화학공학소재연구정보센터
Journal of Crystal Growth, Vol.240, No.3-4, 581-602, 2002
Effect of steady ampoule rotation on radial dopant segregation in vertical Bridgman growth of GaSe
For vertical Bridgman growth of the nonlinear optical material GaSe in an ampoule sufficiently long that flow and dopant transport are not significantly influenced by the upper free surface, we show computationally that steady rotation about the ampoule axis strongly affects the flow and radial solid-phase dopant segregation. Radial segregation depends strongly on both growth rate U and rotation rate Omega over the ranges 0.25 mum s(-1) less than or equal to U less than or equal to 3.0 mum s(-1) and 0 less than or equal to Omega less than or equal to 270 rpm. For each growth rate considered. the overall radial segregation passes through two local maxima as Omega increases, before ultimately decreasing at large P. Rotation has only modest effects on interface deflection. Radial segregation computed using a model with isotropic conductivity (one-third the trace of the conductivity tensor) predicts much less radial segregation than the "correct" model using the anisotropic conductivity, with the segregation decreasing monotonically with Omega. Consideration of a model in which centrifugal acceleration is deliberately omitted shows that, as 9 increases, diminution and ultimately disappearance of the "secondary" vortex lying immediately above the interface is due to centrifugal buoyancy, while axial distension of the larger "primary" vortex above is due to Coriolis effects. These results. which are qualitatively different from those accounting for centrifugal buoyancy. suggest that several earlier computational and analytical predictions of rotating vertical Bridgman growth are either limited to rotation rates sufficiently low that centrifugal buoyancy is unimportant, or are artifacts associated with its neglect. The overall radial segregation depends approximately linearly on the product of 1 - (k) over bar and the growth rate U for the conditions considered, where (k) over bar is the segregation coefficient. (C) 2002 Elsevier Science B.V. All rights reserved.