HR: 1340h
AN: MR23A-0192    [Abstracts]
TI: Thermal Conductivity of Deep Mantle Phases: Grain-size as a Primary Control of Radiative Transfer
AU: Taran, M N
EM: taran@mineral.freenet.kiev.ua
AF: National Academy of Sciences of Ukraine, Institute of Geochemistry, Mineralogy and Ore Formation., Prosp Palladina 34, Kyiv, 03680 Ukraine
AU: * Hofmeister, A M
EM: hofmeist@wustl.edu
AF: Washington U., Dept. of Earth and Planetary Sciences, 1 Brookings Dr., St. Louis, MO 63130 United States
AU: Yuen, D A
EM: davey@krissy.geo.umn.edu
AF: U. Minnesota, ept. of Geology and Geophysics and Minnesota Supercomputing Institute, 310 Pillsbury Dr. S.E, Minneapolis, MN 55455 United States
AU: Yanagawa, T
EM: yanagwtm@jcom.home.ne.jp
AF: Kyushu University, Dept. Earth & Planetary Sciences, 6-10-1 Hakozaki, Fukuoka, 812-8521 Japan
AB: The assumed dependence of thermal conductivity (k) on temperature (T) strongly influences results from mantle convection models due to feedback in the temperature equation. Numerical calculations have show that the exponential dependence of viscosity on T has a lesser influence on the style of mantle convection than does a T$^{3}$ dependence of k. Several constraints can be set on k of the deepest mantle phases through comparison. The phonon component of thermal conductivity (k$_{lat}$) for minerals is provided by laser-flash measurements. For 20 silicates and oxides examined so far, k$_{lat}$ becomes independent of temperature above roughly 1200 to 1900 K. This behavior arises because discrete phonon states only exist at low frequency, and these states are all populated by about 1500 K. Invariant k$_{lat}$(T) at high T is thus universal, and expected for the perovskite (pv) and post-pv phases, regardless of chemical composition. Deep mantle behavior thus depends on diffusion of photons. The formulation for an effective thermal conductivity due to radiative transfer was recently revised to account for physical scattering and emission characteristics of a solid medium. We applied this formulation to phases with very different spectra (perovskite, garnet, olivine), with and without Fe$^{3+}$-Fe$^{2+}$ charge transfer bands, and with and without d-d transitions of Fe$^{2+}$. We find that grain-size and Fe content are key variables, whereas charge transfer and mineral structure have relatively little effect. For grain-size = 0.1 cm and Fe/(Fe+Mg) = 0.1, k$_{rad}$ is roughly 0.1377-0.000612T+6.28T$^{2}$/10$^{7}$ in W/m-K. Lowering grain size and either raising or lowering Fe content lowers k$_{rad}$. Increasing grain-size raises k$_{rad}$ at low T. The high temperature behavior for large grain-size is difficult to predict because absorptions shift and intensify with temperature, but few measurements exist. Transparent materials such as Fe-free phases and minerals with low spin Fe can have large k$_{rad}$ at high T when grain-size nears 1 cm. However, in D", phases should bear Fe, due to interactions with the core, and high spin Fe is stable, as this disordered state is promoted at high T by its large magnetic and electronic entropy (large positive Clapeyron slope). Therefore, k$_{rad}$ of the ppv phase in D" can be modeled with the above equation. The power law dependence on T weakens convection in D". The stability of this layer against convection largely depends on the concentration of Fe ions in the silicate and oxide phases. Incorporation of metal particles similarly increases opacity, making radiative transfer less efficient, and convection more vigorous and time dependent
DE: 8120 Dynamics of lithosphere and mantle--general
DE: 8124 Earth's interior--composition and state (old 8105)
DE: 8130 Heat generation and transport
DE: 5139 Transport properties
DE: 3939 Physical thermodynamics
SC: Mineral and Rock Physics [MR]
MN: 2004 AGU Fall Meeting