Discussion Forum

If these errors seem familiar please tell me how I can rid of them.

I am using Comsol Multiphysics 5.4a to see how electrical conductivity in a doped semiconductor varies with mechanical stress. My goal is to be able to plot curves of stress vs current in a doped p-type material. Since this is a uni-direction problem (I want the solver to solve for stress in the block, calculate conductivity and then show me how current distribution is).. the solver is segregated.

I started by going through both Piezoresistor pressure sensor examples in the library. My first approach was to build a 3D block of p-type piezoresistive Si, apply current source and a ground on two sides and apply a stress and fixed constraint on the sides perpendicular to the direction of current flow. For some reason, the simulation does not work with an applied current source (I specify the terminal face as a current source) though it does work when I define the terminal as a Voltage source. Do you have any thoughts on this? Why would the terminal type affect the ability of the simulation to run?

My second approach was to build a rectangular block of doped Si, apply force/stress and voltage through it, and define the electrical conductivity of the doped p-Si material as a function of stress {i.e. material conductivity= f(solid.x) }, instead of using piezoresistance. But I get a solution error whenever I try to use p-type Si material under piezoresistivity. When I change the material from p-type Si to just Si I do not get matrix errors anymore but the simulation does not converge.

I don't understand why I cannot use p-type Si as a material without getting the matrix error.

I am new with Comsol and this model is a little over the top for me so any help/guidance is much appreciated.