Thermal transport in 2D materials

Dnr:

SNIC 2016/1-423

Type:

SNAC Medium

Principal Investigator:

Andreas Isacsson

Affiliation:

Chalmers tekniska högskola

Start Date:

2016-10-11

End Date:

2017-10-01

Primary Classification:

10304: Den kondenserade materiens fysik

Secondary Classification:

21001: Nanoteknik

Webpage:

http://www.chalmers.se

Allocation

Abstract

This project aims at shedding light on thermal transport and phononics in 2D materials such as graphene and Transition Metal Dichalcogenides(TMD)s. We will perform Molecular Dynamic(MD) simulations to probe the functionality of phononic structures such as thermal rectifiers, thermal waveguides, reconfigurable phononic Bragg gratings, etc. Exploring new degrees of freedom like mechanical and electrical stimuli, structural modifications and chemical functionalization and their effects on the thermal transport in these materials open up new possibilities for practical energy conversion and energy harvesting devices. The computational study involves generation of geometrical devices in atomistic level, energy minimization, thermodynamical equilibration and finally the calculation of phononic properties in MD framework which is implemented in LAMMPS. The structures under study are comprised of several thousand atoms. Thus, simulation times, specially for minimization, equilibration and final time averaging of the thermodynamical quantities of interest, require long runs (in order of multi-nanoseconds) with 1-0.5 femtosecond time step. Importantly the post processing and data extraction mandates having large memory space available. As an example the size of a typical dump file for a graphene with 8000 atoms can reach to 0.5-1 Gbyte. In order to make a more precise prediction of the above mentioned device properties, realistic experimental features e.g. effect of substrates, non ideal interfacing etc. have to be taken into account. As an example the study of thermal resistance of a graphene sheet which is embedded on silicon dioxide or amorphous silicon nitride substrate, will drastically increase the computational burden of the MD simulation in terms of number of atoms, relaxation of substrate, relaxation of graphene on the substrate among others. This necessitates using large number of cores(processors) for small or moderate size devices. The requested amount of computational resource in this proposal is based upon the above mentioned issues for study of preliminary small size devices (5000 - 50000 atoms) of our project.