betaFeedback

WorksPhysics and Astronomy

Transient Surface Oxides Form on Pt(111) - But Vanish During Ammonia Oxidation

Abstract

Ammonia oxidation on platinum catalysts is pivotal for industrial nitric acid production and environmental abatement, yet the role of surface oxides in this process remains debated. Using operando surface X-ray diffraction (SXRD), crystal truncation rod (CTR) analysis, and near-ambient pressure X-ray photoelectron spectroscopy (NAP-XPS), we reveal that Pt(111) does not form stable surface oxides under ammonia oxidation conditions. Instead, transient hexagonal monolayers and a Pt(111)-(8x8) superstructure emerge under oxygen-rich atmospheres and above the catalyst light-off temperature, but vanish upon ammonia exposure. Real-time mass spectrometry and NAP-XPS demonstrate that the reaction proceeds via a Langmuir-Hinshelwood mechanism, where adsorbed NHx and O species availability dictate selectivity toward NO or N2. Reducing the oxygen pressure by an order of magnitude slows the kinetics of oxide growth, only detected after 24 hr, and facilitated by transient and precursor structures.

Introduction

Ammonia oxidation on platinum catalysts is central to nitric acid production by the Ostwald process, where high selectivity toward nitric oxide (NO) is required [1]. The difficulty is that NH3_3 oxidation proceeds through several competing routes that also generate N2_2O and N2_2. In practice, maximizing NO formation requires high temperature, oxygen-rich feeds (high pO2/pNH3p_{\mathrm{O}_2}/p_{\mathrm{NH}_3}), and a surface that remains catalytically active under strongly oxidizing conditions [1, 2]. Beyond nitric acid synthesis, this reaction is environmentally and technologically relevant. Selectivity control between NO and N2_2 affects nitrogen-oxide emissions and enables routes for ammonia abatement [3, 4, 5, 6, 7]. In addition, renewed interest in ammonia as an energy carrier motivates a mechanistic understanding of its oxidation and decomposition chemistry at catalytic interfaces [8, 9]. From a surface-science standpoint, ammonia oxidation is also an attractive model reaction: it couples strong adsorbate interactions with oxidizing environments, and it drives structural and chemical transformations that are both facet-dependent and reversible. Capturing these transformations under reaction conditions is essential, because the active surface is not a static termination of a bulk crystal but a dynamic state that can select pathways and control selectivity, as clearly identified in a recent study of the Pt(100) surface [10]. Industrial catalysts have long relied on Pt–Rh gauges, where Rh additions (typically up to 10%) improve lifetime and reduce platinum loss [2, 11, 12]. Under operating conditions, oxygen-rich and high-temperature environments promote substantial surface evolution, including restructuring and roughening, with direct consequences for activity and selectivity [1, 10, 13]. On Pt–Rh(100), reversible relaxation transitions between nitrogen- and oxygen-rich feeds were reported and linked to electronic-structure changes that affect adsorption [14, 15, 16, 17]. Rh-containing oxides have further been proposed to act as oxygen reservoirs and to stabilize volatile platinum oxides [1, 14], recent work also showed that Rh strongly impinges on the formation of platinum oxides on Pt(100) that are associated with material loss [1, 10]. In particular, the role of Pt-oxide surface phases during ammonia oxidation, and how alloying shifts their stability and kinetics, are still not resolved at the level of detailed structure. Progress on these questions requires operando structural probes that are sensitive to the topmost atomic layers. A key operando study combining surface X-ray diffraction (SXRD) and Bragg coherent diffraction imaging (BCDI) showed that Pt microcrystals reconstruct under oxygen-rich ammonia oxidation, with increased exposure of {111} and {100} facets, consistent with facet-dependent stabilization of oxidized surfaces predicted by theory [18, 19]. While BCDI provides access to particle shape and internal strain fields, it samples a limited region of reciprocal space and does not directly resolve thin crystalline surface oxides or reconstructions. In contrast, SXRD offers a unique and powerful route to surface structure under realistic pressures and temperatures: crystal truncation rods (CTRs) quantify relaxations and roughness at the surface, while superstructure features reveal reconstructions and ordered oxide-related phases with surface specificity. These capabilities make SXRD particularly well suited to link atomic-scale surface structure to catalytic performance in environments where microscopy and ultra-high-vacuum approaches are constrained. In this work, we present an operando investigation of Pt(111) under industrially relevant ammonia oxidation conditions, following our previous study on Pt(100) [10]. In-plane reciprocal-space maps track surface reconstructions and the emergence of oxide-related superstructures. CTR measurements quantify surface relaxation and roughness, and dedicated superstructure rods probe the out-of-plane structure of newly observed surface phases. Complementary near-ambient pressure X-ray photoelectron spectroscopy (NAP-XPS) provides chemical information through O 1s, N 1s, and Pt 4f core-level spectra. By systematically varying the ammonia-to-oxygen ratio and simultaneously monitoring products with real-time residual gas analysis (RGA), we relate surface structure and surface chemistry to catalytic selectivity on Pt(111), with the goal of clarifying which dynamic surface states control the reaction mechanism.

Results

Experimental conditions

The ammonia oxidation on Pt(111) was first measured operando using SXRD, and then with NAP-XPS. Both experiments were performed at 450 °C, a temperature above the catalyst light-off (Figure A.1 - [10, 18]), and at different gas atmosphere as summarised in Table 1. Extended details regarding the experimental setups and sample preparation procedure are available in the Methods section. Argon was used before (1) and after (6) the oxidation cycle to evaluate the reversibility of the catalyst surface. To ensure that any surface reconstruction/relaxation effects are attributed to the simultaneous presence of ammonia and oxygen in the reactor, the sample was first exposed to a high-oxygen atmosphere (2), monitoring potential surface/bulk oxide growth. Introducing ammonia (3) facilitated the correlation of such oxides with the catalyst selectivity, while also studying their stability under reaction conditions. Two different partial pressures of oxygen were employed alongside the same partial pressure of ammonia to investigate the influence of the pO2/pNH3p_{\mathrm{O}_2}/p_{\mathrm{NH}_3} ratio on the catalyst structure, surface species, and selectivity (3, 4). Removing oxygen to keep only ammonia in the reactor (5) enabled the separation of the effect of ammonia’s presence from the combined presence of ammonia and oxygen. Finally, the sample surface was exposed to a reduced pressure of oxygen, to make certain that the observation, when lowering the pO2/pNH3p_{\mathrm{O}_2}/p_{\mathrm{NH}_3} ratio, are due to the combined presence of the reactants. All gases have similar thermal conductivity (Table 4).

OrderTechniquesTotal Pressure (mbar)Ar (mbar)O₂ (mbar)NH₃ (mbar)Targeted Information
1SXRD, NAP-XPS500 / 1500 / 10 / 00 / 0Catalyst state without reactants
2SXRD, NAP-XPS500 / 8.8420 / 080 / 8.80 / 0Growth of surface oxides
3SXRD, NAP-XPS500 / 9.9410 / 080 / 8.810 / 1.1Influence of (NH₃/O₂) ratio and impact on/of potential surface oxides
4SXRD, NAP-XPS500 / 1.65485 / 05 / 0.5510 / 1.1Ammonia adsorption
5SXRD, NAP-XPS500 / 1.1490 / 00 / 010 / 1.1
6SXRD, NAP-XPS500 / 1500 / 10 / 00 / 0Returning to pristine state
7SXRD, NAP-XPS500 / 0.55495 / 05 / 0.550 / 0Growth of surface oxides

Table 1. Combined experimental conditions for SXRD and NAP-XPS. All SXRD experiments were conducted at 500 mbar; NAP-XPS experiments were performed at lower total pressure to increase the electron counts.

StructureInterplanar spacing (Å)Oxygen pressure
80 mbar5 mbar
Time since gas introduction (end of measurement)
03h2309h5300h3404h0308h0015h5722h5624h0825h4326h3627h28
Pt(111)-(6 × 6)-R±8.8°3.01 ± 0.0100000000000
Pt(111)-(6 × 6)-R±8.8°2.92 ± 0.0300xxxxxx0xx
Pt(111)-(6 × 6)-R±8.8°2.87 ± 0.02xx000000000
Pt(111)-(6 × 6)-R±8.8°2.79 ± 0.07xxxx0000000
Pt(111)-(8 × 8)2.69 ± 0.02x0xxxxx0000
Pt(111)-(8 × 8)1.53 ± 0.000nvnv0nvnvnvnvnvnvnv

Table 4. Interplanar spacings computed from signals observed during in-plane reciprocal maps, for different oxygen pressure and exposition times for Pt(111). The three markers (0, x, nv) correspond to observed, non-observed, and non visible signals (i.e. not in the area spanned by the map). The errors on the interplanar spacings are computed by considering the positions of similar peaks in q-space.

Surface oxidation at high oxygen partial pressure

The first reciprocal space in-plane map is under inert atmosphere (Figure 1a), collected after sputtering and annealing under high vacuum (see Methods). All reciprocal space maps in this study are measured at L=0.15L = 0.15 and at 450 °C. The (110) and (21ˉ\bar{1}0) Bragg peaks are visible, together with the intersection of the [0, 1, LL], [0, 2, LL], [1, 1ˉ\bar{1}, LL], [1, 0, LL], and [2, 0, LL] CTRs with the [HH, KK, 0] plane. The sample alignment was corrected after the first map, as it seems that thermal stability was not yet reached from the decreasing intensity gradient (the measurement starting from the reciprocal space centre). New peaks are first observed during surface oxidation at pO2=80p_{\mathrm{O}_2} = 80 mbar (Figure 1b). All corresponding interplanar spacings are given in Table 4. The angle between the peaks circled in grey and white, and the peaks circled in red and purple is equal to 60°, proving the existence of two hexagonal surface structures, each one rotated by ±8.8∘\pm8.8^\circ with respect to the Pt(111) surface unit cell and with a larger in-plane lattice parameter (drawn in green and grey in Figure 1b-Figure 1c). However, no apparent second order peak can be linked to those rotated hexagonal structures, the related intensity is possibly too low at high scattering vector to be detected. The closest rotated and commensurate superstructure is Pt(111)-(6×6)(6 \times6)-R±8.8∘\pm8.8^\circ.

Figure 1. Reciprocal space in-plane maps collected under different atmospheres, computed using the hexagonal lattice of Pt(111). The measurement of the large reciprocal space in-plane map takes about 3 h30 min, and 1 h15 min for the small map. Both measurements under oxygen share the same colormap.

Interestingly, summing two vectors each going from the centre of the reciprocal space to either the white or purple circled peaks yields the position of the black circled peak in Figure 1b. The angle between both vectors is then equal to 42.6°, equivalent to 137.4° in real space. No other peak linked to this unit cell can be identified in this in-plane map. A second map was measured in a smaller region of the reciprocal space 9 h30 min after the introduction of oxygen in the cell. Three additional peaks (circled in black in Figure 1c) describe an epitaxied hexagonal lattice with ∣ahex∣=3.12 A˚|a_{\mathrm{hex}}| = 3.12\,\text{\AA}, the closest commensurate superstructure is Pt(111)-(8×8)(8 \times8). The grey and purple circled peaks have vanished from the reciprocal map, later measurements show that this is due to alignment issues during the oxidation reaction.

Super-structure rods (SSRs) were collected on the peaks belonging to the Pt(111)-(8×8)(8 \times8) structure at [HH, KK] = [0.89, 0], and to the rotated hexagonal structures at [HH, KK] = [0.15, 0.74], to probe their out-of-plane order. A second SSR was measured at [HH, KK] = [1.78, 0] to probe for a potential second-order peak associated to the Pt(111)-(8×8)(8 \times8) structure, approximately 10 h after the measurement at [HH, KK] = [0.89, 0]. The integrated intensity is presented in Figure 2a. The intensity distribution for both Pt(111)-(8×8)(8 \times8) measurements suggests the presence of a multi-layer structure, in this case a surface oxide, while the small peak at L≈0.2L \approx0.2 followed by a constant intensity for the SSR measured at [HH, KK] = [−0.15, −0.74] is characteristic of a monolayer / chemisorbed specie [20]. Reflectivity curves measured in a specular geometry (Figure 2b) were fitted using GenX [21, 22] (see Methods). The position of the oscillations can be linked to the thickness of the oxide layer (toxt_{\mathrm{ox}}), 14.42 Å and 14.24 Å thick in the first and second reflectivity curves, i.e. stable between both measurements.

Figure 2. Super-structure rods (SSRs) measurements for three different [H,K][H, K] positions at pAr=420p_{\mathrm{Ar}} = 420 mbar and pO2=80p_{\mathrm{O}_2} = 80 mbar (a) δt\delta t designates the elapsed time since the introduction of oxygen, until the end of the measurement. The empty circles correspond to the Pt(111)-(8 × 8) structure. X-ray reflectivity curves at pAr=420p_{\mathrm{Ar}} = 420 mbar and pO2=80p_{\mathrm{O}_2} = 80 mbar for different exposure times (b) Curves fitted using GenX are shown as lines. δt\delta t designates the time elapsed since the introduction of oxygen. ρox\rho_{\mathrm{ox}}, toxt_{\mathrm{ox}}, and σox\sigma_{\mathrm{ox}} are the oxide density, thickness, and root mean square roughness. σsub\sigma_{\mathrm{sub}} is the substrate root mean square roughness.

Absence of surface reconstructions and roughness decrease during ammonia oxidation

Ammonia was introduced in the reactor after surface oxidation to investigate the stability of the platinum oxide structures, as well as the potential presence of additional surface structures or surface reconstructions during the reaction (Figure 3a-Figure 3b). None of the previously observed features were measured at pO2/pNH3=8p_{\mathrm{O}_2}/p_{\mathrm{NH}_3}=8 or pO2/pNH3=0.5p_{\mathrm{O}_2}/p_{\mathrm{NH}_3}=0.5, while no additional peaks emerged, the reaction removes the oxide layer grown during initial surface oxidation. A gradual increase of NO is observed in the first hours following the introduction of ammonia in the reactor (Figure B.3), the presence of the surface oxide at the beginning of the condition possible influencing the product selectivity. All reciprocal space in-plane maps during the reaction cycle and back to argon showed a bulk-terminated surface (Figure 3a-Figure 3b-Figure 3c-Figure 3d) that resembles the first map also measured under argon before surface oxidation (Figure 1a), proving that the oxidation cycle reproduces the same surface termination. Complementary X-ray reflectivity curves are presented in Figure 4. In accordance with the results of the in-plane reciprocal space maps in Figure 3, no oxide layer was used in the fitting model. The surface roughness increased to 2.80 Å following the introduction of ammonia (Figure 4b), after reaching 2.14 Å when pO2=80p_{\mathrm{O}_2}=80 mbar (Figure 2). Since the X-ray reflectivity curves are measured at the start of each condition, the high roughness when pO2/pNH3=8p_{\mathrm{O}_2}/p_{\mathrm{NH}_3}=8 can be linked to the removal of the surface oxides. Lowering the amount of oxygen decreases the roughness (Figure 4b). The largest decrease is observed when switching the pO2/pNH3p_{\mathrm{O}_2}/p_{\mathrm{NH}_3} ratio from 8 to 0.5, i.e. during reacting conditions.

Reciprocal space in-plane maps collected under different atmospheres

Figure 3. Reciprocal space in-plane maps collected under different atmospheres, computed using the hexagonal lattice of Pt(111).

X-ray reflectivity curves under different atmospheres during ammonia oxidation

Figure 4. X-ray reflectivity curves under different atmospheres during ammonia oxidation. Curves fitted using GenX are shown as lines. σsub\sigma_{\mathrm{sub}} is the substrate root mean square roughness.

Surface oxidation at low oxygen partial pressure

At the end of the ammonia oxidation cycle, the sample was cleaned with two sputtering and annealing cycles and put under 5 mbar of oxygen to monitor the growth of surface oxides at lower oxygen pressure. Large and small reciprocal space in-plane maps were measured to have the best compromise between a higher temporal resolution of the oxide growth, and the possibility to detect other peaks from corresponding surface unit cells. The larger reciprocal space in-plane maps are shown in Figure 5. The first map starts directly after the stabilisation of the oxygen pressure at 5 mbar in the reactor cell. Low intensity signals are instantly detected (Figure 5a), corresponding to those previously measured shortly after setting a pressure of 80 mbar in the cell (Figure 1c), proving that ammonia effectively impinges on their formation at both reactant pressure ratios. Circles of the same colour are used to indicate peaks at similar positions, the related interplanar spacing is also summarised in Table 4. Some peaks split in a similar pattern, exhibiting three distinct signals around a more diffuse region (clearly visible for the red circled peak in Figure 5d). This splitting can be interpreted as the signature of a Pt(111) surface covered by different domains that exhibit a similar hexagonal in-plane symmetry. These unit cells are on average rotated by ±8.8∘\pm8.8^{\circ} with respect to the Pt(111) surface unit cell, with a higher magnitude of the in-plane lattice parameter, following the same Pt(111)-(6 ×\times 6)-±8.8∘\pm8.8^{\circ} unit cells observed under an oxygen pressure of 80 mbar (Figure 1c). No corresponding second order peak is observed. The grey circled peak becomes fully visible after 4 h, which confirms that the absence of peak in Figure 1c come from alignments problems. A peak can be observed at [H,K]=[0.89,0.89][\mathrm{H}, \mathrm{K}] = [0.89, 0.89] in the large reciprocal space in-plane map collected between 41 min and 4 h 3 min (circled in black in Figure 5b). Its position is the same as in the first large reciprocal space in-plane map measured at pO2=80p_{\mathrm{O}_2} = 80 mbar (Figure 1c), and corresponds to the Pt(111)-(8 ×\times 8) commensurate superstructure. However, no peaks at [0.89,0,0][0.89, 0, 0] and [0,0.89,0][0, 0.89, 0] can yet be detected, before (Figure 1b), and after (Figure 1c) measuring this signal. Such peaks are only detected in the large maps after ≈24\approx24 h of measurements (described by the hexagonal lattice drawn with black dashed lines in Figure 1f). As mentioned before, it is possible to draw a surface unit cell in Figure 1b that includes the peak at [0.89,0.89][0.89, 0.89] together with the white and purple circled peaks. The area sampled during the measurement was extended in the last two maps to observe more signals around the (110) region (Figure 5e-Figure 5f). For each map, the purple and red circled peaks around the Pt(111)-(8 ×\times 8) peaks are separated by 60°, likewise for the grey and white circled peaks. In the last two maps, four doublet of peaks are visible, all split in a similar pattern. During this set of measurement, the grey circled peak did not disappear in contrast with the measurements carried out at pO2=80p_{\mathrm{O}_2} = 80 mbar, which furthermore supports the existence of rotated hexagonal structures that coexist with the Pt(111)-(8 ×\times 8) structure.

Figure 5. Large reciprocal space in-plane maps collected at pAr=495p_{\mathrm{Ar}}=495 mbar and pO2=5p_{\mathrm{O}_2}=5 mbar for different exposure times.

Smaller maps taken with a shorter time interval are presented in Figure 6, a Pt(111)-(8 ×\times 8) peak is detected after ≈23\approx23 h of measurements at [0.89,0.89][0.89, 0.89] (circled in black). The same peak was measured after only 13 hour when pO2=80p_{\mathrm{O}_2} = 80 mbar (Figure 1d), highlighting the importance of the oxygen chemical potential in the surface oxidation. Finally, X-ray reflectivity curves were measured before, 14 h, and 23 h after the introduction of oxygen (Figure 7). The last reflectivity curve was measured just before the detection of the Pt(111)-(8 ×\times 8) structure with SXRD (Figure 6f). Following the information gathered by SXRD, a homogeneous layer on top of Pt(111) was introduced in the fitting model for the measurements at 14 h and 23 h, to represent the rotated structures visible in Figure 6. The oxide layer was necessary for the 23 h curve to achieve a good fit, but only slightly improved the result for the 14 h curve. The layer thickness and density are difficult to estimate since no oscillations are visible, possibly too far in δ\delta. However, this confirms that the oscillations measured at pO2=80p_{\mathrm{O}_2} = 80 mbar in Figure 2 are linked to the Pt(111)-(8 ×\times 8) structure and not to the rotated hexagonal structures, and that a longer exposure time is needed to reach the same thickness. This also further confirms that the red and white peaks are linked to monolayers (also confirmed by the lack of peak in the related SSRs in Figure B.2). A transition in terms of oxide roughness and density is observed between the orange and green curves. Interestingly, the final density is equal to that retrieved at pO2=80p_{\mathrm{O}_2} = 80 mbar (Figure 2), but without the same layer thickness, and roughness. This effect is clearly visible in the reflectivity curve with a decrease of intensity starting at low δ\delta angles. The substrate roughness increases with the elapsed time since the introduction of oxygen, coherent with the creation of a surface oxide.

Figure 6. Small reciprocal space in-plane maps collected at pAr=495p_{\mathrm{Ar}} = 495 mbar and pO2=5p_{\mathrm{O}_2} = 5 mbar for different exposure times.

Figure 7. X-ray reflectivity curves measured in a specular geometry (full lines)at pAr=495p_{\mathrm{Ar}} = 495 mbar and pO2=5p_{\mathrm{O}_2} = 5 mbar for different exposure times. Curves fitted using GenX are shown as lines. δt\delta t designates the elapsed time since the introduction of oxygen. ρox\rho_{\mathrm{ox}}, toxt_{\mathrm{ox}}, and σox\sigma_{\mathrm{ox}} are the oxide density, thickness, and root mean square roughness. σsub\sigma_{\mathrm{sub}} is the substrate root mean square roughness.

Surface roughness and surface relaxation effects

The (2ˉ1Lˉ)(\bar{2}1\bar{L}), (10L)(10L) and (11L)(11L) crystal truncation rods intensities are presented in Figure 8. The evolution of the surface roughness when changing condition is visible from the variation of intensity minimum near L=1.5L = 1.5 (or L=2.5L = 2.5 for the (10L)(10L) CTR). Overall, the intensity distribution is symmetric around those minima, the surface layers lattice parameters are thus close to their bulk values [23].

Figure 8. Evolution of crystal truncation rods under different atmospheres. Only the (21ˉL)(2\bar{1}L) CTR was measured under 500 mbar of argon before the oxidation cycle.

Surface species

The mass spectrometer signals from the NAP-XPS experiment are available in Figure C.4. Because the XPS mass-spectrometer measurements are taken from gases pumped trough the inlet of the XPS analyser, which is closer to the sample surface, it is more representative of the catalyst surface activity. The mass spectrometry data from the SXRD experiment instead reflect the average composition of the gas in the reactor. The production of N2\mathrm{N}_2 and H2\mathrm{H}_2 under pure ammonia measured during the XPS experiment are therefore more visible on average than with the SXRD setup (Figure B.3). The high pO2/pNH3p_{\mathrm{O}_2}/p_{\mathrm{NH}_3} ratio favours the production of NO as expected [1], accompanied by water, N2_2O and N2_2 in low amounts. Despite the relative high amount of oxygen in the reactor, ammonia is still detected. The complete oxidation reaction may be limited by the availability of active sites, the transportation of reactants towards the catalyst surface or the relatively low temperature compared to industrial conditions (>700 ∘C>700\,^{\circ}\mathrm{C}) which can for example impact the active sites turnover frequency. Reducing the pO2/pNH3p_{\mathrm{O}_2}/p_{\mathrm{NH}_3} ratio to 0.5 by lowering the oxygen pressure shifts the reaction selectivity entirely towards N2_2, water is still the main product. Oxygen is undetected near the catalyst surface, fully consumed for the production of N2_2 and H2_2O via the oxidation reaction. Ammonia can be thus considered to be in excess, a new production of H2_2 is measured, which supports that the simultaneous dissociation of NH3_3 is favoured is this condition. When removing oxygen, only the dissociation of ammonia is observed, put into evidence by the presence of hydrogen and absence of water.

O 1s and N 1s Spectra: oxygen speciation and nitrogen adsorbates

The evolution of the N 1s and O 1s XPS spectra for different atmospheres is displayed in Figure 9. Binding energy are given with reference to the Fermi level, and re-normalised at the pre-edge intensity (low binding energy side. All the reported peaks and corresponding species in the discussion below are summarised in Table 2. Subscript aa and gg stand for adsorbed and gas phase respectively.

Figure 9. XPS spectra collected at the O 1s (a) and N 1s (b) levels under different atmospheres at 450 °C with an incoming photon energy of 700 eV. The spectra are normalised and shifted in intensity to highlight the presence of different peaks.

1234567
Partial pressures (mbar)Ar 1000010
NH3_3 001.11.11.100
O2_2 08.88.80.55000.55
Gas signals (decreasing pressure order)ArO2_2O2_2, H2_2O, NO NH3_3, N2_2, N2_2OH2_2O, NH3_3 H2_2, N2_2H2_2, NH3_3 N2_2ArO2_2
N 1s: peak positionsNo dataNo peak402.6 eV404.1 eV 399.8 eV 397.5 eV405.3 eV 400.8 eV 398.4 eV400.4 eV 398.4 eVNo peak
Attributed surface speciesNot indexedN2,g_{2,g} NH3,a_{3,a} Na_aN2,g_{2,g} NH3,g_{3,g} NHx,a_{x,a}NH3,a_{3,a} NHx,a_{x,a}
O 1s: peak positions532.4 eV538.2 eV 537.1 eV 531.4 eV 529.7 eV538.5 eV 537.5 eV 534.0 eV534.0 eV 532.0 eV532.4 eV534.0 eV 532.4 eV538.3 eV 537.2 eV 531.4 eV 529.7 eV
Attributed surface speciesH2_2Oa_aO2,g_{2,g} O2,g_{2,g} OHa_a Oa_aO2,g_{2,g} O2,g_{2,g} H2_2Og_gH2_2Og_g H2_2Oa_aH2_2Oa_aH2_2Og_g H2_2Oa_aO2,g_{2,g} O2,g_{2,g} OHa_a Oa_a

Table 2. Indexing of peaks measured during ammonia oxidation of the Pt(111) surface.

Argon: A peak is visible at 532.4 eV in the O 1s edge, attributed to adsorbed water [24, 25], possibly from contaminants in the gas line.

High oxygen: The oxidation cycle is started by introducing 8.8 mbar of oxygen in the reactor. The presence of gas phase oxygen (O2,g_{2,g}) can be confirmed in the O 1s spectra by a characteristic peak doublet around 538 eV. The positions are shifted by 2.2 eV in binding energy with respect to literature (539.3 eV, 540.4 eV, Avval et al. [26]). This is explained by a change in the work function of the sample-analyser setup, impacting gas species exclusively since not in electrical contact with the system [27]. The broad signal extending from 528.5 eV to 533 eV is hiding various peaks from oxygen species. Fisher et al. [24] report the signal for adsorbed oxygen (Oa_a) and adsorbed hydroxyl groups (OHa_a) at respectively 529.8 eV and 531 eV when exposing the Pt(111) surface to water. Peuckert et al. [28] have studied various oxidised Pt surfaces and indexed a peak at 530.2 eV for Oa_a on Pt(111), and at 531.5 eV for OHa_a on poly-crystalline Pt. Derry et al. [29] report 530.8 eV for Oa_a on Pt(111) during its exposition to oxygen, while Zhu et al. [30] report Oa_a at 529.9 eV when probing the dissociation of NO on the Pt(111) surface. Fantauzzi et al. [31] report oxygen surface species at 529.7 eV during the oxidation of Pt(111) at 225 °C, similarly to Miller et al. [32]. During a recent study of the oxidation of ammonia at different pressures and O2_2/NH3_3 ratio on Pt(111), (2 × 2) chemisorbed oxygen and hydroxyl groups were reported respectively at 529.7 eV and 531.4 eV in 1 mbar of oxygen at 325 °C [33]. However, such surface reconstructions were not observed in the in-plane reciprocal space maps (Figure 1b-Figure 1c). It is probable that hydroxyl groups as well as atomic oxygen are adsorbed on the surface around 531.4 eV and 529.7 eV, respectively.

High oxygen reacting conditions: Introducing 1.1 mbar of ammonia leads to the pO2/pNH3=8p_{\mathrm{O}_2}/p_{\mathrm{NH}_3} = 8 condition. Only a weak signal at 402.6 eV can be detected in the N 1s spectrum. Gas phase nitric oxide (NOg_g), ammonia (NH3,g_{3,g}), nitrogen (N2,g_{2,g}), and nitrous oxide (NO2,g_{2,g}) are visible in the RGA signals (Figure C.4). The main product, NOg_g, is expected between 404.5 eV and 406.7 eV in the N 1s level, observed during reacting conditions with an equal amount of oxygen and ammonia, a total pressure of 1 mbar, and temperature of 325 °C [33]. Since not observed in the O 1s level, the authors have proposed that the related signal is hidden by the important gas phase oxygen features. The same study reports N2,g_{2,g} between 403.9 eV and 404.8 eV, and NH3,g_{3,g} at 400.4 eV. NH3,g_{3,g} was also measured shifted to 400.7 eV under 1 mbar of ammonia, without oxygen present in the cell. Adsorbed nitrogen oxide (NOa_a) is absent, expected to yield peaks between 530 eV and 532 eV in the O 1s level, and between 400.4 eV and 401.3 eV in the N 1s level [30, 34, 35]. It is worth noticing that this region is also populated by water and OHa_a signals. Since NO is one of the main products at this condition, the absence of NOa_a can be explained by short desorbing time from the Pt(111) surface [33]. The peak at 402.6 eV in this study could not be indexed, situated far away in binding energy from any of the reported gas phase peaks, and of adsorbed nitrogen and NHx_x type features. Gas phase water (H2_2Og_g) is visible in the O 1s spectra by a peak at 534 eV, as reported during ammonia oxidation by Weststrate et al. [36], generated by the oxidation reaction. The energy difference between the low energy peak of O2,g_{2,g} and H2_2Og_g, 3.5 eV, is close to the difference reported in literature for pure gas phases, equal to 3.3 eV [26, 37].

Low oxygen reacting conditions: By lowering the oxygen pressure to 0.55 mbar so that pO2/pNH3=0.5p_{O_2}/p_{NH_3}=0.5, nitrogen production is favoured (Figure C.4), three peaks can be detected in the N 1s level. 397.5 eV is characteristic of adsorbed atomic nitrogen (Na_a) on Pt(111) [30, 38]. The energy difference between Na_a, NHa_a, and NH2,a_{2,a} is reported to be approximately 0.95 eV, 1.9 eV in total on Pt(111) [33], which is too few to link the peak at 399.8 eV to NHa_a or NH2,a_{2,a} with respect to the Na_a peak. Since there is a peak at 400.8 eV in the following spectrum that can be linked to gas phase ammonia from similar reported binding energies, the peak at 399.8 eV is attributed to adsorbed ammonia (NH3,a_{3,a}). The peak at 404.1 eV is attributed to N2,g_{2,g}, which is the only measured nitrogen product, and also in accordance with reported binding energies. No more oxygen is detected in the reactor (Figure C.4). Moreover, no OHa_a and Oa_a peaks can be detected in the O 1s level (Figure 9), even though the total pressure was divided by 6, increasing the detection of photo-electrons. The peak at 532 eV is attributed to adsorbed water groups (H2_2Oa_a), reported between 532.2 and 532.9 eV on Pt(111) depending on the surface coverage [24, 25].

Ammonia: Once oxygen is removed, gas phase water disappears from the O 1s level, confirming that water in gas phase is mostly present from the oxidation reaction rather than UHV/gas contaminant. Adsorbed water is still visible from long desorption time, or from contaminants. The energy level differs by 0.5 eV from adsorbed water during reacting conditions, signifying different chemical environments. As observed in the mass spectrometer, the dissociation of ammonia towards nitrogen still occurs, a slightly shifted N2,g_{2,g} peak is reported at 405.3 eV. There is a clear effect of the presence of oxygen on the position of gas phase signals. The large peak linked to atomic nitrogen has disappeared, a peak linked to NH3,g_{3,g} is reported as well as a large and weak peak linked to NHx_x groups.

Argon: The introduction of argon and removal of ammonia increases the H2_2Og_g signal, possibly from contaminants. Some nitrogen rich species are still visible. Just like water, ammonia has the tendency to adsorb well on the inner walls of the chamber resulting in a slow pumping speed.

Low oxygen: Removing argon and introducing 0.55 mbar of oxygen has removed all the N 1s peaks by the oxidation of the leftover NHx_x species. As previously discussed, comparable peaks are observed in the O 1s level as under 8.8 mbar of oxygen, linked to hydroxyl groups, Oa_a, and O2,g_{2,g} but with a higher apparent amount of atomic oxygen species in comparison to hydroxyl groups.

Core level Evolution

The Pt 4f7/2_{7/2} level was also measured, the spectra are normalised by the pre-edge (low binding energy side) intensity and shown in Figure 10.

Figure 10. XPS spectra collected at the Pt 4f7/2_{7/2} level under different atmospheres at 450 °C with an incoming photon energy of 550 eV. A Shirley-type background has been subtracted from all XPS spectra. Normalisation performed first by the background intensity and secondly by the maximum intensity to allow a qualitative comparison between different total pressures. The top-left panel presents the spectra before scaling them to the maximum intensity.

High oxygen: Miller et al. [39] have measured two peaks at 72.1 eV and 73.5 eV under 6.66 mbar of oxygen at 450 °C. The 72.1 eV peak is assigned to (i) oxygen between the metallic surface and surface α-PtO2_2 oxide as reported by Ellinger et al. [40]. The 73.5 eV peak is attributed to (ii) Pt atoms within the trilayer oxide structure. However, both peaks are absent at 0.66 mbar of oxygen at 350 °C, for which the presence of (iii) (2 × 2) chemisorbed oxygen and (iv) "4O" oxide surface stripes are linked to two other peaks, respectively 71.1 eV, and 71.6-71.7 eV. In a following study under 13.33 mbar of oxygen, slightly shifted high intensity peaks at 72.2 eV and 73.6 eV are linked to the presence of surface α-PtO2_2 oxide, fully covering a Pt(111) crystal [32]. The photon energy for both experiments is equal to 275 eV, whereas the photon energy is here equal to 550 eV. Thus, both studies by the Miller group are more sensitive to the surface structure since the emitted photo-electrons have increased inelastic mean free paths. Interestingly the crystal preparation to grow surface α-PtO2_2 oxide is not too far from this study. The Pt(111) crystal was exposed to 13.33 mbar of oxygen for 10 min while cycling the temperature from 25 °C to 525 °C four times. A small peak can be identified near 71.6 eV in the current experiment under 8.88 mbar of oxygen that could correspond to the "4O" oxide stripe structure. Its presence is not certain since the intensity is very low, corresponding potential peaks in the O 1s level cannot be resolved either. No clear peak could be detected near 73.5 eV in the Pt 4f7/2_{7/2} level, but a very small peak can be seen at 72.1 eV. Since the presence of surface α-PtO2_2 oxide was linked to very high intensity peaks by Miller et al. [32], this structure can safely be ruled out in this study. It seems that the temperature cycling is crucial to grow α-PtO2_2 on Pt(111). The peak maxima is shifted by 0.09 eV when introducing 8.88 mbar of oxygen in the cell after pure argon atmosphere. Overall, the lack of clear peak corresponding to the "4O" oxide stripe and α-PtO2_2 phases suggests that the oxygen on the Pt(111) surface is mostly chemisorbed, which could also be why the peak is slightly shifted towards 71.1 eV.

High oxygen reacting conditions: When introducing 1 mbar of ammonia in the reactor, the maximum of the Pt 4f7/2_{7/2} peak is measured back at the position under inert atmosphere, while a new component is measured at 70.5 eV.

Low oxygen reacting conditions: Reducing the pressure of oxygen to 0.55 mbar further increases the intensity of the latter component compared to the maximum peak intensity.

Ammonia: Removing oxygen but keeping ammonia in the reactor removes this peak. Since the intensity of this peak increases when the O2/NH3\mathrm{O}_2/\mathrm{NH}_3 ratio decreases, i.e. when all the oxygen is consumed, it is probably linked to the presence of adsorbed nitrogen species on the platinum surface. The absence of this peak under the presence of ammonia in the cell, for which adsorbed nitrogen can not be detected, supports a link with Na\mathrm{N}_a (Figure 9).

Argon: Only a smaller shift is repeated when 0.55 mbar of oxygen is introduced after 1 mbar or argon, approximately equal to 0.2 eV.

Low oxygen: Both spectra collected under argon before and after the oxidation cycle are very similar, some nitrogen species possibly left on the surface are probably removed by oxygen, which explains the increased intensity after the oxidation cycle near 70.5 eV.

Discussion

Kinetics and transient oxides

The different features observed in the SXRD maps during surface oxidation are summarised and discussed below. First, two Pt(111)-(6 ×\times 6)-R±\pm8.8∘^\circ rotated hexagonal domains appear, at both high (Figure 1b) and low (Figure 5a) oxygen pressure. Secondly, a Pt(111)-(8 ×\times 8) superstructure grows, its second order peaks is detected before the first order peaks at both pressures (Figure 1b-Figure 1c & Figure 5). More time resolved maps at low oxygen pressure allow us to be certain that this peak does not just appear during the map due to long measurement times (Figure 5). These maps also allow us to rule the disappearance of the grey and purple circled peaks in Figure 1c as an alignment problems (large low intensity region), as they are clearly visible in Figure 5. X-ray reflectivity and SSR results indicate that the hexagonal structures are monolayers (Figure 7 and Figure B.2), while the Pt(111)-(8 ×\times 8) superstructure eventually grows into a multi-layer thick epitaxied surface oxide (approximately 15 Å thick in Figure 2b).

Seriani et al. [41] have determined that the most stable surface oxide on the Pt(111) surface is hexagonal α\alpha-PtO2_2. A high uncertainty regarding the lattice parameter in the c⃗\vec{c} direction is reported in experimental and theoretically studies due to a poor crystallisation of the bulk oxide [42], and to underestimation of weak interlayer Van der Waals forces in theoretical studies [43]. The formation of hexagonal or hexagonal surface α\alpha-PtO2_2 on Pt(111) is expected to induce a high amount of in-plane compressive strain in the oxide layer, from the large difference between the Pt-Pt distance on the (111) surface (2.77 Å) and the in-plane lattice parameter of α\alpha-PtO2_2 (3.14 Å). A rotation of 30° is expected to reduce the in-plane strain, resulting in a (2 ×\times 2) arrangement, no such structure was observed in this experiment on Pt(111). Ellinger et al. [40] have reported the existence of surface α\alpha-PtO2_2 on Pt(111) using SXRD at 500 mbar of oxygen, within 245°C and 635°C, and also with a Pt(111)-(8 ×\times 8) superstructure, as previously reported by Ackermann [44]. The retrieved in-plane lattice parameter is equal to 3.15 Å, close to the value estimated by Seriani et al. [41]. The surface α\alpha-PtO2_2 model consists of one unit cell, with an out-of-plane lattice parameter reduced by 15% in comparison with the bulk structure (3.62 Å) [40]. They introduced a displacement of ≈\approx33% in the (a⃗α-PtO2+b⃗α-PtO2\vec{a}_{\alpha\text{-PtO}_2}+\vec{b}_{\alpha\text{-PtO}_2}) direction for the platinum atoms on the top layers with respect to those of the bottom layer. A super-structure rod collected at [H, K] = [0, 0.88] present a peak at L=0.65L = 0.65 (LL computed using the distorted α\alpha-PtO2_2 unit cell). The two SSRs performed in the current study were fitted using two Gaussian peaks sharing the same full width at half maximum (FWHM), and a constant background (Figure 11). The second peak near L=1L = 1 in the SSR at [H, K] = [0.89, 0] comes from a Pt powder signal, while the peak at L=2.4L = 2.4 is from alien signal. A peak at [0.89, 0, 0.65] is visible, but using the Pt(111) surface unit cell, with an out-of-plane lattice parameter of 6.79 Å. This is incompatible with the results presented in Ellinger et al [40] since they used the lattice parameter of the refined α\alpha-PtO2_2 structure, ∣c⃗∣=4.16|\vec{c}| = 4.16 Å. It seems that a different out-of-plane structure is present in the current work.

Figure 11. SSRs and fit results under 80 mbar of O2O_2. μ\mu, and σ\sigma are the peak positions and standard deviations in LL.

XPS studies have also reported α\alpha-PtO2_2 above 6.66 mbar of O2\mathrm{O}_2 at 450°C [32, 39]. As mentioned by Fantauzzi et al. [31], the extended exposure to oxygen necessary to grow the oxide structure makes its characterisation difficult, the structure possibly evolving with time. A large peak at 530.7 eV was attributed to Pt surface oxides during a study at pO2=1p_{\mathrm{O}_2} = 1 bar [45]. Additional simulations are needed to fully understand the structures in this study using the proposed distorted α\alpha-PtO2_2 structure, but also different surface models including the spoked-wheel superstructure identified via scanning tunnelling microscopy (STM) by Van Spronsen et al [45] and Boden et al. [46].

This structure, identified above 1 bar of oxygen, and at approximately 200 °C, also exhibits a (8×8)(8 \times8) coincidence with the Pt(111) lattice.

A more quantitative analysis of the different structures appearing during the exposition to oxygen was performed by integrating the scattered intensity around the Pt(111)-(8 × 8) and two Pt(111)-(6 × 6)-R±8.8° signals present in the (100) region (respectively circled in black, red and white in Figure 6). The average background was subtracted to each reciprocal space voxel before integration. The starting time of each reciprocal space in-plane map is used as estimate for the time since the introduction of oxygen, the evolution of each peak is presented in Figure 12. The growth of the Pt(111)-(8 × 8) peak at [H, K] = [0, 0.89] follows an exponential increase after 23 h of exposition.

Figure 12. Intensity evolution as a function of time since introduction of 5 mbar of oxygen for peaks corresponding to the Pt(111)-(6 × 6)-R±8.8° and Pt(111)-(8 × 8) structures.

The white Pt(111)-(6 × 6)-R±8.8° peak is visible from the start of the exposition to oxygen (Figure 6a), whereas the red Pt(111)-(6 × 6)-R±8.8° peak is only clearly visible after 7 h (Figure 6c), with a lower intensity. Both peaks intensity quickly increases in the first hours of oxygen exposure, and then reach a steady state between 8 h and 23 h, increasing again together with the appearance of the Pt(111)-(8 × 8) signal. The Pt(111)-(6 × 6)-R±8.8° peak intensity then starts to decreases after ≈25 h. Additional studies are needed to understand if whether or not the rotated hexagonal structures act only as precursors for the Pt(111)-(8 × 8) superstructure, or if those peaks are still present after an extended exposure to oxygen. Several SSRs have been measured at the positions of the Pt(111)-(6 × 6)-R±8.8° peaks at low oxygen pressure (5 mbar), presented in Figure B.2. The intensity is constant as a function of LL, similarly to the SSRs presented in Figure 2, which shows that the rotated hexagonal structures are flat monolayers on the Pt(111) surface. No SSR was measured at the position of the Pt(111)-(8 × 8) peak at that condition.

Oxides and ammonia

The presence of α-PtO₂ on Pt(111) has been shown to not hinder the oxidation of CO by Ackermann et al. [44], that would then occur via a Mars-Van Krevelen mechanism [47], inducing a progressive roughening of the platinum surface. The presence of α-PtO2_2 on the Pt(111) surface was also shown to provide favourable special sites that could contribute to the catalytic activity [43]. The low roughness and removal of SXRD signals during ammonia oxidation is in contradiction with a possible Mars-Van Krevelen mechanism involving surface oxides at the current conditions, consistent with the reported Langmuir-Hinshelwood literature mechanism.

The presence of adsorbed water supports a Langmuir-Hinshelwood mechanism with quick stripping of hydrogen from NHx,a_{x,a} species by OHa_a and Oa_a, eventually forming adsorbed water. The de-hydrogenation process must be limiting the catalytic activity due to the lack of available adsorbed oxygen species near adsorbed ammonia, which could be why NH3,a_{3,a} is observed, but not NHx,a_{x,a}. The key importance of available atomic oxygen for the production of NO is supported by the simultaneous absence of Oa_a in the O 1s level, and the absence of NO in the RGA signals (Figure C.4). The presence of adsorbed nitrogen could be due to long residual times on the catalyst before recombination and desorption of N2_2.

The dissociation of ammonia is reported to be slower without the help of oxygen species on Pt(111) [48, 49, 50, 51]. This could explain why such a large peak is observed for NHx,a_{x,a}, and why NH3,g_{3,g} is observed rather than NH3,a_{3,a}, since most of the adsorption sites are probably occupied by NHx,a_{x,a} species. The difficulty to fully dissociate ammonia is also correlated to the production of hydrogen, it is possible that the combination of two hydrogen atoms removed from ammonia to produce H2_2 is slow, and thus occupies part of the adsorption sites in the absence of oxygen.

To investigate potential surface relaxation effects, the three CTRs were fitted together (see Methods), the evolution of the surface roughness and out-of-plane strain is shown in Figure 13. The Pt(111) surface roughness is high under Argon at 450 °C (Figure 13a), probably due to the presence of impurities contained in the argon gas flow. The introduction of oxygen in the cell further increases the surface roughness, as expected from the formation of the different surface oxides visible in the in-plane reciprocal space maps (Figure 1, CTR recorded between c and d). Adding ammonia in the reactor cell, which removes the different surface oxides (Figure 3), also decreases the surface roughness. Finally, the re-introduction of 5 mbar of oxygen increases the surface roughness again (Figure 13a), in accordance with the formation of surface oxides detected during in-plane reciprocal space maps (Figure 5). Low amounts of strain are detected on the surface, almost imperceptible when observing the CTR minimum position in Figure 8. From the fitting results, the topmost layer is already under tensile strain under inert atmosphere. Different reacting conditions are related to the same direction of out-of-plane displacement, higher ammonia to oxygen ratio coincides with higher tensile strain, no changes from tensile to compressive strain are observed during the reaction. The largest evolution in the strain values comes from the introduction oxygen after pure argon, the out-of-plane lattice parameter almost returning to its bulk value. Overall, the presence of oxygen leads to the formation of surface oxides which lowers the surface strain, and increases the surface roughness. Reacting conditions and the presence of ammonia remove the surface oxides, and decrease the surface roughness.

Figure 13. a) Fitting results for roughness parameter β\beta and b) out-of-plane strain ϵz\epsilon_z with respect to bulk lattice parameter, as a function of the experimental conditions. The innermost layer is named A, which has the same lattice parameter as in the bulk. The topmost layer of the Pt(111) single crystal is named B.

Conclusion

Although Pt–Rh alloys and Pt(100) surfaces can undergo reversible oxide formation and reconstructions that affect ammonia oxidation [10, 14], the operando surface state of pure Pt(111) has been less well established under realistic reaction environments. Here we combined surface X-ray diffraction (SXRD), crystal truncation rod (CTR) analysis, and near-ambient pressure X-ray photoelectron spectroscopy (NAP-XPS) with on-line mass spectrometry to directly link surface structure, chemical state, and product distribution during ammonia oxidation on Pt(111). Across the conditions explored, Pt(111) does not sustain a stable surface oxide or reconstruction in the presence of ammonia. Instead, oxygen-rich feeds promote transient ordered phases, including hexagonal monolayers and a Pt(111)-(8 × 8) superstructure, which are removed upon ammonia exposure. The combined NAP-XPS and product analysis are consistent with a reaction regime governed by co-adsorbed NHx_x and oxygen-derived species, with oxygen-rich conditions favouring NO production and ammonia-rich feeds shifting selectivity toward N2_2. The availability of adsorbed atomic oxygen is clearly revealed as the key to NO selectivity on Pt(111). This behaviour contrasts with Pt(100), where epitaxial oxide phases such as Pt3_3O4_4(001) can form at similar oxidation condition, and where difference surface reconstructions drive the catalyst selectivity [10]. These results highlight a pronounced facet dependence in the stability of oxidized surface states on platinum and demonstrate that Pt(111) operates predominantly in an oxide-free regime under ammonia oxidation. More broadly, this work frames ammonia oxidation as an interface problem in which dynamic surface structure and surface chemistry jointly determine function. By resolving ordered surface phases, relaxations, and roughness under near-industrial environments, SXRD provides a uniquely powerful route to establish structure–chemistry relationships at reactive metal/oxide interfaces. The present findings therefore provide experimentally grounded guidance for designing platinum-based catalytic interfaces with improved stability and selectivity under strongly oxidizing operating conditions.

Materials and methods

Surface X-ray Diffraction

The Pt(111) sample was bought from Surface Preparation Laboratory [52], polished to mirror-finish. It is circular, 1 cm large, and 0.2 cm thick (Figure D.6b). The polished surface area of the single crystal is ≈50 mm2^2. The surface X-ray diffraction experiment was performed at the SIXS beamline (SOLEIL synchrotron), at 18.44 keV on the vertical Z-axis diffractometer by means of a flow reactor that operates mirroring ref. [53] (Figure D.6 & D.7) at industry relevant conditions. The large reactor chamber accommodates a sputtering gun, and is connected to two gas inlet and outlet valves. The incoming gas mixture is controlled upstream by mass flow controllers, while the total reactor pressure is set at the outlet. A leak from the reactor chamber to a residual gas analyser (RGA) allows monitoring the reactant and product partial pressures. Two sputtering (30 min) and annealing (800 °C - 5 min) cycles were necessary to obtain a clean and flat surface before starting the SXRD experiment, as verified by the truncated Pt(111) peaks in Figure 1. The sputtering current was maximised by tuning the sample height to ensure that the Ar ion beam covers the entire sample surface. The incident angle during the SXRD experiment was set to 0.3° to increase surface sensitivity. The hexagonal basis set of vectors with ∣a1∣|a_1|, ∣a2∣|a_2| lying in the surface plane and ∣a3∣|a_3| perpendicular to them was used to describe the crystal structure. In terms of the bulk lattice constant aPta_{\mathrm{Pt}} (aPt=3.92a_{\mathrm{Pt}} = 3.92 Å) the lengths of these vectors can be expressed as ∣a1∣=∣a2∣=∣aPt∣/2=2.77|a_1| = |a_2| = |a_{\mathrm{Pt}}|/\sqrt{2} = 2.77 Å and ∣a3∣=3aPt=6.79|a_3| = \sqrt{3}a_{\mathrm{Pt}} = 6.79 Å. Given the Pt(111) surface three-fold symmetry and for the sake of temporal approximation, only a third of the reciprocal space in the (q⃗x,q⃗y)(\vec{q}_x,\vec{q}_y) plane, where q⃗x\vec{q}_x and q⃗y\vec{q}_y are the in-plane components of the scattering vector, was measured during the in-plane reciprocal space maps. The (21L), (10L) and (11L) crystal truncation rods were measured 6 h after the start of each condition, each measurement lasted for 2 h. The rods where numerically integrated using the fit aid module of BINoculars which allows for background corrections [54].

X-ray reflectivity

The specular X-ray reflectivity simulations are conducted using the Parratt algorithm [55]. The classical way of implementing roughness is based on a Gaussian distribution, introducing corrective factors to the electric field amplitudes at the interfaces in accordance with the Nevot-Croce model [56]. For the computation of material refractive indices, the Henke tables [57] are employed. A simple model was used consisting of a semi-infinite slab of platinum, on top of which is present a homogeneous layer of platinum oxide. The density of the layer (ρox\rho_{\mathrm{ox}}) was set free during the minimisation process, between zero and 21.2×10−321.2 \times10^{-3} FU/Å3^{3} (formula unit per cubic Angström, computed from unit cell elements and volume), the value for α\alpha-PtO2_2. The initial intensity is fixed to obtain the best match at low incident angles, the intensity drop after the critical angle is difficult to fit, but the period of the oscillations, as well as the intensity at high angles are in general well represented. The oscillation magnitude can be linked to the density of the oxide layer on top of the Pt(111) surface. The fitted values (7×10−37 \times10^{-3} FU/Å3^{3} and 6.5×10−36.5 \times10^{-3} FU/Å3^{3}) are an order of magnitude below the expected value for a homogeneous layer of α\alpha-PtO2_2. The root mean square roughness of the oxide layer σox\sigma_{\mathrm{ox}} is always zero, whereas the roughness of the Pt(111) surface (σsub\sigma_{\mathrm{sub}}) is of the same order of magnitude for both measurements.

Near Ambient Pressure X-ray Photoelectron Spectroscopy

The same Pt(111) crystal was measured with NAP-XPS at the B07 beamline[58] (Diamond Light Source). Due to the impact of the pressure on the electron transmission, the total pressure was lowered and no longer kept constant by the use of a carrier gas (argon) between each condition. The reactant partial pressure becomes approximately 10 % of the partial pressures used in SXRD, with an equivalent (O2_2/NH3_3) ratio (Table 1). The electron analyser axis was set at normal emission with respect to the sample surface. The cone of the analyser was also used as inlet for mass spectrometry measurements, it has a diameter of 0.3 mm, with a typical working distance of 0.2 mm. The combined energy resolution of the beamline and analyser, while working in NAP conditions, ranged between 0.8 eV and 1.2 eV. The Fermi edge was also measured after each core level to be able to correctly calibrate the binding energy.

Acknowledgements

We would like to offer our most sincere gratitude to Dr. Michele Sauvage for fruitful discussions, and to Dr. Frédéric Picca for his help regarding the installation and development of BINoculars (https://repo.or.cz/hkl.git, https://github.com/picca/binoculars/) at SixS. This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (grant agreement No. 818823). We thank SOLEIL for providing the majority of the beam time required for this work and Diamond Light Source for the additional allocation, and we are grateful to the SixS and B07 teams for their assistance throughout the experiments.

Conflict of interest

The authors declare no competing financial interest.

Reactor calibration

Figure A.1. a) Sample temperature calibration, measured under vacuum and different Ar pressures as a function of heater input current. b) Evolution of partial pressures in the RGA chamber when pO2/pNH3=8p_{\mathrm{O}_2}/p_{\mathrm{NH}_3} = 8. Increasing and decreasing (low transparency) temperature ramp to 650 °C. The ramp is performed with 100 steps, each lasting 10 s.

ArNH₃O₂NON₂ON₂H₂O
25 °C17.725.126.525.917.426.018.6
125 °C22.437.234.033.126.032.826.1
225 °C26.553.141.039.634.139.035.6
325 °C30.368.647.746.241.844.846.2

Table 3. Thermal conductivity of gases used in during ammonia oxidation in mW m−1^{-1} K−1^{-1} [59].

Surface X-ray diffraction

Structure factor measured at 5 mbar O₂ as a function of L

Figure B.2. SSR measured at pO2=5p_{\mathrm{O}_2} = 5 mbar.

Figure B.3. Time dependent partial pressures recorded from a leak in the reactor output by a residual gas analyser (RGA) during the SXRD experiment on the Pt(111) single crystal at 450 °C.

X-ray Photoelectron Spectroscopy

Figure C.4. Evolution of reaction product partial pressures as a function of time during the XPS experiment on Pt(111) at 450 ∘C450\,^{\circ}\mathrm{C}. Transition between conditions are indicated with dashed vertical lines. The glitch in the partial pressure at 28.5 h is due to a re-setting of the total pressure to 1.1 mbar.

Figure C.5. XPS spectra collected at the Pt 4f7/24f_{7/2} level under different atmospheres at 450 ∘C450\,^{\circ}\mathrm{C}. Normalisation performed by the pre-edge intensity.

Experimental setups

Large XCAT reactor, round single crystal on sample holder, and detailed drawing of the XPS chamber

Figure D.6. a) Large XCAT reactor. b) Round single crystal set on sample holder, the crystal presents here a (111)(111) top surface. The surface area of the single crystal is ≈50 mm2\approx50\ \mathrm{mm}^2. c) Detailed drawing of the XPS chamber at the B07 beamline at the Diamond Light Source [58].

Sample holder used for the XCAT catalysis reactor

Figure D.7. Sample holder used for the XCAT catalysis reactor. A PEEK dome, transparent to X rays, is set upon the sample holder during operando measurements.

References

  1. [1]Hatscher, S. T.; Fetzer, T.; Wagner, E.; Kneuper, H.-J. In Handbook of Heterogeneous Catalysis; John Wiley & Sons, Ltd: 2008; Chapter 12.2, pp 2575–2592, DOI: 10.1002/9783527610044.hetcat0130.
  2. [2]Heck, R. M.; Bonacci, J. C.; Hatfield, W. R.; Hsiung, T. H. A new research pilot plant unit for ammonia oxidation processes and some gauze data comparisons for nitric acid process. Industrial & Engineering Chemistry Process Design and Development 1982, 21, 73–79, DOI: 10.1021/i200016a014.DOI
  3. [3]MIT Greenhouse gases climate.mit.edu/explainers/greenhouse-gases (accessed 20/04/2023).
  4. [4]EPA Greenhouse gases overview www.epa.gov/ghgemissions/overview-greenhouse-gases (accessed 20/06/2023).
  5. [5]Solomon, S., Climate change 2007-the physical science basis: Working group I contribution to the fourth assessment report of the IPCC; Cambridge university press: 2007; Vol. 4.
  6. [6]Davidson, E. A. The contribution of manure and fertilizer nitrogen to atmospheric nitrous oxide since 1860. Nature Geoscience 2009, 2, 659–662, DOI: 10.1038/ngeo608.DOI
  7. [7]Kabange, N. R.; Lee, S.-M.; Shin, D.; Lee, J.-Y.; Kwon, Y.; Kang, J.-W.; Cha, J.-K.; Park, H.; Alibu, S.; Lee, J.-H. Multiple Facets of Nitrogen: From Atmospheric Gas to Indispensable Agricultural Input. Life 2022, 12, DOI: 10.3390/life12081272.DOI
  8. [8]Afif, A.; Radenahmad, N.; Cheok, Q.; Shams, S.; Kim, J. H.; Azad, A. K. Ammonia-fed fuel cells: a comprehensive review. Renewable and Sustainable Energy Reviews 2016, 60, 822–835, DOI: 10.1016/j.rser.2016.01.120.DOI
  9. [9]Jeerh, G.; Zhang, M.; Tao, S. Recent progress in ammonia fuel cells and their potential applications. Journal of Material Chemistry A 2021, 9, 727–752, DOI: 10.1039/D0TA08810B.DOI
  10. [10]Simonne, D.; Coati, A.; Vlad, A.; Garreau, Y.; Voisin, B.; Resta, A.; Richard, M. Linking Surface Reconstructions to Catalytic Selectivity on Pt(100) during Ammonia Oxidation. ACS Catalysis 2026, DOI: doi.org/10.1021/acscatal.5c07084.DOI
  11. [11]Kaiser, K. New Industrial Rhodium-Platinum Fabric Catalyst, 271517, 1909.
  12. [12]Handforth, S. L.; Tilley, J. N. Catalysts for Oxidation of Ammonia to Oxides of Nitrogen. Industrial & Engineering Chemistry 1934, 26, 1287–1292, DOI: 10.1021/ie50300a016.DOI
  13. [13]McCabe, R.; Pignet, T.; Schmidt, L. Catalytic etching of platinum in NH₃ oxidation. Journal of Catalysis 1974, 32, 114–126, DOI: 10.1016/0021-9517(74)90165-1.
  14. [14]Resta, A.; Hejral, U.; Blomberg, S.; Albertin, S.; Vlad, A.; Garreau, Y.; Chatelier, C.; Venturini, F.; Ferrer, P.; Held, G.; Grinter, D.; Lundgren, E.; Coati, A. Ammonia Oxidation over a Pt25Rh75(001) Model Catalyst Surface: An Operando Study. Journal of Physical Chemistry C 2020, 124, 22192–22199, DOI: 10.1021/acs.jpcc.0c07128.DOI
  15. [15]Hammer, B.; Nørskov, J. K. Why gold is the noblest of all the metals. Nature 1995, 376, 238–240, DOI: 10.1038/376238a0.DOI
  16. [16]Mavrikakis, M.; Hammer, B.; Nørskov, J. Effect of Strain on the Reactivity of Metal Surfaces. Physical Review Letters 1998, 81, DOI: 10.1016/j.jcat.2020.07.030.
  17. [17]Hammer, B.; Nørskov, J. K. Theoretical surface science and catalysis—calculations and concepts. Advances in Catalysis 2000, 45, 71–129, DOI: 10.1016/S0360-0564(02)45013-4.DOI
  18. [18]Simonne, D.; Richard, M.; Dupraz, M.; Vlad, A.; Garreau, Y.; Chatelier, C.; Voisin, B.; Atlan, C.; Yehya, S.; Levi, M.; Rabkin, E.; Coati, A.; Resta, A. Dynamic structural changes of ensemble and single supported Pt particles during ammonia oxidation. Applied Catalysis B: Environment and Energy 2026, 383, 126055, DOI: https://doi.org/10.1016/j.apcatb.2025.126055.
  19. [19]Seriani, N.; Mittendorfer, F. Platinum-group and noble metals under oxidizing conditions. Journal of Physics: Condensed Matter 2008, 20, 184023, DOI: 10.1088/0953-8984/20/18/184023.DOI
  20. [20]Robinson, I. K.; Tweet, D. J. Surface X-ray diffraction. 1991, 55, 599–651, DOI: 10.1088/0034-4885/55/5/002.DOI
  21. [21]Björck, M.; Andersson, G. GenX: an extensible X-ray reflectivity refinement program utilizing differential evolution. Journal of Applied Crystallography 2007, 40, 1174–1178, DOI: 10.1107/S0021889807045086.DOI
  22. [22]Glavic, A.; Björck, M. GenX 3: the latest generation of an established tool. Journal of Applied Crystallography 2022, 55, 1063–1071, DOI: 10.1107/S1600576722006653.DOI
  23. [23]Robinson, I. K. Crystal truncation rods and surface roughness. Physical Review B 1986, 33, 3830–3836, DOI: 10.1103/PhysRevB.33.3830.DOI
  24. [24]Fisher, G. B.; Gland, J. L. The interaction of water with the Pt(111) surface. Surface Science 1980, 94, 446–455, DOI: 10.1016/0039-6028(80)90018–7.
  25. [25]Kiskinova, M.; Pirug, G.; Bonzel, H. Adsorption and decomposition of H₂O on a K-covered Pt(111) surface. Surface Science 1985, 150, 319–338, DOI: 10.1016/0039–6028(85)90649–1.
  26. [26]Avval, T. G.; Chatterjee, S.; Hodges, G. T.; Bahr, S.; Dietrich, P.; Thiben, A.; Linford, M. R.; Meyer, M. Oxygen gas, O₂(g), by near-ambient pressure XPS. Surface Science Spectra 2022, 2, DOI: 10.1116/1.5100962.
  27. [27]Starr, D. E. In ACS Symposium Series, Vol. 1396; American Chemical Society: 2021, pp 1–17, DOI: 10.1021/bk–2021–1396.ch001.
  28. [28]Peuckert, M.; Bonzel, H. Characterization of oxidized platinum surfaces by X-ray photoelectron spectroscopy. Surface Science 1984, 145, 239–259, DOI: 10.1016/0039–6028(84)90778–7.
  29. [29]Derry, G.; Ross, P. High coverage states of oxygen adsorbed on Pt(100) and Pt(111) surfaces. Surface Science 1984, 140, 165–180, DOI: 10.1016/0039–6028(84)90389–3.DOI
  30. [30]Zhu, J. F.; Kinne, M.; Fuhrmann, T.; Denecke, R.; Steinrück, H. P. In situ high-resolution XPS studies on adsorption of NO on Pt(111). Surface Science 2003, 529, 384–396, DOI: 10.1016/S0039–6028(03)00298–X.DOI
  31. [31]Fantauzzi, D.; Krick Calderón, S.; Mueller, J. E.; Grabau, M.; Papp, C.; Steinrück, H. P.; Senftle, T. P.; van Duin, A. C.; Jacob, T. Growth of Stable Surface Oxides on Pt(111) at Near-Ambient Pressures. Angewandte Chemie - International Edition 2017, 56, 2594–2598, DOI: 10.1002/anie.201609317.DOI
  32. [32]Miller, D.; Sanchez Casalongue, H.; Bluhm, H.; Ogasawara, H.; Nilsson, A.; Kaya, S. Different reactivity of the various platinum oxides and chemisorbed oxygen in CO oxidation on Pt(111). Journal of the American Chemical Society 2014, 136, 6340–6347, DOI: 10.1021/ja411325q.DOI
  33. [33]Ivashenko, O.; Johansson, N.; Pettersen, C.; Jensen, M.; Zheng, J.; Schnadt, J.; Sjåstad, A. O. How Surface Species Drive Product Distribution during Ammonia Oxidation: An STM and Operando APXPS Study. ACS Catalysis 2021, 8261–8273, DOI: 10.1021/acscatal.1c000956.DOI
  34. [34]Kiskinova, M.; Pirug, G.; Bonzel, H. NO adsorption on Pt(111). Surface Science 1984, 136, 285–295, DOI: 10.1016/0039–6028(84)90612–5.
  35. [35]Günther, S.; Scheibe, A.; Bluhm, H.; Haevecker, M.; Kleimenov, E.; Knop-Gericke, A.; Schlögl, R.; Imbihl, R. In Situ X-ray Photoelectron Spectroscopy of Catalytic Ammonia Oxidation over a Pt(533) Surface. The Journal of Physical Chemistry C 2008, 112, 15382–15393, DOI: 10.1021/jp803264v.DOI
  36. [36]Weststrate, C.; Bakker, J.; Rienks, E.; Vinod, C.; Matveev, A.; Gorodetskii, V.; Nieuwenhuys, B. Ammonia oxidation on Pt(410). Journal of Catalysis 2006, 242, 184–194, DOI: 10.1016/j.jcat.2006.06.011.DOI
  37. [37]Linford, M. R.; Meyer, M.; Thiben, A.; Thiben, A.; Linford, M. R. Water vapor, by near-ambient pressure XPS. Surface Science Spectra 2019, DOI: 10.1116/1.5111634.DOI
  38. [38]van den Broek, A.; van Grondelle, J.; van Santen, R. Determination of Surface Coverage of Catalysts: Temperature Programmed Experiments on Platinum and Iridium Sponge Catalysts after Low Temperature Ammonia Oxidation. Journal of Catalysis 1999, 185, 297–306, DOI: 10.1006/jcat.1999.2506.DOI
  39. [39]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. Physical Review Letters 2011, 107, 195502, DOI: 10.1103/PhysRevLett.107.195502.DOI
  40. [40]Ellinger, C.; Stierle, A.; Robinson, I. K.; Nefedov, A.; Dosch, H. Atmospheric pressure oxidation of Pt(111). Journal of Physics Condensed Matter 2008, 20, 1–5, DOI: 10.1088/0953-8984/20/18/184013.DOI
  41. [41]Seriani, N.; Pompe, W.; Ciacchi, L. C. Catalytic oxidation activity of Pt₃O₄ surfaces and thin films. Journal of Physical Chemistry B 2006, 110, 14860–14869, DOI: 10.1021/jp063281r.DOI
  42. [42]Muller, O.; Roy, R. Formation and stability of the platinum and rhodium oxides at high oxygen pressures and the structures of Pt₃O₄, β-PtO₂ and RhO₂. Journal of the Less Common Metals 1968, 16, 129–146, DOI: 10.1016/0022-5088(68)90070-2.DOI
  43. [43]Li, W.; Hammer, B. Reactivity of a gas/metal/metal-oxide three-phase boundary: CO oxidation at the Pt(111)–c(4×2)-2CO/α-PtO₂ phase boundary. Chemical Physics Letters 2005, 409, 1–7, DOI: 10.1016/j.cplett.2005.04.064.DOI
  44. [44]Ackermann, M. Operando SXRD: A New View on Catalysis. Ph.D. Thesis, 2007.
  45. [45]Van Spronsen, M. A.; Frenken, J. W.; Groot, I. M. Observing the oxidation of platinum. Nature Communications 2017, 8, DOI: 10.1038/s41467-017-00643-z.DOI
  46. [46]Boden, D.; Groot, I. M. N.; Meyer, J. Elucidating the Initial Oxidation of Pt(111) Using Large-Scale Atomistic Thermodynamics: A ReaxFF Study. The Journal of Physical Chemistry C 2022, 126, 20020–20027, DOI: 10.1021/acs.jpcc.2c05769.DOI
  47. [47]Mars, P.; van Krevelen, D. Oxidations carried out by means of vanadium oxide catalysts. Chemical Engineering Science 1954, 3, 41–59, DOI: 10.1016/S0009-2509(54)80005-4.DOI
  48. [48]Offermans, W. K.; Jansen, A. P. J.; van Santen, R. A. Ammonia activation on Pt(111): A density functional theory study. Surface Science 2006, 600, 1714–1734, DOI: 10.1016/j.susc.2006.01.031.DOI
  49. [49]Offermans, W. K.; Jansen, A. P. J.; Van Santen, R. A.; Novell-Leruth, G.; Ricart, J. M.; Pérez-Ramírez, J. Ammonia dissociation on Pt(100), Pt(111), and Pt(211): A comparative density functional theory study. Journal of Physical Chemistry C 2007, 111, 17551–17557, DOI: 10.1021/jp073083f.
  50. [50]Imbihl, R.; Scheibe, A.; Zeng, Y. F.; Günther, S.; Kraehnert, R.; Kondratenko, V. A.; Baerns, M.; Offermans, W. K.; Jansen, A. P.; Van Santen, R. A. Catalytic ammonia oxidation on platinum: Mechanism and catalyst restructuring at high and low pressure. Physical Chemistry Chemical Physics 2007, 9, 3522–3540, DOI: 10.1039/b700866j.DOI
  51. [51]Novell-Leruth, G.; Ricart, J. M.; Pérez-Ramírez, J. Pt(100)-catalyzed ammonia oxidation studied by DFT: Mechanism and microkinetics. Journal of Physical Chemistry C 2008, 112, 13554–13562, DOI: 10.1021/jp802489y.
  52. [52]Laboratory, S. P. Home Page https://www.spl.eu (accessed 08/12/2025).
  53. [53]Rijn, R. V.; Ackermann, M. D.; Balmes, O.; Dufrane, T.; Geluk, A.; Gonzalez, H.; Isern, H.; Kuyper, E. D.; Petit, L.; Sole, V. A.; Vermeille, D.; Felici, R.; Frenken, J. W. Ultrahigh vacuum/high-pressure flow reactor for surface x-ray diffraction and grazing incidence small angle x-ray scattering studies close to conditions for industrial catalysis. Review of Scientific Instruments 2010, 81, DOI: 10.1063/1.3290420.DOI
  54. [54]Roobol, S.; Onderwaater, W.; Drnec, J.; Felici, R.; Frenken, J. BINoculars: Data reduction and analysis software for two-dimensional detectors in surface X-ray diffraction. Journal of Applied Crystallography 2015, 48, 1324–1329, DOI: 10.1107/S1600576715009607.DOI
  55. [55]Parratt, L. G. Surface Studies of Solids by Total Reflection of X-Rays. Phys. Rev. 1954, 95, 359–369, DOI: 10.1103/PhysRev.95.359.DOI
  56. [56]Né́vot, L.; Croce, P. Caractérisation des surfaces par réflexion rasante de rayons X. Application à l’étude du polissage de quelques verres silicates. Rev. Phys. Appl. (Paris) 1980, 15, 761–779, DOI: 10.1051/rphysap:01980001503076100.DOI
  57. [57]Henke, B.; Gullikson, E.; Davis, J. X-Ray Interactions: Photoabsorption, Scattering, Transmission, and Reflection at E = 50-30,000 eV, Z = 1-92. Atomic Data and Nuclear Data Tables 1993, 54, 181–342, DOI: 10.1006/adnd.1993.1013.DOI
  58. [58]Held, G.; Venturini, F.; Grinter, D. C.; Ferrer, P.; Arrigo, R.; Deacon, L.; Quevedo Garzon, W.; Roy, K.; Large, A.; Stephens, C.; Watts, A.; Larkin, P.; Hand, M.; Wang, H.; Pratt, L.; Mudd, J. J.; Richardson, T.; Patel, S.; Hillman, M.; Scott, S. Ambient-pressure endstation of the Versatile Soft X-ray (VerSoX) beamline at Diamond Light Source. Journal of Synchrotron Radiation 2020, 27, 1153–1166, DOI: 10.1107/S1600577520009157.DOI
  59. [59]Huber, M. L.; Harvey, H. Thermal conductivity of gases tsapps.nist.gov/publication/get_pdf.cfm?pub_id=907540 (accessed 20/08/2023).

Paper details

vlnc:W_dtphkv6w
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents