1.1 Context

The available radiometric data on the timing and relationships of Variscan magmatism and metamorphism in the Northeast Iberian massifs suggest a late, Upper-Carboniferous age for these events [1] compared for instance to Northwest Iberia [8], which occupies an equivalent position in the Chain, where syn-metamorphic granitoids are older [3]. In this regard, the Guilleries massif provides a good example of a close relationship between magmatism and medium- to high-grade metamorphism [11] to further evaluate this model. The dating of the Variscan intrusions would also help to know the relative age of the two main pervasive foliations found in the more metamorphic part of the massif [2]: S1 (D1) is the oldest cleavage (a slaty cleavage in low-grade areas), which becomes refolded by a strong syn-metamorphic S2 (D2) foliation, whose transpressive or transtensive nature is the subject of debate in different parts of the Chain. A shear-band style, extensional deformation D3 occurs in the studied massif that is linked to retrogression [11].

The Guilleries massif (Fig. 1) is comparable to the Canigó and Albera Variscan massifs [5,15] of the Pyrenean axial zone, but located approximately 50 km south of them. The massif is divided from south to north into three blocks: Osor, Susqueda, and Sant Martí (A, B, and C in Fig. 2). These blocks are separated by NE–SW-striking normal faults that sink their northern limbs, causing the deeper, more metamorphic, and richer in igneous intrusions Osor block to outcrop in the southern part of the massif.

1.2 Leucogranites in the Osor block

1.3 Diorites and tonalites in the Susqueda block

Three leucogranites from the Osor area and one biotite granite and one diorite from the Susqueda area in the Guilleries massif (see locations in Fig. 2) have been dated with U–Pb zircon geochronology using the Shrimp-RG instrument at Stanford University (See Appendix A for analytical techniques). Isotopic data were reduced and plotted using the Squid and IsoplotEx programs of Ludwig [6,7] and are presented in Tables 1 and 2. Uncorrected U–Pb zircon spot data are projected to the Tera–Wasserburg Concordia diagrams (Figs. 3–7) along a model common-Pb line after Stacey and Kramers [14]. For most samples, a preferred weighted mean ²⁰⁶Pb/²³⁸U age is calculated at 2-sigma level of precision and shown in a diagram adjacent to the Tera–Wasserburg diagrams (Figs. 3–7).

The three leucogranites from the Osor area are medium to fine-grained, muscovite-rich, two-mica granites and were collected from subconcordant veins, less than 20 cm thick, intrusive in the metasediments along the Osor River route GI-642. Their UTM (Zone 31N-ED50) coordinates are LGG1: 467211 4644698; LGG2: 467535 4644503; LGG5: 466345 4644509). U–Pb zircon data for sample LGG5, plotted in a Tera–Wasserburg concordia (Fig. 3a), shows that 11 data points, collected form a concordant array, gave a mean ²⁰⁶Pb/²³⁸U age of 301.5 ± 1.7 Ma (Fig. 3b; MSWD = 0.94; 2-sigma). Individual zircon ages used to calculate this ²⁰⁶Pb/²³⁸U age are shown in bold in Table 1. Only one data point (spot-12; dotted-line error ellipse in Fig. 3a) for this sample was excluded from the calculation because it is statistically older, and may represent an area of the zircon with some degree of inheritance. We interpret this mean ²⁰⁶Pb/²³⁸U age as the time of crystallization for this leucogranite vein.

Zircon data for leucogranite sample LGG1 are more complex and indicate the presence of abundant zircon inheritance with both Precambrian (Fig. 4a; ²⁰⁷Pb/²⁰⁶Pb ages) and Palaeozoic ages (Fig. 4b). Individual zircon ages are shown in Tables 1 and 2. There are two concordant data points (Fig. 4b) close to the age obtained for leucogranite samples LGG5 and LGG2 (discussed below), but the youngest single-spot ²⁰⁶Pb/²³⁸U age at 305.3 ± 1.9 Ma (1-sigma) is our best approximation for the crystallization age of this leucogranite sample. We acknowledge the limitation of this age constraint for leucogranite LGG1.

Leucogranite sample LGG2, collected only few hundred metres away from LGG1, has a zircon population with much better U–Pb behaviour; there is no evidence of the presence of inheritance comparing to sample LGG1. In this case, a Tera–Wasserburg Concordia plot (Fig. 5a & b) shows that eight data points collected from a concordant array gave a mean ²⁰⁶Pb/²³⁸U age of 299.0 ± 2.3 Ma (Fig. 5c; MSWD = 1.6; 2-sigma). Individual zircon ages used to calculate the preferred ²⁰⁶Pb/²³⁸U age are shown in bold in Table 1. Three data points (spot-2, 5 & 8; dotted-line error ellipses in Fig. 5a & b) were excluded from the age calculation because of their high U and very high common Pb values. In addition, spot-10 was excluded from the calculation because it is statistically older, and may represent an area of the zircon with some degree of inheritance. We interpret the mean ²⁰⁶Pb/²³⁸U age as the time of crystallization for leucogranite vein LGG2; within the limits of analytical error, this age is the same as the one obtained for leucogranite LGG5.

Sample GRG4, a medium-grained biotite granite, was sampled on the southern bank of the Susqueda reservoir (UTM: 457824 4645624). U–Pb data of 12 zircon points gave a mean ²⁰⁶Pb/²³⁸U age of 305.9 ± 1.5 Ma (Fig. 6c; MSWD = 0.85; 2-sigma). Only one data point (spot-7; dotted-line error ellipse in Fig. 6b) from this sample was excluded from the age calculation because this zircon spot is statistically younger, and may represent initial stages of Pb loss in this particular zircon. We interpret this mean ²⁰⁶Pb/²³⁸U age as the time of crystallization for the Susqueda biotite granite GRG4.

Sample G04SQ, a fine-grained diorite, was collected near the southern end of the Susqueda dam (UTM: 460601 4647607) from an outcrop showing an igneous banding concordant with the host rocks. U–Pb zircon data for this sample, plotted in a Tera–Wasserburg concordia (Fig. 7a), are very concordant, and nine data points gave a mean ²⁰⁶Pb/²³⁸U age of 323.6 ± 2.8 Ma (Fig. 7b; MSWD = 1.9; 2-sigma). Only one data point (spot-4; dotted-line error ellipse in Fig. 7a) from this sample was excluded from the age calculation, because this zircon spot was very high U compared to the other concordant zircon grains. There is no evidence of the presence of zircon inheritance in this sample. Individual zircon ages used to calculate the preferred ²⁰⁶Pb/²³⁸U age are shown in bold in Table 1. We interpret the mean ²⁰⁶Pb/²³⁸U age as the time of crystallization for the Susqueda diorite.

This study shows that the Susqueda subconcordant intrusion and its contact-induced melting in the metapelites constituted a specific event that took place at 323.6 ± 2.8 Ma at the Namurian–Visean boundary [4] at ca. 3 kbar and 760 °C [12]. This would also be the age of D2, whatever the meaning (transpressive or transtensive) of the latter is.

This Susqueda diorite would be equivalent in age and probably in meaning to the earliest intrusive ones in the West Asturian Leonese zone in northwestern Iberia [3]. Consequently, D1 must be older, at least Lower Carboniferous, in northeastern Iberia.

The syn-D2 Susqueda intrusion is clearly separated in time from that of the leucogranites (301.5 ± 1.7; 305.3 ± 1.9; 299.0 ± 2.3 Ma) and of the coeval biotite granite (305.9 ± 1.5 Ma). These Westphalian–Stephanian ages are similar to those of the Albères, Querigut [13] Mont-Louis, and La Junquera [9] granitoids, and several others in the Pyrenean axial zone (B. Laumonier, pers. comm.), which have been related to the extensional collapse of the chain D3. Also related to the latter would be the pre- to syn-D3 leucogranites, D2 and D3 hence being separated by around 20 Ma in this part of the chain.

Another interesting consideration is that leucogranites and biotite granites are almost coeval, the former being slightly younger. The age link between both is an unexpected result of this study, and the reason for dating three different leucogranite veins, in an attempt to confirm their age. The presence of andalusite xenocrysts [2] in the leucogranites suggests that these are, at least in part, the melting-induced products of the biotite granites in the host rocks, at a deeper level than that of the present outcropping. This induced melting could have been similar to that which occurred in the contact of the Susqueda complex around 20 Ma earlier, where 15–20% partial melt was formed, restricted to the inner contact rocks [12].

Deformation can be used to constrain the age of the metamorphism in Osor if it is assumed that D2 is the same age in Osor and Susqueda. In this case, the 323.6 ± 2.8 Ma age of the Susqueda diorite would also be the age of the syn-D2 metamorphic peak in Osor. If the baric peak in Osor at around 5 kbar and 660 ± 40 °C [11], near the muscovite dehydration melting, is about the same age, then there is also a 20-Ma gap between the baric peak and the leucogranites (323 Ma vs. 300 Ma). An age gap of 10 to 15 Ma between the metamorphic baric peak and leucogranite formation is observed in the Himalayas [16], where the latter appears to be formed by decompressional melting of the protoliths [10]. However, in the Guilleries massif, given the spatial and temporal relation between the leucogranites and biotite granites, the heat supplied by the latter seems to have played a critical role in leucogranite genesis in a thermal pulse younger than the high-grade metamorphism in Osor and the likely coeval Susqueda intrusion.

There are still two questions that remain open: one is that, unlike in the case of the Susqueda contact rocks, the heat source involved in the widespread low-P, high-T Osor metamorphism is still unknown; the other is why the late biotite granites induced crustal melting, producing the leucogranites, but there is only limited melting restricted to the contact in the hotter Susqueda diorite and there are no melts linked with the likely coeval sillimanite-grade Osor metamorphism. A hypothetical answer to the latter question would be that Susqueda was intruded in relatively cold middle-crustal rocks, and that melts migrated from the Osor rocks, explaining their restitic signature [11].

The online version of this article contains additional supplementary material. Please visit at doi:10.1016/j.crte.2007.12.006