Improving the coverage area and flake size of ReS₂through machine learning in APCVD

Area of Science:

• Materials Science
• Nanotechnology
• Chemical Engineering

Background:

• Optimizing material synthesis, especially for 2D semiconductors like rhenium disulfide (ReS2), is complex due to numerous growth parameters.
• Chemical vapor deposition (CVD) is a key technique for synthesizing these materials, but controlling its parameters is challenging.

Purpose of the Study:

• To identify optimal CVD growth conditions for maximizing the covered area and crystal size of rhenium disulfide (ReS2).
• To leverage machine learning for understanding the complex interplay between CVD parameters and material properties.

Main Methods:

• Conducted orthogonal experiments varying temperature, carrier gas flux, and precursor position.
• Utilized an artificial neural network to establish correlations between thermodynamic variables.
• Generated contour plots to visualize the impact of parameters on crystal growth.

Main Results:

• Identified optimal CVD parameters for enhanced ReS2 crystal growth.
• Demonstrated the effectiveness of machine learning in predicting and visualizing growth outcomes.
• Established clear correlations between experimental variables and crystal coverage/size.

Conclusions:

• Machine learning significantly enhances the optimization of CVD processes for 2D semiconductors.
• This approach facilitates the large-scale integration of materials like ReS2.
• The study provides a framework for applying AI to complex material synthesis challenges.