[lammps-users] Irradiation in materials

Dear lammps uesrs

I have some problem in calculating the thermal conductivity in irradiated SiC material, and I used the 101010 box with the Tersoff potential to see the differences between the original SiC material and the irradiated SiC material. The EMD method based on the Green-Kubo formular was used to simulate the thermal conductivity (computer heat/flux command). The original SiC is calculated around 10W/(mK), but with setting a velocity of 250m/s (units box) at the top of box, the caluation of thermal conductivity is hundreds than previous result and is not converged.
I see the results and find that the heat flux along the z axis Jz is much bigger than the Jx and Jy, is it something about the limitation of calculating thermal conductivity with many-body potential in LAMMPS or something that I’ve ignored ?

Best wishes

GK includes a kinetic energy contribution from compute ke/atom. If you mean
your system has some number of atoms with very high z velocities, then I suppose
that would bias the KE in that direction. It’s not clear to me what the “right answer”
is for heat flux when you have fast motion of atoms in a particular direction.

There are also issues with manybody potentials
and GK due to the way the virial is computed, which are well discussed in recent
literature. But I don’t think that would affect z differently than x,y.

I’m CCing Aidan in case he has additional ideas.

Steve