The rise of electric vehicles (EVs) in recent years brings attention to an important issue: the efficiency of electric motors. A key challenge is energy loss, mainly from a process called magnetic hysteresis. This happens when the magnetic fields in the motor core, made from soft magnetic materials, repeatedly switch directions. High temperatures can cause partial demagnetization, worsening energy loss. The arrangement of magnetic domains—small magnetic regions—plays a crucial role in how well these materials perform in such environments.
Magnetic domains come in various structures. In some materials, they create intricate zig-zag patterns known as maze domains. These formations can behave unpredictably with temperature changes, impacting energy loss. Unfortunately, existing analytical models struggle to accurately capture this complexity due to various influencing factors like metallographic structure and thermal effects.
To overcome these challenges, researchers from the Tokyo University of Science, led by Professor Masato Kotsugi and Dr. Ken Masuzawa, developed a new model called the entropy-feature-eXtended Ginzburg-Landau (eX-GL) model. This model aims to clarify the energy landscape of maze domains in specific materials, like rare-earth iron garnets. “Conventional models oversimplify real materials. Our physics-based AI framework helps explain the complex temperature-dependent reversal of magnetization,” explains Prof. Kotsugi.
The team studied how magnetization behaves under different temperatures by capturing detailed images of magnetic domains. They applied persistent homology, a modern analytical technique, to extract vital information on the structures from these images. This process creates a digital energy landscape, showing how changes in micromagnetic structures relate to the overall magnetization reversal.
They discovered a unique feature, termed PC1, which effectively characterizes the magnetization reversal process. By linking PC1 with various physical parameters, they pinpointed four energy barriers crucial to the dynamics of this reversal.
This analysis shed light on the energy transfers involved in the reversal process. Remarkably, they found that as maze domains’ wall lengths increase, their complexity grows. This complexity is driven by the interaction of entropy and exchange forces, helping clarify the mechanisms behind the reversal of maze domains.
“Our eX-GL model automates the understanding of complex magnetization processes, uncovering mechanisms that traditional methods may overlook,” adds Prof. Kotsugi. “Moreover, since free energy is a universal concept, this model could be applied to other similar systems.”
This research not only enhances our understanding of magnetic domain behavior but also provides a framework for exploring complex energy landscapes in various magnetic systems.
For more details, you can refer to the full study published in Scientific Reports.
