These comments are intended for the editor and authors.

In this short paper, Fukuda et al. compare the water content of quartz in three granitic samples using infrared spectroscopy. Water has a major affect on the rheology of quartz, but the details of this are complicated and not well understood.  The paper is fairly clearly written, the figures are of excellent quality, and I enjoyed reading the paper and learned some things. I must point out here that while I am familiar with recrystallization of quartz, I have not studied water contents in quartz and the reader of this review should keep that in mind while considering my comments. 

My main concern about the paper is the small sample size (N=3) and use of only one technique to investigate the samples. As far as I know, IR measurements are not particularly expensive or laborious to make, although IR mapping may be relatively novel (I’m not sure). So my sense is that the study is well below the average published contribution in terms of the intellectual rigor involved in its production. Also, other studies have already shown similar results—the authors cite two previous studies (Finch et al., 2016; and Kronenberg et al., 2020) showing decreased water concentration associated with recrystallization in natural samples. Adding another data set like this to the literature is valuable, however I am accustomed in published work to see significantly more data presented and/or a more detailed analysis that makes more progress towards some outstanding question. There does not seem to any specific question being targeted or addressed by the study, other than “how does water content in this shear zone change during recrystallization?”.

I recommend that additional samples are analyzed or some complimentary technique is added to the study before publication. For example, EBSD maps of the samples would allow a quantification of the degree of recrystallization and subgrain formation involved in changing water contents. Also, the authors infer the presence of subgrains in their thick sections, but this could be proven and quantified using EBSD. The authors also infer different types of recrystallization in strongly deformed and weakly deformed samples which I find puzzling—such differences could also be quantified using EBSD.

Below are some additional comments.

13-15. Bulges form on host grains, so I don’t understand how they can also be a few hundred microns distant. Please clarify the language.

14-16. How can it be that small amounts of deformation involve bulging recrystallization, but large amounts of deformation are inferred to have experienced mainly subgrain rotation recrystallization? Is a switch in recrystallization mechanism over time being inferred, or did deformation occur at different conditions in different places (problematic for the study, since there is a tacit assumption that deformation conditions were similar in the three different samples). Alternatively, possibly there is not as much clarity about the deformation mechanisms as the author’s think (EBSD analysis could help a little with this).

20. Language issue:  “is released” is problematic. “can be released” or “was released” would be better. There are scenarios where very dry quartz is recrystallized in the presence of water, and during this process water is added to the quartz. By saying “is” it sounds like the authors are making a universal claim.

29 “several” better than “a few”

Section 2 Samples:  I would also like to see more attention paid to the deformation history of these samples. Are they from a strike slip, thrust, or normal sense deformation environment? Are they foot wall or hanging wall? Do we have information about the temperature of deformation? Is there information about the initial distance between these samples when deformation occurred (how similar do we really expect them to be)? Does previous work in the area suggest that the deformation of the three samples occurred simultaneously (at different strain rates), or was there a sequence of overprinting at progressively lower temperatures?

96. “By imaging” is quite vague.

101-102. These piezometers were calibrated using some samples that also experienced subgrain rotation recrystallization, not just bulging. Also, (line 106) the Cross piezometer uses a subset of the same samples Stipp used, so it’s absurd to say that Cross et al is useful for samples that experienced subgrain rotation recrystallization, while Stipp piezometer is for bulging.  In any case, it doesn’t seem necessary to separate out two kinds of piezometers anyway because only rough estimates of stress are produced.

110-113. Inaccurate language: There is much more to preparing thin sections than by dissolving resin in acetone.

111-112. Similarly, this language about the dial gauge seems incomplete.

150-151. For those ignorant of such things (like myself): can we assume that the same IR calibration for quartz gives accurate values of water content in feldspars? A note and reference here or in the methods would be useful.

155-156. This observation, that water content is less between recrystallized and unrecrystallized areas of the same sample is very important. It strengthens conclusions drawn from sample to sample differences in water content. The authors might want to emphasize this more elsewhere in the paper.

163-164. It is very difficult, perhaps impossible, to truly distinguish a small subgrain from a small grain in a thick section (when the grain size is much smaller than the thickness of the slide). I don’t doubt that there are subgrains, but the language here is misleading.

169-170. Figure 8a is too zoomed out to see any subgrains (it seems from the placement of “(Fig. 8a)” that we are to see subgrains in the figure).

183-184. It is suggested later in the paragraph that small optically invisible inclusions can also provide a similar magnitude IR signal. So why presume that it is the large inclusions that are responsible here? Couldn’t it be the case here that the signal is also mainly from submicroscopic inclusions?  You could substantiate your claim by measuring the density of inclusions as compared to the IR results in the sample (is there a correlation?), or cite some earlier work if this has been addressed before.

185. Change “do not differ” to “do not differ substantially” (means of 500 vs 800 are different)

185-187. This is a big claim (that it must be because of invisible inclusions… how would that happen? show us micrograph images of the two areas?). Can the same technique be used to verify the presence of small inclusions that Stunitz used? 

190-193. I don’t know much about IR detection of water, but I imagine that the signal strength might depend on the size of inclusions (microscopic vs sub-microscopic). If so, it would affect this supposition. Add some information on this if possible.

193-194. Confusing language. Change “are comparable” to “correspond to”

195-197. I’m not following the logic here. Why “therefore”? This idea of redistribution seems odd to me. It seems like, overall, water left the system during deformation, so it seems unlikely it would increase in the host crystals—is that what is being implied (“higher water contents…due to the redistribution of fluid inclusions”)?

210. Section 5.2. It is unclear to me what the main point(s) of this section is. I read it a few times and it feels like a loose collection of thoughts and information from the literature about water in recrystallized quartz. I’m not sure if any new idea is being presented, or what exactly the new data contribute to the previous understanding.

219. “development of dynamic recrystallization” wording is linguistically problematic.  Replace with “development of dynamically recrystallized grains” or “dynamic recrystallization”.

229. “which may be due to the transition from subgrains to recrystallized grains” This could be tested with EBSD.

244. This is a hypothetical statement. Use “were” instead of “is”

244-245. Just a thought (no need to address it): If the water can be “distributed homogeneously in grain boundaries as thin films,” is there any concern that it can be added or lost during the process of making thick sections? What happens if the samples are heated before IR analysis (this might remove grain boundary water)? Could test for this with IR analysis before and after a heating.

252-254. This sentence is confusing me. Why is thought to be “continuously supplied”? What is the significance of “textural modifications” here?

254. “Intracrystaline parts” language is unclear. “Crystal interiors” makes more sense to me (if that is what is meant).

278. Change “original fluid inclusions can be redistributed” to “water from original fluid inclusions can be redistributed”

286-288. I’m not following this logic. Why is diffusion invoked to explain host vs rxld grain water in these samples? All you need is to remove water in recrystallized grains, and the remaining host quartz just holds onto its water (no diffusion necessary except perhaps to redistribute from large to small inclusions).

290-291. The words “that lead to the development of equilibrium texture” may not be necessary, and ideas of “equilibrium“ are often complicated (are they really equilibrium? How would we know?). Also it would be necessary to explain what is meant by “equilibrium texture;” do you mean “equilibrium microstructure” (texture for some people means CPO, but I don’t think it’s what is meant)? Probably best to just trim the sentence by removing these words. 

294.-295. I think it is being suggested that formation of subgrains releases an intermediate amount of water. Can this really be inferred with the available data? Consider also the possibility that measuring a mixture of recrystallized grains and undeformed grains also gives an intermediate value.  It is very difficult (or impossible) to tell, optically, in a thick (100 μm) sample what is a subgrain and what is a recrystallized grain… EBSD could help clarify what is happening, I believe.

297-299. What is the “equilibrium state” of water during deformation? This concept would need to be explored before ending the paper with this sentence. Is it related to temperature, presence of other minerals? Equilibrium between water in quartz and what? Has this been quantified in the literature, and if so what is the significance of the value found (can it be used to quantify anything, water fugacity maybe?)?

629. “addition” not “addiction”

629-633. suggestion: I find the language “former and latter” confusing. Try more direct language, i.e. just say “in the weakly deformed sample” rather than “in the former sample.”

630. It says water distributions in host grains in the undeformed sample…are “not comparable” with the shapes of host grains  (Figs 5–7). This is a strange statement because the size of grains in the undeformed sample is much bigger than the IR maps, so how is this known?

Discussion. What is the bigger significance of these findings? It was already known that water can be released during deformation.

FIgure 1. Label A and B. Need inset showing location in Japan.  The lower map should be clear about what the black lines are…are they faults, intrusive contacts, depositional contacts, roads (they seem to be a mix of these things)? Also the blobs with various deformed status are confusing—is this really the only part of the Wariyama granite that is deformed? Is the deformation related to any of the faults/contacts shown?

Figure 5, 6, 7, 8. Color scale for the IR maps should be the same so that they can be directly compared visually. For example in line 185-187, it says that  regions of two samples have the same water content, but we can’t visually “see” this easily the way the figures are currently colored. This would also help compare figures 6 and 7, which are of the same sample. The change will also help the reader intuitively grasp the main result of the paper (the strongly deformed sample will show up as mostly blue, whereas the undeformed sample will be red).

Some questions I am left with: 

How significant would the change in water content be in terms of rheology of the samples?

The undeformed quartz is heterogeneous in terms of water content. Water weakens quartz. So one might expect that deformation would occur mainly in areas of high water content. However, the host quartz remaining in weakly deformed quartz has more water, on average, than the undeformed quartz—not what you’d expect if recrystallization were focused in the high-water areas.  How do you explain this? Perhaps it is a result of the small data set (i.e. one of the samples is anomalous in terms of water content)

Comments on the manuscript: “Water release and homogenization by dynamic recrystallization of quartz” by Fukuda, Okudaira, and Ohtomo

The manuscript describes FTIR measurements in quartz from granitoids deformed to different finite strains. The authors find that the H2O content decreases systematically with different degrees of dynamic recrystallization due to different finite strains. The manuscript is concisely written, to the point, and the conclusions are supported by the data presented. I recommend that the manuscript is published after minor revisions.

Detailed comments:

Line 34: ad reference: Negre et al. 2021

Line 50: please add “quartz aggregates” instead of “in quartz” to include grain boundaries rather than just grain interiors.

Line 53: better “grain boundary regions” than just grain boundaries, because you consider volumes here (grain boundaries would just be surfaces).

Line 78: instead of using “texture”, it would be better to use “microstructure” or “fabric” here. Texture in a deformation context could be confused with CPO. This applies to the whole text.

Line 155: better: “...host quartz grains...”

Line 156: some recrystallized regions contain H2O higher than 300 ppm H2O (light blue colors correspond to about 400 ppm H2O in Fig. 6).

Lines 162-166: the “subgrains” could be regions with healed microscracks – it is difficult to see from Fig. 7b. Higher H2O contents in such regions would be consistent with microcracks. Please mention the possibility of microcracks  and perhaps discuss them later in the discussion section.

Line 220: add reference Kilian et al. 2016.

This manuscript presents IR absorption spectra of OH stretching bands and maps of water content measured for deformed granite samples from the Wariyama uplift of NE Japan, including a Wariyama granite that is not deformed or only modestly deformed (designated as “almost deformed”), and weakly deformed and strongly deformed granites of the same unit.  The deformation microstructures of quartz in these samples is described, including undulatory extinction, subgrains and recrystallized grains with characteristics of dynamic recrystallization during dislocation creep.  I am pleased to see that the authors have found similar results for water in quartz during recrystallization as our group did for deformed and recrystallized quartz mylonites (Kronenberg et al., 2020).  I will comment, in part, in response to remarks and discussion of other review that suggest that the results here are not new.  I would counter that it is necessary to learn whether recrystallization routinely excludes water from quartz grain interiors or not.  The fact that Finch et al (2016) found reductions in water content with strain in the El Pichao shear zone of NW Argentina, and we found reductions in water contents with recrystallization of the Moine Thrust, NW Scotland is no guarantee that water contents will be reduced in other deformed rocks undergoing dislocation creep and recrystallization.  In other words, I’m pleased to learn that this result was obtained in the present study and with further confirmations, we may be able to generalize the conclusion that water contents always (or normally) decrease with recrystallization.

On the other hand, I find that this manuscript includes new results for coexisting plagioclase grains of the same deformed granites, and I recommend that the authors revise the manuscript for publication, building on these results. Firstly, neither the title nor abstract prepare the reader for the interesting IR results for plagioclase of these deformed granites.  The results section includes IR spectra for plagioclase but the discussion section could address these results further.  Secondly, I cannot tell from the manuscript whether feldspars of these granites are internally deformed or not, and whether they are also recrystallized.  The IR spectra show that coexisting plagioclase grains are very wet, compared with deformed quartz grains, and their spectra appear to include sharp high-wavenumber OH bands due to layer silicate (I assume sericite) inclusions.  In an earlier IR study of Sierra Nevada granites deformed at greenschist grade conditions, Kronenberg et al. (1990) also found that feldspars had higher water contents than quartz grains of the same granites.  However, we found no evidence of plastic deformation of the feldspars so we could not say anything about water weakening of feldspars.  Apparently the temperature of deformation was too low for any plastic deformation of feldspars.

I propose the following additions to this contribution:  1) describe the microstructures of plagioclase relevant to their brittle or plastic deformation, and any recrystallization that might have occurred (if no plastic deformation is evident, that’s OK – simply report your observations),  2) if the plagioclase is recrystallized at grain margins, are they wet in these regions or are they dry?  3) please cite any papers or results that provide constraints on metamorphic temperatures during Wariyama granite deformation.   Again, providing a link between temperature/metamorphic facies, deformation and water contents of feldspars and comparisons with the dislocation glide and creep of quartz would be very useful, and 4) please describe the fluid inclusions and white mica inclusions in plagioclase grains that relate to their large OH absorption bands.  I would be curious if these rocks also reveal evidence of plastic deformation and recrystallization of plagioclase and potential water weakening of feldspars at conditions that favor dislocation creep, subgrain formation, and dynamic recrystallization of quartz.  5)  With such results, the authors can also compare with results reported for other deformed granitic rocks.  For example, Kilian et al. (2016) found that quartz deformed and recrystallized at higher temperature, amphibolite facies conditions are very dry.  Feldspars in those same rocks were observed to be less deformed (appearing as augen) but they were recrystallized at their margins.  The previous study did not characterize OH of feldspars but this study does, so adding the microstructural information is important.

In the following, I comment and ask the authors to correct or clarify technical issues.  First, the ability to measure IR spectra using very small apertures (25 x 25 microns) using a conventional FTIR and IR source (not IR of a synchrotron) is impressive and this leads to nice OH maps of the samples.

It would be useful to the reader to describe the use and characteristics of the glycol phthalate resin used in preparation of thin sections.  I assume this resin was used to inject samples in order to polish the IR samples.  However, I do not know anything about the IR spectra of this substance or whether it can be completely removed from samples after polishing is complete.  Please describe the use of this resin and whether it represents an issue for the spectra of quartz and plagioclase. 

The authors state very clearly their choice of the Paterson relationship between IR absorbance and OH contents for quartz grains, and comparisons between this relationship and other calibrations for quartz.  However, the authors should state clearly that this calibration depends on integrated absorbance, and shape of the OH absorptions.  Was the Paterson relationship and its method of integrating also used for the plagioclase grains?

It has become common practice to report water contents of minerals as weight ppm H2O rather than molar (or atomic) ppm (H/Si) but some OH absorption bands of quartz are due to hydrogen interstitials (OH) and do not represent H2O defects.  In addition, the original measurements by which OH absorption bands of standards have been determined (such as calibrations of Kats 1962) are not necessarily measurements of H2O weight per oxide (SiO2) weight. As I result, I don’t favor this choice of concentration units of measure, even though the authors have every right to use the units of choice that are now in wide usage.  More importantly, the authors should clarify how they map water contents in their maps when some of the map area consists of quartz and some consists of plagioclase.  Have separate calibrations been used for quartz and plagioclase to map water contents?  The contouring is in wt ppm, but I find it difficult to compare these for two minerals with different formula weights.  This becomes even more complicated for plagioclase grains that include a sharp band at 3630 cm-1 due to sericite inclusions, which represent OH of layer silicates and are probably strongly polarized relative to the crystallographic axes of the sericite inclusions.  The methods of determining and plotting H2O over areas of multiple phases should be described.  I wonder if it isn’t safer to map and contour integrated absorbances over OH bands of the mapped area, and infer water contents from these maps, rather than attempt to plot water contents using multiple conversions of OH absorption to water content.

There are a few places that the author might reword the manuscript text. 

For example, in the Samples section, I recommend replacing “by the naked eye” by “as observed in hand specimen”

In the same section, I recommend replacing “as an equivalent circle diameter” by “as the diameter of a circle of equivalent area to the diameter of the grain”

In the section on Analytical Procedure for IR spectroscopy, I assume that the sample translation stage translates the sample in X-Y while the IR beam was fixed in place (not as described by a “beam-moving function”).

In section 5.1 of the Discussion, I am not sure what is meant by “OH dislocations”.  Do you mean OH at dislocation cores?  Or do you mean a fully hydrated dislocation (OH completely saturate dangling bonds of dislocations)?

Referring to the last line of the Section 5.3 of the Discussion, I agree with the authors and doubt that many water contents measured in deformed quartz actually represent equilibrium concentrations.  They are generally too large for equilibrium defects and highly variable spatially.

Caption to Figure 9 – minor typo- “addiction” should be respelled as “addition”

Commenting on Data Availability, are the Jasco data files formatted in a way that readers can download Jasco-formatted IR files and plot the results as open-source files without purchasing special, analysis software?   If so, great.  If not, it would be good to make files available as csv or other open-source file formats.

In summary, this is an interesting contribution that will help us understand deformation of quartz and feldspars at natural strain rates.  The results for quartz appear to support and corroborate previous reports of water losses from shear zones undergoing dislocation creep and recrystallization, and the results for plagioclase are new.  The plagioclase results will have increased impact on the field by including the same detail of microstructural descriptions for plagioclase as for quartz.

The scientific quality of the results presented is very good, but the methods, as discussed above, need to be more clearly stated, and water contents and microstructures of plagioclase need further evaluation and representation, to the same extent as presented for quartz.  The changes I propose are not truly major but they involve more than a simple rewrite, so I characterize them as "major".

I look forward to seeing this paper in print,

Andreas Kronenberg