Geochemistry and Petrogenesis of Dioritic-Gabbroic Pegmatites in the Bulfat Complex, Qala Diza, Northeastern Iraq

The pegmatite dikes and associated plutonic rocks stand as a part of the igneous complexes associated with the Bulfat complex, located in the Zagros Suture Zone (ZSZ), NE Iraq. The Bulfat complex is a part of the ophiolite-bearing terranes that are allochthonous sheets. The complex represents the upper allochthon of the AlbianCenomenian. The study area is located within Bulfat complex of ZSZ, specifically in the northwestern part of this zone and within the second unit of the Penjween-Walash Subzone. The rock samples were collected from pegmatite dike which is located to the northeast of the Darishmana village, the thickness of dike is about 5 m. Electron microprobe analyses (EMPA) of plagioclase in 8 spots of dioritic pegmatites ranges between oligoclase (An18.00-An28.23) and andesine (An32.53An33.62). Bulk whole-rock chemical analysis of fourteen samples using ICP-MS analysis reveals alkalinity Index (AI) of pegmatites to be metaluminous (A/NK > 2). Generally, the silica content in these pegmatites is from 46.70 wt. % to 52.67 wt. %. The relatively flat pattern of REEs is characterized by the slight enrichment of LREEs compared to HREEs indicating the common ancestry of the studied pegmatites. Also, the enrichment of these pegmatites in LILEs (Sr, Pb, Rb) and depletion in HFSEs especially (Nb, Ta, Y) indicate the environment of the island arcs. Moreover, the low ratios of (Rb/Sr)N and (Ba/Sr)N refer to that these pegmatites are derived from a basic origin. Tectonic discriminate diagrams show that the tectonic environment of studied pegmatites is I-type, which is the oceanic island arcs environment of sub alkaline rocks. The pegmatites of the present study have a genesis relationship with intrusions close to them in the study area; these intrusions are gabbros of Wadi Rashid that represent the environment of E-MORB. Moreover, the gabbros of Wadi Rashid and studied pegmatites are part of ophiolite-bearing terranes, they are found within upper allochthon thrust sheet. The current study of pegmatites reflects the oceanic island arcs environment, this indicates the existence of double island arcs, the first adjacent to the Arabian shelf, and the second close to the middle of paleo-ridge. Numerous evidences support that the gabbros of Wadi Rashid being as the likely parent to the studied pegmatites such as geochemistry, tectonogenesis, and the close spatial distribution of the pegmatites to the gabbros of Wadi Rashid. Moreover, the Shareef T. Al-Hamed et al., 65 studied pegmatites appear to entail further dissection mainly due to the fact that the occurrence of dioriticand gabbroic-pegmatites with a small-scale in the single intrusion might have its explanation in the liquid associated immiscibility.

studied pegmatites appear to entail further dissection mainly due to the fact that the occurrence of dioritic-and gabbroic-pegmatites with a small-scale in the single intrusion might have its explanation in the liquid associated immiscibility.

INTRODUCTION
The pegmatite dikes and associated plutonic rocks stand as a part of the igneous complexes associated with the Bulfat complex, located in the Zagros Suture Zone (ZSZ), NE Iraq (Jassim et al.,1982a) (Fig. 1). ZSZ occupies an area about 5000 km2 along Iraq-Iran-Turkey borders and within the second unit of the Penjween-Walash Subzone (Buday and Jassim, 1987) (Fig. 1). The Bulfat complex is a part of the ophiolite-bearing terranes that are allochthonous sheets and the complex represents the upper allochthon of the Albian-Cenomenian (Aswad and Elias, 1988;Aswad, 1999), while the Walash-Naopurdan sequence represents the lower allochthon of the Palaeogene (Aswad et al., 2011). Allochthonous sheets are located over the autochthonous deposits represented by the sediments of the Arabian shelf and Tertiary sedimentary rocks (Aswad, 1999) (Fig.  1).
The Iraqi Bulfat complex (100 km 2 in extents in Iraq) outcrops in the mountain of Bulfat near Qala Diza. It comprises a volcano-sedimentary unit as the Gimo and Sirginil groups, these groups were originally referred to the Bulfat group (Jassim et al., 1982a) (Fig. 1). The Gimo group at Bulfat comprises regionally metamorphosed basalt, diabase, andesite, acidic volcanic, tuffs, calcareous, and argillaceous sediments (Jassim et al., 1982b), as a result of the igneous intrusive, aerosols of hornfels and marble were formed that overprinted the regional metamorphic rocks (Jassim et al., 2006a). Whereas the Sirginil group is located in the northern part of the Bulfat complex, this group consists of pelitic and arenaceous rocks with volcanic flows that have undergone a regional and contact metamorphism (Buday and Suk, 1978), forming the upper part of the Bulfat group (Jassim et al., 2006b) (Fig. 1).
The Bulfat complex is generally composed of plutonic rocks (basic, ultrabasic, intermediate, and acidic-granite), these rocks are mainly composed of gabbro-diorite intrusions and accompanied by syenite and nepheline syenite resulting from late-stage magmatic differentiation (Jassim et al., 2006a), forming many rocks intruded during an early Tertiary (Paleocene-Eocene) (Jassim et al., 2006a). The complex is intruded by minor intrusions and pegmatites (Jassim et al., 2006a). Field investigations of the accessible areas at Darishmana appear to reveal that these pegmatite bodies have intruded the Albian-Cenomenian metasedimentary sequences (Gimo-Qandil formation) (Aswad and Elias, 1988;Aswad, 1999), which experienced medium-grade regional metamorphism, as overprinted by a high-grade contact metamorphism during Paleogene (Jassim et al., 1982a). Peralkaline magma is produced from the assimilation of the calcareous and pelitic host rocks, these results in the production of pegmatites and nepheline syenite (Buda, 1993). The pegmatites are course-grained dikes cutting both the intrusive and surrounding rocks (Jassim et al., 2006a).
The basic intrusions in the Bulfat group have caused contact metamorphism of the surrounding country rocks (Jassim et al., 2006a). The pyroxene-hornblende gabbros are mostly developed in the northeastern part of the complex, forming the high ground of the Bulfat range along Iraqi-Iranian border, and form a large part of the complex (Jassim et al., 2006a) (Fig. 1), while the olivine gabbro rocks are spread in the form of rock bodies containing varying amounts of olivine (Jassim et al., 2006a).
The complex near the village of Pauza (Shaban valley) contains ultramafic rocks represented by peridotite affected by the serpentinization process (Fig. 1). The ultramafic rocks can be divided into three occurrences (Buday and Jassim, 1987): serpentinites, pyroxenites, and Pauza ultramafic body which form 4% of the Bulfat igneous complex, representing the major ultramafic occurrence and exhibiting tectonic emplacement features, tectonic breccia along its western contact.
The study area is located in the northeast of Iraq, specifically 74 km to northeast of Sulaimaniya city. It is located about 13.5 Km to east of Qala Diza subdistric and approximately 4 Km of Darishmana village. The study area extends between longitudes (45°: 13': 27.944") (45°: 19': 43.346")E and latitudes (36°: 14': 43.442") (36°: 10': 28.447")N close to the Iranian border. Geologically, the study area is located within Bulfat complex of ZSZ, specifically in the northwestern part of this zone and within the second unit of the Penjween-Walash Subzone (Buday and Jassim, 1987). The rock samples were collected from pegmatite dike located to the northeast of the Darishmana village ( Fig. 1), whose thickness is about 5 m. The sampling started from the first sample (36º: 11′: 23.3″N 45º: 17′: 6.9″E) and ended with the last sample (36º: 11′: 24.2″N 45º: 17′: 24.6″E) where the elevation from sea level is reached about 1793 m. This paper focuses on pegmatites in the Bulfat complex, throwing light on the variation in the concentrations of the major, trace, and rare earth elements to clarify the petrogenesis, tectonic setting of pegmatites.  (Buday and Jassim, 1987).

METHODOLOGY
Pegmatite as a very coars-grained intrusive igneous rock, whose interlocking grains are usually larger than 2.5 cm in size, any attempt to classify these rocks by means of thin section and chemical classification appears to be unscientific. such methods are discovered to be non-representative of bulk minerals, i.e. the classification zone should be markedly larger than the thin section to calculate the mode analysis (Al-Hamed et al., 2019). To overcome this problem, the staining and digital image processing were used (Al-Hamed et al., 2019). The polished slabs were prepared in the Department of Geology, University of Mosul, Iraq, and stained. The Images have taken for the polished and the stained surfaces by a digital camera. By the ENVI software, they were changed into various image styles to calculate and recalculate the color index M′, quartz, alkali feldspar, and plagioclase proportions (Al-Hamed et al., 2019).
Electron microprobe analyses (EMPA) were carried out at 8 spots to determine the anorthitic component (An mol. %) in D7B and D7C samples utilizing a fully automated, CAMECA SX50 Electron Microprobe at Utah University, USA, fitted with 4 wavelength-dispersive spectrometers, accelerating voltage 15 KeV, beam current 30 nA, and spot size 10μm.
Major, trace, and rare earth elements analysis of fourteen samples were determined by Perkin Elmer Sciex-Elan 6000 (ICP-MS) with a 4-acid digestion at ACME Analytical Laboratories in Canada (Table 1).

RESULTS
Based on the staining and digital image processing that provided by Al-Hamed et al (2019), the classification of the studied pegmatites was melano-diorite and gabbro ( Table 2). The extinction angle of plagioclase was calculated to find out An mol. % in the plagioclase using the Michale Levy method to differentiate between melano-diorite and gabbro (Table 3). The plagioclase composition was andesine in D7B and D7C samples and M' > 50%, indicating that these samples are melanocratic. Moreover, electron microprobe analyses (EMPA) of plagioclase in D7B and D7C samples are shown in Table 4. The anorthitic component (An mol. %) ranges between oligoclase (An18.00-An28.23) and andesine (An32.53-An33.62) for D7B and D7C samples respectively. The difference in plagioclase compositions in the D7B and D7C samples from the other studied samples may be related to the magmatic fractionation, revealing more sodic plagioclase with progressive magma differentiation (Shawna et al., 2003

Geochemistry of Pegmatites:
In order to clarify the geochemical variation of the major oxides and trace elements, the binary variation diagrams using MgO and FeOtotal were used against the major oxides and trace elements, because MgO and FeOtotal are concentrated in the mafic minerals are reflected by the variation in the color index (M '%) which range between (50-63.7 %) ( Table 2), as well as that MgO and FeOtotal are increased in these rocks.

Major Oxides:
Alkalinity Index (AI) refers to that the studied pegmatites are metaluminous (A/NK > 2) (Fig. 2). Generally, the silica range in these pegmatites is from 46.70 wt. % to 52.67 wt. % (Table 1), showing a dispersed correlation with MgO and a negative correlation with FeOtotal (Fig. 3A, B). The silica correlation with FeOtotal reflects the magmatic fractionation of plagioclase, where plagioclase becomes more sodic with progressive magmatic fractionation (Shawna et al., 2003).
Al2O3 and Na2O show two continuous trends with MgO (Fig. 3C, E), the first reflects an increase in Al2O3 and Na2O with increasing MgO, indicating the start of plagioclase fractionation; whereas the second trend shows a decrease in these oxides with increasing MgO, indicating fractionation of clinopyroxene. CaO has also two trends with MgO (Fig. 3G), the first represents an increase in CaO with decreasing MgO, indicating fractionation of plagioclase and depletion of Mg in clinopyroxene with stability of Ca content (Sofy, 2003); whereas the second trend shows an increase in CaO with increasing MgO, reflecting fractionation of clinopyroxene. The correlations of Al2O3, Na2O, and CaO with MgO (Fig. 3C, E, and G) may reflect liquid immiscibility in the single intrusion. Al2O3, Na2O, and CaO show a dispersed correlation with FeOtotal ( Fig. 3D, F, H). TiO2 and MnO have a dispersed correlation with MgO and a positive correlation with FeOtotal (Fig. 3V, X, Y, Z), this reflects that these oxides occur in the Fe-phase minerals as pyrrhotite.

Trace Elements:
Ni shows a positive correlation with MgO (r MgO/Ni = 0.82) and a dispersed correlation with FeOtotal (Fig. 4A, B), this reflects that Ni 2+ replaces Mg 2+ in clinopyroxene (augite) (Rollinson, 1993). Also Sc and V show a dispersed correlation with FeOtotal, indicating the scarcity of their presence in accessory Fe-rich phases (Fig.  4D, F). Whereas, Sc and V have a strong positive correlation with MgO (r MgO/Sc = 0.91 and r MgO/V = 0.83) (Fig. 4C, E) which confirms their presence in one phase (augite) (Mason and Moore, 1982;Wodepohl, 1978). Co has a dispersed relationship with MgO and is positively associated with FeOtotal (r FeO total/Co = 0.68) (Fig. 4G, H), this reflects that Co increases in accessory Fe-rich phases, especially ferric, where Co has a positive correlation with ferric (r = 0.7).
LILEs (Rb, Ba, Sr) show a negative correlation with MgO and a dispersed correlation with FeOtotal (Fig. 5A, B, C, D, E, F), indicating that Rb, Ba, and Sr are associated with felsic minerals (Mason and Moore, 1982) since the felsic minerals in these rocks are dominated by plagioclase, so this mineral is the host of Rb, Ba, and Sr. Cs and Li do not show any clear correlation with MgO and FeOtotal (Fig. 5G, H, V, X), Cs is highly incompatible due to the very large ionic radius but is associated with feldspars and micas (Linnen et al., 2012), while Li has a much smaller ionic radius than that of other alkali metals and is associated with amphiboles and micas by coupled substitution reactions (London, 2005a). Generally, the melts crystallize under equilibrium conditions, whereas pegmatitic melt crystallizes under super-cooled conditions far from the equilibrium (Chakoumakos and Lumpkin, 1990;Morgan and London, 1999;Webber et al., 1999). Therefore, the difference occurs in the fractionation and behaviour of LILEs from what is expected in other melts (London, 2005b).
HFSEs have a medium negative correlation with MgO and a positive with FeOtotal ( Fig. 6A, B, C, D, E, F), indicating their presence in Fe-rich minerals. Zr, Y, and Hf elements show a strong positive correlation with each other (Fig. 7), reflecting their common existence where be formed from a single source of mafic magma by fractional crystallization (Wilson, 1989). Nb and Ta have an unclear correlation with MgO and FeOtotal (Fig. 6G, H, V, X). Th and U show a dispersed correlation with MgO and FeOtotal (Fig. 8), indicating that they are associated with accessory minerals.
Chalcophile elements (Sn, Zn, Cu) show divergent correlations with MgO and FeOtotal (Fig. 8). Sn and Cu show a dispersed correlation with MgO and FeOtotal (Fig.  9A, B, E, F). Zn exhibits a negative correlation with MgO and a strong positive correlation with FeOtotal (r = 0.9) (Fig. 9C, D), this indicates that Zn entered Fe-rich phases, it replaces ferrous due to the similarity of charge and radius (Zn 2+ 0.74Å, Fe 2+ 0.74Å).

Rare Earth Elements (REEs):
The pattern diagram of REEs after normalization with chondrite (Sun and McDonough, 1989) is accomplished (Table 5 and Fig.10).
The ratio (La/Sm)N shows enrichment and depletion in LREE, this ratio ranges between (0.85-2.03) ( Table 7). It is noted that this ratio is greater than one (except D3A sample) (Table 7), which reflects the enrichment of these rocks in the LREEs content.
The ratio (Gd/Yb)N indicates the variation in HREEs content, this ratio ranges between (1.07-1.78) ( Table 7) and these values reflect a limited variance, which explains the homogeneity and flatness in the HREEs.
In general, the ratio (La/Yb)N shows the variation in the REEs behavior, as it indicates enrichment or depletion in LREEs relative to HREEs content, this ratio extends between (1.34-3.80) (Table 7) reflecting the enrichment of these rocks with LREEs compared to HREEs, whereas the ratio (La/Yb)N is consistent with the ratio (La/Sm)N.
Moreover, an anomaly in the behavior of (Eu) element is observed in the REEs pattern, which can be illustrated by using the ratio (EuN / Eu*) through the equation EuN/ Eu * = (Eu)N / (SmN* GdN) 0.5 . Most samples show a small negative anomaly, indicating the fractionation of plagioclase (Zhang et al., 2003;Van Wagoner et al., 2002), while some samples show a positive anomaly, reflecting the accumulation of plagioclase as a result of the magmatic processes (Ottonello et al., 1984;Trubelja et al., 1995) (Table 7).
A negative correlation is observed between Mg# and ΣREEs, where the REEs increase with the decrease of Mg# and this is due to the fractionation process of the mafic minerals (Fig. 10).  Fig. 10: Chondrite-normalized REEs diagram of pegmatites, data from (Sun and McDonough, 1989).  The spider diagram is accomplished after normalization with primitive mantle (Sun and McDonough, 1989) (Table 8).
The spider diagram of the studied pegmatites (Fig. 11) shows a positive anomaly in Sr, Pb, and Rb elements in most samples relative to neighboring REEs (Table 9), with a noticeable flatness in the pattern from Zr to Sm. While Y, Ta, and Nb elements show a negative anomaly in most samples relative to adjacent REEs (Table 9), this anomaly reflects the island arcs environment (Aswad et al., 2013).
Generally, the ratios of (Rb/Sr)N and (Ba/Sr)N are low in these rocks (Table 9), which suggests that these pegmatites are derived from a basic origin, whereas the high ratios of (Rb/Sr)N and (Ba/Sr)N refer to a non-basic origin (Rogers and Greenberg, 1990). Table 8: The concentrations of the elements in the pegmatites normalized with the same concentrations of primitive mantle elements (P.M) from (Sun and McDonough, 1989 Fig. 11: Spider diagram of the elements in the pegmatites after normalization with the same concentrations of primitive mantle elements (P.M) from (Sun and McDonough, 1989). Table 9: The ratios of the elements in the pegmatites normalization with the same concentrations of primitive mantle elements from (Sun and McDonough, 1989).

DISCUTION AND CONCLUSIONS
During the last three decades, many researchers have used various geochemical methods to isolate tectonic environments from each other depending on binary and ternary variation diagrams known as tectonic discriminate diagrams, by relying on some of the major, trace, and rare earth elements. Figure (12A) shows that the studied pegmatites are of igneous origin (I-type). The SiO2-K2O diagram explains the evolution of the volcanic arc from tholeiite to shoshonite, this diagram shows the presence of these pegmatites in the calc-alkaline field (Fig. 12B), which are sub alkaline rocks (Fig. 12C). The studied pegmatites in the mentioned diagrams are consistent with the intrusive rocks close to the study area that are gabbros of Wadi Rashid studied by Aswad et al (2013).
Using (Ta/Yb) against (Th/Yb) of (Pearce, 1982) to sort between volcanic arc basalts (VAB), mid oceanic ridge basalts (MORB), and within plate basalts (WPB), the studied pegmatites veer from MORB field and occur in VAB field of calc-alkaline type (CA) (Fig. 12D). The tectonic discriminate diagrams (Hf-Th-Ta) and (Nb-Zr-Y) (Fig.  13A, B) show most pegmatites are located in the volcanic arc basalts field. Moreover, Fig.(13) shows some samples are located in MORB; this indicates that these rocks were erupted in an extensional stress regime (Aswad et al., 2013). Nevertheless, Fig. 13 emphasizes the volcanic arc environment of studied pegmatites. The SiO2-K2O/Na2O diagram (Fig. 14) shows that the environment of studied pegmatites is oceanic island arcs. Fig. 12: Tectonic environment of the studied pegmatites. A: The molar relationship between A/NK and A/CNK for sorting between the type of alumina saturation from (Maniar and Piccoli, 1989) and type of rocks from (Chappell and White, 1974). B and C: SiO2 vs. K2O diagram to distinguish the type of pegmatites from (Peccerillo and Taylor, 1976) and (Middlemost, 1975) respectively. D: Ta/Yb vs. Th/Yb diagram to distinguish tectonic of the studied pegmatites from (Pearce, 1982). In comparison with gabbros of Wadi Rashid from (Aswad et al., 2013). Fig. 13: Tectonic discriminate ternary diagrams of the studied pegmatites. A: from (Wood, 1980). B: from (Pearce, 1982). In comparison with gabbros of Wadi Rashid from (Aswad et al., 2013).
The relatively flat pattern of REEs is characterized by the enrichment of LREEs compared to HREEs (Fig. 10) indicating the common ancestry origin of the studied pegmatites. Also, the enrichment in LILEs (Sr, Pb, Rb) and depletion in HFSEs especially (Nb, Ta, Y) (Fig. 11) indicate the environment of the island arcs. The enrichment of LILEs is due to their migration by aqueous solutions derived from the subducted oceanic slab to the mantle wedge above it by dehydration, and an indication of the added fluids from the subducted slab to the mantle wedge (Breeding et al., 2004), whereas the HFSEs are immobile through these solutions and remain in the subducted oceanic slab (Breeding et al., 2004). In addition, the pegmatitic melt crystallizes far from equilibrium (Chakoumakos and Lumpkin, 1990;Morgan and London, 1999;Webber et al., 1999). Therefore, the difference occurs in the fractionation and behavior of LILEs from what is expected in other melts (London, 2005b). Moreover, the low ratios of