Phillip D. Ihinger

The influence of bulk composition on the speciation of water in silicate glasses

Infrared spectroscopy was used to determine the concentrations of molecular water and hydroxyl groups in hydrous rhyolitic, orthoclasic, jadeitic, and Ca−Al-silicate glasses synthesized by quenching of melts from elevated presure and temperature. The rhyolitic glasses and some of the Ca−Al-silicate glasses were quenched from water-vapor-saturated melts and used to determine the solubility of water in melts of these compositions. For all compositions studied, hydroxyl groups are the dominant hydrous species at low total water contents, whereas molecular water dominates at elevated water contents. Although the trends in species concentrations in all these compositions are similar, the proportions of the two hydrous species are influenced by silicate chemistry: increasing silica content and K relative to Na both favor molecular water over hydroxyl. Results on rhyolitic glass demonstrate that molecular water is also favored by decreasing temperature at T<850°C. For rhyolitic glasses quenched from vapor-saturated melts, the mole fraction of molecular water is proportional to water fugacity for P(H2O)≤1500 bars, demonstrating that the behavior of molecular water is approximately Henrian at total water contents up to at least several weight percent. Data on water solubility for albitic, orthoclasic, and Ca−Al-silicate melts to higher pressures can also be fit by assuming Henrian behavior for molecular water and can be used to set constraints on the partial molar volume of water in these melts. The demonstration of Henrys law for molecular water in these liquids provides a link between spectroscopic measurements of microscopic species concentrations and macroscopic thermodynamic properties.

New calibration of infrared measurement of dissolved water in rhyolitic glasses

This paper presents a new calibration for infrared analyses of dissolved water and its species concentrations in rhyolitic glasses. The new calibration combines infrared/manometry measurements and infrared study of hydrous rhyolitic glasses heated at different temperatures. The heating experiments show that the ratio of the molar absorptivity of the 5230 cm−1 band to that of the 4520 cm−1 band varies with water concentration. Therefore, earlier calibrations assuming constant molar absorptivities are not accurate. Using our new calibration, total water concentration, and species concentrations can be calculated as follows: (ϱϱ0)C1 = a0−523, (ϱϱ0)C2 = (b0 + b1−523 + b2−452)−452, and C = C1 + C2, where C1, C2, and C are the mass fractions of molecular H2O, H2O present as OH, and total H2O, πϱ0 is the ratio of the density of the hydrous glass to that of the anhydrous glass and is approximately 1 − C, −523 and −452 are the absorbances (peak heights) of the 5230 cm−1 and 4520 cm−1 bands per mm sample thickness and relative to a baseline that was fit by a flexicurve, a0 = 0.04217 mm, b0 = 0.04024 mm, b1 = −0.02011 mm2, and b2 = 0.0522 mm2. The new calibration has a high internal reproducibility in calculating H2Ototal, six times better than the calibration of Newman et al. (1986). We expect the new calibration to be accurate in retrieving H2Ototal for H2Ototal ≤ 5.5 wt% and in retrieving molecular H2O and OH concentrations for H2Ototal ≤ 2.7 wt%. Using the new calibration, the equilibrium coefficient K for the reaction H2O + O = 2OH is independent of H2Ototal (for H2Ototal ≤ 2.4 wt%) at a given temperature and can be expressed as lnK = 1.876 − 3110T, where T is in K. The bulk water diffusivity reported before is not affected by the new calibration, but the molecular H2O diffusivity will be roughly 4–30% greater.

KINETICS OF THE REACTION H2O + O = 2OH IN RHYOLITIC AND ALBITIC GLASSES: PRELIMINARY RESULTS

Exsolution phenomena of kirschsteinite (CaFeSi04) in olivine have been studied by single-crystal X-ray diffraction techniques and scanning electron microscopy (SEM) of oriented polished thin sections (PTS) of three single crystals separated from the Antarctic angrite LEW86010, supplemented by micro-area X-ray diffraction with the Laue method (MXL) by synchrotron radiation (SR) for PTS of a rock chip of LEW8601O. The cell dimensions of the host olivine and exsolved kirschsteinite are a = 4.79(3), b = 10.39(5), and c = 6.06(3) A, and a = 4.87(5), b = 11.14(10), and c = 6.36(5) A, respectively, from the precession photos. The PTS of olivine single crystals oriented parallel to (100) show exsolution lamellae of kirschsteinite up to I 0 ~m in width. The two sets of lamellae are symmetrically related and parallel to (031) and (031). Electron microprobe analysis gave Si02 33.1, Ti02 0.07, Al203 0.03, FeO 49.4, MnO 0.61, MgO 13.4, CaO 2.2, Cr203 0.02, V203 0.01, NiO 0.05 (sum 99.4 wt%) for the host olivine and Si02 33.3, Ti02 0.03, FeO 31.5, MnO 0.39, MgO 5.4, CaO 28.5, Cr203 0.02, NiO 0.05 (sum 99.2 wt%) for the exsolved kirschsteinite. The results from MXL for the olivine crystals on the rock PTS are compatible with the observation on the single crystals that the lamellae are parallel to (031) and (031). The (031) and (031) planes have been known to be twin planes for olivine, and the twinning is by reticular pseudomerohedry based on a quadruple lattice. Although other reported exsolved precipitates in meteoritic olivines exist as inclusions, kirschsteinite in LEW86010 olivine takes the form of lamellae. Our explanation is that LEW86010 olivine is Fe-rich and that lamellar precipitates are more easily formed than inclusions because exsolution lamellae along {031} in Fe-rich olivine maintain lattice coherency.

The speciation of dissolved water in rhyolitic melt

Concentrations of water molecules and hydroxyl groups have been measured in rhyolitic glasses with 0.5% to 5.0% H_2O_t using infrared spectroscopy at room temperature. The glasses were cooled at ~10^2 C/s after having been held at 400 to 600C, for sufficient time for the equilibrium distribution of species to have been reached. The speciation of water in samples with greater than 2.5 wt.% total dissolved water and quenched rapidly from temperatures ≥600C were shown to reequilibrate during quench. However, samples with less than 2.0% water and quenched rapidly from ≤600C, and those with less than 5.5% water and quenched rapidly from ≤500C, did not undergo changes on quench and record the equilibrium species concentrations of the experimental run conditions. Knowledge of the equilibrium speciation of water in samples at lower temperatures can be used to predict the species concentrations at magmatic temperatures and allow us to explore the effect of melt structure on the physical properties of natural hydrous magmas. Ideal mixing models can be used as a rough approximation for modeling the solution of water in rhyolitic melts with less than 2.5 wt.% total water: ln[(X^(melt)_(OH))^2/(X^(melt)_(H_2O_m) X^(melt)_(O))] = ln K = 1.89 ± 0.05 - (3120 ± 40)/T, where X^(melt)_i is the mole fraction of species i on a single oxygen basis, H_2O_m = water molecules, O = anhydrous oxygens, and T is temperature in Kelvin. This fit provides a standard state enthalpy and entropy of ΔH° = 25.9 ± 0.4 kJ/mol and ΔS° = 15.7 ± 0.4 J/mol • K for the mixing of water molecules in rhyolitic melt. At high water contents, either a modification to the infrared calibration or more complex models (such as a regular solution model) are required to fit the data. Our measurements differ with recent studies using in situ measurement techniques that show lesser concentrations of molecular species at magmatic temperatures, and we address concerns associated with the in situ method. Our study on quenched glasses can be applied to natural rhyolites; using measured species concentrations, the “apparent” equilibration temperature can be calculated to within 12°C (2σ uncertainty) which can be used to determine the cooling rate of a naturally quenched rhyolitic glass.

Influence of water on nucleation kinetics in silicate melt

Abstract Steady-state rates and induction times for crystal nucleation in lithium disilicate melt depend exponentially on water content (up to 975 ppm). Time-dependent crystal number densities exhibit self-similar behavior as a function of water content, thereby allowing for a description of all data by a universal curve. This behavior provides a powerful tool for predicting crystallization histories, implies that mechanisms involved in nucleation are accelerated by water, and suggests that no new reaction pathways arise with the addition of water. The influence of water on crystal nucleation cannot be solely explained by the effect of water on viscosity; in addition, the free energy barrier to nucleation is ∼ 8% with an addition of water from 130 to 975 ppm.

The isotopic composition of hydrogen in nominally anhydrous mantle minerals

The common, nominally anhydrous minerals of the Earth’s mantle are an important reservoir for hydrogen in the Earth’s interior due to their ability to incorporate hydroxyl in trace quantities. We present measurements of the isotopic composition (D/H ratio) of hydrogen from upper-mantle derived garnet, enstatite and augite employing a two-step extraction procedure with infrared spectroscopic monitoring of OH removal from the samples. Raw δD values of all samples range from −80 to −130‰, and blank-corrected δD of high-purity separates with <50% blank contribution to the measured gas lie in the range −92 to −113‰. These δD values, together with predictions of fractionation behavior from IR spectroscopic systematics, suggest that the D/H ratio of the nominally anhydrous mineral reservoir may differ from that of other mantle hydrogen in being relatively depleted in deuterium. To the extent that nominally anhydrous minerals control the distribution of water between the Earth’s hydrosphere and mantle, isotopic fractionations between OH in these minerals and hydrous melts or fluids may contribute to the large D/H fractionation between the mantle and the hydrosphere.

Heterogeneous crystal nucleation on bubbles in silicate melt

Abstract Experiments reported herein document heterogeneous crystal nucleation on bubbles in supercooled lithium disilicate melt. Crystalline lithium disilicate (Li2Si2O5) nucleated and grew on small bubbles (∼1 μm) with a one-to-one correspondence between the number of bubbles and crystals (ranging from <102 to ∼105 bubbles/mm3). Crystals grew on large bubbles (>100 μm) only in samples fused in N2, suggesting a chemical control on nucleating efficiency. Bubbles ∼1 μm in diameter served as nucleation sites for polycrystalline lithium disilicate spherulites; bubbles smaller than ∼1 μm served as nucleation sites for the more common ellipsoidal crystalline form. This difference in behavior might be due to the additional surface area available for crystal nucleation on the 1 μm bubbles. Our findings suggest that superliquidus thermal history can influence crystal nucleation via bubble formation induced by supersaturation, and has implications for both natural samples and experimental studies. Heterogeneous crystal nucleation on bubbles may serve as an efficient nucleation mechanism in natural degassing magmas and may aid in the formation of fine-grained groundmasses common to many volcanic rocks. Furthermore, we have documented a new mechanism of spherulite formation in highly supercooled silicate melt, similar to conditions thought to exist during devitrification of natural glasses. The ability of crystals to nucleate on bubbles can be exploited in the production of commercial glass-ceramic materials.

The color of meteoritic hibonite: an indicator of oxygen fugacity

Hibonites similar in composition to those found in Ca-Al-rich inclusions change color from blue, to green, to orange, to nearly colorless as oxygen fugacity is increased at high temperature from below the iron-wustite buffer up to air. The development of the blue color is correlated with the growth of an absorption band at 715 nm in the optical spectra of the hibonites as the oxygen fugacity is reduced. The growth of this band is attributed to the increasing concentration of Ti^(3+) in these hibonites with decreasing oxygen fugacity. The blue hibonites in meteorites reflect equilibration under reducing conditions; we estimate, based on the intensity of the 715 nm band, that the hibonite in the Blue Angel inclusion indicates an oxygen fugacity 4–5 orders of magnitude more oxidizing than that expected in the early solar nebula. This may be due to formation in an anomalously oxidizing region of the nebula or to oxidation during cooling or later alteration. The orange hibonites in Allende reflect oxygen fugacities approximately ten or more orders of magnitude more oxidizing than the expected primitive nebula; this color probably indicates alteration of initially more reduced (blue?) hibonites. The colorless hibonite in the HAL inclusion reflects highly oxidizing conditions and/or its low Ti content.

Determination of relative growth rates of natural quartz crystals

Although the theory describing crystal growth in the geological environment is well established, there are few quantitative studies that delimit the absolute time involved in the growth of natural crystals. The actual mechanisms responsible for the variation in size and shape of individual crystal faces are, in fact, not well understood. Here we describe a micro-infrared spectroscopic study of a single, gem-quality quartz crystal that allows us to measure the size, shape and relative growth rate of each of the crystal faces that are active throughout its growth history. We demonstrate that the abundances of hydrogen-bearing impurities can serve as ‘speedometers’ to monitor the growth rate of advancing crystal faces. Our technique can be applied to crystals from a variety of geological environments to determine their growth histories. Within the electronics industry, the technique might facilitate the production of defect-free synthetic crystals required for high-quality resonators and, ultimately, might allow determination of the absolute time involved in geological processes such as the crystallization of magmas, fluid flow in metamorphism and the sealing of open cracks in earthquake rupture zones.

Influence of hydroxyl on glass transformation kinetics in lithium disilicate melt and a re-evaluation of structural relaxation in NBS 710 and 711

Differential scanning calorimetry (DSC) experiments are presented that document the strong effect that ppm-level hydroxyl concentrations have on glass transformation kinetics in silicate melt. In particular, relaxation times decrease by an order of magnitude and limiting fictive temperatures (Tf∞) decrease ~20 K with an increase in hydroxyl content from 30 to ~1400 ppmw in the lithium disilicate system. Activation energies derived from Arrhenius plots of reciprocal Tf∞ values versus log(quench rate) show no dependence on hydroxyl content. Heat capacity scans of samples with various hydroxyl contents quenched at the same rate collapse to single curves when scans are plotted using (T−Tf∞). Such behavior can be incorporated most simply into the Tool–Narayanaswamy (TN) model of structural relaxation by allowing only the pre-exponential term in the relaxation function to be a function of hydroxyl content. Fits of our data to the TN model were obtained using three different activation energy values: (1) 757 kJ/mol (from cooling rate dependence of Tf∞); (2) 599 kJ/mol (from published viscosity data); and (3) 526 kJ/mol (4-parameter fit). The fit using the activation energy from viscosity measurements provided the best overall fit. In addition, structural relaxation phenomena of NBS 711 and NBS 710, lead silicate and soda-lime silicate compositions, respectively, were re-examined and a significant discrepancy in activation energies for NBS 710 and 711 between this study and earlier work was documented.