Experiments on Atomic-Scale Diffusion of Dopants in Strained and Relaxed Si and SiGe

University Of Houston, Houston TX

Investigators

Abstract

This project addresses diffusion of boron in Si or SiGe, utilizing a new experimental approach. The coefficient of diffusion (D) is normally expressed as a specific combination (D ~ X(squared)g) of two contributing factors: the average migrating distance (X) and hopping frequency (g) of the diffusing boron atoms. For some conditions, e.g., the diffusion of boron in relaxed versus strained SiGe, significant changes in the average migrating distance may be involved. In this project, experiments capable of distinguishing between changes in migrating distances from changes in hopping frequencies will be carried out. The approach involves refined measurements of the profile spreading of boron, starting from a sharply localized layer in a superlattice. Standard diffusion theory with the simplest models indicates that this profile is a Gaussian for all times. When details of the interaction between dopants (boron) and defects in parent materials (SiGe) on the atomic scale is taken into account, it is expected that the profile acquires an exponential character for short times. It is planned to assess the migrating distance and the hopping frequency separately by monitoring the gradual disappearance of the exponential character of the profile as it evolves into a pure Gaussian with increasing annealing time. Doing this at a range of annealing temperatures is expected to elucidate the dependence of these quantities separately as a function of temperature. This information is expected to provide new knowledge of the basic mechanism behind diffusion of boron in Si or SiGe, and assist explanation of many documented anomalies. Besides diffusion of boron in Si, strained or strain-relaxed SiGe systems, the superlattice diffusion marker approach can be extended to other solid-state systems of which the diffusion mechanism is either unknown or in debate. For example, the presence of F, N, C, Al, Ga, and In, etc. may affect the migration of a specific dopant. Attempts to understand and predict such phenomena using ab initio calculations to compute energy barriers under different chemical environments are complicated by interrelations of these atoms and associated strain fields; thus, experimental work is needed to complement computational results. It is anticipated that the experimental approach developed in this project will have significant impact in the field of diffusion of impurities in solids in general. The project includes ab initio calculations of reaction barrier energies under different chemical and/or physical (strain) environments to compare and contrast with measurements. Macroscopic consequences of reaction barrier energies extracted from experiments will be sought and evaluated. A Monte Carlo simulation code will be developed for computer experimentation to lend further support to measured migrating distances and hopping frequencies. %%% The project addresses fundamental research issues associated with electronic/photonic materials having technological relevance and emphasizes the integration of research and education. The project will involve an undergraduate and two graduate students each participating as a team member in conducting research at various stages. The project also benefits instruction in film growth, material modification, ion-solid interaction and materials analysis in chemistry and physics departments. Activities are also devoted to educating high school teachers through special lectures and high school students through a summer internship program. ***