Hydrogen is a promising low-carbon energy carrier, but its low volumetric energy density makes storage and transportation difficult. Methanol offers an attractive alternative because it is liquid under ambient conditions, contains 12.5 wt% hydrogen and can be transported using existing fuel infrastructure. Conventional gas-phase methanol steam reforming generally operates above 300 °C and requires additional equipment to vaporize methanol and water.
Aqueous-phase reforming of methanol can instead generate hydrogen directly from a liquid methanol-water mixture at 150–250 °C, with limited formation of carbon monoxide and methane. Although oxide-supported platinum catalysts can accelerate this reaction, the respective effects of surface hydroxyls, lattice oxygen and relatively inert oxygen species have remained unclear.
A study (DOI: 10.48130/een-0026-0013) published in Energy & Environment Nexus on 07 July 2026 by Hui Zhou's team, Tsinghua University, shows that hydroxyl-rich amphoteric oxides promote hydrogen production, whereas reactive lattice oxygen and inert oxygen species restrict different stages of the reforming process.
The researchers prepared platinum catalysts supported on five oxides - Al2O3, ZrO2, CeO2, TiO2 and SiO2 - using wet impregnation, and evaluated them in an autoclave containing a methanol-water mixture with a molar ratio of 1:3. After reduction under hydrogen, each catalyst was tested from 190 to 250 °C. The team quantified hydrogen production and reforming selectivity by gas chromatography and examined catalyst structure, acidity, basicity, reducibility and metal-support interactions using microscopy, X-ray diffraction, X-ray photoelectron spectroscopy, electron paramagnetic resonance and temperature-programmed techniques.
In-situ diffuse reflectance infrared Fourier transform spectroscopy was then used to follow methoxy, adsorbed carbon monoxide and formate intermediates during the reaction. Activity differed sharply among the supports: hydroxyl-containing amphoteric oxides performed best, reducible oxides showed intermediate activity and inert SiO2 performed worst. At 250 °C, Pt/Al2O3 produced hydrogen at 846.9 μmol gPt-1 s-1, compared with 329.8 μmol gPt-1 s-1 for Pt/CeO2 and 41.4 μmol gPt-1s-1 for Pt/SiO2. Pt/Al2O3 also contained the highest proportion of surface hydroxyls, at 48.7%. Spectroscopic measurements showed that methanol decomposed efficiently at platinum sites on Pt/Al2O3, while surface hydroxyls supplied accessible OH* species that converted adsorbed CO through the water-gas shift reaction.
These hydroxyls could be replenished continuously through water dissociation. On Pt/CeO2, reactive lattice oxygen also supported formate formation, but the formate bonded too strongly, remained after nitrogen purging and obstructed subsequent hydrogen production. On Pt/TiO2 and Pt/SiO2, weak metal-support interactions restricted methoxy conversion into CO*, suppressing methanol decomposition. Tests using alumina and ceria with different morphologies confirmed that oxygen-species chemistry, rather than support shape or platinum particle size alone, controlled the reaction pathway.
Overall, the study establishes surface oxygen species as a key descriptor of catalyst performance in aqueous-phase methanol reforming. The results demonstrate that neither maximum oxygen reactivity nor oxygen-vacancy abundance necessarily produces the most active catalyst. Instead, effective hydrogen generation requires a balanced surface that can activate methanol and water while allowing reaction intermediates to proceed without becoming trapped. By prioritizing accessible, renewable surface hydroxyls and appropriately tuned metal-support interactions, future catalysts could enable safer and more efficient on-demand hydrogen production from liquid methanol.