Introduction

In the Carajás Mineral Province (CMP), located in the southeastern margin of the Amazon Craton, lies a number of Precambrian Cu–Au deposits having different metal associations and ages. Noteworthy are the world class (>200 Mt) Igarapé Bahia, Salobo, Sossego (Sequeirinho ore body), and Cristalino deposits, all with substantial magnetite content, which are Archean in age. They have been classified as iron oxide–copper–gold (IOCG) deposits (Tallarico et al. 2005; Monteiro et al. 2008; Grainger et al. 2008; Xavier et al. 2010; Groves et al. 2010; Moreto et al. 2013), although this interpretation is contentious for the Igarapé Bahia deposit (Villas and Santos 2001; Dreher et al. 2008). Smaller IOCG deposits also occur such as the Neoarchean Bacaba, Bacuri and Castanha (Augusto et al. 2008; Pestilho and Monteiro 2008; Moreto et al. 2013), and the Paleoproteozoic Alvo 118 and Sossego orebodies (Tallarico 2003; Torresi et al. 2012; Moreto et al. 2013), in addition to the minor polimetallic (Cu–Au ± W–Bi–Sn) Águas Claras, Breves, and Estrela deposits, which formed during the Paleoproterozoic (Silva and Villas 1998; Tallarico et al. 2004; Botelho et al. 2005; Xavier et al. 2012).

IOCG-type deposits (Hitzman et al. 1992; Williams et al. 2005) have become prime targets for mineral exploration worldwide because they commonly show high ore grade and tonnage. Classic examples are the Olympic Dam and Ernest Henry deposits in Australia or Candelaria and Punta del Cobre deposits in Chile. Although proposed as a new genetic class of deposits (Hitzman et al. 1992), there is no consensus regarding the geological environment where they form and the sources of fluids, metals, and sulfur.

Regarding the Carajás IOCG deposits, there is much debate about how many Cu–Au mineralization events occurred in Archean-Proterozoic times, the simultaneity of the sulfide deposition in deposits with similar mineralization styles, the origin of high-salinity fluids, the age of mineralization, and the role played by granitic intrusions in the deposit genesis (Lindenmayer 1990; Lindenmayer et al. 2005; Villas and Santos 2001; Requia et al. 2003; Tallarico et al. 2005; Villas et al. 2005, 2006; Chiaradia et al. 2006; Dreher et al. 2008; Monteiro et al. 2008; Carvalho 2009; Xavier et al. 2008, 2010, 2012; Torresi et al. 2012; Moreto et al. 2013). Despite the advances in the knowledge of these deposits, several questions remain unanswered. The metallogenetic study of the Cu–Au Visconde deposit presented here provides insights to help solving some of those controversial topics, especially the nature of the mineralizing fluids and the temporal relationships between the ore formation and the granite emplacement.

The Visconde deposit (20–50 Mt at 1.0 % Cu, 0.28 g/t Au; Benevides Aires, personal communication) is one of the satellite ore bodies of the large Sequeirinho/Sossego deposit located approximately 15 km to the west. Recent studies (Craveiro 2011; Craveiro et al. 2012) have shown that, with respect to the stages and types of alteration, styles of mineralization, and fluid properties, the Visconde deposit is very similar to other Archean Cu–Au deposits of Carajás, especially those located near the boundary between the southern edge of the Carajás domain and the Transition domain, namely the Bacaba, Bacuri, Castanha, Alvo 118, Cristalino, Jatobá, and Sequeirinho/Sossego deposits (Fig. 1). To contribute to understanding of the genesis of the Visconde deposit, this study attempts to define the likely sources of the hydrothermal fluids based on stable isotope (O, H, and S) and fluid inclusion data, in addition to constrain the age of mineralization (Pb isotope).

Fig. 1
figure 1

Geological map of the CMP with the location of the Visconde and other cupro-auriferous deposits (compiled from DOCEGEO 1988; Gomes and Dall'Agnol 2007; Feio et al. 2012)

Full size image

Materials and methods

Core samples from eight boreholes drilled in the Visconde deposit area were studied petrographically, some with scanning electron microscopy coupled with energy dispersive X-ray spectrometry (EDS), at the Geosciences Institute of the Federal University of Pará (IG-UFPA).

For stable isotope analysis, pure mineral fractions were separated manually with the aid of a binocular lens from ground rock samples. These fractions consisted of silicates (32), oxides (4), and sulfides (12). Analyses were performed on albite, magnetite, quartz, and tourmaline (oxygen), actinolite, biotite, chlorite, and epidote (oxygen and hydrogen), chalcopyrite and bornite (sulfur). The isotopic composition of oxygen in silicates (except albite) was determined at the Center for Granite Research -Stable Isotope Laboratory (NEG-LABISE) of the Federal University of Pernambuco. Hydrogen in silicates, oxygen in magnetite and albite, and sulfur in sulfides were determined at the Queen's Facilities for Isotope Research (QFIR) at Queen's University (Canada). The BrF5 method of Clayton and Mayeda (1963) was used to extract oxygen. For hydrogen extraction the procedures followed those described in Kyser and Kerrich (1991). Stable isotope measurements were made using a Thermo-Finnigan Delta V Advantage (LABISE) and a Finnigan MAT 252 mass spectrometer (QFIR), and the results are expressed in the conventional delta per mil notation (δ, ‰). The oxygen and hydrogen isotopic compositions are reported in relation to Vienna Standard Mean Ocean Water (VSMOW) and the sulfur values in relation to Vienna Canyon Diablo Troilite (VCDT).

Fluid inclusion studies were carried out on five doubly polished thin sections prepared from core samples of three drill-holes (VISD-FD 35, VISD-FD 36, and VISD-FD 37) representing the altered granite, the breccia matrix and veins that crosscut the altered gabbro/diorite and metadacite. The studied inclusions were trapped in quartz, scapolite, apatite and calcite. Approximately 500 phase change temperatures were measured in tens of inclusions using a LINKAM THMS600 heating-freezing stage coupled to a conventional petrographic microscope at the Metallogenesis Laboratory of the Geosciences Institute of the Federal University of Pará (UFPA). Fluid compositions were estimated from eutectic temperatures obtained experimentally in simple chemical systems by Borisenko (1977) and Davis et al. (1990).

The geochronological analyses were carried out at the Laboratory of Isotope Geology (Pará-Iso), IG-UFPA. Lead total digestion and leaching methods were used on chalcopyrite ore samples to date the mineralization. The experimental procedure followed the techniques described in Galarza et al. (2008) and Romero et al. (2013). During the total digestion, ≈100 mg of chalcopyrite underwent cold acid attack, and for the leaching procedure, 400 mg were progressively attacked by hot acids, and the resulting solutions were collected in six different steps (L1 to L6). Specific mixtures of tri-distilled HCl, HNO3, and HBr were employed in both methods. Lead was separated by ion-exchange chromatography using an Eichrom® Sr 50–100 μm resin. The lead isotopic ratios were measured using a Thermo-Finnigan Neptune ICP-MS and corrected for mass discrimination by exponential law using the 205Tl/203Tl ratio. The total Pb blank of the analytical procedure ranged from 0.33 to 1.50 ng. Pb–Pb isochron ages were calculated with the ISOPLOT software (v.3.68) of Ludwig (2008). The errors on lead isotopic ratios were quoted at a 95 % confidence level.

Dating of a granitic body with no evidence of mineralization was accomplished by the Pb-evaporation method on zircon monocrysts (Kober 1986, 1987). Different size fractions of zircon grains were obtained by sieving and heavy liquid separation, but only those between 0.075 and 0.180 mm provided suitable grains for analysis. The experimental procedures, lead isotopic determinations, and age calculations are described in detail in Klötzli (1997) and Barreto et al. (2013). The analyses were performed on a TIMS Finnigan MAT 262 mass spectrometer at increasing temperature steps of 1450, 1500, and 1550 °C, with five blocks in each step. Ages were obtained using the weighted mean of 207Pb/206Pb ratios of each block, the most reliable values being those yielded at the highest temperature steps. The ISOPLOT software (v.3.68) of Ludwig (2008) was used for age calculations.

Regional geological setting

The CMP lies within the Archean southern Amazon Craton (Fig. 1). Two tectonic blocks are distinguished in the CMP: the southern Rio Maria Granite-Greenstone Terrane and the northern Carajás Domain. Given that the boundaries between these two tectonic blocks are not well defined, Dall’Agnol et al. (2006) proposed the existence of a Transition Domain, assumed to be an extension of the Rio Maria terrane intensely affected by the Neoarchean magmatic and tectonic events recorded in the Carajás Domain.

The Mesoarchean Rio Maria terrane consists of rocks with ages between 2.97 and 2.87 Ga, which comprise the greenstone-type sequences of the Andorinhas Supergroup and various granites, including trondhjemite-tonalite-granodiorite (TTG) series, Mg-rich granites (Rio Maria sanukitoid suite; Oliveira et al. 2009) and leucogranites (Dall’Agnol et al. 2006; Feio et al. 2012).

The Carajás Domain is largely made up of metavolcanic-sedimentary and meta-sedimentary rocks that were originally deposited in the intracratonic Carajás basin (Gibbs et al. 1986). The basin oldest unit is represented by the mafic-ultramafic rocks, iron formations, and schists of the Rio Novo Group (>2.76 Ga). The overlying Itacaiunas Supergroup (2.76–2.73 Ga) comprises the Igarapé Salobo, Igarapé Bahia, Grão Pará, Igarapé Pojuca, and Buritirama groups, which show different metamorphic grades (greenschist to amphibolite facies) and varying deformation intensities related to the Itacaiúnas transpressive ductile shear zone (Pinheiro and Holdsworth 2000). Amphibolites, paragneisses, metavolcanic rocks, pelitic schists, meta-arkoses, and iron formations are common lithotypes and host most of the Fe, Cu, and Au deposits of Carajás.

The Rio Novo Group rocks are cut by the mafic-ultramafic-layered Luanga Complex (2.76 Ga; Machado et al. 1991) and by metaluminous, alkaline granites that transect units of the Itacaiunas supergroup. The syn- to late-tectonic granites are represented by the Estrela, Serra do Rabo, Cristalino, and Igarapé Gelado intrusions, with ages ranging from 2.76 to 2.73 Ga (Huhn et al. 1999; Barros et al. 2001; Sardinha et al. 2006). The Igarapé Salobo group hosts the ca. 2.5 Ga Itacaiunas and Old Salobo granites (Machado et al. 1991; Souza et al. 1996).

These older rocks are unconformably overlain by the siliciclastic and minor volcaniclastic rocks of the Aguas Claras Formation, with ages varying from 2.68 to 2.64 Ga (Nogueira et al. 1995; Dias et al. 1996; Trendall et al. 1998).

In the Transition domain, the oldest rocks belong to the Pium Complex, which includes granulites formed at ~3.0 Ga and metamorphosed at 2.86 Ga (Rodrigues et al. 1992; Pidgeon et al. 2000), as well as diopside-norites dated at 2745 ± 1 and 2744 ± 1 Ma (Pb–Pb zircon; Santos et al. 2013a, b).The Xingu Complex consists mainly of undifferentiated granites, gneisses and migmatites associated with subordinate amphibolites, metamorphosed at ca. 2.86 Ga (Machado et al. 1991).

A few intrusive suites have been recognized in the Transition Domain. The Neoarchean Plaquê and Planalto suites consist of numerous E-W-oriented granitic bodies, the former (~2.73 Ga, Avelar et al. 1999) being composed of metaluminous to peraluminous, alkaline to calc-alkaline lens-shaped intrusions (Macambira et al. 1996). The latter (2.74–2.73 Ga; Huhn et al. 1999; Sardinha et al. 2004; Feio et al. 2012) comprise subalkaline, metaluminous to weakly peraluminous stocks. The Cateté intrusive suite is represented by mafic-ultramafic intrusions. Geochronological data are only available from a gabbro (Serra da Onça intrusion) and yield a Sm–Nd isochronic age of 2378 ± 55 Ma (Macambira and Tassinari 1998) and a SHRIMP U–Pb age on zircon of 2766 ± 6 Ma (Lafon et al. 2000).

After a long period with no rock record, a magmatic episode took place at 1.88 Ga throughout the CMP during an extensive or transtensive event (Pinheiro and Holdsworth 2000). Several isotropic A-type granites were then emplaced, such as the Cigano, Jamon, Bannach, Musa, Pojuca, Seringa, Serra dos Carajás, and Rio Branco intrusions, some of them mineralized with Cu and Au (Machado et al. 1991; Dall’Agnol et al. 2005; Santos et al. 2013a, b).

The Visconde Cu–Au deposit

The Visconde deposit (Craveiro et al. 2012; Fig. 2) is located within a regional WNW-ESE-trending shear zone that is approximately 60 km long at the contact between a metavolcanic-sedimentary sequence of the Itacaiunas supergroup (Grão Pará Group?) and the basement (Xingu Complex and other granitic units). The Cu–Au Bacaba, Castanha, Alvo 118, Cristalino, Jatobá, and Sossego deposits also occur along that shear zone. The Sossego deposit is composed of the Sequeirinho, Sossego, Pista, Baiano, and Curral ore bodies, with the first two representing more than 85 % of the ore reserves.

Fig. 2
figure 2

Geological map (a) and a N-S cross-section (b) of the Visconde deposit (modified from VALE, unpublished, 2003)

Full size image

The Visconde deposit is hosted by felsic metavolcanic rocks, the Serra Dourada granite, gabbros/diorites and, to a lesser extent, ultramafic rocks (Fig. 3). These rocks are hydrothermally altered and crosscut by mylonitic zones, along which abundant hydrothermal minerals formed, especially biotite. The felsic metavolcanic rocks, described as metadacites according to Craveiro et al. (2012), are gray colored and fine grained and show isotropic to foliated varieties. When foliated, they display fine alternating bands composed of quartz, plagioclase, muscovite or biotite, and/or hornblende. The Serra Dourada granite (2,60 ± 22 Ma, Moreto et al. 2011) is coarse-grained, isotropic, although incipiently foliated in places, and grayish to pinkish in color. Quartz, plagioclase, and K-feldspar are the main mineral constituents, giving the rock a monzogranitic to granodioritic composition with calc-alkaline affinity. This intrusion bears evidence of Na metasomatism indicated by the common occurrence of chessboard albite (Craveiro et al. 2012; Feio et al. 2013).

Fig. 3
figure 3

Main features of altered and mineralized rocks of the Visconde deposit. Simplified description of representative drill cores (VSD-35, VSD-35, and VSD-43)

Full size image

Gabbro/diorites occur as stocks and dikes, and are apparently intrusive in both the metadacites and the Serra Dourada granite. These gabbro/diorites are medium-grained and, when less altered, display primary relict features such as sub-ophitic texture and straight contacts among plagioclase, Mg hornblende, or quartz crystals, in addition to primary magnetite crystals with exsolved ilmenite lamellae.

Ultramafic lenticular bodies occur in some gabbros and the Serra Dourada granite, being represented by magnesite- and talc-rich serpentinites, with lesser amounts of pentlandite, Ni-bearing chalcopyrite and magnetite. Most rocks are strongly oriented in the E-W direction and steep-dipping in response to the development of a ductile–brittle to brittle shear zone in the Neoarchean (2.76–2.65 Ga; Pinheiro and Holdsworth 2000).

The Visconde granite

A slightly altered, highly fractured (Fig. 4a), but nonmineralized granitic body, referred to here as the Visconde granite, occurs in the northern part of the deposit area (Fig. 2). It is reddish pink, isotropic, medium to coarse grained, and displays a heterogranular texture, locally with graphic intergrowth (Fig. 4b, c). Microcline, quartz, plagioclase, and biotite are the major minerals, with a modal composition of syenogranite. Microcline crystals are coarse and commonly perthitic. Quartz crystals are anhedral to subhedral, some with weak undulose extinction. Plagioclase occurs as exsolution lamellae in microcline and as subhedral crystals, some showing kink bands and microfractures. Most biotite occurs interstitially as isolated flakes. Zircon is the main primary accessory mineral, whereas minor sericite, chlorite, epidote, and clay minerals represent the secondary phases, generally concentrated near fractured zones (Fig. 4b, d).

Fig. 4
figure 4

Features of the Visconde granite. a Photograph of an outcrop located at the km 7 along the Sossego mine road. b The pinkish gray color of the Visconde granite contrasting with the dark greenish color of the chloritization halos along the fractures.; c Photomicrograph of the hypidiomorphic granular texture of the rock with a local graphic intergrowth. d Photomicrograph of chlorite lamellas partially replacing microcline. Chlorite is one of the most abundant hydrothermal alteration product in the Visconde granite. Chl chlorite, Mic microcline, Plg plagioclase, Qz quartz

Full size image

This body was initially described as a 1.88 Ga anorogenic granite (Vale 2003). More recently, based on petrographic data, it was tentatively correlated to the Planalto suite (Feio et al. 2012), with age determined at 2.74 Ga (Huhn et al. 1999; Sardinha et al. 2004; Feio et al. 2012). However, the age of the granite remains an open question. Despite having fractured zones potentially favorable to fluid circulation, this granite shows no signs of alkali-metasomatism or mineralization, suggesting that its emplacement took place after the Visconde mineralizing event.

Hydrothermal alteration and mineralization

The hydrothermal alteration evolved in various stages and was largely controlled by tectonic evolution from a ductile to a brittle regime. During the first stage of the early sodic-calcic alteration, the rock underwent a pervasive replacement of primary mineral assemblages. Chessboard albite in the Serra Dourada granite (Fig. 5a) and marialitic scapolite that formed in the gabbros/diorites at the expense of the plagioclase (Fig. 5b) are typical minerals of this stage. Early veins (albite-epidote, scapolite-tourmaline, tourmaline-quartz, and actinolite; Fig. 5c, d) and aggregates of actinolite, locally associated with magnetite (with fine inclusions of ilmenite and rutile; Fig. 6g, h), are the most conspicuous early hydrothermal products. Disseminated chalcopyrite and pyrite are common. More magnetite precipitated later on, locally forming magnetitites. The potassic stage developed next and characteristically produced biotite (Fig. 5e) and K-feldspar. In highly deformed rocks, biotite is so abundant and the mylonitic foliation so pronounced that they resemble a schist (Fig. 5e). In the deformed ultramafic rock lenses, talc and magnesite also formed in addition to biotite (Fig. 5g). The corresponding alteration assemblages also include quartz, allanite, magnetite, chalcopyrite, and molybdenite. Next abundant sulfides (chalcopyrite >> bornite) precipitated, while a new episode of sodic-calcic and magnesian alteration occurred in the outer zones, with albite, epidote, chlorite, calcite, and minor fluorite filling fracture-controlled veins and veinlets (Fig. 4h, i). The hydrothermal history ended with a calcic-magnesian alteration characterized by the precipitation of chlorite, calcite and local gypsum or hematite in veinlets and/or breccias (Fig. 6i).

Fig. 5
figure 5

Photomicrographs of common alteration assemblages at the Visconde deposit. a Chessboard albite in the Serra Dourada granite rocks. b Scapolite aggregate crosscut by a quartz vein. c Early tourmaline-bearing vein hosted by metadacite. d Mg hornblende partially replaced by actinolite (in a biotite-rich altered diorite). e Biotitization zone in the Serra Dourada granite. f Gabbro/diorite affected by mylonitization and replaced by scapolite and biotite. g Mylonitized ultramafic rock with talc, magnesite, and magnetite. h Chessboard albite and chlorite in a millimeter-thick late vein. i Epidote- and calcite-bearing vein in a metadacite

Full size image
Fig. 6
figure 6

Textural features of the main sulfides and oxides at the Visconde deposit. a Chalcopyrite and pyrite disseminations in gabbro/diorite. b Chalcopyrite and pyrite along a mylonitic foliation plane developed in metadacite. c Breccia matrix containing molybdenite partially replaced by chalcopyrite (in gabbro/diorite). d Early pentlandite crystal replaced by chalcopyrite along a cleavage plane. Gabbro/diorite. e Pyrite clasts cemented by chalcopyrite-rich breccia matrix, in Serra Dourada granite. f Apatite clasts immersed in a chalcopyrite- and bornite-rich breccia matrix, in gabbro/diorite. g Rutile inclusions in a magnetite crystal, in gabbro/diorite. h Disseminated ilmenite close to a magnetitite body, in gabbro/diorite. i Needle-like hematite in a late vein cutting a rhyolite dyke

Full size image

The ore occurs as tectonic breccias, veins, veinlets and disseminations (Fig. 6a, b). The breccias form bodies that transect the potassic alteration zones (Fig. 6c–f). These breccias consist of clasts of hydrothermally altered rocks immersed in a matrix of sulfides (locally over 60 %), chiefly chalcopyrite, together with minor bornite, pyrite, chalcocite, covellite, and digenite. In veins and veinlets, the sulfides co-precipitated with sodic-calcic minerals. The paragenetic sequence in the Visconde deposit is summarized in Fig. 7.

Fig. 7
figure 7

Mineral paragenetic sequence of altered and mineralized rocks from the Visconde deposit

Full size image

Isotope geochemistry

Oxygen isotopes

δ 18 O VSMOW data from hydrothermal minerals representative of the different alteration stages recorded in the Visconde deposit are reported in Table 2. The measured δ 18 O VSMOW values in the silicates vary between +2.3 and +11.6 ‰, although for each mineral the range is narrow. The greatest variation is for chlorite (+2.3 to +5.1 ‰), though it does not reach 3 ‰. For magnetite, the values range from −1.3 to +0.5 ‰.

The formation temperatures of some mineral pairs (Table 1) were calculated based on oxygen isotope fractionation factors (Wenner and Taylor 1971; Matsuhisa et al. 1979; Zheng 1991, 1993). Mineral pairs were selected after careful petrographic examination to ensure textural equilibrium and, as much as possible, to have remained unaffected by subsequent alteration stages. The actinolite-magnetite pair from the early aggregates yielded the highest temperatures of 410 ± 21 °C, below the temperature of 470 °C obtained by fluid inclusion microthermometry in scapolite and quartz crystals that precipitated during the early sodic-calcic alteration (see below). A quartz-tourmaline pair from a vein formed during this same alteration stage yielded temperature of 355 °C, whereas the albite-epidote pair from a later vein, coeval with the main sulfide deposition, recorded a lower temperature of 237 °C.

Table 1 Temperatures estimated from oxygen isotopic composition of hydrothermal minerals from the Visconde deposit
Full size table

Hydrogen isotopes

δD VSMOW values from hydroxyl-bearing silicates (actinolite, biotite, epidote, and chlorite) are presented in Table 2. The values for actinolite (early sodic-calcic alteration) range from −47 to −61 ‰ and coincide partly with those for biotite from the potassic zones (−54 to −66 ‰). Chlorite and epidote (later stages) yielded values from −63 to −66 and from −18 to −49 ‰, respectively. Except for epidote, the δD VSMOW results for the other minerals overlap largely.

Table 2 Isotopic composition (δ 18 O VSMOW and δD VSMOW) of silicate minerals, magnetite, and fluid of the Visconde deposit (values in ‰)
Full size table

Isotopic composition of the hydrothermal fluids

The temperatures calculated for each alteration stage were used to estimate the hydrogen and oxygen isotopic composition of the fluids (Table 2), using the fractionation factors for actinolite-H2O, magnetite-H2O, quartz-H2O, tourmaline-H2O, biotite-H2O, epidote-H2O, albite-H2O, and chlorite-H2O (Wenner and Taylor 1971; Bottinga and Javoy 1973; Suzuoki and Epstein 1976; Matsuhisa et al. 1979; Graham et al. 1980, 1984; Zheng 1991, 1993).

For the early sodic-calcic alteration (actinolite and magnetite), which recorded temperatures of 410 ± 21 °C, the \( {\delta}^{18}{O}_{{\mathrm{H}}_2\mathrm{O}} \) and \( \delta {D}_{{\mathrm{H}}_2\mathrm{O}} \) values fall within the ranges +5.6 ± 0.2 to +9.4 ± 0.1 and −32 to −18 ‰, respectively. Lower \( {\delta}^{18}{O}_{{\mathrm{H}}_2\mathrm{O}} \) values (+3.3 to +5.7 ‰) were obtained for the fluids in equilibrium with tourmaline and quartz (at 355 °C). The biotite formation temperature is uncertain. Nevertheless, the fluid isotopic composition was also calculated at 355 °C because, within the paragenetic sequence, it is the closest temperature suggested for the biotite appearance. Thus, fluids related to the potassic alteration show \( {\delta}^{18}{O}_{{\mathrm{H}}_2\mathrm{O}} \) values between +4.8 and +7.2 ‰ and \( \delta {D}_{{\mathrm{H}}_2\mathrm{O}} \) values spread over the −12 to −3 ‰ interval. At the final alteration stages, the temperature was approximately 229 ± 11 °C and the fluids became 18 O depleted (δ 18 O VSMOW = −1.3 ± 0.4 to +4.4 ± 0.4 ‰) but highly variable in δD VSMOW (−28 to −24 ‰ for chlorite and −26 to +5 ‰ for epidote).

Sulfur isotopes

Sulfur isotope compositions (δ 34 S VCDT) were determined in chalcopyrite and bornite samples (Table 3). The results show narrow variation (+0.5 to +3.4 ‰), the higher values tending to be recorded in late vein and breccia sulfides. Disseminated and early vein sulfides reveal compositions slightly above zero (+0.5 to +1.5 ‰) whereas one sample (cpy 24–32) has a negative value (−1.2 ‰). Although chalcopyrite and bornite from sample 24–32 were apparently in equilibrium, attempts to calculate the formation temperature returned unrealistic values.

Table 3 Sulfur isotope composition (δ 34 S CDT) of sulfides from the Visconde deposit (values in ‰)
Full size table

Fluid inclusion microthermometry

Data from fluid inclusions in hydrothermal minerals revealed three distinct aqueous fluids characterized as the following chemical systems: H2O–NaCl–CaCl2 ± MgCl2 (fluid 1), H2O–NaCl–FeCl2 ± MgCl2 (fluid 2), and H2O–NaCl ± KCl (fluid 3). Only found in secondary inclusions, fluids 2 and 3 seem to have circulated after the copper-gold mineralizing event and will not be discussed hereafter.

Fluid 1 was primarily responsible for the wall-rock alteration and mineralization. It was trapped in two-phase (L + V), three-phase (L + V + S) and rare multi-phase (L + V + S1 + S2) fluid inclusions, as observed at room temperature, in quartz, scapolite, apatite, and calcite host minerals. Saturated inclusions, with daughter crystals, were not identified in carbonate. Quartz and calcite samples are from the altered granite (both hydrothermal and recrystallized magmatic grains) and the breccia matrix, respectively. In turn, scapolite and apatite samples are from veins that crosscut the gabbro/diorite and the metadacite, respectively. Those fluid inclusions occur individually or clustered along intra-granular trails, and tend to be preferentially located in the crystal cores, what allow them to be assumed primary or pseudosecondary in origin. Their shapes are mostly subrounded to irregular with dimensions that range from 2 to 15 μm and degree of filling fairly uniform (0.7 to 0.9). The daughter mineral in the three-phase inclusions is most likely halite, whose cube edges vary from 1 to 4 μm, seldom 7 μm. In the multi-phase inclusions, the other daughter mineral is probably hematite with size that does not exceed 3 μm. Figure 8 displays some morphological characteristics of the three- and two-phase inclusions trapped in a quartz and apatite crystals.

Fig. 8
figure 8

Photomicrographs and their respective sketches of fluid inclusions hosted by quartz (A–C) from the altered Serra Dourada granite and apatite (D) from a polyminerallic vein that crosscuts a metadacite. A, A’ Hydrothermal quartz crystal containing inclusions that trapped fluid 1 and are clustered in small groups or along intra-granular trails. The inter-granular trails contain secondary inclusions (fluid 2 or 3). B, B’ Three-phase primary fluid inclusions located at the right upper corner of the quartz crystal pictured in (A). C, C’ An intra-granular trail defined by three- and two-phase fluid inclusions, both primary and pseudosecondary, developed in the left lower corner of the quartz crystal shown in (A). D, D’ Trails of two-phase secondary (?) inclusions that follow basal cleavage planes of a crystal of apatite. Note an isolated primary three-phase fluid inclusion at the center of the figure

Full size image

Cooling revealed that the eutectic temperatures range from −68 to −44 °C, but most measurements (85 %) fall between −58 and −50 °C. This may indicate that NaCl, CaCl2, and MgCl2 are the main solutes, although KCl is also assumed to have been an important component especially in the solutions that accounted for the potassic alteration. Ice melting temperatures for the unsaturated inclusions record values from −21 to −3 °C.

During heating, the three-phase inclusions homogenized into the liquid state by disappearance of the solid. In a few inclusion trapped in quartz, however, the gas bubble was the last phase to disappear. The multi-phase inclusions did not homogenize even at temperature as high as 600 °C (limit of the Linkam thermometric apparatus). Most likely they represent heterogeneous trapping, although the nondissolution of the solids could also be due to kinetic factors or leaking. All two-phase inclusions homogenized into the liquid state.

The homogenization temperatures (Th) cover a wide range (160–478 °C), the highest values corresponding to three-phase inclusions hosted by scapolite and quartz crystals, followed by three-phase inclusions trapped in apatite (355 to 412 °C). The temperature gap (~350–300 °C) may reflect incomplete sampling or the limited number of the analyses done. Below approximately 300 °C most Th record come from two-phase inclusions, including those trapped in calcite crystals, which have Th as high as 290 °C. As will be discussed later, at this stage the mineralizing fluid would have undergone significant dilution. Except for scapolite, whose fluid inclusions did not yield Th less than 190 °C, the other host minerals kept forming at least until ~150 °C.

From a paragenetic point of view, considering the minimum temperatures of trapping, the first generations of scapolite, quartz and apatite formed apparently in this order. Calcite came next after the hydrothermal system having experienced a temperature drop of ~200 °C.

Fluid salinities were estimated for unsaturated and saturated inclusions using the equations proposed by Bodnar (1993) and Sterner et al. (1988), respectively. The calculated values show also a wide variation (56 to 7 wt. % equiv. NaCl). As a rule, the lower the Th, the lower is the salinity, indicating that the fluid may have mixed with colder and less saline solutions as the hydrothermal system evolved.

Geochronology

Pb evaporation of zircon single crystals

The sample selected for dating the Visconde granite was medium grained, isotropic, devoid of fractures, and located at point P2-A (Fig. 2). Eighteen zircon crystals were chosen for analysis. These crystals were colorless, euhedral to subhedral, and elongated, some containing dark inclusions and/or microfractures. The isotopic results are presented in Table 4. Eight crystals presented a Pb signal too low to allow 207Pb/206Pb ratios to be measured or 206Pb/204Pb also too low (206Pb/204Pb < 2500) to permit a reliable common lead correction. The remaining ten grains yielded ages between 2701 ± 4 and 2748 ± 3 Ma., with a mean value of 2732 ± 6 Ma (MSWD = 48; Fig. 9). Considering only the five crystals that provided older ages, a mean value of 2744 ± 5 Ma (MSWD = 4.9) is calculated (Fig. 9).

Table 4 Pb isotopic ratios of zircon crystals from the Visconde granite. (c: corrected from common lead)
Full size table
Fig. 9
figure 9

Single zircon Pb-evaporation age diagram for the Visconde granite showing the main Neoarchean and Paleoproterozoic episodes of granitogenesis in the CMP (deep pink and brown lines). The enclosed light yellow diagram displays the age spectrum using all data presented in Table 1. The vertical bars represent the errors for each zircon grain, and the horizontal green line corresponds to the weighted average crystallization age

Full size image

Pb–Pb in sulfides

As an attempt to determine the age of the mineralization and/or subsequent events that may have affected the ore bodies of the Visconde deposit, seven samples of chalcopyrite representing the main stage of mineralization were selected for Pb–Pb analysis. Three samples are from the sulfide-rich breccias and four from mono- or multi-mineral veins. The Pb isotopic results are shown in Table 5.

Table 5 Pb isotope ratios chalcopyrite samples from the Visconde deposit (2σ = analytical error)
Full size table

The analyses revealed the samples to be very radiogenic. For total digestion, the ranges of the 206Pb/204Pb, 207Pb/204Pb, and 208Pb/204Pb ratios are, respectively, 57.60–256.66, 24.90–61.92, and 56.17–270.98, while for leaching these ratios record values of 32.45 to 291.78, 18.21 to 66.00, and 46.26 to 87.81, respectively. On a 207Pb/204Pb vs.206Pb/204Pb diagram, the samples submitted to total digestion yielded a linear array which corresponds to a calculated age of 2729 ± 150 Ma with MSWD = 3115 (Fig. 10a). A similar value (2736 ± 100 Ma; MSWD = 6131) was obtained using the whole dataset including total digestion and leaching (Fig. 10b).

Fig. 10
figure 10

207Pb/204Pb vs. 206Pb/204Pb diagrams for chalcopyrite from the Visconde deposit. a Total digestion. b All data (digestion and leaching). Analytical error = 2σ

Full size image

Discussions

Evolution and sources of the hydrothermal fluids and metals

The types and stages of hydrothermal alteration at the Visconde deposit are similar to those commonly described for the IOCG deposit class (Hitzman et al. 1992; Williams et al. 2005) and reproduce fairly well what is observed in other Cu–Au deposits of Carajás, as Bacaba, Castanha, Alvo 118, Cristalino, and Sossego (Huhn et al. 1999; Villas et al. 2004; Monteiro et al. 2008; Moreto et al. 2011; Torresi et al. 2012).

Temperatures calculated using the oxygen isotope composition of different mineral pairs indicate that the Visconde hydrothermal system involved hot fluids (410 °C) at the earlier stages and cooler fluids (~240 °C) at the later stages. By the time chlorite precipitated overprinting the biotite-rich zones, the temperature was estimated at 285 °C based on the chlorite geothermometer (Craveiro 2011). In turn, the fluid inclusion data revealed a greater temperature range (>460 to ~160 °C) for the Visconde hydrothermal system, yet in a good agreement with the isotopic results. Concomitantly with the temperature drop, there was not only a decrease in fluid salinity (from 56 to 7 wt. % equiv. NaCl) but also a progressive 18 O depletion in the fluid (+9.4 to −1.3 ‰; Fig. 11).

Fig. 11
figure 11

Range of oxygen isotope composition of the fluids (black bars) that caused alteration and mineralization in the Visconde deposit. The vertical lines on the right represent the temperatures used in the calculations for a given set of minerals. The magmatic and meteoric water fields are from Taylor (1997). Ab albite, Act actinolite, Bt biotite, Chl chlorite, Ep epidote, Mag magnetite, Qz quartz, Tur tourmaline

Full size image

During the early sodic-calcic alteration, \( {\delta}^{18}{O}_{{\mathrm{H}}_2\mathrm{O}} \) values were higher (+3.3 to +9.4 ‰ at 410°–355 °C). Magnetite, in particular, precipitated from a fluid with a more restricted range of \( {\delta}^{18}{O}_{{\mathrm{H}}_2\mathrm{O}} \) (+5.6 to +7.4 ‰ at 410 °C). In the potassic stage, coincident with the development of local ductile shear zones, the \( {\delta}^{18}{O}_{{\mathrm{H}}_2\mathrm{O}} \) values were slightly lower than those of the previous stages (+4.8 to +7.2 ‰ at 355 °C). All these \( {\delta}^{18}{O}_{{\mathrm{H}}_2\mathrm{O}} \) ranges are consistent with fluids derived from magmatic sources (Taylor 1997), especially those responsible for the early alteration types.

The later stages (sodic-calcic and calcic-magnesian alterations), typically formed in veins outside the mineralized breccias, are related to fluids with temperatures of 230 °C and \( {\delta}^{18}{O}_{{\mathrm{H}}_2\mathrm{O}} \) values between −1.3 and +4.4 ‰, suggesting the involvement of surficial water. This hypothesis is corroborated by the fluid inclusion data that show lower salinity for the fluid that had circulated through the deposit rocks at the later stages. The salinity decreased from ~30 to 7 wt. % equiv. NaCl, indicating dilution, although at a constant temperature around 160–200 °C (Fig. 12).

Fig. 12
figure 12

Salinity vs homogenization temperature diagram depicting two important dilution trends for fluid 1, one with a high-temperature drop (460 to 160–200 °C) and the other without significant temperature change. The vertical bars represent the temperature ranges of formation of quartz (Qz), scapolite (Scp), apatite (Ap), and calcite (Cc). The halite saturation curve and critical curve are from Bodnar et al. (1989) and Knight and Bodnar (1989), respectively

Full size image

The \( {\delta}^{18}{O}_{{\mathrm{H}}_2\mathrm{O}} \) and \( \delta {D}_{{\mathrm{H}}_2\mathrm{O}} \) values for fluids related to the early sodic-calcic and potassic alteration zones plot in the metamorphic water field, close to the magmatic field. Some actinolite values are compatible with magmatic waters, whereas chlorite and epidote, corresponding to the later alteration zones, tend to approach the meteoric water line (Fig. 13).

Fig. 13
figure 13

Oxygen and hydrogen isotope composition of fluids responsible for the alteration and mineralization in the Visconde deposit. Next to each alteration type (in parentheses) are the temperatures used in the calculations. Metamorphic, magmatic, and formation water fields are from Sheppard (1986), Taylor (1992), and Taylor (1997), respectively

Full size image

A metamorphic signature is plausible because a regional metamorphic event lasting from 2.76 to 2.73 Ga (Machado et al. 1991; Pinheiro and Holdsworth 2000) is recorded in both the Transition and Carajás domains. This event is related to the inversion of the Carajás basin where the metamorphic conditions range from the lower greenschist to upper amphibolite facies. According to Taylor (1974, 1997), dehydration water is the dominant constituent of metamorphic pore fluids only at high grades, whereas formation waters prevail at lower grades. Some samples from the Visconde deposit plot where the metamorphic and formation water fields overlap (Fig. 13). Formation water could have migrated toward the Visconde deposit area from the nearby Carajás basin, particularly from rocks submitted to less- intense regional metamorphism. Along its pathway, this water interacted with the deposit metamorphic host rocks and its composition was progressively modified. As a result, it acquired a metamorphic isotopic character at the same time that it mixed with pre-existing metamorphic fluids.

Given that the dehydration water generally presents low salinity, the involvement of formation water is also an attractive hypothesis to explain the high salinities (up to 56 wt. % equiv. NaCl) recorded in the Visconde hydrothermal fluids. Waters trapped within the sediment interstices tend to have their salinity increased and to undergo compositional changes during the residence time (Taylor 1997), becoming a potential reservoir for metals and ligands. That hypothesis gains further support from B isotopic data and Cl/Br and Na/Cl ratios, which have indicated the interaction of fluids with evaporite beds in other Carajás IOCG deposits (Xavier et al. 2008, 2010).

A few actinolite and biotite samples from early alteration stage yielded \( \delta {D}_{{\mathrm{H}}_2\mathrm{O}} \) values characteristic of water dissolved in felsic magmas that did not undergo boiling (Taylor 1992). These values may indicate the involvement of magmatic fluids in the Visconde hydrothermal system, especially in the early stages, despite the overlapping with metamorphic fluids. Several authors suggested a magmatic origin for the IOCG mineralizing fluids (e.g., Pollard 2001; Groves et al. 2010). For the Visconde deposit, this is a permissive hypothesis because there is obvious spatial relationship between ore and intrusive bodies of the Planalto suite granites and gabbros/diorites. Mixing of metamorphic with magmatic fluids would then be a reasonable interpretation and, in this case, those intrusive rocks could be most likely magmatic sources.

Assuming the hypothesis of mixing, the fluid composition changed from the sodic-calcic (actinolite) to the potassic (biotite) stage, moving away from the magmatic field due to the greater contribution of metamorphic water to the solutions that precipitated that mica. This mixing resulted in a steep decrease in both temperature (460 to 230 °C) and salinity (58 to 35 wt. % equiv. NaCl) of the fluid responsible for these types of hydrothermal alteration (Fig. 11). From biotitization to chloritization, the \( {\delta}^{18}{O}_{{\mathrm{H}}_2\mathrm{O}} \) and \( \delta {D}_{{\mathrm{H}}_2\mathrm{O}} \) values varied significantly with the progressive depletion in 18 O and D, indicating now mixing with meteoric waters (Fig. 10). In turn, most epidote samples present \( \delta {D}_{{\mathrm{H}}_2\mathrm{O}} \) values fairly distinct and tend to approach that of seawater (Fig. 13). A similar trend has been found in the neighboring Sossego deposit (Monteiro et al. 2008). However, Dilles et al. (1992) claim that the epidote \( \delta {D}_{{\mathrm{H}}_2\mathrm{O}} \) data should be interpreted with caution, since this mineral is extremely susceptible to D fractionation. The later alteration stages seem to have been dominated by surficial water, whose influx was largely favored by the brittle regime that then prevailed. Because the mineralization was contemporaneous with the later hydrothermal stages, dilution and decreasing temperature were the main factors causing abundant sulfide precipitation.

All δ 34 S values of the Visconde chalcopyrite, but a single datum are positive and deviate very slightly from the composition of magmatic sulfur. The narrow δ 34 S range (−1.2 ≤ δ 34 S ≤ +3.4 ‰) allows inferring that sulfur did not undergo any significant degree of isotopic fractionation while being transported by the hydrothermal fluid. Most likely, the fluid was reducing otherwise a larger fractionation and a greater spread of δ 34 S values would be expected (Rye and Ohmoto 1974). The abundant precipitation of magnetite prior to the main sulfide mineralization certainly consumed a large amount of oxygen causing a decrease of the oxidation potential. The oxygen fugacity of the fluid dropped to levels that hindered the co-precipitation of the oxide and sulfide phases during the main ore formation. Under these reducing conditions, the range of δ 34S values of the sulfide may be assumed to represent closely that of the fluid from which it precipitated (Ohmoto and Goldhaber 1997). After the main mineralization event, the O2 fugacity increased as supported by the local occurrence of late gypsum and hematite-bearing veins.

Sulfide precipitation occurred during all alteration stages, but it was more significant after the intense biotitization of the rocks and was contemporaneous with the calcic-magnesian alteration. Thus, fluids transported S species from the very beginning of the Visconde’s hydrothermal history. The Visconde sulfur isotope composition is partly consistent with the mantle source (δ 34 S = 0 ± 1 ‰; Eldridge et al. 1991) and compatible with the interval commonly attributed to sulfur from magmatic reservoirs (δ 34 S = 0 ± 5 ‰; Ohmoto and Goldhaber 1997). Those species most likely were leached from magmatic sulfides present in the metavolcanic rocks from the Carajás basin or ultramafic rocks and granitic intrusions from the Transition domain, or both. On the other hand, the less oxidizing conditions of the Archean atmosphere may not have precluded the extraction of some sulfur from evaporitic sequences. The δ 34SSO42− values in these reservoirs vary from ≈2 to 5 ‰ between ≈3.5 and ≈2.7 Ga (Ohmoto and Goldhaber 1997), being compatible with the S isotope composition of the Visconde deposit.

Compared with other Cu–Au Carajás deposits (Fig. 14), the Visconde sulfide δ 34 S VCDT values are similar to those from the Cristalino, Pista, Bacaba, Castanha, Jatobá, and Salobo deposits (Fig. 13). In general, heavier S isotopic compositions have been found in the Sossego ore bodies (Sequeirinho, Sossego, Baiano, and Curral) and Alvo 118, for which the incorporation of 34 S from surficial sources during the sulfide precipitation is a viable hypothesis (Torresi et al. 2012).

Ages of the Visconde granite and mineralizing event

The set of Pb–Pb data obtained on zircon from the Visconde granite does not allow determining an accurate and reliable age for its emplacement. The high MSWD value (48) of the age (2732 ± 6 Ma) obtained on the whole population of analyzed zircons reveals a complex U–Pb behavior and may reflect the effects of postcrystallization radiogenic lead loss by metamictization. The existence of slightly older inherited cores cannot be ruled out and may also account for the scattering of the results.

Even poorly constrained, the Pb–Pb results allow discarding a Paleoproterozoic age for this granite. The 2748–2704 Ma range obtained for the zircons of the Visconde granite is similar to those of other granites emplaced into the rocks of both the Transition and Carajás domains, such as the Serra do Rabo intrusion and the Planalto suite (Fig. 9). Pb–Pb zircon dating of several intrusions that have been correlated to the Planalto Suite yielded ages from 2747 ± 2 Ma (Huhn et al. 1999) to ~2730 Ma (Sardinha et al. 2004; Feio et al. 2012). A similar U–Pb zircon age of 2743 ± 1.6 Ma was also determined for the Serra do Rabo granite (Sardinha et al. 2006). Although mapped as a different stratigraphic unit, this intrusion mineralogically and texturally resembles the Planalto suite stocks. Recently, younger ages (2729 ± 17, 2710 ± 10, and 2706 ± 5 Ma) have been obtained for zircon grains from Planalto granites using the U–Pb LA-ICP-MS technique (Feio et al. 2012). The fact that the zircon ages of the Visconde granite lie within the range of ages attributed to the crystallization of the Planalto suite suggests that it belongs to this suite, in agreement with what was previously proposed on the basis of petrographic features (Feio et al. 2012; Craveiro et al. 2012).

Pb–Pb dating of the Visconde ore yielded ages of 2.73 ± 0.15 and 2.74 ± 0.10 Ga (Fig. 10a, b), which, despite the large errors, are coherent with the geology of the area and similar to the Archean age of the main Carajás Cu–Au mineralization. Similar ages have been obtained for the Cristalino deposit (2716 ± 36 Ma, Pb–Pb in pyrite, Soares et al. 2001) and for the Sequeirinho ore body of the Sossego mine (2712 ± 4.7 Ma, U–Pb in monazite, Moreto et al. 2013). These ages indicate a Neoarchean mineralization event, but they should be evaluated in the light of other geological events recorded in the Transition domain.

Fig. 14
figure 14

Sulfur isotope composition of sulfides from the Visconde deposit compared with other Cu–Au Carajás deposits. 1, Monteiro et al. (2008); 2, Monteiro et al. (2007); 3, Ribeiro et al. (2009); 4, Torresi et al. (2011); 5, Requia and Fontboté (2001); 6, Tazava and Oliveira (2000); 7, Villas et al. (2001); 8, Dreher et al. (2008). Magmatic field and Archean sulfate (crosshatched) are from Ohmoto and Goldhaber (1997); mantle signature is from Eldridge et al. (1991)

Full size image

In IOCG provinces, it is of particular interest to assess the genetic relationship between mineralization and magmatic activity (e.g., Candelaria in Chile, Olympic Dam in Australia; Mathur et al. 2002; Bastrakov et al. 2007). For the Carajás Cu–Au deposits, available geochronological data, based on Pb–Pb, U–Pb, and Re–Os systematics, have indicated three main mineralization ages, coincident with regional magmatic events that occurred at 2.76–2.70, ca.2.57, and 1.88 Ga. The first refers to the formation of the Estrela Granitic complex, Planalto suite, Plaquê suite, and correlated granites (Soares et al. 2001; Galarza et al. 2008); the second event is represented by the Old Salobo and Itacaiunas granitic stocks (Requia et al. 2003; Tallarico et al. 2005); and the third is represented by the Paleoproterozoic A-type granitic plutonism (Tallarico 2003; Neves et al. 2006; Moreto et al. 2013).

The time interval in which the Planalto suite formed not only includes the age of the mineralization obtained here but also marks a tectonically active period in the Carajás Province. Similarly to the Sossego deposit (Villas et al. 2005; Monteiro et al. 2008), the altered zones in the Visconde deposit display textures indicative that the hydrothermal system developed under a ductile to brittle structural regime. Hence, it is possible that both the emplacement of the Planalto granitic rocks and the Visconde hydrothermal alteration took place during the same tectonic episode. The occurrence of the 1.88 Ga Rio Branco granite (Santos et al. 2013a, b) at approximately 15 km from the Visconde deposit does not a priori rule out a genetic link with the Paleoproterozoic A-type granites. However, in contrast to what was observed for the Sossego and Alvo 118 deposits (Tallarico 2003; Neves et al. 2006; Moreto et al. 2013), no isotopic evidence has been detected indicating that the Visconde Pb–Pb system was affected at that time, and the influence of a Paleoproterozoic event is unlikely.

Conclusions

The Visconde deposit is hosted mainly by metadacites, the Serra Dourada granite, and gabbros/diorites of Meso- and Neoarchean age. The more conspicuous ore bodies occur as breccias supported by a matrix dominantly composed of chalcopyrite. The mineral associations give a Cu–Au-Fe–Ni–ETRL–B–P signature to the ore.

In the outer portions of the ore bodies, hydrothermal alteration zones are well developed. The earlier and more pervasive sodic and sodic-calcic types were controlled by regional shear zones and almost completely changed the primary mineralogy of the rocks. Isotope data of silicate minerals from these alteration zones reveal hot (410 to 355 °C) and 18 O-rich (\( {\delta}^{18}{O}_{{\mathrm{H}}_2\mathrm{O}} \) = +4.2 to +9.4 ‰) fluids dominantly formed by metamorphic waters, most likely derived from the Carajás basin rocks. Fluid inclusion data indicate that the hydrothermal fluids responsible for these alteration zones were very saline (up to 58 wt. % equiv. NaCl) and may have reached temperatures above 460 °C. Some contribution from magmatic sources is permissive, in which case the Planalto suite is assumed to be the most viable source.

Considering that the Itacaiúnas Supergroup units do not host the Visconde deposit, large-scale structures would have enabled migration of the fluids from the basin to the basement. Fluids with similar oxygen isotope composition (\( {\delta}^{18}{O}_{{\mathrm{H}}_2\mathrm{O}} \) = +4.8 to +7.2 ‰ at 356 °C) promoted severe biotitization of the rocks (potassic alteration) that was controlled by local shear zones, as evidenced by the development of mylonitic foliation at different scales. The fluids experienced not only a significant temperature drop, but also a dilution process during which fluid salinity decreased to 30 wt. % equiv. NaCl approximately.

Major precipitation of chalcopyrite (±bornite ± chalcocite ± digenite) occurred later during the brittle regime to which the rocks had been subjected. These sulfides occur mainly in veins and breccias associated with calcic-magnesian minerals, whose isotopic oxygen signature suggests the involvement of connate and meteoric fluids (\( {\delta}^{18}{O}_{{\mathrm{H}}_2\mathrm{O}} \) = −1.3 to +3.7 ‰) at lower temperatures (230 ± 11 °C). As a result, the fluids became progressively less saline (from 30 to 7 wt. % equiv. NaCl) and colder, with temperatures as low as ~160 °C. Thus, dilution and temperature decrease were most likely the main mechanisms that triggered the precipitation of large amounts of sulfides. The sulfur isotope composition (δ 34 S VCDT = −1.2 to 3.4 ‰) indicates relatively reducing conditions for the fluid, which had much of its oxygen consumed by the earlier precipitation of abundant magnetite. This signature is compatible with a magmatic source for sulfur, which could have been either exsolved during the crystallization of a granitic magma or leached from sulfides originally present in pre-existing igneous rocks. Basement Mesoarchean ultramafic, felsic volcanic and granitic rocks could have also supplied components to the hydrothermal system, as the Pb data suggest.

Pb–Pb ages for the Cu–Au mineralization are imprecise (2.73 ± 0.15 Ga and 2.74 ± 0.10 Ga). Nevertheless, they indicate a Neoarchean age for the formation of the Visconde main ore bodies. Despite the large errors, these ages are consistent with the geology of the area and with the 2.75 Ga IOCG mineralizing event that has been recorded in Carajás.

Textural evidence shows that the hydrothermal system was conditioned by a structural regime that evolved from ductile to brittle, which may be indicative of its contemporaneity with the transpressive tectono-thermal event that occurred in the Transition domain during the Neoarchean. Although the geochronological data do not define more precisely the age of the Visconde mineralization, it is very unlikely that it was temporally related to the 2.5 or 1.88 Ga magmatic events, related the formation of other Carajás Cu–Au deposits (Tallarico 2003; Requia et al. 2003; Tallarico et al. 2005; Neves et al. 2006; Moreto et al. 2013).

The types and sequence of the alteration stages, structural controls, association with regional magmatism, style of mineralization, and Fe–Cu–Au metallic signature allow including the Visconde deposit in the IOCG class, similarly to other Cu–Au deposits of the Transition domain.