Dear Daniela,

I have reviewed the manuscript titled “Effect of chemical abrasion of zircon on SHRIMP U/Pb, δ18O, Trace element, and LA-ICPMS trace element and Lu-Hf isotopic analyses” by Kooymans et al. This manuscript presents a major effort to report U-Pb, O and Lu-Hf isotopic systems and various trace elements including OH of zircon, to show if any effect is observed by chemical abrasion (CA) treatment.

It is indeed a very interesting subject, one that I’d like to see published in the near future, but not in the current form. My biggest concern is that the manuscript could not focus on solving the main problems (which were not clearly defined), but instead distractedly explained too many things and presented data that are not essential. In my opinion, the main argument in this paper (in the other word, the most novel component of this manuscript that zirconologists must be interested in) is that CA could produce poorer surface condition on zircon that possibly affect the sputtering process and change U and Pb ion emissions, resulting in scattered SIMS U-Pb data for moderately damaged zircon. The authors did not discuss this well enough but continued discussing other isotopic/elemental data and MPV zircon core data. The other data unchanged by CA are interesting, but unfortunately distracting in the current manuscript, together with the poor writing.

The overall quality of the writing makes the manuscript challenging to navigate. Even the introduction lacks a clear focus on a main purpose, and the motivation and aims of the study are not distinctly described. The manuscript could be significantly shortened by removing repetitive and unnecessary sections. Many paragraphs consist of only a sentence or two without a clear main idea, indicating a need for substantial improvement in paragraph construction. The overall structural organization is notably lacking, contributing to a distracted and unfocused narrative. A more focused and concise writing style would immensely enhance the manuscript.

Furthermore, many sentences in the manuscript appear as guesses or the authors' opinions without thorough discussion or supporting references or evidence. Oftentimes, conclusions, interpretations, or the significance of these statements are missing entirely. The manuscript lacks a clear logic build-up that is necessary for constructing a reasonable and convincing argument. Throughout my reading, I had to keep asking – "WHY?" or "SO WHAT?". I had to spend much more time than necessary to make a guess what the authors meant. I still do not comprehend many sections and may have misunderstood some parts.

I strongly recommend a comprehensive rewrite of the manuscript, with a narrowed focus on SIMS U-Pb data (possibly with O). Considering my suggestion, non-SIMS sections are not much commented below. Expecting a thorough rewriting, I may not provide detailed comments on every aspect I have noted.

Please consider my review anonymous.

Specific comments

McKanna et al (2023, GChron) conducted a comprehensive surface study of CA-zircon, and mentioned “sponge-like texture” in CA-treated high-damage zircon samples. They also reported that “acid …regularly accesses crystal cores to dissolve … interior zones”, implying that the zircon core also can get the sponge texture if it is highly damaged. That is readily critical to SIMS analysis, which is a very good motivation of this manuscript. McKanna’s work is a key paper to introduce the aims of this study.  

Scattered U-Pb data observed from CA-treated MPV zircon rims unquestionably raise concerns about the reliability of SIMS U-Pb data for CA-zircon and necessity of CA in any SIMS work. Two hypotheses were introduced to explain it, but they were only briefly addressed in a sentence or two, L603-610, requiring further elaboration and explanation.

Additional session to re-run the zircon grains where the scattered U-Pb data was a good idea, although would have preferred a more comprehensive double-check, including all the same grains rather than just those deviating from the average U-Pb dates – so that you can properly see if the data are less scattered with >0.05 probability. And the conclusion that sputtering difference on CA-zircon surface (nanospongeform - not visible at all with high mag SE?) is reasonable. But again, I’d like to see more comprehensive discussion about this conclusion.

The argument that CA of moderately-strongly damaged zircon yields poorer SIMS U-Pb data should be consistently maintained throughout the manuscript, not trying to interpret CA inherited zircon data. Without considering the possibility of worse SIMS U-Pb data after CA-treatment more, the authors just kept discussing about the MPV core ages in the geological context, which is insignificant because the MPC core age data are highly likely less reliable than MPU. Need discussion if the CA reaching depth into zircon is throughout the core area too (yes, according to McKanna et al 2023). The probability density diagrams of CA-treated and untreated MPV cores – the latter looks a lot sharper, which may be consistent with the possibility of MPV CA cores provided more scattered U-Pb data, same as on the rims. + I’d like to see the diagram of the source rock if you can specify it.

I also found the higher ratio of the magmatic age (430 Ma) out of total number of core analyses potentially misleading and prone to misinterpretation. 6 vs 18 looks obvious difference, but I believe it may be due to a target selection bias for analysis. As clearly mentioned and illustrated in the CL figures, CA treated MPV zircons have obvious etch channels. It is highly possible that the authors selected better-looking zircons, avoiding visible cracks, inclusions and possibly etch channels too (even unconsciously). As the authors mentioned, the 430 Ma-cores have the least evident etch channels around, making them more likely chosen for anlaysis compared to those surrounded channels (possibly older). What is the ratio of measured cores that show obvious etch channels? - I could identify max 10 cores analysed in the CL image. Additionally, I cannot understand “survivor bias” (L589) explanation.

Those above are the main comments to the SIMS U-Pb data. Some others follow:

For sample selection: No indication of density or magnitude of damage of zircon? It reads like the authors tempted to use U content as a potential indicator in the section 1.2, but the connection between U content and the level of damage is not explained. In the similar regards, it is not clear why MP volcanic zircon was chosen for this study. Clarifying the rationale behind choosing this specific type of zircon would enhance the reader's understanding of the study's context and objectives.

U-Pb internal/external error: you always should report the U-Pb dates with external error propagated, as the published dates and ages are going to be compared to the other ages. Not sure why the authors consider reporting the internal errors only, all the explanations about which make the manuscript at least 1-2 pages longer.

Untreated TEM2 was used as a primary standard for U-Pb (with two other untreated zircon references) for running both CA-treated or untreated unknowns. This kind of test should be done under the same condition, so it is a reasonable question that any other readers might have. What if the U-Pb calibrations are different between CA-treated and untreated zircon? You showed TEM2 has same U-Pb calib curve slope for both preparations in the results, but it was supposed unknown during the analytical strategy stage (and QGNG and OG1 show different calibration slopes, although the authors consider it trivial; I’d like to see the calibration curves illustrated in the manuscript). If it was deliberate to see how untreated standard affects to CA-treated unknown zircon U-Pb calibration, it should be mentioned.

CommonPb correction using 204Pb. L265: “…as 204Pb overcounts were within uncertainty of zero…” …? What else would you do if it is above zero then? And why don’t you use 207Pb-corrected 6/38 age for MPV zircon dates?

Many unnecessary – examples:

L413: “…did not run as smoothly…” unless you are going to point something out for the less smooth session, it is super unnecessary. Every session has its own condition, and you cannot compare them all the time. L271: Truly unnecessary. Calibration slope could be different even session to session on a same instrument.

Many unclear – examples: L24, “…the analyses of chemically abraded materials show excess scatter” OF WHAT?

L107: “These volcanic zircon rims are also lower in U content than S-type granite rims, which often go metamict…” S-type granite “zircon” rim does not have constant U concentration. Which specific S-type granite do you mean? And what is the U content of the MP volcanic zircon? How do you say something is lower than something without showing their data? What is the U concentration threshold to go metamict? Any reference for that?

L149: “We also include four new aliquots for OGC…” Why? Then they are newly measured not from the literature data? And the new data should be properly reported.

Result-discussion mixed up: L287-9, that is one of the aim of this study, not method

L502: “… reduces the uncertainty in the reference ages by 140-290%...” According to Table 1, ratio uncertainty except 91U (why is that?) is certainly smaller including tracer, but not age uncertainties. How come? Age uncertainties are actually larger including tracer.

L541: “…, which is consistent with our SHRIMP U-Pb data” How and what exactly?

L549-551: Not sure.. Fig 2 shows QNGN and OG1 have different calibration slopes between U-C

L552-554: self-calibrated? What about the other standards data using the low-sloped 91500 calibration?

L556-560: that is a pure guess. If you want to say so, you need more examples of low Hf & REE zircon data. Do not argue anything without supporting evidence.

L572-578: Why S-type granite only? Slow(er) crystallization is a general condition to form granite than volcanic rock. And common Pb is not only from Pb-rich inclusion. “For comparison…”: comparison of what and what? Why the granites from Bodorkos et al (2015)? Apparently same igneous ages (~430 Ma), so are they potential co-genetic bodies? There is no explanation. “… statistically significant common Pb”: which means…? “…raw 207Pb/206Pb ratios…”: raw ratios? Same as total 207/206? 207, 206Pb are mostly radiogenic Pb in zircon anyway? “…unusually high common Pb contents” I think it is not uncommon to see common Pb? Apart from many question marks about every single sentence in this paragraph, I do not see why this full paragraph is necessary.

L582-3: show the data. Compare the size of uncertainties or etc.

L608-9: Uh… probably MPV zircon is more damaged than the reference zircons?

L613: total 206Pb… why?

L617-: I don’t get what you try to explain in this paragraph.

L627-633: “This is consistent with…” “consistent” is to compare to the others’ arguments or consensus. To me, this paragraph is a guess, with no other supporting arguments. From the rim data (MPU, MPC), I am convinced that the CA method weakens the SIMS U-Pb validity and think the MPC core dating is less reliable due to the CA-induced surface damage (nanospongeform – according to the authors’ description, L607).

L650-: That is bizarre to publish. I strongly suggest the authors to re-run a session for Temora2 using a same batch. What if the huge d18O difference after CA is true?

L654: MPV zircon scattered d18O may indicate that the source rock/melt was not in equilibrium. Not all S-type granites show scattered zircon d18O!

L657-8: what kind of evidence do you need?

L658: Ti thermometry? Reference? Did you perform the measurement?

L662: “… not all … volcanic zircon cores are detrital” They never are.

L668-: this paragraph should to go the result section

L677-8: I don’t see significant difference in Fig 7 except OG1, which look plotted wrong

Technical corrections

Supplemantary tables:

Check all the table format (not only for the suppl tables)

All negative values should be corrected to zero or “-“

Th/U not a ratio of isotopes (232Th/238U) but total Th/U estimate

Reasonable effective number and decimal places

Oxygen data table only show internal SE (standard error), which should be replaced by fully propagated uncertainties. (or if you argue that no external reproducibility is not necessarily considered because only within-session data are compared, it should be explained in the text)

Figures:

Fig 1: it is useful to have some indications to figure our easily which cores are syn-eruptic and which “cracks” are the etch channels.

Fig 4: readers would like to see the Concordia plots for all standard/samples.

Fig 11: CA treated zircon data should be included too.

Other comments

It does not read a formal scientific manuscript, rather oral speech script or a personal journal in many aspects – use of informal words (best example is SHRIMPing), use of subject adjectives without giving numbers, lots of mistakes, sentences scientifically incorrect…

Lots of mistakes: typos, wrong capital letters, use of spaces, inconsistent expression (U-Pb vs U/Pb, use of acronyms), etc

Always make it clear what uncertainty you indicate – 1se, 1sd, 1s (65% conf) or 2s/ts (95% conf) etc.

SHRIMP – a specific brand name, but a method. Change it to SIMS except where specific SHRIMP IIe and SI are described.

Scientifically incorrect or text-data-plot not matching:

L89, 207/206Pb ratio is almost constant: not necessarily especially for old ones (in that regards, I actually would like to see Pb/Pb data too especially for OG1)

L91: “… SHRIMP can produce 7/6 age with 2‰ precision (really meant permil not percent?)” It cannot be a general comment as 7/6 age precision depends on the zircon age and Pb concentration.

L98: “…old enough for decent counting statistics…” Again, not necessarily. Counting statistics of U-Pb dating depends on U concentration too.

L512: “a piezo stage, automated analyses” why should they be conditions to get better precision?

L513: “…better than the 1-3% value…” Better? you see the exactly that range from TEM2 and 91500 in Table 2 (and larger for the other standards)?

L535: you do not have DR12 and 13 figures; if they are the ones from Magee et al., those fig numbers are not necessary or you need to make it clear.

L554: “For a difference in slope of 0.5, this would yield ages 0.2% older…” Calibration slope does not necessarily increase or decrease ages, as it is depending on where the unknown data sit compared to the standard’s calib curve.

Fig 7E: OH/18O values of OGC are all >0.1 in the suppl table, but they are plotted mostly <0.1 in thei figure.

Fig 12: The repeated session is 210046? It is 170124 session in the fig