ADVANCE: AI-Driven Valuation and Novel Creation of Energy- efficient, Low-Dimensional, High-Entropy Alloys with Superior Catalytic Performance for Hydrogen Production

Coordinator: Zheng LIU

CINTRA – CNRS-International-NTU-Thales Research Alliance

Keywords: High Entropy Alloys (HEA), Artificial Intelligence (AI), 2D and 0D Materials, Material synthesis, material characterization, DFT and CALPHAD simulations, high- performance catalysis, Hydrogen production, Green Deal, materials by design.

The escalating demand for green energy solutions underscores the imperative for efficient hydrogen production methods. Water splitting stands out as a promising technique in this context. While noble metals like platinum and iridium exhibit commendable catalytic prowess, their limited availability and prohibitive costs make them less viable for widespread applications. This proposal spotlights low-dimensional high-entropy alloys (LD-HEAs) as a groundbreaking alternative, unlocking a vast compositional realm to sidestep the constraints of traditional catalysts..

HEAs, distinguished by their multi-element composition in near-equimolar ratios, outshine conventional alloys in terms of performance and stability. The inherent low-dimensionality of LD-HEAs amplifies surface area, optimizes charge transport, and accentuates quantum confinement-traits indispensable for catalysis. Yet, navigating the expansive compositional spectrum of LD-HEAs is not without challenges, encompassing the intricacies of determining optimal compositions, grappling with processing complexities, and the nascent understanding of their nuanced behavior. To navigate these challenges, this research marries artificial intelligence (AI) techniques with hands-on experimental and computational methodologies. The proposal’s objectives are threefold:

1. Use AI for the design and discovery of LD-HEAs, leveraging machine learning (ML) to predict catalytic performance and design algorithms for efficient identification.

2. Fabricate and characterize the structure of AI-recommended LD-HEAs.

3. Assess and study of the electrocatalytic performance of these LD-HEAs.

Accordingly, the proposal is structured into three technical working packages (WPs):

  • WP1 lays the groundwork by curating a comprehensive database for HEAs. This data will be pivotal for crafting an ML model, which, when combined with efficient search algorithms, is designed to identify LD-HEAs with superior catalytic potential.
  • WP2 focuses on the theoretical validation of ML-suggested HEA candidates, assessing their formability and stability. Those that meet the criteria will be fabricated and subjected to in-depth structural analysis using state-of-the-art instrumentation.
  • WP3 pivots to catalytic study, examining the electrocatalytic performance of the produced LD-HEAs using various electrochemical methods, while also elucidating the underlying catalytic mechanism. Insights gleaned from WP2 and WP3 will be channeled back into WP1, finetuning the ML model’s predictive prowess and guiding the pursuit of premier LD-HEAs.

The consortium is a collaboration of three elite laboratories with complementary expertise: i) CINTRA specializes in the fabrication and catalytic evaluation of LD-HEAs; ii) ICMCB excels in computational modelling and AI-driven HEA design; and iii) CEA IRIG MEM offers unparalleled expertise in compositional and structural characterization.

This project envisions a transformative approach to material design, where AI-driven insights guide the development of catalysts. By seamlessly integrating AI with experimental and computational methods, we aim to spotlight LD-HEAs as the next-generation catalysts for water splitting. The implications of this research are profound, with the potential to reshape the catalysis landscape and accelerate the transition to renewable energy solutions. This synergy between AI and HEA design could set a new benchmark in advanced catalyst materials, steering us towards a sustainable future.