Discussion Forum

Hi,

I want to study the mass transport of diluted species under convection and diffusion. My simulations yielded a significant amount of negative concentration. Therefore I created a model for testing purposes:

1. Only diffusion
2. Geometry (see geometry.png) -> a defined flux is applied at the surface. This flux was originally defined by a square pulse (for testing purposes the flux is constant)
3. Time-dependent study

The results show that:

• There is a zone of negative concentration above the surface where the flux is set
• The concentration on the surface where the flux is applied has not a uniform concentration.

What did I try to resolve this:

• Change the mesh. I used the standard mesh from comsol and two custom quad meshes different accuracy. Concentration profiles from the quad-meshes are attached. It seems that the mesh quality only compresses/shifts the negative concentration zone.
• Ramping up the pulse and removing the pulse (constant flux)
• change time-steps (smaller steps)
• change the stabilization methods
• change the magnitude of the flux

Nothing of the above methods did help. I have also attached my test model.

TLDR: Transient study using diffusion of a concentration applied via a flux results in significant regions of negative concentration as well as concentration "hot-spots".

Thanks Claus