Fermi-Dirac statistics

[dalonsoa]

Solcore's PDD solver, as it is very often the case, uses the Boltzmann approximation. While that is good enough for many applications - and usually faster - there are cases where a more rigorous treatment, using the Fermi-Dirac (FD) statistics, will be useful: highly doped materials, quantum wells, etc.

Ideally, using FD statistics should be implemented as a modification of the existing solver, rather than a separate solver.

[tcaduser]

FD integrals here https://github.com/scott-maddox/fdint

[dalonsoa]

Thank you very much for the information. It is certainly an excellent source of information and references, although it will need some tweaking as the PDD solver is written in Fortran with quadruple precision variables.

[tcaduser]

I'm confused what f2py actually does.

For my own code, I am using the Joyce-Dixon approximation in double precision.

https://github.com/devsim/devsim/blob/master/src/MathEval/Fermi.cc

[dalonsoa]

f2py is a Python tool that compiles a Fortran module and makes the functions and variables inside accessible from Python as with any other Python package. You have more information here: https://docs.scipy.org/doc/numpy/f2py/