Composition and Pressure Effects on Partitioning of Ferrous Iron in Iron-Rich Lower Mantle Heterogeneities

Round 1

Reviewer 1 Report

In the manuscript the partitioning of iron between coexisting lower mantle phases - magnesiowüstite, bridgmanite and post-perovskite was studied at lower mantle conditions ranging from 33-128 GPa and 1900-3000 K in the laser-heated diamond anvil cell. Partitioning results are used to model compositions and densities of mantle phase assemblages as a function of pressure, FeO-content and SiO₂-content. Since the dependence of the exchange coefficient on the bulk iron content is the key to the modern composition of dense heterogeneities of the mantle, the experimental studies presented in this work are undoubtedly relevant. Based on the experimental data obtained in this study, an important conclusion for mantle petrology was drawn: iron-rich compositions in the mantle exhibit negative dependence of density on SiO₂-content at all mantle depths.

 

The article is well structured and written clearly. The article is well structured and written clearly. Undoubtedly of interest to researchers of the Earth's mantle and deserves to be published. There are no serious comments on the text of the manuscript.

In the text of the manuscript, some references to the literature used are lost, probably due to the use of the reference manager (lines 188б 205, 208, 215-216 and many others further down to the end of the text).

 

Specific comments:

Lines 204-205: “a single peak attributed to CaIrO3-type post-perovskite (Supplementary Figure S4)”. The figure S4 does not show the peak related to postperovskite of the CaIrO3 type. In any case, none of the peaks is indicated in the figure as pPv.

Figure 2: Why are copper and gallium peaks in the EDS spectra? Cu intensity comes from the copper grid? It is necessary to clarify.

Line 230: “In one sample Fe metal droplets as small as 10s nm…” In the Figure S5, the metal droplets still look larger than 10 nm. Based on the scale bar indicated in the figure, their size is about 30-40 nm and more.

Figure 5: In the previous figures Mws=magnesiowüstite, Sti=stishovite, but in the figure 5 – Mw and St. Please make it uniform.

In section 2. Materials and Methods says that chemical analyses were obtained by electron microscopy. But, Table S1 - Microprobe results for compositions of synthetic olivine starting materials. Why microprobe technique is not listed in methods?

Table S2 is missing in Supplementary Material.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

In the present study, Dorfman and co-authors present new data on iron partitioning between Bridgmanite and Magnesiowüstite, two main mineral phases of the deep mantle. While previous studies have focused mainly San Carlos olivine or a pyrolite composition, here they studied Fe-rich compositions under pressure and temperature conditions of the entire Earth's mantle.

Experiments were performed in Laser Heated Diamond Anvil Cell with both in situ X-ray diffraction and ex situ Transmission Electron Microscopy analyses. This is a nice study, that brings new light on the fate of iron in the deep Earth, as well as on deep mantle heterogeneities .

Since the discussion of the paper is based on few hundredths variation of the KD, I wonder if the author could describe better the EDX measurements in TEM. In particular, have the authors used any K factors collected on references samples in similar conditions ? I understood that they evaluate the error base on the standard deviation between multiple grains. Is it possible that they under evaluate their error bar by the lack of K factors ? Is it possible that this is at the origin of their discrepancy with previous studies such as Tange et al., 2009 ?

 

Author Response

Thank you to the reviewer for the positive evaluation. Yes, the studies referenced from similar conditions, e.g. Auzende at al. 2008 and Sakai et al. 2009, 2010 employ EDX measurements in TEM. When you look at our Fig. 4a, our error bars do appear significantly smaller than those on these previous studies. However, we should also note that since KD is derived from the product of ratios of compositions, the size of the error bars scales with KD. When the oxide phase is very Fe-rich and KD is very small, the same precision on composition measurements makes the error bar look smaller on a linear KD plot.

As we noted in lines 195-196, error bars were computed based on precision of EDX measurements. There are other potential sources of error discussed in the manuscript such as pressure calibration,  nonhydrostatic stress, temperature calibration, temperature differences between studies, and temperature gradients. These sources of uncertainty are not at a level to affect the large-scale findings including the thermodynamic model and model of density of Fe-rich assemblages. But to address the reviewer’s question about comparison to observed compositions of Fe-rich bridgmanite from Tange et al., it’s likely that differences in temperature, run time, or redox state could also contribute to the discrepancy between observations, as well as uncertainty in EDX (see lines 266-268).