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 calc-alkaline (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. 2010, 2007; Jiang 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, calc-alkaline 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; Liu et al. 2018; 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. 2001, 2011; Ramos-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 the geodynamic evolution of this magmatic arc in north-central Mexico (Díaz-Bravo et al. 2021; Garcia et al. 2021).

Fig. 1
figure 1

Present-day map of northern Mexico showing the Cretaceous-Eocene Mexican Magmatic Arc (CEMMA), other post-Eocene magmatism, the location of the COIC (Concepción del Oro Igneous Complex), and the main orogenic trend of the structures associated with the Mexican fold-and-thrust belt

Full size image

The Concepción del Oro Igneous Complex (COIC) consists of a series of magmatic centers (NNW−SSE) emplaced in a carbonate−siliciclastic rock sequence (Late Jurassic–Cretaceous), belonging to the foreland orogenic wedge of the Mexican fold-and-thrust belt (Fig. 1), mainly in the form of laccoliths, lopoliths, sills, and dikes. The COIC is an important historic mining region in the northern part of the state of Zacatecas (Mexico). Gold, silver, and copper mineralizations occur in a porphyritic alteration assemblage associated with these magmatic intrusions (Buseck 1966; Castro-Reino 2004; Rocha-Rocha 2016). Regarding the nature of the parental magma in terms of its alkali concentration, the COIC rocks cluster in the HKCA suite. With this in mind, the general geodynamic context of the COIC is ideal for testing various hypotheses about the origin of the HKCA granitoid. Furthermore, because there are few meaningful petrologic studies in the region (Castro-Reino 2004; Velasco-Tapia et al. 2011), this work provides more robust constraints on the tectonomagmatic evolution of the CEMMA, with a particular focus on north−central Mexico. Thus, the main objectives of the current study are to assess the geochemical and isotopic fingerprints for the plutonic rocks of the COIC and to propose a conceptual petrological model for the more internal magmatic pulses of the CEMMA. These approaches were developed based on original petrographic, geochemical, zircon U–Pb geochronology, and Rb–Sr and Sm–Nd isotopic analyzes.

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-and-thrust 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.

Fig. 2
figure 2

Simplified geological map of the COIC (modified after SGM 2000) and sample location of the plutonic rocks under research. The location and ages of previously dated rocks from the COIC by the zircon U/Pb method are also shown. Ages are given in Ma; sources of data: 1—Valencia (2010), unpublished data; 2—(Ramírez-Peña et al. 2019)

Full size image
Fig. 3
figure 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

Full size image

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 40Ar-39Ar 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 CO-13) 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.

Materials and methods

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.

Table 1 Location and analyzes performed in the rock samples of the Concepción del Oro Complex
Full size table

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 Li2B4O7–LiBO2 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%.

Major element data were processed using the R-based (v.4.2.2; R 2022) cross-platform software shinyNORRRM (v.0.8.6; González-Guzmán et al. 2023) to automatically determine: (a) major element compositions on an anhydrous basis after Fe-oxidation adjustment considering Fe2O3/FeO ratios (Middlemost 1989; Verma et al. 2003) and (b) the normative mineralogy. Adjusted major element data were used for all plots and calculations. The Zr saturation temperatures (TZr) were calculated using GCDkit software (v. 6.0; Janoušek et al. 2006).

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 + N2 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 204Pb count rates were insignificant compared to the 204Hg 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 (206Pb, 207Pb, 208Pb, 232Th, and 238U), 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 29Si 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, αSiO2 was considered as 1 in the calculation, whereas we assumed αTiO2 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 84Sr–149Sm–145Nd tracer. Then, sample digestion was carried out in a closed system (DAS digestion system) with a mixture of concentrated and ultrapure acids (HF, HNO3, and HClO4 (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 87Rb 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 (86Sr/88Sr = 0.1194, 152Sm/147Sm = 1.78308, and 146Nd/144Nd = 0.7219). In addition to the whole–rock analysis, the NBS 987 standard yielded an average 87Sr/86Sr value of 0.710241 ± 28 (1σ). Repeated analysis of the La Jolla Nd standard yielded 143Nd/144Nd of 0.511896 ± 25 (1σ).

Results

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 that are inequigranular in texture. Supplementary material (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.

Fig. 4
figure 4

a–f Microphotographs (crossed polarized light) of selected samples from the COIC. a CO-01; b CO-02; c CO-13; d RE-03; e RE-05; f CO-08. The thin sections showing the typically inequigranular texture and zoned euhedral Pl. Mineral phases in matrix such as Qz, Amp, K-Fsp, Bt, Cpx, and Pl are ≤ 0.5 mm in size. Mineral abbreviations according to Whitney and Evans (2010): Amphibole Amp; Biotite Bt; Clinopyroxene Cpx; K-feldspar Afs; Plagioclase Pl; Quartz Qz; Zircon Zrn

Full size image
Fig. 5
figure 5

QAP classification diagram for the plutonic rocks (Le Maitre et al. 2002). Modal classification of the COIC is based on a minimum ~ 1000 points per thin sections. Fields: 1: quartz-rich granitoids; 2: alkali feldspar granite; 3: granite; 4: granodiorite; 5: tonalite; 6*: quartz alkali feldspar syenite; 7*: quartz syenite; 8*: quartz monzonite; 9*: quartz monzodiorite/quartz monzogabbro; 10*: quartz-diorite/-gabbro/-anorthosite; 6: alkali feldspar syenite; 7: syenite; 8: monzonite; 9: mozodiorite/monzogabbro; 10: diorite/gabbro/anorthosite

Full size image

Whole-rock geochemistry

Major and trace element compositions of the plutonic rocks are given in the Supplementary Material (Table S2). Adjusted (normalized to 100% in an anhydrous base) silica contents vary from 53.7 to 72.5 wt. %, indicating an intermediate-felsic composition. Harker diagrams for major (wt.%) and trace elements (ppm) versus silica (wt.%) are reported in Fig. 6a–k. MgO, TiO2, FeO, CaO, P2O5, Sr, V, and Yb show a negative correlation, whereas HFSE, such as Ta, Th, and Hf, show a positive trend. The bivariate Th-Rb diagram shows a positive correlation trend (Fig. 6l). All granitoid samples, except CO09 (Noche Buena, LOI = 5.49 wt.%), can be considered as geochemically unaltered (LOI = 0.55–1.98 wt.%). The Fe-index parameter (FeO/[FeO + MgO]; Frost and Frost 2008) ranges from 1.27 to 1.87, while Mg# ranges from 38.10 to 50.06. On the TAS diagram, all samples are plotted in the subalkaline range (Fig. 7a). In addition, they display metaluminous features, with A/CNK ratios varying between 0.48 and 0.99 (Fig. 7b), and they follow a well-defined calc-alkaline trend in the ternary AFM diagram (Fig. 7c). This behavior is consistent with that observed in the K2O (wt.%) versus SiO2 (wt.%) diagram (Fig. 7d), in which the majority of granitoids are distributed in the HKCA field (Peccerillo and Taylor 1976). Normative mineralogy (Supplementary Material, Table S3) is characterized by Qz saturation (up to 34 wt.%) accompanied by variable amounts of Kfs (~ 8–24wt.%), Ab (~ 26–36 wt.%), and An (~ 8–24 wt.%), consistent with petrographic results.

Fig. 6
figure 6

Silica variation diagrams showing selected major elements: a–MgO; b–TiO2; c–FeO; d–CaO; e–P2O5, 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 Fig. 5

Full size image
Fig. 7
figure 7

Geochemical classification diagrams using the whole-rock major element composition of the plutonic rocks from the COIC. a Total Alkali (Na2O + K2O; wt.%) vs. SiO2 (wt.%) (TAS) diagram. Discrimination line after Irvine and Baragar (1971). b Shand’s diagram: A/NK (molecular Al2O3/[Na2O + K2O]) vs. A/CNK (molecular Al2O3/[CaO–(1.67 × P2O5) + Na2O + K2O]) (Frost et al. 2001). c AFM ternary diagram in terms of A–Na2O + K2O (wt.%), F–total iron as FeO (wt.%), and M–MgO (wt.%). The line indicates the boundary between tholeiitic and calc-alkaline series according to Irvine and Baragar (1971). d K2O vs. SiO2 (wt.%) diagram after Peccerillo and Taylor (1976), showing a HKCA affinity of the granitoids

Full size image

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 high-field 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.

Fig. 8
figure 8

a Chondrite-normalized REE diagram and b Primordial Mantle-normalized incompatible multi-element diagram for rock samples from the study area. Normalizing values are after McDonough and Sun (1995)

Full size image

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. 9). Raw U−Pb data are listed in the Supplementary Material (Table S4). The U/Pb isotopic results are presented in Tera–Wasserburg diagrams (207Pb/206Pb vs. 238U/206Pb), with uncertainties at the one sigma level. In addition, weighted mean ages were calculated from 238U/206Pb apparent ages for the four rocks. These represent crystallization ages, as well as the minimum emplacement ages. Zircon spots from all samples display Th/U ratios of 0.22–1.61.

Fig. 9
figure 9

Tera-Wasserburg (207Pb/206Pb vs. 238U/206Pb) diagrams of zircon U−Pb dating of the selected plutonic rocks of the Concepción del Oro stock. a Sample CO-01; b Sample CO07; c CO-10; d CO-11. The weighted average ages for the four rocks and representative CL images of dated zircons are also shown. Grey ellipses and bars were excluded from the age calculation. Error ellipses and age uncertainties are at the 2σ level

Full size image

For sample CO-01 (Concepción del Oro), 29 zircons were analyzed. Most dates are slightly discordant. However, due to large uncertainties in 207Pb signals for such young zircons measured with quadropole ICP-MS and additional uncertainties on common Pb corrections, the best estimate for the time of zircon crystallization is the weighted mean calculated from the apparent 206Pb/238U age at 42 ± 0.1 Ma (MSWD = 3.7; 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 206Pb/238U age of 42.6 ± 0.1 Ma (MSWD = 6.4; Fig. 9b). In addition, one laser spot from this sample yields an outlier 206Pb/238U 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 206Pb/238U 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 206Pb/238U age of 42.1 ± 0.1 Ma (MSWD = 2.1; Fig. 9d). Three additional laser spots yielded 206Pb/238U ages of 49.3 (± 0.1), 62.9 (± 0.6), and 176.9 (± 1.2) Ma, respectively, indicating the presence of an inherited (crustal) component.

The trace element compositions in the COIC zircons were determined using LA-ICP-MS, being tabulated in the Supplementary Material (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.

Fig. 10
figure 10

a Median zircon compositions normalized to mid-ocean ridge (MORB) zircon (Grimes et al. 2015). Data for continental arc and post-collisional zircon settings from the compilation of Grimes et al. (2015) are plotted for comparison. b Median zircon compositions normalized to chondrite (normalizing values from McDonough and Sun (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

Full size image

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 87Sr/86Sr and 143Nd/144Nd ratios (Table 2) were calculated at 40 Ma. Age-corrected 87Sr/86Sr(t = 40 Ma) and 143Nd/144Nd(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.

Table 2 Results of the analyzes of Rb–Sr and Sm–Nd isotopic systems determined for this research
Full size table

Discussion

Geochemical affinity

According to the magmatic source or protolith, granitic rocks have commonly been categorized into SIAM types (Pitcher 1982; Whalen et al. 1987; Barbarin 1999; Chappell and White 2001; Frost and Frost 2008). Although the distinction is not always straightforward, they can be classified based on numerous petrological parameters (Chappell and Stephens 1988; Chappell and White 1992, 2001; King et al. 1997; Chappell et al. 1998; Frost et al. 2001; White and Chappell 2004; Ghani 2005; Ghani et al. 2013). Several features of the COIC granitoids strongly suggest that they belong to I-type granites: (i) a lithological association with porphyry-type Fe−Cu−Au mineralization, (ii) the presence of hornblende-bearing mafic enclaves (Fig. 3b), (iii) a Hbl + Bt ± Ttn association in the mineral assemblage (Fig. 4), (iv) a SiO2 range from 53.07 to 68.55 wt.%), (v) a broadly negative correlation between SiO2 and P2O5 (Fig. 6e), (vi) a positive correlation between Th and Rb (Fig. 6l), (vii) a metaluminous character (Fig. 7b; A/CNK < 1, A/NK > 1, AI > 0), (viii) major-element geochemistry belonging to the HKCA series (Fig. 7d), and (ix) patterns slightly enriched in LREEs (La–Nd)-in the chondrite-normalized diagram ([La/Yb]N = 6–19; Fig. 8a).

All these features indicate a magmatic source produced by partial melting from meta-igneous protoliths of intermediate to mafic composition. This fact is confirmed by the behavior of major elements in the Al2O3/(FeOt + MgO)–3CaO–5(K2O/Na2O) ternary diagram (Fig. 11a; Laurent et al. 2014). The COIC rocks are mainly located in the range of mafic melts with high K content in this diagram.

Fig. 11
figure 11

a Al2O3/FeOt + MgO—3CaO—5K2O/Na2O 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 + 0.9Fe2O3/FeO + 0.9Fe2O3 + MgO vs. SiO2 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. SiO2 discrimination diagram (Harris et al. 1986). WPG Within Plate Granite; VAG Volcanic Arc Granite; ORG Ocean Ridge Granite. e Th/Yb vs. Ta/Yb diagram proposed by Pearce (1983). S–Subduction component trend; C–Crustal contamination trend; W–Within plate trend

Full size image

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 + 0.9Fe2O3/FeO + 0.9Fe2O3 + 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. SiO2 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 MORB-normalized 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 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 (TZr) based on the experimental partition coefficient for zircon (DZr) as a function of the parameter M = (Na + K + 2Ca)/(Si × Al) and temperature. Application of this approach to COIC granitoids has shown a TZr = 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 TZr 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; Liu et al. 2018; Wu et al. 2018; Litvinovsky et al. 2021). In both models, additional assimilation by crust-derived materials is not excluded.

Several mineralogical and geochemical clues support the hypothesis of a heterogeneous source composed of crustal- and 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 SiO2 (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 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 K2O 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 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).

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, (87Sr/86Sr)t = 40 Ma = 0.70335, Nd = 6.76 ppm, (143Nd/144Nd)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, (87Sr/86Sr)t = 40 Ma = 0.70535, Nd = 31.41 ppm, (143Nd/144Nd)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 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.

Fig. 12
figure 12

a Age-corrected 87Sr ⁄ 86Sr vs. 143Nd/144Nd correlation diagram of the COIC. (1)Data from this study; (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). Variation of age-corrected (b) 143Nd/144Nd and (c) 87Sr/86Sr ratios among silica content of the COIC. The arrows indicate Fractional Crystallization, Crustal Assimilation, and Hydrothermal alteration trends. BSE–Bulk Silicate Earth; CHUR–Chondritic Uniform Reservoir

Full size image

Most of the Sr and Nd isotopic values produce roughly homogeneous trends with respect to SiO2 showing slightly positive (143Nd/144Nd) and negative (87Sr/86Sr) 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 equilibrium—considering 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 Zr/Hf (36.25–48.84) and 147Sm/144Nd (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, 2022). This period was associated with a supra-subduction magmatic arc that has produced significant volumes of plutonic products (early 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 ~ 32.3 Ma (Late Eocene–Early Miocene; Buseck 1966; Ohmoto et al. 1966 Valencia 2010; Rocha-Rocha 2016; Ramírez-Peña et al. 2019; Diaz-Bravo et al. 2021). During this period, the tectonic settings in north-central Mexico transitioned from compressional to the onset of an extensional regime (Aranda-Gómez et al. 2005; 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.

Fig. 13
figure 13

Schematic sections illustrating the proposed COIC geodynamic model (not to scale). The palaeo-Pacific plate was subducted below the North American plate at an extremely low angle. The progressive flattening of the subducted slab caused the eastward migration of the magmatic arc. The COIC represents one of the innermost distinct magmatic pulses in the Cretaceous-Eocene Mexican Magmatic Arc (CEMMA; Valencia-Moreno et al. (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

Full size image

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. TZr 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.