Introduction

Development of collisional orogens implies low-pressure–high-temperature (LP–HT) metamorphism, crustal anatexis and generation of crustal-derived granitic magmas as late-stage features (e.g. Žák et al. 2011). The study of granitic magmatism is, of great importance to understand the evolution of orogenic belts, providing information regarding the interactions between magmatism and tectonics, as well as intracrustal heat and mass transfer processes.

The Variscan orogenic belt in Europe is a collisional orogen that developed during the complex collision of the Laurussia and Gondwana continents during the Devonian and Carboniferous periods (e.g. Nance et al. 2010; Kroner and Romer 2013). Along the European Variscan Belt, the initial continental collision and related crustal thickening precede the formation of metamorphic core complexes characterized by exhumed mid-crustal migmatites, LP–HT metamorphism and large volumes of granitic magmas (Schulmann et al. 2002, 2008; Žák et al. 2011; Villaros et al. 2018).

Large volumes of granites can also be found in the Iberian Variscan Belt, particularly in the Central Iberian Zone (CIZ), (Ferreira et al. 1987; Dias et al. 1998; Villaseca et al. 1998; Azevedo and Valle Aguado 2013; Valle Aguado et al. 2017; Dias da Silva et al. 2018), a zone which has been interpreted as a section of the Gondwana margin partially underlying a set of allochthonous tectonic slices of continental and oceanic affinities (e.g., Ribeiro et al. 2007; Arenas et al. 2016; Mateus et al. 2016). In the CIZ, granitic plutons were emplaced after the first Variscan compressive event and Barrovian metamorphism, being usually interpreted as the result of syn-orogenic collapse (Escuder-Viruete et al. 1994; Valle Aguado et al. 2005; Dias da Silva et al. 2017). However, some of these complexes are associated with strike-slip shear zones (e.g., Costa et al. 2014; Pereira et al. 2017) raising an interesting and unsolved question concerning the role of the shear movements in the exhumation of anatectic complexes and, consequently, in the timing of the generation and emplacement of the associated Variscan granites during the intracontinental collision stages.

This study focuses on the granites within the Figueira de Castelo Rodrigo–Lumbrales Anatectic Complex (FCR–LAC), constrained by a major strike-slip shear zone: the Juzbado–Penalva do Castelo Shear Zone (JPCSZ; Iglesias and Ribeiro 1981). The FCR–LAC has been the target of recent studies that partially constrained its evolution (Díez Fernández and Pereira 2016, 2017; Pereira et al. 2017; Alves Ribeiro et al. 2017). However, for the complete understanding of this complex, it is still necessary to characterize/quantify the emplacement conditions of these anatectic granitic bodies, namely through the determination of crystallization ages and exhumation rates. Indeed, to date, and besides some past geochronological studies using K–Ar and Rb–Sr methods (Macedo 1988; Ribeiro 2001), only two studies presented U–Pb ages for granites of this complex (Díez Fernández and Pereira 2017; Pereira et al. 2018) although 10 distinct intrusive facies have been described (Ribeiro 2001).

The main objective of this work is to present new U–Pb zircon and apatite ages for distinct FCR–LAC granite facies with the purpose of, for the first time, constrain their crystallization ages and cooling rates, and, therefore, their emplacement conditions within the framework of the Variscan Orogeny. This innovative study aims to bring a better understanding for the tectono-metamorphic evolution of the internal zone of the Iberian Variscides, as well as for the role of intracontinental first-order shear zones in the exhumation of deep settled rocks in collisional orogens worldwide.

Geological setting

The Figueira de Castelo Rodrigo–Lumbrales Anatectic Complex

The Gondwana–Laurussia continental collision sets the beginning of the Variscan orogeny, being responsible for the tectonometamorphic evolution of the Iberian Massif from the Upper Devonian to the late Carboniferous (e.g. Matte 1991; Ribeiro et al. 2007). The effects of continental collision in the Central Iberian Zone (CIZ), started at ca. 370–360 Ma and led to multistage deformation and metamorphic events (Dallmeyer et al. 1997; Martínez Catalán et al. 2014; Dias da Silva et al. 2017; Díez Fernández and Pereira 2017; Pastor-Galán et al. 2019). Three regional tectono-metamorphic events have been identified in the CIZ (e.g., Valle Aguado et al. 2017, and references within): (1) The D1 deformation phase, caused crustal thickening/shortening and Barrovian metamorphism between 365 and 340 Ma (Dallmeyer et al. 1997); (2) The local D2 extensional phase (340–320 Ma; e.g., Martínez Catalán et al. 2014; Gutiérrez-Alonso et al. 2018, and references within) produced flat-lying extensional detachments and related low-dipping foliation preserved in low- to high-temperature/low-pressure metamorphic rocks, including migmatites (see Rodrigues et al. 2013; Dias et al. 2016 and Pereira et al. 2017 for alternative interpretations); (3) the D3 phase favoured the reactivation of several first-order transcurrent shear zones, the emplacement of crustal-derived melts (320–295 Ma; Valle Aguado et al. 2017) and regional retrograde metamorphism, synchronous with regional uplift (e.g., Martínez Catalán et al. 2014).

The FCR–LAC, located in the Iberian Variscan Belt within the autochthonous terrains of the CIZ (Fig. 1), is an anatectic complex composed of migmatites (metatexites and diatexites) that are gradually transforming into S-type, two-mica granites with several distinct facies, that can be differentiated by grain size and by the relative abundance of micas (muscovite and biotite) (Fig. 2a, b). The different facies were named Iγ to Xγ, from oldest to youngest, and their distinction was based on deformation and field relationships (Ribeiro 2001). Some granitic plutons (e.g., IIy) reveal deformation structures concordant with the direction of the Juzbado–Penalva do Castelo shear zone supporting its syn-kinematic nature.

Fig. 1
figure 1

(modified from Silva and Ribeiro 2000)

a Location of the studied area in the Iberian Massif context (modified from Dias et al. 2016); b geological map of the Figueira de Castelo Rodrigo–Lumbrales Anatectic Complex

Full size image
Fig. 2
figure 2

a Medium-grained, two-mica granite (IXγ) intruded by a pegmatitic vein; b fine-grained, porphyritic granite (Xγ); c metatexite exhibiting stromatic texture; d diatexite including a restitic nodule

Full size image

Granitic plutons outcropping in the FCR–LAC have been previously dated yielding a K–Ar age of 319 ± 6 Ma (Macedo 1988) and an U–Pb age of 307.8 ± 3.1 Ma for the São Pedro-Vieiro granite (Díez Fernández and Pereira 2017), as well as an U–Pb age of 318.7 ± 4.8 Ma for the Mêda-Escalhão-Penedono granite (Pereira et al. 2018). The FCR–LAC extends to Spain where the Lumbrales granite, yielded a Rb–Sr whole-rock age of 300 ± 8 Ma (García Garzon and Locutura 1981) re-estimated in 311.2 ± 3.7 Ma using 40Ar/39Ar step-heating micas (Roda-Robles et al. 2009, 2018; Vieira 2010).

Associated with the S-type, two-mica granites occur metatexites and diatexites forming together an anatectic complex. The metatexites exhibit stromatic textures, and occasionally it is noticeable centimetric to millimetric layers of peritectic sillimanite associated with muscovite (Fig. 2c). At times, the metatexites are intersected by leucosomatic veins or pockets of granitic, pegmatitic or diatexitic material. Diatexites show restitic nodules (Fig. 2d), schlieren structures, and, occasionally, ptygmatic folding. When near local shear zones, the diatexites reveal a steep foliation parallel to the shear direction (E–W). Nebulitic textures in diatexites are also occasionally present.

The anatectic complex is 5–15-km wide, delimited by two sinistral, east–west to ENE–WSW trend, first-order shear zones (Fig. 1) that juxtapose the complex onto the low-grade metamorphic units of the Ediacaran–Cambrian Douro-Beiras Supergroup (Sousa and Sequeira 1993) and the Ordovician Armorican Quartzites (Sá et al. 2005; Gutiérrez-Alonso et al. 2007). These two sinistral shear zones that limit the FCR–LAC are the Huebra Shear Zone at the North, and the Juzbado–Penalva do Castelo Shear Zone to the South (Iglesias and Ribeiro 1981; Pereira et al. 2017). The JPCSZ extends for 200 km and its sinistral displacement took place at least during the D3 intracontinental collision stage (Iglesias and Ribeiro 1981; Villar Alonso et al. 2000; Pereira et al. 2017). The last activity of this strike-slip shear zone was dated at 309 ± 2.5 Ma (40Ar/39Ar on synkinematic white micas), in the eastern part of it (in Juzbado; Gutiérrez-Alonso et al. 2015). Recently, Valle Aguado et al. (2017) suggested that the JPCSZ movement in the western termination ceased during the emplacement of the late-tectonic Viseu batholith at ca. 299 Ma (U–Pb zircon ages), indicating diachronism in the shearing propagation.

The JPCSZ has been suggested to have played an important role for the exhumation of the FCR–LAC with an estimated minimum vertical displacement of 7–13 km and a horizontal displacement of 65–100 km by simple shear-dominated transpression during the Variscan D3 orogenic events (Pereira et al. 2017). Yet, it is still not clear when anatexis took place and what mechanisms controlled the genesis and final emplacement of the anatectic complex.

Sample description and analytical techniques

Sample description

Granite samples were collected within the FRC–LAC at the localities indicated in Fig. 1. The IIγ granite crops out between the migmatite belt and the JPCSZ, elongated accordingly with the shear-zone direction (WSW–ENE). This muscovite > biotite granite is fine-grained and reveals meso- and microscopic deformation. The mineral assemblage is defined by quartz + plagioclase + muscovite + biotite ± orthoclase ± chlorite ± zircon ± apatite ± opaque minerals. The IIIγ granite has a porphyroid texture (feldspar megacrysts) within a medium grain-size matrix. It is composed by quartz + plagioclase + microcline + biotite + muscovite ± orthoclase ± fibrolitic sillimanite ± zircon ± apatite ± opaque minerals. The Vγ granite is a biotite > muscovite facies and it has a coarse-grained texture. The mineral assemblage of this granite consists in quartz + plagioclase + potassic feldspar + biotite + muscovite ± chlorite ± fibrolitic sillimanite ± zircon ± apatite ± opaque minerals. The IXγ granite is the most representative facies in the study area. Its grain size ranges between fine- and coarse-grained. The mineral assemblage is represented by quartz + plagioclase + potassic feldspar + biotite + muscovite ± chlorite ± fibrolitic sillimanite ± zircon ± apatite ± rutile ± opaque minerals. The Xγ granite is a muscovite > biotite granite, essentially coarse-grained, at times exhibiting a porphyroid fine-grained texture. Its mineral assemblage is composed of quartz + plagioclase + microcline + muscovite + biotite ± orthoclase ± chlorite ± fibrolitic sillimanite ± zircon ± apatite ± rutile ± opaque minerals.

Analytical techniques

All samples were prepared for the different types of analytical procedures at the Mineral Separation Lab of GeoFCUL—Department of Geology of the University of Lisbon. Zircon and apatite crystals were picked from a 63- to 250-µm fraction, after heavy liquid and electromagnetic separation.

Grains, mounted in epoxy resin mounts, were attached to metallic stubs with thin copper strips, and coated with a 1-nm pulverized gold film. Zircon and apatite grains were observed using a ZEISS EVO10MA scanning electron microscope (SEM) at the University of Portsmouth (UoP). For backscatter electron (BSE) imaging, an accelerating voltage of 20 kV and 700 pA beam current was applied to reveal internal structures in the analyzed grains.

U–Pb isotopic analyses were performed using an ASI RESOlution 193 nm ArF excimer laser coupled to the ANALYTIK JENA Plasma Quant Elite quadrupole ICP-MS at UoP. The detailed instrumental setup and ablation conditions can be found in Supplementary Material. For zircon U–Pb dating, a beam spot size of 20 µm (cores) and 11 µm (smaller grains and rims) was preferred. Beam energy densities used ranged from 2.1 to 2.5 J cm−2, with a 2-Hz repetition rate. As for apatite, a beam spot size between 50 and 20 µm was used, with beam energy densities ranging from 2.8 to 3 J cm−2, and a 3-Hz repetition rate. For additional information on the analytical conditions used for the LA-ICP-MS analyses, see Supplementary Material 1.

Plešovice was used as a primary standard for zircon (337.13 ± 0.13 Ma; Sláma et al. 2007) whereas the Madagascar standard was used for apatite (473.5 ± 0.7 Ma; Thomson et al. 2012). 91500 (1062.4 ± 0.4 Ma; Wiedenbeck et al. 1995), and GJ1 (608.5 ± 0.4 Ma; Jackson et al. 2004) were used as zircon secondary standards, whereas for apatite were used McClure (523.51 ± 2.09 Ma; Schoene and Bowring 2006) and Xuxa (unpublished, ca. 572 Ma, provided by courtesy of C. Lana, Federal University of Ouro Preto) (see Supplementary Material 2). For all the above, reproducibility of the secondary standards was within 2%.

IOLITE 3.31 software package was used for data reduction. A sample-standard bracketing method was used to correct for both instrumental drift and elemental mass fractionation.

For zircon, Wetherill concordia and weighted mean 238U/206Pb ages were calculated using ISOPLOT/EX 4.1 (Ludwig 2003). From the youngest zircon population data, only grains that were 95–105% concordant were used to determine crystallization ages.

For apatite, the isotopic data was processed using VizualAge_UcomPbine DRS and measured 207Pb (Chew et al. 2014). This data reduction scheme allows for common-Pb (Pbcm) correction of the primary standard based on their known radiogenic and variable Pbcm compositions. This correction is then applied to the unknowns.

Apatite Tera–Wasserburg concordia ages were determined using ISOPLOT/EX 4.1 (Ludwig 2003).

Trace element analyses in zircon were performed using the same instrument setup as for U–Pb, and a standard-sample bracketing method to correct for instrumental drift. The following isotopes were analyzed: 29Si, 31P, 39K, 40Ca, 45Sc, 49Ti, 51V, 52Cr, 55Mn, 85Rb, 87Sr, 89Y, 90Zr, 93Nb, 95Mo, 118Sn, 121Sb, 133Cs, 137Ba, 139La, 140Ce, 141Pr, 146Nd, 147Sm, 153Eu, 157Gd, 159Tb, 163Dy, 165Ho, 166Er, 169Tm, 172Yb, 175Lu, 177Hf, 181Ta, 182W, 208Pb, 209Bi, 232Th and 238U.

A laser beam diameter of 40 µm for NIST612, 35 µm for secondary standards and 35–25 µm for unknowns was used, with beam energy densities ranging from 3.8 to 4.1 J cm−2, with a 4-Hz repetition rate (Supplementary Material 1). NIST612 was used as primary standard using concentrations by Jochum et al. (2011), whereas 91500 (Wiedenbeck et al. 1995) and GJ1 (Jackson et al. 2004) were used as secondary standards, as they are relatively homogenous in terms of trace element concentrations. Zr was used as an internal calibration standard, considering zircon stoichiometry (Zr = 43.1 wt%). Most analyzed elements are within 10% and 5% accuracy relative to secondary standards published values, and 49Ti is within 10% uncertainty of reported values (see Supplementary Material 2), excluding uncertainties, with a detection limit of 3 ppm.

Results

Zircon U–Pb ages

U–Pb zircon ages were obtained for five of the granite facies from the FCR–LAC defined by Ribeiro (2001) (Fig. 1; Table 1): São Pedro-Vieiro (IIγ), Ribeira de Massueime-Galegos (IIIγ), Châs-Amargo (Vγ), Mêda-Escalhão (IXγ) and Sta. Comba-Algodres granite (Xγ).

Table 1 LA-ICP-MS U–Pb zircon data of the Figueira de Castelo Rodrigo–Lumbrales anatectic complex granites
Full size table

Zircon external morphology has been widely used in petrogenetic studies, particularly those targeting granitoids (e.g., Pupin 1980; Barbarin 1988; Belousova et al. 2006; Köksal et al. 2008). Zircon morphology depends on the crystallization rate, fluid composition and on the temperature of the crystallization medium (Corfu et al. 2003). Pupin (1980) established a systematics for zircon using the relative development of the prismatic and pyramidal crystal forms, considering the development of prismatic faces mainly related with the temperature of crystallization and the pyramidal faces with chemical factors. These parameters inferred from a zircon population can be helpful to characterize the evolution of a magma system (Corfu et al. 2003). However, the morphology systematics proposed by Pupin (1980) has been questioned by several authors (e.g., Vavra 1990, 1993; Benisek and Finger 1993) which advocate that the zircon morphologies can only reflect the latest stages of growth. Sometimes, it is possible to distinguish igneous and metamorphic zircon from its morphology. In general, euhedral, concentric oscillatory zoning and euhedral, prismatic external morphology are evidences for igneous zircons. Zircon from a high-grade metamorphic environment can exhibit patchwork zoning and multifaceted, equant, tabular external morphology (e.g., Aleinikoff et al. 2006). However, at very high metamorphic grade, these morphological distinctions related with the origin (igneous vs. metamorphic) are not so clear, especially under anatectic conditions (Aleinikoff et al. 2006). Th/U ratios has also been used to distinguish metamorphic (< 0.1) and magmatic (> 0.1) zircons (e.g., Williams et al. 1996; Rubattto and Gebauer 2000), yet its use for such purpose is  debatable (see Discussion).

Considering all these mentioned features, the selection of zircon grains for dating was made taking into account the maximum number of characteristics that was possible to determine/observe.

São Pedro-Vieiro granite (IIγ)

Zircon morphology of the São Pedro-Vieiro granite varies among elongated prismatic and oval shape (Fig. 3a). Occasionally, zircon grains show the development of bipyramidal terminations (most of them 211), whereas others have one pyramid developed in one direction and the opposite edge is rounded. About the zircon prisms, most of them are 100. Regarding zircon internal morphology, oscillatory zoning is rare. All zircon crystals have narrow rims (2–18 µm), but most of them depicting an unzoned core and intermediate zones. At times, a few zircons appear fractured. The Th/U ratio ranges from 0.07 to 0.54.

Fig. 3
figure 3

Back-scattered electron images of representative zircon grains of the five studied granites: a São Pedro-Vieiro granite (IIγ); b Ribeira de Massueime-Galegos granite (IIIγ); c Chãs-Amargo granite (Vγ). d Mêda-Escalhão granite (IXγ); e Sta. Comba-Algodres granite (Xγ)

Full size image

From this granite, we analyzed 22 zircon grains for U–Pb. From 5 younger zircon cores (Table 1) yielding 300 ± 2.2 Ma (MSWD = 1.4) (Fig. 4a), we determined the crystallization age of the IIγ granite, significantly younger than the K–Ar crystallization age (319 ± 6 Ma) obtained by Macedo (1988).

Fig. 4
figure 4

Wetherill concordia diagrams showing the U–Pb crystallization ages for: a São Pedro-Vieiro granite (IIγ); b Ribeira de Massueime-Galegos granite (IIIγ); c Chãs-Amargo granite (Vγ); d Mêda-Escalhão granite (IXγ). e U–Pb crystallization ages of Sta. Comba-Algodres granite (Xγ) are represented by a weighted mean average diagram

Full size image

Recently, Díez Fernández and Pereira (2017) obtained a SHRIMP U–Pb zircon age of 307.8 ± 3.1 Ma (MSWD = 1.8) for the São Pedro-Vieiro granite, which is slightly older than our estimate and outside analytical uncertainty of our measurement.

Eleven zircon grains with concordant inherited cores yield dates ranging from Lower Devonian (400 Ma) to Paleoproterozoic (2000 Ma) (Table 2). From the inherited zircon cores, it was observed 4 younger overgrowths with Variscan age.

Table 2 LA-ICP-MS U–Pb inherited zircon data of the Figueira de Castelo Rodrigo–Lumbrales anatectic complex granites
Full size table

Ribeira de Massueime-Galegos granite (IIIγ)

In general, the Ribeira de Massueime-Galegos granite contains prismatic zircon, although a few crystals have acicular shape with oscillatory zoning and others have oval shapes with rounded terminations. When the oscillatory zoning is present, it is more visible in the thin zircon rims (5–18 µm) surrounding the unzoned cores and, quite rarely, the cores appear with convoluted zoning (Fig. 3b). This feature is typical of zircon growing during high-temperature metamorphism (Corfu et al. 2003). The majority of the zircon grains have 211 pyramids and 100 prisms, and sometimes they appear fractured. The Th/U ratios are low, ranging from 0.01 to 0.06.

The ablation was done in 22 grains of the IIIγ granite. The young concordant dates (Table 1) (1 core and three rims) allowed to estimate the 314.1 ± 2.6 Ma (MSWD = 0.12) crystallization age for this granite (Fig. 4b). Five old zircons provided Cambrian and Ediacaran ages around 485–600 Ma (Table 2).

Chãs-Amargo granite (Vγ)

Zircon grains from the Chãs-Amargo granite have prismatic and oval morphology, sometimes with rounded terminations and narrow rims (4–18 µm) (Fig. 3c). The zircon typology is heterogeneous, since the pyramids range between 211 and 101, and prisms between 110 and 100. The Th/U ratio varies between 0.005 and 0.24.

Thirty zircon grains were analyzed and it was determined a date of 316.2 ± 3.9 Ma (MSWD = 1.10) for the Vγ granite, using 5 younger zircons (Table 1) (4 cores and 1 rim), which is ascribed as its crystallization age (Fig. 4c). From the 30 grains, 16 of them correspond to older concordant dates with ages from Lower Devonian (400 Ma) to Paleoproterozoic (2500 Ma) (Table 2).

Mêda-Escalhão granite (IXγ)

The Mêda-Escalhão granite carries prismatic zircon grains, at times with planar or oval shape. Many crystals have the pyramids developed in one direction and occasionally, they appear fractured. In addition, they have thin rims and faint zoning in both cores and rims (3–26 µm) (Fig. 3d). Most of the zircon grains exhibit 211 pyramids and 100 prisms. The Th/U ratio fluctuates between very low values of 0.004 and 0.13.

The U–Pb dating was carried out on 23 grains. From 4 younger zircon grains (Table 1) (3 cores and 1 rim) was determined a date of 317.4 ± 2.1 Ma (MSWD = 0.097) that corresponds to the crystallization age of the IXγ granite (Fig. 4d).

Pereira et al. (2018) dated this granite facies (Mêda-Escalhão-Penedono massif) to the west of the FCR–LAC, estimating a LA-ICP-MS U–Pb zircon age of 318.7 ± 4.8 Ma (MSWD = 0.22) which is in agreement with the age obtained in this study, within error.

Five concordant old cores stand out, from the group of 23 grains, showing Cambrian and Cryogenian ages from 500 to 750 Ma (Table 2).

Sta. Comba-Algodres granite (Xγ)

Most of the zircon that compose the Sta. Comba-Algodres granite are prismatic and elongated, and a minority have acicular shape. The zircon rims (6–20 µm) are narrow and some of them have an incipient oscillatory zoning in the core (Fig. 3e). A large proportion of the zircon have 101 pyramids and 100 prisms. The Th/U ratio exhibits a large range of values, from 0.02 to 7.1.

Thirty-three zircons were analyzed, resulting in 26 grains with younger dates. From the 26 younger dates, 20 of them defines a cluster of absolute dates between 307 and 320 Ma (Table 1), similar within error between them, whose weighted mean average date is 312.9 ± 1.6 (MSWD = 1.6) (Fig. 4e) which is considered the crystallization age of the Xγ granite. Three concordant older zircon grains provide Neoproterozoic dates from 600 to 750 Ma (Table 2).

Inherited zircon

A general overview of the inherited zircon grains found in the granites (total of 40 concordant ages) allowed to ascribe them to different age groups: 400–500 Ma, 500–650 Ma, 650–850 Ma and 2000–2500 Ma (Table 2). These age distributions help us constraining possible protoliths as the source of the FCR–LAC granites.

Apatite U–Pb ages

In general, the apatite grains appear as elongated prisms, but occasionally show anhedral rounded shapes. Regarding the texture, the apatite grains do not show oscillatory zoning at backscattered electron imaging. However, the BSE imaging revealed several very small zircon inclusions (Fig. 5). Apatite ages analyses (Table 3) are plotted in the Tera–Wasserburg (TW) diagrams due to their high and variable common Pb (Pbcm)/radiogenic Pb ratios, precluding further Pbcm corrections. The TW regression results in the determination of the initial Pbcm composition and cooling age.

Fig. 5
figure 5

Back-scattered electron images of representative apatite grains of four studied granites: a São Pedro-Vieiro granite (IIγ); b Ribeira de Massueime-Galegos granite (IIIγ): c Mêda-Escalhão granite (IXγ); d Sta. Comba-Algodres granite (Xγ)

Full size image
Table 3 LA-ICP-MS U–Pb apatite data of the Figueira de Castelo Rodrigo–Lumbrales anatectic complex granites
Full size table

Regarding the number of analyzed grains, around 25 apatite grains per granite were targeted, and apatite ages were estimated from ca. 22 grains. All these grains fall on an isochron in the TW diagram supporting the idea that they are cogenetic despite their rounded morphologies.

For the five granitic facies mentioned above, the following TW Concordia lower-intercept apatite ages were obtained: IIγ (301.4 ± 2.6 Ma; MSWD = 0.8), IIIγ (288.0 ± 14.0 Ma; MSWD = 3.3), Vγ (306.6 ± 8.5 Ma; MSWD = 3.9), IXγ (307.0 ± 10.0 Ma; MSWD = 3.6), and Xγ (302.6 ± 5.6 Ma; MSWD = 1.9) (Fig. 6). The apatite ages reflect the timing at which the apatite closure temperature (450-550 ºC; Schoene and Bowring 2007) was reached after their crystallization and are, therefore, younger than the equivalent magmatic zircon ages.

Fig. 6
figure 6

Tera–Wasserburg U–Pb lower-intercept apatite ages for the five studied granites: a São Pedro-Vieiro granite (IIγ); b Ribeira de Massueime-Galegos granite (IIIγ); c Chãs-Amargo granite (Vγ); d Mêda-Escalhão granite (IXγ); e Sta. Comba-Algodres granite (Xγ)

Full size image

Discussion

Th/U zircon ratios in S-type granites

Th/U ratios have long been considered as an effective discriminator between metamorphic and magmatic zircon, with the value of 0.1 being a threshold below and above which, respectively, are placed those two types of zircons (e.g., Williams et al. 1996; Rubatto and Gebauer 2000). However, it has been demonstrated that Th/U ratios can not be used as a rule of thumb given that the Th/U ratio of zircon largely depends on the coexistence with Th-rich minerals such as monazite and allanite (Möller et al. 2003; Schaltegger and Davies 2017). For example, Rubatto (2017) showed that the occurrence of high-grade metamorphic rocks with zircons depicting Th/U > 0.1 is not rare, while Yakymchuk et al. (2018) referred a population of metamorphic zircons from Western Australia having a mean Th/U ratio of 0.4. On the other hand, Lopez-Sanchez et al. (2016) reported Th/U < 0.1 for zircon overgrowths of magmatic origin.

Considering all these facts, the reported Th/U ratios (Table 1) might be considered with caution. Indeed, due to the petrogenetic nature of these granites, with the melt being segregated from a HT metamorphic rock, it is expectable that magmatic zircon shows variable Th/U ratios. Under these circumstances, the determined zircon Th/U ratios can not be used as an independent discriminant factor for magmatic vs. metamorphic origin.

The timing of granites emplacement

The FCR–LAC granites have different facies allowing their subdivision in 10 distinct granitic bodies, referred as Iγ to Xγ, from oldest to young, based on the geometries of their contacts (Ribeiro 2001).

In this study, we obtained U–Pb zircon ages for the five most representative granite facies, (IIγ, IIIγ, Vγ, IXγ and Xγ). Many of our ages are identical within error, despite ages spanning from 300.0 ± 2.2 Ma (IIγ) to 317 ± 2.1 Ma (IXγ). In the light of these new results, our data indicate that the FCR–LAC granite suite crystallized between 317 and 313 Ma with the final emplacement of IIγ at ca. 300 Ma, supported by its intrusive field relationships with respect to the anatectic complex (Fig. 4).

It must be emphasized that the complexity of this granitic region can be higher than it can be inferred from the 10 mapped facies. Indeed, for the IIγ granite, besides the age here presented (300.0 ± 2.2 Ma), another age has been obtained, with similar number of zircons used to infer a concordia age, yielding 307.8 ± 3.1 Ma (Díez Fernández and Pereira 2017). This suggests that the IIγ granite can comprise more than one late intrusive body, but reinforces it being a late magmatic episode affecting this region. The crystallization age of the IIγ granite is substantially younger than the age obtained by Macedo (1988) of 319 ± 6 Ma using the K–Ar dating method. However, K–Ar ages are known to be prone to the effects of post-magmatic alteration processes leading commonly to younger ages by Ar loss (e.g. Baksi 1994), but also to older ages as a consequence of preferential K mobility (e.g., Cerling et al. 1985). Indeed, these authors demonstrated that low-temperature alteration involving meteoric water can result in hydrogen exchange by K+ and Na+ without significant alteration of other elements, conditions under which Ar also appears to be less mobile than alkali ions. Also, Mata et al. (2015) noticed that dates determined by the K–Ar method elsewhere for doleritic rocks portraying evidences for alteration, resulted in older dates than those obtained by more robust methods.

The obtained age for IIγ granite confirms the diachronous deformation along this shear zone, progressing from east (Juzbado) to west (Penalva do Castelo), as previously suggested by Pereira et al. (2017). Indeed, in Juzbado, the eastern sector of the JPCSZ, the last shear event has been dated at 309 ± 2.5 Ma (40Ar–39Ar in white micas; Gutiérrez-Alonso et al. 2015). Later, Valle Aguado et al. (2017) showed that in the western sector deformation continued until ca. 299 Ma, which is within error of the date, we obtained for the D3-affected IIγ granite (300.0 ± 2.2 Ma).

In the CIZ, significant granitic plutonism has been considered to occur during syn-, late- and post-D3 stage (e.g., Ferreira et al. 1987; Azevedo and Valle Aguado 2013). Particularly, granites from this studied region have been classified as syn-D3 which has been mainly developed between 310 and 320 Ma (e.g. Ribeiro 2001; Azevedo and Valle Aguado 2013). However, as mentioned before, IIγ granite (300.0 ± 2.2 Ma) was clearly deformed by D3 allowing to consider it as late-D3, while the other dated facies are syn-D3.

Protoliths of the granites

The obtained data for inherited zircon data were compared with published detrital zircon ages from distinct domains of the Douro-Beiras Supergroup in the CIZ, namely the metasediments from the Northern and Southern domains (Orejana et al. 2015). Comparison with the kernel density estimate (KDE) plots for these domains (Fig. 7) puts in evidence the similarities between metasediments of the Northern domain with our data, namely the existence of a main age group of the Lower Cambrian/Cryogenian, with minor Tonian and Paleoproterozoic contributions, and the absence of Mesoproterozoic ages.

Fig. 7
figure 7

Kernel density estimates (KDE) plots of inherited zircon ages for the studied granites (a) and also for the detrital zircons of the Northern (b) and Southern domains (c) of the CIZ (Orejana et al. 2015). The defined bin width is 50 Ma. The selected ages to perform the plots are 206Pb/238U, when the age is < 1000 Ma, and 207Pb/206Pb for ages > 1000 Ma. The KDE plots were performed with the support of the Vermeesch (2018) program (IsoplotR)

Full size image

Moreover, the second representative group of ages, 400–500 Ma (Upper Cambrian/Lower Devonian), seems to be related with the Ollo de Sapo formation that was emplaced in the Iberian Massif between 495 and 470 Ma (García-Arias et al. 2018). These ages suggest that the Douro-Beiras Supergroup metasediments that melted to form these granites also included Ollo de Sapo magmatic rocks.

Cooling and exhumation rates

In this work, we determined U–Pb ages for zircon and apatite occurring in the same rocks (see above), providing an opportunity to assess the cooling history of the granitic rocks where they crystallized. Indeed, closure temperatures (Tc) for the system U–Pb are usually considered to be in excess of 900 °C for zircon (Cherniak and Watson 2000), while for apatite Tc of 450–550 °C has been usually considered (Schoene and Bowring 2007). However, there have been reported apatite closure temperatures above 800 °C in doleritic rocks from the Armorican Massif (Pochon et al. 2016).

The concept of closure temperature was proposed by Dodson (1973) as referring to the temperature of a mineral at the time of its apparent (“freezing”) age. It can be determined from:

$$E/RT_{\text{c}} = \ln \left[ { - AD_{0} RT_{\text{c}}^{2} /\left\{ {a^{2} \left( {{\text{d}}T/{\text{d}}t} \right) \, E} \right\}} \right],$$

where E is the activation energy for the diffusion process, R is the ideal gas constant, Tc is the closure temperature, A is a numerical constant depending on the geometry of the grain (spherical or cylindrical), D0 is the diffusion coefficient at infinitely high temperature, dT/dt is the cooling rate and a is an effective diffusion dimension (i.e., radius in the case of a sphere). This approach has implicit the role of the daughter element volume diffusion over time, which is function of temperature. Moreover, the Tc is also dependent on the cooling rate. The Tc parameter appears on both sides of equation, so the equation is solved through several iterations for a given value of cooling rate. This procedure allows to calculate a consistent set of closure temperatures and cooling rates from the ages of two mineral species. For these calculations, the zircon and apatite diffusion parameters compiled by Hodges et al. (2003) were used.

Ti-in-zircon and the zircon saturation (TZircsat) geothermometers were determined in this study. The Ti-in-zircon geothermometer uses the Ti content in the zircon structure, which is dependent on temperature and independent of pressure, to estimate the magma temperature at the time of zircon crystallisation (Watson et al. 2006). The TZircsat predicts the temperature at which zircon crystallisation begins in a cooling magma, and, on the other hand, the temperatures above which zircon dissolution should occur (Watson and Harrison 1983).

In the case of zircon, we were faced with the problem that the experimental closure temperature (> 900 °C; Cherniak and Watson 2000) is clearly above of those determined by Ti-in-zircon and TZircsat geothermometers (Table 4), indicating that the studied zircon grains grew below their closure temperatures and, consequently, that diffusion processes did not occur after crystallization. Recently, Siégel et al. (2018) have shown that TZircsat is a dynamic variable that changes during magma crystallisation, and, thus, can not be used to constrain magmatic or partial melting temperatures. Consequently, we used the Ti-in-zircon geothermometer (Watson et al. 2006), which is considered as providing very reliable zircon crystallization temperatures, and therefore used as zircon closure temperatures. For the granite facies Vγ, IXγ and Xγ the use of this thermometer led to the following crystallization temperatures: Vγ = 825 ± 22 °C, IXγ = 836 ± 22 °C, and Xγ = 783 ± 31 °C (Table 4).

Table 4 Zircon geothermometers applied to the Vγ, IXγ and Xγ granites
Full size table

The obtained temperatures are endorsed by data published by Pereira et al. (2017) pointing to metamorphic temperatures in excess of 800 °C at the onset of partial melting. These results are also consistent with extensive literature on the melting of pelitic rocks to form anatectic granitoids (e.g. Clemens 2003; Bento dos Santos et al. 2011; Clemens and Stevens 2016).

The next step was to obtain the apatite closure temperature for each granite facies through Dodson’s equation, using Tc = 446 °C as an initial experimental value, as proposed by Cherniak et al. (1991). This provided Tc results between 483 and 465 °C (Table 5).

Table 5 Zircon and apatite estimated closure temperatures for the granites in study and the respective cooling and exhumation rates
Full size table

Combining the zircon and apatite Tc with their respective ages, the cooling rates for the IIIγ, Vγ, IXγ and Xγ granite facies are of 13, 34, 35 and 28 °C Ma−1, respectively (Table 5). It was not possible to infer the closure temperatures and respective cooling rates for the IIγ granite, because the apatite ages are slightly older than the zircon ages obtained for this sample, even considering their errors.

Cooling rates of plutons are highly variable, with published values differing by more than two orders of magnitude (e.g. Chesley et al. 1993; Tsuchiya and Fujino 2000; Meert et al. 2001; Miyazaki and Santosh 2005). It depends essentially on the way heat is exchanged between the hot intrusive body and their surroundings. The heat transfer may be done by conduction or by fluid-assisted advection. Considering that thermal diffusivity is strongly dependent on temperature, with which correlates negatively, and the role of the latent heat of crystallization on lowering the cooling rate of a crystallizing melt (Whittington et al. 2009; Nabelek et al. 2012), the rates of conductive cooling are very low when compared with advective cooling.

Although it was not possible to infer the cooling rate for the IIγ granite, the closeness of the zircon and apatite crystallization ages points to an abrupt cooling, at least until the apatite Tc. This suggests the localized intervention of significant fluid circulation during cooling within the JPCSZ. This hypothesis receives some support from the occurrence of abundant tourmaline in the diatexites and significant iron oxides enrichment in the metasedimentary rocks in contact with this granite facies.

The estimated granite cooling rates are similar for Vγ and IXγ, (34 ± 4.1 and 35 ± 3.1 °C Ma−1) and lower for IIIγ and Xγ granite concerning their absolute value (13 ± 16.5 and 28 ± 39.5 °C Ma−1). Nonetheless, these cooling rates are all identical within error. Comparing these cooling rates, particularly the Vγ and IXγ granites yielding small uncertainties, with high temperature metamorphic complexes (Spear and Parrish 1996; Bento dos Santos et al. 2010, 2014), allows us to infer fast cooling conditions. In geodynamic settings where granitoid cooling is mainly driven by denudation, cooling rates are substantially lower, typically spanning from 0.5 to 5 °C Ma−1 (Ashwal et al. 1999; Munhá et al. 2005; Gallien et al. 2010; Bento dos Santos et al. 2010, 2014; Scibiorski et al. 2015). Cooling rates of the FCR–LAC granites are, therefore, compatible with rapid exhumation mechanisms and shallow crustal emplacement.

Gravity has been considered an important constrainer during the late stages of the collisional orogens, when the tectonic driving orogenic forces diminishes or ceases its influence (Jadamec et al. 2007). This leads to lateral release of potential gravitational energy characterizing the thickened crust, leading to orogenic collapse (Rey et al. 2001 and references therein). The pressure release associated with such collapse can be viewed as one of the triggering mechanisms for crustal partial melting. Yet, mantle to crust heat transfer and intra-crustal radioactive heating might also be considered as significant mechanisms behind crustal melting (Vanderhaeghe 2009). Independent of the cause, partial melting triggers a significant strength/viscosity and density decrease, promoting the exhumation of migmatitic/granitic complexes, which tend to form domes, emplaced at shallow crustal levels (Vanderhaeghe 2009; Vanderhaeghe et al. 2018). This must be the case of the Tormes dome in the CIZ (Escuder-Viruete et al. 1994).

However, it is not yet clear how fast the exhumation caused by orogenic collapse can be (Vanderhaeghe and Teyssier 2001; Scibiorski et al. 2015). Moreover, the interpretation of the fast cooling rates reported on this study can not be done without taking into account that the anatectic complex is delimitated by high-angle crustal-scale shear zones, which juxtapose side-by-side rocks with clearly distinct metamorphic grades, a structural architecture also evidenced by magnetotelluric imaging (see Fig. 7 in Alves Ribeiro et al. 2017). Granites and the associated high-grade metamorphic rocks preserve a low-dipping non-horizontal transport lineation (6°–12°; Pereira et al. 2017), clearly indicative of a significant net vertical mass transfer when the 65–100 km lateral displacement is taken into account. Indeed, the associated migmatites of the FCR–LAC endured a significant tectonic exhumation, corresponding to a combined vertical displacement of 5–8 km (Pereira et al. 2017), which suggest that the granitic rocks within the complex must have endured the same tectonic uplift. Considering the upper Carboniferous geothermal gradient calculated by Pereira et al. (2017) for the FCR–LAC (42 °C km−1), and taking into account the estimated closure temperatures and ages for zircon and apatite, the vertical exhumation of granitic rocks would have been, indeed, of ~ 8 km, thus reinforcing the idea of a common assisted tectonic exhumation for granites and migmatites inside the JPCSZ. Considering these ~ 8 km vertical exhumation required for the determined cooling rates, exhumation rates of 0.3–0.8 mm a−1 are obtained. Such exhumation rates are clearly faster than those inferred for erosional denudation of granitic plutons (0.16 mm a−1; Yuguchi et al. 2017), but similar to granite exhumation rates in strike-slip shear zones (0.6–1 mm a−1; Steenken et al. 2002; Zhang et al. 2004; Annen et al. 2006).

In conclusion, our unprecedented results for cooling and exhumation rates of the FCR–LAC granites clearly support the role of first-order shear zones in assisting the exhumation of mid-crustal rocks as has been described elsewhere (e.g., Steenken et al. 2002; Corsini and Rolland 2009; Schulmann et al. 2008; Bento dos Santos et al. 2010, 2014; Fernández et al. 2013; Díaz-Azpiroz et al. 2014; Pereira et al. 2017), emphasizing the role of the JPCSZ in the emplacement of some Iberian collision-related Variscan granites. This mechanism should also be considered of utmost importance in intracrustal heat transfer, influencing the rheological behavior of the continental crust during and after collisional orogenesis.

Conclusions

The Variscan orogeny dynamics was responsible for the emplacement of several plutono-metamorphic complexes in the Iberian Massif and other sectors of the European Variscan Belt. The Figueira de Castelo Rodrigo–Lumbrales Anatectic Complex (FCR–LAC) is an example of a plutono-metamorphic complex where the granite–migmatite association is delimited by two sinistral, east–west to ENE–WSW trending, first-order shear zones (Juzbado–Penalva do Castelo and Huebra shear zones). New U–Pb zircon crystallization ages were determined for 5 different granite facies of this anatectic complex, yielding ages between 300 and 317 Ma, confirming that most of these granites formed during the syn-D3 magmatic stage, with the exception of one late-tectonic granite facies which is clearly intrusive into the other facies and yet is affected by the late stages of Variscan shearing. Moreover, since these granites are genetically related with migmatites and both formed and exhumed together, these syn-D3 crystallization ages (317 ± 2.1–313 ± 1.9 Ma) represent the maximum age (end of D2 and beginning of D3) and the duration of anatexis (ca. 5 Ma).

The inherited zircon population of these granites suggest that they are melting restites of units with Cadomian (650 to 550 Ma) and Upper Cambrian–Lower Ordovician (495 to 470 Ma) ages. In addition, these zircon age distributions reveal a protolith affinity with the Douro-Beiras Supergroup, as retrieved from the Spanish sector. It should also be noted a remarkable contribution of Upper Cambrian–Lower Ordovician ages, which point out to the contribution of metaigneous rocks of the Ollo de Sapo formation

Combining the U–Pb zircon ages with the U–Pb apatite ages and their respective closure temperatures, it was possible, for the first time, to quantitatively constrain the emplacement conditions of the FCR–LAC granites. Zircon (317 ± 2.1–313 ± 1.9 Ma) and apatite (307 ± 10–288 ± 14 Ma) enable the calculation of cooling rates ranging from 13 to 35 °C Ma−1. The closure temperatures of both geothermometers (zircon and apatite) allowed to estimate the emplacement of the studied granites at approximately 8 km of depth due to a fast exhumation mechanism. Such emplacement conditions are compatible with the transpressive shearing associated with the movement of the Juzbado–Penalva do Castelo shear zone, the most likely mechanism for the exhumation of the FCR–LAC (Pereira et al. 2017).

This novel application of zircon and apatite as petrochronometers, particularly in the CIZ where this approach was used for the first time, has proven to be useful in constraining the emplacement conditions (e.g., crystallization temperatures, cooling and exhumation rates) of syn-tectonic, S-type granitoids. This new type of approach is, therefore, important to understand the behavior of the continental crust during orogenic processes.