Appendix I - Machine Learning Accelerated Discovery, Testing, and Characterization of Light-weight Refractory High Entropy Alloys for Space Application

The goal of this proposal is to accelerate the discovery of light-weight refractory high entropy alloys (RHEAs) with enhanced mechanical performances by integrating first-principles material simulations, data-driven machine learning (ML) models, and experimental validation. Advanced metals and alloys with superior mechanical properties at elevated temperatures remain in high demand for aeronautical applications. Recently, a new category of RHEAs has shown great promise as the potential materials for the high-temperature structural application. The RHEAs usually contain four or five elements from the nine elements in Group IV (Ti, Zr, and Hf), Group V (V, Nb, and Ta), and Group VI (Cr, Mo, and W) with the additions of non-refractory elements such as Al, Si, Co, or Ni. The refractory elements allow the RHEAs to function as load-bearing components at a temperature even higher than that for Ni-based superalloys. However, the traditional RHEAs suffer from limited room-temperature ductility and heavyweight, significantly hindering their manufacturing and applications. To design lightweight RHEAs with balanced specific strength and ductility is of urgency and yet poses remarkable challenges in the complex multicomponent system. Firstly, considering the potentially vast number of alloys with the complexity of compositional space, the typically used sequential trial-and-error experiments are daunting. Secondly, the feasibility of the empirical or semi-empirical rules developed for dilute alloys confronts fundamental issues due to unknown physics coming from many-body interaction. The proposed project will employ a computationally aided materials discovery (CAMDIS) approach, and deploy multiscale ML models to data-mine the connection between electronic and thermodynamic properties of multicomponent alloys with their mechanical properties for the rapid design of light-weight RHEAs. Specifically, we will explore the compositional space of Group IV, V, and VI multicomponent alloys using first-principles based density functional theory (DFT) calculations to compute the defect energetics and identify the alloy compositions that process intrinsic ductile propensity. Multiscale neural network ML models will be developed to integrate heterogeneous simulation and experimental data and learn features and representation in the scale of relevance. The parameters of ML models will be optimized to uncover physical-based mechanisms thus enabling decision-making capabilities in precision experimentation planning. The predicted alloys will be synthesized using high-throughput magnetron co-sputtering, and then characterized to obtain structural, chemical, thermal, and mechanical properties. The new experimental data will be supplied into the dataset to train next-generation ML models for accuracy improvement. ML approaches are well suited to this problem because they can (1) operate with an incomplete understanding of the underlying physics in the novel HEA alloys but still take chemical, physical, and thermodynamic parameters to accelerate learning; and (2) find patterns in observed data, making it possible to model relationships that are unavailable in theory. We envision the proposed ML approaches will significantly accelerate the search of multiphase multi-principal component alloys, providing novel alloy systems to tailor the mechanical performance. The created dataset and the developed ML will allow the promising alloy composition for future computational and experimental investigation towards the design of novel RHEAs for aeronautical applications.

This is a mature technology (TRL 7+) — the realistic path in is usually NASA's Technology Transfer Program: licensing an existing NASA patent, or a Space Act Agreement to use NASA facilities/expertise directly. NASA also runs a startup licensing program with no upfront fee for companies formed to commercialize a specific NASA technology.

None of these are guaranteed paths for this specific project — TechPort itself doesn't have an "apply" button. Reaching out to the contact(s) above with a specific question is usually the fastest way to find out what's actually open.