
The Arrhenius form is well known and commonly used to describe various rate processes such as chemical reactions and diffusion. Its simplicity is intriguing but also somewhat deceptive, since the Arrhenius parameters extracted from the analysis are effective ones, thus not revealing the true microscopic nature of the problem. One demonstrative example is the interpretation of the effective surface diffusion barrier E_{A}, which is commonly measured in experiments and interpreted as the adiabatic potential energy barrier that the particle has to cross to move from one adsorption site to the neighboring one. This interpretation is correct for singleparticle motion under certain conditions, but is without any grounds for manyparticle diffusion where the interaction effects are more pronounced. Nevertheless, the Arrhenius form is often used to describe experimental diffusion data at finite coverages.
The aim of this work [1] is to study the origin of effective surface diffusion barriers E_{A} through Monte Carlo simulations within the latticegas model. In particular, we focus on the importance of memory effects, which originate from the idea that consecutive displacements of a tagged diffusing particle are correlated. It is found that E_{A} of tracer diffusion is strongly affected by memory effects at finite coverages and low temperatures, where about 10%  50% of E_{A} comes from the memory contribution which arises from temperature variations in the memory effects and is not thermally activated. In the barrier of collective diffusion, the memory contribution is also significant but less pronounced than in tracer diffusion. These results demonstrate that in general any effective barrier E_{A} for some quantity describes its rate of change vs. temperature. Only in special cases it clearly corresponds to some particular thermally activated processes.
A more comprehensive report of the memory effects in general will be published elsewhere [2]. Furthermore, related studies of memory effects in various strongly interacting adsorption systems are found in Ref. [3].
[1] I. Vattulainen, Surf. Sci. 412413, L911 (1998).
[2] I. Vattulainen, S. C. Ying, T. AlaNissila, and J. Merikoski, submitted to Physical Review B; Helsinki Institute of Physics preprint HIP199852/Th (1998).
[3] S. C. Ying, I. Vattulainen, J. Merikoski, T. Hjelt, and T. AlaNissila, Phys. Rev. B 58, 2170 (1998).