Machine Learning Tools to Predict Hot Injection Syntheses Outcomes for II–VI and IV–VI Quantum Dots

In
order to allow quantum dots with the desired physical and chemical
properties, the fine control and prediction of size during chemical
syntheses is a challenge that must be addressed. In this work, we
applied machine learning algorithms, with information extracted from
scientific papers, to identify the most important variables in the
synthesis of CdSe, CdS, PbS, PbSe, and ZnSe quantum dots. From the
random forest and gradient boosting machine algorithms, the most influential
parameters on the final diameter of the quantum dots were the time
of reaction, temperature, and metal precursors. Our models were applied
to suggest the best reaction parameters for a desired quantum dot
size. This methodology shall contribute to the quantum dot community
to save time and money while reaching the proper material conditions
for their applications.