Thin film solar cells, which consist of multiple layers such as the PIN layers and the Transparent Conducting Oxide (TCO) layers, are playing a more and more important role in the overall solar cell market owing to the potential of improving light conversion efficiencies (currently on the order of 10$\%$ for production modules). Over the last decade, it has been widely recognized that the surface morphology at each interface, which is characterized by surface root-mean-square roughness and slope, has crucial influence on the light conversion efficiency of thin film solar cells. Therefore, precisely shaping the surface morphologies of different layers in thin film solar cells during the thin film deposition process is a promising way to improve solar cell efficiency. Despite its importance, the computational modeling and control of the surface morphology, especially of the root-mean-square surface mean slope, during the thin film deposition process and its application to improve solar cell performance have not received enough attention.

This dissertation presents a systematic framework for modeling and control of thin film surface morphology in both PIN layers and TCO layers. Specifically, we will present novel definitions for describing surface morphology by introducing both the surface root mean square roughness and slope to describe the surface morphology, study its physical properties and dependence on model parameters such like lattice size, activation energies and temperature, introduce and identify stochastic closed-form equations describing surface morphology, and finally design model predictive controllers to control the surface morphology to desired levels. Extensive simulation results are presented to demonstrate the effectiveness of the proposed modeling and control framework.