Influence of hydrogen sulfide on gas-water interface in underground hydrogen storage: A molecular dynamics study

The problem

The Hydrogen Council (2021) projects that global net-zero emissions by 2050 will require about 660 million tonnes of hydrogen. Storage capacity for this output is inadequate (Perera, 2023; Thiyagarajan et al., 2022; Sekar et al., 2023). Depleted hydrocarbon reservoirs are one candidate for underground storage, because their geology and caprock integrity are already characterised and surface facilities are in place (Perera, 2023; Zivar et al., 2021; Muhammed et al., 2023).

One problem for these stores is hydrogen sulfide (H2S), which in hydrocarbon reservoirs causes souring, corrosion, and stress-cracking of steel (Bai et al., 2014; Dopffel et al., 2021). Microorganisms that reduce sulfate can make it (Muyzer and Stams, 2008). In a hydrogen store, hydrogen is presumed to be the main electron donor for these microbes, which include Desulfovibrio vulgaris and Desulfotomaculum salinum (Tao et al., 2014).

H2S can also change the interfacial tension (IFT) between gas and the water left in a reservoir's pores. IFT affects how gas and water separate in pore spaces and how stored gas is distributed (Guo et al., 2020; Delshad et al., 1986; Isfehani et al., 2023). High-pressure, high-temperature experiments with H2S are hazardous because it is toxic and corrosive (Fauve et al., 2017). Shah et al. (2008) measured H2S-water IFT up to 393 K and 15 MPa. However, IFT for water against H2S mixed with hydrogen, or with methane used as a cushion gas, has received little study. In this study, we simulate these mixtures.

What we did

We used molecular dynamics, which tracks every molecule from the forces between them. Each model held 2048 water molecules between about 1024 gas molecules, in a box 10 nm long. All runs were at 14.5 MPa and 343 K, which we convert to roughly 140 times atmospheric pressure and about 70 °C. These are typical storage conditions (Hassanpouryouzband et al., 2020). We mixed H2S with hydrogen, with methane, and with equal amounts of both. H2S made up 5% to 80% of the gas, at six concentrations. The upper end of this range is unlikely at reservoir scale but may arise locally where microbial activity is high. We also modelled each pure gas against water.

Simulated densities of the pure fluids agreed with NIST reference values, for example 0.67 against 0.70 g/cm³ for H2S. Simulated IFTs against water were 70.3 mN/m for hydrogen, 52.4 for methane, and 6.9 for H2S. Measured values are about 65.0 (Aslannezhad et al., 2023), 53.7 to 55.9 (Ren et al., 2000), and 7.4 mN/m (Shah et al., 2008). By our calculation, the largest difference is about 8%, for hydrogen.

What we found

In our simulations, IFT fell as H2S concentration rose in every mixture. At 5% H2S, IFT fell by about 12% with hydrogen, 9% in the hydrogen-methane mix, and 6% with methane. We calculate that pure H2S against water has an IFT about a tenth of that for pure hydrogen (6.9 against 70.3 mN/m). However, above 80% H2S, further increases in concentration made little difference to IFT.

Gas density profiles show that H2S concentrates in the interface region in every mixture. We attribute this to its absorption into the water, consistent with earlier simulations (Ofori et al., 2022; Dehaghani et al., 2023). The accumulated H2S contributes to the fall in IFT. Hydrogen is depleted near the interface, as van Rooijen et al. (2023) also found, whereas methane density peaks there. This indicates that methane adsorbs at the interface, consistent with earlier studies (Miqueu et al., 2011; Sakamaki et al., 2011; Yang et al., 2017). At 5% H2S, the interface held more H2S than hydrogen in the hydrogen mixture, but more methane than H2S in the methane mixture. This is because methane has a greater affinity for the interface and lowers the accumulation of H2S there. We suggest that this difference explains the larger effect of H2S alongside hydrogen, and that methane adsorption partly counteracts the effect of H2S.

The plateau in IFT above 80% coincides with a limit on H2S at the interface. We define Region II as the layer where gas molecules are in direct contact with water. In both two-gas mixtures, the density of H2S in Region II rose with H2S concentration and levelled off above 80%. In the hydrogen-methane mix, the H2S profiles in Region II at 80% and 100% are nearly indistinguishable. This suggests that Region II saturates with H2S at about 80%.

Water molecules at the interface tend to align parallel to it. This preference is stronger against hydrogen or methane than against H2S. As H2S concentration increased to 80%, the orientation of interfacial water became more random, with little change above that. This may be because H2S makes the boundary between the phases less sharp and the interface partially miscible.

The 9% fall in the hydrogen-methane mix indicates that low H2S concentrations affect IFT even where methane is present as a cushion gas. This is particularly relevant to stores that use only small volumes of cushion gas. These stores have less methane to offset H2S.

Limitations

Our simulations used a single pressure and temperature. Changes in either during reservoir operations may also affect IFT. Other impurities in hydrogen stores, such as carbon dioxide and nitrogen, were not included. We recommend experiments to confirm these trends and to refine the molecular models. The dependence of IFT on gas mixture should also be considered in upscaled pore- scale and continuum models, in particular its relation to relative permeability, sweep efficiency, and mixing.

Glossary

Hydrogen sulfide (H2S) is a toxic, corrosive gas.

Souring is the contamination of reservoir gas and water with H2S.

A cushion gas is gas left permanently in a storage reservoir to maintain its pressure. Methane is one option.

An electron donor is the substance a microorganism oxidises for energy, here hydrogen.

Interfacial tension (IFT) is the energy needed to enlarge the boundary between two fluids, per unit area of boundary.

mN/m (millinewtons per metre) is the unit of IFT.

MPa (megapascal) is a unit of pressure. K (kelvin) is a temperature scale that starts at absolute zero.

nm (nanometre) is a billionth of a metre.

Molecular dynamics is a simulation method that tracks the motion of individual molecules from the forces between them.

NIST is the National Institute of Standards and Technology, which publishes reference properties of fluids.

g/cm³ (grams per cubic centimetre) is a unit of density.

Adsorption is the accumulation of molecules on a surface. Absorption is their uptake into the bulk of the other phase.

The interface region is the layer in which water density changes from 10% to 90% of its bulk value.

Region II is the part of the interface region where gas molecules are in direct contact with water.

Miscible fluids mix completely. A partially miscible interface is less sharply defined.

Relative permeability describes how easily one fluid flows through rock when another fluid also occupies the pores. Sweep efficiency is the fraction of a reservoir reached by an injected fluid.