HR: 11:50h
AN: V32B-07    [Abstracts]
TI: Ti-rich Silicate Perovskite: A New Lower Mantle Phase and the Possible Source of Unradiogenic Hf in Kimberlites and Carbonatites
AU: * Collerson, K D
EM: k.collerson@mailbox.uq.edu.au
AF: Univ. Queensland, Earth Sci., Brisbane, Q 4072 Australia
AU: Terasaki, H
V32B-07 AF: Tohoku Univ., Inst. Mineral. Petrol. & Economic Geol., Sendai, 980-8578 Japan
AU: Ohtani, E
EM: ohtani@mail.tains.tohoku.ac.jp
AF: Tohoku Univ., Inst. Mineral. Petrol. & Economic Geol., Sendai, 980-8578 Japan
AU: Suzuki, A
V32B-07 AF: Tohoku Univ., Inst. Mineral. Petrol. & Economic Geol., Sendai, 980-8578 Japan
AU: Kondo, T
V32B-07 AF: Tohoku Univ., Inst. Mineral. Petrol. & Economic Geol., Sendai, 980-8578 Japan
AB: In an attempt to synthesize the pre-exsolution homogeneous phase proposed by [1] as the protolith of exsolution-textured cpx-ilm xenoliths in kimberlite, we conducted a subsolidus MA experiment at 25 GPa and 1800°C using a natural cpx-ilm xenolith from Monastery kimberlite containing 17% TiO₂ as the starting composition. Phases identified (EPMA, Raman &XRD) were Ti-rich MgSi perovskite, Ti-rich CaSi perovskite and stishovite. TiO₂ contents ranged from 16-18% in the CaTiSiPv to between 12.5 and 25% in the MgTiSiPv. This indicates that an extensive field of solid solution exists in the system MgSiPv - CaSiPv - CaTiPv at pressures greater than 24 GPa [cf. 2,3]. Furthermore, a multi-phase system was observed using XRD in a DAC experiment at 30 GPa and 1800°C. By contrast, the maximum TiO₂ in majorite garnet in this composition is only 5-6% at 18 GPa [4]. Raman spectra for CaTiSiPv vary systematically with Ti content. This observation could have application for interpretation of spectra obtained in subsequent DA experiments on the stability of CaTiSiPv. The protolith of the cpx-ilm xenoliths does not exist as a single homogeneous Ti-rich silicate phase in the upper mantle. However, presence of a Ti-bearing phase in the lower mantle (LM) is inferred from crystals of CaSiPv and CaTiPv that occur in contact with each other in LM diamonds [5]. These were interpreted as reversion products, formed from CaSiPv and CaTiPv solid solution during ascent in kimberlite magma of at P < 9 GPa [2]. However, following [2] these phases must have existed as a single phase at higher pressure. Using the mean composition of CaTiSiPv produced in our experiments, we calculated that this solid solution is likely to involve 0.7 CaSiPv and 0.3 CaTiPv. Ti-rich SiPv in the LM phase may explain the "hidden" low Lu/Hf reservoir required by unradiogenic Hf isotopic compositions in kimberlites and carbonatites [6,7]. [1] Ringwood & Lovering (1970) EPSL,7, 371. [2] Kubo et al., (1997) PCM 24: 488-494. [3] Litasov & Ohtani (2005) PEPI 150: 239-263. [4] Collerson et al., (2004) Fall AGU Abstract Vol. [5] Hayman et al., (2005) CMP 149: 430-445. [6] Bizzarro et al., 2002. Geology 30: 771-774. [7] Nowell et al., 2004. J. Petrol., 45: 1583-1612.
DE: 1025 Composition of the mantle
DE: 1040 Radiogenic isotope geochemistry
DE: 3621 Mantle processes (1038)
DE: 3630 Experimental mineralogy and petrology
SC: Volcanology, Geochemistry, Petrology [V]
MN: Fall Meeting 2005