Introduction

The history of Gondwana supercontinent and its configuration have been the focus of many investigations1,2,3,4,5. The Late Neoproterozoic–Cambrian (Pan-African) orogeny, which can be regarded as a diagnostic feature for reconstructing the Gondwana supercontinent, has been recognized in many Gondwana continental fragments including India, East Antarctica and Africa6,7,8,9. Although Pan-African-aged detrital or inherited/xenocrystic zircons have been found in some Paleozoic to Mesozoic sedimentary or igneous rocks of the South China Block (SCB)10,11,12,13, unequivocal Pan-African magmatic or metamorphic rocks have not been identified in the SCB. Thus, most workers believe that the Gondwana-forming orogeny did not affect the SCB, and whether the SCB was an integral part of the Gondwana supercontinent and, if so, where it was located in Gondwana are still uncertain.

In this article, we report the first evidence for a significant Pan-African metamorphic event in the SCB. The results provide robust constraints on the location of the SCB in the Gondwana supercontinent.

Geological setting and sample description

The SCB is composed of the Yangtze Block to the northwest and the Cathaysia Block to the southeast. The intervening Jiangnan orogen considered as a collisional belt14,15,16,17 (Fig. 1) between the two, with the proposed timing of collision varying from late Mesoproterozoic14 to middle Neoproterozoic15,16,17.

Figure 1
figure 1

Simplified geological map of eastern South China (modified from Zhao and Cawood52), showing the location of the honblendites.

The Yangtze Block is characterized by an Archean-Paleoproterozoic crystalline basement surrounded by late Mesoproterozoic to early Neoproterzoic folded belts, unconformably overlain by unmetamorphosed late Neoproterozoic and younger covers18, 19. The Cathaysia Block mainly consists of Paleoproterozoic and Neoproterozoic rocks20, 21. Predominant Precambrian basement of the Cathaysia Block is bounded by the Jiangshan-Shaoxing Fault (JSF) to the northewest and the Zhenghe-Dapu Fault (ZDF) to the southeast (Fig. 1). It includes metasedimentary and metavolcanic rocks that have been metamorphosed from upper greenschist to granulite facies, in the early Palaeozoic22, 23, and at least partly also during the Pan-African orogeny (see below).

Mafic-ultramafic rocks are exposed as lenses in the Precambrian basement of the Cathaysia Block. The ultramafic rocks include serpentinized peridotite, clinopyroxenite and hornblendite. They are commonly associated with meta-gabbros and basalts of ca. 830–860 Ma24. The ultramafic lenses are hosted in gneisses and mica schists of Neoproterozoic age along the ZDF. They are flattened parallel to the regional metamorphic foliation and thus likely experienced the same deformation as their host rocks24. This study is mainly concerned with the hornblendites exposed in the Shitun village, Zhenghe County, South China (Fig. 1). The hornblendites mainly consist of >95% hornblende that is characterized by anhedral crystal habits. There are small amount of plagioclase, quartz, titanite and opaque minerals. Grain boundaries vary from straight to variably indented (Fig. 2).

Figure 2
figure 2

Field and microscope photographs for the hornblendites of the Cathaysia Block.

Results

Zircon U-Pb geochronology and Hf isotopic study

Zircons from a hornblendite sample (14WY-8-15, N27°20′25.6″, E118°45′14.3″) are mostly rounded to subhedral, transparent and light brown in color. Crystal sizes range from 50 to 150 μm, with length/width ratios ranging from 3:2 to 1:1.

The zircons are interpreted to be metamorphic in origin based on the following: (1) CL images reveal sector zoning that is typical of metamorphic zircons (Fig. 3); (2) The zircons are abundant in the sample, whereas igneous zircons should be rare, if any, in ultramafic igneous rocks; and (3) the hornblendites of this study are demonstrated to be metamorphic rocks (see below). As discussed below, during high-grade metamorphism, zircon can grow in hornblende-rich rocks.

Figure 3
figure 3

Representative CL images of analyzed zircons from the hornblendite.

A total of nineteen zircon grains were analyzed. U-Pb results are given in Table 1. More details (including Hf data discussed below) are given in Supplementary Table S1. The data are all concordant or nearly concordant and give a weighted mean 206Pb/238U age of 533 ± 7 Ma (MSWD = 1.02, n = 19) (Fig. 4). This age is regarded as the metamorphic age of the hornblendite. Th/U ratios dominantly range from 0.32 to 0.74.

Table 1 LA-ICP-MS U-Pb data for the hornblendite (14WY-8-15) from Shitun village, Zhenghe County, South China.
Figure 4
figure 4

Concordia diagram of zircon U-Pb data for the dated hornblendite.

The (176Hf/177Hf)i values are concentrated in a range of 0.282386–0.282572. Their single-stage model ages (T DM1) are of 0.94–1.2 Ga. Zircons from mafic rocks from the same area also have T DM1 of 0.9–1.4 Ga24. These features reflect that the principal magma of these mafic-ultramafic rocks was derived from a mantle source with late Mesoproterozoic to early Neoproterozoic model ages. It is noted that morphology and T DM1 age (~1.0 Ga) of the ~530 Ma detrital zircons10 from the Cathaysia Block are similar to that of the zircons of this study, suggesting that some of the detrital zircons might have been sourced from the hornblendites.

Geochemical features of the hornblendites

Major and trace elements results are given in the Supplementary Table S2. The hornblendites are characterized by relatively low SiO2 (43.8–44 wt.%) and high MgO (9.99–10.5 wt.%), FeOT (14.9–15.4 wt.%) and CaO (12–13.4 wt.%) contents with high Mg# (57.2–58.5) and CaO/Al2O3 ratios (0.96–1.15). They have high TiO2 (3.05–3.39 wt.%) and moderate total alkaline contents (2.05–2.6 wt.%) with Na2O (1.66–2.03 wt.%) higher than K2O (0.33–0.66) contents. The hornblendites also have high compatible element contents (Cr 381–430 ppm, Co 54–56 ppm and Ni 175–197 ppm). Their moderate alkali and high TiO2 contents denote a tholeiitic affinity of their primary magma. On the chondrite-normalized REE diagrams (Fig. DR1) they exhibit fractionated LREE patterns (La/YbN = 4.97–5.34) with weak negative to slightly positive Eu anomalies (Eu/Eu* = 0.95–1.21). Their primitive mantle-normalized diagrams are characterized by moderate enrichments in most trace elements, such as Zr (195–231 ppm), Ta (0.93–1.19), Nb (14.1–17.2 ppm) and Th (0.94–2.57 ppm), with slight depletion to enrichment in Sr (286–895 ppm) (Fig. DR2). Such trace element patterns without significant Nb–Ta depletion relative to La are typical of intraplate tholeiitic basaltic rocks.

Sr-Nd and Re-Os isotopic results are listed in Supplementary Tables S3 and S4. They show high initial 87Sr/86Sr ratios of 0.7081 to 0.7095 and relatively consistent 143Nd/144Nd ratios of 0.512296 to 0.512304. The initial Sr-Nd isotope compositions of the hornblendites are consistent with those from enriched mantle source. The hornblendites have high Re (475–2669 ppt) and Os (293–454 ppt) concentrations. The measured 187Re/188Os and 187Os/188Os ratios range from 7.62 to 42.9 and 0.1831 to 0.4748, respectively. In the plot of Re/Os and Os, they fall into the field of mantle melt, indicating that the hornblendites are not mantle residues (Fig. DR3). In the plot of Os and 187Re/188Os, they fall into the field of OIB (Fig. DR4).

Discussion

Since igneous zircons are absent in the hornblendites, the timing of the emplacement of the ultramafic magma cannot be directly determined. A previous study shows that the rhyolites of the bimodal volcanic rocks from the Mamianshan Group, of which the hornblendites of this study form a part, has yielded a U-Pb zircon age of 818 ± 9 Ma25. In addition, mafic rocks in the same area yield weighted mean SHRIMP 206Pb/238U age of 830–860Ma24. The ultramafic rocks of this study are spatially associated with the mafic rocks and are thus likely also of a Neoproterozoic age, which is consistent with their Hf model ages.

The geochemical features of the hornblendites are characterized by enrichment of high field strength elements such as TiO2, Zr, Ta and Th. Overall OIB-like trace element patterns, without positive Pb and Sr anomalies or negative Nb-Ta anomalies, suggest that they are unlikely to have formed from partial melting of metasomatized lithospheric mantle. Since the TiO2 content of basaltic magma from asthenosphere mantle sources is relatively high compared to magmas from lithospheric mantle sources26, the relatively high contents of TiO2 (3.05–3.39 wt.%), Zr (195–231 ppm) and Zr/Sm ratios (21.1–23.3) in the hornblendites indicate that the ultramafic magma were probably derived from an asthenosphere mantle reservoir. Re-Os isotopic data also shows that the primary magma of the hornblendites have sources similar to OIB. The slightly enriched Sr-Nd isotopic compositions further suggest that the ultramafic magma was from enriched asthenosphere mantle source.

High Sc (38.7–44.2 ppm) and V (425–466 ppm) concentrations in the ultramafic rocks are consistent with an igneous origin27. These samples have relatively low Mg# (56–58) and Ni (175–197 ppm), reflective of certain degree of the fractionation crystallization of the olivine and clinopyroxene. In the discrimination diagrams (Fig. DR5–6), these samples plot in the field of within-plate basalt or within-plate tholeiite, similar to the coeval mafic rocks, suggesting that the mafic-ultramafic magma erupted in a continental rift environment which was presumably triggered by asthenospheric upwelling. Shu et al.24 suggest that the ultramafic rocks, pillow basalts, gabbros, dykes, or more generally the bimodal igneous rocks from the Mamianshan Group, were emplaced and partly accumulated in the same rift basin. The evolution of the rift basin in the Cathaysia Block is coeval with the Neoproterozoic basins in many parts of South China, which is similar to that of the Adelaide rift system in southeastern Australia that was related to the breakup of Rodinia supercontinent28.

There are two possible processes for the formation of the amphibole in the hornblendites. Firstly, they can result from the metasomatic alteration of pyroxene by metamorphic fluids in ultramafic rocks29. Secondly, they can originate as an igneous mineral but were recrystallized during high grade metamorphism30. The hornblendites in this study is composed of >95% amphibole. Petrographic observations and geochemical data suggest that amphibole in the hornblendites is magnesio-hornblende, a common metamorphic mineral (Supplementary Table S5, Fig. 5a). Since the temperature and pressure control the Ti and Al content of the amphibole, the high Ti and Al content indicate a higher temperature and pressure of the amphibolites. These amphiboles plot in the high amphibolite-granulite facies field (Fig. 5b). Thus, the hornblende appears to have originated as an igneous mineral but underwent extensive recrystallization during amphibolite to granulite facies metamorphism.

Figure 5
figure 5

Amphibole chemistry from the hornblendite. (a) Si-Mg/Mg+Fe2+ (after Leak et al.53); (b) Al-Ti diagram (after Jin54).

Ultramafic rocks are silica unsaturated, and common minerals such as zircon are typically absent. However, there are large amount of zircons in the hornblendites of this study. As described above, these zircons with sector zoning are typical metamorphic zircons, which most likely formed during high grade metamorphism. Since the Ti content of the zircon is correlated with the equilibration temperature, the formula log(Tizircon) = 6.01 ± 0.03 − (5080 ± 30)/T can be used to calculate the metamorphic temperature31 (Supplementary Table S6 ). The temperature (~700 °C) is obviously lower than the crystallization temperature of the ultramafic magma, but is consistent with the above metamorphic temperature of the magnesio-hornblende. The crystallographic lattice of hornblende can accommodate Zr in significant amounts, thus Zr is a compatible element in hornblende. Simple calculations show that reaction of hornblende to form non-Zr bearing phases will release sufficient Zr to account for at least some new zircon growth32 and therefore hornblende can be regarded as a source for zirconium during high-grade metamorphism.

Various configurations and models have been proposed for the timing and tectonics of the assembly of the Gondwana supercontinent. Some workers argued that the SCB was an isolated continental block in the paleo-Pacific during the assembly of Gondwana33. However, it has been gradually accepted that the SCB was closely related to the Gondwana assembly in the Late Neoproterozoic to Early Paleozoic34, 35. The ca. 533 Ma metamorphic event documented in the hornblendite indicate that the SCB preserves the record of a major Pan-African orogeny, supporting that the SCB was an integral part of the Gondwana assembly. More significantly, it can help to constrain the location of the SCB in Gondwana.

Available geological data shows that the SCB has great affinity with India or Australia. However, the exact position of the SCB in Gondwana has not been well constrained. For instance, detrital zircon age patterns indicate that the SCB were either adjacent to northern India13, 36 or between India and Australia10,11,12, 34. Paleomagnetic data also allow the SCB being either near Eastern Australia or adjacent to Western Australia and India in late Neoproterozoic to Early Paleozoic37,38,39. Faunal affinities between the SCB and the India-Himalaya region appeared throughout much of the Early Paleozoic40. Comparable stratigraphic records between northern India and the Yangtze Block also exist in the Neoproterozoic to Early Paleozoic41.

The Pan-African orogenic belts in general are believed to have formed between 650–500 Ma. They are further classified into three belts, including the western belt (the Brasiliano-Damara orogen), the central belt (the East African Orogen or Mozambique orogen) and the eastern belt (the Kuunga orogen)7 (Fig. 6). The western belt was mostly consolidated at 630–600 Ma9. The N-S trending East African Orogen (the central belt), developed during the closure of the Mozambique Ocean, is also well-dated between 549–535 Ma42. The Kuunga orogen (the eastern belt) runs along the western margin of Australia, and presumably continues along northern Antarctica43. Parts of Sri Lanka and Madagascar may also belong to the eastern belt44. Meert45 interpreted that the Kuunga orogen was 570–530 Ma, coinciding with the closure of the Mozambique Ocean in the central belt, whereas Squire et al.44 suggested an age of 530–515 Ma. Combining with the above mentioned geological data, the ca. 533 Ma metamorphic event in the SCB most likely belongs to eastern Kuunga orogeny.

Figure 6
figure 6

Gondwana reconstruction diagram showing the location of South China Block at 540–530 Ma (modified after Santosh et al.9).

It should be noted that northern India was not involved in the above mentioned main period of Gondwana-forming orogeny46, although India forms the ‘heart’ of Gondwana47. Therefore, based on the occurrence of the Pan-African orogeny in the Cathaysia Block and that recognized in the western Australia and northern Antarctica, we believe that the Cathaysia Block was more likely connected to west Australia and East Antarctica than to north India at ca. 533 Ma. Also, considering the marked similarity between stratigraphic records in the Yangtze Block and those recognized in northern India, it is suggested that the Yangtze Block is adjacent to northern India. We thus present a modified paleogeographic reconstruction for the SCB in the Gondwana supercontinent (Fig. 6). During the Pan-African period (ca. 533 Ma), the SCB was likely located at the nexus between India, Antarctica and Australia, along the northern margin of East Gondwana, with the Cathaysia Block connecting western Australia whereas the Yangtze Block facing northern India.

Conclusions

Metamorphic zircon from a hornblendite in the South China Block (SCB) yield U-Pb age of 533 ± 7 Ma and Hf model ages from 900 to 1200 Ma. The primary magma of the hornblendites was probably derived from an enriched asthenospheric mantle source triggered by asthenospheric upwelling. The ca. 533 Ma high-grade metamorphism recorded in the hornblendites provides first direct evidence for a major Pan-African orogeny in the SCB. Combined with available paleomagnetic data, faunal affinities and comparative stratigraphic records as well as comparative detrital zircon age patterns, the data indicate that the SCB was likely located at the nexus between India, Antarctica and Australia at ca. 533 Ma, with the Cathaysia Block connecting western Australia whereas the Yangtze Block facing northern India.

Methods

Zircon U-Pb geochronology and Hf isotopic analyses

Measurements of U, Th and Pb isotopes of zircon were conducted using an agilgent 7500a quadruple (Q)-ICPMS attached with a Geolas laser-ablation system equipped with a 193 nm Ar-F-excimer laser at the Hefei University of Technology (HFUT). Zircon U-Th-Pb ratios and absolute abundances were determined relative to the standard zircon 91500. Spot size in the range of 40–50 µm was used for data collection. The standard 91500 and GJ-1 zircons were used to calibrate the U-Th-Pb ratios and absolute U abundances. The instrumental setting and detailed analytical procedure have been described by Yuan et al.48. Uncertainties on single analyses are reported at the 1σ level; mean ages for pooled U-Pb analyses are quoted with a 95% confidence interval. Data reduction was carried out using the Isoplot/Ex 3 software49.

Zircon in-situ Hf isotopic analysis was carried out using a Geolas-193 laser-ablation microprobe at the Guangzhou Institute of Geochemistry (GIG), Chinese Academy of Sciences (CAS). External calibration was made by measuring zircon standard 91500 with the unknowns during the analyses to evaluate the reliability of the analytical data, which yielded a weighted mean 176Hf/177Hf ratio of 0.282307 ± 31 (2σ). This value is in good agreement with the recommended value of 0.282305 ± 12 (2σ). The mean βYb value was applied for the isobaric interference correction of 176Yb on 176Hf in the same spot. The ratio of 176Yb/172Yb (0.5887) was also applied for the Yb correction. Details of Hf isotopic analytical method followed Wu et al.50.

Whole-rock analyses

Rock chips were ground in an agate mill and prepared for whole-rock analysis. Major element oxides were analyzed on fused glass disks with a Rigaku RIX 2000 X-ray fluorescence spectrometer (XRF) at GIG, CAS. Based on the measured values of rock standards (BHVO-1 and AGV-1), the analytical uncertainties are estimated to be better than 3% for all the major elements.

Trace elements were determined by Perkin-Elmer Sciex ELAN 6000 inductively coupled plasma mass spectrometry (ICP-MS) at GIG, CAS. Sample powders were decomposed in a mixture of distilled HF-HNO3 in Savillex Teflon beakers for 6 days at 120 °C. The sample solution was dried and the residue dissolved in 50 ml 1% HNO3 for ICP-MS analysis. A set of international standards including BHVO-1, G-2, GSR-3 and AGV-1 was used to estimate the accuracy and precision of the analyses.

Sr-Nd isotopic analyses were carried out at the Guiyang Institute of Geochemistry, Chinese Academy of Sciences. Sample powders (~100 mg) were dissolved in distilled HF-HNO3 in Savillex Screwtop Teflon beakers at 150 °C overnight. Sr and REE were separated on columns made of Sr and REE resins of the Eichrom Company using 0.1% HNO3 as eluant. Separation of Nd from the REE fractions was carried out on HDEHP columns with a 0.18N HCl as an eluant. Isotopic compositions were determined using a Micro Mass Isoprobe Multi-collector Mass Spectrometer (MC-ICP-MS). The mass fractionation corrections for Sr and Nd isotopic ratios are based on 86Sr/88Sr = 0.1194 and 146Nd/144 Nd = 0.7219, respectively. 87Rb/86Sr and 147Sm/144Nd ratios were calculated using the Rb, Sr, Sm and Nd abundances measured by ICP-MS. The measured 87Sr/86Sr ratio of the (NIST) SRM 987 standard and 143Nd/144Nd ratio of the La Jolla standard are 0.710265 ± 12 (2σ) and 0.511862 ± 10 (2σ), respectively.

Re-Os isotopic compositions were determined on the MC-ICPMS mass spectrometer at the State Key Laboratory of Continental Dynamics, Northwest University, Xi’an, China. The analytical procedure is described in detail in Li et al.51. Samples were spiked with solutions enriched in 190Os and 185Re and digested in reverse aqua regia (2:1 HNO3:HCl) in Carius tubes. Osmium was extracted by carbon tetrachloride solvent extraction and further purified by micro distillation. Rhenium was extracted and purified from the remaining solution by anion exchange using AG 1 · 8 resin (100–200 mesh). Instrumental mass fractionation for Os was corrected by normalizing the measured 192Os/188Os to 3.08271. Oxide corrections were made using 17O/16O = 0.00037 and 18O/16O = 0.002047. Rhenium isotopic abundances were determined after total evaporation of the samples on the filaments. This method eliminates the effect of instrumental mass fractionation and yields isotopic ratios more accurate than conventional NTIMS measurement techniques. Total procedural blanks were ~7 pg for Re, and ~2 pg for Os with 187Os/188Os of ~0.298. Contribution of the blank to measured Os concentrations and 187Os/188Os were <10% and <5% respectively. Precision of 187Os/188Os measurements was better than 0.4% (2σ).

Major oxides of minerals

Chemical compositions of amphiboles of hornblendite were analyzed using the JOEL JXA8230 electron microprobe equipped at the School of Resources and Environment Engineering, HFUT, China. The analytical conditions were 15 kV accelerating voltage, a beam current of 20 nA with an electron beam size of 5 μm and 10–20 s counting time. Standards for this laboratory were natural and synthetic minerals.