Post-Collisional Magmatic Evolution of the Castelo Intrusive Complex, Espírito Santo, Brazil: New U-Pb Geochronological Data and Integration of Petrographic and Isotopic Evidence

1. Introduction

The period of continental collision is marked by intense metamorphism, melting, and continental growth due to crustal accretion [1,2]. Following this period, there is the post-collisional stage, where mafic magmas of mantle origin access the felsic continental crust and can mix, generating different products due to the degree of hybridization and being observed both at the outcrop scale and in thin sections [3,4,5,6]. Many authors consider the mixing process as the main mechanism capable of generating compositional variations in suites of igneous rocks [7,8,9].

Although many authors oppose the mixing process due to the physical difficulties involving contrasting magmas in terms of temperature, viscosity, and degree of crystallinity, among other parameters, the diffusion capacity of the elements allows the mixing mechanism to occur, leading to degrees of homogenization that can be modeled from a geochemical point of view through numerical simulations and experiments [8,9]. In addition, zircon can be used as a powerful petrogenetic tool to trace magma sources [10,11]. The Araçuaí Orogeny, located in southeastern Brazil and involving the states of Espírito Santo and Minas Gerais, records a long evolutionary history marked by different stages of collision, metamorphism, and magmatism. Intense magmatic activity between 530 and 480 Ma [12,13] marks the post-collisional stage. The rocks generated during this period present a broad compositional spectrum, ranging from gabbro/norite to syenogranite, and several features related to magma mixing [14].

The Castelo Intrusive Complex (CIC) is a representative of post-collisional suite, presenting a wide compositional variation ranging from monzogranite to diorite, exhibiting several features related to magma mixing, and some of its units have recently been studied from a geochronological and isotopic point of view [15]. The objective of this work is to present state-of-the-art research developed in the Castelo Intrusive Complex, present the facies map for the intrusive complex, detail the petrographic characteristics of the rocks that compose it, and present new geochronological data (U-Pb).

2. Geological Setting

The Araçuaí Orogeny, together with the Western Congo Orogeny, constitutes an orogenic system developed in the embayment of the São Francisco and Congo paleocontinents during the Brasiliano-Pan-African event, which extended from the Neoproterozoic to the Cambrian-Ordovician boundary, and which resulted in the closure of the Adamastor Ocean and the consolidation of the West Gondwana supercontinent (Figure 1) [16,17], The Araçuaí-Western Congo Orogeny remained united until the opening of the South Atlantic Ocean in the Cretaceous [13,18].

Located in the northern portion of the Mantiqueira Province, the Araçuaí Orogeny is bordered to the north and west by the São Francisco Craton, to the east by the Atlantic Margin, and the south by the Ribeira Orogeny [19,20]. The boundary of the Araçuaí and Ribeira Orogens has been the target of several studies to connect them since no marked discontinuity is observed between them. Instead, a certain continuity is observed between the rocks that compose them, mainly the magmatic arcs, thus forming the Araçuaí-Ribeira Orogenic System (AROS) [20,22,23,24], proposed an evolution model for the Araçuaí Orogeny known as “nutcracker,” divided into five stages: 1) Formation of the precursor Macaúbas basin, 2) Initial convergence stage, 3) Collisional stage, 4) Lateral escape stage of the southern portion, and 5) Gravitational collapse of the Orogeny. Based on a range of field, structural, petrographic, geochemical, geochronological and isotopic data, [12,13] grouped the magmatism of the Araçuaí Orogeny into five supersuites, reflecting different stages of the evolution, known as: G1 (630–580 Ma), G2 (585–540 Ma), G3 (545–500 Ma), G4 and G5 (530–480 Ma).

Supersuite G1 corresponds to the pre-collisional stage, comprising the plutonic arc section. It consists of an expanded calc-alkaline-magnesian series, formed essentially by rocks of granodioritic to tonalitic composition, in addition to mafic enclaves of dioritic composition [13,25,26,27,28,29,30,31]. The supracrustal sequences related to the arc are formed by metavolcanosedimentary successions, corresponding to the Rio Doce Group and other related basins, such as the Nova Venécia and Paraíba do Sul Complexes [25,29]. Supersuite G2 comprises the syn-collisional stage and is composed of S-type granitoids, sub-alkaline to alkaline and peraluminous, while supersuite G3 represents the late collisional stage [25,29,32,33]. The post-collisional stage is represented by two super suites, G4 and G5, which consist of calc-alkaline to alkaline granitoids, free from regional deformation [13,14].

3. Local Geology: G5 Supersuite

Several intrusive bodies mark the post-collisional magmatism of the Araçuaí Orogeny with elliptical, concentric, and oval shapes related to the gravitational collapse of the orogen. These plutons typically have gabbroic/noritic cores surrounded by syenomonzonites, monzonites, and granites and may have external rings of norite and charnockite [13,14,33,34,35,36,37,38,39]. The most striking characteristics of these plutons are the absence of regional foliation, evidence of magmatic flow and features related to magma mixing, such as net-veined granitic intrusions, micro granular mafic enclaves and schlieren-like features in granites [15,19,41,42,44,45,46,47,48,49,50]. These plutons intrude pre- and syn-collisional granitoids and paragneisses migmatitic (Nova Venécia Complex, Paraíba do Sul, Jequitinhonha). Geothermobarometric data indicate emplacement at shallow crustal levels, showing for the plutons in the southern region of Espírito Santo, such as Pedra Azul, Afonso Cláudio, Santa Angélica, Mimoso do Sul, Venda Nova and Castelo, crystallization at pressures of 5.7 to 11.5 kbar, while the plutons located in the north of the state of Espírito Santo and northeast of Minas Gerais, such as Várzea Alegre, Pedra do Elefante, Barra de São Francisco, Padre Paraíso and Medina, present crystallization at pressures of 2.4 to 3.5 kbar [38,45,46,48,51].

The G5 supersuite magmatism, which corresponds to the last regional event recorded in the Araçuaí Orogeny, is important. This magmatism was triggered by the rise of a hot asthenospheric mantle related to the plate breakup during subduction, followed by a delamination process favoring the crustal anatexis. Isotopic data show a considerable variation for the ratios ⁸⁷Sr/⁸⁶Sr (0.702–0.741), ℇNd_(t) (- 3 to - 24), Nd TDM (1.4–2.6 Ga), ℇHf_(t) (+12 to - 24), and Hf TDM ages (1.25–2.4 Ga). Zircon U-Pb ages for the plutons show an age range of 535 to 480 Ma [19,42,43,46,50].

4. Materials and Methods

The bibliographic review covered the period from the 1980s, when studies on post-collisional magmatism of the Araçuaí Orogeny began to be published, to the present day. All information regarding the relief, field relationships between the outcropping rocks (and geological maps), airborne geophysical data, petrographic characteristics, and lithogeochemical, isotopic, and geochronological data were compiled for this review.

Fieldwork in the Castelo Intrusive Complex was conducted between 2019 and 2022. During the fieldwork, the contact relationships between the different types of outcropping rocks in the intrusive complex and their relationships with the host rocks were identified, the different plutonic petrographic facies were determined, as defined by [60], and the facies map of the intrusive complex was prepared. In addition to the mapping work, twenty-eight samples (from the intrusive complex) were collected to prepare thin sections and three samples for U-Pb geochronology in zircon grains.

The samples for geochronology were prepared at the Geological Sample Preparation Laboratory (LGPA) of the State University of Rio de Janeiro (UERJ). The samples were then washed, manually reduced using a sledgehammer and anvil, crushed (jaw crusher), and powered (disc mill). The pulverized material underwent density separation processes, starting with the hydrodynamic separation table and followed by separation using dense liquids (iodide and bromoform). Finally, the concentrates underwent magnetic separation (Franz) and manual sorting. After epoxy preparation (fixation of zircon grains to an epoxy resin), the samples are sent to the polishing process of the mount for subsequent imaging in a QUANTA 250 Scanning Electron Microscope (SEM) to produce backscattered and cathodoluminescent images. The SEM image of the zircon grains allows observation of the internal structure of the mineral and evidence of inherited core, areas of reabsorption, and magmatic zoning to guide the subsequent location of the isotopic analysis. The complete description of the sample preparation at Multilab (UERJ-Brazil) is reported by [58].

The U-Pb analyses on zircon grains were performed at the Multi-user Environmental Laboratory (MultiLab) – UERJ, using a Laser-Induced Plasma Mass Spectrometer (LA-ICP-MS) through the Element-2 equipment. The order of data reading in the equipment was: (1) blank reading, (2) reading of the GJ-1 standard, (3) reading of the GJ-1 standard, (4) reading of the 91500 standard, (5) reading of eighteen unknown grains, (6) reading of the blank.

The grain surface was ablated using laser pulses with a diameter of 30 μm. The material vaporized by the laser was transported in Ar (0.80 L/min) and He (0.55 L/min) for analysis using 700 cycles of 1 second each. The analyses included mass measurements of ²⁰⁴Pb, ²⁰⁶Pb, ²⁰⁷Pb, ²⁰⁸Pb, ²³²Th and ²³⁸U. Hg represents a common contaminant in He and Ar gases, resulting in an isobaric ²⁰⁴Hg interference with the ²⁰⁴Pb mass. The samples were analyzed together with the zircon reference materials GJ-1 (TIMS normalization data ²⁰⁷Pb/²⁰⁶Pb = 608.3 Ma, ²⁰⁶Pb/²³⁸U = 600.7 Ma and ²⁰⁷Pb/²³⁵U = 602.2 Ma [59] and the results obtained during this investigation was 608.5 ± 0.4Ma. For reference material 91,500 (TIMS-ID age for ²⁰⁶Pb/²³⁸U = 1062.4 ± 0.8 Ma and ²⁰⁷Pb/²⁰⁶Pb = 1065.4 ± 0.6 Ma, according to [59], and the results obtained here indicate a crystallization age of 1065 ± 6 Ma.

5. Results

5.1. Castelo Intrusive Complex

The Castelo Intrusive Complex (Figure 2—Geological Map) generally has an elliptical shape and mountainous relief. However, it has some lowered portions in the southern and central regions of the massif. Outcrops are observed in road cuts, quarries, walls, and waterfalls. Blocks and boulders are common and are associated with the exploitation of dimension stones in quarries. The boundary between the pluton and its host rocks is well-marked by geomorphology.

The mapped units in the Castelo Intrusive Complex are (i) Porphyritic monzogranite, (ii) Fine to coarse inequigranular monzogranite, (iii) Granodiorite, (iv) Quartz-monzodiorite, and (v) Diorite (Figure 3) (Table 1). The CIC has a steep sub-vertical escaped relief, forming elongated mountain ranges with rounded tops, mainly in the granite domain outside this domain, where granodiorite, diorite, and quartz-monzodiorite outcrop, and the relief tends to be lower, where the main drainages are located. The topographic contrast between the host rocks and the CIC is that the host rocks are usually presented as gentler mountain ranges, hills, and mountains.

The host rocks of the CIC are (i) biotite gneiss with amphibolite enclaves, (ii) migmatitic paragneiss, and (iii) marble with amphibolite intrusions. Biotite gneiss is the main unit hosting the CIC. It presents discontinuous metamorphic banding, sometimes continuous, spaced millimetrically to centimetrically, consistent with the regional foliation, marked mainly by K-feldspar porphyroblasts in the felsic band and by biotite and amphibole in the mafic band. The migmatitic paragneiss unit presents a gradational contact with the biotite gneiss unit, and its principal characteristic is the features related to migmatization. Marble with amphibolite intrusions occurs in isolated lenses with exploration queries. The main characteristics of this unit are the skarnitization processes and structures, such as concentric folds and boundins.

The monzogranite outcrops are mainly at the edges of the massif (Figure 4 A), presenting only facies variation without compositional change. The edge of the Castelo Intrusive Complex is marked mainly by the fine- to coarse-grained inequigranular monzogranite (Figure 4 B), passing to the porphyritic monzogranite, with K-feldspar phenocrysts of up to 6 cm (Figure 4 C), as it approaches the center of the intrusion. These two facies present micro granular mafic enclaves (Figure 4 D) of different compositions, sizes, and textures, and features analogous to schlieren of various sizes and sometimes oriented.

Mafic dikes are observed occasionally, and syn plutonic mafic dikes with a high degree of hybridization with the host granite are common. Pegmatitic dikes and intrusions of late leucogranites and even mafic rocks are common, generating agmatic features (Figure 4 E). Xenoliths of the host rock are observed in both facies, presenting various sizes and shapes and degrees of assimilation. In the porphyritic facies, the alignment of K-feldspar phenocrysts is joint, indicating igneous flow. The composition of the monzogranite (Figure 5 A to D) of: Microcline (40-30%), plagioclase (30-20%), quartz (35-25%), biotite (25-15%), orthoclase (15-20%), titanite (5%), opaques (5%), allanite (2-5%), zircon (1%), hornblende (<1%). Muscovite (<1%), sericite (<1%), chlorite (<1%) and carbonates (<1%) are secondary minerals.

Granodiorite is the least common lithotype, occurring punctually in the central-southern portion of the CIC. It is usually gray and often yellowish, with an inequigranular texture and fine to medium grain size (Figure 4 G). It presents small plagioclase xenocrysts, mostly rounded and mafic enclaves. It is common to observe clots of mafic minerals and leucocratic injections of granite in the form of dikes or pockets (Figure 4 F). When the grain size is fine, it usually presents a salt and pepper texture due to the contrast in color between the felsic and mafic minerals. The mineralogy consists of: Quartz (35-25%), plagioclase (35-25%), orthoclase (10-15%), biotite (15-25%), hornblende (15-20%), titanite (5%), opaque minerals (5%), apatite (1-5%), zircon (<1%). Sericite (1%) and chlorite (1%) are secondary minerals. It presents plagioclase crystals evidencing compositional zoning (Figure 5 E and F).

Quartz-monzodiorite is the predominant unit in the mixed zone, often associated with monzogranite, in contact with it, and is predominant in the central-southern portion of the CIC (Figure 4 H). It is dark gray to black in color, meso- to melanocratic, equigranular, and fine-grained (Figure 4 I). Coarse-grained granitic intrusions in the form of veins, pegmatites, and pockets are widespread in interdigitated to abrupt contacts (Figure 4 J). From the interaction between the host quartz-monzodiorite and the granitic intrusions, some features related to magma mixing are found, such as chilled margins, features analogous to pillows, rounded, corroded feldspar xenocrysts, and zones of inclusion of mafic minerals in these xenocrysts, in addition to small mixing zones. The mineralogy consists of plagioclase (40-30%), hornblende (30-20%), biotite (20-10%), quartz (10%), orthoclase (10%), titanite (5%), and opaque minerals and apatite (<5%). From a microscopic point of view, this lithotype presents features indicative of mixing, recorded mainly in the plagioclase, such as compositional zoning (Figure 5 G), plagioclase laths with evidence of corrosion at the edges (Figure 5 H), boxy cellular textures (Figure 5 I) and slots of mafic minerals. Other textures related to magma mixing are apatite crystal mix and poikilitic texture.

The diorite occurs in the core of the CIC, in the topographically lowest region, surrounded by large monzogranite scarps and appearing as large enclaves within them. Several coarse-grained granite dikes and pockets with tabular K-feldspar phenocrysts (Figure 4 K) were cut. The diorite is dark gray in color, melanocratic, equigranular, medium to coarse-grained, and has mafic clots of a few centimeters (~ 3 cm) (Figure 4 L). Its composition consists of hornblende (40-35%), plagioclase (35-30%), biotite (25-20%), quartz (8-5%), orthoclase (5%), apatite (5%), opaques (5%), titanite (5%), and zircon (1%). Biotite, hornblende, and apatite aggregates are common, in addition to quartz xenocrysts, which in some cases are surrounded by hornblende crystals, evidencing the ocellar quartz-hornblende texture (Figure 5 J, K, and L).

5.2. Lithogeochemical Characteristics

The lithogeochemical data presented here were compiled from the works of [9,14,62]. The new lithogeochemical data are presented in supplementary material A. Each of the works presents different lithological classifications for the rocks of the CIC. Therefore, we grouped the rocks regarding chemical classification, as observed in the classification diagram R1-R2, adapted from [63] (Figure 6).

The chemical classification diagram (Figure 6) demonstrates the compositional variation of the rocks of the CIC, where syenodiorite, monzodiorite, monzonite, diorite, tonalite, granodiorite, and granite are present. Another parameter that demonstrates the compositional variability of the rocks of the CIC is the SiO₂ content, which varies between 50.19 and 74.14% (Figure 7). Figure 7 shows the behavior of the main oxides concerning the variation in the SiO₂ content, where a possible “compositional gap” (3.72%) was observed between 58.17% and 61.89% of SiO₂.

Figure 8 presents the principal component analysis performed with the main oxides (Figure 8A) and the trace elements (Figure 8B) available in the compiled works. The possible compositional gap demonstrated in the Harker diagrams delimits two large sets of rocks in the CIC, and this same pattern is observed in the principal component analysis (Figure 8A). The group composed of granites, granodiorites, and two tonalite samples (group 1) is strongly influenced by the SiO₂ and K₂O vectors, where the low angle between the vectors demonstrates their high correlation. The second group, composed of syenodiorite, monzodiorite, monzonite, diorite, and two tonalite samples (group 2), is strongly influenced by the vectors CaO, TiO₂, Fe₂O₃, MnO, P₂O₅, and MgO, where the CaO-TiO₂, Fe₂O₃-MnO and P₂O₅-MgO pairs are highly correlated.

The principal component analysis for the trace elements demonstrated that group 1 can be subdivided into two groups (Figure 8B). The behavior of Rb strongly influences one group, with some influence of Ga, and the other group is influenced by the behavior of Zr. Group 2 presents a more dispersed pattern (Figure 8B). However, a set of samples is strongly influenced by Ba and Nb.

Correlation matrices were calculated for the two large groups observed in the principal component analysis. Molar ratios were calculated for each oxide to standardize the elements, converted from % by weight to parts per million (ppm). In Figure 9, the color bar is associated with different r² values. The r² values are generally low to intermediate for both groups, except for the Fe-Si and Fe-Ti pairs in group 1, where the r² values are more significant than 0.8 (Figure 9).

In the multielement diagram with samples normalized to the chondrite [64] (Figure 10), the set of rocks from group 1 shows a more significant negative anomaly in Eu than the samples from group 2. Both groups were observed to present pronounced ETRL enrichment to ETRP. In the multielement diagram with samples normalized to the primitive mantle [65] (Figure 10), the set of rocks from group 2 presents negative anomalies for Ti, Zr, Sr, and Nb and enrichment for Ba, Th, U, La, and Nd. The set of rocks from group 1 presents a pattern with negative anomalies in Ti, P, Sr, and Nb, differing only for a strongly positive anomaly in Zr, in addition to Th and Nd.

The granitoids of the CIC can be classified as ferrous to weakly metaluminous, alkaline-calcic to calc-alkaline, with a positive trend between the silica content and the Na₂O+K₂O-CaO parameter, and meta- to peraluminous saturated in silica, according to the diagrams of [66] (Figure 11). The classification proposed by [66] determines the rocks of group 2 as having originated in alkaline arc environments. (Figure 12).

5.3. U-Pb Results

Three samples from the CIC were analyzed, two of which were related to granites from the northern and western edges of the massif, and the diorite located in the core. Analytical results are presented in supplementary material B. Sample FGIL 01, corresponding to the granite from the northern edge, presents euhedral zircon crystals, prismatic habits, compositional zonation, and homogeneous cores (Figure 13). The sample provided an upper intercept age of 504 ± 6 Ma and MSWD of 0.62 (Figure 14).

Sample FGIL 37 corresponds to the diorite from the massif's core. The zircon crystals are euhedral to subhedral and show prismatic to subrounded habits, compositional zonation, and homogeneous colors (Figure 15). The sample yielded a 514 ± 6 Ma crystallization age and MSWD of 0.032 (Figure 16).

Sample FGIL 45 corresponds to the granite from the western edge of the CIC. The zircon crystals are euhedral and show prismatic habits, sometimes with bipyramidal terminations, compositional zonation, and homogeneous cores (Figure 17). This sample yielded a crystallization age of 529 ± 3 Ma and an MSWD of 2.7 (Figure 18).

6. Discussions

6.1. Geological Units

The Castelo Intrusive Complex (CIC) has been studied since the 1990s, and since then, several mapping works have been carried out in this intrusive complex, each with its specific focus. As a result, different geological maps have been presented in recent decades [14,37,38,40,42,68]. The map presented in this work results from extensive facies mapping work in the CIC, which has been developed to standardize the different maps presented so far. The diversity of rocks present in the CIC is evident in all works.

Monzogranite is the main unit in the CIC and is related to the highest altitudes and steep reliefs. It also presents two facies variations. The fine- to coarse-grained inequigranular monzogranite outcrops on the edge of the massif, in the western portion and south-central. Porphyritic monzogranite outcrops in the innermost portions and on the eastern edge of the CIC, with K-feldspar phenocrysts that reach 6 cm and indicate magmatic flow. Both have the same mineralogy and are cut by pegmatite veins and dikes, in addition to exhibiting xenoliths of the host rock and features related to magma mixing, such as micro granular mafic enclaves, features analogous to schlieren and agmatic structures (Figure 4 and Figure 5).

Granodiorite and quartz-monzodiorite are predominant units in the south-central portion of the massif (Figure 4 and Figure 5). However, quartz-monzodiorite is present in a restricted form in other portions of the CIC. They are mesocratic to melanocratic rocks, fine-grained, frequently cut by coarse-grained granite injections. The granodiorite presents a salt-and-pepper texture and microgranular mafic enclaves, small plagioclase xenocrysts, mostly rounded, and clots of mafic minerals of a few centimeters. The quartz-monzodiorite is the unit that presents the most features related to magma mixing, such as chilled margins, features analogous to pillows, rounded, corroded feldspar xenocrysts, inclusion zones of mafic minerals in the xenocrysts, mixing zones, compositional zoning, plagioclase laths with evidence of corrosion on the edges, boxy cellular textures, synesis, zones of mafic minerals, crystalline inversion, a mix of apatite crystals and poikilitic texture.

Diorite is the unit that outcrops in the core of the massif in a region of low topography surrounded by granite scarps. It is melanocratic, equigranular, medium to coarse-grained, presents mafic clots of a few centimeters, and is frequently cut by coarse-grained granite injections. Aggregates of biotite and hornblende are common, as are quartz xenocrysts, which show the ocellar quartz-hornblende texture when surrounded by hornblende crystals (Figure 4 and Figure 5).

6.2. The Mixing Process: Field and Petrographic Evidences

Associated with the diversity of rocks, field features show an important process: the physical dispersion of contrasting magmas – magma mixing. This same process is observed in other intrusive complexes of the G5 super suite as well as in other plutons in Brazil and the world, for example [14,70,71,72,73]. The main field evidence observed in the CIC that points to a process of magma mixing is the mafic microgranular enclaves, recurrent in the monzogranites, pillow-like dikes, net-veined, and migration of feldspar megacrysts. According to [74], a single texture cannot be used as evidence of magma mixing but rather a set of them in the CIC.

Microgranular mafic enclaves in the host monzogranite are recurrent in the CIC. Many authors consider that microgranular mafic enclaves would be the result of the mixing of felsic and mafic magmas [75,76,77]. However, they can be interpreted as “bubbles” or “globules” that were produced in the mantle and mixed with felsic rocks of crustal origin [78,79,80]. Other authors, however, consider the enclaves as remnants of the partial melting of the rock that gave rise to the granitoids; therefore, the enclaves would be residues of the most refractory minerals [81,82,83]. In the case of the CIC, some textures in the mafic enclaves point to an interaction between two contrasting magmas. The rounded shapes indicate the disintegration of the mafic magma with the felsic magma [80]. The fine grain of the enclaves probably consists of rapid crystallization due to thermal equilibrium between a high-temperature magma and a low-temperature magma [80,84].

The cooled margins and syn-plutonic dikes suggest temperature differences between the magmas that mix [85]. K-feldspar phenocrysts occur in enclaves as xenocrysts. This happens when phenocrysts from the host rock can overcome the edges of microgranular mafic enclaves and crystallize within them [76,77,86]. The mechanisms for a phenocryst to become a xenocryst involve mechanical migration or capture processes [87], where partial dissolution of K-feldspar crystals is common, making them rounded [88].

Crenulated, lobed, and cusp contacts reinforce the existence of different magmas coexisting, which had low viscosity and/or high resistance to flow when they came into contact [77,80,85]. According to [78], net-veined structures form in the portions of the magma chamber where mafic magma is abundant. Examples of net-veined structures were described by [89] in the British Igneous Province. Physical dispersion between contrasting magmas can form pillow-like features, representing portions of the host mafic rock individualized by the intrusion and with a rounded shape. A type similar to those described is the agmatic analogous features, distinguished by the lack of pattern in their shapes and their chaotic appearance. Both features are recognized in the Santa Angélica Intrusive Complex [37,43].

Petrographic evidence such as zoned crystals, box cellular and synnesius textures in plagioclase, corroded and rounded crystals, ocellar quartz, mafic mineral clots, acicular apatite and poikilitic crystals are indicative of mixing between mafic and felsic magmas. Zonation is indicative of sudden variations in crystallization conditions, such as composition and nucleation, and growth rates of crystals, representing an excellent record of thermodynamic variations during magma mixing [90,91,92]. The boxy cellular texture in plagioclases can be interpreted as the result of a high growth rate and low nucleation rate. The cooling of the environment caused by heat transfer from the heat to the cooler magma during mixing may be ideal for forming the texture [90].

Plagioclase with poikilitic texture is also formed by thermal disequilibrium with the matrix of fine-grained mafic rocks. Plagioclase crystals present inclusions of minerals such as quartz, hornblende, biotite, and opaques [90]. Fine graining and plagioclase laths are common in hybrid rocks, which form due to relatively rapid crystallization, favoring a high nucleation rate [90,93]. The formation of the synneusis texture occurs by the joining of plagioclase crystals suspended in the magma. Quartz mantled by ocellar hornblende consists of a coarse-grained and rounded quartz xenocryst with inclusions of tiny hornblende crystals on its rim. [75] and [90] highlight the mixture of two systems, one more felsic with early quartz crystals and the other more mafic systems containing tiny hornblende crystals. Acicular apatite crystals are formed due to the rapid cooling (quenching) of mafic magma incorporated into a felsic host rock, which already has prismatic apatite crystals [90,94]. Mafic clots comprise hornblende, biotite, opaques, titanite, and apatite crystals.

6.3. The Source Evidences

The lithological variety of the CIC is also expressed in the lithogeochemical data set, especially in the wide range of SiO₂ contents (50.19% to 74.14%) and in the variation observed in the R1-R2 diagram for the classification of plutonic rocks, where the rocks were classified as syenodiorite, monzodiorite, monzonite, diorite, tonalite, granodiorite, and granite. The silica variation observed in the Harker diagrams demonstrated a “compositional gap,” which reinforces the interaction of magmas from different sources that interacted with each other [9,77,90,92].

The principal component analysis (Figure 8) reinforces the existence of two large groups of rocks, group 1 being composed of granites, granodiorites, and tonalites, and group 2 being composed of granodiorites, mozo diorites, monzonites, diorites, and tonalites. The principal component analysis performed with trace elements demonstrated the existence of 2 subgroups among the rocks of group 1, one being strongly controlled by the concentrations of Rb and the other controlled by the concentrations of Zr. In group 2, a diorite, monzonite, and monzodiorite subgroup presents low dispersion, and the other samples present a dispersed pattern in the diagram.

The correlation matrices between the principal components reveal the prominent role of the magma mixing process. The low values observed in general (colors in shades of blue) show a strong correlation between the elements and the effectiveness of the mixing process. The strong correlation between the elements is facilitated in the mixing process by the diffusion process of the chemical elements present in the different magmas. Chemical exchanges are suggested by the advection process, where the contact area between the interacting magmas increases exponentially as a function of time. Consequently, chemical diffusion becomes progressively more efficient [77,95,96,97,98].

The normalized multi-element diagram to chondrite [64] shows similarities between the two groups of rocks of the CIC. Among them, the negative Eu anomaly was more pronounced for group 1 and enrichment in ETRL, with pronounced fractionation. The multielement diagram normalized to the primitive mantle [65] shows some similarities for the two sets of rocks, with very similar values in the content of Ti, Nb, Sr, K, Ba, Th, and U. This demonstrates geochemical affinity between the different types of rock, and the crustal contamination in the most mafic members is attested by the enrichment of elements such as K, Ba, Rb, Sr and the ETRL.

Finally, the classification diagrams demonstrate that the rocks of the CIC are ferrous to weakly metaluminous, alkaline-calcic to calc-alkaline, with a positive trend between the silica content and the Na₂O+K₂O-CaO parameter, and meta- to peraluminous saturated in silica, according to the diagrams of [66], and originated in alkaline arc environments.

In addition to the magma mixing process, which presents much evidence, the assimilation process is also a factor to be considered for the rocks of the CIC. Many xenoliths of the host rock are observed in the massif. The main host rock of the CIC is orthogneiss, followed by biotite-garnet gneiss with migmatization features and marble. The host orthogneiss of the CIC was defined by [24] as the Caxixe Batholith, constituting rocks of granodioritic to granitic composition, with tonalites and gabbros present, calc-alkaline, metaluminous to weakly peraluminous, enrichment in ETRL and patterns similar to the rocks of the CIC in the multielement diagrams normalized to the primitive mantle. In addition, they present positive and negative values of εHf(t) and εNd.

6.4. U-Pb and Lu-Hf Signatures

The U-Pb and Lu-Hf analyses for the rocks of the CIC are presented here, and the data reported by [15] are used here to constrain the emplacement timing of the study rocks. Three samples of monzogranite, one sample of quartz-diorite, and one sample of hololeucocratic dike. The U-Pb zircon ages yielded five new ages for the CIC, showing a peak of magma production around 500 ± 15 Ma, ranging from 527 ± 3 Ma to 427 ± 5 Ma. Three analyses were performed on zircon grains from the monzogranitic component, one on a quartz-diorite sample and one on a hololeucocratic dike sample.

The analyzed samples show positive and negative values of ƐHf_(t), with maximum values between +12.70 and -39.5. The model ages (TDM) presented range from the Paleoproterozoic to the Neoproterozoic, showing a pattern of Paleoproterozoic ages for negative ƐHf_(t) values. In contrast, the Neoproterozoic ages are related to positive ƐHf_(t) values. Another characteristic of the samples analyzed is that they all indicate mantle and crustal sources.

Table 2. Summary of U-Pb geochronological and Lu-Hf isotopic data from the CIC compiled from [68] and [69].

SAMPLE	ROCK	U-Pb AGE	MSWD	176Hf/177Hf	ƐHf (t)	TDM AGES (Ga)
FGIL 06 A	Monzogranite	499 ± 4	1.12	0.28142 to 0.28268	+10.7 and -13.2	0.75 to 2.11
FGIL 11 A	Monzogranite	524 ± 5	1.4	0.28134 to 0.28275	+ 12.6 and – 22.0	0.64 to 2.60
FGIL 17 A	Monzogranite	521 ± 14	0.49	0.28142 to 0.28300	+11.3 and -24.0	0.72 to 2.71
FGIL 06 B	Qtz-Monzodiorite	486 ± 12	0.43	0.28181 to 0.28281	+12.57 and -22.7	0.65 to 2.64
FGIL 06 C	Hololeucocratic dyke	432 ± 32	1.17	0.28155 to 0.28281	+12.7 and -24.0	0.64 to 2.71

Table 2. Summary of U-Pb geochronological and Lu-Hf isotopic data from the CIC compiled from [68] and [69].

SAMPLE	ROCK	U-Pb AGE	MSWD	176Hf/177Hf	ƐHf (t)	TDM AGES (Ga)
FGIL 06 A	Monzogranite	499 ± 4	1.12	0.28142 to 0.28268	+10.7 and -13.2	0.75 to 2.11
FGIL 11 A	Monzogranite	524 ± 5	1.4	0.28134 to 0.28275	+ 12.6 and – 22.0	0.64 to 2.60
FGIL 17 A	Monzogranite	521 ± 14	0.49	0.28142 to 0.28300	+11.3 and -24.0	0.72 to 2.71
FGIL 06 B	Qtz-Monzodiorite	486 ± 12	0.43	0.28181 to 0.28281	+12.57 and -22.7	0.65 to 2.64
FGIL 06 C	Hololeucocratic dyke	432 ± 32	1.17	0.28155 to 0.28281	+12.7 and -24.0	0.64 to 2.71

The new U-Pb data in zircon confirm that the peak of magma production in the CIC corresponds to the post-collisional stage of the Araçuaí Orogeny, between 530 and 480 Ma [36,38,45,47,48,49,50]. Isotopic studies of ⁸⁷Sr/⁸⁶Sr and ¹⁴³Nd/¹⁴⁴Nd presented by [38] and references therein suggested that the gabbroic rocks of the G5 supersuite are the product of partial melting of a previously enriched mantle during the onset of subduction in the Araçuaí Orogeny. A previously enriched mantle may be one of the sources for the generation of basaltic magmas, which is partially corroborated by the low ¹⁷⁶Hf/¹⁷⁷Hf ratios obtained by [15] for the quartz-monzodiorite of the CIC. However, [15] also obtained zircon grains with high ¹⁷⁶Hf/¹⁷⁷Hf ratios and positive εHf_(t) values for this same quartz-monzodiorite from the CIC, which demonstrates that a depleted mantle component is also present. Here, an important discussion arises regarding the mantle source of the G5 supersuite: does only one mantle reservoir act as a source, or, as with magmatism associated with mantle plumes, is more than one mantle reservoir present?

The significant variation between the ƐHf_(t) values, which reach around 30 units, can be explained by the interaction process of magma originating from the crust and the mantle. This variation is observed within each sample analyzed, corroborating that the zircons in these rocks have distinct isotopic compositions and were subsequently agglutinated in the same volume of rock, preserving this considerable isotopic variation [100,101]. Some plutons of the Araçuaí Orogeny also exhibit positive and negative values of ƐHf_(t), as is the case of the Afonso Cláudio, Barra de São Francisco and Santa Angélica Intrusive Complexes [34,45,50] corroborating the fact that different sources of magmas were present in the formation of these massifs, and which are also confirmed by field, petrographic and lithogeochemical evidence, unlike that suggested for the Várzea Alegre and Venda Nova Intrusive Complexes [103], which do not take into account the evidence of magma mixing, including that suggested by other authors for the same massifs [38,45,48,49].

The TDM ages range from 2.7 Ga to 1.41, and together with the ƐHf_(t) values, they corroborate with different magma sources for the CIC, one mantle source, and at least two crustal sources. The negative ƐHf_(t) values observed in all samples show Neoproterozoic TDM ages, ranging from 0.64 Ma to 1.18 Ma, thus constituting an important juvenile reservoir. The ƐHf_(t) values between -26 and -22 show Archean to Paleoproterozoic TDM ages (2.6 to 2.4 Ga), while the samples with ƐHf_(t) values between -17 to -8 are essentially Paleoproterozoic (2.2 Ga to 1.7 Ga).

[46] and [50] discuss the possibility of different crustal sources for the Pedra Azul and Santa Angélica Intrusive Complexes, respectively. As observed in the CIC, an older crustal source appears in both. Therefore, some candidates appear as possible older crustal sources [16] report Archean and Paleoproterozoic TDM ages, in addition to positive and negative values of ƐHf_(t) for rocks of the Caparaó Complex, showing similarities with the signatures observed in the CIC. Other possible candidates that are presented are the Juiz de Fora and Pocrane Complexes, basements of the Araçuaí Orogeny [21,51,104]. The rocks of the Juiz de Fora and Pocrane Complex, which represent the basement of the Araçuaí Orogeny, also present positive and negative values for ƐHf_(t) and ƐNd_(t), in addition to TDM ages in the Archean to Paleoproterozoic transition.

An important Paleoproterozoic candidate for the crustal source of the CIC is the granitoids of the G1 supersuite [39]. Hf and Nd data for rocks of the Estrela Orthogneiss in the Castelo region indicate values ranging from -9.2 to -4.7 and Paleo to Mesoproterozoic TDM ages ranging from 2.2 to 1.4 Ga [105]. The host of the CIC, the Caxixe batholith [24], presents U-Pb ages in zircon around 850 Ma, correlated with the Serra da Prata magmatic arc of the Ribeira Orogeny [32]. These rocks present positive ƐHf_(t) values, between +10 and +14, and a Neoproteorozoic model age. [33] also present for the Caxixe batholith, rocks compatible with the G1 supersuite of the Araçuaí Orogeny, between 630 and 580 Ma, with negative Hf values, between -4 and -13, and TDM ages between 1.33 and 1.67.

7. Conclusions

The Castelo Intrusive Complex (CIC) reveals a complex interaction between magmas of different compositions and sources, evidenced by both petrographic and structural features and lithogeochemical and isotopic data. Mixing magmas is widely supported by microgranular mafic enclaves, specific textures such as ocellar quartz and K-feldspar phenocrysts, and field features such as syn-plutonic dikes and pillow-type volcanic rocks.

In addition to the mixing of magmas, the isotopic data suggests the participation of multiple mantle and crustal sources in the genesis of the CIC, with significant variations in the values of ƐHf_(t) and TDM model ages. These data corroborate the hypothesis of contribution from enriched and depleted mantle reservoirs and multiple crustal sources. These data indicate a complex interaction of post-collisional magmatic processes in the context of the Araçuaí Orogeny, reinforcing the importance of investigating multiple reservoirs to understand the tectonic-magmatic evolution of this Orogeny.

This mechanism may have been facilitated by the delamination of the very thick lithosphere with significant elevations with high gravitational potential and the enrichment of the mantle through the slab at the end of the collisional stage. These elements lead to lateral flow that facilitated the intrusion of basic magmas formed by the partial melting of an asthenospheric mantle and crustal magmas generated by the heat provided by the rising mantle. In this sense, the process of decompression of the lower crust provided sufficient heat for crustal anatexis, where later magmatic mechanisms of physical dispersion (mingling) and chemical diffusion (mixing) favored hybridization. However, it is considered that this was not a process that occurred all at once; the subtle differences in the chemical and isotopic signatures shown in the rocks of the ICC lead to the belief that several pulses and different portions of the crust were responsible for the formation of the magmas that originated the ICC.