Support ionic potential dictates molybdenum carbide formation and reverse water gas shift activity

The reverse watergas shift reaction—converting CO₂ and H₂ into CO—is a key step for carbon recycling, as CO serves as a crucial feedstock for FischerTropsch synthesis, methanol production, and hydroformylation. However, activating CO₂ and achieving high CO selectivity remain challenging. In a study published in ENG. Chem. Eng., researchers at Tianjin University and collaborators demonstrate that support properties critically determine the insitu formation of active molybdenum carbide from MoO₃ precatalysts, providing a rational framework for catalyst design.

The team first evaluated seven commercial early transition metal oxides. MoO₃ and WO₃ exhibited the highest CO₂ conversions (40.8 % and 33.9 %, respectively), while others showed much lower activity. Thermodynamic calculations revealed that only MoO₃ and WO₃ possess negative Gibbs free energies for carburization by CO. XPS confirmed that Mo₂C formed on the surface of spent MoO₃ (6.8 %), while WO₃ formed W₂C (67.5 %). No carbides were detected on other oxides, establishing a strong correlation between catalytic activity and carbide formation.

To further enhance performance, Mo was supported on six distinct oxides (MgO, γAl₂O₃, SiO₂, TiO₂, ZrO₂, and CeO₂) at 8.4 wt % loading. Mo/SiO₂ exhibited the highest CO₂ conversion (46.5 % at optimal 25.2 wt % loading) and forward rate (388.8 molCO·molmetal⁻¹·h⁻¹), while Mo/CeO₂ and Mo/MgO showed the lowest activities (9.4 % and 7.9 % conversion, respectively). All catalysts maintained 100 % CO selectivity.

XPS analysis of spent catalysts transferred without air exposure revealed that Mo₂C formed on all supported catalysts, but surface concentrations varied dramatically. The forward rate correlated positively with Mo₂C surface content. Kinetic studies showed that active catalysts (Mo/SiO₂, Mo/ZrO₂, Mo/TiO₂, Mo/γAl₂O₃) exhibited similar reaction orders (CO₂: ~0.31, H₂: ~0.66) and activation energies (42.6±5.9 kJ·mol⁻¹), indicating a common Mo₂C active phase. In contrast, Mo/CeO₂ and Mo/MgO displayed distinctly higher activation energies (69.0 and 77.9 kJ·mol⁻¹, respectively), consistent with solidsolutiondominated active sites resistant to carburization.

The ionic potential of the support was identified as a quantitative descriptor for metal oxidesupport interaction strength. On acidic supports with high ionic potential (SiO₂, ZrO₂, TiO₂, γAl₂O₃), weak metal oxidesupport interaction preserves MoO₃ crystallites that readily undergo insitu carburization to Mo₂C. On basic supports with low ionic potential (MgO, CeO₂), strong metal oxidesupport interaction promotes stable solid solutions that suppress carburization. Specific surface area serves as a secondary factor: on amphoteric supports with comparable ionic potential, higher surface area strengthens metal oxidesupport interaction and reduces Mo₂C formation.

Notably, lowering the calcination temperature from 500 °C to 300 °C for Mo/CeO₂ eliminated the CeMo solid solution, restored MoO₃ crystallites, and increased CO₂ conversion fivefold (from 5.2 % to 27.4 %), confirming that inhibiting solid solution formation enhances performance.

This work establishes ionic potential as a practical descriptor for support selection, providing a clear theoretical foundation for designing highperformance Mobased RWGS catalysts.