Volume 100, Issue 8-9

Scientists sometimes have the idea that “data are eternal,” i.e., that our scientific observations long outlive our hypotheses and ideas based on such. In this Editorial, we make use of work by science historians Chang (2004), Danielson and Graney (2014) and others, to show that data have a limited period of usefulness-a date of expiration so to speak. Beyond that date (probably mostly unknown beforehand), data either are (1) no longer of interest, because the problems that motivated their collection are resolved or no longer of concern, or (2) because new technologies render them obsolete. Scientific progress in any era is thus defined as the art of making observations that are “good enough”-so as to develop the “middle-level” and other theories, which as discussed by Chang (2004), appear to have lasting value.

The equilibrium model of prograde metamorphism, in which rocks are regarded as departing only negligibly from equilibrium states as they recrystallize, has generated a wealth of petrologic insights. But mounting evidence from diverse approaches and observations over a range of scales has revealed that kinetic impediments to reaction may prevent metamorphic rocks from attaining rock-wide chemical equilibrium along their prograde crystallization paths. To illustrate the resulting potential for inaccurate interpretation if kinetic factors are disregarded, we briefly review several case studies, including: out-of-sequence, metastable, and displaced isograds in contact aureoles; paragenetic sequences documenting overstepped, disequilibrium reaction paths; patterns of compositional zoning in garnet demonstrating partial chemical equilibrium; petrologic incongruities between observation and thermodynamic prediction; and inhibited reaction progress revealed by petrologically constrained numerical simulations of garnet crystallization. While the equilibrium model provides an indispensable framework for the study of metamorphic systems, these examples emphasize that all reactions require departures from rock-wide equilibrium, so all rocks must traverse kinetically sensitive reaction paths during recrystallization. Mindfulness of the potential significance of kinetic influences opens new avenues for petrologic investigation, thereby enhancing both analysis and interpretation.

Many studies exist on magmatic volatiles (H, C, N, F, S, Cl) in and on the Moon, within the last several years, that have cast into question the post-Apollo view of lunar formation, the distribution and sources of volatiles in the Earth-Moon system, and the thermal and magmatic evolution of the Moon. However, these recent observations are not the first data on lunar volatiles. When Apollo samples were first returned, substantial efforts were made to understand volatile elements, and a wealth of data regarding volatile elements exists in this older literature. In this review paper, we approach volatiles in and on the Moon using new and old data derived from lunar samples and remote sensing. From combining these data sets, we identified many points of convergence, although numerous questions remain unanswered. The abundances of volatiles in the bulk silicate Moon (BSM), lunar mantle, and urKREEP [last ~1% of the lunar magma ocean (LMO)] were estimated and placed within the context of the LMO model. The lunar mantle is likely heterogeneous with respect to volatiles, and the relative abundances of F, Cl, and H 2 O in the lunar mantle (H 2 O > F >> Cl) do not directly reflect those of BSM or urKREEP (Cl > H 2 O ≈ F). In fact, the abundances of volatiles in the cumulate lunar mantle were likely controlled by partitioning of volatiles between LMO liquid and nominally anhydrous minerals instead of residual liquid trapped in the cumulate pile. An internally consistent model for lunar volatiles in BSM should reproduce the absolute and relative abundances of volatiles in urKREEP, the anorthositic primary crust, and the lunar mantle within the context of processes that occurred during the thermal and magmatic evolution of the Moon. Using this mass-balance constraint, we conducted LMO crystallization calculations with a specific focus on the distributions and abundances of F, Cl, and H 2 O to determine whether or not estimates of F, Cl, and H 2 O in urKREEP are consistent with those of the lunar mantle, estimated independently from the analysis of volatiles in mare volcanic materials. Our estimate of volatiles in the bulk lunar mantle are 0.54-4.5 ppm F, 0.15-5.3 ppm H 2 O, 0.26-2.9 ppm Cl, 0.014-0.57 ppm C, and 78.9 ppm S. Our estimates of H 2 O are depleted compared to independent estimates of H 2 O in the lunar mantle, which are largely biased toward the “wettest” samples. Although the lunar mantle is depleted in volatiles relative to Earth, unlike the Earth, the mantle is not the primary host for volatiles. The primary host of the Moon’s incompatible lithophile volatiles (F, Cl, H 2 O) is urKREEP, which we estimate to have 660 ppm F, 300-1250 ppm H 2 O, and 1100-1350 ppm Cl. This urKREEP composition implies a BSM with 7.1 ppm F, 3-13 ppm H 2 O, and 11-14 ppm Cl. An upper bound on the abundances of F, Cl, and H 2 O in urKREEP and the BSM, based on F abundances in CI carbonaceous chondrites, are reported to be 5500 ppm F, 0.26-1.09 wt% H 2 O, and 0.98-1.2 wt% Cl and 60 ppm F, 27-114 ppm H 2 O, and 100-123 ppm Cl, respectively. The role of volatiles in many lunar geologic processes was also determined and discussed. Specifically, analyses of volatiles from lunar glass beads as well as the phase assemblages present in coatings on those beads were used to infer that H 2 is likely the primary vapor component responsible for propelling the fire-fountain eruptions that produced the pyroclastic glass beads (as opposed to CO). The textural occurrences of some volatile-bearing minerals are used to identify hydrothermal alteration, which is manifested by sulfide veining and sulfide-replacement textures in silicates. Metasomatic alteration in lunar systems differs substantially from terrestrial alteration due to differences in oxygen fugacity between the two bodies that result in H 2 O as the primary solvent for alteration fluids on Earth and H 2 as the primary solvent for alteration fluids on the Moon (and other reduced planetary bodies). Additionally, volatile abundances in volatile-bearing materials are combined with isotopic data to determine possible secondary processes that have affected the primary magmatic volatile signatures of lunar rocks including degassing, assimilation, and terrestrial contamination; however, these processes prove difficult to untangle within individual data sets. Data from remote sensing and lunar soils are combined to understand the distribution, origin, and abundances of volatiles on the lunar surface, which can be explained largely by solar wind implantation and spallogenic processes, although some of the volatiles in the soils may also be either indigenous to the Moon or terrestrial contamination. We have also provided a complete inventory of volatile-bearing mineral phases indigenous to lunar samples and discuss some of the “unconfirmed” volatile-bearing minerals that have been reported. Finally, a compilation of unanswered questions and future avenues of research on the topic of lunar volatiles are presented, along with a critical analysis of approaches for answering these questions.

Most models of early lunar evolution predict that the anorthositic highlands crust is the result of plagioclase flotation on a magma ocean. However, the lunar highlands crust typically contains 4 wt% FeO and so is more mafic than the strict definition of the anorthosites thought to comprise it. We used new Clementine-based mineral maps of the Moon as inputs to a series of mixing models that calculate the abundance and distribution of major highland rock types and shed light on three possible sources of excess mafic material in the lunar highlands: mafic (15 vol% mafic minerals) anorthosites, postmagma ocean igneous activity, and mafic basin ejecta. Mixing models that feature pure anorthosites like the purest anorthosite (PAN) described by Ohtake et al. (2009) and Pieters et al. (2009) are most compatible with the data. They allow us to place an upper limit of 10-20 vol% mantle material that could be mixed with the primary highlands crust. The upper limit on mantle material indicated by the mixing models is significantly lower than the 30-40 vol% mantle material expected from simple geometric calculations of the major lunar basins’ excavation cavities based on an excavation cavity depth/diameter ratio of 1/10; this discrepancy allows us to conclude that the excavation cavities of the three largest lunar basins may have been significantly shallower than those of the smaller basins. Our results are consistent with excavation cavity depth/diameter ratios for these largest basins in the range of 0.035 to 0.06, which agrees with previous gravity measurements by Wieczorek and Phillips (1999).

Experimental degassing of H-, F-, Cl-, C-, and S-bearing species from volatile-bearing magma of lunar composition at low pressure and f O₂ close to the quartz-iron-fayalite buffer (QIF) indicates that the composition of the fluid/vapor phase that is lost changes over time. A highly H-rich vapor phase is exsolved within the first 10 min of degassing leaving behind a melt that is effectively dehydrated. Some Cl, F, and S is also lost during this time, presumably as HCl, HF, and H 2 S gaseous species; however much of the original inventory of Cl, F, and S components are retained in the melt. After 10 min, the exsolved vapor is dry and dominated by S- and halogen-bearing phases, presumably consisting of metal halides and sulfides, which evolves over time toward F enrichment. This vapor evolution provides important constraints on the geochemistry of volatile-bearing lunar phases that form subsequent to or during degassing. The rapidity of H loss suggests that little if any OH-bearing apatite will crystallize from surface or near surface (~7m) melts and that degassing of lunar magmas will cause the compositions of apatites to evolve first toward the F-Cl apatite binary and eventually toward end-member fluorapatite during crystallization. During the stage of loss of primarily H component from the melt, Cl would have been lost primarily as HCl, which is reported not to fractionate Cl isotopes at magmatic temperatures (Sharp et al. 2010). After the loss of H-bearing species, continued loss of Cl would result in the degassing of metal chlorides, which have been proposed as a mechanism to fractionate Cl isotopes (Sharp et al. 2010). After the onset of metal chloride degassing, the δ 37 Cl of the melt would necessarily increase to +6 (82% Cl loss), +8 (85% Cl loss), and +20‰ (95% Cl loss) at 1, 4, and 6 h, respectively, which was approximated using a computed trajectory of δ 37 Cl values in basalt during degassing of FeCl 2 . This strong enrichment of 37 Cl in the melt after metal chloride volatilization is fully consistent with values measured for the non-leachates of a variety of lunar samples and would be reflected in apatites crystallized from a degassing melt. Our results suggest that a range in δ 37 Cl from 0 to >20‰ is expected in lunar apatite, with heavy enrichment being the norm. While 95% loss in the initial Cl content of the melt (280 ppm Cl left in the melt) would cause an increase to +20‰ in δ 37 Cl, the ability to measure this increase in a lunar sample is ultimately dependent upon the starting Cl abundances and whether or not a mechanism exists to concentrate the remaining Cl such that it can be subsequently analyzed with sufficient accuracy. Therefore, the higher the starting Cl abundances in the initial melts, the heavier δ 37 Cl values that can be measurably preserved. Importantly, such enrichments can occur in spite of high initial hydrogen contents, and therefore, our experiments demonstrate that elevated values of δ 37 Cl cannot be used as supporting evidence for an anhydrous Moon. Furthermore, if the H-bearing vapor has a significant H 2 component, this process should also result in strong enrichment of δD in the residual magmas that reach the lunar surface or near-surface environment. Apatites within some mare basalts exhibit elevated δD of 1000 ‰ depending on the initial value (Tartese and Anand 2013) in addition to the δ 37 Cl values, but elevated δ 37 Cl values are accompanied by only modest enrichments in δD in apatites from samples of the highlands crust (McCubbin et al. 2015a).

Interface coupled dissolution-reprecipitation reactions (ICDR) are a common feature of fluid-rock interaction during crustal fluid flow. We tested the hypothesis that ICDR reactions can play a key role in scavenging minor elements by exploring the fate of U during the experimental sulfidation of hematite to chalcopyrite under hydrothermal conditions (220-300 °C). The experiments where U was added, either as solid UO 2+x (s) or as a soluble uranyl complex, differed from the U-free experiments in that pyrite precipitated initially, before the onset of chalcopyrite precipitation. In addition, in UO 2+x (s)- bearing experiments, enhanced hematite dissolution led to increased porosity and precipitation of pyrite+magnetite within the hematite core, whereas in uranyl nitrate-bearing experiments, abundant pyrite formed initially, before being replaced by chalcopyrite. Uranium scavenging was mainly associated with the early reaction stage (pyrite precipitation), resulting in a thin U-rich line marking the original hematite grain surface. This “line” consists of nanocrystals of UO 2+x (s), based on chemical mapping and XANES spectroscopy. This study shows that the presence of minor components can affect the pathway of ICDR reactions. Reactions between U- and Cu-bearing fluids and hematite can explain the Cu-U association prominent in some iron oxide-copper-gold (IOCG) deposits.

A structural change in Fe 2 SiO 4 spinel (ringwoodite) has been found by synchrotron powder diffraction study and the structure of a new high-pressure phase was determined by Monte-Carlo simulation method and Rietveld profile fitting of X‑ray diffraction data up to 64 GPa at ambient temperature. A transition from the cubic spinel structure to a body centered orthorhombic phase (I-Fe 2 SiO 4 ) with space group Imma and Z = 4 was observed at approximately 34 GPa. The structure of I-Fe 2 SiO 4 has two crystallographically independent FeO 6 octahedra. Iron resides in two different sites of sixfold coordination: Fe1 and Fe2, which are arranged in layers parallel to (101) and (011) and are very similar to the layers of FeO 6 octahedra in the spinel structure. Silicon is located in the sixfold coordination in I-Fe 2 SiO 4 . The transformation to the new high-pressure phase is reversible under decompression at ambient temperature. A martensitic transformation of each slab of the spinel structure with translation vector generates the I-Fe 2 SiO 4 structure. Laser heating of I-Fe 2 SiO 4 at 1500 K results in a decomposition of the material to rhombohedral FeO and SiO 2 stishovite. FeKβ X‑ray emission measurements at high pressure up to 65 GPa show that the transition from a high spin (HS) to an intermediate spin (IS) state begins at 17 GPa in the spinel phase. The IS electron spin state is gradually enhanced with pressure. The Fe 2+ ion at the octahedral site changes the ion radius under compression at the low spin, which results in the changes of the lattice parameter and the deformation of the octahedra of the spinel structure. The compression curve of the lattice parameter of the spinel is discontinuous at ~20 GPa. The spin transition induces an isostructural change.

FTIR and RAMAN spectroscopic methods were used to study the ordering of non-stoichiometric nano-magnesium aluminate spinels (MgOnAl 2 O 3 , 0.4 < n < 12) synthesized using a combustion synthesis method. It was established that the degree of structural disorder (i.e., the inversion parameter, i) can be quantified using the intensities of the γ 1 and γ 5 IR modes or 670 and 723 cm -1 Raman shifts. The results indicated that the as-synthesized materials were heavily disordered and obey earlier conclusions that the defect chemistry of non-stoichiometric spinels is dominated by clusters formed from anti-site defects. Analysis of the temperature dependency of cation distribution in the Mg- and Alrich samples showed that the spinel phase moved toward equilibrium upon increases in temperature. Where decomposition occurred, the disordered level decreased at temperatures up to 1000 °C. Above this temperature, the order level dropped far below the expected equilibrium value and the γ 3 mode (a mode that is characterized for ordered structures, such as a natural spinel) that appears. These findings, together with Raman results of partly decomposed Al-rich samples, support the hypothesis that a MgAl 2 O 4 -γ-Al 8 / 3 □ 1/3 O 4 solid solution comprises of series of complex micro-phases with considerable short-range order.

The Jahn-Teller effect at Cu 2+ in cuprospinel CuFe 2 O 4 was investigated using high-pressure single-crystal synchrotron X‑ray diffraction techniques at beamline BL10A at the Photon Factory, KEK, Japan. Six data sets were collected in the pressure range from ambient to 5.9 GPa at room temperature. Structural refinements based on the data were performed at 0.0, 1.8, 2.7, and 4.6 GPa. The unit-cell volume of cuprospinel decreases continuously from 590.8(6) to 579.5(8) Å 3 up to 3.8 GPa. Least-squares fitting to a third-order Birch-Murnaghan equation of state yields the zero-pressure volume V 0 = 590.7(1) Å 3 and bulk modulus K 0 = 188.1(4.4) GPa with K′ fixed at 4.0. The structural formula determined by electron microprobe analysis and site occupancy refinement is represented as T(Fe 3+ 0.90 Cu 2+ 0.10 )M(Fe 3+ 1.10 Fe 2+ 0.40 Cu 2+ 0.50 )O 4 . Most of the Cu 2+ are preferentially distributed onto the octahedrally coordinated (M) site of the spinel structure. With pressure, the arrangement of the oxygen atoms around the M cation approaches a regular octahedron. This leads to an increase in the electrostatic repulsion between the coordinating oxygen ions and the 3d z 2 orbital of M Cu 2+ . At 4.6 GPa, a cubic-tetragonal phase transition is indicated by a splitting of the a axis of the cubic structure into a smaller a axis and a longer c axis, with unit-cell parameters a = 5.882(1) Å and c = 8.337(1) Å. The tetragonal structure with space group I4 1 /amd was refined to R1 = 0.0332 and wR2 = 0.0703 using 38 observed reflections. At the M site, the two M-O bonds parallel to the c-axis direction of the unit cell are stretched with respect to the four M-O bonds parallel to the a-b plane, which leads to an elongated octahedron along the c-axis. The cubic-to-tetragonal transition induced by the Jahn-Teller effect at Cu 2+ is attributable to this distortion of the CuO 6 octahedron and involves Cu 3d z 2 orbital, ab initio quantum chemical calculations support the observation. At the tetrahedrally coordinated (T) site, on the other hand, the tetrahedral O-T-O bond angle increases from 109.47° to 111.7(7)°, which generates a compressed tetrahedral geometry along the c-axis. As a result of the competing distortions between the elongated octahedron and the compressed tetrahedron, the a unit-cell parameter is shortened with respect to the c unit-cell parameter, giving a ca′ ratio (a′ = √2 a) slightly greater than unity as referred to cubic lattice (ca′ = 1.002). The ca′ value increases to 1.007 with pressure, suggesting further distortions of the elongated octahedron and compressed tetrahedron.

To investigate how large volumes of silicic melts segregate to form granitic plutons, we conducted a case study of a zoned pluton, in which SiO 2 increases from intermediate (69 wt%) to highly silicic compositions (74 wt%) toward the contact with metasedimentary wallrock in the outer 25 m of the pluton. All other major, minor, and trace elements vary systematically with SiO 2 and indicate that outward increasing SiO 2 is due to a decrease in mafic elements and minerals. Whole-rock oxygen isotopes and elemental variation diagrams do not support mixing with wallrock as an explanation for the Si-rich boundary layer. Instead, mafic enclaves, which are common in the pluton, also decrease in abundance in the outer 25 m of the pluton, suggesting a mechanical origin for the Si-rich boundary layer. The coupling of mechanical and geochemical boundary layers, combined with geochemical modeling, indicate that the silica-rich, enclave-poor boundary layer formed by hindered settling or compaction of a crystal-rich (crystal fractions >60%) magmatic mush. Segregation of melts at high crystal fraction is known to be a slow process. However, petrography and Zr-based thermometry indicate that the residual Si-rich liquids were water-saturated. Water decreases melt viscosity, which helps expulsion, but equally importantly, water also delays much of the latent heat release to late in the thermal and crystallization history of a cooling magma. We show that the higher the water content, the longer the time interval over which a magma chamber resides at the stage when water-saturated, high-silica liquids form, allowing sufficient time for exfiltration of silicic liquids before the magma body freezes.

Granitoid lithic clasts from the 0.7 ka Kaharoa eruption at the Tarawera volcano (Okataina Volcanic Complex, Taupo Volcanic Zone, New Zealand) provide insight into the processes of rhyolite formation. The plutonic lithic clasts of the Kaharoa eruption consist of (1) quartz phenocrysts, which are often grouped into clusters of two to eight quartz grains, (2) plagioclase phenocrysts (mostly ~An 40 with up to An 60 cores), and (3) interstitial alkali feldspar. Quartz orientations obtained through electron backscatter diffraction (EBSD) methods show that 78% of the 82 analyzed clusters have at least one pair of quartz grains with the dominant dipyramidal faces matched. Variations in cathodoluminescence (CL) zoning patterns of the quartz suggest that quartz clusters came together after initial crystal growth and that many quartz crystals were subject to one or more resorption events. The process of quartz crystals with different magmatic histories coming together into common relative orientations to form clusters is indicative of oriented quartz synneusis and suggests a history of crystal accumulation. The quartz clusters are interpreted to have formed as part of a crystal cumulate mush within a shallow magma chamber where quartz crystals rotated into contact along their dominant dipyramidal faces during hindered settling and/or compaction. The preservation of oriented quartz clusters from the Kaharoa plutonic lithics thus provides evidence for synchronous, shallow pluton formation from a cumulate mush during active silicic volcanism. This result is consistent with models whereby meltrich, high-silica rhyolite formation occurs via interstitial melt extraction from a low-silica rhyolite mush in the shallow crust.

Apatite-melt partitioning experiments were conducted in a piston-cylinder press at 1.0-1.2 GPa and 950-1000 °C using an Fe-rich basaltic starting composition and an oxygen fugacity within the range of ΔIW-1 to ΔIW+2. Each experiment had a unique F:Cl:OH ratio to assess the partitioning as a function of the volatile content of apatite and melt. The quenched melt and apatite were analyzed by electron probe microanalysis and secondary ion mass spectrometry techniques. The mineral-melt partition coefficients (D values) determined in this study are as follows: D F Ap-Melt = 4.4-19, D Cl Ap-Melt = 1.1-5, D OH Ap-Melt = 0.07-0.24. This large range in values indicates that a linear relationship does not exist between the concentrations of F, Cl, or OH in apatite and F, Cl, or OH in melt, respectively. This non- Nernstian behavior is a direct consequence of F, Cl, and OH being essential structural constituents in apatite and minor to trace components in the melt. Therefore mineral-melt D values for F, Cl, and OH in apatite should not be used to directly determine the volatile abundances of coexisting silicate melts. However, the apatite-melt D values for F, Cl, and OH are necessarily interdependent given that F, Cl, and OH all mix on the same crystallographic site in apatite. Consequently, we examined the ratio of D values (exchange coefficients) for each volatile pair (OH-F, Cl-F, and OH-Cl) and observed that they display much less variability: K dCl-F Ap-Melt = 0.21± 0.03, K dOH-F Ap-Melt = 0.014 ± 0.002, and K dOH-Cl Ap-Melt = 0.06 ± 0.02 . However, variations with apatite composition, specifically when mole fractions of F in the apatite X-site were low (X F < 0.18), were observed and warrant additional study. To implement the exchange coefficient to determine the H 2 O content of a silicate melt at the time of apatite crystallization (apatitebased melt hygrometry), the H 2 O abundance of the apatite, an apatite-melt exchange K d that includes OH (either OH-F or OH-Cl), and the abundance of F or Cl in the apatite and F or Cl in the melt at the time of apatite crystallization are needed (F if using the OH-F K d and Cl if using the OH-Cl K d ). To determine the H 2 O content of the parental melt, the F or Cl abundance of the parental melt is needed in place of the F or Cl abundance of the melt at the time of apatite crystallization. Importantly, however, exchange coefficients may vary as a function of temperature, pressure, melt composition, apatite composition, and/or oxygen fugacity, so the combined effects of these parameters must be investigated further before exchange coefficients are applied broadly to determine volatile abundances of coexisting melt from apatite volatile abundances.

Arseniosiderite and yukonite are among the important arsenate minerals occurring as secondary alteration products in relation to the oxidation of arsenopyrite and arsenian pyrite and as discrete grains in some gold ores, mine tailings, and contaminated soils. Characteristics of these Ca-Fe arsenate species are not well known and our understanding of the conditions promoting their formation and dissolution is limited. Long- and short-range structural characteristics and thermodynamic properties of the Ca-Fe arsenates forming in the Ca-Fe(III)-As(V)-NO 3 system were determined to better predict the mineralogical transformations taking place in neutralized sludge and tailings environments, and their influence on arsenic mobilization. Yukonite and arseniosiderite readily form from solutions with highly variable compositions at a wide pH range from slightly acidic to alkaline conditions. Calcium concentrations corresponding to molar Ca/(Ca+Fe+As) ratios as low as 0.1 appear to be adequate for their formation. Our experimental results confirm observations in natural settings and mine tailings where scorodite is progressively replaced by yukonite and arseniosiderite. The initial amorphous precipitates made of small oligomeric units of edge-sharing FeO 6 octahedra with bridging arsenate evolve to yukonite through the establishment of corner linkages between the FeO 6 chains. Yukonite represents a nanocrystalline precursor and Ca-deficient variety of arseniosiderite. Formation of arseniosiderite is kinetically controlled with faster development of crystallinity at neutral to slightly acidic pH and slower kinetics under alkaline conditions. Calorimetric measurements provided an enthalpy of formation value of -1950.3 ± 3.1 kJ/mol and standard entropy of 237.4 ± 4.4 J/(mol·K) for arseniosiderite [with composition Ca 0.663 Fe 1.093 (AsO 4 )(OH) 1.605 ·0.827H 2 O], the corresponding Gibbs free energy of formation is -1733 ± 3.4 kJ/mol. A rough estimate of the thermodynamic properties of yukonite is also provided. Arseniosiderite is a stable arsenate between pH 3.5 and 7.5 in solutions saturated with respect to soluble Ca minerals such as calcite, gypsum, anorthite, or Ca-montmorillonite. Arsenic release from mine wastes and contaminated soils can be effectively controlled by arseniosiderite and the conditions promoting its formation such as lime-treatment leading to gypsum saturation in ferric arsenate solutions would prove to be desirable for stabilizing arsenic in the form of arseniosiderite in mine wastes.

High-pressure single-crystal X‑ray diffraction experiments were conducted on natural cordierite crystals with composition Mg 1.907(18) Fe 0.127(6) Al 4.01(2) Si 4.96(3) Na 0.026(3) O 18.12(9) using a synchrotron X‑ray source. The samples were compressed at 300 K in a diamond-anvil cell to a maximum pressure of 15.22(15) GPa with a neon pressure-transmitting medium and a gold pressure calibrant. We observed a recently described orthorhombic to phase transition, as well as a further transition to a second triclinic phase. We solved and refined both new triclinic phases in space group P1, and designate them cordierite II and III. The structures of cordierite II and III were refined at 7.52(3) and 15.22(15) GPa, respectively. The lattice parameters at these pressures are a = 15.567(3), b = 9.6235(4), c = 9.0658(6) Å, α = 89.963(5)°, β = 86.252(10)°, and γ = 90.974(8)° for cordierite II, and a = 8.5191(19), b = 8.2448(3), c = 9.1627(4) Å, α = 85.672(4)°, β = 85.986(7)°, and γ = 70.839(10)° for cordierite III. Across the phase transitions there is a significant reduction in the length of the a-axis (~2 Å per phase transition), whereas both the b- and c-axis remain largely unchanged. Cordierite II has fourfold- and fivefold-coordinated Si and Al, while cordierite III has fourfold-, fivefold-, and sixfold-coordinated Si, fourfold- and fivefold-coordinated Al, and fivefold- and sixfold-coordinated Mg. The sequence of high-pressure phases shows increasing polymerization of coordination polyhedra. These results, together with other recent studies, suggest that mixed four-, five-, and sixfold coordination states may occur more commonly in silicate structures compressed at 300 K than previously recognized.

Pseudotachylytes resulted from frictional melts associated with ultramylonites in high-grade metapelitic rocks from the Ivrea-Verbano zone in the Southern Alps (Northern Italy) were studied with focus on the deformation microstructures in zircon. The aims were to investigate the characteristics of zircon deformation in seismic zones, and to recognize specific microstructures generated in zircon during earthquakes, which could be useful for mineral dating of paleo-seismic events; helps to understand how seismic energy is released at depth and interacts with metamorphic processes. The interior of polished zircon grains ranging from 30 to 150 mm in length were investigated with optical microscope and scanning electron microscope (SEM) techniques, including secondary electron (SE), backscattered electron (BSE), forward scattered electron (FSE), cathodoluminescence (CL) imaging, and crystallographic orientation mapping by electron backscatter diffraction analysis (EBSD). Grains were studied in situ and as separated fractions embedded in epoxy disks. Among different cataclastic and crystal-plastic deformation microstructures in zircon we identified characteristic planar deformation bands (PDBs), planar fractures (PFs), and curviplanar fractures (CFs). Planar deformation bands in zircon are crystallographically controlled planar lattice volumes with misorientation from the host grain, which varies from 0.4° to 2.7°. PDBs are usually parallel to {100} crystallographic planes, have width from 0.3 to 1 mm and average spacing of 5 mm in 2D sections. Planar deformation bands appear as contrast lamellae in orientation contrast images and in EBSD maps, and in rare cases can be observed with the optical microscope. PDBs form in specifically oriented grains due to high differential stresses, high temperatures, and high strain rates generated in seismically active environment and/or due to shearing in the vicinity of frictional melts. Discovered structures represent a result of crystal-plastic deformation of zircon grains with operating dislocations having <100>{010} glide system and <001> misorientation axis, therefore, they can be classified as a new type 4 lattice distortion pattern, according to the existing classification for zircon (Piazolo et al. 2012; Kovaleva et al. 2014). We have demonstrated that formation of planar fractures in zircon takes place not only during impacts, but also in seismically active zones. We observe at least two cases of formation of PFs with {100} orientation: (1) as a result of evolution of PDBs and (2) as micro-cleavage. This study demonstrates that planar microstructures in terrestrial zircon do not exclusively form during impact events, but also as a result of seismic events at depth due to unusually high differential stress, strain rate, and temperature. According to the new findings, PDBs in zircon from the deep-crust are supposed to represent newly recognized evidence of seismicity.

Phengite exsolution in omphacite from the Yematan eclogite, North Qaidam UHP belt, NW China, is described. Mineralogical investigations show that the precursor omphacite in the Yematan eclogite contains up to 1.16 wt% K 2 O and ~10 000 ppm H 2 O. Experimental studies document this omphacite to be formed at pressures higher than 6 GPa (at 900 °C). The pressure-temperature conditions of 3.68 GPa and 892 °C for phengite exsolution in omphacite associated with garnet during the exhumation were obtained by using Grt-Omp-Ph geothermobarometer. We conclude that omphacitic-clinopyroxene in subducted eclogites may act as a robust medium to transport H 2 O and potassium deep into the interior of the Earth. This study suggests that the Yematan eclogite in North Qaidam UHP metamorphic belt, NW China, may have been subducted and exhumed from depths of more than 200 km.

To understand the origin of observed low velocities in the crustal portion of subducting plates, we performed in situ measurements of elastic wave velocities of anorthite at temperatures up to 1373 K at pressure of ~1 GPa and up to 773 K at 2.0-7.0 GPa. A fine-grained polycrystalline anorthite, which was synthesized using a gas pressure apparatus, was used for the measurements. The high-pressure experiments were performed using the multi-anvil apparatus installed on beamline BL04B1 at SPring-8. The elastic wave velocity was measured by the ultrasonic pulse method with synchrotron X‑ray radiographic imaging and X‑ray diffraction techniques. At ~1.0 GPa, elastic wave velocities exhibited a sharp temperature-induced kink at ~500 K. Below 500 K, the elastic wave velocities decrease with increasing temperature. In contrast, above 500 K, the elastic wave velocities show an increasing trend in the range of 500-900 K, and then revert back to a decreasing trend at above 900 K. We also found a pressure-induced velocity anomaly of anorthite. At 300-373 K, ν P is constant up to 4 GPa, but decrease above 4 GPa with increasing pressure, while ν S decreases monotonously with increasing pressure. These elastic anomalies are considered to be attributable to the tilting behavior of the corner-sharing TO 4 (T = Al, Si) tetrahedra in three-dimensional frameworks of anorthite. Our results suggest the presence of plagioclase feldspar has the potential to causes low-velocity anomaly in the subducting oceanic crust when it survives as a metastable phase in the slab at higher pressure and lower temperature conditions.

Molecular dynamic simulations offer promise as an essential tool, complementary to experiments, for expanding the reach of computational thermodynamics in igneous petrology by evaluating excess thermodynamic properties of multicomponent silicate melts. However, we present evidence suggesting that current practices in simulation may not achieve the precision needed to predict complex phase equilibria relevant to modeling the Earth’s interior evolution. We highlight the importance of quantification of the chemical short-range order in terms of cation-cation pairs in the melt and its impact on different kinetic aspects of molecular dynamic simulations. We analyze published molecular dynamic simulation studies of silicate melts to identify specific criteria and best practices for achieving and demonstrating equilibrium and producing accurate results. Finally, we propose a list of experimental and numerical investigations that need to be performed in the future to ensure full consistency between these two approaches to reduce the gap in our fundamental understanding of silicate melts between the atomic level and the macroscopic scale.

To understand clay mineral transformations in hydromorphic conditions in the red earth sediments in Xuancheng, south China, clay mineralogy was investigated using X‑ray diffraction (XRD) and highresolution transmission electron microscopy (HRTEM). The XRD results indicated that clay minerals in the hydromorphic soils were illite, kaolinite, smectite, vermiculite, and mixed-layer illite/smectite and illite/smectite/kaolinite. Changes of the kaolinitic reflections under the various conditions suggested that the kaolinitic phase is a mixed-layer structure having kaolinite layers randomly interstratified with illite and smectite layers. HRTEM observation showed that 10 Å illite layers interstratified with both 15 Å smectite layers and 7 Å kaolinite layers in clay particles, confirming the occurrence of illite/smectite/kaolinite (I/S/K) three-component mixed-layer clays. The lattice fringes of the I/S/K clays appeared corrugated and vanishing, and also exhibited variable thickness along a lattice fringe, which were consistent with changes from illite to smectite, from smectite to kaolinite, and from illite to kaolinite, respectively. Hydromorphic conditions in the Xuancheng soils led simultaneously to the direct transformation of illite to kaolinite and the transformation of illite to smectite to kaolinite in the pedogenic processes, and the formation of I/S/K three-component mixed-layer clays as intermediate products of these processes.

Using an in situ electrical method and the multi-anvil apparatus, we determined the melting curve of sodium chloride (NaC1) up to ~20 GPa, with an estimated uncertainty of ±40 K. Our results agree well with the existing data up to 6.5 GPa. At higher pressures, the melting temperatures from this study are as much as 200 K higher than those from an experimental study using the diamond-anvil cell (DAC), and are up to 500 K lower than those from theoretical studies using molecular dynamics (MD). The discrepancies may originate from surface melting in the DAC measurements, which underestimate the melting temperature, and from superheating in MD calculations, which over-predict the melting temperature. Fitting our results to the Simon equation yield (T/T 0 )4.5 = (P - P 0 )/0.6 + 1, where T and T 0 are the melting temperatures at P and P 0 , respectively, with T 0 = 1073.6 K, T in K and P in GPa. The Simon equation fits the experimental data within uncertainties and therefore can be used to interpolate the melting curve. Using the equation of state (EoS) of NaCl at 300 K, the results are fitted to the Kraut-Kennedy equation in the form of T/T 0 = (V 0 - V)/V 0 ·4.37 + 1, where T (in K) and T 0 (= 1073.6 K) are the melting temperatures at V and V 0 (at 0.0001 GPa), respectively. At pressures above 14 GPa, the experimental data deviate from the Kraut-Kennedy equation fit toward lower temperatures, probably because the volume dependence of the Grüneisen parameter was ignored in the equation. The Gilvarry-Lindemann equation Tm ~ 1.689·f 2 ·Θ 0 2 ·(V 0 /V) 2(γ-1/3) provides a satisfactory fit to the melting curve of NaCl between 0 and 19 GPa if the exponent q in the volume dependence of the Grüneisen parameter γ = γ 0 ·(V/V 0 ) q is allowed to deviate from one. Given that the melting curve of NaCl up to 6.5 GPa is well established, monitoring the melting of NaCl offers an efficient alternative for pressure calibration of large-volume high-pressure apparatus for Earth science applications.

Two-dimensional X‑ray diffraction data contain information about not only the type of mineral phases present in an assemblage, but also the textural or grain size relationships between minerals in a sample. For minerals within a certain grain size range, ~0.1 to 100 μm, the appearance and characteristics of a Debye ring can reveal the mean grain size of a sample. In this contribution, using mineral and rock samples of known grain size ranges, we investigate the applicability of calculating the grain size of a material using a two-dimensional X‑ray diffraction crystallite size analysis method for micrometersized materials. A radial integration technique was used to derive the number of grains contributing to produce diffraction spots in the Debye ring. Monomineralic pyroxene and magnetite samples of known grain size ranges were analyzed, and the calculated grain size was observed to broadly correlate with the sample size except at the upper and lower extremes. To evaluate the technique on broader geological materials, polymineralic basalt samples with known grain size ranges were analyzed, and the calculated grain sizes did not correlate with the size of the rock fragments, but did correlate closely with the size of the individual mineral grains. Using a Bruker D8 Discover X‑ray diffractometer with a 300 μm nominal incident beam diameter, the effectiveness of the applied method appeared limited to the grain size range of ~15-63 μm for monomineralic samples. The method is further limited by several complicating factors and assumptions, including the requirement for the crystallite size to correlate with the sample grain size. The effective range of this method will vary with different instrumental and experimental conditions. When applying this method to calculate the grain size of geological materials, the calculated result should be interpreted as a minimum estimate of the grain size.

To test if hydrogen incorporation by ionic diffusion can occur between a volatile-rich kimberlitic liquid and forsterite, results of high-pressure and high-temperature experiments using a piston-cylinder apparatus at 1200-1300 °C and 1 GPa for durations of 1 min, 5 h, and 23 h, are reported here. Kimberlitic liquid in the system CaO-MgO-Al 2 O 3 -SiO 2 -CO 2 -H 2 O and synthetic forsterite single crystals were chosen as a first simplification of the complex natural kimberlite composition. Unpolarized Fourier transform infrared spectroscopy was used to quantify the concentrations of OH in the crystallographically oriented forsterite. Scanning electron microscopy, electron backscattered diffraction, electron microprobe analyses, and transmission electron microscopy were performed to identify the run products. After 5 and 23 h, a forsterite overgrowth crystallized with the same orientation as the initial forsterite single crystal. The kimberlitic liquid has crystallized as micrometer-scale euhedral forsterite neocrystals with random crystallographic orientations, as well as a nanoscale aluminous phase and a calcic phase. Despite theoretical water-saturation of the system and long duration, none of the initial forsterite single crystals display signs of hydration such as hydrogen diffusion profile from the border toward the center of the crystal. Most likely, the presence of CO 2 in the system has lowered the H 2 O fugacity to such an extent that there is no significant hydration of the starting forsterite single crystal or its overgrowth. Also, the presence of CO 2 enhances rapid forsterite crystal growth. Forsterite growth rate is around 2 × 108 mm3/h at 1250 °C. These experimental results suggest a deep mantle origin of the high OH content found in natural mantle-derived xenoliths transported in kimberlites, as reported from the Kaapvaal craton. In agreement with previous studies, it also points out to the fact that significant hydration must take place in a CO 2 -poor environment.

Determination of gem provenance is a topic of research in the gemological community for financial, security, and societal reasons. Laser-induced breakdown spectroscopy (LIBS) and multivariate analysis have the potential to revolutionize the field of gem provenance. This study acquired LIBS spectra from 569 rough sapphire and ruby specimens from 21 localities in 11 countries. The spectra were analyzed using the multivariate technique partial least-squares regression (PLSR) in separate algorithms for sapphires and rubies. Each algorithm consists of a series of PLS models. Each model compares the spectra from a locality of interest to the spectra from all other localities in the database. Success rates, as determined by the percent of correct provenance identifications, are 98.9% (sapphire) and 96.0% (ruby) for country of origin and 97.9% (sapphire) and 95.4% (ruby) for deposit of origin. Individual deposits are not recognized by the concentrations of a few elements; rather, the unique compositional signature of each deposit consists of the ratios of many elements, primarily Ca, Zr, Fe, Ba, Mt, Ti, Sr, Si, Cr, H, C, and Li, some of which may reside in inclusions. This work demonstrates that determination of country or deposit of origin may be related to a quantitative measure with a high level of success.

We report high-pressure Raman and Brillouin spectroscopy results measured in the (010) plane of a natural antigorite single crystal. We find that structural changes at >6 GPa lead to (1) an intensity crossover between Raman modes of the Si-O-Si bending vibrations, (2) changes of the compression behavior of Raman modes related to the SiO4 tetrahedra, (3) changes of the pressure derivative of the Raman shifts associated with OH stretching vibrations, (4) the emergence of a new Raman band in the OH spectral region, (5) a softening of the elastic constants c 33 and c 11 , and (6) a directional change of the slowest compressional wave velocity in the a-c plane. In addition to the structural insights at high-pressure, the unique characteristics of our single-crystal sample allows for first direct measurements of the acoustic velocity anisotropy in a plane perpendicular to the basal a-b plane. Comparison to previously published data indicates that the elastic anisotropy of antigorite strongly depends on the FeO and/or Al 2 O 3 content. In contrast, it seems not to be affected by increasing temperature as inferred from an additional high-temperature experiment performed in our study. These constraints are important for the interpretation of seismic anisotropy observations in subduction zone environments.

Anisotropy of magnetic susceptibility (AMS) is often used as a proxy for mineral fabric in deformed rocks. To do so quantitatively, it is necessary to quantify the intrinsic magnetic anisotropy of single crystals of rock-forming minerals. Amphiboles are common in mafic igneous and metamorphic rocks and often define rock texture due to their general prismatic crystal habits. Amphiboles may dominate the magnetic anisotropy in intermediate to felsic igneous rocks and in some metamorphic rock types, because they have a high Fe concentration and they can develop a strong crystallographic preferred orientation. In this study, the AMS is characterized in 28 single crystals and 1 crystal aggregate of compositionally diverse clino- and ortho-amphiboles. High-field methods were used to isolate the paramagnetic component of the anisotropy, which is unaffected by ferromagnetic inclusions that often occur in amphibole crystals. Laue imaging, laser ablation-inductively coupled plasma-mass spectrometry, and Mössbauer spectroscopy were performed to relate the magnetic anisotropy to crystal structure and Fe concentration. The minimum susceptibility is parallel to the crystallographic a*-axis and the maximum susceptibility is generally parallel to the crystallographic b-axis in tremolite, actinolite, and hornblende. Gedrite has its minimum susceptibility along the a-axis, and maximum susceptibility aligned with c. In richterite, however, the intermediate susceptibility is parallel to the b-axis and the minimum and maximum susceptibility directions are distributed in the a-c plane. The degree of anisotropy, kʹ, increases generally with Fe concentration, following a linear trend: kʹ = 1.61 × 10 -9 Fe - 1.17 × 10 -9 m 3 /kg. Additionally, it may depend on the Fe 2+ /Fe 3+ ratio. For most samples, the degree of anisotropy increases by a factor of approximately 8 upon cooling from room temperature to 77 K. Ferroactinolite, one pargasite crystal and riebeckite show a larger increase, which is related to the onset of local ferromagnetic (s.l.) interactions below about 100 K. This comprehensive data set increases our understanding of the magnetic structure of amphiboles, and it is central to interpreting magnetic fabrics of rocks whose AMS is controlled by amphibole minerals.

In this paper, we describe an in situ non-destructive technique to identify areas within zircon crystals that have experienced fluid exchange. We show that Raman spectroscopy combined with electron microprobe trace-element analysis can be used to pinpoint areas in altered, complexly zoned, and metamict zircon that record the original magmatic compositions. These techniques are developed on a suite of Paleoproterozoic zircon crystals from the Scourie dike swarm in the Archean gneiss terrane of NW Scotland that are known to be anomalously low in δ 18 O. We show that zircons from the Scourie dikes record extremely low-δ 18 O isotopic compositions down to approximately -3‰, which reflect their magmatic values. Zircon populations from the dikes have a range in δ 18 O from low values (<0‰) up to ~5‰ with no obvious relationship between oxygen isotopic composition and cathodoluminescence. Raman spectra from the zircons show evidence for fluid interaction in some areas of the grains because of partial recrystallization and atypical radiation damage properties. Electron microprobe analysis in the same areas documents high (>1000 ppm) U concentrations and high (>1000 ppm) abundances of non-formula Ca. When the combined Raman and electron microprobe data are used to filter the oxygen isotopic data, the fluid altered areas of the grains show consistently higher δ 18 O values than the areas without evidence for fluid alteration. The low-δ 18 O values therefore reflect the original magmatic composition of the grains and indicate that the Scourie dike magmas were low in 18 O. We suggest that these non-destructive techniques should be used prior to SIMS analysis of complex zircons to target the least disturbed areas of the grains.

To evaluate the accuracy of Fe 3+ and Fe 2+ ratios in silicate glasses determined by Mössbauer spectroscopy, we examine in detail the temperature (47-293 K) of Mössbauer spectra for two andesitic glasses, one quenched at 1 atm, 1400 °C (VF3) and the other at 3.5 GPa, 1600 °C (M544). Variabletemperature Mössbauer spectra of these two glasses are used to characterize the recoilless fraction, f, by two different methods-a relative method (RM) based on the temperature dependence of the ratios of Fe 3+ and Fe 2+ Mössbauer doublets and the second based on the temperature dependence of the center shift (CS) of the doublets. The ratio of the recoilless fractions for Fe 3+ and Fe 2+ , C T , can then be used to adjust the observed area of the Mössbauer doublets into the Fe 3+ /∑Fe ratio in the sample. We also evaluated the contributions of non-paramagnetic components to the Fe in the glasses by determining the influence of applied magnetic field on sample magnetization. Finally, for the VF3 glass, we determined the Fe 3+ /∑Fe independently by wet chemical determination of the FeO content combined with careful electron microprobe analyses of total Fe. Recoilless fractions determined with the CS method (CSM) are significantly smaller than those determined with the relative method and suggest larger corrections to room-temperature Fe 3+ /∑Fe ratios. However, the RM determinations are believed to be more accurate because they depend less on the assumption of the Debye harmonic model and because they produce more nearly temperature-independent estimates of Fe 3+ /∑Fe ratios. Non-linear responses of sample magnetizations to applied magnetic fields indicate that the glasses contain a small (0.4-1.1% for VF3) superparamagnetic component that is most likely to be nanophase precipitates of (Fe,Mg)Fe 2 O 4 oxide, but corrections for this component have negligible influence on the total Fe 3+ /∑Fe determined for the glass. For the VF3 glass, the Fe 3+ /∑Fe produced by uncorrected room-temperature Mössbauer spectroscopy [0.685 ± 0.014 in two standard deviation (2σ)] agrees within 3% of that determined by wet chemistry (0.666 ± 0.030 in 2σ). The Fe 3+ /∑Fe corrected for recoilless fraction contributions is 0.634 ± 0.078(2σ), which is 7.5% lower than the uncorrected room-temperature ratio, but also agrees within 5% of wet chemical ratio. At least for this andesitic glass, the room-temperature determination of Fe 3+ /∑Fe is accurate within analytical uncertainty, but room-temperature Mössbauer determinations of Fe 3+ /∑Fe are always systematically higher compared to recoilless-fraction corrected ratios.

Aluminous bridgmanite (Al-Bm) is the dominant phase in the Earth’s lower mantle. In this study, the Mössbauer spectra of an Al-Bm sample Mg 0.868 Fe 0.087 Si 0.944 Al 0.101 O 2.994 were recorded from 65 to 300 K at 1 bar. The temperature dependence of the center shift was fitted by the Debye model and yielded the Debye temperatures of 305 ± 3 K for Fe 2+ and 361 ± 22 K for Fe 3+ . These values are lower than those of Al-free bridgmanite by 17 and 24%, respectively, indicating that the presence of Fe and Al increases the average Fe-O bond length and weakens the bond strength. At 300 K, the calculated recoil-free fractions of Fe 2+ (0.637 ± 0.006) and Fe 3+ (0.72 ± 0.02) are similar and therefore the molar fractions of Fe 2+ and Fe 3+ are nearly the same as the area fractions of the corresponding Mössbauer doublets. At 900 K, the calculated recoil-free fractions of Fe 3+ is 46% higher than that of Fe 2+ , implying that the molar fraction of Fe 3+ is only 41% for a measured spectral area fraction of 50%, and that the area fractions of iron sites may change with temperature without any changes in the valence state or spin state of iron. We infer that Fe 3+ accounts for 46 ± 2% of the iron in the Al-Bm and it enters the A site along with Al 3+ in the B site through the coupled-substitution mechanism. An Fe 2+ component with large quadrupole splitting (~4.0 mm/s) was observed at cryogenic conditions and interpreted as a high-spin distorted iron site.

The Phepane Dome is a circular outcrop of metasedimentary rock within the Eastern Lobe of the Bushveld Complex hypothesized to have formed as a diapir, when underlying wallrock rose into the overlying magma body. Interactions between the metasedimentary rock of the Phepane Dome and the magma of the Bushveld Complex were investigated through measurements of oxygen and lithium isotopic compositions, determination of mineral modes and major-element mineral compositions, cathodoluminescence imaging, and dihedral angle analysis. Evidence from cathodoluminescence imaging and dihedral angle analysis suggest that heat transfer during diapirism caused partial melting and complete recrystallization of the Phepane Dome metasedimentary rock. Oxygen isotope analysis of samples from traverses spanning the contact between metasedimentary and igneous rocks demonstrates that relatively minimal exchange of oxygen (over distance ~4 m) occurred across the contact between the Phepane Dome and the surrounding Bushveld magma. The lithium concentrations and isotopic compositions of metasedimentary rock are significantly different from the associated igneous rocks. Lithium isotope analysis of samples from traverses across the contact demonstrates exchange of Li over somewhat greater distances (~60 m) than oxygen, consistent with evidence that suggests a higher diffusivity of Li than most major elements. Models of oxygen diffusion through intergranular melt and aqueous fluid are used to place maximum constraints on the duration of diffusive exchange across the contact, resulting in estimates ranging from 5 kyr to 5 Myr. These values are consistent with previous estimates of the duration of crystallization of the Bushveld Complex and Phepane diapir development.

We simulated the redox decomposition of magnesium-siderite at pressures and temperatures corresponding to the top of the Earth’s D″ layer (135 GPa and 2650 K). It transforms into new phases, with unexpected stoichiometry. We report their crystal structure, based on single-crystal synchrotron radiation diffraction on a multi-grain sample, using a charge-flipping algorithm. Mg 2 Fe 2 (C 4 O 13 ) is monoclinic, a = 9.822(3), b = 3.9023(13), c = 13.154(5) Å, β = 108.02(3)°, V = 479.4(3) Å 3 (at 135 GPa). It contains tetrahedrally coordinated carbon units, corner-shared in truncated C 4 O 13 chains. Half of the cations are divalent, and half trivalent. The carbonate coexists with a new iron oxide, Fe 13 O 19 , monoclinic, a = 19.233(2), b = 2.5820(13), c = 9.550(11) Å, β = 118.39(3)°, V = 417.2(5) Å 3 (at 135 GPa). It has a stoichiometry between hematite, Fe2O3, and magnetite, Fe 3 O 4 . The formation of these unquenchable phases indicates, indirectly, the formation of reduced-carbon species, possibly diamond. These structures suggest the ideas that the mineralogy of the lower mantle and D″ region may be more complex than previously estimated. This is especially significant concerning accessory phases of fundamental geochemical significance and their role in ultra-deep iron-carbon redox coupling processes, as well as the iron-oxygen system, which certainly play an important role in the lower mantle mineral phase equilibria.