Search

In this study we present the first Mg isotope data that record the fate of Mg during mineralisation of atmospheric CO2 in ultramafic mine tailings. At the Woodsreef Asbestos Mine, New South Wales, Australia, weathering of ultramafic mine waste sequesters significant amounts of CO2 in hydromagnesite [Mg5(CO3)4(OH)2·4H2O]. Mineralisation of CO2 in above-ground, sub-aerially stored tailings is driven by the infiltration of rainwater dissolving Mg from bedrock minerals present in the tailings. Hydromagnesite, forming on the surface of the tailings, has lower δ26Mg (δ26MgHmgs = −1.48 ± 0.02‰) than the serpentinised harzburgite bedrock (δ26MgSerpentinite = −0.10 ± 0.06‰), the bulk tailings (δ26MgBulk tailings = −0.29 ± 0.03‰) and weathered tailings containing authigenic clay minerals (δ26MgWeathered tailings = +0.28 ± 0.06‰). Dripwater (δ26MgDripwater = −1.79 ± 0.02‰) and co-existing hydromagnesite (δ26MgHmgs = −2.01 ± 0.09‰), forming in a tunnel within the tailings, and nodular bedrock magnesite [MgCO3] (δ26MgMgs = −3.26 ± 0.10‰) have lower δ26Mg than surficial fluid (δ26Mg = −0.36‰) and hydromagnesite. Complete dissolution of source minerals, or formation of Mg-poor products during weathering, is expected to transfer Mg into solution without significant alteration of the Mg isotopic composition. Aqueous geochemical data and modelling of saturation indices, along with Rayleigh distillation and mixing calculations, indicate that the 26Mg-depletion in the drip water, relative to surficial water, is the result of brucite dissolution and/or precipitation of secondary Mg-bearing silicates and cannot be assigned to bedrock magnesite dissolution. Our results show that the main mineral sources of Mg in the tailings (silicate, oxide/hydroxide and carbonate minerals) are isotopically distinct and that the Mg isotopic composition of fluids and thus of the precipitating hydromagnesite reflects both isotopic composition of source minerals and precipitation of Mg-rich secondary phases. The consistent enrichment and depletion of 26Mg in secondary silicates and carbonates, respectively, underpins the use of the presented hydromagnesite and fluid Mg isotopic compositions as a tracer of Mg sources and pathways during CO2 mineralisation in ultramafic rocks.

Pre- to syn-eruptive fluids released by the magmatic system forming the non-welded rhyolitic Kos Plateau Tuff ignimbrite, Greece, were characterized using fluid inclusions entrapped in quartz from granitic clasts entrained from the roof or walls of the feeding magma reservoir. This fluid was initially of intermediate density and relatively low salinity (3–11 wt% NaClequivalent). Following significant pressure drop(s), it separated into immiscible brine and CO2-bearing vapor occurring as texturally less mature inclusions. The transition from single- to two-phase-fluid occurred at magmatic conditions, as supported by high fluid inclusion homogenization temperatures (ca. 700 °C) and simultaneous trapping of fluid inclusions with silicate melt inclusions. The large recorded pressure drop together with texturally immature fluid inclusion shapes suggests a possible direct relation between the fluid phase change and the incipient 161 ka caldera-forming eruption. All inclusion types occur predominantly along (pseudo-) secondary trails in the granitic clasts, indicating efficient entrapment in microfractures formed at high crystallinity, as attested to by elevated concentrations of incompatible elements in the silicate melt inclusions. Fast quenching of the inclusions due to eruption prevented significant post-entrapment chemical modification, providing pristine samples of magmatic fluid compositions. Element concentrations in co-existing fluids and silicate melt were analyzed using laser ablation inductively-coupled plasma mass spectrometry and compared in view of element transporting capabilities and magmatic-hydrothermal ore formation. Economically important metals such as Cu, Zn, Mo, and W were extracted from the silicate melt by the fluid. In particular, Cu concentrations in the single-phase fluid in the range of 100–500 ppm agree well with previous experimental and modeling studies and, therefore, represent values expected for primary magmatic fluids. Most elements got further enriched in the brine upon phase separation, except Li, B, and As, which may have contributed significantly also to the vapor phase. Considering the similarity between vapor and intermediate-density fluid in terms of salinity, however, the metal-enriched brine phase only comprised a minor fraction of the bulk fluid.

The oxygen and clumped isotope signatures of Mg calcites are routinely used as environmental proxies in a broad range of surroundings, where Mg calcite forms either by classical nucleation or via an amorphous calcium magnesium carbonate (ACMC) precursor. Although the (trans)formation of ACMC to Mg calcite has been identified for an increasing number of settings, the behavior of both isotope proxies throughout this stage is still unexplored. In the present study ACMC (trans)formation experiments were carried out at constant pH (8.30 ± 0.03) and temperature (25.00 ± 0.03 °C) to yield high Mg calcite (up to 20 mol% Mg). The experimental data indicate that the oxygen isotope values of the amorphous and/or crystalline precipitate (δ$^{18}$O$_{prec}$, analyzed as Mg calcite) are affected by the (trans)formation pathway, whereas clumped isotopes (Δ$_{47prec}$ = Δ$_{47Mg-calcite}$) are not. The oxygen isotope evolution of the solid phase can be explained by the instantaneous trapping of the isotopic composition of the aqueous (bi)carbonate complexes. This entrapment results in remarkably high 10$^{3}$ln(α$_{prec-H2O}$) values of ∼33‰ at the initial ACMC formation stage. During the ACMC transformation process the oxygen isotope equilibrium is approached rapidly between Mg calcite and water (Δ$^{18}$O$_{Mg calcite-water}$ = 30.3 ± 0.4‰) and no isotopic memory of the initial to the final Mg calcite at the end of the experiment occurs. The implications for oxygen and clumped isotope signatures of Mg calcite formed via ACMC are discussed in the aspects of various scenarios of (trans)formation conditions and their use as environmental proxies.

Barium isotope fractionation during calcite and aragonite inorganic precipitation was studied in mixed flow reactors as a function of precipitation rate at 25 °C and pH = 6.3 ± 0.1. The measured Ba isotope fractionation that occurs between calcite and the forming fluid in the investigated range of calcite growth rates (10-8.0 ≤ Rp(calcite) ≤ 10−7.3 mol/m2/s) is insignificant. Barium isotope fractionation between aragonite and the fluid decreases with increasing precipitation rate from Δ137/134Baaragonite-fluid = +0.25 ± 0.06‰ for Rp(aragonite) ≤ 10−8.7 mol/m2/s to −0.10 ± 0.08‰ for Rp(aragonite) = 10−7.6 mol/m2/s, thus reflecting preferential incorporation of either heavy or light Ba isotopes in aragonite at slow and fast growth rates, respectively. The dependence of Ba isotope fractionation on aragonite growth rate is well described by the surface reaction kinetic model developed by DePaolo (2011) when the values +0.27‰ and −2.0 ± 0.2‰ are used for the equilibrium and kinetic Ba isotope fractionation factor, respectively. The enrichment of aragonite in the heavier Ba isotopes is consistent with the equilibrium fractionation factor of +0.34‰, calculated here between Ba-substituted aragonite and Ba2+ (aq), from first-principles calculations. This positive fractionation is related to a shorter average Ba[sbnd]O bond length in the aragonite structure while the coordination number does not change much (i.e. 9). The lack of isotope fractionation between the Ba aquo ions and the 6-coordinated Ba in calcite likely suggests that the coordination reduction required for the incorporation in the lattice of Ba adsorbed at calcite growing sites proceeds without isotope fractionation with the fluid. Otherwise the precipitated calcite should have been enriched in heavy isotopes by ∼0.17‰, as predicted by first-principles calculations. These results are the first experimental measurements of Ba isotope fractionation during inorganic calcite and aragonite mineral formation and set the basis for understanding the mechanisms controlling Ba isotope composition in CaCO3 minerals that is an essential perquisite for application of this isotopic system in natural samples.

In contrast to fairly good knowledge of dissolved ( 0.45 µm) fluxes of carbon, nutrients and metals from the land to the ocean, colloidal (1 kDa−0.45 µm) forms of solutes are rarely quantified. This is especially true for Siberian rivers draining into the Arctic Ocean: because of organic-rich soils, colloidal fractions of elements are high and may sizably impact coastal biological processes. However, the main environmental parameters such as seasons, river size, climate, permafrost distribution and landscape parameters of the watershed controlling colloidal distribution of organic carbon (OC) and metals remain totally unknown. Here we used on-site centrifugation combined with ultrafiltration via 3, 30 and 100 kDa pore size membranes and 1 kDa dialysis to characterize colloidal size fractionation of OC and metals in 32 western Siberian rivers, ranging in size from 10 to 150,000 km2 watershed area, across a climate and permafrost gradient (from absent to continuous permafrost). The dominant forms of OC and metals was low molecular weight LMW

International audience; Over millions of years, the interaction of marine basalt with percolating seawater in low-temperature ocean floor hydrothermal systems leads to the formation of calcite and aragonite. The presence of these minerals in marine basalts provides evidence for substantial CO2 fixation in these rocks. Here, we report on laboratory experiments to study this process under enhanced CO2 partial pressures (pCO2) at 130 °C. Mid-ocean-ridge-basalt (MORB) glass was reacted with North Atlantic Seawater charged with CO2 in batch experiments lasting up to 7 months. For experiments initiated with seawater charged with ~ 2.5 bar pCO2, calcite and aragonite are the first carbonate minerals to form, later followed by only aragonite (± siderite and ankerite). For experiments initiated with seawater charged with ~ 16 bar pCO2, magnesite was the only carbonate mineral observed to form. In total, approximately 20 % of the initial CO2 in the reactors was mineralized within five months. This carbonation rate is similar to corresponding rates observed in freshwater-basalt-CO2 interaction experiments and during field experiments of the carbonation of basalts in response to CO2-charged freshwater injections in SW-Iceland. Our experiments thus suggest that CO2-charged seawater injected into submarine basalts will lead to rapid CO2 mineralization. Notably, at pCO2 of tens of bars, magnesite will form, limiting the formation of Mg-rich clays, which might otherwise compete for the Mg cation and pore-space in the submarine basaltic crust. This suggests that the injection of CO2-charged seawater into subsurface basalts can be an efficient and effective approach to the long-term safe mineral storage of anthropogenic carbon.

International audience; Storage of Xe at depth in silicate minerals has recently been proposed to explain the low Xe abundance in the Earth’s and Mars’ atmospheres compared to other noble gases (the so-called ‘Missing Xenon’ issue). Evidences for incorporation, and thus reactivity of Xe in olivine at high pressure and high temperature are based on variations in cell parameters and the appearance of a new Raman band. To constrain the, so far only hypothetical, Xe incorporation mechanism in olivine, we theoretically investigated models of Xe-bearing olivine using density functional theory. Three types of incorporation mechanisms are tested: Xe for Si and Xe for Mg substitutions, and interstitial Xe. Xe for Si substitution, implying an oxidation of Xe, is found to be the only mechanism consistent with experimental observations, leading to an increase of cell parameter a and the appearance of a new Raman band around 720–750 cm−1 associated with XeO stretching vibrations. Raman spectroscopy makes it possible to identify Xe incorporation site, even at low Xe content, due to high Xe polarizability. An estimation of Xe content in olivine, based on present work and previous in situ experimental results, shows that up to 0.4 at.% Xe could be stored in olivine at depth.

International audience; The present study investigated the weathering and transport mechanisms of Fe in the Amazon River. A particular emphasis was placed on Fe partitioning, speciation, and isotopic fractionation in the contrasting waters of the Solimões and Negro rivers and their mixing zone at the beginning of the Amazon River. Samples collected in the end-member rivers and thirteen sites distributed throughout the mixing zone were processed through frontal vacuum filtration and tangential-flow ultrafiltration to separate the different suspended solid fractions, i.e., particulate (P\textgreater0.45μm and P\textgreater0.22μm), colloidal (0.22μm\textgreaterC\textgreater5kDa) and truly dissolved elements (TD\textless5kDa). The Fe isotopic composition and electron paramagnetic resonance (EPR) species were measured on these different pore-sized fractions. The acidic and organo-Fe-rich waters of the Negro River displayed dissolved and colloidal fractions enriched in heavy isotopes (~1.2\textperthousand, in δ57Fe values relative to IRMM-14), while the particulate fractions yielded light isotopic compositions of -0.344\textperthousand for P\textgreater0.22μm and -0.104\textperthousand for P\textgreater0.45μm fractions). The mineral particulate-rich waters of the Solimões River had dissolved and colloidal fractions with light isotopic composition (-0.532\textperthousand and -0.176\textperthousand, respectively), whereas the particulate fractions yielded δ57Fe values close to those of the continental crust (i.e., -0.029\textperthousand for P\textgreater0.22μm and 0.028\textperthousand for P\textgreater0.45μm). Ten kilometers downstream from the Negro and Solimões junction, the concentrations of colloidal and dissolved Fe species deviate markedly from conservative mixing. A maximum Fe loss of 43μg/L (i.e., 50% of the dissolved and colloidal Fe) is observed 110km downstream from the rivers junction. The contrasting Negro and Solimões Rivers isotopic compositions along the pore-sized water fractions is attributable to the biogeochemical processes involving different types of upland soils and parental materials. For instance, the isotopic composition of colloidal and dissolved Fe from the Negro River are consistent with Fe oxidation and complexation mechanisms at the interface between waterlogged podzols and river networks, as supported by strong organo-Fe complexes signals observed by EPR. Conversely, the particulate and colloidal fractions from the Solimões River have δ57Fe consistent with strong mechanical erosion in the Andean Cordillera and upland soils, as evidenced by high concentrations of Fe3+-oxides sensu lato measured by EPR. The massive dissolved and colloidal Fe removal is associated with the evolution of the physical and chemical composition of the waters (i.e., ionic strength) during mixing, which influences organo-Fe3+ and Fe3+-oxyhydroxides stability. Several models are discussed to explain Fe non-conservative behavior, including dissociation of organo-Fe complexes and the subsequent formation of solid Fe3+-oxyhydroxides and semiquinone free radicals, as evidenced by EPR spectra demonstrating that organo-Fe signals decrease as Fe3+-oxyhydroxides and free radicals signals increase. As in estuarine regions, the mechanisms involving Fe transfer and loss in the mixing zone has a negligible effect on the bulk water Fe isotopic composition. This result suggests that a tropical basin similar to the Amazon River Basin delivers to the ocean waters with an Fe isotopic composition similar to that of the Earth\textquoterights continental crust.

The formation constants of magnesium bicarbonate and carbonate ion pairs have been experimentally determined in dilute hydrothermal solutions to 150 °C. Two experimental approaches were applied, potentiometric acid–base titrations at 10–60 °C and spectrophotometric pH measurements using two pH indicators, 2-naphthol and 4-nitrophenol, at 25 and 80–150 °C. At a given temperature, the first and second ionization constants of carbonic acid (K1, K2) and the ion pair formation constants for MgHCO 3 + ( aq ) ( K MgHCO 3 + ) and MgCO3(aq) ( K MgCO 3 ) were simultaneously fitted to the data. Results of this study compare well with previously determined values of K1 and K2. The formation constants of MgHCO 3 + ( aq ) and MgCO3(aq) ion pairs increased significantly with increasing temperature, with values of log K MgHCO 3 + = 1.14 and 1.75 and of log K MgCO 3 = 2.86 and 3.48 at 10 °C and 100 °C, respectively. These ion pairs are important aqueous species under neutral to alkaline conditions in moderately dilute to concentrated Mg-containing solutions, with MgCO3(aq) predominating over CO 3 2 - ( aq ) in solutions at pH >8. The predominance of magnesium carbonate over carbonate is dependent on the concentration of dissolved magnesium and the ratio of magnesium over carbonate. With increasing temperature and at alkaline pH, brucite solubility further reduced the magnesium concentration to levels below 1 mmol kg−1, thus limiting availability of Mg2+(aq) for magnesite precipitation.

Multiple temperature-age constraints for eight H-chondrite samples have been used to provide insight into the thermal history of their parent-body through combination with numerical models of thermal evolution assuming internal heating by 26Al and conductive cooling. The effect of spreading accretion out over time is the principal focus of this work. A wide range of body size and date of accretion is systematically tested for different values of accretion rate in order to quantify and illustrate the parameter space that is consistent with the available thermo-chronological data. We conclude that the H-chondrite samples considered have a thermal history consistent with a parent body that at some stage had a concentric ‘onion-shell’ internal structure. That body had a radius no larger than 130 km, and accretion most probably took place over a time interval on the order of 0.0–0.2 Myr, approximately 2 Myr after CAI condensation. In any case, the time interval of accretion is unlikely to have been more than 0.5 Myr supporting evidence in favour of rapid accretion, possibly through reassembly of the fragments of an earlier generation of bodies. Furthermore, the H-chondrites studied here are inferred to have come from a wide depth range within the body where they experienced metamorphism, indicating that preservation of the onion-shell structure is unlikely. The presence of an insulating regolith does not modify this conclusion, as appropriate thermal histories for the three H6 samples considered cannot be reproduced at depths near the surface. Asteroid 6-Hebe may be the parent body of the H-chondrites, but the high bulk density of the latter is difficult to reconcile with a ‘rubble-pile’ structure of pure H-chondrite material. Finally, optimized thermal histories are used to constrain the temperatures characterizing boundaries between petrological types (800, 1000, and 1140 K for the H3/4, H4/5, and H5/6 boundaries respectively). In detail, the type 6 samples studied here are inferred to have experienced a similar peak temperature (on the order of 1180 K), but spent different times at that temperature, while samples of lower metamorphic grade experienced different peak temperature, but spent more or less the same time within 10 K of peak temperature (on the order of 1 Ma). These results constitute a quantitative framework within which variations in petrographic and textural properties may be interpreted.

The solubility of bismuth oxide (α-Bi 2 O 3(s) ; bismite) in near-neutral sodium perchlorate solutions at 65 and 80 °C, and pure water from 150 to 600 °C, P Sat to 800 bar was studied using various batch-reactor techniques and in situ XAS spectroscopy. The solubility of Bi 2 O 3(s) follows a similar trend to Sb 2 O 3(s) (senarmontite), which has been interpreted in terms of a neutral Sb(OH) 3(aq) complex. Thus a similar neutral complex, Bi(OH) 3(aq) , is inferred for Bi. XANES spectroscopy confirms that the Bi(OH) 3(aq) complex carries a stereochemically active lone electron pair, and EXAFS data suggest that the geometry of the complex changes little over the temperature range 380–610 °C at 800 bar, with three oxygen neighbors at ∼2.08 A. The solubility data obtained in this study are used in conjunction with thermodynamic properties for α-Bi 2 O 3(s) to obtain thermodynamic parameters for Bi(OH) 3(aq) within the framework of the revised Helgeson–Kirkham–Flowers (HKF) equation of state. Speciation calculations using these new properties indicate that, similarly to arsenic and antimony, bismuth is transported predominantly as a neutral hydroxide complex in a wide range of temperature, pressure, and fluid compositions. In contrast to arsenic and antimony, bismuth is much less soluble in typical hydrothermal fluids in the form of hydroxide complexes, and high temperatures (⩾400 °C) are required for significant Bi transport by aqueous fluids. These results are consistent with the common association between Bi mineralization and magmatism.

This work is aimed at quantification of forsteritic olivine (Fo 92 ) dissolution kinetics in batch and mixed-flow reactors in the presence of aerobic gram-negative bacteria ( Pseudomonas reactans HK 31.3) isolated from an instrumented well located within a basaltic aquifer in Iceland. The release rate of mineral constituents was measured as a function of time in the presence of live and dead cells in constant-pH (4–9), bicarbonate-buffered (0.001–0.05 M), nutrient-rich and nutrient-free media in batch reactors at 0–30 atm of CO 2 partial pressure (pCO 2 ). In batch reactors at 30 atm pCO 2 , 0.1 M NaCl and 0.05 M NaHCO 3 the rates were weakly affected by the presence of bacteria. In nutrient media, the SEM observation of reacted grains revealed the presence of biofilm-like surface coverage that does not modify Mg and Si release rate at the earlier stages of reaction but significantly decreased the dissolution after prolonged exposure. Olivine dissolution rates measured in flow-through reactors are not affected by the presence of dead and live bacteria at pH ⩾9 in 0.01 M NaHCO 3 solutions. In circumneutral, CO 2 -free solutions at pH close to 6, both live and dead bacteria increase the dissolution rate, probably due to surface complexation of exudates and lysis products. In most studied conditions, the dissolution was stoichiometric with respect to Mg and Si release and no formation of secondary phases was evidenced by microscopic examination of post-reacted grains. Obtained results are consistent with known molecular mechanism of olivine dissolution and its surface chemistry. Overall, this work demonstrates negligible effect of P. reactans on olivine reactivity under conditions of CO 2 storage in the wide range of aqueous fluid composition.

The hydrous magnesium carbonates, nesquehonite (MgCO3.3H2O) and dypingite (Mg5(CO3)4(OH)2.5(H2O)), were precipitated at 25 oC in batch reactors from aqueous solutions containing 0.05 M NaHCO3 and 0.025 M MgCl2 and in the presence and absence of live photosynthesizing Gloeocapsa sp. cyanobacteria. Experiments were performed under a variety of conditions; the reactive fluid/bacteria/mineral suspensions were continuously stirred, and/or air bubbled in most experiments, and exposed to various durations of light exposure. Bulk precipitation rates are not affected by the presence of bacteria although the solution pH and the degree of fluid supersaturation with respect to magnesium carbonates increase due to photosynthesis. Lighter Mg isotopes are preferentially incorporated into the precipitated solids in all experiments. Mg isotope fractionation between the mineral and fluid in the abiotic experiments is identical, within uncertainty, to that measured in cyanobacteria bearing experiments; measured δ26 ranges from -1.54‰ to -1.16‰ in all experiments. Mg isotope fractionation is also found to be independent of reactive solution pH and Mg, CO32-, and biomass concentrations. Taken together, these observations suggest that Gloeocapsa sp. cyanobacterium does not appreciably affect magnesium isotope fractionation between aqueous fluid and hydrous magnesium carbonates.

The vertical variability in mineralogical, chemical and isotopic compositions observed in large river suspended sediments calls for a depth-integration of this variability to accurately determine riverine geochemical fluxes. In this paper, we present a method to determine depth-integrated chemical particulate fluxes of large rivers, based on river sampling along depth-profiles, and applied to the Amazon Basin lowland tributaries. The suspended particulate matter (SPM) concentration data from depth-profiles is modeled for a number of individual grain size fractions using the Rouse model, which allows to predict the grain size distribution of suspended sediment throughout the whole river cross-section. Then, using (1) the relationship between grain size distribution and the Al/Si ratio (2) relationships between the Al/Si ratio and the chemical concentrations, the chemical composition of river sediment is predicted throughout the river cross-section, and integrated to yield the depth-integrated chemical particulate flux for a number of chemical elements (e. g. Si, Al, Fe, Na, REEs, ... ). For elements such as Al, Fe, REEs, Th, the depth-integrated flux is around twice as high as the one calculated from river surface sample characteristics. For Na and Si, the depth-integrated flux is three times higher than the "surface" estimate, due to the enrichment of albite and quartz at the bottom of the river. Depth-integrated (87)Sr/(86)Sr composition of suspended sediment, also predictable using this method, differs by more than 10(-3) from the surface sample composition. Finally, potential implications of depth-integrated estimates of Amazon sediment chemistry are explored. Depth-integration of particulate (87)Sr/(86)Sr isotopic ratios is necessary for a reliable use of Sr isotopes as a provenance tracer. The concept of steady-state weathering of a large river basin is revisited using depth-integrated sediment composition. This analysis shows that, in the Amazon Basin river, the previously observed discrepancy between (1) weathering intensities of channel surface sediment and (2) silicate-derived dissolved fluxes is only slightly accounted for by the vertical variability of suspended sediment weathering intensities. This observation confirms that most large rivers basins are not eroding at steady-state. (C) 2011 Elsevier Ltd. All rights reserved.

Anthropogenic and natural climate change affect processes in the atmosphere, biosphere, hydrosphere, and pedosphere. The impact of climate on soil evolution has not been well-explored, largely due to slow rates and the complexity of coupled processes that must be observed and simulated. The rates of mineral weathering in loess deposited 23 kyr ago and experiencing soil formation for 13 kyr are explored here using the WITCH model for weathering and the GENESIS model for climate simulation. The WITCH model, which uses rigorous kinetic parameters and laws with provision for the effect on rates of deviation from equilibrium, can successfully simulate the depletion profiles in the soil for dolomite and albite if soil CO2 is assumed to rise over the last 10 kyr up to about 30–40× the present atmospheric pressure, and if the solubility product of the Ca-smectite is assumed equal to that of an Fe(III)-rich Ca-montmorillonite. Such simulations document that dissolution behavior for silicates and carbonates are strongly coupled through pH, and for Ca-smectite and feldspars through dissolved silica. Such coupling is not incorporated in simple geometric and analytical models describing mineral dissolution, and therefore probably contributes to the long-standing observation of discrepancies among laboratory and field mineral dissolution rates.

The stability of celestine (SrSO 4 ) in marine sediments has been investigated through the calculation of its saturation index at the in situ temperature and pressure, using the entire ODP/IODP porewater composition data base (14,416 samples recovered from sediments collected during 95 ODP and IODP Legs) that has been thoroughly corrected for missing data and inconsistencies. The porewater in situ pressure has been obtained from depth data, and the in situ sample temperature has been calculated from the bottom seawater temperature and from the measured thermal gradient. When the latter is unavailable, a default value of 35 °C/km or an assumed gradient similar to that of a site located close to the study area has been used. Molarities have been converted to molalities through the calculation of the porewater densities, which have represented by an empirical function of the total dissolved solids load up to 150 g/L. Ocean bottom waters are largely undersaturated with respect to celestine. Yet sediment porewater saturation is not uncommon: it is reached in 83 boreholes (i.e. about 10% of all boreholes) drilled during 22 ODP/IODP Legs. Celestine equilibrium can be reached through two different non-exclusive causes: a strontium increase in porewater linked to the dissolution of Sr-rich aragonite or biogenic calcite and precipitation of Sr-poor calcite (carbonate recrystallization), or an increase in the strontium and sulfate porewater concentrations linked to a salinity increase due to the presence of brines. Sediments at most of the sites exhibiting celestine equilibrium share common characteristics such as a carbonate-rich lithology (typically higher than 80 wt.%) and a low organic carbon content (generally below 1 wt.%). These results indicate that modification of porewaters during burial diagenesis can easily lead to celestine saturation, especially in carbonate-rich sediments. We then briefly discuss former interpretations of the presence of celestine in ancient and recent marine sediments, as well as the consequences of the incorporation of celestine formation in diagenetic models using the porewater Sr content and Sr isotopic composition.

The goal of this study was to highlight the occurrence of an additional proton-promoted weathering pathway of carbonate rocks in agricultural areas where N-fertilizers are extensively spread, and to estimate its consequences on riverine alkalinity and uptake of CO2 by weathering. We surveyed 25 small streams in the calcareous molassic Gascogne area located in the Garonne river basin (south-western France) that drain cultivated or forested catchments for their major element compositions during different hydrologic periods. Among these catchments, the Hay and the Montousse, two experimental catchments, were monitored on a weekly basis. Studies in the literature from other small carbonate catchments in Europe were dissected in the same way. In areas of intensive agriculture, the molar ratio (Ca + Mg)/HCO3 in surface waters is significantly higher (0.7 on average) than in areas of low anthropogenic pressure (0.5). This corresponds to a decrease in riverine alkalinity, which can reach 80% during storm events. This relative loss of alkalinity correlates well with the NO 3 - content in surface waters. In cultivated areas, the contribution of atmospheric/soil CO2 to the total riverine alkalinity (CO2 ATM-SOIL/HCO3) is less than 50% (expected value for carbonate basins), and it decreases when the nitrate concentration increases. This loss of alkalinity can be attributed to the substitution of carbonic acid (natural weathering pathway) by protons produced by nitrification of N-fertilizers (anthropogenic weathering pathway) occurring in soils during carbonate dissolution. As a consequence of these processes, the alkalinity over the last 30 years shows a decreasing trend in the Save river (one of the main Garonne river tributaries, draining an agricultural catchment), while the nitrate and calcium plus magnesium contents are increasing. We estimated that the contribution of atmospheric/soil CO2 to riverine alkalinity decreased by about 7–17% on average for all the studied catchments. Using these values, the deficit of CO2 uptake can be estimated as up to 0.22–0.53 and 12–29 Tg 1 yr−1 CO2 on a country scale (France) and a global scale, respectively. These losses represent up to 5.7–13.4% and only 1.6–3.8% of the total CO2 flux naturally consumed by carbonate dissolution, for France and on a global scale, respectively. Nevertheless, this loss of alkalinity relative to the Ca + Mg content relates to carbonate weathering by protons from N-fertilizers nitrification, which is a net source of CO2 for the atmosphere. This anthropogenic CO2 source is not negligible since it could reach 6–15% of CO2 uptake by natural silicate weathering and could consequently partly counterbalance this natural CO2 sink.

The dissolution and precipitation rates of boehmite, AlOOH, at 100.3 °C and limited precipitation kinetics of gibbsite, Al(OH)3, at 50.0 °C were measured in neutral to basic solutions at 0.1 molal ionic strength (NaCl + NaOH + NaAl(OH)4) near-equilibrium using a pH-jump technique with a hydrogen-electrode concentration cell. This approach allowed relatively rapid reactions to be studied from under- and over-saturation by continuous in situ pH monitoring after addition of basic or acidic titrant, respectively, to a pre-equilibrated, well-stirred suspension of the solid powder. The magnitude of each perturbation was kept small to maintain near-equilibrium conditions. For the case of boehmite, multiple pH-jumps at different starting pHs from over- and under-saturated solutions gave the same observed, first order rate constant consistent with the simple or elementary reaction: Al ( OOH ) ( cr ) + H 2 O ( l ) + OH - ⇄ Al ( OH ) 4 - . This relaxation technique allowed us to apply a steady-state approximation to the change in aluminum concentration within the overall principle of detailed balancing and gave a resulting mean rate constant, (2.2 ± 0.3) × 10−5 kg m−2 s−1, corresponding to a 1σ uncertainty of 15%, in good agreement with those obtained from the traditional approach of considering the rate of reaction as a function of saturation index. Using the more traditional treatment, all dissolution and precipitation data for boehmite at 100.3 °C were found to follow closely the simple rate expression: R net,boehmite = 10 - 5.485 { m OH - } { 1 -exp ( Δ G r / RT ) } , with Rnet in units of mol m−2 s−1. This is consistent with Transition State Theory for a reversible elementary reaction that is first order in OH− concentration involving a single critical activated complex. The relationship applies over the experimental ΔGr range of 0.4–5.5 kJ mol−1 for precipitation and −0.1 to −1.9 kJ mol−1 for dissolution, and the pHm ≡ −log(mH+) range of 6–9.6. The gibbsite precipitation data at 50 °C could also be treated adequately with the same model: R net,gibbsite = 10 - 5.86 { m OH - } { 1 -exp ( Δ G r / RT ) } , over a more limited experimental range of ΔGr (0.7–3.7 kJ mol−1) and pHm (8.2–9.7).

The Merouane Chott, located in southeastern Algeria, is an ideal natural system for studying mineral dissolution and precipitation rates because it (1) undergoes annual cycles of filling and complete evaporation, and (2) has a simple, well defined hydrology. The major element concentrations of Merouane Chott lake waters were measured weekly from January to June 2003. These concentrations are used together with estimates of the Chott lake volume and infiltration rates to calculate the temporal evolution of the total mass of these major elements. Element precipitation rates are generated by dividing the first derivative of total mass of each element with respect to time by the Chott surface area. Mass balance considerations yield precipitation rate estimates for halite, calcite, and gypsum during the complete evaporation of the lake. These rates are compared with the saturation indexes of these minerals to deduce the degree to which they are consistent with laboratory measured rates available in the literature.

International audience; Sorption of the 14 Rare Earth Elements (REE) by basaltic rock is investigated as a function of pH, ionic strength and aqueous REE concentrations. The rock sample, originating from a terrestrial basalt flow (Rio Grande do Sul State, Brazil), is composed of plagioclase, pyroxene and cryptocrystalline phases. Small amounts of clay minerals are present, due to rock weathering. Batch sorption experiments are carried out under controlled temperature conditions of 20°C with the < 125 ?m fraction of the ground rock in solutions of 0.025 M and 0.5 M NaCl and at pH ranging from 2.7 to 8. All 14 REEs are investigated simultaneously with initial concentrations varying from 10-7 to 10-4 mol/L. Some experiments are repeated with only europium present to evaluate possible competitive effects between REE. Experimental results show the preferential retention of the heavy REEs at high ionic strength and circumneutral pH conditions. Moreover, results show that REE sorption increases strongly with decreasing ionic strength, indicating two types of sorption sites: exchange and specific sites. Sorption data are described by a Generalised Composite (GC) non-electrostatic model: two kinds of surface reactions are treated, i.e. cation exchange at >XNa sites, and surface complexation at >SOH sites. Total site density (>XNa + >SOH) is determined by measuring the cation exchange capacity (CEC = 52 µmol/m2). Specific concentrations of exchange sites and complexation sites are determined by fitting the Langmuir equation to sorption isotherms of REE and phosphate ions. Site densities of 22 ± 5 and 30 ± 5 µmol/m2 are obtained for [>XNa] and [>SOH], respectively. The entire set of REE experimental data is modelled using a single exchange constant (log Kex = 9.7) and a surface complexation constant that progressively increases from logK = -1.15 for La(III) to -0.4 for Lu(III). The model proves to be fairly robust in describing other aluminosilicate systems. Maintaining the same set of sorption constants and only adjusting the site densities, we obtain good agreement with literature data on REE/kaolinite and REE/smectite sorption. The Generalised Composite non-electrostatic model appears as an easy and efficient tool for describing sorption by complex aluminosilicate mineral assemblages.

Steady-state talc dissolution rates, at far-from-equilibrium conditions, were measured as a function of aqueous silica and magnesium activity, pH from 1 to 10.6, and temperature from 25 to 150 °C. All rates were measured in mixed flow reactors and exhibited stoichiometric or close to stoichiometric dissolution. All measured rates at pH > 2 obtained at a fixed ionic strength of 0.02 M can be described to within experimental uncertainty using r + = s BET A A a H + 2 a Mg 2 + 1 / 4 + A B exp ( - E A / RT ) where r+ signifies the BET surface area normalized forward talc steady-state dissolution rate, sBET denotes the BET surface area of talc present in the reactor, AA and AB refer to pre-exponential factors equal to 5.0 × 10−9 and 0.8 × 10−9 mol/cm2/s, respectively, EA designates an activation energy equal to 45 kJ mol−1, R represents the gas constant, T denotes absolute temperature, and ai refers to the activity of the subscripted aqueous species. The first term of this rate expression is consistent with talc dissolution rates at acidic pH being controlled by the detachment of partially liberated silica tetrahedral formed at talc edge surfaces from the exchange of Mg2+ for two protons. Corresponding atomic force microscopic observations confirms that dissolution proceeds by the removal of T-O-T sheets from talc edges. At pH ⩽ 2, the Mg2+ for proton exchange is so extensive that talc T-O-T sheets break apart leading to increased surface area and accelerated rates, whereas rates appear to be pH independent at pH ⩾ 7.

The stability of yttrium–acetate (Y–Ac) complexes in aqueous solution was determined potentiometrically at temperatures 25–175 °C (at Ps) and pressures 1–1000 bar (at 25 and 75 °C). Measurements were performed using glass H+-selective electrodes in potentiometric cells with a liquid junction. The species YAc2+ and YAc 2 + were found to dominate yttrium aqueous speciation in experimental solutions at 25–100 °C (log [Ac−] YAc 3 0 into the Y–Ac speciation model was necessary. The overall stability constants βn were determined for the reaction Y 3 + + n Ac - = YAc n 3 - n β n = YAc n 3 - n [ Y 3 + ] · [ Ac - ] n · γ YAc n 3 - n γ Y 3 + · γ Ac - n where terms in square parentheses stand for the molal concentrations [mol/kg H2O], and γi is the activity coefficient of the ith species. With temperature increase from 25 to 200 °C, the values of log β1, log β2 and log β3 increase from 2.27 to 4.51, 3.86 to 7.86 and 4.63 to 10.24, respectively. Increase of pressure results in decrease of the stability of all three complexes: log β1 decreases by 0.21 and 0.37, log β2 by 0.40 and 0.73, and log β3 by 0.57 and 1.01 when pressure increases from 1 to 1000 bar at 25 and 200 °C, respectively. However, the effect of pressure on Y speciation is much less pronounced than that of temperature and results only in a decrease of the predominance field of YAc 3 0 in high-pressure solutions. The calculated values of the partial molal thermodynamic functions for the overall formation reactions are: ΔrS∘ = 23.4, 42.5, and 54.5 cal/mol K, Δ r C p ∘ = 73.2 , 122.2, and 182.5 cal/mol K, and ΔrV∘ = 16.1, 30.9, 44.3 cm3/mol for YAc2+, YAc 2 + and YAc 3 0 , respectively. The revised Helgeson–Kirkham–Flowers (HKF) equation of state parameters were generated for Y–Ac complexes.

The sorption of Eu(III) onto kaolinite and montmorillonite was investigated up to 150 °C. The clays were purified samples, saturated with Na in the case of montmorillonite. Batch experiments were conducted at 25, 40, 80 and 150 °C in 0.5 M NaClO4 solutions to measure the distribution coefficients (Kd) of Eu as a trace element (

The dacite pumice erupted from Mt. Pinatubo on June 15, 1991 (whole-rock, rhyolitic groundmass glasses and homogenized melt inclusions) has been analyzed using inductively coupled plasma-mass spectrometry (ICP-MS), nanosecond and femtosecond laser ablation ICP-MS and secondary ion mass spectrometry (SIMS) to evaluate its ore-forming potential. Data suggest that adakite magmas are metal-rich and concentrate ore metals during magmatic differentiation. Sulfides segregate in limited amounts under the hydrous, oxidizing conditions typical of adakitic magmas resulting in incompatible behavior for Au (6–22 ppb), Cu (26–77 ppm), and Pb, Mo, As, and Sb in melts of dacitic to rhyolitic compositions. Metal transfer from this adakite magma to the coexisting aqueous phase was favored by the peraluminous composition of the rhyolitic melt and high aqueous chloride concentrations. Mass balance calculations suggest that the pre-eruptive aqueous phase could have extracted a minimum of 100 t Au and 5 × 105 t Cu from the Mt. Pinatubo magma. Our data suggest that intrusives having adakitic signatures are genetically associated with Au–Cu and Cu–Mo mineralization, auriferous porphyry copper deposits, and epithermal gold veins. High H2O, Cl, Sr/Y, Pb/Ce, Mo/Ce, As/Ce and Sb/Ce in Mt. Pinatubo melts reflect the contribution of deep fluids derived from subducted sediments and altered MORBs in the dacite genesis. The slab-derived fluids carrying mobile elements are likely responsible for the enrichment of adakite magmas in gold, associated metals and H2O, and may explain the exceptional ore-forming potential of adakite magmatism.

International audience; Adsorption of germanium on goethite was studied at 25 °C in batch reactors as a function of pH (1 12), germanium concentration in solution (10?7 to 0.002 M) and solid/solution ratio (1.8 17 g/L). The maximal surface site density determined via Ge adsorption experiments at pH from 6 to 10 is equal to 2.5 ± 0.1 ?mol/m2. The percentage of adsorbed Ge increases with pH at pH < 9, reaches a maximum at pH ? 9 and slightly decreases when pH is further increased to 11. These results allowed generation of a 2-pK Surface Complexation Model (SCM) which implies a constant capacitance of the electric double layer and postulates the presence of two Ge complexes, >FeO Ge(OH)30 and >FeO GeO(OH)2?, at the goethite-solution interface. Coprecipitation of Ge with iron oxy(hydr)oxides formed during Fe(II) oxidation by atmospheric oxygen or by Fe(III) hydrolysis in neutral solutions led to high Ge incorporations in solid with maximal Ge/Fe molar ratio close to 0.5. The molar Ge/Fe ratio in precipitated solid is proportional to that in the initial solution according to the equation (Ge/Fe)solid = k × (Ge/Fe)solution with 0.7 ? k ? 1.0. The structure of adsorbed and coprecipitated Ge complexes was further characterized using XAFS spectroscopy. In agreement with previous data on oxyanions adsorption on goethite, bi-dentate bi-nuclear surface complexes composed of tetrahedrally coordinated Ge attached to the corners of two adjacent Fe octahedra represent the dominant contribution to the EXAFS signal. Coprecipitated samples with Ge/Fe molar ratios >0.1, and samples not aged in solution (

International audience; Comparison of measured far-from-equilibrium dissolution rates of natural glasses and silicate minerals at 25 °C and pH 4 reveals the systematic effects of crystallinity and elemental composition on these rates. Rates for both minerals and glasses decrease with increasing Si:O ratio, but glass dissolution rates are faster than corresponding mineral rates. The difference between glass and mineral dissolution rates increases with increasing Si:O ratio; ultra-mafic glasses (Si:O ? 0.28) dissolve at similar rates as correspondingly compositioned minerals, but Si-rich glasses such as rhyolite (Si:O ? 0.40) dissolve ?1.6 orders of magnitude faster than corresponding minerals. This behaviour is interpreted to stem from the effect of Si O polymerisation on silicate dissolution rates. The rate controlling step of dissolution for silicate minerals and glasses for which Si:O > 0.28 is the breaking of Si O bonds. Owing to rapid quenching, natural glasses will exhibit less polymerisation and less ordering of Si O bonds than minerals, making them less resistant to dissolution. Dissolution rates summarized in this study are used to determine the Ca release rates of natural rocks at far-from-equilibrium conditions, which in turn are used to estimate their CO2 consumption capacity. Results indicate that Ca release rates for glasses are faster than those of corresponding rocks. This difference is, however, significantly less than the corresponding difference between glass and mineral bulk dissolution rates. This is due to the presence of Ca in relatively reactive minerals. In both cases, Ca release rates increase by ?two orders of magnitude from high to low Si:O ratios (e.g., from granite to gabbro or from rhyolitic to basaltic glass), illustrating the important role of Si-poor silicates in the long-term global CO2 cycle.

The solubility of gold was measured in KCl solutions (0.001–0.1 m) at near-neutral to weakly acidic pH in the presence of the K-feldspar-muscovite-quartz, andalusite-muscovite-quartz, and pyrite-pyrrhotite-magnetite buffers at temperatures 350 to 500°C and pressures 0.5 and 1 kbar. These mineral buffers were used to simultaneously constrain pH, f(S2), and f(H2). The experiments were performed using a CORETEST flexible Ti-cell rocking hydrothermal reactor enabling solution sampling at experimental conditions. Measured log m(Au) (mol/kg H2O) ranges from −7.5 at weakly acid pH to −5.9 in near-neutral solutions, and increases slightly with temperature. Gold solubility in weakly basic and near-neutral solutions decreases with decreasing pH at all temperatures, which implies that Au ( HS ) 2 is the dominant Au species in solution. In more acidic solutions, solubility is independent of pH. Comparison of the experimentally measured solubilities with literature values for Au hydrolysis constants demonstrates that at 350°C AuHS ( aq ) 0 dominates Au aqueous speciation at the weakly acidic pH and f(S2)/f(H2) conditions imposed by the pyrite-pyrrhotite-magnetite buffer. In contrast, at temperatures >400°C AuHS ( aq ) 0 becomes less important and AuOH ( aq ) 0 predominates in weakly acid solutions. Solubility data collected in this study were used to calculate the following equilibrium reaction constants: Au ( cr ) + H 2 S ( aq ) 0 = AuHS ( aq ) 0 + 0.5 H 2 ( g ) , K ( AuHS 0 ) and Au ( cr ) + H 2 S ( aq ) 0 + HS − = Au ( HS ) 2 − + 0.5 H 2 ( g ) , K ( Au ( HS ) 2 − ) The calculated log K ( AuHS 0 ) values are −5.25 ± 0.35 and −5.92 ± 0.36 at 350°C (500–1000 bars) and 400°C/1000 bars, respectively. The log K ( Au ( HS ) 2 - ) values are −1.68 ± 0.42 at 350°C/500 bars, −1.40 ± 0.41 at 350°C/1000 bars, −2.14 ± 0.42 at 400°C/ 500 bars, −1.75 ± 0.40 at 400°C/1000 bars, −1.87 ± 0.40 at 450°C/1000 bars, and −1.80 ± 0.40 at 500°C/1000 bars. Equilibrium constants for both reactions are retrograde with respect to temperature, but this effect is more pronounced for AuHS0(aq) than for Au(HS)2−. The K values’ dependence on temperature, as well as the calculated values of K ( Au ( HS ) 2 - ) are consistent with the data of Benning and Seward (1996) . The calculated values of K ( AuHS 0 ) are between those determined by Benning and Seward (1996) and Gibert et al. (1998) .

International audience; Geospeedometry allows to estimate the cooling rate (sinit) of metamorphic rocks at the beginning of the cooling history using diffusion data. But the choice of a diffusion activation energy (E) and a preexponential factor (D0) from experimental results can be difficult. We propose a method to obtain E directly from the rock itself by studying the variation of the average concentration of elements or isotopes (?C?) as a function of mineral grain size (d). An appropriate value of D0 can then be estimated using an existing compensation rule, a linear relationship between log D0 and E. Consequently, uncertainties on sinit are markedly reduced. All parameters of this analytical model and their sensitivity on sinit can be estimated from ?C? of the mineral grains under study. As a test we apply our model to a study by Edwards and Valley (1998)**** on 18O/16O fractionation between diopside and calcite in Adirondacks marbles, and find a cooling rate in agreement with previous works, without choosing experimental values for E and D0.

International audience; The microstructure of monazite was studied using scanning electron microscopy (SEM), electron microprobe analysis (EMP), X-ray diffraction patterns (XRD), and transmission electron microscopy (TEM). Four well-characterized monazites were investigated, having very different concordant U-Pb ages (24 to 1928 Ma), and up to ~15wt.% ThO2, and 0.94wt.% UO2. The SEM and EMP analyses of polished single crystal fragments reveal the absence of significant chemical zoning. XRD and TEM investigations show that the monazites are not metamict, despite their old ages, very high abundances of radionuclides, and hence, high time-integrated radiation doses. Except for the youngest one, the monazite crystals are composed of a mosaic of crystalline but slightly distorted domains. This structure is responsible for the presence of (1) mottled diffraction contrasts on the TEM, and (2) a second structural phase (B), with very broad reflections in the XRD patterns. Older monazites receive higher self-irradiation doses, and hence, they contain higher amounts of this B-phase. For the 1928 Ma monazite, XRD reveals only the broad reflections of phase B, implying that the whole monazite was affected by radiation damage that resulted in total distortion of the lattice. It is concluded that radiation damage in the form of amorphous domains does not accumulate in monazite because selfannealing heals the defects as they are produced by alpha-decay damage. The only memory of irradiation-induced defects is the presence of distorted domains. As the diffusion rate of Pb in an undisturbed monazite lattice is extremely low, Pb loss due to volume diffusion out of the monazite lattice is virtually impossible. This is considered as one reason why almost all monazites have concordant U-Th-Pb ages. Moreover, as long-term self-irradiation effects are limited in monazite, we consider this phase as a good candidate for the storage of high-level nuclear waste under the aspect of its high resistance to irradiation.

International audience; We have used a new thermodynamic model of barium and calcium sulfate solubilities in multicomponent electrolyte solutions (Monnin, 1999) to investigate the stabilities of barite and anhydrite in seawater or in marine sediment porewaters at high temperature and pressure. As a further test supplementing those previously carried out during model development, we have calculated the temperature at which standard seawater becomes saturated with respect to anhydrite. The model predicts that, upon heating at 500 bars, standard seawater becomes saturated with respect to anhydrite at 147 ± 5°C, which compares well with the literature value of 150°C (Bishoff and Seyfried, 1978). At 20 bars the calculated saturation temperature is 117 ± 3°C. This points to a non negligible pressure effect even at moderate pressures. We have calculated the barite and anhydrite saturation indices for the in situ temperatures and pressures, from the composition of porewaters collected at ODP Sites 855, 856, 857, 858, 1035 and 1036 during ODP Legs 139 and 169 (Juan de Fuca and Gorda ridges, NE Pacific). Calculated saturation indices for porewater samples collected at depths corresponding to temperatures between 70° and 110-120°C at an in situ pressure of about 260 bars yield equilibrium values for anhydrite and barite. Saturation indices of samples collected at depths where the temperature exceeds 110-120°C, however, yield values indicating supersaturation with respect to anhydrite and undersaturation with respect to barite. This result is consistent with the redissolution of anhydrite during cooling, leading to the well documented sampling artifact affecting porewater compositions in high temperature marine sediments: anhydrite dissolution increases the porewater sulfate content, which in turn induces a loss of barium from solution through barite precipitation (the common ion effect). We postulate that this redissolution occurs in sediment samples for which the in situ temperature exceeds 110-120°C: below this limit anhydrite remains at equilibrium or does not have time to significantly dissolve before porewaters are sampled.

The solubility of boehmite, γ-AlOOH, was measured in a modified hydrogen-electrode concentration cell, which provided continuous in situ measurements of hydrogen ion molality over the range of ionic strengths from 0.1 to 5.0 mol · kg−1 (NaCl) at temperatures from 100 to 290°C. A series of conventional solubility measurements was also carried out in acidic solutions over the same temperature range (i.e., pH was not monitored, but rather calculated from mass balance). The combined results yielded the molal solubility quotients, Qs0 and Qs4 for the equilibria: AlOOH(cr) + 3H+ ⇄ Al3+ + 2H2O Qs0 = [Al3+]/[H+]3 AlOOH(cr) + 2H2O ⇄ Al(OH)4− + H+ Qs4 = [Al(OH)4−][H+]In the regression of each isothermal data set, the values for the first hydrolysis quotient, Al3+ + H2O ⇄ Al(OH)2+ + H+ Q1,1 = [Al(OH)2+][H+]/[Al3+]were fixed according to a previous potentiometric study (Palmer and Wesolowski, 1993) . Moreover, for one series of titrations at 0.1 mol · kg−1 ionic strength at 150°C, the remaining two solubility quotients, Qs2 and Qs3, were determined simultaneously from the regression. However, at all other conditions, the values of Qs2 and Qs3 were also fixed at values consistent with the corresponding 0.03 mol · kg−1 ionic strength results (Benezeth et al., 2001) by invoking the isocoulombic assumption (i.e., an assumption of minimal ionic strength dependence for reactions with no net change in charge). The stability field with respect to pH of the Al(OH)2+ and Al(OH)30 aqueous species were found to be very narrow, and hence assumptions concerning their stabilities had little effect on the predicted shape of the solubility profiles at high ionic strength. Global fits were made of the log Qs0 and log Qs4 values as functions of temperature and ionic strength after combining with corresponding values from an analogous study at 0.03 mol · kg−1 ionic strength (Benezeth et al., 2001) , as well as appropriate constants taken from the literature. The fits were further constrained by inclusion of gibbsite solubility data ( Wesolowski 1992 , Palmer and Wesolowski 1992 after adjustment for the relative free energies of formation of gibbsite and boehmite. This treatment ensured that a continuous empirical model exists for Al3+ and Al(OH)4− from ambient conditions to 300°C and infinite dilution to five mol · kg−1 ionic strength.

The solubility of pure synthetic boehmite (γ-AlOOH) has been measured over a wide range of pH (2–10, depending on temperature), and temperature (100–290°C) at 0.03 mol · kg −1 ionic strength (NaCl) in a hydrogen-electrode concentration cell (HECC), which provided continuous in situ measurement of hydrogen ion molality (Palmer et al. 2001) . A least-squares regression of the results obtained was used to determine the molal solubility quotients (Q s0 to Q s4 ) of boehmite. The solubility products (Q sn ) were extrapolated to infinite dilution (K sn ), permitting calculation of the thermodynamic properties of aqueous species of aluminum for comparison with previous works, such as Bourcier et al. (1993) and Castet et al. (1993) . These results are generally consistent with the latter study conducted near infinite dilution, although some significant differences are apparent, particularly near the solubility minimum.

Institute GEMOC Key Centre, Department of Earth andPlanetary Sciences, Macquarie University, Sydney, NSW2109, AustraliaOn a global scale, several authors have reported good correla-tions between physical erosion fluxes and chemical weatheringfluxes (Gaillardet et al., 1999). This linkage have several origins,one being that the active floodplains can act as a biogeochemicalreactor in which the sediments produced in mountainous zonesand transported by the rivers, once deposited in foreland basins,are deeply weathered.We have tested this idea in the floodplain of the Rio Beni, inthe upper Madeira Basin, in Bolivia.Extreme physical denudation rates are found in the BolivianAndes, in particular due to the abundance of poorly lithifiedshales and of extensive outcrops of fluviolacustrine deposits inthe Upper Andes. These river sediments are exported and 40%is deposited in the foreland basin of the Beni R. where extensivefloodplains are submitted to wet tropical conditions.We compare here monthly collected river samples upstream(Rurrenabaque) and downstream (Riberalta) the Rio Beni flood-plain and averaged dissolved fluxes (over one year). Concentra-tions of major elements are corrected from atmospheric inputs.The main result is that the floodplain has almost no effect onthe chemical denudation rates of the Rio Beni, except loweringthem because of an increase in the drainage surface area.Although long-term weathering in the floodplain is more intensethan in the Andes, the flux of dissolved material released doesnot impact the Rio Beni chemical denudation rates. The differentelements show particular behavior that we will discussed in detail.We conclude that the foreland basin of the Rio Beni acts aboveall as a deposition area and does not contribute to the actualchemical weathering rates. This conclusion supports the idea thatchemical erosion is essentially active in mountainous areas andthat the high temperatures of the plain do not increase the chem-ical denudation rates. In the Beni river basin, tectonic character-istics (uplift and subsidence zone) appear as the first-orderfactors controlling chemical weathering fluxes.

International audience; Not Available

International audience

International audience; The ion-interaction model of PITZER (1973), is very effective in deriving stability relationships at high concentrations for the system Na - Cl - HCO 3 - CO 3 - OH - H 2 O . The solubility products of the main sodium carbonates have been calculated from solubility data between 5 and 50°C. The stability diagram in log p co2 -- temperature coordinates and the invariant points deduced from the newly determined data are in good agreement with the most recent measurements. These results are used to calculate the activities of the major dissolved species in Lake Magadi brines (Kenya). The thermodynamic treatment confirms the main conclusions reached earlier by Eugster (1970, 1980) mainly from field observations. Trona precipitation occurs at equilibrium while natron is likely to form when the temperature decreases below 25°C. After the salt deposition the CO 2 supply from the atmosphere is too slow to allow equilibrium between the atmosphere and the brines. In the next stages of evaporative concentration thermonatrite and halite precipitate. The deposition of the latter salts along with the observed HCO - 3 depletion suggest that fractional crystallization is likely to control trona deposition.