Published August 8, 2016 | Version v1
Journal article

On the channel width-dependence of the thermal conductivity in ultra-narrow graphene nanoribbons

  • 1. Department of Electrical Engineering, University of Kashan, Kashan 87317-53153 (Iran, Islamic Republic of)
  • 2. School of Engineering, University of Warwick, Coventry CV4 7AL (United Kingdom)

Description

The thermal conductivity of low-dimensional materials and graphene nanoribbons, in particular, is limited by the strength of line-edge-roughness scattering. One way to characterize the roughness strength is the dependency of the thermal conductivity on the channel's width in the form Wβ. Although in the case of electronic transport, this dependency is very well studied, resulting in W6 for nanowires and quantum wells and W4 for nanoribbons, in the case of phonon transport it is not yet clear what this dependence is. In this work, using lattice dynamics and Non-Equilibrium Green's Function simulations, we examine the width dependence of the thermal conductivity of ultra-narrow graphene nanoribbons under the influence of line edge-roughness. We show that the exponent β is in fact not a single well-defined number, but it is different for different parts of the phonon spectrum depending on whether phonon transport is ballistic, diffusive, or localized. The exponent β takes values β < 1 for semi-ballistic phonon transport, values β ≫ 1 for sub-diffusive or localized phonons, and β = 1 only in the case where the transport is diffusive. The overall Wβ dependence of the thermal conductivity is determined by the width-dependence of the dominant phonon modes (usually the acoustic ones). We show that due to the long phonon mean-free-paths, the width-dependence of thermal conductivity becomes a channel length dependent property, because the channel length determines whether transport is ballistic, diffusive, or localized.

Additional details

Identifiers

Publishing Information

Journal Title
Applied Physics Letters
Journal Volume
109
Journal Issue
6
Journal Page Range
vp.
ISSN
0003-6951
CODEN
APPLAB

Optional Information

Notes
(c) 2016 Author(s)