Dear Lammps Dev,
After spending some times in this research field, I have a doubt regarding DFT and MD
1. For examples, I have a structure of a crystal at 300K, if I minimize it at DFT, is it the structure still at the 300K configuration? (What I understand is minimization is at 0K).
2. Lets say I have a crystal and 300K and its potential, is it possible I heat it to a finite tempearature, e.g. 1000K, then I put it in DFT to find is electronics properties at 1000K?
Regards
min
|
Dear Lammps Dev,
After spending some times in this research field, I have a doubt regarding DFT and MD
DFT (i assume you are referring to density functional theory) and MD (= molecular dynamics)
describe two very different things. you can do MD with DFT (to derive the forces)
1. For examples, I have a structure of a crystal at 300K, if I minimize it at DFT, is it the structure still at the 300K configuration? (What I understand is minimization is at 0K).
no. however, only a complete minimization to the global minimum will
result in a true 0K structure. this is not easy. starting from a thermalized
configuration, a minimization usually leads into a local minimum. only
system with a very small number of degrees of freedom can always go
to the global minimum of potential energy.
2. Lets say I have a crystal and 300K and its potential, is it possible I heat it to a finite tempearature, e.g. 1000K, then I put it in DFT to find is electronics properties at 1000K?
yes and no. it will be an approximation. since forces in a classical model
are rarely identical with those from a DFT calculation. also, depending
on your material, you may have to consider thermal excitation of the
electrons in the process. a DFT based single point calculation is effectively
a minimization in a similar vain, as you were asking about in question 1
and thus compute the electronic structure of the ground state. if your
system is metallic or a semiconductor, that may not be sufficient. but
that is a question for a different mailing list.
you should also keep in mind, that at finite temperature
you don’t have a single conformation that is representative
but rather an ensemble of configurations.
you probably should re-read some chapters in your text books
on both electronic structure calculations and statistical mechanics.
cheers,
axel.