Revealing Variable Dependences in Hexagonal Boron Nitride Synthesis via Machine Learning

Wafer-scale monolayer two-dimensional (2D) materialshave beenrealized by epitaxial chemical vapor deposition (CVD) in recent years.To scale up the synthesis of 2D materials, a systematic analysis ofhow the growth dynamics depend on the growth parameters is essentialto unravel its mechanisms. However, the studies of CVD-grown 2D materialsmostly adopted the control variate method and considered each parameteras an independent variable, which is not comprehensive for 2D materialsgrowth optimization. Herein, we synthesized a representative 2D material,monolayer hexagonal boron nitride (hBN), on single-crystalline Cu(111) by epitaxial chemical vapor deposition and varied the growthparameters to regulate the hBN domain sizes. Furthermore, we exploredthe correlation between two growth parameters and provided the growthwindows for large flake sizes by the Gaussian process. This new analysisapproach based on machine learning provides a more comprehensive understandingof the growth mechanism for 2D materials.

Automated summary

Wafer-scale monolayer two-dimensional (2D) materials have been achieved through epitaxial chemical vapor deposition (CVD) in recent years. To scale up the synthesis of 2D materials, a comprehensive analysis of the growth dynamics and its dependence on the growth parameters is important. However, most studies on CVD-grown 2D materials have considered each parameter as an independent variable, which is not comprehensive for growth optimization. In this study, monolayer hexagonal boron nitride (hBN) was synthesized on single-crystalline Cu(111) using CVD, and the growth parameters were varied to regulate hBN domain sizes. The correlation between two growth parameters was explored, and a more comprehensive understanding of the growth mechanism for 2D materials was provided through machine learning.