Magnetic Properties and Redox State of Impact Glasses: A Review and New Case Studies from Siberia

Round 1

Reviewer 1 Report

This work is dedicated to the study of impact glass as a result of high-speed impacts. The following are presented as objectives: 1. Study of the magnetic properties of several types of impact glass found in Siberia, making a total of approximately 30 new analysed samples; 2 - Presentation of new results; 3 - Revision of the oxidation state and petrography of the magnetic minerals present in the impact glass; 4 - Evaluation of the potential applications of impact glasses in the studies of paleomagnetism, magnetic anomalies. 

With this work authors obtain/ reinforce some interesting conclusions, namely:

1. Purely paramagnetic behaviour related to the redox state, essentially restricted to Fe2+;

2. Magnetite appears as the main magnetic carrier, its typical size being within superparamagnetic and/or near mono-domain state;

3. Definition of petrogenic conditions including redox, discrimination with glasses from other regions, paleomagnetism and magnetic anomalies.

The work is generally well written and organized, although a final reading is suggested. The authors were very thorough in the information provided regarding methods and results. The interpretation of the results is careful and detailed enough. The number of figures and tables is appropriate.

Finally, this work contributes to a better perception of the usefulness and limits of the use of magnetic properties for the characterization of impact glass. I recommend its publication after minor revisions.

Some specific comments

Figure 2. The overlap of the various colours brings some uncertainty regarding the amplitude of the standard deviation. Consider adopting an alternative, for example, an indication of the mean and respective error bars for the standard deviation ate the right side of the graph.

Figure 3. Nothing is mentioned relative to graph f)

Figure 9. Part of the right side of the figures is omitted and / or badly formatted. Figure 9c is never mentioned (line 365?)

Figure 10b. The figure display is badly formatted.

Table 3. Header is badly formatted.

Section 4.2.2. It is mentioned a number of 27 analysed samples, which does not correspond to value shown in table 2.

Line 347. Give the meaning of FWHM

Author Response

response in italics

Figure 2. The overlap of the various colours brings some uncertainty regarding the amplitude of the standard deviation. Consider adopting an alternative, for example, an indication of the mean and respective error bars for the standard deviation ate the right side of the graph.

 modified

Figure 3. Nothing is mentioned relative to graph f)

 modified

Figure 9. Part of the right side of the figures is omitted and / or badly formatted. Figure 9c is never mentioned (line 365?)

 modified

Figure 10b. The figure display is badly formatted.

  modified

Table 3. Header is badly formatted.

 modified

Section 4.2.2. It is mentioned a number of 27 analysed samples, which does not correspond to value shown in table 2.

 for Popigai only samples with hysteresis data are reported in table 2 (precised in text).

Line 347. Give the meaning of FWHM

 done

Reviewer 2 Report

Review for manuscript ‘Magnetic properties and redox state of impact glasses: a review and new case studies from Siberia’ geosciences-463067

By Pierre Rochette *, Natalia S. Bezaeva *, Andrei Kosterov, Jérôme Gattacceca, Victor L. Masaitis, Dmitry D. Badyukov, Gabriele Giuli, Giovanni O. Lepore, Pierre Beck. 

Rochette et al present new detailed data on glasses from Siberia and a review magnetic properties of ejecta glasses. I am not an expert in these rock, their magnetic properties, or the research/literature which preceded this study. Therefore I conducted this review with the aim of ensuring a robust data set is presented and that it is well presented. I suggest an expert in magnetic properties of these peculiar rocks needs to also view the manuscript. 

I have annotated the PDF(roughly) to indicate where there are problems. I have assumed that the reference list has already bene thoroughly checked for consistency by the journal. 

Overall the manuscript is easy to read and well written. In a few places it degrades and becomes what feels like a collection bullet points. There are also odd spacing issued particularly with the ‘X’. I assume these will be taken care of when the manuscript is typeset. 

The authors should think carefully about whether to set the manuscript in present or past tense. At the moment it is in both. I have marked a few instances where I found them. 

The figures are inconsistent in terms of units, style and some have been cropped and are missing data/labels. 

If hysteresis loops do not close (such as in figure 3) then the results are compromised. Check the results of 3a for example. With these noisy data the Hc estimates will be poor and in all likely hood wrong. Likewise Mrs data derived form IRM will be noisy too and therefore the Mrs/Ms ratio presented in table 2 will be compromised.

It should be possible to measure FORCs on these samples? Has this been done? It would be insightful especially for samples 3c. 

Where are the IRM curves from which the Mrs was determined? We need to see these in a figure as well. 

How do you determine the ppm concentration form an Ms value. This needs to be described more clearly… or its already there and I missed it.  

Below are point by point comments on major issues:

Fig1. Upper case -9 in susceptibility

Fig3. a the loop does not close. Remeasure this. Samples c and d are reversed in the text. 

Line 249-251. This needs more explaining. how does 10Be fit in? It is a bit out the blue. 

Line 289-312. This block of text is messy and difficult to read. I suggest reworking it or compressing it into discrete sections.

Line 309. you need a reference or some evidence before stating metastable phase of defects. 

Fig4. Add samples labels to the curves. This will make it easier to follow. The hump in 4(a?) at 220°C looks to be as significant as the one labeled ta 320°C. what is it? 

Please make the style of these the same as other figures including consistent symbol sizing and style. Upper aces -6 in the axes labels)

Fig 5. Make figure style consistent with other figures.

Fig 7. Elaborate more in the figure caption. What does this figure show us? What does it all mean. 

line 342. are these spectra (plural)?

Line 361-366 is too long. Compress this. 

Fig 9. Make figure style consistent with the others. Expand the figure caption. The figure is cut off and data are missing. 

Lin 398. K/Ar dating is often suspect as well. 

Fig 10. Make figure style consistent with the others. The figure is cut off and data are missing. Expand the figure caption…. what do the low temp data tell us?

Table 5. Where are the km? Are these in brackets?

fig 12. This is not a histogram… its a bunch of coloured curves… what does the figure tell us.. what does it relate to and what does it mean!?! This needs a better caption.

Line 506. How do superparamagnetic grains affect an NRM. By definition these do not contribute to the NRM. Perhaps there is something I missed here. Please tell me or elaborate in the manuscript. 

Comments for author File: Comments.pdf

Author Response

point in annotated pdf have been corrected in the new ms.

response in italics:

The figures are inconsistent in terms of units, style and some have been cropped and are missing data/labels. 

Modified. Concerning the use of both Kelvin and Celsius, this is common practice in rock magnetism to report low and high temperature in these two units, respectively.

If hysteresis loops do not close (such as in figure 3) then the results are compromised. Check the results of 3a for example. With these noisy data the Hc estimates will be poor and in all likely hood wrong. Likewise Mrs data derived form IRM will be noisy too and therefore the Mrs/Ms ratio presented in table 2 will be compromised.

This is the case only for fig.3a, with a very low Ms. We indeed do not place much confidence on the hysteresis parameters of this sample (not tabulated but just used to roughly estimate magnetite amount)

It should be possible to measure FORCs on these samples? Has this been done? It would be insightful especially for samples 3c. 

We agree. However, this is a review where we briefly report the Popigai samples, that deserve to be more thoroughly studied in a forthcoming paper.

Where are the IRM curves from which the Mrs was determined? We need to see these in a figure as well. 

Well again we feel such details do not belong to the present review.

How do you determine the ppm concentration form an Ms value. This needs to be described more clearly… or its already there and I missed it.  

Explanation (Ms value for pure magnetite and metal) added in text

Below are point by point comments on major issues:

Fig1. Upper case -9 in susceptibility

modified

Fig3. a the loop does not close. Remeasure this.

the data presented is already a stack of 5 loops; Ms is only a few percent of total magnetization due to paramagnetism. There is no way to make it better (I have been using this VSM for over twenty years on thousands of mostly paramagnetic samples) ! Note added in caption: (note that strong paramagnetism prevented to obtain a well defined closed loop)

Samples c and d are reversed in the text. 

 Text modified

Line 249-251. This needs more explaining. how does 10Be fit in? It is a bit out the blue. 

It is just to cite the paper about these microtektites; 10Be mention has been cancelled.

Line 289-312. This block of text is messy and difficult to read. I suggest reworking it or compressing it into discrete sections.

We separated several paragraphs. The point is that the magnetic mineralogy of these samples is very peculiar and deserves more investigations to be fully understood. That is why may be you found the message unclear.

We previously wrote in the caption “The magnetic mineralogy of Popigai glasses is obviously complex (see Fig.4b with suggestion of up to 4 different phase transition) and deserves further dedicated investigations.”. Now this cautionary point is moved to text.

Line 309. you need a reference or some evidence before stating metastable phase of defects. 

Reference to martian meteorite (with similar problematic) added

Fig4. Add samples labels to the curves. This will make it easier to follow. The hump in 4(a?) at 220°C looks to be as significant as the one labeled ta 320°C. what is it? 

We don’t know! See above cautionary note. To be conservative in this sample the 220 and 320 humps are hardly significant (as stated in text)

Please make the style of these the same as other figures including consistent symbol sizing and style. Upper aces -6 in the axes labels)

modified

Fig 5. Make figure style consistent with other figures.

Fig 7. Elaborate more in the figure caption. What does this figure show us? What does it all mean. 

Caption developed. Meaning is described in text

line 342. are these spectra (plural)?

modified

Line 361-366 is too long. Compress this. 

We disagree, 7 lines to fully report on the magnetic properties of a special glass that was never reported before seems not that long.

Fig 9. Make figure style consistent with the others. Expand the figure caption. The figure is cut off and data are missing. 

modified

Lin 398. K/Ar dating is often suspect as well. 

We agree

Fig 10. Make figure style consistent with the others. The figure is cut off and data are missing. Expand the figure caption…. what do the low temp data tell us?

modified

Table 5. Where are the km? Are these in brackets?

Yes (added in caption).

fig 12. This is not a histogram… its a bunch of coloured curves… what does the figure tell us.. what does it relate to and what does it mean!?! This needs a better caption.

Implemented (sorry there was an error in figure order). To be able to put the four distributions on the same graph we prefer not to use the standard histogram convention so call it distribution. But the abundance is a number per bin, normalized to maximum value.

Line 506. How do superparamagnetic grains affect an NRM. By definition these do not contribute to the NRM. Perhaps there is something I missed here. Please tell me or elaborate in the manuscript. 

OK modified to “However, high proportion of grains near the superparamagnetic threshold can lead to unreliable NRM record”

Reviewer 3 Report

This paper provides a summary review of literature data on magnetic susceptibility and remanence in impact glasses, supplemented by new measurements. The authors’ main conclusion is that these impact glasses can be divided in 2 groups, showing predominantly paramagnetic behavior on the one side, and ferromagnetic behavior on the other. They also discuss possible explanations for this bimodal distribution in magnetic properties; however, it does not seem straightforward which glasses are paramagnetic or ferromagnetic. The data presented is interesting, however given the review character of this study, the new data needs to be thoroughly integrated with existing data.

The following suggestions would help to increase the review character of the paper:

1. Please introduce all types of glasses with their definition and main characteristics. This could be part of the introduction text, which already introduces tectites, but does not provide a comprehensive and systematic overview of all types of impact glasses. Alternatively, it could be added as a table.

2. Provide a figure of the glasses studied and described in this paper, showing their geographic distribution, and potentially also other defining characteristics, such as their position in/distance from the impact crater.

3. Can you provide a maximum grain size for these glasses?

4. Histograms of susceptibility and remanence, or a scatter plot of susceptibility vs remanence for all samples from this study as well as previous studies would help the reader to (1) compare measurements made in different studies on similar samples, (2) compare and distinguish between properties of different glasses, and (3) support the author’s statement of a bimodal distribution in susceptibility.

5. In the Methods section, it seems that different types of measurements have been made on different groups of samples. Can you explain how the methods to be applied were chosen? I think part of the explanation is given in the Results section, but it belongs to the Methods. It would also be helpful to add a table showing which data has been measured for which group of samples.

6. The current Results section includes both Methods (and justifications why specific methods have been used), Results (i.e. observations) and Discussion (i.e. interpretation of own data, and comparison to other data/interpretation of differences to previous data). Please rearrange to the relevant sections of the paper, or rename this section to Results and Discussion. In the latter case, the section now called Discussion could be renamed to Implications for Formation Processes.

7. Please discuss issues that are specific to these samples, such as the small sample size, in terms of representativeness (how representative are samples of 1-10 mg as mentioned in L84?), and the need to have accurate mass normalization (my experience with the SM150 is that repeat measurements of sample mass differ by up to 5%, resulting in errors in the mass-normalized susceptibility – has this been checked?)

8. Chapter 4.1 is called ‘essentially paramagnetic glass’, but then goes on to describe titanomagnetite. Please rename or add sub-titles.

9. On Fig 1, please add the results from previous studies, either as points, or as susceptibility ranges if the Fe content is not known.

10. Table 1: Please show this data also in a figure

11. Based on your own measurements, and the review of literature data, can you provide a set of quality criteria for magnetic measurements on impact glasses?

Author Response

The following suggestions would help to increase the review character of the paper:

1. Please introduce all types of glasses with their definition and main characteristics. This could be part of the introduction text, which already introduces tectites, but does not provide a comprehensive and systematic overview of all types of impact glasses. Alternatively, it could be added as a table.

We have provided a general definition of impact glasses, and then described tektites. Other varieties of impact glasses are also evoked (the ones ejected ballistically and the ones injected in the target or mixed in the impactites). We feel adding a table would be indigestible (there is already 3 tables in the paper) and that categorizing the range of impact glasses is a bit arbitrary.

2. Provide a figure of the glasses studied and described in this paper, showing their geographic distribution, and potentially also other defining characteristics, such as their position in/distance from the impact crater.

World map added. Distance to impact crater is in most case not mapable (but see tektites).

3. Can you provide a maximum grain size for these glasses?

That’s a hard task that may lead to lengthy discussion! This is a review on magnetic properties, not a general review on impact glasses! We added in text: Typical size for ballistically ejected glasses is centimetric, with maximum size hardly exceeding 10 cm.

4. Histograms of susceptibility and remanence, or a scatter plot of susceptibility vs remanence for all samples from this study as well as previous studies would help the reader to (1) compare measurements made in different studies on similar samples, (2) compare and distinguish between properties of different glasses, and (3) support the author’s statement of a bimodal distribution in susceptibility.

The data on the 24 different glasses we have is very heterogeneous; some glass have 3000 samples, some only 3; we feel that plotting all data like this will be hardly legible and not very useful. Partial data is plotted in various figures (2,3, 12) .

In the Methods section, it seems that different types of measurements have been made on different groups of samples. Can you explain how the methods to be applied were chosen? I think part of the explanation is given in the Results section, but it belongs to the Methods. It would also be helpful to add a table showing which data has been measured for which group of samples.

Again we feel this is too complex. The data we report has been acquired by the cited authors or by us on several decades and thus with a large variety of techniques. 

6. The current Results section includes both Methods (and justifications why specific methods have been used), Results (i.e. observations) and Discussion (i.e. interpretation of own data, and comparison to other data/interpretation of differences to previous data). Please rearrange to the relevant sections of the paper, or rename this section to Results and Discussion. In the latter case, the section now called Discussion could be renamed to Implications for Formation Processes.

OK done

7. Please discuss issues that are specific to these samples, such as the small sample size, in terms of representativeness (how representative are samples of 1-10 mg as mentioned in L84?), and the need to have accurate mass normalization (my experience with the SM150 is that repeat measurements of sample mass differ by up to 5%, resulting in errors in the mass-normalized susceptibility – has this been checked?)

yes we made reproducibility tests. Also we measured the same batch of tektites (about 200) using both SM150 and MFK1. SM150 is not that bad (dispersion only slightly larger). About small masses it’s implicit in our text that large masses are more representative: “measure… on much larger i.e. more representative samples ». Still we need to mention literature data, which seems representative for tektite even for 10 mg samples, due to the very homogeneous nature of these paramagnetic glasses.

8. Chapter 4.1 is called ‘essentially paramagnetic glass’, but then goes on to describe titanomagnetite. Please rename or add sub-titles.

By essentially paramagnetic we mean that >95% of susceptibility is paramagnetic. But ferromagnetic impurities can be detected using remanence, or in rare anomalous samples.

9. On Fig 1, please add the results from previous studies, either as points, or as susceptibility ranges if the Fe content is not known.

We feel the message of this figure (now 2) will be blurred by doing this. Using a consistent dataset (obtained with the same instrument and measurement session) ensure that no measurement bias is responsible for the tendencies observed.  Table 1 allows comparison of our data with previous studies. We feel our data, being cross-calibrated and obtained on larger dataset and with modern instruments is more robust.

10. Table 1: Please show this data also in a figure

this data is partly represented in Fig.2, 3, 12 (new numbering).

11. Based on your own measurements, and the review of literature data, can you provide a set of quality criteria for magnetic measurements on impact glasses?

Difficult request. We do not feel at ease making such general statements as it depends on each case study. Making general statements such as “one must check heterogeneity using large number of samples, adapt sample mass to sensitivity, be careful about sample holder, etc., does not seem very useful.”