DI33A-1117
Calibration of the infrared molar absorption coefficients by elastic recoil detection analysis (ERDA) for H measurements in olivine and clinopyroxene crystals and rhyolitic glasses
Fourier transform infrared (FTIR) spectroscopy is the most widely applied technique for measuring hydrogen in nominally anhydrous minerals (NAMs) and silicate glasses. FTIR is rapid, sensitive, widely available and gives information on the bonding environment of H-bearing species. H determination relies on the Beer-Lambert law and therefore requires constraints on the applicable molar absorption coefficient, ε. Values of ε may be derived only from independent absolute methods. These ε are now reasonably well known for glasses, but to date determinations of ε applicable to NAMs are extremely limited and subject to uncertainties. Most notably, the Paterson (1982) calibration gives H contents in olivine that are a factor of 2.5- 3.5 lower than those suggested by the Bell et al. JGR 2003 calibration. We performed elastic recoil detection analysis (ERDA) on a range of samples that had been previously analyzed by FTIR, including natural rhyolitic glasses (1430-1772 ppm H2O), natural and synthetic olivine (0-910 ppm), natural orthopyroxene (38-147 ppm), and natural clinopyroxene crystals (0-490 ppm). ERDA is a nuclear microprobe technique that yields matrix-independent absolute determinations of H concentration. A 3 MeV beam of 4He is employed at high spatial resolution (50 × 200 microns). The detection limit, determined from analysis of dry minerals is 150±20 ppm H2O, too great for analysis of many NAMs from the upper mantle, but applicable to H-rich natural and synthetic NAMs. For glasses, synthetic olivines, and clinopyroxenes, we found good proportionality between the measured ERDA hydrogen concentration and the linear (rhyolite) or integrated (minerals) absorbance measured by FTIR. The ε found for rhyolite (103±9 l/mol per cm) is close to that of 88±2 l/mol per cm given by Dobson et al. (GCA, 1989). For clinopyroxene, we obtain ε 47010±6070 l/mol per cm2, slightly larger than 38300±1700 l/mol per cm2 found by Bell et al. (Am. Min. 1995). Finally, for olivine we obtain ε 34330±4000 l/mol per cm2 (to be compared to the 28450±1830 l/mol per cm2 of Bell et al. JGR 2003). Thus, our results confirm the previous conclusion of Bell et al. (2003) that the Paterson (1982) calibration yields strong underestimates of H in olivine.
DI33A-1118
Water and the Oxidation State of Global Arc and MORB Magmas
Subduction zone magmas may derive from sources that have been subject to hydrous, oxidizing fluids, either within the mantle wedge or via alteration of oceanic lithosphere. Despite its fundamental importance, oxidation state is a difficult parameter to constrain from bulk erupted lavas since the initial magmatic oxidation state may change during late-stage degassing and eruption. To investigate the relationship between water content and oxidation state, we present here μ-XANES analyses of Fe speciation (Fe3+/ΣFe) in natural basaltic glasses from submarine pillows and olivine-hosted melt inclusions (MI) from a variety of global tectonic settings. Melt inclusions, trapped before large-scale degassing and crystallization take place, bypass the pitfalls of conventional bulk-rock techniques (e.g., `auto-oxidation' due to H outgassing) by providing direct samples of undegassed, H2O-rich glasses that are ideal for microbeam analysis. Samples capture a full range of natural, undegassed, magmatic H2O contents from mid-ocean ridges (0.2-0.5 wt%), back-arc basins (~1 wt%), and global arc volcanoes (0.3-6.4 wt%). In the arc MI, these new data show Fe3+/ΣFe ratios consistently higher (>QFM) than in the global MORBs (Cottrell et al., this meeting), and a general increase in Fe3+/ΣFe with increasing dissolved H2O content. Neither the effects of post-entrapment crystallization, nor H diffusion through the host olivine, can fully explain this global trend. The data do suggest that post-entrapment Fe diffusion might affect the Fe content of olivine-hosted MI. Some MI indicate ferrous Fe contents too low to be in equilibrium with their olivine hosts, suggesting up to 20% loss of ferrous Fe from some MI. Accounting for these losses, however, does not significantly change the observation that Fe3+/ΣFe correlates with magmatic H2O. This global correlation suggests that melt oxidation state is directly related to magmatic H2O content, and that the high H2O fluxes through subduction zone mantle wedges may thus play a central role in oxidizing arc magmas.
DI33A-1119
Micro-XANES Determination Fe Speciation in Natural Basalts at Mantle-Relevant fO2
We demonstrate that the oxidation state of iron (Fe3+/ΣFe) can be determined with a precision of ±0.02 (10% relative) on natural basalt glasses at mantle-relevant fO2 using Fe K-edge X-ray absorption near edge structure (XANES) spectroscopy. This is equivalent to ±0.25 log unit resolution relative to the QFM buffer. Precise determination of the oxidation state over this narrow range (Fe3+/ΣFe=0.06-0.30) and at low fO2 (down to QFM-2) relies on appropriate standards, high spectral resolution, and highly reproducible methods for extracting the pre-edge centroid position. We equilibrated natural tholeiite powder in a CO/CO2 gas mixing furnace at 1350°C from QFM-3 to QFM+2 to create six glasses of known Fe3+/ΣFe, independently determined by Mössbauer spectroscopy. XANES spectra were collected at station X26A at NSLS, Brookhaven Natl. Lab, in fluorescence mode (9 element Ge array detector) using both Si(111) and Si(311) monochromators. Generally, the energy position of the 1s→3d (pre-edge) transition centroid is the most sensitive monitor of Fe oxidation state using XANES. For the mixture of Fe oxidation states in these glasses and the resulting coordination geometries, the pre-edge spectra are best defined by two multiple 3d crystal field transitions. The Si(311) monochromator, with higher energy resolution, substantially improved spectral resolution for the 1s→3d transition. Dwell times of 5s at 0.1eV intervals across the pre-edge region yielded spectra with the 1s→3d transition peaks clearly resolved. The pre-edge centroid position is highly sensitive to the background subtraction and peak fitting procedures. Differences in fitting models result in small but significant differences in the calculated peak area of each pre-edge multiplet, and the relative contribution of each peak to the calculated centroid. We assessed several schemes and obtained robust centroid positions by simultaneously fitting the background with a damped harmonic oscillator (DHO) function and pre-edge features with two Gaussians over a sub-sample of the pre-edge region (7110-7120 eV). We found that the relation between Fe3+/ΣFe and the centroid energy is non-linear over this fO2 range, which is expected if the coordination environment changes with oxidation state. ΔQFM is linearly related (R2=0.99) to the centroid position. This new calibration allows the oxidation states of natural mantle melts to be discriminated with high spatial resolution (9μm). We apply the new calibration to determination of Fe3+/ΣFe in natural basaltic glasses and olivine-hosted glass inclusions (Cottrell et al. & Kelley et al., this meeting).
DI33A-1120
Volatile (H, C, Cl, S) Concentrations in Ocean Island Basalt Glasses From Pitcairn and the Society Islands
Submarine lavas from Pitcairn and the Society Islands offer a unique opportunity to investigate volatile contents in melts derived from mantle OIB sources, and in particular to constrain the composition and origin of the EM-1 and EM-2 mantle endmembers. We present here the results of a study on volatile abundances in dredged glasses from Pitcairn and the Society Islands, two well-documented OIB localities with EM-1 and EM-2 isotopic compositions. Samples for this study consist of submarine glasses, dredged from 485-2829 m depth and representing a range of melt compositions (42.7-59.7 wt% SiO2, 1.08-11.69 wt% MgO). The EM-2 lavas of the Society Islands have H2O concentrations between 0.9-2.6 wt%, whereas Pitcairn's EM-1 lavas typically contain 0.8-1.8 wt% H2O. Both locations have CO2 abundances that range from low (<190 ppm for Society, <115 ppm for Pitcairn) to below detection, suggesting significant degassing of CO2 ± H2O. H2O/Ce ranges from 55-170 for Pitcairn lavas, and the highest H2O/Ce ratios are observed in samples with the highest CO2 concentrations (least degassed). Samples with H2O/Ce lower than typical MORB or EM sources (<100) may have volatile contents modified by magmatic degassing. Sulfur concentrations (200-1000 ppm for Society, 400-1450 ppm for Pitcairn) may be largely controlled by sulfide saturation. Concentrations of Cl in Society (650-1400 ppm) and Pitcairn (400-1100 ppm) basalts correlate inversely with 87Sr/86Sr, and do not show evidence of Cl assimilation. Cl/K ratios vary from 0.03 to 0.07 for both sample suites, ranging from only slightly higher than enriched MORB (0.01-0.02) to within the HIMU field (0.04-0.08). Pitcairn lavas with the lowest Cl/K have the strongest EM-1 signature (lowest 206Pb/204Pb). Similarly, Society lavas with the lowest Cl/K have the strongest EM-2 signature (highest 87Sr/86Sr). We suggest that these represent primary mantle signatures.
DI33A-1121
Carbon Solubility in Core Melts in Shallow Magma Ocean Environment and its bearing on Distribution of Carbon between Deep Earth Reservoirs
Carbon affects the melting phase relations of mantle rocks [1] and core metal [2], influences the physical properties of molten silicates and metals, and also has significant effect on partitioning of other key elements between various deep Earth phases. But the carbon budget of Earth's deep mantle and core is poorly constrained due to lack of knowledge of behavior of carbon during core formation. In order to determine the storage capacity of dissolved carbon in metallic core melts and to put constraints on partitioning of carbon between silicate mantle and metallic core, we have determined the solubility of carbon in molten core metal at P- T conditions relevant for a shallow magma ocean.Experiments are performed at 2 GPa and to 2500 °C using a piston cylinder apparatus. Pure Fe-rod or a mixture of Fe-5.2%Ni loaded into graphite capsules were used as starting materials. Al coated run products are analyzed by EMP. Carbon concentration of 5.8 ± 0.4 wt.% at 2000 °C, 6.5 ± 0.9 wt.% at 2250 °C, and 7.5 ± 1.2 wt.% at 2500 °C are measured in quenched iron melt saturated with graphite. The trend of C solubility versus temperature for Fe-5.2 wt.% Ni melt, within analytical uncertainties, is similar to that of pure Fe.We have compared our solubility data and an estimate of the current carbon content of the mantle with the carbon content of core melts and residual mantle silicates respectively, derived from equilibrium batch or fractional segregation of core liquids, to constrain the partition coefficient of carbon between silicate and metallic melts in a magma ocean, DC. Translation of the limits of DC, derived from our solubility data, on calculation of carbon content of the residual silicate shows that the observed mantle concentration of carbon is too low to be matched by the process of shallow magma ocean fractionation of carbon between metal and silicate in a chondritic protoearth. If carbon solubility in liquid Fe does not change strongly as a function of pressure, this may indicate the presence of a hidden carbon- rich mantle reservoir untapped by oceanic volcanism. For the entire range of possible bulk Earth carbon content from chondritic to subchondritic and for the mantle carbon content of 50-1000 ppm, DC of 10-4- 100 are derived. But for 1000 ppm bulk Earth carbon, DC is 10-2-100. Using the complete range of possible DC for a magma ocean at ~2200 °C, we predict maximum carbon content of the Earth's core to be 6-7 wt.% and a preferred value of 0.25 ± 0.15 wt.% carbon for a bulk Earth carbon concentration of 1000 ppm. Based on our estimate, the core is likely one of the most enriched terrestrial reservoirs of carbon with concentration as high as 0.4 wt.%, which likely is at least an order of magnitude higher than that of the average mantle. The higher carbon content of OIBs compared to MORBs thus may derive in part from core contributions to mantle plumes.[1] Dasgupta, R. and Hirschmann, M.M. 2006, Nature 440, 659- 662. [2] Wood, B.J. 1993, Earth Planet Sci Lett 117, 593-607.
DI33A-1122
Limestone Assimilation by Basaltic Magmas: an Experimental re-Assessment and Application to Italian Volcanoes
Based on field observations, Daly proposed hundred years ago that limestone assimilation could generate silica undersaturated magma. Discarded by successive studies this process is here re-assessed by an experimental survey and is proposed as a differentiation mechanisms operating in the plumbing system of Italian Plio- Quaternary volcanoes. The widespread presence of sedimentary limestone in their basement, the abundant high temperature skarns in their eruptive products and their important CO2 emissions make carbonate assimilation in the plumbing system of Italian volcanoes a central question, which, however, has been poorly addressed by specific studies. Experimental results of Ca-Mg carbonate assimilation by hydrated mafic magmas in the range 1050-1150°C, 0.1– 500 MPa are here reported. Two types of experiments have been performed in internally heated pressure vessels to simulate magma-carbonate interactions. In the first type we characterized equilibrium phase diagram of basaltic system as a function of variable amount of added carbonate (up to 20 wt percent of the total charge). In all experiments, carbonates completely breakdown and no immiscible carbonate melts are observed. MgO and CaO are essentially incorporated in clinopyroxene and olivine, while CO2 is partitioned between the fluid phase and the silicate glass, with a strong preference for the fluid. The major effect of carbonate incorporation on liquidus phase equilibria is to favor the massive crystallization of Ca-rich clinopyroxene (accompanied by leucite for some starting magma compositions) and the consumption of the other phases crystallizing in carbonate-free conditions (olivine, plagioclase, Fe-Ti oxides). Crystallization of pyroxene from carbonate consumes magmatic SiO2 leading to silica undersaturated residual liquid. Such desilication trend is recognized in several magmatic series emitted in Italy strongly suggesting that assimilation of carbonate is an important regional process. The second type of experiments focused on the non-equilibrium interactions between carbonate and hydrated basalts. Similarly to diffusion couple experiments, two cylinders, one of hydrated basaltic glass and one of carbonate (either calcite or dolomite) were juxtaposed at high temperature during few hours. Magma desilication is also observed together with clinopyroxene growth but the formation of high temperature skarn in the carbonate end-member complicate further the mineralogical assemblage. The comparison between equilibrium and non equilibrium interactions is proposed and allows a better interpretation of the differentiation process occurring in the plumbing system of Italian volcanoes.
DI33A-1123
Density of Carbonated Magmas and Stability of Carbonatite and Kimberlite at the Earth's Upper Mantle and Transition Zone
Seismological and electrical conductivity studies show that the presence of LVZ atop of the 410-km discontinuity which reveal the possible existence of a melt at this boundary [e.g., Reveanugh and Sipkin, 1994; Toffelmier and Tybruczy, 2007]. The anomalies from these studies support conceptual model [Bercovici and Karato, 2003]. Density measurements of anhydrous basaltic melts indicate that it is denser than the surrounding mantle near 410-km depth [Ohtani and Maeda, 2001]. Hydrous peridotitic and basaltic melts are denser than peridotite at the top of the 410 km discontinuity and therefore can be accumulated at the base of the upper mantle [Sakamaki et al., 2006]. CO2 is one of the important volatile in the mantle and it could be also important to constraints the conceptual models experimentally for the explanation of LVZ near a 410 km depth. In the present study, we have measured the density of carbonated basaltic melt at pressures from 16-20 GPa and 2573 K by using sink-float experiment using a diamond marker. We determined the partial molar volume of CO2 in magmas at around 20 GPa. Using the partial molar volume estimated by several authors in the lower pressure range, the compression behavior of the partial volume in magmas can be expressed by the Vinet equation of state with K= 16GPa and dK/dP= 5.2. Using the pressure dependency of the partial molar volume of CO2 in magmas, we can estimate the density of various carbonated magmas at high pressure. Our results show that the basaltic melt can contain up to ~3.5 wt% CO2 and the peridotite melt can contain up to ~4.0 wt% CO2 to be denser than the surrounding mantle at the top of the 410 km discontinuity. These amounts of CO2 are comparable with the amount of H2O in the hydrous basaltic (~3.0 wt%) and peridotitic (~6.7 wt%) melts, which is stable atop of the 410 km discontinuity [Sakamaki et al., 2006]. However carbonated melt can be formed only at significant degree of melting of mantle materials (e.g., peridotite or eclogite), whereas at low degree of partial melting (1-5 %) carbonatite melt, which is thought to be significantly less denser than peridotite, is formed [Dasgupta and Hirschmann, 2006, 2007]. The melt formed by higher degrees of melting of carbonated peridotite is kimberitic containing ~17-32 wt% of CO2 in the pressure range of the bottom of the upper mantle from 10 GPa to 20 GPa [Ghosh et al., 2005]. The present results indicate that the kimberlitic melt formed by partial melting of the carbonated mantle is less dense than the surrounding mantle, and it can ascend even from the depths of the base of the upper mantle and transition zone. This is consistent with the existence of mantle xenoliths containing diamond with majorite or perovskite inclusions in some kimberites. If we combine our data with hydrous basaltic melt [Sakamaki et al., 2006] and consider the linear mixing between H2O and CO2 then the basaltic melt with 1.5 wt% H2O and ~1.3 wt% CO2 and peridotitic melt with 3.3 wt% H2O and 2.0 wt% CO2could be stable at the top of 410 km discontinuity.
DI33A-1124
Ordering kinetics in double carbonates and implications for processes at subduction zones
The identity of carbonate mineral species that are present on the solidus of carbonate bearing lithologies is the critical information that determines the solidus phase relations and element partitioning. Experimental determinations of pressure and temperature conditions of dolomite breakdown reaction, dolomite = magnesite + aragonite (or calcite), carried out over the last five years show disparities that have been attributed to the disorder state in dolomite starting materials and subsequent equilibration during the experiments. We have conducted high-pressure, in-situ synchrotron experiments at SPring8 (BL04B1) using multi-anvil apparatus to determine ordering state of dolomite during this reaction. Four dolomite breakdown experiments were conducted at pressures, from 3 to 7 GPa and temperatures, from 600 to 1200°C. We have crossed the dolomite breakdown reaction at three points between 4 and 5.5 GPa and have determined that dolomite is stable at 7.4 GPa and 1090°C. Our results agree with quench studies of Luth (2001 CMP141:222) and Buob et al. (2006 AmMim91:435) at 5 GPa and above. We disagree with those of Sato and Katsura (2001 EPSL184:529) and Morlidge et al. (2006 CMP152:365). Our data point at 4 GPa does not fall on any extrapolated curve of any previous investigation. Ordering state of dolomite in our experiments are consistent with those of Luth (2001) and results presented by Antao et al. (2004 AmMin89:1142) show a systematic offset, suggesting disagreement on pressure and temperature calibration. We have also observed that kinetics of disordering is relatively rapid in the order of tens of minutes. On the contrary reversal kinetics (i.e. to create ordered state) is sluggish. Only small fraction of reversal ordering has occurred within 2 hours of heating. We consider that this asymmetric kinetics of the order- disorder reaction is a cause of disparity of the previous quench experiments
DI33A-1125
Carbonated Peridotite Xenoliths From the Western Canadian Cordillera
Carbonated mantle is inferred as the source of numerous alkaline magmas. However, carbonate is rarely expressed modally in mantle-derived peridotitic xenoliths. Most carbonate interpreted as primary (i.e. not an alteration product) occurs as interstitial patches associated with second generation crystallization, or as inclusions within crystals. Here we report on a unique occurrence of primary mantle-derived carbonate preserved in spinel peridotite xenoliths within a 19 Ma basanite dike. The dike intrudes Jurassic volcaniclastic rocks near the Intermontane - Coast Belt boundary in western British Columbia. The peridotite xenoliths are concentrated in a 4 meter interval where the dike is narrowest (1 m width); they are ≤40 cm in diameter, and are dominated by lherzolite, with less abundant harzburgite, dunite, and websterite. Thermometry (Brey & Köhler) produces xenolith equilibrium temperatures of 850 to 1000 °C, corresponding to approximately 35 to 50 km depth on an average geotherm for warm, thin Cordilleran-style lithosphere. The basanite also contains primary magmatic calcite occurring as a late stage groundmass crystallization product and as crystalline aggregates lining or filling vesicles. Both habits have isotopic compositions indicating a mantle origin (δ18O = 14 to 15, δ13C = -4 to -5, 87Sr/86Sr = 0.7040) and are isotopically distinct from secondary carbonate in the adjacent country rocks (δ18O = -5, δ13C = -5, 87Sr/86Sr = 0.7047).Calcite occurs in all but three of the suite of 50 xenoliths examined. The calcite appears both as grains that appear to have been in textural equilibrium with neighboring minerals (granoblastic texture), and as intergranular or fracture-filling veins. Calcite grains are ≤0.3 mm in size. The calcite comprises up to 2 % of xenoliths; calcium content (Ca / Ca + Mg + Fe) is ~0.90, contrasting with ~0.98 for the basanite's calcite. Carbonate C and O isotopes (whole rock leachate) indicate a range of δ18O = 10 to 12, δ13C = -3 to -6 for the suite. Ongoing studies are establishing the conditions required for preservation of this unique xenolith-hosted mantle-derived carbonate and the implications for source region and transport processes.
DI33A-1126
Sulfides in the Garnet Pyroxenite xenoliths from Oahu, Hawaii
Oahu is known for its garnet bearing xenoliths that occur in the Honolulu Volcanics. Clinopyroxene is the dominant minerals of these rocks, and modes of other silicate minerals â€" orthopyroxene, olivine, garnet, amphibole, and phlogopite vary considerably. Ilmenite and Spinels of diverse variety also occur (Keshav et al. 2007, J. Petrol.). In this report we present new electron microprobe and LA-ICPMS data on the sulfides that are always present in these xenoliths although they make up only trace amounts. In terms of morphology and mode of occurrence the sulfides can be divided fundamentally into two types â€" Type I occurs as poikilitic inclusions in the silicate phases mostly in clinopyroxene and Type II occurs in the interstitial spaces between the silicates, along grain boundaries and along cracks within individual silicate grains. Sizes of both types vary considerably. Type I sulfides are generally globular and appear to have formed from immiscible sulfide melts that got enclosed by the silicate minerals that grew from the main body of silicate melt. Keshav et al. (2007) estimate the average solidus temperatures of garnet pyroxenites from Oahu to range from 1215 to 1600°C (average 1325°C) at 3-5 GPa. Therefore, the Type I sulfides are high temperature sulfides that formed above the silicate solidus. Type II sulfides take various forms â€" from vein-like to dendritic. Compositionally, both types include Ni rich pyrrhotites (Ni content varies from 3-5 wt%) and monosulfide solid solutions(MSS). The MSS are divided into Ni rich MSS containing as much as 20 wt% of Ni, the average is 15 wt% while the Ni poor MSS has 5-9 wt% of Ni in it. We have limited data on PGE so far but the Type II sulfides have a very low PGE content. Two recent papers have noted that Hawaiian plume-derived shield tholeiites are too rich in Ni for a given SiO2% to be produced by partial melting of a peridotite and called for an unusual Ni-rich pyroxenite source in which the large Ni content is locked in clinopyroxene. The many experiments that have been conducted on pyroxenites have not been able to generate such high Ni clinopyroxenes. We propose that the Ni actually comes from the high Ni monosulfide solid solutions similar to those in the pyroxenites studied. However, we do not think that these pyroxenites are the source of Hawaiian shield lavas because their isotopic composition is distinct from shield lavas.
DI33A-1127
Deep Slab Subduction and Dehydration and Their Geodynamic Consequences: New Insights From Seismology and Mineral Physics
Fluids play important roles in the dynamics and evolution of the Earth, such as lowering the melting temperature of the mantle, transporting elements, enhancing diffusion and creep, and possibly changing the location of phase boundaries such as the 410 and 670 km discontinuities. Fluids are also intimately linked to a variety of earthquake faulting processes. In this work we present new pieces of evidence from seismology and mineral physics for the existence of significant low-velocity anomalies in the deep part of the upper-mantle wedge and the transition zone that are caused by fluids from the deep subduction and deep dehydration of the Pacific and Philippine Sea (PHS) slabs under western Pacific and East Asia. The PHS slab is found to subduct down to the mantle transition zone depth though the seismicity within the slab occurs only down to 200-300 km depths. The PHS slab dehydration has contributed to the formation of the Okinawa Trough and the Quaternary and active volcanism in SW Japan. Combining with the convective circulation processes in the mantle wedge, deep dehydration of the subducting Pacific slab has affected the morphology of the subducting PHS slab and its seismicity under SW Japan. The Pacific slab is found to stagnate in the mantle transition zone under East Asia, which has contributed to the formation of the continental rift system and intraplate volcanism in Northeast Asia (such as the active Changbai and Wudalianchi volcanoes). Slow anomalies are also found in the mantle under the subducting Pacific slab, which may represent (a) small mantle plumes, (b) mantle upwellings associated with the deep slab subduction, or (c) slab dehydration associated with deep earthquakes caused by the reactivation of large faults in the slab which are generated by the large normal-faulting earthquakes in the oceanic plate near the trench and are preserved during the slab subduction. http://www.aob.geophys.tohoku.ac.jp
DI33A-1128
Evolution of a hydrous silicate melt layer above the mantle transition zone
Recent research suggests that a dense hydrous silicate melt may form and accumulate at the 410-km discontinuity due to dehydration melting. A recent model by Leahy and Bercovici (2007) predicts that accumulated melt could form a steady-state structure that efficiently recycles water into the transition zone via a mechanism that relies upon viscous spreading of the melt along the discontinuity into regions of downgoing mantle flow. In these regions, melt is advected into the wadsleyite stability field, and the ensuing crystallization enriches the melt in water relative to the overlying olivine. The overlying solid then melts to restore equilibrium. This mechanism is hypothesized to permit the melt to spread further into mantle downwelling regions, allowing the establishment of a steady-state water and silicate cycle. Here we present a dynamic investigation of this mechanism to address whether the melt reaction or gravitational collapse governs the evolution of the melt layer towards steady state, whether transient or steady state solutions are more relevant on time scales controlling upper mantle chemistry, and whether the steady solutions are stable.
DI33A-1129
Precipitation of Excess Hydrogen in Olivine During Cooling Under Pressures: An Experimental Study
Water (hydrogen) content in olivine transported from the upper mantle is used to infer the water content in the upper mantle (e.g., Bell and Rossman 1992). However, since hydrogen diffusion is known to be fast, processes of hydrogen loss need to be examined. In many literature, diffusion loss (or gain) of hydrogen is usually considered, but in addition to diffusion loss, hydrogen could also precipitate inside of olivine as small inclusions. Consider an upward transport of olivine-bearing rock that originally contained a large amount of hydrogen in the deep interior. As this rock is transported to the shallow region, the solubility limit of hydrogen will decrease because of the reduction of pressure (and temperature) (Kohlstedt et al. 1996, Zhao et al 2004). Consequently, excess hydrogen will precipitate to form water bubbles and/or hydrous minerals as inclusions. Frequently observed submicron-scale inclusions of hydrous minerals (Khisina and Wirth 2002, Kitamura et al. 1987) may correspond to these precipitation products. If that is the case, hydrogen content corresponding to these minerals should not be excluded when estimating the hydrogen content of a sample in the Earth's upper mantle. However, kinetics of precipitation of hydrogen from olivine have not been investigated in the laboratory. We have conducted a series of experimental study in which we annealed hydrogen-saturated olivine single crystals in two different P- T conditions. The starting material was an olivine crystal in which ~1,135 H/106Si (70 wt ppm H2O) was dissolved at P= 3.5 GPa and T=1,573 K. A small piece of this crystal (0.5 mm3) was placed in a multianvil at P=3.5 GPa and either at T= 873K or 1,173K with oxygen fugacity, fO2, buffered by the Ni-NiO solid-state reaction and silica activity, aSiO2, buffered by the presence of orthopyroxene powder in contact with the crystal. Annealing experiments were conducted up to 72 hours. Hydroxyl concentrations were determined from infrared spectra obtained from polished thin sections from crack-free regions of 100 x 100 μm. The hydroxyl concentration at the OH-stretching region around 3678 cm-1 increases systematically with increasing time at 873 K, whereas this band is not detected in samples annealed at 1,173 K. The peak(s) at 3678 cm-1 corresponds to the OH-stretching vibration in hydrous minerals such as serpentine (Mellini et al. 2002, Hofmeister and Bowey 2006). We conclude that the water in the upper mantle not only diffuse out and disappear during the ascent (cooling and depressurization), but also is bounded in hydrous minerals (e.g. serpentine).
DI33A-1130
Magma Dynamics at Mid-Ocean Ridges by Noble Gas Kinetic Fractionation: Assessment of Magmatic Ascent Rates and Mantle Composition
Topical scientific literature on magma degassing at mid-ocean ridges more and more focuses on exsolution processes occurring under conditions that are far from thermodynamic equilibrium between bubbles and silicate melt. Indeed, the dynamics of magma ascent and decompression can be faster than that of CO2 diffusion into bubbles, in which case the diffusivity ratios among volatiles are the main control of the composition of the exsolving gas phase. We have developed a model of bubble growth in silicate melts that calculates the extent of both CO2 supersaturation and kinetic fractionation among noble gases in vesicles in relation to the decompressive rate of basaltic melts. The model predicts that, due to comparable Ar and CO2 diffusivity, magma degassing at low pressure fractionates both He/Ar and He/CO2 ratios by a similar extent, while the slower CO2 diffusion at high pressure causes early kinetic effects on Ar/CO2 ratio and dramatically changes the degassing paths. By using this tool, we have reviewed the global He-Ar-CO2 dataset of fluid inclusions in mid-ocean-ridge glasses. We display that non-equilibrium fractionations among He, Ar and CO2, driven by their different diffusivities in silicate melts, are common in most of the natural conditions of magma decompression and their signature strongly depends on pressure of degassing. The different geochemical signatures among suites of data coming from different ridge segments mainly depend on the depth of the magma chamber where the melt was stored. Moreover, variations inside a single suite emerge due to the interplay between variable ascent speed of magma and cooling rate of the emplaced lava. As a result, two data groups coming from the Pito Seamount suite (Easter Microplate East ridge), showing different degree of CO2 supersaturation and He/Ar fractionation, provide ascent rates which differ by ten folds or even more. The large variations in both the He/CO2 and Ar/CO2 ratios at almost constant He/Ar, displayed in products coming from the Mid-Atlantic Ridge 24°N segment and the Rodriguez Triple Junction, require magma storage and degassing processes occurring at high-pressure conditions. In contrast, the simultaneous increase in both He/CO2 and He/Ar of the East Pacific Rise and South-East Indian Ridge data sets suggests the dominance of low-pressure fractionation, implying that the shallow magma chambers are at a lower depth than those of the Mid-Atlantic Ridge 24°N and Rodriguez Triple Junction. Our conclusions support the presence of a relationship between spreading rate and depth of high-temperature zones below ridges, and are consistent with the depth of magma chambers as suggested from seismic studies. Finally, the non-equilibrium degassing model provides striking constraints on the compositions of noble gases and carbon in mantle-derived magmas. Our results dispense in fact with the supposed need for He-Ar-CO2 heterogeneities in the upper mantle, because the degassing of a single, popping-rock-like primary magma is able to explain all the available data.
DI33A-1131
Volatile Transport and Accumulation Timescales Modeled using 210Pb-226Ra Disequilibrium in Tephras From Mount Pinatubo Volcano, Philippines: Moving From the Timescales of 222Rn to 210Pb
We measured 210Pb-222Rn-226Ra disequilibria in samples representing a time sequence through the June 15, 1991 cataclysmic eruption tephras of Mount Pinatubo volcano, Philippines. Previous U- series degassing studies have interpreted 210Pb excess to indicate rapid volatile transport (differential 222Rn motion) and subsequent volatile accumulation. Mount Pinatubo deposits show evidence for volatile saturation and the presence of a pre-eruptive volatile phase. In addition, intensive monitoring at Mount Pinatubo prior to eruption provides gas emissions levels which allow us to decouple gas accumulation from passive degassing. Because the 1991 eruption of Mount Pinatubo was well studied, Mount Pinatubo is a uniquely constrained system with which to test hypotheses for the generation of 210Pb-226Ra disequilibria. Analyzed tephra units have (210Pb/226Ra)0 values that fall outside of equilibrium given 2σ analytical errors. These samples exhibit 210Pb excesses with (210Pb/226Ra)0 ranging from 1.079 to 1.119. The highest excess occurs in the 1992 dome sample extruded after the climactic eruption. Differential gas motion and gas accumulation are typically called upon to generate 210Pb excess; however, we find that physical bubble rise and gas fluxing through a permeable media are an ineffective means of 222Rn transport in Pinatubo's dacitic magma reservoir. Calculated Stokes rise velocities show that bubbles in dacite magma rise slower than is needed to allow for a flux of 222Rn large enough to generate measurable disequilibria in the 210Pb -226Ra system. Instead, we present a conceptual model in which 210Pb -226Ra disequilibria is established during basaltic under-plating of the magma reservoir. Basaltic magma has a viscosity low enough to allow for differential gas motion and the mobilization of 222Rn. 222Rn decays to 210Pb, imparting a 210Pb signal to melt at the bubble walls. The timescale of gas transport and accumulation is then constrained not by the half-life of 222Rn (3.8 days) but rather by the half-life of 210Pb (22.6 years). Our results suggest that the transport of buoyant melts containing exsolved volatiles is a physically viable way to preserve fractionation signals in the 210Pb - 226Ra system and alleviates the need for gas transport on very short time-scales. Further analyses of major and trace elements, including Li measurements, are currently in progress to provide additional constraints on volatile fluxing.
DI33A-1132
Volatile-rich komatiitic and picritic melt inclusions in Cr-spinel beach sand from Gorgona Island, Colombia
Volatile content of komatiite is a key to constrain thermal evolution of the deep Earth. We report volatile contents with major and trace element compositions of melt inclusions (MIs) in chromian spinel (Cr-spinel) from beach sands of Gorgona Island, Colombia. Gorgona Island is ~90 Ma volcanic island, where picrites and the world-youngest komatiites occur. As Cr-spinel is dense and rigid oxide mineral that crystallizes only at early stages of crystallization, it is considered to be a superior container for retaining primitive melt, even including volatiles. Volatile (H2O, CO2, S, F and Cl) and trace element (K2O, Sr, Y, Zr, Nb, Ba, La, Ce, Sm, Dy, Yb) compositions of ~80 MIs were analyzed by SIMS (Cameca-1280 and 3f, respectively) at WHOI. MIs in the Cr-spinel from Gorgona Is. are classified into three types by their host Cr-spinel compositions such as low-Ti (P-type), high-Ti with high-Cr# (BK-type) and high-Ti with low-Cr# (K-type). MIs of P-type, BK-type and K-type are mostly in compositional ranges of picrite, high TiO2 komatiite (some basalt) and low TiO2 komatiite in Gorgona Island, respectively. Water content of P-type MIs is variable, ranging from 0.05 to 0.9 wt%, whereas those of BK and K-type MIs are limited (< 0.1 wt%). On the other hand, CO2 contents of BK-type and K- type MIs are highly scattered (40 to 4200 ppm), whereas that of P-type is relatively constant at ~200 ppm. All MIs with high CO2 content (>500 ppm) do not contain (shrinkage) bubbles and many of them are low in K2O. H2O/K2O, CO2/K2O, S/K2O and F/K2O ratios are positively correlated with Y/Sr ratios, indicating degassing trends of melt at crystallization, magma mixing and/or assimilation. Undegassed H2O/K2O, CO2/K2O, S/K2O and F/K2O ratios of komatiitic (picritic) melt are estimated to be ~10 (~40), ~80 (n.d.), ~7(~3) and ~1(~0.5), respectively, which are much higher than those estimated for the depleted source mantle of the MORB [1.6, 0.7, 1.6 and 0.2, respectively; Salters, V. & Stracke, A. (2004), Composition of the depleted mantle. Geochem. Geophys. Geosys. 5 (2003GC000597)]. The results suggest that Gorgona komatiite and picrite magmas were derived from volatile-rich sources. CO2 degassing might also have contributed to eruption of high-density magmas to the surface. In addition, H2O, S, F and Cl contents in MIs in olivine from a picrite were identical to those of P-type MIs in Cr-spinel, but CO2 in olivine-hosted MIs were considerably lower (~50 ppm) than those in Cr-spinel. This indicates that entrapment pressure for MIs in Cr-spinel is likely to be greater than that for MIs in olivine. Therefore, in order to evaluate the volatile contents of undegassed magmas from oceanic islands, melt inclusions in Cr-spinel beach sand could be very useful.
DI33A-1133
Noble Gases Analyses of Samples Synthesized at High P and T in a Multi Anvil Press Device: Protocol and Implications
Noble gases (He-Ne-Ar-Kr-Xe) in mantle-derived samples allow an undisputable tracing of different sources of materials. Concerning the deep mantle part, the study of noble gases suggests that a "primordial" component (which is non or partially degassed) exists. Nevertheless, this conclusion is challenged by several observations, both geophysical and geochemical, suggesting that contrariwise the mantle is now totally depleted, degassed or renewed by convection. Furthermore, the lack of experimental data disables quantitative modelling of geochemistry processes. It is still unknown how much the fractionations are dependent on the conditions on pressure, temperature and chemical composition in the mantle. Recent studies [1-3] suggest a more incompatible behavior for noble gases in comparison to their parent element (K for Ar, U + Th for He) in very specific conditions of pressure, temperature, and chemical composition. Nevertheless, those studies focus on only particular compositions or pressures or only one single noble gas. No exhaustive studies (of all nobles gases at different pressures, temperatures and compositions) were accomplished on this subject so far. We set up a new experimental protocol allowing the analyses of rare gases in samples synthesized under mantle conditions, at high pressures and temperatures. This new protocol associates the use of a gas loading device [4], a multi-anvil press device (INSU MAP, Clermont-Ferrand, France), a laser ablation coupled to mass- spectrometer for the noble gases analysis (excimer laser, λ = 193 nm), and a 3D profilometry device to quantify the amount of ablated material. We will present an application of these methods on the noble gases partitioning between solid and liquid natural phases in the 3-5 GPa pressure range and for temperature of 1400 to 1600°C. [1] E.M. Chamorro, R.A Brooker, J.-A Wartho, B.J. Wodd, S.P. Kelley and J.D. Blundy. Ar and K partitioning between clinopyroxene and silicate melt to 8 GPa. Geochimica et Cosmochimica Acta, 66: 507-519, 2002. [2] S.W. Parman, M.D. Kurz, S.R. Hart and T. L. Groove. Helium solubility in olivine and implication for high 3He/4He in ocean island basalts. Nature, 437: 1140-1143, 2005. [3] V.S. Heber, R.A. Brooker, S.P Kelley and B.J. Wood. Crystal-melt partitioning of nobles gases (helium, neon, argon, krypton and xenon) for olivine and clinopyroxene. Geochimica et Cosmochimica Acta, 71: 1041-1061. [4] S.L. Boetcher, Q. Guo and A. Montana. A simple device for loading gases in high-pressure experiments. American Mineralogist, 74: 1383-1384, 1989.