Topological data analysis: Unveiling heat conduction in amorphous germanium

09/09/2024

Persistent homology reveals structural features influencing thermal properties in disordered materials

Dr. Akagi, the corresponding author of this research paper

With no periodic atomic configurations, amorphous materials exhibit unique properties, distinct from those of crystalline materials, that promise novel applications—provided that the physical model bridging the microscopic structure and the macroscopic properties of these materials can be understood.

However, the lack of periodic structure is also a major challenge to elucidating which microscopic structural features impact macroscopic properties such as thermal conduction.

“Conventional methods (e.g., X-ray diffraction) provide convoluted information on the local atomic structures of amorphous materials,” explains Kazuto Akagi, a member of an AIMR research team. “Without a clear idea of the local network of atoms, determining the features that impact thermal properties is practically impossible.”

In a recent article1, Akagi and co-workers overcame this challenge with a new approach that used topological data analysis (TDA) based on persistent homology to analyze the structural features of amorphous germanium (a-Ge), providing a clearer understanding of the atomic network's influence on thermal conductivity.

Using the persistent homology (i.e., a mathematical framework to quantify the given discrete data based the “holes” within) to investigate a-Ge structures obtained both by transmission-electron microscopy and by ab-initio molecular dynamics simulations, the team uncovered the presence of larger atomic rings (five or six vertices with longer bond length) that may play a key mechanistic role in a-Ge heat conduction.

“In a subsequent paper2, our team used the same method to generate the structural descriptors that affect CO2 adsorption in metal-organic frameworks—a type of material completely different from a-Ge—demonstrating our approach’s versatility,” says Akagi. “Further, both results1,2 illustrated how the microstructures reflecting composition and fabrication process affect the macroscopic properties of materials through physical or chemical models.”

(Author: Patrick Han)

References

  1. Wu Y.-J., Akagi K., Goto M., and Xu Y. Topological data analysis of TEM-based structural features affecting the thermal conductivity of amorphous Ge International Journal of Heat and Mass Transfer 221, 125012 (2024). | article
  2. Akagi K., Naito H., Saikawa T., Kotani M. and Yoshikawa H. Linear regression model for metal–organic frameworks with CO2 adsorption based on topological data analysis Scientific Reports 14, 12021 (2024). | article

This research highlight has been approved by the authors of the original article and all information and data contained within has been provided by said authors.