Introduction

The genetic link between magmatism and porphyry-copper-style mineralization in active continental margins is well documented (Titley and Beane 1981; Hedenquist and Lowenstern 1994). Intrusion-related hydrothermal systems get their thermal energy and variable amounts of volatiles, metals and other components largely from subduction-related magmas emplaced at shallow levels of the Earth's crust (Cathles 1981; Sawkins 1990).

The Andes of central Chile host some of the world's largest porphyry copper deposits, such as El Teniente, Río Blanco-Los Bronces and Los Pelambres. These deposits were formed during the Miocene-Pliocene when both shallowing of the subduction angle and crustal thickening occurred (Skewes and Stern 1996). Skewes and Stern (1994) suggested that exsolution of copper-bearing magmatic fluids responsible for brecciation, alteration and mineralization at this time was produced by a rapid decrease of lithostatic pressure.

In this paper we examine the geochemical composition and the magma sources of the porphyry intrusions associated with the mineralization at the Los Pelambres Cu deposit (31°43′S, 70°29′W), located in the southern part of the Andean flat-slab segment, 190 km north of Santiago (Fig. 1). Previous work on the Los Pelambres deposit reported data on the age of the mineralization (e.g., Mathur et al. 2001), the nature and extent of the associated hydrothermal alteration/mineralization (Sillitoe 1973; Atkinson et al. 1996), and isotopic signatures of the fluids related to brecciation events (Skewes and Stern 1996). However, there is no information about the origin of the magmas that generated this porphyry system. For this reason we constrain geochemically the nature of its source, and propose a petrogenetic model consistent with the tectonic setting at the time of its formation. In order to put the Los Pelambres rocks in a more regional context, a geochemical comparison with barren magmatic units of similar age (Aconcagua volcanic rocks and La Gloria pluton) and located in the same Late Miocene belt, is presented.

Fig. 1.
figure 1

Map showing the main tectonic features of the southeastern Nazca plate and the Chilean flat-slab segment (28–33°S). The position of the predicted path of the Juan Fernández hotspot chain at 10 Ma (bold black line), and the volcanic gap separating the Central Volcanic Zone (CVZ) and the Southern Volcanic Zone (SVZ) are indicated. The locations of the Los Pelambres porphyry copper deposit, La Gloria pluton and the Aconcagua andesites are also shown. The Wadati-Benioff zone contours were taken from Isacks (1988)

Full size image

Tectonic setting

The Los Pelambres deposit is located in the "flat-slab" segment (28–33°S) of the Chilean Andes, where Recent volcanism is absent (Barazangi and Isacks 1976; Fig. 1). The southernmost end of the flat-slab segment changes gradationally into the "normal-slab" segment of the southern Chilean Andes (33–46°S), characterized by a subduction angle of about 30°, where a continuous belt of modern active volcanoes occurs.

Tectonic and magmatic studies (Kay et al. 1987, 1991; Allmendinger et al. 1990; Reynolds et al. 1990; Kay and Abruzzi 1996; Kay and Mpodozis 2002) concerning the Miocene to Recent evolution of the present-day "flat-slab" segment of the Chilean Andes have shown that the subducted oceanic slab shallowed since ~26 Ma. During the Early Miocene (~27–20 Ma) this segment had a subducted slab geometry similar to that currently observed in the normal-slab segment at 35°S, and a crustal thickness of 35–40 km (Kay et al. 1991; Kay and Abruzzi 1996; Kay and Mpodozis 2002). The shallowing of the subduction zone progressed from Middle to Late Miocene (20–5 Ma), accompanied by crustal thickening, a substantial decrease in the astenospheric wedge thickness, eastward migration and broadening of the arc, and cessation of the magmatic activity over the flat-slab (~5 Ma). During this period, the crustal thickness along the flat-slab segment was different from north to south. The northern (~28–29°S) and central (~30–31°S) regions reached a crustal thickness exceeding 55 km by the end of the Miocene, while the southern region of the flat-slab (31–33°S) reveals a thinner crust (~30–35 km) for the same period (Kay et al. 1991). Moreover, the crust below the Middle to Late Miocene magmatic belt south of 33°S never thickened to as much as 40 km. South of 33°S the magmatic arc migrated eastward during the Pliocene to its current position in the high Andes, where crustal thickness of more than 50 km occurs (Stern and Skewes 1995).

These changes in the subduction geometry between 28–33°S have been related to the collision and subduction of the Juan Fernández Ridge (JFR) hotspot chain (Pilger 1981; Nur and Ben-Avraham 1981, Yáñez et al. 2001, 2002). The cited authors modeled the flat-slab geometry as an effect of the subduction of a buoyant hotspot chain. In this scenario, the contribution of the oceanic lithosphere to the arc magma source would have increased with the decreasing volume of mantle wedge resulting from progressive slab flattening.

Geology of the Los Pelambres porphyry intrusions

The Los Pelambres porphyry copper mineralization is hosted by an intrusive complex emplaced within a sequence of andesitic rocks of the Los Pelambres Formation (Late Cretaceous) (Rivano and Sepúlveda 1991; Atkinson et al. 1996; Fig. 2). The intrusive complex consists of a main tonalite stock and porphyritic bodies, and a small number of post-mineralization andesite and aplite dikes. Magmatic/hydrothermal breccia pipes also occur within the deposit. Detailed petrographic descriptions and characterization of these bodies were presented by Atkinson et al. (1996).

Fig. 2.
figure 2

Geological map of the Los Pelambres porphyry copper deposit, modified from Schultz (1997)

Full size image

Tonalites and tonalite porphyries

The tonalites form a stock which hosts the main mineralization. They have a medium grain size (0.5–3 mm), composed of a subequigranular hypidiomorphic intergrowth of plagioclase (normally zoned, An30–40), biotitized hornblende and biotite, with minor quartz and perthitic K-feldspar as interstitial grains. Zircon and apatite are common accessory phases, and sulfides (chalcopyrite, bornite and pyrite) occur as disseminated grains within altered intrusions. Irregular tonalite porphyry bodies are recognized throughout the stock. They have plagioclase phenocrysts, disseminated biotite and biotitized hornblende. The phenocrysts (~1 cm) are surrounded by a fine-grained quartz/K-feldspar matrix (0.05–0.15 mm).

Quartz diorite and quartz monzodiorite porphyries

These porphyries occur as dikes and irregular bodies, and are distributed throughout the central portion of the tonalite stock. Two types of dioritic porphyries are recognized: quartz diorite (Porphyry A) and quartz monzodioritic (Porphyry B) porphyries. Quartz diorite porphyry consists of plagioclase and biotitized hornblende (2–5 mm) phenocrysts in a biotite trachytic groundmass. Quartz monzodiorite porphyry is composed of plagioclase, biotite and biotitized hornblende immersed in an biotite-bearing aplitic matrix. All these porphyritic rocks are mineralized.

Post-mineralization dikes

Post-mineralization magmatic events are represented by a late andesite dike (plagioclase phenocrysts in an aphanitic matrix), and aplitic dikes (fine grained quartz and K-feldspar).

Mineralization

Hypogene and supergene alteration and mineralization at Los Pelambres have been described in detail by Atkinson et al. (1996), who recognized multiple events of ore deposition during the life span of the hydrothermal system. Early hypogene mineralization occurs as disseminated sulfides (chalcopyrite/bornite, traces of molybdenite and pyrite) in alteration halos around veins of a quartz stockwork, associated with potassic alteration. Late mineralization is defined by pyrite veins with sericitic halos. K-Ar dating on hydrothermal biotite (potassic alteration) of the tonalite stock yielded ages of 9.74±0.16 and 9.96± 0.18 Ma (Sillitoe 1973). An average K-Ar secondary biotite age of 9.9±1.0 Ma was obtained from the tonalite and some porphyry facies (Atkinson et al. 1996). Recent Re-Os dating in early and late molybdenite yielded mineralization ages of 10.75±0.05 Ma and 10.40±0.05 Ma, respectively (Mathur et al. 2001).

Geochemistry of the Los Pelambres intrusions

Major and trace elements

Major and trace element compositions of selected Los Pelambres intrusive rocks (tonalite, tonalite porphyry, and quartz monzodiorite porphyry) were determined by a combination of XRF, AAS, INAA, ICP-MS and DCP-AES techniques at Bondar Clegg Laboratories, Canada. Representative analyses suitable for petrological work (LOI<1.5 wt%) are given in Table 1.

Table 1. Representative major and trace element abundances in Los Pelambres rocks. T, Tonalite; T-P, tonalite porphyry; QDM-P, quartz-diorite porphyry
Full size table

The Los Pelambres rocks (tonalite, tonalite porphyry and quartz monzodiorite porphyry) form a suite covering a SiO2 range from 62–72 wt% (average SiO2 of 63.3 wt%), and are characterized by high Al2O3 (15.2–17.8 wt%) and Na2O (4.8–6.6 wt%) abundances, with K2O/Na2O<1 (0.23–0.71) and an Al2O3/(CaO+ Na2O+ K2O) molar ratio of about 1.0. They have low MgO between 1.2–2.7 wt%, and the magnesian numbers (#Mg = 100*MgO/[MgO+Fe2O3], molar) are moderately high, ranging from 38 up to 75. Major element variation diagrams show a negative correlation for Al2O3, CaO, Fe2O3, MgO, TiO2 and P2O5, while Na2O and K2O indicate incompatible behavior for the whole silica range (Fig. 3).

Fig. 3.
figure 3

Major elements variation diagram of samples from the Los Pelambres intrusions. The compositional fields of the coeval La Gloria pluton and Aconcagua volcanic rocks according to data in Cornejo (1990) and Kay et al. (1991)

Full size image

Trace element abundances of the Los Pelambres rocks show high Sr (306–750 ppm) and low Y (1.59–6.55 ppm), with high Sr/Y ratios (~100–300) (Fig. 4a). Cr (1–17 ppm), Ni (2–12 ppm) and Nb (1–10 ppm) contents are low (except for quartz monzodiorite porphyry sample LP-44, where Cr=46 ppm). Chondrite-normalized rare earth element (REE) patterns of the Los Pelambres rocks are strongly fractionated, with light rare earth element (LREE) enrichment and heavy REE (HREE) depletion (low Yb: 0.143–0.592 ppm; high LaN/YbN: ~25–60; Fig. 4). REE patterns display a steep negative slope with an inflection at Tb, and no Eu anomalies (Eu/Eu* ~1) are observed (Fig. 5).

Fig. 4.
figure 4

Sr/Y versus Y, and LaN/YbN versus YbN discrimination diagrams for the Los Pelambres intrusions, La Gloria pluton and Aconcagua volcanic rocks. The adakitic and ADR (Andesite-Dacite-Rhyolite) fields were taken from Martin (1999)

Full size image
Fig. 5.
figure 5

. Chondrite-normalized REE diagram of Los Pelambres rocks. Fields of La Gloria and Aconcagua samples are also shown. Normalizing factors after Nakamura (1974)

Full size image

Sr-Nd isotopic compositions

Sr and Nd isotopic compositions were determined for three samples of the Los Pelambres tonalite stock (F-51, LP-48, LP-75; Table 2). The Sr-Nd analytical determinations were performed at the Centro de Instrumentación Científica, Universidad de Granada, Spain, using a Finnegan MAT 262 thermal ionization mass spectrometer (TIMS) with variable multicollector and RPQ. 87Sr/86Sr was normalized using 88Sr/86Sr=8.375209, with a 0.0007% (2σ) reproducibility under successive determinations of the NBS-987 dissolved standard. The 143Nd/144Nd ratios, normalized using 146Nd/144Nd=0.7219 showed precision better than 0.0016% (2σ) calculated from successive measurements of the WSE powder standard. Initial ratios were calculated assuming a Late Miocene age of 10 Ma. The Sr and Nd isotopic isotopic compositions of the Late Miocene Los Pelambres intrusives reveal initial 87Sr/86Sr ratios between 0.70439 and 0.70465, while 143Nd/144Nd ratios range from 0.512619 to 0.512635. In terms of εNd, the Los Pelambres rocks exhibit values close to zero.

Table 2. Sr-Nd isotope data for the Los Pelambres porphyries, La Gloria pluton and Cerro Aconcagua andesites
Full size table

Comparison with contemporaneous barren igneous rocks

The geochemical composition of the Los Pelambres rocks is contrasted with the contemporaneous Late Miocene La Gloria pluton (9.8 Ma, cf. Cornejo 1990) and the Cerro Aconcagua andesites (10.3–9 Ma; cf. Kay and Mpodozis 2002), selected as representative magmatic units located in the southern portion of the current flat-slab Andean segment. The La Gloria pluton is an equigranular granodiorite to quartz-monzodiorite epizonal barren intrusion, extending over 100 km2, that is located 40 km east of Santiago City (ca. 33° 30′S; Cornejo and Mahood 1997). The Late Miocene Cerro Aconcagua magmatic center is located on the Chile-Argentina border (32°40′S, 70°W), 120 km northeast of Santiago. Chemical data from the La Gloria pluton are mostly taken from Cornejo (1990). Additional chemical and Sr-Nd isotopic compositions were obtained for sample G-318 of this pluton (Table 2). Chemical and isotopic data from the Cerro Aconcagua andesitic lava flows and breccias (upper level) were taken from Kay et al. (1991).

When compared with the barren La Gloria and Aconcagua magmatic rocks, fresh samples from the Los Pelambres rocks display distinct chemical differences, mainly noticeable in their trace element signature. They are Al2O3 and Na2O-enriched, together with higher Sr/Y and LaN/YbN ratios, and the strongly fractionated, HREE-depleted chondrite-normalized patterns contrast with the rather flat, less fractionated REE patterns from La Gloria and Aconcagua (Figs. 3, 4, 5). The Los Pelambres rocks have Sr initial ratios similar to those reported for the La Gloria pluton and the Aconcagua andesites (Table 2). On the other hand, the εNd values of the Los Pelambres rocks are similar to that of the Aconcagua andesites, but lower than that obtained on the La Gloria pluton.

Fig. 6.
figure 6

Ab–An–Or normative diagram (Baker 1979) showing the fields of Los Pelambres and La Gloria rocks. Los Pelambres samples show the typical trodhjemitic character recognized in adakitic rocks

Full size image

The adakite-like signature of the Los Pelambres rocks: an oddity in the Late Miocene magmatism of central Chile

Adakitic affinity

The term adakite was introduced by Kay (1978) and has been used to describe high-Al and Na-rich andesitic to dacitic, extrusive or intrusive rocks with a high Sr content (>600 ppm), strongly fractionated REE patterns (HREE depleted, LREE enriched), among other features, which were interpreted as resulting from slab melting where garnet and hornblende are residual phases. With the exception of the low MgO contents, the Los Pelambres rocks display an adakitic major and trace element geochemical affinity, following the criteria defined by Defant and Drummond (1990), Drummond and Defant (1990), Drummond et al. (1996) and Martin (1999). One of the most relevant chemical features of this particular type of magmatism can be seen in the (Sr/Y) vs. Y and (LaN/YbN) vs YbN discrimination diagrams (Fig. 4), where the Los Pelambres intrusive rocks plot well within the adakitic field. The Na-rich rocks of Los Pelambres show the typical trondhjemitic character recognized in adakitic rock suites, when plotted on an Ab–An–Or normative diagram (Barker 1979) (Fig. 6). Similary, the low Nb/Ta (3.0–8.0) and high Zr/Sm (32–60) ratios are comparable with those recorded in TTG gneisses and modern adakites (Foley et al. 2002). On the other hand, the Sr–Nd isotopic composition of the Los Pelambres porphyries is more radiogenic than most typical adakites, whose Sr-Nd values are close to MORB (144Nd/143Nd>0.5129 and 87Sr/86Sr <0.705; Martin 1999). These isotopic differences to true adakites are attributable to participation of crustal components in flat-slab subduction magmatism (see below).

Under subduction settings constrained by particular P–T–H2O conditions (P≥5 kbar, T≥750 °C, >10 wt% H2O), young (≤25 Ma), mafic oceanic lithosphere melts before reaching dehydration, generating adakitic magmas with a MORB-like isotopic signature, instead of typical calc-alkaline arc andesite-dacite-rhyolite suites, originating by partial melting of a metasomatized mantle wedge (Drummond et al. 1996; Martin 1999; Prouteau et al. 1999). However, adakite-type melts may also result from partial melting of overthickened mafic lower crust equilibrated with a garnet-hornblende residual mineralogy (e.g., Kay et al. 1987, 1991; Petford and Atherton 1996; Kay and Mpodozis 2001).

The source of the Los Pelambres intrusions

The identification of the magma source of porphyry copper systems has been the subject of long-standing controversy. Particulary interesting is the debate generated after Oyarzún et al.'s (2001) model for the formation of the adakite-like Late Eocene-Early Oligocene porphyry copper systems of northern Chile by melting of the oceanic slab (Richards 2002; Rabbia et al. 2002). Because the adakitic signatures are not exclusively derived by slab melting and have also been explained by crustal participation either as a source contaminant or as a protholith after crustal thickening (e.g., Petford and Atherton 1996; Kay and Mpodozis 2002), a combination of geochemical and geodynamic evidence is needed to better understand its origin. In the following discussion we present such evidence to constrain the source of the Los Pelambres rocks.

With the exception of adakites recognized along the Austral Volcanic Zone (49–54°S; Stern et al. 1984; Futa and Stern 1988; Kay et al. 1993; Stern and Kilian 1996), where young, hot and buoyant oceanic lithosphere (<24 Ma) is subducted under a relatively thin crust (<35 km), no modern adakites have been documented in the Chilean Andes. A recent thermal model given by Gutscher et al. (2000) has shown that slab melting before dehydration is viable under flat-slab conditions similar to those recognized further north along the Andes. The authors tested the model in the central Andes flat-slab segment to explain adakite-type rocks erupted between 10 and 4 Ma, but this was questioned by Kay and Mpodozis (2002), based on geological and geochemical arguments. They argued that the Eocene age of the Nazca plate that is subducting beneath the central Chile Andes (Yáñez et al. 2001) is too old to favor slab melting, and indicated that these rocks are better explained by contamination of arc-derived magmas by deep garnet granulite and eclogite rocks of a thickened crust or by crust tectonically incorporated (subduction erosion) into the mantle source. The explanation based on a thickened crust is unlikely for the case of the studied adakite-like rocks because the crustal thickness for the Los Pelambres region may not have reached more than 35 km at the time of the ore deposit formation (cf. Kay et al. 1991). In addition, melting of an overthickened mafic lower crust implies major crustal processes, which might have a regional expression. On the contrary, the adakite-like signature of Los Pelambres rocks is an oddity along the Late Miocene belt of Central Chile. For example, the contemporaneous La Ramada (c. 32°S; see Kay and Mpodozis 2002 for geochemical characteristics) and Aconcagua volcanic centers, and the La Gloria pluton have a typical arc geochemical composition with moderate La/Yb ratios and differ largely from the adakitic signature of the Los Pelambres intrusions.

The Late Miocene adakite-like rocks from Los Pelambres are closely related in time and space with changing subduction conditions. In fact, tectonic reconstructions by Yáñez et al. (2001, 2002) revealed that the JFR hotspot chain migrated from north to south along the Chilean margin since the Early Miocene to its actual position in central Chile. These reconstructions show that the oblique interaction between a NE segment of the ridge and the trench migrated rapidly southward along the Chilean margin since the Early Miocene in northern Chile (19°S) to the Late Miocene (~ 10 Ma) at the latitude of Los Pelambres, where a W-E segment of the JFR hotspot track started to subduct (Fig. 1). This caused a dramatic slowdown of the southward migration of the ridge-trench interaction and coincided with the termination of the frontal arc andesitic volcanism (Kay and Mpodozis 2002). In this scenario we favor an alternative explanation based on the particular case of partial melting of rocks of the W-E segment of the JFR hotspot chain, when it started to migrate slowly southward, allowing to build up a sufficient volume of magma to ascend. These locally restricted tectonic conditions may explain why adakite-like rocks were not found in contemporaneous igneous rocks formed in the same belt. During the formation of the Los Pelambres deposit, the age of the subjacent hotspot rocks was about 10 Ma older (Yáñez et al. 2001), thus it is likely that the hotspot source rocks were at a sufficient temperature to reach partial melting before dehydration was completed. The latter assumption is based on the low Nb/Ta and high Zr/Sm ratios, which are considered to result from melting of an amphibolite or garnet-amphibolite source (Foley et al. 2002). The isotopic differences observed between the Los Pelambres rocks and the typical depleted adakites can be explained by the incorporation of more radiogenic crustal material tectonically dragged by the subducting ridge to the Los Pelambres magma source depth (Kay and Mpodozis 2002) and/or the inheritance of a more radiogenic OIB signature of the source. It is noteworthy that the tectonic erosion mechanism would have been operative in Central Chile since 15 Ma (Stern 1991). The decreasing thickness of the mantle wedge by combination of both slab shallowing and incorporation of tectonically eroded crustal material to the source region, could explain the lower MgO contents of the Los Pelambres rocks compared to true adakites. In fact, the high MgO contents of adakites have been interpreted as due to slab melt—mantle wedge interaction (Kay 1978; Kay et al. 1993; Sen and Dunn 1994).

Is the subducting lithosphere the source of metals?

There has been growing interest in adakitic magmatism and its relation to copper and gold mineralization during the last decade. An association between adakites and ore deposits has been documented, for example, at Mount Pinatubo in Luzon, Philippines, where small Pliocene to Quaternary porphyry copper-gold deposits occur in the vicinity of the summit of the volcano (Sillitoe and Gappe 1984; Malihan 1987; Imai et al. 1993). In addition, adakites in east Mindanao, Philippines, are associated with Plio-Pleistocene copper and gold porphyry/epithermal systems (Maury et al. 1996; Sajona and Maury 1998). Pasteris (1996) proposed that "open to sulfur" systems (e.g., with huge amounts of sulfur released to the atmosphere), such as the adakitic Mount Pinatubo volcano, represent aborted or failed porphyry copper deposits.

In the Chilean Andes, which host the largest concentration of world-class copper deposits on Earth, adakite-like rocks have been recognized associated with Oligocene porphyry copper deposits, suggesting a metallogenic connection between this particular magmatism and porphyry copper formation (Thiéblemont et al. 1997). Furthermore, a controversial causal relationship between adakitic magmatism and the size of porphyry copper deposits in northern Chile has been suggested by Oyarzún et al. (2001). They proposed that Late Eocene-Early Oligocene giant porphyry copper deposits such as Chuquicamata, are related to adakitic, highly oxidized, water-rich melts, and suggested that these melts were eventually derived from a slab source under flat subduction conditions, evolving as closed systems at depth. The genetic link between oceanic lithosphere and the origin of the Los Pelambres deposit seems to be clearer than that given by Oyarzún et al. (2001) for the Chilean Late Eocene-Early Oligocene porphyry copper deposits, because the geodynamic setting is better constrained. In the case studied here, we recognized a close temporal and spatial relationship between a particular ocean floor structure and the formation of the Los Pelambres rocks. In fact, the robust paleotectonic reconstructions of the floor of the Pacific Ocean and geochemical evidence of the Los Pelambres igneous rocks, suggest that its magma source could be related to the JFR subduction and equilibrated under garnet-amphibolite facies. The fluid release on breakdown of an amphibole-bearing residual mineralogy to garnet-bearing assemblages during the melting process has been considered of fundamental importance for the formation of the large central Andean ore deposits (Kay et al. 1999; Kay and Mpodozis 2001). In this scenario, recycling of metals associated with both the magmatic activity of a subducted hotspot chain and crustal material tectonically incorporated into the source region could be a relevant process to explain the origin of the mineralization in the Los Pelambres porphyry copper deposit. However, this mechanism does not explain why contemporaneous rocks such as the nearby La Gloria intrusion apparently lack mineralization.

Although the Los Bronces-Rio Blanco and El Teniente porphyry copper systems located southward (33 and 34°S, respectively) have steep REE patterns and high Sr contents (Kay et al. 1999; Rabbia et al. 2000), the proposed genesis of the Los Pelambres rocks cannot be extended to them. These deposits are younger (Early Pliocene; cf. Serrano et al. 1996; Stern and Skewes 1995) and emplaced beyond the influence area of the JFR as a source region. In the case of the Los Bronces-Rio Blanco deposit, the high La/Yb ratios that are exhibited by small volumes of late intrusive rocks, correlate with higher SiO2 values, which has been interpreted as due to late stage fractional crystallization involving amphibole (López-Escobar 1982).