Geochronology, geochemistry and tectonic implications of early Carboniferous plutons in the southwestern Alxa Block

Abstract

The southeastern Central Asian Orogenic Belt (CAOB) records the assembly process between several micro-continental blocks and the North China Craton (NCC), with the consumption of the Paleo-Asian Ocean (PAO), but whether the S-wards subduction of the PAO beneath the northern NCC was ongoing during Carboniferous–Permian time is still being debated. A key issue to resolve this controversy is whether the Carboniferous magmatism in the northern NCC was continental arc magmatism. The Alxa Block is the western segment of the northern NCC and contiguous to the southeastern CAOB, and their Carboniferous–Permian magmatism could have occurred in similar tectonic settings. In this contribution, new zircon U–Pb ages, elemental geochemistry and Sr–Nd isotopic analyses are presented for three early Carboniferous granitic plutons in the southwestern Alxa Block. Two newly identified aluminous A-type granites, an alkali-feldspar granite (331.6 ± 1.6 Ma) and a monzogranite (331.8 ± 1.7 Ma), exhibit juvenile and radiogenic Sr–Nd isotopic features, respectively. Although a granodiorite (326.2 ± 6.6 Ma) is characterized by high Sr/Y ratios (97.4–139.9), which is generally treated as an adikitic feature, this sample has highly radiogenic Sr–Nd isotopes and displays significantly higher K₂O/Na₂O ratios than typical adakites. These three granites were probably derived from the partial melting of Precambrian continental crustal sources heated by upwelling asthenosphere in lithospheric extensional setting. Regionally, both the Alxa Block and the southeastern CAOB are characterized by the formation of early Carboniferous extension-related magmatic rocks but lack coeval sedimentary deposits, suggesting a uniform lithospheric extensional setting rather than a simple continental arc.

1. Introduction

The Phanerozoic Central Asian Orogenic Belt (CAOB), one of the largest long-lived accretionary orogens worldwide, is situated to the north of the Tarim–North China cratons (Fig. 1a) and formed by complex subduction, accretion and collision processes related to the consumption of the Paleo-Asian Ocean (PAO), with significant crustal growth (Han et al. 1997, 2011; Jahn et al. 2000; Wu et al. 2003; Windley et al. 2007; Xiao et al. 2018). The southeastern CAOB records the Palaeozoic amalgamation between the North China Craton (NCC) in the south and Mongolia, Hunshandake and Songliao blocks within the CAOB in the north (Xu et al. 2013; Zhao et al. 2018; Zhou et al. 2018). The Permian–Early Triassic Solonker suture (Solonker–Xar Moron–Changchun suture) contains the youngest ophiolites within the southeastern CAOB and is usually regarded as the terminal closure site of the PAO (Eizenhöfer & Zhao, 2018; Wilde & Zhou, 2015; Xiao et al. 2003). However, when and how the PAO finally closed in the southeastern CAOB is still controversial, and different opinions can be grouped into three models.

Fig. 1. (a) Tectonic location of the Alxa Block. (b) Schematic geological map showing the distribution of Palaeozoic intrusions and ophiolitic mélanges in the Alxa Block (modified after Dan et al. 2014). (c) Simplified geological map of the southwestern Alxa Block (modified after Wang et al. 2020).

In the first set of models, the subduction of the PAO was continuous from the early Palaeozoic Era to Late Permian–Early Triassic time and led to the successive accretion of micro-continental blocks and magmatic arcs to the northern NCC, with the northern margin of the NCC as a continental arc during Carboniferous–Permian time and the Solonker suture as the final closure site of the PAO (e.g. Xiao et al. 2003, 2009 b, 2018; Zhang et al. 2014, 2016d ). The second set of models propose the Late Devonian–early Carboniferous closure of the PAO, with the southeastern CAOB in a post-collisional setting since then (e.g. Xu et al. 2013; Tong et al. 2015; Zhang et al. 2015 b). The third set of models infer that the large-scale PAO closed before the Late Devonian Epoch, but a new orogenic cycle began with intra-continental rifting within the southeastern CAOB during early Carboniferous time and resulted in the formation of a Red-Sea-like limited ocean basin, with the Solonker suture marking its closure during the Early Triassic Epoch (e.g. Zhang et al. 2015 a; Luo et al. 2016; Pang et al. 2016; Zhao et al. 2017; Xu et al. 2018). In the third model, the lithospheric extension may be triggered by slab break-off (Kozlovsky et al. 2015; Zhang et al. 2012 a) and enhanced by slab avalanche-driven wet mantle upwelling rising from the hydrous mantle transition zone (Wang et al. 2015 a, 2016 a).

To test the likelihood of one of these geodynamic models, a key question is whether the Carboniferous–Permian tectono-magmatic activity of the southeastern CAOB was dominated by the continued S-wards subduction of the PAO or by lithospheric extension. Accordingly, the tectonic setting of the Carboniferous magmatism in the northern margin of the NCC, either continental arc or lithospheric extension, can provide insights into the terminal evolutionary history of the southeastern CAOB.

The Alxa Block, also known as the Alxa Tectonic Belt (Song et al. 2018), connects the NCC to the east and the Tarim Craton to the west and lies between the CAOB to the north and the North Qilian Orogen to the south (Fig. 1a). Although this block is largely covered by deserts, numerous Phanerozoic plutons intruding into Precambrian metamorphic basement rocks crop out in its southwestern and northeastern parts (Fig. 1b). These plutons are mostly Palaeozoic in age, spanning Middle Ordovician–Early Devonian time (c. 458–394 Ma) and end Late Devonian–end Permian time (c. 359–252 Ma; Fig. 2a). Notably, this age pattern is quite similar to that of the southeastern CAOB (including the northern NCC), which includes two magmatic stages of middle Cambrian–Middle Devonian time (c. 508–386 Ma) and end Late Devonian–end Permian time (c. 362–252 Ma; Fig. 2b), indicating an operation of comparable tectonic processes. Further, the early magmatic stage in the southwestern Alxa Block could also be related to the North Qilian Orogen (Duan et al. 2015; Zhang et al. 2017 a; Wang et al. 2020), but the Qilian orogenesis ended before the Late Devonian Epoch (Xiao et al. 2009 a; Song et al. 2013). The Carboniferous magmatism within the Alxa Block was therefore most likely related to the tectono-magmatic activity of the southeastern CAOB.

Fig. 2. Statistical histograms of zircon U–Pb ages of Palaeozoic magmatic rocks in the (a) Alxa Block (data from this study and Qin, 2012; Tang, 2015; Gong et al. 2018 a; Zhang et al. 2018d ; Liu et al. 2019; Pan, 2019; Song et al. 2019; Chen et al. 2020; Wang et al. 2020; Zhao et al. 2020) and (b) the southeastern Central Asian Orogenic Belt (data from Wang et al. 2015 b).

In this study, new geochronological, elemental and isotopic geochemical analyses of three early Carboniferous plutons in the southwestern Alxa Block are presented. These results, combined with regional correlations, suggest a lithospheric extensional setting rather than a simple continental arc for the development of early Carboniferous magmatism in both the Alxa Block and the southeastern CAOB.

2. Geological background

The Alxa Block is separated from the CAOB by the Enger Us Fault to the north, and from the North Qilian Orogen to the SW by the Longshoushan Fault (Fig. 1b). It is traditionally considered as the western part of the northern NCC (Fig. 1a), either the western part of the Yinshan Block (e.g. Zhao et al. 2005, 2012; Wan et al. 2006; Wang et al. 2016 b, 2019 a) or the western extension of the Khondalite Belt (e.g. Geng et al. 2010; Zhang et al. 2013 a; Zhang & Gong, 2018). However, a close affinity of the Alxa Block to the Tarim or South China cratons had also been proposed (e.g. Tung et al. 2007; Yuan & Yang, 2015; Song et al. 2017), and the amalgamation of this block with the NCC might have taken place during early–middle Palaeozoic time (Dan et al. 2016; Zhang et al. 2016c ), although no ophiolitic mélanges have been recognized between them until now. Nevertheless, in any of the proposed models, the Alxa Block has been considered as part of the northern NCC, having been amalgamated at least since the Carboniferous Period.

Three ophiolitic mélanges have been reported in Alxa area (Fig. 1b). Two of them crop out in the NE, including the c. 302 Ma Enger Us and the c. 275 Ma Quagan Qulu ophiolitic mélanges, with their basaltic rocks exhibiting normal mid-ocean-ridge basalt (N-MORB) and boninite-like geochemical features (Zheng et al. 2014), respectively. The Tepai ophiolitic mélange in the SW is also characterized by boninite-like basaltic rocks, but its formation age is either c. 278 Ma (Zheng et al. 2018) or c. 437–448 Ma (Pan, 2019).

The southwestern Alxa Block between the Longshoushan Fault and the Badain Jaran Desert involves the NW–SE-trending Beidashan and Longshoushan–Helishan mountains (Fig. 1c). The widespread Precambrian basement rocks in this area include the Neoarchean Beidashan complex (Gong et al. 2012; Zhang et al. 2013 a) and Palaeoproterozoic Longshoushan Group (Tung et al. 2007; Gong et al. 2011). They consist of amphibolite- to greenschist-facies metamorphosed igneous and sedimentary rocks and are overlain unconformably by Neoproterozoic greenschist-facies meta-sedimentary rocks (Zhang & Gong, 2018). Recently, syenite of age c. 1.87 Ga and granitic gneiss of age c. 1.2 Ga were recognized in the Helishan area (Song et al. 2017; Wang et al. 2019 b).

Lower Palaeozoic sedimentary rocks in the southwestern Alxa area crop out only to the south of the Longshoushan Fault (Fig. 1c). They are known as the Dahuangshan Formation and are composed of unmetamorphosed or greenschist-facies marine clastic and carbonate rocks (Zhang et al. 2016 a). In contrast, the upper Carboniferous–middle Permian sedimentary rocks are widely distributed (Fig. 1c). The upper Carboniferous succession consists of interbedded volcanic and clastic rocks in the lower part and shallow-marine bioclastic limestones and sandstones in the upper part, and is conformably overlain by lower–middle Permian strata, which include, from bottom to top, conglomerates, pebbly coarse sandstone, sandstone and siltstone, with volcanic interlayers. Mesozoic terrigenous clastic rocks are extensively distributed in this area (Fig. 1c).

Phanerozoic plutons are voluminous and widely exposed in the southwestern Alxa Block (Fig. 1c), with two magmatic periods of Middle Ordovician–Early Devonian and early Carboniferous–late Permian. Plutons of the earlier period are generally felsic granitoids (Qin, 2012; Wei et al. 2013; Tang, 2015; Liu et al. 2016 b; Zhou et al. 2016; Zhang et al. 2018d ; Wang et al. 2020), with only a few dolerite dykes (c. 424 Ma) in eastern Longshoushan (Duan et al. 2015). In contrast, plutons of the later period are widely distributed and include peridotite, gabbro, diorite, tonalite, granodiorite, monzogranite and granite (Chen et al. 2013; Jiao et al. 2017; Liu et al. 2017; Xue et al. 2017; Gong et al. 2018 a, b; Huo, 2019; Song et al. 2019). In addition, several Triassic plutons crop out in the western Beidashan (Fig. 1c; Gu, 2012).

3. Samples and petrography

In this study, three granitic plutons were investigated and sampled in the southwestern Alxa Block; all are massive and salmon-pink to off-white in colour (Fig. 3). A medium- to coarse-grained alkali-feldspar granite in western Beidashan (17WAL-07; Fig. 1c) is composed of quartz (c. 30%), plagioclase (c. 20%), alkali-feldspar (c. 40%), biotite (c. 10%) and minor hornblende (Fig. 3b). The other two plutons are located in Longshoushan to the north of Shandan County (Fig. 1c). One is medium-grained granodiorite (17WAL-35) and composed of quartz (c. 20%), plagioclase (c. 40%), alkali-feldspar (c. 20%) and biotite (c. 20%; Fig. 3d). The other sample is coarse-grained monzogranite (17WAL-39), with similar mineral assemblage of quartz (c. 25%), plagioclase (c. 25%), alkali-feldspar (c. 30%) and biotite (20%; Fig. 3f). Accessory minerals of zircon, apatite and titanite are present in all three plutons.

Fig. 3. Field photographs and mineral assemblages under microscope (cross-polarized light) of the studied late early Carboniferous plutons in the southwestern Alxa Block. (a, b) 17WAL-17, alkali-feldspar granite; (c, d) 17WAL-35, granodiorite; (e, f) 17WAL-39, monzogranite. Afs – alkali-feldspar; Bt – biotite; Hbl – hornblende; Pl – plagioclase; Qtz – quartz.

4. Analytical methods

4.a. Whole-rock major- and trace-element analyses

Fresh granitoid samples were first crushed and then ground to 200 mesh in a tungsten carbide cup and ball mill, and then analysed geochemically at the National Research Center of Geoanalysis, China Geological Survey. Whole-rock major-element oxides were measured using a Malvern Panalytical Axios PW4400 x-ray fluorescence spectrometer (XRF), and the analytical uncertainties are generally between 1% and 5%. The concentrations of trace and rare earth elements were determined by a PerkinElmer NexION 300Q inductively coupled plasma mass spectrometer (ICP-MS), with analytical precision generally better than 5%.

4.b. Zircon U–Pb dating

Zircon grains were firstly separated by conventional heavy liquid and magnetic techniques, and then hand-picked under a binocular microscope. The selected zircon crystals were mounted in epoxy resin and polished to half thickness. Potential analytical spots were determined based on morphological features and internal structures of zircons on optical and cathodoluminescence (CL) images. Zircon U–Pb analyses on mineral separates from the three samples were conducted in Tianjin Institute of Geology and Mineral Resources, China Geological Survey, China. A Thermo Fisher Scientific multi-collector inductively coupled plasma mass spectrometer (MC-ICP-MS; Neptune) was coupled to a New Wave 193 nm ArF excimer laser ablation system. Detailed procedures are reported by Cui et al. (2012). Zircon standard GJ-1 was employed as an external standard (Jackson et al. 2004), and measurements of zircon standard Plešovice, which was used as an unknown, yielded a weighted mean ²⁰⁶Pb/²³⁸U age of 335.5 ± 2.6 Ma (n = 12; 2σ). This result is in good agreement with the recommended value within error (337.13 ± 0.37 Ma; Sláma et al. 2008). The corrections of common lead were carried out using the method of Andersen (2002). Concordia diagrams and ages were obtained using ISOPLOT 4.15 (Ludwig, 2012). Uncertainties of individual measurements were at the 1σ level, but the weighted mean ages and concordia diagrams were given at the 2σ level (95% confidence level).

4.c. Sr–Nd isotopic analyses

The whole-rock Sr and Nd isotopic compositions were determined using a Finnigan MAT-262 mass spectrometer and a Nu Plasma high-resolution MC-ICP-MS, respectively, at the Institute of Geology, Chinese Academy of Geological Sciences, China. The measured ⁸⁷Sr/⁸⁶Sr ratio of the SrCO₃ standard SRM 987 was 0.710243 ± 0.000012 (2σ), in good agreement with the recommended value within error (0.710251 ± 0.000018; Coombs et al. 2004). Two standards of JMC Nd₂O₃ (reference value = 0.511137 ± 0.000008; Jahn et al. 1980) and GSB 04-3258-2015 (certified value = 0.512438; Tang et al. 2017) were employed during Nd isotopic analyses, with measured ¹⁴³Nd/¹⁴⁴Nd ratios of 0.511123 ± 0.000010 and 0.512441 ± 0.000012 at the 2σ level, respectively. Detailed analytical procedures for both Sr and Nd isotopic compositions are described by Tang et al. (2021). All measured ratios were corrected for mass fractionation by normalizing to ⁸⁸Sr/⁸⁶Sr = 8.37521 and ¹⁴⁶Nd/¹⁴⁴Nd = 0.7219, respectively.

5. Results

Whole-rock major- and trace-element concentrations, LA-ICP-MS zircon U–Pb data and Sr–Nd isotopic compositions are given in online Supplementary Tables S1–S3 (available at http://journals.cambridge.org/geo), respectively.

5.a. Whole-rock major and trace elements

All three plutons have high SiO₂ (68.49–77.01 wt%) and K₂O + Na₂O (8.07–8.25 wt%; Fig. 4a) and low MgO (0.17–0.82 wt%) and MnO (0.03–0.06 wt%), show peraluminous features (A/CNK = 1.04–1.13), and belong to the high-K calc-alkaline series (Fig. 4b). Alkali-feldspar granite 17WAL-07 and monzogranite 17WAL-39 display lower CaO (0.59–0.61 wt%), higher K₂O (K₂O/Na₂O = 1.35–1.55), higher total rare earth element (REE) concentrations (257.58–275.96 ppm) and distinct negative Eu anomalies (δEu = 0.17–0.37; Fig. 5a), with enrichments in large-ion-lithophile elements (LILEs; e.g. Cs, Rb, Th and Pb) and depletions in Nb, Ta, Ba and Sr (Fig. 5b). In comparison, granodiorite 17WAL-35 displays relatively higher CaO (1.52–2.51 wt%) and lower total REE concentrations (114.30–206.16 ppm), with significantly enriched light rare earth elements (LREEs; (La/Yb)_N = 35.75–56.24) and positive Eu anomalies (δEu = 1.12–1.14; Fig. 5c). Moreover, it is characterized by enriched LILEs (Cs, Rb, Ba, Th, Pb and Sr) and depleted high-field-strength elements (HFSEs; Y, Yb and Lu), with negative Nb–Ta and positive Zr–Hf anomalies, respectively (Fig. 5d).

Fig. 4. (a) Na₂O + K₂O versus SiO₂ and (b) K₂O versus SiO₂ diagrams for the early Carboniferous plutons in the Alxa Block. Data sources include Wang et al. (2015 b), Dan et al. (2016), Liu et al. (2016 a) and Xue et al. (2017).

Fig. 5. (a, c) Chondrite-normalized REE patterns and (b, d) primitive mantle-normalized trace-element diagrams for the late early Carboniferous plutons in the southwestern Alxa Block. Compositions of C1 chondrite and primitive mantle after Sun & McDonough (1989).

5.b. Zircon U–Pb ages

Zircon grains from the studied samples are transparent, euhedral and short columnar or prismatic in shape. They exhibit well preserved concentric magmatic oscillatory zoning, with a few inherited zircon cores appearing occasionally in samples 17WAL-35 and 39 (Fig. 6). For alkali-feldspar granite 17WAL-07, all 24 spots are concordant and cluster together (Fig. 7a). Their Th/U ratios are 0.33–0.51 and they yield a concordia age of 331.6 ± 1.6 Ma (mean square weighted deviation (MSWD) = 4.2; 2σ, decay-constant errors included), which is consistent with the weighted mean ²⁰⁶Pb/²³⁸U age (331.7 ± 1.5 Ma; MSWD = 1.01; 2σ). With the exception of four discordant spots (16, 17, 18 and 22), concordant analyses of the other 21 granodiorite 17WAL-35 spots have Th/U ratios of 0.36–0.83 but form two age clusters (Fig. 7b). The older population includes 17 spots with a weighted mean ²⁰⁶Pb/²³⁸U age of 344.1 ± 2.2 Ma (MSWD = 1.40; 2σ; Fig. 7b1) and the younger population includes 4 spots with a weighted mean ²⁰⁶Pb/²³⁸U age of 326.2 ± 6.6 Ma (MSWD = 1.50; 2σ; Fig. 7b2). Furthermore, monzogranite 17WAL-39 has six discordant spots (1, 8, 11, 16, 17 and 21) and one concordant age cluster (Fig. 7c), which yields a consistent concordia age of 331.8 ± 1.7 Ma (MSWD = 4.8; 2σ, decay-constant errors included) and weighted mean ²⁰⁶Pb/²³⁸U age of 331.9 ± 1.7 Ma (MSWD = 0.88; n = 17; 2σ), with Th/U ratios of 0.43–1.03.

Fig. 6. Cathodoluminescence (CL) images of representative zircon grains from the studied late early Carboniferous plutons in the southwestern Alxa Block.

Fig. 7. (a–c) Concordia diagrams showing LA-ICP-MS zircon U–Pb data of the studied late early Carboniferous plutons in the southwestern Alxa Block (all the diagrams and calculations are at the 2σ level).

5.c. Whole-rock Sr–Nd isotopes

The ⁸⁷Rb/⁸⁶Sr and ¹⁴⁷Sm/¹⁴⁴Nd ratios of three granitic samples were calculated using the measured whole-rock Rb, Sr, Sm and Nd concentrations. The alkali-feldspar granite (17WAL-07; t = 332 Ma) has the lowest initial ⁸⁷Sr/⁸⁶Sr (0.700128) and highest initial ¹⁴³Nd/¹⁴⁴Nd (0.512219) ratios among the three plutons, with positive ϵ_Nd(t) value (0.16) and Mesoproterozoic Nd model age (T _DM = 1207 Ma; Fig. 8). The initial ⁸⁷Sr/⁸⁶Sr ratio of the granodiorite (17WAL-35; t = 326 Ma) is low (0.705102), and its initial ¹⁴³Nd/¹⁴⁴Nd ratio and ϵ_Nd(t) value are 0.511358 and −16.80, respectively (Fig. 8). As its ƒ_Sm/Nd (−0.59) significantly deviates from that of the average crust (−0.40; DePaolo et al. 1991), both T _DM (1847 Ma) and T _DM2 (2446 Ma) were calculated. For the monzogranite (17WAL-39; t = 332 Ma), its initial ⁸⁷Sr/⁸⁶Sr and ¹⁴³Nd/¹⁴⁴Nd ratios are 0.717670 and 0.511706, respectively, with negative ϵ_Nd(t) value (−9.85) and Palaeoproterozoic T _DM2 (1889 Ma; Fig. 8).

Fig. 8. Sr–Nd isotopic features of early Carboniferous plutons in the Alxa Block. Symbols and data sources as for Figure 4.

6. Discussion

The well preserved concentric magmatic oscillatory zoning (Fig. 6) and high Th/U ratios (0.33–1.03) of dated zircon grains indicate their magmatic origin (Corfu et al. 2003); the concordia and weighted mean ²⁰⁶Pb/²³⁸U ages are therefore interpreted as crystallization ages (Fig. 7). Because several spots from the older age cluster of granodiorite (17WAL-35) are located within the inherited zircon cores (e.g. spot 24 in Fig. 6b), the younger age cluster is employed. The three granitic plutons in the southwestern Alxa Block were therefore formed during late early Carboniferous time (c. 332–326 Ma).

6.a. Petrogenesis of the studied late early Carboniferous granitic plutons

The alkali-feldspar granite (17WAL-07) and monzogranite (17WAL-39) have similar geochemical features, such as high K₂O + Na₂O (8.10–8.25 wt%), FeO^T (1.49–1.51 wt%) and FeO^T/MgO (4.38–8.87), low CaO (0.59–0.61 wt%), MgO (0.17–0.43 wt%) and P₂O₅ (< 0.06 wt%), high total REE concentrations (257.58–275.96 ppm) with V-type REE patterns (Fig. 5a), and strongly depleted Ba and Sr (Fig. 5b). These characteristics indicate A-type granite nature, which can be clearly identified on the discrimination diagrams (e.g. Fig. 9b, c; Whalen et al. 1987; King et al. 1997). A-type granites may originate from the fractionation of mantle-derived basaltic magmas (Eby, 1990, 1992; Bonin, 2007), the mixing of mantle- and crust-derived magmas (Yang et al. 2006), or the partial melting of crust at high temperatures (Whalen et al. 1987; King et al. 1997; Wu et al. 2002). If rhyolitic magmas were derived from fractional crystallization of coeval basaltic magmas, the two components would commonly be spatially and temporally associated (Whitaker et al. 2008). If the plutons had their origin by magma mixing, then they would have intermediate compositions with the presence of profuse mafic microgranular enclaves (MMEs; Yang et al. 2006, 2007; Zhang et al. 2016 b), although the MMEs may be also cogenetic with their host granitoids (Zhang & Zhao, 2017). The two A-type granites in the southwestern Alxa Block are rhyolitic in composition (Fig. 4a), but no MMEs were observed (Fig. 3a, e) and their coeval mafic intrusions crop out far away in the northeastern Alxa Block (Wang et al. 2015 b; Liu et al. 2016 a). They are also characterized by high SiO₂ (73.89–77.01 wt%) and K₂O/Na₂O (1.35–1.55) and are peraluminous (A/CNK = 1.04–1.13), similar to aluminous A-type granites with continental crustal sources (King et al. 1997). Moreover, the alkali-feldspar granite has low positive ϵ_Nd(t) value (0.16) and Mesoproterozoic Nd model age (1207 Ma; Fig. 8b), which is close to the protolith crystallization age of a granitic gneiss in the Helishan (c. 1200 Ma; Song et al. 2017). Its unusually low initial ⁸⁷Sr/⁸⁶Sr value (0.700128; Fig. 8a) may be caused by the strong depletion of Sr (Fig. 5b), as the initial ⁸⁷Sr/⁸⁶Sr value was calculated based on the measured whole-rock Sr concentration. The monzogranite has radiogenic Sr–Nd isotopes (Fig. 8a) and a Palaeoproterozoic Nd model age (1889 Ma; Fig. 8b). The Palaeoproterozoic basement rocks are commonly observed in Longshoushan (Tung et al. 2007; Gong et al. 2011), in addition to a c. 1872 Ma syenite in Helishan (Wang et al. 2019 b). The two aluminous A-type granites were therefore most probably the high-temperature partial melts of Palaeo- and Mesoproterozoic crustal materials.

Fig. 9. (a) Tectonic discrimination diagrams of Rb versus (Y + Nb) for the early Carboniferous felsic plutons in the Alxa Block (Pearce, 1996). (b) Plot of (K₂O + Na₂O)/CaO versus Zr + Nb + Ce + Y and (c) plot of Ce versus 10 000×Ga/Al for A-type granites (Whalen et al. 1987). (d) Nb–Y–Ce diagram for distinguishing between A1 and A2 granites (Eby, 1992). Symbols and data sources as for Figure 4.

The granodiorite (17WAL-35) is also high-K calc-alkaline (Fig. 4b) and weakly peraluminous (A/CNK = 1.07–1.08) and has depleted HREEs and HFSEs (Fig. 5c, d). It is chemically characterized by high Sr (522.0–918.0 ppm) and low Y (5.36–6.56 ppm) and Yb (0.62–0.75 ppm) concentrations, with high Sr/Y ratios (97.4–139.9). Although high Sr/Y ratio (> 40) usually occurs in adakitic rocks, the high K₂O contents (3.59–4.12 wt%) and K₂O/Na₂O ratios (0.80–1.03) of this granodiorite are more ‘continental’ than typical adakites (Defant & Drummond, 1990; Martin et al. 2005; Moyen, 2009). The coexistence of negative Nb–Ta and positive Zr–Hf anomalies (Fig. 5d) and highly radiogenic Sr–Nd isotopes (Fig. 8a) also suggest a continental crustal source (Rudnick & Gao, 2003). The enrichments of Eu, Ba and Sr are attributed to the large proportion of plagioclase (c. 40%), whereas the low Y concentration may suggest the presence of garnet in the residue, so that the high Sr/Y ratios indicate a deeper crustal level of magma source (Ducea et al. 2015). In addition, c. 2.5 Ga basement rocks and magmatic activity are commonly observed in the southwestern Alxa Block (Zhang et al. 2013 a; Zhang & Gong, 2018; Wang et al. 2019 b), which is coeval with the two-stage Nd model age of this granodiorite (c. 2446 Ma; Fig. 8b). This granodiorite of high Sr/Y ratio may therefore have its origin in the partial melting of upper Neoarchean lower crust.

6.b. Tectonic setting of the early Carboniferous magmatism in the Alxa Block

Two different tectonic processes accounting for the early Carboniferous magmatism within the Alxa Block were proposed previously: continental arc magmatism induced by the S-wards subduction of the PAO (Liu et al. 2016 a; Xue et al. 2017; Gong et al. 2018 a), or the collision and amalgamation between the Alxa Block and the NCC (Zhang et al. 2013 b; Dan et al. 2016). Noticeably, whether a Palaeozoic suture between the Alxa Block and the NCC existed or not is still in debate, especially with no associated ophiolitic mélanges observed (e.g. Dan et al. 2016; Zhang & Gong, 2018; Wang et al. 2019 b), and the early Carboniferous magmatic rocks are widely distributed, rather than along a linear trend in the eastern margin of the Alxa Block (Fig. 1b), so they are less likely attributed to such an amalgamation process. Furthermore, the argument of continental arc magmatism is mainly based on their arc-like geochemical signatures, such as calc–alkaline characteristics (Fig. 4b), negative Nb–Ta anomalies and high Sr/Y ratios (e.g. Liu et al. 2016 a; Xue et al. 2017). However, these signatures can also be inherited from magma sources (Wang et al. 2016 a), and most granites of high Sr/Y ratio in this area exhibit high K₂O/Na₂O ratios (0.92–3.70), positive Zr–Hf anomalies and radiogenic Nd–Hf isotopes, indicating derivation by the partial melting of lower continental crust (Fig. 8a; Dan et al. 2016; Xue et al. 2017); this can occur not only in continental arc belts but also in lithospheric extensional environments.

It is noteworthy that the early Carboniferous plutons within the Alxa Block are mostly basic or acidic in silica content (Fig. 4), resembling bimodal associations. The felsic plutons plot not only in volcanic arc but also in within-plate and post-collision granite fields (Fig. 9a), with most of them exhibiting radiogenic Sr–Nd isotopes (Fig. 8a). They are characterized by the coexistence of A-type granites, peraluminous granites and calc-alkaline I-type granitoids (Dan et al. 2016; Liu et al. 2016 a; Xue et al. 2017; Zheng et al. 2019), which mostly occur in extensional settings (Maniar & Piccoli, 1989). A-type granites usually indicate high-temperature anatectic conditions related to asthenospheric upwelling in a lithospheric extensional setting (Whalen et al. 1987; Eby, 1992). The mafic plutons plot mostly in the MORB and within-plate basalt fields, similar to the rift-related Basin-and-Range basalts (Fig. 10), and display juvenile or weakly radiogenic Sr–Nd isotopes (Fig. 8a). It is noteworthy that several of the mafic plutons in the northeastern Alxa Block have hornblende as the dominant mafic mineral and resemble appinitic intrusions in geochemistry (Wang et al. 2015 b). Generally, mafic appinitic melts were most likely produced by the partial melting of subduction-modified sub-continental lithospheric mantle (Fig. 10c) and the melting may be triggered by asthenospheric upwelling following slab break-off or delamination after a subduction event (Murphy, 2013). The generation of both the mafic and felsic early Carboniferous plutons within the Alxa Block therefore most likely resulted from the asthenospheric upwelling at that time. Although an upwelling asthenosphere may also occur in a continental arc setting, continental arc magmatism is typically characterized by linear tracks within a specific tectonic unit and dominated by andesitic rocks, with continued major elemental compositions from basalts to rhyolites but without compositional gaps (Ducea et al. 2015). Evidently, this is not the case for the early Carboniferous plutons within the Alxa Block (Figs 1b, 4a), meaning that their formation in a continental arc is less likely, but rather more likely in a lithospheric extensional setting.

Fig. 10. Petrogenetic discrimination diagrams of (a) V–(Ti/1000) (Shervais, 1982), (b) (Zr/Y)–Zr (Pearce & Norry, 1979), (c) (La/Ba)–(La/Nb) (Saunders et al. 1992), and (d) (Zr/Sm)–(Sr/Nd)–(Ti/V) (Wang et al. 2016 a) for the early Carboniferous mafic rocks in the Alxa Block. The Basin-and-Range rift-related basalt field refers to Wang et al. (2016 a). Symbols and data sources as for Figure 4.

Furthermore, A-type granites are a good indicator of lithospheric extension, but the specific extensional setting could be varied (Sain et al. 2017), including not only rift-related (intraplate) extension (Whalen et al. 1987; Eby, 1992) but also back-arc extension (Karsli et al. 2012; Bickford et al. 2015). The two early Carboniferous aluminous A-type granites in the southwestern Alxa Block are A2 type (Fig. 9d) and therefore represent magmas derived from continental crust that has been through an orogenic cycle of arc magmatism and collision (Eby, 1992). The geochemical similarities between early Carboniferous mafic plutons in the Alxa Block and Basin-and-Range basalts (Fig. 10), which were generated in back-arc extensional setting to the Sierra Nevada arc (Cousens et al. 2019), also suggest a subduction-related tectonic setting. In back-arc extensional setting, the asthenospheric upwelling could be induced by the foundering of arc root during the roll-back process of subducting slab (DeCelles et al. 2009; DeCelles & Graham, 2015). Another possibility is the intra-continental extensional setting, because the sub-continental lithospheric mantle and lower continental crust of the Alxa Block had been modified by subduction during Middle Ordovician–Early Devonian time (Liu et al. 2016 b; Zhou et al. 2016), and the subduction-related geochemical signatures of later magmas may be inherited from the subduction-modified magma sources (Wang et al. 2016 a). Moreover, the extension-related rock associations of calc-alkaline I-type granites, aluminous A2-type granites and peralkaline granites were present in the southwestern Alxa Block from late Silurian–Early Devonian time, following earlier arc magmatism and implying post-collisional setting (Wang et al. 2020). In addition, the cyclical magmatic flare-ups and lulls within each Palaeozoic magmatic stage of the Alxa Block (Fig. 2a) are quite similar to those of Cordilleran arcs in terms of time span and frequency (DeCelles et al. 2009), but the magmatic hiatus between the two magmatic stages is relatively too long for one single subduction event. The two magmatic stages of the Alxa Block may therefore represent two orogenic cycles and the early Carboniferous extension, as the initiation of the second orogenic cycle, may suggest intra-continental extensional setting. Although more geological evidence is urgently needed to discriminate between the two kinds of extensional settings, a simple continental arc model is less likely for the early Carboniferous magmatism within the Alxa Block.

Additionally, continental arc magmatism is usually accompanied by syn-arc sedimentation in fore-arc or back-arc basins (Ducea et al. 2015), but lower Carboniferous strata are absent from the Alxa Block based on available geological reports. Although a few outcrops in the northern Alxa Block were previously identified as lower Carboniferous deposits, they were recently reassigned as lower–middle Permian strata (Zhang et al. 2018c ). By contrast, the upper Carboniferous–middle Permian strata are widely distributed. The sedimentary facies show a distinct change from terrestrial alluvial fan and delta in the lower stratigraphic sections to platform, littoral and shallow-marine in the upper stratigraphic sections, with abundant fossils (e.g. plants, fusulinids, brachiopods, corals) and volcanic interlayers (Bu et al. 2012; Han et al. 2012; Yin et al. 2016; Song et al. 2018). Such a transgression sequence is consistent with the further development of the lithospheric extension.

6.c. Tectonic implications for the development of southeastern CAOB

Even if the Alxa Block was separated from the NCC during the Precambrian Eon, sedimentologic, magmatic and structural evidences (Li et al. 2012 a; Dan et al. 2016; Zhang et al. 2013 b, 2016 c) all suggest that their amalgamation occurred before early Carboniferous time. Palaeomagnetic studies also suggest that the Precambrian micro-continental blocks within the southeastern CAOB (e.g. Mongolia, Songliao and Hunshandake blocks) may have already accreted to the northern NCC by early Carboniferous time (Pruner, 1992; Li et al. 2012 b; Zhao et al. 2013; Zhang et al. 2018 a). Furthermore, the Palaeozoic magmatic episodes of the Alxa Block and the southeastern CAOB (including the northern margin of the NCC) are very similar (Fig. 2), indicating comparable tectonic processes. Consequently, the whole region had been experiencing a uniform tectonic regime since early Carboniferous time and, if there was on-going S-wards subduction of the large-scale PAO at that time, the arc-trench system was most likely located to the north of these micro-continental blocks.

Regionally, the early Carboniferous is the initial period of the second magmatic stage (Fig. 2), and magmatic rocks during this period are characterized by the mafic–ultramafic complexes in northern Inner Mongolia (Jian et al. 2012; Zhang et al. 2015c ; Li et al. 2018), the appinitic intrusions in the northern NCC (Zhou et al. 2009; Zhang et al. 2012 a; Wang et al. 2015 b), the calc-alkaline I-type and peraluminous granites with crustal origins throughout the southeastern CAOB (Bao et al. 2007; Zhang et al. 2007, 2011; Liu et al. 2009, 2016 a; Blight et al. 2010; Dan et al. 2012; Xue et al. 2017), and the A-type granites newly identified in the southwestern Alxa Block (this study). Such rock associations are commonly associated with asthenospheric upwelling in lithospheric extensional setting. Although some of the basaltic rocks from the mafic–ultramafic complexes exhibit subduction-related geochemical features (Jian et al. 2012; Zhang et al. 2015c ; Li et al. 2018), these features can also be imprinted by crustal contamination (Xia, 2014) or inherited from magma sources that have been modified by earlier subduction fluids or melts (Wang et al. 2016 a). Further, the coeval intrusions are widely distributed (Xu et al. 2014) rather than along one or two specific ribbons as would be expected for a magmatic arc, supporting their formation in an extensional tectonic setting. Moreover, if this lithospheric extension occurred in back-arc, then the remnants of the large-scale PAO may be represented by the early Carboniferous Erenhot–Hegenshan ophiolitic mélanges to the north of the micro-continental blocks (Zhang et al. 2015c ; Li et al. 2018). Otherwise, the early Carboniferous extension of the southeastern CAOB was probably developed in an intra-continent environment and may represent the initiation of the second orogenic cycle (Xu et al. 2018).

In addition to the intrusions, the early Carboniferous sedimentary rocks are mostly absent from the southeastern CAOB, indicating regional uplift related to asthenospheric upwelling during the initial stage of the lithospheric extension. The Carboniferous metamorphic rocks are high-temperature–low-pressure and show a clockwise P–T path, involving pre-peak heating with slight decompression, peak and post-peak cooling stages, also suggesting an extension process (Zhang et al. 2018 b).

Subsequently, the late Carboniferous–Permian magmatism in the southeastern CAOB became intense (Fig. 2) with the formation of the widespread bimodal volcanic rocks, continental basaltic intrusions, calc-alkaline I-type granites, peraluminous S-type granites, A-type granites and several peralkaline magmatic belts (e.g. Jahn et al. 2009; Zhang et al. 2012 b, 2015 b, 2016 d, 2017 b; Pang et al. 2016, 2017; Zhao et al. 2016 a; Ji et al. 2018; Wang et al. 2021 b), implying further development of the early Carboniferous extension. This is also consistent with the occurrence of many late Carboniferous–Permian mafic dykes (Fig. 3a) with MORB or within-plate basalt geochemical signatures in this region (Lin et al. 2014). Accordingly, the late Carboniferous–Permian Solonker, Enger Us and Quagan Qulu ophiolitic mélanges (Jian et al. 2010; Zheng et al. 2014), which contain MORB-type intrusions, continental basalts and terrigenous sediments (Luo et al. 2016; Shi et al. 2016), may represent the newly opened limited ocean basins and mark the strongest extension (Xu et al. 2014, 2018). The late Carboniferous–Permian sedimentary sequences are also widely exposed throughout the southeastern CAOB. They vary from plant fossil-bearing terrigenous clastic rocks to shallow-marine clastic and carbonate depositions, with basal conglomerates, and are transgression sequences related to regional extension (Zhao et al. 2016 b; Ji et al. 2020; Wang et al. 2021 a).

To summarize, we propose a lithospheric extensional process rather than a simple continental arc for the tectono-magmatic development of the southeastern CAOB during early Carboniferous time (Fig. 11). The early Carboniferous extension-related magmatism and the absence of coeval sedimentary successions may reflect the onset of asthenospheric upwelling and regional uplift, and therefore mark the initiation of the lithospheric extension. Nevertheless, the asthenospheric upwelling could be induced by either slab roll-back or slab break-off of the subducted PAO; more geological, geochemical, geophysical and palaeontological evidence is therefore needed to further constrain the specific tectonic setting of this extension, either back-arc or intra-continental.

Fig. 11. Extensional tectonics of the Alxa Block and the southeastern CAOB during early Carboniferous time. (a) Micro-continental blocks within the southeastern CAOB had already been accreted to the northern NCC (Alxa Block) before early Carboniferous time. (b) During early Carboniferous time, the asthenospheric upwelling induced by either the roll-back or the break-off of the subducted PAO slab heated both the subduction-modified lithospheric mantle and the overlying crust, leading to the generation of the mafic and felsic plutons, respectively.

7. Conclusions

The early Carboniferous (c. 332–326 Ma) granodiorite with high Sr/Y ratio, A-type monzogranite and A-type alkali-feldspar granite in the southwestern Alxa Block were most likely formed by partial melting of Neoarchean, Palaeoproterozoic and Mesoproterozoic crustal sources heated by upwelling asthenosphere in an lithospheric extensional setting. According to regional geological correlations, a uniform lithospheric extensional setting, either back-arc or intra-continental, but not a simple continental arc, is suggested for both the Alxa Block and the southeastern CAOB during early Carboniferous time, with the development of extension-related magmatism and the absence of coeval sedimentary rocks.