Zircon U–Pb Dating and Lu–Hf Isotopic Composition of Some Granite Intrusions in Northern and Central Portugal: Constraints on the Emplacement Age and Nature of the Source Rocks

Abstract

Freixo de Numão (FNG) and Capinha (CG) granites are prominent intrusions in the Douro Group (northern Central Iberian Zone, CIZ) and Beiras Supergroup (southern CIZ) metasediments, respectively. U-Pb dating revealed crystallization ages of 306 ± 2 Ma for FNG and 301 ± 3 Ma for CG, whereas Lu–Hf systematics has shown ɛHf_t values ranging from −4.5 to +0.6 and from −5.5 to +0.3 in FNG and CG autocrysts, respectively, suggesting that they originate from heterogeneous crustal anatectic melts, but a direct mantle-derived material contribution can also be considered. The isotopic data of inherited zircon cores of both granites, with ɛHf_t values ranging from −16.8 to +8.4 in FNG, and from −19.4 to +10.1 in CG, are compatible with a derivation from heterogeneous Neoproterozoic metasedimentary sources, consisting of juvenile and recycled crustal materials, comparable to those of the wall rocks. However, the less evolved initial ¹⁷⁶Hf/¹⁷⁷Hf signature of magmatic zircons of both granites requires more immature metasediments/juvenile materials as main sources for the parental magmas. In fact, for FNG, the high Sr and Ba whole-rock content, and the upper Cambrian inheritance highlight the involvement of a metaigneous protolith in its genesis.

1. Introduction

The study of felsic magmatism, especially of granites, is crucial for better understanding the evolution and differentiation processes of the continental crust, e.g., [1,2]. Northern and central Portugal, being part of the Western European Variscan Belt, are characterized by the occurrence of a huge volume of granites, which are intrusive in the thick sequences of upper Precambrian and lower Palaeozoic metasedimentary rocks. The Variscan magmatism in the NW Iberian Peninsula comprises a wide range of granites in terms of mineralogy, texture, and deformation [3,4,5]. Most of them are two-mica (S-type) granites, but there are also I-type granites, and others that exhibit transitional signatures [4,6,7,8]. In general, these S-type granites are richer in primary muscovite than biotite, are strongly peraluminous (A/CNK ≥ 1.1, [9]) and were mostly generated by partial melting of Neoproterozoic–middle Cambrian heterogeneous metasediments and upper Cambrian–Lower Ordovician metaigneous sources [10,11,12,13]. I-type granites are biotite-rich, with very scarce or complete absence of primary muscovite, and of lower crustal origin [4,8].

Zircon (ZrSiO₄) is a common accessory mineral in nature, occurring in a wide variety of sedimentary, igneous, and metamorphic rocks. Its structure can incorporate a diversity of minor and trace elements, which enable its use in a wide range of geochemical research studies, namely those involving the U–Pb and Lu–Hf isotopic systems, e.g., [14,15,16]. Indeed, hafnium easily enters the zircon crystal lattice and due to its low Lu/Hf ratios, low solubility in silicic magmas and high resilience to deformation, hydrothermal alteration, and deep weathering processes, the initial ¹⁷⁶Hf/¹⁷⁷Hf isotopic ratios in this mineral tend to be preserved through time [17]. This fact has promoted the use of the Lu–Hf radiogenic isotopic system as an important petrogenetic tracer, as it can provide information on the age, magma nature and evolutionary processes of the source rocks, e.g., [18,19,20,21].

The Lu–Hf and Sm-Nd isotopic systems share similar characteristics. However, due to the greater incompatibility of Hf when compared to that of Lu and to the shorter half-life of ¹⁷⁶Lu than that of ¹⁴⁷Sm, the fractionation of Lu/Hf is twice as high as that of Sm/Nd during melting processes in the mantle [22]. Therefore, the Hf isotopes can provide a more precise identification of the mantle source domains, as well as an extraordinary record of continental crustal evolution [23,24,25,26]. The diffusion rate of Hf in the zircon crystal lattice is very low, allowing the preservation of zonation patterns caused by changes in the Hf isotope composition of the magma from which the zircon crystallized, e.g., [27]. Thus, the combination of U–Pb dating and Lu–Hf isotopic data can contribute to the knowledge of the precise emplacement time, the nature of the granitic magma source and its compositional evolution, and thus provide essential information for modeling crustal evolution.

The present paper presents new U–Pb LA-Q-ICP-MS geochronological and unprecedented Lu–Hf LA-MC-ICP-MS isotopic data of magmatic zircons and inherited zircon cores from the peraluminous Freixo de Numão (FNG) and Capinha (CG) granites, located near Vila Nova de Foz Côa and Fundão areas, in northern and central Portugal, respectively (Figure 1). These two circumscribed granitic suites make part of two major groups of late- to post-kinematic granites, which occur in distinct places inside the Central Iberian Zone intruding different metasedimentary sequences. The purpose of this research is to date the time of the emplacement of these intrusions during the Variscan orogeny, and to constrain the nature of the magma sources.

2. Geological Setting

The Variscan (or Hercynian) orogeny is part of a major geological mountain-forming event the Variscan-Alleghanian orogen [28], which occurred between 400 and 275 Ma ([29] and references therein). The Variscan-Alleghanian orogen stretches from the southeastern United States to eastern Europe and fragments of it are also detected in western and northwestern Africa [30]. The orogen has originally been produced by the collision between Gondwana, Laurentian and Baltic cratons, as well as the intervening microcontinents ([31] and references therein). However, this study will focus on the Variscan orogeny in the European terrains.

The Variscan orogeny was the major and most representative tectonic event of Western Europe and is currently explained by the oblique collision between Laurussia and Gondwana supercontinents, during the late Devonian and most of the Carboniferous, which involved the closure of the Rheic ocean and was followed by a post-thickening extension process, from the middle Carboniferous to the Permian [30,32,33,34,35,36,37,38,39,40,41,42].

The Iberian Variscan Belt is subdivided into six paleogeographic zones, each having specific tectonic, geologic, metamorphic, and stratigraphic characteristics [43,44,45] as follows: (i) Cantabrian Zone, (ii) West Asturian-Leonese Zone, (iii) Galicia-Trás-os-Montes Zone, (iv) Central Iberian Zone (CIZ), (v) Ossa-Morena Zone, and (vi) South Portuguese Zone. The Galicia-Trás-os-Montes Zone corresponds to allochthonous domains emplaced over the CIZ (autochthonous domain).

During the Variscan collision, three main ductile deformation stages (D₁, D₂, and D₃) and two extension periods (E₁ and E_2, pre- and post-D₃, respectively) have been recognized [46,47,48,49,50].

The early folds (D₁) were dated at 359–336 Ma and the emplacement of the allochthonous terrains occurred around 343–321 Ma (D₂) [35]. Subsequently, the thickened crust underwent extensional collapse (E₁, 321–315 Ma). D₃ folds were dated at 315–305 Ma [35,51]. A new extensional event E₂ (305–280 Ma [50]) occurred after the D₃ deformation period, with the development of a set of conjugate strike-slip faults (NNW–SSE, NNE–SSW and ENE–WSW) in higher crustal levels, under greenschist facies retrograde conditions [46,52,53,54,55]. However, in the lower crustal levels, the high temperatures could have persisted as the result of the high thermal gradients from the D₂ deformation period and the intrusion of early synkinematic granitoids [35,36,37,38,39,40,41,42,43,44,45,46,47,48,49].

Based on geological data and U–Pb emplacement ages, different stages of Variscan intrusive magmatism were considered, with respect to the D₃ deformation phase: pre-D₃ (331–315 Ma); syn-D₃ (315–310 Ma); late-D₃ (310–305 Ma), late to post-D₃ (ca. 300 Ma) and post-D₃ (<299 Ma) (Figure 1a) [48,56,57,58,59,60]. According to [61] and references therein, the two first stages correspond to syn-orogenic granitoids (older than 310 Ma) and the others to post-orogenic granitoids (310–285 Ma).

The Vila Nova de Foz Côa-Freixo de Numão region is located in the CIZ, northern Portugal [43] (Figure 1b) and is characterized by the occurrence of the Douro Group (northern CIZ of [62]) metasedimentary sequence, composed of flysch-like alternations of metagreywackes, slates and metaconglomerates, with carbonates levels at the base, belonging to the upper Ediacaran–middle Cambrian [63,64,65,66,67]. In this area, two regional Variscan deformation phases are recognized (D₁ and D₃; [46,47]). The D₁ was responsible for the most representative structures, originating WNW–ESE-trending folds with subvertical axial planes and a pervasive axial planar cleavage. It is worth mentioning that Lower Ordovician subvolcanic felsic porphyritic rocks can occur in dykes cutting the Douro Group metasediments, as is the case of Mateus-Vila Real area [68]. The Freixo de Numão pluton (FNP; Figure 1b) is intrusive into the metasediments, filling pre-existing late-Variscan shear zones and/or extensional fractures. FNG is zoned and composed of two distinct cogenetic granitic facies: (a) the Freixo de Numão granite (FNG, main granite facies)–biotite-rich, medium- to coarse-grained, porphyritic granite with rare metasedimentary and mafic microgranular enclaves, with no oriented fabric visible at macro- and meso-scale (Figure 2a–c); and (b) the Frei Tomé granite (FTG, less abundant)–two-mica, fine-grained granite with sulfides that occurs in small and NNW–SSE elongated outcrops within the FNG (Figure 2d). The FNP has a circular shape, discordant to the regional D₁–D₃ foliation (N60°W), where both granites have no oriented structure, and clear and sharp contacts with the metasediments [69]. A narrow thermal metamorphic aureole made up of biotite ± andalusite was developed around the intrusion, indicative of low-pressure and high-T conditions superimposed to a previous low-grade regional metamorphism (chlorite to biotite zones).

The Fundão-Capinha region lies within the CIZ, central Portugal, and is dominated by the Beiras Supergroup Neoproterozoic metasedimentary sequence (southern CIZ metasediments of [62]), consisting of a monotonous flysch-like succession of metagreywackes, slates and metaconglomerates, without carbonate intercalations [67,70]. In the studied area, this metasedimentary sequence displays a dominant NW–SE trending foliation (S₁) with vertical or subvertical dips (80°–85° SW) and voluminous intrusive granite batholiths. The Fundão-Capinha region is considerably affected by the major NNE–SSW Vilariça-Manteigas strike-slip fault system. The Capinha granite (CG, Figure 1c) has sharp and regular intrusive contacts at the east with metasediments from the Beiras Supergroup and at the west with the coarse-grained porphyritic biotite-rich Pêro Viseu-Seia granite. The CG intrusion displays a small circular shape, discordant to the regional foliation (N60°W). In contrast to FNP, the emplacement of CG did not cause thermal metamorphism in the metasedimentary rocks. The CG is a homogeneous two-mica, medium-grained, slightly porphyritic granite (Figure 2e), with no apparent structural deformation, suggesting that it was passively emplaced [70,71].

Figure 1. (a) Geological map of northern and central Portugal showing the distribution of Portuguese granites (adapted from [48]). (b) Simplified geological map of the Freixo de Numão Pluton and corresponding host rocks. (c) Simplified geological map of the Capinha granite (CG) and its host rocks.

Figure 1. (a) Geological map of northern and central Portugal showing the distribution of Portuguese granites (adapted from [48]). (b) Simplified geological map of the Freixo de Numão Pluton and corresponding host rocks. (c) Simplified geological map of the Capinha granite (CG) and its host rocks.

Figure 2. Macroscopic features of the Freixo de Numão, Frei Tomé and Capinha granites. (a) Yellowish biotite-rich medium to coarse-grained porphyritic granite of FNG. (b,c) Mica schist xenolith and mafic biotite-rich microgranular enclaves observed in FNG. (d) Yellowish two-mica, fine-grained, aphyric granite of Frei de Tomé. (e) Muscovite–biotite medium-grained slightly porphyritic granite cut by an aplite vein (contact of granite-aplite displayed in the circle) from CG.

Figure 2. Macroscopic features of the Freixo de Numão, Frei Tomé and Capinha granites. (a) Yellowish biotite-rich medium to coarse-grained porphyritic granite of FNG. (b,c) Mica schist xenolith and mafic biotite-rich microgranular enclaves observed in FNG. (d) Yellowish two-mica, fine-grained, aphyric granite of Frei de Tomé. (e) Muscovite–biotite medium-grained slightly porphyritic granite cut by an aplite vein (contact of granite-aplite displayed in the circle) from CG.

3. Petrography and Whole-Rock Geochemistry

A summary of the mineralogical and whole-rock compositions of the studied granite rocks is shown in Table 1. Full data, analytical methods, and geochemical characteristics of FNG and CG can be found in [71,72,73].

3.1. Petrography and Microstructures

The FNG yields a hypidiomorphic inequigranular medium- to coarse-grained porphyritic texture and a strong brittle deformation. It contains quartz, microcline, plagioclase (albite–oligoclase) and biotite > muscovite. The accessory mineralogy is composed of apatite, zircon, monazite, xenotime, titanite, ilmenite, pyrite, chalcopyrite, sphalerite, bismuthinite, stannite, tantalite-(Fe), cassiterite, ilmenite, rutile and other sulfides composed of a complex mixture of Sn-Cu-Fe [72]. In FNG, the microstructural studies allowed verification that the most common microstructures observed belong to both magmatic and low-T solid-state domains [72,74]. Their spatial distribution is very homogeneous in the whole studied igneous body (for detailed information see Table S1).

The FTG is essentially composed of scarce anhedral haematite surrounded by a dominant quartz ± sericite matrix. Very tiny anhedral crystals of xenotime (<10 μm) occur as inclusions in the haematite.

The CG displays a typical hypidiomorphic inequigranular medium-grained, incipient porphyritic texture and contains quartz, K-feldspar (microcline >> orthoclase), plagioclase (albite–oligoclase), and muscovite > biotite. Apatite, rutile, zircon, monazite and opaque minerals are accessory minerals. The opaque mineralogy is dominated by ubiquitous ilmenite and subordinate amounts of pyrite, arsenopyrite, and hematite [71].

In both granites, there are local alteration effects such as chloritization of biotite and sericitization of plagioclase. The microstructural study in the CG displayed the predominance of magmatic to submagmatic microstructures; however, a continuous transition was observed from magmatic to submagmatic microstructures in the core of the suite to low-T solid-state microstructures in the peripheries [71,74] (for detailed information see Table S1 [75,76,77,78,79]).

3.2. Geochemical Characterization

According to the classification criteria proposed by [80], these granite rocks are magnesian, with the exception of FTG that is ferroan (Figure S1). They mainly belong to the alkali-calcic series, but FNG also has a calc-alkalic signature (Figure S1 [9,80]). The Shand’s molar A/CNK = Al₂O₃/(CaO + Na₂O + K₂O) ratio index ranges from ~1.12 to ~1.45, and shows that all granites are peraluminous (Figure S1). In the Rb-Sr-Ba ternary diagram [81] (Figure S2), CG and FTG correspond to granites more evolved than FNG, which in turn plots within the field of anomalous granite due to its enrichment in Sr contents. Regarding Sn and W contents, CG has 9–24 ppm Sn and 6–12 ppm W, FNG has 5–18 ppm Sn and 1.8–4.2 ppm W, and FTG has 6 ppm Sn and 3.5 ppm W. These Sn and W contents are very similar to those found by [82,83] for unaltered tin-bearing granites.

The FNG shows moderate REE contents (ΣREE~175 to 230 ppm), highly fractionated chondrite-normalized REE patterns, with (La/Lu)_N ranging from 25.69 to 30.29, and Eu negative anomalies with (Eu/Eu)_N varying between 0.42 and 0.57 (Figure S2 [81,84]). The REE pattern of FTG is subparallel to those of FNG, but it is poorer in all REE (38.49 ppm) and has a higher Eu negative anomaly [(Eu/Eu)_N = 0.38] (Figure S2) [72]. The CG is enriched in ΣLREE (61.86–69.41 ppm) when compared to ΣHREE (7.85–9.6 ppm), showing moderately fractionated chondrite-normalized REE patterns, with (La/Lu)_N varying from 10.16 to 13.31, and Eu negative anomalies with (Eu/Eu)_N ranging between 0.23 to 0.28 [71] (Figure S2).

Table 1. General characteristics of the Freixo de Numão (FNG), Frei Tomé (FTG) and Capinha (CG) granites [71,72].

Granite	Rock Variety	Enclaves	Mineral Association	Geochemical Classification
Freixo de Numão (FNG)	Biotite-rich, medium- to coarse-grained, porphyritic	Metasedimentary and mafic microgranular	Qz ± Kfs (essentially Mc) ± Pl (Oligoclase-Andesine and Ab) ± Bt >> Ms I ± Ap ± Zrn ± Mnz ± Rt ± Ttn ± opaques ± Chl ± Ser ± Ep ± Cal	Peraluminous (A/CNK 1~1.18) Alkalic-calcic/Calc-alkalic Magnesian Syenogranite (R1-R2 diagram 2) P2O5 = 0.39–0.48 wt. %
Frei Tomé (FTG)	Two-mica (muscovite > biotite), fine-grained granite with sulfides	Absent	Qz ± Kfs ± Pl ± Ms I > Bt ± opaques ± Xtm ± Ap	Peraluminous (A/CNK 1~1.23) Alkalic-calcic Ferroan Alkali granite (R1-R2 diagram 2) P2O5 = 0.23 wt. %
Capinha (CG)	Two-mica (muscovite > biotite), medium-grained, incipient porphyritic granite	Absent	Qz, Kfs (Mc >> Orthoclase), Pl (Ab-Oligoclase), ± Ms I > Bt ± Ap ± Zrn ± Rt ± opaques ± Chl ± Ser ± Kln	Peraluminous (A/CNK 1~1.35) Calc-alkalic to alkali-calcic Magnesian Syenogranite (R1-R2 diagram 2) P2O5 = 0.33–0.38 wt. %

¹ molar A/CNK = Al₂O₃/(CaO + Na₂O + K₂O); ² R₁ = 4Si − 11(Na + K) − 2(Fe + Ti); R₂ = 6Ca + Mg + Al [85]; Key: Qz, quartz; Kfs, K-feldspar; Mc, microcline; Pl, plagioclase; Ab, albite; Bt, biotite; Ms, muscovite; Ap, apatite; Zrn, zircon; Mnz, monazite; Xtm, xenotime; Rt, rutile; Ttn, titanite; Chl, chlorite; Ser, sericite; Ep, epidote; Cal, calcite; Kln, kaolinite; A/CNK, aluminium saturation [9].

Table 1. General characteristics of the Freixo de Numão (FNG), Frei Tomé (FTG) and Capinha (CG) granites [71,72].

Granite	Rock Variety	Enclaves	Mineral Association	Geochemical Classification
Freixo de Numão (FNG)	Biotite-rich, medium- to coarse-grained, porphyritic	Metasedimentary and mafic microgranular	Qz ± Kfs (essentially Mc) ± Pl (Oligoclase-Andesine and Ab) ± Bt >> Ms I ± Ap ± Zrn ± Mnz ± Rt ± Ttn ± opaques ± Chl ± Ser ± Ep ± Cal	Peraluminous (A/CNK 1~1.18) Alkalic-calcic/Calc-alkalic Magnesian Syenogranite (R1-R2 diagram 2) P2O5 = 0.39–0.48 wt. %
Frei Tomé (FTG)	Two-mica (muscovite > biotite), fine-grained granite with sulfides	Absent	Qz ± Kfs ± Pl ± Ms I > Bt ± opaques ± Xtm ± Ap	Peraluminous (A/CNK 1~1.23) Alkalic-calcic Ferroan Alkali granite (R1-R2 diagram 2) P2O5 = 0.23 wt. %
Capinha (CG)	Two-mica (muscovite > biotite), medium-grained, incipient porphyritic granite	Absent	Qz, Kfs (Mc >> Orthoclase), Pl (Ab-Oligoclase), ± Ms I > Bt ± Ap ± Zrn ± Rt ± opaques ± Chl ± Ser ± Kln	Peraluminous (A/CNK 1~1.35) Calc-alkalic to alkali-calcic Magnesian Syenogranite (R1-R2 diagram 2) P2O5 = 0.33–0.38 wt. %

¹ molar A/CNK = Al₂O₃/(CaO + Na₂O + K₂O); ² R₁ = 4Si − 11(Na + K) − 2(Fe + Ti); R₂ = 6Ca + Mg + Al [85]; Key: Qz, quartz; Kfs, K-feldspar; Mc, microcline; Pl, plagioclase; Ab, albite; Bt, biotite; Ms, muscovite; Ap, apatite; Zrn, zircon; Mnz, monazite; Xtm, xenotime; Rt, rutile; Ttn, titanite; Chl, chlorite; Ser, sericite; Ep, epidote; Cal, calcite; Kln, kaolinite; A/CNK, aluminium saturation [9].

4. Sampling and Analytical Methods

4.1. Sampling and Optical Characterization of Zircons

Sampling for obtaining zircon concentrates of the Freixo de Numão and Capinha started with the collection of 17 kg and 15 kg of non-weathered rock, respectively. Both samples were collected from active quarries located inside the igneous suites.

In order to obtain a zircon concentrate for U–Pb geochronology and Lu–Hf isotopic analysis, a precise and extensive procedure was performed at the Department of Geosciences, Environment and Spatial Planning laboratories (Faculty of Sciences, University of Porto–FCUP). In a first stage, zircons were recovered after crushing and sieving the samples, which was followed by gravimetric (Wilfley table), magnetic (Frantz Isodynamic Magnetic Separator, Model L-1) and heavy liquid (bromoform) separations. Finally, under a magnifying lens, zircon grains were selected by random handpicking from zircon fractions (Figure S3). Three zircon fractions were selected according to their morphology, color and absence of inclusions, fractures and metamictisation. After the handpicking, the zircons were mounted in epoxy resin to make a probe and, afterwards, polished to expose their central portion for further steps.

Optical characterization of zircons was performed using a Leica DM LSP polarizing microscope with transmitted light, equipped with an electronic camera with the LAS EZ software 2.0.0 at the Department of Geosciences, Environment and Spatial Planning, and Institute of Earth Sciences (FCUP). Scanning Electron Microscopy with Energy Dispersive Spectroscopy (SEM-EDS) was performed at the Materials Centre (University of Porto) using a FEI Quanta 400 FEGESEM/EDAX Genesis X4M instrument. The SEM was operated at 15 kV in high vacuum mode, with manual aperture, and 3.5 nm beam spot sizes. Detailed descriptions of the equipment and its specifications are presented in [86].

4.2. Zircon U–Pb Dating

U-Th-Pb isotope and elemental measurements of zircon grains were conducted by laser ablation quadrupole inductively coupled plasma mass spectrometry (LA-Q-ICP-MS) at the SGIker-Geochronology and Isotope Geochemistry Facility of the University of the Basque Country (Spain). The samples were ablated with a Resolution M50 193 nm UV ArF excimer laser system coupled to a Thermo Fisher iCAP Qc quadrupole-based ICP-MS instrument with enhanced sensitivity through a dual pumping system. Instrument operating parameters and conditions of the laser ablation used for individual zircon analyses fell within those reported in Table S2. Measurements were made using a 16 μm laser beam focused on the polished zircons, with repetition rates of 5 Hz and laser fluence at the target of ca. 2.5 J/cm². Ablations were performed in helium and nitrogen, and the aerosol mixture was mixed with argon before entering the plasma source. The signals of ²⁰²Hg, ²⁰⁴(Pb + Hg), ²⁰⁶Pb, ²⁰⁷Pb, ²⁰⁸Pb, ²³²Th and ²³⁸U masses were acquired and the occurrence of common Pb in the samples was monitored by the evolution of the ²⁰⁴(Pb + Hg) signal intensity. Single analyses consisted of 20 s of background integration with laser off followed by 30 s integration with the laser firing and a 30 s delay to wash out the previous sample (ca. 10 s for 6 orders of magnitude) and prepare the next analysis. Data reduction was carried out with Iolite v. 4.3 [87] and VizualAge [88] using GJ-1 zircon standard [89] for calibration and Plesovice zircon [90] and BB40 zircon [91] as secondary standards. For each analysis, the time resolved signal of single isotopes and isotope ratios was monitored and carefully inspected to verify the presence of perturbations related to inclusions, fractures, mixing of different age domains or common Pb. The concentrations of U-Th-Pb were calibrated relative to the certified contents of GJ-1 zircon standard [92]. Concordia diagrams (2σ error ellipses), concordia ages (95% confidence level) and probability plots were produced using Isoplot 3.70 [93].

4.3. Lu–Hf Isotopes

Lu–Hf isotopic analyses of zircon grains were conducted by laser ablation multi-collector inductively coupled plasma source mass spectrometry (LA-MC-ICP-MS) at the FIERCE (Frankfurt Isotope & Element Research Center), Goethe Universitat, Frankfurt am Main, following the method described by [94]. The ¹⁷⁶Lu decay constant of 1.867 × 10⁻¹¹ yr⁻¹ was used in all calculations, e.g., [95,96]. Mantle evolution models compatible with this value for the decay constant were used: Chondritic ¹⁷⁶Hf/¹⁷⁷Hf = 0.282785 and ¹⁷⁶Lu/¹⁷⁷Hf = 0.0336 [97] and the depleted mantle of [98], modified to the λ ¹⁷⁶Lu and chondritic composition used, which produces a present-day ¹⁷⁶Hf/¹⁷⁷Hf value of 0.28325 (εHf = +16.4) from chondritic initial hafnium at ¹⁷⁶Lu/¹⁷⁷Hf_DM = 0.0388 [17]. The samples were ablated with a RESOlutionArF excimer laser (COMpex Pro 2) laser system with a two-volume ablation cell (Laurin Technic S155). Instrument operating parameters and conditions of the laser ablation used for individual zircon analyses are reported in Table S3.

Two-stage model ages for New Crust (T_NC) were calculated using the measured ¹⁷⁶Lu/¹⁷⁷Hf of each spot (first stage = age of zircon), a value of 0.0113 for the average continental crust (second stage), and a juvenile crust (NC) ¹⁷⁶Lu/¹⁷⁷Hf and ¹⁷⁶Hf/¹⁷⁷Hf of 0.0384 and 0.28314, respectively, following the proposal presented by [99].

5. Results

5.1. Zircon Morphology and Structure

The zircon concentrates were observed under a polarizing microscope (transmitted light) in order to characterize the zircon grains. In both concentrates, apatite and monazite were also identified. The abundance of zircon is higher in FNG than in CG, which is supported by their whole-rock Zr content of ca. 169 ppm and ca. 77 ppm, respectively [71].

The zircon autocrysts from FNG and CG display a high degree of transparency and consist of euhedral prisms with terminal pyramidal faces, which range from colorless to light brown. The prisms can be quite long (aspect > 4:1) and rarely show melt, monazite, rutile and apatite inclusions. However, short-prismatic morphologies were also identified in these magmatic zircons (aspect ratio < 4:1), especially in CG (Figure 3). Most of the zircon autocrysts show concentric oscillatory zoning, but nebulitic structures were also identified.

The inherited zircon cores are a fairly common feature of these granites, occurring in stubby prisms (aspect ratio < 3:1) or equant multifaceted crystals (Figure 3). The inherited zircon cores are, broadly, colorless/pale yellow and, generally, rimmed by newly grown fine-zoned and light brown colored rims. Their internal structure is variable and can be homogeneous or exhibit nebulitic or oscillatory patterns. Sometimes, the regular growth zoning can be interrupted by textural discontinuities along which the original pattern is resorbed and succeeded by the deposition of new-growth zoned zircon. Mainly in CG, there are zircon crystals that were most probably incorporated from host Variscan granites during transit and emplacement, and, therefore, they are termed as xenocrysts [100]. These xenocrysts share many of the features described for those with inherited zircon cores.

5.2. U–Pb Zircon Ages

A total of 74 analyses were carried out in zircons of an FNG sample, but 26 revealed more than 5% discordance and, therefore, were discarded. For CG, out of a total of 60 analyses in zircons, only 32 were considered concordant. The % concordance was calculated as: [(²⁰⁶Pb/²³⁸U age)/(²⁰⁷Pb/²³⁵U age)] × 100 [101].

Discordance is usually due to Pb loss, addition of common Pb or ablation of different domains within the zircon. The ²⁰⁶Pb/²³⁸U age was taken for interpretation of all inherited zircon cores younger than 1000 Ma and the ²⁰⁷Pb/²⁰⁶Pb ages for older grains, e.g., [102].

U–Pb isotopic data from FNG and CG are presented in Table S4.

5.2.1. Freixo de Numão Granite

The geochronological results for FNG are presented in Figure 4. Most of the analyses yield ages around 300 Ma, although there are a few older and younger values. Among the 48 results with concordance higher than 95%, there is an age gap between 480 and 330 Ma (Figure 4a,b), which was used to divide the data into two main groups: (1) younger than 330 Ma (grey bars in Figure 4b,c) and (2) older than 480 Ma (green bars in Figure 4b).

Thirty-eight concordant analyses, performed in zircon autocrysts, provided an apparent bimodal statistical distribution (Figure 4c), excluding the outlier of 323 Ma (zircon FNG-76a). The two identified subgroups (a₁ and b₁; Figure 4d) were then used in the determination of the following Concordia ages: 284 ± 4 Ma for the group a₁ (n = 6, highlighted in blue) and 306 ± 2 Ma for the group a₂ (n = 32, highlighted in red) (Figure 4c).

Alternatively, a weighted average age of 307.0 (+3.0/−3.0) Ma has been obtained from a coherent group of 26 zircons, using the TuffZirc algorithm of [103] implemented in Isoplot [104] (Figure 4e).

Ten analyses in inherited zircon cores yielded ages older than 480 Ma (Figure 4b). The main group of ages is Neoproterozoic (70%, 935–546 Ma), but there are also Paleoproterozoic (10%, 2155 Ma) and upper Cambrian (20%, 495–489 Ma). The Neoproterozoic population is characterized by a dominant cluster of ages in the Cryogenian–Ediacaran (650–546 Ma). Some inherited zircon cores are affected by a network of thin fractures (e.g., FNG-76a), providing evidence of resetting events that may reflect an interaction with the younger Variscan melts (Table S5).

5.2.2. Capinha Granite

The geochronological results for CG are shown in Figure 5, where the 33 results, with concordance higher than 95%, exhibit an age gap between two main groups. One group is dominated by ages around 300 Ma (grey bars in Figure 5a,b) and the other by ages older than 500 Ma (green bars in Figure 5b).

The group with ages around 300 Ma is made up of 23 analyses with concordant ages, which nevertheless show a wide dispersion, blocking any possible attempt to obtain a single Concordia age for the entire data set. This fact could be due to the coexistence of older Variscan xenocrysts or, alternatively, to the possibility of ablation-induced mixing between domains of different ages. In fact, three analyses out of the 23 performed were discarded from the calculations due to their strong relation to inherited zircon cores in the BSE images (namely CAP-84a, CAP-91 and CAP-98b, Table S5).

The analysis of the remaining 19 ages points to a bimodal distribution, subdividing the data into two main groups (Figure 5c): a younger group that yields a Concordia age of 301 ± 3 Ma (highlighted in red, Figure 5c,d), and an older group that yields a Concordia age of 317 ± 3 Ma (highlighted in blue, Figure 5c,d). Using the TuffZirc algorithm of [103], a weighted average age of 303.5 (+10.5/−4.5) Ma was obtained for a coherent group of 16 zircons (Figure 5e).

The group with ages older than 500 Ma consists of 14 analyses and is dominated by Neoproterozoic ages (93%), with a main cluster in the Cryogenian–Ediacaran (695–544 Ma), and a major peak at ca. 620–617 Ma. There are also Tonian zircons (779–739 Ma) and one from the Mesoproterozoic (7%, 1095 Ma). Some of these inherited zircon cores are surrounded by slightly younger overgrowths, such as in CG-108, CG-113 and CG-117 grains.

5.3. Lu–Hf Isotopic Data

The Lu–Hf isotopic results of magmatic zircons and inherited zircon cores of granites are summarized in Table S6.

5.3.1. Freixo de Numão Granite

The zircon autocrysts from FNG show initial ¹⁷⁶Hf/¹⁷⁷Hf ratios ranging from 0.282469 to 0.282623 (εHf_t = −4.5 to +0.6; Table S6 and Figure 6a) and New Crust (T_NC) from 1.35 to 1.09 Ga. Within these zircon autocrysts, just 5% present positive ɛHf_t values (Figure 6b).

Two inherited zircon cores of the FNG give ²⁰⁶Pb/²³⁸U ages from 495 to 489 Ma (upper Cambrian) and show initial ¹⁷⁶Hf/¹⁷⁷Hf ratios from 0.282507 to 0.282427 (εHf_t = +1.2 to −1.7; Table S6 and Figure 6a). Most inherited zircon cores in granite FNG give Neoproterozoic ²⁰⁶Pb/²³⁸U ages in the range of 935 to 546 Ma, but there is also a zircon with a ²⁰⁶Pb/²³⁸U age of 2155 Ma. The initial ¹⁷⁶Hf/¹⁷⁷Hf ratios for these zircons range from 0.281145 to 0.282612, corresponding to εHf_t from −16.8 to +8.4 (Table S6 and Figure 6b).

5.3.2. Capinha Granite

Initial ¹⁷⁶Hf/¹⁷⁷Hf ratios and εHf_t values range from 0.282441 to 0.282591 and from −5.5 to +0.3 for zircon autocrysts from CG (Table S6 and Figure 6a). The two-stage model ages for T_NC for CG range from 1.44 to 1.14 Ga. Among these magmatic zircons, only 8% display positive ɛHf_t values (Figure 6b).

The most representative age fraction among inherited zircon cores in the CG is of Neoproterozoic age (²⁰⁶Pb/²³⁸U ages ranging from 779 to 544 Ma). The initial ¹⁷⁶Hf/¹⁷⁷Hf ratios for these inherited zircon cores range from 0.281830 to 0.282694, corresponding to εHf_t of −19.4 to +9.1 (Table S6 and Figure 6a,b). However, there is also an older zircon, with a ²⁰⁶Pb/²³⁸U age of 1095 Ma and an initial ¹⁷⁶Hf/¹⁷⁷Hf ratio of 0.282343 (εHf_t = +10.1) (Table S6 and Figure 6b).

6. Discussion

6.1. Geochronological Constraints on the Petrogenesis of the Granites

The most representative group of concordant zircon autocrysts of the FNG give a Concordia age of 306 ± 2 Ma (group 2 of Figure 4c,d), which is considered the best estimate for the magmatic crystallization age. This age coincides with the weighted average age of 307.0 (+3.0/−3.9) Ma obtained with the TuffZirc algorithm of [103].

The age obtained for FNG in this study is consistent with field relationships, structural data and ductile deformation described by [73], suggesting a late- to post-D₃ passive ascent and emplacement crystallization. Nevertheless, a restricted group of younger grains (n = 6, group 1 of Figure 4c,d) yields evidence of Pb loss, presumably caused by the thermal overprint of a non-outcropping post-orogenic intrusion at 284 ± 4 Ma, which is probably related to the NNE–SSW Vilariça-Manteigas fault system. This would be in line with the geological context of Penacova-Régua-Verín fault, located to the west, where several examples of late- to post-orogenic (299–290 Ma) felsic Variscan granites are known, e.g., [8]. A previous geochronological study on FNG, carried out by [48], provided a younger result with a Rb-Sr whole-rock isochron yielding an age of 253 ± 19 Ma, which, at that time, led to the classification of FNG as a post-D₃ granite. However, care must be taken in the interpretation of this result since, under subsolidus conditions, the closing of the Rb-Sr whole-rock system on a local scale may be affected by the efficient action of hydrothermal fluids, e.g., [105].

Figure 6. Hypothetical Hf isotope evolution of different terrestrial reservoirs (adapted from [19]). CHUR, Chondrite Uniform Reservoir broadly taken as representative of Bulk Silicate Earth (BSE). (a) ¹⁷⁶Hf/¹⁷⁷Hf versus U–Pb ages (Ma) obtained in zircons from the Freixo de Numão and Capinha granites, and (b) the same diagram in epsilon Hf notation. The dashed band, representing the crustal evolution of the analyzed Variscan zircons (excluding outliers), was drawn by forcing growth curves of a system with a ¹⁷⁶Lu/¹⁷⁷Hf ratio of 0.0113, corresponding to the average continental crust, through the zircon initial ¹⁷⁶Hf/¹⁷⁷Hf ratios, as suggested by [106].

Figure 6. Hypothetical Hf isotope evolution of different terrestrial reservoirs (adapted from [19]). CHUR, Chondrite Uniform Reservoir broadly taken as representative of Bulk Silicate Earth (BSE). (a) ¹⁷⁶Hf/¹⁷⁷Hf versus U–Pb ages (Ma) obtained in zircons from the Freixo de Numão and Capinha granites, and (b) the same diagram in epsilon Hf notation. The dashed band, representing the crustal evolution of the analyzed Variscan zircons (excluding outliers), was drawn by forcing growth curves of a system with a ¹⁷⁶Lu/¹⁷⁷Hf ratio of 0.0113, corresponding to the average continental crust, through the zircon initial ¹⁷⁶Hf/¹⁷⁷Hf ratios, as suggested by [106].

For the CG, no dating was available in the literature, although it has been classified as a post-orogenic granite (310 to 290 Ma), because it is intrusive in the Pêro Viseu-Seia late-D₃ granite (U–Pb: 304.1 ± 3.9 Ma [59,83]; 307.7 ± 7.8 Ma to 305.2 ± 4.4 Ma [49]) and also based on field relationships with D₃ structures and deformation patterns. The concordant U–Pb data of magmatic zircons of CG reveal a significant age dispersion (Figure 5a,b), which may indicate the incorporation of zircon Variscan xenocrysts (with an estimated age of 317 ± 3 Ma, Figure 5c,d) [107,108]. This estimated age matches with the emplacement of the syntectonic (syn-D₃) S-type muscovite–biotite leucogranite in the Aguiar da Beira region, at 317.0 ± 1.1 Ma [109], which in all respects resembles some intrusions that crop out to the NW of the studied area. The BSE imaging of these zircon xenocrysts has shown that they are generally characterized, at least, by two distinct stages of formation (Figure 3). Therefore, the best age estimate for the CG, provided by the younger group of magmatic zircons, is 301 ± 3 Ma autocrysts (Figure 5c), which is in accordance with the weighted average age of 303.5 (+10.0/−4.0) Ma obtained with the TuffZirc algorithm of [103].

Concerning the emplacement conditions of both granites, the magnetic fabrics indicate that they should be coeval with the late- to post-D₃ granites, since, as mentioned above, the studied granites do not display any internal structure or deformation pattern. The ascending/emplacement and development of magnetic or magmatic fabrics were, therefore, conditioned by (a) NW–SE structures related to sinistral WNW–ESE sinistral shear zones, in the case of the FNG [73] and (b) the intersection of NNE–SSW and NNW–SSE fracturing systems, for the CG [72].

6.2. Magma Source(s)

6.2.1. Inferences from Petrography and Whole-Rock Geochemistry

The major element composition of the studied granites indicates that the biotite > muscovite FNG has a transitional alkali-calcic to calc-alkalic geochemical signature, whereas the muscovite > biotite CG belongs to the alkali-calcic series. Both granites are peraluminous, with average A/CNK of ~1.18 and ~1.35 for FNG and CG, respectively. These granites also have ilmenite (and absence of magnetite), fractionated chondrite-normalized REE patterns, low Na₂O (2.89 wt.% for FNG and 3.09 wt.% for CG) and high K₂O (5.21 wt.% for FNG and 4.4 wt.% for CG) and SiO₂ (68.9 wt.% for FNG and 72.76 wt.% for CG), highlighting their affinity to S-type magmas, which is also supported by the presence of metasedimentary enclaves in the FNG [110,111]. Considering that metapelitic rocks have CaO/Na₂O < 0.5, in contrast to metagreywacke or metaigneous rocks with CaO/Na₂O = 0.3–1.5, [111] have used this ratio to deduce the source composition of peraluminous granites. The FNG displays high CaO/Na₂O (0.44–0.66) suggesting a metaigneous or metagreywacke protolith, whereas the low CaO/Na₂O (0.17–0.23) of the CG seems to point to the partial melting of metapelite-dominated sources. Moreover, the lower peraluminosity and higher mean Sr and Ba whole-rock contents of FNG (ca. 250 ppm and 741 ppm, respectively) [73] are also in line with a partial melting process involving a heterogeneous crust, made up of feldspar-rich components [107,112,113]. In the local context, such mineralogy can be found in: (1) immature samples of metasediments from the outcropping Desejosa and Ervedosa do Douro formations, which contain plagioclase (rare K-feldspar only occurs in the latter formation [114]); (2) non-outcropping upper Cambrian-Lower Ordovician felsic porphyry dykes, comparable to those that cut the Douro Group metasediments at Mateus-Vila Real [68], Nabo-Carrazeda de Ansiães [69] and Sabrosa-Castedo [115], with Sr and Ba varying from 111 to 198 ppm and from 320 to 868 ppm, respectively [116].

Although the whole-rock geochemical signatures of the FNG and CG seem to indicate a major role of supracrustal protoliths in the genesis of these magmas, the FNG also contains mafic microgranular enclaves, which point to some interaction between mantle-derived magmas and felsic crustal melts, enough to generate a somewhat more primitive geochemical signature. In fact, the coeval input of mafic magma into the crust has been considered by several authors as the source of hybrid magmas and the driving mechanism for S-type melt generation in the CIZ, e.g., [49,117,118]. However, other authors argue that the generation of the S-type granite suites is dominated by continental crust recycling, rather than by magma mixing with basic melts, e.g., [42,119,120]. This agrees with thermal models that suggest that, due to the high internal radiogenic heat production of the existing crustal rocks and to the increased crustal thickness during the Variscan collision, granite magmatism in the CIZ did not require the addition of heat from mantle sources, e.g., [121,122].

A qualitative estimation of the melt formation temperatures associated with unfractionated granitic magmas can be obtained from the Al₂O₃/TiO₂ ratio, since magmas with low values are generated at higher temperatures than those with high Al₂O₃/TiO₂ ratios [111]. In this sense, FNG (Al₂O₃/TiO₂ = 28.95) likely originated at a higher temperature than CG (Al₂O₃/TiO₂ = 67.27). Temperature estimations in granitic magmas can also be inferred from the zircon saturation equation [123], assuming equilibrium conditions. The average zircon saturation temperature (Tzr) is 783 °C for FNG and 746 °C for CG, indicating a higher degree of partial melting for FNG [123]. However, these Tzr values are overestimated since there are inherited zircon cores in both granites [124].

Considering that CG plots near the strongly differentiated granite field in the Rb-Ba-Sr diagram of [81], its Sn content can reach 24 ppm and that it has an REE pattern with a significant Eu negative anomaly, it likely resulted from a magmatic differentiation process.

6.2.2. Inferences from U–Pb Isotopic Data

The distributions of ²⁰⁶Pb/²³⁸U and ²⁰⁷Pb/²⁰⁶Pb ages of the inherited zircon cores of both FNG and CG show predominant Neoproterozoic age populations (Figure 4b and Figure 5b), with main clusters in the Cryogenian–Ediacaran and scarcity in the Tonian, as well as in older ages, making it clear that there is a strong record of Pan-African and Cadomian major episodes of zircon crystallization, e.g., [125,126] and references therein. The presence of inherited zircon cores of upper Cambrian ages was only recognized in FNG. With the exception of these latter ages, the other ages show a general overlap with geochronology data from detrital zircons of metasedimentary formations equivalent to the FNG and CG wall rocks (Figure 7a). Examples of this similarity in the Precambrian inheritance include the Ediacaran metapelitic sediments from the Douro Group (northern CIZ autochthonous domain) [127,128], the late Ediacaran metagreywackes of the Penacova area [102,126] and, to a lesser extent, of the Sabugal region [127], both belonging to the Beiras Supergroup (Figure 7b). The similitude is extensible to the southern CIZ metasediments from Spain, which are characterized by a high percentage of 690 to 540 Ma ages [65] (Figure 7b). Therefore, considering the existing U–Pb data, it is admissible to infer that these metasediments can be considered potential sources of the inherited zircon cores recorded in FNG and CG.

Regarding the Cambrian ages, it turns out that they are within the time interval of the pre-Variscan magmatic events recognized in the northern domain of the Iberian autochthon (mainly in Spain), during the upper Cambrian and Lower Ordovician, around 495–475 Ma, e.g., [65,129,130]. In northern Portugal, there is less voluminous evidence of this magmatism, but a subvolcanic porphyry dyke, with a U–Pb age of 478.0 ± 1.7 Ma [68], cuts Douro Group metasediments near Mateus-Vila Real. Similar situations were also identified in the Nabo-Carrazeda de Ansiães [69] and Sabrosa-Castedo [117] areas, which may support the involvement of similar lithologies in the origin of the FNG magma.

However, the role played by contamination during emplacement should also be considered, and care must also be taken in the determination of potential granite sources, since they are not usually the outcropping metamorphic rocks, but geological formations located at deeper crustal levels [123,131].

6.2.3. Inferences from Lu–Hf Isotopic Data

The magmatic zircons from FNG show a wide range of negative εHf_t values (Figure 6b and Table S6), indicating that the granitic magma was most probably formed by melting of a relatively enriched crustal source. However, the occurrence of some magmatic zircons with slightly positive εHf_t values (up to +0.6) allows the contribution of a depleted mantle-derived component in the genesis of FNG, in agreement with the presence of mafic microgranular enclaves and its transitional alkali-calcic to calc-alkalic geochemical character [73].

The inherited zircon cores from the FNG and detrital zircons of the chlorite phyllite (belonging to Ervedosa do Douro Formation, with maximum depositional ages of ca. 578 Ma) [127] have similar initial ¹⁷⁶Hf/¹⁷⁷Hf ratios (Figure 8a and Table S6). This is particularly so regarding the domains of Cryogenian–Ediacaran and Tonian zircons, generally with low radiogenic _ƐHf_t values (up to +4) and a Paleoproterozoic grain (2155 Ma) with highly negative εHf_t (−13). This similitude supports the conclusions drawn from the U–Pb data that a metasediment, with these characteristics, could have been one of the sources involved in the genesis of granite FNG, as has also been established for the early syn-D₃ granite of Cabeça Boa-Carrazeda de Ansiães, northern Portugal [127]. Furthermore, the Hf isotope ratios of the analyzed Variscan zircons (excluding outliers) are coherent with crustal growth-curves that converge towards the Cambrian and Neoproterozoic inheritances (Figure 6), supporting the occurrence of successive recycling events from related crustal sources. However, the presence of other components cannot be discarded. In fact, the average εHf₃₀₆ for the detrital zircons of the chlorite phyllite is ca. −19, which is much more evolved than the magmatic zircons of FNG with an average εHf₃₀₆ of approximately −3, suggesting that the melt Hf isotopic ratios were not exclusively controlled by the detrital zircons of this metasediment. In this sense, other materials may have been involved in the genesis of this granite, such as juvenile Pan-African and Cadomian materials that likely constitute a cryptic stratigraphic infrastructure of most of the Iberian Massif [126,132], but also felsic metaigneous protoliths, as evidenced by the presence of an upper Cambrian inheritance [133,134,135].

With respect to the last hypothesis, there is, in fact, a considerable similarity in the U–Pb and Lu–Hf signatures of inherited zircon cores of the FNG and those of the restite-rich granites of the Sotosalbos complex (yielding Variscan ages in the range of 390–335 Ma [134]) and felsic granulite xenoliths of the lower crustal levels from the Spanish Central System (Figure 8a). According to refs. [8,131,135], these rocks were most probably formed through the partial melting/migmatization of Cambro-Ordovician orthogneisses, with the residuum being represented by the felsic granulite xenoliths. Their Hf isotope signatures are mainly controlled by lower Paleozoic and Neoproterozoic zircons with highly radiogenic Hf isotope compositions (Figure 8a).

The compatibility in the U–Pb and Lu–Hf signatures can also be partially extended to the I-type granites from the Spanish Central System batholith [136] and to the post-tectonic type-3 granitoids from Montes de Toledo batholith [132]. The most striking feature is the presence of Lower Ordovician inherited magmatic zircons, sometimes surrounding Ediacaran cores, with εHf_t variable from −13.6 to +5.0 (Spanish Central System) and from −13.8 to +0.3 (Montes de Toledo). This fact, together with the low peraluminousity and higher CaO, Na₂O, Sr, Y and HREE contents of these granitoids, makes the exclusive involvement of a metasedimentary source in their genesis improbable, and hence highlights the involvement of a Lower Ordovician metaigneous protolith in the melting process [132,137].

It is also admissible that part of the inherited zircons of FNG could come from contamination by the host metasediments during the emplacement of the granite body, not reflecting the source rocks, which agrees with the fact that the isotopic composition of granitic magmas derived from a source at depth does not necessarily have a one-to-one relationship to the equivalent metamorphic rocks at the level of emplacement [131,132].

Magmatic zircons from granite CG show a wide range of negative εHf_t values (Figure 6b and Table S6), suggesting that it could have been derived from the melting of a heterogeneous enriched crust. The existence of some magmatic zircons with slightly positive εHf_t values (up to +0.3) suggests a limited mixing with a mantle-derived depleted component. This possibility might seem unlikely due to the absence of mafic microgranular enclaves and the alkali-calcic geochemical composition of the CG [71]. However, considering that the geochemical composition of the CG may have been affected by a certain degree of magmatic differentiation, that hypothesis may already be feasible.

The widespread presence of Cryogenian, Ediacaran and, to a lesser extent, Tonian inherited zircon cores (779–544 Ma) with high radiogenic εHf values (up to +9) in CG points to a major involvement of a Neoproterozoic heterogeneous crustal source with an important contribution of juvenile Pan-African and Cadomian materials. The metagreywackes from the Beiras Supergroup could be considered rocks representative of the CG magma sources, as they contain highly radiogenic (εHf > +9) detrital zircons within the 810–672 Ma age range [127]. On the contrary, positive εHf values (>+4) have not been found in Neoproterozoic detrital zircons of the chlorite phyllite of the Douro Group [127], precluding the involvement of this metasediment in the genesis of CG. The overlapping of the CG Variscan zircon autocrysts and the Neoproterozoic inherited zircon cores within the same crustal growth-curve is indicative of a succession of recycling events from a common crustal protolith, but, as in the FNG, the presence of other components cannot be excluded (Figure 6). In fact, it should be mentioned that the average εHf₃₀₁ of the detrital zircons from the metagreywacke is ca. −24, which is significantly less radiogenic than the magmatic zircons of CG with an average εHf₃₀₁ at approximately −3. In this sense, the contribution of juvenile sources or more immature metasediments, enriched in juvenile components, in the genesis of the CG parental magma would be required, e.g., [62,132].

It is also noticeable that the Neoproterozoic ages and initial ¹⁷⁶Hf/¹⁷⁷Hf ratios of the inherited zircon cores of CG are almost coincident with those of the late- to post-D₃ granite of Monte Margarida-Sabugal, central Portugal [83,127] and partially match the signatures of the post-tectonic type-1 and type-2 highly peraluminous granitoids from Montes de Toledo batholiths, central Spain [132] (Figure 8b). These granites are intrusive in the southern CIZ metasediments, and their zircon inheritances show distinctive ages and Hf isotope compositions when compared to intrusions from the northern CIZ (Figure 8a,b). This may be considered as further evidence of the well-recognized contrasting geological, geochemical, and isotopic signatures between the northern and southern CIZ regions, e.g., [62] and references therein.

On the other hand, it can be assumed that the inherited zircon cores of the CG might come from metasedimentary contamination during emplacement, not reflecting the source of melt.

6.2.4. Crustal Residence Times

The two-stage Hf model ages T_NC for inherited zircon cores of the FNG (Table S6) indicate crustal residence times of 3243 Ma in the Paleoproterozoic zircon, 2263–963 Ma in the Neoproterozoic zircons and 1391–1235 Ma in the upper Cambrian zircons. The T_NC mean value for the dominant age cluster in the Cryogenian–Ediacaran (650–546 Ma) is 1512 Ma (Tables S4 and S5), which is within the range of whole-rock depleted mantle Nd model ages obtained for Douro Group metasediments (1483–1730 Ma; [60]). The T_NC mean value is significantly older than the Cryogenian–Ediacaran crystallization ages, indicating that these zircons represent a magmatic event that reworked crustal rocks, ultimately associated with a juvenile crustal component of about 1500–1700 Ma. Thus, magmatism in the Central Iberian Zone, during the interval from 650 to 546 Ma, can be considered mainly a crustal reworking event, rather than of crustal growth [24]. This is in accordance with the Nd isotopic signatures of metapelite samples from the northern CIZ in central Spain, which have provided a relatively old range of Nd model ages and indicate a major contribution of recycled crustal components [62].

The two-stage Hf model ages T_NC for inherited zircon cores of the CG (Table S6) show crustal residence times of 1277 Ma in the Mesoproterozoic zircon and ranging from 2481 to 846 Ma in the Neoproterozoic zircons. The mean T_NC value for the dominant age cluster in the Cryogenian–Ediacaran (695–544 Ma) is 1341 Ma (Table S6), which is within the range of Nd model ages of southern CIZ metasedimentary rocks from central Spain (1200 to 1560 Ma, with a mean value of 1380 Ma [62]) and the Albergaria-a-Velha area, Portugal (1170 to 1590 Ma, with a mean value of 1330 Ma [137]), and very close to those from the Raiva locality, Portugal (1249 to 1328 Ma, with a mean value of 1300 Ma [138]).

According to [62,139], the isotopic signatures and the lower Nd model ages of the metasediments from the southern CIZ, when compared to those from the northern CIZ, support the idea of a major contribution of juvenile sources. Nevertheless, the higher juvenile contribution in southern CIZ metasediments most likely does not reflect a more mafic composition, when compared to the metapelites from the northern CIZ [62].

The Cryogenian–Ediacaran inherited zircon cores from CG and the detrital zircons from the metagreywacke with ages ranging from 779 to 544 Ma plot mainly above the εHf_t CHUR evolution line (Figure 6b), indicating major magmatic episodes in the source area during the Neoproterozoic, which probably developed either along or in the vicinity of the northern Gondwana margin, in a continental arc setting related to the Cadomian orogeny [103]. These ages also include one of the main peaks of rapid global crust growth, as indicated by [24,140].

The two-stage Hf model ages T_NC for zircon autocrysts of FNG and CG show crustal residence times ranging from 1.35 to 1.09 Ga and from 1.44 to 1.14 Ga, respectively, which are very similar to those of their Neoproterozoic (and upper Cambrian in FNG) inheritance (Figure 8). As mentioned in Section 6.2.3, their growth-curves overlap, suggesting a genetic relationship between them, in which the granites would represent melts extracted from Neoproterozoic crustal rocks (either metasedimentary or magmatic).

7. Conclusions

U–Pb LA-Q-ICP-MS geochronological and Lu–Hf LA-MC-ICP-MS isotopic analyses of magmatic zircons and inherited zircon cores from the Variscan peraluminous Freixo de Numão (FNG) and Capinha (CG) granites, located in northern and central Portugal, respectively, were undertaken to investigate their emplacement age and constrain possible magma sources.

• U–Pb geochronology suggests that FNG and CG are both late- to post-kinematic in relation to the last ductile deformation phase of the Variscan orogeny (D₃), with crystallization ages of 306 ± 2 Ma for FNG and 301 ± 3 Ma for CG. However, care must be taken with the interpretation of individual ages, since there are dispersions of several Ma among the concordant zircon population of each granite;
• A few FNG zircon autocrysts yield ages younger than the calculated magmatic age, most probably due to a partial resetting event caused by the thermal overprint of a non-outcropping intrusion, whereas in CG the incorporation of Variscan xenocrystic zircons (with an estimated age of 317 ± 3 Ma) matches that of the emplacement of the syntectonic (syn-D₃) S-type muscovite–biotite leucogranite in the Aguiar da Beira region, which is similar to some intrusions that crop out to the NW of the studied area;
• Inherited zircon cores from FNG and CG yield a high percentage of Cryogenian and Ediacaran zircons (650–546 Ma for FNG and 695–544 Ma for CG), but also contain smaller proportions of Tonian, Mesoproterozoic and Paleoproterozoic zircons, which match quite closely the age patterns from the northern CIZ and southern CIZ host metasediments;
• Hf isotopic data of autocrystic zircons of FNG and CG show similar average values, but have a wide εHf_t range, indicating that they could have been derived from heterogeneous crustal anatectic melts. The mixing of melt batches of mantle-derived and crustal sources should also be considered, given the presence of mafic microgranular enclaves and the alkali-calcic to calc-alkalic affiliation of FNG. Their absence in CG can be due to its probable differentiation from a more mafic precursor;
• The U–Pb age and εHf_t values of inherited zircon cores of FNG and CG are compatible with a derivation from heterogeneous Neoproterozoic metasedimentary sources that include both juvenile and recycled crustal materials. Therefore, zircons of the host Douro Group and Beiras Supergroup metasedimentary sequences could be seen as components involved in the partial melting process that led to the formation of FNG and CG, respectively. Nevertheless, in both cases, the magmatic zircons require a much less evolved source than the metasedimentary sequences, and, therefore, the detrital zircon cores can be viewed as a contaminant during ascent and emplacement (xenocrysts) and not the main source for the melt. Thus, more juvenile sources must be considered in the genesis of the studied granites. In fact, the involvement of a metaigneous protolith in the origin of the FNG magma is attested by its high Sr and Ba whole-rock composition and the presence of an upper Cambrian zircon inheritance;
• There is a considerable similarity in the U–Pb and Lu–Hf signatures of inherited zircon cores of the FNG and those of the: (a) restite-rich granites of the Sotosalbos complex and felsic granulite xenoliths of the lower crustal levels from the Spanish Central System; (b) I-type granites from the Spanish Central System batholith; (c) post-tectonic type-3 granitoids from the Montes de Toledo batholith (central Spain); (d) early syn-D₃ granite of Cabeça Boa-Carrazeda de Ansiães, northern Portugal, as well as between CG and those of the: (a) late- to post-tectonic type-1 and type-2 granitoids of the Montes de Toledo batholith and (b) late- to post-D₃ granite of Monte Margarida-Sabugal, central Portugal. Ultimately, this fact agrees with the recognized geological, geochemical and isotopic contrast between the Neoproterozoic metasedimentary rocks from the northern and southern CIZ, but also with a plausible involvement of metaigneous sources in the melting process.