Petrochemical tests were conducted on 5 samples from Qianhe and the north of Xinxian county for major and trace elements analyses. Experiments were carried out at the State Key Laboratory of Isotope Geochemistry, Guangzhou Institute of Geochemistry, Chinese Academy of Sciences. Major elements were determined by the Varian Vista-PRO ICP-AES (CCD full-spectrum direct reading ICP-AES), with an analytical error (RSD) of less than 5%. Detailed procedures can be found in (Goto and Tatsumi, 1994). Trace elements were determined by the PE-Elan 6000 ICP-MS in the ultra-purification laboratory of the Guangzhou Institute of Geochemistry, Chinese Academy of Sciences, with a RSD of less than 10%. Detailed procedures can be found in (Liu et al., 1996).

Zircon samples were processed before U-Pb dating to get rid of mixtures. First, ca. 500 g of rock sample was weighed out and crushed into 1 cm³ grains. Then these grains were put into a 20 cm-size stainless steel bowl and were ground repetitively for 3–5 s each time in an XZW100 vibrating mill until they were broken into 60 mesh. Dust was removed and an aluminum-made batea was used to concentrate heavy minerals. Then through magnetic and electromagnetic processing, magnetic minerals were removed and refined zircon was obtained. Finally, zircon grains with good crystals were hand picked under binocular for dating (Zhang et al., 2004).

LA-ICP-MS zircon U-Pb dating was conducted at the Key Laboratory of Isotope Geochronology and Geochemistry, Guangzhou Institute of Geochemistry, Chinese Academy of Sciences. Zircon grains were fixed by epoxy resin and made into laser ablation sample targets. The sample targets were ground and polished until the central surface of the zircon was exposed. Carbon dust was sprayed on zircon for CL imaging and then was wiped off for LA-ICP-MS U-Pb analysis (single spot laser ablation was used with a diameter of the laser spot of 31 μm and at a firing frequency of 8 Hz). Detailed analytical procedures can be found in (Tu et al., 2011). ²⁰⁴Pb data were used to correct Pb data in zircon. The errors of single data points were within 1σ. The ²⁰⁶Pb/²³⁸U age was obtained, whose weighted averages were at the 95% confidence level. Experimental results were plotted into concordia diagrams using the software of Isoplot 3.0 (Ludwig, 2012).

As shown in Table 1, results were obtained through the geochemical analysis of major elements. It can be seen that the paragneiss in the Qianhe area has a high content of silicon (SiO₂ = 76.10–76.42%), aluminum (Al₂O₃ = 11.58–11.91%), and alkali elements (Na₂O + K₂O = 8.35–8.50%) but shows a deficiency in titanium (TiO₂ = 0.09–0.10%) and calcium (CaO = 0.13–0.26%). Its potassium content is comparatively high (Na₂O/K₂O = 0.57–0.76) and Aluminous Saturation Index (A/CNK) is in the range of 1.35–1.38, suggesting that the rock is peraluminous.

The paragneiss in the north of Xinxian county is consistent with the one in the Qianhe area in terms of chemical composition, as it is rich in silicon (SiO₂ = 68.29–75.93%), aluminum (Al₂O₃ = 12.58–16.06%) and alkali elements (Na₂O + K₂O = 8.14–8.63%) but poor in titanium and calcium (CaO = 0.14–0.53%). By comparison, its content of potassium (Na₂O/K₂O = 0.79–1.19) is lower than that of the paragneiss in the Qianhe area. As the A/CNK is in the range of 1.43–1.85, it is evident that the rock is peraluminous.

The results of trace element analysis are shown in Table 1. The total amount of REE was high in the Qianhe paragneiss (QH-1, QH-2, and QH-3), up to 117.55–216.89 μg/g. As shown in the chondrite-normalized REE diagram (Figure 3), LREEs were enriched as La/Yb was between 4.27 and 4.86 and the negative Eu anomaly was evident (δEu is 0.12–0.13) (Sun and McDonough, 1989). Besides, its pattern curve was skewed to the right and had an obvious V valley. As shown in the primitive mantle-normalized trace element spider diagram (Figure 4), elements such as Rb, K, Th, and U were enriched in the paragneiss in the Qianhe area whereas elements such as Ba, Nb, Sr, and Ti were depleted.

The total amount of REEs was also high in the Xinxian paragneiss (XX4-1, XX4-2), up to 122.97–240.37 μg/g. Its chondrite-normalized REE plot was similar to that of the paragneiss of the Qianhe area; however, the content of heavy rare Earth elements (HREE) was higher with La/Yb in the range of 4.00–11.71. Besides, the negative Eu anomaly was relatively indistinct with δEu of 0.53–0.73. As shown in the spider diagram (Figure 4), elements such as Rb, K, Ba, Th, and U were enriched in the paragneiss in the north of Xinxian county whereas elements such as Nb, Sr, and Ti were depleted.

Zircon U-Pb dating was conducted on Samples QH-1 and QH-1-A from the Qianhe area and Samples XX4-1 and XX4-2 from the north of Xixian county. The CL image (Figure 5) shows that zircon grains were of similar sizes with a length to width ratio of about 1:2. All zircons contained detrital zircon and most of them had a clear core-edge double structure. Detrital zircon in the core of zircons had oscillatory zoning. For most zircons, the edge was darker than the core (Table 2).

Five points on the detrital zircon of the metamorphic composite zircon sample QH-1 were selected for the determination of Th and U concentrations (Figure 5A). The ratio of Th/U suggests that these zircons were formed during the Proterozoic at the ages of 801, 793, 756, 743, and 727 Ma, respectively. Five points on the edge of the metamorphic composite zircon were dated at 251, 228, 226, and 211 Ma, respectively (Figure 6).

A total of 17 points on the detrital zircon of the metamorphic composite zircon sample QH-1-A were selected for the determination of Th and U concentrations (Figure 5B). Four age ranges were obtained: 913, 811-803, 784-724, and 684 Ma. One point on the edge of the metamorphic zircon was dated at 352 Ma.

A total of 24 points on the detrital zircon of Xinxian metamorphic composite zircon samples were selected for the determination of Th and U concentrations (Figure 5C). These points have a wide age range: 456–1,727 Ma. One point on the edge of the metamorphic zircon was dated at 256 Ma.

U-Pb dating results for the edge of the metamorphic zircon and the metamorphic composite zircon were more concentrated, with ages ranging from 211 to 352 Ma and concordant ages of 233 ± 24 Ma. Besides, the Th/U ratios of these seven points are in the range of 0.08–0.37 while three of them are below 0.1, significantly smaller than the Th/U ratio of magmatic zircon (generally greater than 0.4) and closer to that of metamorphic zircon (Bryant et al., 2002). As suggested by the U-Pb dating results, the detrital zircons in the samples were mainly formed during the Neoproterozoic (564–884 Ma), with their ages mainly in the range of 705-811 Ma. However, some detrital zircons were formed in the Paleozoic and Middle Proterozoic as shown by dating results.

Two results were obtained from zircon U-Pb dating. The first one was about the edge of metamorphic zircon, whose age was in the range of 211–352 Ma and concordant age was 233 ± 24 Ma. The other result was about detrital zircon in metamorphic zircon, whose age ranged from 456 to 1,727 Ma, mostly between 705 and 811 Ma.

The edge of the metamorphic zircon in samples was dated to the Triassic, suggesting that paragneiss in Xinxian county experienced high-pressure metamorphism during the Triassic. It was consistent with a large amount of previous research on the metamorphic age of the Dabie-Sulu UHP metamorphic belt which confined the age to the Triassic. Hacker et al. (1998) suggested that the metamorphism of the Dabie-Sulu UHP metamorphic belt reached its peak at 245 Ma through zircon dating for paragneiss. Liu et al. (2006) found through U-Pb dating for the edge of metamorphic zircon with coesite that the metamorphic event of the Dabie-Sulu UHP metamorphic belt took place during 240–225 Ma. Our work suggests that the edge of metamorphic zircon is consistent with the data reported by previous researchers, providing further evidence that paragneiss in Xinxian county experienced ultrahigh-pressure metamorphic events during the Triassic.

The dating results of detrital zircon showed that many samples were dated to the middle of the Neoproterozoic, which accounted for about 72% of the total number of samples. Their age range of 700–800 Ma was consistent with the characteristic zircon age of rocks in the Yangtze block (Rowley et al., 1997; Hacker et al., 1998). Considering no large-scale magmatic activity was recorded in the Middle Neoproterozoic in the North China craton, the protolith of the paragneiss in Xinxian county was hence supposed to originate from the Yangtze block. As the magmatic activity was caused by the breakdown of the Rodinia supercontinent, the protolith of the paragneiss of Xinxian county may contain granite formed in the Middle Neoproterozoic in the Yangtze block. In addition, part of detrital zircon in Sample XX4-2 was dated to the Paleozoic (456–532 Ma), which was consistent with data obtained by previous researches of the paragneiss of the UHP belt (Hacker et al., 1998; Liu et al., 2004). However, such age data were in small amount in previous researches and were considered to be the result of the loss of Pb, or due to the fact that the core and edge of zircon were not separated. Hence, these data were not widely discussed or constraining the age by the fugacity state of the paragneiss and orthogneiss (Fu-lai and Xue, 2007). Additionally, (Yuan et al., 2015) found that the core of metamorphic zircon in paragneiss in Luotian of Hubei province was mostly dated to the Paleozoic. No signs of Pb loss could be found, and the boundary between the margin and core was also clear on CL images. Further, considering parts of zircon were dated to the Paleozoic in our work and they had the Th/U ratio close to the ratio range of zircon of magmatic origin (Dao-gong et al., 2001), part of the protolith of the paragneiss was supposed to form later than 456 Ma. As some detrital zircon in the samples was dated to the Middle Proterozoic, the source material of the protolith could possibly come from the crustal residual of the Middle Proterozoic in the Yangtze block rather than from the Paleozoic sedimentary cover.

With regard to the protolith of the paragneiss of the Dabie-Sulu UHP metamorphic belt, it was considered in previous research to be mainly granite formed by the magma coming from the Rodinia rift in the Neoproterozoic at the north edge of the Yangtze block (Zheng et al., 2003). Based on geochemical analysis of the paragneiss in the Dabie belt and Cretaceous granite, the source materials were likely to come from the north edge of the subducting Yangtze block (Frost and Frost, 2011). Therefore, some researchers proposed that the paragneiss in the Dabie belt and Cretaceous granite were produced at different horizons of the Yangtze block (Shu-guang et al., 1997).

The samples of paragneiss from Xinxian had a high SiO₂ content, up to 68.29–76.42%, and their A/CNK reached 1.35–1.85, indicating they are peraluminous. The alkali content ranged between 8.14 and 8.63%, in which the content of potassium is relatively high as the ratio of Na₂O/K₂O is between 0.57 and 1.19. The content of calcium was low and in the range of 0.13–0.53% (Wong et al., 2009; Frost and Frost, 2011). Light rare Earth elements (LREEs) were enriched in the samples and negative Eu anomaly was distinct. Elements such as Rb, K, Th, and U were enriched whereas elements such as Nb, Sr, and Ti were depleted. This distribution pattern suggests that the protolith was formed by the materials of the continental crust (Chen et al., 2006). Felsic magma generated by partial melting of felsic rock under high pressure can extract Si, K, Na, LILEs, and LREEs (Hermann and Green, 2001), and hence, the magma can become enriched in silicon, alkali, aluminum and LREEs while depleted in calcium and magnesium and iron. As Gu found LILEs and LREEs were depleted in eclogite (Gu et al., 2013), it suggests that the paragneiss in Xinxian county might be partially melted, like the felsic magma formed under high pressure. The continuous process from sub-solidus dehydration to partial melting within the UHP gneissic rocks is also evidenced by experimental and petrological observations in Shuanghe paragneiss (Liu and Wu, 2013). According to the geochemical characteristics of the samples, the protolith is likely to be sedimentary rock that has been partially melted during the subduction and retreat of the Yangtze block (Wallis et al., 2005).

By volume, the gneiss occupies between 70 and 80% of the Dabie-Sulu complex (e.g., Dao-gong et al., 2000; Wang et al., 2000; Hui-fen et al., 2001). Of these, 40–60% comprise foliated, garnet-bearing, A-type granitic rock (Shu-tian and You, 2005). These rocks are associated with jadeite-quartzites and marbles that formed from quartz sandstone and limestone respectively. According to the zircon U-Pb dating results, the protoliths of these rocks are controlled by Paleozoic Dabie - Sulu basin. Distributions of the major-elements of these rocks are high SiO₂, alkali-rich, high Na₂O and low Al₂O₃. The mineral component is more rich in alkaline ferromagnesian. The sedimentary origins of these protoliths appear to be neritic facies, some of which are clastic sediments (quartz sandstone to arkoses) and chemical sediments (limestone). These are characteristic of the continental slope environment. The Dabie-Sulu ocean basin was not a mature ocean; it lacked deep-water sedimentation, such as siliceous deposits and turbidites. It evidently lacked contemporaneous pillow lavas and cumulates of ophiolite rock assemblages which is why we have not found remnants of ophiolitic melanges (Wang and Cong Bo-lin, 1998). As the Dabie-Sulu ocean basin was an aborted oceanic basin, it narrowed, with a resultant increase in salinity. This ultimately led to the oceanic basin sediments becoming increasingly alkaline. This provided the necessary material conditions for the formation of alkaline mafic minerals (sodic pyroxenes and amphiboles) during the Late Triassic metamorphism of these rocks.