Whole-rock major and trace element analyses were conducted at the ALS Laboratory Group (Guangzhou) Co., Ltd. Major elements were determined using X-ray fluorescence spectrometry (XRF) with a PANalytical PW2424 instrument, with a relative deviation (RD) of less than 5%. Trace elements were analyzed using inductively coupled plasma mass spectrometry (ICP-MS) with an Agilent 7,900 instrument, with a relative deviation (RD) of less than 10%. The standard samples used were GSR3 and GSR5. Detailed procedures are available in Li et al. (2005).

Two fresh two-mica monzogranite samples (XW-9 and 05QS-4) were selected for in-situ zircon U–Pb dating. The samples were crushed to 100–120 mesh, and zircons were separated using heavy liquid and magnetic separation methods. Zircons with good crystal shapes and clear zonation were handpicked under a binocular microscope, mounted in epoxy resin, and polished to expose their cores. Cathodoluminescence (CL) and back-scattered electron (BSE) imaging were used to select U–Pb spots. This work was completed at Langfang Chenchang Rock Mineral Testing Technology Service Co., Ltd. Zircon U–Pb isotopic dating was performed using a laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) system at the State Key Laboratory of Nuclear Resources and Environment, East China University of Technology. The laser ablation system was a GeoLasHD 193 nm excimer laser, and the ICP-MS instrument was an Agilent 7,900. The laser ablation spot size was 32 μm, with a frequency of 5 Hz and an energy density of 6 J/cm², using He as the carrier gas. Isotopic ratio corrections were made using the 91,500 zircon as an external standard (the recommended age value 1062.4 Ma), the weighted average age of standard zircons 91,500 was 1060.6 ± 5.4 Ma (MSWD = 0.88, n = 12). The TEM (417 Ma) zircon as a monitoring sample. Data processing was performed using the ICPMSDataCal program (Liu et al., 2008, 2010), and concordia diagrams were generated using the Isoplot 3.0 program (Ludwig, 2003). Detailed parameters and procedures can be found in Jackson et al. (2004).

Zircon Lu–Hf isotopic analysis was carried out at Nanjing Hongchuang Geological Exploration Technology Service Co., Ltd. The analysis was performed on and near the U–Pb dating spots using a GeoLasHD 193 nm laser ablation system coupled with an Agilent 7900a ICP-MS. The laser spot size was 40 μm, with a frequency of 10 Hz. During the testing process, the average values of ¹⁷⁶Hf/¹⁷⁷Hf for standard zircon 91,500 and CJ-1 were 0.282310 ± 0.000005 (2σ, n = 14) and 0.282005 ± 0.000014 (2σ, n = 7), respectively, which were in accordance with the recommended ¹⁷⁶Hf/¹⁷⁷Hf values within the error range (0.282305 ± 0.000003, 2σ; 0.282015 ± 0.000019, 2σ) (Elhlou et al., 2006; Wu et al., 2006; Blichert-Toft, 2008). The parameters used for ε_Hf(t) calculations were as follows: decay constant of ¹⁷⁶Lu = 1.865×10⁻¹¹ yr⁻¹ (Scherer et al., 2001), chondritic uniform reservoir (CHUR) values of ¹⁷⁶Hf/¹⁷⁷Hf = 0.282772 and ¹⁷⁶Lu/¹⁷⁷Hf = 0.0332 (Blichert and Albarède, 1997), and crustal model age (T_DM2) calculations using ¹⁷⁶Lu/¹⁷⁷Hf = 0.015 (Griffin et al., 2002). Isotopic ratio corrections were made using the Plesovice zircon (337.00 ± 0.37 Ma) as an external standard (Sláma et al., 2008), and common lead corrections were made using the ComPbCorr#3.17 program (Andersen, 2002).

Sr–Nd isotopic analysis was conducted at Nanjing Hongchuang Geological Exploration Technology Service Co., Ltd. The samples were weighed according to their Rb–Sr and Sm–Nd contents, and dissolved in Teflon bombs with mixed ⁸⁷Rb–⁸⁴Sr and ¹⁴⁹Sm–¹⁵⁰Nd spikes using HF, HNO₃, and HClO₄. Rb, Sr, Sm, and Nd were separated using ion exchange resin columns. Measurements were made using a Thermo Fisher Neptune XT thermal ionization mass spectrometer (TIMS). The analysis process maintained a blank level of Rb, Sr < 100×10⁻¹² and Sm, Nd < 50×10⁻¹². Detailed analytical procedures can be found in Li et al. (2016). During the sample analysis, the measured ⁸⁷Sr/⁸⁶Sr ratio of the international standard NBS987 was 0.710288 ± 0.000028 (2σ, n = 9), and the ¹⁴³Nd/¹⁴⁴Nd ratio of the JNdi-1 standard was 0.512109 ± 0.000012 (2σ, n = 9).

Muscovite electron probe microanalyses were performed at the State Key Laboratory of Nuclear Resources and Environment, East China University of Technology. Large, clean, and flat muscovite grains were selected under a microscope for EPMA. The instrument used was a JEOL JXA-8530 with an accelerating voltage of 15 kV, a current of 20 nA, a beam diameter of 5 μm, and a counting time of 20 s. The standards used were olivine (Si), rutile (Ti), almandine (Al), garnet (Fe), wollastonite (Mn), diopside (Mg, Ca), jadeite (Na), orthoclase (K), apatite (P), LiF (F), and scapolite (Cl), with corrections made using the ZAF method. The mica cation numbers and related parameters were calculated based on 22 oxygen atoms, and the Li₂O content was calculated as [2.1/(0.356+MgO)]-0.088 (Tischendorf et al., 1999).

Twenty-five of each sample zircon crystals were selected from two-mica monzogranite samples (XW-9 and 05QS-4) for LA-ICP-MS U–Pb isotopic dating. The CL images and structures of the zircons are shown in Figures 7A, B. The U–Pb isotopic dating results are presented in Supplementary Table S2, and the concordia diagrams and zircon chondrite-normalized REE patterns are shown in Figure 8.

Zircon crystals from the sample XW-9 are mostly euhedral, prismatic, with lengths ranging from 70 to 120 µm and length-to-width ratios of 1:1 to 3:1. The Th content ranges from 199 to 475 ppm, U content from 174 to 703 ppm, and Th/U ratios from 0.32 to 2.26, significantly higher than those of metamorphic and hydrothermal zircons (<0.1) (Belousova et al., 2002). The zircons exhibit clear oscillatory zoning and some have core structures, positive Ce anomalies, and heavy REE enrichment (Figure 8E), indicating typical magmatic origin. Data points with concordance less than 90% were excluded (point 10 and 24), while inherited zircon ages of 280 Ma (point 5), 277 Ma (point 9), 336 Ma (point 15), 440 Ma (point 17), and 449 Ma (point 20) were identified as xenocrysts or captured zircons. The remaining 18 points, with concordance greater than 90% (Figure 8A), yielded a weighted mean age of (241.2 ± 1.8) Ma (MSWD = 0.77, n = 18) (Figure 8B), representing the crystallization age of the rock.

Zircon crystals from the sample 05QS-4 are mostly euhedral prismatic, with lengths ranging from 50 to 100 µm and length-to-width ratios of 1:1 to 2:1. The content of zircon ranges from 44 to 482 ppm, U content from 141 to 1,511 ppm, and Th/U ratios from 0.22 to 1.59, showing geochemical characteristics which are similar to sample XW-9 (Figure 8F), indicating magmatic origin. Data points with concordance less than 90% were excluded (point 9 and 25), while inherited zircon ages of 408 Ma (point 6) were identified as xenocrysts or captured zircons, 223 Ma and 230 Ma (point 10 and 22) are relatively younger. The remaining 20 points, with concordance greater than 90% (Figure 8C), yielded a weighted mean age of (238.3 ± 1.7) Ma (MSWD = 0.21, n = 20) (Figure 8D), representing the crystallization age of the rock.

The zircon Hf isotopic data for the Xiekeng pluton are presented in Supplementary Table S3. Given the high closure temperature of the zircon Lu–Hf isotopic system, these data provide important constraints on the genesis and evolution of the zircons (Scherer et al., 2000). Lu–Hf isotope analysis was performed on the basis of zircon U–Pb laser denudation points. The results show ¹⁷⁶Hf/¹⁷⁷Hf values ranging from 0.00045 to 0.00230 (<0.02), indicating minimal radiogenic Hf accumulation, with initial values representing the present Hf isotopic composition. Zircon crystals from the sample XW-9 have ¹⁷⁶Hf/¹⁷⁷Hf values ranging from 0.28224 to 0.28236, with ε_Hf(t) values of −13.36 to −9.12 and two-stage model ages (T_DM2) of 1.86–2.13 Ga. Zircon crystals from the sample 05QS-4 have ¹⁷⁶Hf/¹⁷⁷Hf values ranging from 0.28220 to 0.28236, with ε_Hf(t) values of −14.93 to −9.29 and two-stage model ages (T_DM2) of 1.86–2.20 Ga. The ε_Hf(t) values of Xiekeng pluton are consistent with the Indosinian granites in south Gannan. In the Age vs. ε_Hf(t) (Figure 9) diagram, the samples plot above the lower crust evolution trend, indicating a crustal origin for the granitoid magmas.

The Sr–Nd isotopic data for the Xiekeng pluton are shown in Supplementary Table S4. The samples have high (⁸⁷Sr/⁸⁶Sr)_i values (0.720302–0.726407) and low ε _Nd(t) values (−11.2 to −10.9), with two-stage model ages (T_DM2) ranging from 1.90 to 1.93 Ga. In the (⁸⁷Sr/⁸⁶Sr)_i vs. ε_Nd(t) diagram (Figure 10A), the samples plot in the Mayuan Group area, in the T_DM2 vs. ε_Nd(t) diagram (Figure 10B), they plot in the Mesoproterozoic and Paleoproterozoic sedimentary rock areas, and in the age vs. ε_Nd(t) diagram (Figure 10C), they plot in the Mayuan Group area. The ε_Nd(t) values are consistent with those of Indosinian granites in South China (Figure 10D). It can be inferred that the source rock of the Xiekeng pluton is derived from metamorphosed sedimentary rocks.

Muscovite electron microprobe analysis results are shown in Supplementary Appendix S5. The muscovite has SiO₂ contents of 46.02–49.04 wt%, TiO₂ contents of 0.47–0.98 wt%, Al₂O₃ contents of 28.40–32.46 wt%, FeOt contents of 2.76–6.90 wt%, MgO contents of 0.72–1.86 wt%, Na₂O contents of 2.76–6.90 wt%, and K₂O contents of 9.39–10.18 wt%. The calculated cation numbers are Si⁴⁺ 6.33–6.58, Ti⁴⁺ 0.05–0.10, Al^Ⅳ 1.42–1.67, Al^Ⅵ 3.13–3.47, Fe²⁺ 0.31–0.80, Mg²⁺ 0.15–0.37, Na⁺ 0.05–0.17, classifying them as highly siliceous muscovite (Figures 11A,B). The average crystal chemical formula for muscovite in the Xiekeng pluton is K_1.69Na_0.10Fe_0.52Mg_0.21Al_3.31 [Al_1.53Si_6.47]O₂₀(OH)₄, indicating it is not an ideal pure muscovite KAl₂ [AlSi₃O₁₀](OH)₂, but rather an intermediate product of an isomorphous series with paragonite NaAl₂ [AlSi₃O₁₀](OH)₂ and phengite KAl(Fe, Mg) (Si₄O₁₀) (OH)₂)₇ (Figure 11C).

Under the microscope, the muscovite grains are large with clear edges, showing good crystal habits (Miller et al., 1981; Speer, 1984), indicating primary muscovite characteristics. The muscovite in the peraluminous granite coexisting with biotite, which shows replacement textures, does not significantly differ in composition from primary muscovite (Tao et al., 2014). Generally, primary muscovite has higher TiO₂ contents (>0.4%) (Miller et al., 1981), and the TiO₂ content of muscovite in the Xiekeng biotite monzogranite ranges from 0.47 to 0.96 wt%, indicating primary muscovite characteristics. In the Mg²⁺-Na⁺-Ti⁴⁺ diagram (Figure 11D), the samples plot within or near the primary muscovite field.

The specific formation age of the Xiekeng pluton has not been previously reported. In this study, the LA-ICP-MS zircon U–Pb ages of the two-mica monzogranite samples were determined to be 241.2 ± 1.8 Ma and 238.3 ± 1.7 Ma, indicating an Middle Triassic magmatic event, representing typical early Indosinian magmatism in South China, earlier than the Indosinian period granites that have been discovered in the area (Guo et al., 2011; Wang, 2015; He et al., 2017; Li J. et al., 2021). The presence of inherited zircon ages of 277 Ma, 280 Ma, and 336 Ma suggests Hercynian thermal events in the study area, while inherited zircon ages of 456 Ma and 467 Ma reflect the influence of the Caledonian tectonic activities in southern Jiangxi. Indosinian granites in South China can be divided into two stages: the early Indosinian (228–243 Ma, peak around 236 Ma) and the late Indosinian (214–220 Ma, peak around 214–218 Ma). This indicates that South China experienced two significant tectono-magmatic events during the Indosinian orogeny (Wang et al., 2007). Statistical analysis of the ages of Triassic granites in South China shows peak ages at 220 Ma and 236 Ma, consistent with the above stages (Figure 12).

Granites are generally classified into I-type, S-type, A-type, and M-type based on their genesis (Chappell and White, 1974; Qiu et al., 2008). Different genetic types of granites usually have distinct magma sources and tectonic settings. M-type granites are derived from the differentiation of basaltic magmas and are rare in nature. I-type granites are characterized by metaluminous to slightly peraluminous compositions, derived from igneous sources, and typically contain hornblende. They have relatively low initial Sr isotope ratios (<0.708), FeOt generally less than 1, and P₂O₅ negatively correlates with SiO₂. A-type granites have low oxygen fugacity, low water content (<2%), low P₂O₅ content (Chappell, 1999; Wang et al., 2000), and are primarily formed in extensional settings, containing alkaline dark minerals (e.g., aegirine, riebeckite), high melting temperatures (900°C), high 10000Ga/Al (>2.6), and high (Zr+Nb+Ce+Y) >350 ppm with high FeOt/(FeOt+MgO) ratios. S-type granites are rich in primary garnet, cordierite, muscovite, and have high A/CNK (>1.1) and isotopic signatures indicating derivation from metasedimentary sources.

The Early Indosinian (243–228 Ma) and Late Indochina (243–228 Ma) granites in SCB exhibit a high SiO₂ content ranging from 65 to 79 wt% and K₂O/Na₂O ratios between 1.03 and 3.89, accompanied by A/CNK ratios of 1.0–1.45. These granites are predominantly classified as high-potassium calc-alkaline peraluminous types. In terms of rock classification, I- and S-type granites were primarily distributed during the early Indochina period, whereas I-, S-, and A-type granites were prevalent during the late Indochina period (Xu et al., 2003; Ding et al., 2006; Dong et al., 2010; He et al., 2010; Guo et al., 2011; Sun et al., 2011; Mao et al., 2013a; Ren et al., 2013; Zhao et al., 2013a, b; Jiao et al., 2015; Wang, 2015; Gao, 2016; Wang H. Z. et al., 2018; Wang W. B. et al., 2018; Qing et al., 2020; He, 2021; Han, 2023; Sun, 2023). Regional Early Indosinian granites are mainly distributed in Hainan and Guangxi, while Late Indosinian granites are mainly distributed in Jiangxi, Hunan and surrounding areas (Figure 1B). The Xiekeng pluton primarily comprises quartz, plagioclase, potassic feldspar, muscovite, and biotite, with geochemical characteristics of high A/CNK values (1.11–1.27), indicating strongly peraluminous S-type granite (Figure 4C). The Xiekeng pluton has high 10000Ga/Al ratios (2.83–3.23) and enrichment in high field strength elements such as Th, U, Ta, and Gd, which shows similarities with A-type granites (Li X. W. et al., 2010). With increasing research on granites, it has been found that highly differentiated I- and S-type granites can also exhibit characteristics similar to A-type granites, such as 10000Ga/Al > 2.6 and enrichment in high field strength elements (Li and Li, 2007; Wu et al., 2017). As the degree of granitic magma differentiation increases, elements like Li and Rb increase, while Cr, Ni, Sr, and Ba decrease in the residual melt (Lee and Morton, 2015), resulting in lower K/Rb, Zr/Hf, and Nb/Ta ratios (Ballouard et al., 2016). Therefore, K/Rb, Zr/Hf, and Nb/Ta ratios are important indicators of the degree of magmatic differentiation. The K/Rb ratios of the Xiekeng pluton range from 8.40 to 11.04, Nb/Ta ratios from 2.38 to 9.24, and Zr/Hf ratios from 25.42 to 35.37, indicating significant magmatic differentiation and classifying the pluton as highly fractionated granite. In the 10000Ga/Al vs. Zr and (Zr+Nb+Ce+Y) vs. (Na₂O+K₂O)/CaO diagrams (Figures 13A,B), the samples plot within the highly fractionated granite field, indicating they are highly fractionated I- or S-type granites. The P₂O₅ content (0.14–0.18 wt.%) does not show significant variation with SiO₂, consistent with the evolution trend of S-type granites (Chappell, 1999). In the Rb vs. Th variation diagram (Figure 13C), the samples exhibit trends consistent with S-type granites. The presence of muscovite, a characteristic mineral of S-type granites, and the A-C-F diagram (Figure 13D) positioning of the samples within the S-type granite field further support this classification. Whole-rock zircon saturation temperatures range from 780°C to 806°C (avg. 791°C), consistent with S-type granite zircon saturation temperatures (Watson, 1979; Zhao et al., 2013b). Thus, the Xiekeng pluton is interpreted as a highly fractionated S-type granite.

Highly fractionated S-type granites are characterized by the absence of the typical pressure indicator mineral amphibole (Wang et al., 2007), which complicates efforts to constrain the depth of intrusion. Primary muscovite in peraluminous granites can be used as a pressure indicator to calculate magmatic emplacement pressures, and subsequently, the depth of emplacement (Tao et al., 2014). The muscovite barometer, first proposed by Velde (1965) and later modified by Anderson (1996) based on Si content (calculated with 11 oxygen atoms), is used here: P (kbar) = −2.6786Si²⁺ 43.975Si⁴⁺ 0.01253T (°C) – 113.9995. Using a solid-liquid line temperature of 650°C for calculations, the results (Supplementary Table S5) show that the formation pressures of the Xiekeng two-mica monzogranite range from 6.48 to 9.89 kbar (avg. 8.35 kbar). The emplacement depth, calculated using P = ρ·g·H (ρ = 2.7×10³ kg/m³, g = 9.8 m/s²), ranges from 24 to 36 km (avg. 30 km). The presence of large, clean, and well-formed primary muscovite grains, similar in size to quartz, potassic feldspar, and plagioclase, indicates crystallization under high-pressure conditions.

Petrographic and geochemical characteristics suggest that the Xiekeng pluton is an S-type granite, derived from the melting of metasedimentary rocks (Chappell et al., 1987). The recent study conducted by Roberts et al. (2024) indicates that the zircons from S-type granites in the Himalayas exhibit the lowest Ce/U and Th/U ratios, which overlap with those observed in metamorphic zircons. This finding further supports that S-type granites are derived from metamorphic rocks. Additionally, the trace element compositions of zircons within the Xiekeng intrusion closely resemble those characteristic of Phanerozoic strongly peraluminous granites (Figure 14A). In the Log (Ce/U) vs. Log (Th/U) diagram, a significant majority of data points fall within the ‘hybrid’ S-type granite region, suggesting that the origin of the Xiekeng pluton is linked to metamorphic rock sources. The CaO/Na₂O ratio can help determine the source rock composition: CaO/Na₂O > 0.3 indicates a source rock rich in plagioclase and poor in clay minerals (sandy rocks), while CaO/Na₂O < 0.3 indicates a source rock poor in plagioclase and rich in clay minerals (muddy rocks) (Sylvester, 1998). The CaO/Na₂O ratios of the Xiekeng pluton range from 0.14 to 0.36 (Figure 14C), with three samples having CaO/Na₂O > 0.3 (sandy rocks) and six samples having CaO/Na₂O < 0.3 (muddy rocks). The Rb/Sr ratios range from 5.81 to 23.3, and Rb/Ba ratios from 1.58 to 9.71. In the Rb/Sr vs. Rb/Ba diagram, all the samples fall within the shale area (Figure 14D). However, due to the high degree of magmatic differentiation and subsequent late hydrothermal processes, elements such as K, Na, Ca, Rb, Sr, and Ba are particularly susceptible to migration throughout the rock-forming process. Therefore, the Al₂O₃/TiO₂ vs. CaO/Na₂O and Rb/Sr vs. Rb/Ba diagrams may not be accurate to distinguish the material composition of source rocks. Most scholars believe that the Indosinian highly differentiated S-type granites in South China originated from metamorphic mudstones (Wang et al., 2007; Mao et al., 2013b; Chen et al., 2007; Gao et al., 2018; Du et al., 2022). In the Eu/Eu* vs. (La/Yb)_N diagram (Figure 14E), they fall within the crustal source range, Mg^# values range from 13 to 35 (avg = 28 < 40), and the samples fall within the pure crustal range, indicating no mantle source involvement (Figure 14F) (Rapp and Watson, 1995). Zircon Lu–Hf isotopes and whole-rock Sr–Nd isotopes indicate the granites primarily originated from the deep melting or re-melting of ancient crustal materials (Zhu et al., 2009). Thus, we conclude that the source rock of the Xiekeng pluton is likely composed of crustal clay-rich mudstones.

Trace elements are crucial for determining granite source evolution (Zhao and Zhou, 1997). The Nb/Ta ratios of the Xiekeng pluton range from 2.38 to 9.24 (avg. 7.15), significantly lower than those of chondrites (19.9) and continental crust (13.4) (Taylor and Mc Lennan, 1995), indicating Nb depletion. The Zr/Hf ratios range from 25.42 to 35.37 (avg. 31.79), lower than those of chondrites (34.3) and continental crust (36.7) (Figure 15A) (Chen and Yang, 2015), indicating significant magmatic differentiation. The Rr/Sr ratios range from 5.81 to 23.3 (avg. 9.15), significantly higher than the average global upper crustal values of 0.31 (Gao et al., 1999) for eastern China and 0.32 (Taylor and Mc Lennan, 1995) globally. The Rb/Nb ratios range from 9.22 to 14.66 (avg. 12.19), higher than the global upper crustal Rb/Nb ratio of 4.5 (Taylor and Mc Lennan, 1995). The Nb/U ratios range from 1.55 to 11.09 (avg. 6.44), and Ce/Pb ratios from 1.90 to 4.03 (avg. 2.66), similar to continental crust values (Nb/U = 9, Ce/Pb = 4), suggesting a source rock of mature continental crust. The presence of trapped zircons in the magma also indicates that the Xiekeng pluton originate from the melting of ancient crust. The Zr and Zr/Nb ratios show little variation (Figure 15B), indicating magmatic differentiation influence. The positive correlation between Ba and Sr suggests potassic feldspar crystallization (Figure 15C), while the negative correlation between Sr and Rb/Sr indicates plagioclase fractionation (Figure 15D). The negative correlation between TiO₂ and P₂O₅ suggests the crystallization of apatite and ilmenite (Figure 15E). In the SiO₂ vs. FeOt/(FeOt+MgO) diagram (Figure 15F), the samples show iron enrichment characteristics.

The tectonic evolution of South China during the Indosinian period is highly complex, with significant debate regarding the dynamics of magmatic activities. Various models have been proposed, including: (1) Early Mesozoic intracontinental orogeny (Hsü et al., 1988, 1990), (2) intracontinental crustal stacking and thickening (Wang et al., 2002; Wang Y. J. et al., 2005), (3) subduction of the Paleo-Pacific Block (Charvet et al., 1994; Gilder et al., 1996; Wang Q. et al., 2005; Li and Li, 2007), and (4) lithospheric extension and mafic magma underplating (Guo et al., 1998). While these models partially explain the genesis of granites, they do not fully account for the dynamic background of Indosinian magmatism in South China.

Increasing evidence suggests that the Mesozoic South China Block was subjected to multi-directional compression and crustal thickening (Zhou, 2003; Sun et al., 2005), followed by lithospheric extension and thinning (Guo et al., 1997; Jin et al., 2017). The Triassic tectonic evolution of South China can be divided into two stages: compressional deformation during the Early and Middle Triassic (ages mainly concentrated at 228–243 Ma) and extensional deformation during the Late Triassic (ages mainly concentrated at 206–220 Ma) (Zhang et al., 2012), corresponding to the early and late Indosinian stages, respectively (Wang et al., 2007). The majority of early granites are classified as strongly peraluminous S-type granites, which are thought to have formed in a collisional tectonic setting (Yu et al., 2007; Zhao et al., 2013a; Zhao et al., 2013b). In contrast, the late granites may have developed under post-collisional granite tectonic mechanisms (Zhou et al., 2006; Wang et al., 2007; Mao et al., 2011). From 267 to 262 Ma, the Indosinian orogeny began, with the Indochina Block moving northwards and colliding with the South China Block, reaching its zenith during 258–243 Ma, the Paleo-Tethys Ocean closed (Carter et al., 2001), and the South China Block experienced multi-directional compression from surrounding blocks (Mao et al., 2014), leading to crustal thickening to ≤50 km and metamorphism near the suture zone and magmatic activity at around 240 Ma. During 240–225 Ma, the South China Block collided with the North China Block, affecting the entire region and causing widespread folding and thrusting of Devonian to Middle Triassic sedimentary covers, followed by crustal extension (Chen, 1999). Within 10 to 20 Ma following the collision between the Indochina Block and the South China Block, it is likely that the southeastern coastal region of South China entered a phase of stress relaxation. This observation aligns with previous studies indicating that the thickened crust experienced thermal-stress relaxation within a similar timeframe of 10–20 million years (Patinño Douce et al., 1990; Sylvester, 1998), resulting in partial melting of the middle-lower crust and forming early-stage peraluminous granites (Zhou, 2003). Thermal-stress concentration during thickening preferentially released, causing small-scale mafic magma underplating at the base of the thickened crust. This underplated mafic magma, through thermal conduction with surrounding rocks, reached thermal equilibrium within 5–20 Ma, and under the background of further lithospheric thinning, induced large-scale crustal re-melting, forming late-stage granites.

During the Early Mesozoic, the South China Block underwent significant Indosinian tectonic events, colliding and amalgamating with the Sibumasu Block along the Red River south of the Jinsha-Mojiang-Songma line. Subsequently, the South China Block collided with the North China Block, forming the Qinling-Dabie orogenic belt (Figure 17A). Yu et al. (2007) identified two distinct metamorphic-magmatic events (249–225 Ma and 225–207 Ma) slightly lagging behind the two Block collision events, reflecting the large-scale metamorphic orogenic response to the Indosinian tectonic events. Qing et al. (2020) proposed that South China was in a syn-collisional stage from 258 to 231 Ma, transitioning to a post-collisional stage after 231 Ma. The statistics of Indosinian granite (232–253 Ma) in South China Block also support the above views (Figure 16). The Xiekeng pluton formed around 240 Ma, during the syn-collisional stage, and in the (Y+Ta) vs. Rb and R/30-Hf-3Ta geochemical tectonic discrimination diagrams (Figure 16), the samples plot in the syn-collisional field, indicating formation under compressional tectonics. The presence of biotite-quartz schist xenoliths and stress-induced quartz undulose extinction and muscovite bending further support a regional compressional strain regime. We anticipate that intraplating of basaltic magmas would trigger deep (ca.30 km depth) dehydration melting of the early Paleozoic granitoids, forming the early Indosinian Xiekeng S-type granites (Figure 17B).

It is traditionally believed that the parental rocks for ion-adsorption REE deposits formed mainly during the Yanshanian period (150–190 Ma). However, recent discoveries have identified Indosinian parental rocks for REE deposits, such as the Jiaping deposit in Guangxi (239–254 Ma), the Nan’an-Renhe pluton in Guangxi (260 Ma) (Fu et al., 2022), the Qinfang pluton in Guangxi (250 Ma), the Qingxi pluton in southern Jiangxi (220–225 Ma) (Yu et al., 2012; Wang, 2015; He et al., 2017), the Wuliting granite in southern Jiangxi (237 Ma) (Qiu et al., 2004), and the Lincang granite in Yunan (218–242 Ma) (Lu et al., 2019; Lan et al., 2021). The threshold for ion-adsorption REE deposits is 150 ppm, and higher REE contents in the parental rocks favor mineralization (Zhang et al., 2024). The Xiekeng pluton has total REE contents ranging from 224.71 to 353.12 ppm, significantly exceeding this threshold, with geochemical characteristics similar to those of typical Indosinian ion-adsorption REE parental rocks (Supplementary Appendix S6). Approximately 10 km north of the Xiekeng pluton, the Shitouping ion-adsorption REE deposit has been identified. The primary ore-forming parental rock is characterized as high-silicon and high-potassium calc-alkaline granite. The average thickness of the weathering crust measures 9.52 m, with an average grade of SRE₂O₃ at 0.088%. The predominant rare earth minerals include bastnäsite-(Ce), synchysite-(Ce), synchysite-(Y), columbite, xenotime, and other rare earth-rich minerals (Cao et al., 2024). During the weathering process, easily weathered rare earth minerals undergo decomposition and release REE³⁺, while clay minerals adsorb REE³⁺. The pH value exhibits a gradual increase from the surface soil layer to the semi-weathered layer. In the upper section of the weathering crust, pH levels are acidic; H⁺ desorb a portion of REE³⁺, which subsequently migrates to lower regions of the weathering crust along with water flow. As pH increases, clay mineral adsorption of REE³⁺ intensifies while its migration capacity diminishes-ultimately leading to precipitation and enrichment in deeper layers of the weathering crust (Zhu et al., 2022). Due to their stronger migration ability compared to LREE³⁺, HREE³⁺ tends to be preferentially adsorbed by clay minerals when pH approaches neutrality. This phenomenon results in HREE³⁺ being enriched in deeper strata and contributes to fractionation between light and heavy rare earth elements. Consequently, a “light on top and heavy at the bottom” distribution pattern is observed for rare earth elements within the profile of the weathering crust (Li and Zhou, 2020). A statistical analysis of the parental rocks associated with typical ion-adsorption REE deposits in South China indicates that these deposits are primarily derived from partial melting of metamorphosed conglomeratic sandstone and metamorphosed mudstone (Yang et al., 2024). The Xiekeng pluton is characterized as a high-potassium calc-alkaline to shoshonitic granite, exhibiting significant differentiation along with a high initial strontium ratio (I_Sr) value ranging from 0.720302 to 0.726407, low ε_Nd(t) values (−11.2 to −10.90), and a two-stage model age (T_DM2) estimated at 1.90 to 1.93 Ga. The source rocks is mainly composed of sedimentary rocks. During the formation process, it is conducive to the crystallization and precipitation of rare earth minerals such as apatite, monazite, xenotime, fluocerite, and bastnäsite (Figure 3), and the rich rare earth mineral types are similar to those of the Shitouping deposit. The region in southern Jiangxi is characterized by low hills with minimal elevation differences, conducive to forming complete weathering crusts, adequate rainfall and humid climate (temperatures 18°C–21°C, rainfall 1,500–2000 mm/year), well-developed vegetation preventing erosion of the weathering crust (Figure 2C), and weakly acidic to neutral pH values that favor REE enrichment (Wang D. H. et al., 2013). These factors provide excellent conditions for forming ion-adsorption REE deposits. Therefore, the Xiekeng pluton has significant potential for ion-adsorption type REE mineralization, and further investigation is necessary.