Experimental Petrology

In this study, we show how veined lithospheric mantle is involved in the genesis of ultrapotassic magmatism in cratonic settings. We conducted high pressure experiments to simulate vein + wall rock melting within the Earth's lithospheric mantle by reacting assemblages of harzburgite and phlogopite-rich hydrous mantle xeno-liths. These comprised a mica-, amphibole-, rutile-, ilmenite-, diopside (MARID) assemblage at 3–5 GPa and 1325–1450 °C. Melting of the MARID assemblages results in infiltration of melt through the harzburgite, leading to its chemical alteration. At 3 and 4 GPa, melts are high in K 2 O (> 9 wt%) with K 2 O/Na 2 O > > 2 comparable to anorogenic lamproites. Higher pressures and temperatures (5 GPa/1450 °C) lead to increasing MgO contents of the melt and to some extent lower K 2 O contents (5–7 wt%) at equally high K 2 O/Na 2 O ratios. Our experiments provide insights into the role of alkalis in nickel-partitioning (D Ni) between olivine and ultrapotassic melt. We observe that the high contents of Na, K, and Al are indicative of high D Ni values, implying that the melt polymerization is the dominant factor influencing the olivine/melt nickel partitioning. The change of D Ni as a function of melt composition results in a pressure independent, empirical geothermometer: = − ⎛ ⎝ ⎜ ⎜ + ⎡ ⎣ + ⎤ ⎦ + ⎞ ⎠ ⎟ ⎟ + ° () liq liq liq T (C) 88.14·ln [((Na O K O)]· 1 ·[liq(Al O)] [ (MgO FeO)] ·D 1906.2 SiO TiO 2 2 2 2 2 3 Ni Element oxides represent the composition of the glass (in wt%), and D Ni is the liquid/olivine Ni-partitioning coefficient. We propose that this geothermometer is applicable to all natural silicate melts that crystallized olivine in a temperature interval between 1000 and 1600 °C. Application to glass-olivine pairs from calc-alkaline settings (Mexico), MORB (East Pacific Rise), and OIB (Hawaii) yielded reasonable values of 996–1199 °C, 1265 °C, and 1330 °C, respectively.

Amphibole crystallization from hydrous high-magnesium andesite (Shiveluch volcano, Kamchatka) melt has been studied experimentally at pressures 2, 3, 5 kbar. Slightly subliquidus temperatures have been selected for amphibole crystallization. Amphiboles with close to the equilibrium composition have been obtained at low undercoolings in short 4 hours experiments. We select octahedral aluminium content and sum of high charged cations (Ti4+, Fe3+) all evaluated with the 13eCNK method as master parameters reflecting pressure of amphibole formation. Asymptotic maximum of Al^{VI} content at the sum Ti+Fe^{3+} =0 as a function of pressure was calculated following negative linear correlation trend for each run of the original and published experimental data sets. The derived novel Al^{VI} in amphibole barometer is calibrated in the pressure range of 2-12 kbar for andesitic and basaltic magmas. Reliability and validity of the new geobarometer were tested on the amphibolized ultramafic xenoliths from the Shiveluch Volcano (Kamchatka) and Dish Hill (California). In both cases our estimates appear to be equal within the uncertainty of the method to the predictions done by two-pyroxene geo-thermo-barometer (Putirka, 2008).

Garnet is the most commonly used mineral in thermobarometry, whereas zircon is the most robust chronometer to date high-grade metamorphic rocks. To provide a basis for correlation of zircon and garnet growth, we determined experimentally the trace element partitioning between zircon, a hydrous granitic melt and garnet at 20 kbar and 800-1000°C for P, Y, rare earth elements (REE), Zr, Hf, Th and U. In respect to melt, zircon preferentially incorporates all investigated elements apart from REE with atomic number lower than Sm. At this pressure and in the chosen composition, the distribution coefficient between zircon and melt (D Zrn/Melt ) for REE increases with increasing atomic number of the REE and with decreasing temperature. D Yb Zrn/Melt is ∼20 at 1000°C, but more than an order of magnitude higher at 800°C. The solubility of Zr in hydrous granitic melts buffered by zircon is about a factor of two lower at 20 kbar than what has been previously established for mid-crustal pressures. Large garnet produced in the experiments allowed determination of garnet/melt trace element partitioning (D Grt/Melt ) at temperatures of 800-1000°C, conditions relevant for partial melting of crustal rocks. There is a systematic increase in D REE Grt/Melt with decreasing temperature. Zircon contains significantly more heavy-REE than garnet at temperatures of 800-850°C. Zircon/garnet partition coefficients of heavy-REE decrease with increasing temperature from D Lu Zrn/Melt of 12 at 800°C to 1.4 at 1000°C. Middle-REE partitioning is close to unity for the whole investigated temperature range. Different empirically determined zircon/garnet partition coefficients from granulites and ultra-high temperature granulites can potentially be explained by the experimentally determined change of partitioning as a function of temperature. These data can assist in establishing equilibrium between garnet and zircon zones in natural rocks, and in the construction of detailed pressure-temperature-time paths in high-grade metamorphic rocks.

Hau¨y coined the term eclogite, meaning ''chosen rock'', in 1822, but de Saussure had already observed rocks of this type in the Alps four decades earlier. Throughout the 19th century, the origin of eclogite remained an enigma, in spite of great progress in our knowledge of this rock. The first chemical analyses, carried out around 1870, showed that its bulk composition was the same as gabbro. Therefore, eclogite was thought to be either an igneous rock of gabbroic composition or a metamorphosed gabbro. This second hypothesis became preferred when progressive transitions were observed between gabbros and eclogites. In 1903, simply by comparing the molar volumes of gabbroic and eclogite parageneses, Becke inferred that eclogite was the high-pressure equivalent of gabbro. In 1920, eclogite was involved in the conception of the metamorphic facies by Eskola. However, a few researchers denied the existence of an eclogite facies, and claimed that high stress instead of high lithostatic pressure could generate eclogites. In the 1960s, consideration of the water pressure parameter also favoured the belief that eclogite was simply the anhydrous equivalent of amphibolite. Finally, eclogite was definitely considered as a high-pressure metamorphic rock following the development of experimental petrology and the application of thermodynamics.

Experimental investigations and thermodynamic calculations of the phase relations of a cordierite-rich monzogranite from the Cabeza de Araya batholith (Cáceres, Spain) have been performed to understand the formation of cordierite. The experiments failed to crystallize cordierite in the pressure range 200-600MPa, in the temperature range 700-975ºC and for different water activities (melt water contents between 2 and 6 wt.%). In contrast, clinopyroxene and orthopyroxene (absent in the natural mineral rock assemblage), together with biotite, were observed as ferromagnesian assemblage in a wide range of experimental conditions. Thermodynamic calculations, using the software PERPLE_X, describe the formation of cordierite only at 200 and 400MPa and very low water contents, and the amount of cordierite formed in the models is always below 3.5 vol.%. The results indicate that cordierite is not in equilibrium with the bulk rock compositions. The most probable explanation was that cordierite nucleated and crystallized from a melt that is not in equilibrium with part of the mineral assemblage present in the magma. This " non-reactive " mineral assemblage was mainly composed of plagioclase. The silicate melts from which cordierite crystallized was more Al-rich and K-rich than the silicate melt composition in equilibrium with the bulk composition. One possible process for the high Al content of the silicate melt is related to assimilation and partial melting of Al-rich metasediments. An exo-perictetic reaction is assumed to account for both textural and geochemical observations. On the other hand, hybridization processes typical for calc-alkaline series can also explain the high proportions of " non-reactive " minerals observed in relatively high temperature magmas. This study clearly demonstrates that silicate melts in a crystal mush can depart significantly from the composition of melt that should be in equilibrium with the bulk solid assemblage.

Zircon was grown from trace-element doped hydrous peralkaline rhyolite melts with buffered oxygen fugacities in cold-seal experiments at 0.1 and 0.2 GPa and 800 °C and piston-cylinder experiments at 1.5 GPa and 900–1300 °C. Zircon and glass were present in all run products, and small monazite crystals were present in eight of the 12 experiments. Average diameters of zircon crystals ranged from 5 to 20 lm at 800 °C to 30–50 lm at 1300 °C. Zircon crystals have thin rims, and adjacent glass has a narrow (1 um thick) compositional boundary layer. Concentrations obtained through in-situ analysis of cores of run product zircon crystals and melt pools were used to calculate trace-element partition coefficients D zircon/melt for In most cases Lu was the most (D 12–105) and La the least (0.06–0.95) compatible elements. D values from this study fall within the range of previously measured values for Rare Earth Elements (REE). However, D values measured experimentally show less fractionation than those recently measured using natural phenocryst/matrix pairs. For example, D Lu /D La measured experimentally in this study range between 27 and 206 compared to a value of 706,522 for a natural zircon/dacite pair [Sano, Y., Terada, K., and Fukuoka, T. 2002 High mass resolution ion micro-probe analysis of rare earth elements in silicate glass, apatite and zircon: lack of matrix dependency. Chem. Geol. 184, 217– 230]. Although D values from this study show good agreement with the lattice strain model, D values from natural pheno-cryst/matrix pairs combined with measured zircon compositions better reproduce host-rock (magma) compositions of igneous rocks. They also yield more reasonable estimates of magma compositions when combined with compositions of ''out-of-context " zircons. For example, compositions of the Hadean detrital zircons from Jack Hills, Australia yield LREE-enriched mag-mas when combined with D values from phenocryst/matrix pairs yields, but yield LREE-depleted magmas when experimentally determined D values are used. We infer that experimentally measured D zircon/melt values represent disequilibrium partitioning resulting from rapid zircon growth during short laboratory timescales. Rapid growth causes development of observed diffusive boundary layers in the melt adjacent to zircon crystals. D values from phenocryst/matrix pairs are therefore recommended for petrogenetic modeling.

One of the most outstanding characteristics of some granodioritic to granitic rocks is the presence of K-feldspar megacrysts. For instance, granodiorites and monzogranites of the Spanish Central System batholith present variable amounts of large (up to 10cm in length) euhedral K-feldspar crystals. The porphyritic textures, the euhedral shape, the alignment of plagioclase and biotite inclusions and the magmatic fabrics point to a magmatic origin for these megacrysts.
This work presents a phase equilibria study in a high-K2O granodioritic system. A series of experiments were conducted with a granodioritic composition (GEMbiot) to study the crystallization sequence at the emplacement conditions in the Gredos massif, i.e. 4 H2O wt.% and 0.4GPa. Experimental results show that orthopiroxene is the liquidus phase at 1010ºC, which reacts with the H2O-rich melt to stabilize biotite between 980 and 940ºC. Plagioclase crystallizes at around 910ºC, and K-feldspar crystallizes in the matrix between 750 and 700ºC when the crystal fraction is around 0.5. However, at 850ºC, a pelite-doped experiment shows euhedral K-feldspar (≈5vol%) in both the reactive xenolith domain together with cordierite and the granodioritic domain, where the K2O wt.% rose from 4.5 in the normal experiment to 5.9 in the doped experiment. These results suggest that the bulk-assimilation process promotes the bulk and heterogeneous K2O enrichment in a huge granodioritic magma volume, which triggers an early crystallization of K-feldspar megacrysts. Because of this early crystallization of the megacrysts, the magmatic foliations defined by K-feldspar megacrysts are formed during and after the emplacement processes and are highly influenced by tectonic kinematics.

Alkaline lavas occur globally in subduction-related volcanic arcs. Conventional models for the origin of these lavas typically invoke a multi-stage process in which mantle wedge peridotite, enriched in phlogopite and/or amphibole due to prior metasomatism, partially melts during infiltration by fluids and melts derived from subducted oceanic lithosphere. However, geochemical systematics in the majority of subduction-related alkaline lavas require physical mixing of subducted components and peridotite prior to partial melting. This can be explained by the mélange diapir model, which predicts the generation of arc magmas during advection of buoyant material from the slab-wedge interface into the mantle wedge below arcs. Here we report results from experiments in which natural mélange materials were partially melted at upper mantle conditions to produce alkaline magmas. Partial melts produced in our experiments have trace-element abundance patterns that are typical of alkaline arc lavas, such as enrichment in large ion lithophile elements (LILEs) and depletion in Nb and Ta. These results favor generation of alkaline magmas in the arc and backarc regions of subduction zones by partial melting of mélange materials rather than previously metasomatized peridotite.

The investigation of deeply subducted eclogites from the Dora-Maira massif, Western Alps reveals that accessory minerals are important hosts for trace elements. Rutile contains most of the bulk rock Ti, Nb and Ta while zircon hosts nearly all Zr and Hf. More than 90% of the bulk rock light rare earth elements (LREE) and Th and about 75% of U are incorporated in accessory allanite. Phengite is the most interesting major mineral because it hosts more than 95% of Rb, Ba and Cs. Synthesis piston cylinder experiments in a model crustal composition in the range 2.0 -4.5 GPa, 680 -1150 jC doped with trace elements demonstrate that accessory allanite forms at the expense of the major mineral zoisite at temperatures above 700 jC and 2.0 GPa. Allanite is stabilised by the incorporation of light rare earth elements and was found up to temperatures of 1050 jC and to pressures of at least 4.5 GPa. Disappearance of allanite is caused by dissolution in a coexisting hydrous granitic melt and is not related to the breakdown of any major phase. Experimentally determined element partitioning between allanite and a hydrous granitic melt at 900 jC, 2.0 GPa yield D LREE all/melt of about 200, D Th all/melt of 60 and D U all/melt of 20. These results, combined with literature data, are used to estimate light rare earth element saturation levels of subduction zone liquids as a function of temperature. Allanite is a residual phase up to temperatures of at least 900 and 1000 jC for metabasalts and metasediments, respectively, at V 10% partial melting. Allanite is therefore capable of controlling LREE and to a lesser extent Th contents for the critical temperature range of subduction zone liquid extraction. Low temperature fluids coexisting with allanite are not capable of transporting significant amounts of light rare earth elements and therefore arc lavas showing enrichment of these elements probably originate from a mantle source that experienced metasomatism by a hydrous granitic melt. D

Liquid immiscibility between carbonate and silicate melts is a possible mechanism for the genesis of carbonatites and associated silica-undersaturated rocks . Compelling textural evidence for liquid immiscibility is displayed by carbonate globules associated with silicate melts of melilititic compositions at numerous carbonatite provinces [2]. These textures include sharp curved menisci against a silicate melt, budding and coalescing of globules, and dumb-bell and amoeboid forms. One possibility is that unmixing occurs as magma evolves in the shallow mantle or crust . However, carbonate globules are sometimes found in deeply sourced mantle xenoliths and as inclusions in xenocrystic olivines with primitive compositions . They are therefore possible examples of near-primary carbonatitic melts derived at high pressure in the mantle. Experiments show that immiscibility can occur in partial melts of carbonated eclogite at 3 GPa [5]. To investigate the possibility of a primary origin for carbonatite-silicate melts at the carbonated peridotite solidus, we have conducted multi-anvil experiments in the system CMASK-CO 2 at 3 GPa. Model peridotite compositions in the CMAS system are prepared as glass, with carbonate added as a combination of K 2 CO 3 , CaCO 3 , and MgCO 3 . Samples are contained in welded platinum capsules. In our preliminary experiments at 1300 • C we have produced an assemblage of forsterite + garnet + clinopyroxene in equilibrium with two melts; one melt is carbonatitic (< ∼5wt% SiO 2 ) and the other approximates a kimberlitic composition (> 20wt% SiO 2 ). We are currently attempting to saturate these experiments with orthopyroxene and determine phase relations at the solidus model mantle lherzolite.

To extend our knowledge of partitioning of hydrogen between nominally anhydrous mantle minerals and basaltic melts to higher pressures, we conducted 8 experiments at 3–5 GPa and 1350–1440 °C on a modified mid-ocean ridge basalt (MORB) + 2 wt.% H2O. H contents in olivine, orthopyroxene, clinopyroxene, and garnet were measured by counting 16O1H ions with low-blank SIMS, and total H2O concentrations in coexisting glasses were measured by FTIR, SIMS, and electron microprobe. The glasses have water contents ranging from 4.1 to 8.9 wt.%. Eleven measured pyroxenes have H2O concentrations ranging from 430 to 1430 ppm, two garnets have concentrations of 73 and 260 ppm, and one olivine has a concentration of 29 ppm. Mineral/melt hydrogen partition coefficients are as follows: DHcpx/melt ranges from 0.014–0.021, DHopx/melt ranges from 0.009–0.019, DHgt/melt ranges from 0.001–0.0033, and DHol/melt has a value of 0.0006. Mineral/mineral hydrogen partitioning data are as follows: DHcpx/opxranges from 1.2–2, DHcpx/gt has values of 5.5 and 14, and a single DHcpx/ol has a value of 27. DHpyx/melt correlates with the concentration of Al2O3 in pyroxene and with the concentration of Al on tetrahedral sites. Combining the mineral/melt hydrogen partitioning data from this study, and estimates of mineral modes and compositions along the solidus, we estimate that DHperid/melt increases from 0.006 at 1 GPa up to 0.01 near the garnet lherzolite/spinel lherzolite transition, owing to the maximum in Al2O3 concentration in pyroxenes. With increasing pressure DHperid/melt for garnet lherzolite diminishes to 0.005 at 5 GPa, as pyroxenes become less aluminous and less abundant. As a result, incipient hydrous melts beneath oceanic ridges with bulk mantle H2O contents of 50–200 ppm will have solidus partial melts with H2O contents between 0.5 and 3.8 wt.%, while oceanic island mantle with 300–1000 ppm will have solidus partial melts with H2O content between 3 and 20 wt.%.

Cancrinite-group minerals are framework aluminosilicates (feldspathoids) characterized by a network of (Al,Si)O 4 tetrahedra having fully ordered Al and Si cations. Their structure consists of parallel six-membered rings of alternating AlO 4 and SiO 4 tetrahedra . The hexagonal symmetry (P6 3 ) results from the stacking of these six-membered rings in an ABAB... sequence. This stacking gives rise to large continuous channels parallel to the 6 3 axis and chains of smaller undecahedral (elevenfold) cages along the threefold axes. The large channels contain extra-framework cations (Na + , Ca 2+ , K + ), anions (CO 3 2-, OH -), and H 2 O molecules whereas the undecahedral cages hold [Na•H 2 O] + clusters Hassan and Busseck 1992). Phoenix and Nuffield (1949) defined the cancrinite group of minerals as (Na,K,Ca,Al) 6-8 [Si 6 Al 6 O 24 ](SO 4 ,CO 3 ) 1-2 •1-5H 2 O, and proposed an ideal formula for the carbonate end-member of cancrinite as Na 6 Ca 2 [Si 6 Al 6 O 24 ](CO 3 ) 2 •3H 2 O. They recognized the difficulties of dealing with the chemistry of this mineral, due to numerous possible substitutions, vacant positions, and variable zeolitic water content. Given the role that the carbonate salts play in the chemistry of cancrinite, Eitel (1922),

Melting experiments carried out at 1-atm and at 2 kbar on mid-ocean ridge basalts dredged from the mid-Atlantic ridge near the Kane Fracture Zone (KFZ, 22 ~ to 25 ~ N. latitude) provide a basis for evaluating the role of crystal fractionation in generating compositional variability observed in "normal" mid-ocean ridge basalt. The 1-atm olivine-plagioclase-clinopyroxene saturation boundary for KFZ lavas defines a path in mineral projection schemes and in oxide-oxide diagrams that is displaced from the same experimentally determined boundaries in FAMOUS (Grove and Bryan 1983) and Oceanographer Fracture Zone (Walker et al. 1979) basalts. The glass margins of sparsely phyric KFZ lavas record small amounts of near surface, low pressure fractional crystallization, and their glass and bulk rock compositions are similar. An important signature of low pressure differentiation is recorded in the quenched glass margins of moderately phyric KFZ lavas compared to their bulk rock compositions, and the glass has evolved along low-pressure fractionation paths that are similar to those produced in the 1-arm experiments. Many of the lavas have retained phenocrysts in equilibrium proportions, so that their bulk rock compositions represent liquid compositions. When the effects of near-surface differentiation and crystal accumulation are removed from the Kane data set, and only liquid compositions are considered, a suite of basalt magmas can be identified that forms a trend in mineral component projection schemes parallel to the 1-atm olivplag-cpx multiple saturation boundary, but displaced from it toward olivine. These basalts have only olivine and plagioclase as phenocrysts, and are well removed from clinopyroxene saturation at low pressure, The compositional variation can not be generated by mixing any primary liquid composition with a low pressure liquid that has evolved along the oliv-plag-cpx multiple saturation boundary. Major and trace element models of this trend using olivine, plagioclase and clinopyroxene as fractionating phases match the compositional variability. This compositional trend is generated by fractionation at pressures greater than 2 kbar, but within the plagioclase stability field. A review of the data for other normal MORB suites from this part of the mid-Atlantic ridge reveals a similar elevated pressure fractionation signature which persists when the effects of low pressure magma mixing are removed from the data set.

The 238U/204Pb ratio of the Earth's mantle is conventionally seen as reflecting the loss of lead to the Earth's core, implying that this element is siderophile. A contrasting view holds that lead may be depleted because it did not fully condense at the accretion temperatures of the early Earth. Experiments reported here show that lead is indeed quite siderophile at high temperature. At the same time, it is noted that no plausible set of physical conditions exists where the present-day abundances of all moderately volatile elements of the mantle could have been in chemical equilibrium with a metallic iron melt. It is argued that a large proportion of the moderately volatile elements, including Pb, were added late, during the waning stages of core formation. In consequence, caution is advised in using the 238U/204Pb and the Pb isotope ratios of the mantle to date the differentiation of the Earth in silicate mantle and metallic core (Hofmann, 2001).

Advances in field observations and experimental petrology on anatectic products have motivated us to investigate the geochemical consequences of accessory mineral dissolution and nonmodal partial melting processes. Incorporation of apatite and monazite dissolution into a muscovite dehydration melting model allows us to examine the coupling of the Rb-Sr and Sm-Nd isotope systems in anatectic melts from a muscovite-bearing metasedimentary source. Modeling results show that (1) the Sm/Nd ratios and Nd isotopic compositions of the melts depend on the amount of apatite and monazite dissolved into the melt, and (2) the relative proportion of micas (muscovite and biotite) and feldspars (plagioclase and K-feldspar) that enter the melt is a key parameter determining the Rb/Sr and 87 Sr/ 86 Sr ratios of the melt. Furthermore, these two factors are not, in practice, independent. In general, nonmodal partial melting of a pelitic source results in melts following one of two paths in Nd -87 Sr/ 86 Sr ratio space. A higher temperature, fluid-absent path (Path 1) represents those partial melting reactions in which muscovite/biotite dehydration and apatite but not monazite dissolution play a significant role; the melt will have elevated Rb/Sr, 87 Sr/ 86 Sr, Sm/Nd, and Nd values. In contrast, a lower temperature, fluid-fluxed path (Path 2) represents those partial melting reactions in which muscovite/biotite dehydration plays an insignificant role and apatite but not monazite stays in the residue; the melt will have lower Rb/Sr, 87 Sr/ 86 Sr, Sm/Nd, and Nd values than its source. The master variables controlling both accessory phase dissolution (and hence the Sm-Nd system), and melting reaction (and hence the Rb-Sr systematics) are temperature and water content. The complexity in Sr-Nd isotope systematics in metasediment-derived melts, as suggested in this study, will help us to better understand the petrogenesis for those granitic plutons that have a significant crustal source component.

A model metasomatized lherzolite composition contains phlogopite and pargasite, together with olivine, orthopyroxene, clinopyroxene and spinel or garnet as subsolidus phases to 3 GPa. Previous works established that at z 1.5 GPa, phlogopite is stable above the dehydration solidus, determined by the melting behaviour of pargasite and coexisting phases. At 2.8 GPa, melts with residual phlogopite + garnet lherzolite mineralogy at 1195 jC and with garnet lherzolite mineralogy at 1250 jC are both olivine nephelinite with K/Na (atomic) = 0.51 and K/Na = 0.65, respectively. Recent work shows that melting along the dehydration (fluid-absent) solidus of the phlogopite + pargasite lherzolite at pressures < 1.5 GPa is very different with the presence of phlogopite, decreasing the solidus below that of pargasite lherzolite. At 1.0 GPa, both phlogopite and pargasite disappear at temperatures at or slightly above the solidus. The compositions of two melts at 1.0 GPa, 1075 jC (with different water contents), in equilibrium with residual spinel lherzolite mineralogy are silica-saturated trachyandesite ( f 5% melt fraction, f 3% H 2 O) to silica-oversaturated basaltic andesite ( f 8% melt fraction, 4.5% H 2 O). Both compositions may be classified as 'shoshonites' on the basis of normative compositions, silica-saturation, and K/Na ratio. Decompression melting of metasomatized lithospheric lherzolite with minor phlogopite and pargasite may produce primary 'shoshonitic' magmas by dehydration melting at f 1 GPa, 1050 -1150 jC. Such magmas may be parental to Proterozoic batholithic syenites occurring in Brazil. D

Forsterite crystals were synthesised under water saturated conditions at 1400°C and 1.5 GPa doped with trace amounts of either B, Al, Sc, Ti, V, Cr, Mn, Fe, Co, Ga, Y, Zr, In, Sm, Gd, Dy, Tm, or Lu. The common and intense hydroxyl stretching bands in the infrared spectra of spinel peridotite olivine, at 3572 and 3525 cm −1 , were only reproduced in the presence of Ti. Those samples where the trace element substitutes as the trivalent cation on the Mg 2+ site were identified from a systematic variation in concentration with the trivalent ionic radius. The hydroxyl region of all samples is essentially identical except for between 3300 and 3400 cm −1 . This region is characterised by one or more bands, with the energy of the most intense feature being correlated with the ionic radius of the trivalent cation. The integrated intensity of these hydroxyl bands also correlates with the concentration of the trivalent cation. These correlations provide unambiguous evidence that bands, or peaks, in this region correspond to water at defect sites associated with trivalent cations. "Trivalent peaks" are sometimes observed in samples of mantle olivine and most likely indicate water associated with Fe 3+ . The water at this site is not incorporated under normal mantle conditions and should not be included in estimates of the water capacity of mantle olivine. These results emphasise the importance of identifying the infrared signature of different water substitution mechanisms.

Apatite/aqueous fluid partition coefficients have been measured for Ce, Gd, Yb and Sr at 1.0 GPa and 1000°C using radiotracers and gamma-ray spectroscopy. Analysis of clean apatite separates from 3-5-day piston cylinder experiments allowed calculation of KD'S from mass balance. Partition coefficients from forward experiments with doped fluids and reversal experiments with doped crystals show good agreement. Partition coefficients for apatite/H20 range from 15 to 33 in the order Gd > Ce > Sr> Yb, corresponding to a convex.up REE pattern characteristic of natural apatites. If Sr serves as an analog of Eu 2+, then Eu 2+ should behave similarly to REE and any (-) Eu anomalies in fluid precipitated/equilibrated apatites must result from inheritance rather than from apatite fractionation. The elements Be and Cs partition strongly into the fluid, so that low levels in clean apatite separates permitted only estimates of maximum Ki>'s. Addition of dissolved albite to the fluid decreased KD'S of REE and Sr 2-3 ×, implying complexing of REE + Sr with silicates dissolved in the fluid. Apatite/H20 KD'S are similar to apatite/silicate melt KD'S, except that in the latter case Eu (Sr) is less compatible than other REE. At 1.0 GPa and 1000 °, the high KD-values for REE and Sr combined with the insolubility of apatite in near-neutral pH aqueous fluids suggest that apatite may strongly influence the behavior of these elements during metaso-matism by an aqueous fluid. Mass-balance modeling using measured partition coefficients and solubilities demonstrate this, and also show that compatible elements do not reach equilibrium with the primary fluid until the mass ratio fluid/ rock z_ 1. Thus, caution is required when inferring compositions of primary fluids from solid mineral assemblages.

The partitioning of Pt in sulphide melt (matte) has been studied as a function of fS 2 and fO 2 at 1200 and 1300°C. The results show that the solubility of Pt in mattes increases strongly with increasing fS 2 and decreases weakly with increasing fO 2 . The increase in Pt solubility with increasing fS 2 is attributed to Pt dissolving in the melt as a sulphide species and the weak inverse dependence of Pt solubility on fO 2 to the diluting effect of increasing O in the melt at high fO 2 . These results, coupled with measurements of Pt solubility in silicate melts taken from the literature, allow the calculation of Pt matte/silicate-melt partition coefficients (D matte-sil Pt ) for a range of conditions pertinent to the formation of Pt-rich horizons in layered intrusions. The calculated values range between 10 7 and 10 11 , depending on fO 2 and fS 2 , several orders of magnitude higher than previously published values. Our preferred value for D matte-sil Pt for conditions appropriate to the Merensky Reef is 10 7 and for the Stillwater Pt-rich horizon 10 8 . The new results are consistent with the magmatic hypothesis for Pt-rich horizons in layered intrusions.

This experimental study examines relationships between alternative evolution paths of basaltic liquids (the so-called Bowen and Fenner trends), and silicate liquid immiscibility. Synthetic analogues of natural immiscible systems exhibited in volcanic glasses and melt inclusions were used as starting mixtures. Conventional quench experiments in 1atm gas mixing furnaces proved unable to reproduce unmixing of ferrobasaltic melts, yielding instead either turbid, opalescent glasses, or crystallization of tridymite and pyroxenes. In contrast, experiments involving in situ high-temperature centrifugation at 1000g (g ¼ 9Á8 m/s 2 ) did yield macroscopic unmixing and phase separation. Centrifugation for 3^4 h was insufficient to complete phase segregation, and resulted in sub-micron immiscible emulsions in quenched glasses. For a model liquid composition of the Middle Zone of the Skaergaard intrusion at super-liquidus temperatures of 1110^11208C, centrifugation produced a thin, silicic layer (64Á5 wt% SiO 2 and 7Á4 wt% FeO) at the top of the main Fe-rich glass (46 wt% SiO 2 and 21 wt% FeO). The divergent compositions at the top and bottom were shown in a series of static runs to crystallize very similar crystal assemblages of plagioclase, pyroxene, olivine, and Fe^Ti oxides. We infer from these results that unmixing of complex aluminosilicate liquids may be seriously kinetically hampered (presumably by a nucleation barrier), and thus conventional static experiments may not correctly reproduce it. In the light of our centrifuge experiments, immiscibility in the Skaergaard intrusion could have started already at the transition from the Lower to the Middle Zone. Thus, magma unmixing might be an important factor in the development of the Fe-enrichment trend documented in the cumulates of the Skaergaard Layered Series.

This chapter summarizes the arguments for and against large-scale chemical differentiation of the mantle and examines the variations in mineral chemistry that need to be accounted for when placing constraints on the compositional and thermal structure of the lower mantle using seismic velocity observations. Various lines of evidence have been used to argue that the lower mantle is chemically distinct in major elements from the upper mantle. One of the few approaches that may clarify this would be to compare observed seismic velocities with estimates computed from mineral physical models. Building a realistic model for the effects of composition on the seismic velocities of lower mantle minerals remains a key goal in mineral physics. The dominant lower mantle minerals are all solid solutions. An important aspect of measured bridgmanite Fe3+ contents is that they appear to be nominally independent of oxygen fugacity.

We have experimentally determined the solidus of an alkali-bearing carbonated peridotite (with 5 wt.% CO 2 ) between 10 and 20 GPa. Based on K-deficit in all low-temperature runs we assumed that some melt could be present in the low temperature runs and the true solidus of an alkali-bearing carbonated peridotite is placed below 1200°C. However, based on the disappearance of magnesite and the appearance of the visible quenched melt coexisting with silicate phases, the 'apparent' solidus, which may be applicable for peridotite with low alkali contents, was identified. The 'apparent' solidus temperature increases from~1380°C at 10 GPa to~1525°C at 15 GPa and the 'apparent' solidus curve becomes almost flat from 15 GPa to 20 GPa, where it is located near 1550°C. At 10 GPa, the 'apparent' solidus of carbonated peridotite is~550°C lower than the solidus of CO 2 -free natural anhydrous peridotite. The solidus of the present study was also~120°C lower than the solidus determined by Melting in the Earth's deep upper mantle caused by carbon dioxide. Nature, 440, 659-662.] for natural carbonated peridotite. The drop in the solidus temperature is mainly due to the effect of alkalis (Na 2 O, K 2 O). The melt near the 'apparent' solidus has high CO 2 (N 40 wt.%) and contains b6.0 wt.% SiO 2 , b 0.30 wt.% Al 2 O 3 and b 0.25 wt.% TiO 2 . The composition of near-solidus partial melt is close to that observed at 6-10 GPa in the CMS-CO 2 and CMAS-CO 2 systems, and natural carbonated peridotite, with some variations in Ca/Mg-ratio. High alkali contents in measured and calculated partial melts are consistent with the compositions of deepseated fluids observed as inclusions in diamonds and may be consistent with the compositions of parental melt, reconstructed for natural magnesiocarbonatite. We have demonstrated that magnesiocarbonatite-like melt can be generated by partial melting of carbonated peridotite at pressure up to at least 20 GPa. The generation of calciocarbonatite and ferrocarbonatite is unlikely to be possible during melting of carbonated peridotite in the deep mantle.

We present experimental determinations of the influence of H2O on partial melting of garnet peridotite (+1.5, 2.5, and 5 wt. % added H2O) at 3.5 GPa and 1200–1450°C. Experiments produced complex polyphase regions of quenched melt and equilibrium partial melt compositions were reconstructed by combined EMP and LA-ICP-MS analyses. Mass balance-derived melt fractions (F) range from 0.18 to 0.33 and dissolved water contents range from 4.5 to 23.5 wt. %. One exceptional experiment quenched glass, allowing independent verification of H2O concentration by FTIR. The influence of H2O on melt production is quantified by the temperature difference required to achieve a given F under dry and wet conditions, ΔT, which is controlled by the H2O concentration in partial melts. Melts with 1.5, 5, 10, and 15 wt. % H2O yield ΔT values of 50, 150, 250, and 320°C, respectively, consistent with a cryoscopic parameterization that assumes 3 oxygens per mole of silicate melt. Based on this parameterization, we calculate that beneath oceanic ridges, peridotite H2O storage capacity increases from 0 to 240 ppm from 66 to 110 km depth. For H2O to be solely responsible for melting in the oceanic low velocity zone (LVZ) at least 5.7 wt. % H2O must be dissolved in the melt at 110 km, and considerably more (e.g., 15 wt.% at 220 km) is required for melting throughout the entire observed interval. The addition of H2O results in 3.5 GPa partial melts of garnet peridotite (normalized anhydrous) that are SiO2 and Al2O3 poor (43–50 and 9–11.5 wt. %, respectively), and MgO and CaO rich (18–27 and 7–12 wt. %, respectively) when compared to anhydrous analogues. These effects become highly pronounced deep in the upper mantle, and are opposite to the effect of H2O on melt compositions in the spinel stability field, potentially owing in part to OH− association with network modifying cations in high pressure, depolymerized melts and in part to low-temperature stabilization of garnet, which enhances CaO/Al2O3.

This experimental study examines relationships between alternative evolution paths of basaltic liquids (the so-called Bowen and Fenner trends), and silicate liquid immiscibility. Synthetic analogues of natural immiscible systems exhibited in volcanic glasses and melt inclusions were used as starting mixtures. Conventional quench experiments in 1atm gas mixing furnaces proved unable to reproduce unmixing of ferrobasaltic melts, yielding instead either turbid, opalescent glasses, or crystallization of tridymite and pyroxenes. In contrast, experiments involving in situ high-temperature centrifugation at 1000g (g ¼ 9Á8 m/s 2 ) did yield macroscopic unmixing and phase separation. Centrifugation for 3^4 h was insufficient to complete phase segregation, and resulted in sub-micron immiscible emulsions in quenched glasses. For a model liquid composition of the Middle Zone of the Skaergaard intrusion at super-liquidus temperatures of 1110^11208C, centrifugation produced a thin, silicic layer (64Á5 wt% SiO 2 and 7Á4 wt% FeO) at the top of the main Fe-rich glass (46 wt% SiO 2 and 21 wt% FeO). The divergent compositions at the top and bottom were shown in a series of static runs to crystallize very similar crystal assemblages of plagioclase, pyroxene, olivine, and Fe^Ti oxides. We infer from these results that unmixing of complex aluminosilicate liquids may be seriously kinetically hampered (presumably by a nucleation barrier), and thus conventional static experiments may not correctly reproduce it. In the light of our centrifuge experiments, immiscibility in the Skaergaard intrusion could have started already at the transition from the Lower to the Middle Zone. Thus, magma unmixing might be an important factor in the development of the Fe-enrichment trend documented in the cumulates of the Skaergaard Layered Series.

The determination of the exact protolith from which every leucogranite is derived may be a hard task. The trace element composition of both the leucogranite and their likely source rock can be a way to establish this relationship. This hypothesis is checked in an experimental study on the partial melting of two Iberian orthogneisses, the Ollo de Sapo and the Tormes Dome Gneisses. Several runs between 700 and 900°C at 0.6 and 1.0 GPa, with added water ranging from 2% wt. to 10% wt. have been carried out in piston-cylinder devices. Their results show that two gneisses of virtually the same major and trace element composition yield partial melts of closely similar major element composition but extremely different trace element composition, above all REE. These differences are produced because of several factors, mainly the ASI value of melts (which in turn depends on minor differences in the composition in the main oxides of the gneisses) and the phosphorus hosted in the feldspars of the gneisses. The consequence of these minor changes is that the dissolution of the trace-element-bearing accessories, mainly apatite, can be delayed until higher temperatures are achieved, so melts formed under these circumstances are depleted in REE, Th and Y. Gneisses of relatively low aluminosity and phosphorus-rich feldspars will produce REE-poor melts saturated in P 2 O 5 at low temperatures because neither apatite nor other phosphates dissolve. In contrast, gneisses of relatively high aluminosity and phosphorus-poor feldspars will produce REE-rich melts subsaturated in P 2 O 5 at low temperatures. For this reason, using the trace element composition of peraluminous granites as the only way to establish their protoliths is not recommended.

The origin of subduction-related magmas is still a matter of debate in the Earth Sciences. These magmas are characterised by their distinctive trace element compositions compared to magmas from other tectonic settings, e.g. mid-ocean ridges or rifts. The distinct trace element composition of these magmas is generally attributed to alteration of the source region by a contaminating agent: either a silicate melt or a hydrous fluid, possibly chlorine-enriched.In this study, we have used μPIXE (proton induced X-ray emission) to analyse synthetic samples obtained from a micro-experimental petrology study that aims to determine the partitioning behaviour of two key elements, Ba and Pb, between silicate melt and both pure water and saline fluids. Our experiments were performed at high-pressure (>0.34–1.53 GPa) and high-temperature (697–1082 °C) in a hydrothermal diamond anvil cell, that was used as a transparent rapid quench autoclave. We observed that at high pressure and temperature, in the presence of pure water, Ba and Pb are not strongly fractionated into one phase or the other. The partition coefficient of Pb is ranging from 0.46 to 1.28. Results from one experiment performed at 0.83 GPa and 847 °C, in the presence of a saline fluid indicate that the presence of Cl induces strong fractionation of Pb and moderate fractionation of Ba both into the silicate melt. In addition, our data indicate that Cl is strongly partitioned into the fluid phase.

We have determined the near-solidus melt compositions for peridotite MM-3, a suitable composition for the production of mid-ocean ridge basalt (MORB) by decompression partial melting, at 1 and 1Á5 GPa. At 1 GPa the MM-3 composition has a subsolidus plagioclase-bearing spinel lherzolite assemblage, and a solidus at $ 12708C. At only $ 58C above the solidus, 4% melt is present as a result of almost complete melting of plagioclase. This melting behaviour in plagioclase lherzolite is predicted from simple systems and previous experimental work. The persistence of plagioclase to 40Á8 GPa is strongly dependent on bulk-rock CaO/Na 2 O and normative plagioclase content in the peridotite. At 1Á5 GPa the MM-3 composition has a subsolidus spinel lherzolite assemblage, and a solidus at $ 13508C. We have determined a near-solidus melt composition at $ 2% melting within 108C of the solidus. Near-solidus melts at both 1 and 1Á5 GPa are nepheline normative, and have low normative diopside contents; also they have the highestTiO 2 , Al 2 O 3 and Na 2 O, and the lowest FeO and Cr 2 O 3 contents compared with higher degree partial melts. Comparison of these near-solidus melts with primitive MORB glasses, which lie in the olivine-only field of crystallization at low pressure, indicate that petrogenetic models involving aggregation of near-fractional melts formed during melting at pressures of 1Á5 GPa or less are unlikely to be correct. In this study we use an experimental approach that utilizes sintered oxide mix starting materials and peridotite reaction experiments. We also examine some recent studies using an alternative approach of melt migration into, and entrapment within 'melt traps' (olivine, diamond, vitreous carbon) and discuss optimal procedures for this method.

alkaline melts, modelled on glasses found in many mantle xenoliths, The compositions of silicate melts potentially in equishow that part of this compositional range is saturated with librium with the Earth's mantle have long been of interest harzburgite (or possibly lherzolite) under anhydrous conditions. to petrologists and geochemists. Much useful information Under C-O-H fluid-saturated conditions with X H2O =0·5, phloabout such melts has come from the study of melt gopite mica is present along with anhydrous phases similar to those inclusions in phenocrysts in volcanic rocks and in the found under dry conditions. Phlogopite is the sole liquidus phase minerals of ultramafic mantle xenoliths. In addition, when X H2O =1·0. At X H2O =0·5 and 3·0 GPa, garnet, kyanite many mantle xenoliths contain silicate glasses as a discrete and carbonate minerals appear as near-liquidus phases and the phase, rather than (or in addition to) inclusions (see shape of the liquidus surface is reminiscent of that of the carbonated Appendix A for references). In this paper, we will refer peridotite solidus. Saturation of these liquids with harzburgite, and to discrete-phase xenolith glasses simply as xenolith comparisons with calculated melt silica activities, suggests that these glasses and those found as inclusions in minerals as liquids would face no chemical or thermal obstacles to circulating inclusion glasses. Xenolith glasses have been reported amongst and coexisting with harzburgitic mantle. Also, there is from mantle xenoliths hosted by alkaline, mafic magmas textural evidence that these melts may be mobile. Accordingly, these both from intra-plate and from subduction-related teckinds of liquids could act as cryptic metasomatic agents. If mantle tonic settings. The glasses have a wide range of com-at~45-90 km depth is pre-enriched in low-melting-temperature positions and textural relations: they are found as components, and probably volatiles, via the ascent and percolation intergranular blebs, irregular patches, veins, and thin of alkaline, mafic liquids (along geotherms that cross inflections in films that appear to wet all grain faces. Most glasses the solidus of CO 2 -bearing peridotite), then subsequent low-degree account for, at most, only a few volume percent of their partial melts could yield the liquids that are ultimately trapped as host xenolith, and some much less than 1 vol. %. xenolith glasses.

HAL is a multidisciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L'archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d'enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

The distribution of heat-producing elements (HPE) potassium (K), uranium (U), and thorium (Th) within planetary interiors has major implications for the thermal evolution of the terrestrial planets and for the inventory of volatile elements in the inner solar system. To investigate the abundances of HPE in Mercury’s interior, we conducted experiments at high pressure and temperature (up to 5 GPa and 1900 °C) and reduced conditions (IW-1.8 to IW-6.5) to determine U, Th, and K partitioning between metal, silicate, and sulfide (Dmet/sil and Dsulf/sil). Our experimental data combined with those from the literature show that partitioning into sulfide is more efficient than into metal and that partitioning is enhanced with decreasing FeO and increasing O contents of the silicate and sulfide melts, respectively. Also, at low oxygen fugacity (log fO2 &lt; IW-5), U and Th are more efficiently partitioned into liquid iron metal and sulfide than K. Dmet/sil for U, Th, and K increases with decr...

Crystallization experiments were conducted on dry glasses from the Unzen 1992 dacite at 100-300 MPa, 775-875 C, various water activities, and fO 2 buffered by the Ni-NiO buffer. The compositions of the experimental products and natural phases are used to constrain the temperature and water contents of the lowtemperature and high-temperature magmas prior to the magma mixing event leading to the 1991-1995 eruption. A temperature of 1050 AE 75 C is determined for the high-temperature magma based on two-pyroxene thermometry. The investigation of glass inclusions suggests that the water content of the rhyolitic lowtemperature magma could be as high as 8 wt % H 2 O. The phase relations at 300 MPa and in the temperature range 870-900 C, which are conditions assumed to be representative of the main magma chamber after mixing, show that the main phenocrysts (orthopyroxene, plagioclase, hornblende) coexist only at reduced water activity; the water content of the post-mixing dacitic melt is estimated to be 6 AE 1 wt % H 2 O. Quartz and biotite, also present as phenocrysts in the dacite, are observed only at low temperature (below 800-775 C). It is concluded that the erupted dacitic magma resulted from the mixing of c. 35 wt % of an almost aphyric pyroxenebearing andesitic magma (1050 AE 75 C; 4 AE 1 wt % H 2 O in the melt) with 65 wt % of a phenocryst-rich low-temperature magma (760-780 C) in which the melt phase was rhyolitic, containing up to 8 AE 1 wt % H 2 O. The proportions of rhyolitic melt and phenocrysts in the low-temperature magma are estimated to be 65% and 35%, respectively. It is emphasized that the strong variations of phenocryst compositions, especially plagioclase, can be explained only if there were variations of temperature and/or water activity (in time and/or space) in the low-temperature magma.

Two-layer experiments where a hydrous mafic glass is attached to a powdered tonalite have been performed at conditions of 850 to 1000 8C and 6 to 10 kbar. The aim of these experiments is to simulate the effects of fluids released from a crystallizing hydrous magma emplaced within a tonalitic continental crust. This process has been inferred from field observations and geochemical relationships in Archaean cratons where the first K-rich granite batholiths were developed. The results of these experiments indicate that conditions of 950 8C and 6 kbar, about 50 vol.% of granodiorite melt is generated in the tonalitic crust. The process implies the transfer of H 2 O, originally dissolved into the mafic melt, together with potassium to the adjacent tonalite. It is proposed that the Archaean granodiorites (of tonalite-trondhjemite-granodiorite (TTG) complexes) result of recycling of older tonalitic crust, and that their generation required the intrusion of mantle-derived mafic magmas. These mafic magmas are sanukitoid in composition and their water content is higher than the water coordinated by hydrous phases such as biotite and amphibole. A simple model to determine the fertility of granitic crust to produce new granite batholiths based on the water content of the hydrous mafic magmas emplaced into the continental crust is presented. This model indicates that for conditions prevailing in the lower and middle crust, the volume of granitic melt may be twice the volume of basic magma. D

Compositions of primary basalt magmas were backcalculated from Miocene Mg-rich basalts in the Japan Sea back arc basin and the NE Japanese volcanic front, in order to assess the spatial variation in physicochemical conditions for magma production during opening of the back arc basin. Chemical compositions of these magmas together with available liquidus experimental data suggest that the primary basalt magmas both in the Japan Sea (BABB magma) and the NE Japanese volcanic front (VFB magma) were equilibrated with harzburgitic mantle materials at almost identical pressure and temperature conditions. The observation that the BABB magma is more depleted in Fe and some incompatible trace elements and more enriched in A1 than the VFB magma then suggests that these magmas may have been derived from source mantle materials with different chemical compositions. An equilibrium melting model further demonstrates that the BABB magma was produced by a smaller degree of partial melting of a more depleted source mantle material (compositional equivalent to abyssal harzburgites) than the VFB magma. Provided that the varying degrees of depletion in the source mantle materials were manifestations of mantle upwelling efficient beneath the back arc side region, the more depleted BABB magma source may have been a plausible geochemical reservoir in the deeper portion of the mantle wedge.

in liquidus temperature as water migrates from the mafic to the felsic magma. We refer to this process as " chemical quenching " and suggest that some textures associated with natural mafic–felsic interactions are not simply cooling-driven in origin, but can be caused by exsolution of vola-tiles adjacent to an interface, whether a water-undersatu-rated felsic magma (as in our experiments) or a fracture.

Northwest Africa (NWA) 1068 is one of the few olivine-phyric shergottites (e.g., NWA 1068, Larkman Nunatak [LAR] 06319, and Roberts Massif [RBT] 04262) that is not depleted in light rare earth elements (LREE). Its REE pattern is similar to that of the basaltic shergottite Shergotty, suggesting a possible connection between the olivine-phyric and the basaltic shergottites. To test this possible

1] In the Wadi Haymiliyah of the Oman ophiolite (Haylayn block), discordant wehrlite bodies ranging in size from tens to hundreds of meters intrude the lower crust at different levels. We combined investigations on natural wehrlites from the Wadi Haymiliyah section with an experimental study on the phase relations in a wehrlitic system in order to constrain the petrogenesis of the crustal wehrlites of the Oman ophiolite. Secondary ion mass spectrometry analyses of clinopyroxenes from different wehrlite bodies imply that the clinopyroxenes were crystallized from tholeiitic, mid-ocean ridge (MORB)-type melts. The presence of primary magmatic amphiboles in some wehrlites suggests a formation under hydrous conditions. Significantly enhanced 87 Sr/ 86 Sr isotope ratios of separates from these amphiboles imply that the source of the corresponding magmatic fluids was either seawater or subduction zone-related. The experiments revealed that under wet conditions at relatively low temperatures, a MORB magma has the potential to produce wehrlite in the ocean crust by accumulation of early olivine and clinopyroxene. These show typically high Mg# which is a consequence of the oxidizing effect of the prevailing high aH 2 O. First plagioclases crystallizing after clinopyroxene under wet conditions are high in An content, in contrast to the corresponding dry system. Trace element compositions of clinopyroxenes of those wehrlites from the Moho transition zone are too depleted in HREE to be in equilibrium with present-day MORB, implying a genetic relation to the V2 lavas of the Oman ophiolite, which are interpreted to be the result of fluidenhanced melting of previously depleted mantle. We present a model on the petrogenesis of the crustal wehrlites in an upper mantle wedge above an initial, shallow subduction zone at the beginning of the intraoceanic thrusting.

Seismic anomalies in subducted slabs, including low velocity
zones and shear wave splitting, have often been related to
hydrous regions. Phase D (MgSi2H2O6, 10–18 wt.% H2O) may
be the ultimate water carrier in hydrous subducted peridotite
and its seismic properties are thus essential for interpreting
observed anomalies in terms of hydration. Here, we report the
sound velocities and elasticity of Mg- and Al-Fe-bearing phase
D single-crystals measured by Brillouin spectroscopy. The room
conditions adiabatic bulk and shear modulus are: KS0 = 154.8
(3.2) GPa and m = 104.3(2.1) GPa for Mg-phase D and
KS0 = 158.4(3.9) GPa and m = 104.7(2.7) GPa for AlFephase
D, suggesting minor effect of cation substitution on the
elasticity of phase D. Based on the seismic velocity data, we
found that 16 vol.% of AlFe-phase D in hydrous subducted
peridotite with 1.2 wt.% H2O could provide a plausible
explanation for the negative velocity anomalies of 3% observed
in fragments of the Tonga slab below the transition zone.