Breakdown of hydrous ringwoodite to pyroxene and spinelloid at high P and T and oxidizing conditions

Abstract

To get deeper insight into the phase relations in the end-member system Fe2SiO4 and in the system (Fe, Mg)2SiO4 experiments were performed in a multi-anvil apparatus at 7 and 13 GPa and 1,000–1,200°C as a function of oxygen fugacity. The oxygen fugacity was varied using the solid oxygen buffer systems Fe/FeO, quartz–fayalite–magnetite, MtW and Ni/NiO. The run products were characterized by electron microprobe, Raman- and FTIR-spectroscopy, X-ray powder diffraction and transmission electron microscopy. At fO2 corresponding to Ni/NiO Fe-ringwoodite transforms to ferrosilite and spinelloid according to the reaction: 9 Fe2SiO4 + O2 = 6 FeSiO3 + 5 Fe2.40Si0.60O4. Refinement of site occupancies in combination with stoichiometric Fe3+ calculations show that 32% of the total Fe is incorporated as Fe3+ according to \( \left ( {{\text{Fe}}_{1.60}^{2 + } {\text{Fe}}_{0.40}^{3 + } } \right)^{\text{VI}} \left ( {{\text{Si}}_{0.60} {\text{Fe}}_{0.40}^{3 + } } \right)^{\text{IV}} {\text{O}}_{4} . \) From the Rietveld refinement we identified spl as spinelloid III (isostructural with wadsleyite) and/or spinelloid V. As we used water in excess in the experiments the run products were also analyzed for structural water incorporation. Adding Mg to the system increases the stability field of ringwoodite to higher oxygen fugacity and the spinel structure seems to accept higher Fe3+ but also water concentrations that may be linked. At oxygen fugacity corresponding to MtW conditions similar phase relations in respect to the breakdown reaction in the Fe-end-member system were observed but with a strong fractionation of Fe into spl and Mg into coexisting cpx. Thus, through this strong fractionation it is possible to stabilize very Fe-rich wadsleyite with considerable Fe3+ concentrations even at an intermediate Fe–Mg bulk composition: assuming constant K D independent on composition and a bulk composition of x Fe = 0.44 this fractionation would stabilize spl with x Fe = 0.72. Thus, spl could be a potential Fe3+ bearing phase at P–T conditions of the transition zone but because of the oxidizing conditions and the Fe-rich bulk composition needed one would expect it more in subduction zone environments than in the transition zone in senso stricto.