Volume 94, Issue 2-3

Phase equilibria of Fe 2 O 3 have been studied up to 58 GPa and 1400 K using the Kawai-type multi anvil apparatus equipped with sintered diamond anvils. Identification of phases and pressure determination has been carried out by means of in situ X-ray observation using synchrotron radiation at SPring-8. Hematite (phase I) successively transforms to the Rh 2 O 3 (II)-type structure (phase II) and then to an orthorhombic structure (phase III) with increasing pressure. The transformations of hematite into high-pressure phases have been observed only at temperatures higher than 500 K, which is not concordant with previous results obtained by using the diamond anvil cell. Volume changes accompanied by the I-II and II-III transformations are calculated to be -2.8 and -5.0%, respectively. The phase boundary between I and II phases and that between II and III have been proposed to be P (GPa) = -0.015 T (K) + 44.2 and P (GPa) = -0.005 T (K) + 48.7, respectively. Possible correlation between a Mott transition and the phase stabilities may be concealed at room temperature due to slow reaction kinetics of the structural transformations.

A new ammonium-bearing trioctahedral mica (suhailite) has been found in gneisses from the Betic Cordillera (Spain). Suhailite appears as aggregates of golden grains unoriented with respect to the main foliation. It shows weak pleochroism from light to darker yellow and low birefringence (0.028). X-ray data indicate the presence of two compositional variations: a NH 4 -rich phase (with a basal spacing of 10.40-10.44 Å) and a NH 4 -K intermediate phase (with basal spacing of 10.20-10.26 Å). Average composition, as deduced from analyses obtained at the scale of the scanning electron microscope is [Ca 0.04 Na 0.07 K 0.35 (NH 4 ) 0.55 ](Al 0.42 Ti 0.22 Fe 1.33 Mn 0.01 Mg 0.71 ) Σ=2.70 (Si 2.67 Al 1.33 )O 10 (OH) 2 . Thermal data indicate that maximum NH 4 detachment occurs at 502 °C, suggesting a thermal stability similar to tobelite. Textural data indicate that unoriented golden grains consist of fine intergrowths of annite and suhailite and suggest that suhailite formed from primary red annite during the annite to fibrolite transformation.

Inclusions of polycrystalline K-feldspar aggregates after K-cymrite (KAlSi 3 O 8 ·nH 2 O) were discovered in garnet from a Dulan eclogite in the Qaidam ultrahigh-pressure (UHP) terrane, western China. The eclogite consists of garnet (Alm 56 Grs 23 Prp 20 Sps 01 ), omphacite (Jd 35 Aeg 6 Aug 59 ), and minor rutile and apatite. The 20 to 200 μm inclusions vary in shape from prismatic, hexagonal to rounded, and exhibit palisade and mosaic textures. Host garnets show radial fractures, similar to those surrounding quartz pseudomorphs after coesite. Some inclusions consist of almost end-member K-feldspar (Or99-100Ab0-1) polycrystalline aggregates, whereas others are composed of >90 vol% K-feldspar (Or96-99Ab1-4) with minor secondary albite occurring along the margins of the inclusions. Raman spectra of K-feldspar crystalline aggregates vary slightly reflecting various degrees of Si-Al ordering, and show a Raman peak at ~390-395 cm -1 , typical for cymrite structure. These characteristics of the K-feldspar polycrystalline inclusions imply the presence of former K-cymrite in the Dulan eclogite formed at >3 GPa at ~720 °C. The occurrence of K-cymrite in UHP eclogite is significant because of its potential as an important carrier of crustal K and H 2 O to the upper mantle.

The uranyl peroxide, studtite (UO 4 ⋅4H 2 O, C2/c, Z = 4), is expected to form as a consequence of alpha radiolysis of water in contact with spent nuclear fuel (SNF) in a geologic repository. Investigation of its stability is, therefore, of critical importance because secondary U(VI) phases may incorporate trace amounts of radionuclides and thus retard their mobility away from a repository site. To examine the effect of ionizing radiation on uranyl peroxides, electron-beam irradiation experiments have been conducted on two synthetic uranyl peroxides: studtite and metastudtite (UO 4 ⋅2H 2 O, Immm, Z = 2). All experiments were done using a transmission electron microscope (TEM) with an acceleration voltage of 200 kV at room temperature. The fluence required to completely amorphize studtite was 0.51-1.54 × 1017 e/cm 2 , which is equivalent to an absorbed dose of 0.73-1.43 × 107 Gy. Metastudtite becomes amorphous at a higher absorbed dose (1.31 × 107 Gy) than studtite, most likely because it contains fewer water molecules in its structure. These uranyl peroxides partially amorphize at doses that are one-tenth of the dose required for complete amorphization. With continued irradiation, uraninite nanocrystals form that are a few nanometers in diameter, at 4-20 × 1010 Gy. In a geologic repository, for spent nuclear fuel, the estimated absorbed doses due to ionizing radiation may be as high as 108-1011 Gy after 106 years. This is well in excess of doses in the laboratory experiments that caused the uranyl peroxides to become amorphous and decompose.

At high temperature, coarse-grained (bulk) rutile is well established as the stable phase of TiO 2 , and nanophase anatase, thermodynamically stable relative to nanophase rutile, transforms irreversibly to rutile as it coarsens. The lack of experimental heat-capacity data for bulk anatase below 52 K lends uncertainty to its standard entropy and leaves open a slight possibility that anatase may have a thermodynamic stability field at low temperature, as suggested by some theoretical calculations. In the present study, the molar heat capacities of rutile and anatase were measured from 0.5 K to about 380 K. These data were combined with previously measured high-temperature heat capacities, and fits of the resulting data set were used to generate C P ° ,m , Δ 0 T S m °, Δ 0 T H m °, and Δ 0 T G m ° values at smoothed temperatures between 0.5 and 1300 K for anatase and 0.5 and 1800 K for rutile. Using these new data and the enthalpy of transformation between anatase and rutile at 298 K, the change in Gibbs free energy for the transition between anatase and rutile from 0 to 1300 K was calculated. These calculations reveal that the transformation from bulk anatase to bulk rutile is thermodynamically favorable at all temperatures between 0 and 1300 K, confirming that bulk anatase does not have a thermodynamic stability field. Implications for the natural occurrence of these two minerals in terrestrial, lunar, and planetary settings are discussed. In particular, anatase requires low-temperature aqueous conditions for its formation and may be a reliable indicator of such conditions in both terrestrial and extraterrestrial settings.

Cobalt and nickel in olivine from lunar, martian, and terrestrial basalts show systematic differences that appear to be related to planetary parentage. We present results from a series of 1 atm experiments designed to examine the partitioning of Co and Ni in olivine over a range of conditions relevant to planetary basalts. Our results confirm the suggestions of previous workers that the partitioning of Co is such that an increase in D Co ol with crystallization is balanced by a decrease in Co concentration in the melt, resulting in constant Co in olivine, either within olivine showing otherwise normal igneous zoning, or within a suite of basalts related by fractionation. It is evident that the partitioning of Co and Ni into planetary olivine is influenced by such intensive variables as temperature, oxygen fugacity, and composition. Although separating these effects is not possible, our data demonstrate that over the range of oxygen fugacity of lunar, martian, and terrestrial basalts, oxygen fugacity has no significant effect on Co and Ni partitioning into olivine. We propose that an increase in melt polymerization with decreasing temperature, coupled with the preference of Co and Ni for the octahedral sites in olivine, is the mechanism by which D Co ol and D Ni ol increase with crystallization. With our experimental results as a baseline, we interpret data from Co and Ni in olivine from lunar, martian, and terrestrial basalts and conclude that the cause of increases in Co concentrations in olivine, either core-to-rim or within a co-genetic suite and observed primarily in martian and terrestrial basalts, is enrichment of the melt as crystallization progresses. We hypothesize that this enrichment is attributable either to crystallization of a phase in which Co is incompatible (such as plagioclase) or replenishment of the igneous system with Co-enriched melt.

Trace element mineral-magma partitioning models are important in understanding processes by which basaltic magmas are generated. Partitioning models for nickel have been extrapolated from their original applicability for the Earth’s mantle to compositions appropriate for other planets, notably the Moon and Mars. Before partitioning models can be extrapolated to explain nickel concentrations in planetary rocks, these models need to be verified thermodynamically and experimentally using planetary basaltic compositions. Experiments conducted in this study on the Martian Gusev Adirondack-class basalt, Humphrey, with 1 wt% nickel in the magma have shown that Ni affects its liquidus phase relations. By stabilizing olivine to higher temperatures, Ni increases the liquidus temperature. These experiments have shown that the Hart and Davis (1978) model based on iron-free systems cannot be extrapolated to planetary, iron-rich, basaltic systems. This work verifies the independence of the Jones (1984, 1995) and Beattie et al. (1991) models from temperature and pressure effects and suggests extrapolation to planetary compositions is justified but needs further verification. Furthermore, these experiments support the Longhi and Walker (2006) hypothesis that at high temperature nickel may be incompatible.

Lorentz transmission electron microscopy has been used to study fine-scale exsolution microstructures in ilmenite-hematite, as part of a wider investigation of the lamellar magnetism hypothesis. Pronounced asymmetric contrast is visible in out-of-focus Lorentz images of ilmenite lamellae in hematite. The likelihood that lamellar magnetism may be responsible for this contrast is assessed using simulations that incorporate interfacial magnetic moments on the (001) basal planes of hematite and ilmenite. The simulations suggest qualitatively that the asymmetric contrast is magnetic in origin. However, the magnitude of the experimental contrast is higher than that in the simulations, suggesting that an alternative origin for the observed asymmetry cannot be ruled out. Electron tomography was used to show that the lamellae have lens-like shapes and that (001) planes make up a significant proportion of the interfacial surface that they share with their host.

The open-branched tetrahedral chain of sapphirine-group minerals contains tetrahedra with 1-3 bridging O atoms (Q 1 , Q 2 , and Q 3 ). The topology implies that short-range order (SRO) driven by bonding enthalpies forces some long-range order (LRO). This is an ideal system to investigate the interplay between SRO, LRO, and configurational entropy. Results are presented of Monte Carlo simulations of ordering patterns involving Be, Al, and Si on the branched chain. The “Loewenstein’s Rule” enthalpy term favoring AlOSi links is insufficient to obtain the experimental LRO occupancies. Additional site-dependent potentials repelling Si from Q 1 and Q 2 and driving asymmetry between the two Q 3 sites are needed. All enthalpy terms are close to 2 RT in magnitude, so the observed LRO cannot be regarded as arising dominantly from Si-avoidance. A “Markov Braid” method is presented of calculating configurational entropies for the simulations. The substantial SRO present is shown to reduce Sconfig by only 17-18%. Khmaralite is related to sapphirine but contains ca. 1.5 Be per 12 tetrahedra and shows different LRO of {Be, Al, Si}. The minority cation Be is concentrated in Q 3 sites. If the exchange enthalpy ΔHAlSi = -2 RT, then ΔHBeAl and ΔHBeSi must be -6.5 to -7.5 RT to reproduce the observed mean Q for Be, Al, and Si. Khmaralite has a doubled chain period relative to sapphirine. This superstructure enables the formation of tetrahedral Be-octahedral Fe 2+ clusters in the structure. A small number of site-specific potentials allows the LRO of the khmaralite chain to be reproduced, but “Be-avoidance” remains by far the dominant interaction. Site occupancies were extrapolated to Be-rich compositions. More than 1.7 Be atoms per 12 tetrahedra is predicted to drive Be into new sites and create unfavorable BeOBe links, and to drive Si from Q 3 into the Q 1 sites. This is likely to be why the related mineral welshite, which can incorporate higher Be contents, has a quite different Be ordering pattern and symmetry.

The ammonium analogues of the high-pressure potassium-bearing silicate phases K-hollandite, K-Si-wadeite, K-cymrite, and phengite were synthesized in the system (NH 4 ) 2 O(-MgO)-Al 2 O 3 -SiO 2 - H 2 O [N(M)ASH] using multi-anvil and piston-cylinder equipment. Syntheses included NH4-hollandite (NH 4 AlSi 3 O 8 ) at 12.3 GPa, 700 °C; NH 4 -Si-wadeite [(NH 4 ) 2 Si 4 O 9 ] at 10 GPa, 700 °C; NH 4 -cymrite (NH 4 AlSi 3 O 8 ⋅H 2 O) at 7.8 GPa, 800 °C; and NH 4 -phengite [NH 4 (Mg 0.5 Al 1.5 )(Al 0.5 Si 3.5 )O 10 (OH) 2 ] at 4 GPa, 700 °C. Run products were characterized by SEM, FTIR, and powder XRD with Rietveld refinements. Cell parameters of the new NH 4 end-members are: a = 9.4234(9) Å, c = 2.7244(3) Å, V = 241.93(5) Å 3 (NH4-hollandite); a = 6.726(1) Å, c = 9.502(3) Å, V = 372.3(1) Å 3 (NH 4 -Si-wadeite); a = 5.3595(3) Å, c = 7.835(1) Å, V = 194.93(5) Å 3 (NH 4 -cymrite). NH 4 -phengite consisted of a mixture of 1M, 2M1, 2M2, 3T, and 2Or polytypes. The most abundant polytype, 2M1, has cell dimensions a = 5.2195(9) Å, b = 9.049(3) Å, c = 20.414(8) Å, β = 95.65(3)°, V = 959.5(5) Å 3 . All unit-cell volumes are enlarged in comparison to the potassium analogues. Substitution of NH 4 for K does not cause changes in space group. NH 4 incorporation was confirmed by the appearance of NH 4 -vibration modes ν 4 and ν 3 occurring in the ranges of 1397-1459 and 3223-3333 cm -1 , respectively. Ammonium in eclogite facies metasediments is mainly bound in micas and concentrations may reach up to a few thousand parts per million. It can be stored to greater depths in high-pressure potassium silicates during ongoing subduction. This possibly provides an important mechanism for nitrogen and hydrogen transport into the deeper mantle.

Highly oriented aragonite tablets have been found in the nacre layers of molluscan shell (or mother of pearl). In this article, we show that highly organized aragonite rods can be prepared over a broad range of pH values (1.5 to 6.9) and in the absence of any bio- or organic macromolecules. The organized rods were characterized by XRD, FTIR, FESEM, TEM, SAED, and EDX techniques. FESEM results reveal that the mesoscale aragonite rods are not only assembled with aragonite microrods end-to-end, and side-to-side, but are also partially fused to one another, forming flat, faceted surfaces, i.e., mesocrystal structure. TEM and SAED analyses confirm that the organized rods have the same crystallographic symmetry as single-crystal aragonite, and thus the self-assembly process is energetically favorable. Similar assembly processes also occur for the mineral strontianite of the aragonite group, revealing the occurrence of a general self-assembly process. The driving force controlling the self-assembly process may originate from the inherent anisotropic dipole-dipole interactions between the assembled units. Such dipole interaction may generally occur in biomineralization of nacre layers in molluscan shell, and orchestrate aragonite nanocrystals in an aragonite tablet to coherently orient and array. Furthermore, the dipole-dipole interactions may also contribute to the co-orientation of the aragonite tablets in the same nacreous column. Therefore, our experimental results may provide insight into biomineralization mechanisms. It appears that biological genetic and crystallochemical factors may synergistically operate in biomineralization.

Quartz veins in high-pressure (HP) to ultrahigh-pressure (UHP) rocks are the products of fluid-rock interaction, and thus provide insight into fluid processes in subduction zones. In this paper, we report an integrated study of mineral inclusion, trace-element, U-Pb age, and Lu-Hf isotope compositions of hydrothermal zircon grains from a quartz vein within an UHP eclogite outcrop from the Hong’an area, western Dabie Mountains. These data are used to decipher the age, conditions of formation, and source of fluid for zircon formation during the exhumation of UHP rocks. Zircon grains from the vein have perfect euhedral shape, and show sector zoning or weak zoning, indicating that they precipitated from the aqueous fluid responsible for the vein formation. Raman spectroscopy analysis reveals that the zircon grains contain inclusions of garnet, omphacite, rutile, quartz, and H 2 O, implying that they crystallized from aqueous fluid under HP eclogite-facies conditions. The zircon grains show low Th/U and Lu/Hf ratios, nearly flat HREE patterns, absent Eu anomalies and low LREE contents. These characteristics are consistent with their precipitation in the presence of garnet and epidote, and absence of feldspar, and thus suggest that trace-element concentrations in hydrothermal zircon are controlled by co-precipitation of mineral assemblages. Crystallization temperatures of 670 to 712 °C, which were calculated using the Ti content of zircon, are consistent with their formation under eclogitefacies conditions and may correspond to the temperature of the infiltrating fluid. The weighted mean 206 Pb/ 238 U age of 224.7 ± 1.3 Ma is taken as the best estimate for the age of quartz-vein formation and records aqueous fluid flow during the early exhumation stage of UHP rocks. The zircon grains in the quartz-vein have Hf compositions similar to those in the host eclogite, which demonstrates isotopic equilibrium between fluid and rocks and that the fluid-rock ratio was likely low.

Under slightly acidic conditions and 25 °C, the interaction between phosphate-rich aqueous solutions and gypsum cleavage fragments results in the surface precipitation of brushite (CaHPO 4 ·2H 2 O) crystals, which grow epitaxially on the (010) surface of gypsum. Using an A-centered unit-cell setting for both brushite (Aa) and gypsum (A2/a), the epitaxial relationship implies matching of the planes (010) of both structures and correspondence between equivalent crystallographic directions within these planes. The crystal habit of the overgrowing brushite is thin tabular to laminar on {010} with {111} and {111} as side faces and a clear elongation on [101]. There are two orientations of the brushite plates on the gypsum surface related to each other by a twofold axis on [010]. Thus, the overgrowth is an aggregate of parallel brushite crystals that may be twin-related, with the twofold axis as the twin law. During the interaction, gypsum dissolution is coupled with brushite growth until saturation with respect to both minerals is reached. A model of this thermodynamically driven dissolution-crystallization process is presented using the geochemical code PHREEQC. The epitaxial relationships are explained by comparing the bond system and the crystallographic properties of both minerals.

The real crystal structure of fukalite, Ca 4 Si 2 O 6 (OH) 2 (CO 3 ), was solved by means of the application of order-disorder (OD) theory and was refined through synchrotron radiation diffraction data from a single crystal. The examined sample came from the Gumeshevsk skarn copper porphyry deposit in the Central Urals, Russia. The selected crystal displays diffraction patterns characterized by strong reflections, which pointed to an orthorhombic sub-structure (the “family structure” in the OD terminology), and additional weaker reflections that correspond to a monoclinic real structure. The refined cell parameters are a = 7.573(3), b = 23.364(5), c = 11.544(4) Å, β = 109.15(1)°, space group P21/c. This unit cell corresponds to one of the six possible maximum degree of order (MDO) polytypes, as obtained by applying the OD procedure. The derivation of the six MDO polytypes is presented in the Appendix1. The intensity data were collected at the Elettra synchrotron facility (Trieste, Italy); the structure refinement converged to R = 0.0342 for 1848 reflections with I > 2σ(I) and 0.0352 for all 1958 data. The structure of fukalite may be described as formed by distinct structural modules: a calcium polyhedral framework, formed by tobermorite-type polyhedral layers alternating along b with tilleyitetype zigzag polyhedral layers; silicate chains with repeat every fifth tetrahedron, running along a and linked to the calcium polyhedral layers on opposite sides; and finally rows of CO3 groups parallel to (100) and stacked along a.

The microstructural features of a biotite (Mg-rich annite) crystal from dacite rocks of Džep, Serbia, were studied by X-ray diffraction topography (XRDT) and high-resolution and analytical transmission electron microscopy (HRTEM and ATEM, respectively). The average chemical composition, obtained by electron microprobe analysis (EMPA) and secondary-ion mass spectrometry (SIMS) data, is (K 0.87 Na 0.05 Ca 0.01 ) Σ=0.93 (Fe 2+ 1.36 Mg 1.25 Ti 0.22 Al 0.14 Mn 0.03 Li 0.01 ) Σ=3.01 (Si 2.84 Al) Σ=4.00 O 10 [(OH) 1.53 O 0.35 F 0.10 Cl 0.02 ] Σ=2.00 . Inhomogeneous regions from either different polytypes or twins were mapped by XRDT, and these images were used to guide sampling for TEM analysis. Three stacking arrangements, each belonging to subfamily-A of mica polytypes, were identified by selected area electron diffraction (SAED) patterns and HRTEM images: (1) dominant semi-random 1Mr-n(120°), including an occasional repetition of long-period inhomogeneous stacking sequences belonging to the 3T structural series; (2) highly faulted 2M 1 , sometimes with short-range-ordered long-period inhomogeneous sequences based on the 2M 1 structural series; and (3) perfectly ordered 3T stacking repeating over long distances (several micrometers). ATEM data revealed that the microchemical composition of 3T differs from that of the host matrix. The 3T is enriched in interlayer cations (K and Na), and depleted in Si and octahedral cations. No evidence of Li was detected. Compositional control on the stability of 3T stacking is suggested.

Crystals of U- and Th-doped fluor-, chlor-, and strontium-apatite have been synthesized from phosphate-halide-rich melts, and their structures were refined at room temperature with single-crystal X-ray diffraction intensities to R = 0.0167-0.0255. Structure refinements of U-doped fluorapatites indicate that U substitutes almost exclusively into the Ca2 site with site occupancy ratios UCa 2 /UCa 1 that range from 5.00 to 9.33. Similarly, structure refinements of Th-doped fluorapatites indicate that Th substitutes dominantly into the Ca 2 site with ThCa 2 /ThCa 1 values that range from 4.33 to 8.67. Structure refinements of U-doped chlorapatites show that U is essentially equally distributed between the two Ca sites with UCa 2 /UCa 1 values that range from 0.89 to 1.17. Results for Th-doped chlorapatites show that Th substitutes into both Ca 1 and Ca 2 sites with ThCa 2 /ThCa 1 values that range from 0.61 to 0.67. In the Th-doped strontium-apatites with F and Cl end-members, Th is incorporated into both the Ca 1 and Ca 2 sites. The range of ThCa 2 /ThCa 1 values is 0.56 to 1.00 for the F end-member, and 0.39 to 0.94 for the Cl end-member. XANES measurements of the U-doped samples indicate that U in fluorapatite is tetravalent, whereas in chlorapatite it is heterovalent but dominantly hexavalent. According to our calculation, the volume of the Ca 2 polyhedron increases by about 5.8% from fluorapatite to chlorapatite, but that of Ca 1 polyhedron increases by only 0.59%. We speculate that the much greater size of the Ca 2 polyhedron in chlorapatite may diminish the selectivity of this position for U and Th. The incorporation of U and Th into fluorapatite results in a decrease in the size of both Ca polyhedra, but the incorporation of U and Th into chlorapatite results in an increase in the volume of both Ca polyhedra. We suggest that the preference of U and Th for both Ca sites in chlorapatite is attributable to the large increase in size and distortion of the Ca 2 polyhedron upon substitution of Cl for F.

An in-situ, high-pressure, single-crystal X-ray diffraction study has been carried out at room temperature up to 7.42 GPa on a crystal of ingersonite, ideally Ca 3 Mn 2+ Sb 5+ 4 O 14 , from the type material. Ingersonite is isostructural with the synthetic weberite-3T polytype and related to the pyrochlore structure-type. Owing to the P range investigated and the quality of data, a second-order Birch- Murnaghan Equation of State (EoS) is the best approximation to describe the ingersonite volume evolution with P. The refined EoS parameters are V 0 = 810.6(1) Å 3 and K 0 = 154.5(2.4) GPa. The behavior of ingersonite with pressure is almost isotropic and the decrease of the unit-cell volume is mainly due to the kinking of the polyhedra rather than their volume decrease. The overall mean distances are quite constant, indicating virtually no compressibility of both the A and B polyhedra in the P range investigated. However, some geometrical changes in the pyrochlore-like AB 3 layer can be observed and compared with those observed in synthetic pyrochlore compounds. The largest change is observed for the z atomic coordinate of the O 2 atom. Using an anion-centered polyhedral description, O 2 is the only O atom that is asymmetrically located in an octahedral interstice, this feature being the most remarkable difference between the structure of ingersonite (i.e., weberite-3T type, space group P3 1 21) and that of zirconolite-3T (pyrochlore structure type, space group P3 1 21), where all the O atoms occupy the tetrahedral interstices of a cubic A 2 B 2 array. With the increase of pressure, the O 2 atom migrates from the A 4 B 2 octahedral cavity toward the adjacent AB 3 tetrahedral cavity, suggesting that a transition from weberite-3T to zirconolite-3T structure type could occur at pressures higher than 11 GPa.

Garnet solid solutions have been synthesized across the skiagite-Fe-knorringite (Fe 3 2+ Fe 2 3+ Si 3 O 12 - Fe 3 Cr 2 Si 3 O 12 ) binary join. Such compositions reflect a simple Cr-Fe 3+ exchange on the octahedral sites, with Fe 2+ occupying the adjacent dodecahedral sites. Solid solution is complete across the join and the correct stoichiometry was verified by Mössbauer spectroscopy. A symmetric fit to the molar volume data yields 119.77(2) cm 3 /mol [unit-cell parameter ao =11.6736(7) Å] for the Fe-knorringite end-member and a small negative excess volume, W v = -0.76(15) cm 3 /mol. Combining this result with literature data reveals that binary joins involving octahedral site substitutions exhibit significantly different behavior than those where substitution occurs on the dodecahedral sites. In the former case, WV is usually negative, whereas the latter joins have positive deviations from ideal behavior. Therefore, we conclude that the garnet structure responds in a fundamentally different way when accommodating different cations on the dodecahedral or octahedral sites. Mössbauer spectra of skiagite-Fe-knorringite garnets do not exhibit any significant asymmetry in the [8] Fe 2+ doublet. Fe 3+ /ΣFe values determined at room temperature and 80 K confirm the general applicability of the recoil-free fraction correction factors reported by Woodland and Ross (1994) for mixed-valence garnets. Coexisting spinels in some samples are either binary Fe 3 O 4 -Fe 2 SiO 4 or Fe 3 O 4 -FeCr 2 O 4 solid solutions. Very little mutual solubility is apparent suggesting a significant solvus may exist between the silicate and Cr-bearing spinel series.

A new synthesis technique allows large fluid inclusions to be produced at conditions under which one normally would obtain only small inclusions, with the additional advantage that the timing of fluid entrapment can be controlled. In the first step, primary fluid inclusions are grown at P-T conditions under which it is relatively easy to produce large inclusions. In the second step, which can be performed either during the same experiment or during a separate experiment after a desired time of pre-equilibration, some of the primary inclusions produced during the first step are re-opened by in-situ fracturing, causing partial replacement of the inclusion content. The sample is then left at high pressure and temperature until the cracks leading to the re-opened inclusions are healed. To test the method and quantify the efficiency of fluid replacement in the re-opened inclusions, we produced primary inclusions at 700 °C/200 MPa and re-opened them during a second experiment in the presence of a compositionally different fluid at 500 °C/70 MPa. Laser-ablation ICP-MS (LA-ICP-MS) analyses of eight primary and 11 refilled inclusions demonstrate that in the latter more than 97% of the original fluid was replaced by new fluid. Thus, the refilled inclusions are representative of the surrounding fluid at the time of in situ fracturing.

Sodium-rich montmorillonite, Na-exchanged montmorillonite, and Na-exchanged nontronite form intercalate complexes with methane hydrate, identified by a characteristic d(001) value of ~2.2 nm. The upper stability of both Na-rich montmorillonite-methane-hydrate complexes is nearly identical to that of methane hydrate, whereas that of Na-exchanged nontronite-methane-hydrate complex is ~1 °C lower. The low-temperature stability of these complexes is controlled by dehydration reactions of the montmorillonite and nontronite. At temperatures of 2 °C, the d(001) value of the montmorillonite complex decreases step-wise with decreasing temperature from ~2.2 nm at 2 °C to 1.6 nm at ≤-5 °C, indicating that H 2 O is progressively expelled from the interlayer. All methane is probably expelled at ~0 °C. The d(001) value of the nontronite complex did not show a similar step-wise reduction and, consequently, the lower stability of this complex is not well established. We conclude that under conditions of reduced salinity, smectite may sufficiently swell and intercalate with methane hydrate in an intermediate to deep-ocean floor environment. Consequently, these smectite-methane-hydrate complexes in the sub-ocean-floor surface may store substantial quantities of carbon.

First-principles calculations and molecular-dynamics simulations are employed in this study to further our understanding of the crystal structure and orientational order of carbonate ions in vaterite, which is the least stable polymorph of calcium carbonate. The structural details of vaterite have been controversial but they are prerequisite for investigating and understanding the processes involving vaterite crystal nucleation, growth, and stabilization at a molecular level. The first-principles calculations, using density functional theory with the plane-wave pseudopotential method, are carried out to calculate relative thermodynamic stabilities of proposed structures. Molecular-dynamics simulations with classical empirical potentials, larger computational cells, and fully flexible models at different temperatures are performed to investigate the orientation and order-disorder of the carbonate ions. The results show that the previously accepted structure with disordered CO 3 ions, which are randomly distributed over three orientations parallel to the c axis, is only metastable. By applying a temperature annealing technique to the molecular dynamics simulations, a more stable structure with fully ordered carbonate ions is found that has a hexagonal superstructure. The space group of this newly derived vaterite structure is P6 5 22 (no. 179) with Z = 18 and cell dimensions of √3 times in a, and 3 times in c of the previous suggested disordered structure. Comparison of experimental observations of X-ray diffraction patterns, enthalpies of transformation to calcite, and volume change by heat treatment with our theoretical calculations indicates that freshly made vaterite is often carbonate-disordered and metastable and can fully or partially transform to a carbonate-ordered structure by aging and heating.

An unusual group of mantle minerals including about 70-80 species has been recovered from podiform chromitites of the Luobusa ophiolite, Qusong County, Tibet, China. All of the minerals were hand picked from heavy mineral separates of the chromitite. The minerals include diamond, coesite, moissanite, wüstite, intermetallic compounds, Os-Ir alloys with diamond inclusions, Fe-silicides, and a new mineral, qusongite. Qusongite is associated with chromian chlorite, calcite, (W,Ti)C and (Ti,W) C alloys, and chromite. It occurs as angular grains generally 4-8 μm in diameter, but some are as large as 0.2 × 0.3 × 0.25 mm. The grains are opaque and steel-gray with a metallic luster and grayish-yellow reflection. The empirical formula (based on 2 atoms) is W 1.006 Cr 0.02 C 0.992 , and the simplified formula is WC. Qusongite has a hexagonal structure and belongs to space group P6m2, with a = 2.902(1) Å, c = 2.831(1) Å, c:a = 0.9775, V = 20.05 (1) Å 3 , Z = 1.

A new mode of arsenic incorporation into arsenian pyrite has been discovered. Electron microprobe analyses and elemental maps of arsenian pyrite from Pueblo Viejo, Dominican Republic, show that its chemical composition varies from (Fe 0.998 As 0.003 ) 1.001 S 2 to (Fe 0.963 As 0.050 Cu 0.003 Ag 0.001 ) 1.017 S 2 and that arsenic is inversely correlated with Fe. High-resolution transmission electron microscopy (HRTEM) images show that some arsenic in this pyrite is present as nanoscale inclusions of amorphous As- Fe-S in a matrix of arsenian pyrite. The amorphous inclusions display negative facets with a cubic or rectangular morphology, typical of negative inclusions of pyrite and arsenopyrite or marcarsite, respectively, and they are oriented parallel to the lattice fringes (100) of the arsenian pyrite matrix. Elemental maps collected by high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) show that the inclusions have a higher content of arsenic than the surrounding pyrite, and TEM-EDX analyses indicate approximate atomic proportions of 62 at% S, 28 at% Fe, and 10 at% As for the inclusions, which are near the minimum melting temperature in the As-Fe-S ternary. These observations suggest that the inclusions were trapped as liquids during growth of the surrounding arsenian pyrite. Although not a new mineral, this constitutes a third form for arsenian pyrite, which has previously been shown to contain arsenic as either As 1- or As 3+ .

Through a combination of 1 H MAS NMR, 1 H → 29 Si → 1 H double cross-polarization (CP) MAS NMR and 27 Al → 1 H CP MAS NMR, different OH species [SiOH, AlOH, and (Ca,Mg)OH (free OH)] have been unambiguously identified for hydrous Ca,Mg-(alumino)silicate glasses. This confirms my earlier speciation assignments made partially on the basis of 1 H chemical shift arguments. The dissolution mechanisms of water in both Al-free silicate and aluminosilicate glasses (quenched melts) are fundamentally similar. For relatively polymerized compositions, it involves dominantly the formation of TOH species (T: Si, Al) through the rupture of T-O-T linkages, in addition to molecular H 2 O; for more depolymerized compositions containing network-modifying cations of large field strength (e.g., Ca, Mg), free OH species are also important.

Combined experimental, modeling, and analytical results indicate that the rapid acidification of dilute waters in contact with nominally Fe 2+ -sulfate minerals (FeSO 4 ·nH 2 O) is caused by Fe 3+ hydrolysis, which occurs when low levels (<1 mol%) of a contaminant Fe 3+ -sulfate phase are dissolved along with the FeSO 4 ·nH 2 O. This rapid acidification has previously been attributed to hydrolysis by Fe 2+ . However, dissolution experiments performed using ZnSO 4 ·nH 2 O, in which the Zn 2+ cation has a higher hydrolysis constant (log K = -8.96) than Fe 2+ (log K = -9.5), failed to produce significant changes in solution pH. We present the results of geochemical modeling simulations confirming that FeSO 4 ·nH 2 O dissolution alone cannot explain the experimentally observed change in pH from 5.65 to 3.50. Nor can the experimental observations be explained by oxidation of Fe 2+ to Fe 3+ in solution. Instead, our experimental results can be best explained by modeling the incorporation of <1 mol% Fe 3+ contamination from any number of Fe 3+ or mixed valence Fe-sulfate phases, including anhydrous Fe 3+ 2 (SO 4 ) 3 , coquimbite, kornelite, römerite, bilinite, copiapite, or ferricopiapite, all of which are reasonable candidate phases for oxidative breakdown products of FeSO 4 ·nH 2 O. Laboratory Mössbauer spectra are consistent with up to 0.6 mol% of the total Fe in the sample to be present as Fe 3+ . Although the doublet has parameters that are not diagnostic of any specific Fe 3+ -sulfate, they do help constrain its identification. These results demonstrate that minor contamination of labile Fe 2+ sulfates by Fe 3+ can have dramatic effects on solution chemistry that should be considered when studying reactions relevant to acid mine drainage waste sites and other localities where Fe-sulfate minerals occur, such as the surface of Mars.