Geochronological, Geochemical and Sr-Nd-Hf Isotopic Studies of the A-type Granites and Adakitic Granodiorites in Western Junggar: Petrogenesis and Tectonic Implications

: Late Carboniferous magmatism in the Western Junggar region of the Central Asian Orogenic Belt (CAOB) provides a critical geological record of regional tectonic and geodynamic history. In this study, we determined the zircon U-Pb isotopic compositions, bulk-rock Sr-Nd-Hf isotopic compositions, and major and trace element geochemistry of two granitic bodies in the Western Junggar, with the aim of constraining their emplacement ages, magmatic origin, and geodynamic signiﬁcance. Radiometric ages indicate that the plutons were emplaced during the Late Carboniferous (322–307 Ma). Plutons in the North Karamay region are characterized by high Sr content (347–362 ppm) and low Y content (15.3–16.7 ppm), yielding relatively high Sr / Y ratios (20.8–23.7). They show consistent Yb (1.68–1.85 ppm), Cr (16–19 ppm), Co (7.5–8.1 ppm) and Ni (5.9–6.6 ppm) content, similar to that of modern adakites. The Hongshan plutons are characterized by high SiO 2 (69.95–74.66 wt%), Na 2 O (3.26–3.64 wt%), and K 2 O (4.84–5.16 wt%) content, low Al 2 O 3 (12.02–12.84 wt%;) and MgO (0.13–0 18 wt%) content, and low Mg # values (0.16–0.22). This group shows a clear geochemical a ﬃ nity with A-type granites. All of the studied granitoids have positive ε Nd(t) ( + 4.89 to + 7.21) and ε Hf(t) ( + 7.70 to + 13.00) values, with young T DM (Nd) 806–526 Ma) and T DM (Hf) (656–383 Ma) ages, indicating a substantial addition of juvenile material. The adakitic granodiorites in the North Karamay region were likely generated via partial melting of thickened lower crust, while the A-type granites in the Hongshan area may have been derived from the melting of lower-middle crust in an intra-oceanic arc, which consists mainly of oceanic crust. The emplacement of these granitoids represents a regional magmatic “ﬂare up”, which can be explained by the rollback of a subducting slab.


Introduction
Magmatic rocks represent crucial windows into regional tectonic processes and events, and can provide important constrains on the dynamics of the deep asthenosphere [1][2][3][4][5]. The production of subduction-related magmatic rocks can be attributed to several distinct geodynamic mechanisms, The Western Junggar is located in the southern part of the CAOB, and provides an ideal natural laboratory to study the evolution of the orogen (Figure 1b; [23,31,43]). Granitoid intrusions are widespread in the Western Junggar region, and are characterized by highly-depleted Nd isotopic signatures (εNd(t) = +6.4 to +9.2) [4,[27][28][29][30]44,45]. However, the origin of the granitic intrusions in the Western Junggar is controversial. They have been proposed to reflect either a subduction-related island arc setting [13,23,[46][47][48], or a post-collision extensional regime [4,44,45]. These two models

Zircon U-Pb Dating
Zircon grains were extracted using standard density and magnetic separation techniques. Cathodoluminescence (CL) images of the separated zircon grains were obtained using an FEI NOVA NanoSEM 450 scanning electron microscope, equipped with a Gatan Mono CL4 cathodoluminescence system, at the State Key Laboratory of Continental Tectonics and Dynamics, Institute of Geology, Chinese Academy of Geological Sciences. Zircon U-Pb isotope and trace element analyses were carried out simultaneously using an Agilent 7500a ICP-MS equipped with a 193 nm GeoLas 2005 laser ablation system, at the State Key Laboratory of Geological Processes and Mineral Resources, China University of Geoscience, Beijing, Wuhan. Analyses were conducted with a beam diameter of 32 µm, 5 Hz repetition rate, and energy of 10-20 J/cm 2 . Zircon standard 91500 was analyzed twice for each five sample analyses, and used to calibrate isotope fractionation. The NIST 610 glass standard was also analyzed once for every ten analyses, in order to correct for time-dependent drift in the sensitivity or mass discrimination of the instrument. Details of the instrumental conditions, analytical procedures, and data reduction process are given in Liu et al. [72]. Calculated ages and concordia diagrams were generated using Isoplot/Ex 2.49 software [73].

Sr-Nd Isotope Analysis
Sample powders were spiked with mixed isotope tracers and dissolved in Teflon capsules with a mixture of HF and HNO 3 prior to Sr and Nd isotope analysis. Sr and rare earth elements (REEs) were separated using Eichrom resin columns, with 0.1% HNO 3 as elutant. Separation of Nd from the REE fraction was carried out using a HDEHP column, with a 0.18 N HCl elutant. Isotopic measurements were conducted via thermal ionization mass spectrometry (TIMS) at the Experimental Test Center, Tianjin Institute of Geology and Mineral Resources. Mass fractionation corrections for Sr and Nd isotopic ratios were based on 86 Sr/ 88 Sr and 146 Nd/ 144 Nd ratios of 0.1194 and 0.7219, respectively. The 87 Sr/ 86 Sr ratios of the NBS987 and NBS607 isotopic standards, and the 143 Nd/ 144 Nd ratios for the BCR-1 and La Jolla isotopic standards, were measured as 0.710240 ± 15 (2σ), 1.20032 ± 30 (2σ), 0.512663 ± 9 (2σ), and 0.511862 ± 7 (2σ), respectively. Sample preparation and analytical procedures followed those of Zhang et al. [74].

Major and Trace Element Analyses
The whole rock major and trace element compositions of the studied samples were analyzed at the State Key Laboratory of Ore Deposit Geochemistry, Institute of Geochemistry, Chinese Academy of Sciences. Fresh chips of bulk sample were powdered to 200-mesh using a tungsten carbide ball mill. Subsamples for trace element analysis were digested in high-pressure Teflon bombs at 190 • C for 48 h, using a mixture of HF and HNO 3 . Major elements were analyzed using a Rikagu RIX 2100 x-ray fluorescence (XRF) spectrometer, and trace elements were analyzed using an Agilent 7500a inductively coupled plasma mass spectrometer (ICP-MS). United States Geological Survey and international rock standards (BHVO-2, AGV-2, BCR-2 and GSP-1) were used to assess data quality and measurement repeatability. The analytical precision and accuracy for most major and trace elements were better than 5% and 10%, respectively [75].

In Situ Zircon Hf Isotope Analysis
Zircon Hf isotope analyses were conducted using a Neptune Plus multicollector ICP-MS instrument (Thermo Fisher Scientific, Germany) coupled to a Geolas 2005 excimer ArF laser ablation system (LambdaPhysik, Göttingen, Germany), at the State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences, Wuhan, China. All data were acquired in single-spot ablation mode, with a spot size of 44 µm. Each measurement consisted of 20 s of background signal acquisition, followed by 50 s of ablation signal acquisition. Detailed analytical procedures and operating conditions for both the laser ablation system and the MC-ICP-MS instrument are described in Hu et al. [76]. Offline selection and integration of analyte signals, and mass bias calibrations, were conducted using the ICP-MS DataCal software [77].

Zircon U-Pb Ages
Two samples from the North Karamay and Hongshan plutons were selected for zircon U-Pb dating ( Table 1). The zircon grains in these granite samples are pale yellow in color, transparent, euhedral to subhedral, and 80-100 µm in size. All zircons exhibit bright cathodoluminescence, with clear concentric oscillatory zoning ( Figure 4). The Th and U concentrations for all analyzed sites ranged from 129 to 2231 ppm and from 218 to 1772 ppm, respectively, with relatively high Th/U ratios (0.64-1.53). These characteristics are typical of zircons that are magmatic in origin. All U-Pb measurements plotted very close to the concordant line on the 206 Pb/ 238 U vs. 207 Pb/ 235 U concordia diagram ( Figure 5). These analyses yield ages of 308 ± 5 Ma for the Hongshan granitoids, and 323 ± 3 Ma for the North Karamay granitoids.      (Figure 7a,b) likely implies plagioclase differentiation during petrogenesis [30]. Primitive mantle-normalized trace element spider diagrams (Figure 7a) show that the Hongshan samples are enriched in Rb, U, and Pb, with pronounced negative Ba, Sr, and Eu anomalies. The major and trace elements of the studied grantoids are list in Tables 2 and 3.

North Karamay Granitoids
All of the North Karamay granitoids plot in the adakite field on the Sr/Y vs. Y and (La/Yb) N vs. (Yb) N discrimination diagrams (Figure 9a,b; [78,87]), indicating an adakitic affinity. The genesis of adakitic magmas remains the subject of debate, and several major hypotheses have been proposed, including (1) assimilation and low-pressure fractional crystallization in basaltic parent magmas [88], (2) partial melting of thickened lower crust (high SiO 2 and K 2 O contents and low MgO, Cr, and Ni contents, [89][90][91][92]), (3) partial melting of delaminated lower crust (high Cr, Co and Ni contents; [93][94][95][96]), (4) mixing of basaltic and adakitic magmas [97], (5) partial melting of young, hot, recently subducted oceanic crust [98], and (6) melting of subducted continental crust [99]. The studied adakitic granodiorites show none of the compositional trends characteristic of low-pressure or high-pressure assimilation fractional crystallization (AFC) (Figure 10), and it is therefore unlikely that they represent the product of low-pressure or high-pressure fractional crystallization. The North Karamay granitoids have very low Mg # values (0.37-0.38), Cr (<16.0 ppm), Co (<8.1 ppm) and Ni (<6.6 ppm) content, meaning they are unlikely to have been formed by melt-mantle interaction, which usually occurs when delaminated lower crust melt interacts with mantle peridotite (Figure 11a-c; [93,94]). The studied granitoids do show highly positive εNd(t) and εHf(t) values, which are distinctly different from those of adakites formed via partial melting of continental crust; in Tibet, these typically show low, negative εNd(t) and εHf(t) values [99].  No mafic enclaves have been observed in the adakitic granites, suggesting that they did not form via mixing of basaltic and adakitic magmas. The adakitic granodiorites also have lower εNd(t) values (+4.82 to +4.95) than the A-type granites (+6.70 to +7.21) (Figure 8a), which represent the most crust-derived, felsic end member in the studied region. This rules out the possibility that the adakitic granodiorites are the product of mixing between basaltic and adakitic magmas. Therefore, we suggest that the adakitic granodiorites were most likely generated by partial melting of the lower crust. This scenario explains both the relatively high SiO 2 and K 2 O content, and the low MgO, Cr, and Ni content.

Hongshan Granitoids
The Hongshan granitoids are characterized by high alkaline content, as well as high Fe, Zr, and Nb content. They display prominent negative anomalies in Ba, Sr, P, Eu, and Ti (Figure 6a,b), and high 10 4 Ga/Al ratios, similar to typical A-type granites [101,102]. All samples from the Hongshan plutons plot in the A-type field on the 10 4 Ga/Al vs. Zr and 10 4 Ga/Al vs. (K 2 O + Na 2 O) discrimination diagrams (Figure 12a,b; [101]). The depleted Nd-Hf isotopic compositions of the Hongshan granitoids (εNd(t) = +6.70 to +7.21; εHf(t) = +10.79 to +13.00, Tables 4 and 5) imply that they are unlikely to have been generated via partial melting of crystalline basement rocks, which typically have distinctly negative εNd(t) values in the Western Junggar [103]. Other potential Carboniferous sources in the Western Junggar region include Early Carboniferous arc-type basalts and basaltic andesites [55] and the Late Carboniferous Hatu tholeiitic basalts [104]. However, the Sr and Nd isotope ratios of the Hongshan granites are much greater than those of the contemporary Hatu basalts (Figure 13a). Therefore, the Hongshan A-type granites could not have been generated via fractional crystallization of mantle-derived arc basalt melts. These data suggest that these A-type granites likely formed via partial melting of the oceanic crust, based on the following lines of evidence: (1) the negative Eu anomalies and high concentrations of HREEs; (2) the field encompassing the Hongshan A-type granites partially overlies that of the North Xinjiang ophiolites (Figure 13a,b). Therefore, we proposed that the Hongshan A-type granites are derived from partial melting of the oceanic crust.
Previous studies have demonstrated that partial melting of oceanic crustal components stored in the lower crust could provide the parent magma for A-type granites [48]. The lower and middle crust of the Western Junggar intra-oceanic arc would likely have consisted of the arc itself, sitting on top of basaltic oceanic crust [20]; it may also have included underthrust or accreted oceanic crustal components [105,106]. Therefore, we propose here that the A-type granites are mainly derived from melting of the lower and middle crust of an intra-oceanic arc, which largely consisted of oceanic crust.

Tectonic Implications
All of the studied granitoids plot in the "VAG" field on the Rb vs. (Yb + Ta) and Rb vs. (Y + Nb) tectonic discrimination diagrams (Figure 14a,b), indicating that they are likely to have been formed during island-arc magmatism. Our zircon U-Pb data indicate a Late Carboniferous (ca. 322-307 Ma) magmatic "flare up" occurred in the Western Junggar. Asthenospheric sources of heat are required to explain the formation of the studied granites, and two geodynamic models have been proposed: subduction of an oceanic ridge [13,47,48,107], or the rollback of a subducting slab [85,86]. Both of these processes are capable of generating adakites and A-type granites in a subduction setting [12,48,98,[108][109][110][111][112]. If this ridge subduction model is accurate, the distribution of adakites and A-type granites would imply a spreading ridge oriented sub-parallel to the subduction trench, and perpendicular to the suture zone. However, The Devonian-Carboniferous granitoids in the Karamay region define a narrow, linear zone of magmatism parallel to the subduction zone, which is inconsistent with a ridge subduction model. In addition, the adakitic rocks in a slab window are typically characterized by high Cr, Co, and Ni content [85], whereas the Karamay adakitic granites show generally low Cr, Co, and Ni content. Moreover, the adakitic granodiorites and A-type granites are almost contemporaneous with arc volcanic rocks, which does not support the ridge subduction model [86].
An alternative explanation is that asthenospheric upwelling resulted from the rollback of a subducting slab of Junggar oceanic crust. This scenario would also have provided the required high heat flow to drive partial melting of the subducting crust ( Figure 15). In this model, the lithosphere is negatively buoyant, causing the subducting slab to sink vertically [113,114]. If the slab sinks into the mantle faster than it converges with the overriding plate, the subduction trench is expected to 'roll back' from the overriding plate [115]. Slab rollback could induce upwelling of the asthenosphere to compensate for the loss in volume of the mantle wedge [86,116]. During the Late Silurian to Early Devonian, the oceanic crust underlying the Junggar Ocean was being consumed beneath the Karamay volcanic arc. With ongoing extension during the rollback of the Junggar slab, the crust and lithospheric mantle became progressively thinner. Upwelling of the asthenosphere might have triggered partial melting of the oceanic crust and overlying lower and middle crustal rocks, generating adakitic magmas and A-type granitic magmas [117,118]. Therefore, we conclude that slab rollback is the most likely explanation for the Late Silurian to Early Devonian magmatic "flare up" in the Western Junggar.

1.
The two studied granitoids from the Western Junggar can be classified as an adakitic granodiorite and an A-type granite. Both were emplaced in the period from 322-307 Ma.

2.
The adakitic granodiorites were likely generated from the partial melting of thickened lower crust. The A-type granites may have formed from partial melting of lower and middle crust in an intra-oceanic arc setting, where the melted material mainly consisted of oceanic crust.

3.
Slab rollback appears to have played an important role in the generation of arc-related igneous rocks, and in the continental growth of the Central Asian Orogenic Belt.