Introduction

Granitoid plutons are essential constituents of collision belts. They are characterized by a large compositional diversity arising from different source compositions, variable melting conditions, and complex chemical and physical interactions between mafic and felsic magmas, fractional crystallization and crustal contamination (e.g. DePaolo 1981; Zorpi et al. 1991; Roberts and Clemens 1993; Galan et al. 1996; Thompson and Connolly 1995; Altherr et al. 1999, 2000; Altherr and Siebel 2002). There is a strong link between mineralogy, geochemical and isotopic composition and the geodynamic setting of granitoids. For this reason, compositionally well-characterized granitoids of known age may constrain the evolution and development of the continental crust through geological time (Barbarin 1999). Although numerous granitoid plutons have been extensively studied in terms of geochemistry and petrology, the transition from subduction to post-collisional extension still attracts attention and is often debated. The relationships among the stages of collision, the nature of the produced melts and their chemical composition are controversially discussed in magmatic provinces such as the Alpine-Himalayan orogenic belt. This belt embraces various arc, collision and post-collision geologic settings and magmatic rocks were formed during each of these settings. In the Alpine–Himalayan belt, Turkey as a zone of interaction between two major tectonic plates, Eurasia and Gondwanaland, lies in an important geodynamic position. The Pontide unit (Ketin 1966) of Turkey includes various eruptive and intrusive rocks, constituting the widespread Eastern Pontide Terrane, many of which are related to the convergence of these two plates (Fig. 1).

Fig. 1
figure 1

Simplified map showing the main granitoid distribution in the Eastern Pontides (modified from Gedik at al. 1992 and Güven 1993). Tectonic map of Turkey (from Ketin 1966), and major structures of the Eastern Pontides

Full size image

The Eastern Pontides (Ketin 1966), extending from the Black Sea coast to about 200 km south and straddling the North Anatolian Fault, form the northern margin of Anatolia (Fig. 1). The Eastern Pontides provide insight into paleo-island arc and long-term crustal evolution from pre-subduction rifting, through arc volcanism and plutonism to post-subduction alkaline volcanism (Akın 1978; Şengör and Yılmaz 1981; Akıncı 1984). The present-day geological setting of the Eastern Pontides is mainly the result of three main Neo-Tethyan volcanic cycles during the Jurassic, Upper Cretaceous and Eocene (Adamia et al. 1977; Eğin et al. 1979; Kazmin et al. 1986; Korkmaz et al. 1995; Çamur et al. 1996; Arslan et al. 1997, 2000). A number of studies (Tokel 1977; Akın 1978; Yılmaz 1981; Ercan and Gedik 1983; Robinson et al. 1995; Genç and Yılmaz 1995; Çamur et al. 1996; Yılmaz and Boztuğ 1997; Okay and Şahintürk 1997; Arslan et al. 1997; Şen et al. 1998) have stressed the different temporal, spatial and compositional relationships between the three Neo-Tethyan convergence systems.

In the Pontides of Turkey, particularly in the Eastern Pontides, the magmatic arc has been interpreted as related to the evolution of the Tethyan Ocean between Palaeozoic and Tertiary times (Şengör and Yılmaz 1981). In the western part of the Eastern Pontides, the Jurassic-Lower Cretaceous sequence has been interpreted to have formed during a rifting event and to indicate a passive continental margin, while during the Late Cretaceous both the eastern and western parts of the Eastern Pontides became an active continental margin (Yılmaz et al. 2000). The collision of the Eastern Pontides island arc with the Anatolides-Taurides along the İzmir-Ankara-Erzincan Suture Zone in the south occurred during Palaeocene–Early Eocene compressional events (Okay and Şahintürk 1997; Kazmin et al. 2000). Subsequent regional extension, only a few million years following the collision, was associated with post-collisional magmatism in the Eastern Pontides.

Recent studies (Boztuğ et al. 2003, 2004, 2006) have identified intrusive units derived from different geodynamic environments, with different ages, compositions and emplacement levels in the crust. These range from Early Cretaceous (Delaloye et al. 1972; Giles, 1974; Taner, 1977; Gedikoğlu, 1978; Moore et al. 1980; Boztuğ et al. 2003) through Late Cretaceous (Taner, 1977; Moore et al. 1980; Jica, 1986; Okay and Şahintürk, 1997; Yılmaz et al. 1997; Kaygusuz 2000; Köprübaşı et al. 2000, Boztuğ et al. 2006) to Eocene (Karslı 2002; Arslan et al. 2004; Boztuğ et al. 2004; Şahin 2005; Topuz et al. 2005; Boztuğ et al. 2006) in age. Rocks comprise low-K tholeites (sometimes high-K), calc-alkaline metaluminous granitoids, peraluminous leucogranites, silica-oversaturated alkaline syenites and monzonites (Yılmaz and Boztuğ, 1996; Boztuğ et al. 2003) which formed during arc-collisional, syn-collisional crustal thickening and post-collisional extensional regimes (Yılmaz and Boztuğ, 1996; Okay and Şahintürk, 1997; Yılmaz et al. 1997; Boztuğ et al. 2003; Yeğingil et al. 2002). The Early/Late Cretaceous to Early Palaeocene arc-related granitoids have calc-alkaline, metaluminous I-type character, and show characteristics of magma mingling–mixing between coeval mafic and felsic magma sources (Okay and Şahintürk 1997; Yılmaz et al. 1997; Boztuğ et al. 2003). The syn-collisional granitoids, which intrude the arc-related granodiorites and granites, are biotite leucogranites with peraluminous S-type chemical and mineralogical compositions (Boztuğ et al. 2003). The Eocene post-collisional extension-related intrusive rocks form small stocks of diorites and quartz diorites/monzodiorites of calc-alkaline to tholeitic affinity, and syenites to monzonites of alkaline affinity (Yılmaz and Boztuğ 1996; Boztuğ et al. 2003; Arslan and Aslan 2005). In the Torul region of the Eastern Pontides, arc-related magmatism developed under a compressional regime and is characterized by the predominance of calc-alkaline granitoids.

In this paper, we present a comprehensive set of new geochemical, Sr, Nd and oxygen isotope data on the Torul pluton, eastern Pontide magmatic arc. This arc comprises several granitoid plutons of different sizes. Although some of the Eastern Pontide arc granitoids were geochemically studied and radiometrically dated (Fig. 1), not much work was done to constrain their magma sources and magma production processes.

Geological setting

The Eastern Pontides can be divided into three zones: the northern, southern and axial zones (Bektaş et al. 1999). These zones have different lithological characteristics and are separated by E–W, NE–SW and NW–SE oriented fault zones that define the block faulted tectonic style of the eastern Pontides (Bektaş and Çapkınoğlu 1997). The northern zone of the eastern Pontides contains extensive Upper Cretaceous and Tertiary volcanic rocks (Pontide magmatic arc) cropping out mainly along the Black Sea coast. Palaeozoic plutonic and metamorphic rocks and their sedimentary cover sequences characterize the southern zone. The top of the succession consists of Eocene sediments with volcanic lithoclasts and macro foraminifera. In the axial zone, Mesozoic Alpine type peridotites and gabbro-diorites occur.

The Torul pluton covers an area of approximately 40 km2 and is an elliptical body with the long axes extending NE–SW. The pluton separates the southern and northern zone of the Eastern Pontides (Kaygusuz and Şen 1998) (Fig. 2). The southern zone comprises lower Jurassic volcanic and pyroclastic rocks at the bottom and Eocene volcanic rocks at the top. Cretaceous to Palaeocene sediments are intercalated into the volcanic series. The northern volcanic series is made up of Late Cretaceous basic and acidic volcanic rocks interbedded with sedimentary layers. According to field observations, the Torul pluton cuts upper Jurassic, early Cretaceous and late Cretaceous formations, and is cut by aplitic, dacitic and andesitic dykes (Kaygusuz 2000). A K–Ar age of 72.1 ± 3.6 Ma has been obtained for a quartz monzonite sample from the pluton (Jica 1986).

Fig. 2
figure 2

Location and geological map of the investigated area (modified from Kaygusuz 2000)

Full size image

The Torul pluton is a composite pluton which can be subdivided into three major units. These units are, from oldest to youngest: (1) small stocks of syenogranites (77.9 ± 0.3 Ma, this study), (2) granodiorites, biotite hornblende monzogranites, quartz monzodiorites, quartz monzonites, hornblende biotite monzogranites (72.1 ± 3.6 Ma; Jica 1986), and, (3) small stocks of quartz diorites which cut the granodiorites and biotite hornblende monzogranites representing the youngest subunit within the Torul pluton. The syenogranites and the quartz diorites form small outcrops, whereas the granodiorites, biotite hornblende monzogranites, quartz monzodiorites, quartz monzonites and hornblende biotite monzogranite make up the main volume of the pluton (Fig. 2).

Analytical techniques

More than 300 samples were collected from the Torul Pluton. The modal mineralogy of 267 samples was determined by point counting with a Swift automatic counter fitted to a polarizing microscope. On each thin-section a total of 1,000–1,250 points were counted. Modes were normalized to 100% (Table 1).

Table 1 General petrographic features of the various rock-types from the Torul pluton
Full size table

After macroscopic and microscopic investigation, 65 samples were selected for major, trace and rare earth element analyses. Major and trace elements were determined by wavelength dispersive XRF at the Department of Geology, University of New Brunswick (Canada), by full automated energy dispersive XRF at the Geological Engineering Department, Karadeniz Technical University (Turkey), and by ICP-emission spectrometry and ICP-mass spectrometry at ACME, Analytical Laboratories Ltd., Vancouver (Canada), using standard techniques. Major and trace elements were analysed by ICP after 0.2 g of rock-powder was fused with 1.5 g LiBO2 and dissolved in 100 ml 5% HNO3. Loss on ignition (LOI) was determined on the dried samples heated at 1,000°C. Rare earth element analyses were performed by ICP-MS at ACME Analytical Laboratories Ltd., Vancouver. Detection limits range from 0.01 to 0.1 wt% for major oxides, from 0.1 to 10 ppm for trace elements, and from 0.01 to 0.5 ppm for the rare earth elements.

Mineral analyses were carried out at the University of New Brunswick (Canada) with a JEOL JSM-6400 scanning electron microscope equipped with a Link eXL energy dispersive analyser and a single wavelength dispersive channel. X-ray analyses were done at an acceleration potential of 15 kV and sample currents of 2.5 nA, using a live time of 100 s for energy dispersive data acquisition. Data were reduced with the Link ZAF procedure using a combination of mineral (orthoclase-K, albite-Na, hornblende-Al, olivine-Mg, pyroxene-Si, K, Ca) and metal (Fe, Ti) standards. Analytical results are presented in Tables 2 and 3. Detection limits are generally about 0.1 wt%.

Table 2 Microprobe analysis of plagioclase and K-feldspar from the Torul pluton
Full size table
Table 3 Microprobe analysis of hornblende, biotite and pyroxene from the Torul pluton
Full size table

Rb–Sr, Sm–Nd and oxygen isotope analyses were performed at the Institute of Geosciences, Tübingen. For the radiogenic isotope study, very fine grained (<10 μm) sample powders were digested in 52% HF for 4 days at 140°C on a hot plate. Rb-Sr and Sm–Nd whole-rock samples as well as Rb–Sr biotite samples were spiked with mixed 149Sm–150Nd and 87Rb/84Sr tracer solutions. Digested samples were dried and redissolved in 6N HCl, dried again and redissolved in 2.5N HCl. After digestion, Rb, Sr and the light rare earth elements were isolated by conventional ion exchange chromatography with a 5 ml resin bed of Bio Rad AG 50W–X12, 200–400 mesh. Sm and Nd was separated from other rare-earth elements on quartz columns using 1.7 ml Teflon powder coated with HDEHP as cation exchange medium. All isotopic measurements were made by Thermal Ionization Mass Spectrometry, on a Finnigan MAT 262 mass spectrometer in static mode. The 87Sr/86Sr and 143Nd/144Nd isotope ratios were normalized to 86Sr/88Sr =  0.1194 and to 146Nd/144 Nd =  0.7219, respectively. The LaJolla Nd standard gave a 143Nd/144 Nd ratio of 0.511867 ± 0.000018 (2σ standard deviation, reference value 0.511850). The NBS 987 Sr standard yielded an 87Sr/86 Sr ratio of 0.710266 ± 0.000020 (2σstandard deviation, reference value 0.710240). Total procedural blanks (chemistry and loading) were <200 pg for Sr and <80 pg for Nd. 143Nd/144Nd ratios are quoted in the εNd notation as deviations from a chondritic reference (CHUR) with present-day 143Nd/144Nd ratio of 0.512638. Based on external reproducibility of ten separate standard loads of granite G2 (USGS standard material), uncertainties of 1.5% for 87Rb/86 Sr ratio and 0.05% for the 87Sr/86Sr (2σ confidence level) were used for biotite age calculation. The uncertainty of the initial isotope ratio resulting from age correction varies between 0.02 and 0.03% (for formulae see Siebel et al. 2005).

Oxygen was extracted from approximately 10 mg of dried whole-rock powder using modified versions of the standard techniques of the flow scheme silicate line (Clayton and Mayeda 1963) at 600°C by employing BrF5 as reagent. Quantitative oxygen yields were between 95 and 100%. Oxygen was converted to CO2 at a graphite rod heated by a Pt-coil and CO2 was analysed for its 18O/16O ratio with a Finnigan MAT 251 gas source mass spectrometer at the University of Tübingen. The isotopic ratios are reported in the δ-notation relative to Vienna standard mean ocean water (V-SMOW). All analyses were duplicated and analytical precision is better than ±0.2‰. A δ18O value of +9.7 ± 0.1‰ (1σm) was obtained for the NBS–28 quartz standard and all data are normalized to this value.

Results

Petrography and field relations

Petrographic characteristics of the Torul granitoids are presented in Table 1. The QAP (quartz–alkali feldspar–plagioclase) modal classification (Streckeisen 1976) is used here to address the rock types in Torul pluton (Fig. 3).

Fig. 3
figure 3

Classification based on modal compositions of the Torul granitoid (Streckeisen 1976)

Full size image

Syenogranites are the oldest rocks, and form two circular intrusions near the centre of the pluton in areas with smooth surfaces where the relief has been largely eroded away. Granodiorites and biotite hornblende monzogranites constitute the bulk of the pluton (Fig. 2). The granodiorites are grey to light grey in colour and have a fine to medium-grained texture, which is feldspar porphyritic near the contact to the country rocks. Contacts between the granodiorites and the biotite hornblende monzogranites are gradational. The biotite hornblende monzogranites are medium-grained and pink to pinkish grey in colour. The quartz monzodiorites, quartz monzonites and biotite hornblende monzogranites occur within the centre of the elliptically shaped pluton (Fig. 2). Internal contacts between all these intrusions are gradational but the quartz monzodiorites and quartz monzonites have sharp contacts with the granodiorites and biotite–hornblende monzogranites. The quartz monzodiorites are medium-grained and grey to dark grey in colour. The quartz monzonites are grey to light grey in colour and medium-grained but grain size decreases towards the contacts. The hornblende biotite monzogranites are fine to medium-grained and hypidiomorphic to porphyritic with feldspar crystals, and pink to pinkish grey in colour. The young quartz diorites comprise dark green to dark grey rocks forming small outcrops at the northern and southern central margin of the pluton where they intrude the granodiorites and biotite hornblende monzogranites (Fig. 2). Contacts between the quartz diorites and these rocks are sharp. The quartz monzodiorites, granodiorites and monzogranites also contain mafic microgranular enclaves (MMEs). The contacts between the enclaves and the host rocks are sharp.

Rock samples are generally holocrystalline, fine to medium grained, porphyric, poikilitic, myrmekitic, anti-rapakivi and rarely micrographic (Fig. 4a) in texture. Towards the contact with the volcanic country rocks, the granitoids possess fine-grained textures; towards the centre of the pluton medium-grained textures predominate. Porphyric textures are generally seen close to the contact of the volcanic country rocks, especially in the quartz diorites, rarely in the syenogranites and in the contact zones of the different rock types. Plagioclase, K-feldspar, quartz, hornblende, biotite, pyroxene and actinolite-tremolite are major mineral phases in all rock types; titanite, allanite, apatite, zircon, epidote and opaque minerals form accessory minerals; secondary minerals comprise chlorite, calcite, sericite and clay minerals.

Fig. 4
figure 4

a–h Microphotographs showing certain textural features of the Torul granitoids and associated rocks: a graphic texture; b oscillatory zoning in the plagioclase; c poikilitic texture in plagioclase in which some large plagioclase may contain small plagioclase crystals; d K-feldspar megacrystals; e biotite, plagioclase and opaque mineral inclusions in large K-feldspar; f plagioclase and opaque mineral inclusions in large hornblende crystals; g small plagioclase and hornblende in large biotite crystals; h accicular apatite crystals in the plagioclase, and poikilitic biotite with fine-grained plagioclase in MMEs (Pl Plagioclase, Kf K-feldspar, Bi Biotite, Hb Hornblende, Q Quartz, Ap Apatite, Op Opaque minerals)

Full size image

Plagioclase mainly forms subhedral to anhedral, normally and reversely zoned prismatic and lath-shaped crystals. Grain size variations range from 0.2 mm for inclusions to 2.5 mm for large crystals. A wide range in anorthite composition (An23 to An53, Table 2) can be found. The anorthite component decreases from rim to centre of the intrusion. Plagioclase mainly shows oscillatory zoning (Fig. 4b), albite twinning and prismatic-cellular growth. Normally zoned plagioclase has ∼An46 in the centre and ∼An26 at the rim. A myrmekitic texture is observed at the grain boundaries between orthoclase and plagioclase. Plagioclase also shows anti-rapakivi texture, mantled by orthoclase. Some plagioclase crystals have poikilitic textures, in which large plagioclase crystals (up to 2.5 mm) may contain small crystals of plagioclase (Fig. 4c), hornblende, biotite and apatite. Some large plagioclase crystals are altered to sericite and clay minerals.

K-feldspar forms anhedral, rarely subhedral crystals of perthitic orthoclase. Large phenocrysts (up to 3 mm) are heterogeneously distributed within the various rock types (Fig. 4d). K-feldspar composition is characterized by a variation in orthoclase content ranging from Or72 to Or95 (Table 2). The mineral shows a poikilitic texture in which abundant quartz, biotite, hornblende, plagioclase and opaque minerals are included in large K-feldspar oikocrysts (Fig. 4e). Alteration to clay minerals is more common in large K-feldspar crystals than in plagioclase.

Quartz is anhedral in shape and fills interstices between other minerals. It generally shows undulose extinction and its grain size becomes increasingly smaller in the contact zones between the different rock types.

Hornblende occurs as small euhedral to subhedral tabletish prismatic crystals. Its composition varies from Mg-hornblende through actinolitic hornblende to actinolite. Mg-number (Mg# = atomic ratios Mg/(Mg + Fe), where Fe is total iron) varies between 0.7 and 0.8 (Table 3). Hornblende is abundant in the quartz diorites and granodiorites. Towards the contact with the country rocks and in the transition zones between the different intrusions, the mineral is altered into chlorite, calcite and actinolite. In the quartz diorites, actinolitic hornblende and actinolite occur, whereas magnesium-hornblende is found mainly in the centres of the other intrusions. Large hornblende crystals (2–2.5 mm) may contain small plagioclase and biotite inclusions (Fig. 4f).

Biotite is abundant in the monzogranites and syenogranites. It is euhedral and subhedral reddish-brown and forms prismatic crystals and lamellas. In some samples, especially in those from the contact zones, biotite is altered into chlorite, epidote and opaque mineral along the cleavage planes. Biotite is Mg-rich (Fe2+/Fe2+ + Mg =  0.3−0.5, Table 3), and some large crystals (2–2.5 mm) may contain small plagioclase and opaque minerals showing poikilitic texture (Fig. 4g).

Pyroxene forms subhedral to anhedral crystals and is found less frequently than the other mafic minerals. Its composition is mainly augite and rare diopside. It is characterized by varying wollastonite content, ranging from 26 to 47% (Table 3). Some minerals are altered into uralite and calcite.

Titanite forms euhedral and subhedral crystals in all rock types but was not observed in the quartz diorites. Allanite occurs as euhedral, reddish crystals in all rock types except the granodiorites and quartz diorites. Needle-like crystals of apatite are mainly found in plagioclase. Zircon is seen as short euhedral and prismatic crystals within all rock types except the quartz diorites.

The MMEs of different sizes (0.5 to 9 cm) are ubiquitous in the Torul pluton. The MMEs are semi-rounded to ovoid in shape and very fine-grained compared to the host rock. They have quartz monzodiorite, monzodiorite and diorite compositions. MMEs have similar textural and mineralogical features as their host rocks. They consist of plagioclase, K-feldspar, quartz, hornblende, biotite, pyroxene, accompanied by accessory apatite, zircon and opaque minerals (magnetite). Plagioclase is subhedral, andesine (An42-49, optically) in composition, and has oscillatory zoning and albite-law twinning. Hornblende is the most abundant mafic mineral. Biotite is abundant at contacts between enclave and host rock, and may show chloritization. Acicular apatite, 0.2 mm in maximum length, is a frequent mineral in most MMEs (Fig. 4h). Such crystals are described from mafic melt globules trapped in silicic magmas and often cited as evidence of quenching (Wyllie et al. 1962).

Major and trace elements

Chemical analyses of representative samples from the Torul pluton are given in Tables 4 and 5. In the diagram of Debon and Le Fort (1982), the samples plot in quartz diorite/gabbro, granodiorite, quartz monzodiorite, quartz monzonite, adamellite and granite fields (Fig. 5), consistent with the QAP classification (Fig. 3). The term “adamellite” is obsolete and monzogranite is used instead (Le Maitre et al. 1989). The MMEs plot in quartz monzodiorite and diorite/gabbro fields (Fig. 5). Quartz diorites, MMEs, the other rock types (granodiorites, biotite hornblende monzogranites, quartz monzodiorites, quartz monzonites and hornblende biotite monzogranites) and syenogranites display distinct ranges of SiO2 contents (52–55, 53–57, 57–68 and 70–74 wt%, respectively) (Table 4; Fig. 6). Quartz diorites have lower SiO2 content than the MMEs and the other rock types. Harker plots of selected major and trace elements (Fig. 6a–r) show systematic variation in element concentration. The aluminium saturation index, ASI [molar Al2O3/(CaO +  Na2O +  K2O)], increases with increasing SiO2 from 0.84 to 1.09 (Fig. 6a). MMEs have similar ASI values, 0.78–0.92, as their host rocks. Apart from the quartz diorites, which belong to the medium-K series, the other rock types and the MMEs plot into the high-K field (Fig. 6b). For some major and trace elements (Fig. 6b, e, f, j, k, n, p), variation trends of the granodiorites and biotite hornblende monzogranites differ distinctly from those of the quartz monzodiorites, quartz monzonites and hornblende biotite monzogranites. Quartz diorites have the lowest K2O, Na2O, Rb, Th, Ba, Zr, Ce and Nb concentrations whereas syenogranites are depleted in CaO, MgO and Sr (Fig. 6b, e, f, h, k–r).

Table 4 Whole-rock major (wt%) and trace (ppm) element analyses of representative samples and CIPW norms of the granitoids from the Torul pluton
Full size table
Table 5 Rare earth elements analyses (ppm) of representative samples from the Torul pluton
Full size table
Fig. 5
figure 5

Chemical nomenclature diagram (Debon and Le Fort 1982) for selected samples from the Torul pluton. See Fig. 3 for explanation

Full size image
Fig. 6
figure 6

a–r SiO2 (wt%) versus major oxides (wt%) and trace elements (ppm) variation diagrams for samples from the Torul pluton. a ASI versus SiO2 with field boundaries between I-type and S-type of Chappell and White (1974) and peraluminous and metaluminous fields of Shand (1947). b K2 O versus SiO2 diagram with field boundaries between medium-K, high-K and shoshonitic series of Peccerillo and Taylor (1976). ASI (aluminium saturation index) = molar Al2O3/(Na2 O + K2 O + CaO). See Fig. 3 for explanation

Full size image

Chondrite-normalized rare earth element patterns of all rock types have concave-upward shape and are characterized by negative Eu anomalies with Eu/Eu* ranging from 0.59 to 0.84 (Fig. 7a–h; Table 5). MMEs display more negative Eu anomalies (Eu/Eu* = 0.46–0.63) than the other rock types, indicating plagioclase fractionation. The quartz diorites are less enriched in LREE compared to the other rock types and the MMEs. There is no systematic relationship between SiO2 content and magnitude of Eu anomaly. General trends of the primitive mantle normalized element concentration diagrams are shown in Fig. 8a–h. The quartz diorites are less enriched in incompatible elements when compared to the other rock types and MMEs. All rocks show enrichment of large ion lithophile elements (LILEs), depletion of high field strength elements (HFSEs) and prominent positive Pb anomalies. The depletion in HFSEs is best expressed by negative Nb–Ta anomalies. In addition, negative P, Hf and Ti anomalies and subdued negative Y anomalies are seen in the granodiorites, biotite hornblende monzogranites, quartz monzodiorite, quartz monzonites and the hornblende biotite monzogranites (Fig. 8c–g). The quartz diorites are depleted in Hf whereas the syenogranites are characterized by depletion of P and Ti.

Fig. 7
figure 7

a–h Chondrite normalized rare earth element patterns (normalizing values from Taylor and McLennan 1985) for selected samples from the Torul pluton. See Fig. 3 for explanation

Full size image
Fig. 8
figure 8

a–h Primitive mantle normalized trace element patterns (normalizing values from Sun and McDonough 1989) for selected samples from Torul pluton. See Fig. 3 for explanation

Full size image

Rb-Sr and Sm-Nd isotopes

Rb-Sr and Sm-Nd isotopic data are listed in Table 6 and plotted in Fig. 9a–c. Rb–Sr dating of biotite minerals separated from the syenogranites gave an age of 77.9 ± 0.3 Ma. (Fig. 10) slightly older than the K-Ar age of 72.1 ± 3.6 Ma reported for a quartz monzonite sample in an earlier study (Jica 1986). Initial Sr ratios and ɛNd(i) values were calculated using an age of 77 Ma for all rock types. Granodiorites, biotite hornblende monzogranites, quartz monzodiorites, quartz monzonites and hornblende biotite monzogranites show a small range in Sr–Nd values (initial 87Sr/86Sr from 0.7058 to 0.7072 and ɛNd(i) from −3.1 to −4.1). This group comprises rocks with different chemical composition. Two out of three syenogranite samples are displaced towards lower initial 87Sr/86Sr ratios (0.7034 and 0.7042) but contain the same ɛNd(i) values as the other samples. The quartz diorite samples belong to the most mafic rock type of the Torul pluton, have higher initial 87Sr/86Sr ratios (0.7075 and 0.7079) and lower ɛNd(i) values (−4.5 and −5.3) compared to the other samples.

Table 6 Rb-Sr, Sm-Nd and δ18O isotope data of granitoids from the Torul pluton
Full size table
Fig. 9
figure 9

a ɛNd(i) values versus 87Sr/86Sr(i) ratio; b and c εNd(i) and 87Sr/86Sr(i) versus SiO2, respectively. ɛNd(i) and 87Sr/86Sr(i) values are calculated for an age of 77 Ma

Full size image
Fig. 10
figure 10

87Rb/86Sr versus 87Sr/86Sr isochron diagram for a whole-rock and two biotite fractions from the syenogranite (sample 63b) of the Torul pluton. Numbers in brackets indicate grain size fractions of the biotite concentrates

Full size image

Oxygen isotopes

Whole-rock oxygen isotopic data are listed in Table 6 and plotted in Fig. 11a, c. δ18O values of the Torul granitoids vary between +3.5 and +8.2‰, similar to those commonly found in I-type granitoids (e.g. Clarke 1992). Samples from granodiorites, biotite hornblende monzogranites, quartz monzodiorites, quartz monzonites and hornblende biotite monzogranites show an even smaller range of δ18O values between +4.4 and +7.3. In the ɛNd(i)– δ18 O diagram (Fig. 11a), these samples define a weak trend of slightly increasing δ18O with decreasing ɛNd(i). A slightly positive correlation between δ18O values and SiO2 is observed for granodiorites, biotite hornblende monzogranites, quartz monzodiorites, quartz monzonites and hornblende biotite monzogranites (Fig. 11c). Such a trend, however, does not exist for the quartz diorite and syenogranite samples.

Fig. 11
figure 11

a ɛNd(i)value versus δ18O isotopic ratio; b δ18O values versus 87Sr/86Sr(i); c δ18O versus SiO2; ɛNd(i) and 87Sr/86Sr(i) values are calculated for an age of 77 Ma

Full size image

The measured δ18O values may reflect contributions from both mantle and crustal material. Intermediate and acid crustal igneous rocks are enriched in 18O relative to mantle rocks and the weak correlation between ɛNd(i), δ18O and silica (Fig. 11) is in line with this interpretation. Some samples are lower in δ18O than average mantle rocks (+5 to +6 ‰, e.g. Harmon and Hoefs 1995). Such low values can be attributed to contamination of the magmas by assimilation of hydrothermally altered rocks. Alternatively, the low δ18O values might be explained by hydrothermal alteration of the rocks through meteoric fluids at the subsurface. Meteoric fluids are depleted in δ18O relative to silicate rocks and open-system water–rock exchange can produce a decrease in δ18O in the rock phase.

Fractional crystallization

Major and trace element variation trends bear evidence that fractional crystallization has taken place during the formation of the Torul granitoids (Fig. 6). This is also supported by the depletion in Ba, Sr, P, and Eu (Fig. 8). Depletion in P results from removal of apatite during fractional crystallization. Negative Eu anomalies (Fig. 7) require fractionation of plagioclase and/or K-feldspar. Fractionation of feldspar would also result in depletion of Sr and Ba. Negative Eu anomalies and a decrease of Sr with increasing silica (Fig. 6 l) demonstrate that plagioclase was an important fractionating phase. The decrease of Al2O3, TiO2, CaO, MgO, Fe2O3, P2O5, Sr, Ni and increase of K2O, Rb with increasing silica (Fig. 6b–g, i–k) is related to the fractionation of plagioclase, amphibole, pyroxene, apatite and titanite. The increase of K2O and Rb with increasing silica indicates that K-feldspar and biotite were not early fractionation phases. This is in line with the late appearance of both minerals in the crystallization sequence.

Thermobarometry

Apatite and zircon saturation temperatures (Harrison and Watson 1983; Watson and Harrison 1983; Hanchar and Watson 2003; Miller et al. 2003) and temperatures based on the hornblende-plagioclase geothermometer (Blundy and Holland 1990; Holland and Blundy 1994) are given in Table 7. Although hornblende is a widely used mineral for thermobarometric studies in calc-alkaline igneous rocks (e.g. Stein and Dietl 2001), it alters easily in most magmatic environments, and the obtained pressures and temperatures do not always reflect the conditions during crystallization.

Table 7 Pressure and temperature calculations for granitoids from the Torul pluton obtained from zircon and apatite saturation temperature and plagioclase-hornblende geothermometry
Full size table

Within the Torul granitoids both Zr and P concentrations decrease with increasing SiO2 content (Fig. 6 i, p). With the exception of the quartz diorite, zircon and apatite saturation temperatures range from 850 to 900°C. The temperatures obtained for the quartz diorite are around 800°C. Since zircon and apatite are early crystallizing phases the saturation temperatures may represent near-liquidus relations. Calculated hornblende–plagioclase temperatures (670–760°C) are much lower compared to the zircon and apatite saturation temperatures. This can be explained by the lower closure temperature of hornblende or by subsequent alteration of this mineral. There is evidence for both. Subhedral hornblende crystals in the sample indicate the late-stage crystallization of hornblende and microprobe analysis show that some of the hornblendes are partially altered into tremolites.

The Torul pluton contains the critical mineral assemblage of, quartz + plagioclase + K-feldspar + hornblende + biotite + apatite + zircon + titanite + Fe–Ti oxide, for application of the Al-in-hornblende barometer (Hammastrom and Zen 1986; Andersen and Smith 1995; Hollister et al. 1987). Using the calibration of Anderson and Schmidt (1995), pressures between 1 and 2.5 kbar were obtained in the granodiorites, biotite hornblende monzogranites, quartz monzodiorites, quartz monzonites and hornblende biotite monzogranites. These values correspond to emplacement depths between ∼3 and 8 km.

Source rocks of the Torul granitoids

Dehydration melting experiments of various source rock lithologies (Patiño Douce and Johnston 1991; Wolf and Wyllie 1994; Patiño Douce and Beard 1996; Patiño Douce and McCarthy 1998) yield partial melts of distinct chemical signatures that allow for discrimination between compositionally different protoliths (Altherr and Siebel 2002; Topuz et al. 2005). Compositional differences of magmas produced by partial melting of different source rocks, such as amphibolites, tonalitic gneisses, metagraywackes, metapelites and amphibolites, under variable melting conditions, may be visualized in terms of molar oxide ratios. Dehydration melting of metapelites and metagreywackes (Rapp et al. 1991; Rapp 1995; Rapp and Watson 1995) yields higher K2O/Na2O, Mg#, Al2O3/(FeOtotal + MgO + TiO2), (Na2O + K2O)/(FeOtotal + MgO + TiO2) and lower CaO + FeOtotal + MgO + TiO2 values compared to the investigated rocks (Fig. 12). The chemical composition of the Torul pluton would be broadly compatible with an origin by dehydration melting from mafic lower crustal rocks (e.g. amphibolites). In addition, the syenogranites have slightly higher values of K2O/Na2O, Al2O3/(MgO + FeOtotal), Al2O3/(FeOtotal + MgO + TiO2), (Na2O + K2O)/ (FeOtotal + MgO + TiO2), whereas the quartz diorites have slightly higher Na2O + K2O + Al2O3 + CaO + FeOtotal + MgO + TiO2 values and Mg# compared to the other rock types in Torul pluton (Fig. 12).

Fig. 12
figure 12

a–f Chemical composition of the Torul granitoids. Outlined fields denote compositions of partial melts obtained in experimental studies by dehydration melting of various bulk compositions. MB metabasalts (bold-solid line); MA meta-andesites (dotted line); MGW metagreywackes (dashed line); MP metapelites (solid line); AMP amphibolites (bold-solid line). Data sources: Vielzeuf and Holloway (1988), Patiño Douce and Johnston (1991), Rapp et al. (1991), Gardien et al. (1995), Rapp (1995), Rapp and Watson (1995), Patiño Douce and Beard (1996), Stevens et al. (1997), Skjerlie and Johnston (1996), Patiño Douce (1997, 1999), Patiño Douce and McCarthy (1998). See Fig. 3 for explanation

Full size image

Mode of emplacement

Over the years, various modes of magma ascent and emplacement mechanisms, such as diapirism (Marsh 1982; Bateman 1984), stoping (Daly 1933; Marsh 1982), ballooning (Pitcher and Read 1963; Bateman 1985; Ramsay 1989) and dyking (Petford, 1996; Brown and Tryggvason 2001; Haederle and Atherton 2002) have been identified. The contacts between the Torul pluton and the country rocks are predominantly sharp and discordant, the contact facies are finer-grained, and the shape of the pluton is elliptic. The textures are massive, porphyritic and granophyric, and the granitoids contain abundant country rock xenoliths at the endocontact. There is a widespread foliation texture developed, especially at the margins of the pluton. All of these features show that the Torul pluton emplaced into shallow crustal depths either by a stoping type of ascent or by ballooning.

It is believed that subduction-related fracture tectonic played an important role during the emplacement of the granitoids within the Pontide magmatic arc (Gedikoğlu 1978). In general, the long axis of most granitic plutons shows good agreement with the major tectonic directions (NE–SW, NW–SE). These directions correspond to the two main fracture alignments in the Eastern Pontides as defined by Bektaş and Çapkınoğlu (1997), suggesting the important role of fractures during pluton emplacement. It seems therefore likely that the Torul pluton emplaced along a NE–SW trending fracture line during the Upper Cretaceous.

Tectonic setting

The Torul granitoids are medium- to high-K, calc-alkaline rocks enriched in LILEs such as Rb, Th, U and K with respect to the HFSEs, especially Nb and Ti (Fig. 8). Apart from the quartz diorite, all rock types show a significant positive Pb anomaly. Magmas with these chemical features are generally believed to be generated in subduction-related environments (e.g. Floyd and Winchester 1975; Rogers and Hawkesworth 1989; Sajona et al. 1996). Applying the discrimination criteria of Pearce et al. (1984), all samples plot in the fields of volcanic arc granites (VAG) and syn-collisional granites (Syn-COLG) in the Nb versus Y diagram (Fig. 13a), whereas in the Rb versus (Y+Nb) diagram, the samples plot in the VAG field (Fig. 13b). Lower-silica samples (quartz diorites, 52–55 wt% SiO2) fall within the VAG field, whereas the high-silica samples (70–74 wt% SiO2) plot around the boundary between the VAG and COLG fields. This shift is likely due to progressive magma differentiation (Förster et al. 1997).

Fig. 13
figure 13

a Nb-Y and b Rb-(Y+Nb) discrimination diagrams (Pearce et al. 1984) for samples from the Torul pluton, showing the fields of volcanic-arc granites (VAG), syn-collisional granites (Syn-COLG), within-plate granites (WPG) and ocean-ridge granites (ORG). See Fig. 3 for explanation

Full size image

It has been demonstrated (Brown et al. 1984) that the abundances of incompatible elements in granites can be correlated with the degree of arc maturity. Increase of the Rb/Zr ratio with increasing Nb and Y content is in accordance with the arc maturity, from primitive to mature. A comparison of the Torul pluton with arc-type granitoids is performed on the Nb versus Rb/Zr and Y versus Rb/Zr diagrams (Fig. 14). All rock types from the pluton plot in the normal arc fields (Fig. 14a, b), except the quartz diorites which plot at the primitive arc-normal arc transition (Fig. 14a).

Fig. 14
figure 14

a Nb versus Rb/Zr and b Y versus Rb/Zr diagrams (after Brown et al. 1984) for samples from the Torul pluton. See Fig. 3 for explanation

Full size image

Discussion

The Eastern Pontide belt represents a Mesozoic to Early Cenozoic paleo-subduction zone. The generation of subduction-related granitoids reached a peak during Late Cretaceous to Paleocene. The composition of the granitoids shifts from tonalitic-trondhjemitic through calc-alkaline to alkaline with increasing crustal thickness from Jurassic to Late Eocene (Köprübaşı et al. 2000).

The geochemical features of the Torul pluton (i.e. depletion of Ba, Sr, Nb, and Ti, enrichment of Rb, Th, K, Pb and La) are also compatible to those of typical crustal melts, e.g. granitoids of the Lachlan Fold belt (Chappell and White 1992). Several experimental studies (Wolf and Wyllie 1994; Rapp and Watson 1995) have shown that extremely high temperatures in excess of ∼1,100°C are needed to produce mafic metaluminous low-silica (∼58 wt%) melts by dehydration melting of metabasic crustal rocks. Such melts are generally characterized by low Mg# (∼44) and high Na2O contents (∼4.3 wt %). The Torul intrusions have low Mg# (38 to 14) and high Na2O contents (2.3–3.8 wt%). Compositional diversity among crustal magmas may arise in part from different source compositions, but variation of intrinsic parameters, such as pressure, temperature, oxygen fugacity, or water content during partial melting also plays an equally important role (e.g. Beard et al. 1994; Wolf and Wyllie 1994; Patiño Douce and Beard 1996; Thompson 1996; Stevens et al. 1997; Altherr et al. 2000). These parameters control the degree of partial melting and the stability fields of residual mineral phases (e.g. plagioclase, biotite, hornblende, orthopyroxene, garnet) that buffer resultant melt composition. Partial melts derived from mafic sources, for example, have lower K2O/Na2O, Al2O3/(MgO + FeOtotal), Mg#, Al2O3/(FeOtotal + MgO + TiO2), (Na2O + K2O)/(FeOtotal + MgO + TiO2), CaO/(FeOtotal + MgO + TiO2) and higher Na2O + K2O + Al2O3 + CaO + FeOtotal + MgO + TiO2 than those derived from metapelites and metagraywackes as seen in the Torul pluton (Fig. 12).

Trace element characteristics can yield important constraints on the depth of magma segregation and subsequent fractionation processes. Crustal lithologies show a depth-dependent mineralogy. For any fixed bulk-rock composition, the amount of garnet increases and the amount of plagioclase decreases with increasing depth. Certain trace element abundances and ratios can be used as indicators for the relative amounts of garnet, clinopyroxene, hornblende and plagioclase as fractionating phases either during magma segregation from the source region or subsequent fractionation (e.g. Altherr et al. 1999). Fractionation of garnet will lower the relative abundances of HREE, Y, and Sc in the remaining melt phase. Plagioclase fractionation will result in lower abundances of Ba and Sr, low Sr/Nd ratios, and negative Eu anomalies in the chondrite-normalized REE patterns of the melts. Fractionation of hornblende will cause an increase in LREE/HREE in the residual melt, but a concave-upward shape (e.g. Romick et al. 1992) should characterize the resulting chondrite-normalized REE pattern of the melt. The felsic syenogranites (SiO2-rich) have relatively high abundances of Yb, low abundances of Sr, and low values of Sr/Y, Sr/Nd, Zr/Y, and (Tb/Yb)cn. Chondrite-normalized REE patterns of these samples display variable negative Eu anomalies and are characterized by moderate concave upward shapes (Fig. 7b). These features preclude the presence of substantial amounts of garnet either during partial melting or as part of the fractionating assemblage during an AFC process (DePaolo 1981). Compared to the other rock types the quartz diorites display similar abundances of large ion lithophile elements (LILE), but higher abundances of Sr, lower abundances of Yb, and higher ratios of Sr/Y, Sr/Nd, Zr/Y and (Tb/Yb)cn. Chondrite-normalized REE patterns of these rocks display slightly negative Eu anomalies and are characterized by minor concave upward shapes (Fig. 7h), implying some feldspar and hornblende fractionation during magma differentiation.

The Torul pluton has textures that indicate magma mingling and mixing processes. The presence of MMEs may indicate mingling of coeval mafic (dioritic) and felsic magmas (e.g. Didier 1973; Vernon 1990; Didier and Barbarin 1991; Barbarin and Didier 1992). Similar mineral assemblages, mineral compositions and existence of strong correlations between major and trace elements show the MME and host granitoids, although concentrations of major and trace elements are different. The resorbed nature of plagioclase megacrysts in the enclaves and the similarity in anorthite contents between these crystals and host rock plagioclase crystals suggest a common origin. Most probably, plagioclase, together with K-feldspar and quartz from the granitic host magma was mechanically transferred to the mafic globules during the hybridization process. Oscillatory-zoned plagioclases (Fig. 4b), coexistence of two types of plagioclase phenocrysts, irregular changes of anorthite contents within plagioclase, poikilitic textures (Fig. 4c, e–g), acicular hornblende, blade and acicular biotite (Fig. 4b), acicular apatite in plagioclase (Fig. 4h), prismatic-cellular plagioclase growth and K-feldspar megacrysts in mafic microgranular enclaves (Barbarin 1988; Lesher 1990; Hibbard 1991; Baxter and Feely 2002) possibly record the mixing of coexisting mafic and felsic magmas. Since original compositions of the enclaves (former globules of mafic melt) were probably modified by interaction with the felsic host magma (as suggested for other occurrences; e.g. Waight et al. 2000), a more thorough discussion of the genesis of the microgranular enclaves is beyond the scope of this study.

Conclusions

The Torul pluton is a composite granitoid pluton which belongs to the medium to high-K, calc-alkaline series. Intrusion contact relationship, mineralogical, geochemical and isotopic characteristics allows defining three major suites. According to their emplacement chronology, these are (1) syenogranites, followed by (2) granodiorites, biotite hornblende monzogranites, quartz monzodiorites, quartz monzonites, hornblende biotite monzogranites and (3) quartz diorites.

The Torul pluton displays fractionated chondrite-normalized REE patterns with pronounced negative Eu anomalies. These features combined with the decrease of Al2O3, TiO2, CaO, MgO, Fe2O3, P2O5, and Sr with increasing silica suggest that the granitoids have undergone fractionation of plagioclase, amphibole, pyroxene, apatite and titanite. The geochemical and isotopic composition of the Torul pluton is broadly compatible with an origin by dehydration melting from a metabasaltic source.

Major and trace element patterns indicate that the Torul pluton consists of volcanic arc granitoids. The rocks represent a period of (immature) arc development and probably formed during an early stage of convergent plate boundary tectonics. Magmatism is related to subduction of the Neo-Tethyan Ocean beneath the Eurasian plate along the Izmir-Ankara-Erzincan suture zone during the Cretaceous.