Introduction

Widespread Late Mesozoic plutonic and volcanic rocks that range in composition from mafic to acidic characterize the eastern Yangtze craton, eastern China. The magmatism has been traditionally thought to be related to the subduction of the paleo-Pacific plate underneath the eastern China continent (Zhou and Li 2000; Li and Li 2007). On the contrary, Zhai et al. (1996) proposed that a continental rift setting was responsible for the magmatism. Recent investigations have led to recognition of a number of adakitic intrusions over the eastern Yangtze craton, whose origin has been interpreted as derivation from partial melting of a thickened lower crust or delaminated lower crust in the lithospheric mantle (Xu et al. 2002; Wang et al. 2004, 2006).

The Daye region represents the westernmost and largest magmatic region within the eastern Yangtze craton (Fig. 1). Previous isotopic dating, mainly by K–Ar and partly Rb–Sr and conventional U–Pb methods, yielded ages for most intrusive rocks ranging from 240 to 90 Ma (Shu et al. 1992; Su and Liu 1994). These data, if correct, would imply that the Daye region had a protracted magmatic history lasting from the early Triassic to the late Cretaceous. The plutonic rocks can be broadly grouped as diorites and quartz diorites (Shu et al. 1992), with the latter having an adakite affinity (Wang et al. 2004). Genesis of the adakitic quartz diorites has been interpreted as derivation from melting thickened or delaminated mafic lower crust. The aim of this study was to provide a comprehensive understanding in the age, genesis and geodynamic control of the magmatism of the region. In this contribution, we first present zircon U–Pb ages by sensitive high-resolution ion microprobe (SHRIMP) and laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) to precisely date the magmatism. We further use geochemical and Sr–Nd isotope data and results of geochemical modeling to constrain the petrogenesis of the plutonic rocks and show that fractional crystallization of metasomatised mantle-derived magmas has been an important mechanism in generating the adakitic quartz diorites. Finally, we discuss the geodynamic setting that was responsible for the extensive magmatism in light of existing geological, geochronological and geochemical data.

Fig. 1
figure 1

Schematic illustration of the seven magmatic regions from the eastern Yangtze craton (after Zhai et al. 1992). TLF Tan-Lu Fault, XGF Xiangfan-Guangji Fault, CYF Changzhou-Yangxin Fault

Full size image

Geological background

The Yangtze craton is separated to the north by the Dabie-Sulu Orogen from the North China craton and to the south by a Neoproterozoic suture from the Cathaysia block (insert of Fig. 1). The eastern part of the craton is confined or cut through by several deep faults, mainly the Xiangfan-Guangji, Yangxin-Changzhou, and Tan-Lu Faults (Fig. 1). Interpretations of geophysical data reveal the existence of the Yangtze River Fault (Fig. 1), which extends for ca. 450 km from Daye in the southwest to Zhenjiang in the northeast (Chang et al. 1991). The eastern Yangtze craton contains seven magmatic regions, which are, from southwest to northeast, the Daye, Jiurui, Anqing, Luzong, Tongling, Ningwu, and Ningzhen (Fig. 1). The Daye region is defined by three major faults, namely the Tan-Lu, Xiangfan-Guangji, and Ma-Tuan Faults (Fig. 2).

Fig. 2
figure 2

Geological map of the Daye region (modified from Shu et al. 1992). Also shown are the sampling sites for U–Pb geochronology and the respective ages

Full size image

The stratigraphy in Daye consists of Ordovician to middle Triassic marine sedimentary rocks (>10,000 m thickness), late Triassic to Cenozoic continental deposits, and Cretaceous volcanic assemblages (Fig. 2). The pre-Triassic strata were intensely folded due to the continental collision between the Yangtze and North China cratons during the middle Triassic (Hubei Bureau of Geology, Mineral Resources 1990). Extensive volcanism of the early Cretaceous time resulted in >2,000 m thick volcanic associations, which are grouped, from the base to the top, as the Majiashan, Lingxiang and Dasi Formations (Hubei Bureau of Geology, Mineral Resources 1990). The volcanic rocks have K–Ar ages ranging from 130 to 113 Ma (Shu et al. 1992). Recent SHRIMP zircon U–Pb dating of one dacitic sample from the Dasi Formation yielded a weighted mean 206Pb/238U age of 128 ± 1 Ma (Xie et al. 2006). In the SiO2 versus K2O + Na2O diagram, rocks of the Dasi Formation display typical bimodal pattern (Xie et al. 2006).

Several major intrusions were emplaced into the Devonian to early Triassic sedimentary rocks (Fig. 2), including the E’Cheng and Lingxiang diorite, the Jinshandian syenodiorite (hereafter diorite), the Tieshan and Yangxin quartz diorites, and the Yinzu complex consisting of quartz diorite and a small volume of gabbro-diorite. More than 30 granodiorite stocks, mostly less than 1 km2 of exposure, occur surrounding or intruding the plutons. The Jinshandian diorite consists of the Jinshandian pluton in the east and the Wangbaoshan pluton in the west (Fig. 2). The Tieshan and Lingxiang intrusions contain abundant mafic enclaves composed of variable amounts of plagioclase, hornblende, biotite, pyroxene, and K-feldspar, and accessory titanite, magnetite, and apatite. Monzonite, diorite, and gabbro-diorite dykes, mostly 0.4–2 m wide, intrude the E’cheng, Tieshan, and Jinshandian plutons. The plutons all contain abundant accessory minerals, consisting of zircon, apatite, titanite, and magnetite (Shu et al. 1992).

Analytical methods

U–Pb geochronology

Three samples, each from the Tieshan, Yangxin and Yinzu quartz diorite (Fig. 2) were selected for SHRIMP zircon U–Pb dating. Zircons were handpicked under a binocular microscope after conventional liquid and magnetic separation. About 50–60 grains were mounted in an epoxy resin disc together with chips of the 417-Ma Canberra standard TEMORA (Black et al. 2003). The discs were polished, cleaned, and then gold coated. Prior to isotopic analysis, all grains were photographed under transmitted- and reflected-light, and subsequently examined using the cathodoluminescence (CL) image technique. U–Pb isotopes were analyzed on a SHRIMP II hosted in the Institute of Geology, Chinese Academy of Geological Sciences (Beijing), with procedures similar to Compston et al. (1992). During the SHRIMP analysis, spot size was averaged ca. 30 μm and each spot was rastered over 120 μm for 3 min to remove common Pb on the surface or contamination from the gold coating. The TEMORA zircon standard was used for interelement fractionation correction, and the 572-Ma standard SL13 for determining U, Th, and Pb concentrations. Corrections of common-Pb were applied using the 204Pb-correction method.

Two samples from the Lingxiang and Jinshandian diorites, and the other two from adjacent monzonite and gabbro-diorite dykes (Fig. 2), were collected for LA-ICP-MS zircon U–Pb dating. Analyses were performed on an Elan 6100 DRC ICP-MS, hosted at the Department of Geology, Northwest University (Xi’an, China). The GeoLas 200 M laser ablation system was used for laser ablation, which is equipped with a 193-nm ArF excimer laser and a homogenizing, imaging optical system. Approximately, 30-μm spot-size was adopted in this study. Helium was used as a carrier gas to enhance transport efficiency of ablated materials. Both standards and samples were arranged on the gas line in order to improve transfer efficiency and reduce elemental fractionation. The 207Pb/235U ratio was calculated from the values of 207Pb/206Pb and 206Pb/238U. Common Pb was not subtracted in this work due to the large determination uncertainty of 204Pb, as noted elsewhere (Andersen 2002; Jackson et al. 2004). Analyses of the standard TEMORA 1 as an unknown yielded a weighted 206Pb/238U age of 415 ± 4 Ma (MSWD = 0.112, n = 24) (Yuan et al. 2004), which is in good agreement with the recommended ID-TIMS age of 416.8 ± 0.2 Ma (Black et al. 2003). Details of analytical procedures have been described in Yuan et al. (2004).

Whole-rock geochemical analysis

Major and trace elemental analysis of 28 fresh whole-rock samples were made at the Department of Earth Science, University of Hong Kong. Samples were crushed and powdered in an agate mill. Major oxides were determined by wavelength-dispersive X-ray fluorescence spectrometry of fused glass beads using a Philips PW2400 spectrometer. Trace elements, including REE, were determined using a VG Plasma-Quad Excel ICP-MS after a 2-day closed beaker digestion using a mixture of HF and HNO3 acids in high-pressure ‘‘bombs’’ (Qi et al. 2000). Pure elemental standard solutions were used for external calibration and BHVO-1 and SY-4 were used as reference materials. The accuracies of the XRFS analyses are estimated at ~2% for elements with concentrations greater than 0.5 wt.% and 5% for those >0.1 wt.%. The uncertainties of the ICP-MS analyses are estimated to be better than ±5% for most trace elements with concentrations >10 ppm, and ±10% for those <10 ppm.

Sr–Nd isotopic determinations

For Rb–Sr and Sm–Nd isotope analyses, 100 mg powder (10–20 μm) of each sample were spiked with mixed isotope tracers (87Rb, 84Sr, 149Sm, 150Nd), and digested in Teflon capsules with a mixture of distilled HF and HClO4 acids. Strontium and REE were separated and purified using cation exchange columns of AG50 W X-12 (200–400 mesh resin). Samarium and Nd were separated and purified using a second exchange column packed with Teflon powder with an exchange medium of HDEHP. Isotopic measurements were conducted on a Finnigan MAT-262 mass spectrometer at the Institute of Geology and Geophysics, Chinese Academy of Sciences (Beijing). Strontium was loaded with a Ta–HF activator on preconditioned Ta filaments and measured in single filament mode. Neodymium was loaded on preconditioned Re filaments and measured in a Re double filament configuration. Mass fractionation corrections for Sr and Nd isotopic ratios were based on values of 86Sr/88Sr = 0.1194 and 146Nd/144Nd = 0.7219. Analyses of standard BCR-1 yielded averages of 87Sr/86Sr = 0.704998 ± 12 (2σ), 143Nd/144Nd = 0.512658, 147Sm/144Nd = 0.1411, Sm = 6.621 ppm, and Nd = 28.40 ppm. Total procedural blanks were <100 pg for Sr and <50 pg for Nd, and the estimated analytical uncertainties of 147Sm/144Nd and 87Rb/86Sr ratios are <0.5%.

Results

U–Pb geochronology

SHRIMP U–Pb analytical data are summarized in Table 1 and graphically illustrated in the concordia diagrams (Fig. 3). Errors on individual analyses are cited as 1σ, and the weighted mean 206Pb/238U ages are quoted at the 95% confidence level.

Table 1 SHRIMP zircon U–Pb data of the quartz diorite plutons from Daye, eastern China
Full size table
Fig. 3
figure 3

SHRIMP U–Pb concordia diagrams for the Yinzu (a), Yangxin (b), and Tieshan (c) quartz diorites

Full size image

Sample YZ3 was taken from the Yinzu quartz diorite. Zircons are colorless and generally clear or weakly cloudy, with excellent prism and pyramid faces (Fig. 3a). The crystals range in size from 150 to 300 μm, and their elongation ratios are mostly 2:1 to 3:1. Twelve analyses on ten zircons yielded 206Pb/238U ages ranging from 160.9 ± 4.9 to 143.6 ± 6 Ma (Table 1) and constitute a coherent group with a weighted mean 206Pb/238U age of 151.8 ± 2.7 Ma (Fig. 3a).

Sample FJS2 was collected from the Yangxin quartz diorite. Zircons are euhedral and 100–200 μm long, with an elongation ratio of 1.5–2:1 (Fig. 3b). Most of them are cloudy and only a few are clear. We analyzed 14 spots on 12 zircons, yielding 206Pb/238U ages of 146.8 ± 4.4 to 134.6 ± 4.7 Ma (Table 1). All 14 analyses form a concordant homogeneous group with a weighted mean 206Pb/238U age of 138.5 ± 2.5 Ma (Fig. 3b).

Sample TS2 was selected from the Tieshan quartz diorite. Zircons are euhedral, colorless or pale to grey, and generally 200–280 μm long, with an elongation ratio of 1–1.5:1. CL images reveal oscillatory zoning (Fig. 3c). Fourteen analyses were carried out on 11 grains, yielding 206Pb/238U ages ranging from 146.3 ± 5 to 127.6 ± 4 Ma (Table 1) and form a coherent group with a weighted mean age of 135.8 ± 2.4 Ma (Fig. 3c).

The LA-ICP-MS zircon analytical results are summarized in Table 2 and illustrated in Fig. 4. Zircons have oscillatory zoning typical of magmatic origin, similar in morphology and texture to those used for SHRIMP analysis. Sample LX3 was selected from the Lingxiang diorite. Nineteen analyses of 20 yielded 206Pb/238U ages ranging from 143 ± 1 to 138 ± 1 Ma, forming a coherent group that gave a weighted mean 206Pb/238U age of 141.1 ± 0.7 Ma (Fig. 4a). The remaining one has a slightly younger age of 136.6 ± 0.9 Ma (LX3-F.09; Table 2).

Table 2 LA-ICPMS zircon U–Pb data of diorite plutons and dykes from Daye, eastern China
Full size table
Fig. 4
figure 4

Concordia diagrams of LA-ICP-MS zircon U–Pb data for the Lingxiang (a) and Jinshandian diorites (b), and late-stage dykes (c, d)

Full size image

Sample WBS3 was taken from the Wangbaoshan pluton of the Jinshandian diorite (Fig. 2). Twenty analyses on 18 zircon crystals yielded ages that cluster in two groups (Fig. 4b). The first group comprises 14 analyses ranging in age from 136 ± 2 to 128 ± 2 Ma, with a weighted mean 206Pb/238U age of 132.4 ± 1.3 Ma that is interpreted as the crystallization age for the Jinshandian diorite. The second group comprises six analyses, whose ages are 124 ± 1 to 121 ± 1 Ma. The younger ages resulted possibly from analytical uncertainties due to signal instability.

Sample WBS2 was collected from a 1.5-m-wide, NE-striking fine-grained monzonite dyke that intruded the Wangbaoshan pluton. Nineteen analyses were carried out on 18 zircon grains. One analysis yielded an old age of 141 ± 2 Ma that is identical to the age of sample LX3 (Fig. 4a) and thus interpreted as an inheritance from the nearby Lingxiang diorite or contamination by the sample LX3 during sample preparation. The remaining 18 analyses form two coherent groups that have weighted mean 206Pb/238U age of 127.5 ± 1.6 Ma (n = 14) and 118.9 ± 2.2 Ma (n = 4), respectively (Fig. 4c). The weighted mean age of the first group is interpreted as the crystallization age of the dyke, whereas the second group may have resulted from signal instability during analysis.

Sample WBS4 was selected from a 1.2-m-wide gabbro-diorite dyke emplaced into the late Jurassic conglomerate, 200 m southeast of the Wangbaoshan pluton (Fig. 2). Twenty analyses from 19 zircons yielded ages ranging from 125 ± 2 to 119 ± 2 Ma (Table 2). The weighted mean 206Pb/238U age of 121.5 ± 0.6 Ma (Fig. 4d) is interpreted as the crystallization age for the gabbro-diorite dyke.

Whole-rock geochemistry

Major and trace elemental contents are listed in Table 3 and illustrated in Figs. 5, 6 and 7. The rocks have silica between 52.3 and 69.9 wt.% and total alkali of 4–9.6 wt.% and can be broadly grouped as diorite (Lingxiang and Jinshandian plutons) and quartz diorite (Yangxin, Tieshan, and Yinzu plutons) (Fig. 5a). Most samples plot in the high-K field in the SiO2 versus K2O diagram (Fig. 5b). They have A/CNK ratio [molar Al2O3/(CaO + Na2O + K2O)] mostly between 0.73 and 0.96 (Table 3) and A/NK ratio [molar Al2O3/(Na2O + K2O)] of 1.3–2.3, indicating a metaluminous affinity (Fig. 5c), consistent with the presence of normative diopside in most samples (Table 3). The Na2O/K2O ratios of 1.3–3 (except samples YZ2 and JS1) indicate the rocks are sodic (Table 3), as demonstrated by the K–Na–Ca cationic ternary plot in which they follow a TTG differentiation rather than a calc-alkaline trend (Fig. 5d). Rocks have MgO and Mg# [100 × molar Mg/(Mg + Fe) ratio] of 0.64–3.46 wt.% and 21–48, respectively. Samples JS1, TS3 and YZ2 are somehow different from others of the same pluton (Fig. 5). Both diorites and quartz diorites have high Al2O3 (15.1–19.3 wt.%), Na2O (3.0–7.2 wt.%), Sr (634–2087 ppm), but low Rb (27–119 ppm), Ni (5–28 ppm), Cr (10–34 ppm), Nb (7–26 ppm) and Zr (42–166 ppm). The quartz diorite samples show an adakite geochemical affinity, marked by low Y (7–22 ppm), Yb (0.5–1.8 ppm), and Sc (5–15 ppm) and high Sr/Y (46–128) and La/Yb (32–63) ratios.

Table 3 Major and trace elemental abundances of plutonic rocks from Daye, eastern China
Full size table
Fig. 5
figure 5

Geochemical characteristics of the plutonic rocks in the Daye region. a Total alkali versus SiO2 diagram; b K2O versus SiO2 diagram; c A/NK versus A/CNK plot; and d Cationic Ca–Na–K diagram. The TTG and CA trends in d follow Barker and Arth (1976) and Nockolds and Allen (1953), respectively. A/NK = [molar Al2O3/(Na2O + K2O)], A/CNK = [molar Al2O3/(CaO + Na2O + K2O)], TTG Trondhjemite–Tonalite–Granodiorite, CA Calc-Alkaline

Full size image
Fig. 6
figure 6

Selected oxide and trace element Harker’s plots for the Daye diorites and quartz diorites. Symbols as in Fig. 5

Full size image
Fig. 7
figure 7

Chondrite (Boynton 1984) normalized rare earth element patterns (a, b) and primitive mantle (McDonough and Sun 1995) normalized spiderdiagrams (c, d) for the plutonic rocks of the Daye region. The shaded area denotes the Cretaceous potassic to ultra-potassic basalts from Daye and other regions of the eastern Yangtze craton (Wang et al. 2001, 2006; Liu et al. 2002; Yan et al. 2008)

Full size image

Figure 6 illustrates the Harker diagrams of selected major and trace elements taking SiO2 as differentiation index. With exceptions of samples TS3, JS1 and YZ2, the rocks show a linear or curvilinear trend in most Harker diagrams, suggesting close petrogenetic relationships between the diorites and quartz diorites. Major oxides including Al2O3, FeOT, MgO, CaO, TiO2 and P2O5 and trace elements such as Sc, Yb, V, and Y decrease with increasing SiO2, indicating fractionation of hornblende, plagioclase, apatite, and Fe–Ti oxides.

Both diorite and quartz diorite exhibit similar, sub-parallel chondrite-normalized REE patterns, characterized by strong fractionation between LREE and HREE, without significant Eu anomalies (Fig. 7a, b). The quartz diorites are more HREE depleted than the diorites, marked by (La/Yb)N values of 21–43 and 9–22 and YbN of 3–11 and 9–15, respectively (Table 3; Fig. 7a, b). In the primitive mantle-normalized spiderdiagram (Fig. 7c, d), both quartz diorites and diorites are enriched in highly incompatible elements such as LREE and large ion lithophile elements (LILE), but depleted in high field strength elements (HFSE), with prominent negative anomalies of Ta–Nb and Ti. The quartz diorites are more so than the diorites, with (Rb/Lu)N ratios of 67–80 and 36–50, respectively (not shown).

Sr–Nd isotope data

The Rb–Sr and Sr–Nd isotope data are reported in Table 4 and Fig. 8. Data of the 141-Ma Tongshankou granodiorite (Li et al. 2008) and the 128-Ma basalts of the Dasi Formation (Xie et al. 2006) from the same region are also included for comparison. Initial 87Sr/86Sr and 143Nd/144Nd ratios were calculated using the crystallization age of each intrusion. In the εNd(t) versus (87Sr/86Sr)t diagram (Fig. 8), both quartz diorites and diorites lie in the enriched extension of the mantle array defined by mantle-derived rocks. The samples have similar, radiogenic, relatively homogeneous Sr–Nd isotopic compositions, characterized by (87Sr/86Sr)t = 0.7058–0.7091, (143Nd/144Nd)t = 0.512019–0.512294, and εNd(t) = –3.4 to −8.8. These values are strikingly distinct from the lower crust of the Yangtze and North China cratons (Ames et al. 1996; Jahn et al. 1999) and the young upper continental crust (Taylor and McLennan 1985) as well. They also differ significantly from the Archean Kongling orthogneiss of the Yangtze craton (Ames et al. 1996). Instead, the Sr–Nd isotope data overlap those of the early Cretaceous basalts or their plutonic equivalents from Daye (Xie et al. 2006), Luzong (Liu et al. 2002; Wang et al. 2006), Ningwu (Wang et al. 2001), and Tongling regions (Yan et al. 2008) of the eastern Yangtze craton (Fig. 1), which have been demonstrated to have an enriched-mantle affinity.

Table 4 Sr–Nd isotopic compositions of magmatic rocks from the Daye region
Full size table
Fig. 8
figure 8

Initial 87Sr/86Sr versus εNd(t) for the Daye diorites and quartz diorites. Reference data sources: Dasi Formation basalts (Xie et al. 2006), Luzong and Ningwu basalts (Wang et al. 2001, 2006; Liu et al. 2002), mafic intrusions of the eastern Yangtze craton (Yan et al. 2008), Archean Kongling gneiss (Ames et al. 1996), upper continental crust (UCC; Taylor and McLennan 1985). Other fields cited from Jahn et al. (1999). Also shown are the mixing calculations for source and melt contaminations using the parametric equations 13.5 of Faure (1986). The mixing hyperbolas were modeled by mixing average compositions of lower continental crust (LCC) and UCC (Rollinson 1993; Rudnick 1995) with those of N-MORB (Rollinson 1993; Sun and McDonough 1989). See Fig. 2 for locations of referenced areas

Full size image

Discussions

Large-scale Late Mesozoic magmatism

The present geochronological data demonstrate that the latest Jurassic Yinzu quartz diorite, and presumably together with the undated gabbro-diorite of the Yinzu complex, marks the initiation of Late Mesozoic magmatism in the Daye region. There was a magmatic hiatus of ca. 10 million years until the emplacement of the Lingxiang diorite (141.1 ± 0.7 Ma), which was immediately followed by the formation of the Yangxin (138.5 ± 2.5 Ma) and Tieshan quartz diorite (135.8 ± 2.4 Ma). The Jinshandian diorite (132.4 ± 1.3 Ma) represents the end of major plutonic magmatism, although dykes emplacement persisted. Despite we were not able to date the E’cheng diorite, a phlogopite K–Ar age of 137 Ma (Su and Liu 1994) for the related Cheng-Chao iron skarn deposit (Fig. 2) indicates a similar, early Cretaceous age for its emplacement.

Extensive volcanism took place shortly after the plutonic magmatism in the Daye region. K–Ar dating on biotite and whole-rock samples yielded ages for the volcanism in the 130–113 Ma range (Hubei Bureau of Geology, Mineral Resources 1990; Shu et al. 1992). Recently, Xie et al. (2006) reported a SHRIMP zircon U–Pb age of 128 ± 1 Ma for the Dasi Formation dacite, close to the emplacement ages of the dykes from this study. Collectively, it is concluded that plutonic magmatism in the Daye region started in the late Jurassic (ca. 152 Ma) and peaked in the early Cretaceous (141–132 Ma), which was followed by volcanic eruption and dyke emplacement (Fig. 9a).

Fig. 9
figure 9

Age distribution of the Late Mesozoic magmatic rocks in the Daye region (a) and other regions of the eastern Yangtze craton (b). See text for data sources and explanations

Full size image

Geochronological studies from other regions of the eastern Yangtze craton have yielded results comparable with the Daye regions (Fig. 9b). Similar to Daye, the 151.8 ± 2.6 Ma Shatanjiao quartz monzonite (SHRIMP U–Pb age; Di et al. 2005) represents the earliest magmatic event in the Tongling area (Fig. 1). Pervasive magmatism succeeded in the interval of 142.8 ± 2.2 to 135.8 ± 1.1 Ma, constrained by SHRIMP and LA-ICP-MS zircon U–Pb (Xie et al. 2008, and references therein) and hornblende 40Ar/39Ar ages of granitoid plutons (Wu et al. 1996). Similar ages (142.3 ± 1.6–134.7 ± 2.3 Ma) were also obtained for the magmatic rocks and associated ore deposits from the Anqing and Jiurui regions (Fig. 1) (Chen et al. 1991; Wu and Zhou 1997; Mao et al. 2004). The Luzong and Ningwu regions (Fig. 1) are characterized by widespread shoshonitic to high-K calc-alkaline volcanic rocks ranging in composition from basalts to rhyolites, coexisting with small volumes of gabbros, diorites, quartz diorites, syenites, and A-type granites (Chang et al. 1991; Xing and Xu 1994, 1996; Wang et al. 2001; Liu et al. 2002). Available age data indicate that both the volcanic rocks and the mafic and alkaline intrusions were emplaced mainly in the 136–118 Ma range (Xu and Xing 1994; Liu et al. 2002; Wang et al. 2006). Taken together, existing age data demonstrate that extensive magmatism throughout the eastern Yangtze craton was broadly coeval and marked by an early plutonic magmatism of 143–132 Ma and subsequent volcanic eruptions and dyke emplacement during 130–113 Ma (Fig. 9b).

Petrogenesis of the Daye plutonic rocks

Mantle-derived origin and genetic link between the diorites and quartz diorites

The Sr–Nd isotopic data (Table 4) of the present rocks are strikingly distinct from the Archean basement rocks of the Yangtze craton and the lower crust of both the Yangtze and North China cratons (Fig. 8), precluding their formation by direct remelting of any crustal sources. Instead, the Sr–Nd isotopes are indistinguishable from the contemporaneous, mantle-derived basalts (130–120 Ma) and mafic intrusions (143–125 Ma) of the eastern Yangtze craton (Wang et al. 2001, 2006; Liu et al. 2002; Yan et al. 2008). Combined with the geochemical features (Figs. 5, 6, 7), we suggest that intrusions from the Daye region were most likely generated by partial melting of an enriched mantle source. This is corroborated by the similarity of Sr–Nd isotopes between the Daye intrusions and an EM-II source that has resulted from metasomatism of subcontinental lithospheric mantle by hydrous slab melts or fluids (Geldmacher et al. 2008). The homogeneous isotopic compositions of individual intrusions (Table 4) and lack of a correlation between SiO2 and εNd(t) values (not shown) also indicate that assimilation of crustal materials, if any, was not evident in the magmatic process. This is partly confirmed by the lack of old inherited zircons in light of both CL images and spot U–Pb analyses of zircon grains (Figs. 3, 4).

It is noted that both diorites and quartz diorites have prominent negative Nb–Ta anomalies (Fig. 7c, d), typical signature of subduction-related magmas (Sajona et al. 1993). When normalized to the average of oceanic plagiogranites (ocean ridge granite) reported by Pearce et al. (1984), the rocks resemble arc granites but clearly differ from within-plate granites (not shown). This seems to indicate that the rocks have an arc affinity. Such anomalies, however, were also frequently observed in rocks derived from enriched subcontinental lithospheric mantle (Whalen et al. 1996; Jahn et al. 1999). Absence of arc magmatism in the Daye region implies that the observed Nb–Ta anomalies reflect an enriched lithospheric mantle source that resulted from interactions with subduction-related melts and/or fluids. The radiogenic Sr–Nd isotopic data of the Daye rocks, which are compatible with the contemporaneous, mantle-derived potassic to ultra-potassic basalts and mafic intrusions (Fig. 8), confirm this view. The early Cretaceous basalts and mafic intrusions over the eastern Yangtze craton all display an enriched-mantle affinity (Wang et al. 2001, 2006; Liu et al. 2002, 2006; Xie et al. 2006; Yan et al. 2008), demonstrating that modification of the mantle beneath the eastern Yangtze craton has been pervasive. Timing and nature of the subduction and related metasomatism are not clear, but we envisage two possible causes here. One is the collision of the Cathaysia block towards the Yangtze craton during the ca. 1,000-Ma Jinning orogeny (Yang et al. 1986). The other may relate to the Mesozoic flat-slab subduction of the paleo-Pacific plate of ca. 1000 km diameter underneath the eastern China continent, beginning at middle Jurassic (Zhou and Li 2000) or early Triassic (Li and Li 2007).

The similarity in chondrite-normalized REE and the primitive mantle-normalized trace element patterns (Fig. 7) and the well comparable Sr–Nd isotopic compositions (Fig. 8) between the diorites and quartz diorites indicate that they are genetically related. The linear or sub-linear trend of selected major and trace elements in the Harker diagrams (Fig. 6) demonstrate that the quartz diorites may have been formed via fractional crystallization of magmas represented by the diorites. Close temporal and spatial relations of diorites and quartz diorites provide additional weight of evidence for their genetic link. For example, the Lingxiang diorite (141.1 ± 0.7 Ma) and the Tongshankou granodiorite (140.6 ± 0.6 Ma; Li et al. 2008) were emplaced simultaneously in close proximity (Fig. 2).

Genesis of the diorites

As discussed earlier, the dioritic rocks in the Daye region were most likely derived from an enriched mantle source. However, direct partial melts of such a source would be of basaltic composition rather than andesitic (Nicholls and Ringwood 1972; Green 1973; Wilson 1989). Although experimental data showed that silica-rich melts might be produced by small degree partial melting of mantle peridotite (Baker et al. 1995), Mg# of such melts in equilibrium with mantle olivine would range from 70 to 80 (Falloon et al. 1997). Much lower Mg# is seen for the Lingxiang and Jinshandian diorites (Table 3), indicating that the rock compositions measured cannot represent primary melts that were in equilibrium with peridotitic mineral assemblages. The low Cr and Ni contents in the samples (Table 3) are compatible with this view. We therefore suggest that the present dioritic rocks represent differentiated liquids that evolved from a parental basaltic magma by fractional crystallization of olivine in depth.

Geochemical modeling was performed to examine the genesis of the Daye diorites by partial melting of peridotitic mantle and subsequent olivine crystallization. During the modeling, three mantle compositions were chosen: (1) DM (depleted mantle) with YbN = 2.9 and (La/Yb)N = 0.5 (McCulloch and Bennett 1994); (2) EM (slightly enriched mantle) with YbN = 3.5 and (La/Yb)N = 6.2 (Graviou et al. 1988), and (3) RS1, a phlogopite- and pargasite-bearing lherzolite sample with YbN = 2 and (La/Yb)N = 11.3 (Menzies et al. 1987). For trace element modeling, we applied the Shaw’s batch melting equation C L  = C O/[D + F(1−D)] (Shaw 1970), where C O and C L are trace element abundances respectively in the source and liquid, D means bulk partition coefficients, and F denotes degree of partial melting. The partition coefficients were taken from the GERM Kd database (Nielsen 2006).

Figure 10a illustrates the curves for partial melting of DM (with 5% spinel and 6% garnet), EM and sample RS1 (5% spinel, 6% garnet plus 2% amphibole and 2% phlogopite). It is clear that partial melting of a depleted mantle, regardless the residual mineral assemblage, cannot generate magmas with La/Yb ratios as high as observed in the Daye diorites. The more likely source thus would be an enriched mantle source. Most samples of the Jinshandian diorite may be explained by 22% partial melting of enriched mantle equivalent in composition to sample RS1, following considerable amounts of olivine fractionation (Fig. 10a). However, sample JS1, which differs from other samples by high CaO, Ba, Hf, Rb, Sr, Ta, U, low Na2O/K2O and presence of normative nepheline (Table 3), corresponds to 20% olivine fractionation from a parental magma generated by 17% partial melting of EM. The Yinzu gabbro-diorite would require 35% fractional crystallization from a parental magma derived from 36% partial melting of an EM source. Samples of the Lingxiang diorite might be explained by variable degree of partial melting of an EM source, with different amounts of olivine fractionation of the partial melt. Coexistence of diorites and broadly coeval basalts, together with abundance of mafic enclaves in the Lingxiang diorite, support the geochemical modeling results.

Fig. 10
figure 10

Chondrite-normalized (Sun and McDonough 1989) (La/Yb) versus Yb diagrams for the Daye diorites and quartz diorites. a Partial melting curves of DM (5% Spinel 6% Gt), EM and sample RS1 (5% spinel, 6% garnet plus 2% amphibole and 2% phlogopite). Labeled tick marks indicate per cent partial melting of spinel- and garnet-bearing model curves and fractional crystallization of olivine (FC Ol). b Partial melting curves of lower crust and tholeiitic oceanic crust leaving garnet or amphibole in the residue, calculated by using Shaw’s batch melting equation (Shaw 1970) and partition coefficients in Martin (1987), Rollinson (1993) and Nielsen (2006). The fields of adakites and low Al TTD are from Drummond and Defant (1990). Other symbols as in Fig. 6. Other symbols as in Fig. 6. See text for explanation. DM depleted mantle, EM enriched mantle, RS1 garnet peridotite sample in Menzies et al. (1987), TTD Trondhjemite–Tonalite–Dacite

Full size image

Other experimental studies demonstrated that diorites might be formed through partial melting of an oceanic crust (Rushmer 1991; Sen and Dunn 1994; Rapp and Watson 1995). The same modeling procedure was applied to test the possibility of this mechanism in the formation of the Daye diorites, assuming the mafic source with (La/Yb)N = 5 and YbN = 6–17 (Fig. 10b). It appears that the Lingxiang diorites may be generated by ~30% partial melting of an oceanic crust transformed into garnet-free amphibolite. However, when applying this degree of partial melting of an amphibolite source (N-MORB: 35.8 ppm Y and 113 ppm Sr; Hofmann 1988) to the batch melting equation of Shaw (1970), the expected melt has 33 ppm Y, 140 ppm Sr, and Sr/Y ratio of 4, which are distinctly different from the Lingxiang diorites (17–23, 634–749, and 30–39 ppm, respectively). Moreover, experimental studies have shown that partial melting of low-K tholeiites under both water-saturated and water-undersaturated conditions generate liquids that are peraluminous or transitional between metaluminous and peraluminous (Beard and Lofgren 1991), which is in contrast with the metaluminous nature of the Lingxiang diorites (A/CNK <0.84, without normative corundum).

In light of the results of geochemical modeling and the Sr–Nd isotope data, we conclude that the Daye diorites were derived from an enriched mantle source. Composition of the source can be calculated by mixing DM (Workman and Hart 2005) with different proportions of low-silica adakites (LSA), high-silica adakites (HSA), Archean TTGs, and sanukitoids (Martin et al. 2005). The best results were obtained for a metasomatised mantle (MM) produced by mixing 90% DM and 10% HSA. Samples of both Jinshandian and Lingxiang diorites and gabbro-diorite of the Yinzu complex have trace element patterns parallel to the MM (Fig. 11), providing an additional argument for this common source. The patterns agree well with the formation of the Daye diorites by ~15–25% partial melting of a slab-melt modified mantle source, followed by 40–70% fractional crystallization of olivine at mantle depth, consistent with results shown in Fig. 10a. Occurrences of ophiolites in the southeastern margin of the Yangtze craton and the 968-Ma slab-derived Xiwan adakites that intruded the ophiolites provide direct geological evidence for an ancient oceanic crust subduction process and hence mantle metasomatism by slab melts and/or fluids (Li and Li 2003).

Fig. 11
figure 11

Primitive mantle normalized (Sun and McDonough 1989) multi-elemental diagrams for the Daye diorites and the metasomatised mantle (MM). The MM is assumed as mixing of 90% deplete MORB mantle (Workman and Hart 2005) with 10% slab-derived melt having high-silica adakite (HSA) composition (Martin et al. 2005). The modeling was based on the Shaw’s batch melting (Shaw 1970) and Rayleigh fractional crystallisation equations (Rayleigh 1896). The partition coefficients are from Martin (1987), Rollinson (1993) and Nielsen (2006). The modeling was performed assuming a two-stage process: (1) partial melting of a MM source, and (2) fractional crystallization of olivine in the resultant magma. The MM source is assumed to have phlogopite (2%), amphibole (2%), garnet (0–5%), and spinel (0–5%) as accessory phases

Full size image

Genesis of the quartz diorites

A striking feature of the quartz diorites is that they have geochemical signatures similar to adakites, such as high Sr, Al2O3, Na2O/K2O, low Y, Yb, Sc, and high Sr/Y and La/Yb ratios (Table 3). Adakites are a class of intermediate to acidic volcanic or plutonic rocks characterized by high Sr (>400 ppm), Sr/Y (>40) and La/Yb (>20), and low Y (<18 ppm) and Yb (<1.9 ppm) (Defant and Drummond 1990). They have been interpreted to be products of partial melting of young subducting oceanic crust, with hybridization of the slab melts with mantle peridotite while they ascend (Kay 1978; Defant and Drummond 1990; Sajona et al. 1993; Peacock et al. 1994). Recent studies have also shown that many igneous rocks possess geochemical features similar to adakites but are apparently unrelated to slab melting. These are termed as adakitic rocks as suggested by Castillo (2006). Alternative mechanisms have been proposed to explain the generation of adakitic rocks. These mainly include (1) partial melting of a peridotitic mantle that has been modified by slab melts (Martin et al. 2005, and references therein), (2) melting of thickened (Chung et al. 2003; Petford and Atherton 1996; Wareham et al. 1997) or delaminated mafic lower crust (Kay and Kay 1993; Xu et al. 2002; Gao et al. 2004), and (3) fractional crystallization of mantle-derived magmas (Castillo et al. 1999; Macpherson et al. 2006).

Some of the quartz diorite samples in the Daye region appear to fit partial melting curves of a tholeiitic source metamorphosed to garnet-free amphibolite to 10% garnet amphibolite (Fig. 10b). These rocks, however, have radiogenic Sr–Nd isotopic compositions (Table 4; Fig. 8), indicating that a slab-melting model is not feasible for their genesis as slab-derived adakites are generally characterized by primitive Sr and Nd isotopes (87Sr/86Sr < 0.7045) (Kay 1978; Defant and Drummond 1990; Bourdon et al. 2002). Lack of coeval arc magmatism in the Daye region also argues against a slab-melting model. Moreover, slab melts would acquire elevated MgO contents and Mg# via interactions with mantle peridotite during their ascent (Martin et al. 2005). The opposite is observed in the Daye quartz diorites, which have much lower MgO (0.6–2.0 wt.%) and Mg# (21–48) (Table 3) relative to slab-derived adakites.

Adakitic rocks by partial melting of a peridotitic mantle wedge whose composition has been modified by reactions with felsic slab-melts were termed as LSA, marked by low silica (56 wt.%), high MgO (5.15 wt.%) and Mg# (61), and elevated Cr (157 ppm) and Ni (103 ppm) contents (Martin et al. 2005). The adakitic rocks in the Daye region show different compositions compared to LSA, containing higher silica (60–71 wt.%) but much lower MgO (0.68–1.96 wt.%), Mg# (21–48), Cr (12–31 ppm) and Ni (5–28 ppm). High-Mg andesites, commonly associated with adakitic rocks by partial melting of mantle peridotite contaminated by slab melts (Tatsumi 2006), are also lack in either Daye or other magmatic regions of the eastern Yangtze craton. These suggest that the present adakitic rocks cannot represent the direct melts of peridotitic mantle metasomatised by slab-derived magma.

Generation of adakitic rocks via partial melting of thickened basaltic lower crust where garnet is a stable and residual phase has been demonstrated by experimental studies of hydrous metabasalt at high P–T conditions (Rapp et al. 1991; Sen and Dunn 1994; Rapp and Watson 1995) and examined in several localities (Atherton and Petford 1993; Muir et al. 1995; Chung et al. 2003). This mechanism was also introduced to explain the genesis of the Yinzu quartz diorite (Wang et al. 2004). The present crustal thickness of the Daye region is merely 30–31 km (Zhai et al. 1992). Considering that the volcanic associations and shallow-seated porphyry ore deposits in the region have been well-preserved (Shu et al. 1992; Li et al. 2008), it is reasonable to assume a magnitude of denudation less than 3–4 km since the early Cretaceous. This indicates that the early Cretaceous crustal thickness in the Daye region must not be more than 40 km, condition insufficient for the lower crust to be transformed into garnet amphibolite or eclogite. In addition, lack of eclogite or eclogite-bearing assemblages throughout the eastern Yangtze craton, either as outcropped blocks or xenoliths entrained in Cretaceous basalts, does dot support existence of eclogitized lower crust and thus its melting as an effective way to generate the adakitic rocks.

Wang et al. (2004) invoked partial melting of delaminated lower crust in the mantle lithosphere or thickened mafic lower crust to explain the genesis of the adakite-like Yinzu quartz diorite and the Tongshankou granodiorite (Fig. 2). Melts of delaminated lower crust will inevitably interact and equilibrate with the mantle peridotite while they ascend, resulting in significantly elevated concentrations of Cr, Ni, and MgO (Rapp et al. 1999; Gao et al. 2004). The low abundances of these elements in the present adakitic rocks are inconsistent with such interactions. The Daye region has long been regarded as a backland area of the Dabie Orogen (Hubei Bureau of Geology, Mineral Resources 1990), indicating that it had much thinner crust than the latter during the early Cretaceous when the adakitic intrusions were formed. This in turn implies that the Daye area was less susceptible to delamination of lower crust than the Dabie Orogen at that time. Interestingly, the 143–142-Ma adakitic intrusions from the core of the Dabie Orogen that contain low to very low Cr (2.1–24 ppm) and Ni (1.4–9.6 ppm) have been demonstrated to be formed by partial melting of thickened lower crust, rather than delaminated lower crust (Wang et al. 2007). This indicates that the lower crust of the Dabie Orogen was not significantly delaminated during the earliest Cretaceous. In this regard, we infer that the adakitic intrusions of the Daye region were less likely to be produced by partial melting of delaminated lower crust.

Preclusion of those three possibilities led us to favor fractional crystallization of mantle-derived parental magmas (Castillo et al. 1999; Macpherson et al. 2006) to explain the genesis of the Daye adakitic rocks. Richards and Kerrich (2007) stressed that the key adakitic geochemical signatures, notably low Y and Yb concentrations and high Sr/Y and La/Yb ratios, are more frequently generated by this mechanism. Selected Harker plots (Fig. 6) confirm that fractional crystallization has been important in the magmatic processes, with removal of Al-, Fe-, Mg-, Ca-, Ti-, P-, Y-, Yb- and V-rich phases such as hornblende, plagioclase, magnetite and apatite. The SiO2 versus Sr/Y and YbN diagrams (not shown), where the values of YbN decrease and Sr/Y increase with increasing SiO2, are consistent with such a mechanism in generating the adakitic quartz diorites. When the most incompatible elements Ba and Rb are plotted against the compatible element Ni (Fig. 12), a fractional crystallization trend becomes more evident. Such a trend is indicative of fractionation of hornblende and magnetite that have partition coefficients for Ni much greater than 1 in intermediate to acidic liquids (Nielsen 2006). Fractionation of hornblende and plagioclase also agrees with the slightly negative Eu anomalies in quartz diorites (Fig. 7b).

Fig. 12
figure 12

Plots of the most incompatible elements Ba and Rb versus the compatible element Ni, indicating that fractional crystallization (FC), rather than partial melting (PM) dominated the genesis of the adakitic quartz diorites of the Daye region. Data of the Yinzu quartz diorite are from Wang et al. (2004) and this study, while the Yinzu gabbro after Wang et al. (2004)

Full size image

To provide more details on the fractional process, we performed geochemical modeling by employing a classical mass-balance equation system that was solved using the algorithm of Störmer and Nicholls (1978). The theoretical modeling was performed assuming the differentiation of a parental magma (C O) into an evolved/differentiated liquid (C L). The accuracy of the adjustment of the theoretical model to the data is expressed by the sum of the squares of the residuals Σr 2 = Σ(m i  − c i )2, where m i and c i denote measured and calculated concentrations of oxide i, respectively. The mineral compositions used in the calculation were from Deer et al. (1983). Because (1) the Yinzu complex (151.8 Ma) consists of gabbro-diorite and quartz diorite that were emplaced at least 10 Myr earlier than the other intrusions (141–132 Ma), and (2) the Jinshandian diorite is somewhat different in geochemistry from the Lingxiang diorite and the Yinzu gabbro-diorite (Figs. 5, 10), we divided the diorites into two subsets, with the Lingxiang diorite plus Yinzu gabbro-diorite as subset 1 and the Jinshandian diorite as subset 2. Accordingly, we modeled the formation of the Yinzu quartz diorite by fractional crystallization of the Yinzu gabbro-diorite (group A) and formation of the Tieshan quartz diorite by fractional crystallization of the Jinshandian diorite (group B), separately. To model the behavior of the group A, sample 01YZ-4-1 (Wang et al. 2004) was chosen as C O and YZ3 as C L, whereas JS3 and TS2 are C O and C L for group B. The composition of both C O and C L were recalculated to 100% on anhydrous basis; but MnO was not considered as its concentration is negligible (Table 3).

The results for the modeling of both groups are shown in Table 5. The best fit of computed model to the analytical data is obtained for crystallization of a mineral assemblage of hornblende, plagioclase, magnetite and traces of apatite for group A, and of hornblende, plagioclase, K-feldspar, magnetite and apatite for group B; the degree of fractional crystallization is 64 and 56%, respectively. In group B, the behavior of Zr is accounted for only when small amounts (0.03%) of zircon are added in the fractional phases. The calculated cumulates are consistent with the inferred fractional assemblages from the Harker diagrams (Fig. 6) and agree with the major phenocryst phases observed in the quartz diorites. The reliability of the modeling is indicated by the calculated oxide contents (C L′) that agree well with the observed values of the selected samples YZ3 and TS2, with statistical error (Σr 2) of 0.2 and 0.7 (Table 5). The second step in quantification of fractional crystallization is to reintroduce the computed modal compositions of the cumulate and the degree of fractional crystallization in trace element modeling. The equation chosen for fractional crystallization is that of Rayleigh (1896): C L  = C O × F (D − 1), where C L is the calculated concentration of a trace element in the differentiated liquid, C O is the concentration of a trace element in the parental magma, F = 1 − FC (FC is the degree of fractionation, with 0 < FC ≤ 1) and D is the bulk distribution coefficient (Rollinson 1993). For 64% (group A) and 56% (group B) fractional crystallization, the calculated trace element contents show perfect agreement with the analytical values (Table 5). The modeling results are corroborated by the very good fit of the present samples to fractional crystallization curves in the Y versus Sr/Y and normalized Yb versus La/Yb diagrams (Fig. 13). Coexistence of adakitic quartz diorite and non-adakitic gabbro-diorite in the Yinzu complex and presence of mafic enclaves in the Tieshan and Lingxiang intrusions, consisting of plagioclase, hornblende, biotite, pyroxene, and K-feldspar, and accessory titanite, magnetite, and apatite, lend additional support to fractional crystallization in the formation of quartz diorites.

Table 5 Major and trace element modeling for the Daye intrusions
Full size table
Fig. 13
figure 13

Y versus Sr/Y (a) and YbN versus (La/Yb)N (b) diagrams illustrating the genesis of the quartz diorites via fractional crystallization (FC) of mainly hornblende plus plagioclase from parental dioritic magmas similar in composition to the Yinzu, Lingxiang or Jinshandian diorites. In both figures, curves of mantle and crustal assimilation and fractional crystallisation (AFC) were calculated following DePaolo (1981). For mantle AFC we used a MORB-like (Sun and McDonough 1989) magma and the lower continental crust (Rudnick and Fountain 1995) as contaminant. For crustal AFC, we used the Yinzu gabbroic–diorite and the Jinshandian diorite as original liquid and the lower continental crust as contaminant. In the AFC models, we considered mass-assimilation ratios of 2 and 1.5 for mantle and crustal AFC, and 80% (20%) amphibole and 60% (40%) plagioclase as fractionating phases, respectively

Full size image

Figure 14 illustrates the petrogenetic model for the plutonic rocks of the Daye region. We envisage four stages in their genesis: (1) an early episode of hybridization, in which the depleted mantle was metasomatised by slab melts (HAS); (2) ~15–40% partial melting of this metasomatised mantle, generating basic magmas that evolved by (3) 40–70% fractional crystallization of olivine at mantle depth, forming the Daye diorites; and (4) low pressure fractional crystallization of cumulates composed of amphibole + plagioclase ± K-feldspar + magnetite + apatite ± zircon, resulting in the adakitic quartz diorites.

Fig. 14
figure 14

Schematic diagram showing the successive petrogenetic processes that gave rise to the plutonic rocks of the Daye region. MM metasomatised mantle; DM depleted mantle; FC fractional crystallization; PM partial melting; Ol olivine; Hb hornblende; Pl plagioclase; Kf K-feldspar; Mgt magnetite; Apt apatite. See text for explanation

Full size image

Regional synthesis and geodynamic implications

Combined with previous isotopic ages, the present geochronological data suggest that pervasive magmatism in the Daye region and throughout the eastern Yangtze craton took place coevally in the latest Jurassic to early Cretaceous (Fig. 9). Voluminous and broadly synchronous magmatism is also widespread in the eastern part of the Cathaysia block (Zhou and Li 2000), in the Dabie-Sulu Orogen (Ma et al. 1998; Jahn et al. 1999; Xu et al. 2007), and in the northern and eastern margins of the North China craton (Wu et al. 2005a, b), forming a huge NNE-striking magmatic belt in eastern China. The geochemical and Sr–Nd isotope data of the magmatic rocks, together with the coexistence of coeval felsic to mafic rocks, indicates large scale melting of subcontinental lithospheric mantle during the Late Mesozoic time. Mafic granulite xenoliths entrained in Cenozoic basalts from several localities around eastern China are considered as analogue of mantle-derived magmas ponded at the base or lower part of crust (Yu et al. 2003; Huang et al. 2004). SHRIMP zircon U–Pb dating has constrained crystallization of the protolith of the granulite xenoliths mainly in the 159–120 Ma range (Fan et al. 1998; Wilde et al. 2003; Huang et al. 2004), coincident with the timing of extensive magmatism in the eastern Yangtze craton and the entire eastern China continent. Widespread occurrences of strong, sub-horizontal, continuous layered seismic reflectors in the lower crust beneath the eastern Yangtze craton (20–33 km depth), interpreted as reflecting basaltic underplating (Lü et al. 2004), provide further evidence for extensive partial melting of the lithospheric mantle.

Large-scaled mantle melting requires abnormally high heat flow during the early Cretaceous, which was most likely envisaged in an extensional environment. Presence of syenite, A-type granite, shoshonitic basalts, and metamorphic core complexes that have been dated between 131 and 120 Ma (Xing and Xu 1994; Wang et al. 2001; Liu et al. 2002; Grimmer et al. 2003; Wang et al. 2006; Yan et al. 2008), implies that large areas of the eastern Yangtze craton was dominated by lithospheric extension during that time period. Nevertheless, extension should have been initiated since the earliest Cretaceous or latest Jurassic, based on geochronological data of numerous mafic to acidic intrusions (Jahn et al. 1999; Xu et al. 2004; Wang et al. 2006; Yan et al. 2008). As for other regions of eastern China, such an extension event is indicated by pervasive Au–Cu–Fe–Mo mineralization (Li et al. 2003; Mao et al. 2004), wide distribution of faulted basins and metamorphic core complexes (Tian et al. 1992; Davis et al. 2002; Liu et al. 2005), and transtensional movement of the continental Tan-Lu Fault (Zhu et al. 2005). The mechanism driving the Jurassic–Cretaceous extension may have been the removal and thinning of the lithospheric mantle and upwelling of asthenosphere, as indicated by geophysical data (Chen et al. 2008) and studies of mantle-derived xenoliths sampled by Paleozoic and Cenozoic basalts (Menzies et al. 1993; Xu et al. 2000).

Conclusions

Extensive plutonic magmatism in the Daye region started in the latest Jurassic (ca. 152 Ma) and peaked in the early Cretaceous (141–132 Ma), which was followed by volcanic eruption and dyke emplacement in the 130–113 Ma interval. The plutonic rocks can be grouped as diorites and quartz diorites that show close spatial–temporal relationships. Geological, geochemical, and Sr–Nd isotopic data indicate that the diorites were formed by partial melting of enriched lithospheric mantle that had been modified by slab melts and/or fluids, followed with variable degree of olivine fractionation at mantle depth. The quartz diorites have geochemical signatures similar to adakites, which were most likely generated by fractional crystallization of an assemblage consisting of hornblende, plagioclase, magnetite, titanite and apatite, from parental dioritic magmas represented by the diorites. This study thus confirms that fractional crystallization of mantle-derived magmas may have been a more common mechanism in the formation of adakitic rocks than previously thought. The fact that voluminous magmas were emplaced broadly coevally over the eastern Yangtze craton and the eastern China continent indicates a consistent geodynamic setting responsible for their formation. A synthesis of geological, geochronological and geochemical data demonstrates that the large-scaled magmatism was driven by extensive lithospheric extension, which was most likely caused by lithosphere thinning.