Prof. Jakubi and Prof. Israel, thank you for your responses.  Yes, I am interested in exploring a different diffusion coefficient in the film as compared to the bulk after understanding the simpler case of the same diffusion coefficient.

Prof. Israel, thank you also for your clarifications of the problem and excellent suggestions.  I adopted them immediately, but have follow up questions

For the argument of the exponent, is d in the denominator?  Also, would the pre-exponential term be M/sqrt(Pi d t)?  M = int(C(x),x = -infinity .. infinity).   This references Crank  2nd ed. p 13

I'm not sure that the impermeable boundary is captured by the solution, as I graphed the function with respect to x and t, for a given h and d.  At t=0 the box profile with width of 2h looks fine.  At short t, (ie 1 second), the concentration at x = -h decreases - with an impermeable boundary, I think it should remain constant due to reflection and superposition until the source becomes depleted at extended times.

Thank you very much for your thoughts and help.  This is a very interesting problem.

S Carroll, student