Digital hydrogen platform (DigHyd): a rigorously curated database for hydrogen storage materials empowered by AI-assisted literature mining
6 affiliations
- 1Advanced Institute for Materials Research (WPI-AIMR), Tohoku University, Sendai, 980-8577, Japan
- 2Unprecedented-scale Data Analytics Center, Tohoku University, Sendai, 980-8578, Japan
- 3Frontier Research Institute for Interdisciplinary Sciences (FRIS), Tohoku University, Sendai, 980-8577, Japan
- 4Department of Materials Engineering, The University of Tokyo, Tokyo, 113-8656, Japan
- 5Institute for Materials Research, Tohoku University, Sendai, 980-8577, Japan
- 6Institute of Fluid Science, Tohoku University, Sendai, 980-8577, Japan
Abstract
Solid-state hydrogen storage materials are promising candidates for safe and compact hydrogen storage; however, data-driven discovery in this field remains limited by the availability of large-scale, well-curated datasets. Here, we present the Digital Hydrogen Platform (DigHyd: www.dighyd.org ), a rigorously curated database comprising > 4,000 experimental literature sources and > 30,000 data entries on hydrogen storage materials, constructed through AI-assisted literature mining combined with human-in-the-loop validation. In addition to gravimetric hydrogen storage density (w), DigHyd also covers thermodynamic parameters, specifically the enthalpy (∆H) and entropy (∆S) changes associated with hydrogenation reactions, primarily defined as M+1/2H2rightleftharpoons MH . These parameters were obtained by manually analyzing multi-temperature pressure-composition-temperature (PCT) data using van’t Hoff analysis. By focusing on ∆H and ∆S rather than fixing equilibrium pressure at a single temperature, DigHyd enables flexible evaluation of equilibrium behavior under application-specific operating conditions. Statistical analyses reveal distinct distributions of thermodynamic parameters across material classes, together with broad compositional variability within representative hydride systems. As a representative application, we performed composition-based symbolic-regression modeling using a curated single-phase or near-single-phase subset, which achieved predictive performance comparable to state-of-the-art black-box models while providing compact, physically interpretable relationships for w, Peq,RT (equilibrium pressure at room temperature), ∆H, or ∆S. The resulting descriptor map identifies recurring physical factors, including host atomic mass, lattice geometry, elastic stiffness, metal filling factor, and electronegativity-derived descriptor correlations, which rationalize the trade-off between w and Peq,RT . Overall, DigHyd provides a rigorously curated thermodynamic dataset that serves as a reliable basis for data-driven analyses of hydrogen storage materials and supports systematic exploration of structure–property relationships.
Related
- Machine Learning-Empowered Electromagnetic Wave Absorbing Materials: From Forward Prediction to Generative Inverse Design
- AGAPI-Agents: An Open-Access Agentic AI Platform for Accelerated Materials Design on AtomGPT.org
- Hydrogen-based materials for energy storage and conversion
- Artificial intelligence and machine learning in composite materials: A comprehensive literature review and bibliometric analysis
- Advances in catalytic hydrogen storage materials: progress, breakthroughs, and challenges toward scalable applications.
BibTeX
@article{jang2026digital,
title = {{Digital hydrogen platform (DigHyd): a rigorously curated database for hydrogen storage materials empowered by AI-assisted literature mining}},
author = {Seong‐Hoon Jang and Di ZHANG and Xue Jia and Hung Ba Tran and Linda Zhang and Ryuhei Sato and Yusuke Hashimoto and Toyoto Sato and K. Konno and Shin‐ichi Orimo and Hao Li},
journal = {Applied Physics A},
year = {2026},
month = sep,
doi = {10.1007/s00339-026-09903-6},
url = {https://doi.org/10.1007/s00339-026-09903-6},
note = {CC BY},
}
RIS
TY - JOUR
TI - Digital hydrogen platform (DigHyd): a rigorously curated database for hydrogen storage materials empowered by AI-assisted literature mining
AU - Seong‐Hoon Jang
AU - Di ZHANG
AU - Xue Jia
AU - Hung Ba Tran
AU - Linda Zhang
AU - Ryuhei Sato
AU - Yusuke Hashimoto
AU - Toyoto Sato
AU - K. Konno
AU - Shin‐ichi Orimo
AU - Hao Li
PY - 2026
DA - 2026/09/04
JO - Applied Physics A
DO - 10.1007/s00339-026-09903-6
UR - https://doi.org/10.1007/s00339-026-09903-6
AB - Solid-state hydrogen storage materials are promising candidates for safe and compact hydrogen storage; however, data-driven discovery in this field remains limited by the availability of large-scale, well-curated datasets. Here, we present the Digital Hydrogen Platform (DigHyd: www.dighyd.org ), a rigorously curated database comprising > 4,000 experimental literature sources and > 30,000 data entries on hydrogen storage materials, constructed through AI-assisted literature mining combined with human-in-the-loop validation. In addition to gravimetric hydrogen storage density (w), DigHyd also covers thermodynamic parameters, specifically the enthalpy (∆H) and entropy (∆S) changes associated with hydrogenation reactions, primarily defined as M+1/2H_2rightleftharpoons MH . These parameters were obtained by manually analyzing multi-temperature pressure-composition-temperature (PCT) data using van’t Hoff analysis. By focusing on ∆H and ∆S rather than fixing equilibrium pressure at a single temperature, DigHyd enables flexible evaluation of equilibrium behavior under application-specific operating conditions. Statistical analyses reveal distinct distributions of thermodynamic parameters across material classes, together with broad compositional variability within representative hydride systems. As a representative application, we performed composition-based symbolic-regression modeling using a curated single-phase or near-single-phase subset, which achieved predictive performance comparable to state-of-the-art black-box models while providing compact, physically interpretable relationships for w, P_eq,RT (equilibrium pressure at room temperature), ∆H, or ∆S. The resulting descriptor map identifies recurring physical factors, including host atomic mass, lattice geometry, elastic stiffness, metal filling factor, and electronegativity-derived descriptor correlations, which rationalize the trade-off between w and P_eq,RT . Overall, DigHyd provides a rigorously curated thermodynamic dataset that serves as a reliable basis for data-driven analyses of hydrogen storage materials and supports systematic exploration of structure–property relationships.
SN - 0947-8396
SN - 1432-0630
ER -