V31A-1402 0800h
Hydrothermal Diamond Anvil Cell Investigations Into the Alumina-Silica-Water System up to 1073 K and 4 GPa
Understanding fluid chemistry in the subduction zone environment is key to unraveling the details of element transport from the slab to the surface. Solubility of different mineral assemblages in predominantly water-rich fluid along with pressure and temperature conditions control the chemical structure of the aqueous fluid and govern the transport opportunities for various chemical components away from the subducting slab. In-situ Raman experiments were performed in the alumina-silica-water system in an externally heated Bassett-type hydrothermal diamond anvil cell in the Department of Geological Sciences at the University of Michigan. Natural quartz samples (from the Owl Creek Mountains, Wyoming) were used as the silica source and synthetic ruby was used for the alumina source. Temperatures inside the diamond cell were monitored using type-K thermocouples wrapped around the diamonds and the pressure calibrated by the Raman shift of diamond or quartz or the fluorescence of ruby depending on conditions. Raman measurements of the aluminosilicate fluid show the presence of multiple alumina, silica, and mixed species. As predicted by calculations an aluminosilicate specie possibly of the form (HO)$_{3}$SiOAl(OH)$_{3}$$^{2-}$ as well as the silica monomer and dimer specie were observed in the aluminosilicate fluid. There also appeared to be at least one hydrous alumina specie based on the presence of a Raman peaks at 228 cm$^{-1}$, 339 cm$^{-1}$ and 970 cm$^{-1}$ in the fluid and a comparative analysis between Raman peaks in aqueous fluid in the silica-water, alumina-water, and alumina-silica-water systems. Solid phases formed during experiments (diaspore, kyanite) were confirmed with Raman spectroscopy.
V31A-1403 0800h
Calcite Dissolution Kinetics and Solubility in NaCl-CaCl$_{2}$-MgCl$_{2}$ Brines up to 1 bar pCO$_{2}$ and 80\deg C
Sedimentary basins can contain close to 20% by volume pore fluids that are commonly classified as brines. These fluids can become undersaturated with respect to calcite as a result of migration, dispersive mixing or by anthropogenic injection of CO$_{2}$. Both the solubility of calcite and its dissolution kinetics are currently poorly constrained in these concentrated solutions. This study measured calcite solubility and dissolution rates in geologically relevant NaCl-CaCl$_{2}$-MgCl$_{2}$ synthetic brines (50,000 to 200,000 mg L$^{-1}$ TDS). The EQPITZER calculated calcium carbonate ion activity product at steady-state was in reasonable agreement (\pm10%) with K$_{cal}$ in low concentration brines but systematically exceeded K$_{cal}$ with increasing brine concentration. The deviation was most strongly correlated with calcium activity and was independent of magnesium concentration. This has been interpreted as an uncertainty in the carbonate ion activity coefficient perhaps related to errors in the calcium interaction parameters with the carbonic acid system under these conditions. The dissolution rate dependency on brine concentration, pCO$_{2}$ (0.1 to 1 bar) and temperature (25\deg C to 80\deg C) was modeled using the empirical rate equation R=k(1-\Omega) $^{n}$. Relative to an apparent kinetic solubility it was found that n could be assumed to be first order over the range of degree of disequilibrium investigated (\Omega = 0.2 to 1.0). Rates increased with increasing pCO$_{2}$ as did the sensitivity to brine concentration. At 0.1 bar rates were independent of concentration (k = 9.0\pm1.0\times10$^{-3}$ moles m$^{-2}$ hr$^{-1}$). However, at higher partial pressures rates were linearly correlated to TDS. At 1 bar pCO$_{2}$ and 25\deg C the rate constant can be described by the linear regression k(moles m$^{-2}$ hr$^{-1}$) = 0.053 - 1.7\times10$^{-7}$(TDS), R$^{2}$ = 0.996. The specific effects of Ca$^{2+}$, Mg$^{2+}$ and ionic strength were tested in addition to the inhibitory effect of the presence of 1000 mg L$^{-1}$ SO$_{4}$$^{2-}$. A roughly three fold increase in rate was measured at 80\deg C relative to 25\deg C. The relatively high activation energy (E$_{a}$ = 20 kJ mol$^{-1}$) along with a stirring rate independence suggest the dissolution is dominated by surface controlled processes. These findings may offer important implications to reaction-transport models in carbonate bearing saline reservoirs.
V31A-1404 0800h
Thermodynamic Properties of Sulfatian Apatite: Constraints on the Behavior of Sulfur in Calc-Alkaline Magmas
The Gibbs free energy of hydroxyellestadite [Ca10(SiO4)3(SO4)3(OH)2] was estimated using mineral equilibria applied to analyzed assemblages from the experimental charges of Luhr (1990). The apatite analyses of Peng et al. (1997) were used in conjunction with new analyses of the oxides and silicates in this study. An ideal mixing model was employed for apatite combined with mixing models from MELTS (Ghiorso & Sack, 1994) and Gibbs free energy data from Robie & Hemingway (1995) for the other crystalline phases. The resultant equation of the Gibbs free energy vs. T for hydroxyellestadite is as follows: DG°T(elem) = [2.817(T - 273) - 11831]/1000 kJ/mol, T in K. The calculated entropy for hydroxyellestadite is 1944 J/mol.K at 1073 K and 2151 J/mol.K at 1227 K. Independent estimates of the entropy of hydroxyellestadite obtained with the method of Robinson & Haas (1983) are within 5% of these values. The thermodynamic data on hydroxyellestadite were used to calculate the locus of the reactions: 2Ca10(SiO4)3(SO4)3(OH)2 + 7S2 + 21O2 = 20CaSO4 + 6SiO2 + 2H2O 6Ca10(SiO4)3(SO4)3(OH)2 + 102SiO2 + 20Fe3O4 = 60CaFeSi2O6 + 6H2O + 9S2 + 37O2 2Ca10(SiO4)3(SO4)3(OH)2 + 10Mg2Si2O6 + 14SiO2 = 20CaMgSi2O6 + 2H2O + 3S2 + 9O2 in fO2-fS2 space at fixed P-T. Application of these equilibria to apatite zoned in sulfate from oxidized granitoids reflects a drop in fS2 by more than 1 log unit during its growth. The zoning is interpreted to represent the removal of a magmatic vapor phase during crystallization of these plutons. Removal of sulfur from magmas by hydrothermal fluids is important to the ore-forming process and to the production of acid sulfate aerosols during eruption of oxidized magmas. Preservation of sulfatian apatite may yield data on the sulfidation states of ancient flood basalts such as the Deccan Traps of India and the Parana basalts of Brazil to address the environmental impact of these giant eruptions.
V31A-1405 0800h
Solubility of apatite in H$_{2}$O-NaCl and silicate-bearing solutions at 0.7-3.0 GPa, $800\deg$C
Apatite is a major reservoir for the rare-earth elements (REE) in the earth's crust. However, little is known about its solubility in metamorphic fluids. We measured the solubility of apatite in H$_{2}$O-NaCl and silicate bearing fluids at $800\deg$C and 1.0-2.0 GPa using a piston-cylinder apparatus with NaCl-graphite furnaces. A single Durango fluor-apatite crystal was loaded into a 1.6 mm OD Pt inner capsule, which was crimped and then placed in a 3.5 OD Pt outer capsule with ultra pure H$_{2}$O and NaCl or powdered albite. Solubility was determined by the weight loss of the apatite grain after 24 hrs. In the H$_{2}$O-NaCl experiments, total dissolved solids (TDS) were initially below detection (0.4 millimolal) between X$_{NaCl}$= 0 and X$_{NaCl}$= 0.025. At X$_{NaCl}$= 0.035, solubility was 3.3(0.2) millimolal (errors are 1s), and it increased to 57.5(0.4) millimolal at X$_{NaCl}$= 0.526. Our results show that there is an enhancement in apatite solubility with increasing pressure in pure H$_{2}$O. Solubility is initially below detection at $<$1.0 GPa, but increases to 0.8(0.2) and 1.2(0.2) millimolal at 1.5 GPa and 2.0 GPa respectively. At 1.0 GPa, apatite solubility in H$_{2}$O-albite (Ab) fluids increases from 0.5(0.2) millimolal at X$_{albite}$= 0.004 to 1.4(0.2) at X$_{albite}$= 0.007. Monazite occurs as $<$1 $\mu$m subhedral grains on apatite surfaces in all runs, but its weight is insignificant and does not effect the solubility measurements. Our results demonstrate that, at a given P and T, apatite solubility rises with increasing concentrations of NaCl and albite, suggesting that the additional components promote mass transport of dissolved apatite components during metamorphism. Concentrated brines may explain the origin of apatite veins associated with in high-grade granulite terranes and ore bodies. The formation of monazite demonstrates that the REE are less soluble in H$_{2}$O-NaCl and silicate bearing solutions. This probably explains textures in which monazite mantles apatite, which are common in granulite metamorphic terranes, such as the Kiirunavaara magnetite-apatite ore.
V31A-1406 0800h
Solubility of corundum and quartz in the system Al$_{2}$O$_{3}$-SiO$_{2}$-NaCl-H$_{2}$O at deep-crustal metamorphic conditions: 800 $\deg$C and 10 kbar
The solubility of corundum in H$_{2}$O-NaCl-SiO$_{2}$ fluids has been measured at 800 $\deg$C and 10 kbar. Solubility was determined by weight loss after equilibration of polished single crystals of synthetic corundum with fluids in welded Pt envelopes. Fluid compositions were X$_{NaCl}$=0 to 0.5 and SiO$_{2}$ up to quartz saturation at a given X$_{NaCl}$. The experiments were 1 to 7 days duration in 1.91-cm-diameter piston-cylinder apparatus with NaCl pressure medium. Kyanite and sillimanite are slightly more stable than quartz + corundum at these P-T conditions, but did not appear. The solubility in initially pure H$_{2}$O is very small: m$_{Al2O3}$ (mol Al$_{2}$O$_{3}$/kg H$_{2}$O) = 0.00135, but increases rapidly as the fourth root of X$_{NaCl}$ to 0.01 at X$_{NaCl}$=0.1 and 0.015 at halite saturation (X$_{NaCl}$\sim$0.6). Quenched solutions were pH neutral. The data were modeled with simple mixing models involving NaAl(OH)$_{4}$, Al(OH)$_{3}$, and any of their several dehydration and dissociation products, as well as Na$^{+}$, Cl$^{-}$, and H$_{2}$O. The best-fit reactions are: X$_{NaCl}$$<$0.3: Al$_{2}$O$_{3}$(cr) + Na$^{+}$ + H$_{2}$O = NaAlO(OH)$_{2}$ + AlO$^{+}$ X$_{NaCl}$$>$0.3: Al$_{2}$O$_{3}$(cr) + Na$^{+}$ = NaAlO$_{2}$ + AlO$^{+}$ Silica in solution further amplifies NaCl-enhancement of corundum solubility. Al$_{2}$O$_{3}$ molality at metastable quartz saturation rises rapidly from 0.006 at X$_{NaCl}$=0, to 0.038 at X$_{NaCl}$=0.1, and then more gradually to 0.045 at X$_{NaCl}$=0.3. Quenched solutions from SiO$_{2}$- and NaCl-rich compositions are acid (pH$\sim$1). The dissolution reaction, assuming a neutral NaAlSi species, must therefore be similar to: Al$_{2}$O$_{3}$(cr) + 2Na$^{+}$ + {\it x}Si(OH)$_{4}$ = 2NaAl Si$_{x/2}$O$_{x/2+y}$(OH)$_{x-2y+4)}$ + ({\it x}+2{\it y}-5)H$_{2}$O + 2H$^{+}$. Simple-mixing modeling suggests that the average Si/Al ratio ({\it x}/2) and number of non-bridging O$^{2-}$ ({\it y}) of the dominant Al solute species at 0.03$\leq$XNaCl$\leq$0.6 and quartz saturation are given by the formula: NaAlSi$_{2.75}$O$_{6.38}$(OH)$_{2.24}$, close to the NaAlSi$_{3}$O$_{8}$-H$_{2}$O join. Small amounts of spherical globules of aluminosilicate glass appeared close to quartz saturation at X$_{NaCl}$$\geq$0.1 It was not proven in our experiments whether a small field of stable silicate liquid intervenes near quartz saturation, or whether the melt phase was metastable. The great enhancement of Al$_{2}$O$_{3}$ solubility in the presence of NaCl and solute silica indicates that, under some conditions of high-grade metamorphism, Al$_{2}$O$_{3}$ is a substantially mobile component. NaCl solutions infiltrating peraluminous rocks may become quite acidic.
V31A-1407 0800h
Experimental constraints on H$_{2}$O activity in high-pressure metamorphic brines
Subduction of crustal materials is accompanied by metamorphic reactions liberating fluids. Fluid inclusions in eclogite minerals range from dilute solutions to chloride-rich brines; however, the effect of salinity variations on the stability of hydrous phases in subduction zones is poorly understood. To address this problem, we carried out reversed piston-cylinder experiments on the equilibrium (1) paragonite = jadeite + kyanite + H$_{2}$O at 700$\deg$C, 1.5-2.5 GPa, in the presence of H$_{2}$O-NaCl fluids. The experiments were conducted using fluids with nominal starting compositions: X(H$_{2}$O)=1.0, 0.90, 0.75 and 0.62. At X(H$_{2}$O) = 1.0, the equilibrium lies between 2.25 and 2.30 GPa. Lowering X(H$_{2}$O) decreases the pressure of paragonite breakdown to 2.10 - 2.20 GPa at X(H$_{2}$O) = 0.90 and 1.85-1.90 GPa at X(H$_{2}$O) = 0.75. The experiments at X(H$_{2}$O) = 0.62 yielded albite + corundum at $\leq$1.60 GPa, and jadeite + kyanite at $\geq$1.70 GPa. The shift in the equilibrium pressure constrains a-X relations in the sytem H$_{2}$O-NaCl and indicates that a(H$_{2}$O) varies as nearly the square of its mole fraction. The results are consistent with the extrapolated non-ideal activity model of Aranovich and Newton (1996, CMP, 125, 200). Our results permit use of appropriate paragonite-bearing or -absent assemblages to quantify a(H$_{2}$O) in high-P metamorphic environments, such as the Austroalpine units in the Western Alps. For example, jadeite and kyanite in a metapelite from Val Savenca in the Sesia Lanzo Zone formed during the Eo-Alpine high-P metamorphic event at 1.7-2.0 GPa, 550-650$\deg$C. The absence of paragonite requires a fluid with low a(H$_{2}$O) of 0.3-0.6, which could be due to the presence of saline brines. Petrologic investigations of Sesia Lanzo eclogites from Val Ianca show that paragonite occurs as inclusions in garnet cores but gives way to omphacite + kyanite toward rims, suggesting a decrease in a(H$_{2}$O) from $\sim$1.0 to $<$0.81 (Tropper and Essene, 2002, SMPM, 82, 487-514). During the subsequent hydration in the late eclogite-facies at slightly lower P, a(H$_{2}$O) increased to 0.84-0.98, stabilizing the assemblage paragonite + omphacite. The new results can also be used to compare a(H$_{2}$O) from phase-equilibrium studies with fluid-inclusion constraints. Paragonite-bearing eclogites from the Mt. Emilius unit in the Western Alps interacted with fluid with a(H$_{2}$O)=0.62-0.72, based on fluid-inclusion data from Scambelluri et al. (1998, Lithos, 43, 151) during Eo-Alpine high-P metamorphism of 1.1-1.3 GPa, 450-550$\deg$C (Dal Piaz et al., 1983, Am. J. Sci., 283A, 438). Calculation of reaction (1) with reduced a(H$_{2}$O) shows no incompatibilities with the observed phase assemblage. This study shows that paragonite + cpx + kyanite-bearing rocks have great potential for constraining a(H$_{2}$O).
V31A-1408 0800h
Chemical and Mineralogical Patterns Along a Late-Archean Crustal Profile in Southern India: Signatures of a High Temperature Brine?
Several features are present in a regional dehydration zone in Tamil Nadu, Southern India that have been attributed to the action of high temperature brines in smaller scale more localized dehydration zoneselsewhere (Harlov and Forster, 2002. J Pet. 43: 769-824). Many of these same features can also be explained by a model in which dehydration is the result of partial melting during high-grade metamorphism. An 8-km thick section of late Archean, mid to lower crust was studied along a 100 km traverse from Krishnagiri to Salem. Intermediate to felsic orthognesises make up the bulk of the terrane. In the northern amphibolite-facies zone these contain hornblende and biotite. First clinopyroxene and then orthopyroxene appear in the central transitional zone and become more abundant southwards as biotite and hornblende abundances decrease. Orthopyroxene is the dominant ferromagnesian silicate in the southern zone and garnet is widespread. Among the features of this traverse that can be explained by the migration of a high temperature brine upwards through the crust are: 1) the presence of K-feldspar grain boundary veins that increase in abundance with increasing metamorphic grade , 2) monazite grains within and on the rim of apatites in orthopyroxene-bearing gneisses, 3) increases in F concentrations in apatite and biotite and increases in the calculated HF-fugacity/H2O-fugacity ratios with increasing metamorphic grade, 4) depletion in whole rock Th and U, 5) depletion in both whole rock Rb and Rb in biotite southwards with increasing grade, 6) relatively high oxygen fugacities (2.5 log units above QFM) in orthopyroxene-bearing rocks, 7) increases in the abundance of hematite-rich hemo-ilmenites southwards with increasing grade, and 8) redistribution of REE from titanite in amphibolite-facies gneisses to apatite in granulite-facies gneisses. Monazite rims on apatite and K-feldspar grain boundary veins are also consistent with a model of metamorphism involving dehydration melting of intermediate igneous rocks. However, the presence of monazite inclusions in apatite and the detailed pattern of trace-element depletion are not as easily explained by this model. The formation of these features by brines acting on a regional scale would require a source of large amounts of these fluids in the lower crust or upper mantle.
V31A-1409 0800h
Fluid Induced Dehydration of the Mafic Lower Crust from Amphibolite to Granulite Facies: Nature and Experiment
Natural evidence for the role of low H$_{2}$O activity fluids (CO$_{2}$ or concentrated brines) in the dehydration of H$_{2}$O-rich, mafic amphibolite facies rocks to H$_{2}$O-poor, Opx-bearing granulite facies rocks (700-900 C and 500-1000 MPa) for both highly localised dehydration zones (CO$_{2}$; cm's) as well as regional terranes (brines; km's) (Harlov and Förster 2002, J Petrol, 43, 769) include the presence of Kfs micro-veins along Qtz-Plg grain boundaries; Plg grains metasomatised in a K-rich fluid; Mnz and/or Xn inclusions in the FAp grains; Bt enriched in Ti, F, and Cl; and FAp enriched in Cl and F. These features are not seen in the "source" amphibolite facies terrane along the same traverse. When log(fHF/fH$_{2}$O) for either Bt or FAp is plotted as a function of the distance from the fluid/heat source, a uniform decrease in log(fHF/fH$_{2}$O) is observed across the granulite to amphibolite facies traverse suggesting the presence of a uniform low H$_{2}$O activity uniform fluid front. Dehydration experiments (900 C; 1000 MPa; 3 weeks; Au capsule; quenched) involving a cylinder of natural tonalitic Bt gneiss (Plg, Qtz, Bt) (220 mg) and a concentrated KCl brine (20-30 % H$_{2}$O; 70-80 % KCl) (8 mg) placed at the base of the cylinder have been conducted in the piston cylinder apparatus (CaF$_{2}$ setup). Micro-veins primarily of Kfs, with some evidence of partial melting, formed along Qtz/Plg grain boundaries though only where Bt and Qtz were in contact. Here the Bt reacted with Qtz to form numerous small Opx and Cpx grains as well as minor Ilm from the 2-3 wt % of TiO$_{2}$ present in the Bt. The two principle reactions responsible for both the formation of the Kfs micro-veins as well as the pyroxenes include: (1) An (in Plg) + Qtz + KCl (in fluid) = Kfs + CaCl$_{2}$ (in fluid) and (2) Bt + Qtz = Opx + Kfs + H$_{2}$O. The same experiment performed under the same P-T conditions involving either a concentrated NaCl brine (20-30 % H$_{2}$O; 70-80 % NaCl) or a CO$_{2}$-rich fluid (80 % CO$_{2}$, 20 % H$_{2}$O) or a fluid absent dry melt resulted in micro-veins approximating a granitic composition along Qtz/Plg grain boundaries with numerous small Opx grains + minor Ilm forming along biotite grain boundaries, again only when the Bt and Qtz were in contact. Due to an absence of KCl in these three cases, only reaction (2) was relevant.
V31A-1410 0800h
Evidence for KCl Brine Metasomatism in Mafic Xenoliths from the Lower Crust
The occurrence of Kfs micro-veins in high-grade felsic gneisses has been interpreted as evidence for migrating alkali-rich fluids in the deep crust which may be responsible for melt-absent dehydration of high-grade rocks from amphibolite to granulite facies. Here we present new data on the occurrence of K-feldspar in mafic granulites found as xenoliths in recent alkaline lavas from Western Sardinia, Italy. The xenoliths originated from the underplating of subduction-related basaltic liquids which underwent cooling and recrystallisation in the deep crust (T = 850-900 C, P = 800-1000 MPa). They consist of quartz-free metagabbronorites characterised by Opx + Cpx + Plg porphyroclasts (An$_{50-66}$) in a granoblastic recrystallized matrix composed of Pyx + Plg (An$_{56-72}$) + Fe-Ti oxides +/- Kfs +/- Bt +/- FAp +/- Ti-Prg. Texturally, the Kfs occurs in a variety of different modes. These include: (1) rods, blebs, and irregular patches in a random scattering of Plg grains, (2) micro-veins along Plg-Plg and Plg-Pyx grain rims, (3) myrmekite-like intergrowths with Ca-rich Plg along Plg-Plg grain boundaries, and (4) discrete anhedral grains (sometimes microperthitic). The composition of each type of Kfs is characterized by relatively high Ab contents (16-33 mol.%). Increasing An content in the Plg towards the contact with the Kfs micro-vein and myrmekite-like intergrowths into the Kfs along the Plg-Kfs grain boundary are also observed. Small amounts of Bt (TiO$_{2}$ = 4.7-6.5 wt.%; F = 0.24-1.19 wt.%; Cl = 0.04-0.20 wt.%) in textural equilibrium with the granulite facies assemblage is present in both Kfs-bearing and Kfs-free xenoliths. These Kfs textures, e.g. the randomly scattered and relatively rare antiperthitic plagioclase grains, suggest a likely metasomatic origin due to solid-state infiltration of KCl-rich fluids/melts. The presence of such fluids is supported by the FAp in these xenoliths which is enriched in Cl (Cl = 6-50 % of the total F+Cl+OH). These lines of evidence suggest that formation of Kfs in the mafic xenoliths reflects metasomatic processes, requiring an external K-rich fluid source, which operated in the lower crust during in situ high-T recrystallization of relatively dry rocks.
V31A-1411 0800h
Processes and Timing estimate of Metasomatism in peridotitic xenoliths from Orapa kimberlites, Botswana
The evolution of cratonic lithosphere has been of debate for a long time. High SiO2 contents in whole rock, presence of hydrous minerals and enrichment in magmaphile elements suggest the important role of metasomatism in the history. Peridotitic xenoliths from Orapa kimberlites, Botswana, mainly classified into three main groups (Coarse Low-T type, Mid-T type and Deformed High-T type). Some peridotitic xenoliths from these groups include garnets which show well preserved chemical heterogeneities resulting prior to kimberlite entrapment and its intrusion. These Garnets show several zonation patterns regarding major elements like Cr, Ca, Ti, Fe and Mg. Most of such zonations are along the lherzolite trend defined by CaO and Cr203 contents especially in Deformed High-T type xenoliths. In addition, Cr2O3 and TiO2 line profiles show well-developed U-shape or S-shape zonation patterns which are taken as prototype of resorption or regrowth of garnet. These zonation patterns indicate resorption or growth by metasomatic fluid infiltrated through cratonic lithosphere beneath Orapa. Because ideal resorption patterns depend on diffusion coefficiencys and the timing of infiltration of fluid, we estimate the timing of the infiltration of metasomatic fluids by comparison with the analyzed chemical patterns and ideal simple simulations. REE patterns of coexisting garnet and clinopyroxene suggest that the rims of these minerals are well equilibrated but cores are not always so. This fact indicates the process of garnet-clinopyroxene equilibration by metasomatic fluid (silicate, carbonate or kimberlitic melt). Using these new data, we constructed one of possible rough story about metasomatism in cratonic lithosphere beneath Orapa
V31A-1412 0800h
Heat Capacity of Hollandite-Type KAlSi3O8 and Related Phase Equilibria
The behavior of potassium in the deep mantle is of great significance from geophysical and geochemical points of view. Several experimental studies of phase transitions of stoichiometric KAlSi3O8 have been undertaken (Ringwood et al., 1967; Liu, 1978; Kinomura et al., 1975; Yagi et al., 1994; Urakawa et al., 1994; Fasshauer et al., 1998; Akaogi et al., 2004). Hollandite-type KAlSi3O8 is suggested to be a possible reservoir of potassium in the Earth's lower mantle. Tutti et al. (2001) showed in experiments that hollandite-type KAlSi3O8 was a stable phase under the pressure of at least 2200 km deep in the lower mantle and suggested that it is important host for potassium in that region. However, because of the lack of low-temperature heat capacity data of hollandite-type KAlSi3O8, the S-a(298) is not well determined, hence the phase equilibria related to hollandite-type KAlSi3O8 are still not well constrained in terms of slopes of the reactions. In this study, we first synthesized hollandite-type KAlSi3O8 from sanidine glass as starting material at ca. 10 GPa and 1273K using a multi-anvil device at the Univ. of Minnesota. The hollandite structure was confirmed by XRD analysis. The heat capacity was measured in the range of 5-303 K using a PPMS at Salzburg. The low-temperature heat capacity data merge smoothly with data measured on hollandite-type KAlSi3O8 between 160 and 700 K by Akaogi et al. (2004). The entropy of hollandite-type KAlSi3O8 is calculated as 147.6 J/mol.K at 298.15 K. In combination with the existing thermodynamic and experimental data, the phase boundary between hollandite-type KAlSi3O8 and wadeite, kyanite, and coesite is more accurately fixed.
V31A-1413 0800h
Experimental Study of Trace Element Partitioning Between Immiscible Magmatic Brine and Granitic Melt
Highly concentrated magmatic brines can exsolve in silicate magmas of variable composition. The exsolutions are documented in fluid and melt inclusions trapped in magmatic minerals during growth, and the occurrences of the brines are most common in silicic magmas of granitic-rhyolitic composition. Here we present the results of experimental study of element partitioning between immiscible brine (chloride melt) and aluminosilicate melt of granitic composition. Starting mixture was prepared from equal amounts of synthetic haplogranitic glass and mixed reagent-grade chlorides of alkalis and alkaline earths. The mixture was doped by trace elements, at about 300 ppm each. Experiments were performed in an internally heated pressure vessel mounted on a centrifuge, which allowed an effective in situ centrifuge separation of immiscible melts during the run. Quenched run products were split along the meniscus, and the chloride and the silicate layer analysed separately by bulk solution-based ICP MS. Two runs were carried out: the first one at 900 C, 100 MPa, no water added, and the second at 800 C, 100 MPa, with about 10 percent of water. The two-liquid Nernst partition coefficients (D) calculated as a ratio of weight concentrations in the chloride melt and concentrations in the granitic melt were the following (first run - second run): Rb (2.3 - 2.7); Cs (2.8 - 3.3); Sr (6.1 - 10.2); Ba (6.0 - 9.7); Y (0.75 - 0.71); La (6.4 - 10.3); Sm (3.1 - 3.4); Eu (6.1 - 8.5); Gd (2.0 - 2.1); Ho (0.97 - 0.93); Lu (0.53 - 0.51); Zr (0.001 - 0.003); Hf (0.0008 - 0.002); Pb (0.17 - 0.79). All the other REE were also analysed, and their D values decrease almost linearly from La to Lu. D value of Eu is anomalous (by a factor of 2.4-3), although oxygen fugacity was not controlled and Eu is likely to be in the oxidised trivalent form. The addition of water increased D values of Zr and Hf by a factor of 2.5 and that of Pb by a factor of 4-5, while the other elements included in the study were not much affected. In general, the solubility of HFSE in the brine appears to be very low. The brine-silicate melt partitioning noticeably decouples the "geochemical twins", such as Y-Ho and Zr-Hf (see the D values above). Thus, the separation of the brine in natural magmas and brine-induced metasomatism of country rocks may produce a distinct trace element signature in volcanic and plutonic rocks affected by the processes.
V31A-1414 0800h
On the Effects of NaCl on Convective Fluid-Flow in Magmatic-Hydrothermal Systems
Previous studies of convective fluid-flow and heat transport in magmatic-hydrothermal systems have approximated the fluid as pure H$_2$O, ignoring the major concentrations of dissolved NaCl. The presence of NaCl in water has profound effects on the thermodynamics and hydrodynamics of hydrothermal systems. NaCl-H$_2$O fluids can separate into a high-density, high-salinity liquid (brine) phase and low-density, low-salinity vapor phase at pressures and temperatures well above the critical point of pure water. This process is particularly common in magmatic-hydrothermal systems and likely a key driver for the formation of the world's major ore deposits of Cu, Mo, and Au. During convective transport of heat and NaCl, so-called double-diffusive, double-convective systems form because heat diffuses faster than salt, but is advected at a slower rate. The presence of NaCl can hence stabilize the flow, preventing convection cells to form, or amplify flow-instabilities, leading to chaotic convection behavior. Using numerical simulations, we discuss the effects of NaCl on hydrothermal convection at geologically realistic pressure, temperature, and salinity conditions. The results show that five general flow-patterns, ranging from single-phase diffusive to multiphase convective, can be identified. Classical parameters such as the Rayleigh number or buoyancy ratio fail to predict the evolution of convection patterns, because they do not account for the non-linearity of the fluid properties and are intrinsically not defined at two-phase conditions. Geological implications are that the transport of NaCl, and probably that of many other ore-forming metals, is maximized at single-phase conditions. Because of its high density and low volume fractions, the brine is immobile at two-phase conditions and sub-lithostatic pressures. Many geochemical fractionation processes and ore-mineral precipitation, however, occur near the transition from single to two-phase flow.
V31A-1415 0800h
Cu-Cl-Salt-Hydrate Melts and Their Possible Role in Ore-forming Systems
We found direct evidence for a Cu-Cl-salt-hydrate melt in equilibrium with a complex system of silicate melts in a magnesiohastingsite-dominant cumulate from the TUBAF Seamount near Lihir Island (Papua New Guinea). We observed four distinct exsolution-induced features in the trachyandesitic intercumulus melt: 1) exsolution of a Cu-Fe sulfide melt, 2) exsolution of a Cu-salt-hydrate melt, 3) exsolution of a second foiditic silicate melt, and 4) exsolution of a vapor phase. The Cu-Cl-salt-hydrate melt crystallized into polycrystalline monomineralic aggregates of clinoatacamite. The features, which testify the formation of the clinoatacamite as an exsolution from the silicate melt, are presented. There is no indication of late hydrothermal alteration or weathering in the sample. We compare different possibilities to explain the formation of the xenolith and show, that the xenolith had formed independently from its parent magma in the oceanic crust. Even though no data on the solubility in Cu-Cl-silicate melt systems are available we suggest, with respect to other salt-melt - silicate-melt systems, an upper temperature limit of 800 $\deg$C for the unmixing of the salt-hydrate-melt. The observation of these unmixing processes sheds new light on the modus operandi of how metal-enriched magmatic fluids enter hydrothermal systems. We present different scenarios, how these fluids may affect the formation of ore deposits, like epithermal gold deposits or copper porphyry systems.
V31A-1416 0800h
Constraints on Variable Ag:Au:Cu Ore-Metal Ratios in Felsic Arc-Magmas
Silver:gold:copper ratios are variable in porphyry-type ore systems. In an attempt to better understand why, we have employed experimental techniques to determine how silver and copper, and gold from previous experiments, are sequestered in felsic magmas. To this end, we are performing sealed silica tube experiments on the equilibria among pyrrhotite-magnetite-silver alloy at 800C and at vapor pressure. Run times for the preliminary experiments were 144 hours; runs had magnetite/pyrrhotite ratio of 4. The source of silver in the runs was AgCl. Analysis of reconnaissance experiments demonstrates the stability of magnetite, pyrrhotite and a silver sulfide solid solution under the conditions of the experiments. Equilibrium concentrations of ore metals in the run products are ~3000 ppm Ag and 3500 ppm Cu in the pyrrhotite. However, the concentrations in magnetite are significantly different: 100 ppm Ag and ~20 ppm Cu. Like copper and gold (Jugo et al., 1999; Lithos), silver is concentrated in pyrrhotite relative to magnetite. The equilibrium Ag-sulfide composition in the run products is Ag53Fe8Cu3S36, with a mole fraction of Ag2S of 0.74. The log fS2 is approximated as ~ -4. The mole fraction of Ag in an ideal metal solid solution in equilibrium with an ideal model Ag2S solid solution, and a log fS2 of -4, is ~0.4. By analogy with Au, the substitution of Ag into pyrrhotite may occur as an AgFeS2 component. The substitutional mechanism for Ag in magnetite is not clear: silver may substitute as AgFe(3+)(Fe(2+))-2, but may also be present in defects in the magnetite structure. The partition coefficient (D(po/mt)) for approximately 30 for Ag. The partition coefficient for Au is higher (~120) based on the data of Simon et al. (2003; Am. Min,) and Jugo et al. (1999; Lithos). These data can be combined with data on the solubility of Ag in silicate melts to calculate mineral-melt partition coefficients. These data suggest that the role of pyrrhotite crystallization in felsic magmas is considerably more important than magnetite in sequestering Ag relative to Au. Therefore, the fractionation of magnetite relative to pyrrhotite is important in the generation of the variable silver:gold ratios in silicate melts.
V31A-1417 0800h
Fluid Evolution in the Nepheline Syenites of the Ditrau Alkaline Massif, Romania
The Ditrau Alkaline Massif (Romania) is situated in the central part of the Eastern Carpathians, as an intrusion in the Bukovina nappe system of the Mesozoic crystalline zone. Nepheline syenites are the most abundant rocks occurring in the central and eastern part of the Massif, and representing the youngest intrusion of the complex. The nepheline syenite is composed of perthitic feldspars, nepheline, biotite, amphibole, pyroxene and titanite as primary minerals, and sodalite, cancrinite, calcite, analcite as secondary minerals formed at the expense of nepheline. Petrographic observations and fluid inclusion studies were performed on nepheline syenites in order to examine the effect of residual magmatic fluids on the alteration of nepheline to secondary minerals listed above. The alteration of nepheline to secondary minerals is obvious from textural relationships and comparison of the compositions of the minerals. Fluid inclusion studies provide evidence for the role of highly saline fluids in the incongruent transformation reactions (nepheline to sodalite and/or cancrinite and/or analcite). The fluids, in most cases, can be modeled by the H$_{2}$O-NaCl system with various NaCl contents; however inclusions with more complex fluid (containing also K, Ca, CO$_{3}$, etc. besides H$_{2}$O and NaCl) composition are abundant in the nepheline. The alteration process is supported by the presence of fluid inclusions in cancrinite, showing lower salinity compared to those in nepheline. During the crystallization period of the nepheline syenites the rock was in equilibrium with a salty solution whose salinity increased with time, mostly by the loss of H$_{2}$O to produce H$_{2}$O-bearing minerals like amphiboles and micas. One possible interpretation of the fluid inclusions and textural observations is that nepheline alteration to sodalite, cancrinite and analcite was associated with increasing salinity of the fluids with time.
V31A-1418 0800h
Evidence for Extremely Large Lava Flows on Ontong Java Plateau from High Precision Measurements of Volatiles and Major Elements in Natural Glasses
Magmas of Ontong Java Plateau (OJP) display little geochemical variation, having only a few widely dispersed magma types (Mahoney et al. , 1993). (Here we define magma type as all lavas that have evolved by similar extents of melting of a similar mantle source, and have undergone similar polybaric fractionation histories). In this study, we use high precision microprobe measurements of Cl, K, S, H2O, CO2 and major elements in glasses to show that magmas from widespread locations on OJP are identical in composition and are probably from the same eruption and quite possibly from the same series of lava flows. By same eruption, we mean the quasi-continuous issuance of magma from a continuous chamber over a time period that is insufficient for further differentiation or assimilation. By same lava flow, we mean lavas that have issued from the same or nearby vents and were part of a sequence that that was continuously molten at the surface or beneath a crust. Cl concentrations are controlled by assimilation that takes place fairly late at shallow levels in the magma chamber. The amount of assimilation and Cl content of assimilated material control Cl contents of magmas, and are expected to be highly variable in this stochastic process. It is inconceivable that magmas erupted at different times would have precisely the same Cl content, even if they have the same major element chemistry from identical cotectic evolution. The clearest case of distant lavas being from the same eruption is the Kroenke-type lavas from ODP holes 1187A and 1185B, about 140 km apart. The lavas form roughly 150 flow units of about 1 meter average thickness, which we feel are multiple surges of lava from a quasi-continuous eruption. Glass compositions (major elements and volatiles) do not vary more than analytical uncertainties within each hole. Differences between the two holes are also less than analytical uncertainties. Averages of 4 samples from each of the two holes are: Cl 750 vs 732 ppm; s.d.=15 ppm. S 988 vs 969 ppm; s.d.=5 ppm. K 616 vs 608 ppm; s.d.=22 ppm. Counting precision is ±15 ppm (2 sigma) for each element in each single glass analysis. Precision is better than accuracy. Similar CO2 contents in the glasses of the separate holes (Roberge et al., 2004), despite several hundred meters difference in their current reconstructed basement depth suggests that the lavas were erupted from the same vents and flowed for long distances: that they are not merely separate flows of a large eruptive episode that issued from a single magma chamber that had many vents. This observation does not mean that dissolved CO2 contents cannot be used to estimate paleoeruption depths, but that caution is required. The volume of the eruption exceeds that of any known submarine eruption by a great deal. If its lateral dimensions are similar to the length between the holes, then its area would be >200 km2 and its volume would be >3000 km3. That such widely separated lavas have identical chemistry means that there was virtually no cooling over 10s of km of flow. This requires an insulating layer above a rapidly flowing magma (Gregg and Fornari, 1998). We are currently evaluating whether other widely separated recoveries on OJP are also from single lava flows. Possible matches include 1185B (lower) and 1186A, which are about 150-200 km apart. They would indicate even larger eruptive volumes. It is possible that certain glasses from 807C are the same as a thin layer in 1185B. References: Gregg and Fornari, JGR 103, 27517. Mahoney et al. Proc. ODP: Sci Results 130, 3-22, 1993. Roberge et al. Geol. Soc. Lond. Spec. Pub 229, 239-257, 2004.
V31A-1419 0800h
Limited Boron Isotopic Variation Between Caldera-Forming and Post-Caldera Low-$\delta$$^{18}$O Rhyolites from Yellowstone Caldera
Post-collapse rhyolites from within Yellowstone caldera show prominent excursions to $\delta$$^{18}$O values (VSMOW) as low as $\sim$0 $\permil$ that provide strong evidence for significant non-lithospheric oxygen input into magmas, presumably via infiltration of meteoric waters (1). Very little, however, is known about the behavior of other stable isotope systems, such as boron, in low $\delta$$^{18}$O rhyolites and their potential for unraveling the mechanisms by which brines and magmas interact. Ion microprobe measurements of quartz-hosted melt inclusions from two low-$\delta$$^{18}$O rhyolite flows (South Biscuit Basin SBB and Middle Biscuit Basin MBB) yielded average $\delta$$^{11}$B values (NIST SRM 951) of -1.5 and -0.9 $\permil$, respectively. These values overlap within error with those for glassy obsidian matrix from SBB and MBB. Melt inclusions from two caldera-forming tuff eruptions (Lava Creek Tuff LCT and Huckleberry Ridge Tuff HRT), known to have 'normal' oxygen isotopic compositions, also yielded indistinguishable $\delta$$^{11}$B values of -1.8 $\permil$. Recent petrologic studies (1) suggested that bulk remelting of hydrothermally altered volcanic rocks, specifically HRT, in the down-dropped roof of the magma chamber produced the low-$\delta$$^{18}$O magmas. The lack of strong boron isotopic variations (within $\pm$2$\permil$) between 'normal' and low-$\delta$$^{18}$O rhyolites, however, contrasts with published evidence for strong $^{11}$B-depletion in hydrothermal altered rhyolite encountered in Yellowstone drill-wells ($\delta$$^{11}$B = -9.7 $\permil$; 2). This implies that boron isotopic fractionation due to interaction with hydrothermal fluids was either absent in the source region of the SBB and MBB magmas, or became masked due to subsequent processes. From preliminary mixing calculations it is concluded that assimilation of $^{11}$B- and $^{18}$O-depleted rocks by fresh rhyolite recharge could be a compositionally and thermally viable alterative to bulk remelting. (1) I. N. Bindeman and J. W. Valley (2001) J Petrol 42, 1491-1517; (2) M. R. Palmer and N. C. Sturchio (1990) Geochim Cosmochim Acta 54, 2811-2815
V31A-1420 0800h
Regional-scale Free Thermohaline Convection in Multiply Faulted Sedimentary Basins: theoretical results from computational modelling
Free thermohaline convection resulting from fluid density variations due to spatial and temporal changes in temperature and salinity has tremendous implications for many geological processes. There are many examples that investigate thermally induced free convection with no account taken of the effects of salinity. There is also a vast body of contaminant hydrological studies that examine the importance of salinity-driven free convection in enhancing hydrodynamic mixing of a dense solute with less dense ambient groundwater under isothermal conditions. Only a few studies have simultaneously considered the heat- and salinity-induced buoyancy effects on fluid flow but ignored discrete fractures pervasive in the earth materials. We have recently developed a finite element computer package that fully incorporates the hydrodynamic coupling between discrete fractures and surrounding porous media. The computer software has been employed to investigate the importance of salinity in controlling ore-forming fluid flow in sedimentary basins. This paper introduces for the first time the relevant details of the computational algorithm. A number of unpublished numerical case studies for the McArthur basin in northern Australia are also added to illustrate the applicability of the software.
V31A-1421 0800h
Chemical Characterization of Brines from Selected Oil Fields, Tabasco, Mexico
Thirteen brine samples were recovered from nine oil-producing wells in the Agave (Cretaceous) and Saramako (Cretaceous and Tertiary) oil fields. These samples were analyzed for major and trace elements as well as O and D isotopic compositions. The goal of this study was to compare the possible links between oil-related brines enclosed within Cretaceous and Tertiary productive horizons that were thought to have similar origin oils. The salinity of the Saramako Cretaceous and Tertiary horizons is very constant, around 30000 ppm, one to six times lower than the salinities found in the Agave Cretaceous Field (from 45000 to 170000 ppm). Major ion chemistry suggests that brines are in equilibrium with the host rock. One of the main difference, besides Mg, resides in the S concentrations, were Agave samples present lower concentrations, probably related to the presence of abundant sulfides in the aquifer's rock. Halogen (Br, Cl) systematics indicates a different origin for the Saramako and Agave brines. The Saramako samples halogen composition plot near normal seawater both in the Na/Cl vs Cl/Br (molar ratios) and the Cl vs Br (ppm) plots. The Agave halogen data scatter near and underneath the seawater evaporation line in the Na/Cl vs Cl/Br (molar ratios), suggesting that these fluids could represent seawater evolved past the point of halite precipitation. The Cl vs Br (ppm) plot indicates that these fluids undergone some degree of mixing with low-salinity fluids, probably seawater. The presence of two different groups of data suggests the compartment of the aquifer. The \deltaD and \delta18O data show strong differences between the Saramako and Agave brines. The Saramako brine \delta18O and \deltaD isotopic compositions are +2.1% (VSMOW) and -13.8% respectively. The Agave samples have a \delta18O composition from +4.3% to +6.0% and \deltaD isotopic composition from -20.0% to -12.6%. Differences in \delta18O compositions between Saramako and Agave brines indicate that the latter fluids were in equilibrium with the host dolostones at certain temperature. Saramako brine composition indicates a near-pristine seawater origin. Agave brines chemical composition suggests an origin related with a bittern formed after evaporation of seawater past the point of halite precipitation, subsequently mixed with seawater. Oxygen isotopic compositions reflect both different host rock and water to rock interaction phenomena, whereas hydrogen isotopic composition can be modified by diagenetic reactions.
V31A-1422 0800h
Fluid Origin of the Stratabound Fluorite and Celestite Deposits in Coahuila, Mexico.
The Mesozoic units that outcrop north of Coahuila State in northern Mexico host numerous celestite, celestite-fluorite and fluorite deposits, mainly enclosed in the Cretaceous limestone units. Celestite and celestite-fluorite deposits are lens-shaped bodies, up to 2 m thick and a total length exceeding 500 meters, intercalated within the carbonates of the Cupido Formation (Aptian) or equivalents. Celestite-free fluorite bodies, mainly enclosed in the Aurora Formation (Albian-Cenomanian), appear as sub-concordant lenses with abundant evidences of hydraulic fracturing, usually near low-angle fractures. The celestite brine halogen composition on a Cl/Br vs Na/Br molar ratio plot on the trend defined by the evaporation of seawater, while in a ppm Cl vs Cl/Br plot away from the seawater evaporation line but parallel to it, indicating that the solution was mainly modified by dilution. The fluids involved in the genesis of the La Encantada fluorite deposit probably are evolved seawater that undergone some degree of evaporation. Mixing of fluids is clearly delineated by both microthemometric and halogen data, where salinities calculated are too low for the halogen ratios found, indicating a mixing with a low-salinity end-member. Hydrocarbon-bearing fluid inclusions repeatedly show the presence of solid bitumen trapped along with heavy oils, indicative of thermal degradation. Both, mixing and thermal degradation of hydrocarbon-rich fluids along with hydraulic fracture of the host rock points to an "in situ" organic matter maturation, due to the mixing of a saline, oxidant, sulphate and CaCl2-rich "bittern", which probably transported fluorite, with an organic-matter rich fluid present in the cretaceous carbonates. It is noteworthy that these previous results suggest that both celestite and fluorite-dominated deposits formed from brines from similar origin, showing the same halogen systematics. Strontium enrichment can be explained by leaching of Sr from the carbonate series during diagenesis, probably related with the inversion from aragonite to calcite. Transport of fluor is favored and maximized by (basinal) brines enriched in Ca2+, probably originated after seawater evaporation or evaporite dissolution prior to F leaching. Then, the distribution of celestite-dominant or fluorite-dominant deposits can be controlled by both the sort of rocks locally present in the basement as well as the depth of brine circulation.
V31A-1423 0800h
The Effect of Chlorine on the Rheology of Na$_{2}$O-Fe$_{2}$O$_{3}$-Al$_{2}$O$_{3}$-SiO$_{2}$ Melts
Because of the high fluid mobility of chlorine, the mantle wedge in subduction zone settings may be enriched in chlorine by the fluids released by the subducted altered oceanic crust plus sediments. The presence of chlorine (and other halogens) will affect the solidus temperature and the rheology of melts, thus influencing the magma evolution, eruption and degassing. Very little is known about the solubility mechanism and diffusivity of Cl in silicate melts and even less is known about the effect of Cl on viscosity. The present study addresses the effect of the halogen chlorine on the viscosity of silicate melts. The shear viscosities of Cl-bearing melts in the system Na$_{2}$O-Fe$_{2}$O$_{3}$-Al$_{2}$O$_{3}$-SiO$_{2}$ were determined over the temperature range 550-950C at room pressure in air. Viscosities were determined using the micropenetration technique in the range of 10$^{8.5}$ to 10$^{12.0}$ Pa s. The compositions are based on addition of Fe$_{2}$O$_{3}$ or FeCl$_{3}$ to aluminosilicate glasses with a fixed amount of SiO$_{2}$ (67 mol %). Although there was loss of Cl$^{-}$ during the glass syntheses, no loss occurred during the viscometry experiments. It is to be expected that Cl$^{-}$ takes the structural position of an oxygen, and thus reduces the polymerization of the melt structure, and therefore the viscosity of the melt; as F$^{-}$ does. Our measurements show that, depending upon the melt composition, the addition of Cl$^{-}$ will either increase or decrease the viscosity of the melt. In the present melts at least 20% of the iron exists as network modifying, viscosity reducing Fe$^{2+}$; while the rest exists as network forming Fe$^{3+}$. It is proposed here that the different effects of chlorine on viscosity are due to the preferred Cl$^{-}$-Fe$^{2+}$$_{NBO}$ bonding together with the different structure of peralkaline and peraluminous melts. In peralkaline aluminosilicate melts, the addition of Cl$_{2}$O$_{-1}$ will destroy 2 NBOs and create one BO if Cl$^{-}$ bonds primarily to the Fe$^{2+}$ creating non-bridging oxygens. This would result in an increase in viscosity. In peraluminous melts, the addition of Cl$_{2}$O$_{-1}$ may result in Cl$^{-}$ bonds to the charge-balancing Fe$^{2+}$, creating 2 new tri-clusters (assuming (Al$^{3+}$, Fe$^{3+}$)Si$_{2}$O$_{5}$ tri-clusters exist). The viscosity should then decrease due to the weaker bonds of the charge balancing Fe$^{2+}$-Cl$^{-}$ units. The preference of Cl$^{-}$ to form bonds to the NBO-forming Fe$^{2+}$ is indicated by the small amount of Cl$^{-}$ soluble in the peraluminous melt structure in comparison to that soluble in the peralkaline structure.
V31A-1424 0800h
Cycling of Halogens Through the Cascadia Subduction Zone and the Cascade Volcanic arc: Insights From Thermal Springs in Central Oregon, USA
We use concentrations of halogen elements and isotopic ratios of 36Cl/Cl and 129I/I to examine the cycling of halogen elements through the Cascadia subduction zone and volcanic arc. The samples come from thermal springs in the most volcanically-active segment of the Cascades Range (north-central Oregon) and mineral springs in the forearc (Willamette Valley), and are compared to halogen concentrations in pore fluids from the accretionary prism. Mass-balance, element-ratio, and isotopic considerations suggest that a significant part of the Cl, as well as other halogens discharged from the thermal springs are derived from magma degassing. Given published estimates of the subducted halogen flux, mass balance suggests that only about 3% of the Cl, 1.5% of the Br and 0.5% of the I are discharged via thermal springs along the arc. These fluxes are low compared to those in other volcanic arcs and may indicate that most devolatilization from the slab occurs at shallow depths, perhaps related to the elevated temperatures of the subducted Juan de Fuca plate. Total Cl and I discharge along the 260-km segment of the arc is one order of magnitude lower than fluxes from fumaroles in individual volcanoes along the arcs of Central America, Japan, and New Zealand. If our halogen mass-balance calculations are indicative of water loss from the subducting Juan de Fuca Plate, then this would imply a relatively dry sub-arc mantle wedge, and limited melting. Our estimate is in good agreement with the observation that the degree of volcanic vigor in the Cascade Range is low compared to other volcanic arcs surrounding the Pacific Rim, expressed in the paucity of Holocene volcanoes, low-frequency of known eruptions and large volumes of basaltic andesite compared to the more silicic compositions in other arcs.
V31A-1425 0800h
Fluid Flow in the Main and Fore Arc of Central America: $^{129}$I as Tracer of Subduction Processes
Iodine combines the most biophilic behavior with the largest ionic radius among the halogenes, which, together with the presence of the long-lived cosmogenic isotope, $^{129}$I (T$_{1/2}$ = 15.7 Myr), makes it an excellent tracer for fluid movement and the origin of organic material. We have used this system for the detection of recycled marine sediments in subduction zones by determining the isotopic composition in geothermal and volcanic fluids. $^{129}$I /I ratios in fluids collected from main arc settings in Central America, Japan, the Andes and New Zealand result in ages, which are site specific and are compatible with derivation from recycled marine sediments. We report here first results for fluids collected from mud-volcanoes in the marine section of the Central American fore arc and compare them to earlier results from the main arc (Snyder and Fehn, 2002). Shallow pore water samples were collected during several cruises, organized by the SFB 574 at Kiel University (e.g. Hensen et al., 2004). The water depths of the sampling sites ranged from 1000 m at the shelf to 5400 m close to the axis of the Central American Trench. Iodine is enriched by factors of 1000 or more in these fluids compared to sea water and occurs together with methane in these locations. Initial results for $^{129}$I /I ratios in these fluids are between 400 and 200x10$^{-15}$, giving ages between 30 and 45 Ma, which are not compatible with derivation from the currently subducting marine sediments ($<$25 Ma) in this area. These ages are also considerably older than those found in the main arc ($<$20 Ma), indicating different sources for fluids in main arc and fore arc. Our results suggest that the dominant source of iodine in main arc fluids are currently subducting marine sediments, while iodine in fore arc fluids is derived from older sources, probably located in the continental margin wedge, demonstrating fundamental differences in fluid transport between fore arc and main arc settings. Snyder and Fehn, GCA, 66, 3827 (2002) Hensen et al., Geology, 32, 201 (2004)
V31A-1426 0800h
Halogens and Sulfur Across the Kamchatka Arc
Volatile (S, Cl, F), major and trace elements across the Kamchatka arc were studied by electron microprobe and SIMS analysis in homogenized melt inclusions in olivine from arc basalts collected along the E-W transect of the Kamchatka arc, from Eastern Volcanic Front (EVF) to back arc region. Halogens show the opposite behavior across the arc. Chlorine concentrations in melts from EVF and Central Kamchatka Depression (CKD) are similar (750 to 1100 ppm) and higher than for Sredinniy Ridge back arc (SR, 500 ppm). In contrast, fluorine increase from EVF and CKD melts (300 to 450 ppm) to the SR region (800 ppm), increasing the F/Cl ratio from arc front to back arc by a factor of five. Sulfur shows the general decrease from the arc front to back arc, but the highest concentrations (2500 ppm) occur in the CKD melts. Volatile element - trace element ratios in general correlate with fluid-mobile elements (B, Li) and suggest successive changes in fluid compositions with increasing slab depth. In detail, we identify three distinct fluid compositions. EVF-fluids are highly enriched in B, Cl and chalkophile elements. This fluid carries additionally LILE (U, Th, Ba, Pb), F, S and LREE (La, Ce). This fluid is dominating EVF sources and is reduced in the CKD melts and is not documented in SR region. The second CKD-fluid affects the Kluchevskaya Group of volcanoes and is enriched in S and U, S/K2O, U/Th and depleted in Cl/S ratio. This fluid unusually enriched in 87Sr, 18O and 11B. The third SR-fluid observed for back arc volcanoes is highly enriched in F, Li and Be, carrying also LILE, LREE and probably some Nb, which results in correlations of F/Ce with Li/Yb, Li/Dy, Sr/Y, Ba/Y, and Nb/Yb. This fluid starts to be released below CKD and becomes the dominant fluid in the back arc region. We argue that the dehydration of different water-rich minerals at different depths from the arc front (low P-T conditions) to back arc (high P-T conditions) explains the difference in fluid composition and the decoupling of B (and Cl) and Li (and F) across the Kamchatka arc. Seawater component could be a significant additional source for chlorine in Kamchatka melts.
V31A-1427 0800h
Does Fluid-Induced Eclogitization of Subducted Lower Oceanic Crust Produce the Slab Component of Arcs?
To investigate fluid-induced transformation processes and associated trace element mobilization, co-genetic gabbros and eclogites of Zambia have been used. The rocks represent relics of subducted lower oceanic crust and gradual stages of the prograde gabbro-to-eclogite transformation are preserved by disequilibrium textures of incomplete reactions. No evidence for prograde blueschist- or amphibolite-facies mineral assemblages was found in the eclogites. Instead, fine-grained intergrowths of eclogite-facies minerals replacing plagioclase indicate the direct eclogitization of gabbroic precursors. Eclogitization occurred at 630-690°C and 2.6-2.8GPa and was accompanied by a channelized fluid flow that produced veins of the peak metamorphic assemblage. Evidences for aqueous fluids with variable salinities, in cases up to brine compositions, were found. Although all of the mafic rocks were subducted, only those gabbros that were infiltrated by fluids were eclogitized. Hence, the eclogites and their veins represent relict fluid pathways through subducted oceanic crust, providing direct evidence of channelized fluid flow within a slab. The gabbros and eclogites have MORB-like trace element patterns and initial Nd and Hf isotope compositions. In some eclogites, however, the LREE have been strongly fractionated from the HFSE and HREE, an effect that cannot be of magmatic origin but must have occurred during metamorphism. Eclogitization was limited by fluid availability, and the fluid flow through the rock is the most likely mechanism for LREE fractionation. Model fluid-rock ratios reveal that the fractionated rocks reacted with an amount of fluid up to 80% of their mass to create the most depleted REE patterns. The lower gabbroic part of the oceanic crust is an unlikely source for such a large volume of fluid and thus we hypothesise that the fluid originated in the underlying serpentinised lithospheric mantle. If, after triggering eclogitization, the resulting LREE-rich, HFSE+HREE-poor slab fluid reaches the zone of partial melting in the mantle wedge, it may contribute significantly to the arc signature. We will evaluate whether the trace element mobilization during fluid-induced eclogitization could be generally responsible for producing the slab component in arc magmas.