The Valpelline Series (Dent-Blanche Tectonic System, Western Italian Alps) is a sector of lower continental crust, which consists of Permian migmatitic metapelite with different mineral assemblages (i.e., garnet-, cordierite- and orthopyroxene-bearing), minor amphibolite and marble, intruded by aplite and pegmatite. Widespread melt production in metapelite and locally in amphibolite occurred during the development of the regional foliation. The PT conditions during migmatisation, estimated using conventional geothermobarometers, range between 800–900 °C and 0.5–0.8 GPa, with a difference of up to ∼50 °C between cordierite- and orthopyroxene-bearing migmatites, the latter reaching higher temperatures. The Valpelline Series shows rock types, metamorphic assemblages, PT conditions and published ages of high-temperature regional metamorphism like the archetypal lower crust section of the Ivrea-Verbano Zone in the Southern Western Alps. The latter likely represents an external portion of the same extending lower crust, at the onset of the Tethyan rifting due to lithospheric extension and asthenospheric rising.

Investigating the tectono-metamorphic evolution of exposed amphibolite- to granulite-facies rocks is crucial for understanding the formation and evolution of the middle- to lower continental crust (Bohlen, 1987; Hacker et al.2015; Touret et al.2022) within different geodynamic settings, i.e., from collisional-related crustal thickening to post-collisional or back-arc lithospheric thinning (Jiao et al.2023 and references therein). In these deeper crustal sectors, melt production and migration contribute to the growth and differentiation of the continental crust, shaping its composition and structure and influencing the deformation regime (Hawkesworth et al.2020). Comprehending the litho-structural features of the lower continental crust is an important first step for the implementation of PT–D–t paths and geochemical surveys to unveil crustal differentiation processes and provide insights into their tectonic framework.

In the Western Alps, there are pre-alpine lower crustal slices such as the II Zona Diorito-Kinzigitica (IIDK; Carraro et al.1970), the Ivrea-Verbano Zone (IVZ; Quick et al.2003; Redler et al.2012, 2013; Ewing et al.2013, 2015) and the Valpelline Series (VP; Gardien et al.1994; Manzotti & Zucali, 2013; Caso, 2023), which have been affected by widespread high-temperature (hereafter HT) metamorphism and are now fragmented and found in different portions of the Western Alpine belt. The HT evolution is related to the post-Variscan collisional collapse and/or Permian lithospheric extension (Lardeaux & Spalla, 1991; Marotta & Spalla, 2007; Schuster & Stüwe, 2008; Roda et al.2018; Roda et al.2023). In particular, the VP in the Dent-Blanche Tectonic System (DBTS; Manzotti et al. 2014; Manzotti & Ballèvre, 2017), located in the Valle d’Aosta (Western Italian Alps), represents a well-preserved exposure of a lower continental crustal section that has largely escaped the Alpine-age metamorphic overprint. The VP is traditionally compared to the IVZ and the IIDK, all of which are part of the Adria paleomargin. This comparison is based on lithological affinities and similar dominant HT pre-alpine tectono-metamorphic imprint. In the past decades, several works have dealt with the HT evolution of the VP (Diehl et al.1952; Gardien et al.1994; Manzotti & Zucali, 2013) but they lack a detailed description and distinction of the different rock types and structures or they are localised to restricted areas of the unit (Diehl et al.1952; Nicot, 1977; Pennacchioni & Guermani, 1993; Gardien et al.1994; Manzotti & Zucali, 2013). Therefore, this contribution provides a detailed meso- and microscale description of VP rocks and their relations with the deformation history. Preliminary PT estimations related to HT metamorphism and melt-production stages have been obtained both from metapelite and metabasite. A first characterisation of the thermal structure of the VP is provided, discussing the implications for the thermal conditions of the extended Permian lithosphere. Ultimately, a preliminary comparison between the VP and the IVZ crustal section is given: it compares lithologies, structures, thermal conditions and available ages of metamorphism and magmatism discussing them within the regional context.

Geological setting

The Western Alps (Fig. 1a) originated from the subduction of the Jurassic Ligurian Piedmont Ocean and subsequent collision between the Adriatic and European plates from Cretaceous to Eocene times (Schmid et al.2004; Beltrando et al.2010; Angrand & Moutherau, 2021). The Western Alps consist of three main structural domains (Bousquet et al. 2008): (i) the southern external domain, corresponding to the Adria-derived domains, i.e., the Southern Alps; (ii) the external domain, representing the European foreland (Helvetic domain) and (iii) the axial zone between the Penninic Front to the north and the Insubric Line to the south (Schmid et al.1989; Handy et al.2010). The axial zone includes remnants of the subducted oceanic lithosphere of the Piedmont–Ligurian Ocean (i.e., Piedmont Zone; Pognante, 1979; Dal Piaz, 1999; Zanoni et al.2016; Luoni et al.2020; Caso et al.2021) and the continental crust of Adria (Giuntoli & Engi, 2016; Austroalpine Domain) and European pertinence (Penninic Domain). The DBTS (Manzotti et al.2014; Manzotti & Ballèvre, 2017) is a continental klippe of the Austroalpine domain, which is divided into the Dent-Blanche s.s. to the northwest and the Mont-Mary Nappe (Fig. 1b) to the southwest separated by the Rosain-Cignana Shear Zone (Fig. 1b; Manzotti et al.2014). The Dent-Blanche s.s. is composed of a Paleozoic basement corresponding to the Arolla and the Valpelline series (Argand, 1906; Diehl et al.1952; Manzotti & Zucali, 2013; Manzotti et al.2014). The Alpine-aged metamorphism is more pervasive in the Arolla Series ranging from greenschist- to blueschist-facies conditions (Roda & Zucali, 2008; Manzotti et al.2020), while in the VP, the overprint is weak and localised in decametric- to hectometric deformation zones. The pre-alpine metamorphism is mostly preserved in the VP and is related to Variscan and post-Variscan evolution of HT and pressures conditions, ranging from kyanite to sillimanite stability field (Gardien et al.1994; Manzotti & Zucali, 2013).

The VP consists of high-grade gneiss and migmatites associated with pegmatitic dikes, lenses of Grt–Cpx-amphibolite, basic granulite and marble (Diehl et al.1952; Nicot, 1977; Gardien et al.1994; Manzotti & Zucali, 2013). Regarding the pre-Alpine evolution, the first recognised tectono-metamorphic event (pre-D2, T = 700 ± 50 °C and P = 0.57 ± 0.11 GPa; Manzotti & Zucali, 2013) is marked by a relict Spre-2 foliation (referring to Manzotti & Zucali, 2013 nomenclature of the foliations and stages) preserved within metabasite and marble boudins enclosed within metapelite. Gardien et al. (1994) detected the oldest stage preserved within garnet cores and in the kyanite stability field (T = 700–800 °C and P = 0.9–1.0 GPa). The Spre-2 foliation, developed during the pre-D2 stage and preserved in the metabasic boudins, is wrapped by the S2 foliation. The S2 foliation developed during the subsequent D2 stage and is associated with melt production (T = 810 ± 40 °C and P = 0.7 ± 0.01 GPa; Gardien et al.1994; Manzotti & Zucali, 2013). The D2 stage has been dated to Permian (Zucali et al.2011; Pesenti et al.2012; Kunz et al.2018). During the D3 stage (T = 800 ± 30 °C and P = 0.6 ± 0.07 GPa; Manzotti & Zucali, 2013), the S2 is deformed by isoclinal folds, leading to the local development of an S3 axial plane foliation, marked by the growth of biotite and sillimanite, both at the meso- and microscale. Gardien et al. (1994) constrained the D3 event linked to the biotite–sillimanite S3 foliation at T = 750–800 °C and P = 0.55–0.6 GPa. Moreover, they recognised a further migmatitic stage related to the cordierite growth, estimated at T = 650–700 °C and P = ∼0.35–0.45 GPa. S2 and S3 are parallel to the lithological boundaries, strike NE-SW and dip at high-angle (Manzotti & Zucali, 2013). The D4 stage is locally associated with the development of a mylonitic foliation (S4) marked by chlorite and white mica, under greenschist-facies conditions at T = 430 ± 30 °C and P = 0.3–0.5 GPa (Gardien et al.1994; Manzotti & Zucali, 2013). Locally, faults and cataclastic rocks related to brittle deformation also occur (Pennachioni & Guermani, 1993; Roda & Zucali, 2008; Manzotti & Zucali, 2013).

The relative chronology of the tectono-metamorphic stages has been inferred by combining meso- and microstructural analysis, considering the geometric relationships between different structural elements (Passchier & Trouw, 2005; Spalla et al.2005). The deformation phases are named using a numerical progression (e.g., D1, D2, etc.), in which S- indicates foliations, A- fold axes and a.p. axial planes. Metamorphic assemblages linked to each D- stage have been inferred correlating each mineral phase with their respective structural element, following the guidelines of Johnson and Vernon (1995), Passchier and Trouw (2005) and Spalla et al. (2005). Different mineral generations have been distinguished using a numerical progression (e.g., Bt I, Bt II, etc.) to emphasise the chronological relationships between minerals and structures. Mineral abbreviations used in the description, except for Wm (white mica) and Oamp (orthoamphibole), are after Whitney and Evans (2010).

Major elements whole-rock compositions have been acquired by the ICP–AES (Inductively Coupled Plasma Atomic Emission Spectroscopy) at ‘Bureau Veritas Minerals’ laboratories in Vancouver (Canada) and through FUS–ICP (Fusion-Inductively Coupled Plasma Mass) at Activation Laboratories LTD. in Ontario (Canada). Some compositions have been obtained by combining modal estimations from optical microscope observation and mineral compositions from the literature (e.g., Nicot, 1977; Manzotti & Zucali, 2013). Moreover, data from Nicot (1977) and Diehl et al. (1952) acquired through X-ray fluorescence have been also reported. All the analyses are in Supplementary Material 1 available online.

Mineral chemical analysis has been performed through a Jeol JXA-8200 Superprobe WDS (wavelength-dispersive spectrometer, accelerating voltage of 15 kV, beam current of 15 nA) operating at the Dipartimento di Scienze della Terra ‘Ardito Desio’ – Università degli Studi di Milano (Italy). For trace-element analyses in rutile (i.e., Zr, Cr, V, Nb, U, Y, Fe, Al and Si), a beam current of 100 nA has been used. The calculation of interference between V and Ti and V and Cr has been also applied. Mineral formulae have been calculated using pyMin v.2023 (Zucali, 2023), which uses the recalculation rules of Deer et al. (1996). Amphiboles are classified according to Hawthorne et al. (2012) classification, and pyroxenes after Morimoto et al. (1988). Structural formulae have been calculated on the basis of 22 oxygens for biotite, 12 for garnet, 8 for plagioclase, 23 for amphibole, 6 for pyroxene and 18 for cordierite. FeOtot accounts for FeO and Fe2O3. XMg is defined as Mg/(Mg+Fe2+) for biotite.

The VP is dominated by metapelites, with subordinated metabasite and metacarbonatic rocks occurring as lenses within metapelites or interlayered with them (Fig. 2). The lithologies have also been discriminated due to their mineralogy (Fig. 3a and b and Table 1) and bulk rock chemical variations.

Metapelite

Migmatite gneiss

Migmatite gneiss covering ∼43 % of the VP at the map-scale (Fig. 2; Table 1) can be divided into four sub-types: (i) biotite + garnet + sillimanite-migmatite gneiss; (ii) biotite + garnet +  sillimanite ± cordierite-migmatite gneiss; (iii) biotite +  garnet +  sillimanite ± orthopyroxene-migmatite gneiss and (iv) fine-grained biotite-migmatite gneiss (Table 1).

Bt+Grt+Sil-migmatite gneiss is the most abundant sub-type, occurring as metric to decametric outcrops, interlayered with black amphibolite, sillimanite-gneiss, felsic granulite and marble. They show a gneissic structure made by a regular layering between millimetric- to pluricentimetric-thick brownish domains of biotite, garnet and sillimanite (i.e., melanosomes) and pluricentimetric-thick white quartz-feldspathic leucosomes (Fig. 4a and b). Melanosomes constitute a continuous to locally spaced foliation, defined by the elongation of biotite and sillimanite crystals making mm-thick layers. Leucosomes constitute plurimillimetric- to pluricentimetric-thick levels, locally appearing as lens-shaped decimetric pods wrapped by the melanosomes or cutting the main foliation. In some outcrops, they wrap decimetric fine-grained metabasic lenses. In detail, leucosomes are made of fine-grained quartz, plagioclase and garnet (blue star in Fig. 4a); locally centimetric K-feldspar occurs (red star in Fig. 4a). The modal abundance of sillimanite in melanosomes varies across the outcrops, with some nearly devoid or dominated by biotite. Conversely, in some locations, sillimanite modal abundance and grain size increase, reaching up to 2 mm in length. At the microscale, the gneissic structure of migmatite is defined by the isorientation of idiomorphic millimetric biotite and fibrous to locally prismatic sillimanite (Fig. 4b) with minor ilmenite. Biotite locally occurring as larger deformed relict grains shows an intense reddish-brown colour and very fine-grained inclusions of rutile. Garnet commonly displays rounded inclusions of biotite, quartz, sillimanite and minor plagioclase, K-feldspar and rutile, concentrated in the core. Garnet is wrapped by the biotite + sillimanite-rich main foliation. Relict deformed biotite and fibrous sillimanite, with quartz films in between, are present in microsites within skeletal garnet porphyroblasts. In some samples, garnet appears as anhedral millimetric blasts forming intergrowths with lobate quartz. The latter commonly shows a granoblastic structure with triple junctions in the less deformed or annealed leucosomes. A ‘chessboard’ structure is visible in the high-strain domains indicating deformation temperatures >700 °C (Stipp et al. 2002; Law, 2014). Quartz also forms irregular sub-millimetric to millimetric films among grain boundaries, such as garnet or biotite. Plagioclase in leucosomes is idiomorphic and sub-millimetric to plurimillimetric, often showing a sharp undeformed twinning. It is in mutual contact with quartz and when present, K-feldspar.

At the Lac Mort area (N 45°55’8” E 7°29’8” WGS84; Fig. 2), the aspect of migmatite gneiss is slightly different: the structure becomes more complex with folded leucosomes forming a ptygmatic structure. Moreover, the number of metabasite lenses and aplitic or pegmatitic dykes here increases. Sillimanite, which together with biotite highlights the main foliation, is fine-grained (∼1 mm in length) with a fibrous habit.

The second sub-type of migmatite gneiss, Bt+Grt+Sil±Crd-migmatite gneiss, crops in a few spots along the valley (see yellow stars in Fig. 2). Cordierite appears as dark grey porphyroblasts of up to 3 cm within leucosomes or in the melanosomes wrapped by the biotite + sillimanite-rich foliation (Fig. 4c). At the microscale, cordierite is sub-idiomorphic or interstitial, millimetric to plurimillimetric in size, showing widespread fractures along which fine-grained aggregates of white mica and chlorite occur (Fig. 4d). It includes sub-millimetric rounded quartz, biotite and sillimanite. Moreover, cordierite (as well as garnet) is locally overgrown by late-stage biotite and sillimanite.

Bt+Grt+Sil±Opx-migmatite gneiss (Fig. 4e) crops out in one locality in the valley (N 45°51’55”; E 7°24’13”; red star in Fig. 2). These rocks show a gneissic structure defined by the alternance of discontinuous millimetric biotite-rich layers and plurimillimetric to pluricentimetric leucosomes. Plurimillimetric green to grey orthopyroxene porphyroblasts occur mostly in the leucosomes, together with garnet, and are wrapped by the biotite foliation (Fig. 4f). Sillimanite modal abundance is lower with respect to the other migmatite gneiss types. At the microscale, the foliation forms slightly discontinuous layers mainly made of biotite lamellae wrapping leucosomes. Orthopyroxene is characterised by the presence of rounded inclusions of quartz and biotite. Locally, biotite forms millimetric symplectitic intergrowths with quartz, wrapping or growing over orthopyroxene porphyroblasts.

Fine-grained Bt-migmatite gneiss occurs as metric to plurimetric-thick bodies (not visible at the map-scale), defining domains alternating with biotite + sillimanite-rich foliation and leucosomes or folded decimetric levels within Bt+Grt+Sil-migmatite gneiss (Fig. 4g). They are abundant at the Lac Mort and Dzovenno areas (Fig. 2). These rocks consist of sub-millimetric quartz, plagioclase, garnet and biotite, the latter defining a weak gneissic fabric whose orientation is parallel to the foliation of the surrounding migmatite gneiss. Locally, their main foliation is crossed by millimetric- to pluricentimetric leucosomes (Fig. 4g). At the microscale, this lithology shows a gneissic to granoblastic structure made of sub-millimetric quartz and plagioclase showing contacts with triple junctions. Garnet constitutes millimetric subhedral blasts with few rounded quartz inclusions. Biotite is mainly interstitial (Fig. 4h) but locally occurs as subhedral to idiomorphic millimetric lamellae defining a weak foliation.

Sillimanite-gneiss

Sillimanite-gneiss occurs as hectometric elongated bodies striking parallel to the axial plane of the isoclinal folds in migmatite gneiss, i.e., with an NE–SW elongation. They are in contact with migmatite gneiss, and locally with black amphibolite and felsic granulite (Fig. 2). At the outcrop scale, the contact of these rocks with migmatite gneiss and felsic granulite can be both sharp and gradual. Sillimanite-gneiss are coarse-grained rocks with a foliated structure consisting of mm-thick layers of plurimillimetric prismatic to locally fibrous sillimanite crystals and minor biotite wrapping around garnet porphyroblasts, the latter with size reaching up to 2 cm in diameter. In this rock type, garnet is characterised by a cm-thick dark-purple core and a millimetre-thick pink rim (Fig. 4i). The sillimanite-rich foliation is pervasive and leucosomes are thinner and discontinuous with respect to those within migmatite gneiss.

At the microscale, the main foliation consists of isoriented idiomorphic prismatic sillimanite (Fig. 4j), along with minor biotite lamellae, and sub-millimetric ilmenite and graphite. Garnet cores are inclusions-rich with fine-grained quartz, graphite and sillimanite inclusions, that locally determine an internal foliation, discordant to the external pervasive foliation of the rock. Conversely, garnet rim is inclusions-poor (i.e., biotite, plagioclase, quartz, sillimanite and rutile). Inclusions within the rim are coarser with respect to those within garnet core and generally have a rounded shape. Locally, a millimetric film made of plagioclase and quartz marks the boundary between garnet crystals and the biotite + sillimanite foliation. Graphite occurs as millimetric to plurimillimetric grains in the cm-sized garnet strain shadows, along with sillimanite.

Felsic granulite

Felsic granulite crops out mainly in the central-southern part of the unit, as metric to decametric bodies interlayered with migmatite gneiss and sillimanite-gneiss and parallel to the strike of the foliation pervasive in migmatite gneiss at the map-scale (Fig. 2). The contact of these rocks with migmatite gneiss and sillimanite-gneiss is mostly gradual. In some outcrops, felsic granulites also occur within the neck of metabasite boudins. At the mesoscale, felsic granulite shows a granoblastic structure made of ∼5 mm-sized garnet porphyroblasts surrounded by a white plagioclase-rich matrix (Fig. 4k). Locally, garnet-rich domains highlight a weak layering. Their fabric is locally crosscut by decimetric-thick pegmatitic dykes, made of coarse-grained grey K-feldspars and quartz. At the microscale, garnet is idiomorphic and with sharp edges (Fig. 4l), and rounded sub-millimetric inclusions of quartz, biotite, rutile and minor zircon. Moreover, quartz occurs also as sub-millimetric and irregular films among garnet porphyroblasts. Rutile is particularly abundant in this rock type with respect to other lithologies: it is present as fine-grained inclusions within garnet porphyroblasts or occurs as millimetre-sized blasts in the plagioclase-rich matrix (Fig. 4l). Biotite constitutes deformed millimetric lamellae, in relatively low modal abundance with respect to other metapelites. Sillimanite shows a plurimillimetric prismatic habitus, wrapping around garnet. The plagioclase-rich matrix is often overgrowth by fine-grained aggregates of white mica and epidote (Fig. 4l).

Metabasite

Basic granulite

Basic granulite is mainly found two localities in the valley (N 45°55’8” E 7°29’8” and N 45°50’10”; E 7°21’47”) as plurimetric to pluricentimetric boudins (not mappable at the 1:10,000 map-scale; Table 1; Fig. 5a) and ellipsoidal lenses elongated parallel to the foliation pervasive in migmatite gneiss, or as close to isoclinally folded layers. These bodies are locally crosscut by a complex network of leucosomes and aplitic or pegmatitic dykes. In the rest of the unit, basic granulite crops out as smaller lenses (from 5 to 10 cm of thickness) within migmatite gneiss, sillimanite-gneiss and felsic granulite, locally separated by biotite + garnet-rich plurimillimetric to pluricentimetric portions (see zoom within Fig. 5a). The internal structure of basic granulite is granoblastic, locally gneissic, characterised by weak isorientation of fine-grained grey-greenish aggregates of pyroxenes and whitish plagioclase blasts. The orientation of their foliation is often discordant with respect to the pervasive foliation in migmatite gneiss. The contact between basic granulite and metapelite is marked by millimetric selvedges made of brown amphibole and biotite. Here, the modal abundance of amphibole gradually increases at the expense of pyroxene, up to becoming an amphibolite. Locally, a centimetric pale greyish amphibole occurs within the amphibole-rich layers, with no preferential orientation (Fig. 5b).

At the microscale orthopyroxene and minor clinopyroxene, orthoamphibole, brown clinoamphibole and minor biotite and quartz are recognised (Fig. 5c): (i) orthopyroxene and clinopyroxene occur as anhedral sub-millimetric blasts; (ii) orthoamphibole is interstitial and colourless with twinning; (iii) plagioclase domains occur as grey to brownish fine-grained aggregates that consist of white mica and epidote (Fig. 5c) and (iv) brown amphibole forms sub-millimetric anhedral crystals growing after pyroxenes. Locally, plurimillimetric garnet porphyroblasts occur in plagioclase-rich domains with a subhedral and locally skeletal habitus including rounded sub-millimetric plagioclase, amphibole and minor quartz showing no preferential orientation.

Brown amphibolite

Brown amphibolite occurs as metric boudins (not mappable at the 1:10,000 map-scale), lenses or folded layers within migmatite gneiss (Table 1; Fig. 5d). These rocks have a fine-grained layered gneissic structure made by the isorientation of brown amphibole and minor plurimillimetric plagioclase aggregates, defining a weak foliation. Pegmatites hosting plagioclase and biotite occur at the contact between these lenses and the surrounding migmatite gneiss or locally crosscutting their fabric.

At the microscale, brown amphibole occurs as subhedral crystals marking the main foliation, together with interstitial sub-millimetric clinopyroxene and orthoamphibole blasts (Fig. 5e). Plagioclase marking the main fabric is replaced by brownish fine-grained aggregates of white mica and epidote. Ilmenite occurs as an accessory phase and marks the main foliation. At the contact of these bodies with pegmatites or leucosomes, clinopyroxene and ‘fresh’ plagioclase blasts increase in modal abundance together with euhedral biotite, while brown amphibole becomes absent. Locally, plurimillimetric sites made by granoblastic quartz, plagioclase and skeletal garnet similar to those occurring in basic granulites have been detected.

Black amphibolite

These rocks are widespread along the valley as decametric to hectometric bodies, interlayered with metapelites. At the map-scale, these bodies are elongated NE–SW, parallel to the strike of the structures pervasive in metapelites (Fig. 2). Black amphibolite shows a layered structure made by the regular alternance between cm-thick black amphibole-rich and leucocratic pluricentimetric plagioclase-rich levels (Fig. 5f and g). Amphibole shape-preferred orientation (SPO) marks the main foliation (i.e., parallel to the regional foliation). Locally, up to 10 – 15 cm in diameter reddish garnet porphyroblasts occur within quartz-feldspathic layers (Fig. 5g). Aggregates of plagioclase, quartz and minor clinopyroxene generally rim these large garnets (Fig. 5g). Clinopyroxene also makes 1–2 cm-thick layers parallel to the main foliation or may occur as coarse-grained aggregates in leucocratic domains, generally wrapped by the main foliation. Locally black amphibolite levels are characterised by a pluricentimetric greenish selvedge at the contact with migmatite gneiss, where amphibole is absent, and pyroxene increases in modal abundance with respect to ‘fresh’ amphibolites. Leucocratic domains locally crosscut the main foliation of these rocks. Moreover, black amphibolites are crosscut by up to 10 cm-thick Qz+Pl+Amp-pegmatite.

At the microscale amphibole occurs as sub-idiomorphic blasts, marking the main foliation, together with minor sub-millimetric biotite lamellae (Fig. 5h). Clinopyroxene is present both as sub-millimetric relict blasts and as up to cm-thick aggregates in plagioclase and quartz-rich domains. Locally, leucocratic domains are characterised by the presence of a few millimetric calcite blasts. Titanite occurs as an accessory mineral marking the main foliation (Fig. 5h).

Metacarbonates

This group comprises pure and impure marbles, and minor calcsilicates-bearing rocks (Fig. 5i; Table 1). Metacarbonates at the map-scale occur as decametric to hectometric bodies whose elongation is parallel to the NE–SW strike of the other lithologies (Fig. 2). At the outcrop scale, they occur both as interlayered with metapelites and as metric boudins within migmatite gneiss. Marbles are generally grey and locally yellowish, with a massive structure, which becomes locally gneissic due to the slight isorientation of calcite aggregates or quartz. Locally, greenish millimetric crystals of olivine can be recognised. Calcsilicate-bearing rocks occur mainly in the centre-northern part of the unit; they show a brownish-yellowish colour and are characterised by a foliated structure given by the SPO of calcite aggregates and thin quartz levels. In particular, the latter highlight disrupted fold hinges (Fig. 5i) or occur as up to 20 cm-clasts in the calcite-rich matrix.

At the microscale marble is made of sub-millimetric to millimetric calcite blasts, locally showing large, deformed twinning. Sub-millimetric anhedral olivine and clinopyroxene occur (Fig. 5j). Quartz and plagioclase are locally present together with calcite; plagioclase is overgrown by fine aggregates of white mica. In correspondence with some grain boundaries, thin films made of quartz occur (red arrow in Fig. 5j).

Aplite and pegmatite

Aplite and pegmatite are widespread throughout the unit and are found in all lithologies, mainly striking NNW–SSE and NW–SE. They form dykes up to 1 m-thick, parallel to the pervasive foliation at the map- and outcrop-scale or cutting the main foliation with complex patterns, made by different crosscutting generations of dykes. In particular, three main types of dykes have been distinguished based on their composition and structure: (i) Qz–Kfs-bearing aplites which occur in m-thick bodies in migmatite gneiss (Fig. 5k and l) and metacarbonate discordant to their main foliation; (ii) Pl–Bt-bearing pegmatite, occurring as 10 cm- to 1 m-thick dykes within brown and black amphibolites, or at the contact between these and migmatite gneiss and (iii) Amp-bearing pegmatites which occur in black amphibolite, cutting their main foliation. The first type forms 5 to 40 cm-thick dykes, with an isotropic structure made of plurimillimetric to centimetric sub-idiomorphic quartz, plagioclase and K-feldspar, plus minor biotite and muscovite. The second type shows an isotropic structure made of pluricentrimetric lamellae of biotite and whitish plagioclase crystals. Amp-bearing pegmatites have an isotropic structure made of dark green plurimillimetric to centimetric idiomorphic crystals in a whitish quartz and plagioclase matrix.

D1 stage

The D1 is well recognisable at the mesoscale in basic granulite and brown amphibolite as an S1 foliation, which occurs within metre-sized boudins, having various orientations to the main foliation and elongation of main lithological boundaries (Fig. 2). The S1 foliation has scattered orientations, with three main clusters: at Lac Mort, it dips towards the S, while at Oyace (N 45°51’1” E 7°22’54”) toward both W–SW and E (Fig. 2).

In basic granulite, the S1 foliation is a discontinuous layering marked by the isorientation of pyroxene- and plagioclase-rich domains. At the outcrop scale, the S1 is geometrically discordant to the long axis of the boudin and to the penetrative foliation that commonly wraps them (Fig. 6a). At the microscale, the S1 is marked by the weak SPO of Opx I, Bt I (Fig. 7) and minor orthoamphibole. Pl I occurs as granoblastic domains, locally replaced by dark grey fine-grained aggregates.

In brown amphibolite, the S1 foliation is defined by alternating brown amphibole-rich and plagioclase-rich mm-thick layers, where brown amphibole shows an SPO parallel to the layering (Fig. 6b). Biotite SPO may also contribute to the layering, while clinopyroxene occurs as isolated grains within amphibole-rich layers. At the microscale, Amp I is brown hornblende, and Cpx I occur as interstitial grains in amphibole-rich layers or are associated with Pl I microdomains.

In metapelites, a remnant of this phase is testified by Grt I inclusions-rich cores in bistadial garnets within sillimanite-gneiss. In particular, inclusions of Bt I, Sil I, Qz I and Pl I locally are isoriented marking a foliation discordant to the external one. However, the rounded Bt I, Qz I, Pl I and Sil I inclusions within Grt II and Crd are not always oriented (Fig. 6c and d).

D2 stage

During the D2, the main foliation (S2) develops, and the main lithological boundaries are parallelised both at map- and mesoscale. The S2 foliation, the lithological boundaries, and the alignment of basic granulite and brown amphibolite boudins tend to strike NE–SW and dip vertically toward NW or SE (Fig. 2). Sillimanite-gneiss, felsic granulite and black amphibolite bodies interlayered with migmatite gneiss are elongated parallel to the S2 (Fig. 2). Metre-scale D2 folds are recognised, folding the S1 foliation and the lithological boundaries between amphibolite or basic granulite and migmatite gneiss (Fig. 6b). A2 fold axes are scattered but with two main clusters dipping at a low angle toward the NE and SW. Axial planes (a.p.2) mainly dip toward the S–SE with relatively high dip angles (Fig. 2).

The S2 is the main foliation in migmatite gneiss and is represented by the alternation between leucosomes and melanosomes (Fig. 6e). In cordierite-bearing migmatite gneiss melanosomes are defined by idiomorphic Bt II, locally in apparent textural equilibrium with Grt II (Fig. 6f), and minor fibrous Sil II. In garnet- and orthopyroxene-migmatite gneiss the S2 is made by melanosomes whereas biotite has a relict aspect and is deformed. Conversely, in sillimanite-gneiss, the S2 is almost completely transposed or preserved within microlithons at the microscale. In felsic granulite, the S2 foliation is marked by the weak isorientation of garnet-rich layers. At the microscale, the S2 foliation is marked by the SPO of Bt II and minor Sil II in millimetric to plurimillimetric layers, alternating with the leucosomes (Fig. 6f). Grt II, Crd and Opx II porphyroblasts grow in metapelites (Fig. 7). Grt II occurs as inclusions-poor porphyroblasts within migmatite gneiss and sillimanite-gneiss (Fig. 6f), as the rim of inclusions-rich Grt I cores (Fig. 6c), or as blasts in equilibrium with cordierite or orthopyroxene (Opx II). In felsic granulite, the D2 is testified by the gentle isorientation of Grt II + Pl II domains marking the S2.

In black amphibolite, the S2 is made by alternate dark amphibole- plus clinopyroxene-rich and quartz-feldspathic levels (Fig. 6g). The quartz-feldspatic portions of these amphibolites locally show irregular patterns, crosscutting the penetrative S2 foliation; in these portions, up to 10 cm garnet or pluricentimetric clinopyroxene porphyroblasts occur. At the microscale, the S2 is marked by the isorientation of the dark Amp II, Cpx II and Pl II + Qz II + Cc in the quartz-feldspathic plurimillimetric to pluricentimetric lithons. S2 is also locally marked by Ttn blasts.

In marble, the S2 is marked by the SPO of calcite and quartz-rich aggregates, with minor clinopyroxene or olivine. In calcsilicates, the S2 foliation is locally mylonitic, wrapping around up to 30 cm quartz-rich porphyroclasts.

During the D2 stage partial melting of metapelites and, in a minor amount, of black amphibolite occurs. In metapelites, this is mainly testified by the presence of leucosomes rich in quartz and feldspars marking the S2 foliation. In black amphibolite, it could be testified by the presence of quartz and plagioclase-rich domains, which in some cases, crosscut the main foliation (Fig. 6g).

D3 stage

The D3 deforms the previous S2 foliation producing tight to isoclinal folds at the meso- and map-scale, with rounded to acute hinges, locally disrupted, leading to the local transposition of the S2 and the development of S3 foliation. At the map and mesoscale, the S3 is pervasive only in the sillimanite-gneiss and is locally parallel to the axial plane of the folds deforming the S2. It dips mainly towards E–SE and NW with high dip angles. A3 fold axes plunge mainly towards N–NE and E (Fig. 2).

At the mesoscale, in migmatite gneiss, the D3 deformation phase is locally testified by close to isoclinal folds that deform the S2 foliation (Fig. 6h), locally transposed into a S3, the latter marked by the SPO of Bt III and Sil III. Leucosomes, likely produced during the D2, occur also within syn-D3 fold hinges or neck boudins (Fig. 6e). They are made of perfectly igneous granitic texture, but locally the grains show preferred orientation. In sillimanite-gneiss, the S2 is completely transposed, and the pervasive S3 is defined by the SPO of Sil III and minor Bt III (Fig. 6i) wrapping around the centimetric Grt I + Grt II porphyroblasts. However, some garnet porphyroblasts show a rim rich in fine sub-millimetric sillimanite inclusions, which outline a foliation concordant and parallel to the external S3, suggesting that garnet is also syn-kinematic with respect to the D3 phase (Figs. 6j and 7). At the Lac Mort area, asymmetric folds deform migmatite gneiss, also affecting cordierite blast aggregates. During this stage sillimanite and biotite locally grow after garnet or cordierite. In some cases, fine-grained intergrowths of Bt III + Qz II over Opx II occur (Figs. 6k and 7).

In black amphibolite, the D3 deformation phase is only locally detectable as close to isoclinal folds deforming the main S2 foliation.

Aplites and pegmatites intrude all the rock types locally along their main foliation, but in most cases, they crosscut the D2 and D3 structures. However, some of the dykes are also folded, probably those emplaced during the late-D3 stage (Fig. 6l).

D4 stage

This stage is localised in some hm-thick belts (Fig. 2) as an S4 foliation made by white mica and chlorite or as ‘coronitic’ domains where the blastesis occurs without deformation, better recognisable at the microscale. In metapelites, fine-grained white mica aggregates grow over cordierite along its fractures, or over plagioclase and sillimanite. Chlorite partly pseudomorphs biotite and garnet along fractures. Fine-grained aggregates of epidote and white mica grow after plagioclase both in metabasite and metapelite. Locally, brown amphibole is characterised by a thin rim of green amphibole.

Inferred stable mineral associations

Based on the relationships between the mineral phases and the observed marking fabrics, it is possible to summarise the inferred mineral associations stable during each deformation phase (Fig. 7).

The D1 in metapelite is defined by the stable association: Bt I + Sil I + Qz I + Pl I + Kfs I + Grt I ± Rt I.

In basic granulites is defined by Opx I + Cpx I + Qz I + Pl I + Oamp I ± Bt I ± Amp I and by Amp I + Cpx I +  Pl I + Qz I ± Bt I ± Oamp I in brown amphibolite.

The D2 is defined by the stable associations: Grt II + Crd + Opx II + Pl II ± Rt II + melt (in metapelite) or Bt II + Grt II + Crd + Pl II + melt and Amp II + Pl II + Cpx II + Grt II ± Ttn ± melt (in black amphibolite).

The D3 is characterised by the stable association: Sil III + Bt III ± Qz II ± Rt III ± Ilm ± Gf.

According to microstructural constraints, during the late D3, the melt produced during the D2 crystallises (Qz III + Pl III + Kfs II).

Metapelite

Metapelites show mineralogical differences confirmed by whole-rock chemical variations. In particular, SiO2 in migmatite gneiss covers a relatively wide range of values from ∼42 to 75 wt%, while in sillimanite-gneiss this range is smaller, i.e., from ∼46 to 60 wt% (Fig. 8). Felsic granulite shows a SiO2 content of ∼43 wt%, similar to sillimanite-gneiss (Fig. 8). Alkali content is similar among all analysed metapelites, between ∼3 and ∼6%, except for some modal-derived bulk compositions having relatively low values (<2%) or higher, up to ∼8% (Fig. 8a). Al2O3 wt% has highest values in correspondence of sillimanite-gneiss, while decreases in most of the migmatite gneiss samples (Fig. 8b). Pegmatite has a relatively low concentration of Al2O3 and a higher one of SiO2 of ∼72 wt% (Fig. 8b). The iron content (Fe2O3 wt%) of migmatite gneiss varies between ∼3 and 20 wt%, depending on garnet modal abundance; in sillimanite-gneiss, it ranges from ∼7 to 14 wt% (Fig. 8c). Pegmatite has the lowest iron content (∼2%). The MgO and CaO (wt%) contents are constant through different samples: they range from ∼1 to 7 wt% (Fig. 8d) and ∼0.1% to ∼5% (Fig. 8e), respectively. A distinction between migmatite and sillimanite-gneiss is evident from the TiO2 values: it is lower in migmatite gneiss in which it ranges between ∼0.1 and 2%, while is higher in sillimanite-gneiss (∼1.4 to 3 wt%; Fig. 8f).

Metabasite

Whole-rock analyses on metabasites confirm the differences among the described groups: black amphibolite shows a SiO2 wt% content ranging between ∼40 and 47%, while brown amphibolite at 45 wt%; basic granulites show higher values, ranging from ∼45 to 55 wt% (Fig. 8). The alkali content varies between ∼1 and 6.8 wt%, with the lowest values in basic granulite samples from Nicot (1977). Al2O3 ranges between ∼4 and 22% in all metabasites (Fig. 8b). Fe2O3 wt% content is higher in black amphibolite (∼13 – 20 wt%) and lower in brown amphibolite (∼12%) and basic granulite (∼7 – 15%; Fig. 8c). MgO content is between ∼4.8 and 7 wt% in most of the metabasites, with higher values (up to ∼17%; Fig. 8d) in brown amphibolite and two samples of basic granulite. The difference among metabasite is evident also looking at the CaO content: the highest values are in correspondence with black amphibolite and brown amphibolite (up to 11 wt%) and decrease in basic granulite (Fig. 8e). TiO2 ranges between ∼0.4 and 6%, with the highest values in black amphibolite (Fig. 8f).

According to the microstructural relationships described above, mineral phases were selected for chemical analyses performed by the electron microprobe analyser. In this section, mineral chemistry is described and discussed with respect to the bulk rock system. All data from this work and Nicot (1977) are in Appendixes 15 in the Supplementary Material available online.

Biotite composition in metapelites varies depending on the lithotype, mostly between Grt-only metapelite (i.e., sillimanite-gneiss and migmatite gneiss), and cordierite- and orthopyroxene-bearing migmatite gneiss, but without a clear distinction between biotite formed during different stages (i.e., between Bt I, Bt II and Bt III distinguished from their microstructural position; Fig. 9a). The lowest Ti a.p.f.u. (atoms per formula units) values correspond to biotites from Bt+Grt+Sil±Crd-migmatite gneiss, which are characterised by Ti contents from 0.27 to 0.69, with most of the biotite having Ti content in the 0.35–0.60 range and an average of 0.44 (1σ = 0.09). Biotite from Bt+Grt+Sil-migmatite gneiss has Ti ranging between ∼0.35 and 0.65 a.p.f.u. and has slightly higher average values (0.47, 1σ = 0.08). Biotite from Bt+Grt+Sil±Opx-migmatite has the highest content in Ti, between 0.55 and 0.79. Ti values of biotite from sillimanite-gneiss range between 0.20 and 0.70, but with two clusters at 0.25–0.45 and 0.4–0.7, depending on the sample. The Ti a.p.f.u. content values in biotite from felsic granulite are scattered and range between ∼0.30 and 0.65. Ti values in biotite from fine-grained Bt-migmatite gneiss range between ∼0.35 and 0.65, with an average of 0.49 a.p.f.u. for the felsic granulites and 0.42 for the gneiss (Fig. 9a). The XMg varies between 0.50 and 0.72 in Bt+Grt+Sil±Crd-migmatite gneiss and between 0.57 and 0.66 in the Bt+Grt+Sil±Opx-migmatite gneiss. In Bt+Grt+Sil-migmatite gneiss there are two clusters, the first one has XMg between ∼0.40 and 0.55 and the second has values in the 0.53 and 0.77 range. Biotite in sillimanite-gneiss has XMg between 0.57 and 0.77, while in felsic granulite, it is 0.68–0.70 (Fig. 9a). XMg in biotite from fine-grained Bt-migmatite gneiss range between 0.55 and 0.63 (Fig. 9a).

Biotite from black amphibolite has Ti between 0.34 and 0.40 a.p.f.u., and XMg of 0.46 and 0.47. In brown amphibolite there are two main clusters: (i) Ti values between 0.53 and 0.66 a.p.f.u. and XMg from 0.54 to 0.65; (ii) Ti a.p.f.u. from 0.35 to 0.38 and XMg 0.79–0.80. The former is in correspondence with biotite marking the S1 foliation, the latter is in correspondence with biotite growing at the boundary with the leucosome of surrounding metapelite (Fig. 9b).

Garnet from all metapelites is Fe-rich, with FeO wt% greater than 30. Generally, single garnet crystals are relatively chemically homogeneous grains, with only local zoning toward the rim, especially in contact with biotite. Zoning is more pronounced in sillimanite-gneiss where it is possible to distinguish a garnet core and garnet rim. In detail, garnet from Bt+Grt+Sil-migmatite gneiss has almandine content (Alm) ranging from ∼65 to ∼80% and pyrope (Prp) between 15 and 30%. Garnets from Bt+Grt+Sil±Crd-migmatite gneiss show Alm content of ∼65–76%; Prp from 20 to 30% and grossular (Grs) and spessartine (Sps) are ∼2–3% and ∼1–3% respectively. Garnets from Bt+Grt+Sil±Opx-migmatite gneiss are more homogeneous (Alm ∼60 and 65%) and are higher in CaO, with Grs from 6 to 11% (Fig. 9c). Garnet in fine-grained Bt-migmatite gneiss has Alm content between 70 and 75% and Prp from 20 to 26%; Grs and Sps are constant at ∼20% (Fig. 9c). In sillimanite-gneiss, there is an increase in the Fe content and Mg decrease from Grt I (cores) to Grt II (rims), with Alm content varying from core to rim from ∼60 to ∼80%, and Prp content increasing from ∼15 to ∼35%. Grs and Sps range between 5–7% and 1–2%, respectively (Fig. 9c). Felsic granulite garnets are also relatively homogeneous: the Alm content is between ∼59–69% while the Prp content is between ∼29–34%; Grs and Sps are 5–8% and 1–2% respectively (Fig. 9c). Garnet analyses from metabasite have been taken from Nicot (1977). Garnets from black amphibolite are Alm-rich, but relatively enriched in the Grs component with respect to metapelite. Alm content ranges from ∼62 to 77%, while Prp content varies only between ∼24 and 26%. Basic granulite garnets show two clusters of composition: one is richer in the Prp component, and one is richer in the Alm and Grs+Sps components (Fig. 9d).

Feldspars from Bt+Grt+Sil-migmatite correspond to K-feldspar and plagioclase: the former has Or (orthoclase) component ranging from 0.87 to 0.92, the latter has An component ranging between 0.25 and 0.54. Plagioclase in Bt+Grt+Sil±Crd-migmatite gneiss varies from pure Ab crystals to having An (anorthite component) between 0.20 and 0.42. K-feldspar has Or content ranging from 0.76 and 0.81 (Fig. 9e).

Bt+Grt+Sil±Opx-migmatite gneiss plagioclase has An component ranging from 0.20 to 0.43; K-feldspar has Or component ranging between ∼0.80 and 0.90. In sillimanite-gneiss, they have two clusters: one has An of 0.46, and the other is richer in Ab and has lower An content (0.03–0.20). K-feldspars have Or content of ∼0.90 (Fig. 9e). Feldspar from felsic granulite corresponds only to plagioclase. An values are between ∼0.40 and 0.50 (Fig. 9e). Plagioclase from black amphibolite, marking the S2 foliation, has An content ranging from ∼0.18 to 0.35. Those from brown amphibolite show two clusters of compositions: one is almost Ab-pure and corresponds to plagioclase marking the S1 foliation, and the second has An ranging between 0.50 and 0.56 corresponding to recrystallised plagioclases at the boundary with metapelite leucosome (Fig. 9f).

Amphibole from black amphibolite, considering both our new data and those from Nicot (1977), has Ca content approximately of 1.7 a.p.f.u. FeTOT + Mg a.p.f.u. content is 4.1–4.4 a.p.f.u (Fig. 10a). AlTOT varies between ∼1.7 and 2.7 a.p.f.u. while the Ti content ranges between 0.15 and 0.26 a.p.f.u (Fig. 10b). Amphibole from brown amphibolite has Ca content varying between 1.41 and 1.89 a.p.f.u. FeTOT + Mg a.p.f.u. range from 4.1 to 4.6 (Fig. 10a). AlTOT ranges from 1.7 to 2 (Fig. 10b). Amphiboles from black amphibolite correspond to pargasite, those from brown amphibolite correspond to Mg-hornblende and minor pargasite. Most of the amphiboles in basic granulite have almost no Ca, and correspond to orthoamphibole, apart from a few grains having Ca of ∼1.8 a.p.f.u. Fe2+ in orthoamphibole is of ∼2.4 a.p.f.u. while in the Ca-rich one is ∼0.9 a.p.f.u. (Fig. 10a). Orthoamphibole has AlTOT and Ti relatively low compared to clinoamphibole from other lithologies (Fig. 10b).

Pyroxene in migmatite gneiss is enstatite (Fig. 10c and d). Fe is 0.8 a.p.f.u., Mg is 1.16 a.p.f.u. In black amphibolite, clinopyroxene is diopside (Fig. 10d), with low Fe2+ and Mg contents. Fe ranges between 0.22 and 0.52 a.p.f.u., while Mg ranges from 0.58 to 0.74 a.p.f.u. (Fig. 10c). Al varies between 0.07 and 0.41 a.p.f.u. while Na ranges from 0.05 to 0.13 a.p.f.u. Brown amphibolite has both clino- and orthopyroxene. The former corresponds to those with relatively low Fe2+ and Mg a.p.f.u. values (Fig. 10c) and the latter to those with higher values. Both clino- and orthopyroxenes mark the S1 foliation. Al varies between 0.03 and 0.11 while Na reaches 0.03 a.p.f.u. Pyroxenes from basic granulite correspond to orthopyroxene, with relatively high Fe2+ and Mg a.p.f.u. values of ∼0.75 and ∼1.20, respectively (Fig. 10c). They fall in the compositional range of enstatite (Fig. 10d).

Cordierite from migmatite gneiss shows a slight chemical zoning: XMg (Mg/Mg+Fe) varies from 0.65 to 0.77, i.e., from the core towards the rim. Fe2+ a.p.f.u. values range from 0.45 to 0.71, while Mg varies between 1.30 and 1.53 a.p.f.u. (Fig. 10e).

Rutile in sillimanite-gneiss has Nb concentrations covering a wide range of values from ∼200 to ∼2600 ppm and Zr concentrations from ∼400 up to ∼2000 ppm. Cr concentrations vary between ∼200 and ∼700 ppm. Rutiles from Bt+Grt+Sil-migmatite gneiss are more homogeneous than those from sillimanite-gneiss. Nb varies from ∼800 to ∼1600 ppm (except for one grain which has 2500 ppm of Nb) and Zr from ∼440 to ∼1400 ppm (Fig. 10f). Cr concentrations range from ∼250 to ∼450 ppm (Fig. 10f). Rutiles of felsic granulite have Fe concentration ranging between ∼1000 and 6000 ppm but with a main cluster between ∼1000 and 2000 ppm. Zr content ranges between ∼500 and 2500 ppm. Nb concentration is between ∼100 and 2000 ppm, while Cr is quite constant at about 600 ppm with a maximum of ∼1600 ppm.

Grt–Bt exchange

The Grt–Bt exchange between garnet and biotite has been applied on three metapelite samples, on a total of 20 pairs. For these pairs have been used both analyses from garnet core and included biotite, or garnet rim and neighbouring biotite in the matrix or in garnet strain shadows. The Ferry and Spear (1978) calibration has been chosen since garnet is mainly Fe- and Mg-rich. Sillimanite-gneiss shows temperatures ranging from 662 to 736 °C, from three pairs. Bt+Grt+Sil±Crd-migmatite gneiss has temperatures ranging from 556 to 730 °C while those from Bt+Grt+Sil±Opx-migmatite gneiss, taken from four pairs are between 680 and 770 °C (Fig. 11a; Table 2).

Zr-in-rutile

The temperature of rutile crystallisation has been calculated based on its Zr content following the calibration of Tomkins et al. (2007). Since the Zr-in-rutile thermometer is pressure-dependent, a value of 0.7 GPa has been used for the calculations. This pressure was estimated for the migmatitic events by Manzotti and Zucali (2013) and Gardien et al. (1994) based on conventional geobarometry. A total of 203 results have been obtained: 112 from sillimanite-gneiss, 43 from Bt+Grt+Sil-migmatite gneiss, and 48 from felsic granulite. Sillimanite-gneiss yielded temperatures ranging from 691 to 818 °C (see Table 2 for average T °C and errors), while Bt+Sil+Grt-migmatite gneiss temperatures vary between 683 and 773 °C (Figs. 11b and 12; Table 2). In felsic granulite, temperatures range between 667 and 843 °C (Fig. 11b).

Ti-in-biotite

The Ti-in-Bt thermometer (Henry et al.2005) has been applied to biotites from three different microstructural domains (Bt I, Bt II and Bt III) on 18 samples (Fig. 11c). Four samples of sillimanite-gneiss collected at different localities have been analysed (139 biotite analyses used). Except for one sample (belonging to the west of Vessonaz stream) having temperatures between 649 and 727 °C, sillimanite-gneiss have values between 721 and 812 °C (see Table 2 for average T °C for each sample and error). Five samples of Bt+Grt+Sil-migmatite gneiss have been analysed (167 biotite analyses used) giving temperatures between 676 and 784 °C. Five Bt+Grt+Sil±Crd-migmatite gneiss samples (179 biotite analyses used) have temperatures ranging from 654 to 789 °C. Temperatures in Bt+Grt+Sil±Opx-migmatite gneiss (27 biotite analyses used) are quite higher: the two samples have temperatures ranging from 700 to 799 °C and from 760 to 812 °C pointing to higher average T °C (Table 2). Fine-grained Bt-migmatite gneiss (25 analyses used) yielded temperatures between 715 and 764 °C. Biotite from felsic granulite has been analysed only on one sample (from the Perquis area) due to its low modal abundance. Temperatures (13 biotite analyses used) are from 729 to 784 °C (Table 2). Since the chemical composition between different biotite generations is similar, the resulting temperatures do not show significant variations (Fig. 12).

Ti in-amphibole

The Ti-in-Amp calibration of Liao et al. (2021) has been used on 34 analyses of black amphibolite (plus 18 analyses from Nicot, 1977) and 51 analyses of brown amphibolite (Fig. 11d). Black amphibolite has temperatures ranging from 746 to 880 °C with two main clusters at 767 ± 19 °C and 844 ± 21 °C (Figs. 11d and 12; Table 2). Brown amphibolite has relatively higher temperatures ranging from 858 to 916 °C (Figs. 11d and 12; Table 2). Among the Nicot (1977) samples, D48 gave temperatures between 791 and 859 °C with a mean of 825 ± 48 °C; D16 gave temperatures between 695 and 893 °C with a mean of 783 ± 74 °C. B78 gives temperatures ranging between 750 and 901 °C with a mean value of 823 ± 68 °C (Figs. 11d and 12; Table 2).

Pressures have been obtained applying two geobarometers, i.e., the Grt–Crd barometer (Bhattacharya, 1986) on one cordierite-bearing sample and the Pl–Amp barometer (Molina et al.2015) on one black amphibolite sample.

Grt–Crd geobarometer

The Grt–Crd barometer (Bhattacharya, 1986) has been used on the 33SC sample, corresponding to a Bt+Sil+Grt±Crd-migmatite gneiss. Three pairs of Grt–Crd (using cordierite rim and core, without significant P differences) have been used for the calculation, both with the minimum and maximum T °C estimated through the Ti-in-Bt thermometer on the same sample (i.e., 704 and 769 °C, respectively). Both temperatures used yielded similar values, ranging between 0.55 and 0.58 GPa, with an error of 0.01 GPa.

Pl–Amp geobarometer

The Pl–Amp geobarometer of Molina et al. (2015) has been used on a total of 17 pairs. Even for this sample, we used both the minimum and maximum T (i.e., 767 and 879 °C) for the P estimation, which was calculated through the Ti-in-Amp thermometer. Obtained pressure ranges between 0.8 ± 0.08 GPa (at 767 °C) and 1.1 ± 0.10 GPa (at 879 °C).

The VP is well suited to investigate: (i) pervasive fabrics at unit scale during melt production and migmatisation; (ii) the thermal state of the lower crust during the melt-producing and melt-consuming stages and (iii) the tectono-metamorphic evolution of the Adria lower crust and its geodynamic significance, comparing the VP with the well-studied case of the IVZ.

Pervasive fabrics associated with melt production and migmatisation at the unit scale

Differently to previous studies, which were merely lithological (i.e., Diehl et al.1952; the Monte Cervino Sheet n.70 of Geological Map of Italy; Dal Piaz et al. 2016) or restricted to more localised areas (Manzotti & Zucali, 2013), this work extends the description of rock types and associated structures at the whole unit scale.

This work detailed the distribution of partially melted metapelites with variable textures and %vol of leucosomes, with decreasing melt abundance: migmatite gneiss, sillimanite-gneiss and felsic granulite. Metabasites have also been thoroughly described and separated in terms of bulk rock chemistry and marking structures: (i) the metric-scale basic granulite and brown amphibolite lenses preserve the oldest recognised structures (i.e., S1) while (ii) black amphibolites interlayered with metapelites mark the regional foliation (S2).

Three deformation phases (D1, D2 and D3) related to the HT evolution of the VP have been recognised. The oldest structures and stable assemblages (D1 stage) preserved in basic granulite and brown amphibolite boudins as an S1 foliation (green foliation traces in Fig. 13) are locally discordant to the S2 foliation pervasive in metapelites (green traces in Fig. 13). The evidence of D1 is detected just in a few localities in the valley: (i) in brown amphibolite metre-scale lenses at Oyace; (ii) in basic granulites and brown amphibolites decametric to metric-scale lenses and boudins in Thoules (as previously also described by Manzotti & Zucali, 2013 as Spre-2) and (iii) at Lac Mort area within basic granulites and brown amphibolites lenses or folded levels. The S1, marked by the stable association of Amp I + Opx I + Cpx I + Pl I, is not accompanied by an extensive production of melt, except for rare microscale Pl + Grt-rich domains, which may indicate localised melting.

The D2 tectono-metamorphic stage is related to the development of the S2 regional foliation pervasive at the map-scale in metapelites, black amphibolites and metacarbonatic rocks (blue traces in Fig. 13). The S2 is marked in migmatite gneiss by Bt+Sil+Grt (i.e., melanosomes) and quartz-feldspathic domains. These domains, interpreted as leucosomes, are parallel to the S2 regional foliation at the micro-, meso- and map-scales and testify to widespread melt production at different scales during the D2 stage. Leucosomes, which mark the D2-related fabrics, contain garnet, cordierite and orthopyroxene (Fig. 13). These three phases are interpreted as peritectic products related to melt production. Garnet occurs widely in the VP rocks, while cordierite and orthopyroxene are detected only in some localities (see yellow and red stars in Fig. 13). In detail, cordierite occurs with garnet; orthopyroxene is also together with garnet, but an Opx–Crd-bearing assemblage has not been detected.

The S2 foliation is reactivated during the D3 stage, leading to its local transposition and development of the S3 axial plane foliation (Fig. 13). The S3 is pervasive in sillimanite-gneiss and in correspondence of m-wide localised high-strain domains in the migmatite gneiss; it wraps garnet, cordierite and orthopyroxene. Gardien et al. (1994) pointed out the presence of cordierite growing after regional foliation development. Conversely, our observations highlight that cordierite develops during the D2 stage and is wrapped by the S3 sillimanite-rich foliation (Caso, 2023). Due to the local isoclinal geometry of the D3 folds, the S3 (black lines in Fig. 13) and S2 are parallel, striking NE–SW.

Litho-structural relationships with melt distribution in the lower crust

Analysing the geometrical relationships between different fabric elements distributed in diverse rock types is fundamental for understanding the generation and mobilisation of melt throughout the crust. Taking the VP as an example, each rock type in the lower crust exhibits diagnostic features and structures revealing information about different stages of the continental crust evolution and partial melting processes (i.e., melt production, subsequent differentiation in melt-rich and restitic portions, melt extraction and melt consumption; Kreigsman & Hensen, 1998; Waters, 2001; White & Powell, 2002; Kreigsman & Alvarez-Valero, 2010; Dyck et al.2020; Yakymchuk, 2020).

As mentioned above, basic granulite (Fig. 14a) and brown amphibolite boudins (Fig. 14b and c), enclosed within metapelites and preserving the earlier mineral assemblages and structures, show little evidence of melt production. It is possible that the small amount of melt, eventually produced during the D1 or D2 stages, has been completely extracted thanks to the synchronous deformation, leaving a highly restitic rock made by amphibole, clino- and orthopyroxene and minor plagioclase.

The majority of metapelites in the studied area correspond to migmatite gneiss, characterised by a relatively high modal abundance of leucosomes (> 30%). Here, the boundary between leucosomes and fine-grained restitic lenses is frequently marked by biotite and sillimanite-rich selvedges, which may suggest limited melt extraction from these rocks and significant retrogression (Sawyer, 2008). The rounded quartz and biotite inclusions within garnet (i.e., Bt I and Qz I; Fig. 14d) provide evidence for its growth during melt production. This is possible according to the reaction Bt + Sil + Qz = Grt + Kfs + melt (Fig. 14d; Kriegsman & Alvarez-Valero, 2010; White et al.2014). Relict and deformed Bt I and fibrous Sil I near or included in garnet porphyroblasts and separated by melt films (Fig. 14e) highlight the remnants of the biotite-dehydration melting reaction (Le Breton & Thompson, 1988). The same interpretation can be attributed to the few relict biotite and sillimanite grains re-oriented parallel to the S2. The extensive partial melting during D2 is also supported by the presence of myrmekites between quartz and plagioclase (Fig. 14f), diagnostic of melt production (e.g., Kreigsman & Alvarez-Valero, 2010). Felsic granulites made by garnet-rich domains, which mark the S2 foliation, contain biotite or sillimanite in very low modal abundance. The few remnants of melt-related microstructures are recognisable only at the microscale as peritectic inclusions-rich garnet and plagioclase, and thin melt films along grain boundaries. They likely preserve the granulite-facies peak assemblage (i.e., Pl + Grt ± Rt), which may be due to significant melt-loss that prevented pervasive retrogression to lower temperature assemblages (e.g., White & Powell 2002). According to this hypothesis, felsic granulite could represent melt extraction during the D2 (Voshage et al.1990; Harlov & Wirth, 2000; White & Powell 2002). On the other hand, it also could be possible these rocks derived from a granitic mica-poor protolith which may not have produced a considerable (< 10%) amount of melt (Kreigsman & Alvarez-Valero, 2010; Sawyer, 2010).

Folded leucosomes within syn-D3 fold hinges testify migration, during D3, of the syn-D2 melt towards the thickened hinges of these folds. Moreover, leucosomes are also localised within neck boudins. The transposition of the S2 foliation is mainly localised in the sillimanite-gneiss, which is characterised by a pervasive continuous foliation (i.e., the S3) marked by plurimillimetric prismatic sillimanite and minor biotite (Sil III and Bt III). This strong enrichment in sillimanite (Fig. 14g and h) that forms the pervasive foliation in sillimanite-gneiss could be produced in two ways: (i) destabilisation of white mica according to the reaction Ms + Qz = Sil + Kfs + melt or (ii) during retrogression in the presence of melt by back-crossing the reaction Bt + Sil + Qz + Pl = Grt + Crd + Kfs + melt (Waters, 1988; Waters, 2001; Kreigsman & Alvarez-Valero, 2010; White et al.2014). The S3 sillimanite-rich foliation wraps peritectic mineral as garnet (Manzotti & Zucali, 2013) and cordierite (Caso, 2023), likely represents the back-reaction with melt (Spear, 1994; Kriegsman & Alvarez-Valero, 2010) involving garnet and cordierite. The intergrowths between quartz and biotite marking S3 and wrapping cordierite or orthopyroxene (Fig. 14i) are also commonly interpreted as back-reaction products (Sawyer, 1999; Waters, 2001; Kreigsman & Alvarez-Valero, 2010). Therefore, S3 may represent a preferential pathway for both melt migration and its interaction with the peritectic minerals.

The quartz-feldspathic layers or lenses, found in black amphibolite, contain clinopyroxene and/or garnet porphyroblasts, which could represent peritectic minerals as the product of partial melting through fluid-absent incongruent hornblende breakdown. Melt production in metabasic rocks needs sufficiently high temperatures to take place (> 800 °C; Rushmer, 1991; Wyllie & Wolf, 1993; Kunz et al.2014) similar to those obtained through the Ti-in-amphibole thermometry. Since leucosomes of black amphibolite are parallel to S2 foliation, it can be inferred that melting has occurred during the D2 stage in amphibolites as well.

Thermal state during melt-production and migmatisation

Temperatures derived from different geothermometers differ depending on the methods used and vary across different localities. However, systematic variations exist among the different lithologies. Particularly evident is the temperature difference of ∼100 °C between metapelites (peak temperature ca. 760 °C) and amphibolites (peak temperature ca. 860 °C, locally up to 900 °C; Figs. 11 and 12). This marked difference may occur since biotite was no longer stable at the maximum reached temperature and therefore was not able to register peak temperatures like amphibole did. In fact, the biotite-dehydration melting (here testified by relicts Bt I) occurs at lower temperatures (Le Breton & Thompson, 1988) than amphibole-dehydration melting. Temperatures derived from biotite marking the S3 foliation growing together with sillimanite after garnet, cordierite or orthopyroxene that reacted with melt during cooling after peak (Bt III; Stüwe, 1997; Kreigsman & Hensen, 1998; Brown, 2002; Kreigsman & Alvarez Valero, 2010), are equal to those from Bt I and Bt II. Most of the temperatures derived from the garnet–biotite Fe–Mg exchange thermometer (Ferry & Spear, 1978) provide lower values (with a minimum of ∼560 °C). These values are related to late diffusional resetting after peak metamorphic conditions, and thus after D2/D3 stages (Pattison et al.2003 and references therein).

A temperature variation is visible among the different migmatite gneisses (i.e., Grt-, Crd- and Opx-bearing), with the highest temperatures calculated for Opx-bearing gneiss (up to ∼810 °C) and the lowest ones in Crd-bearing migmatites (750 °C; Fig. 15). However, these differences are relatively small, with values close to the error of each method, i.e., ±50 °C at > 700 °C (see Table 2 standard deviation and Henry et al.2005).

Partial Alpine retrogressions could also explain temperature variations in metapelites, where such retrogressions are localised along hectometre-wide Alpine-age deformation horizons (Fig. 2) of widespread greenschist-facies overprinting (Manzotti et al.2014; Dal Piaz et al.2016). These deformation zones could separate different domains within the VP that underwent different PT conditions, explaining the temperature variation between Crd- and Oxp-bearing migmatites. According to the present map of rock distribution and Alpine horizons (Figs. 2 and 13), the reviewed temperature differences do not seem to be controlled by the Alpine horizons. Indeed, similar differences are recorded at the Lac Mort area, north of one km-scale Alpine horizons and at the Oyace area (Fig. 13).

Tectono-metamorphic evolution of the Valpelline series

The VP preserves a complex polyphasic evolution spanning from the pre-Alpine to Alpine age, under amphibolite-/granulite- to greenschist-facies conditions. The oldest remnants of the earlier tectono-metamorphic stage (D1) have been detected in the basic granulite and amphibolite boudins within metapelites. An earlier subsolidus stage before partial melting occurring during the D1 or early D2 is preserved in garnet cores as relict biotite, sillimanite, plagioclase and rutile, included in porphyroblasts. Manzotti & Zucali (2013) provided thermobarometric estimates of this event at T = 700 ± 50 °C and P = 0.57 ± 0.1 GPa (MZ1 in Fig. 15). Different values have been obtained by Gardien et al. (1994), who fixed the PT condition of their oldest stage at higher pressure (G1 in Fig. 15) from kyanite relicts within garnet cores (P = 0.9–0.1 GPa).

The D2 stage during which partial melting occurs, is accompanied by garnet, cordierite or orthopyroxene growth in metapelites yielding different temperatures (considering the Ti-in-biotite thermometer). The temperature range of 650–780 °C obtained from cordierite-migmatites intersects the biotite-dehydration melting reaction (Le Breton & Thompson, 1988; Spear, 1994; Waters, 2001; White et al.2014). Pressures obtained using the Grt–Crd barometer (Bhattacharya, 1986) of 0.58 ± 0.01 GPa and fall within the P range defined by the peak assemblage of a similar cordierite-metapelite from the IVZ (Redler et al.2012) using thermodynamic modelling (orange field in Fig. 15). Temperatures obtained from orthopyroxene-bearing migmatites using the Ti-in-biotite thermometer, ranging between 700 and 830 °C, also fit with the peak assemblage and PT conditions detected in similar samples in the IVZ by Redler et al. (2012) (blue field in Fig. 15) but at higher pressure (P = 0.8 GPa) respect to the cordierite-bearing ones. Temperatures yielded by black amphibolite using the Ti-in-Amp thermometer of Liao et al. (2021) range between 800 and 930 °C. These values, ∼100 °C higher than metapelites agree with the destabilisation of hornblende at ∼T > 900 °C (dashed black line in Fig. 15), meaning that amphibolite recorded higher temperature conditions. Pressures obtained from black amphibolite for the D2 stage applying the Amp–Pl geobarometer (Molina et al.2015) range between ∼0.7 and 1.2 GPa (Fig. 15).

The thermal conditions during the D2 and D3 stages in metapelites are similar since the Ti-in-Bt thermometry applied on biotite marking both S2 and S3 foliations yielded similar T °C values. In fact, these values should be considered only as minimum temperatures (up to 800 °C) related to the D2 and D3 phases, i.e., slightly before biotite destabilisation due to incongruent melting (D2) or related to melt consumption and back reaction with peritectic minerals (D3) producing biotite. The T °C range of black amphibolites coincides with one of the D2 and D3 stages of Manzotti & Zucali (2013; MZ2 and MZ3 in Fig. 15) and the PT peak conditions of similar cordierite- and orthopyroxene-bearing metapelites from the IVZ from Redler et al. (2012). Taking this into account is possible to constrain the PT conditions for melt production and consumption. The T °C range is comprised between ∼800 and 900 °C and the P between ∼0.5 and 0.8 GPa. The field of orthopyroxene-bearing migmatites from the Val Strona (IVZ), determined by Redler et al. (2012, 2013), can be used as upper pressure limit.

The inferred metamorphic conditions and PT paths for the VP lower continental crust are compatible with an evolution typical to many high-grade terranes related to extensional geodynamic settings: prograde to peak metamorphic conditions reaching and overstepping the suprasolidus conditions, followed by nearly isobaric heating (Jiao et al.2023 and references therein). This could explain the close PT conditions and similar mineral assemblage stability during the D2 and D3 events. However, the detection of kyanite in garnets by Gardien et al. (1994) shifting the early path to higher pressure conditions followed by decompression heating may highlight a more complex evolution probably linked to Variscan orogenic crustal thickening during the Carboniferous (Carosi et al. 2020; Schulmann et al. 2022; Roda et al. 2023) and subsequent orogenic collapse and lithospheric extension during late Carboniferous-Permian (Lardeaux & Spalla, 1991; Marotta & Spalla, 2007; Schuster & Stuwe, 2008; Roda et al.2018; Ewing et al.2023).

Geodynamic implications for the Adria lower continental crust: comparison with the Ivrea-Verbano Zone lower crustal section

The IVZ (Southern Alpine Domain) and the IIDK (Austroalpine Domain), both belonging to the Adria microplate, have several similarities with the VP in terms of rock types, P–T conditions, and ages of metamorphism.

Metapelites, felsic granulite (usually named ‘stronalites’; Redler et al.2012) and black amphibolite with localised partial melting (Kunz et al.2014) occur both in the VP and IVZ. The bulk rock compositions of both VP and IVZ metapelites and metabasite are also similar (Fig. 8). However, an important difference with the IVZ lies in the distribution of the three migmatite types (i.e., Grt-, Crd- and Opx-bearing) and in the map-scale continuity of the isograds, with higher-grade conditions approaching the Insubric Line (i.e., from the Ms- to Opx isograds). In fact, the VP rock types crop out without a clear crustal zonation unlike the IVZ, where the different parageneses reflect a clear metamorphic zonation. Moreover, prograde muscovite occurring in IVZ migmatites is lacking in the VP ones. The only remnant of prograde dehydration melting reactions is the relict biotite included within peritectic phases or deformed biotite grains scattered within the rock matrix.

The thermal state described in the IVZ with similar methods used in the present work (i.e., the Zr-in-Rt; Ewing et al.2013) is comparable with the one described in the VP. However, in the IVZ, higher temperatures reaching 1000 °C have been locally detected close to the contact with the Mafic Complex (Barbooza & Bergantz, 2000; Luvizotto & Zack, 2009). Conversely, the VP never reached the same UHT conditions, suggesting that no mafic bodies were in place close by and large enough to produce such temperatures in the VP rocks. For this reason, the VP represents a good case study to investigate the regional Permian metamorphism without the effects of contact metamorphism.

The few available ages for the VP metapelites obtained through U-Pb geochronology on zircon and monazite yielded late Carboniferous and Permian ages (Zucali et al. 2011; Pesenti et al.2012; Kunz et al.2018), similar to those for the IVZ (Barbooza et al.1999; Peressini et al.2007; Ewing et al.,2013; Williams et al.,2022). In particular, Kunz et al. (2018) provided two groups of ages by dating zircons from VP migmatites of ∼290–280 Ma and ∼270 Ma. Permian ages are also found in the Arolla Series granitoids and orthogneiss (∼290 Ma; Bussy et al.1998; Manzotti et al. 2018). This geochronological link between the VP HT metamorphism and magmatism in the neighbouring Arolla Series could explain partial melting occurring in the VP lower crust and the emplacement of granitic intrusions in the Arolla Series upper crust (Kunz et al.2018).

This work provides a detailed lithological and structural overview of the VP, with a focus on its Permian tectono-metamorphic evolution. The obtained results are summarised as follows:

  • The VP represents a lower continental crust sector, belonging to the Dent-Blanche Nappe in the Western Alps that preserves pre-Alpine structures and mineral assemblages.

  • The pervasive distribution of migmatite-bearing structures in the VP rocks testifies that this unit was part of a km-scale migmatitic terrane involved in high-grade metamorphism during Permian lithospheric extension and asthenosphere rising, linked with abundant melt production, which migrated from its source to shallower crustal levels.

  • Migmatite gneiss shows three different mineral assemblages i.e., (i) Grt-bearing; (ii) Crd–Grt-bearing and (iii) Opx–Grt-bearing associated with slightly different PT conditions.

  • The VP shows rock types, PT conditions and ages of partial melting similar to IVZ, but lacks the presence of potential heat sources like the Mafic Complex occurring in the IVZ associated with locally higher peak-T.

  • PT conditions of the migmatisation in the VP are between 800–900 °C and 0.5–0.8 GPa.

To view supplementary material for this article, please visit https://doi.org/10.1017/S0016756824000037

We would like to thank Prof. Federico Farina and Dr. Marco Filippi for their suggestions on the first draft of the manuscript. Dr. Francesco Nosenzo and an anonymous reviewer are also thanked for their constructive reviews and suggestions. Many thanks also to Dr. Simon Schorn for his editorial work and useful suggestions. The results here presented have been developed in the frame of the MIUR Project ‘Dipartimenti di Eccellenza 2017—Le Geoscienze per la società: risorse e loro evoluzione (work package 3, tasks 3.3 and 3.4)’. The work was partly supported by the Italian Ministry for Universities and Research through the project ‘Dipartimenti di Eccellenza 2023-27’.

This research has been funded by ‘PSR2020_MRODA’.

The authors declare no conflict of interest.