Volume 81, Issue 1-2

The {1̅014} surfaces of optical quality calcite (Iceland spar) were examined in air using scanning force microscopy (SFM) immediately after cleavage and during the hours and days that followed. In the absence of a visible contacting solution phase, spontaneous rearrangement of the surface was observed. Images with nanometer-scale resolution showed the formation of hillocks and holes on terraces and cleavage Steps; thickness or depth varied from one to ten calcite monolayers (3-30 Å). The rate of surface change varied with geochemical System Parameters such as humidity and partial pressure of N 2 and CO 2 , but instrument Parameters such as imaging force, tip composition, scanning rate, and reimaging frequency had almost no effect. On the basis of previous work documenting the existence of surface-hydration species and a layer of molecular water on samples exposed only to air, we interpret that the observed process results from dissolution and reprecipitation within an invisible layer of water that is adsorbed from air following cleavage. The dynamic nature of calcite surfaces has important implications in conceptual models for the behavior of adsorbed trace metals in unsaturated porous media or arid climates and also for attack on “dry” statues and building stones under acidic atmospheres.

Phase transitions similar to those observed at low temperatures in lawsonite, CaAl 2 [Si 2 O 7 ] - (OH) 2 · H 2 O, were recorded at ~ 95 and ~ 150°C in hennomartinite, SrMn 2 [Si 2 O 7 ](OH) 2 · H 2 O. The room temperature structure of the natural phase was refined from X-ray singlecrystal data in space group P2 1 cn and that of the > 150 °C phase at 245 °C in space group Cmcm. The structures are isotypic to the low-temperature and room temperature structures of lawsonite. The P2 1 cn structure of hennomartinite exhibits twinning parallel to (100) with additional disorder mainly involving the H positions. As indicated by a split O position (O5), disorder and twinning must also be considered in the Cmcm structure. Similar to lawsonite, the driving force of the phase transitions is apparently the development of cooperative hydrogen bonds. At low temperatures the H atoms shift toward neighboring O atoms forming strong hydrogen bonds. The transition temperatures were determined by monitoring the intensities of critical reflections over the temperature range from -163 to 245°C. The low-temperature boundary of the Cmcm phase is characterized by the appearance of the 017 reflection (among other C-centering forbidden reflections) below 150(5) °C. The high-temperature boundary of the twinned room temperature phase is characterized by a dramatic enhancement of the intensity of the 406 reflection below 95(5) °C. Between 95 and 150°C (in analogy to lawsonite) the average space group Pmcn is proposed, which is the mediating sub-super group between P2 1 cn and Cmcm. After one heating and cooling cycle, splitting of hk0 reflections (observed in ω scans) below ~ 100°C indicates that the true, relaxed room temperature structure represents a monoclinic derivative of the natural, initially determined P2 1 cn average structure. Moreover, a decrease of the a unit-cell parameter and the unit-cell volume during the 245°C data collection suggests a temperature-controlled structural evolution of hennomartinite from disorder and small twin domains in the natural phase to large twin domains caused by annealing at elevated temperatures.

Coherent inelastic neutron scattering measurements have been performed on a natural garnet pyrope from the Dora Maira Massif, Italian Western Alps, [(Mg 0.92 Fe 0.05 Ca 0.03 ) 3 Al 2 Si 3 O 12 , space group Ia3̅d, Z = 8, a = 11.459 Å] to determine the acoustic phonon-dispersion relations at room temperature. Phonon measurements have not previously been made on natural garnets, although the dispersion curves form a basis for developing and testing physically correct interatomic potentials for use in lattice dynamical models. Acoustic phononspropagatingin the high-symmetry[100] and [110] directions were observed using the PRISMA spectrometer installed at the ISIS Neutron Spallation Source, U.K. The inelastic scans were performed around the (660), (080), (860), (10,6,0), and (0,10,0) Bragg points: Two transverse acoustic and one longitudinal acoustic phonon branches were measured. The observed acoustic phonon frequencies are in good agreement with those calculated on the basis of a semiempirical interatomic potential model. As a check, the elastic constants obtained from the slopes of the acoustic phonon branches are comparable with previous measurements obtained by Brillouin spectroscopy, although their large estimated errors reflect the complexity of the experimental measurements of phonon-dispersion curves on pyrope garnets.

Kaolinite-rich samples were selected for 27 Al and 29 Si MAS NMR study to explore the effect of Fe on their spectra and provide insight into the nature of Fe ordering in kaolinite. Initial characterization by Chemical analysis, TEM, X-ray diffraction, and magnetic mass susceptibility (x*) measurement was conducted to obtain five samples of high purity and with a range of Fe content. Secondary iron oxide and hydroxide phases were extracted using an HCl treatment. TEM study of the samples before and after treatment revealed the effective removal of secondary Fe phases and pristine kaolinite crystal habits. MAS NMR experiments included measurement of 29 Si spin-lattice relaxation times (T 1 ) at 6.36 T and quantitative 27 Al measurements at 8.46 T. The 29 Si T 1 studies show a general increase in T 1 with decreasing Fe content. It is probable that the relaxation mechanism occurs dominantly through the dipole-dipole interaction with unpaired electron spins in Fe Centers of the dioetahedral sheet. Data fitting indicates that spin relaxations are not best deseribed by a single T 1 . Improved fitting using double exponential or power-law behavior to describe 29 Si spin-relaxation phenomena indicates a heterogeneous distribution of Fe centers. The relatively longer T 1 , values for two samples suggest their dioctahedral Fe domains may have a more clustered distribution within the kaolinite structure. Results from 27 Al studies indicate very good eorrelation between Fe content, x g , and spinning sideband (SSB) intensity. There is very little eorrelation between total integrated 27 Al intensity and Fe content. When 27 Al NMR intensity variations are compared with estimates of 27 Al NMR signal loss predicted by a paramagnetic line-broadening wipeoutsphere model that uses an idealized kaolinite structure with regularly ordered dioetahedral Fe sites, the two samples appear to have a more clustered distribution of Fe. The relative increase in SSB intensity is consistent with an increase in the number of antiferromagnetic or ferrimagnetic domains.

A Monte Carlo method was developed to determine the distribution of Al and Si cations from 29 Si MAS NMR data for framework aluminosilicates. The method was used to analyze data obtained from Cs-exchanged leucite in its high-temperature cubic phase and from cubic analcime samples. It was also used to determine the number of Al-O-Al linkages in cordierite for different degrees of Al-Si order. The results for cordierite were correlated with the calorimetric data of Carpenter et al. (1983) to estimate the energy for the reaction (Al-O-Al) + (Si-O-Si) → 2(Al-O-Si) of ~40 kJ/mol.

High-pressure X-ray powder diffraction data for MgSiO 3 ilmenite have been collected up to 28 GPa using a diamond-anvil cell in an energy-dispersive configuration at the storage ring DCI at the LURE (Orsay, France). H 2 O was used as a pressure-transmitting medium. Pressures were measured using the ruby-fluorescence method or estimated from the equation of state of ice VII. Both pressure determinations are in good agreement up to 15 GPa but diverge at higher pressures because of the development of nonhydrostatic stresses and pressure gradients marked by an increase of the ruby-fluorescence line width. No phase changes are observed during compression or decompression. The compression of MgSiO 3 ilmenite is anisotropic, with the c axis being twice as compressible as the a axis. The high-pressure volume data were fitted to a third-order Birch-Mumaghan equa­tion of state and are consistent with the former determination of K 0 = 212 GPa by Brillouin scattering; the data yield Kʹ 0 = 7.5(10) if the ruby-fluorescence pressures are used or Kʹ 0 = 5.6(10) if the pressures obtained from the equation of state of ice VII are used; the latter value is preferred because pressures were determined on the X-ray spot. The effects of nonhydrostatic stresses and pressure gradients on the determination of equations of state are discussed.

Pressure-volume X-ray diffraction data were collected at eight pressures to 17.2 GPa for synthetic single-crystal forsterite, Mg 2 SiO 4 . Birch-Mumaghan equation-of-state param­eters were determined to be K 0 = 125(2) GPa and Kʹ 0 = 4.0(4). To ensure hydrostatic conditions, we monitored the diffraction-peak shapes while the sample was in different pressure med.ia: 4:1 methanol-ethanol, He, and Ar. The diffraction peaks for forsterite in 4:1 methanol-ethanol at 12.2 GPa were broadened to such an extern that we believe that any data collected at similar pressures and in this medium may be suspect and should be examined critically. The diffraction-peak profiles for the sample in He were Sharp and well-formed, suggesting He provides an excellent medium at pressures ≤ 17.2 GPa. However, upon decompression, the diffraction peaks of the sample in He also showed significant broadening. This broadening suggests that with increasing pressure He diffused into the crystal, but with the release of pressure the He was unable to diffuse out quickly. The diffraction peaks regained their sharpness after several months.

Ti-rich potassium-richterite samples were synthesized at 900 °C and 1 kbar P H₂O and characterized by Fourier-transform infrared spectroscopy (FTIR) and Rietveld structure refinement. The infrared spectrum of end-member Ti-free potassium-richterite shows a main band at 3734 cm -1 assigned to an OH anion bonded to three octahedral Mg cations and directed toward K at the A site facing a tetrahedral ring of composition Si 6 . When Ti is introduced into the structure, two new overlapping components appear in the infrared spectrum. Using a Symmetrie Gaussian model, they can be resolved into two single bands at 3711 and 3698 cm -1 . These two bands are assigned to OH groups facing rings of tetrahedra with compositions Si 5 Ti and Si 4 Ti 2 , respectively. Their intensities are linearly related to the Rietveld-refined Ti content at T2 up to 0.8 apfu (atoms per formula unit). The analysis of relative band intensities indicates short-range disorder of Si and Ti over the T2 sites.

Synthetic Co 3 Al 2 (SiO 4 ) 3 garnet and (Mg,Ni) 3 Al 2 (SiO 4 ) 3 garnet with Ni/(Ni + Mg) = 0.15 - 0.18 have been studied by single-crystal X-ray diffraction and optical spectroscopy. The X-ray data confirm that Co 2+ and Ni 2+ occupy the large distorted cubic (dodecahedral) site of the structure. Structure refinement indicates that the Ni 2+ ion is probably displaced out of the center of this site and statically or dynamically disordered onto the 48g equipoint at (1/8,y, 1/4+y). This is consistent with the lattice constant of the Ni-bearing garnet of 11.4717(7) Å being significantly larger than the lattice constant of pyrope (11.459 Å), although the ionic radius of Ni 2+ is usually smaller than that of Mg 2+ . The lattice constant of Co 3 Al 2 (SiO 4 ) 3 garnet is 11.4597(2) Å. Because of the unusual coordination geometries of Ni 2+ and Co 2+ , the optical spectra of the garnets are fundamentally different from those of other Co 2+ - and Ni 2+ -bearing silicates. The following crystal-field Parameters were estimated from spectroscopic data: crystal-field Splitting Δ = 4430 cm -1 (Co 2+ ) and 4210 cm -1 (Ni 2+ ); Racah parameter B = 890 cm -1 (Co 2+ ) and 1080 cm“ 1 (Ni 2+ ). These data yield an exceptionally low crystal-field stabilization energy (CFSE) of Ni 2+ in garnet of 3370 cm -1 , which explains the very low Ni contents of garnets in equilibrium with olivine. The CFSE of Co 2+ in garnet of 5320 cm -1 is similar to values for pyroxenes and olivine. Therefore, the partition coefficient of Co 2+ between these phases should be close to unity, as is observed.

To explore the effect of pressure on the stacking-sequence of the chlorite cookeite, a systematic TEM study of stacking microstructures has been conducted on natural samples from pegmatitic and metamorphic environments representing a large pressure field over a relatively narrow temperature range. The low-P and low-T conditions (1-5 kbar, 280 - 350 °C) are represented by late-crystallized cookeite from pegmatites and by cookeite from low-temperature metapelites. For the high-P conditions (5-16 kbar, 280-430 °C), cookeite samples from metapelites and metabauxites were selected. Combining selected-area electron diffraction (SAED) and high-resolution transmission electron microscopy (HRTEM) imaging shows that the polytypes evolve from a fully semirandom type to fully ordered types under increasing total pressure. For both polytypes the repeat distance (from one to five layers) of the basic structures and the diversity of repeats increase with pressure. Above 10 kbar, long-period polytypes (LPP), up to at least a 14-layer repeat, coexist with perfectly ordered basic structures. Neither the unique type of intralayer stacking (Ia) nor the single-layer subcell volume measurements improve understanding of the role of pressure on the ordering process. Clearly, the regular polytypes of cookeite behave as polymorphs that are remarkably sensitive to total pressure.

Serpentinite from Lancaster County, Pennsylvania, consists of a variety of fine-grained serpentine minerals, chlorite, randomly interstratified serpentine-chlorite, and a series of phases based on regular interstratification of Serpentine and chlorite (S x C y , where x and y are integers). Within the resolution of the AEM technique, all layer silicates have the same Mg-rich, Al-rich, Cr-rich, and Fe-poor compositions. Regularly interstratified serpentine-chlorite minerals are frequently intimately intergrown with serpentines that have repeat distances identical to those of the regular interstratifications. Thus, dozyite (S 1 C 1 , β = 90°) is intimately associated with serpentine with three-layer octahedral order (I,I,II). Longer period polysomes (S 2 C 1 , S 1 C 2 , S 2 C 2 , S 1 C 3 , S 3 C 2 , and S 1 C 4) all with β = 90°) are each accompanied by serpentines with equivalent c-axis periodicities. S x C y phases apparently form by selective growth of lbb chlorite units from I,II octahedral sequences in long-period serpentines. All microscopic structural evidence is consistent with the formation of regular interstratifications by tetrahedral inversion within existing serpentine. Atomic resolution images reveal that the tetrahedral sheet is displaced by a/3 where it inverts to form the 2:1 layer. A ±a/3 shift is required for hydrogen bonding between OH of the newly formed brucite-like interlayer and O atoms of the 2:1 layer. The sense of the shift is determined by the strong interactions between the octahedral cations in the brucite-like interlayer and the Si in the 2:1 layer (direct superimposition, previously described as a type-a interaction, is strongly unfavorable). Distortion at the inversion point probably lengthens Si-O bonds in the next tetrahedra, facilitating relocation of Si on the other side of the basal O plane. Reversal of the octahedral slant in the 2:1 layer occurs because the +a/3 tetrahedral shift necessitates repositioning of O and OH coordinating octahedral cations, requiring movement of octahedral cations from type-II to type-I positions. Except in the 2:1 layers, the stacking and octahedral slants are inherited. The result is a series of regular interstratifications characterized by a single octahedral slant (specifically, Ibbb,I) and b/3 stacking disorder. This analysis reveals the importance of cation-cation interactions in determining the relative stability of pairs of 1:1 layers and in controlling the detailed structures of layer silicates formed in solid-state serpentine-to-chlorite reactions. Because similar constraints apply to formation of Serpentine from chlorite by direct structural modification, the common 1T lizardite polytype may be produced from both IIbb and Ibb chlorites.

The phonon spectra of IR-active vibrations in alkali feldspars have been systematically investigated in the spectral range 50-2000 cm -1 at temperatures between 20 and 900 K. Samples with high degrees of Al-Si Order display IR absorption line profiles with spectral line widths that are systematically smaller than those for samples with low degrees of Al-Si order. Phonon frequencies depend in a nonlinear manner on the Na-K content, indicating nonideal mixing properties. The mode Grüneisen parameters are largest for low-frequency lattice modes and Si-O stretching modes; some negative Grüneisen parameters were found at phonon energies below 430 cm -1 . The temperature evolution of the phonon spectra shows clear indications of the C2/m-C1̅ transition in Al-Si disordered Na-rich samples. An anomaly for highly Al-Si ordered microcline at 270 K is confirmed.

Nanpingite, ideally CsAl 2 (AlSi 3 )O 10 (OH) 2 , is a Cs end-member dioctahedral mica, crystallizing in C2/c with a = 9.076(3), b = 5.226(2), c - 21.41(5) Å, and β = 99.48(6)°. The crystal structure of type nanpingite was refined to R = 0.058 using single-crystal X-ray diffraction methods. Nanpingite is a 2M 2 polytype dioctahedral mica rather than a 2M 1 polytype as reported originally. In comparison with the interlayer cations in common muscovite and paragonite, Cs + in nanpingite increases the [001] interlayer Separation between adjacent 2:1 layers but has little effect on the lateral (a, b) cell dimensions. The existence of the rare 2M 2 polytype is attributed to this large interlayer Separation, which minimizes the repulsion of the super-imposed (along [001]) O anions in the basal planes of neighboring tetrahedral layers. The 2M 2 polytype involves ditrigonal prismatic coordination of the interlayer Cs + , in contrast to the octahedral coordination of interlayer K + and Na + in muscovite-2M 1 and paragonite-2M 1 . The tetrahedral rotation in nanpingite is the smallest among dioctahedral micas and is also caused by the incorporation of the large interlayer cation Cs + .

Si K- and L-edge X-ray absorption near-edge structure (XANES) of SiO 2 -P 2 O 5 and Na 2 O-SiO 2 -P 2 O 5 glasses containing P 2 O 5 above 30 mol% were investigated using synchrotron radiation. Both Si K- and L-edge spectra indicate that Si remains fourfold coordinated ( [4] Si) with O in these glasses until the content of P 2 O 5 reaches about 32 mol%, at which sixfold coordinated Si ( [6] Si) first appears. The Proportion of [6] Si increases qualitatively with increase in the content of P 2 O 5 . However, several P 2 O 5 -rich glasses contain [4] Si only, possibly pointing to a dependence of [6] Si content on quench rate. These results are consistent with 29 Si MAS NMR spectra for silicate-phosphate glasses of similar composition. To estimate further the relative proportions of [4] Si and [6] Si in these glasses using Si K-edge spectra, model composite materials of a-SiO 2 , containing [4] Si only, and c-SiP 2 O 7 , containing [6] Si only, were used to establish the correlation of area ratio for [6] Si and [4] Si edge features with bulk composition. The regression equation may be used for semiquantitative estimation of relative proportions of [6] Si and [4] Si in glasses and other materials of unknown structure with compositions similar to those of the present glass systems.

We studied pyrite spherules built of radiating crystals to establish the relationships among microstructure, composition, and macroseopic appearance. Results of high-resolution transmission electron microscopy (HRTEM) imaging and selected-area electron diffraction (SAED) show that many pyrite crystals contain planar faults perpendicular to one of the [001] axes. A comparison between HRTEM micrographs and images simulated for defect model structures indicates that the faults can be interpreted as single (101) layers of marcasite that disrupt the regular sequence of (002) layers in pyrite. Such a marcasite layer can be described as a boundary between two pyrite crystals that are related by a 2 1 screw axis parallel to [100]. Energy-dispersive X-ray spectrometry (EDS) indicates that such disordered pyrite crystals contain about 3 at% As. However, the distribution of As is uniform over heavily faulted and fault-free regions, indicating the As content is not related to the occurrence of marcasite lamellae in pyrite.

Newly measured viscosities of a low H 2 O-content granitic melt and a low H 2 O-content albitic melt have been combined with viscosity data from the literature to create two new models for the calculation of granitic melt viscosities at crustal pressures and temperatures between 700 and 900 °C; one model is purely empirical, and the other is based on config­urational entropy theory. When the molecular weight of the anhydrous melt is calculated on the basis of eight O atoms, or assumed to be 260 gram formula weight (gfw), and the total H 2 O concentration is expressed by its mole fraction, the viscosity, η, of per- and metaluminous granitic melts, ~ 70-76 wt% SiO 2 , can be calculated from the following empirical equations: For X tot H₂O ≤ 0.25, log η = (0.0292649T - 53.198903)X tot H₂O - 0.0129207T + 24.2973 and for X tot H₂O > 0.25, log η = - 2.2977826X tot H₂O - 0.0095110T + 15.6293 where viscosity is in pascal-seconds, and temperature, T, is in kelvins. The first equation expresses the effects of H 2 O addition on melt viscosity at low H 2 O contents where the predominant H 2 O species in the melt is OH - (as demonstrated in previous studies) and small additions of H 2 O have large effects on melt viscosity; above X tot H₂O ≈ 0.25, molecular H 2 O is the dominant species of H 2 O, the addition of which has only modest effects on melt viscosities. This empirical model is applicable to granitic melts with total H 2 O con­tents between 0.3 and 12.3 wt%. Following configurational entropy theory, viscosities of the same hydrous granitic melts can be calculated by modeling the viscosities as ternary mixtures of various mole fractions of molecular H 2 O, X H₂O , OH - , X OH⁻ , and silicate melt, X silicate melt : ln η = -15.398 + [181.622/{T[S conf ,(X silicate melt ) - R (C H₂O X H₂O ln H₂O + C OH⁻ X OH⁻ ln X OH⁻ + C silicate melt X silicate melt ln X silicate melt )]}], where S conf , the configurational en­tropy calculated from the viscosity measurements, is 4.567 × 10 -3 J/(mol · K); R is the gas constant, 8.314 J/(mol · K); C H₂O is -1.899 × 10 -3 ; C OH⁻ is -1.531 × 10 -3 ; and C silicate melt , is 4.913 × 10 -3 . Mole fractions of H 2 O and OH - are estimated from previously published measurements of species abundances in quenched rhyolitic glasses, and the molecular weight of the silicate melt is 260 gfw. Application of this model is limited to melts with a maximum of approximately 6 wt% total H 2 O because X H₂O -X OH⁻ speciation information is not available at higher H 2 O contents. Numerical values in the above equation are not considered to have any physical or chemical significance. Calculated granitic melt viscos­ities deviate on average from measured viscosities by -0.03 log units for the empirical model and -0.2 log units for the configurational entropy model. At H 2 O concentrations below 4 wt% and above 7 wt%, these models for viscosity estimation arc significantly better than previous models, which can be in error by one to two orders of magnitude. Using either of these new models, the viscosities of granitic melts with 1-3 wt% H 2 O are calculated to be up to two orders of magnitude lower than previously thought. This results in a one to two order-of-magnitude increase in the viscosities of melt extraction from source regions, transport of melts through the crust, and Stokes settling of crystals.

The Taylor Creek Rhyolite, a group of coeval, mid-Tertiary, silica-rich rhyolite lava domes in southwestern New Mexico, is notable for recording bulk-rock evidence of minor, yet easily measurable, contamination of its source magma reservoir resulting from assimilation of Proterozoic roof rock. Most of the evidence is recorded in trace element concentrations and 87 Sr/ 86 Sr i ratios, which are far different in uncontaminated magma and roof rocks. Hornblende phenocrysts and biotite xenocrysts also record the effects of contamination. Electron microprobe analyses show that all hornblende grains are zoned to Mg- rich and Fe- and Mn-poor rims. Rim MgO content is typically about 10 wt% greater than core MgO content. Other hornblende constituents are not measurably variable. Biotite xenocrysts, trace mineral constituents, are present only in the domes that are most contaminated, as judged by bulk-rock variations in trace element concentrations and 87 Sr/ 86 Sr i . Biotite grains are invariably partly to almost completely altered. Microprobe analyses of the cores of the least-altered grains show that large variations in Fe and Mg and that biotite contains 2-20 times as much Mg as fresh biotite phenocrysts in other silica-rich rhyolite lavas. Fe and Mg are negatively correlated in hornblende and biotite, consistent with mixing two end-member compositions. The mass ratio of contaminant to magma was probably less than 1:100, and major constituents, including Al, were not measurably affected in hornblende. Al-in-homblende barometry yields essentially a constant calculated pressure of about 1.5 kbar, which is consistent with the interpretation that all contamination occurred in a boundary zone about 300 m thick at the top of the magma reservoir.

Alpha-track imaging (ATI) provides a map of the surface distribution of B + Li across a polished thin section and can reveal textural detail that is otherwise not accessible. It is also useful for studying minerals such as cordierite, which can be difficult to distinguish. In minerals from the Peña Negra anatectic complex, Spain, the image density decreases in the order tourmaline ≫ sillimanite ≈ muscovite ≈ sericite > chlorite > cordierite > biotite > feldspar ≈ quartz. The sillimanite image comes mostly from B, the cordierite and biotite images reflect the content of Li, and muscovite and sericite contain both. The Peña Negra rocks contain less B and Li than many European Hercynian granitic complexes, and the anatectic process retained these elements in the restites rather than concentrating them in the granitic melt.

Rare-element-enriched granitic pegmatites in and near the Lamotte and Lacome plutons of the Archean Preissac-Lacorne batholith (Quebec) are zoned from a least fractionated beryl type in the parental pluton through a transitional spodumene + beryl type at the edge of the plutons to a most fractionated spodumene type in the country rocks. Columbite-tantalite is the only primary niobium tantalum oxide mineral; its composition is dominated by Nb, Ta, Fe, and Mn. Ti is present in minor quantities (≤ 0.1 atoms per formula unit). In both plutons, the Ta/(Ta + Nb) and Mn/(Mn + Fe) of columbite-tantalite correlate positively with the degree of pegmatite evolution. These correlations are interpreted to reflect the greater solubility of Mn relative to Fe and of Ta relative to Nb in the melt. In the Lacorne suite, there is a much greater increase in Mn/(Mn + Fe) over Ta/(Ta + Nb) of columbite-tantalite with pegmatite evolution. In pegmatite bodies in the Lacorne suite, where columbite-tantalite crystallization was accompanied by garnet, the Ta/(Ta + Nb) of columbite-tantalite increases at constant Mn/(Mn + Fe) values. We interpret this to indicate that garnet saturation buffered the activities of Fe and Mn in the liquid. The degree of order of columbite-tantalite decreases with the evolution of the pegmatite suites and reflects increased cooling rates of the more evolved, and more distal, bodies of pegmatite.

Morphologies of primary rhyolitic melt inclusions in three successive populations of quartz phenocrysts in a single volcanic unit illustrate the change of inclusion shape after trapping. The 14 Ma Badlands lava flow on the Owyhee Plateau of southwestern Idaho contains two magmas with three distinct quartz populations differing in size and mor­phology. In tiny quartz crystals that nucleated shortly before the eruption, large melt inclusions retain their initial, irregular shapes, whereas the smallest inclusions show mature negative crystal shapes. Inclusions in a population of larger quartz crystals that nucleated earlier all show either negative crystal shapes or faceted shapes transitional to negative crystals. All inclusions in a third population of large corroded quartz crystals have mature negative crystal shapes regardless of inclusion size. In silicic magmatic systems, irregularly shaped inclusions in quartz may imply trapping just before eruption; analyses of such inclusions should have the greatest likelihood of revealing the magmatic volatile condi­tions driving the eruption. Electron microprobe analyses show that maturation had no effect on the composition of the trapped melt. After eruption, however, slow cooling led to coarse crystallization and loss of silica from the melt; when these inclusions were revitrified in the laboratory, they did not regain all lost silica and did not become completely homogeneous. Revitrification of inclusions that cooled more quickly showed no such loss of silica to the host. Thus, given appropriate cooling conditions, even very old (Precambrian?) silicic melt inclusions may be suitable for microbeam analysis after any necessary revitrification in the labora­tory.

Scapolites are tetragonal framework aluminosilicates with the general formula M 4 T 12 O 24 A, the major constituents being M = Na + Ca, T = Si + Al, and A = Cl + CO 3 + S-bearing anions, along with minor Fe, Sr, Ba, K, and H. Substitutions among the M, T, and A sites are independent and constrained only by net charge balance and maximum site occupancy. The cell parameters vary primarily with framework chemistry and Al-Si order and are not a strong function of the M- or A-site substitutions. End-member marialite has a = 12.06(1) and c = 7.551(5) Å, and meionite has a = 12.20(1) and c = 7.556(5) Å. Changes in cell- parameter trends in the series occur at Al 3.6 Si 8.4 O 24 and Al 4.7 Si 7.3 O 24 and are correlated with the locations of P4 2/ n-I4/m phase transitions. The scapolite group can be divided into three isomorphous series: 9.0 > Si > 8.4, 8.4 > Si > 7.3, and 7.3 > Si > 6.0, each with its own patterns of Al-Si order. Because scapolites with the same Si content have variable cell parameters, there might be variable degrees of Al-Si order or a minor, secondary influence of the M- or A-site substituents.

Coesite inclusions have been discovered in dolomite from carbonate-bearing eclogite in the Dabie UHP terrane. The eclogite boudins are enclosed in calc-silicate rocks that contain quartz aggregates after coesite in the matrix. The carbonate-bearing eclogite is com­posed of garnet, coesite or quartz, dolomite, magnesite, omphacite, rutile, ±minor phengite, apatite, and zircon, and was formed at ~760 °C, >28 kbar, and 0.01 < X CO₂ <0.1. In addition to coesite inclusions in dolomite, inclusions of coesite and quartz pseudo-morphs after coesite are abundant in garnet. One inclusion of quartz + calcite aggregates after coesite and aragonite (?) was identified in garnet. Radial fractures developed in garnet around the coesite inclusions but are absent in the dolomite. Dolomite contains about 14 mol% ankerite component, whereas magnesite contains about 2 mol% calcite and 23 mol% siderite. This is the first reported occurrence of coexisting magnesite + dolomite + possible aragonite + coesite in a single eclogitic sample and is consistent with experimentally determined stabilities of these phases at mantle conditions. Systematic mineral parageneses of eclogite, marble, and country-rock gneisses support the previous conclusion that supracrustal rocks were subducted to mantle depths of > 100 km.

Conventional and laser microprobe measurements of S isotope ratios were made on a suite of eight natural sulfide samples to develop a well-characterized, isotopically homogenous set of standards for in situ laser microprobe analysis. All the samples are from massive sulfide deposits in either high-grade metamorphic terranes or igneous systems. The data indicate that at the conventional sampling scale (>1 mm), the samples are isotopically homogeneous within the analytical uncertainty of the conventional technique (±0.2‰). At the laser microprobe sampling scale (70-100 μm), the samples are slightly heterogeneous, reflecting either actual isotope variation, imprecision associated with the laser microprobe technique, or both. A set of calibration curves was developed to stan­dardize accurately laser microprobe data generated on unknown samples. This represents a fundamental advance in in situ laser microprobe technology and is a rigorous analytical method for accurately determining δ 34 S CDT values from extremely small sample domains within sulfide minerals.

Analysis of Ca in olivine grains close to boundaries with Ca-rich phases is inherently problematic because of the effects of secondary X-ray fluorescence of Ca from the adjoining Ca-rich phase. Commercial fluorescence-correction methods assume chemical homoge­neity throughout the entire interaction volume and yield enhanced Ca concentrations for the olivine in such cases. We have determined the effects of Fe content of the olivine and Ca content of the adjoining phase on the magnitude of secondary fluorescence effects during electron microprobe analysis of Ca in olivine. Significant errors in Ca determination were observable at horizontal beam-interface distances ranging from 30 μm for an Fe-free olivine-diopside couple to 70 μm for an olivine (Fo 90.7 )-calcite couple. In general, the beam-interface distance required to suppress secondary fluorescence of Ca X-rays from the adjoining phase increases with increasing Fe content of the olivine and Ca content of the neighboring phase. Analysis of an olivine (Fo 90.7 ) wedge mounted on a standard thin-section glass slide (5.06 wt% Ca) showed significant secondary fluorescence of Ca to a wedge depth approaching 20 μm. It is recommended that olivine grains be separated and remounted in a Ca-free matrix for the determination of Ca concentrations in olivine by electron microprobe analysis.

Fluid-absent melting is believed to be an important process in the generation of melts in the lower crust and upper mantle. Breakdown of phlogopite makes H 2 O available and thus controls the conditions at which fluid-absent melting occurs. Both F and Ti in biotite have been shown to affect strongly the thermal stability of biotite. To model better the fluid-absent melting of assemblages containing phlogopite, the reaction F-phlogopite + quartz + rutile = enstatite + melt has been studied experimentally. Experiments were performed at 7, 10, and 15 kbar using a natural F-rich phlogopite with a starting com­position of F/(F + OH) = 0.43 and Mg/(Mg + Fe) (in molar proportions) = 0.94. Results indicate that the thermal stability of F-rich phlogopite + quartz + rutile is extended by as much as 450 °C relative to the KMASH system and by 300 °C relative to the Ti-free F-KMASH system. Approach to equilibrium in the experiments was assessed by conver­gence of results of melting and crystallization experiments. Phlogopite compositions from experimental products show that, although F-rich phlogopite incorporates relatively little Ti (2-3 wt% TiO 2 ), the combination of F and Ti increases the stability of phlogopite to considerably higher temperatures (~300 °C) than that of either component alone. Melts formed by the fluid-absent melting of F-rich phlogopite + quartz + rutile at temperatures > 1000 °C are granitic and strongly peraluminous. The compositions of these melts suggest that the formation of metaluminous to peralkaline A-type granites by fluid-absent melting of halogen-enriched sources is unlikely.

Randomly interstratified serpentine-chlorite (Sp-Ch) is ubiquitous in sandstones of the subsurface Tuscaloosa Formation. The Sp-Ch occurs as grain coatings, pore fillings, pel- oids, and infillings. The presence of peloids and infillings suggests that the Sp-Ch originated as odinite, a 7 Å mineral that forms at the sediment-seawater interface in shallow, tropical marine sediments. We conclude that during the very earliest stages of diagenesis, odinite peloids and infillings transformed to mixed-layer Sp-Ch, forming grain coatings with pres­ervation of some of the original odinite textures. The formation of grain-coating Sp-Ch during early diagenesis apparently preserved high primary porosity in medium-grained specimens. Fine-grained and silty sandstones, however, contain poorly formed grain coatings and are tightly cemented by quartz, implying that grain size also has some effect on authigenic quartz growth. With increasing burial depth between 1702 and 6216 m, the proportion of Sp layers in Sp-Ch decreases from 20.7 ± 1.1 to 1.3 ± 0.1%, indicating that Sp layers transform to Ch layers with increasing diagenetic grade. The polytype of the Sp-Ch is essentially Ibb at depths <2000 m. At depths >2000 m, the polytype is randomly interstratified Ibb-Iaa, and the proportion of Ibb layers decreases from ~ 100 to 49% over the interval studied. At depths <2000 m, orthohexagonal Sp transforms to Ibb Ch, but at depths >2000 m, orthohexagonal Sp transforms to Iaa Ch. There is essentially no change in Sp-Ch chemical composition with depth, implying that temperature is the dominant control on the struc­tural transformations. Mean crystallite thickness also remains relatively constant with increasing depth, indicating that diagenetic transformations in Sp-Ch do not involve crystal-growth mechanisms typical of most clay minerals. Rather, the mineralogic transformations appear to proceed on a layer-by-layer basis, with individual layers dissolving and reprecipitating within the confines of the crystallites.

Fluor-ferro-leakeite is a new amphibole species from the Canada Pinabete pluton, Questa, New Mexico, U.S.A.; it occurs in association with quartz, alkali feldspar, acmite, ilmenite, and zircon. It forms as anhedral bluish black crystals elongated along c and up to 1 mm long. It is brittle, H= 6, D meas = 3.37 g/cm 3 , D calc = 3.34 g/cm 3 . In plane-polarized light, it is strongly pleochroic, X = very dark indigo blue, Y = gray blue, Z = yellow green; X ∧ c = 10° (in β obtuse), Y = b, Z ∧ a = 4° (in β obtuse), with absorption X > Y > Z. Fluor-ferro-leakeite is biaxial positive, α = 1.675(2), β = 1.683(2), γ = 1.694(1); 2V = 87(2)°; dispersion is not visible because of the strong absorption. Fluor-ferro-leakeite is monoclinic, space group C2/m, a = 9.792(1), b = 17.938(1), c = 5.3133(4) Å, β = 103.87(7)°, V= 906.0(1) Å 3 , Z = 2. The ten strongest X-ray diffraction lines in the powder pattern are [d(I,hkl)]: 2.710(100,151), 2.536(92,2̅02), 3.404(57,131), 4.481(54,040), 8.426(45,110), 2.985(38,2̅41), 2.585(38,061), 3.122(29,310), 2.165(26,261), and 1.586(25,4̅03). Analysis by a combination of electron microprobe, ion microprobe, and crystal-structure refinement (Hawthorne et al. 1993) gives SiO 2 51.12, Al 2 O 3 1.13, TiO 2 0.68, Fe 2 O 3 16.73, FeO 8.87, MgO 2.02, MnO 4.51, ZnO 0.57, CaO 0.15, Na 2 O 9.22, K 2 O 1.19, Li 2 O 0.99, F 2.87, H 2 O calc 0.60, sum 99.44 wt%. The formula unit, calculated on the basis of 23 O atoms, is (K 0.23 Na 0.76 )(Na 1.97 Ca 0.03 )(Mg 0.46 Fe 2+ 1.14 Mn 2+ 0.59 Zn 0.07 Fe 3+ 1.93 -Ti 0.08 Al 0.02 Li 0.61 )(Si 7.81 Al 0.19 )O 22 (F 1.39 OH 0.61 ). A previous crystal-structure refinement (Hawthorne et al. 1993) shows Li to be completely ordered at the M3 site. Fluor-ferro-leakeite, ideally NaNa 2 (Fel 2+ 2 Fe 3+ 2 Li)Si 8 O 22 F 2 , is related to leakeite, NaNa 2 (Mg 2 Fel 3+ 3 Li)Si 8 O 22 (OH) 2 , by the substitutions Fe 2+ → Mg and F → OH.

Sanidine crystals precipitating from lamproitic magmas characteristically contain ap­preciable amounts of Fe 3+ . Sanidine grains from the Leucite Hills lamproites (Wyoming) have compositions in which up to 70 mol% of the KAlSi 3 O 8 molecule is replaced by KFe 3+ Si 3 O 8 , thus constituting a new natural feldspar component. Sanidine is among the last phases to crystallize in the groundmass of Leucite Hills lamproites and rarely exceeds 150 μm in length. Nearly all grains are zoned in Fe, ranging from 2.5 wt% Fe 2 O 3 in the cores to 19.6 wt% Fe 2 O 3 at rims adjacent to quenched glass. The Fe-rich rims are relatively narrow on (010) crystal faces in comparison with (001) faces (maximum width 10 μm), probably reflecting faster- (001) and slower-growing (010) crystallographic directions. A near-perfect correlation is obtained by combining Ti, Mg, and excess Si cations with Fe 3+ in tetrahedral substitution for Al. Insufficient numbers of R 2+ cations exist to balance the high concentrations of Mg through an R 2+ Mg 2+ R 1+ -1 R 3+ -1 substitution, suggesting the exchange Mg 2+ Si 4+ Al 3+ -2 in 2KAlSi 3 O 8 , analogous to that found in synthetic leucite and kalsilite. Selected-area electron diffraction and convergent-beam electron diffraction show that the Fe-rich sanidine rims have a C2/m structure with cell dimensions of a = 8.68 ± 0.15, b = 13.14 ± 0.23, c = 7.31 ± 0.15 Å, α = 90.0 ± 0.54, β = 116.0 ± 0.46, γ = 90.0 ± 0.33°, and a calculated cell volume of 747 ± 21 Å 3 . These crystallographic data are in very close agreement to those obtained previously for synthetic KFe 3+ Si 3 O 8 . The unusually high concentrations of Fe 3+ , Ti, and Mg in the sanidine rims are likely the consequence of rapid crystallization during quenching of an unusually Si- and K-rich, yet Al-deficient residual magma.

The crystal structure of namuwite, (Zn,Cu) 4 SO 4 (OH) 6 · 4H 2 O, trigonal, a = 8.331(6), c = 10.54(1) Å, V = 634(1) Å 3 , P3̅, Z= 2, has been solved by direct methods and refined to an R index of 7.9% using MoΚα X-ray data. The fundamental building unit of the structure is a gently modulated sheet of edge-sharing (Zn,Cu) octahedra parallel to (001), with one vacancy for every six occupied octahedral sites. The sheets are decorated on both sides by SO 4 and [Zn(OH) 3 (H 2 O)] 1- tetrahedra; the latter are located above and below the vacant octahedral sites. Adjacent tetrahedral-octahedral-tetrahedral (T-O-T) layers are held together by hydrogen bonds to interstitial H 2 O molecules. In synthetic Zn 4 SO 4 (OH) 6 · 4H 2 O successive octahedral layers are displaced by ½[a + b]. In namuwite, this is not the case, which suggests that the presence of Cu in the octa­hedral sites prevents this from happening. Namuwite undergoes a reversible transition, probably to the pentahydrate phase, at approximately room temperature.

Bechererite is a new mineral found on dump material from the Tonopah-Belmont mine, Maricopa County, Arizona, U.S.A. It occurs as thin elongated crystals, commonly in hemispherical aggregates, and it is associated with rosasite, hydrozincite, willemite, smith- sonite, paratacamite, and boleite. Bechererite is light green, transparent with vitreous lus­ter, and has perfect cleavage parallel to {001}. It has a density of 3.45(5) g/cm 3 and a Mohs hardness of 2-3. Optically, it is uniaxial (-) with ε = 1.611(1) and ω = 1.705(1). The empirical formula (based on 21 O atoms) is (Zn 4.93 Cu 1.01 )Zn 2 (OH) 13 (S 1 . 10 Si 0.95 )[O 5.30 (OH) 2 . 70 ], simplified to (Zn,Cu) 6 Zn 2 (OH) 13 [(S,Si)(O,OH) 4 ] 2 . The strongest lines in the X-ray powder diffraction pattern are [d obs (Å)(I obs ,hkl)] 7.37(100,001), 3.623(25,111̅), 3.282(30,012), 2.724(30,120), 2.556(50,121), and 1.572(20,410). The crystal structure was solved by direct methods and refined to R = 0.041 on the basis of single-crystal X-ray diffraction data of 281 unique reflections up to (sin θ)/λ = 0.48 Å -1 [space group P3̅, a = 8.319(2), c = 7.377(3) Å, V= 442.1(1) Å 3 , Z = 1]. The atomic arrangement of bechererite is characterized by sheets of edge-sharing octahedra with ordered vacancies, interconnected by Zn 2 O 7 groups. Isolated XO 4 tetrahedra (X = S, Si) share a common comer with one O atom of the octahedral sheet. A complex system of hydrogen bonds occurs in bechererite.

Ab initio calculations using the local density approximation and pseudopotentials were made on forsterite from 0 to 700 kbar. Our calculations are generally consistent with almost all the experimental cell-parameter and bond-length data. In contrast to some experimental data, however, we find no evidence for abrupt changes in compressional mechanism at any pressure.

This work reports X-ray photoelectron spectroscopy (XPS) and low-energy electron diffraction (LEED) results for an as-grown (100) surface of pyrite (FeS 2 ) cleaned by low-energy He + ion bombardment and thermal annealing. LEED suggests that the atomically clean pyrite surface prepared in this manner is unreconstructed. XPS S 2p data for this surface show no evidence for a monosulfide species coexisting with the disulfide moiety that characterizes pyrite. This experimental observation contrasts with many prior studies that prepared pyrite surfaces by cleavage or mechanical cutting and observed a monosul­fide species. We argue that a pyrite surface prepared directly from the as-grown surface may better represent a natural surface than does a preparation method such as cleavage.