A First Look at Hydrogen Generation in an Ultramafic Rock with Micro-CT and SEM-BEX
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Condensed Matter. The day in condensed matter: materials, phases and the measurements under them.
Kai: I'm Kai.
Mira: And I'm Mira, here with Lev.
Kai: We're looking at the paper: "A First Look at Hydrogen Generation in an Ultramafic Rock with Micro-CT and SEM-BEX".
Mira: Here's the gist of the paper. Natural hydrogen (H2) generated by water–rock interaction in ultramafic rocks is increasingly recognised as a potentially important primary energy resource, but the pore-scale processes that control the initiation and early transport of a free gas phase remain poorly constrained The time-resolved 4D imaging captures the transition from a fully liquid-saturated pore space to the appearance and growth of a distinct gas phase after an ∼8 h induction period Observations provide qualitative pore scale evidence relevant to natural hydrogen generation in the subsurface Experimental Setup and Methodology The study utilized an in situ X-ray micro-tomography experiment where an ultramafic granular pack saturated with KI-doped brine was heated to 100◦C under specific pressure conditions The setup involved confining the sample within a Hassler-type sleeve under an external confining pressure of 10 bar while maintaining a pore pressure of 4 bar KI was used as an X-ray contrast agent to enhance the attenuation of the fluid phase, maximizing the contrast between rock grains, brine, and gas Time-resolved 3D images were acquired during the heating phase at a spatial resolution of 6.277 µm to capture inter-granular porosity and nucleation events Observation of Gas Phase Evolution The time series of segmented gas phase shows a three-stage evolution: (i) no resolvable gas from 0–8 h, (ii) rapid gas generation and connectivity growth between 8 and 23 h, and (iii) slower, diffusion-limited growth between 23 and 30 h Gas bubble nucleation was first observed at 8 hours near the top of the sample column The bulk gas saturation increased to Sg ≈ 0.39 by 23 hours, corresponding to a rate of change of ∆Sg/∆t ≈ 2.6 × 10−2 h−1 between those times Mineralogical and Chemical Evidence SEM-BEX imaging revealed textural alteration and local changes in elemental signals between reacted and unreacted material Post-reaction SEM imaging showed more irregular, roughened surfaces with clusters of elongate to platy crystals protruding from the grain surface BEX spot analyses indicated that post-reaction spots showed a broader spread with locally elevated Al and Ca, suggesting partial dissolution of Mg-rich primary phases Implications for Natural Hydrogen Systems The results suggest that substantial water–rock reaction and H2 production can occur before a free gas phase becomes detectable at the pore scale The delay before the first visible bubbles is consistent with a timescale of 6–20 hours, which brackets the observed induction period Once nucleation thresholds are crossed, gas connectivity can be established over relatively short timescales compared to the overall reaction history This behaviour suggests that zones of intense reaction may remain “invisible” in terms of free gas for extended periods The proof-of-concept experiment demonstrates that time-resolved 4D micro-CT can monitor the emergence and early evolution of a gas phase generated by water–rock reaction in an ultramafic granular pack The three-stage evolution is consistent with delayed supersaturation, nucleation and subsequent diffusion-limited growth within the imaged volume The study provides qualitative constraints on pore-scale behaviour regarding gas nucleation and connectivity during low-temperature alteration Future work should extend this approach to a broader range of mineralogies, pressures, temperatures and flow regimes REFERENCE arXiv:2512.09466v2
cond-mat.other: 22 Dec 2025 Page 1 --- manuscript submitted to code-mat.other REFERENCE Natural hydrogen (H2) generated by water–rock interaction in ultramafic rocks is increasingly recognised as a potentially important primary energy resource, but the pore-scale processes that control the initiation and early transport of a free gas phase remain poorly constrained Here we present an in situ X-ray micro-tomography experiment in which an ultramafic granular pack of ultramafic from Papua Province, Indonesia, saturated with KI-doped brine, is heated to 100◦C with a pore pressure of 4 bar under 10 bar confining pressure inside a micro-CT scanner Time-resolved 4D imaging captures the transition from a fully liquid-saturated pore space to the appearance and growth of a distinct gas phase after an ∼8 h induction period Bubbles first nucleate near the top of the sample before becoming distributed throughout the imaged volume as a connected ganglia The nucleating gas phase is most plausibly dominated by molecular hydrogen (H2) generated by low-temperature fluid–rock reaction, as indicated by independent hydrogenpresence detectors, although we cannot yet fully exclude minor contributions from other gases SEM-BEX imaging reveals textural alteration and local changes in elemental signals between reacted and unreacted material Observations provide spatially and temporally resolved evidence for gas generation during low-temperature alteration of ultramafic grains and demonstrate that pore-scale imaging can directly link water–rock reaction kinetics, gas generation and multiphase flow behaviour in natural hydrogen systems Page 2 --- manuscript submitted to code-mat.other REFERENCE Plain Language Summary Hydrogen gas can be produced naturally when certain ultramafic, iron-rich rocks formed in Earth’s mantle react with water There is growing interest in whether this “natural hydrogen” could be used as a low-carbon energy source, but we still know very little about where the gas forms inside the rock and how it starts to move In this study, we packed grains of an ultramafic rock into a small cylinder, saturated them with salty water, and heated the system while it sat inside a 3D X-ray scanner (micro-CT) At first, all of the pores contained only liquid As the experiment continued, dark patches appeared and grew in the images, showing that a separate gas phase was forming Two independent measurements of the gas produced during the experiment are consistent with hydrogen, although this single experiment cannot fully describe all of the reactions taking place After imaging, we examined the reacted grains under an electron microscope and saw evidence that the minerals had altered and some elements had moved Together, these observations demonstrate that we can directly visualise gas generation inside reactive rocks at relatively low temperatures Page 3 --- manuscript submitted to code-mat.other REFERENCE 1 Introduction Natural geologic hydrogen (H2) is emerging as a potentially vast, carbon-free primary energy resource that could complement, or in some settings substitute, engineered “green” H2 production Global compilations of gas analyses from seeps, boreholes and mines show that H2 is far more widespread than previously appreciated and occurs in many tectonic settings, often as a major component of the gas phase (Zgonnik, 2020) Recent system-scale syntheses formalise the concept of a “H2 system”, in which H2 is generated, migrates and accumulates in traps in an analogous way to hydrocarbons (Jackson et al., 2024; Prinzhofer & Deville, 2015) This has led to growing industrial interest in “native” or “natural” hydrogen, with order assessments suggesting that alteration-driven systems (e.g.
Kai: areas of serpentinisation) could sustain commercially significant fluxes over geological timescales (Gaucher, 2020), and to proposals for engineered “rock-based” hydrogen concepts (white and “orange” hydrogen) that explicitly exploit water–rock reactions in mafic and ultramafic rocks as part of the low-carbon hydrogen portfolio (Osselin et al., 2022) Page 4 --- manuscript submitted to code-mat.other REFERENCE Among the diverse proposed sources of natural H2, low- to high-temperature serpentinization of ultramafic rocks is generally considered the most efficient and volumetrically important (Oze & Sharma, 2007; McCollom & Bach, 2009; Jackson et al., 2024) In this process, Fe2+-bearing primary minerals such as olivine and pyroxene react with water to form Fe3+-bearing phases (e.g., magnetite, Fe(III)-serpentine), generating H2 as a by-product (Sleep et al., 2004; Holm et al., 2015) Thermodynamic and kinetic models predict that H2 yields depend sensitively on temperature, water-to-rock ratio, mineralogy, water chemistry, and redox state (McCollom & Bach, 2009; McCollom et al., 2016; Ely et al., 2023) while field observations from ophiolites and oceanic and continental settings (e.g., Samail ophiolite, Oman; Lost City hydrothermal field) demonstrate that serpentinization can sustain long-lived fluxes of H2-rich, hyperalkaline fluids (Schrenk et al., 2013) More recently, work in the Samail ophiolite Oman has shown that significant H2 can also be generated during late, low-temperature (< 200◦C) alteration of already serpentinized peridotite, extending the window of H2 production to cooler, near-surface conditions (Ellison et al., 2021) Page 5 --- manuscript submitted to code-mat.other REFERENCE Experimental studies have been crucial in constraining the rates and controls of H2 generation, but they are dominated by batch or flow-through reactor experiments containing powders or finely crushed rock Low-temperature (30–150◦C) experiments on partially serpentinized peridotite and olivine powders show H2 production over days to months, with rates and yields strongly influenced by grain size, the presence of magnetite and fluid composition More recently, flow-through experiments on olivine under hydrothermal conditions have demonstrated sustained H2 generation under continuous fluid circulation and further quantified the coupling between serpentinization, reaction progress and H2 release However because powders maximise reactive surface area and eliminate a connected pore space, they provide little insight into where in a consolidated rock H2 actually nucleates, how rapidly it forms a separate gas phase, or how pore-scale connectivity and mineral heterogeneity control the transition from dissolved to free gas Page 6 --- manuscript submitted to code-mat.other REFERENCE Core-scale reactive transport experiments bridge part of this gap by preserving rock fabric and permeability while allowing continuous geochemical monitoring Yet in most such studies the rock behaves as a “black box”: H2 production is inferred from outlet fluid chemistry or headspace analyses, and any gas phase that forms inside the pore network remains invisible Even in well-characterised natural systems such as the Samail ophiolite, constraints on H2 generation and transport rely primarily on fluid chemistry, isotopes and mineralogical proxies rather than direct imaging of the gas phase itself As a result, the fundamental pore-scale mechanisms governing the initiation and early transport of H2, bubble nucleation sites, snap-off events, critical gas saturation and connectivity are still poorly constrained Page 7 --- manuscript submitted to code-mat.other REFERENCE In parallel, a large and rapidly growing body of work has used X-ray micro-tomography (micro-CT) to image dynamic multiphase flow and reactive transport in rocks Timeresolved (4D) micro-CT has been applied to dissolution and precipitation in carbonates–3 Similar techniques have been used to quantify residual trapping and ganglion dynamics during CO2–brine flow, to measure capillary pressure–saturation relationships, and to calibrate pore-network and direct numerical simulations These studies demonstrate that 4D imaging can resolve micron-scale fluid–fluid and fluid–solid interfaces, capturing non-linear displacement patterns, snap-off and reconnection events that strongly influence macroscopic transport properties Only recently has this pore-scale imaging toolkit been applied to H2 itself MicroCT studies of H2–brine flow in sandstones show that H2 behaves as a non-wetting phase in many reservoir-relevant conditions, with significant capillary trapping, strong sensitivity to rock heterogeneity and subtle but important differences from CO2–brine systems Page 8 --- manuscript submitted to code-mat.other REFERENCE Despite these advances, a critical observational gap remains at the intersection of serpentinization geochemistry and pore-scale multiphase flow physics To our knowledge, there are no prior studies that directly image the nucleation and growth of an H2 gas phase as it is produced within an ultramafic rock under hydrothermal conditions Existing serpentinization experiments constrain bulk H2 generation rates but lack spatially resolved information on gas formation, while pore-scale H2 imaging studies focus on storage scenarios with externally supplied gas In this study, we address this gap using a proof-of-concept in situ experiment that combines time-resolved 4D micro-CT imaging with post-mortem SEM and BEX analyses of an ultramafic granular pack reacting with brine at 100◦C and 4 bar pore pressure Page 9 --- manuscript submitted to code-mat.other REFERENCE 2 Materials and Methods 2.1 Sample Selection and Preparation The experiment utilized an ultramafic rock sample selected for its potential for lowtemperature serpentinization The sample was collected from the mantle sequence of the Cyclops Ophiolite located in the Cyclops Mountains of Papua Province, Indonesia (alongthe northern coast of the island of New Guinea) Petrographic inspection shows extensive alteration consistent with serpentinization, with relict olivine grains preserved within a fine-grained alteration matrix The rock was crushed and dry sieved to isolate the coarse fraction, retaining only grains between 0.6-1 mm Page 10 --- manuscript submitted to code-mat.other REFERENCE 2.2 Experimental Setup and Protocol All experiments were conducted at a temperature of 100◦C The pore space was saturated with KI-doped brine at a pore pressure of 4 bar, while the core holder was loaded within a Hassler-type sleeve under an external confining pressure of 10 bar Unless otherwise stated, pressures quoted in this study refer to pore pressure The granular sample was loaded into a custom-built X-ray transparent core holder designed for high-pressure and high-temperature flow experiments The initial saturation protocol was designed to eliminate trapped air and ensure full liquid saturation prior to heating First, the sample was flushed with CO2 at ambient conditions The system was then confined at 10 bar pressure Subsequently, the sample was flooded with a brine solution composed of 5 wt% potassium iodide (KI) at a pore pressure of 4 bar to keep the brine in the liquid phase KI was used as an X-ray contrast agent to enhance the attenuation of the fluid phase, thereby maximizing the contrast between rock grains (grey), the brine (bright white), and any generated gas (black) Page 11 --- manuscript submitted to code-mat.other REFERENCE At this concentration KI is not expected to strongly modify the low-temperature alteration reactions compared to chloride brines, and we did not observe any textural or compositional evidence for abundant KI salt precipitates in the post-reaction samples The system was flooded for 1000 Pore Volumes (PV) to ensure the complete dissolution of the initial CO2 into the brine, ensuring that any gas phase observed later resulted from reaction rather than trapped shielding gas Following the brine saturation phase, flow was halted (static conditions with no net through-flow), and the temperature of the sample was raised to 100◦C to initiate the hydrothermal reaction Page 12 --- manuscript submitted to code-mat.other REFERENCE During the first 8 hours the images were taken every 50 mins with no H2 detected until scan 11 After the onset of gas formation had been captured, the imaging interval was increased and the sample was imaged again only at 23 and 30 hours Each 3D image took approximately 50 mins to acquire with 800 projections at 100 KeV and 9.9 Watts and 2 frames/s in an Rx EasyTom micro-CT scanner Page 13 --- manuscript submitted to code-mat.
Paper summary: Kai: So, to recap what we just discussed about "A First Look at Hydrogen Generation in an Ultramafic Rock with Micro-CT and SEM-BEX", the core thesis is that they used time-resolved 4D X-ray micro-tomography to image gas formation in a reacting ultramafic rock <ref:2512.09466#pg1,A First Look at Hydrogen Generation in an Ultramafic Rock with Micro>.
Mira: They claim that this experiment shows a gas phase appears during the low temperature reaction of brine saturated ultramafic grains, and they use electron microscopy to show mineral alteration and element redistribution consistent with fluid rock reaction.
Lev: The setup involved heating an ultramafic granular pack from Papua Province, Indonesia, which was saturated with KI-doped brine at four bar pore pressure under ten bar confining pressure inside a micro-CT scanner.
Kai: The key observation is that the time-resolved 4D imaging captures the transition from a fully liquid-saturated pore space to the appearance and growth of a distinct gas phase after an eight hour induction period <ref:2512.09466#pg1>.
Mira: They also point out that bubbles first nucleate near the top of the sample before becoming distributed throughout the imaged volume as a connected ganglia.
Lev: The nucleating gas phase is most plausibly dominated by molecular hydrogen, H2, generated by low-temperature fluid rock reaction, which they indicate is supported by independent hydrogen presence detectors.
Kai: But they also mention that they cannot yet fully exclude minor contributions from other gases in that initial observation.
Mira: The SEM-BEX imaging reveals textural changes and local changes in elemental signals between the reacted and unreacted material, which supports the idea of fluid rock reaction happening at the pore scale.
Lev: This whole study provides spatially and temporally resolved evidence for gas generation during low temperature alteration of ultramafic grains, demonstrating that pore scale imaging can directly link water rock reaction kinetics to gas generation and multiphase flow behaviour in natural hydrogen systems.
Conclusion: Kai: So, looking at the title "A First Look at Hydrogen Generation in an Ultramafic Rock with Micro-CT and SEM-BEX", the authors are essentially showing us a proof of concept for monitoring gas generation in these systems.
Mira: The implication is that they've shown that the three-stage evolution—no gas, rapid growth, then diffusion limited growth—is consistent with delayed supersaturation, nucleation, and subsequent diffusion limited growth within the imaged volume.
Lev: This suggests that zones of intense reaction might remain invisible in terms of free gas for extended periods before we can actually see them in a micro-CT scan.
Kai: So, what does this mean for us listening outside of this specific paper? It means that when you're looking at natural hydrogen systems, you can expect a delay before the first visible bubbles appear.
Mira: And once those nucleation thresholds are crossed, gas connectivity can establish itself over relatively short timescales compared to the overall reaction history.
Lev: This is valuable because it gives us constraints on where and how gas forms in these rocks, moving beyond just inferring it from fluid chemistry or bulk measurements.
Hannah P. Menke, Zaid Z. Jangda, Max Webb, Jim Buckman, Amy Gough
Institute of GeoEnergy Engineering, Heriot-Watt University
cond-mat.other
Submitted: 2025-12-10
Updated: 2026-10-28
Comments: 24 pages, 7 figures, 6 tables, including Supporting Information. Revised to include additional hydrogen-generation measurements and petrographic characterisation
License: http://creativecommons.org/licenses/by-nc-nd/4.0/
Importance score: 75/100
The gist: Natural hydrogen (H2) generated by water–rock interaction in ultramafic rocks is increasingly recognised as a potentially important primary energy resource…
Key concepts
- Water–Rock Interaction
- This is a chemical reaction where water reacts with minerals in ultramafic rocks, such as olivine. This process is a key mechanism for generating natural hydrogen gas, which is important for potential energy resources.
- Micro-CT Imaging
- This technique uses X-rays to create detailed 3D images of the rock's internal structure at a very small scale (6.277 µm). It allows scientists to visualize how the liquid changes into gas and track its movement inside the tiny pores over time.
- Gas Phase Evolution
- The study tracked three stages of gas formation: no visible gas initially, rapid growth between 8 and 23 hours, and slower growth later. This evolution helps scientists understand the timing of when hydrogen bubbles nucleate and how they connect throughout the rock pores.
Terminology
Summary
Natural hydrogen (H2) generated by water–rock interaction in ultramafic rocks is increasingly recognised as a potentially important primary energy resource, but the pore-scale processes that control the initiation and early transport of a free gas phase remain poorly constrained The time-resolved 4D imaging captures the transition from a fully liquid-saturated pore space to the appearance and growth of a distinct gas phase after an ∼8 h induction period Observations provide qualitative pore scale evidence relevant to natural hydrogen generation in the subsurface
Experimental Setup and Methodology
The study utilized an in situ X-ray micro-tomography experiment where an ultramafic granular pack saturated with KI-doped brine was heated to 100◦C under specific pressure conditions The setup involved confining the sample within a Hassler-type sleeve under an external confining pressure of 10 bar while maintaining a pore pressure of 4 bar KI was used as an X-ray contrast agent to enhance the attenuation of the fluid phase, maximizing the contrast between rock grains, brine, and gas Time-resolved 3D images were acquired during the heating phase at a spatial resolution of 6.277 µm to capture inter-granular porosity and nucleation events
Observation of Gas Phase Evolution
The time series of segmented gas phase shows a three-stage evolution: (i) no resolvable gas from 0–8 h, (ii) rapid gas generation and connectivity growth between 8 and 23 h, and (iii) slower, diffusion-limited growth between 23 and 30 h Gas bubble nucleation was first observed at 8 hours near the top of the sample column The bulk gas saturation increased to Sg ≈ 0.39 by 23 hours, corresponding to a rate of change of ∆Sg/∆t ≈ 2. p. 8 39, corresponding to an average rate of change ∆Sg/∆t ≈ 2. 6 × 10−2 h
−1 between 8 and 23 hours (Table 1b; Figure 5).Where the paper says it
Mineralogical and Chemical Evidence
SEM-BEX imaging revealed textural alteration and local changes in elemental signals between reacted and unreacted material Post-reaction SEM imaging showed more irregular, roughened surfaces with clusters of elongate to platy crystals protruding from the grain surface BEX spot analyses indicated that post-reaction spots showed a broader spread with locally elevated Al and Ca, suggesting partial dissolution of Mg-rich primary phases
Implications for Natural Hydrogen Systems
The results suggest that substantial water–rock reaction and H2 production can occur before a free gas phase becomes detectable at the pore scale The delay before the first visible bubbles is consistent with a timescale of 6–20 hours, which brackets the observed induction period Once nucleation thresholds are crossed, gas connectivity can be established over relatively short timescales compared to the overall reaction history This behaviour suggests that zones of intense reaction may remain “invisible” in terms of free gas for extended periods
Conclusions
The proof-of-concept experiment demonstrates that time-resolved 4D micro-CT can monitor the emergence and early evolution of a gas phase generated by water–rock reaction in an ultramafic granular pack The three-stage evolution is consistent with delayed supersaturation, nucleation and subsequent diffusion-limited growth within the imaged volume The study provides qualitative constraints on pore-scale behaviour regarding gas nucleation and connectivity during low-temperature alteration Future work should extend this approach to a broader range of mineralogies, pressures, temperatures and flow regimes
REFERENCE
arXiv:2512.09466v2 [cond-mat. p. 0 uk
–1–
arXiv:2512. 09466v2 [cond-mat.Where the paper says it
--- Page 1 ---
manuscript submitted to code-mat.other
REFERENCE
Abstract
Natural hydrogen (H2) generated by water–rock interaction in ultramafic rocks is increasingly recognised as a potentially important primary energy resource, but the pore-scale processes that control the initiation and early transport of a free gas phase remain poorly constrained Here we present an in situ X-ray micro-tomography experiment in which an ultramafic granular pack of ultramafic from Papua Province, Indonesia, saturated with KI-doped brine, is heated to 100◦C with a pore pressure of 4 bar under 10 bar confining pressure inside a micro-CT scanner Time-resolved 4D imaging captures the transition from a fully liquid-saturated pore space to the appearance and growth of a distinct gas phase after an ∼8 h induction period Bubbles first nucleate near the top of the sample before becoming distributed throughout the imaged volume as a connected ganglia The nucleating gas phase is most plausibly dominated by molecular hydrogen (H2) generated by low-temperature fluid–rock reaction, as indicated by independent hydrogenpresence detectors, although we cannot yet fully exclude minor contributions from other gases SEM-BEX imaging reveals textural alteration and local changes in elemental signals between reacted and unreacted material Observations provide spatially and temporally resolved evidence for gas generation during low-temperature alteration of ultramafic grains and demonstrate that pore-scale imaging can directly link water–rock reaction kinetics, gas generation and multiphase flow behaviour in natural hydrogen systems
--- Page 2 ---
manuscript submitted to code-mat.other
REFERENCE
Plain Language Summary
Hydrogen gas can be produced naturally when certain ultramafic, iron-rich rocks formed in Earth’s mantle react with water There is growing interest in whether this “natural hydrogen” could be used as a low-carbon energy source, but we still know very little about where the gas forms inside the rock and how it starts to move In this study, we packed grains of an ultramafic rock into a small cylinder, saturated them with salty water, and heated the system while it sat inside a 3D X-ray scanner (micro-CT) At first, all of the pores contained only liquid As the experiment continued, dark patches appeared and grew in the images, showing that a separate gas phase was forming Two independent measurements of the gas produced during the experiment are consistent with hydrogen, although this single experiment cannot fully describe all of the reactions taking place After imaging, we examined the reacted grains under an electron microscope and saw evidence that the minerals had altered and some elements had moved Together, these observations demonstrate that we can directly visualise gas generation inside reactive rocks at relatively low temperatures
--- Page 3 ---
manuscript submitted to code-mat.other
REFERENCE
1 Introduction
Natural geologic hydrogen (H2) is emerging as a potentially vast, carbon-free primary energy resource that could complement, or in some settings substitute, engineered “green” H2 production Global compilations of gas analyses from seeps, boreholes and mines show that H2 is far more widespread than previously appreciated and occurs in many tectonic settings, often as a major component of the gas phase (Zgonnik, 2020) Recent system-scale syntheses formalise the concept of a “H2 system”, in which H2 is generated, migrates and accumulates in traps in an analogous way to hydrocarbons (Jackson et al., 2024; Prinzhofer & Deville, 2015) This has led to growing industrial interest in “native” or “natural” hydrogen, with order assessments suggesting that alteration-driven systems (e. p. 1 , 2024; Prinzhofer & Deville, 2015), and multi-scale reviews highlight the key
migration pathways and timescales from the planetary to basin scale (Lodhia et al. This has led to growing industrial interest in “native” or “natural” hydrogen, with first-
–2– order assessments suggesting that alteration-driven systems (e. areas of serpentinisation) could sustain commercially significant fluxes over geological timescales (Gaucher,
2020), and to proposals for engineered “rock-based” hydrogen concepts (white and “orange” hydrogen) that explicitly exploit water–rock reactions in mafic and ultramafic rocks
as part of the low-carbon hydrogen portfolio (Osselin et al.Where the paper says it
Where the paper says it
--- Page 4 ---
manuscript submitted to code-mat.other
REFERENCE
Among the diverse proposed sources of natural H2, low- to high-temperature serpentinization of ultramafic rocks is generally considered the most efficient and volumetrically important (Oze & Sharma, 2007; McCollom & Bach, 2009; Jackson et al., 2024) In this process, Fe2+-bearing primary minerals such as olivine and pyroxene react with water to form Fe3+-bearing phases (e. p. 2 Among the diverse proposed sources of natural H2, low- to high-temperature serpentinization of ultramafic rocks is generally considered the most efficient and volumetrically important (Oze & Sharma, 2007; McCollom & Bach, 2009; Jackson et al. In this process, Fe2+-bearing primary minerals such as olivine and pyroxene react with
water to form Fe3+-bearing phases (e. In this process, Fe2+-bearing primary minerals such as olivine and pyroxene react with
water to form Fe3+-bearing phases (e. , magnetite, Fe(III)-serpentine), generating H2
as a by-product (Sleep et al. , magnetite, Fe(III)-serpentine), generating H2
as a by-product (Sleep et al. , 2004; Holm et al. , 2004; Holm et al. Thermodynamic and kinetic models predict that H2 yields depend sensitively on temperature, water-to-rock ratio, mineralogy, water chemistry, and redox state (McCollom & Bach, 2009; McCollom et al. Thermodynamic and kinetic models predict that H2 yields depend sensitively on temperature, water-to-rock ratio, mineralogy, water chemistry, and redox state (McCollom & Bach, 2009; McCollom et al. , 2016;
Ely et al. , 2016;
Ely et al. , 2023), while field observations from ophiolites and oceanic and continental settings (e. , 2023), while field observations from ophiolites and oceanic and continental settings (e. , Samail ophiolite, Oman; Lost City hydrothermal field) demonstrate that serpentinization can sustain long-lived fluxes of H2-rich, hyperalkaline fluids (Schrenk et
al. , Samail ophiolite, Oman; Lost City hydrothermal field) demonstrate that serpentinization can sustain long-lived fluxes of H2-rich, hyperalkaline fluids (Schrenk et
al. More recently, work in the Samail ophiolite Oman has shown that significant
H2 can also be generated during late, low-temperature (< 200◦C) alteration of already
serpentinized peridotite, extending the window of H2 production to cooler, near-surface
conditions (Ellison et al.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
--- Page 5 ---
manuscript submitted to code-mat.other
REFERENCE
Experimental studies have been crucial in constraining the rates and controls of H2 generation, but they are dominated by batch or flow-through reactor experiments containing powders or finely crushed rock Low-temperature (30–150◦C) experiments on partially serpentinized peridotite and olivine powders show H2 production over days to months, with rates and yields strongly influenced by grain size, the presence of magnetite and fluid composition More recently, flow-through experiments on olivine under hydrothermal conditions have demonstrated sustained H2 generation under continuous fluid circulation and further quantified the coupling between serpentinization, reaction progress and H2 release However because powders maximise reactive surface area and eliminate a connected pore space, they provide little insight into where in a consolidated rock H2 actually nucleates, how rapidly it forms a separate gas phase, or how pore-scale connectivity and mineral heterogeneity control the transition from dissolved to free gas
--- Page 6 ---
manuscript submitted to code-mat.other
REFERENCE
Core-scale reactive transport experiments bridge part of this gap by preserving rock fabric and permeability while allowing continuous geochemical monitoring Yet in most such studies the rock behaves as a “black box”: H2 production is inferred from outlet fluid chemistry or headspace analyses, and any gas phase that forms inside the pore network remains invisible Even in well-characterised natural systems such as the Samail ophiolite, constraints on H2 generation and transport rely primarily on fluid chemistry, isotopes and mineralogical proxies rather than direct imaging of the gas phase itself As a result, the fundamental pore-scale mechanisms governing the initiation and early transport of H2, bubble nucleation sites, snap-off events, critical gas saturation and connectivity are still poorly constrained
--- Page 7 ---
manuscript submitted to code-mat.other
REFERENCE
In parallel, a large and rapidly growing body of work has used X-ray micro-tomography (micro-CT) to image dynamic multiphase flow and reactive transport in rocks Timeresolved (4D) micro-CT has been applied to dissolution and precipitation in carbonates–3 Similar techniques have been used to quantify residual trapping and ganglion dynamics during CO2–brine flow, to measure capillary pressure–saturation relationships, and to calibrate pore-network and direct numerical simulations These studies demonstrate that 4D imaging can resolve micron-scale fluid–fluid and fluid–solid interfaces, capturing non-linear displacement patterns, snap-off and reconnection events that strongly influence macroscopic transport properties Only recently has this pore-scale imaging toolkit been applied to H2 itself MicroCT studies of H2–brine flow in sandstones show that H2 behaves as a non-wetting phase in many reservoir-relevant conditions, with significant capillary trapping, strong sensitivity to rock heterogeneity and subtle but important differences from CO2–brine systems
--- Page 8 ---
manuscript submitted to code-mat.other
REFERENCE
Despite these advances, a critical observational gap remains at the intersection of serpentinization geochemistry and pore-scale multiphase flow physics To our knowledge, there are no prior studies that directly image the nucleation and growth of an H2 gas phase as it is produced within an ultramafic rock under hydrothermal conditions Existing serpentinization experiments constrain bulk H2 generation rates but lack spatially resolved information on gas formation, while pore-scale H2 imaging studies focus on storage scenarios with externally supplied gas In this study, we address this gap using a proof-of-concept in situ experiment that combines time-resolved 4D micro-CT imaging with post-mortem SEM and BEX analyses of an ultramafic granular pack reacting with brine at 100◦C and 4 bar pore pressure
--- Page 9 ---
manuscript submitted to code-mat.other
REFERENCE
2 Materials and Methods
2.1 Sample Selection and Preparation
The experiment utilized an ultramafic rock sample selected for its potential for lowtemperature serpentinization The sample was collected from the mantle sequence of the Cyclops Ophiolite located in the Cyclops Mountains of Papua Province, Indonesia (alongthe northern coast of the island of New Guinea) Petrographic inspection shows extensive alteration consistent with serpentinization, with relict olivine grains preserved within a fine-grained alteration matrix The rock was crushed and dry sieved to isolate the coarse fraction, retaining only grains between 0.6-1 mm
--- Page 10 ---
manuscript submitted to code-mat.other
REFERENCE
2.2 Experimental Setup and Protocol
All experiments were conducted at a temperature of 100◦C The pore space was saturated with KI-doped brine at a pore pressure of 4 bar, while the core holder was loaded within a Hassler-type sleeve under an external confining pressure of 10 bar Unless otherwise stated, pressures quoted in this study refer to pore pressure The granular sample was loaded into a custom-built X-ray transparent core holder designed for high-pressure and high-temperature flow experiments The initial saturation protocol was designed to eliminate trapped air and ensure full liquid saturation prior to heating First, the sample was flushed with CO2 at ambient conditions The system was then confined at 10 bar pressure Subsequently, the sample was flooded with a brine solution composed of 5 wt% potassium iodide (KI) at a pore pressure of 4 bar to keep the brine in the liquid phase KI was used as an X-ray contrast agent to enhance the attenuation of the fluid phase, thereby maximizing the contrast between rock grains (grey), the brine (bright white), and any generated gas (black)
--- Page 11 ---
manuscript submitted to code-mat.other
REFERENCE
At this concentration KI is not expected to strongly modify the low-temperature alteration reactions compared to chloride brines, and we did not observe any textural or compositional evidence for abundant KI salt precipitates in the post-reaction samples The system was flooded for 1000 Pore Volumes (PV) to ensure the complete dissolution of the initial CO2 into the brine, ensuring that any gas phase observed later resulted from reaction rather than trapped shielding gas Following the brine saturation phase, flow was halted (static conditions with no net through-flow), and the temperature of the sample was raised to 100◦C to initiate the hydrothermal reaction
--- Page 12 ---
manuscript submitted to code-mat.other
REFERENCE
During the first 8 hours the images were taken every 50 mins with no H2 detected until scan 11 After the onset of gas formation had been captured, the imaging interval was increased and the sample was imaged again only at 23 and 30 hours Each 3D image took approximately 50 mins to acquire with 800 projections at 100 KeV and 9.9 Watts and 2 frames/s in an Rx EasyTom micro-CT scanner
--- Page 13 ---
manuscript submitted to code-mat.
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Better predictive modeling of natural hydrogen systems can be achieved by incorporating pore-scale connectivity thresholds into kinetic models to predict whether H2 generation will lead to detectable free gas or remain trapped in dissolved states.
-
AI systems can be trained to interpret time-resolved micro-CT data from reactive rocks to automatically identify the three distinct regimes of bubble evolution: no resolvable gas, rapid gas generation and connectivity growth, and slower diffusion-limited growth.
-
Improved AI can predict the characteristic timescale for H2 accumulation by linking literature values for H2 solubility at 100 degrees Celsius to experimentally derived reaction rates and surface areas within the imaged pore volume.
-
Systems can be developed to correlate BEX elemental maps with textural changes observed in SEM images, specifically identifying localized secondary alteration products such as
platy surface crystals
that indicate where mineralogical reorganisation occurs near pores. -
AI can perform automated analysis of gas ganglia statistics, including cluster size distributions and percolation metrics, to quantify the degree of pore-scale connectivity established by the generated H2 phase over time.
Abstract
We present a 4D (time-lapse) micro-CT experiment that visualizes gas emergence and growth during water-rock reaction in an ultramafic granular pack at 100 C. Repeated scans resolve the onset of gas exsolution after an initial delay, followed by rapid bubble growth, coalescence, and evolving gas connectivity. Image-derived metrics quantify changes in bulk gas saturation and gas clustering through time. Hydrogen was detected independently in the pump effluent and by an indicator strip in the confining sleeve. A companion reactor experiment using the same parent rock at 100 C yielded an estimated net production of 0.65 mmol H 2 over approximately three weeks, accounting for dissolved hydrogen and gas removed during sampling. Complementary SEM and multi-element backscattered electron and X-ray (BEX) imaging of pre- and post-reaction grains indicates localized alteration textures consistent with reaction-driven H 2 production. These observations constrain pore-scale gas accumulation and multiphase transport relevant to natural hydrogen accumulations.
Related papers
- Topological Magnetism in 2D and 3D: Skyrmions and Hopfions. Classical and quantum approaches
- Exact analytical edge states in the extended Su-Schrieffer-Heeger model
- Quantitative Theory for the Amplitude of Fluorescence Quantum Beats from Geminate Triplet-Pair Fusion
- Nonlinear Static Screening of Positive Charges in an Electron Gas: Contact Hartree Potential
- Plasmascopy of ultrafast hot charge carriers
- Non-Hermitian Skin Effect from Radiative Coupling in a Reciprocal Chiral Medium