[lammps-users] Si thermal conductivity

Dear LAMMPS users,

I need to get the thermal conductivity of hydrogenated nanoporous-Silicon. I have few questions about:

  1. I am starting with Silicon bulk. The input files contains: 4x4x4 diamond structure, Tersoff potential, Green-Kubo approach, correlation length = 200 ps, simulation time = 10 ns. Time step = 0.5 fs.
    Many published works achieved a value as large as 270 W/ (m K) whereas I got only about 30.
    Does anyone get the correct value (even if it overestimates the experimental value) by using the built-in ave/correlate function?

  2. Does anyone build a H.tersoff file which properly describes the SiH interaction (maybe from the paper F. de Brito Mota,J. Appl. Phys., Vol. 86, No. 4, 15 August 1999) ?

Thank you!


Did you try any of the other methods discussed in section 4.20
of the documentation?


Hi Giuseppe,

What temperature are you talking about?
Second question: Did you try SW potential?
I also had small values with Tersoff potentials.
Attached please find thermal conductivity as a function of temperature
calculated for Si 6x6x6 sample.


total_k_t_Si.ps (16.5 KB)

Hi German,

T = 300K. I noted that you got 270 W/(K m) with Tersoff potential. Did you use the LAMMPS built-in function ave/correlate and trap? I haven’t tried SW, yet.


Hi Giuseppe,
The results I sent were obtained with Stillinger-Weber (SW) and MEAM potentials.
For Tersoff I obtained much smaller values.
I didn't use built-in function ave/correlate and trap, used heat flux
values and then calculated autocorrelations outside of the code.