Low thermal conductivity for single layer hBN

LAMMPS LAMMPS Beginners
1

Hi all,

I am trying to calculate the thermal conductivity of a single sheet of hexagonal boron nitride but the thermal conductivity values that I get are 10-15x lower than expected. However, when I run the same setup (only a .data file change) for single layer graphene, the thermal conductivity values are within expectations

I relax the system in npt then nvt, then run it again in nve before implementing the GK algorithm. The barostat and thermostat fixes are unfixed between and after the npt and nvt runs.

Any ideas why I’m facing issues? Thanks in advance

log.lammps (41.7 KB)
SLBN50x50.data (41.5 KB)
SLBN-GK.in (4.0 KB)

2

Have you looked up the paper describing the Tersoff potential you are using?
Does it say anything about the kind of systems it has been parameterized for?
If yes, do they include single sheets of hBN?

3

Have you looked up the paper describing the Tersoff potential you are using?
Does it say anything about the kind of systems it has been parameterized for?
If yes, do they include single sheets of hBN?

Thanks for the reply. Yes I have

The resulting κ of the model “graphene” system is ∼20% larger than that of BNWG.
… a Tersoff type empirical interatomic potential which is reparameterized to represent experimental structure and phonon dispersion of 2D hexagonal BN…

This is an excerpt from the source paper - Sevik, C., Kinaci, A., Haskins, J.B. and Çağın, T., 2011. Characterization of thermal transport in low-dimensional boron nitride nanostructures. Physical Review B , 84 (8), p.085409.

And also

We also note that in previous works the thermal conductivity of pristine single layer h-BN calculated via EMD at 300 K was predicted to be 400 W/m-K using the Tersoff potential parameterised by Sevik et al

This is an excerpt from the paper I am trying to replicate - Mortazavi, B., Pereira, L.F.C., Jiang, J.W. and Rabczuk, T., 2015. Modelling heat conduction in polycrystalline hexagonal boron-nitride films. Scientific reports , 5 (1), p.13228

4

Then my only remaining suggestion is to study the paper(s) with the reference results and see if you can find significant differences in the description of the simulation setups and settings to yours. Failing that, I would try to contact the authors and ask for more details and potentially the exact input data.