Accelerated screening of colloidal nanocrystals using artificial neural network-assisted autonomous flow reactor...

Area of Science:

• Materials Science
• Nanotechnology
• Chemical Engineering

Background:

• Colloidal semiconductor nanocrystals offer tunable properties for advanced applications like optoelectronics and bioimaging.
• Current methods for synthesizing nanocrystals are resource-intensive and slow, hindering rapid discovery.
• Optimizing synthesis pathways is crucial for unlocking the full potential of these materials.

Purpose of the Study:

• To develop and demonstrate an autonomous experimentation platform for accelerated synthesis and optimization of colloidal nanocrystals.
• To enable simultaneous optimization of nanocrystal properties, such as polydispersity and absorption wavelengths.
• To gain insights into nucleation-growth kinetics for enhanced monodispersity.

Main Methods:

• An integrated platform combining machine learning, automated flow reactors, and inline spectroscopy was employed.
• An ensemble neural network was trained using data from autonomous experiments to predict synthesis outcomes.
• Indium phosphide (InP) nanocrystals were synthesized autonomously with a focus on optimizing polydispersity for target absorption wavelengths.

Main Results:

• The autonomous system successfully synthesized indium phosphide nanocrystals with optimized polydispersity across the visible spectrum.
• Minimal experiments (under 50 within 2 days) were required to achieve desired nanocrystal properties.
• The trained machine learning model accurately predicted synthesis outcomes across the entire parameter space.
• The study provided new insights into nucleation-growth kinetics for achieving highly monodisperse nanocrystals.

Conclusions:

• Autonomous experimentation platforms offer a viable and efficient solution for accelerating the discovery and optimization of colloidal nanocrystals.
• Machine learning-guided synthesis significantly reduces the time and resources needed for materials discovery.
• The developed approach facilitates the tailored synthesis of nanocrystals with specific optoelectronic properties and improved monodispersity.