Upper-mantle Conditions Research Articles

Magmatic water contents determined through clinopyroxene: Examples from the Western Canary Islands, Spain

AbstractWater is a key parameter in magma genesis, magma evolution, and resulting eruption styles, because it controls the density, the viscosity, as well as the melting and crystallization behavior of a melt. The parental water content of a magma is usually measured through melt inclusions in minerals such as olivine, a method which may be hampered, however, by the lack of melt inclusions suitable for analysis, or postentrapment changes in their water content. An alternative way to reconstruct the water content of a magma is to use nominally anhydrous minerals (NAMs), such as pyroxene, which take up low concentrations of hydrogen as a function of the magma's water content. During magma degassing and eruption, however, NAMs may dehydrate. We therefore tested a method to reconstruct the water contents of dehydrated clinopyroxene phenocrysts from the Western Canary islands (n = 28) through rehydration experiments followed by infrared and Mössbauer spectroscopy. Employing currently available crystal/melt partitioning data, the results of the experiments were used to calculate parental water contents of 0.71 ± 0.07 to 1.49 ± 0.15 wt % H2O for Western Canary magmas during clinopyroxene crystallization at upper mantle conditions. This H2O range is in agreement with calculated water contents using plagioclase‐liquid‐hygrometry, and with previously published data for mafic lavas from the Canary Islands and comparable ocean island systems elsewhere. Utilizing NAMs in combination with hydrogen treatment can therefore serve as a proxy for pre‐eruptive H2O contents, which we anticipate becoming a useful method applicable to mafic rocks where pyroxene is the main phenocryst phase.

Geochemical composition, petrography and 40Ar/39Ar age of the Heldburg phonolite: implications on magma mixing and mingling

Differentiated magmatic rocks such as trachyte and phonolite are volumetrically subordinate to mafic volcanic rocks within the Cenozoic Central European Volcanic Province (exceptions are the East Eifel and the Rhon volcanic fields). Within the volcanic field of the “Heldburg dike swarm” (Heldburger Gangschar), the phonolite of the Burgberg near Heldburg represents the only known occurrence of differentiated magmatic rocks. However, the Heldburg phonolite is famous foremost for containing mantle xenoliths (spinel lherzolite). Former studies proposing a cogenetic relationship between the phonolite and the peridotites concluded that the phonolite magma must have evolved under upper mantle conditions. Herewith, we present petrographic and geochemical evidence for magma mixing and mingling in the Heldburg phonolite melt due to the intrusion of mantle-derived basanitic magma, which is exposed today as dikes at the foot of the Heldburg Burgberg. During this process, the mantle xenoliths were introduced into the phonolite melt as they all contain rims of basanitic magma. Extensive mingling features (e.g., schlieren layers, load casts, flame structures, mafic enclaves) are developed, indicating that the basanite and the zoned phonolitic body were melts at the time of mixing. These petrographic and geochemical indications of two coeval melts of different composition are substantiated by 40Ar/39Ar dating, revealing identical ages of ca. 15 Ma.

Experimental constraints on the electrical anisotropy of the lithosphere-asthenosphere system.

The relative motion of lithospheric plates and underlying mantle produces localized deformation near the lithosphere-asthenosphere boundary. The transition from rheologically stronger lithosphere to weaker asthenosphere may result from a small amount of melt or water in the asthenosphere, reducing viscosity. Either possibility may explain the seismic and electrical anomalies that extend to a depth of about 200 kilometres. However, the effect of melt on the physical properties of deformed materials at upper-mantle conditions remains poorly constrained. Here we present electrical anisotropy measurements at high temperatures and quasi-hydrostatic pressures of about three gigapascals on previously deformed olivine aggregates and sheared partially molten rocks. For all samples, electrical conductivity is highest when parallel to the direction of prior deformation. The conductivity of highly sheared olivine samples is ten times greater in the shear direction than for undeformed samples. At temperatures above 900 degrees Celsius, a deformed solid matrix with nearly isotropic melt distribution has an electrical anisotropy factor less than five. To obtain higher electrical anisotropy (up to a factor of 100), we propose an experimentally based model in which layers of sheared olivine are alternated with layers of sheared olivine plus MORB or of pure melt. Conductivities are up to 100 times greater in the shear direction than when perpendicular to the shear direction and reproduce stress-driven alignment of the melt. Our experimental results and the model reproduce mantle conductivity-depth profiles for melt-bearing geological contexts. The field data are best fitted by an electrically anisotropic asthenosphere overlain by an isotropic, high-conductivity lowermost lithosphere. The high conductivity could arise from partial melting associated with localized deformation resulting from differential plate velocities relative to the mantle, with subsequent upward melt percolation from the asthenosphere.

Structural and elastic anisotropy of crystals at high pressures and temperatures from quantum mechanical methods: The case of Mg2SiO4 forsterite.

We report accurate ab initio theoretical predictions of the elastic, seismic, and structural anisotropy of the orthorhombic Mg2SiO4 forsterite crystal at high pressures (up to 20 GPa) and temperatures (up to its melting point, 2163 K), which constitute earth's upper mantle conditions. Single-crystal elastic stiffness constants are evaluated up to 20 GPa and their first- and second-order pressure derivatives reported. Christoffel's equation is solved at several pressures: directional seismic wave velocities and related properties (azimuthal and polarization seismic anisotropies) discussed. Thermal structural and average elastic properties, as computed within the quasi-harmonic approximation of the lattice potential, are predicted at high pressures and temperatures: directional thermal expansion coefficients, first- and second-order pressure derivatives of the isothermal bulk modulus, and P-V-T equation-of-state. The effect on computed properties of five different functionals, belonging to three different classes of approximations, of the density functional theory is explicitly investigated.

Na-Ca carbonates synthesized under upper-mantle conditions: Raman spectroscopic and X-ray diffraction studies

The new Na4Ca(CO3)3, Na2Ca3(CO3)4 and Na2Ca4(CO3)5 compounds were synthesized in the system Na2CO3–CaCO3 in multianvil experiments at 6 GPa and characterized by Raman spectroscopy. In addition, the Na2Ca3(CO3)4 compound was studied using in situ energy dispersive and single-crystal X-ray diffraction. Single bands in the CO32− symmetric stretching region ( v1 ) and out-of-plane bending region ( v2 ) in the Na4Ca(CO3)3 Raman spectrum suggest a single crystallographically distinct carbonate group in the structure. In contrast, the spectra of Na2Ca3(CO3)4 and Na2Ca4(CO3)5 show two and three bands, respectively, in both the symmetric stretching region ( v1 ) and out-of-plane bending region ( v2 ), suggesting more than one crystallographically distinct carbonate group in the unit cell. Raman activity in the forbidden v2 mode and multiple bands are observed in the in-plane bending region ( v4 ) for the three compounds, proving the reduction of site symmetry of the CO32− ions with the loss of the threefold rotation axis ( D 3h → D 2h or C s). Such a decrease in symmetry suggests distortion of the group itself, but may be attained by rearrangements of the coordinated metal cations as in the aragonite-group carbonates. At 6.5 GPa and 1000°C, the structure of Na2Ca3(CO3)4 was found to be orthorhombic or monoclinic with a β angle close to 90° and the lattice parameters: a = 7.3357(6) A, b = 8.0377(9) A, and c = 31.5322 (32) A, with V = 929.59(14) A3. No structural changes were observed during pressure decrease down to 1 GPa, while a discontinuous increase in unit-cell parameters and volume was observed upon decompression from 1 GPa at room temperature. This indicates a pressure-induced phase transition to a structurally related ambient-pressure phase. The abnormally long c -parameter and proximity of the β-angle to 90° of Na2Ca3(CO3)4 at ambient conditions suggest that, in the monoclinic system, the metric symmetry is higher than the Laue symmetry, which is a common sign for merohedral twinning.

Carbon dioxide in silicate melts at upper mantle conditions: Insights from atomistic simulations

The detail of the incorporation of carbon dioxide in silicate melts at upper mantle conditions is still badly known. To give some theoretical guidance, we have performed first-principle molecular dynamics simulations (FPMD) to quantify the speciation and the incorporation of carbon dioxide in two CO2-rich silicate melts (~20wt.% CO2 at 2073K and 12GPa), a basaltic and a kimberlitic composition chosen in the CaO–MgO–Al2O3–SiO2 system. In the basaltic composition, carbon dioxide is incorporated under the form of a minority population of CO2 molecules and a prevailing population of carbonate ions (CO32−). In contrast, the amount of CO2 molecules is found to be very small in the kimberlitic melt. Moreover, a new (transient) species has been identified, the pyrocarbonate ion C2O52− issued from the reaction between CO2 and CO32−. With regard to the structure of the CO2-bearing melts, it is shown that the carbonate ions modify the silicate network by transforming some of the oxygen atoms into bridging carbonates, non-bridging carbonates, and free carbonates, with a distribution depending on the melt composition. In the basaltic melt a majority of carbonate ions are non-bridging or free, whereas in the kimberlitic melt, most of the carbonate ions are under the form of free carbonates linked to alkaline earth cations. Surprisingly, the addition of CO2 only has a weak influence on the diffusion coefficients of the elements of the melt. The consequence is that the strong enhancement of the electrical conductivity reported recently for carbonated basalts (Sifré et al., 2014, Nature 509, 81), can be reproduced by simulation only if one assumes that the ionic charges assigned to the elements of the melt depend, in a non-trivial way, on the CO2 content. Finally, a comparison of the FPMD calculations with classical molecular dynamics simulations using an empirical force field of the literature (Guillot and Sator, 2011, GCA 75, 1829) shows that the latter one needs some improvement.

Melting of siderite to 20 GPa and thermodynamic properties of FeCO3-melt

The siderite (FeCO3) melting curve is determined through multi-anvil experiments at 6–20GPa, and 1300–1870°C. The experiments define a melting curve with a Clapeyron slope steepening from 45 to 18°C/GPa but without backbend at upper mantle conditions, i.e., siderite is denser than FeCO3-melt (FeCO3L). The melting curve fits Tm=1037(44)+70.0(88)*P−1.43(37)*P2 (valid from 5 to 20GPa) where pressure is in GPa and temperature in °C. Siderite melting is not stoichiometric, minor quench magnetite was always observed and is interpreted as the result of partial redox dissociation of FeCO3L leading to dissolved Fe3+ and CO2 in the carbonate melt. At pressures below ~6.8GPa, siderite does not melt but decomposes through an auto redox dissociation reaction to magnetite, a carbon polymorph and CO2. From the experimental determination of the pure siderite melting curve, we calculate thermodynamic properties of the FeCO3L end-member, which reproduce the siderite melting curve in P–T space better than 10°C. The metastable 1atm melting temperature is calculated to 1012°C. Siderite has the lowest melting temperature of the Ca–Mg–Fe carbonates, its melting curve may cross the mantle geotherm at transition zone pressures. The stability of siderite is not only dependent on pressure and temperature but also strongly on oxygen fugacity (fO2). Model calculations in P–T-fO2 space in a Fe–C–O2 system show that the siderite or FeCO3-melt maximum stability is always reached at conditions of the CCO buffer. Our experimental and thermodynamic data constitute a cornerstone to model carbonate melting in the deep Earth, necessary to understanding the deep carbon cycle.

Experimental reactions between olivine and orthopyroxene with phonolite melt: implications for the origins of hydrous amphibole + phlogopite + diopside bearing metasomatic veins

Amphibole + phlogopite + diopside bearing veins are observed in a large number of upper mantle xenoliths, but the composition of the melt that forms them is poorly constrained. Recent data from the Heldburg Phonolite, Central Germany, has shown that phonolite melt will react with olivine and orthopyroxene xenocrysts to form reaction rims of amphibole + phlogopite + diopside at mid-lower crustal pressures. This is the first example of where a melt has reacted with peridotite to form the mineralogy of the metasomatic veins. It is therefore necessary to explore whether a phonolite melt could be the parent melt that forms amphibole + phlogopite + diopside metasomatic veins. Experimental reactions between single crystals of olivine and orthopyroxene with phonolite melt were conducted at upper mantle conditions of 1.0–1.5 GPa and 900–1,000 °C. Melt water contents were varied from anhydrous to >12 wt. H2O. Olivine reacts to form phlogopite reaction rims with overgrowths of diopside 1,000 °C. Orthopyroxene reacts to form amphibole with epitaxial diopside overgrowths <1,000 °C. No reaction rims form when the bulk melt H2O is lower than ~3.8 wt%. Pressure has little effect over the small range tested. These experiments reproduce reaction rims on olivine and orthopyroxene observed in the Heldburg Phonolite, Central Germany, and suggest that a relatively narrow range of temperatures and melt water contents is required for rim formation. The compositions of rim amphibole, phlogopite and diopside from the experiments have very similar compositions to those from Heldburg but do not match those from metasomatic veins. Phenocrysts from Heldburg are similar to the metasomatic veins, suggesting that a phonolite could potentially form the veins if vein formation is dominated by crystallization rather than reaction and replacement of wall rock phases.

Single-crystal elasticity of the deep-mantle magnesite at high pressure and temperature

Magnesite (MgCO3) is considered to be a major candidate carbon host in the Earth's mantle, and has been found to exist as an accessory mineral in carbonated peridotite and eclogite. Studying the thermal elastic properties of magnesite under relevant pressure–temperature conditions of the upper mantle is thus important for our understanding of the deep-carbon storage in the Earth's interior. Here we have measured the single-crystal elasticity of a natural magnesite using in situ Brillouin spectroscopy and X-ray diffraction in a diamond anvil cell up to 14 GPa at room temperature and up to 750 K at ambient pressure, respectively. Using the third-order Eulerian finite-strain equations to model the elasticity data, we have derived the aggregate adiabatic bulk, KS0, and shear moduli, G0, at ambient conditions: KS0=114.7(±1.3) GPa and G0=69.9(±0.6) GPa. The pressure derivatives of the bulk and shear moduli at 300 K are (∂KS/∂P)T=4.82(±0.10) and (∂G/∂P)T=1.75(±0.10), respectively, while their temperature derivatives at ambient pressure are (∂Ks/∂T)P=−24.0(±0.2) MPa/K and (∂G/∂T)P=−14.8(±0.7) MPa/K. Based on the thermal elastic modeling of the measured elastic constants along an expected normal upper-mantle geotherm and a cold subducting slab, magnesite exhibits compressional wave (VP) anisotropy of approximately 46–49% and shear wave (VS) splitting of 37–41% that are much larger than those of major constituent minerals in the Earth's upper mantle including olivine, pyroxene, and garnet. The modeled aggregate VP and VS velocity in moderately carbonated peridotite and eclogite containing approximately 10 wt.% magnesite (approximately 5 wt.% CO2) show minimal effects of magnesite on the seismic profiles of these rock assemblages at upper mantle conditions, suggesting that the presence of magnesite is likely difficult to be detected seismically. However, due to its unusually high VP and VS anisotropies, magnesite with strong preferred orientations may exhibit sufficient VP and VS anisotropies that can have significant influences on seismic anisotropies of the regionally carbonated upper mantle.

The density of andesitic melts and the compressibility of dissolved water in silicate melts at crustal and upper mantle conditions

Many magmatic processes, including the extraction of partial melts, fractional crystallization and buoyancy overpressure, depend on the density of the silicate melt phase. Because of its low molecular weight, water has a disproportionately large effect on melt density and accurate data on the effect of water on melt density is required to model density-dependent processes. However, the volumetric properties of the hydrous component at crustal and mantle conditions remain debated. In this study, we determined the density of dry and hydrous (0–9 wt% H2O) haplo-andesitic melts from 0.7 to 4.0 GPa and 1600 to 1930 K using the X-ray absorption method in a Paris–Edinburgh press to determine the partial molar volume of water at high pressure. Our in situ data were combined with atmospheric pressure data to derive the first experimental equation of state for andesitic melts valid for crustal and subduction zone conditions, with KT=17.2 GPa and K′¯=6.3−2.5+3.3 at 1273 K. In addition, the partial molar volume and compressibility of dissolved water in andesitic melts were derived from the in situ data, with V0¯=19.2−1.9+2.4 cm3/mol, K¯T=2.7−2.2+7.6 GPa and K′¯=3.6−2.7+13.5 at 1273 K. Combined with available literature data for other melt compositions, our results support a compositionally independent partial molar volume of water in silicate melts, at least within the currently attainable precision of the in situ measurements. This observation greatly simplifies the construction of future models for magma density.

Experimental determination of carbon isotope fractionation between graphite and carbonated silicate melt under upper mantle conditions

Carbon isotope fractionation between graphite and carbonated silicate melt was determined at 5 GPa and in the temperature range between 1400 and 1900 °C. High pressure experiments were carried out in the carbon-saturated model harzbergite system (Enstatite–Magnesite–Olivine–Graphite), where carbonated silicate melt and graphite were the two stable carbon-bearing phases in the run products. Carbonated silicate melting resulted in an isotopic fractionation between graphite and carbon in the silicate melt, where the carbon in the melt is 13C enriched than co-existing graphite. 13C enrichment in carbonate melt when compared to graphite was further confirmed in experiments where redox melting between olivine and graphite produced carbonate melt as well as carbonate reduction experiments to form graphite. Although a quantitative estimate of fractionation between carbonate melt and graphite could not be obtained, our results indicate that mantle melting in the presence of graphite can result in progressive 13C carbon isotope enrichment in carbonate melt and depletion in graphite, which can be an alternate explanation for the carbon isotopic heterogeneity observed in the mantle derived carbon.

Hydration of mantle olivine under variable water and oxygen fugacity conditions

The incorporation of H into olivine is influenced by a significant number of thermodynamic variables (pressure, temperature, oxygen fugacity, etc.). Given the strong influence that H has on the solidus temperature and rheological behavior of mantle peridotite, it is necessary to determine its solubility in olivine over the range of conditions found in the upper mantle. This study presents results from hydration experiments carried out to determine the effects of pressure, temperature, and the fugacities of H2O and O2 on H solubility in San Carlos olivine at upper mantle conditions. Experiments were carried out at 1–2 GPa and 1,200 °C using a piston-cylinder device. The fugacity of O2 was controlled at the Fe0–FeO, FeO–Fe3O4, or Ni0–NiO buffer. Variable duration experiments indicate that equilibration is achieved within 6 h. Hydrogen contents of the experimental products were measured by secondary ion mass spectrometry, and relative changes to the point defect populations were investigated using Fourier transform infrared spectroscopy. Results from our experiments demonstrate that H solubility in San Carlos olivine is sensitive to pressure, the activity of SiO2, and the fugacities of H2O and O2. Of these variables, the fugacity of H2O has the strongest influence. The solubility of H in olivine increases with increasing SiO2 activity, indicating incorporation into vacancies on octahedral lattice sites. The forsterite content of the olivine has no discernible effect on H solubility between 88.17 and 91.41, and there is no correlation between the concentrations of Ti and H. Further, in all but one of our experimentally hydrated olivines, the concentration of Ti is too low for H to be incorporated dominantly as a Ti-clinohumite-like defect. Our experimentally hydrated olivines are characterized by strong infrared absorption peaks at wavenumbers of 3,330, 3,356, 3,525, and 3,572 cm−1. The heights of peaks at 3,330 and 3,356 cm−1 correlate positively with O2 fugacity, while those at 3,525 and 3,572 cm−1 correlate with H2O fugacity.

Elasticity of grossular-andradite solid solution: an ab initio investigation.

Grossular and andradite are garnet end-members stable under upper mantle conditions. We perform ab initio simulations to investigate the dependence of the bulk modulus on chemical composition of the grossular-andradite solid solution, Ca3Fe(2-2x)Al(2x)(SiO4)3. All-electron local basis sets of Gaussian-type orbitals and the hybrid B3LYP density functional are used. Our calculations predict a linear modulus-composition trend, in contrast to previous conjectures based on "heterogeneous" experimental measurements. We estimate the largest deviation from linearity to be about 0.5 GPa under ambient conditions, and to progressively reduce to less than 0.2 GPa at pressure P = 20 GPa. The bulk modulus is computed over the whole composition range 0 ≤x≤ 1 following two independent approaches: fitting energy-volume data to an equation-of-state and calculating elastic tensors. Results from the two methods are in perfect agreement, assuring consistency and high numerical accuracy of the adopted algorithms.

Rheology of fine‐grained forsterite aggregate at deep upper mantle conditions

AbstractHigh‐pressure and high‐temperature deformation experiments on fine‐grained synthetic dunite (forsterite aggregate) were conducted to determine the dominant deformation mechanism in the deep upper mantle. The sintered starting material has 90% forsterite, 10% enstatite, and an average grain size of ~1 µm. Deformation experiments were performed using a deformation‐DIA apparatus at pressures of 3.03–5.36 GPa, temperatures of 1473–1573 K, and uniaxial strain rates of 0.91 × 10−5 to 18.6 × 10−5 s−1 at dry circumstances &lt;50 H/106Si. The steady state flow stress was determined at each deformation condition. Derived stress‐strain rate data is analyzed together with that reported from similar but low‐pressure deformation experiments using flow law equations for diffusion creep (stress exponent of n = 1, grain‐size exponent of p = 2) and for dislocation‐accommodated grain‐boundary sliding (GBS‐disl, n = 3, p = 1). The activation volume for diffusion creep (V*dif) and for GBS‐disl (V*GBS) of dunite is determined to be 8.2 ± 0.9 and 7.5 ± 1.0 cm3/mol, respectively. Calculations based on these results suggest that both diffusion creep and dislocation creep play an important role for material flow at typical deformation conditions in the Earth's asthenospheric upper mantle whereas the contribution of GBS‐disl is very limited, and dislocation creep is the dominant deformation mechanism during the deformation of olivine in sheared peridotite xenolith. Though these conclusions are not definitive, these are the first results on potential deformation mechanisms of forsterite aggregate based on extrapolation in the pressure, temperature, stress, and grain‐size space.

Experimental constraints in the CMAS system on the Ca-Eskola content of eclogitic clinopyroxene

The seismic X-Discontinuity, occasionally observed at ~300 km depth, has been attributed partly to the exsolution of the calcium-Eskola (Ca 0.5 □ 0.5 AlSi 2 O 6 , CaEs) component from eclogitic clinopyroxene upon reaching the coesite–stishovite transition. To test if this is a viable mechanism, we have undertaken high-pressure experiments between 4 and 11.5 GPa at 1000–1350 °C in order to further investigate the factors controlling CaEs incorporation in clinopyroxene at upper mantle conditions up to coesite–stishovite transition. Bulk compositions had pyroxene stoichiometry in the simple system CaO–MgO–Al 2 O 3 –SiO 2 ± Na 2 O (CMAS ± Na), which restricts possible end-member components to diopside (di), calcium-Tschermaks (CaTs), calcium-Eskola (CaEs), and clinoenstatite (en), ±jadeite. All run products had a typical eclogitic mineral assemblage with clinopyroxene always in equilibrium with garnet and coesite or stishovite, and in some runs kyanite. Clinopyroxene is non-stoichiometric due to the defect-bearing CaEs component. CaEs and CaTs contents vary systematically with P , T and bulk composition. CaEs reaches up to 0.22 mol % at 4 GPa and 1200 °C, and for a given bulk composition CaEs decreases continuously with pressure from 4 to 11.5 GPa. Our experiments, along with experiments performed with more complex compositions reveal different systematics when Na is present, since it is only accommodated in clinopyroxene. Incorporation of a jadeite component drives down the CaTs content and also influences the CaEs content. Generally, maximizing the CaEs content requires a coexisting free SiO 2 phase and an elevated Al 2 O 3 content, but it is not clear if kyanite must be part of the assemblage, or whether buffering of Al by garnet is adequate. At P > 9 GPa, the systematics of clinopyroxene composition change as garnet becomes majoritic and also begins to accommodate significant Na. In both simple and complex bulk compositions CaEs content in clinopyroxene decreases steadily from ~4 GPa so that no sharp change occurs at pressures corresponding to the coesite–stishovite phase transition. Calculations reveal that only small (1–2 wt %) amounts of free stishovite can be exsolved from CaEs-bearing clinopyroxene, which are insufficient to produce a large enough impedance contrast to explain a significant contribution to the X-Discontinuity. If the X-Discontinuity is related to the coesite–stishovite phase transition in eclogite, as proposed in the literature, then free SiO 2 must be already present in the mineral assemblage rather than being exsolved from CaEs-bearing clinopyroxene upon entering the stability field of stishovite.

The microstructural record of porphyroclasts and matrix of partly serpentinized peridotite mylonites – from brittle and crystal-plastic deformation to dissolution–precipitation creep

Abstract. We present microfabrics in high-pressure, metamorphic, partly serpentinized peridotite mylonites from the Voltri Massif, in which porphyroclasts and matrix record independent deformation events. The microfabrics are analysed using polarization microscopy and electron microscopy (SEM/EBSD, EMP). The mylonites contain diopside and olivine porphyroclasts originating from the mantle protolith embedded in a fine-grained matrix consisting mainly of antigorite and minor olivine and pyroxene. The porphyroclasts record brittle and crystal-plastic deformation of the peridotite at upper-mantle conditions and differential stresses of a few hundred MPa. After the peridotites became serpentinized, deformation occurred mainly by dissolution–precipitation creep resulting in a pronounced foliation of the antigorite matrix, crenulation cleavages and newly precipitated olivine and pyroxene from the pore fluid at sites of dilation, i.e. in strain shadows next to porphyroclasts and folded fine-grained antigorite layers. Antigorite reveals a pronounced associated shape preferred orientation (SPO) and crystallographic preferred orientation (CPO) with the basal (001) cleavage plane oriented in the foliation plane. In monomineralic antigorite aggregates at sites of stress concentration around porphyroclasts, a characteristically reduced grain size and deflecting CPO as well as sutured grain boundaries indicate also some contribution of crystal-plastic deformation and grain-boundary migration of antigorite. In contrast, the absence of any intragranular deformation features in newly precipitated olivine in strain shadows reveals that stresses were not sufficiently high to allow for significant dislocation creep of olivine at conditions at which antigorite is stable. The porphyroclast microstructures are not associated with the microstructures of the mylonitic matrix, but are inherited from an independent earlier deformation. The porphyroclasts record a high-stress deformation of the peridotite with dislocation creep of olivine in the upper mantle probably related to rifting processes, whereas the serpentinite matrix records dominantly dissolution–precipitation creep and low stresses during subduction and exhumation.

Partitioning of copper between olivine, orthopyroxene, clinopyroxene, spinel, garnet and silicate melts at upper mantle conditions

Previously published Cu partition coefficients (DCu) between silicate minerals and melts cover a wide range and have resulted in large uncertainties in model calculations of Cu behavior during mantle melting. In order to obtain true DCumineral/melt values, this study used Pt95Cu05 alloy capsules as the source of Cu to experimentally determine the DCu between olivine (ol), orthopyroxene (opx), clinopyroxene (cpx), spinel (spl), garnet (grt) and hydrous silicate melts at upper mantle conditions. Three synthetic silicate compositions, a Komatiite, a MORB and a Di70An30, were used to produce these minerals and melts. The experiments were conducted in piston cylinder presses at 1.0–3.5GPa, 1150–1300°C and oxygen fugacities (fO2) of from ∼2log units below to ∼5log units above fayalite–magnetite–quartz (FMQ). The compositions of minerals and quenched melts in the run products were measured with EMP and LA-ICP-MS. Attainment of equilibrium is verified by reproducible DCu values obtained at similar experimental conditions but different durations. The results show that DCu for ol/, opx/, spl/ and possibly cpx/melt increase with increasing fO2 when fO2>FMQ+1.2, while DCu for cpx/ and spl/melt also increase with increasing Na2O in cpx and Fe2O3 in spinel, respectively. In the investigated P–T–fO2 conditions, the DCumineral/melt values are 0.04–0.14 for ol, 0.04–0.09 for opx, 0.02–0.23 for cpx, 0.19–0.77 for spl and 0.03–0.05 for grt. These results confirm that Cu is highly incompatible (DCu<∼0.2) in all the silicate minerals and oxides of the upper mantle with the exception of the high-Fe spinel, in which Cu is moderately incompatible (DCu=0.4–0.8) and thus Cu will be enriched in the derived melts during mantle partial melting and magmatic differentiation if sulfide is absent. These experimental DCu values are used to assess the controls on Cu behavior during mantle melting. The model results suggest that MORBs and most arc basalts must form by sulfide-present melting at relatively reduced conditions, while high Cu (>70ppm) arc basalts may form at oxidized, sulfide-absent conditions, which is consistent with the possibility of some high fO2 regions present in the arc mantle.

Experimental determination of siderite stability at high pressure

The stability field of siderite has been determined up to 10 GPa. Decarbonation of siderite occurs at pressures below 6 GPa with a Clapeyron slope of about 0.0082 GPa/K. At higher pressure, we observed direct melting of siderite without decarbonation. The melting temperature is about 1550 °C at 10 GPa. Our experimental results, compared with previous studies on the decomposition curve of magnesite, indicate that Fe has a significant effect on the stability of magnesite-siderite solid solutions under upper mantle conditions. The reaction products are strongly dependent on the oxygen fugacity of the system. The disproportionation reaction during decomposition of siderite might be an important mechanism to explain the stability of carbon as graphite (diamond) in the Earth’s mantle.

Post-entrapment modification of volatiles and oxygen fugacity in olivine-hosted melt inclusions

The solubilities of volatiles (H2O, CO2, S, F, and Cl) in basaltic melts are dependent on variables such as temperature, pressure, melt composition, and redox state. Accordingly, volatile concentrations can change dramatically during the various stages of a magma's existence: from generation, to ascent through the mantle and crust, to final eruption at the Earth's surface. Olivine-hosted melt inclusions have the potential to preserve volatile concentrations at the time of entrapment due to the protection afforded by the host olivine against decompression and changes to the oxidation state of the external magma. Recent studies, however, have demonstrated that rapid diffusive re-equilibration of H2O and oxygen fugacity (fO2) can occur within olivine-hosted melt inclusions. Here we present volatile, hydrogen isotope, and major element data from dehydration experiments and a quantitative model that assesses proposed mechanisms for diffusive re-equilibration of H2O and fO2 in olivine-hosted melt inclusions. Our comprehensive set of data for the behavior of common magmatic volatiles (H2O, CO2, F, Cl, and S) demonstrates that post-entrapment modification of CO2, and to a lesser extent S, can also occur. We show that the CO2 and S concentrations within an included melt decrease with progressive diffusive H2O loss, and propose that this occurs due to dehydration-induced changes to the internal pressure of the inclusion. Therefore, deriving accurate estimates for pre-eruptive CO2 and S concentrations from olivine-hosted melt inclusions requires accounting for the amount of CO2 and S hosted in vapor bubbles. We find, however, that Cl and F concentrations in olivine-hosted melt inclusions are not affected by diffusive re-equilibration through the host olivine nor by dehydration-induced pressure changes within the melt inclusion. Our results indicate that measured H2O, CO2 and S concentrations and Fe3+/ΣFe ratios of included melts are not necessarily representative of the melt at the time of entrapment and thus are not reliable proxies for upper mantle conditions.

Experimental chlorine partitioning between forsterite, enstatite and aqueous fluid at upper mantle conditions

Cl partition coefficients between forsterite, enstatite and coexisting Cl-bearing aqueous fluids were determined in a series of high pressure and temperature piston cylinder experiments at 2GPa between 900 and 1300°C in the system MgO–SiO2–H2O–NaCl–BaO–C±CaCl2±TiO2±Al2O3±F. Diamond aggregates were added to the experimental capsule set-up in order to separate the fluid from the solid residue and enable in situ analysis of the quenched solute by LA-ICP-MS. The chlorine content of forsterite and enstatite was measured by electron microprobe, and the nature of hydrous defects was investigated by infrared spectroscopy. Partition coefficients show similar incompatibility for Cl in forsterite and enstatite, with DClfo/fl=0.0012±0.0006, DClen/fl=0.0018±0.0008 and DClfo/en=1.43±0.71. The values determined for mineral/fluid partitioning are very similar to previously determined values for mineral/melt. Applying the new mineral/fluid partition coefficients to fluids in subduction zones, a contribution between 0.15% and 20% of the total chlorine from the nominally anhydrous minerals is estimated.Infrared spectra of experimental forsterite show absorption bands at 3525 and 3572cm−1 that are characteristic for hydroxyl point defects associated with trace Ti substitutions, and strongly suggest that the TiO2 content of the system can influence the chlorine and OH incorporation via the stabilization of Ti-clinohumite-like point defects. The water contents for coexisting forsterite and enstatite in some runs were determined using unpolarized IR spectra and calculated water partition coefficients DH2Ofo/en are between 0.01 and 0.5.