Oxidation of Silicon Carbide by O₂and H₂O: A ReaxFF Reactive Molecular Dynamics Study, Part I

Abstract

Simulations of the initial oxidation process of a SiC surface exposed to O2 and H2O molecules was studied with ReaxFF, an atomically detailed reactive molecular dynamics method that naturally models the breaking and forming of bonds. In this work, the ReaxFF forcefield was first expanded by training it with new quantum mechanics data of the binding energy, equation of state, and heat of formation of the SiC crystal, along with data from earlier studies that describes Si – Si, Si – O, and Si – H interactions. This expanded ReaxFF forcefield is capable of simultaneously describing both Si–C–O and Si–O–H bonding interactions. Using the forcefield, oxidation simulations were performed at various temperatures (in the range of 500 to 5000 K), and the trajectories were analyzed. The analyses showed that SiC gradually transforms into the oxides of silicon with simultaneous formation of a graphite-like layer. In presence of excess O2, the graphite-like layer was further oxidized to CO and CO2. We also analyzed the trajectories with two-atom and three-atom clusters to quantitatively track the oxidation progress. This analysis clearly showed Si–O and C–C bond formation at the expense of O–O and Si–C bond consumption indicating SiC oxidation with simultaneous formation for carbon-like structure. Consumption of SiC with O2 was found to be faster than that with H2O. We have also reported the oxidation simulation of SiC with a mixture of H2O and O2. Oxidation proceeded effectively as a two-part sequence, with O2 first oxidizing the SiC, followed then by H2O.

Electric-field-induced structural changes in water confined between two graphene layers

“…The connectivity in the entire system is recalculated in every iteration and non-bonded interactions (e.g., van der Waals) are calculated between all atom pairs [17][18][19] . Numerical simulations are carried out using the LAMMPS code 20 which includes the ReaxFF method 21 .…”

Atomistic Modelling of Si Nanoparticles Synthesis

“…(2) and (3), respectively, while σ refers to their root-mean-square dispersion. DFT: density functional theory; ND: present optimization of ReaxFF; VD: parametrization for Si-based materials available in the literature [7].…”

“…The atomistic model adopted to simulate the early stages of the gas phase condensation of Si nanoparticles consists of classical molecular dynamics simulations based on a reactive force field (ReaxFF) originally developed by van Duin [5], where the parameters were recently optimized [7] for Si-containing systems. ReaxFF is flexible, computationally efficient, and has been successfully applied to a large variety of material modelling problems [8][9][10][11][12][13][14][15].…”

“…Starting from a parametrization of the FF for Si based materials (VD) available in the literature [7] and designed to describe the bulk structure of SiC and SiO 2 , we re-optimized the FF for the specific systems under study in order to consider surface effects of small Si clusters. These effects can be fundamental in the nucleation and growth processes, leading to nanoparticles with maximum diameter of about 1 nm.…”

“…The t-set adopted was given by geometries and energies of a number of model systems selected among the structures optimized by the DFT-BH searches, namely: Si 2 , Si 3 , Si 4 , Si 5 , Si 6 , Si 8 , Si 10 , Si 12 (two structures), see Figure 2, to which the basic unit of silicon diamond cubic crystal was added. Starting from the original parametrization (VD) [7], we optimized only the relevant parameters responsible for the flexibility of the Si clusters at high temperature against our t-set: (i) Si-Si bond (14 parameters); (ii) Si-Si-Si valence angle (5 parameters); (iii) Si-Si-Si-Si torsion angle (5 parameters). These latest parameters were not present in the original parametrization, but were added in this new force field to get a better agreement with the t-set data.…”