3.1 Sampling

Samples were obtained from two sub-horizontal drill cores from the underground facilities that intersect the BFZ045 fault (Fig. 2a). The analysed samples were selected from (i) a 2 m continuous section along drill core PH28, which was drilled at a depth of 433 m b.s.l. and is oriented ESE–WNW and (ii) drill core PH16 from the “Demonstration Facilities” tunnels at 420 m depth, described in Aaltonen et al. (2016). The PH28 samples were selected from the 99–101 m interval of the drill core, with core distance measured from ESE to WNW, which includes the fault core and the proximal damage zone. The selected interval of the drill core was cut in half parallel to the stretching lineation and perpendicular to the mylonitic foliation and subsampled at regular intervals of 2 cm. Nineteen polished thin sections were added to three samples from PH16 previously described in Aaltonen et al. (2016). An additional thin section from drill core PH21 (drilled at the same depth of PH16) was used to estimate the T of the graphite formation in the host rock using Raman spectroscopy of carbonaceous material (see Sects. 3.2 and 4.5).

3.2 Microstructural observations and electron backscatter diffraction (EBSD) analysis

Deformation microstructures were studied using petrographic and scanning electron microscopy (SEM). SEM and EBSD analyses were performed at the Plymouth Electron Microscopy Centre using a JEOL LV6610 SEM and a JEOL 7001 FEG-SEM. Thin sections used for electron backscatter diffraction (EBSD) analysis were polished with colloidal silica before being carbon coated. Data were acquired on a NordlysNano and a NordlysMax EBSD detector (Oxford Instruments). Working conditions during acquisition of the EBSD patterns were 20 kV, 20 mm working distance, 70^∘ sample tilt, and high vacuum. AZtec software was used for pattern indexing on rectangular grids with step sizes of 0.7, 1, and 1.8 µm. EBSD patterns were processed with the Channel 5 software (Oxford Instruments), and noise reduction was performed following the procedure suggested in Bestmann and Prior (2003).

Differential stresses during mylonitic creep were estimated using the recrystallized grain size piezometer for quartz by Cross et al. (2017). The method relies on the separation between relict and recrystallized grains based on the grain orientation spread (GOS), which is a measure of the internal strain of a grain defined as the average misorientation angle between each pixel in a grain and that grain's mean orientation (Wright et al., 2011). Further details on the presentation of the EBSD data and on the recrystallized grain size piezometer for quartz by Cross et al. (2017) are presented in the Supplement S1.

3.3 Mineral chemistry and Raman spectroscopy

Major element mineral chemistry of chlorite and white mica was measured with electron microprobe analysis (EMPA) at the Department of Earth Sciences, University of Milan, Italy. Carbon coated thin sections were probed with a JEOL 8200 Super Probe equipped with five wavelength-dispersive spectrometers (WDS). Working conditions were 15 kV of probe voltage, 5 nA of current on sample, and 1 µm beam diameter. Natural minerals were used for standardization; measurement times were 30 s on peaks and 10 s on backgrounds of the X-ray lines.

Raman spectroscopy was applied for feldspar and opaque phase identification and for carbonaceous material (CM) characterization. Data acquisition was conducted at the Department of Chemistry, University of Padua (Italy), using a Thermo Scientific DXR MicroRaman spectrometer, equipped with a 532 nm depolarized laser. Raman analysis was also carried out at the Department of Mathematical, Physical and Computer Sciences of the University of Parma (Italy) using a Jobin Yvon Horiba LabRam spectrometer equipped with He–Ne laser (emission line 473.1 nm) and motorized xy stage. Spectra were acquired from polished thin sections using a laser power of 5 mW, spectrograph aperture 25 µm pinhole, and 50× or 50× short-distance objective. The estimated spot size was 1–2 µm in diameter and the spectral resolution was 2 to 4.4 cm⁻¹, with an acquisition time of 30–90 s. Feldspar composition was classified on the basis of the acquired Raman spectra, as suggested in Freeman et al. (2008), using a comparison with standard Raman spectra from the RRUFF Project database (Lafuente et al., 2016). To assure a good statistical analysis of the CM structural heterogeneity, only samples with >10 CM spectra were taken into consideration. Omnic software (Thermo Fisher Scientific) was used for Raman spectrum decomposition, using the software's Lorentzian/Gaussian function, following the procedure described in Kouketsu et al. (2014). Raman analytical conditions are detailed in the Supplement S1. A discussion about the analytical limitation of the chosen methods for temperature estimation is also presented in the Supplement S1.

Figure 3Polarized light microscope images of the damaged host rock (a) and representative quartz microstructures (b–c). Cross-polarized light is used. (a) Host rock showing the original magmatic textures and mineral assemblage. White arrow shows a chlorite-rich shear band oriented at a low angle to the mylonitic foliation of BFZ045. (b) Quartz with wide extinction bands (WEBs) and undulatory extinction. WEBs are bounded by sets of fluid inclusion trails (FI trail, dashed lines). Intercrystalline deformation bands and well-developed FI trails developed subparallel to the mylonitic foliation. (c) Polygonal recrystallized quartz grains, with small grain size (∼20 µm), forming bands oriented subparallel to the mylonitic foliation. The white arrow shows sutured grain boundaries between magmatic quartz grains, indicative of bulging. Foliation trace is projected as a white line onto the images from the adjacent foliated host rock.

Download

4.1 BFZ045 fault zone structure

A schematic representation of the BFZ045 fault geometry obtained from underground field mapping and detailed characterization along the horizontal PH28 drill core is represented in Fig. 2a. A 10–100 cm wide mylonite with a sinistral sense of shear, oriented 61∕075 (dip ∕ dip direction), hosts a sub-vertical network of cohesive cataclasites with an average orientation of 87∕095 (Fig. 2b). Rodding of quartz and feldspar in the mylonite defines a stretching lineation oriented 10∕168 (plunge ∕ trend; Fig. 2b, c). Slickensides with chlorite mineral striations are abundant throughout BFZ045 and the average orientation of the striations is 07∕169; stepped slickensides indicate a dominant sinistral sense of shear, although striations associated with dextral kinematics have also been observed (Aaltonen et al., 2016; Nordbäck and Mattila, 2018).

A network of veins typically filled with chlorite, quartz, and calcite, as well as chlorite-filled fractures, overprint the mylonite and host rock (Fig. 2a; Aaltonen et al., 2016). The damage zone is localized between 0.5 and 1 m from the fault core, and it is characterized by an increased fracture density towards the fault core, with the average spacing between fractures decreasing from 3 to ca. 0.5 cm. With the term “fault core”, we refer here to the brittle core of BFZ045 defined by cohesive cataclasites that overprint the mylonite. This means that the damage zone of BFZ045 affects both the migmatitic host rock and the BFZ045 mylonite.

In Fig. 2d, representative samples of core PH28 are shown from left to right in a sequence from the damage zone in the host rock to the fault core. The host rock is a coarse-grained veined migmatite (Fig. 2dI). The extent of damage in the host rock was defined by the farthest occurrence of chlorite-filled fractures, identified microscopically as deformation bands associated with visible slip. A total of 12 fractures longer than 5 cm can be observed along the core, of which 7 are within a distance of 10 cm from the fault core (west side of the core). The contact between the host rock and the mylonite is sharp. The millimetre-spaced mylonitic foliation is defined by a compositional layering of alternating quartzofeldspathic domains and mica-rich domains (Fig. 2dII). Multiple slip surfaces marked by 0.5–10 cm thick cataclastic domains overprint the mylonite along the foliation (Fig. 2dIII). Locally, phyllosilicates and trails of opaque minerals define thin (<1 mm thick) anastomosing foliation planes within the cataclasites, which wrap around subangular fragments of the mylonitic precursor. Along a thin (<5 mm thick) slip surface, two pseudotachylyte injection veins intruding the mylonite at a high angle have been observed, which demonstrates the transient seismogenic behaviour of BFZ045. Calcite veins (1–3 mm thick) locally overprint the fault core both at high angle to the foliation and along the slip surfaces.

Figure 4Microstructure of the BFZ045 mylonite. Cross-polarized light is used. (a) The mylonitic foliation is defined by the alternation of quartz-rich and mica-rich bands. Fractured porphyroclasts of feldspars are preferentially located in mica-rich domains. (b) Quartz ribbons in the shear zone boundary are stretched along the foliation and show typical core-and-mantle microstructure, with recrystallization localized at the grain boundaries. Ribbons also contain well-developed WEBs. (c) Completely recrystallized quartz ribbon. The recrystallized grains show a shape preferred orientation indicating a sinistral sense of shear. Thin muscovite (Ms) layers define the mylonitic foliation, together with the elongated and recrystallized quartz domains. (d) Quartz veins along the foliation, filling a mode I fracture at a distance of ∼1 cm from the cataclasite fault core. Quartz grain elongation in the veins is normal to the vein wall and to the foliation. (e) High-magnification view of the quartz vein. Bulges along the grain boundaries, and subgrains within the grains, are visible.

Download

4.2 Petrography and microstructures

4.2.2 Mylonite

The relative mineral abundances in the mylonite are slightly different from those in the host rock and consist of 50 % quartz, 20 % white mica + chlorite, and 30 % K-feldspar + plagioclase. Accessory phases are graphite, rutile, anatase, and apatite, which are typically found associated with chlorite to form black seams. The spaced mylonitic foliation is defined by a compositional layering between alternating millimetre-thick quartz bands and narrower (0.2 to 1 mm thick) mica- and feldspar-rich bands (Fig. 4a). Porphyroclasts of K-feldspar are up to 7 mm in size and show asymmetric pressure shadows filled with chlorite + muscovite ± feldspars (albite and K-feldspar), with a geometry indicative of a sinistral sense of shear. Feldspars in the mylonite present brittle behaviour, with domino-type fragmented porphyroclasts. Neoblasts of K-feldspar are commonly observed in the pressure shadows of porphyroclasts or together with fine-grained white micas along the mylonitic foliation.

Figure 5Microstructure of the BFZ045 brittle fault core. Cross-polarized light (a–c), plane-polarized light (a), and scanning electron microscope backscatter electron images (SEM-BSE) (d–e). (a) Cataclasite with characteristic pseudotachylyte injection veins (arrow). The cataclasite matrix is enriched in chlorite and Ti oxides. (b) Contact between the mylonite (upper left corner) and the cataclasite. Quartz forms almost entirely recrystallized polycrystalline ribbons. (c) Detail of a subangular polycrystalline clast of quartz in the cataclasite. The trace of the mylonitic foliation is still visible in the clast, and it is only slightly rotated with respect to the trace of the foliation in the mylonite. The surrounding matrix is a fine mixture of white mica and plagioclase. (d) Pseudotachylyte injection vein. Rounded quartz clast (dark grey) and Ti oxides (white) are surrounded by a chlorite- and mica-rich ultrafine-grained matrix. (e) Quartz + chlorite aggregates filling fractures within and strain shadows around plagioclase porphyroclast in the mylonite. The clast is slightly rotated within the protocataclasite. The white line in the top right corner indicates the trace of the mylonitic foliation.

Download

Quartz in the mylonite presents various degrees of recrystallization (Figs. 4b, c). In zones where ductile deformation is less prominent (e.g. samples PH28_3, PH16_1; Fig. 2), quartz retains a coarser grain size (1–5 mm), and it forms slightly asymmetric (sigmoidal shaped) grains. Internally, the grains display patchy undulatory extinction, well-developed blocky to elongated WEBs (100–300 µm wide), and discrete intracrystalline bands (<200 µm wide) of bulges and recrystallized grains preferentially oriented subparallel to the foliation and at ca. 45^∘ from the foliation, measured anticlockwise (Fig. 4b). The host grains contain small subgrains (<60 µm), which, towards the boundaries of the host grain, transition to aggregates of recrystallized grains of sizes comparable to the subgrains, forming typical core-and-mantle microstructures (Fig. 4b).

Figure 6EBSD data of quartz from the mylonitic fault core. In all the EBSD maps, black lines correspond to high-angle boundaries (misorientation >10^∘), white lines to low-angle boundaries (misorientation between 2^∘ and 10^∘), and red lines to Dauphiné twin boundaries (misorientation of 60^∘ around the c axis). (a) Quartz ribbons and intracrystalline bands of recrystallized grains (Fig. 5a). Cross-polarized light is used. The box indicates the EBSD map shown in (b)–(d) and (g). (b–c) Grain size map (diameter of the equivalent circle in micrometres, µm), the higher magnification in (c) highlights the presence of subgrains in the large relict quartz grains and in the recrystallized grains. (d) Grain orientation spread (GOS) for each grain coloured relative to the GOS threshold (black line) between recrystallized (blue) and relict (red) grains. (e) Histogram of grains size distribution of grains in the intercrystalline bands. (f) Histogram of subgrain size distribution of subgrains in the relict quartz from the intercrystalline bands (light red in d). (g) Grain relative orientation distribution (GROD) maps were used to estimate visually the subgrain size in the quartz ribbon.

Download

In zones of complete recrystallization, quartz forms highly elongated polycrystalline ribbons (up to 0.5–1 mm thick and up to 2 cm long) parallel to the foliation (Fig. 4c). The recrystallized grains locally define a shape preferred orientation (SPO in Fig. 4c) inclined with 10 to 30^∘ with respect to the trace of the foliation, consistently with the bulk sinistral sense of shear. Adjacent to the cataclastic fault core (sample PH16_3, Fig. 2), ataxial/unitaxial (i.e. with no visible median line; Bons, 2012) quartz veins occur parallel to the mylonitic foliation (and to the layers of recrystallized quartz) and contain grains elongated normal to the vein boundary (i.e. normal to the foliation; Fig. 4d). The vein crystals range in length from 200 to 400 µm and have a maximum thickness of 150 µm measured parallel to the vein. Quartz in the veins shows undulatory extinction and bulges at the grain boundaries (Fig. 4e) indicative of crystal plastic deformation. The recrystallized quartz in the mylonite surrounding the vein has a finer grain size than the one in the mylonite described in Fig. 4a–c, which is located farther away from the cataclastic core. Veins of radiate chlorite are observed cutting the mylonitic foliation at a high angle (∼60^∘).

Figure 7EBSD data of quartz from a recrystallized ribbon in the mylonite. (a) Polycrystalline ribbons of recrystallized quartz grains elongated parallel to the mylonitic foliation (Fig. 5b). Cross-polarized light is used. The box indicates the EBSD maps shown in (b)–(d). Colour coding of the boundaries like in Fig. 6. (b, c) Grain size map (diameter of the equivalent circle in micrometres, µm) and detail (c) showing that the larger grains contain subgrains of the same size as the surrounding finer grains. (d) GOS map showing that the GOS values are mostly under the threshold, indicative of high degree of recrystallization. (e) Histogram of the grain size distribution for recrystallized and relict grains. (f) Histogram of the subgrain size distribution in the relict quartz grains identified in (d) and (e). (g) Pole figure of the c-axis orientation of recrystallized and relict grains, colour coded like the GOS map in (d) (equal area, lower-hemisphere projection).

Download

4.3 EBSD and grain size analysis of quartz

4.3.1 Mylonite

EBSD analysis of the mylonite was conducted on sample PH16_1, which is located at the mylonitic shear zone boundary at a distance of 4 cm from the brittle fault core (Fig. 2a). EBSD maps were acquired from intracrystalline bands of recrystallized grains within an elongated millimetre-sized quartz grain (Fig. 6a) and from a highly recrystallized quartz layer along the mylonitic foliation (Fig. 7a). The recrystallized grain size within the intracrystalline bands ranges from 5 to 60 µm (Fig. 6b, c). The grain orientation spread (GOS) within the recrystallized bands varies between 0 and 8.4^∘, with a threshold value of 1^∘ between the recrystallized grains and the relict grains when analysing the trade-off curve proposed by Cross et al. (2017). The average grain size of recrystallized grains (GOS <1^∘) is 16±7 µm, whereas relict grains (GOS >1^∘) have an average grain size of 25±9 µm. Relict grains contain subgrains of an average size of 17±7 µm (Fig. 6f). Quartz grain hosting the intracrystalline band shows subgrains of approximately 25–50 µm in size, which is comparable to the size of the coarser recrystallized grains observed in the intracrystalline bands (Fig. 6c). The size of the subgrains in the host quartz was estimated visually with the aid of grain relative orientation distribution maps (GROD, Fig. 6g).

Figure 8EBSD data of quartz from a foliation-parallel vein in the mylonite near the contact to the cataclasite. Colour coding of the boundaries like in Fig. 6. (a) Quartz vein along the foliation. Quartz in the mylonite shows a strong SPO consistent with the sinistral sense of shear of BFZ045. Cross-polarized light is used. (b) Grain size map (diameter of the equivalent circle in micrometres, µm). (c) Details of previous maps. Bulges and subgrains (white arrows) of similar size of the bulges are evident within the quartz vein. (d) GOS map of quartz in the vein and of the surrounding mylonite. The GOS threshold value of 1.94^∘ separates relict grains (red) from recrystallized grains (blue). Grey stars indicate relict grains plotted in (f) for subgrain size estimates. (e) Histogram of the grain size distribution in the mylonite, with relict and recrystallized grains separated with the GOS method. (f) Histogram of the subgrain size distribution (red) in relict quartz of the mylonite (grey). (g) Histogram of the recrystallized grain size (red) and subgrains size (blue) distribution in the vein. (h–i) Pole figures of the c-axis orientation of the recrystallized and relic quartz in the mylonite (h) and in the veins (i) (equal area, lower-hemisphere projection).

Download

In the recrystallized quartz layer (Fig. 7a), quartz grain shapes range from equant to elongate parallel to the foliation, with grain sizes ranging from 5 to 87 µm (Fig. 7b, c). GOS analysis identified a threshold value of 1.56^∘ to separate recrystallized and relict grains (Fig. 7d). Average grain size of the recrystallized grains is 18±8 µm, while relict grains have an average size of 28±11 µm (Fig. 7e). The relict grains contain subgrains of an average size of 17±7 µm (Fig. 7c, f). The crystallographic preferred orientation (CPO) of the c axis of the relict grains and recrystallized grains forms a single girdle consistently inclined with the sinistral sense of shear of the sample (Fig. 7g). The EBSD-calibrated recrystallized grain size piezometer for quartz by Cross et al. (2017) was used to estimate the differential stresses during plastic flow in the mylonite. The estimated differential stress is 73–80 MPa for the average recrystallized grain size of 16–18 µm.

Figure 9EBSD data of quartz from a clast in the cataclasite. Colour coding of the boundaries like in Fig. 6. (a) The analysed quartz clast (Fig. 5f). The white rectangle indicates the EBSD map shown in (b)–(d). (b)–(c) Grain size map (diameter of the equivalent circle in micrometres, µm). The map highlights the presence of subgrains in the coarser elongated quartz grains with size comparable to the surrounding finer quartz. (d) GOS map. (e) Histogram of the grain size distribution of the recrystallized (blue) and relict (red) grains. (f) Pole figure of the c-axis orientation of recrystallized and relict grains. (equal area, lower-hemisphere projection, colour coded as GOS map).

Download

4.3.2 Veins parallel to the mylonitic foliation

The quartz vein parallel to the foliation shown in Figs. 5c and 8a was analysed to identify possible evidence of crystal-plastic deformation and dynamic recrystallization. Grain shape varies from fibrous with elongation perpendicular to the vein wall to more equant and less elongate. Grain boundaries of vein crystals are straight to lobate, which is the latter most commonly observed in association with fine recrystallized grains and bulges (<15 µm in size) (Fig. 8b, c). Irrespective of their shape, most of the grains contain low-angle boundaries and Dauphiné twins. The low-angle boundaries are typically arranged to define polygonal to slightly elongated domains of ∼10 µm in size, comparable to that of the surrounding recrystallized grains in the mylonite (Fig. 8c).

Quartz in the mylonite flanking the vein shows a fine grain size (<20 µm), with only a small fraction of coarser grains (30–60 µm). GOS analysis indicates that dynamic recrystallization is pervasive. The mean recrystallized grain size is 10±3 µm, and the relict average grain size is 20±9 µm, with a GOS threshold value of 1.94^∘ (Fig. 8d, e). Selected relict grains (size >40 µm) in the mylonite contain subgrains with sizes in the range of 5 to 25 µm (Fig. 8f). Finer grains observed within the vein also present a GOS value below the threshold, which suggests that they represent the recrystallized fraction in the quartz vein (Fig. 8c, f). In the quartz vein, GOS analysis indicates that the average subgrain size is 24±7 µm, although the largest fraction of subgrains is smaller than 15 µm, i.e. similar in size to the recrystallized grains in the flanking mylonite and in the vein itself. The c-axis CPO of the recrystallized grains in the mylonite forms a single girdle synthetically inclined with the sinistral sense of shear (Fig. 8h). The c-axis CPO of recrystallized grains in the veins overlaps with the one of the relict grains (Fig. 8i). The differential stress estimated from the average recrystallized grain size in the mylonite (10 µm) is 106 MPa.

Figure 10Light microscopy and SEM backscattered electron (BSE) images of characteristic microstructural domains and mineral assemblages used for geothermobarometry estimates. (a) Graphite flakes in association with radiate chlorite in the host rock. Plane-polarized light is used. (b) Detail of graphite grains along the mylonitic foliation of BFZ045. Plane-polarized light is used. (c) Chlorite in pressure shadows around a K-feldspar porphyroclast in the mylonite. White mica and a fine grained recrystallized K-feldspar assemblage is common along the foliation. (d) Light microscope image of a microstructure similar to (c) (cross-polarized light). (e–f) SEM BSE images of radiate chlorite used for chlorite thermometry in the cataclasite matrix (e) and in a vein cutting the mylonitic foliation (f). The trace of the mylonitic foliation in (f) is oriented ca. NW–SE.

Download

4.3.3 Cataclasite

We analysed a largely recrystallized quartz clast in the cataclasite from sample PH28_10 (Fig. 9a). The selected clast is rotated of less than 10^∘ with respect to the adjacent mylonitic foliation. Quartz grain size in the clast ranges from 4 to 60 µm. The coarser grains are elongated parallel to the foliation, show bulges and fine recrystallized grains at their boundaries, and contain a high density of low-angle boundaries (Fig. 9b). The low-angle boundaries define small polygonal domains of a size comparable to the one of the recrystallized grains found at the grain boundaries (Fig. 9c). The GOS map in Fig. 9d identifies two grain size distributions, separated by a GOS threshold value of 3.23^∘. The recrystallized grains (average grain size: 8±4 µm) form equigranular aggregates at the boundaries of the coarser (average grain size: 17±10 µm) elongated relict grains. The c-axis CPO of the recrystallized grains and of the relict grains is the same, showing two maxima at an intermediate position between the centre of the pole figure and its periphery, and consistently inclined with the sinistral sense of shear of the sample (Fig. 9f). The differential stress estimated from the average recrystallized grain size in the clast (8 µm) is 123 MPa. Although the map has been acquired from a clast, these microstructures and recrystallized grain sizes are similar to those of the mylonite in the immediate vicinity (<2 cm) of the brittle fault core (Fig. 5b, c).

Figure 11Results of P–T estimates. (a) Carbonaceous material Raman geothermometer. Average R2 ratio (refer to text for explanation) for graphite-rich mylonitic and host rock samples was measured to derive a T estimate using the method by Beyssac et al. (2002) and Aoya et al. (2010). (b) Estimated P of mylonitization using the Si-in-phengite barometer (Massonne and Schreyer, 1987) for the average T of 440 ^∘C obtained with the carbonaceous material Raman thermometry. Red line shows the total spread of the Si values obtained. Black square shows the (c) chlorite compositional diagram based on Hey (1954). (d) Chlorite formation temperature estimated for mylonitic foliation, veins, and cataclasite using the method by Lanari et al. (2014).

Download

4.4 Mineral chemistry, Raman spectroscopy, and pressure–temperature (P–T) conditions of deformation

We estimated the P–T conditions of mylonitic deformation using Raman spectroscopy of carbonaceous material (RSCM), chlorite thermometry, and phengite barometry. Carbonaceous material was observed as grains and aggregates ranging in size from ∼50 to ∼200 µm in the host rock (sample PH21_1, Fig. 10a) and as smaller grains (20–50 µm) along the mylonitic foliation (sample PH16_1–2, Fig. 10b), along chlorite- and muscovite-rich layers. Data from Raman peak deconvolutions are reported in the Supplement S1. We estimated a maximum temperature of 530 ^∘C ± 50 ^∘C for the carbonaceous material in the host rock (Fig. 11a) (using the thermometer calibration for a laser wavelength of 514 nm; Beyssac et al., 2002), and a lower T of 440 ^∘C ± 50 ^∘C for the mylonite (Aoya et al., 2010, using the thermometer calibration for a laser wavelength of 532 nm; Fig. 11a).

Table 1Representative compositions of muscovite from the BFZ045 mylonite.

Download Print Version | Download XLSX

The pressure during mylonitization was estimated using the Si-in-phengite geobarometer (Massonne and Schreyer, 1987). Representative compositions of white mica are listed in Table 1. The full dataset of chemical compositions of white micas is reported in the Supplement S1. White mica composition was measured for grains parallel to the foliation associated with neoblasts of K-feldspar, as the application of the Si‐in‐phengite geobarometer requires stability of K-feldspar (Fig. 10c, d). Structural formulae were calculated assuming 11 oxygens. The range of Si apfu (atom per formula unit) in the probed white mica grains is 3.12–3.16. This compositional range indicates a pressure of 2–4 kbar for the mylonitization (Fig. 11b), considering the average temperature of 440 ^∘C derived for the mylonite with the graphite thermometry.

Chlorite composition was determined for (i) chlorite grains intergrown with quartz and muscovite in the strain shadows around feldspar porphyroclasts in the mylonite (Fig. 10c, d), (ii) chlorite flake aggregates in the cataclasite quartz matrix (Fig. 10e), and (iii) radiate chlorites filling veins at high angle to the mylonitic foliation (Fig. 10f). The structural formula of chlorite was calculated based on 14 oxygens, and representative compositions are shown in Table 2. The full dataset of chemical compositions for chlorite is reported in the Supplement S2.

Table 2Representative compositions of chlorite from different domains of BFZ045.

Download Print Version | Download XLSX

Chlorites along the mylonitic foliation have an Si content between ∼2.60 and 2.70 apfu, Al content between 2.74 and 2.88 apfu, and they are moderately Fe rich with a XFe ($\mathrm{XFe}=\mathrm{Fe}/(\mathrm{Fe}+\mathrm{Mg})$) between 0.62 and 0.82. Chlorites in the mylonitic sample from PH_16 have a more narrow range of XFe, between 0.57 and 0.73 (see the Supplement S1). In the cataclasite chlorites Si content ranges between 2.53 and 2.772 apfu, and Al contents have a wider range of 2.61–2.94 apfu (Supplement S1). The cataclastic chlorite is richer in Fe compared to the mylonite, with a XFe range between 0.71 and 0.85. The radial chlorite filling the veins cutting the mylonite has an Si content between 2.51 and 2.780 apfu, Al between 2.74 and 3.00 apfu, and XFe varies between 0.69 and 0.81. In general, BFZ045 chlorites have a aphrosiderite-ripidolite composition and the microprobe results show that the composition of distinct chlorite generations is similar (Fig. 11c). The CHL(2) semi-empirical thermometer by Lanari et al. (2014) was applied to each EMPA analysis of chlorite with Si <3 apfu and (Na + K +  Ca) <0.1 apfu. FeO was used as Fe total, and $a_{{\mathrm{H}}_{\mathrm{2}})\mathrm{O}}=\mathrm{1}$ and $a_{{\mathrm{SiO}}_{\mathrm{2}}}=\mathrm{1}$ were assumed. The estimated temperature for the mylonite ranges from ∼380 to 500 ^∘C (limit of the used thermometer) for an assumed P of 3.5 kbar (Fig. 11d), with a temperature variation of 10 ^∘C every 0.5 kbar increment. Temperatures estimated for the cataclasite matrix and for the radiated chlorite in the veins have a larger range (∼300–500 ^∘C, Fig. 11d). Compositions yielding T lower than 400 ^∘C are more frequent, especially in the cataclasite, but no clear relationship could be discerned between temperature and microstructural position of the chlorites.

Figure 12Schematic representation of the microstructural evolution of BFZ045. (a–d) Grey lines: traces of metamorphic foliation in the host rock. In the schematic fault scheme evolution, red displays the active deformation process, grey represents the inactive deformation processes, and orange suggests a transition in the deformation. (a) The development of mylonite was punctuated by the emplacement of quartz vein (b). (c) Ductile deformation localized toward the centre of the mylonitic fault core in an overall narrowing shear zone, and it was followed by formation of cataclasite, chlorite veins, and pseudotachylyte (d). (e) Schematic summary of the quartz recrystallized grain size, differential stresses, and P–T conditions of deformation for BFZ045 derived in the present study, in relation to the fault core geometry. Each sample is coloured differently to indicate the spatial position of the results described in Sect. 4 of this paper.

Download

5.1 The sequence of deformation events in BFZ045: ductile–brittle deformation history

Our microstructural observations are consistent with the general conclusion that brittle deformation along BFZ045 exploited a ductile (mylonitic) precursor (Nordbäck and Mattila, 2018; Skyttä and Torvela, 2018). Veins, cataclasites, and pseudotachylytes are localized along the mylonitic fabric of BFZ045, and only minor evidence of brittle deformation (mostly in the form of fractures filled by chlorite) is present outside of the mylonitic fault core (Figs. 2a, 3a). The first stage of deformation of BFZ045 is represented by the development of a localized N–S-trending mylonitic foliation (Figs. 2, 4a–c). Mylonitic creep was punctuated by transient brittle events, with the opening of extensional fractures along the mylonitic foliation filled by quartz veins (Fig. 4d, e). Veining was again followed by mylonitic deformation, as indicated by dislocation creep and dynamic recrystallization microstructures of quartz in the veins. Mylonite and veins were then overprinted by brittle deformation that formed, centimetre-thick cataclasites (Fig. 5a–c) and a <0.5 cm thick pseudotachylyte (Fig. 5d) that, together, form the brittle fault core of BFZ045. Cataclasites and pseudotachylytes were not overprinted by mylonitic creep, which might indicate that they formed under conditions favourable to predominantly brittle deformation along BFZ045.

The observed parallelism between stretching lineation in the mylonite and chlorite slickenlines in the cataclasites suggest that the ductile–brittle deformation history of BFZ045 occurred under a prolonged strike-slip regime with sinistral kinematics (Fig. 2b,c). This conclusion is consistent with the model of the brittle evolution of SW Finland proposed by Mattila and Viola (2014), which attributes the sinistral kinematics of NNW–SSE-trending sub-vertical faults to the ductile–brittle transition stage of the basement at ca. 1.75 Ga (“Stage 1” of deformation in Mattila and Viola, 2014). BFZ045 experienced later reactivations during the prolonged brittle history of the SW Finland basement, as indicated by calcite veins cutting the brittle fault core and by (rare) slickenlines with dextral kinematics observed in the underground exposures (Aaltonen et al., 2016). Local dextral kinematics along BFZ045 are potentially consistent with “Stage 2” of deformation by Mattila and Viola (2014) at 1.7–1.6 Ga. These later features, however, are not discussed further in this paper as they are subordinate to, and did not obliterate, the earlier history.

Table 3Parameters used for λ–σ failure mode diagrams.

Download Print Version | Download XLSX

5.2 Deformation mechanisms and stress history during mylonitic creep of BFZ045

Microstructural observations show that ductile deformation of the mylonite BFZ045 was accommodated in quartz by dislocation creep, while feldspars experienced dominantly brittle behaviour with only limited neocrystallization in pressure shadows (Fig. 10c). Given the lack of crystal plastic deformation and recovery features in the interior of feldspar porphyroclasts, we interpret the neocrystallization in pressure shadows as a possible indication of dissolution–precipitation, which has been commonly reported in feldspars deforming at mid-crustal conditions (Stünitz and Gerald, 1993; Menegon et al., 2008; Eberlei et al., 2014; Torgersen et al., 2015; Giuntoli et al., 2018; Hentschel et al., 2019).

Quartz recrystallization microstructures (bulges at grain boundaries and within intracrystalline bands, and core and mantle microstructures with subgrains of comparable size to that of the recrystallized grains; Figs. 4, 5) suggest that bulging and subgrain rotation were the dominant recrystallization mechanisms (Hirth and Tullis, 1992; Stipp and Kunze, 2008). The average recrystallized grain size of quartz ranges between 8 and 18 µm, which falls within the <40 µm size value representative of bulging as the dominant recrystallization mechanism (defined as slow grain boundary migration coupled to localized subgrain rotation at the mantle of the host grain; Stipp et al., 2010). In all the studied samples, GOS analysis distinguished two quartz grain size populations, in which the coarser (relict) grain size contains subgrains of size comparable to the average size of the finer recrystallized grains (Figs. 7–9). The CPO of relict and recrystallized grains is the same, and this is consistent with the host-controlled development of a CPO during subgrain rotation recrystallization (e.g. Stünitz et al., 2003).

A key observation in the BFZ045 mylonite is the decrease in recrystallized grain size of quartz from the shear zone boundary (16–18 µm) towards the centre (8–12 µm; Figs. 7–9). The coarser (16–18 µm) grain size is the most representative of the partially recrystallized quartz ribbons (Fig. 7) and of the nearly completely recrystallized quartz ribbons (Fig. 8) in the BFZ045 mylonite at distances ≥4 cm from the brittle fault core, and it also occurs as relict grain size in quartz clasts embedded in the cataclasite, where it is overprinted by the finer (8–12 µm) recrystallized grains (Fig. 10). We interpret this overprint and the overall decrease in recrystallized grain size to reflect a progressive increase in stress and strain rate towards the shear zone centre during mylonitic creep (Kidder et al., 2016). Throughout the samples the CPO of recrystallized quartz is consistently inclined with a sinistral sense of shear (Fig. 6–9); this is interpreted as evidence that samples developed under constant kinematic conditions in BFZ045. Relict quartz grains of 25–28 µm in size are associated with the coarser (16–18 µm) grain size (Figs. 7–8), and un-recrystallized portions of quartz contain subgrains of similar size (≥25 µm, Fig. 7b–d). We consider that the 25–28 µm grain size population might represent an early, dynamic recrystallization event within BFZ045, which was later overprinted by recrystallization occurring under progressively increasing differential stress.

Deformation and stress history of BFZ045 can be summarized in different stages. After the regional metamorphic peak and migmatite formation (i.e. 660–700 ^∘C, 3.7–4.2 kbar; Tuisku and Kärki, 2010), the basement of Olkiluoto was affected by different stages of ductile deformation (D₂-D₄, Aaltonen et al., 2010) under a metamorphic retrograde path toward greenschist facies conditions. The progressive change in P–T condition was favourable to the mobilization and subsequent recrystallization of carbonaceous material (e.g. Kirilova et al., 2018). We thus interpret the estimated metamorphic temperature of 530 ^∘C ± 50 ^∘C in the host rock as the temperature of remobilization of carbonaceous material during retrograde metamorphism that culminated in the D₄ deformation stage. During the final stage of D₄ a localized N–S-trending mylonitic foliation developed in the migmatites (Fig. 12a). Mylonitic creep took place under differential stresses increasing from ∼54 to ∼80 MPa, as determined from the recrystallized grain size of 16–18 µm overprinting grains of 25–28 µm. Foliation-parallel quartz veins crystallized transiently along the mylonitic foliation (Fig. 12b) and were overprinted by crystal plastic deformation and dynamic recrystallization under progressively higher stress (80–120 MPa, Figs. 8, 12c) during strain localization towards the centre of the shear zone. We are aware of the uncertainties and limitations of the palaeopiezometric calibrations, and our estimated flow stresses must be taken with care. However, the similarity between subgrain size and grain size, as well as the consistent sinistral asymmetry of the quartz c-axis CPO of relict and recrystallized grains, in all the maps support that dynamic recrystallization occurred during the sinistral strike-slip movement of BFZ045 under progressively increasing differential stresses.

We estimate that mylonitization of BFZ045 (Fig. 12a–c) occurred at ∼450 ^∘C and 3.5 kbar, consistent with the retrograde greenschist facies conditions attributed to the final stages of D₄ deformation in the SW Finland basement at 1.81–1.77 Ga (Lahtinen et al., 2005; Mänttäri et al., 2010; Skyttä and Torvela, 2018). Considering an average crustal density of 2.7 g cm⁻³, the depth of mylonitic deformation and of transient veining in BFZ045 corresponds to approximately 13 km.

The mylonite and quartz veins were eventually overprinted by more pervasive brittle deformation that formed chlorite veins, centimetre-thick cataclasites, and a <0.5 cm thick pseudotachylyte that, together, form the brittle fault core of BFZ045 (Fig. 12d). We attempted to estimate the temperature of chlorite veins cutting the mylonitic foliation and of chlorite in the cataclastic matrix (Fig. 11), but we obtained a wide temperature range of ∼300–500 ^∘C that is unable to constrain the precise T of the final brittle deformation event(s). Although the conditions of formation of cataclasites and pseudotachylytes cannot be tightly constrained, it is interesting to note that the T estimates from the chlorite in the veins are generally higher than 400 ^∘C. This suggests that chlorite veining occurred early in the deformation history of BFZ045 (i.e. at T≥400 ^∘C) under overall ductile conditions, as also supported by the precipitation of chlorite + quartz aggregates in micro-veins and in strain shadows (Fig. 5e). This is consistent with the results by Marchesini et al. (2019), who estimated a temperature of at least 350 ^∘C for the early stages of deformation and emplacement of quartz-chlorite veins along the dextral fault zone BFZ300 (conjugate to BFZ045).

5.3 Conceptual model of the fault slip behaviour of BFZ045 at the base of the seismogenic zone

Using a λ–σ failure mode diagram (Cox, 2010), we propose a conceptual model of the structural evolution of BFZ045 (Fig. 13). Strain rates during mylonitic creep were estimated using the dislocation creep flow laws of quartz by Hirth et al. (2001). Parameters and assumptions used for the calculation of the strain rates and of the λ–σ failure envelopes are listed in Table 3.

According to the failure envelope, differential stresses <80 MPa are necessary for extensional and hybrid fractures to occur. In our model, this stage of fault evolution is represented by the ∼54 MPa creep recorded by quartz grain and subgrain sizes of 25–28 µm (Fig. 7b–d). The failure envelope also shows that extensional failure and vein formation required transient high fluid pressure reaching lithostatic values (λ=1; Fig. 13b, stage 2). Although the overall geometric stress conditions during ductile deformation of BFZ045 are expected to generate en-echelon vein systems oblique to the mylonitic foliation, the foliation-parallel veins are consistent with mode I opening mode due to hydrofracturing. A regional rotation of the stress field appears unlikely, given the constant orientation of the stretching lineation and slickenlines in the core samples and the consistent asymmetry of the pre- and post-vein quartz c-axis CPO (Figs. 7–9). Thus, a transient reorientation of the stress field in the fault zone must be invoked to explain the switch from viscous creep to mode I fracturing along the mylonite foliation.

Veining was then followed by a decrease in fluid pressure (e.g. Sibson, 1989, 1993; Cox, 1995) and a switch back to mylonitic creep under progressively higher stress and strain rate conditions (Fig. 13c, stage 3). The failure mode diagram indicates that these higher stresses are expected to result in brittle shear failure mode of the fault for a pore fluid pressure factor λ>0.8. The cataclasite and pseudotachylytes in the brittle fault core likely represent the final product of BFZ045 deformation under progressively higher differential stresses and fluid pressure across the BDTZ. However, the time span between the high strain rate mylonitic creep and the formation of the brittle fault core is unknown and so are the exact P–T conditions of cataclasites and pseudotachylytes. Thus, our model for stage 3 remains somewhat speculative.

The increase in stress towards the shear zone centre (Fig. 12c) may reflect the rheological evolution of a shear zone that is narrowing with exhumation under progressively decreasing T from the ductile to the brittle crust (e.g. Kohlstedt et al., 1995; Behr and Platt, 2011). Alternatively, it might be attributed to external stress loading from seismic faulting in the overlying upper crust (i.e. seismic loading; e.g. Küster and Stöckhert, 1999; Trepmann and Stöckhert, 2003; Trepmann et al., 2017; Trepmann and Seybold, 2019). The two models are not mutually exclusive and are both equally valid to explain our observations. We are unable to discern between the two options, due to the limitations of our P–T estimates. However, the presence of pseudotachylytes along the BFZ045 fault core indicates that the fault was capable of generating earthquakes, and the seismogenic behaviour of other faults in the Olkiluoto basement has been previously discussed (Marchesini et al., 2019, and references therein). This suggests that transient seismic loading might have triggered the localized increase in creep rate during mylonitization of BFZ045.