TRACER DIFFUSION IN A RANDOM BARRIER MODEL - THE CROSSOVER FROM STATIC TO DYNAMIC DISORDER

In earlier investigations, we have shown that in a frozen-in random ba rrier environment the diffusive behavior of a thermally activated trac er particle shows a crossover from anomalous to normal diffusion, gove rned by the percolation threshold of the underlying lattice and the de gree of randomness of these barriers. Changes due to a periodic renewa l of the environment were not considered. In the present work, we use an analysis within the framework of the effective medium approximation , and Monte Carlo simulations, to study the crossover from a ''frozen in'' static to dynamically updated random barrier disorder with changi ng temperature T, and find a temperature transition to a qualitatively different type of diffusive behavior of the tracer particle. It turns out that the Arrhenius relationship of diffusion coefficient D on T i s replaced by a linear one at a crossover temperature T-c, which itsel f depends on the frequency of environmental renewal omega with a power law: T(c)(proportional to)omega(delta), with delta=0.21+/-0.02. In th e linear regime below T-c, where we find that the tracer movement is h ighly correlated, the average effective activation energy for diffusio n [E-a] is equal to the thermal energy of the tracer, [E-a]approximate to kT, while for T>T-c (Arrhenian regime) the random walks are practi cally uncorrelated and [E-a] is constant, given by the mean value of t he barrier heights probability distribution, (E) over bar. These resul ts are found to be independent of the particular type of probability d istribution which is used for the barrier heights. [S1063-651X(98)0731 0-3].