Search

In Extended Data Table 1 of this Letter, some of the elastic constants were reported incorrectly. This occurred as a result of an error in the script used to generate the numbers. The values of the elastic constants at room pressure cited in the manuscript on page 544 were derived using the same erroneous script, and the correct values and 1σ-uncertainties in the last given digit are C11 = 461.3(17) GPa instead of 462.7(17) GPa; C22 = 509.7(26) GPa instead of 504.9(26) GPa; C33 = 425.7(5) GPa instead of 426.6(5) GPa; C44 = 188.8(6) GPa instead of 188.4(6) GPa; C55 = 166.5(4) GPa instead of 166.6(4) GPa; C66 = 127.2(17) GPa instead of 129.7(17) GPa; C12 = 141.7(14) GPa instead of 140.2(14) GPa; C13 = 130.0(11) GPa instead of 132.2(11) GPa; and C23 = 161.0(12) GPa instead of 159.3(12) GPa. These errors do not affect any of the conclusions and we apologize for any confusion this may have caused. Extended Data Table 1 and the room-pressure values in the text have been corrected online. The Supplementary Information of this Author Correction contains the original, incorrect Extended Data Table 1, for transparency.

The high-pressure phases Fe4O5 and Fe506 have recently been added to the list of known iron oxides. As mixed -valence phases, it has been suggested that they could form in the Earth's mantle once the dominant minerals become saturated in ferric iron. The possibility that Fe4O5 could exist in the mantle is also supported by the fact that it forms extensive solid solutions with both Mg2+ and Cr3+. In this study, we present the results of high-pressure and high-temperature multi -anvil experiments performed between 5 and 24 GPa at 1000-1400 degrees C aimed at constraining the stability field of the Fe4O5 phase. We combine these results with published phase equilibria, equation of state and Fe Mg partitioning data to estimate the thermodynamic properties of Fe4O5, Fe5O6 and the (Mg,Fe)(2)Fe2O5 solid solution. Using our thermodynamic model, the oxygen fugacity at which the high-pressure iron oxides become stable is calculated and the redox stability of (Mg,Fe)(2)Fe2O5 in an assemblage of olivine and pyroxene is calculated as a function of the bulk Fe/(Fe + Mg) ratio. Fe4O5 and (Mg,Fe)(2)Fe2O5 are stable at oxygen fugacities higher than the diamond stability field and are, therefore, unlikely to be found as inclusions in diamonds. The stability field of Fe5O6, on the other hand, extends to oxygen fugacities compatible with diamond formation. Using the Mg Fe solid solution model, we show that Fe4O5-structured phases would be restricted to aluminium -poor environments in the mantle such as dunites or silica iron oxide rich sediments transported into the mantle via subduction.

The crystal structure of synthetic Ca2Sb2O7 has been investigated by in situ X-ray powder diffraction methods within the temperature range 298-1273 K. Ca2Sb2O7 is orthorhombic, space group Imma, over the entire temperature range and shows only general thermal expansion. The unit-cell parameters show a monotonous increase as a function of temperature. Only slight changes occur in the individual bond distances without evidence of abrupt structural changes. The mean coefficients of thermal expansion along the various axes are: alpha(a) = 9.49 x 10(-6)K(-1), alpha(b) = 10.11 x 10(-6) K-1, alpha(c) = 7.56 x 10(-6) K-1 and the bulk thermal expansion coefficient alpha(v) = 27.20 x 10(-6) K-1.