Experimental Measurement of the Rheology of Eclogite and its Constituent Minerals

Harry W. Green, II (hgreen@ucrac1.ucr.edu)¹, Junfeng Zhang¹, Zhenmin Jin² & Shuyan Jin²

¹ IGPP, University of California, Riverside, CA 92521, USA

² Faculty of Earth Sciences, China Univ. of Geosciences, Wuhan, CHINA

Eclogite, the high-pressure equivalent of basalt/gabbro, consists of approximately equal amounts of garnet and pyroxene, commonly with minor quartz and rutile. The limited experimental data available previously for garnets and pyroxenes have shown them to exhibit very high flow stresses, suggesting that this rock type should be significantly stronger than mantle peridotite. The expected very high strength, coupled with restriction of stability of the eclogitic mineral assemblage to P > 2 GPa at elevated temperatures, has inhibited previous motivation and ability to measure eclogite rheology. However, with the discovery of continental rocks subducted to and exhumed from depths > 120 km corresponding to the diamond stability field (P > 4 GPa), and microstructural evidence from naturally deformed eclogites suggesting that the pyroxene specific to these rocks (omphacite) may be significantly weaker than those for which rheological data exist (diopside and enstatite), a need has developed to understand the flow properties of these rocks. We have measured the rheology of a reconstituted eclogite and its constituent minerals (grain-size ~30-50 microns) at 3 GPa; 1400-1600K in a nominally dry environment. We find that eclogite has a flow stress comparable to that of harzburgite, that omphacitite is much weaker, and that garnetite is extraordinarily strong. These results explain why many natural eclogites exhibit evidence of extensive flow and suggest that delamination of the oceanic crust from the underlying mantle during subduction due to contrasting rheologies is unlikely in the shallow mantle but quite possible in the mantle transition zone if garnet is abundant in former oceanic crust but not in the underlying cold mantle lithosphere. The high strength of polycrystalline silicate garnet suggests that the lower portion of the mantle transition zone (~500-700 km) may be a layer of enhanced viscosity, conceivably the strongest region of the mantle, depending on the abundance of garnet.

The Effect of Rutile on High-Field-Strength-Element Behaviour in the Crust-Mantle System

Trevor H. Green (trevor.green@mq.edu.au)

GEMOC, Dept. Earth and Planetary Sciences, Macquarie University, N.S.W. 2109, Australia

New partition coefficients (Ds) determined at high pressure for Nb, Ta, Zr and Hf between rutile and silicate and carbonatitic liquids and aqueous fluid are combined with literature data to assess the impact of rutile on HFSE behaviour. Rutile decouples Nb and Ta from Zr and Hf to varying extent, depending on the melt or fluid composition, from D_Nb,Ta/D_Zr,Hf of 3-4 for basanite to 17 for rhyodacite to 3-12 for carbonatite to 4-12 for fluid. All rutile/melt pairs except for one (a highly silicic composition) have D_Nb/D_Ta <1, indicating that such melts will have higher Nb/Ta than their source, if rutile is residual in the melting process. In contrast, highly silicic melts will have low Nb/Ta. It is significant that rutile from eclogites from kimberlites has median Nb/Ta of 30, and is possibly residual from extraction of a low Nb/Ta silicic melt as the oceanic crust was subducted. The rutile/fluid Nb and Ta results show D_Nb/D_Ta <1. Thus fluid does not appear to be crucial in causing the low Nb/Ta for continental crust, and available D data suggest that an effective fractionation mechanism capable of producing this low Nb/Ta is rutile residual to highly silicic melts. The results for Zr and Hf indicate that rutile will have much less effect on Zr/Hf than on Nb/Ta, so that these two element pairs are also decoupled in terms of their ratios. In contrast to this proposal for rutile controlling Nb/Ta in continental crust, low Nb/Ta ratios (« chondritic value of 17) in some island arc volcanics may result from a 2-stage melting process and not necessarily involve rutile. This may occur because major residual silicate minerals controlling fractionation in the first melting stage (clinopyroxene, orthopyroxene and garnet) all have D_Nb/D_Ta <1, leaving a source for later melting with Nb/Ta « 17.

Bulk Moduli and P-V-T Data of the High-Pressure Phases Topaz-OH, Al₂SiO₄(OH)₂, and Phase Pi, Al₃Si₂O₇(OH)₂

Klaus-Dieter Grevel (klaus-dieter.grevel@ruhr-uni-bochum.de), Detlef Wilhelm Faßhauer & Silvia Rohling

Ruhr-Universität Bochum, Institut für Geologie, Mineralogie und Geophysik, D-44780 Bochum, Germany

The high-pressure phases topaz-OH, Al₂SiO₄(OH)₂, and phase Pi, Al₃Si₂O₇(OH)₂, were investigated by Wunder et al. (1993a, 1993b) in much detail. In order to enable thermodynamic calculations in the high-pressure part of the chemical system Al₂O₃-SiO₂-H₂O, bulk moduli as well as P-V-T data of these phases were determined using a multi-anvil-X-ray-apparatus (MAX-80, Peun et al. 1995) which is installed at synchrotron beam-line F.2 of the DORIS III storage ring of HASYLAB (Desy, Hamburg). Several experimental runs were carried out for both phases, first compressing the samples at room temperature up to a distinct pressure (maximum pressure: 7.5 GPa) and then heating them up to a maximum temperature of about 800°C. For all X-ray diffraction experiments synthetic material was used which was mixed with vaseline as pressure medium; NaCl served as standard for pressure calibration. The bulk moduli of the phases investigated were calculated using the Birch-Murnaghan equation of state (EOS); k' was assumed to be equal 4: topaz-OH: 142.77 GPa, (sum) = 1.47 GPa, V⁰ = 355.5 ų phase Pi: 133.13 GPa, (sum) = 2.47 GPa, V⁰ = 307.95 ų Fixing these values the parameters (dK/dT)_P and a of the temperature dependent Birch-Murnaghan EOS were calculated: topaz-OH: (dK/dT)_P = -0.0009 GPa K^-1, (sum) = 0.0034 GPa K^-1; a = 2.069e-5 K^-1, (sum) = 0.070e-5 K^-1 phase Pi: (dK/dT)_P = -0.0313 GPa K^-1, (sum) = 0.0095 GPa K^-1; a = 2.925e-5 K^-1, (sum) = 0.298e-5 K^-1 Obviously, the bulk modulus of phase Pi is strongly temperature dependent, while the bulk modulus of topaz-OH remains almost constant over the entire P-T range of investigation.

Peun T, Zinn P, Lauterjung J & Hinze E, Bochumer geol. u. geotech. Arb, 44, 139-143, (1995).

Wunder B, Medenbach O, Krause W & Schreyer W, Eur. J. Mineral, 5, 637-649, (1993a).

Wunder B, Rubie DC, Ross II CR, Medenbach O, Seifert F & Schreyer W, Am. Mineral, 78, 285-297, (1993b).