Applied Geochemistry (v.58, #C)

Southwestern Ontario has a legacy of unplugged oil and gas wells drilled in the late 1800s and early 1900s before the advent of government regulatory controls. A number of these wells exhibit artesian flow of salty and/or sulphurous water at the surface, creating a liability to landowners and a possible threat to potable groundwater aquifers and the surface environment. Cost-effective plugging of these wells is hindered by incomplete drilling records regarding well depths, well construction, and geologic information on the sources of the leaking fluids. This study seeks to fill this knowledge gap by characterizing distinct geochemical ‘fingerprints’ for each of the major bedrock water-producing zones that may be the possible sources of the leaking fluids. To this end, over 130 groundwater samples were collected and analyzed for a broad suite of isotopic parameters (δ 18O, δ 2H, δ 34SSO4, δ 18OSO4, δ 13CDIC, 87Sr/86Sr, δ 37Cl and δ 81Br), and this dataset was combined with available data from previous studies in the area. A Bayesian mixing model, SIAR, was applied to these data to develop a statistical tool for identifying the probable source(s) of leaking fluids. Several hypothetical samples and one real-world example are presented here to demonstrate the model’s performance.

Riverine transfer of anthropogenic and natural trace metals to the Gulf of Lions (NW Mediterranean Sea) by C. Dumas; W. Ludwig; D. Aubert; F. Eyrolle; P. Raimbault; A. Gueneugues; C. Sotin (14-25).
During 2005–2011, two contrasting river systems in the Gulf of Lions (southern France) have been monitored for particulate trace metal (PTM) loads (Cd, Co, Cr, Cu, Ni, Pb, Zn). These are the Rhone River, one of the largest Mediterranean rivers and a significant source of man-made contaminants to this semi-enclosed sea, and the Tet River, a small river which is representative for the numerous coastal rivers in the Mediterranean area. The objectives were to understand the transport mechanisms of PTM and to identify specific anthropogenic pressures in the corresponding drainage basin. Concentrations of all elements are highly associated with the concentrations of suspended particulate matter (SPM), but relative metal enrichment in these particulates can be variable depending on the considered elements, rivers and hydrological conditions. Concentrations of Co, Cr and Ni are generally more constant, whereas concentrations of Cd, Cu, Pb and Zn depict much greater variability, likely the result of anthropogenic contributions. Because of its large basins size and particle mixing during transport, variability in the Rhone River is reduced compared to the Tet River, which underlines the interest of small rivers for the study of transport processes. In both rivers, increasing water discharge and SPM concentrations generally reduce PTM enrichment, but flushing of enriched PTM during floods can have a strong positive impact on the average PTM budgets. Through the definition of natural background concentrations for both catchments, we further propose discrimination of the total PTM fluxes into their anthropogenic and natural counterparts. When subtracting the natural counterparts, specific anthropogenic pressures consistent with the prevailing land use practices can be identified. Greater specific anthropogenic fluxes of Cu in the Tet River (0.8 kg km−2  yr−1 compared to 0.6 kg km−2  yr−1 for the Rhone River) point to inputs from agricultural sources, whereas greater specific anthropogenic fluxes of Cd and Pb in the Rhone River (respectively 0.009 and 0.7 kg km−2  yr−1 compared to 0.005 and 0.4 kg km−2  yr−1 for the Tet River) indicate inputs from industrial sources. Also detailed analyses of the PTM fluxes during floods were helpful for source identification of PTM. In the Rhone River, comparison of two major floods showed that Pb is mobilised from the upper and more industrialised basins parts. In the Tet River, strong association of anthropogenic PTM and organic matter during floods indicates significant contributions of waste water point sources.

Reactive transport processes occurring during nuclear glass alteration in presence of magnetite by D. Rébiscoul; V. Tormos; N. Godon; J.-P. Mestre; M. Cabie; G. Amiard; E. Foy; P. Frugier; S. Gin (26-37).
Display OmittedIn this study, we have investigated and clarified the processes occurring during the alteration of SON68 glass – the reference nuclear glass for the waste arising from reprocessing of spent fuel from light water reactors – at 50 °C in Callovo-Oxfordian clay groundwater in presence of magnetite. Magnetite is known to be one of the iron corrosion products expected to be present in the vicinity of glass in geological disposal conditions. The effects of the amount of magnetite relative to the glass surface and the transport of aqueous species during glass alteration were studied. A first series of experiments was focused on the effect of various magnetite amounts by mixing and altering glass and magnetite powders. In a second series of experiments, magnetite was separated from the glass by a diffusive barrier in order to slow down the transport of aqueous species. Glass alteration kinetics were analyzed and solids were characterized by a multiscale approach using Raman Spectroscopy, Scanning and Transmission Electron Microscopy, Energy-Dispersive X-ray and Scanning Transmission X-ray Microscopy coupled with Fe L2,3-edge and C K-edge NEXAFS.It appears that glass alteration increases with the amount of magnetite and that the transport of aqueous species is a key parameter. Several processes have been identified such as (i) the silica sorption on the magnetite surface, (ii) the precipitation of Fe-silicates in the vicinity of the glass (iii) the precipitation of SiO2 on the magnetite surface, (iv) the incorporation of Fe within the alteration layer. Process (iv) was not frequently observed, suggesting local variations in geochemical conditions. Moreover, this process is strongly influenced by the transport of aqueous species as indicated by the morphology and composition of the alteration layers. Indeed, when glass and magnetite are homogeneously mixed, the glass alteration layer consists of a gel enriched in Fe having the same Fe(II)/Fe(III) ratio as in magnetite. When both materials are separated by a diffusive barrier, the glass alteration layer consists of a porous gel (not enriched in iron) in presence of a mixture of Fe-silicates with Fe having the same valence as in magnetite, rare-earth precipitates and phyllosilicates. These results suggest that Fe incorporation within the alteration layer changes depending on the distance and the time required for dissolved Fe originating from the magnetite to reach the glass.

Display OmittedMany studies have proposed that silicic acid and phosphate (PV) can displace arsenic sorbed to iron oxides leading to elevated As concentrations in aquatic systems. While surface complexation models are adept at quantifying sorption to synthetic oxides in laboratory systems their application to complex natural systems remains challenging. In this study we provide a systematic approach to developing a robust use of models for understanding AsV distribution in natural systems in which hydrated iron oxides are the main adsorptive phase. The Waikato River provides a useful laboratory for this work because it contains high H4SiO4, AsV and PV loadings due to geothermal and agricultural inputs. A 15 min oxalate extraction and a 48 h ethylenediaminetetraacetic acid (EDTA) extraction of river sediment contained the same ratios of As:Fe, P:Fe and Si:Fe. Both of these extracts target the poorly ordered iron oxide phases (typically ferrihydrite) and by following the release of elements over time in the EDTA extraction it was possible to demonstrate that the extracted As, P, and Si were associated with the ferrihydrite. This demonstrates for the first time that a single oxalate extraction can quantify ferrihydrite sorbed H4SiO4, As and PV and provides a basis to quantify the role of these ligands in inhibiting AsV sorption to sediments. The measured concentrations of ferrihydrite sorbed AsV, PV and H4SiO4 for the Waikato River suspended sediment allow for the informed selection of appropriate model parameters for applying the Diffuse Layer Model to the system. In this way it was possible to quantify the effect of the individual components in the river water on AsV sorption. This study provides an explanation for the observation that the proportion of sorbed As in the Waikato River is generally significantly lower than that observed in rivers closer to the world average concentrations. More generally the study provides a method to quantify the role of individual water chemistry components on AsV distribution in natural systems.

Reconstructing the evolution of Lake Bonney, Antarctica using dissolved noble gases by Rohit B. Warrier; M. Clara Castro; Chris M. Hall; Fabien Kenig; Peter T. Doran (46-61).
Lake Bonney (LB), located in Taylor valley, Antarctica, is a perennially ice-covered lake with two lobes, West Lake Bonney (WLB) and East Lake Bonney (ELB), which are separated by a narrow ridge. Numerous studies have attempted to reconstruct the evolution of LB because of its sensitivity to climatic variations and the lack of reliable millennial-scale continental records of climate in this region of Antarctica. However, these studies are limited by the availability of accurate lacustrine chronologies. Here, we attempt to better constrain the chronology of LB and thus, the evolution of past regional climate by estimating water residence times based on He, Ne and Ar concentrations and isotopic ratios in both WLB and ELB. 3He and 4He excesses up to two and three orders of magnitude and 35–150 times the atmospheric values are observed for WLB and ELB samples, respectively. In comparison, while measured 40Ar/36Ar ratios are atmospheric (∼295.5) in ELB, WLB samples display 40Ar/36Ar ratios of up to ∼315 reflecting addition of radiogenic 40Ar. Both 4He and 40Ar excesses clearly identify the addition of subglacial discharge (SGD) from underneath Taylor Glacier into WLB at depths of 25 m and 35 m. He isotopic ratios suggest that He excesses are predominantly crustal (>93%) in origin with small mantle contributions (<7%). These crustal 4He and 40Ar excesses are used together with basement rock production rates of these isotopes to derive first-order approximations of water residence times for both lobes. Numerous factors capable of affecting water residence times are evaluated and corrected 4He and 40Ar water ages are used to place further constrains into the reconstruction of both WLB and ELB history. Combined 4He and 40Ar ages in WLB suggest maximum water residence times of ∼250 kyrs BP. These results support the presence of remnant water from proglacial lakes that existed during Marine Isotope Stage 7 (160–240 kyrs) in WLB, in agreement with previous studies. In comparison, 4He ages in ELB are much younger (<27 kyrs BP) and display a complex evolutionary history that is very different from WLB. 4He ages also suggest that the ELB ice cover formed significantly earlier (∼1.5 kyrs BP) than previously reported. The timing of these hydrologic changes in ELB appears to correspond to regional and global climatic events that are recorded in both the Taylor Dome ice-core record as well as in other Dry Valley Lakes.

A conceptual model was developed for a hydrogeological flow system in the southern Outaouais Region, Quebec, Canada, where the local population relies heavily on groundwater pumped from shallow overburden aquifers and from deeper fractured crystalline bedrock. The model is based on the interpretation of aqueous inorganic geochemical data from 14 wells along a cross-section following the general flow direction, of which 9 were also analysed for isotopes (δ18O, δ2H, 3H, δ13C, 14C) and 4 for noble gases (He, Ne, Ar, Xe, Kr). Three major water types were identified: (1) Ca–HCO3 in the unconfined aquifer as a result of silicate (Ca-feldspar) weathering, (2) Na–Cl as a remnant of the post-glacial Champlain Sea in stagnant confined zones of the aquifer, and (3) Na–HCO3, resulting from freshening of the confined aquifer due to Ca–Na cation exchange. Chemical data also allowed the identification of significant mixing zones. Isotope and noble gas data confirm the hypothesis of remnant water from the Champlain Sea and also support the hypothesis of mixing processes between a young tritium-rich component with an older component containing high 4He concentrations. It is still unclear if the mixing occurs under natural flow conditions or if it is induced by pumping during the sampling, most wells being open boreholes in the bedrock. It is clear, however, that the hydrogeochemical system is dynamic and still evolving from induced changes since the last glaciation. As a next step, the conceptual model will serve as a basis for groundwater flow, mass transport and geochemical modelling to validate the hypotheses developed in this paper.

Dissolution of altered tuffaceous rocks under conditions relevant for CO2 storage by Yutaro Takaya; Kentaro Nakamura; Yasuhiro Kato (78-87).
We conducted CO2–water–rock interaction experiments to elucidate the dissolution characteristics and geochemical trapping potential of three different altered andesitic to rhyolitic tuffaceous rocks (Tsugawa, Ushikiri and Daijima tuffaceous rock) relative to fresh mid-ocean ridge basalt. The experiments were performed under 1 MPa CO2 pressure to reproduce the water–rock–CO2 interactions in CO2 storage situations. Basalt showed high acid neutralization potential and rapid dissolution of silicate minerals. Two of the tuffaceous rocks (Ushikiri and Daijima) showed relatively high solubility trapping potential, mainly due to the dissolution of carbonate minerals in the andesitic Ushikiri tuffaceous rock and the ion-exchange reaction with zeolite minerals in the rhyolitic Daijima tuffaceous rock. The mineral trapping potential of the Ushikiri tuffaceous rock was found to be relatively high, due to the rapid dissolution of Mg- and Ca-bearing silicate minerals. Our experimental results suggest that regions of porous and andesitic tuffaceous rock hold global promise as CO2 storage sites.

Key factors controlling the gas adsorption capacity of shale: A study based on parallel experiments by Jijun Li; Xintong Yan; Weiming Wang; Yanian Zhang; Jianxin Yin; Shuangfang Lu; Fangwen Chen; Yuanlin Meng; Xinwen Zhang; Xiang Chen; Yongxin Yan; Jingxiu Zhu (88-96).
This article performed a series of parallel experiments with numerical modeling to reveal key factors affecting the gas adsorption capacity of shale, including shale quality, gas composition and geological conditions. Adsorption experiments for shales with similar OM types and maturities indicate that the OM is the core carrier for natural gas in shale, while the clay mineral has limited effect. The N2 and CO2 adsorption results indicate pores less than 3 nm in diameter are the major contributors to the specific surface area for shale, accounting for 80% of the total. In addition, micropores less than 2 nm in diameter are generated in large numbers during the thermal evolution of organic matter, which substantially increases the specific surface area and adsorption capacity. Competitive adsorption experiments prove that shale absorbs more CO2 than CH4, which implies that injection CO2 could enhance the CH4 recovery, and further research into N2 adsorption competitiveness is needed. The Langmuir model simulations indicate the shale gas adsorption occurs via monolayers. Geologically applying the adsorption potential model indicates that the adsorption capacity of shale initially increases before decreasing with increasing depth due to the combined temperature and pressure, which differs from the changing storage capacity pattern for free gases that gradually increase with increasing depth at a constant porosity. These two tendencies cause a mutual conversion between absorbed and free gas that favors shale gas preservation. During the thermal evolution of organic matter, hydrophilic NSO functional groups gradually degrade, reduce the shale humidity and increase the gas adsorption capacity. The shale quality, gas composition and geological conditions all affect the adsorption capacity. Of these factors, the clay minerals and humidity are less important and easily overshadowed by the other factors, such as organic matter abundance.

This study presents geochemical results of springs draining a small humid tropical watershed composed of geologically dolomitic marbles in Central Sri Lanka. Water samples were investigated for their major ion chemistry, water stable isotope composition (δ2HH2O and δ18OH2O) and isotope composition of dissolved inorganic carbon (δ13CDIC). From major ion chemistry Ca2+ and Mg2+ were the dominant cations and were balanced mostly by HCO3 . All collected spring water samples scattered around the local meteoric water line with values from −31.9‰ to −46.8‰ for δ2HH2O and from −5.5‰ to −7.4‰ for δ18OH2O against VSMOW. This indicates local groundwater recharge pathways by regional precipitation rather than water from deeper aquifer systems. Concentrations of dissolved inorganic carbon (DIC) ranged from 0.91 to 9.38 mM L−1 and δ13CDIC ranged from −22‰ to −14‰ against VPDB with an average of −16‰. Samples of spring water from carbonate rocks had increased DIC and δ13CDIC together with elevated pH values. Combined δ13CDIC and Ca2+ and Mg2+ contents suggest that groundwater evolution was dominated by dissolution of dolomitic marble. This unexpected weathering process was favored by intense rain and high ambient temperatures and excessive CO2 production in tropical soils. The intense weathering resulted in karst structures with high hydraulic conductivities that rendered the terrain tectonically less stable.

Ra-226 and Rn-222 in saline water compartments of the Aral Sea region by Georg Schettler; Hedi Oberhänsli; Knut Hahne (106-122).
The Aral Sea has been shrinking since 1963 due to extensive irrigation and the corresponding decline in the river water inflow. Understanding of the current hydrological situation demands an improved understanding of the surface water/groundwater dynamics in the region. 222Rn and 226Ra measurements can be used to trace groundwater discharge into surface waters. Data of these radiometric parameters were not previously available for the study region. We determined 222Rn activities after liquid phase extraction using Liquid Scintillation Counting (LSC) with peak-length discrimination and analyzed 226Ra concentrations in different water compartments of the Amu Darya Delta (surface waters, unconfined groundwater, artesian water, and water profiles from the closed Large Aral Sea (western basin).The water samples comprise a salinity range between 1 and 263 g/l. The seasonal dynamics of solid/water interaction under an arid climate regime force the hydrochemical evolution of the unconfined groundwater in the Amu Darya Delta to high-salinity Na(Mg)Cl(SO4) water types. The dissolved radium concentrations in the waters were mostly very low due to mineral over-saturation, extensive co-precipitation of radium and adsorption of radium on coexisting solid substrates.The analysis of very low 226Ra concentrations (<10 ppq) at remote study sites is a challenge. We used the water samples to test and improve different analytical methods. In particular, we modified a procedure developed for the α-spectrometric determination of 226Ra after solid phase extraction of radium using 3M Empore™ High Performance Extraction Disks (Purkl, 2002) for the analysis of the radionuclide using an ICP sector field mass spectrometer. The 226Ra concentration of 17 unconfined groundwater samples ranged between 0.2 and 5 ppq, and that of 28 artesian waters between <0.2 and 13 ppq. The ICP-MS results conformed satisfactorily to analytical results based on γ-measurements of the 222Rn ingrowth after purging and trapping on super-cooled charcoal. The 226Ra concentrations were positively correlated with the salinity and the dissolved NaCl concentrations. The occurrence of unusually high 226Ra activities is explained by radium release from adsorption sites with increasing salinity. The inferred spatial variability of 222Rn in the Aral Sea and of 222Rn and 226Ra in the groundwater of the Amu Darya Delta is discussed in the context of our own previous hydrochemical studies in the study sites. Relatively low 222Rn activities in the unconfined GW (1–9.5 Bq/l) indicate the alluvial sediments hosting the GW to be a low-238U(226Ra) substrate. Positive correlations between U and 226Ra, and U and 222Rn are likely related to locally deposited Fe(Mn)OOH precipitates. The 222Rn activity of the GW, however, distinctly exceeds the 222Rn concentration in the Aral Sea (10 mBq/l), in principle, making 222Rn a sensitive tracer for the inflow of GW. The high water volume of the Large Aral Sea and wind induced mixing of its water body, however, hamper the detection of local groundwater inflow.

Uranium migration and retention during weathering of a granitic waste rock pile by F. Boekhout; M. Gérard; A. Kanzari; A. Michel; A. Déjeant; L. Galoisy; G. Calas; M. Descostes (123-135).
This study investigates the post-mining evolution of S-type granitic waste rocks around a former uranium mine, Vieilles Sagnes (Haute Vienne, NW Massif Central, France). This mine was operated between 1957 and 1965 in the La Crouzille former world-class uranium mining district and is representative of intra-granitic vein-type deposits. 50 years after mine closure and the construction and subsequent re-vegetation of the granitic waste rock pile, we evaluate the environmental evolution of the rock pile, including rock alteration, neo-formation of U-bearing phases during weathering, and U migration. Vertical trenches have been excavated through the rock pile down to an underlying paleo-soil, allowing the investigation of the vertical differentiation of the rock pile and its influence on water pathways, weathering processes and U migration and retention. Arenization dominantly drives liberation of U, by dissolution of uraninite inclusions in the most alterable granitic minerals (i.e. K-feldspar and biotite). Retention of U in the matrix at the base of the waste rock pile, and in the underlying paleo-soil most likely occurs by precipitation of (nano-) uranyl phosphates or a combination of co-precipitation and adsorption reactions of U onto Fe (oxy)hydroxides and/or clay minerals. Even though U-migration was observed, U is retained in stable secondary mineral phases, provided the current conditions will not be modified.

Flow-through experiments on water–rock interactions in a sandstone caused by CO2 injection at pressures and temperatures mimicking reservoir conditions by Farhana Huq; Stefan B. Haderlein; Olaf A. Cirpka; Marcus Nowak; Philipp Blum; Peter Grathwohl (136-146).
Flow-through experiments were performed in a newly designed experimental setup to study the water–rock interactions caused by CO2 injection in sandstones obtained from the Altmark natural gas reservoir under the simulated reservoir conditions of 125 °C and 50 bar CO2 partial pressure. Two different sets of experiments were conducted using CO2-saturated millipore water and CO2-saturated brine (41.62 g L−1 NaCl and 31.98 g L−1 CaCl2·2H2O), mimicking the chemical composition of the reservoir formation water. The major components in the sandstone were quartz (clasts + cement), feldspars, clay minerals (illite and chlorite), and cements of carbonates and anhydrite. Fluid analysis suggested the predominant dissolution of anhydrite causing increased concentrations of calcium and sulfate at early time periods at non-equilibrium geochemical conditions. The Ca/SO4 molar ratio (>1) indicated the concurrent dissolution of both calcite and anhydrite. Dissolution of feldspar and minor amounts of clay (chlorite) was also evident during the flow-through experiments. The permeability of the sample increased by a factor of two mostly due to the dissolution of rock cements during brine injection. Geochemical modeling suggests calcite dissolution as the major buffering process in the system. The results may in future studies be used for numerical simulations predicting CO2 storage during injection in sandstone reservoirs.

Dynamic interplay between uranyl phosphate precipitation, sorption, and phase evolution by P. Sumudu Munasinghe; Megan E. Elwood Madden; Scott C. Brooks; Andrew S. Elwood Madden (147-160).
Natural examples demonstrate uranyl-phosphate minerals can maintain extremely low levels of aqueous uranium in groundwaters due to their low solubility. Therefore, greater understanding of the geochemical factors leading to uranyl phosphate precipitation may lead to successful application of phosphate-based remediation methods. However, the solubility of uranyl phosphate phases varies over >3 orders of magnitude, with the most soluble phases typically observed in lab experiments. To understand the role of common soil/sediment mineral surfaces in the nucleation and transformation of uranyl phosphate minerals under environmentally relevant conditions, batch experiments were carried out with goethite and mica at pH 6 in mixed electrolyte solutions ranging from 1–800 μM U and 1–800 μM P. All experiments ended with uranium concentrations below the USEPA MCL for U, but with 2–3 orders of magnitude difference in uranium concentrations. Despite the presence of many cations that are well known to incorporate into less soluble autunite-group minerals, chernikovite rapidly precipitated in all experiments containing U and P, except for solutions with 1 μM U and 1 μM P that were calculated to be undersaturated. Textures of uranyl phosphates observed by AFM and TEM indicate that nucleation was homogenous and independent of the initial mineral content. Comparison of time-course U and P concentrations from the experiments with thermodynamic modeling of solution equilibria demonstrated that aqueous uranium concentrations in the experimental systems evolved as increasingly sparingly soluble uranyl phosphate phases nucleated over time, with sorption accelerating the transition between phases by influencing solution chemistry. Aqueous uranium concentrations consistent with partially dehydrated (meta-) autunite were achieved only in experiments containing goethite and/or mica. These dynamic nucleation-growth-sorption-nucleation-growth-sorption cycles occur over the time scales of weeks, not hours or days at room temperature. Lab experiments and field-based investigations of uranium phosphate should consider these or longer time scales for the greatest long-term relevance.