Abstract:Real-time monitoring and control of epitaxial growth has been recognized as an enabling technology for volume manufacturing of epitaxial-based compound semiconductor optoelectronic and electronic devices (lasers, microwave transistors etc.). Conventional approach which is based on pre-growth calibration does not allow real-time compensation for the drift of process parameters that occurs during the growth process (growth time varies from I to 10 hours). By monitoring the deposited material directly and continuously, the new approach in this thesis based on model-reference system and least square phase not only ensure higher growth reproducibility, but also eliminate calibration runs completely. To date, in situ non-invasive normal incidence laser reflectance, pyrometric interferometry and diffused reflection spectroscopy are increasing being used for in situ growth calibrations for both molecular beam epitaxy and metal-organic vapor phase deposition growths. The physical original of non-invasive normal incidence laser reflectance, pyrometric interferometry, diffused reflectance spectroscopy has been studied. The interpretation of sample temperature from the measured diffused reflectance spectroscopy spectrum is influenced by interference effect in the heterostructure. The compensation algorithm with the aid of curvature has been investigated. And our experiments have discovered that temperature drift during MBE growth of heterostructures is caused by change of sample emissivity as well as radiative heating from the Knudsen cells. The identification of these mechanisms will lead to a more precise modeling of sensor signal and thus improving the performance of the closed-loop control system. Using ‘signal phase’ and a model-reference system, we have achieved continuous growth rate estimation with thickness resolution of 30 nm on both Pyrometric Interferometry (PI) and Laser Reflectance (LR). This is nearly five-fold improvement over the previous technique. This new method also achieved improved accuracy (thickness error < 2%) over conventional methods (error ∼4–7%). In addition, the new method shows excellent dynamic tracking of growth rate when the growth rate is intentionally changed by as much as 18%. Finally, the automation of MBE system and measurement synchronization during wafer rotation are also designed and investigated in this thesis.

Knowledge-infused Process Monitoring for Quality Improvement in Solar Cell Manufacturing Processes

A Condition Change Detection Method for Solar Conversion Efficiency in Solar Cell Manufacturing Processes

An Automatic Condition Detection Approach for Quality Assurance in Solar Cell Manufacturing Processes

Optimal Online Soft Sensor for Product Quality Monitoring in Propylene Polymerization Process

Enhancing solar cell production line monitoring through advanced statistical analysis

Numerical Analysis of Efficiency Enhancement in Plasmonic Thin-Film Solar Cells by Using the Silvaco Tcad Simulator

Advancement in Treating Some Features of CIGS Thin Film Solar Cells During Manufacturing

The digital twin of the quality monitoring and control in the series solar cell production line

A Combined Numerical Modeling and Machine Learning Approach for Optimization of Mass-Produced Industrial Solar Cells

Machine learning with knowledge constraints for process optimization of open-air perovskite solar cell manufacturing

Quantitative o perando visualization of the energy band depth profile in solar cells

Discovering Process Dynamics for Scalable Perovskite Solar Cell Manufacturing with Explainable AI

Ensemble Modeling of in Situ Features for Printed Electronics Manufacturing with in Situ Process Control Potential

Quantifying Infiltration for Quality Control in Printed Mesoscopic Perovskite Solar Cells: A Microscopic Perspective

Photoluminescence Imaging for the In-Line Quality Control of Thin-Film Solar Cells

Understanding Causalities in Organic Photovoltaics Device Degradation in a Machine-Learning-Driven High-Throughput Platform

In-Situ Self-Monitoring of Real-Time Photovoltaic Degradation Only Using Maximum Power Point: the Suns-Vmp Method

Machine Learning for Ultra High Throughput Screening of Organic Solar Cells: Solving the Needle in the Haystack Problem

Intelligent Monitoring and Control System During Molecular Beam Epitaxy Growth

Preliminary empirical results suggesting the mapping of dynamic in situ process signals to real-time wafer attributes in a plasma etch process

Utilizing Machine Learning and Diode Physics to Investigate the Effects of Stoichiometry on Photovoltaic Performance in Sequentially Processed Perovskite Solar Cells