Search

The production rate and isotopic composition of H2 derived from radiolytic reactions in H2O were measured to assess the importance of radiolytic H2 in subsurface environments and to determine whether its isotopic signature can be used as a diagnostic tool. Saline and pure, aerobic and anaerobic water samples with pH values of 4, 7, and 10 were irradiated in sealed vials at room temperature with an artificial γ source, and the H2 abundance in the headspace and its isotopic composition were measured. The H2 concentrations were observed to increase linearly with dosage at a rate of 0.40 ± 0.04 molecules (100 eV)−1 within the dosage range of 900 to 3500 Gray (Gy; Gy = 1 J Kg−1) with no indication of a maximum limit on H2 concentration. At ∼2000 Gy, the H2 concentration varied only by 16% across the experimental range of pH, salinity, and O2. Based upon this measured yield and H2 yields for α and β particles, a radiolytic H2 production rate of 10−9 to 10−4 nM s−1 was estimated for the range of radioactive element concentrations and porosities typical of crustal rocks. The δD of H2 ( δ D = ( ( D / H ) sample / ( D / H ) standard − 1 ) × 1000 ) was independent of the dosage, pH (except for pH 4), salinity, and O2 and yielded an αDH2O-H2 of 2.05 ± 0.07 (αDH2O-H2 = (D/H)H2O to (D/H)H2), slightly less than predicted radiolytic models. Although this radiolytic fractionation value is significantly heavier than that of equilibrium isotopic exchange between H2 and H2O, the isotopic exchange rate between H2 and H2O will erase the heavy δD of radiolytic H2 if the age of the groundwater is greater than ∼103 to 104 yr. The millimolar concentrations of H2 observed in the groundwater of several Precambrian Shields are consistent with radiolysis of water that has resided in the subsurface for a few million years. These concentrations are well above those required to support H2-utilizing microorganisms and to inhibit H2-producing, fermentative microorganisms.

In this study, compositions and δ13C and δ2H isotopic values of hydrocarbon gases from 5 mines in the Witwatersrand basin, South Africa, support the widespread occurrence of microbially produced methane in millions of years-old fissure waters. The presence of microbial methane is, to a large extent, controlled by the geologic formations in which the gases are found. Samples from the Witwatersand Supergroup have the largest microbial component based on δ13C and δ2H signatures and CH4/C2+ values. Based on mixing between a microbial CH4 component and a more 13C-enriched and 2H-depleted C2+-rich end member, conservative estimates of the % contribution of microbial CH4 to the gas samples range from >90% microbial CH4 at Beatrix, Masimong, and Merriespruit, to between 5 and 80% microbial CH4 at Evander, and

Concentrations and isotopic ratios of dissolved noble gases, 36Cl, δD and δ18O in water samples from the ultra-deep gold mines (0.718 to 3.3 km below the surface) in the Witwatersrand Basin, South Africa, were investigated to quantify the dynamics of these ultra deep crustal fluids. The mining activity has a significant impact on the concentrations of dissolved gases, as the associated pressure release causes the degassing of the fissure water. The observed under saturation of the atmospheric noble gases in the fissure water samples (70–98%, normalized to ASW at 20°C and 1013 mbar) is reproduced by a model that considers diffusive degassing and solubility equilibration with a gas phase at sampling temperature. Corrections for degassing result in 4He concentrations as high as 1.55 · 10−1cm3STP4He g−1, 40Ar/36Ar ranging between 806 and 10331, and 134Xe/132Xe and 136Xe/132Xe ratios above 0.46 and 0.44, respectively. Corrected 134(136)Xe/132Xe and 134(136)Xe/4He-ratios are consistent with their production ratios, whereas the nucleogenic 4He/40Ar, and 134(136)Xe/40Ar ratios generally indicate that these gases are produced in an environment with an average [U + Th]/K-content 2–3 times above that of crustal average. In two scenarios, one considering only accumulation of in situ produced noble gases, the other additionally crustal flux components, the model ages for 14 individual water samples range from 13 to 168 Ma and from 1 to 23 Ma, respectively. The low 36Cl-ratios of (4–37) · 10−15 and comparatively high 36Cl-concentrations of (8–350) · 10−15 atoms 36Cl l−1 reflect subsurface production in secular equilibrium indicating an age in excess of 1.5 Ma or 5 times the half-life of 36Cl. In combination, the results suggest residence times of the fluids in fissures in this region (up to 3.3 km depth) are of the order of 1–100 Ma. We cannot exclude the possibility of mixing and that small quantities of younger water have been mixed with the very old bulk.

The interpretation of 39Ar recoil loss from clay minerals during application of the 40Ar/39Ar analytical technique is controversial. To further explore the mechanism for 39Ar recoil loss, we quantified the effects of backscattering where the recoiled 39Ar atom fails to reenter the coherent clay particle upon crossing an incoherent boundary. A numerical simulation with the computer program, SRIM, was conducted. The results indicate that the recoil range distribution is significantly influenced by the incident angle between the attacking direction and c-axis (perpendicular to the plane of clay surface). 39Ar concentration profiles were generated for two spatial analogues to natural illite/smectite (1) a one-component system composed of the K-rich particles with thicknesses ranging from 5 to 100 nm, and (2) a two-component system composed of alternation of the K-rich and K-poor particles having thicknesses ranging from 5 to 100 nm. The models yield 14–15%, 9–10%, 2–3%, and 1–2% backscattered 39Ar for particle thicknesses of 5 nm, 10 nm, 50 nm, and 100 nm, respectively. The initial 39Ar concentration exerts very little influence on the fraction of backscattered 39Ar. If the particles are separated by incoherent boundaries along which 39Ar atoms freely migrate, then the backscattered 39Ar must contribute to the 39Ar recoil loss. The estimates of 39Ar backscatter loss are generally consistent with published experimental data. Good agreement is obtained for samples having a particle thickness larger than 50 nm. The model underestimates the observed 39Ar recoil loss by a factor of two for samples with a particle thickness less than 50 nm. Our results indicate that 39Ar recoil loss in clays is largely controlled by backscattering at the incoherent boundary.