Integrated U-Pb and Hf zircon and whole-rock Nd isotopes studies of Devonian granitic rocks from Sierra de San Luis (Sierras Pampeanas, Argentina): Petrogenetic implications

. Previous geochronological data indicate a protracted Devonian magmatic activity developed in the Sierra de San Luis, Sierras Pampeanas of Argentina, with three major crystallization events: 393±3, 384±2, and 377±2 Ma. Previously reported whole-rock Sm-Nd isotopes data define two average distinctive ε Nd t values: -1.37 and -3.47, and they are consistent with new data presented here. The first signature is assumed for a parental magma with dominant metasomatized subcontinental lithospheric mantle (SCLM) source, whereas the second signature could represent a parental magma derived of a lower continental crust source hybridized with magmas of the first signature. Notably, the new zircon Hf isotopes performed on the same zircon domains that were previously dated, indicate that the contribution of the source was variable over time. In situ Hf in zircon is relevant to evaluate the compositional evolution of the Devonian granitic magmas in the Sierra de San Luis, since the high variability of the ε Hf t values recorded in zircons indicate that the calculated ε Nd t values for the samples can only be interpreted as a final picture of the petrogenetic process. Zircon Hf isotopes data suggest that the zircon crystallized from a magma with variable composition, recording two major events, yielding two ε Hf t signatures: (1) -3.54 and (2) -6.85. A third composition, yield a less representative ε Hf t value of -5.44, and represent a ε Hf t signature (3).

LA-MC-ICP-MS zircon U-Pb ages from a sample (CHA-101) of the largest granitic unit from the Las Chacras-Potrerillos batholith, Sierra de San Luis, yield threes distinctive ages (outer error range): 393±3, 384±2, and 377±2 Ma, suggesting a protracted magmatic activity with three major crystallization events for this magmatism (Dahlquist et al., 2019). Dahlquist et al. (2019) indicate that the older Late Devonian ages could indicate recycling of antecrystic zircons formed during early magma crystallization during the construction of a longlived magma reservoir, while the youngest age is considered the emplacement age. This postulated protracted magmatic activity is appropriated to consider a prolonged zircon crystallization recording different magma compositions.
In this paper, we report the first study integrating previous in situ U-Pb and new Hf isotope data from magmatic zircon together with previous and new whole-rock Sm-Nd isotope data for two granitic rocks of the Las Chacras-Potrerillos and Renca batholiths, Sierra de San Luis, Argentina. This study is relevant to evaluate the compositional changes using Hf values of zircon crystallized in granitic magmas, derived from two potential sources previously identified using whole-rock Nd isotopes data.

Devonian granitic foreland magmatism in Sierra de San Luis
A Devonian foreland magmatism is located between 31° 00′ and 33°30′ S, in the present-day Sierras de Córdoba and Sierra de San Luis of Argentina. (Fig. 1). This Devonian foreland magmatism is characterized by development of large intracontinental batholiths and the absence of arc magmatism on the western margin of the plate at this latitude (Dahlquist et al., 2021). Emplacement of these magmas took place mainly into metamorphic basement formed

Introduction
A feature of many studies is an implicit assumption that all zircons present in the host igneous rock are autocrysts, that is, crystallized from the surrounding melt. However, it has long been recognized that zircons present in an igneous rock can be inherited either from the surrounding country rock or source region (xenocrysts), or from earlier stages of magmatism in the magmatic source or in the plumbing system (antecrysts) (e.g., Miller et al., 2007).
Numerous works use different age populations of zircon to calculate an age, although individual ages show remarkable variations, displaying a relevant difference of ca. 15-20 Ma. This range of ages strongly suggest the presence of autocrysts and antecrysts in the crystallized rocks as indicated for different works (e.g., Miller et al., 2007;Siégel et al., 2018;Dahlquist et al., 2019;Moreno et al., 2020). The presence of antecrysts suggests that the origin and emplacement processes of granitic magmas can last much longer than a million years. It has been recently shown by high precision geochronology that many plutons have been assembled in several magmatic pulses that can last several million years (e.g., Tuolumne suite (USA), Coleman et al., 2004; the Altiplano-Puna Volcanic Complex (Argentina-Bolivia-Chile), De Silva and Gosnold, 2007;Queensland (Australia), Siégel et al., 2018;Cordillera Real (Bolivia), Iriarte et al., 2021). In this context, more precise geochronological studies are needed to understand this problem and ultimately the development of the continental crust. On the other hand, during this extended petrogenetic process different chemical and/or isotopic changes are expected for the magmas, like those suggested by Kemp et al. (2007) for the granitoids of the Lachland Fold Belt, SW Australia, where the zircon of a same rock display a notable spectrum of εHf t values (up to 10ε units).
FIG. 1. Simplified regional geological map of Sierra de San Luis displaying the studied granitic pluton included in the Las Chacras-Potrerillos and Renca batholiths and eastern plutons (modified from López de Luchi et al., 2017;Morosini et al., 2017;Dahlquist et al., 2019). The studied samples were georeferenced using GIS software. Geochronological data for the studied samples and those referred in the text are included. In violet letter, geochronological data for the studied samples in this work. Previous geochronological data: *Muñoz et al. (2022), ** Stuart-Smith et al. (1999). In bracket, range of determined age values. Granite and Monzonite suites are from López de Luchi et al. (2017). In blue letter samples with whole-rock Sm-Nd data reported in this work. Inset: Simplified regional geological map of central NW Argentina, showing the Devonian and Carboniferous magmatic development of the pre-Andean margin of SW Gondwana (Dahlquist et al., 2018a(Dahlquist et al., , 2021  during the Early Paleozoic Pampean and Famatinian orogenies (e.g., Steenken et al., 2011;Casquet et al., 2018;Muñoz et al., 2022). Devonian granites were discordantly emplaced in medium to high grade metamorphic country rocks (e.g., Rapela et al., 1998;Christiansen et al., 2019) overprinted by a strong shear deformation leading to formation of a major mylonitic belt (Siegesmund et al., 2004;Semenov and Weinberg, 2017 and references therein). Integrated petrological, geochemical, isotopic, and U-Pb zircon geochronology data of this Devonian foreland magmatism of Sierras Pampeanas of Argentina, represented for the granitoids of Sierras de Córdoba and San Luis (Fig. 1), have been discussed in previous works (e.g., López de Luchi et al., 2017;Muñoz et al., 2022;Dahlquist et al., 2019Dahlquist et al., , 2021 and references therein). Recently, a geodynamic framework for the generation of the Devonian magmatism foreland was postulated by Dahlquist et al. (2021), and an overview of the origin of these magmas can be found in that work.
According to López de Luchi et al. (2017), the granitic rocks of the Sierra de San Luis define two distinctive suites, called Monzonite suite (<65 wt%) and Granite suite (>65 wt%), emplaced in shallow conditions, ranging from 3.3 to 4.7 kbar (Iannizzotto and López de Luchi, 2012;Muñoz et al., 2022). As noted by López de Luchi et al. (2017), classification of the studied Devonian granitoids is subject to debate because they are I-to A-type hybrid granites. Whole-rock Nd isotopes data have been reported by some authors (e.g., López de Luchi et al., 2017;Dahlquist et al., 2019), but zircon Hf isotopes data remain absent for these granitic rocks. Based on whole-rock geochemistry and Sm-Nd isotopes data, previous work identifies two sources for the parental magmas: 1) metasomatized subcontinental lithospheric mantle (SCLM) and 2) a lower continental crust source hybridized with magmas derived from SCLM (López de Luchi et al., 2017;Dahlquist et al., 2019).
The first U-Pb SHRIMP zircon age was reported by Stuart-Smith et al. (1999) for a granitoid of the Renca batholith. Subsequently, U-Pb SHRIMP and LA-MC-ICP-MS zircon ages were reported by Dahlquist et al. (2019) for granitic rocks of the Las Chacras-Potrerillos and Renca batholiths, respectively. Recently, a U-Pb LA-MC-ICP-MS zircon age for the El Hornito pluton was reported by Muñoz et al. (2022).
As noted in Section 1, a common feature of a studied granitic rock from the Las Chacras-Potrerillos batholith is that the individual U-Pb zircon ages mostly vary between ca. 393 and 377 Ma, displaying a relevant time spam of ca. 15 My (previous and new U-Pb zircon LA-MC-ICP-MS data are indicated in Fig. 1). The SHRIMP zircon age reported by Stuart-Smith et al. (1999) is a weighted mean age of 393±5 Ma, with individual values ranging from 405 to 381 Ma, whereas the LA-MC-ICP-MS zircon age of Muñoz et al. (2022) is a Concordia age of 385±2 Ma with a moderately high MSWD=2.2 (1σ, recalculated age to 2σ is 385±5, n=9, with MSWD=2.1, and it is shown in Fig. 1), with individual values varying between 396 and 375 Ma that evidence a relevant difference of ca. 20 My. Notably, the individual U-Pb zircon values reported by Stuart-Smith et al. (1999) are mostly close to 400 Ma, but similar individual values have not been found in our analyses and are also absent in the work of Muñoz et al. (2022).

Analytical methodology
New Nd isotopic analyses for the samples CHA-216, CHA-218, and REN-223 (location in Fig. 1), representative of the main granitic units of the Las Chacras-Potrerillos and Renca batholiths, were carried out at the Geochronology and Isotope Geochemistry Centre of the Complutense University of Madrid, Spain. Isotopic analyses were made on an automated multicollector TIMS-PhoenixR mass spectrometer. Analytical uncertainties are estimated to be 0.006% for 143 Nd/ 144 Nd and 0.1% 147 Sm/ 144 Nd. Replicate analyses of the JNdi-1 Ndisotope standard yielded an average 143 Nd/ 144 Nd ratio of 0.512108±0.000012 (2σ) with n=6. 143 Nd/ 144 Nd was normalized to 146 Nd/ 144 Nd=0.7219. Sm, and Nd concentrations in ppm were determined by ICP-MS to calculate the 147 Sm/ 144 Nd ratios.
SiO 2 , and Sm and Nd data for the samples CHA-218, CHA-216, and REN-223 are reported in table 1, were analyzed at the Geosciences Institute of the University of Campinas (UNICAMP), Brazil. The analyses were led on a Philips PW 2404 X-ray fluorescence spectrometer and on a Thermo (Xseries2) quadrupole ICP-MS equipped with Collision Cell Technology (CCT), respectively, using procedures described in Cardoso et al. (2019). Analytical information can be found in Cardoso et al. (2019). Previous and new Sm-Nd data are reported in table 1. The decay constant used in the calculations is λ 147 Sm=6.54×10 −12 year −1 recommended by the IUGS Subcommission for Geochronology (Steiger and Jäger, 1977 (2017), excluding isotopes data of Sato et al. (2001) reported in that work because the location of the analyzed samples is unknown and the Devonian age is uncertain (see Morosini et al., 2017, Fig. 2  Zircon Hf-isotopes data were obtained from the samples CHA-101 and REN-103 (Granite suite) previously dated by Dahlquist et al. (2019). Description about the characteristics of the analyzed zircon grains as well as the separation and concentration mineral is carried out in Dahlquist et al. (2019). In situ LA-MC-ICP-MS Lu-Hf isotope analyses were also conducted at the Geochronological and Isotopic Geochemical Research Centre, Sao Paulo University, Brazil using a Laser Analyte Excite-Photon Machines (Teledyne) -193 nm coupled to a Thermo-Finnigan Neptune MC-ICP-MS with nine Faraday collectors. The Lu-Hf isotopic analyses reported here were performed on the same zircon domains that were previously dated. Complete analytical description as well as typical laser operating conditions, analysis routine, correction data from Morales Cámera et al. (2020). Lu-Hf results are reported in table 2.

U-Pb LA-MC-ICP-MS zircon ages
As referred in Section 1, a common characteristic of the studied sample CHA-101 (as well as other samples referred in Section 2) is the relevant time spam of ca. 15 My display between the individual U-Pb zircon analyses. Therefore, we report combined diagrams as suggested by Siégel et al. (2018) to discriminate age populations, in order to identify zircon antecrysts and autocrysts (e.g., combined linearized probability plot and weighted mean age diagrams). To distinguish autocrysts from antecrysts (or xenocrysts), we use a two-stage approach. First, the number of zircon age populations are defined using linearized probability plots to identify younger and older age analyses. Any data points that are not within the calculated regression line are excluded from the populations. Linearized probability plots therefore provide a visual basis whereby multiple zircon age populations can be recognized. Is relevant to clarify that for discriminating age populations it is mandatory to work with a large group of data, thus the obtained results will have statistical significance. Subsequently, weighted mean ages of each zircon age population can then be calculated and interpreted as autocrysts, antecrysts or xenocrysts. Three age groups from sample CHA-101 were used by Dahlquist et al. (2019) to calculate three consistent weighted mean ages (outside error limits): 379±2, 385±2, and 392±2 Ma ( Fig. 2A-C). Notably, when a weighted mean is calculated using all the individual U-Pb zircon data from sample CHA-101, an unacceptable MSDW=13 is obtained (Fig. 2D).
Using previous geochronological data from Dahlquist et al. (2019) we report new linearized probability plot (Isoplot/Ex 4.15 Ludwig, 2008) diagrams ( Fig. 3A-D), where three age groups as those reported by Dahlquist et al. (2019) are clearly distinguished using now mathematical support. In addition, a linearized probability plot diagram considering the whole data-set from sample CHA-101 yields a regression line with a relatively high error (±14 Ma) for the calculated slope (Fig. 3D). The figure 3A-D reported in this work gives robustness to previously calculated ages, and confirm a protracted Devonian magmatic event in Sierra de San Luis.

Sm-Nd whole-rock isotope data
Previous whole-rock Sm-Nd isotope data reported by López de Luchi et al. (2017)

Hf in zircon isotope data
Using previous U-Pb zircon crystallization ages referred to above, εHf t values were calculated from the new Lu-Hf zircon isotope data reported in table 2. Zircons from CHA-101 and REN-103 have variable but negative εHf t values, ranging from -12.61 to -1.57 and from -4.75 to -2.46, respectively (Table 2). In the case of sample CHA-101, although all εHf t values are negative, distinctive εHf t values are obtained from the different individual ages (Fig. 4): εHf t values mostly range from -1.57 to -4.27 (n=4, a single εHf t value yield -8.76) for ages ranging from 376 to 379 Ma, and from -4.53 to -8.49 (n=5, a single εHf t value yield -12.61) for ages ranging from 380 to 386 Ma. Older ages ranging from 391 to 393 Ma yield contrasting εHf t values: -9.56, -3.65, and -2.94 (Table 2).
Two ages of 393±3 and 353±4 Ma were determined for the sample REN-103. The first age was interpreted as a crystallization age, whereas the younger age as the one resulted from a reheating event at ca. 350 Ma (see Dahlquist et al., 2019 and Section 5.1). The εHf t values from the younger individual ages (i.e., ~350 Ma) vary between -3.77 and -5.61, and they are comparable to the εHf t values reported for the Devonian crystallization ages of the samples CHA-101 and REN-103 (Table 2 and

U-Pb LA-MC-ICP-MS and SHRIMP zircon ages
Previous analyses (Section 4.1.) based on geochronological data, suggest a protracted magmatic activity during the building of the Las Chacras-Potrerillos batholith, and ongoing studies are focused on determining if this process is applicable to the Devonian magmatism of the Sierra de San Luis. This interpretation assumes a prolonged time in a deep hot zone maturation with a subsequent migration and emplacement of the granitic magma in shallow levels, as postulated by different authors (e.g., Kemp et al., 2007;Alasino et al., 2017;Macchioli Grande et al., 2020). This conceptual model is consistent with the studies of Dahlquist et al. (2018b) on the Veladero granitic stock in Western Sierras Pampeanas of Argentina (Fig. 1), which indicate that the granitic magma passed through ∼10 km of continental crust (from ca. 17 to 7 km) before reaching its final emplacement level at shallow conditions.
On the other hand, as noted by Dahlquist et al. (2019) the younger age of 353±4 in the sample REN-103 (Renca batholith) is comparable to previous ages determined in micas crystallized in intragranitic pegmatites as well as cooling ages from biotite hosted in granitoids of the Las Chacras-Potrerillos batholith, being interpreted by Dahlquist et al. (2019) as the result of a subsequent heating event rather than a crystallization age. In particular, previous K-Ar ages from biotite crystallized in the Devonian granitoids of Sierra de San Luis, reveal cooling ages ranging from ca. 375 to 345 Ma (López de Luchi et al., 2017). Therefore, younger ages than ca. 375 Ma are assumed as ages affected by Pb loss, with subsequent resetting of the isotopic clock.  figure 4 for sample CHA-101, with starts representing averages. In the average calculation for the εHf t signature (2), the εHf t value of -12.61 (Table 2) was exclude because this value was not considered in the weighted mean diagram of figure 4F (see discussion in the text). An "anomalous" εHf t value of -15.45 (Table 2) for the sample REN-103 is not projected.

In situ Hf zircon and whole-rock Nd isotopes
Based on whole-rock geochemistry and Sm-Nd isotopes data, previous works identify two sources for the parental magmas: 1) metasomatized subcontinental lithospheric mantle (SCLM) and 2) a lower continental crust source hybridized with magmas derived of SCLM (López de Luchi et al., 2017;Dahlquist et al., 2019). As described in Section 4.2, the average εNd t value for the Monzonite suite is -1.37, and two distinctive εNd t values are recognized for the samples of the Granite suite: -3.47 and -1.32. The second value is only observed in three samples (n=17) being undistinguishable of the εNd t values reported for the Monzonite suite, and could indicates greater participation of the SCLM source in the hybridization process for these samples of the Granite suite.
The protracted magmatic activity postulated by Dahlquist et al. (2019) and verified in this work, along with the contribution of two sources for the parental magma of the Granite suite, is appropriated to postulate a prolonged zircon crystallization recording different magma compositions.
Furthermore, the figure 4A-B show that during the prolonged petrogenetic process the magma underwent Hf isotopic compositional changes, with variable εHf t values for the analyzed zircon of the sample CHA-101, which displays a single εNd t value of -3.12 as shown in figure 5. Therefore, the εNd t value for this sample CHA-101 can only be interpreted as a final picture of the petrogenetic process, and the compositional changes (and participation of different sources) cannot be assessed. Similar conclusions were obtained by Kemp et al. (2007) for the granitoids of the Lachlan Fold Belt of SW Australia, where the zircon of the studied samples derived from two postulated sources (SCLM and continental crust) yield different zircon Hf and O values for a given εNd t value.
The zircon Hf values determined from sample REN-103 (Renca batholith) are insufficient to carry out a mathematical analysis like the one carried out for sample CHA-101 (Las Chacras-Potrerillos batholith). However, the epsilon Hf t for zircons of sample REN-103 can be compared with those εHf t values reported above for the sample CHA-101. The epsilon Hf t of the sample REN-103 shows values comparable with the εHf t of the sample CHA-101 (signature 1 and 3), but the εNd t is -1.12 (Fig. 5). The limited analysis number from zircon of the sample REN-103 does not allow deeper interpretations, although preliminarily a dominant signature represented for (1) and (3) could be considered as well as the presence of a relatively homogeneous source, but excluding the signature (2), which indicate a marked contribution of continental crust.
The zircon εHf t values for the younger ages from sample REN-103 ranging from 343 to 356 Ma, are comparable to those values reported for magmatic zircon (Fig. 5). This indicates that the U-Pb system would have been affected by the heating event but the 177 Hf/ 176 Hf ratio remained mostly unchanged, in a similar way as suggested by different works (e.g., Farina et al., 2014;Dahlquist et al., 2020 and references therein). As it is known Hf and Pb are strongly compatible and incompatible in zircon, respectively. Consequently, it is expected that Hf will be preserved and Pb excluded from zircon grains, leading to Pb loss without eventual modification of the 177 Hf/ 176 Hf ratio.

Conclusions
1. The geochronological data strongly suggest the presence of zircon antecrysts and autocrysts with distinctive Hf composition. Combined individual U-Pb zircon ages and Hf analysis are relevant to evaluate the compositional evolution of granitic magmas in Sierra de San Luis. 2. Whole-rock Nd isotope data permit distinguishing two εNd t signatures for the granitic rocks of the Las Chacras-Potrerillos and Renca batholiths: ca. -1.37 and -3.47. However, Hf zircon data from sample CHA-101 indicate that the composition of the magma was not uniform and underwent variable isotopic changes during the petrogenetic process.
In addition, it also seems to suggest the variable participation of different sources. In particular, the calculated εNd t values for the samples CHA-101 and REN-103 can only be interpreted as a final picture of the petrogenetic process. 3. In general, two main εHf t signatures were recognized from sample CHA-101: (1) -3.54 and (2) -6.85. A less representative εHf t signature (3) of -5.44 is also recognized. The εHf t signature (1) is mostly dominant in older and younger ages (376-379 and 391-393 Ma), whereas signature (2) and (3) (1) and (3) and suggest less compositional variation than sample CHA-101 during the petrogenetic process. A relatively homogeneous source, but excluding signature (2), could be inferred for the sample REN-103. 4. The age of 353 Ma calculated from some zircons of sample REN-103, indicates that the U-Pb system would have been affected by a heating event, but the 177 Hf/ 176 Hf ratio remained mostly unchanged.