[hda]

I have a microchannel partitioned in two sections with a perforated membrane of a finite thickness. A dilute mixture of oxygen, glucose and water is given from inlet.

The mixture can flow through the perforation as it is. In addition, there is molecular diffusion (but no advection) through the membrane solid. I want to make the membrane allow diffusion of water and oxygen but not of glucose.

How can we model this?

Here is what I tried but it did not work:
1) Set the membrane and both of the sections to be fluid zones of "mixture" material.
2) Set the velocity components in the membrane to zero using "fixed values" tab. This will stop advection.
3) Using a UDF, set the mass diffusion coefficient of glucose inside the membrane zone to a very small number, say 1e-20 (setting it to zero gives a warning). However, set the coefficients of oxygen and water to their appropriate values. This will stop diffusion of glucose while allowing oxygen and water to diffuse.

Outside the membrane zone all species have appropriate values for their mass diffusion coefficients (order of 1e-10).

Unfortunately, when I run this, I see glucose diffuse right into the membrane. Setting its mass diffusion coefficient exactly to zero (despite the warning) won't help.

It seems like UDF works fine because I see in the contour plot whatever value I assign for diffusion coefficient of glucose in the membrane. Fixed values for velocities also work, since the contour plots show zero velocity inside the membrane.

I thought that this trick would stop the diffusion and advection mechanisms for glucose in the membrane but it did not work.

Any ideas why did this fail?