Professor Ma's research interests are in the areas of heat transfer, fluid dynamics and species transport as applied to growth of photonic semiconductor crystals. This research involves the analytical and numerical modelling of transport phenomena during various crystal growth processes in the presence of magnetic and electric fields. The modelling of these phenomena is extremely important for support of on-going experimental efforts in order to guide process optimization for the production of high-quality semiconductor ingots which serve as bases for optoelectronic devices. Materials of interest are photonic semiconductor crystals, namely, doped or alloyed III-V semiconductors, alloyed II-VI semiconductors, and high-purity silicon.

The quality of any semiconductor is directly related to the uniformity of its composition, or compositional homogeneity, because devices are produced on surfaces of thin wafers sliced from a semiconductor crystal. When only a few devices were fabricated on a wafer, large compositional deviations could be tolerated because the devices were large enough to see the local average. Even a small local compositional variation can lead to the malfunction of a device. With the recent growth in the marketplace and with the use of micro- and nano-scale devices, there is a critical need for compositional homogeneity. Therefore models which accurately predict the transport of species during crystal growth are needed in order to determine the optimal processes which minimize segregation in the photonic semiconductor. The development of such models are critically important for several reasons. (1) Currently, high-quality crystals of compound or alloyed semiconductors can be grown by molecular methods but their production rates are roughly one-thousand times smaller than those for bulk methods such as the Czochralski, Bridgman, vertical gradient freezing or floating-zone methods. Recent rapid advances in optoelectronics have led to a great demand for more and larger crystals, so that molecular methods cannot be used to produce the quantity suitable for future needs. (2) Current rejection rates of integrated circuits and optoelectronic devices are extremely high due to unacceptable compositional variations in the wafer. This has enormous economic consequences because the only way to determine whether a wafer is viable is to fabricate and test the entire device. (3) Because molten semiconductors are opaque and very high in temperature, experimental measurements of transport variables during growth are extremely difficult to obtain so that little data are available. With the essentially infinite number of possible combinations of controllable process parameters, an empirical trial-and-error search is not practical. Therefore, model predictions of the crystal's composition need to be compared with measurements of concentrations in wafers sliced from a crystal in order to guide process optimization.

The segregation of dopants during silicon or III-V crystal growth, and the segregation of species during alloyed III-V or alloyed II-VI crystal growth depend on the convective and diffusive transport of the components in the melt during crystal growth. Since molten semiconductors are excellent electrical conductors, externally-applied magnetic and electric fields are used to control the melt motion in order to improve the crystal's quality. Major objectives of our research are to understand, predict and control the segregation during various processes for different semiconductor materials. Our models provide accurate predictions of the unsteady species transport for the entire period of time needed to grow a crystal, and are used to systematically search over all possible combinations of controllable process parameters in order to determine the optimal processes which minimize segregation in the crystal.

ABSTRACTS