Volume 92, Issue 2-3

The usefulness of cathodoluminescence (CL) in mineralogy and petrology is largely due to its high sensitivity to variations in the chemical compositions of minerals. Difficulty in interpretation of CL images, however, often limits their use beyond the identiÞ cation of growth zones, which are then analyzed using other, more readily quantifiable, techniques. The use of scanning electron microscopy (SEM) CL in hyperspectral imaging mode, combined with simultaneously acquired X-ray composition mapping, extends the technique by allowing the separation of spectral features and their correlation with elemental composition. In this paper, we describe the use of such measurements in conjunction with multivariate statistical analysis to automatically identify and characterize zones in calcite and zircon. We demonstrate that this novel combination of techniques significantly increases the effective- ness of CL as a diagnostic tool for Earth science applications.

Cryptoperthitic alkali feldspars within the Proterozoic Klokken syenite have been pervasively altered deuterically to patch perthites whose constituent subgrains display very fine-scale zoning in optical cathodoluminescence (CL). Electron-probe analyses and hyperspectral maps were acquired from several areas of patch perthite to potentially identify the centers responsible for CL emission and to access information that the subgrains might provide about the late-stage geological evolution of the pluton. Each hyperspectral map is composed of tens of thousands of CL spectra covering a wavelength range of 350.850 nm, and images can be formed from any desired wavelength band. The patch perthite subgrains have two emission bands: one in the blue (~460 nm) and the other at red to infrared wavelengths (~690.725 nm). Cathodoluminescence images formed using both bands show that the optically visible zoning is developed only at blue wavelengths. Comparisons of emission intensities with electron-probe analyses show that the blue band is activated by Ti3+ and less than ~25 ppm of the trace element is required for optically detectable luminescence. Adjacent Ab- and Or-rich subgrains can have identical patterns of zoning at blue wavelengths, indicating that they crystallized from the deuteric fluid simultaneously, but in most cases, differences in zoning within any one area of patch perthite indicate that subgrains grew at different times and slowly relative to the frequent changes in concentrations of Ti within the deuteric fluid. The red to infrared CL emission band is inferred to be activated by Fe 3+ . Ab-rich feldspar with the greatest intensities of long-wavelength emission and FeO concentrations crystallized from a late-stage and relatively low-T fluid that obtained a proportion of its Fe from mafic grains. The generally poor correlation between variations in intensities of the two emission bands within any one subgrain indicates that concentrations of Ti and Fe in the deuteric fluid varied independently of each other, and one trace element has no sensitizing or quenching effect on emission from the other center. The maximum length of zones within individual subgrains (~0.5 mm) and maximum separation of subgrains with similar zoning patterns (~1 mm) helps to constrain the length and interconnectivity of the thin films of deuteric fluid that mediated alteration

The luminescence of three natural samples of leucophanite (ideally, NaCaBeSi 2 O 6 F) has been investigated with ion-beam excitation and by electron spin resonance (ESR). A blue luminescence, ascribed to a defect associated with SiO 4 or BeO 4 tetrahedra, and an orange luminescence, ascribed to an Mn 2+ center, dominates the emission. Further luminescence centers in the investigated wavelength range (200.1100 nm) include those related to Gd 3+ , Dy 3+ , Tb 3+ , Sm 3+ , Eu 3+ , Tm 3+ , and Nd 3+ . In spite of a sheet-like structure, leucophanite is relatively resilient to the ion implantation and β-irradiation, compared with other minerals, but loss of luminescence intensity as a function of implantation show that some permanent defects are formed. No centers change their emission energy as a function of temperature in the investigated interval (40.300 K), but due to crystal-field interaction, a broadening of the Mn 2+ emission band is observed with increasing temperature, and with low activator concentration. The variation in activator concentration is clearly seen from a pronounced concentration quenching. ESR data (9.7 and 188 GHz) show that Mn is only present in one crystallographic site and that the REE-richest sample has an additional signal with g ~ 2.003. No new defects could be seen with ESR in a β-irradiated sample.

Infrared luminescence is recognized in many minerals; however, few reported examples of infrared cathodoluminescence (CL) imaging are cited in the literature. Low-light digital web cameras can be modified by removing their infrared blocking filters to allow the imaging of minerals and rocks in the near-infrared (NIR). Preliminary investigation of a limited number of minerals has shown that many display cathodoluminesce in the NIR. Of the minerals found to have NIR-CL, the feldspars are notable as they commonly reveal hidden details of growth and alteration not easily observed with polarized light microscopy (PLM) or normal visible CL. Cold-cathode devices for CL have been a standard part of petrography labs for over 30 years. With the development of SEM systems for CL work, many of these older units have fallen into disuse. With the recent availability of low-cost consumer digital cameras ranging from low-light web cameras to consumer-grade megapixel cameras, most of these older CL units may be brought back into service to provide a powerful tool for petrographic research. The ease of capture of digital imagery, combined with the open-source image-processing software ImageJ, can produce high-quality images of visible and NIR cathodoluminescence of minerals in thin sections. Benefits of consumer digital cameras include the low cost, an order of magnitude below cooled cameras, and the speed and simplicity of digital imaging over film. This paper will discuss the use of modified web cameras and consumer-grade megapixel cameras for image capture and processing of visible CL and NIR-CL images.

Aluminum phosphate-sulfate (APS) minerals formed around the Athabasca unconformity-type deposits and those from their Australian counterparts are chemically very similar showing the same continuum between the diagenetic Sr-rich APS minerals of the barren sandstones and the LREE-rich composition of the APS minerals in the hydrothermally altered sandstone. The P- and LREE-rich compositions were controlled by the transport and the redistribution of P and LREE elements released from the dissolution of phosphate minerals (principally monazite) in the basement rocks and in the basin during the syn-ore alteration processes. The S/Sr ratio measured in the APS minerals from unaltered sandstone away from the unconformity and any mineralization is preserved during the syn-ore alteration processes suggesting that the fluids involved in both the deep burial diagenetic processes and the syn-ore alteration system were derived from a similar diagenetic reservoir in both the Athabasca and Kombolgie regions. The trioctahedral chlorite host-rock alteration around the Australian basement-hosted U deposits, as compared to the illite and sudoite associated with the Athabasca basement-hosted, along with the more LREE-rich APS compositions in the Australian deposits, suggests that the pH and oxygen fugacity (f O₂ ) of the syn-ore fluids differed in the alteration systems of the two regions at the time of the U deposition.

Studies of the hardness and toughness of minerals have historically focused on minerals of the Mohs scale, although, with the exceptions of quartz, orthoclase, and calcite, Mohs phases are not common rock-forming minerals. We report new hardness (H), toughness (resistance to fracture, K IC ), and indentation modulus (E*) data obtained by microhardness and depth-sensing indentation (DSI, or nanoindentation) experiments for common metamorphic minerals: sillimanite, kyanite, andalusite, garnet, quartz, and orthoclase feldspar. Because the experimental techniques involve indentation-induced cracking as well as depth-sensing indentation, the new data set can be used to investigate a range of plastic behavior for minerals in the crust and mantle. The three Al 2 SiO 5 polymorphs have similar H values (~10.12 GPa): kyanite has the largest values and andalusite the smallest. These values are similar to that of quartz (~12 GPa) and greater than that of orthoclase (~7 GPa). Garnet H values vary with composition: for grossular, H ~ 13 GPa, and for almandine-pyrope, H ~ 15 GPa. Although H values for the minerals we analyzed span a range of ~10 GPa, most fracture toughness values are between 1.1.8 MPa·m½. Garnet is much harder than Al 2 SiO 5 , but has a similar to slightly lower K IC (grossular ~1.2 MPa·m½; almandine-pyrope ~1.4 MPa·m½; andalusite 1.8 MPa·m½; sillimanite 1.6 MPa·m½); kyanite K IC is difficult to measure owing to the ease with which kyanite cleaves. Garnet has properties similar to those of cubic zirconia (ZrO 2 ), which we measured as a reference. Another reference mineral, periclase (MgO), has the lowest H (~5 GPa) and the highest K IC (~4 MPa·m½) of minerals we measured. Among the silicates, E* varies significantly from orthoclase (~89 GPa) to quartz (~117 GPa) to garnet (245.260 GPa), and Al 2 SiO 5 has intermediate values: kyanite ~ 186.253 GPa, sillimanite ~ 207 GPa, andalusite ~ 232 GPa.

Inspection of available experimental data reveals log-linear compensation effects between activation energies and pre-exponential factors for Ar, H, Pb, and Sr diffusion in a wide array of minerals. As a result, diffusion of Ar, H, Pb, and Sr converges to the same rates, respectively, at isokinetic temperatures in these minerals. Ionic porosity, Z, defined as the fraction of the unit-cell volume in a mineral not occupied by ions, is a measure of atomic packing density in silicate, carbonate, and phosphate minerals. Experimental diffusion parameters exhibit first-order correlations with ionic porosity, which proxies for mean metal-oxygen bond length/strength in minerals. An empirical kinetics-porosity model systematizes Ar, H, Pb, and Sr diffusion in minerals for which experimental diffusion data exist. For Ar and H diffusion, linear correlations are documented between activation energy and total ionic porosity. Combination of these correlations with diffusional compensation effects, which are also documented, yields empirical relationships among elemental diffusivity, total ionic porosity, and temperature. Linear correlations are also observed between experimental diffusion coefficients for Pb and Sr at given temperatures and calculated ionic porosities. For most minerals, the empirical predictions are remarkably consistent with experimental data, which strengthens the link between crystal chemistry and diffusion kinetics.

Landau theory has been used to develop expressions for the elastic anomalies that accompany octahedral tilting transitions in perovskites that are associated with the M and R points of the Brillouin zone. The master equation is a 246 Landau potential with saturation terms that provides phenomenological descriptions of transition sequences from a parent cubic structure through tetragonal or rhombohedral intermediates to orthorhombic or monoclinic product structures. Data from the literature have been used to determine values for all the coefficients required to generate a quantitative description of the Pm3̄ m ↔ I4/mcm transition in SrTiO 3 , which is taken as a model system. Solutions to the Landau expansion have been adapted to include the general influence of hydrostatic pressure and non-hydrostatic stress on transition temperature and the evolution of the order parameter. Critical examination of elastic constant data from the literature reveals inconsistencies between the results of measurements on tetragonal samples using ultrasonic rather than Brillouin scattering methods. An internally consistent data set has, nevertheless, been assembled. Good qualitative agreement was obtained between the general pattern of calculated and observed variations of all the single crystal elastic constants, and semi-quantitative agreement was obtained for C 11 , C 33 , C 12 , and C 13 . Some inconsistencies remain in relation to the temperature dependence of the square of the soft mode frequencies in the tetragonal phase, which follow the square of the order parameter rather than its inverse susceptibility, but the 246 potential seems to provide a good description of the structural evolution of SrTiO 3 over a wide temperature interval up to the cubic-tetragonal transition point.

A Landau free energy expansion has been developed in a first attempt to describe the Pm 3̄ m ↔ I4/mcm octahedral tilting transition in perovskite across a complete binary solid solution, CaTiO 3 - SrTiO 3 (CST). Only two parameters, the critical temperature and the coefficient, λ 4 , for coupling between the order parameter and the tetragonal strain are given a composition-dependence. The best match between observed and calculated variations of the tetragonal strain as a function of temperature in CaTiO 3 (CST0) and as a function of composition at room temperature is found for a model of the solid solution in which the transition is close to tricritical for the composition range CST0 to ~CST90. The change in transition character from second order (246 potential) to tricritical is accounted for by an increase in |λ 4 | between SrTiO 3 (CST100) and ~CST90. The Landau potential is used to calculate the variations of single-crystal and bulk elastic constants through the Pm3̄ m ↔ I4/mcm transition as functions of both pressure and composition. This model for the CST solid solution permits the classical softening effects of strain/order parameter coupling to be separated quantitatively from the much larger superelastic softening observed for twinned crystals of tetragonal SrTiO 3 and CST. Data from the literature for relative changes of elastic properties measured at low frequencies have been rescaled to absolute values for comparison with calculated values of Youngʼs modulus. These appear to show that, below the temperature for domain wall pinning, some 5.10% of softening of the isotropic Youngʼs modulus could still be due to the effects of twin wall movements.

Bulk and shear moduli of polycrystalline samples of perovskites with different compositions across the CaTiO 3 -SrTiO 3 solid solution have been measured at ambient conditions and in-situ at high pressures by pulse-echo ultrasonic methods. The samples were prepared as dense pellets by hot pressing synthetic powders at ~7.5 GPa and ~1000 °C. Any variations of bulk modulus due to phase transitions are small, but signiÞ cant anomalies have been observed in the shear modulus at ambient conditions. These are associated with a sequence of symmetry changes Pm3̄ m → I4/mcm → Pbcm → Pnma with increasing CaTiO 3 content. Comparison with variations in elastic properties predicted using Landau theory suggests that a substantial part of the elastic softening observed in tetragonal samples could be due to anelastic contributions from transformation twin walls. This additional softening does not occur in orthorhombic samples, and the transition from tetragonal to orthorhombic symmetry results in a stiffening of the shear modulus. No overt evidence was found for a phase transition I4/mcm ↔ Pnma at high pressures in Ca 0.35 Sr 0.65 TiO 3 but small changes in the trends of both bulk and shear moduli in the range 2.5.3 GPa could be due either to a different transition or a change in compression mechanism. A Pm3̄ m ↔ I4/mcm transition at ~2 GPa in Ca 0.05 Sr 0.95 TiO 3 shows the same form of softening as observed for the transition as a function of composition. A simple model of twin wall contributions to the compliance of tetragonal samples failed to match the observed variations that, alternatively, seem to follow ΔG ∞ q 4 where ΔG is the change in shear modulus and q 4 the driving order parameter for the Pm3̄ m ↔ I4/mcm transition. Analogous elastic behavior is expected to occur in (Mg,Fe)SiO 3 and CaSiO 3 perovskites at high pressures and temperatures.

Compositions of oxysalt, sulfosalt, seleniosalt, tellurisalt, and fluosalt minerals all follow the same generalizable patterns with regard to hardness or softness of cations, and with regard to ionic potential. For example, the hard F 1- anion forms simple fluosalt minerals (fluoborates and fluosilicates) with hard cations. The relatively hard O 2- anion forms simple oxysalts (e.g., sulfates and arsenates) largely with hard to intermediate cations. On the other hand, the softer S 2- anion forms sulfosalt minerals almost entirely with intermediate to soft cations, and the still softer anions Se 2- and Te 2- form seleniosalt and tellurisalt minerals with an even more predominately soft suite of cations. Thus, across all these classes of minerals, harder anions generally form minerals with harder cations, and softer anions form minerals with softer cations. The compositions of simple oxysalt minerals of one inter-radical cation vary greatly, but they do not include minerals in which both the inter-radical cation and intra-radical cation (e.g., Ca 2+ and S 6+ in CaSO 4 ) have high ionic potential. Their compositions thus reach, but do not exceed, a limit on a plot of the ionic potential of cations in these minerals. The same general relationship, wherein no simple minerals contain both inter-radical and intra-radical cations of high ionic potential, also exists among sulfosalts, seleniosalts, tellurisalts, and fluosalts. These limits of ionic potential of cations are greatest in oxysalts and progressively less in fluosalts, sulfosalts, seleniosalts, and tellurisalts. The limits thus decrease, both through the 2- anions and across the hard anions, from more negative ionic potential of the anion to less so, seemingly because anions of less negative ionic potential have lesser capability to shield cations from each other and thereby preclude cation-cation repulsion. Among the simple oxysalts of one inter-radical cation, three minerals near the oxysalt limit form less commonly, or with greater inhibition, than one might expect. For example, thermodynamic calculations indicate that magnesite should precipitate abundantly from seawater, but it does not instead, in low-temperature systems Mg 2+ only forms simple carbonate minerals with a cation of lesser ionic potential (Ca 2+ ). Similarly, the chemistry of Earth.s crust suggests that (Al,Fe 3+ ) 2 (SO 4 ) 3 would be a common mineral, but instead analogous hydrous minerals with additional cations of lesser ionic potential are common. Finally, SiO 2 minerals crystallize at distinctly low temperatures in magmatic systems, and they either fail to nucleate or precipitate as a hydrous disordered mineral in many low-temperature systems. In all three cases, cation-cation repulsion seemingly inhibits formation of what might otherwise be more common or more readily forming minerals

To explore the effect of variable CO 2 concentrations on the solidus of natural carbonated peridotite, we determined near-solidus phase relations of three different nominally anhydrous, carbonated lherzolite bulk compositions at 6.6 GPa. Starting mixes (PERC, PERC2, and PERC3) were prepared by adding variable proportions of a carbonate mixture that has the same Ca:Mg:Fe:Na:K ratio as the base silicate peridotite [MixKLB-1: Mg no. = 89.7; Ca no. = molar Ca/(Ca + Mg + Fe*) = 0.05]. For all three bulk compositions, the subsolidus assemblage includes olivine, orthopyroxene, clinopyroxene, garnet, and magnesite solid solutions. Above the solidus, crystalline carbonate disappears and quenched Fe, Na-bearing dolomitic carbonatite melts were observed. For PERC3 (1.0 wt% bulk CO 2 ; Na 2 O/CO 2 weight ratio = 0.30), the observed solidus is between 1190 and 1220 °C; for PERC (2.5 wt% bulk CO 2 ; Na 2 O/CO 2 = 0.12), it is between 1250 and 1275 °C; and for PERC2 (5.0 wt% bulk CO 2 ; Na 2 O/CO 2 = 0.06), it is between 1300 and 1330 °C. At 6.6 GPa, experimental solidi of natural magnesite peridotites are 100.200 °C lower than the CMAS-CO 2 solidus, chiefly owing to the fluxing effect of alkalis, and solidus temperatures increase with increasing bulk CO 2 (i.e., decreasing bulk Na 2 O/CO 2 ), owing to dilution of Na 2 O in near-solidus melt. The effects of Mg no. and Ca no. on carbonated peridotite solidi appear to be less significant than that of Na 2 O/CO 2 . Trends of decreasing solidus temperature with increasing Na 2 O/CO 2 and with decreasing CO2 indicate that natural mantle peridotite with ~100.1000 ppm bulk CO 2 will have solidus temperatures ~20° to ~100° lower than that determined experimentally. The solidus of peridotite drops discontinuously by ~600 °C (at 6.6 GPa) at the CO 2 bulk concentration (~5 ppm) at which carbonate is stabilized, but then varies little with increasing bulk CO 2 . This result contrasts with the effect of H 2 O, which lowers the solidus continuously with increasing concentration.

Time-resolved Rietveld refinements using synchrotron X-ray diffraction (XRD) have documented real-time changes in unit-cell parameters in response to cation substitution in synthetic Na-birnessite. Potassium- and Ba-birnessite, like Na-birnessite, were found to have triclinic symmetry. Rietveld analyses of the XRD patterns for K- and Ba-exchanged birnessite revealed decreases in the a, c, and β unit-cell parameters, with a decrease of 1.7 and 0.5%, respectively, in unit-cell volume relative to Na-birnessite. Fourier electron difference syntheses revealed that the changes in the configuration of the interlayer species, and the charge, size, and hydration of the substituting cations, serve as the primary controls on changes in unit-cell parameters. Split electron density maxima with centers at (0 0 0.5) were present for Na, K, and Ba end-members; however, with increased substitution of K + for Na + , the axis connecting the split-site maxima rotated from an orientation parallel to the b-axis to along the a-axis. Substitution of Ba 2+ for Na + did not result in rotation, but splitting of the interlayer site was more pronounced.

The heat capacity, CP, of a synthetic hydrous cordierite of composition Mg 1.97 Al 3.94 Si 5.06 O 18 ·0.625H 2 O was measured for the first time using precise adiabatic calorimetry in the temperature range from 6 to 300 K. Hydrous Mg-cordierite was obtained by hydrothermal treatment of anhydrous Mg-cordierite (Paukov et al. 2006) at 4 kbar and 600 °C for 24 hours. The synthetic product was characterized using X-ray diffraction and powder IR spectroscopy. Rietveld refinement gives a = 17.060(2) Å, b = 9.721(1) Å, and c = 9.338(1) Å with V = 1548.7(3) Å 3 and Δ = 0.25, and the IR spectrum shows only the presence of Class I-Type I H 2 O in the channel cavities. Small C P anomalies were observed at 272.98 ±0.03 K and 239.43 ±0.13 K, which are thought to be related to very small amounts of H 2 O occurring in tiny fluid inclusions and to surface H 2 O, respectively. From the heat-capacity data on hydrous Mg-cordierite, various thermodynamic functions were calculated and are presented in table form. The calculated partial molar entropy for one mole of H 2 O in hydrous Mg-cordierite at 298.15 K and 1 bar is 80.5 J/(mol·K). The partial molar volume for H 2 O in hydrous Mg-cordierite at 298 K and 1 bar is zero. The C P results, together with published heat-capacity data on three different zeolites, permit a comparison and analysis of their heat-capacity behavior. The heat-capacity behavior of H 2 O molecules in zeolites is more similar to that of ice at T < 300 K and not to gaseous H 2 O, which can be attributed to the presence of hydrogen-bonded H 2 O molecules. In contrast, the heat-capacity behavior for the .quasi-free. H 2 O molecule in cordierite is more similar to that of a free H 2 O molecule in the gaseous state between approximately 100 and 300 K. At T < 100 K, the energies of low-energy modes, especially external H 2 O translations, determine heat-capacity behavior. Model heat capacities for H 2 O in cordierite were calculated using the Einstein model and using as input data the results from inelastic neutron-scattering measurements on hydrous Mg-cordierite (Winkler and Hennion 1994). Reasonable agreement between experiment and calculations can be achieved using three H 2 O translational modes, one of which is hypothetical, and two librational H 2 O modes. The experimental spectra do not appear to show all six external H 2 O modes and further vibrational spectroscopic study is required to determine their energies. At T > 300 K, the heat capacity for H 2 O is the smallest in steam with values increasing in hydrous beryl and cordierite, to H 2 O in various zeolites, and finally to liquid H 2 O. This behavior may reflect the nature of the hydrogen bonding and the energies of internal H 2 O stretching modes, which decrease in energy with increasing hydrogen-bonding strength in the various systems

The nature of the chemical interactions between organic species and mineral surfaces and interlayers has significant implications for many geochemical processes and for medical and materials applications, but the molecular-scale characteristics of such interactions are poorly understood. Here we describe an experimental investigation of citrate-hydrotalcite (HT) that is designed to investigate the structural environments and dynamical behavior of carboxylic species interacting with the protonated surfaces of positively charged materials such as layered double hydroxides (LDHs). X-ray diffraction, compositional analysis, and thermal analysis confirm successful intercalation of trivalent citrate anions into the interlayers of the Mg/Al hydrotalcite using an ion-exchange method. The basal spacing changes from 8.9 Å at relative humidity (R.H.) = 0% to 11.3 Å at R.H. = 75%, to 18.9 Å at 100% R.H., consistent with reorientation of the citrate from parallel to the hydroxide layers at low water contents to perpendicular to them at higher water contents. At the largest water contents, the citrate may lose contact with one of the hydroxide layers, in contrast to the behavior of LDHs containing small, inorganic species. The similarity of the 13 C nuclear magnetic resonance (NMR) chemical shifts for citrate-HT and sodium citrate solution demonstrates the absence of grafting (covalent bond formation) between the citrate molecules and the metal hydroxide layers, and the dominance of Coulombic and H-bond interactions in stabilizing the structure. Decreasing 13 C NMR peak widths, decreasing signal/noise ratios in the 1 H- 13 C CPMAS NMR spectra, and increasing basal spacing with increasing R.H. are consistent with increasing dynamical disorder of the interlayer citrate sites in a progressively more water-rich interlayer

The crystal structure of a terrestrial 6H-SiC moissanite has been refined in the P6 3 mc S.G. from area detector single crystal X-ray data, down to an R-index on the observed reflections of 0.0205. The cell parameters refined over all the collected reß ections are a = 3.0810(2) and c = 15.1248(10) Å. The average Si-C bond lengths are 1.8898 Å, with average bonds along the stacking direction (1.8993 Å) slightly longer than those along the bilayer (1.8862 Å). The interlayer distances, defined as the distances along [0001] between Si-Si layers, which may occur either in cubic (c) or hexagonal (h) configurations, are maximal at the c-h interface (2.5270 Å) and minimal at the h-c interface (2.5165 Å), entailing that the h-bilayer is not equidistant from either c-bilayers. All the tetrahedral angles are identical within the experimental error and close to the ideal value of 109.47°, but those at the c-h interface, where a significant distortion of 0.15° is recorded. Finally, the anisotropic displacement factors are utterly very small, identical among different atoms within the experimental error, and significantly spherical. It thus appears that the 6H-SiC structure is affected by a slight relaxation along the [0001] stacking direction with respect to the ideal cubic structure, and that the relaxation is mainly accomplished at the c-h interface, i.e., at the twin-like boundary, where a bilayer in cubic configuration links a bilayer in antiparallel, hexagonal configuration. As far as we know this is the first crystal structure refinement of a natural 6H-SiC moissanite. Possible implications on the polytype stability in the light of these results are briefly discussed.

Mullite samples doped with cobalt were derived from diphasic gels with constant atomic ratio (Al + Co)/Si = 3:1, where 0, 1, 2, and 3 at% of aluminum was replaced by cobalt. X-ray powder diffraction showed that the samples contained mullite phase and some amount of α-Al 2 O 3 (for pure and doped samples) and CoAl 2 O 4 (for doped samples). Cobalt doping caused an increase in unit-cell parameters of the mullite phase. Transmission electron microscopy and energy dispersive X-ray spectroscopy were used for sample microanalysis and determination of the chemical composition of the Co-doped mullite phase. The Rietveld method was performed for quantitative phase analysis of the samples and for structure refinement of the mullite phase in the samples. It was found that a small amount of Co 2+ , 0.36 wt%, substituted for Al 3+ in the AlO 6 octahedra of the mullite structure. Simultaneously, the same amount of tetrahedral Al 3+ ions was likely substituted by Si 4+ in the (Al,Si)O 4 tetrahedra for the purpose of charge compensation. The remaining cobalt reacted with alumina forming CoAl 2 O 4 , and dissolved in the glassy phase. The proposed formula for Co-doped mullite is Co y Al 4+2x-2y Si 2 - 2x+y O 10-x

Thermal expansion of Cr-doped (0, 2.30, 4.74, 6.97, and 9.17 wt% Cr 2 O 3 ) mullites derived from single-phase precursors was studied by in situ (480.940 °C) powder X-ray diffraction. Chromium doping caused an increase in unit-cell parameters. The strongest expansion due to Cr incorporation into mullite was noticed along the c-axis, followed by a and b. In the temperature range from 480 to 940 °C, the lattice parameters expanded almost linearly for all prepared mullite samples. Thermal expansion of Cr-doped mullite was smaller than of pure mullite. Both undoped and Cr-doped mullites displayed strongest thermal expansion along the b-axis, followed by c and a. Coefficients of thermal expansion along the a, b, and c-axes displayed minima for mullite doped with 6.97 wt% Cr 2 O 3 .

Armbrusterite, ideally K 5 Na 6 Mn 3+ Mn 2+ 14 [Si 9 O 22 ] 4 (OH) 10 ·4H 2 O, is a new silicate of potassium, sodium, and manganese found in a thin cancrinite-aegirine-microcline vein within urtite at Mt. Kukisvumchorr. The mineral occurs in intimate association with raite. Other associated minerals are lamprophyllite, mangan-neptunite, pectolite, vinogradovite, calcite, molybdenite, galena, sphalerite, and fluorite. Armbrusterite occurs as split, curved crystals and spherulites (≤2 mm diameter). The mineral is translucent (transparent in thin fragments), dark reddish-brown. It has vitreous luster and light-brown streak. Cleavage is perfect on (001) and the fracture is uneven. Mohs hardness is about 3.5. In transmitted light, the mineral is reddish-brown, with strong pleochroism: X = light yellowish-brown, Y and Z = dark reddish-brown; dispersion r > v, weak. Armbrusterite is biaxial (.): α = 1.532(2), β = 1.560(2), γ = 1.564(2) (for λ = 589 nm), 2V varies from 10° to 20°. Optical orientation: X is perpendicular to (001). The mean chemical composition determined by electron microprobe and the Penfield method (for H 2 O) is (wt%): Na 2 O 5.26, MgO 0.19, Al 2 O 3 0.04, SiO 2 56.02, K 2 O 6.13, CaO 0.26, TiO 2 0.04, MnO 23.62, Mn 2 O 3 2.07, FeO 0.65, ZnO 0.20, H 2 O 4.1, sum. 98.58. Empirical formula calculated on the basis of Si = 36 is K 5.03 Na 6.55 (Mn 2+ 12.86 Mn 3+ 1.01 Fe 2+ 0.35 Mg 0.18 Ca 0.18 Zn 0.09 Al 0.03 Ti 0.02 ) Σ=14.72 [Si 36 O 88 ](OH) 10.10 ·3.75 H 2 O. Armbrusterite is monoclinic, C2/m, a = 17.333(2), b = 23.539(3), c = 13.4895(17) Å, β = 115.069(9)°, V = 4985.4(11) Å 3 , Z = 2. The strongest X-ray powder-diffraction lines are [d in Å, (I), (hkl)]: 12.28 (100) (001), 4.10 (10) (003), 3.562 (10) (113, 261), 3.260 (18) (114), 3.117 (13) (203), 3.077 (54) (004), 2.622 (10) (371). The crystal structure of armbrusterite was refined to R1 = 0.085 on the basis of 3960 unique observed reflections. The structure is based upon double silicate [Si 9 O 22 ] layers consisting of 5-, 6-, 7-, and 8-membered tetrahedra rings. The layers are linked via octahedral sheets formed by Na and Mn octahedra. The interior of the double silicate layers is occupied by K + cations and H 2 O molecules. The mineral is named in honor of Thomas Armbruster (b. 1950; University of Berne) for his outstanding contribution to structural mineralogy and crystallography, especially to the study of Mn-rich minerals.

The elastic properties of ceramics and a mix of minerals can be calculated in a simple way including some properties of the interfaces between grains. While the Hashin Shtrikman averaging method has bounds, which are strictly correct for many realistic physical conditions, one finds empirical data which lie outside these bonds. One possible interpretation is that the scaling of the volume proportions in mineral assemblies is not truly represented by the nominal volume proportion f of each phase. It is argued that interfacial effects scale with f(1-f). In the case of certain assemblies, it is shown that the scaling is entirely with f 2 rather than f. For intermediate cases, a more realistic scaling replaces f in the relevant averaging schemes by f(1-S) + S f 2 where S weighs the effect of the non-linearities in the volume expansion of the averaging schemes.

Optical and near infrared absorption spectra of ferropericlase Mg 0.88 Fe 0.12 O have been measured to 84 GPa. Under ambient conditions, the spectrum shows two crystal field bands of high-spin Fe 2+ at 8922 and 12 533 cm -1 , which shift to higher frequencies with increasing pressure (dν/dP = 50.7 and 85.5 cm -1 /GPa). Simultaneously, the intensity of the high-frequency band continuously decreases until it vanishes around 40 GPa, suggesting a quenching of the Jahn-Teller effect. Between 51 and 60 GPa, the absorption spectrum changes drastically. Two new bands appear at 60 GPa at 9728 and 14 592 cm -1 with frequency shifts at higher pressures of dν/dP = 23.8 and 21.0 cm -1 /GPa, respectively. If the change in optical spectra between 51 and 60 GPa were interpreted as being due to spin-pairing, the crystal field parameters of low-spin Fe 2+ at 60 GPa would be Δ = 10 546 cm -1 and B = 377 cm -1 . This would imply that the main cause of spin-pairing is not the increase in crystal field splitting Δ, but the stronger covalency of the Fe-O bond as seen in the reduction of the Racah parameter B. Even at 84 GPa, ferropericlase is by no means opaque. In particular, the inferred spin-pairing transition between 51 and 60 GPa reduces radiative thermal conductivity only by about 15%. Spin-pairing in ferropericlase is therefore unlikely to have major consequences for the temperature distribution or the mode of convection in the lower mantle. The absorption edge of the high-pressure phase appears to be deeper in the UV than for the low-pressure phase, which could imply a reduced electrical (polaron) conductivity

We present results of an acoustic emissions (AE) study of single crystals of chromian-clinochlore during compression to 10 GPa at room temperature. Distinct AE are detected associated with the type Ia → Ib transition at 6 GPa. Analysis of AE source locations and first motions at transducers are consistent with a rapid collapse of the c-axis of the sample and AE originating within the sample and not the surrounding pressure medium. This is the first time that AE have been detected directly from a phase transformation in the multi-anvil press and opens new possibilities for kinetic studies and studies of deep-seismogenesis

We discovered a new CaGe 2 O 5 polymorph in high-pressure experiments (above 8 GPa). The phase is orthorhombic, space group P bam , with a = 7.306(2), b = 8.268(2), c = 5.714(1) Å, V = 345.2(1) Å 3 , and Z = 4. The new phase, which we call post-titanite CaGe 2 O 5 , is the high-pressure polymorph of titanite CaGe 2 O 5 . The structure of this new polymorph is based on a network of 5- and 6-coordinated Ge polyhedra and 8-coordinated Ca atoms. Following the germanate analog to silicate, post-titanite CaSi 2 O 5 could be expected to form at high-pressure conditions and thus might exist in Earthʼs mantle.

Natural jarosites selected for study have mixed domains of jarosite, KFe 3 3+ (SO 4 ) 2 (OH) 6 , and natrojarosite, NaFe 3 3+ (SO 4 ) 2 (OH) 6 . Minor elements include Al in the octahedral B-site, and P, As, Mo, and V in the tetrahedral T-site. High abundances of As are detected in some samples. Oscillatory zoning of Na and K in a subset of these samples has been explored using BSE images, X-ray maps, and a 1 μm EDS beam. Our work shows that zoning is composed of less than micrometer-sized bands of near end-member compositions. This agrees with the XRD work of Desborough et al. (2006), where 32 natural hypogene and supergene jarosites were found to be mixtures of near end-member compositions and showed <5% solid solution. This indicates a wide solvus (miscibility gap) between jarosite and natrojarosite. It also suggests that special crystallization effects are active in solid-solution and aqueoussolution interactions. Here, the optimal conditions for the crystallization of end-member compositions are at low temperatures (<100 °C), and when the two end-members of a binary solid-solution series have different solubilities, as do jarosite and natrojarosite (Glynn 2000). These conditions are commonly found in supergene environments, and are best illustrated by spectacular oscillatory zoning of jarosite and natrojarosite in samples from the Apex Mine, Arizona and Gold Hill, Utah.