Abstract
Understanding the total structure of noble metal nanoclusters has been a major dream of colloid chemists, however, relatively low synthesis efficiency is still the main issue of cluster research. In a recent publication in Chemistry - A European Journal, Shuxin Wang and co-workers reported the development of a machine learning (ML) guided automated high-throughput synthesis platform, which has been crucial for accelerating the synthesis of clusters. The authors ingeniously implemented a cyclical feedback mechanism involving high-throughput synthesis and machine learning. A series of nanoclusters with atomically accurate structures were successfully synthesized, namely, [Au41Cu66(SC6H11)44](SbF6)3, Au40Cu34(4-S-PhF)40, Au40Cu34(4-S-PhF)40, Au18Cu32(3,5-C8H9S)36, [Au6Cu6(SPh)12]n, Cu-NC respectively, are abbreviated as Au41Cu66,
Au40Cu34, Au18Cu32, Au18, Au6Cu6, Cu-NC. This approach not only significantly optimized the synthesis pathway, but also offered profound insights into the influence of various reaction variables, thereby illuminating fresh perspectives for advancing research in the realm of metal nanoclusters.
Keywords
Metal nanocluster, Machine learning, High-throughput synthesis, Precise structures