Volume 104, Issue 2

The Re-Os isotopic system is largely considered the geochronometer of choice to date partial melting of terrestrial peridotites and in constraining the evolution of Earth’s dynamics from the mantle viewpoint. While whole-rock peridotite Re-Os isotopic signatures are the core of such investigations, the Re-Os dating of individual peridotite minerals—base metal sulfides (BMS) and platinum group minerals (PGM)—that are the main hosts for Re and Os in the mantle peridotites came into play two decades ago. These nanometric-micrometric BMS and PGM display an extreme complexity and heterogeneity in their 187 Os/ 188 Os and 187 Re/ 188 Os signatures that result from the origin of the BMS±PGM grains (residual vs. metasomatic), the nature of the metasomatic agents, the transport/precipitation mechanisms, BMS±PGM mineralogy, and subsequent Re/Os fractionation. Corresponding whole-rock host peridotites, typically plot within the 187 Os/ 188 Os and 187 Re/ 188 Os ranges defined by the BMS±PGM, clearly demonstrating that their Re-Os signatures represent the average of the different BMS ±PGM populations. The difference between the 187 Os/ 188 Os ratios of the least radiogenic BMS±PGM and the respective host peridotite increases with the fertility of the peridotite reflecting the increasing contribution of metasomatic BMS±PGM to the whole-rock mass balance of Re and Os concentrations and Os isotope compositions. Corollaries to these observations are that (1) BMS may provide a record of much older partial melting event, pushing back in time the age of the lithospheric mantle stabilization, (2) if only whole-rock peridotite Re-Os isotopic measurements are possible, then the best targets for constraining the timing of lithospheric stabilization are BMS-free/BMS-poor ultra-refractory spinel-bearing peridotites with very minimal metasomatic overprint, as their 187 Os/ 188 Os signatures may be geologically meaningful, (3) while lherzolites are “fertile” in terms of their geochemical composition, they do not have a “primitive,” unmodified composition, certainly in terms of their highly siderophile elements (HSE) and Re-Os isotopic systematics, and (4) the combined Re-Os isotopic investigations of BMS and whole-rock in BMS-rich mantle peridotites would provide a complementary view on the timing and nature of the petrological events responsible for the chemical and isotopic evolution and destruction of the lithospheric mantle. In addition, the 187 Os/ 188 Os composition of the BMS±PGM (both residual and metasomatic) within any single peridotite may define several age clusters—in contrast to the single whole-rock value—and thus provide a more accurate picture of the complex petrogenetic history of the lithospheric mantle. When coupled with a detailed BMS±PGM petrographical study and whole-rock lithophile and HSE systematics, these BMS age clusters highlight the timing and nature of the petrological events contributing to the formation and chemical and isotopic evolution of the lithospheric mantle. These BMS±PGM age clusters may match regional or the local crustal ages, suggesting that the formation and evolution of the lithospheric mantle and its overlying crust are linked, providing mirror records of their geological and chemical history. This is, however, not a rule of thumb as clear evidence of crust-mantle age decoupling also exist. Although the BMS ±PGM Re-Os model ages push back in time the stabilization of lithospheric mantle, the dichotomy between Archean cratonic and circum-cratonic peridotites, and post-Archean non-cratonic peridotites and tectonites is preserved. This ability of BMS±PGM to preserve older ages than their host peridotite also underscores their survival for billions of years without being reset or reequilibrated despite the complex petrogenetic processes recorded by their host mantle peridotites. As such, they are the mantle equivalents of crustal zircons. Preservation of such old signatures in “young” oceanic peridotites ultimately rules out the use of the Re-Os signatures in both oceanic peridotites and their BMS to estimate the timescales of isotopic homogenization of the convecting mantle.

In the mantle, base metal sulfides have been proposed as the main host for many chalcophile and siderophile elements. This includes elements such as Pb, Se, and Te, which are often used as tracers of processes ranging from planetary accretion to mantle melting. We present in situ measurements of these elements, along with As, Sb, Ag, Au, and Cl, in abyssal peridotite sulfides to provide constraints on the storage of these elements in the mantle. A total of 152 sulfides from 11 peridotites and 1 pyroxenite from the Gakkel and Southwest Indian ridges were analyzed. The sulfides are pentlandites, some of which contain either discrete chalcopyrite domains or Cu-rich intergrowths. Trace-element concentrations in 108 unaltered sulfides range from 2 to 36 ppm Pb, 45 to 250 ppm Se, <4 to 360 ppm Te, <1.5 to 1900 ppm As, 2 to 420 ppm Sb, 2 to 340 ppm Ag, 2 to 770 ppb Au, and 0.2 to 1000 ppm Cl. Tellurium abundances are highly variable within sulfides, which is likely due to the presence of telluride micro- or nano-phases. Based on morphology, composition, and the absence of monosulfide solid solution, the sulfides are interpreted to have formed by fractional crystallization from sulfide melt during conductive cooling of the mantle beneath the ridge axis. The average sulfide Pb concentration of 4 ppm can be reproduced by >90% fractional crystallization from a sulfide melt. The remaining sulfide melt, which is modeled to contain 800 ppm Pb, will dissolve into silicate melt as it rises through the mantle due to the increasing solubility of sulfur in silicate melt as pressure decreases. However, the amount of sulfide melt that remains after fractional crystallization is too low (mode of <0.005%) to contribute a significant amount of Pb to mid-ocean ridge basalts. We conclude that sulfides are not the main host for mantle Pb, even prior to the onset of any melting, and that the majority of mantle Pb is stored in silicate phases.

Hibonite (CaAl 12 O 19 ) is a constituent of some refractory calcium-aluminum inclusions (CAIs) in carbonaceous meteorites, commonly accompanied by grossite (CaAl 4 O 7 ) and spinel. These phases are usually interpreted as having condensed, or crystallized from silicate melts, early in the evolution of the solar nebula. Both Ca-Al oxides are commonly found on Earth, but as products of high-temperature metamorphism of pelitic carbonate rocks. We report here a unique occurrence of magmatic hibonitegrossite-spinel assemblages, crystallized from Ca-Al-rich silicate melts under conditions [high-temperature, very low oxygen fugacity ( f O2 )] comparable to those of their meteoritic counterparts. Ejecta from Cretaceous pyroclastic deposits on Mt Carmel, N. Israel, include aggregates of hopper/skeletal Ti-rich corundum, which have trapped melts that crystallized at f O2 extending from 7 log units below the iron-wustite buffer (ΔIW = –7; SiC, Ti 2 O 3 , Fe-Ti silicide melts) to ΔIW ≤ –9 (native V, TiC, and TiN). The assemblage hibonite + grossite + spinel + TiN first crystallized late in the evolution of the melt pockets; this hibonite contains percentage levels of Zr, Ti, and REE that reflect the concentration of incompatible elements in the residual melts as corundum continued to crystallize. A still later stage appears to be represented by coarse-grained (centimeter-size crystals) ejecta that show the crystallization sequence: corundum + Liq → (low-REE) hibonite → grossite + spinel ± krotite → Ca 4 Al 6 F 2 O 12 + fluorite. V 0 appears as spheroidal droplets, with balls up to millimeter size and spectacular dendritic intergrowths, included in hibonite, grossite, and spinel. Texturally late V 0 averages 12 wt% Al and 2 wt% Mn. Spinels contain 10–16 wt% V in V 0 -free samples, and <0.5 wt% V in samples with abundant V 0 . Ongoing paragenetic studies suggest that the f O2 evolution of the Mt Carmel magmatic system reflects the interaction between OIB-type mafic magmas and mantle-derived CH 4 +H 2 fluids near the crust-mantle boundary. Temperatures estimated by comparison with 1 atm phase-equilibrium studies range from ca. 1500 °C down to 1200–1150 °C. When f O2 reached ca. ΔIW = –7, the immiscible segregation of Fe,Ti-silicide melts and the crystallization of SiC and TiC effectively desilicated the magma, leading to supersaturation in Al 2 O 3 and the rapid crystallization of corundum, preceding the development of the hibonite-bearing assemblages. Reports of Ti-rich corundum and SiC from other areas of explosive volcanism suggest that these phenomena may be more widespread than presently realized, and the hibonite-grossite assemblage may serve as another indicator to track such activity. This is the first reported terrestrial occurrence of krotite (CaAl 2 O 4 ), and of at least two unknown Zr-Ti oxides.

The plasticity of the mantle is still not well constrained, and satisfactory mineral-physics-based rheological laws are still missing. Despite olivine being the major component of the upper mantle, it is still debated which deformation mechanism (dislocation creep, diffusion creep, grain boundary sliding) dominates deformation. High-pressure research developments (state-of-the-art presses, synchrotron experiments, and so on) as well as competitive analysis utilities (software analysis, microscopy, and so on) allow considering intra- and intergranular mechanisms (grain boundary sliding accommodated by diffusion/dislocation creep) simultaneously. To study the contribution of individual deformation mechanism to the overall deformation in the upper mantle, we deformed polycrystalline forsterite at 3.5–5.0 GPa, 1000–1200 °C, 2 × 10 –5 s –1 at different strains in a 6-axis Mavo press. Split-cylinder experiments allowed to characterize an “internal” surface of the sample before and after the deformation experiments. Intra- and intergranular deformation was tracked using a focus ion beam milled reference grid on this surface. Grain internal misorientation where obtained from electron backscatter diffraction (EBSD) data. Both techniques suggest the dominance of intragranular deformation, in agreement with the fact that the samples have been deformed in the dislocation creep regime, as usually defined. Moreover, strain markers and out-of-plane displacements of grains provide the first microstructural evidence for a contribution of grain boundary sliding to plastic deformation at upper mantle pressure. Whether these displacements are grain boundary sliding or involve grain boundary migration cannot be clarified, given the resolution of the strain markers. Our EBSD data suggest that grain boundary processes become increasingly relevant at temperatures above 1100 °C and ensure homogenous plastic strain distribution in the aggregate.

A full complement of standard state thermodynamic properties (ΔfG298.1∘,ΔaGT,i∘,   S298.1∘,  and  CP∘)$\left( {{\Delta }_{f}}G_{298.1}^{\circ },{{\Delta }_{a}}G_{\text{T,i}}^{\circ },\,\,\,\text{S}_{298.1}^{\circ },\,\,\text{and}\,\,C_{P}^{\circ } \right)$ has been determined for a magnesian chamosite [Fe-Chl(W)] and a ferroan clinochlore [Mg-Chl] investigated by calorimetry and low-temperature hydrothermal experiments; this makes these two samples the only natural chlorites whose complete set of thermochemical properties have been reported. ΔfG298.1∘${{\Delta }_{f}}G_{298.1}^{\circ }$ for Mg-Chl and Fe-Chl (W) have been determined to be –8161.76 ± 32.50 and –7278.97 ± 21.50 kJ/mol, respectively. Ternary molecular chlorite solid solution modeling approaches have been developed for Al-rich and Si-rich chlorites; unlike available atomic site-mixing chlorite solid-solution models, a molecular model obviates the need for the adoption of a putative structural chemistry. The calculated excess entropy of mixing in the ternary system exhibits a curvilinear dependence on composition and at 25 °C, G ex ss vary from about –72 to 413 kJ/mol implying a significant deviation from ideality. The effect of di-trioctahedral substitutions was evaluated by modeling the solid solutions in the quaternary amesite-chamosite-clinochlore-sudoite system for aluminous chlorites; excess functions (S ex , G ex ) calculated for these quaternary and ternary solid solutions are marginally different, inherently validating the ternary model. The molecular solid solution model further unmasks significant deficiencies in the available database of standard state thermodynamic properties of chlorites. Finally, pursuant to the recent recognition that green rusts probably play significant roles in the cycling of iron through sedimentary sequences, the neoformation of authigenic iron chlorites from green rusts has been examined; green rusts will readily transform to berthierine and Fe-chlorites except under oxidizing conditions atypical of aquatic environments and ferrugineous sediments.

Pyroxene exsolutions and associated Fe–Ti oxides and spinels are described in a sample of olivine gabbro representing the Middle Zone of the Panzhihua layered intrusion, Southwest China, part of the Emeishan LIP. High-angle annular dark-field scanning transmission electron microscope imaging, electron diffraction, and energy dispersive spectroscopy reveal complex multi-stage exsolution relationships in the host clinopyroxene. The studied assemblage is common in gabbroic rocks and comprises subcalcic diopside and lamellar clinoenstatite (<1 wt% Ca). Two sets of exsolved clinopyroxene lamellae are observed. Only one is, however, well developed as lamellae oriented approximately parallel to (801) of diopside, making an angle of ~10 to 11° with the (100) planes, or the c axis, of both phases. These are the so-called “100” lamellae with a perfect fit along a -crystallographic axes when viewed down to [010] zone axis. Crosscutting exsolutions of Fe–(Ti) oxides are relatively common throughout the same host clinopyroxene. Apart from ilmenite and magnetite with variable Ti-content, hercynite is a minor yet ubiquitous phase. The nanoscale study indicates a sequence of fine-scale processes: from higher- T (~1030–1100 °C): (I) (clino)enstatite exsolutions in low-Ca diopside; followed by (II) slightly Ca-richer diopside overgrowths and high- T titanomagnetite exsolution in diopside; to lower- T (<450 °C) (III) titanomagnetite exsolutions into ulvöspinel + magnetite; followed by (IV) sub-solidus re-equilibration in clinopyroxenes and among Fe–Ti oxides + hercynite. Using exact phase boundary theory, pressures of lamellar exsolution within the host diopside are estimated as ~2 GPa with an error of ± ≤1 GPa. The present study of complex exsolutions in clinopyroxene demonstrates that a nanoscale approach can help constrain P - T - X evolution during formation of layered intrusions.

In this study, we measured the sound velocities of single-crystal periclase by Brillouin light scattering (BLS) combined with in situ synchrotron X‑ray diffraction (XRD) up to ~30 GPa and 900 K in an externally heated diamond-anvil cell (EHDAC). Our experimental results were used to evaluate the combined effects of pressure and temperature on the elastic moduli of single-crystal periclase using third-order Eulerian finite-strain equations. All of the elastic moduli increased with increasing pressure but decreased with increasing temperature, except the off-diagonal modulus C 12 , which remained almost constant up to ~30 GPa and 900 K. The derived aggregate adiabatic bulk and shear moduli ( K S 0, G 0 ) at ambient conditions were 162.8(±0.2) and 130.3(±0.2) GPa, respectively, consistent with literature results. The pressure derivatives of the bulk [( ∂KS / ∂P ) 300 K ] and shear moduli [( ∂G / ∂P ) 300 K ] at ambient conditions were 3.94(±0.05) and 2.17(±0.02), respectively, whereas the temperature derivatives of these moduli [( ∂KS / ∂T ) P and ( ∂G / ∂T ) P ] at ambient conditions were –0.025(±0.001) and –0.020(±0.001) GPa/K, respectively. A comparison of our experimental results with the high-pressure ( P ) and high-temperature ( T ) elastic moduli of ferropericlase (Fp) in the literature showed that all the elastic moduli of Fp were linearly correlated with the FeO content up to approximately 20 mol%. These results allowed us to build a comprehensive thermoelastic model for Fp to evaluate the effect of Fe-Mg substitution on the elasticity and seismic parameters of Fp at the relevant P - T conditions of the lower mantle. Our modeling results showed that both the increase of the Fe content in Fp and the increasing depth could change the compressional wave anisotropy ( AV P ) and shear wave splitting anisotropy ( AV S ) of Fp in the upper parts of the lower mantle. Furthermore, using our modeling results here, we also evaluated the contribution of Fp to seismic lateral heterogeneities of thermal or chemical origin in the lower mantle. Both the thermally induced and Fe-induced heterogeneities ratios ( R S/P = ∂ ln V S / ∂ ln V P ) of Fp from 670 to 1250 km along a representative lower mantle geotherm increased by ~2–5% and ~15%, respectively. The thermally induced R S/P value of Fp20 is ~30% higher than Fp10, indicating that the Fe content has a significant effect on the thermally induced R S/P of Fp. Compared to the seismic observation results ( R S/P = 1.7–2.0) in the upper regions of the lower mantle, the Fe-induced R S/P value of Fp is more compatible than the thermally induced R S/P value of Fp20 (the expected composition of Fp in the lower mantle) within their uncertainties. Thus, we propose that Fe-induced lateral heterogeneities can significantly contribute to the observed seismic lateral heterogeneities in the Earth’s lower mantle (670–1250 km).

We present a suite of low strain deformation experiments conducted on polycrystalline San Carlos olivine in a deformation DIA apparatus at temperatures ranging from 440 to 1106 °C at pressures between 3.8 and 4.6 GPa. The deformation behavior was monitored using in situ diffraction of white synchrotron X‑rays. The experiments were conducted at a slow strain rate of ~5 × 10 –6 /s so as to allow the initial elastic behavior to be closely monitored. For each experiment, we fit the diffraction data using elastic plastic self-consistent (EPSC) models. We find that to model the experiments we must incorporate an isotropic deformation mechanism that permits a small amount of non-elastic deformation during the initial elastic portion of the experiment. This deformation mechanism mimics the observed reduction in the elastic modulus as a function of temperature and permits us to better model the remainder of the stress strain curve. The critical resolved shear stresses (CRSS) for slip obtained from these models compare well with those measured in single-crystal deformation experiments

Antigorite dehydration experiments were performed under ambient pressure using a non-isothermal thermogravimetric analysis. Antigorite, with a grain size of 5−10 μm, was analyzed using heating rates of 10, 15, 20, and 25 K/min at temperatures of up to 1260 K. The results show that the progress of the dehydration reaction varies with the heating rate, and the dehydration reaction of antigorite occurs within a temperature range of 800–1050 K. Several models were used to fit the dehydration results, and the double-Gaussian distribution activation energy model (2-DAEM) yielded the best fit to the experimental data. The dehydration kinetics of antigorite follow 2-DAEM, and there is a compensation effect between the pre-exponential factor and the average activation energy. The activation energy of the first step of antigorite dehydration stretches over a wide interval; the second step has a significantly higher activation energy, distributed over a narrower interval. We determined that the release rate of H 2 O is 8.0×10 –5 and 2.1×10 –3 m 3 fluid m 3 rock s –1 at 893 and 973 K, respectively, which are near the onset temperature for the isothermal dehydration reaction. Our results indicate that antigorite dehydration is fast enough to induce mechanical instabilities that may trigger seismicity in the lower plane of the double seismic zone.

Elastic properties of Fe alloys are critical in constraining the compositions of planetary bodies by comparing to the planetary observations. The sound wave velocities and density of an Fe 5 Si (9 wt% Si) alloy in body-centered cubic (bcc) structure were measured by combining an ultrasonic technique with synchrotron X‑ray radiography at pressure ( P ) and temperature ( T ) conditions of 2.6–7.5 GPa and 300–1173 K, respectively. At room temperature, it is observed that adding Si to bcc-Fe increases the compressional wave velocity ( v P ) but decreases the shear wave velocity ( v S ). At high temperatures, we observed a pronounced effect of pressure on the v S - T relations in the Fe 5 Si alloy. The v P -density (ρ) relationship of the Fe 5 Si alloy is found to follow the Birch’s law in the P-T range of this study, whereas the v S -ρ relation exhibits complex behavior. Implications of these results to the lunar core and the Mercurian core are discussed. Our results imply that adding Si to a pure Fe lunar core would be invisible in terms of v P , but exhibit a decreased v S . Including Si in a sulfur-rich lunar core would display an increased v P and a decreased ρ. Our density and sound wave velocity model provide lower and upper limit for a Si-bearing lunar core if 1–3 wt% Si content of enstatite chondrite is taken as compositional analog. A Si-rich (>9 wt%) Mercurian core model is derived to satisfy newly observed moment of inertia values by Messenger spacecraft.

Most frequently arsenic is nominally monovalent (As 1– ) in pyrite (FeS 2 ) and substituted for S. Nominally trivalent arsenic (As 3+ ) has been reported previously in hydrothermal Peruvian pyrite and was considered to be substituted for Fe based on the negative correlation between the concentrations of the two elements. Here, we provide the first observation of the incorporation of As 3+ in goldfieldite [Cu 12 (As,Sb,Bi) 2 Te 2 S 13 ] and As 5+ in colusite [Cu 26 V 2 (As,Sb) 4 Sn 2 S 32 ] inclusions in As 1– -pyrite from high-sulfidation deposits in Peru. This information was obtained by combining spatially resolved electron probe (EPMA), synchrotron-based X‑ray fluorescence (SXRF), and absorption spectroscopy (micro-XANES and micro-EXAFS) with new high energy-resolution XANES spectroscopy (HR-XANES). The two Cu sulfide inclusions range from several to one hundred micrometers in size, and the As 3+ /As 5+ concentration varies from a few parts per million (ppm) to a maximum of 17.33 wt% compared to a maximum of 50 ppm As 1– in pyrite. They also contain variable amounts of Sn (18.47 wt% max), Te (15.91 wt% max), Sb (8.54 wt% max), Bi (5.53 wt% max), and V (3.25 wt% max). The occurrence of As 3+ /As 5+ -containing sulfosalts in As 1– -containing pyrite grains indicates that oxidizing hydrothermal conditions prevailed during the late stage of the mineralization process in the ore deposits from Peru. From an environmental perspective, high concentrations of potentially toxic As, contained in what appear to be non-As-bearing pyrite, may pose a heretofore unrecognized threat to ecosystems in acid mine drainage settings. More generally, the combination of techniques used in this study offers a new perspective on the mineralogy and crystal chemistry of hazardous elements in pyrite, such as highly toxic and little studied thallium.

Lunar apatites contain hundreds to thousands of parts per million of sulfur. This is puzzling because lunar basalts are thought to form in low oxygen fugacity ( f O2 ) conditions where sulfur can only exist in its reduced form (S 2– ), a substitution not previously observed in natural apatite. We present measurements of the oxidation state of S in lunar apatites and associated mesostasis glass that show that lunar apatites and glass contain dominantly S 2– , whereas natural apatites from Earth are only known to contain S 6+ . It is likely that many terrestrial and martian igneous rocks contain apatites with mixed sulfur oxidation states. The S 6+ /S 2– ratios of such apatites could be used to quantify the f O2 values at which they crystallized, given information on the portioning of S 6+ and S 2– between apatite and melt and on the S 6+ /S 2– ratios of melts as functions of f O2 and melt composition. Such a well-calibrated oxybarometer based on this the oxidation state of S in apatite would have wide application.