MR33A-1836
The thermodynamics and roughening of solid-solid interfaces
We consider the morphological stability of an interface separating two non-hydrostatically stressed solids.
The solids are allowed to exchange mass by undergoing a phase transformation in which one solid is
transformed into the other. Thermodynamic relations for the transformation of a mass element are derived
and a stability analysis of the morphology of the propagating interface is carried out. In the majority of solid-
solid phase transformation processes the rearrangement of the material structure is accompanied by a
change in density (including density change due to altered deformability). For that reason we consider
examples where density is an important order parameter that quantitatively characterizes the difference
between the two phases.
We consider two types of phase transitions underlying the kinetics, first order and second order, which result
in fundamentally different behaviors at the phase boundary. A first order phase transition occurs when the
two phases have different referential densities and we show that it results in a morphological instability along
the boundary whereas the second order phase transition stabilizes the boundary. Relevant examples include
dry recrystallization, isochemical transformations and coherent precipitates. Numerical simulations are
performed in the non-linear regime to investigate the evolution and roughening of the interface. It is shown
that even small contrasts in the referential densities of the solids may lead to the formation of finger like
structures aligned with the principal direction of the far field stress.
http://folk.uio.no/joachma/Research.html
MR33A-1837
Dependence of Thermal Conductivity on Water Saturation of Sandstones
Information on thermal conductivity of rocks and soils is essential in applied geothermal and hydrocarbon
maturation research. In this study, we investigate the dependence of thermal conductivity on the degree of
water saturation. Measurements were made on five sandstones from different outcrops in Germany. In a first
step, we characterized the samples with respect to mineralogical composition, porosity, and microstructure by
nuclear magnetic resonance (NMR) and mercury injection. We measured thermal conductivity with an optical
scanner at different levels of water saturation. Finally we present a simple and easy model for the correlation
of thermal conductivity and water saturation.
Thermal conductivity decreases in the course of the drying of the rock. This behaviour is not linear and
depends on the microstructure of the studied rock. We studied different mixing models for three phases:
mineral skeleton, water and air. For argillaceous sandstones a modified arithmetic model works best which
considers the irreducible water volume and different pore sizes. For pure quartz sandstones without clay
minerals, we use the same model for low water saturations, but for high water saturations a modified
geometric model. A clayey sandstone rich in feldspath shows a different behaviour which cannot be explained
by simple models. A better understanding will require measurements on additional samples which will help to
improve the derived correlations and substantiate our findings.
MR33A-1838
Water Flow in Fractured Rock Under Changeable Confining Pressure
Fluid flow in fractured rocks is largely governed by the properties of the fractures. In turn, fracture geometry can be affected by confining pressure. There is thus a feedback between fluid flow and rock properties. To better understand the interaction between fluid pressure and fracture properties, we have integrated results from experimental, analytical and numerical methods. In the experiments, we monitor the change in permeability as confining pressure changes. The analytical models describe fluid-solid coupling using fluid mass conservation and solid equilibrium equations. These coupling equations are simulated numerically in a system with a porous matrix cross-cut with fractures; numerical results are used to interpret the experimental measurements. Measurements were performed on fractured sandstones. We find that fracture permeability changes by between 50% and 70% for changes in confining pressure from 6MPa to 14MPa. Permeability decrease is a negative exponential function of confining pressure as fracture aperture depends strongly on confining pressure. Saturating pore pressure, on the other hand, has little effect on permeability. The numerical simulations allow us to identify the location and nature of changes in flow that accompany changes in fracture properties as the confining pressure changes.
MR33A-1839
Determination of Hydrogen Diffusion Coefficients in Fused Silica From 23 to 250°C Using Raman Spectroscopy
The oxygen buffer technique is routinely used in experimental studies of redox sensitive geochemical reactions at elevated pressures (P) and temperatures (T). However, this technique is limited to T above about 400°C due to the low permeability of sample containers (Pt or Ag-Pd alloys) to hydrogen at lower T. Preliminary results of Chou et al. (Geochim. Cosmochim. Acta, 2008, doi:10.1016/j.gca.2008.07.030) indicate that the use of fused silica capillary (FSC) container may extend this technique to lower T. In this study, hydrogen diffusion coefficients (D) in FSC were determined from 23 to 250°C by measuring the loss of hydrogen from the FSC containers (0.3 mm OD, 0.1 mm ID, and ~10 mm long) with Raman spectroscopy using CO2 as an internal standard. First, CO2 was loaded cryogenically in a FSC capsule (Chou et al., ibid.). The capsule was then inserted in a protective ceramic tube, sealed in a gold capsule containing Fe powder and water, and heated at 300°C under 100 MPa of Ar external P in a cold-seal pressure vessel for several days allowing H2 to diffuse into the capsule. After quench, the Raman spectra were collected and the initial relative concentration of hydrogen in the silica capsule was derived from the peak height ratios between H2 (near 587 cm-1) and CO2 (near 1387 cm- 1). The sample capsule was then heated at a fixed T at one atmosphere to let H2 diffuse out of the capsule, and the changes of hydrogen concentration were monitored by Raman spectroscopy after quench. This process was repeated with different heating durations at 23 (room T), 50, 102, 157, 200, and 250°C. The values of D (in cm2s-1) in FSC were obtained by fitting the observed changes of hydrogen concentration to an equation based on Fick's second law. Our results can be represented by: Ln D (±0.14) = (-39810/RT) - 9.5491 (r2 = 0.9985) where R is the gas constant, and T in Kelvin. The slope corresponds to an activation energy of 39.81 kJ/mol. Our D values are about a half order of magnitude lower than those extrapolated from the values of Lee et al. (1962, J. Chem. Phys., 36, 1062) measured between 300 and 500°C, and about a half order of magnitude higher than those reported by Berrer (1941, Diffusion in and through solids, Cambridge Univ. Press, 141). The D in FSC determined at 23°C is about three times higher than that of Pt at 500°C, indicating that FSC is a suitable membrane for hydrogen at T below 400°C, even at room T, and has a great potential for studying redox reactions at these T, especially for the systems containing organic material.
MR33A-1840
Fluorine Diffusion in Titanite: Preliminary Results from Experiments
Recent experiments have yielded new data on fluorine-hydroxyl diffusion in titanite. Experiments utilized natural low-F titanite crystals, sectioned into wafers, polished to 1 μm alpha alumina and finished with colloidal silica. Prepared crystals were placed with polished faces in contact with a F-rich diffusion source. Each titanite-source assembly was then sealed in an evacuated Si glass ampoule and heated to temperatures between 600 and 900 °C, or sealed in an Ag envelope and heated in a cold-seal pressure vessel to temperatures between 600 and 906 °C at 50 to 100 kPa. Several F sources were employed: PbF, NaF, CaF2, LiF, and fluoroapatite. PbF and NaF strongly reacted with the titanite crystals and their capsules, particularly at higher temperatures. CaF2 and LiF were less destructive, but reacted with and etched the titanite faces, precluding analysis. Titanite retained a polished and largely unaltered surface only in those runs with an apatite source. We noted only a slight change in color (green to brown, likely due to oxidation of minor amounts of Fe present in the titanite) in these runs. We analyzed the polished faces of crystals from successful experiments using nuclear reaction analysis (NRA), employing the 340 keV resonance for the nuclear reaction 19F(p,αγ) 16O to measure fluorine. The results show elevated fluorine at the surface and diminishing concentrations with depth. Fitting a simple complementary error function to each concentration profile yields a diffusivity, assuming a model of a semi-infinite material with a surface maintained at constant concentration. For the runs at near atmospheric pressure, diffusion coefficients of 2×10-21m2/s and 6×10-21m2/s are obtained at temperatures of 850°C and 889°C, respectively. A single cold seal experiment at 906°C suggests a diffusivity of 3×10-21 m2/s. While preliminary, these data suggest that fluorine-hydroxyl exchange in titanite is more sluggish than in apatite, biotite, and tremolite.
MR33A-1841
REE Diffusion on Quartzite Grain Boundaries: Preliminary Results from Experiments
Two novel experimental configurations were used to characterize REE diffusion along grain boundaries. Both configurations involve juxtaposing a concentrated REE source phase with a synthetic quartzite containing a disseminated sink phase (~5 vol. %). The quartzite was synthesized from a mixture of powdered natural quartz (22-75 μm) fired at 1000 °C and combined with polycrystalline fragments of Dy2O3 or small crystals of synthetic xenotime. These mixtures were annealed for 3 days at 1000 °C and 1GPa in nickel capsules within a piston cylinder apparatus. This produced a ~100 μm grain size quartzite containing dispersed, 10-40 μm Dy2O3 - Dy2SiO5 or xenotime particles. Polished sections of the quartzite containing the Dy-phase were paired with Pr2O3 powder, and those of the xenotime quartzite were coupled with DyPO4 powder. These diffusion couples were run in the piston cylinder at temperatures between 1000 and 1300 °C, 1 GPa pressure, for durations between 1 day and 1 week. In both the quartzite synthesis and diffusion experiments, all materials were prepared to minimize water in the materials. Cathodoluminescence imaging of the run products suggests that that REE diffusion occurs mainly along grain boundaries, with some penetration into the quartz crystals by lattice diffusion. For runs with T >1200 °C, EPMA evaluation of the sink particles reveals a decrease in concentration of the diffusant as a function of distance from the boundary. In the oxide experiments, the penetration of Pr varied from 370 μm at 1300 °C for 24 hours, to 245 μm at 1250 °C for 48 hours, to 150 μm at 1200 °C for 72 hours. In a single phosphate experiment, Dy penetrated to 240 μm at 1300 °C in 6 hours. In both types of experiments, those with T < 1200 °C have thus far failed to show measurable penetration of the diffusant for durations up to a week.
MR33A-1842
Characterization of Fe-Mg Interdiffusion Along Dry Quartzite Grain Boundaries Using the Dispersed Sink Method
Grain boundary diffusion was characterized using a new experimental approach that involves migration of a diffusant from a crystalline source along dry grain boundaries of a polycrystalline rock analog, where it is 'intercepted' by dispersed sink particles. This approach involves two steps: (1) pre-synthesis at elevated P-T conditions in a piston-cylinder apparatus of a rock analog exhibiting an equilibrium microstructure of the major phase (quartz in this case) along with much smaller, dispersed sink grains; and (2) diffusion experiments in which the rock containing the dispersed sink phase is placed against a source of diffusant that is compatible in the sink phase but incompatible in the major minerals of the rock of interest. In an effort to characterize Fe-Mg interdiffusion along quartz grain boundaries, two quartzites containing different sink phases were synthesized at elevated P-T conditions: one containing ~5% dispersed fayalite and the other containing ~10% dispersed enstatite. The presynthesized quartzites were cut into slabs and polished. Diffusion experiments were set up by juxtaposing the polished surfaces against sources of Fe or Mg diffusant, which included MgF2, MgO or polished slabs of San Carlos olivine (~Fo90). In some experiments, diffusion couples were formed by placing polished slabs of the fayalite- and enstatite-bearing quartzites against one another. The juxtaposition of source and sink materials resulted in strong chemical potential gradients in Fe and Mg, but the only available pathway for interdiffusion of the source and sink is the quartzite grain boundaries. After the diffusion experiments, the diffusion couples were sectioned and the sink phases were analyzed for the diffusant by electron microprobe. The time-integrated grain-boundary flux was evaluated by measuring the total number of diffusant atoms accumulated in the sink in a given time. Assuming local partitioning and diffusive equilibrium between sink phases and grain boundaries, the diffusant concentration in the sink phases will reflect (i.e., be proportional to) the concentration profile of the diffusant in the grain boundaries throughout the bulk sample. Microprobe analyses and BSE images of the sink phases demonstrate significant and efficient transport of diffusants along the quartz grain boundaries. In an experiment in which a slab of fayalite-bearing quartzite was placed against MgF2 and run at 1125°C for two hours, MgO contents of the fayalite grains show that Mg was transported ~200 μm from the diffusant source. This transport distance suggests an effective grain boundary diffusivity on the order of 10-12 m2s-1. Interestingly, the 8 day experiments do not show significantly greater transport distance, but at a given distance from the source, the fayalite grains from the longer experiments have a higher concentration of MgO. Preliminarily we suggest that the grain boundaries become 'saturated' in the diffusant, effectively shutting down grain boundary transport. Similar processes may be important in the preservation of diffusive halos surrounding minerals in natural rocks. In experiments that placed a fayalite-bearing quartzite against an enstatite-bearing quartzite, the MgO content of fayalite grains 20 μm from the interface is the same as that of the starting material. Compared to the diffusion experiments that used Mg-rich sources, the characteristic transport distance is approximately an order of magnitude smaller. This preliminary result suggests that grain boundary transport may depend upon the diffusant concentration in the source.
MR33A-1843
Ti Diffusion in Pyroxene
Diffusion of titanium has been characterized in natural enstatite and diopside under buffered conditions and in air. The sources of diffusant for the enstatite experiments were mixtures of Mg, Si and Ti oxide powders, which were combined and heated at 1300°C overnight, and then thoroughly mixed with synthesized enstatite powder and heated for an additional day at 1300°C. Sources for diopside experiments were prepared similarly, using Ca, Mg, Si, and Ti oxide powders combined with synthesized diopside powder, with heating of source materials at 1200°C. Buffered experiments were prepared by enclosing source material and pyroxene (polished and pre-annealed under conditions comparable to those to be experienced in the experiment) in AgPd or platinum capsules, placing the metal capsule in a silica glass capsule with a solid buffer (to buffer at NNO or IW) and sealing the assembly under vacuum. Some experiments on enstatite were run in air; sample and source were placed in Pt capsules and crimped shut. Prepared capsules were then annealed in 1 atm furnaces for times ranging from 8 hours to a few months, at temperatures from 950 to 1200°C. The Ti distributions in the pyroxene were profiled with Rutherford Backscattering Spectrometry (RBS). The following Arrhenius relation is obtained for Ti diffusion in a natural enstatite, for diffusion normal to the (210) cleavage face (950 - 1150°C, experiments run in air): DTi = 1.9×10-10 exp(-300 ± 44 kJ mol-1/RT) m2 sec-1. Diffusion under NNO and IW-buffered conditions is similar to that for experiments run in air, suggesting little dependence of Ti diffusion on oxygen fugacity. There is also little evidence of anisotropy, as diffusion normal to (001) does not differ significantly from diffusion for the other orientation. Preliminary findings for Ti diffusion in diopside suggest diffusivities similar to those for enstatite. Ti diffusivities in enstatite are similar to those of the trivalent REEs (Cherniak and Liang, 2007), but more than two orders of magnitude slower than those of Fe-Mg (ter Heege et al., 2006) and Cr (Ganguly et al., 2007). These respective variations may reflect the interplay of cation size and charge, or may point to the substitution of Ti on the tetrahedral site. Measurements of diffusion under a broader range of conditions and for other high field strength elements are underway to better interpret these findings. Major and trace element zoning in pyroxenes have been observed in residual peridotites and mafic cumulates. The large differences in cation mobility among Ti, Cr, and Fe-Mg in pyroxene may allow us to distinguish the dominant process that gives rise to the chemical disequilibria. In contrast to those produced by subsolidus reequilibration during cooling, the apparent diffusive boundary layer thicknesses as measured by major and trace elements in a pyroxene grain are not sensitive to the respective cation diffusion rates if zoning is produced by magmatic processes that involves dissolution- precipitation. Examples of zoning in pyroxenes produced by magmatic and subsolidus processes will be discussed. Ganguly et al. (2007) GCA 71, 3915-3925; ter Heege et al. (2006) Eos Trans. AGU 87, Fall Mtg. Suppl. MR21A-0004; Cherniak and Liang (2007) GCA 71, 1324-1340
MR33A-1844
Diffusion of REE, Hf and Sr in Olivine
We have determined diffusion coefficients of the rare earth elements Ce, Nd, Sm, Eu, Lu, and also of Sr and Hf, in single crystals of natural olivine at atmospheric pressure, at an oxygen fugacity of 10-5 Pa and a temperature of 1275 °C. Sources of diffusants were thin films of olivine composition doped with the relevant elements. Thin films were produced by PLD (pulsed laser deposition) and RBS (Rutherford backscattering) was used to measure thickness and stoichiometry of the films as well as to analyze the concentration profiles. The concentration profiles were numerically fitted to yield the following diffusion coefficients (D, in m2/s): log DCe: -19.61 ± 0.21; log DNd: -19.54 ± 0.11; log DSm: -19.15 ± 0.05; log DEu: -19.10; log DLu: -19.00, log DHf: -20.23 ± 0.07; log DSr: -18.7. Diffusion coefficients of the rare-earth elements increase from Ce to Lu, demonstrating the role of ionic radius in controlling diffusion because all REE are trivalent. The tetravalent and divalent cations hafnium and strontium diffuse an order of magnitude slower and faster, respectively, than the REE in olivine. This highlights the important influence of ionic charge on diffusion rates. The diffusion coefficients of the REE are slower by a few orders of magnitudes than the diffusion rate of Cr in olivine [1]. The rates found in this study are slower than those assumed by a model [2] for compositional modification of melt inclusions in olivine. Use of our data in their calculations indicates that it will take longer to modify the composition of melt inclusions in olivine (millions of years rather than thousands of years) but the fractionation of HREE from LREE will be larger. [1] Ganguly J, Ito M (2006) Geochim Cosmochim Acta, 70, 799-809. [2] Cottrell E, Spiegelman M, Langmuir CH (2002) Geochem Geophys Geosyst, doi:10.1029/2001GC000205
MR33A-1845
D-H Interdiffusion in Olivine: Preliminary Results
Studying interdiffusion rates of deuterium and hydrogen in olivine will improve our understanding of point defects and electrical conductivity in this important mantle mineral. Our preliminary results indicate that D-H interdiffusion occurs at a rate similar in magnitude to that of chemical diffusion of hydrogen in San Carlos olivine. Three experiments were performed on 500-micron spheres of San Carlos olivine previously annealed and hydrated with ~30 ppm-wt H2O. These olivine spheres were pressurized to 2 GPa, at 900-950 C, for 1-2 hours in a 99.9% D2O bath. Deuterium was introduced as a traceable species of hydrogen. We used the Cameca 6f SIMS at Arizona State University to measure hydrogen and deuterium across samples from each of these experiments. The experiment performed at the lowest temperature (900 C) and for the shortest duration (1 hr) had not achieved equilibrium allowing us to fit the deuterium diffusion profile. Least-squares fitting of the 900 C and 1 hour data to an expression for non steady-state diffusion through a sphere yields a diffusion coefficient of ~6e-12 m2/s. Additionally, nominally anhydrous San Carlos enstatite used as a silica buffer in the same experiments allows us to estimate chemical diffusion (or incorporation rate) of deuterium through enstatite. We roughly measured chemical diffusion of deuterium through enstatite at 900 C and 950 C to be ~2e-12 m2/s and ~6e-12 m2/s respectively. These preliminary measurements are not yet tightly constrained but are useful for optimizing run conditions of future diffusion experiments on olivine and enstatite.
MR33A-1846
Si and O diffusion in (Mg,Fe)2SiO4 wadsleyite and ringwoodite and its implication for rheology of the mantle transition zone
Si and O diffusion rates on polycrystalline (Mg,Fe)2SiO4 wadsleyite and ringwoodite have been determined at pressures between 16 to 22 GPa and temperatures between 1673 to 1873K. High pressure experiments were conducted using a Kawai-type multi-anvil high pressure apparatus. Pre-synthesized polycrystalline wadsleyite or ringwoodite were used as starting materials. Diffusing sources of 29Si and 18O enriched (Mg,Fe)2SiO4 thin film were coated on the surface of wadsleyite and ringwoodite by a pulsed laser deposition (PLD) in order to ensure the deposition of stoichiometric thin films. The diffusion profiles were obtained by a depth-profiling mode using a secondary ion mass spectrometry (SIMS). The obtained all diffusion profiles were composed of volume and grain-boundary diffusion regimes. Therefore, Arrhenius relations in volume and grain-boundary diffusion rates in wadsleyite and ringwoodite have been determined simultaneously. Their diffusion rates are characterized as follows: In (Mg,Fe)2SiO4 wadsleyite with 20-80 wt. ppm H2O, Dv(Si)= 2.79x 10-8 [m2/s] exp(-409 [kJ/mol]/RT),deltaDGB(Si)= 1.31x10-15 [m3/s] exp(-327 [kJ/mol]/RT), Dv(O)= 3.04x10-11 [m2/s] exp(-291 [kJ/mol]/RT), deltaDGB(O)= 1.62x10-17 [m3/s] exp(-244 [kJ/mol]/RT). In (Mg,Fe)2SiO4 ringwoodite with 130-220 wt. ppm H2O, Dv(Si)= 3.33x10-6 [m2/s] exp(-483 [kJ/mol]/RT), deltaDGB(Si)= 5.70x10-14 [m3/s] exp(-402 [kJ/mol]/RT), Dv(O)= 2.87x10-9 [m2/s] exp (-367 [kJ/mol]/RT), deltaDGB(O)= 7.85x10-18 [m3/s] exp (-246 [kJ/mol]/RT). The results show that Si diffusion rates are slower than O diffusion rates and previously reported Mg-Fe interdiffusion rates in both (Mg,Fe)2SiO4 wadsleyite and ringwoodite. Si is likely to be rate-controlling species in high-temperature creep involving diffusion creep and climb-controlled dislocation creep. Compared viscosities for diffusion and dislocation creep estimated from Si diffusion data with mantle viscosity inferred from geophysical observations, the mantle viscosity is explained by a grain size of 1 mm in diffusion creep regime and a stress of 0.1-1 MPa in dislocation creep regime. Consequently, plausible grain size and stress in the mantle transition zone may be 1 mm and 0.1-1 MPa.
MR33A-1847
Mobility of the trace incompatible elements under the uppermost mantle conditions
Trace element abundances of the oceanic mantle source regions are useful information to decipher the material circulation and transfer during the Earth's history. Here we report the new result of the reaction experiments between basalt and peridotite, and basalt and olivine aggregates at the uppermost mantle conditions. In order to investigate the mobility of elements migrated from basalt to peridotite and olivine and understands the behavior of incompatible trace elements under upper mantle conditions, reactions experiments between basalt (JB-1) and peridotite (JP-1) , and basalt and olivine aggregates (San Carlos) were examined at 1-3 GPa and 1073-1373K using a multi-anvil high pressure apparatus. After reaction experiments, SEM-EDS and LA-ICPMS were used to determine the concentrations of the major and trace elements. We found that the reaction zone consisting of garnet (pyrope-rich) and orthopyroxene (Fe-rich) was formed at the interface between the basalt and peridotite or olivine aggregates (-3 GPa). The textual and compositional observations suggest that the reaction zone act as a barrier to prevent the rapid melt intrusion and the major elements transfer through this zone. However, most trace elements reached deep inside the subsolidus peridotite zone in the peridotite-basalt experiments. Depth profiling data over 100 micron into the peridotite are fitted to the semi-infinite diffusant model equation. The obtained effective diffusion coefficients are likely to correspond to the grain boundary diffusion mechanism. Their values indicated the following order: Na, Rb, K > Ba, Sr, La, Ce, Zr, Th, Nb. D(Na) is 3.3-0.4×10-12m2/s at 1373K, which is about 4 times larger than D(Zr). This result is suggestive of the difference of trace elements mobility as a possible process to produce the chemical heterogeneity in the Earthfs interior. In this meeting, we will discuss the effective diffusion coefficients obtained from the basalt-olivine experiments and its temperature dependence.
MR33A-1848
The effect of water and iron content on electrical conductivity of upper mantle rocks.
Geophysical observations (MT and GDS) show the conductivity anomaly which may be related to the presence of water and melting. Recently, several researchers have estimated the water content in the transition zone (Huang et al. 2005; Yoshino et al. 2008) and the upper mantle (Wang et al.2006; Yoshino et al. 2006) by electrical conductivity methods. They may underestimate the water content, especially, Yoshino et al did too much underestimate. However, other coexisting phases such as pyroxene and its high-pressure polymorphs may also contribute to the bulk conductivity of the mantle. To test this hypothesis, we measured the electrical conductivity of upper mantle rocks- dunite, pyroxenite and lherzolite at ~ 2-3 GPa and ~1273-1573 K using impedance spectra within a frequency range of 0.1~1000000 Hz. The oxygen fugacity was controlled by a Mo-MoO2 solid buffer. The results show that the electrical conductivity of lherzolite and pyroxenite are ~ half and one order of magnitude higher than that of dunite. These differences were interpreted through a preliminary model involving water and iron content effects on the electrical conductivity. We extrapolated our results and compared the results with some of geophysical observations of the upper mantle. Our results indicate the maximum water content in oceanic upper mantle is as high as ~ 0.09wt % and suggest that pyroxenes dominate the bulk conductivity of upper mantle in hydrous conditions. These results indicated that our model with various water contents could explain the conductivity anomaly in the oceanic upper mantle without involving the presence of partial melt at these depths. This work was supported by national natural science foundation of china (40774036); the special grant from the president of Chinese Academy of Sciences and Graduate University of Chinese Academy Sciences.
MR33A-1849
Electrical Conductivity of Micas at High Temperatures
Electrical conductivity, along with seismic velocity, gives us clues to infer constituent materials and temperatures in the Earth's interior. Dry rocks have been considered to be electrically insulating at crustal temperatures. Observed high conductivity has been ascribed to the existence of fluids. However, Fuji-ta et al. (2007) recently reported that a dry gneiss shows relatively high conductivity (10-4-10-3 S/m) at the temperature of 300-400°C, and that it is strongly anisotropic in conductivity. They suggested that the alignment of biotite grains governs conductivity of the gneiss sample. Electrical properties of rock forming minerals are still poorly understood. We thus have measured electrical properties of biotite single crystals up to 700°C. In order to get a good understanding of conduction mechanisms, measurements have been also made on phlogopite and muscovite, which are common micas with similar crystallographic structures. Thin plates parallel to cleavages (thickness~0.1mm) were prepared from mica single crystals. Electrical impedance was measured by 2-electrode method. The specimen was kept in nitrogen or argon atmosphere. The conductivity measured parallel to cleavages is higher than that measured perpendicular to cleavages by 3-4 orders of magnitude. However, no significant difference in the activation energy of conductivity was observed between two directions. The activation energy of conductivity is ~50 kJ/mol for biotite and ~100 kJ/mol for phlogopite and muscovite. The conductivity of biotite is higher than those of phlogopite and muscovite by several orders of magnitude at the same temperature. The conductivity of biotite parallel to cleavages is ~10-1 S/m at 400°C. The conductivity of biotite increases irreversibly by heating. The irreversible change was not significant below 450°C. Remarkable increase is observed at the temperature of 450-550°C. No significant change was observed in the second heating. Such an increase in conductivity was not observed on phlogopite and muscovite. Brindley and Lemaitre (1987) reported that Fe2+ in biotite changes to Fe3+ at the temperature of 450-650°C. The increase in conductivity can be related to this change. Since ions Fe2+ and Fe3+ occupy equivalent sites in biotite, electrons can move easily. The hopping of electrons must be the dominant conduction mechanism of biotite.