Hydrogen is a clean energy carrier, but its large scale use hinges on safe, high density storage. Solid state hydrogen storage (metal hydrides, complex hydrides, metal–organic frameworks) is promising, yet traditional experimental and computational methods are slow and costly.
The Digital Hydrogen S platform, developed by the team, consolidates data from over 1000 peer reviewed publications. It contains more than 3000 unique material entries and exceeds 254,000 structured records across ten material classes (Mg based, AB₅, high entropy alloys, borohydrides, MOFs, etc.). Unlike earlier single modal databases, Digital Hydrogen S integrates multimodal data: numerical thermodynamic/kinetic parameters, PCT (pressure composition temperature) curves, kinetic curves, and synthesis metadata.
Data analytics reveal persistent gaps. Clustering analysis of V–Ti–Cr, Mg–RE–Ni, and Ti–Fe/Ti–Mn alloys shows that almost no material simultaneously meets the US Department of Energy targets for onboard storage: gravimetric capacity >5.5 wt% (2025) and 6.5 wt% (ultimate), delivery temperature –40 to 85 °C, and delivery pressure 5–12 bar. V–Ti–Cr alloys suffer from low capacity (<4 wt%); Mg–RE–Ni alloys have high capacity (>6 wt%) but dehydrogenation temperatures above 500 K; Ti–Fe/Ti–Mn alloys have too low capacity and too high pressure. The data are also unevenly distributed, causing out of distribution problems for ML models.
ML is now widely applied. Models (random forest, gradient boosting, Gaussian process regression, graph neural networks) predict hydrogen storage capacity, formation enthalpy, equilibrium pressure, and even full PCT curves with R² often >0.95. For example, a GPR model for AB₂ alloys achieved R² = 0.969 for hydride formation enthalpy. High throughput computation combined with ML has screened millions of candidate alloys and MOFs, identifying promising materials such as a V based MOF with 9.0 wt% H₂ at 77 K and 150 bar.
For mechanistic understanding, neural network potentials (NNPs) offer near DFT accuracy at greatly reduced cost. NNPs have simulated H₂ dissociation on Cu surfaces, hydrogen diffusion in Pd(111), and dehydrogenation of MgH₂ slabs, capturing subsurface H₂ formation and long time dynamics. Current limitations include poor handling of long range electrostatics, lack of transferability across materials, and the absence of a general NNP for complete absorption desorption cycles.
The review outlines a forward roadmap: (1) build open access multimodal databases combining numbers, text, spectra, and images; (2) develop multimodal foundation models integrating experiment, computation, and data; (3) implement application driven inverse design (generative models, genetic algorithms); (4) construct generalized NNPs covering full hydrogen storage cycles. These advances will accelerate the transition from empirical to intelligent materials discovery, supporting the clean energy transition.
ENGINEERING Chemical Engineering
Experimental study
Not applicable
Toward intelligent design of solid-state hydrogen storage: trends, challenges, and machine learning insights
15-Mar-2026