Catalysts come in two kinds. Homogeneous catalysts are dissolved molecules that float among the reactants, meeting them in solution. Heterogeneous catalysts are solid surfaces that reactants adsorb onto, react, and leave. The distinction is fundamental to how catalysis is taught, researched, and industrially designed. You pick one or the other, then optimize within the category.
Vinyl acetate is produced at roughly 7 million metric tons per year, mostly using palladium-based catalysts on solid supports. The standard industrial account is heterogeneous catalysis: ethylene and acetic acid meet on a palladium surface, react in the presence of oxygen, and vinyl acetate desorbs. The surface does the work.
Harraz, Lodaya, Tang, and Surendranath at MIT (Science, 2025) found that this account is half right — but the other half is not a refinement. The palladium catalyst cycles between two forms. The solid metal activates oxygen. The dissolved palladium ions activate ethylene and acetic acid. The interconversion between them is driven by electrochemical corrosion: the solid palladium corrodes into soluble species, driven by reaction with oxygen that converts to water. The molecular form does the organic chemistry; the surface does the oxygen chemistry; and the transition between them is corrosion.
The corrosion rate is the rate-limiting step of the entire process. The speed of vinyl acetate production is controlled not by either catalytic step but by the speed at which the catalyst dissolves and reforms.
For decades, the field treated this system as heterogeneous catalysis with homogeneous contamination, or vice versa. The dissolved palladium species were regarded as leached catalyst — degradation, not mechanism. The corrosion was the thing to be minimized. The finding inverts this: the corrosion IS the mechanism. The system requires its own degradation to function, and the rate of degradation governs the rate of production.
The structural insight is about categorical boundaries. Homogeneous and heterogeneous catalysis are treated as distinct types — different chapters, different conferences, different optimization strategies. This system is neither. Or rather, it is both, sequentially. The catalyst is the interconversion — the cycling between dissolved and solid forms. The boundary between categories is the mechanism.
This reframes what counts as catalyst failure. If corrosion is the productive step, then a catalyst that doesn't corrode doesn't catalyze. Durability, normally the primary engineering goal, would kill the reaction. The design criterion flips: you don't want a catalyst that resists dissolution. You want one that dissolves at the right rate.
The question of whether vinyl acetate synthesis is homogeneous or heterogeneous was never a confusion to be resolved. It was a fact about the mechanism that the categories couldn't express.