Dear Syed,

welcome,

In the regions with high gradient you have to make the grid sufficiently fine to calculate the performance parameters such as the potential, the carrier concentrations and the electric field gradients with suffice accuracy and density such that one can compute the currents correctly.

In the practice you start with relatively coarse grid and make it finer till the solutions are stabilized. In the sense the solution i differs only negligible deviation from the previous errors.

For validation you can solve numerically structures which has an analytical solution.

Best wishes

Dear Professor Zekry,

Thank you for your advise. The problem I am having is, as I mentioned above I am currently simulating with 2A0 grid thickness. As I go finer as in 1A0 grid thickness, the simulations can not converge. In that case can I consider 2A0 as my lower limit? The solutions are having issues with convergence well before they are becoming stabilized.

Thank you.

I do not know how one can solve the convergence problem. However, i think that if the dynamic range of the parameters is large the computer can over flow. One solution may be is to scale down these parameters. I mean one can normalize them by dividing them on reference values. One may also use compressed parameters.

Best wishes

Thank you Dr. Zekry