There has been growing demand for high performance nanochips due to the Fourth Industrial Revolution. However, semiconductor manufacturing leaders are struggling to meet the need due to the current challenges and requirements, such as displacement of vertically stacked architectures and stricter film quality. Despite leveraging revolutionary technology such as atomic layer deposition (ALD), the development of more advanced techniques are necessary to address these challenges. Recently, atomic layer etching, as a counterpart of ALD, and area-selective atomic layer deposition, as a bottom-up nanopatterning technique, have recently gained momentum due to their potential abilities to satisfy the stringent quality specification and assemble nanostructures in a self-aligned manner. Thus, in this dissertation, a series of in silico research using multiscale computational fluid dynamics modeling have been performed to provide comprehensive understandingand insight to industry with respect to operation and robust control in order to further miniaturize the dimension of chips and precisely stack 3D features on semiconductors.

Facilitated by the increasing importance and demand of thin-film materials, plasma-enhanced atomic layer deposition (PEALD) has gained tremendous industrial interest as it offers a way to efficiently deposit thin-films with ultra-high conformity. Despite the variety of PEALD processes, there lacks a fundamental and general methodology to understand, characterize and control a realistic PEALD process system. To fully understand the PEALD process, a series of studies have been carried out in our previous work. First, a kinetic Monte-Carlo (kMC)-based microscopic model that describes the surface dynamics was implemented and a multiscale CFD model was developed to characterize the PEALD process. Additionally, a corresponding multiscale data-driven model was derived to efficiently explore the optimal operating condition in the process domain based-on an elementary cost-analysis. The successful development of the aforementioned models enables the further study into the process control of PEALD where disturbances in operating conditions are present. In this work, an integrated control scheme using a proportional-integral (PI) controller and a run-to-run (R2R) controller is proposed and implemented. Using the developed multiscale CFD model, the PEALD process under typical disturbances is simulated, and the controllers are applied in the process domain. The result demonstrates the successful mitigation of disturbances in operating pressure, inlet molar fraction and gas feeder temperature under the combined effort of both controllers.