German Title: Neue experimentelle Daten zur Rekalibrierung der Fe-Ti-Oxid Thermo-Oxybarometer
Despite their modest modal abundances, Fe-Ti oxides are important accessory minerals in igneous rocks. They are major carriers of rock magnetism and petrologic indicators of temperature and redox conditions during the magmatic stage (Buddington & Lindsley, 1964). The titanomagnetite-ilmenitess (Tmt-Ilmss) thermo-oxybarometer has been widely used to retrieve information, e.g. on redox states in the Earth’s mantle, processes in magma chambers, or the crystallisation conditions of lunar or martian basalts. Current thermodynamic formulations of the Tmt-Ilmss thermo-oxybarometer (Andersen & Lindsley, 1988; Ghiorso & Sack, 1991) fail to reproduce experimental results, in particular at the high T relevant for basaltic assemblages. This is shown by applying the formulations to experimental literature data. The shortcomings of the existing models are partly due to the limited T-fO2-range of experiments used for the original calibrations. The effect of small contents of Mg, Al, etc. typical for natural Tmt and Ilmss was only partly evaluated, and non-stoichiometry of the phases (in particular Tmt) was not considered at all. To support a revised version of the Tmt-Ilmss thermo-oxybarometer we have synthesised Tmt+Ilmss assemblages in the system Fe-Ti-Mg-Al-O at 1 bar, 1000-1300°C and under a large range of fO2 (NNO-5 to +5). A large data set has also been provided for Ilmss coexisting with pseudobrookitess (Psbss). This forms the basis for an experimental calibration of a thermodynamic Ilmss-Psbss (thermo-)oxybarometer model. The data base on Tmt coexisting with Ilmss used to calibrate the previous models has now been extended towards higher T (i.e. 1000-1300°C) and to a larger range of fO2. The compositions of coexisting Fe-Ti oxides in the system Fe-Ti-O have been determined by EMP as a function of T and fO2. The experimental data derived within this study has been compared to the existing models. The structural transition within the ilmenite-hematite solid solution series causes an abrupt change of the slope of the experimentally determined isotherms, which gets increasingly pronounced with decreasing T. Andersen & Lindsley (1988) did not account for the transition, and consequently their model isotherms do not show the related bend. Ghiorso & Sack (1991) in contrast did account for the Ilm-Hem transition, however their model heavily relies on old experimental data involving sample compositions that were not controlled by microanalytical methods. Synthetic Tmt compositions are generally Ti-richer than predicted from the models. At low fO2 and high T, the discrepancies between synthetic Tmt data and the models are presumably due to non-stoichiometric (i.e. cation vacant) Tmt. Tmt non-stoichiometry has been investigated in the system Fe-Ti-O at 1100-1300°C for Tmt in assemblages in which maximum (Tmt+Ilmss) versus minimum (Tmt+Wus, Tmt+Fe) vacancy concentration (v.c.) is expected. Three methods have been used to detect and quantify v.c.: (1) Because cation deficiency involves a decrease of Tmt cell dimensions at constant Ti/(Ti+Fe), lattice parameters have been determined by Rietveld analysis. (2) At constant Ti/(Ti+Fe), cation vacancies are associated with higher Fe3+/Fetot. The latter have been determined by electron energy-loss spectroscopy on selected samples. (3) Annealing of a non-stoichiometric high-T Tmt at lower T (950°C) under oxygen-conserving conditions induces vacancy relaxation. This comes along with compositional changes in the phases, which result in the formation of Ilmss exsolution lamellae and rims in Tmt, i.e. an increase in the Ilmss proportion. The results of the three approaches indicate that Tmt coexisting with Ilmss has only low v.c. or is very close to stoichiometry at XUsp<0.7 at all investigated T. At XUsp>0.7, v.c. increases with increasing XUsp and T to about 2 cat% for 1300°C at XUsp close to 1. At T<=1100°C, i.e. at T relevant for crystallisation conditions of basic rocks, Tmt is close to stoichiometry for all XUsp. As the revised version of the Tmt-Ilmss thermo-oxybarometer also will involve high-T calibration data, it is important to account for Tmt non-stoichiometry in its solid solution models. To approach natural compositions, Tmt+Ilmss assemblages have been produced in the system Fe-Ti-Mg-Al-O. Mg-Al bearing Fe-Ti oxide compositions have been projected onto the Ilm-Hem and Mag-Usp binaries. X’Usp and X’Ilm of Mg-Al bearing phases are in reasonable agreement with the system Fe-Ti-O. The experimental isotherms determined for the system Fe-Ti-O can in principle also be applied to Mg-Al bearing samples. This provides a new, preliminary and graphical version of the Tmt-Ilmss thermo-oxybarometer, which has been tested on experimental data by Toplis & Carroll (1995). Compared to the T-fO2 estimates derived with the previous models, the new, graphical thermo-oxybarometer brings significant improvements, especially in the T estimates.
|Supervisor:||Lattard, Prof. Dr. Dominique|
|Date of thesis defense:||29 May 2006|
|Date Deposited:||17 Jul 2006 11:46|
|Faculties / Institutes:||Fakultät für Chemie und Geowissenschaften > Institut für Geowissenschaften|
|Subjects:||550 Earth sciences|
|Uncontrolled Keywords:||Titanomagnetit , Fe-Ti-Oxide , Oxybarometer , Thermometer , Petrologie|