Radiolytically reworked Archean organic matter in a habitable deep ancient high-temperature brine

Abstract

Investigations of abiotic and biotic contributions to dissolved organic carbon (DOC) are required to constrain microbial habitability in continental subsurface fluids. Here we investigate a large (101-283 mg C/L) DOC pool in an ancient (>1Ga), high temperature (45-55 °C), low biomass (10²-10⁴ cells/mL), and deep (3.2 km) brine from an uranium-enriched South African gold mine. Excitation-emission matrices (EEMs), negative electrospray ionization (-ESI) 21 tesla Fourier-transform ion cyclotron resonance mass spectrometry (FT-ICR MS), and amino acid analyses suggest the brine DOC is primarily radiolytically oxidized kerogen-rich shales or reefs, methane and ethane, with trace amounts of C₃-C₆ hydrocarbons and organic sulfides. δ²H and δ¹³C of C₁-C₃ hydrocarbons are consistent with abiotic origins. These findings suggest water-rock processes control redox and C cycling, helping support a meagre, slow biosphere over geologic time. A radiolytic-driven, habitable brine may signal similar settings are good targets in the search for life beyond Earth.

Fig. 1. Dissolved organic and inorganic carbon (DOC/DIC) and δ¹³C, Δ¹⁴C signatures.

a δ¹³C (circles) and Δ¹⁴C measurements (triangles), relative to Vienna Pee Dee Belemnite (V-PDB), for DOC and DIC pools from (b). Reproducibility of δ¹³C is ±0.3‰ for Moab Khotsong and Kloof samples, and ±2.5‰ for Kidd Creek. δ¹³C_DOC ranges include: (i) reported values for DOC of Witwatersrand Basin fracture waters (full range -24 to -57‰^,), (ii) extractable organic carbon from Vaal Reef, (iii) extractable organic carbon from the Kimberley Shales. δ¹³C_DIC ranges include: (iv) acetogenesis or methanogenesis (full range of −7 to 12‰^,), (v) reported values for DIC of Witwatersrand Basin fracture waters (full range of -12 to -43‰^,) (vi) calcites from Moab Khotsong. δ¹³C_DIC of 0.6‰ for Transvaal Dolomites is out of range. b Quantity of DOC and DIC in mg C/L of the Moab Khotsong fracture fluids compared to Kloof Mine and Kidd Creek brine (Kidd Creek DIC (0.7 mg C/L) is smaller than other plotted samples). Error bars represented as relative standard deviation (% RSD) and are ±1% for Moab Khotsong and Kloof Mine samples and ±20% Kidd Creek, and in some cases, are smaller than the symbols. Percentages in dashed boxes indicate the percentage of low molecular weight organic acid (acetate + formate + propionate + oxalate) comprising the DOC pool. Apparent ages based on Δ¹⁴C from (a) are shown in years above each carbon pool ± relative error. Blue ages for DIC of the 1200-level include correction for carbonate dissolution upon aquifer recharge.

Fig. 2. Excitation emission matrices (EEMs).

EEMs shown are for the 95-level brine, 101-level brine, 1200-level dolomite fluid, and Vaal Reef extract. Scales display fluorescence intensity values with contours at 0, 0.02, and 5. Diagonal features represent first and second order Raman scatter. Peak regions are identified by corresponding lettered circles^,: A,N – humic-like; B – tyrosine protein-like; T – tryptophan protein-like; M – autochthonous microbial DOC. The large white area from ~480 to 600 nm emission in the 95-level spectrum is due to the presence of fluorescein dye in this sample and the scale has been adjusted to exclude the high fluorescence of this region. The full spectra for the 95-level 2018 to 2020 spectra with both this excluded vs. included fluorescein signal can be seen in Supplementary Information Fig. S2.

Fig. 3. Van Krevelen diagrams from negative electrospray ionization 21 tesla Fourier-transform ion cyclotron resonance mass spectrometry (FT-ICR MS).

Samples shown include the 95 and 101-level brines compared to the 1200-level dolomite fluid, and an extract from the Vaal Reef. Dashed red lines and arrows indicate regions of increasing (1) aliphatic, (2) aromatic, and (3) condensed aromaticity of molecules. Circles included on the 95-level diagram represent typical H/C vs. O/C regions for broad organics classes.

Fig. 4. Percent relative abundance of organic species.

Major classes O_x, N₁O_x, and S₁O_x obtained through negative ion ESI 21 tesla FT-ICR MS for 95 and 101-level brines compared to the 1200-level dolomite fluid and an extract of the kerogen-rich Vaal Reef.

Fig. 5. Chromatographs of volatile organic species from headspace gas in the 95-level brine.

The top panel displays intensity of volatile species’ peaks from 0 to 3 · 10⁷ counts while the bottom panel displays species’ peaks of lower intensity (1.6 · 10⁴ to 6.0 · 10⁴ counts) over a 19.75-minute retention range.

Fig. 6. Isotopic values (‰) for hydrogen and carbon of C_1-4 hydrocarbon gases.

Fluid systems including Moab Khotsong brines (95 (red) and 101 (green) levels), other fracture fluids previously sampled from the Witwatersrand Basin (Dreifontein—DR548, Mponeng—MP104, Kloof—KL739), and fluids from the Canadian Shield (Copper Cliff—CCS4546) including fluids from Kidd Creek Mine (KC7792) where residence times of over 1 Ga were identified (Sherwood Lollar et al.). Error bars represent ±0.5‰ and ±5‰ relative to Vienna Pee Dee Belemnite (V-PDB) and Vienna Standard Mean Ocean Water (V-SMOW) for δ¹³C and δ²H, respectively. Specific isotopic values and gas compositional data for C_1-4 hydrocarbons are included in Supplementary Material Table S1.

Fig. 7. Geographic and stratigraphic location of Moab Khotsong mine sampling sites.

a Location of the Moab Khotsong mine (along with Carltonville-based mines mentioned in this study) within the Witwatersrand Basin of South Africa (b) Stratigraphic location of facture fluid sampling sites in the Moab Khotsong mine showing the relationship of the boreholes in the Transvaal dolomite aquifer [1200-level; 1.2 km], and West Rand quartzites [95-level; 2.9 km and 101-level; 3.1 km], relative to major reef [VCR – Ventersdorp Contact Reef] and shale zones.