Towards Pareto optimal high entropy hydrides via data-driven materials discovery

Abstract

The ability to rapidly screen material performance in the vast space of compositionally complex (high entropy) alloys is of critical importance to efficiently identify optimal hydride candidates for various use cases. Given the prohibitive complexity of first principles simulations and large-scale sampling required to rigorously predict hydrogen equilibrium in these systems, we turn to compositional machine learning models as the most feasible approach to screen on the order of 10,000s of candidate equimolar alloys. We critically show that machine learning models can predict hydride thermodynamics and capacities with reasonable accuracy and that explainability analyses capture the competing trade-offs that arise from feature interdependence. We can therefore elucidate the multi-dimensional Pareto optimal set of materials (i.e., where two or more competing objective properties can't be simultaneously improved) and provide rapid and efficient down-selection of the highest priority candidates for more time-consuming density functional theory investigations and experimental validation. Selected target materials were experimentally synthesized, characterized, and tested amongst an international collaboration group to validate the proposed novel hydrides. Additional top-predicted candidates are suggested to the community for future synthesis efforts, and we conclude with an outlook on improving the current approach for the next generation of computational HEA hydride discovery efforts.

Large Destabilization of (TiVNb)-Based Hydrides via (Al, Mo) Addition: Insights from Experiments and Data-Driven Models

“…15 Consequently, there is now a push to rationally design HEAs for hydrogen storage that can maximize capacity while minimizing desorption enthalpy (and required desorption temperature). 8,16 A number of HEAs with a bcc structure possess outstanding volumetric hydrogen capacities, as reported for TiVZrNbHf, 9 TiVZrNbHf 0.5 17 TiZrNbHfTa, 18 and Ti 4 V 3 NbCr 2 . 19 The inclusion of lightweight elements (such as Mg, Al) to improve gravimetric capacities has also been explored.…”

“…For this study, we released v0.0.5 of the ML-ready HydPARK database, which augments v0.0.4 with the comprehensive AB 2 metal hydride thermodynamic data recently presented in ref 51. Gradient boosting tree models were trained to predict the enthalpy of the hydriding reaction, ΔH, using the identical strategy described in ref 16, with scripts to reproduce them provided here (https://github.com/ mwitman1/HEAhydrideMLv2). 4).…”

Towards Pareto optimal high entropy hydrides via data-driven materials discovery

“…Training data. Our models are trained on the experimental metal-hydride thermodynamic data contained in v0.0.4 of the ML-HydPARK database, 40 which includes recent literature data on high entropy alloy hydrides 26,30,32,35,37,41 and metal hydrides for compression. 15,47 The thermodynamic quantities available in the database are hydrogen per metal saturation capacity…”

“…To address some of these opportunities and further drive discovery with data-driven methods, we rst improve upon our previously developed hydride thermodynamic models. v0.0.4 of the ML-HydPARK training database 40 was augmented to contain additional HEA hydride thermodynamic properties from literature studies when the necessary pressure-composition-temperature (PCT) measurements were performed to extract the enthalpy and entropy of hydrogen desorption (DH and DS) and saturation capacity (H/ M). 26,30,32,35,37,41 Thermodynamic properties from metal hydrides investigated for hydrogen compression were also added to the training data.…”

Simple local environment descriptors for accurate prediction of hydrogen absorption and migration in metal alloys

“…This class of materials is especially attractive due to immense opportunities for tuning hydrogen solubility and preferential absorption sites by altering chemical composition and crystal structure. [31][32][33][34] While some recently discovered HEAs demonstrate excellent hydrogen resistances 35,36 , others can be used for hydrogen storage applications. 33,37 It was shown that the mechanical properties of HEAs strongly depend on hydrogen concentration and the distribution of hydrogen atoms within a metal matrix.…”