Introduction

Metapelites containing staurolite, garnet, andalusite, and sillimanite as part of the mineral assemblage have been a subject of debate concerning the evolution of these rocks, particularly regarding the metastability of aluminium silicate polymorphs (andalusite and sillimanite) and staurolite (e.g., García-Casco and Torres-Roldán 1996, 1999; Verdecchia et al. 2013; Mezger and Régnier 2016). Several pressure–temperature (P–T) trajectories can be proposed to explain the formation of these intricate mineral associations, including either a singular event or polymetamorphic evolution. In each of these instances, the reactive bulk composition undergoes progressive changes in the proportion of metastable and stable minerals along a P–T path (e.g., Lanari and Engi 2017). Part of this evolution could be preserved in garnet, whose crystallochemical characteristics resemble a 'black box' where some aspects of the prior metamorphic evolution may be retained (e.g., Caddick et al. 2010; Baxter et al. 2013, 2017; Lanari and Engi 2017; Yakymchuk et al. 2017; Giuntoli et al. 2018; Bovay et al. 2021; Verdecchia et al. 2022; Nerone et al. 2023). Thus, in samples of metapelites with compositionally zoned garnet, a step-by-step fractionation method (Evans 2004) is one of the most used strategies for P–T path reconstruction. However, other factors requiring evaluation for their potential impact in the progressive metamorphism include diffusion (Caddick et al. 2010), interface-coupled dissolution–reprecipitation (Ague and Axler 2016; Giuntoli et al. 2018; Verdecchia et al. 2022), and the degree of overstepping in reactions (Spear et al. 2014). These processes could explain discrepancies between the mineral assemblages and mineral chemistries predicted from phase equilibria modelling and what is observed. Additional factors are the reliability of activity-composition models predicting the chemical behaviour of minerals and uncertainties in thermodynamic databases (see Palin et al. 2016; Gervais and Trapy 2021).

In a portion of northwestern North Patagonia (Argentina), micaceous schists are mainly composed of staurolite, andalusite, garnet, sillimanite, biotite, quartz, and plagioclase, with an absence of cordierite. The schists are part of the Cushamen Complex (in the sense of Serra-Varela et al. 2022) and represent an opportunity to analyse a complex mineral assemblage previously associated with disequilibrium or polymetamorphic contexts (García-Casco and Torres-Roldán, 1996; Pattison et al. 1999; Tinkham et al. 2001; Gaidies et al. 2006; Jerabek et al. 2008; Verdecchia et al. 2013). In this study, we present a detailed petrological analysis of the micaceous schist unit of the Cushamen Complex, combining mineral chemistry, X-ray compositional maps, conventional thermobarometry, and phase equilibria analysis (conventional pseudosections and phase equilibria through Bingo-Antidote routine from XMapTool program development by Lanari et al. 2014a). The resulting pressure–temperature (P–T) path for this schist was constrained by testing the fractionation of different minerals at various stages of the P–T path (Marmo et al. 2002; Evans 2004; Lanari and Engi 2017).

The igneous-metamorphic basement of North Patagonia (North Patagonian Andes, Patagonian Precordillera and North Patagonian Massif) was involved in the Paleozoic evolution of the southwestern margin of Gondwana (see Rapela et al. 2021; Serra-Varela et al. 2022; Oriolo et al. 2023, between others). In this region, two main metamorphic complexes were distinguished: pre-Devonian metamorphic rocks associated with the Colohuincul Complex (Serra-Varela et al. 2019, 2022) and the Devonian-to-Carboniferous Cushamen Complex (Hervé et al. 2005, 2018; Marcos et al. 2020; Murra et al. 2022; Serra-Varela et al. 2022; Dicaro et al. 2023). Little is known about the structural and metamorphic evolution of either complex (Giacosa et al. 2004; García-Sansegundo et al. 2009; García et al. 2018; Serra-Varela et al. 2019; Rapela et al. 2024) and their P–T conditions (Martínez et al. 2012; Oriolo et al. 2019; Marcos et al. 2020).

Geological setting

The northern region of Patagonia is constituted by different morphostructural regions among which ancient basement rocks crop out in the North Patagonian Andes (NPA), the Patagonian Precordillera (PP), and the North Patagonian Massif (NPM) (see Giacosa et al. 2022; Fig. 1a). The schist unit studied here is part of the Patagonian Precordillera, where Paleozoic igneous and metamorphic rocks crop out discontinuously over several kilometres (Hervé et al. 2005; Pankhurst et al. 2006; Heredia et al. 2016; Marcos et al. 2020; Giacosa et al. 2022; Serra-Varela et al. 2022 and references therein; Fig. 1a). The igneous rocks were constrained by middle and late Paleozoic ages defining two different igneous belts (Pankhurst et al. 2006; Rapela et al. 2021; Serra-Varela et al. 2021; Oriolo et al. 2023). The older one corresponds to plutonic rocks ranging from gabbros to granites related to a Devonian-to-early Carboniferous NNW–SSE continental arc with ages ranging from 400 to 329 Ma (Varela et al. 2005; Pankhurst et al. 2006; Hervé et al. 2018; Rapela et al. 2021, 2024; Serra-Varela et al. 2021; Fig. 1a). Late Carboniferous–Permian plutonic rocks are found as part of a younger igneous belt with ages ranging from 318 to 253 Ma (von Gosen and Loske 2004; Varela et al. 2005; Pankhurst et al. 2006; Lagorio et al. 2021; Fig. 1a). Metamorphic rocks are found as scattered outcrops throughout the Patagonian Precordillera and the western North Patagonian Massif (Varela et al. 1991, 2005; Dalla Salda et al. 1994; Cerredo 1997; Marcos et al. 2020; Giacosa et al. 2022; González and Giacosa 2022) and were originally grouped together as the Cushamen Formation by Volkheimer (1964), recently redefined as the Cushamen Complex (Serra-Varela et al. 2022). The type locality near the town of Cushamen and along the Río Chico is represented by low-to-high-grade metasedimentary successions (Volkheimer 1964; Giacosa et al. 2004; Fig. 1b). Lithologies in the area are mostly phyllites, micaceous and quartz-micaceous schists, and sillimanite gneiss. Interbedded metaquarzite and orthogneiss layers are scarce as calc-silicates and amphibolitic bodies (see Giacosa et al. 2004).

Fig. 1
figure 1

a Simplified geological map modified from Serra-Varela et al. (2021) showing the distribution of the main Paleozoic igneous and metamorphic rocks in northwestern Patagonia. Ages according to: 1Pankhurst et al. (2006); 2Hervé et al. (2005); 3Hervé et al. (2018); 4Martínez et al. (2012); 5Lucassen et al. (2004); 6Serra-Varela et al. (2019); 7Marcos et al. (2020); 8Dicaro et al. (2023); 9Murra et al. (2022). Location of the North Patagonian Andes (NPA), Patagonian Precordillera (PP) and North Patagonian Massif (NPM) according to Giacosa et al. (2022). b Simplified geological map of the studied area modified from Giacosa et al. (2004) showing the main lithostratigraphic units. Geochronological data from 1Pankhurst et al. (2006) and 2Hervé et al. (2018). c Field photograph from the studied area showing the main lithologies. d Outcrop of the igneous intrusive unit showing in black dashed lines the tectonic foliation affecting this unit

Full size image

In general, sediments of the protoliths of the metasedimentary units of the complex would have been deposited between Silurian and Devonian time (ca. 440 and 369 Ma) based on U–Pb detrital zircon ages and 87Sr/86Sr isotopic ratios in carbonates (Pankhurst et al. 2006; Hervé et al. 2018; Marcos et al. 2020; Murra et al. 2022; Dicaro et al. 2023). Devonian metamorphic ages (360–380 Ma) have been calculated using U–Pb TIMS on titanite and U–Pb SHRIMP on metamorphic zircon rims (Lucassen et al. 2004; Pankhurst et al. 2006). Younger Carboniferous–Permian metamorphic ages of ca. 311 and 285 Ma were obtained from a U–Th–Pb CHIME age of monazite (Murra et al. 2022) and U–Pb LA–ICP–MS metamorphic rims in zircon (López de Luchi et al. 2022), respectively. In addition, the metasedimentary succession of the Cushamen Complex near the town of Cushamen was intruded by a tonalitic pluton dated at ca. 329 ± 4 Ma that was posteriorly deformed (Fig. 1b; U–Pb SHRIMP in zircon; Pankhurst et al. 2006).

Analytical method

Sampling and whole rock geochemical analysis

A representative sample of the andalusite–garnet–staurolite–sillimanite schist unit from the Cushamen region (MNP-28055, 42°08′32.5″ S–70°32′29.7″ W; see Fig. 1b) was selected for whole rock geochemical analysis of the major elements, mineral chemistry (spot analysis and X-ray elemental maps), thermometry calculation, and phase equilibrium analysis (P–T and T–X pseudosection constructions).

For whole-rock geochemistry, about 2 kg of material was collected. Powders for whole rock analyses were obtained by crushing in a tungsten mill at the laboratory LABGEO of Centro de Investigaciones en Ciencias de la Tierra (CICTERRA, CONICET-Universidad Nacional de Córdoba, Argentina). These were analysed by X-ray fluorescence at the Laboratorio del Instituto de Geología y Minería (Universidad Nacional de Jujuy, Argentina) on a Rigaku FX2000 spectrometer with Rh tube. The sample was fused with lithium tetraborate for major element analyses. Operating conditions were 50 kV and 45 mA.

Mineral chemistry analyses

In-situ chemical mineral analyses were conducted with a JEOL JXA-8230 microprobe (EPMA) instrument at the Laboratorio de Microscopía Electrónica y Análisis por Rayos X (LAMARX) of Universidad Nacional de Córdoba (Argentina), operated at 15 kV accelerating voltage and 10 nA beam current (on Faraday cup) in phyllosilicates, and 20 nA in nominally anhydrous minerals. Beam diameter was set from 2 to 10 μm depending on mineral grain size, with count time of 10 s on peak, and 5 s at each background position. To minimize migration of Na and K during analyses (e.g., Morgan and London 2005), these elements were analysed first in each analytical run, 5 s on peak and 2.5 s on background. The instrument was internally calibrated against synthetic and natural mineral reference materials: Si (forsterite, albite), K (orthoclase) Al (anorthite, albite), Fe (hematite), Mg (forsterite), Na (albite, anorthoclase), Mn (pyrolusite), Ti (ilmenite), Sr (celestite), Ba (baryte), Cr (chromite), Ca (anorthite, diopside, wollastonite), and P (libethenite). Raw element data were ZAF-corrected using the phi-rho-Z algorithm implemented in proprietary JEOL software. Corrected element contents were converted to oxides assuming stoichiometry; FeOT represents ‘total iron’. The concentration of Fe2O3 in garnet was stoichiometric calculated from measured FeOT using the procedure of Droop (1987). Mineral compositions are expressed in weight percent (wt%) of the main element oxides and are recast into numbers of cations in atoms per formula units (apfu) following recommendations by Papike (1987, 1988) and Deer et al. (2013), for muscovite and biotite (11 oxygen; Rieder et al. 1998), plagioclase and K-feldspar (8 oxygen; Smith and Brown 1988), and garnet (12 oxygen; Grew et al. 2013). In addition, X-ray maps of elements were performed using wavelength-dispersive spectrometry (WDS) in microprobe and processes using the free access software XMapTools 3.4 (Lanari et al. 2014a). The standardization converts X-ray intensity (collected photons per analysed element per pixel) into oxide weight percent concentrations using analysis obtained from microprobe (punctual analysis) as internal standards (see Lanari et al. 2014a). The use of compositional X-ray maps allows for the following: (1) accurately evaluating chemical variations within the garnet crystal and selecting the profile that best represents these variations and (2) reconstructing the reactive bulk composition of microdomains, which is utilized during the Bingo-Antidote routine of XMapTools (see below).

Empirical to semi-empirical thermometry

The thermometer based on the Ti content of biotite (Henry et al. 2005) was applied. This has a precision of ± 24 °C for temperatures between 480 and 600 °C. The studied sample meets the minimum criteria imposed by these authors for the use of this thermometer: a metaluminous rock with titanium saturation, evidenced by the presence of rutile and ilmenite, as well as graphite, which suggests a low oxygen fugacity.

In the chemical analyses of chlorite, empirical and semi-empirical thermometers were utilized, employing the spreadsheet developed by Verdecchia et al. (2019). This spreadsheet primarily incorporates thermometers designed for quartz–H2O-rich systems, and the results can be found in Table S1. For the semi-empirical calibrations proposed by Inoue et al. (2009) and Lanari et al. (2014b; Chl1 thermometer), it is necessary to fix the Fe3+ content in chlorite. To achieve this, the Fe3+ concentration was estimated using the comparative method introduced by Inoue et al. (2018). This method involves comparing the temperature outcomes obtained from a calibration, where all FeTotal are considered as FeO (Inoue et al. 2018) with another method, where the Fe3+ content is explicitly known (Inoue et al. 2009). We also employed the empirical thermometers developed by Cathelineau (1988) and Jowett (1991), which operate under the assumption that temperatures exhibit a direct correlation with the Al content within chlorite. This relationship is particularly applicable at pressures below 6 kbar (see Vidal et al. 2016). It is important to distinguish between the two thermometers: the calibration proposed by Cathelineau (1988) is based on the AlIV content, whereas that of Jowett (1991) considers the influence of the FeTotal/(FeTotal + Mg) ratio.

Phase equilibrium modelling

The P–T pseudosections were calculated in the simplified chemical system MnNCKFMASH (MnO–Na2O–CaO–K2O–FeO–MgO–Al2O3–SiO2–H2O) using Theriak-Domino software (de Capitani and Brown 1987; de Capitani and Petrakakis 2010) and the internally consistent thermodynamic data set DS5 of Holland and Powell (1998, updated 2004). The mineral solution models used for these calculations were: plagioclase (Holland and Powell 2003), garnet and biotite (White et al. 2005), white mica (Coggon and Holland 2002), chlorite, chloritoid, cordierite and staurolite (Holland and Powell 1998), orthopyroxene (White et al. 2002), and melt (White et al. 2007). Zoisite, albite, kyanite, sillimanite, andalusite, and H2O were considered as pure phases. The effect of Fe3+ was not considered, since low oxygen fugacity was suggested by the presence of graphite. A low presence of Fe3+ is also consistent with the presence of rutile and scarce ilmenite in the mineral assemblage (see Diener and Powell 2010; Forshaw and Pattison 2021). In a similar way, Ti was not included, since we considered that rutile is the main host of TiO2. The amount of H2O defined in the bulk composition was set to be in excess for modelling in subsolidus conditions. The amount of calcium was corrected due to the presence of abundant apatite in the sample; thus, Ca equivalent to 3.33 mol of P2O5 from the whole-rock composition was subtracted.

Results

Field relations and structural features of Cushamen Complex

In the study area, micaceous and quartz–micaceous schists, metaquartzites, and low-grade metamorphic rocks (phyllites and metapsammites) were identified. The alternation of levels of different composition (e.g., Al-rich vs Al-poor layers) represents preservation of the compositional bedding (S0) of the protoliths. Low-grade metamorphic rocks are exposed in the northeast region along Río Chico, whereas schists are found in most of the study area (Fig. 1b). Metamorphism increases from low- to mid-grade towards the southwest, crossing through the biotite–white mica, garnet, and sillimanite zones. In the western region of the study area (Fig. 1b), metasedimentary packages (ca. 4 km2) of micaceous and quartz–micaceous schists are exposed. In these packages, andalusite-–garnet–staurolite schist with centimetre andalusite (chiastolite) porphyroblasts was identified (Fig. 2a).

Fig. 2
figure 2

Field photograph and photomicrographs of the analysed sample. a Field photograph of the andalusite–garnet–staurolite–sillimanite schist. b Photomicrograph of the schist showing the relationship of the two foliations recognized in the sample (S1 and S2). c Sillimanite prisms in the rim of large andalusite crystals replaced by Ms3. d Large staurolite porphyroblast with sigmoidal inclusion pattern and the presence of garnet in its mantle zone. e Type-a garnet showing Grt1 in its core and Grt2. Red dashed lines showing the sigmoidal pattern of inclusions in mantle of Grt2. f Type-c garnet forming a cluster and staurolite cluster in pressure shadow of large andalusite crystals. Mineral abbreviations according to Whitney and Evans (2010)

Full size image

The metasedimentary succession exhibits two foliations: S1 and S2. In the micaceous schist, the most penetrative structure is the S2 foliation, whereas S1 is locally preserved and recognized in microlithons (see below). The S2 foliation strikes W–E to WNW–ESE (approximately N86°–110°) and dips mainly between 50° and 88° to both NE and SW (Fig. 1b). This is similar to the strike of the main structure recognized by Giacosa et al. (2004), but they referred to this as S1 foliation. Furthermore, centimetre-to-decimetre wavelength tight folds were identified affecting quartz–feldspar veins in the metasedimentary successions. These have axial planes concordant to the S2 foliation, with axes plunging mainly towards NW. Moreover, S2 was affected by kink-bands and open folds at the millimetre scale, without associated metamorphic mineral growth.

In the western region of the study area, the micaceous and quartz–micaceous schists are intruded by granitic bodies which are represented by small, foliated stocks, and dyke swarms (Fig. 1c). The dykes range from 1 to 10 m wide and show systematic NW–SE strike (Fig. 1c). Both types of intrusive bodies are composed of quartz, plagioclase, K-feldspar, biotite, and muscovite and are affected by a tectonic foliation (Fig. 1d) concordant to the S2 foliation in the schists. Contacts between the dykes and the metamorphic sequence are mainly sharp and parallel concordant to the main metamorphic foliation (Fig. 1b).

Mineralogy and microstructural features in the schist

This schist is composed of porphyroblasts of andalusite (chiastolite), staurolite, and garnet inside a matrix of sillimanite (prismatic and fibrolite), biotite, muscovite, quartz, and plagioclase. The accessory minerals are rutile, ilmenite, apatite, monazite, zircon, magnetite, tourmaline, and graphite. Kaolinite + fine white mica aggregates and chlorite were identified as alteration products of plagioclase (albitization) and garnet, respectively. The microfabric is mainly characterized by the development of S2 foliation, defined by aligned flakes of biotite (Bt2), muscovite (Ms2) and minor rutile in addition to fibrolite aggregates. S2 shows a spaced morphology defined by domains with different proportions of phyllosilicate (modal amount of ~ 40% to ~ 90%) relative to quartz and plagioclase, with transitional contacts (Fig. 2b). Locally, in deformation shadows around andalusite and staurolite poikiloblast, an early crenulated S1 foliation was identified (Fig. 2b), defined by aligned flakes of biotite (Bt1) and muscovite (Ms1) and scarce rutile (up to 1 mm of size). Andalusite poikiloblasts are up to 5 cm long and show small inclusions of graphite, biotite, ilmenite, and quartz. They are subhedral, prismatic, and preserve inclusion patterns with hourglass shape, typical of chiastolite. Some cores of andalusite crystals have a pleochroic pink colour, likely due to Fe-rich composition, but almost all crystals are colourless. The S2 foliation surrounds flattened andalusite prisms, suggesting a pre-tectonic origin of the andalusites. The andalusite is partially to almost completely replaced by coarse-grained flakes of muscovite (Ms3; Fig. 2c) and minor biotite (Bt3, concentrate in the rim of andalusite crystals), both up to 2 mm in size. Sillimanite prisms are found as inclusions in Ms3 in the mantle zone of andalusite pseudomorphs (Fig. 2c). The Ms3 is also seen cutting the S2 foliation in phyllosilicate-rich domains. Plagioclase up to 2 mm occurs in pressure shadows of andalusite porphyroblasts, together with biotite, fibrolite, and quartz.

Garnet, staurolite, and fibrolite are mostly found around, or in contact with, andalusite poikiloblasts (Fig. 2a), suggesting preferential nucleation on the edges of andalusite prisms. Staurolite and garnet were also observed, but in minor proportion, in the matrix relatively far from the andalusite prisms. Staurolite occurs as euhedral to subhedral poikiloblastic crystals up to 5 mm with slightly sigmoidal inclusion patterns continuous with the external S2 foliation (Fig. 2d). Oriented inclusions associated with Si (internal foliation) are composed of graphite, quartz (< 0.2 mm), rutile, and scarce fibrolite (sillimanite).

Garnet is present in low volume (~ 3%) and in three textural varieties. The first (a) is characterized by large subhedral-to-euhedral porphyroblasts (1.5–2 mm) with inclusions which are sometimes arranged zone in the form of a ring, leaving the inner core free of inclusions. In some cases, these inclusions show a straight pattern (Si) that progressively shifts to sigmoidal towards the mantle of the garnet. The outer part of the garnet mantle is frequently inclusion-free. The transition between a straight inclusion pattern and a sigmoidal one is interpreted as a limit between a Grt1 (core) and a Grt2 (mantle) (Fig. 2e), as evidenced by the chemical compositions (see Mineral Chemistry section). The inclusions in the core or surrounding rim include quartz, rutile, scarce ilmenite, graphite, and apatite. The sigmoidal pattern of inclusions (Si) is linked to development of the external S2 foliation, in a similar way to the staurolite porphyroblasts. The second garnet type (b) occurs as small (< 1 mm), isolated subhedral-to-anhedral crystals with few or no inclusions. This variety is dispersed in the matrix and crosscuts the S1 foliation in strain shadows around staurolite porphyroblasts (Fig. 2b). Symmetric-to-asymmetric strain shadows, composed of quartz ± biotite, are locally preserved in these garnets. The third type (c) is characterized by clusters or aggregates of small garnets (0.1–0.3 mm) with few inclusions, mostly located in the crystal cores (Fig. 2f). These clusters are found adjacent to others of small, similarly arranged staurolite crystals, as well as aggregates of fibrolite and minor proportion of biotite. These aluminous mineral assemblages are located in contact zones between andalusite and staurolite porphyroblasts. The staurolite and garnet clusters are characterized by an increase in the size of individual crystals from the core to the outer edge of the aggregates (Fig. 2f). Their inclusions include fibrolite, apatite and aligned crystals of ilmenite. In addition, there are scarce small inclusions (up to 150 µm) of staurolite and andalusite in the garnet crystals from this aggregate. Finally, a few garnet crystals (up to 500 µm) are recognized as inclusions in the mantle zone of staurolite porphyroblasts (Figs. 2d and 5b).

The quartz–plagioclase aggregates in the matrix show equigranular interlobate shape. Quartz lenses up to 8 mm were identified and can be interpreted as remnants of deformed and dismembered quartz veins. Anhedral magnetite aggregates are found in the matrix or as filling fractures in andalusite crystals. The magnetite is interpreted as post-S2.

Mineral chemistry

Garnet, staurolite, muscovite, biotite, plagioclase, chlorite, andalusite, and sillimanite were analysed to characterize the composition and the chemical variations within these minerals. In addition, chemical variations in different textural contexts were examined for some of these minerals. Representative analyses are listed in Supplementary Table S1 and the complete list can be consulted in Supplementary Table S2.

Garnet

Garnets from different textural contexts (types a, b, and c) were analysed through punctual analysis and X-ray compositional maps. In this way, a representative garnet porphyroblast from type-a is shown in Fig. 3a–d. Two compositional profiles were reconstructed showing the variation of spessartine (XSps), grossular (XGrs), pyrope (XPrp), and almandine (XAlm) ratios: profile-1 with punctual analysis measured in garnet and profile-2 from standardized X-ray maps (Fig. 3e, f). In these profiles, three zones are recognized in this garnet crystal: the core (Grt1), mantle (Grt2), and rim (Grt3) zones.

Fig. 3
figure 3

a–d Compositional X-ray maps for XAlm (a), XSps (b), XGrs (c), and XPrp (d), from representative type-a garnet, type-a garnet compositional profile from point analyses (e), and from X-ray map (f). Location of the analysed spots is shown in (a)

Full size image

The profile-1 shows a symmetrical normal bell-shaped zoning from core (Grt1) to rim (Grt3) for XSps, Mn/(Mn + Mg + Ca + FeTotal) but inverse zoning for XAlm, FeTotal/(Mn + Mg + Ca + FeTotal). Moreover, XGrs and XPrp, Ca/(Mn + Mg + Ca + FeTotal) and Mg/(Mn + Mg + Ca + FeTotal), respectively, present a more oscillatory and complex pattern. In this profile, no signs of garnet resorption in the rim are present. Grt1 shows a progressive increase of XGrs from core outward, whereas XPrp is mostly homogeneous. Thus, the composition of Grt1 varies from the core outward, from [XAlm 0.745–XSps 0.139–XPrp 0.080–XGrs 0.035] to [XAlm 0.747–XSps 0.122–XPrp 0.078–XGrs 0.053]. In Grt2, XGrs shows a decrease toward the rim, while XPrp increases, with values of [XAlm 0.750–XSps 0.112–XPrp 0.083–XGrs 0.056] in the inner zone to [XAlm 0.813–XSps 0.057–XPrp 0.096–XGrs 0.033] in the rim. Finally, the Grt3 is characterized by an increase in XGrs and a decrease in XPrp toward rim, with [XAlm 0.839–XSps 0.023–XPrp 0.092–XGrs 0.045] in the rim. On the other hand, profile-2 shows evidences of resorption of Grt1. It is characterized by asymmetrical bell-shaped zoning in XAlm and XSps profiles, where a compositional jump was identified at the boundary between Grt1 and Grt2 (Fig. 3f).

The type-b garnet shows a strain shadow composed of biotite and quartz (Fig. 4a). This crystal preserved a prograde compositional variation from core to rim, with a composition of [XAlm 0.821–XSps 0.047–XPrp 0.098–XGrs 0.033] in the core and [XAlm 0.840–XSps 0.021–XPrp 0.090–XGrs 0.050] in the rim (Fig. 4b–d). This composition is very similar to compositional progressive change between the inner zones to rim in the Grt3 from type-a garnet.

Fig. 4
figure 4

a Backscattered electron (BSE) image of type-b garnet. b–d X-ray maps and compositional profile from representative type-b garnet. Location of the analysed spots is shown in (a). e–h X-ray maps and compositional profile from type-c garnet from a cluster. Location of profile in (e)

Full size image

The composition of type-c garnet was defined from three analysed crystals characterized by a nearly homogeneous composition with mean values of [XAlm 0.842–XSps 0.023–XPrp 0.097–XGrs 0.038] (Fig. 4e–i). This composition is very similar to the edge of rim of Grt3 in a- and type-b garnets.

In staurolite, two inclusions of garnet were chemically characterized. The larger crystal (500 µm of size) has a core with composition of [XAlm 0.821–XSps 0.044–XPrp 0.100–XGrs 0.036] and rim of [XAlm 0.844–XSps 0.023–XPrp 0.094–XGrs 0.039]. This composition resembles that of Grt3 in b- and type-c garnets. On the other hand, a second smaller inclusion (200 µm) has a homogeneous composition of [XAlm 0.847–XSps 0.022–XPrp 0.093–XGrs 0.038], similar to such rim of Grt3 in type-c garnet.

Staurolite

In general, staurolite porphyroblasts have almost homogeneous composition along crystals, as can be seen in chemical profiles (Fig. 5 and Table S1). Commonly, it records XMg = Mg/(Mg + FeTotal) ratios of 0.14–0.16. However, slight variations are recognized such as a weakly decreasing trend of XMg from 0.15 to 0.14 that was observed from core to rim (Fig. 5). In the cluster of staurolite aggregates, fine-grained staurolite from the core of the cluster have an XMg of 0.13–0.14, whereas in the grains located outward from the cluster, XMg is 0.14–0.15. In all the analyses of staurolite, the ZnO content is low with values of 0.06–0.36 wt% without a clear pattern of variation between the core and rim of the crystals.

Fig. 5
figure 5

a–d Backscattered electron (BSE) image and profiles from different staurolite porphyroblasts showing the variation in XMg

Full size image

Biotite

Two biotite crystals included in andalusite cores (Bt1) have XMg = Mg/(Mg + Fe) ratios of 0.37–0.38 and TiO2 values of 1.49–1.51 wt% In the matrix, Bt2 is in cleavage or micaceous domains (composed of muscovite, biotite ± rutile) and in quartz–plagioclase–biotite domains. Bt3 is in biotite–muscovite aggregates linked to replacement of andalusite crystals. In the cleavage domains, Bt2 has 1.34–2.10 wt% TiO2 (mean of 1.81 wt%, n = 17) and XMg of 0.37–0.43 (mean of 0.41, n = 17), whereas in quartz–plagioclase–biotite domains, it has 1.72–1.95 wt% TiO2 (mean of 1.84 wt%, n = 6) and XMg of 0.40–0.42 (mean of 0.41, n = 6). In muscovite–biotite aggregates in the contact between andalusite and the matrix, as pseudomorphs of andalusite, Bt3 has 1.65–1.88 wt% TiO2 (mean of 1.76 wt%, n = 4) and a XMg of 0.40–0.41. Finally, two biotite crystals (Bt2) located in strain shadows of type-b garnet have 1.38–1.69 wt% of TiO2 and XMg of 0.41. These values suggest that biotite is mostly homogeneous in composition, but Bt2 and Bt3 have slightly higher values of XMg and TiO2.

Muscovite

Muscovite was identified as an inclusion in the core of andalusite porphyroblasts (Ms1) within the cleavage domains in the matrix (Ms2), and also in biotite–muscovite aggregates (Ms3, linked to replacement of andalusite crystals). One analysis of Ms1 records 3.03 apfu Si, 0.26 wt% TiO2, XMg = Mg/(Mg + Fe) ratio of 0.41 and Na/(Na + Ca + K) of 0.23. On the other hand, Ms2 in cleavage domain has 3.00–3.12 apfu Si (mean of 3.05 apfu, n = 19), 0.17–0.53 wt% TiO2 (mean of 0.44 wt%, n = 19), XMg of 0.36–0.49 (mean 0.42, n = 19) and Na/(Na + Ca + K) 0.14–0.20 (mean 0.18, n = 19). Ms3 in biotite–muscovite aggregates after andalusite, has 3.01–3.07 apfu Si (mean 3.04 apfu, n = 7), 0.43–0.58 wt% TiO2 (mean 0.51 wt%, n = 7), XMg of 0.38–0.48 (mean 0.42, n = 7) and Na/(Na + Ca + K) 0.16–0.19 (mean 0.18, n = 7). In similar way to biotite, the Ms1 has lower content of TiO2 (0.26 wt%) than Ms2 and Ms3.

Plagioclase

The plagioclase is classified as an oligoclase in composition. In the matrix (cleavage and quartzo-feldspathic domain), it has anorthite values of XAn = Ca/(Ca + Na + K) of 0.23–0.28 (mean 0.24, n = 12). Plagioclase included in andalusite cores has XAn of 0.22–0.27 (mean 0.24, n = 6). Locally, the plagioclase is replaced by a near albite composition concentrated in patches with XAn of 0.01–0.02.

Chlorite

Chlorite was identified as inclusions in the core of andalusite crystals and replacing garnet. In the first case, the composition is characterized by XMg = Mg/(Mg + Fe) of 0.39 and AlIV of 1.44 apfu, whereas chlorite after garnet has XMg 0.40–0.44 and of AlIV 1.44–1.54 apfu.

Andalusite–sillimanite

A porphyroblast of chiastolite (andalusite) was analysed the pink area in the core zone and in the white to grey mantle zone. The mantle has 0.19–0.37 wt% Fe2O3 (total) and < 0.06 wt% MgO, whereas the pink core has 1.24–1.37 wt% Fe2O3 and 0.32–0.36 wt% MgO. Sillimanite was analysed near the contact with an andalusite porphyroblast; it has 0.14–0.24 wt% Fe2O3 and < 0.04 wt% MgO.

Metamorphic evolution from microstructure and mineral chemistry evidence

Combining textural relations and chemical data it is possible to reconstruct the mineral growth history in the studied schist, necessary for conventional thermobarometry and phase equilibrium analysis. Three main mineral growth events or episodes are identified: M1, M2 and M3 (Fig. 6). Evidence for M1 is represented by biotite (Bt1)–muscovite (Ms1)–quartz–plagioclase. The subsequent M2 is defined by decussate andalusite together with garnet (Grt1), which are pre-tectonic with respect to S2 foliation (Fig. 6). The chemistry of Grt1 is characterized by a core with [XAlm 0.745–XSps 0.139–XPrp 0.080–XGrs 0.035] whereas this composition progressively changes outwards to [XAlm 0.747–XSps 0.122–XPrp 0.078–XGrs 0.053] near the metamorphic peak of M2. Finally, the main metamorphic event is M3, represented by garnet (Grt2-3)–staurolite–sillimanite–biotite (Bt2)–muscovite (Ms2)–plagioclase–quartz (Fig. 6), with partial resorption of Grt1 in the earliest stage, before Grt2 growth. The early stage of Grt2 is preserved in the inner part of mantle in type-a garnet [XAlm 0.750–XSps 0.112–XPrp 0.083–XGrs 0.056] that changes progressively outwards (represented by the core of type-b garnet and the outer zone of type-a mantle), with composition XAlm 0.815–0.821, XSps 0.047–0.055, XPrp 0.098–0.100 and XGrs 0.031–0.033. Finally, a later stage is represented by Grt3 with composition [XAlm 0.839–0.842, XSps 0.021–0.023, XPrp 0.090–0.097 and XGrs 0.038–0.050], identified in rims or edges of garnet crystals of all three types. Inclusions of type-b and type-c garnets are observed near the rim of the staurolite porphyroblast, indicating that staurolite formation is synchronous with Grt3 and the rims of Grt2 (Fig. 6). The slight decrease in XMg from the core (0.15–0.16) to the edge (0.13–0.14) of staurolite crystals may be associated with a weak but progressive change comparable to the composition shift from Grt2 to Grt3. The growth of large flakes of muscovite (Ms3) and minor biotite (Bt3) as pseudomorphs of andalusite, with muscovite crosscutting the S2, is likely related to a post-S2 stage within the M3 event. This interpretation is supported by the similar composition of Bt3 and Bt2, and Ms3 and Ms2, respectively.

Fig. 6
figure 6

Relationship between the foliations recognized in the sample and the mineral assemblage. Metamorphic events related to porphyroblast growth interpreted in this study are represented from M1 to M3

Full size image

Empirical to semi-empirical thermometry

To assess different P–T conditions for the stages of metamorphic evolution, empirical to semi-empirical thermometers were applied. First, biotite was analysed using Ti in biotite content based on Henry et al. (2005). For the M1 event, Bt1 included in the andalusite core suggests temperatures of 534–536 °C (± 24 °C, including the calibration error; see Table S1). On the other hand, biotite (Bt2 and Bt3) from M3 were analysed in different textural contexts. Bt2 in cleavage domains yields 577 ± 35 °C, in quartz–plagioclase–biotite domains 583 ± 26 °C, and in strain shadows around type-b garnet 538 ± 45 °C. Bt3 in biotite–muscovite aggregates after andalusite gives 571 ± 27 °C (Tables S1 and S2).

Empirical and semi-empirical thermometry were both applied in chlorite included in the core of andalusite crystals and chlorite after garnet. The empirical thermometers of Cathelineau (1988) and Jowett (1991) indicate values of 401 and 408 °C, respectively, for chlorite in andalusite, whereas semi-empirical thermometers of Inoue et al. (2009), Inoue et al. (2018) and Lanari et al. (2014b; equation with Fe3+ included) return results of 341, 319 and 431 °C respectively. In chlorite after garnet, thermometers of Cathelineau (1988) and Jowett (1991) given 402–434 °C and 410–440 °C, respectively, whereas semi-empirical thermometers yield 373–397 °C, 349–378 °C and 435–474 °C, respectively.

Phase equilibrium modelling

To reconstruct the P–T evolution of the schist unit, P–T pseudosections were constructed for subsolidus conditions ranging from 450 to 700 °C and from 1 to 10 kbar (Fig. 7). The initial bulk composition used for thermodynamic modelling was derived from the major-element whole-rock geochemistry of sample MNP-28055 (Table 1).

Fig. 7
figure 7

Pseudosections in the MnNCKFMASH system calculated for subsolidus conditions. Numbers in brackets represent the variance for each assemblage. a P–T diagram showing the upper limit of andalusite stability (blue) and the appearance of garnet (brown). b P–T diagram from (a) showing XGrs, XAlm, XSps, and XPrp ratios isopleths. P–T conditions obtained from isopleth composition convergence are indicated by black stars. Uncertainties ± 50 °C and ± 1 kbar (after Powell and Holland 2008). c T–XAnd diagram calculated at 3.5 kbar showing XGrs, XAlm, XSps, and XPrp isopleths for Grt2 mantle. d XGrs, XAlm, XSps, and XPrp ratios isopleths for the following stage of garnet (Grt1) growth (solid lines) and garnet (Grt2) rim (dashed lines) without considering fractionation of any mineral (see text for details). P–T pseudosections in the MnNCKFMASH system calculated for subsolidus conditions for Grt2 mantle (e) and rim (f). For both pseudosections fractionation of garnet was considered from point P2.1 for (e) and from point P2.2 for (f). The bulk compositions used for phase equilibria analysis are listed in Table 1

Full size image
Table 1 Bulk composition for thermodynamic modelling
Full size table

The P–T conditions of M1 could not be succesfully evaluated from the isopleth projection: XMg of Bt1, Si (apfu) of Ms1 and XAn of Pl, possibly due to chemical re-equilibration of the phyllosilicates in response to successive metamorphic events.

P–T conditions of the M2 event

Given the preservation of mineral associations linked with M1, it is likely that most of the rock underwent chemical equilibration during the M2 event. Consequently, the reactive bulk composition utilized for modeling the M2 event was the same as that employed for M1 (Table 1).

For this event, different P–T stages were constrained from composition of the core and rim of Grt1, in the presence of andalusite which is stable at pressures below 4.5 kbar (see Cesare et al. 2003). Almandine (XAlm), grossular (XGrs), spessartine (XSps), and pyrope (XPrp) compositional isopleths were projected. The composition of core of Grt1 [XAlm 0.745–XSps 0.139–XPrp 0.080–XGrs 0.035] is constrained at 563 °C and 3.3 kbar inside the P–T field where andalusite, garnet, staurolite, biotite, muscovite, plagioclase, and quartz are stable (point P1.1 in Fig. 7b). At this point, the predicted modal amount of Grt1 is less than 1%. The P–T conditions associated with the rim of Grt1 were not calculated as the composition was not preserved due to resorption before the growth of Grt2 (M3 event).

P–T conditions of the M3 event

Two approaches were used to calculate peak metamorphic P–T conditions of M3 event. First, these were estimated with AvPT multi-equilibrium method (Powell and Holland 1994), considering the mineral assemblage represented by garnet, staurolite, biotite, muscovite, quartz, plagioclase and sillimanite. The activity composition of solid solutions was calculated with AX2 software (Powell and Holland 1994), and P–T calculations were performed using THERMOCALC version 3.33 (Powell et al. 1998) and the DS55 data set (Holland and Powell 1998, updated in November 2003). The resulting P–T conditions were 650 ± 27 °C and 6.6 ± 1.1 kbar (cor = 0.906, sigfit = 0.86; excluded phases Eas, Pg, Ab, and Fe–Cel; Table 2).

Table 2 Results from AvPT routine showing the independent set of reactions calculated, the activities for each member and the codes of the mineral analyses used
Full size table

As an alternative, P–T pseudosections were constructed to decipher the P–T path from the compositions preserved in Grt2 (inner to outer zone) and Grt3. Textural evidence indicates that most of the andalusite associated with Grt1 was metastable during M3, so a T–XEBC pseudosection was constructed within the temperature range of 450 to 700 °C, maintaining a constant pressure of 3.5 kbar. Here, EBC represents the effective bulk composition in relation to the fractionation of andalusite. Thus, X0 represents the entire modal amount of andalusite as part of the reactive composition, whereas X1 indicates no participation of andalusite in the reactive bulk composition during the M3 event. The modal amount of andalusite was adjusted based on the P–T conditions obtained for the M2 event (563 °C and 3.3 kbar). On this pseudosection XAlm, XSps, XGrs, and XPrp isopleths for the start of Grt2 growth were projected [XAlm 0.750–XSps 0.112–XPrp 0.083–XGrs 0.056] (Fig. 7c). The variation in andalusite content has no effect on the composition of garnet, as indicated by the nearly parallel XAlm, XSps, XGrs and XPrp isopleths (Fig. 7c). The negligible impact of andalusite modal abundance, coupled with the very low volume of Grt1 predicted (core of garnet, 563 °C and 3.3 kbar), allows assumption of the same reactive bulk composition from both the M1 and M2 events for calculations relevant to the initial stage of the M3 event (see Table 1).

In the calculated P–T pseudosection, the initial stage of growth of Grt2 [XAlm 0.750–XSps 0.112–XPrp 0.083–XGrs 0.056] was constrained at 553 °C and 3.5 kbar (P2.1 in Fig. 7b), within a P–T field composed of plagioclase, muscovite, biotite, chlorite, staurolite, garnet, and quartz. On the other hand, the P–T conditions linked to the progressive chemical change of Grt2 toward its rim could not be successfully replicated using the same reactive bulk composition as for the beginning growth stage of Grt2. However, the rim composition of Grt2 [XAlm 0.813–XSps 0.057–XPrp 0.096–XGrs 0.033] was not predicted within the P–T space (Fig. 7d). This is likely due to compositional fractionation of garnet, as described in previous studies (e.g., Marmo et al. 2002; Evans 2004; Lanari and Engi 2017, and references therein). Thus, the early formed Grt2 should not be considered as part of the reactive bulk composition. The non-reactive garnet was fractionated from the bulk composition based on its composition and mode predicted at the P–T conditions of the initial growth stage of Grt2 (553 °C and 3.5 kbar). From this new reactive bulk, the P–T conditions for the rim composition of Grt2 [XAlm 0.813–XSps 0.057–XPrp 0.096–XGrs 0.033] are 630 °C and 4.7 kbar (P2.2 in Fig. 7e), within the stability field representing the mineral association of sillimanite, garnet, biotite, muscovite, plagioclase and quartz, but not staurolite (Fig. 7e). The absence of staurolite may suggest its metastability in conditions close to the thermal peak of M3 or may simply be attributed to uncertainties in the method (see Palin et al. 2016), as the obtained P–T conditions are very close (within ~ 15 °C) to the stability limit of staurolite (Fig. 7e).

Finally, the peak metamorphic conditions are related to the Grt3 composition [XAlm 0.839, XSps 0.023, XPrp 0.092, XGrs 0.045], which was recognized in all three garnet types. As in the approach taken to Grt2, fractionation of the early garnet was used to derive the reactive bulk composition. Thus, P–T conditions of 620–635 °C and 4.9–5.6 kbar were obtained from a small area resulting of the intersection of XAlm, XSps, XGrs and XPrp isopleths. These conditions are located within a stability field with staurolite, sillimanite, muscovite, biotite, plagioclase and quartz (P2.3 in Fig. 7f). The peak metamorphic conditions of M3 derived through phase equilibrium modelling are similar to those obtained using the multiequilibrium approach (650 ± 27 °C and 6.6 ± 1.1 kbar).

Discussion

The effect of bulk composition in mineral predicted composition

The selection of a reactive bulk composition for the calculation of P–T pseudosection diagrams plays a major role in the outcome of the analysis of a particular rock (Lanari and Engi 2017 and references therein). The growth of porphyroblasts can cause chemical fractionation of the reactive bulk composition, especially in the case of minerals with low diffusion rates at intermediate temperature subsolidus conditions, such as garnet (Stüwe 1997; Marmo et al. 2002; Evans 2004; Gaidies et al. 2006; Lanari and Engi 2017; Lanari et al. 2017). Lanari and Engi (2017) suggest that fractionation becomes significant when the garnet mode exceeds 2 vol%. In the studied schist unit, compositional fractionation resulting from andalusite, and garnet growth was evaluated to assess whether either or both phases had any influence on the metamorphic evolution of the rock. Andalusite was chosen because it occurs as centimetric porphyroblasts with a high modal content (~ 7%), while garnet was selected, because it exhibits compositional zoning linked to the P–T evolution of this rock. As argued above the fractionation of andalusite from the bulk composition does not play a major role in the predicted composition of garnet. In constrast, the pseudosections resulting from our studies in the schist indicate that garnet porphyroblasts should be considered in the fractionation of the bulk composition, despite the low volume of this mineral in the sample (~ 3%). Hence, integrating the compositional fractionation method for mineral phases with low modal quantities can be considered pertinent in modeling phase equilibria. As pointed out by other authors (Evans 2004; Lanari and Engi 2017; and references therein), this has relevance in the accurate reconstruction of the P–T trajectory of rocks exhibiting compositional zoning in refractory minerals and multistage mineral growth (e.g., of garnet), especially in polyphase or polymetamorphic contexts. Relying solely on interpretations or reconstructions of the P–T path based on a single reactive composition is not sufficient in these instances; such an approach could lead to significant discrepancies in the results due to the omission of fractionation considerations.

Although garnet composition has been reproduced for most of the P–T path for M3, the XMg composition of staurolite does not show the same behaviour. Textural and chemical relations suggest that staurolite located in aggregates or clusters is synchronous with Grt3. In these aggregates, fine-grained staurolite in the core of the clusters has a composition of XMg ranging from 0.13 to 0.14, whereas the outer grains record XMg values of 0.14 to 0.15. This is compatible with an increase in pressure (Fig. 7f), where an XMg value of 0.15 is constrained at a minimum pressure of approximately 6.4 kbar. However, the XMg values in staurolite at the P–T conditions constrained for Grt3 (620–635 °C and 4.9–5.6 kbar) range from 0.10 to 0.12, lower than those observed. This inconsistency was evaluated using Bingo-Antidote software (Lanari and Hermann 2021) in a selected and representative area of the MNP-28055 sample, where X-ray elemental maps were obtained. Bingo-Antidote uses an iterative optimization of P–T based on a local bulk composition calculated from the standardized X-ray maps (Duesterhoef and Lanari 2020). This approach was employed to investigate whether the bulk composition exerted any influence on the composition of the mineral phases, with a particular focus on staurolite. From this bulk composition, the Bingo-Antidote method calculates P–T pseudosections using Theriak-Domino software (Lanari and Duesterhoef 2019). The results obtained include the calculations of model quality factors comparing the predicted assemblage (Qasm), volume (Qvol), and mineral composition (Qcmp) with the corresponding observed parameters.

For the pseudosection calculation, the Bingo-Antidote was configured for a simplified MnNCKFMASH system, using the DS5 data set of Holland and Powell (1998, updated 2004). The optimal P–T conditions derived by minimizing Gibbs free-energy are 623 °C and 5.3 kbar, with a total quality parameter (Qtotal) of 92.4% (Fig. 8), practically the same as the ones calculated with pseudosections. Under these conditions, the predicted vs. measured quality parameters are high, where the parameter of mineral association (Qasm) is 100%, volume mode (Qvol) is 87.6% and the mineral composition (Qcmp) is 89.62%. Bingo-Antidote software has the function of calculating a model-quality factor (Qcmp) for each mineral. This factor compares the composition of a phase obtained from the X-ray compositional map with the composition calculated for a specific P–T point using phase equilibrium modelling. When analysing the P–T maps of the model-quality factor for staurolite (Fig. 8), Qcmp values match the calculated conditions 100%, with the observed composition being the same as the calculated composition. Thus, it can be deduced that the bulk composition has an important influence on the predicted composition of this phase.

Fig. 8
figure 8

Results from Bingo-Antidote add-on of XMapTools 3.4. a Mineral map of the selected area and model quality factors resulting from optimal P–T conditions calculated with the software. b Map of quality-factor Qcmp for staurolite. The white star reflects the optimal P–T conditions obtained with antidote routine c map of quality-factor Qcmp for garnet (Grt2). d Map of quality-factor Qcmp for biotite. e Map of quality-factor Qcmp for muscovite

Full size image

P–T path for the andalusite–garnet–staurolite–sillimanite bearing schist

The macro- and micro-structural features suggest three main episodes of mineral growth: M1, M2 and M3 (Fig. 6). The M1 event was synchronous with the D1 (S1) episode and resulted in inclusions in staurolite and garnet porphyroblasts or in strain shadows, mainly around andalusite crystals. M2 is represented by decussate growth of andalusite together with the core of type-a garnet (Grt1). Finally, M3 is linked to D2 (S2) and the development of a higher grade metamorphic mineral assemblage (peak conditions) composed of staurolite, garnet (Grt2; Grt3), biotite, sillimanite (fibrolite), plagioclase, muscovite and quartz. Syn-kinematic inclusion patterns corresponding to S2 are preserved in staurolite and garnet (Grt2). A final episode of mineral growth is represented by post-S2 genesis of decussated Ms3 and Bt3 around and replacing the rims of large andalusite porphyroblasts.

The combination of mineral association, textural and chemical information from the analysed schists, together with phase equilibrium modelling led to reconstruction of the P–T path of this rock. Four stages of crystal growth can be recognized from textural analyses combined with mineral chemistry. The first event (M1) is recorded as biotite or muscovite inclusions in large andalusite porphyroblasts and formed the relict S1 foliation. Temperature conditions for this event were estimated at 535 ± 24 °C using Ti in biotite thermometer (Henry et al. 2005) for Bt1, included in the core of andalusite (Fig. 9). Unfortunately, no other constraints could be used for the M1 event since we interpreted the Ms1 and some Bt1 as partially re-equilibrated during the M2 event. Therefore, the temperature of approximately 535 °C should be regarded as a maximum for the event. The second event (M2) is characterized by the appearance of a first stage of garnet growth (Grt1) together with andalusite porphyroblasts. The composition of Grt1 can be found in the cores of major garnet (type-a). The P–T conditions of this stage are 563 °C and 3.3 kbar (P1.1 in Fig. 9).

Fig. 9
figure 9

Summary of P–T conditions calculated for M2 and M3 metamorphic events. Andalusite, sillimanite, and kyanite fields are shown. The violet arrow represents the inferred P–T path for M2 metamorphism while the red arrow represents the P–T path for M3. XMg isopleths for staurolite composition are shown. P1.1, 2.1, 2.2 and 2.3 were calculated by phase equilibrium modelling (see Figs. 8 and 9) whereas PBA represents the result from Bingo-Antidote routine and AvPT the conditions were calculated with the AvPT multiequilibrium method. Results from Ti in biotite thermometer (Henry et al. 2005) are also displayed for Bt1 and Bt2

Full size image

The compositional shift in the transition from Grt1 to Grt2 is evident through changes in the chemical patterns observed in standardized X-ray element maps, particularly in XAlm and XSps ratios. This variation is associated with resorption of the early Grt1; together with the subsequent initial growth of Grt2 this marks the onset of the next metamorphic event (M3). The P–T conditions associated with the beginning of the M3 event were calculated as 553 °C and 3.5 kbar (P2.1 in Fig. 9), representing the early growth of Grt2. On the other hand, the progressive chemical variation of Grt2 from inner to outer zones of crystals (type-a and type-b garnets) indicate conditions of 630 °C and 4.7 kbar (P2.2 in Fig. 9). The last stage of garnet growth is represented by Grt3 (type-c and rims of type-a and type-b garnets), at around 620–635 °C and 4.9–5.6 kbar (P2.3 in Fig. 9), similar to results from the multiequilibrium approach (650º and 6.6 kbar; AvPT in Fig. 9). The peak of M3 is represented by a mineral assemblage of garnet (Grt3), sillimanite, staurolite, biotite (Bt2), muscovite (Ms2), plagioclase, and quartz. Ti in biotite thermometer (Henry et al. 2005) for Bt2 indicates temperatures of 577 ± 26 °C and 583 ± 10 °C lower than those calculated with P–T pseudosections. However, the differences in the results between these methods falls within the overlapping uncertainty of each (± 24 °C in Ti thermometry and ± 50 °C in pseudosections). Bingo-Antidote software from XMapTools (Lanari and Hermann 2021) was applied in a selected area of sample MNP-28055 to calculate the P–T conditions of the last stage of garnet growth (Grt3). This method utilizes standardized X-ray elemental maps to determine the equilibrium P–T conditions: 623 °C and 5.3 kbar were calculated for Grt3 growth. (PBA in Fig. 9), consistent with values obtained through the construction of pseudosections.

Finally, the retrograde stage is characterized by the growth of chlorite, replacing ferromagnesian minerals such as staurolite and biotite. The estimated temperatures in chlorite found within the andalusite cores (341–431 °C) are inconsistent with the estimated conditions obtained from the Ti content in biotite thermometer applied to Bt1 (~ 535 °C). This suggests that the chlorite results from post-Bt1 retrograde growth are not associated with the thermal peak of M1. The chlorites resulting from the alteration of garnet and staurolite record the retrograde stage of the M3 event, limited to conditions of 349–474 °C.

Regional implications: the metamorphic evolution of Cushamen Complex

The metasedimentary successions of the Cushamen Complex in the Patagonian Precordillera record a complex P–T evolution. For the study area, where the Cushamen Formation was originally defined (Volkheimer 1964), at least three metamorphic events were recognized: M1, M2, and M3. Comparing these metamorphic events with the ones observed by Marcos et al. (2020) north of the studied area (Comallo locality), the M2 event in Cushamen can be correlated with their Carboniferous M1 event. This event was associated with magmatism and deformation for which Marcos et al. (2020) obtained intermediate temperature (~ 500 °C) and low pressure (< 4.5 kbar) in Comallo. The M3 event in the study area can be correlated with main metamorphic M2 event preserved in Comallo in terms of P–T conditions.

Whereas the age and characteristics of M1 in the Cushamen area are unknown, the low-pressure mineral association and the calculated P–T conditions of M2 event point to an isobaric metamorphic event (Fig. 9). The emplacement of intrusive igneous rocks in the area could be responsible as the main heat source in low pressure conditions with the metamorphic development of andalusite porphyroblasts. The major plutonic body located in the southeast of the area has been dated at 329 ± 4 Ma (Fig. 1b; U–Pb in zircon; Pankhurst et al. 2006). Furthermore, the development of the S2 foliation in the schists coincides with the tectonic foliation in the igneous rocks, and both could be attributed to the same deformational event, related to the M3 event recorded in the schists. The M3 metamorphic event is post-mid-Carboniferous, constrained from igneous ages, most likely close to the Carboniferous–Permian boundary (311–300 Ma), as a regional metamorphic event is found in other North Patagonian localities with similar lithologies (e.g., Pankhurst et al. 2006; Heredia et al. 2016; Serra-Varela et al. 2022, and references therein). This regional metamorphic event has been associated with contrasting geodynamic settings involving crustal thickening (e.g., Pankhurst et al. 2006; Heredia et al. 2016; Serra-Varela et al. 2022; Oriolo et al. 2023).

Conclusions

The micaceous schist analysed in this study preserves an extensive part of the P–T evolution of the Cushamen Complex. Macrostructural, microstructural and mineral chemistry evidence from the micaceous schist suggest a polymetamorphic evolution represented by the M1 (D1–S1), M2 and M3 (D2–S2) events.

The M1 is less representative and preserved than the other events and is defined by biotite (Bt1), muscovite (Ms1), quartz and plagioclase. Oriented phyllosilicates and elongated quartz grain define a S1 foliation. Ti content in biotite thermometry suggests maximum temperatures of ~ 535 °C.

M2 is an essentially thermal event linked to local intrusion of mid Carboniferous plutons (ca. 329 Ma). It is represented by andalusite and garnet (Grt1). The near P–T peak conditions for M2 were constrained at ~ 3.3 kbar and ~ 563 °C.

The main M3 event is represented by the mineral assemblage garnet (Grt2–Grt3), staurolite, sillimanite, biotite (Bt2), muscovite (Ms2), plagioclase and quartz. A P–T path with increasing temperature and pressure was reconstructed from phase equilibrium modelling, with conditions from ~ 3.5 kbar and ~ 553 °C to ~ 4.9–5.6 kbar and ~ 620–635 °C. The M3 event was close in time to the Carboniferous–Permian limit, ca. 300–311 Ma, and the metamorphism is geotectonically related to a crustal thickening stage.

The selection of the reactive bulk composition is a critical step that significantly influences the calculations of pseudosection diagrams and the reconstruction of the P–T path, particularly in rocks with complex mineral assemblages like those encountered in the studied metasedimentary unit. This study underlines the significance of garnet composition fractionation, even when it occurs in very low modal quantities. Discrepancies between observed and predicted compositions and mineral assemblages in phase equilibrium modelling could be attributed to the definition of the reactive bulk composition. Consequently, the use of software tools such as Bingo-Antidote highlights the influence of bulk composition on predicted mineral compositions, accentuating the importance of adopting comprehensive approaches in P–T trajectory reconstructions.