Page 112 - Read Online

Mooraj et al. J Mater Inf 2023;3:4 https://dx.doi.org/10.20517/jmi.2022.41 Page 37 of 45

applications. Finally, high-throughput material characterization is highlighted for rapid understanding of
the relationships between composition, microstructure, and material properties. This review article serves as
a guideline for developing workflows that can efficiently discover new high-performance HEAs. To this end,
several research frontiers in the field are put forward:

1. Machine learning (ML) techniques can provide predictions of massive design space, but there currently
exists a shortage of robust training sets for HEA compositions. Further investment in high-throughput
computational techniques that can produce these robust databases, such as CALPHAD, first-principles
calculations, and molecular dynamics simulation, is needed. Once these databases are sufficiently
established, ML techniques can provide highly reliable predictions of the phase constitution for unknown
compositions. They can even predict bulk properties such as yield strength and density.

2. Additive manufacturing provides a means to rapidly produce bulk samples of varying compositions and
microstructures. However, AM materials are prone to defects that can significantly deteriorate performance.
Further studies, including in-situ characterization during 3D printing, are needed to better characterize the
small-scale physics, in-situ alloying chemistry, and macroscale defect formation to reduce the work needed
in preliminary optimization.

3. Data collection and analysis of material characterization techniques need to be further automated to
enable high-throughput characterization of enormous materials libraries without significant time
investments from researchers. Such techniques as phase, composition and microstructure characterization
may need to be carried out in parallel to maximize the efficient use of equipment with overlapping
functionalities, such as SEM with EBSD capabilities. Additionally, data processing automation is critically
needed to rapidly characterize the vast number of compositions that are explored in high-throughput
experiments.

DECLARATIONS
Acknowledgments
This work is based upon research conducted at the Center for High Energy X-ray Sciences (CHESS), which
is supported by the National Science Foundation under award (DMR-1829070). The authors are grateful to
Katharine Shanks at CHESS for her support in data acquisition and analysis at the ID3A beamline.

Authors’ contributions
Writing: Mooraj S
Manuscript supervision and editing: Chen W

Availability of Data and Materials
Not applicable.

Financial support and sponsorship
Chen W acknowledges the support from National Science Foundation (DMR-2004429) and UMass
Amherst Faculty Startup Fund.

Conflicts of interest
All authors declare that there are no conflicts of interest.