Volume 103, Issue 11

The garnet-biotite-sillimanite enclaves from El Hoyazo are quenched anatectic metapelites found within peraluminous dacites (Betic Cordillera, SE Spain), representing a residual lower crust in the area after 40–60% of melt extraction. Anatexis occurred concomitantly with deformation in a regional metamorphic setting during the Upper Miocene at the base of the continental crust. Previous studies have provided detailed information on the pressure-temperature evolution, the sequence of melting reactions, and associated melt proportions and compositions. They show that enclaves mostly record peak metamorphic assemblages, mineral compositions, and, likely, microstructures, with minor changes upon entrapment within the dacite magma and rapid ascent and extrusion. The enclaves still preserve a proportion of the primary melt, that solidified to glass in abundant melt inclusions (MI) and matrix melt, permitting the study of the microstructural relationships between melt and residue. This study focuses on the geometry of the glass network at the microscale that, combined with the previously reported anatectic history, helps shed light on the mechanisms and history of melt drainage from these rocks. A representative sample of the enclaves was investigated by synchrotron μ-XRF and scanning electron microscopy to map the distribution of glass and minerals on three thin sections cut perpendicularly to the foliation. The combination of major and trace element μ-XRF distribution maps and detailed backscattered electron images evidence the presence of a pervasive and mostly interconnected glass network through the studied centimeter-scale sections. Interconnection is due to the crosscutting of films and glass-rich domains oriented parallel and at high angle with foliation. Although enclaves lost ~40–60% of melt, they still contain ~10–15% of glass, with a considerable proportion of it stored within the Mix, which is an aggregate of micrometer-sized fibrolitic sillimanite and glass. The distribution of glass (former melt) is not in textural equilibrium with the solid residue and resembles the interconnected network of deformation bands observed in migmatites of anatectic terranes at the mesoscale. Microstructural studies of melt pseudomorphs in migmatites and granulites of anatectic terranes are scarce, but the following remarkable interpretations can be made combining our observations of these enclaves: melt formed an interconnected network during anatexis that permitted melt segregation and extraction, though melt-residue textural disequilibrium is the rule rather than the exception. The proportion of melt present in residual migmatites can be much higher than the permeability threshold for crustal protoliths; in this particular study, two reasons for this might be that (1) melt was still being produced and flowing through the residual migmatite right before disaggregation and inclusion within the host dacite, where additional melt drainage was impeded by the hydrostatic stress field, and (2) a particular microstructure produced at the onset of anatexis, such as the Mix, acted as a trap for melt impeding or delaying melt segregation.

Low oxygen fugacity ( f O 2 ) in the lunar interior (one log unit below the iron-wüstite buffer [IW-1]) offers the possibility that stable Fe-metal and sulfide phases exist as restites within lunar mare basalt source regions. Metal and sulfide phases have high metal-melt and sulfide-melt partition coefficients for chalcophile, siderophile (>100), and highly siderophile elements (>>100 000; HSE: Os, Ir, Ru, Rh, Pt, Pd, Re, Au). If these phases are residual after mare basalt extraction, they would be expected to retain significant quantities of these elements, likely generating non-chondritic HSE inter-element ratios, including Re/Os in the silicate magma. If such phases were present, then the estimated HSE abundances of the bulk silicate moon (BSM) would be proportionally higher than current estimates (0.00023 ± 2 × CI chondrite), and perhaps closer to the bulk silicate earth (BSE) estimate (0.009 ± 2 × CI chondrite). Here I show that relationships between elements of similar incompatibility but with siderophile (W), chalcophile (Cu), and lithophile tendencies (Th, U, Yb) do not deviate from expected trends generated by magmatic differentiation during cooling and crystallization of mare basalts. These results, combined with chondrite-relative HSE abundances and near-chondritic measured 187 Os/ 188 Os compositions of primitive high-MgO mare basalts, imply that lunar mantle melts were generated from residual metal- and sulfide-free sources, or experienced complete exhaustion of metal and sulfides during partial melt extraction. Evidence for the loss of moderately volatile elements during lunar formation and early differentiation indicates that the BSM is >4 to 10 times more depleted in S than BSE. Because of an S-depleted BSM, mare basalt melts are unlikely to have reached S saturation, even if sulfide concentration at sulfide saturation (SCSS) was lowered relative to terrestrial values due to low lunar f O 2 . In the absence of residual sulfide or metal, resultant partial melt models indicate that a lunar mantle source with 25 to 75 μg/g S and high sulfide-melt partition coefficients can account for the chondritic-relative abundances of the HSE in mare basalts from a BSM that experienced <0.02% by mass of late accretion.

In this paper, we report the results of the first study focused on the thermal stability and dehydration dynamics of the natural zeolite mineral ferrierite. A sample from Monastir, Sardinia [(Na 0.56 K 1.19 Mg 2.02 Ca 0.52 Sr 0.14 ) (Al 6.89 Si 29.04 )O 72 ·17.86H 2 O; a = 19.2241(3) Å; b = 14.1563(2) Å; c = 7.5106(1) Å, V = 2043.95(7) Å 3 ] was investigated by thermogravimetric analysis and in-situ synchrotron X‑ray powder diffraction. Thermogravimetric data show that H 2 O release begins already in the range 50–100 °C and is complete at ~600 °C. The results of the structure refinements performed in Immm space group by Rietveld analysis with data collected up to 670 °C show that ferrierite belongs to the group of zeolites that do not undergo phase transitions. Upon heating to 670 °C, ferrierite behaves as a non-collapsible structure displaying only a slight contraction of the unit-cell volume (Δ V = –3%). The unit-cell parameter reductions are anisotropic, more pronounced for a than for b and c (Δ a = –1.6%; Δ b = –0.76%; Δ c = –0.70%). This anisotropic response to a temperature increase is interpreted as due to the presence in the ferrierite framework of five-membered ring chains of SiO 4 tetrahedra, which impart a higher structural rigidity along b and c . Upon dehydration we observe: (1) the gradual H 2 O loss, beginning with the molecules hosted in the 10MR channel, is almost complete at 670 °C, in good agreement with the TG data; (2) as a consequence of the decreased H 2 O content, Mg and K migrate from their original positions, moving from the center of the 10MR channel toward the walls to coordinate the framework oxygen atoms. The observation of transmission electron microscopy selected-area electron diffraction patterns revealed defective crystals with an occasional and moderate structural disorder. Beyond providing information on the thermal stability and behavior of natural ferrierite, the results of this work have significant implications for possible technological applications. These data allow for comparison with the dehydration kinetics/mechanisms of the corresponding synthetic phases, clarifying the role played by framework and extra-framework species on the high-temperature behavior of porous materials with ferrierite topology. Moreover, the information on the thermal behavior of natural ferrierite can be used to predict the energetic performances of analogous synthetic Si-pure counterparts, namely “zeosil-electrolyte” systems, under non-ambient conditions. Specifically, the very high thermal stability of ferrierite determined in this study, coupled with the baric behavior determined in other investigations, suggests that the “Si-FER–electrolyte” system may be an excellent candidate for use as an energy reservoir. Indeed, ferrierite exhibits the so-called “spring behavior,” i.e., upon compression in water or in an electrolyte solution, it converts the mechanical energy into interfacial energy, and—when pressure is released—it can completely restore the supplied mechanical energy accumulated during the compression step.

Sakhaite, ca. Ca 48 Mg 16 (BO 3 ) 32 (CO 3 ) 16 (HCl,H 2 O) 2 , is a rare rock-forming borate-carbonate mineral typically occurring in high-temperature, low-pressure calcareous skarns. It forms a complete solid solution with harkerite, ca. Ca 48 Mg 16 [AlSi 4 (O,OH) 16 ] 4 (BO 3 ) 16 (CO 3 ) 16 (HCl,H 2 O) 2 . The solid solution can be described with the general formula Ca48(Mg,Fe,Mn)16(CO3)16[AlaSi5−a(O,OH)16]y(BO3)32−4y(HCl,H2O)n $$\begin{array} {} {\rm{C}}{{\rm{a}}_{48}}{({\rm{Mg}},{\rm{Fe}},{\rm{Mn}})_{16}}{({\rm{C}}{{\rm{O}}_3})_{16}}{[{\rm{A}}{{\rm{l}}_a}{\rm{S}}{{\rm{i}}_{5 - a}}{({\rm{O}},{\rm{OH}})_{16}}]_y}{({\rm{B}}{{\rm{O}}_{\rm{3}}})_{32 - 4y}}{({\rm{HCl}},{{\rm{H}}_2}{\rm{O}})_n}\end{array} $$ where y max = 8 and n max = 16 – y . In this study, we examine samples of sakhaite and harkerite from four localities worldwide: Titovskoye deposit, Sakha Republic, Russia (type locality for sakhaite); Solongo B deposit, Buryatia Republic, Russia; Camas Malag, Skye, Scotland (type locality for harkerite); as well as a sakhaite-like mineral from the Kombat Mine, Tsumeb. The Si:B ratios of the samples ranged from that of end-member sakhaite (containing B only) to that of end-member harkerite (Si:B = 1:1), with several intermediate compositions. All samples were deficient in B relative to the ideal composition, implying significant substitution for borate groups. The Si:Al ratio of silicate-containing samples ranged from the ideal 4:1 to 4:1.5, implying substitution of Al at the Si site. The cubic unit-cell parameter was found to increase linearly with increasing Si content, except for the sakhaite-like mineral from Tsumeb. This mineral was found to have significant substitution of Pb for Ca (0.4–0.5 apfu) and was poor in Cl, which in most sakhaite and harkerite samples occupies the interstitial site surrounded by four borate groups. This interstitial site in the Tsumeb samples appears to be, instead, mainly occupied by H 2 O, which may qualify the mineral as a distinct species.

Attempts have been made to use unpolarized infrared analyses on unoriented anisotropic crystals of nominally anhydrous minerals to determine H contents, rather than using the more demanding polarized techniques that are more accurate (given that a reliable calibration is available). In this context, different approaches have been either empirically or theoretically proposed for the quantification; however, the involved accuracy has not been systematically documented by experimental work of both polarized and unpolarized analyses. In this study, we present a careful evaluation of experimentally grown, gem-quality OH-bearing olivine, clinopyroxene, and orthopyroxene single crystals. The samples were prepared for polarized and unpolarized infrared analyses, and the obtained spectra were used to estimate the H 2 O contents. We show that, regardless of the applied protocol, a single unpolarized determination is inadequate for quantitative analysis and the uncertainty could be up to ~80%. The unpolarized method of Paterson (1982), by considering the linear absorbance intensity either through a single analysis or by averaging the data from multi-grain analyses, commonly underestimates the H 2 O content, by a factor of up to ~6. The other unpolarized calibration method by using the averages of integrated absorbances of unoriented grains is in general of good accuracy, mostly within ±25% even for analyses on 2 grains (with perpendicular indicatrix sections), and the accuracy is even better if as many as 10 gains of random orientations are involved, e.g., within ±10%. Therefore, the latter method may be safely applied to quantify H in anisotropic minerals if a reasonable number of randomly oriented grains are chosen for the analyses. However, the uncertainty is non-systematic, and both underestimates and overestimates of H are possible depending upon orientation. These results provide a basis for quantifying H-species in anisotropic minerals and for documenting the quantitative effect of H on the physical properties of the host phases.

Physical properties of liquid Fe-S alloys (from 10 to 50 at%S) under high pressure were investigated by in situ X‑ray diffraction (up to 5 GPa and 1900 K) and by ab initio calculations. The local structure of Fe-S liquid alloys clearly shows how S modifies the local arrangement of the Fe atoms. Density has been extracted from the diffuse scattering by minimization of the oscillation in the short distance of the radial distribution function g ( r ). Two different formalisms for the P-V-T-X equation of state are presented to model density and sound velocity as a function of pressure, temperature, and sulfur content. Based on these results, Moon’s core composition is discussed. This coherent data set will serve as a thermodynamically consistent ground for modeling the core of small telluric planets and large icy satellites.

We wished to advance the knowledge of speciation among volatiles during melting and crystallization in the Earth’s interior; therefore, we explored the nature of carbon-, nitrogen-, and hydrogen-bearing species as determined in COHN fluids and dissolved in coexisting aluminosilicate melts. Micro-Raman characterization of fluids and melts were conducted in situ while samples were at a temperature up to 825 °C and pressure up to ~1400 MPa under redox conditions controlled with the Ti-TiO 2 -H 2 O hydrogen fugacity buffer. The fluid species are H 2 O, H 2 , NH 3 , and CH 4 . In contrast, under oxidizing conditions, the species are H 2 O, N 2 , and CO 2 . The equilibria among silicate structures (Q-species) and reduced carbon and nitrogen species are, 2NH 3 + 4Q n ⇋ 2Q n–1 (NH 2 ) + 2Q n–1 (OH), and 2CH 4 + 4Q n ⇋ 2Q n–1 (CH 3 ) + 2Q n–1 (OH). The Q n and Q n–1 denote silicate species with, respectively, n and n–1 bridging O atoms. The formulation in parentheses, (NH 2 ), (CH 3 ), and (OH), is meant to indicate that those functional groups replace one or more oxygen in the silicate tetrahedra. There is no evidence for O-NH 2 or O-CH 3 bonding. Therefore, a solution of reduced C- and N-species species in the COHN system results in depolymerization of silicate melts. The Δ H values derived from the X NH 2 / X NH 3 and X CH 3 / X CH 4 evolution with temperature, respectively, were 8.1 ± 2.3 kJ/mol and between –4.9 ± 1.0 and –6.2 ± 2.2 kJ/mol. The fluid/melt partition coefficients, K fluid/melt , of the reduced species, H 2 O, H 2 , NH 3 , and CH 4 , remain above unity at all temperatures. For example, for carbon it is in the 6–15 range with a Δ H = –13.4 ± 2.4 KJ/mol. These values compare with a 0.8–3 range with Δ H = –19 ± 2.4 kJ/mol in N-free silicate-COH systems. The K fluid/melt values for reduced nitrogen and molecular hydrogen are in the 6–10 and 6–12 range with Δ H values of –5.9 ± 0.9 and = 8 ± 6 kJ/mol, respectively. A change in redox conditions during melting and crystallization in the Earth sufficient to alter oxidized to reduced carbon- and nitrogen-bearing species will affect all melt properties that depend on melt polymerization. This suggestion implies that changing redox conditions during melting of a COHN-bearing mantle can have a profound effect on physical and chemical properties of melts and on melting and melt aggregation processes.

Impacts on undifferentiated asteroidal bodies provide invaluable information for understanding the solar system evolution. Nevertheless, dating early and small-scale impact events is technically challenging. Uranium–lead (U–Pb) systematics of U-bearing phases within shock-induced melt veins (SMVs) of chondrites may be significantly disturbed by localized heating, thus providing opportunities to date these impact events. As one of the major U hosts in chondrites, apatite in the Suizhou (L6) chondrite has been compositionally and structurally modified to varying degrees by shock metamorphism. Apatite grains in the host remained largely intact during the impact and have recorded the initial thermal cooling time (~4550 Ma) on their parent body. Apatite grains in regions less than 100 μm bordering the SMVs or in relatively “cold” regions within the SMVs were partly decomposed to tuite, driven by the localized transient heating within the SMVs. Their U–Pb systematics were disrupted to varying extents. Apatite in regions close to the center of the SMVs has been completely transformed to tuite, which yields an age of 4481 ± 30 Ma (2σ), providing an upper limit to the impact event. This study clearly demonstrates that by integrating in situ U–Pb isotope analysis with detailed microstructural and compositional analysis of phosphates, it is possible to deduce the timing of early and small-scale celestial impact events, and hence create a more comprehensive understanding of the impact history of the solar system.

Olivine is a key constituent in the silicate Earth; its composition and texture informs petrogenetic understanding of numerous rock types. Here we develop a quantitative and reproducible method to measure olivine composition in three dimensions without destructive analysis, meaning full textural context is maintained. The olivine solid solution between forsterite and fayalite was measured using a combination of three-dimensional (3D) X‑ray imaging techniques, 2D backscattered electron imaging, and spot-analyses using wavelength-dispersive electron probe microanalysis. The linear attenuation coefficient of natural crystals across a range of forsterite content from ~73–91 mol% were confirmed to scale linearly with composition using 53, 60, and 70 kV monochromatic beams at I12-JEEP beamline, Diamond Light Source utilizing the helical fly-scan acquisition. A polychromatic X‑ray source was used to scan the same crystals, which yielded image contrast equivalent to measuring the mol% of forsterite with an accuracy of <1.0%. X‑ray tomography can now provide fully integrated textural and chemical analysis of natural samples containing olivine, which will support 3D and 3D+time petrologic modeling. The study has revealed >3 mm domains within a large crystal of San Carlos forsterite that varies by ~2 Fo mol%. This offers a solution to an outstanding question of inter-laboratory standardization, and also demonstrates the utility of 3D, non-destructive, chemical measurement. To our knowledge, this study is the first to describe the application of XMT to quantitative chemical measurement across a mineral solid solution. Our approach may be expanded to calculate the chemistry of other mineral systems in 3D, depending upon the number, chemistry, and density of end-members.

Olivine-rich mafic volcanic rocks (picrites) are a common and important part of ocean island and flood basalt volcanism. Despite their primitive bulk compositions (high MgO, FeO, Mg#, and low SiO 2 ), olivine-rich magmas are typically interpreted as the result of the addition of olivine from cumulate zones into more evolved basaltic liquids (MgO ≤ ~8 wt%). There are commonly two texturally distinct olivine populations in picrites: type 1 grains with planar dislocation (kink) bands, subgrains, or undulose extinction; and type 2 grains that lack these optical textures. Type 1 olivine is similar in texture to olivine from tectonized ultramafic rocks, suggesting that these textures result from plastic deformation, likely within cumulate zones. However, recently it has been proposed that type 1 olivine could also result from growth phenomena or crystal-crystal collisions. In the Kilauea Iki picrite samples used in this study, type 1 grains make up only 10–20% of the modal olivine; however they make up 30–65% of the total olivine by volume due to their large size. Therefore, type 1 grains make a large contribution to the overall composition of Kilauea Iki picrites. A combination of textural (optical defects, crystal size distributions, and minor element zoning) and geochemical analyses (trace element concentrations and diffusion of minor elements) suggests that type 1 and type 2 olivine grains have experienced distinctly different petrological histories and that they are antecrysts and autocrysts, respectively. Differences between type 1 and type 2 olivine are evident in the abundances of slow diffusing trace elements (Al, P, Ti, V), which are likely inherited from their distinct parent magmas. Type 1 and type 2 grains also define different slopes in crystal size distributions, and constraints from diffusion of P and Cr suggest that type 1 grains have longer magmatic residence times than type 2 grains. Type 1 grains likely derive from deformed cumulates within the plumbing system of Kilauea volcano, and our work supports the hypothesis that picrites from Kilauea Iki are formed by the accumulation of antecrystic olivine in more evolved basaltic liquid. Our work further supports models that type 1 olivine textures are formed during plastic deformation within cumulate zones and are not the result of growth phenomena. Our methods can be applied to other olivine-rich volcanic rocks to test the cumulate model for the formation of type 1 olivine textures, which are relatively common in picritic and related rocks from other settings.

Olivine group minerals are ubiquitous in extrusive igneous rocks and play an important role in constraining equilibria for samples in the upper mantle and above. All Raman spectra of the olivine group minerals in the solid solution between forsterite (Fo, Mg 2 SiO 4 ) and fayalite (Fa, Fe 2 SiO 4 ) have a high-intensity doublet between 800 and 880 cm –1 . Previous studies used small sample suites with limited compositional ranges and varying spectrometers to relate energy shifts of these two bands to Mg/Fe contents. In this work, Raman spectra of 93 olivine samples were acquired on either Bruker’s 532 nm (laser wavelength) Senterra or BRAVO (785/852.3 nm) spectrometer. This paper compares the two-peak band shift univariate method with two multivariate methods: partial least squares (PLS) and the least absolute shrinkage operator (Lasso). Data sets from several instruments are also examined to assess the most accurate method for predicting olivine composition from a Raman spectrum. Our 181-spectra PLS model is recommended for use when determining olivine composition from a Raman spectrum. For Raman spectra of mixed phases where only the olivine doublet can be identified, composition can best be determined using the position of the peak ca. 838–857 cm –1 through use of the regression equation %Fo = –0.179625 x 2 + 310.077 x –133 717 (where x = DB2 centroid in units of cm – 1 ). In situ methods for predicting mineral composition on planetary surfaces are critically important to extraterrestrial exploration going forward; of these, Raman spectroscopy is likely the best, as shown by the impending deployment of several Raman instruments to Mars (ExoMars and Mars 2020). More broadly, application of machine learning methods to spectral data processing have implications to multiple fields that use spectroscopic data.

In addition to dehydroxylation, thermal decomposition of Fe(II)-bearing chlorites also involves dehydrogenation. Dehydrogenation does not require the presence of oxygen and readily occurs in an inert gas atmosphere via electron transfer between the OH group and octahedral Fe(II). The reaction results in releasing one hydrogen atom that forms an H 2 gas upon diffusing out of the crystallite, and leaves one structural Fe(II) oxidized, to keep the charge balance. Dehydrogenation removes structural hydrogen reducing the amount of OH groups available for dehydroxylation thus H 2 O released during heating. In the present study, the dehydrogenation was tracked thermogravimetrically (TG) for pure chlorites. Clinochlore, Fe-clinochlore, and Mg-chamosite were investigated under various isothermal and ramp-heating conditions under pure nitrogen flow. Thermally altered Mg-chamosite was analyzed ex-situ using Mössbauer spectroscopy, X‑ray diffraction (XRD), and infrared spectroscopy. Dehydrogenation and dehydroxylation were found to occur simultaneously, but by independent mechanisms during the heating of all Fe(II)-containing chlorites. The extent of these reactions was tracked using a combination of total mass loss and degree of Fe(II) oxidation. The dehydrogenation/dehydroxylation ratio varied with heating conditions. The slower the ramp-heating rate, thus the longer time at elevated temperatures before dehydroxylation starts, the greater the dehydrogenation that precedes dehydroxylation. Each studied chlorite had its unique range of isothermal temperatures where dehydrogenation can be enhanced with only minor dehydroxylation. For Mg-chamosite, a narrow range of isothermal temperatures, 390–410 °C, caused—after 48 h of heating—the oxidation of almost 70% of Fe(II), with respect to a maximum ~20% of oxidized Fe(II) after dehydroxylation-dominated ramp heating. Any lower or higher isothermal temperatures than the optimum resulted in a lower amount of Fe(III) and greater total mass losses. Enhanced dehydrogenation led to the formation of a discrete population of a ferric (oxy) chamosite phase, observed in XRD patterns with a d -space of 13.91 Å, coexisting with 14.17 Å of the original chamosite. The dehydroxylation and dehydrogenation of chamosite at 450 °C resulted in the broadening of 00 l XRD reflections interpreted as a mixed-layer phase that consisted of original, dehydroxylated, and dehydrogenated layers. Each particular heating protocol enhancing either dehydroxylation or dehydrogenation resulted in different compositions of the product formed after chlorite structure breakdown at 1000 °C. Even with ramp heating, dehydrogenation can occur, especially with Fe(II)-rich chlorites, decreasing the total mass loss. The procedure for the determination of “structural water” content and “loss of ignition” and “total mass loss” commonly measured in rocks and minerals by thermogravimetric methods can be questioned for Fe(II)-bearing chlorites and hence in the case of all Fe(II)-containing phyllosilicates. In geological conditions, if dehydrogenation occurs prior to chlorite dehydroxylation, the quantity of “structural water” transported within chlorites to the metamorphic environment in subduction zones can be reduced even by 50%. If occurring in natural conditions, Fe(II)-oxidizing dehydrogenation reaction questions the applicability of chlorite in geothermometry. Increased Fe(III)/total Fe ratio results in the miscalculation of chlorite formation temperature as much as by hundreds of degrees Celsius. The presence of Fe(III) as a result of dehydrogenation should be considered for all Fe(II)-bearing phyllosilicates.

We quantify the metamorphic pressure-temperature ( P-T ) conditions for a newly discovered silicaundersaturated high-pressure granulite (HPG) from the Central Maine Terrane (CMT) in northeastern Connecticut, U.S.A. The rocks lie within the Acadian-Neoacadian orogenic belt (Devonian) and form part of the Brimfield Schist. The Brimfield and the adjacent Bigelow Brook Formation contain silica-saturated rocks that have previously been shown to have undergone ~1000 °C metamorphism. The pressure was less well constrained at ≥ ~1 GPa. Silica-undersaturated rocks hold underutilized potential for pinpointing peak metamorphic conditions, particularly pressure, because of their resilience to melting and the variety of refractory minerals they contain. The typical silica-undersaturated mineral assemblage is garnet + spinel + corundum + plagioclase + K-feldspar + biotite + ilmenite. Leucosomes are syenites consisting of two feldspars ± biotite. Plagioclase is commonly antiperthitic, particularly in feldspathic domains surrounding peritectic garnet; such garnet crystals reach ~10 cm in diameter. Alkali feldspars are perthitic. The rocks contain remarkable ellipsoidal spinels as much as 5.5 cm long comprising discrete crystallographic domains hosting crystallographically oriented lamellae of a Fe-Ti phase, most likely ilmenite. Corundum is usually colorless, but can also be found as sapphire in shades of pink, purple, and blue, particularly in antiperthite-rich domains surrounding large garnets. Some sapphires are concentrically color zoned. We carried out a P-T estimation using ternary feldspar reintegration thermometry of metamorphic antiperthites together with pseudosection modeling. Samples texturally and chemically record near-eclogite facies equilibration at minimum conditions of ~1040 °C and ~1.8 GPa, establishing the CMT in northeastern CT as the first known HPG locality in the U.S. These results are consistent with high P 2 O 5 levels found in garnet (0.18 wt%), Ti-in-biotite thermometry, regional sillimanite pseudomorphs after kyanite, and preliminary experimental work on melt inclusions in garnet (Ferrero et al. 2017). The leucosomes provide strong evidence that partial melting of silica-undersaturated rocks at HPG conditions can produce syenitic magmata. Strongly melt-depleted silica-undersaturated rocks may also be protoliths for garnet + spinel + corundum xenoliths reported from kimberlites. The presence of HPG gneisses demonstrates that the large-scale thrusts of the CMT sample the deepest roots of the orogenic belt (60–70 km), and perhaps even deeper subduction zone lithologies as well.

Elastic geobarometry for host-inclusion systems can provide new constraints to assess the pressure and temperature conditions attained during metamorphism. Current experimental approaches and theory are developed only for crystals immersed in a hydrostatic stress field, whereas inclusions experience deviatoric stress. We have developed a method to determine the strains in quartz inclusions from Raman spectroscopy using the concept of the phonon-mode Grüneisen tensor. We used ab initio Hartree-Fock/Density Functional Theory to calculate the wavenumbers of the Raman-active modes as a function of different strain conditions. Least-squares fits of the phonon-wavenumber shifts against strains have been used to obtain the components of the mode Grüneisen tensor of quartz (γ1m$ \begin{array} {} \gamma _1^{\rm{m}}\end{array} $ and γ3m$ \begin{array} {} \gamma _3^{\rm{m}}\end{array} $) that can be used to calculate the strains in inclusions directly from the measured Raman shifts. The concept is demonstrated with the example of a natural quartz inclusion in eclogitic garnet from Mir kimberlite and has been validated against direct X‑ray diffraction measurement of the strains in the same inclusion.

A novel Mg-bearing iron oxide Mg0.87(1)Fe4.13(1)2+Fe43+O11$ \begin{array} {} {\rm{M}}{{\rm{g}}_{0.87(1)}}{\rm{Fe}}_{4.13(1)}^{2 + }{\rm{Fe}}_4^{3 + }{{\rm{O}}_{11}}\end{array} $ was synthesized at 12 GPa and 1300 °C using a large volume press. Rietveld structural analysis was conducted with a laboratory X‑ray diffraction pattern obtained at ambient conditions. The crystal structure, which has one oxygen trigonal prism site and four octahedral sites for the cations, was found to be isostructural with Ca 2 Fe 7 O 11 . The unit-cell lattice parameters are a = 9.8441(5) Å, b = 2.8920(1) Å, c = 14.1760(6) Å, β = 99.956(4)°, V = 397.50(3) Å 3 , and Z = 2 (monoclinic, C 2/ m ). Mg and Fe cations are disordered on the trigonal prism site and on two of the four octahedral sites, and the remaining Fe is accommodated at the other two octahedral sites. The present structure is closely related to the other recently discovered Fe oxide structures, e.g., Fe 4 O 5 and Fe 5 O 6 , by distortion derived either from incorporation (Fe 4 O 5 ) or removal (Fe 5 O 6 ) of an edge-shared FeO 6 single octahedral chain in their structures. The present synthesis at deep upper mantle conditions and the structural relationships observed between various novel Mg-Fe oxides indicate that a series of different phases become stable above 10 GPa and that their relative stabilities (Fe 2+ /Fe 3+ ) must be controlled by oxygen fugacity.