Tokyo University of Science Discovers New Class of Ferromagnetic Quasicrystals with Machine Learning, AI Accelerates Magnetic Material Development

July 12, 2026

Digital Journal (Tokyo University of Science) Japan

Overview

A research team at Tokyo University of Science has successfully developed a new class of ferromagnetic quasicrystals by applying machine learning. This AI-guided approach enabled the prediction of promising alloy compositions for stable ferromagnetic quasicrystals and systematic investigation of their structural and magnetic properties, without relying on traditional trial-and-error methods. This breakthrough significantly accelerates the development of advanced magnetic materials for next-generation devices.

In Depth

Key Findings

A research team at Tokyo University of Science has made a groundbreaking discovery and development: a new class of ferromagnetic quasicrystals, enabled by the innovative application of machine learning (ML) technology. This AI-guided approach efficiently predicted promising alloy compositions for stable ferromagnetic quasicrystals and systematically explored their complex structural and magnetic properties, without resorting to traditional trial-and-error methods. This breakthrough holds the potential to dramatically accelerate the design and discovery processes for advanced magnetic materials.

Technical / Clinical Details

The research team began by training ML models with extensive data on the complex atomic arrangements of quasicrystals and their resulting magnetic properties. The AI models extracted subtle structural features and electronic state patterns common to quasicrystals exhibiting ferromagnetism from this data. Based on this learning, the AI proposed candidate alloy compositions, previously unknown, that had a high probability of forming stable ferromagnetic quasicrystals. In traditional materials science, the exploration of quasicrystals has largely depended on random trial and error, making it extremely difficult to find quasicrystals with specific functionalities like ferromagnetism. The AI’s predictive capabilities significantly narrowed down this search space, allowing researchers to focus on promising candidates. Indeed, quasicrystals with compositions predicted by the AI were synthesized, and their ferromagnetic properties and stability were confirmed through X-ray diffraction and magnetic property measurements. This demonstrated the effectiveness of AI-proposed alloy compositions for designing practical ferromagnetic quasicrystals.

Background & Context

Quasicrystals are materials with unique structures where atoms are arranged regularly but non-periodically, and their distinct electronic states and surface properties have led to research into applications such as thermoelectric materials, catalysts, and low-friction coatings. However, ferromagnetic quasicrystals are exceedingly rare, and their exploration has been extremely challenging. Ferromagnetic quasicrystals could possess unique properties not found in conventional crystalline magnetic materials, potentially opening new avenues for the development of next-generation magnetic memory, high-sensitivity sensors, and spintronic devices. This AI-enabled discovery addresses a significant bottleneck in this exploration, greatly expanding the frontier of quasicrystalline materials science.

Strategic Significance & Outlook

This ML-driven method for discovering ferromagnetic quasicrystals is a versatile approach that can be applied to the exploration of other functional quasicrystals and a broader range of new magnetic materials. The research team aims to further refine this AI framework, enhancing its predictive accuracy and discovery efficiency. Specifically, they plan to advance AI-driven design to further optimize specific magnetic properties of ferromagnetic quasicrystals, such as saturation magnetization, Curie temperature, and magnetic anisotropy. The widespread adoption of this technology is expected to dramatically shorten the materials development cycle, enabling the faster market introduction of groundbreaking magnetic materials that will significantly impact the next-generation electronics and information technology sectors.

Source: https://www.digitaljournal.com/article/machine-learning-unlocks-a-new-class-of-magnetic-materials/

Get our weekly technology intelligence — free

Receive an infographic that lets you judge at a glance whether each field’s analysis report is worth reading.

Subscribe Free — Weekly Tech Intelligence

By subscribing, you’ll receive Troy-Technical’s weekly technology intelligence newsletter.