Displacement of inactive phases by the reactive regime in a lattice gas model for a dimer-monomer irreversible surface reaction

A lattice gas model for the catalytic reaction A+B₂{→}AB is used for the study of the adsorbate displacement on fully covered surfaces with B and A species (i.e., poisoned phases) by the reactive state. The geometry used mimics narrow channels of width L whose ends are in contact with B₂ and A sources. Also, the channels are in contact with a reservoir of A and B₂ species in the gas phase, PA and 1-PA being the adsorption probabilities of such species, respectively. The displacement of the B-poisoned phase is slower than the A-poisoned one; however, both displacements stop at certain critical probabilities, PAᶜ¹(L) and PAᶜ²(L), respectively, which depend on the channel width. Extrapolation to the asymptotic regime L{→}{∞} leads us to conclude that such critical probabilities can be correlated with the critical points P1A and P2A at which the model exhibits irreversible phase transitions between a stationary reactive regime and B- and A-poisoned states. During the displacement of A-poisoned phases, it is found that the interface width w²(t) of the reactive regime diverges for PA{}PAᶜ² but remains bounded for higher PA values. Just at the critical point one has w²(t){∝}t^B, with {B}{→}1 for L{→}{∞}. Above the critical threshold, the time and adsorption probability dependence of w² is well described using general dynamic scaling arguments.