Parametrization of a Reactive Force Field (ReaxFF) for Molecular

Jun 22, 2017 - The algorithm is based on a sequence of classical reactive molecular dynamics (RMD) at high temperature and Basin Hopping(39) (BH) sear...
4 downloads 14 Views 3MB Size
Article pubs.acs.org/JCTC

Parametrization of a Reactive Force Field (ReaxFF) for Molecular Dynamics Simulations of Si Nanoparticles Giovanni Barcaro,† Susanna Monti,‡ Luca Sementa,† and Vincenzo Carravetta*,† †

CNR-IPCF, Institute of Chemical and Physical Processes, via G. Moruzzi 1, I-56124 Pisa, Italy CNR-ICCOM, Institute of Chemistry of Organometallic Compounds, via G. Moruzzi 1, I-56124 Pisa, Italy



ABSTRACT: A novel computational approach, based on classical reactive molecular dynamics simulations (RMD) and quantum chemistry (QC) global energy optimizations, is proposed for modeling large Si nanoparticles. The force field parameters, which can describe bond breaking and formation, are derived by reproducing energetic and structural properties of a set of Si clusters increasing in size. These reference models are obtained through a new protocol based on a joint high temperature RMD/low temperature Basin Hopping QC search. The different procedures of estimating optimal force field parameters and their performance are discussed in detail.

1. INTRODUCTION Silicon-based materials are the most important semiconducting components of the microelectronics industry, which is presently focused on developing new miniaturization techniques based on the structural and electronic properties of nanometer aggregates. A careful control of the synthesis of these nanoparticles is fundamental to design high performance devices. This is provided by bottom-up gas phase synthesis with tuned controls for driving the nucleation and growth processes.1−3 Accurate experimental techniques capable of characterizing all the phases of the nanoparticle production at the nanometer scale are rare and hardly provide sounded data. Appropriate computational protocols at the atomic and mesoscale levels have the potential to reveal valuable information for interpreting the experimental data and for predicting and driving the production procedures. As a matter of fact, numerical simulations can identify important effects regarding the growth mechanism of Si nanoclusters and disclose the physical chemical properties of Si-based materials at the nanoscale. Careful examination of the literature reveals that most of the data about structure and properties of small Si clusters (