Evaluating the Role of Iron-Rich (Mg,Fe)O in Ultralow Velocity Zones

Round 1

Reviewer 1 Report

This manuscript reports on the equation of state of Mw94 and Mossbauer spectroscopy to characterize the spin state of Fe in Mw. This is combined with previous work on other Mw composition (by the same group) to calculate seismic velocities of aggregates of Mw, bridgmanite and calcium perovskite.  These results are compared to seismic observations of ULVZ using both forward and inverse modeling of aggregate elastic properties.  These results show that Fe rich Mw can reproduce seismic properties of ULVZ and thus this is a viable alternative to a partial melt model of ULVZs.

The paper is well written and provides interesting observations.  I actually reviewed a previous version of this manuscript for Geophysical Research Letters where I thought it was worthy of publication.  Unfortunately it seems that it did not make it through publication at GRL.   In any case I am pleased to see that author addressed my previous comments in this draft and think that this is both acceptable and appropriate for publication in Minerals

I have only one small comment which is that there are downward and upward triangles to indicate Voigt and Reuss bounds for the Coral Sea in figure 2b.   I assume (from table 2) that there are also corresponding values for South Atlantic ULVZ compositions.  It seems that these might as well be added to be consistent across the board.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

The study of Dobrosavljevic et al. is divided nto two parts: the high-pressure structural and magnetic investigation of magnesiowüstite at embient temperature, and the use of the obtained results for seismic modelling. For this second part, although not an expert myself, I fell that the authors know their stuff very well and present the results convincingly.

For the first part, however, I have certain objections. In particular:

(1) From the presented XRD patterns in the Supplement, I personally do not see the Bragg peak splitting, characteristic of the B1-rhombohedral transition. In the 34.5 GPa pattern for example, the assigned MWR(003) doublet could be the MWc(001) and the B1-NaCl(200) peak. Can this assignement be excluded?

(2) Claiming also that use of neon can bring down the B1-rhombohedral transition pressure by 10-15 GPa in (Mg,Fe)O compared to helium does not appear plausible. Numerous studies in the literature show the similarities between helium and neon investigations up to 40-50 GPa. Hence, the difference in the transition pressure (if there is a transition at all) should be explained with different arguments.

(3) The XRD peaks in the 65.4 GPa  XRD pattern are weak and vanishing. The SMS studies took place beyond these pressure values (90-120 GPa), where I can only assume that no XRD signal would be detectable anymore. I would like the authors to consider the possibility that they have caused pressure-induced disorder/amorphization in their sample at these pressures, and if so, how would this affect the interpretation of the SMS results?

(4) The incorporation of NaCl in the (Mg,Fe)O sample causes any effects in the structural/magnetic properties?

(5) Back-extrapolation of the volume of the rhombohedral (Mg,Fe)O phase at zero pressure should yield high uncertainties in the obtained values, thus creating a relatively large range of V0 values to be used in the subsequent modelling.

All in all, I recommend the paper for publication, but only after the aforementioned points have been resolved.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

The manuscript addresses the question of the nature of the ultra-low velocity zones (ULVZs) at the base of the lower mantle. They authors refine the room temperature equation of state of iron-rich magnesiowüstite and combine it with the literature data in order to verify the seismic signature of magnesiowüstite at the core-mantle boundary. Then they apply the inverse modeling in order to quantify the potential amount of magnesiowüstite in the two observed ULVZs.

The manuscript is well written and, although the proposed forward models extensively use extrapolations from measurements on slightly different compounds and/or phases, these extrapolations seem justified by the absence of the experimental data. I, therefore, recommend publication of the manuscript after addressing the points below.

The authors are trying to construct a quantitative model of the composition of the ULVZs together with the estimated uncertainties of these models. The authors calculate the sound velocities and densities of magnesiowüstite with errors propagated from the relevant experimental results. Yet the errors for calculated values should come not only from the experimental errors but also from the use of the parameters measured for different phases/compositions. For instance, the authors use V_D of the rhombohedral phase to constrain shear elastic behavior of cubic magnesiowüstite, use elastic parameters of (Mg₀₆Fe_0.94)O to constrain densities of Mw84, use temperature corrections for pure MgO to calculate the sound velocities at high temperature. What additional uncertainties arise from these parameters “adopted” from other phases? The authors should try to estimate this or if it is not possible to discuss in more detail the influence of these errors on sound velocities they found and on the  models. The authors found a very high mismatch between the elastic properties they measured with Ne pressure medium and He pressure medium (far above the error of the fits). Therefore, the authors stress the necessity of experiments in He pressure medium for the proper determination of equations of state. I see two potential problems here. Firstly, are the conditions at the base of the lower mantle (quasi-)hydrostatic? If so, secondly, are the measurements of elastic parameters for other phases, which they use for calculations (Br, CaPv, Mw78), performed in the He pressure medium? If not then how reliable are these parameters? The authors demonstrate by their own measurements that the difference between measurements in Ne and He of, for example, the bulk modulus can be as high as 30% (155 vs 197 GPa). The choose of priors on Bg concentration and prior windows (lines 277-282) looks artificial. It is not clear why for one of the ULVZs the authors try to keep pyrolitic proportions, whereas for another one they allow it to deviate from pyrolitic. Either the same constraint should be kept for the entire inverse modeling or much clearer explanation of the particular prior concentrations should be done.

Minor comments:

Line 88: helium gas at 25000 psi

Shouldn’t we use the International System of Units (SI)?

Line 93: Calibration was performed...

Calibration of what?

Line 106: Peaks for Mw94 and NaCl…

Where does NaCl come from? It was never mentioned in the description of the pressure chamber.

Figure 2: compositions with Mw concentrations ranging from 8.3% 264 (Mw94), 10.2% (Mw84), and 10.6% (Mw78) up to 50%

It is hard to understand from the picture how the picture is related to concentrations. It would help to add an additional axis that shows a change of Mw concentration (for a least one composition) or to add a supplementary figure (speed reduction vs Mw concentration).

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

The authors have answered all of my comments convincingly. I recommend the paper for publication as is.