Introduction

Skarn deposits are important sources of metals such as W, Au, Fe, Cu, Zn, Mo, and Sn. These rocks are complex and form by magmatic fluids interacting with reactive Ca-(Mg)-rich country rocks, leading to the development of unique calc–silicate mineral assemblages (Zharikov 1970; Einaudi et al. 1981). The process of skarn formation involves the metasomatic addition of various components from an igneous source, such as Cl, F, Fe, W, S, Si, and Al. The formation of the metasomatic mineral assemblage is controlled by the variation of CO2 in the magmatic fluid coupled with other parameters such as temperature, pressure, fluid chemistry, and protolith composition (Korzhinskii 1968; Kwak 1987; Robb 2005; Lefebvre et al. 2019a, b; Romer et al. 2022). Nevertheless, some authors have recently been discussing the formation of skarn during regional metamorphism (e.g. the Hämmerlein skarn deposit; Lefebvre et al. 2019a, b; Romer et al. 2022). In this scenario, skarn formation is linked to the peak regional metamorphism of a pre-enriched sedimentary protolith, and mineralization is a result of metal remobilization due to the subsequent emplacement of granites.

The Borborema Province embraces a major metallogenetic region known as the Seridó Mineral Province (Fig. 1a). The Seridó Mineral Province contains W–(Mo)–(Au) skarn deposits (Souza Neto et al. 2008), mineralized pegmatites with Nb–Ta–(Li), rare earth elements (REEs), and gems such as Paraíba tourmaline (Beurlen et al. 2008, 2014), as well as other mineralization styles including mafic–ultramafic-hosted emerald (Santiago et al. 2019; Araújo Neto et al. 2023), orogenic Au (Araújo et al. 2002), and banded Fe formations (Sial et al. 2015; Cavalcante et al. 2016). The Seridó Mineral Province features more than 700 skarn occurrences in an area of approximately 20,000 km² (Santos et al. 2014). Tungsten exploration in the province has been developed since the 1940s and approximately 60,000 tons of WO3 have been produced (Souza Neto et al. 2008).

Fig. 1
figure 1

(a) Geological setting of the Borborema Province in northeastern Brazil highlighting the Seridó Mineral Province (modified from Hollanda et al. 2017). (b) Simplified geological map of the Seridó belt modified from Hollanda et al. (2017) showing the location of the main skarn deposits. Ages for granitic magmatism from Caramuru, Angicos, Cerro Corá and Picuí are from Hollanda et al. (2017); Acari and Totoró dates are from Archanjo et al. (2013); Macambira date is from this study. Re‒Os dates for the mineralization are from Hollanda et al. (2017)

Full size image

Three skarn deposits stand out in the Seridó Mineral Province: Bodó, Brejuí, and Bonfim. The Bodó mine is located in the homonymous district, a 12-km-long belt where a few minor W-skarn occurrences are also present (Fig. 2). The Bodó deposit is currently the highest-grade tungsten mine in the Seridó Mineral Province, with ore reserves estimated at 9 Mt at a grade of 2% WO3. The Bonfim and Brejuí ore reserves are estimated at 5 and 11 Mt, respectively, sharing average grades of 0.5-1.0% WO3 (Santos et al. 2014). Tungsten is the main product in most skarn bodies, with Mo and, in particular, Au as important by-products, with the exception of the Itajubatiba skarn deposit, in which Au is the only metal exploited (Souza Neto et al. 1998, 2008).

Fig. 2
figure 2

Geological map highlighting the regional-scale folding system that controlled the W–Mo mineralization in the Bodó district. Sampling locations (BDD-17 for U‒Pb dating; BDD-17, BDD-9, and DH10-3 for 40Ar/39Ar dating; BDD-1.6, BDD-15 for thermodynamic modeling), as well as major W-skarn occurrences in the Bodó district are indicated

Full size image

Skarns and pegmatites have historically been genetically linked to the late Ediacaran–Cambrian granitic magmatism that extensively intruded along the Seridó Belt (Silva et al. 1995; Souza Neto et al. 2008; Santos et al. 2014). This connection is due to their close spatial relations, and it is assumed that high-K calc-alkaline granitic magmas were the fluid source for skarn formation. Additionally, the late crystallization of granitic magmas is believed to have played a role in the emplacement of granitic pegmatites. Nevertheless, Re‒Os dating of molybdenite from the three main deposits (Brejuí, Bonfim, and Bodó) in the Seridó Mineral Province (Hollanda et al. 2017) and available U‒Pb zircon (ca. 600 − 526 Ma), U-Pb columbite-tantalite (ID-TIMS dates of ca. 515 − 509 Ma), and 40Ar/39Ar mica (ca. 520 − 500 Ma) data from granites and granitic pegmatites (Araújo et al. 2005; Baumgartner et al. 2006; Archanjo et al. 2013; Hollanda et al. 2015, 2017; Souza et al. 2016; Strmić Palinkaš et al. 2019) have so far failed to establish a link between magmatism and skarn formation, as the ages of the deposits do not match those of nearby igneous bodies.

The objective of this study is to investigate the geological processes and geochemical conditions that lead to the formation of W-skarn mineralization in the Bodó mineral district. Thus, we aim to elucidate the general principles governing the formation of such mineral deposits, providing insights that can be applied to unravel the source of fluids related to skarn formation and the metamorphic and metasomatic evolution of skarn deposits. With this in mind, we investigate the potential connection between skarn mineralization and magmatic-fluid sources via geochronology of a nearby granite pluton and granitic pegmatite. We also present carbon and oxygen isotope data to model the role of magmatic fluids during the metasomatism of marbles and the genesis of the skarns in the Bodó district. Lithogeochemical and mineral chemistry data were used to construct T-XCO2 pseudosections to constrain the prevailing metamorphic/metasomatic conditions during the evolution of the magmatic-hydrothermal system.

Geological setting

The Borborema Province and the Seridó Belt

The Borborema Province in northeastern Brazil (Almeida et al. 1981) is a complex geologic terrain formed during the assembly of the Western Gondwana. The province underwent significant deformation, metamorphism, and magmatism during the Brasiliano orogeny in the Neoproterozoic (Brito Neves et al. 2000; Ganade de Araujo et al. 2014; Caxito et al. 2020, 2021). The amalgamation of Gondwana during the Neoproterozoic resulted in the formation of a continental-scale transcurrent shear zone system (Vauchez et al. 1995; Archanjo et al. 2021; Fossen et al. 2022). The major E-W-trending Patos and Pernambuco shear zones divide the Borborema Province in three tectonic sub-provinces or domains: Northern, Southern, and Transversal (or Central) (Brito Neves et al. 2000; Van Schmus et al. 2011; Oliveira and Medeiros 2018).

The Archean and Paleoproterozoic cratonic nuclei of the Borborema Province are overlain by Neoproterozoic mobile belts. The Seridó Belt is a NE-trending metasedimentary sequence located in the northern portion of the province. It was deposited in the Ediacaran period and later deformed in the Brasiliano orogeny (625–510 Ma; Caxito et al. 2020). The tectonic setting is characterized by dextral transpressive deformation, where NE-trending shear zones are interconnected with the regional Patos shear zone (Corsini et al. 1991; Archanjo et al. 2013).

The Seridó Group (Fig. 1b) is an Ediacaran terrigenous sedimentary sequence unconformably deposited over an Archean–Paleoproterozoic migmatite–gneiss basement (Dantas et al. 2004, 2013; Hollanda et al. 2011; Souza et al. 2016; Ferreira et al. 2020). The Jucurutu Formation of the Seridó Group is characterized by paragneiss with marble lenses of meters to tens of meters in length and minor felsic–intermediate meta-volcanic rocks, mica schist and, subordinately, quartzite lenses (Caby et al. 1995; Van Schmus et al. 2003), and meter-thick layers of banded iron formation (BIF) (Sial et al. 2015). The Jucurutu paragneiss grades laterally and upward to a thick, turbidite-like sequence consisting dominantly of mica schist of the Seridó Formation with minor marble and calc-silicate lenses (Caby et al. 1995; Jardim de Sá et al. 1995; Van Schmus et al. 2003). Neodymium isotope data indicate that the transition between the Jucurutu and the Seridó formations was gradational and conformable, with increasingly radiogenic compositions toward the top. Youngest detrital zircon dates indicate that the deposition of the Seridó Group occurred during the early Ediacaran, at ca. 650–610 Ma (Van Schmus et al. 2003; Hollanda et al. 2015). The intrusion of granite-diorite plutons in the metapelites of the Seridó Formation occurred after ca. 597 Ma (e.g., Totoró pluton; Archanjo et al. 2013; Hollanda et al. 2015, 2017). This indicates that the deposition of the entire succession ceased before 600 Ma, with the distal Seridó Formation overlapping the platformal graywacke–carbonate–BIF association of the Jucurutu Formation during a marine transgression.

At the base of the Seridó Group lies the Equador Formation, which is made up of quartzite and oligomictic to polymictic metaconglomerate. This formation is laterally discontinuous and assigned to an intermediate stratigraphic position between the Jucurutu and Seridó formations (Caby et al. 1995; Jardim de Sá et al. 1995; Van Schmus et al. 2003). A recent study on the detrital zircon provenance of the Seridó Group showed no zircons younger than 1.8 Ga in the quartzite or metaconglomerate of the Equador Formation and strongly negative εNd zircon values (< -22.0). These data differ from the values of the Jucurutu–Seridó succession which provided abundant Neoproterozoic zircon dates and εNd from − 11.0 to -4.5 (Hollanda et al. 2015). The oldest detrital zircon of the Jucurutu Formation coincides with those of the regional basement rocks, whereas in the Seridó-Jucurutu sequence detrital zircons indicate a prominent Neoproterozoic source. Thus, this change in sedimentary provenance may indicate a regional unconformity between the deposition of the Equador sedimentary rocks and the other units (Hollanda et al. 2015).

The metamorphic conditions of the Seridó paleobasin reached 620 °C and 2.5-4.0 kbar (Lima 1987, 1992; Souza 1996; Souza et al. 2007; Cioffi et al. 2021), with locally higher temperatures near the granitic intrusions. The emplacement of granitoids and diorites in the Seridó Belt was related to voluminous felsic to mafic magmatism emplaced during the pre- to post-collisional stages of the Brasiliano orogeny (Archanjo et al. 2013). According to their geochemical affinities, these rocks can be divided into (i) shoshonitic (K-rich) gabbro and diorite, (ii) high-K oxidized calc-alkaline granite (s.l.) with grain sizes and textures varying from porphyritic to medium-grained equigranular, and (iii) A-type quartz syenite and syenogranite (Jardim de Sá 1994; Nascimento et al. 2015). Minor occurrences of calc-alkaline granite and charnockite complete the geochemical diversity of the Brasiliano magmatism in the northeastern Borborema Province. U-Pb ages range from 597 to 525 Ma (Archanjo et al. 2013; Souza et al. 2016; Hollanda et al. 2017). Later granitic pegmatite dikes and sills intruded supracrustal rocks of the Seridó Group (Silva et al. 1995; Beurlen et al. 2008, 2014, 2019). ID-TIMS U–Pb columbite-tantalite dates constrain their emplacement to the Cambrian period, mostly at ca. 515–509 Ma (Baumgartner et al. 2006), whereas U–Pb dates on monazite and 40Ar/39Ar dates on muscovite extend this interval to ca. 525–502 Ma (Araújo et al. 2005; Baumgartner et al. 2006; Beurlen et al. 2008; Strmić Palinkaš et al. 2019).

Skarns and W–Mo–Au mineralization in the Seridó Mineral Province

The major skarns from the Seridó Mineral Province (Brejuí, Bonfim, and Bodó) are centimeter- to meter-thick stratabound deposits developed within the marble or along the lithologic contact between the marble and paragneiss of the Jucurutu Formation (Souza Neto et al. 2008). Although these deposits do not strictly share the same mineral assemblage, they encompass prograde and retrograde metasomatic assemblages, and a later phase of Fe–Cu and Bi–Te ± Au mineralization. The prograde mineral assemblage is defined by diopside–grossular–plagioclase ± amphibole (the latter can also be a retrograde phase), whereas the retrograde mineral assemblage includes epidote, clinozoisite, vesuvianite, and zeolite (Souza Neto et al. 2008; Santos et al. 2014). The main ore minerals are scheelite and molybdenite, with minor pyrite and chalcopyrite in Brejuí and Bodó. In the Bonfim deposit, the W–Mo–Au mineralization is associated with bismuthinite, bismite, native bismuth, joseite, chalcopyrite, and sphalerite. In contrast, in the Itajubatiba deposit native gold is often associated with pyrrhotite, chalcopyrite, pyrite, and magnetite (Souza Neto et al. 2008, 2012). Re‒Os molybdenite dates provided Ediacaran‒Cambrian ages for the mineralization in Brejuí (ca. 554 Ma), Bonfim (ca. 524 Ma), and Bodó (ca. 510 Hollanda et al. 2017).

The Bodó mine is centered on one of several marble lenses deformed along a sequence of NE-trending anticline–syncline fold systems developed due to syn-Brasiliano transpressional kinematics (Fig. 2). Locally, skarns and hydrothermally modified marble lie stratigraphically in contact with the Jucurutu paragneiss. A few igneous bodies are exposed around the mine area. A fine-grained biotite leucogranite shows incipient solid-state foliation and abundant pegmatite and aplite dikes of varying width are mapped on the surface and underground. The Macambira granite is a subcircular body exposed a few kilometers to the northeast. This granite is a K-rich calc-alkaline pluton in which biotite syenogranite and monzogranite are the prevailing facies (Silva and Souza 2012).

Sampling and analytical methods

The samples were collected from a variety of locations within the Bodó district to ensure that the data reflected the overall characteristics of skarn and marble as accurately as possible. In the Bodó mine, samples were collected from outcrops, drill cores, and accessible mine galleries. At the other skarn occurrences, samples were collected only from outcrops. A summary of the analyzed samples can be found in ESM Table S1.

U-Pb zircon and 40Ar/39Ar mica geochronology

The geochronologic investigation targeted (i) one representative sample of the Macambira pluton for U‒Pb (zircon) and 40Ar/39Ar (biotite) dating; (ii) a granitic pegmatite dike in the Bodó mine area collected for 40Ar/39Ar (muscovite) dating; and (iii) a mylonite collected for 40Ar/39Ar (muscovite) dating. A graphical representation of the drill cores investigated is provided in ESM Fig. S1.

Zircons were extracted from the bulk rock samples by conventional hydraulic, magnetic, and density separation (heavy liquids), then handpicked, mounted, and polished, and finally imaged in transmitted light and on a scanning electron microscope in cathodoluminescence (CL) mode. U‒Pb isotope analyses were performed on the SHRIMP IIe hosted at the Institute of Geosciences, University of São Paulo (Brazil). The reader is referred to Sato et al. (2014) for details of the instrumentation and data acquisition and reduction. The isotopic ratios of U and Pb of the unknown samples and the TEMORA standard (Black et al. 2003, 2004) were measured in sets of six mass scans, with TEMORA being analyzed after every third unknown analysis. The standard provided a concordia age of 416.8 ± 1.4 Ma (2σ, n = 18, MSWD = 0.035). U, Th, and Pb abundances and Pb isotope ratios were normalized to the SL13 zircon standard (U = 238 ppm). Common Pb correction was performed using the measured 204Pb and assuming the Pb isotope composition after Stacey and Kramers (1975) model. Raw data were reduced using SQUID 2.5, and U-Pb dates were calculated using IsoplotR v5.2 software (Vermeesch 2018). Uncertainties on the calculated concordia dates are reported at 1σ confidence level. The complete U‒Pb dataset is provided in ESM Table S2.

Mica concentrates from samples and Fish Canyon sanidine standard were irradiated for 14 h in the CLICIT facility of the TRIGA reactor at Oregon State University (USA). After radioactive cooling, mica grains and standards were loaded onto a Cu disk, sealed in a 10− 10 vacuum chamber, and heated under increasing temperature steps using a Nd-YVO4 laser (532 nm) connected to an ARGUS VI (Thermo Scientific) mass spectrometer at the Isotope Geology Center (https://uspmulti.prp.usp.br/centrais/dashboard/122), Institute of Geosciences, University of São Paulo. Corrections for the instrumental blank were made simultaneously with the data acquisition. Data visualization, corrections for nuclear interference, mass discrimination, and irradiation parameter J, as well as date calculation were performed with ArArCALC v2.6 (Koppers 2002) and IsoplotR (Vermeesch 2018). The 40Ar/39Ar data are provided in ESM Table S3.

Bulk-rock, mineral, and isotope geochemistry

The bulk geochemical compositions of two impure marble samples were obtained for use in pseudosection analyses. They were collected at sites BBD-1.6 and BDD-15 in the Bodó mine area and in the Cinzas occurrence, respectively (Fig. 2). These samples were also analyzed for mineral chemistry, together with three skarn samples collected at site BDD-22 and along drill core DDH-10. A powder aliquot of 30 g obtained from a 5 × 5-cm fragment of rock was fused with lithium tetraborate and analyzed for major element compositions by X-ray fluorescence spectrometry using the facilities of the Technological Characterization Laboratory (https://uspmulti.prp.usp.br/public/centrais/4) at the Department of Mining and Petroleum Engineering, University of São Paulo. The loss on ignition (LOI) was estimated from the weight difference after ignition of 2 g of rock at 1020 °C. The complete bulk-rock analyses are reported in ESM Table S4.

Back-scattered electron (BSE) images were taken using a Variable Pressure Environmental Scanning Electron Microscope (VPESEM) Quanta 250 model coupled to an Oxford Energy Dispersive X-ray Spectrometer (EDS) at the Institute of Geosciences, University of São Paulo. The thin sections were uncoated, and the imaging conditions were an acceleration voltage of 20 kV, a working distance of 10 mm, a beam size of 5 μm, and a low pressure of 60 Pa.

Mineral chemistry of amphibole, calcite, clinopyroxene, epidote, feldspar, garnet, and scapolite were performed on 30 μm-thick carbon-coated thin sections by wavelength dispersive spectrometry (WDS) on the JEOL JXA-FE-8530 electron microprobe of the Geoanalítica Center (https://uspmulti.prp.usp.br/public/centrais/74) of the Institute of Geosciences, University of São Paulo. The analyses were performed at 15 kV acceleration voltage and a 20 nA beam current, with a beam size of 5 μm. Analytical parameters and calibration standards used for WDS analyses are listed in ESM Table S5.

Finally, carbon and oxygen isotope analyses were obtained from ca. 200 mg bulk aliquots of 44 agate-milled calcite samples from marbles and skarns. The samples were reacted with 100% H3PO4 at 72 °C under a He-rich atmosphere in a GasBench II device coupled to a DeltaPlus Advanced (Thermo Scientific) isotope ratio mass spectrometer at the Isotope Geology Center (https://uspmulti.prp.usp.br/centrais/dashboard/124), Institute of Geosciences, University of São Paulo. δ18O and δ13C data were measured in reference to the V-PDB (Vienna – Pee Dee Belemnite) standard and stated errors were 0.05‰ and 0.07‰, respectively. δ18O V-PDB was converted to δ18O V-SMOW (Vienna Standard Mean Ocean Water) using the equation δ18O V-SMOW = 1.03091 * δ18O V-PDB + 30.91 (Coplen et al. 1983). The complete C-O isotope data are reported in ESM Table S6.

Thermodynamic modeling

T-XCO2 pseudosections were constructed using Perple_X 6.9.1 (Connolly 1990, 2009) with the internally consistent thermodynamic dataset tc-ds55, and Holland and Powell’s (1998; revised 2004) equation of state for H2O–CO2. Sample BDD-15 was modeled in the NKCFMAS-HC system, whereas sample BDD-1.6 was modeled in the CFMAS-HC system. For the sample BDD-15, Fe3+ was ignored because the its content allocated due to the charge balance in the structural formula in the observed mineral assemblage was insignificant. However, the O2 content in sample BDD-1.6 was estimated to account for the presence of Fe3+ in the analyzed garnets. The bulk composition was adjusted through charge balance to generate garnet with the calculated andradite content.

Both T-XCO2 pseudosections were calculated for fluid-saturated conditions using the following solid solutions: amphibole (Diener et al. 2007), biotite (Tajčmanová et al. 2009), clinopyroxene (Holland and Powell 1998), dolomite (Holland and Powell 1998), calcite (Anovitz and Essene 1987), garnet (White et al. 2000), plagioclase (Newton et al. 1980), and scapolite (Kuhn 2005).

Results

Petrography and mineral chemistry

Bodó district marbles

The Bodó district comprises several discontinuous marble lenses and skarn occurrences and deposits following the sequence of the fold system. The marble occurs in different colors, such as gray, white, green, and orange, and is characterized by a massive (Fig. 3a-c) or incipient to well-developed foliation (Fig. 3d-e). The marble lenses exhibit heterogeneous mineralogy with some parts mostly composed of carbonates and others with substantial enrichment in silicate minerals representing up to 15% vol. (Fig. 3a, e).

Fig. 3
figure 3

Marbles (a-e) and skarns (f-i) from the Bodó district. (a) Greenish and orange marble at the surface of the Bodó mine. Wollastonite and discontinuous/irregularly shaped garnet + clinopyroxene bands. (b) Color varieties of massive marbles. Samples BDM11-1, BDM11-4, and BDM4-6, respectively. (c) Gray marble from the same outcrop as the marble from Fig. 4a with wollastonite and garnet-rich bands. (d) Marble with pervasive foliation from the Cinzas lens with phlogopite and minor diopside. (e) Foliated marble with phlogopite, scapolite, tremolite and diopside from the same lens as in Fig. 4d. (f) Massive skarn consisting mainly of garnet, quartz, and late epidote. (g) Banded skarn composed of garnet, vesuvianite and quartz. (h) Porous skarn consisting mainly of altered garnet and epidote, showing large grayish scheelite crystals. Sample BD-4. (i) Retrograde skarn composed mainly of epidote and vesuvianite with late orange calcite vugs developed over prograde skarns containing garnet. Sample BDD-N409-1. Abbreviations: Cpx = clinopyroxene, Ep = epidote, Grt = garnet, Phl = phlogopite, Qz = quartz, Scp = scapolite, Sch = scheelite, Ves = vesuvianite, Wol = wollastonite

Full size image

The main mineral assemblage in marbles is diverse and includes amphibole, clinopyroxene, garnet, phlogopite, and scapolite, with smaller amounts of epidote, magnetite, plagioclase, titanite, and wollastonite. Among them, epidote, garnet, scapolite, titanite, and wollastonite are characteristic of the metasomatic stage. The composition of amphibole, calcite, clinopyroxene, epidote, garnet, phlogopite, plagioclase, and scapolite is shown in ESM Table S7.

Calcite is almost pure in composition, nearly equigranular, and displays a coarse-grained mosaic texture. Mica is often aligned, although it does not always show prominent foliation (Fig. 4a-b). It has a uniform phlogopitic composition (XMg between 0.866 and 0.893). Clinopyroxene occurs as individual crystals or as aggregates and may show inclusions of calcite, quartz, and phlogopite (Fig. 4b-c). It is classified as diopside (Di79 − 92Hd7−21Jo0−1, with XMg ranging from 0.791 to 0.929; ESM Fig. S2a). Amphibole often presents a decussate texture with inclusions of calcite and rarely phlogopite and diopside (Fig. 4d). It can also occur included in diopside. The amphibole is classified as tremolite and shows high XMg values ranging from 0.900 to 0.969 (ESM Fig. S2b). Epidote forms medium- to coarse-grained xenoblastic crystals around diopside and garnet (Fig. 4e).

Fig. 4
figure 4

Photomicrographs of marbles from the Bodó district showing the occurrence of the main silicate phases. (a) Crystals of phlogopite in equilibrium with calcite. (b) Diopside in contact with and including phlogopite crystals. (c) Small-grained crystals of diopside. (d) Tremolite between calcite crystals. (e) Garnet is replaced by interstitial epidote. (f) Garnet encompassing diopside crystals. (g) Porphyroblast of scapolite consuming plagioclase, tremolite and calcite. (h) Wollastonite invading calcite and quartz. All photomicrographs were taken with transmitted cross-polarized light (XPL). Abbreviations: Cal = calcite, Di = diopside, Ep = epidote, Grs = grossular, Mt = magnetite, Phl = phlogopite, Pl = plagioclase, Qz = quartz, Scp = scapolite, Tr = tremolite, Wo = wollastonite

Full size image

Garnet commonly exhibits xenoblastic to hypidioblastic textures with inclusions of calcite and quartz and is often in contact with or partially includes diopside (Fig. 4f). It has a high grossular component (Gr70 − 84) and is also enriched in andradite and almandine components (Adr8 − 21Sps+Alm6 − 13; ESM Fig. S2c) with no chemical differences between crystal cores and rims. Scapolite occurs as large crystals with a poikiloblastic texture and overgrew after plagioclase (An29 − 30), calcite, and tremolite, often including diopside, tremolite, and quartz (Fig. 4g). It is classified as meionite (Mei57 − 69, EqAn between 0.537 and 0.636). Wollastonite occurs as aggregated acicular crystals (Fig. 4h).

Bodó district skarns

The prograde skarn assemblage is composed of diopside, grossular, quartz, plagioclase, titanite, and apatite, while the most common retrograde minerals include epidote, vesuvianite, calcite, quartz, wollastonite, and prehnite (Fig. 5). The composition of skarn minerals is given in ESM Table S8.

Fig. 5
figure 5

Photomicrographs of the main skarn mineral assemblages. (a) Diopside in contact with plagioclase replaced by white mica. (b) Hypidioblastic to idioblastic grossular crystals of the first generation (grossular I) and interstitial quartz with pervasive late calcite. (c) Diopside inclusions in a large garnet crystal. All pyroxene borders in contact with garnet are irregular, indicating pyroxene corrosion prior to garnet deposition. (d) Large interstitial epidote between coarse-grained garnets, both containing clinopyroxene inclusions. (e) Hypidioblastic porphyroblasts of vesuvianite and quartz occur interstitially between large garnet crystals, with a significant amount of fine clinopyroxene inclusions within the former and latter minerals. (f) Vesuvianite–quartz assemblage showing garnet inclusions, all exhibiting prehnite and calcite filling microfractures and vugs. (g) Veins of epidote crosscutting vesuvianite. (h) Grossular II with quartz inclusions and retrograde mineral assemblage composed of prehnite, wollastonite and calcite. All photomicrographs were taken with transmitted, cross-polarized light (XPL). Abbreviations: Cal = calcite, Di = diopside, Ep = epidote, Grs = grossular, Pl = plagioclase, Prh = prehnite, Qz = quartz, Ves = vesuvianite

Full size image

The skarns are dominated by second-generation, coarse garnet crystals comprising more than 90% of the rock volume in some portions. Garnet often occurs in straight grain boundary with quartz (Fig. 3f-h), making it challenging to establish an accurate relationship with the preceding mineral assemblages. Garnet-bearing skarns locally exhibit a porous texture (Fig. 3h). This texture is correlated with higher ore grades. The textural relationship between mineral phases indicates a complex evolution for the studied rocks, as the former assemblage was overprinted by the second generation of grossular (II). First-generation grossular (I) is distinguished from grossular II by its smaller grain size and hypidioblastic to idioblastic texture. Grossular I occurs with diopside, plagioclase, and quartz, representing the first minerals formed during the prograde stage (Fig. 5a-b). As in marbles, grossular is the most abundant garnet endmember, while other components occur in minor proportions (Gr70 − 84Adr5−21Alm+Sps6 − 13; ESM Fig. S2c). The composition of clinopyroxene corresponds to diopside (Di70 − 92Hd7−28Jo1−4; ESM Fig. S2a), and the XMg values range between 0.716 and 0.929. Centimeter-scale garnet crystals (grossular II) and quartz overprint the previous mineral assemblage (Fig. 5c-e). The mineral assemblage that formed prior to grossular II exhibits corroded borders and textural evidence of disequilibrium caused by interaction with the garnet-forming fluid (Fig. 5c-d).

The retrograde stage comprises recrystallization and deposition of calcite, vesuvianite, epidote/clinozoisite, quartz, wollastonite, and prehnite (Figs. 3i and 5e-h). Vesuvianite occurs as euhedral to subhedral porphyroblastic crystals, often in straight contact or partially including grossular II (Fig. 5e). Vesuvianite commonly contains multiple tiny inclusions of quartz and diopside (Fig. 5e). Epidote/clinozoisite partially replaced diopside, garnet, and vesuvianite (Fig. 5d-f) and filled fractures within these minerals (Fig. 5g). Epidote grains also filled interstitial spaces between garnet crystals and often show diopside inclusions (Fig. 5d). Epidote has XFe between 0.102 and 0.246. Wollastonite occurs as acicular crystals near the marble contact and often shows quartz inclusions (Fig. 5h). Prehnite filled fractures in garnet and occurs as porphyroblasts encompassing crystals of plagioclase, epidote, garnet, and calcite (Fig. 5h). Quartz is present in all stages of the Bodó skarn formation. Pervasive opal with garnet and scheelite inclusions deposited in the final stage. Late zeolite is commonly associated with opal.

The ore stage is mainly represented by scheelite (I and II) and molybdenite with minor pyrite and chalcopyrite. Scheelite I (Fig. 6a-b) occurs as fine-grained crystals and may be completely or partially enclosed by grossular II, which shows corroded borders. Scheelite I represents the main ore phase. Scheelite II (Fig. 6c-d) features crystals that reach 8 cm in size and is often associated with porous skarn. Scheelite II may be surrounded by opal and exhibits corroded borders. Molybdenite occurs in association with quartz, either in veins or disseminated in the skarn matrix, alongside disseminated pyrite and chalcopyrite.

Fig. 6
figure 6

(a) SEM backscattered electron (BSE) image of first generation scheelite (I) included in grossular of the second generation (II). (b) SEM BSE image of a scheelite I crystal in irregular contact with grossular II. In (c) and (d), coarse scheelite II crystals encompassed by a mass of opal. (c) and (d) Photomicrographs taken under transmitted, cross-polarized light (XPL). Abbreviations: Grs = grossular, Opl = opal, Sch = scheelite

Full size image

To summarize, the evolution from marbles to skarns in the Bodó district involved the formation of a primary metamorphic silicate assemblage containing phlogopite, tremolite, and diopside (Fig. 7). Subsequently, metasomatism led to the deposition of wollastonite, grossular (I and II), and scapolite in the marble. The metasomatism involved skarn formation and was characterized by the crystallization of diopside, grossular I, plagioclase, scheelite I, and molybdenite in the prograde stage. The prograde-stage minerals were subsequently overprinted by grossular II. The formation of vesuvianite, epidote, calcite, wollastonite, and prehnite occurred in the retrograde stage, along with the deposition of scheelite II and sulfide minerals.

Fig. 7
figure 7

Paragenetic sequence of marble and skarn in the Bodó district

Full size image

Geochronology

The Macambira pluton (sample BDD-17) is a medium- to coarse-grained biotite syenogranite. Zircons exhibit euhedral bipyramidal shapes and are 150–300 μm long, with aspect ratios ranging from 2 to 4. They commonly show oscillatory or parallel zoning (Fig. 8a). Some grains exhibit weak compositional variations and enrichment in U in rims compared to cores. Apatite is commonly present as inclusions in these zircons. After excluding the analyses that were discordant or characterized by large errors in their individual analyses, eleven analyses yield a Concordia and weighted mean date of 534.9 ± 3.6 Ma (2σ; Fig. 8b-c).

Fig. 8
figure 8

Cathodoluminescence images of analyzed zircon grains (a) and U-Pb zircon dates (b-c) of the Macambira granite (sample BDD-17). (a) U‒Pb concordia diagram and (b) 206Pb/238U weighted-mean diagram. In (a) analyzed spots are indicated with 206Pb/238U dates (in Ma) at 1σ errors. In (b-c) grey ellipses and unfilled rectangles are the excluded samples from date calculation

Full size image

The 40Ar/39Ar dates were obtained on biotite from the Macambira pluton and on muscovite from a pegmatite dike (sample BDD-9) and a mylonite (sample DH10-3). Data are shown in Fig. 9 and summarized in Table 1. The two biotite aliquots provided similar spectra with pronounced 40Ar* loss at lower and intermediate temperatures, representing 30–45% of the total 39Ar released. The sample BDD-17 yielded two distinct ‘mini-plateaus’ (Fig. 9a) whose weighted mean dates are 491.08 ± 2.35 Ma (0.48% error; 42% 39Ar) and 488.74 ± 2.65 Ma (0.54% error; 25% 39Ar), which overlap at the 2σ level. Together, these data provided a weighted mean date of 490.65 ± 0.67 Ma (Fig. 9b).

Table 1 Summary of 40Ar/39Ar geochronology data for the studied samples. Two aliquots were analyzed for each sample
Full size table
Fig. 9
figure 9

40Ar/39Ar age spectra of the two aliquots and weighted mean diagrams for the two combined replicates of biotite from the Macambira granite in (a) and (b); muscovite from a pegmatite dike in (c) and (d); and muscovite from a mylonite in (e) and (f). In the weighted mean diagrams, the gray rectangular boxes represent outliers

Full size image

The 40Ar/39Ar dates on muscovite from pegmatite (BDD-9) and mylonite (DH10-3) samples overlap within analytical error (Fig. 9c). The replicates of the BDD-9 muscovite yielded plateau dates of 501.75 ± 2.36 Ma (0.47% error; 76% 39Ar) and 501.87 ± 2.62 (0.52% error; 86% 39Ar), and provided a weighted mean date of 501.63 ± 0.59 Ma (Fig. 9d); this date is indistinguishable from the dates of the two DH10-3 aliquots from the mylonite (Fig. 9e), 498.24 ± 2.39 Ma (0.48% error; 86% 39Ar) and 501.41 ± 2.31 Ma (0.46% error; 76% 39Ar). Together, these data yield an integrated weighted mean date of 500.24 ± 0.57 Ma (Fig. 9f). The muscovite from the mylonite is richer in K (K/Caaverage ~112) than the muscovite from the pegmatite (K/Caaverage ~21).

Carbon and oxygen stable isotopes

The carbon and oxygen isotopic compositions of a total of 44 calcite samples from marbles and skarns were analyzed (ESM Table S6; ESM Fig. S3). Previous data from Nascimento et al. (2004) and Sial et al. (2015) has been compiled. The calcite from regional marbles are characterized by δ13C values between + 10.08‰ and − 1.17‰ and δ18O values between + 20.08 and + 27.43‰ (ESM Fig. 3). The isotopic compositions of calcite from skarns and marbles of the Bodó district exhibit a more restricted δ13C range, from + 7.72 to -7.46‰, but a broader δ18O range, from + 20.87‰ to of + 9.16‰.

T-XCO2 pseudosections of marble samples

To evaluate the effects of possible contact metamorphism and subsequent metasomatism, two isobaric T–XCO2 pseudosections were calculated at 3 kbar (Fig. 10) using the bulk-rock compositions of marbles with different mineral assemblages. The pressure chosen is compatible with metamorphic (Lima 1992; Souza et al. 2007; Cioffi et al. 2021) and skarn-forming conditions estimated for the Seridó Belt (Souza Neto et al. 2008). Sample BDD-15 consists of calcite, phlogopite, tremolite, diopside, scapolite, and plagioclase, whereas sample BDD-1.6 includes calcite, garnet, diopside, quartz, and plagioclase.

Fig. 10
figure 10

Pseudosections for samples (a) BDD-15 and (b) BDD-1.6. (a) T-XCO2 conditions were estimated using intersections of the compositional isopleths of biotite, plagioclase, clinopyroxene and amphibole. The peak mineral assemblage is indicated by the gray polygon. The gray arrow represents a schematic retrograde cooling path. The black arrow indicates the change in CO2 needed to form scapolite without a temperature change. (b) Intersections of compositional isopleths of garnet, clinopyroxene and calcite indicate that the field best represents the high-temperature stage (gray polygon). The gray arrow represents a schematic path derived from an increase in H2O. Amp = amphibole, An = anorthite, Bt = biotite, Cal = calcite, Cpx = clinopyroxene, Dol = dolomite, Ep = epidote, Grt = garnet, Hem = hematite, Kfs = k-feldspar, Phl = phlogopite, Pl = plagioclase, Qz = quartz, Scp = scapolite, Ves = vesuvianite, Wo = wollastonite, Zo = zoisite

Full size image

In the pseudosection for sample BDD-15 (Fig. 10a), the mineral assemblage observed in the rock relates to 675 °C at XCO2 = 0.9 to temperatures below 500 °C at XCO2 ≤ 0.1. Above this field, scapolite is absent, and it may represent the first assemblage. Phlogopite occurs under a wide range of temperatures. The compositional isopleths of XMg in phlogopite and diopside, and of XAn in plagioclase indicate a range of peak stabilities at ∼650 °C. Although amphibole is present under almost all conditions, the XMg amphibole isopleths indicate stability only below ~ 575 °C. The scapolite isopleths were not contoured because of the compositional discrepancies between the modeled scapolite and the analyzed mineral.

In sample BDD-1.6 (Fig. 10b), the mineral assemblage is represented by a wide range of stability conditions in the calculated pseudosection. Garnet with the observed composition (indicated by XGrs isopleths) occurs only at low values of XCO2. The combination of diopside and calcite isopleths shows that the clinopyroxene formed at higher temperatures (~ 600 °C) and at higher XCO2 than the garnet. Therefore, garnet formation required a decrease in XCO2 (< 0.2).

Discussion

C–O isotope modeling

The primary control on skarn formation in the Bodó district was lithologic. With minor exceptions, most of the skarns mapped in this area exist within marble lenses interlayered with paragneiss of the Jucurutu Formation or along the marble/paragneiss interface. As argued above, the marble of the Bodó district is variably affected by metasomatic modifications promoted by interactions with magmatic fluids on a regional scale. The original metamorphic mineral assemblage was likely dominated by calcite with phlogopite > tremolite or diopside, with the presence of diopside instead of tremolite being dictated by variations in the temperature conditions. In contrast, the metasomatic minerals are garnet ± wollastonite or scapolite.

Based on a comprehensive geochemical and isotopic (C and Sr) study of the well-preserved marble lenses of the basal Jucurutu Formation and a few preserved examples from the upper Seridó Formation schists, Nascimento et al. (2004, 2007) showed that regional metamorphism was an isochemical process, retaining the original (i.e., sedimentary) isotopic signature of the carbonate protoliths along many of the studied stratigraphic profiles even under amphibolite-facies conditions. They showed that the δ13C values of the Jucurutu marble lenses vary broadly, from + 11‰ to near 0‰, with similar δ18O value intervals. These isotopic data, together with the results obtained for the marble lenses in this work and for the skarns in the Bodó district, are plotted in Fig. 11. For clarification, the variation in the δ13C values of the regional marbles obtained by Nascimento et al. (2004, 2007) was sorted in accordance with a palinspastic restoration of their stratigraphic position within the sedimentary succession: The marble lenses occurring in the lower stratigraphic levels of the Jucurutu Formation were referred to by those authors as the ‘1st level’ group, while the uppermost lenses were referred to as the ‘3rd level’ group. Lenses within the upper Seridó schists were undivided, and their isotopic compositions were not included here. We drew mixing curves with variable CO2 mole fractions to model the extent of magmatic fluid–carbonate rock interactions, assuming that the initial δ13C-δ18O rock compositions corresponded to the average values of the primary compositions of the 1st - and 3rd -level groups, magmatic component featured values of δ13C = -8‰ and δ18O = + 6‰ according to Bowman (1998), and T = 550 °C. The curves were calculated according to the models of Taylor (1974, 1977) using the isotopic fractionation models proposed by Ohmoto and Ryes (1979) and Zheng (1999) for carbon and oxygen, respectively. The calculations were performed using the Alpha Delta web tool of Beaudoin and Therrien (2009).

Fig. 11
figure 11

δ18O versus δ13C diagram for samples from the Bodó District compared to other major deposits from the Seridó Belt and the regional marble of the Jucurutu Formation. The three Jucurutu levels are highlighted in the graph according to the interpretation by Nascimento et al. (2004). The gray curves represent mixing lines for fluid/rock interactions at 550 °C from initial values of 10‰ and 25‰ for δ13C and δ18O, respectively, whereas the dashed gray lines represent initial values of 3.5‰ and 25‰ for δ13C and δ18O, respectively. Ticks indicate fluid: rock ratios for closed and open systems, with values of 1/0.69, 10/2.39 and 100/4.61, respectively, following the same sequence on all curves. Field 1 represents primary magmatic waters from Sahlström et al. (2019) after Taylor (1974, 1986, 1992), Sheppard (1981), Valley (1986), and Bowman (1998). The D dotted lines indicate decarbonation paths. Arrow E indicates the qualitative effect of exchange with δ13C-depleted graphite or CH4-rich fluids

Full size image

Several findings arise from the C–O data in the diagram. First, the carbon and oxygen isotopic compositions of most of the studied calcite samples, whether from marble or skarn, plot on mixing lines departing from either the regional 1st - or 3rd -level compositions toward the magmatic endmember.

Most skarn calcite samples exhibit a depletion trend consistent with a fluid‒rock interaction model since carbon and oxygen depletion in heavy isotopes cannot be entirely produced by decarbonation/devolatilization processes. All mixing lines were determined using different XCO2 values, and the effect of decarbonation was modeled for temperatures of 300 and 550 °C (Fig. 11). In this model, all carbon can be released as CO2, while only 40% of oxygen can be liberated from the rock (calc-silicate limit; Valley 1986). The interaction of externally derived fluids from magmatic sources is common in contact metamorphism and skarn systems (Bowman 1998; Shin and Lee 2003; Timón et al. 2007; Orhan et al. 2011). The data in Fig. 11 reveal that different stratigraphic levels of the regional marbles in the Jucurutu Formation are needed to fit the modeled mixing lines between sedimentary and magmatic endmembers. This interpretation is supported by the two distinct groups of mixing lines along which almost all the samples plot, indicating the need for different sedimentary signatures. Moreover, calcite from all deposits shows a trend toward the igneous field, suggesting interactions with magmatic fluids.

Two samples of late calcite formed during the retrograde stages of the Bodó District skarns also plot within the magmatic field, indicating a pure igneous signature (Fig. 11). The data dispersion across the multiple curves may suggest that marble, skarn, and late calcite interacted with fluids over a wide range of XCO2 values, varying from 0.5 to almost 0.01, and at different fluid: rock ratios, which is plausible since skarn systems are widely heterogeneous (Meinert et al. 2005; Gaspar et al. 2008; Ismail et al. 2014). This interaction was probably not coeval throughout all the marble and skarn lenses and could have happened under different temperature conditions over time. A higher temperature than the one used in the modeling would shift the mixing lines to the left, whereas a lower temperature would lead to a shift to the right. Nonetheless, the fluid–rock interaction model was calculated using only a temperature of 550 °C for simplification purposes, which is a reasonable estimation of the prograde–retrograde transition. Samples that plot below the mixing lines must result from an exchange between magmatic fluid and δ13C-depleted graphite or from the interaction with CH4-rich fluids, indicating the presence of organic matter. All the alternatives above could result in a trend following arrow E in Fig. 11.

Linking magmatic events to skarn formation in the Bodó district

Tungsten skarns are commonly associated with differentiated calc-alkaline metaluminous to peraluminous granites of crustal origin (Meinert et al. 2005; Hulsbosch et al. 2016; Legros et al. 2020). Skarn formation in the Seridó Mineral Province has been classically linked to the regional-scale intrusion of high-K calc-alkaline granitic plutons that are usually spatially associated with the main deposits (Souza Neto et al. 2008; Santos et al. 2014). One of the best studied districts of the Seridó Mineral Province, Brejuí-Boca de Lage district, which includes the Brejuí deposit and nearby occurrences (Cavalcante et al. 2016), had been inferred to be associated with the intrusion of the Acari batholith. However, zircons from the porphyritic and equigranular granitic facies that constitute the bulk of the batholith were dated at ca. 577 Ma and 572 Ma, respectively (Archanjo et al. 2013), in clear dissociation with the Re-Os molybdenite date at ca. 554 Ma obtained by Hollanda et al. (2017).

The Re‒Os system in molybdenite is a robust geochronometer for mineralization because it is a resilient system during high-grade metamorphism and deformation (Stein et al. 2001) with a relatively moderate to high closure temperature (~ 500 °C; Suzuki et al. 1996). Thus, Re‒Os dates are widely used to determine the crystallization age of molybdenite during hydrothermal stages (Stein et al. 2001; McInnes 2005; Chiaradia et al. 2013) and are integrated with U‒Pb and 40Ar/39Ar dates on igneous bodies to link magmatism and hydrothermal mineralization (Li et al. 2020, 2021; Zhang et al. 2020).

Magmatic activity throughout the Bodó district started at ca. 534 Ma while the thermal activity in the district ended at ca. 490 Ma (Fig. 12). The W–Mo mineralization occurred at 510 ± 2 Ma, as constrained by Re‒Os dating of molybdenite (Hollanda et al. 2017). Considering the closure temperature of the Re-Os system and the temperature for the Bodó district skarns obtained in this work, the Re-Os date can be attributed to the transition from the prograde stage to the retrograde stage. This date is significantly younger than the ca. 534 Ma date of the nearest granite, the high-K calc-alkaline magmatic Macambira pluton. This indicates that the magmatic fluids related to the mineralization were not derived from this granite body. In contrast, the Re-Os molybdenite date matches a set of U‒Pb ID TIMS dates in the range between 515 and 509 Ma on columbite-tantalite from homogeneous and heterogeneous pegmatites exposed within the Seridó Mineral Province (Baumgartner et al. 2006). Since the U–Pb systematics in columbite-tantalite are comparable to the U–Pb systematics in zircon in a way that both represent crystallization ages (e.g., Romer and Wright 1992; Romer and Smeds 1994), the dates obtained by Baumgartner et al. (2006) are the best estimates for the crystallization age of the pegmatitic magmas in the Seridó Mineral Province. Therefore, considering the available data so far, these dates are used here to establish a connection between the source of magmatic-hydrothermal ore fluids with pegmatites. The source of heat and fluid for skarn formation, however, requires further investigation. In this sense, some possibilities involve skarn formation during regional metamorphism (Lefebvre et al. 2019a, b; Romer et al. 2022) or during Ediacaran granitic magmatism, followed by later metal enrichment and/or remobilization as a result of granitic pegmatite emplacement.

Fig. 12
figure 12

Summary of Cambrian geochronological data from the Bodó district and U‒Pb ages of columbite-tantalite from other pegmatites in the Seridó Mineral Province (data from Baumgartner et al. 2006). Re‒Os molybdenite age from Hollanda et al. (2017)

Full size image

The 40Ar/39Ar plateau date of 501.9 ± 2.7 Ma obtained for a local E–W-trending pegmatite dike reflects the cooling age of this volatile-rich magma. The coincidence of the 40Ar/39Ar dates in muscovite and biotite from pegmatite, granite, and mylonite rocks at 500 − 490 Ma corresponds to the cooling stage of the Seridó Belt (Corsini et al. 1998) with temperatures between 310 and 280 °C (assuming experimental cooling rates ranging from 10 to 1 °C/Ma; Harrison et al. 1985), and the late reactivation of shear zones on a regional-to-local scale.

Metamorphic/metasomatic evolution

Phase equilibria modeling of skarns is challenging due to their chemical, textural, and mineralogical heterogeneity. This has led to an extrapolation of the interpretations obtained from the marbles, as both skarn and marble share a common mineralogy. The results obtained through pseudosection modeling show that the peak temperature ranged between 650 and 600 °C at different XCO2 depending on the main mineral assemblage (Fig. 10). These temperature estimates agree with the geothermometry calculations obtained by Sallet et al. (2021) for the Seridó schists and Jucurutu gneiss at the contact with the pegmatites. In sample BDD-15, petrographic observations indicate that scapolite formed after the main silicate phases represented by the field composed of diopside–phlogopite–tremolite–plagioclase–calcite. The scapolite growth encompassing the other mineral phases suggests a fluid-induced process and provides evidence of a Cl-rich fluid, as shown by electron probe micro-analysis (Cl ranging from 0.109 to 0.237 a.p.f.u.). Scapolite-bearing rocks usually form in the presence of CO2-rich fluids (Aitken 1983; Kwon et al. 2020), supporting the results of the presented pseudosection, in which an increase in XCO2 marks the transition towards the scapolite field under near-isothermal conditions (black arrow in Fig. 10). The incursion of acicular crystals of wollastonite in the marble suggests a process induced by fluids commonly associated with Si-rich metasomatism (Fig. 4h). The retrograde path, marked in the pseudosection by the compositions of amphibole and phlogopite, indicates a temperature decrease to below ~ 575 °C.

The mineral assemblage of sample BDD-1.6 shows slightly different XCO2 conditions for the formation of garnet. Although clinopyroxene is stable under a wide range of T-XCO2 conditions, compositional garnet isopleths are limited to a narrow field at lower XCO2. The proposed evolution path started at ~ 600 °C and was characterized by the crystallization of the diopside–phlogopite–plagioclase–calcite association. Grossular II could only have formed in response to a decrease in XCO2 induced by H2O influx in the system. The assemblage composed of clinopyroxene–grossular I–calcite–plagioclase–quartz can be stable over a wide range of XCO2 values, whereas the late grossular II assemblage is stable only under low XCO2 (< 0.2) conditions.

Including the skarns in these interpretations, fluid-induced processes (e.g., metasomatism) evidently played an important role in the genesis of the studied rocks. An open system setting is supported by the presence of monomineralic garnet zones (Korzhinskii 1965; Putnis and Austrheim 2010), cm-sized garnet porphyroblasts (Fig. 5c-e), and vesuvianite, wollastonite, garnet, and scapolite with textural evidence of growth along corroded borders.

The decrease in XCO2 in the fluid (from 0.6 to > 0.1) led to the replacement of early pyroxene by garnet in the marble and skarn (Fig. 10). This large variation in XCO2 could be explained by the heterogeneities observed in skarn systems and is consistent with the data obtained from carbon and oxygen stable isotopes, which indicate that XCO2 can range from approximately 0.01 to above 0.5 (Fig. 11).

Conclusions

The data obtained in this study lead to the following conclusions

  • C and O stable isotope signatures of calcite from the Jucurutu marble lenses did not change during regional metamorphism. In contrast, depletion in isotopically heavy C and O occurred during metasomatic events triggered by the circulation of hydrothermal fluids of magmatic origin. Consequently, mixing lines between the regional Jucurutu marble and a magmatic endmember are consistent with a fluid–rock interaction model in which the observed carbon and oxygen isotope depletions could not have been produced by devolatilization alone.

  • Wide variation in C and O isotope values, and corresponding range in fluid XCO2 at different fluid: rock ratios, indicates a system with significant variability in its fluid composition during the formation of the different mineral assemblages.

  • T-XCO2 pseudosections produced for marbles indicate a peak metamorphic temperature of 650–600 °C at different XCO2 values depending on the main mineral assemblage. The formation of coarse garnet required a decrease in XCO2 and, consequently, a higher H2O influx than required for clinopyroxene formation. In contrast, scapolite in the Bodó district marbles would have formed at almost isothermal conditions at ~ 650 °C with an increase in XCO2. The transition from prograde to retrograde stage started at ~ 550 –500 °C as indicated by the crystallization of the second generation of garnet (i.e., garnet II) and tremolite.

  • The integration of new U‒Pb zircon (536.6 ± 3.4 Ma) and 40Ar/39Ar biotite (490.48 ± 1.00 Ma) dates for the Macambira granite, as well as a 40Ar/39Ar muscovite date for a local pegmatitic dike (501.36 ± 0.59 Ma), coupled with a previously reported Re‒Os molybdenite date (510 ± 2 Ma) for the Bodó deposit, and U‒Pb zircon dates for the Seridó Mineral Province pegmatites (515–509 Ma) reveal that skarn mineralization was likely coeval with the emplacement of the pegmatitic dikes and stocks, while the 40Ar/39Ar dates represent cooling ages for the regional magmatic-hydrothermal system.

  • The data presented in this study contribute to unraveling the complex metamorphic, metasomatic, and magmatic processes related to W-skarn mineralization in the Seridó Mineral Province. The results shed light on the skarn and marble formation conditions, the role of fluid-rock interactions, and ore-forming fluid sources. The integration of novel and published data provides insights for further investigation of the magmatic-hydrothermal system, and the potential for metal exploration in the Seridó Mineral Province.