Hydrogen oxidation on oxygen-rich IrO₂(110)

ABSTRACT

We investigated the adsorption and oxidation of H₂ on O-rich IrO₂(110) using temperature programmed reaction spectroscopy (TPRS) and density functional theory (DFT) calculations. Our results show that H₂ dissociation occurs efficiently on O-rich IrO₂(110) at low temperature and initiates from an adsorbed H₂ σ-complex on the coordinatively-unsaturated Ir atoms (Ir_cus). We find that on-top oxygen atoms (O_ot), adsorbed on the Ir_cus sites, promote the desorption-limited evolution of H₂O during subsequent oxidation of the adsorbed hydrogen on IrO₂(110) while suppressing reaction-limited production of H₂O via the recombination of bridging HO groups (HO_br) (~500 to 750 K) during TPRS. The desorption-limited TPRS peak of H₂O shifts from ~490 to 550 K with increasing O_ot coverage, demonstrating that O_ot atoms stabilize adsorbed OH and H₂O species. DFT predicts that molecularly-adsorbed H₂ dissociates on O-rich IrO₂(110) at low temperature and that the resulting H-atoms redistribute to produce a mixture of HO_br and HO_ot groups, with equilibrium favouring HO_ot groups. Our calculations further predict that subsequent H₂O evolution occurs through the recombination of HO_br/HO_ot and HO_ot/HO_ot pairs, and that these reactions represent desorption-limited pathways because the dissociative chemisorption of H₂O is favoured over molecular adsorption on IrO₂(110). The higher stability of HO_ot groups and their preferred formation causes the higher-barrier HO_ot/HO_ot recombination reaction to become the dominant pathway for H₂O formation with increasing O_ot coverage, consistent with the experimentally-observed upshift in the H₂O TPRS peak temperature.

Graphical abstract

KEYWORDS:

• Hydrogen oxidation
• IrO₂
• iridium oxide
• dissociative chemisorption
• water desorption
• sigma complex

Introduction

Understanding the interactions of hydrogen with IrO₂ surfaces is important for improving applications of electrocatalytic water splitting and developing IrO₂-based catalysts that can efficiently convert light alkanes to value-added products. Prior studies show that IrO₂ exhibits high activity towards the electrocatalytic conversion of water to hydrogen and oxygen [1-4], and thus motivate fundamental investigations of the chemical properties of well-defined IrO₂ surfaces towards species that are involved in water-splitting chemistry (e.g., H₂, O₂, H₂O). Recent studies also demonstrate that the IrO₂(110) surface efficiently promotes the C-H bond cleavage of methane, ethane and propane at low temperatures (<150 K) [5-7], and that the resulting alkyl groups react at higher temperature during temperature-programmed reaction spectroscopy (TPRS). Our prior results show that surface H-atoms can have a strong influence on the chemistry of alkyl groups on IrO₂(110), and thus further motivate efforts to clarify the fundamental interactions between hydrogen and IrO₂ surfaces.

Late transition-metal oxides can be highly active for oxidizing H₂, hydrocarbons and other compounds, provided that the surfaces expose pairs of coordinatively-unsaturated (cus) metal and oxygen atoms [8,9]. The stoichiometric termination of rutile RuO₂(110) and IrO₂(110) surfaces exposes equal quantities of cus-metal and oxygen atoms, where the cus-oxygen atoms are referred to as bridging O-atoms (O_br) because each O_br atom occupies a bridging site relative to two underlying metal atoms in the oxide lattice. Oxygen atoms can also bond directly on-top of the cus-metal atoms to produce a singly-coordinate species referred to as an on-top oxygen atom (O_ot). Prior studies show that O₂ dissociation occurs efficiently on RuO₂(110) at room temperature, and can generate variable coverages of O_ot atoms that coexist with Ru_cus and O_br atoms [8-10]. These studies reveal that both O_br and O_ot atoms can be highly reactive, but that O_ot atoms are generally more active as H-atom acceptors. Studies to clarify the reactivity of O_br vs. O_ot atoms are important for developing atomic-level descriptions of oxidation catalysis promoted by IrO₂(110).

Past investigations demonstrate that stoichiometric (s-) RuO₂(110) is reactive for oxidizing H₂ and that O_ot atoms significantly enhance the surface activity [11-17]. At low temperature, hydrogen dissociation initiates from H₂ molecularly-adsorbed on Ru_cus atoms, where the H₂ molecule datively bonds with a Ru_cus atom to produce an adsorbed H₂ σ-complex [11,16-19]. Experiments show that H₂O evolves in a reaction-limited TPRS feature (~500 to 700 K) during H₂ oxidation on s-RuO₂(110), such that H₂O forms at temperatures higher than that required for chemisorbed H₂O to desorb from the surface. The rate-limiting step in H₂ oxidation on s-RuO₂(110) has been attributed to the decomposition of stable HO_br groups [12]. In contrast, O_ot-atoms promote the desorption-limited evolution of H₂O (~400 K) during H₂ oxidation on RuO₂(110) [17], demonstrating that the O_ot atoms provide low-energy pathways for H₂O formation. Knapp et al. report that O_ot atoms efficiently harvest H-atoms from O_br atoms, resulting in the formation of chemisorbed H₂O_ot species at low temperature [13-15]. We have recently found that H₂ adsorption and oxidation occurs on s-IrO₂(110) by a similar mechanism as reported for s-RuO₂(110) [20]. A key difference is that H₂ interacts more strongly with Ir_cus compared with Ru_cus atoms, resulting in a higher molecular H₂-binding energy on IrO₂(110) and facile H₂ dissociation at 90 K. After H₂ dissociation, H₂O evolves from IrO₂(110) in a reaction-limited TPRS feature (~400–750 K) attributed to the recombination of HO_br groups.

In the present study, we investigated H₂ oxidation on O-rich IrO₂(110) using TPRS and DFT calculations. We find that O_ot atoms on IrO₂(110) promote H₂ oxidation during TPRS by facilitating desorption-limited H₂O evolution but that O_ot atoms also stabilize HO_ot and H₂O_ot species, resulting in higher barriers for H₂O desorption from O-rich vs. stoichiometric IrO₂(110). We present evidence that H-atoms transfer rapidly from O_br to O_ot atoms on IrO₂(110) and that the desorption-limited evolution of H₂O occurs through recombination of HO_ot/HO_br as well as HO_ot/HO_ot pairs, with the latter pathway becoming increasingly important with increasing O_ot coverage.

Experimental details

Details of the UHV analysis chamber with an isolatable ambient-pressure reaction cell utilized in the present study have been reported previously [5]. Briefly, the Ir(100) crystal employed in this study is a circular disk (9 mm × 1 mm) that is attached to a liquid-nitrogen-cooled, copper sample holder by 0.015” W wires that are secured to the edge of crystal. A type K thermocouple was spot welded to the backside of the crystal for temperature measurements. Resistive heating, controlled using a PID controller that varies the output of a programmable DC power supply, supports linearly ramping from 80 to 1500 K and maintaining the sample temperature. Sample cleaning consisted of cycles of Ar⁺ sputtering (2000 eV, 1.5 μA) at 1000 K, followed by annealing at 1500 K for several minutes. The sample was subsequently exposed to 5 × 10⁻⁷ Torr of O₂ at 900 K for several minutes to remove surface carbon, followed by flashing to 1500 K to remove final traces of oxygen. We considered the Ir(100) sample to be clean when we obtained sharp low-energy electron diffraction (LEED) patterns consistent with the reconstructed (5 × 1) structure and did not detect impurities using AES and detected negligible CO and CO₂ production during flash desorption after adsorbing oxygen.

We generated an IrO₂(110) film by exposing Ir(100) to 5 Torr O₂ (Airgas, 99.999%) for a duration of 10 min (3 × 10⁹ Langmuir) in the ambient-pressure reaction cell at a surface temperature of 765 K. Our ambient-pressure reaction cell is designed to reach elevated gas pressure while maintaining UHV in the analysis chamber [5]. After preparation of the oxide film, we lowered the surface temperature to 600 K, and then evacuated O₂ from the reaction cell and transferred the sample back to the UHV analysis chamber. We exposed the film to ~23 L of O₂ while cycling the surface temperature between 300 and 650 K to fill oxygen vacancies that may have been created during sample transfer from the reaction cell to the analysis chamber. This procedure produces a high-quality IrO₂(110) surface that has a stoichiometric surface termination and consists of ~10 layers of IrO₂(110), corresponding to a thickness of 3.2 nm.

Figure 1 shows a ball and stick model of IrO₂(110) partially covered with O_ot atoms. The IrO₂(110) surface unit cell is rectangular, with bulk-terminated dimensions of a = 3.16 Å and b = 6.36 Å, and the surface consists of alternating rows of Ir_cus and O_br atoms along the [001] direction. Each of these surface species has a single dangling bond due to the decrease in bond coordination relative to bulk IrO₂; the Ir_cus atoms have 5-fold coordination whereas bulk Ir atoms have 6-fold coordination, and O_br atoms have 2-fold coordination whereas bulk O-atoms have 3-fold coordination. One may consider that formation of the stoichiometric termination of IrO₂(110) involves cleaving equal numbers of Ir-O and O-Ir bonds, and thus generating equal quantities of Ir_cus and O_br atoms at the surface. On the basis of the IrO₂(110) unit cell, the areal density of Ir_cus atoms and O_br atoms is equal to 37% of the Ir(100) surface atom density of 1.36 × 10¹⁵ cm². Since Ir_cus atoms are active adsorption sites, we define 1 ML as equal to the density of Ir_cus atoms on the IrO₂(110) surface. On-top oxygen atoms (O_ot) bond directly on Ir_cus atoms and also expose a dangling bond perpendicular to the surface (Figure 1). Similar to the behaviour of oxygen on RuO₂(110), TPD measurements show that O_ot atoms on IrO₂(110) are less stable that O_br atoms with O_ot atoms desorbing between 400 and 650 K and O_br atoms desorbing in a TPD peak near 950 K [5,20].

Figure 1. Model representation of top and side views of the IrO₂(110) structure with an O_ot atom. The Ir_cus, Ir_6f, O_br, O_ot and O_3f atoms are indicated, where Ir_6f and O_3f corresponds to 6-fold and 3-fold coordination. Red and blue atoms represent O and Ir atoms, respectively, and the O_ot atom is shown in orange.

We investigated the adsorption and reactivity of H₂ (Airgas, 99.999%) on O_ot-covered (‘O-rich’) IrO₂(110) using TPRS. After a hydrogen exposure, we positioned the sample in front of a shielded mass spectrometer at a distance of ~5 mm and heated at a constant rate of 1 K/s until reaching a sample temperature of 800 K. Because the IrO₂(110) starts to decompose above 650 K, it was necessary to prepare a fresh IrO₂(110) film for each adsorption/reaction experiment. Reproducibility in our TPRS results provides evidence that we can repeatedly generate IrO₂(110) films with nominally the same surface structure and composition. We quantified adsorbate coverages and desorption yields from TPRS using established procedures as described in the SI.

Computational details

All plane wave DFT calculations were performed using the projector augmented wave pseudopotentials [21] provided in the Vienna ab initio simulation package (VASP) [22,23]. The Perdew-Burke-Ernzerhof (PBE) exchange-correlation functional [24] was used with a plane wave expansion cutoff of 400 eV. We used four layers to model the IrO₂(110) film which has an ~12 Å thick slab. The PBE bulk lattice constant of IrO₂ (a = 4.54 Å and c = 3.19 Å) is used to fix the lateral dimensions of the slab. The bottom two layers are fixed, but all other lattice atoms are allowed to relax during the calculations until the forces are less than 0.05 eV/Å. A vacuum spacing of ~25 Å was included, which is sufficient to reduce the periodic interaction in the surface normal direction. In terms of system size, a $2\times 4$ unit cell with a corresponding $2\times 2\times 1$ Monkhorst-Pack k-point mesh is employed. We also have performed select tests of H₂, O₂ and H₂O adsorption energies with a large k-point mesh of 4 $\times$ 4 $\times$ 1. The difference with the larger k-point mesh is less than ~ 0.1 kJ/mol. Unless otherwise noted, our DFT calculations were performed for a single H₂ molecule adsorbed within the $2\times 4$ surface model of IrO₂(110), and corresponds to an H₂ coverage equal to 12.5% of the total density of Ir_cus atoms and 25% of the Ir_cus density within one Ir_cus row. In the present study, we define the binding energy, $E_b$, of an adsorbed H₂ molecule on the surface using the expression,

$E_b=\left(E_{H_2}+E_{surf}\right)-E_{H_2/surf}$

where $E_{H_2/surf}$ is the energy of the state containing the adsorbed H₂ molecule, $E_{surf}$ is the energy of the bare surface, and $E_{H_2}$ is the energy of an isolated H₂ molecule in the gas phase. From the equation above, a large positive value for the binding energy indicates a high stability of the adsorbed H₂ molecule under consideration. We evaluated the barriers for H₂ oxidation on the oxygen-rich IrO₂(110) surface using the climbing nudged elastic band (cNEB) method [25]. All the activation energy barriers that we report here have an imaginary vibrational frequency at the transition state. All reported binding energies and energy barriers are corrected for zero-point vibrational energy.

Experimental results

Chemisorption of O₂ on IrO₂(110) film at 85 K versus 300 K

Figure 2(a,b) show O₂ TPD spectra obtained after adsorbing various quantities of oxygen on IrO₂(110) at temperatures of 85 K and 300 K, respectively. Total oxygen coverages (atomic + molecular species) are specified in Figure 2 and given in units of ML of O-atoms. After saturating IrO₂(110) with oxygen at 85 K, the O₂ TPD spectrum exhibits two main features, one below 200 K and the other between ~400 and 650 K (Figure 2(a)). We have previously assigned these low and high temperature TPD features to the desorption of molecularly-adsorbed O₂ and the recombinative desorption of O_ot atoms on the Ir_cus rows, respectively [20]. Analogous O₂ desorption behaviour from RuO₂(110) supports these assignments [8-10]. Furthermore, we find that a maximum of 0.86 ML (O-atom basis) of O₂ desorbs in the features below 650 K, in agreement with the jamming coverage predicted by the random sequential adsorption (RSA) model for immobile dimers adsorbing on a linear array of sites. We note that an average of 0.11 ± 0.02 ML of H₂O desorbs from the O₂-exposed surfaces due to the adsorption of H₂ from the vacuum background before or during the O₂ exposures (Supporting Information, Figure S1). Initial TPRS experiments with ¹⁸O₂-exposed IrO₂(110) show that hydrogen reacts nearly exclusively with on-top oxygen when hydrogen is the limiting reactant. Thus, since the background hydrogen uptake is less than the lowest O_ot coverage that we studied, it is reasonable to assume that the initial O_ot coverage is equal to the sum of the H₂O and O₂ desorption yields below 650 K. Here, we describe the evolution of the O₂ TPD spectra with increasing oxygen coverage on IrO₂(110) and compare the behaviour with that reported previously for oxygen adsorbed on RuO₂(110).

Figure 2. O₂ TPD spectra obtained after exposing a s-IrO₂(110) film to varying quantities of O₂ at (a) 85 K and (b) 300 K.

After adsorption at 85 K, O₂ initially desorbs in a TPD feature (δ₂) centred at ~570 K that broadens and shifts to ~550 K as the on-top oxygen coverage initially increases (Figure 2(a)). A second feature (δ₁) develops and downshifts from ~500 to 470 K with increasing oxygen coverage above ~0.30 ML. The broadening of the O₂ TPD feature towards lower temperatures suggests that neighbouring O_ot atoms interact repulsively, and thereby lower the energy barrier for recombinative O_ot desorption. Prior studies also report strong repulsive interactions between O_ot atoms on RuO₂(110) [26,27], and a significant broadening of the O₂ TPD feature with increasing O_ot coverage [28]. The total desorption yield in the δ₁ + δ₂ feature is equal to 0.46 ML at saturation of the on-top oxygen layer at 85 K. Adsorption of O₂ at 300 K also produces a δ TPD feature that saturates at an O_ot coverage of 0.44 ML (Figure 2(b)).

Molecular oxygen begins to desorb in several, overlapping TPD features (γ_{1 –} γ₄) below ~300 K, with the features developing sharply as the oxygen coverage increases above ~0.35 ML at 85 K after the δ₁ feature first appears. We label the O₂ TPD maxima at ~102, 125, 150 and 180 K as γ_1, γ_2, γ₃ and γ₄ and note that these features populate nearly uniformly when they first become evident but that the γ₂ peak at 125 K becomes dominant as the coverage increases towards saturation. The small shoulder on the trailing edge of the γ₄ peak may arise from O₂ adsorbed on defect sites or a minority phase within the oxide layer. The appearance of multiple, low-temperature TPD features may indicate that molecularly-adsorbed O₂ adopts several binding geometries and that the binding energies are also sensitive to the local environment at the high oxygen coverages at which these species exist on IrO₂(110). Our results demonstrate that O₂ dissociation to produce O_ot atoms is facile on the IrO₂(110) surface. This behaviour is analogous to O₂ on RuO₂(110); however, an important difference is that O₂ dissociation produces a lower saturation coverage of O_ot atoms on IrO₂(110) compared with RuO₂(110) (~0.45 ML vs. 0.86 ML) in UHV. An implication is that molecular adsorption becomes favoured over O₂ dissociation at lower O_ot coverage on IrO₂(110) relative to RuO₂(110). Also, molecularly adsorbed O₂ desorbs in a well-defined peak from RuO₂(110) [10], whereas molecular O₂ desorbs in multiple TPD features from IrO₂(110). This difference is likely a consequence of the larger amount of molecularly-adsorbed O₂ that desorbs from IrO₂(110) rather than dissociating.

Lastly, the appearance of the low temperature O₂ TPD features (<~ 200 K) mainly after the recombinative O₂ TPD feature (~400 to 600 K) saturates demonstrates that molecularly adsorbed O₂ predominantly dissociates on IrO₂(110) at coverages up to ~0.35 ML and may further indicate that the molecularly adsorbed species dissociate at temperatures below ~200 K, i.e., below the temperatures at which O₂(ad) species desorb. Our DFT calculations support this conclusion as they predict a barrier of only 14 kJ/mol for O₂(ad) to dissociate on IrO₂(110) for an O₂ coverage of 0.25 ML (O-atom basis). A possible implication is that the desorption of molecularly adsorbed O₂ occurs only when O_ot atoms are also present in large amounts on the Ir_cus rows. Further study is needed to examine how O_ot atoms influence the binding of O₂(ad) species on the IrO₂(110) surface.

TPRS spectra of H₂ oxidation on O-rich IrO₂(110)

Figure 3(a) shows H₂ and H₂O TPRS traces obtained after exposing a saturation amount of H₂ to IrO₂(110) surfaces held at 85 K and initially covered with different quantities of on-top oxygen. We prepared the O-rich IrO₂(110) surfaces by first saturating with O₂ at 85 K, followed by flashing the surface to temperatures selected to desorb a specific fraction of the on-top oxygen. Hydrogen oxidation on stoichiometric-IrO₂(110) produces H₂O over a wide temperature range that can be decomposed into a TPRS peak near 490 K (α₁) and a broad feature between ~500 and 750 K (α₂). We have previously attributed the α₁ and α₂ features to the desorption-limited evolution of H₂O (see Figure S2) and the reaction-limited recombination of HO_br groups, respectively [20]. At saturation of the H₂ layer on s-IrO₂(110), the H₂ TPRS spectrum also exhibits small peaks at ~200 and 530 K that have been assigned to H₂ desorbing from a molecularly-adsorbed σ-complex state and the recombination of H-atoms, respectively. We estimate an H₂ saturation coverage of ~0.68 ML on s-IrO₂(110) at 85 K and find that more than 90% of the initially adsorbed H₂ σ-complexes dissociate during TPRS with the majority oxidizing to H₂O.

Figure 3. (a) H₂ and H₂O TPRS spectra obtained after exposing a saturation amount (0.5 L) of H₂ to IrO₂(110) surfaces held at 85 K and initially covered with different quantities of on-top oxygen. (b) O₂ TPD spectra obtained from initial (solid) and H₂-saturated (dashed) IrO₂(110) surfaces with initial on-top oxygen coverages of 0.42 ML (blue) and 0.65 ML (orange). (c) Initial H₂ coverage (black), H₂O TPRS yields (red) and H₂ TPRS yields (green) as a function of the initial O_ot coverage.

Our TPRS results show that on-top oxygen species promote the desorption-limited evolution of H₂O (α₁) during H₂ oxidation on IrO₂(110) while suppressing reaction-limited H₂O production (α₂). This finding demonstrates that on-top oxygen species provide low-energy pathways for the oxidation of H₂ to adsorbed H₂O or HO species on IrO₂(110) such that these adsorbed species form at temperatures lower than that at which adsorbed water desorbs during TPRS. Increasing the initial O_ot coverage to 0.10 ML causes the H₂ TPRS peaks to nearly vanish while the desorption-limited H₂O peak at 490 K intensifies proportionally and the reaction-limited H₂O feature remains at about the same intensity as observed during H₂ oxidation on s-IrO₂(110). The desorption-limited H₂O peak continues to intensify as the on-top oxygen coverage increases from 0.10 to ~0.40 ML while the reaction-limited H₂O feature concurrently diminishes to a negligible level. The desorption-limited H₂O peak also shifts from 490 to ~540 K with increasing O_ot-coverage to about 0.40 ML, and the H₂ TPRS peaks completely vanish once the initial O_ot coverage surpasses ~0.20 ML. The enhancement in the desorption-limited H₂O yield shows that O_ot atoms on IrO₂(110) promote H₂ oxidation to adsorbed H₂O or HO below the temperatures at which H₂O desorption becomes appreciable during TPD. Further, the substantial upshift in the H₂O peak temperature suggests that increasing quantities of the adsorbed H₂O or HO products are stabilized as the initial O_ot coverage increases. The desorption-limited H₂O peak upshifts to 550 K but also diminishes sharply as the initial O_ot coverage increases above 0.40 ML.

Figure 3(b) shows O₂ TPD traces obtained from initial vs. H₂-exposed IrO₂(110) surfaces with initial on-top oxygen coverages of 0.42 and 0.65 ML. We exposed each O-rich IrO₂(110) surface to a dose of H₂ (0.5 L) that is sufficient to saturate the s-IrO₂ (110) surface with H₂ at 90 K, and note that the H₂ saturation coverage decreases with increasing initial O_ot coverage due to site blocking, as discussed further below. Comparison of the O₂ TPD traces (solid vs. dashed) reveals that the leading-edge component (δ₁) of the O_ot-atom recombination feature preferentially diminishes during H₂ oxidation. This behaviour suggests that partial hydrogenation of O_ot-covered IrO₂(110) reduces the population of O_ot-atoms that are located in the interior of O_ot-atom chains where they are destabilized by neighbouring O_ot atoms. Below, we report DFT calculations which predict that the hydrogenation of O_ot atoms to HO_ot groups is thermodynamically favourable and could thus eliminate repulsive interactions among neighbouring O_ot atoms. While the O₂ TPD traces provide clear evidence that O_ot atoms are reactive towards H₂, they are inconclusive about the reactivity of O_2,ot because the traces exhibit only small decreases in the molecular O₂ desorption yields after H₂ adsorption. The O₂ TPD data also demonstrate that a smaller quantity of H₂ adsorbs and oxidizes for initial on-top oxygen coverage of 0.65 ML compared with 0.42 ML. This behaviour is expected because H₂ dissociation requires vacant Ir_cus sites and the coverage of such sites decreases as the initial coverage of on-top oxygen species increases. We note, however, that the H₂ coverage decreases more sharply than expected as the on-top oxygen coverage increases above ~0.4 ML. We discuss the sharp decrease in H₂ uptake in more detail below.

Figure 3(c) shows the initial H₂ coverage and the H₂O and H₂ desorption yields obtained during TPRS as a function of the initial on-top oxygen coverage on IrO₂(110), where the initial H₂ coverage is taken as the sum of the H₂O and H₂ TPRS yields. The figure also shows the temperatures at which the surface was heated during preparation of the initial on-top oxygen layer. The yield data shows that all the adsorbed H₂ oxidizes to H₂O during TPRS when the initial O_ot-coverage is greater than ~0.10 ML and the H₂ coverage is saturated. The initial H₂ coverage first decreases slowly but exhibits an abrupt decrease at an O_ot coverage of ~0.4 ML before settling at a value of ~0.16 ML at saturation of the on-top oxygen layer. The H₂ coverage is expected to decrease in proportion to the coverage of O_ot atoms since these species occupy Ir_cus sites and prevent H₂ adsorption and dissociation. Molecularly-adsorbed O_2,ot species may also be effective in blocking H₂ adsorption, considering that similar amounts of molecular O₂ desorb during TPRS performed with and without adsorbed H₂ (Figure 3(b)).

We attribute the abrupt decrease in H₂ uptake at an O_ot coverage of ~0.4 ML to the dissociative adsorption of background H₂ before preparation of the on-top oxygen layer and the effect of preparation temperature on the site occupancy of the resulting H-atoms. Background H₂ adsorption may also cause the H₂ coverage to settle at a value of 0.16 ML at saturation of the on-top oxygen layer, rather than decreasing further (Figure 3(c)). We find that an average of 0.16 ± 0.03 ML of H₂ dissociatively adsorbs on IrO₂(110) from the vacuum background when the on-top oxygen layer is prepared by heating an O₂-saturated surface to selected temperatures. The amount of H₂ uptake from the background is consistent with an initial H₂ dissociative adsorption probability near unity. Our prior study shows that H₂ dissociates efficiently to produce H_ot atoms and HO_br groups on s-IrO₂(110) at 85 K, whereas the subsequent hopping of H_ot atoms onto O_ot or O_br atoms involves a higher energy barrier of 85 to 90 kJ/mol according to DFT. This prediction suggests that H_ot atoms will occupy Ir_cus sites until the surface reaches a sufficient temperature (~350–400 K) to promote H_ot hopping to O-atoms. Consistent with this idea, we have previously presented evidence that saturating s-IrO₂(110) with H₂ at 85 K, followed by heating to 380 K enables the surface to accommodate additional H₂ at 85 K because H_ot atoms transfer to O_br sites and thereby vacate Ir_cus sites needed for H₂ adsorption [20]. In the present study, we find that the abrupt decrease in H₂ coverage near [O_ot] ~ 0.40 ML coincides with a decrease in the O_ot-preparation temperature below about 450 K (Figure 4(c)). This sudden decrease in the H₂ uptake likely occurs because the rate of H-transfer from Ir_cus to O atoms decreases abruptly below ~450 K, resulting in a proportional decrease in the coverage of Ir_cus sites available for H₂ adsorption at 85 K. In support of this interpretation, we find that the H₂ uptake decreases at a more uniform rate with increasing O_ot coverage in TPRS experiments for which we prepared the O_ot layer by performing variable O₂ exposures at 300 K (Figures S3 and S4). Overall, the H₂ uptake behaviour provides further evidence that H₂ oxidation on O-rich IrO₂(110) initiates from the H₂ σ-complex state and thus requires vacant Ir_cus sites to proceed at the conditions studied.

Figure 4. Images showing computed structures of molecular and dissociated H₂O chemisorbed on IrO₂(110) and binding energies defined relative to the specified initial states. (a) H₂O_ot species and (b) a HO_br/HO_ot pair resulting from molecular vs. dissociative chemisorption of an H₂O molecule on s-IrO₂(110). (c) HO_br/HO_ot-H₂O_ot dimer and (d) HO_br/HO_ot + HO_br/HO_ot dimer with binding energies per H₂O defined relative to 2H₂O(g) + s-IrO₂(110). (e) O_ot-H₂O_ot species and (f) a HO_ot/HO_ot pair resulting from molecular vs. dissociative chemisorption of an H₂O molecule next to an O_ot atom on IrO₂(110). (g) HO_ot-H₂O_ot species and (h) HO_ot/HO_br+ HO_ot species resulting from molecular vs. dissociative chemisorption of an H₂O molecule next to a HO_ot group on IrO₂(110).

Computational results

Oxygen adsorption and dissociation on IrO₂(110)

We performed DFT calculations to investigate the adsorption of O₂ on s-IrO₂(110). We examined O₂ on atop and bridge sites of the Ir_cus row with different orientations and our calculations predict that the most favourable configuration of adsorbed O₂ is a flat-lying, peroxy-species in which each O-atom bonds with an adjacent Ir_cus atom. The peroxy-species is the preferred configuration of molecularly-adsorbed O₂, and has a computed binding energy of 182 kJ/mol at an O₂ coverage of 0.125 ML (O-atom basis) in the 4 × 2 IrO₂(110) model (see Computational Details), which corresponds to 0.50 ML of O-atoms along an Ir_cus row. The computed binding energy of the peroxy species is higher than the binding energies (~50 kJ/mol) suggested by the O₂ TPD peaks observed below 200 K, due to both coverage effects as well as overbinding of molecularly adsorbed O₂ predicted by the PBE functional [29]. Notably, recent results provide evidence that PBE is more accurate for predicting the binding energies of adsorbed atomic oxygen [30]. We find that the adsorbed peroxy species can dissociate to produce O_ot atoms by overcoming a barrier of only 14 kJ/mol and that the exothermicity of reaction is 190 kJ/mol relative to the energy of a gas-phase O₂ molecule. The calculations thus predict that O₂ should dissociate efficiently at low coverage to generate O_ot atoms. Figure S5 shows the computed energy pathway and molecular images for O₂ dissociation on s-IrO₂(110). We also find that the reaction exothermicity decreases to 157 kJ/mol for dissociation of 0.50 ML of adsorbed O₂, demonstrating that the O_ot-Ir_cus binding becomes weaker at high O_ot coverage. These results are consistent with our experimental observations that O₂ dissociates efficiently on IrO₂(110) at low coverage and that the O_ot-atoms become less stable with increasing coverage. We leave to future work a more detailed DFT study of O₂ dissociation with increasing coverage, and instead focus on the effect of O_ot atoms on H₂ oxidation processes on the IrO₂(110) surface.

Water chemisorption on IrO₂(110)

We investigated H₂O chemisorption on IrO₂(110) using DFT to determine whether H₂O adsorbs molecularly or dissociatively at low coverage and characterize the interactions between H₂O and co-adsorbed O_ot atoms and HO_ot groups. A key aim of these calculations is to obtain information that can aid in understanding the mechanism for H₂ oxidation on O-rich IrO₂(110). Figure 4 shows images of computed configurations of H₂O and HO groups adsorbed next to different adsorbates on IrO₂(110) and the corresponding binding energies determined using PBE. We also examined other configurations of different orientations for H₂O and OH; however, we found that configurations forming hydrogen bonds (see Figure 4) are more energetically stable than other configurations.

Our calculations predict that an isolated H₂O_ot molecule binds on IrO₂(110) by positioning its O-atom on-top of an Ir_cus atom and adopting an orientation that allows one of the H-atoms to hydrogen-bond with a neighbouring O_br atom (Figure 4(a)). The large binding energy of the chemisorbed H₂O molecule (136 kJ/mol) is indicative of a strong, localized bonding interaction between H₂O and IrO₂(110). We predict that the dissociative chemisorption of H₂O on s-IrO₂(110) occurs by proton transfer to an O_br atom to produce a HO_br/HO_ot pair in which the HO_ot group aligns its H-atom towards the O_br atom of the HO_br group (Figure 4(b)). We find that dissociation of an isolated H₂O_ot molecule on s-IrO₂(110) occurs with a negligible barrier, and is exothermic by 167 kJ/mol relative to the gaseous H₂O reference state. According to DFT, the dissociative chemisorption of H₂O on s-IrO₂(110) is favorable over molecular adsorption by 31 kJ/mol at low coverage. We also investigated the stabilities of H₂O-derived dimers on s-IrO₂(110), and find that the dissociated dimer, 2(HO_br/HO_ot), is more stable than the mixed HO_br/HO_ot-H₂O_ot dimer on IrO₂(110) (Figure 4(c,d)). However, the dissociated dimer is only ~3 kJ/mol more stable than the mixed dimer (per H₂O molecule), compared with the larger stability of a single HO_br/HO_ot pair relative to a single H₂O_ot species (~31 kJ/mol). This comparison reveals that hydrogen-bonding between HO_ot and H₂O_ot acts to stabilize the HO_br/HO_ot-H₂O_ot dimer, and causes the mixed and dissociated dimers to have similar stabilities on IrO₂(110).

Our calculations predict that an O_ot atom stabilizes a neighbouring H₂O_ot molecule on IrO₂(110) and that dissociative chemisorption of H₂O is favoured over molecular adsorption on the O-rich surface. The H₂O_ot molecule adsorbed next to an O_ot atom adopts a flat-lying orientation with one H-atom directed towards the O_ot atom and the other directed towards an O_br atom, and has a binding energy of 178 kJ/mol relative to the gaseous H₂O reference state (Figure 4(e)). Our results indicate that the O_ot-H₂O_ot interaction stabilizes the H₂O_ot molecule by 42 kJ/mol relative to H₂O_ot adsorbed on s-IrO₂(110) at low coverage. Dissociative chemisorption of H₂O on the O_ot-IrO₂(110) surface produces a HO_ot/HO_ot pair and is exothermic by 197 kJ/mol (Figure 4(f)), corresponding to a stabilization of 19 kJ/mol over molecular adsorption. Our calculations predict that dissociative chemisorption is slightly favoured (by ~4 kJ/mol) over molecular adsorption when H₂O adsorbs next to a HO_ot group on IrO₂(110) (Figure 4(g,h)). This situation is similar to dimer formation on s-IrO₂(110) (Figure 4(c,d)). Interestingly, we find that H₂O adsorption next to a HO_ot group is more stabilizing than adsorption next to an HO_ot/HO_br pair (Figure 4(c,g)), with the added stability equal to about 12 kJ/mol. Key findings of our calculations are that H₂O dissociation is favoured on IrO₂(110) at low H₂O coverage, and that H₂O achieves stronger binding on O-rich IrO₂(110).

Comparison with prior studies reveals similar characteristics in H₂O adsorption on IrO₂(110) and RuO₂(110). For example, the dissociative adsorption of a single H₂O molecule is favoured over molecular adsorption on both oxides though DFT predicts that the HO_br/HO_ot pair is only ~3 kJ/mol more stable than a single H₂O_ot molecule on RuO₂(110) [31] compared with 31 kJ/mol for IrO₂(110). A recent study also shows that formation of a HO_ot/HO_ot pair is favoured during H₂O adsorption on O-rich RuO₂(110) with an exothermicity that is ~20 kJ/mol higher than adsorption of a H₂O molecule that is separated from the O_ot atom [32]; the exothermicity is 60 kJ/mol for the analogous comparison of HO_ot/HO_ot vs. H₂O_ot on O-rich IrO₂(110) (Figure 4(a,f)). The larger energy differences between molecular and dissociative H₂O adsorption demonstrate that IrO₂(110) is more active than RuO₂(110) in promoting H₂O deprotonation.

Mechanism for H₂ oxidation on s-IrO₂(110)

We recently investigated the oxidation of H₂ on s-IrO₂(110) using TPRS and DFT, and find that the reaction mechanism can be summarized using a few key steps as shown in Table 1. Our prior results demonstrate that H₂ adsorbs strongly on Ir_cus sites (step 1) and that the molecularly-adsorbed H₂ species (H_2,ot) readily dissociates via H-transfer to an O_br atom, resulting in H_ot and HO_br species (step 2). We find that H₂ binds to the Ir_cus atom through a strong dative bonding interaction, resulting in a molecularly-adsorbed H₂ σ-complex with a computed binding energy of 78 kJ/mol. DFT calculations predict that the H_2,ot species dissociates with a barrier of only 13 kJ/mol and that the reaction is exothermic by 113 kJ/mol. These energetics suggest that H₂ dissociation on s-IrO₂(110) occurs with high probability at temperatures as low as 85 K. According to DFT, H_ot atom hopping to an O_br site is the predominant reaction step after H₂ dissociation, and will nominally cause all the H_ot atoms to convert to HO_br groups prior to H₂O formation. The hopping of a H_ot atom to an O_br site is exothermic by 36 kJ/mol and features a barrier of 85 kJ/mol.

Table 1. Key reaction steps involved in H₂ oxidation on s-IrO₂(110) as identified by DFT. Energy barriers (E_a) and energy changes of reaction (ΔE) are shown for each step.

Description	Reaction	Ea (kJ/mol)	ΔE (kJ/mol)
1) Adsorption and H2 σ-complex formation	H2(g) → H2,ot	–	−78
2) Dissociation of adsorbed H2	H2,ot + Obr → Hot + HObr	13	−113
3) Hopping of H-atom from Ircus to Obr site	Hot + Obr → HObr	85	−36
4) Water formation from HObr groups	2HObr → H2O(g) + Ov	–	220–260

We predict that the recombination of HO_br groups to produce gaseous H₂O can occur by several pathways that are endothermic by between ~220 and 260 kJ/mol [20]. These pathways are the most energetically demanding among the sequence of H₂ oxidation steps that we identified, and are expected to control the overall rate of H₂ oxidation to H₂O over s-IrO₂(110). The large barriers for HO_br conversion to gaseous H₂O are generally attributable to the preference for H₂O dissociation and strong binding of HO_br groups on s-IrO₂(110) as well as energetic penalties associated with the creation of bridging oxygen vacancies (O_v). Our prior DFT calculations also predict that the barriers for H-atom diffusion along the Ir_cus and O_br rows are quite large, with values of 199 and 213 kJ/mol, respectively, and comparable to the barriers for H₂O formation. An implication is that H-atoms will be effectively trapped near the original H₂ dissociation site (Ir_cus and adjacent O_br atoms) at moderate temperature and that diffusion away from these sites can only begin to occur at appreciable rates at temperatures at which H₂O formation and desorption also occurs at significant rates.

Mechanism for H₂ oxidation on O-rich IrO₂(110)

We used DFT to identify viable reaction steps that govern H₂ oxidation on O-rich IrO₂(110), and determine how O_ot-atoms promote the conversion of adsorbed H₂ to water. Our calculations reveal two major differences in the H₂ oxidation mechanism on O-rich vs. stoichiometric IrO₂(110). First, we find that the recombination of HO groups to produce gaseous H₂O can occur by lower energy pathways in the presence of O_ot atoms compared with s-IrO₂(110). Our calculations also predict that the diffusion of H-atoms is highly facile from O_br to O_ot atoms and also along the O_ot rows. The fast diffusion favours the formation of HO_ot groups and thus water production by low energy pathways. Below, we describe the reaction steps involved in H₂ oxidation on O-rich IrO₂(110) as determined from DFT calculations.

Table 2 lists key reaction steps involved in H₂ oxidation on O-rich IrO₂(110) as well as the energy barriers and heats of reaction determined from DFT calculations. Figures S6–S11 show energy diagrams and molecular images for each of these reaction steps. We focused mainly on systems with an H₂ coverage of 0.25 ML and an O_ot coverage of 0.25 or 0.50 ML along an Ir_cus row. Exploring numerous H and O_ot coverages and configurations lies outside of the scope of the present study. Our calculations predict similar energetics for the first three reaction steps on O-rich and stoichiometric IrO₂(110). Adsorption of H₂ on an Ir_cus atom located next to an O_ot atom produces a strongly-bound H₂ σ-complex with a binding energy of 85 kJ/mol. The calculations predict that dissociation of the H_2,ot species is highly facile via reaction with either an O_br or an O_ot atom, with barriers less than 20 kJ/mol in each case, but that dissociation via H-transfer to an O_ot atom is more exothermic by 30 kJ/mol. The calculations also predict that H_ot hopping onto a O_ot atom to produce an HO_ot group must overcome a barrier of 89 kJ/mol and is exothermic by 84 kJ/mol. This reaction barrier is similar to that required for H_ot to hop onto an O_br atom (85 kJ/mol) but the reaction exothermicity is higher for the H_ot + O_ot reaction compared with the H_ot + O_br reaction (89 vs. 36 kJ/mol). Overall, we find similar barriers but larger exothermicity for HO_ot vs. HO_br formation as a result of H₂ dissociation and H_ot atom hopping.

Table 2. Key reaction steps involved in H₂ oxidation on O-rich IrO₂(110) as identified by DFT. Energy barriers (E_a) and energy changes of reaction (ΔE) are shown for each step.

Description	Reaction	Ea (kJ/mol)	ΔE (kJ/mol)
1) Adsorption and H2 σ-complex formation	H2(g) + Oot → H2,ot+ Oot	–	−85
2) Dissociation of adsorbed H2	H2,ot + Oot → Hot + HOot	19	−143
3) Hopping of H-atom from Ircus to Oot site	Hot + Oot → HOot	89	−84
4A) Recombination of HObr with a HOot group	HOot + HObr → H2O(g) + Obr	–	167
4B) Recombination of HOot groups	2HOot → H2O(g) + Oot (2HOot → H2Oot + Oot) (H2Oot + Oot → H2O(g) + Oot)	– (26) –	196 (18) (178)
Hopping of a H-atom from HObr to an Oot site	HObr + Oot → Obr + HOot	9	−27
Hopping of a H-atom from HOot to an Oot site	HOot + Oot → Oot + HOot	17	0

Our calculations predict that OH recombination reactions are the most energetically-demanding reaction steps during H₂ oxidation on O-rich IrO₂(110), demonstrating that the H₂O formation steps are rate-limiting in H₂ oxidation on O-rich as well as stoichiometric IrO₂(110). Compared with s-IrO₂(110), the H₂O formation reactions are more facile on O-rich IrO₂(110) and occur by the reverse pathways identified computationally for the dissociative chemisorption of H₂O on IrO₂(110) (Figure 4). This finding is consistent with our conclusion that H₂O evolution is desorption-limited during H₂ oxidation on O-rich IrO₂(110). Among the surface configurations that we studied, the recombination of an HO_ot group with an HO_br group to generate gaseous H₂O and an O_br atom is the most favoured pathway for H₂O formation on O-rich IrO₂(110) and has a thermochemical barrier of 167 kJ/mol (Table 2, step 4A). This reaction is the reverse of the dissociative chemisorption of an H₂O molecule on s-IrO₂(110) (Figure 4(b)). Table 2 also shows that the recombination of HO_ot groups to generate gaseous H₂O and an O_ot atom has an overall thermochemical barrier of 196 kJ/mol (step 4B), and may proceed through the formation of a H₂O_ot-O_ot intermediate. We speculate that HO_ot + HO_br recombination contributes mainly at low O_ot coverage, but that recombination of adjacent HO_ot groups becomes increasingly important at higher O_ot coverage. Such a change in the dominant H₂O formation pathway may explain, at least in part, the upshift in the desorption-limited H₂O TPRS feature that we observe with increasing initial O_ot coverage (Figure 3(a)).

Our calculations predict that H-atom hopping is highly facile from O_br to O_ot atoms and also between adjacent O_ot atoms. Hydrogen atom hopping between O_ot atoms has a computed barrier of 17 kJ/mol, and H-atom hopping from an O_br to an O_ot atom has a barrier of 9 kJ/mol and is exothermic by 27 kJ/mol (Table 2). These fast diffusion processes can enable a quasi-equilibrium to be established between the HO_ot and HO_br groups in local regions of the surface. Equilibrium between HO_ot and HO_br groups favors formation of the HO_ot species at the temperatures of interest (< 750 K) because the HO_br + O_ot → O_br + HO_ot reaction is significantly exothermic (ΔE = −27 kJ/mol). The extent of such equilibration may depend sensitively on the local configurations and species coverages, since other surface species diffuse more slowly compared with H-atom hopping from O_br/O_ot to O_ot atoms [20].

Our DFT calculations suggest a scenario wherein H₂ dissociates efficiently on O-rich IrO₂(110) to produce H_ot and HO species. The resulting surface H-atoms will redistribute via site-hopping processes to produce a mixture of HO_ot and HO_br groups prior to H₂O formation, where the equilibrium composition of this mixture favours HO_ot groups below ~750 K. Water formation then ensues by several HO recombination steps, the most favourable of which is H-transfer from an HO_br to a HO_ot group and evolution of H₂O (step 4A, Table 2). The contributions from the various HO recombination reactions will depend on the relative H and O_ot coverages. For saturation H₂ coverage, as investigated in our experiments, an increasing fraction of H₂O will be produced by the more facile HO_ot + HO_br reaction as the O_ot coverage initially increases, while HO_br recombination will make a decreasing contribution. However, since HO_ot groups are energetically favoured over HO_br groups, water formation by HO_ot + HO_ot recombination (step 4B, Table 2) will begin to dominate above an intermediate O_ot coverage. This scenario would cause the H₂O production rate at a given temperature to initially increase with increasing O_ot coverage as the HO_ot + HO_br reaction makes an increasing contribution, but the H₂O production rate would then decrease with further increases in O_ot coverage because the barrier for HO_ot + HO_ot recombination is higher than that for HO_ot + HO_br recombination (Table 2). Such behaviour is qualitatively consistent with our experimental observations that the desorption-limited H₂O TPRS peak replaces the broad reaction-limited H₂O feature and also shifts towards higher temperature with increasing O_ot coverage (Figure 3(a)).

The present study reveals that H₂ oxidation involves analogous steps on O-rich IrO₂(110) and RuO₂(110). Prior work specifically demonstrates that H-transfer from O_br to O_ot atoms is facile and leads to the production of H₂O_ot species at low temperature, followed by the desorption-limited evolution of H₂O during TPRS [12,15,17]. A recent study of H₂O adsorption on O-rich RuO₂(110) supports interpretations of the H₂ oxidation mechanism on RuO₂(110) as well as IrO₂(110), demonstrating that H-transfer from O_br to O_ot atoms involves a low barrier (~20 kJ/mol) and that HO_ot groups are favoured over HO_br groups by ~16 kJ/mol [32]. Those authors also predict that a HO_ot/HO_ot pair is more stable than a HO_br/HO_ot pair on RuO₂(110) by ~17 kJ/mol. These findings are similar to the results of the present study, though our calculations predict larger energy differences between HO_ot and HO_br groups on IrO₂(110). Notably, Jacobi et al. also find that the desorption-limited H₂O TPRS peak upshifts from ~400 to 425 K with increasing O_ot coverage during H₂ oxidation on RuO₂(110) [12]. We find that the desorption-limited H₂O TPRS peak shifts from 490 to 550 K with increasing O_ot coverage on IrO₂(110) (Figure 3(a)). Comparison suggests a higher energy barrier for H₂O desorption as well as larger stabilization of an HO_ot/HO_ot vs. an HO_br/HO_ot pair on IrO₂(110) compared with RuO₂(110). These findings are consistent with DFT results, and support the interpretation that desorption-limited evolution of H₂O from O-rich RuO₂(110) also occurs by the mechanism that we propose here, namely, that HO_br/HO_ot recombination is important during H₂ oxidation at low O_ot coverage, but that the higher-barrier HO_ot/HO_ot recombination step becomes increasingly dominant as the O_ot coverage increases.

Summary

We investigated the adsorption and oxidation of H₂ on O-rich IrO₂(110) using TPRS and DFT calculations. We find that the dissociative chemisorption of hydrogen occurs efficiently on O-rich IrO₂(110) at low temperature and that O_ot atoms facilitate the subsequent oxidation of adsorbed hydrogen. Our results show that O_ot atoms promote the desorption-limited evolution of H₂O during H₂ oxidation on IrO₂(110) while suppressing reaction-limited H₂O production arising from recombination of HO_br groups. The desorption-limited H₂O TPRS peak initially appears at 490 K but shifts to 550 K with increasing O_ot coverage, demonstrating that O_ot atoms stabilize adsorbed water species. We find that the H₂ uptake on IrO₂(110) decreases with increasing O_ot coverage and interpret this behaviour as evidence that H₂ dissociation initiates from the adsorbed H₂ σ-complex state and thus requires Ir_cus sites to proceed under the conditions studied.

Our DFT calculations predict that O_ot atoms promote H₂ oxidation on IrO₂(110) and that H₂O evolution via desorption-limited processes is rate-determining in H₂ oxidation on O-rich IrO₂(110). The calculations predict that dissociative chemisorption of H₂O is favoured over molecular adsorption on both stoichiometric and O-rich IrO₂(110), and produces HO_br/HO_ot and HO_ot/HO_ot pairs, respectively. Desorption-limited evolution of H₂O from IrO₂(110) thus occurs by the recombination of HO_br/HO_ot or HO_ot/HO_ot pairs, with the latter involving a higher energy barrier (196 vs. 167 kJ/mol) due to the greater stability of HO_ot compared with HO_br groups. According to DFT, H₂ adsorbs molecularly on Ir_cus sites as a strongly-bound σ-complex and dissociates efficiently at low temperature to produce H_ot and HO groups. Our calculations suggest that the H-atoms rapidly redistribute on the surface to generate a mixture of HO_br and HO_ot groups, with equilibrium favouring HO_ot formation at temperatures below at least 750 K. Water evolution subsequently occurs on O-rich IrO₂(110) by the desorption-limited recombination of HO_br/HO_ot and HO_ot/HO_ot pairs. Our results suggest that the lower energy HO_br/HO_ot recombination pathway will be important at low O_ot coverages, while the higher-energy HO_ot/HO_ot recombination pathway will become dominant at higher O_ot coverages due the preferable formation of HO_ot groups prior to water formation. This prediction is consistent with our experimental finding that the desorption-limited H₂O TPRS peak shifts to higher temperature with increasing O_ot coverage.

Supporting information

Estimates of adsorbate coverage and product yields from TPRS; H₂O TPD spectra obtained from O₂-exposed IrO₂(110) at 90 K; H₂O TPD spectra from s-IrO₂(110) after H₂O adsorption at 85 K; TPRS for H₂ oxidation on IrO₂(110) with O_ot-layer prepared at 300 K; H₂ oxidation yields as a function of initial O_ot coverage prepared by heating to selected temperatures vs. O₂ adsorption at 300 K; Computed energy pathway and molecular images for O₂ dissociation on IrO₂(110); Computed energy diagrams and molecular images for key steps involved in H₂ oxidation on O-rich IrO₂(110).

Acknowledgments

We acknowledge the Ohio Supercomputing Center for providing computational resources. We gratefully acknowledge financial support for this work provided by the Department of Energy, Office of Basic Energy Sciences, Catalysis Science Division through Grant DE-FG02-03ER15478.

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary material

Supplemental data can be accessed here.

Additional information

Funding

This work was supported by the Office of Basic Energy Sciences [DE-FG02-03ER15478].