Geochemistry, Geochronology, and Hf-S-Pb Isotopes of the Akechukesai IV Maﬁc-Ultramaﬁc Complex, Western China

: The Akechukesai IV maﬁc–ultramaﬁc complex, located in the western segment of the eastern Kunlun orogenic belt (EKOB), represents a newly-discovered complex, containing Ni ores at grades of up to 0.98% Ni. It is dominated by olivine pyroxenite, pyroxenite, and gabbro units. The gabbros are enriched in lithophile elements (e.g., Rb, U, and K) and light rare-earth elements (LREE), with negative anomalies in high ﬁeld-strength elements, except Zr, Ta. Nb / Ta( ∼ 5) and Zr / Hf ( ∼ 10) ratios lower than the primitive mantle and chondrites, respectively, indicate the inﬂuence of the mantle metasomatic process or fractionation of accessory mineral phases. Zircon U–Pb dating of the gabbro yielded an age of 423.9 ± 2.6 Ma, indicating that the complex formed contemporaneously with the Xiarihamu Ni deposit (423 ± 1 Ma). The gabbro has negative ε Hf(t) values ( − 11.3 to − 1.2) with corresponding T DM1 ages of 1535–1092 Ma. The vein-like and disseminated mineralization (i.e., pyrite and pyrrhotite) have δ 34 S values of 13.1% (cid:24) –13.4% (cid:24) and 5.0% (cid:24) –8.5% (cid:24) , respectively, suggesting that the magmas that formed the complex assimilated crustal sulfur. They yield 206 Pb / 204 Pb, 207 Pb / 204 Pb, and 208 Pb / 204 Pb values of 17.323–18.472, 15.422–15.626, and 37.610–38.327, respectively, indicating Pb derived from multiple sources (i.e., mantle crustal sources). Geochemical and Hf–S–Pb isotopic characteristics suggest that the complex formed from a primitive magma derived by partial melting of a spinel- and garnet-bearing lherzolite mantle at variable degree of 5%–10%. This source region was geochemically enriched by previous interaction with slab-related ﬂuids. Tectonic reconstruction suggests that the Akechukesai IV complex was generated in a post-collisional extensional environment.


Introduction
The eastern Kunlun orogenic belt (EKOB) is located along the northern margin of the Tibetan Plateau in western China [1] (Figure 1a): Previous researches determined that the orogenic events are associated with the subduction of the proto-and paleo-Tethyan oceans [4,5]. This Silurian-Devonian tectonism also generated widespread magmatic rocks, including granites of slightly variable ages and minor amounts of mafic-ultramafic intrusions [6][7][8][9]. The origin of the latter remain poorly investigated despite the fact that some of them are mineralized. For instance, Xiarihamu intrusion contains 1.07 million tons of Ni reserves [7].
The Akechukesai Ⅳ mafic-ultramafic complex was first discovered by researchers from the Jilin University during a geological mapping in the western segment of the EKOB in 2015 (Figure 1b). It is similar to the previously discovered AkechukesaiⅠand Xiarihamu mafic-ultramafic complexes in terms of petrological features and metallic mineral assemblage [unpublished data]. The Xiarihamu deposit, one of the largest and best-known magmatic Ni-Cu deposits in this region [6,10,11], was the first Ni-Cu deposit discovered in the EKOB, and formed between late Silurian and Early Devonian (424-406 Ma) [6,7,11]. The discovery of this deposit stimulated more interest in evaluating the potential of the Cu-Ni mineralization in this region. One of the successful activities is the discovery of four mineralized mafic-ultramafic intrusions in the Akechukesai district. On the basis of the analysis of experimental results [unpublished data], these complexes, including the recently -discovered complex IV, could potentially host significant amounts of sulfide ores. Here, we present new geochronological, geochemical, and Hf-S-Pb isotopic data in order to understand its genesis. It provides an excellent example of regional mineralization. The results can contribute to the understanding of mafic-ultramafic intrusion related Cu-Ni deposit.

Geological Setting
The EKOB is an important area of the Tibetan Plateau, testifying to an extensive magmatic activity in the western China, and covers an area of ~1000 km 2 (Figure 1b). It contains three major fault zones, namely the North and Middle Kunlun faults and the South Kunlun-Anyemaqen fault. These three fault zones tectonically split the EKOB into three sub-tectonic units, namely the North Kunlun belt, the Middle Kunlun belt, and the South Kunlun belt [12].
The North Kunlun belt is dominated by the Paleoproterozoic-Mesoproterozoic Jinshuikou Group gneiss and the Ordovician-Silurian Tanjianshan Group marble [12,13]. It also contains voluminous high-K calc-alkaline granites [12]. The Middle Kunlun belt is an extensively exhumed belt that contains granitoid units, the Jinshuikou Group as well as small amounts of mafic-ultramafic intrusions. The South Kunlun belt is well-preserved and is dominated by Proterozoic to Triassic units Previous researches determined that the orogenic events are associated with the subduction of the proto-and paleo-Tethyan oceans [4,5]. This Silurian-Devonian tectonism also generated widespread magmatic rocks, including granites of slightly variable ages and minor amounts of mafic-ultramafic intrusions [6][7][8][9]. The origin of the latter remain poorly investigated despite the fact that some of them are mineralized. For instance, Xiarihamu intrusion contains 1.07 million tons of Ni reserves [7].
The Akechukesai IV mafic-ultramafic complex was first discovered by researchers from the Jilin University during a geological mapping in the western segment of the EKOB in 2015 (Figure 1b). It is similar to the previously discovered Akechukesai I and Xiarihamu mafic-ultramafic complexes in terms of petrological features and metallic mineral assemblage [unpublished data]. The Xiarihamu deposit, one of the largest and best-known magmatic Ni-Cu deposits in this region [6,10,11], was the first Ni-Cu deposit discovered in the EKOB, and formed between late Silurian and Early Devonian (424-406 Ma) [6,7,11]. The discovery of this deposit stimulated more interest in evaluating the potential of the Cu-Ni mineralization in this region. One of the successful activities is the discovery of four mineralized mafic-ultramafic intrusions in the Akechukesai district. On the basis of the analysis of experimental results [unpublished data], these complexes, including the recently -discovered complex IV, could potentially host significant amounts of sulfide ores. Here, we present new geochronological, geochemical, and Hf-S-Pb isotopic data in order to understand its genesis. It provides an excellent example of regional mineralization. The results can contribute to the understanding of mafic-ultramafic intrusion related Cu-Ni deposit.

Geological Setting
The EKOB is an important area of the Tibetan Plateau, testifying to an extensive magmatic activity in the western China, and covers an area of~1000 km 2 (Figure 1b). It contains three major fault zones, namely the North and Middle Kunlun faults and the South Kunlun-Anyemaqen fault. These three fault zones tectonically split the EKOB into three sub-tectonic units, namely the North Kunlun belt, the Middle Kunlun belt, and the South Kunlun belt [12].
The North Kunlun belt is dominated by the Paleoproterozoic-Mesoproterozoic Jinshuikou Group gneiss and the Ordovician-Silurian Tanjianshan Group marble [12,13]. It also contains voluminous high-K calc-alkaline granites [12]. The Middle Kunlun belt is an extensively exhumed belt that contains granitoid units, the Jinshuikou Group as well as small amounts of mafic-ultramafic intrusions. The South Kunlun belt is well-preserved and is dominated by Proterozoic to Triassic units as well as as well as rare Paleozoic and Mesozoic granitic rocks. These units include the Mesoproterozoic Wanbaogou Group, thought to represent part of an oceanic basalt plateau [12].
The Akechukesai Ⅳ mafic-ultramafic complex is located in the Middle Kunlun belt adjacent to the North Kunlun fault (Figure 1b). Unlike the Xiarihamu and Shitoukengde deposits, the area that hosts the complex contains numerous outcrops of the Tanjianshan Group, whereas the Jinshuikou Group outcrops only within the northeastern part of the district. The structure is dominated by NW-SE trending faults, especially in the northeast of the region. The intrusive units in this area are dominated by four mafic-ultramafic complexes in the form of small intrusive bodies.

Geology of the Akechukesai Ⅳ Complex
The Akechukesai mafic-ultramafic complex is divided into four complexes numbered Ⅰ, Ⅱ, Ⅲ , and Ⅳ (Figure 2), all hosting Ni mineralization. The Akechukesai I mafic-ultramafic complex was the first to be discovered, and zircon U-Pb dating suggests it was formed at 422 ± 10 Ma (unpublished data). This research focuses on the Akechukesai Ⅳ mafic-ultramafic complex, which outcrops over an area of approximately 0.1 km 2 and is hosted by marbles of the Tanjianshan Group. It contains olivine pyroxenite, pyroxenite and gabbro units in an increasing order of volume. Recent explorations (ZK17-43 and ZK17-48 drillholes) intercepted two orebodies with thickness of 51m and 57m and average grades of 0.98% and 0.97% Ni, respectively. Other drillholes in the area also revealed Ni ores. The Ni mineralization identified to date is primarily hosted by pyroxenite units, although Cu of cutoff grades were also identified within several drillholes in pyroxenites. Overall, the Akechukesai Ⅳ appears to be the most economically prospective complex in this region, as a result of the high Ni grades.
The Akechukesai Ⅳ has undergone post-magmatic chlorite hydrothermal alteration. Sulphides within the complex are dominated by pentlandite and pyrrhotite with minor chalcopyrite. The ore minerals in the complex have idiomorphic-hypidiomorphic and inequigranular textures (Figure 3a, b). The sulphides are hosted by vein-like structures and are also disseminated in host rocks; crustal contamination-related evidence of interaction between the magmas that formed the contaminated rocks and the surrounding marble is ubiquitous in the field and cores (Figure 3c-f). Gabbro samples were used for geochemical, geochronological, and Hf isotopic analysis during this study. These gabbros are charcoal grey, massive, medium-to fine-grained, and contain plagioclase (~50% in volume), clinopyroxene (~25% in volume), amphibole (~10% in volume), and biotite (~10% in volume) Recent explorations (ZK17-43 and ZK17-48 drillholes) intercepted two orebodies with thickness of 51m and 57m and average grades of 0.98% and 0.97% Ni, respectively. Other drillholes in the area also revealed Ni ores. The Ni mineralization identified to date is primarily hosted by pyroxenite units, although Cu of cutoff grades were also identified within several drillholes in pyroxenites. Overall, the Akechukesai IV appears to be the most economically prospective complex in this region, as a result of the high Ni grades.
The Akechukesai IV has undergone post-magmatic chlorite hydrothermal alteration. Sulphides within the complex are dominated by pentlandite and pyrrhotite with minor chalcopyrite. The ore minerals in the complex have idiomorphic-hypidiomorphic and inequigranular textures (Figure 3a,b). The sulphides are hosted by vein-like structures and are also disseminated in host rocks; crustal contamination-related evidence of interaction between the magmas that formed the contaminated rocks and the surrounding marble is ubiquitous in the field and cores (Figure 3c-f). Gabbro samples were used for geochemical, geochronological, and Hf isotopic analysis during this study. These gabbros are charcoal grey, massive, medium-to fine-grained, and contain plagioclase (~50% in volume), clinopyroxene (~25% in volume), amphibole (~10% in volume), and biotite (~10% in volume) with minor amounts of chlorite (~5% in volume). The plagioclase commonly has polysynthetic twins and the pyroxene is partially altered by chlorite (Figure 4a-d).

Analytical Methods
For gabbro whole-rock samples, altered surfaces were removed before being crushed to ~200 mesh. X-ray fluorescence (XRF; pw1401/10) equipped with ICP-MS (Agilent 7500a) and fused-glass disks were utilized to determine the major and trace element compositions (including Fe2O3T content), after digestion of samples in Teflon bombs by HF-HNO3. The analysis was performed at the testing center of Jilin University, Changchun, China. Standard samples for analysis were AGV-1 with minor amounts of chlorite (~5% in volume). The plagioclase commonly has polysynthetic twins and the pyroxene is partially altered by chlorite (Figure 4a-d).

Analytical Methods
For gabbro whole-rock samples, altered surfaces were removed before being crushed to ~200 mesh. X-ray fluorescence (XRF; pw1401/10) equipped with ICP-MS (Agilent 7500a) and fused-glass disks were utilized to determine the major and trace element compositions (including Fe2O3T content), after digestion of samples in Teflon bombs by HF-HNO3. The analysis was performed at the testing center of Jilin University, Changchun, China. Standard samples for analysis were AGV-1

Analytical Methods
For gabbro whole-rock samples, altered surfaces were removed before being crushed to~200 mesh. X-ray fluorescence (XRF; pw1401/10) equipped with ICP-MS (Agilent 7500a) and fused-glass disks were utilized to determine the major and trace element compositions (including Fe 2 O 3 T content), after digestion of samples in Teflon bombs by HF-HNO 3 . The analysis was performed at the testing center of Jilin University, Changchun, China. Standard samples for analysis were AGV-1 (andesite), BCR-2 (basalt), and BHVO-1 (basalt), which indicated that the precision of whole-rock samples was better than 5%. For the determination of FeO content, we used the Potassium dichromate volumetric method. This method follows the standards of GB/T 14505 and GB/T 14506.1 of the People's Republic of China.
The separation process of zircons from gabbros used for U-Pb dating conforms to standard at the regional geology survey, Langfang city, China. Then, transmitted, reflected light and cathodoluminescence (CL) images were obtained to uncover their internal characteristics.
U-Pb dating was performed at the Yanduzhongshi testing company, Beijing, China. The experimental apparatus used was an Aurora M90 ICP-MS interfaced with a New Wave UP213 laser ablation system. Helium and argon served as the carrier and make-up gas, respectively. They needed to be mixed evenly before entering the ICP. The spot diameter for analysis was 30 µm. In order to standardize the instrument, zircon 91,500 and Nist 610 were selected. The detailed procedure and method for common Pb correction are from Yuan et al. [14] and Andersen [15]. Isotopic data and age calculations were performed by ICP-MS, following Liu et al. [16] and Ludwig [17].
In situ zircon Hf isotope analyses were implemented using a Neptune multi-collector ICP-MS equipped with a NewWave UP213 laser-ablation system, at the Chinese Academy of Geological Sciences, Beijing, China. Hf isotopic measurements were done on the same spots previously analyzed for U-Pb with laser spot 30 µm in diameter, repetition rates of 8 Hz, laser beam energy density of 16 J/cm 2 , and an ablation time of 31s. Zircons 91,500 and GJ-1 were used to optimize the analytical instruments. The present-day chondritic ratios of 176 Hf/ 177 Hf = 0.282772 and 176 Lu/ 177 Hf = 0.0332 were adopted to calculate εHf(t) values [18]. In the calculation of Hf model ages, 176 Lu/ 177 Hf and 1 76 Hf/ 177 Hf were 0.282772 and 0.28325, respectively [18]. Wu et al. [19] and Guo et al. [20] have previously discussed the laboratory apparatus, testing process, and data capturing process.
Nine sulphides (i.e., pyrite and pyrrhotite) were obtained from the vein-like and disseminated ores for S and Pb isotope analyses, which were done at the test center of Beijing Research Institute of Uranium Geology. A Finnigan MAT 251 mass spectrometer was utilized to analyze the S isotopic compositions and uncertainty was ±0.2% for δ 34 S V-CDT . The reported data are relative to Vienna Canon Diablo Troilite (V-CDT) sulfide. As for Pb-isotopic compositions of sulfides, 50 mg of powder was dissolved in a mixed solution of hydrofluoric and perchloric acid, followed by elution through anion exchange resin, to separate Pb. The Pb isotopic compositions were measured on an ISOPROBE-T Thermal Ionization Mass Spectrometer with the standard NBS SRM 981. The precisions for 204 Pb/ 206 Pb and 208 Pb/ 206 Pb ratios were better than 0.005.

Major-and Trace-Element Compositions
The major-and trace-element compositions of the samples analyzed during this study are given in Table 1   The gabbros on primitive-mantle-normalized multi-element variation diagrams are enriched in large-ion lithophile elements (LILE; e.g., Rb, U, and K), but negative anomalies in Nb, P, and Ti ( Figure 5b). The shaded area in Figure 5b includes samples from the Akechukesai I mafic-ultramafic complex, showing the similarity of compositions of samples from these two complexes.

Zircon U-Pb Ages and Hf Isotopic Data
The results of U-Pb dating are given in Table 2. The zircons within the gabbro are primarily colorless to charcoal grey, are columnar and contain homogeneous but faintly developed zoning. They appear striped and free of inclusions under cathodoluminescence (CL) imaging and are 150-200 µm long with aspect ratios of 1:1-1.5:1. They contain 176-1617 ppm Th and 533-3421 ppm U, yielding Th/U ratios of 0.33-0.73, indicating they are of magmatic origins. A total of 17 zircons yielded a weighted-mean age of 423.9 ± 2.6 Ma (MSWD = 0.26;1σ), indicating the timing of formation of the gabbro ( Figure 6). The results of in situ Hf isotopic analysis are given in Table 3. A total of 14 zircons yielded εHf (t = 424 Ma) values and associated T DM1 ages of −1.2 to −11.3 and 1535-1092 Ma, respectively. These negative εHf values contrast with the positive εHf values previously determined for the Xiarihamu and Akechukesai I mafic-ultramafic complexes (Figure 7), indicating that ancient crustal material may have been incorporated into the source.  The results of in situ Hf isotopic analysis are given in Table 3. A total of 14 zircons yielded Hf(t = 424 Ma) values and associated TDM1 ages of -1.2 to -11.3 and 1535-1092 Ma, respectively. These negative Hf values contrast with the positive Hf values previously determined for the Xiarihamu and Akechukesai I mafic-ultramafic complexes (Figure 7), indicating that ancient crustal material may have been incorporated into the source.      Data sources for 1.8Ga, 2.5 Ga, and 3.0Ga lines are from Wu et al. [23]. Gray shadow areas are from Li et al. [6] and a compilation of literature data.  Data sources for 1.8Ga, 2.5 Ga, and 3.0Ga lines are from Wu et al. [23]. Gray shadow areas are from Li et al. [6] and a compilation of literature data.   [24], Allegre et al. [25], Zindler and Hart [26].       6. Discussion

Crustal and Source Contamination
Fieldwork in the study area identified an unusual lithology located at the contact between pyroxenite and marble units. This lithology is here termed contaminated rock (Figure 3e), and appears to have formed during reactions between calcium-rich marble and mantle-derived magma. The geochemistry of the gabbros in the study area also provide evidence of assimilation of crustal material, as the variations in the ratios of elements with similar distribution coefficients in the gabbros cannot be explained by partial melting and fractional crystallization alone. As such, these elemental ratios are likely to reflect crustal contamination [27]. The mantle has a uniform Ce/Pb ratio of 25 ± 5 compared to typical crustal material, which has Ce/Pb ratios < 15 [28]. The samples from the Akechukesai IV mafic-ultramafic complex have Ce/Pb ratios of 1.80-2.48, much lower than the value expected for the mantle and similar to typical crustal compositions, indicating that crustal contamination occurred.
Low Ce/Pb ratios in continental crust are usually thought to be the result of Pb enrichment during arc magmas [29], potentially by preferential extraction of Pb from subducted oceanic crustal material and the transfer of this Pb into the mantle wedge via a fluid phase generated during slab dehydration [30]. In addition, the samples from the Akechukesai IV complex have Nb/U ratios of 2.10-2.65, much lower than those of oceanic basalts (~50; [21]), bulk silicate Earth (~32; [21]), and continental upper crust (~9; [31]), but similar to arc volcanic rocks (0.3-9.0; [32]). If intrusions were generated only by contamination of the mantle-derived end-member by continental crust, the ratios of Nb/U should be intermediate between the mantle-derived end-member and continental crust. This also suggests that the primitive magma of these intrusions was originated from a metasomatized mantle source [33]. Nb/Ta ratios are ∼5, different from the range of primitive mantle (17.5 ± 2; [34]) and Zr/Hf ratios (∼10) deviate from the values of crust and chondrites (33-40; [35]), implying the effect of mantle metasomatic processes or fractionation of accessory mineral phases [36].
A Ba/Th versus Th (ppm) diagram can be used to identify the agent involved in the metasomatism of magma sources [37]. High Ba/Th ratios and low Th concentrations generally indicate fluid metasomatism. The samples from the Akechukesai IV mafic-ultramafic complex have high Ba/Th ratios but contain low concentrations of Th (Figure 10a), tend to be consistent with the trend of fluid metasomatism, suggesting that they record fluid metasomatism [37]. Low Th/Yb ratios commonly record fluid-related metasomatism. These samples also have relatively low Th/Yb and Ba/La values (Figure 10b), again suggesting the involvement of fluid metasomatism in their genesis [37]. Métrich et al. [38] suggested that sulfur was one of the most mobile elements to be transferred by fluids from subducted slabs to mantle wedges, as evidenced by research of a close association between S and H 2 O-rich fluid undertaken on Batan Island. This reflects the fact that fluid metasomatism can dissolve and transfer soluble elements from subducted slab (Mclnnes et al. [39]). The partitioning of sulfur between different S-bearing phases during mantle melting generates a strongly heterogeneous distribution of bulk sulfur concentrations in the resultant melt as well as variations in the 34 S composition of the metasomatized mantle source [38]. Addition of S is often linked to the generation of magmatic sulphide mineralization as the addition of significant amounts of S may be more conducive to sulfur saturation of the magma [7]. The Akechukesai IV mafic-ultramafic complex has δ 34 S values (5.0% -13.4% ) that are higher than the expected value for the mantle (0% ± 2% ), suggesting the complex contains additional S, most likely derived from subducted oceanic crustal material. The vein-like and disseminated sulphides have significantly different δ 34 S values, suggesting that the vein-like sulphides contain more sulfur. However, the two ore types have similar Pb isotopic compositions, suggesting they contain Pb derived from the same source. The disseminated and vein-like ores plot between orogenic belt and mantle Pb sources in Figure 9a, however, two vein-like ore samples plot between orogenic belt and upper crustal sources. The two ore types plot between orogenic belt and mantle sources in Figure 9b. Nevertheless, two disseminated samples plot between orogenic belt and lower crustal sources. This suggests that most Pb in these samples was derived from the mantle, with the crustal Pb present added as a result of orogenic processes.

Mantle Source Affinities
The presence of Nb-Ta anomalies within primitive-mantle-normalized multi-element variation diagrams can provide evidence of source characteristics [31]. Samples of the Akechukesai IV mafic-ultramafic complex have negative Nb and Ta anomalies (Figure 5b) relative to both OIB and MORB (positive Nb and Ta anomalies) that suggests the magmas that formed this complex were not derived from primary OIB or MORB sources [40].
In addition, these samples have chondrite-normalized REE patterns that are different from those of N-MORB (Figure 7a). The Hf model ages (T DM ) of zircons may represent the time at which they detached from depleted mantle, if the magma source has been unmodified, the crystallization time of the zircons ought to be approximately equal to the model ages. The Akechukesai IV intrusions have Hf model ages (T DM1 = 1535-1092 Ma) older than their crystallization ages (t = 424 Ma). Compared to the Hf model ages (T DM1 = 610-875 Ma) of the Xiarihamu intrusions, the older Hf model ages (T DM1 = 1535-1092 Ma) of zircons from Akechukesai IV intrusions than their crystallization ages (t = 424 Ma) may indicate that the magma source had experienced high levels of enrichment [7]. The Akechukesai IV mafic-ultramafic samples plot in the arc volcanic field on a Ce versus Ce/Pb diagram (Figure 11a) and in the E-MORB field on a Nb/Yb versus Zr/Yb diagram (Figure 11b), whereas Zr/Hf is~10, significantly different from an E-MORB source (~43; [41,42]), similar to the range of depleted mantle peridotite (~10; [43]). However, these samples plot near the MORB area of 207   Previous research indicates that magmatic rocks generated in post-collisional extensional settings (see Section 6.3) may also have the geochemical characteristics of magmas developed in island arc or active continental margin settings (Aldanmaz et al., [51]). This indicates that magma sources can retain characteristics derived from previously subducted oceanic crust, including Nb, P, and Ti depletions and enrichment in the LILE (Rb, U, and K). This has been confirmed by previous research within the EKOB on the Xiarihamu complexes by Peng et al. [7], and on the Weibao volcanic rocks by Zhong et al. [52]. The fact that the Akechukesai complex formed in a post-collisional setting (see Section 6.3) suggests that it may have inherited characteristics from earlier subduction in this area, as is the case for the Xiarihamu complexes and the Weibao volcanics. This would suggest that the low εHf (t) values of the gabbros in the study area reflect subduction-related metasomatism in a similar fashion to the Xiarihamu mafic-ultramafic complexes (εHf (t) = 4-10; T DM1 = 610-875Ma) and the Weibao volcanic rocks (εHf (t) = -2.6 to +7.5). Lower negative values indicate that ancient crustal composition may have also been incorporated. Couzinié et al. [53] and Franciosi et al. [54] also determined that post-collisional mafic magmatism in the Variscan French Massif Central and proposed a hybrid source containing subducted components and depleted-mantle material. This is also the case for the Xiarihamu mafic-ultramafic complexes and the Weibao volcanic rocks (Peng et al., [7]; Zhong et al., [52]). The next logical step in furthering our understanding of the evolution of the Akechukesai complex would be to quantify the relative contributions of these two end-members.
Previous research has determined that garnet and spinel have different REE partition coefficients that can be used to identify whether magmas are derived from garnet-or spinel-bearing mantle material [55]. The Dy/Yb ratios of melts originating from mantle material within the spinel stability field are generally < 1.5, whereas melts derived from garnet-bearing mantle material generally have Dy/Yb ratios > 2.5 (Yang et al. [56]). In addition, melts with (Dy/Yb) N ratios < 1.06 are indicative of a magma source that contains spinel (Blundy et al. [57]). The samples from the study area have Dy/Yb ratios of 1.9-2.0 and (Dy/Yb) N ratios of 1.2-1.3, indicating they were derived from mantle material containing both garnet and spinel.
Plotting the samples from the study area on a La/Yb versus Yb diagram can provide information on source mineralogy and the degree of partial melting during magma genesis. The proportions of minerals used in this melting model are from McKenzie and O'Nions [58] and Rollinson [59]. Two curves are plotted in Figure 12 to enable the characterization of the mineralogy of the mantle source for the Akechukesai IV mafic-ultramafic complex. The samples from the complex plot between curves reflecting a mix of 5% melting of garnet lherzolite combined with 5% melting of spinel lherzolite and 10% melting of the same combination of the two lithologies. This indicates that the magmas that formed the complex derived from variable degrees of partial melting of mantle material located within both the spinel and garnet stability fields.

Geodynamic Implications
Previous research indicates that magmatic rocks generated in post-collisional extensional settings may also have the geochemical characteristics of magmas developed in island arc or active continental margin settings (Aldanmaz et al., [51]). In the EKOB, the youngest magmatism associated with early Paleozoic subduction event has been identified as 436Ma [61]. Peng et al. [7] have considered that continent collision ceased at 430Ma. Zircon U-Pb dating indicates that the Akechukesai IV mafic-ultramafic complex formed at 423.9 ± 2.6 Ma, coeval with the Akechukesai I mafic-ultramafic complex, which formed in a post-collisional environment (422 ± 10 Ma; unpublished data). Some samples from the Akechukesai IV mafic-ultramafic complex contain biotite and amphibole, indicating that the primitive magma that formed the complex was hydrous. This hydrous magma could have formed by decompression-related melting of a hydrous region of the mantle (Gao et al. [62]).These evidences indicate that the Akechukesai IV mafic-ultramafic complex may also be generated in a post-collisional environment.
Some of the mafic-ultramafic complexes and A-type granitoids in the EKOB have similar ages, as exemplified by the Xiarihamu I mafic-ultramafic complex (423 ± 1 Ma; Wang et al. [63]), the Xiarihamu II mafic-ultramafic complex (424 ± 1 Ma; Peng et al. [7]), and the Baiganhu A-type moyite and monzonitic granites (ages of 422 ± 3 and 421 ± 3.7 Ma, respectively; Li et al. [64]). The Wulonggou A2-type granite, which formed at 426-424 Ma, represents the earliest evidence for a post-orogenic tectonic setting in this area (Xin et al. [9]). The presence of these coeval middle Silurian mafic-ultramafic complexes and A-type granites provide evidence of formation in a post-collisional tectonic environment.
The EKOB experienced two stages of post-collisional mantle-derived magmatism during the middle Silurian and Late Triassic. The mineralized and prospective mafic-ultramafic complexes discovered in the EKOB to date all formed during the middle Silurian, namely the Xiarihamu, Akechukesai, Shitoukengde, and Langmuri mafic-ultramafic complexes. Although some Late Triassic mafic-ultramafic complexes have been identified in this area (e.g., the Kaimuqi and Xiaojianshan complexes), these complexes appear to be unmineralized. As such, research has focused on the middle Silurian mafic-ultramafic complexes in this area. Among these, the Xiarihamu complexes, which host a super-large Cu-Ni deposit, are primarily studied in this area, and have characteristics that are representative of the other mafic-ultramafic complexes formed at this time. Combined with the regional tectonic evolution, Wang et al. [63] and Peng et al. [7] suggest that the Xiarihamu mafic-ultramafic complexes formed in a post-collisional tectonic setting associated with

Geodynamic Implications
Previous research indicates that magmatic rocks generated in post-collisional extensional settings may also have the geochemical characteristics of magmas developed in island arc or active continental margin settings (Aldanmaz et al., [51]). In the EKOB, the youngest magmatism associated with early Paleozoic subduction event has been identified as 436Ma [61]. Peng et al. [7] have considered that continent collision ceased at 430Ma. Zircon U-Pb dating indicates that the Akechukesai IV mafic-ultramafic complex formed at 423.9 ± 2.6 Ma, coeval with the Akechukesai I mafic-ultramafic complex, which formed in a post-collisional environment (422 ± 10 Ma; unpublished data). Some samples from the Akechukesai IV mafic-ultramafic complex contain biotite and amphibole, indicating that the primitive magma that formed the complex was hydrous. This hydrous magma could have formed by decompression-related melting of a hydrous region of the mantle (Gao et al. [62]). These evidences indicate that the Akechukesai IV mafic-ultramafic complex may also be generated in a post-collisional environment.
Some of the mafic-ultramafic complexes and A-type granitoids in the EKOB have similar ages, as exemplified by the Xiarihamu I mafic-ultramafic complex (423 ± 1 Ma; Wang et al. [63]), the Xiarihamu II mafic-ultramafic complex (424 ± 1 Ma; Peng et al. [7]), and the Baiganhu A-type moyite and monzonitic granites (ages of 422 ± 3 and 421 ± 3.7 Ma, respectively; Li et al. [64]). The Wulonggou A2-type granite, which formed at 426-424 Ma, represents the earliest evidence for a post-orogenic tectonic setting in this area (Xin et al. [9]). The presence of these coeval middle Silurian mafic-ultramafic complexes and A-type granites provide evidence of formation in a post-collisional tectonic environment.
The EKOB experienced two stages of post-collisional mantle-derived magmatism during the middle Silurian and Late Triassic. The mineralized and prospective mafic-ultramafic complexes discovered in the EKOB to date all formed during the middle Silurian, namely the Xiarihamu, Akechukesai, Shitoukengde, and Langmuri mafic-ultramafic complexes. Although some Late Triassic mafic-ultramafic complexes have been identified in this area (e.g., the Kaimuqi and Xiaojianshan complexes), these complexes appear to be unmineralized. As such, research has focused on the middle Silurian mafic-ultramafic complexes in this area. Among these, the Xiarihamu complexes, which host a super-large Cu-Ni deposit, are primarily studied in this area, and have characteristics that are representative of the other mafic-ultramafic complexes formed at this time. Combined with the regional tectonic evolution, Wang et al. [63] and Peng et al. [7] suggest that the Xiarihamu mafic-ultramafic complexes formed in a post-collisional tectonic setting associated with closure of the proto-Tethyan Ocean. The dehydration of the subducted slab generated a water-rich fluid phase that migrated into and subsequently metasomatized (and transported S into) the overlying mantle wedge. The thickness of this oceanic plateau material (up to 36 km) impeded normal subduction, potentially causing break-off of the subducted slab [12]. This would allow hot asthenospheric mantle material to ascend through the resulting slab window, causing shallower decompression melting. This slab break-off model is consistent with the linear distribution of mafic-ultramafic complexes along the middle Kunlun fault.

Conclusions
1. The Akechukesai IV mafic-ultramafic complex was formed at 423.9 ± 2.6 Ma and it was just one of the mafic-ultramafic complexes emplaced in the middle Silurian within the post-collisional tectonic framework of the EKOB.
2. Hf-S-Pb isotopes and geochemical data indicate that primitive magma has a MORB affinity and is derived from a mantle source that is metasomatized by subduction-related fluids. The older T DM1 ages than crystallization ages indicate that the original source experienced some enrichment. The magma source is composed of a mix between 5% melted garnet lherzolite-5% spinel lherzolite and 10% melted garnet lherzolite-and 10% spinel lherzolite.
3. Addition of high amounts of sulfur may have had a significant role in driving S saturation, leading to the precipitation of sulfide minerals.