Introduction

The Altaids, also known as the Altaid Tectonic Collage (Şengör et al. 1993) or the southern Central Asian Orogenic Belt (CAOB), was formed by the largest and longest-lived accretion and convergence of the Kazakhstan, Mongolia and Tarim–North China collage systems (Xiao et al. 2015a). The Paleo-Asian Ocean (PAO) recorded in the Altaids was a long-lived ocean, which was opened during the breakup of the Rodinia supercontinent in the Mesoproterozoic–Neoproterozoic (ca. 1300–700 Ma) (Coney 1992; Dobretsov et al. 2003; Li et al. 2008a; Wan et al. 2018), and its subduction made the Central Asian continent generating considerable continental growth. Therefore, a systematical study on the earlier evolution history of the PAO will shed light on not only the tectonics and continental growth of the Altaids but also the reconstruction of the Rodinia supercontinent.

While the closure timing of the PAO has been extensively studied and relatively well constrained either in the Permian (Liu et al. 2020) or even in the Triassic (Ao et al. 2021; Xiao et al. 2015b), the start of the PAO is not well constrained. Several different views about the start of the PAO have been proposed (Berzin and Dobretsov 1994; Khain and Bozhko 1988; Mossakovskii et al. 1993; Şengör et al. 1993; Șengör and Natal'in 1996), with possible ages only from the late Neoproterozoic to Paleozoic, and reconstructed the general historical framework of the PAO at ∼1000–650, ~ 650–510, and ~ 510–450 Ma (Khain et al. 2003). The evidence found in an ophiolite in the accretionary wedge as well as corresponding arc formation in the northern Altaids indicates the evolution of the PAO had begun at least ~ 1020 Ma ago (Rytsk et al. 2007). The ophiolites at 917 ± 14 Ma in northern Transbaikalia (Gordienko et al. 2009) and at 800 ± 3 Ma in northern Mongolia (Kuzmichev et al. 2005) demonstrate the existence of the PAO since the early Neoproterozoic.

However, the location of the PAO’s major oceanic basin and tectonic history of the CAOB remain controversial, particularly because there is no direct evidence for its early evolution, which hampers a better understanding of the tectonics of the Altaids and the reconstruction of the Rodinia supercontinent. Recently, He et al. (2018) reported in situ O and Hf isotopes of zircons from the Mesoproterozoic (ca. 1.4 Ga) metabasic amphibolites and granitic rocks from the Chinese Central Tianshan microcontinent in the southern Central Asian Orogenic Belt (CAOB), which provides indirect evidence for Mesoproterozoic (ca. 1.4 Ga) crustal growth constituting a considerable Paleoproterozoic supracrustal component in the Central Tianshan block.

Adakites are usually considered to be derived from the partial melting of metamorphosed mafic rocks or the fractional crystallization of arc-related mafic magmas (Hastie et al. 2021) and therefore, are considered as direct diagnostic evidence for subduction and continental growth. This work focuses on our newly discovered adakite-like gneissic granite at Xingxingxia in the Eastern Tianshan, NW China, in the southern Altaids. In this study, we present zircon U–Pb age and in situ Hf isotopic composition and whole-rock geochemistry of the Xingxingxia gneissic granite, which exhibits adakitic characteristics and is Neoproterozoic in age (ca. 869 Ma). The new discovery of Neoproterozoic adakitic rocks enables us to address the tectonic position of the major Paleo-Asian Ocean and its mature stage with the formation of the subduction zone at ca. 870 Ma, together with its implications for the reconstruction of the Rodinia supercontinent.

Geological background and sampling

The vast Altaids is situated between the Siberian Craton to the north and the Tarim Block to the south and was formed by the amalgamations and accretions of massive continental margin arcs, microcontinents, island arcs, ophiolites, accretionary wedges (Windley et al. 2010; Xiao et al. 2004, 2010; Zuo et al. 1991). The Tianshan orogenic belt is located in the southern Altaids, and is mainly composed of the North Tianshan suture zone, Yili block, Central Tianshan block and South Tianshan suture zone from the north to the south (Fig. 1a) (Charvet et al. 2011; Kröner et al. 2008; Wang et al. 2014b; Windley et al. 2007; Xiao et al. 2010a).

Fig. 1
figure 1

a Simplified tectonic map of the Central Asian Orogenic Belt (CAOB: modified after (Jahn et al. 2000; Kröner et al. 2008; Şengör et al. 1993; Xiao and Santosh 2014). b Sketch map of the Tarim craton showing Precambrian geology and surrounding orogens, including the extensive Mesoproterozoic–Neoproterozoic (~ 1436–872 Ma) arc magmatism, the Eastern Tianshan and Beishan (modified after (He et al. 2019; Xiao and Santosh 2014; Xu et al. 2013; Zhao et al. 2021)

Full size image

With a long-term accretionary orogeny, several E–W-striking accretionary belts and faults were developed from north to south and the Eastern Tianshan of the Tianshan orogenic belt is composed of several blocks juxtaposed by ophiolitic mélanges (Mao et al. 2021b). In the Eastern Tianshan, the Central Tianshan block is separated from the Tarim block to the south and the North Tianshan tectonic zone to the north by the Kumishi–Hongliuhe suture and the Aqikkudug–Weiya suture, respectively (Guo and Li 1993; Shu et al. 2004) (Fig. 1b).

In the Beishan, several ophiolitic belts were recognized and regarded as the suture zones between the blocks/arcs (Liu and Wang 1995; Nie 2002; Wang et al. 2014a; Xiao et al. 2010b; Zuo 1990; Zuo et al. 1991), which include the Hongshishan, Xingxingxia, Hongliuhe and Liuyuan belts that occur along large-scale faults named after these ophiolitic belts, respectively (Fig. 2). The Hanshan block was proposed to be the eastern part of the eastern Central Tianshan (Liu and Wang 1995; Xiao et al. 2010b; Zuo et al. 1991).

Fig. 2
figure 2

The schematic geological map of Xingxingxia region in Eastern Tianshan (Mao et al. 2019; Xiao et al., 2004b) showing the locations of the gneissic granite

Full size image

The Central Tianshan block comprises mainly of Precambrian basement with minor amounts of the early and late Paleozoic volcano-sedimentary formations (Hu et al. 2000; Li and Kusky 2007; Liu et al. 2004; Shu et al. 2004; Xiao et al. 2004; Yang et al. 2008) and a number of granitoid intrusions (Fig. 2). The basement contains the Mesoproterozoic Xingxingxia, the Neoproterozoic Tianhu and the Kawabulake Groups, and was subjected to greenschist to amphibolite facies metamorphism, which shows fault and/or unconformable contact relationships to each other (Gao et al. 1993). The widely distributed granitoid intrusions of Proterozoic to Triassic ages mostly share a strong affinity to I-type granites (Gu 2006; Gu et al. 1999; Hu et al. 2007; Wang et al. 2006; Wu et al. 2006; Zhang et al. 2005; Zhang et al. 2007b).

The Xingxingxia Group consisted of banded and augen gneissic granites, marbles, amphibolites, migmatites, quartzites and schists that have experienced metamorphism from greenschist to amphibolite facies, or even to granulite facies at some localities (Li et al. 2007, 2008b; Liu et al. 2004). Hu et al. (2000) suggested that the gneissic granite and amphibolites in the Xingxingxia Group are metamorphosed granitic–mafic plutons or volcanic rocks. The Tianhu Group, located between the Tianhu and Xingxingxia area, which is composed of metamorphic volcanic rocks, carbonates and clastic rocks, with the main components of amphibolite schists, plagioclase gneisses, chloritic quartz schists and migmatites. It shows fault unconformable contact with the Kawabulag Groups (Deng et al. 2017; Liu et al. 2004). The Kawabulake Group generally shows as a shallow sea silica-rich calcium–magnesium carbonate sedimentary sequence interspersed with terrigenous clastic rocks and partly rich in phosphorus, and represented by a suite of metamorphic rocks, mainly marbles, crystalline limestones, silicified limestones, wollastonite schists, and quartz schists (Chen 2006).

The gneissic granite in this study crops out in the south of the Xingxingxia town (E95°8′3.05441″, N41°46′57.28908″), displaying intrusive contacts with the Xingxingxia Group and fault contacts with the Jijitaizi mélange. The gneissic granite is medium- to coarse-grained in texture and gneissic in structure (Fig. 3a, b), with light layers composed of granular deformed and broken quartz and feldspar (Fig. 3c, d).

Fig. 3
figure 3

Petrographic photos of the studied samples. a and b Close-ups of the Xingxingxia gneissic granite specimens; c photomicrograph of the representative sample 20YY22-1 (in cross-polarized light); d photomicrograph of the representative sample 20YY22-1 (in plane-polarized light). Bt biotite, Pl plagioclase, Kfs K-feldspar, Qtz quartz

Full size image

Analytical methods

Major elements were determined by X-ray fluorescence spectrometry (XRF), and trace elements by inductively coupled plasma techniques (ICP) at the Geological Test and Analysis Center of the Beijing Research Institute of Uranium Geology. Detailed procedures can be found in Mao et al. (2018).

The geochronological experiments on zircon U–Pb were adopted at Beijing Quick-Thermo Science & Technology Co., Ltd., using an ESI New Wave NWR 193UC (TwoVol2) laser ablation system connected to an Agilent 8900 Inductively Coupled Plasma Mass Spectrometer (ICP-MS). The analytical procedures were followed Ji et al. (2020). The spot size and frequency of the laser were set to 20 µm and 5 Hz, respectively. Individual zircon grains (mounted and polished in epoxy) were ablated in a constant stream of He that is mixed downstream with N2 and Ar before entering the torch region of the ICP–QQQ. After warmup of the ICP–QQQ and connection with the laser ablation system, the ICP-MS is first tuned for robust plasma conditions by optimizing laser and ICP–QQQ setting, monitoring 232Th16O+/232Th+ ratios (always ≤ 0.2%) and 238U+/232Th+ ratios (always between 0.95 and 1.05) while ablating NIST SRM 610 in line scan mode. The 91,500 zircons were used as primary reference material for all U–Pb age determinations, while zircon PleŠovice was used as secondary reference. NIST610 glass was used to calibrate trace element with internal standard major element Si. The standards were analyzed two times before and after each analytical session including 6–8 spots on minerals. Background subtraction and correction for laser downhole elemental fractionation were performed using the Iolite data reduction package within the Wavemetrics Igor Pro data analysis software (Paton et al. 2011). Concordia plots were processed using ISOPLOT 4.15. Detailed operating conditions and data acquisition procedures were described by Ji et al. (2020).

Zircon Lu–Hf in situ Lu–Hf isotope measurements were performed using a Thermo Finnigan Neptune-plus MC–ICP-MS fitted with a J-100 femto-second laser ablation system Applied Spectra Inc., housed at the Beijing Chron Technology Co., LTD, Beijing, China, which follow the analytical procedures and calibration methods are similar to those described by Wu et al. (2006). Zircons were ablated for 31 s at a repetition rate of 8 Hz at 16 J/cm2, and ablation pits were ~ 30 μm in diameter. During analysis, the isobaric interference of 176Lu on 176Hf was negligible due to the extremely low 176Lu/177Hf in zircon (normally < 0.002). The mean 173Yb/172Yb value of individual spots was used to calculate the fractionation coefficient (βYb) and then to calculate the contribution of 176Yb to 176Hf. An isotopic ratio of 173Yb/172Yb = 1.35274.

Whole-rock geochemistry: major and trace elements

Eight samples from the gneissic granite are geochemically analyzed and the results are listed in Table 1.

Table 1 Major and trace elements of the Xingxingxia gneissic granite
Full size table

The samples from the gneissic granite exhibit homogeneous geochemical characteristics, including high SiO2 (70.66–73.92 wt.%), K2O (2.80–3.93 wt.%), Na2O (3.48–4.11 wt.%) and Al2O3 (14.07–15.09 wt.%), and low TiO2 (0.19–0.28 wt.%), MgO (0.41–0.67 wt.%) and TFe2O3 (1.34–1.95 wt.%). They display high-K calc-alkaline features (Fig. 4a, b) and fall into the similar fields as the lower crust-derived adakitic rocks from the Cordillera Blanca batholith (Petford and Atherton 1996) and Kanguer accretionary complex (Mao et al. 2021a) in the K2O–SiO2 diagram. Their A/CNK values range from 1.02 to 1.08 (Fig. 4b), suggesting their weakly peraluminous affinities.

Fig. 4
figure 4

 a K2O vs. SiO2 diagram (Peccerillo and Taylor 1976); The adakitic intrusive rocks in this study show K2O contents similar to those of the Cordillera Blanca batholith (Petford and Atherton 1996) and Kanguer accretionary complex (234–242.5 Ma) (Mao et al. 2021a, b), but higher than those of the slab-melted adakites in the Eastern Tianshan (Mao et al. 2019). b Al/(Na + K) vs. Al/(Ca + Na + K) diagram after (Maniar and Piccoli 1989)

Full size image

The samples have low Cr (6.64–12.93 ppm) and Ni (1.20–11.54 ppm), while their anomaly in Sr (202.6–343.4 ppm) and low Y (3.00–6.36 ppm) result in high Sr/Y ratios (47–114). In the primitive mantle-normalized multi-element patterns (Fig. 5a), the gneissic granite samples show obvious enrichment in Rb, Th, K, Pb, and Sr and depletion in Nb, Ta, P, and Ti, with slightly positive Zr–Hf anomalies. In the chondrite–normalized element rare earth element (REE) patterns, they are highly enriched in LREE((La/Yb)N = 7.64–9.86), and show positive Eu anomalies (Eu/Eu* = 0.96–1.45) (Fig. 5b).

Fig. 5
figure 5

εHf(t) a The primitive mantle-normalized multi-element diagrams and b the chondrite-normalized rare earth element (REE) patterns for the gneissic granite. The data for the N-type MORB, E-MORB and OIB are from Sun and McDonough (1989)

Full size image

Zircon U–Pb ages and in situ Lu–Hf isotopes

One gneissic granite sample (20YY22) was analyzed for zircon U–Pb ages and in situ Lu–Hf isotopic compositions. The analytical results are listed in Tables 2 and 3. The spot location for Lu–Hf isotopes was the same as that for zircon U–Pb analyses. Zircon grains separated from the gneissic granite are transparent, euhedral to subhedral and 80–100 μm long, with length/width ratios of 1:1 to 3:1.

Table 2 U–Pb isotopic data for multigrain analyses of zircons from gneissic granite
Full size table
Table 3 Hf isotopic data for the zircons of the Xingxingxia gneissic granite
Full size table

Twelve zircons from the gneissic granite yielded unconcordant ages due to the Pb-loss. The remaining eight zircons have concordant ages and exhibit distinct oscillatory zoning with high Th/U ratios (0.47–0.93), consistent with a magmatic origin. They yielded 206Pb/238U ages of 856–889 Ma with a weighted mean age of 869 ± 9 Ma (MSWD = 2.5, 95% conf.) (Fig. 6a, b), suggesting that the gneissic granite emplaced in the Neoproterozoic.

Fig. 6
figure 6

a K2O vs. SiO2 (Peccerillo and Taylor 1976); K2O (Petford and Atherton 1996) (Mao et al. 2021a, b), (Mao et al. 2019). b Al/(Na + K) vs. Al/(Ca + Na + K) (Maniar and Piccoli 1989) a Zircon representative CL image and U–Pb Concordia diagram and b age histogram of the Xingxingxia gneissic granite

Full size image

Zircons from the gneissic granite possess similar Hf isotopic compositions, with 176Hf/177Hf ratios of 0.282232–0.282387 and positive εHf (t) values (+ 1.9 to + 4.7) (Fig. 7; Table 3), indicating that the rock was mainly derived from the juvenile materials in the Paleo- to Mesoproterozoic Fig. 8.

Fig. 7
figure 7

Diagram of zircon εHf(t) versus age (1989)

Full size image
Fig. 8
figure 8

a Sr/Y vs. Y diagram. The fields for adakites and classical island arc magmatic rocks are from (Defant and Drummond 1990); b (La/Yb)N vs YbN diagram. The fields for the adakites and classical island arc magmatic rocks are from Martin (1999)

Full size image

Discussion

Petrogenesis of the adakitic rocks

Adakites are a type of intermediate-acid magmatic rock with distinctive geochemical properties and significant geodynamic and metallogenetic implications (Wang et al. 2020). It is used to describe andesitic–rhyolitic extrusive to intrusive rocks having features including SiO2 ≥ 56%, Na rich (Na2O ≥ 3.5%), high Al (Al2O3 ≥ 15%) and Sr (> 400 ppm), low Y (≤ 18 ppm), strongly fractionated REE patterns (HREE depleted), positive Eu and/or Sr anomalies, as well as high Sr/Y and (La/Yb)N ratios (≥ 40 and 20, respectively). The Neoproterozoic Xingxingxia gneissic granite (869 ± 9 Ma) from the eastern Central Tianshan (NW China) contains high Sr/Y (47–114) and low levels of Y (3.00–6.36 ppm) and Yb (0.23–0.62 ppm), sharing strong geochemical similarities to adakitic rocks (Defant and Drummond 1990).

Adakites could form as a result of slab melting in a variety of modern subduction-arc settings, which may be related to subduction of young and hot oceanic crust (Calmus et al. 2003; Defant and Drummond 1990; Guivel et al. 1999; Wang et al. 2007; Yogodzinski et al. 1995), active ridge subduction (Aguillón-Robles et al. 2001; Bourgois and Michaud 2002; Johnston and Thorkelson 1997; Rogers et al. 1985; Reich et al. 2003; Thorkelson and Breitsprecher 2005), initiation of subduction (Sajona et al. 1993), flat subduction (Gutscher et al. 2000), and partial melting of lower crust of thickened arc (Defant 2002; Petford and Atherton 1996). However, several other mechanisms were also proposed for their generation, such as the differentiation of mantle-derived basaltic magmas (Castillo et al. 1999; Macpherson et al. 2006), magma mixing (Chen et al. 2013), and partial melting of lower crustal, hydrated mafic rocks in magmatically or tectonically thickened crust or partial melting of delaminated lower crust (Atherton and Petford 1993; Jiang et al. 2012; Kay and Kay 2002; Kay and Mpodozis 2002; Kay et al. 1993; Muir et al. 1995; Peacock et al. 1994; Windley et al. 2010; Xiong et al. 2001; Zhang et al. 2002; Zhao et al. 2006).

The Xingxingxia gneissic granite contains high SiO2 (> 70 wt.%) and low MgO (< 0.7 wt.%) contents, as well as a narrow range of major and trace element compositions (Table 1). Additionally, the pluton lacks a mafic endmember and contemporaneous mafic rocks are absent in the study area, implying hypotheses of fractional crystallization of mantle-derived magmas or mixing of mantle- and crust-derived magmas inadequately account for its genesis.

Furthermore, the low Mg# (37–42), Cr (6.64–12.93 ppm) and Ni (1.20–11.54 ppm) contents of the Xingxingxia gneissic granite are also inconsistent with their derivation by slab melting. The gneissic granite has relatively high K2O contents, which is comparable to those of the adakitic rocks derived from the thickened lower curst of the Andes arc in the Peru (Petford and Atherton 1996) and the Dananhu arc (Mao et al. 2021a) (Figs. 4a and 9a) but higher than the slab-derived adakites in the Eastern Tianshan (Mao et al. 2021a, b) (Fig. 4a). In combination with its moderate positive Hf isotopic compositions, we suggest that the gneissic granite formed as a result of partial melting of juvenile lower crust of the eastern Central Tianshan arc. The very low HREE and relatively low CaO, TFe2O3 and Al2O3 contents of the Xingxingxia gneissic granite further imply that it was produced from thickened lower crust (Yang et al. 2020). In the Rb vs. Y + Nb diagram, the Xingxingxia gneissic granite plots in the field of volcanic granites (Fig. 9b).

Fig. 9
figure 9

a MgO vs. SiO2 diagram for the adakitic rocks and experimental basaltic partial melts. Data sources and legends are the same as in Fig. 5a; b Discrimination diagram of tectonic setting for the Xingxingxia gneissic granite. Symbols are the same as those in Fig. 4

Full size image

Implications for long-lived subduction and crustal growth

To better constrain the source and tectonic setting of the Xingxingxia adakitic rock, it is necessary to compare it with those from other orogens where the similar thickened lower crust-derived adakitic rocks were reported (Fig. 10), such as the Early Cretaceous adakitic rocks in Eastern China that were derived from thickened or delaminated lower crust (Wang et al. 2004; Xu et al. 2002). Our adakitic rock tends to be more depleted in isotopes than those adakitic rocks formed in intra-continental or collisional settings in which the orogen experienced tectonically thickening or lower-crustal delamination.

Fig. 10
figure 10

Paleogeographic reconstructions of the northern Rodinia at ca. 860 Ma after (Zhao et al. 2021). The reconstructions are based on the paleolatitude of each block also accounting for geological evidence. Final closure of Jiangnan–Central Tarim Ocean, approaching amalgamation of the South China and Tarim, and the (relative) paleogeographic position the Central Tianshan separated from the Tarim and South China by the Circum–Rodinia Subduction Zone were reconstructed between ~ 870 and 820 Ma. This map is in Mollweide projection

Full size image

As another lower-crustal melting example, the Late Miocene Cordillera Blanca batholith adakitic rocks in Peru were generated by the thickened Andean arc (Petford and Atherton 1996) and the magmas formed by partial melting of newly underplated materials that were quickly melted following the formation of underplate to produce the Cordillera Blanca Na-rich suite (Petford and Atherton 1996). This case for the generation of adakitic rocks, indeed, reflects the process of continental crustal growth. The Xingxingxia gneissic granite geochemically comparable to the adakitic rocks of the Cordillera Blanca batholith (Fig. 4a), and it has relatively depleted Hf isotopic composition (Fig. 7), representing the product of subduction-crustal growth during the Neoproterozoic.

The geochemical properties of the Xingxingxia gneissic granite, which include enrichment of LILE and LREE and depletion of HFSE and HREE with evidently negative Nb and Ta anomalies, are consistent with those of magma in an island arc setting (Fig. 5).

Therefore, the southern PAO had its accretionary tectonic history that may have been extended into the Neoproterozoic. Our newly found adakitic rocks indicate that the PAO may have evolved into a mature stage at ca. 870 Ma along the Central Tianshan Arc. This is consistent with previous studies suggesting that the PAO may have begun at least ~ 1020 Ma ago (Rytsk et al. 2007; Gordienko et al. 2009; Kuzmichev et al. 2005).

It is well known that the PAO generally became younger southward, and in the Late Paleozoic PAO had several branches in the southern Altaids, including the ones around the Central Tianshan Arc (Şengör et al. 1993; Windley et al. 2007; Xiao et al. 2015a, b). Our work, together with some previous data, shows that the branches around the Central Tianshan Arc had a long evolutionary history at least from ca. 870 Ma lasting to the Late Paleozoic and even Triassic (Chen et al. 2019; Ao et al. 2021; Mao et al. 2021a). Actually, other PAO branches north to the Tianshan, such as the Erqis and other sutures in the Junggar and adjacent areas may have been closed earlier (Wang et al. 2012; Xiao et al. 2015a; Li et al. 2015; Yang et al. 2015; Song et al. 2020; Guy et al. 2020). Therefore, the oceanic slabs beneath the Central Tianshan arc could have represented one of the major ocean(s) of the PAO with long-lived subduction-related accretion, which generated a substantial foundation for considerable Phanerozoic continental growth (Şengör et al. 1993; Windley et al. 2007; Xiao et al. 2015a, 2018).

Tectonic scenarios and implications for northern Rodinia reconstruction

Based on the identification of 1.4–0.9 Ga high grade metamorphism (Hu et al. 2006, 2010; Spencer et al. 2017) and paleomagnetic reconstructions (Evans 2013; Li et al. 2008a; Wen et al. 2018), the configuration of the earliest Rodinia orogenesis could be the reconstruction of the Umkondo large igneous province with the aggregated of Kalahari, Amazonia, Congo, and other cratons of the southern Rodinia by the time of ca. 1.1 Ga, which has been referred to as “Umkondia” (Choudhary et al. 2019). Additionally, during the Grenvillian orogenesis (1.3–1.0 Ga), the core of Rodinia was rebuilt with the assembly of Australia–East Antarctica, Amazonia (of Umkondia), and Baltica, which were located along the western, eastern, and northeastern margins of the Laurentia, respectively (Hoffman 1999; Li et al. 2008a).

Even the Cathaysia and South Tarim, potentially adjacent or connected to each other, likely joined Rodinia to the north of Australia in the early Neoproterozoic. The Central Tianshan, South China and Tarim had not yet amalgamated as discrete cratons (Cawood et al. 2013; Wen et al. 2018). The evidence for the subduction of the Jiangnan–Central Tarim Ocean is found in the Neoproterozoic arc magmatism along multiple active margins of Yangtze, Cathaysia, and North Tarim, demonstrating that the Beishan, as a part of the Central Tianshan, is still separated from Tarim and South China by the Circum–Rodinia subduction zone after the closure of the Jiangnan–Central Tarim Ocean between 870 and 820 Ma (Fig. 10) (Yao et al. 2019; Zhao et al. 2011).

The Eastern Tianshan consists of multiple arcs, accretionary complex belts and microcontinental blocks and is crucial for understanding the earlier accretionary and collisional processes of the southern Altaids (Muhtar et al. 2020; Wang et al. 2014a; Xiao et al. 2015a). The polarity of this subduction zone or whether two-sided subduction was developed remain unresolved due to the scarcity of evidence for the Central Tarim suture. There are several competing models for the collision between Yangtze and Cathaysia, ranging from northwestward subduction beneath the Yangtze (Yan et al. 2019; Zhao et al. 2011) to southeastward subduction beneath the Cathaysia (Cawood et al. 2013; Yao et al. 2019) to two-sided subduction beneath both Yangtze and Cathaysia (Zhao 2015). In any case, there is evidence following peripheral accretion of the Yangtze and North Tarim, a circum–Rodinia subduction girdle with arc magmatism developed along the external margins of the South China, Tarim, and Australia (Cawood et al. 2018) (Fig. 10).

While the Eastern Tianshan includes several microcontinents or arcs with Precambrian basement, the adakite may suggest that its tectonic affinity was proposed as the following three main scenarios (A, B and C) in view of the complicated evolutionary histories of the Rodinia and Gondwana. They are the three potential kinematic solutions for the final assembly of Rodinia along its northern margin.

In scenario A (Andean-type arc), the adakite was intruded into the magmatic arc or back-arc position of the northern Tarim carton. This occurred as the consequence of retreating subduction in the North Tarim block at ca. 730 Ma, with the opening of the Proto Tethys Ocean and South and North Altyn Oceans resulted by the breakup of the Rodinia (Ma et al. 1997; Wang et al. 2022) (Fig. 10).

In scenario B, there was a Japan-type arc composed of subduction-related rocks mostly on an oceanic plate with some recycled materials in the PAO (Fig. 10).

In scenario C (Mariana-type arc), the subduction also possibly occurred in an intra-oceanic setting in the vast PAO like what happens today along the Mariana subduction zone in the western Pacific Ocean, developing juvenile magmatic arcs in an oceanic plate (e.g., Yao et al. 2021).

Subduction zone develops various types of island arc system, including the Andean type, Japan type, Mariana type, and Alaska type, depending on the attributes of the upper plate in subduction system (Xiao et al. 2010). The Andean type, Japan type, and Mariana type can be listed as the scenarios A, B and C, respectively. The Alaska type is taken from the northern North American craton where the Alaska extends to connect the Aleutian arc south to the Bering Sea, which is a representative kind of mixture type of the Andean type, Japan type, and Mariana type.

Based on the chemical characteristics, such as obvious negative anomalies in Nb, Ta and Ti elements, and Hf and Nd isotopic compositions, all the Xingxingxia granitoids in 1.4 Ga and 0.9 Ga from the Western and Eastern Tianshan were suggested to be generated in a continental margin tectonic setting and formed by remelting of Paleoproterozoic rocks, followed by advanced fractional crystallization, probably forming a part of the Rodina during the early Neoproterozoic period (Hu et al. 2006, 2010).

The Hf isotopes in this study show relatively lower positive εHf(t) signature, contrasting to an intro-oceanic subduction-arc system like Mariana type, which contains highly positive εHf(t) feature close to juvenile or depleted mantle source. The northern margin of the Precambrian Tarim carton, on the contrary, shows dominantly negative εHf(t) and εNd(t) isotopic feature due to incorporating large amounts of reworked crustal material in an Andean-type subduction (Ge et al. 2014; He et al. 2013; Long et al. 2015; Zhang et al. 2007a), e.g., the lower crust-derived adakitic granites (795–820 Ma) in the northeastern Tarim have the εNd(t) values varying from  – 12.7 to  – 17.3 (Zhang et al. 2007a). Given the fact that the Central Tianshan was constructed by newly deformed arcs, accretionary complexes and ancient microcontinental blocks, juvenile and reworking (remelting) material could serve as the source for the composite isotopic compositions and result in a relative (intermediate) positive εHf(t) feature (Hu et al. 2000). Therefore, we exclude scenario A and favor scenario B in which the arc on which our adakite was mostly a Japan-type one. Of course, it may also have been an Alaska type.

On the basis of the foregoing facts and arguments, we can postulate that, when the adakitic rock of this study was intruded at ca. 869 ± 9 Ma, the Central Tianshan arc had not yet amalgamated with the Tarim block to the south (Fig. 10), which had relatively juvenile crust in the Neoproterozoic. On the other hand, the Eastern Tianshan as a peripheral island arc in the PAO developed to the north of the Yangtze and North Tarim, was assembled along the northern margin of the Rodinia, heralding the long-lived of arcs in the PAO following the final amalgamation of the Neoproterozoic supercontinent.

Our work suggests that there was an arc in this part of the PAO as a Japan type or as an Alaska type. Therefore, the arc growth in the Xingxingxia area was most possibly taken root in the Central Tianshan, indicating that the subduction of the PAO might have taken place in the Neoproterozoic (Fig. 11).

Fig. 11
figure 11

Schematic tectonic model for the evolution of the adakites in the Xingxingxia area, eastern Tianshan (modified after Mao et al. (2012))

Full size image

Conclusions

  1. (1)

    Our new zircon U–Pb dating reveals that the Xingxingxia gneissic granite from the eastern Central Tianshan (NW China) was formed in the Neoproterozoic (869 ± 9 Ma).

  2. (2)

    The geochemical data indicate that the Xingxingxia gneissic granite has high Sr/Y, low Y and Yb and strongly fractionated REE patterns, indicating an adakitic affinity. It was derived from thickened juvenile arc lower crust and represents the products of Neoproterozoic subduction and corresponding continental crustal growth.

  3. (3)

    The subduction of the Paleo-Asian oceanic slab beneath the Central Tianshan occur at least in the Neoproterozoic, which resulted in the crustal thickening and melting of the thickened crust. Therefore, it is a solid and key evidence to show that the tectonic position of the major Paleo-Asian Ocean was located in the South/North Tianshan, where the paleo-oceanic basin evolved into its mature stage with the subduction beneath the Central Tianshan at ca. 870 Ma.

  4. (4)

    Our work also sheds light on the reconstruction of the Rodinia and the interaction of the Paleo-Asian Ocean with the Jiangnan–North Tarim Ocean.