Binary mixing of lithospheric mantle and asthenosphere beneath Tengchong volcano, SE Tibet: evidence from noble gas isotopic signatures

ABSTRACT

The Miocene–Quaternary Tengchong volcanic field in SE Tibet was generated after the main stage of continental collision between India and Eurasia. Consensus on the origin of Tengchong volcanism has not yet been achieved. In this study, we analysed the He-Ar isotopic compositions of olivine and pyroxene phenocrysts, whole-rock major and trace elements and Sr-Nd-Pb isotopic compositions. The ³He/⁴He ratios range from 4.1 to 8.2 Ra (Ra = 1.4 × 10⁻⁶), overlapping the values of mid-ocean ridge basalt (7.0–9.0 Ra) and subcontinental lithospheric mantle (SCLM, 5.2–7.0 Ra). The low ³He/⁴He (< 7.0 Ra) basalts have high ⁸⁷Sr/⁸⁶Sr (average of 0.707928) and La/Yb (average of 22.0) ratios and low Nb/La (average of 0.36) ratios, whereas the high ³He/⁴He (> 7.0 Ra) basalts exhibit relatively low ⁸⁷Sr/⁸⁶Sr (average of 0.706708) and La/Yb (average of 17.0) ratios and relatively high Nb/La (average of 0.52) ratios. These observations indicate that the primitive magmas originated from a mixture of metasomatized SCLM and enriched asthenospheric mantle. The SCLM was likely metasomatized by the subducted Neo-Tethyan oceanic plate, while the asthenospheric mantle was enriched by the subducting Indian oceanic plate. The increasing trend of ³He/⁴He ratios and decreasing trend of ⁸⁷Sr/⁸⁶Sr ratios over time suggest that the contribution of the metasomatized SCLM decreased after the late Pleistocene relative to that of the enriched asthenosphere, reflecting progressive lithospheric extension and thinning. Our results reveal that magmatic He isotopes can be used to constrain deep dynamic processes.

Graphic abstract

KEYWORDS:

• Intraplate volcano
• tengchong
• helium isotope
• trace element
• SCLM

1 Introduction

The convergence between the Indian and Eurasian plates has resulted in intense seismic activity, strike-slip fault zones, southward movement of the Burma-Tengchong Terrane and widespread Cenozoic volcanism. The Tengchong volcanic field is located within the Tibet-Yunnan Fold System, which is separated from the South China Block by the Red River fault to the east and from the Burma Central Lowlands to the west (Zhou et al. 2012) (Figure 1). The Tengchong volcanic field is located in a pull-apart basin due to the stress from NNE-NE compression and WNW-NW extension (Wang et al. 2007). Volcanism in this field initiated at approximately 8 Ma, significantly after the main stage of continental collision. The field is dominated by basalts, basaltic andesites, andesites and dacites that are similar to the arc volcanic rocks in central Myanmar (Zhu et al. 1983; Lee et al. 2016). However, different trace element patterns and Sr-Nd isotope compositions require different mantle sources (Zhou et al. 2012; Guo et al. 2015; Lee et al. 2016).

Figure 1. Major tectonic units in the collision and subduction zones between the Indian and Eurasian plates. (a) Overview map showing the main tectonic settings. The red triangles denote Quaternary volcanoes in Myanmar and Yunnan, China. The purple stars denote the great earthquakes with magnitudes larger than Mw 7.5 around Tengchong. The black lines mark the major faults in and near the Tibetan Plateau (Mo et al. 2006). The red rectangle indicates the study area. (b) Location of the Tengchong volcanic field in southwestern China (modified from Zhou et al. 2012). (c) Distribution of magmatic rocks from different periods in the Tengchong volcanic field (modified from Cheng et al. 2020).

Three main models have been proposed to explain the source and origin of the Tengchong volcanism. The first model suggests that the volcanism originated from the subcontinental lithospheric mantle (SCLM) metasomatized by the subducted Neo-Tethyan oceanic plate (Zhao and Fan 2010; Zhang et al. 2012; Cheng et al. 2018). The second model proposes that the Tengchong magma was generated from asthenospheric mantle that was enriched by the subducting Indian plates (Chen et al., 2002; Tian et al. 2018; Duan et al. 2019; Cheng et al. 2020). In the third model, melts are thought to have resulted from the dehydration of the stagnant slab in the mantle transition zone (MTZ), generating melts in the overlying asthenosphere (Lei et al. 2009; Huang et al. 2015).

Noble gas isotopes, particularly He isotopes, are able to resolve mantle sources. For Martelli et al. (2004), (2008)) observed clear covariations in He and Sr isotopes in Italian Plio-Quaternary volcanism. They attributed the He-Sr isotopic covariation to binary mixing between an asthenosphere and an enriched mantle end-member that was metasomatized by the subducted plate. Dodson et al. (1998) analysed the He, Sr, and Nd isotopes in Tertiary basalts from the Basin and Range Province of western America. They found that the lowest ³He/⁴He basalts were characterized by high ⁸⁷Sr/⁸⁶Sr and La/Nb ratios. In contrast, high ³He/⁴He basalts had low ⁸⁷Sr/⁸⁶Sr and La/Nb ratios. Based on these characteristics, Dodson et al. (1998) suggested binary mixing between the asthenosphere and lithosphere.

Previous He isotope studies of hot springs gas revealed mantle ³He/⁴He ratios (0.2–5.9 Ra, average 2.7 Ra), which are higher in the central Tengchong basin than in the eastern and western regions (Zhao et al. 2012; Cheng et al. 2014). However, the He isotope compositions of these fluids from hot springs can be easily contaminated by crustal radiogenic He and thus cannot reflect true magmatic values. Therefore, the He isotopes of gases from hot springs are not able to reveal the mantle sources for the Tengchong volcano.

To determine the mantle sources of the melts beneath the Tengchong volcanic field, we first report He-Ar isotopes from olivine and pyroxene phenocrysts combined with whole-rock geochemical and Sr-Nd-Pb isotope characteristics from a suite of late Miocene–Holocene basalts. This study may provide a new understanding of the origin of volcanism and constrain the Cenozoic mantle dynamics in SE Tibet associated with plate subduction, collision and extension.

2 Geological setting and samples

The Indian slab subducted eastward beneath Myanmar following the closure of the Neo-Tethys Ocean and the collision between the Indian and Eurasian plates during 60–40 Ma (Leech et al. 2005; Zhu et al. 2005; Copley et al. 2010). The volcanism of the Tengchong field began in the late Miocene (7.2 Ma), with the most recent eruption occurring in AD 1609 (Fan et al. 1999). The volcanic rocks overlie Palaeozoic-Mesozoic sedimentary rocks and Mesozoic-Cenozoic granites (Guo et al. 2015). Tengchong volcanic deposits formed primarily in four periods: (1) late Miocene–Pliocene basalts (8.0–2.7 Ma), which are named Group 1; (2) early Pleistocene trachyandesites and dacites (2.7–0.8 Ma), Group 2; (3) late Pleistocene basalts and trachyandesites (0.8–0.2 Ma), Group 3; (4) Holocene basaltic trachyandesites and trachyandesites (< 0.2 Ma), Group 4 (Cheng et al. 2020). The locations of volcanism have shifted over time from the western and eastern margins to the centre of the Tengchong basin. Cheng et al. (2018) analysed the basalts in marginal and central basins and found that the marginal basalts are more alkaline than the central basalts. In addition, the basalts in central Tengchong show lower ⁸⁷Sr/⁸⁶Sr and higher ¹⁴³Nd/¹⁴⁴Nd ratios than those in marginal areas (Zou et al. 2017; Cheng et al. 2018). Both element and isotopic characteristics suggest that the tectonic setting played a role in governing the chemistry of the erupted lavas.

We collected forty fresh volcanic rocks from the Tengchong volcanic field that dating from the late Miocene to the Holocene (Figure 2). Ten rock samples are from Group 1; three are from Group 2; seven are from Group 3; and twenty are from Group 4. Major and trace elements were measured in all samples. Furthermore, we selected fifteen typical rocks for Sr-Nd-Pb isotope examination by considering their age, type and location. Then, ten porphyritic samples were chosen for He-Ar isotope analysis.

Figure 2. Photomicrographs of minerals and inclusions in the Tengchong volcanic rocks. (a) Columnar jointed basalt in Wuhe County. (b) Basalt with a vesicular structure. (c-d) Olivine, plagioclase and clinopyroxene phenocrysts in magmatic rocks. (e) Inclusions in an olivine phenocryst. (f) Magnified images outlined by the red rectangle in (e), showing fluid and melt inclusions in the olivine phenocryst.

3 Methods

3.1 Whole-rock major and trace elements

Whole-rock major and trace elemental analyses were carried out at the Key Laboratory of Crustal Dynamics, China Earthquake Administration. For major element analysis, the rock samples were ground to 200 mesh in an agate mortar. Then, ~0.5 g powder was mixed with ~5.0 g Li₂B₄O₇ and heated to obtain fused glass beads. Major elements in the fused glass beads were measured using an Axios-Minerals sequential X-ray fluorescence (XRF) spectrometer. The loss on ignition (LOI) of the samples was evaluated after heating the sample for two hours at a constant temperature of 1000°C and thirty minutes of cooling to ambient temperature. For the major elements greater than 1 wt.%, the uncertainty is less than 1%.

For trace element examination, the whole-rock powders (25 mg) were dissolved in distilled 1.5 ml HNO₃ and 1.5 ml HF in Teflon capsules and then heated at 50°C for 12 hours. The solutions were evaporated to dryness at 130°C, and the residue was again digested with 1.5 ml HNO₃ and 1.5 ml HF. The solutions were heated at 170°C for 72 hours and then evaporated to dryness. The residue was redissolved in 3 ml HNO₃ at 150°C for 5 hours to completely dissolve the samples, and the obtained solutions were diluted in 1% HNO₃ to 50 ml for analysis. The solutions were measured by inductively coupled plasma-mass spectrometry (ICP-MS) (ELEMENT Co. Ltd.). Three replicates and two standards (GSR2 and GSR3) were analysed to evaluate the entire procedure and instrument stability. The repeated analysis has a deviation of less than 4%. The analysed standards are in agreement with the recommended values, and the deviation is less than 5% for the measured values.

3.2 Whole-rock Sr-Nd-Pb isotopes

Sr, Nd and Pb isotopic ratios were analysed at the Institute of Geochemistry, Chinese Academy of Sciences. For Sr-Nd isotope analyses, 100 mg samples were decomposed by a mixture of HF-HNO₃-HClO₄ in Teflon capsules. The separation of Sr and Nd was achieved by using the classical two-step ion exchange chromatographic method (Guo et al. 2015), and the collected Sr and Nd fractions were evaporated and dissolved in 2% HNO₃ for subsequent analysis on a Neptune plus mass spectrometer. The mass fractionation corrections for Sr and Nd isotopic ratios were based on ⁸⁶Sr/⁸⁸Sr = 0.1194 and ¹⁴⁶Nd/¹⁴⁴Nd = 0.7219, respectively. Two international standard samples (NBS987 and JNdi-1) were measured to evaluate the instrumental stability for mass fractionation calibration. In our analysis, the international standard NBS987 had a ⁸⁶Sr/⁸⁸Sr ratio of 0.710245 ± 0.000016 (n = 5), and JNdi-1 had a ¹⁴³Nd/¹⁴⁴Nd ratio of 0.512117 ± 0.000012 (n = 5). The whole procedural blank was less than 2 × 10⁻¹⁰ g for Rb-Sr isotopic analysis and 5 × 10⁻¹¹ g for Sm-Nd isotopic analysis.

For Pb isotopic analysis, 100–150 mg sample powders were dissolved in Teflon capsules by a mixture of HF-HNO₃. Pb was separated by using AG1-X8 resin (100–200 mesh) with HBr eluent. Pb isotopic ratios were measured on a Neptune plus mass spectrometer and corrected based on replicate analyses of the international standard NBS981. Repeated analyses of the NBS981 standard yielded ²⁰⁶Pb/²⁰⁴Pb = 16.94046 ± 0.00064 (n = 5), ²⁰⁷Pb/²⁰⁴Pb = 15.48914 ± 0.00063 (n = 5) and ²⁰⁸Pb/²⁰⁴Pb = 36.71602 ± 0.00165 (n = 5).

3.3 Noble gas isotopes

Two steps were adopted to obtain purified phenocrysts. In the first step, olivine and pyroxene phenocrysts (80–120 μm) were separated from the matrix by magnetic and heavy liquid techniques and then manually picked under a binocular microscope. In the second step, the minerals were cleaned in an ultrasonic bath with 5% HNO₃, deionized water, ethanol and acetone to eliminate the possible alteration items on the surface or in the cracks of the phenocrysts. After these two steps, each sample was dried, and approximately 2 g of phenocrysts for each sample was loaded into the crusher chamber.

The noble gas isotope analyses were conducted at the Institute of Geology and Geophysics, Chinese Academy of Sciences, using a Noblesse mass spectrometer (He et al. 2011). Under vacuum conditions, the samples in the crusher chamber were heated to 120°C for 72 hours to remove any adsorbed atmospheric noble gas. The phenocrysts in the vacuum chamber were crushed 15 times within 5 minutes using a piston device with a pressure of approximately 1500 psi. Mechanical crushing is better for extracting gases trapped in inclusions while avoiding radiogenic and cosmogenic gases being emitted from the crystal lattice. The extracted noble gases were introduced into a purifying system and underwent separation. With one cold finger (at liquid nitrogen temperature) and four SAES Zr-Al getters (two at room temperature, and the other two at 450°C), argon was trapped on charcoal at liquid nitrogen temperature, and then helium isotopes were analysed. The argon was released at room temperature and imported into the mass spectrometer for analysis.

Blank experiments without sample loading were conducted prior to the noble gas sample analyses, and the blank experimental results are ⁴He < 1.0 × 10⁻¹⁰ cm³ standard temperature and pressure (STP) and ⁴⁰Ar < 1.0 × 10⁻⁹ cm³ STP. The results were used to correct all isotopic measurements. The artificial helium standard gas HESJ (³He/⁴He = 20.63 ± 0.10 Ra; Matsuda et al. 2002) and air were used as standards to normalize the measured He and Ar isotopic ratios, and the measurement uncertainties are expressed in 1 σ, which are generally better than ± 0.6 Ra for ³He/⁴He ratios and ± 7.9 for ⁴⁰Ar/³⁶Ar ratios.

4 Results

4.1 Whole-rock major and trace elements

These volcanic rocks are mainly high-K calc-alkaline basalts, trachybasalts, basaltic trachyandesites, and trachyandesites (supplementary Table S1; Figure 3). The SiO₂ and MgO contents show large variations and are 47.6–64.2 wt.% and 2.0–8.1 wt.%, respectively. Samples from Group 1 and Group 3 are more mafic than those from Group 2 and Group 4 (Figure 3a). As the MgO content decreases, the SiO₂, TiO₂, total FeO (TFeO), CaO/Al₂O₃, Ni and Cr contents generally decrease, while the K₂O and Na₂O contents increase (supplementary Figure S1). The Tengchong basalts (squares in supplementary Figure S1) do not show clear relationships between MgO and other major oxides.

Figure 3. Whole-rock major element compositions of the Tengchong volcanic rocks. (a) Diagram of the total alkali (K₂O+Na₂O) versus SiO₂. The classification boundaries are from Le Bas et al. (1986) and Le Maitre et al. (1989). TB and BTA are the abbreviations for trachybasalt and basaltic trachyandesite, respectively. (b) Diagram of K₂O versus SiO_2. The boundaries are from Peccerillo and Taylor (1976) and Rickwood (1989). The symbols used in this figure are identical to those in other figures.

The primitive mantle-normalized incompatible trace element pattern (Figure 4a) shows positive anomalies in large ion lithophile elements (LILEs, e.g. K, Rb and Ba) and negative anomalies in high field strength elements (HFSEs, Nb, Ta, Zr, Hf and P). These patterns are similar to those of Indian sediments. Negative Nb-Ta-Ti anomalies and positive Pb anomalies indicate subduction-related volcanism (Figure 4a). The Tengchong basalts are enriched in light rare earth elements (LREEs) (Figure 4e). The differentiated magmas from Group 2 and Group 4 exhibit higher total rare earth element (REE) contents and strong negative Eu anomalies.

Figure 4. Diagram of primitive mantle-normalized trace elements (a-d) and chondrite-normalized rare earth elements (e-h). The normalization factors are from Sun and McDonough (1989). The red curve is the database for ocean Island basalts (OIBs) (Sun and McDonough 1989); the black curve is the database for Indian sediments (Plank and Langmuir 1998); the green curve is the database for enriched mantle (EM) with type-II features (Workman et al. 2004); and the blue curve is the database for normal mid-ocean ridge basalts (N-MORB) (Sun and McDonough 1989).

4.2 Sr-Nd-Pb isotopes

The fifteen samples selected for isotope analysis are basalt, trachybasalt, basaltic trachyandesite, trachyandesite and dacite. They exhibit high ⁸⁷Sr/⁸⁶Sr ratios (0.705945–0.708851), relatively low ¹⁴³Nd/¹⁴⁴Nd ratios (0.512145–0.512535), and moderately high ²⁰⁷Pb/²⁰⁴Pb (15.61950–15.67448) and ²⁰⁸Pb/²⁰⁴Pb (38.59086–39.15366) ratios (see supplementary Table S1). In the ¹⁴³Nd/¹⁴⁴Nd versus ⁸⁷Sr/⁸⁶Sr diagram (Figure 5b), the samples are located between the EM-I and EM-II end-members. In standard Pb isotope diagrams (Figure 5c), the basalts plot between the EM-I and EM-II end-members, and they lie to the left of the Northern Hemisphere Reference Line (NHRL). The relationship between the ⁸⁷Sr/⁸⁶Sr ratio and SiO₂ content is commonly used to determine the influence of crustal contamination on volcanic rocks (Li et al. 2016). For the volcanic rocks with SiO₂ > 52.5%, ⁸⁷Sr/⁸⁶Sr ratios show a positive correlation with SiO₂ contents, which indicates that these samples have undergone crustal contamination. However, for the samples with SiO₂ < 52.5%, a positive correlation between ⁸⁷Sr/⁸⁶Sr and SiO₂ contents is not found, implying that crustal contamination has no significant effect on these samples. Therefore, the ⁸⁷Sr/⁸⁶Sr variation in the basalts is mainly attributed to the heterogeneity of the magma source.

Figure 5. Sr-Nd-Pb isotopic ratios of Tengchong volcanic rocks. Plots (a-d) are diagrams of ⁸⁷Sr/⁸⁶Sr versus SiO₂, ¹⁴³Nd/¹⁴⁴Nd versus ⁸⁷Sr/⁸⁶Sr, ²⁰⁷Pb/²⁰⁴Pb versus ²⁰⁶Pb/²⁰⁴Pb, and ²⁰⁸Pb/²⁰⁴Pb versus ²⁰⁶Pb/²⁰⁴Pb, respectively. The compositions of depleted MORB mantle (DMM) are from Workman and Hart (2005); those of EM-I and EM-II are from Hofmann (1997); and the Northern Hemisphere Reference Line (NHRL) is from Hart (1984). Gray dots in (a) are the samples in the literature (Chen et al., 2002; Zhou et al. 2012; Guo et al. 2015; Zou et al. 2017; Cheng et al. 2018, 2020; Tian et al. 2018). The isotopic compositions of two geochemical end-members in plot (b) are based on Miller et al. (1999) and Cheng et al. (2020). In plot (b), the dotted curve denotes two-component mixing.

4.3 Helium and argon isotopes

The isotopic compositions of noble gases are presented in supplementary Table S2. The ⁴He concentrations range from 2.5 × 10⁻⁹ to 60.2 × 10⁻⁹ ccSTP/g, and the ³He/⁴He ratios range from 4.1 to 8.2 Ra (Figure 6a). The ⁴⁰Ar/³⁶Ar ratios range from 376.9 to 872.3 in basalt phenocrysts and the ratio is 549.9 in pyroxene phenocryst. These values fall within the typical range of He-Ar ratios associated with subduction-related volcanism (Hilton et al. 2002; Sano and Fischer 2013). The calculated ⁴He/⁴⁰Ar* ratios range from 0.02 to 1.33.

Figure 6. He-Ar isotope diagram for Tengchong volcanic rocks. (a) ³He/⁴He versus ⁴He/⁴⁰Ar*. The ranges of MORB and SCLM are from Graham (2002) and Gautheron and Moreira (2002), respectively. Green diamond denotes the pyroxene phenocryst and the other samples are all olivine phenocrysts. (b) ³He/⁴He versus ⁴He concentrations. (c) ³He/⁴He versus Zr/Y. The sample in the red dashed circle represents magmas that underwent crustal assimilation.

5 Discussion

5.1 Variation in ³He/⁴He ratios in basalts

The ³He/⁴He ratios of the Tengchong basalts range from 4.1 to 8.2 Ra, which overlap the range of mid-ocean ridge basalts (MORBs), SCLM, and many subduction-related fields (Martelli et al. 2004). To better illustrate the values of ³He/⁴He representing the nature of the magma source, it is necessary to clarify the potential influences on the variation in ³He/⁴He ratios during the ascension and after the eruption of magma. Four factors may affect the variation: (1) element fractionation during magma degassing; (2) crustal contamination during magma ascension through the crust; (3) radiogenic ⁴He from U and Th decay within the crystal lattice; and (4) cosmogenic ³He accumulation.

First, during magma degassing, light nuclides diffuse faster than heavy nuclides; therefore, ³He spreads more rapidly than ⁴He, and He migrates faster than Ar (Ozima and Podosek 2002). Based on this mechanism, the effects of noble gas diffusion fractionation are typically estimated using ³He/⁴He versus ⁴He/⁴⁰Ar* (the definition of ⁴⁰Ar* is presented in supplementary Table S2). Figure 6a illustrates that the variation in the ³He/⁴He ratio does not correlate with that of the ⁴He/⁴⁰Ar* ratio, suggesting that all basalts exhibit nearly the same diffusion fractionation. Therefore, diffusion fractionation is not an important factor influencing the variation in the ³He/⁴He ratio.

Second, investigations of Sr isotopes have shown that crustal contamination does not have a clear effect on basalts (Figure 5a). Crustal contamination also does not typically affect olivine ³He/⁴He ratios, as melt inclusions are trapped prior to residence in the crust (Stuart et al. 2000). In addition, based on the fact that the Zr/Y ratio increases with the degree of crustal assimilation (Conly et al. 2005; Li et al. 2017), the effect of crustal assimilation on the basalts in Group 1 and Group 3 can be excluded because the ³He/⁴He ratio from these basalts does not show an evident relationship with the Zr/Y ratio (Figure 6c).

Third, the decay of radioactive U and Th within the crystal lattice might yield radiogenic ⁴He, causing a decrease in the ³He/⁴He ratio (Xu et al. 2014; Dai et al. 2016). However, the crushing of phenocrysts under vacuum conditions during sample preparation extracted noble gas mainly from inclusion, which are free of the radiogenic ⁴He produced within the crystal lattice (Stuart et al. 2000). In addition, as shown in Figure 6b, we do not observe a clear negative correlation between the ³He/⁴He ratio and the ⁴He concentration, suggesting that radiogenic ⁴He is unlikely to reduce the ³He/⁴He ratio.

Fourth, Figure 6b shows that samples TC2041 and TC2050 (the locations are presented in supplementary Table S1) exhibit low ⁴He contents and slightly high ³He/⁴He ratios. Cosmogenic ³He can result in this phenomenon, but this origin is excluded according to the analysis method and sampling location. Specifically, cosmogenic ³He is mainly in the lattice of minerals, and the vacuum crushing of phenocrysts during sample treatment prevents the release of cosmogenic ³He (Carracedo et al. 2019). In addition, TC2042 and TC2050 were collected at a roadcut, so these samples were not exposed to cosmic rays for a long period and hence did not accumulate cosmogenic ³He. Thus, we conclude that the ³He/⁴He ratios of the basalts can represent the nature of the magma sources and were not affected by later processes.

5.2 Mantle geochemistry

Previous studies have proposed three different types of models to explain the origin of Tengchong volcano, such as dehydration of the stagnant slab in the MTZ (e.g. Lei et al. 2009), enriched asthenosphere (e.g. Cong et al. 1994; Guo et al. 2015; Zou et al. 2017; Tian et al. 2018; Duan et al. 2019; Cheng et al. 2020), and metasomatized SCLM (e.g. Cheng et al. 2018). The dehydration of the stagnant slab in the MTZ can be excluded. On the one hand, the existence of a subducted slab in the MTZ is uncertain. According to the receiver function study of Xu et al. (2018), the 660-km discontinuity beneath Tengchong is clearly depressed. This phenomenon is attributed to the stagnant slab in the MTZ based on the negative Clapeyron slope of the transition from ringwoodite to perovskite and magnesiowüstite. However, the stagnant slab in the MTZ is not imaged by teleseismic S-wave tomography (Zhang et al. 2018). On the other hand, trace element data indicative of slab dehydration, such as K/U, Ba/Th and Pb/U, do not support this as a mechanism. The Wudalianchi volcanic field (NE China) is related to the dehydration of stagnant slabs in the MTZ (Yang and Faccenda 2020), and high K/U (~50,000), Ba/Th (~400) and Pb/U (~16.7) ratios are observed in Wudalianchi volcano (Wang et al. 2017). The averages of these ratios in Tengchong volcano are 3186, 52, and 9.5, respectively. Evidently, these ratios are quite different from those in Wudalianchi volcano. Therefore, it is unlikely that the Tengchong magma was derived from the dehydration of the stagnant slab in the MTZ.

Most studies proposed that Tengchong magma was generated from the enriched asthenosphere that was metasomatized by the subducting Indian plates based on the whole-rock compositions and Sr-Nd-Pb isotopes. The enriched asthenosphere usually has MORB-like ³He/⁴He ratios (7–9 Ra; Poreda and Craig 1989; Martelli et al. 2004). However, some basalts show low ³He/⁴He ratios (4.1–6.4 Ra) (supplementary Table S2). This phenomenon may indicate that the enriched asthenosphere is not the only source for the magma of the Tengchong volcano.

The MORB-like high ³He/⁴He cannot be decreased to 4.1–6.4 Ra because of two reasons. On the one hand, the subducted oceanic plate provides very little He to the mantle wedge, resulting in little influence on lowering the ³He/⁴He ratio (e.g. Poreda and Craig 1989; Martelli et al. 2004). On the other hand, postmetasomatic radiogenic He ingrowth is unlikely to reduce the ³He/⁴He from 8.0 Ra to 4.1 Ra. Assuming a high U concentration of 200 ppb and Th concentration of 950 ppb in the metasomatites (Martelli et al. 2004), this would produce radiogenic ⁴He of 2.57 × 10⁻⁶ ccSTP/g in 50 Ma (beginning of the Indian plate subduction). This process would lower the ³He/⁴He from 8.0 Ra to 6.8 Ra, but it is unable to reduce the ³He/⁴He to 4.1 Ra, if the premetasomatic mantle had MORB He concentration (Sarda and Graham 1990). Thus, the subducted Indian oceanic plate is unlikely to result in a low ³He/⁴He ratio (4.1–6.4 Ra) for the volcanic rocks in the Tengchong area.

Besides the contribution from the enriched asthenosphere, we consider that another component contributes to the Tengchong magma source. The correlations between ⁸⁷Sr/⁸⁶Sr and Nb/La, La/Yb (Figure 7) provide strong evidence that the magma source is primarily composed of two different endmembers with distinct isotopic and trace element properties. In addition, the temporal variations in isotopic composition in Tengchong volcanic rocks further imply two-component mixing (described below), which is comparable to that observed in the southwestern Basin and Range Province (e.g. DePaolo and Daley 2000). In this place, isotopic variation with time is explained as the mixing of asthenosphere and lithosphere.

Figure 7. ⁸⁷Sr/⁸⁶Sr versus Nb/La and La/Yb in the Tengchong basalts. Filled squares denote data from this study, and empty squares are from the literature (Zhang et al. 2012; Zou et al. 2017; Cheng et al. 2018, 2020; Tian et al. 2018).

Based on the above discussion, we suggest that the SCLM is probably another component for Tengchong magma source, which is supported by two pieces of evidence. First, the presence of residual hydrous mineral phases (e.g. amphibole and phlogopite) (supplementary Figure S2) indicates that the Tengchong magmas were in part derived from the lithospheric mantle. These minerals are not stable in the hot and anhydrous convecting asthenosphere (Tang et al. 2012). In contrast, they are stable in the lithospheric mantle (Class and Goldstein 1997). Therefore, the Tengchong samples were generated from mixing melts derived from the asthenosphere and lithospheric mantle. Second, the sample TC2043 shows the lowest ³He/⁴He ratio of 4.1 Ra. After the metasomatic enrichment of U and Th from the crustal materials because of plate subducting mechanism, the decay of U and Th causes radiogenic ⁴He ingrowth in the SCLM source, which may lower the ³He/⁴He ratios (e.g. Dodson and Brandon 1999; Martelli et al. 2004). The plate subducting mechanism is related to the Neo-Tethys Ocean subduction, which will be discussed in the subsequent section. According to Martelli et al. (2004), under the conditions of an average U and Th concentration of 1.4 and 10 ppm (Tengchong primitive magma), and 1–9% melting (Zhou et al. 2012; Guo et al. 2015), we can infer that the U and Th concentrations of the present SCLM beneath Tengchong are 14–126 ppb and 100–900 ppb, respectively. Assuming batch melting and perfectly incompatible behaviour of U and Th (D = 0), and the highest inferred U and Th concentrations (126 and 900 ppb, respectively), and an upper time limit of metasomatism beginning at 140 Ma, radiogenic ⁴He ingrowth based on theoretical calculation is about 7.3 × 10⁻⁶ ccSTP/g. This ⁴He ingrowth may reduce the ³He/⁴He ratios in SCLM from 6.1 Ra to 4.1 Ra (formula for this theoretical calculation is from Dodson and Brandon (1999)), which implies that the low ³He/⁴He ratio (4.1 Ra) is possibly derived from the metasomatized SCLM.

5.3 Mixed source for Tengchong volcano

Many studies suggest that the interaction between lithospheric mantle and asthenospheric mantle plays an important role in intraplate basalt formation (e.g. Tang et al. 2006; Konrad et al. 2016; Lee et al. 2021). The relationships of ⁸⁷Sr/⁸⁶Sr versus Nb/La and ⁸⁷Sr/⁸⁶Sr versus La/Yb of the basalts (Figure 7) imply that the Tengchong volcanic rocks were generated from the mixing of two-endmembers. The Nb/La and La/Yb ratios are able to investigate the nature of the magma source (DePaolo and Daley 2000; Ionov et al. 2002; Tang et al. 2012). Lithosphere mantle-derived melts often show distinct lower Nb/La and higher La/Yb ratios than asthenospheric melts (DePaolo and Daley 2000; Tang et al. 2006, 2012). The Nb/La and La/Yb ratios in Tengchong basalts are 0.3–0.8 and 11.6–34.8, respectively. These values demonstrate the mixing between asthenospheric mantle and lithospheric mantle (Zhou et al. 2012; Duan et al. 2019; Lee et al. 2021) (Figure 8a). Furthermore, magmatic He and Sr isotopes can provide important information about mantle sources (e.g. Martelli et al. 2008). The He-Sr mixing model is frequently utilized to interpret the genesis of continental basalts (Dodson et al. 1998; Tang et al. 2006; Konrad et al. 2016; Lee et al. 2021). Our measured He-Sr isotopes reveal that the Tengchong magmas are a hybrid of two-endmembers: a metasomatized SCLM and an enriched asthenospheric mantle (Figure 8b).

Figure 8. Diagrams of a two-component mixing model between metasomatized SCLM and enriched asthenosphere using He-Sr isotopes and trace elements. (a) Plot of La/Yb versus Nb/La in the Tengchong basalts (modified from Watson, 1993; Zhou et al. 2012), and filled squares denote data from this study, and open squares are from the literature (Zhang et al. 2012; Zou et al. 2017; Cheng et al. 2018, 2020; Tian et al. 2018). (b) Plot of ³He/⁴He versus ⁸⁷Sr/⁸⁶Sr. (c) Plot of ³He/⁴He versus Nb/La. (d) Plot of ³He/⁴He versus La/Yb. For metasomatized SCLM, the value of ⁸⁷Sr/⁸⁶Sr is approximately 0.717 (Miller et al. 1999); the values of ³He/⁴He are approximately 4.1–7.0 Ra (this study; Gautheron and Moreira 2002; Lee et al. 2021); the ratio of Nb/La can be 0.3 (the lowest value in Tengchong); and the La/Yb ratio can be 35 (the highest value in Tengchong). For enriched asthenosphere, the value of ⁸⁷Sr/⁸⁶Sr can be 0.705 (Cheng et al. 2020), the ³He/⁴He ratio is approximately 8.0 Ra (Poreda and Craig 1989; Graham 2002), the Nb/La ratio is approximately 0.9 (the highest value in Tengchong), and the La/Yb ratio is approximately 5.0 (the lowest value in Tengchong). The solid lines represent the two-component mixing relationships.

One sample (TC2043) located in the lithospheric area (Figure 8a) exhibits low ³He/⁴He ratio (4.1 Ra) and high ⁸⁷Sr/⁸⁶Sr ratio (0.708851) (Figure 8b). The ³He/⁴He ratios of off-craton SCLM range from 5.2 to 7.0 Ra. The low ³He/⁴He ratio (4.1 Ra) in our study may be caused by radiogenic ⁴He ingrowth in the SCLM after the metasomatic enrichment of U and Th. According to trace element modelling, Cheng et al. (2018) suggested the metasomatized SCLM contributing to Tengchong magma based on the consistency between the constructed metasomatized SCLM model and the measured trace element values. A similar conclusion is also drawn by Zhao and Fan (2010) through the composition of Sr-Nd-Pb isotopes between Nabang metamorphic rocks and the Tengchong basalt. They attributed the metasomatized SCLM to the early subduction of the Neo-Tethyan oceanic plate.

Three basalts display ³He/⁴He ratios ranging from 7.5 to 8.2 Ra and relatively high ⁸⁷Sr/⁸⁶Sr ratios (0.706386–0.707024) (Figure 8b). Such MORB-like ³He/⁴He and relatively high ⁸⁷Sr/⁸⁶Sr ratios in our study may indicate that the asthenosphere beneath Tengchong was metasomatized by He-poor fluids from the subducted slab. Previous studies mainly used Sr-Nd-Pb-Hf-Mg isotopes to investigate the nature of the enriched asthenospheric component. The basalts in previous studies are characterized by high Sr, moderate Pb, and relatively low Nd isotopes. Some researchers have considered that continental material is suitable to interpret these characteristics (e.g. Chen et al., 2002; Zhang et al. 2012). However, considering the low Mg isotopes of the basalts from Tengchong volcano, others have proposed that the metasomatized mantle is probably caused by a subducted Indian oceanic slab (Tian et al. 2018). Based on the Sr-Nd-Pb-Hf isotopes and element ratios, some studies have claimed that the mantle is metasomatized by both Indian continental and oceanic lithosphere (Zhou et al. 2012). Clearly, the source of metasomatism is controversial. We suggest that the source of metasomatized asthenosphere is the subducted Indian oceanic plate based on two pieces of evidence. The first piece of evidence is that the samples with high ³He/⁴He and relatively high ⁸⁷Sr/⁸⁶Sr ratios (Figure 8b) suggest clay-rich sediments on the subducting plate (e.g. Poreda and Craig 1989). The second piece of evidence comes from the tomographic images of the subducted plate. According to Yao et al. (2021), the dip of the subducted plate beneath Tengchong is approximately 60° in the upper mantle. This steep dip of the subducted plate is also demonstrated by both Lei et al. (2009) and Li et al. (2008) through teleseismic tomography. Steep dips are common in the subduction zones related to oceanic plate subduction, such as the Adman Sea, Sumatra, Java and Japan (e.g. Chen et al. 2015; Hall and Spakman 2015; Mishra et al. 2020; Liu et al. 2021). This dip is much larger than that beneath southern Tibet, where the Indian continental plate subducts (Chen et al. 2017). Thus, the enriched asthenosphere is most likely metasomatized by the subducted Indian oceanic plate.

The mixing models of ³He/⁴He-Nb/La and ³He/⁴He-La/Yb can also be used to investigate the mixing between the enriched SCLM and the asthenosphere (e.g. Dodson et al. 1998; Lee et al. 2021). The metasomatized SCLM and enriched asthenosphere are fertile mantle components. Thus, for the metasomatized SCLM end-member, the values of ³He/⁴He are approximately 4.1–7.0 Ra (this study; Gautheron and Moreira 2002; Lee et al. 2021). According to Dodson et al. (1998) and Lee et al. (2021), the Nb/La ratio can be 0.3 (the lowest value in Tengchong), and the La/Yb ratio can be 35 (the highest value in Tengchong). For the enriched asthenosphere end-member, the ³He/⁴He ratio is approximately 8.0 Ra (Poreda and Craig 1989; Graham 2002), the Nb/La ratio is approximately 0.9 (the highest value in Tengchong), and the La/Yb ratio is approximately 5.0 (the lowest value in Tengchong). Figure 8c-8d show that the Tengchong basalts are roughly located between these two end-members. This conclusion of two-endmember mixing can also be supported by the basalt Sr-Nd isotope mixing trend, in which the basalts are distributed along the mixing curve between the metasomatized SCLM and enriched asthenosphere (Figure 5b).

5.4 Isotopic evolution and its geodynamic implication

Combined with previous geochronologic studies (Mu et al. 1987; Guo et al. 2015; Zou et al. 2017; Li et al. 2019; Cheng et al. 2020), we find significant changes in ³He/⁴He and ⁸⁷Sr/⁸⁶Sr ratios in Tengchong basalts over time (Figure 9a). From 7.2 Ma to 0.8 Ma, the ³He/⁴He ratios in basalts gradually increase from 4.1 to 8.2 Ra, and the ⁸⁷Sr/⁸⁶Sr ratios decrease from 0.708013 to 0.708690 (average of 0.708304) to 0.705269–0.706728 (average of 0.706113). This phenomenon suggests that the nature of magma sources changes over time. In the early stage (7.2 Ma), the contribution from the metasomatized SCLM is significant, but this contribution decreases over time. In the late stage, the enriched asthenosphere becomes crucial for two-endmember mixing.

Figure 9. Variations in ³He/⁴He (Ra) and ⁸⁷Sr/⁸⁶Sr ratios in the Tengchong basalts as a function of time (Ma). In plot (a), filled squares denote data from this study, and the open green squares are from Zhang et al. (2021b). In plot (b), the open squares are from the literature (Zhang et al. 2012; Zou et al. 2017; Cheng et al. 2018, 2020; Tian et al. 2018). The ages of samples are based on previous studies (Mu et al. 1987; Li et al. 2019; Guo et al. 2015; Zou et al. 2017; Cheng et al. 2020).

The mechanism for the variation in He-Sr isotopes (Figure 9) is mainly caused by the lithospheric extension, as is depicted in Figure 10. The Burma-Tengchong terrane underwent west-east extension starting at 8 Ma (e.g. Chen et al., 2002; Tapponnier et al. 1982; Harrison et al. 1992; Socquet and Pubellier 2005; Mo et al. 2006; Zhou et al. 2012). This extension resulted in the thinning of the lithosphere, which has facilitated the easy upwelling of deep asthenospheric mantle and its mixing with shallow lithosphere. The thinning of the lithosphere is also demonstrated by seismic tomography and gas geochemistry. The inversion of S-wave receiver functions shows that the lithosphere-asthenosphere boundary (LAB) is at a depth of 70–80 km beneath the Tengchong volcanic field, which is significantly shallower than those beneath surrounding areas (Hu et al. 2012). In addition, the extremely intense release of mantle-derived volatiles is observed in the Tengchong volcanic field by Zhao et al. (2012) and Zhang et al. (2021a). Zhao et al. (2012) consider that this phenomenon is caused by the thinning of the lithosphere. The thinning of the lithosphere has probably occurred in Tengchong since 8.0 Ma, which results in the component variation in the two end-members.

Figure 10. Schematics illustrating the magma origin beneath Tengchong volcano. The magma of Tengchong volcano is from two end-member mixing of metasomatized SCLM and enriched asthenospheric mantle. At 7.2 Ma, the extension of the lithosphere resulted in lithospheric thinning. At 0.8 Ma, the thin lithosphere allowed the easy upwelling of deep asthenospheric mantle and resulted in a significant contribution from the enriched asthenosphere.

6 Conclusions

Based on the new geochemical and isotopic data presented here, the following conclusions can be drawn:

1. He-Sr isotope and trace element ratios indicate that the Tengchong primitive magmas were formed by melt mixing from the metasomatic SCLM and enriched asthenospheric mantle.

2. The SCLM was probably metasomatized by the subducted Neo-Tethyan oceanic plate, and the produced melts are characterized by low ³He/⁴He (< 7.0 Ra) and Nb/La (average of 0.36) ratios and high ⁸⁷Sr/⁸⁶Sr (average of 0.707928) and La/Yb (average of 22.0) ratios.

3. The asthenospheric mantle was enriched by the subducting Indian oceanic plate, and the produced melts are characterized by high ³He/⁴He (> 7.0 Ra) and Nb/La (average of 0.52) ratios and low ⁸⁷Sr/⁸⁶Sr (average of 0.706708) and La/Yb (average of 17.0) ratios.

4. Clear evolution of He-Sr isotopes is observed from the late Miocene to the Holocene. This isotopic evolution over time reflects lithospheric extension and thinning, which led to the upwelling of deep asthenospheric mantle and its mixing with lithospheric mantle.

Highlights:

• ³He/⁴He ratios in Tengchong volcanic rocks range from 4.1 to 8.2 Ra

• Both metasomatized SCLM and enriched asthenosphere contribute to Tengchong volcano

• He-Sr isotopic variation reflects progressive lithospheric extension and thinning

Acknowledgments

The authors are grateful to the Editor Robert J. Stern and two anonymous reviewers for their valuable suggestions that have improved the original paper. We greatly appreciate the discussions with Zhiheng Ren and Yadong Wu, which greatly enriched the content of this article. K. C. conducted sampling and wrote the manuscript. S. L. conceived and executed this research project. H. H., F. S. and X. X. helped to evaluate the results and conclusions. Z. C., J. L., D. Y. and W. W assisted with manuscript preparation. F. S. contributed to the noble gas isotopic analysis and interpretation. All authors contributed to the writing of the manuscript.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed here.

Additional information

Funding

This work was supported by the National Institute of Natural Hazard, Ministry of Emergency Management of China [ZDJ2019-18]; State Key Laboratory of Lithospheric Evolution, Institute of Geology and Geophysics, Chinese Academy of Sciences [SKL-K202101]; National Natural Science Foundation of China [42064004]; Key Laboratory of Petroleum Resources Research, Gansu Province [SZDKFJJ20211001]; Program of the National Key R&D on Monitoring, Early Warning and Prevention of Major Natural Disaster, China [2017YFC1500301].