Introduction

The Sierras Pampeanas geological province of northwestern Argentina contains abundant granitoid massifs generated during the Famatinian orogenic cycle (for details see Rapela et al. 2001a; Miller and Söllner 2005). Most of these Famatinian granitoids are related to the main subduction phase of this cycle (e.g. Pankhurst et al. 2000; Rapela et al. 2001a; Miller and Söllner 2005) and have Early-Middle Ordovician ages (e.g. Pankhurst et al. 1998, 2000; Söllner et al. 2001; Höckenreiner et al. 2003) (Fig. 1a). These granitoids are distributed along two sub-parallel, NNW–SSE trending belts: a main calc-alkaline I-type belt towards the southwest, and an inner peraluminous and S-type belt towards the northeast (Fig. 1a).

Fig. 1
figure 1

a General geological map of the Sierras Pampeanas of NW Argentina with the main lithologies; sierras considered in the text are named. b General geology of the Sierra de Velasco; c Geological map of the central part of the Sierra de Velasco showing the Huaco, Sanagasta and La Chinchilla granites, with locations of dated samples; Bt biotite, Ms muscovite, Crd cordierite, Mzgr monzogranite, Ton tonalite, Grd granodiorite

Full size image

Additionally, numerous younger granites of Middle Devonian to Lower Carboniferous age are also present in the Sierras Pampeanas (e.g. Brogioni 1987, 1993; Rapela et al. 1991; Grissom et al. 1998; Llambías et al. 1998; Saavedra et al. 1998; Siegesmund et al. 2004; Dahlquist et al. 2006) (Fig. 1a). The genesis of these granites is not well constrained, and they have been alternatively considered as products of a crustal reheating process during a final phase of the Famatinian cycle, (e.g. Grissom et al. 1998; Llambías et al. 1998; Höckenreiner et al. 2003; Miller and Söllner 2005) or part of a separate cycle called Achalian (e.g. Sims et al. 1998; Rapela et al. 2001a; Siegesmund et al. 2004; López de Luchi et al. 2007).

The Sierra de Velasco is located in the central region of the Sierras Pampeanas (Fig. 1a) and consists almost entirely of rocks of granitoid composition, making it the largest granitic massif of this geological province. The Sierra de Velasco granitoids have generally been regarded as part of the Famatinian inner peraluminous S-type belt (e.g. Rapela et al. 1990; Toselli et al. 1996, 2000; Pankhurst et al. 2000), with the exception of the southern portion of the sierra which seems to correspond to the main calc-alkaline I-type belt (Bellos et al. 2002; Bellos 2005) (Fig. 1a, b). However, field studies carried out in the northern (Báez et al. 2002; Báez and Basei 2005) and central (Grosse and Sardi 2005; Grosse et al. 2005) parts of the sierra indicate the presence of younger undeformed granites (Fig. 1b), possibly belonging to the late-Famatinian, or Achalian, granite group. Recent U–Pb age determinations have confirmed that the northern undeformed granites are of Lower Carboniferous age (Báez et al. 2004; Dahlquist et al. 2006). The central undeformed granites have yet to be dated.

The goal of this study is to determine the absolute ages and the geochemistry of the undeformed granites located in the central part of the Sierra de Velasco. To this end, we have carried out U–Pb dating on monazite and whole-rock elemental and Sr–Nd isotopic geochemical analyses. The obtained data are used to place constraints on the possible magma sources and geotectonic setting of these granites, and to discuss regional implications.

Geological setting: the Sierra de Velasco

The Sierra de Velasco is dominated by rocks of granitoid composition. Low grade metamorphic rocks are only present as small outcrops along the eastern flank of the sierra (Fig. 1b, c). These phyllites and mica schists have been correlated with the La Cébila Formation, located in the Sierra de Ambato (González Bonorino 1951; Espizua and Caminos 1979). Recent discovery of marine fossils in this formation constrains its age to the Lower Ordovician (Verdecchia et al. 2007), in agreement with detrital zircon geochronology (Rapela et al. 2007).

The granitoid units of the Sierra de Velasco have been reviewed and described by Toselli et al. (2000, 2005) and Báez et al. (2005). Two groups can be distinguished (Fig. 1b): older deformed granitoids (here referred to as metagranitoids) and younger undeformed granites. The metagranitoids are the most abundant rocks. They are weakly to strongly foliated, depending on the degree of deformation. The main variety consists of strongly peraluminous porphyritic two-mica-, garnet-, sillimanite- and kyanite-bearing meta-monzogranites (Rossi et al. 2000, 2005). Subordinate varieties include strongly peraluminous porphyritic biotite–cordierite meta-monzogranites and moderately peraluminous coarse- to medium-grained biotite meta-granodiorites and meta-tonalites. In the southern part of the sierra, the main lithologies are metaluminous to weakly peraluminous biotite-hornblende meta-granodiorites and meta-tonalites (Bellos 2005) (Fig. 1b). Two U–Pb SHRIMP determinations indicate Lower Ordovician ages for the metagranitoids (481 ± 3 Ma, Pankhurst et al. 2000; 481 ± 2 Ma, Rapela et al. 2001b).

All of the metagranitoids are cut by several NNW–SSE trending mylonitic shear zones (Fig. 1b). No age determinations exist of these shear zones in the Sierra de Velasco. However, similar mylonitic shear zones in other areas of the Sierras Pampeanas have been dated, with ages varying between the Upper Ordovician and the Upper Devonian (Northrup et al. 1998; Rapela et al. 1998; Sims et al. 1998; López et al. 2000; Höckenreiner et al. 2003). The precise Sm–Nd age of 402 ± 2 Ma (Höckenreiner et al. 2003) obtained on syntectonically grown garnet from mylonites of the Sierra de Copacabana (Fig. 1a), which can be traced directly into the Sierra de Velasco (López and Toselli 1993; Söllner et al. 2003), can be considered the best age estimate of mylonitization in this range.

The undeformed granites crop out in the northern and central-eastern parts of the sierra (Fig. 1b). Toselli et al. (2006) have grouped these granites in the Aimogasta batholith. The northern San Blas and Asha granites intrude the older metagranitoids and cross-cut the mylonitic shear zones (Báez et al. 2002; Báez and Basei 2005). They are moderately to weakly peraluminous porphyritic two-mica monzogranites. Existing U–Pb ages are 334 ± 5 Ma (conventional U–Pb method on zircon, Báez et al. 2004) and 340 ± 3 Ma (U–Pb SHRIMP on zircon, Dahlquist et al. 2006) for the San Blas granite, and 344 ± 1 Ma (conventional U–Pb method on monazite, Báez et al. 2004) for the Asha granite.

In restricted areas, the granitic rocks are unconformably overlain by continental sandstones and conglomerates of the Paganzo Group (Salfity and Gorustovich 1984), of Upper Carboniferous to Permian age, deposited during regional uplift of the Sierras Pampeanas. Unconsolidated Tertiary-recent sediments, related to Andean tectonics, locally fill basins and form fluvial terraces and cones.

The Huaco, Sanagasta and La Chinchilla granites

The central-eastern region of the Sierra de Velasco is formed mainly by two large granitic massifs, the Huaco granite (HG) and the Sanagasta granite (SG) (Fig. 1c) (Grosse and Sardi 2005). These granites consist of adjacent, sub-elipsoidal bodies with dimensions of approximately 40 × 30 km for the HG and 25 × 15 km for the SG. Additionally, a small stock of around 2 × 2 km, named La Chinchilla stock (LCS), has been recognized in the central area of the HG (Fig. 1c) (Grosse et al. 2005).

The HG and the SG intrude into the older metagranitoids and mylonites and are not deformed. The contacts are sharp and the granites truncate both the structures of the metagranitoids and the mylonitic shear zones, and contain enclaves of both of these host rocks. These field relationships indicate that the granites are younger than both the crystallization of the metagranitoids and their deformation. The contact between the HG and the SG is irregular and transitional, suggesting that the two granites have similar ages and consist of two coeval magmatic pulses. The transitional area between the two granites is of ∼100–200 m; in Fig. 1c the contact between the granites was drawn along this transitional zone. The LCS clearly intrudes into the HG and is thus younger. The contacts are sharp and straight, and aplitic dykes from the LCS commonly cut through the HG.

Both the HG and the SG are rather homogeneous porphyritic syeno- to monzogranites. They are characterized by abundant K-feldspar megacrysts up to 12 cm long (generally between 2 and 5 cm) set in a medium- to coarse-grained groundmass of quartz, plagioclase, K-feldspar, micas and accessory minerals. The megacrysts are usually oriented, defining a primary magmatic foliation.

The HG consists in grayish-white K-feldspar megacrysts (30–36 vol.%) and a groundmass of anhedral quartz (25–39%), subhedral plagioclase laths (An10–23) (18–31%), interstitial perthitic K-feldspar (2–14%), dark brown to straw-colored biotite (4–10%) and muscovite (2–6%). Accessory minerals include apatite (up to 0.5%), zircon, monazite and ilmenite, all of which are generally associated with, or included in, biotite.

The SG contains pink K-feldspar megacrysts (33–37%) that are occasionally mantled by plagioclase generating a Rapakivi-like texture. The groundmass consists in anhedral quartz (23–34%), subhedral plagioclase laths (An18–24) (17–33%), interstitial perthitic K-feldspar (2–17%), and dark brown to straw-colored biotite (3–10%). Muscovite is absent or very scarce (0–2%). Accessory minerals are commonly found included in biotite. Apatite is less abundant than in the HG, whereas zircon, monazite and especially the opaque minerals (both ilmenite and magnetite) are more frequent. In addition, titanite and allanite are sometimes present.

Both the HG and the SG commonly contain small and rounded mafic microgranular enclaves. These generally have ovoid shapes, elongated parallel to the magmatic flow direction. The enclaves are fine- to very fine-grained equigranular tonalites and quartz-diorites. They contain abundant biotite (15–50%) forming small, subhedral crystals. Opaque minerals and acicular apatite are common. The enclaves usually contain much larger xenocrysts of quartz, feldspar or biotite, and have chilled margins, suggesting partial assimilation and homogenization with the enclosing granites.

Pegmatites and aplites are very common in these granites, specially in the HG. The larger pegmatites are zoned and belong to the rare-element class, beryl type, beryl-columbite-phosphate sub-type with a hybrid LCT-NYF affiliation (Galliski 1993; Sardi 2005; Sardi and Grosse 2005). The HG also contains a small outcrop of an orbicular granite (Quartino and Villar Fabre 1962; Grosse et al. 2006b).

The LCS is a medium-grained, equigranular to slightly porphyritic, monzogranite. It shows a weak textural zonation determined by a progressive increase in grain size towards the center of the stock, where a slight porphyritic texture is present (up to 10% of K-feldspar megacrysts). Mineralogically, the LCS consists of quartz (37–42%), plagioclase (almost pure albite, An1–2) (25–33%), K-feldspar (19–34%), discolored, very pale brown to pale red-brown biotite (4–9%), anhedral and irregularly shaped fluorite (up to 1%) and small quantities of zircon, monazite, opaque minerals and very scarce apatite. Beryl is occasionally present as euhedral prismatic crystals.

Microprobe analyses (Grosse et al. 2006a) indicate that the biotites of the HG and the SG have compositions ranging from Fe-biotites to siderophyllites (according to the classification diagram of Tischendorf et al. 1997) and have high Fe/(Fe + Mg) ratios (0.76–0.82), typical of evolved granites. In the discrimination diagram of Nachit et al. (1985), they plot in the calc-alkaline field. Biotites from de LCS have very high Fe/(Fe + Mg) ratios (0.94–0.97) and are Li-rich. They classify mainly as zinnwaldites and also as protolithionites in the classification diagram of Tischendorf et al. (1997).

Zircons of the HG and the SG have similar morphologies. They correspond mainly to the S17–19 and S22–23 types of Pupin (1980), which are characteristic of calc-alkaline series granites. On the other hand, the zircons of the LCS are different, with morphologies mostly of the P5-type of Pupin (1980), of primitive alkaline affiliation. The San Blas granite, in the north of the sierra (Fig. 1b), has the same zircon typology as the LCS.

No previous U–Pb age determinations exist of the HG and the SG, while the LCS has not been previously dated by any method. K–Ar and Rb–Sr geochronological studies have been carried out on granites of the Sierra de Velasco, which in some cases correspond to the HG or SG (see compilation in Linares and González 1990). The ages in these studies are very variable, spanning from the Ordovician to the Permian, probably due to the inherent problems of the methods used (low closure temperature, Ar loss, etc.).

Analytical methods

U–Pb geochronology

U–Pb geochronology was carried out at the Department of Earth- and Environmental Sciences, Ludwig-Maximilians-Universität, Munich, Germany. Heavy mineral concentrates, mainly zircons and monazites, were obtained using standard crushing, magnetic separation, and heavy-liquid techniques. For each analyzed sample around 50 monazite crystals were handpicked. Chosen crystals were yellow, translucent, anhedral to subhedral and lacked inclusions and fractures. We chose to analyze monazites because this mineral generally does not contain inherited cores and does not suffer radiogenic Pb loss at low temperatures, both common problems in zircons (see Parrish 1990 for discussion). Additionally, the closing temperature of monazite, although slightly lower than that of zircon (for details see Romer and Rötzler 2001), is sufficiently high to maintain the system unperturbed by low-temperature post-crystallization events.

The monazite fractions were cleaned with purified 6 N HCl, H2O and acetone, and then deposited in Teflon inserts together with a mixed 205Pb–233U spike. Subsequently, samples were dissolved in autoclaves, heated at 180°C, for 5 days using 48% HF and subsequently 6 N HCl. The U and Pb of the samples were separated using small 50 μl ion exchange columns with Dowex raisin AG 1 × 8 100–200 mesh. The isotopic ratios of Pb and U were determined with a thermal ionization mass spectrometer (TIMS) Finnigan MAT 261/262. Pb isotopes were measured in static mode and U isotopes in dynamic mode. Standards (NBS 982Pb and U500) were used for measurement control. U–Pb data was treated using the PBDAT 1.24 (Ludwig 1994) and ISOPLOT/Ex 2.49x (Ludwig 2001) programs. Errors quoted are at the 2σ confidence level. The corrections for initial non-radiogenic Pb was obtained following the model of Stacey and Kramers (1975). The U decay constants proposed by the IUGS (Steiger and Jäger 1977) were used for the age calculations. Mass fractionation was corrected using 0.13 ± 0.06%/a.m.u. for Pb and 0.05 ± 0.04% per a.m.u for U. Together with the samples, a procedural blank was analyzed to determine the level of contamination. For Pb blank corrections a mean value of 0.2 ng and an isotopic composition of 208Pb/204Pb = 38.14; 207Pb/204Pb = 15.63; 206Pb/204Pb = 18.15 was used. Long term measured standards gave values of: NBS 982 (Pb): 208Pb/206Pb = 0.99474 ± 0.00013 (0.013%, 2σm, n = 94); U500 (U): 238U/235U = 1.00312 ± 0.00027 (= 0.027%, 2σm, n = 14).

Whole-rock major and trace element geochemistry

Whole-rock geochemistry was determined at the universities of Oviedo (major elements) and Huelva (trace elements), Spain. Major elements were analyzed by X-ray fluorescence (XRF) with a Phillips PW2404 system using glass beads. The typical precision of this method is better than ±1.5% relative. Trace elements were analyzed by inductively coupled plasma mass spectrometry (ICP-MS) with an HP-4500 system. Samples were dissolved using a mixture of HF + HNO3 (8:3), a second dissolution in HNO3 after evaporation and final dissolution in HCl. The precision and accuracy for most elements is between 5 and 10% relative (5–7% for Rb, Sr, Nd and Sm) and was controlled by repeated analyses of international rock standards SARM-1 (granite) and SARM-4 (norite). Details on the method can be found in de la Rosa et al. (2001).

Sr and Nd isotope geochemistry

Sr and Nd isotope analyses were carried out at the Department of Earth- and Environmental Sciences, Ludwig-Maximilians-Universität, Munich, Germany. The analyzed powders were the same as those used for major and trace element analyses. For the determination of concentrations and for comparison with the ICP-MS data, a mixed Sm–Nd spike was added to 12 samples. For the remaining samples, and for all Rb–Sr calculations, the concentrations obtained by ICP-MS were used.

Samples (approximately 0.1 g each) were dissolved on a hot plate (140°C) during 36 h using a mixture of 5 ml of HF 48% + HNO3 (5:1). Sr and REE were separated using ion exchange columns with Dowex AG 50W raisin. Nd and Sm were then separated from the total REE fractions using smaller ion exchange columns with bis(2-ethylhexyl)phosphoric acid (HDEHP) and Teflon powder. The isotopic ratios of Sr, Nd and Sm were determined with a thermal ionization mass spectrometer (TIMS) Finnigan MAT 261/262. Standards were used for measurement control (NBS987, AMES Nd and AMES Sm). All errors used are at the 95% (2σ) confidence level. Mass fractionation was corrected normalizing the isotopic ratios to 88Sr/86Sr = 8.3752094 for Sr, 146Nd/144Nd = 0.7219 for Nd, and 148Sm/152Sm = 0.4204548 for Sm. CHUR constants used for ɛNd calculation were 143Nd/144Nd = 0.512638 (Goldstein et al. 1984) and 147Sm/144Nd = 0.1967 (Peucat et al. 1988). One-step model ages were calculated following Goldstein et al. (1984) (with 143Nd/144Nd (DM) = 0.51315 and 147Sm/144Nd (DM) = 0.217) and two-step model ages were calculated following Liew and Hofmann (1988) (with 143Nd/144Nd (DM) = 0.513151, 147Sm/144Nd (DM) = 0.219 and 147Sm/144Nd (CC) = 0.12). During the period of analyses, the measured standards gave the following average values: NBS987 (Sr): 87Sr/86Sr = 0.710230 ± 0.000013 (0.0018%, 2σm, n = 8); AMES (Nd): 143Nd/144Nd = 0.512131 ± 0.000007 (0.0013%, 2σm, n = 10); AMES (Sm): 149Sm/147Sm = 0.91262 ± 0.00016 (0.018%, 2σm, n = 3).

U–Pb monazite geochronology

Monazite fractions of six samples were analyzed, three of which correspond to the Sanagasta granite (SG), two to the Huaco granite (HG), and one to the La Chinchilla stock (LCS). Locations of the analyzed samples are shown in Fig. 1c. Table 1 shows the analytical results.

Table 1 U–Pb monazite data of the three studied granites of central-eastern Sierra de Velasco
Full size table

In the U–Pb concordia diagram (Fig. 2), two of the six analyzed samples are concordant whereas the other four are discordant, three of which plot above the concordia (phenomenon called “reverse discordance”) and one below. Reverse discordance in monazite has been observed by many authors and seems to be a common phenomenon in this mineral (Parrish et al. 1990, and references therein). Schärer (1984) suggests that reverse discordances are owed to an excess in 206Pb due to the decay of 230Th, an intermediate product in the decay chain of 238U to 206Pb, incorporated in significant amounts in the crystal during crystallization of monazite, because this mineral is a carrier of Th. This might be valid for sample 7703Mo, which is slightly reverse discordant (Fig. 2). However, samples 7365Mo, 7381Mo and 7369Mo are strongly reverse and normally discordant, respectively (Fig. 2). These samples probably suffered loss of U (7365Mo, 7381Mo) and radiogenic Pb (7369Mo).

Fig. 2
figure 2

U–Pb Concordia diagram of monazites from the three studied granites of central-eastern Sierra de Velasco. Two samples correspond to the Huaco granite (HG: 7365Mo and 7381Mo), three to the Sanagasta granite (SG: 7369Mo, 7379Mo and 7703Mo) and one to the La Chinchilla stock (LCS: 7740Mo). See text for further explanations. Plotted error ellipses and quoted errors are at the 2σ confidence level

Full size image

The two samples of the HG are strongly reverse discordant, probably due to loss of U (U contents: 6,135 and 10,129 ppm) (Fig. 2). 207Pb/206Pb ages of both samples are equivalent within limits of errors at 350 ± 5 and 358 ± 5 Ma. These ages are interpreted as the best estimate for crystallization of the HG. Recently, Söllner et al. (2007) have carried out LA-ICP-MS U–Pb age determinations on zircons of sample 7365 of the HG, obtaining a main crystallization age of 354 ± 4 Ma, thus confirming the monazite 207Pb/206Pb ages. In addition, many of these zircons have non-detrital inherited cores with Ordovician ages, suggesting significant participation of Ordovician metagranitoids in the formation of the HG (Söllner et al. 2007).

Only one of the three samples of the SG (sample 7379Mo) gives a concordant age of 352.7 ± 1.4 Ma (degree of discordance = 1.5%, Fig. 2). Sample 7703Mo is slightly reverse discordant at 350.3 ± 1.2 Ma (207Pb/235U age), whereas sample 7369Mo is strongly discordant at 38.0 ± 0.1 Ma (206Pb/238U age; 207Pb/206Pb age = 342 ± 8 Ma) (Fig. 2), suggesting loss of radiogenic Pb, possibly related to the very high measured U content (30,483 ppm) and the presence of dim and/or fractured crystals. All three data points, including the origin, fit a regression line with an upper intercept of 340 ± 26 Ma (MSWD = 3.8). The concordant age of 352.7 ± 1.4 Ma of sample 7379Mo is interpreted as the most precise and adequate age of crystallization of the SG.

Sample 7740Mo of the LCS is concordant at 344.5 ± 1.4 Ma (degree of discordance = 1.2%, Fig. 2), which is interpreted as dating the time of crystallization of the LCS.

Geochemistry

Major and trace elements

Table 2 shows 31 whole-rock major and trace element chemical analyses of the studied granites; 13 analyses correspond to the HG, 10 to the SG, 4 to the LCS and 4 to mafic microgranular enclaves of the HG and the SG (see also Grosse et al. 2007). For comparison, the average composition of the border and central facies of the San Blas granite are also shown (calculated from 13 analyses of Báez 2006).

Table 2 Chemical analyses of the studied granites and mafic enclaves of central-eastern Sierra de Velasco, and average composition of the San Blas granite of northern Sierra de Velasco
Full size table

The HG and the SG are characterized by a high and restricted SiO2 range of 69.7–74.7% (wt%). With slightly lower average SiO2, the SG has somewhat higher Fe2O tot3 , MgO, TiO2 and CaO concentrations than the HG, although both granites are poor in these oxides. They are, on the other hand, rich in alkalis (generally >8%), specially in K2O (generally >5%). Both granites are peraluminous; the HG is mainly moderately peraluminous (Alumina Saturation Index, A/CNK, = 1.06–1.18), whereas the SG is weakly peraluminous (A/CNK = 1.01–1.09).

In major element variation diagrams (Fig. 3), both granites show similar, poorly defined correlations. Fe2O tot3 , MgO and TiO2 decrease with increasing SiO2 suggesting fractionation of mafic phases, mainly biotite. Al2O3, CaO and P2O5 also decrease, suggesting fractionation of plagioclase and apatite, respectively, whereas Na2O and K2O do not correlate well with SiO2.

Fig. 3
figure 3

Variation diagrams of major oxides versus SiO2. In the K2O diagram the divisions are from Rickwood (1989)

Full size image

The HG and the SG can be distinguished well in an A/CNK versus SiO2 diagram (Fig. 4a) and in the A–B diagram of Debon and Le Fort (1983) (Fig. 4b), due to the different variations in peraluminosity: it decreases with differentiation in the HG, while it increases with differentiation in the SG. These opposite tendencies can be explained by fractionation of muscovite in the HG (which will strongly decrease the peraluminosity of the remaining melt due to its high peraluminosity) and the absence of this mineral in the SG (where the increase in peraluminosity is due mainly to the fractionation of plagioclase, whose A/CNK = 1).

Fig. 4
figure 4

a A/CNK versus SiO2 diagram; b A–B diagram of Debon and Le Fort (1983) with the peraluminous granite classification of Villaseca et al. (1998a): H-P, highly peraluminous; M-P, moderately peraluminous; L-P, low peraluminous; Felsic, felsic peraluminous. Symbols as in Fig. 3. In b fields correspond to the two facies of the San Blas granite: SB-B, border facies; SB-C, central facies (data from 13 analyses of Báez (2006))

Full size image

The LCS is weakly peraluminous (A/CNK = 1.05–1.06; except one sample with a value of 1.13) with elevated SiO2 contents (75.4–76.2%). It has very low contents of CaO, P2O5 and ferromagnesian oxides, specially MgO (Fig. 3). The alkali contents are similar to that of the HG and SG, although the LCS has higher Na2O and lower K2O concentrations (Fig. 3).

All three granites are ferroan (Fig. 5a), alkali-calcic to slightly calk-alkalic (Fig. 5b) and potassium-rich (Fig. 3). In the A–B diagram of Debon and Le Fort (1983), with the fields of Villaseca et al. (1998a) (Fig. 4b), the HG is classified as moderately peraluminous, the SG as low peraluminous and the LCS as a felsic peraluminous granite.

Fig. 5
figure 5

Granite classification diagrams of Frost et al. (2001). a FeOtot/(FeOtot + MgO) versus SiO2. b Na2O + K2O − CaO versus SiO2. Symbols as in Fig. 3

Full size image

The HG and the SG have similar overall trace element concentrations; high contents of several LIL (Li, Rb, Cs) and HFS (Nb, Ta, Th, U) elements and of Zn and Ga, and low contents of Sr and Ba. The HG has higher concentrations in Li, Rb, Cs and U, whereas the SG is richer in Sr, Ba, Y, Th and the REE. These differences suggest that the HG is more evolved than the SG, in accordance with their major element contents. The more evolved character of the HG is also indicated by its lower K/Rb ratios (HG = 115–165; SG = 155–233) and higher Rb/Sr ratios (HG = 3.6–8.0; SG = 1.8–5.9).

The LCS is characterized by elevated concentrations of many trace elements and very low concentrations of some others. Compared to the HG and the SG, the LCS is enriched in Li, Rb, Nb, Ta, U, Y and the HREE, and it is impoverished in Sr, Ba and Eu. The enrichment in U is especially noticeable, with values (18–69 ppm) much higher than that of normal granites (4 ppm, Rogers and Adams 1969). The LCS also has very low K/Rb ratios (54–98) and very high Rb/Sr ratios (16–72). Although F was not analyzed, the presence of fluorite suggests high F contents. The major and trace element contents of the LCS indicate that it is a highly evolved leucogranite.

The HG and the SG define similar patterns in a multi-element spider diagram normalized to the primitive mantle (McDonough et al. 1992) (Fig. 6). They both show Rb, Th and Ta positive anomalies, Ba and Sr negative anomalies, and smoothly decreasing values between La and Lu (Fig. 6). Compared to the composition of the crust of Taylor and McLennan (1985), the HG and the SG have higher concentrations in most elements (Rb, Th, U, K, Nb, Ta and the LREE), similar Ba, Zr, Y and HREE contents, and a lower Sr value (Fig. 6). The LCS shows, in general terms, a similar pattern as the HG and the SG, but is distinguished by the stronger U positive anomaly and Ba and Sr negative anomalies, and by the higher concentrations in Nb, Ta and specially Y and the HREE (Fig. 6).

Fig. 6
figure 6

Multi-element variation diagram, normalized to the primitive mantle (McDonough et al. 1992), of the studied granites (compositional fields) and of the continental crust (Taylor and McLennan 1985)

Full size image

The HG and the SG show similar chondrite-normalized REE patterns, although the SG has higher overall concentrations (Fig. 7). The REE of both granites are moderately fractionated, with (La/Yb) N values between 8 and 15. They both have moderate negative Eu anomalies (Eu/Eu* between 0.2 and 0.5), suggesting, together with the Sr and Ba negative anomalies (Fig. 6), plagioclase fractionation at the source. The LCS shows sub-horizontal, weakly fractionated patterns with (La/Yb) N values between 1 and 5, and very strong negative Eu anomalies (Eu/Eu* between 0.01 and 0.16).

Fig. 7
figure 7

Chondrite-normalized (Nakamura 1974) REE diagram of the studied granites (compositional fields)

Full size image

According to major and trace element geotectonic discrimination diagrams, the compositions of the HG and SG are generally compatible with a post-orogenic setting. In the major element diagrams of Maniar and Piccoli (1989), both granites mostly plot in the post-orogenic granite field (Fig. 8). In the trace element diagrams of Pearce et al. (1984), the HG and the SG typically plot on or close to the triple junction of the volcanic arc, syn-collisional and within-plate granite fields (Fig. 9). This area of the diagrams is typical of post-orogenic or post-collisional granite compositions, as shown by the post-collisional granite field added by Pearce (1996) in the Rb versus Y + Nb diagram (Fig. 9), and discussed by Förster et al. (1997). In these trace element diagrams, the LCS plots in the within-plate granite field (Fig. 9). Figure 10 shows that the geochemistry of the studied granites is only partially similar to that of A-type granites (e.g. Whalen et al. 1987; Eby 1992). The studied granites have high Ga/Al ratios and Nb concentrations, comparable to those of A-type granites (Fig. 10). However, their Zn, Y, Zr and Ce concentrations (and the Zr + Nb + Ce + Y sum, used by Whalen et al. 1987, as a discriminator) are lower than the values considered typical of A-type granites and more similar to those of I- and S-type granites (Fig. 10).

Fig. 8
figure 8

Major element geotectonic discrimination diagrams of Maniar and Piccoli (1989). a A/NK versus A/CNK. b FeOtot/(FeOtot + MgO) versus SiO2. Symbols as in Fig. 3. Fields: IAG island arc granites, CAG continental arc granites, CCG continental collision granites, POG post-orogenic granites, RRG rift related granites

Full size image
Fig. 9
figure 9

Trace element geotectonic discrimination diagrams of Pearce et al. (1984).a Rb versus Y + Nb. b Nb versus Y. Symbols as in Fig. 3. Fields: ORG, ocean ridge granites; VAG, volcanic arc granites; Syn-COLG, syn-collisional granites; WPG, within-plate granite; the post-collisional field (post-COLG) in a is from Pearce (1996)

Full size image
Fig. 10
figure 10

Whalen et al. (1987) discrimination diagrams of A-type granites. Fields: I-S, I- and S-type granites; Fel, I- and S-type felsic and fractionated granites. A is the composition of the average A-type granite according to Whalen et al. (1987). Dotted-line fields are the compositional fields of 16 A-type granites studied by Eby (1992). Symbols as in Fig. 3

Full size image

The San Blas granite, in the north of the sierra, is characterized by a more mafic (SiO2 = 68–71%), moderately peraluminous (A/CNK = 1.09–1.20) border facies and a felsic (SiO2 = 74–76%), weakly peraluminous (A/CNK = 1.02–1.08) central facies (Báez 2006) (Table 2; Fig. 4b). This granite presents the same alkali-calcic, ferroan and K-rich affinity as the studied granites. The trace and REE element contents of the San Blas granite are more similar to those of the LCS than to those of the HG and SG (Table 2). As with the LCS, the San Blas granite has high concentrations in Rb, Th, U, Nb, Ta, Y and the HREE, and plots mostly in the within-plate granite field in the Pearce et al. (1984) diagrams. It can be distinguished from the LCS by its higher LREE contents and less pronounced negative Eu anomaly (Table 2).

The four analyzed mafic enclaves have very variable major element compositions that can be attributed to different degrees of hybridization with the host granite. Two enclaves have low SiO2 contents of 55% and can be considered the least hybridized enclaves, while another has an intermediate SiO2 content of 63% and the fourth has a high SiO2 content of 70%. In comparison with the granites, the mafic enclaves have higher concentrations in ferromagnesian elements, CaO, Na2O and P2O5, and lower K2O contents. They are weakly to moderately peraluminous (A/CNK = 1.05–1.18). The mafic enclaves have trace element contents similar to the HG and the SG, although with higher concentrations in Zr, Y, the HREE and several transitional elements such as Sc, V and Zn (Table 2). They also produce similar REE patterns, but with somewhat lower (La/Yb) N values (5–10), due to higher HREE contents, and less pronounced negative Eu anomalies (Eu/Eu* = 0.25–0.34).

Nd isotopes

Table 3 shows 22 whole-rock Nd isotope analyses of the studied granites: eight analyses correspond to the HG, six to the SG, three to the LCS, two to mafic microgranular enclaves of the HG and SG, and three analyses are of the San Blas granite. The initial ɛNd values are calculated using the obtained U–Pb ages: 354 Ma for the HG, 353 Ma for the SG and 345 Ma for the LCS. For the mafic enclaves, the age of the host granite is considered, whereas for the San Blas granite the U–Pb SHRIMP age of 340 Ma obtained by Dahlquist et al. (2006) is used.

Table 3 Sm–Nd and Rb–Sr isotopic data of the studied granites and mafic enclaves of central-eastern Sierra de Velasco, and of the San Blas granite of northern Sierra de Velasco
Full size table

The HG and the SG have similar ɛNd values between −2.1 and −4.3 (average −3.3 ± 0.7). The LCS has somewhat higher ɛNd of −0.6 to −1.4, roughly comparable to the values shown by the San Blas granite (−1.4 to −1.7). The two analyzed mafic enclaves have even higher values of 0.6 and −0.6, suggesting a different, more primitive source.

Figure 11 is a diagram showing the initial ɛNd of the studied granites in comparison with values (recalculated to 350 Ma) of several potential crustal protoliths that outcrop in the Sierras Pampeanas (see figure caption for sources of data). It can be clearly seen that all of the potential crustal protoliths considered have lower ɛNd than the studied granites: the peraluminous S-type metagranitoids of Sierra de Velasco, and neighboring sierras of Copacabana and Fiambalá, have ɛNd between −5.5 and −7.4; the metaluminous I-type metagranitoids of southern Sierra de Velasco and Sierra de Chepes have ɛNd between −5.2 and −7.1; the low grade metasediments of the eastern flank of Sierra de Velasco and of other localities have ɛNd between −8.2 and −9.4; the medium to high grade metasediments of sierras of Aconquija and Quilmes have ɛNd between −4.7 and −9.6; and the medium to high grade metasediments of Sierra de Córdoba have ɛNd between −4.6 and −9.1. The more radiogenic ɛNd of the studied granites suggests that a more primitive component was involved in their genesis.

Fig. 11
figure 11

ɛNd(T) values (T = intrusion age) of the Huaco, Sangasta, La Chinchilla and San Blas granites and of two mafic enclaves included in the Huaco and Sanagasta granites, and the ranges of ɛNd(T) values (T = 350 Ma) of potential crustal protoliths that outcrop in the Sierras Pampeanas: peraluminous S-type metagranitoids of sierras of Velasco, Copacabana and Fiambalá (9 samples from Höckenreiner et al. 2003; our unpublished data); metaluminous I-type metagranitoids of southern Sierra de Velasco and Sierra de Chepes (12 samples from Pankhurst et al. 1998; our unpublished data); low-grade metasediments of the eastern flank of Sierra de Velasco and from other localities (7 samples from Pankhurst et al. 1998; Bechio 2000; Höckenreiner et al. 2003; our unpublished data); medium- to high-grade metasediments of sierras of Aconquija and Quilmes (7 samples from Bechio 2000; our unpublished data); medium- to high-grade metasediments of Sierra de Córdoba (12 samples from Rapela et al. 1998). See text for further explanations

Full size image

Nd model ages were obtained using one- and two-stage model calculations (Goldstein et al. 1984 and Liew and Hofmann 1988, respectively; see Table 3). The two-stage Nd model ages are considered more appropriate as it assumes an average Sm/Nd ratio for the crustal evolution prior to granite formation. Nd-model ages from crustal rocks can be seen as mixing ages representing different provenances of the source rocks. The two-stage model ages of the studied granites vary between 1,400 and 1,100 Ma (Table 3). These model ages are significantly lower than those determined for both I- and S-type Ordovician metagranitoids of neighboring ranges (∼1,500–1,700 Ma; e.g. Pankhurst et al. 1998; Höckenreiner et al. 2003) and also lower than the typical values of the Proterozoic–Lower Paleozoic basement of the Sierras Pampeanas and of the southern Central Andes (∼1,600–1,800 Ma, one-stage Nd model ages; e.g. Rapela et al. 1998; Bock et al. 2000; Lucassen et al. 2000). The higher ɛNd and lower Nd model ages of the studied granites compared to the Ordovician metagranitoids in particular, and the Proterozoic–Lower Paleozoic basement in general, can be explained by the participation of a Carboniferous depleted mantle-derived magma component in the generation of the granites. Taking into account that the non-detrital cores of zircons from the Huaco Granite are, without exception, Ordovician in age (Söllner et al. 2007), a Middle to Late Proterozoic crustal source for the studied granites can be discarded.

Sr isotopes

Sr isotopic ratios were determined for the same 22 samples analyzed for Nd (Table 3). The measured Sr ratios are very high and variable, defining pseudo-isochrons (not shown) with unrealistically old ages and low initial 87Sr/86Sr ratios. Even when recalculated for the time of crystallization using the U–Pb data, the initial 87Sr/86Sr ratios are high and variable (Table 3), suggesting incomplete Sr isotopic homogenization during melt generation and/or late- to post-magmatic disturbance of the Rb–Sr isotope system. Only the samples from the younger San Blas granite feature seemingly primary, homogeneous and low initial 87Sr/86Sr ratios (Table 3). Presentation of the data in a coupled Sr–Nd isotopic correlation diagram is inconsistent and confusing and hence, has been omitted.

The mafic enclaves have high initial 87Sr/86Sr ratios similar to their host granites (Table 3), suggesting an increase of their primary ratios due to interaction with the enclosing granitic melts. On the other hand, their ɛNd are considerably higher than those of the granites (Table 3; Fig. 11), suggesting that the enclaves were able to retain partly or totally their primary Nd isotopic ratios. This decoupling behavior between Nd and Sr isotopes is common in mafic enclaves (e.g. Holden et al. 1991; Lesher 1990, 1994; Pin et al. 1990; Allen 1991) and has been explained as a consequence of differences in diffusion velocities (e.g. Lesher 1994) or of retention of Nd in early crystallizing phases (e.g. Petford et al. 1996).

Discussion

Magma sources

The negative ɛNd values of the HG and the SG (between −2.1 and −4.3; see Table 3 and Fig. 11) suggests a mainly crustal source. In addition, the presence of non-detrital inherited cores with Ordovician ages in zircons of the HG (Söllner et al. 2007) strongly suggests that this crustal source consisted in Ordovician metagranitoids equivalent to those outcropping in the Sierra de Velasco. Other data compatible with a main crustal source for both granites are the high and restricted SiO2 contents; the peraluminous character; the high concentrations in several typically crustal LIL elements such as Li, Rb and Cs; the low contents in CaO and ferromagnesian elements; and the lack of associated mafic intrusions.

However, the ɛNd values of the HG and the SG are clearly higher (and their Nd model ages are clearly lower) than the values of several potential crustal protoliths of the Sierras Pampeanas, and in particular of both the I- and S-type Ordovician metagranitoids (ɛNd values between −5.5 and −7.3; see Fig. 11). Thus, the participation of a more primitive component (i.e. a component with ɛNd values higher than those of the granites) is required to satisfy the Nd isotopic ratios. Although the origin of this more primitive material is unknown, it can be speculated that it consisted in mantle-derived melts, injected into the crust during crustal extension and mantle upwelling in a post-orogenic setting (see Sect. 7.2). Other data compatible with the participation of a more primitive component are the presence of mafic microgranular enclaves; the zircon typology corresponding to calc-alkaline granites of mixed crust-mantle origin (Pupin 1980); the calc-alkaline composition of biotites (Nachit et al. 1985); and the high concentrations in many HFS elements such as Nb, Ta and Th.

Comparison of the chemical compositions of the HG and the SG with experimental melts derived from crustal rocks is consistent with the participation of a more primitive and mafic component. In the A–B diagram of Debon and Le Fort (1983) (Fig. 12), the compositions of the studied granites are plotted together with the fields of experimental melts obtained from several lithologies, their respective starting materials, and the compositional fields of possible crustal protoliths outcropping in the region (see figure caption for sources of data).

Fig. 12
figure 12

A–B diagram with compositions of the studied granites (symbols as in Fig. 3) compared with experimental melts (blank fields), their respective starting materials (rectangles), and potential protoliths outcropping in the region (solid fields). The experimental melts and starting materials are: P&H, felsic metapelites of Patiño Douce and Harris (1998); P&J, mafic metapelites of Patiño Douce and Johnston (1991); CAS, peraluminous metagreywackes of Castro et al. (1999); H&J, peraluminous orthogneiss of Holtz and Johannes (1991); CON, metaluminous dacite of Conrad et al. (1988). The black circles are the melts obtained by 50:50 assimilation at low and medium pressures of metagreywackes with a gabbro (Castro et al. 1999). Potential protoliths are: felsic metapelites from sierras of Velasco and Ambato (Espizua and Caminos 1979; our unpublished data); more mafic metapelites from Sierra de Ancasti (Willner et al. 1990); metagreywackes from Sierra de Ancasti (Willner et al. 1990); S-type metagranitoids (MG-S) from Sierra de Velasco (Rossi et al. 2005; our unpublished data); I-type metagranitoids (MG-I) from Sierra de Velasco (Bellos 2005). See text for further explanations

Full size image

The starting materials used in the felsic and mafic metapelite experiments are comparable, respectively, to the low-grade felsic metapelites and the higher-grade mafic metapelites that outcrop in the region (Fig. 12). The melts obtained from both types of metapelites are strongly peraluminous and leucocratic to mesocratic (Fig. 12). The strong peraluminosity of these melts and the generally lower CaO and ferromagnesian content is not compatible with the compositions of the HG and the SG. On the other hand, the metaluminous dacite used by Conrad et al. (1988), which is comparable to the I-type metagranitoids of the region, generates metaluminous to weakly peraluminous melts with varying ferromagnesian contents (Fig. 12). In the A–B diagram, these melts seem to have a somewhat similar composition to the SG (Fig. 12). However, the melts are much poorer in K2O and richer in Na2O and CaO than the SG, and, additionally, the ɛNd values of the I-type metagranitoids are clearly lower than those of the studied granites (see Fig. 11), implying that they can not be considered an exclusive source.

The metagreywackes used by Castro et al. (1999) and, in particular, the peraluminous orthogneiss used by Holtz and Johannes (1991) have similar compositions to the S-type Ordovician metagranitoids of the Sierra de Velasco (Fig. 12), which, as discussed above, can be considered the main crustal source of the HG and the SG. The experimental melts obtained from these lithologies are strongly to moderately peraluminous and leucocratic, their compositions differing from those of the HG and the SG (Fig. 12). The higher content in ferromagnesian elements and the lower peraluminosity of the HG and, in particular, the SG, compared to these experimental melts (Fig. 12), suggests additional participation of mafic material in the genesis of the granites. Castro et al. (1999) carried out melting-assimilation experiments between the metagreywackes and a gabbro as the mafic component. The resulting melts show a strong increase in ferromagnesian element contents and a decrease in peraluminosity that varies as a function of pressure. The hybrid melts produced at low-to-medium pressure are similar to the composition of the SG (Fig. 12). The composition of the HG could be obtained if (1) less quantity of the mafic component is assimilated; or (2) the assimilation process takes place at higher pressures. The later possibility is preferred to explain the compositional difference between the HG and the SG because both granites have similar ɛNd values and an equal abundance of mafic enclaves are observed in the field suggesting a similar proportion of crustal and mafic components. Different P–T conditions during fusion have been suggested, for example, by Villaseca et al. (1998b) to explain the compositional differences between two types of Hercynian peraluminous granites of Spain, referred to as PS- and PI-types, that have similar ɛNd values. The composition of these granites are comparable to those of the HG and the SG, respectively.

The common mafic enclaves found in the HG and the SG, with ɛNd values close to zero (see Fig. 11), can be considered remnants of the more primitive mafic component. A possible mechanism for the assimilation of mafic enclaves by a granitic magma has recently been experimentally tested by García-Moreno et al. (2006). They have shown that dissolution of mafic enclaves in a granitic magma is a viable mechanism, specially during ascent of the magma, producing an increase in CaO, FeO and MgO in the granitic magma. They suggest that this mechanism is responsible for the generation of the “mixed series” monzogranites of the Central Extremadura Batholith of Spain. The similar composition of the HG and the SG with these granites, as well as the absence of contemporaneous input of mafic magmas and the lack of enclave-host granite Nd equilibration (García-Moreno et al. 2006), all seem to favor such a process.

The obtained data thus suggest that both the HG and the SG have a mixed source, composed by a main crustal source, consisting in Ordovician peraluminous metagranitoids, and a minor more mafic and primitive source, possibly mantle-derived. The main metagranitoid-derived melts assimilated the more mafic and primitive material, thereby reducing the peraluminosity and increasing the ferromagnesian element contents and ɛNd values of the resulting melts. Hybridization between crustal and mantle magmas has been proposed by López de Luchi et al. (2001, 2007) as a mechanism for the generation of Devonian granites in the Sierra de San Luis, which possibly belong to the same magmatic event as the studied granites (see Sect. 7.3).

The mineralogy (presence of albite, zinnwaldite and fluorite) and geochemistry (e.g. high SiO2 and low CaO and ferromagnesian contents, high Rb/Sr and low K/Rb ratios, very strong negative Eu anomaly) of the LCS indicates that it is a highly evolved leucogranite and suggests fractionation by means of a F-rich fluid phase (e.g. Muecke and Clarke 1981; Dostal and Chatterjee 1995). Fluid fractionation with F as a complexing agent tends to produce strong enrichments in LIL and HFS elements (e.g. Taylor et al. 1981; Dostal and Chatterjee 1995; Irber 1999; Frindt et al. 2004), as is observed in the LCS (enriched in Li, Be, Y, Nb, Ta, U, the HREE). The LCS can be classified as a specialized granite, following the criteria of Tischendorf (1977), and is similar to the “P-poor” sub-type of topaz granites proposed by Taylor (1992) and to some A-type granites (e.g. Collins et al. 1982; Whalen et al. 1987).

The source of the LCS is difficult to determine due to its strongly evolved nature. This type of granite can be considered either as a final highly fractionated phase of a larger granite (e.g. Whalen and Currie 1990), i.e. the HG, or the product of melting of a different source (e.g. Manning and Hill 1990; Förster et al. 1995). The obtained data suggests that the LCS was generated from a different, more primitive source than the HG: its zircon typology is different from that of the HG and corresponds to that of alkaline granites of mantle origin (Pupin 1980); it has strong enrichments in many HFS elements such as Nb, Y, Pb and the HREE that do not follow the trends of the HG; and it has considerably higher ɛNd values (see Fig. 11).

The LCS and the San Blas granites share many characteristics (presence of fluorite, similar zircon typology, highly evolved nature, strong enrichments in many trace elements, high ɛNd values) which suggest that both of these granites originated from a similar source, more primitive or with a greater abundance of mantle-derived components compared to HG and the SG. The LCS and the San Blas granite are younger than the HG and SG by about 10–15 Ma, suggesting an increasing participation of primitive material with time during magma generation.

Geotectonic setting

Field relationships and the obtained Lower Carboniferous (∼350 Ma) ages of the HG, SG and LCS clearly indicate that they are younger than the main Ordovician Famatinian magmatism (∼460–490 Ma) and the deformation events that regionally affected the Ordovician metagranitoids and generated afterwards the spatially restricted mylonitic shear zones (e.g. Höckenreiner et al. 2003; Miller and Söllner 2005). If the age of 402 ± 2 Ma determined by Höckenreiner et al. (2003) is taken as the age of formation of the mylonites, and if these are interpreted as the result of a final compressive phase of the Famatinian orogeny (e.g. Miller and Söllner 2005), then the studied granites were generated approximately 50 Ma after this final compressive phase. This considerable age difference suggests that the studied granites are not orogenic (and thus are not related with active margin, collisional or post-collisional settings), but were instead generated during either a post-orogenic or anorogenic period, in both cases in a within-plate setting (following the classification scheme of geotectonic periods and settings of Liégeois 1998).

The geochemistry of the studied granites is more compatible with a post-orogenic period than with an anorogenic one. Their homogeneous compositions, high and restricted SiO2 contents, high-K and alkali-calcic to calc-alkalic affinity, low to moderate peraluminosity and their generally moderate LILE and HFSE enrichments are characteristic of post-orogenic granites, such as the “Caledonian” granites of Pitcher (1983), the “post-orogenic” granites of Maniar and Piccoli (1989) and the “high-K calc-alkaline” granites of Barbarin (1999). In particular, the studied granites are similar to the “post-orogenic alkaline granites (PA-type)” of Dawei et al. (1996) and to the “post-orogenic alkali-calcic granites with moderate LILE and HFSE enrichments” of Bonin et al. (1998).

The studied granites, specially the LCS and the San Blas granite, have several characteristics of A-type granites, such as a ferrous tendency (high Fe/Mg ratios), low CaO, MgO, Sr and Ba contents and high concentrations in certain trace elements such as Nb, Ta, HREE and Ga (e.g. Collins et al. 1982; Whalen et al. 1987; Eby 1990, 1992). However, the granites do not have alkaline tendencies, they are not metaluminous or peralkaline, and their HFS element concentrations are generally not high enough (in the sense of Whalen et al. 1987; see Fig. 10). Therefore, the studied granites are more comparable with post-orogenic A-type granites than with “true” anorogenic A-type granites. The partial similarity with A-type granites can be caused by the mixed origin of the studied granites, with some participation of mantle-derived material, which is considered predominant during the anorogenic period (e.g. Bonin 2004). The more primitive nature of the younger LCS and San Blas granite compared to the HG and the SG suggests an increasing mantle involvement with time and could be indicating a transition from a post-orogenic towards an anorogenic period.

The generation of the studied granites during a post-orogenic period implies a relationship with the previous orogeny. Thus, it can be considered that the studied granites were generated as a consequence of the extensional collapse of the Famatinian orogen. An extensional post-orogenic period, established during the collapse of the orogen (e.g. Dewey 1988), is favorable for the generation and emplacement of large granitic bodies (e.g. Barbarin 1999; Bonin 2004). Crustal extension during this phase favors mantle upwelling via slab-breakoff or slab-distortion (e.g. Davies and von Blanckenburg 1995), or delamination (e.g. Black and Liégeois 1993), which in turn favors crustal melting and magma generation due to its heat input. The crustal melts can also assimilate some of the mantle-derived material, as would have been the case of the studied granites. A further heat input could have been radiogenic heat from the previously thickened crust, as has been suggested for the generation of Variscan granites in Iberia (Bea et al. 1999) and Bohemia (Gerdes et al. 2000).

Regional implications

The obtained Lower Carboniferous ages of the HG, SG and LCS permits to correlate them with the Carboniferous San Blas and Asha granites of northern Sierra de Velasco (Fig. 1b). According to the existing U–Pb age data, these two granites have ages similar to the LCS, and they are approximately 10–15 Ma younger than the HG and the SG. Taken together, all of these granites occupy a total area of approximately 1,500 km2 (25–30% of the Sierra de Velasco), demonstrating the importance of Lower Carboniferous magmatism in this region of the Sierras Pampeanas.

The obtained ages also allow to include the studied granites in the Middle Devonian to Lower Carboniferous granite group of the Sierras Pampeanas (i.e. the post-Famatinian or Achalian granite group). These granites are located along the previous Ordovician Famatinian magmatic arc and also to the east, both in central and eastern Sierras Pampeanas (Fig. 13). The HG and the SG share many characteristics with the granites of this group, such as nearly circular shapes, general lack of pervasive solid-state deformation, discordant relationships with the host rocks, shallow emplacement in the upper crust, syeno- to monzogranitic compositions, high SiO2 and K2O contents, porphyritic textures and abundance of pegmatites and aplites (e.g. Lira and Kirschbaum 1990; Grissom et al. 1998; Llambías et al. 1998; Pinotti et al. 2002; López de Luchi et al. 2007). These similarities suggest a common origin for the granites of this group. They were possibly generated during a regional crustal heating event (e.g. Miller and Söllner 2005) that occurred during a Late Devonian–Early Carboniferous post-orogenic period.

Fig. 13
figure 13

Distribution of the Middle Devonian–Lower Carboniferous granites of the Sierras Pampeanas with existing U–Pb ages suggesting northward migration of the magmatism in a direction approximately parallel to the Ordovician Famatinian arc (modified from Pankhurst et al. 2000) and the main mylonitic shear zones

Full size image

Granites with Devonian–Carboniferous ages confirmed by U–Pb geochronology are found in two main areas of the Sierras Pampeanas: southeast (sierras of Córdoba and San Luis) and north-central (sierras of Velasco and Fiambalá) (Fig. 13). In addition, other granites possibly belonging to this group crop out in the sierras of Vinquis and Zapata (e.g. Rapela et al. 1996), and in the sierras of Los Llanos, Chepes and Ulapes (according to unpublished field observations and mapping carried out by one of the authors, F. Söllner) (Fig. 13), although U–Pb age determinations of these granites are lacking.

The scarce existing U–Pb geochronological data suggests that the spatial distribution correlates with a temporal variation (Fig. 13). The southeastern granites have older, Middle to Upper Devonian ages, whereas the north-central granites have younger, Lower Carboniferous ages (Fig. 13). Even within the north-central area, there seems to be a general age decrease towards the north (Fig. 13). These U–Pb age variations suggest a possible northward migration of the magmatic activity, in a direction approximately parallel to the Ordovician Famatinian arc and to the main shear zones (Fig. 13). This possible migration could have been related to the mechanisms of magma generation. For example, it could be associated with a progressive delamination of the crust accompanied by upwelling of the upper mantle from south to north. The velocity of this magmatic migration can be calculated to about 10 km/Ma over a distance of more than 600 km and a time span of ∼70 Ma. Further geochronological and geochemical data of this granite group is needed to verify this magmatic migration and identify its causes.

On the other hand, the studied Carboniferous granites cannot be deemed part of the Gondwanan Cycle (considered to extend from the Carboniferous to the Cretaceous; e.g. Pankhurst and Rapela 1998) because the Carboniferous Gondwanan granites (e.g. Llambías 1999), (1) are located far to the west, in the Cordillera Frontal, Precordillera and western Sierras Pampeanas; (2) are younger, mostly of Upper Carboniferous age; and (3) define a granodioritic to tonalitic calc-alkaline series typical of a subduction setting.

Conclusions

The data presented here allow to draw the following conclusions:

  1. 1.

    The central-eastern part of the Sierra de Velasco is composed by the undeformed Huaco and Sangasta granites, and the La Chinchilla stock, of Lower Carboniferous age. These granites, together with the Carboniferous granites of the north of the sierra, make up about 25–30% of the sierra, indicating the importance of Carboniferous magmatism in this area, as opposed to previous studies which considered that the Sierra de Velasco consisted mostly in Famatinian inner-cordilleran S-type granites of Ordovician to Lower Devonian age (e.g. Rapela et al. 1992; Toselli et al. 1996; Pankhurst et al. 2000).

  2. 2.

    The HG and the SG have a mixed source, mainly crustal but with participation of a more primitive, possibly mantle-derived, component. The main crustal component can be attributed to Ordovician peraluminous metagranitoids such as those outcropping around the granites. The LCS, and the San Blas granite, derive from a more primitive source, suggesting an increase with time in the participation of the primitive component during magma genesis.

  3. 3.

    The studied Lower Carboniferous granites were generated during a post-orogenic geotectonic period in a within-plate setting. They were produced during a regional crustal heating phase in response to the collapse of the previous Famatinian orogen, extension of the crust and mantle upwelling.

  4. 4.

    The studied granites are part of the group of Middle Devonian–Lower Carboniferous granites of the Sierras Pampeanas. Many of these granites were probably generated during a post-orogenic period. The distribution and U–Pb ages of these granites suggest a northward arc-parallel migration of the post-orogenic magmatism with time.