oxygen. The growth initially seems to be inhibited directly above the particles, but the stoichiometric TiO2 layers ultimately also come to cover the platinum, as shown in the right schematic of the final buried state.

As mentioned above, spectroscopy supports this picture: In the XPS data shown in Figure h, the Pt 4f peak area is significantly diminished after 30 min O2 annealing (compare the black curve of the as-sintered nanoparticles with the green curve after O2 annealing). This effect is even more pronounced in the more surface-sensitive XPS spectra acquired at 45° emission angle (dashed curves), in line with the schematic picture we draw in Figure f. In the Ti 2p region shown in Figure g, the Ti3+ component after sintering Pt (black) is much stronger than for the pristine surface (blue) due to the reduced encapsulating overlayer. This reduction is largely reversed after annealing in O2, though the normal-emission spectrum (green, solid) still exhibits more Ti3+ than the pristine surface (blue). However, the more surface-sensitive XPS spectra at 45° emission angle (dashed green) reveal that near-surface Ti is already fully reoxidized to the bulk state at this point, suggesting that the additional Ti3+ seen in normal emission is located in the vicinity of the buried Pt particles.

Figure i shows the time evolution of the Pt 4f peak area in proportion to the Ti 3s peak area while annealing in oxygen under normal emission, with the corresponding NAP-XPS spectra shown in Figure S6a. Unlike the clusters in Figure , the Pt signal from the nanoparticles does not begin to decline until several minutes after reaching 600 K. This can easily be understood by comparison with the STM data: New TiO2 layers first need to grow up around the Pt particles, and actual burial of the particles only becomes visible in normal emission geometry after they are already situated in pits. There is no indication of Pt oxidation at any point throughout this measurement seriesthe Pt 4f peak shape remains completely unchanged. However, in the final buried state, an unspecific rising background is observed. This is characteristic for photoelectrons originating at buried species, which have more opportunities to inelastically scatter on the way to the detector, and is most clearly seen in the Pt 4d peak in Figure S7.

Finally, Figure j shows LEIS data acquired in the same experiment, sampling different regions of the sample each time to avoid damage from the ion beam in the region where XPS was acquired. The homogeneity of Pt coverage and reduction state was confirmed by XPS throughout. As expected, the spectrum acquired directly after depositing Pt (pink) is dominated by Pt, and the Pt component is drastically diminished after UHV sintering (black) due to SMSI encapsulation. A background-level LEIS signal is still observed in the region leading up to the Pt peak position. This can be assigned as a reionization background due to He scattering at subsurface Pt. After annealing in oxygen (green line in Figure j), this low Pt signal is completely suppressed, again in good agreement with deeper burial of the particles.

3.4. Pt Nanoparticles in Near-Ambient Oxygen Pressures

Next, we investigate the same Pt nanoparticles while again increasing the O2 pressure. Since we already know from the cluster study that the support growth in 0.1 mbar O2 prevents us from obtaining meaningful images by STM, we now focus on spectroscopy. Figure shows the effect of 0.1 mbar O2 on Pt nanoparticles supported on both LR-TiO2 (Figure a–c) and HR-TiO2 (Figure d–f). As before, 7 ML Pt were deposited on each sample and sintered by annealing at 1000 K in UHV. As we have reported previously, XPS and LEIS both clearly show that the as-sintered particles are encapsulated on HR-TiO2, but not on LR-TiO2: The Ti 2p peak exhibits a strong Ti3+ component on HR-TiO2 (Figure d) but not on LR-TiO2 (Figure a), and the LEIS Pt signal decreases only weakly after sintering on LR-TiO2 (Figure c), but much more strongly on HR-TiO2 (Figure f).

5.

5

Open in a new tab

Effect of NAP O2 exposure on Pt nanoparticles as a function of TiO2 reduction state, with Pt/LR-TiO2 in the top row and Pt/HR-TiO2 in the bottom row. XPS spectra (monochromatic Al Kα, normalized to low-binding-energy background) of (a,d) the Ti 2p, and (b,e) the Pt 4f and Ti 3s regions at room temperature in UHV after sintering at 600 K (black), at room temperature in 0.1 mbar O2 (orange), and at room temperature in UHV after annealing at 600 K in 0.1 mbar O2 (green). The inset in (b) compares Pt 4f peak shapes before and after O2 annealing, with peaks aligned and scaled to match their low-binding-energy edge and the difference spectrum shown in orange. The inset in (e) shows the time evolution of the Pt 4f region while heating in oxygen, with (blue) the area ratios of integrated Pt 4f and Ti 3s peaks and (red) the temperature. (c,f) LEIS (1025 eV He+ ions, 132.5° scattering angle, normalized to full range) acquired in the same experiments as the XPS data.

Similar to the effect of 0.1 mbar O2 on Pt clusters (Figure ), the Pt 4f peak for nanoparticles on LR-TiO2 already shifts substantially relative to the Ti 3s peak when exposed to 0.1 mbar O2 at room temperature (orange curve in Figure b), from 70.6 to 71.2 eV. The effect of annealing at 600 K in O2 is less pronounced than for the clusters, but there are still clear indications of Pt oxidation: Directly comparing the peak shapes before and after oxygen treatment, as shown in the inset to Figure b, reveals a sharp oxidized component at 72.2 eV. The main Pt 4f peak also seems to shift slightly by another 0.1 eV to 71.3 eV. The binding energy difference between the two components is much smaller than for the clusters, which is why we assign this as partial oxidation to Ptδ+.

The LEIS data, shown in Figure c, confirms that while the Pt nanoparticles are oxidized, they are not becoming encapsulated. The Pt peak after oxidation (green) is only marginally lower than the one directly after sintering, and both are significantly higher than what is typically observed for SMSI-encapsulated Pt (compare to the black curve in Figure f). Somewhat unusually, we find that a peak appears at 470 eV kinetic energy, close to the main oxygen peak. We assign this to double-scattering of He+ ions from two oxygen atoms, which results in less overall energy loss due to smaller scattering angles. , The only chemical species that would yield a peak at a similar position with a single scattering event is fluorine, which was never detected in XPS. Similarly, the Pt peak always exhibits a double line shape, which we also assign to double-scattering.

On HR-TiO2, exposure to 0.1 mbar O2 at room temperature strongly modifies the appearance of the Ti 2p peak in the Ti3+ region (compare the black and orange curves in Figure d), likely due to interaction with the reduced SMSI overlayer. However, the Pt 4f peak (Figure e) is affected much less by O2 exposure than on LR-TiO2, exhibiting only a slight broadening and a tiny shift from 70.7 to 70.8 eV. When heating to 600 K, however, the effect is well in line with that observed for Pt clusters (Figure ) and for nanoparticles at low O2 pressure (Figure ). Pt quickly becomes deeply buried in oxidatively grown TiO2 layers, as seen by the rapid decay of the Pt 4f signal, shown in the inset to Figure e. The individual NAP-XPS spectra are shown in Figure S6b. After annealing in oxygen at 600 K for 15 min, no Pt signal is detectable any more, and the Ti 2p peak appears completely bulk-like. LEIS spectra (Figure f) also show the same trends as were already seen at lower pressure (Figure j), with no more detectable Pt after oxygen treatment.

3.5. Single-Crystalline vs Powder Supports

The experiments we have presented thus far illustrate the key advantage of working with the extremely well-defined single crystalline model systems: even small changes in the support oxidation state, in particle size and oxygen pressure can completely change the oxidation behavior of the Pt/TiO2 system, which we can only disentangle in detail by changing one parameter at a time. However, it is equally important to critically evaluate how well the findings from such models can be translated to realistic powder catalysts, where, e.g., the amount of Pt compared to TiO2, the oxidation state and the crystallinity might be rather different. In this section, we directly compare the Pt/TiO2(110) single crystal model system with a pellet of Pt-loaded P25 powderthe same sample for which electron microscopy studies have been reported in the literature , in NAP-XPS and NEXAFS experiments. Figure S8 shows NEXAFS spectra (Ti L2,3 edge) of the powder sample, Pt/LR-TiO2(110) and Pt/HR-TiO2(110). The curves for the two single crystals differ only slightly, confirming that the bulk crystal structure is the same in both cases. The powder sample also shows similar features but in a different intensity ratio, which could indicate an anatase component in this sample.

NAP-XPS data from the Pt-loaded P25 powder are shown in Figure a. Unlike for the single crystal samples, all XPS data had to be acquired at elevated temperatures in gas atmosphere due to poor conductivity of the pressed pellet at RT in UHV. The top (black) spectrum in Figure a was acquired in 1 mbar H2 at 870 K to mimic the reductive UHV annealing step while avoiding charging of the powder sample. Pt appears fully metallic, with a similar peak shape as observed for nanoparticles on the single-crystalline samples. The peak shape of this metallic component was kept fixed for all further fits.

6.

6

Open in a new tab

NAP-XPS spectra (hν = 650 eV, normalized to Ti 3s peak) showing the Pt 4f and Ti 3s region (a) for a pellet of Pt/P25 powder catalyst as used in ref and (b) for Pt on HR-TiO2. In (a), all spectra were acquired at elevated temperatures on the powder catalyst pellet due to poor conductivity at room temperature. The bottom (pink) spectrum was acquired under the same conditions as the middle (green) one, after cooling to room temperature, then heating back up to 900 K in oxygen. In (b), Pt was deposited on an HR-TiO2(110) single crystal, sintered at 600 K (top left, black), then exposed to 0.1 mbar O2 at room temperature (middle, orange) and heated to 600 K in O2. The bottom (green) spectra were acquired after returning to room temperature and UHV. The right panel in (b) shows a second run of O2 annealing, acquired after annealing the same sample at 600 K in UHV again; the dashed gray line is the original as-sintered spectrum (top left, black) for comparison.

After cooling to room temperature, replacing the H2 gas with 0.1 mbar O2 and heating to 900 K (green curve in Figure a), a clear oxidized Pt 4f component appears at 72.6 eV. The main metallic peak shifts from 70.6 to 71.4 eV, placing it very close to the 71.3 eV seen for Pt nanoparticles on LR-TiO2(110) (Figure b). Interestingly, this initial oxidation is not fully stable: After cooling the sample back to room temperature, then heating again to nominally the same conditions as before, platinum appears to be less oxidized (pink curve in Figure a). While the main peak remains at 71.4 eV, the oxidized component is less pronounced, and now appears at 72.3 eV. No further modification of this end state was observed, and we conclude that the original stronger oxidation (green curve) is due to a metastable configuration, which we can also mimic in our model system, as shown below in Figure b. It is worth noting that in the final configuration, the distance between the main peak and the oxidized component is 0.9 eV, in excellent agreement with that observed for Pt nanoparticles on LR-TiO2 (Figure b).

3.6. Effect of Particle Crystallinity

For a direct comparison in the same NAP-XPS instrument as used for Pt-loaded P25 powder, Pt was again deposited onto HR-TiO2. Based on SESSA simulations, we estimate the coverage to be about 3 ML. In this experiment, the Pt/HR-TiO2 sample was sintered at only 600 K, leading to a smaller average particle size than in the nanoparticle experiments presented above. As a result, the oxidation behavior in 0.1 mbar O2 at 600 K shown in Figure b also lies in between that observed for clusters (Figure ) and for nanoparticles (Figure ): Despite the high coverage, Pt becomes strongly oxidized, which was not observed on well-crystallized nanoparticles. Interestingly, the Pt signal is also not completely suppressed even after 30 min in O2 (green curve in first run of Figure b).

It becomes clear that this is due to the less crystalline nature of the original state when following up with another cycle of the same experiment, on the same sample, shown in the right panel of Figure b. Here, the top (black) spectrum was acquired after once more annealing the sample at 600 K in UHV, directly subsequent to the experiment shown in the left panel of Figure b. This results in a fully metallic Pt peak, though significantly lower than the initial as-sintered state (shown again as a dashed gray line for comparison). Repeating the exposure to oxygen and annealing results in much less oxidation than in the first O2 annealing step, presumably due to the particles becoming more well-sintered and crystalline during the first round of oxidation and reduction. This metastable oxide formation is comparable in the HR-TiO2 and the powder sample. In all three samplesPt/P25, Pt/HR-TiO2 and Pt/LR-TiO2(110)the NEXAFS spectra (shown in Figure S9) remain largely unchanged throughout the experiments, indicating that there is no major bulk restructuring. That being said, after finishing the entire round of experiments, the 0.5 mm thick TiO2 crystal had lost most of its original oblique-black color, indicating bulk titania reoxidation.

4. Discussion

Two main themes emerge from all the results presented above: First, the degree to which platinum can be oxidized varies with particle size and oxygen pressure. Second, we describe a new deactivation mechanism, namely the oxidative deep burial of platinum in the support, which is not limited by direct interactions with the particles, in contrast to the shallow, self-limiting encapsulation in the nonclassical SMSI. , This burial is surprisingly fast even at only 600 K, indicating a high rate of bulk diffusion for Ti interstitials. The dependence on the availability of Ti interstitials obviously limits the comparability of powder TiO2 to single crystals, which have a much larger bulk reservoir.

4.1. Size-Dependence of Pt Oxidation

We find substantial differences when we directly compare the oxidation of Pt nanoparticles and subnanometer clusters on LR-TiO2 upon exposure to 0.1 mbar O2. Since the bulk of the larger Pt nanoparticles remains metallic, a much larger fraction of Pt atoms is oxidized for clusters (Figure ) than for nanoparticles (Figure b). In addition, there is also a clear qualitative difference in the degree of oxidation: the peak-to-peak separation between the least and most oxidized components is significantly larger for clusters (2.4 eV) than for nanoparticles (0.9 eV), with the most oxidized component for clusters appearing at 74.8 eV. While peaks are typically shifted to higher binding energy for small clusters due to final-state effects, − making it difficult to directly assign oxidation states, the peak-to-peak separation still clearly indicates the formation of a highly oxidized Pt species, which we assign to Pt2+. The nanoparticles, in contrast, are only oxidized to an intermediate Ptδ+ state: A similar oxidized peak at +0.9 eV with respect to Pt0 was observed in a previous NAP-XPS study on bulk Pt(111) after annealing in 0.5 Torr (0.67 mbar) O2, and was interpreted as a surface oxide. Bulk-like platinum oxide with a highly oxidized Pt 4f component, comparable to that seen here for the clusters, was only formed when annealing Pt(111) to 720 K in 5 Torr (6.67 mbar) O2. Pt can likely achieve a higher oxidation state in clusters than in nanoparticles because the already low-coordinated surface atoms can restructure relatively easily into highly O-coordinated motifs. This is in good agreement with previous studies of small Pt clusters on ceria, where theory predicts the formation of PtO x clusters in oxygen-rich conditions, with all Pt–Pt bonds replaced by Pt–O–Pt. ,, Interestingly, cluster oxidation appears to be facile even at room temperature, with initial Pt oxidation seen in XPS on both LR-TiO2 and HR-TiO2 when Pt clusters are exposed to 0.1 mbar O2 (orange curves in Figure b,d).

The hypothesis that the degree of Pt oxidation is largely dependent on its initial coordination is further supported by the behavior of less well-crystallized Pt nanoparticles, sintered at a lower temperature (Figure b). Here, strongly oxidized components are observed even for very high Pt loadings, more comparable to Pt clusters than to fully crystallized nanoparticles, sintered at 1000 K. We also found a similar behavior with highly oxidized components in previous work, where submonolayer Pt coverages were sintered at 800 K. We interpret this observation as a stronger interaction of oxygen with under-coordinated Pt, which would be present both in clusters and in not yet fully crystallized nanoparticles or films. Cycling oxygen and UHV annealing ultimately reduces the degree to which particles can be oxidized (Figure b), likely because Pt sinters to larger and more crystalline particles in the process. Direct comparison to large P25-supported Pt nanoparticles (Figure a) also supports this view, as these behave much like the well-crystallized Pt nanoparticles on LR-TiO2 (Figure b) and HR-TiO2 after long O2 annealing (Figure b), and do not exhibit any highly oxidized components.

Finally, we also consider the formation of a Ti–Pt–O mixed oxide when small Pt clusters are annealed in NAP O2. In nanoparticles, Ti–Pt alloying has been reported under reducing conditions, , but oxidizing atmospheres typically lead to resegregation of TiO x in a “non-classical SMSI” effect. This is expected based on thermodynamic considerations: The standard enthalpy of formation ΔfH0 of rutile TiO2 (939 kJ/mol) is about three times higher per Ti atom than that of bulk TiPt3 (298 kJ/mol), and we are not aware of any reported Ti–Pt–O bulk oxide phase. Overall, while we cannot fully exclude the possibility of a (transient) mixed phase, we conclude that the oxidation of Pt to PtO x clusters is more likely to proceed without Ti–Pt mixing.

4.2. Oxidative Deactivation of Pt on HR-TiO2

A previous study investigated the CO oxidation activity of Pt clusters on LR-TiO2 and HR-TiO2 at low pressures and came to the conclusion that HR-TiO2 is less active by 2 orders of magnitude because Pt competes for oxygen with the support. Clearly, our results show that Pt on (single-crystalline) HR-TiO2 is a poor oxidation catalyst, due to oxidative TiO2 layer growth leading to rapid burial of Pt particles at all pressures and sizes.

It is plausible that this buried state is not the thermodynamic minimum, but rather a result of kinetic limitations. Notably, TiO2 does not appear to preferentially overgrow the Pt nanoparticles. At low oxygen pressure, the reduced SMSI overlayer remains intact even as the surrounding TiO2 becomes oxidized (Figure S5). This suggests that replacing or overgrowing the encapsulation layer with stoichiometric TiO2 may be energetically costly because of interface strain, which the thin overlayer can more easily compensate than (1 × 1)-periodic, bulk-like TiO2. Only once the particles are situated at the bottom of newly grown pits, which also incur an energy cost, do they become overgrown by bulk-like TiO2 layers. From an energetic point of view, it is likely that placing the particle on top of the TiO2 crystal, rather than as a subsurface inclusion, would be preferred.

Interestingly, the only instance in which Pt on HR-TiO2 is not fully buried after annealing in 0.1 mbar O2 is the partially sintered high Pt coverage, shown in Figure b. As discussed above, Pt in this configuration appears more susceptible to oxidation, indicating that an encapsulation layer is absent or at least incomplete. We can speculate, then, that more oxidized, larger Pt particles can better resist the burial. This is in agreement with our previous study where size-selected Pt10 clusters were deposited intact. The enhanced sintering of Pt clusters in 0.1 mbar O2 (Figure ) implies faster diffusion rates, which may allow at least some fraction of Pt to “float” on the growing TiO2 layers. Such a floating effect would be made significantly more difficult by a complete SMSI encapsulation layer. We conclude that the buried configuration of Pt particles is likely metastable, and arises only because oxidative layer growth on the bare TiO2 is facile compared to Pt particle diffusion.

More generally, we speculate that the same mechanisms as described here for Pt on TiO2 would also lead to burial of metal particles on other reducible oxide supports which exhibit SMSI encapsulation and cation-dominated bulk defect chemistry, such as magnetite (Fe3O4) , as well as other spinel and rutile oxides. It would be interesting to contrast this to substrates where bulk reduction is primarily defined by oxygen vacancy defects, such as ceria. Presumably, no oxidative burial would be observed there, as substrate reoxidation would not involve the growth of new layers.

4.3. TiO2 Reoxidation Rate

The extremely fast burial of Pt in near-ambient pressure oxygen, to the point where it is not detectable by XPS any more, is somewhat surprising. The NIST Electron Effective-Attenuation-Length Database predicts an inelastic mean free path (IMFP) for rutile of 2.8 nm at the kinetic energy of the Pt 4f 7/2 peak (1415 eV with Al Kα excitation). Attenuating the Pt signal below 5% (1%) of its initial value would take 7.3 nm (11.2 nm) of added rutile thickness, requiring a substantial number of Ti atoms to be drawn from the bulk. This is accomplished in 0.1 mbar O2 within 5 min of reaching 600 K on HR-TiO2 (see inset to Figure d).

The depth from which Tiint can reach the surface while annealing depends on the Tiint diffusion barriers one assumes, but may be surprisingly large: With a 0.5 eV barrier, − a simple random-walk model of noninteracting Tiint, discussed in more detail elsewhere, yields a characteristic diffusion length of 0.14 mm for only 5 min at 600 K. A more thorough calculation for the number of Tiint reaching the surface within a given time is given in the Supporting Information. We calculate that up to 22.7 nm of new TiO2 layers may be grown in 5 min at 600 K if every Tiint that randomly diffuses to the surface is oxidized. Conversely, we can use the same model to find an upper limit for the Tiint bulk diffusion barrier, E B = 0.62 eV, above which the observed Pt attenuation could not be explained (see Supporting Information for details).

This diffusion model does not account for any additional barriers in rearranging the newly oxidized TiO2 moieties at the surface into new bulk-like TiO2 terraces, − or a possible higher barrier for diffusion from the subsurface to the surface. It seems likely that these would be rate-limiting for surface layer growth, at least initially. However, the estimates we have made are extremely conservative, and the calculated maximum bulk diffusion barrier of 0.62 eV is still relatively close to the previously reported 0.5 eV. − The possibility thus still remains that bulk Tiint transport limits the layer growth from the beginning even on HR-TiO2, and it should quickly become dominant for more stoichiometric samples. Even with a barrier on the order of 0.6 eV, however, diffusion and bulk reoxidation clearly proceed much faster than intuition might suggest. This has severe implications for the comparison between bulk rutile crystals and TiO2 powder samples.

In a previous study, some of the present authors had observed that submonolayer Pt coverages were buried less quickly at high than at intermediate oxygen pressure. The stoichiometry of the TiO2 samples was not strictly controlled in that work, but based on preparation and sample color, we estimate that they were slightly more reduced than the LR-TiO2 used here, and much less reduced than HR-TiO2. The rate of burial was also compared as a function of oxygen dose, rather than of time. With such a description, the present data would similarly suggest a “faster” burial at low pressure: The Pt signal decreases about an order of magnitude more slowly in 0.1 mbar O2 (Figure ) than in 10–6 mbar O2 (Figure ), but this is easily compensated by the 5 orders of magnitude pressure difference. That being said, the availability of gas-phase oxygen may no longer be rate-limiting to the reoxidation process at 600 K in 0.1 mbar O2, as implied when referring to the oxygen dose: The availability of Tiint, together with possible barriers for surface reorganization, are sufficient to limit the growth rate to the experimentally observed value. We therefore conclude that although reoxidation rates depend linearly on the oxygen pressure in the 10–6–10–7 mbar range, this is not the case at elevated pressure, where Tiint availability instead becomes rate-limiting.

It is interesting to frame this in terms of a “pressure gap”: When the system is in a diffusion-limited regime (i.e., lower Tiint flux than O2 impingement at the surface), the TiO2 growth rate becomes largely insensitive to pressure. However, the crossover from pressure-limited to diffusion-limited kinetics is exactly what causes a pressure gap to appear: The linear relationship between oxygen pressure and TiO2 growth rate, identified by varying p O2 in the UHV-compatible range, cannot be extrapolated indefinitely to higher pressures.

4.4. Validity of Model Systems and Lessons for Powder Catalysts

The deep burial of Pt through oxidative layer growth is clearly not directly transferrable to Pt nanoparticles supported on powder TiO2, simply because there is no comparably large bulk reservoir of Ti3+ interstitials. Highly reduced rutile TiO2–x still only contains a relatively small amount of excess Ti, with x on the order of 4 × 10–4. This implies that the volume of a TiO2 particle can only increase by about 0.04% when moving from highly reduced to a fully oxidized stoichiometry. Due to the extremely fast Tiint diffusion and the large volume of TiO2, this can still amount to hundreds of nanometers on millimeter-sized single crystals. In contrast, TiO2 powder samples have a much smaller TiO2 to Pt volume ratio and will thus rapidly deplete their bulk of excess Tiint, limiting the growth.

The flipside of this difference in bulk reservoir is that we can also expect powders to rapidly change their bulk oxidation state in reducing or oxidizing environmentsunlike single crystals. For example, if we assume TiO2 particles with a radius of 100 nm, then 0.04% growth amounts to only 0.2 Å (!), and would be finished in seconds at 600 K. Surface reaction barriers may limit this rate somewhat. Nonetheless, on a typical experimental time scale, we would expect that there is essentially no delay between changing the chemical potential of the gas environment, and the stoichiometry of a powder TiO2 sample following suit. This has important implications for the electronic interactions of metal particles and their supports, but also for the study of SMSI encapsulation and de-encapsulation: If the TiO2 stoichiometry is strongly coupled to the gas atmosphere, the direct effects of the gas atmosphere on supported metal particles are impossible to disentangle from the effects of changing metal–support interaction.

Single crystal model systems can alleviate this issue. Due to the much larger bulk reservoir, the support stoichiometry can be kept approximately constant throughout a typical experiment, while the gas environment can be switched in a matter of minutes. This allows a distinction between how a change in the gas atmosphere (i) directly affects the Pt particles themselves, and (ii) indirectly affects the particles by modifying the supporting oxide. For example, an oxidizing atmosphere does not appear to modify the reduced encapsulation layer on HR-TiO2 (Figure S5), and even extended UHV annealing at 1100 K does not lead to encapsulation of Pt particles on LR-TiO2. In both cases, we can conclude that the interaction of the particles with an oxidizing or reducing environment is not enough in itself: It takes a modification of the TiO2 support to create or remove an overlayer.

We conclude that model studies on single crystal supports can be extremely helpful for mechanistic understanding, but care must be taken to control all relevant variables. When using single crystals as a stand-in for powder samples, a good rule of thumb is that the single crystal stoichiometry should mirror the one expected for the powder under the given conditions. Here, we find that under oxidizing conditions, Pt nanoparticles on LR-TiO2 behave very similarly to particles on P25, which fits with the assumption that the powder is fully oxidized in 0.1 mbar O2. In contrast, HR-TiO2 is the better model for powder TiO2 under reducing atmospheres. Precisely controlling the stoichiometry of oxide single crystals, rather than allowing it to drift over many sputtering and annealing cycles, is thus essential in order to get reproducible and interpretable results. We predict that this holds not only for TiO2, but for all reducible oxide supports, in particular when they exhibit rapid ion transport.

5. Conclusions

We have presented a systematic study of Pt oxidation on rutile TiO2 where we control the particle size, oxygen pressure, and support stoichiometry. On near-stoichiometric TiO2, both Pt nanoparticles and subnanometer clusters become oxidized in 0.1 mbar O2 at 600 K. Clusters are oxidized even at room temperature, and their oxidation at 600 K is much more pronounced than that of nanoparticles, both in terms of the fraction of oxidized Pt and of the highest oxidation state. In general, the degree of Pt oxidation is largely dependent on its coordination or crystallinity.

On reduced TiO2, we have described the rapid burial of Pt particles by the TiO2 support through oxidative layer growth. The high rate of support reoxidation implies rapid Ti interstitial diffusion. The significantly lower number of Tiint available in nanoscale TiO2 powder grains thus suggests that TiO2 bulk stoichiometry can be modified within minutes for powder samples. Model systems must thus be prepared with well-defined stoichiometries to fit the state of a real catalyst under the respective atmosphere.

Supplementary Material

ja5c14353_si_001.pdf (3.6MB, pdf)

Acknowledgments

We gratefully acknowledge Hannes Frey for providing the powder sample and thank Marc Willinger and Friedrich Esch for fruitful discussions. This work was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under Germany’s Excellence Strategy EXC 2089/1–390776260 and through the project CRC1441 (project number 426888090, subproject A02). This project received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (grant agreement no. 850764). This research used resources of the Advanced Light Source, which is a DOE Office of Science User Facility under contract no. DE-AC02-05CH11231. This research was funded in whole or in part by the Austrian Science Fund (FWF) [FK, grant no. J 4811 N]. For the purpose of open access, the author has applied a CC BY public copyright license to any Author Accepted Manuscript (AAM) version arising from this submission.

Glossary

Abbreviations

HR-TiO2

highly reduced TiO2

LEIS

low-energy ion scattering

LR-TiO2

low-reduced TiO2

ML

monolayer

NAP

near-ambient pressure

NEXAFS

near-edge X-ray absorption fine structure

RT

room temperature

SMSI

strong metal–support interaction

STM

scanning tunneling microscopy

UHV

ultrahigh vacuum

XPS

X-ray photoelectron spectroscopy.

The data is available on: 10.5281/zenodo.17233126

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacs.5c14353.

  • The following files are available free of charge. Description of SESSA simulations, discussion of bulk diffusion lengths during TiO2 reoxidation, supplementary figures (PDF)

⊥.

Institute of Applied Physics, TU Wien, 1040 Vienna, Austria

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

The authors declare no competing financial interest.

References

  1. Engel, T. ; Ertl, G. . Elementary Steps in the Catalytic Oxidation of Carbon Monoxide on Platinum Metals. In Adv. Catal.; Eley, D. D. , Pines, H. , Weez, P. B. , Eds.; Academic Press, 1979; Vol. 28, pp 1–78. 10.1016/S0360-0564(08)60133-9. [DOI] [Google Scholar]
  2. Rodriguez J. A., Goodman D. W.. High-pressure catalytic reactions over single-crystal metal surfaces. Surf. Sci. Rep. 1991;14(1):1–107. doi: 10.1016/0167-5729(91)90002-F. [DOI] [Google Scholar]
  3. Van Spronsen M. A., Frenken J. W. M., Groot I. M. N.. Surface science under reaction conditions: CO oxidation on Pt and Pd model catalysts. Chem. Soc. Rev. 2017;46(14):4347–4374. doi: 10.1039/C7CS00045F. [DOI] [PubMed] [Google Scholar]
  4. Schlögl R.. Heterogeneous Catalysis. Angew. Chem., Int. Ed. 2015;54(11):3465–3520. doi: 10.1002/anie.201410738. [DOI] [PubMed] [Google Scholar]
  5. Tao F., Salmeron M.. Surface restructuring and predictive design of heterogeneous catalysts. Science. 2024;386(6724):eadq0102. doi: 10.1126/science.adq0102. [DOI] [PubMed] [Google Scholar]
  6. Xu L., Papanikolaou K. G., Lechner B. A. J., Je L., Somorjai G. A., Salmeron M., Mavrikakis M.. Formation of active sites on transition metals through reaction-driven migration of surface atoms. Science. 2023;380(6640):70–76. doi: 10.1126/science.add0089. [DOI] [PubMed] [Google Scholar]
  7. Kaiser S., Maleki F., Zhang K., Harbich W., Heiz U., Tosoni S., Lechner B. A. J., Pacchioni G., Esch F.. Cluster Catalysis with Lattice Oxygen: Tracing Oxygen Transport from a Magnetite (001) Support onto Small Pt Clusters. ACS Catal. 2021;11(15):9519–9529. doi: 10.1021/acscatal.1c01451. [DOI] [Google Scholar]
  8. Kaiser S., Plansky J., Krinninger M., Shavorskiy A., Zhu S., Heiz U., Esch F., Lechner B. A. J.. Does Cluster Encapsulation Inhibit Sintering? Stabilization of Size-Selected Pt Clusters on Fe3O4(001) by SMSI. ACS Catal. 2023;13(9):6203–6213. doi: 10.1021/acscatal.3c00448. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Lykhach Y., Kozlov S. M., Skála T., Tovt A., Stetsovych V., Tsud N., Dvořák F., Johánek V., Neitzel A., Mysliveček J., Fabris S., Matolín V., Neyman K. M., Libuda J.. Counting electrons on supported nanoparticles. Nat. Mater. 2016;15(3):284–288. doi: 10.1038/nmat4500. [DOI] [PubMed] [Google Scholar]
  10. Campbell C. T.. Electronic perturbations. Nat. Chem. 2012;4(8):597–598. doi: 10.1038/nchem.1412. [DOI] [PubMed] [Google Scholar]
  11. Tauster S. J.. Strong metal-support interactions. Acc. Chem. Res. 1987;20(11):389–394. doi: 10.1021/ar00143a001. [DOI] [Google Scholar]
  12. Tauster S. J., Fung S. C., Garten R. L.. Strong metal-support interactions. Group 8 noble metals supported on titanium dioxide. J. Am. Chem. Soc. 1978;100(1):170–175. doi: 10.1021/ja00469a029. [DOI] [Google Scholar]
  13. Pesty F., Steinrück H.-P., Madey T. E.. Thermal stability of Pt films on TiO2(110): evidence for encapsulation. Surf. Sci. 1995;339(1):83–95. doi: 10.1016/0039-6028(95)00605-2. [DOI] [Google Scholar]
  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(16):3646–3649. doi: 10.1103/PhysRevLett.84.3646. [DOI] [PubMed] [Google Scholar]
  15. Bennett R. A., Pang C. L., 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(18):4688–4696. doi: 10.1021/jp0138328. [DOI] [Google Scholar]
  16. Bowker M., Stone P., Morrall P., Smith R., Bennett R., Perkins N., Kvon R., Pang C., Fourre E., Hall M.. Model catalyst studies of the strong metal–support interaction: Surface structure identified by STM on Pd nanoparticles on TiO2(110) J. Catal. 2005;234(1):172–181. doi: 10.1016/j.jcat.2005.05.024. [DOI] [Google Scholar]
  17. Fu Q., Wagner T., Olliges S., Carstanjen H.-D.. Metal–Oxide Interfacial Reactions: Encapsulation of Pd on TiO2 (110) J. Phys. Chem. B. 2005;109(2):944–951. doi: 10.1021/jp046091u. [DOI] [PubMed] [Google Scholar]
  18. Beck A., Huang X., Artiglia L., Zabilskiy M., Wang X., Rzepka P., Palagin D., Willinger M.-G., van Bokhoven J. A.. The dynamics of overlayer formation on catalyst nanoparticles and strong metal-support interaction. Nat. Commun. 2020;11(1):3220. doi: 10.1038/s41467-020-17070-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Tauster S. J., Fung S. C.. Strong metal-support interactions: Occurrence among the binary oxides of groups IIA–VB. J. Catal. 1978;55(1):29–35. doi: 10.1016/0021-9517(78)90182-3. [DOI] [Google Scholar]
  20. Qin Z. H., Lewandowski M., Sun Y. N., Shaikhutdinov S., Freund H. J.. Encapsulation of Pt Nanoparticles as a Result of Strong Metal–Support Interaction with Fe3O4(111) J. Phys. Chem. C. 2008;112(27):10209–10213. doi: 10.1021/jp801756q. [DOI] [Google Scholar]
  21. Xu M., Peng M., Tang H., Zhou W., Qiao B., Ma D.. Renaissance of Strong Metal–Support Interactions. J. Am. Chem. Soc. 2024;146(4):2290–2307. doi: 10.1021/jacs.3c09102. [DOI] [PubMed] [Google Scholar]
  22. Frey H., Beck A., Huang X., van Bokhoven J. A., Willinger M. G.. Dynamic interplay between metal nanoparticles and oxide support under redox conditions. Science. 2022;376(6596):982–987. doi: 10.1126/science.abm3371. [DOI] [PubMed] [Google Scholar]
  23. Matsubu J. C., Zhang S., Derita L., Marinkovic N. S., Chen J. G., Graham G. W., Pan X., Christopher P.. Adsorbate-mediated strong metal–support interactions in oxide-supported Rh catalysts. Nat. Chem. 2017;9(2):120–127. doi: 10.1038/nchem.2607. [DOI] [PubMed] [Google Scholar]
  24. Monai M., Jenkinson K., Melcherts A. E. M., Louwen J. N., Irmak E. A., Van Aert S., Altantzis T., Vogt C., van der Stam W., Duchoň T., Šmíd B., Groeneveld E., Berben P., Bals S., Weckhuysen B. M.. Restructuring of titanium oxide overlayers over nickel nanoparticles during catalysis. Science. 2023;380(6645):644–651. doi: 10.1126/science.adf6984. [DOI] [PubMed] [Google Scholar]
  25. Herzing A. A., Kiely C. J., Carley A. F., Landon P., Hutchings G. J.. Identification of Active Gold Nanoclusters on Iron Oxide Supports for CO Oxidation. Science. 2008;321(5894):1331–1335. doi: 10.1126/science.1159639. [DOI] [PubMed] [Google Scholar]
  26. Crampton A. S., Rötzer M. D., Landman U., Heiz U.. Can Support Acidity Predict Sub-Nanometer Catalyst Activity Trends? ACS Catal. 2017;7(10):6738–6744. doi: 10.1021/acscatal.7b01844. [DOI] [Google Scholar]
  27. Slavinskaya E. M., Stadnichenko A. I., Quinlivan Domínguez J. E., Stonkus O. A., Vorokhta M., Šmíd B., Castro-Latorre P., Bruix A., Neyman K. M., Boronin A. I.. States of Pt/CeO2 catalysts for CO oxidation below room temperature. J. Catal. 2023;421:285–299. doi: 10.1016/j.jcat.2023.03.004. [DOI] [Google Scholar]
  28. Gao F., McClure S. M., Cai Y., Gath K. K., Wang Y., Chen M. S., Guo Q. L., Goodman D. W.. CO oxidation trends on Pt-group metals from ultrahigh vacuum to near atmospheric pressures: A combined in situ PM-IRAS and reaction kinetics study. Surf. Sci. 2009;603(1):65–70. doi: 10.1016/j.susc.2008.10.031. [DOI] [Google Scholar]
  29. Eads C. N., Wang W., Küst U., Prumbs J., Temperton R. H., Scardamaglia M., Schnadt J., Knudsen J., Shavorskiy A.. Resolving active species during the carbon monoxide oxidation over Pt(111) on the microsecond timescale. Nat. Commun. 2025;16(1):1216. doi: 10.1038/s41467-025-56576-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Wang H., Liu J.-X., Allard L. F., Lee S., Liu J., Li H., Wang J., Wang J., Oh S. H., Li W., Flytzani-Stephanopoulos M., Shen M., Goldsmith B. R., Yang M.. Surpassing the single-atom catalytic activity limit through paired Pt-O-Pt ensemble built from isolated Pt1 atoms. Nat. Commun. 2019;10(1):3808. doi: 10.1038/s41467-019-11856-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Henderson M. A.. A surface perspective on self-diffusion in rutile TiO2 . Surf. Sci. 1999;419(2):174–187. doi: 10.1016/S0039-6028(98)00778-X. [DOI] [Google Scholar]
  32. Aono M., Hasiguti R. R.. Interaction and ordering of lattice defects in oxygen-deficient rutile TiO2–x . Phys. Rev. B. 1993;48(17):12406–12414. doi: 10.1103/PhysRevB.48.12406. [DOI] [PubMed] [Google Scholar]
  33. Onishi H., Fukui K. i., Iwasawa Y.. Atomic-Scale Surface Structures of TiO2(110) Determined by Scanning Tunneling Microscopy: A New Surface-Limited Phase of Titanium Oxide. Bull. Chem. Soc. Jpn. 1995;68(9):2447–2458. doi: 10.1246/bcsj.68.2447. [DOI] [Google Scholar]
  34. Li M., Hebenstreit W., Gross L., Diebold U., Henderson M. A., Jennison D. R., Schultz P. A., Sears M. P.. Oxygen-induced restructuring of the TiO2(110) surface: a comprehensive study. Surf. Sci. 1999;437(1):173–190. doi: 10.1016/S0039-6028(99)00720-7. [DOI] [Google Scholar]
  35. Li M., Hebenstreit W., Diebold U.. Morphology change of oxygen-restructured TiO2(110) surfaces by UHV annealing: Formation of a low-temperature (1 × 2) structure. Phys. Rev. B. 2000;61(7):4926–4933. doi: 10.1103/PhysRevB.61.4926. [DOI] [Google Scholar]
  36. Kraushofer F., Krinninger M., Kaiser S., Reich J., Jarosz A., Füchsl M., Anand G., Esch F., Lechner B. A. J.. The influence of bulk stoichiometry on near-ambient pressure reactivity of bare and Pt-loaded rutile TiO2(110) Nanoscale. 2024;16(38):17825–17837. doi: 10.1039/D4NR01702A. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Bowker M., Bennett R. A.. The role of Ti3+ interstitials in TiO2(110) reduction and oxidation. J. Phys.: Condens. Matter. 2009;21(47):474224. doi: 10.1088/0953-8984/21/47/474224. [DOI] [PubMed] [Google Scholar]
  38. Bennett R. A., Stone P., Price N. J., Bowker M.. Two (1 × 2) Reconstructions of TiO2(110): Surface Rearrangement and Reactivity Studied Using Elevated Temperature Scanning Tunneling Microscopy. Phys. Rev. Lett. 1999;82(19):3831–3834. doi: 10.1103/PhysRevLett.82.3831. [DOI] [Google Scholar]
  39. Bennett R. A., Stone P., Bowker M.. Pd nanoparticle enhanced re-oxidation of non-stoichiometric TiO2: STM imaging of spillover and a new form of SMSI. Catal. Lett. 1999;59(2–4):99–105. doi: 10.1023/a:1019053512230. [DOI] [Google Scholar]
  40. Petzoldt P., Eder M., Mackewicz S., Blum M., Kratky T., Günther S., Tschurl M., Heiz U., Lechner B. A. J.. Tuning Strong Metal–Support Interaction Kinetics on Pt-Loaded TiO2(110) by Choosing the Pressure: A Combined Ultrahigh Vacuum/Near-Ambient Pressure XPS Study. J. Phys. Chem. C. 2022;126(38):16127–16139. doi: 10.1021/acs.jpcc.2c03851. [DOI] [Google Scholar]
  41. Crumlin E. J., Liu Z., Bluhm H., Yang W., Guo J., Hussain Z.. X-ray spectroscopy of energy materials under in situ/operando conditions. J. Electron Spectrosc. Relat. Phenom. 2015;200:264–273. doi: 10.1016/j.elspec.2015.06.008. [DOI] [Google Scholar]
  42. Grass M. E., Karlsson P. G., Aksoy F., Lundqvist M., Wannberg B., Mun B. S., Hussain Z., Liu Z.. New ambient pressure photoemission endstation at Advanced Light Source beamline 9.3.2. Rev. Sci. Instrum. 2010;81(5):053106. doi: 10.1063/1.3427218. [DOI] [PubMed] [Google Scholar]
  43. Wendt S., Sprunger P. T., Lira E., Madsen G. K. H., Li Z., Hansen J. Ø., Matthiesen J., Blekinge-Rasmussen A., Lægsgaard E., Hammer B., Besenbacher F.. The Role of Interstitial Sites in the Ti3d Defect State in the Band Gap of Titania. Science. 2008;320(5884):1755–1759. doi: 10.1126/science.1159846. [DOI] [PubMed] [Google Scholar]
  44. Smekal W., Werner W. S. M., Powell C. J.. Simulation of electron spectra for surface analysis (SESSA): a novel software tool for quantitative Auger-electron spectroscopy and X-ray photoelectron spectroscopy. Surf. Interface Anal. 2005;37(11):1059–1067. doi: 10.1002/sia.2097. [DOI] [Google Scholar]
  45. Werner, W. ; Smekal, W. ; Powell, C. J. , NIST Database for the Simulation of Electron Spectra for Surface Analysis (SESSA), SRD 100, Version 2.1. 2017. [Google Scholar]
  46. Van Spronsen M. A., Frenken J. W. M., Groot I. M. N.. Observing the oxidation of platinum. Nat. Commun. 2017;8(1):429. doi: 10.1038/s41467-017-00643-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Tougaard S.. Improved XPS analysis by visual inspection of the survey spectrum. Surf. Interface Anal. 2018;50(6):657–666. doi: 10.1002/sia.6456. [DOI] [Google Scholar]
  48. Průša S., Linford M. R., Vaníčková E., Bábík P., Pinder J. W., Šikola T., Brongersma H. H.. A practical guide to interpreting low energy ion scattering (LEIS) spectra. Appl. Surf. Sci. 2024;657:158793. doi: 10.1016/j.apsusc.2023.158793. [DOI] [Google Scholar]
  49. Niehus H., Heiland W., Taglauer E.. Low-energy ion scattering at surfaces. Surf. Sci. Rep. 1993;17(4):213–303. doi: 10.1016/0167-5729(93)90024-J. [DOI] [Google Scholar]
  50. Ruus R., Kikas A., Saar A., Ausmees A., Nõmmiste E., Aarik J., Aidla A., Uustare T., Martinson I.. Ti 2p and O 1s X-ray absorption of TiO2 polymorphs. Solid State Commun. 1997;104(4):199–203. doi: 10.1016/S0038-1098(97)00300-1. [DOI] [Google Scholar]
  51. Wertheim G. K., Dicenzo S. B., Youngquist S. E.. Unit Charge on Supported Gold Clusters in Photoemission Final State. Phys. Rev. Lett. 1983;51(25):2310–2313. doi: 10.1103/PhysRevLett.51.2310. [DOI] [Google Scholar]
  52. Tamura K., Kudo M., Owari M., Nihei Y.. X-Ray Photoelectron Diffraction (XPED) Studies of Platinum on Titanium Dioxide (110) Surface. Chem. Lett. 1986;15(11):1921–1924. doi: 10.1246/cl.1986.1921. [DOI] [Google Scholar]
  53. Steinrück H.-P., Pesty F., Zhang L., Madey T. E.. Ultrathin films of Pt on TiO2(110): Growth and chemisorption-induced surfactant effects. Phys. Rev. B. 1995;51(4):2427–2439. doi: 10.1103/PhysRevB.51.2427. [DOI] [PubMed] [Google Scholar]
  54. Perco D., Pozzo M., Berti A., Loi F., Lacovig P., Lizzit S., Kartouzian A., Heiz U., Alfè D., Baraldi A.. Limitations in Determining Oxidation States in Condensed Matter at the Subnanometric Scale. J. Am. Chem. Soc. 2025;147(25):21501–21511. doi: 10.1021/jacs.5c02242. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Miller D. J., Öberg H., Kaya S., Sanchez Casalongue H., Friebel D., Anniyev T., Ogasawara H., Bluhm H., Pettersson L. G. M., Nilsson A.. Oxidation of Pt(111) under Near-Ambient Conditions. Phys. Rev. Lett. 2011;107(19):195502. doi: 10.1103/PhysRevLett.107.195502. [DOI] [PubMed] [Google Scholar]
  56. Quinlivan Domínguez J. E., Neyman K. M., Bruix A.. Stability of oxidized states of freestanding and ceria-supported PtOx particles. J. Chem. Phys. 2022;157(9):094709. doi: 10.1063/5.0099927. [DOI] [PubMed] [Google Scholar]
  57. Beard B. C., Ross P. N.. Platinum-titanium alloy formation from high-temperature reduction of a titania-impregnated platinum catalyst: implications for strong metal-support interaction. J. Phys. Chem. 1986;90(26):6811–6817. doi: 10.1021/j100284a020. [DOI] [Google Scholar]
  58. Chase, M. W. NIST-JANAF Thermochemical Tables. 4 ed.; National Institute of Standards and Technology: Gaithersburg, MD, 1998. [Google Scholar]
  59. Meschter P. J., Worrell W. L.. An investigation of high-temperature thermodynamic properties in the Pt-Ti system. Metall. Trans. A. 1976;7(2):299–305. doi: 10.1007/BF02644471. [DOI] [Google Scholar]
  60. Bonanni S., Aït-Mansour K., Harbich W., Brune H.. Effect of the TiO2 Reduction State on the Catalytic CO Oxidation on Deposited Size-Selected Pt Clusters. J. Am. Chem. Soc. 2012;134(7):3445–3450. doi: 10.1021/ja2098854. [DOI] [PubMed] [Google Scholar]
  61. Tober S., Creutzburg M., Arndt B., Krausert K., Mattauch S., Koutsioubas A., Pütter S., Mohd A. S., Volgger L., Hutter H., Noei H., Vonk V., Lott D., Stierle A.. Observation of iron diffusion in the near-surface region of magnetite at 470 K. Phys. Rev. Res. 2020;2(2):023406. doi: 10.1103/PhysRevResearch.2.023406. [DOI] [Google Scholar]
  62. Powell C. J., Jablonski A.. The NIST Electron Effective-Attenuation-Length Database. Journal of Surface Analysis. 2002;9(3):322–325. doi: 10.1384/jsa.9.322. [DOI] [Google Scholar]
  63. Lundy T. S., Coghlan W. A.. Article. J. Phys. Colloques. 1973;34(C9):C9-299. doi: 10.1051/jphyscol:1973953. [DOI] [Google Scholar]
  64. Venkatu D. A., Poteat L. E.. Diffusion of titanium of single crystal rutile. Mater. Sci. Eng. 1970;5(5):258–262. doi: 10.1016/0025-5416(70)90014-5. [DOI] [Google Scholar]
  65. Lee D. K., Yoo H. I.. Unusual oxygen re-equilibration kinetics of TiO2−δ . Solid State Ionics. 2006;177(1):1–9. doi: 10.1016/j.ssi.2005.10.016. [DOI] [Google Scholar]
  66. Zhang Z., Lee J., Yates J. T., Bechstein R., Lira E., Hansen J. Ø., Wendt S., Besenbacher F.. Unraveling the Diffusion of Bulk Ti Interstitials in Rutile TiO2(110) by Monitoring Their Reaction with O Adatoms. J. Phys. Chem. C. 2010;114(7):3059–3062. doi: 10.1021/jp910358w. [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

ja5c14353_si_001.pdf (3.6MB, pdf)

Data Availability Statement

The data is available on: 10.5281/zenodo.17233126

Articles from Journal of the American Chemical Society are provided here courtesy of American Chemical Society

RESOURCES