the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 29 May 2024
Dislocation creep near the frictional-viscous transition in blueschist: experimental constraints
Abstract. Mafic oceanic crustal rocks at blueschist facies conditions are an important rheological component of subducting slabs and the interface at subduction plate boundaries. However, the mechanical properties and deformation mechanisms of glaucophane, a rheologically-controlling sodic amphibole in blueschists, are poorly constrained. To investigate its mechanical and microstructural properties, we conducted general shear constant rate and strain rate stepping experiments on glaucophane aggregates using a Griggs apparatus at temperatures of 700–750 °C, shear strain rates of ~3x10-6 to 9x10-5 s-1, varying grain sizes, and a confining pressure of ~1.0 GPa. The constant rate experiments show an initial stage of grain-size-dependent strain hardening followed by weakening associated with brittle slip along cleavage planes, kink-band development, cataclasis resulting in a fine-grained matrix, and dislocation glide. These experiments evolved to a steady-state stress that did not depend on starting grain size, showing evidence for subgrain development and dynamic recrystallization by bulge nucleation, interpreted to reflect dislocation creep with limited recovery by climb. The mechanical behavior and microstructures of glaucophane in our experiments are consistent with experiments on other low-symmetry minerals as well as microstructural observations from natural blueschists. The strain rate stepping experiments were used to develop a dislocation creep flow law for glaucophane with values of A = 2.23 x 105 MPa-n s-1, n = 3, and Q = 341 ± 37 kJ/mol. A deformation mechanism map comparing our dislocation creep flow law to an existing flow law for blueschist diffusion creep indicates dislocation creep should activate at lower temperatures, higher stresses and larger diffusion lengthscales. Viscosities predicted by our flow law for a typical subduction strain rate of 1 x 10-12 s-1 lie between quartz and eclogite dislocation creep for the blueschist stability field, implying that mafic oceanic crustal rocks remain strong relative to quartz-rich metasediments all along the subduction interface.
- Preprint
(35211 KB) - Metadata XML
-
Supplement
(27170 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-1507', Anonymous Referee #1, 25 Jun 2024
As main component in blueschist-facies mafic oceanic crust, the quantification of glaucophane’s mechanical behaviour at high-pressure, high-temperature conditions has been the centre of the present study by Hufford et al. To understand its deformation behaviour, the authors conducted several hydrostatic experiments in addition to a series of deformation tests using a Griggs-deformation apparatus. The microstructures of the recovered samples were carefully studied using analytical techniques such as polarised light microscopy and electron-backscatter diffraction. The experimental and microstructural data are all of high quality and support the interpretation proposed by the authors. Furthermore, the well written and structured text makes it easy to follow the authors’ reasoning. I enjoyed reading the manuscript, acknowledge the amount of experimental work presented, and find that the present study is a valuable contribution, which will be of interest to a broad geoscientific community. Therefore, I am highly supportive of this work and suggest publication in EGUsphere after minor revision.
A few minor comments:
Influence of grain size/porosity
The observed relation between grain size of the starting material and peak stress is very interesting (Fig. 2b). Could it, however, be that this apparent relation is only indirectly caused by grain size but rather a product of differences in porosity? As you started with powders, I assume that there is some initial porosity present in the starting material. This initial porosity could depend on grain size or rather the grain size distribution in the sample and influence the peak stress of the samples (e.g., Hirth & Tullis, 1991). Renner et al. (2007) experimentally showed that the measured strength of their quartz-calcite mixtures mainly depended on the amount of quartz, which itself positively correlates with porosity (Figs. 2; 6 in Renner et al., 2007). Even if the authors find no such relation between porosity and strength in their sets of experiments, it would be informative to provide measured/estimated porosities present in the starting material. Related to this comment, I find the statement in line 380-385 a bit too provocative as it does not include any discussion on other aspects that may influence rock strength such as existing textures that could act as planes of weakness.
Hydrostatic tests
It would be good to know whether or not the selected durations for hydrostatic pressing are somewhat related with the duration of deformation or rather with the duration of hot-isostatic pressing (HIP) prior to the onset of deformation. As noted in the manuscript, the hot-pressed powders are important to have an idea about the initial sample thickness, grain size, initial microstructure, etc., with which I fully agree. However, no explanation is given, why different run durations were chosen and it would be interesting to know on which basis you decided to run the hydrostatic tests for 24-88 h. Furthermore, the HIP runs were only conducted at 700 ˚C, although the deformation tests were run at 725 and 750 ˚ C as well. In the supplementary material, it is written that glaucophane becomes unstable at temperatures higher than 725 ˚C and that the run durations were chosen accordingly to avoid an impact of glaucophane breakdown on the mechanical data. I agree that this information belongs to the supplementary material. However, I find that the authors should explain in a bit more detail why they chose these hot-pressing durations/conditions in the main text.
Other comments:
Lines 197-199 and Fig. 2c: I agree that the reproducibility is very good for the 750 ˚C test (shear stress at 1st and 4th step), but for the other runs conducted at 700 and 725 ˚C, the reproducibility is far less striking. Please add some more explanation on how you decided that the sample reached steady state.
Lines 250-253: It would be helpful if you could be more specific and provide information on the experimental conditions and the type of material used in the studies by AveLallemant (1978) and Reynard et al. (1989).
Lines 272-275: Just a suggestion: The word ductile can be confusing as I have the impression that this term means different things in different geoscientific communities. Perhaps use: “transition from localized to distributed deformation”?
Citation: https://doi.org/10.5194/egusphere-2024-1507-RC1 -
RC2: 'Comment on egusphere-2024-1507', Anonymous Referee #2, 22 Jul 2024
Comments on the manuscript: “Dislocation creep near the frictional-viscous transition blueschist: Experimental constraints. By Hufford, Tokle, Behr, Morales, and Madonna.
The manuscript describes a set of experiments on polycrystalline glaucophane performed in a solid deformation apparatus. A flow law is derived from the data. The manuscript is concisely written and contains interesting experimental data and observations that seem to be worth publishing. However, there are some major problems with inferring the main conclusion of the present manuscript from the data set presented, i.e. that dislocation creep should be the dominant deformation mechanism in the experimentally deformed samples. For this reason, I do not see the manuscript fit for publication. I suggest to re-write the manuscript with different conclusions after further data collection and to re-submit a completely revised manuscript before it can be published.
The major problems of the manuscript are as follows:
(1) Dislocation creep is inferred for the deformed samples after an initial stage of peak stress and subsequent weakening during which grain size reduction occurs. Dislocation creep is a grain size independent deformation mechanism, so grain size should not matter for the activation of this deformation mechanism. Therefore, it is inconceivable why an initial grain size reduction should be required to activate dislocation creep in these samples. Furthermore, it is questionable whether a flow law can be derived from such experiments that are ambiguous or difficult to interpret.
(2) The authors infer that steady state deformation by dislocation creep is partially achieved by re-orientation of grains for easy slip after early stages of deformation. However, the documented CPO´s of samples with low to high strain and even hydrostatic initial treatment are not substantially different from one another in terms of a distribution of maxima for the different crystal orientations.
(3) The same authors have published a paper recently where they infer that diffusion creep in glaucophane is the rate-limiting deformation mechanism in blueschist samples at temperatures of 650-700C, while the deformation temperatures in the present study is 700-750C at the same pressures and similar strain rates. Flow stresses at 700C are similar in both studies, so how can in one case the rate limiting deformation mechanism be dislocation creep and in the other diffusion creep? Admittedly, the composition of samples in the two studies is different, glaucophane is inferred to be strain rate controlling in the former study, and for glaucophane very different deformation mechanisms are inferred in both studies.
(4) In the present study glaucophane is deformed well outside its thermodynamic stability field. In the light of the fact that glaucophane is a high pressure mineral, it is not clear why such a low pressure was used in the deformation experiments? (The solid medium apparatus typically can be operated routinely at pressures of 1.5 GPa or more) In any case, reconstitution of the glaucophane material is likely to involve some chemical changes (e.g. H2O content), because according to earlier studies, breakdown of glaucophane should occur. Even subtle changes in composition may introduce other driving potentials than strain energy for recrystallisation, which in such a case might not be strictly “dynamic”. Dislocation creep depends on the operation of recovery processes such as dynamic recrystallisation, so that this process (and to demonstrate its operation) is of key importance. In this manuscript, only light micrographs are presented, no SEM backscatter images that may indicate chemical changes. A thorough documentation of microstructures by SEM-BS or EDS maps is required to demonstrate that recrystallisation takes place without chemical changes.
Detailed comments:
Line 24: certainly, one or more works by P. Agard should be cited here.
Line 33: “hierarchy” instead of “heirarchy”
Fig. S2: the labels of the grain size ranges are partly missing (two brown coloured ranges). Furthermore, the 63 to 355 micron grain size fraction (according to the label) includes smaller grains, too. Please correct the figure or the text.
Lines 80-89 and lines 31-42 in supplementary data: the procedure employed here appears problematic for reasons concerning the stability of glaucophane. In a careful early experimental study of glaucophane stability by Maresch (1977) it was found that at 1 GPa the stability of glaucophane extends to about max. 525C. This value is about 200C lower than the temperatures employed in this study. In addition, it is unclear why relatively low pressures of 1 GPa were employed here, when the solid medium apparatus allows higher pressures to be applied routinely. More recent studies by Graham et al. (1989) demonstrated that the H2O content of glaucophane may vary systematically in Na- and other amphiboles with P and T. Corona et al. (2013) and Cheng et al. (2019) and other studies find solid solution effects well below the temperature range of the breakdown reaction of glaucophane. Only the breakdown melting conditions are cited in the manuscript, but other processes at much lower temperature conditions will be relevant here, too. From all of these studies it is obvious that at the experimental conditions employed here glaucophane is deformed well outside its thermodynamic stability field. Even if major decomposition reactions are not observed as indicated in the supplementary data, a more or less subtle shift in composition (e.g. by a change in H2O content) should be detectable in SEM-BS images or EDS maps. Some compositional shift is documented in Fig. S3. Such compositional changes may have important consequences for the recrystallisation process and its driving potentials. As dislocation creep depends on, e.g., recrystallisation as a recovery process, additional driving potentials, such as chemical potentials, affect the reconstitution process, so that the recrystallisation might not be termed “dynamic” (i.e. induced by strain energy only). In any case, the light micrographs presented in this manuscript as the only microstructural documentation are in many ways inadequate to document the full range microstructural features of the samples. Not only is the resolution insufficient, but also any chemical information is missing. This study requires thorough documentation of the microstructures by SEM-BS and possibly EDS maps.
Lines 117-118: Fig. 1 does not show a CPO, Fig. 4 shows CPO´s.
Lines 118-119: these light micrographs and their magnification are insufficient and inadequate to show the features required to document the details of this study (see comment above). SEM-BS and possibly ED maps are required here. For example, compositional changes near the grain boundary regions or along cracks, etc. would indicate additional processes that take place in these samples.
Lines 119-120: the strengths of CPO´s, appears to be approximately the same in all pole figures. The (001) pattern is dominated by a broad single maximum, not a girdle.
Lines 120-122: this sentence is obscure and difficult to understand: If there is crushing during hot pressing, how can the grain size distribution of hot pressed samples correlate with that of the initial grain fraction before hot pressing? Please explain. Furthermore, Fig. 1 does not show grain size distributions.
Line 138: the red arrow points to a discrete low angle boundary - this is not undulatory extinction (which is a change in crystal orientation over a distributed region).
Line 145: there is no undulatory extinction in Fig. 5d, only discrete low angle boundaries.
Lines 146-148: the maxima of (110) and (100) are inclined in an antithetic sense with respect to the shear sense in Fig. 4 for the peak stress sample. Consequently, the weak girdle for (001) is also inclined and not parallel to the shear plane (certainly not parallel to the shear direction!).
Line 152: it is not clear why the pull-apart structures and cracked zones should be kink-bands? Pull-apart and cracked zones adequately describes the observation. Again, it is impossible to observe undulatory extinction in these micrographs.
Line 155: again, there is no obvious indication for kink bands.
Lines 156-158: the point maxima in (110) and (100) are indeed subnormal to the shear plane, while the [001] direction is subparallel to the shear direction. All of these orientations appear slightly rotated from the peak stress orientation in terms of geometry, but they do not differ in, e.g., type or strength of the CPO.
Line 161-171: the low angle boundaries are interpreted here as subgrain boundaries in the sense of ordered dislocation substructures, whereas their morphology and general appearance does not differ form the low angle misorientations described above (which are interpreted as cracking features). There is no evidence presented here why these low angle boundaries should not be interpreted as healed fractures, like the same features above.
Lines 171-175: the (010) maximum is more or less the same as in the other pole figures, only it is more central to the pole figure. In the weakening sample pole figure the (010) point maximum is even better defined than in the steady state sample. The expression of maxima is mainly due to the much larger number of points for the pole figure (3 times as many in the steady state sample).
Lines 186-187: please define: what is “fine grained”? Obviously, there is a wide range of grain sizes in these samples. It is critical to know whether there is a defined range of small grains of, e.g., similar size in all samples or not, i.e., whether the size ranges, especially for the small grains, are identical in different samples or not. If dynamic recrystallization had caused the fine grain sizes, their size should be identical in all samples, because the flow stresses are very similar (piezometric relationship), and they should show a very narrow size range.
Lines 195-205: what is the significance of dashed and solid lines in Fig. 2d? This should be explained in the figure caption.
Lines 208-213: recrystallization is inferred as a grain size reduction mechanism here, while cracking, etc. is inferred to the place in larger grains in the same samples (e.g. crushing at large grains). It is unclear why cracking as a grain size reduction process is not considered here - why should the grain size reduction process change in these samples at the same conditions of P,T, and strain rate?
218-220: the microstructural features of slip on cleavage planes, grain size reduction, and kinking are inferred as the result of dislocation activity without any evidence given. The initial grain size reduction during hot pressing and after peak stress is inferred to take place by fracturing (referred to as “crushing”) above. The observed features in Fig. 6 are consistent with healed cracks producing internal misorientation in grains (“subgrains”), etc. There is no need to infer another processes than cracking (e.g., dislocation processes) if the microstructures are similar in the same samples at the same conditions of deformation. The crushed grains in peak stress samples are termed “core mantle structures” here (Fig. 6) - what are the data or observations to distinguish these core mantle structures from the different interpretation above (“grain crushing”)?
Lines 222-223: how can fewer slip systems be available, if the CPO-type (Fig. 4) looks more or less the same for coarse grained and steady state samples?
Lines 226-227: cleavage slip and granular rotation are consistent with brittle deformation processes, not necessarily with dislocation creep.
Line 228: dislocation creep with limited climb should result in strain hardening, not steady state. Cleavage slip is a type of frictional deformation, not a plastic process. It is difficult to conceive that a steady state stress results from a “transition to dislocation creep”. Why should “a transition stage” from different grain sizes to a fine grain size produce a steady state stress? The reasons given for the transition do not support the interpretation, because:
(a) stress exponents of 2 to 3 (observed here) are lower than 3 to 5 expected for dislocation creep.If frictional processes operate in conjunction with dislocation creep, the expected stress exponent should even be higher than 3-5, certainly not lower.
(b) dislocation creep is a grain size independent deformation mechanism, so grain size should not affect its efficiency and thus the rheology. For this reason, the inference of dislocation creep as the dominant deformation mechanism in the steady state part of deformation following a grain size reduction does not make sense.
(c) “subgrain formation” can easily be explained by non-plastic processes. Core-mantle-structures of Fig. 6 resemble grain crushing structures in “lower strain” samples or hot pressed samples, i.e., brittle deformation processes are inferred for similar microstructures at lower strain.
(d) there are no systematic changes in the CPO patterns that might account for a change in deformation mechanism.Lines 234-236: no clear evidence is given for inferring the recovery process of bulge nucleation.
Lines 301-304: this statement is quite correct; the cited studies do not document any deformation process that causes the CPO types. For this reason it is interesting to refer to these papers, but their CPO description does not contribute to the understanding or interpretation of the observed CPO patterns.
Lines 315-317: The operation of a (100)[001] slip system has not been demonstrated by observations of this study. The CPO´s shown in Fig. 4 do not change substantially from hydrostatic to steady state samples. Fig. 6 explains some of the “subgrain” misorientations by this and other slip systems, but their resolved shear stresses are extremely low for the large grain in the kinematic framework of the sample, so that it is, for geometric reasons alone, questionable that dislocation activity has caused their origin.
Lines 344-347: Reynard et al. (1989) show an image of small glaucophane grains but do not demonstrate that these have formed by dynamic recrystallisation.
Lines 355-359: As pointed out above, the range of n-values is lower than 3 to 5, so that dislocation creep (in conjunction with other deformation mechanisms such as friction) is unlikely. In the light of ambiguous microstructures presented here, together with an ambiguous database of n-values, the inferred deformation mechanism is one of the least probably deformation mechanisms that can be inferred from this data set. The activation energy given for dislocation creep is indistinguishable from or even lower than that given for diffusion creep by the same authors. This would be a very unusual result compared to known flow laws for silicates where the activation energy of diffusion creep typically is lower than that of dislocation creep, not the same or higher.
Lines 418-420: this statement is questionable, because blueschists contain substantial amounts of weak minerals (white mica, chlorite) that will deform and accommodate strain at considerably lower stresses than glaucophane at low temperatures, so that conditions for dislocation creep may not be reached in glaucophane.
,Citation: https://doi.org/10.5194/egusphere-2024-1507-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 337 | 70 | 23 | 430 | 22 | 14 | 17 |
- HTML: 337
- PDF: 70
- XML: 23
- Total: 430
- Supplement: 22
- BibTeX: 14
- EndNote: 17
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1