2003 Seattle Annual Meeting (November 2–5, 2003)

Paper No. 8
Presentation Time: 2:45 PM

SHOCK DATA FOR MGSIO3: SUPER-DENSE MELT ABOVE CORE-MANTLE BOUNDARY


AKINS, Joseph A., Seismological Laboratory, California Institute of Technology, 252-21, Pasadena, CA 91125, AHRENS, Thomas J., Seismological Laboratory, California Institute of Technology, 252-21, Pasadena, CA and ASIMOW, Paul D., Division of Geological and Planetary Sciences, California Institute of Technology, 170-25, Pasadena, CA 91125, tja@caltech.edu

Shock Hugoniot data for Sri Lankan enstatite (En99.2,Fs0.6,Wo0.2) (48-206 GPa) and synthetic enstatite glass [(Si/Mg)=1.04] (47-110 GPa), taken with static data of others, define 10 regimes along the crystal Hugoniot: 1) elastic to 6.7 GPa; 2) deformational, 6.7-15 GPa; 3) enstatite & majorite, 15-34 GPa; 4) majorite, 34-48 GPa; 5) majorite & akimotoite (ilmenite structure), 48-70 GPa; 6) akimotoite,70-105 GPa; 7) akimotoite & perovskite structure, 105-110 GPa; 8) perovskite, 110-~170 GPa; 9) perovskite & liquid 170-175 GPa; 10) liquid > 175 GPa. Hugoniot density-pressure-temperature data in the majorite, akimotoite & perovskite regimes agree closely with calculated values from Brillouin, ultrasonic and high-pressure x-ray data. Previously these high-pressure phases were identified in impact-induced veins in naturally shocked bronzite-bearing chondrites and laboratory shock-recovery experiments. Between 170 & 175 GPa the 3.6 – 5.4 % increase in density (supported by 3 of our data) is unmatched by known ultra-high pressure polymorphism or disproportionation reactions. We propose that this density increase accompanies melting (similar to the crystal quartz Hugoniot, where we also have shock-temperature measurements showing that super-heated stishovite at 115 GPa,~7000 K melts to a denser, 5000 K state). Melting in regime 9 is consistent with lower-pressure laser-heated diamond cell results (Sweeney & Heinz and Boehler), which fix a peak on the MgSiO3 perovskite melting curve at ~ 5700 K and ~ 105 GPa. We estimate perovskite melting at 5600 K at 130 GPa (upon shocking MgSiO3 glass), decreasing further to ~5100 K at ~170 GPa (upon shocking crystal enstatite). Taking into account thermal expansivity (1x10-5 / K) indicates that molten MgSiO3 is 2-3 % denser than the solid at lowermost mantle conditions. This may account for the stability of partially molten silicate in the ULVZ (proposed by Helmberger, Garnero and Williams).