Machine learning for the development of new materials for a magnetic tunnel junction

Area of Science:

• Materials Science
• Computational Materials Design
• Alloy Development

Background:

• Increasing the number of constituent elements in alloys and compounds is crucial for enhancing material properties.
• Ternary alloys like Neodymium-Iron-Boron (NdFeB) and Cobalt-Iron-Boron (CoFeB) are vital in permanent magnets and spintronics.
• The complexity of multi-elemental alloys necessitates advanced discovery methods beyond manual exploration.

Purpose of the Study:

• To propose a standardized process for developing novel multi-elemental materials.
• To leverage machine learning for predicting promising alloy candidates.
• To outline requirements for improving the materials development workflow.

Main Methods:

• Utilizing machine learning algorithms to predict candidate multi-elemental alloys.
• Employing ab initio calculations for screening predicted candidates.
• Integrating quantum annealing to enhance machine learning adoption in materials development.

Main Results:

• Identification of a systematic approach for accelerated materials discovery.
• Demonstration of machine learning's capability in predicting complex alloy compositions.
• Highlighting the synergy between computational methods and experimental validation.

Conclusions:

• A standardized process integrating machine learning, ab initio calculations, and quantum annealing can significantly advance materials development.
• This computational approach is essential for exploring the vast compositional space of multi-elemental alloys.
• Further refinement of this process will broaden the applicability of machine learning in creating next-generation materials.