Robot-assisted colloidal nanocrystalline digital manufacturing platform

Recently, the team of Yu Xuefeng and Zhao Haitao of the Institute of Advanced Materials Science and Engineering, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, together with the University of Science and Technology of China and the Australian National University, published a research paper entitled A Robotic Platform for Synthesis of Colloidal Nanocrystals in Nature Synthesis. This work is the first time to build a robot-assisted colloidal nanocrystalline digital manufacturing platform by integrating data mining, data-driven automated synthesis, machine learning, and reverse design, which is expected to liberate researchers from traditional trial-and-error experiments and labor-intensive characterization, and realize the digital preparation of colloidal nanocrystals.

Nanocrystals have great application potential in energy, optics, photochemistry, electrochemistry, optoelectronics and biomedicine. The physical and chemical properties of nanocrystals are closely related to their morphology and size, while traditional trial-and-error experiments and intensive characterization require a lot of time and effort, which restricts the research and development of nanocrystals. To this end, the research team integrates data-driven automatic synthesis, robot-assisted controllable synthesis, morphology-oriented reverse design and other technologies to build a robot-assisted colloidal nanocrystalline digital intelligent manufacturing platform, so as to break through the limitations of current nanocrystalline controllable synthesis research. Among them, the automation platform consists of three modules: automated synthesis module, automated characterization module and collaborative robot, each module contains several sub-modules, with high-throughput synthesis, sample storage, in situ optics, spectroscopic characterization and other functions (Figure 1). The research team took two typical colloidal nanocrystals as research examples, one is gold nanorods, which are currently widely studied in the field of biosensing detection, and the other is perovskite nanocrystals with application potential in the field of new energy and optical detection.

Figure 1.Robot-assisted colloidal nanocrystalline digital manufacturing automation platform

To automate synthesis, researchers data mine the literature to provide initial selection of key synthesis parameters. For gold nanorods, 1300 reported literature on gold nanorod synthesis were data mined, and their key parameters were sorted horizontally to obtain robot execution parameters, orthogonal experiments and high-throughput experiments were designed to verify that important parameters for the morphology regulation of gold nanorods were obtained. For double perovskite, through data mining of other perovskite-related literature, 48 solvents and 61 surfactants that can be potentially used to adjust the size and morphology of double perovskites were screened, and the screening of solvents and surfactants was quickly realized by combining high-throughput in situ synthesis and characterization.

Furthermore, through the design of univariate, two-factor and three-factor experiments, high-throughput synthesis, in situ optical characterization (RGB value acquisition), in situ spectroscopy and ectopic characterization (transmission electron microscopy, scanning electron microscopy) were carried out to obtain large sample data and small sample data, combined with machine learning, the relationship model between synthesis key parameters (structure directing agent) and absorption spectra and the relationship between absorption spectrum and nanocrystalline size were obtained (Figure 2). By accumulating data samples, the model is further improved. In addition, based on the large-sample color information (RGB) of the two materials, it is also possible to build a model of the relationship between color information and nanocrystalline size. This model can be used as another indicator for rapid identification of nanocrystal size. Thanks to the construction of these models, the input of target product size information can feed back the synthesis key parameters (structural directers), enabling efficient reverse design and synthesis of nanocrystals (Figure 3). Therefore, this work is promising in the field of nanocrystalline synthesis by data-driven robots.

Figure 2.Machine learning model of controllable synthesis in situ representation and verification of ectopic representation

Figure 3.Experimental database, machine learning model and reverse synthesis of nanocrystals

Shenzhen Advanced Institute is the first communication unit. The research work has been supported by the National Natural Science Foundation of China, the Natural Science Foundation of Guangdong Province, the Natural Science Foundation of Shenzhen Municipality, and the Shenzhen-Hong Kong-Macao Science and Technology Program. (Source: Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences)

Related paper information:

Special statement: This article is reproduced only for the need to disseminate information, and does not mean to represent the views of this website or confirm the authenticity of its content; If other media, websites or individuals reprint and use from this website, they must retain the “source” indicated on this website and bear their own legal responsibilities such as copyright; If the author does not wish to be reprinted or contact the reprint fee, please contact us.

Source link

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button