INTRODUCTION

The Central Asian Orogenic Belt (CAOB) has long been considered as one of the largest accretionary orogens consisting of island arcs, ophiolitic belts, accretionary complexes, and Precambrian terranes (Kozakov et al., 1976, 1999; Sengör et al., 1993; Kovach et al., 2004; Windley et al., 2007; Kröner et al., 2014; Kozakov and Azimov, 2017). Accompanied by the southward subduction and closure of the Paleo-Asian Ocean, extensive continental growth occurred in the CAOB from Neoproterozoic to Mesozoic (Khain et al., 2003; Jahn et al., 2004; Windley et al., 2007; Kröner et al., 2014, 2017; Xiao et al., 2015). However, there is currently no consensus on the timing of the final closure of the Paleo-Asian Ocean, with different studies placing the estimate from Late Devonian to Late Triassic (Jahn et al., 2004; Xiao et al., 2009; 2015; Windley et al., 2007; Han et al., 2011; Kröner et al., 2014, 2017; Zhang et al., 2015a, 2015b).

The CAOB is located to the south of the Siberia-East Europe cratons and the north of the Tarim-Alxa-North China Cratons (Fig. 1a). It is generally accepted that the Paleo-Asian Oceanic closure is recorded by the Tianshan-Solonker suture zones, which extend along the northern edges of the Tarim-Alxa-Tarim cratons (Xiao et al., 2015). The Alxa is a key junction between the Tarim and the North China, with the former to its west and the latter to its east (Fig. 1a). Compared to its two flanking geographical neighbors, the Alxa Block has received significantly less research attention. One of the geological features of the Alxa Block is the presence of abundant late Paleozoic to Mesozoic igneous rocks, which mostly comprise granitoids, and to a lesser extent, mafic to intermediate intrusions, with a small proportion of lava (Liu et al., 2017a, 2017b; Fig. 1b). These igneous rocks could offer important insights into the tectonic evolution of the Paleo-Asian Ocean as a record of its southward subduction process. However, previous studies have focused on crust-derived granitoids and less attention has been paid to mantle-derived basaltic rocks.

Fig. 1.
figure 1

(a) Simplified tectonic sketch map of the Central Asian Orogenic Belt (CAOB), showing the Location of the Alxa Block (modified after Liu et al. (2017a, 2017b)). (b) Geological map of the Alxa Block (modified after Liu et al. (2017a, 2017b)). SLB – Shalazhashan belt; NLB – Nuru-Langshan belt; ① – Enger Us fault; ② – Badain Jaran fault; ③ – Langshan fault; ④ – Longshoushan Fault.

Full size image

In this paper, we investigate the Jigede gabbro intrusion from the Shalazhashan tectonic belt in the northern Alxa Block. New laser ablation–inductively coupled plasma–mass spectrometry (LA-ICP-MS) zircon U-Pb dating, whole-rock geochemical and Sr-Nd isotopic data of the intrusion are obtained to explore its petrogenesis and the associated tectonic setting. These results were subsequently used to constrain the timing of the final closure of the Paleo-Asian Ocean.

GEOLOGICAL CONTEXT AND SAMPLING

The Alxa Block is separated from the southernmost CAOB by the Enger Us fault (Fig. 1b). The Enger Us ophiolite suite distributed along the fault, represents the site of the final Paleo-Asian oceanic closure in the Alxa area (Wang et al., 1994; Zheng et al., 2014). The Alxa Block is divided by the Badain Jaran fault into two tectonic belts, including the Shalazhashan belt in the north and the Nuru-Langshan belt in the south (Fig. 1b). The Quagan Qulu ophiolite suite along the fault was produced in a back-arc setting, in response to the Paleo-Asian oceanic subduction (Wang et al., 1994; Wu et al., 1998; Zheng et al., 2014). The Alxa Block is underlain by Precambrian rocks and Late Paleozoic strata, with the absence of Early Paleozoic strata (Fig. 1b). There are abundant Paleozoic to Mesozoic magmatic rocks in the block, including predominant granitoids and secondary mafic-intermediate intrusions as well as lavas (Fig. 1b).

The magmatic rocks in the Shalazhashan tectonic belt are dominated by late Paleozoic to Triassic granitoids (301 to 247 Ma), with minor amounts of diorites and gabbros (264 to 249 Ma) (Liu et al., 2017a, 2017b). The strata in the belt is composed mainly of late Paleozoic sedimentary rocks represented by the Late Carboniferous to Early Permian Amushan Formation. The Amushan Formation can be subdivided into a lower sequence of lava-interbedded clastic rocks with zircon U-Pb ages of 320 to 302 Ma, and upper sequences of shale, sandstone, and conglomerate (Lu et al., 2012; Liu et al., 2017a, 2017b).

In this contribution, we investigate the Jigede intrusion, which was emplaced in a Permian granite (Fig. 2). The samples collected from the intrusion are medium to fine-grained gabbros, which consisted of plagioclase (60–70%), clinopyxene (~10–20%), biotite (~10–15%), and amphibole (~5%) (Fig. 3). We collect eight representative samples from the Jigede intrusion for whole-rock geochemical and Sr-Nd isotopic analysis. Sample 16CG-20 was collected for zircon separation and U-Pb dating. The detailed sampling locations are illustrated in Fig. 2.

Fig. 2.
figure 2

Geological map of the Jigede intrusion. The sampling locations are also shown.

Full size image
Fig. 3.
figure 3

Representative photomicrographs of the Jigede gabbro. Bt—biotite, Cpx—clinopyxene, Hb—hornblende, Pl—plagioclase.

Full size image

ANALYTICAL METHODS

Zircon LA-ICP-MS U-Pb Dating

We separated zircons from the Sample 16CG-20 through conventional heavy liquid and magnetic separation methods, and then handpicked them with a binocular microscope. Representative zircon crystals were prepared and placed on a glass slide and covered with epoxy resin in a cylindrical frame. The surface of this side was polished for further cathodoluminescence (CL) investigations. CL images were obtained using a GATAN MINI probe placed on the JSM 6510 electron microprobe, with an accelerating potential of 10 kV at Beijing GeoAnalysis Co. Ltd., China. LA‑ICP-MS U-Pb analyses for these zircons were conducted at the State Key Laboratory for Mineral Deposits Research, Nanjing University (SKLMDR, NJU), determined by an Agilent 7500a ICP-MS equipped with a New Wave Research 213 nm laser ablation system. Analytical procedures are described by Xu et al. (2009) and Jackson et al. (2004) in detail. We corrected mass discrimination of the mass spectrometer and residual elemental fractionation through calibration against a homogeneous standard zircon, GEMOC/GJ-1 (608.53 ± 0.37 Ma, n = 8, 2σ; Jackson et al., 2004). The standard zircon Mud Tank (TIMS = 732 ± 5 Ma, n = 5, Black, Gulson, 1978; LA-ICP-MS 206Pb/238U = 732.4 ± 1.4 Ma, n = 359 at 95% confidence, Jackson et al., 2004) was determined as an independent control on reproducibility and instrument stability. We evaluated common Pb contents and corrected common Pb following a method introduced by Andersen (2002).

Major and Trace Elements, and Whole-rock Sr-Nd Isotope Analysis

Major element contents were analyzed using XRF on fused glass beads at SKLMDR, NJU, with precision better than 5%. When it comes to analyzing trace elements, for each sample, about 50 mg of powder was dissolved in s screw-top Teflon beaker, mixed with HF/HNO3 mixture acid, and then heated at about 160°C for 48 h. We determined whole-rock trace element contents through a Finnigan Element II inductively coupled plasma mass spectrometry (ICP-MS) at the State Key Laboratory of Continental Dynamics, Northwest University (SKLCD, NWU), with precision better than 10%. Detailed methods for trace elements analyses are presented in (Gao et al., 2003).

The Sr-Nd isotopic ratios were measured using a Finnigan Triton TI thermal ionization mass spectrometer (TIMS) at SKLMDR, NJU. For Sr-Nd isotopic analyses, about 100 mg of powder was dissolved in Teflon beakers with a HF + HNO3 mixture acid, and Sr and Nd were then separated and purified by conventional cation-exchange technique. Detailed analytical procedures are elaborated by (Pu et al., 2004, 2005). The mass fractionation corrections about 87Sr/86Sr and 143Nd/144Nd ratios are on the basis of 86Sr/88Sr of 0.1194 and 146Nd/144Nd of 0.7219, respectively. During the analyses, measurements for the La Jolla standard gave 143Nd/144Nd of 0.511842 ± 4 (2σ, n = 5), and for NBS-987 Sr standard gave 87Sr/86Sr of 0.710260 ± 10 (2σ, n = 30). Total analytical blanks were 5 × 10−11 g for Sm-Nd and (2–5) × 10−10 g for Rb-Sr.

RESULTS

LA-ICP-MS zircon U-Pb Dating

The zircons separated from 16CG-20 ranged from 50 to 100 µm in length and exhibited no oscillatory zoning (Fig. 4). They have high Th/U ratios in the range of 0.20 and 0.71 (Table 1), indicating magmatic origins (Williams et al., 1996; He et al., 2018). All U-Pb data that we obtained clustered on the concordant curves (Fig. 4) and produced a weighted mean 206Pb/238U age of 262 ± 4 Ma (1σ, MSWD = 1.4) for 16CG-20 (Fig. 4), suggesting that the Jigede intrusion was emplaced during the middle Permian.

Fig. 4.
figure 4

LA-ICP-MS zircon U-Pb concordant curves and typical CL images of representative zircons collected from the sample 16CG-20.

Full size image
Table 1.   LA-ICP-MS U-Pb isotopic analysis for zircons from the Jigede gabbro (sample 16CG-20)
Full size table

Bulk Geochemistry and Sr-Nd Isotope

The Jigede intrusion has a low SiO2 content of 49.80 to 52.22 wt % (Table 2; Fig. 5). Compared to the experimental partial melts of mantle rocks, all the samples contain similar abundance of MgO (11.15–12.15 wt %), low contents of TiO2 (0.24–0.37 wt %) and Fe2O3T (4.87–5.41 wt %), as well as high levels of CaO (11.45–12.75 wt %), Al2O3 (14.18–17.08 wt %) and Mg# (0.81–0.83) (Table 2, Fig. 6). Meanwhile, the samples are enriched in LREEs relative to HREEs, with positive Eu anomalies (Fig. 7a). In PM-normalized spidergrams, all samples exhibit notably negative Th, Nb, Zr-Hf, Ti anomalies as well as positive U and Sr anomalies (Fig. 7b). Compared to mid-ocean-ridge basalt (MORB), the Jigede gabbros are depleted in REEs and HFSEs, but enriched in LILEs (Fig. 7b). These mafic samples have low initial 87Sr/86Sr in the range of 0.7046 to 0.7054, and positive εNd(T) values of + 1.8 and + 4.8 (Table 3).

Table 2.   Whole-rock major element oxides (wt %), trace elements (ppm) and REE (ppm) abundances of selected rocks samples
Full size table
Fig. 5.
figure 5

TAS-diagram (Le Maitre, 2002) of the Jigede intrusion.

Full size image
Fig. 6.
figure 6

Plots of (a) MgO vs. Total FeO, (b) TiO2 vs. total FeO and (c) CaO vs. Al2O3 for the Jigede intrusion. The fields of some experimental melts of depleted and fertile peridotites are also exhibited (Falloon et al, 1988).

Full size image
Fig. 7.
figure 7

(a) Chondrite-normalized (Sun, McDonough, 1989) REE patterns and (b) primitive mantle-normalized (Sun, McDonough, 1989) trace element patterns for the Jigede intrusion.

Full size image
Table 3.   Sr-Nd isotopic components of the studied intrusions
Full size table

DISCUSSION

Crustal Assimilation

The Jigede intrusion contains basaltic components with low SiO2 abundances of 49.80–52.22 wt % (Table 2). During the emplacement of the basaltic magma, continental components may contaminate and modify the initial melt. In this regard, the ratios of isotopes and strongly incompatible elements can serve as indicators of crustal contamination as they are not disturbed by pure fractional crystallization. Continental crust, especially upper crust, is relatively enriched in La, but depleted in Nb, with a low Nb/La ratio (0.39) (Rudnick, Gao, 2003). However, the Jigede gabbros exhibit lower Nb/La ratios (0.13 to 0.19) than upper continental crust (Fig. 8a), implying that the depletion of Nb is not associated with crustal contamination. Furthermore, the Nb/La ratios of all samples were found to be relatively constant and showed no apparent correlation with their SiO2 levels (Fig. 8a). The low initial Sr isotopic ratios, positive εNd(T) values as well as the slightly positive correlation between εNd(T) and SiO2 also suggested that the contamination, if any, is relatively minor (Fig. 8b). Most importantly, significant crustal contamination is not reconcilable with the markedly negative Zr-Hf anomalies in the spider diagrams (Fig. 7b).

Fig. 8.
figure 8

Plots of (a) Nb/La, (b) εNd(T) and SiO2 for the Jigede intrusion. UCC: upper continental crust. The data of UCC are from Rudnick and Gao (2003).

Full size image

Mineral Cumulation

Figure 3 shows a partly-densified, texturally equilibrated mosaic of plagioclase. The petrographic feature provided clear evidence of plagioclase accumulation in the Jigede intrusion (Hunter, 1996; Fig. 3). This is further supported by the positive Sr and Eu anomalies in the REE and trace element patterns (Fig. 7). Compared to those experimental partial melts of mantle rocks, the Jigede gabbros have high CaO and Al2O3 contents (Fig. 6c), which also entailed the cumulation of plagioclase. Relative to MORB, the low abundances of REEs and HFSEs abundances in the Jigede gabbros can be also attributed to plagioclase cumulation due to their strongly incompatibility in the mineral (Fig. 7). In the thin section, we did not observe any significant accumulation of mafic minerals (Fig. 3). The positive correlation between Fe2O3T, MgO, TiO2 and SiO2 further precluded the cumulation of mafic minerals and Ti-Fe oxides (Fig. 9).

Fig. 9.
figure 9

Plots of SiO2 vs. (a) TiO2, (b) MgO and (c) \({\text{F}}{{{\text{e}}}_{2}}{\text{O}}_{3}^{{\text{T}}}\) for the Jigede intrusion.

Full size image

Source Nature

Strongly incompatible element ratios and Sr-Nd isotopic components are not generally disturbed by fractional crystallization or cumulation of minerals. Thus, they can be used to infer source nature. Relative to HREEs, all the samples are evidently enriched in LREEs ((La/Yb)n = 1.52–2.13) (Fig. 7a) and exhibit significantly different (87Sr/86Sr)i (0.7046 to 0.7054) and εNd(T) (+ 1.8 to + 4.6) from MORB (Table 3). Thus, the Jigede intrusion might originate from a modified subcontinental lithospheric mantle or an ocean-island-basalt (OIB)-like mantle rather than a MORB-like mantle source. On the other hand, the low contents of incompatible elements and the depletion of Nb, Zr-Hf as well as Ti in the PM-normalized spidergrams argue against an OIB-like mantle source (Fig. 7b; Sun and McDonough, 1989). Experimental studies revealed that the partial melts of fertile peridotite have higher TiO2 contents (>1 wt %) than those (<1 wt %) derived from depleted peridotite at a given FeOT content (Fig. 6b). The Jigede intrusion exhibits a low TiO2 level in the range of 0.24–0.37 wt %, indicating a depleted mantle source. The greatest possibility is a lithospheric mantle source. The relative depletion of these HFSEs has been interpreted as an indicator of a subduction process (Thirlwall et al., 1994). Taken together, we tentatively conclude that the Jigede intrusion was most likely to have originated from a subcontinental lithospheric mantle source modified by subducted slab-derived components (fluids or melts). Furthermore, both HFSEs and REEs (e.g., Th, Y, Zr, Hf, Ti and Nb) are more immobile in fluids than in melts (Kepezhinskas et al., 1997; Zhao and Zhou, 2007). Thus, relative to LILEs, the depletion of Th, Zr-Hf and Ti in the trace element patterns indicates the introduction of slab fluids into the overlying mantle source (Fig. 7b; Liu et al., 2018). The positive U anomaly and high Th/U ratio can be attributed to preferential partitioning of U into aqueous fluid coexisting with subducted eclogitic slab (Brenan et al., 1995a, 1995b). The fluid involvement is also supported by the presence of hornblende (water-bearing mineral) in the mineral assemblages. REEs ratios can be used to explore the formation depths of basaltic magmas (McKenzie and O’Nions, 1991). The low (Dy/Yb)n ratios (1.13 to 1.32) coupled with flat HREE pattens (Fig. 7a) indicated that the formation of the initial melts came from the spinel stability field (Liao et al., 2015; Liu et al., 2015). Thus, melting depth is not more than 80 km (Robinson and Wood, 1998). Such a shallow depth is reconciled with the lithospheric source. In sum, the Jigede intrusion was likely derived from partial melting of a shallow lithospheric mantle source that had been modified by slab-released fluids, with subsequent cumulation of plagioclase.

Tectonic Implications

Previous studies have identified a giant flare-up of Permian magmatic rocks in the Alxa Block (Shi et al., 2012; Zhang et al., 2013; Dan et al., 2014; Lin et al., 2014, 2017a, 2017b). The predominance of granitoids (ca. 280 Ma) has been considered to be reconciled with a silicic igneous province. Dan et al. (2014) further proposed that the Alxa silicic igneous province was produced by the ca. 280 Ma Tarim mantle plume activities. Indeed, the Alxa silicic igneous province is smaller (ca. 0.05 Mkm2) than other typical silicic igneous provinces, which generally cover an area extent of > 0.10 Mkm2 (Bryan, 2007). More importantly, high-T continental flood basalts, komatiitic sequences, and mafic intrusions are not absent in the Alxa Block. Thus, the occurrence of Permian magmatic rocks cannot be attributed to mantle plume activities. Shi et al. (2014) advocated that the Permian magmatic rocks were emplaced in a post-collisional setting. Several Middle to Late Permian (266 to 250 Ma) intrusions (Baogeqi, Hurentaolegai, Wuliji and Sharijimiao) from the Shalazhashan belt are bimodal rock associations with high-K calc-alkaline affinities, considered to have been emplaced in post-collisional settings (Shi et al., 2014). However, bimodal magmatic suites are not restricted to the post-collision setting. It is also plausible for the emplacement of bimodal magmatic suites in a back-arc basin or hinterland of an active continental margin above the subducted slab (Liu et al., 2015). Recently, on the basis of whole-rock Nd and in-situ zircon Hf isotopic data on magmatic rocks, a significant variation occurring at between 280 and 265 Ma has been identified (Liu et al., 2017a, 2017b). The authors proposed that the sudden change in the Nd-Hf isotopic ratios was in response to a tectonic switch from an oceanic subduction setting to a post-collisional setting. In addition, slab roll-back can also lead to the isotopic change (Sun et al., 2017).

The Jigede intrusion is characterized by HFSEs depletions and negative Th-Nb-Ti anomalies, which are inconsistent with the geochemical features of a plume-related OIB (Fig. 7b). Instead, its petrogenesis revealed that its lithospheric mantle source had been strongly modified by slab-derived fluids. All the samples exhibited arc geochemical signatures, indicative of an oceanic subduction setting. They exhibit relatively low Ti/V ratios (< 20) that can be reconciled with volcanic arc basalts (Fig. 10). We thus argue that the emplacement of the Permian magmatic rocks was more likely to have occurred in subduction setting. This is also supported by the following evidence. First, SHRIMP zircon U-Pb dating of the gabbros from the Quagan Qulu ophiolite yielded a weighed mean age of 275 ± 3 Ma (Zheng et al., 2014). They are significantly enriched in LILEs relative to HFSEs and LREEs, which suggests that they were emplaced in a back-arc basin. Furthermore, dating of deformed porphyries from the Langshan area resulted in emplacement ages of 291.7 to 284.7 Ma (Lin et al., 2014). The deformation was likely associated with subduction-induced compression. The timing of the final closure should be later than the Early Permian. Second, the Late Carboniferous to Middle Permian Bijiertai, Honggueryulin, and Qinggele mafic-ulramafic rocks (306 to 262 Ma) display evident arc geochemical characteristics, implying that the Paleo-Asian oceanic subduction has continued to the Middle Permian (Feng et al., 2013). Third, the Late Paleozoic sedimentary sequence and paleontological assemblages on the northern and southern sides of the Enger Us ophiolite belt are obviously different (Wang et al., 1994). And there is Late Triassic continental molasse exposed in the Alxa Block (Wang et al., 1994), indicating the final closure of the Paleo-Asian Ocean occurred during Late Permian to Late Triassic. To sum up, the Permian magmatic rocks were emplaced in a subduction setting, and the final closure of the Paleo-Asian Ocean did not occur prior to the Middle Permian.

Fig. 10.
figure 10

V vs. Ti/1000 tectonic discriminant diagram (Shervais, 1982) for the Jigede intrusion. Also shown are the fields for volcanic arc basalts (VAB), mid-ocean-ridge basalt (MORB)/back-arc basin basalts (BABB), continental flood basalts (CFB), and ocean-island basalts (OIB)/alkali basalts (AB) (Rollinson, 1993).

Full size image

CONCLUSIONS

(1) The Jigede intrusion from the Shalazhanshan belt was emplaced in the Middle Permian (~262 Ma).

(2) The Jigede intrusion is gabbro and exhibits arc geochemical characteristics. It was likely to have resulted from partial melting of a shallow lithospheric mantle source modified by slab-released fluids, with subsequent cumulation of plagioclase.

(3) The emplacement of the Jigede intrusion was in response to the Paleo-Asian oceanic subduction.

(4) The final closure of the Paleo-Asian ocean did not occur prior to the Middle Permian.