Machine-learning assisted epitaxy analysis

Introduction: High-quality epitaxial materials are crucial for their integration into various device applications, including quantum information technologies. Key characteristics such as stoichiometry, surface roughness, and crystallinity are essential indicators of thin-film material quality. While numerous characterization techniques have been developed to measure these properties with high accuracy, subtle details often remain undetected by the human eye. By applying machine learning to characterization data, we aim to uncover these hidden details and extract additional, precise information to enhance the quality of epitaxial materials.

Objective: The primary objective of this internship is to explore and investigate the use of machine-learning techniques in epitaxy analysis as an advanced characterization method. The goal is to optimize the quality of epitaxial materials to unprecedented levels. With many techniques existing for the analysis of characterization data (PCA, K-means clustering, deep learning models, etc.), the optimal method for this desired result must be found and deployed within the existing data analysis infrastructure.

Methodology: This research will focus on the comprehensive analysis of existing characterization data using machine learning algorithms. The aim is to identify and interpret critical details that are not easily detectable through conventional methods. By leveraging machine learning, we hope to gain deeper insights into the properties of epitaxial materials and develop strategies to improve their quality. Research acceleration will be enabled by the streamlining of this data analysis flow.

Expectations from the Candidate:  The internship candidate is expected to:

• Getting familiar with material characterization techniques to monitor the crystal quality
• Acquire and/or prepare experimental data for training and testing machine learning algorithms.
• Implement, deploy, test and document machine learning assisted data analysis.
• Willingness to learn, being involved in technical discussions and group meetings
   

Candidate Profile:  The ideal candidate possesses a solid background in materials engineering / physics and machine learning applied to these domains.

Support and Guidance:  The research will be supported and guided by experts from different domains within imec, ensuring a multidisciplinary approach to problem-solving and leveraging a wealth of knowledge and resources. Your daily interactions will be supported by two PhD students with hands on experience in the project.

Daily supervision: Danut Dinu,  Andries Boelen

Promotors: Clement Merckling, Christian Haffner