Volume 92, Issue 5-6

This paper reports a single-crystal FTIR spectroscopic study of vishnevite, ideally [Na 6 (SO 4 )] [Na 2 (H 2 O) 2 ](Si 6 Al 6 O 24 ), a member of the cancrinite group of feldspathoids. The study was done on several crystals from various geological occurrences. Infrared spectra show that most samples, and in particular the specimens from the holotype locality at Vishnevye Mountains (Urals, Russia), contain molecular CO 2 as the main carbon species in the structural pores, while the specimens from Loch Borolan (Scotland) were found to be CO 3 -rich. Polarized-light measurements show that the linear CO 2 molecules are oriented perpendicular to the crystallographic c axis. Structure refinement of sample Pi4 from Latium (Italy) shows usual H 2 O···Na···H 2 O sequences within the undecahedral cages; however, difference Fourier maps suggest the presence of additional protons in the channels, possibly forming OH groups. The FTIR spectra show three absorptions in the 3800.3200 cm -1 region. The first one, at 3590 cm -1 is strongly polarized for E||c while the second, at 3535 cm -1 , behaves almost isotropic. These two bands are assigned to the stretching vibrations of an asymmetric water molecule in the structural cages. The third broad absorption at 3320 cm -1 , is predominantly polarized for E||c and is assigned to additional OH groups in the channels. Detailed microspectroscopic mapping showed several samples from Latium (Italy) to be zoned with respect to the CO 2 /CO 3 content, thus pointing to a possible use of the volatile content of these minerals for petrological modeling.

We analyzed the microstructure, microchemistry, and microhardness variations across the architectural elements of the shells of the brachiopod species Megerlia truncata and Terebratalia transversa with scanning electron microscopy (SEM), electron backscatter diffraction (EBSD), laser-ablation inductively coupled plasma mass-spectrometry (LA-ICP-MS), and Vickers microhardness indentation (VMHI). The brachiopod valves consist of two principal layers of distinct calcite biomineralization: a thin, nanocrystalline, outer, hard protective layer with VMHI values exceeding 200 HV and a much thicker, inner, secondary layer of a hybrid organic-inorganic fiber composite material. The secondary layer is further structured into two sublayers, an outer part with VMHI values varying between 110 and 140 HV, and a softer inner part (70 < HV < 110). Whereas the size of the calcite crystals within the primary layer varies between a few tens of nanometers and 2 μm, calcite crystals within the secondary layer are fibrous, commonly reaching lengths exceeding 150 μm. Cross sections of these fibrous crystals are spade shaped, their dimensions being about 5 × 20 μm. The fibers are aligned parallel to each other. They are single crystals with their morphological fiber axes pointing almost parallel to the shell vault. The crystallographic orientation of the morphological fiber axes, however, is arbitrary within the a-b plane of the calcite lattice, whereas the c-axis (hexagonal unit-cell setting) is perpendicular to the morphological fiber axes and thus parallel to the radial vector of the valve vault. This morphology strongly indicates that fibrous growth is controlled by confinement within a cell in an organic matrix and not by attachment of biomolecules to specific crystallographic faces. We observe inhomogeneous Sr 2+ and Mg 2+ concentrations in the shell calcite within the 0.1.0.9 wt% range. Design of the shell appears to be highly optimized for mechanical performance. Crystal morphology and orientation as well as incorporated organic matter are structured hierarchially at different length levels forming a hybrid organic-inorganic fiber composite architecture.

Phase relationships in the system MgO-Al 2 O 3 -SiO 2 -Cr 2 O 3 (MASCr) were investigated experimentally from 1250 to 1560 °C using a 1-atm vertical quench furnace on 12 oxide mixtures covering a wide range of chemical compositions. Spinel, corundum, mullite, and sapphirine were found to display a wide range of solid solutions toward CR-rich end-members, whereas enstatite, cordierite, forsterite, and glasses were always Cr-poor. Glasses formed in relatively MgO-rich bulk compositions from 1250 °C, in assemblage with cordierite, enstatite, forsterite, and spinel, suggesting a lowering of the eutectic minimum in MASCr compared to available melting relationships in the model system MAS. Although unexpected, the relatively low-temperature melting of Cr-bearing mixtures is supported by topological constraints in the subsystems SACr and MSCr. The extent of Cr solubility in mullite has been found to reach ~14 wt% at 1560 °C, and 33 wt% in sapphirine at 1340 °C, which are the highest Cr values reported so far for these minerals. Liquidus surfaces are constrained by glass compositions that are displaced systematically toward silica compared to corresponding liquids in the ternary system MAS. As a consequence, the primary crystallization fields of Cr-spinel, Cr-sapphirine, and Cr-mullite are found to face the cristobalite/tridymite primary fields. A tentative liquidus diagram for the system MgO-Al 2 O 3 -SiO 2 -Cr 2 O 3 is proposed, which could serve as a basis for the interpretation of the high-temperature evolution of planetary materials as well as for the development of new ceramic materials.

The effect of the anisotropic connectivity within the layers of the post-perovskite, CaIrO 3 , structure are examined using a Rigid Unit Mode and group theory approach, and possible commensurate tilt structures are given. It is shown that the anisotropic bonding of the sheet may give rise to preferred directions of buckling for such structures and anisotropy in the speed of sound.

A broad (FWHM ≈ 7300 cm -1 ) intense band at ~21 700 cm -1 in the optical absorption spectra of natural Fe 2+ , Fe 3+ -rich garnets is attributed to electronic intervalence charge-transfer transitions (IVCT), VIIIFe 2+ + VIFe 3+ → VIIIFe 3+ + VIFe 2+ . In Fe 3+ , Fe 2+ -bearing garnets of predominantly almandine compositions, this band causes yellowish tinges in addition to the pink color, typical of pure Fe 3+ -free almandines. In garnets from deeper-seated mafic granulites from kimberlite pipes in Siberia with high skiagite (Fe 3 2+ Fe 2 3+ Si 3 O 12 ) contents, IVCT causes intense brownish-yellow colors. The relatively high energy of the band (~21 700 cm -1 ) compared to diverse minerals showing IVCT between Fe 2+ and Fe 3+ in adjacent octahedral sites, is attributed to the charge-transfer transition taking place between Fe 2+ and Fe 3+ in non-equivalent, dodecahedral and octahedral sites of the garnet structure. Band intensity is directly correlated with the product of Fe 2+ and Fe 3+ as measured by Mössbauer spectroscopy. The energy of the IVCT band is nearly independent of temperature, whereas its intensity decreases slightly with increasing temperature. Pressure induces a weak shift of the band to lower energies, Δν/ΔP ≈ -75 cm -1 /GPa, but intensity of the bands remains practically unchanged. Such temperature and pressure dependencies are quite different from those in other minerals showing IVCT between Fe 2+ and Fe 3+ in equivalent octahedral positions of structure.

The development of numerical modeling of reactive transport relies on the availability of thermodynamic properties for the solid, surface, aqueous, and vapor species stable at the conditions of interest. The lack of experimental studies and comprehensive activity-composition models severely limits the predictive capabilities of these models in systems involving highly saline fluids or low-density volatile-rich fluids. X-ray absorption near-edge structure spectroscopy (XANES) is a powerful technique to study the speciation of transition metals in aqueous fluids: it is element specific; sensitive to the oxidation state of the metal, the ligand field and the coordination geometry of the complex; and it is suitable for measuring trace amounts of metals (<1 wt%) in solutions over a wide pressure and temperature range. Formation constants for the aqueous complexes of transition metals can be determined from a series of XANES spectra obtained on solutions containing a constant amount of the metal of interest and variable concentrations of a ligand. The method relies on a non-linear, least-squares-fitting approach, with full distribution of species calculations based on a complete thermodynamic model for the experimental system under consideration. The technique is particularly suitable for following octahedral to tetrahedral transitions among weak chloride complexes of transition metals. The log K for the reaction Zn 2+ (octahedral) + 4Cl - = ZnCl 4 2- (tetrahedral) at 25 °C is retrieved to be 0.1(6), within error of the accepted literature value. The same method applied to Fe 2+ -chloride complexes shows that the log K for the reaction Fe 2+ (octahedral) + 4Cl - = FeCl 4 2- (tetrahedral) increases from -6.2(6) at 25 °C to -2.9(3) at 150 °C. This study confirms that tetrahedral chloride complexes play an important role in Fe transport in hypersaline brines especially at elevated temperatures, and shows that XANES is well suited to study systems that may be difficult to study with other techniques.

The polytypes of birnessite with a periodic stacking along the c* axis of one-, two-, and threelayers are derived in terms of an anion close-packing formalism. Birnessite layers may be stacked so as to build two types of interlayers: P-type in which basal O atoms from adjacent layers coincide in projection along the c* axis, thus forming interlayer prisms; and, O-type in which these O atoms form interlayer octahedra. The polytypes can be categorized into three groups that depend on the type of interlayers: polytypes consisting of homogeneous interlayers of O- or P-type, and polytypes in which both interlayer types alternate. Ideal birnessite layers can be described by a hexagonal unit cell (ah = bh ≈ 2.85 Å and γ = 120°) or by an orthogonal C-centered cell (a = √3 b, b h ≈ 2.85 Å, and γ = 90°); and, hexagonal birnessite polytypes (1H, 2H1, 2H2, 3R 1 , 3R 2 , 3H 1 , and 3H 2 ) have orthogonal analogs (1O, 2O 1 , 2O 2 , 1M 1 , 1M 2 , 3O 1 , and 3O 2 ). X-ray diffraction (XRD) patterns from different polytypes having the same layer symmetry and the same number of layers per unit cell exhibit hkl reflections at identical 2θ positions. XRD patterns corresponding to such polytypes differ only by their hkl intensity distributions, thus leading to possible ambiguities in polytype identification. In addition, the characteristics of the birnessite XRD patterns depend not only on the layer stacking but also on the presence of vacant layer sites, and on the type, location, and local environment of interlayer cations. Several structure models are described for birnessite consisting of orthogonal vacancy-free or of hexagonal vacancy-bearing layers. These models differ by their stacking modes and by their interlayer structures, which contain mono-, di-, or trivalent cations. Calculated XRD patterns for these models show that the hkl intensity distributions are determined by the polytype, with limited influence of the interlayer structure. Actual structures of phyllomanganates can thus be approximated by idealized models for polytype identification purpose. General rules for this identification are formulated. Finally, the occurrence of the different polytypes among natural and synthetic birnessite described in the literature is considered with special attention given to poorly understood structural and crystalchemical features.

Apatite fission-track length data are used extensively for thermal history inversion. However, several studies have documented instances of poor reproducibility of length data. We address this problem by using c-axis projection to normalize track lengths for crystallographic angle in the extensive laboratory annealing data set acquired by Barbarand et al. (2003a, 2003b). A new simplification reduces the c-axis projection model from six to four fitted parameters. Normalizing for track angle using c-axis projection improves every aspect of length measurement reproducibility examined. It accelerates convergence of mean length in single analyses; increases consistency among replicate measurements by a single analyst; enhances consistency of measurements of the same mounts by different analysts; and improves the match between analyses conducted with and without Cf-irradiation. C-axis projection is also shown to enhance the thermal sensitivity of length data. Based on these results, we assert that c-axis projection is a good means of compensating for observer bias, although it does not overcome differences caused by experimental error.

Two abiding issues impact the reliability of apatite fission-track analysis and thermal history inversion in particular: reproducibility of track-length measurements and variability of annealing kinetics. In the companion to this paper, we addressed the first by demonstrating that using c-axis projection to normalize track lengths for crystallographic angle improves reproducibility among measurements acquired by Barbarand et al. (2003a, 2003b). We continue here by looking at the effect of c-axis projection on extrapolation of laboratory measurements to geological time scales. We find that c-axis projection improves agreement between predictions of empirical annealing models fit to measurements by Barbarand et al. (2003b) analysts 1 and 3, further corroborating its effectiveness in ameliorating observer bias. Furthermore, these annealing models closely match predictions from the Carlson et al. (1999) annealing data set, which was obtained by a different analyst using a different etching protocol. Normalizing for angle thus allows us to create a single annealing model that encompasses both data sets and etching techniques. By combining the different compositional varieties of apatite in the two data sets, the resulting model provides an improved basis for estimating apatite kinetic properties in unknowns using proxies such as composition, etch-figure dimension, and unit-cell parameters. Predictions from a fanning curvilinear model form accord with conventional expectations of annealing at high and low temperatures, and its use for thermal history inversion will not introduce spurious late cooling events.

We present new Fourier Transform Infrared Spectroscopy (FTIR) and ion microprobe/secondary ion mass spectrometry (SIMS) analyses of 1H in 61 natural and experimental geological samples. These samples include 8 basaltic glasses (0.17 to 7.65 wt% H 2 O), 11 rhyolitic glasses (0.143 to 6.20 wt% H 2 O), 17 olivines (~0 to 910 wt. ppm H 2 O), 9 orthopyroxenes (~0 to 263 wt. ppm H 2 O), 8 clinopyroxenes (~0 to 490 wt. ppm H 2 O), and 8 garnets (~0 to 189 wt. ppm H 2 O). By careful attention to vacuum quality, the use a Cs + primary beam, and a resin-free mounting technique, we routinely achieve hydrogen backgrounds equivalent to less than 5 ppm by weight H 2 O in olivine. Compared to previous efforts, the new calibration extends to a wider range of H 2 O contents for the minerals and is more reliable owing to a larger number of standards and to characterization of anisotropic minerals by polarized FTIR on oriented crystals. When observed, discrepancies between FTIR and SIMS measurements are attributable to inclusions of hydrous minerals or fluid inclusions in the crystals. Inclusions more commonly interfere with FTIR analyses than with SIMS, owing to the much larger volume sampled by the former. Plots of H 2 O determined by FTIR vs. ( 1 H/ 30 Si) × (SiO 2 ), determined by SIMS and electron microprobe (EMP) yield linear arrays and for each phase appear to be insensitive to bulk composition. For example, basalt and rhyolite calibration slopes cannot be distinguished. On the other hand, calibration slopes of different phases vary by up to a factor of 4. This reflects either phase-specific behavior of 1 H/ 30 Si secondary ion ratios excited by Cs + ion beams or discrepancies between phase-specific FTIR absorption coefficient schemes.

The effects of cation substitution and ordering in the otavite (CdCO 3 ).magnesite (MgCO 3 ) solid solution have been investigated on samples synthesized at 1 GPa in the temperature range 500-800 °C for run durations up to 120 h. A complete, disordered solid solution, with R3̄c symmetry, was obtained at 800 °C, whereas Mg and Cd show partial ordering within the cadmium dolomite stability field, with R3̄ symmetry, at intermediate compositions in the temperature range 500-650 °C. Rietveld refinements for X-ray diffraction data show that variation of the a-axis is linear as a function of composition, independent of the degree of order, whereas the c-axis shows a positive deviation from linearity as a function of composition, decreasing with increasing degree of order. Octahedral bond distances of the 800 °C series vary linearly with composition. Site occupancies were used to determine the long-range order parameter, Q, for samples with R3̄ symmetry. Q 4 varies linearly as a function of temperature, suggesting a tricritical phase transition with a critical transition temperature T c of 702(10) °C. EDXTEM compositional microanalyses of samples within two-phase regions are in good agreement with Rietveld refinements, and allow better constraint of phase boundaries.

The local mixing and ordering behavior of the otavite-magnesite solid solution has been characterized using infrared powder absorption spectroscopy in the spectral region 50.2000 cm -1 at room temperature. Lattice modes due to Cd translations (75-200 cm -1 ) and Mg translations (200-350 cm -1 ) were both observed in the IR spectra for samples of intermediate composition. Peak positions of Cd translations remain constant as a function of composition, whereas those of Mg translations decrease in wavenumber with increasing otavite content. This suggests that the otavite-magnesite solid solution displays a combination of one-mode and two-mode behavior, most likely due to the large difference in atomic weight between magnesium and cadmium. Vibrational bands relating to the CO 3 2- groups in the spectral region 600-900 cm -1 vary linearly as a function of composition and are not sensitive to the degree of order. The vibrational band at ~1400-1450 cm -1 also varies linearly with composition, but R3 ordered samples show a marked increase in frequency compared to their R3c counterparts. Positive deviations from linearity are observed for the effective line width determined using the autocorrelation method (expressed in terms of Δcorr values) for both lattice modes and bending vibrations of the CO 3 2- molecular groups, whereas a linear variation as a function of composition of the Δcorr values of the CO 3 2- stretching mode is observed for the 800 °C solid solution. Only in the region 600.900 cm.1 is an effect of order observed as a reduction in Δcorr values with respect to the sample with R3c symmetry. The difference in Δcorr values between the ordered and disordered samples was used to determine the local order parameter, q, which is consistent with a tricritical order-disorder transition.

Olivine/melt partitioning of the transition metal cations, Fe 2+ , Mn 2+ , Co 2+ , and Ni 2+ , together with Mg 2+ and Ca 2+ , has been examined experimentally as a function of melt composition at ambient pressure. Melt structure was inferred from bulk-chemical composition, existing structural data, and 57Fe resonant absorption Mössbauer spectroscopy. Under isothermal conditions, K olivine/melt D(i-Mg) = (C i /C Mg )olivine/(C i /C Mg ) melt , is an exponential function of melt NBO/T for i = Ca 2+ , Mn 2+ , Co 2+ , and Ni 2+ . For i = Fe 2+ , the relationship is parabolic with maximum K olivine/melt D(Fe2+-Mg) lt-values at NBO/T near 1. At constant melt NBO/T, K olivine/melt D(i-Mg) increases systematically with decreasing cation radius, an effect that is more pronounced the more polymerized the melt. The K olivine/melt D(i-Mg) melt is also a positive and linear function of Na/(Na + Ca) of Al-free melts. This latter effect results from changes in Qn-species abundance governed by Na/(Na + Ca) of the melts. The enthalpy of the exchange equilibrium, i olivine + Mg melt = i melt + Mg olivine , derived from the temperature-dependence of K olivine/melt D(i-Mg) , is also a positive function of the ionic radius of the cation. The relationship of enthalpy to melt polymerization also depends on cation radius. The K olivine/melt D(Fe2+-Mg) lt does not, however, follow this trend possibly because the bond distance, dFe 2+ -O, in the melts depends on melt composition. The cations examined in this study are network-modifiers in silicate melts at ambient pressure. The solution behavior of network-modifying cations in melts is governed by the extent of steric hindrance near nonbridging oxygen, which in turn affects the energetics of metal-nonbridging oxygen bonds. Those structure effects, in turn, are related to the type of Q n -species, their abundance, and on the Al-distribution between the Q n -species. It is suggested, therefore, that the observed variations of mineral/melt partition coefficients with melt composition can be understood by considering bulk polymerization (NBO/T), the distribution of Al 3+ among coexisting Q n -species, and the distribution of network-modifying cations among nonbridging in these Q n -species.

The electron microprobe compositions and crystal structure of seven members of the grandidierite- ominelite (MgAl 3 BSiO 9 -Fe 2+ Al 3 BSiO 9 ) series with X = (Fe 2+ + Mn + Zn)/(Fe 2+ + Mn + Zn + Mg) ranging from 0.00 to 0.52 were studied to determine the geometric effects of Fe substitution for Mg on the crystal structures. Calculating Fe 3+ from the electron microprobe analyses gave 0.04.0.06 Fe 3+ apfu, but such small amounts at the Al sites could not be detected in the refinements. Regression equations derived from single-crystal X-ray diffraction data show that b increases by 0.18 Å from X = 0.1. The crystal structure refinements show that the most significant changes involve the (Mg,Fe2+)O5 polyhedron, which increases in volume by 0.36 Å 3 (5.0%), largely as a result of expansion of the MgFe- O5, -O2, and -O6 (×2) bond distances, which increase by 0.09 (4.4%), 0.06, and 0.04 Å, respectively. Other significant changes include increasing O1-MgFe-O2 (3.44°) and -Al3-O5a angles (1.9°) and a decreasing O6-MgFe-O6b (.2.20°) angle. Significant increases are also seen in the lengths of the O1-O2 (0.13 Å) and O6-O5a (×2) (0.09 Å) edges. The SiO 4 tetrahedra appear to respond to changes in the surrounding polyhedra by changing O-Si-O angles such that the tetrahedral angle variance and mean tetrahedral quadratic elongation increase with X. The BO 3 triangles appear to behave as relatively invariant units in the crystal structure. Regression equations obtained from the MgFe-O bond distances were used to determine a radius for V Fe 2+ of 0.70 Å. Although our samples show little Mn, the presence of Mn 2+ at the MgFe site would be expected to cause more distortion than an equivalent amount of Fe 2+ . Substitution of Zn likely would have little effect. The presence of Cr 3+ at any of the Al sites would be expected to increase the size of the coordination sphere, but the substitution of P 5+ for Si at the Si sites would most likely decrease the Si-O bond distances.

We have established a method to determine the oxidation state of Fe using the electron microprobe “the flank method.” This method has the possibility to analyze simultaneously the majorelement chemistry on the same spot. The method is based on FeL X-ray emission spectra that show a concomitant change of both the intensity and the wavelength of the soft FeLα and FeLβ emission lines. The method has been developed for garnets using well-characterized synthetic samples, and an empirical correction scheme has been established for self-absorption of the FeL X-ray emission. This enables the unambiguous correlation of measured intensity ratios of FeLα/FeLβ X-ray emission to the oxidation state in garnet. The flank method can be used with routine microprobe settings (e.g., 15 kV, 60 nA) and, for garnets, with minimum beam size. This way, Fe 3+ /ΣFe and major and trace elements can be determined within 10 minutes on the microscale. The resulting accuracy of Fe 3+ /ΣFe is then ±0.04 for garnets at 10 wt% total Fe. Application to natural garnet megacrysts with 8 wt% total Fe and Fe 3+ known from Mössbauer spectroscopy yields excellent agreement between the two methods. For practical application, the calibration can be achieved by using three natural standards and fitting the data to Fe 2+ and Fe 3+ with linear equations. Through repeated measurements on homogeneous samples, the error in Fe 3+ /ΣFe becomes approximately ±0.02 (1σ). Optimization of analytical parameters such as beam current, measurement time, and repetition of analyses, will further lower the uncertainty.

During structure solution of Ag-, Cu-bearing minerals it is quite common to observe disorder. Ag + and Cu + , indeed, can occur in different, but overlapping sites. The typical way to deal with these kind of minerals in structure determination is to use a split-atom model. This approach, however, has several disadvantages and may give rise to ambiguities. A solution to the problem can be the use of higher order tensor elements in the expression of the structure factors (the .non-harmonic approach.). This alternative approach gives, in cases of highly overlapping electron densities, an equivalent description of the split-atom model. The non-harmonic approach based upon a Gram-Charlier development of the atomic displacement factors can be useful in mineral sciences for the determination of still unknown structures. In addition, such an approach can be used to refine known structures with suspiciously high R values and/or high isotropic displacement parameters for the silver or copper atoms. As an example of the application of this method, we have reinvestigated the crystal structure of samsonite, Ag 4 MnSb 2 S 6 .

A new superstructure was found in mordenite (Na 5.59 Ca 1.80 Al 9.19 Si 38.81 O 96 ·nH 2 O) from Challis Valley, Idaho-a zeolite widely used in previous studies. The occurrence of the superstructure reflections were observed in 12 specimens from two rock samples, at the midpoint of layers in the oscillation photographs around the a* and b* axes, but not around the c* axis. The apparent 2a, 2b, and c axes are orthogonal. In spite of yielding an apparent orthorhombic cell, careful observation of the intensities revealed that the superstructure is monoclinic with twice the volume of the orthorhombic cell: a’ = 27.356(9), b’ = 27.356(9), c’ = 7.517(2) Å, γ′ = 97.14(4)°, and V′ = 5582(4) Å 3 . The presence of the superstructure was examined for specimens from ten other localities and a synthetic sample but there was no evidence for extra spots of the superstructure, indicating that the occurrence is not common. The large displacement factors of O8 oxygen, and its associated 180° T-O-T angle that is energetically unfavorable, are basic to the mordenite structure. In the monoclinic superstructure the O8 oxygen is decomposed into several asymmetric O atoms by the symmetry reduction. The large displacement smearing of O8 oxygen perpendicular to the straight T-O-T, are explained by the superposition of decomposed O atoms. The extra spots are predominantly observed in the reciprocal planes with l = odd in the same way as diffuse streaks. The observation of extra spots and diffuse streaks suggests that there are three kinds of domains in the mordenite, one in the ordered form and two others in the random form of c/2 linear displacements. The domain in the ordered form represents the superstructure. Thus, we propose a model for the superstructure in which the c/2 displacement occurs alternatively in the two basic unit cells slightly modified in the superstructure, yielding the periodic arrangements.

A series of experiments with and without sample retreatment was performed on a starting mixture of dolomite plus quartz at 0.5 GPa and 600 °C in a CO 2 -H 2 O fluid with a mole fraction of CO 2 of 0.2 for durations up to 582 h. The initial reaction of dolomite and quartz led to rapid formation of talc and calcite instead of tremolite and calcite by the intended reaction: 5 dolomite + 8 quartz + H 2 O = tremolite + 3 calcite + 7 CO 2 . With continued treatment, the talc + calcite + quartz assemblage gradually reacts to form increasing amounts of tremolite and diopside with the eventual loss of quartz and nearly complete loss of talc. The detailed structure of the amphibole and the nature of the Mg enrichment were revealed using AEM analysis of individual amphibole crystals. The most abundant defects are triple-chain chain multiplicity faults (CMFs), which appear as isolated lamellae with single periodicity in short-duration experiments and increasingly as groups of lamellae with variable multiplicity and periodicity in longer duration experiments and especially in a long-duration experiment without retreatment. In this latter experiment, a calcian clinojimthompsonite domain extending 15 unit cells along the b axis was observed. The sample after 582 h and five retreatments shows, on average, 3.5% true solid solution with Mg expressed as the Mg-cummingtonite (MC) component after correction for the presence of CMFs. This sample is thought to most closely approach the equilibrium composition for the amphibole. The sample after 250 h with three retreatments has about 10%, whereas that after 582 h without retreatment has 0.6% MC component, the latter having a relatively high density of CMFs. This study affirms the importance that precursor non-amphibole biopyriboles play in the formation of tremolitic amphibole.

In this paper, we report results of experiments on the dissolution of octahedral, pseudo-dodecahedral, and cubic natural diamond crystals in water-containing carbonate and silicate systems at high-pressure and high-temperature conditions in the diamond stability field. The dissolution agents used include CaCO 3 , CaMg(CO 3 ) 2 , CaMgSi 2 O 6 , and kimberlite from the Udachnaya pipe, Yakutia, with addition of distilled water. The obtained diamond dissolution forms were studied using scanning electron microscopy and double-beam interferometry. A quantitative analysis of rounded diamonds was carried out by the photogoniometry method. The experimental data show that diamonds change their morphology from octahedrons, dodecahedrons, and cubes to tetrahexahedroids when dissolved in water-containing systems. Octahedron transforms into tetrahexahedroid when the weight loss is 20.25%; cube, when the loss is >50%; and pseudo-dodecahedron passes into tetrahexahedroid when the weight loss is as low as 10%. Comparison of crystal morphology, surface features, and goniometric data of diamond dissolution forms produced in water-containing systems and of rounded natural diamonds showed their complete identity. It has been established that the morphological variations of rounded natural diamonds depend on the initial habit of the crystals and the degree of their dissolution. With the significant dissolution of the starting crystals the dissolution forms of initial octahedrons, pseudo-dodecahedron, and cubes are similar. The evolution of the diamond crystals morphology is terminated with the formation of tetrahexahedroid with curvature parameters АВ = 36°07′, СD = 13°15′, and DD = 13°15′. The obtained quantitative data allowed us to construct a scheme for the morphological evolution of natural diamond crystals during their dissolution.

The present paper reports changes to the existing nomenclature for minerals belonging to the pearceite-polybasite group. Thirty-one samples of minerals in this group from different localities, with variable chemical composition, and showing the 111, 221, and 222 unit-cell types, were studied by means of X-ray single-crystal diffraction and electron microprobe. The unit-cell parameters were modeled using a multiple regression method as a function of the Ag, Sb, and Se contents. The determination of the crystal structures for all the members of the group permits them to be considered as a family of polytypes and for all members to be named pearceite or polybasite. The main reason for doubling the unit-cell parameters is linked to the ordering of silver. The distinction between pearceite and polybasite is easily done with an electron microprobe analysis (As/Sb ratio). A hyphenated italic suffix indicating the crystal system and the cell-type symbol should be added, if crystallographic data are available. Given this designation, the old names antimonpearceite and arsenpolybasite are abandoned here and the old names pearceite and polybasite, previously defined on a structural basis (i.e., 111 and 222), are redefined on a chemical basis. The old name pearceite will be replaced by pearceite-Tac, antimonpearceite by polybasite-Tac, arsenpolybasite-221 by pearceite-T2ac, arsenpolybasite-222 by pearceite-M2a2b2c, polybasite-221 by polybasite-T2ac, and polybasite-222 by polybasite-M2a2b2c. Since all polytypes are composed of two different layers stacked along [001]: layer A, with general composition [(Ag,Cu) 6 (As,Sb) 2 S 7 ] 2- , and layer B, with general composition [Ag 9 CuS 4 ] 2+ , the chemical formulae of pearceite and polybasite should be written as [Ag 9 CuS 4 ][(Ag,Cu) 6 (As,Sb) 2 S 7 ] and [Ag 9 CuS 4 ][(Ag,Cu) 6 (Sb,As) 2 S 7 ], respectively, instead of (Ag,Cu) 16 (As,Sb) 2 S 11 and (Ag,Cu) 16 (Sb,As) 2 S 11 , as is currently accepted. The new nomenclature rules were approved by the Commission on New Minerals and Mineral Names of the International Mineralogical Association.

A detailed investigation of microstructures of authigenic illite has been carried out using highresolution TEM. Near-atomic images of 1M and 2M 1 polytypes of illite with incidence [100] and [110] were observed. We unambiguously determine 1M illite in both individual particles and microtwinned domains. Newly formed 2M 1 domains are surrounded by former packets of 1Md or 1M illite. Such intergrowth of different illite-polytypic domains on a scale of several layers suggests that they are metastable. The lateral-coherent 1M and 2M 1 illite domains are aligned along the same orientation and crystallographically continuous normal to [001]*. This is a new phenomenon observed for polytypic intergrowth and transition. The intralayer-stepwise change of stagger direction in a series of successive layers and consequently the interlayer-stacking angle change resulting in polytypic transition are hypothesized as an “intralayer-stepwise transition” model. It explains the process of illite polytypic transition in the solid state.

Peralkaline and peraluminous glasses close to the metaluminous join in the systems Na 2 O-Al 2 O 3 -SiO 2 (NAS), CaO-Al 2 O 3 -SiO 2 (CAS), and MgO-Al 2 O 3 -SiO 2 (MAS) have been examined with Raman spectroscopy. At least three different SiO 2 contents in each of the systems (NAS, CAS, and MAS) have been studied. Each series of glasses spans the metaluminous join at constant silica content. The spectra of glasses in all three systems show changes consistent with a continuous decrease in abundance of depolymerized species and an increase in fully polymerized species as compositions change from peralkaline to peraluminous. There is no evidence for maxima or minima of these abundances across the metaluminous joins for any of the studied series. These observations confirm previous suggestions that non-bridging O atoms are a general feature of .fully polymerized. glasses, and that a population of Al exists in the melt structure that is not associated with a charge-balancing cation, even in peralkaline compositions. With decreasing peralkalinity, new components appear in the Raman spectra of peraluminous NAS and CAS glasses, but such changes are much less obvious in the MAS system. These additional components are most likely related to the presence of “excess” Al (i.e., Al not associated with a chargebalancing cation). The structural role of this “excess” Al could be in the form of (Al, Si)-triclusters with different Al/Si-ratios depending on composition, but simple distorted AlO 4 -tetrahedra and possibly higher-coordinate Al 3+ cannot be ruled out from the Raman spectra alone. The implications of these results for isothermal variations in viscosity are considered. It is concluded that the diversity and complexity of viscosity variations observed in different aluminosilicate systems may be a consequence of the fact that the difference in bond strength between bridging and non-bridging O atoms in the Ca and Mg-bearing systems is small compared to that in the Na-bearing system.

The crystal structure of ingersonite, [Ca 2.93 Mn 2+ 1.06 Fe 2+ 0.01 ][Sb 5+ 3.95 Mg 0.05 ]F 0.15 O 13.85 , has been solved and refined in the space group P3121 [a = 7.282(2), c = 17.604(4) Å, V = 808.4(3) Å3, Z = 3] to R = 2.32% for 2219 Fo > 4σ(Fo) using MoKα X-ray data. The structure of ingersonite is isostructural with the synthetic weberite-3T polytype and related to the pyrochlore structure type. Both ingersonite and pyrochlore structures can be described as a sequence of pairs of polyhedral layers (named M and N), stacked along [111] and [001], respectively. In terms of the cation sites, M and N layers have general formula AB 3 and A 3 B, respectively, where B are the octahedral cations forming the B 2 X 6 framework of the pyrochlore structure and A are the larger, interstitial cations forming eightfold polyhedra in pyrochlore. In ingersonite, the M layers occur at z ~ 1/6, 1/2, and 5/6: The B octahedra are occupied by Sb 5+ and share corners to form a pseudohexagonal tungsten bronze (HTB) motif with the A position occupied by octahedral Mn2+ at the center of the pseudohexagonal rings. N layers occur at z ~ 0, 1/3, and 2/3, with A = Ca and B = Sb 5+ : Isolated B octahedra share 6 edges with 6 eightfold A polyhedra, to form a continuous sheet similar to the analogous layer in pyrochlore. The stacking of successive pairs of M and N layers in ingersonite is the same as in pyrochlore. Nonetheless, a difference in the relative position between M and N layers in ingersonite and pyrochlore is observed. The crystal-chemical relationships with other pyrochlore-related minerals are outlined.

Natural Al-saturated chlorite having a wide range of Mg:Fe ratios [Mg/(Mg + Fe) = 0.357 to 0.943] was studied to determine the effect of Mg-Fe substitution on the unit-cell parameters. With a nearly constant Al content [Al T /(Al T + Mg + Fe) = 0.362 to 0.416, where Al T represents total Al contents in both tetrahedral and octahedral sites], the only major variable affecting the unit-cell dimensions of natural chlorite is the Mg:Fe ratio. The value of about 0.4 for the Al content represents apparent Al saturation for chlorites in metamorphosed rocks. Unit-cell parameters were determined by least-squares refinement from 39 chlorite samples obtained over a large range in metamorphic grades, from X-ray data derived from Gandolfi pattern simulations (via single crystals and a single-crystal diffractometer) from this study and from traditional powder diffractometer data from the literature. Nine of the samples were analyzed for their Fe 3+ contents (= 0.147 to 0.304 atoms) by Mössbauer spectroscopy; chemical compositions were generally obtained by electron-microprobe analysis. Statistical tests show that the change of the crystallographic a and b axes are linear over the range of Mg:Fe ratios studied, whereas the crystallographic c axis and the β angle have no significant relationship to these Mg:Fe ratios. Equations relating the effect of Mg:Fe ratio on a and b are: a (Å) = −0.092 x + 5.408; x = Mg/(Mg + Fe) and b (Å) = − 0.162 x + 9.370, respectively. Electrostatic models involving Pauling.s electrostatic valency principle for Al-saturated chlorite were determined assuming: (1) Al-Al avoidance for Al-substituted tetrahedra, (2) no vacancies, and (3) no Al 3+ substitution in the M1 sites. Electrostatic restrictions occur for R 3+ → R 2+ substitutions in the M2 sites of the 2:1 layer. These restrictions produce an Al saturation in chlorite at Al T /(Al T + Mg + Fe) ratios of near 0.4. This saturation limit requires that 37.5% of the tetrahedral sites (1.5 out of 4 sites) are occupied by Al, and the electrostatic charge (electrostatic valency units, e.v.u.) close to +1 e.v.u. is favored for the interlayer to offset the net negative charge on the 2:1 layer.

Cr K-edge X-ray absorption near-edge structure (XANES) spectra were recorded on Cr:MgO periclase and Cr:(Mg,Fe)O ferropericlase synthesized at different pressures (4 and 12 GPa) and temperatures (1200 to 1400 °C) at reducing oxygen fugacity conditions (~iron-wüstite buffer IW to IW - 2), and on Cr:MgSiO 3 perovskite with 0.5 wt% Cr 2 O 3 . 57Fe Mössbauer spectra were collected on the Fe-containing samples. The aim of the study was to determine the Cr oxidation state in phases found in the Earth.s lower mantle, and to examine the possible relationship with the Fe oxidation state in the same materials. To calculate the amount of Cr 2+ , the intensity of the shoulder at the low-energy side of the edge crest was quantified using the area of the corresponding peak in the derivative XANES spectra (Berry and O.Neill 2004). In Cr:(Mg,Fe)O the relative Cr 2+ content reached at most 12.5% but results from Mössbauer spectroscopy combined with chemical composition data suggest that some Cr 2+ oxidized during cooling through the reaction Cr 2+ + Fe 3+ → Cr 3+ + Fe 2+ . In iron-free Cr:MgO, the Cr 2+ content is much higher and reaches ~40%. In Cr:MgSiO 3 perovskite with 0.006 Cr pfu (similar to estimated lower mantle abundance), chromium is mainly divalent.

Hydrogen incorporation in aluminous orthopyroxene may control the generation of melt and dominate the seismic properties at the base of the Earth.s lithosphere. To clarify the substitution mechanism of H, we have synthesized, characterized, and refined the crystal structure of this potentially significant variant of orthopyroxene. The experimentally produced crystals are small needles up to approximately 20 × 20 × 100 μm in size. Electron microprobe chemical analysis indicates about 11.7 wt% Al 2 O 3 . FTIR spectra indicate 7500 ppmw H 2 O with absorbance features qualitatively similar to natural mantle orthopyroxenes. TEM imagery indicates that the phase is pure orthopyroxene with low concentrations of defects and inclusions. Cell-parameter refinement from single-crystal X-ray diffraction gives a = 18.1876(7) Å; b = 8.7352(7) Å; c = 5.1789 (5) Å, V = 822.79(11) Å 3 , which is 1.2% smaller than pure Mg anhydrous orthoenstatite. The crystal structure has been refined from single-crystal X-ray intensity data measured using a rotating anode X-ray generator, micro-focused X-ray beam, and CCD detector system. The refined structure indicates about 5% vacancy in M2 and significant Al occupancy in both M1 and T2, consistent with its composition, (Mg 0.95,0.05 ) M2 , (Mg 0.79 Al 0.21 ) M1 , (Al 0.25 Si 0.75 ) T2 Si T1 O 6 . The existence of hydrous orthopyroxene in the mantle could absorb water released from olivine on decompression to delay the onset of melting in the spinel stability region in mantle peridotite compositions.