Michael E. Berndt

Reduction of CO2 during serpentinization of olivine at 300 °C and 500 bar

CO 2 reduction processes occurring during experimental serpentinization of olivine at 300 °C and 500 bar confirm that ultramafic rocks can play an important role in the generation of abiogenic hydrocarbon gas. Data reveal that conversion of Fe(II) in olivine to Fe(III) in magnetite during serpentinization leads to production of H 2 and conversion of dissolved CO 2 to reduced-C species including methane, ethane, propane, and an amorphous carbonaceous phase. Hydrocarbon gases generated in the process fit a Schulz-Flory distribution consistent with catalysis by mineral reactants or products. Magnetite is inferred to be the catalyst for methanization during serpentinization, because it has been previously shown to accelerate Fischer-Tropsch synthesis of methane in industrial applications involving mixtures of H 2 and CO 2 . The carbonaceous phase was predominantly aliphatic, but had a significant aromatic component. Although this phase should ultimately be converted to hydrocarbon gases and graphite, if full thermodynamic equilibrium were established, its formation in these experiments indicates that the pathway for reduction of CO 2 during serpentinization processes is complex and involves a series of metastable intermediates.

Phase equilibria constraints on the chemistry of hot spring fluids at mid-ocean ridges.

Abstract Recent advances in experimental and theoretical geochemistry have made it possible to assess both homogeneous and heterogeneous equilibria involving a wide range of aqueous species at temperatures and pressures appropriate to model hydrothermal alteration processes at mid-ocean ridges. We have combined selected aspects of the chemistry of hot spring fluids with constraints imposed by a geologically reasonable assemblage of minerals in the system Na2O-K2O-CaO-MgO-FeO-Fe2OrAl2O3-SiO2-H2O-HCl-H2S to assess the effect of temperature on the composition of the aqueous phase and the activities of mineral components in plagioclase and epidote solid solutions. Assuming ƒO 2(g) and ƒS 2(g) controlled by pyrite-pyrrhotite-magnetite equilibria, a constant dissolved Ca concentration, and a dissolved Cl concentration equivalent to that of seawater, increasing temperature from 250 to 400°C at 500 bars results in systematic changes in the composition of mineral phases, which in turn constrain pH and the distribution of aqueous species. The model predicts that dissolved concentrations of Fe, SiO2, K, H2S, and H2 increase, while Na and pH(25°c) decrease with increasing temperature. pH(in-situ) decreases slightly with increasing temperature, and has a value of 5.16 at 400°C. Dissolved Mg concentrations do not exceed 1 mmolal at any temperature investigated. Allowing for differences in pressure and total dissolved Cl, the predicted effect of temperature on fluid chemistry is in good agreement with results from basalt alteration and plagioclase + epidote + quartz recrystallization experiments, and in some cases, with results from hot spring fluids at mid-ocean ridges. Some vent fluids, in particular NGS (EPR, 21°N) and vent-4 (EPR, 11°N), however, reveal pH and/or dissolved Fe, H2S, H2, and SiO2 concentrations, which are difficult to reconcile with measured temperatures in comparison with results of temperature dependent mineral solubility calculations, and suggest that these fluids have lost heat by conduction on ascent to the seafloor. That many hot spring vent fluids are characterized by variable degrees of conductive heat loss renders measured temperatures unreliable as indicators of the maximum temperature of subseafloor hydrothermal alteration processes. The implications of this are significant for hot spring fluids which reveal large Cl variations relative to seawater, since likely mechanisms to account for such variability typically require temperatures in excess of those inferred for subseafloor reaction zones by simply correcting measured temperatures for the effects of adiabatic cooling.

Boron, bromine, and other trace elements as clues to the fate of chlorine in mid-ocean ridge vent fluids

Fluids from mid-ocean ridge hot springs typically have Cl concentrations which depart significantly from seawater values. These variations may be due in part to phase separation processes and/ or precipitation and dissolution of chloride-bearing minerals. Both of these processes likely produce systematic and recognizable variations in the distributions of trace elements which should be evident in vent fluid chemistries. To better understand how supercritical phase separation can affect trace element distributions, we conducted an experiment involving a Na-Ca-K-Cl fluid containing trace quantities of Sr, Ba, B, Li, and Br, which was allowed to separate into vapor and brine phases at 425, 440, and 450°C by systematically adjusting pressure. All of the measured trace elements were concentrated into the brine phase relative to the vapor phase. The relative order of partitioning into the brine was Ba > Sr > Ca > K > Na,Cl > Li > Br > B. Comparing the experimentally determined distribution of trace elements, especially B and Br, and Cl, with analogous data for vent fluids leads to the interpretation that both phase separation and mineral precipitation processes occur in ridge crest hot spring systems. Fluids with Cl greatly different from seawater may require a phase separation process at temperatures greater than 450°C to explain the BSW (boron from a seawater source) and Br data. Furthermore, there is evidence that nearly all of the fluids venting from ridge crest systems, whether phase separation is indicated or not, appear to have experienced some degree of precipitation of a Cl mineral which may be similar to that formed in hydrothermal experiments. Relative concentrations of trace alkali and Brock (boron derived from rock) in vent fluids suggest that fluids interact directly with recently cooled magma prior to phase separation but close to or within a cracking front region located slightly above the brittle/ductile transition zone surrounding the magma chamber. Vent fluids in such systems represent the combined chemistries of mixed brine, vapor, and hydrothermal fluid which has reequilibrated with the host rock at temperatures much less than those required for phase separation.

Plagioclase and epidote buffering of cation ratios in mid-ocean ridge hydrothermal fluids: Experimental results in and near the supercritical region

Abstract Experiments have been performed with Na-Ca-K-Cl fluids of seawater chlorinity and diabase, basalt, and plagioclase bearing mineral mixtures at 350–425 °C and 250–400 bars to help constrain hydrothermal alteration processes at mid-ocean ridges. Dissolved Ca, Na, and pH for all experiments responded systematically to differences in dissolved SiO2 concentrations and the compositions of plagioclase reactants. Diabase alteration at low fluid/rock mass ratios (0.5 to 1) produces fluids undersaturated with respect to quartz during hydration of primary olivine and orthopyroxene, whereas basalt alteration under similar conditions yields fluids slightly supersaturated with respect to quartz during breakdown of glass to smectite and amphibole. Fluid chemistry in all experiments appears to approach a partial equilibrium state with the albite and anorthite components in plagioclase and approaches a pH consistent with plagioclase alteration to epidote. Trace element data from vent fluids, specifically B and Sr, together with major element chemistry (Ca, Na, SiO2, pH), provides evidence that the reaction zone for “black-smoker” fluids at mid-ocean ridges is composed of only slightly altered diabase and is characterized by small amounts of epidote, nearly fresh plagioclase and clinopyroxene, and partially to completely hydrated olivine and orthopyroxene. Fluids reacting with this rock may be undersaturated with respect to quartz so pressure estimates based on the quartz geobarometer should be regarded as minimums. Using equilibrium between plagioclase, the dominant reactant, and epidote, the dominant reaction product in experiments, we estimate that temperatures in reaction zones are in excess of 375°C for most vent systems. These temperatures are higher than measured vent temperatures, suggesting that hotspring fluids commonly loose heat during ascent to the seafloor.

Chloride depletions and enrichments in seafloor hydrothermal fluids: Constraints from experimental basalt alteration studies

Na-K-Ca-Cl fluid was reacted with diabase at 400° and 425°C, 400 bars and fluidrock mass ratio of 0.5 to assess the relative mobility of dissolved Cl. Fluids from the present experiments reveal relatively large time and temperature dependent decreases in Cl; temperature increase enhances Cl removal, whereas reaction progress has the opposite effect. These changes are not due to boiling and/or phase separation. At 400° and 425°C, 400 bars, Cl removal from solution is accompanied by Ca and Na fixation, respectively. Both experiments produced mixed-layer chlorite/smectite, clinozoisite and albite(?). Olivine was not detected by X-ray diffraction or petrographic analysis of run products from either experiment, however, we propose that the time dependent changes in dissolved Cl result from olivine replacement by an Fe-hydroxy chloride phase followed by hydration effects. Quenched and washed alteration products, however, did not reveal anomalous Cl, which suggests that the Cl-bearing phase is characterized by retrograde solubility. This is consistent with the relative magnitude of Cl fixation observed for experiments at both temperatures and from results of an isobaric cooling experiment. The experiments provide evidence for the non-conservative behavior of Cl during hydrothermal alteration of basalt. The data are particularly important in light of the Cl-depleted nature of some ridge crest hot spring fluids, and suggest temperatures of formation for these fluids of approximately 410°C, assuming sub-seafloor pressure of 400 bars. In addition, the retrograde solubility of the Cl-bearing phase responsible for Cl fixation during high temperature basalt alteration, may help to explain Cl enrichment in hot spring fluids which have conductively cooled below 350°C

Calibration of BrCl fractionation during subcritical phase separation of seawater: Possible halite at 9 to 10°N East Pacific Rise

Abstract Experiments were conducted to assess Br Cl fractionatiodn during phase separation of seawater at 400°C and 250 to 275 bars. These conditions are applicable to the 9–10°N EPR system where low Cl concentrations and low Br Cl rations in vent fluids have been attributed to phase separation of seawater at conditions below the critical point of seawater (408°C, 300 bars). The level of Br Cl fractionation observed in experiments is well below that needed to account for Br Cl systematics at EPR 9–10°N. Based on our experimental results, we propose an alternative model involving dissolution and precipitation of halite to account for the anomalous Br Cl data at 9–10°N. Halite can be predicted to form at pressure and temperature conditions prevailing during eruption of magma at mid-ocean ridges. Subsequent changes in physical conditions, however, would induce halite to dissolve, and thus, decrease Br Cl ratios of vent fluids.

Boron isotope fractionation during supercritical phase separation

Abstract Boron isotopic fractionation between vapor and brine phases separated under supercritical conditions has been experimentally examined using a Na-Ca-K-Cl fluid. Experiments were conducted at 425, 440, and 450°C. Phase separation was controlled by adjusting pressure. Fractionation between the coexisting phases was less than 0.5%.. In contrast, fractionation between trigonal and tetrahedral B, at these temperatures, has been predicted to be approximately 8%.. The lack of fractionation suggests that the trigonal/tetrahedral speciation of B is similar in the two phases. Although this result may not be general for other fluid compositions, it is relevant to ridge crest hydrothermal solutions and indicates that the boron isotopic compositions of these solutions reflect the proportions of B from seawater and crustal sources and not phase separation.

Hydrogen isotope systematics of phase separation in submarine hydrothermal systems: Experimental calibration and theoretical models

Abstract Hydrogen isotope fractionation factors were measured for coexisting brines and vapors formed by phase separation of NaCl/H 2 O fluids at temperatures ranging from 399–450°C and pressures from 277–397 bars. It was found that brines are depleted in D compared to coexisting vapors at all conditions studied. The magnitude of hydrogen isotope fractionation is dependent on the relative amounts of Cl in the two phases and can be empirically correlated to pressure using the following relationship: 1000 In α (vap-brine) = 2.54(±0.83) + 2.87 (±0.69) × log (Δ P ), where α (vap-brine) is the fractionation factor and Δ P is a pressure term representing distance from the critical curve in the NaCl/H 2 O system.

Calcium and sodium exchange during hydrothermal alteration of calcic plagioclase at 400°C and 400 bars

A series of experiments were conducted to measure the effect of plagioclase composition on the composition of coexisting Cl-dominated hydrothermal fluids. Assemblages containing quartz and natural plagioclase with a composition of either An60 or An82 were reacted with Na-Ca-Cl fluids at 400°C and, 400 bars using flexible cell hydrothermal facilities. Plagioclase remained homogeneous during these experiments despite the fact that homogeneous solid solutions with intermediate compositions are likely metastable with respect to mechanical mixtures at these conditions. It was found that Ca and Na exchange between fluids and minerals appears to be reversible, and highly dependent on plagioclase composition but mCa++m2Na+ ratios remained constant with increasing Cl for a given plagioclase composition. In fact, a solid solution model which approaches ideal behavior appears to account best for the chemistry of fluids coexisting with An60 and An82 plagioclase. These data may have a direct bearing on hydrothermal alteration processes responsible for the chemistry of hot spring fluids at mid-ocean ridges, which appear to have equilibrated with metastable plagioclase solid solutions of composition An60 to An70.

The stability of gold polysulfide complexes in aqueous sulfide solutions: 100 to 150°C and 100 bars

Abstract Gold solubility was measured in sulfur-saturated aqueous solutions containing small amounts of dissolved NaHS and NaOH or HCl at 100–150°C and 100 bars. Addition of NaOH and HCl to sulfur-saturated solutions creates systematic variation in pH and f o 2 and changes the relative abundances of dissolved sulfur species. Strong correlations between gold solubility and the abundances of aqueous polysulfide species suggest that Au-polysulfides (AuS n S − where n = 2–7) are the dominant Au-bearing complexes in sulfur-saturated solutions. A preliminary assessment of our data suggests that Au-polysulfide species likely dominate the transport and deposition of Au at temperatures from 100–150°C in a broad region surrounding the sulfur saturation field in pH- f o 2 - f s 2 space. Although Au(HS) − 2 complexes clearly dominate gold solubility in sulfide solutions significantly undersaturated with respect to sulfur, our results suggest that the highest gold solubilities should occur in solutions coexisting with elemental sulfur.