William J. Debnam

A procedure to visualize the melt-solid interface in Bridgman grown germanium and lead tin telluride

Abstract An X-ray and gamma-ray technique has been developed that enables the melt-solid interface to be observed during crystal growth in a Bridgman furnace. This technique has been used to observe the interface movement and shape in germanium and in lead tin telluride.

Properties of GaN grown on sapphire substrates

Epitaxial growth of GaN on sapphire substrates using an open-tube growth furnace has been carried out to study the effects of substrate orientation and transfer gas upon the properties of the layers. It has been found that for the (0001) substrates, surface appearance was virtually independent of carrier gas and of doping levels. For the (1 ¯102) substrates surface faceting was greatly reduced when He was used as a transfer gas as opposed to H2. Faceting was also reduced when the GaN was doped with Zn and the best surfaces for the (1 ¯102) substrates were obtained in a Zn-doped run using He as the transfer gas. The best sample in terms of electrical properties for the (1¯102) substrate had a mobility greater than 400 cm2 V−1 sec−1 and a carrier concentration of about 2 × 1017 cm−3. This sample was undoped and used He as the transfer gas. The best (0001) sample was also grown undoped with He as the transfer gas and had a mobility of 300cm2V−1 sec−1 and a carrier concentration of 1 × 1018 cm−3.

Magnetically damped convection and segregation in Bridgman growth of PbSnTe

Abstract The effects of an axially imposed magnetic field on convection and solute segregation during Bridgman growth of a non-dilute multicomponent system Pb 0.8 Sn 0.2 Te were studied using a finite-element model. The model considers heat and mass transport, fluid motion, solid/liquid-phase change and magnetic damping. The main objectives are to provide a quantitative understanding of the complex transport phenomena during solidification in a magnetic field, to provide estimates of the required magnetic field strength for low gravity growth, and to assess the role of magnetic damping for space and earth growth control. Numerical results for both vertical and horizontal growth configurations are presented. In addition to full-scale simulation, a revised scaling analysis is also presented.

Interface shapes during vertical bridgman growth of (Pb, Sn)Te crystals

Abstract Melt-solid interfaces obtained during vertical Bridgman growth of (Pb´, Sn)Te crystals were investigated with a quenching technique. The shapes of these interfaces, revealed by etching longitudinally cut sections, were correlated with the composition variations determined by electron microprobe analysis. These experiments demonstrated that the interface shape can be changed from concave to convex by moving its location from the edge of the cold zone into the hot zone. The metallography and microsegregation near the melt-solid interface were analyzed in detail. A sharp change in composition above the interface indicated the existence of a diffusion boundary layer of 40–90 μm thick. This small diffusion boundary layer is consistent with strong convective mixing in the (Pb, Sn)Te melt.

Growth rates and interface shapes in germanium and lead tin telluride observed in-situ, real-time in vertical Bridgman furnaces

Using the advanced technology developed to visualize the melt-solid interface in low Prandtl number materials, crystal growth rates and interface shapes have been measured in germanium and lead tin telluride semiconductors grown in vertical Bridgman furnaces. The experimental importance of using in-situ, real time observations to determine interface shapes, to measure crystal growth rates, and to improve furnace and ampoule designs is demonstrated. The interface shapes observed in-situ, in real-time were verified by quenching and mechanically induced interface demarcation, and they were also confirmed using machined models to ascertain the absence of geometric distortions. Interface shapes depended upon the interface position in the furnace insulation zone, varied with the nature of the crystal being grown, and were dependent on the extent of transition zones at the ends of the ampoule. Actual growth rates varied significantly from the constant translation rate in response to the thermophysical properties of the crystal and its melt and the thermal conditions existing in the furnace at the interface. In the elemental semiconductor germanium the observed rates of crystal growth exceeded the imposed translation rate, but in the compound semiconductor lead tin telluride the observed rates of growth were less than the translation rate. Finally, the extent of ampoule thermal loading influenced the interface positions, the shapes, and the growth rates.

Effects of supercooling in the initial solidification of PbTe-SnTe solid solutions

Abstract Deviations from compositions anticipated by the thermal equilibrium phase diagram have been observed in Bridgman-grown crystals of Pb 1− x Sn x Te, in the first to freeze region of the boule. A set of experiments were conducted to determine the extent of thermal supercooling of Pb 1− x Sn x Te in a Bridgman-like configuration. The results of the compositional profiles and the supercooling measurements are consistent with a diffusionless transformation occurring at the onset of solidification and the length of uncontrolled growth is inversely related to the temperature gradient of the furnace.

Oscillation phase relations in a Bridgman system

The thermal conditions under which convection in a tin melt undergoes the transition from steady flow to time-dependent flow are investigated. Previously, it has been shown that crystals that are directionally solidified from a time-dependent melt will have a larger defect density and increased chance of compositional striations. Thus, it is important for this transition to be well characterized. The experimental results to be presented were obtained from a two-zone Bridgman furnace with a middle insulation zone. Thermocouples were placed on the axis and on the outer wall of a cylindrical vitreous silica tube which contained molten tin. The phase relations between temperature oscillations measured at different positions in the cell are discussed. Fourier transforms are used to investigate the increasing complexity of convection as the temperature gradient is increased.

A technique for measuring the heat transfer coefficient inside a Bridgman furnace

Knowledge of the amount of heat that is conducted, advected and radiated between an ampoule and the furnace is important for understanding vertical Bridgman crystal growth. This heat transfer depends on the temperature, emissivities and geometries of both the furnace and ampoule, as well as the choice of ambient gas inside the furnace. This paper presents a method which directly measures this heat transfer without the need to know any physical properties of the furnace, the ampoule, or the gaseous environment. Data are given for one specific furnace in which this method was used.

Heat transfer measurements in the Bridgman configuration

Abstract A heat transfer measuring device was developed to help characterize a vertical Bridgman furnace. The actively heated device, the heat transfer sample HTX, measures the temperature difference created between it and the furnace wall as a function of input power. This allows the heat transfer coefficient to be measured without having to know the physical properties of the sample, the ambient gas, or the furnace. However, if some of these physical properties are known, then the HTX can be used to determine others. Heat transfer measurements were made under a variety of conditions. The convective heat loss was varied by making measurements in air, argon, helium, and vacuum environments. The radiative heat loss was varied by using fused silica, oxidized inconel, and platinum foil sleeves placed over the HTX heater. A convection coefficient for air and the emissivities of the different sleeves were determined.

Modelling melt-solid interfaces in Bridgman growth

Abstract Doped epoxy models with abrupt interfaces were prepared to test radiographic and computer enhancement procedures used to study the images of melt–solid interfaces during crystal growth in Bridgman furnaces. A column averaging procedure resulted in improved images that faithfully reproduced the positions and shapes of interfaces even at very low density differences. These techniques were applied to lead tin telluride growing in Bridgman furnaces.