fig1

Figure 1. Coupled GCMC–EKMC workflow for simulating potential-dependent *CO/*H co-adsorption and Cu(100) surface reconstruction under eCO₂R conditions. Top left: Initial ordered Cu(100) surface (~35 nm lateral dimension)^[30]. Top right: The GCMC loop samples the occupancy of each surface site (empty, *CO-occupied, or *H-occupied) under fixed electrochemical parameters (U, pH, T, pCO) via Metropolis-criterion acceptance until convergence: a proposed move is accepted with probability min{1, exp(-ΔG_tot/k_BT)}, where ΔG_tot = G_totⁱ⁺¹ - G_totⁱ, so that moves lowering the total free energy are always accepted, and those raising it are accepted with Boltzmann probability^[30]. Bottom left: Final reconstructed Cu(100) surface (~35 nm), in which adsorbate-driven Cu atom migration produces a disordered morphology qualitatively distinct from the initial slab^[30]. Bottom right: The converged GCMC coverage is passed to an EKMC loop, which propagates the surface in time by distinguishing fast (CO and H diffusion) from slow (Cu migration) events^[30]. All panels reproduced with permission from Ref.^[30]. Copyright © 2026 American Chemical Society. GCMC–EKMC: Grand canonical Monte Carl and environmental kinetic Monte Carlo; eCO₂R: electrochemical CO₂ reduction.