Diff14.pptVIP

Module 14: Correlation Effects in Diffusion DIFFUSION IN SOLIDS Professor Martin Eden Glicksman Professor Afina Lupulescu Rensselaer Polytechnic Institute Troy, NY, 12180 USA Outline Markov sequences and correlated walks Jumping probability estimate Estimates of the correlation factor: pure materials Computation of correlation factors in pure crystals Exercises Markov Sequences and Correlated Walks Jumping Probability Estimate Jumping Probability Estimate Jumping Probability Estimate Jumping Probability Estimate Jumping Probability Estimate Jumping Probability Estimate Correlation for Self-diffusion in Pure Crystals Estimates of f (Compaan Haven) Exercises Exercises Key Points Random walk theory is based on Markov processes: independent steps that lack “memory.” Memory effects reduce the efficiency of random walks. Such effects are termed correlation. Generally, three diffusing “elements” are required for correlation effects to be significant. The correlation factor represents the efficiency of non-Markovian random walks associated with diffusion. For defect-assisted self-diffusion, the correlation factor depends on the crystal structure. Interstitial diffusion in dilute systems is not correlated. Correlation factors for self-diffusion in pure materials are well known, and typically may be approximated as f=1-2/Z. ? meg/aol ‘02 Reduction in mean-square displacement for non-Markov random walks may be introduced through an exchange efficiency or correlation factor. For self–diffusion in a pure crystal The square displacement of an n-step random walk can be described as Substituting and comparing the result Correlated exchanges between “tagged” tracer atom and a lattice vacancy on an adjacent crystal plane The effective frequency, ?R, of a tracer atom jumping to the right is The alternating series of conditional exchange probabilities can be gathered as two summations Jumping Probability Estimate The correlation factor, f, may be reintroduced where Note: