Equations for "Water-Holding Materials" derived from Machine-Learning

November 11, 2025

Establishing a Method for designing Polymer Materials Enabling for Contribute to One Health

Researchers at the RIKEN CSRS used a data-driven approach to develop a methodology for generating equations that contributes to designing of new polymer materials. This achievement is expected to accelerate the creation of new materials through the design of hydrogels—"water-holding" materials used in products such as disposable diapers—and, in turn, contribute to the realization of One Health.

The complexity of the three-dimensional molecular networks of polymer materials makes it extremely difficult to predict their physical properties using physical equations alone. In this study, the researchers focused on two crucial functions of hydrogels: swelling and ligand interactions. The team acquired data from analytical methods such as time-domain nuclear magnetic resonance (TD-NMR) and multi-resonance nuclear magnetic resonance spectroscopy, as well as RDKit molecular descriptor data, among others, and applied a machine learning technique known as symbolic regression. This led to the successful derivation of a clear design equation that incorporates the factors related to these functions. The equation incorporated not only water molecules but also an important newly identified factor: the mobility of polymer chains. This opens the way for rapid and rational design that simultaneously fulfills multiple hydrogel functions, based on extensive data. Furthermore, the findings from this study hold the potential to accelerate the design of not only hydrogels but also other polymer materials (e.g., heat-resistant polymers, lightweight yet strong structural materials, and electrically conductive next-generation semiconductors), enabling these materials to fully maximize their properties.

 

Original article

ACS Materials Letters　doi: 10.1021/acsmaterialslett.5c00957

M. Okada, W. Zhu, Y. Amamoto, J. Kikuchi,

"Data-driven formulation based on integrated symbolic regression of hydrogel swelling and molecular interactions".

Contact

Jun Kikuchi
Team Director
Environmental Metabolic Analysis Research Team