On the geodynamics of the Alpine collisional granitoids from Central Anatolia: petrology, age and isotopic characteristics of the granitoids of the Ekecikdağ Igneous Association (Aksaray/Turkey)

ABSTRACT Granitoids of the Ekecikdağ Igneous Association (Central Anatolia/Turkey) are products of collisional–post-collisional magmatism in the Ekecikdağ area. These granitoids are granodiorite, microgranite and leucogranite. Field relations of granodiorites with microgranites is obscured, but leucogranites intrude both rock types. Mean zircon laser ablation (LA)-ICP-MS 206Pb-238U ages of granodiorites and microgranites are 84.52 ± 0.93 Ma and 80.7 ± 1.6 Ma, respectively, and age of leucogranites is suggested as 80 Ma, based on field relations combined with 206Pb/238U and Rb-Sr ages. Crystallisation temperatures of granodiorites, microgranites and leucogranites are 728°C-848°C, 797°C-880°C, 704°C-809°C, respectively. Geochemical characteristics including Sr-Nd isotopic evidences infer a non-cogenetic character, as there is a high crustal contribution in I-type granodiorite sources, a crustal source with insignificant and significant mantle inputs in S-type microgranites and leucogranites, respectively. LA-ICP-MS Lu-Hf isotope data from zircons reveal their crustal nature (εHf(t): −1.3 ± 0.5 to −8.8 ± 0.5). Crustal melting linked to the Alpine thickening during the Late Cretaceous led to formation of heterogeneous sourced granitoids with crustal dominated sources in the Ekecikdağ area. Understanding of the nature and evolution of collisional Ekecikdağ granitoids is not only important to put contribution in the geodynamic evolution of Central Anatolia and surrounding Alpine area, but also to better understand systematics of collisional magmatic systems.

Granodiorites are the main magmatic phase in the study area, and form large spheroidal greyish outcrops. It is characterised by phaneritic texture with abundant biotite and amphibole contents and coarse K-feldspar crystals in the field (Figure 2(a)). Sub-angular to rounded mafic microgranular enclaves are also characteristic of the granodiorites (Figure 2(b)). In the study area, especially in the northeastern part, there are no in-situ granitic outcrops. The loose soil in this area, however, is very different from the Oligocene-Miocene cover units, and consists of minerals, especially K-feldspar existing in the granodiorite. Hence these areas are mapped as granitic soil that is formed by decomposition of the granodiorite (Figure 1).
Microgranites are found in the NW of the study area, near Sinandı and Gökkaya villages and in the east near Yanyurt village (Figure 1). They are characteristically dark grey and microgranular with abundant biotite (Figure 2(c)), and are cut by the leucogranites (Figure 2(d)), but their relation with the granodiorites is obscured in the field by the cover units.
Gabbroic rocks in the study are intruded by the Ekecikdağ granitoids, and are described as dismembered ophiolitic rocks as in other parts of the CACC (e.g. Koçak, Isık, Arslan, & Zedef, 2005;Köksal, Toksoy-Köksal, & Göncüoğlu, 2017;Toksoy-Köksal et al., 2001;Toksoy-Köksal, Oberhaensli, & Göncüoğlu, 2009;Yalınız et al., 1999). The field relation of the granitoids with the metamorphic basement units is not observed in the study area. However, in nearby regions, intrusive contacts of the granodiorites and the leucogranites with the metamorphic rocks of the CACC can be observed. The cover units are Oligocene-Miocene horizontally layered fine-grained clayey-tuffaceous lacustrine deposits.

Petrography
The petrographical characteristics of the granitoids, along with their mineral chemistry data, were presented in detail by Toksoy- Köksal (2016). Here the summary of the petrographical features are given. Amphibole with ferro-and magnesio-hornblende composition is exclusively observed in the granodiorites. In addition, the granodiorites consist of quartz, orthoclase (generally kaolinized subhedral, perthitic crystals with Or 85-98 composition), plagioclase (rarely serizited, zoned subhedral with An 15-49 content), biotite (rarely chloritized, subhedral, transitional to Mg-biotite) as main phases with zircon, titanite, allanite (having euhedral crystals up to 1 mm in size), apatite and opaque minerals as accessory phases (Figure 3(a)). K-feldspar phenocrysts produce a porphyritic texture. Hornblendedominant microgranular enclaves are frequently observed within the granodiorites.

Analytical methods
In the scope of this study, 5 microgranite, 5 granodiorite, 12 leucogranite and 2 mafic microgranular enclave samples were analysed for whole-rock geochemistry in Acme Analytical Laboratories (Canada) and Central Laboratory of Middle East Technical University (Turkey). A subset of these samples, five from microgranites, five from granodiorites and four from leucogranites were also analysed for their wholerock Sr and Nd isotope ratios. Moreover, on zircon crystals, LA-SF-ICP-MS U-Pb age measurements were performed at the Geological Survey of Denmark and Greenland (GEUS, Denmark) and Lu-Hf isotope analyses at the Goethe-University Frankfurt (Germany). Major, trace and rare earth elements (REE) are measured after LiBO 2 /Li 2 B 4 O 7 fusion with ICP-AES, and after acid digestion (HNO 3 5%) with ICP-MS. Detection limits are 0.01 wt. % for SiO 2 , Al 2 O 3 , MgO, CaO, Na 2 O, K 2 O, MnO, TiO 2 , 0.04 wt. % for Fe 2 O 3 , 0.001-0.002 wt. % for P 2 O 5 and Cr 2 O 3 and 0.10 wt. % for LOI. Detection limits for trace and REE are 8 ppm in V, 1 ppm in Ba, 0.5 ppm in Sr, Gd and W, 0.3 ppm in Nd, 0.1 ppm in Cs, Hf, Nb, Rb, Ta, U, Y, Zr, Th, La and Ce, 0.05 ppm in Sm, Dy, Yb, 0.03 ppm in Er, 0.02 ppm in Pr, Eu and Ho, 0.01 ppm in Tb, Tm, Lu. Analytical precision is about 0.05-0.15% for major elements, and 0.5 to 1.5% in trace elements and REE. Strontium (Sr) and (Neodymium) Nd isotope analyses were performed at the Radiogenic Isotope Laboratory of Central Laboratory/Middle East Technical University (METU), (Ankara) after the methods described by Köksal et al. (2017) Nd/ 144 Nd = 0.511846 ± 5 (n = 3). No bias correction was applied for Sr and Nd analyses. Quality control of the isotope analyses was checked by applying the same procedures to the USGS rock standards. During the period of analyses, the AGV-1 USGS standard gave 87 Sr/ 86 Sr = 0.703993 ± 12 (n = 2) and 143 Nd/ 144 Nd = 0.512784 ± 3 (n = 2) and the G2 USGS standard gave 87 Sr/ 86 Sr = 0.709775 ± 10 (n = 2) and 143 Nd/ 144 Nd = 0.512224 ± 3 (n = 2), respectively. Detailes of U-Pb geochronology were given in Köksal et al. (2012), Köksal et al. (2013) and Gürsu et al. (2015). Zircon crystals were obtained through standard concentration methods (i.e. crushing, grinding, sieving, heavy mineral concentration by water table, heavy liquids and electro-magnetic procedures, handpicking under microscope), mounted in epoxy, polished up to half thicknesses and their back-scattered electron (BSE) images were taken by using a Philips XL 40 SEM at GEUS. In-situ LA-SF-ICP-MS U-Pb zircon measurements were performed on a ThermoFinnigan Element2 high resolution magnetic sector field ICP-MS coupled with a Merchantek New Wave 213 nm UV laser ablation system at the Geological Survey of Denmark and Greenland (GEUS) from pre-evaluated spots on the selected grains regarding their rim/core structures and absence of secondary effects like recrystallisation. Lutetium-Hafnium isotope analyses, on the other hand, were performed at Goethe-University Frankfurt with a Thermo-Finnigan Neptune multi-collector ICP-MS by static multicollection during 60 s of ablation with a . General petrographical view of (a) the granodiorite containing hornblende, plagioclase, quartz, orthoclase, biotite, (b) two mica microgranite with biotite and muscovite, and quartz, orthoclase, microcline, plagioclase and (c) primary muscovite in the leucogranite (hb: hornblende, bio: biotite, musc: muscovite, plag: plagioclase, orth: orthoclase, qtz: quartz).
spot size of 40 μm on selected grains after evaluation of U-Pb data. Discordance percentages of the U-Pb analyses were determined from the following formula; Granodiorites: The analysed zircon crystals have generally long prismatic with rare short prismatic forms, and their habits are euhedral to subhedral with oscillatory zoning, which includes inherited cores ( Figure 4). Ten spots from nine zircons have high discordia % (> 10 %) and their U and Pb contents are extremely high, with values that range respectively from 267 to 7091 ppm and 3 to 93 ppm (Table 1). In the remaining spots U and Pb values have discordance % lower than 10 %, and vary respectively from 65 to 749 ppm and 2 to 43 ppm (Table  1). Th/U ratios of the analysed spots having low discordia % (< 10) vary from 0.11 to 2.41. The rim and core ages of the analysed spots range between 83.0 ± 2.0 Ma and 762 ± 27 Ma. On the 206 Pb/ 238 U-207 Pb/ 235 U concordia diagram, no concordant age has been determined ( Figure 5) but the weighted ages of the 12 spots indicate 84.52 ± 0.93 Ma (MSWD = 1.3) ( Figure 5), and are evaluated as the crystallisation age of the granodiorites.
Microgranites: 26 spots on 14 zircons are measured during U-Pb dating analyses of the microgranites. Three spots discordia % on three zircons exceed 10 % and cannot be used in the age studies. The analysed zircons are euhedral-subhedral, long to short prismatic, and have well-developed oscillatory zones ( Figure 6). Th/U ratios vary from 0.41 to 1.60 indicating a typical magmatic origin. The younger ages from the rims of the selected zircons range from 79.0 ± 2.0 Ma to 86.0 ± 3.0 Ma, whereas the core ages vary from 618 ± 24 to 1008 ± 16 Ma (Table 1). No concordia ages between the younger ages have been determined, but a 80.7 ± 1.6 Ma (MSWD = 2.1 weighted mean age from eight spots is interpreted as reflecting the crystallisation age (Figure 7).
Leucogranites: Fourteen spots on the six long prismatic-euhedral oscillatory zoned zircons were analysed to determine their ages ( Figure 8). Only three of the 14 spots have low discordance % (< 10 %) and two of them yield 82.7 ± 1.0 Ma (MSWD = 1.6) concordia ages and the oldest inherited age is determined as 156.0 ± 6.0 Ma in the sample. The condordia age (82.7 ± 1.0 Ma) obtained from two spots is interpreted as their crystallisation age (Figure 9(a)). The determined 206 Pb/ 238 U age of the leucogranites from two spots are older than the age of the microgranites. The leucogranites, however, cut the microgranites and the granodiorites in the field (Figure 2(d)). Therefore, a further Rb-Sr isochron age was determined as 75.3 ± 4.7 Ma despite having a high MSWD value (Figure 9(b)). The high MSWD value could be due to slight alteration in a sample of the leucogranite. Therefore the oldest crystallisation age for the leucogranites is suggested as 80 Ma and this value is used in the calculation of initial isotopic data.

Whole-rock geochemistry
Results of whole-rock analyses are presented in Table 2. These data were examined by Toksoy-Köksal (2016)except enclave samples EK-9 and EK-15but in this study the data are re-evaluated with additional interpretations. The mobility of the major elements was checked on basis of the alteration indices proposed by Hughes (1973) and Spitz and Darling (1978). On K 2 O+Na 2 O versus (K 2 O/(K 2 O+Na 2 O))*100 diagram (Hughes, 1973), all granitic samples plotted within the range of the igneous spectrum/weakly altered areas, whereas they are placed within the fresh to weakly altered area on an Al 2 O 3 /Na 2 O versus Na 2 O diagram (Spitz & Darling, 1978) (Figure 10). The alteration diagrams indicate that all the samples are free of Na-K metasomatism/alteration that can be used in the further petrogenetic studies.
All the granitoids have high-K calc-alkaline character but their aluminium indexes are different (Toksoy- Köksal, 2016). The granodiorites show metaluminous to peraluminous character, but the microgranites and the leucogranites are peraluminous, with higher ANK/ ACNK ratios for the microgranites. On the Zr-TiO 2 diagram (Figure 11(a)), the petrographically described granodiorites and microgranites plot in the granodiorite area, whereas the leucogranites are placed in the granite area (Figure 11(a)). These geochemical based rock types are supported by plots in the granodiorite and granite fields on the R1 versus R2 diagram of the De La Roche, Leterrier, Grandclaude, and Marchal (1980) (Figure 11b). On the SiO 2 versus major element variation diagrams, the leucogranite samples have higher SiO 2 , Na 2 O, K 2 O but lower Al 2 O 3 , Fe 2 O 3 (T) , MgO, CaO, TiO 2 , P 2 O 5 contents than the granodiorites Table 1. LA-ICP-MS U-Pb age data of zircons from the EIA granitoids.  are no steady continuous trends observed among the granitoid rocks ( Figure 13). The plots against TiO 2 except for P 2 O 5 ( Figure 13) for the granodiorite and the microgranite give almost parallel trends. As in the Harker diagrams, no well-defined trends are observed from the granodiorites and the microgranites to the leucogranites on the SiO 2 against Rb, Sr, Nb, Th, Zr and La diagrams. In all the variation plots, the leucogranites separate clearly from the others, while no clear relation between the granodiorites and the microgranites is evident. The granitoids other than the leucogranites have roughly similar Rb, La but clear differences in Sr, Nb, Th, Zr ( Figure 12). The granodiorites are lower in Sr, Nb, Zr but higher in Th than the microgranites.  Figure 14). However, the granodiorites with lower Nb and higher Th contents have a lower Nb/Th ratio that separates them from the microgranites.
The granitoids commonly exhibit enrichment in Th, U, K, Pb, LREE and depletion in Ba, Nb, Sr, P, Eu, Ti elements with flat middle rare earth element (MREE) and heavy rare earth element (HREE) patterns on the chondrite and primitive mantle normalised spider diagrams (after Sun & Mcdonough, 1989). However, positive and negative element patterns of the leucogranites are pronouncedly different from the microgranites and the granodiorites ( Figure 15). The leucogranites have higher Th, U contents, lower LREE and more pronounced negative anomalies in Eu, Ti than the granodiorites and the microgranites ( Figure 15). On the chondrite normalised REE diagram, all studied samples show LREE enrichment compared to  * Weighted mean aged used, ** Concordia ages used, *** Discordia % > 10 not used in the interpretations MREE and HREE. However, the granodiorites and the microgranites have higher enrichment in LREE levels with minor depletion of Eu on a chondrite normalised diagram, whereas the leucogranites with a similar pattern in LREE display lower values with distinctive depletion of Eu. The leucogranites are relatively enriched in HREE ( Figure 15). Even though similar enrichment and depletion patterns for both of the microgranites and the granodiorites are present, the microgranites are more enriched in Ba, Sr, P but has lower Th, U, Pb values than the granodiorites. Moreover, the MREE of the microgranites are higher than the granodiorites.
Geochemical plots on tectonic discrimination diagrams by Pearce, Harris, and Tindle (1984), Batchelor and Bowden (1985), Maniar and Piccoli (1989) and Thiéblemont and Cabanis (1990) are used to better understand environments of formation for the granitoids. All of the granitoids plot at the junction point of Volcanic Arc Granites (VAG), Syn-collisional granites (Syn-Colg) and Within Plate Granites (WPG) on the Rb-Nb+Y diagram, that infers a post-collisional system (Figure 16(a)). On the R1-R2 diagram, they are close to syn-collision and post-orogenic areas, whereas they plot within the area of the post-collisional granite area on the Al 2 O 3 -SiO 2 diagram (Figure 16(b,c)). The granitoids concentrate in the post-collisional/syn-subduction area on the Y-Rb-Nb diagram (Figure 16(d)). To sum up, all of the Ekecikdağ granitoids might have been generated in a collisional to post-collisional system, but they still display differences in geochemical features.

Whole-rock Sr and Nd isotopes
Strontium-Nd isotopic study was conducted to shed light on the source characteristics (Table 3; Figure 17). The initial 87 Sr/ 86 Sr and 143 Nd/ 144 Nd values of the samples are 0.709538-0.712353 and 0.512114-0.512183 for the granodiorites, 0.713279-0.713744 and 0.512086-0.512097 for the microgranites, and 0.707138-0.709765 and 0.512143-0.512230 for the leucogranites, respectively (Table 3). ƐNd (T) values of the samples range from −6.76 to −8.11 in the granodiorites, −8.62 to −8.75 in the microgranites, and −5.95 to −7.67 in the leucogranites (Table 3). Whole-rock Sr and Nd isotope data from the granitoids display high initial 87 Sr/ 86 Sr (higher than 0.707) and low εNd values (lower than −5.95) and plot in the  area between EMI, EMII and crustal sources (Figure 17). Among the Ekecikdağ granitoids, the microgranites have the most enriched source component, while the leucogranites have the least. Collectively Sr-Nd isotopic data point to enriched sources ( Figure 17).

Zircon Lu-Hf isotope systematics
In-situ LA-MC-ICP-MS Lu-Hf isotope analyses were performed on the selected zircon crystals (Figures 4, 6, 8) from the EIA granitoids. The initial 176 Hf/ 177 Hf ratios were calculated based on the U-Pb ages of the given   points, and results are presented in Table 4. The low initial 176 Hf/ 177 Hf ratios along with low εHf (t) values (i.e. <0.28268; <-1.3) infer crustal derived sources ( Figure 18). Furthermore, T DM Hf ages corresponding to the Late Cretaceous LA-ICP-MS analyses provide a minimum age for the source material of about 1.08-1.43 Ga, that almost coincides with the T DM Nd ages range of the EIA granitoids (i.e. 1.37-1.60 Ga, Table 3). T DM Hf ages from the inherited zircon analyses, on the other hand, are older, ranging from 1.51 to 2.59 Ga.

Zircon and apatite saturation temperatures
Determination of crystallisation temperature and pressure of the Ekecikdağ granitoid is difficult due to limited mineral assemblages suitable for geothermobarometric calculations. Despite the limitations, geothermometric calculations were conducted by Toksoy- Köksal (2016) that resulted in magmatic to sub-solidus temperatures (granodiorite: 615-783°C, microgranite: 464-565°C, leucogranite: 456-482°C). In this study, therefore, zircon and apatite saturation temperatures from the whole rock element geochemical data were used to determine the magma tempratures of the studied samples using the formulae of Watson and Harrison (1983) and Harrison and Watson (1984). The zircon saturation temperatures used in this study are assumed to be free of the inheritance in the cores. The zircon saturation tempratures are calculated from the following formula;  (Harrison & Watson, 1984) Temperatures from zircon saturation calculations are 750°C-770°C (average 758°C) for the granodiorites, 786°C-808°C (average 797°C) for the microgranites,   (Hughes, 1973) and (b) Al 2 O 3 /Na 2 O against Na 2 O diagram (Spitz & Darling, 1978). The area defined by dashed line represent the data compiled from the Ağaçören Intrusive Suite granitoids (9 samples from Kadıoğlu et al., 2003; 18 samples from Kadıoğlu et al., 2006; 6 samples from Köksal et al., 2012).         (Continued) and 644°C-734°C (average 704°C) for the leucogranites. Their apatite saturation temperatures vary from 763°C to 883°C (average 850°C) for the granodiorites, from 790°C to 947°C for the microgranites (880°C), and from 769°C to 856°C (average 809°C) for the leucogranites. The zircon and apatite saturation temperatures imply that the magma temperatures are in range of 758°C and 850°C for the granodiorites, in range of 797°C-880°C for the microgranites, and in range of 704°C-809°C for the leucogranites. The temperature ranges increase from the leucogranite, granodiorites to microgranites (Table 5). Lower temperatures estimated by Toksoy- Köksal (2016) infer recrystallisation at a late stage during slow cooling. Also, the values calculated from the zircon and apatite saturation methods are reliable, reflecting crystallisation temperatures.

Discussion
Intrusive rocks in the Ekecikdağ area display different granitoid types, which evolved in a short period in the Late Cretaceous. U-Pb crystallisation ages of the granodiorites and the microgranites were determined as 84.52 ± 0.93 Ma and 80.7 ± 1.6 Ma, respectively, whereas the age of the leucogranites is suggested as 80 Ma relative age, based on the field relations and considering the 206 Pb/ 238 U and Rb-Sr ages. These granitoids not only show distinct ages but also differ from each other in their field, petrographical, mineralogical, whole-rock element and even isotopic geochemical characteristics. In spite of the variable petrological characteristics, they are combined to form an igneous association in a particular geodynamic setting. The petrological investigation of the EIA using new geochronological and isotopic data is crucially important as a contribution to the geodynamic evolution of Central Anatolia.
1999; Boztuğ, 1998Boztuğ, , 2000Düzgören-Aydın et al., 2001). The other view, on the other hand, states that the Inner Tauride ocean, separating the Tauride carbonate platform from the Anatolide block, was subducted beneath the CACC and Andean-type style magmatic arc granitoids were derived (Deniz & Kadıoğlu, 2016;Görür et al., 1984;Kadıoğlu et al., 2003;Okay et al., 1996;Şengör & Yılmaz, 1981). Kadıoğlu et al. (2003) interpreted the granitoids in the Ağaçören Intrusive Suite (AIS), which seems to be the northwest continuation of the EIA, as a product of a magmatic arc evolved at the active margin of the western edge of the CACC. Moreover, Kadıoğlu et al. (2006) determined the 40 Ar/ 39 Ar age of the AIS granitoids as 77.7 ± 0.3 Ma and suggested that partial melts were formed by the injection of metasomatized upper mantle melts into the continental crust. These partial melts result in the interaction of mantle-and crustal-derived magmas accompanied by the combined assimilation-fractional crystallisation, mixing and mingling processes and generated the calc-alkaline magmas as sources of these Figure 16. Geochemical plots of the granitoids on the tectonodiscrimination diagrams (a) Rb vs. Y+ Nb (syn-coln: syn-collisional granitoids, WPG: within plate granites, VAG: volcanic arc granites, ORG: orogenic granites, Pearce et al., 1984), (b) R2 vs. R1 (Batchelor & Bowden, 1985) and (c) Al 2 O 3 vs. SiO 2 (RRG+CEUG: rift related granitoids and continental epiorogenic uplift granitoids, POG: post-orogenic granitois, IAG+CAG+CCG: island arc granitoids, continental arc granitoids and continental collisional granite, Maniar & Piccoli, 1989), (d) Y/44-Rb/100-Nb/16 (Thiéblemont & Cabanis, 1990) (Symbols and explanation of the area defined by dashed line are given in Figure 10). granitoids (Kadıoğlu et al., 2006). Köksal et al. (2012), on the other hand, presented LA-ICP-MS U-Pb zircon age data yielding an age range from 84 to 74 Ma for the AIS granitoids and described them as collisional to post-collisional formed due to crustal thickening after obduction of the ophiolitic rocks onto the TAP. Therefore, there is no consensus on petrogenesis of the granitoids in the Ağaçören area. For that reason, the previous geochemical data from the AIS granitoids are compared with those of the EIA granitoids to better understand the nature of the granitoids within that part of the CACC. The whole-rock element geochemical data compiled from Kadıoğlu et al. (2003Kadıoğlu et al. ( ), (2006 and Köksal et al. (2012) for the granitoids from the Ağaçören area well compare with data of the EIA granitoids ( Figure 10-16). Moreover, the Sr-Nd isotope data from the Ağaçören granitoids ([ 87 Sr/ 86 Sr] T : 0.709143 to 0.715191; εNd: −6.48 to −8.84) (Köksal et al., 2012) are also comparable to those of the EIA granitoids. Tectonic discrimination diagrams of the EIA granitoids reveal their collisional to post-collisional character ( Figure 16). This is in accordance with the description of the other CACC granitoids with collisional to post-collisional nature (e.g. Boztuğ, 1998Boztuğ, , 2000Göncüoğlu & Türeli, 1994;Köksal et al., 2012Köksal et al., , 2013. Although all three rock types show sub-alkaline and high-K calc-alkaline features, their alumina indexes are different (Toksoy- Köksal, 2016). The mafic enclave-rich granodiorites have hornblende and plagioclase as essential phases and titanite as an accessory mineral. The granitoids are metaluminous to peraluminous, and reveal an I-type granitoid nature (granitoid formed by partial melting of metamagmatic crust, e.g. Chappell, White, & Wyborne, 1987). The microgranites and the leucogranites with two micas (biotite + primary muscovite), in contrast, are peraluminous and imply that these rocks are S-type granitoids not bearing hornblende (granitoids formed from sedimentary protoliths; e.g. Chappell & White, 1974). Both microgranites and leucogranites are peraluminous, whereas the leucogranites display lower ACNK values due to the presence of orthoclase and the albitic composition of plagioclase. EPMA analyses of biotites support the peraluminous character of the microgranites (Toksoy- Köksal, 2016).
SiO 2 contents of the granodiorites and the microgranites are similar in range, but lower than the leucogranites. The relation of the granodiorites and the microgranites to the leucogranites due to decreasing Al 2 O 3 , Fe 2 O 3 (T) , MgO, CaO contents but increasing SiO 2 may infer fractional crystallisation ( Figure 12). However, on the diagrams of TiO 2 against Fe 2 O 3 (T) / Al 2 O 3 , Fe 2 O 3 (T) , MgO and P 2 O 5 , there are no continuous trends observed through the granodiorites, the microgranites and the leucogranites representing a fractional crystallisation (Figure 13). Trends of Zr against SiO 2 and TiO 2 show that the leucogranites are separate from the granodiorites and the microgranites (Figure 12, 13). Moreover, the granodiorites and the microgranites, with similar Zr/Ti values, display markedly different patterns to the leucogranites on the Zr/Ti against Th, Nb, Nb/Th, Nb/Y, Zr/Y, La/Yb diagrams ( Figure 14). The overlapping plots on the diagrams of SiO 2 against Al 2 O 3 /Fe 2 O 3 (T) may imply that the granodiorites and the microgranites have a cogenetic origin than the leucogranites (Figure 12). The granodiorites and the microgranites have also overlapping values of Nb/Y, Zr/Y, La/Yb that may infer derivation/evolution from the same source and then fractionation ( Figure 14). However, the granodiorites with lower Nb and higher Th have a lower Nb/Th ratio that separates the granodiorites from the microgranite. Moreover, the plots of TiO 2 against Figure 17. Initial 143 Nd/ 144 Nd vs. 87 Sr/ 86 Sr diagram of the Ekecikdağ granitoids. BSE: Bulk Silicate Earth (Workman & Hart, 2005), EM I: Enriched Mantle I (Richardson, Erlank, Duncan, & Reid, 1982), Enriched Mantle II (Workman et al., 2004), DMM: Depleted MORB Mantle (Workman & Hart, 2005), ACC: Average Continental Crust (Hofmann, 1997), LC: Lower Crust (Villaseca, Downes, Pin, & Barbero, 1999), Oceanic Basalt Suites (Zindler & Hart, 1986).    give two almost parallel trends for the granodiorites and the microgranites, which opposes their cogenetic origin ( Figure 12). Moreover, no magmatic differentiation trends are observed from the granodiorites and the microgranites linked to the leucogranites on the SiO 2 versus Rb, Sr, Nb, Th, Zr and La variation diagrams ( Figure 12). Bivariate plots clearly display a good separation of the Ekecikdağ granitoids from each other that infers distinct genetic relations. Moreover, mineral-based petrological interpretations by Toksoy-Köksal (2016) strongly oppose a cogenetic relationships of the granitoids, especially as fractionation of the microgranite from the granodioritic magma. This assumption is supported by the crystallisation temperature of the microgranites calculated from zircon and apatite saturation formulations that is higher than the granodiorites. This means that the microgranites were not fractionated from the granodioritic melt. Moreover, there are some whole-rock elemental and isotopic inputs that oppose their cogenetic character. Although all the granitoids have well-characterised positive anomalies in Th, La, Ce, Nd, Sm, U, K, Pb and negative anomalies in Nb, Zr, Eu, Ti, flat MREE and HREE on chondrite and primitive mantle normalised multi-element diagrams (Figure 15), the microgranites are enriched in Ba, Sr, P but depleted in Th and U compared to the granodiorites, even though similar enrichment and depletion levels are observed for both rock types. Furthermore, the leucogranites differ from the granodiorites and the microgranites by pronounced enrichment of Th, U and HREE, and depletion of Sr, P, Eu and Ti and LREE ( Figure 15).
To clarify and understand genetic relationships of the rocks, the very incompatible elements (VICE: Th, U, Nb, Zr, Hf, Ti, Y) and moderately incompatible elements (MICE) are used in this study. Because VICE profoundly fractionate in melt residue but fractionate much less in the melt, they therefore preserve the same ratios as in the original source (Anderson, 2007;Hofmann, 2003), thus their ratios can reflect heterogeneity in the source. Moreover, use of moderately incompatible elements (MICE) relative to VICE may also reveal source heterogeneity (Hofmann, 2003;Hofmann & Jochum, 1996). In this study, VICE/MICE ratios (Nb/Yb) are plotted against Zr/Ti, Ti/Y, Nb/Th, La/Yb and ƐNd to throw light on source relations (Figure 19). The pronounced distinct plots on the Nb/Yb against Zr/Ti, Ti/Y and La/Yb and ƐNd diagrams infer that the source difference among the rock types are clear (Figure 19). The higher Zr/Ti and ƐNd values with lower Ti/Y and La/Yb of the leucogranites shows immediately that its source is different from both the granodiorites and the microgranites. The microgranites and the granodiorites have overlapping Zr/Ti, Ti/Y and La/Yb (Figure 19), while the ƐNd of the granodiorites is higher than the microgranites but lower than the leucogranites. Moreover, where VICE/ MICE ratios are plotted against 143 Nd/ 144 Nd ( Figure 19) the leucogranites have higher Zr/Ti but lower Ti/Y, La/ Yb, and low to moderate Nb/Th ratios against 143 Nd/ 144 Nd, which infer that they have a significantly different source from the granodiorites and the microgranites ( Figure 19). Nb/Th and ƐNd of the granodiorites overlap with the leucogranites but are different from the microgranites. Thus the microgranites might have a different source from the granodiorites. The pronounced negative Eu anomalies ((Eu/Eu*) N ) in the leucogranites (0.09-0.36), the microgranites (0.63-0.78) and the granodiorites (0.53-0.74) reflect fractional crystallisation of plagioclase. (La/Yb) N , (La/ Sm) N and (Gd/Yb) N ratios change from 8.19 to 25.78 (average = 13.89), 4.75 to 9.17 (average = 6.39), 1.15 to 1.51 (average = 1.32) for the granodiorites, 9.12 to 14.23 (average = 12.01), 3.87 to 4.19 (average = 4.01), 1.65 to 2.20 (average = 1.96) for the microgranites, 1.28 to 8.0 (average = 2.52), 1.95 to 4.68 (average = 3.09), 0.40 to 0.99 (average = 0.64) for the leucogranites, respectively. High (La/Yb) N ratios for the granodiorites and the microgranites infer high fractionation of REE and the presence of garnet (Hauri, Wagner, & Grove, 1994;Petermann, Hirschmann, Hametner, Gunther, & Schmidt, 2003) and/or hornblende (Reichardt & Weinberg, 2012;Tiepolo & Tribuzio, 2008) in the source. Thus, the magma source could be garnet lherzolite or a meta-magmatic crustal source with garnet and hornblende (Toksoy- Köksal, 2016).
Mineral based petrological evidence infers that the S-type microgranites and leucogranites were derived by partial melting of a meta-sedimentary crustal source ( Figure 18 in Toksoy- Köksal, 2016). In the origin of the I-type hybrid granodiorite, however, mixing of mantle and meta-magmatic and meta-sedimetary crust derived magmas might have played a role (Toksoy- Köksal, 2016). The I-type character of the granodiorites may reveal partial melting of a metamagmatic crustal source (e.g. Chappell et al., 1987) while the S-type features of the microgranites and the leucogranites may reflect partial melting of a meta-sedimentary crustal source (Chappell & White, 1974). However, evolution of all the data explains that it is too simple to classify the Ekecikdağ granitoids as Sand I-type (Toksoy- Köksal, 2016). Crustal contribution is signficant in the origin of the I-type granodiorites while a mantle contribution is present in the source of the mainly S-type leucogranites and microgranites. Therefore, heterogeneous sources derived from both mantle and crust (Chen, Xu, Chen, & Yu, 2016;Gray, 1984;Keay, Collins, & Mcculloch, 1997) are present in the protolith of the Ekecikdağ granitoids. Studies suggest mantle-derived magma in the origin of the S-type granitoids (e.g. Maas, Kamenetsky, Nicholls, & Steele, 2001), and reworking of crustal sedimentary material with mantle-like source (e.g. Kemp et al., 2007). A mixture of crustal rocks, including felsic-intermediate magmatic, metamorphic and sedimentary rocks may contribute to the source of the I-type granitic rocks (Chen et al., 2016). Hence a higher contribution from a mantle and/or meta-magmatic crustal source in origin of the Ekecikdağ granodiorites is possible. When these data collectively evaluated, the classical S-type and I-type classification does not seem to accurately reflect the petrogenesis of the Ekecikdağ granitoids. Because in the granodiorites with mainly Itype characteristics there is a significant crustal contribution, while in the S-type leucogranites and microgranites there is a mantle contribution. For this reason, as suggested by some authors like Gray (1984) and Keay et al. (1997) both S-, and I-type Ekecikdağ granitoids should have been derived from heterogenous source including mantle and crustal components. This is the case already inferred by some other studies disclosing S-type granitoids with mantle contribution (e.g. Maas et al., 2001) or I-type granitoids formed by reworking of sedimentary material (e.g. Kemp et al., 2007). Oxygen and sulphur isotope geochemistry displays a significant crustal source signature in the genesis of post-collisional granitoids in Central Anatolia (Boztuğ & Arehart, 2007). As a result, granitoids in the Ekecikdağ region can be assessed mainly as crustal granitoids having a mantle contribution.
Zircon Lu-Hf analyses reveal the robust petrogenetic nature of the host granitoids. The narrow range of initial 176 Hf/ 177 Hf ratios of Late Cretaceous analyses (i.e. 0.282500-0.282683) and εHf (t) values (i.e. −1.3 to −7.8) indicate no evidence for juvenile Cretaceous input. Moreover, these initial zircon Hfisotope data indicate that the Ekecikdağ granitoids were not derived from melting of juvenile crust formed from depleted mantle since εHf (t) values are below zero and so not typical of a mantle source (e.g. Amelin, Lee, Halliday, & Pidgeon, 1999). The Ağaçören granitoids with zircon εHf (t) data ranging from −4.1 to −8.8 (Köksal et al., 2012) are comparable with zircon εHf (t) data, which correspond to Late Cretaceous, of the Ekecikdağ granitoids. The unradiogenic Hf-isotope character of the zircons from the Ekecikdağ granitoids, along with the unradiogenic Neodymium whole-rock isotope data, away from the MORB and island arc volcanics sources, even below the typical deep sea sediment data, and close to the average sediment data, also infer a crust dominated source (Figure 18) (e.g. Chauvel, Lewin, Carpentier, Arndt, & Marini, 2008). Whole-rock Sr-Nd isotope data also support enriched sources with significant crustal contribution ( Figure 17). The larger range of initial 176 Hf/ 177 Hf ratios (i.e. 0.281706-0.282683) and εHf (t) values (i.e. −16.1 to +0.2) corresponding to older ages (Table 4), contrariwise, indicates heterogeneity of the older sources. Zircon Lu-Hf isotope data consequently indicate traces of reworked crust in the petrogenesis of the Ekecikdağ granitoids. This phenomenon is also suggested for the Ağaçören granitoids located to the NW of the study area (e.g. Köksal et al., 2012). Therefore, it is suggested that the Ağaçören Intrusive Suite comprising granitoids, which are contemporaneous and show similarities in geochemical data Figure 19. Bivariate plots of VICE/MICE ratios against 143 Nd/ 144 Nd and Nb/Y (symbols are given in Figure 10).
including whole-rock Sr-Nd and zircon Lu-Hf isotopes with the EIA granitoids, is likely to be have the same petrogenetical and geodynamical characteristics with the Ekecikdağ Intrusive Association.
As a result, this study reveals that the EIA granitoids were generated in a collisional to post-collisional setting with a pronounced crustal signature. These granitoids are suggested to be formed from partial melting of crustal dominated sources triggered by crustal thickening during the closure of the İzmir-Ankara-Erzincan branch of the Neotethyan Ocean by obduction of ophiolitic units on to the passive edge of the Tauride-Anatolide Platform.

Conclusions
The EIA granitoids, namely granodiorite, microgranite and leucogranite, show distinct field, petrographical, whole-rock element and isotopic geochemical characteristics. The main magmatic phase in the area is respresented by the granodiorites, which exhibit a mainly I-type character, with essential hornblende, biotite and plagioclase contents, and abundant mafic microgranular enclaves. The microgranites and the leucogranites, on the other hand, are S-type two-mica granitoids. The geochemical data collectively show that these three granitoid types are non-cogenetic.
Whole-rock geochemistry, including Sr-Nd isotope data, reveals that the S-type microgranites with high 87 Sr/ 86 Sr and low εNd have a crustal source with an insignificant mantle input, while the S-type leucogranites display higher mantle input than the microgranites. The I-type granodiorite, on the other hand, has higher 87 Sr/ 86 Sr and lower εNd with respect to typical I-type granitoids, thus indicating higher crustal contribution in its petrogenesis.
Mean LA-ICP-MS zircon 206 Pb/ 238 U ages for the Ekecikdağ granitoids are reported as 80 to 85 Ma. Moreover, LA-ICP-MS Lu-Hf isotope data obtained from the rims and younger zones of the zircons from the Ekecikdağ granitoids reveal their crustal nature (i.e. εHf (t) : −1.3 ± 0.5 to −8.8 ± 0.5).
The zircon and apatite saturation methods give the crystallisation temperatures for the EIA granitoids as 728°C-848°C for the granodiorites, 797°C-880°C for the microgranites and 704°C-809°C for the leucogranites.
As a result, it is suggested that the EIA granitoids were formed by partial melting of a crustal dominated igneous source. However, all the granitoids have a mantle input, which is the highest in granodiorites and the lowest in the microgranites. Understanding of the nature and evolution of the collisional granitoids in Central Anatolia is important not only to establish its contribution to the geodynamic evolution of Central Anatolia and surrounding Alpine area, but also to understand the systematics of collisional magmatic systems.