Localized incipient dehydration in antigorite serpentinite during deformation experiments at subduction zone conditions
Abstract. Intermediate-depth earthquakes are commonly linked to slab dehydration, with dehydration embrittlement of antigorite proposed as a dominant mechanism of brittle failure. However, experimental results increasingly suggest that classical dehydration embrittlement alone may not fully explain intermediate-depth seismicity. Recent work shows that localized incipient dehydration can occur in antigorite due to chemical heterogeneities and stress, forming olivine-rich veins that are also observed in natural serpentinites. These sites may represent regions of strain localization where failure can develop as deformation proceeds. To determine whether deformation promotes localized incipient dehydration relative to hydrostatic conditions, and whether this process contributes to antigorite brittle failure, we investigated the incipient dehydration of antigorite using a six-ram multi-anvil apparatus under both hydrostatic and deviatoric stress conditions. Experiments were conducted across the antigorite stability field at pressures of 3 GPa, temperatures of 530–712 °C, and strain rates of 10-4- 10-5 s-1. Localized incipient dehydration occurs within the antigorite stability field during both static and deformation experiments. Nanocrystalline clusters, veins, and networks containing olivine and pyroxene are observed in all experiments. Localized dehydration is enhanced during deformation experiments, leading to the development of distinct microstructures relative to static conditions. However, this process promotes failure only at fast strain rates (~10⁻⁴ s⁻¹), suggesting that incipient or even complete dehydration alone is insufficient to cause embrittlement at laboratory conditions.
Dear Editor and authors,
This paper provides new results of static versus deformation experiments that aimed at exploring the relationships between antigorite dehydration and deformation processes, particularly to test the hypothesis of dehydration embrittlement to account for intermediate-depth Earthquakes. This type of study is particularly welcome to better understand such a type of earthquakes, origin of which is still enigmatic and highly debated (I am myself quite doubtful about this hypothesis). Although some references are missing, the paper is well written and correctly organized with the description of new features, such as the production of olivine veins through coeval dehydration and deformation, which deserve to be published. However, I see a critical bias in the experimental setup that, to me, compromises the publication. Indeed, the authors describes an incipient dehydration of antigorite within its stability field at 3 GPa and 530-560 °C, i.e., quite far from the antigorite-out boundary. Although discussing this point, they do not provide any rational explanation of this feature, which strongly suggests that dehydration results here from an experimental artefact. In the absence of an oxygen buffer within the platinum capsule, it is indeed very likely that some hydrogen escaped from the capsule during the experiment, giving rise to a loss of H2O, which would explain why antigorite is changing into olivine at grain boundaries. Although interesting to explore the feedbacks between deformation and antigorite dehydration, the amount of water released has nothing to see with the one occurring during antigorite dehydration within the field of antigorite-out, for which a massive H2O released is expected. Hence, focusing on the role of incipient (i.e., minor) dehydration is here not applicable to the conditions of intermediate-depth Earthquakes, particularly to test the hypothesis of dehydration embrittlement. I also have other minor comments below.
With my Best Regards,
Jacques Précigout
Minor comments/concerns
- Introduction or method section: a PT graph of the antigorite stability field with respect to the experimental conditions would be welcome to be added to introduce the topic and investigations.
- Line 24: some important references are missing, including Raleigh and Paterson (1965), Yamasaki and Seno (2003) and Peacock (2001).
- Line 79: "Backscatter", not "backscattered". And what do you mean by "Qualitative EBSD"?
- Figure S3: Upper- or lower-hemisphere pole figures? Both are written. The Pole figures also do not look like the ones you can get using MTEX (they look like a screenshot of the Aztec software). But if you used MTEX, I strongly recommend to change the color bar and add isocountours on pole figures, so that any fabric (if applicable) will bring better out.
- Line 143: "Backscattered", not "Back-scattered".
- Line 148: Again, it's confusing what type of software you have used to treat your EBSD data (MTEX or AZTEC)?
- Line 206: Maybe I missed the point, but I did not find the duration of the experiment M909.
- Line 229: Needle-like crystals.
- Figure 6: The sample number is M923, not M928, right?
- Line 288: "Cluster length IN the bin".
- Figure 9c: 10-6 or 10-5/s
- Line 305: "this is caused BY the transformation..."
- Figure 10: Please, make sure that planes and axes are properly written, i.e., not with curly bracket. For olivine (orthorhombic), it should be only axes, so between [], and for antigorite (monoclinic), it should be planes for a and b, and axis for c ((100), (010), [001]). Also, what are Fx, Fy and Fz?
- Line 316: "This study IS caused..."
- Line 356: Again, the formalism of the axes/planes is confusing. Please check.
- Line 364: the reference of Nagaya et al. (2014; EPSL) is missing.
- Line 384: Using "fast" or "slow" to define the change in strain rate is confusing and not appropriate, because in any case, the experimental conditions are way faster than the expected strain rates in subduction zones (except for co-seismic events).
- Discussion: I often found repetitions in this section. For instance, you mention the same thing in line 390 and line 410. I recommend to rework the discussion to avoid this.
- Line 416: I am just wondering if some EBSD could be done on the olivine veins, which would be very instructive in terms of deformation mechanisms.