Magmatic source, petrogenesis, and tectonic setting of the Concepción del Oro Igneous Complex: a geochemical and isotopic (Rb–Sr and Sm–Nd) study of a late complex of the Cretaceous–Eocene Mexican Magmatic Arc

During the Mexican fold-and-thrust belt tectonics, the inboard migration of long-term cordilleran magmatism typifies the geological setting during the Cretaceous–Paleogene period. The Concepción del Oro Igneous Complex (COIC) represents one of the most internal and isolated magmatic pulses that occurred during this magmatic activity. In this work, plutonic rocks from the COIC were studied using petrography, U–Pb geochronology, whole rock and zircon geochemistry, and Rb–Sr and Sm–Nd isotopes. Petrographic observations reveal mainly fine-to-coarse-grained granodioritic composition for most plutonic masses. U–Pb zircon analyzes of four samples from the largest plutonic center of the COIC yielded ages ranging from 42.6 ± 0.2 Ma to 41.8 ± 0.1 Ma. The granodiorites exhibit geochemical features typical of I-type, high-K calc-alkaline, Magnesian rocks. The chondrite-normalized REE diagram summarizes enrichment of LREE ([La/Yb]N = 5.94–19.19) and moderate to slightly negative Eu anomalies (Eu/Eu* = 0.63–0.94). In addition, the mantle-normalized spider diagram shows enrichment of LILE, such as Rb and Ba, and depletion of HFSE, such as Nb, Ta, Ti, and P. Whole-rock and zircon geochemistry indicate a post-collisional setting with an inherited magmatic arc fingerprint. Ti-in-zircon temperatures and zircon saturation thermometry suggest a magmatic origin from evolved and relatively cold melts (~ 700 °C). Moreover, the systematic variations in the ratios of highly incompatible elements, such as La/Sm (6.54–3.79) and Rb/Sr (0.32–0.10), and relatively narrow ranges in Zr/Hf (36.25–48.84) and 147Sm/144Nd (0.0941–0.1261), suggest fractional crystallization as the main petrogenetic process involved in the formation of the COIC rocks. Whole-rock isotopic data suggest a magma source mixing juvenile material with older continental crust, as indicated by 87Sr/86Sr(t = 40) ratios between 0.704583 and 0.707783, 143Nd/144Nd(t = 40) ratios varying in the ranges of 0.512478–0.512702 (εNd(t = 40) = from + 2.18 to − 2.10), and TDM(Nd) ranging from 1.03 to 0.62 Ga. We conclude that the parental source of the COIC was derived from partial melting of the lower crust triggered by emplacement of an underplated magma at the base of the crust during the end of an orogenic cycle.


Introduction
The rocks grouped under the generic term "granitoids" are the major lithologic component of the Earth's continental crust. Notwithstanding this fact, high-K calcalkaline (HKCA) type rocks require special attention, not only because they are spatially and temporally widespread (e.g., Liégeois et al. 1998;Rottura et al. 1998;Ferré and Leake 2001;Karsli et al. 2010Karsli et al. , 2007Jiang et al. 2013;Li et al. 2013;Deschamps et al. 2017;Wu et al. 2018;Liu et al. 2019;Litvinovsky et al. 2021), but also because their petrogenesis and geological settings are still the subject of considerable controversy. Roberts and Clemens (1993) affirmed that HKCA geochemical fingerprint could be derived mainly from the partial melting of hydrous, calcalkaline to HKCA (mafic or intermediate) metamorphic crustal protoliths. More recently, petrogenetic models for the origin of HKCA igneous rocks (predominantly I-type granitoids) have been proposed, including the anatexis of mafic lower crust, due to additional heating by underplated basaltic magma in continental arcs or in collision settings (e.g., Jiang et al. 2013;Xiong et al. 2016). In addition, several authors have attempted to define HKCA granitoids as products of fractionation of a hybrid magma derived from variable proportions of crustal and mantelic components in arc-related and post-collisional settings (Liégeois et al. 1998;Wu et al. 2018;Litvinovsky et al. 2021). Therefore, geochemistry and Sr-Nd isotopic compositions are useful proxies to infer the magmatic precursors (crust or mantle) involved in this granitoid petrogenesis.
Recently, Valencia-Moreno et al. (2021) introduced the term Cretaceous-Eocene Mexican Magmatic Arc (CEMMA; Fig. 1) referring to the sustained magmatic arc activity along northern Mexico in response to a flat-slab subduction that occurred in North America during the Late Cretaceous-Eocene period (Clark et al. 1982;Urrutia-Fucugauchi and Bermea 1997;English and Johnston 2004;Liu et al. 2010;Copeland et al. 2017). The spatial and temporal distribution of CEMMA activity probably began in the Late Cretaceous near to the paleotrench (~ 90 Ma), followed by arc widening that reached its maximum extent in the late Eocene (~ 40 Ma) and extended ~ 1000 km eastward. During the initial arc-front migration, a calcic to calc-alkaline volcano-plutonic regime with a high magmatic production rate developed in the western zone. However, outcrops of plutonic masses farther east decrease volumetrically (Fig. 1), and HKCA and alkaline compositions are common in these rocks (Damon 1981;Clark et al. 1982;McDowell et al. 2001;Staude and Barton 2001;Amato et al. 2017;Copeland et al. 2017;Ferrari et al. 2018;Elizondo-Pacheco et al. 2022). Although several magmatic intrusions have been thoroughly studied to propose a petrogenesis model as well as a geodynamic scenario for the CEMMA in northwestern Mexico (González León et al. 2000;Valencia-Moreno et al. 2001Ramos-Velázquez et al. 2008;Roldán-Quintana et al. 2009;Pérez-Segura et al. 2013;Mahar et al. 2016;González-León et al. 2017), few efforts have been made to combine geochemical and isotopic data to help understand

Geological settings and previous studies
The COIC is located in an arcuate, symmetrical anticlinorium system, oriented ~ E−W toward to ~ NW−SE directions, forming a north−south trending sequence of elongated ranges, including: (i) La Caja, (ii) Santa Rosa, and (iii) Santa Rita (Fig. 2). The emplacement of the COIC is considered to be a post-tectonic event following a thin-skinned deformation pulse (Turonian-Campanian) of the Mexican fold-andthrust belt. The magmatic pulses were likely intruded into the anticlinorium structures using a back thrust system as a feeder channel (Fitz-Díaz et al. 2018;Ramírez-Peña et al. 2019). Most of the exposed intrusive rocks are granodioritic in composition. Hornblende-bearing mafic microgranular enclaves appear scattered within the plutonic intrusions without spatial regularity or homogeneous shapes (Fig. 3a,  b). COIC granitoids typically exhibit nonoriented fabrics but massive, fine-grained, and foliated varieties occur along shear zones. Furthermore, porphyritic aggregates are scattered as patches in the plutonic masses.
Country rocks are represented by the Oxfordian-Kimmeridgian Zuloaga carbonates. The areas surrounding large intrusions have been partially metamorphosed into skarn, hornfels, and marble, forming thermal aureoles. In contrast, the contact zones between minor intrusions (i.e., small stocks, dykes, and sills) and sedimentary rocks are relatively sharp and unmetamorphosed. The COIC consists of at least five major magmatic centers: (i) Melchor Ocampo, (ii) Noche Buena, (iii) El Colorado-La Pachona, (iv) Concepción del Oro-Providencia, and (v) Santa Rosa (Fig. 2). The COIC is composed mainly of plutonic rocks, but small amounts of volcanic masses are also exposed along the complex.
Since the early sixteenth century, mining activities have been developed in the Concepción del Oro and Providencia plutons (Bergeat 1910). It is not surprising, therefore, that the first efforts to document the geology in this area were focused on ore deposits (e.g., Rogers et al. 1956;Sawkins 1964;Buseck 1966;Ohmoto et al. 1966;Rye 1966). Subsequently, the sedimentological and structural features of the folded and thrusted mesozoic (Late Jurassic-Cretaceous) sequences and their relationship with the regional deformation event were documented (Antuñano et al. 2000;Ocampo-Díaz et al. 2016;Pinzon-Sotelo et al. 2019;Ramírez-Peña et al. 2019). Castro-Reino (2004) documented three magmatic pulses for Concepción del Oro and neighboring centers, suggesting a common magmatic source and similar petrogenetic models. Based on the mineral and whole-rock dating, a regional activity period from Late Eocene to Early Oligocene has been proposed for the COIC.
Geochronological data for COIC magmatism have been documented in the previous literature. K-Ar dating in biotite, muscovite, and adularia was reported for the Concepcion del Oro and Providencia plutons, covering a time span between ~ 40 and ~ 34.5 Ma (Buseck 1966;Ohmoto et al. 1966). Rb-Sr isochrons for these minerals give ages in a range from ~ 44.0 to ~ 41.0 Ma (Ohmoto et al. 1966). Re-Os dating of molybdenite from the Peñasquito pluton yielded mineralization ages between 35.27 ± 0.18 and 34.97 ± 0.17 Ma (Rocha-Rocha 2016). In addition, K-feldspar and biotite dating from this ore deposit yielded 40 Ar-39 Ar ages varying from ~ 33.95 to ~ 32.82 (Rocha-Rocha 2016). Later, Valencia (2010) conducted a U−Pb in zircon geochronological survey on behalf of Goldcorp Inc., dating small stocks and diatremes exposed in the vacinity of the Peñasquito ore deposit. According to the latter work, COIC crystallization and mineralization events occurred between ~ 46.0 and ~ 33.4 Ma. Ramírez-Peña et al. (2019) reported U−Pb LA−ICPMS ages in zircon from five COIC samples. Two of them (CO-03 and CO-05) were collected in the Concepción del Oro and Providencia plutons, yielding best ages of 42.3 +0.5 / −0.1 and 42.8 +0.8 / −0.6 Ma, respectively. Also, separated zircons from sample CO-08 (Noche Buena) yielded a U−Pb age of 36.80 +0.3 / −0.4 Ma. A dyke subordinated to the Santa Rosa pluton (sample  1 3 was dated by the U − Pb method at 32.2 +0.2 / −0.3 Ma, whereas an andesitic lava flow exposed near the Terminal de Providencia pluton yielded an age of 41.0 +0.2 / −0.7 Ma. Recently, Diaz-Bravo et al. (2022) reported a U−Pb zircon age (42.6 ± 0.2 Ma) for a quartz-monzodiorite from the main intrusive body of the COIC.

Sampling and petrography
Fifteen plutonic rocks were sampled (3-5 kg) during the fieldwork (Table 1; Fig. 2). Petrographic analyzes were conducted using a Leica DM750P microscope (Facultad de Ciencias de la Tierra, UANL) to document texture, mineral assemblage, and modal composition (approximately 1,000 points per sample). Photomicrographs were taken with a digital camera (ICC50 HD) coupled with LAS EZ software (Leica Microsystems). Mineral abbreviations according to Whitney and Evans (2010) have been used in this paper. Fig. 3 a, b Field photographs of granodiorite outcrop sampled in the study area (Concepción del Oro pluton). Note the dimensions of (b) the mafic magmatic enclave (MME) hosted in the plutonic mass. Diameter of coin for scale is 20 mm

Whole-rock geochemical analyzes
Whole-rock geochemical analyzes were performed for the fifteen studied samples. During the sample preparation, any distinguishable weathered portions were removed. The samples were pulverized to 200-mesh using an agate mill. The resulting powder was quartered and used for geochemical and Sr−Nd isotopic analysis. Major element compositions were acquired through the analysis of fused glass discs using a scanning wavelength dispersion X-ray fluorescence (WD-XRF) spectrometer (Siemens SRS 3000) with an Rh-anode X-ray tube as the radiation source. The instrument is installed at the Instituto de Geología, UNAM, Mexico. Furthermore, the sample powder was fused with a 10% dilution in a Li 2 B 4 O 7 -LiBO 2 mixture (1:1) using a Claisse flux automatic melting machine equipped with Pt-5% Au crucibles and molds. Loss on ignition (LOI) was determined gravimetrically by heating 1 g of the sample powder at 1,000 °C for at least 1 h in a muffle furnace. International and in-house standards and replicate analyzes were performed in the same batch to estimate the reliability of the analyzes. The details of the methodology, data acquisition, and reproducibility are summarized in Lozano and Bernal (2005). Relative analytical uncertainties were better than 2%.
Thirty-two trace elements, including rare earth elements (REEs), were determined using a Thermo X Series II inductively coupled plasma-mass spectrometer (ICP-MS). Chemical preparations and analytical procedures were conducted at LEI-UNAM, Mexico. The reagents and samples were handled under clean-room conditions. In addition, the sample dissolution was achieved in steel-jacketed Teflon Parr ® bombs under high pressure at 190 °C for four days. Data acquisition procedures and typical reproducibility are consistent with those described by Mori et al. (2007). Analytical accuracy in trace element analysis was also checked by routine analysis of five international rock standards. The average concentration and standard deviation of several analyzes of these standards in the same laboratory are summarized in Mori et al. (2009). The accuracy of trace element analysis in this study is within 5-10%.

LA-ICP-MS zircon analysis
Four samples (CO-01, CO-07, CO-10, and CO-11) were considered for U−Pb zircon geochronology. Sample preparation and zircon separation were performed at the Departamento de Geología, CICESE, Mexico. Zircon grains were separated from ground samples using conventional methods, such as wet shaking table (Wifley ® ) and magnetic separation by a Frantz ® instrument, followed by hand selection under a binocular microscope. The separated zircons were mounted on epoxy resin and polished down to expose the interior of the grains. CL color imaging was performed at LEI-UNAM using a luminoscope (ELM-3R) that includes a cooled CCD camera, a cold cathode discharge tube, and a vacuum chamber. Laser ablation inductively coupled plasma mass spectrometry (LA-ICPMS) was also conducted at LEI-UNAM, using the methodology described by Solari et al. (2010). Ablation procedure was performed with a Resonetics Resolution M-50 excimer laser system using a spot diameter of 25-35 μm, and the drill depth is ~ 15-20 μm for a total mass ablated of approximately 70-80 ng for each analysis. This material was fed with a He + N 2 gas mixture into the plasma source of a Resonetics M50 workstation coupled to a Thermo X series II quadrupole ICP-MS. The Pleišovice reference zircon (337 Ma;Sláma et al. 2008) was used to correct the instrumental drifts. Given that the low 204 Pb count rates were insignificant compared to the 204 Hg contained in the carrier gasses, common Pb correction was performed employing the algebraic method of Andersen (2002). The age distribution, its uncertainties, and the conventional diagrams were plotted using IsoplotR (v. 3.8;Vermeesch 2018).
In addition to the isotopic information of the U-Th-Pb geochronology ( 206 Pb, 207 Pb, 208 Pb, 232 Th, and 238 U), trace element compositions of all zircons were determined during isotopic analysis using the same Thermo X series II quadrupole. The trace elements analyzed included all REE as well as P, Ti, Nb, Y, and Hf. The standard glass NIST 610 was also used to recalculate element concentrations, using 29 Si as the internal standard. Estimates of Ti-in-zircon temperature followed the calibration method of Watson and Harrison (2005). Quartz was present in all samples studied. Therefore, α SiO2 was considered as 1 in the calculation, whereas we assumed α TiO2 as 0.5 (Watson and Harrison 2005;Ferry and Watson 2007). The TZT computer program (Visual Basic-based software) was used to perform the calculation (Dardier et al. 2021).

Rb-Sr and Sm-Nd isotopic analyzes
Whole-rock isotopic analyzes were performed on 12 samples using the isotope dilution method (ID). The Rb, Sr, Sm, and Nd isotopes were analyzed on the same aliquot to minimize potential uncertainties due to sample powder heterogeneity. Sample preparation and chemical procedures were performed in the ultra-clean laboratory facilities of the Departamento de Geología, CICESE, Mexico. The standard methodology is described in González-Guzmán et al. (2016) and Weber et al. (2018). Approximately 100 mg of powdered whole-rock aliquot was weighed into a digestion vessel and spiked with a mixed 84 Sr-149 Sm-145 Nd tracer. Then, sample digestion was carried out in a closed system (DAS digestion system) with a mixture of concentrated and ultrapure acids (HF, HNO 3 , and HClO 4 (4: 1: ~ 0.5) for 5 days at 185 °C. After spike equilibration, a portion of the resulting solution was divided and spiked with an 87 Rb tracer for Rb isotopic analysis. Elemental separation from the sample solution was performed using the conventional ion exchange chromatography technique. The purification of Rb was done separately. For Sr, Sm, and Nd aliquots, mass analyzes were performed in a Finnigan MAT 262 mass spectrometer (TIMS) equipped with a variable-collector system (with eight Faraday cups) in static mode. Rubidium measurements were made with the NBS-NIST single-collector mass spectrometer. Both spectrometers were part of the facilities of LUGIS-UNAM. The full procedural blanks were 52 pg of Nd and 80 pg of Sr. Data reduction was performed offline, normalizing to an accepted constant isotope ratio using the exponential law ( 86 Sr/ 88 Sr = 0.1194, 152 Sm/ 147 Sm = 1.78308, and 146 Nd/ 144 Nd = 0.7219). In addition to the whole-rock analysis, the NBS 987 standard yielded an average 87 Sr/ 86 Sr value of 0.710241 ± 28 (1σ). Repeated analysis of the La Jolla Nd standard yielded 143 Nd/ 144 Nd of 0.511896 ± 25 (1σ).

Petrography
At the outcrop scale, the studied rocks are light gray due to the conspicuous presence of Fsp and Qz and have mostly fine-to coarse-grained (0.1-8 mm) crystalline varieties  (Table S1) contains a description of the collected rock samples.
The typical mineralogical assemblage of the COIC consists of Pl, Kfs, and Qz as major phases, whereas Hbl, Cpx, Bt, Ap, Zrn, and Opq occur as accessory minerals in variable amounts ( Fig. 4a-f). The modal abundance proportions of Qz, Kfs, and Pl was plotted in a Streckeisen ternary diagram (Fig. 5;La Maitre et al. 2002). Most of the samples spread over the granodiorite field. Nonetheless, three samples can be classified as quartz-monzodiorite. Plagioclase is the most abundant mineral, which generally forms euhedral crystals complexly zoned with well-developed polysynthetic twinning ( Fig. 4a, b). In addition, it also occurs as poikilitic phenocryst that rarely contains inclusions, surrounded by other major and accessory minerals; therefore, Pl is interpreted to have been the first phase to crystallize from the magmatic melt. Orthoclase is the main K-Fsp phase (Fig. 4f), characterized by either Carlsbad or Baveno twinning modes, clearly distinguishing it from the undulatory extinction present in quartz. Microcrystals and the interstitial growth of Qz only occur in sparse areas (Fig. 4c), indicating that this phase appears during the late crystallization stages. Mafic phases include Hbl, Bt, and, less commonly, Cpx ( Fig. 4d-f). Hornblende occurs as interstitial crystals with yellow to brownish pleochroism. Pink to orange Cpx specimens generally appear rimmed by Hbl, indicating a crystal−liquid reaction. Brownish, green Bt flakes often exhibit a pebbly texture and pleochroic halos around Zrn, due to the likely high U contents in these minerals (Fig. 4d). Main accessory phases in addition to Zrn are scattered and may include Ap, Opq minerals, and Ttn. Alteration effects, such as Fsp albitization and Bt chloritization are rare.
The REE and other immobile trace elements were normalized to chondrite and primitive mantle using the values reported in McDonough and Sun (1995) (Fig. 8). The Chondrite-normalized REE patterns display slightly fractionated REE patterns ([La/Yb] N = 6-19) and significant to weakly negative Eu anomalies (Eu/Eu* = 0.63-0.94; Eu/ Eu* = ([Eu] N /([Sm] N + [Gd] N ) 1/2 ), indicating the variable plagioclase fractionation. Primitive mantle-normalized multielement patterns are characterized by a zigzag distribution with an enrichment in large ion lithophile elements (LILE), including Rb, K, and Pb, and a depletion in highfield strength elements (HFSE) such as Th, Nb, Ta, P, and Ti, which is coherent with a subduction-related source and the typical mobility behavior of both LILE and HFSE.

Zircon geochemistry and U−Pb geochronology
The separated zircons are colorless or buff to transparent, euhedral to subhedral, and elongated to stubby, ranging in length from ~ 200 to 400 µm. In the CL images, they exhibit oscillatory zoning typical of magmatic grains (insets in  Fig. 9a). Isotopic ratios of 27 zircons from the granodiorite CO-07 (Providencia pluton) show a similar distribution to the other samples from the study area. These zircons yield a weighted mean 206 Pb/ 238 U age of 42.6 ± 0.1 Ma (MSWD = 6.4; Fig. 9b). In addition, one laser spot from this sample yields an outlier 206 Pb/ 238 U age date at 189.2 ± 4.5 Ma. Twenty-seven out of 28 laser spots from granodiorite CO-10 (Concepción del Oro Pluton) yielded a weighted mean 206 Pb/ 238 U age of 41.8 ± 0.1 Ma (MSWD = 6.0; Fig. 9c). The isotope ratios of 26 zircons from granitoid CO-11 define a weighted mean 206 Pb/ 238 U age of 42.1 ± 0.1 Ma (MSWD = 2.1; Fig. 9d). Three additional laser spots yielded 206 Pb/ 238 U ages of 49.3 (± 0.1), Fig. 6 Silica variation diagrams showing selected major elements: a-MgO; b-TiO 2 ; c-FeO; d-CaO; e-P 2 O 5 , and trace elements: f-Sr; g-V; h-Ta; i-Th; j-Hf). Note a marked decreasing trend in the concentrations of the major elements, an increase in the concentration of some trace elements, and a decrease in others, which suggest a fractionation trend with the magmatic evolution. (l) Rubidium is plotted against Th (ppm). For symbols, see SiO 2 (% wt.)  (Table S5). Median values were plotted on MORB-normalized multielement (Fig. 10a) and chondrite-normalized REE (Fig. 10b) diagrams (normalization data: Grimes et al. (2015) and McDonough and Sun (1995), respectively). Most of the patterns overlap in the multielement diagram, being characterized by strong negative P anomalies, strong La peaks, and moderate to weak positive Eu anomalies. The REE diagram shows homogeneous patterns, typical of igneous zircons, showing a smooth increase from La to Lu, strong positive Ce anomalies (Ce/Ce* ([Ce] N /([La] N + [Pr] N ) 1/2 ) = 14-120), and negative Eu anomalies (0.35-0.60). HREE enrichment yields high [Lu/Gd] N ratios (mean = 40) and an average value of 61 for [Sm/La] N ratios. The tectonic setting discrimination diagrams by Grimes et al. (2015), based on U, Nb, Gd, and Ce normalized to Yb (Fig. 10c, d), reveal an affinity to magmatic arc zircons.

Rb-Sr and Sm-Nd isotopic data
Based on the literature consensus and our new geochronological data on the period of regional magmatic activity and parent/daughter ratios calculated for the same aliquots, initial 87 Sr/ 86 Sr and 143 Nd/ 144 Nd ratios (Table 2) were calculated at 40 Ma. Age-corrected 87 Sr/ 86 Sr (t = 40 Ma) and 143 Nd/ 144 Nd (t = 40 Ma) ratios span in the 0.70458-0.70778 and 0.51248-0.51270 (εNd (t) = from + 2.2 to − 2.1) ranges, respectively, whereas Nd model ages (TDM [Nd] ) were calculated between 1.03 and 0.62 Ga.
The COIC granitoids also exhibit some distinctive geochemical features of subduction-related rocks. The decoupling of LREE−HREE and a marked enrichment of LILE relative to HFSE indicate that the magma source was previously enriched in mobile elements by fluids released from a subducted slab (e.g., Pearce 1983;Hawkesworth et al. 2003;Baier et al. 2008). Moreover, the Fe-index parameter (FeO + 0.9Fe 2 O 3 /FeO + 0.9Fe 2 O 3 + MgO; Frost and Frost 2008) has revealed a probable association to magnesian granitoids (Fig. 11b), a feature largely related to a subduction environment (Frost et al. 2001). Notwithstanding the clear arc-related fingerprint, a syn-to post-orogenic environment is also hypothesized for the larger COIC stocks based on magma emplacement mechanisms (Ramírez-Peña et al. 2019). In this sense, their K-rich character is another remarkable geochemical feature of the COIC. Although potassic magmatism is typically associated with post-collisional environments (e.g., Roberts and Clemens 1993;Barbarin 1999), a genetic association with a mature continental arc environment cannot be ruled out. Furthermore, the COIC rocks in the Rb -(Y + Nb) discrimination diagram (Pearce et al. 1984) are in the Volcanic Arc Granite (VAG) field, close to the triple junction with the Within Plate Granite (WPG) and Syn-Collisional Granite (Syn-COLG) fields (Fig. 11c). Considering also the post-COLG field (Pearce 1996), all points indicate a post-collisional environment. This fingerprint is confirmed using several classical discrimination diagrams, such as the Rb/Zr vs. SiO 2 diagram ( Fig. 11d; Harris et al. 1986).
Some geochemical features of the zircon specimens, such as (i) the Th/U ratios (avg. = 0.43; Supplementary Material, Table S4) and (ii) the chondrite-normalized REE patterns, characterized by an increase in concentration with the atomic number accompanied by a prominent positive Ce anomaly (Fig. 10b), suggest a typical magmatic affinity (Rubatto 2002;Hoskin and Schaltegger 2003). Most MORBnormalized zircon multi-element patterns from COIC rocks (Fig. 10a) resemble those of zircon from post-collisional rocks characterized by U and Ce enrichments and positive Eu anomalies (Grimes et al. 2015). A post-collisional condition in a continental arc tectonic setting for the magmatic zircons was also confirmed applying the U/Yb vs. Nb/Yb and Ce/Yb vs. Gd/Yb (Grimes et al. 2015).

Geothermometry
Petrographic features of zircon and plagioclase, as well as zircon saturation thermometry suggest that COIC rocks could be considered as low-temperature I-type granites (King et al. 1997;Chappell et al. 1998;White and Chappell 2004). For example, petrographic analysis has observed  McDonough and Sun (1995) complexly zoned plagioclase crystals, often with distinctly corroded cores (Fig. 4), whereas CL images have revealed isometric or rounded zircon cores with well-developed magmatic oscillatory zonation (Fig. 9). These attributes are undoubtedly related to a slow homogenization process at low magmatic temperatures.
Zircon saturation thermometry using whole-rock data (Watson and Harrison 1983;Miller et al. 2003;Boehnke et al. 2013) provides a simple and robust estimate for magma temperatures (T Zr ) based on the experimental partition coefficient for zircon (D Zr ) as a function of the parameter M = (Na + K + 2Ca)/(Si × Al) and temperature. Application of this approach to COIC granitoids has shown a T Zr = 680 -814 °C (average ~ 713 °C; Supplementary Material, Table S6), which is considered as the minimum temperature range for magma emplacement. Accordingly, these estimates agree with previously reported T Zr for typical I-type granites (King et al. 1997). In addition, titanium-in-zircon thermometry (Ferry and Watson 2007;Schiller and Finger 2019) has been applied to assess the zircon crystallization temperature (e.g., Buret et al. 2016;Cisneros de León et al. 2021). A similar range of titanium-in-zircon temperature has been observed for the Concepción del Oro-Providencia and Santa Rosa granitoids (632 ± 84 °C [1σ] to 730 ± 140 °C [1σ]; Supplementary material, Table S5). Despite the highest average temperature at 832 ± 39 °C (1σ) established in the zircon specimens separated from sample CO08 (Noche Buena pluton), the relatively low regional temperature evidenced by the overall COIC zircon chemistry suggests an origin from evolved and relatively cold melts (~ 700 °C). This hypothesis has also been supported by the negative Eu anomaly observed in REE patterns (Fig. 10b), confirming the plagioclase fractionation.

Magmatic source and petrogenesis
Two petrogenetic processes have been considered to constrain the origin of HKCA I-type magmas: (i) the partial melting of hydrous K-rich meta-basaltic to intermediate protoliths (lower crust) under relatively high pressure conditions (Roberts and Clemens 1993;Ferré and Leake 2001;Altherr and Siebel 2002;Jiang et al. 2013) and (ii) partial melting of the lower crust with the addition of mantle-derived fluids and/or melts, generating a K-rich magma (Hildreth and Moorbath 1988;Liégeois et al. 1998;Rottura et al. 1998;Castro 2014;Wu et al. 2018;Litvinovsky et al. 2021). In both models, additional assimilation by crustderived materials is not excluded.
Several mineralogical and geochemical clues support the hypothesis of a heterogeneous source composed of crustaland mantle-derived components, explaining the magmatic origins of the COIC. Relatively low Mg# values, low Ni (mean ~ 6.1 ppm) and Cr (mean ~ 10.3 ppm) contents, and a wide range of SiO 2 (53.71-72.53 wt.% water free) indicate that the magmatic source is significantly more differentiated than any magma in equilibrium with the upper mantle. Therefore, the origin of the granitoids cannot be attributed to pure mantle melting. The role of the continental crust in the generation of the high-K calc-alkaline granitic magmas can be more directly assessed from xenoliths trapped in the nearby Quaternary maar fields from San Luis Potosi in central Mexico (Schaaf et al. 1994). This xenolith collection includes mafic to intermediate garnet (± hbl)-bearing  (1995). c Nb/Yb vs. U/Yb tectono-magmatic zircon discrimination diagram. Fields in diagrams, Mantle-and magmatic arc-zircon arrays proposed by Grimes et al. (2015). d Ce/Yb vs. Gd/Yb discrimination diagram for zircon sourced from Arc-MORB-OIB settings and kimberlite (Grimes et al. 2015). Arrows reflect the change in Gd/Yb and Nb/Yb expected for various fractionation processes or sourced melts. Mid-Ocean Ridge Basalt MORB; Ocean Island Basalt OIB; Garnet Grn; Titanite Ttn; Zircon Zrn Rb/ 86 Sr are < 0.5%. Uncertainties reported on Sr and Nd measured isotope ratios are in-run 2σ√n analytical errors in last decimal place, where n is the number of measured isotopic ratios. Initial isotope values were calculated at t = 40 Ma, using λ 147 Sm = 0.654 χ 10-11 (Lugmair and Marti 1977) and λ 87 Rb = 1.393 ˣ 10-11 (Nebel et al. 2011 lithologies in granulite facies (940 ± 60 °C and 7-11 kbar). The petrographic data, initial Sr-Nd isotope ratios, and estimated P-T for the xenoliths suggest that they were assimilated from the lower crust (Schaaf et al. 1994). However, these potential magma precursors have low K 2 O contents (0.07-1.05 wt.%) and, hence, these rocks can be excluded as the only source of the COIC granitoids. Nonetheless, mafic crust was likely present and thoroughly involved during the Eocene anatexis. Together with other crustal components anatexis may have been triggered by emplacement of basaltic magma underplated from an enriched mantle in the arc setting environment related to the CEMMA (Solari et al. 2022). Schaaf et al. (1994) also reported upper mantle xenoliths, including the nodules of hornblende pyroxenite within alkaline host rocks and spinel pyroxenites. These xenoliths were considered to be part of metasomatized mantle fragments or cumulates from a previous magma underplating event at the upper mantle-lower crust boundary. We suggest Fig. 11 a Al 2 O 3 /FeO t + MgO-3CaO-5K 2 O/Na 2 O ternary diagram (Laurent et al. 2014). The fields represent the composition of melts derived from tonalites, metasediments, low-and high-K mafic protoliths. b (FeO + 0.9Fe 2 O 3 / FeO + 0.9Fe 2 O 3 + MgO vs. SiO 2 diagram (Frost and Frost 2008) showing fields for ferroan and magnesian granitoids. c Trace element discrimination diagram (Rb vs Y + Nb) from Pearce et al. (1984). Field of Post-collisional Granite (post-COLG) proposed by (Pearce 1996). d Rb/Zr vs. SiO 2 discrimination diagram (Harris et al. 1986 SiO 2 (% wt.) that the magmatic source of the COIC granitoids is a product of direct melting of mixed crustal−mantle rocks beneath central Mexico. Consequently, the fertility of the mafic lower crust is controlled by LILE-enriched fluids released from the hydrated mineral phases, such as K-rich white mica, carried by magma from the partial melt of an enriched mantle (source region). The Nb/Yb vs. Th/Yb diagram (Pearce 2008;Fig. 11e) shows that the ratios of immobile trace elements for the COIC rocks follow a sub-vertical trend that lies above the MORB-OIB array, indicating a combined source involving enriched mantle and crustal components in a continental arc scenario. Moreover, it is noteworthy that the Nd-isotopic data suggest a mixture of a juvenile crustal component and a significant contribution from older material, as indicated by the larger scatter of TDM [Nd] (0.62-1.03 Ga).

Fe-index
In the conventional Sr-Nd isotope diagram (Fig. 12a), the majority of the COIC granitoids are distributed along the mantle array, showing a well-defined trend. We used a simple two-component mixing model to evaluate the likely source components (mantellic input + lower crust) and the relationship between them involved in their genesis (DePaolo 1988): the mantle component is assumed to be similar to a spinel pyroxenite (JP7, Sr = 48.0 ppm, ( 87 Sr/ 86 Sr) t = 40 Ma = 0.70335, Nd = 6.76 ppm, ( 143 Nd/ 144 Nd) t = 40 Ma = 0.51279) from the Quaternary volcanic field to the south of the study area mentioned above (Schaaf et al. 1994). The second endmember is assumed to be like sample CO03 (Concepción del Oro stock, Sr = 471.09 ppm, ( 87 Sr/ 86 Sr) t = 40 Ma = 0.70535, Nd = 31.41 ppm, ( 143 Nd/ 144 Nd) t = 40 Ma = 0.51254) because it is one of the most evolved granitic rocks. Sr-Nd isotopic modeling suggests that crust-mantle interactions have played a key role in the generation of the granitoids emplaced at the COIC. The Sr-Nd isotopic ratios of the granitoids are like the regional mafic lower crust (Schaaf et al. 1994), sharing (2) Data from Castro-Reino (2004); (3) Data from Valencia-Moreno et al (2021). The dashed line represents a model for simple two end-member mixing. Sources of data: Spinel pyroxenite and mafic lower crust (Mesa Central, Mexico) are reported in Schaaf et al. (1994); Mantle array and DM component are reported in Zindler and Hart (1986   Melchor Ocampo (1) Santa Rosa (1) CEMMA (3) Noche Buena (1) El Colorado-La Pachona (1) (a) a clear evolution trend in Fig. 12a. It is reasonable to deduce that the crustal-derived components are the main source of the protoliths, but with the involvement of mantle-enriched materials.
Most of the Sr and Nd isotopic values produce roughly homogeneous trends with respect to SiO 2 showing slightly positive ( 143 Nd/ 144 Nd) and negative ( 87 Sr/ 86 Sr) correlations, respectively. Such trends can be explained by variable crustal assimilation. The absence of stronger slopes in these diagrams (Fig. 12b, c) indicates that isotopic equilibriumconsidering the magma-underplating model-was achieved during the magmatic processes in the Melting, Assimilation, Storage, and Homogenization (MASH)-zone (e.g., Hildreth and Moorbath 1988;Smithies et al. 2011) prior to fractionation. It is noteworthy that the abrupt shift in the observed trends observed in Figs. 12b and c, caused by the apparently most evolved rock (sample CO-09, Noche Buena pluton and some reported analyzes from Castro-Reino (2004)), suggests that a limited assimilation occurred during the magma emplacement at the shallow crust levels. However, a more likely alternative explanation is that this rock experienced extensive post-magmatic hydrothermal alteration. As result of natural leaching (Andrade et al. 1999;Verma et al. 2018), these samples developed characteristic alteration features: (i) a high LOI percent, (ii) a CaO enrichment, (iii) a Sr and Pb depletion, and (iv) a Sr isotopic composition toward more radiogenic values.
On the other hand, a systematic diminution of Eu/Eu* (0.94-0.63), La/Sm (6.54-3.79), and Rb/Sr (0.32-0.10) ratios relative to Mg# (not shown) and narrow ranges for immobile trace element and isotopic ratios such as ) and 147 Sm/ 144 Nd (0.0941-0.1261) indicate a genetic relationship among the COIC plutonic rocks and identifies fractional crystallization as the predominant petrogenetic evolution process involved for their genesis.

Geodynamic implications
The COIC magmatic activity has been directly associated, both spatially and temporally, with the eastward subduction of the oceanic Farallon plate (paleo-Pacific plate) beneath the southern part of North America (Castro-Reino 2004;Valencia-Moreno et al. 2021;Diaz-Bravo et al. 2022). From ~ 90 to 50 Ma, the magmatic activity triggered by this geodynamic scenario was distributed mainly along a northwest-trending belt, close to the plate margin (McDowell et al. 2001;Staude and Barton 2001;González-León et al. 2017;Diaz-Bravo et al. 2021. This period was associated with a supra-subduction magmatic arc that has produced significant volumes of plutonic products ( (2021)). The magmatic source is the metasomatized (enriched) lithospheric mantle that generates K-rich magmas and juvenile material beneath the COIC (underplating). Magma underplating in this interval contributed to crustal growth, including adding mantle materials to lower crust by intra-crustal differentiation and remelting of the Mexican basement domain (MASH-zone: melting + assimilation + storage + homogenization-zone). Furtthermore, back-arc extension or intra-arc rifting was triggered by the roll-back of the palaeo-Pacific plate. With ongoing extension and upwelling of the asthenosphere, syn-to post-Eocene magmatism occurred. Melts ascended and were geochemically fractionated within the continental crust through trans-crustal faults that acted as feeder channels CEMMA; Valencia-Moreno et al. 2021). For the Late Cretaceous, most authors agree that the subducted Farallon plate underneath the North American plate evolved to a flat geometry. The Mexican Fold and Thrust Belt deformation within the continental interior is associated with this flat subduction geometry (English and Johnston 2004;Fitz-Díaz et al. 2018).
Given that CEMMA magmatic activity in central Sonora and Sinaloa ended at ~ 50 Ma (e.g., Staude and Barton 2001;González-León et al. 2017), the COIC emplacement occurred during the last phase of the regional magmatic pulse, after an apparent space−time hiatus of the magmatic arc. Including the dated samples reported here, the timing of the exposed plutonic rocks belonging to the COIC displays ages that range from ~ 45.3 to ~ 32.  Ferrari et al. 2018). Consequently, we correlate the formation of the COIC with the end of an orogenic cycle. In fact, the dichotomy between the arc and post-collisional geochemical fingerprint of the COIC may symbolize the geodynamic transition from a syn-orogenic to a post-orogenic setting (i.e., late-orogenic environment). In our geological model (Fig. 13), the magmas were initially originated in the source region, close to the mantle wedge, as partial melting of the asthenospheric mantle induced by aqueous fluids/ melts liberated by the subducted slab. Subsequently, these ascending masses caused a large-scale melting of the lower crust. Underplating and intraplating magmas into the lower crust resulted in voluminous magma generation, magma mingling (generating mafic enclaves), differentiation, and the formation of the I-type HKCA magmas generating the plutonic suite of the COIC. Finally, magmas possibly rose and evolved within the continental crust through trans-crustal faults, which acting as feeder channels. Because there is a tectonic control on magma emplacement in the COIC (Ramírez-Peña et al. 2019), the onset of the crustal extension, triggered by the rollback geodynamics of the Farallon (paleo-Pacific) plate, could play a fundamental role in the magma-feeding mechanism of the complex.

Conclusions
It has been demonstrated that the COIC plutonic rocks were emplaced during the late-Eocene-early Oligocene and show strong correlations between major and trace elements, forming a suite, sharing the geochemical affinity of I-type HKCA granitic rocks and suggesting co-genetic relationships between them within the late phases of magmatic activity of the CEMMA. T Zr and Ti-in-zircon temperatures calculated for these rocks indicate that they are derivated from relatively evolved and cold melts. COIC magmatism was derived from two endmembers: (i) a lower mafic crust whose anatexis may have been triggered by emplacement of (ii) mantle-derived magma fluids enriched in incompatible elements during the first stages of the rollback-slab geodynamics in northern Mexico. These magmas ascended to higher crustal levels and underwent fractional crystallization and limited assimilation of host rock materials. The COIC plutonic suite was formed in a late-orogenic setting and was likely linked in response to a transition from an arc-related setting to a post-collisional extensional setting in the late stages of the Mexican fold-and-thrust belt tectonic event.
Funding This research was supported by the CONACyT project "Magmatismo, deformación y metalogenia Laramide: análisis de la subducción y el papel de la litósfera en el norte de México" (Grant No. V49528-F).

Data availability
The authors declare that the data supporting the findings of this study are available within the paper and its Supplementary Information files. In case raw data files are needed in another format, they are available from the corresponding author upon reasonable request.

Declarations
Conflict of interest All the authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.