Open Access

ARTICLE

Numerical Simulation of Liquid Phase Diffusion Growth of SiGe Single Crystals under Zero Gravity

by M. Sekhon¹, N. Armour¹, S. Dost¹

1 Crystal Growth Laboratory, Department of Mechanical Engineering, University of Victoria, Victoria, BC Canada V8W3P6.
2 Corresponding author: sdost@me.uvic.ca

Fluid Dynamics & Materials Processing 2013, 9(4), 331-351. https://doi.org/10.3970/fdmp.2013.009.331

Abstract

Liquid Phase Diffusion (LPD) growth of Si_xGe_1-x single crystals has been numerically simulated under zero gravity. The objective was to examine growth rate and silicon concentration distribution in the LPD grown crystals under diffusion dominated mass transport prior to the planned LPD space experiments on the International Space Station (ISS). Since we are interested in predicting growth rate and crystal composition, the gravitational fluctuation of the ISS (g-jitter) was neglected and the gravity level was taken as zero for simplicity.
A fixed grid approach has been utilized for the simulation. An integrated top-level solver was developed in OpenFOAM to carry out numerical simulations for the melting and solidification periods of the LPD growth process. The solver employs the well-known enthalpy method for modeling the initial melting process and uses the virtual front-tracking method, originally developed to model dendritic growth. This simulates the solidification as driven by saturation and precipitation as is the case for this solution growth technique. The melting simulation determines the initial conditions for growth interface, temperature, and concentration. The solver then calculates the onset of solidification, the evolution of the growth interface. In addition, the concentration and temperature fields are calculated in the melt and grown crystal.
The present simulation results agree qualitatively with the radial and axial silicon distributions in the grown crystals of the Earth-bound experiments, and also with those previously predicted numerically. The computed total growth rate also agrees quantitatively with that of the experiment. However, the simulation shows slight differences in the interface shapes and predicts faster initial growth rate. Such a small discrepancy is expected since the contribution of natural convection in the melt was not included in the present simulation. A well-design LPD space experiment may shed light on this prediction.

---

Keywords

---