Volume 81, Issue 7-8

Molecular orbital calculations on various aluminosilicate Q 3 T-OH and bridging O spe­cies were performed to model atomic structural changes on mineral surfaces that occur as a function of pH. Calculated vibrational frequencies are reported for the terminal T-O, T-OH, and O-H bonds of the central T cation as a test of our models, and the predicted frequencies compare well with experimental vibrational spectra of aluminosilicates. Optimized bond lengths and T-O-T angles to the central Q 3 Si 41 and Al 3+ cations in these molecules change significantly as the T-OH bond is protonated and deprotonated. Pro­tonation of terminal bonds tends to shorten and strengthen the remaining three T-O br bonds that would attach the central cation to the bulk mineral. This result is in contrast to the T-O br weakening that has been suggested previously as a mechanism for proton- assisted dissolution (e.g., Furrer and Stumm 1986; Wieland et al. 1988). Proton affinities (PA) of the Q 3 T-OH, T-OH 2 , and T-OH-T species are predicted to help delineate the process) of proton-assisted dissolution in quartz and feldspars. Theoretical results predict that the PAs of Al-OH 2 and Al-OH-Si species are comparable; thus, as Al-OH 2 surface species become stable, protons are also energetically favored to attach to bridging O atoms. In addition, we found that substitution of Al 3+ for Si 4+ in the second-nearest-neighbor site of a Q 3 Si-OH can increase the calculated PA, although the effect is diminished by the presence of a charge-balancing Na + cation. This result has implications for models that attempt to describe the behavior of aluminosilicate surfaces in terms of the component oxides. Addition of Na + to charge balance molecules strongly affects calculated structures, proton affinities, and vibrational spectra. The role of charge-balancing cations was often omitted in previous theoretical studies of aluminosilicates, but the magnitude of the charge­balancing effect could alter the results and conclusions of earlier calculations in this area.

A large-scale computer simulation of Al-Si ordering and the corresponding development of a twin microstructure in alkali feldspars is reported for the first time. In the simulation, the order-disorder transition is driven by long-range, strain-mediated interactions between ordering atoms. These interactions are the result of an elastic accommodation of two tetrahedra of different sizes, AlO 4 and SiO 4 , in the feldspar framework structure. The elementary step in changing the Al-Si distribution in this model is the interchange of two Al and Si atoms occupying neighboring comer-sharing tetrahedra. A combined Monte Carlo-lattice-relaxation method was used. An initially homogeneous disordered sample was “annealed” at a temperature, T, below the transition. The mesoscopic microstructure at T < T c is dominated by the spontaneous strain of the triclinic structure. The formation and subsequent coarsening of the pericline twins as developed in a thin slab parallel to the {010} plane was observed; the orientations of twin domain boundaries deviate signif­icantly on a local scale from the macroscopic direction predicted by the compatibility condition.

A Born-von Kármán rigid-ion, lattice-dynamical model has been applied to diopside with the use of empirical potentials derived by fitting the vibrational frequencies of a group of nesosilicates and oxides (including quartz). The Raman and infrared spectra are satisfactorily interpreted by these calculations; there is also good agreement with anisotropic atomic displacement parameters (ADPs) derived from accurate crystal-structure refinement at various temperatures. The agreement with theoretical and experimental data of independent origin confirms the physical meaning of both experimental and calculated ADPs; consequently, the correlation tensors between the displacement of different atoms, which can be evaluated in this way and cannot be obtained from Bragg diffraction measurements, are also reliable, thereby allowing bond-length correction for libration in the most general case. As for other oxides and silicates, the zero-point motion contribution amounts to about 35% of the total at room temperature. Hirshfeld’s rigid-bond test applied to the SiO 4 groups is in agreement with our calculations. The values of thermodynamic functions such as the specific heat and entropy at different temperatures are also satisfactorily reproduced, although a rigid-ion model has been used and the empirical potentials were not specifically fitted to the properties of the substance under study: For this purpose, a simple treatment of anharmonic expansion is proposed. Therefore, the possibility of transferring empirical potentials to widely different groups of minerals and that of developing a routine application to derive accurate values for various properties of pure minerals at various temperatures, starting only from crystal data, are confirmed.

The CaGeO 3 and SrGeO 3 perovskites and CaGeO 3 , SrGeO 3 , and MnGeO 3 pyroxenoids have been studied at high-temperature by X-ray diffraction (XRD) and X-ray absorption fine-structure spectroscopy (XAFS). The diffraction studies show that the back-transfor­mation of the perovskite begins at lower temperature for SrGeO 3 (~500 K) than CaGeO 3 (~945 K). An intensity reduction of the Bragg lines and the presence of a diffuse band in the diffraction patterns show the occurrence of transient amorphous phases. This transient phase contains Ge in fourfold and sixfold coordination after the loss of CaGeO 3 and SrGeO 3 perovskites, respectively. The recrystallization of the stable pyroxenoids occurs at higher temperature in a second step of the transformation. For these compounds, the anharmonic character of the first Ge-O bond was extracted from the XAJFS analysis. For the perovskites, we observed that the bulk thermal expansion and the Ge-O bond anharmonicity are closely connected. Both parameters are higher for cubic SrGeO 3 and simultaneously increase near 525 K. when CaGeO 3 changes symmetry. For the pyroxenoids, our calculated tetrahedral thermal expansion is lower than for the bulk, with Ge-O bond anharmonicity very low. Close to the melting point, anomalous motion of the tetrahedral chains is suggested by a significant increase of the Ge-O bond length.

Independent isobaric data on thermal expansion and isothermal compressibility data for lawsonite, CaAl 2 Si 3 O 7 (OH) 2 ·H 2 O, were determined between 23 and 598 °C and 0.001 and 37.7 kbar using single-crystal X-ray diffraction in a microfumace and a diamond- anvil cell, respectively. The crystal structures of lawsonite were also refined from intensity data collected at 23, 444, and 538 °C and 0.5 and 28.7 kbar. Both expansion and compression patterns are slightly anisotropic, with minor changes along c with respect to a and b. Unit-cell dimensions vary linearly with T and P: α a = 1.26(4) × 10 -5 , α b = 1.12(4) × 10 -5 , α c = 7.6(3) × 10 -6 °C -1 , and β a = 3.4(1) × 10 -4 , β b = 3.0(1) × 10- 4 , β c = 2.8(2) × 10 -4 kbar -1 . Bulk modulus, calculated as the reciprocal of cell-volume compressibility, is 1100(40) kbar. The cavities of the framework accommodating Ca and H 2 O molecules change by only about ± 5% in the investigated T and P ranges. Thus, at room pressure H 2 O molecules can be hosted in the lawsonite structure at least up to above 500 °C, where the reduction of reflection intensities shows the beginning of dehydration and breakdown of the phase. Like other dense phases, structural changes with T and P essentially affect bond lengths, whereas interpolyhedral variations mainly concern the Si-O-Si′ angle between tetrahedral pairs, which increases with temperature and decreases with pressure. The present data define the “geometric” equation of state for lawsonite on the basis of cell-volume variations: V/V 0 = 1 + 3.13(9) × 10 -5 T - 9.1(3) × 10 -4 P, where T is in degrees Celsius, P is in kilobars, and the α/β ratio is 34 bar/°C. This indicates that the cell volume of lawsonite remains unchanged with geothermal gradients of about 10 °C/km, a condition actually observed in down-going subduction slabs. Therefore, the results of high- T and high-P structure refinements are in agreement with results from multi-anvil exper­iments and confirm that lawsonite is a good candidate for carrying water down to mantle depths.

An X-ray structure-refinement procedure was developed to characterize orthopyroxene from the Serra de Magé meteorite. In the studied sample, the orthorhombic phase coexists with exsolution lamellae of C2/c augite, parallel to (100), with a* Opx = a* Aug , [101]* Opx = c* Aug , b* Opx = b* Aug . Diffraction maxima of the monoclinic phase overlap those of the ortho­rhombic phase with h + l = 2n, yielding violations of the extinction conditions for space group Pbca and simulating the lower symmetry space group P2 1 ca. Structural parameters of the Pbca phase together with the W cpx parameter, which expresses the fraction of the C2/c phase present, were refined first. F c values were calculated using the equation with [F c ]C px values taken from the structure refinement of an augite. Observed F 0 values were then corrected by subtracting the calculated contri­bution of the monoclinic phase with use of the equation and used for a further refinement of the orthorhombic phase. The final residual indices, R obs , were 2.11 and 1.78% for two crystals with different augite contents. This refinement pro­cedure confirms Pbca aS the corect space group and provides more accurate structural parameters for the Serra de Magé orthopyroxene.

The cell parameters of deuterated gypsum have been extracted from Rietveld analysis of powder neutron diffraction data within the temperature range 4.2-320 K. The thermal expansion of gypsum is highly anisotropic along the b axis principally because of the effect of the D2···O1 hydrogen bond. The high-temperature limits for the expansion coefficients for the cell edges a, b, and care 3.98 × 10 -6 , 4.36 × 10 -5 , and 2.53 × 10 -5 K -1 , respectively, and for the cell volume the limit is 6.96 × 10 -5 K -1 . The β angle displays oscillatory variation, reflecting a change in the influence of at least two coupled processes. Empirical data analysis, within this temperature range, results in α β = 1.251 × 10 -6 sin(0.0116T + 0.311) K -1 . = 1.251 × 10 -6 sin(0.0116T + 0.311) K -1 .

Previous work has described the structural and diffraction criteria for distinguishing between tv-1M, cv-1M, and m-1M illite varieties, where tv-1M illite corresponds to a structure with trans vacant (tv) 2:1 layers, cv- 1M illite consists of cis-vacant (cv) 2:1 layers, and m-1M illites consist of 2:1 layers with random distributions of octahedral cations over trans and cis sites within each layer. Detailed analysis of calculated and experimental X-ray diffraction (XRD) patterns from illites and mixed-layered illite-smectites (I/S) with a range of cis- and trans-vacant layer interstratification, and consideration of unit-cell parameters for pure tv-1M and cv-1M, provides insight into the nature of the diffraction effects from the interlayered structures with different types and contents of rotational stacking defects. Structural fragments of the 2M 1 and 2M 2 polytypes exist within 1 Md structures in which rotational disorder is dominated by n·120 and n·60°, respectively. Such structural frag­ments constitute coherent-scattering domains and affect the peak positions of the diag­nostic 11 l reflections for rotationally disordered I/S and illites in different ways depending on the proportion of cis- and trans-vacant 2:1 layers. Simple techniques that do not require computer programs were developed for the semi-quantitative determination of the proportion of tv and cv 2:1 layers in I/S and illites where such layers are interstratified. These techniques may help to reveal the diversity of I/S and illite samples relating to the conditions of their formation and transformation.

Natural apatites used in fertilizer industries often contain trace amounts of Cd, which may reach concentrations of several tens to a few hundred parts per million. Cd is not eliminated during the production of phosphate fertilizers, and its concentration in the final product can exceed environmental norms. Knowledge of the chemical state of Cd in apatite ores is a prerequisite for the design of technical processes of extraction. In the present study, Cd K-edge EXAFS spectroscopy was used to investigate the structural environment of Cd present in sedimentary apatite ores from West Africa. These apatites are fluorinated and contain goethite, quartz, and crandallite as ancillary phases detected by X-ray diffraction or EXAFS spectroscopy. Cd K-edge EXAFS spectra for two natural samples were analyzed and compared with those for Cd-containing reference minerals, including hydroxylapatite, goethite, otavite, and crandallite. A good spectral resemblance was observed between natural products and synthetic apatite containing small amounts of Cd. This spectral likeness indicates that the majority of Cd atoms are diluted in the apatitic framework and do not form Cd 10 (PO 4 ) 6 (OH,F) 2 clusters. This finding was confirmed by quantitative analysis of the EXAFS spectra, which indicated that Cd atoms are surrounded by nearest 0 atoms at 2.33 Å, next-nearest P atoms at ~3.53 Å, and a third-nearest shell of Ca atoms at ~4.02 Å. A comparison of these data with those obtained for synthetic apatites allowed us to assess that Cd occupies both Ca crystallographic sites with a slight preference for the Ca2 site.

Transformations induced by light in realgar and β-As 4 S 4 have been studied by means of X-ray diffraction (powder and single-crystal) and Raman laser spectroscopy. Both poly­chromatic and monochromatic light alter realgar and the β phase into pararealgar. By lowering the reaction rate (long-wavelength pass filter, low laser power) an intermediate product, corresponding to the so-called “χ phase,” can be observed during the alteration process from β-As 4 S 4 to pararealgar (polychromatic filtered light and laser light) and that from realgar to pararealgar (only with polychromatic filtered light). The powder pattern of the χ phase is fairly similar to that of β-As 4 S 4 but differs from realgar, and it can be indexed on the basis of an expanded β-phase unit cell. Lattice parameters are as follows: a = 9.758(5), b = 9.522(5), c = 9.074(5) Å, 0 = 100.84(5)°, and a = 9.757(4), b = 9.512(7), c = 9.089(4) Å, 0 = 100.97(3)° for the χ phase obtained from alteration of realgar and β-As 4 S 4 , respectively. The formation of the χ phase is preceded by a strong anisotropic increase of the unit-cell volume of realgar and the β phase. The effects of light exposure on lattice parameters were studied using single-crystal X-ray diffraction: In realgar a and c sin β increase linearly with increasing exposure times, whereas b remains substantially unchanged. In the β-As 4 S 4 crystals, b and c sin β increase while a decreases so that, after 1560 min, the unit-cell dimensions match those of the χ phase. Raman spectroscopy, which allows monitoring of the alteration process in real time, indicates the presence of the same As 4 S 4 cage molecule in the χ phase as in realgar and β -As 4 S 4 and confirms the similarity between the structure of χ and that of β. However, a less-ordered crystal struc­ture and a lower molecular symmetry is suggested from analysis of the spectroscopic results.

Mueller’s model has been widely applied to modeling kinetic experimental data on the ordering-disordering of cations between two nonequivalent sites. This model is valid only for pure or nearly pure binary systems. For ordering-disordering involving three or more cations (multiple cation or multi-cation) between two sites, Mueller’s methodology, which is based on a two-cation exchange reaction, yields no explicit general solution. On the basis of a single-cation exchange reaction, we present an alternative kinetic model for multi-cation ordering-disordering in minerals with two nonequivalent sites. This model is not only suitable for binary systems but is also valid for multi-cation ordering-disordering at two nonequivalent sites. In addition, two kinetic coefficients for each individual cation can be easily obtained using nonlinear parameterization. A comparison of reported experimental data with theoretical calculations has shown that our model can fit both binary and multi-cation ordering-disordering very well and can also explain and predict many kinetic features observed in experiments.

Quartz overgrowths from Keuper sandstones of the Paris basin were examined using cathodoluminescence (CL) microscopy and spectroscopy coupled with a scanning electron microscope (SEM). With the use of standard cold CL equipment, it was observed that the emission of authigenic quartz is much less intense than that of detrital quartz grains, but a reversal of intensity was observed with scanned CL in the 200-800 nm range. The main CL emission band of diagenetic quartz is at 330-340 nm in the UV range, with other bands in the visible range. The determination of trace element distributions in authigenic quartz by in situ analyses and by SIMS imagery reveals the correlation of the 330-340 nm emission band with the highest Al and Li contents. The cathodoluminescence emission of diagenetic quartz in the UV range appears to be influenced by the coupled substitution of Al and Li into the crystal. Two explanations are suggested: (1) Al, Li, or both serve as activators, and (2) the incorporation of Al and Li causes lattice defects that lead to an enhancement of the intrinsic luminescence. Trace element analyses and fluid-inclusion studies revealed that quartz overgrowths precipitated from a fluid that partially originated from an Li-enriched primary brine derived from eastern Triassic evaporites of the Paris basin. The specific CL emission band in the UV range can be linked to the diagenetic environment. The chemistry of the fluids appears to be the essential parameter, whereas precipitation temperatures do not influence the occurrence of such emissions.

We have determined major (Si, Zr, Hf), minor (Al, Y, Fe, P), and trace element (Ca, Sc, Ti, Ba, REE, Th, U) concentrations and Raman spectra of a zoned, 200μm zircon grain in lunar sample 14161,7069, a quartz monzodiorite breccia collected at the Apollo 14 site. Analyses were obtained on a thin section in situ with an ion microprobe, an electron microprobe, and a laser Raman microprobe. The zircon grain is optically zoned in birefringence, a reflection of variable (incomplete) metamictization resulting from zo- nation in U and Th concentrations. Variations in the concentrations of U and Th correlate strongly with those of other high-field-strength trace elements and with changes in Raman spectral parameters. Concentrations of U and Th range from 21 to 55 ppm and 6 to 31 ppm, respectively, and correlate with lower Raman peak intensities, wider Raman peaks, and shifted Si-O peak positions. Concentrations of heavy rare earth elements range over a factor of three to four and correlate with intensities of fluorescence peaks. Correlated variations in trace element concentrations reflect the original magmatic differentiation of the parental melt ~4 b.y. ago. Degradation of the zircon structure, as reflected by the observed Raman spectral parameters, has occurred in this sample over a range of α-decay event dose from ~5.2 × 10 14 to 1.4 × 10 15 decay events per milligram of zircon, as calculated from the U and Th concentrations. This dose is well below the ~10 16 events per milligram cumulative dose that causes complete metamictization and indicates that laser Raman microprobe spectroscopy is an analytical technique that is very sensitive to the radiation-induced damage in zircon.

This contribution deals with the crystal chemistry of phlogopite and Fe 3+ -rich phlogo- pite from the Tapira alkaline-carbonatite complex (Brazil) to assess the petrological sig­nificance and genetic conditions of these rocks. The Tapira complex consists of a layered intrusion composed mainly of ultramafic rocks (dunite, wehrlite, clinopyroxenite, bebedourite, gamet-magnetitite, perovskite-magnetitite, and glimmerite) with subordinate car- bonatite. The wide range of textural, optical, and crystal-chemical characteristics of phlog­opite is related to the variation of f O₂ , a H₂O , and a CO₂ as well as magma bulk-chemical composition during fractional crystallization. Phlogopite from alkaline-silicate rocks (rang­ing from dunite to bebedourite) is characterized by fairly constant Al content, moderate [4] Fe 3+ substitution, and variable amounts of Ti. The [4] Fe 3+ substitution, accompanied by crystals showing reverse pleochroism, increases during fractional crystallization. These features correspond to crystallization at low pressure and high and a H₂O in the presence of moderate saturation in Ti-bearing phases, Al 2 O 3 in the magma, or both. Phlogopite from silicate-carbonatite rocks, classified as ferriphlogopite on the basis of strong reverse pleochroism related to [4] Fe 3+ tetrahedral substitution, also presents low to very low Al, Fe 2+ , and Ti contents. These features suggest very high f O₂ , H 2 O, and CO 2 conditions in the presence of strong saturation in Ti-bearing phases as well as very low Al 2 O 3 content in the liquid. The crystal-structure refinements of Tapira phlogopite show that Fe 3+ substitutes for Si in tetrahedral sites; Fe distribution is completely disordered, so the resulting space group is C2/m. The octahedral-site composition is similar to that of phlogopite, the octahedral sites being preferentially occupied by Mg. The presence of Fe 3+ in the tetrahedral sheet enlarges the whole structure. This enlargement is reflected by an increase in cell-edge lengths and a decrease in β-angle values. The increase in distortion of the tetrahedral ring (a angle up to ~ 11°) is necessary for the tetrahedral and octahedral sheets to fit together.

Exsolved pairs of hornblende and cummingtonite from two localities in southern New England, U.S.A., have been studied using transmission electron microscopy (TEM) to measure the crystallographic orientation of the exsolution lamellae. Both the cummingtonite and the hornblende show multiple generations or stages of exsolution. Electron microprobe analyses gave compositions of the pre-exsolved (averages) and the coarsest exsolved amphiboles. In amphiboles, two orientations of lamellae are usually present and nearly parallel to {100} and {101̄} of the host. The observed lamellar orientations are consistent with orientations predicted by the optimal phase-boundary (OPB) theory and are a function of small differences in a, c and β of the host and exsolved amphiboles ( Δa = a host- a lamella , Δβ = β host - β lamella , Δc = c host - c lamella ). Because these lattice parameters vary with T and composition, the precise orientations of the lamellae are controlled by these variables. The widths of determined exsolution lamellae varied from microscopic (micrometer size) to submicroscopic (nanometer size). The TEM images show, in some cases, multiple generations of lamellae with lamellar orientations near {101̄} differing up to 4.1°, whereas orientations near {100} differed up to 4°. For two pairs of coexisting amphiboles, the T dependence of the lattice parameters from 25 to 600 °C was measured using a Guinier camera. The OPB calculations indicate that for all the samples the different lamellar generations formed between about 780 and 300 °C (±80 °C, on the basis of maximum errors of the lattice-parameter determinations). Lamellae of the same generation all showed nearly the same width. Exsolution temperatures could not be derived from the "100" lamellae because relatively small variations in Δc, which controls exact orientation of the "100" lamellae, could not be measured accurately enough.

Hydrothermal bulk-loss experiments employing radiogenic Ar ( 40 Ar ∗ ) were performed to determine whether 40 Ar ∗ diffusivity in biotite increases with Fe content. Diffusion laws determined for intermediate and Fe-rich biotite assuming single-domain diffusion (infi­nite-cylinder geometry) are remarkably similar: Fe-mica biotite (X anitee = 0.71) D = 0.40 +096 -0.28 exp[-(50500 ± 2.2)/RT] and Cooma biotite (X anite = 0.54) D = 0.075 +0.049 -0.021 exp [-(47100 ± 1.5)/RT]. The nearly identical results for Fe-mica biotite and Cooma biotite and their similarity to those from previous studies indicate that most biotite grains of intermediate composition possess comparable 40 Ar ∗ diffusion properties. Because limited grain breakage and volumetrically minor recrystallization is unavoidable during hydrothermal heating in bulk diffusion experiments, these diffusion laws necessarily provide upper limits to 40 Ar ∗ loss by intercrystalline diffusion. The measured rates of 40 Ar ∗ loss from biotite agree reasonably well with expectations based on single-domain volume dif­fusion using infinite-cylinder geometry when experimental uncertainties are taken into account. However, lack of information regarding 40 Ar ∗ gradients within the hydrothermally treated mica prevents us from precluding more complex diffusion mechanisms in­volving high diffusivity pathways. In this paper we consider the significance of bulk-loss 40 Ar ∗ diffusion experiments and discuss how diffusion parameters determined in the lab­oratory may be applied to thermochronology provided suitable constraints are available.

A thermodynamic formulation for the sorption of H 2 O in clinoptilolite has been ob­tained from analysis of equilibrium data collected by thermogravimetry on near end- member Ca-, Na-, and K-exchanged natural clinoptilolite (Fish Creek Mountains, Neva­da). Temperature and pressure of the experiments ranged from 25 to 250 °C and 0.2 to 35 mbar H 2 O vapor pressure. Equilibrium of three clinoptilolite species was successfully formulated with the following expression for the Gibbs free energy of hydration as a function of temperature and pressure: Δμ Hy /T = Δμ 0 Hy /T 0 + ΔH̅⁰ Hy (1/T - 1/T 0 ) - 3R[ln(T/T 0 ) + (T 0/ T - 1)] + R ln[θ/(1 - θ)P] +W 1 /T·θ+W 2 /T· θ² where R is the gas constant, P is the vapor pressure of H 2 O, W 1 and W 2 are the excess mixing parameters, and 6 is the ratio H 2 O/(maximum H 2 O) with maximum water contents for the K, Na, and Ca end-members of 13.49, 15.68, and 16.25 wt%, respectively. The molar Gibbs free energy of hydration for calcium, sodium, and potassium clinoptilolite is -36.13 ± 3.02, -29.68 ± 3.77, and -25.53 ± 1.37 kJ/mol H 2 O, respectively. The molar enthalpy of hydration for these phases is -76.92 ± 2.88, -74.19 ± 3.46, and -67.78 ± 1.25 kJ/mol H 2 O. The thermodynamic formulation is applied to the occurrence of clinoptilolite at Yucca Mountain, Nevada, where the proposed emplacement of nuclear waste would lead to heating of clinoptilolite-bearing tuffs. Rock units with abundant clinoptilolite (or by anal­ogy other hydrous phases) would remain significantly cooler than units with anhydrous minerals and would evolve a substantial volume of water.

Phase relations in the pseudotemary system CaO-CoO-SiO 2 have been established at 1323 K. Three quaternary oxides were found to be stable: CaCoSi 2 O 6 with clinopyroxene (Cpx), Ca 2 CoSi 2 O 7 with melilite (Mel), and CaCoSiO 4 with olivine (Ol) structures. The Gibbs energies of formation of the quaternary oxides from their component binary oxides were measured using solid-state galvanic cells incorporating yttria-stabilized zirconia as the solid electrolyte in the temperature range of 1000-1324 K. The results can be sum­marized as follows: CoO (rs) + CaO (rs) + 2SiO 2 (Qtz) → CaCoSi 2 O 6 (Cpx), ΔG 0 f = - 117920 + 11.26T (± 150) J/mol CoO (rs) + 2CaO (rs) + 2SiO 2 (Qtz) → Ca 2 CoSi 2 O 7 (Mel), ΔG 0 f = -192690 + 2.38T (±130) J/mol CoO (rs) + CaO (rs) + SiO 2 (Qtz) → CaCoSiO 4 (Ol), ΔG 0 f = -100325 + 2.55T (±100) J/mol where rs = rock salt (NaCl) structure and Qtz = quartz. The uncertainty limits correspond to twice the standard error estimate. The experimentally observed miscibility gaps along the joins CaO-CoO and CaCoSiO 4 - C O2 SiO 4 were used to calculate the excess free energies of mixing for the solid solutions Ca χ -CO 1-χ -O and (Ca γ -Co 1-γ )CoSiO 4 : ΔG E = X(1 - X)[31975X + 26736(1 - X)] J/mol and ΔG E = 23100 (±250) Y(1- Y) J/mol. A T-X phase diagram for the binary CaO-CoO was computed from the thermodynamic information; the diagram agrees with information available in the literature. The computed miscibility gap along the CaCoSiO 4 -Co 2 SiO 4 join is associated with a critical temperature of 1389 (± 15) K. Stability fields for the various solid solutions and the quaternary compounds are depicted on chemical-potential dia­grams for SiO 2 , CaO, and CoO at 1323 K.

The heat capacity of a three-phase sample of Fe 0.990±0.005 O, Fe 3 O 4 , and Fe (mole fractions 0.915, 0.078, and 0.007, respectively) has been measured by adiabatic shield calorimetry at temperatures from 13 to 450 K. Fe 0.99 O and magnetite are formed as a metastable intermediate on heating of a quenched nonstoichiometric wüstite with composition Fe 0.9374 O. The small amount of Fe present stems from the second disproportionation reaction, in which the stable two-phase mixture of Fe and magnetite is formed from Fe 0.99 O. The value of ΔS m (Fe 0.99 O, 298.15 K), 60.45 J/(K·mol), is derived from the entropy of the three- phase sample and recommended standard entropies of Fe and magnetite. The character of the magnetic order-disorder transition changes with composition and is strongly co­operative in Fe 0.99 O, with T n ≈ 191 K. A minor, seemingly higher order transition is observed at ~ 124 K. It is caused by the Verwey transition in the magnetite. This mag­netite, formed in metastable equilibrium with Fe 0.99 O, is presumably more Fe-rich than magnetite in stable equilibrium with Fe.

Plagioclase, the most abundant phenocryst at Mount Pinatubo, displays varying textures and compositions within the 1991 eruption products. In June 7-14 dome-forming ande­site, plagioclase phenocrysts show prominent rims with higher MgO (0.04-0.06 wt%), Fe 2 O 3 T (0.6-0.8 wt%), and K 2 O at given An than the interiors. The compositions of the rims are identical to those of microlites, which are abundant in the groundmass glass. White dacitic pumice, the most voluminous product of the June 15 eruption, contains partly corroded plagioclase phenocrysts but no prominent rims and no microlites. The interiors of phenocrysts from the dacitic pumice and the dome-forming andesite are re­markably similar in terms of textures and compositions. They show oscillatory zoning (mostly An 35-60 ), low MgO (<0.02 wt%) and Fe 2 O 3 T (0.10-0.30 wt%), and similar K 2 O at given An. This similarity indicates that the two types of plagioclase phenocrysts formed in the same rhyolitic melt. The oscillatory zoning likely formed by temperature fluctuations in the convecting magma and incorporation and degassing of external fluids. A portion of the felsic magma (~800 °C) mixed with a mafic melt (~ 1000-1100 °C) to become an andesitic magma that extruded to form the June 7-14 dome. All plagioclase phenocrysts in the andesite were derived from the felsic magma. The mixing caused de stabilization of phenocrysts, forming sieve textures, dusty zones, and partial resorption. Extrusion of the mixed magma resulted in overgrowths on once-resorbed phenocrysts and the nucleation of plagioclase microlites in the groundmass glass. In the unmixed, remaining portion of the felsic magma, some plagioclase underwent partial resorption but did not develop overgrowths. This lack of overgrowths and the absence of microlites in the ground- mass glass of the June 15 dacitic pumice indicate rapid magma ascent during the eruption and a short time span between the injection of a mafic melt and the cataclysmic eruption, supporting the linkage between the two.

Electron microprobe analysis of amphiboles from Pargas, Finland, shows up to 2.42 Ca apfu (atoms per formula unit), far in excess of the usual maximum value of 2.00 Ca apfu observed in amphiboles. The most Ca-rich composition is (K 0.12 Na a0.35 Ca 0.52 )(Ca 1.92 - Mn 2+ 0.05 Fe 2+ 0.03 )(Mg 3.73 Fe 2+ 0.43 Ti 0.02 Al 0.82 )(Si 5.63 Al 2.37 )O 22 [(OH) 0.46 F 1.54 ]. Fluor-cannilloite, ideally CaCa 2 (Mg 4 Al)(Si 5 Al 3 )O 22 F 2 , is a new amphibole species in which Ca is dominant at the A site. Fluor-cannilloite occurs as isolated anhedral grains and small granular aggregates in a marble together with calcite, muscovite, anorthite, aluminous diopside, pyrope, and fluorite. It is grayish green with a grayish white streak, brittle, H = 6, D meas = 3.05 g/cm 3 , and shows perfect {110} cleavage. In plane-polarized light, it is pale greenish gray to almost colorless, rather cloudy and mottled in appearance, and is not pleochroic. Fluor-cannilloite is biaxial positive, α = 1.611(2), β = 1.616(2), γ = 1.633(2), 2V = 49(2)°, dispersion weak r > ν . It is monoclinic, space group C2/m, a = 9.826(4), b = 17.906(9), c = 5.301(3) Å, β = 105.41(4)°, V = 899.2 Å 3 , Z = 2. The ten strongest X-ray diffraction lines in the powder pattern [d in Å(I)(hkl)] are 8.936(4)(020), 8.355(3)(110), 3.366(4)(131), 3.107(7)(310), 2.686(10)(151), 2.578(10)(061), 2.328(4)(3̄ 51), 2.165(6)(171), 2.036(5)(202), 1.435(9)( 4̄ .10.1). The structures of fluor-cannilloite and another Ca-rich amphibole from Madagascar were refined to R indices of ~1.5% using reflection intensities collected with MoK α X radiation. Site populations were assigned from the refined site-scattering values, consid­erations of mean bond lengths, and the unit formulas calculated from the electron micro­probe analyses. Both the electron microprobe data and the site-scattering refinements show significant Ca (up to 0.52 apfu) together with Na and K at the A site. Both A2 and Am sites are occupied, with the A2 site having the higher electron density. Local bond-valence considerations indicate that Ca is equally partitioned between the A2 and Am sites, where­as K occupies Am and Na occupies A2.

Cl- and SO 4 -bearing cancrinite-like minerals from Italian localities were studied by means of EPMA, X-ray diffraction, and IR spectroscopy to relate their chemical composition, IR data, and structural features. The framework of the cancrinite-like minerals consists of six-membered rings of alumino silicate tetrahedra stacked along c. Six types of subunits resulting from different stacking sequences can be identified: free channel, cancrinite, sodalite, losod, liottite, and giuseppettite cages. The free channel is possible only with an AB sequence; in fact, the insertion of a C layer interrupts the channel forming the various cages. The occurrence of these structural subunits seems to correspond to the different chemistries of the cancrinite-like minerals, especially their anion content. According to this model, the SO 4 group seems to playa major role. With the exception of giuseppettite, it tends to fill completely the available voids within the frameworks.