Self-diffusion constants in silicon: Ab initio calculations in combination with classical rate theory
Chi-Ok Hwang
|
Journal of Chemical Physics
125/22
(2006)
We demonstrate that local-density approximation in combination with the dynamical matrix method is a plausible method for calculating diffusion constants in solids. Especially we compute the diffusivity of the neutral self-interstitial in silicon bulk. The climbing image nudged elastic band method is used for the energy barrier and the transition state atomic configuration. The diffusion prefactor is obtained by using a classical rate theory, the dynamical matrix method. We compare with the diffusivity from another alternative way, ab initio molecular-dynamics simulations, at 1500K. They are in good agreement.
This work was supported by the Korea Research Foundation Grant funded by the Korean Government (MOEHRD) (KRF-2005？005-J01103). Also, this work was supported partly by the Korean Ministry of Information & Communication through Information Technology Research Center Program and Information Technology Professor Support Program supervised by Institute of Information Technology Assessment, and partly by the Korean Ministry of Science and Technology through Korea Institute of Science and Technology Evaluation and Planning. In addition, the author would like to acknowledge support from Korea Institute of Science and Technology Information under “The Seventh Strategic Supercomputing Applications Support Program” with Dr. Sang-Min Lee as the technical supporter. The use of the computing system of the Supercomputing Center is also greatly appreciated. Finally, the author gives special thanks to Professor Graeme Henkelman for his help in the dynamical matrix method through the forum: http://theory.cm.utexas.edu/forum.
- 초록
We demonstrate that local-density approximation in combination with the dynamical matrix method is a plausible method for calculating diffusion constants in solids. Especially we compute the diffusivity of the neutral self-interstitial in silicon bulk. The climbing image nudged elastic band method is used for the energy barrier and the transition state atomic configuration. The diffusion prefactor is obtained by using a classical rate theory, the dynamical matrix method. We compare with the diffusivity from another alternative way, ab initio molecular-dynamics simulations, at 1500K. They are in good agreement.
This work was supported by the Korea Research Foundation Grant funded by the Korean Government (MOEHRD) (KRF-2005？005-J01103). Also, this work was supported partly by the Korean Ministry of Information & Communication through Information Technology Research Center Program and Information Technology Professor Support Program supervised by Institute of Information Technology Assessment, and partly by the Korean Ministry of Science and Technology through Korea Institute of Science and Technology Evaluation and Planning. In addition, the author would like to acknowledge support from Korea Institute of Science and Technology Information under “The Seventh Strategic Supercomputing Applications Support Program” with Dr. Sang-Min Lee as the technical supporter. The use of the computing system of the Supercomputing Center is also greatly appreciated. Finally, the author gives special thanks to Professor Graeme Henkelman for his help in the dynamical matrix method through the forum: http://theory.cm.utexas.edu/forum.
- 초록
We demonstrate that local-density approximation in combination with the dynamical matrix method is a plausible method for calculating diffusion constants in solids. Especially we compute the diffusivity of the neutral self-interstitial in silicon bulk. The climbing image nudged elastic band method is used for the energy barrier and the transition state atomic configuration. The diffusion prefactor is obtained by using a classical rate theory, the dynamical matrix method. We compare with the diffusivity from another alternative way, ab initio molecular-dynamics simulations, at 1500K. They are in good agreement.
This work was supported by the Korea Research Foundation Grant funded by the Korean Government (MOEHRD) (KRF-2005？005-J01103). Also, this work was supported partly by the Korean Ministry of Information & Communication through Information Technology Research Center Program and Information Technology Professor Support Program supervised by Institute of Information Technology Assessment, and partly by the Korean Ministry of Science and Technology through Korea Institute of Science and Technology Evaluation and Planning. In addition, the author would like to acknowledge support from Korea Institute of Science and Technology Information under “The Seventh Strategic Supercomputing Applications Support Program” with Dr. Sang-Min Lee as the technical supporter. The use of the computing system of the Supercomputing Center is also greatly appreciated. Finally, the author gives special thanks to Professor Graeme Henkelman for his help in the dynamical matrix method through the forum: http://theory.cm.utexas.edu/forum.
More
