All the studied zircons are euhedral, elongate, inclusion–free, and transparent. LA–ICP–MS zircon U–Pb data are given in Table 2, and representative zircon CL images and analyzed spots are shown in Figure 3. The zircons have euhedral crystal shapes and exhibit oscillatory or broad zoning. The zircons from the quartz monzonite porphyry (sample 17MD22–1) and MMEs (18NM42–4) all have high concentrations of Th and U and high Th/U ratios (0.38–0.80 and 0.49–0.85, respectively; Table 2), indicating their magmatic origin (Corfu et al., 2003).

Twenty-three analyses from sample 17MD22–1 yield concordant 206Pb/²³⁸U ages of 211–219 Ma (Table 2) and a weighted mean age of 215 ± 1 Ma (MSWD = 0.34; Figure 3A ), which is identical to their Tera-Wasserburg U–Pb lower intercept age of 215 ± 1 Ma (MSWD = 0.85), were within the acceptable error. Thus, the 215 ± 1 Ma is interpreted as the crystallization age of the Mogetong quartz monzonite porphyry. Fifteen analyses from sample 18NM42–4 exhibit 206Pb/²³⁸U show ages ranging from 208 Ma to 223 Ma and yield a concordant age of 212 ± 2 Ma with a weighted mean age of 209 ± 2 Ma (MSWD = 0.67; Figure 3B and Table 2). These two ages were consistent with the acceptable error, so the concordant age of 212 ± 2 Ma represented the crystallization age of MMEs. Thus, the above results indicate the zircons of MMEs have nearly identical crystallization age to that of the zircons in host rocks.

Geochemical analyses of representative samples are listed in Table 3. The Mongetong quartz monzonite porphyries have a narrow range of SiO₂ contents (63.31–65.74 wt%), moderate contents of TiO₂ (0.47–0.68 wt%), Al₂O₃ (15.19–16.02 wt%), FeO^T (2.54–3.83 wt%), MgO (1.33–2.67 wt%), and high contents of total alkaline (Na₂O + K₂O = 7.03–8.19 wt%). The quartz monzonite porphyries define a sub–alkaline trend in the total alkalis–silica (TAS) diagram (Figure 4A) and a high-K calc–alkaline trend with high K₂O contents (2.83–3.34 wt%; Table 3 and Figure 4B). These samples are metaluminous with A/CNK values of 0.94–1.02 (Table 3), and it is worth noting that the Mongetong quartz monzonite porphyries exhibit high Mg^# values (47–55). The MMEs show mafic-intermediate composition with SiO₂ of 49.73–56.94 wt%, TiO₂ of 1.14–1.20 wt%, FeO^T of 6.00–7.93 wt%, MgO of 3.88–3.91 wt%, and exhibit much higher Na₂O/K₂O values (2.18–3.23) and similar Mg^# (47–54).

The Mongetong quartz monzonite porphyries are characterized by enrichment in light rare earth elements (LREEs) and depletion in heavy rare earth elements (HREEs) [(La/Yb)_N = 14.73–32.87; Figure 5A]. The host rocks and MMEs all display high total REEs (RREE = 148–222 ppm) and slightly negative Eu anomalies (Eu/Eu* = 0.8–0.9). To sum up, the host rocks and MMEs exhibit similar trace elements patterns, exhibiting overall enrichment in large ion lithophile elements (LILE; e.g., Th, U, K) and LREE, and depleted in high field strength elements (HFSE; e.g., Nb, Ta, Ti and p; Figure 5B). Importantly, the quartz monzonite porphyries have high contents of Sr (462–729 ppm) and Ba (752–1,019 ppm), but low contents of Y (9.14–15.7 ppm) and Yb (0.73–1.39 ppm) and low values of Sr/Y (30–57) and La/Yb (27–49), resembling typical adakites or adakitic rocks.

Zircon Lu–Hf isotopic data are shown in Table 4. Fifteen spots were analyzed for the sample 17MD22–1 and all the analyses have similar initial 176Hf/¹⁷⁷Hf values of between 0.282690 and 0.282753. The zircons exhibit weakly juvenile Hf isotopes (εHf(t) = 1.80–4.03), with Hf isotopic model ages ranging from 1.00 Ga to 1.14 Ga (Table 4; Figure 6).

Several Triassic adakitic rocks have been reported in the East Kunlun orogenic belt, and most of them exhibit crystallization age of Middle–Late Triassic (Yuan et al., 2009; Xiong et al., 2014; Chen et al., 2017; Zhang et al., 2017; Liang et al., 2021; Zhong et al., 2021), thus providing a window to constrain the magma-tectonic evolution of East Kunlun Paleo-Tethyan orogen. The studied Mongetong quartz monzonite porphyries have adakitic affinities, e.g., intermediate SiO₂, high Sr, low Y and Yb contents, and high Sr/Y and La/Yb ratios (Table 3; Figure 7). Their high SiO₂ contents and moderate (CaO + Na₂O) values show that the studied pluton is a high–silicon adakite like the other Triassic adakitic rocks in East Kunlun (Table 3; Martin et al., 2005; Castillo, 2012).

Adakites or adakitic rocks were originally recognized in Cenozoic Aleutian island arcs where their generation is associated with the subduction of relatively young (≤25 Ma) oceanic slabs (Defant and Drummond, 1990). Slab melting of subducted oceanic slabs has been proposed to account for the genesis of adakitic rocks (Defant and Drummond, 1990; Drummond et al., 1996; Ma et al., 2014; Zheng et al., 2021). However, several other mechanisms have also been proposed to account for the origin of adakitic rocks, such as partial melting of the thickened basaltic lower crust (Atherton and Petford, 1993; Chung et al., 2003; Wang et al., 2007a; Yu et al., 2018, Yu et al., 2019), partial melting of mantle peridotite metasomatized by slab melt under water–bearing condition (Martin et al., 2005; Castillo, 2008), magma mixing between felsic and basaltic magmas (Qin et al., 2010; Foley et al., 2013), low–pressure fractional crystallization of parental basaltic magmas (Castillo, 2012; Chen et al., 2016), and high–pressure fractional crystallization (involving garnet) of mafic magmas (Martin et al., 2005; Macpherson et al., 2006; Chiaradia, 2009). Adakitic magmas derived from melting of subducted oceanic slabs usually have the following geochemical characteristics: depleted or relative juvenile radiogenic isotopes with young model ages, and sodic series with relatively low K₂O content and K₂O/Na₂O values (Defant and Drummond, 1990; Martin et al., 2005; Wang et al., 2007b). The Mogetong adakitic quartz monzonite porphyries have old zircon Hf isotopic model ages (1.00–1.14 Ga; Figure 6) and they exhibit high K₂O/Na₂O values (0.57–0.81) with high-K calc-alkaline affinities (Figure 4, 8A), which is identical to the felsic rocks derived from partial melting of the ancient lower continental crust in East Kunlun (Xiong et al., 2014; Ding et al., 2015; Li et al., 2021). Moreover, the studied adakitic rocks have weak juvenile zircon Hf isotopes (εHf(t) = 1.80–4.03), which is much lower than the East Kunlun Paleo-Tethyan MORB (εHf(t) = 15–20; Huang et al., 2014; Hu et al., 2016), and it is unlikely that the Mogetong adakitic quartz monzonite porphyries were derived from basaltic slab melting. Besides, phase equilibrium modeling and experimental studies reveal that melting of either pristine or altered oceanic basaltic slab usually leads to the generation of high-SiO₂ and strongly sodic adakitic melts (Hernández-Uribe et al., 2019; Li et al., 2022), and the melts are usually rich in SiO₂ (>68 wt%) and Al₂O₃ (>18 wt%), but poor in FeO and MgO (<0.2 wt%), which is not the case of the studied Mogetong adakitic rocks (e.g., potassium, low contents of Al₂O₃ (15.19–16.02 wt%) and moderate contents of SiO₂ (63.31–65.74 wt%) and MgO (1.33–2.67 wt%); Figures 4, 8 and Table 3). Besides, the partial melts from fresh or hydrothermally altered MORB usually have (La/Yb)_N >194 and (La/Yb)_N >15, respectively, and those from altered MORB are characterized by strong negative Th anomalies (Hernández-Uribe et al., 2019), which are not the case of the studied Mogetong adakitic rocks (e.g., (La/Yb)_N = 18–33; enrichment in Th; Figure 5B and Table 3).

Besides, adakitic magmas derived from partial melting of thickened basaltic lower crust without crust-mantle interaction would have relatively low MgO contents and Mg^# values, and these melts are usually characterized by low Mg^# values (<40) regardless of the melting degree (Atherton and Petford, 1993; Castillo, 2012). The Mogetong adakitic quartz monzonite porphyries have high Mg^# values (Mg^# = 47–55; Figures 8B–D), which is distinct from the thickened basaltic lower crust–derived adakitic rocks. Conversely, adakitic rocks formed by high-pressure partial melting of continental crust would have relatively high (Gd/Yb)_N ratios (>5.8) (Huang and He, 2010), however, this is not the case of the Mogetong adakitic rocks ((Gd/Yb)_N = 2.14–3.09). The Mogetong adakitic quartz monzonite porphyries show partial melting trends rather than fractional crystallization trends (Figure 9A), indicating fractional crystallization of parental basaltic magmas may not be the most critical factor accounting for the genesis of the studied rocks. Typical adakitic rocks derived by fractional crystallization of basaltic magmas usually have low SiO₂ content (Martin et al., 2005; Chiaradia, 2009; Castillo, 2012), which is different from the Mogetong adakitic rocks as well as other Triassic adakitic rocks in East Kunlun (Figures 8B,D). In addition, given that high–pressure fractional crystallization involving garnet usually causes a decrease in HREEs and Y contents, the Sr/Y and Dy/Yb ratios in the residual magmas should increase with increasing SiO₂ and MgO content (Macpherson et al., 2006; Laurent et al., 2013). However, the Mogetong adakitic quartz monzonite porphyries do not show such a trend in the Sr/Y versus MgO and Dy/Yb versus SiO₂ diagrams (Figures 9B,C). During low–pressure fractional crystallization involving hornblende and plagioclase, the net effect of hornblende and plagioclase fractionation is a decrease in δEu with increasing SiO₂ content, which is observed in most typical arc adakites (Martin et al., 2005; Moyen, 2009), but is not present in the Mogetong adakitic rocks (Figure 9D).

Here, we propose that magma mixing between felsic and basaltic magmas is the most possible mechanism accounting for the generation of the studied Mogetong quartz monzonite porphyries. Their moderate SiO₂ contents (63.31–65.74 wt%) but high Mg^# values (47–55), as well as weak juvenile Hf isotopes, support the model of crust-mantle interaction (Figures 6, 8D). The studied rocks exhibit much more depleted zircon Hf isotope than those derived from the ancient continent crust (Figure 6), indicating the significant contribution of mantle-derived magma. Furthermore, the high field strength element ratios of these samples also support the crust–mantle mixing model, such as their Nb/Ta values (average 14.87) ranging between the average continental crust (Nb/Ta = 11; Sun and McDonough, 1989) and the average primary mantle (Nb/Ta = 17.7; Taylor and McLennan, 1985). The correlation trend between La and La/Yb also confirms the major role of magma mixing (Figure 9A). Besides, the presence of the MMEs further indicate the occurring of magma mixing, although several petrogenetic models have been proposed for the origin of MMEs (Qin et al., 2010; Xiong et al., 2012, 2014), including 1) restites, 2) cognate fragments of cumulates, 3) quenched mafic melts derived from the mantle, and 4) mixing of mafic and felsic magmas. The MMEs in the Late Triassic Mogetong pluton have granitic textures with euhedral plagioclase, biotite and amphibole, subhedral K-feldspar, and anhedral quartz (Figure 2F), which does not support the cumulates and quenched mafic melts model but indicates that the MMEs crystallized from a melt. The MMEs have mafic-intermediate compositions with slightly silica contents of 49.73–56.94 wt%, which also contradicts the idea of the MMEs being cognate fragments of mafic cumulates and quenched mafic melts. Moreover, the euhedral magmatic zircons with clear igneous oscillatory and identical U-Pb ages to their host rocks also preclude the restite model, but support the magma mixing model (Figure 3). These features, as well as the fact of different trace elemental patterns between the MMEs and their host rocks, support the magma mixing model for the origin of the studied Mogetong pluton and MMEs, and the contemporary most mafic enclave (sample 18NM42–9) could represent the basaltic end-member. The Mogetong adakitic rocks have variable and positive values of εHf(t) (1.80–4.03) with moderate Mg^# (47–54), which also indicate a mantle component in their source. However, petrological and geochemical evidence indicates that the contribution of mantle material is not significant. For example, the MMEs are sporadically developed in the Mogetong adakitic pluton and the host rocks have low contents of Cr (14–59 ppm) and Ni (8–30 ppm). More importantly, the Hf isotopes of the studied rocks are much lower than those of the underplating primitive basaltic magma (εHf(t) = 8–15; Hu et al., 2018) or the Paleo-Tethyan MORB in the East Kunlun (εHf(t) = 15–20; Huang et al., 2014; Hu et al., 2016). The continental crust-like Nb/Ta values of the Mogetong adakitic rocks (14–16) are much lower than the MORB-like Nb/Ta values of the MMEs (17–20), which also support this suggestion (Taylor and McLennan, 1985; Sun and McDonough, 1989).

The tectonic evolution of East Kunlun Paleo-Tethyan orogen during the late Paleozoic to early Mesozoic is still controversial and the key problem is the timing of the tectonic transition from oceanic slab subduction to continental collision (Li et al., 2013a, Li et al., 2018; Liu et al., 2017; Dong et al., 2018; Xiong et al., 2019; Yu et al., 2020; Zhang et al., 2021). Since there is an obvious genetic link between the magma source and tectonic setting of adakitic rocks, the detailed petrogenesis study could contribute to our understanding of the evolution of Triassic East Kunlun Paleo–Tethyan orogen. The above discussion on the petrogenesis shows that the Mogetong adakitic magma was originated from crust-mantle mixing, indicating that there was a significant contribution of mantle material in the Late Triassic. This is further supported by the zircon Hf isotopes of Triassic adakites (Figure 10). A comprehensive comparison of the Hf isotopic composition of Triassic adakites in East Kunlun orogen shows that there is a significant juvenile trend in the Late Triassic (Figure 10). Besides, the published data shows that there were two episodic adakitic magmatisms in the East Kunlun orogen, i.e. the Middle Triassic and Late Triassic, however, the latter is the main metallogeny stage for the porphyry–and/or skarn–type deposits (Yuan et al., 2009; Xiong et al., 2014; Chen et al., 2017; Zhang et al., 2017; Liang et al., 2021; Zhong et al., 2021). The above characteristics all indicate that the East Kunlun was in a special tectonic setting during the Late Triassic.

Through an integrated geochemical comparison among the Triassic adakites in East Kunlun orogen, we propose a revised tectonic evolution model for the East Kunlun Paleo-Tethyan orogen. As mentioned above, the crust-mantle magma mixing is the main genetic mechanism for the Late Triassic adakites in the East Kunlun, thus, revealing the properties of the late Triassic mantle-derived magma is the key to understanding the tectonic setting of Late Triassic adakitic magmatism. Previous studies have shown that mid-ocean ridge subduction, slab break–off, and slab roll–back and lithospheric delamination could trigger the underplating of mantle-derived magma and subsequent crustal melting and magma mixing (Martin et al., 2005; Wang et al., 2007b; Tang et al., 2010; Keller and Schoene, 2018). At present, however, Triassic mafic magmatic rocks with N–MORB affinities, back-arc basalts, back-arc basin, and the temporal–spatial migration trend of Triassic magmatism have not been found or reported in the East Kunlun orogen, precluding the model of mid-ocean ridge subduction and slab rollback (Hu et al., 2016; Xiong et al., 2019). Besides, lithospheric delamination-related magma will interact with mantle peridotite, forming adakitic magmas with high contents of Ni and Cr (Huang et al., 2008; Qin et al., 2010), but the Mogetong adakitic quartz monzonite porphyries do not exhibit high Cr and Ni contents.

Meanwhile, recent studies have shown that whole-rock Sr/Y and La/Yb ratios of felsic rocks can effectively reveal continent crustal thickness (Chapman et al., 2015). To eliminate mafic rocks generated in the mantle and to avoid the influence of magma mixing and/or highly fractional crystallization on the calculation results, the data set was filtered to include rocks with moderate SiO₂ content (55–68 wt%) and MgO content (one to four wt%). Data with Rb/Sr > 0.2 or Rb/Sr < 0.05 were further discarded, providing a trace elemental filter for mantle-derived rocks or rocks formed by melting of pre-existing metasedimentary framework rocks (Chapman et al., 2015). Using these empirical correlations, we investigate temporal variations of crustal thickness in the East Kunlun orogen, and our results suggest that crustal thickness increased during the Early-Middle Triassic and remained constant during the Late Triassic (Figure 11). The calculated results also indicate that the continent crustal thinning of East Kunlun did not occur during the Late Triassic, precluding the model of lithospheric delamination in the Late Triassic. Here, we propose that slab break-off may occur in the Late Triassic, which induced the underplating of mantle-derive mafic magma and promoted the melting of thickened crust and subsequent magma mixing (Figure 12). The Late Triassic OIB-like mafic igneous rocks and mafic dyke swarms in East Kunlun may be the response of slab break-off (Hu et al., 2016; Xiong et al., 2019). The slab break–off model also matches the linear distribution of Late Triassic adakites along the Eastern Kunlun Paleo-Tethyan suture zone (Dong et al., 2018). Besides, mantle-derived mafic magma underplating in the break–off model not only provides heat, but also provides ore-forming metals, which may be the key to the Late Triassic mineralization flare-up in the Eastern Kunlun orogen (Zhang et al., 2017; Zhong et al., 2021). Several pieces of evidence support the idea that late Triassic Mogetong adakites were formed in a post–subduction extensional setting, such as mantle–derived high Nb–Ta rhyolite (213 Ma; Ding et al., 2011), crustal–derived shoshonitic volcanic rocks (228–215 Ma; Hu et al., 2016), and coeval mafic dyke (226–210 Ma; Xiong et al., 2011, 2019; Hu et al., 2016; Liu et al., 2017). The angular unconformity between Late Triassic land facies volcanic-sedimentary rocks, e.g., Babaoshan Formation red clastic sedimentary rocks and E’lashan Formation volcanic rocks, and underlying sea-land facies strata is the records of tectonic transition from oceanic subduction to post-subduction collision or extension (Li et al., 2013a).

This study also suggests that the reworking of the ancient lithosphere and subsequent magma mixing are the important mechanisms for the generation of granitoids and continental crust growth. It is previously well accepted that the growth and evolution of the continental crust in orogen is mostly dominated by felsic magmatism, such as the Central Asian Orogenic Belt (Li et al., 2013b; Kröner et al., 2014), and the late Paleozoic to early Mesozoic felsic igneous rocks in the Central Asian Orogenic Belt are characterized by much depleted Sr–Nd–Hf isotopic compositions, indicating juvenile materials played a key role in crustal growth in Central Asia (Jahn et al., 2000; Li et al., 2013b; Xiao and Santosh, 2014). However, our study shows that the isotopic compositions of the Triassic adakites in East Kunlun are different from those of the Central Asian Orogenic Belt but identical to those derived from the melting of Precambrian continental crust with a contribution of mantle-derived magma, implying a different crustal evolution mechanism in East Kunlun (Xiong et al., 2012, 2014; Yu et al., 2017; Li et al., 2018). We propose that the reworking of a dominantly ancient continental crust with certain addition of lithospheric mantle materials is another major mechanism for the generation of felsic igneous rocks and the evolution of the continental crust.