Return to search

Influence of static magnetic fields and solutal buoyancy on silicon dissolution into germanium melt

Elemental semiconductors like silicon and germanium have been used since the beginning of the electronics industry. Silicon has dominated research and production and thus silicon based devices can be produced at the lowest cost using the most mature technology. While dopants can be used to tailor the electric properties of the semiconductor within certain limits, more flexibility is gained using compound semiconductors such as silicon-germanium. The electric properties of a compound semiconductor are highly dependant on the composition, which in turn is influenced by the dissolution reaction and flow characteristics during the growth process.

Liquid phase diffusion (LPD) is a solution growth technique that has been proposed to grow silicon-germanium seed crystals for other growth techniques. The dissolution of silicon is a limiting factor for the growth rate in LPD and also Bridgman growth techniques. Investigation of the dissolution process is aimed at increasing the growth rate while still maintaining maximum uniformity of the crystal composition. To accomplish this, a static magnetic field was utilized in experiments done by Armour. The experimental results showed that a top seeded configuration without magnetic fields leads to a diffusion driven process and homogeneous dissolution, while the addition of a strong 0.8 Tesla magnetic field resulted in non-uniform and slightly increased dissolution. This work is complementary to the experimental investigation and aims to help understand the influence of magnetic fields on silicon dissolution.

For this work, an OpenFOAM magnetohydrodynamics application including heat and species transport and three different magnetic force models has been developed and validated. The simulations done show that an isothermal state is reached within 90 seconds if no temperature gradient is imposed. Additional simulations with a temperature gradient helped to rule out a possible thermal leak in the experimental system, confirming that it must have been close to isothermal. Since the solutal expansion coefficient of has not been measured properly to the Author's knowledge, two possible values for the expansion coefficient have been considered. It has been found that the exact value of the solutal expansion coefficient does not have a great influence on the results of this work.

Identiferoai:union.ndltd.org:uvic.ca/oai:dspace.library.uvic.ca:1828/1900
Date26 November 2009
CreatorsKidess, Anton
ContributorsDost, Sadik
Source SetsUniversity of Victoria
LanguageEnglish, English
Detected LanguageEnglish
TypeThesis
RightsAvailable to the World Wide Web

Page generated in 0.0016 seconds