1 s2.0 S0009254123003984 Main

Chemical Geology 638 (2023) 121698
Contents lists available at ScienceDirect
Chemical Geology
journal homepage: www.elsevier.com/locate/chemgeo
Natural hydrogen in low temperature geofluids in a Precambrian granite,

South Australia. Implications for hydrogen generation and movement in the
upper crust
Julien Bourdet a, *, Claudio Delle Piane a, Cornelia Wilske b, c, Dirk Mallants b, Axel Suckow b,
Danielle Questiaux c, Christoph Gerber b, Punjehl Crane b, Alec Deslandes b, Laure Martin d,
Matvei Aleshin d
a
CSIRO Energy, 26 Dick Perry Avenue, Kensington, WA 6151, Australia
b
CSIRO Environment, Waite Road – Gate 4, Glen Osmond, SA 5064, Australia
c
The University of Adelaide, School of Physical Sciences, North Terrace, Adelaide, SA 5000, Australia
d
The University of Western Australia, Centre for Microscopy, Characterisation and Analysis, WA 6009, Australia
A R T I C L E I N F O A B S T R A C T
Editor: Oleg Pokrovsky Natural hydrogen (H2) has the potential to be a low-carbon fuel and energy source, playing a crucial role in
achieving a clean, secure, and affordable energy future. However, our understanding of the generation, migra
Keywords: tion, and accumulation of H2 in the subsurface remains poorly constrained. Increasing evidence suggests that H2
Natural hydrogen is abundantly and widely present in various geological settings, making them viable targets for exploration of this
Granite
emerging resource. Reports have indicated the presence of hydrogen-dominated natural gas in vintage boreholes
Fluid inclusion
located in the Yorke Peninsula and Kangaroo Island, South Australia. These boreholes were drilled in Cambrian
Raman spectroscopy
Magnetite sedimentary cover over Precambrian terranes that host iron ore, copper or gold ore deposits. To gain insight into
Hematite these occurrences, rock samples were examined from the Roxby Downs granite, a Mesoproterozoic pluton
Water-rock interaction associated with the Hiltaba Suite event (1.59 Ga), belonging to the same magmatic origin as the Yorke Peninsula
Water radiolysis basement.
Energy resource Our investigation revealed metasomatic fluid circulation, evidenced by alteration of feldspar and mafic grains,
as well as discrete cemented fractures within the granite. In some of these fracture-fill cements, we discovered
water and gas inclusions containing hydrogen. The formation of the hydrogen-bearing cements occurred at
paleo-temperatures ranging from 170 ◦ C to the present-day 55 ◦ C. Analysing the oxygen isotope values of the
quartz cements and considering their fluid inclusion temperatures, a marine water signature as the source of
water in equilibrium with the quartz cements was identified.
To constrain the fluid history, neon and argon isotope signatures from the fluid inclusion gases were measured
from the granite and separated quartz grains. Nucleogenic neon isotopes increased with depth along specific
production lines. While this is consistent with long-lived fluid within the granite and increasing fluid residence
with depth, separated quartz grains show less variation with depth and suggest that younger fluids were
introduced, triggering alteration of less robust minerals, visible in bulk granite. Petrographic analyses identified
hydrolytic alterations of mafic minerals into magnetite, and subsequently magnetite to hematite, as sources of
hydrogen and contributors to the observed increase in pore water salinity. Our findings highlight that natural
hydrogen can be produced and retained in granites at low temperatures, expanding beyond the commonly re
ported alteration processes observed in mafic and ultramafic rocks.
1. Introduction low environmental impact. In this context, the production of low-cost

natural hydrogen is an attractive option, but the industry appetite re
Sustainable energy production requires decarbonised sources and mains emerging. Production of natural hydrogen gas reservoirs is, so far,
* Corresponding author.
E-mail address: julien.bourdet@csiro.au (J. Bourdet).
https://doi.org/10.1016/j.chemgeo.2023.121698
Received 7 June 2023; Received in revised form 16 August 2023; Accepted 29 August 2023
Available online 30 August 2023
0009-2541/© 2023 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
J. Bourdet et al. Chemical Geology 638 (2023) 121698
limited to one example in Mali (Prinzhofer et al., 2018; Maiga et al., Reference Library and prepared for petrographic and mineralogical
2023), where the hydrogen is converted to electricity for local usage. analysis, fluid inclusion, in-situ oxygen isotope measurements and neon
Emanation of hydrogen from subsurface has been, however, identified in and argon noble gas extracts and quantification of fluid inclusion
a variety of localities and geological contexts such as offshore mid extracts.
oceanic ridges (Charlou et al., 2002), volcanoes (Aiuppa et al., 2011),
onshore ophiolitic terranes (Vacquand et al., 2018), granites (Sherwood 2.1. Automated mineral analysis
Lollar et al., 2014; Nivin, 2016; Warr et al., 2018; Truche et al., 2021), at
the surface of sedimentary basins near circular depressions (Zgonnik Modal mineralogical analysis was conducted on polished thin sec
et al., 2015; Prinzhofer et al., 2019; Frery et al., 2021; Moretti et al., tions; prior to examination, the samples were coated with an electrically
2021) and tectonic fault zones (Sato et al., 1984). conductive 5 nm thick carbon layer. For quantitative mineralogy the
Geogenic H2 can be generated from ultrabasic rocks, iron-rich cra samples were analysed using the Tescan Integrated Mineral Analyser
tons and uranium-rich rocks (Gaucher, 2020). In granites, H2 generation (TIMA). TIMA is an automated mineralogy system for fast analysis of
has been associated to the radiolytic dissociation of water and to water- rock samples providing particle-by-particle quantitative mineralogical
rock hydration reactions (Sherwood Lollar et al., 2014; Warr et al., 2019; data (Hrstka et al., 2018). It operates by combining back scattered
Truche et al., 2021; Leila et al., 2021; Boreham et al., 2021). Truche electron (BSE) imaging and X-ray energy dispersive spectroscopy (EDS)
et al. (2021) documented the hydrothermal alteration of amphibole at analysis to identify minerals, determine their abundance and create
elevated temperatures 280 ◦ C–400 ◦ C, while serpentinization of mafic mineral distribution images. For this study, the measurement parame
and ultramafic rock can take place at lower temperatures (Mayhew ters were set to a spatial resolution defined by beam steps of 12 μm. For
et al., 2013; Neubeck et al., 2014; Okland et al., 2014). Natural gener each pixel of the map, the BSE and EDS signals from four independent
ation and migration of hydrogen from granitic basements at lower EDS detectors are collected and matched with available mineral libraries
temperatures is however a more attractive prospect as it would signify to create mineral maps. Results are expressed in mass percentage of
more common and widespread occurrences. individual minerals.
Hydrogen-dominated natural gas has been reported in vintage bore
holes in the Yorke Peninsula and Kangaroo Island in South Australia 2.2. X-ray diffraction mineralogy
(Boreham et al., 2021), above Precambrian terranes where Paleo
proterozoic and Mesoproterozoic felsic and mafic plutons constitute the XRD patterns were recorded with a PANalytical X’Pert Pro Multi-
basement of Cambrian sedimentary series, and in proximity with iron, purpose Diffractometer using Fe filtered Co Kα radiation, automatic
copper or gold ore deposits. This area is the southern part of the Olympic divergence slit, 2◦ anti-scatter slit and fast X’Celerator Si strip detector.
metallogenic belt, a Cu–Au Province that stretches over 500 km along The diffraction patterns were recorded from 3 to 80◦ in steps of 0.017◦ 2
the eastern margin of the Gawler Craton (Fig. 1). We investigated cores theta with a 0.5 s counting time per step for an overall counting time of
from Blanche 1 borehole drilled into the Roxby Downs granite, a pluton approximately 35 min.
of the Mesoproterozoic Hiltaba Suite event in South Australia. Blanche 1 Qualitative analysis was performed on the XRD data using in-house
is a continuously cored geothermal well, 8 km west of the giant Olympic XPLOT and HighScore Plus (from PANalytical) search/match software.
Dam Iron Oxide copper gold uranium mine located on the northern part Quantitative analysis was performed on the XRD data using the com
of the Olympic metallogenic belt. mercial package TOPAS from Bruker AXT. The results are normalised to
The Roxby Downs pluton is a A-type granite emplaced at ~1593.87 100%, and hence do not include estimates of unidentified or amorphous
± 0.21 Ma as part of the regionally extensive Hiltaba Suite magmatism materials.
(Cherry et al., 2018) in an intracontinental setting (Jagodzinski et al.,
2021). It intruded into the marginally older Gawler Range Volcanics 2.3. Fluid inclusion characterisation in thick sections
(1594.73 ± 0.30 Ma, Cherry et al., 2018), the Palaeoproterozoic Wal
laroo Group and Donington Suite granitoids (~1850 Ma). The Roxby Granites were prepared as double polished 100 μm thick sections
Downs Granite is generally porphyritic medium to coarse-grained sye glued on a petrographic glass slide using acetone-dissolvable superglue.
nogranite to monzogranite. It is also classified as metaluminous and While still glued on the glass slides, the rock wafers were cut in 5 mm
strongly oxidized. The granite is iron-rich, with hematite and magnetite, squares using a diamond saw fitted to a Dremel drill. Separation of the
enriched in F, Rb and high-field-strength elements (U, Th, Zr, Nb and double polished sections from the glass was performed by a three hours
Ce), relatively undeformed and with minor mafic to ultramafic intrusive acetone bath. Each wafer square surface was then carefully cleaned
rocks. Blanche 1 is a geothermal exploration borehole completed in using methanol and optical tissues.
2005 to a total depth of 1934 m. Below 718 m of sedimentary cover, it The microthermometry technique was performed using a standard
intersected 1218 m of variably altered syenogranite and monzogranite transmitted light petrographic microscope with Linkam MDS600
lithologies. Downhole petrophysical evaluation found an average tem heating-freezing stage and T95 controller connected to a LNP95 cooling
perature gradient of around 60 ◦ C/km in the overburden sediments and system. The stage was calibrated using final ice melting temperatures of
30 ◦ C/km in the granite and a measured maximum downhole temper H2O-NaCl synthetic inclusions in quartz of known salinities at 100 ppm
ature of 85.3 ◦ C at 1934.2 m (Green Rock Energy, 2006). The granite and 30,000 ppm, and a Linkam CO2 inclusions standard giving triple
body intersected by Blanche 1 is estimated to cover an area of about point at − 56.6 ◦ C. The precision and accuracy on temperature mea
1180 km2 with an average thickness of about 6000 m as interpreted from surements is ±0.1 ◦ C.
deep 2D reflection seismic surveys. The Horiba LabRam HR Evolution Raman spectrometer was used to
We report the presence of hydrogen gas contained in fluid inclusions detect hydrogen and methane using a 532 nm single frequency 100 mW
in cemented fractures of two granite samples collected at the depths of diode laser, providing 12 mW at the focus point through a 100×
918 m and 1319 m from the Blanche 1 borehole. An array of techniques objective. Grating settings were 1800 groves/mm for gas detection. The
was used to characterise the paleo-fluids and cementation processes to signal was collected using a 1024 × 256 pixels Peltier cooled CCD
estimate the conditions and key parameters responsible for the genera Synapse detector, sensitive in the 300–1050 nm range. The 4156 cm− 1
tion and migration of hydrogen in the granitic basement rocks. peak position of hydrogen was calibrated at increasing densities be
tween 1 and 300 bar using pure calibration H2 gas (Coregas) in a high-
2. Methods pressure optical cell (Chou et al., 2005; Li et al., 2018; Chen and Chou,
2022) and a reference calibration xenon lamp. The peak positions were
The samples were collected at the South Australia Drill Core derived on Labspec 6 software (Horiba) after a baseline treatment and
2
Fig. 1. Geological map of the main Archaean to Early Mesoproterozoic rock units (SARIG website) in South Australia. Blanche 1 is in the northern part of the
Olympic Dam Province. Three legacy wells that delivered reservoir gas samples dominated by H2 are located on the southern part of the Olympic Dam province. For
the units not presented in the legend, refer to the SARIG website.
3
using the Gaussian-Lorentzian peak finder tool. This allowed the deri cooled two-stage water trap at − 15 ◦ C, two steel cryotraps held at 180 K
vation of the hydrogen density and pressure of the gas inclusions at room and 30 K respectively, a getter, another steel cryotrap and a charcoal
temperature. cryotrap. Helium and neon are separated on the charcoal cryotrap with
A Carbide CB5 (Light Conversion) pulsed femtosecond laser system He desorbing at 40 K and Ne at 120 K. Heavy noble gases are separated
(6 W) was used to stimulate vapour bubble nucleation in the metastable from reactive gases with getters and Ar is separated from Kr and Xe on a
liquid state of the monophasic inclusions by means of single ultrashort stainless steel cryotrap at 40 K. Measurements of all isotope ratios were
laser pulses. Pulse duration 290 fs using a 515 nm with harmonic performed on a Thermo Fisher Scientific Helix-MC plus mass spec
generator, in TEM00 mode (M2 < 1.2). trometer and calibrated with air and in-house reference standards.
2.4. Oxygen isotope analyses by LG-SIMS 3. Characterisation of the host rock and fractures
18
O/16O measurements were carried out using a Cameca IMS 1280 Based on previous characterisation work presented by Wilske et al.
housed at the Centre for Microscopy Characterisation and Analysis at the (2022), we selected two fracture-bearing samples extracted from cores
University of Western Australia. Analyses were carried out in cut thick from 918 m and 1319 m depth. Both rocks have close mineralogical
sections embedded in epoxy together with quartz standards NBS 28 composition, consist of K-feldspar, quartz and albite phenocrystals
(Hut, 1987) and UWQ-1 (Kelly et al., 2007). The resin mounts were (Table 1) and are classified as monzogranite at 1319 m and syenogranite
thoroughly cleaned with detergent, ethanol and distilled water in an at 918 m. Petrographic analyses show that plagioclase have been almost
ultrasonic bath and coated with a 20-nm-thick Au coating prior to SIMS completely albitized and K-feldspar crystals commonly show albite
analyses. exsolution (perthite). Some albite crystals present altered cores with
A static 50pA Cs+ beam with an impact energy of 20 keV was focused illitic clay material in-fill (sericite). Biotite is totally altered into Fe-
to a ~ 3 μm spot on the sample surface. A normal-incidence electron gun chlorite (chamosite). Clusters of chamosite are commonly associated
was used for charge compensation. Instrument parameters were as fol with iron oxides crystals (hematite in the 918 m and magnetite in the
lows: magnification of 130× between the sample and field aperture, 400 1319 m sample), anatase, apatite and zircon.
μm contrast aperture, 2500 μm field aperture, 90 μm entrance slit, 500 Thin section analyses of the sample from 918 m depth, reveal rare,
μm exit slits, and a 30 eV band pass for the energy slit with a 5 eV gap. subvertical 10 to 100 μm wide veins cutting through several grains and
Secondary O− ions were accelerated to 10 keV and oxygen isotopes were cemented with quartz and barite, and minor galena, pyrite, chalcopyrite
analysed simultaneously with a mass resolving power of ~2430 using a and K-feldspar. Cathodoluminescence (CL) imaging shows that the vein
Faraday Cup detector (L1 for 16O) and an electron multiplier (H2 for cements have darker luminescence compared to the magmatic quartz
18
O) on the multicollection axis. Spots were pre-sputtered for 90 s before hosts (Fig. 2). The luminescence of the CL in veins can be uniform in
automated peak centering in the field and contrast apertures. Analyses places and vary in other parts of the veins, suggesting that multiple fluid
consisted of 60 × 4 s cycles, which gave an average internal precision of circulation and cement growth events occurred. The cement is also
0.25‰ (2 SD mean). Bracketing of reference materials allowed for characterised by an abundance of inclusions when transecting magmatic
instrumental mass fractionation (IMF) and drift corrections. Instru quartz crystals, significantly higher than in the quartz cement through
mental mass fractionation was corrected using UWQ-1 (12.33 ± 0.14‰, felspar crystals. Small inclusions of hematite and rare sulfides (galena)
Kelly et al., 2007) and accuracy was checked using NBS28 (Hut, 1987). are present in low amounts in quartz cemented veins. Micrometric
External precision (standard deviation of the mean) on UWQ-1 was cracks in magmatic quartz, parallel to the main fractures, are also
0.26‰ on mount from the 1319 m depth sample and 0.64‰ on mount commonly present. Hematite crystals contain residual magnetite, as
from the 918 m depth sample. NBS28 returned δ18O of 9.54 ± 1.35‰ (n evidence of martitization (magnetite to hematite alteration).
= 12, 2SD) and 9.56 ± 1.10‰ (n = 18, 2SD), which are consistent with The core at 1319 m presents sub-horizontal 100 μm wide fractures,
published value for the standard (9.57 ± 0.1‰). Corrected 18O/16O cemented by a combination of quartz, albite, chlorite, calcite and rare
ratios are reported in δ18O notation, in per mil variations relative to fluorite (Fig. 2). The mineralogy of the cement is strongly controlled by
VSMOW. Average values are calculated as simple mean for each quartz the mineralogy of the vein walls, for instance fractured quartz are
population and uncertainties are calculated as the standard deviation of cemented with quartz while fractured feldspars are often cemented with
the mean and returned at the 95% confidence level. albite, calcite, chlorite, with rare occurrences of fluorite. Quartz cement
in fractures in quartz is revealed by CL and characterised by dark
2.5. Stable noble gas isotope analysis of fluid inclusions from bulk granite luminescence contrasting with the brighter magmatic quartz. Fractures
and purified quartz consist commonly of a single vein, but rarer multiple coalescing thinner
veins also occur.
An in-house high vacuum crushing system was used to extract fluid
inclusions from minerals at room temperature. It consists of a stainless- 4. Fluid inclusions in fracture-fill cements and microcracks
steel container with a bellow on top holding the crush rod inside and an
outlet valve connectable to the noble gas machine port. The bottom 4.1. 918 m
blank flange holds multiple stackable tungsten carbide (WC) sample
plates in between which the sample grains are crushed. Three types of The quartz cements in veins at 918 m contain assemblage of di-
grains were prepared for crushing: (i) granite grains of 0.85–1 mm, (ii) phasic liquid-vapour (L-V) aqueous inclusions and assemblage of
hand-picked and chemically cleaned quartz grains of 1–2 mm, and (iii) monophasic liquid (L) and di-phasic (L-V) aqueous inclusions. Their
chemically purified quartz grains of 0.85–0.71 mm Each type had a total homogenisation temperatures (TH) vary widely between 54 ◦ C and
sample weight between 1.1 and 2.5 g. After a crusher container is filled 150 ◦ C (Fig. 3A). The TH values within an assumed assemblage
with a sample, it is evacuated to 10− 10 bar (high vacuum) and heated up commonly show 20 ◦ C to 30 ◦ C variability which indicates that in
to <80 ◦ C for ca. 24 h to desorb attached gases from the mineral surfaces clusions may not be part of the same assemblages or that some post-
and the inner surfaces of the crusher container. For measurements, the entrapment re-equilibrations have occurred. The assumption of an in
evacuated crusher container is attached directly to the inlet of the clusion to belong to an assemblage is not always obvious as cemented
preparation system, and after a leak check, an external manual press veins seem to have formed in multiple events in a same array. Never
moves the crush rod towards the top sample plate to crush the samples theless, the temperatures at which the veins formed can be estimated
with up to 10 tons of applied load. All gases and fluids are released into with some degree of confidence from the minimum homogenisation
the automated gas purification system where they pass through a Peltier- temperatures of the fluid inclusions in each assemblage. The veins
4
therefore formed at a minimum of 54 ◦ C to 121 ◦ C, likely during episodic
Mineralogy of the core samples at 918 m and 1319 m derived from automated mineral mapping of thin sections and quantitative X-ray diffraction (XRD) analysis (wt%) of bulk micronized sample. Nd: not detected.
Unclassified
events. Salinity of the aqueous inclusions are consistently very high
according to the final ice melting temperatures (Tm ice) measured be
8.6
8.9
tween − 15 ◦ C down to − 33 ◦ C, corresponding to total salinities above
20 wt% (NaCl Eq.) (Fig. 3C). During heat treatment, in some inclusions,
the bubble shape and movement seem to be obstructed by a large clear
Barite
0.1
Nd
Nd
Nd
solid, possibly salt, not systematically present in inclusions of a same
assemblage. Measured eutectic temperatures between − 28 to − 52 ◦ C are
Calcite indicative of a mix of salt systems consistent with NaCl-CaCl2-H2O
0.5
Nd
Nd
Nd
(Steele-MacInnis et al., 2011). Eutectic below − 50 ◦ C could indicate
some contents of LiCl which has very low eutectic temperature (Monnin
Thorite
et al., 2002). The hydrohalite melting temperature necessary to calcu

<0.1
0.1
Nd
Nd
late the absolute CaCl2 and NaCl quantities were not attempted due to
the small size of the inclusions. Hydrogen was frequently detected by
Zircon
Raman spectroscopy in the vapour phase of di-phasic water inclusions at

0.1
0.2
Nd
Nd
room temperature (additional illustration in supplementary material),
indicative of the presence of dissolved hydrogen in water when the fluid
Apatite
inclusions were trapped.

0.1
0.1
Nd
Nd
Healed microcracks in quartz crystals contain several types of as

semblages. We observed assemblages of gas inclusions with minor or no
visible water (Fig. 4C). This type indicates entrapment of free gas with
Anatase
0.4
0.2
0.3
0.2
Rutile/
low abundance of liquid water. Another type contained assemblages of

gas inclusions with variable amounts of water, diphasic L-V water in
clusions with small bubbles and monophasic liquid water inclusions at
Magnetite
room temperature (Fig. 4D). This type indicates entrapment of free gas
1.4
Nd
with high abundance of liquid water (heterogenous trapping). After

bubble nucleation using a femto-second laser stimulation, the mono
phasic water inclusions provided homogenisation temperatures between
Hematite
50 ◦ C and 87 ◦ C. Two salinities of 20 and 22 wt% (NaCl eq.) were

1.5
0.3
derived from final ice melting. The analysis of the gas in the gas in
clusions by Raman spectroscopy indicated the presence of hydrogen
only. The pressures were derived assuming pure H2 vapour and vary
oxide
0.3
1.4
from 3 to 81 bars with half of the gas inclusions between 30 and 60 bars.
Iron
It is anticipated that H2 could diffuse from the inclusions through the

quartz host grains with time. Diffusion should occur at the same rate for
Chlorite
all the gas inclusions within the assemblage. However, pressure vari
1.7
1.0
2.1
1.1
ability between inclusions of a same assemblage is often large. The cause

for low H2 pressure in some gas inclusions is therefore unclear. However,
it appears that the maximum hydrogen pressure measured in a gas in
Muscovite/
clusion of 81 bar along with the minimum TH of 50 ◦ C for the water

0.8
6.6
0.4
2.4
inclusions in the same assemblage, are a clear indication of entrapment

illite
at current day hydrostatic pressure and temperature. The values around

60 bars are compatible with current hydrostatic pressures (estimated at
Oligoclase
85–92 bar) and temperatures of around 150 to 180 ◦ C.

0.1
0.6
Nd
Nd
4.2. 1319 m
Feldspar
The quartz cements in veins at 1319 m contain assemblage of di-

12.5
16.3
5.7
9.1
Na-K
phasic (L-V) liquid-rich aqueous inclusions. Their homogenisation

temperatures range from 176 ◦ C to 260 ◦ C (Fig. 3B). Secondary in
clusions within cements, forming trails of inclusions, are also di-phasic
Albite
19.8
23.6
21.4
27.9
(L-V) liquid-rich aqueous inclusions and their homogenisation temper

atures range from 115 ◦ C to 199 ◦ C. For primary or secondary fluid in
Microcline
clusions, the TH values within an assumed assemblage commonly show

35.6
24.6
25.5
19.7
20 to 30 ◦ C variability, with no visible morphological evidence of re-

equilibration (Vityk and Bodnar, 1995). Misattribution as primary or
secondary is therefore possible and the attribution is used as a guide to
Quartz
indicate relative timing of fluid entrapment. The salinity of the aqueous

26.9
30.0
29.3
30.6
inclusions in primary inclusions are often close to 5 wt% (NaCl eq.) and
more rarely reaches higher values up to 15 wt% (NaCl eq.), according to
the final ice melting temperatures measured between − 3.0 ◦ C down to
1319
1319
918
918
Depth
− 10.6 ◦ C. In secondary assemblages, salinities vary widely between

(m)
assemblages with values from 5 wt% (NaCl eq.) up to 25 wt% (NaCl eq.)
(Tm ice = − 3.0 ◦ C to − 22.7 ◦ C). Only few eutectic temperatures were
Method
EDS
EDS
Table 1
SEM-
SEM-
XRD
XRD
measured between − 26 to − 46 ◦ C, indicative of a NaCl-CaCl2-H2O mix

of salt system. None of the primary or secondary tested inclusions
5
Fig. 2. Scanning electron microscopy images of cemented fractures in samples from depths of 918 m and 1319 m. A: Back-scattered electron (BSE) image of
inclusion-rich fracture transecting quartz and feldspar grains and filled with quartz and barite cement. B: Cathodoluminescence (CL) image of the same field of view
in A highlighting the secondary nature of the quartz fracture-fill (darker) in contrast to the bright signal from the quartz host grains. C: BSE images of a multimineral-
cemented fracture in cross-cutting feldspar and quartz grains and cements by fluorite, calcite and quartz, yellow square shows a CL image of the same field of view
showing dark quartz fracture-fill cement. D: False colour elemental map overlay on BSE image highlighting the spatial distribution of selected elements within the
fracture. Alb.: albite; Fl.: fluorite. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
contained hydrogen in the vapour phase, indicative of an absence or gas inclusions using the 4156 cm− 1 peak position range between 106
very low level of dissolved hydrogen in the water when formed. and 325 bar. The minor methane contents play an important role in the
Calcite cements contain di-phasic (L-V) liquid-rich aqueous inclusion density change rendering the pressure derivation invalid for methane
assemblages with TH values from 167 ◦ C to 221 ◦ C with elevated sa co-bearing inclusions, and pressures should only be derived where
linities varying between assemblages with values from 13 wt% (NaCl methane or other gases are absent. Entrapment at current day hydro
eq.) up to 24 wt% (NaCl eq.) (Tm ice = − 8.0 ◦ C to − 22.0 ◦ C). Hydrogen in static pressure and temperature of pure hydrogen gas would provide
the vapour phase could not be tested due to the fluorescence of the inclusions at about 107–111 bar (the minimum value is considering
calcite under laser illumination. current day water table level and the maximum value is considering the
Fluorite cements contained fluid inclusion assemblage constituted by water table level at the surface), which is close to the minimum pressure
gas inclusions, vapour-rich aqueous inclusions, and liquid-rich aqueous derived in the all-gas inclusions.
inclusions (Fig. 5). This type of assemblage is typical of two phases,
liquid and vapour, and evidence for heterogenous entrapment of a fluid 5. Identification of fracture fluid origins by in-situ oxygen
that reached its boiling point, leading to phase separation. Quantifica isotopes
tion of homogenisation temperatures was only attempted on the liquid-
rich inclusions showing the smallest bubble sizes, returning values of Oxygen isotopes values in the 1319 m sample were collected across
155 ◦ C and 177 ◦ C. These represent true entrapment temperatures due to two analytical transects including the ~100 μm-thick quartz-filled
the fluid being at its bubble point. Their salinities were 21 wt% and 10 fractures and the magmatic quartz host grain (Fig. 6). Additional
wt% (NaCl eq.) respectively (Tm ice = − 18.0 ◦ C and − 6.5 ◦ C). Raman points have been analysed from the tips of the fracture, within a large
spectroscopy revealed hydrogen in the vapour in the two-phase L-V feldspar crystal (Fig. 6). The magmatic domains show a mean δ18O value
aqueous inclusions and the all-vapour inclusions. Low methane content of 9.5 ± 1.2‰ (n = 14), which is lower than in the fractures, where a
was also detected in some inclusions (both in V or L-V), while not mean δ18O of 12.2 ± 1.9‰ (n = 19) is observed.
detected elsewhere. Pressure derivations assuming a pure H2 vapour in In the 918 m sample, 2 profiles were collected across multiple
6
Fig. 3. Fluid inclusion thermometry data for the samples at 918 m and 1319 m depths. A: homogenisation temperature (TH) histogram of the water inclusions in
quartz veins and in microcracks in quartz crystals in the sample from 918 m. B: homogenisation temperature histogram in quartz, calcite, and fluorite cements in
veins in the sample from 1319 m. C: cross-plot of water inclusion homogenisation temperature and total salinity data measured in fracture-fill cements and
microcracks in quartz crystals.
fractures. The magmatic domains show a mean δ18O of 9.0 ± 1.7‰ (n = By using the fractionation factor for oxygen isotopes between water
21), lower than in the fractures, characterised by a mean δ18O of 22.6 ± and quartz, the temperature estimated from the primary fluid inclusions,
4.8‰ (n = 22) (Fig. 7). and the oxygen isotope composition of quartz in the same veins, we
The temperature dependence of oxygen isotope fractionation be derived the oxygen isotope value of water in equilibrium with quartz
tween quartz and water has been established through modelling and cement in the fractures (Fig. 8). In the 1319 m sample, the mean δ18O
experiments (Pollington et al., 2016; Vho et al., 2019). Magmatic quartz value of quartz in fractures from profile C is 11.2 ± 0.3‰. Given the
from 1319 m and 918 m depths return mean δ18O values of 9.5‰ and temperature range of 210–244 ◦ C of the fluid inclusions in quartz at this
9.0‰, respectively. At magmatic temperature of the granite emplace location and applying pressure corrections for hydrostatic and litho
ment, isotope fractionation between phases is neglectable and O iso static regime to those temperatures (details on the pressure correction
topes recorded in the magmatic quartz grains can be used as an estimate can be found in Supplementary Material), the range of δ18O values of
of the bulk granitic composition. water in equilibrium with the fracture quartz can be derived and is found
7
(caption on next page)
8
Fig. 4. Gas inclusion assemblage in quartz microcracks. A: view of the magmatic quartz grain in transmitted light optical microscopy. B: SEM image in cath
odoluminescence of the yellow rectangle area in A showing thin healed cracks in magmatic quartz containing gas inclusion FIAs. C: gas inclusion FIA (FIA1) in
transmitted light optical microscopy photomicrograph. D: monophasic (L), diphasic (L-V) and gas inclusion FIA (FIA2) in transmitted light optical microscopy
photomicrograph. E: Raman spectra of H2 region near 4156 cm− 1 showing xenon (Xe) calibration lamp peak positions (grey vertical lines). (For interpretation of the
references to colour in this figure legend, the reader is referred to the web version of this article.)
to be from − 2.5 to 1.5 ‰ (Fig. 8). Furthermore, the total salinities of 0.2 with the deeper sample towards a higher nucleogenic contribution.
fluid inclusions range between 5.4 and 6.2 wt%, which are some of the Data obtained from separated quartz fractions are located along
lowest values in this study (Fig. 3). The range of values obtained for the different nucleogenic lines with smaller 21Ne/22Ne ratios. Isotope dif
parent water involved in the quartz cement precipitation, which is ferences of samples from different depths decrease (samples become
around 0‰, suggests a marine signature (considering a constant more similar in triple Ne isotope composition and plot closer together)
seawater δ18O over time, Tartèse et al., 2017), and is consistent with the the more purified the quartz fraction is. The petrographic observations
relatively low fluid inclusion salinities. and the Tima mineral maps showed that zircon and apatite, commonly
In the sample from 918 m, quartz cements in fractures are charac containing the alpha particle emitters uranium and thorium, were
terised by a mean δ18O of 22.8 ± 4.6‰, and of 23.5 ± 1.4‰ in profile C. spatially associated with iron oxides and chamosite. These mafic en
Given the temperature range for the fluid inclusions on profile C cement claves were quantified using the TESCAN mineral association algorithm
of 54–84 ◦ C and their pressure correction applied for hydrostatic and as more often surrounded by feldspar than quartz (see Supplementary
lithostatic conditions, the δ18O of water in equilibrium with the fracture Material). These observations explain that the spread in results of triple
quartz can be calculated and is evaluated as ranging from − 2.5 to 3 ‰. Ne isotope ratios from bulk granite grains is larger along the nucleogenic
The range of values obtained for the parent water involved in the quartz lines as compared to those from crushed purified quartz only. This data
cement precipitation in this sample also suggests a marine source. The distribution indicates that the nucleogenic neon extracted from fluid
total salinities of fluid inclusions in the profile C cement range between inclusions is produced in-situ while the fluid is already trapped as fluid
19 wt% to 23 wt%, which are some of the highest values in this study inclusion. This leads to a higher amount and isotopic variety of
(Fig. 3), indicating that the fluid has chemically evolved from a marine nucleogenic neon in granite aggregates due to the statistically closer
seawater salinity. proximity of fluid inclusions in feldspar to alpha emitters contained in
the mafic enclaves. The purified and crushed quartz fraction suggests a
6. Argon and triple neon isotope systematics of fluid inclusions dominated nucleogenic production of 22Ne that may be explained from
in bulk granite and separated quartz grains for fluid age fluorine-rich brine trapped in the inclusions in quartz and may refer to
characterisation early water in the granite. This implies that quartz better preserved the
fluid inclusions at both depths and is less affected by mineral alteration.
Granite samples from depths of 918 m and 1319 m were analysed for Nucleogenic production in this study shows different ratios compared to
the stable noble gas isotopes of neon and argon contained in fluid in those found in Witwatersrand (Lippmann-Pipke et al., 2011) and
clusions. We analysed (1) the sieved granite fraction of 1 to 0.85 mm, (2) average crust. The granite composition in our study has much higher
hand-picked and chemically cleaned quartz grains of the 1 to 2 mm content in F (620 ppm at 1285.2 m, 1380 ppm at 1504.2 m, 1200 ppm at
sieved fractions, and (3) chemically separated quartz grains of the 0.85 1908.1 m, SARIG data) compared to the 48–346 ppm values from
to 0.71 mm sieved fraction. Based on the microscopy observations, the Lippmann-Pipke et al. (2011). This difference may be critical to explain
quartz grains contain multiple fractures filled by secondary quartz the lower 21Ne/22Ne and 20Ne/22Ne ratios in our study, with a relative
cement, characterised by a darker luminescence than the magmatic higher production of 22Ne.
quartz in SEM-CL. Therefore, the crystals contain a variety of fluid in Argon has three natural isotopes: the radiogenic isotope 40Ar pro
clusion entrapment events. duced by 40K decay and the two primordial isotopes 38Ar and 36Ar
The neon isotope system consists of three nucleogenic isotopes: 20Ne without radiogenic sources. In contrast to the neon isotopes, radiogenic
(produced by 17O(α, n)20Ne), 21Ne (produced by 18O(α, n)21Ne or 24Mg 40
Ar production depends only on the concentration of 40K, not on any
(n, α)21Ne) and 22Ne (produced by 19F(α, n)22Na(β+)22Ne or 25Mg(n, other target element for nucleogenic production. The 40Ar/36Ar ratio of
α)22Ne) (Kennedy et al., 1990). The greater the nucleogenic component, bulk granite is much greater in the deeper sample at 1319 m (7200 and
i.e. the greater the distance between the triple neon isotope composition 8240, respectively, for two aliquots) than in the shallower sample (5400
and the isotopic composition of air in Fig. 9A, the older the fluid. The and 5500 for two aliquots) (Fig. 9B). The extracted quartz fractions show
slope of the nucleogenic lines (after Kennedy et al., 1990; Ballentine and much smaller 40Ar/36Ar ratios (1100 and 1400, respectively), not un
Burnard, 2002; Kendrick and Burnard, 2013) depends on the concen surprising due to the trace levels of potassium in the quartz as compared
tration of uranium and thorium and of F, O and Mg in very close vicinity to the high K contents in feldspar of the bulk granite.
(μm) to the fluid inclusion. For these reactions uranium and thorium Both Neon and Argon isotope data indicate that the 918 m granite
deliver the alpha particles and neutrons (α and n respectively in the sample has a larger admixture of younger secondary fluids compared to
nucleogenic formulas above). However, alpha particles in rock have a the 1319 m granite sample. This suggests more recent rock alterations at
reach of only few μm. Therefore only if both the U and Th and the target 918 m, consistent with observations (i) from CL imaging and fluid in
nuclides of F, O and Mg are in the fluid inclusion or less than a few clusion data that multiple fluid circulation and cement growth events, at
micrometres away from the fluid inclusions, they can contribute to the lower temperatures have occurred at 918 m, (ii) a higher level of
neon isotopes in the fluid inclusions. oxidation recorded by the iron oxide in the 918 m sample where
The smaller the x-axis intercept value of the lines is, the more magnetite has been replaced by hematite.
dominant is the nucleogenic pathway producing 22Ne. Pure fluorite (not
containing any Mg) would produce no 21Ne but only 22Ne and 20Ne 7. Implications for the geogenic hydrogen system of granites
(Kendrick et al., 2011; Kennedy et al., 1990) and plot on a vertical line
with older samples downward, therefore result in an intercept close to 7.1. Radiolitic H2
that of air (around 0.03). However, due to this very local nuclide pro
duction, variations between different measured aliquots and different The decay of uranium, thorium, and potassium in granite produces
mineral fractions are expected. The data points from the bulk granite radiogenic helium and argon, and hydrogen and oxygen by the radiol
crush plot along a nucleogenic production line with a 21Ne/22Ne ratio of ysis of the contained water. Research by Lin et al. (2005), Sherwood
9
Fig. 5. A: transmitted light photomicrographs of gas inclusion and water inclusion assemblage in fluorite fracture-fill cement in granite sample from 1319 m (shown
in Fig. 2). B: calibration data of the 4156 cm− 1 peak shift with pressure. C: Raman spectra of H2 region near 4156 cm− 1 showing Xenon calibration lamp peak
positions (grey vertical lines).
10
Fig. 6. Scanning electron microscopy images in back-scattered electron (BSE) mode (top left) and cathodoluminescence (CL) mode (middle left) placing the oxygen
isotope measurement points in areas A to E along the quartz cemented fracture in sample from 1319 m. Orange, blue and green dots indicate δ18O measurements
spots in, respectively, magmatic grain, cement and across magmatic grain and cement. The values of δ18O are plotted in the bottom left diagram. Photomicrographs
in transmitted white light on the right show primary aqueous inclusions with their homogenisation temperatures (TH) and total salinities (S); scale bars are 10 μm.
(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Lollar et al. (2014) and Warr et al. (2019) provides information on the standard pressure) over the period of 1.59 Ga (details are given in
production of hydrogen, noble gas and H2/4He ratios that can be ex Supplementary Material). The pore water salinity would increase, from
pected from felsic rocks, based on various Canadian Precambrian the disappearance of water due to radiolysis over the 1.59 Ga, from 5.0
granites. The Roxby Downs granite is comparatively richer in radioac wt% to 5.4 wt% (NaCl eq.). The salinity increases to values above 20 wt
tive elements uranium, thorium and potassium, with respective average % (NaCl eq.) observed in the fluid inclusion record in the cements can
values of 16 ppm, 59 ppm and 5.6 wt% (SARIG public data in Blanche 1), therefore not be explained by the radiolysis of the pore water alone.
but lower porosity (see further) than any of the examples from Warr The hydrogen production rate by radiolysis is increased at high pore
et al. (2019). However, the abundances in those elements may not be water salinity. The fluid inclusions in cements showed increasing
original and rock alteration and enrichment processes may have affected salinity with cooling of the rock to values above 20 wt% (NaCl eq.).
their concentrations, potentially altering the H2 production by radiol Using the modified radiation chemical yield (GiH2) parameters for 5 M
ysis. According to Dmitrijeva et al. (2019), K and Th are both typical of NaCl solution (22.6 wt% NaCl) from Tarnas et al. (2018), we derived
non-mineralized felsic and mafic intrusive rocks. Therefore, their con values of production of 9.9 × 10− 9 mol.m− 3.yr− 1. This new rate would
centrations are unlikely to have been substantially changed. The values create 15.7 mol.m− 3 or 352 L.m− 3 (at standard pressure) over the period
of U and Th indicate highly radioactive granite, but the Th/U ratio is on of 1.59 Ga. It is likely that the salinity increase seen in the fluid inclusion
average 4.2, consistent with unaltered values. Furthermore, Skirrow record occurred over a large period of time. Therefore, the H2 produc
et al. (2007) dated the ages of hydrothermal minerals in IOCG alteration tion rate by radiolysis can be estimated to be initially of 7.1 × 10− 9 mol.
assemblages in the Olympic Cu-Au-(U) province from 1598 ± 7 to 1575 m− 3.yr− 1 (based on an initial low salinity), and increasing to 9.9 × 10− 9
± 7 Ma. Ciobanu et al. (2013) dated U-bearing hematite ages of 1590 ± mol.m− 3.yr− 1 (based on 22.6 wt% NaCl). These rates fall within the
8 Ma and 1577 ± 5 Ma at Olympic Dam. Overall, those dates support the higher end of the range of radiolytic rates calculated by Warr et al.
views that mineralization is coeval with the emplacement of the Gawler (2019) for rocks in the Canadian Precambrian Shield, which range from
Range Volcanics and associated Hiltaba Intrusive Suite. Therefore, we 2.2 × 10− 9 to 1.1 × 10− 8 mol.m− 3.yr− 1. Additionally, they exceed the
assume that the contents of radioactive elements at Blanche 1 are close values reported by Boreham et al. (2021) for the Hiltaba Suites granite
to the original, on average, and if enriched, the enrichment may be close locations, which range from 4.3 to 5.4 × 10− 9 mol.m− 3.yr− 1.
to the age of the pluton. The radiolytic production of hydrogen can also be examined by the
Considering pre-decay U and Th abundances derived from the values contributions of the different elements (details are given in Supple
above, and the K abundance (K is kept at 5.6 wt% as 40K is calculated mentary Material). For pure water, uranium (U) accounts for 45% of the
from the 39K/40K abundance ratio), the gas porosity measurement of hydrogen production, thorium (Th) accounts for 38%, and potassium (K)
Esteban et al. (2022) at Blanche 1 in the granite of 0.3% at about 1319 accounts for only 17%. However, when 5 M NaCl brines are subjected to
m, an initial water salinity at 5.0 wt%, based on the fluid inclusions data, radiolysis, the values shift to 28%, 54%, and 18%, respectively, high
we derive a production of H2 of 7.1 × 10− 9 mol.m− 3.yr− 1 (160 L.km− 3. lighting the significant role of Th in highly saline brines such as those
yr− 1) which equates to a production of 11.3 mol.m− 3 or 253 L.m− 3 (at found in some parts of the Roxby Downs granite. It should be noted that
11
Fig. 7. Scanning electron microscopy images in back-scattered electron (BSE) mode (top left) and cathodoluminescence (CL) mode (top middle) placing the oxygen
isotope measurement points in areas A to D along a quartz and anhydrite cemented fracture in sample from 918 m. Orange, blue and green dots indicate δ18O
measurements spots in, respectively, magmatic grain, cement and across magmatic grain and cement. The values of δ18O are plotted in the bottom diagram.
Photomicrographs in transmitted white light on the right show primary aqueous inclusions with their homogenisation temperatures (TH) and total salinities (S). Blue
dashed lines indicate transition between CL zones in quartz cement. δ18O measurements spots of transects C and D are positioned on the photomicrographs. (For
interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
these values assume a homogeneous distribution of elements in the rock, total oxidation of the iron induced the precipitation of magnetite or
which may be more accurate for K than U and Th. Therefore, the hematite, with increasing amounts of hydrogen production. In contrast,
contribution of K may be underestimated. the absence of oxidation resulted in the precipitation of chamosite
With a H2 solubility in pure water below 0.1 mol/kg at 100 bar and without hydrogen production.
temperatures below 100 ◦ C, and even lower solubilities in brine, the At Blanche 1 in the Roxby Downs granite, biotite relics are commonly
maximum total radiolytic H2 production is not dissolvable in the limited replaced by chamosite, muscovite (or sericite), with minor titanium
volume of pore water per cubic meter of rock (at 0.3% porosity), sug oxide, quartz, and rare fluorite (Fig. 10). Iron oxides are often found in
gesting that, overtime the appearance of a hydrogen vapour phase, also association with chamosite, as well as apatite and zircon as compact
containing noble gases, is expected due to water radiolysis alone. clusters (Fig. 10A and C), designated as mafic enclaves by Krneta et al.
(2017). Their distribution suggests grain replacements of a pre-existing
mafic or iron-rich minerals containing some calcium and some zirco
7.2. Hydrolytic H2 nium, possibly a ferro-hornblende such as hastingsite (NaCa2(Fe2+ 3+
4 Fe )
(Si6Al2)O22(OH)2), as described in the Proterozoic rapakivi granite by
Iron-rich felsic rocks such as Precambrian Roxby Downs granite are Suikkanen et al. (2019), or arfvedsonite (Na3(Fe2+ 3+
4 Fe )Si8O22(OH)2) as
expected to produce additional hydrogen through water-rock interac investigated by Truche et al. (2021).
tion (WRI) such as hydration of Fe2+-bearing minerals (Sherwood Lollar The results of the simulations of Murray et al. (2020) on the granite
et al., 2014; Warr et al., 2019; Truche et al., 2021). A study by Murray from of the Soultz-sous-Forêts suggest that the petrography observed at
et al. in 2021 simulated the dissolution of biotite in the granitic base Blanche 1 may be the result of two stages of hydrothermal alteration
ment of the Soultz-sous-Forêts geothermal site in the Upper Rhine process. First, a reducing fluid would result in chamosite precipitation,
Graben, France. The simulations used brine (8.5 wt% Eq. NaCl) at 165 ◦ C followed by a more oxidant fluid that would cause magnetite
with different oxidation potentials. The results showed that moderate to
12
Fig. 8. Diagram of the quartz-water oxygen

isotope fractionation as a function of tem
perature (Pollington et al., 2016). The curves
represent isovalues of δ18O of the quartz
formed from parent waters with δ18O values
displayed on the x-axis and at temperature
conditions on the y-axis. The coloured areas
are showing the temperature of the primary
fluid inclusions near the transects C across
fractures in samples 1319 m (blue) and 918
m (orange). Using the values measured in the
cements along the transects and the fluid
inclusion homogenisation temperatures
(grey areas) with applied pressure correc
tions, the ranges of δ18O of the water are
graphically constrained. The zones a and b
refer CL zones as shown in Fig. 5. (For
interpretation of the references to colour in
this figure legend, the reader is referred to
the web version of this article.)
Fig. 9. Neon three-isotope plot (A) and Argon three-isotope plot (B) for Blanche 1 from depths of 918 m and 1319 m: (A) Nucleogenic neon 21Ne/22Ne (intersection
point with the x-axis at 20Ne = 0) is discussed having ratios of 0.47–0.52 (Kennedy et al., 1990; Ballentine and Burnard, 2002; Kendrick and Burnard, 2013). All
nucleogenic lines start at air composition. Solar wind and MORB lines refer to Graham (2002). A crustal line from fluid inclusion measurements in vein quartz from
Witwatersrand by Lippmann et al. (2003) is shown for comparison. (B) A triple Ar isotope plot similar to Seltzer et al. (2021) shows granite samples and quartz
samples plot along the radiogenic line of 40Ar production by 40K decay with small deviations of the 38Ar/36Ar ratio from air resulting from mass fractionation during
the gas purification process.
precipitation. The fluid induced a transition from a low oxidation po minium (Pb2+₂Pb4+O₄) precipitation in absence of organic matter.
tential associated with the mafic grain destabilization and replacement Therefore, water radiolysis and its production of oxygen can initiate
by chamosite, to high oxidation potential fluid responsible for the mafic different reactions of oxidation of Fe2+ into Fe3+ (Dubessy et al., 1988).
grains replacement into iron oxides and associated with generation of The production of oxygen by water radiolysis may play a role in the shift
hydrogen. The petrographic assessments of the rock samples showed from low to higher oxidation potential of the fluid. However, based on
that iron oxides consisted in most instances of clusters suggesting grain the water radiolysis rate previously calculated for our samples and on
replacements. Simple mass balance calculations, if all the magnetite their magnetite contents (1.5 wt%), the maximum radiolitic oxygen
(1.5 wt%) has been produced by the alteration of a ferro-hornblende, placed in magnetite is 1.5 wt% to 2 wt%, considering pure water or 5 M
suggest an initial abundance of the amphibole in the granite of 5.8 wt NaCl brine respectively. Therefore, another source of oxygen is required,
%, in the case of hastingsite, and 4.4 wt% in the case of arfvedsonite. that we can assume to be the water itself, leading to hydrogen produc
Oxygen associated with H2 production by water radiolysis was tion via WRI, following the reaction (1).
shown by Dubessy et al. (1988) and Savary and Pagel (1997), based on
Fe2+ (aq) + H2 O(l) = Fe3+ (aq) + H2(g) + OH − (1)
fluid inclusion analysis in natural fission reactors. The oxidation state of (aq)
the fluid increases as a consequence of water radiolysis. They observed Replicating the procedure of Truche et al. (2021), the mass balance
that oxygen was in most cases rapidly consumed by hematite and of H2 production from FeII oxidation into FeIII in a cubic meter of rock
13
Fig. 10. Scanning electron microscopy images (A, B and C) in back-scattered electron (BSE) mode with false colours for Si, Al, Mg, Fe and K, showing the mineral
assemblages resulting from the alteration of mafic minerals such as biotite and possibly ferro-hornblende. The chamosite-sericite and magnetite by-products suggest a
change in the oxidation potential of the fluid. Image D shows a reflected light optical photomicrograph of the yellow contoured area in C, illustrating the magnetite
alteration in hematite. Apt: apatite; Qz: quartz; Zr: zircon. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version
of this article.)
containing 6 wt% of ferro-hornblende can produce a maximum of 171 production from the oxidation of ferrous amphibole to magnetite and
mol of H2 if all the FeII is oxidized into FeIII following the reaction (1). from magnetite to hematite, aligning with petrographic observations in
We should subtract to that amount the potential FeII oxidation into FeIII the rock samples. Warr et al. (2019) presented H2 production values
using the radiolytic oxygen, a reaction that does not produce hydrogen, from WRI in various rock types, including felsic (granitic), mafic crys
leaving a maximum amount of H2 production to about 160 mol per cubic talline, sedimentary, and metamorphic rocks, based on noble gas (He
meter of granite, and an equal amount of consumed H2O. and Ar), hydrogen, nitrogen, and hydrocarbon gas analyses from the
At 918 m, the clusters of iron oxides are hematite with relics of Canadian Precambrian Shields. We converted their WRI H2 production
magnetite, that can be interpreted as evidence of replacement of the values, expressed as mol.m− 3.yr− 1, into mol.m− 3 (total production over
magnetite by hematite, or martitisation. Geymond et al. (2022) exper the age of the rock) by multiplying them by the age of the rocks. The
imentally produced hydrogen during magnetite oxidation by water at calculated H2 production range from their data varied from 2.2 mol.m− 3
low temperatures (80 ◦ C). Therefore, this further oxidation of magnetite to a maximum of 120 mol.m− 3. Therefore, while our values are higher,
to hematite in some zones within the Roxby Downs granite can have they are not significantly different from the maximum mass-balanced H2
taken place at near current day temperature. This process of magnetite production rates. Another study by Albers et al. (2021) investigated
oxidation by water provides an additional source of production of serpentinization in ultramafic Mid-Ocean Ridges (MORs) and reported
hydrogen, following the reaction (1). The mass balance of H2 production H2 yields of 420 to 1050 mol.m− 3 at high temperatures (>200 ◦ C) and
would lead in this case to a maximum of 85 mol of additional H2 per 175 to 525 mol.m− 3 at low temperatures (<200 ◦ C) (assuming a rock
cubic meter of rock and an equal amount of consumed H2O. The total density of 3.5 kg.dm− 3). Comparing these data, the rates calculated from
amount of H2 production considering radiolytic and WRI sources are the Roxby Downs granite are lower, as expected, but surprisingly high
therefore expected to be 256 mol.m− 3 (5828 L.m− 3). This rate is for felsic rocks.
considering that a portion of the FeII oxidation into FeIII proceeds with The consumption of water over time by reaction (1) is a driver for the
radiolytic oxygen rather than oxygen from the water. The calculated pore water salinity to increase. If we still consider a fixed pore volume of
values of 160 mol.m− 3 and 265 mol.m− 3, derived from mass balance 3 L per cubic meter of rock (0.3% porosity), with a rate of H2 production
calculations assuming complete reactions, represent the maximum H2 of 160 and 86 mol.m− 3, the FeII oxidation into FeIII by water would
14
consume the entirety of the water molecule stock. The fluid inclusion The apatite fission track data at ~430–400 Ma and ~ 350–330 Ma are
analysis in the fracture-fill quartz cement, calcite and fluorite, and in the interpreted to be caused by the Alice Springs Orogeny. They also define a
magmatic quartz microcracks revealed that over time salinity increased central ‘young’ thermal corridor where Blanche 1 is located and where
to halite saturation (Fig. 3) and remained near halite saturation at thermal events occurred in more recent times (~200 Ma). While we
temperatures from 150 ◦ C down to 50 ◦ C in the sample from 918 m cannot date the fluids or the cemented fractures and microcracks, the
depth. fluid inclusion analysis within the quartz cements and microcracks at
Similar salinity increase calculations were made by Brady et al. 918 m indicates low temperatures to current conditions. This paleo-fluid
(2019) involving the alteration of biotite into chlorite in granite. The record represents a young fluid entrapment, which commonly contains
alteration of biotite into chlorite consumes water which in turn raises dissolved hydrogen in water or immiscible free hydrogen vapour.
the concentration of dissolved species. These two processes illustrate Therefore, despite the relatively young fluid entrapment, hydrogen is
that water consumption through water-rock interaction is likely com still contained in the fluid, and the granite actively produces and con
plex and salinity increase in pore water alone cannot be a diagnostic for tains hydrogen generated through both radiolysis and WRI.
hydrogen generation. However, hydrogen-bearing gas inclusion assem As the fluid cooled, it initiated a phase separation or boiling, liber
blages with little amount of salt saturated water, to no visible water ating a free vapour phase containing hydrogen, water vapour, and noble
trapped along microcracks in quartz grains also reveal the appearance of gases. The gas and water inclusion assemblages in fluorite suggest that
a free H2 vapour phase in the granite pore space. In few instances, some this phase separation began between 155 ◦ C and 170 ◦ C. The gas and
water inclusions contained large solids which very likely be halite. water inclusion assemblages in quartz microcracks at 918 m showed that
Those observations and their interpretations tend to confirm the disap free gas was present at the current-day temperature of about 50 ◦ C. This
pearance of the pore water being completely consumed by the WRI to fluid phase separation is likely promoted during the cooling of the fluid,
form magnetite and hematite and form H2, and by other possible mineral which contained a significant amount of dissolved gas responsible for
alteration (biotite to chlorite). passing the bubble point curve of the fluid (see Fig. 11A).
One possible factor that may influence the release of hydrogen
7.3. Geofluids evolution in the granite vapour from the fluid during boiling is a transition from a lithostatic
pressure regime to a hydrostatic pressure regime, as illustrated in
The fracture cements and microcracks, along with their fluid inclu Fig. 11B. As the pore fluid pressure decreases, the ability of the fluid to
sion temperatures, salinities, and hydrogen contents, describe a cooling contain dissolved gas decreases as well. This could result in the fluid
fluid with increasing dissolved species. Despite differences in the tem passing the bubble point, transitioning from a monophasic liquid to a
perature and salinity of the fluid inclusions, the oxygen stable isotope diphasic liquid-vapour state. The pressure readings from the fluid in
signatures of the quartz cements in the two samples suggest a common clusions in fluorite at 1319 m and quartz microfractures at 918 m sug
marine parent water. This exotic fluid within a granitic pluton implies a gest that the boiling was likely governed by hydrostatic pressure, and
metasomatic or hydrothermal alteration event introducing seawater into there is no evidence to suggest that a pressure regime transition influ
the pluton, likely a very long time ago, as indicated by the neon and enced the process. However, the cooling of the fluid is clear in the fluid
argon isotopes. The existence of hydrothermal alteration involving inclusion data.
marine water has been already invoked in the geochemical in
vestigations of IOCG deposits by Bastrakov et al. (2007) and Barton 7.4. Associated gases
(2014). The source of the seawater may originate from the intruded
sedimentary Paleoproterozoic siltstone-dominated metasediments of the The methane detected in the gas and water inclusion assemblage in
Wallaroo Formation (Bastrakov et al., 2007). The persistence of the fluorite at entrapment temperatures above 150 ◦ C reveals the occur
same marine signature at a range of temperatures, even after significant rence of Fischer-Tropsch or Sabatier reactions, which are abiotic pro
cooling of the pluton, suggests a long fluid residence time. Additionally, cesses involving CO2 or CO and H2 consumption (Neubeck et al., 2011;
the increase in salinity resulting from water consumption during water- Monnin et al., 2021). The presence of calcite in the same veins as the
rock interaction such as the alteration of biotite to chlorite, amphibole to fluorite confirms the equilibrium of the fluid with carbonates, which can
magnetite and magnetite to hematite, supports the interpretation of a act as sources or sinks of carbon. Neubeck et al. (2011) showed exper
closed system after the hydrothermal water influx. The oxidation of imentally the abiotic production of methane in the presence of CO2 and
magnetite to hematite, and the presence of barite (Ba-sulfate) in the 918 mafic grain alteration and/or magnetite as a catalyst, or source of H2 by
m sample suggests that at this depth a subsequent higher oxidation Fe2+ to Fe3+ oxidation with water (Truche et al., 2021; Geymond et al.,
potential was reached compared to the rock at 1319 m depth. 2022). Wan et al. (2021) showed that this equilibrium can be reversed
The analysis of neon and argon isotopes from fluid inclusion extracts by methane oxidation into CO2 with iron oxide reduction. The 918 m
from crushed granite fragments shows that the noble gas signatures at sample contains barium sulfates which has often been reported in IOCG
918 m suggest poorer fluid isolation compared to the 1319 m granite deposits. We can speculate on the influence of a sulfur system in
sample. The crushed quartz grains though seem to suggest a similar fluid conjunction with hydrogen vapour. H2 can react with sulfate minerals to
isolation between the two samples. The difference in noble gas records form H2S (Truche et al., 2009) or with sulfide and promote the disso
from the fluid inclusions is explained by the feldspar contribution, which lution of secondary minerals such as calcite by lowering the pH, that
can be interpreted as having suffered stronger alteration compared to may create porosity. These processes are beyond the scope of this study
quartz. The poorer isolation could indicate more intense alteration, or but highlight the importance of the associated dissolved gases, such as
additional younger rock alteration, recorded by a larger proportion of CO2 or H2S, and the oxidation potential of the fluid in the presence of
the inclusions in feldspar at 918 m than at 1319 m, as evidenced by the iron oxides, which can consume H2 or compromise the hydrogen pro
neon and argon isotope analysis. duction potential. They are also critical in understanding the fluid’s
Hall et al. (2018) measured ages of thermal events in this area and ability to create or plug pore space to contain hydrogen.
defined thermal activities recorded by K-feldspar 40Ar/39Ar ages
(~350–150 ◦ C) that are likely related to hydrothermal alteration at 8. Conclusions
~1000–650 Ma in the Neoproterozoic Adelaide Rift Complex. Apatite
fission track (~60–120 ◦ C), zircon (U-Th-Sm)/He (~180–200 ◦ C), and Roxby Downs granite was investigated to constrain the mechanisms
apatite (U-Th-Sm)/He (~45–75 ◦ C) ages reveal regional low tempera involved in the production of hydrogen trapped in fluid inclusions in
ture thermal events at ~1000 Ma, ~430–400 Ma, 350–330 Ma, and ~ veins. The fluid inclusion dataset derived from cemented fractures de
200 Ma, in addition to localised thermal events during the Cretaceous. scribes the evolution of cooling metasomatic fluids within the granite,
15
Fig. 11. Temperature versus depth diagram showing the present-day temperatures measured at Blanche 1, the primary and secondary fluid inclusion temperatures
measured in fracture-fill cements in samples from 918 m and 1319 m (black dots), the liquid-vapour saturation curves for saline solutions (0.5 to 25 wt% NaCl) and
the bubble point curves or isopleths for water containing 0.5 mol% and 1 mol% dissolved H2 (from Seward and Franck, 1981). Diagram A illustrates the cooling effect
that induces a reduction of hydrogen solubility in water for a given depth and may result in passing the bubble point of the pore fluid inducing phase separation.
Diagram B illustrates the pore fluid pressure drop associated with transitioning from a lithostatic to a hydrostatic pressure regime, and the H2-water isopleth shifts to
deeper depth. For a given depth, a pressure drop induces a decrease in hydrogen solubility in water which may result in passing the bubble point of the pore fluid
inducing phase separation.
starting from temperatures above 250 ◦ C, moderately saline and devoid The hydrogen may also be emanating from the granite to the surface
of detectable hydrogen, to the current temperature of 55 ◦ C and asso over time. Despite all these scenarios, free hydrogen vapour phase was
ciated with salt saturation, dissolved H2, and free H2 vapour. The for still present in fluid inclusions along microcracks in quartz and trapped
mation of an H2-rich vapour phase was initiated by the lowering of H2 at current-day temperatures. Precambrian granites that have undergone
solubility in the water due to a cooling of the granite and the salinity early hydrothermal alteration mobilising water from older sedimentary
increase. units such as those near the IOCG deposits may be particularly fertile for
The δ18O values of the quartz cements were consistent with parent hydrogen generation due to their enrichment in uranium, thorium and
water of marine origin, and the metasomatic fluids were possibly potassium, and more importantly by adding water in the plutons, a
following a fracture network as pathways into the granite, lending necessary condition to sustain water-rock interaction.
support for the proposed origin of H2 in the H2-rich gas recovered from
the Yorke Peninsula Ramsay Oil Bore 1 (Boreham et al., 2021). The Declaration of generative AI and AI-assisted technologies in the
circulation of these secondary fluids was possibly triggered with major writing process
structural events that affected the area. The preservation of the δ18O
marine signature of the fluid along with increasing salt concentration During the preparation of this work, the authors used chatGPT (GPT
and low temperatures indicated that the fluid was not being replaced at 3.5) in order to increase readability. After using this tool, the authors
a fast rate and that the fluid had a long residence time in the granite. reviewed and edited the content as needed and take full responsibility
However, the analysis of neon and argon isotopes from fluid inclusion for the content of the publication.
extracts from crushed granite fragments shows that the noble gas sig
natures at 918 m suggest poorer fluid isolation compared to the 1319 m
granite sample. Declaration of Competing Interest
We identified multiple sources of hydrogen, including radiolytic,
oxidic hydrolysis of mafic minerals; and low temperature oxidic The authors declare that they have no known competing financial
magnetite martitisation. The radiolytic H2 production rates in the Roxby interests or personal relationships that could have appeared to influence
Downs granite are between 7.1 × 10− 9 mol.m− 3.yr− 1 to 9.9 × 10− 9 mol. the work reported in this paper.
m− 3.yr− 1 which equates to a production between 11.3 and 15.7 mol.m− 3
or 253 to 352 L.m− 3 (at standard conditions) over the period of 1.59 Ga. Data availability
Oxidation reactions of FeII to FeIII by water during the alteration of pre-
existing mafic mineral formed hydrogen, magnetite or hematite leading Data will be made available on request.
up to 268 mol.m− 3. The reactions involving the consumption of water
drive an increase in pore water salinity. Therefore, the fluid inclusion Acknowledgments
revealing large salinity increase to halite saturation over time, remain
ing consistently high at cooling temperatures from 150 ◦ C to 50 ◦ C, the We would like to extend our appreciation to the reviewers, Laurent
dissolved hydrogen in water inclusions, the hydrogen gas inclusions and Truche, Chris Boreham, and an anonymous reviewer. The authors
the presence of magnetite and hematite mineralization in the granite are acknowledge the South Australia Drill Core Reference Library for
together indicative of an active hydrogen generation system within the providing inspection (5164) and sampling authorisations (samples
Roxby Downs granite. Hydrogen may accumulate in the granite or is 3580018 and 3580022). The authors are also grateful for the CSIRO
possibly mineralogically converted or consumed by bacterial activity. Interchange program that gave the opportunity to work on this research.
We thank Mark D Raven, Peter G Self and Rodrigo Gomez-Camacho for
16
the XRD mineralogical analysis, and Jelena Markov for her contribution Hrstka, T., Gottlieb, P., Skala, R., Breiter, K., Motl, D., 2018. Automated mineralogy and
petrology-applications of TESCAN Integrated Mineral Analyzer (TIMA). J. Geosci. 63
to figure sketching.
(1), 47–63.
Hut, G., 1987. Consultants’ group meeting on stable isotope reference samples for
Appendix A. Supplementary data geochemical and hydrological investigations. International Atomic Energy Agency
(IAEA).
Jagodzinski, E.A., Cordero, R.R., Polcyn, A.D., 2021. Hydrogen storage in metal-organic
Supplementary data to this article can be found online at https://doi. frameworks: a review. J. Mater. Res. Technol. 11, 1665–1679. https://doi.org/
org/10.1016/j.chemgeo.2023.121698. 10.1016/j.jmrt.2021.02.016.
Kelly, J.L., Fu, B., Kita, N.T., Valley, J.W., 2007. Optically continuous silcrete quartz
cements of the St. Peter Sandstone: high precision oxygen isotope analysis by ion
References microprobe. Geochim. Cosmochim. Acta 71, 3812–3832.
Kendrick, M.A., Burnard, P., 2013. Noble gases and halogens in fluid inclusions: A
journey through the earth’s crust. In: Burnard, P. (Ed.), The Noble Gases as
Aiuppa, A., Shinohara, H., Tamburello, G., Giudice, G., Liuzzo, M., Moretti, R., 2011.
Geochemical Tracers. Springer, Berlin Heidelberg, Berlin, Heidelberg, pp. 319–369.
Hydrogen in the gas plume of an open-vent volcano, Mount Etna, Italy. J. Geophys.
Kendrick, M.A., Honda, M., Oliver, N.H.S., Phillips, D., 2011. The noble gas systematics
Res. 116, B10204.
of late-orogenic H2O-CO2 fluids, Mt Isa, Australia. Geochim. Cosmochim. Acta 75
Albers, E., Bach, W., Pérez-Gussinyé, M., McCammon, C., Frederichs, T., 2021.
(6), 1428–1450.
Serpentinization-driven H2 production from continental break-up to mid-ocean ridge
Kennedy, B.M., Hiyagon, H., Reynolds, J.H., 1990. Crustal neon: a striking uniformity.
spreading: unexpected high rates at the West Iberia margin. Front. Earth Sci. 9
Earth Planet. Sci. Lett. 98 (3), 277–286.
https://doi.org/10.3389/feart.2021.673063.
Krneta, S., Ciobanu, C.L., Cook, N.J., Ehrig, K., Kontonikas-Charos, A., 2017. The Wirrda
Ballentine, C.J., Burnard, P.G., 2002. Production, release and transport of noble gases in
Well and Acropolis prospects Gawler Craton, South Australia: insights into evolving
the continental crust. Rev. Mineral. Geochem. 47 (1), 481–538.
fluid conditions through apatite chemistry. J. Geochem. Explor. 181, 276–291.
Barton, M.D., 2014. Iron oxide(–Cu–Au–REE–P–Ag–U–Co) systems. In: Holland, H.D.,
https://doi.org/10.1016/j.gexplo.2017.08.004.
Turekian, K.K. (Eds.), Treatise on Geochemistry, Second edition. Elsevier,
Leila, M., Levy, D., Battani, A., Piccardi, L., Segvic, B., Badurina, L., Pasquet, G.,
pp. 515–541. ISBN 9780080983004. https://doi.org/10.1016/B978-0-08-095975-
Combaudon, V., Moretti, I., 2021. Origin of continuous hydrogen flux in gas
7.01123-2.
manifestations at the Larderello geothermal field, Central Italy. Chem. Geol. 585,
Bastrakov, E.N., Skirrow, R.G., Davidson, G.J., 2007. Fluid evolution and origins of iron
120564.
oxide Cu-Au prospects in the Olympic Dam district, Gawler Craton, South Australia.
Li, L., Zhang, X., Luan, Z., Li, Y., Li, Y., 2018. Raman vibrational spectral characteristics
Econ. Geol. 102, 1415–1440. https://doi.org/10.2113/gsecongeo.102.8.1415.
and quantitative analysis of H2 up to 400◦ C and 40 MPa. J. Raman Spectrosc. 49
Boreham, C.J., Edwards, D.S., Czado, K., Rollett, N., Wang, L., Van Der Wielen, S.,
(11), 1722–1731. https://doi.org/10.1002/jrs.5420.
Champion, D., Blewett, R., Feitz, A., Henson, P., 2021. Hydrogen in Australian
Lin, L.H., Slater, G.F., Sherwood-Lollar, B., Lacrampe-Couloume, G., Onstott, T.C., 2005.
natural gas: occurrences, sources and resource. APPEA J. 61 (1), 163–191. https://
The yield and isotopic composition of radiolytic H2, a potential energy source for the
doi.org/10.1071/AJ2044.
deep subsurface biosphere. Geochim. Cosmochim. Acta 69 (4), 893–903. https://doi.
Brady, P., Lopez, C., Sassini, D., 2019. Granite hydrolysis to form deep brines. Energies
org/10.1016/j.gca.2004.07.021.
12, 2180.
Lippmann, J., Stute, M., Torgersen, T., Moser, D.P., Hall, J.A., Lin, L., Borcsik, M.,
Charlou, J., Donval, J., Fouquet, Y., Jean-Baptiste, P., Holm, N., 2002. Geochemistry of
Bellamy, R.E.S., Onstott, T.C., 2003. Dating ultra-deep mine waters with noble gases
high H2 and CH4 vent fluids issuing from ultramafic rocks at the Rainbow
and 36Cl, Witwatersrand Basin, South Africa. Geochim. Cosmochim. Acta 67 (23),
hydrothermal field (36◦ 14′N, MAR). Chem. Geol. 191, 345–359. https://doi.org/
4597–4619.
10.1016/S0009-2541(02)00134-1.
Lippmann-Pipke, J., Lollar, B.S., Niedermann, S., Stroncik, N.A., Naumann, R., van
Chen, Y., Chou, I.-M., 2022. Determination of H2 densities over a wide range of
Heerden, E., Onstott, T., 2011. Neon identifies two billion-year-old fluid components
temperatures and pressures based on the spectroscopic characterization of Raman
in Kaapvaal Craton. Chem. Geol. 283 (3–4), 287–296.
vibrational bands. Appl. Spectrosc. 76 (7), 841–850.
Maiga, O., Deville, E., Laval, J., Prinzhofer, A., Boubacar Diallo, A., 2023.
Cherry, A.R., Ehrig, K., Kamenetsky, V.S., McPhie, J., Crowley, J.L., Kamenetsky, M.B.,
Characterization of the spontaneously recharging natural hydrogen reservoirs of
2018. Precise geochronological constraints on the origin, setting and incorporation
Bourakebougou in Mali. Sci. Rep. 13, 11876. https://doi.org/10.1038/s41598-023-
of ca. 1.59 Ga surficial facies into the Olympic Dam Breccia Complex, South
38977-y.
Australia. Precambrian Res. 315, 162–178.
Mayhew, L.E., Ellison, E.T., McCollom, T.M., Trainor, T.P., Templeton, A.S., 2013.
Chou, I.-M., Burruss, R.C., Lu, W., 2005. A new optical capillary cell for spectroscopic
Hydrogen generation from low-temperature water-rock reactions. Nat. Geosci. 6 (6),
studies of geologic fluids at pressures up to 100 MPa. In: Chen, J., Wang, Y., Duffy, T.
478–484.
S., Shen, G., Dobrzhinetskaya, L.F. (Eds.), Advances in High-Pressure Technology for
Monnin, C., Dubois, M., Papaiconomou, N., Simonin, J.-P., 2002. Thermodynamics of the
Geophysical Applications. Elsevier, pp. 475–485. https://doi.org/10.1016/B978-
LiCl + H2O system. J. Chem. Eng. Data 47 (6), 1331–1336. https://doi.org/
044451979-5.50026-0.
10.1021/je0200618.
Ciobanu, C.L., Wade, B.P., Cook, N.J., Schmidt Mumm, A., Giles, D., 2013. Uranium-
Monnin, C., Quéméneur, M., Price, R., Jeanpert, J., Maurizot, P., Boulart, C., Donval, J.-
bearing hematite from the Olympic Dam Cu-U-Au deposit, South Australia: a
P., Pelletier, B., 2021. The chemistry of hyperalkaline springs in serpentinizing
geochemical tracer and reconnaissance Pb-Pb geochronometer. Precambrian Res.
environments: 1. The composition of free gases in New Caledonia compared to other
238, 129–147.
springs worldwide. J. Geophys. Res. Biogeosci. 126 e2021JG006243.
Dmitrijeva, M., Ehrig, K.J., Ciobanu, C.L., Cook, N.J., Verdugo-Ihl, M.R., Metcalfe, A.V.,
Moretti, I., Brouilly, E., Loiseau, K., Prinzhofer, A., Deville, E., 2021. Hydrogen
2019. Defining IOCG signatures through compositional data analysis: a case study of
emanations in intracratonic areas: New guidelines for early exploration basin
lithogeochemical zoning from the Olympic Dam deposit, South Australia. Ore Geol.
screening. Geosciences 11 (3), 145.
Rev. 105, 86–101.
Murray, J., Clément, A., Fritz, B., Schmittbuhl, J., Bordmann, V., Fleury, J.M., 2020.
Dubessy, J., Cathelineau, M., Poinssot, C., Fourcade, S., 1988. Radiolysis evidenced by
Abiotic hydrogen generation from biotite-rich granite: a case study of the Soultz-
H2-O2 and H2-bearing fluid inclusions in three uranium deposits. Geochim.
sous-Forêts geothermal site, France. Appl. Geochem. 119, 104631. ISSN 0883-2927.
Cosmochim. Acta 52, 1155–1167.
https://doi.org/10.1016/j.apgeochem.2020.104631.
Esteban, L., Dewhurst, D.N., Akanji, S.P., Sarout, J., Josh, M., Wilske, C., Crane, P.,
Neubeck, A., Duc, N.T., Bastviken, D., Crill, P., Holm, N.G., 2011. Formation of H2 and
Giwelli, A., Mallants, D., 2022. Influence of mineral alteration intensity of Hiltaba
CH4 by weathering of olivine at temperatures between 30 and 70 ◦ C. Geochem.
Granite and its vertical distribution on physicochemical containment for borehole
Trans. 12 (1), 6.
disposal of intermediate level waste in Australia. In: International High-Level
Neubeck, A., Duc, N.T., Hellevang, H., Oze, C., Bastviken, D., Bacsik, Z., Holm, N.G.,
Radioactive Waste Management Conference, Technical Session, Thursday,
2014. Olivine alteration and H2 production in carbonate-rich, low temperature
November 17, 2022, Kaibab. ANS Winter Meeting.
aqueous environments. Planet. Space Sci. 96, 51–61.
Frery, E., Langhi, L., Maison, M., Moretti, I., 2021. Natural hydrogen seeps identified in
Nivin, V.A., 2016. Free hydrogen-hydrocarbon gases from the Lovozero loparite deposit
the North Perth Basin, Western Australia. Int. J. Hydrog. Energy 46 (61),
(Kola Peninsula, NW Russia). Appl. Geochem. 74, 44–55.
31158–31173. https://doi.org/10.1016/j.ijhydene.2021.07.023.
Okland, I., Huang, S., Thorseth, I.H., Pedersen, R.B., 2014. Formation of H2, CH4 and N-
Gaucher, E., 2020. New perspectives in the industrial exploration for native hydrogen.
species during low-temperature experimental alteration of ultramafic rocks. Chem.
Elements 16 (1), 8–9.
Geol. 387, 22–34.
Geymond, U., Ramanaidou, E., Lévy, D., Ouaya, A., Moretti, I., 2022. Can weathering of
Pollington, A.D., Kozdon, R., Anovitz, L.M., Georg, R.B., Spicuzza, M.J., Valley, J.W.,
banded iron formations generate natural hydrogen? Evidence from Australia, Brazil
2016. Experimental calibration of silicon and oxygen isotope fractionations between
and South Africa. Minerals 12 (2), 163. https://doi.org/10.3390/min12020163.
quartz and water at 250◦ C by in situ microanalysis of experimental products and
Graham, D.W., 2002. Noble gas isotope geochemistry of mid-ocean ridge and ocean
application to zoned low δ30Si quartz overgrowths. Chem. Geol. 421, 127–142.
island basalts: characterization of mantle source reservoirs. Rev. Mineral. Geochem.
Prinzhofer, A., Tahara Cissé, C.S., Diallo, A.B., 2018. Discovery of a large accumulation
47 (1), 247–317.
of natural hydrogen in Bourakebougou (Mali). Int. J. Hydrog. Energy 43 (38),
Green Rock Energy Limited, 2006. Blanche 001 Well Completion Report, 317 pp. http
19315–19326.
s://sarigbasis.pir.sa.gov.au/WebtopEw/ws/samref/sarig1/wci/Record?r=0
Prinzhofer, A., Moretti, I., Françolin, J., Pacheco, C., D’Agostino, A., Werly, J., Rupin, F.,
&m=1&w=catno=2028485.
2019. Natural hydrogen continuous emission from sedimentary basins: the example
Hall, J.W., Glorie, S., Reid, A.J., Collins, A.S., Jourdan, F., Danišík, M., Evans, N., 2018.
of a Brazilian H2-emitting structure. Int. J. Hydrog. Energy 44 (12), 5676–5685.
Thermal history of the northern Olympic Domain, Gawler Craton; correlations
between thermochronometric data and mineralizing systems. Gondwana Res. 56,
90–104. https://doi.org/10.1016/j.gr.2018.01.001.
17
Sato, M., Sutton, A.J., McGee, K.A., 1984. Anomalous hydrogen emissions from the San conditions: Implication for the nuclear waste storage. Geochim. Cosmochim. Acta 73
Andreas fault observed at the Cienega Winery, Central California. Pure Appl. (16), 4824–4835. ISSN 0016-7037. https://doi.org/10.1016/j.gca.2009.05.043.
Geophys. 122 (2), 376–391. Truche, L., Bourdelle, F., Salvi, S., Lefeuvre, N., Zug, A., Lloret, E., 2021. Hydrogen
Savary, V., Pagel, M., 1997. The effects of water radiolysis on local redox conditions in generation during hydrothermal alteration of peralkaline granite. Geochim.
the Oklo, Gabon natural fission reactors 10 and 16. Geochim. Cosmochim. Acta 61 Cosmochim. Acta 308, 42–59. https://doi.org/10.1016/j.gca.2021.05.048.
(20), 4479–4494. Vacquand, C., Deville, E., Beaumont, V., Guyot, F., Sissmann, O., Pillot, D., Arcilla, C.,
Seltzer, A.M., Krantz, J.A., Ng, J., Danskin, W.R., Bekaert, D.V., Barry, P.H., Prinzhofer, A., 2018. Reduced gas seepages in ophiolitic complexes: evidences for
Kimbrough, D.L., Kulongoski, J.T., Severinghaus, J.P., 2021. The triple argon isotope multiple origins of the H2-CH4-N2 gas mixtures. Geochim. Cosmochim. Acta 223,
composition of groundwater on ten-thousand-year timescales. Chem. Geol. 120458. 437–461.
Seward, T.M., Franck, E.U., 1981. The system hydrogen - water up to 440◦ C and 2500 Vho, A., Lanari, P., Rubatto, D., 2019. An internally-consistent database for oxygen
bar pressure. Ber. Bunsenges. Phys. Chem. 85 (1), 2–7. isotope fractionation between minerals. J. Petrol. 60 (11), 2101–2129. https://doi.
Sherwood Lollar, B., Onstott, T.C., Lacrampe-Couloume, G., Ballentine, C.J., 2014. The org/10.1093/petrology/egaa001.
contribution of the Precambrian continental lithosphere to global H2 production. Vityk, M.O., Bodnar, R.J., 1995. Textural evolution of synthetic fluid inclusions in quartz
Nature 516 (7530), 379–382. during reequilibration, with applications to tectonic reconstruction. Contrib.
Skirrow, R.G., Bastrakov, E.N., Barovich, K., Fraser, G.L., Creaser, R.A., Fanning, C.M., Mineral. Petrol. 121, 309–323. https://doi.org/10.1007/BF02688246.
Raymond, O.L., Davidson, G.J., 2007. Timing of iron oxide Cu-Au-(U) hydrothermal Wan, Y., Bourdet, J., Hu, W., Kang, X., Heath, C., Qiu, Y., Gao, W., Wang, X., 2021.
activity and Nd isotope constraints on metal sources in the Gawler Craton, South Experimental investigations on the thermochemical oxidation of n-alkane and
Australia. Econ. Geol. 102 (8), 1441–1470. alcohol compounds by MnO2 and Fe2O3 at temperatures up to 325 ◦ C. Chem. Geol.
Steele-MacInnis, M., Bodnar, R.J., Naden, J., 2011. Numerical model to determine the 559, 119982 https://doi.org/10.1016/j.chemgeo.2020.119982.
composition of H2O–NaCl–CaCl2 fluid inclusions based on microthermometric and Warr, O., Sherwood Lollar, B., Fellowes, J., Sutcliffe, C.N., McDermott, J.M., Holland, G.,
microanalytical data. Geochimica et Cosmochimica Acta 75 (1), 21–40. https://doi. Mabry, J.C., Ballentine, C.J., 2018. Tracing ancient hydrogeological fracture
org/10.1016/j.gca.2010.10.002. ISSN 0016–7037. network age and compartmentalisation using noble gases. Geochim. Cosmochim.
Suikkanen, E., Rämö, O.T., Ahtola, T., Halkoaho, T., Hanski, E., Voutilainen, M., 2019. Acta 222, 340–362. https://doi.org/10.1016/j.gca.2017.10.022.
Clinopyroxene episyenites in a Proterozoic rapakivi granite, SE Warr, O., Giunta, T., Ballentine, C., Sherwood, B., 2019. Mechanisms and rates of helium,
Finland—recrystallization textures, mass transfer and implications for the petrology argon, and hydrogen production and accumulation in fracture fluids in Precambrian
of A-type granite complexes. Mineral. Petrol. 113 (6), 727–743. https://doi.org/ Shield environments. Chem. Geol. 530, 119322 https://doi.org/10.1016/j.
10.1007/s00710-019-00684-w. chemgeo.2019.119322.
Tarnas, J.D., Mustard, J.F., Sherwood Lollar, B., Bramble, M.S., Cannon, K.M., Wilske, C., Delle Piane, C., Bourdet, J., Suckow, A., Deslandes, A., Gerber, C., Crane, P.,
Palumbo, A.M., Plesa, A.-C., 2018. Radiolytic hydrogen production on Noachian Questiaux, D., Spooner, N., Mallants, D., 2022. Noble gas composition of deep rocks
Mars: implications for habitability and atmospheric warming. Earth Planet. Sci. Lett. to interpret provenance and residence time of fluids at a granite site in Australia. In:
502, 133–145. WM 2022 Symposia; 06 to end of 10 Mar 2022; Virtual. CSIRO; 2022. 1. CSIRO:
Tartèse, R., Chaussidon, M., Gurenko, A., Delarue, F., Robert, F., 2017. Warm Archean EP2021-1892. http://hdl.handle.net/102.100.100/437325?index=1.
oceans reconstructed from oxygen isotope composition of early-life remnants. Zgonnik, V., Beaumont, V., Deville, E., Larin, N., Pillot, D., Farrell, K.M., 2015. Evidence
Geochem. Perspect. Lett. 3, 55–65. for natural molecular hydrogen seepage associated with Carolina bays (surficial,
Truche, L., Berger, G., Destrigneville, C., Pages, A., Guillaume, D., Giffaut, E., Jacquot, E., ovoid depressions on the Atlantic Coastal Plain, Province of the USA). Progress Earth
2009. Experimental reduction of aqueous sulphate by hydrogen under hydrothermal Planet. Sci. 2 (1), 31.
18

1 s2.0 S0009254123003984 Main

Uploaded by

Copyright:

Available Formats

1 s2.0 S0009254123003984 Main

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1 s2.0 S0009254123003984 Main

Uploaded by

Copyright:

Available Formats

Chemical Geology 638 (2023) 121698

Contents lists available at ScienceDirect

Natural hydrogen in low temperature geofluids in a Precambrian granite,

1. Introduction low environmental impact. In this context, the production of low-cost

therefore formed at a minimum of 54 ◦ C to 121 ◦ C, likely during episodic

et al., 2002). The hydrohalite melting temperature necessary to calcu­

Raman spectroscopy in the vapour phase of di-phasic water inclusions at

inclusions were trapped.

Healed microcracks in quartz crystals contain several types of as­

low abundance of liquid water. Another type contained assemblages of

with high abundance of liquid water (heterogenous trapping). After

50 ◦ C and 87 ◦ C. Two salinities of 20 and 22 wt% (NaCl eq.) were

It is anticipated that H2 could diffuse from the inclusions through the

ability between inclusions of a same assemblage is often large. The cause

clusion of 81 bar along with the minimum TH of 50 ◦ C for the water

inclusions in the same assemblage, are a clear indication of entrapment

at current day hydrostatic pressure and temperature. The values around

85–92 bar) and temperatures of around 150 to 180 ◦ C.

The quartz cements in veins at 1319 m contain assemblage of di-

phasic (L-V) liquid-rich aqueous inclusions. Their homogenisation

(L-V) liquid-rich aqueous inclusions and their homogenisation temper­

clusions, the TH values within an assumed assemblage commonly show

20 to 30 ◦ C variability, with no visible morphological evidence of re-

indicate relative timing of fluid entrapment. The salinity of the aqueous

− 10.6 ◦ C. In secondary assemblages, salinities vary widely between

measured between − 26 to − 46 ◦ C, indicative of a NaCl-CaCl2-H2O mix

(caption on next page)

Fig. 8. Diagram of the quartz-water oxygen

You might also like

et al., 2002). The hydrohalite melting temperature necessary to calcu

Healed microcracks in quartz crystals contain several types of as

(L-V) liquid-rich aqueous inclusions and their homogenisation temper