Computational studies of reacting flows with applications in nanoscale materials synthesis
The focus of this dissertation is on the development of fundamental models describing the vapor-phase synthesis of compound semiconductor nanostructures and on the optimal design of reactors used for manufacturing them.
Metalorganic chemical vapor deposition (or MOCVD) of compound semiconductors is a very common process in the microelectronics industry that involves flow of a mixture of organometallic precursors, diluted in hydrogen carrier gas, over a heated substrate resulting in the deposition of a thin single-crystalline film. In such a process, complex transport phenomena (flow, heat and mass transfer) are coupled with gas-phase and surface reactions between the organometallic precursors, their decomposition byproducts, and the hydrogen carrier gas. The development of predictive process models for such a process is hampered by its inherent complexity and the difficulty in obtaining in-situ measurements of species concentrations in a non-intrusive manner.
In this study, fundamental models of transport phenomena in vertical rotating-disk MOCVD reactors were coupled with kinetic models describing gas-phase and surface reactions during the growth of GaAs thin films. The models were validated by comparing numerical simulations with experimental data for GaAs thin film growth by MOCVD. The vertical rotating disk MOCVD reactor is a very common reactor configuration used for deposition of compound semiconductors. A stream of hydrogen carrying the film precursors flows downwards towards a heated rotating susceptor where the substrates to be coated are placed. Typical design criteria for such reactors include the elimination of buoyancy- and inertia-driven flow recirculations, uniformity of thickness and composition of the deposited films over large-area substrates, and the ability to grow atomically-abrupt interfaces between layers of different film materials to enable the manufacturing of advanced optoelectronic devices.
A comprehensive parametric study of vertical rotating-disk CVD reactors was performed using fundamental reaction-transport models and generalized design criteria were developed for: (1) The elimination of flow recirculations based on the Reynolds number, rotational Reynolds number, Grashof number, and the aspect ratio of the reactor (inlet to susceptor distance divided by susceptor diameter). (2) The maximization of compositional uniformity of deposited films based on Péclet and Damköhler numbers for reactor designs that meet the first criterion. The approach developed in this dissertation can be used for the optimal design of reactors used for growing multilayer structures of compound semiconductors, such as GaAs, (Al,Ga)As, GaN and (Al,Ga)N.
Compound semiconductor nanocrystals (also known as “quantum dots”) exhibit unique size-dependent properties, such as size-dependent luminescence, when their size becomes smaller than the wavelength of an electron-hole pair (or exciton). The tunability of their optoelectronic properties makes such materials suitable for emerging applications in the fields of ultrasensitive biological detection, high-density electronics, solar energy conversion, and photocatalysis. At present, control of particle growth rate and size distribution can be achieved only through experimental trial and error that is both expensive and time-consuming. Vapor-phase synthesis offers superior control on materials purity and elimination of solvents, when compared to liquid-phase techniques, and is also more compatible with existing processes in the semiconductor industry.
The vapor-phase synthesis of II-VI compound semiconductor nanocrystals, such as CdSe or ZnSe, poses challenging reactor design and optimization problems related to control of the particle size and size distribution. The focus of this part of the dissertation was the development of predictive models describing the nucleation and growth of compound semiconductor nanocrystals in a counter-flow jet reactor. Fundamental models of transport phenomena were coupled with population balance models describing the nucleation of nanoparticles and their growth by particle-particle coalescence. The models were validated using experimental results from a laboratory reactor and were used to perform a detailed parametric study of ZnSe nanocrystal synthesis. The effects of operating conditions on the size and size distribution of the nanocrystals were studied and optimal operating conditions were identified that yield nanocrystals with narrow size distribution and average sizes below the quantum confinement threshold for ZnSe, which is 9nm.