Re: Full band simulation

 NTDinh wrote:
> Hello,
>
> I am alone doing Full band Monte Carlo simulation but I have some
> problems, please give me some suggestions:
> 1. Which mechanism electrons use to transfer from one conduction band
> to another one?
> 2. How can we determine the index of conduction band that electrons
> belong to? (in first Brillouin zone)
> 3. Do I need to consider DOS together bandstructure.
> 4 All above related to question that how can I incorporate the band
> structure into the program code. It is also a question.
1. Probably primarily phonon scattering, perhaps some impurity 
scattering, maybe others depending upon which material you have in mind. 

But that's the fun of doing a Monte Carlo thesis; you get to work out 
all those Fermi Golden Rule expressions.

2. Decide which band each electron starts in and keep track of the index 
as it evolves in free-flight and through each scattering event.

3. The density of states, the bandstructure, and the electron dynamics 
are all essentially the same thing.  Have an explicit model of your band 
structure, keep track of the k of each particle (NOT the velocity!), 
find the velocity from (grad k) E, and do all statistical sums over k. 
Then you have automatically included the DOS.

4. Again, the first step is to design the data structure that will hold 
the bandstructure.  Then write functions that evaluate velocity, 
scattering rates, etc. using that data.  At this point you are well on 
you way to a object-oriented design.  Of course this can't be 
retrofitted into the legacy code, so you throw that away, and rewrite 
the Monte Carlo algorithm.  There isn't that much content to it, anyway.

- Bill Frensley