The Pressure Stability of Chlorite in Synthetic Peridotites and Consequences on the Subduction of Water

Patrizia Fumagalli (patrizia@expe.terra.unimi.it) & Stefano Poli (stefano@biko.terra.unimi.it)

Dipartimento di Scienze della Terra, Via Botticelli 23, 20133 Milano, Italy

Phase relationships in hydrous peridotites at high pressure represent one of the primary inputs for subduction zone modelling. Although hydrous peridotites have been extensively investigated in the MgO-SiO₂-H₂O (MSH) and MgO-Al₂O₃-SiO₂-H₂O (MASH) systems, where chlorite and antigorite are the major hydrous phases, more complex systems are almost unexplored. This study focuses on the high pressure stability and relevance of chlorite in the model peridotite system Na₂O-CaO-FeO-MASH at subsolidus conditions.

Multianvil experiments were performed at the Dipartimento di Scienze della Terra, Milan. Three synthetic bulk compositions corresponding to peridotites at different degree of depletion were investigated: a harzburgite (Hébert et al., 1983), Tinaquillo lherzolite (Robinson & Wood, 1998) and a particularly Al-enriched peridotite (peridotite D, Mysen & Boettcher, 1975). Gels seeded with chlinochlore, diopside, forsterite and pyrope were used as starting materials and run at 2.0-6.0 GPa and 650°C-800°C. Runs lasted up to 400 hours. All experiments were performed at fluid saturated conditions and buffered with graphite. Run products were characterized by X-ray powder diffraction, backscattered and secondary electron images and microprobe analyses.

In the Al-enriched peridotite, the amphibole + chlorite bearing assemblage was found at 2.2 GPa, 700°C. At 2.0 GPa, 800°C the assemblage amphibole + garnet + orthopyroxene + olivine is stable. With increasing pressure chlorite was found with clinopyroxene + olivine + garnet at 4.2 GPa, 680°C. A Dense Hydrous Magnesium Silicate, the 10Å phase, appears with clinopyroxene + olivine + garnet at the expense of chlorite at 4.8 and 5.3 GPa (680°C) suggesting the H₂O-conserving reaction enstatite + chlorite = forsterite + pyrope + 10Å phase. The anhydrous assemblage olivine + garnet + orthopyroxene + clinopyroxene was found at 4.6 GPa, 750°C.

However, X-ray powder diffraction, BSE and SEM images, and microprobe analyses of homogeneous "10Å phase" domains (up to a few tens of µm large) which contain more than 0.5 atoms per formula unit of aluminium, indicate a mixed layered chlorite-10Å phase. The occurrence of a "mixed-layered 10Å phase" at pressures above chlorite stability promotes H₂O transfer to DHMS and H₂O transport to the deep mantle (> 200 km depth).

Hébert R, Bideau Dand Hekinian R, Earth and Planetary Science Letters, 65, 107-125, (1983).

Robinson JAC & Wood BJ, Earth and Planetary Science Letters, 164, 277-284, (1998).

Mysen BO & Boettcher AL, Journal of Petrology, 16, 520-548, (1975).

Attempt to Synthesize Grandidierite-Ominelite Solid Solution

Noboru Furukawa (furukawa@earth.s.chiba-u.ac.jp)

Department of earth sciences, Faculty of science, Chiba University, 1-33,Yayoi-cho,Inage-ku,Chiba-shi, CHIBA 263-8522 Japan

Recently, very high iron content Grandidierite(Ominelite, X_Fe = 0.93) were found out in Omine district, central Japan. I attempt to synthesize Grandidierite(Gdd)-Ominelite(Omi) solid solution. Standard cold-seal autoclaves and internal-heated pressure vessel were used. Starting materials were grandidierite(Gdd)(X_Fe=0) and ominelite(Omi)(X_Fe=1) composition gel or oxide mix, and Gdd₂₅-Omi₇₅(X_Fe=0.75), Gdd₅₀-Omi₅₀(X_Fe=0.5), Gdd₇₅-Omi₂₅(X_Fe=0.25) gels was prepared. Oxygen fugacity condition were close to NNO. Unfortunately, Ominelite synthesis has not been succeeded. Obtained most iron-rich Gdd-Omi solid solutions were X_Fe = 0.3 to 0.4, at 900C, 100 MPa and X_H2O was close to 0.1. T<700°C, Gdd-Omi solid solution are not synthesized and Tur + Crm ± Al₄B₂O₉ ± Mt ± Qtz was crystallize (X_Fe from 0 to 1). At higher temperature, other borosilicates or borate were crystallized (e.g. Werdingite, Hulsite, Al₈Si₂B₂O₁₉ (boron mullites), Al₄B₂O₉) and almost sample included magnetite. Therefore Ominelite synthesis will need low oxygen fugacity condition, so that Fe in Ominelite is all ferrous.

Olesch & Seifert, N Jb Mineral Monatshefte, 11, 513-518, (1976).

Werding and Schreyer, Boron Mineralogy, Petrology and Geochemistry (Mineral. Soc. Am.), 33, 139-141, (1996).

Measurements of Ferrous and Ferric Iron Activities in Silicate Melts: The Effect of fO₂ and Water

Fabrice Gaillard (gaillard@cnrs-orleans.fr), Michel Pichavant (pichavan@cnrs-orleans.fr) & Bruno Scaillet (bscaille@cnrs-orleans.fr)

ISTO, CRSCM, 1 A rue de la ferollerie, 45071 Orléans Cedex 2, France

An important data base of ferrous iron activities measurements covering simple, reduced and anhydrous basaltic to andesitic compositions has been established. As most Arc-derived magmas are oxidised and contain several wt% of H₂O, we propose to extend this data base toward oxidised and hydrated rhyolitic melts compositions. Rapid quench experiments were performed in a 1 atm vertical furnace and in IHPV from 1000 to 1200°C, 0.1 to 200 MPa and for fO₂ ranging from QFM-2 to NNO+6. Experiments consisted of equilibrating a Q-Ab-Or-FeO melt mixed with Ir-Fe powders. Iron was added to a silicate base in a concentration varying from 0.5 to 8.5 wt% FeO. Water was also added in different amounts from 0 to 6.5 wt%. After experiments, glasses, crystals and Ir-Fe alloys compositions were determined by EMPA. Ferrous iron is measured by wet chemistry and ferric iron by difference between total and ferrous iron. Water content in glasses was measured by FTIR. Iron activities are determined through the following equilibria: Fe^(Fe-Ir) + 1/2 O₂ <=> FeO^(liquid) 2 Fe^(Fe-Ir) + 3/2 O₂ <=> Fe₂O₃ ^(liquid) The constant of these equilibria are respectively taken from Coughlin (1954) and extrapolated from Haematite thermodynamic properties. In spite of the difference between our experimental strategy and the previous ones, an excellent accord is obtained. The method is also consistent with magnetite-liquid equilibria. Results show that (gamma)FeO follows an Henrian law for all experimented conditions. The Henrian constant is strongly dependant on Fe₂O₃/FeO. The temperature effect on (gamma)FeO seems weak as found for simple systems. Water addition slightly increase (gamma)FeO under reducing conditions and has the opposite effect for higher Fe₂O₃/FeO. Activity-composition relationships for Fe₂O₃^(liquid) are not linear and show a strong negative departure from ideality. (gamma)Fe₂O₃^(liquid) increases with increasing Fe₂O₃/FeO and decreases when water content increases.

Coughlin JP, Bureau of mines, contributions to the data on the theoretical metallurgy, 542, 1-78, (1954).