O(N) algorithms in tight-binding molecular-dynamics simulations of the electronic structure of carbon nanotubes

TL;DR

This paper presents O(N) algorithms and parallelization techniques for efficient tight-binding molecular-dynamics simulations of carbon nanotubes, improving accuracy and demonstrating electronic properties of different nanotube chiralities.

Contribution

It introduces buffer-based basis functions and analyzes buffer size effects, enhancing the accuracy and efficiency of O(N) tight-binding simulations for carbon nanotubes.

Findings

Buffer size significantly affects simulation accuracy and computational cost.

The method accurately distinguishes metallic and semiconducting nanotubes.

Parallelization improves simulation speed for large systems.

Abstract

The O(N) and parallelization techniques have been successfully applied in tight-binding molecular-dynamics simulations of single-walled carbon nanotubes (SWNTs) of various chiralities. The accuracy of the O(N) description is found to be enhanced by the use of basis functions of neighboring atoms (buffer). The importance of buffer size in evaluating the simulation time, total energy, and force values together with electronic temperature has been shown. Finally, through the local density of state results, the metallic and semiconducting behavior of (10x10) armchair and (17x0) zigzag SWNT s, respectively, has been demonstrated.