A MONTE-CARLO SIMULATION OF THE CO OXIDATION ON PROBABILISTIC FRACTALS

We study a model of the CO oxidation on percolation clusters which represent the surface of our system in the Monte Carlo simulation. We observe two phase transitions where the surface is completely covered (poisoned) by one species (in our system CO or O). These phase transitions are described by their order and the values of the mole fraction y(CO) of CO in the gas phase at y1 (O poisoning) and at y2 (CO poisoning). The interval (y1,y2) represents the reactive regime. The influence of the occupation probability p for generating a spanning cluster on the underlying square lattice, of the diffusion of CO and of the lattice size on the value and the character of the kinetic phase transitions is studied. Increasing p leads to a shift Of Y2 to larger values of y(CO) for all ratios of adsorption to diffusion events but the value of y1 is increased to a larger value of y(CO) only if diffusion is not allowed. In the case of diffusion the value of y1 is maximally independent of p. A change in the character of the phase transition at y2 from first order on regular lattices to second order is observed on percolation clusters (p greater-than-or-equal-to p(C) = 0.592 75). Here pc is the percolation threshold on the square lattice. The character becomes again first order if diffusion is allowed or if p exceeds 0.97. The observed effects may be understood as a result of the different ramification of the percolation clusters which strongly depends on p. The cluster structures are characterized by pair correlation functions. The lattice size shows for p near unity no significant effect on the values of y1 and Y2 but its influence increases with decreasing p. We will show that the fractal dimension of the surface is not a useful quantity for predicting the positions and the character of the phase transitions for the reaction system studied here.