Lecturer Daniel Puckwood of Kyoto University and Professor Taro Ichisugi of Tokyo Institute of Technology have succeeded in creating guidelines for predicting the arrangement of molecules on a metal substrate using machine learning.
With the demand for further miniaturization and higher integration of electronic devices, "self-organization" of molecules is attracting attention.Molecular self-organization is a phenomenon in which molecules attached to a substrate are attracted to each other by an intermolecular attractive force and aggregate to spontaneously form a minute structure (supramolecular structure).For the development of nanoelectronics, there is a possibility that it can be used to make minute electrical wiring (nanowires) and supramolecular structures that can be used as electronic devices, and research activities are becoming active.However, there is no guideline for spontaneously assembling molecules into the desired structure, and it has been difficult to develop them into applications.
In this research, Lecturer Puckwood, who specializes in mathematical science and theoretical chemistry, conducted joint research with Professor Ichisugi, who specializes in materials science. Utilizing "unsupervised machine learning", we created guidelines for assembling molecules on a substrate as desired.Unsupervised machine learning is the process of comparing various objects with a computer and classifying them into common features.This time, we learned how the chemical characteristics of the molecule and the assembly process of the molecule are related, and summarized the results graphically.Then, by analyzing this figure, guidelines were derived.
From this guideline, it is possible to predict what kind of molecule should be used when forming a linear supramolecular structure that can be used as an electric wiring, for example.This result is expected to lead to the formation of necessary parts for microdevices and accelerate the development of nanoelectronics.
Paper information:[Nature Communications] Materials informatics for self-assembly of functionalized organic precursors on metal surfaces