Control and Monitoring of Microstructural Defects in Thin Film Deposition

University Of California-Los Angeles, Los Angeles CA

Investigators

Abstract

PI: Panagiotis Christofides Institution: UCLA Proposal Number: 0652131 Title: Control and Monitoring of Microstructural Defects in Thin Film Deposition Intellectual merit: Microelectronic devices are fabricated through a series of processing steps including numerous thin film deposition processes. Owing to current trends towards decreased device dimensions, thin film properties, such as surface roughness and amount of microstructural defects have recently emerged as important film quality variables which strongly influence the overall electrical and mechanical properties of microelectronic devices. While significant research work has been recently done in achieving precise regulation of thin film surface roughness during deposition, at this point, the problem of controlling thin film microstructural defects (which can be naturally thought of as the problem of minimizing thin film porosity) despite its importance have received little attention. Thin film internal microstructure strongly influences film electrical properties since the vacancies in the microscopic structure of a thin film provide free sites for undesired electrical static charge, high leakage current and longer latency, thereby lowering transistor operating speed. For example, in the case of gate dielectrics, it is important to reduce thin film porosity as much as possible and prohibit the development of holes close to the interface. Motivated by these considerations, the objective of this research is to develop a systematic framework for the design of feedback control systems for real-time control of thin film microstructural defects using an integration of multiscale (i.e., coupled macroscopic/microscopic) models and measurements. Since thin film microstructure is determined by microscopic processes like atom adsorption, desorption and migration, a key element of the research is the introduction of a new method for the construction of stochastic dynamic models, which are suitable for controller design and real-time controller implementation, using data from multiscale process models that predict the effect of controllable process variables on thin film porosity. On the basis of these stochastic models, nonlinear and predictive control theory will be developed and used to produce practically-implementable, feedback control systems that enforce the desired stability, performance and robustness specifications in the closed-loop system. Specifically, the research will focus on the following projects: 1. Multiscale modeling of thin film growth capturing the relationship between controllable (macroscopic) process variables and thin film porosity. 2. Construction of stochastic dynamic models that describe the effect of controllable process variables on the evolution of thin film porosity using multiscale process model data. 3. Design of estimation and feedback control systems for real-time thin film porosity minimization using the stochastic dynamic models and process measurements. 4. Design of fault-detection filters for assessing actuator/sensor/controller abnormal behavior and controller reconfiguration strategies for dealing with abnormal process events. 5. Applications to thin film growth processes using detailed multiscale models and realistic porosity specifications. The research will be carried out in close collaboration with engineers and scientists at Intel Corporation. Broader impact: These control methods are expected to improve the operation and performance of semiconductor manufacturing processes, increase process yield and reliability, and minimize the negative economic impact of failures on overall process operation. The research will address the design of feedback control and estimation systems accounting explicitly for multiscale process behavior and the occurrence of actuator/sensor/controller faults, as well as addressing the integration of fault-detection and control strategies. The incorporation of research results into education and the publication of a book will benefit educators and students through the development and offering of advanced-level courses in process control and multiscale modeling. The development of software, short courses and workshops and the collaboration with Intel Corporation will be the means for transferring the results of this research to the industrial sector. Moreover, the research will benefit from and contribute to a number of educational initiatives and innovations in the UCLA campus in the area of systems, dynamics and control.