Spectral descriptors for bulk metallic glasses based on the thermodynamics of competing crystalline phases

TL;DR

This paper introduces a spectral descriptor based on the thermodynamics of crystalline phases to predict the glass-forming ability of metallic alloys, enabling accelerated discovery of new metallic glasses.

Contribution

It develops a robust model using spectral decomposition of crystalline phases' features to predict metallic glass formation, based on first-principles calculations and a new confusion heuristic.

Findings

More than 17% of binary alloys could be potential glass formers.

The spectral descriptor effectively predicts glass-forming ability.

The approach overcomes key bottlenecks in discovering new metallic glasses.

Abstract

Metallic glasses have attracted considerable interest in recent years due to their unique combination of superb properties and processability. Predicting bulk metallic glass formers from known parameters remains a challenge and the search for new systems is still performed by trial and error. It has been speculated that some sort of "confusion" during crystallization of the crystalline phases competing with glass formation could play a key role. Here, we propose a heuristic descriptor quantifying confusion and demonstrate its validity by detailed experiments on two well-known glass forming alloy systems. With the insight provided by these results, we develop a robust model for predicting glass formation ability based on the spectral decomposition of geometrical and energetic features of crystalline phases calculated ab-initio in the AFLOW high throughput framework. Our findings indicate that the formation of metallic glass phases could be a much more common phenomenon than currently estimated, with more than 17% of binary alloy systems being potential glass formers. Our approach is capable of pinpointing favorable compositions, overcoming a major bottleneck hindering the discovery of new materials. Hence, it is demonstrated that smart descriptors, based solely on the energetics and structure of competing crystalline phases calculated from first-principles and available in online databases, others the sought-after key for accelerated discovery of novel metallic glasses.