Search

The efficiency of three slurries was tested for treating of acidity caused by pyrite weathering. Slurries of BauxsolTM, powdered limestone, and brucite [Mg(OH)2] were sprayed onto 10-ton masses of pyritic aggregate, and the acidity and sulfate concentrations of outflows were monitored for 21.5 months. After addition in identical procedures, neutralization of the outflow to a pH of above 6 was achieved by Bauxsol for about 1 day, by limestone for 4 days, and by Mg(OH)2 for the 652 days of the testing. Limestone control of pH was limited by armoring of the CaCO3 by gypsum [CaSO4·2H2O] and by restricting of water flow to channels so that most of this slurry remained unreacted. Channels were not evident in the Mg(OH)2-treated aggregate. With the limestone and Mg(OH)2 slurries, sulfate concentrations were controlled by crystallization of gypsum, or at higher concentration of sulfate by hexahydrite [MgSO4⋅6H2O]. The limestone slurry limited sulfate concentration to approximately 25 percent less than that occurring with the Mg(OH)2 slurry. An ideal slurry might include Mg(OH)2 for neutralization, limestone for limiting sulfate concentration, and other components to alter rheological behavior. Thixotropic behavior that favors initial dispersion of the slurry solids yet reduces loss in outwash and also limits channelization of water flow is best.

Electrochemical cells are reviewed and a new design is evaluated for potentiometric pH measurements to above 300 jC. The new design system minimizes the effects of metal corrosion on measured pH. In addition, a recently developed [Zhou, X.Y., Lvov, S.N., Ulyanov, S.M., 2003. Yttria-Stabilized Zirconia Membrane Electrode, US Patent #6, S17, 694] flow-through, yttriastabilized zirconia (YSZ) pH sensor has been further tested. The Nernstian behavior and precision of the YSZ electrode were evaluated by measuring the potentials vs. H2–Pt electrode at 320 and 350 jC. Also, using the YSZ electrode, the association constants of HCl(aq) at 320 and 350 jC have been determined from the potentials of a HCl(aq) solutions at 0.01 to 0.001 mol kg � 1 . The results, pK320= � 1.46F0.46 and pK350= � 2.35F0.25, in good agreement with literature data, both demonstrate the effective use of the cell and YSZ electrode for pH measurements to about F0.05 pH units, and confirm the Nernstian behavior of the YSZ electrode in acidic HCl solutions up to 350 jC. Commercial YSZ tubes available for high-temperature pH sensing are, however, far from ideal because of irregular compositions, phase structures, and interstitial materials. A consequence is the premature structural decay of YSZ tubes in acidic solutions at elevated temperatures. In spite of the longterm decay, YSZ sensors respond rapidly to changes in pH, apparently limited only by the rate of mixing of solutions within the cell. This system makes the measurement feasible above 300 jC of mineral hydrolysis equilibrium constants and their free energy changes within uncertainties of about F1.0 kJ. D 2003 Elsevier Science B.V. All rights reserved.

The kinetics and mechanism of analcime formation from precursor Na-clinoptilolite (Cpt-Na) and Na-mordenite (Mor-Na) were investigated from 125–225 °C, pressures up to 34.5 MPa, and pH = 9.2–10.7. By using batch and flow-through experimental methodologies, compositions of solids and solutions were monitored to track reaction progress, determine rates of analcime nucleation and growth, and evaluate the rate dependency of these processes on aqueous supersaturation. Analcime formation proceeds via a clinoptilolite (or mordenite) dissolution→analcime nucleation→analcime growth sequence. Synthetic analcime crystals are similar in morphology (trapezohedron to cubic trapezohedron) and composition (Si/Al = 2.1–2.7) to sedimentary analcimes formed during the low-grade alteration of crustal rocks, evidence that the experimental reaction mechanism is similar to that in natural environments. Rates of analcime nucleation were approximated by evaluating the time-dependence of the size and number of particles and range between 109.80 and 1011.88 per h per cm3 at 150° and 225 °C, respectively. The nucleation rate is a function of temperature and degree of supersaturation: ln rate = 56.76 – 15978.9/ T + 2.99 × 10−4·Δ G r, where Δ G r is the free energy change of analcime precipitation and T is temperature in Kelvins. This rate equation is consistent with an apparent activation energy of analcime nucleation ( E a,n) of 132.8 ± 8.3 kJ/mol. Although conditions were thermodynamically favorable for analcime formation at 125 °C, nucleation was not detected after 144 hours. These data suggest that diagenetic timing of the clinoptilolite to analcime transformation is principally controlled by kinetics rather than by thermodynamic equilibrium. Rates of analcime growth were estimated by measuring particle size distributions. Average growth rates ranged from 0.15 μm/h at 150 °C to 0.396 μm/h at 225 °C, and are consistent with an apparent activation energy of analcime growth ( E a,g) of 77.1 ± 9.4 kJ/mol. These nucleation and growth parameters are combined to successfully model the evolution of analcime particle size distributions. The experimental methods developed in this study demonstrate the use of hydrothermal flow-through methods in the study of zeolite transformations.

The formation pathways of pyrite are controversial. Time resolved experiments show that in reduced sulphur solutions at low temperature, the iron monosulphide mackinawite is stable for up to 4 months. Below 100°C, the rate of pyrite formation from a precursor mackinawite is insignificant in solutions equilibrated solely with H2S(aq). Mackinawite serves as a precursor to pyrite formation only in more oxidised solutions. Controlled, intentional oxidation experiments below 100°C and over a wide range of pH (3.3–12) confirm that the mackinawite to pyrite transformation occurs in slightly oxidising environments. The conversion to pyrite is a multi-step reaction process involving changes in aqueous sulphur species causing solid state transformation of mackinawite to pyrite via the intermediate monosulphide greigite. Oxidised surfaces of precursors or of pyrite seeds speed up the transformation reaction. Solution compositions from the ageing experiments were used to derive stability constants for mackinawite from 25°C to 95°C for the reaction: FeS(s) +2 H + ⇔ Fe 2+ + H 2 S The values of the equilibrium constant, logKFeS, varied from 3.1 at 25°C to 1.2 at 95°C and fit a linear, temperature-dependent equation: logKFeS=2848.779/T−6.347, with T in Kelvin. From these constants, the thermodynamic functions were derived. These are the first high temperature data for the solubility of mackinawite, where Fe2+ is the dominant aqueous ferrous species in reduced, weakly acidic to acidic solutions.

Pyrite formation under hydrothermal conditions has been studied using in situ time-resolved X-ray diffraction (XRD). This study employed two different synchrotron X-ray sources (National Synchrotron Light Source (NSLS), Advanced Photon Source (APS); Brookhaven and Argonne National Laboratories, USA, respectively) and two types of reaction cells (capillary and hydrothermal autoclave type) to examine reactions in the Fe–S system under both anoxic and controlled oxic conditions. Starting materials were mackinawite slurries equilibrated in reduced, H 2 S (aq) solutions and heated to a maximum of 190°C. The results show that under fully anoxic conditions, mackinawite persists to at least 120°C, while aerated (oxic) slurries transform rapidly to greigite, pyrite, magnetite and goethite. Possible reaction mechanisms are discussed in light of these results. These experiments are the first application of previously described in situ reaction cells to anoxic and oxic reactions in the Fe–S system. Finally, the capabilities of each cell type and their applicability to other systems are discussed.

The hydration state of Na- and K-exchanged clinoptilolite from Castle Creek (Idaho, U.S.A.) has been measured by a pressure titration method to 300 °C and PH2O

In a shallow, fjord-like estuary, the Upper and Lower Basins of the Pettaquamscutt River (Rhode Island) each contain water columns stratified into oxic and anoxic plus sulfidic layers. Both water columns contain iron monosulfides and pyrite from just below the oxic-anoxic interface down to the sediment-water interface. Concentrations of iron monosulfides increase with increasing depth in the water columns, but concentrations of pyrite are relatively constant with depth. Suspended pyrite is present only as framboids, whereas the underlying sediments contain dominantly framboids but also euhedral grains and infilled framboids. Based on pyrite textures, S-isotopic compositions of water-column and sediment sulfides, and settling rate calculations, about 70% of the pyrite burial flux to the Pettaquamscutt sediments is accounted for by framboids forming at the oxic-anoxic interface and settling to the sediment-water interface. The remaining pyrite burial flux is secondary pyrite nucleation and growth within the sediments represented by euhedral grains and infilled framboids. The S-isotopic composition of pyrite in the Pettaquamscutt sediments, and presumably of pyrite in the sediments of other modern anoxic basins and analogous ancient sediments, therefore, contains a component originating near the oxic-anoxic interface and a component added below the sediment-water interface. Pyrite framboids form in the water column where there is high supersaturation with respect to pyrite and slight undersaturation with respect to the iron monosulfides. We propose that framboid nucleation and growth occurs in the water column, adjacent to, and below, the oxic-anoxic interface, where supplies of ferrous species, sulfide species, and suitable electron acceptors are available. Infilled and overgrown framboids from in the sediments as a result of surface nucleation and continued pyrite growth on framboids that originate above the sediment-water interface.

Pyrite framboid formation may be the result of four consecutive processes: (1) nucleation and growth of initial iron monosulfide microcrystals; (2) reaction of the microcrystals to greigite (Fe 3 S 4 ; (3) aggregation of uniformly sized greigite microcrystals, i.e., framboid growth; and (4) replacement of greigite framboids by pyrite. The uniform morphology, uniform size range, and ordering of the microcrystals in individual framboids, as well as the range of observed framboid structures from irregular aggregates to densely packed spherical aggregates and polyframboids, are consequences of these processes. Using DLVO theory (Derjaguin, Landau, Verwey, and Overbeek), we have evaluated the stability of colloidal, iron monosulfide suspensions with ionic strengths typical of marine and lacustrine waters. In addition to van der Waals attractive and double-layer repulsive forces, a term is included to account for the ferrimagnetic properties of greigite. Numerical models predict that magnetically saturated greigite particles >0.1 μm in diameter will rapidly aggregate in either marine or fresh water. The aggregation model is in agreement with the sequence of greigite formation followed by pyrite framboid formation established in a previous experimental study (Sweeney and Kaplan, 1973) and is consistent with the occurrence of framboids composed of other magnetic minerals, e.g., greigite, magnetite, and magnesioferrite. Based on the temperature-dependent magnetic properties of greigite and aging experiments in hydrothermal solutions, this mechanism for framboid formation via precursor greigite could operate to temperatures of ∼200°C, consistent with the occasional occurrence of pyrite framboids in the paragenesis of metalliferous ore deposits.

Pyrite formation has been investigated at 70°C and pH 6–8 by aging precipitated, disordered mackinawite, Fe9S8, and greigite, Fe3S4, in solutions containing aqueous H2S, HS−, Sx2−, S2O32−, SO32−, colloidal elemental sulfur, and the organic sulfur species thiol, disulfide, and sulfonate. Pyrite formed in all experiments where unoxidized iron monosulfides were aged with species containing zero-valent sulfur, i.e., polysulfides and colloidal elemental sulfur, but not with hydrogen sulfide (or bisulfide), the sulfoxy anions, or the organic sulfur species. Pyrite formation also occurred in experiments where the starting monosulfides were air-exposed prior to aging in sulfide solutions, or when air was bubbled through a reaction vessel containing iron monosulfides suspended in a sulfid sulfide solution. The experiments indicate the rate of conversion from iron monosulfides to pyrite is not only a function of solution chemistry (i.e., pH and aqueous speciation), but also depends on the surface oxidation state of the precursor iron monosulfides. Measurements of δ34S of reactants and products from pyrite-forming experiments suggest that the conversion from iron monosulfides to pyrite may proceed via loss of ferrous iron from, rather than via addition of zero-valent sulfur to, the precursor monosulfides. The sulfur isotopic composition of pyrite in sedimentary environments should reflect the sulfur isotopic composition of the precursor iron monosulfide plus sulfur sources incorporated during surface-controlled growth processes. Pyrite forms produced in this study ranged from poorly developed octahedral grains, in experiments where initial pyritization rates were the slowest, to framboidal aggregates in experiments where initial pyritization rates were the fastest. Although greigite formation occurred in experiments that produced framboids, not all experiments that produced greigite led to framboid formation. The formation of pyrite with framboidal texture is apparently favored when iron monosulfides rapidly convert to pyrite.

Pyrite framboids are densely packed, generally spherical aggregates of submicron-sized pyrite crystals. In this study, a survey was made of framboid size distributions in recently deposited sediments from euxinic (Black Sea; Framvaren Fjord, Norway; Pettaquamscutt River Estuary, Rhode Island, USA), dysoxic (Peru Margin), and oxic (Wallops Island, Virginia, USA; Great Salt Marsh, Delaware, USA) environments. Pyrite framboids in sediments of modern euxinic basins are on average smaller and less variable in size than those of sediments underlying dysoxic or oxic water columns. Down-core trends indicate framboid size distribution is a sediment property fixed very early during anoxic diagenesis, generally within the top few centimeters of burial. Size distributions in modern sediments are comparable with those in ancient sedimentary rocks, evidence that framboid size is preserved through advanced stages of diagenesis and lithification. It is proposed that where secondary pyrite growth is limited, as to preserve primary pyrite textures, framboid size distribution may be used to indicate whether fine-grained sedimentary rocks were deposited under oxic or anoxic conditions. The Crystal Size Distribution Theory relates framboid size to growth time and rate. On the basis of this theory, the characteristic smaller sizes of framboids in sediments of modern euxinic basins reflect shorter average growth times relative to oxic or dysoxic environments. In euxinic environments, framboid nucleation and growth occurs within anoxic water columns, and growth times are, on average, shorter because of hydrodynamic effects than when framboid nucleation and growth occurs within anoxic sediment porewaters underlying oxic water columns. A maximum framboid growth time of 0.4 years is indicated for framboids forming in the water columns of euxinic basins.

Ligands responsible for zinc and lead mobility in solutions forming Mississippi Valley-type (MVT) deposits have not yet been identified and this deficiency precludes formulation of detailed geochemical models of the genesis of MVT deposits. A suitable ligand must be identified that is capable of complexing a minimum of 10 ppm of zinc and of lead. Nevertheless, many constraints on the geochemical environment during formation of the Upper Mississippi Valley (UMV) district in Illinois, Wisconsin, and Iowa, have been identified. Ore characteristics, such as sphalerite color-banding, require that deposition occur in a reducing environment at near-equilibrium conditions. Conditions for the ore-transporting fluid are T = 125 ± 25°C, pH = neutral ± 1, log a O 2 = −55 to −43, and total dissolved reduced sulfur about 0.001 to 0.01 m . Models for MVT ore formation fall into three classes, labeled herein as the sulfur addition, sulfate reduction, and reduced sulfur models. Comparison of these three models for MVT ore formation with geochemical constraints on the ore-forming fluids of the Upper Mississippi Valley district leads to the conclusion that the reduced sulfur model applies best to this district. Experiments were carried out to test the chemical compatibility of a variety of organic ligands with the geochemical constraints of the reduced sulfur model. Solubility tests were conducted in sealed quartz tubes with a series of short-chain mono- and dicarboxylic acids at 125°C, pH 6.1 (near neutral), total dissolved reduced sulfur = 0.002 m , and ligand concentrations of 0.1 or 0.05 m . Formic, acetic, propionic, n-butyric, iso-butyric, n-valeric, oxalic, malonic, succinic, methyl-succinic, glutaric, and adipic acids were tested, in addition to six thiocarboxylic acids (thiolactic, 3-mercaptopropionic, mercaptosuccinic, 3,3′-thiodipropionic, thioglycolic, and thiodiglycolic acids). Results show that the monocarboxylic acids and the thiocarboxylic acids are weak ligands for zinc and cannot play a role in MVT deposit genesis. Several dicarboxylic acids are moderately strong complexing ligands. Oxalic acid produced solutions with more than 13.6 to 22.0 ppm Zn dissolved under experimental conditions. Average values for formation constants for 1:1 Zn(II) organic complexes ( I = 0 m ) at 125°C were determined for malonic acid (log β 1 = 8.30); succinic acid (log β 1 = 9.05); glutaric acid (log β 1 = 10.70); methyl-succinic acid (log β 1 = 10.85); and adipic acid (log β 1 = 10.67). For zinic oxalate, the average values of log β 1 ( I = 0 m ) = 11.95 ± 0.40 at 125°C. Oxalate, although a strong ligand for Zn(II), is precluded from contributing to MVT formation because it readily precipitates as calcium oxalate in the presence of calcium ion. In general, none of the short-chain mono- or dicarboxylic acids nor the thiols tested form zinc complexes strong enough to contribute to MVT deposit genesis. Other longer-chain and thiocarboxylic acids as well as humic and fulvic acids remain to be investigated as transporting agents in forming these deposits.

Experiments to measure the rates of dissolution and precipitation for cristobalite (xtb), S i O 2 ( x t b ) + 2 H 2 O ( 1 ) ⇌ k − k + S i ( O H ) 4 ( a q ) were carried out from 150–300°C. These experiments consisted of monitoring the isothermal approach to chemical equilibrium of aqueous solutions and synthetic cristobalite in a closed-system. The dissolution and precipitation rate constants for cristobalite, in water, are given by ln k + = − 0.9 − Δ E a + / ( R ⋅ T ( K ) ) ln k - = − 0.16 − Δ E a - / ( R ⋅ T ( K ) ) The activation energy of dissolution, ΔEa+, is 68.9 ± 11 kJ·mol−1. The activation energy of precipitation, ΔEa−(52.9 ± 10 kJ·mol−1), is in agreement, within the given uncertainty, with published activation energies for other silica polymorphs. The same activation energy of precipitation for different silica polymorphs is predicted by Transition State Theory (TST). Using experiments that were allowed to most closely approach equilibrium, steady-state concentrations were approximated at 150°C from initially supersaturated solutions and at 200°C from both supersaturated and undersaturated solutions. From the resulting steady-state conditions, equilibrium constants were derived for the above reaction, pK150 = 2.22 and pK200 = 2.10. These values are in close agreement with published data. The results show that cristobalite may precipitate from hydrothermal solutions if the concentration of Si(OH)4 exceeds that at cristobalite saturation, and is less than that of amorphous silica saturation and if there are cristobalite nuclei present.

The formation of pyrite and marcasite from solutions between 100 and 300°C has been examined experimentally, and the solubility product for the initial precipitate upon mixing of Fe2+ and H2S solutions has been determined. Below 300°C, pyrite and marcasite form via an FeS precursor. The precursor phase is crystalline, nearly stoichiometric FeS with a solubility product of 102.9±0.2 ( K = a Fe 2 +a H 2S (a H + ) 2 ). It reacts progressively to mackinawite, hexagonal pyrrhotite, and/orgreigite before forming pyrite or marcasite. In the presence of elemental sulfur, thiosulfate, or polysulfides, the rate of reaction is extremely fast (minutes at 150°C). Without these sulfur species and only hydrogen sulfide or bisulfide present, the conversion rate drops significantly ( 1 h at 300°C). In the absence of any aqueous sulfur source, the conversion proceeds at an indetectably slow rate. Because the rate of direct FeS2 nucleation is insignificant in (slightly) acidic solutions below 300°C, most pyrite and marcasite in hydrothermal ores form via the conversion of an FeS precursor.

The formation of pyrite and marcasite from a FeS precursor has been examined experimentally. In aging experiments at 65°C, the conversion of precursor amorphous FeS depends on these geologically relevant variables: concentration of sulfur-contributing species, acidity, redox state, time, Fe(II)/S(−II) ratio in solution, and addition of an organic ligand (citrate). The results indicate that pyrite and marcasite formation proceed at a significant rate only if intermediate sulfur species (i.e., polysulfides, polythionates, or thiosulfate) are present in solution. In the absence of any sulfur contributor or with only hydrogen sulfide or bisulfide present, no FeS2 formed within 16 days. Sulfidation of the precursor proceeds through progressively more sulfur-rich, Fe-S phases: am FeS (Fe1.11S-Fe1.09S) → mackinawite (FeS0.93-FeS0.96) → greigite (Fe3S4) → pyrite/marcasite FeS2. Greigite is absent under very reduced environments. The conversion sequence found in this study is in good agreement with iron-sulfide distribution patterns found in modern marine sediments.

Reaction paths for nucleation and growth of pyrite and marcasite from solution have been investigated experimentally. Conditions were chosen to avoid the precipitation of metastable Fe-S phases which can act as precursors for FeS2 formation. The experiments indicate that FeS2 nucleation is extremely slow below 100°C. Instead of FeS2 nuclei, the reaction of ferrous ions and polysulfide ions produces initially amorphous FeS. Although the nucleation of FeS2 is inhibited below 100°C, pyrite and marcasite can grow from solutions devoid of polysulfides and undersaturated with respect to possible Fe-S precursor phases. The inability of pyrite to rapidly nucleate explains high supersaturation with respect to pyrite and marcasite in anoxic environments. Although pyrite is the stable Fe-S phase in these environments, it will not control the Fe2+ and H2S (or HS−) concentrations until its growth rate exceeds the dissolution rate of far more soluble, metastable FeS precursor phases.

This Final Report includes a summary and discussion of results obtained under this project on the solubilities in subcritical aqueous solutions of Mont St. Hilaire analcime, Wikieup analcime, and Castle Creek Na-clinoptilolite. Also included here are the methods and results of hydrothermal flow-through experiments designed to measure the rates of Na-clinoptilolite dissolution and precipitation at 125{degree}C. In this report, high-temperature solubility measurements made in our lab are integrated and discussed along with the low-temperature measurements made at Yale University. The final report prepared by the group at Yale University (Lasaga et al.) includes a synthesis of dissolution rate measurements made between 25{degree} and 125{degree}C on the Na-clinoptilolite.

The principal effort has been concentrated on the preparation of clean clinoptilolite, quartz, and boehmite and then reaction of the natural clinoptilolite solid solution to the Naendmember, plus measurements of the endmember solubility to derive an accurate equilibrium constant for the clinoptilolite dissolution reaction, correctly speciated. We are very pleased with the consistency between the best calorimetrically measured and modeled equilibrium constants and those we have determined from 125{degrees}C to 265{degrees}C. These results now provide a basis for relating measurements of reaction rates to departures from equilibrium.

Pyrite and marcasite were precipitated by both slow addition of aqueous Fe2+ and SiO32− to an H2S solution and by mixing aqueous Fe2+ and Na2S4 solutions at 75°C. H2S2 or HS2− and H2S4 or HS4− were formed in the S2O32− and Na2S4 experiments, respectively. Marcasite formed at pH < pK1 of the polysulfide species present (for H2S2, pK1 = 5.0; for H2S4, pK1 = 3.8 at 25°C). Marcasite forms when the neutral sulfane is the dominant polysulfide, whereas pyrite forms when mono-or divalent polysulfides are dominant. In natural solutions where H2S2 and HS2 are likely to be the dominant polysulfides, marcasite will form only below pH 5 at all temperatures. The pH-dependent precipitation of pyrite and marcasite may be caused by electrostatic interactions between polysulfide species and pyrite or marcasite growth surfaces: the protonated ends of H2S2 and HS2 are repelled from pyrite growth sites but not from marcasite growth sites. The negative ions HS2 and S22− are strongly attracted to the positive pyrite growth sites. Masking of 1πg* electrons in the S2 group by the protons makes HS2 and H2S2 isoelectronic with AsS2− and As22−, respectively (Tossellet al., 1981). Thus, the loellingitederivative structure (marcasite) results when both ends of the polysulfide are protonated. Marcasite occurs abundantly only for conditions below pH 5 and where H2S2 was formed near the site of deposition by either partial oxidation of aqueous H2S by O2 or by the reaction of higher oxidation state sulfur species that are reactive with H2S at the conditions of formation e.g., S2O32− but not SO42−. The temperature of formation of natural marcasite may be as high as 240°C (Hannington and Scott, 1985), but preservation on a multimillion-year scale seems to require post-depositional temperatures of below about 160°C (Rising, 1973; McKibben and Elders, 1985).

The solubility of gold was measured in aqueous sulfide solutions at pH from 3 to 8, 150° to 350°C, and at pressures determined by the liquid-vapor pressure of the solution, with oxidation state fixed or buffered by either sulfate-sulfide equilbria or H2(g). High solubilities were measured in solutions with near neutral pH with a maximum measured gold concentration of 0.036 m (7224 mg/kg) at 350°C in a solution containing 0.66 m H2S and 0.28 m NaHS. The results are consistent with the aqueous complex Au(HS)2−. Log equilibrium constants for the reaction Au + H 2 S ( aq ) + HS − = Au ( HS ) 2 − + 1 2 H 2 ( g ) at 150°, 200°, 250°, 300°, and 350°C were determined as −2.39 ± 0.2, −1.89 ± 0.2, −1.56 ± 0.3, −1.35 ± 0.3, and −1.22 ± 0.2, respectively. These values are in reasonable agreement with published data at both 25°C and elevated temperatures. The high stability of Au(HS)2− indicates that geologically significant quantities of gold can be transported in typical hydrothermal solutions. Calculated gold solubility for the Ohaaki geothermal system in New Zealand shows that Au(HS)2− can easily account for the measured hydrothermal gold concentration. Gold may be precipitated from solution by both pH and redox changes. In addition, decreasing the activity of sulfide in solution is an effective mechanism for gold deposition. Analysis of the effect of temperature on the solubility of gold shows that a decrease in temperature may increase or decrease solubility. Deposition by cooling depends upon the pH-oxidation state path of the solutions.

Rate laws have been determined for the aqueous oxidation of pyrite by ferric ion, dissolved oxygen and hydrogen peroxide at 30°C in dilute, acidic chloride solutions. Fresh, smooth pyrite grain surfaces were prepared by cleaning prior to experiments. Initial specific surface areas were measured by the multipoint BET technique. Surface textures before and after oxidation were examined by SEM. The initial rate method was used to derive rate laws. The specific initial rates of oxidation (moles pyrite cm−2 min−1) are given by the following rate laws (concentrations in molar units): rsp,Fe3+ = −10−9.74M0.5Fe3+M−0.5H+ (pH 1–2)rsp,o2 = −10−6.77M0.5O2 (pH 2–4)rsp,h2o2 = −10−1.43MH2O2 (pH 2−4) An activation energy of 56.9 ± 7.5 kJ mole−1 was determined for the oxidation of pyrite by dissolved oxygen from 20–40°C. HPLC analyses indicated that only minor amounts of polythionates are detectable as products of oxidation by oxygen below pH 4; the major sulfur product is sulfate. Ferric ion and sulfate are the only detectable products of pyrite oxidation by hydrogen peroxide. Hydrogen peroxide is consumed by catalytic decomposition nearly as fast as it is by pyrite oxidation. SEM photomicrographs of cleaned pyrite surfaces indicate that prior to oxidation, substantial intergranular variations in surface texture exist. Reactive surface area is substantially different than total surface area. Oxidation is centered on reactive sites of high excess surface energy such as grain edges and corners, defects, solid and fluid inclusion pits, cleavages and fractures. These reactive sites are both inherited from mineral growth history and applied by grain preparation techniques. The geometry and variation of reactive sites suggests that the common assumption of a first-order, reproducible dependence of oxidation rates on surface area needs to be tested.

Anion exchange separation of polythionates and thiosulfate by HPLC (high-performance liquid chromatography) followed by differential pulse polarographic determination provides a systematic and sensitive analytical method for mixtures of these aqueous sulfur species. The time necessary for the partial separation of thiosulfate and polythionates is 12 min. Polarographic determination of each ion can be performed within 5-15 min except for the time necessary for sulfitolysis (20 min) with an error of less than +/- 10% for 10/sup -5/ to 10/sup -3/ M of thiosulfate and polythionates.

The solubility of Ag2S (acanthite/argentite) was measured in vapor-saturated aqueous sulfide solutions at 25–300°C, total sulfide = 0.2−1.4 m, and pH25° = 5.8−7.3. Under these conditions, Ag2S was found to dissolve according to the following reaction: 1 2 Ag 2 S(s) + 1 2 H 2 S(aq)+HS −= Ag(HS) − 2 for which the following log K's were obtained: −3.82 ± 0.10 (25°C), −3.26 ± 0.10 (100°C), −2.91 ± 0.10 (150°C), −2.55 ± 0.10 (200°C), −2.32 ± 0.10 (250°C), and −2.11 ± 0.10 (300°C). These data are in good agreement with previous work at 20°C ( Schwarzenbach and Widmer , 1966), and 100–180°C ( Melent'yev et al., 1969), but not with the dinuclear complex stoichiometries recently obtained by Sugaki et al. (1987). Although Seward (1976) has shown that considerable silver can be dissolved as chloride complexes, our data indicate that Ag(HS)2− is the dominant silver species in hydrothermal fluids with near-neutral to alkaline pH, relatively low oxidation state, high total sulfide, and T Brown , 1986). Silver transported as Ag(HS)2− is precipitated in response to a decrease in aqueous sulfide concentration, a change in pH away from the pK1 for H2S, or, in the native silver field, a decrease in oxidation state. Cooling is a less important depositional mechanism, particularly for solutions whose pH is buffered by feldspar alteration reactions.

In unmineralized, carbonate rocks of the Wisconsin district, the Zn content is less than 4.7 ppm and varies with the clay content. Therefore, by subtracting an amount proportional to the clay content from the analysis of unweathered mineralized samples, the remainder is the amount of Zn added hydrothermally. These adjusted analyses reveal that the lateral extent of primary hydrothermal dispersion is a function of the width of the adjacent sulphide vein and may extend to at least 53 m from a major orebody. The adjusted concentration and its gradient are indicators of the distance to, and size of, veins in this district. The utility of this relationship for exploration depends upon optimizing several factors to maximize the probability of detecting any orebody in the area searched at minimum cost. To detect orebodies with a minimum width of 1 m, the optimum sampling interval was found to be 9 m with a threshold concentration of 5 ppm Zn (adjusted value). This combination achieves a 95% probability of detecting any orebody over 1 m thick within the area sampled. There is a 27% probability that a consecutive pair of samples might be anomalous due to sampling within the narrow dispersion of two random fractures or minor veins. A second cycle of samples taken midway between each pair of anomalous samples eliminate such false anomalies. Where samples can be collected across the trend of orebodies, either through potentially transecting drill core or along apparently barren drifts, the cost of testing an apparently barren 300-m section is roughly equivalent to that of 12 m of diamond drilling. This procedure complements a drilling programme by greatly increasing the probability of detecting orebodies at minimum cost for those areas where the initial, background metal concentration can be reliably established either from its constancy in the host rock or by a normalizing procedure such as that used here.

A differential rate equation for silica-water reactions from 0–300°C has been derived based on stoichiometry and activities of the reactants in the reaction SiO2(s) + 2H2O(l) = H4SiO4(aq) ( ∂a H 4 SiO 4 ∂t ) P.T.M. = ( A M )(γ H 4 SiO 4 )(k+a SiO 2 a 2 H 2 O − k_a H 4 SiO 4 ) where ( A M ) = (the relative interfacial area between the solid and aqueous phases/the relative mass of water in the system), and k+ and k− are the rate constants for, respectively, dissolution and precipitation. The rate constant for precipitation of all silica phases is log k − = − 0.707 − 2598 T (T, K) and Eact for this reaction is 49.8 kJ mol−1. Corresponding equilibrium constants for this reaction with quartz, cristobalite, or amorphous silica were expressed as log K = a + bT + c T . Using K = k + k − , k was expressed as log k + = a + bT + c T and a corresponding activation energy calculated: a b c Eact(kJ mol -1) Quarts 1.174 -2.028 x 103 -4158 67.4–76.6 α-Cristobalite -0.739 0 -3586 68.7 β-Cristobalite -0.936 0 -3392 65.0 Amorphous silica -0.369 -7.890 x 10-4 3438 60.9–64.9 Upon cooling a silica-saturated solution below the equilibrium temperature, the decreasing solubility of silica causes increasing super saturation, which tends to raise the precipitation rate, but the rate constants rapidly decrease, which tends to lower the precipitation rate. These competing effects cause a maximum rate of precipitation 25–50°C below the saturation temperature. At temperatures below that of the maximum rate, silica is often quenched into solution by very slow reaction rates. Consequently, the quartz geothermometer will give the most accurate results if samples are taken from the hottest, highest flow rate, thermal springs which occur above highly fractured areas.

Solubilities are useful only if measured in defined systems where at least the minimum number of intensive variables required by the Gibbs' phase rule for invariancy are either fixed, measured, or controlled in the experiment. The array of determined intensive variables and any necessary buffers should be selected also to provide maximum resolution in identifying stoichiometries of solute species and in evaluating equilibrium constants of solubility-controlling reactions. Simultaneous measurements of volumetric properties of solutions and gases present during solubility measurements are practical and simplify thermodynamic analysis of solubilities. Equilibration is best demonstrated by equal solubilities for the same conditions approached from opposite directions. In solubility experiments at high pressures and temperatures, continuous analyses are possible using spectra, cell potentials, or radioactive tracers but they are uncertain indicators of total solute concentration. Methods providing only a single data point per experiment, by nutrient weight-loss or by analysis of quenched fluids, are comparatively slow and are not reversible. Better are periodic sampling and analysis of quenched, filtered fluids from two types of experimental systems. Condensed fluids are investigated effectively with an external-fluid-supported flexible cell but vapor-containing or supercritical experiments are less complex in a fixed-volume, rocking vessel. In both systems the parent solution remains virtually isothermal and isobaric during sampling, replicate sampling is possible, and reversing of reactions is simple. The less-complicated, fixed-volume system also permits P-V-T measurements on liquids and gases together with sampling. The stoichiometry of the dominant solute species is given, commonly with satisfactory resolution, by the exponential dependence of the solubility on the concentration of each potential ligand. The exponent closely approximates the stoichiometric coefficient of the ligand in the solute species. From these ligand and solute concentrations and tabulated or calculated activity coefficients, activities are obtained for dominant solute species, ligands, and gases corresponding to each solubility measurement. These activities permit the calculation of an equilibrium constant for each principal reaction representing equilibration among the solute, its dissolved species, and reacting ligands. The resulting constants, based on individual solubility measurements may then be compared for consistency both isothermally and polythermally. Measurements of solubilities at high temperatures and pressures are now impeded primarily by the lack of (1) well-calibrated buffers of oxidation state and acidity with known reaction rates for hydrothermal use, and (2) inert high-strength alloys that are resistant to hydrogen embrittlement, nonreactive with acidic and other high temperature fluids, and with mechanical properties compatible with use in the bodies or liners of reaction vessels.

Poor resolution ΔG ƒ 0 forS2− affects hydrothermal complexing arguments, evaluation of sulfide solubility products, and calculation of aqueous sulfur speciation. The second dissociation constant for H2S (logKHS−) if well known could be used to determine this free-energy. The logKHS− has been evaluated here by extrapolating the thermodynamic data for the dissociation of polysulfides (H2Sn) as a function of the reciprocal chain length ( 1 n ). The extrapolation method is supported by standard weak acid theory. The extrapolation is further based on flocculation experiments which determined the equilibrium constant for the protonation of a colloidal sulfur surface, which is a reaction analog for the protonation of an infinite sulfur chain, H2S∞. The derived log KHS− value, −18.51 ± 0.56 (20°C), implies that S2− is never a dominant aqueous species. This new value has been used to derive alternative solubility products of some metal sulfides.

Manganese at equilibrium in seawater occurs dominantly as Mn2+ and inorganic complexes at a concentration ratio of about 1:0.72; solubility decreases exponentially with increasing pH or Eh. However, the nodule oxides birnessite and todorokite are at least four orders of magnitude undersaturated relative to the Mn concentrations of seawater, and are metastable relative to hausmannite and manganite. This apparent lack of equilibrium is explicable by the mechanism of precipitation. Surfaces assist Mn precipitation by catalyzing equilibration between dissolved and reactive O2 and simultaneously also by adsorbing ionic Mn species. The effective Eh at the surface becomes 200–400 mV above that of seawater; the oxidation rate of Mn increases about 108 ×, and the activation energies for Mn oxidation decrease ~ 11.5 kcal/mole. Consequently, marine Mn nodules and crusts form by adsorption and catalytic oxidation of Mn2+ and ferrous ions at nucleating surfaces such as sea-floor silicates, oxyhydroxides, carbonates, phosphates and biogenic debris. The resulting ferromanganese surfaces autocatalyze further growth. In addition, Mn-fixing bacteria may also significantly accelerate accretion rates on these surfaces. Mn which accumulates in submarine sediments may be diagenetically recycled in response to steep solubility gradients causing upward migration from more acidic and reducing horizons toward the sea floor. In contrast, the concentrations of the predominant ferric complexes, Fe(OH)30 and Fe(OH)4−, are relatively less sensitive to the Eh's and pH's found in this environment; Fe is therefore not as readily recycled within buried sediments. Consequently, Fe is not so effectively enriched on the sea floor, although it precipitates more readily than Mn because seawater is saturated in amorphous Fe(OH)3. The metastable, perhaps kinetically-related, Mn oxides of nodules have a characteristic distribution: birnessite predominates in oxidizing environments of low sedimentation rate and todorokite where sedimentation rates and diagenetic Mn mobility are higher. Surface adsorption and cation substitution within the disordered birnessite-todorokite structure account for the high trace element content of Mn nodules.

Single crystals of cinnabar up to 2 mm in diameter and free of foreign, contaminating, solvent anions were grown hydrothermally from aqueous sodium bisulfide solutions in a rocking autoclave at 26° to 200°C. Of the conditions tested, optimum results were obtained when both the solubility and solubility gradient, controlled by the composition of the bisulfide solution and temperature, were highest. Stoichiometry of the cinnabar crystals is controlled by fixing temperature and fS2 during growth, fS2 being dependent on pH and bisulfide concentration.

Evidence from natural occurrences and from published syntheses indicates that the sphalerite-wurtzite inversion is not an invariant reaction at 1020°C, 1 atm but is a univariant function of fs2 and temperature. Single crystals of zinc sulfide, grown in aqueous NaOH solutions which were selected to control fs2, have shown that a univariant boundary exists between sphalerite and wurtzite near 500 atm from 465 ± 4°C to 517 ± 2°C over a corresponding calculated fs2 range of 10−9.5 to 10−8.7 atm. Direct determination of fs2 for coexisting sphalerite and wurtzite, made by passing H2 + H2S mixtures over ZnS powder, gave fs2 of 10−5 atm at 890°C, between 10−5.5 and 10−6.4 atm at 800°C, and between 10−6.5 and 10−8.5 atm at 700°C. The fs2 -dependence of this phase change demonstrates that wurtzite is sulfur-deficient relative to sphalerite. Nonstoichiometry in zinc sulfide is also indicated by its color and by published luminescence studies, electrical measurements, and chemical analyses. Electrical measurements show the defects to be zinc vacancies in sphalerite and sulfur vacancies in wurtzite. The combined range of nonstoichiometry in ‘ZnS’ is on the order of 0.9 at per cent. The hypothesis that wurtzite might be formed metastably at low temperatures due to oxygen substitution for sulfur is untenable. No oxygen was detected by cell-edge measurements of sphalerite or wurtzite from the hydrothermal experiments. In the gas-mixing experiments in which wurtzite was produced from sphalerite below 900°C, oxygen was not present. Wurtzite is thermodynamically stable at lower fs2 than sphalerite. Above about 250°C, this stability field lies well outside of the normal sulfidation state encountered in ore-forming environments; however, at lower temperatures wurtzite may be deposited in highly reducing environments. Because it is stable throughout the geologically most important pH range within 2 or 3 units of neutrality, highly acid solutions are not necessary for its precipitation as advocated by Allen et al. (1914).