Simulation of Atomic-Level Defects in Ceramics
W. J. Weber and R. Devanathan, Pacific Northwest National Laboratory
Research Objectives
The aim of the present study is to improve our understanding of the production and evolution of atomic-level defects in ion-implanted ceramics and advanced semiconductors, such as silicon carbide (SiC). The simulations are needed because the small time and distance scales of these processes preclude direct experimental observation.
The results of our work will be used as inputs for kinetic Monte Carlo simulations and analytical models to extend our knowledge of defect processes to real-world time and distance scales.
Computational Approach
Our molecular dynamics code is written in FORTRAN 90 and uses the Parallel Virtual Machine (PVM) message-passing library. This enables the simultaneous calculation of forces acting on several million atoms, using the combined speed and memory of as many as 512 processing elements (PEs). The interactions between atoms are described using a combination of Tersoff and first-principles repulsive potentials. The computer used is the NERSC Cray T3E.
Accomplishments
We have calculated the energy needed to knock atoms in SiC out of their lattice sites along various crystallographic directions. This value, known as the displacement threshold energy, is a fundamental parameter in radiation damage models and is needed to verify the reliability of our interatomic potentials. Our results show that the threshold energy is 20 and 35 eV for C and Si, respectively, in excellent agreement with experiments and first-principles calculations.
Currently we are simulating several high-energy (10-30 keV) Si recoils along three crystallographic directions in SiC to understand the statistics of the damage cascade. The simulation cell contains 2 million atoms, and each calculation has to be run for more than 20,000 time steps.
Our calculations show that anti-site defects are produced by energetic recoils and may play an important role in the amorphization of SiC. In addition, the mixing in the cascade is much less in SiC than in Si, which helps explain the different cascade morphologies of these two materials. The figure shows the distribution of defects in a 10-keV cascade in SiC.
Vacancies (green), interstitials (red), replacements (blue), and anti-site defects (black) created by a 10 keV Si recoil in SiC. The box shown is about 130 x 90 x 90 Å. The smaller and larger circles represent C and Si defects, respectively.
Significance
Silicon carbide (SiC) is an advanced semiconductor that performs better than Si in radiation and high-temperature environments. It has potential applications in space stations, advanced fighter aircraft, proposed fusion reactors, and the petroleum and automotive industries. In addition, SiC-based composite materials are being considered for structural applications in fusion reactors. Both semiconductor doping by ion-implantation and long-term operation in a radiation environment result in the production of atomic-scale defects by the displacement mechanism. The evolution of these nonequilibrium defects determines the ultimate performance of the material. Our study provides much-needed information on the formation and migration of defects in SiC.