Search

A series of Secondary Ion Mass Spectrometry (SIMS) sessions to measure hydrogen (H) in Martian meteorite minerals was completed using the Cameca 6f SIMS and NanoSIMS 50L at Arizona State University (ASU). During these sessions, a “creeping” phenomenon has occurred, where the edges of samples pressed in indium are covered by a metal alloy. We summarize these observations herein, present a collection of preliminary data, and discuss explanations and concerns for future SIMS work. We conclude the report with a research plan for further study.

We want to define the H2O content ([H2O]) and hydrogen (H) isotope composition of meteoritic material from Mars [1-3] with motivation to understand Mars’ volatile history, constrain geochemical signatures of interior water reservoirs (i.e. the Martian mantle) and explore effects of planetary (e.g. planet formation, magma ocean degassing) and local (e.g. volcanic degassing, impact melting and degassing) processes on H incorporated in minerals. Secondary ion mass spectrometry (SIMS) allows multiple avenues to address these questions. However, application to (1) precious astromaterials and (2) low level H measurements, pose specific challenges that are further complicated when combined. We present preliminary data of a multi-approach (SIMS vs. NanoSIMS) study of H in Roberts Massif 04262 (RBT 04262), an enriched lherzolitic shergottite with nonpoikilitic (NP) and poikilitic (P) lithologies [4]. We analyze olivine, pyrox-ene, and melt inclusions to compare indigenous mantle water, with impact-generated maskelynite to investigate H signatures due to shock.

Meteorites from Mars lack field context but chemical and chronologic studies have revealed remarkable links between nakhlites and chassignites. A widely held consensus is that nakhlites and chassignites originated from a large, single differentiated flow or shallow intrusive [1-5]. An Ar-Ar study assumed multiple flows based on resolvable age differences between meteorites [6], but did not address the possibility of differential cooling in a large, shallowly emplaced intrusion [1]. REE abundances in pyroxenes from nakhlites and Chassigny led [7] to argue for derivation of these rocks from distinct magmas. Volatile abundances (F, Cl, OH) in chlorapatites indicated that the entire suite of nakhlites and chassignites experienced hydrothermal interaction with a single fluid supporting a single body origin [4]. The discovery of a new chassignite, NWA 8694, extended the Mg# range from 80-54, providing a closer link to nakhlites but revealed the petrological difficulty of fractionating a single body of liquid to yield a series of olivine cumulates with such a large Mg# range [8]. When mafic magmas are emplaced into the crust, crustal assimilation can impart distinct elemental signatures if the country rock has experienced sedimentary or hydrothermal processing [9]. In this work, we used Sn abundances of nakhlites and chassignites to show that these rocks were crystallized from distinct magma batches, providing vital contextual clues to their origin.

Oxygen fugacity and water content are crucial parameters for many chemical and physical properties of the Earth's mantle, for example bearing on fluid type, melting initiation, and deformation. However, the exact behaviour of Fe3+ and H during melting and metasomatism is still under debate. Here, the Fe3+/ΣFe ratio (Mössbauer and EMP) and water content (FTIR) of peridotite minerals are examined in mantle xenoliths from Kilbourne Hole (KH), NM, and Dish Hill (DH), CA (USA). These spinel peridotites have compositions consistent with partial melting with variable degrees of metasomatism (undetectable to cryptic to modal). Pyroxenites also allow to examine melt-rock reactions. Bulk-rock Fe2O3 content of the KH peridotites correlates with indices of melting (positive with bulk-rock Al2O3 and Cpx Yb contents, and negative with spinel Cr#) confirming that Fe3+ behaves as an incompatible element during melting. Correlations of the Fe3+/ΣFe ratio of minerals with these indices, however, indicates that Fe3+ is incompatible in Cpx but compatible in Opx and spinel during melting. Water contents in olivine, Cpx and Opx from most KH peridotites can be explained by partial melting and correlate negatively with the Fe3+/ΣFe ratio of spinel and Opx but positively with that of Cpx. This indicates partial control of Fe3+ on the incorporation of H in pyroxene, but not related to a redox equilibrium in Cpx. The higher Fe3+/ΣFe ratio of spinel in the metasomatized KH and DH peridotites, and in the pyroxenites confirms that oxidation characterizes modal metasomatism. Metasomatism, however, is not necessarily accompanied by water addition.

Water in Earth’s mantle affects processes like magmatism and plate tectonics. Experiments show that CO2-rich fluids lower the water solubility in olivine, implying that CO2-rich melts/fluids may dehydrate the lithosphere during metasomatism. To test this hypothesis, we report water concentrations (by polarized FTIR) of olivines, orthopyroxenes (OPX) and clinopyroxenes (CPX) from Savai’i (Samoa) and Lanzarote (Canary Islands) peridotite xenoliths with evidence of carbonatite metasomatism. Savai’i peridotites are highly depleted harzburgites and dunites with spinel Cr# (Cr/(Cr+Al)) ranging from 0.4 to 0.76 (estimated degree of melting: 191.5%). Strong Light Rare Earth Element (LREE) enrichments with Ti and Zr depletions in OPX and CO2-rich fluid inclusions (via Raman spectroscopy) are consistent with carbonatite metasomatism. Olivine, OPX and reconstructed bulk rock water concentrations (0.67-3.8, 17-89 and 4-26 ppm H2O, respectively) are low and show no apparent relationship with extent of carbonatite metasomatism. Calculated water concentrations of melts in equilibrium with Savai’i OPX (OPX/melt partitioning of water 0.0063 to 0.011) are, on average (0.540.32 wt% H2O), lower than host Samoan lavas (0.63 to 1.5 wt% H2O), despite the LREE enrichments in OPX. Lanzarote peridotites are also highly depleted (degree of melting from spinel Cr#: 171.8%).Water concentrations are low in olivines (1.7-5.3 ppm H2O) and variable in pyroxenes (OPX: 42-103 ppm H2O; CPX: 105-301 ppm H2O), and show no apparent correlation with indicators of carbonatite metasomatism. Both Savai’i and Lanzarote peridotites show negative correlations between water and degree of melting (i.e. Mg/(Mg+Fe), Cr#), suggesting melt depletion rather than metasomatism may have influenced their water concentrations. Calculated water concentrations of melts in equilibrium with Lanzarote CPX (average 1.90.75 wt% H2O; CPX/melt partitioning of water 0.011 to 0.012) are similar to those for Western Canaries lavas (average 1.80.31 wt%; CPX/melt partitioning of water 0.016 to 0.021) inferred from their CPX phenocrysts. However, calculated Ce concentrations in such melts (352 to 378 ppm; CPX/melt partitioning of Ce 0.07) are an order of magnitude greater than the lavas, and similar to carbonatites. This leads to H2O/Ce to be an order of magnitude lower in the inferred melts (26 to 57) than estimates for Western Canary lavas (280150). These low H2O/Ce ratios may suggest H2O loss from CPX during ascent, but the lack of strong water diffusion gradients in Lanzarote minerals does not support this. Instead we hypothesize that carbonatite metasomatism resulted in greater enrichment of Ce over H2O. Assuming carbonatite magmas are water rich, this implies a lower partitioning of water between minerals and melts during metasomatism, as suggested by experiments. Our data suggests carbonatite metasomatism does not result in significant re-hydration of the lithosphere, in contrast to silicate metasomatism as previously observed in Hawaiian peridotites.

Olivine-hosted melt inclusions (MIs) may retain trapped parent magma compositions as well as record progressive differentiation while magmas crystallize and ascend towards the surface [1,2 and references therein]. Major element compositions of the MIs, especially Fe and Mg, can be affected by post-entrapment re-equilibration with their host olivine [1,2]. Therefore, Fe/Mg ratio correction is required to obtain MI bulk compositions following equilibrium with their host olivine. Partition coefficients of most of the trace elements in olivine are very low (i.e. DOL/melt<0.001). Thus, ratios of trace elements of olivine-hosted MIs are unlikely to be affected by post-entrapment re-equilibration and no correction is necessary [2]. Hence, tracking trace element behavior in MIs may constrain the composition of the parent magma and its evolution yielding insights on magma differentiation of shergottites. However, analyzing MIs for chemical compositions is a challenging task due to their low abundances and small sizes. Using a highly sensitive and precise micro-beam technique is essential to examine olivine-hosted MIs in order to measure trace element abundances. For this purpose, laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) is an excellent tool due to its wide range of laser spot sizes (1-150 μm), ability to obtain raster analysis (several mm2) and lower detection limits (0.1 ppb) [3].

Impact cratering is an important geological process that occurs on every rocky body in the solar system. It alters the texture and mineralogy of rocks via shock metamorphism. The peak shock pressures experienced by a rock are traditionally evaluated using qualitative optical methods however, quantitative methods do exist. One such method was developed by Uchizono et al., who used X-ray Diffraction (XRD) to measure lattice strain (ε) in several artificially shocked olivine grains using XRD peak broadening as a function of tan θ, where θ is the diffraction angle. They plotted the ε values against the known peak shock pressures experienced by the olivine grains. Using this calibration curve, the precise shock pressure experienced by a grain of olivine can be determined using its measured ε value. Another method was developed by McCausland et al. and Izawa et al., who used in situ XRD to measure strain-related mosaicity (SRM) of olivine in several ordinary chondrites and enstatite in enstatite chondrites, respectively. They plotted these results against the shock stage estimates for these meteorites. Using these plots, meteorites can be assigned to shock stage bins by measuring the SRM of olivine and/or enstatite. Both methods are useful for evaluating shock metamorphism, however, they have limitations. Uchizono et al.’s calibration curve has been successfully applied to martian meteorites, however it can only be applied to olivine-bearing rocks. McCausland et al.’s and Izawa et al.’s SRM method is uncalibrated and is limited to binning meteorites by shock stage. This work aims to expand on both methods by creating calibration curves for clinopyroxene (CPX): one for ε, similar to Uchizono et al.’s calibration curve for olivine, and one for SRM. This will extend the application of shock calibration methods to a greater variety of rock types. Preliminary results are presented herein.

Hydrogen in nominally anhydrous minerals (NAMs) in meteorites provides insight to mantle sources of indigenous water on differentiated bodies: e.g. Peslier et al. 2017 [1], including Mars [2-4]. However, all meteorite samples, including Martian shergottites, record impact events as fractures, deformation, silicate darkening, shock melt veins and pockets, etc. The effect of shock on hydrogen in NAMs is poorly constrained, and must be understood prior to using these data to infer planetary indigenous water. Here we present water contents and D/H ratios (calculated as dD, i.e. the variation of the D/H ratio relative to a standard, in this case sea water "SMOW") in pyroxene, olivine and maskelynite in the olivine-phyric shergottite Larkman Nunatak 06319 (LAR 06319) as a function of proximity to impact melt. While the results suggest impact may have a role in fractionating H isotopes, the magmatic signature of H2O in Mars can be preserved in some pyroxene.

Water, in the form of structurally bound hydrogen in the crystal lattice of nominally anhydrous minerals (NAMs), strongly influences many important physical processes on terrestrial planets and planetary objects. Water enhances the rates of plastic deformation and controls the degree of partial melting in silicate rocks, which influences the generation of melt and therefore the nature of planetary volcanism. Water has also been experimentally demonstrated to influence the nature of lattice preferred orientation in deformed aggregates, and thus may be important in the interpretation of seismic anisotropy data collected from planetary bodies, such as from the current InSight mission on Mars. Therefore, much attention has been focused on characterizing the distribution and concentration of water in the planets and rocky bodies of our solar system.

Germanium (Ge) and Zn enrichment in martian sedimentary rocks has been reported from rocks at Gale Crater, showing concentrations of Ge from tens to hundreds ppm [1]. The Ge concentrations in martian meteorites are significantly lower (0.5-2.5 ppm) [2]. Our recent studies [3-4] have revealed that Ge is lost from shergottites due to volatility. Recent experimental studies confirm that Ge and Zn are both significantly volatile under magmatic conditions [5-7]. Further, Ge is moderately incompatible during magmatic differentiation [8] so Ge contents in olivines or pyroxenes increase during igneous fractionation in nakhlites and chassignites [4]. Shergottites for which Ge abundances had been determined included rocks with ages of 150-600 Ma, while the enrichments reported from Gale Crater rocks likely occurred over 3 Ga ago. The recent discovery of two unpaired ancient (2.4 Ga) depleted shergottites, NWA 7635 [9] and NWA 8159 [10], afforded the prospect of obtaining an extended history of martian volcanic outgassing. Both of the ancient shergottites are depleted in incompatible elements and share a similar GCR exposure age to younger depleted shergottites implying derivation from a single, long-lived (>2 Ga) volcanic center [9].

Knowing the distribution and origin of water in terrestrial planets is crucial to understand their formation, evolution and the source of their atmospheres and surface water. Mantle D/H ratios may be used to determine what type of material contributed water to the terrestrial planets [1]. However, other processes, magmatic or surface alteration processes, can also modify D/H ratios, and for Mars, we only have samples from the crust, as meteorites. The D/H ratio of igneous phases of Martian meteorites is generally explained in terms of the mixing contributions of two reservoirs: surficial with high D/H (dD > 700 ‰) related to interaction with the martian atmosphere (dD ~ 5000‰), and mantle-derived with lower D/H (dD < 500 ‰ but the exact value is still debated)[2]. However, our present study evidences that H loss in clinopyroxene during degassing can significantly fractionate H isotopes and increase their D/H ratios. In situ analyses of H isotopes, and of water, major and trace element contents were performed on the pyroxenes of 5 nakhlites. Nakhlites are clinopyroxenites that likely originated from the same lava flow or shallow magma chamber. Water contents decrease (380 to <5 ppm H2O) with increasing dD (-268 to 4860 ‰). Significant influence from spallation, exchange with the martian atmosphere, shock, surface alteration, and hydrothermal processes is ruled out. Together with the evidence of less water at the edge of individual pyroxene grains compared to their interior, we interpret this correlation as the result of preferential diffusive loss of H relative to D from the already crystallized pyroxenes during ascent of the partially-crystallized magma. Similar H isotope fractionations have been observed in another nominally anhydrous mineral, garnet, during experimental dehydration [3]. These results emphasize that caution is warranted when interpreting H isotope analyses of igneous, nominally anhydrous minerals in terms of planetary processes.

Final Paper and not the abstract is attached. Introduction: Germanium is generally thought to follow Si in its geochemical behavior, but little data has existed to rigorously understand the behavior of Ge in the Martian mantle. Germanium is known to be more siderophile than Si, and its partitioning into the martian core has been studied by. Typical abundances in igneous martian meteorites range from 0.5-3 ppm, a larger range than what is observed in terrestrial basalts (1.5 +/- 0.1 ppm). In situ measurements by the MER and MSL rovers have revealed a surprisingly large range in Ge abundances in surface rocks (30-650 ppm), but many igneous rocks and soils are shown to have greater than 30 ppm Ge (the detection limit of the APXS). Recently, reported that shergottite minerals showed a depletion of Ge with increasing fractionation, while nakhlites and chassignites exhibited Ge behavior compatible with closed system igneous differentiation. They interpreted their observations as tentatively indicating volcanic outgassing of Ge from shergottites (but not from nakhlite-chassignites) with recondensation of the Ge vapor into soils and breccias. Recent experimental studies show that Ge is significantly volatile from magmas. In this study, we followed up on those results by analyzing minerals in five nakhlites and five shergottites by laser ablation ICP-MS (LA-ICP-MS). See Attached

Water in the mantle influences melting, viscosity, seismic velocity, and electrical conductivity. The role played by water in the long-term stabilization of cratonic roots is currently being debated. This study focuses on water contents of mantle minerals (olivine, pyroxene and garnet) from xenoliths found in kimberlites of the Archean Slave craton. 19 mantle xenoliths from central Lac de Gras, and 10 from northern Jericho were analyzed by FTIR for water, and their equilibration depths span the several compositional layers identified beneath the region. At both locations, the shallow peridotites have lower water contents in their olivines (11-30 ppm H2O) than those from the deeper layers (28-300 ppm H2O). The driest olivines, however, are not at the base of the cratonic lithosphere (>6 GPa) as in the Kaapvaal craton. Instead, the deepest olivines are hydrous (31-72 ppm H2O at Lac de Gras and 275 ppm H2O at Jericho). Correlations of water in clinopyroxene and garnet with their other trace element contents are consistent with water being added by metasomatism by melts resembling kimberlite precursors in the mantle approx.0.35 Ga ago beneath Lac de Gras. The northern Jericho xenoliths are derived from a region of the Slave craton that is even more chemically stratified, and was affected at depth by the 1.27 Ga Mackenzie igneous events. Metasomatism at Jericho may be responsible for the particularly high olivine water contents (up to 300 ppm H2O) compared to those at Lac de Gras, which will be investigated by acquiring trace-element data on these xenoliths. These data indicate that several episodes of metasomatic rehydration occurred in the deep part of the Slave craton mantle lithosphere, with the process being more intense in the northern part beneath Jericho, likely related to a translithospheric suture serving as a channel to introduce fluids and/or melts in the northern region. Consequently, rehydration of the lithosphere does not necessarily cause cratonic root delamination and these peridotites may represent localized metasomatic zones - the wall rocks to kimberlite magma passage.

The origin of the incompatible trace element (ITE) enriched compositions of shergottites has been a point of contention for decades [1-2]. Two scenarios have been proposed, the first is that enriched shergottite compositions reflect an ITE-enriched mantle source, whereas in the second, the ITE enrichment reflects crustal contamination of mantle-derived parent magmas. Evidence supporting the first scenario is that the ITE-enriched shergottite compositions are consistent with the outcomes of magma ocean crystallization [3], and that Os-Nd isotope relationships for shergottites cannot be explained by realistic crustal contamination models [4]. In contrast, Cl and S isotopes are consistent with shergottite magmas interacting with Mars crust [5,6], and ITE-enriched olivine-hosted melt inclusions and interstitial glass are found in depleted shergottite Yamato 980459 [7]. These findings indicate that some level of crustal interaction occurred but the question of whether ITE-enrichments in some bulk shergottites reflect crustal contamination remains open. Recently, a Mars crustal breccia meteorite has been found, NWA 7034 and its paired stones, that is our best analogue to an average of Mars ancient crust [8-10]. This allows for better constraints on crustal contamination of shergottite magmas. We modeled magma-crust mixing and assimilation-fractional crystallization (AFC) using ITE-depleted shergottite compositions and bulk NWA 7034 and its clasts as end-members. The results of these models indicate that crustal contamination can only explain the ITE-enriched compositions of some bulk shergottites under unusual circumstances. It is thus likely that the shergottite range of compositions reflects primarily mantle sources.

The amount and distribution of water in the oceanic mantle lithosphere has implications on its strength and of the role of volatiles during plume/lithosphere interaction. The latter plays a role in the Earth's deep water cycle as water-rich plume lavas could re-enrich an oceanic lithosphere depleted in water at the ridge, and when this heterogeneous lithosphere gets recycled back into the deep mantle. The main host of water in mantle lithologies are nominally anhydrous minerals like olivine, pyroxene and garnet, where hydrogen (H) is incorporated in mineral defects by bonding to structural oxygen. Here, we report water concentrations by Fourier transform infrared spectrometry (FTIR) on olivine, clino- and orthopyroxenes (Cpx & Opx) from spinel peridotites from the Pali vent and garnet pyroxenite xenoliths from Aliamanu vent, both part of the rejuvenated volcanism at Oahu (Hawaii). Pyroxenes from the Aliamanu pyroxenites have high water concentrations, similar to the adjacent Salt Lake Crater (SLC) pyroxenites (Cpx ~400-500 ppm H2O, Opx ~200 ppm H2O). This confirms that pyroxenite cumulates form water-rich lithologies within the oceanic lithosphere. In contrast, the Pali peridotites have much lower water concentrations than the SLC ones (<25 ppm vs. 50-96 ppm H2O respectively) despite being relatively fertile with >10% modal Cpx and low spinel Cr# (0.09-0.10). The contrast between the two peridotite suites is also evident in their trace elements and radiogenic isotopes. The Pali Cpx are depleted in light REE, consistent with minimal metasomatism. Those of SLC have enriched light REE patterns and Nd and Hf isotopes consistent with metasomatism by alkaline melts. These observations are consistent with heterogeneous water distribution in the oceanic lithosphere that may be related to metasomatism, as well as relatively dry peridotites cross-cut by narrow (?) water-rich melt reaction zones.

The issue of whether crustal contamination has affected the lithophile trace element budget of shergottites has been a point of contention for decades. The evaluation has focused on the enriched shergottite compositions as an outcome of crustal contamination of mantle-derived parent magmas or, alternatively, the compositions of these stones reflect an incompatible trace element (ITE) enriched mantle source.

Most volatile studies of Mars have targeted its surface via spacecraft and rover data, and have evidenced surficial water in polar caps and the atmosphere, in the presence of river channels, and in the detection of water bearing minerals. The other focus of Martian volatile studies has been on Martian meteorites which are all from its crust. Most of these studies are on hydrous phases like apatite, a late-stage phase, i.e. crystallizing near the end of the differentiation sequence of Martian basalts and cumulates. Moreover, calculating the water content of the magma a phosphate crystallized from is not always possible, and yet is an essential step to estimate how much water was present in a parent magma and its source. Water, however, is primarily dissolved in the interiors of differentiated planets as hydrogen in lattice defects of nominally anhydrous minerals (olivine, pyroxene, feldspar) of the crust and mantle. This hydrogen has tremendous influence, even in trace quantities, on a planet's formation, geodynamics, cooling history and the origin of its volcanism and atmosphere as well as its potential for life. Studies of hydrogen in nominally anhydrous phases of Martian meteorites are rare. Measuring water contents and hydrogen isotopes in well-characterized nominally anhydrous minerals of Martian meteorites is the goal of our study. Our work aims at deciphering what influences the distribution and origin of hydrogen in Martian minerals, such as source, differentiation, degassing and shock.

The fact that Archean and Proterozoic cratons are underlain by the thickest (>200 km) lithosphere on Earth has always puzzled scientists because the dynamic convection of the surrounding asthenosphere would be expected to delaminate and erode these mantle lithospheric "keels" over time. Although density and temperature of the cratonic lithosphere certainly play a role in its strength and longevity, the role of water has only been recently addressed with data on actual mantle samples. Water in mantle lithologies (primarily peridotites and pyroxenites) is mainly stored in nominally anhydrous minerals (olivine, pyroxene, garnet) where it is incorporated as hydrogen bonded to structural oxygen in lattice defects. The property of hydrolytic weakening of olivine [4] has generated the hypothesis that olivine, the main mineral of the upper mantle, may be dehydrated in cratonic mantle lithospheres, contributing to its strength. This presentation will review the distribution of water concentrations in four cratonic lithospheres. The distribution of water contents in olivine from peridotite xenoliths found in kimberlites is different in each craton (Figure 1). The range of water contents of olivine, pyroxene and garnet at each xenolith location appears linked to local metasomatic events, some of which occurred later then the Archean and Proterozoic when these peridotites initially formed via melting. Although the low olivine water contents (<10 ppm wt H2O) at > 6 GPa at the base of the Kaapvaal cratonic lithosphere may contribute to its strength, and prevent its delamination, the wide range of those from Siberian xenoliths is not compatible with providing a high enough viscosity contrast with the asthenophere. The water content in olivine inclusions from Siberian diamonds, on the other hand, have systematically low water contents (<20 ppm wt H2O). The xenoliths may represent a biased sample of the cratonic lithosphere with an over-­abundance of metasomatized peridotites with high water contents. The olivine inclusions, however, may have been protected from metasomatism by their host diamond and record the overall low olivine water content of the cratonic lithosphere. Water may thus still play a role in cratonic keel longevity.

Two of the latest Martian meteorites found in Antarctica, paired olivine-phyric shergottites LAR 12240 and LAR 12095, are described in order to decipher their petrological context, and place constraints on the geological history of Mars. This project identifies all phases found in LAR 12240 and 12095 and analyzes them for major and trace elements. The textural relationships among these phases are examined in order to develop a crystallization history of the magma(s) that formed these basalts.

NWA 856 or Djel Ibone, is a basaltic shergottite discovered as a single stone of 320 g in South Morocco in April, 2001. This meteorite is fresh, i.e. shows minimal terrestrial weathering for a desert find. No shergottite discovered in North Africa can be paired with NWA 856. The purpose of this study is to constrain its crystallization history using textural observations, crystallization sequence modeling and in-situ trace element analysis in order to understand differentiation in shergottite magmatic systems.

Determining the volatile budget of the interior of Mars is crucial for our understanding of that planet's formation, geodynamics, cooling history and the origin of its volcanism and atmosphere as well as its potential for life. Surficial water is evident from spacecraft and rover data in polar caps and the atmosphere, in the presence of river channels, and in the detection of water-bearing minerals. Meteorites, however, are our best candidates for estimating the amount of water present at depth, even if all are crustal samples. The last 10 years have seen a blooming of studies measuring water and halogens in Martian meteorites. The bulk of these studies target phosphate, a typically late-stage phase in the igneous Martian meteorites that potentially would concentrate incompatible element hydrogen (H quantified traditionally as "water", i.e., H2O concentrations in weight) near the end of the crystallization sequence. However, determining the amount of water, F, and Cl in the magma from which a phosphate crystallized from is not straightforward and in most instances not possible. On the other hand, phosphates have turned out to be very useful in identifying hydrothermal processes that could have added water while or after the magma flowed and crystallized. Another caveat of analyzing Martian meteorite phases for water is that shocked phases such as maskelynite and impact melts appear to have incorporated water from the Martian atmosphere, as evidenced by high H isotope ((delta)D) signatures, and therefore their water concentrations cannot be interpreted in terms of deep planetary processes. The best candidates for estimating the water content of the Martian interior have been melt inclusions (glass or amphibole-bearing) which the enclosing mineral (usually olivine) would have prevented from exchanging volatiles with the surroundings after crystallization. Even some of these, however, have high (delta)D, meaning they were affected by H exchange via impact events or with crustal reservoirs or hydrothermal fluids. Here, nominally anhydrous minerals (pyroxene, olivine, plagioclase, or maskelynite) in orthopyroxenite ALH 84001 and selected nakhlites are analyzed for water and major elements, in order to determine 1) whether they contain any water; 2) if they do, what controls its distribution (crystallization, degassing, hydrothermal or impact processes); and 3) if any of these measurements can be used to infer the water contents of the parent magma and their mantle sources. A shock-reverberation experiment was also performed on terrestrial orthopyroxenes (opx) to simulate the heavily shocked conditions of ALH 84001 (> 31 GPa [17]).

Peridotite xenoliths from the Rio Grande Rift (RGR) are being analyzed for H (sub 2) O contents by FTIR (Fourier Transform Infrared) as well as for major and trace element compositions. Nine samples are from the Rio Puerco Volcanic Field (RP) which overlaps the central RGR and southeastern Colorado Plateau; seventeen samples are from Kilbourne Hole (KH) in the southern RGR. Spinel Cr# (Cr/(Cr+Al)) (0.08-0.46) and olivine Mg# (Mg/(Mg plus Fe)) (0.883-0.911) of all RGR samples fall within the olivine-spinel mantle array from [1], an indicator that peridotites are residues of partial melting. Pyroxene H (sub 2) O in KH correlate with bulk rock and pyroxene Al (sub 2) O (sub 3).The KH clinopyroxene rare earth element (REE) variations fit models of 0-13 percent fractional melting of a primitive upper mantle. Most KH peridotites have bulk-rock light REE depleted patterns, but five are enriched in light REEs consistent with metasomatism. Variation in H (sub 2) O content is unrelated to REE enrichment. Metasomatism is seen in RP pyroxenite xenoliths [2] and will be examined in the peridotites studied here. Olivine H (sub 2) O contents are low (less than or equal to 15 parts per million), and decrease from core to rim within grains. This is likely due to H loss during xenolith transport by the host magma [3]. Diffusion models of H suggest that mantle H (sub 2) O contents are still preserved in cores of KH olivine, but not RP olivine. The average H (sub 2) O content of Colorado Plateau clinopyroxene (670 parts per million) [4] is approximately 300 parts per million higher than RGR clinopyroxene (350 parts per million). This upholds the hypothesis that hydration-induced lithospheric melting occurred during flat-slab subduction of the Farallon plate [5]. Numerical models indicate hydration via slab fluids is possible beneath the plateau, approximately 600 kilometers from the paleo-trench, but less likely approximately 850 kilometers away beneath the rift [6].

Water in the mantle influences melting, metasomatism, viscosity and electrical conductivity. The Alligator Lake mantle xenolith suite is one of three bimodal peridotite suites from the northern Canadian Cordillera brought to the surface by alkali basalts, i.e., it consists of chemically distinct lherzolites and harzburgites. The lherzolites have equilibration temperatures about 50 C lower than the harzburgites and are thought to represent the fertile upper mantle of the region. The harzburgites might have come from slightly deeper in the mantle and/or be the result of a melting event above an asthenospheric upwelling detected as a seismic anomaly at 400-500 km depth. Major and trace element data are best interpreted as the lherzolite mantle having simultaneously experienced 20-25% partial melting and a metasomatic event to create the harzburgites. Well-characterized xenoliths are being analyzed for water by FTIR. Harzburgites contain 29-52 ppm H2O in orthopyroxene (opx) and (is) approximately140 ppm H2O in clinopyroxene (cpx). The lherzolites have H2O contents of 27-150 ppm in opx and 46-361 ppm in cpx. Despite correlating with enrichments in LREE, the water contents of the harzburgite pyroxenes are low relative to those of typical peridotite xenoliths, suggesting that the metasomatic agents were water-poor, contrarily to what has been suggested before. The water content of cpx is about double that of opx indicating equilibrium. Olivine water contents are low ((is) less than 5 ppm H2O) and out of equilibrium with those of opx and cpx, which may be due to H loss during xenolith ascent. This is consistent with olivines containing more water in their cores than their rims. Olivines exclusively exhibit water bands in the 3400-3000 cm-1 range, which may be indicative of a reduced environment.

Nominally anhydrous minerals such as olivine, pyroxene and garnet can accommodate tens to hundreds of ppm H2O in the form of hydrogen bonded to structural oxygen in lattice defects. Although in seemingly small amounts, this water can significantly alter chemical and physical properties of the minerals and rocks. Water in particular can modify their rheological properties and its distribution in the mantle derives from melting and metasomatic processes and lithology repartition (pyroxenite vs peridotite). These effects will be examined here using Fourier transform infrared spectrometry (FTIR) water analyses on minerals from mantle xenoliths from cratons, plume-influenced cratons and oceanic settings. In particular, our results on xenoliths from three different cratons will be compared. Each craton has a different water distribution and only the mantle root of Kaapvaal has evidence for dry olivine at its base. This challenges the link between olivine water content and survival of Archean cratonic mantle, and questions whether xenoliths are representative of the whole cratonic mantle. We will also present our latest data on Hawaii and Tanzanian craton xenoliths which both suggest the intriguing result that mantle lithosphere is not enriched in water when it interacts with melts from deep mantle upwellings (plumes).

Diogenites are achondrites that are part of the HED (howardite, eucrite, diogenite) meteorite group thought to originate from asteroid Vesta. This suite of igneous rocks offers a glimpse of early planetary differentiation and subsequent igneous processes. While eucrites represent asteroidal basaltic crust and howardites the impact brecciated surface, diogenites are samples of the mantle and lower crust. Most of them are orthopyroxene (Opx) dominated cumulates, although harzburgites and rare dunites have also been found. The majority of diogenites are impact breccias. This study describes NWA 5480, a pristine, i.e. hardly altered and minimally shocked, harzburgitic diogenite.

Martian meteorite RBT 04262 is in the shergottite class. It displays the two lithologies typically found in "lherzolitic shergottites": one with a poikilitic texture of large pyroxene enclosing olivine and another with non-poikilitic texture. In the case of RBT 04262, the latter strongly ressembles an olivine- phyric shergottite which led the initial classification of this meteorite in that class. RBT 04262 has been studied with regards to its petrology, geochemistry and cosmic ray exposure and belongs to the enriched oxidized end-member of the shergottites. Studies on RBT 04262 have primarily focused on the bulk rock composition or each of the lithologies independently. To further elucidate RBT 04262's petrology and use it to better understand Martian geologic history, an in-depth study of its melt inclusions (MI) is being conducted. The MI chosen for this study are found within olivine grains. MI are thought to be trapped melts of the crystallizing magma preserved by the encapsulating olivine and offer snapshots of the composition of the magma as it evolves. Some MI, in the most Mg-rich part of the olivine of olivine-pyric shergottites, may even be representative of the meteorite parent melt.

The detection of indigenous water in mare basaltic glass beads has challenged the view established since the Apollo era of a "dry" Moon. Since this discovery, measurements of water in lunar apatite, olivine-hosted melt inclusions, agglutinates, and nominally anhydrous minerals have confirmed that lunar igneous materials contain water, implying that some parts of lunar mantle may have as much water as Earth's upper mantle. The interpretation of hydrogen (H) isotopes in lunar samples, however, is controversial. The large variation of H isotope ratios in lunar apatite (delta Deuterium = -202 to +1010 per mille) has been taken as evidence that water in the lunar interior comes from the lunar mantle, solar wind protons, and/or comets. The very low deuterium/H ratios in lunar agglutinates indicate that solar wind protons have contributed to their hydrogen content. Conversely, H isotopes in lunar volcanic glass beads and olivine-hosted melt inclusions being similar to those of common terrestrial igneous rocks, suggest a common origin for water in both Earth and Moon. Lunar water could be inherited from carbonaceous chondrites, consistent with the model of late accretion of chondrite-type materials to the Moon as proposed by. One complication about the sources of lunar water, is that geologic processes (e.g., late accretion and magmatic degassing) may have modified the H isotope signatures of lunar materials. Recent FTIR analyses have shown that plagioclases in lunar ferroan anorthosite contain approximately 6 ppm H2O. So far, ferroan anorthosite is the only available lithology that is believed to be a primary product of the lunar magma ocean (LMO). A possible consequence is that the LMO could have contained up to approximately 320 ppm H2O. Here we examine the possible sources of water in the LMO through measurements of water abundances and H isotopes in plagioclase of two ferroan anorthosites and one troctolite from lunar highlands.

Hydrogen is present as a trace element in olivine and pyroxene and its content distribution in the mantle results from melting and metasomatic processes. Here we examine how these H contents can be disturbed during decompression. Hydrogen was analyzed by FTIR in olivine and pyroxene of spinel peridotite xenoliths from Salt Lake Crater (SLC) nephelinites which are part of the rejuvenated volcanism at Oahu (Hawaii) [1,2]. H mobility in pyroxene resulting from spinel exsolution during mantle upwelling Most pyroxenes in SLC peridotites exhibit exsolutions, characterized by spinel inclusions. Pyroxene edges where no exsolution are present have less H then their core near the spinel. Given that H does not enter spinel [3], subsolidus requilibration may have concentrated H in the pyroxene adjacent to the spinel exsolution during mantle upwelling. H diffusion in olivine during xenolith transport by its host magma and host magma ascent rates Olivines have lower water contents at the edge and near fractures compared to at their core, while the concentrations of all other chemical elements appear homogeneous. This suggests that some of the initial water has diffused out of the olivine. Water loss from the olivine is thought to occur during host-magma ascent and xenolith transport to the surface [4-6]. Diffusion modeling matches best the data when the initial water content used is that measured at the core of the olivines, implying that mantle water contents are preserved at the core of the olivines. The 3225 cm(sup -1) OH band at times varies independantly of other OH bands, suggesting uneven H distribution in olivine defects likely acquired during mantle metasomatism just prior to eruption and unequilibrated. Diffusion times (1-48 hrs) combined with depths of peridotite equilibration or of magma start of degassing allow to calculate ascent rates for the host nephelinite of 0.1 to 27 m/s.

Spinel peridotite and garnet pyroxenite xenoliths from Hawaii provide important clues about the composition of the oceanic lithosphere, and can be used to assess its contribution to mantle heterogeneity upon recycling. The peridotites have lower bulk H2O (approximately 70-114 ppm) than the MORB source, qualitatively consistent with melt depletion. The garnet pyroxenites (high pressure cumulates) have higher H2O (200-460 ppm, up to 550 ppm accounting for phlogopite) and low H2O/Ce ratios (less than 100). The peridotites have relatively light Fe-isotopes (delta Fe -57 = -0.34 to 0.13) that decrease with increasing depletion, while the pyroxenites are significantly heavier (delta Fe-57 up to 0.3). The observed xenolith, as well as MORB and OIB total Fe-isotope variability is larger that can be explained by existing melting models. The high H2O and low H2O/Ce ratios of pyroxenites are similar to estimates of EM-type OIB sources, while their heavy delta Fe-57 are similar to some Society and Cook-Austral basalts. Therefore, recycling of mineralogically enriched oceanic lithosphere (i.e. pyroxenites) may contribute to OIB sources and mantle heterogeneity. The Fe(3+)/Sigma systematics of these xenoliths also suggest that there might be lateral redox gradients within the lithosphere, between juxtaposed oxidized spinel peridotites (deltaFMQ = -0.7 to 1.6, at 15 kb) and more reduced pyroxenites (deltaFMQ = -2 to -0.4, at 20-25kb). Such mineralogically and compositionally imposed fO2 gradients may generate local redox melting due to changes in fluid speciation (e.g. reduced fluids from pyroxenite encountering more oxidized peridotite). Formation of such incipient, small degree melts could further contribute to metasomatic features seen in peridotites, mantle heterogeneity, as well as the low velocity and high electrical conductivity structures near the base of the lithosphere and upper mantle.

The East African Rift System was initiated by the impingement of the Afar mantle plume on the base of the non-cratonic continental lithosphere (assembled during the Pan-African Orogeny), producing over 300,000 kmof continental flood basalts approx.30 Ma ago. The contribution of the subcontinental lithospheric mantle (SCLM) to this voluminous period of volcanism is implied based on basaltic geochemical and isotopic data. However, the role of percolating melts on the SCLM composition is less clear. Metasomatism is capable of hybridizing or overprinting the geochemical signature of the SCLM. In addition, models suggest that adding fluids to lithospheric mantle affects its stability. We investigated the nature of the SCLM using Fourier transform infrared spectrometry (FTIR) to measure water content in mantle xenoliths entrained in young (1 Ma) basaltic lavas from the Ethiopian volcanic province. The mantle xenoliths consist dominantly of spinel lherzolites and are composed of nominally anhydrous minerals, which can contain trace water as H in mineral defects. Eleven mantle xenoliths come from the Injibara-Gojam region and two from the Mega-Sidamo region. Water abundances of olivines in six samples are 1-5ppm H2O while the rest are below the limit of detection (<0.5 ppm H2O); orthopyroxene and clinopyroxene contain 80-238 and 111-340 ppm wt H2O, respectively. Two xenoliths have higher water contents - a websterite (470 ppm) and dunite (229 ppm), consistent with involvement of ascending melts. The low water content of the upper SCLM beneath Ethiopia is as dry as the oceanic mantle except for small domains represented by percolating melts. Consequently, rifting of the East African lithosphere may not have been facilitated by a hydrated upper mantle.

The mantle lithosphere beneath the cratonic part of continents is the deepest (> 200 km) and oldest (>2-3 Ga) on Earth, remaining a conundrum as to how these cratonic roots could have resisted delamination by asthenospheric convection over time. Water, or trace H incorporated in mineral defects, could be a key player in the evolution of continental lithosphere because it influences melting and rheology of the mantle. Mantle xenoliths from the Lac de Gras kimberlite in the Slave craton were analyzed by FTIR. The cratonic mantle beneath Lac de Gras is stratified with shallow (<145 km) oxidized ultradepleted peridotites and pyroxenites with evidence for carbonatitic metasomatism, underlain by reduced and less depleted peridotites metasomatized by kimberlite melts. Peridotites analyzed so far have H O contents in ppm weight of 7-100 in their olivines, 58 to 255 in their orthopyroxenes (opx), 11 to 84 in their garnet, and 139 in one clinopyroxene. A pyroxenite contains 58 ppm H2O in opx and 5 ppm H2O in its olivine and garnet. Olivine and garnet from the deep peridotites have a range of water contents extending to higher values than those from the shallow ones. The FTIR spectra of olivines from the shallow samples have more prominent Group II OH bands compared to the olivines from the deep samples, consistent with a more oxidized mantle environment. The range of olivine water content is similar to that observed in Kaapvaal craton peridotites at the same depths (129-184 km) but does not extend to as high values as those from Udachnaya (Siberian craton). The Slave, Kaapvaal and Siberian cratons will be compared in terms of water content distribution, controls and role in cratonic root longevity.

Kilbourne Hole (NM, USA) and Dish Hill (CA, USA) mantle xenoliths sample continental mantle in two different tectonic settings. Kilbourne Hole (KH) is located in the Rio Grande rift. Dish Hill (DH) is located in the southern Mojave province, an area potentially affected by subduction of the Farallon plate beneath North America. FTIR analyses were obtained on well characterized pyroxenite, dunite and wehrlite xenoliths, thought to represent crystallized melts at mantle depths. PUM normalized REE patterns of the KH bulk-rocks are slightly LREE enriched and consistent with those of liquids generated by < 5% melting of a spinel peridotite source. Clinopyroxenes contain from 272 to 313 ppm weight H2O similar to the lower limit of KH peridotite clinopyroxenes (250-530 ppm H2O). This is unexpected as crystallized melts like pyroxenites should concentrate water more than residual mantle-like peridotites, given that H is incompatible. PUM normalized bulk REE of the DH pyroxenites are characterized by flat to LREE depleted REE profiles consistent with > 6% melting of a spinel peridotite source. Pyroxenite pyroxenes have no detectable water but one DH wehrlite, which bulk-rock is LREE enriched, has 4 ppm H2O in orthopyroxene and <1ppm in clinopyroxene. The DH pyroxenites may thus come from a dry mantle source, potentially unaffected by the subduction of the Farallon plate. These water-poor melts either originated from shallow oceanic lithosphere overlaying the Farallon slab or from continental mantle formed > 2 Ga. The Farallon subduction appears to have enriched in water the southwestern United States lithospheric mantle further east than DH, beneath the Colorado plateau.

The lunar meteorite NWA 3163 (paired with NWA 4881, 4483) is a ferroan, feldspathic granulitic breccia characterized by pigeonite, augite, olivine, maskelynite and accessory Tichromite, ilmenite and troilite. Bulk rock geochemical signatures indicate the lack of a KREEP- derived component (Eu/Eu* = 3.47), consistent with previously studied lunar granulites and anorthosites. Bulk rock chondrite-normalized signatures are however distinct from the anorthosites and granulites sampled by Apollo missions and are relatively REE-depleted. In-situ analyses of maskelynite reveal little variation in anorthite content (average An% is 96.9 +/- 1.6, 2 sigma). Olivine is relatively ferroan and exhibits very little variation in forsterite content with mean Fo% of 57.7 +/- 2.0 (2 sigma). The majority of pyroxene is low-Ca pigeonite (En57Fs33Wo10). Augite (En46Fs21Wo33) is less common, comprising approximately 10% of analyzed pyroxene. Two pyroxene thermometry on co-existing orthopyroxene and augite yield an equilibrium temperature of 1070C which is in reasonable agreement with temperatures of 1096C estimated from pigeonite compositions. Rb-Sr isotopic systematics of separated fractions yield an average measured Sr-87/Sr-87 of 0.699282+/-0.000007 (2 sigma). Sr model ages are calculated using a modern day Sr-87/Sr-86 Basaltic Achondrite Best Initial (BABI) value of 0.70475, from an initial BABI value Sr-87/Sr-86 of 0.69891 and a corresponding Rb-87/Sr-97 of 0.08716. The Sr model Thermomechanical analysis (TMA) age, which represents the time of separation of a melt from a source reservoir having chondritic evolution, is 4.56+/-0.1 Ga. A Sr model T(sub RD) age, which is a Rb depletion age and assumes no contribution from Rb in the sample in the calculation, yields 4.34+/-0.1 Ga (i.e. a minimum age). The Ar-Ar dating of paired meteorite NWA 4881 reveals an age of c. 2 Ga, likely representing the last thermal event this meteorite experienced. An older Ar-40/Ar-39 age of c. 3.5 Ga may record the thermal event which produced the granulitic texture. Additional chronological constraints will be provided by Sm-Nd systematics. Ferroan Anorthosites like NWA 3163 have been interpreted to represent direct lunar magma ocean (LMO) crystallization products. If this is the case, trace element concentrations in NWA 3163 primary mineral phases should be in equilibrium with residual LMO liquids present during crystallization of those phases. Results from petrogenetic modeling suggest that the NWA 3163 protolith did not form from crystallization of an initially LREE depleted LMO but rather require an initially chondritic LMO with early garnet crystallization. Furthermore, a two-stage crystallization model where plagioclase crystalized prior to pyroxene (93% vs. 99.5% of LMO crystallization) is implied.

Water is believed to be a key factor controlling the long-term stability of cratonic lithosphere, but mechanisms responsible for the water content distribution in the mantle remain poorly constrained. Water contents were obtained by FTIR in olivine, pyroxene and garnet for 20 well-characterized peridotite xenoliths from the Udachnaya kimberlite (central Siberian craton) and equilibrated at 2-7 GPa. Water contents in minerals do not appear to be related to interaction with the host kimberlite. Diffusion modeling indicates that the core of olivines preserved their original water contents. The Udachnaya peridotites show a broad range of water contents in olivine (6.5 +/- 1.1 to 323 +- 65 ppm H2O (2 sigma)), and garnet (0 - 23 +/- 6 ppm H2O). The water contents of olivine and garnet are positively correlated with modal clinopyroxene, garnet and FeO in olivine. Water-rich garnets are also rich in middle rare earth elements. This is interpreted as the result of interaction between residual peridotites and water rich-melts, consistent with modal and cryptic metasomatism evidenced in the Siberian cratonic mantle. The most water-rich Udachnaya minerals contain 2 to 3 times more water than those from the Kaapvaal craton, the only craton with an intact mantle root for which water data is available. The highest water contents in olivine and orthopyroxene in this study (>= 300 ppm) are found at the bottom of the lithosphere (> 6.5 GPa). This is in contrast with the Kaapvaal craton where the olivines of peridotites equilibrated at > 6.4 GPa have < 1 ppm H2O. The latter "dry" olivine may make the base of the Kaapvaal cratonic root strong and thus protects it from erosion by the convective mantle The calculated viscosity for water-rich Udachnaya peridotites at > 6 GPa is lower or similar (8.4× 10(exp 16) to 8.0× 10(exp 18) Pa./s) to that of the asthenosphere (<= 3.7x10(exp 18) Pa./s ). Such lithologies would not be able to resist delamination by the convecting asthenosphere. However, seismology studies as well as the high equilibration pressures of our samples indicate that the Udachnaya cratonic lithosphere is 220-250 km thick. Consequently, the water-rich peridotites are likely not representative of the overall Siberian cratonic lithosphere. Their composition is linked to spatially limited melt metasomatism in mantle regions above asthenospheric upwellings responsible for the kimberlite magmatism prior to their ascent and eruption.

Water dissolved as trace amounts in anhydrous minerals has a large influence on the melting behavior and physical properties of the mantle. The water concentration of the oceanic mantle is inferred from the analyses of Mid-Ocean Ridge Basalt (MORB) and Oceanic Island Basalt (OIB). but there is little data from actual mantle samples. Moreover, enriched mineralogies (pyroxenites, eclogites) are thought as important sources of heterogeneity in the mantle, but their water concentrations and their effect on the water budget and cycling in the mantle are virtually unknown. Here, we analyzed by FTIR water in garnet clinopyroxenite xenoliths from Salt Lake Crater, Oahu, Hawaii. These pyroxenites are high-pressure (>20kb) crystal fractionates from alkalic melts. The clinopyroxenes (cpx) have 260 to 576 ppm wt H2O, with the least differentiated samples (Mg#>0.8) in the 400-500 ppm range. Orthopyroxene (opx) contain 117-265 ppm H2O, about half of that of cpx, consistent with other natural sample studies, but lower than cpx/opx equilibrium from experimental data. The pyroxenite cpx and opx H2O concentrations are at the high-end of on-and off-craton peridotite xenolith concentrations and those of Hawaiian spinel peridotites. In contrast, garnet has extremely low water contents (<5ppm H2O). There is no correlation between H2O in cpx and lithophile element concentrations. Phlogopite is present in some samples, and its modal abundance shows a positive correlation in Mg# with cpx, implying equilibrium. However, there is no correlation between H2O concentrations and or the presence of phlogopite. These data imply that cpx and opx may be at water saturation, far lower than experimental data suggest. Reconstructed bulk rock pyroxenite H2O ranges from 200-460 ppm (average 331 +/- 75 ppm), 2 to 8 times higher than H2O estimates for the MORB source (50-200 ppm), but in the range of E-MORB, OIB and the source of rejuvenated Hawaiian magmas. The average bulk rock pyroxenite H2O/Ce is 69 +/-35, lower than estimates of the MORB source (approx 150) or FOZO, C (200-250) mantle component, but consistent with "dry" EM sources (<100). These data suggest that a metasomatized, refertilized oceanic lithosphere that contains pyroxenitic veins (e.g. the lower part of an oceanic plate, where ascending melts can become trapped and crystallize), will have both higher water concentrations and low H2O/Ce, and may contribute to EM-type OIB sources, like that of Samoa basalts. Therefore, a low H2O/Ce mantle source may not necessarily be "dry".

The water content distribution of the upper mantle will be reviewed as based on the peridotite record. The amount of water in cratonic xenoliths appears controlled by metasomatism while that of the oceanic mantle retains in part the signature of melting events. In both cases, the water distribution is heterogeneous both with depth and laterally, depending on localized water re-enrichments next to melt/fluid channels. The consequence of the water distribution on the rheology of the upper mantle and the location of the lithosphere-asthenosphere boundary will also be discussed.

Lunar meteorites are crucial to understand the Moon s geological history because, being samples of the lunar crust that have been ejected by random impact events, they potentially originate from areas outside the small regions of the lunar surface sampled by the Apollo and Luna missions. The Apollo and Luna sample sites are contained within the Procellarum KREEP Terrain (PKT, Jolliff et al., 2000), where KREEP refers to potassium, rare earth element, and phosphorus-rich lithologies. The KREEP-rich rocks in the PKT are thought to be derived from late-stage residual liquids after approx.95-99% crystallization of a lunar magma ocean (LMO). These are understood to represent late-stage liquids which were enriched in incompatible trace elements (ITE) relative to older rocks (Snyder et al., 1992). As a consequence, the PKT is a significant reservoir for Th and KREEP. However, the majority of the lunar surface is likely to be significantly more depleted in ITE (84%, Jolliff et al., 2000). Lunar meteorites that are low in KREEP and Th may thus sample regions distinct from the PKT and are therefore a valuable source of information regarding the composition of KREEP-poor lunar crust. Northwest Africa (NWA) 3163 is a thermally metamorphosed ferroan, feldspathic, granulitic breccia composed of igneous clasts with a bulk anorthositic, noritic bulk composition. It is relatively mafic (approx.5.8 wt.% FeO; approx.5 wt.% MgO) and has some of the lowest concentrations of ITEs (17ppm Ba) compared to the feldspathic lunar meteorite (FLM) and Apollo sample suites (Hudgins et al., 2011). Localized plagioclase melting and incipient melting of mafic minerals require localized peak shock pressures in excess of 45 GPa (Chen and El Goresy, 2000; Hiesinger and Head, 2006). NWA 3163, and paired samples NWA 4481 and 4883, have previously been interpreted to represent an annealed micro-breccia which was produced by burial metamorphism at depth in the ancient lunar crust (Fernandes et al., 2009). This is in contrast to the interpretation of Hudgins et al. (2009) where NWA 3163 was interpreted to have formed through contact metamorphism. To further constrain its origin, we examine the petrogenesis of NWA 3163 with a particular emphasis on in-situ measurement of trace elements within constituent minerals, Sm-Nd and Rb-Sr isotopic systematics on separated mineral fractions and petrogenetic modeling.