Petrogenesis and tectonic setting of the Early Cretaceous granitoids in the eastern Tengchong terrane, SW China: Constraint on the evolution of Meso-Tethys

Granites, as results of tectono-thermal events, display great diversity due to the variety of their sources, evolution processes, and emplacement within different tectonic regimes and geodynamic environments (Barbarin, 1999). Therefore, granites can provide important insights into tectonic settings and crust-mantle interaction within orogenic belts (Xu et al., 2008). Accordingly, the Early Cretaceous granitoids widespread in the Tengchong terrane, SW China, are prime records for understanding their tectonic environments and Sn mineralization related to synchronous magmatism (Cong et al., 2011a, 2011b; Qi et al., 2011; Luo et al., 2012; Xu et al., 2012; Cao et al., 2014; Zhu et al., 2015; Xie et al., 2016; Fang et al., 2018; Zhang et al., 2018; Qi et al., 2019). On the basis of close similarity in geochemical characteristics and chronology of these igneous rocks, geologists proposed that the Tengchong terrane is most likely linked with the Lhasa terrane and experienced similar tectonic histories since the Early Paleozoic (Xu et al., 2008, 2012; Xie et al., 2016; Qi et al., 2019). Compared with the magmatism in the Lhasa terrane, the Early Cretaceous magmatism in the Tengchong terrane is considered as the eastern extension of the Gangdese batholith which is located in the southern Lhasa terrane (Xie et al., 2016; Qi et al., 2019). However, the dynamic setting of Early Cretaceous magmatism in the eastern Tengchong terrane and its tectonic affinity remain debated. For instance, some researchers suggest that abundant Early Cretaceous magmatism in the Tengchong terrane was related to the southward subduction of the Bangong-Nujiang Meso-Tethyan Ocean (Qi et al., 2011; Xu et al., 2012; Zhu et al., 2015, 2017a, 2017b, 2018; Qi et al., 2019). Others propose post-collisional settings to interpret the generation of these igneous rocks (Yang et al., 2006; Luo et al., 2012; Xu et al., 2012; Cao et al., 2014). Recently, identification of mafic enclaves (Cong et al., 2010; Zhang et al., 2018; Qi et al., 2019) and the Middle Jurassic Myitkyina ophiolitic suite suggest that the Bangong-Nujiang Ocean in the Tibetan Plateau extended southward into the Myitkyina-Mogok area in Myanmar, to the west of the Tengchong terrane (see Fig. 1B; Liu et al., 2016a, 2016b), rather than Gaoligong shear zone (see Fig. 1B). Subsequently, the geodynamic setting responsible for the Early Cretaceous magmatism and tectonic affinity of the Tengchong terrane needs to be re-evaluated.

In this study we present new zircon laser ablation–inductively coupled plasma–mass spectrometry (LA-ICP-MS) U-Pb dating data, whole-rock geochemical, and Nd isotopic results for Early Cretaceous granites (124 Ma) from the Tengchong terrane. Combined with previous data of granitoids and mafic enclaves in this area, we provide new insights into their petrogenesis, and the dynamic setting of the Early Cretaceous magmatism, to understand the evolution of Meso-Tethys and the tectonic affinity of the Tengchong terrane.

The Tengchong terrane is located in the southern margin of the Tibetan Plateau (see Fig. 1A) and is bound by the Gaoligong shear zone to the east and the Myitkyina suture to the west (Replumaz and Tapponnier, 2003; see Fig. 1B). Previous studies on the paleogeographic evolution of eastern Tethys (Cocks and Torsvik, 2013; Metcalfe, 2013; Burrett et al., 2014) suggest that the Tengchong terrane was located in the northern margin of the Gondwana supercontinent during the early Paleozoic, and accreted to the Eurasia plate in the late Mesozoic. Before an amalgamation of Tengchong and Baoshan terranes, there was a Bangong-Nujiang Meso-Tethyan Ocean between the two terranes. This ocean had been closed before the Early Cretaceous (e.g., Fan et al., 2018, and references therein) and now is represented by the east-west–trending Bangong-Nujiang suture zone, which crosses the central Tibetan Plateau and the west of the Tengchong terrane and to the south it is correlated with the Shan boundary in Burma (see Fig. 1B). The closure of the Neo-Tethys (leading to the Yarlung-Tsangpo suture) happened after the Late Cretaceous (Xu et al., 2012). The Meso-Neo Proterozoic metamorphic basement, termed the Gaoligongshan Group, is considered as the oldest geological unit in the Tengchong terrane, yielding zircon U-Pb ages of 1053–635 Ma and 490–470 Ma, according to Song et al. (2010) and is mainly composed of amphibolite, gneiss, quartzite, schist, marble, and slate (Zhao et al., 2016a, 2016b). The overlying Paleozoic strata sequence consists of largely glacial-marine diamictite, sandstone, and limestone (Fig. 1C). Numerous Cretaceous–Cenozoic igneous rocks intruded this terrane, which is covered by Paleogene–Quaternary volcanic rocks. Magmatism of this period induced the absence of the Cretaceous–Eocene sedimentary strata (Liu et al., 2009). Two stages of magmatism in the Tengchong terrane have been reported in previous studies: (1) Late Cretaceous to early Cenozoic magmatism (75–65 Ma, Xie et al., 2016) and (2) early Cenozoic magmatism (55–47 Ma, Xu et al., 2012; He et al., 2019a). The terrane can be divided into the eastern and western Tengchong by the Dayingjiang fault (Fig. 1C) and is also cut by several ductile shear zones such as the NNE-trending dextral Nabang shear zone and the prominent N-S–striking dextral Gaoligong shear zone, which, according to ⁴⁰Ar/³⁹Ar dating, separates the Tengchong terrane from the Baoshan terrane and deformed ca. 18–13 Ma (Fig. 1C; Lin et al., 2009; Xie et al., 2016; Cao et al., 2019).

Eleven samples for this study were collected from the Xishanjiao pluton in the central Tengchong terrane (Fig. 2). Samples consist of medium-coarse grained biotite monogranites (XSJ-1, XSJ-2, XSJ-3, XSJ-4, XSJ-5, XSJ-8, XSJ-9) and biotite granodiorites (XSJ-6, XSJ-10, XSJ-11, XSJ-12). A summary of the locations, lithology, ages, and ε_Hf(t) of the samples (including this study) from the Tengchong terrane is listed in Table 1. The representative petrographic features with fine-grained texture and massive structure for granites of this study are shown in Figure 3. Major minerals are quartz (∼30 vol%), K-feldspar (∼20 vol%), plagioclase (∼30 vol%), and biotite (∼15 vol%). Accessory minerals are hornblende, zircon, and apatite. Plagioclase are partly altered to sericite and kaolinite (see Figs. 3D and 3E).

Sample ages were calculated based on LA-ICP-MS U-Pb dating of zircons from the Xishanjiao granites (XSJ-05, XSJ-10). Zircon grains were extracted by heavy liquid and magnetic separation, before being handpicked under a binocular microscope for mounting in epoxy resin. To identify their internal structure and to choose potential target sites for the U-Pb analysis, cathodoluminescence (CL) images (Fig. 4) were obtained using a scanning electron microscope at the Institute of Geochemistry, Chinese Academy of Sciences (IGCAS), Guiyang, China. Measurements of U, Th, and Pb isotopes were conducted using an Agilent 7500a LA-ICP-MS at the State Key Laboratory of Ore Deposit Geochemistry (SKLOG), IGCAS. A GeoLasPro laser-ablation system and an Agilent 7700x inductively coupled plasma–mass spectrometry (ICP-MS) were combined for these experiments. A beam size of 33 μm was used for all samples; zircon 91500 (1062 Ma) and NIST SRM 610 with Si were used as internal standards for external calibration, and Zr was used as internal standard for other trace elements (Liu et al., 2010). GJ-1 (600 Ma) and Plešovice (337 Ma) were treated as quality control for geochronology. Operational and analytical methods are detailed by Li et al. (2009). Measured compositions were corrected for common Pb using non-radiogenic ²⁰⁴Pb. As corrections were sufficiently small to be insensitive to the choice of common Pb composition, an average of present-day crustal composition (Stacey and Kramers, 1975) was used for common Pb, assuming that it is largely related to surface contamination introduced during sample preparation. Uncertainties relating to individual analysis in the data tables are reported at a 1δ level, and mean ages for pooled U/Pb (and Pb/Pb) analyses are quoted at a 95% confidence interval. Data reduction was conducted using the Isoplot/Ex v. 2.49 program (Ludwig, 2001).

The granites were crushed to less than 200 mesh for whole-rock major and trace elemental analyses. Major oxides were measured using an Axios PW4400 X-ray fluorescence spectrometer on fused glass beads at Wuhan Sample Solution Analytical Technology Co., Ltd., China. The analyses were monitored by international standard references AGV-1, BCR-2, and BHVO-1, which showed that the analytical errors are <2%. Trace elements were analyzed using ICP-MS at the State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences (Wuhan), following procedures established by Liu et al. (2008).

High-precision Nd isotope measurements were performed on a Triton thermal ionization mass spectrometer at SKLOG, IGCAS. Samples for Nd isotopic analyses were dissolved in an acidic mixture of 0.5 mL 40 wt% HNO₃ and 1.0 mL HF 40 wt% in Teflon bombs that were steel-jacketed and placed in the oven at 195 °C for three days. Digested samples were dried down on a hotplate, reconstituted in 1.5 mL of 1.5N HCl before ion exchange purification. Sample chemical separation was conducted following Pu et al. (2005). Mass fractionation corrections for Nd isotopic ratios were based on ¹⁴⁶Nd/¹⁴⁴Nd = 0.7219. The ¹⁴³Nd/¹⁴⁴Nd ratios of the La Jolla and JNDI-1 Nd standard solutions were 0.511841 ± 3 (2σ) and 0.512104 ± 5 (2σ) (Tanaka et al., 2000; Weis et al., 2006).

Two representational samples (XSJ-05, XSJ-10) were selected from the Xishanjiao granites in the Tengchong terrane for zircon U-Pb dating. Seventeen and sixteen reliable U-Pb age analytical spots were obtained from XSJ-05 and XSJ-10, respectively. Complete U-Pb isotopic data for these zircons are presented in Table 2. LA-ICP-MS data for XSJ-05 and XSJ-10 yielded weighted mean average ²⁰⁶Pb/²³⁸U ages of 124.2 ± 0.5 Ma (1σ, mean square weighted deviation [MSWD] = 0.15) and 124.3 ± 0.6 Ma (1σ, MSWD = 0.11) (Fig. 4), respectively. In the CL images (Fig. 4), most zircon grains show oscillatory zoning, typical of igneous zircons (Wu and Zheng, 2004). All of the analyzed spots have high U (251–1093 ppm) and Th (150–1164 ppm) contents with Th/U ratios of 0.49–2.48 (greater than 0.1), which suggests that these zircons have a magmatic origin (Belousova et al., 2002; Wu and Zheng, 2004). Therefore, we interpret these zircon U-Pb ages as the timing of crystallization of the Xishanjiao granites.

Major and trace elemental data of all granite samples are listed in Table 3. In the total alkali versus silica diagram (Fig. 5A), seven samples are plotted in the granite area whereas the others are in the granodiorite area. The Xishanjiao granites have variable contents of SiO₂ (65.8–74.6 wt%), K₂O (1.82–4.89 wt%), K₂O/Na₂O (0.38–1.29 wt%), and K₂O+Na₂O (6.91–8.21 wt%) and relatively high contents of TiO₂ (0.23–0.59 wt%), CaO (0.18–3.41 wt%), P₂O₅ (0.03–0.14 wt%), Fe₂O₃^T (1.67–4.51 wt%), and MgO (0.12–1.64 wt%). They also have relatively low Al₂O₃ contents (12.7–15.1 wt%), A/CNK indexes [molar Al₂O₃/(CaO+Na₂O+K₂O)] (0.96–1.25; Fig. 5B), and Rittmann indexes (σ) ranging between 1.60 and 2.06, suggesting that these granites are metaluminous to weakly peraluminous and high-K calc-alkaline series (Fig. 5C). Trace element analysis of the Xishanjiao granites show similar chondrite-normalized rare-earth element (REE) patterns (Fig. 6A) inclining to the right, showing relatively high total REE contents of 114–202 ppm, enrichment in light REE (102–179 ppm) with relatively negative Eu anomalies (δEu = 0.32–0.66), light REE/high REE ratios of 3.82–10.52, (La/Yb)_N ratios of 2.41–11.35, and Sr/Y ratios of 2.06–6.97. The primitive mantle-normalized trace element spider diagram (Fig. 6B) shows relatively marked depletions in large ion lithophile elements (LILE) such as Ba and Sr and high field strength elements (HFSE) such as Th, U, Nb, P, Ti. On the Harker diagrams (Fig. 7), the Al₂O₃, Fe₂O₃^T, CaO, MgO, TiO₂, and P₂O₅ contents for the Xishanjiao samples exhibit decreasing trends with increasing SiO₂ contents, which is consistent with previously published available data.

The results of Nd isotopic compositions of eleven granite samples are presented in Table 4. The initial ε_Nd(t) values were calculated at t = 124 Ma. In addition, depleted mantle Nd model ages (T_DM) were calculated using the De Paolo’s model (1981). Nd isotopic compositions of the Xishanjiao granites are characterized by relatively homogenous ε_Nd(t) values ranging from −10.86 to −8.64, with corresponding two-stage depleted-mantle Nd model ages (T_2DM) ranging from 1.49 to 1.68 Ga.

Generally, granites are divided into I-, S-, and A-type granites in the literature (Chappell and White, 1974; Loiselle and Wones, 1979). The most important feature of S-type granites is that they are always peraluminous with high A/CNK (>1.1) and generally accompanied by Al-rich minerals such as cordierite or muscovite. The excess Al is considered to be hosted in Al-rich biotite (A/CNK = 1.3–1.5 in biotite, according to Zen, 1986) and the compositions become less peraluminous with increasing SiO₂ (e.g., Chappell et al., 1987). In contrast, most I-type granites, which are derived from preexisting intracrustal igneous source rocks, are metaluminous (A/CNK < 1.0) and always contain amphibole, although some more felsic I-type granites are weakly peraluminous (Chappell et al., 1987, 2012). A-type granites are characterized by high-temperature and anhydrous minerals (e.g., pyroxene, arfvedsonite, and riebeckite), enriched in REE (except Eu), SiO₂, K₂O, Fe/Mg contents, Zr, Nb, Ga, Y (Zr+Nb+Ce+Y > 350 ppm) and high molar Ga/Al ratios (10000*Ga/Al > 2.6) but low Al₂O₃, CaO, Ba, Sr, and Eu contents (Loiselle and Wones, 1979; Collins, 1982; Chappell et al., 1987). With the exception of two samples plotted in the field of fractionated granites or A-type granites, the majority of Xishanjiao samples plotted in the field of unfractionated granites in the FeO^T+MgO, (K₂O+Na₂O)/CaO versus Zr+Nb+Ce+Yb diagrams (Figs. 8A and 8B), indicating that they are probably I- or S-type granites rather than A-type granites. This is further supported by the low contents of REE (114–202 ppm), Nb, Zr (119–194 ppm, <220 ppm), and lack of mafic alkaline minerals (e.g., arfvedsonite and riebeckite). However, the Xishanjiao granites have variable SiO₂ contents (65.8–74.2 wt%), high Na₂O (3.18–4.87 wt%), and A/CNK indexes (0.96–1.04) indicating metaluminous to weakly peraluminous compositions, similar to the geochemical characteristics of I-type granites. Furthermore, various trends of trace elements such as Rb, Y, and Th can be used to distinguish I- or S-type granites. In the Rb versus Th and Y diagrams (Figs. 8C and 8D), the plots of the Xishanjiao samples show a distinct I-type trend, which is also evidenced by the decrease of P₂O₅ with the increase of SiO₂ (Fig. 7) and the occurrence of amphibole (Fig. 3D). Therefore, based on the mineralogical and geochemical results, we suggest that the Xishanjiao granites are I-type granites, which is consistent with the results reported previously for the Tengchong terrane (Yang et al., 2006; Cong et al., 2011a, 2011b; Qi et al., 2011, 2019; Luo et al., 2012; Cao et al., 2014; Zhu et al., 2015, 2018; Fang et al., 2018; Zhang et al., 2018). We have summarized data of the Early Cretaceous igneous rocks from the Tengchong terrane and find that they show similar trends and chemical characteristics. For instance, all samples reveal I-type trends in the plot of P₂O₅ versus SiO₂ (Fig. 7) and in the Rb versus Th and Y diagrams (Figs. 8C and 8D), suggesting that Early Cretaceous igneous rocks have I-type affinity with variable degrees of fractional crystallization. In contrast, the anomalous two samples (XSJ-08, XSJ-09) in this study have high contents of SiO₂ (73.6–74.6 wt%), K₂O (4.24–4.89 wt%), A/CNK indexes (1.25), and low contents of CaO (0.18–0.31 wt%) and MgO (0.12%–0.14%), indicating that they may be the result of highly fractionated magma, which often reveal A-type affinity in geochemical characteristics (Wu et al., 2017).

I-type granites are considered to be derived from intra-crustal igneous rocks in the continental crust (Chappell, 1998, 1999; Chappell and White, 1974), and can be used to constrain the nature of magmatic source. Based on experimental studies and natural research, three main recognized models for the generation of I-type granites have been proposed: (1) partial melting of mafic to intermediate metaigneous crustal rocks with or without contributions of mantle-derived materials in the presence of various amounts of water (e.g., Beard and Lofgren, 1989; Roberts and Clemens, 1993; Griffin et al., 2002; Weissman et al., 2013); (2) fractional crystallization of mafic melts, which has similar compositions to the mafic xenoliths, producing intermediate and felsic magmas (e.g., DePaolo, 1981; Chiaradia, 2009; Be’eri-Shlevin et al., 2010; Weissman et al., 2013; Lee and Bachmann, 2014; He et al., 2019b); and (3) mixing of crust- and mantle-derived melts in different proportions to produce various granitoid magmas (e.g., Roberts and Clemens, 1993; Droop et al., 2003; Kemp et al., 2007; Zhu et al., 2009; Weissman et al., 2013; Zhao et al., 2019). The Xishanjiao granites are enriched in LILEs (Rb, Th, U) and depleted in HFSEs (Ba, Nb, P, and Ti), suggesting the importance of crustal rocks in their magma sources (Roberts and Clemens, 1993). Except for the two anomalous samples, the Xishanjiao granites show a wide range of Mg^# values (30.5–41.9) and a metaigneous source in the plot of Mg^# versus SiO₂ (Fig. 9A). Meanwhile, they reveal the result of partial melting of amphibolites in the plot of (Na₂O+K₂O)/(FeO^T+MgO+TiO²) versus Na₂O+K₂O+FeO^T+MgO+TiO₂ (Fig. 9B). The relatively low Al₂O₃/TiO₂ ratios (25.3–55.9) suggest that metapelite is also involved in their source (Fig. 9C; Sylvester, 1998). Nd isotopic compositions can also be used to provide constraints on their source components (Champion and Bultitude, 2013; He et al., 2019a). The negative ε_Nd(t) values (−8.64 to −10.86) with T_2DM ranging from 1.49 to 1.68 Ga are similar to previously published Early Cretaceous granites in the Tengchong terrane, with ε_Nd(t) values ranging from −13.26 to 1.51 (see Fig. 9D; e.g., Yang et al., 2006; Zhu et al., 2015, 2018; Zhang et al., 2018), implying a predominantly Mesoproterozoic crustal source (metaigneous or amphibolite) with little involvement of mantle-derived materials. Therefore, we suggest that the Xishanjiao granites are from partial melting of metaigneous crustal rocks with minor contributions of mantle-derived materials.

Compared to Nb/Ta ratios (13.4, according to Rudnick and Gao, 2003) in continental crust, the significantly low Nb/Ta ratios (5.28–9.77) of the Xishanjiao granites suggest they experienced appreciable fractional crystallization of apatite, which is supported by the observed negative correlations of P₂O₅ and TiO₂ versus SiO₂ (Fig. 7), as well as the depletion of P and Ti in the primitive-mantle normalized trace elemental spider diagram (Fig. 6B). The negative anomaly δEu (0.32–0.66) in chondrite-normalized REE patterns (Fig. 6A) indicate the fractionation of plagioclase, which is evidenced by the increasing of K₂O with the increase of SiO₂ whereas Na₂O slightly decreases (see Fig. 7). Because the removal of non-K minerals (e.g., plagioclase) can cause the increase of K₂O content in melt. The flat HREE pattern (Fig. 6A) and relatively high Y content indicate that the source melting occurred at pressures below the garnet stability field (Rapp and Watson, 1995) and that fractional crystallization of amphibole has taken place (Jaques and Green, 1980). Therefore, we suggest that the Xishanjiao granites have experienced extensive fractional crystallization of plagioclase, amphibole, and apatite.

According to our summary in Figures 7–9, the Early Cretaceous igneous rocks in the Tengchong terrane show similar geochemical characteristics and variations with the Xishanjiao granites in this study, implying that the Early Cretaceous magmatism probably originated from a similar source and occurred in a similar geodynamic setting to the Xishanjiao granites in this study.

Previous comparisons of the Tengchong terrane with the Lhasa terrane suggested similarities in magmatic activities, stratigraphy, and paleobiogeography. Subsequently, it has been proposed that the Tengchong terrane is the southern extension of the Lhasa terrane, both of which experienced similar tectono-magmatic histories since the Early Paleozoic (Xu et al., 2008, 2012; Xie et al., 2016; Qi et al., 2019). Accordingly, tectonic models of Bangong-Nujiang suture zone in Tibet have been adopted to interpret the Early Cretaceous magmatism in the Tengchong terrane. However, several different models have been proposed, being: (1) low-angle or flat northward subduction of Neo-Tethys oceanic lithosphere (Ding, 2003; Kapp et al., 2005; Liu et al., 2017); (2) southward subduction and slab break-off of the Bangong-Nujiang oceanic slab (Zhu et al., 2015, 2017a, 2017b; Fang et al., 2018; Zhang et al., 2018; Qi et al., 2019); or (3) post-collisional tectonic setting following the closure of the Meso-Tethyan Ocean (Xu et al., 2012; Cao et al., 2014). Based on the data presented in this study, combined with previously published data from the region, we proposed a completely different view on the tectonic setting and provide constraints on the dynamic setting and tectonic evolution of the Tengchong terrane.

As described above, the Xishanjiao plutons are metaluminous to weakly peraluminous, high-K, calc-alkaline I-type granites (see Figs. 5A and 5C). Two tectonic settings have been recognized to interpret the formation of high-K, calc-alkaline, I-type granites: (1) subduction-related continental arc setting such as Andean type, where a large number of igneous rocks generated by hybrid magma derived from the mantle and crust occur, or (2) post-collision extensional setting caused by decompression following crustal thickening (Roberts and Clemens, 1993). Here we propose that the following observations provide an important constraint on the geodynamic setting for the Early Cretaceous magmatism in the Tengchong terrane. (1) The absence of Jurassic and Cretaceous strata in the whole Tengchong terrane (YNBGMR, 1990) likely indicates that regional uplift and erosion had been in progress during this time, which is also supported by the occurrence of purplish-red sandstone produced in the continental facies environment (Zhang et al., 2018). (2) Some depleted mantle-derived diabased and mafic enclaves with emplacement ages at ca. 122–115 Ma have been identified recently in both the Tengchong terrane (Cong et al., 2011a; Zhu et al., 2017b; Fang et al., 2018; Zhang et al., 2018; Qi et al., 2019) and the Mogok metamorphic belt of Myanmar (Chen et al., 2016) indicating that the onset of the extension-related mantle-derived magmatism occurred during the Early Cretaceous. (3) The relatively high crystal temperatures (most zircon saturation temperature T_Zr = 805–844 °C; see Fig. 10A) of granitoids also strongly support a mantle contribution to the generation of the Early Cretaceous magmatism, which is also supported by the more variable εHf ranging from -18.88 to 10.26 (see Fig. 10B). Collectively, all the above features point to the fact that the Meso-Tethys Ocean basin had been closed at least in the Early Cretaceous in the Tengchong region and a post-collisional extensional regime related to the upwelling and subsequent partial melting of the depleted asthenospheric mantle had occurred. Hence, the most feasible mechanism responsible for such scenario can be ascribed to the post-collisional lithospheric extension. Recent identification of Jurassic Myitkyina ophiolite (173 Ma; see Fig. 1A) to the west of Tengchong terrane suggests that the eastern belt in Burma represents the southern continuation of the Bangong-Nujiang suture rather than Gaoligong shear zone (Liu et al., 2016a). Accordingly, Early Cretaceous granitoids in the east Tengchong terrane is closely related to the evolution of Meso-Tethys rather than the Gaoligong shear zone between the Tengchong and Baoshan terranes. Therefore, the Tengchong terrane is probably the southern extension of the South Qiangtang terrane rather than the Lhasa terrane. Moreover, we prefer the model of rollback of Meso-Tethys slab (see Fig. 11) to interpret the generation of Early Cretaceous granitoids produced in a post-collisional lithospheric extension. In this model, a flat or low-angle subduction of the Bangong Meso-Tethys oceanic slab—like Andes-type in the South America—has occurred, which has been supported by the previous ages of magmatic rocks from the Tengchong terrane showing a distinct magmatic gap at 142–130 Ma (Zhang et al., 2017). Because of long-term high temperature and pressure beneath the Tengchong terrane, the subduction Bangong Meso-Tethys slab converted to anomaly heavy eclogite phase and roll-back happened and triggered subduction angle steeper. When the slab rolled back, hot asthenospheric material upwelled and enhanced decompression melting in the overriding plate, forming various granitoids with involvements of mantle materials in extensional environments. This process is consistent with the formation of magmatic rocks in central Tibet (Zhang et al., 2017).

Geochronological and geochemical data for the Early Cretaceous granitoids in the eastern Tengchong terrane, coupled with previous publications, allow us to reach the following conclusions:

• (1) The Xishanjiao pluton was emplaced at ca. 124 Ma in the Early Cretaceous period.

• (2) The Xishanjiao granites, similar to most of Early Cretaceous granitic rocks in the Tengchong terrane, are dominated by I-type granites mainly derived from partial melting of the Mesoproterozoic metaigneous rocks with minor sediments in the lower crust.

• (3) The Early Cretaceous magmatism in the Tengchong terrane was related to post-collisional slab rollback and represented the product of the evolution of the Myitkyina Meso-Tethys Ocean in eastern Myanmar (Burma) rather than the Bangong-Nujiang Ocean or Gaoligongshan shear zone between the Tengchong and Baoshan terranes.

This study was supported by National Natural Science Foundation of China (grant numbers 41702084, 41872089, and 41903032), Yunnan Department of Science and Technology application of basic research project (grant number 2017FD063), and Geology Discipline Construction Project of Yunnan University (grant number C176210227).