Petrogenesis and tectonic implications of the early Mesozoic granitoids in the northern Alxa region, Central Asian Orogenic Belt

Abstract

The northern Alxa region is located in the central segment of the southern Central Asian Orogenic Belt. Many controversies and deficiencies still exist regarding the magma source characteristics, petrogenesis and tectonic regimes during the late Palaeozoic – early Mesozoic period within this region. This study presents whole-rock compositions and zircon U–Pb and Lu–Hf isotopic data for three early Mesozoic I- and A-type granitic plutons occurring in the northern Alxa region. The Haerchaoenji and Chahanhada I-type granitoids yielded zircon ²⁰⁶Pb–²³⁸U ages of 245 ± 5 Ma and 245 ± 2 Ma, respectively. The variable positive zircon ϵ_Hf(t) values between +1.8 and +11.8, with young T_DM ages of 425–837 Ma, indicate that these I-type granitoids were mainly derived from juvenile crustal materials. The Wulantaolegai pluton has a zircon ²⁰⁶Pb–²³⁸U age of 237 ± 2 Ma and is classified as having high-K calc-alkaline A-type affinity. Furthermore, the positive zircon ϵ_Hf(t) values of the Wulantaolegai granite range from +3.3 to +8.7 with young T_DM ages of 545–778 Ma, suggesting the involvement of a juvenile crustal source as well. Furthermore, the major-element compositions of the Chahanhada and Wulantaolegai granites suggest the input of metasedimentary components. Geochemically, the Haerchaoenji and Chahanhada I-type granitoids show an arc affinity, while the Wulantaolegai granite exhibits a post-collisional affinity. However, with regional data, we suggest that the Haerchaoenji and Chahanhada I-type granitoids were also emplaced in a post-collisional setting, and the arc affinity was probably inherited from recycled subduction-related materials. These lines of evidence obtained in this study enable us to argue that the Palaeo-Asian Ocean in the central segment of the Central Asian Orogenic Belt closed before Middle Triassic time.

1. Introduction

The Central Asian Orogenic Belt (CAOB), which is located in northcentral Asia from the Uralides to the Pacific Ocean (e.g. Şengör et al. 1993; Jahn et al. 2004; Windley et al. 2007; Li et al. 2013; Xiao et al. 2015; Liu et al. 2016) (Fig. 1a), has been regarded as one of the world’s largest and most complex accretionary orogens (Şengör et al. 1993; Windley et al. 2007; Xiao et al. 2009; Wilhem et al. 2012). The CAOB has been widely considered to have undergone long-lived, giant orogenic processes driven by the evolution and closure of the Palaeo-Asian Ocean (PAO) during the Neoproterozoic to Mesozoic period (Şengör et al. 1993; Jahn et al. 2004; Cope et al. 2005; Windley et al. 2007; Shen et al. 2009; Zhang et al. 2009; Cai et al. 2011 a,b; Li et al. 2013, 2016 a,b, 2017; Xu et al. 2013; Xiao et al. 2013, 2015; Wang et al. 2017; He et al. 2018; Song et al. 2018 a,b; Chen et al. 2019; Zhao et al. 2020).

Fig. 1. (a) Schematic geological map of the Central Asian Orogenic Belt (modified after Liu et al. 2017). (b) Geological map of the northern Alxa region (modified after 1:200 000 geological maps from BGMRIM, 1991).

Numerous studies have focused on the multi-stage evolution of the PAO and CAOB, with significant progress made (e.g. Şengör et al. 1993; Windley et al. 2007; Xiao et al. 2009, 2015; Wilhem et al. 2012; Eizenhöfer et al. 2014; Eizenhöfer & Zhao, 2018). However, the timing of the final closure of the PAO is still debated, with estimates ranging from Late Devonian to Triassic time (e.g. Charvet et al. 2011; Xu et al. 2013; Eizenhöfer et al. 2014, 2015 a,b; Xiao et al. 2015, 2018; Zhang et al. 2015 a,b; Zhang, W. et al. 2015; Shi, G. Z. et al. 2016; Yin et al. 2016; Song et al. 2018 b). These controversies are mainly due to: (1) the various objects studied, such as the late Palaeozoic magmatic rocks (e.g. Shi et al. 2012, 2014 a,b; Xiao et al. 2015; Liu et al. 2017, 2018; Song et al. 2018 a), tectonic deformation and regional unconformity (e.g. Tang, 1990; Xu et al. 2013; Xu, X. Y. et al. 2014), or detrital zircon indicators (e.g. Chen et al. 2019; Song et al. 2018 b, 2021; Niu et al. 2021); (2) limited study areas (different segments probably closed at diverse times); and (3) relatively poor study in some areas because of execrable natural conditions, e.g. the northern Alxa region. Actually, the CAOB evolved with multiple convergences and the accretion of many orogenic components during multiple phases of amalgamation (Xiao et al. 2015), i.e. the closure of the PAO was probably diachronous. Furthermore, previous studies of magmatic rocks mainly focused on the Tianshan–Beishan in the western segment (e.g. Yang et al. 2014; Zhang, W. et al. 2015; Tian et al. 2017) or Inner Mongolia in the eastern segment (e.g. Jian et al. 2008, 2010; Chen et al. 2009; Xu et al. 2013; Li et al. 2016 a,b, 2017; Shi, Y. R. et al. 2016; Zhao, P. et al. 2017) along the southern CAOB. Much less is known, however, about the central segment of the southern CAOB (the northern Alxa region), which is a crucial junction between the North China Block (NCB) and the Tarim Block (Fig. 1a). It has hampered us from better understanding the evolutionary history of the PAO and subsequent development of the CAOB. In the central segment of the southern CAOB, the late Palaeozoic magmatic rocks are widely exposed and have attracted the attention of many scholars (e.g. Shi et al. 2014 a,b; Zhang et al. 2017; Liu et al. 2017, 2018; Song et al. 2018 a; Zhao et al. 2020). Nevertheless, the timing of tectono-magmatic switching from an arc-related to a post-collisional process is still actively debated. Previous research indicated that this region was in a subduction setting during most of the late Palaeozoic period (e.g. Shi et al. 2014 a,b; Zhang et al. 2017; Liu et al. 2017, 2018; Song et al. 2018 a; Zhao et al. 2020). Therefore, the earliest Mesozoic should be a key period in the evolution of the PAO and probably provides significant information to constrain the tectonic switch from a subduction setting to a post-collisional setting.

Thus, this research focused on the earliest Mesozoic magmatic rocks, which have been rarely reported, exposed in the northern Alxa region of the central segment of the southern CAOB. We report new geochronological, geochemical and isotopic data from three early Mesozoic granitoids in the northern Alxa region and evaluate their petrogenesis and tectonic implications, in order to decipher the evolution of the central segment of the southern CAOB.

2. Geological background

The northern Alxa region is situated in western Inner Mongolia, which borders the NCB to the east separated by the Zunnbayan fault belt and the Langshan fault belt (Fig. 1a) (Huang et al. 1999; Geng & Zhou, 2010; Zhang et al. 2013), and the North Qilian Orogen to the southwest separated by the Longshoushan fault belt (Liu et al. 2017). Largely covered by the Badain Jaran desert, the Alxa Block exposes sporadic Precambrian rocks, Palaeozoic to Mesozoic volcanic and intrusive rocks, and Phanerozoic sedimentary rocks. The Alxa Block is generally considered to be a Precambrian block belonging to the westernmost part of the NCB at present (Fig. 1a). Based on palaeontology, sedimentary sequences and magmatic events, some researchers have argued that the northern Alxa region comprised a complete trench–arc–basin system during late Palaeozoic time (Wang et al. 1994; Zhang et al. 2013). In this region, there are two significant ophiolite belts, i.e. the Qagan Qulu Ophiolite Belt and the Enger Us Ophiolite Belt (Fig. 1b). The Enger Us Ophiolite Belt (∼302 Ma) is regarded as the major suture of the PAO in the northern Alxa region (BGMRIM, 1991; Wang et al. 1994; Wu et al. 1998; Xie et al. 2014; Zheng et al. 2014), and the Qagan Qulu Ophiolite Belt (∼275 Ma) is considered to have been generated in a back-arc setting (Wu et al. 1998; Zheng et al. 2014). Based on these two sutures and the Yagan fault belt, the northern Alxa region can be further subdivided into four units (from north to south): the Yagan Tectonic Belt (YTB), the Zhusileng–Hangwula Tectonic Belt (ZHTB), the Zongnaishan–Shalazhashan Tectonic Belt (ZSTB) and the Nuoergong–Honggueryulin Tectonic Belt (NHTB) (Wu & He, 1992, 1993) (Fig. 1b).

The ZSTB extends southwestward to the Badain Jaran desert and northeastward to the south of the Solonker region in a nearly ENE–WSW direction (Fig. 1b). To the south, the ZSTB borders the NHTB separated by the Qagan Qulu Ophiolite Belt. Northward, the Enger Us fault separates the ZSTB from the ZHTB. Palaeozoic–early Mesozoic plutons are widely exposed in the ZSTB, including voluminous calc-alkaline granitoids and minor gabbro–diorites (e.g. Shi, G. Z. et al. 2016). The geochemical characteristics show that the late Palaeozoic plutonic rocks were mainly involved in the subduction process of the PAO (e.g. Wang et al. 1994; Shi et al. 2014 a,b). The early Mesozoic plutonic rocks are mainly medium–fine-grained monzogranite and K-feldspar granite, which intruded into the pre-Mesozoic rocks as small stocks or branches (Wang et al. 1994). Minor Precambrian rocks are also exposed in the ZSTB, which are mainly composed of metamorphosed supracrustal rocks and meta-intrusive rocks with an age of 1.4∼1.5 Ga (Shi, X. J. et al. 2016). The lower Palaeozoic sedimentary rocks are absent, while the upper Palaeozoic sedimentary rocks are more prevalent, represented by the upper Carboniferous – lower Permian Amushan Formation (BGMRIM, 1991; Bu et al. 2012; Lu et al. 2012; Zheng et al. 2014; Zhang & Zhang, 2016). The lithology of the lower and middle sections of the Amushan Formation is obviously different from that of the upper section, suggesting a significant tectonic event occurred (Shi et al. 2014 a; Liu et al. 2017). The Jurassic sequences are sporadically exposed, which are composed of coarse-grained clastic rocks. By contrast, the Cretaceous sequences are more developed, characterized by volcaniclastic rocks.

3. Field observations and sampling

3.a. Field observations

In the ZSTB, the late Palaeozoic – Early Triassic intrusive rocks constitute the principal part of the Zongnaishan–Shalazhashan Mountain (NXBG, 1980 a,b, 1982, 2001; Zhang et al. 2013; Xie et al. 2021). The majority of the Triassic intrusive rocks in this region are controlled by E–W or NW-directed faults and are emplaced into the late Palaeozoic granitoids (Fig. 1b) (NXBG, 1980 a,b, 1982, 2001). Furthermore, these Triassic plutons are mainly exposed as small-scale stocks or branches, and mainly consist of granite, monzogranite and granodiorite (NXBG, 1980 a,b, 1982; Zhang et al. 2013; Shi et al. 2014 a; Zhang, Z. P. et al. 2016; Zhao, Z. L. et al. 2016). In this study, we conducted detailed studies on three plutons (the Haerchaoenji, Wulantaolegai and Chahanhada plutons) in the ZSTB. Mafic enclaves associated with these plutons were not observed during the field studies. The locations of the investigated plutons are shown in Figures 1 and 2.

Fig. 2. Geological map of the (a) Zongnaishan and (b) Shalazhashan areas.

The Haerchaoenji pluton is the largest pluton in the southwestern Zongnaishan area with an outcrop area of ∼100 km² (NXBG, 1982). The shape of this pluton is complex, and it is mainly exposed as branches and dykes. The strike of this pluton is mostly near N–S, implying that the rock mass intruded along a N–S-directed fault (Yebuerhai Fault) (NXBG, 1982). This pluton intruded the Precambrian gneiss and the Palaeozoic granitoids, and is unconformably covered by the Middle Jurassic strata in the south (Fig. 2a) (NXBG, 1982). The Haerchaoenji pluton is dominated by medium–fine-grained granite, biotite granite and granodiorite (NXBG, 1982). The Wulantaolegai pluton intruded into the upper Carboniferous strata (Fig. 2b) and is dominated by medium-grained granite and monzonitic granite (NXBG, 1980 a, 2001). The Wulantaolegai pluton is exposed as a rock branch. The Chahanhada pluton is located in the eastern Shalazhashan area, and trends in a NE–SW direction with an outcrop area of ∼12 km² (NXBG, 1980 b). This pluton is in the form of an elliptical stock and intrudes the late Palaeozoic granitoids (Fig. 2b). The Chahanhada pluton is unconformably covered by Lower Cretaceous strata (Bayingebi Fm) in the south and east areas (Fig. 2b) (NXBG, 1982). The main rock types are granite and monzonitic granite with a medium to coarse-grained granitic texture (Fig. 3g).

Fig. 3. Field photographs and photomicrographs showing petrographic features of the studied samples. (a–c) YE-17-69; (d–f) YE-17-78; (g–i) YE-17-88. Mineral abbreviations: Pl – plagioclase; Qtz – quartz; Kf-K – feldspar; Bt – biotite. Length of hammer for scale is 290 mm; length of hammer head for scale is 175 mm.

3.b. Sampling

A total of 16 samples were collected from the Haerchaoenji, Wulantaolegai and Chahanhada plutons for systematic zircon U–Pb–Hf isotopic and whole-rock geochemical analysis. The detailed description of these samples is carried out below.

The samples (YE-17-69, 69-1, 69-2, 69-3, 69-4) from the Haerchaoenji pluton are light grey, homogeneous, undeformed medium-grained granodiorites (3–5 mm) (Fig. 3a). The major mineral assemblages are quartz (∼25 vol. %), plagioclase (∼45–55 vol. %), K-feldspar (∼10–15 vol. %) and biotite (∼5–10 vol. %) (Fig. 3b, c), while the main accessory minerals are zircon, apatite and titanite. The plagioclases are subhedral–euhedral and show polysynthetic twinning (Fig. 3b). Most of the K-feldspars are subhedral to anhedral and show features of alteration on their surfaces (Fig. 3c). Some quartz crystals exhibit an anhedral granular texture among other minerals with wavy extinction, indicating dynamic recrystallization (Fig. 3c). Sub- to anhedral biotite is characterized by strong pleochroism, and it occasionally appears as mineral aggregates.

The samples (YE-17-78, 78-1, 78-2, 78-3, 78-4) from the Wulantaolegai pluton are pale red, fine–medium-grained granite (Fig. 3d), primarily composed of K-feldspar (∼30–35 vol. %), quartz (∼35 vol. %) and plagioclase (20–25 vol. %), with minor biotite (∼3 vol. %) (Fig. 3e, f) and accessory minerals (e.g. zircon, magnetite, titanite and apatite). K-feldspars are euhedral or subhedral and show relatively strong alteration. In addition, some K-feldspars show the distinctive feature of gridiron twinning. Quartz crystals are anhedral with rounded borders, while plagioclases are euhedral with polysynthetic twinning (Fig. 3e, f).

The samples (YE-17-88, 88-1, 88-2, 88-3, 88-4, 88-5) from the Chahanhada pluton are pale red, homogeneous medium-grained (3–5 mm) granites (Fig. 3g). Quartz (∼35–40 vol. %), K-feldspar (∼35–40 vol. %), plagioclase (∼20–27 vol. %) and biotite (∼1–2 vol. %) (Fig. 3h, i) are the major minerals. Zircon, apatite and titanite are the main accessory minerals. The K-feldspars show obvious evidence of alteration. The plagioclases are zoned with idiomorphic plates and show polysynthetic twinning. The quartz grains exhibit an anhedral granular texture among other minerals and have wavy extinction (Fig. 3h, i).

4. Analytical methods

4.a. Whole-rock major and trace elements

Whole-rock major and trace elements of the studied samples were analysed at the State Key Laboratory of Continental Dynamics, Northwest University, China. Fresh chips of whole-rock samples were powdered to ∼200 mesh using a tungsten carbide ball mill. Major elements were analysed using a Rigaku RIX 2100 X-ray fluorescence (XRF) spectrometer, and trace elements were analysed by an Agilent 7500a inductively coupled plasma mass spectrometer (ICP-MS) using United States Geological Survey (USGS) and international rock standards (BHVO-2, AGV-2, BCR-2 and GSP-1). For the trace-element analysis, sample powders were digested using an HF + HNO₃ mixture in high-pressure Teflon bombs at 190 °C for 48 hours. The analytical precision and accuracy for most of the major and trace elements is better than 5 % and 10 %, respectively (Liu et al. 2007).

4.b. Zircon Lu–Hf isotopic analyses

In situ zircon Hf isotope analysis was undertaken on a Nu Plasma HR multi-collector ICP-MS (Nu Instrument Ltd, UK) equipped with a GeoLas 2005 193 nm ArF excimer laser-ablation system. Analysis was carried out using a beam size of 44 μm and helium was used as a carrier gas. The laser repetition rate was 10 Hz and the energy density applied was 15–20 J cm⁻². Instrumental conditions and data acquisition methods were described by Zhao, Y. et al. (2017). Time-dependent drifts of Lu–Hf isotopic ratios were corrected using a linear interpolation according to the variations of 91500 and GJ-1. A decay constant of 1.867 × 10^–11 a⁻¹ for ¹⁷⁶Lu (Albarède et al. 2006) and the present chondritic ratios of ¹⁷⁶Hf/¹⁷⁷Hf = 0.282772 and ¹⁷⁶Lu/¹⁷⁷Hf = 0.0332 (Blichert-Toft & Albarède, 1997) were adopted to calculate ϵ_Hf(t) values (ϵ_Hf(t) = ((¹⁷⁶Hf/¹⁷⁷Hf)_s − (¹⁷⁶Lu/¹⁷⁷Hf)_s × (e^λt − 1))/((¹⁷⁶Hf/¹⁷⁷Hf)_CHUR,0 − (¹⁷⁶Lu/¹⁷⁷Hf)_CHUR × (e^λt − 1)) − 1) × 10 000; Wu et al. 2007). Bea et al. (2018) proposed that the best strategy to calculate the Hf T_DM is to use the analytically determined whole-rock Lu/Hf ratio as a proxy for the source Lu/Hf. In this study, we use the analytically determined whole-rock Lu/Hf ratio as described by Bea et al. (2018).

4.c. Zircon U–Pb geochronology

Zircon grains for U–Pb dating were extracted by using a combined technique of heavy liquid and magnetic separation, and then handpicked under a microscope, mounted in epoxy resin and polished until the centres of the zircon grains were exposed. Cathodoluminescence (CL) images were taken to reveal their internal structures and select the suitable U–Pb dating spots by using a Quanta 400FEG environmental scanning electron microscope.

Laser-ablation ICP-MS (LA-ICP-MS) zircon U–Pb dating was carried out at the State Key Laboratory of Continental Dynamics, Northwest University, China. The U–Pb dating was conducted on an Agilent 7500a ICP-MS instrument equipped with a 193 nm ArF excimer laser and a homogenizing imaging optical system. A fixed spot size of 32 μm with a laser repetition rate of 6 Hz was adopted throughout this study. Helium was used as carrier gas to provide efficient aerosol delivery to the torch. The standard silicate glass NIST 610 was used to optimize the instrument to obtain maximum signal intensity (²³⁸U signal intensity >460 cps/ppm) and low oxide production (ThO/Th <1 %). The ICP-MS measurements were carried out using time-resolved analysis operating in fast peak jumping mode and DUAL detector mode using a short integration time. ²⁰⁷Pb/²⁰⁶Pb, ²⁰⁶Pb/²³⁸U, ²⁰⁷Pb/²³⁵U and ²⁰⁸Pb/²³²Th ratios were calculated using the GLITTER 4.0 program (Macquarie University). The zircon 91500 was used as an external standard with a recommended ²⁰⁶Pb–²³⁸U age of 1065.4 ± 0.6 Ma (Wiedenbeck et al. 1995) for correction of both instrumental mass bias and depth-dependent elemental and isotopic fractionation. U, Th and Pb concentrations were calibrated by using ²⁹Si as an internal standard and NIST SRM 610 as an external standard. Concordia diagrams and weighted mean calculations were made using the Isoplot program (version 3.0) (Ludwig, 2003).

5. Analytical results

5.a. Whole-rock geochemistry

In this research, field investigation and photomicrographs reveal that these intermediate–acid intrusive rocks have rarely been affected by regional metamorphism. Major- and trace-element compositions of selected granitoids from the study area are listed in online Supplementary Material Table S1.

The samples from the Haerchaoenji granodiorite have SiO₂ = 63.10–65.80 wt %, total Fe₂O₃ = 3.86–4.65 wt %, Na₂O = 4.52–4.77 wt %, K₂O = 1.94–2.15 wt %, MgO = 1.55–1.91 wt %, Mg no. = 48–49 and CaO = 3.78–4.14 wt % (online Supplementary Material Table S1). In the plot of total alkalis versus SiO₂, these samples all fall into the subalkaline series field (Fig. 4a). In the plot of K₂O versus SiO₂, all samples fall into the medium-K calc-alkaline field (Fig. 4b). These granodiorites collected from the Haerchaoenji pluton are metaluminous to slightly peraluminous, with A/CNK (molecular ratio of Al₂O₃/(CaO + Na₂O + K₂O)) ratios ranging from 0.97 to 1.01 (Fig. 4c). In addition, these samples show enrichment of light rare earth elements (LREEs) ((La/Yb)_N = 27.13–41.31) and no obvious Eu anomalies (Eu = 0.95–1.02) in the chondrite-normalized REE diagrams (Fig. 5). They also exhibit depletion of Nb, Ta and Ce, and enrichment of Ba, Th, U and Pb contents in the primitive mantle-normalized spider diagrams (Fig. 5).

Fig. 4. (a) (Na₂O + K₂O) versus SiO₂, (b) K₂O versus SiO₂ and (c) A/NK versus A/CNK plots for investigated samples from the ZSTB. The field boundaries in the three diagrams are from Irvine & Baragar (1971), Peccerillo & Taylor (1976) and Maniar & Piccoli (1989), respectively. Previous data of ZHTB and ZSTB are cited from Shi et al. (2014 a) and Zhang et al. (2017).

Fig. 5. (a, c, e) Chondrite-normalized REE patterns; the normalization values of chondrite are from Taylor & McLennan (1985). (b, d, f) Primitive mantle-normalized trace-element patterns; data for primitive mantle are from Sun & McDonough (1989).

The samples of the Wulantaolegai granite show SiO₂ = 68.6–70.70 wt %, total Fe₂O₃ = 1.76–1.90 wt %, Na₂O = 5.63–6.30 wt %, K₂O = 3.52–3.74 wt % and CaO = 0.89–1.12 wt % (online Supplementary Material Table S1). In addition, they have low MgO contents of 0.23–0.25 wt % and Mg no. values of 23–24. Theses granites are light peraluminous, with A/CNK from 1.0 to 1.08 (Fig. 4c). In the plot of K₂O versus SiO₂, all samples fall into the high-K calc-alkaline field (Fig. 4b). In the chondrite-normalized REE diagrams, the granite samples show enrichment of LREEs ((La/Yb)_N = 5.87–6.66) and negative Eu anomalies (δEu = 0.80–0.82) (Fig. 5). They also exhibit depletion of Ba, Nb, Ce and Sr, and enrichment of Rb, Th, U and Pb contents in the primitive mantle-normalized spider diagrams (Fig. 5).

The samples from the Chahanhada granite show SiO₂ = 72.76–77.70 wt %, total Fe₂O₃ = 0.98–1.26 wt %, Na₂O = 4.05–4.77 wt %, K₂O = 3.20–3.77 and CaO = 0.35–0.51 wt % (online Supplementary Material Table S1). In addition, they have low MgO contents of 0.26–0.36 wt % with Mg no. values of 38–40. These granites are peraluminous, with an A/CNK from 1.15 to 1.17 (Fig. 4c). In the plot of K₂O versus SiO₂, all samples fall into the medium to high-K calc-alkaline field (Fig. 4b). Chondrite-normalized REE patterns of these samples show enrichment of LREEs ((La/Yb)_N = 10.27–12.93) with obvious Eu anomalies (δEu = 0.49–0.53) (Fig. 5). They exhibit depletion of Ba, Nb, Ce, Sr and Eu, and enrichment of Rb, Th, U, La, Pb and Nd contents in the primitive mantle-normalized spider diagrams as well (Fig. 5).

5.b. U–Pb zircon geochronological data

The results of zircon LA-ICP-MS U–Pb dating are presented in online Supplementary Material Table S3. The zircons separated from the granodiorite (YE-17-69) and granites (YE-17-78, YE-17-88) are mostly colourless, transparent and well crystallized, with grain diameters of 200–300 μm, 150–200 μm and 50–120 μm, respectively (Fig. 6). The length/width ratios of the zircon grains range from 1:1 to 5:1 (YE-17-69), 1:1 to 3:1 (YE-17-78) and 1:1 to 2:1 (YE-17-88), respectively. The CL images revealed that the selected zircons display clear oscillatory zoning and platy structures (Fig. 6). All zircon grains are euhedral to subhedral with prismatic to sub-prismatic shapes (Fig. 6). Moreover, the relatively high Th/U ratios of the three samples (0.43–1.34, 0.37–0.62 and 0.47–1.30, respectively) also suggest a magmatic origin (Hoskin & Schaltegger, 2003). The ²⁰⁶Pb–²³⁸U weighted average ages of concordant points are 245 ± 5 Ma (MSWD = 0.56, N = 19) for YE-17-69, 237 ± 2 Ma (MSWD = 0.43, N = 25) for YE-17-78 and 245 ± 2 Ma (MSWD = 0.38, N = 17) for YE-17-88.

Fig. 6. (a) Cathodoluminescence (CL) images of representative zircons of investigated samples from the ZSTB. (b) Chondrite-normalized REE patterns of the zircons from investigated samples.

5.c. Zircon Lu–Hf results

The zircon grains that were previously analysed by U–Pb methods were also analysed for Lu–Hf isotopes on the same spot, and the results are listed in online Supplementary Material Table S2. Fifteen spots on zircons selected from sample YE-17-69 yielded variable ϵ_Hf(t) values between +1.8 and +6.4 (Fig. 7), with Hf model ages (T_DM) of 636–837 Ma, and initial ¹⁷⁶Hf/¹⁷⁷Hf ratios from 0.282676 to 0.282807. Fifteen spots on zircons selected from sample YE-17-78 showed variable ϵ_Hf(t) values ranging from +3.3 to +8.7 (Fig. 7), corresponding to T_DM ages varying from 545 to 778 Ma, with the initial ¹⁷⁶Hf/¹⁷⁷Hf ratios ranging from 0.282712 to 0.282864. Fifteen spots on zircons from sample YE-17-88 yielded positive ϵ_Hf(t) values ranging from +5.5 to +11.8 (Fig. 7), corresponding to young T_DM ages from 425 to 729 Ma, and the initial ¹⁷⁶Hf/¹⁷⁷Hf ratios varied between 0.282776 and 0.282955.

Fig. 7. Zircon Hf isotopic compositions of intrusive rocks from the CAOB. ZSTB – Zongnaishan–Shalazhashan Tectonic Belt; ZHTB – Zhusileng–Hangwula Tectonic Belt; NHTB – Nuoergong–Honggueryulin Tectonic Belt. The ϵ_Hf(t) values are cited from Shi et al. (2012, 2014 a,b), Dan et al. (2014, 2015, 2016), Ye et al. (2016), Zhang, W. et al. (2016), Liu et al. (2017) and Zhao et al. (2020).

6. Discussion

6.a. Geochronological framework of the ZSTB

The geochronological data are important to constrain the magmatic event and further understand the tectonic evolution of the northern Alxa region. In this study, the obtained zircon U–Pb ages are considered to reflect the timing of magmatic events. The zircon U–Pb dating of the samples from three plutons in the ZSTB yielded weighted mean ²⁰⁶Pb–²³⁸U ages of 237∼245 Ma (Fig. 8). These dates provide robust evidence for the presence of early Mesozoic magmatism in the northern Alxa region. Furthermore, we collected previously reported magmatic events in the ZSTB (e.g. Zhang et al. 2013; Liu & Zhang, 2014 a,b; Shi et al. 2014 a,b; Yang et al. 2014; Shi, G. Z. et al. 2016; Zheng et al. 2016; Xie et al. 2021) and revealed several magmatic episodes in the ZSTB (Fig. 9; online Supplementary Material Table S4). Although such late Palaeozoic – early Mesozoic magmatism is successive, the statistical data display three main age peaks at c. 270, 250 and 228 Ma (Fig. 9). When these age data are combined, they show multi-stage magmatism in the ZSTB (Fig. 9), implying a long-lived magmatism from late Palaeozoic to early Mesozoic times in response to a prolonged subduction, collision and extension in the central segment of the CAOB.

Fig. 8. Zircon U–Pb concordia diagrams and histograms for investigated samples.

Fig. 9. Histogram of zircon U–Pb ages of the Phanerozoic magmatism in the ZSTB, northern Alxa region. Data sourced from online Supplementary Material Table S4.

6.b. Genetic type

Granitoids are commonly classified into I-, A-, S- and M-types based on their source compositions, mineral assemblages and geochemical features (Chappell & White, 2001; Bonin, 2007). The Haerchaoenji granodiorite and Chahanhada granite are similar to typical I-type granitoids. Specifically, these granitoids are metaluminous to weakly peraluminous and medium-K to high-K calc-alkaline with A/CNK and A/NK ratios of 0.97–1.17 and 1.24–1.74, respectively. These features suggest that they represent an I-type or A-type granitoid rather than an S-type (Chappell & White, 1992; Zhang et al. 2017; Zhao et al. 2020). Moreover, these granitoids have relatively lower 10 000 Ga/Al ratios (1.86–2.34) and Zr + Nb + Ce + Y contents (269.88–347.60 ppm) than A-type granitoids (Whalen et al. 1987) (Fig. 10a–d). The negative correlation between P₂O₅ and SiO₂ appears to follow the I-type trend (Fig. 10e). The relatively low Zr and Ce contents of the samples also suggests that these rocks are I-type granitoids. This conclusion can be further supported by the Na₂O versus SiO₂ diagram (Fig. 10f).

Fig. 10. Petrogenetic discrimination diagrams for early Mesozoic granitoids in the ZSTB. (a) (K₂O + Na₂O/CaO) versus Zr + Nb + Ce + Y. (b) FeO^T/MgO versus Zr + Nb + Ce + Y. (c) Zr versus 10 000Ga/Al. (d) FeO^T/MgO versus 10 000Ga/Al. (e) P₂O₅ versus SiO₂. (f) Na₂O versus SiO₂ (a–d are after Whalen et al. 1987, and e, f are after Chappell & White, 1992).

However, the Wulantaolegai granite has characteristics more similar to A-type granitoids. These samples have high K₂O + Na₂O, FeO^T/MgO, Zr and Ga/Al ratios, which are consistent with those of A-type granitoids (e.g. Dan et al. 2014; Ao et al. 2019). In addition, the samples have higher 10 000*Ga/Al (2.67–2.74) and Zr + Nb + Ce + Y (1051–1230 ppm), and plot into the A-type granitoid field on the discrimination diagrams (Fig. 10a–d). Thus, the Haerchaoenji granodiorite and Chahanhada granite are considered to be I-type granitoids, while the Wulantaolegai granite is classified as A-type granitoid.

6.c. Temperature–pressure conditions of melting

Zircon saturation thermometry can be used to make an approximate estimate of the temperature of crustal-derived silicic magmas at the early stage of crystallization (Hui et al. 2021 and references therein). Zircon saturation temperatures (T _Zr) of magma are estimated using zirconium concentrations of melt using the equation from Boehnke et al. (2013). Based on the Zr content of the studied samples, the T _Zr ranged from 728 to 747 °C (av. 738 °C) in the Haerchaoenji granodiorites, 928 to 981 °C (av. 955 °C) in the Wulantaolegai granites, and 740 to 792 °C (av. 778 °C) in the Chahanhada granites. The mean values of T _Zr in the Haerchaoenji granodiorite and Chahanhada granite are consistent with those in typical I-type granites (781 °C, e.g. Chappell & White, 1992). The mean value of T _Zr in the Wulantaolegai granite points to a hot granitoid (T_Zr >800 °C; Miller et al. 2003), which is consistent with that of A-type granites (Watson & Harrison, 1983).

With respect to pressure, low Sr contents and Sr/Y ratios, as well as negative Eu anomalies in the Wulantaolegai and Chahanhada granites (online Supplementary Material Table S1), reflect low-pressure conditions of the magma source region (e.g. Martin et al. 2005). The coupled observations of the two pressure-dependent ratios, namely Sr/Y and La/Yb, point to a low pressure as well (Fig. 11). The low-pressure conditions inferred for these granites are consistent with their high silica contents as well (e.g. Blundy & Cashman, 2001). In contrast, the Haerchaoenji granodiorites exhibit high Sr and Ba contents with high Sr/Y ratios, low Y and heavy rare earth element (HREE) contents, implying high-pressure conditions (pressure >12 kbar) (e.g. Patiño Douce, 1999; Martin et al. 2005; Liu et al. 2016).

Fig. 11. Plots of (a) La/Sm versus La; (b) Zr/Nb versus Zr; (c) Sr/Y versus Y; and (d) Sr/Y versus La_N/Yb_N (a, b are after Xie et al. 2021, and c, d are after Castro et al. 2011).

6.d. Petrogenesis and magma source

6.d.1. The Haerchaoenji and Chahanhada I-type granitoids

The Haerchaoenji granodiorite and Chahanhada granite are calc-alkaline and peraluminous I-type granitoids, which could be formed by: (1) partial melting of pre-existing igneous rocks in the crust (Clemens et al. 2011; Topuz et al. 2019; Xie et al. 2021); (2) mixing of mantle-derived magmas with crustal-derived materials (Clemens et al. 2009); and (3) assimilation and fractional crystallization processes of mantle-derived basaltic melts (Barth et al. 1995; Quelhas et al. 2020).

The investigated samples have Rb/Sr = 0.80–1.00, K/Rb = 326.32–435.50 and Zr/Hf = 35.66–46.95, which differs from the high Rb/Sr (> 5), low K/Rb (110) and low Zr/Hf (20) ratios of fractionated granitoids (Wu et al. 2020). The fractional crystallization of mafic melts would leave large amounts of mafic–ultramafic cumulates (Clemens et al. 2011), which is obviously different from the field investigation. This supposition is also evidenced by the absence of xenocrystic zircons in the investigated granitoids. In addition, these samples from the Haerchaoenji granodiorite and Chahanhada granite show low MgO (0.26–1.91), Cr (4.25–17.04) and Ni (2.61–6.49) contents and moderate Mg no. values (38–49), similar to those of magma formed by partial melting of thickened lower crust instead of fractional crystallization from the mantle directly (Ao et al. 2019; Yomeun et al. 2022). Furthermore, the positive correlation of La/Sm versus La and Zr/Nb versus Zr presented by the studied rocks can be produced by either magma mixing or partial melting rather than fractional crystallization (Fig. 11). Commonly, the magma mixing model can generate massive mafic enclaves and geochemical variations (Kemp et al. 2007). As mentioned above, there are no mafic microgranular enclaves discovered in the field investigation. The studied samples do not show obvious geochemical variations either. In the Mg no. versus SiO₂ diagram, these samples are also not in conformity with the magma mixing trend (Fig. 12c). The ϵ_Hf(t) values of the Haerchaoenji and Chahanhada granitoids are distinct from the variable ϵ_Hf(t) values of granitoids formed by magma mixing (usually from negative to positive; Griffin et al. 2002). The zircon trace elements of the Haerchaoenji and Chahanhada granitoids have medium Th and U contents, indicating a crustal affinity as well. Thus, the Haerchaoenji and Chahanhada granitoids were probably generated by partial melting of pre-existing crustal basements.

Fig. 12. (a) Al₂O₃/(MgO + FeO^T + TiO₂) versus Al₂O₃ + MgO + FeO^T + TiO₂ (Patiño Douce, 1999). (b) (Na₂O + K₂O)/CaO versus Na₂O + K₂O + CaO (Patiño Douce, 1999). (c) Mg no. versus SiO₂ diagram (after Zhu, R. Z. et al. 2018; reference fields after Patiño Douce, 1999; Wolf & Wyllie, 1994). (d) (Na₂O + K₂O)/(FeO^T + MgO + TiO₂) versus (Na₂O + K₂O + FeO^T + MgO + TiO₂) (Patiño Douce, 1999). (e) CaO/(MgO + FeO^T + TiO₂) versus (CaO + MgO + FeO^T + TiO₂) (Patiño Douce, 1999). (f) Rb/Ba versus Rb/Sr (Patiño Douce, 1999).

Partial melting of different source rocks would generate compositional variations in the magmas that could be visualized in terms of major-element compositions (Altherr et al. 2000). The major-element compositions of the Haerchaoenji granodiorites (e.g. high Na₂O and Al₂O₃, medium CaO, low MgO, etc) are similar to those of the intermediate to granitic rocks generated by the partial melting of basaltic (mafic) rocks (Rapp & Watson, 1995; Patiño Douce, 1999). In the major-element feature diagrams (Fig. 12a–e), the granodiorites display a similarity with the experimental melts of amphibolite-bearing mafic rocks (Patiño Douce, 1999; Lu et al. 2016, 2017). The low Rb/Ba (0.07–0.08) and Rb/Sr (0.08–0.10) ratios indicate basalt-derived components as well (Fig. 12f). The low Th/La ratios (<0.5) of these granodiorites are also consistent with those of the products yielded by partial melting of mafic crustal sources. The positive zircon ϵ_Hf(t) values between +1.8 and +6.4 (Fig. 7), with young T_DM ages of 636–837 Ma, indicate that the granodiorites were mainly derived from Neoproterozoic juvenile mafic crustal materials. In contrast, the Chahanhada granite samples have relatively high Al₂O₃/TiO₂ ratios (52.54–93.76), A/CNK values (1.00–1.17) and low CaO/Na₂O ratios (0.09–0.18), suggesting the derivation from a parental magma that was probably generated by the partial melting of a metasedimentary source (Sylvester, 1998; Zhu, R. Z. et al. 2018). In the source discrimination diagrams (Fig. 12), the Chahanhada granites plot into the fields of metagreywacke and metapelite melts. Actually, it is common that the source of I-type granites involves mature sedimentary materials (Zhu, Y. et al. 2018). However, the positive zircon ϵ_Hf(t) values ranged from +5.5 to +11.8 (Fig. 7), with young T_DM ages of 425–729 Ma, suggesting the significant involvement of juvenile crustal materials. Thus, the Chahanhada granites might have originated from juvenile crust with the input of metasedimentary components.

6.d.2. The Wulantaolegai A-type granite

The Wulantaolegai granite displays the features of A-type granite, which is generally attributed to: (1) differentiation of mantle-derived alkaline basalts (Turner et al. 1992; Mushkin et al. 2003); (2) partial melting of crustal materials at high temperatures (Collins et al. 1982; King et al. 1997), and (3) a combination of crustal and mantle sources, i.e. crustal assimilation and fractional crystallization of mantle-derived magmas, or magma mixing of mantle-derived melts and crustal magmas (Kemp et al. 2005). The Mg no. values and Cr and Ni contents of the Wulantaolegai granites are much lower than those of the mantle-derived melts (Mg no. = 73–81, Cr >1000 ppm, Ni >400 ppm) (Wilson, 1989). The Nb/Ta (8 on average) and Zr/Hf (43 on average) ratios of the Wulantaolegai granites in this study are consistent with those of the crust. The low Nb/Y (0.38–0.41) and Rb/Y (2.37–2.71) ratios also suggest a lower crustal source (Rudnick & Fountain, 1995). Furthermore, the Wulantaolegai granites have higher Y/Nb (2.45–2.60, >1.2), i.e. A2-type granite affinities (Eby, 1992; Frost & Frost, 2011), which also suggests that the magmas were derived from continental crust or underplated basaltic protoliths (Eby, 1992). So far, coeval mantle-derived mafic rocks have not been recognized in the study area. The absence of mafic microgranular enclaves in the Wulantaolegai pluton does not support the model of a combination of crustal and mantle sources. In the Mg no. versus SiO₂ diagram (Fig. 12c), the Wulantaolegai granite samples are not in conformity with the magma mixing trend. In the La/Sm versus La and Zr/Nb versus Zr diagrams, the Wulantaolegai granite samples also display the feature of partial melting processes rather than magma mixing or fractional crystallization (Fig. 11). The Wulantaolegai granite samples have positive ϵ_Hf(t) values ranging from +3.3 to +8.7, indicating a magma source from juvenile crustal basement rather than a mixed source. In addition, the zircon saturation temperatures of the Wulantaolegai granite indicate high-temperature conditions. Thus, the model of partial melting of juvenile crustal materials at high temperatures is reasonable for the petrogenesis of the Wulantaolegai granite.

The relatively low Sr (57.50–62.60 ppm) and high HREE contents, and weakly fractionated HREEs and low Sr/Y ratios (1.72–1.99) suggest these rocks were mainly derived from a crustal source above the garnet stability depth (Cai et al. 2011 b), and the high Rb/Y ratios (2.37–2.71) and low Nb/Y ratios (0.38–0.41) display the approach to the upper crustal source (Taylor & McLennan, 1985). These features suggest that the source region of these granites is relatively shallow. In the source discrimination diagrams (Fig. 12), the Wulantaolegai granites variably fall into the overlapping fields of the partial melts of metagreywackes, psammite and meta-igneous rocks. The positive zircon ϵ_Hf(t) values between +3.3 and +8.7 (Fig. 7) with young T_DM ages of 545–778 Ma suggest the involvement of juvenile mafic crust for the Wulantaolegai granites, which is similar to other granitoids in the southern CAOB (e.g. Shi et al. 2014 a; Xie et al. 2021). The low Th/La ratios (<0.5) of the A-type granites in this study are also consistent with that of the partial melting products of mafic crustal sources. The variable zircon ϵ_Hf(t) units of these granites were probably caused by some recycled sediments in the magma source.

6.e. Tectonic setting and geological implications

6.e.1. Tectonic setting

In this study, the investigated granitoids display the common features of volcanic arc granites, such as the depletion of Nb, Ta and enrichment of large ion lithophile elements with low Sr/Y and (La/Yb)_N (e.g. Zhao, Y. et al. 2017; Xie et al. 2021). On the Th/Yb versus Ta/Yb and Ta*3–Rb/30–Hf ternary diagram, these granitoids also display the affinity of volcanic arc granitoids, analogous to a subduction-related compressional setting (Fig. 13a, b). On the tectonic discrimination diagrams, the Haerchaoenji and Chahanhada granitoid samples plot in the volcanic arc field, while the Wulantaolegai samples show trends from the arc to post-collisional fields (Fig. 13a–f). These findings suggest that these granitoids either formed in a subduction-related setting, or a post-collisional setting with arc-like geochemical signatures which are inherited from a previous arc source. In this study, we prefer a post-collisional setting with arc affinity based on the following regional data: (1) the magmatism ranging from late Carboniferous to middle–late Permian times exhibits a marked petrogenetic, geochemical and isotopic transition and trends from the subduction to post-collisional fields (e.g. Zhang et al. 2013; Shi et al. 2014 a,b; Xu, D. Z. et al. 2014; Yang et al. 2014; Zheng et al. 2014; Chen et al. 2015; Xie et al. 2015; Liu et al. 2017, 2018); (2) the regional unconformity and the change of sedimentary facies also suggest that a significant tectonic event happened during early–middle Permian time (Zhang, 2019); (3) the palaeomagnetic, provenance and palaeontological studies further suggest that the PAO in the northern Alxa region closed before earliest Mesozoic time (Fig. 14a) (Pu et al. 2013; Huang et al. 2018; Zhang et al. 2018).

Fig. 13. Tectonic setting discrimination diagrams for the early Mesozoic granitoids in the ZSTB. (a) Ta*3–Rb/30–Hf ternary plot (Harris et al. 1987). (b) Th/Yb versus Ta/Yb (Gorton & Schandl, 2000). (c) Rb versus Y + Nb (Pearce et al. 1984; Pearce, 1996). (d) Ta versus Yb (Pearce et al. 1984). (e) Rb versus Ta + Yb (Pearce et al. 1984). (f) Nb versus Y (Pearce et al. 1984). Syn-COLG – syn-collision granites; VAG – volcanic arc granites; WPG – within plate granites; ORG – ocean ridge granites.

Fig. 14. (a) Palaeogeographic reconstructions of Eastern Asian blocks (modified after Huang et al. 2018). (b, c) Diagrams illustrate the tentative tectonic scenario showing the Middle Triassic evolution of the ZSTB and adjacent areas. IC – Indochina Block; MOB – Mongolian Block; NCB – North China Block; NQ – North Qiangtang block; SCB – South China Block; Si – Sibumasu block; SQ – South Qiangtang block.

Therefore, the Middle Triassic granitoids in the ZSTB are interpreted as post-collisional granites (Fig. 14b, c). Furthermore, in the scenario of subduction and subsequent continental collision processes, asthenospheric mantle upwelling would be inevitable owing to slab roll-back or break-off (Ersoy et al. 2017; Collins et al. 2020). The Wulantaolegai A-type granite was probably generated by an extensional setting in response to slab break-off during the final amalgamation (Fig. 14c).

6.e.2. Geological implications

As mentioned above, extensive studies have been carried out on the closure of the PAO, producing a large quantity of data and competing models (e.g. Xiao et al. 2013, 2015, 2018; Eizenhöfer et al. 2014, 2015 a,b; Li et al. 2015, 2016 a, 2017; Liu et al. 2017, 2018, 2019 a,b; Han & Zhao, 2018; Eizenhöfer & Zhao, 2018; Du et al. 2019; Shen et al. 2019; Zheng et al. 2019; Niu et al. 2021). Generally, the western segment of the PAO closed along the Tianshan Orogen during the Carboniferous–early Permian period (e.g. Han & Zhao, 2018; Zheng et al. 2019). However, the eastern segment of the PAO closed during late Permian to Middle Triassic times along the Solonker Suture Belt (e.g. Eizenhöfer et al. 2014, 2015 a,b; Li et al. 2015, 2016 a, 2017; Eizenhöfer & Zhao, 2018). Recent studies demonstrated that the central segment of the PAO closed at c. 280–265 Ma (Liu et al. 2016, 2017, 2018; Zhao et al. 2018), which is also consistent with this study. Combining these data together, we still tend to support the scissor-like closure manner, which is in accordance with previous studies (e.g. Boucot et al. 2013; Xiao et al. 2015; Zhao et al. 2018; Han & Zhao, 2018; Shen et al. 2019). This conclusion is also supported by the constraints from sedimentary strata (Zhao, Y. L. et al. 2016; Liu et al. 2019 a; Du et al. 2019), syn-collisional magmatic rocks (Wang et al. 2015; Chen et al. 2017; Ma et al. 2017), structural evidence (Xiao et al. 2015) and plate reconstruction (Domeier & Torsvik, 2014; Domeier, 2018).

In order to decipher the nature of the different tectonic units of the northern Alxa region, we collected comprehensive Hf isotopic data in this region (Fig. 7) (Shi et al. 2012, 2014 a,b; Dan et al. 2014, 2015, 2016; Ye et al. 2016; Zhang, W. et al. 2016; Liu et al. 2017; Zhao et al. 2020). It turns out that the magmatic rocks from the ZHTB and ZSTB have the most positive to low negative ϵ_Hf(t) values and relatively young Hf model ages (Fig. 7), suggesting a juvenile nature for the basement (Shi et al. 2014 a,b; Zhao et al. 2020). Significantly, these characteristics are similar to those of the granitoids in the CAOB (Guo et al. 2007; Cao et al. 2011, 2012; Meng et al. 2011; Li et al. 2012, 2013). However, the magmatic rocks from the southernmost NHTB display negative ϵ_Hf(t) values and ancient Hf model ages (Fig. 7), indicating an ancient nature for the basement (Zhang, J. J. et al. 2015; Ye et al. 2016). Therefore, the juvenile nature of the ZHTB and ZSTB is similar to the CAOB (Shi et al. 2014 a,b; Zhang, J. J. et al. 2015; Xie et al. 2021), but is different from the Alxa Block (NHTB). This conclusion is further reinforced by whole-rock Nd isotopic studies of the Phanerozoic granitoids and volcanic rocks (e.g. Dolgopolova et al. 2013; Shi et al. 2014 a), and obvious differences in magmatism record and Precambrian rock constitution (e.g. Geng & Zhou, 2010, 2011; Shi et al. 2014 a). Thus, the boundary of the CAOB and Alxa Block is most likely the border between the ZSTB and NHTB (Badain Jaran fault or Qagan Qulu Ophiolite Belt) rather than the Enger Us belt previously proposed (e.g. Shi, 2015; Zhang, J. J. et al. 2015). On a larger scale, this boundary is most likely the central segment of the Tianshan–Solonker suture zone, which connects the northern CAOB with the southern Tarim and North China cratons.

7. Conclusion

1. (1) New LA-ICP-MS zircon U–Pb dating results have revealed the Middle Triassic magmatism in the Zongnaishan and Shalazhashan areas: the Haerchaoenji granodiorite (245 ± 5 Ma), the Wulantaolegai granite (237 ± 2 Ma) and the Chahanhada granite (245 ± 2 Ma). This study and previous data provide evidence of a prolonged mafic–intermediate magmatism in the ZSTB related to the subduction and closure of the PAO.

2. (2) The Haerchaoenji granodiorite and Chahanhada granite are classified as I-type granitoids, while the Wulantaolegai granite is considered to be an A-type granite. They were probably derived from partial melting of juvenile crustal materials, inferred from the variable positive Hf isotopic signature and young T_DM model ages. The major-element compositions of the Chahanhada granite and Wulantaolegai granites suggest input of a metasedimentary component as well.

3. (3) Based on the compilation of magmatic, sedimentary, palaeomagnetic and palaeobiogeographic evidence, we propose that the Middle Triassic granitoids in this study were formed in a post-collisional setting, and the arc affinity was probably inherited from recycled subduction-related materials.

4. (4) The findings of this study support the scissor-like closure mode of the PAO as well as the different tectonic affinities of the ZHTB + ZSTB and NHTB.