Abstract

We consider the problem of determining analytically the effective diffusion constant Deff for diffusion in an inhomogeneous medium, described by the diffusion equation deltatP = deltax[D(x) delta xP] with a position-dependent diffusion coefficient D(x). As there is no translational invariance, two different natural definitions are possible: From the long-time behaviour of the variance, Deff = D = lim((Delta x)2)/2t. From the large-distance behaviour of the mean first passage time T(+or-x mod 0) to reach exit points at +or-x starting from the origin, Deff=D=limx to infinity x2/2T(+or-x mod 0). In general, D not=D/. We find an exact formula for D and examine a number of interesting special cases. If D(x) tends to finite limits D+or- as x to +or- infinity , then D is simply the arithmetic mean (D++D-)/2. In the important case of a periodic D(x), we find that D is the harmonic mean of D(x) in a period. We also give an argument suggesting that D=D in this case. If D(x) is piecewise constant in an arbitrary fashion, with a finite number of discontinuities, and tends to D+or- and x to +or- infinity , then D=(D++D-)/2 as before, but D=(D+D-)1/2+(1-2/pi)(D+1/2-D-1/2)2<=D (the equality obtaining only if D+=D-). Thus D is the geometric mean of D+ and D- plus a 'correction' term. We also illustrate the significant role of inhomogeneities in determining Deff with the help of a simple example involving a discrete-time random walk on a chain.

Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call

Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.