This paper presents U and Th concentration data and uncertainties for select magnetite and Al-spinel samples with a focus on applicability for (U-Th)/He thermochronology. The primary novel contribution of this study is quantifying the reproducibility between wet chemistry dissolution and laser ICPMS results for samples with different concentration levels of U and Th. In the process of completing the study, they test the impact of matrix matched standards for LA-ICPMS analyses. Overall, this study provides a very helpful and useful scientific contribution on our understanding of the analysis and systematics of U and Th in magnetite and spinel. These are very difficult analyses and the techniques are still in the early stages of becoming more widely applicable. Work like this helps push the method forward and has appeal beyond (U-Th)/He dating (for example, economic geology research is also interested in the trace element chemistry of magnetite and spinel and analytical methods).

Overall I think this manuscript makes an original contribution worthy of publication. Before it is ready to be published, I have several comments, suggestions, and questions for clarification. 

Specific comments:

1) More sample information should be provided. These tests were run on 2 natural magnetite samples, 1 natural spinel sample, and 2 magnetite synthetic samples. Magnetite grain habits and inclusions suites can vary significantly between samples. Spinel chemistry can vary significantly as well. It is very possible that different magnetite and spinel samples will have different behaviors in dissolution and/or different analytical challenges in terms of matrix effects and U and Th concentrations. The more these samples are characterized in terms of their crystal habit, age, zonation, inclusion suites and any other known geochemistry, the better for future comparison as more studies include more samples. Table 1 is helpful and 2.1.1 and 2.1.2 have some important background information.

1.  Either the main text or the appendix should include more documentation of the sample history and any known geochemical, mineralogy or petrologic characteristics.
2.  The study would also benefit from adding photographs of the samples before and after crushing.
3. The spinel sample says it is Mg 0.65, Fe 0.35 in Line 84 – how was this determined?
4. XRD determined the synthetic magnetite is 85% magnetite and 15% goethite. Why are the XRD results not included in the appendix?

2) All samples were powdered prior to analysis. Table 1 documents different powdered grain sizes. As far as I know, it is not common to powder aliquots before dissolution during routine (U-Th)/He analysis. A few questions:

1. Is there any evidence that powder grain size impacts U+Th recovery after wet chemistry dissolution? Or was there any observed relationship between powder grain size and laser ablation conditions (pit size, efficiency, matrix effects)?
2. The Issua sample is a mixture of magnetite, quartz, and actinolite. Does that mean these analyses included a mixture of these minerals or was the magnetite isolated (I assume not based on Table 1)? If it is a mixture, then what is the justification of using the sample to compare with other magnetite? Would such a mixture ever be used for (U-Th)/He analysis?
3. Is the recommendation of this paper that magnetite and spinel (U-Th)/He should powder samples after degassing and before dissolution? If not, then are the results here translatable to dissolving whole grains? What are the recommendations or warnings to people who may try to do this with whole grains (which is more common for U-Th/He analysis)?

3) A significant portion of the manuscript assesses the potential sources for data dispersion, but there is no discussion of the impact of inclusions or intergrown minerals on the results. One of the known issues with magnetite and spinel (and other opaque phases) is that internal inclusions can be present in unknown quantities and can contribute He, or U-Th-Sm, and/or not be fully dissolved, etc. Prior work tries to get around this by using microCT to screen for and avoid inclusions. Here, some of the samples are reported to include mineral phases other than magnetite (Issua and the synthetic magnetite) determined by XRD.

1. Were the Rocher Blanc magnetite tested for inclusions or intergrown minerals either by microCT or XRD? What about the Al_Spl?
2. For this study, how could intergrown phases or inclusions impact the dispersion in the data? How would this vary between the wet chemistry technique and LA-ICPMS? Please include a greater discussion on the possibility for these effects within discussion section. A recent study that showed the impact of inclusion in magnetite on He concentration is Hofmann et al., 2021 “Exposure dating of detrital magnetite using 3He enabled by microCT and calibration of the cosmogenic 3He production rate in magnetite” in GChron.

4) Spinel dissolution can be quite challenging. It would be very helpful to include in the appendix the exact procedure used for others to reference and reproduce. The text mentions that some spinel took multiple rounds of acid attack.

1.  Did the time it took to dissolve spinel trend with data accuracy or reproducibility? It would be very helpful to know if it impacts U and Th recovery or sample loss. If it doesn’t impact the data, that would be very comforting to document. If it does impact the data, it will be important to know. It seems that this study can address this question.

5) Many of these analyses are very low concentration and close to blank level. Blanks are not reported. Please add any blank or standard data to the main tables or appendix. Without blanks it is not possible to assess the measurements (were the blank corrected?) and without knowing the blanks reproducibility, it is not possible to propagate the full uncertainty on the measurements, which is central to the study.

Line Comments (some may be repetitive with the comments above):

45: “is very little soluble in minerals” should be corrected to “He is not very soluble in minerals” or “has low solubility in minerals”

53: Sentence starting “In addition, well characterized…” is clunky and should be rewritten.

75-85 (2.1.1.): I’m left wanting more information on the samples. Please include more details. Also, how was the spinel composition determined? Microprobe?

93: The natural samples and synthetic samples have different grains sizes after powdering. Does this difference in grain size make a difference in the analyses?

100: The samples were ground up before dissolution. Is this a requirement for dissolution? What mass was dissolved per aliquot? Is powdering samples reasonable for typical (U-Th)/He analysis or would it need to be modified?

125-129: Was there a trend in dissolution steps vs U+Th recovery for spinel? Does it affect the accuracy of the measurement? What are microbombs?

134: Spinel can contain variable amounts of Fe, Al, Mg, and Cr beyond what is listed here. The chemistry likely makes a difference in the way it dissolves and potentially could relate to U+Th concentrations. It’s beyond the scope of this study, but worth noting here that there is a solid solution to consider.

134: “The direct analyze” should be “The direct analysis”

135: Can these elements be removed via column chemistry? Do your results suggest that is an important step to avoid matrix effects?

159-163: This is a very interesting observation (that the powdered samples are 100x higher in U+Th than measured by Schwartz et al., 2020). You say it could be contaminated with U+Th. Is that during preparation? Or is it possible that powdering grains included a lot of inclusions that Schwartz et al 2020 avoided by CT scanning their grains prior to analysis? How much sample was powdered to produce the homogenous, enriched U+Th in this study? Was the same powder split and used to make the pellet for LA-ICPMS?

169: “The dispersion is more important for Th” – do you mean more important or larger?

171: Have you considered that the larger Th dispersion is due to Th falling out of solution? This is often a problem for wet chemistry analyses. Alternatively, Th wash out times on ICPMS can take significantly longer than U and other elements. Sometimes Th takes a long time to reach the detector compared to U and other elements. Could either of these issues be a possibility for the Th uncertainty?

Figure 2:  Th dispersion appears to be concentration dependent, but also sample dependent. Your IF-G sample appears to have the highest dispersion and lowest concentrations which makes sense with analytical limits on uncertainties. But it is also a sample with three intergrown minerals. Could some of the dispersion be due to heterogeneous mixtures or nugget effects? How much sample was homogenized and how large are the aliquots that were analyzed?

Line 206: What does “contrasted values” mean?

Figure 4: I note that the sample weights that I asked about in my previous comments are plotted here. Can sample weights be added to a results table so that the reader can reference it more easily throughout the text?

240: Interesting that the glass standards made the LA-ICPMS RB samples 30% higher than the wet chemistry method, which are already 100x higher than Schwartz et al., 2020. Can you expand more on why this matrix effect causes higher concentrations (rather than lower or dispersed)?

261: There are other studies that have performed LA-ICPMS on spinel in the literature that should be cited here and can be used to discuss how others have matrix matched their standards or any implications your study has on these prior studies. For example, Colas et al., 2014 “Fingerprints of metamorphism in chromite: New insights from minor and trace elements” in Chemical Geology is one but there are others as well.

290-300: This section of the discussion offers no reference to the impact of inclusions on dispersion in magnetite and spinel. This should be added.

297: The laser ablation parameters should be reported in a table in the text or appendix. How many spots per sample? Were spots averaged? Were some samples/spot sizes under the detection limits? What was the variation in U and/or Th recovery with spot size?

310: Do you think the dispersion (20%) is primarily an analytical limitation or a geologic limitation?

356-362: This discussion on the location of U in the synthetic magnetite is super interesting. A big question is whether U can be incorporated into the crystal structure or is adsorbed onto the surfaces and the magnetite grows around it. It can have important impact on dissolution and He production. I wonder if this discussion can be moved into the main text?

Review of Corre et al., submitted to Geochronology

This manuscript explores the measurement of U and Th in magnetite and spinel samples to assess the expected uncertainties with modern analytical techniques at different concentration levels. Magnetite and spinel tend to contain low concentrations of U and Th, which makes it difficult to use them for geochronology, but the ubiquity of magnetite and spinel, and the relatively high closure temperate of helium make them interesting target phases. This manuscript presents useful guidance on how to optimize measurements and outlines the limitations of modern analytical techniques. It represents a significant step towards exploring the potential of magnetite and spinel (U-Th)/He dating and making it possible to use these techniques routinely.

In my opinion, this manuscript is well-written, and the data is presented effectively, but it could benefit from minor revisions, as outlined below. The topic fits within the scope of Geochronology, but I would suggest changing the article type to “Technical Note” since it fits the description of that category more than that of a “Research article”.

I agree with the comments by RC1, and I will only mention additional points below:

Line 42: Change “fault” to “faulting”.

Line 45: Change “radiogenic” to “radioactive”.

Line 46: Change “neighbor” to “neighboring”.

Line 56: The quoted number of “0.0012%” is incorrect; 0.0012 is the fraction (not percentage) of the contribution of Sm to the effective U concentration (eU), which equates to 0.12%. The exact contribution of Sm to the radiogenic budget depends on the sample-dependent U/Sm and Th/Sm ratios. I have seen some samples with low U and Th concentrations and relatively high Sm concentrations in which Sm did contribute to the total amount of ⁴He to a level above that of the measurement uncertainty. For most samples, the contribution of Sm to the measured amount of ⁴He is negligible, but can’t be ruled out a priori. I agree with not discussing Sm in the manuscript, but the reasons should be clarified.

Table 1: Pictures of the samples in addition to the descriptions would be helpful.

Line 93: Was the goethite removed before analyzing the magnetite? Were U and Th partitioned between the magnetite and goethite? Please discuss this possibility and mention any data you might have.

Line 100: Why were the samples ground into a powder? The stated goal is to assess uncertainties as a result of single-aliquot dating of magnetite grains, but this process homogenizes the sample similarly to a two-aliquot approach. As a result, all intra-sample variability is homogenized, which could be due to a U-Th zonation, or could be true age heterogeneity. If the sample has true age zonation, homogenizing the material would result in a meaningless average age. Therefore, this is very different to the single-aliquot approach usually employed for these types of samples. Please discuss what differences can exist and how the results will be relevant for a single-aliquot approach.

Line 132: I do not understand this sentence: “The quantitative determination of U-Th abundances can therefore hardly be led on too diluted solutions…”. Please re-write to clarify.

Line 135: Can you matrix-match your standard solutions to counter these matrix effects? Is this an effective strategy or would removing Fe (like suggested by RC1) produce better results? Did you employ this technique here?

Line 158: “45.62±3.40” and “116.01±12.60” contain too many significant figures. Uncertainties shouldn’t exceed two significant figures, and the measurement should be rounded accordingly.

Line 159: A possible contamination is very concerning. What are the procedural blank levels for these measurements? How many procedural blanks were run? Do the

Figure 1: Add 1:1 line to make deviations more apparent.

Figure 2: I’m not sure what the point of breaking the axis is in subfigure (a). There are no values >60% so the axes could just end at 60%. For (b) and (c), adding a line for the minimum uncertainty derived from counting statistics would be helpful. This would show the magnitude of other sources of error, e.g., matrix effects. The analytical trends of uncertainties increasing rapidly below 0.5 ppm are in agreement with my own experience working with similar instruments.

Table 2: There is a mismatch between the number of digits for the measurements and that of the uncertainties. Keep the uncertainties to either one or two significant figures and adjust the rounding of the main values accordingly. Use the same number of significant figures for the mean values and CVs. Give the full sample names and their abbreviations in the table to make it easier to reference. Also, change “Aluminons” to “Aluminous”. The absolute measured amounts of U and Th, as well as the measured (or weighed?) Fe-oxide mass should be given for each sample, along with the results of procedural blanks. This would allow a comparison of the measurement and the blank level/detection limit.

Section 3.2: The wording in this section is a bit unclear and should be revised.

Line 189: Change “samples” to “sample”.

Line 191: Change “those” to “this”.

Line 192: Delete “in mind”.

Lines 207-209: Did you consider the stability of Th in the solution as a possible cause for dispersion? Th is known to be “sticky”, and a high level of acidity needs to be maintained to keep it in solution. Typically, this is done with 5-10% HNO₃ and/or by adding a small quantity of HF to the solution. Was the dilution done with water or an acid mixture? How was Th stabilized during dilution? Discuss this here and add a detailed description of the dilution procedure to section 2.3.2.

Line 208: Explain what you mean by “over dilution”. As the solution is diluted, the U and Th count rates are going to diminish, but matrix effects are going to be reduced. Relative to what do you define the “over” dilution?

Lines 208-209: The observed natural variability in U and Th concentrations is similar to that of other iron oxides, such as hematite and goethite (see, for example, Hofmann et al., 2020, Chemical Geology). This natural variability, which can be true age inhomogeneity in some samples, highlights the importance of single-aliquot ages that sample small volumes, such as with conventional laser-heating of aliquots in metal packets or laser-ablation.

Lines 257-258: Adjust significant figures as above.

Section 4.3: This is a very helpful section!

Lines 288-291: Add references to the relevant literature for these effects.

Line 294: Do you mean (>10%)?

Line 308: The hyphenation of “in-situ” is inconsistent throughout the manuscript.