A metal surface must hold the reactant strongly enough to catch it, but weakly enough to let it react. Tune the binding energy and watch a kinetic Monte Carlo lattice trace out the Sabatier volcano.
Pick a mission; its status updates live from the mean-field rates.
The science
Model. A molecule A (CO₂ in the picture) arrives on an empty site at rate ka = ra p, desorbs at kd = ν exp(−Eb/kBT) and reacts (H₂ is assumed to be in excess, products H₂O and CO leave at once) at kr = ν exp(−Ea/kBT), with ν = 10¹³ s⁻¹ and ra = 10⁶ s⁻¹ (model values, not measurements).
BEP relation. Ea = E0 + α Eb: a more strongly bound adsorbate is harder to convert. This is the Brønsted-Evans-Polanyi linear free-energy relation, here with illustrative constants.
Steady state. Mean field gives θ = ka/(ka + kd + kr) and turnover frequency TOF = kr θ. Setting d ln TOF/dEb = 0 gives exactly kd = ka α/(1 − α), i.e. Eb* = kBT ln[ν (1 − α)/(α ka)] - the top of the Sabatier volcano. Weak binding: bare surface, rate limited by arrival. Strong binding: surface blocked and barrier too high.
Kinetic Monte Carlo. The lattice (10 × 8 sites) uses the Gillespie (BKL) algorithm: the waiting time is −ln(u)/Rtot and the event is chosen in proportion to its rate. Sites do not interact, so KMC must agree with the mean-field result; the checks below confirm it.
Simplifications. One adsorbed species on a perfect surface; no adsorbate-adsorbate interactions, no H₂ coverage, no steps or defects, no reverse reaction, no entropy terms. The pictured intermediates (CO₂*, formate-like species) are illustrative. Not a calculation for any real metal [перевірити за джерелом before comparing with a real catalyst]. Educational virtual model. Models built with Blender, scene by Three.js.