Zircon grains were isolated using density and magnetic separation techniques. Zircons were mounted into an epoxy resin disk and polished to expose their surfaces. After photographing in both reflected and transmitted light, CL (cathode luminescence) imaging was taken by a JSM6510 SEM attached to a Gatan CL detector. Zircon U-Pb isotopic analyses were carried out on the transparent zircons without fracture or mineral inclusion using a Thermo X2 ICP-MS housed at the Testing Center of Shandong Bureau of China Metallurgical Geology Bureau, Jinan, Shandong province. The mounted zircon grains were ablated using an attached Coherent Geolas Pro 193 nm laser ablation system with a spot diameter of 30 μm. Ablation occurred in intervals of ten sample zircons, directly preceded and followed by two 91,500 standard zircons and one artificial glass 610. The analyzed results of the 91,500 fall into the ranges of the long-term test values of the Lab, with the uncertainty of 1.5% (1 RSD) for most of the ²⁰⁶Pb/²³⁸U measurement and are listed in Supplementary Table S1. Detailed instrument conditions and data acquisition have been described by Li et al. (2016a). The ICPMSDataCal 8.0 (Liu et al., 2010) was used to select offline raw data, integrate background and analytical signals, and time drift correct and quantitative calibrate U-Pb isotopes. The common lead correction was made following the method of Anderson (2002). The age distributions are visually compared using probability density plots (PDPs; Ludwig, 2003). Weighted average age calculation was made using Isoplot 3.23 (Ludwig, 2003).

All samples were prepared by crushing them in an agate shatter box. Major elements were analyzed using an Axios 4.0 X-ray fluorescence spectrometer (XRF) at the Testing Center of Shandong Bureau of China Metallurgical Geology Bureau, Jinan, Shandong province, following the procedures described by Franzini et al. (1972), with analytical precision better than 1%. Rare earth elements and other trace elements from the rocks were analyzed using ICP-MS (Thermo Fisher icap Q) techniques at the Testing Center of Shandong Bureau of China Metallurgical Geology Bureau, Jinan, Shandong province. The precisions for most elements are better than 5% (1SE).

Zircons Hf isotope analyses were carried out using a Neptune Plus MC-ICP-MS equipped with a GeoLas Pro193 nm laser at MiDeR, Nanjing University. The diameter is 44 μm with a pulse rate of 10 Hz, and beam energy of 5 J/cm². Only zircon grains with concordant ages and suitable sites for Hf analyses were analyzed. The zircon standard 91,500 was analyzed during the analytical session, which yielded a¹⁷⁶Hf/¹⁷⁷Hf ratio of 0.282310 ± 9 (N = 34). The calculation of εHf (t) value and T_DM used the following parameters: ^λLu = 1.867 × 10⁻¹¹ a⁻¹ (¹⁷⁶Lu/¹⁷⁷Hf) _CHUR = 0.0336 and (¹⁷⁶Hf/¹⁷⁷Hf) _CHUR = 0.282785, (¹⁷⁶Lu/¹⁷⁷Hf) _DM = 0.0384 and (¹⁷⁶Hf/¹⁷⁷Hf) _DM = 0.28325 (Griffin et al., 2000). The average crustal ¹⁷⁶Lu/¹⁷⁷Hf value used for TDM calculations is 0.015 (Griffin et al., 2002).

The Guzhang granite is characterized by high SiO₂ (71.14–74.07 wt%), FeO^t (2.35–4.37 wt%), K₂O (3.90–4.64 wt%), Na₂O (2.67–3.02 wt%), moderate P₂O₅ (0.14–0.24 wt%), Al₂O₃ (13.04–13.71 wt%) and low CaO (0.37–0.81 wt%), MgO (0.52–0.62 wt%) contents (Supplementary Table S1, Figure 6). It shows strongly peraluminous character, with obvious high ASI values which range from 1.21 to 1.38. Samples from the Guzhang granite were plotted in the calcic-alkalic ferroan A-type field due to the high FeO^t/(FeO^t + MgO) ratios (0.82–0.88) (Figure 7; Frost and Frost, 2011). The Shadi pluton has SiO₂ contents that range from 66.76 to 75.07 wt%, with lower FeO^t (1.32–3.44 wt%), CaO (0.44–0.54 wt%), moderate P₂O₅ (0.12–0.22 wt%) and higher K₂O (4.30–6.14 wt%) and Na₂O (2.92–5.02 wt%) contents. The Shadi granite has mildly high ASI values (1.12–1.22) and also belongs to the peraluminous series. It has FeO^t/(FeO^t+MgO) values varying from 0.85 to 0.95, slightly higher than the Guzhang granites. Both Guzhang and Shadi plutons have obvious high FeO^t/(FeO^t+MgO) values and low Cao and MgO contents compared with coeval granitic rocks in the orogeny at the same SiO₂ (Figures 6A,B,D).

Samples from Guzhang have higher concentrations of high-field-strength elements (HFSEs, e.g., Nb, Zr, Figures 6F,G) when compared with other coeval granites in the Nanling area (Xu and Xu, 2015). Primitive mantle normalized spider diagrams exhibit strong enrichment in Rb, Th, U, and Nd as well as clear negative anomalies in Ba, Nb, Sr, P, Eu, and Ti. The Rare earth element (REE) abundances vary from 152 to 212 ppm with light REE enrichment, significant negative Eu anomalies, and flat-sloping heavy REE patterns (Figures 8A, B). Samples from the Shadi have similar trace element features with the Guzhang granite, which are featured by the enrichment of HFSEs and depletion of Ba, Nb, Sr, Eu, and Ti. The total REEs of Shadi granites have larger variations than that of Guzhang granites, from 67 to 254 ppm. It also shows strong negative Eu abnormally on the chondrite normalized REEs patterns (Figures 8C, D). The heavy depletion of Eu, Ba, and Sr in both plutons indicates that the feldspars played a significant role during the magma generation, either as residuals in the source or as fractionated minerals.

The majority of granites can commonly be classified as I-, S-, or A-types based on mineral combinations and geochemical characteristics (Bonin, 2007). Chappell and white (1974) divided the granites in the Lachlan fold belt into I- and S-type. The I-type granites are characteristic of low SiO₂ contents, ASI values (normally <1.1), depleted isotopes (e.g., Sr, Nd, Hf), and the appearance of hornblendes, derived from mainly meta-igneous rock sources. The S-type granites are featured by high ASI values (normally >1.1), enriched isotopes with the appearance of garnet and/or cordierite, and are suggested to be mainly derived from meta-sedimentary rocks. It is worth noting that the mineralogical and geochemical characteristics mentioned above are not solid standards for the classification of granite types. For example, aluminum-rich minerals such as primary garnet can also be found in weak peraluminous and even meta-aluminous granites (Chappell, 1999). The geochemical characteristics of A-type granites differ from those of I- and S-type granites (e.g., high contents of HFSEs and Ga/Al ratios), and the geochemical parameters used to define A-type granite are unambiguous. (Loiselle and wones, 1979; Eby, 1990; Creaser et al., 1991; Bonin, 2007).

Even the high ASI values and the existence of biotite and minor muscovite are similar to the future of some typical S-type granites (Bonin, 2007), the high formation temperatures, FeO^tot/(FeO^tot+MgO) ratios, and limited inherited zircons (3 inherited zircons of 76 grains) and other features that will be discussed next preclude the S-type origin. The most noticeable characteristics of Guzhang and Shadi plutons are as follows: 1) high FeO^tot/(FeO^tot+MgO) ratios and ASI values (Figures 6, 7); 2) low contents of Al₂O₃, CaO, MgO (Figure 6); 3) mildly negative to positive Zircon Hf isotopic composition (Figure 5B); 4) high HFSEs concentrations and Ga/Al ratios; 5) the occurrences of interstitial biotites and perthitic microcline; 6) lack of inherited zircons. The FeO^t/(FeO^t+MgO) ratios of samples from both plutons are plotted in the ferroan A-type granites field (Figure 7C, defined by 175 A-type granites worldwide; Frost et al., 2001) and are distinguishable compared with the magmatic rocks that formed during 460–400 Ma in Nanling area (Figure 6B). Most of the samples were plotted near or in the field of A-type granites on different chemical diagrams because of the high HFSEs concentrations and 10,000*Ga/Al ratios (Figure 9).

Besides the distinct A-type granite geochemical features, these two plutons also have high Zr saturation temperatures. The Guzhang granite has whole rock Zr saturation temperatures from 820°C to 840°C (Supplementary Table S1, M[(Na + K + 2*Ca)/(Al*Si)]=1.1–1.2; Watson and Harrison, 2005) and four of six samples from the Shadi granite (M=1.2–1.4) have temperatures from 813°C to 845°C (Figure 10A). The other two low-temperature samples (ca. 750°C) from the Shadi pluton have the highest SiO₂ and lowest Zr (68 and 80 ppm) and Ti contents, which may cause by fractional crystallization. The zircon Ti thermometry is also used to calculate the zircon crystallization temperature. A new correction of Ti in zircon temperature calculation was made by Schiller and Finger (2019), and they suggest the zircon temperatures in most A-type granites were underestimated. The zircons from Guzhang and Shadi granite have saturation temperatures from 700°C to 1000°C and 700°C–950°C, respectively (Figure 10B). Combined with the occurrences of interstitial biotites and perthitic microcline, it may be concluded that both plutons evolved under anhydrous conditions at high temperatures. All of the aforementioned characteristics suggest that these two granitic plutons are ferroan A-type.

There is no consensus on the origin of the A-type granites, especially on the source (Bonin, 2007). Even Frost and Frost (2011) suggested using the term “Ferroan granites” instead of “A-type granites” for some reason, the term “A-type granite” is still more popular than the “Ferron granites” nowadays. The main petrogenetic models for the A-type granite include 1) Partial melting of lower crust rocks that had been granulitized during an earlier thermal event (Collins et al., 1982; Feng et al., 2014); 2) Direct differentiation from the mantle-derived basaltic magma (Frost and Frost, 2011; Girei et al., 2019; Jiang et al., 2022); 3) Partial melting of the quartzofeldspathic rocks (Patiño Douce, 1997; Bogaerts et al., 2006; Kong et al., 2018; Li et al., 2018a; Li et al., 2018b; Li et al., 2018c); 4) Magma mixing from anatectic granitic and mantle-derive mafic magmas (Dahlquist et al., 2010; Cámera et al., 2017).

The granulite model with pre-extraction of I-type granitic melt was argued by some scholars for such melts would have low Fe/Mg and (Na₂O+K₂O)/Al₂O₃ ratios that differ from typical A-type granites (e.g., Creaser et al., 1991; Frost and Frost, 2011; Jiang et al., 2022). However, some studies suggested that the early Paleozoic A-type granites in the Wuyi-Nanling-Yunkai orogen were derived from mainly granulitic rock sources, such as granulitic meta-igneous rocks (Cai et al., 2017), granulitic meta-sedimentary rocks (Feng et al., 2014) or granulite that had extracted S-type granitic magma (Xin et al., 2020). The main reason is due to the low water affinity of the granulite which seems as a suitable source for the anhydrous A-type magmas. We are not intended to connect the Guzhang and Shadi plutons with the granulite source, since it is not a necessary factor to generate the A-type granites (Creaser et al., 1991; Skjerlie and Johnston, 1993).

The Guzhang granite has negative zircon ε _Hf(t) values (−8.7—−3.0) that are unlikely derived from a coeval mantle-derived mafic magma (Figure 5B). The Shadi granite has higher ε _Hf(t) values that overlap with the ranges of coeval mafic rocks, which can be derived from the differentiation of the basaltic rocks isotopically (Figure 5B). However, A-type granites that derived from differentiation of the basaltic magma are generally metaluminous and commonly found in association with mafic intrusions (Mccurry et al., 2008). Both Guzhang and Shadi plutons are peraluminous and not found to be associated with mafic intrusion or mafic enclaves. Thus, they are not likely derived from the direct fractionation of mafic magmas. Besides, zircon ε _Hf(t) values from both plutons have large ranges, up to 10 ε units. The large ranges in zircon ε _Hf(t) are either caused by magma mixing (Griffin et al., 2002) or by disequilibrium melting during the crustal anataxis (Tong et al., 2021). The magma mixing model is not favored here because mafic magmas mostly have low ASI values (<0.8) and the magma produced by pure mixing can hardly achieve the strong peraluminous affinity of our samples, while source mixing is more plausible.

Partial melting of mainly quartzofeldspathic rocks during crustal anataxis is a possible way to generate the A-type granites (Frost and frost, 2011). For instance, dehydration melting of F-rich tonalitic gneiss at mid-crustal pressures can generate strong peraluminous A-type granites (Figure 7; Skjerlie and Johnston, 1993). The Guzhang granite has mildly negative ε _Hf(t) values (−8.7—−3.0), which are higher than most coeval granitic plutons, especially the typical S-type granites (ε _Hf(t) = −34.2—−0.2; n=963, median = −8.4; Xin et al., 2020) and also the Neoproterozoic basement rocks in the west Wuyishan area. (Figure 11). Thus, it is unlikely derived from a source that is dominated by old metasedimentary rock. Zircon from the Shadi pluton has positive ε _Hf(t) values up to +6.3, indicating that it originated from the partial melting of more immature source rocks rather than sedimentary source rocks.

The Neoproterozoic juvenile rocks in the Jiangnan Orogen, like the tonalitic to rhyolitic magmatic rocks from Shuangxiwu and Pingshui area, seem a good candidate as the source of these A-type granites (Chen et al., 2009; Li et al., 2009). Neoproterozoic inherited zircons with depleted Hf isotopic characteristics were found in the Shadi (ε _Hf (t=960 Ma) = 9.2 ± 0.6) and the Guzhang pluton (ε _Hf (t=950 Ma) = 7.3 ± 0.5), both of which occur within the ranges of zircons from the tonalitic rocks at the eastern Jiangnan Orogen (Figure 11; Li et al., 2009). Besides, the Beiwu volcanic rocks with tonalitic composition (See Supplementary Table S2 in Li et al., 2009) have major elements similar to the starting materials used in the dehydration melting experiment in Skjerlie and Johnston, 1993. Thus, the dehydration melting of source rocks that are similar to the juvenile tonalitic rocks from the east Jiangnan Orogen can generate peraluminous Shadi A-type granite, geochemically and isotopically. Therefore, we suggest these two peraluminous granitic plutons were most likely formed by the dehydration melting of Neoproterozoic juvenile tonalitic rocks with minor metasedimentary rocks at medium crust pressure, with the source of Guzhang granite containing more old metasedimentary rocks (Figure 12).

The orogeny of this early Paleozoic belt was first recognized by the mass distributions of the regional metamorphic rocks (greenschist, amphiboles, and minor granulites, Figure 1). Systematic geochronology studies of those metamorphic rocks were formed at 460–436 Ma (Yu et al., 2005; Yu et al., 2014; Li et al., 2010; Tong et al., 2021). The occurrence of high-pressure metamorphic rocks revealed crustal thickening events in the belt (Li et al., 2010; Yu et al., 2014; Tong et al., 2021). The crustal thickening events were mainly suggested to be triggered by continental collision (Charvet et al., 2010; Li et al., 2010; Yao et al., 2012; Wang et al., 2013; Shu et al., 2014; Kong et al., 2021; Zhao et al., 2022). The youngest metamorphic zircon U-Pb ages from the Yiyang granulite are at ca. 435 Ma (Yu et al., 2014) and this age may mark the end of the collision event since no metamorphic rocks with younger ages have been reported. However, the accurate transformational age of this belt from compression to extension is still debated.

A-type granites had long been recognized as derived at extensional settings (e.g., Collins et al., 1982; Forst and Forst, 2011). In collisional orogens, syn-collisional crustal thickening has been followed by delamination of substantial amounts of mantle lithosphere at the post-collisional stage (Davies and von Blanckenburg, 1995). The post-collisional extension provides a suitable environment for hot mafic magma upwelling and also the heat for melting the source rocks of A-type granites (Jiang et al., 2022). Thus, the formation ages of A-type granites are often suggested to be the time that the region changes into a strongly extensional geological background. A-type granites in SCB have previously been dated between 415 and 400 Ma, hence it was hypothesized that the Wuyi-Nanling-Yunkai orogeny transitioned from syn-orogenic crustal thickening to a post-orogenic thinning at 415 Ma. (e.g., Xin et al., 2020).

However, most studies considered the hornblende bearing I-type granites (e.g., Guiyang granite formed at 442 Ma, Zhang et al., 2015) and coeval mafic rocks (e.g., Chayuanshan high Mg basalts formed at 435 Ma, Yao et al., 2012) were produced in an extensional setting that caused by lithospheric delamination. Huang and Wang (2019) also suggested that the occurrence of massive high-K calc-alkaline I-type granites in the Xuefengshan area can exemplify the transitional period from compression to extension, with the ages of ca. 430 Ma. Besides, the gneissic S-type granites in the belt mainly have a peak age of 440 Ma, and massive S-type granites have a peak age of 430 Ma. Those massive S-type granites were also suggested to be derived during the orogenic collapse stage (Feng et al. (2014). The ages of Shadi and Guzhang A-type granites in this study are consistent with the conclusion from the S-, I- type granites and the coeval mafic rocks. Thus, we suggest that the setting of Wuyi-Nanling-Yunkai orogen had changed from compression to extensional at least at ca. 430 Ma, which is 15 Ma earlier than the previous conclusion from A-type granites.