Anomaly in the behavior of silicon from free energy analysis: a computational study

Abstract

Silicon shows a very different trend while melting. Melting has remained a challenging subject from a long time. Especially, predicting the melting temperature of any solid substance still exists as a problem in many cases. Recently, various studies and new rules and set of parameters have simplified things, but its mechanism is yet to be studied properly and there still does not exist any generalized concept regarding this. Also, there are certain anomalies in silicon, which makes it’s melting and phase transition mechanism more difficult to understand and predict. It is essential to study about its behavior, which also gives us an insight into melting behaviors of nanowires, clusters, defect and surfaces. This project is an attempt to study the mechanism of melting using classical molecular dynamics simulation, to define a set of parameters that could help us predict the behavior of silicon at any temperature and also its phase transition mechanism. In order to understand the phenomenon, it is important to know the interaction potential governing the silicon system. Stillinger-Weber potential is a good model for Si atoms which takes into account two and three particle interactions. Melting of Silicon atoms is studied using Molecular Dynamics Simulation with the help of LAMMPS software. Heating and quenching processes is implemented on a system of Si atoms. Variations of various parameters like density, volume per atom, potential energy has been studied. Free energy gap connecting phases is estimated with the help reversible thermodynamic route. Supercritical path is constructed with the help of more than one reversible thermodynamic path. The best of my knowledge, this is first attempt to implement pseudo-supercritical reversible thermodynamic path for a system whose solid volume is higher than liquid volume at phase transition point. It has been found that melting in Si occurs in three stages. It involves pre-melting, melting and relaxation. The potential energy and density appear to be steady, indicating pre-melting. Melting temperature can be predicted using Gibbs free energy. Gibbs free energy calculation involve with thermodynamics integration and multiple histogram reweighting(MHR) method. The density and potential energy drops abruptly, for temperature more than 1750 K indicating complete loss of crystallinity.