MR21C-01
Thermal conductivity of MgO at lower mantle conditions from ab initio calculations.
Geodynamical models indicate that thermal conductivity plays a significant role in determining the structure and dynamics of the Earth's mantle. In addition, because the thermal conductivity of the mantle regulates the heat flux from the core, it also effects magnetic field generation. In spite of this, the thermal conductivity of mantle minerals at the relevant pressures is unknown, due to difficulties in making measurements at such extreme conditions. In view of this, we have calculated the lattice thermal conductivity of MgO using a reverse perturbation steady-state non-equilibrium molecular dynamics method, with the forces driving the dynamics being computed within density functional theory. In the first case, calculations were carried out at temperatures of 1000 K and 2000 K at ambient pressures, in order to allow compare with experimental results. Following this values were calculated at 2000 K and 4000 K at 136 GPa. Calculations for MgSiO3 perovskite are ongoing. The method will be discussed and results presented.
MR21C-02 INVITED
The Effects of Crystal-plastic Deformation on Zircon Geochemical Systems
The widespread usage of trace element, radiogenic isotope and stable isotope geochemistry of zircon (ZrSiO4) for constraining the nature and timing of Earth processes throughout geological history is underpinned by a thorough understanding of the mechanisms by which zircon geochemistry may be modified. One potentially important but generally overlooked mechanism, is the modification of zircon diffusivity via the formation of deformation-related fast-diffusion pathways throughout the mineral lattice. Electron backscatter diffraction (EBSD) allows quantitative resolution of intra- and intergrain variations in crystallographic orientation, and has facilitated the recognition of low-angle subgrain boundaries and lattice misorientation microstructure in zircon. Dislocation slip systems have been determined for zircon via analysis of EBSD data using simple geometric models for low-angle boundary formation. Assessment of the mean local misorientation has allowed direct comparison of the deformation microstructure with trace element and/or isotopic analyses for zircon via lower spatial resolution ion probe (SHRIMP) techniques. Analysis of several naturally deformed zircons from different environments indicates localized deformation- related geochemical modification with a variety of enrichment/depletion relationships. The most extreme changes occur along low-angle boundaries, with a strong boundary misorientation angle control. In all examples given, the deformation-related geochemical modification of zircon requires trace element mobility via fast diffusion pathways at much higher rates than predicted by known volume diffusion parameters. The exact nature of geochemical modification in zircon depends on many factors, such as trace element compatibility in the zircon lattice, exchange with an external reservoir, and/or possible trace element stabilization of low-angle boundaries, P-T conditions of deformation, and, for the U-Pb system, the timing of deformation relative to initial crystallisation. These results provide some new constraints on geological processes, for example allowing U-Pb systematics to be directly linked to the timing of impact deformation. Therefore, the analysis of deformed zircon provides a new approach to unlocking zircon as a geological tool. However, quantification of the rates and exact mechanisms of deformation-related fast-pathway diffusion in zircon remains a challenge for future research.
MR21C-03 INVITED
Wet, dry, Dopants and Defects - an Integrated View of Diffusion in Olivine as a Prototypical Silicate
Knowledge of diffusion coefficients of different species in minerals is a necessary pre-requisite for understanding and modeling the compositional evolution of rocks. Their use to determine time scales of various geological processes, to understand the reaction mechanisms and rheological behavior of rocks and minerals, and to evaluate the significance of dates and temperatures obtained from geochronometers and geothermometers, is becoming routine. However, many aspects of diffusion behavior of minerals remain enigmatic. Two of these are: (1) Do the presence of small amounts of impurities ("dopants") affect diffusion rates, and if yes, how? (2) Water is known to enhance transport rates, but exactly at what conditions and how does the wet to dry transition occur? Can it be predicted? These questions can be approached by developing an appropriate description of the diffusion process in terms of point defect thermodynamics. The large body of systematic diffusion data and natural observations in olivine makes it the mineral of choice to explore this avenue. Diffusion coefficients for various cations as well as O are known typically as a function of temperature, and often as a function of other variables such as pressure, oxygen fugacity or water fugacity. We developed a quantitative point defect model that allows the role of trace elements (i.e. dopants in the terminology of material science) in controlling transport properties to be quantified [1]. In addition we explored the nature of wet to dry transition of these transport properties in olivine [2]. We have now combined these approaches to develop an integrated scheme for describing point defect chemistry and transport properties of olivine containing arbitrary trace elements, in the presence or absence of water. The approach is perfectly general and can be extended to any other mineral provided enough data are available. Oxygen fugacity is known to affect transport rates of Fe-bearing silicates but the extent of this influence depends on T and fO2. At low T (e.g., < 1000 °C) a transition to a new diffusion mechanism, independent of fO2, occurs in impurity bearing natural olivines due to the presence of Al3+ and Cr3+ (Dohmen and Chakraborty, 2007). An extension of this model allows us to treat the role of water and the wet to dry transition in an exactly analogous, and self consistent, manner. The dependence of diffusion rates on fO2, fH2O and concentrations of various trace elements can be obtained without needing to empirically determine the critical fH2O at which the "wet" to "dry" transition occurs for each kinetic / transport property. We find that the transition occurs over a significant range of pressure and over this interval the kinetic behavior is transitional between wet and dry. We also find that although dopants play a key role in defining the transport properties of various semiconductors, in natural materials such as olivine their influence is limited because simultaneous incorporation of multiple trace elements of different ionic charges largely offset the effects of each other. We are left with Fe3+ (via oxygen fugacity) and H as the two key "impurities" that can commonly and significantly affect transport rates in olivines; their influence can now be calculated and predicted quantitatively. [1] Dohmen R, Chakraborty S (2007) Phys Chem Mineral, 34, 409-430. [2] Costa F, Chakraborty S (2008) Phys Earth Planet Int, 166, 11-29.
MR21C-04
Fe-Mg Interdiffusion Coefficients in Clinopyroxene: Experimental Determinations Using Nanoscale thin Films
Temperature dependent partitioning of Fe and Mg between various ferromagnesian silicates and melts constitutes the basis of the most common and well calibrated geothermometers. Of these, the geothermometers involving clinopyroxene (e.g. Cpx with Grt, Opx or melt) are some of the most sensitive and widely used. However, widespread compositional zoning found in clinopyroxenes implies diffusion rates are sluggish and for proper application of geothermometers as well as geospeedometers to meteorites, ultramafic mantle rocks, various volcanic and plutonic igneous rocks and metamorphic rocks from the granulite facies, diffusion coefficients need to be known. We report here preliminary Fe-Mg interdiffusion coefficients in clinopyroxene (Cpx) (diopside rich crystals of composition Di95He5). Diffusion couples were prepared by depositing a thin film (~ 20-100 nm) of olivine source material (Fo30) by pulsed laser deposition (PLD) onto an oriented, polished and pre- heated surface of a diopside crystal under vacuum. Samples were annealed for 10 to 400 hours at 800 to 1000°C, 1 atmosphere total pressure under a controlled oxygen fugacity of about 10-16 bar. Film thickness and compositional profiles were measured using Rutherford backscattering Spectroscopy (RBS) on reference and annealed samples. Concentration-depth profiles of Fe (up to 300 nm) were extracted from the RBS spectra and fitted numerically. At the lowest temperature studied by us (800°C), Fe-Mg interdiffusion coefficient in this clinopyroxene is found to be 2.2 x 10-22 m2/s (log D = -21.66). The diffusion coefficients found in this study are consistently about a half an order of magnitude slower than the corresponding rates found in orthopyroxenes (ter Heege et al., Fall AGU 2007). Taken together with data from the literature, this indicates that DFe-Mg decreases in the order olivine > garnet > orthopyroxene > clinopyroxene. Consequently, the closure of various geothermometers involving clinopyroxene will be controlled by diffusion in this phase.
MR21C-05 INVITED
Grain boundaries as rapid diffusion paths for incompatible elements in the Earth's mantle
Grain boundary segregation (GB) in polycrystalline materials, that is, the enrichment of certain elements at GBs, is a common phenomenon in any polycrystalline material. Our investigation of structure and chemistry of GBs in geo-materials has revealed that: (i) incompatible elements are highly concentrated at the GBs, which are from natural and synthetic aggregates [1-4], (ii) the concentration of the elements at the GBs is correlated with the elements concentrations in grain matrices (= crystal lattice), showing the achievement of equilibrium elemental partitioning between the two regions [3,4], (iii) those partition coefficients are comparable to the coefficient between crystal and melt [4], (iv) chemical and structural analyses indicate that the width of the GBs are approximated as 2 mono-atomic layers (= 0.7-0.8 nm) [4]. We can predict the effect of segregation on GB diffusion based on the partition coefficient, βi (= element concentration of i at GB/grain matrix). The diffusivity of element i through bulk, fluid-free polycrystalline material, Di(bulk) can be expressed in terms of the diffusivities of i in the grain matrix, Di(GM) and along the GB, Di(GB), as Di(bulk)=Di(GM)+ βi*(3w/d)*Di(GB) (where d and w are grain diameter and width of GB, respectively). The equation states that the concentration of i at a GB is always larger by a factor of βi than the concentration in the grain matrix at local equilibrium between the two regions. If Di(GB)/Di(GM) = 104, GB diffusion and lattice diffusion are expected to contribute equally to the transport of element i through a rock when βi*(3w/d)*104 = 1. Accordingly, if the range of βi*(3w/d)* 104 > 1 for a particular ion, then GB diffusion should be the dominant mechanism of mass transport through a rock. To confirm this prediction, the relationship between Di(GB) and βi should be investigated. We will show our experimental results of the diffusion on submicron mineral aggregates synthesized by our newly developed technique. Ref: [1] Hiraga et al. (2002) CMP 144:163; [2] Hiraga et al. (2003) Am Min 88:1015; [3] Hiraga et al. (2004) Nature 427:699; [4] Hiraga T, Kohlstedt DL (2007) GCA 71:12661
MR21C-06
Silica Transport and Cementation in Quartz Aggregates
Silica transport and cementation in quartz aggregates have been experimentally investigated. Starting materials include a natural quartz arenite (Pocono sandstone), sized clasts of synthetic quartz, and sized grains of disaggregated natural sandstones. Experimental charges consisted of amorphous silica powder (~25 mg), AlCl3 powder (~3 mg), 25 wt% NaCl brine solution (~20 mg), and the starting material (~150 mg). The charges were weld-sealed in gold capsules and run in cold-seal pressure vessels at 300°C to 600°C at 150 MPa confining pressure for up to 4 weeks. Detailed calibrations of the furnaces indicate the maximum temperature variation across the length of the sample charges (3-7mm) was <5°C, and typically <3°C. After the experiments, samples were vacuum impregnated with epoxy containing a blue dye and sawn in half along the long axis of the sample charge. The nature and amount of silica transport and cementation in the samples was determined by a combination of Cathodoluminescence (CL), Light Microscopy (LM), and Scanning Electron Microscopy (SEM). Photomosaics of the samples were collected and the amount of cement, porosity, and average grain sizes were determined by point-counting. The cement was easily recognized from the quartz grains by the difference in luminescence. The experiments indicate that the presence of amorphous silica results in rapid silica cementation in quartz aggregates (e.g., up to 12% cement by volume in 4 weeks at 450°C). The amount of cementation is a function of substrate type, time, temperature, and ionic strength of the brine. The rate of silica transport through the length of the experimental charge appears to be limited by the silica solubility and its rapid depletion by cementation. Although most of the cement was derived from the amorphous silica, evidence for local dissolution-precipitation was observed. The experiments demonstrate that the mobility of silica, and consequent precipitation of cement, does not require a temperature or pressure gradient as is commonly assumed. Rather, the only requirement is a concentration gradient, which is much easier to maintain in a variety of geologic environments. In addition, we have begun to investigate the important role of iron oxides on silica transport and cementation. Preliminary results show the amount of cementation is increased in the presence of iron oxides, which is most likely due to an increase in silica solubility.
MR21C-07
Modeling the Time-dependent Changes in Electrical Conductivity of Basaltic Melts With Redox State
The electrical conductivity σ is an efficient probe of mass transfer processes within silicate melts and magmas. Little attention has been given to the influence of redox state (fO2) on the melts conductivity. We present an experimental setup allowing electrical conductivity measurements for basaltic melts under variable fO2. We demonstrate a significant dependence of σ with fO2, allowing to characterize in situ the mechanisms and kinetics of redox changes in the melt. Experiments were conducted on basalts from Pu'u 'O'o, Hawaii, and Mt.Vesuvius, Italy. Measurements were performed cylindrical glass samples (OD: 6mm, ID: 1mm, L: 8mm) using an impedance spectrometer. Experiments were conducted in a 1atm vertical furnace, from 1200°C to 1400°C. Variable gas atmosphere (air, CO2 or CO-CO2 gas mixtures) were used, imposing ΔNNO from -1 to +7. Electrical conductivities were determined for the two melts at constant fO2, different T (constant fO2) and constant T, different fO2 (variable fO2) obtained by changing the gas composition. Isothermal reduction and oxidation cycles were performed. Glasses quenched from different T and fO2 conditions were analyzed by electron microprobe, the FeO concentration was determined by wet chemistry. In constant fO2 experiments, a small but detectable effect of fO2 on σ is evidenced. At 1300°C, the difference in the Kilauea sample conductivity between reduced (ΔNNO=-1) and oxidized (ΔNNO=+7) fO2 is <1(ohm.m)-1, the sample being more conductive when reduced. The temperature dependence of σ was fitted using Arrhenian equations, the activation energy Ea being 100kJ/mol. Sodium was identified as the main charge carrier in the melts. The fO2-effect on σ can thus be attributed to the influence of the Fe2+/Fe3+ ratio on sodium mobility. The fO2-dependence of σ was included in the model of Pommier et al.(2008), allowing the conductivity of natural melts to be calculated as a function of T, P, H2O, and fO2. Variable fO2 experiments confirmed the increase in σ when reducing the melt. At 1200°C, for both reduction-oxidation cycles, a stable value of σ following a change in fO2 is reached in 15hours, while 2hours are needed at 1400°C. The real-time changes in σ of basaltic melts following fO2 step changes were monitored. The time-dependent changes in σ are interpreted in terms of kinetics processes due to redox reequilibration between melt and gas. The evolution of σ with time can be fitted using a diffusion-limited process for reduction in CO-CO2 gas mixtures and oxidation in air. However, a reaction at the gas-melt interface probably rate limits oxidation in CO2. Reduction and oxidation rates are similar and increase with T. Oxidation-reduction rates calculated from the analysis of the conductivity evolution with time range from 10-9 to 10-8m2/s for the T range 1200-1400°C. These reaction rates are in agreement with typical alkali diffusion coefficients in basaltic melts. However, the high value of Ea (230kJ/mol) calculated from the T dependence of the oxidation-reduction rates agrees with the Ea for alkali-Earth elements. Furthermore, microprobe analyses document the existence of alkali-Earth cation fluxes during oxidations and reductions. Such cation migration probably occurs to charge-balance electron fluxes in the melt, in agreement with the study of Cooper et al. (1996). Our results suggest that the migration of alkali and alkali-Earth elements rate-limits the redox state changes in basaltic melts, and that redox mechanisms are not restricted to oxygen chemical diffusion. A discussion of chemical vs tracer oxygen diffusion studies is proposed.
MR21C-08
Three-Dimensional Melt Distribution in Partially Molten Rocks
Interpreting the geochemical and geophysical observations at ocean ridges requires knowledge of permeability and melt distribution of partially molten rocks. Recent developments in 3-D modeling of complex flows in porous media lead to estimates of the permeability of partially molten rock to complement values obtained through laboratory measurements, theoretical analysis and numerical simulations the primary tools used to estimate the permeability of partially molten rock. Quantification of melt channel connectivity requires high-resolution data on 3-D melt distribution. In this study, we conducted X-ray synchrotron microtomography experiments to characterize the 3-D melt distribution in synthetic, partially molten aggregates and natural mafic crystalline rocks. Textually equilibrated partially molten samples containing mantle olivine plus 5, 10, 20% MORB were fabricated experimentally at 1.5 GPa and 13508222;3352;0C. Natural rocks include glassy partially crystalline nodules from Hawaii and mafic cumulates from the Rum Layered Intrusion. We obtain textual information such as the pore geometry (throat size and dihedral angle) and channel connectivity from the microtomography images using MATLAB and AVISO. Comparison between these samples provides a quantitative characterization of how melt fraction and lithologic composition affect melt distribution. The tomographic data are incorporated into our network models to obtain macroscale transport properties. The tomographic study thus provide constraints on rates of melt migration and melt extraction within the partially molten regions beneath ocean ridges.