Normal heat conductivity in two-dimensional scalar lattices

TL;DR

This study shows that heat conduction in certain 2D scalar lattices is convergent, correcting previous claims of divergence by using larger simulated systems and ribbon geometries to obtain more reliable results.

Contribution

The paper demonstrates that previous reports of divergent heat conduction in 2D lattices are artifacts, providing a new simulation approach with ribbons to accurately assess convergence.

Findings

Heat conduction in 2D lattices with on-site potential is convergent.

Simulating ribbons of small width yields reliable 2D heat conduction data.

Longer systems are needed to observe convergence in 2D compared to 1D.

Abstract

The paper revisits recent counterintuitive results on divergence of heat conduction coefficient in two-dimensional lattices. It was reported that in certain lattices with on-site potential, for which one-dimensional chain has convergent conductivity, for the 2D case it turns out to diverge. We demonstrate that this conclusion is an artifact caused by insufficient size of the simulated system. To overcome computational restrictions, a ribbon of relatively small width is simulated instead of the square specimen. It is further demonstrated that the heat conduction coefficient in the "long" direction of the ribbon ceases to depend on the width, as the latter achieves only 10 to 20 interparticle distances. So, one can consider the dynamics of much longer systems, than in the traditional setting, and still can gain a reliable information regarding the 2D lattice. It turns out that for all considered models, for which the conductivity is convergent in the 1D case, it is indeed convergent in the 2D case. In the same time, however, the length of the system, necessary to reveal the convergence in the 2D case, may be much bigger than in its 1D counterpart.