V41A-1350 0800h
Partial Molar Volumes for Lanthanide Sesquioxides in Sodium Silicate Melts
Lanthanides are of great interest in igneous petrology as trace indicators of magmatic processes that control the origin and evolution of igneous rocks. A key to the petrogenetic modelling of magmatic processes and to determine the phase diagrams of lanthanide host phases is the accurate determination of the physico-chemical and thermodynamic properties of lanthanide-containing materials, such as the volumetric properties of lanthanide-bearing silicate melts. Therefore, we have undertaken to provide a new reliable volumetric data set for lanthanide-bearing silicate melts which allows the available models in the literature to be extended to lanthanide-bearing melts. For this purpose, the densities of various lanthanide-bearing silicate melts distributed along various pseudo-binary joins, where the end-members are Na-disilicate and one of the lanthanide sesquioxides (i.e., Ce$_{2}$O$_{3}$, Pr$_{2}$O$_{3}$, Nd$_{2}$O$_{3}$, Sm$_{2}$O$_{3}$, Eu$_{2}$O$_{3}$, Gd$_{2}$O$_{3}$, Tb$_{2}$O$_{3}$, Dy$_{2}$O$_{3}$, Ho$_{2}$O$_{3}$, Er$_{2}$O$_{3}$, Tm$_{2}$O$_{3}$ and Yb$_{2}$O$_{3}$), have been measured using the double-bob Archimedean method. The present results show that the addition of any lanthanide to Na-disilicate leads to an increase in the melt density and that the melt density increases with increasing atomic number of the lanthanide. From the present density data set, the molar volumes of these melts have been calculated and the partial molar volumes of each lanthanide sesquioxide in these melts have been determined using a linear regression through each pseudo-binary join (i.e., Na-disilicate - lanthanide sesquioxide). This study indicates ideal behaviour with respect to the molar volume (i.e., a linear variation of the molar volume along each pseudo-binary join) for Na-silicate melts containing up to 10 mol% of lanthanide oxide. Comparison between the partial molar volumes of lanthanide sesquioxides obtained in this study and the molar volumes of molten lanthanide sesquioxides given in the literature raise the possibility however that this ideality is not maintained along the entire Na-disilicate - lanthanide sesquioxide pseudo-binary joins. Excess volumes of mixing appear to be required to describe the combined volumetric data set.
V41A-1351 0800h
A partial molar volume for ZnO in silicate melts.
Trace elements in igneous petrology have, in comparison with major elements, a relevance in the petrogenetic modelling of magmatic differentiation that far outweighs their relative abundance. Optimal use of the information contained in trace element variations within igneous phases requires an accurate description of their partitioning behaviour as a function of phase composition and structure, as well as temperature and pressure. In this manner, the partial molar thermodynamic properties of trace elements in silicate melts may contribute to the petrogenetic modelling of such systems. With this in mind, a series of investigations into the partial molar properties of trace elements in silicate melts have been carried out in recent years. Here we extend this work to the analysis of the volumetric properties of ZnO in silicate melts. Densities of 8 Zn-bearing silicate melts have been determined in air in the temperature range of 1363 to 1850 K. The compositional joins investigated (sodium disilicate (NS2) - ZnO; anorthite-diopside 1 bar eutectic (AnDi) - ZnO; and diopside - petedunnite) were chosen based on the pre-existing experimental density data set, their petrological relevance and to provide a test for significant compositionally induced variations in the structural role of ZnO. The ZnO concentrations investigated range up to 25 mol% for sodium disilicate, 20 mol% for the anorthite-diopside 1 atm eutectic and 100 mol% petedunnite. Molar volumes and expansivities of all melts have been derived. The molar volumes of the present liquids all decrease with increasing ZnO content. The partial molar volume of ZnO derived here from the volumetric measurements for each binary system is the same within error. A multicomponent fit to the volumetric data for all compositions yields a value of 14.141(0.730) cm$^{3}$.mol$^{-1}$ at 1300 K. We find, herewith, no volumetric evidence for compositionally-induced coordination number variations for ZnO in alkali-bearing versus alkali-free silicate melts.
V41A-1352 0800h
Temperature Independent Thermal Expansivities of Silicate Melts in the System Anorthite-Wollastonite-Gehlenite (CAS) system
Calcium aluminosilicate melts are, in addition to their model role in geochemistry, also important for both the glass and stone wool industry. Contributions to the PVT equation of state of such melts are needed for geochemical and geophysical modelling, as well as for providing tests of structure-property relationships for magma. The temperature-independent thermal expansivities of ten melts included in the anorthite-wollastonite-gehlenite (An-Wo-Geh) compatibility triangle were determined on glassy and liquid samples using a combination of calorimetry and dilatometry. The melts have either 0.5 or 1 non-bridging oxygens per tetrahedraly coordinated cations. The volumes at room temperature were derived from density measurements using the Archimedean buoyancy method. Each sample had a cooling-heating history of 10-10 K/min at 298K and a precise dimension of the samples allowed calculation of the density from the sample geometry at room temperature. The thermal expansion coefficient of the glass from 298K to the glass transition interval was measured by a dilatometer and the heat capacity was measured using a differential scanning calorimeter over a 298-1135K temperature range The thermal expansion coefficient and the heat flow was determined at a heating rate of 10 K/min on glasses which was previously cooled at 10 K/min. Supercooled liquid molar thermal expansivities were indirectly determined by combining differential scanning calorimetric and dilatometric measurements assuming that kinetics of enthalpy and shear relaxation are equivalent. This low-temperature combined determination of supercooled liquid density, molar volume and molar expansivities was tested against high-temperature data obtained by using the Lange-Carmichael (1987), Lange (1997) and Courtial-Dingwell (1999) model. The best linear fit provides a combination of data presented in this study and high temperature data calculated using the Courtial-Dingwell (CAS) model. This dilatometric/calorimetric method of liquid molar expansivity determination greatly increases the temperature range accessible for thermal expansion measurements. These results contrast strongly with those obtained for geologic multicomponent melts, as well as anorthite-diopside eutectic and diopside melts, which exhibit a clear temperature-dependence of expansivity. The temperature-dependence of the thermal expansivity of melts is herewith confirmed to be a sensitive function of composition. This leads us to speculate that its origins may indeed lie in the temperature-dependence of the coordination number of specific cations with temperature.
V41A-1353 0800h
In Situ Density Measurement of Basaltic Melts at High Pressure by X-ray Absorption Method
Density of silicate melt at high pressure is one of the most important properties to understand magma migration in the planetary interior. However, because of experimental difficulties, the density of magma at high pressure is poorly known. Katayama et al. (1996) recently developed a new in situ density measurement method for metallic melts, based on the density dependency of X-ray absorption in the sample. In this study, we tried to measure the density of basaltic melt by this absorption method. When X-ray is transmitted to the sample, the intensity of the transmitted X-ray beam (I) is expressed as follows; I=I$_{0}$exp(-$\mu$$\rho$t), where I$_{0}$ is the intensity of incident X-ray beam, $\mu$ is the mass absorption coefficient, $\rho$ is the density of the sample, and t is the thickness of the sample. If t and $\mu$ are known, we can determine the density of the sample by measuring I and I$_{0}$. This is the principle of the absorption method for density measurement. In this study, in order to determine t, we used a single crystalline diamond cylinder as a sample capsule, diamond is less compressive and less deformable so that even at high pressure t (thickness of the sample at the point x) is expressed as follows; t = 2*(R$_{0}$$^{2}$-x$^{2}$)$^{1/2}$, R$_{0}$ is the inner radius of cylinder at the ambient condition, and x is distance from a center of the capsule. And diamond also shows less absorption so that this make it possible to measure the density of silicate melt with smaller absorption coefficient than metallic melts. In order to know the $\mu$ of the sample, we measured both densities ($\rho$) and absorptions (I/I$_{0}$) for some glasses and crystals with same composition of the sample at the ambient condition, and calculated as fallows; $\mu$=ln(I/I$_{0}$)/$\rho$. Experiments were made at the beamline (BL22XU) of SPring-8. For generation of high pressure and high temperature, we used DIA-type cubic anvil apparatus (SMAP180) there. We used tungsten carbide anvils with the edge-length of 6 mm. The energy of monochromatic X-ray beam was 25 keV and the beam size was reduced to 0.1*0.1 mm$^{2}$ by two slits. Intensities of X-ray beam were measured by ion chambers. The starting material was a glass with the MORB composition (SiO$_{2}$-Al$_{2}$O$_{3}$-FeO-MgO-CaO-Na$_{2}$O). Experiments were made from 1 atm to 5GPa, from 300 to 1873 K. We measured the density of basaltic glass, crystals (eclogite) and melt. A density error of this method is less than 2 %. We calculated the bulk modulus of the glass at 773K, crystals at 1273K and melt at 1873 K, and obtained K$_{glass}$(773K)=46(4) GPa, K$_{crystals}$(1273K)=100(7) GPa, K$_{melt}$(1873K)=16.5(1.5) GPa assuming K$\prime$=4. This K$_{melt}$(1873K) value is consistent with the previous study by the sink-float method (Ohtani and Maeda (2001); K=18.7(2.1) GPa). We can conclude this method is applicable for silicate melts.
V41A-1354 0800h
Redox Viscosity of Iron Rich Silicate Melts - Martian Mantle Analogues.
The dependence of shear viscosity on the oxidation state of ferrosilicate melts has been measured using the concentric cylinder method and a gas mixing furnace. Two different simple Fe-bearing systems have been studied to date: (i) anorthite-diopside eutectic composition (AnDi) with variable amount of Fe (up to 20 wt%) as a basalt analogue and (ii) sodium disilicate (NS2 up to 30 wt % Fe). In addition, the compositional range has been extended to include the more complex SNC meteorite composition, a composition more relevant to Mars. The measurements were performed under air, CO2 and CO2-CO mixture at 1 atm and in a temperature range of 1300 to 1350 \°C. The experimental procedure involve a continuous measurement of viscosity during stepwise reduction state. The melt was reduced by flowing CO2 and then successively reducing mixtures of CO2-CO through the alumina muffle tube. Gas flow rates were electronically controlled using Tylan mass flow controllers and oxygen fugacity was directly measured using a sensor and calculated with Nernst equation. The composition and oxidation state of the melt was monitored by obtaining a melt sample after each redox equilibrium step. The melts were sampled by dipping an alumina rod into the sample and drawing out a drop of liquid, which was then plunged into water for quenching. The resulting glasses were analyzed by electron microprobe, and the volumetric potassium dichromate titration were employed to determine FeO. In addition, the redox dependence of viscosity of our samples have been compared with data from literature (Mysen et al. 1985, Dingwell and Virgo, 1988; Dingwell 1989, Dingwell 1991). The viscosity of all melts investigated herein decreases with melt reduction. The viscosity decrease is, in general, a nonlinear function of oxidation state expressed as Fe2+/Fetot and can be fitted using logarithmic equation.
V41A-1355 0800h
Viscosity Measurement of a Hydrous MORB Melt at High Pressure
Viscosity of silicate melt is a fundamental property for understanding the chemical differentiation of the Earth. The change of viscosity with temperature and pressure reflects the change of the structure of the melt. Therefore, it is important to measure the viscosity of silicate melt at high pressure. The viscosity of many silicate melts has been investigated to date, such as the basalt melt (Fujii and Kushiro, 1977; Ando et al., 2003). The influence of volatile component on viscosity of silicate melt is a matter of debate. Therefore, we measured the viscosity of a hydrous MORB melt at high pressure. The viscosity was measured by the in situ falling sphere viscometry using X-ray radiography at SPring-8 synchrotron facility (BL04B1) in Japan. We used the synthetic glass powder of MORB plus gibbsite of which composition was simplified to a 7-component system, SiO2-Al2O3-FeO-MgO-CaO-Na2O-H2O. Water content was 8 wt%. Experiments were carried out in the temperature range of 1463 to 1752 K, and in the pressure range of 0.7 to 5.1 GPa. It is turned out that the viscosity of a hydrous MORB melt has a weaker temperature dependency than the olivine tholeiite melt (Fujii and Kushiro, 1977) at about 0.7GPa. Generally, the more polymerized melt has a larger temperature dependency of viscosity. This result suggests that water depolymerizes the structure of the silicate melt. This is consistent with the viscosity of the melt in the albite-water system (Dingwell, 1987).The viscosity of the hydrous MORB melt is smaller than that of anhydrous MORB (Ando et al., 2003) at the same pressure and the temperature, due to incorporation of water in the silicate melt structure. The effect of water is less remarkable than that in the albite-water system (Dingwell, 1987) because the structure of hydrous MORB melt is depolymerized. The viscosity of anhydrous MORB decreases with pressure up to ca. 2.5 GPa, but it increases with pressure at above 2.5 GPa (Ando et al., 2003). The viscosity of the hydrous MORB melt, on the other hand, increase with increasing pressure like other depolymerized melt, such as diopside (Reid et al., 2003).
V41A-1356 0800h
Studies of Diffusion, Atomic Hopping Frequency and Site Residence Times in Molten SiO2 by Molecular Dynamics
Computer modelling of silicate melts enables the study of pressure-temperature conditions not easily obtainable by traditional experimentation (e.g. 1). Diffusion in melts under various conditions is critical to our understanding of a variety of processes such as melt crystallisation, magma mixing and the behaviour of trace elements during magma ascent that underpins the field of igneous petrogenesis. Understanding of diffusion mechanisms and activation energies also provides information on changes in melt structure. In the present paper, the diffusion of silicon and oxygen in molten silica has been investigated by molecular dynamics using a modified BKS potential (2). A range of melt temperatures and pressures was studied with a view to understanding the relationship between temperature, pressure, diffusion and melt structure. At each P-T point studied, the system was equilibrated for between 1 million and 40 million steps of 1fs depending on the conditions, with data being collected over the same time range. The potential was adjusted to overcome problems with instability in the particle velocities at high temperature. The simulations were run at the Oxford University Supercomputing centre, UK. Systems of 144, 288, 576 and 1152 particles were investigated. In addition, two different sets of periodic boundary conditions were used - cubic and truncated octahedral. The latter was found to provide a better ratio of simulated time to compute time. We have previously reported a pronounced non-linearity in the temperature dependence of diffusion in molten SiO$_{2}$ (3). This suggests at least two diffusion mechanisms with differing activation energies which operate to differing extents at lower and higher temperatures. It is this relationship between diffusion coefficients and diffusion mechanism/melt structure that we have investigated in the present paper. In addition, we have extended the study to include a wider range of pressures. Detailed examination of trajectory data, as well as radial distribution function (rdf) and density data provides a picture of the structure and dynamics of SiO$_{2}$ over a range of conditions. We have developed a method for analysing residence times and 'hop' distances under varying conditions. The plots found at the URL accompanying this abstract compare diffusion of an oxygen atom at 3000K and 4000K over a 10$^{6}$ step (1ns) run. Peaks represent 'hopping', troughs residence in 'sites'. At temperatures of 4000K and above the concept of a discreet 'rattle and hop' diffusion mechanism breaks down to be replaced with a plasma-style situation of more continuous random movement. (1) Fraser DG, Cagin T, Demiralp E, Goddard WA, III, "New transferable interatomic potentials for simulating melting of Mg silicates near the base of mantle," A.G.U. 1998. (2) Van Beest BWH, Kramer GJ, Van Santen RA (1990) Force fields for silicas and aluminophosphates based on ab-initio calculations. Phys Rev Lett 64: 1995. (3) Gemmell AL, Refson K, Fraser DG. Molecular Dynamics Simulations of Diffusion in a Silica Melt. EOS Trans AGU 84(46), Fall Meet. Suppl., Abstract V11D-0528, 2003.
http://www.earth.ox.ac.uk/~alastair
V41A-1357 0800h
Halogen Chemical Diffusivities in Silicate Melts
Halogens may exert a significant influence on the physico-chemical properties and the structure of silicate glasses and melts, as well as on their phase relations. Furthermore, the geochemistry of halogens from volcanic systems potentially provides valuable information on the nature and efficiency of the dagssing process in subduction zone volcanism. Knowledge of the transport properties of halogens in silicate melts is a necessary prerequisite in order to model the information contained in halogen concentrations of eruptive products and volcanic gases in terms of the potential influence of kinetics in controlling degassing. Towards this end, chemical diffusion of halogens (fluorine, iodine, chlorine and bromine) has been invesigated in melts in the system Na-Fe-Si-O-(F,Cl,Br,I) over a wide range of temperature (450 - 1400°C) using diffusion couple techniques. Halogens were added in the form of FeF3, FeCl3, FeI2 or FeBr3. Starting melts were fined by stirring for several hours at 1000-1100°C using a concentric cylinder viscometer. The synthesis temperature was restricted to 1100°C to limit the volatilization of halogens. Melted and doubly polished discs were then put into platinum tubes (5mm diameter), where the halogen-rich sample was located at the bottom, and sealed by welding. During the experiments the temperature was monitored with a thermocouple located at the vicinity of the capsule. Run durations were between 30 minutes and 1 hour. The recovered samples were analyzed using an electron microprobe in order to determine the diffusion profiles of the halogens. The experiments for the I-containing samples were conducted between 450 and 1025 °C and for a run duration of 30 to 45 min, for the Cl-containing samples between 800 and 1100°C and 45 to 60 min, while the Br-containing materials were investigated between 750 and 1000°C for 45 to 60 min. The preliminary results suggest a significant range of at least 3 orders of magnitude between the diffusion coefficients for F, Cl, Br and I at 1000°C. This raises the possibility of significant kinetic disequilibrium during foaming and rapid degassing of magma prior to and during eruption. The range also implies that the diffusion of halogens under these conditions in magma is intrinsic in nature and not controlled by melt viscosity.
V41A-1358 0800h
The Melting Curve of Carbon Dioxide and Implications for the Fluid Equation-of-State
Despite being a dominant volatile in magma, the thermodynamic properties of carbon dioxide remain problematic from mid-crustal to mantle conditions of pressure and temperature. Several published equations of state are neither in agreement nor do they always match existing high-pressure data. In order to investigate the properties of fluid carbon dioxide, the melting curve of supercritical carbon dioxide has been measured up to 5.5 GPa. When coupled with a robust thermodynamic model for solid carbon dioxide, these data allow a test of various proposed equations-of-state for the fluid. In order to match the melting curve, both the densities predicted by the Pitzer and Sterner (1994) equation-of-state and those predicted by Span and Wagner (1996) must be decreased by a few percent. This finding is in accord with shock data at higher pressures ($>$10 GPa) which also suggest the necessity of lower fluid densities. The equation-of-state of Frost and Wood (1997), based on measurements of graphite/CO2 equilibria, gives substantially higher densities and is inconsistent with the shock data. Along the cooler geotherms associated with older slabs the solid phase of carbon dioxide is stable into the lower mantle.
V41A-1359 0800h
Bubble nucleation in highly viscous silicate melts during instantaneous decompression from high pressure
Vulcanian eruptions and dome collapses involve disruption of cooled, degassed, and thus highly viscous, magma. Theoretical models (e.g., Toramaru, 1995) argue that in highly viscous melts, in which bubble growth is negligible, bubble nucleation should result in exponentially increasing number densities of bubbles as viscosity ($\eta$) increases. To investigate how bubbles nucleate in highly viscous melts, we performed experiments in which water-saturated, high-silica rhyolitic melts were decompressed suddenly from high pressure, held at lower pressure for various amounts of time, and then quenched rapidly. The decompressions were carried out at 500 to $700\deg$ C to achieve viscosities up to $>$10$^{8}$ Pa s. Most bubbles nucleated on faces of Fe-Ti oxide crystals, and so nucleation is heterogeneous. We found that when $\eta$ $<$ 10$^{6}$ Pa s, nucleation occurs in less than 10 seconds, producing up to $>$10$^{8}$ cm$^{-3}$ of bubbles. At similar super-saturations, nucleation takes between 10 and 60 seconds to occur when $\eta$ increases to $\sim$10$^{7}$ Pa s, but produces similar number densities of bubbles as seen at lower viscosity. The delay in nucleation increases dramatically with higher viscosity, with nucleation not occurring after 360 seconds when $\eta$ = 10$^{7.5}$ Pa s. Our preliminary results thus show that viscosity suppresses the rate of nucleation, but so far we have not found that number densities increase significantly. Continued experiments will explore how long nucleation can be suppressed at relatively high viscosities, and whether number densities exponentially rise when nucleation finally occurs. The long delay in nucleation may suggest that most disruption that occurs in eruptions of cooled or degassed magma results from expansion of pre-existing bubbles. Toramaru A., JGR, v 100, p 1913-1931, 1995.
V41A-1360 0800h
Quantifying the distribution of bubble sizes in volcanic rocks: Generalized statistical formulation and application to vesicular lavas
We have developed a generalized analytical and computational approach to analyze the full spectrum of observed bubble size distributions in volcanic rocks. All natural bubble populations can be characterized within the logarithmic family of statistical distributions. Populations are typically described with best fit functions. We have generalized this approach to employ complimentary use of distribution density and exceedance forms for probability. This results in very accurate estimates of distribution moments and bubble number density for each mode in any distribution. We demonstrate that transformation linearization of logarithmic distributions can be accomplished by changing the scale of measurement units. This enables the investigator to visualize distribution characteristics and thus eliminates the problem of observational bias resulting from limitations in instrument resolution, sample size, etc. In an initial application of this analytical technique, we have studied bubble populations in volcanic rocks from a collection of vesicular basalts from the Colorado Plateau. The samples reveal a variety of mono- and polymodal distributions within the logarithmic family of statistical functions. Within this family, most populations have a bimodal log-normal distribution, but the others are represented by mono or bimodal log logistic, and Weibull distributions. We present eleven bubble population types based on distribution function, mode location and intensity. Trends within some of these population types can be interpreted as evolution of vesiculation processes. Larger bubble size modes can result from coalescence of smaller bubbles within the distribution. In the course of coalescence larger modes grow at the expense of smaller modes. Another type of trend leads to monomodal Weibull (or exponential) distributions as a result of superposition of multiple log normal distributions if modes are sufficiently close in size and intensity.
V41A-1361 0800h
A Hand-made Gas Permeameter for Permeability Measurement of Small Samples of Natural and Experimental Volcanic Materials.
Gas permeability in vesiculating magma, in which connected bubble network is developing, is an essential physical property controlling behavior of volcanic eruptions, since the gas permeability varies drastically in vesiculating processes during magma ascent. Although there are several studies on gas permeability of vesiculating magma, they have been limited in permeability measurements of natural samples, and their numerical simulations. For further understanding of gas permeability development in vesiculating magma, the permeability measurement on experimental products produced by vesiculating experiments is an effective approach. However, since the size of experimental run products is generally from 1 mm to 1 cm scale, they are too small to be measured by using commercial gas permeameter. In this study, we constructed a hand-made gas permeameter to measure permeability of small samples such as experimental run products. The hand-made permeameter can measure permeability in the wide range from 10$^{-17}$ to 10$^{-10}$ m $^{2}$ within the precision of one order for mm scale samples. Nitrogen gas is used as a working gas in this measurement system. The permeability is calculated by steady gas flow rate at fixed pressure difference up to 15000 Pa (ca. 0.15 atm). The pressure difference is measured with accuracy of 10 Pa by a water column manometer. Gas flow rate is converted to water flow rate in an acrylic container and the water flow seeping from the tube into a beaker is monitored by an electric balance. We confirmed the accuracy in permeability values by measuring gas flow in stainless capillary tube (15 mm in length and 100 mm in inner diameter). We carried out flow measurement at 1.8\times10$^{2}$-1.4\times10$^{4}$ Pa in pressure difference and 3.0\times10$^{-10}$-3.6\times10$^{-8}$ m$^{3}$/s in flow rate. For this flow rate, Reynolds number of the gas flow is estimated to be 10$^{-2}$-10$^{0}$. Therefore, the gas flow can be assumed to be Poiseuille flow. Although the difference between the measured and calculated flow rates increases with decreasing flow rate, the discrepancy is about 40 % at the maximum for the flow rate more than 10$^{-10}$ m$^{3}$/s. Therefore, measurement of the pressure difference and gas flow rate in this measurement system is precise enough to determine the permeability within 0.4 log unit. Using this measurement system, permeabilities of four air-fall pumice and scoria were measured. The results are in good agreement with a trend obtained from permeability measurement of pyroclastic materials by Klug and Cashman (1996). This consistency also supports the validity of this permeameter. This permeability measurement system can be constructed easily at a very low cost, and is expected to be a useful tool to measure permeability of small volcanic materials and experimental run products.
V41A-1362 0800h
A High-load, High-temperature Deformation Apparatus For Volcanological Studies
The need for an adequate understanding of the nature and extent of physico-chemical processes involved in explosive volcanism is considerable. In recent years much effort has been concentrated on rhyolitic melts under conditions relevant to explosive volcanism. Especially the description of single-phase rhyolitic melt properties such as the temperature, pressure and compositional dependence of viscosity has been greatly improved. Yet, modelling of the emplacement and eruption of silicic domes is still hampered by the lack of a sufficiently accurate rheological database for multi-phase lavas with crystals and vesicles and no simple expression can be used to describe their rheology. We have developed a unique high-load, high-temperature deformation apparatus for studying in situ the non-Newtonian flow behaviour of magmas. The apparatus accommodates samples that are up to 100 mm in diameter and 100 mm long, and can be used to run constant displacement rate and constant load experiments. The rig is ideal for volcanological studies because it uses experimental conditions that closely match those found in volcanic processes: temperature (25 to $1300 \deg$), stress (0 to $>$ 500 MPa), strain rates (10-6 to 10-2 s), and total strain (0 to 100%). The apparatus still has to be optimised, but we can already present some preliminary results. To study the flow behaviour of a reference melt, we performed a viscosity study at a constant temperature on a "NIST 710a" soda-lime composition. The sample was placed between the two pistons of the apparatus and heated up to the desired temperature ($609 \deg$) above the calorimetric glass transition temperature ($550 \deg$). After allowing the system to reach thermal equilibrium (6 hours) in a parallel plate type experiment, load was applied (10, 50, 100, 200, 250 kN), holding that load for several 10 seconds. For applied loads $<$ 100 kN (which correspond to stresses $<$ 80 MPa), the stress versus strain rate relationship always behaves linearly and the viscosity remains constant with time (Newtonian flow regime). If the applied load was $>$ 200 kN (which corresponds to stresses $>$ 160 MPa), the stress versus strain rate relationship curves and the apparent viscosity decrease with time (non-Newtonian flow regime). If the applied load is raised to 250 kN "hot" cracks are produced and the sample is partly fragmented. A major advance is, that we can use sample sizes in the range of several 10$^{5}$ mm$^{3}$, which means we are able to measure natural samples that contain large phenocrysts in cm size, like the Unzen conduit material. Finally, with this facility, it will be possible to measure reliable temperature distributions due to viscous heating in situ during the deformation process, using several thermocouples inside the sample. Experiments on obsidian are underway.
V41A-1363 0800h
Rheology and timescales of welding
We describe results from 15 high-temperature, constant strain rate and constant load deformation experiments on natural pyroclastic materials that simulate welding. Experiments were run on unconfined samples at temperatures between 835$\deg$ and 900$\deg$C. Samples comprised 4.3 cm diameter, $\sim$6 cm length cores of sintered Rattlesnake Tuff rhyolite ash. Porosity of starting materials is $\sim$78$%$. The experiments used uniaxial load stresses of 0.2 to 5 MPa which corresponds to overburden depths of $<$ 200 m in ignimbrite deposits. The experimental results track strain (porosity loss) and strain rate as a function of time at fixed conditions (load and temperature). Our results show that deformation of pyroclastic material has a strain dependent rheology. The effective viscosity ($\eta_{e}$) of the samples increases during the experiment as strain acccumulates and porosity ($\phi$) is reduced. We describe this behaviour using the relationship: (1) log $\eta_{e}$ = log $\eta_{o}$ - $\alpha$ [$\phi$/(1-$\phi$)]. where effective viscosity is related to the viscosity of the framework material (melt), the sample porosity, and a fit-parameter for the material ($\alpha$). Our experimental work suggests a value of 0.63 for compaction of natural pyroclastic materials. Equation 1 is the basis for an empirical equation that describes the total strain during viscous compaction as a function of original porosity ($\phi_{o}$), the viscosity of framework melt ($\eta_{o}$),load ($\sigma$) and ${\bf time}: (2) $\epsilon$ = $\phi_{o}$ + (1-$\phi_{o}$)/$\alpha$ $\times$ ln [($\alpha \sigma \Delta$t)/($\eta_{o}$ (1-$\phi_{o} ) + exp$[-($\alpha \phi_{o}$)/(1 - $\phi_{o} ) ] ]. In this relationship, the values of $\phi_{o}$ and $\eta_{o}$ are physical properties of the specific deposit and load relates to the thickness of the deposit and the position (depth) of the sample. Eq. 2 can be used to predict $\epsilon$ {\it vs.} time paths to compare against the original experimental data and to model natural deposits. By rearranging the above equation to isolate time ($\Delta$t) we predict the times required for strain accumulation (reduced $\phi$) during welding of natural pyroclastic deposits. We show that the timescales of welding for even moderate emplacement temperatures, relative to glass transition temperatures, can be very short (i.e., days) and within an order of magnitude of the timescales of deposition or assembly of large ignimbrite sheets.
V41A-1364 0800h
Flow Banding in Volcanic Rocks: A Record of Multiplicative Magma Deformation
Banding in obsidian from Big Glass Mountain (BGM), Medicine Lake volcano, California and Mayor Island (MI), New Zealand provide a record with a 1/wavenumber power-spectral density and multifractal characteristics. The samples are compositionally homogeneous, with banding defined by variable microlite content (BGM) or vesicularity (MI). In both samples banding formation is well explained by continuous deformational reworking of magma and a concurrent change in crystallinity (vesicularity) that is a small random multiple of the total amount already present. Banding formation therefore represents a multiplicative process. We complement our spectral and multifractal analysis with several null-hypothesis test and propose repeated brittle deformation, concurrent development of textural heterogeneity (microlite content or vesicularity), reannealing, and viscous deformation as a viable process for the formation of flow banding in these samples. A brittle deformational component in a simple flow geometry, for example during magma ascent in the volcanic conduit, provides a suitable mechanism for (1) the spatial redistribution of textural heterogeneity over a broad range of length scales; (2) enhanced open-system magma degassing via a temporary network of highly permeable cracks and fractures; and (3) the development of spatially variable microlite content (vesicularity) through variable degassing of magma located at different proximities to cracks and fractures.
V41A-1365 0800h
Petrologic and Dynamic Importance of Flow Banding in Obsidian Lavas
One of the intriguing characteristics of effusive obsidians is the abundance of flow banding, or micrometer to centimeter-scale variations in microlite concentration. As these features arise from degassing, crystallization, and deformation processes, flow bands must contain important information regarding the chemical and physical evolution of obsidian magmas. Relatively little is known about the origin of this feature, and information on the relative rheologic properties of microlite-rich and poor bands is currently unavailable. In this paper, we present: 1) textural measurements on microlitic flow bands, 2) H2O concentrations, and 3) calorimetric measurements on flow bands of variable microlite content from several late Holocene obsidian flows. The goals are to better understand the mechanism of flow band formation and how these bands affect flow rheology and emplacement dynamics. Flow banded obsidians from Obsidian Dome (OD), Big Glass Mountain (BGM), and Big Obsidian Flow (BOF), are the focus of this study. Petrographic analysis shows that all obsidians contain microlites of pyroxene, feldspar, and oxide. However, the relative abundances of these phases vary dramatically within particular samples and between analyzed suites. Flow bands are therefore classified as 1) modal, wherein adjacent bands have the same mineral assemblage but contain different volume fractions, size distributions, and/or number densities of constituent phases, or 2) mineralogic, wherein adjacent bands differ by virtue of their constituent mineral assemblages. Banding in obsidians from both OD and BOF is dominantly modal, although rare bands display mineralogic differences defined by the presence or absence of plagioclase microlites. BGM obsidians tend to be modal in character, containing pyroxene microlites whose size and number densities vary across bands. Crystal size distributions measured on BGM obsidians reveal significant differences in the size and shape of microlite populations between adjacent flow bands. In general, the microlite size is largest in microlite-poor bands. Calorimetric and H2O content measurements of microlite-rich and poor bands are currently underway. Preliminary results indicate that texturally-distinct bands underwent similar cooling and degassing histories just prior to quenching during dome emplacement. Textural differences, may therefore originate in the conduit prior to extrusion.
V41A-1366 0800h
``Sedimentary'' Structures in a Layered Granodiorite: A Window Into Physical Conditions Present During the Development of the Tuolumne Intrusive Suite, Sierra Nevada California
The presence of sedimentary-type features exposed in layered Tuolumne Intrusive Suite (TIS) outcrops provides a unique opportunity to assess fluid dynamic processes associated with both pluton assembly and magma chamber processes. Previous analyses of layered mafic intrusions, experimental and mathematical modeling of igneous layering, analogs in sedimentary rocks, and layering in granitic intrusions, have not defined a set of physical conditions that exist during intermediate granitic pluton development. Using constraints provided by field, petrographic, and geochemical relations among the layered outcrops in the TIS field area, both the dynamic conditions and rheological state necessary to form these "sedimentary" structures can be identified. Similar layer orientations at many different outcrops and sharp upper and lower layer/host rock contacts suggest that the layers occupy shallowly dipping, near-planar, "slot-like" openings in the outer members of the TIS. The TIS layers appear to originate by a combination of flow sorting and crystal settling due to the modal grading of the cm-scale layers, cumulate mineral textures, and mafic glomercrysts. Groups of layers with similar features demonstrate the repetition of this process. Examination of reoriented hornblende crystals along fault surfaces shows that slumps form due to the addition of overlying layers when the melt fraction is between 20 and 50%. Given the presence of (1) dikes full of K-feldspar megacrysts that cross-cut the layers but not the host rock, (2) resorbed crystal edges, (3) zoned hornblende crystals with pyroxene cores and (4) adjacent groups of layers that have features different from one another; the layering conditions such as magmatic temperature, water content, and direction of layer deposition must change periodically.
http://www.ess.washington.edu/Program/physpet/personal_page-sierra.html
V41A-1367 0800h
Pressure Dependence of Viscosity of Hydrous Rhyolitic Melts
Knowledge of viscosity of silicate melts is critical to the understanding of igneous processes, such as bubble growth, magma fragmentation, etc. Numerous measurements have been made on the viscosities of natural and synthetic silicate melts, but only a few investigations have been carried out on the pressure dependence of viscosity of silicate melts. A new method using the cooling-rate method based on the interconversion of two water species in silicate melts (Zhang et al., 2003) is applied to investigate the pressure dependence of viscosity of hydrous rhyolitic melt at near glass-transition temperatures. All the samples used so far are natural obsidian glass with about 0.8 % H$_{2}$O by weight. The experiments were at 3 GPa and 1 GPa. A prerequisite for viscosity estimation using the method of Zhang et al. (2003) is to know the temperature dependence of the equilibrium constant K of the interconversion reaction at a given pressure. Equilibrium experiments were verified by reversal experiments. Equilibrium constant changes with pressure nonlinearly. At 600 \deg C, ln(K) is almost constant from one bar (-1.69) to 1 GPa (-1.72), but increases from one bar to 3 GPa (-1.56). Cooling rates varied from 100 \deg C/s to 0.1 \deg C/s in the cooling-rate experiments. Viscosity (in Paùs) at the apparent equilibrium temperature of the hydrous species reaction (i.e., glass transition temperature) is obtained as 10$^{11.45}$/q where q is cooling rate in K/s. So the range of viscosity inferred from this method can be 3 orders of magnitude. Data show that viscosity of hydrous rhyolitic melt increases with pressure nonlinearly. At 600 \deg C, the viscosity increases by 0.3 log units from 1 bar to 1 GPa, and by 1.5 log units from one bar to 3 GPa. It appears that in terms of the pressure dependence of viscosity, rhyolite with 0.8 wt% water at low temperature range behaves as a depolymerized melt such as diopside melt. The equilibrium and cooling rate data suggest a structure change occurring between 1 GPa to 3 GPa, leading to a sudden change in the equilibrium constant and viscosity.