Precision in evaluating space charge density (semi.rho)

Please login with a confirmed email address before reporting spam

I am curious: The space charge density is calculated from the difference in doping and carrier densitites which are of the order of 1E24/m^3 with a resulting space charge density at least in some areas being 1E4/m^3 or similar. How is this implemeted in the COMSOL semiconductor module (as far as I know the general calculations are double precision?).

Can this issue affect the convergence for high doping densities or high injection of both electrons and holes into a region (e.g. optoelectronic heterostructure).

Thanks, Kresten


4 Replies Last Post 18/04/2021, 18:25 GMT-4
Robert Koslover Antennas, Waveguides, Electromagnetics

Please login with a confirmed email address before reporting spam

Posted: 3 weeks ago 14/04/2021, 11:05 GMT-4

Are you sure that the net space charge density is actually being computed in the way you described? (I, for one, would not attempt to code it that way.) See https://www.comsol.com/semiconductor-module for more info.

Are you *sure* that the *net* space charge density is actually being computed in the way you described? (I, for one, would not attempt to code it that way.) See https://www.comsol.com/semiconductor-module for more info.

Please login with a confirmed email address before reporting spam

Posted: 3 weeks ago 14/04/2021, 14:06 GMT-4

That is what they write in the documentation and how it is evaluated in the model so when plottet it gives a limited "dynamic range".

However, looking at the equation view in more detail it looks like the space charge density itself is actually not used for the further calculations so it is just computed for plotting. In that case things make sense.

That is what they write in the documentation and how it is evaluated in the model so when plottet it gives a limited "dynamic range". However, looking at the equation view in more detail it looks like the space charge density itself is actually not used for the further calculations so it is just computed for plotting. In that case things make sense.

Dave Greve Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 3 weeks ago 14/04/2021, 22:31 GMT-4

Often (usually?) the semiconductor equations are formulated in terms of the quasi-Fermi levels, not the charge densities.

I don't have the semiconductor module, but you can see what variables Comsol solves for by looking under Solver Configurations > Solution > Dependent Variables. All other quantities will be computed from these variables.

Often (usually?) the semiconductor equations are formulated in terms of the quasi-Fermi levels, not the charge densities. I don't have the semiconductor module, but you can see what variables Comsol solves for by looking under Solver Configurations > Solution > Dependent Variables. All other quantities will be computed from these variables.

Please login with a confirmed email address before reporting spam

Posted: 3 weeks ago 18/04/2021, 18:25 GMT-4

I am using the quasifermilevel formulation but that doesn't change the problem of evaluating Poissons equation.

I am using the quasifermilevel formulation but that doesn't change the problem of evaluating Poissons equation.

Reply

Please read the discussion forum rules before posting.

Please log in to post a reply.

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.