Search

The distribution of water in differentiated Solar System bodies depends on many factors including size, distance from the Sun, and how they incorporated water. Most of this water is likely locked as hydrogen in mantle minerals and could amount to several Earth oceans worth in mass for the largest planets. An essential compound for the development of life, water also has a tremendous influence on planetary evolution and volcanism. Only Earth has an active exchange of water between surface and mantle. Surface water on other differentiated bodies mostly results from degassing by volcanoes whose mantle sources are inherited from magma ocean processes early in their history. Airless bodies also acquire surface water by impacts, spallation, and from the solar wind.

Whether hydrogen incorporated in nominally anhydrous mantle minerals plays a role in the strength and longevity of the thick cratonic lithosphere is a matter of debate. In particular, the percolation of hydrogen-bearing melts and fluids could potentially add hydrogen to the mantle lithosphere, weaken its olivines (the dominant mineral in mantle peridotite), and cause delamination of the lithosphere's base. The influence of metasomatism on hydrogen contents of cratonic mantle minerals can be tested in mantle xenoliths from the Slave Craton (Canada) because they show extensive evidence for metasomatism of a layered cratonic mantle. Minerals from mantle xenoliths from the Diavik mine in the Lac de Gras kimberlite area located at the center of the Archean Slave craton were analyzed by FTIR for hydrogen contents. The 18 peridotites, two pyroxenites, one websterite and one wehrlite span an equilibration pressure range from 3.1 to 6.6 GPa and include samples from the shallow (≤145 km), oxidized ultra-depleted layer; the deeper (∼145–180 km), reduced less depleted layer; and an ultra-deep (≥180 km) layer near the base of the lithosphere. Olivine, orthopyroxene, clinopyroxene and garnet from peridotites contain 30–145, 110–225, 105–285, 2–105 ppm H2O, respectively. Within each deep and ultra-deep layer, correlations of hydrogen contents in minerals and tracers of metasomatism (for example light over heavy rare-earth-element ratio (LREE/HREE), high-field-strength-element (HFSE) content with equilibration pressure) can be explained by a chromatographic process occurring during the percolation of kimberlite-like melts through garnet peridotite. The hydrogen content of peridotite minerals is controlled by the compositions of the evolving melt and of the minerals and by mineral/melt partition coefficients. At the beginning of the process, clinopyroxene scavenges most of the hydrogen and garnet most of the HFSE. As the melt evolves and becomes enriched in hydrogen and LREE, olivine and garnet start to incorporate hydrogen and pyroxenes become enriched in LREE. The hydrogen content of peridotite increases with decreasing depth, overall (e.g., from 75 to 138 ppm H2O in the deep peridotites). Effective viscosity calculated using olivine hydrogen content for the deepest xenoliths near the lithosphere-asthenosphere boundary overlaps with estimates of asthenospheric viscosities. These xenoliths cannot be representative of the overall cratonic root because the lack of viscosity contrast would have caused basal erosion of lithosphere. Instead, metasomatism must be confined in narrow zones channeling kimberlite melts through the lithosphere and from where xenoliths are preferentially sampled. Such localized metasomatism by hydrogen-bearing melts therefore does not necessarily result in delamination of the cratonic root.

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. The nakhlites represent a suite of minimally shocked meteorites that likely originated from lava flows from a single volcano or from a shallow intrusion or sill complex on Mars. Measuring the water contents and D/H ratios of their igneous minerals allows identification of phases that have preserved their magmatic hydrogen, and therefrom permits estimation of the water content of their mantle source. Pyroxene, olivine, melt inclusions and mesostasis of five nakhlites (NWA 998, Nakhla, Y 000593, MIL 03346 and NWA 6148) were analyzed in situ for water contents and H isotopes, and major and trace element contents. No water was detected in olivine grains except in Y 000593. The water content of pyroxenes is highly heterogeneous within individual grains and between grains within a single meteorite. Water concentrations in pyroxene ( After ruling out significant influence from spallation, exchange with the martian atmosphere, shock, surface alteration, and hydrothermal processes, the H data of the pyroxenes can be explained by degassing and crystallization processes. Degassing is consistent with a decrease of water content from pyroxene interior to edge. Fractionation of H isotopes during degassing results in increases of δD during H loss from pyroxene but in decreases in δD during H2O-OH loss from a melt. Consequently, the low-water content, high-δD of most pyroxenes is best explained by degassing after the pyroxenes had crystallized. All melt and plagioclase inclusions analyzed are located in degassed pyroxenes and are also degassed. The lower δD of the mesostasis (24 ± 131‰) compared to that of the least-degassed pyroxenes (430 ± 172‰) is likely the result of melt degassing and interaction with hydrothermal fluids. Magmatic H, however, has been preserved in each nakhlite in some pyroxenes that are characterized by >15 ppm H2O and δD

The distribution of water in the upper mantle plays a crucial role in the Earth's deep water cycle, magmatism, and plate tectonics. To better constrain how these large-scale geochemical systems operate, peridotite and pyroxenite mantle xenoliths from Kilbourne Hole (KH) and Rio Puerco (RP) along the Rio Grande Rift (NM, USA) were analyzed for water, and major and trace element contents. These xenoliths sample a lithosphere whose composition was influenced by subduction and rifting, and can be used to examine the effects of melting, metasomatism, and sub-solidus equilibration on the behavior of water. The first result is that in KH xenoliths, olivines underwent negligible H loss during xenolith ascent, i.e. preserved their mantle water contents. These olivine water contents are used to calculate mantle viscosities of 0.5–184 · 1021 Pa·s. These viscosity values are more than 40 times higher than those of the asthenosphere and show that KH peridotites represent samples from the lithosphere. The preservation of olivine water contents is exceptional for off-cratonic xenoliths, and the KH peridotites provide the first estimate of the average concentration of water in Phanerozoic continental mantle lithosphere at 81 ± 30 ppm H2O. The mantle lithosphere beneath the Rio Grande rift is nevertheless heterogeneous with water contents ranging from

New measurements of Earth's building blocks point to a simpler water source

Water influences geodynamic processes such as melting, deformation and rheology, yet its distribution in the oceanic upper mantle is primarily known indirectly from melt inclusions and glasses of erupted mantle melts (i.e. mid-ocean ridge and ocean island basalts). To better constrain the mechanisms influencing the distribution of H2O in the mantle, particularly regarding the role of metasomatism, we analyzed 15 peridotite xenoliths from Savai‘i and two dunite xenoliths from Ta‘ū (Samoa) for structural H2O (by polarized Fourier transform infrared spectroscopy), and major and trace element concentrations. Clinopyroxenes from the Ta‘ū dunites show trace element concentrations consistent with equilibration with their host lavas, but lower H2O contents than expected. Savai‘i peridotites are highly depleted harzburgites (melt depletion ≥17 %). They show strong evidence of transient metasomatism by both carbonatite and silicate melts, with highly variable Ti and Zr depletions and light rare earth element enrichments. However, despite metasomatism the H2O concentrations in olivines (0 − 4 ppm H2O) and orthopyroxenes (17 − 89 ppm H2O) are among the lowest reported in oceanic xenoliths, but higher than expected for the estimated degree of depletion. In general, H2O concentrations vary less than those of other incompatible trace elements in these samples. Transects across mineral grains show generally homogeneous distributions of H2O, indicating no significant H2O loss or gain during ascent. Raman spectroscopy on inclusions in minerals shows the presence of CO2 but an absence of molecular H2O. This agrees with the absence of H2O concentration variations between inclusion-rich and -poor domains in minerals. The above data can be explained by transient metasomatism along grain boundaries, now recorded as planes of inclusions within annealed grains. Fast diffusion of hydrogen (but not lithophile elements) from the inclusions into the host mineral phase will simultaneously enrich H2O contents across the grain and lower them in the inclusion-rich domains. The result is highly variable metasomatism recorded in lithophile elements, with smaller magnitude H2O variations that are decoupled from lithophile element metasomatism. Comparison with xenoliths from Hawai‘i shows that evidence for metasomatism from lithophile elements alone does not imply rehydration of the oceanic lithosphere. Instead, H2O concentrations depend on the overall amount of H2O added to the lithosphere through metasomatism, and the proximity of sampled material to areas of melt infiltration in the lithosphere.

It is widely accepted that the Apollo 17 orange glass beads are of volcanic origin. The degree of degassing of the glass beads depends on their cooling rates, so the estimation of volatile contents in the parental magmas of these lunar pyroclastic glasses also depends on the cooling rates. The cooling rate can be estimated using the calorimetric properties of the glass across the glass transition. In this study, a series of heat capacity measurements were carried out on hand-picked lunar volcanic orange glass beads during several cycles of heating to temperatures above their glass transition using a differential scanning calorimeter. The cooling rate of orange glass beads (sample 74220,867) was calculated to be 101 K/min using the correlation between glass cooling rates and fictive temperatures estimated from their heat capacity–temperature paths. This cooling rate is close to the lower end of the range that best fits the diffusion profiles of several volatile species in the glass beads, and at the upper end of the cooling-rate range recorded in glasses quenched subaerially on Earth. The cooling rate is likely to be controlled by the cooling environment (cooling medium) such that the lunar volcanic glass beads could have been cooled in a gaseous medium released from volcanic eruptions on the Moon, but not during “free flight” in vacuum. The existence of a gas medium suggests, in turn, that there may have been at least a short-lived or episodic atmosphere on the early Moon at around 3.5 Ga.

Knowing the amount and timing of water incorporation into the Moon has fundamental implications for our understanding of how the Earth–Moon system formed. Water has been detected in lunar samples but its abundance, distribution and origin are debated. To address these issues, we report water concentrations and hydrogen isotope ratios obtained by secondary ion mass spectrometry (SIMS) of plagioclase from ferroan anorthosites (FANs), the only available lithology thought to have crystallized directly from the lunar magma ocean (LMO). The measured water contents are consistent with previous results by Fourier transform infrared spectroscopy (FTIR). Combined with literature data, δD values of lunar igneous materials least-degassed at the time of their crystallization range from −280 to +310‰, the latter value being that of FAN 60015 corrected for cosmic ray exposure. We interpret these results as hydrogen isotopes being fractionated during degassing of molecular hydrogen (H_2) in the LMO, starting with the magmatic δD value of primordial water at the beginning of LMO being about −280‰, evolving to about +310‰ at the time of anorthite crystallization, i.e. during the formation of the primary lunar crust. The degassing of hydrogen in the LMO is consistent with those of other volatile elements. The wide range of δD values observed in lunar igneous rocks could be due to either various degrees of mixing of the different mantle end members, or from a range of mantle sources that were degassed to different degrees during magma evolution. Degassing of the LMO is a viable mechanism that resulted in a heterogeneous lunar interior for hydrogen isotopes.

The concentration and distribution of water in the Earth has influenced its evolution throughout its history. Even at the trace levels contained in the planet’s deep interior (mantle and core), water affects Earth’s thermal, deformational, melting, electrical and seismic properties, that control differentiation, plate tectonics and volcanism. These in turn influenced the development of Earth’s atmosphere, oceans, and life. In addition to the ubiquitous presence of water in the hydrosphere, most of Earth’s “water” actually occurs as trace amounts of hydrogen incorporated in the rock-forming silicate minerals that constitute the planet’s crust and mantle, and may also be stored in the metallic core. The heterogeneous distribution of water in the Earth is the result of early planetary differentiation into crust, mantle and core, followed by remixing of lithosphere into the mantle after plate-tectonics started. The Earth’s total water content is estimated at $18_{-15}^{+81}$ times the equivalent mass of the oceans (or a concentration of $3900_{-3300}^{+32700}~\mbox{ppm}$ weight H2O). Uncertainties in this estimate arise primarily from the less-well-known concentrations for the lower mantle and core, since samples for water analyses are only available from the crust, the upper mantle and very rarely from the mantle transition zone (410–670 km depth). For the lower mantle (670–2900 km) and core (2900–4500 km), the estimates rely on laboratory experiments and indirect geophysical techniques (electrical conductivity and seismology). The Earth’s accretion likely started relatively dry because it mainly acquired material from the inner part of the proto-planetary disk, where temperatures were too high for the formation and accretion of water ice. Combined evidence from several radionuclide systems (Pd-Ag, Mn-Cr, Rb-Sr, U-Pb) suggests that water was not incorporated in the Earth in significant quantities until the planet had grown to $\sim60\mbox{--}90\%$ of its current size, while core formation was still on-going. Dynamic models of planet formation provide additional evidence for water delivery to the Earth during the same period by water-rich planetesimals originating from the asteroid belt and possibly beyond. This early delivered water may have been partly lost during giant impacts, including the Moon forming event: magma oceans can form in their aftermath, degas and be followed by atmospheric loss. More water may have been delivered and/or lost after core formation during late accretion of extraterrestrial material (“late-veneer”). Stable isotopes of hydrogen, carbon, nitrogen and some noble gases in Earth’s materials show similar compositions to those in carbonaceous chondrites, implying a common origin for their water, and only allowing for minor water inputs from comets.

Lunar granulitic meteorites provide new constraints on the composition and evolution of the lunar crust as they are potentially derived from outside the Apollo and Luna landing sites. Northwest Africa (NWA) 3163, the focus of this study, and its paired stones NWA 4881 and NWA 4483, are shocked granulitic noritic anorthosites. They are petrographically and compositionally distinct from the Apollo granulites and noritic anorthosites. Northwest Africa 3163 is REE-depleted by an order of magnitude compared to Apollo granulites and is one of the most trace element depleted lunar samples studied to date. New in-situ mineral compositional data and Rb–Sr, Ar–Ar isotopic systematics are used to evaluate the petrogenetic history of NWA 3163 (and its paired stones) within the context of early lunar evolution and the bulk composition of the lunar highlands crust. The NWA 3163 protolith was the likely product of reworked lunar crust with a previous history of heavy REE depletion. The bulk feldspathic and pyroxene-rich fragments have 87Sr/86Sr that are indistinguishable and average 0.699282 ± 0.000007 (2σ). A calculated source model Sr TRD age of 4.340 ± 0.057 Ga is consistent with (1) the recently determined young FAS (Ferroan Anorthosite) age of 4.360 ± 0.003 Ga for FAS 60025, (2) 142Nd model ages for the closure of the Sm–Nd system for the mantle source reservoirs of the Apollo mare basalts (4.355–4.314 Ga) and (3) a prominent age peak in the Apollo lunar zircon record (c. 4.345 Ga). These ages are ∼100 Myr younger than predicted timescales for complete LMO crystallization (∼10 Myrs after Moon formation, Elkins-Tanton et al., 2011). This supports a later, major event during lunar evolution associated with crustal reworking due to magma ocean cumulate overturn, serial magmatism, or a large impact event leading to localized or global crustal melting and/or exhumation. The Ar–Ar isotopic systematics on aliquots of paired stone NWA 4881 are consistent with an impact event at ⩾ 3.5 Ga. This is inferred to record the event that induced granularization of NWA 3163 (and paired rocks). A later event is also recorded at ∼2 Ga by Ar–Ar isotopes is consistent with an increase in the number of impacts on the lunar surface at this time (Fernandes et al., 2013). Northwest Africa 3163 and its paired stones therefore record a c. 2.4 Gyr record of lunar crustal production, metamorphism, brecciation, impacts and eventual ejection from the lunar surface.

The mantle is the major reservoir of Earth's water, hosted within Nominally Anhydrous Minerals (NAMs) (e.g., Bell and Rossman, 1992; Peslier et al., 2010; Peslier, 2010; Nestola and Smyth, 2015), in the form of hydrogen bonded to the silicate's structural oxygen. From whence cometh this water? Is the water in these minerals representative of the Earth's primitive upper mantle or did it come from melting events linked to crustal formation or to more recent metasomatic/re-fertilization events? During diamond formation, NAMs are encapsulated at hundreds of kilometers depth within the mantle, thereby possibly shielding and preserving their pristine water contents from re-equilibrating with fluids and melts percolating through the lithospheric mantle. Here we show that the NAMs included in diamonds from six locales on the Siberian Craton contain measurable and variable H_2O concentrations from 2 to 34 parts per million by weight (ppmw) in olivine, 7 to 276 ppmw in clinopyroxene, and 11–17 ppmw in garnets. Our results suggest that if the inclusions were in equilibrium with the diamond-forming fluid, the water fugacity would have been unrealistically low. Instead, we consider the H_2O contents of the inclusions, shielded by diamonds, as pristine representatives of the residual mantle prior to encapsulation, and indicative of a protogenetic origin for the inclusions. Hydrogen diffusion in the diamond does not appear to have modified these values significantly. The H_2O contents of NAMs in mantle xenoliths may represent some later metasomatic event(s), and are not always representative of most of the continental lithospheric mantle. Results from the present study also support the conclusions of Peslier et al. (2010) and Novella et al. (2015) that the dry nature of the SCLM of a craton may provide stabilization of its thickened continental roots.

Water and other trace element concentrations in olivine (1–39 ppm H2O), orthopyroxene (10–150 ppm H2O), and clinopyroxene (16–340 ppm H2O) of mantle xenoliths from the Labait volcano, located on the edge of the Tanzanian craton along the eastern branch of the East African Rift, record melting and subsequent refertilization by plume magmas in a stratified lithosphere. These water contents are at the lower end of the range observed in other cratonic mantle lithospheres. Despite correlations between water content and indices of melting in orthopyroxene from the shallow peridotites, and in both olivine and orthopyroxene from the deep peridotites, water concentrations are too high for the peridotites to be simple residues. Instead, the Labait water contents are best explained as reflecting interaction between residual peridotite with a melt having relatively low water content (

The distribution of water concentrations in the oceanic upper mantle has drastic influence on its melting, rheology, and electrical and thermal conductivities and yet is primarily known indirectly from analyses of OIB and MORB. Here, actual mantle samples, eight peridotite xenoliths from Salt Lake Crater (SLC) and one from Pali in Oahu in Hawaii were analyzed by FTIR. Water contents of orthopyroxene, clinopyroxene, and the highest measured in olivine are 116–222, 246–442, and 10–26 ppm weight H2O, respectively. Although pyroxene water contents correlate with indices of partial melting, they are too high to be explained by simple melting modeling. Mantle-melt interaction modeling reproduces best the SLC data. These peridotites represent depleted oceanic mantle older than the Pacific lithosphere that has been refertilized by nephelinite melts containing

Constraining the distribution and mobility of H in olivine, the main mineral of the upper mantle, is crucial to our understanding of Earth’s geodynamics because this trace element influences melting, rheology, and electrical and thermal conductivities of peridotite. For this purpose, the olivines from fresh and well-characterized peridotite xenoliths from Salt Lake Crater and Pali (Oahu, Hawaii), representing samples of the oceanic mantle lithosphere, were analyzed by FTIR. Water concentrations decrease from core to edge and near fractures of olivine grains, and are best interpreted as H loss during xenolith ascent to the surface in its host magma. Diffusion modeling of these profiles allowed the calculation of diffusion times, which were in turn used to estimate the average ascent rates of the xenolith host nephelinite at 0.2–25.3 m s−1. These rates are similar to those of continental basaltic magmas. Diffusion modeling further shows that the water contents at the core of olivines are preserved mantle values and are heterogeneous within each xenolith. In addition, the discrepant behavior of the 3225 cm−1 OH band (due to H in a Mg vacancy) relative to the other OH bands (in particular the Ti-H defect) along profiles evidences that H is heterogeneously distributed amongst olivine defects. These defect profiles are modeled to calculate that the diffusion rate of the Mg-H defect is about 1.3–6.8 times faster than that of the Ti-H defect. The heterogeneous distribution of H in the mantle between olivine cores in single xenoliths and within olivine grains testifies of a state of disequilibrium for water in these samples. The Salt Lake Crater peridotite olivines record two processes; recent metasomatism by a melt bringing water followed by water loss during ascent in the host magma, neither having lasted long enough for water to reach equilibrium. The observed decoupling between the heterogeneous distribution of H and the homogeneous distribution of lithophile elements suggests that the process of water addition to the peridotite via incipient melt metasomatism was likely interrupted by the host nephelinite removing the samples from the mantle and bringing them to the surface.

Mapping the compositional variability of the Earth's mantle is fundamental for understanding mantle dynamics, crustal recycling and melt generation. The geochemistry of intraplate oceanic basalts in particular points to a heterogeneous convecting mantle, often explained by variable proportions of fertile (pyroxenite, eclogite) and depleted (peridotitic) domains. A key parameter necessary to constrain the Earth's deep processes is water because it influences melting and affects the physical properties of the Earth's mantle. However, there are only a few direct water determinations on samples of the oceanic mantle. Here we report water concentrations by FTIR on garnet pyroxenite xenoliths from Salt Lake Crater vent, Oahu, Hawaii, in order to constrain the role of lithological variability in the distribution of water in the upper oceanic mantle. The clinopyroxenes have 260 to 576 ppm H2O and the orthopyroxenes about half these amounts. Curiously, garnets have water concentrations below detection limits ( 160). The discrepancy between the low H2O/Ce ratios in the bulk pyroxenites and high H2O/Ce in the equilibrium melts is consistent with experimental data that predicts 4 to 5 times higher partition coefficient for Ce than H in these pyroxenes. Therefore, the process of high-pressure crystallization in the oceanic lithosphere will create pyroxene-rich lithologies which are paradoxically, both “wet” (i.e., high H2O concentrations) and “dry” (low H2O/Ce ratios) compared to the source of their parental melts. Phlogopite is present as a trace phase in these rocks (

The 187Os/188Os compositions in peridotite samples from the sub-continental lithospheric mantle (SCLM) can be used to constrain the timing of melt extraction and potentially test the link between large-scale mantle melting and juvenile crust production. The SCLM has often experienced a complex history such that some lithophile elements such as REEs (rare earth elements) in these rocks typically record overprinting during metasomatism. New 187Os/188Os, major and trace element compositional data were obtained on sixteen Dish Hill peridotite xenoliths (California, USA) and are used to examine these issues. The samples show strong correlations between 187Os/188Os and indicators of melt depletion such as Lu abundance in clinopyroxene, modal abundance of clinopyroxene, bulk rock Al2O3 and the Cr# (Cr/(Cr + Al)) in spinel. These relationships indicate that metasomatism did not compromise the 187Os/188Os systematics. The data appear to form two melt depletion trends consistent with Re depletion model ages (TRD) obtained from the two Al2O3 versus 187Os/188Os trends are 2.1 ± 0.5 Ga and 1.3 ± 0.3 Ga (±95% conf.). It has been suggested that the SCLM under Dish Hill may be fragments of oceanic lithosphere emplaced as the result of Farallon plate subduction during the Late Cretaceous ( Luffi et al., 2009 ). However, the strong melt depletion trends, major element compositions and Re-depletion ages are not consistent with the interpretation of this suite of xenoliths having an oceanic lithospheric origin. Rather, the 2.1 Ga age coincides with Nd model ages of 2–2.3 Ga ( Bennett and DePaolo, 1987 , Ramo and Calzia, 1998 ) for the overlying Mojavia crustal province. The 1.3 Ga age is consistent with large-scale A-type magmatism in the nearby region at this time that is purported to be the result of mantle plume melting processes. Therefore, data from this study point to the SCLM under Dish Hill being formed by two ancient mantle-melting events, which could be the result of interleaving SCLM at depth. These interpretations indicate that the primary mechanism of SCLM formation under Dish Hill was through stabilization following partial melting in the convecting mantle that also results in contemporaneous juvenile crust production.

Water has been detected on the lunar surface and attributed to delivery by impacts and the solar wind to a dry early Moon. Spectroscopic detections of water in lunar anorthosites from the Apollo collection suggest that a significant amount of water is indigenous to the Moon.

International audience; A subject of continuing debate is how the Earth's lithospheric portion of the upper mantle has remained the thickest (> 200 km) and oldest (> 3 Gy) beneath cratons and is yet surrounded by a vigorously convecting asthenosphere. It is generally admitted that water is a key parameter in the strength and longevity of cratonic roots, because olivine, the main phase of the lithospheric mantle, becomes stronger if its water content decreases. Expanding upon the work presented in Novella et al. (2015) and Taylor et al. (2016), we report new water contents for additional olivine inclusions in diamonds together with the trace-element composition for all olivine inclusions, as well as for mantle xenoliths from various kimberlite pipes located on the Siberian craton. The olivine diamond inclusions from this study have systematically low-water contents (< 50 ppmw H2O), moderate to high forsterite (e.g., Fo91–94) contents and low Ni, Co, and Zn ppm contents (e.g., < 2848, < 108, and < 47 ppm, respectively). In contrast, olivines from Siberian craton mantle xenoliths have a wide range of water contents (6–323 ppmw H2O) and extend to lower-Fo (91–92), Ni, Co, and Zn-rich compositions, compared to the diamond inclusions. Depleted incompatible trace-element concentrations in olivine (0.1–0.001 × Primitive Mantle) advance our hypothesis for the protogenetic origins for the majority of Siberian diamond inclusions. These observations are consistent with the peridotite xenoliths as representing a part of the cratonic lithosphere that has experienced melt re-fertilization, which has also transported water. The olivine diamond inclusions, on the other hand, preserve “micro-samples” of an initial, dry cratonic lithosphere, mostly resulting from melting events. These inclusions are likely sourced from the initial cratonic mantle lithosphere, which thereby, resisted delamination over time, due to its buoyancy and strength, imparted from melt and water depletion, respectively. And thus, our data provides a major argument that the kimberlite-hosted mantle xenoliths may be more metasomatized than common rocks at the base of the Siberian and other cratonic roots away from kimberlite fields.

Iron isotopic compositions potentially provide a powerful new tracer of planetary formation and differentiation processes and of secular and spatial changes in mantle oxidation state. However, the processes governing iron isotope fractionation in igneous rocks remain poorly understood. Here we show that there are significant variations in the iron isotope compositions (δ57/54Fe) of mantle rocks (0.9‰) and minerals (olivines 0.6‰, clinopyroxenes 0.97permil; and orthopyroxenes 0.8‰), with spinels showing the greatest total variation of 1.7‰. Positive linear functional relationships with slopes that are, within error, equal to unity are found between the δ57/54Fe values of coexisting orthopyroxene, clinopyroxene and olivine, strongly suggesting that the δ57/54Fe values of these minerals reflect intra-sample mineral-mineral isotopic equilibrium. Positive correlations between the δ57/54Fe values of silicate minerals and spinels also exist, although they are more scattered, which could be caused by late disturbance of mineral-spinel isotopic equilibrium. Bulk-rock, clinopyroxene and spinel δ57/54Fe values correlate with chemical indices of both melt extraction and oxidation. Iron isotope fractionation during spinel-facies partial melting is investigated using simple models, which demonstrate that the maximum expected fractionation between melt and residue will be ∼0.5‰, with the residue becoming isotopically light relative to the melt and to the initial source region. Hence melt extraction, in combination with significant changes in mantle oxidation state, may be an explanation for Fe isotopic variations in mantle peridotites. Metasomatism of the sub-arc mantle by iron-rich silicate melts originating from the subducting slab may also explain the light bulk-sample δ57/54Fe values of some arc peridotites (-0.2‰ to -0.6‰), but mass-balance calculations require these metasomatic agents to have extreme δ57/54Fe values (e.g. -3.0‰). The large differences in the δ57/54Fe values of garnet and spinel facies rocks are likely to be caused by the contrasting behaviour of Fe3+ during melting in the spinel and garnet facies. However, there is little difference in the δ57/54Fe values of MORB and OIB, despite the fact that OIB are considered, on the basis of incompatible element abundances, to arise dominantly by melting in the garnet stability field. Given that iron is a relatively compatible element, the similarities in the δ57/54Fe values of MORB and OIB provide strong evidence that MORB and OIB are both dominated by melting in the spinel facies. © 2005 Elsevier B.V. All rights reserved.

The oxygen fugacity of the mantle exerts a fundamental influence on mantle melting, volatile speciation, and the development of the atmosphere. However, its evolution through time is poorly understood. Changes in mantle oxidation state should be reflected in the Fe 3+ /Fe 2+ of mantle minerals, and hence in stable iron isotope fractionation. Here it is shown that there are substantial (1.7 per mil) systematic variations in the iron isotope compositions (δ 57/54 Fe) of mantle spinels. Spinel δ 57/54 Fe values correlate with relative oxygen fugacity, Fe 3+ /ΣFe, and chromium number, and provide a proxy of changes in mantle oxidation state, melting, and volatile recycling.

Water and trace element contents were measured by FTIR and laser ablation-ICPMS on minerals from peridotite xenoliths in kimberlites of the Kaapvaal craton from Finsch, Kimberley, Jagersfontein (South Africa), Letseng-La-Terae, and Liqhobong (Lesotho) mines. The peridotites record a wide range of pressure, temperature, oxygen fugacity, and metasomatic events. Correlations between water content or OH vibration bands with major, minor and trace elements in pyroxene and garnet precludes disturbance during xenolith entrainment by the host kimberlite magma and indicate preservation of mantle water contents. Clinopyroxene water contents (150–400 ppm H2O, by weight) correlate with those of orthopyroxene (40–250 ppm). Olivines ( Peslier et al., 2008 , Peslier et al., 2010 ) and garnets have 0–86 and 0–20 ppm H2O, respectively. Relations in individual xenolith suites between the amount of water and that of incompatible elements Ti, Na, Fe3+ and rare earths in minerals suggests that metasomatism by oxidizing melts controls the water content of olivine, pyroxene and garnet. At pressures ⩽5.5 GPa, hydrous, alkaline, siliceous fluids or melts metasomatized Liqhobong and Kimberley peridotites, producing high water contents in their olivine, pyroxenes and garnet. At higher pressures, the percolation of ultramafic melts reacting with peridotite resulted in co-variation of Ca, Ti and water at the edge of garnets at Jagersfontein, and the overall crystallization of garnet with lower water contents than those in the original peridotites. The upward migration of these ultramafic melts through the lithospheric mantle also increased the water content of olivines with decreasing pressure at Finsch Mine. H2O/Ce ratios of melts in equilibrium with Kaapvaal peridotites range from 100 to 20,000 and the larger values may indicate metasomatism in subduction zone settings. Metasomatic events in Kaapvaal peridotites are thought to have occurred from the Archean to the Mesozoic. However, circumstantial evidence suggest that the metasomatic events responsible for setting the water contents may date from the Archean at Kimberley and from the Proterozoic at Jagersfontein. Combined water with Lu–Hf and Sm–Nd isotopic data at Finsch ( Lazarov et al., in press-a ) and with Ar–Ar phlogopite ages at Liqhobong ( Hopp et al., 2008 ) indicates that water addition by metasomatic melts occurred in the Proterozoic. Water contents of mantle minerals in Kaapvaal xenoliths measured here have been preserved since that time and can consequently be used in modelling viscosity and longevity of cratonic roots since at least the mid-Proterozoic.

Siderophile element concentrations were measured by LA-ICP-MS in metals and sulfides from five aubrite meteorites. Siderophile element patterns in aubrites are either similar to those in metal from enstatite chondrites, or can be derived by crystallization from metallic liquids derived by partial melting of E chondrites. Some metal grains in Mt. Egerton, Cumberland Falls, and Aubres show moderate to severe depletion in compatible highly siderophile elements (Re, Os, Ir, Ru) which are consistent with solid metal/liquid metal differentiation of enstatite chondrite-like metal. Metals from chondrite inclusions in Cumberland Falls show more extremely fractionated patterns than those from the aubritic matrix, potentially hinting at fractionation and partial melting processes affecting not only the aubrite parent body, but the chondrite body from which the inclusions were derived as well. Models using experimental partition coefficients show that aubrite metal chemically corresponds to solid metal segregated during differentiation of primary metallic liquids of EH/EL composition that contained both substantial S- and C-contents. This result is consistent with a genetic link between enstatite chondrites and aubrites, but as to whether aubrites were derived from the same body(ies) as enstatite chondrites, or have their origin in multiple, and potentially separated bodies, cannot be answered unequivocally with chemical or isotopic data alone.

– Northwest Africa (NWA) 5298 is an evolved basaltic shergottite that has bulk characteristics and mineral compositions consistent with derivation from an oxidized reservoir in Mars. Chemically zoned clinopyroxene (64.5%, augite and pigeonite), with interstitial lath-shaped plagioclase (29.4%, An40 to An55), constitutes the bulk of this meteorite. The plagioclase has been converted by shock to both isotropic maskelynite and spherulitic, birefringent feldspar representing a quenched vesicular melt. The remainder of the rock consists of minor amounts of Fe-Ti oxides (ilmenite and titanomagnetite), phosphates (merrillite and apatite), silica polymorph, fayalite, pyrrhotite, baddeleyite, and minor hot desert weathering products (calcite and barite). Oxygen fugacity derived from Fe-Ti oxide thermobarometry is close to the quartz-fayalite-magnetite (QFM) buffer indicating that the late stage evolution of this magma occurred under more oxidizing condition than those recorded in most other shergottites. Merrillite contains the largest abundances of rare earth elements (REE) of all phases, thereby controlling the REE budget in NWA 5298. The calculated bulk rock REE pattern normalized to CI chondrite is relatively flat. The evolution of the normalized REE patterns of the bulk rock, clinopyroxene, plagioclase, and phosphate in NWA 5298 is consistent with closed-system chemical behavior with no evidence of crustal contamination or postcrystallization disturbance of the REE contents of these phases.

Combined 147Sm–143Nd and 176Lu–176Hf chronology of the martian meteorite Larkman Nunatak (LAR) 06319 indicates an igneous crystallization age of 193 ± 20 Ma (2σ weighted mean). The individual 147Sm–143Nd and 176Lu–176Hf internal isochron ages are 183 ± 12 Ma and 197 ± 29 Ma, respectively, and are concordant with two previously determined 147Sm–143Nd and 87Rb–87Sr internal isochron ages of 190 ± 26 Ma and 207 ± 14 Ma, respectively ( Shih et al., 2009 ). With respect to the 147Sm–143Nd isotope systematics, maskelynite lies above the isochron defined by primary igneous phases and is therefore not in isotopic equilibrium with the other phases in the rock. Non-isochronous maskelynite is interpreted to result from shock-induced reaction between plagioclase and partial melts of pyroxene and phosphate during transformation to maskelynite, which resulted in it having unsupported 143Nd relative to its measured 147Sm/144Nd ratio. The rare earth element (REE) and high field strength element (HFSE) compositions of major constituent minerals can be modeled as the result of progressive crystallization of a single magma with no addition of secondary components. The concordant ages, combined with igneous textures, mineralogy, and trace element systematics indicate that the weighted average of the radiometric ages records the true crystallization age of this rock. The young igneous age for LAR 06319 and other shergottites are in conflict with models that advocate for circa 4.1 Ga crystallization ages of shergottites from Pb isotope compositions, however, they are consistent with updated crater counting statistics indicating that young volcanic activity on Mars is more widespread than previously realized ( Neukum et al., 2010 ).

Olivine, pyroxene and garnet are nominally anhydrous but can accommodate tens to hundreds of parts per million (ppm) H2O or “water” in the form of protons incorporated in defects in their mineral structure. This review concerns the amount of water in nominally anhydrous minerals from mantle and mantle-derived rocks: peridotites, eclogites, megacrysts, basalts and kimberlites. Trends between internal and external parameters and water content in olivine, pyroxene, and garnet of mantle rocks allow us to identify what controls their H intake. The water content of pyroxenes and garnets in peridotites appears to depend primarily on mineral chemistry, while that of olivines may more readily reflect water activity and water fugacity conditions in the mantle. In eclogites, both mineral chemistry and metamorphic pressure control the water intake of pyroxene and garnet. The water content of minerals in crystallized melts (basalt and kimberlite phenocrysts, xenocrysts, and megacrysts) is determined by the degree of differentiation, the amount of water in the parent melt, and degassing. Basalt and cumulate minerals from Martian meteorites may be as water-rich as their Earth equivalents. No water has been detected at present in Moon minerals but low amounts in degassed basaltic glasses signify that deep Moon reservoirs may still retain water. The presence of water in mantle minerals, typically amounts of tens to hundreds of ppm, enhances their deformation properties. Water contents of peridotite minerals in the oceanic upper mantle, in and around cratons, and in subduction zones may have tremendous influence on Earth's geodynamics.

Cratons, the ancient cores of continents, extend laterally for hundreds of kilometres, and are underlain to depths of 180 km to 250 km by mantle roots that are chemically and physically distinct from the surrounding mantle. But how these large portions of the mantle can stay isolated for so long from mantle convection has remained an open question. Anne Peslier and colleagues show that olivine components in peridotite xenoliths from the lithosphere–asthenosphere boundary region of the Kaapvaal Craton mantle root are water-poor, thereby providing sufficient viscosity contrast with the underlying asthenosphere to explain how they have resisted recycling into Earth's mantle. Cratons, the ancient cores of continents, extend laterally for hundreds of kilometres, and are underlain to depths of 180–250 km by mantle roots that are chemically and physically distinct from surrounding mantle. But how can these roots stay so isolated from mantle convection? Here it is shown that olivine in peridotite xenoliths from the lithosphere–athenosphere boundary region of the Kaapvaal craton mantle root is water-poor, providing sufficient viscosity contrast with the underlying asthenosphere to explain the root's stability. Cratons, the ancient cores of continents, contain the oldest crust and mantle on the Earth (>2 Gyr old)1. They extend laterally for hundreds of kilometres, and are underlain to depths of 180–250 km by mantle roots that are chemically and physically distinct from the surrounding mantle2,3,4. Forming the thickest lithosphere on our planet, they act as rigid keels isolated from the flowing asthenosphere5; however, it has remained an open question how these large portions of the mantle can stay isolated for so long from mantle convection. Key physical properties thought to contribute to this longevity include chemical buoyancy due to high degrees of melt-depletion and the stiffness imparted by the low temperatures of a conductive thermal gradient2,6,7. Geodynamic calculations, however, suggest that these characteristics are not sufficient to prevent the lithospheric mantle from being entrained during mantle convection over billions of years6,7. Differences in water content are a potential source of additional viscosity contrast between cratonic roots and ambient mantle owing to the well-established hydrolytic weakening effect in olivine8,9,10, the most abundant mineral of the upper mantle. However, the water contents of cratonic mantle roots have to date been poorly constrained. Here we show that olivine in peridotite xenoliths from the lithosphere–asthenosphere boundary region of the Kaapvaal craton mantle root are water-poor and provide sufficient viscosity contrast with underlying asthenosphere to satisfy the stability criteria required by geodynamic calculations9. Our results provide a solution to a puzzling mystery of plate tectonics, namely why the oldest continents, in contrast to short-lived oceanic plates, have resisted recycling into the interior of our tectonically dynamic planet.

The Larkman Nunatak (LAR) 06319 olivine-phyric shergottite is composed of zoned megacrysts of olivine (Fo 76–55 from core to rim), pyroxene (from core to rim En 70 Fs 25 Wo 5 , En 50 Fs 25 Wo 25 , and En 45 Fs 45 Wo 10 ), and Cr-rich spinel in a matrix of maskelynite (An 52 Ab 45 ), pyroxene (En 30–40 Fs 40–55 Wo 10–25 ,), olivine (Fo 50 ), Fe–Ti oxides, sulfides, phosphates, Si-rich glass, and baddeleyite. LAR 06319 experienced equilibration shock pressures of 30–35 GPa based on the presence of localized shock melts, mechanical deformation of olivine and pyroxene, and complete transformation of plagioclase to maskelynite with no relict birefringence. The various phases and textures of this picritic basalt can be explained by closed system differentiation of a shergottitic melt. Recalculated parent melt compositions obtained from melt inclusions located in the core of the olivine megacrysts (Fo >72 ) resemble those of other shergottite parent melts and whole-rock compositions, albeit with a lower Ca content. These compositions were used in the MELTS software to reproduce the crystallization sequence. Four types of spinel and two types of ilmenite reflect changes in oxygen fugacity during igneous differentiation. Detailed oxybarometry using olivine-pyroxene-spinel and ilmenite-titanomagnetite assemblages indicates initial crystallization of the megacrysts at 2 log units below the Fayalite-Magnetite-Quartz buffer (FMQ – 2), followed by crystallization of the groundmass over a range of FMQ – 1 to FMQ + 0.3. Variation is nearly continuous throughout the differentiation sequence. LAR 06319 is the first member of the enriched shergottite subgroup whose bulk composition, and that of melt inclusions in its most primitive olivines, approximates that of the parental melt. The study of this picritic basalt indicates that oxidation of more than two log units of FMQ can occur during magmatic fractional crystallization and ascent. Some part of the wide range of oxygen fugacities recorded in shergottites may consequently be due to this process. The relatively reduced conditions at the beginning of the crystallization sequence of LAR 06319 may imply that the enriched shergottite mantle reservoir is slightly more reduced than previously thought. As a result, the total range of Martian mantle oxygen fugacities is probably limited to FMQ − 4 to − 2. This narrow range could have been generated during the slow crystallization of a magma ocean, a process favored to explain the origin of shergottite mantle reservoirs.

Less Old Martian Meteorite The oldest Martian meteorite known, ALH84001, was thought to be a remnant of primordial martian crust formed during solidification of an early magma ocean. Using isotope data, Lapen et al. (p. 347 ) revised the crystallization age of this meteorite from 4.51 billion years to 4.09 billion years ago, meaning that this rock cannot be a fragment of primordial crust that escaped the period of intense bombardment that occurred between 4.25 and 4.10 billion years ago. The revised age also suggests that magmatism was ongoing in Mars for a large part of its history and that ALH84001 was actually formed during the heavy bombardment period, just before the martian core dynamo stopped and the planetary magnetic field was lost.

Fourier transform infrared spectrometry (FTIR) analyses of olivines from peridotite xenoliths found in southern African kimberlites indicate 0 to 80 ppm H2O concentrations. OH absorbance profiles across olivine grains show homogeneous H contents from core to edge for most samples. In one sample the olivines are H-free, while another has olivines characterized by lower H contents at the grain edges compared to the cores, indicating H loss during transport of the xenolith to the surface. Flat or near-flat H profiles place severe constraints on the duration of H loss from olivine grains, with implications for kimberlite magma ascent rates. Diffusion equations were used to estimate times of H loss of about 4 h for the sample with heterogeneous olivine H contents. Resulting kimberlite ascent rates are calculated to be 5–37 m s−1 minimum, although these estimates are highly dependent on volatile contents and degassing behavior of the host kimberlite magma. Xenolithic olivines from alkali basalts generally have lower H contents and more pronounced H diffusion profiles than do those from kimberlites. This difference is likely caused by higher magma temperatures and lower ascent rates of alkali basalts compared to kimberlites.

Olivines in spinel-peridotite mantle xenoliths from Simcoe (Washington State, USA) and Mexico were analyzed by Fourier-transform infrared spectrometry (FTIR) to determine their water contents. The main OH absorbance peaks of most samples are located between 3600 and 3450 cm − 1 (Group I), with a few samples having minor peaks between 3450 and 3100 cm − 1 (Group II). Olivines from one Mexican sample have larger peaks in Group II than in Group I. Most of these OH peaks are predicted by experimental data from the literature in the appropriate range of silica activities and iron contents. A few high-forsterite olivines, however, have mainly Group I peaks which at these low iron contents is characteristic of low-silica activity. Because these olivines coexist with orthopyroxene in the peridotite, buffering silica activity at relatively high values, their FTIR spectra may reflect disturbance of their hydrogen by melts or fluids, most probably associated with the host magma. In eight out of nine samples for which measurement at the olivine edges was possible, water contents are higher in the grain centers than at their edges, with cross-sections showing typical diffusion profiles. Moreover, water concentrations in some samples increase with olivine size. Loss of hydrogen from the olivine during xenolith transport to the surface is likely responsible for these variations. These water-concentration gradients allowed calculation of the duration of hydrogen loss, which ranges from 18 to 65 h, corresponding to host mafic-alkalic magma ascent rates of 0.2–0.5 m s − 1 . The highest measured water contents in olivines from individual xenoliths range from 0 to 6.8 ppm and increase with those of clino- and orthopyroxenes. Differences in hydrogen partition coefficients between olivine and pyroxenes from our data and from experiments suggest that the analyzed olivines lost at least 40% of their water during ascent from the mantle. Olivine water contents do not correlate with partial melting indices, but samples with high olivine water contents generally have low clinopyroxene La/Yb ratios and low spinel Fe 3+ /ΣFe ratios and resultant oxygen fugacities, and vice-versa. Metasomatism by fluids or melts and the ambient oxygen fugacity of the mantle may have played roles in the original incorporation of hydrogen into these olivines, but such primary signals have probably been obscured by later hydrogen loss. The systematically lower water contents of olivines in Mexican and Simcoe xenoliths relative to those from cratonic xenoliths may mainly reflect lower host-magma ascent velocities for mafic alkalic magmas compared to kimberlites. Calculated whole-rock water contents for the studied spinel-peridotite xenoliths range from 2.5 to 154 ppm. If 150 ppm were representative of the water content in the entire upper mantle (to 410 km), the amount of water stored there can be speculated to be only about 0.06 times the equivalent mass of Earth's oceans.

Seven alkali basalt centers in the southern Canadian Cordillera INTRODUCTION contain mantle xenolith suites that comprise spinel Cr-diopside Spinel peridotite xenoliths occur in many alkali basalt peridotites, spinel augite-bearing wehrlites and orthopyroxene-poor centers of the Canadian Cordillera, a Phanerozoic conlherzolites, and minor pyroxenites. The Cr-diopside peridotites vergent orogen bordering the western edge of the Caappear to be residues of the extraction of Mg-rich basalts by up to nadian craton. The Cordillera formed during the late 15% partial melting (median 5–10%) of a pyrolite-like source in Jurassic and Cretaceous through the successive accretion the spinel stability field. The xenoliths are similar to other mantle of tectonic terranes of various lithologies (mainly arc xenolith suites derived from beneath convergent continental margins, related) to the stable margin of ancestral North America but are less depleted, less oxidized, and have lower spinel mg(e.g. Monger et al., 1982; Gabrielse & Yorath, 1991). number than peridotites found in fore-arc settings. Their dominant However, the nature of the lithospheric mantle beneath high field strength element depleted character, however, is typical of these terranes and its tectonic origin are poorly conarc lavas, and may suggest that fluids or melts circulating through strained. Moreover, the southern part of this continental the Canadian Cordillera lithosphere were subduction related. Modmargin is bordered by an active subduction zone (Gaeling using MELTS is consistent with the augite-bearing xenoliths brielse et al., 1991), and thus the lithosphere may have being formed by interaction between crystallizing alkaline melts and been modified by subduction-related melts or fluids. The peridotite. Assimilation–fractional crystallization modeling suggests mantle xenoliths thus represent samples of the lithospheric that the trace element patterns of liquids in equilibrium with the mantle at the transition between cratonic continental augite xenoliths may represent the initial melts that reacted with the lithosphere and oceanic mantle. The major element peridotite. Moreover, the compositions of these melts are similar to composition of the mantle lithosphere is of interest as a those of some glasses observed in the mantle xenoliths. Melt–rock potential source of magmas and also because it may react interaction may thus be a viable mechanism for the formation of with magmas migrating through it on their way to the Siand alkali-rich glass in peridotites. surface.

Correction to: Space Sci Rev DOI This article has been corrected. Figure 3 was initially published with erroneous axis titles in Fig. 3B and 3D where the $x$ axis should be Cpx H2O, instead of Grt H2O.

Siberian cratonic mantle; International audience; The processes that control water distribution in nominally anhydrous minerals from peridotites are twofold. Melt depletion will remove water while metasomatism can potentially add water to these minerals. These processes can lead to a wide range of outcomes in water contents, which in turn could play a role in mantle rheology and long-term cratonic root stability. To examine these complexities, water concentrations in minerals from well-characterized peridotites from the Udachnaya kimberlite in the central Siberian craton were analyzed by FTIR. The peridotites span a complete top to bottom cross-section of typical cratonic lithospheric mantle (2-7 GPa and 700-1400 °C). Diffusion modeling of water content profiles across olivine grains shows that water loss during decompression is limited to the 100 μm rims of olivines; the cores preserved their mantle water contents. Water contents range from 6 to 323 ppm wt H2O in olivine, 28-301 ppm H2O in orthopyroxene (opx), 100-272 ppm H2O in clinopyroxene (cpx) and 0-23 ppm H2O in garnet. Melting modeling cannot reproduce the high water contents of cratonic mantle peridotites and any potential partial melting trend must have been erased by later events. The water contents of minerals, however, are correlated with modal abundances of clinopyroxene and garnet, bulk rock FeO, TiO2 and SiO2 as well as with light and middle rare earth elements in clinopyroxene and garnet. These relationships are best interpreted as interaction of residual, melt-depleted peridotites with silicate melt, which produced modal and cryptic metasomatism. Importantly, the water enrichment in the Siberian cratonic mantle took place prior to kimberlite magmatism and eruption. Water addition by metasomatism occurred from pressures >4 GPa all the way to the base of the cratonic root below central Siberia, but was limited to shallower levels (

Osmium isotope data from lherzolite and harzburgite xenoliths of the northern Canadian Cordil- lera provide constraints on the genesis and age of the lithospheric mantle in a typical off-cratonic continental setting. The 187 Os/ 188 Os ratios of the lherzolites show a positive correlation with Al 2O3 and heavy rare earth elements (HREE), which probably reflects Re/Os fractionation during various degrees of mantle melting, followed by a long period of radiogenic ingrowth. These observations are consistent with melting during the Proterozoic. Harzburgite Os isotopic ratios, however, plot above the regional correlation of the lherzolites. A positive correlation between their Os isotopic ratios and 1/Os concentrations suggests that they are the end result of the introduction of metasomatic agents with low Os contents, but high 187 Os/ 188 Os ratios, into the lithospheric mantle. These fluids or melts may have originated from a region of anomalously slow mantle detected seismically (Frederiksen et al., 1998) below harzburgite-rich xenolith localities (Shi et al., 1998). Alternatively, the radiogenic Os-bearing metasomatic agents may have been related to subduction processes along the western margin of the Canadian Cordillera, as has been previously suggested to explain the high Os isotopic ratios of xenoliths from the northern US Cordillera (Brandon et al., 1996). Copyright © 2000 Elsevier Science Ltd

The 187 Os / 188 Os ratios of lherzolites from eight xenolith suites from the Canadian Cordillera show a correlation with Al2O3 and heavy rare earth elements (HREE). The best interpretation of these correlations appears to be ancient melt depletion followed by a long period of radiogenic ingrowth. The 187 Os / 188 Os–Lu correlation is used to calculate an Os model age of 1.12±0.26 Ga for the lithospheric mantle throughout the Canadian Cordillera. This single melting age suggests that the mantle lithosphere now underlying the entire Canadian Cordillera may have formed by melting events closely spaced in time. This is consistent with seismic evidence of the extension of crustal basement under much of the Canadian Cordillera that is independent of the upper-crustal terranes overlying it. Indeed, this Proterozoic Os model age for the mantle contrasts with the younger formation ages (Nd model ages and U–Pb ages of zircons) of most crustal terranes of the region which are around 0.5 Ga. Early Proterozoic basement is exposed only in southeastern British Columbia and has the same age (1.9 to 2.3 Ga) as the ancestral North American crust, but is older than the Os model age of the mantle lithosphere underlying the Canadian Cordillera. The Canadian Cordilleran mantle is thus probably not a simple extension of the North American cratonic lithosphere beneath the adjacent mobile orogenic belt of the Canadian Cordillera. The difference in age between the formation of the Canadian Cordillera upper-crust and the formation of the underlying mantle suggests that this mantle lithosphere does not represent the mantle roots of the crustal terranes overlying it. Instead, these crustal terranes were thrust onto the mantle lithosphere during Canadian Cordillera orogeny. This contrasts strongly with Archean cratonic zones and Early Proterozoic belts where oldest crustal rocks and mantle may have the same formation age.

[1] Nominally anhydrous minerals (e.g., olivine, clinopyroxene, and orthopyroxene) in peridotite xenoliths collected from the Colorado Plateau and southern Basin and Range in western North America were systematically analyzed by Fourier transform infrared spectroscopy for water contents. Measured water contents range from 2 to 45 ppm for olivine, from 53 to 402 ppm for orthopyroxene, and from 171 to 957 ppm for clinopyroxene. The Colorado Plateau has the highest water contents (up to 45 ppm H2O in olivine, 402 ppm H2O in orthopyroxene, and 957 ppm H2O in clinopyroxene), while San Carlos in the southern Basin and Range has the lowest water contents (up to 4 ppm H2O in olivine, 82 ppm H2O in orthopyroxene, and 178 ppm H2O in clinopyroxene). With the exception of San Carlos, the olivine and pyroxenes from all other localities (Dish Hill, Grand Canyon, and Navajo) have water contents close to or higher than that inferred for the fertile asthenospheric mantle. We interpret the high water contents measured here to have been introduced into the base of the lithospheric mantle by rehydration associated with the subduction of the Farallon plate beneath North America during the early Cenozoic. Application of an updated flow law for dislocation creep of wet olivine to lithospheric mantle conditions beneath the Colorado Plateau predicts that for a given background shear stress, hydration alone can result in approximately 1 order of magnitude drop in the effective viscosity at the base of the lithosphere. If viscosity alone is used to distinguish the lithosphere from underlying asthenosphere, this suggests that hydration could have resulted in more than 10 km of lithospheric thinning. Viscosity reduction and lithospheric thinning of even larger extents (up to ∼100 km) are predicted when thicker lithosphere (such as Archean cratons) and larger water contents (up to water-saturated conditions) are considered. If our interpretations are correct, the implications of our study go beyond western North America and hint at a possible way of recycling continental mantle, including cratonic mantle, back into the convecting mantle.