ABSTRACT

Hydrogen is a key contributor to meeting clean energy and net-zero goals due to its high energy potential and lack of emissions. With demand projected to reach 150 Mt by 2030, global efforts are focused on developing large-scale storage solutions and exploring natural subsurface hydrogen to support a sustainable energy transition. Geological formations, such as depleted hydrocarbon reservoirs and saline aquifers, offer promising solutions for long-term storage, while naturally occurring hydrogen provides a potential continuous supply. However, a limited understanding of hydrogen interactions with subsurface formations poses significant challenges. These interactions can activate complex rock-fluid physical interactions and geochemical and microbial processes that affect hydrogen mobility, retention, purity, and seal integrity.

To address these challenges and gain insights into the intricate interaction between hydrogen and geological materials, extensive simulations ranging from the molecular to the biogeochemical scales are carried out. A series of Monte-Carlo and Molecular dynamics simulations are performed to analyze the distribution of hydrogen and its dynamic properties in representative geological minerals’ pores, including quartz, calcite, kaolinite, illite, and montmorillonite, under non-reactive conditions. These analyses consider coexisting fluids including hydrogen, carbon dioxide, methane, heavy alkanes, and brine with different salinity under diverse reservoir thermodynamic considerations. Key interaction properties, including adsorption, Henry coefficients, adsorption energy, fluid distribution, wettability, intercalation, diffusion, and radial distribution function, are thoroughly examined.

Building upon the non-reactive molecular-scale insights, significant biogeochemical analyses are conducted to understand the effects of biogeochemical reaction kinetics for the determination of hydrogen loss, mineral dissolution-precipitation, pH variations, mineral saturation changes, and growth of bacteria under an abiotic environment with specified geological conditions. To enhance predictive capabilities and support future assessments, machine learning, particularly deep learning, algorithms are trained on the datasets to develop models for automated quantification of hydrogen interaction properties.

The outcome of this integrated approach reveals that thermodynamic conditions, mineralogical composition, and fluid chemistry critically influence hydrogen’s physical interactions and transport behavior in the subsurface. These factors govern hydrogen retention capacity and mobility across different geological minerals. Furthermore, the study confirms that hydrogen intercalation into hydrated clay minerals is thermodynamically unfavorable at various reservoir conditions. Moreover, geochemical and biogeochemical reactivity highly contributes to hydrogen loss, purity degradation, microbial growth, and mineral dissolution–precipitation dynamics, all of which can affect the long-term stability and integrity of hydrogen-bearing systems.

Together, these multiphysical and reactive interactions provide a scientific foundation for evaluating the feasibility of underground hydrogen storage and advancing the understanding of natural hydrogen occurrences. This knowledge is essential for developing safe, efficient, and sustainable subsurface hydrogen utilization strategies.