Phase relations of Fe-Mg spinels including new high-pressure post-spinel phases and implications for natural samples

Phase relations of magnesioferrite-magnetite solid solutions (Mg, Fe2+) Fe23+O4 were investigated at pressures of 9-23 GPa and temperatures of 1200-1600 degrees C. Our new results indicate that the assemblage Mg2Fe2O5 + Fe2O3 reconstitutes to a hp-MgFe2O4 phase at 20 GPa and 1300-1500 degrees C. The stability field of hp-MgFe2O4 begins at similar to 1300 degrees C and widens to higher temperature. At lower temperature (1200-1300 degrees C) Mg2Fe2O5 + Fe2O3 breaks down to the new phase assemblage Mg3Fe4O9 + Fe2O3 with its stability field expanding to higher pressures and temperatures at the expense of hp-MgFe2O4. The Mg3Fe4O9 phase has the same crystal structure that recently reported for Fe7O9, and thus represents the Mg-end-member. From powder X-ray diffraction, we find that hp-MgFe2O4 has a structure consistent with an orthorhombic unit cell belonging to the Pmcn space group (no. 62). However, it could have undergone a transformation from a different structure during decompression. Experiments conducted with a Mg0.5Fe0.52+Fe23+O4 composition demonstrate that the addition of Fe2+ significantly changes the topology of the phase relations compared to the MgFe2O4 end-member system. At 10-11 GPa and 1000-1600 degrees C, Mg0.5Fe0.5Fe2O4 breaks down to the assemblage MgFeFe2O5 + Fe2O3, with the phase boundary described by: P (GPa) = 2.0 x 10(-3) x T (degrees C) + 8.2. No stability field for the constituent oxides [i.e., (Mg, Fe) O + Fe2O3] exists, in contrast to that observed for the MgFe2O4 end-member. The stability of the assemblage MgFe2+ Fe23+O5 + Fe2O3 is limited at higher pressures and appears to pinch out to higher temperatures. At 15-16 GPa and temperatures up to 1350 degrees C, this assemblage reconstitutes to form a hp-Mg0.5Fe0.5Fe2O4 phase. However, at higher temperatures a new assemblage of (Mg, Fe)(3)Fe4O9 + Fe2O3 appears. The occurrence of such compositions suggests that solid solution may be complete across the Mg3Fe4O9-Fe7O9 binary. Our results further demonstrate that phase relations even in simple Fe-Mg oxides can become complex at high pressures and temperatures and that phases with various novel stoichiometries (i.e., Mg3Fe4O9) may become stable. In addition, this study has implications for natural samples by helping to place constraints on the range in pressure and temperature at which a given sample formed. For instance, magnetite or magnesioferrite entrapped as inclusions in diamond could have either have crystallized directly, or formed from precursor phases at depths that exceed the stability of the spinel-structured phases. Evidence for such high-pressure transformations can potentially be found by investigating micro-textures.