Blueschist mylonitic zones accommodating syn-subduction exhumation of deeply buried continental crust: the example of the Rocca Canavese Thrust Sheets Unit (Sesia–Lanzo Zone, Italian Western Alps)

The Rocca Canavese Thrust Sheets Unit (RCTU) is a subduction-related mélange that represents the eastern-most complex of the Sesia–Lanzo Zone (SLZ), bounded by the Periadriatic (Canavese) Lineament that separates the Alpine subduction complex from the Southalpine domain. The RCTU is limited to the south by the Lanzo Massif (LM) and to the east by the Eclogitic Micaschists Complex (EMC). Particularly the tectonic contact area of the RCTU, adjacent to the neighbouring SLZ and the LM is characterised by a 100–200-m-thick mylonitic to ultra-mylonitic zone (MZ) that was active under blueschist-to greenschist-facies conditions. Despite the dominant mylonitic structure, some rocks (garnet-bearing gneiss, garnet-free gneiss and orthogneiss) still preserve pre-mylonitic parageneses in meter-sized domains. The scarcity of superposed structures and the small size of relicts impose a detailed microstructural analysis supported by chemical investigation to reconstruct the tectono-metamorphic history of the MZ. Therefore, we integrated the classical meso- and microstructural analysis approach with a novel quantitative technique based on the Quantitative X-Ray Map Analyzer (Q-XRMA), used to classify rock-forming minerals starting from an array of X-ray elemental maps, both at whole thin section and micro-domain scale, as well as to calibrate the maps for pixel-based chemical analysis and end-member component maps, relevant for a more robust conventional geothermobarometer application as well for calculating reliable PT pseudosections. Pre-Alpine relicts are garnet and white mica porphyroclasts in the garnet-bearing gneiss and biotite and K-feldspar porphyroclasts in garnet-free gneiss and orthogneiss, respectively, providing no PT constraints. The Alpine evolution of the MZ rocks, has been subdivided in three deformation and metamorphic stages. The first Alpine structural and metamorphic equilibration stage (D1 event) occurred at a pressure of ca. 1.25–1.4 GPa and at a temperature of ca. 420–510 °C, i.e. under blueschist-facies conditions. The D2 event, characterised by a mylonitic foliation that is pervasive in the MZ, occurred at ca. 0.95–1.1 GPa and ca. 380–500 °C, i.e. under epidote-blueschist-facies conditions. The D2 PT conditions in the MZ rocks are similar to those predicted for the blocks that constitute the RCTU mélange, and they overlap with the exhumation paths of the EMC and LM units. Therefore, the RCTU, EMC and LM rocks became coupled together during the D2 event. This coupling occurred during the exhumation of the different tectono-metamorphic units belonging to both continental and oceanic lithosphere and under a relatively cold thermal regime, typical for an active oceanic subduction zone, pre-dating Alpine continental collision.


Introduction
The rocks belonging to different complexes of the Austroalpine domain in the Western Alps, originally part of the thinned continental margin of the Adria plate, recorded heterogeneous metamorphic climax conditions and P-T-d-t paths during the Alpine evolution (e.g. Dal Piaz et al. 1972;Compagnoni 1977;Gosso 1977;Dal Piaz et al. 2003;Babist et al. 2006;Dal Piaz 2010;Roda et al. 2012;Lardeaux 2014;Zucali et al. 2020). However, all the inferred deformation and structural histories indicate a common evolution under a depressed thermal regime, which is compatible with active subduction Spalla et al. 1996;Bousquet 2008;Meda et al. 2010;Roda et al. 2010;Spalla et al. 2010;Roda et al. 2012). The Sesia-Lanzo Zone (SLZ) is the largest complex of the Western Austroalpine (Fig. 1a, b) and has been interpreted as a tectonic mixture of mainly continent-derived rocks originally formed in a subduction channel (Pognante 1989a;Zucali et al. 2002;Ferraris and Compagnoni 2003;Spalla and Zulbati 2003;Meda et al. 2010;Roda et al. 2012;Regis et al. 2014;Giuntoli and Engi 2016). The SLZ has been divided into four complexes, which heterogeneously recorded the Alpine evolution ( Fig. 1b): Eclogitic Micaschists (EMC), Gneiss Minuti (GM), II Dioritico-Kinzigitica (IIDK) and Rocca Canavese Thrust Sheets Unit (RCTU). The latter (Fig. 1c) consists of a mixture of meter-to hundreds of metersscale slices of mantle-and continental crustal-derived rocks, enclosed in a serpentinite-rich matrix ( Fig. 1c; Pognante, 1989a;Spalla and Zulbati 2003;Barnes et al. 2014;Cantù et al. 2016;Roda et al. 2018Roda et al. , 2020. The RCTU is considered an example of a subductionrelated mélange formed during the subduction of the Alpine Tethys below the Adria plate (Roda et al. 2018. This interpretation is based on the contrasting tectono-metamorphic evolution recorded by different rock slices enclosed in the serpentinites. In particular, the Alpine metamorphic history of the slices is characterised by an eclogite-to lawsonite-blueschist-facies conditions evolution (Roda et al. 2018), in contrast with the serpentinite-rich matrix that is characterised by a dominant metamorphic assemblage indicating lower pressure conditions. Roda et al. (2018) suggested that the coupling occurred during the exhumation, under a cold thermal regime.
The Periadriatic (Canavese) Lineament is a system of brittle-ductile faults with dextral shear sense (Schmid et al. 1987) and limits the RCTU eastward and separates the Alpine subduction complex from the Southalpine domain that never was involved in the subduction system. The RCTU is limited to the south by the Lanzo Massif (LM). This study focuses on the tectonic contact zone of the RCTU with the neighbouring EMC and the LM units that is mapped as Mylonitic Zone (MZ) (Fig. 1c). This MZ is a 100-200-m-thick mylonitic to ultra-mylonitic zone, dynamically recrystallized under blueschistto greenschist-facies conditions (Spalla and Zulbati 2003;Cantù et al. 2016;Roda et al. 2018). Gneisses and orthogneiss of the MZ preserve pre-mylonitic parageneses in meter-sized domains.
This contribution aims to constrain the tectono-metamorphic evolution of the MZ by selecting those volumes that better preserved early stages of the Alpine evolution. However, the scarcity of superposed structures and the small size of relics impose a detailed microstructural analysis supported by chemical investigations to reconstruct the early tectono-metamorphic imprints that pre-date the Alpine exhumation history (e.g., Spalla and Zucali 2004;Ortolano et al. 2014b;Zucali et al. 2015;Ortolano et al. 2018;Corti et al. 2019). In this work we integrate the classical meso-and microstructural analysis approach (e.g., Williams 1985;Spalla 1993;Lardeaux and Spalla 1990;Passchier et al. 1990;Johnson and Vernon 1995;Zucali et al. 2002;Spalla et al. 2005;Gosso et al. 2010;Hobbs et al. 2010;Salvi et al. 2010;Gosso et al. 2015;Ortolano et al. 2015;Corti et al. 2017) on gneisses (6 samples collected) and orthogneiss (3 samples collected) of the MZ with novel quantitative techniques, to quantify the tectonometamorphic conditions MZ rocks. In particular we applied the Quantitative X-Ray Map Analyzer (Q-XRMA; Ortolano et al. 2018), used to classify rock-forming minerals starting from an array of X-ray elemental maps, both at whole thin section and microdomain scale, as well as to calibrate the maps for pixelbased chemical analysis and end-member component maps, relevant for calculating reliable pseudosections.  Coordinate system WGS 84-UTM 32 N. Redrawn after Cantù et al. 2016 andRoda et al. 2018. d Geological cross-sections (see traces in the geological map). Redrawn after Cantù et al. (2016) and Roda et al. (2018) Finally, we compare the P-T-d evolution of the MZ with that of adjacent SLZ and LM rocks in order to understand when all these tectono-metamorphic units coupled during the Alpine evolution.

Rock types and mesostructures within the MZ and the RCTU
According to Cantù et al. (2016) the MZ is composed of gneisses, micaschists, metagranitoids, silicate-bearing marbles, and glaucophanites ( Fig. 1c). At the map-scale (Fig. 1c), only gneisses and glaucophanites are distinguishable. Metagranitoids, gneisses, micaschists and silicate-marbles occur as meter-to decametre-size pods or lenses and cannot be represented in the map of Fig. 1c. Gneisses (Fig. 2a) and micaschists consist of white mica, quartz, epidote and garnet with minor feldspar, chlorite, blue and green amphiboles. Metagranitoids (Fig. 2a, b) consist of fine-grained clinopyroxene, white mica, quartz, K-feldspar and garnet with minor epidote. Silicate-bearing marbles (Fig. 2c) comprise millimetre to centimetrethick carbonate-rich layers that alternate with chloritoid and white mica layers and contain totally replaced lawsonite porphyroblasts. Glaucophanites (Fig. 2c, d) contain blue amphibole, white mica, and minor epidote, quartz and garnet, and locally pseudomorphosed lawsonite porphyroclasts.
In the RCTU, four groups of structures were detected (Pognante 1989a, b;Spalla and Zulbati 2003;Cantù et al. 2016;Roda et al. 2018). S1 RCTU foliation stands for a micro-scale foliation marked by white mica, amphibole and quartz shape preferred orientation (SPO) and mainly preserved in metapelites. The D2 RCTU event transposed the original lithostratigraphy, and it caused boudinage of metagabbros and glaucophanites enclosed in serpentinites and metapelites, respectively. The S2 RCTU mylonitic foliation is the most pervasive fabric in all rocks of the RCTU. The group of structures associated with D3 RCTU consists of centimetre to meter-scale folds (AP3 RCTU axial plane), associated with the S3 RCTU axial plane foliation, generally oriented vertical and striking NE. Finally, D4 RCTU -related structures consist of a centimetre-spaced crenulation that has been detected only in metapelites.
In the MZ, however, only three groups of structures were detected at outcrop-scale (Cantù et al. 2016). The first group of structures is defined by a mylonitic to ultra-mylonitic foliation (Figs. 1, 3). The orientation of this mylonitic foliation (S2 MZ foliation) strongly clusters around a steep average dip to the NE ranging between 90° and 70° (Fig. 3). The S2 foliations recorded in RCTU and EMC generally exhibit a dip to the ENE with an angle between 45° and 70°, progressively curving when approaching the mylonitic zone (Cantù et al. 2016) where they become parallel to the mylonitic foliation (Figs. 1d, 3;Cantù et al. 2016;Roda et al. 2018). In gneisses, S2 MZ is marked by quartz-rich layers that alternate with white mica and amphibole-rich layers (Fig. 2a, b). Porphyroclasts of clinopyroxene, garnet and locally lawsonite are wrapped by the S2 MZ foliation in glaucophanites, in places exhibiting S-C' structures ( Fig. 2e). In the orthogneisses, D2 MZ structures comprise rootless folds and S-C foliations (S2 MZ ) marked by clinopyroxene trails, white mica films and garnet-rich layers. The shear sense of the mylonitic zone cannot be confidently interpreted because the kinematic indicators are rare and contrasting. Furthermore, at least two successive ductile deformations overprinted this D2 group of structures. Some kinematic indicators suggest a dextral shear sense for the MZ at the contact with the LM, but the change in orientation of the S2 foliations approaching the mylonitic zone suggests a sinistral shear sense for the MZ at the contact with the RCTU and EMC (Figs. 1, 3).
The second group of structures in the MZ (D3 MZ ) consists of tight to isoclinal folds at various scales (AP3 MZ and A3 MZ ), ranging from centimetres to tens of meters. In glaucophanites D3 MZ is associated with the development of the S3 MZ crenulation cleavage or disjunctive foliation (Figs. 2f,3). S3 MZ commonly occurs at a low angle with respect to S2 MZ (Fig. 3) and is marked by the shape preferred orientation (SPO) of green amphibole, white mica, chlorite, plagioclase, epidote and quartz. The S3 MZ orientation is more scattered (Fig. 3), due to the dispersion induced by D4 MZ centimetre-scale crenulation, locally associated with a rough cleavage. The D4 MZ structures in the MZ are generally less pervasive, and only locally associated with a decimetre-scale crenulation (Fig. 3).

Microstructures
At the micro-scale, three rock types belonging to the MZ preserve parageneses and structures that predate the D2 event. Note that from now on the mentioned metamorphic (M-) and deformation (D-) events only refer to the MZ. We analysed garnet-bearing (3 samples) and garnetfree gneisses (5 samples), and a sample of orthogneiss. Within the three rock types, D2 and D3 could clearly be separated on the basis of geometrical coherence and mineral associations (M2, M3; see Table 1). Furthermore, two older parageneses (i.e., pre-1 and 1) were distinguished at the micro-scale on the basis of their mineral associations and microstructural setting ( Table 1). The pre-1 paragenesis is characterized by few relicts only and it cannot be associated with a clear metamorphic (M-) or deformation (D-) event. Therefore, we here use the generic index pre-1 to indicate that this paragenesis formed before the D1 and M1 event. On the other hand, the term "paragenesis 1" has a well-defined structural and mineralogical relationship with the subsequent events (D2 and D3). Therefore this term is associated with a single metamorphic and deformation event of Alpine age (M1, D1).
Minerals are numbered according to their first occurrence. Minerals stable through successive stages keep the number of their first occurrence while minerals always stable (e.g. Qz) have no reference number. Mineral abbreviations follow Whitney and Evans (2010), except for white mica, which is abbreviated as Wm.
The third stage (III, M2 event) is associated with the development of the S2 foliation and is characterised by Wm3, Grt2b, Gln1, Ep2, Qz, Pl1 and rare Ttn2. The S2 films are marked by the SPO of sub-millimetre-scale Wm3, Grt3, Gln1, Ep2 and Ttn2, while the microlithons mainly consist of Qz (Table 1). Gln1 is also enclosed

Fig. 3
Schmidt stereographic projections, lower hemisphere, of the poles of the S2 and S3 foliations, axial planes (AP3, black dots) and axes (A3, red dots) of D3 folding, and D4 decimetre-scale crenulation for the MZ, the RCTU, the EMC and the LM. Density contours are also reported for S2 and S3 foliations. In the MZ, the orientation of poles to S2 and S3 foliations strongly clustered around SW with a dip angle between 0 and 25 within Grt1 boudin necks (Fig. 4g, h), with fibres parallel to S2 films. M2 is also associated with symplectitic exsolutions of Pl1 in Jd1 and the replacement of Ep1 by Ep2. Finally, M3 consists of exsolutions of Cpx1 along Jd1 fractures.
The last stage (IV, M3 event) is characterised by Cpx2, Pl1 and Qz. Cpx2 marks the boundary between the crystals in the Jd + Qz aggregates (Fig. 5h), and exsolutions of Pl1 can be observed within fractures of Jd crystals ( Fig. 5f ).
The third stage (III, M2 event) is associated with the S2 foliation, which is marked by the SPO and the lattice preferred orientation (LPO) of Wm2 crystals, Jd2, Qz and rare Ttn2 (Fig. 6a, b, e, f and Table 1). Jd2 also occurs as coronas of Jd1( Fig. 6b-d, f ). The last stage (IV, M3 event) is represented by exsolutions of Cpx1 along Jd fractures.

Stage Event Structures Mineral association
Grt-bearing gneiss Grt-free gneiss

Methodology
Microstructural analysis aimed at detecting microstructural sites is well suited for mineralogical-chemical investigations. Compositional variations of mineral phases of three analysed samples ( Fig. 1c) are detailed in Tables 2,  3 and 4 and illustrated in diagrams showing the compositional mineral trends, also significant for inferring the physical conditions of metamorphism. Representative mineral analyses from gneisses and orthogneisses are shown in Tables 5, 6 and 7, respectively. Mineral compositions have been determined using a Jeol, JXA-8200 electron microprobe (WDS, accelerating voltage of 15 kV, beam current of 15 nA), operating at the Dipartimento di Scienze della Terra "A. Desio" of Università degli Studi di Milano. Natural silicates have been used as standards and the results were processed for matrix effects using a conventional ZAF procedure (Armstrong and Buseck 1975). Mineral formulae are calculated on the basis of 12 oxygens for garnet, 6 for pyroxene, 23 for amphibole, 11 for mica and biotite, 12.5 for epidote, 8 for feldspar and 20 for titanite. Fe 3+ was recalculated based on the charge balance (Droop 1987). The classification of amphiboles follows IMA 2012 recommendations (Hawthorne et al. 2012;Oberti et al. 2012) and we used a spreadsheet proposed by Locock (2014). The classification of pyroxenes is after Morimoto (1988). The analysis of the metamorphic assemblages related to the different fabrics was performed by using Quantitative X-Ray Map Analyzer (Q-XRMA; Ortolano et al. 2018). Q-XRMA has been used to classify rock-forming minerals starting from an array of X-ray elemental maps, both at thin section and microstructural domain scale, as well as to calibrate the maps for pixel-based chemical analysis and end-member component maps, by using spot chemical analyses as internal standards (De Andrade et al. 2006). The Q-XRMA image processing tool, based on several image analysis functions implemented within ArcGIS software, has been progressively used in different geosciences-related issues (e.g., Li et al. 2010;DeVasto et al. 2012;Pradhan 2013;Ortolano et al. 2014a, b;Belfiore et al. 2016;Fiannacca et al. 2017;Berrezueta et al. 2019;Corti et al. 2019Corti et al. , 2020Ortolano et al. 2020a, b).
The X-ray maps of three microdomains from garnetbearing gneiss (ZR4), garnet-free gneiss (CR115 samples), and orthogneiss (ZR19 sample) have been explored. The ZR4 and ZR19 microdomains were acquired with a dwell time of 160 ms and resolution of 500 × 500 with pixel size equalled to 2 µm. The CR115 microdomain were acquired with a dwell time of 40 ms, image resolution of 870 × 870, and pixel size of 0.50 µm.
The Q-XRMA procedure is divided into three different cycles: (i) the first cycle is useful to classify mineral phases at thin section and microdomain scale as well as to infer the associated modal percentage by a multivariate statistical data handling of the X-ray maps through the Principal Components Analysis (PCA) and the supervised Maximum Likelihood Classification (MLC; Ortolano et al. 2014b); (ii) the second cycle performs a deeper analysis of selected mineral phases for detecting mineral zoning and to calibrate X-ray maps thanks to a series of image analysis functions based on a multiple linear regression analysis. As a result, the element concentration values can be calculated for each pixel related to each investigated mineral phase; (iii) the third cycle allows to manage the calibrated X-ray maps to obtain the maps of end-member components, as well as to quantify the chemical variations within each mineral phase. For each calculated pixel, the stoichiometry is in agreement with the structural formula of corresponding mineral species. This method allows defining a refined analysis of the metamorphic assemblages that characterize the different fabrics.

Garnet-bearing gneiss
The first generation of garnet (Grt1a, Fig. 7a, b) displays an Alm composition, with 10 to 20% of Py (Fig. 8a, b). The Mn content ranges between 0.1 and 0.3 a.p.f.u.  (Tables 2 and 5), resulting in a Sps content < 10%. Adr and Grs contents are negligible. Grt1b that fills microfractures within Grt1 (Fig. 7a, b) has a composition similar to that of Grt1a, but with a lower Py content (< 10%, Fig. 8a, b). During the M1/D1 event a new generation of garnet growths (Grt2) and occurs in the core of garnets that are disposed along the S2 foliation, and as inner corona of Grt1a and Grt1b (Fig. 7a, b). Grt2 has a Ca content that ranges from 1 to 1.3 a.p.f.u., and the Fe 2+ content ranges from 1.3 and 1.7 a.p.f.u. (Tables 2 and  5). This results in an Alm content of ca. 50% and a Grs content of ca. 40% (Fig. 8a, b), with negligible contents of Py, Sps and Adr (Table 2). During M2/D2 a new garnet growth occurs (Grt3) at Grt1a and Grt1b rims and as coronas of Grt2 (Fig. 7a, b). The Ca content increases in Grt3 with respect to Grt2, and ranges from 1.2 to 1.5 a.p.f.u. This results in a similar content of Grs and Alm of about 40-50% ( Fig. 8a, b, and Tables 2 and 5). Py, Sps and Adr contents remain negligible (Tables 2 and 5). Wm1 has a Ms composition with a Si 4+ content of ca. 3 a.p.f.u., and the Ti content is below the detection limit ( Fig. 8c, d, and Tables 2 and 5). Wm2 has a phengite composition (Fig. 8c, d), with the Si 4+ content from 3.4 to 3.6 a.p.f.u. and the Ti content below the detection limit (Fig. 8c, d and Table 2). The same composition of Wm2 is shown by the rim of Wm1 (Fig. 8c, d). Wm3 still has a phengite composition although with a generally lower Si 4+ content with respect to Wm2 (Fig. 8c, d, and Tables 2 and 5).
Jd1 occurs during M1/D1 and has a X Jd content ≥ 80% (Figs. 7c and 8e, f ), and the Na content ranges from 0.85 to 1 a.p.f.u. (Figure 8e, f, and Tables 2 and 5). During M2/D2 symplectites of Pl1 in Jd1 occur with albite composition (Tables 2 and 5). Small exsolutions of Aeg in Jd1 occur during M3/D3 and show an increase in the Fe 3+ and Ca contents with respect to the Jd1, with maximum Q and Aeg components of 40 and 70%, respectively (Figs. 7c, d and 8e, f ). Ep1 and Ep2 have similar Fe 3+ and Al VI content and negligible amount of Mn (Tables 2 and 5).

Garnet-free gneiss
Wm1 occurs during M1/D1 and has a phengite composition (Fig. 9a, b) with the Si 4+ content that ranges from 3.45 to 3.5 a.p.f.u.; the Ti content is below the detection limit (Fig. 9a, b, and Tables 3 and 6). Wm2 still has a phengite composition with a lower Si 4+ content with respect to Wm1 (Fig. 9a, b, and Tables 3 and 6).
During M1/D1 two different generations of Cpx can be compositionally distinguished (Fig. 9a, b). The first one (Jd1) has a X Jd content of 80% (Fig. 9c, d), the Na content varies from 0.93 to 1 a.p.f.u., and the Mg + Ca + Fe content is between 0 and 0.16 (Fig. 9c, d, and Tables 3 and  6). The second one (Cpx1) has an omphacite composition with a Na content of ca. 0.5 and a Mg + Ca + Fe content of ca. 0.8 (Fig. 9c, d, and Tables 3 and 6). Jd2, that usually occurs at the rim of Jd1 (Fig. 10a,b) has a lower X Jd content with respect to Jd1 (i.e., between 65 and 80%), and the Na content ranges from 0.8 to 1 a.p.f.u. (Figure 9c, d, and Tables 3 and 6). During M3/D3 exsolutions of Aeg (Cpx2) occur along Jd grain boundaries (Fig. 10a, b), and they show an increase in the Fe 3+ and Ca contents with respect to the Jd1 and Jd2, with maximum Q and Aeg components of 40 and 80%, respectively (Fig. 9c, d, and Tables 3 and 6). Exsolutions of Pl1 in Jd1 and Jd2 are albite in composition (XAb 95%).
The Fe 3+ content in Ep1 varies from 0.6 to 0.8 a.p.f.u., and the Al VI content ranges from 2.2 to 2.4 a.p.f.u. (Tables 3 and 6). Lower Fe 3+ content characterises Ep2 composition (from 0.5 to 0.6 a.p.f.u., Tables 3 and 6). The Mn content is very low for both of the Ep generations (Tables 3 and 6).

Orthogneiss
Grt1 is mainly Grs in composition, with a maximum of 10-20% of Adr content (Fig. 11a, b, and Tables 4 and 7). The Ca content ranges from 2.75 to 2.85 a.p.f.u., the Fe 3+ content from 0.2 and 0.38 a.p.f.u., the Mn content from 0.1 to 0.2 a.p.f.u. and the Mg and Fe 2+ contents are below the detection limit (Tables 4 and 7). Wm1 has a phengite composition (Fig. 11c, d) with the Si 4+ content that ranges from 3.5 to 3.65 a.p.f.u.; the Ti content below the detection limit (Tables 4 and 7). Wm2 still has a phengite composition, although with a lower Si 4+ content with respect to Wm1 (from 3.35 to 3.45 a.p.f.u., Fig. 11c, d, and Tables 4 and 7).
Jd1 has a X Jd content ≥ 85% (Figs. 11e, f and 12), and the Na content ranges from 0.92 to 1 a.p.f.u. (Fig. 11e, f, and Tables 4 and 7). The maximum Fe + Ca + Mg content is 0.15 a.p.f.u. Jd2 has a lower XJ d amount with respect to Jd1 (between 50 and 85%, Fig. 11e, f ), and the Na ranges from 0.84 to 1 a.p.f.u. (Fig. 11e, f, 12, and Tables 4 and 7). As in the gneisses, exsolutions of Cpx1 in Jd occurs (Fig. 12) and show an increase in the Fe 3+ content with respect to Jd1 and Jd2, with an Aeg component of 40% (Figs. 11c, f and 12).
6 Pressure and temperature conditions of metamorphism

Methodology
Micro-structural analysis indicates that the investigated rocks preserve evidence of superposed structural and metamorphic re-equilibrations (even in volumes as small as that of a thin section) and this allows for the definition of paragenetic sequences. It also permits the identification of favourable sites, where microstructures suggest attainment of grain-scale equilibrium for inferring PT conditions during each re-equilibration stage. Physical conditions of successive re-equilibration stages have been inferred by the application of quantitative image-assisted well calibrated independent thermobarometers and compared and refined by the calculation of quantitative image-assisted thermodynamic modelling, which permitted to calculate PT pseudosections for specific effective bulk rock chemistries (Stüwe and Ehlers 1996;Evans 2004;Zuluaga 2005;Zeh 2006;Ortolano et al. 2014b). We estimated PT conditions on garnet bearing gneiss (ZR4) and orthogneiss (ZR19) since the garnet-free gneiss (CR115) lacks adequate mineral associations to be applied for geothermometry and for obtaining useful data from the pseudosection. However, garnet bearing gneiss and orthogneiss sufficiently cover a broad origin of rocks, from metapelites to metaintrusives. Calibrated image on garnet bearing gneiss and orthogneiss permitted to apply geobarometers based on the X Jd content in Cpx (Holland, 1980) and Si 4+ content in Wm (Massonne and Schreyer 1987). Temperature estimates have been inferred from two independent calibrations of Grt-Wm pairs (Green and Hellman 1982;Wu et al. 2002) for garnet-bearing gneiss and orthogneiss (Table 8).
Conventional geothermobarometers were then refined by phase diagram computation (i.e., Perplex or Thermocalc; Connolly and Petrini 2002;Connolly 2005), here stepwise reinforced by the calculation of quantitative effective bulk rock chemistries supported by image analysis of X-ray maps (Stüwe and Ehlers 1996;Evans 2004;Zuluaga 2005;Zeh 2006;Ortolano et al. 2014b). In the above-mentioned view, the thermodynamic modelling of metamorphic rocks can be considered the most useful tool to quantify the PTX evolution of orogenic processes. The robustness of the modelled systems is based on the correct definition of the real reacting volume of each recognized metamorphic equilibrium (i.e., the effective reactant volume, ERV, Ortolano et al. 2020a, b), which depends on the objective interpretation of textural equilibria. The choice of the most simplified chemical system capable of describing all observed parageneses, the most appropriate solid solution models and the most meaningful effective bulk rock chemistry is at the base of the pseudosection computation method. This method can be susceptible to different sources of uncertainty which can be derived, in general, from the propagation of the uncertainty in the thermodynamic data used for the compilation of the database or in the solid solution models parameters, while, in the specific application of the Perplex software, also from the effects of the adopted pseudocompound approximation strategy of calculation (Connolly 2005). Another potential source of uncertainty is the influence of chemical fractionation induced by mineral phases, that, reaching a specific critical size, are potentially able to subtract a specific percentage of constituents oxides from the effective bulk-rock chemistry (e.g. Angì et al. 2010;Ortolano et al. 2014aOrtolano et al. , 2020b. A useful technique for obtaining more reliable PT constrains from the pseudosection calculation method involves the numerical extraction of petrological data from the image analysis of an array of X-ray maps of entire thin sections or microdomains (e.g., Grosch et al. 2014;Ortolano et al. 2014b;Fazio et al. 2018;Lanari et al. 2019;Ortolano et al. 2020a, b). In this work, we used the outputs obtained by the Q-XRMA to identify the proper ERVs and the associated effective bulk compositions (EBCs) to be used in the pseudosection calculation for the garnet-bearing gneiss and orthogneiss samples. Perplex software package has been chosen as the phase diagram computing system (Connolly 1990;Connolly and Petrini 2002;Connolly 2005). The computed pseudosection span a PT range of T = 250-650 °C and P = 0.4-1.5 GPa, considered suitable to enclose the PT constraints of the observed parageneses. The Na 2 O-TiO 2 -MnO-K 2 O-MgO-FeO-Fe 2 O 3 -CaO-Al 2 O 3 -SiO 2 -H 2 O system was used to model the garnet-bearing gneiss and the orthogneiss whereas the MnO component was excluded from the system in the case of the garnet-free gneiss. SiO 2 and H 2 On have been considered in excess, and the CORK fluid equation of state (EOS) of  was chosen to model the fluid phase behaviour. The adopted solid solution models are: (a) Ca-Fe 2+ -Mg-Al-Fe 3+ garnet model  Holland and Powell (1996) clinopyroxene model; (e) the amphibole model of Dale et al. (2000) with no Ca, K, Ti or Mn solution; (f ) a binary plagioclase solid solution Newton et al. (1980); (g) a binary feldspar model (Waldbaum and Thompson, 1968); (h) Fe 2+ -Mg-Mn stilpnomelane from Massonne (2008); (i) Ti-Fe-Mg-Mn biotite from Powell and Holland (1999) extended to cover Fe and Mn-solution; (j) the magnesioferrite/magnetite model; (k) the ilmenite model after White et al. (2000) was selected in the case of the orthogneiss sample. Ideal ilmenite-geikielite-pyrophanite solution was chosen for the garnet-bearing gneiss; and finally, (l) the chlorite model extended from  was applied only for the garnet-bearing gneiss pseudosection. In order to simplify the pseudosection computation, paragonite, rutile, essenite, Ca-tschermaks pyroxene together with the dependent endmembers of the chlorite solution model were excluded from the calculation. All obtained pseudosections have quartz as the stable phase in each field.

Table 3 Intervals of compositional variations in minerals marking different metamorphic stages of garnet-free gneiss
The PT constraints were obtained by the use of the isopleths intersections approach (e.g., Evans 2004;Cirrincione et al. 2008;Angì et al. 2010), taking care to use at least three independent chemical isopleth intersections for defining reliable constraints and considering an average estimated error of ± 30 °C and ± 0.1 GPa (e.g., Cirrincione et al. 2008). The procedure has been carried out stepwise, verifying the effective reactant volume operating during the different evolutionary stages of the metamorphic history.

Garnet-bearing gneiss
Numerical elaboration of major elements X-ray maps of the entire thin section (350 × 250 pixels), supported by an opportunely selected garnet micro-domain (500 × 500 pixels), has been adopted to extrapolate modal percentages of the representative rock-forming minerals (Table 9) as well as to calibrate the mineral compositions of the micro-domain pixel constituents (Figs. 7,8). The obtained mineral compositions observable in the shaded areas of Fig. 8 were used in turn for conventional geothermobarometer applications as well as to define the EBC involved during the various steps of the tectono-metamorphic evolution. Such an EBC was calculated following the method of Ortolano et al. (2014a) by considering Grt1a as no longer in equilibrium with the surrounding minerals and normalising the Grt1b and Grt2 modal amounts derived by the microdomain image analysis, to the total of the entire thin section (Table 10). Ascertained the impossibility to constrain the pre-M1 metamorphic event, conventional geothermobarometry, applied to calibrated mineral composition maps, permitted to obtain a minimum P range from 1.24 to 1.36 GPa intersecting X Jd content in Cpx (Holland 1980) and an Si 4+ content in Wm (Massonne and Schreyer 1987) for the syn-D1 assemblage. Temperature estimates have been then inferred using the two independent calibrations of Grt-Wm pairs (Green and Hellman 1982;Wu et al. 2002) obtaining T ranging from 430 to 520 °C (Fig. 13a).
These PT conditions well integrate garnet isopleths thermobarometry on PT pseudosection calculated for the first obtained EBC, allowing obtaining PT estimates for D1 event at 520 °C and at slight P conditions of 1.28 GPa (Fig. 13b). The obtained syn-to late-D1 metamorphic events match trivariant and quadrivariant pseudosection PT space characterised by parageneses of Wm + Grt + Cpx (with high Jd content) +Amp + Ttn + Ab (Fig. 13b). This field well fit the observed parageneses consisting in Wm +Grt + Jd + Ttn (Table 1) unless for the occurrence of computed Amp although characterised by a very limited volume percentage (ranging from 8 to 3 vol %).
The observed syn-D2 mineral association of the D2 event accounts for the disappearance of Jd and the occurrence of Gln and Pl (Table 1). Garnet isopleths thermobarometry, constrained via compositional range derived from calibrated X-Ray maps, permitted to constrain the average P for the D2 event at 1.04 GPa, with a T of 490 °C (Fig. 13b). The specific computed pseudosection PT field highlights a Wm + Grt + Jd + Gln + Bt + Ab assemblage, differing from the observed parageneses in view of the occurrence of biotite instead of the observed epidote. Conventional geothermobarometry derived from over 200 white-mica garnet pairs of neighbouring calibrated pixels and ca. 250 white-mica phengite content geobarometer, permitted to constrain a minimum P of 1.0 ± 0.1  GPa for a T of 460 ± 30 °C linked with the D2 event (Fig. 13a). As for the pre-M1 mineral associations, the syn-D3 mineral assemblage does not permit any PT constraints for this deformational event.

Orthogneiss
Mineral modal amounts constrained by image analysis of an opportunely selected microdomain were obtained after classification and calibration of ca. 250,000 pixels, representative of several minerals (Figs. 11, 12). The quantitative image analysis of this selected microdomain permitted also to define the ERV useful to calculate, in turn, the proper EBC for this sample (Table 11). The intersection PT field obtained via conventional geothermobarometers applied on the calibrated pixel analyses, permitted to derive minimum P estimates for the D1 event at ca. P = 1.3 ± 0.1 GPa (Holland, 1980;Massonne and Schreyer, 1987) for T of 495 ± 25 °C (Green and Hellman 1982;Wu et al. 2002; Table 8 and Fig. 14a). Furthermore, conventional geothermobarometers permitted to also constrain the D2 event for this sample, intersecting all the PT fields deriving from the Green and Hellman (1982) and Wu et al. (2002) thermometers and the Holland (1980) and Massonne and Schreyer (1987) barometers, obtaining minimum P ranging from 0.94 to 1.06 GPa at T of 405 ± 25 °C (Table 8 and Fig. 14a).
Finally, the computed PT pseudosection calculated trough the obtained EBC permitted to define PT-estimates for the syn-D2 event lie in the compositional field Wm + Grt + Cpx + Ttn + Ab + Mic (Fig. 14b). Grt1 and Kfs1 do not show any rim or exsolution evidence, therefore they can be considered to still be in equilibrium with the other phases during D2. The content of Ab in the pseudosection is estimated of about 20 vol %. However, we do not detect Pl in these rocks. We estimate the PT conditions of ca. P = 1.07 GPa at an average T of 400 °C for D2 (Fig. 16). Again, the mineral association of D3 does not permit any PT estimate for this stage.

Discussion and conclusions
The integrated application of meso and microstructural and petrologic investigations, assisted by the use of the Quantitative X-Ray Map Analyzer (Q-XMRA) allows the definition of detailed deformation and metamorphic stages of the rocks of the Mylonitic Zone (MZ), which characterise the tectonic contact between the RCTU and the adjacent complexes of SLZ and the LM. Q-XMRA indeed represents, a valid tool to extrapolate reliable modal percentages of the representative rock-forming minerals from X-ray maps, as well as to calibrate the main mineral rock constituents. This approach permitted to use the composition of the calibrated X-ray maps to meaningfully apply conventional geothermobarometers as well as for the more suitable definition of EBCs for pseudosection computation, that, in turn, permitted to project mineral isopleths useful to obtain further PT conditions (Fig. 15). The first deformation stage (I) in Grt-bearing gneiss (Table 1) Spalla et al. (1983Spalla et al. ( , 1991; IIDK- Lardeaux and Spalla (1991)) and related to the pre-Alpine granulite-facies metamorphism Rebay and Spalla, 2001;Spalla et al. 2014;Roda et al. 2019). On the contrary, this kind of pre-1 metamorphic assemblage has never been observed      (Table 8) estimated from quantitative image-assisted calibrated independent thermometers and barometers for garnet-bearing gneiss. b Pseudosection calculated in the Na 2 O-TiO 2 -MnO-K 2 O-MgO-FeO-Fe 2 O 3 -CaO-Al 2 O 3 -SiO 2 -H 2 O system for Grt-bearing gneiss and PT conditions of D1 (blue) and D2 (green) events (Table 8) estimated from the intersections of garnet end-members isopleths (i.e., almandine-black; grossular-white; pyrope-blue; andradite-green). The assemblage lists are reported in the Additional file 1: SM1 in the RCTU. Although this mineral association can suggest a HT metamorphism for the pre-1 stage, geochronological data is needed to infer a pre-Alpine age. The scarcity of pre-1 relicts in garnet-free gneiss and orthogneiss does not allow for an interpretation about their pre-1 history. However, the abundance of Jd in the Alpine assemblages suggests the occurrence of Pl during the pre-1 event. Therefore, we can speculate for a possible granite protolith for orthogneiss and an aplite composition for the protolith of Grt-free gneiss. The Alpine evolution is characterised by three deformation and metamorphic events (II, III and IV, Table 1), and it is similar for all the studied samples. The integration between conventional geothermobarometry and isopleths on pseudosections allows for the estimation of reliable PT conditions. Although the pressures estimated with conventional geobarometry have to be considered minimum pressures given the barometers used (Holland 1980;Massonne and Schreyer 1987), the intersection of three independent chemical isopleths confirms those values. Therefore, the first Alpine structural and metamorphic equilibration event (II, D1 stage) occurred at a pressure of ca. 1.25-1.4 GPa and a temperature of ca. 420-510 °C (Fig. 15), i.e. under blueschist-facies condition. D1 is followed by a T and P decrease during D2, down to ca. 0.95-1.1 GPa and 380-500 °C (Fig. 15), i.e. to epidote-blueschist-facies condition. The metamorphic assemblage of D3 does not allow for the definition of PT conditions, although the occurrence of Aeg exsolutions within Jd may suggest lower pressure conditions than D2 (Popp and Gilbert 1972).
The D1 peak conditions recorded by the MZ rocks ( Fig. 15) are more similar to those of the RCTU (1a, Roda et al. 2018Roda et al. , 2020 than those of EMC (Pognante 1989b;Rebay and Messiga 2007). In any case, the peak conditions of the MZ rocks are different from those of the RCTU D1b conditions, where lawsonite occurred in the mineral assemblage (Roda et al. 2018). The D2 stage, characterised by the S2 MZ mylonitic foliation pervasive overall the MZ, occurred under PT conditions of ca. 0.95-1.1 GPa and at ca. 400-500 °C. The S2 foliations recorded in RCTU and EMC progressively curve approaching the mylonitic zone (Cantù et al. 2016) and become parallel to the S2 MZ mylonitic foliation. All the S2 foliations are marked by Gln, Wm and Ep and occurred under blueschist facies conditions (Fig. 15). Therefore, we can conclude that the D2 fabric detectable in the MZ is conceivably contemporaneous with the D2 event in the RCTU and EMC, when the different sheets composing the RCTU and the EMC coupled together. For what  concerns the S2 foliations recorded in the MZ and in the serpentinites of the LM, the scarcity of structural data in this portion of the LM and the poorly diagnostic metamorphic assemblages of the S2 LM , marked by serpentine, make the correlation between the two foliations less reliable. However, more westward the contact between the LM and the Sesia-Lanzo Zone is characterized by a folding between gneiss (SLZ) and serpentinites (LM), contemporaneous with the development of an axial-plane foliation marked by Gln + Grt + Rt and occurred under blueschist-facies conditions (Blake et al. 1980;Kienast and Pognante 1988 (Rebay and Messiga 2007) and Orco (Pognante 1989b); LM-Lanzo Massif (Pelletier and Müntener 2006). The age of metamorphic peak conditions of the LM are after Rubatto et al. (2008). Metamorphic facies after Ernst and Liou (2008): GS: greenschist; Ep-A: epidote-bearing amphibolite; BS: blueschist; AM: amphibolite; EC: eclogite; HGR: high pressure granulite; GR: granulite. Geotherms: 3a) warm subduction zones, 3b) cold subduction zones (Cloos 1993); Vi: stable geotherms (England and Thompson 1984) for continental crust LM rocks coupled together during D2 event, by ductile shear zones developed under epidote-blueschist-facies conditions, and later evolved as a single tectono-metamorphic unit. Figure 16 displays a cartoon of the principal stages of the evolution. The North Lanzo Massif is considered as representing a portion of mantle deriving from an oceancontinent transition zone (OCT) formed during Triassic-Jurassic rifting (Müntener et al. 2005;Piccardo 2010;Guarnieri et al. 2012;McCarthy and Müntener 2015) while the EMC represents continental crust of the Adria plate (Fig. 16a). According to our interpretation both the Lanzo Massif and the EMC became subducted down to different depths within one and the same subduction channel (Fig. 16b) during a long-lasting subduction episode lasting from 67 to 35 Ma and taking place under a cold thermal regime (Fig. 15). This is consistent with oceanic subduction of the Alpine Tethys during Alpine convergence (see e.g., Gerya and Stöckhert 2005;Meda et al. 2010;Roda et al. 2012;Giuntoli et al. 2018;Roda et al. 2020). This scenario differs from that proposed by of Babist et al. (2006) who proposed a multi-stage subduction and exhumation scenario, compatible with the evolution of the Alps at a larger scale (see Handy et al. 2010). In this alternative scenario a first event occurred when the western corner of upper Adriatic plate, represented by the Sesia Zone, became tectonically eroded and subducted to 60-65 km depth already by 75-65 Ma ( Fig. 13b; Duchêne et al. 1997;Rubatto et al. 1999;Konrad-Schmolke et al. 2006) and then exhumed to 10-15 km depth within the nascent Western Alpine accretionary wedge by 63 Ma. This was closely followed by a second subduction event that involved oceanic subduction in this part of Alpine Tethys and that took place in Paleocene-Eocene time.
However, ages of 70-80 Ma have been proposed for the prograde path of ZSZ during subduction on the basis of the garnet growth modelling (Skora et al. 2009;Assanelli et al. 2020) and older ages than the usual age range reported for peak high-pressure metamorphism in the Piemonte Liguria ophiolites of the Western Alps (ca. 59-45 Ma; Berger and Bousquet 2008;Villa et al. 2014) were recently reported for the Zermatt-Saas Zone: Rebay et al. (2018) document U-Pb ages of around 65 Ma for zircon fringes in serpentinites. Conversely, Vho et al. (2020) refer to ages younger than usual for metamorphism in the Sesia Zone near our working area (63-52 Ma, obtained from Rb-Sr dating of phengite published by Inger et al. 1996). Based on these age data we propose a continuous long lasting oceanic subduction   (Roda et al. 2018. The metamorphic peak of the LM is dated at 55-46 Ma (Rubatto et al. 2008) while no geochronological data are available for the EMC in the SLZ in our working area. The coupling between EMC, RCTU and LM occurred via the formation of shear zones, such as the MZ, during the exhumation between 40 and 35 Ma (Fig. 16c) and still under a cold thermal regime (Fig. 15). From 40 Ma onward EMC, LM and RCTU evolved as a single tectono-metamorphic unit until the continental collision of the European plate at 30 Ma (Fig. 16d).

Fig. 16
Conceptual representation of the evolution of Rocca Canavese Thrust Sheets Unit (RCTU), Eclogitic Micaschists Complex (EMC) and Lanzo Massif (LM) within the tectonic frame of the South-Western Alps. a Hyperextended passive margin of Adria Plate before the Alpine subduction. The North Lanzo Massif is considered as a portion of mantle deriving from an ocean-continent transition zone (OCT) formed during Triassic-Jurassic rifting (Müntener et al. 2005;Piccardo 2010;Guarnieri et al. 2012;McCarthy and Müntener 2015), the South Lanzo Massif is considered a portion of the oceanic lithosphere of the Alpine Tethys (Piccardo 2010) and the EMC represents the continental crust of the Adria plate. b During subduction of the Alpine Tethys below the Adria plate, EMC and LM are subducted within the subduction channel, at different depths, and under a cold thermal regime. The RCTU is interpreted as a mélange formed in the subduction channel (Roda et al. 2018;Luoni et al. 2020;Roda et al. 2020). The LM metamorphic peak is dated at 55-46 Ma (Rubatto et al. 2008) while no chronological data is available for the EMC in this part of the SLZ. c The coupling between EMC, RCTU and LM occurred along the shear zones of the MZ, during the exhumation within the subduction channel, still under a cold thermal regime and formed a single tectono-metamorphic unit. d During the collision between Adria and European plates, LM, EMC and RCTU are involved in the orogenic wedge (i.e., a tectonic mixture of oceanic and continental rocks buried at different depths and exhumed to crustal levels) up to reach the present-day structural configuration ▸