Multicomponent Transition Metal Carbides: Synthesizability and Hardness via MACE ML Potentials

npj Computational Materials Switzerland

Overview

Researchers have developed a machine-learned interatomic potential framework to predict the thermodynamic stability and hardness of multicomponent transition metal carbides. By screening over 1500 equiatomic compositions, the study identifies a new family of stacking-ordered phases with lower formation energies than disordered structures. This approach enables the identification of synthesizable, hard materials for extreme-environment applications.

In Depth

Key Findings

The study identifies a new class of stacking-ordered carbides that are accessible only in multicomponent compositions, specifically those mixing group 4/5 and group 6 metals. These stacking-ordered phases exhibit formation energies significantly lower than those of disordered rocksalt or hexagonal structures. Additionally, the research determines that the group number of constituent metals governs both the stability and hardness of the resulting carbides.

Technical Details

The researchers fine-tuned the MACE machine-learned interatomic potential using approximately 28,000 density functional theory (DFT) calculations. This training set covered the composition space of carbon and transition metals from groups 4–6. The resulting model achieved formation energy errors of ~10 meV/atom for thermodynamically relevant structures while utilizing only 20% of the training data.

The screening process involved over 1500 equiatomic compositions across hcp, hexagonal, and rocksalt prototypes. The methodology combined elasticity-based hardness surrogates with free energy models. Findings indicate that short-range order contributes only a few meV/atom to the free energy, allowing disordered phases to be treated as random solid solutions during high-temperature screening of rocksalt carbides.

Background and Context

Multicomponent transition metal carbides are highly promising for applications in extreme environments. However, the primary challenge has been identifying specific compositions that are simultaneously hard and synthesizable. To validate the machine-learning predictions, the researchers used DFT calculations, which corroborated the existence of the newly discovered stacking-ordered phases.

Significance and Outlook

This research provides a scalable framework for screening multicomponent materials with target properties across a full nine-component space. The alignment between synthesizability predictions at 1500 °C and experimental reports for single-phase and multiphase carbides suggests that the identified stacking-ordered phases should be experimentally accessible. This methodology accelerates the discovery of high-performance carbides for industrial use in extreme conditions.

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.

• Your email and selected fields are used only to deliver the newsletter.
• We never share your information with third parties.
• You can unsubscribe anytime via the link in each email.

See our Privacy Policy for details.

Agree & Continue ▶

Takes about a minute · Unsubscribe anytime