Answers are grounded in this paper and cite their sources.
Predicting reaction selectivity on corrugated catalytic interfaces remains a major challenge when observed pathway bifurcations cannot be rationalized by static, single-site activation barriers alone. Here, by coupling neural-network-accelerated gas–solid nanoreactor molecular dynamics (GS-NMD) with first-principles calculations and microkinetic modeling, we elucidate the atomistic origin of the temperature-dependent ethylene polymerization switch on corrugated Fe3C(102). We demonstrate that product divergence is dictated by the spatiotemporal clearance of dissociated hydrogen adatoms rather than intrinsic C–C coupling barriers. The corrugated carbide surface functions as a spatially coupled nanoscale tandem architecture, wherein exposed top-layer Fe ridges drive primary C–C and C–H transformations, while adjacent subsurface valleys serve as kinetically accessible hydrogen reservoirs. This dynamic framework quantitatively captures key scanning tunneling microscopy observables, including the shift from boundary-confined direct insertion to terrace-wide dehydrogenative oligomerization, simulated apparent intermediate heights, and distinct chain morphologies. Mechanistically, sluggish low-temperature H migration causes local ridge autopoisoning, suppressing C–C coupling and enforcing direct coupling of intact monomers. At elevated temperatures, dynamic adsorbate fluctuations open a facile drainage pathway into the valley sink, continuously evacuating ridge sites to sustain dehydrogenation. These results highlight the indispensable role of multisite byproduct transport across surface topography, offering a generalizable strategy to model and design complex heterogeneous catalysts beyond static single-ensemble descriptors.
No takes yet. Share an insight, caveat, or question.
Song et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: