[lammps-users] Query on two temperature model

LAMMPS LAMMPS Mailing List Mirror
1

Steve,

I also have another question. Is it possible to use ttm model to calculate electron contribution to thermal conductivity in metals using non-equilibrium MD?

Thanks
Ajit

2

This does sound possible. I am not aware of this having been done before.

Paul

3

Ajit, this is feasible with the AtC TTM package although we have not ever used it for the type of calculation you specify. Be aware that your result we be quite sensitive to the electron-phonon coupling model you use so follow Paul's suggestion and make sure you have a reliable approach from the literature. You will also need to take care how you set up the mesh used for computing the electron temperature field.
Jeremy

4

Jeremy,

Thanks for your help. However, i didnot understand what u meant by “AtC” TTM package.

I also want to know whether fix/NVT can be used for equilibrating the lattice before moving to NVE or only fix/langevin has to be used for that purpose.

Thanks
Ajit

5

The documentation is on the main web site here:

http://lammps.sandia.gov/doc/fix_atc.html

You will have to build and link the package yourself. As for your question, thermostatting in AtC is done with a Gaussian isokinetic thermostat, velocity rescaling thermostat, or it can run NVE. I have used NVT in other MD work and shifted to NVE, but this is just personal experience. Other users have tried other equilibration approaches.
Jeremy

6

Ajit,

The kappa_e (electronic thermal conductivity) will not be sensitive to
structure in this type of calculation.
It will be external parameter inserted into model.
As I understand, there is no method to calculate thermal conductivity
or similar electrical resistivity from first principles
unless your system in ballistic regime

German