Experimental validation of (cadmium,zinc)telluride crystal growth model

The results from a series of temperature profiling and crystal growth experiments are used to validate a global finite element model of the (Cd,Zn)Te vertical Bridgman-Stockbarger crystal growth system. A parameter of special interest is the shape of the solid-liquid interface, because it is thought to be a key factor controlling nucleation at the wall, thus determining whether a large single crystal can be achieved. Crystal growth was carried out with a steep gradient of approximately 2°C/mm. Examination of the resulting grain structure suggests that nucleation at the wall was avoided except at the latter stage of growth. Solute concentration or furnace temperature asymmetry may have contributed to this result. A global heat transfer model based on the experimental apparatus is used to predict the shape and position of the solid/liquid interface. The model includes the furnace, furnace liner, ampoule support structure, crucible, and the solid and molten (Cd,Zn)Te. The effects of conductive and radiative heat transfer as well as the release of latent heat at the interface are included. Model results are reported with a 95% confidence band due to the uncertainty in material properties data used in the model. Experimental results are also reported with appropriate 95% confidence intervals. The level of combined experimental and modeling uncertainty is used as a criteria for model validation, and comparisons between experimental and modeled results show that validation is achieved. However, the direction of interface curvature, whether convex or concave, cannot be predicted due to uncertainty in the material properties. The result is also computed using a local model which treats only crucible and solid and molten (Cd,Zn)Te, and uses an assumed temperature profile as a crucible boundary condition. The interface shape predicted using this model cannot be differentiated from that predicted by the global model, given the large modeling uncertainties.