1 Introduction

The Qinghai-Tibet Plateau has played a significant role in recent Earth evolution and climate change, but the complicated geological processes of the Qinghai-Tibet Plateau are still hotly debated in recent years (Ma et al. 2022; Zhang et al. 2017). There is a consensus that the origin of the Qinghai-Tibet Plateau can be attributed to the collision between India and Asia during the Late Cenozoic period (Haider et al. 2013; Hetzel et al. 2011; Liu et al. 2014; Zhang et al. 2004), following Mesozoic continental accretion, subduction, and crustal thickening (Decelles et al. 2007; Murphy et al. 1997).

The Lhasa terrane (LT) represents the final continental fragment integrated into the plateau during the Late Mesozoic era (Chen et al. 2015; Yin and Harrison 2000; Zhu et al. 2011) and has gone through noteworthy S–N crustal thickening and shortening during the Late Cretaceous before the collision between India and Asia (Decelles et al. 2007; Kapp et al. 2007; Murphy et al. 1997; Pan et al. 2014; Yin and Harrison 2000). Responding to tectonic processes, Late Cretaceous magmatism occurred in the LT and has been thought to be related to either the subduction of the Neo-Tethyan Ocean lithosphere or to crustal thickening (Chen et al. 2015; Ma and Yue 2010; Pan et al. 2014). Recently, the Late Cretaceous (~ 80 Ma) felsic magmatism reported in the LT is thought to be related to the subducted Neo-Tethys oceanic slab (Chapman et al. 2018; Wang et al. 2019, 2021). Some workers have argued that the Late Cretaceous (~ 90 Ma) adakitic, Mg-rich, and K-rich rocks in the LT resulted from the lithospheric delamination underneath the thickened crust after the collision of Lhasa-Qiangtang Blocks (Chen et al. 2015; Ji et al. 2021; Li et al. 2013; Liu et al. 2017; Sun et al. 2015a, b, 2020; Wang et al. 2014; Yi et al. 2018; Zhao et al. 2008), but this explanation has been challenged based on evidence from zircon xenocrysts in Lhasa ultrapotassic magmas (Liu et al. 2014). It is noteworthy that the majority of Late Cretaceous magmatism in central Lhasa terrane (CLT), northern Lhasa terrane (NLT), and southern Qiangtang terrane (SQT) has been attributed to the reworking of juvenile crust through partial melting of pre-existing underplated lower crust (Wang et al. 2021; Tang et al. 2019). Moreover, the role of mantle-derived magmas and whether crustal reworking or growth occurred in the SE Lhasa during the Late Cretaceous period (ca. 100–80 Ma) are poorly known. Therefore, Late Cretaceous potassium-rich high Ba–Sr granodiorites may offer valuable insights into magmatic evolution associated with mantle-derived materials. In this paper, our studies provide bulk-rock chemical, isotopic and in situ zircon Hf isotopic and U–Pb age data to constrain: (1) the petrogenesis of high Ba–Sr granodiorites; and (2) the possible geodynamic processes at depth after final Lhasa-Qiangtang amalgamation in the Qinghai-Tibet Plateau.

2 Geological setting and petrography

The Qinghai-Tibet Plateau constitutes a significant component of the Tethyan-related orogenic collage of the predominantly EW trending Himalaya, Qiangtang, Lhasa, Songpan-Ganzi, and Kunlun-Qiadam terranes. These terranes are demarcated by the Yarlung-Zangbo (YZ), Bangong-Nujiang (BN), Kunlun, and Jinsha suture zone (Fig. 1a) (Yin and Harrison 2000).

Fig. 1
figure 1

a Simplified tectonic map of China. b Map of southern Qinghai-Tibet Plateau showing sample locations and major magmatic features (from Chiu et al. 2009). c Regional geology of Yonglaga (from Wang et al. 2021). The geochronology data are summarised in Table 5

Full size image

The LT lies between the BN suture to the northern and the YZ suture to the southern (Fig. 1a) (Chiu et al. 2009). Its detachment from Gondwana occurred during the Permian to Triassic period, succeeded by a northward movement and eventual convergence with the Qiangtang terrane (QT) during the Late Jurassic period to Early Cretaceous period (Kapp et al. 2005, 2007; Zhu et al. 2011). The LT has been shown to contain three distinct magmatic belts (Fig. 1b) (Chiu et al. 2009): extensive Linzizong volcanic successions and Gangdese batholiths, spanning from the Cretaceous to the Early Tertiary, are widely distributed in the southern Lhasa terrane (Harris et al. 1990; Ji et al. 2009a; Lee et al. 2009; Mo et al. 2007; Wang et al. 2022a, b; Wen et al. 2008a, b), and are considered to be the result of northward subduction of Neo-Tethyan oceanic lithosphere. There are also plentiful Mesozoic igneous rocks in the sub-terranes of NLT and CLT (Chiu et al. 2009; Chu et al. 2006; Guynn et al. 2006; Harris et al. 1990; Xu et al. 1985; Zhu et al. 2009a), the former associated with closure of the Bangong-Nujiang Tethyan ocean has been associated with the northern magmatic belt (Chen et al. 2014; Chiu et al. 2009; Hu et al. 2022; Qu et al. 2012; Sui et al. 2013; Zhu et al. 2009a, b, 2011, 2013, 2015). About the development of the Meso-Tethys, it is commonly believed that the Meso-Tethys oceanic plate underwent northward subduction beneath the QT margin located in the south (Li et al. 2013, 2016; Liu et al. 2017; Ji et al. 2021; Zhang et al. 2017). The collision between the QT and the LT during the Late Cretaceous followed the closure of Meso-Tethys and resulted in the thickening of the southern margin of the QT and the northern margin of the LT. The subsequent delamination of thickened lithosphere triggered a series of post-collisional magmatic events (Chen et al. 2015; Li et al. 2013; Ma et al. 2010; Meng et al. 2014; Qu et al. 2006; Sun et al. 2015a, b, 2020; Wang et al. 2014; Yi et al. 2018; Yu et al. 2011; Liu et al. 2017).

In SE Lhasa, a wide range of Cretaceous granitoids crop out in a NW–SE belt SW trending of the BN suture zone (Chiu et al. 2009; Pan et al. 2004; Zhu et al. 2009b), enclosed by the metamorphic rocks of Proterozoic, Devonian, and Carboniferous-Permian periods. These granitoids are predominantly found as batholiths, primarily distributed in the Bomi and Chayu regions (commonly referred to as Bomi-Chayu batholiths) (Chiu et al. 2009; Lin et al. 2013). They mainly comprise monzogranites and granodiorites, with minor occurrences of mafic enclaves and dioritic veins (Pan et al. 2004; Zhu et al. 2009b). This study focuses on the Yonglaga granitoid (Fig. 1c), located in the east of Gongshan County. It has a rounded, elongated shape aligned with the SE section of the Bomi-Chayu batholiths, positioned between the BN suture and the Jiali fault (Fig. 1b, c) (Chiu et al. 2009; Zhu et al. 2009b).

As is typical in the SE Lhasa Block, the granodiorites are poorly exposed. They are emplaced into Paleozoic (mainly Carboniferous) metasedimentary rocks, are undeformed and their margins are sharp but rarely chilled. High-Mg basaltic diorites are closely associated in the western margin of the granodiorites, also with sharp contacts, and both facies are elongated along the structural fabric of the surrounding metasediments (Fig. 1c). The Yonglaga high Ba–Sr granodiorites are fine-grained and mainly composed of quartz 12%–15%, alkali feldspar 28%–33%, sodic plagioclase 40%–45%, hornblende ~ 5% and biotite ~ 5%, with accessory minerals such as zircon, apatite, titanite, magnetite, and other Ti–Fe oxides (Fig. 2).

Fig. 2
figure 2

Field photographs and photomicrographs of the Yonglaga high Ba–Sr granodiorites in the SE Lhasa Block, China

Full size image

3 Analytical methods

Bulk-rock samples were trimmed to remove weathered surfaces and cleaned with deionized water. They were then crushed and powdered through a 200 mesh screen using a tungsten carbide ball mill. Major elements were analyzed by X-ray fluorescence (XRF) spectrometry (Rikagu RIX 2100) at the Guizhou Tuopu Resource and Environmental Analysis Center, Institute of Geochemistry, Chinese Academy of Sciences in Guiyang, China. Analyses of USGS and Chinese national rock standards (GSP-2, BCR-2, and AGV-2) reveal analytical precision and accuracy for major elements typically better than 5%. Trace elements were determined by inductively coupled plasma mass spectrometry (ICP-MS) at the Guizhou Tuopu Resource and Environmental Analysis Center, Institute of Geochemistry, Chinese Academy of Sciences in Guiyang, China. A Bruker Aurora M90 ICP-MS was used, following the methodology described by Qi et al. (2000). Powders were dissolved in a high-pressure PTFE bomb using an HNO3 + HF mixture at 185 °C for 36 h. The ICP-MS analyses have a relative accuracy and precision of approximately ± 5% to ± 10% for the majority of elements.

Bulk-rock Sr–Nd isotope measurements were conducted at the State Key Laboratory of Continental Dynamics, Northwest University, Xi'an, China, using a Nu Plasma HR multi-collector mass spectrometer, following a methodology similar to that described in Chu et al. (2009). Sr and Nd isotopic fractionation was adjusted to 87Sr/86Sr = 0.1194 and 146Nd/144Nd = 0.7219, respectively. JMC-Nd and NIST SRM-987 were used as certified reference standard solutions for the ratios of 143Nd/144Nd and 87Sr/86Sr, respectively. AGV-2, GSP-2, and BHVO-2 were utilized as the reference materials.

The Yonglaga pluton was the source of two ~ 5 kg samples at different sampling locations, from which zircons were extracted separately. The separation of zircon grains was achieved using conventional techniques involving heavy liquid and magnetic methods. Selected zircon grains were manually chosen and mounted on epoxy resin discs, followed by polishing carbon coating. The internal structure was assessed using cathodoluminescence (CL) before conducting U–Pb analyses. Zircon U–Pb analyses used laser ablation ICP-MS with an Agilent 7500a ICP-MS instrument and a 193-nm laser at the State Key Laboratory of Continental Dynamics, Northwest University in Xi’an, China. The analytical method was based on the approach of Yuan et al. (2004). The GLITTER program was used to determine the ratios of 206Pb/238U and 207Pb/206Pb, which were subsequently adjusted for accuracy using the Harvard zircon 91500 as an external calibration standard. These adjustment factors were subsequently utilized to rectify any potential discrepancies caused by instrumental mass bias and variations in elemental and isotopic distribution at different depths. The method described in Andersen (2002) was applied to assess the common Pb contents. ISOPLOT (version 3.0) (Ludwig 2003) was employed for age calculations and plotting of Concordia diagrams. The errors provided in tables and figures correspond to a confidence level of 2σ.

In situ zircon Hf isotopic analyses used a Neptune MC-ICP-MS with a 193-nm laser. The laser operated at a repetition rate of 10 Hz and an energy level of 100 MJ, while the spot sizes were maintained at 32 μm. The detailed analytical technique follows Yuan et al. (2008). During the analysis, the 176Hf/177Hf and 176Lu/177Hf ratios of the standard zircon (91500) were found to be approximately 0.282294 ± 15 (2σ, n = 20) and 0.00031 respectively, which are similar to the widely accepted values of 0.282302 ± 8 and 0.282306 ± 8 (2σ) obtained using the solution method (Goolaerts et al. 2004). The definitions of εHf(t) value,  fLu/Hf ratio, single-stage model age (TDM1), and two-stage model age (TDM2) are provided with reference to Zheng et al. (2008).

4 Results

Major and trace element results can be found in Table 1, the whole-rock Sr–Nd isotopic results are listed in Table 2, and zircon Hf isotopes are presented in Table 3. The zircon U–Pb ages can be accessed from the Supplementary Dataset Table.

Table 1 The major (wt%) and trace element (ppm) results of rocks from Yonglaga pluton
Full size table
Table 2 The whole-rock Rb–Sr and Sm–Nd isotopic data for rocks from Yonglaga pluton
Full size table
Table 3 Single-grain zircon Hf isotopic data for Yonglaga pluton
Full size table

4.1 Zircon U–Pb age

Relevant zircons are examined in the Gongshan region in SE Lhasa (Fig. 3a). Detailed information on sampling locations, lithology, and dating outcomes can be found in Figs. 1, 2, and 3 as well as the Supplementary Dataset Table.

Fig. 3
figure 3

a CL images of representative zircon grains from the high Ba–Sr Yonglaga granitoids, SE Tibet. b LA-ICP-MS zircon U–Pb concordia diagram of representative zircon grains from the high Ba–Sr Yonglaga granitoids, China

Full size image

Zircon grains from the Yonglaga granodiorite (YLG15-1) are euhedral, prismatic, and mostly have clear oscillatory zones (Fig. 3a), typical of magmatic zircons. The twenty-eight reliable analyses show high Th (435–2382 ppm) and U (538–1361 ppm) with Th/U ratios of 0.61–1.80. The data form a coherent group yielding 206Pb/238U ages from 85.2 ± 0.8 to 89.4 ± 1.0 Ma, with a weighted mean age of 87.32 ± 0.43 Ma (MSWD = 1.7, n = 28, 2σ), taken to indicate the crystallization age (Fig. 3b).

4.2 Major and trace element geochemistry

The granodiorites have comparatively uniform SiO2 (63.17wt%–64.56wt%), with high Na2O (3.85wt%–4.29wt%) and K2O contents (4.10wt%–4.82wt%) (Fig. 4a, b). They plot within the domains of high Ba–Sr granites and high Sr/Y and adakitic rocks in the CLT, NLT and SQT, and show a shoshonitic affinity which is similar to the typical high Ba–Sr rocks (Fig. 4a, b). They also have low TiO2, CaO, Fe2O3T, MnO, P2O5, and MgO with Mg# of 44.2–45.6 (Fig. 4c and 7), but exhibit moderate Al2O3 contents of 16.13–16.59 wt% with A/CNK (molar Al2O3/CaO + Na2O + K2O) ratios of 0.92–0.97 (Fig. 4d). Figure 5a shows positive anomalies for Rb, U, Th, Pb, and Nd, and depletions in P, Ti, Nb, and Ta. Figure 5b shows high total REE contents (297–362 ppm) and fractionated LREE/HREE with high (La/Yb)N (27.7–34.7) and small negative Eu anomalies (δEu = 0.83–0.88). Both the major and trace elements share similar characteristics with the typical high Ba–Sr granites and Late Cretaceous adakites plus high Sr/Y rocks in the CLT, NLT, and SQT (Fig. 5, 6 and 7).

Fig. 4
figure 4

a Na2O + K2O versus SiO2 diagram; b K2O vs SiO2 diagram; c Mg# versus SiO2 systematic diagram (from Moyen and Martin 2012); d A/NK versus A/CNK diagram for the high Ba–Sr Yonglaga granitoids in the SE Lhasa. Experimentally obtained compositions of partial melts by dehydration melting of different source rocks under crustal P–T conditions (0.5–1.5 GPa, 800–1000 °C, Patiño Douce and Beard 1995). The data of Late Cretaceous adakites and high Sr/Y rocks are summarized in Table 5

Full size image
Fig. 5
figure 5

Primitive-mantle-normalized trace element spider diagram and Chondrite-normalized REE patterns (a, b) for the high Ba–Sr Yonglaga granitoids in the SE Lhasa. The primitive mantle and chondrite values are from Sun and McDonough (1989)

Full size image
Fig. 6
figure 6

a Sr–Rb–Ba (from Tarney and Jones 1994); b Rb–Ba–Sr; c La/Nb–La; and d Th/Nb–Th diagrams for the Late Cretaceous high Ba–Sr granodiorites in the SE Lhasa, China. Fields of low Ba–Sr granitoids are based on data data from Fowler and Henney (1996) and Fowler et al. (2001), other fields of high Ba–Sr, adakites and high Sr/Y rocks are based on data as same as in Fig. 4

Full size image
Fig. 7
figure 7

The Harker diagram. The data of Late Cretaceous Yonglaga high-Mg basaltic rocks from Zhu et al. (2024). Symbols as in Fig. 4

Full size image

4.3 Whole-rock Sr–Nd and zircon Hf isotopes

Whole-rock Sr–Nd isotope results are listed in Table 2. All εNd(t) values and initial 87Sr/86Sr isotopic ratios (ISr) are computed for the crystallization age and are very similar. The high Ba–Sr granodiorites (sample YLG15-1, -2, and -3) have ISr ratios ranging from 0.707254 to 0.707322 and εNd(t) values ranging from –2.8 to –3.6, with TDM values of 0.96–1.02 Ga (Fig. 8b).

Fig. 8
figure 8

a Histograms of initial Hf isotope ratios; b The εNd(t) values versus initial 87Sr/86Sr diagram for the high Ba–Sr Yonglaga granitoids in the SE Lhasa. The data of Depleted Mantle and Arc-sources mantle from Fowler et al. (2008). And BNS MORB data from Chen et al. (2014). Basement-derived melts (Zhu et al. 2011), lower continental crust (Miller et al. 1998), and upper continental crust (Harris et al. 1988). c The εNd(t) values versus εHf(t) values diagram. The terrestrial array is after Vervoort et al. (2011) [εHf(t) = 1.55 × εNd(t) + 1.21]. The Nd and Hf contents of depleted MORB mantle (DMM) (Workman and Hart 2005), global subducting sediment (GLOSS) (Plank 2014), lower crust (Rudnick and Gao 2014) and upper crust (Vervoort et al. 2011). The Nd/Hf isotopic compositions of DMM (Li et al. 2016), GLOSS [εNd(t) =  − 8.9, εHf(t) =  + 2 ± 3] (Chauvel et al. 2008), lower crust (Hao et al. 2016) and upper crust (Vervoort et al. 2011)

Full size image

Zircons Lu–Hf data are presented in Table 3 and Fig. 8a, with initial 176Hf/177Hf ratios and εHf(t) values based on crystallization age. The zircons display a range of Hf isotopic compositions, from − 4.55 to + 13.01 (24 zircons, including two negative ones), and Hf model ages of 319–1440 Ma. These contrast markedly from the Nd data described above, signifying a clear decoupling of Hf and Nd isotope systems (Fig. 8c).

5 Discussion

5.1 Petrogenesis of the Yonglaga high Ba–Sr granodiorites

The Yonglaga granodiorites analyzed in this study have high alkali contents (Na2O + K2O = 8.18wt%–8.73wt%, Table 1) and LREEs (282–347 ppm), Sr (653–783 ppm), and Ba (1346–1531 ppm), low Nb, Ta and HREEs with an absence of negative Eu anomalies (δEu = 0.83–0.88) (Fig. 5), all of which suggest distinction from classical A-type, I-type, and S-type granites, but similarity with typical high Ba–Sr granitoids (Fowler and Henney 1996; Fowler et al. 2001, 2008; Jiang et al. 2006, 2012; Peng et al. 2013; Qian et al. 2002; Ye et al. 2008). Although they could be compared with adakites in many geochemical characteristics, such as the high Sr and LREEs, low HREEs, Nb, and Ta plus lack of negative Eu anomalies, there are some significant differences: (1) the notably higher alkali contents (Na2O + K2O > 8.2) and higher K2O/Na2O ratios (0.99–1.25) than adakites (K2O/Na2O < 0.5); (2) a shoshonitic affinity rather than the calc-alkaline trend of adakites (Fig. 4b); (3) the presence of coeval high-Mg basaltic rocks which do not show adakitic signature (Polat and Kerrich 2001; Zhu et al. 2024). Thus, the typical high Ba–Sr granite signature can still be recognized despite clear overlap with adakitic rocks (Fig. 6a, b) (Tarney and Jones 1994; Ye et al. 2008).

5.1.1 The origin of high Ba–Sr granodiorites

Tarney and Jones (1994) initially suggested three genetic models for high Ba–Sr granites: (i) the lower veined lithospheric mantle is enriched by small amounts of carbonatitic melts from the asthenosphere through penetration; (ii) partial melting of the ocean plateaus or subducted ocean islands; (iii) hydrous partial melting of underplated mafic magmas. Other explanations have been proposed, including partial melting of thickened mafic lower crust with participation of minor LILE-rich appinitic magma (Table 4) (Ye et al. 2008); low degree partial melting of the sub-continental lithospheric mantle (Jiang et al. 2006) with subsequently metasomatized by minor crustal materials contamination (Fowler and Henney 1996; Fowler et al. 2001, 2008); continental slab-derived melts (Jiang et al. 2012), and subduction-related melts or fluids (Qian et al. 2002; Peng et al. 2013). Given this ongoing debate, it is important to determine whether the high Ba–Sr granodiorites are derived from the mantle or crust and investigate any interaction in between.

Table 4 Summary of some typical high Ba–Sr rocks in the world
Full size table

The presence of a mafic intrusion near the granitoid pluton area. The mafic xenolith and enclave were also not detected in the granodiorites, suggesting that a crust-mantle magma mixing model lacked geological evidence. Zhu et al. (2024) have reported the geochemical characteristics of the high-Mg basaltic rocks in the Yonglaga area. The age relationships and sharp contacts in the field between the high-Mg basaltic rocks and the granodiorites, and the notable compositional gap all suggest that the granodiorites are unlikely to have formed by simple crystal fractionation of these basaltic magmas (Fig. 7). Geochemically, the Yonglaga granodiorites display a weak correlation between SiO2 and other major elements (Fig. 7), indicating a limited role for fractional crystallization during magma evolution. On the plots of La/Nb vs. La and Th/Nb vs. Th diagram (Fig. 6c, d), the Yonglaga granodiorites show clear linear trends, indicating that a partial melting process that plays a significant role in their formation rather than the fractional crystallization processes. Limited crustal contamination is suggested by the significantly higher Ba (1346–1525 ppm) and Sr (653–783 ppm) than the average Ba (390 ppm) and Sr (325 ppm) of the continental crust (Rudnick and Fountain 1995), plus Rb/Nb ratios that do not show any obvious correlation with SiO2. The highest LOI (1.70 wt%) sample has lower Ba (1346 ppm) and Sr (658 ppm), suggesting a similarly limited role for alteration in defining the characteristic chemistry of the rocks.

The collision zone between LT and QT underwent a gentle collision and slab detachment of a slab during 140–110 Ma (Chen et al. 2014; Qu et al. 2012; Sui et al. 2013; Zhao et al. 2008; Zhu et al. 2009a, b, 2011, 2013, 2015), the non-marine facies of the BN suture have been present since approximately 118 Ma (Kapp et al. 2005, 2007), and the unaltered plutons that intrude ophiolitic mélanges of the BN suture have zircon U–Pb ages ranging from 116 to 112 Ma (Hu et al. 2022), indicating that the BNTO subduction-related activity would have been unlikely during generation and intrusion of the high Ba–Sr granodiorites. Therefore, the Yonglaga high Ba–Sr granodiorites are post-collisional shoshonitic granitoids during the Late Cretaceous, precluding a source related to the partial melting of subducted ocean islands or plateaus.

The Yonglaga granodiorites have low Mg# (44.2–45.3) values, low Cr (7.65–14.6 ppm) and Ni (5.69–8.50 ppm), and elevated K2O/Na2O ratios (> 0.9). These characteristics are inconsistent with a direct mantle origin but the remelting of mafic lower crust (Mg# 52–60, Pan et al. 2014) may produce suitable crust-derived magma. The crustal nature of the granodiorites is further supported by the Fig. 5a, which exhibits depletion in P, Nb, and Ti, as well as enrichment in K and Sr, akin to those observed in average compositions of continental crust (Rudnick and Fountain 1995). Additionally, the Ce/Pb (7.3–8.7), Th/U (5.4–6.8), and Nb/U (2.2–2.7) ratios exhibit similarities to those observed in the lower continental crust (Foley et al. 2002). Experimental findings have indicated that partial melting mafic lower crust could generate magmas with elevated (La/Yb)N and Sr/Y ratios and weakly Eu anomalies at the condition of depth ≥ 40 km and pressure ≥ 1.2 GPa (Fig. 5 and 9) (Petford et al. 1996; Rapp and Watson 1995), with garnet in the residual assemblage (Petford and Gallagher 2001; Rapp et al. 1999, 2002). In summary, the petrogenesis of Yonglaga high Ba–Sr rocks can be attributed to partial melting of the mafic lower crust.

Fig. 9
figure 9

Lu/Hf versus Th/La and Lu/Hf versus Th/Yb diagram (a, b) (from Zhu et al. 2023). Light grey represents oceanic arc magmas with pelagic sediment input and light yellow represents continental arc magmas with terrigenous sediment input. Associated data from GEOROC (http://georoc.mpch-mainz.gwdg.de/georoc). The (La/Yb)N versus (Yb)N and Sr/Y versus Y diagram (c, d). The yellow rectangular symbols represent the high Ba–Sr granites in this study. The light grey rectangular symbols are for high Ba–Sr rocks in different terranes of the world. The light blue symbols are Late Cretaceous adakites in the central and northern Lhasa terranes and the lightly green are other Late Cretaceous high Sr/Y rocks in the central and northern Lhasa and southern Qiangtang terranes

Full size image

5.1.2 The magma source and decoupling of Hf and Nd isotope systems

Previous work demonstrated that the Nd–Hf isotope decoupling is common in the magmatic rocks of the SQT region during both the Early and Late Cretaceous periods (Sun et al. 2021; Wang et al. 2021). Our Sr–Nd–Hf isotopic signature is consistent with the range of granites and adakites in the SQT, CLT, and NLT during the Late Cretaceous, and also exhibits vital Nd–Hf isotope decoupling (Fig. 8). Magma resulting from partial melting of the ancient lower crust exhibits more depleted Sr–Nd isotopes and demonstrates limited decoupling between Nd and Hf isotopes [ΔεHf(t) = 0.7–2.4; Sun et al. 2021]. In contrast, our granodiorites exhibit enriched initial 87Sr/86Sr (ISr) ranging from 0.707254 to 0.707322, negative εNd(t) values ranging from − 2.8 to − 3.6, and a range of εHf(t) values spanning from − 4.55 to + 13.01, indicating partial melting of the juvenile lower crust (ΔεHf(t) = 3.0–10.3; Sun et al. 2021).

The negative bulk-rock εNd(t) values associated with positive εHf(t) values, together with the very different model ages derived from each isotopic system, signal unusual decoupling characteristics of Nd–Hf isotopic system (Chauvel et al. 2008, 2009). Disequilibrium melting processes and mantle source inheritance are both possible causes (Tang et al. 2014; Sun et al. 2020). Because our zircons are euhedral, prismatic and mostly have clear oscillatory zoning (Fig. 3a), the disequilibrium melting process is unlikely (Tang et al. 2014; Sun et al. 2021; Wang et al. 2021). However, Nd–Hf isotope decoupling can also indicate a source involving subducted sediment. This model may be consistent with the genesis of the granodiorites which have noticeable Zr and Hf depletion plus variable zircon Hf isotopic compositions but consistent bulk Sr–Nd isotopes (Figs. 5a and 8c) (Tang et al. 2014). The “zircon effect” has also been used to interpret other igneous rocks during the Cretaceous. For example, the Namuqie granite [ISr = 0.7058 to 0.7067, εNd(t) =  − 1.1 to − 0.8, εHf(t) =  + 2.7 to + 9.5] in the SQT, which is derived by partial melting of pre-existing juvenile arc crust involving a higher proportion of mature crust (20%–40%) (Wang et al. 2021). Another is the Duolong andesite in the SQT [ISr = 0.7045–0.7071, εNd(t) =  − 1.8 to + 3.6, εHf(t) =  + 1.3 to + 12.9], which is derived from partial melting of the juvenile crust involving altered oceanic basaltic crust, mantle wedge peridotite, and subducted sediments (Sun et al. 2021). Thus, the spatial and temporal distributions of Yonglaga high Ba–Sr granodiorites are consistent with derivation by partial melting juvenile arc crust, indicating the similarity of the juvenile arc crust in the SE Lhasa and the SQT.

5.1.3 The reworking of metasomatized juvenile crust

The granodiorites have notably high K2O contents (4.10wt%–4.82wt%), K2O/Na2O ratios (1.00–1.24), P2O5 contents (0.34wt%–0.41wt%), total REE (297–362 ppm) and LREE/HREE ratios (Fig. 4 and 5), indicating a shoshonitic affinity. High Ba–Sr and shoshonitic granitoids can be generated through analogous geological processes (Bersan et al. 2020; Tarney and Jones 1994; Fowler et al. 2008). The high Ba–Sr granites spatially overlap and adakitic rocks in SE Lhasa (Fig. 4, 5, and 9c, d). The Yonglaga granodiorites have relatively high Sr/Y (30.92–38.18) and (La/Yb)N (27.7–34.7) ratios, but they are not plotted into the field of adakitic rocks. There has been considerably debated, whether the high Sr/Y and (La/Yb)N geochemical indicators can accurately indicate crustal thickening (Moyen 2009; Wang et al. 2022a, b). The lack of significant correlations between SiO2 and Sr/Y, as well as SiO2 and Dy/Yb in the Yonglaga granodiorites suggests that the high Sr/Y (30.9–38.2) characteristic observed in the Yonglaga granodiorites cannot be solely explained by dominant processes like fractional crystallization involving garnet/amphibole, or by magma mixing. The derived melts from thickened crust will exhibit elevated Sr/Y and (La/Yb)N values with a positive correlation between La/Sm and Dy/Yb ratios (Wang et al. 2023). The lack of such a positive correlation in our samples suggests that the garnet-bearing thickened crust cannot be considered as the source of high Sr/Y characteristic. Therefore, the Yonglaga granodiorites have high Sr/Y and (La/Yb)N values, which can be attributed to inheritance from a magma source rather than crustal thickening (Moyen 2009). Further, for melting taking place within the rutile stability range, the resultant melts would display Nb/Ta ratios (> 16) that are higher than those of the source (Wang et al. 2023). Low Nb/Ta ratios (< 14) in Yonglaga samples suggest that the depletion of Nb, Ta, and Ti in Yonglaga granodiorites relates to derivation mafic arc lower crustal sources lacking residual rutile. On the other hand, the Yonglaga granodiorites exhibit fractionated LREE/HREE signature and discernable flat HREE-depleted patterns that supports the presence of residual garnet ± amphibole in the source (Patino-Douce 1999) (Figs. 5b and 10a, b).

Fig. 10
figure 10

Major element compositions (wt%) of the Yonglaga high Ba–Sr granodiorites plotted as a ratio between two variables vs. the sum of the same variables (a, b). Compositions of melts generated experimentally by dehydration melting of a wide range of bulk compositions (Patiño Douce 1999). The Ba/Y versus Nb/Y and (Hf/Sm)N versus (Ta/La)N diagram (c, d) (La Flèche et al. 1998)

Full size image

The geochemical characteristics of the Yonglaga granodiorites are similar to post-collisional suites that have been affected by prior subduction events with subsequent metasomatism of the lithospheric mantle. The granodiorites have high Ba/Y (60.2–74.8) and low Nb/Y ratios (0.60–0.70), suggesting the influence of fluid-related metasomatism (Fig. 10c). Similarly, the (Hf/Sm)N versus (Ta/La)N diagram also shows that the source has undergone fluid-related metasomatism (Fig. 10d). Therefore, the high-K calc-alkaline to shoshonitic affinity and high Ba–Sr signature of our samples can be explained by fluid-related metasomatism in the mantle source of the parent mafic magmas.

The contribution of mature crustal material in the mantle source is further evidenced by the uniformly enriched Sr–Nd isotopes [ISr = 0.707254–0.707322, εNd(t) =  − 2.8 to − 3.6], similar to high Ba–Sr rocks from the Caledonian in British which have been explained by mixing between depleted-enriched mantle and 5%–10% sediment (Fowler et al. 2008, noted on Fig. 8b). The high Th/Yb (15.2–20.7) and La/Sm (10.3–12.9) ratios in Yonglaga granodiorites also indicate a significant contribution from subducted sediments in their metasomatized source (Bersan et al. 2020). The presence of low Lu/Hf (0.05–0.06) ratios combined with elevated Th/La (0.38–0.49) and Th/Yb (15.3–20.8) implies the dominance of terrigenous contributions rather than pelagic inputs (Fig. 9a and b). The reaction between terrigenous sediment and dunite leads to the formation of phlogopite pyroxenites, which could produce high-K melts with enriched LILEs (Förster et al. 2019). In addition, the carbonate or carbonatite involvement to explain the significant increase in Sr (~ 1000 ppm) and LREE (La/Yb approximately 10) has been proposed (Fowler et al. 2008; Tarney and Jones 1994). The relatively high P2O5 values (0.34wt%–0.41wt%) in our samples well support the possibility of carbonatite metasomatism (Rudnick et al. 1993).

In summary, the reworking of the juvenile mafic arc lower crust (amphibolite source) constitutes a plausible mechanism for generating the potassic high Ba–Sr Yonglaga granodiorites. The juvenile mafic crust itself was formed through the melting of subduction-related metasomatized mantle, including a contribution of terrigenous sediments.

5.2 Geodynamic implications

An increasing amount of post-collision silicic magmatism has been discovered to be associated with continental growth, particularly the high Ba–Sr granitoids derived from the mantle (Gómez-Frutos et al. 2023; Fowler et al. 2008). However, the significance of the relationship between crust-derived high Ba–Sr granitoid and continental growth is often overlooked due to its indirect association. Numerous Late Cretaceous granitoids (ca. 120–100 Ma), derived from the melting of juvenile crust, have been documented in the SQT (Li et al. 2013; Liu et al. 2014, 2017; Sun et al. 2021). The Yonglaga granodiorites show similar zircon Hf isotopic signature (εHf(t) values =  − 4.55 to + 13.01) with these (εHf(t) values =  − 1.3 to + 13.6), suggesting the presence of similar juvenile crust in SE Lhasa (Liu et al. 2014, 2017). Sun et al. (2021) have argued that the juvenile lower crust in the SQT is a mélange of MORB, mantle wedge peridotites, and subducted sediments. The melting of such mélange protoliths plays an important role in Nd–Hf isotopic decoupling and adakitic characteristics observed in the Late Cretaceous SQT rocks. Although there are recycled sediments at the source, the dominant depleted zircon Hf isotopic composition still implies significant crustal growth.

The petrogenesis of Yonglaga high Ba–Sr granodiorites and the closure of BNTO studies indicate the occurrence of post-collision extension during the Late Cretaceous in SE Lhasa (Hu et al. 2022; Kapp et al. 2005, 2007). According to the extensional setting, upwelling of asthenosphere mantle material could cause partial melting of the lithospheric mantle and result in the formation of mafic magma. The Late Cretaceous Yonglaga basaltic rocks represent underplated mafic magma in the lower crust, which could provide a significant amount of heat that leads to partial melting of the juvenile crust and subsequently resulting in the formation of granodiorite (Zhu et al. 2024). The whole-rock Zr saturation temperature (794–805 ℃) of the Yonglaga granodiorites further indicates the contribution of underplated mafic magma.

A model involving lithospheric delamination has been widely proposed to explain the post-collisional magmatic activity, in accordance with the inferred crustal thickening during the Late Cretaceous (He et al. 2019; Li et al. 2013; Lu et al. 2019). Abundant adakitic rocks with high Mg and/or Mg# have been found in central Xizang (Table 5, Fig. 1a), including the NLT and CLT (Cao et al. 2022; Chen et al. 2015; Li et al. 2013; Ma and Yue 2010; Meng et al. 2014; Qu et al. 2006; Sun et al. 2015a, 2020; Wang et al. 2014; Yu et al. 2011; Yi et al. 2018; Zhao et al. 2008), and SQT (Li et al. 2013, 2016; Liu et al. 2014, 2017). These have significantly positive zircon Hf isotope signatures, suggesting a depleted mantle source. Recent studies have provided evidence for a westward migration of small-scale lithospheric delamination from the CLT and NLT and the SQT during this period (Li et al. 2016; Wang et al. 2023; Yi et al. 2018). A plausible geodynamic mechanism is lithospheric foundering by Rayleigh–Taylor instability (Houseman and Molnar 1997; Li et al. 2016; Yi et al. 2018). Sun et al. (2020) have demonstrated that localized mantle convection has partially eradicated the lithospheric mantle keel, leading to crustal and lithospheric mantle thinning during the LT and SQT collision. Such partial lithospheric delamination would have resulted in an increased geothermal gradient within the lithospheric mantle, thereby inducing the melting of the thinned crust. However, the Yonglaga granodiorites and coeval mafic rocks do not show adakitic signature (Polat and Kerrich 2001; Zhu et al. 2024). As above mentioned, the petrogenesis of the Yonglaga granodiorites shows that the crust thickness was normal in SE Lhasa during the Late Cretaceous. In addition, few reports of adakites in the SE Lhasa regionally. When the lithosphere undergoes delamination, the thinned crust becomes more susceptible to partial melting and results in the formation of large areas of magmatic rocks, but the scattered and small-volume igneous rocks are distributed in the SE Lhasa. Therefore, our study did not uncover any evidence to support the mechanism of both crust thickening and delamination in the SE margin of Lhasa. It is worth noting that our samples exhibit similar Hf isotopic characteristics to coeval adakites in the CLT and NLT (Fig. 11a), indicating that the similar source from the juvenile crust. But the relative lower (La/Yb)N (27.7–34.7) and Sr/Y (30.9–38.2) ratios than coeval adakites derived from partial melting thickened lower crust in CLT and NLT, our samples also are plotted out of the field of adakitic rocks (Figs. 9c, d and 11b). Thus, it could be not inferred the crust thickening setting but implies the normal crustal setting in the SE Lhasa terrane. Considering that our samples were collected from the SE margin of the Lhasa terrane, a heterogeneous crustal thickness could have occurred during the Late Cretaceous. The central and northern parts of the Lhasa terrane are thickened but the SE margin could be normal (Fig. 12a).

Table 5 Summary of Late Cretaceous magmatisms in the central and northern Lhasa Terranes and southern Qiangtang block
Full size table
Fig. 11
figure 11

a The Zircon Hf components vs. zircon U–Pb data of in the central and northern Lhasa and southern Qiangtang terranes; b the Sr/Y and (La/Yb)N ratios versus longitude data of Late Cretaceous adakites and high Ba–Sr granodiorites in Lhasa terrane

Full size image
Fig. 12
figure 12

a The model of the Late Cretaceous magmatism in the Lhasa terrane; b the model of the Late Cretaceous magmatism in the Lhasa and southern Qiangtang terranes following final Lhasa-Qiangtang amalgamation

Full size image

Recently, the extension could have been driven by the far-field effects of subduction of the Neo-Tethys oceanic crust (Wang et al. 2019, 2021). The BN slab break-off during the Late Cretaceous (~ 110 Ma) and driven by the far-field subduction of the Neo-Tethys oceanic crust (Chapman et al. 2018). The roll-back of the Neo-Tethyan oceanic slab during the Late Cretaceous (~ 85 Ma) could result in scattered and small-volume igneous rocks in the LT and SQT (Wang et al. 2021). In addition, most of the Late Cretaceous felsic igneous rocks are found from the SQT to the Gangdese zone. They were formed at various magmatic temperatures (600–900 °C), indicating that their formation could have been controlled by the mechanism of the Neo-Tethyan oceanic slab (Wang et al. 2021). This mechanism has caused the reworking of juvenile crust and triggered shoshonitic magmatism in central and eastern Lhasa during the Late Cretaceous, which is consistent with the sporadic occurrence and limited volume of Late Cretaceous (ca. 100–80 Ma) magmatic rocks in the SE Lhasa (Wang et al. 2019, 2021). Therefore, the far-field effect of Neo-Tethyan oceanic slab roll-back during the Late Cretaceous period could lead to crustal melting and thus result in the formation of the Yonglaga high Ba–Sr granodiorites (Fig. 12b).

6 Conclusions

  1. (1)

    The high Ba–Sr granodiorites were emplaced at 87 Ma, derived from the reworking of the juvenile mafic arc lower crust that itself formed by melting of metasomatized mantle.

  2. (2)

    Their isotopic characteristics are similar to the coeval juvenile crust-derived magmatic rocks in the southern Qiangtang terrane, suggesting that the decoupled Nd–Hf isotopic system has been inherited from a complexly metasomatized mantle source, which implies the growth of continental crust.

  3. (3)

    The underplated mafic magma provides a significant amount of heat, leading to partial melting of the juvenile crust and subsequently formation of the high Ba–Sr granodiorite.

  4. (4)

    Both the Late Cretaceous adakitic rocks and high Ba–Sr granodiorites have similar juvenile crustal sources, but they could be derived from heterogeneous thickness. There is crustal thickening and delamination in the center of the Lhasa terrane, and the crust thickness remains normal at the SE margin.

  5. (5)

    The Late Cretaceous igneous rocks in Yonglaga were formed in a post-collisional extensional setting, which was possibly triggered by the far-field effects of the subduction of the Neo-Tethyan oceanic crust.