| Chemical vapor deposition (CVD) is extensively used to produce high quality films critical to modern technology. Furthermore, CVD promises to meet the ever-increasing requirements for new and advanced materials in the production of electronic devices, optical coatings, and wear and corrosion protection for engines and tools. The useful properties of these coatings are largely determined by their micro-structure. One of the advantages of CVD is that it allows the tailoring of these properties. While existing systems regulate the temperature, pressure and gas flow rates, feedback control strategies to achieve a desired film micro-structure have not been developed.; The development of a feedback system to control CVD requires understanding the dynamics and physics of the process. This dissertation develops a model suitable for the design of a real-time, closed-loop system to control the micro-structure of protective coatings produced by CVD. Based on a lumped control volume approach, the model captures the dominant dynamics and physics of both the gas phase and surface phenomena. The effects of reactant depletion due to deposition in the equipment are also included. While many of these aspects had been extensively studied separately under steady state conditions, little work has been aimed at integrating them and at studying their input/output dynamics.; Predictions of the nonlinear model are shown to be consistent with results from steady state experiments on the CVD of titanium nitride, an important material for protective coatings. The dynamic response of the system is analyzed. It is found that for cases in which depletion of the gas mixture is low, the bulk reactant partial pressures determine the behavior of the deposition rate with minimal influence of the boundary layer and surface mechanisms. At higher temperatures the mixture may be depleted and diffusion through the boundary layer can play a significant role.; The model is used to investigate mechanisms that determine the relative performance of alternative actuators and sensors of CVD coatings. The relationship between micro-structure and process variables is presented from the perspective of control systems. Effects that measurable physical variables and inputs can have in controller performance are discussed and a controller is designed. The proposed controller's performance is evaluated through simulations. |
