Although pegmatites are volumetrically minor in the upper continental crust, these rocks host abundant rare metal deposits (e.g., Li, Be, Rb, Ta, and Nb). Pegmatites can be formed either by extensive fractional crystallization of granitic magmas or by low-degree partial melting of metasedimentary rocks. The Mao’ershan–Yuechengling composite batholith in the Nanling Range in the South China Block (SCB) is of early Paleozoic–Triassic age (440–381 and 236–204 Ma, respectively). Recently, hundreds of pegmatites associated with Nb, Ta, Be, Rb, and Li mineralization have been identified in this batholith. These pegmatites are hosted by granitic wall rocks. However, the relationships between the pegmatites and granitic wall rocks are not well constrained. To address this, we investigated the Mao’antang (MAT) and Tongzuo (TZ) pegmatites and their biotite granite wall rocks in the middle part of the Yuechengling pluton. Laser ablation inductively coupled plasma mass spectrometry zircon U–Pb ages revealed that the MAT pegmatites formed during the Permian (269 Ma) and Triassic (231 Ma) and that the MAT biotite granite wall rock records two stages of magmatic activity (271 and 231 Ma) that are coeval with the pegmatites. The TZ pegmatites probably formed during the Triassic (235 Ma), and the TZ biotite granite wall rock formed during the Silurian (435 Ma). The MAT biotite granite and pegmatites (εNdt=12.0 to −10.6; εHft=8.0 to −1.0), TZ pegmatites (εNdt=10.4 to −6.1), and TZ biotite granites (εNdt=9.1 to −8.7; εHft=7.7 to −4.1) have enriched whole-rock Nd and zircon Hf isotopic compositions that are similar to those of early Paleozoic and Triassic S-type granites in the SCB. In addition, the whole-rock Pb isotopic compositions of the MAT and TZ pegmatites and granites are distributed along the upper crust evolution line. We suggest that the MAT and TZ biotite granites were mainly derived from Paleoproterozoic metasedimentary rocks in the middle crust. The MAT pegmatites are fractional crystallization products of the MAT biotite granites, whereas the TZ pegmatites were formed by fractional crystallization of hidden parental S-type granites. We propose that the MAT and TZ pegmatites have potential for rare metal (Nb, Ta, Be, and Li) mineralization, as they record high degrees of fractional crystallization. The MAT and TZ areas in the middle of the Yuechengling pluton are promising targets for rare metal exploration.

Strategic resources and critical materials are of significance to science, technology, and the economy and have the greatest risk of supply disruptions. These materials include Li, Be, Ta, Nb, and rare earth elements [1, 2]. Pegmatites are an important source of rare metal elements (Li, Be, Ta, and Nb), and they are texturally complex igneous rocks marked by some combination of coarse and variable crystal sizes, mineralogical zonation, prominent anisotropy of crystal orientations from their margins inward, and crystals with skeletal, radial, and graphic-intergrowth habits [3]. Pegmatites commonly have a granitic composition with high contents of fluxing components such as H2O, F, B, P, and alkalis, and they are therefore generally considered to represent the final products of extreme differentiation of an evolving granite system (e.g., [3]). However, in the field, there are occasions when no granite plutons occur near pegmatites, as observed for the Qinghe pegmatites in the Altai Orogen (NW China; [4]). An alternative model suggests that pegmatites may be formed as result of thrusting-enhanced low degrees of melting of metapelitic rocks at relatively low temperatures (amphibolite-facies conditions in the middle crust) during compressional deformation in an orogenic belt [46]. Studies of the petrogenesis of rare metal pegmatites and granitic wall rocks have economic significance, as the results serve as a guide during prospecting for rare metal ore deposits [7].

The Nanling Range covers six provinces (Guizhou, Guangxi, Hunan, Guangdong, Jiangxi, and Fujian) in the southern part of South China, and it is endowed with various W, Sn, Bi, and Sb deposits [8]. Recent geological surveys in the western Nanling Range have identified several hundred pegmatites in the Mao’ershan–Yuechengling composite batholith. These pegmatites are associated with Nb, Ta, Be, Li, and Rb mineralization [9]. However, the ages of these pegmatites have not been well constrained, and the genetic relationships between the pegmatites and their granitic wall rocks remain unclear. In this paper, we present the results of a detailed petrographic, geochronological, and geochemical study (including whole-rock Sr–Nd–Pb and zircon Hf isotopic data) of the Mao’antang (MAT) and Tongzuo (TZ) pegmatites and their biotite granite wall rocks in the middle part of the Yuechengling pluton. Our results enable us to constrain the petrogenesis of the pegmatites and biotite granites and to provide important insights into the relationships between magma generation and rare metal mineralization in the Nanling Range.

The South China Block (SCB) comprises the Yangtze Block to the northwest and the Cathaysia Block to the southeast (Figure 1(a)), which were amalgamated during a Neoproterozoic continent–continent collisional orogeny (i.e., the Jiangnan Orogen; [10] and references therein). The basement of the Cathaysia Block consists of Paleoproterozoic–Mesoproterozoic–Neoproterozoic metasedimentary rocks that include two-mica schists, mica–quartz schists, and biotite–plagioclase granulites and metaigneous rocks (amphibolites; [1113]). The basement of the eastern Yangtze Block is the Mesoproterozoic Lengjiaxi Group, which consists of greenschist- to amphibolite-facies meta-tholeiitic basalts and basaltic andesites with subordinate pelites [14]. The metamorphic basement of the SCB was overlain by late Neoproterozoic to early Paleozoic Nanhua rift sediments [15]. An early Paleozoic intracontinental orogeny closed the Nanhua rift, resulting in deformation and metamorphism of the pre-Devonian rocks [16]. In response to the early Paleozoic intracontinental orogeny, large amounts of granites and minor mafic rocks were formed in the Wuyi, Yunkai, Wugong, and Baiyun domains ([10, 16, 17]; Figure 1(b)). During the Middle or Late Devonian to Permian, the SCB evolved into a stable littoral–neritic depositional environment where Brachiopoda-, coral-, and fusulina-bearing carbonates (limestones and dolomites), black cherts, and minor sandstones and mudstones accumulated ([15] and references therein). During the Middle–Late Triassic, a series of thrust and strike-slip faults, ductile shear zones, and widespread granites developed in the SCB ([18] and references therein). Some previous studies have suggested that the Triassic orogeny resulted from the collision between the Indochina Block and SCB [1821]. Other studies have proposed that the Triassic orogeny and associated magmatism were related to flat slab subduction of the Paleo-Pacific Plate [22]. Moreover, Mao et al. [23] suggested that the Triassic orogeny and associated magmatism might have been caused by the collision between the Southeast Asian continental margin and Paleo-Pacific Plate. The tectonic setting of the SCB changed gradually from one of compression to one of extension during the late Mesozoic [21]. In the Jurassic and Cretaceous, the SCB was characterized by magmatic flare-ups, and 80% of the SCB igneous rocks (which have a total exposed area of >200,000 km2) were formed during this period [21, 24].

The Mao’ershan–Yuechengling granitic batholith lies near the Anhua–Luocheng Fault (Figure 1(b)). It extends northeast from north Guangxi Province to southern Hunan Province and has an outcrop area of ~3400 km2 (Figures 1(b) and 2). Most of the Mao’ershan–Yuechengling granites were formed during the early Paleozoic at 440–381 Ma [2830]. These early Paleozoic granitic rocks were emplaced into the core of the Mao’ershan–Yuechengling anticlinorium, which is made up of Neoproterozoic–Silurian slates, pelitic schists, felsic volcaniclastics, metasandstones, and marbles. In some outcrops, the Mao’ershan–Yuechengling granites are overlain by Middle Devonian sedimentary rocks ([30] and references therein). The early Paleozoic Mao’ershan–Yuechengling granites were intruded by younger Triassic granites with ages of 236–204 Ma ([31, 32] and references therein). The Mao’ershan–Yuechengling batholith is divided into the Mao’ershan and Yuechengling plutons by the Xinzi Fault. On the surface, the outcrops of these two plutons are separated by Devonian–Cretaceous sedimentary rocks, but at depth, they are connected ([29] and references therein).

The Mao’antang (MAT) and Tongzuo (TZ) pegmatites and their granitic wall rocks are located in the northeastern part of Ziyuan County in the middle of the Yuechengling pluton (Figure 2). The MAT pegmatites are gray in color and typically have massive structures and inequigranular textures. They consist of quartz (40–45 vol.%), K-feldspar (5–10 vol.%), albite (35–40 vol.%), tourmaline (<3 vol.%), and zinnwaldite (<3 vol.%) with minor garnet, muscovite, biotite, allanite, and monazite (Figures 3(a)–3(d)). The MAT pegmatites are mainly muscovite pegmatites, with subordinate biotite and two-mica pegmatites, which generally occur in fractures and cavities in the biotite granite wall rocks (Figure 3(a); Supplementary Figures 1a–1b). In addition, the MAT pegmatites are related to Nb, Ta, Be, and Li mineralization. They contain Nb–Ta–Be–Li oxide minerals, such as columbite-group minerals, microlite, wodginite, beryl, and spodumene ([9] and references therein). From the outer to inner zones of the MAT pegmatites, the microcline and muscovite contents decrease, whereas the albite contents increase. As such, the Nb, Ta, and Be contents are gradually enriched from the northwest (outer zone) to southeast (inner zone) ([9] and references therein). The wall rocks of the MAT pegmatites are massive biotite granites that consist of quartz, K-feldspar, albite, biotite, and minor muscovite (Figure 3(a); Supplementary Figures 1a–1b).

The TZ pegmatites are gray in color and typically have massive structures and inequigranular–porphyritic textures. They consist of quartz (30–35 vol.%), K-feldspar (25–30 vol.%), albite (15–20 vol.%), garnet (5–10 vol.%), and muscovite (5 vol.%) with minor monazite and chlorite (Figures 3(f)–3(h)). From the outer to inner zones, the TZ pegmatites change gradually from coarse-grained albite pegmatite to medium-grained graphic albite pegmatite, medium–fine-grained albite pegmatite, and quartz pegmatite (Figure 3(f); Supplementary Figures 1c–1d). The TZ pegmatites are also related to Nb, Ta, and Be mineralization and contain columbite-group minerals, microlite, wodginite, and beryl ([9] and references therein). The wall rocks of the Tongzuo pegmatites are gneissic biotite granites composed of quartz, plagioclase, biotite, and minor muscovite (Figure 3(e); Supplementary Figures 1e–1f).

Zircon U–Pb dating and Lu–Hf isotope analyses, whole-rock major and trace element geochemical analyses, and Sr–Nd–Pb isotope analyses were carried out at the Guangxi Key Laboratory of Hidden Metallic Ore Deposits Exploration, Guilin University of Technology, China.

3.1. Zircon U–Pb Geochronology

Zircon grains were separated from ~2 kg rock samples using conventional heavy liquid and magnetic separation techniques. Zircon grains were handpicked and mounted on an epoxy resin disc and polished prior to analysis with a laser ablation inductively coupled plasma mass spectrometer (LA–ICP–MS). Cathodoluminescence (CL) imaging of the zircon grains was performed at the Chongqing Yujin Technology Co., Ltd., China. The U–Pb isotopic compositions of zircon grains from the MAT and TZ pegmatites and granite wall rocks were analyzed using an Agilent 7500 LA–ICP–MS. Laser ablation was undertaken at a constant energy of 80 mJ, a repetition rate of 6 Hz, and a spot size of 32 μm. Helium was used to carry ablated material to the ICP–MS. Element corrections were determined relative to the standard glass NIST 610 [33]. During our analytical work, the zircon standard Plešovice yielded a weighted-mean 206Pb/238U age of 337.1±0.6 Ma (2σ; MSWD=0.10; n=52), which is within error of the recommended value of 337.1±0.4 Ma [34]. The age calculations were undertaken using ICPMSDataCal (version 8.4; [35]), and the concordia plots were made using Isoplot 3.75 [36].

3.2. Major and Trace Element Analyses

Rock samples were first examined under an optical microscope. Selected whole-rock samples were broken into small chips and cleaned ultrasonically in 4 N hydrochloric acid for 30 min to remove any altered material and then washed twice with distilled water. The small rock chips were dried and handpicked to remove visible contamination. The powder was used for analyses of major and trace elements as well as Sr–Nd–Pb isotopes. Major elements were analyzed on fused glass beads using a ZSX Primus II X-ray fluorescence (XRF) spectrometer with analyticaluncertainties<5% ([37] and references therein). Trace elements were analyzed with Agilent 7500cx ICP–MS equipment, following the procedures described by Liu et al. [37]. Analytical precisions for most elements were better than ±2–5% (relative standard deviation, RSD). Standardization was achieved using the United States Geological Survey (USGS) standards BHVO, AGV, W-2, and G-2 and national rock standards GSR-1, GSR-2, and GSR-3 [38].

3.3. Sr–Nd–Pb Isotope Analyses

Whole-rock Sr, Nd, and Pb isotope analyses were conducted using a Neptune plus multicollector- (MC-) ICP–MS. The details of the analytical procedures used have been described by Liu et al. [37]. The measured average 87Sr/86Sr ratio of the NBS987 standard was 0.710281±16 (2σ; n=11), and the measured average 143Nd/144Nd ratio of the Shin Etsu JNdi-1 standard was 0.512092±8 (2σ; n=2), which are in good agreement with the recommended values of 0.710248±12 and 0.512107±12, respectively [39]. All measured 87Sr/86Sr and 143Nd/144Nd ratios were fractionation-corrected to 86Sr/88Sr=0.1194 and 146Nd/144Nd=0.7219, respectively. Four repeat analyses of the SRM981 Pb standard yielded mean values of 206Pb/204Pb=16.933±0.0008, 207Pb/204Pb=15.485±0.0008, and 208Pb/204Pb=36.676±0.0019, which agree with the values reported by Todt et al. [40].

3.4. In Situ Zircon Lu–Hf Isotope Analyses

In situ zircon Lu–Hf isotope measurements were undertaken using a Neptune MC-ICP–MS with a beam size of 44 μm and a laser frequency of 10 Hz. Details of the instrumental conditions and data acquisition procedures have been described by Wu et al. [41]. The isobaric interference of 176Lu on 176Hf was negligible, due to the extremely low value of 176Lu/177Hf in the zircon (normally <0.002). During our analyses, the GJ-1 reference zircon was analyzed as an unknown sample, and it yielded a weighted-mean value of 176Hf/177Hf=0.281989±0.000034 (2σ; n=22), which is in good agreement with the recommended value of 0.282000±0.000005 [42].

4.1. Zircon U–Pb Ages

The results of the zircon U–Pb isotope analyses are given in Supplementary Table 1. Five rock samples, including one MAT biotite granite wall rock, two MAT pegmatites, one TZ biotite granite wall rock, and one TZ pegmatite, were selected for the zircon U–Pb dating. Zircon crystals in these samples have lengths of 150–350 μm and length/width ratios of 1 : 1–3 : 1 (Figures 4(a)–4(e)). Zircon grains in the MAT biotite granite wall rock and pegmatites and in the TZ biotite granite wall rock all exhibit well-developed oscillatory zoning (Figures 4(a)–4(d)) and high Th/U ratios (0.15–1.88; Supplementary Table 1), which indicate a magmatic origin [43]. However, zircon grains in the TZ pegmatite TZ-09 show dark and unclear CL images (Figure 4(e)), which suggest that sample TZ-09 underwent late-stage alteration by hydrothermal fluids.

Zircons from MAT wall rock sample MAT-01 yielded two weighted-mean 206Pb/238U ages of 230.9±1.5 Ma (2σ; MSWD=0.79; n=13; Figure 5(a)) and 271.0±2.0 Ma (2σ; MSWD=0.37; n=12; Figure 5(a)), which indicate that this wall rock records two magmatic events. The zircons from MAT pegmatite samples MAT-03 and MAT-04 yielded weighted-mean 206Pb/238U ages of 269.1±1.2 Ma (2σ; MSWD=1.12; n=19; Figure 5(b)) and 230.5±1.5 Ma (2σ; MSWD=0.23; n=12; Figure 5(c)), respectively. Zircons from the TZ biotite granite wall rock yielded a weighted-mean 206Pb/238U age of 434.7±1.4 Ma (2σ; MSWD=0.57; n=45; Figure 5(d)). Five zircon grains from TZ pegmatite sample TZ-09 yielded a weighted-mean 206Pb/238U age of 234.7±4.8 Ma (2σ; MSWD=1.4; n=5; Figure 5(e)), which probably represents the crystallization age of the TZ pegmatite, while another six zircon grains from the same sample (TZ-09) yielded a weighted-mean 206Pb/238U age of 261.0±2.9 Ma (2σ; MSWD=0.05; n=6; Figure 5(e)), which we interpret to be an inherited age. To summarize, the MAT pegmatites formed during the Permian (269 Ma) and Triassic (231 Ma), the MAT biotite granite wall rocks record two stages of magmatic activity (271 and 231 Ma) that were coeval with the Permian and Triassic pegmatites, the TZ pegmatite probably formed during the Triassic (235 Ma), and the TZ biotite granite wall rock formed during the Silurian (435 Ma).

4.2. Major and Trace Element Geochemistry

4.2.1. MAT and TZ Pegmatites

Whole-rock major and trace element data for the MAT pegmatites and TZ pegmatites and their granite wall rocks are given in Supplementary Table 2. The MAT pegmatites have high contents of SiO2 (70.2–73.5 wt.%), K2O (4.2–5.5 wt.%), and Na2O (0.5–3.9 wt.%), with K2O>Na2O, and they plot in the high-K calc-alkalic to shoshonitic fields on the K2O versus SiO2 diagram (Figures 6(a), 6(b), and 6(d)). Their A/CNK (molar Al2O3/CaO+Na2O+K2O) values range from 1.14 to 2.68 and plot in the peraluminous field (Figure 6(c)). The TZ pegmatites have a wide range of SiO2 (45.0–98.7 wt.%), K2O (0.01–11.3 wt.%), and Na2O (0.04–5.65 wt.%) contents (Figures 6(a), 6(b), and 6(d)), and their A/CNK values range from 0.47 to 4.68, plotting in the metaluminous to peraluminous fields (Figure 6(c)). Three TZ pegmatite samples (TZ-07, TZ-08, and TZ-09) have low SiO2 contents (<61 wt.%; normalized on an anhydrous basis). We do not plot these three pegmatites in Figure 6, because they are cumulates of K-feldspar (SiO2~65 wt.%), albite (SiO2~60 wt.%), and muscovite (SiO2~46 wt.%) (Figured 3(f)–3(g); [26]).

The MAT pegmatites have a wide range of total rare earth element (ΣREE) contents (1.28–203 ppm), and they are characterized by moderately light REE-enriched (La/YbCN=2.836.0) chondrite-normalized patterns with significant negative Eu anomalies (Eu/Eu=EuCN/SmCN×GdCN1/2=0.110.30) (Figure 7(a)). In addition, the REE patterns of the MAT pegmatites have a gull-wing shape and show tetrad effects (Figure 7(a)). The TZ pegmatites also have a wide range of ΣREE contents (1.00–382 ppm). Their (La/Yb)CN values range from 0.02 to 10.8 (Figure 7(c)), and they are characterized by moderate to significant negative Eu anomalies (Eu/Eu=0.010.44; Figure 7(c)).

On primitive-mantle-normalized plots, the MAT and TZ pegmatites are enriched in Rb, Th, U, and Pb; are slightly to significantly depleted in Ba, Sr, Zr, and Ti; and show small negative Nb anomalies and slight Ta positive anomalies (Figure 7(b)). The MAT and TZ pegmatites have Nb/Ta ratios of 3.72 to 6.92 and 1.12 to 8.99 and Zr/Hf ratios of 5.34 to 33.8 and 0.80 to 34.5, respectively (Supplementary Table 2).

4.2.2. MAT and TZ Biotite Granite Wall Rocks

The MAT wall rock sample (MAT-01) has a relatively low SiO2 content of 62.8 wt.% (normalized on an anhydrous basis), which is similar to the diorite (Figure 6(a)). However, sample MAT-01 has obviously lower MgO (0.97 wt.%) and higher Rb (1026 ppm) contents than typical diorites in Guangzhou region of South China (MgO=2.862.90 wt.%; Rb=212217 ppm; [26]). When granitic rocks experience supergene weathering and hydrothermal alteration, their SiO2 contents decrease as biotite (SiO2~40 wt.%) is replaced by chlorite (SiO2~28 wt.%) [51]. This is consistent with the negative Ce anomaly of the MAT biotite granite (Figure 7(a)). Negative Ce anomalies in granites have been interpreted to be related to chemical weathering processes [52]. Therefore, we regard sample MAT-01 to be a biotite granite rather than a diorite. In addition, the MAT biotite granite wall rock has a K2O content of 5.83 wt.% and plots in the shoshonitic field on the K2O versus SiO2 diagram (Figure 6(b)). Moreover, it has A/CNK values of 2.95 and plots in the peraluminous field (Figure 6(c)). The TZ biotite granite wall rocks have SiO2 contents of 66.1–71.8 wt.%, K2O contents of 4.47–7.07 wt.%, and Na2O contents of 4.47–7.07 wt.% (Figures 6(a), 6(b), and 6(d)). They plot in the high-K calc-alkalic to shoshonitic fields on the K2O versus SiO2 diagram (Figure 6(b)). Their A/CNK values range from 1.08 to 1.12, plotting in the peraluminous field (Figure 6(c)).

The MAT biotite granite wall rock has a higher ΣREE content (411 ppm) than the MAT pegmatites (Figure 7(a)). The MAT biotite granite wall rock is characterized by slightly light REE-enriched (La/YbCN=5.64) chondrite-normalized patterns with moderate negative Eu anomalies (Eu/Eu=0.21; Figure 7(a)). The chondrite-normalized REE pattern of the MAT biotite granite wall rock is generally parallel to the REE patterns of the MAT pegmatites (Figure 7(a)). The TZ biotite granite wall rocks have variable ΣREE contents (164.4–498 ppm), and they are characterized by obviously light REE-enriched (La/YbCN=14.429.4) chondrite-normalized patterns (Figure 7(c)) and moderate to significant negative Eu anomalies (Eu/Eu=0.150.51) (Figure 7(c)).

On primitive-mantle-normalized plots, the MAT and TZ biotite granite wall rock is enriched in Rb, Th, U, and Pb and slightly depleted in Ba, Sr, Zr, Hf, and Ti, with small negative Nb anomalies and slight Ta positive anomalies (Figure 7(b)). The trace element pattern of the MAT biotite granite wall rock is generally parallel to the trace element pattern of the MAT pegmatites (Figure 7(b)), and it has Nb/Ta and Zr/Hf ratios of 3.44 and 33.0, respectively (Supplementary Table 2). The TZ biotite granite wall rocks have Nb/Ta ratios that range from 9.54 to 11.6 and Zr/Hf ratios that range from 36.7 to 39.6 (Supplementary Table 2).

4.3. Sr–Nd–Hf–Pb Isotopic Compositions

4.3.1. MAT and TZ Pegmatites

Initial Nd–Pb isotopic ratios were calculated based on the U–Pb ages of zircons in the MAT pegmatites (231 and 269 Ma) and in the TZ pegmatite (235 Ma). Initial Hf isotopic ratios were calculated using the corresponding zircon single-crystal ages. As the MAT and TZ pegmatites have exceptionally high 87Rb/86Sr ratios (57.4 to 260; Supplementary Table 3), we did not calculate their initial 87Sr/86Sr ratios. The MAT pegmatites have whole-rock εNdt values of −12.0 to −10.7 with two-stage Nd model ages ranging from 2.0 to 1.9 Ga (Supplementary Table 2; Figure 8(a)). The TZ pegmatites have whole-rock εNdt values of −10.4 to −6.1, with two-stage Nd model ages ranging from 1.9 to 1.5 Ga (Supplementary Table 2; Figure 8(a)). The ca. 230 Ma MAT pegmatite (MAT-04) yields εHft values of −8.0 to −4.0, with two-stage Hf model ages ranging from 1.8 to 1.5 Ga, whereas the ca. 270 Ma MAT pegmatite (MAT-03) has a wider range of εHft values from −7.8 to −1.0 and two-stage Hf model ages ranging from 1.8 to 1.4 Ga (Supplementary Table 3; Figure 8(b)). The MAT pegmatites have (206Pb/204Pb)i ratios of 18.555–18.796, (207Pb/204Pb)i ratios of 15.728–15.736, and (208Pb/204Pb)i ratios of 38.562–38.789 (Supplementary Table 4; Figure 8(c)). The TZ pegmatite has (206Pb/204Pb)i ratios of 16.665, (207Pb/204Pb)i ratios of 15.628, and (208Pb/204Pb)i ratios of 35.157 (Supplementary Table 4; Figure 8(c)).

4.3.2. MAT and TZ Biotite Granite Wall Rocks

Initial Nd–Pb isotope ratios were calculated based on the zircon U–Pb ages of the MAT biotite granite (271 Ma) and TZ biotite granite (435 Ma). The MAT and TZ biotite granite wall rocks have high 87Rb/86Sr ratios (5.6–154; Supplementary Table 2), which preclude the determination of precise whole-rock initial 87Sr/86Sr ratios. The MAT biotite granite wall rock has an εNdt value of −10.6, with a two-stage Nd model age of 1.9 Ga (Supplementary Table 2; Figure 8(a)). The TZ biotite granite wall rocks have whole-rock εNdt values of −9.1 to −8.7, with a consistent two-stage Nd model age of 1.9 Ga (Supplementary Table 2; Figure 8(a)). The ca. 230 Ma zircons of the MAT biotite granite wall rock (MAT-01) yield εHft values of −8.0 to −2.1 with two-stage Hf model ages ranging from 1.8 to 1.4 Ga, whereas the ca. 270 Ma zircons of MAT-01 show relatively higher εHft values of −3.9 to −1.0 and two-stage Hf model ages ranging from 1.5 to 1.4 Ga (Supplementary Table 3; Figure 8(b)). The TZ biotite granite wall rock (TZ-03) has zircon εHft values of −7.7 to −4.1, with two-stage Hf model ages ranging from 1.9 to 1.7 Ga (Supplementary Table 3; Figure 8(b)). The MAT biotite granite wall rock (MAT-01) has (206Pb/204Pb)i ratios of 18.338, (207Pb/204Pb)i ratios of 15.718, and (208Pb/204Pb)i ratios of 38.728 (Supplementary Table 4; Figure 8(c)). The TZ biotite granite wall rocks have (206Pb/204Pb)i ratios of 17.410–17.941, (207Pb/204Pb)i ratios of 15.659–15.691, and (208Pb/204Pb)i ratios of 37.450–38.487 (Supplementary Table 4; Figure 8(c)).

5.1. Magma Source and Evolution

5.1.1. MAT and TZ Biotite Granites

The whole-rock trace element and Nd–Pb isotopic compositions and zircon Hf isotopic compositions of the MAT and TZ biotite granites suggest that their source was dominated by Paleoproterozoic metasedimentary rocks in the middle crust. The MAT and TZ biotite granites have REE and trace element patterns that are similar to those of metasedimentary rock-derived, early Paleozoic, and Triassic granites in the SCB (Figures 7(a) and 7(b)). In particular, the Th contents (27.4–81.5 ppm) and Th/La ratios (0.60–1.00) of the MAT and TZ biotite granites are similar to those of the early Paleozoic and Triassic granites (Th=7.5684.2 ppm; Th/La=0.311.47; [16, 20, 48]). In addition, the MAT and TZ biotite granites also have enriched whole-rock Nd and zircon Hf isotopic compositions that are similar to those of early Paleozoic and Triassic S-type granites with Nd model ages of T2DM=1.9 Ga and Hf model ages of T2DM=2.41.7 Ga in the SCB (Figures 8(a) and 8(b)). Moreover, the whole-rock Pb isotopic compositions of the MAT and TZ biotite granites plot along the upper crustal evolution line (Figure 8(c)).

We conclude from the above observations that the MAT and TZ biotite granites were mainly derived by partial melting of metasedimentary rocks in the middle crust. In general, metasedimentary rock-derived granitic magmas can be generated by either fluid-absent mica breakdown reactions or water-saturated melting of plagioclase [57]. In such a situation, a fluid-absent mica breakdown reaction would generate granitic melts with K2O>Na2O, whereas water-saturated melting of plagioclase would produce granitic melts with K2O<Na2O ([58] and references therein). Therefore, the fact that the K2O contents (4.5–7.0 wt.%) of the MAT and TZ biotite granites are higher than their Na2O contents (0.6–3.0 wt.%; Supplementary Table 2) indicates that they were formed by melting as a result of fluid-absent mica breakdown in the middle crust.

The MAT biotite granite sample (MAT-01) and one TZ biotite granite sample (TZ-03) have SiO2 contents (62.8–65.9 wt.%; Figure 6(a)) that are lower than those of experimental pure crustal melts (SiO270 wt.%; [59]) produced by incongruent dehydration melting. This indicates, therefore, that metaigneous rocks or mantle-derived mafic magmas with SiO2<70 wt.% were also present in the metasedimentary source area of the MAT and TZ biotite granites. The MAT and TZ biotite granites have relatively low Mg# values (27.4–34.5; Mg#=molarMgO/molarMgO+molarFeOT×100; assuming FeOT=0.8998×Fe2O3T) and do not contain mafic microgranular enclaves (Figures 3(a) and 3(e); Supplementary Figures 1a–1b). We suggest, therefore, that other subordinate metaigneous rocks were also present in the metasedimentary source area of the MAT and TZ biotite granites.

5.1.2. MAT and TZ Pegmatites

The whole-rock Nd–Pb isotopic, zircon Hf isotopic, and mineral compositions of the MAT and TZ pegmatites indicate that their primary magmas were derived mainly by partial melting of Paleoproterozoic–Mesoproterozoic metasedimentary rocks in the middle crust. On an εNdt versus age diagram, the MAT and TZ pegmatites plot within the field of the Longquan–Mayuan groups (i.e., part of the basement of the Cathaysia Block) that have Nd model ages of T2DM=2.01.5 Ga (Figure 8(a); Supplementary Table 2). In addition, on an εHft versus age diagram, the MAT pegmatites all plot below the chondritic evolution line, with Hf model ages of T2DM=2.41.8 Ga (Figure 8(b); Supplementary Table 3). The whole-rock Pb isotopic compositions of the MAT and TZ pegmatites are distributed along the upper crustal evolution line (Figure 8(c)). Furthermore, the MAT and TZ pegmatites contain euhedral tourmaline, muscovite, and garnet (Figures 3(c), 3(d), 3(g), and 3(h)), which generally crystallize from Al-saturated granitic magmas [60].

The whole-rock trace element characteristics of the MAT and TZ pegmatites suggest they experienced high degrees of fractional crystallization. The MAT and TZ pegmatites have relatively high contents of Nb (1.69–1649 ppm), Ta (0.23–183 ppm), Rb (7.38–5699 ppm), and Li (4.55–485 ppm; Supplementary Table 2), which were gradually enriched during fractional crystallization. Rubidium is a strongly incompatible element, whereas Sr and Ba are compatible in plagioclase and K-feldspar, respectively. Therefore, with fractional crystallization of a granitic magma, Rb is gradually concentrated in the residual magmas, whereas Sr and Ba are reduced in the residual magmas because of the removal of plagioclase and K-feldspar ([61] and references therein). Consequently, the Rb contents, Rb/Sr ratios, and K/Rb ratios can be used to evaluate the degree of fractional crystallization. In addition, in granitic melts, Ta5+ with an ionic radius of 0.73 Å forms a stronger covalent bond with oxygen than Nb5+ with its smaller ionic radius of 0.70 Å, which results in the relatively higher Ta concentration compared with Nb [62]. Therefore, the Nb/Ta ratios of granites gradually decrease with the degree of fractional crystallization [63, 64]. Given that Nb is more compatible in biotite than Ta, fractional crystallization of biotite could also decrease the Nb/Ta ratios in the residual melts [65]. Moreover, during fractional crystallization, Zr4+ with an ionic radius of 0.80 Å in zircon is gradually replaced by Th4+, which has a similar ionic radius of 0.95 Å [66]. Therefore, the Nb/Ta and Zr/Hf ratios would decrease during fractional crystallization [25]. On the Rb–Ba–Sr triangular diagram, the MAT and TZ pegmatites all plot in the highly fractionated granite field (Figure 9(a)). In addition, the plots of the MAT and TZ pegmatites on the K/Rb versus SiO2, Nb/Ta versus Rb/Sr, and Zr/Hf versus Rb/Sr diagrams indicate that these rocks represent high degrees of fractional crystallization (Figures 9(b)–9(d)).

The MAT pegmatites are hosted by the MAT biotite granite (Figure 3(a); Supplementary Figures 1a–1b), which indicates these rocks have a cogenetic relationship [3]. In addition, the MAT pegmatites and biotite granite have broadly parallel REE and trace element patterns (Figures 7(a) and 7(b)). Therefore, the MAT pegmatites are possibly the fractional crystallization products of the MAT biotite granite [69]. The results of trace element fractional-crystallization modeling, using the MAT biotite granite (MAT-01) as the initial magma, suggest that MAT pegmatite MAT-04 could have been formed by the removal of small amounts of biotite (<10%; Figures 10(a) and 10(b)). Likewise, using MAT biotite granite MAT-01 as the initial magma, four other MAT pegmatites (MAT-02, 03, 05, and 06) could have been formed by the removal of 30–50% K-feldspar (Figures 10(a) and 10(b)). In addition to major rock-forming minerals, the MAT pegmatites could have been formed by fractionation of apatite, titanite, monazite, and allanite, as revealed by plots of whole-rock (La/Yb)CN versus La and zircon La/Yb versus La (Figures 10(c) and 10(d)), assuming the MAT biotite granite was the parental magma. Moreover, the plots of zircon Ta versus Nb for the MAT pegmatites also suggest that fractional crystallization of columbite-group minerals occurred (Figure 10(e)).

The TZ pegmatites are significantly younger (235 Ma) than the TZ biotite granites (435 Ma; Figures 5(d) and 5(e)). Therefore, the TZ pegmatites may not have a genetic relationship with their biotite granite wall rocks, and the parental granites that are related to the TZ pegmatites may not be exposed at the surface. Nonetheless, the zircon Ta versus Nb diagram for the TZ pegmatites suggests they underwent fractional crystallization of columbite-group minerals (Figure 10(f)). Because the MAT and TZ pegmatites represent high degrees of fractional crystallization, we suggest that these pegmatites have the potential for rare metal (Nb, Ta, Be, and Li) mineralization. During fractional crystallization, the Nb, Ta, Be, and Rb contents of the MAT and TZ granitic magmas would have increased, along with a gradual increase in Li and F contents [63]. Therefore, we conclude that the MAT and TZ areas in the middle of the Yuechengling pluton are promising targets for rare metal exploration.

5.2. Heat Source

Based on a recent review of regional geological data ([72] and references therein), it has been suggested that the early Paleozoic intracontinental orogeny of the SCB can be divided into three stages involving double thickening of the crust (460–435 Ma), orogenic collapse (441–415 Ma), and postorogenic (<415 Ma). Therefore, the early Paleozoic TZ biotite granites probably formed during the double-thickening or orogenic collapse stages. After the Middle or Late Devonian, the entire South China region evolved into a stable littoral–neritic depositional environment, characterized by a carbonate depositional platform with limestones, dolomites, black cherts, and minor sandstones and mudstones ([15] and references therein). Therefore, the late Paleozoic MAT biotite granite possibly formed in a postorogenic setting. The SCB was in a compressional tectonic setting during the Middle–Late Triassic [1820, 22, 23]. As a consequence, the Late Triassic MAT and TZ pegmatites probably formed during the double thickening of the crust. Heat is generated via the decay of radioactive elements in the sedimentary rocks (e.g., [20, 73]), possibly enhanced by heating from mantle-derived mafic magmas in adjacent regions [74], induced dehydration partial melting of the metasedimentary, and associated subordinate metaigneous rocks. This generated the primary magmas of the MAT and TZ biotite granites, as well as the hidden parental granites of the TZ pegmatites (Figures 11(a)–11(c)). The MAT pegmatites were fractional crystallization products of the MAT biotite granite (Figures 11(b)–11(d)). The TZ pegmatites were formed by fractional crystallization of hidden parental granites (Figures 11(c) and 11(d)).

  • (1)

    LA–ICP–MS zircon U–Pb dating demonstrates that the MAT pegmatites formed during the Permian (269 Ma) and Triassic (231 Ma). The MAT biotite granite wall rock records two stages of magmatic activity (271 and 231 Ma) that were coeval with the Permian and Triassic pegmatites. The TZ pegmatites probably formed during the Triassic (235 Ma), whereas the TZ biotite granite wall rock formed during the Silurian (435 Ma)

  • (2)

    The MAT and TZ biotite granites were derived mainly from Paleoproterozoic metasedimentary rocks in the middle crust. The MAT pegmatites are the fractional crystallization products of the MAT biotite granite, whereas the TZ pegmatites were formed by fractional crystallization of hidden parental S-type granites

  • (3)

    Because the MAT and TZ pegmatites record high degrees of fractional crystallization, they have the potential for rare metal (Nb, Ta, Be, and Li) mineralization

Data are uploaded as supplementary files and can be found when required. All the data are obtained at the Guangxi Key Laboratory of Hidden Metallic Ore Deposits Exploration, Guilin University of Technology, China.

This is a contribution to the Guangxi Key Mineral Resources Deep Exploration Talent Highland.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Jin-Heng Liu of the Guangzhou Institute of Geochemistry, Chinese Academy of Sciences, is thanked for helpful suggestions and discussions. Financial support for this research was provided by the Natural Science Fund for Distinguished Young Scholars of Guangxi Province (Grants 2018GXNSFFA281009 and 2019GXNSFFA245005), National Natural Science Foundation of China (Grants 92055208, 41862003, and 42173067), Natural Science Foundation of Guangxi Province for Young Scholars (Grants 2022GXNSFBA035538 and 2022GXNSFBA035463), Guangxi Science Innovation Base Construction Foundation (Grant ZY21195031), Fifth Bagui Scholar Innovation Project of Guangxi Province (to Ji-Feng Xu), and Scientific Research Foundation of Guilin University of Technology (Grant RD2100001797).

Exclusive Licensee GeoScienceWorld. Distributed under a Creative Commons Attribution License (CC BY 4.0).