Review of Nevskaya et al. “Dissolution-precipitation creep in polymineralic granitoid shear zones in experiments I: strain localization mechanisms”

This paper reports on a series of experiments conducted on cores of granitic ultramylonites at 1.2 GPa confining pressure and 650˚C. Two experimental series were performed, one where the rocks were pre-fractured during pressurization and one where deformation has only started once the rocks reached the desired P-T conditions. In both cases, the authors use extensive and detailed microstructural observations down to the TEM scales to argue that dissolution precipitation creep was dominantly accommodating deformation at the experimental conditions. The paper is very well written and logically organized, the figure quality is superb. I have only a few minor comments which the authors might find useful to further improve the manuscript.

Figure 3 – what I find rather interesting is the fact that both the pre-cracked as well as intact rocks essentially reach the same strength, or maybe the pre-cracked samples are actually a little stronger. What influences the peak strength in your opinion, could you expand the discussion to address this topic? It seems that the pre-existing fracture just doesn’t influence strength at all anymore at the elevated P-T conditions of the experiments, i.e. the rocks are firmly above the brittle-ductile transition (following the Kohlstedt et al. 1995 and Rutter 1986 definition of BDT as localized vs. delocalized deformation) and sliding on preexisting fractures is actually more difficult then deforming the bulk? But if the fine grain sizes that are so important for the strain localization form already during the fracturing and are present presumably earlier than in the non-fractured experiments why does the strength remain unchanged?

Section 4.1.2, discusses the differences between mono-mineralic and poly-mineralic rocks undergoing dissolution precipitation creep and the fact that it is driven by flux of matter form a source to a sink. In the classical treatment of grain boundary sliding in diffusion creep by for example Raj and Ashby 1971 the differential stress introduces variations in the normal stress acting at a grain boundary depending on their orientation and these variations introduce chemical potential gradients driving the flux of atoms towards the boundaries in compression and vacancies to grain boundaries in tension. The slowest diffusing species then sets the strain rate as charge neutrality has to be maintained within the bulk crystal. While this treatment is derived for simple monoatomic metals it can be expanded to more complex materials if all the point defect reactions are known.  On lines 343 – 345 you claim though that “In our polyminerlaic system, the transport and source/sink terms are not defined, and the introduction of a chemical driving force will be necessary”. I am confused by this framing. In my understanding you will still have sources and sinks and a chemical potential gradients based on the normal stress acting on the grain boundaries, and in addition you will have chemical potential gradients related to the differences in activities of various species depending on their concentration in the grains making polymineralic mix. So in other words I think you could at least in theory define source and sink terms for individual species diffusing and a chemical driving force is always present due to the differential stress. I agree that the problem gets quickly very complex and so quantitative treatment is currently not possible to my knowledge. A more nuanced framing here will avoid confusion in my opinion.

Also in this section on lines 350 – 355 you discuss the importance of porosity permeability and only a couple paragraphs later in the manuscript you start talking about advection of the fluid. I suggest you highlight here already the important implication of your observations and the necessity to account for fluid advection in DPC models.  In all classical DPC models I know of, the fluid is always treated as a stationary phase through which diffusion occurs and no advection of the fluid is invoked so your observation could motivate further theoretical developments.  

minor edits:

Figure 2 – can you add a scale bar please?

Line 165 – I agree with the statement but would mention here the existence of torsion experiments in solid medium apparatuses that reach very high strains (e.g. Cross and Skemer 2017)

Figure 4c – it is really neat how the microstructure in the highly localized zone is very similar to the mylonitic microstructure in the surrounding material, just on a shorter length scale. Just an observation.   

Line 242 - …if shortening results also IN material…. I would say in and not by

Line 242-243 …and possibly these areas are influenced by external boundary conditions.  I would re-phrase this as the whole sample is influenced by boundary conditions in my opinion. What about “these areas are most influenced by the boundary conditions of our experimental set-up” or something along these lines?

Line 274 – the median grain size reported here (145 nm) is applicable only for the zone of extreme strain localization along the previous fracture zone I assume. The grain size outside of the zone is presumably close to the starting material? Slight re-phrasing here might help to avoid potential confusion.

301 – typo here – previous (not precious)

303 - …chemical changes and -gradients… The dash is extra?

356 - …the diffusive transport length at least the grain size… length IS at least…?

388 – missing a space between 1.2 and GPa

Section 4.2.1 – in this section you always say that grain size is stable at certain strain rates, pressures and temperatures but shouldn’t you also include stresses?

474 - missing spaces between 260 nm and 500 MPa here

531 – missing a bracket after Menegon 2008 and space between that bracket and Hence.

551 – 552 – you have a typo here as you report velocities (m/s) and not strain rates (/s)

References:

Cross, A. J., and P. Skemer. "Ultramylonite generation via phase mixing in high‐strain experiments." Journal of Geophysical Research: Solid Earth 122.3 (2017): 1744-1759.

Kohlstedt, D. L., Evans, B., & Mackwell, S. J. (1995). Strength of the lithosphere: Constraints imposed by laboratory experiments. Journal of Geophysical Research: Solid Earth, 100(B9), 17587-17602.

Rutter, E. H. (1986). On the nomenclature of mode of failure transitions in rocks. Tectonophysics, 122(3-4), 381-387.

Raj, Rishi, and M. F. Ashby. "On grain boundary sliding and diffusional creep." Metallurgical transactions 2 (1971): 1113-1127.

Really great work!

Sincerely,

Matej Pec

Review of “Dissolution-precipitation creep in polymineralic granitoid shear zones in experiments I: strain localization mechanisms” – Nevskaya and others.

The paper reports the results of two sets of deformation experiments on natural mylonitic granitoids, highlighting the dominant role of Diffusion/Dissolution-Precipitation Creep during experimental deformation of fine-grained, polymineralic geologic materials. The manuscript presents a solid and well-structured study of significant interests for the Structural geology and Rock deformation communities, and suitable for Solid Earth. My expertise in experimental deformation is limited, and thus I focused my attention to the microstructural data presentation, analyses and implications/extrapolation to natural conditions. I really appreciate the detailed microstructural and multiscale analysis, as well as the interpretation supported by the presented data. However, I have several major comments that I believe should be taken into account by the Authors to further strengthen the clarity and impact of the manuscript. I hope the authors will consider and discuss these points in a revised version of the paper. I truly apologize for the very late posting of this comment.

Major comments

1. Considering the “three-times larger strain rates” of Type I vs. Type II experiments, the results highlight the extreme efficiency of pre-existing fractures (and/or any other type of mechanical discontinuity – i.e., “sharp contacts”) in steering strain localization. Yet, the inferred deformation mechanisms are the same in both experiments; and even though the shortening of the sample is the same, the volumetric proportion of the actively deforming sample is different in Type I and Type II experiments. The final results are rather counterintuitive because in Type I experiments, the DPC mechanisms are volumetrically limited to the pre-existing fracture volume, and the deformation runs faster. Whereas, in Type II experiments, where the DPC mechanisms occur apparently over the entire sample (or over a larger volume), runs slower…? Is this just a misleading effect of comparing shear strain rates to bulk strain rates, or is it related to geometry of strain localization, or geometry of the sample?

1. Type II experiments and starting grain size: the Authors state that their finer-grained samples might be better suited to activate DPC at conditions at which frictional deformation is otherwise expected, and this statement is kind of misleading in my opinion. In Type II experiments, grain size reduction is in any case observed, before the (partial) activation of efficient DPC. Furthermore, Type II experiments seem to show a complex evolution in terms of localization, and thus in the accommodating deformation mechanisms – and evidence for localized brittle deformation (in form of initial localized discontinuities – shear bands, and/or diffuse microfracturing of certain phases in Fig. 4j-k) need to be taken into account to explain the kickstart (at least in part) of the local grain size reduction on which DPC and GBS are then developed. I am totally fine with the following microstructural and mechanical evolution, but it is hard to believe that microfracturing did not contributed in the incipient stages of diffuse sample deformation [If the black openings in Fig. 4j are just due to unloading, then please replace the image with something less misleading].

1. Porosity and Type I/Type II experiments: it has to be acknowledged and discussed that the two samples differ in terms of timing and efficiency of porosity development (as partially already discussed in Section 4.1.4), as well as “water” availability: Type I samples are characterized by an initial fracture which is very efficient in redistributing the added water into the sample and directly into the deformation zone. Whereas, Type II samples can only rely on the intergranular diffusion and GBS-based advection, which is to be expected way less efficient than the fracture-controlled redistribution of Type I samples. Therefore, even though all the sample have virtually the same content of “added water”, each sample Type might also differ in terms of water availability and thus, the resulting deformation mechanisms, microstructures, and strain rates might be dependent on it.

Minor comments

Line 21: “classical” is not so clear – what do you mean? Sigmoidal? Distributed over the sample?

Fig. 1 caption: “where less phase mixing occurs”, I would delete this to avoid misconceptions, the lesser degree of mixing could be due to the scale of observation and the relative grain size. Same comment for Line 85.

Line 75: delete double brackets on the reference citation.

Line 186-187/Fig. 4e: could you please provide an image of the same microstructure at higher magnification to see indeed the new elongated and more abundant grains?

Lines 193-194: could you please explain then the microstructure shown in Fig. 5d? Is this brittle fabric due to stress-unloading? Please add a short description in the figure caption.

Line 196-198: “Feldspar become more mixed” indicating Fig. 5c. The microstructure shown in 5c is a rather typical replacement/reaction microstructure between Kfs and Ab, and thus it does not support the previous statement.

Line 237-8: could you please indicate where the data about aspect ratios come from?

Line 301: “precious” typo – “previous”?

Lines 444-450: I think that the work of Tokle and Hirth 2021 is here mis-cited, given that their intention was not to adopt paleowattmeters to diffusing systems…?

Lines 507-510: please, highlight again that, even though performed at compatible P-T conditions, the strain rates are far away from natural geologic (sub-seismic) strain rates.

Many congrats on the very nice piece of science,

Best

Alberto Ceccato