Universal machine learning aided synthesis approach of two-dimensional perovskites in a typical laboratory

Yilei Wu,Chang-Feng Wang,Ming-Gang Ju,Qiangqiang Jia,Qionghua Zhou,Shuaihua Lu,Xinying Gao,Yi Zhang,Jinlan Wang

DOI: https://doi.org/10.1038/s41467-023-44236-5

IF: 16.6

2024-01-02

Nature Communications

Abstract:Abstract The past decade has witnessed the significant efforts in novel material discovery in the use of data-driven techniques, in particular, machine learning (ML). However, since it needs to consider the precursors, experimental conditions, and availability of reactants, material synthesis is generally much more complex than property and structure prediction, and very few computational predictions are experimentally realized. To solve these challenges, a universal framework that integrates high-throughput experiments, a priori knowledge of chemistry, and ML techniques such as subgroup discovery and support vector machine is proposed to guide the experimental synthesis of materials, which is capable of disclosing structure-property relationship hidden in high-throughput experiments and rapidly screening out materials with high synthesis feasibility from vast chemical space. Through application of our approach to challenging and consequential synthesis problem of 2D silver/bismuth organic-inorganic hybrid perovskites, we have increased the success rate of the synthesis feasibility by a factor of four relative to traditional approaches. This study provides a practical route for solving multidimensional chemical acceleration problems with small dataset from typical laboratory with limited experimental resources available.

multidisciplinary sciences

What problem does this paper attempt to address?

This paper aims to address the complex and multidimensional synthesis challenges in materials science, particularly focusing on the experimental synthesis of two-dimensional silver-bismuth organic-inorganic hybrid peroxides. Traditional trial-and-error methods are inefficient in designing new materials, while machine learning (ML) techniques, although helpful in predicting material properties, still have limited applications in guiding practical synthesis. The researchers proposed a comprehensive framework that integrates high-throughput experiments, chemical prior knowledge, and ML techniques, such as subgroup discovery and support vector machines, to guide experimental synthesis of materials and quickly screen out compounds with high synthesis possibilities. With this approach, they successfully improved the success rate of two-dimensional silver-bismuth peroxides and highlighted the importance of molecular topological structures in the formation of these materials. However, the imbalance of experimental data and the limitations of availability still remain the main challenges. In the paper, the research team conducted high-throughput experiments on 80 amine compounds, developed descriptors for specific problems, and used ML models to improve the success rate of predictions. Ultimately, they successfully synthesized 8 out of 8 predicted two-dimensional silver-bismuth peroxides, validating the effectiveness of the strategy.