Tuning the activities of cuprous oxide nanostructures via the oxide-metal interaction

The study addresses how oxide–metal interaction (OMI) governs the structure, stability, and catalytic activity of interfacial sites at cuprous oxide–metal (Cu2O–M) interfaces. While oxide layers often form at the surfaces of bimetallic catalysts under oxidizing conditions and can provide unique catalytic properties, the atomic nature of interfacial sites and how OMI tunes their behavior remains unclear. Cu2O–metal interfaces are of particular interest because Cu centers are dynamic (variable valence and coordination) and coordinatively unsaturated Cu+ cations can be active in oxidation reactions. The authors pose the question: for structurally similar Cu2O monolayer nanostructures on different noble metal substrates (Pt(111), Au(111), Ag(111)), how does OMI influence their thermal stability and activity for CO oxidation, and can a simple electronic descriptor capture this effect?

Prior work established that under oxidizing conditions, less noble components in bimetallic catalysts segregate to form self-limited oxide layers, with related concepts such as strong metal–support interaction (SMSI) and the catalytic uniqueness of oxide–metal interfaces. Coordinatively unsaturated cations, including Cu+, can be highly active (e.g., for ethylbenzene oxidation to acetophenone). Cu-based alloys with group VIII metals (notably Pt) are excellent for oxygen reduction, while alloys with group IB metals (Au, Ag) excel in selective hydrocarbon oxidation. Interfacial copper oxide layers have been observed on PtCu, AuCu, and AgCu under mild oxidation conditions, but atomistic understanding of their active sites and how OMI modulates reactivity is incomplete. The literature also highlights that CO strongly poisons Pt-group metals by blocking O2 activation, whereas Au and Ag bind CO and O2 weakly, complicating low-temperature CO oxidation. These contexts motivate a controlled, atomic-scale study of Cu2O–M interfaces and their OMI.

Model catalysts: Monolayer-thick Cu2O nanostructures (NSs) were grown on Pt(111), Au(111), and Ag(111) single crystals under UHV by evaporating Cu in O2 (∼1×10−7 mbar) at cryogenic temperatures followed by annealing in O2 (~500 K). Alternative preparations included NO2-assisted growth on Au(111) and near-ambient O2 oxidation of Cu/Ag(111) with subsequent UHV annealing to remove Ag–O species. Characterization: Atomic structures were imaged with element-specific STM (ES-STM) using LT-STM at 78 K and NAP-STM at 300 K to follow reactions up to tens of mbar. X-ray photoelectron spectroscopy (XPS) (Mg Kα for single crystals; Al Kα for powders) confirmed chemical states. Reactivity studies (model surfaces): CO adsorption and CO oxidation were probed on Cu2O/Pt(111), Cu2O/Au(111), and Cu2O/Ag(111) under UHV (down to 5×10−8 mbar) and near-ambient pressures (up to ~100 mbar) at 300 K, monitoring reduction fronts and alloy formation. Regeneration (oxidation) of reduced surfaces was performed with O2 at controlled pressures and temperatures. Powder catalysts: PtCu/CB, AuCu/CB, AgCu/CB, and monometallic Pt/CB, Au/CB, Ag/CB were prepared by co-impregnation (nominal Pt 4 wt%, Cu:Pt = 1:4; analogous for Au, Ag) on Vulcan XC-72 carbon. Catalytic CO oxidation was conducted in a fixed-bed reactor with 1% CO, 20% O2, 79% He at 20 mL/min over 20 mg catalyst (space velocity 60,000 mL g−1 h−1), heating from room temperature to 573–673 K (1 K/min). Quasi in situ XPS (reactor attached to XPS) analyzed post-reaction states. Computations: Spin-polarized DFT (VASP, PAW, PBE-GGA, 400 eV cutoff) modeled freestanding and M(111)-supported Cu2O nanoribbons (Cu5O6 stoichiometry; O-terminated zig-zag edges). Transition states were found using CI-NEB. CO gas-phase Gibbs free energy accounted for pressure/entropy (UHV 1×10−7 mbar correction). Calculated metrics included CO oxidation pathways and barriers, O2 dissociation barriers at interfaces, adhesion energy (Eadh), oxygen vacancy formation energy (Eov/Eovf) referenced to CO/CO2, surface rumpling (Δh), and d-band centers (εd) for M(111) and edge Cu+ states to establish scaling relations.

• Structure and stability: ES-STM revealed that monolayer Cu2O NSs on Pt(111), Au(111), and Ag(111) share a Cu3O2 honeycomb lattice (Cu2O(111)-like) with O-terminated zig-zag edges; 5–7 defects appear especially on Ag(111). Thermal stability is substrate-dependent: on Pt(111) Cu2O begins decomposing at ~470 K and fully decomposes by ~630 K (with Cu diffusion/alloying), whereas on Au(111) and Ag(111) Cu2O remains intact after 600 K annealing and coalesces only near 700 K.
• CO adsorption/reactivity (model): On Cu2O/Pt(111), CO adsorbs strongly on exposed Pt sites (including through Cu2O ring holes). At 300 K, 5×10−8 mbar CO initiates Cu2O reduction; 1×10−7 mbar CO for 5 min fully reduces a submonolayer, forming triangular Cu3Ox clusters atop a PtCu3-like alloy layer (apparent height ~2 Å). On Cu2O/Au(111), no CO adsorption at 78 K; at 300 K reduction requires elevated CO pressure (~0.5 mbar at edges; ~6 mbar when Au is fully covered), forming a Cu–Au surface alloy. On Cu2O/Ag(111), no appreciable CO oxidation at 300 K even at 10 mbar; zig-zag edges remain inactive up to ~48 mbar, with reduction only from domain boundaries above ~48 mbar.
• Regeneration (oxidation): Pt–Cu alloy re-oxidizes to ordered Cu2O NSs with O2 as low as 5×10−8 mbar at >400 K. Au–Cu requires ~1 mbar O2 at 300 K for Cu2O growth; on Ag(111), Cu oxidizes to Cu2O in ~0.17 mbar O2, but concurrent Ag oxidation complicates the system.
• Powder catalysts (CO oxidation light-off): Under 1% CO/20% O2/He, PtCu/CB shows significant activity just above 300 K; Pt/CB shows negligible conversion below ~400 K (CO desorption temperature). AuCu/CB is more active than Cu2O powder and Au/CB, with onset >300 K but lower activity than PtCu/CB. Ag/CB is active under O2-rich conditions; forming CuO/Ag interfaces (AgCu/CB post-reaction) diminishes activity relative to Ag/CB. Post-reaction XPS: Cu remains as Cu2O in PtCu/CB and AuCu/CB, but converts to CuO in AgCu/CB.
• Reactivity order and mechanism: Activity trend for CO oxidation is Cu2O/Pt > Cu2O/Au > Cu2O/Ag. A dual-site mechanism operates: CO adsorbs on metal sites and reacts with neighboring lattice O from Cu2O; O2 can dissociate at the interface (especially on Pt).
• DFT mechanistic energetics: The rate-determining step (CO2 formation) barriers are 0.42 eV (Cu2O/Pt), 0.60 eV (Cu2O/Au), and 0.86 eV (Cu2O/Ag). O2 dissociation at the Cu2O/Pt interface has a low barrier (0.39 eV) and is thermodynamically favorable; it is less favorable on Cu2O/Au and Cu2O/Ag. Adhesion energies Eadh for nanoribbons do not correlate with activity (Pt: −0.76 eV; Ag: −0.71 eV; Au: −0.58 eV). Oxygen vacancy formation energies (Eovf) at Cu2O edges strongly correlate with activity and are much more negative on supported NSs (−1.33 to −2.35 eV) than on freestanding edges (−0.39 eV), indicating O removal is facilitated by OMI. A clear scaling exists between Eovf and the substrate d-band center εd: stronger Cu+–M electronic interaction (εd closer to EF) gives more negative Eovf and higher activity, following Pt > Au > Ag. Geometric descriptors like surface rumpling (Δh: Pt 0.46 Å > Ag 0.40 Å > Au 0.34 Å) do not track reactivity.
• Overall: OMI enhances interfacial oxygen activity and stabilizes oxygen vacancies in Cu2O NSs, enabling low-temperature CO oxidation and mitigating CO poisoning on Pt-group metals.

The work demonstrates that for structurally similar monolayer Cu2O nanostructures, the metal substrate dictates catalytic behavior via oxide–metal interaction. On Pt(111), strong OMI both lowers the barrier for interfacial O removal by CO and enables facile O2 activation, establishing a rapid redox cycle at very low pressures. On Au(111), weaker OMI suffices for CO oxidation at near-ambient CO but requires higher O2 pressures to regenerate Cu2O. On Ag(111), even weaker OMI cannot sustain an active Cu+ interfacial state; Cu tends to over-oxidize to CuO and the Cu2O zig-zag edges remain largely inactive under similar conditions. DFT links these trends to the energetics of forming interfacial oxygen vacancies (a key step in CO2 formation) and to the electronic interaction between edge Cu+ and the metal, captured by the substrate d-band center. This electronic descriptor rationalizes why Pt-group metals outperform group IB metals for interfacial CO oxidation and provides a predictive handle to tune OMI by choice of substrate. The findings address the central question of how OMI controls interfacial site reactivity and stability, and they highlight interface design as a strategy to alleviate CO poisoning on Pt by supplying reactive lattice oxygen adjacent to CO adsorption sites.

This study establishes that oxide–metal interaction governs both the activity and thermal stability of Cu2O nanostructures supported on noble metals. Despite identical Cu2O edge structures, CO oxidation activity follows Cu2O/Pt > Cu2O/Au > Cu2O/Ag, correlating with increasingly favorable interfacial oxygen vacancy formation energies and stronger Cu+–metal electronic coupling. The substrate d-band center emerges as a simple descriptor for OMI, predicting the ease of O removal and overall catalytic performance. Interfacial dual-site catalysis on Pt circumvents CO poisoning and enables low-temperature CO oxidation with facile redox cycling. These insights inform the rational design of Cu-based alloy catalysts and broader oxide–metal interfaces for oxidation reactions. Future work could extend the descriptor-based approach to other oxide/metal combinations, explore facet and thickness effects, probe operando dynamics under realistic feeds (including moisture and varying stoichiometry), and incorporate alloying or strain to tune εd and OMI.

• Model systems emphasize monolayer Cu2O on ideal single-crystal M(111) substrates; real catalysts have diverse facets, defects, and particle sizes that may modify OMI and kinetics.
• Near-ambient STM and UHV conditions differ from industrial environments; pressure and composition gaps may influence adsorption coverages and reaction pathways.
• On Ag(111), concurrent Ag oxidation under NAP conditions convolutes interpretation of the Cu2O/Ag interface reactivity.
• DFT with PBE and nanoribbon models may not capture all correlation/dispersion effects; absolute energetics can carry systematic errors, though trends are consistent.
• Powder catalysts exhibited some ripening; morphological changes could impact activity independently of OMI.
• The proposed εd descriptor was validated for Pt, Au, Ag; extrapolation to other metals/alloys requires further verification.