Francois Farges (farges@uuniv-mlv.fr)1, Max Wilke (max@geo.uni.potsdam.de), Pierre-Emmanuel Petit (ppetit@esrf.fr)2 & Gordon E. Brown, Jr. (gordon@pangea.stanford.edu)3
1 Lab. des Géomatériaux, Unviersité de Marne la Vallée, 77454 Marne la Vallée cedex 2, France
2 ESRF, BP 220, 38042 Grenoble cedex, France
3 Dept. of Geology, Stanford University, Stanford, CA 94305-2115, USA
Fe K-edge pre-edge and XANES spectra for 35 crystalline Fe(II) and Fe(III)-minerals were collected under high resolutions at the ESRF Facility (Grenoble, France) on beamline ID26 using Si(220) double-crystal monochromator. In these compounds, Fe(II) sites are tetrahedral, square planar, trigonal bipyramidal, octahedral and hexahedral. For Fe(III), tetrahedral, trigonal bipyramidal and octahedral sites were investigated The pre-edge spectra show one feature ([4]Fe(III), [5]Fe(III)), 2 feature ([4]Fe(II), [5]Fe(II), [6]Fe(III)) and 3 feature ([6]Fe(II)). Combination of multiplets and l-DOS calculations are in agreement wit these observations. Consideration of spin-orbit interactions and crystal-field splitting (10q) of ~1 eV are able to reproduce accurately the measured pre-edges. The most robust pre-edge information in terms of Fe(III)/ Fe ratio is given by their centroid and total intensity (i.e., the sum of the integrated normalized height for the 1s -> 3d/4 p transitions). The pre-edge centroids for ferrous and ferric iron were measured absolutely in energy, at 7113.0(1) and 7114.4(1) eV, respectively, suggesting a separation between Fe(II) and Fe(III) ~1.4(1) eV. The pre-edge features increase in intensity with decreasing site centrosymmetry. Therefore, derivation of coordination information from these pre-edge intensity can be difficult. The splitting of the pre-edge feature decreases with lower Fe(II)-coordination, and from Fe(II) to Fe(III), in agreement with crystal-field theory. In order to measure the variation in pre-edge information with variable Fe(III)/ Fe, the pre-edge feature for 60 mixtures between Fe(II) and Fe(III) were investigated. The pre-edge information can vary quite linearly with Fe(III)/ Fe at a constant Fe-redox or Fe-coordination. However, non-linear variations are observed when both coordination and redox vary with Fe(III)/ Fe. Redox determinations can then be done at ±10 mol.%, provided that the site geometry for each redox state is constrained. Application of these methods to 12 minerals containing variable/unknown amounts of Fe(II)/Fe(III) (magnetite, vesuvianite, franklinite, rhodonite, etc.) supports these models.
Joerg Fechner (j.fechner@uni-koeln.de)1, Herbert Palme (palme@min.uni-koeln.de)1 & Sumit Chakraborty (sumit.chakraborty@rz.ruhr-uni-bochum.de)2
1 Institut fuer Mineralogie und Geochemie, Zuelpicher Str. 49b, 50674 Koeln, Germany
2 Institut fuer Mineralogie, Universitaetsstr. 150, 44780 Bochum, Germany
The two oxides SiO2 and MgO make up more than 80% of the mass of the Earth's mantle. The chemical activites of these two components in melts from the mantle of the Earth can be determined experimentally relative to a standard state following a relatively simple procedure.
Olivine (ol) and orthopyroxene (opx) are stable mantle minerals down to a depth of ~400 km and during melting, they coexist with a melt up to ~70% partial melting. The coexistence of these two minerals uniquely defines the activities of SiO2 and MgO relative to a given standard state. The activities of the two oxides can be determined by equilibrating ol and opx with Pd-metal (Chamberlin et al., 1994) and comparing the results with those from equilibration with pure MgO and SiO2 standards. The ratio (Si in Pd) ol-opx mix /(Si in Pd) pure SiO2 gives the activity of SiO2 in the ol-opx mix (i.e. mantle) at the conditions of the experiment (same procedure for MgO).
We have performed reversal experiments at temperatures ranging from 1000°C to 1400°C and at atmospheric pressure. At 1200°C and 1400°C we found SiO2 activities of 0.81 to 0.89 and MgO activities of 0.17 to 0.18. At lower temperatures equilibrium was not achieved. Calculated activities from these data, at 1 GPa (assuming molar volume independent of T and p) and for olivine compositions of fo90 (using interaction parameter of Kawasaki & Matsui 1983) are: 0.67 (1250°C) to 0.75 (1400°C) for SiO2 and 0.19 (1250°C) to 0.18 (1400°C) for MgO, respectively. Using melt composition data of Baker et al., 1994, activity coefficients for SiO2 and MgO in melts in equilibrium with fertile upper mantle peridotite MM3 were calculated to be 1.10 and 1.48 for SiO2 for melt fractions of 2% and 27% and 0.98 (2%) to 1.15 (27%) for MgO, respectively.
Further experiments at high pressues are planned to measure these activities directly.
Chamberlin L, Beckett JR & Stolper E, Contrib. Mineral. Petrol, 116, 169-181, (1994).
Kawasaki T & Matsui Y, Geochim. Cosmochim. Acta, 47, 1661-1679, (1983).
Baker MB & Stolper EM, Geochim. Cosmochim. Acta, 58, 2811-2827, (1994).
Index of EMPG VIII Volume
Further EMPG VIII Information
Index of the Journal of Conference Abstracts
Cambridge Publications Home Page