Physicochemical Conditions of the Devonian-Jurassic Continental Deep Biosphere Tracked by Carbonate Clumped Isotope Temperatures of Granite-Hosted Carbonate Veins

,


Introduction
Understanding the 'deep biosphere' or microbial community living deep within sediments and igneous fracture systems is relevant to how life can be sustained in environments different from the Earth's surface, which may apply to other planets [1].Investigations over the last two decades confirmed that microorganisms thrive in oligotrophic fracture systems [2], for instance, by consuming or producing methane [3] an important constituent of fluids in fractured crystalline rocks [4].The production of microbial methane results in a kinetic effect that fractionated carbon isotope ratios (expressed in δ 13 C VPDB , [5,6]), leading to methane with low δ 13 C values [7], when oxidized in a later stage, leads to 13 C-depleted bicarbonate ions, that can be incorporated into carbonate minerals [8].Recent studies of extremely 13 C-depleted (down to -125‰ [9,10]) deep fracture-lining carbonate within granitic host rocks at Laxemar, Götemar, and Forsmark (hereafter 'LGF'), Sweden, suggest that the 13 C-depleted values formed due to the incorporation of bicarbonate into calcite from anaerobic oxidation of methane (AOM) reactions, a model supported by the analysis of fatty acids shown to be of microbial descent [9,10].Those previous publications have applied the secondary ions mass spectrometry (SIMS) technique within the same samples revisited in the current study, which thus involves samples with confirmed AOM-related 13 C-depletion.
Our goal is to investigate the LGF veins further to constrain the long-term environmental conditions of the deep biosphere processes.We apply carbonate clumped isotope thermometry on authigenic microbially induced calcite within veins in a granitic host rock.Carbonate clumped isotope thermometry is based on measuring the abundance of 13 C- 18 O bonds (Δ 47 ) in the carbonate lattice relative to a stochastic distribution of this isotopologue [11]).Carbonate clumped isotope thermometry only depends on the temperature of the fluid, not its oxygen isotopic composition, thus enabling applications to faults and fracture settings where the fluid composition is often weakly constrained [5,12,13].However, previous studies have identified various extents of isotopic disequilibrium (driving apparent high temperatures) associated with methane oxidation in cold seeps [14][15][16].Thus, the specific objectives of our study were to (a) explore whether microbial activity (methane oxidation) may have caused the disequilibrium signals in carbonate clumped isotope composition, then captured by the carbonate clumped isotope thermometry, and (b) assuming the carbonate clumped isotope temperatures accurately recorded the temperature of precipitation or no kinetic effect, to determine the temperature of the ancient microbial processes that resulted in 13 C-depleted calcite.
The LGF area is selected in part because previous work here demonstrated geochronologically constrained isotopic signs of ancient microbial activity in veins and because of the availability of fluid inclusion results [10] that can be directly compared to carbonate clumped isotope temperatures.Our results demonstrate that carbonate clumped isotope thermometry is a valuable tool for interpreting the paleoenvironment of the fossil and fracture-hosted deep biosphere.

Sites and Materials
Nine calcite samples used in this study come from the Svecokarelian orogeny unit of the Fennoscandian Shield at three different locations situated in the southeastern part of Sweden: Forsmark, Götemar, and Laxemar (Figure 1).Forsmark is situated approximately 120 km north of Stockholm, and Laxemar and Götemar are located about 300 km south of Stockholm.
The Svecokarelian unit has been divided into six tectonic domains [18] based on differences in the timing of tectonic activity or the character and intensity of the strain (Figure 1).Forsmark is situated in tectonic domain 2 [18] that contains broad belts of rocks that were affected by a strong Paleoproterozoic ductile deformation.These ductile high-strain belts, which strike approximately WNW-ESE to NW-SE, are commonly folded and were affected mainly by lower ductile strain.The belts also contain retrograde deformation zones along which deformation occurred in both the ductile and brittle regimes.The Laxemar and Götemar areas are located inside the tectonic domain 5 [18] that formed after the complex geological evolution observed at Forsmark took place, and therefore, the intrusive rocks at Laxemar and Götemar are better preserved [17].Paleoproterozoic intrusive rocks belonging to the Transscandinavian Igneous Belt (TIB) dominate the area and show variable composition (granite to quartz monzodiorite to diorite-gabbro), grain size, and texture that resulted from magmamixing processes.The Götemar granite is a Mesoproterozoic A-type granite that intruded at ~1430 Ma, and the related hydrothermal circulation resulted from the TIB wall rock, as seen by greisens and abundant veins [19].
Sample locations, including boreholes and one outcrop, are dominated by Proterozoic (1.43-1.89Ga, [10]) crystalline rocks.The samples were collected from open, semiopen, and sealed fractures (veins) and analyzed using a microthermometry technique for fluid inclusion and intracrystal SIMS analysis for carbon, oxygen, and sulfur isotopes previously [10].Several generations of fracture-infilling minerals were recognized in the study areas [20], and microscale radiometric dating by [10] puts the calcite veins primarily within the Devonian-Carboniferous periods (402-355 Ma), with some overgrowth of calcite in the Forsmark area and Götemar showing Jurassic ages which are 173:2 ± 7:6 Ma and 160:0 ± 3:3/5:6 Ma, respectively [10,21].Sample "KLX 14A 80" yields an Rb-Sr (isochrones of calcite-K-rich clay mineral pairs) age of 393 ± 15 Ma [22].During this period, the fracture approached the surface through exhumation, and extension was prevalent [23].At all sites, an early period of Proterozoic precipitation of a high-temperature mineral assemblage, which includes epidote, was followed by a Mesoproterozoic period of hydrothermal precipitation of different, lower temperature minerals, including adularia (older generation), hematite, prehnite, and calcite, followed by several pulses of abundant Phanerozoic precipitation of calcite, pyrite, adularia, clay minerals, and quartz in veins [20].

Analytical Methods
Carbonate clumped isotope measurements were performed in the Stable Isotope Laboratory at Imperial College London.Powdered calcite samples were used to perform 3-7 replicate measurements per sample.Each replicate measurement consisted of eight acquisitions with seven cycles per acquisition.For five out of nine samples, each replicate analysis was purified by passage through a conventional manual vacuum line with multiple cryogenic traps and a Porapak-Q trap [5,24].For the other samples, a fully automated inhouse prototype IBEX (Imperial Batch Extraction) system was used.The IBEX uses helium as a carrier gas but a trap configuration similar to our conventional manual vacuum line.
A total of 3-3.5 mg (for automated purification line) or 6-6.5 mg (for manual purification line) calcite powders were reacted in 105% anhydrous orthophosphoric acid bath for 10 minutes under vacuum to liberate analyte CO 2 .After passing multiple traps, the purified CO 2 was processed using a modified Thermo Finnigan MAT-253 mass spectrometer with additional Faraday cups registered for ions with 44-49 m/z to enable carbonate clumped isotope thermometry measurements.

Geofluids
Data processing was done using "Easotope," a software for reducing and normalization complex isotopic data, including bulk isotopes of carbon and oxygen and carbonate clumped isotopes [25].The raw Δ 47 values were corrected for nonlinearity using the heated gas line method [26] for the samples measured on the conventional vacuum line and pressure baseline correction (PBL) for the samples measured on the IBEX [27].The nonlinearity corrected Δ 47 values were then translated into the absolute reference frame of [28]) using an internal standard (Carrara marble, "ICM", with a Δ 47 of 0.409‰), and published interlaboratory standards known as "ETH1-ETH4" ( [29,30].We calibrated the values of the ETH standards internally against equilibrated gases [31]: our values are different from reported values using a Thermo Kiel IV device ( [29,30] but similar to values analyzed on dual inlet systems ( [32,33].Finally, values for Δ 47CDES were corrected for acid digestion at 90 °C by adding 0.082‰ [34] and converted into carbonate clumped isotope temperatures (T-Δ 47 ) using a temperature calibration developed in our laboratory (½Δ47 = 0:04028 ± 0:00076 × 106/T2 + ð0:23776 ± 0:00759Þ, where T is in Kelvin; [31]).This equation includes a range of calcite precipitated between 25 and 250 °C [35] reprocessed using our internal values for the standards and the 'IUPAC-recommended 17 O correction' parameters [36] which assume the values for ratios of 13/12 C, 18/16 O, 17/16 O in VPDB standard, and a mass-dependent slope λ that relates δ 17 O and δ 18 O values (following [37]).Bulk carbon and oxygen isotopic composition was also measured in the same samples and are reported in the per mill (‰) unit on the VPDB scale.The δ 18 O water VSMOW can be calculated using the calcite δ 18 O converted to the VSMOW scale, and temperature was derived from carbonate clumped isotopes following the equation from [38] (½δ18Owater ½VSMOW = ð18:03 × 103Þ/T − 32:42, where T is in Kelvin).

Carbonate Clumped Isotope Thermometry: Kinetic
Isotope Effect or Useful Proxy to Constrain Deep-Biosphere Processes?Carbonate clumped isotope thermometry only depends on the fluid's temperature, not its isotopic composition, making this, at least in theory, a more direct proxy for precipitation temperatures.However, since the clumped isotope thermometer is also based on equilibrium thermodynamics, it potentially can reflect either equilibrium conditions or suffers from the same kinetic effects as oxygen isotope thermometry [39].Furthermore, the carbonate clumped isotopes in microbially related carbonates were previously suggested to show kinetic effects significantly impacting temperature estimates [14][15][16].Since carbonate mineral precipitation in our study area is driven by microbially mediated reactions [10], it is thus essential to verify whether the processes that drive the kinetic effects are significant in the settings where our samples are located in which methane oxidation occurs.This study compares two independent paleothermometers in the same samples, i.e., fluid inclusion data and carbonate clumped isotope thermometry (e.g., [40,41]).These two proxies rely on entirely different properties: fluid inclusions are based on microscopic bubbles of liquid and gas trapped within the crystal, whereas carbonate clumped isotopes rely on the degree of 'clumping' of different isotopes.We thus have an ideal dataset to check whether or not the T − Δ 47 is at equilibrium with the precipitation temperature of the mineral.We note that for most of the samples, the T − Δ 47 is in agreement with the reported lower estimates of fluid inclusion homogenization temperatures, except for one sample: KLX04 513 (Table 1 and Figure 4).The fluid inclusions data show homogenization temperatures of 68-96 °C and high salinity (up to 22 wt.%CaCl 2 eq.) in the older (inner) calcite phase, but a lower temperature single phase (<50 °C) with lower salinity (2-4 wt.% CaCl 2 eq.) in the younger (outer) phases of the calcite coatings [22].We have omitted sample KLX04 669 with no salinity data existing.Carbonate clumped isotope analysis requires relatively large sample volumes, and mixing between the different generations of calcites is, therefore, possible: here, it appears that the carbonate clumped isotope signal is dominated by the lower temperature calcite zone for most samples, which is consistent with petrographic observations (see inset image in Figure 5).
Because the temperatures resulting from both carbonate clumped isotope thermometry and the fluid inclusion are consistent, we conclude that no measurable kinetic fractionation by microbial activity can be detected, and our samples are in equilibrium condition.We suggest that this equilibrium condition is due to the slow methane reaction at a relatively low temperature and occurring over long time scales, thus allowing sufficient time for isotopic equilibration and leading to an equilibrium state [16,42].
More importantly, we note that the relationship between carbon isotope and carbonate clumped isotope values as a proxy for environmental temperatures is significant: in the LGF sites, the lower δ 13 C calcite appears in the relatively cold vein-system or lower carbonate clumped isotope derived temperatures (Figure 2).Microbes can be classified based on their optimal growth temperatures: simply thermophiles (50-64 °C), extreme thermophiles (65-79 °C), and hyperthermophiles (80 °C and beyond, [43].Our observation suggests that simply thermophilic microorganisms dominated at depth in the LGF system, though we cannot completely rule out the influence of extreme and hyperthermophile microorganisms in some samples.Moreover, earlier studies have  5 Geofluids microbial biomarkers in these mineral aggregates, in the form of preserved organic molecules of fatty acids that are specific for sulfate reducers that occur with methane oxidizers, which offer additional support for interpretations of in situ methane oxidation [9,10].
A significant advantage of the new carbonate clumped isotope results compared to the fluid inclusion data is that we have better constraints on temperatures below 80 °C.This turns out to be crucial information, as we demonstrate that the lower temperature precipitates indeed display the more negative carbon isotope values commonly associated with microbial activity.Thus, interpretation based on the carbonate clumped isotopes offer a highly valuable geothermometer for studying the deep biosphere through time.

Constraining the Environment of Palaeozoic Microbial
Calcite Precipitation within Veins.The previous fluid inclusion studies [10,22] determined that the fluid circulating within the fracture system was a sulfate-rich brine developed in the deep bedrock aquifer below thick sedimentary successions, which gave the underlying bedrock average burial temperature of ~154 °C achieved between 366 Ma and 224 Ma, and these successions were subsequently uplifted and eroded, as evidenced by thermochronological (U-Th)/ He investigations [44].Our new results help to understand further the temperature and nature of the fluids that sustained microbial organisms within the veins: the reconstructed oxygen isotope signature of the brine is between -8.6 and 0.5‰ (see Figure 3).Furthermore, two groups can be recognized in a δ 18 O fluid VSMOW versus T − Δ 47 plot (see Figure 3).The first group has higher T − Δ 47 (86 ± 11 to 98 ± 10 °C) and higher δ 18 O fluid (−1:8 ± 0:5 to 0:9 ± 1:4‰ VSMOW), resembling the δ 18 O of seawater (~0‰ VSMOW), corresponding to samples dominated by the earlier generation of calcite [22].The higher δ 18 O fluid-VSMOW combined with salinity values from fluid inclusions of up to 22 wt.%CaCl 2 eq.indicated that the fluid is a subsurface brine.Based on Rb-Sr isochrones of calcite-K-rich clay mineral pairs (e.g., KFM14A:80 m, 393+/-15 Ma), this calcite phase was dated to the Devonian and Carboniferous periods [22].The T-Δ 47 are above regional sea surface temperatures estimates from δ 18 O VPDB in conodont apatite and brachiopods from the Eifelian to early Famennian (22 to 32 °C,   6 Geofluids [45].This confirms that brine must have circulated below the thick sedimentary successions at burial depths, which is consistent with the temperature constraints on exhumation histories from the earlier study [23].The second group of data shows lower T-Δ 47 (51 ± 5 to 65 ± 2 °C) and lower δ 18 O fluid VSMOW value (−0:9 ± 0:6 to −6:7 ± 1:9‰ VSMOW), corresponding to samples dominated by the later generation (younger than ~260 Ma) of calcite [22].The lower δ 18 O fluid VSMOW signature combined with salinity up to 4 wt.%CaCl 2 eq. is surprising, as subsurface brines have more often a positive δ 18 O fluid VSMOW .This new finding leads us to propose that the brine is mixed with a significant portion of meteoric water modified chemically ( [46,47] through exchange with surrounding minerals, organic matter, and gases [48].The main argument for this interpretation is that the δ 18 O fluid VSMOW of the samples is interpreted as containing microbial fauna, which reflects typical modern meteoric water δ 18 O.The high salinity can be explained by dissolution at a depth of evaporite minerals from the overlying formations, placing an upper age limit for this calcite phase in the upper Silurian Öved-Ramsåsa Group in the south Sweden [49].The δ 13 C VPDB depletion would indicate that most of the microbial processes leading to cal-cite precipitation occurred relatively near-surface, where waters are warm but not hot, and perhaps also charged in organic carbon.In summary, the new carbonate clumped isotope data presented here suggest that the evolution of calcite veins in the LGF area started with early burial fractures acting as conduits for subsurface fluids, leading to the precipitation of the first calcite generation.These early fractures must have been reactivated during the Variscan uplift and served as conduits for meteoric fluids; this is when a microbial 'deep biosphere' community was established within the fractures and leading to the precipitation of a late calcite generation (Figure 5).

Conclusions
From the LGF dataset, we conclude that (1) the combined carbonate clumped isotope and fluid inclusion temperature dataset seems a priori to rule out any significant kinetic effects of microbial processes on carbonate clumped isotope temperatures, specifically in the relatively low-temperature setting.This is important as it indicates that carbonate clumped isotopes could be applied more broadly to the study 7 Geofluids of the unexplored ancient deep biosphere.(2) All microbial related calcite with low δ 13 C VPDB reflecting anaerobic oxidation of methane or organic matter precipitated at T-Δ 47 below 60 °C, which suggests that the microbial community was composed of 'simply thermogenic' organisms.
(3) The earlier generation of calcite with higher δ 18 O fluid- VSMOW combined with salinity values from fluid inclusions of up to 22 wt.%CaCl 2 eq.indicated that the fluid is a subsurface brine, and (4) the fluid that sustained microbial activity was probably a near-surface meteoric in origin but modified to brine type by interaction with evaporite minerals and that the timing of the deep biosphere existence was related to the Variscan uplift when meteoric fluids circulated through reactivated fractures.A broader conclusion is that carbonate clumped isotope thermometry significantly improves precision for estimating temperatures and in the LGF case study has proven to be a valuable tool for understanding environmental processes related to the deep biosphere.

Figure 1 :
Figure 1: Fennoscandian Shield tectonic domain (figure modified from Koistinen et al., 2001 in the previously mentioned reference[17]).The six tectonic domains have been separated based on differences in the timing of tectonic activity or the character and intensity of the strain.Forsmark is situated in tectonic domain 2, and Laxemar and Götemar areas are inside tectonic domain 5.The index map (bottom right) shows the location (yellow spots) of Forsmark on the northern side and Laxemar and Götemar on the southern side.

Figure 2 :
Figure 2: Plot between δ 13 C calcite VPDB and temperature derived from carbonate clumped isotopes with standard error bar.The association of lower temperatures with more depleted δ 13 C calcite VPDB is interpreted as representing the optimal growth temperatures of the organisms.

Figure 3 :
Figure 3: Plot between δ 18 O fluid VSMOW and temperature derived from carbonate clumped isotopes with standard error bar of nine samples from Forsmark, Laxemar, and Götemar.The diagonal dash line indicates the δ 18 O calcite VPDB .

Figure 4 :
Figure 4: Plot of temperature derived from clumped isotope analysis (T Δ47 ) and temperature derived from fluid inclusion analysis (T FI ).

Figure 5 :
Figure5: Simplified model of the evolution of paleofluid circulation along different fracture sets.The earlier phase of calcite generation happens in a setting similar to the top picture, whereas the later calcite phase is precipitated in a setting represented by the bottom picture.The inset image is modified after[10] and shows the coexistence of different sets of calcite veins (separated by a dashed yellow line) in the BSE-SEM images of polished calcite crystals.The later calcite phase represents the most significant volume of the mineral precipitate, which may dominate the carbonate clumped isotope signal.The numbered red dots on the inset image indicate sample points from the previous study.
L: aqueous liquid; V: vapor; prim: primary inclusion; sec: secondary inclusion.L (+V) * * * = a vapor bubble arises after cooling just before temperature of complete ice melting.