Modelling and control of silicon and germanium thin film chemical vapor deposition

This study integrates in situ robust and efficient fundamental models and non-invasive optical sensors with state-of-the-art estimation and model predictive control techniques in order to grow unusual and aggressive Si_1−xGe _x alloy films.; Fundamental kinetic models are developed that describe the complicated behavior of the silicon/germanium deposition system. The kinetic models are incorporated into material balances yielding a physical model of the film growth rate as a function of reactor environment variables. The model is validated by comparison to experimental data.; An on-line spectral ellipsometer is installed on the deposition reactor and a dynamic sensor model that characterizes the optical properties of the growing film is presented. The sensor model is integrated with the kinetic and process models that describe the reactor and predicts the ellipsometric parameters Ψ and Δ during film growth.; An extended Kalman filter is developed that utilizes the process model and the sensor model and estimates uncertain system states given ellipsometric measurements. The filter provides state estimates of film growth rate and composition in real time. These measurements are utilized to update the optimization problem solved by the regulator.; A model predictive controller is presented that regulates the various reactor variables, such as precursor gas flow rates, substrate temperature, and reactor pressure, in order to grow an aggressive Si_{1− x}Ge_x alloy film with precise control over film growth rate and composition. Feedback from the optical sensors rejects unmeasured disturbances and eliminates run-to-run variability allowing for precise control over the film qualities and for the fabrication of novel devices with exact specifications.; The model predictive controller is tested by growing aggressive Si _1−xGe_x alloy composition profiles in which flow rates of several component gases and the reactor temperature are varied simultaneously in order to achieve the profile of interest.; The films grown utilizing feedback from the optical sensor are compared to the films grown using open-loop recipes, which is the current industrial practice. These films are compared quantitatively utilizing off-line characterization techniques, such as auger spectrometry and secondary-ion-mass-spectrometry (SIMS) analysis.