Discussion Forum

Hello everyone,

I am trying to simulate diffusion and reaction in a simple 2D square geometry with an interface inside that divides the geometry into two parts. Each subdomain has a different diffusivity. One subdomain has only diffusion and the other subdomain has both diffusion and reaction occurring in it.

Subdomain 1(diffusion only) contains the bulk fluid, which diffuses into Subdomain 2(diffusion-reaction) through the interface alone.

The boundary condition at the interface is a discontinuity of flux condition because I have reaction at the interface which is consuming the reactant.

The boundary condition is of the form:
(as shown in COMSOL): -n.(N1-N2)=N0

Ni=-Di*grad(c) where grad(c) is the gradient of the concentration. Therefore, I get:

n.(D1*grad(c)-D2*grad(c))=N0=kc

Since N0 represents consumption by reaction, it can be written in terms of the surface rate, which is kc (for example).

So I should be inputting kc in the N0 field in the boundary settings of the diffusion module in COMSOL. However, doing so does not lead to convergence and produces garbage. On the other hand, if I input N0= -kc, it gives me results that I expect to see. However, physically, the flux discontinuity equation does not make sense any longer, because a negative sign for N0 implies production of the component of interest.

After looking at the input fields in the boundary settings tab in COMSOL, I realized that the field for N0 asks for an "INWARD FLUX" to be input. I am wondering if this is the only reason why a negative sign is required for my reaction term- in that the reaction occurs on the surface, it does not have any direction as the fluxes do, but COMSOL is expecting to see an inward flux only, and hence the negative sign is required. Any clarification on this would be highly appreciated.

Thanks,

Sanjeev