Local structure and defect segregation on the tilt grain boundaries in silicon

TL;DR

This study uses atomistic and ab initio simulations to analyze the structure and defect segregation behavior of various tilt grain boundaries in silicon, revealing geometric correlations and proposing a simple segregation energy model.

Contribution

The paper introduces new asymmetric tilt grain boundary structures in silicon and develops a geometric model to estimate defect segregation energies.

Findings

All boundary structures maintained tetrahedral coordination after optimization.

Segregation energy correlates with local boundary geometry.

Proposed a simple geometric model for defect segregation energy estimation.

Abstract

We present the results of atomistic and ab initio simulation of several different tilt grain boundaries (GB) in silicon. The boundary structures obtained with genetic algorithm turned out to have no coordination defects, i.e. all silicon atoms restored their tetrahedral coordination during the structure optimisation. That concerns previously known symmetric {\Sigma} 5 (130), {\Sigma} 3 (211) and {\Sigma} 29 (520) boundaries as well as previously unknown asymmetric {\Sigma} 9 (-255)/(-211), {\Sigma} 3 (-255)/(211) and {\Sigma} 13 (790)/(3 11 0) structures. We have performed an extensive study of defect segregation on the boundaries, including neutral vacancy and carbon, phosphorus and boron impurities. A clear correlation between the segregation energy of the defect and local geometry of the boundary site where the defect is segregated has been revealed. We suggest a simple purely geometric model for evaluation of approximate segregation energies of the listed defects.