Advertisement

Topics in Catalysis

, Volume 61, Issue 12–13, pp 1313–1322 | Cite as

Oxygen Assisted Morphological Changes of Pt Nanosized Crystals

  • Sylwia Owczarek
  • Sten V. Lambeets
  • Cédric Barroo
  • Robert Bryl
  • Leszek Markowski
  • Thierry Visart de Bocarmé
Original Paper
  • 166 Downloads

Abstract

Thermal faceting of clean and oxygen-covered Pt nanocrystals was investigated at the nanoscale by means of field ion microscopy (FIM) and field emission microscopy (FEM) in the 500–700 K temperature range. FIM and FEM are used to study the morphology of the crystal prepared in the form of a sharp tip. The tip extremity is observed with nanoscale lateral resolution and corresponds to a suitable model of a single nanoparticle of a real catalyst. By contrast to similar studies on iridium, palladium and rhodium, small oxygen exposures (~ 10 L) and annealing treatments at 700 K did not lead to strong surface modifications. The field ion micrograph was similar to the pattern obtained for the nanocrystals annealed under vacuum conditions, revealing only low index {001} and {111} facets. For higher oxygen doses, i.e. ≥ 100 L, and in field-free conditions, the flat {100}, {111} and {113} facets were developed after annealing the tip at 700 K, which was attributed to the formation of oxide layers. For comparison, the surface modification was studied under oxygen-rich conditions but in the presence of an electric field at 700 K. The results showed that only former reconstruction was observed regardless of oxygen doses. These results are also promising in the frame of engineering catalysts since different gas exposure may lead to the extension or shrinking of specific facets, which may impact the efficiency of the catalyst.

Keywords

Platinum Nanocatalysis Model catalysis Surface reconstruction Field emission techniques Field ion microscopy 

1 Introduction

The platinum group metals (PGM), comprising Pt, Pd, Rh, Ir, are of high importance in the modern chemical industry, especially in the automotive pollution control by catalysis and fuel cell technology. Platinum is still the primary element of many catalytic processes, e.g. the oxidation of NH3 in nitric acid synthesis, the hydrosilylation of alkenes and aryl alkynes, the hydrogenation of benzene and cyclohexene, as well as the direct decomposition and oxidation of alcohols [1, 2, 3, 4, 5, 6, 7]. In the field of automobile exhaust catalysis, Pt is used to oxidise CO and unburnt hydrocarbons to CO2 and H2O [8, 9, 10]. Under reactive environments, the catalytic processes occur according to a well-accepted Langmuir–Hinshelwood mechanism [11]. As a result, the catalyst may undergo significant morphological and structural changes that can in turn influence the chemical composition of the adsorbed layers. During the past decades, many studies focusing on the surface structures of 2D single Pt crystals have been performed. For instance, it has been shown that upon exposure to O2 or CO, surfaces of various Pt single crystals reconstruct into superstructures—(√3×√3)R30°, (2 × 2), c(4 × 2), (10 × 2) to cite some but a few—depending on the temperature and surface coverage of the reactive gas [12, 13, 14, 15, 16, 17, 18, 19, 20].

In applied formulations of catalysts, the catalytic materials consist of nanometer-sized particles [21]. Depending on their geometric structure, the catalysts expose a variety of facets, usually with different catalytic reactivities for a given reaction [22, 23]. Low-coordinated atoms can interact with reactants and serve as active sites for breaking chemical bonds [24, 25, 26]. However, due to the mobility of the atoms within the timeframe of the experiments, the shape and local structure of the nanoparticles are evolving towards the equilibrium shape of the catalyst crystals corresponding to the minimum of total surface free energy of the system. It has been shown that the surface morphology can evolve with a temperature treatment and/or the presence of adsorbates [27]. Strong interactions between adsorbate and substrate atoms can enhance the anisotropy in local surface energy, leading to a rearrangement of the substrate atoms through reconstruction and faceting processes [28]. Moreover, the very first atomic layers can also undergo changes in their chemical nature due to the formation of metal oxide layers or subsurface species [29, 30, 31]. If the reaction is structure-sensitive, these modifications of the catalyst may have a significant impact on specific reaction steps, and subsequently on the catalytic activity. Using field ion microscopy (FIM), Voss et al. showed that morphological changes of the 3D Pt nanocrystal occur in presence of NO gas at 520 K [32]: the quasi-hemispherical shape of the apex of the tip-sample transforms into a truncated pyramid with slopes of {111}-orientations. This result was explained by the growth of {111} planes at the expense of high-index planes. The transient reconstruction of the Pt nanocrystals was also observed during the ongoing NO2 + H2 reaction at 390 K followed by field emission microscopy (FEM) [33]. The oscillating reaction, visible as periodic “surface explosions”, i.e. high increase of brightness signal in the FE images, first occurred on {011} planes and then expanded along the <100> zones over the (001) central pole. Oscillating reactions were also observed for H2 + O2, CO + O2 and NO2 + H2 reactions on Pt nanocrystals in the 300–400 K temperature range [34, 35, 36, 37, 38]. These examples prove that the synergistic effects of the temperature and gas composition have a strong impact on the local surface reactivities, showing the consequences of the morphological evolution of the catalysts nanoparticles.

In the present work, we used field emission and field ion microscopies to investigate the oxygen adsorption on single Pt nanocrystals. The aim of this work is to address how the effect of adsorbates alters the shape and morphology of Pt catalyst nanoparticles. Both techniques enable in situ studies of reconstructions of the apex of the specimens, which closely approximate a single nanoparticle of catalyst, before, after and even during the ongoing reactions and adsorption processes. These studies are performed with nanoscale or possibly atomic lateral resolution [33]. The electric field required for imaging may modify the local chemistry between the adsorbate and the substrate, as well as between metal atoms [39]. In this case, this additional parameter can provoke changes in the shape of the nanocrystal, by faceting formation, and affect every stage of the catalytic reaction. Therefore, we investigated the adsorption of oxygen on Pt sample in the presence and in the absence of electric field to better understand the morphological changes. In the latter case, the exposure of oxygen in field-free conditions enables to approach more realistic conditions.

2 Experimental Section

Experiments were performed in an ultrahigh vacuum (UHV) system with a base pressure of ~ 10−8 Pa. The UHV chamber is equipped with FIM and FEM facilities. Pt-emitter tips were produced by electrochemical etching from a Pt wire (purity: 99.95% − 0.1 mm diameter) in a molten salt mixture of NaCl and NaNO3 (1:4, w/w) at ~ 520 °C with 2 VDC. After etching, the samples were rinsed in distilled water, and both their profile and integrity were verified by optical microscopy. The tips were then introduced in the microscope where they underwent cycles of in situ treatments: thermal annealing at 600 K, field evaporation and Ne+ ion sputtering at 200 K [40]. The samples were imaged at low temperature (60–70 K) with atomic resolution by FIM and with nanoscale resolution by FEM. This procedure produced clean, atomically smooth, and very sharp metallic specimens. The radius of curvature was estimated by net counting of the number of atomic layers between two facets of known crystallographic orientations in the field ion micrograph [41]. The radius of curvature of the samples used in this work varies between 15 and 25 nm.

After characterisation of the well-developed sample by FIM and FEM, the oxygen gas (purity 99.998%, Messer) was introduced into the system through a calibrated leak valve. In this work, two distinct series of experiments were performed to study the morphological changes of Pt samples. In the first set of experiments, the Pt sample was kept at 300 K and exposed to oxygen doses ranging from 10 to 5000 L (1 Langmuir = 1 L = 1.33 × 10−4 Pa s) in field-free conditions. The samples were annealed at a temperature of 500–700 K under vacuum conditions. These specific annealing temperatures correspond to usual working temperatures encountered in automotive catalytic converters [42]. In the second set of experiments, the samples were held at 700 K and exposed to oxygen gas in the presence of an electric field. After that, the samples were cooled down to cryogenic temperature to perform further FIM characterisation with atomic resolution. These two series of experiments raise the question of the influence of the electric field on reconstruction processes. To ensure a clean starting point prior to each oxygen dosing, the Pt samples were cleaned by Ne+ ion sputtering and field evaporation after every single experiment (of selected O2 dosing and annealing temperature). This cleaning procedure was applied to remove oxygen species from the tip apex, from the shank and the possibly present subsurface oxygen-species near the apex, i.e. oxygen atoms located below the surface, in the very first atomic layers. As additional precaution, before starting a new experiment, we ensured that the cryostat reached room temperature to prevent the desorption of any residual physisorbed oxygen from the cooling system. This also prevented the formation of thick oxygen layer on the cryostat, the sample holder and the base of tip that could seriously affect the imaging and the evaluation of the oxygen amounts dosed on the sample.

All the field ion micrographs presented in this paper were acquired using Ne gas (purity 99.999%, Praxair) at a pressure of ~ 3 × 10−3 Pa at cryogenic temperature (50–70 K). The specimens were cooled down using a compressed helium cryostat and temperatures was accurately measured by using a Ni–NiCr thermocouple at the basis of the tip-sample. The micrographs with atomic-scale resolution were recorded with a high sensitivity back-illuminated CCD camera (Princeton Instruments, VersArray 512B-XP, 512 × 512 pixels, 16 bits per pixel).

3 Results and Discussion

The first part of this section focuses on the characterisation of the initial state of the clean Pt nanosized crystals by FIM, as well as the observation of thermally-induced reconstructions. The second part aims at investigating the adsorption-induced reconstructions in field-free conditions. The third and last part corresponds to the description of sample reconstruction obtained during oxygen adsorption in presence of an electric field.

3.1 Characterisation of Clean and Thermally-Treated Pt Samples

Each Pt sample was characterised before any physicochemical treatment to observe the initial structure of the field emitter tip. Then, the evolution of the Pt nanocrystal morphology was studied as a function of the annealing temperature under vacuum conditions and in absence of electric field to strictly probe thermally-induced reconstructions. These pre-characterisations are necessary to distinguish the presence of adsorbate-induced reconstructions from bare thermally-induced reconstructions. Figure 1a represents a high-resolution FI micrograph of a (001)-oriented Pt sample. The quasi-hemispherical shape of the tip was obtained after applying the cleaning procedure described in Sect. 2. The field ion pattern reveals the presence of multiple facets with different crystallographic orientations. Atomic resolution is clearly reached on the most open facets. The main facets are designated by their Miller indexes, and the radius of curvature of the Pt tip is estimated to ~ 17 nm. Such atomically smooth sample is used as the starting point for further experiments.

Fig. 1

Characterisation of the sample and thermally-induced reconstructions. a Field ion micrographs of a clean (001)-oriented Pt sample; evolution of the FI pattern after annealing at b 500 K for 10 min, c 650 K for 2 min, d 700 K for 2 min. The evolution of the morphology upon annealing at increasing temperatures shows a strong faceting of the sample, which translates into the formation of {111} and (001) main facets, and the presence of missing row formation between these facets. Imaging conditions: F = 36 V nm−1, T = 70 K

The clean Pt tip was then annealed at different temperatures under vacuum conditions and in absence of electric field. The corresponding FI patterns are represented in Fig. 1b–d. Samples annealed at mild temperatures, up to around 600 K, for 10 min resulted in highly disordered patterns, as can be seen in Fig. 1b. Annealing process increases the mobility of surface atoms which can therefore occupy the most stable sites and form ordered structures. However, if the temperature is too low and/or the annealing time too short, the energy of the system might not be sufficient to reach an ordered structure: the atoms still undergo diffusion but do not occupy the most stable adsorption sites, resulting in a disordered structure. The first manifestation of a surface reconstruction with a discernible ordered structure was observed at 650 K. The micrograph, presented in Fig. 1c, reveals small but easily distinguishable {111} and (001) facets separated by closed-packed rows of ordered atoms over {113} and {011} regions. Annealing at T = 700 K caused the extension of these {111} and (001) facets. No other main facets could be recognised in the field ion micrograph as seen in Fig. 1d. By comparing Fig. 1a–d, a clear surface restructuration is observed due to the annealing conditions, which translates into extension of the {111} and (001) facets, and formation of ordered structures between these facets. From a general point of view, the reconstruction of a sample is driven by the trend of the system to reach its minimum of total free energy [43].
$${F_{tot}}=\mathop \sum \limits_{m} \mathop \smallint \limits_{{{A_m}}}^{~} \gamma \left( {{{\vec {n}}_m}} \right).dA$$
(1)
where Ftot corresponds to the total free energy of the system, A the area of the facets, γ the surface tension and \({\vec {n}_m}\) the crystallographic orientation of the facet.

In the case of our experiments, the appearance of these facets is in agreement with theoretical studies [44]. The Gibbs free surface energies γ of Pt, calculated in Ref. [45], is approximately 0.093 eV Å−2 for a (111) plane and 0.098 eV Å−2 for a hexagonally reconstructed (001) planes. Moreover, the authors claimed that the (1 × 5) structure on Pt(001) surface is energetically more favourable than the (1 × 1) structure with γ = 0.113 eV Å−2. Even though atomic resolution was achieved in this experiment, the internal surface structure of the (001) plane cannot be identified: inner-atomic arrangement of the densest and flat facets remains usually invisible, which is attributed to the locally smaller electric field strength as compared to more open facets.

The {111} and (001) facets are separated by a few visible atomic rows. According to previous works [29, 46, 47], these local reconstructions correspond to missing-row type (MR) formation, that were identified for (011), (113), and (112) surfaces. Extensive studies of the adsorbent-free single crystals have shown that hill-and-valley structure with (1 × 2) configuration (periodicity) is energetically more favourable as compared to unreconstructed (1 × 1) surface and may occur on (113) and (112) facets [48, 49], while the clean Pt(110) prefers to reconstruct into (1 × n) with n = 2–4 [50, 51] upon annealing treatment. The MR structures were also observed and studied on 3D nanocrystals for Ir [38], Rh [29] and Pt [52] at 450 K. For the equilibrium shape of the Pt 3D-crystal obtained from calculated values of γ, in accordance with the Wulff construction, {111} and {100} planes dominate the morphology. However, the presence of relatively small, elongated {110} with (1 × 2) periodicity and, surprisingly, {112} planes were reported. The γ(113)MR(111) ratio is about 1.16, while surface energy ratio between (112) and (111) is slightly lower (γ(112)/(111) = 1.11) [45]. The (112) surface ([3(111) × (100)] in step notation) is formed by (111) terraces separated by monatomic (100) steps, whereas the (113) surface ([2(111) × (100)] in step notation) has two-atom-wide (111) terraces. This result highlights that the differences between the two surfaces are not considerable. This implies that they are likely to occur on Pt nanocrystals. Nevertheless, from our observations, the {112}, {113} and {011} planes do not exist independently on field emitter tip upon annealing at 700 K. The straight-line steps, visible in the micrograph, are attributed to formed {111} terraces and {100} steps located between low-Miller index planes, as shown in Fig. 1d and schematically in Fig. 2.

Fig. 2

Schematic representation (side-view) of the missing row formation from (1 × 1) to (1 × 3) structure on a (113)-type fcc crystal plane

The sequence of patterns obtained by FIM shown in Fig. 1 depicts transition from an almost hemispherical shape of the initial crystal to a polyhedral shape. Although the annealing temperature was low to achieve the well-developed, close to polyhedron near-equilibrium crystal shape (as it was obtained before for clean Ir crystals [38]), the presented results indicate that the final shape of the annealed Pt crystal consist exclusively of {111} and {100} facets divided by rounded edges built of terraces and steps of {111} and/or {100} structure. The observed results suggest that the annealed-equilibrium shape of Pt nanocrystal is similar to those of Ir and Rh.

3.2 Oxygen Adsorption on Pt Nanocrystals in Field-Free Conditions

It is well-known that the presence of adsorbates can lead to significant changes in surface free energies and, consequently, can induce surface reconstructions such as faceting [37, 38, 44]. The gain in energy, driving the morphological changes of the surface, can originate from the molecule–substrate interactions [44]. Moreover, a vicinal surface is required to supply a sufficient density of steps. These processes are energetically expensive and would normally require higher annealing temperatures than those used in this study [44]. In this section the effect of oxygen adsorption on Pt nanocrystals and the annealing at high temperatures are presented. Figure 3 demonstrates the evolution of the surface morphology of the (001)-oriented Pt specimen exposed to 10 L of oxygen at 300 K as a function of the annealing temperatures. Adsorption and annealing procedures are performed under field-free conditions. Upon annealing up to 620 K, the presented FI patterns did not change significantly (Fig. 3a) as compared to the experiments of thermal annealing in absence of oxygen (see Fig. 1b): the high degree of disorder is still observed. Only the small (001) central pole and hardly distinguishable {111} planes can be identified in this micrograph. The lack of visible atoms arrangement and additional facets can be associated with the relatively low annealing temperature at which the mass transport was not sufficient for the development of larger facets within the timeframe of the experiment.

Fig. 3

Characterisation of a Pt sample after low-dose oxygen exposure. FI patterns of Pt tip exposed to 10 L (4.7 × 10−5 Pa·30 s) at 300 K and annealing at a 620 K (imaging conditions: F = 43 V nm−1, T = 66 K); b 650 K (F = 44 V nm−1, T = 64 K); and c 700 K (F = 38 V nm−1, T = 63 K). Low doses of oxygen do not provoke significant changes in the shape of the nanocrystal. The FE pattern is similar to that of a Pt tip after heating treatment (see Fig. 1). Dark regions in the image represent relatively flat low Miller index planes: {100} and {111} and reconstructed {110} and {113} planes

Sample annealing at 650 K, presented in Fig. 3b, caused a significant extension of the dark regions located in the centre and on the edge of the image corresponding to (001) and {111} facets, respectively. These relatively well-developed low Miller index planes are separated by multiple lines of bright spots. Further heating at 700 K led to a considerable increase in size of the (001) and {111} planes. Still, no other facets can be unambiguously recognised. Local reconstructions, in the form of stripe-like structures, appeared on the {011} and {113} planes and are clearly visible. Previous studies reported that the Pt(110) surface presents a tendency to undergo a MR reconstruction under oxidising conditions. Oxygen species adsorb on the next nearest neighbour positions, on the opposite side of the Pt atom, along the row forming different phases, i.e. (1 × 2), c(2 × 4) or (2 × 2) [17, 18]. For the stepped (113) surface, oxygen atoms occupy positions at intervals of two or three platinum atoms along the [110] direction with an (1 × 2) periodicity [53]. Such arrangement of oxygen atoms allows the stabilisation of the structures consisting in terraces and steps on the platinum surfaces. It is expected that this process also takes place in our experiments.

The FI micrographs of Pt obtained after small doses of oxygen and annealing treatments are similar to the patterns obtained for annealing treatment in absence of oxygen, presented in Fig. 1d. Strong faceting did not occur in presence of adsorbed oxygen and the morphological changes are not observed under our experimental conditions, which is a striking difference with the results obtained for other PGMs. The faceting process of 3D nanocrystals under mild annealing conditions has been obtained on Ir, Pd and Rh [29, 37, 38, 54]. These studies showed that oxygen-induced reconstructions of nanocrystals took place after low-temperature annealing (< 600 K) and at lower doses (< 10 L) of dioxygen. Our results confirm that, as compared to other PGMs, Pt nanosized crystallites present a lower reactivity towards oxygen gas [55, 56].

In the line of the reported results, the sample was exposed to higher oxygen doses, up to 100 L. The outcome, presented in Fig. 4a, shows structural alterations of the surface: the dark elongated areas, visible in the micrograph, correspond to (001) and {111} planes. In comparison to the results described above, (001) and {111} facets are extended and the terraces and steps located between these specific facets are not visible. This result points to the instability of this structure at higher surface coverage of oxygen. An additional ring is formed by protruding atoms on the left-hand side (marked by an arrow in Fig. 4a). It is supposed that this additional dark region separated by bright spots correspond to the initial stages of {113} planes flattening.

Fig. 4

Characterisation of a Pt sample after high-dose oxygen exposure. Field ion micrographs of a (001)-oriented Pt tip exposed to an oxygen dose of a 100 L (1.33 × 10−4 Pa·100 s) (imaging conditions: F = 31 V nm−1, T = 72.0 K), and b 1100 L (1.47 × 10−3 Pa·100 s) (imaging conditions: F = 29 V nm−1, T = 67.7 K). High oxygen doses induce strong faceting of the surface, where flat (001), {011} as well as additional {113} facets are formed. Same field ion micrographs as in (b) imaged with a higher imaging field to observe the underlying structure c F = 31 V nm−1; and d F = 34 V nm−1

To provide better insights on the formation of the flat {113} planes, higher doses of oxygen gas were introduced into the reaction chamber. The morphology of the nanocrystal seen in Fig. 4a remained unchanged for O2 exposures up to 500 L. However, after higher oxygen exposure, 1100 L in the case presented Fig. 4b, further morphological changes can be observed: the FI pattern exhibits four dark and relatively flat planes located between the central (001) pole and the four {111} peripheral poles. Calculated angular distances between the centres of these newly formed facets and the (001) central pole suggest that they are of {113} orientation. The identification of the orientation of extended planes occurring after the faceting process can be performed by a careful field evaporation procedure. Indeed, the removal of a few atomic layers leads to a reduction in size of the enlarged dark areas, as can be seen in Fig. 4c, d showing the sequence of FI pattern during field evaporation. This process initially occurs on the disordered {012} planes, where an increased number of bright spots, corresponding to protruding atoms, are observed after increasing the imaging field from 29 to 31 V nm−1 (Fig. 4c). It can be concluded that field evaporation induces a slight decrease in size of the flat circular regions. When the electric field was raised by some 3 V nm−1, a few tens of atomic layers were removed from the tip and the crystal facets were recovered, eventually (Fig. 4d).

The formation of {113} facets on a particle after oxidation process was already observed in previous studies [45]. This process is associated with the higher adsorption energy of oxygen species in comparison to the other high-indexed surfaces. In the high-coverage regime, the open {113} facets are stabilised due to the formation of an oxide layer. This is in agreement with previous work obtained by FIM [33] where O-covered Pt{113} facets appeared darker because of a lower ionisation probability of the imaging gas [57]. Extensive investigations on (110), (111) and (110) extended surfaces with complementary techniques proved the presence of oxide species with different stoichiometries: PtO, Pt3O4, α-PtO2, and β-PtO2 [58, 59, 60, 61, 62]. From the standard enthalpies of formation of the different bulk Pt oxides, i.e. PtO (− 71 kJ mol−1), Pt3O4 (− 268 kJ mol−1), and PtO2 (− 134 kJ mol−1), a rough estimation of their relative stability may lead to the following trend: Pt3O4 > PtO2 > PtO [63, 64]. According to computational and experimental investigations, the α-PtO2 phase is thermodynamically favoured and PtO2-like structures develop during the metal oxidation, independently of the surface orientation [20, 65, 66]. Considering our results, we suggest that the oxygen dose is sufficient to form an oxide layer over the Pt specimen. Under the used experimental conditions, the oxide region can be created near the surface. However, the coverage and the structure type are not accessible using FIM technique. For this purpose, the 1D atom probe, combining a field ion microscope with a mass spectrometer, should be used to provide the local stoichiometry of that surface oxide as well as an estimation of its thickness [67]. Those data could possibly be correlated with high-resolution transmission electron microscopy. Nevertheless, it turns out to be a fact that the thin oxide layer formed on {113} leads to the stabilisation of these flat facets, as illustrated in Fig. 4b.

3.3 Oxygen Adsorption in the Presence of Electric Field

The next step towards the understanding of the interaction of oxygen gas with Pt field emitter tips consists in the study of oxygen exposure in the presence of an electric field. The electric field may indeed cause changes in the local chemistry between the adsorbate and the substrate [38, 68], and thus acts on relevant catalytic processes, e.g. adsorption, dissociation, diffusion. In these experiments, we used the FEM mode with an imaging field of ~ 3–5 V nm−1 during oxygen adsorption on the Pt specimens. The intensity of the field is lower than in FIM mode (> 30–35 V nm−1) [41] and its influence on the local chemistry is thus less pronounced. The behaviours observed while influenced by a steady electric field could however mimic the conditions encountered in studies over supported catalysts, such as metal/metal-oxide catalysts, where charge transfer from the catalytic nanoparticles to the oxide support occurs [69].

As in the previous sections, Pt samples were firstly characterised by FIM and FEM imaging (see Fig. 5a, b). Afterwards, the temperature was raised from 70 to 700 K. We mention that hydrogen gas was transiently introduced in the chamber, during FE imaging, to get rid of the possible presence of impurities. When hydrogen gas was evacuated, oxygen gas was injected in the chamber at a pressure of 5.0 × 10−5 Pa. This caused a decrease of the global brightness in the FE pattern, at the constant electric field, in less than a minute, corresponding to an oxygen dose of 30 L. Adsorbed oxygen is known to increase the local work function and to inhibit electron field emission [70]. Increasing the imaging field allowed to observe a reconstruction on the field emission pattern which now exhibits a high brightness of {113} and {011} planes under oxygen-rich state (Fig. 5c), in contrast with the FE pattern of clean Pt tip. Moreover, the brightness of the central (001) facet increased, but not systematically. The same results were observed in the 30–50 L oxygen-dose range. The origin of these changes in the FE pattern could be explained by surface reconstructions, but also by the presence of subsurface oxygen [71], which is in line with the assumption of the early stages of oxide layers formation, as discussed in the previous section. A progressive FI imaging (at 70 K and in absence of oxygen gas), presented in Fig. 5d–h, allows to investigate the reconstruction. As can be seen in Fig. 5d, {111} regions are remarkably dark as compared to the rest of the micrograph. The regions surrounding these facets show both undefined bright and elongated spots, sign of the presence of protruding atoms (probably oxygen atoms), and stripes, giving this square-like shape to the FI image. Figure 5e–h depict the evolution of the FI pattern as a function of the intensity of the electric field: an increased voltage of about 100 V radically changes the aspect of the FI micrograph (Fig. 5e). The very bright and irregularly arranged spots occur on {011} regions as well as between them and the centre of the tip. On the contrary, an accurate central circle of atoms, corresponding to the borders of the well-defined (001) plane and the edges of the {111} facets, are formed by spots of low brightness. After further rising the imaging voltage up to 5.9 kV, the surface structure begins to recover its initial form, as can be seen in Fig. 5f–g. As compared to a clean Pt nanocrystal, depicted in Fig. 5h, the rough surface presents only (001) plane and the enlarged dark {111} facets are well visible. Interestingly, the striped-like structures, reminding MR type reconstructions, are observable between the (011) facets and vicinal (012) facets. Furthermore, the micrographs also reveal MR type reconstructions between {111} and (001) facets (marked by a red arrow in Fig. 5e). Obtaining a clean and smooth surface of the nanoparticle requires an imaging voltage higher than 5.9 kV to trigger the field evaporation (Fig. 5h). It must be noted that higher doses of oxygen (> 100 L) produce a FE pattern similar to that observed in Fig. 5b. This observation indicates the metastability of the reconstruction under these specific reaction conditions. The transformations are striking considering the conditions applied (< 30 L). At this point, the effect of the electric field during the oxygen exposure cannot be neglected, even though its relatively low intensity in FEM (~ 2–3 V nm−1). The reconstructions obtained at low doses of oxygen in presence of an electric field are similar to those obtained at low dose of oxygen in absence of electric field: they both present only {111} and (001) facets. As discussed earlier, in absence of electric field, higher oxygen doses induce further changes. Contrarily, in presence of the electric field, higher oxygen doses do not cause any further changes to the structure of the crystal. These differences at high oxygen doses are associated with the presence of the electric field and may be an indication that the electric field may arrest the formation of the flat {113} planes. Further work is however necessary to confirm this hypothesis. The presence of surface oxide formation could also be further explored by using atom probe tomography by exposing Pt samples for longer time at 700 K, in presence of a low electric field, and determine the possible formation of Pt oxides deeper in the sample.

Fig. 5

a FI image of a clean Pt surface (imaging conditions: V = 5.3 kV, T = 74 K); b FE image of the same clean Pt surface. The main four bright spots correspond to the {210} facets (imaging conditions: V = 0.54 kV, T = 75 K); c modification of the FE pattern after O2 exposition at \({P_{\text{O}{_2}}}\) = 5.0 × 10−5 Pa (imaging conditions: V = 0.46 kV, T = 700 K); d FI image of the reconstructed Pt surface induced by O2 exposure. The square-shape and the presence of missing rows on {113} regions can be observed (imaging conditions: V = 5.1 kV, T = 117 K); e–g FI images after a slow field evaporation causing the missing rows structures disappearance and recovering of the initial structure by increasing the voltage (imaging conditions: T = 70 K and e V = 5.2 kV; f V = 5.8 kV; g V = 5.9 kV); h oxygen exposition leads to the formation of a refractory layer. The evaporation of this layer by field effect, finally reveals a clean and smooth surface, available for a new cycle of oxygen dosing (imaging conditions: V = 5.0 kV, T = 67 K)

4 Conclusions

This work provides new insights for a better understanding of the morphological and structural changes induced by oxygen dosing on platinum model catalysts. When a Pt tip-sample is exposed to low doses (< 10 L) of oxygen at 300 K and subsequently annealed under UHV conditions at 700 K, FIM imaging of the crystal did not reveal noticeable modifications as compared to an annealed Pt tip in absence of oxygen: only low index {100} and {111} planes are made visible. At higher oxygen exposures, a surface oxide layer can be formed and causes drastic changes of the surface structure: {100}, {111} and {113} facets extend to the detriment of vicinal orientations upon annealing at T = 700 K for 100 s. Experiments were performed in presence of an electric field during the oxygen dosing. These experiments also revealed field emission pattern changes which are likely to be the consequence of morphological changes. However, within these experimental conditions, i.e. temperature of 700 K and in presence of electric field of 3 V nm−1, the presence of subsurface oxygen species and their influence on the overall brightness is not excluded. Further FIM characterisation highlights the presence of similar reconstructions in the case of low oxygen exposures (< 50 L). Higher exposures (> 100 L) do not lead to any significant changes of the FE pattern, which indicates that the reconstructions shown here, under the reported experimental conditions, are on the way towards the final equilibrium shape.

The obtained results are relevant to understand the reactivity and stability of PGMs that operate under oxygen-rich conditions. The formation of the oxide layers on the Pt surface is evident from our experiments, causing modifications of the tip apex. Further experiments are in progress to determine the extent and the stoichiometry of the oxide layer and how the catalytic reactions proceed in its presence.

Notes

Acknowledgements

S.O., C.B., R.B. and T.V.d.B. thank Wallonia-Brussels International for the Bilateral Cooperation Agreement, and the Bilateral Cooperation between the Fonds de la Recherche Scientifique (F.R.S.-FNRS) and the Polish Academy of Sciences (PAN). S.V.L. and C.B. thank the F.R.S.-FNRS for financial support (PhD grant from FRIA and Postdoctoral fellowship from FNRS, respectively). This work was supported by a research grant from University of Wroclaw (No. 1425/M/FD/15).

References

  1. 1.
    King DA, Woodruff DP (1982) The chemical physics of solid surfaces and heterogeneous catalysis, vol 4. Elsevier, AmsterdamGoogle Scholar
  2. 2.
    Freund HJ (2002) Surf Sci 500:271–299CrossRefGoogle Scholar
  3. 3.
    Arblaster JW (2005) Platin Met Rev 49:141–149CrossRefGoogle Scholar
  4. 4.
    Seriani N, Jin Z, Pompe W, Colombi Ciacchi L (2007) Phys Rev B 76:155421CrossRefGoogle Scholar
  5. 5.
    Bernhardt TM, Heiz U, Landman U (2007) Nanocatalysis. Springer, BerlinGoogle Scholar
  6. 6.
    Croy JR, Mostafa S, Liu J, Sohn YH, Heinrich H, Roldan Cuenya B (2007) Catal Lett 191:209–216CrossRefGoogle Scholar
  7. 7.
    Mostafa S, Croy JR, Heinrich H, Roldan Cuenya B (2009) Appl Catal A 366:353–362CrossRefGoogle Scholar
  8. 8.
    Rinnemo M (1997) Surf Sci 376:297–309CrossRefGoogle Scholar
  9. 9.
    McCrea KR, Parker JS, Somorjai GA (2002) J Phys Chem B 106:10854–10863CrossRefGoogle Scholar
  10. 10.
    Engel T, Ertl G (1979) Adv Catal 28:1–78Google Scholar
  11. 11.
    Laidler KJ, Meiser JH (1982) Physical chemistry, Benjamin/Cummings Pub. Co., Menlo ParkGoogle Scholar
  12. 12.
    Sandert M, Imbihl R, Schuster R, Barth JV, Ertl G (1992) Surf Sci 271:159–169CrossRefGoogle Scholar
  13. 13.
    Parkinson CR, Walker M, McConville CF (2003) Surf Sci 545:19–33CrossRefGoogle Scholar
  14. 14.
    Miller DJ, Öberg H, Kaya S, Sanchez Casalongue H, Friebel D, Anniyev T, Ogasawara H, Bluhm H, Pettersson LGM, Nilsson A (2011) Phys Rev Lett 107:195502CrossRefPubMedGoogle Scholar
  15. 15.
    Devarajan SP, Hinojosa JA Jr, Weaver JF (2008) Surf Sci 602:3116–3124CrossRefGoogle Scholar
  16. 16.
    Ertl G, Neumann M, Streit KM (1977) Surf Sci 64:393–410CrossRefGoogle Scholar
  17. 17.
    Walker AV, Klötzer B, King DA (1998) J Chem Phys 109:6879–6888CrossRefGoogle Scholar
  18. 18.
    Freyer N, Kiskinova M, Pirug G, Bonze HP (1986) Surf Sci 166:206–220CrossRefGoogle Scholar
  19. 19.
    Helveg S, Li WX, Bartelt NC, Horch S, Lægsgaard E, Hammer B, Besenbacher F (2007) Phys Rev Lett 98:115501CrossRefPubMedGoogle Scholar
  20. 20.
    Li WX, Österlund L, Vestergaard EK, Vang RT, Matthiesen J, Pedersen TM, Lægsgaard E, Hammer B, Besenbacher F (2004) Phys Rev Lett 93:146104CrossRefPubMedGoogle Scholar
  21. 21.
    Somorjai GA, Aliaga C (2010) Langmuir 26:16190–16203CrossRefPubMedGoogle Scholar
  22. 22.
    Shao M, Peles A, Shoemaker K (2011) Nano Lett 11:3714–03719CrossRefPubMedGoogle Scholar
  23. 23.
    Tian N, Zhou ZY, Sun SG (2008) J Phys Chem C 112:19801–19817CrossRefGoogle Scholar
  24. 24.
    Somorjai GA, Park JY (2008) Chem Soc Rev 37:2155–2162CrossRefPubMedGoogle Scholar
  25. 25.
    Jin M, Zhang H, Xie Z, Xia Y (2012) Energy Environ Sci 5:6352–6357CrossRefGoogle Scholar
  26. 26.
    Zhang H, Jin M, Xia Y (2012) Angew Chem Int Ed 51:7656–7673CrossRefGoogle Scholar
  27. 27.
    Madey TE, Chen W, Wang H, Kaghazchi P, Jacob T (2008) Chem Soc Rev 37:2310–2327CrossRefPubMedGoogle Scholar
  28. 28.
    Yoshida H, Matsuura K, Kuwauchi Y, Kohno H, Shimada S, Haruta M, Takeda S (2011) Appl Phys Express 4:065001CrossRefGoogle Scholar
  29. 29.
    Medvedev KV, Suchorski Y, Voss C, Visart de Bocarmé T, Bär T, Kruse N (1998) Langmuir 14:6151–6157CrossRefGoogle Scholar
  30. 30.
    Dicke J, Rotermund HH, Lauterbach J (2000) Surf Sci 454–456:352–357CrossRefGoogle Scholar
  31. 31.
    Sadeghi P, Dunphy K, Punckt C, Rotermund HH (2012) J Phys Chem C 116:4686–4691CrossRefGoogle Scholar
  32. 32.
    Voss C, Kruse N (1995) Appl Surf Sci 87/88:134–139CrossRefGoogle Scholar
  33. 33.
    Barroo C, Gilis N, Lambeets SV, Devred F, Visart de Bocarmé T (2014) Appl Surf Sci 304:2–10CrossRefGoogle Scholar
  34. 34.
    Gorodetskii VV, Elokhina VI, Bakker JW, Nieuwenhuys BE (2005) Catal Today 105:183–205CrossRefGoogle Scholar
  35. 35.
    Genty E, Jacobs L, Visart de Bocarmé T, Barroo C (2017) Catalysts 7(5):134CrossRefGoogle Scholar
  36. 36.
    Bär T, Visart de Bocarmé T, Kruse N (2000) Surf Sci 454–456:240–245CrossRefGoogle Scholar
  37. 37.
    Bryl R, Olewicz T, Visart de Bocarmé T, Kruse N (2010) J Phys Chem C 114:2220–2226CrossRefGoogle Scholar
  38. 38.
    Bryl R, Olewicz T, Visart de Bocarmé T, Kruse N (2011) J Phys Chem C 115:2761–2768CrossRefGoogle Scholar
  39. 39.
    McEwen JS, Gaspard P, De Decker Y, Barroo C, Visart de Bocarmé T, Kruse N (2010) Langmuir 26:16381–16391CrossRefPubMedGoogle Scholar
  40. 40.
    Barroo C, De Decker Y, Visart de Bocarmé T, Kruse N (2014) J Phys Chem C 118:6839–6846CrossRefGoogle Scholar
  41. 41.
    Müller EW, Tsong TT (1969) Field ion microscopy: principles and applications. Elsevier, New YorkCrossRefGoogle Scholar
  42. 42.
    Bagot PAJ, Cerezo A, Smith GDW (2007) Surf Sci 601:2245–2255CrossRefGoogle Scholar
  43. 43.
    Visart de Bocarmé T, Kruse N (2001) Top Catal 14:35–42CrossRefGoogle Scholar
  44. 44.
    Chen Q, Richardson NV (2003) Prog Surf Sci 73:59–77CrossRefGoogle Scholar
  45. 45.
    Seriani N, Mittendorfer F (2008) J Phys Condens Matter 20:184023CrossRefGoogle Scholar
  46. 46.
    Voss C, Kruse N (1998) Surf Sci 409:252–257CrossRefGoogle Scholar
  47. 47.
    Kruse N, Gaussmann A (1993) Appl Surf Sci 67:160–165CrossRefGoogle Scholar
  48. 48.
    Yamanaka T, Xue QK, Kimura K, Matsushima T, Hasegawa Y, Sakura T (2000) Jpn J Appl Phys 39:3562–3565CrossRefGoogle Scholar
  49. 49.
    Jenkins SJ (2001) Surf Sci 494:59–65CrossRefGoogle Scholar
  50. 50.
    Zhu T, Sun SG, van Santen RA, Hensen EJM (2013) J Phys Chem C 117:11251–11257CrossRefGoogle Scholar
  51. 51.
    Foiles SM (1987) Surf Sci 191:L779–L786CrossRefGoogle Scholar
  52. 52.
    Lin RJ, Fu TY (2012) Surf Interface Anal 44:658–661CrossRefGoogle Scholar
  53. 53.
    Yamanaka T, Matsushima T, Tanaka SI, Kamada M (1996) Surf Sci 349:119–128CrossRefGoogle Scholar
  54. 54.
    Voss C, Gaussmann A, Kruse N (1993) Appl Surf Sci 67:142–146CrossRefGoogle Scholar
  55. 55.
    Wang T, Schmidt LD (1981) J Catal 71:411–422CrossRefGoogle Scholar
  56. 56.
    Li T, Marquis EA, Bagot PAJ, Tsang SCE, Smith GDW (2011) Catal Today 175:552–557CrossRefGoogle Scholar
  57. 57.
    Suchorski Y (1998) Ultramicroscopy 73:139–145CrossRefGoogle Scholar
  58. 58.
    Muller O, Roy R (1968) J Less-Common Met 16:129–146CrossRefGoogle Scholar
  59. 59.
    Punnoose A, Seehra MS, Wende I (2001) Fuel Process Technol 74:33–47CrossRefGoogle Scholar
  60. 60.
    Wang CB, Lin HK, Hsu SN, Huang TH, Chiu HC (2002) J Mol Catal A 188:201–208CrossRefGoogle Scholar
  61. 61.
    Seriani N, Pompe W, Ciacchi LC (2006) J Phys Chem B 110:14860–14869CrossRefPubMedGoogle Scholar
  62. 62.
    Weaver JF (2013) Chem Rev 113:4164–4215CrossRefPubMedGoogle Scholar
  63. 63.
    Ono LK, Yuan B, Heinrich H, Roldan Cuenya B (2010) J Phys Chem C 114:22119–22133CrossRefGoogle Scholar
  64. 64.
    Samsonov GV (1982) The oxide handbook, 2nd edn. Plenum Publishing Corporation, New YorkCrossRefGoogle Scholar
  65. 65.
    Weaver JF, Kan HH, Shumbera RB (2008) J Phys: Condens Matter 20:184015Google Scholar
  66. 66.
    Ellinger C, Stierle A, Robinson IK, Nefedov A, Dosch HJ (2008) Phys Condens Matter 20:184013CrossRefGoogle Scholar
  67. 67.
    Moors M, Visart de Bocarmé T, Kruse N (2007) Catal Today 124:61–70CrossRefGoogle Scholar
  68. 68.
    Voss C, Kruse N (1998) Surf Sci 416:L1114–L1117CrossRefGoogle Scholar
  69. 69.
    Visart de Bocarmé T, Chau TD, Kruse N (2007) Surf Interface Anal 39:166–171CrossRefGoogle Scholar
  70. 70.
    Derry GN, Ross PNA (1985) J Chem Phys 82:2772–2778CrossRefGoogle Scholar
  71. 71.
    Lambeets SV, Barroo C, Owczarek S, Genty E, Gilis N, Kruse N, Visart de Bocarmé T (2017) J Phys Chem C 121:16238–16249CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media, LLC, part of Springer Nature 2018

Authors and Affiliations

  1. 1.Institute of Experimental PhysicsUniversity of WrocławWrocławPoland
  2. 2.Chemical Physics of Materials and Catalysis (CPMCT), Faculty of SciencesUniversité Libre de Bruxelles (ULB)BrusselsBelgium
  3. 3.Interdisciplinary Center for Nonlinear Phenomena and Complex Systems (CENOLI), Faculty of SciencesUniversité Libre de Bruxelles (ULB)BrusselsBelgium

Personalised recommendations