Machine Learning-Assisted Synthesis of Two-Dimensional Materials.

Area of Science:

• Materials Science
• Nanotechnology
• Computational Chemistry

Background:

• Two-dimensional (2D) materials possess unique properties for electronics, optoelectronics, and energy storage.
• Controllable synthesis of 2D materials remains a significant challenge in materials science.
• Machine learning (ML) offers a powerful approach to accelerate materials discovery and synthesis.

Purpose of the Study:

• To explore synthesis parameters for 2D molybdenum disulfide (MoS2) using ML.
• To evaluate the performance of different ML algorithms in predicting successful MoS2 synthesis.
• To identify critical factors influencing the chemical vapor deposition (CVD) growth of MoS2.

Main Methods:

• Investigated synthesis parameters for MoS2 using four ML algorithms: XGBoost, Support Vector Machine (SVM), Naïve Bayes (NB), and Multilayer Perceptron (MLP).
• Assessed model performance using metrics including recall, specificity, and accuracy.
• Determined feature importance for MoS2 growth, focusing on reaction temperature, Ar gas flow rate, and reaction time.

Main Results:

• XGBoost demonstrated superior performance with over 88% prediction accuracy and an AUROC of 0.91.
• Reaction temperature was identified as a crucial parameter influencing MoS2 growth.
• ML analysis optimized feature importance, highlighting key synthesis variables.

Conclusions:

• ML-assisted materials preparation significantly reduces experimental time and trial-and-error.
• This approach provides valuable insights for the efficient synthesis of 2D materials.
• The findings pave the way for accelerated development and application of advanced 2D materials.