Band 83, Heft 5-6

Single-crystal and powder X-ray diffraction data were collected to characterize the mac­roscopic solid-solution and cation-ordering behavior in the system augite-jadeite (low ac- mite content). We examined 28 natural pyroxenes with compositions on the join augite- jadeite and with different degrees of order. Annealing experiments were carried out to obtain crystals with different degrees of order (P2/n) and complete disorder (C2/c) at compositions between 35 and 60% Jd. Three synthetic C2/c pyroxenes with composition Di 80 Jd 20, Di 60 Jd 40 . and Di 50 Jd 50 were also examined. The long-range order parameters Q M1 and Q M2 of the M1 and M2 sites were obtained by a minimization procedure combining single-crystal X-ray diffraction data and chemical analyses. For both C2/c and P2/n pyroxenes, the a, b, c lattice parameters and unit-cell volume, as well as tetrahedral and octahedral mean bond distances depend linearly on composition. Only the angle β of ordered omphacites slightly deviates from the linear trend of the C2/c samples. The out-of-plane tilting of the basal face of tetrahedra is sen­sitive to the different degrees of order.

Infrared spectra of ordered (P2/n) and disordered (C2/c) pyroxenes belonging to the join augite-jadeite were investigated at room temperature in the range 70-1400 cm -1 . The spectra change as a function of composition and the phase transition P2/n → C2/c produces an increase in band widths. The autocorrelation function quantified changes in line widths due to these effects. Phonons at low frequencies are very sensitive to both changes in composition and degree of order and are indicative of a non-ideal mixing behavior for the C2/c solid solution. Phonons at high frequencies depend on the average composition of the samples and not on local configurational changes related to the order-disorder transition. High degrees of local heterogeneity were suggested by the Δ corr values of the C2/c omphacites at intermediate compositions. This effect is most evident in the low frequency region of the spectra, corresponding to modes involving, primarily, the M cations. The changes in frequency due to ordering, Δω, and the change in line widths δ(Δ corr ) 100-200 for the spectral region 100-200 cm -1 and δ(Δ corr ) 210-800 for the range 210-800 cm -1 have been used to characterize the state of local order at different compositions. The spectra yield information related to the local structural states of augite-jadeite pyroxenes, down to a unit-cell length scale, which may then be compared with the average structure determined by X-ray diffraction studies of the same samples.

A polycrystalline sample of MgSiO 3 -orthopyroxene was hot pressed at a pressure of 4 GPa and temperature of 975 °C in a multi-anvil apparatus. The recovered specimen has a bulk density within 1% of the X-ray density and compressional and shear wave velocities within 1% of the Hashin-Shtrikman averages of the isotropic velocities calculated from the single-crystal elastic moduli. Compressional and shear wave travel times were mea­sured at pressures up to 10 GPa at room temperature using the phase comparison method of ultrasonic interferometry in a 1000 ton uniaxial split-cylinder apparatus (USCA-1000). The velocities and elastic moduli monotonically increase with pressure; discontinuous be­havior is not observed. Both the compressional velocity and bulk modulus, however, ex­hibit pronounced non-linear dependence on pressure. Fitting a fourth-order Eulerian finite- strain equation of state yield values of the bulk modulus and its first and second pressure derivatives, K 0 = 104(2) GPa, K′ 0 = 10.9(5), and K″ 0 = -1.6(2) GPa -1 . In contrast, the shear velocity and modulus are linear with pressure, yielding values of G 0 = 74.9(1.5) GPa and G′ 0 = 1.6(1), when fit to a third-order finite-strain equation of state. A P-V trajectory calculated from these continuous measurements of K vs. P to 8 GPa is in agreement with extant static compression data for this material, without requiring a discontinuous change in K at 4 GPa. The velocities of MgSiO 3 -orthopyroxene increase more rapidly with pressure than those for Mg 2 SiO 4 -olivine (especially for P-waves), such that these two mineral phases are virtually indistinguishable in their velocities at 200 km depth, unless the velocity- temperature dependence of the two phases is dramatically different.

Examination of X-ray diffraction intensities for FeO collected in situ at high pressure and high temperature reveals that the atomic arrangements of FeO in the hexagonal structure are not the same as those in the simple NiAs-type structure (B8) where Fe takes the place of Ni. The observed diffraction intensity can be explained by adding an anti-B8 component (where Fe takes the place of As). Substitution of Fe and O atoms for each other is crystallochemically unique. The exchange of Fe and O positions provides a critical measure of the change in chemical bonding. Our conclusion is consistent with the observed transition of FeO from an ionically bonded structure to a strongly covalent and metallic one. First-principles electronic structure computations using the linearized augmented plane wave (LAPW) method with the generalized gradient approximation (GGA) indicate that both B8 and anti-B8 should be antiferromagnetic, but only anti-B8 should be an insulator. The GGA and local density approximation (LDA) incorrectly compute anti-B8 as the ground state structure.

The crystal structures and M-site populations of a series of unheated and heat-treated, near-binary members of the actinolite series were determined from single-crystal X-ray diffraction data. Small, systematic differences in crystal structure (mean and individual bond lengths, octahedral distortions, and Mg-Fe site preferences) between the actinolite and cummingtonite series are documented and rationalized in terms of the bond-valence model. Mg-Fe 2+ site preferences among the M1, M2, and M3 sites in the actinolite series are small. Between M1 and M3, site preference (Fe 2+ slightly in favor of M1) is indepen­dent of temperature. Mg is favored in M2 over M1 in unheated samples (less strongly than in cummingtonite-grunerite), but Fe 2+ and Mg are virtually disordered in samples re­equilibrated at 700 °C. Site preference between M2 and M3 and possibly also between M1 and M2 appears to undergo a reversal as a result of heat treatment, but consideration of all sources of error renders this conclusion a tentative one. although the observation could be explained in terms of a bond valence deficiency. The ratio of Fe 2+ /Mg atoms on M4 cannot be accurately determined, but refinements indicate that, like cummingtonite-gru­nerite. it is larger than on all other M sites. Expansion of the cell dimensions with an increase in Fe/Mg ratio follows a pattern similar to that of thermal expansion, and a and (3 are influenced by cummingtonite-grunerite solid solution. The data are consistent with a cell volume of 942.8 Å 3 (or molar volume of 2.8388 J/bar) for end-member ferro- actinolite. a first approximation that assumes a reciprocal, quadrilateral ΔV of zero.

The rigid unit mode analysis was initially developed to understand the phase transitions in aluminosilicate minerals containing corner-linked tetrahedra. Here the model is applied to a range of minerals with crystal structures that can be described as frameworks of linked octahedral and tetrahedral units, including garnet, sillimanite, titanite, and ellenbergerite. Consistent with a constraint analysis, there are no rigid unit modes in these minerals. Generalizing these results suggests that there will not be any easy ways to distort structures that have frameworks containing octahedral units. This result explains why polyhedraltilting displacive phase transitions are not common in such minerals, whereas they occur in most aluminosilicate minerals containing only tetrahedral units. It also explains that it will be necessary for the tetrahedra and octahedra to distort when solid solutions are formed.

Within the framework of the Adam-Gibbs (configurational entropy) theory of viscosity, it is shown that for a given composition, the ratio of parameters B e (a temperature inde­pendent constant) to S c (T g ) (the configurational entropy at the glass transition) is propor­tional to the height of the average potential energy barrier to viscous flow (Δμ) and the size of the rearranging domains at the glass transition [z*(T g )]. The ratio B e /S c (T g ) is eval­uated for several silicate and aluminosilicate compositions of variable polymerization. It is found that the ratio B e /S c (T g ) shows simple compositional variations that correspond closely to those that may be expected qualitatively for the height of the potential energy barrier to viscous flow. Assuming that z(T g ) is a constant for all compositions, the avail­able data for B e /S c (T g ) are parameterized as a function of Δμ. The physical basis of this parameterization will therefore allow extension to more complex systems as additional data become available. The A e term in the Adam-Gibbs equation (another temperature independent constant) shows only minor compositional variation (-2.6 ± 1). but the vari­ation that does exist is found to be a linear function of B e/ tetrahedron. The proposed parameterizations of B e /S c (T g ) and A e are shown to be sufficient for estimating the glass transition temperature to within 15-20 K. Calculated glass transition temperatures may be combined with existing models for viscosities in the range 1O-1O 5 Pa·s. Interpolation pro­vides the whole viscosity curve and thus also an estimate of the departure from Arrhenian behavior. Although further work is necessary to verify) and extend the parameterizations to compositions of direct geological relevance, this work represents a step toward a fully generalizable predictive model of silicate melt viscosity based within a physical framework.

The results of multi-anvil melting experiments are reported for a range of compositions in the system CaO-MgO-Al 2 O 3 -SiO 2 . The liquidus crystallization fields for forsterite, or­thopyroxene. clinopyroxene. and garnet have been mapped out at 10 GPa. as have their intersections at various cotectics. The composition of the liquid that is multiply saturated in forsterite. orthopyroxene, clinopyroxene. and garnet has been also determined to within ±0.5-1.0 wt% (2σ). and the result is in excellent agreement with a previous estimate (Herzberg 1992). These experiments confirm that the effect of pressure is to reduce Al 2 O 3 and increase MgO and SiO 2 in magmas formed by the melting of garnet lherzolite with increasing pressure (Herzberg 1992). Melting experiments in the system MgO-SiO 2 also have been performed to constrain how pressure affects the compositions of liquids that are saturated in harzbiugite [L + Ol + Opx]. Experiments in both CaO-MgO-Al 2 O 3 -SiO 2 and MgO-SiO 2 demonstrate that there is a maximum normative olivine content to liquids formed by initial or advanced melting of peridotite in the upper mantle, and this occurs at 7 to 8 GPa. For most peridotites that undergo decompression melting in a plume, clino­pyroxene and garnet are the first crystalline phases to melt out and. with a few important but rare exceptions, the experimentally constrained liquids are unlike most volcanic rocks. Advanced anhydrous melting will yield liquids with a residual harzburgite mineralogy [L + Ol + Opx]. and these liquids are similar in composition to most komatiites with Cre­taceous and 2700 Myr Archaean ages.

Startling results were obtained in kinetic experiments that varied the initial heating history of the mineral reaction: 1 dolomite + 2 quartz → 1 diopside + 2CO 2 . The experiments were earned out at 5 ± 0.05 kbar. X CO ₂= 0.9. and 680 ± 3 °C. which is 65 °C higher than that of equilibrium. The heating history was varied from the conventional procedure [about I h reported by Lüttge and Metz (1991)] by various techniques. The experimental temperature in externally heated cold-seal pressure vessels is attained within 10 min by preheating with a furnace miming at 1000 °C, and quickly switching to a furnace set at the temperature of the experiment when T reaches 670 °C. Percent conversion, measured by CO 2 production as a function of time for the period 5 to 288 hours, is linear and fit by the equation, α = kt, with k = 5.64 10 -2 [%conversion h]. Compared with results produced by the conventional heating procedure, our conversion rate is slower and results show an order of magnitude less scatter. The variation in the magnitudes of the rates measured after the fixed experimental conditions were reached for different heating procedures as well as the large differences in the degree of the scatter of the experimental rate data for different heating procedures are contrary to earlier claims made by most experimentalists and required a rethinking of our kinetic interpretation of the data. A kinetic model developed to simulate the initial period of a mineral reaction experiment provided a reasonable explanation of many of the unusual facets of the experimental re­sults. The model follows the variation of concentration, ΔG of the reaction, surface area, nucleation rates, and volume of the product as a function of time for the different heating histories used in the experimental study. The key problem ignored in earlier work is the significant effect of a complex nucleation history during the heating conditions on the entire subsequent kinetics, even though the latter are earned out for much longer times and under constant P and T conditions. Results show that faster heating rates cause faster nucleation rates, and the duration of the nucleation period is shorter. Two different sce­narios were compared: (1) a reaction product growing independently of the reactant surface (as may occur in hornfelses with fine grain size); and (2) a reaction product growing on the surface of a reactant. The latter case leads to a significantly reduced reaction rate for faster heating rates (as was observed experimentally), because the surface of the reactant is covered by an armoring rim of product crystals early in the experiment. Application of the results of this study to natural systems suggests how the kinetics of mineral reactions behave quite differently if there are different nucleation conditions, even if the P-T-X conditions of the rock are the same.

The Gibbs free energy (ΔG 0 (1) ) of the reaction kaolinite ↔ dickite was generated from solubility measurements of natural kaolinite and dickite performed in acid solutions at temperatures ranging from 150° to 300 °C under vapor-saturated conditions. The ΔG 0 (1) values increase from -0.620 ± 0.150 to -0.218 ± 0.210 kcal/mol with increasing tem­perature from 150 to 300 °C. Regression of these data yields a value of -0.90 ± 0.10 kcal/mol for ΔG 0 (1) at 25 °C. The standard Gibbs free energy of formation (ΔG 0 f,298 ) of dickite deduced from ΔG 0 (1) and the ΔG 0 f,298 of kaolinite (Zotov et ah. in preparation) is -908.36 ± 0.40 kcal/mol. The results obtained in this study indicate that kaolinite is metastable rel­ative to dickite at temperatures to at least 350 °C. It follows that the timing of observed kaolinite to dickite transformations in diagenetic and many hydrothermal systems is con­trolled by the kinetics of this reaction rather than thermodynamic equilibria.

Mössbauer spectrometry of natural samples of tourmaline shows the presence of Fe 2+ and Fe 3+ either in the Y (9b) or the Z (18c) sites, with intervalent iron delocalized between these sites. To understand the distribution of iron in natural samples, tourmaline was syn­thesized at temperatures from 400 to 700 °C, under a pressure of 100 MPa, by the trans­formation of an Fe 2+ -rich natural chlorite in the presence of Na 0.5 K 0.5 -feldspar, boric acid, and H 2 O in stoichiometric proportions. The oxygen fugacity, f O₂ , of most experiments was buffered by solid assemblages including hematite (HM), nickel-nickel oxide (NNO), and quartz-fayalite-iron (QFI). It was possible to synthesize tourmaline with Fe 2+ in the Z site and Fe 2+ or Fe 2+ + Fe 3+ in the Y site. These site occupancies are similar to those observed in many natural samples. For each f O₂ buffer, Mossbauer spectrometry shows that Fe 3+ /Fe 2+ in the Y site is cor­related positively with T and f O₂ for T <550 °C. For higher temperatures, a negative cor­relation is observed. These relationships reflect structural changes involving the localiza­tion of Fe 2+ in the Z site and an increasing amount of Al in the Y site. Fe 3 +/total Fe can be represented by general equations of the type [%Fe 3+ ] = e (a log f O₂ + b) . The curves corre­sponding to each temperature are roughly parallel and equidistant. Thus, Fe tourmaline is a potential geothermometer and an f O₂ indicator.

Partially sericitized tourmaline from a pegmatite was investigated using high-resolution transmission electron microscopy (HRTEM). Fine-grained muscovite crystals form extensive narrow veinlets preferentially developed along the {110} and {100} cleavages of tourmaline, indicating that a cleavage-controlled alteration mechanism was dominant. HRTEM images show that tourmaline-muscovite interfaces parallel well-defined {110} and {100} of tourmaline. In general, (001) of muscovite parallels the c axis of tourmaline, but otherwise tourmaline and replacing muscovite lack crystallographically oriented relationships. The muscovite consists of numerous 100 to 1000 Å thick subparallel packets, and the angles between (001) of muscovite and (110) of tourmaline are highly variable. Aluminous minerals other than muscovite were not observed as alteration products of tourmaline, suggesting that the tourmaline reacted directly to form muscovite; the alteration apparently involved residual fluids in which K + was available and silica was not deficient.

High-resolution transmission electron microscopy reveals dense Brazil-twin boundaries on the unit-cell scale in length-slow chalcedony. The twins can be considered as the result of stacking of left- and right-handed quartz with a (101) twin composition plane. Although most twin sequences on the unit-cell scale are nonperiodic, moganite-type domains result where they are periodic. It is proposed that the twins formed during rapid crystallization rather than as transformation products of a precursor phase such as moganite. The twin boundaries are energetically less stable than twin-free areas and may indicate non-equilibrium crystallization at a high supersaturation of aqueous SiO 2 species in the parent fluid.

Analytical transmission electron microscopy (ATEM) was used to quantitatively measure sub-micrometer compositional profiles produced experimentally by Fe-Mg interdiffusion experiments in olivine. Although analysis of minerals by ATEM is common, compositional profiles suitable for quantitative modeling of diffusion are not generally measured in minerals with ATEM. To demonstrate the suitability of ATEM for diffusion studies in minerals, we have investigated Fe-Mg interdiffusion in experimentally annealed olivine. Because the compositional gradients were induced under well-controlled laboratory conditions, the accuracy of the measurements could be tested by comparing compositional profiles measured by both ATEM and EMPA as well as by retrieving diffusion coefficients from both TEM and EMPA data. The agreement in diffusion coefficients shows that point defect equilibration in the interfacial region of the diffusion couple occurs extremely rapidly at 1200 °C. The ability to obtain diffusion data from such short anneals enables various experiments that were not previously possible-for example, to study diffusion rates at high pressures where long f O₂ buffered anneals are not generally feasible. ATEM profile measurement is O 2 compared with other techniques such as SIMS and RBS and some limitations and applications of ATEM profile measurements are also discussed.

X-ray absorption fine structure spectroscopy (XAFS) was used to determine arsenic (As) oxidation state, local coordination (to a radius of ≈7 Å around As), and the relative proportion of different As species in model compounds and three California mine wastes: fully oxidized tailings (Ruth Mine), partially oxidized tailings (Argonaut Mine), and roast­ed sulfide ore (Spenceville Mine). Mineralogy was characterized by Rietveld refinement of X-ray powder diffraction patterns. The spatial distribution of As in the mine wastes (at several micrometers spatial resolution) was determined by electron microprobe analyses. X-ray absorption near edge structure (XANES) analysis indicates that As 5+ is the dominant oxidation state in the mine samples, but mixed oxidation states (nominally As 0 and As 5+ ) were observed in the Argonaut Mine waste. Non-linear, least-squares fits of mine waste EXAFS (Extended XAFS) spectra indicate variable As speciation in each of the three mine wastes: As 5+ in the Ruth Mine sample is sorbed on ferric oxyhydroxides and aluminosil­icates (probably clay) in roughly equal portions. Tailings from the Argonaut Mine contain ≈20% As bound in arsenopyrite (FeAsS) and arsenical pyrite (FeS 2-x Asx) and ≈80% As 5+ in a precipitate such as scorodite (FeAsO 4· 2H 2 0); however, no precipitate was detected by X-ray diffraction or microprobe analysis, suggesting that the phase is poorly crystalline or has low abundance (total As in sample = 262 ppm). Roasted sulfide ore of the Spenceville Mine contains As 5+ substituted for sulfate in jarosite [KFe 3 (SO 4 ) 2 (OH) 6 ] or incorporated in the structure of an unidentified Ca - or K-bearing phase, and As 5+ sorbed to the surfaces of hematite or ferric oxyhydroxide grains. Detemiination of solid-phase As speciation in mine wastes by XAFS spectroscopy is a valuable first step in the evaluation of its bio­availability, because the mobility and toxicity of As compounds vary with oxidation state. As bound in precipitates, as in the Argonaut mine sample, is considered to be less available for uptake by organisms than when sorbed on mineral surfaces or coprecipitated with poorly crystalline phases, as found for the Ruth and Spenceville mine wastes.

Polarized infrared absorption spectra of thin, oriented single-crystal slabs of pectolite and serandite were recorded between 4000 and 350 cm -1 at 298 and 83 K. The spectra of both minerals show a broad absorption region parallel to the silicate chains (b direction) that is centered around 1000 cm -1 which is interrupted by a transmission window, and which is superimposed by sharp silicate, lattice, and overtone modes. This band is assigned to the OH stretching mode consistent with the alignment of the O-H O hydrogen bond parallel to b and the short O···O distance of 2.45-2.48 Å that was found in previous X- ray structure refinements. At 1396 cm -1 (pectolite) and 1386 cm -1 (serandite) an OH bending mode is observed in the IR spectra parallel to c. At low temperatures, this mode shifts up to higher frequencies (1403 cm -1 at 83 K in pectolite). whereas the down-shift of the OH stretching mode cannot be observed due to the extremely broad band width. The slightly higher energy of the bending mode in pectolite indicates a slightly stronger hydrogen bond with respect to serandite. However, the bond length in serandite is slightly shorter than that in pectolite. An asymmetric O-H ··· O bond is confirmed in pectolite and serandite through comparison with different materials with similar, very strong hydrogen bonds and low-energy OH stretching modes.

Many known occurrences of the zeolites erionite and offretite have been characterized by electron probe microanalysis, X-ray powder diffraction, and optical microscopy. For the first time, a substantial amount of experimentally consistent and homogeneous chemical and crystallographic data have been evaluated for these natural zeolites. Systematic analysis of the data, performed by statistical multivariate analysis, leads to the following conclusions: (1) the two zeolites have well-defined compositional fields in the chemical space describing the extraframework cation content, best illustrated in a Mg-Ca(1Na)- K(+Sr+Ba) diagram; (2) no discrimination is possible on the basis of the framework Si/Al ratio because of the extensive compositional overlap between the two species, however the Si-Al content in the framework tetrahedra is the major control on the unit-cell volume dimensions, particularly in erionite; (3) the crystal chemistry of the Mg cations is a major factor in controlling the crystallization of the mineral species; (4) cation compositions at the boundary of the recognized compositional fields might be due to chemical averaging of two-phase intergrowths, although these mixed-phase occurrences are much less common than previously thought; (5) the sign of optical elongation is not a distinctive character of the two phases, it is related to the Si/Al ratio in the framework tetrahedra of each zeolite type and cannot be used for identification purposes; (6) the zeolite mineral species epitaxially overgrown on levyne in all cases is identified as erionite; in a few cases offretite was found to be overgrown on chabazite; (7) erionite samples epitaxially overgrown on levyne are substantially more Al-rich and Mg-poor than the erionite samples associated with other zeolites.

This study clarifies the crystal structure variations and relationships in the zeolites erionite and offretite. The crystal structure analyses used Rietveld analysis of X-ray powder diffraction data, obtained both by synchrotron radiation and conventional X-ray sources, and on diffraction patterns obtained by transmission electron microscopy. The framework Al atoms in erionite are preferentially located in the single six-membered ring of tetrahedra (T2 site), whereas the Si-Al distribution is essentially disordered on the tetrahedral framework sites in offretite. In both zeolites, the center of the cancrinite cage is always occupied by K cations in similar amounts in both minerals. The erionite cage in erionite is occupied by Ca and Na atoms distributed on four distinct cation sites. Mg ions can be present up to 0.8 atoms per cell and are located between the Ca1 and Ca3 sites and very close to the Ca2 site on the symmetry axis. In offretite, the Mg site is located on the trigonal axis of the gmelinite cage and the Mg atoms are bonded to a variable number (five or six) of H 2 O molecules, depending on the site population. The structural data and TEM analysis clearly show that the crystal chemistry of the extraframework cations, Mg in particular, is a major factor controlling whether erionite or offretite crystallizes, the Si-Al distribution in the framework, and the possible stacking intergrowths of the two minerals.

The new mineral tschortnerite, ideally Ca 4 (K,Ca,Sr,Ba) 3 Cu 3 (0H) 8 [Si 12 Al 12 0 4S ]·ϰH 2 0, ϰ > 20, occurs as well-formed cubes up to a maximum size of 0.15 mm in a Ca-rich xenolith at the Bellberg volcano near Mayen, Eifel, Germany. The light blue, transparent crystals are optically isotropic, n = 1.504(2). Microprobe analysis (in weight percent) gave CaO 13.10, CuO 9.64, SrO 4.49, BaO 1.93, K 2 O 1.37, Fe 2 O 3 0.30, Al 2 O 3 25.21, SiO 2 30.25, H 2 O (calc, by difference) 13.71. The empirical formula based on 48 O atoms within the tetrahedral net is Ca 5.60 Sr 1.04 K 0.70 Ba 0.30 Cu 2.90 Fe 0.09 Al 11.85 Si 12.06 O 48 (OH) 8.44· 14.01H 2 O. Tschörtnerite is cubic, space group Fm3̅m [a = 31.62(1) Å, V = 31614 Å 3 , Z = 16]. The density is -D meas = 2.1 g/cm 3 , D calc = 2.10 g/cm 3 . Single-crystal X-ray investigations showed that tschortnerite is a zeolite; the structure contains interconnection of double six-rings, double eight-rings, sodalite cages, truncated cubo-octahedra, and previously unknown 96-membered cages (tschörtnerite cage). A new structural unit is the [Cu 12 (OH) 2 JCa 8 O 24 (H 2 O) 8 cluster centered within the truncated cubo-octahedron. The cluster is formed by a rhomb- dodecahedron-like arrangement of comer connected CuO 4 squares, the eight CaO 7 poly- hedra are branched. The sodalite cage houses Ca 4 (OH) 4 O 12 clusters of edge-sharing CaO 6 octahedra. Half-occupied (K,Ca,Sr,Ba) positions were located in the basal and top face of the double eight-rings, i.e., the border to the tschortnerite cage. Within the large tschort­nerite cage only H 2 O molecules were localized.

Crystals of sodium heptasilicate (Na 8 Si 7 O 18 ) have been grown at 9 GPa, 1000 0 C using the MA6/8 superpress at Edmonton. The X-ray structure was determined at room pressure (R = 5.8%). Sodium heptasilicate is trigonal with a = 7.180(1) Å. α = 87.26(1)°, space group Rbu671, Z= I, and D x = 3.009 g/cm 3 . The structure contains isolated six-membered UDUDUD rings of SiO 4 tetrahedra interconnected by SiO 6 octahedra with [6] Si: [4] Si =1:6, giving a structural formula of Na 8 Si[Si 6 0 18 ]. The mixed [4] Si, [6] Si framework is undersatu­rated because only one of the non-bridging O atoms in the SiO 4 tetrahedron is shared with the SiO 6 octahedron. Six Na cations per formula unit (pfu) are in eightfold coordination with O, and two are in a distorted and compressed octahedral coordination. Sodium hep­tasilicate belongs to the homologous series Na 2k Na 2(m-k) Si m-k [Si n-m+k O 2n+m ]. with k < m < ti and (n - m + k) ≥ (3/2)(m - k). The X-ray structure is consistent with microprobe analysis data and revises the earlier characterization of this phase as (ξ-Na 2 Si 2 O 5 . The sodium heptasilicate structure doesjiot fiilly_relax with decompression and undergoes incip­ient displacive transformation (R3̅m -> R3̅) with minor positional disorder of Na cations.

Pretulite is a new Sc phosphate with zircon-type structure from the phyllite-micaschist hosted hydrothermal lazulite-quartz veins in the Lower Austroalpine Grobgneis complex, eastern Austria. The new species is the Sc-dominant analogue of xenotime-(Y) and occurs as an accessory mineral in all investigated lazulite specimens, forming anhedral to euhedral crystals up to 200 μm long with the dominant form {211}. It is associated with lazulite, fluorapatite, chlorapatit, quartz, muscovite, clinochlore, paragonite, kyanite, pyrophyllite, augelite, wardite, hydroxylherderite, goyazite, florencite-(Ce), xenotime-(Y), bearthite, m- tile, pyrite, corundum, and an AlO(OH)-phase. Premlite is translucent to transparent with an adamantine luster, colorless to pale pink, uniaxial positive with ω = 1.790 (5), € = 1.86 (I) and shows a weak orange fluorescence at 254 nm as well as a bright blue cathodoluminescence. Mohs hardness is about 5. The empirical formula for pretulite (based on four O atoms) is (Sc 0.98 Y 0.02 ) 1.00 P 1.00 O 4.00 . It contains variable amounts of Y with Y/(Y + Sc) = 0.5 - 3.2 mol% and traces of Yb, Er, and Dy. The space group is I4 1 /amd with a = 6.589 (1) Å, c = 5.806 (1) Å, V = 252.1 (1) Å 3 , d calc = 3.71 g/cm 3 , Z = 4. The four strongest lines in the X-ray powder pattern are d 200 = 3.293 (100), d 112 = 2.4636 (42), d 112 = 1.6927 (45), d 332 = 1.3697 (15) Å. The crystal structure was refined using 108 reflections to R(F o ) = 0.018. Cell parameters and average M-O distances confirm a small amount of Y sub­stituting Sc in the eightfold-coordinated M-position. The formation of pretulite is attributed to a moderate enrichment of Sc in lazulite-rich domains of the veins (about 180 ppm) and the inability of the accompanying minerals to incorporate larger quantities of Sc in their crystal structure. The name is after the mountain Pretul, Fischbacher Alpen, Styria, Austria.

The crystal structure of a synthetic NaAlSiO4 modification has been solved and refined to an R index of 0.020 for 2745 independent reflections. The compound is hexagonal with space group symmetry P6 1 , a = 9.995(2) Å and c = 24.797(4) Å. The crystal showed twinning by merohedry according to m 210 , which was accounted for in the calculations. The phase was named trinepheline, following prior studies, because the length of its c lattice parameter is three times the length of the c parameter in nepheline, whereas the a parameter is about 10 Å in both phases. The crystal structure is characterized by layers of six-membered tetrahedral rings of exclusively oval conformation. The rings are built up by regularly alternating AlO 4 and SiO 4 tetrahedra. The stacking of the layers parallel to the c axis results in a three-dimensional network containing channels that are occupied by the Na cations. Although structural similarities with respect to tridymite derivatives can be found, hexagonal trinepheline represents a new type of stuffed tridymite that is not a simple superstructure of nepheline.

Boralsilite, the first natural anhydrous Al-B-silicate, is a high-temperature phase in peg­matites cutting granulite-facies metapelitic rocks at Larsemami Hills, Prydz Bay, east Ant­arctica (type locality) and Almgjotheii in the contact aureole of the Rogaland Intrusive Complex, southwestern Norway. Stable assemblages include: (1) quartz-potassium feld- spar-boralsilite-schorl/dravite (Larsemaim Hills); (2) potassium feldspar-plagioclasefAn 22 )- boralsilite-werdingite-dunioitierite-grandidierite (Almgjotheii); (3) quartz-potassium feld- spar-boralsilite-dumortierite-andalusite ± sillimanite (Ahngjotheii). Boralsilite is estimated to have formed between 600 and 750 °C and 3-5 kbar at conditions where P H₂O < P tot . The name is from the composition, boron, aluminum, and silicon. Representative electron and ion microprobe (SIMS) analyses of Larsemaim Hills are: SiO 2 10.05 [12.67]; Al 2 O 3 71.23 [69.15]; FeO 0.48 [1.10]; MgO below detection [0.23]; BeO 0.004 [0.094]; B 2 O 3 19.63 [18.11] wt%, totals 101.39 [101.35] wt% where the numbers in brackets were de­termined from Almgjotheii material. However, the SIMS B 2 O 3 values appear to be system­atically too high; boron contents calculated assuming B + Si = 8 and O = 37 atoms per formula unit (apfu) yield B 2 O 3 18.53 wt% corresponding to Fe 0.08 Al 15.98 B 6.09 Si 1.91 O 37 ideally Al 16 B 6 Si 2 O 37 for Larsemaim Hills. The analogous composition of Mg 0.07 Fe 0.18 - Al 15. 66 Be 0.04 B 5.565 Si 2.435 0 37 for Almgjotheii appears to result from solid solution of boralsilite with sillimanite (or Al 8 B 2 Si 2 O 19 ) and subordinate werdingite. Boralsilite forms prisms up to 2 nun long || b> and 0.25 mm across and is commonly euhedral in cross section. It is colorless and prismatic cleavage is fair. Optically, it is biaxial (+); at λ = 589 nm, the Larsemaim Hills material has α = 1.629(1), β = 1.640(1), γ= 1.654(1), 2V meas = 81.8 (6), r > ν extremely weak, and γ||b. It is monoclinic, space group C2/m with lattice parameters for Larsemaim Hills, of a = 14.767(1), b= 5.574(1), c = 15.079 (1) Å, β= 91.96(1)°.F = 1240.4 (2) Å 3 , Z= 2. and P calo = 3.07 g/cm 3 .

Single crystals of MnSiO(SiO 4 ) with the titanite structure together with MnSiO 3 clinopyroxene were synthesized from a MnO-SiO 2 oxide mixture at 1000 8C and 9.2 GPa in a multi-anvil press. The crystal structure of MnSi 2 O 5 [space group C2/c, a = 6.332(1) Å , b = 8.161(1) Å , c = 6.583(1) Å , β = 114.459(3)8, and V = 309.66 Å 3 ] was refined at room temperature from single-crystal X-ray data to R1 = 2.23%. The monoclinic MnSi 2 O 5 phase has the titanite aristotype structure and is similar to the monoclinic Ca-analogue CaSi 2 O 5 . Si occurs in compressed octahedral coordination, replacing Ti in titanite, and in tetrahedral coordination as an orthosilicate group. Mn has a distorted sevenfold coordination with Mn-O distances between 2.086 and 2.365 Å .

23 Na and 133 Cs NMR T 1 relaxation data for wet paste samples made from illite and either NaCl or CsCl solutions show that there is significant adsorption of the Na and Cs onto the surface. The Na and Cs are in the NMR rapid exchange regime (exchange frequencies > approximately 10 4 Hz), allowing the T 1 values to provide useful information about surface occupancies. For 133 Cs, the data can be modeled to yield values for the fraction of Cs in the sample adsorbed onto surface sites and for the density of Cs surface occupancy. The maximum surface density observed for 0.1 M CsCl solutions is approximately 0.035 atoms/Å 2 , essentially identical to the density of inner-sphere Cs sites determined from magic-angle-spinning nuclear magnetic resonance (MAS NMR) spectra of dried samples observed at a relative humidity of approximately 35%. The similarity of these results suggests that the relatively simple compositional dependence of the 133 Cs T 1 data is due to most of the surface Cs being in inner-sphere complexes with a relatively constant average T 1 when bulk H 2 O is present. In contrast, the 23 Na T 1 data cannot be modeled in this way, consistent with the idea that its sorption on illite is dominated by outer-sphere complexes (Stern and/or Gouy layers) with a range of T 1 values. The similarity of the 23 Na T 1 behaviors of pastes made from NaCl solutions and either illite or silica gel (which has no permanent negative charge due to isomorphic substitution) is also consistent with this idea.