Introduction

The North China Craton (NCC) is well known for the Early Cretaceous giant igneous event, which culminated at approximately 125 Ma and is related to the subduction of the Paleo-Pacific Plate (Wu et al. 2005a, b; Sun et al. 2007). As indicated by the Hess Rise and Japanese island chains, the subduction direction of the Paleo-Pacific Plate changed by approximately 80° in 125–120 Myr ago (Wessel and Kroenke 1997; Koppers et al. 2001, 2003; Sun et al. 2007). In addition, large-scale lode gold (Au) mineralization formed in 125–120 Myr ago in the NCC and are generally regarded as typical orogenic lode Au deposits (Wang et al. 1998; Groves et al. 1998; Goldfarb et al. 2001; Qiu et al. 2002; Yang and Zhou 2002; Li et al. 2005).Moreover, destruction of the NCC occurred in the Mesozoic with a peak age of approximately 125 Ma (Wu et al. 2005a; Zhu et al. 2011). These events indicate that 125 Ma is an important point in time, raising the question of whether the tectonic setting was the same before and after 125 Ma.

The NCC is also one of the oldest known cratons, and its location within the Eurasian continental collage strongly favors investigation into lithospheric delamination (Gao et al. 2004; Wu et al. 2005a, b). Recent research has suggested that the NCC has undergone extensive lithosphere destruction since the Paleozoic (Wu et al. 2005a, b), as evidenced by data from early Paleozoic diamondiferous kimberlites within the Mengyin and Fuxian areas and Cenozoic mantle xenoliths in this region, all of which provide evidence for a reduction in the thickness of the subcontinental lithospheric mantle in this area from a maximum value of ~ 200 km to a thickness of 60–120 km (Fan and Menzies 1992; Menzies et al. 1993; Menzies and Xu 1998; Griffin et al. 1998; Wu et al. 2005a, b). Destruction n of the NCC is supported by geochronological and petrogenetic data from igneous rocks (Gao et al. 2004; Wu et al. 2005a; Wang et al. 2007; Cai et al. 2013; Yang and Santosh 2015; Dai et al. 2016) and xenoliths in this region (Xu et al. 2013; Wei et al. 2015a, b; Tang et al. 2014), high-resolution seismic wave tomographic modeling of crust and mantle in this region (Huang and Zhao 2006; Santosh 2010; Zhu et al. 2011, 2012), gravity, geoid, and topographic data (Xu et al. 2016) for the NCC, and the presence of a series of extensional structures within this region (Lin et al. 2011; Shen et al. 2011; Liu et al. 2011; Li et al. 2012; Charles et al. 2013; Zhu et al. 2015; Fu et al. 2016). However, although significant knowledge has been gained from previous research into the delamination of the NCC, there are still hot debates such as the driving force of delamination (Gao et al. 2009).

This study aims to address the aforementioned problems by investigating the characteristics of Early Cretaceous magmatism at approximately 125 Ma within the NCC. Herein, we present new petrological, whole-rock geochemical, zircon U–Pb geochronological, and zircon Lu–Hf isotopic data for Early Cretaceous intrusive rocks from the Wulong region of the Liaodong Peninsula in the NCC.

Geological setting and sample descriptions

Geologically, eastern China is split into the Xing’an–Mongolian Orogenic Belt and the NCC in the north, the Dabie–Sulu Orogenic Belt in the central part, and the Yangtze Craton and Southeastern China Orogenic Belt in the south (Fig. 1a) (Wu et al. 2005a, b). The NCC is one of the oldest known cratons globally and contains crustal rocks as old as 3.8 Ga (Liu et al. 1992). This craton records Phanerozoic subduction events and collisions with adjacent blocks (Windley et al. 2010; Ma et al. 2016), including southward-directed Paleozoic subduction of the Paleo-Asian oceanic plate (Xiao et al. 2003), deep Triassic subduction of Yangtze Craton continental crust (Li et al. 1993), southward-directed Devonian–Cretaceous subduction of the Mongol–Okhotsk oceanic plate (Kravchinsky et al. 2002; Cogné et al. 2005; Donskaya et al. 2013), and Mesozoic–Cenozoic–present-day subduction of the (Paleo-)Pacific Plate (Müller et al. 2013; Xu 2014).

Fig. 1
figure 1

a Simplified geological map of eastern China, showing major tectonic units. b Tectonic subdivisions of the Liaodong Peninsula; c geological map showing distribution of Mesozoic intrusions in the Liaodong Peninsula (after Wu et al. 2005a, b)

The Liaodong Peninsula is located in the northeastern edge of the NCC, and is bounded to the west by the NNE–SSW-trending Tan–Lu Fault zone, to the east by the western margin of the Pacific Plate, and to the north by the Xing’an–Mongolian Orogenic Belt (Fig. 1a). The Liaodong Peninsula contains three tectonic units, namely, the Archean Liaonan Block in the south, the Archean Liaobei Block in the north, and the central Paleoproterozoic Liaoji Orogenic Belt (Fig. 1b) (Wu et al. 2005a, b). Large quantities of Mesozoic intrusive rocks have been identified, including Jurassic granites, Cretaceous intrusions and a few Triassic syenite in NCC (Fig. 1c) (Wu et al. 2005a, b; Yang et al. 2007). Moreover, recently some Early-Late Triassic mafic dykes are found (Tang et al. 2018; Wu et al. 2019a, b; Fang et al. 2020). In addition, the Liaodong Peninsula is a famous gold province (Wang et al. 2018). Among those gold deposits, the Wulong orefield is an important mining area in the Liaodong Peninsula.

The Wulong orefield mainly includes the Wulong and Sidaogou gold deposits, which is located about 30 km west of Dandong, Liaoning Province. The strata of Wulong region are relatively simple and occur sporadically among the widespread in intrusions, including Paleoproterozoic Liaohe Group (Pt1) and Quaternary (Q) (Fig. 2). The magmatic intrusion was intense and developed from mafic to acid rocks, mainly including late Jurassic biotite monzogranite, late Jurassic two-mica monzogranite, late Jurassic pegmatite, early Cretaceous quartz diorite, early Cretaceous porphyritic granodiorite, and so on (Fig. 2).

Fig. 2
figure 2

Detailed geological map of the Wulong region

For our study, we examined 6 groups of intrusions from the Sanguliu and Wulongbei pluton in Wulong region. Five groups of acidic rocks and one group of mafic rocks were identified and the sampling locations are labeled in Fig. 2. The granodiorite (D18-B) from the Sanguliu (Fig. S1a, b in the electronic supplementary material (ESM)) contains plagioclase (45 vol%), K-feldspar (15 vol%), quartz (25 vol%), biotite (10 vol%), and amphibole (5 vol%) with medium-fine subhedral granular texture and massive structure. In addition, the mafic microgranular enclaves (MMEs) were found in the granodiorites (Fig. S1a, c, d in the ESM).

The Wulongbei pluton is exposed to the northeast of the Wulong region. The pluton covers approximately 450 km2 and intrudes into the Jurassic Wulong pluton. According to field observations, the Wulongbei pluton can be divided into three periods. The early period products include gabbro, (porphyritic-like) granodiorite, and (porphyritic-like) monzogranite. Note that the porphyritic-like monzogranite constitutes the main portion of rocks in the Wulongbei pluton. Quartz syenite and (porphyritic-like) syenogranite comprise the second period and are mainly distributed in the north of the Wulongbei pluton, while quartz diorite is the third-period product that invades all the aforementioned rocks. The gabbro (D18-E) has a gabbroic texture and mainly contains plagioclase (65 vol%) and clinopyroxene (30 vol%) with opaque mineral (5 vol%) (Fig. S1e, f in the ESM). The porphyritic-like biotite monzogranite (D18-M) with gray-white, porphyritic-like structure and massive structure, consists of plagioclase (35 vol%), K-feldspar (30 vol%), quartz (25 vol%), biotite (8 vol%), and amphibole (2 vol%) (Fig. S1g, h in the ESM). The porphyritic-like monzogranite (D18-N) has porphyritic-like texture and massive structure, and is composed of plagioclase (35 vol%), K-feldspar (35 vol%), quartz (25 vol%), biotite (3 vol%), and amphibole (2 vol%) (Fig. S1i, j in the ESM). The porphyritic-like syenogranite (D18-P) with gray-white, porphyritic-like structure and massive structure, consists of plagioclase (20 vol%), K-feldspar (50 vol%), quartz (25 vol%), biotite (3 vol%), and amphibole (2 vol%) (Fig. S1k, l in the ESM). The syenogranite (D18-Q) consists of plagioclase (22 vol%), k-feldspar (55 vol%), quartz (20 vol%), biotite (3 vol%). Moreover, it shows gray-white, medium-fine subhedral granular texture, massive structure (Fig. S1m, n, o in the ESM).

Analytical methods

Zircon U-Pb dating

Separation of zircon grains were operated by using a combined heavy liquid and magnetic technique. Then, zircons were handpicked under a binocular microscope before being mounted in epoxy resin and polished to expose the center of individual zircons. Cathodoluminescence (CL) imaging was used to determine the internal structures of zircons and to choose the target areas for laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) analysis. All zircons were imaged using transmitted and reflected light microscopy, and CL. A spot diameter of 32 µm was used for each analysis, and the analytical procedures followed Yuan et al. (2004). Furthermore, U-Pb ages were calculated using Isoplot (Ludwig 2003).

Whole‐rock geochemistry analysis

For whole-rock geochemical analysis, the samples, after the removal of weathering surfaces, were crushed in an agate mill to ~ 200 mesh. X-ray fluorescence (XRF; PW1401/10) using fused-glass disks and inductively coupled plasma mass spectrometry (ICP-MS) (Agilent 7500a with a shield torch) were used to measure the major and trace elements compositions, respectively. The analytical precision for major elements is better than 5%. For trace elements, the relative deviation is less than 5% when their composition is more than 10 µg/g, otherwise, the relative deviation is less than 10%.

Hf isotopes

The analysis of in-situ zircon Lu-Hf isotope was based on LA-ICP-MS zircon dating and refer to the cathodoluminescence (CL) images at the place of original age determination or around. Zircon Hf isotopic analyses were undertaken by Nepture-plus MC-ICP-MS and NewWave UP213 laser system. Instrument conditions and data acquisition methods have been suggested by Wu et al. (2006). The pulse frequency, denudation diameter and laser beam energy are respectively 8 Hz, 45 µm and 16 J·cm− 2.

Analytical results

Zircon U-Pb dating

Six groups samples in the Wulong region were selected for zircons U-Pb ages. All zircons have clear magmatic oscillatory zoning and possess high Th/U ratios (0.16–2.86), indicating a magmatic origin (Hoskin 2003). Zircon U-Pb data are presented in Table S1 in the ESM and representative CL images are shown in Fig. 3. As mentioned above, 125 Ma is an important point in time. Therefore, six intrusive rock are divided into two groups according to the crystallization ages.

Fig. 3
figure 3

LA-ICP-MS U-Pb concordia diagrams, representative zircon CL images (red circles indicate the locations of LA-ICP-MS U-Pb analyses, and blue circles mark the location of Lu-Hf isotopic analyses) and weighted average diagrams

Group 1

Group 1 has older crystallization ages, including the granodiorite, the gabbro, the porphyritic-like biotite monzogranite and the porphyritic-like monzogranite. Twenty-five spots from samples D18-B-N of the granodiorite from the Sanguliu, Wulong region show similar apparent 206Pb/238U ages, and form a coherent group, which yields a weighted mean 206Pb/238U age of 129 ± 1 Ma (mean standard weighted deviation (MSWD) = 0.16) (Fig. 3a). The magmatic zoning (Fig. 3a) and high Th/U ratios (0.79–1.87) (Table S1 in the ESM) of these zircons indicates that 129 ± 1 Ma can be interpreted as the crystallization age of the granodiorite.

Thirty zircons from samples D18-E-N, gabbro from the Wulong region, show well-defined magmatic zoning and Th/U ratios (0.16–2.87) (Fig. 3b, Table S1 in the ESM), suggesting a magmatic origin. Twenty-nine analyses that plot on the concordia form a tight cluster yielding a weighted mean 206Pb/238U age of 126 ± 1 Ma (MSWD = 0.35) (Fig. 3b). This age is considered the crystallization age of the gabbro.

Twenty-five spots from 25 zircons of the porphyritic-like biotite monzogranite (D18-M-N) from the Wulong region plot on or near the concordia forming a tight cluster with a weighted mean 206Pb/238U age of 127 ± 1 Ma (MSWD = 0.14) (Fig. 3c). Their magmatic oscillatory zoning and relatively high Th/U ratios (0.27–2.07) (Fig. 3c, Table S1 in the ESM) suggest that 127 ± 1 Ma can be interpreted as the age crystallization age of porphyritic-like biotite monzogranite.

Twenty zircon grains, show well-defined magmatic zoning, obtained from the porphyritic-like monzogranite (D18-N-N) were used for the analysis. The measured Th and U concentrations of the grains vary from 702 to 9514 ppm and from 860 to 5947 ppm, respectively, while the Th/U ratios ranged between 0.35 and 1.69 (Table S1 in the ESM), indicating a magmatic origin. The analytical results for of the grains plot on or near the U-Pb concordia line, yielding a weighted-mean 206Pb/238U age of 126 ± 1 Ma (MSWD = 0.11) (Fig. 3d), which is interpreted to present the crystallization age of the porphyritic-like monzogranite.

Group 2

Group 2 has a bit younger crystallization ages, including the porphyritic-like syenogranite and syenogranite. The 28 zircon analyses(D18-P-N) yield a weighted mean 206Pb/238U age of 123 ± 1 Ma (MSWD = 0.29) (Fig. 3e). The well-preserved magmatic zoning (Fig. 3e) and rather high Th/U ratios (0.64–1.66) (Table S1 in the ESM) of these zircons suggest that this age can represent the best estimate for the crystallization age of the porphyritic-like syenogranite.

Zircons from the syenogranite (D18-Q-N) are 100–150 µm long with width/length of ~ 1:2 − 1:3. They have changing Th (290–3326 ppm), U (357–2361 ppm), combined with Th/U ratios of 0.75–1.91 (Table S1 in the ESM), are in accord with magmatic origins. There are altogether 29 analyses and yield a mean age of 124 ± 1 Ma with a MSWD of 0.17 (Fig. 3f). We interpreted 124 ± 1 Ma as the crystallisation age of the the syenogranite.

Geochemistry

Group 1

The results of major and trace element analyses are presented in Table 1. The gabbro samples have low SiO2 (47.4–48.0 wt%) contents, but high MgO (7.67–9.17 wt%) contents and Mg# values of 74–75. These gabbro samples also show relatively high Al2O3 (18.3–19.9 wt%) and CaO (12.7–13.0 wt%) contents, while low FeOT/MgO ratios (0.83–0.94). Correspondingly, the samples mainly lie within sub-alkaline gabbro field (Fig. 4a). In the SiO2 vs. FeOT/MgO diagram, all samples plot in the boundary between tholeiitic series and calc-alkali series (Fig. 4c), and classified as medium-K calc-alkaline series in the K2O vs. SiO2 diagram (Fig. 4d).

Table 1 Chemical compositions of rocks from the Liaodong Peninsula
Fig. 4
figure 4

Major-element geochemical plots for the rocks. a Na2O + K2O vs. SiO2 (after Wilson 1989; the rock series boundary is after Irvine and Baragar 1971); b A/NK vs. A/CNK (after Maniar and Piccoli 1989), A/CNK = Al2O3/(CaO + Na2O + K2O), A/NK = Al2O3/( Na2O + K2O); c FeOT/MgO vs. SiO2(after Miyashiro 1974); d K2O vs. SiO2 (after Gill 1981). the same symbols will be used in Figs. 789, 10 and 11 below

The gabbro samples have low total REE (35.68–38.05 ppm) contents and show slight LREE enrichment((La/Yb)N = 6.61–8.20) (Fig. 5a). In addition, the samples have negative Eu anomalies, with Eu/Eu* values of 0.70–0.73. In the primitive mantle-normalized trace elements spider diagram (Fig. 5b), the samples of gabbro are strongly enriched in large ion lithophile elements (LILE) (Rb, K, Pb and Sr), U and depleted in high-field strength elements (HFSE) (Nb, Ce, P, and Ti). Moreover, the gabbro samples have relatively low Cr (42.07–54.31 ppm) and Ni (17.07–18.87 ppm) contents.

Fig. 5
figure 5

(a, c) Chondrite-normalized REE-pattern (normalization values after Boynton 1984) and (b, d) primitive-mantle normalized trace elements spider (normalization values after Sun and Mc Donough 1989) diagrams

The granitoids (the granodiorite, the porphyritic-like biotite monzogranite and the porphyritic-like monzogranite) samples have a relatively narrow range of SiO2 (67.2–70.5 wt%) and Al2O3 (15.1–16.0 wt%). These samples have K2O = 3.35–4.65 wt% and Na2O = 3.46–4.70 wt%, with high K2O/Na2O (0.71–1.34) ratios. While they have variable MgO contents (0.51–1.64 wt%), with Mg# values of 30–53. In general, they belong to the sub-alkaline series (Fig. 4a). The molar Al2O3 / (CaO + NaO2 + K2O (A/CNK) ratios range from 0.96 to 1.17, showing metaluminous-weakly peraluminous characteristics (Fig. 4b). And the samples mainly lie within the high-K calc-alkaline field on the K2O vs. SiO2 diagram (Fig. 4d).

The granitoids (the granodiorite, the porphyritic-like biotite monzogranite and the porphyritic-like monzogranite) samples have relatively high total REE (111–197 ppm) contents and present parallel chondrite normalized REE diagram characterized by relative enrichments of LREE ((La/Yb)N = 16.52–35.36) and nearly flat HREE with a slight Eu anomalies (Eu/Eu*=0.69–1.13, average of 0.89) (Fig. 5a). On the primitive mantle-normalized trace elements spider diagram, they show prominent negative Nb, Ta, P and Ti anomalies, and positive K and Pb anomalies (Fig. 5b). Specifically, the porphyritic-like monzogranite is enriched in Sr (343–537 ppm) and have high Sr/Y ratios (52.9–83.2).

Group 2

The porphyritic-like syenogranite and the syenogranite have high SiO2 (71.2–76.0 wt%), K2O (4.04–4.93 wt%), Na2O (3.60–4.27 wt%), and total alkali (Na2O + K2O = 8.03–8.59 wt%) content. In addition, the K2O/Na2O ratios ranged from 0.95 to 1.36, and the A/CNK ratios are determined to be between 0.98 and 1.07. However, low content levels of MgO (0.14–0.47 wt%), with Mg# ranging from 21 to 32, are observed. In the TAS diagram, all of the sample plots are located in the sub-alkaline series field (Fig. 4a) and are subsequently classified as metaluminous to weakly peraluminous on the A/NK vs. A/CNK diagram (Fig. 4b).

The porphyritic-like syenogranite and the syenogranite have coherent REE patterns characterized by relative enrichments of LREE ((La/Yb)N= 14.41–34.64) and nearly flat HREE, with positive Eu anomalies (Eu/Eu* = 1.01–1.45) (Fig. 5c). On the primitive mantle-normalized trace elements spider diagram, they show remarkable enrichments of LILE (such as Rb, K, Pb) relative to HFSE and LREE, with negative Ta, Nb and Ti anomalies (Fig. 5d).

Zircon Lu-Hf isotopes

Based on the zircon U-Pb dating from LA-ICP-MS, six samples are analyzed for Lu- Hf isotopic of zircon. The analytical data of zircon Lu-Hf isotopes is shown in Table 2; Fig. 6. All measurement points have relatively low 176Lu/177Hf ratios (0.000472–0.001876), indicating that the zircon had no significant accumulation of radioactive Hf after its formation. So, the 176Lu/177Hf ratios can represent the Hf isotope composition of the of the system at the time of its formation (Wu et al. 2007).

Table 2 Lu-Hf isotopic compositions of ziircon grains from rocks from the Liaodong Peninsula
Fig. 6
figure 6

εHf(t) vs. age (Ma) diagrams, compared with Zijinshan alkaline intrusive complex (Ying et al. 2007), Longbaoshan quartz syenite (Lan et al. 2011), Laiwu quartz monzonite (Lan et al. 2013) and Qianshan A-type granite (Yang et al. 2006)

Ten zircon grains from the granodiorite (D18-B-N, 129 Ma) are analyzed for Lu-Hf isotopic compositions and yield εHf(t) values between − 21.0 and − 17.7, with two-stage Hf model ages (TDM2) from 2.29 to 2.50 Ga. Ten zircon grains from the gabbro (D18-E-N, 126 Ma) yield εHf(t) values of -14.3 to -9.4, and old TDM2 model ages (1.77 to 2.08 Ga). Ten zircon grains from the porphyritic-like biotite monzogranite (D18-M-N, 127 Ma) yield variable εHf(t) values (-16.7 to -11.1) and old TDM2 model ages (1.88 to 2.23 Ga). Analyses of ten zircon grains from the porphyritic-like monzogranite (D18-N-N, 126 Ma) yield εHf(t) values ranging from − 19.3 to -13.5, with TDM2 model ages from 2.03 to 2.39 Ga. Ten zircons of the porphyritic-like syenogranite (D18-P-N, 123 Ma) have εHf(t) values of -15.4 to -11.8, and their TDM2 model ages fall between 1.92 and 2.14 Ga. Analyses of ten zircon grains from syenogranite (D18-Q-N, 124 Ma) yield low εHf(t) values (-17.5 to -14.4), with similarly TDM2 model ages (2.08 to 2.27 Ga).

Discussion

Petrogenesis

In general, granites can be divided into the subdivisions S-type, I-type, and A-type. (Chappell and White 1974, 1992; Whalen et al. 1987). The following unfavorable characteristics indicate that the S-type granites affinity required to be excluded. First, S-type granites normally show strong peraluminous features (Chappell 1999), while the Wulong granitoids are metaluminous to weakly peraluminous with A/CNK < 1.1 (Fig. 4b). Second, the studied granitoids lack Al-rich minerals, such as cordierite and muscovite (Fig. S1g,h,k,i in the ESM), which is inconsistent with S-type granitoids. Moreover, they have low P2O5 (0.04–0.07 wt%) content, and P2O5 shows a negative correlation with SiO2 (Fig. 7d) (Chappell and White 1992).

Fig. 7
figure 7

a Zr vs. 10,000 Ga/Al discrimination diagram (after Whalen et al. 1987); b Nb vs. 10,000 Ga/Al discrimination diagrams (after Whalen et al. 1987); c FeOT vs. (Zr + Nb + Ce + Y) discrimination diagram (after Whalen et al. 1987); d P2O5 vs. SiO2 diagram

The porphyritic‐like monzogranite and gabbro

The porphyritic-like monzogranite samples contain high contents of Sr (435–537 ppm) and low contents of Y (5.83–9.15 ppm) and have high Sr/Y (58.7–83.2) and La/Yb (39.3–51.4) ratios. These features suggest that the porphyritic-like monzogranite has adakitic affinities (Defant and Drummond 1990). Moreover, in the two widely used (La/Yb)N vs. YbN (Fig. 8a) and Sr/Y vs. Y (Fig. 8b) discrimination diagrams, all samples plot into the field of adakite. Therefore, this study identified the porphyritic-like monzogranite as typical adakitic granite.

Fig. 8
figure 8

a LaN/YbN vs. YbN diagram (after Defant and Drummond 1990 and Wang et al. 2007); Sr/Y-Y (Martin 1986; Wang et al. 2007); c K2O/Na2O vs. K2O diagram (after Wang et al. 2007); d La/Yb vs. La diagram for the porphyritic-like monzogranites (after Wang et al. 2007)

Adakites were initially interpreted to have formed from magmas generated by the partial melting of subducting oceanic crustal material and are typically enriched in Na relative to K (Na2O/K2O > 2) (Defant and Drummond 1990). The porphyritic-like monzogranite is K-rich (K2O/Na2O = 1.01–1.34) and therefore differ from typical slab-derived adakites. K-rich adakites have been identified in eastern China (Zhang et al. 2001; Xu et al. 2002), the Songpan-Garzê Fold Belt of the eastern Tibetan Plateau (Xiao et al. 2007; Zhang et al. 2006), the Dabieshan Orogen (Xiao and Clemens 2007; Wang et al. 2007) and Cordillera Blanca (Atherton and Petford 1993; Petford and Atherton 1996). The porphyritic-like monzogranite is similar to the early Cretaceous K-rich adakites in Dabieshan Orogen, especially Tiantangzhai adakitic granites (Fig. 9). Previously, genesis models were constructed in order to account for the generation of K-rich adakites, which included: the partial melting of slab melt-modified peridotitic mantle wedge material (Moyen et al. 2001; Martin et al. 2005), the partial melting of delaminated lower-crustal material (Gao et al. 2004), assimilation and fractional crystallization process (AFC) (Castillo et al. 1999; Macpherson et al. 2006), magma mixing (Guo et al. 2007; Streck et al. 2007), and the partial melting of thickened lower crust (Petford and Atherton 1996; Chung et al. 2003). The porphyritic-like monzogranite samples have relatively low Mg# values and contain low contents of MgO, Cr (average of 32 ppm), and Ni (average of 9 ppm) (Table 1), all of which precludes a melt–mantle interaction type origin for these intrusions, including the partial melting of slab melt-modified peridotitic mantle wedge material and the partial melting of delaminated lower-crustal material (Martin 1999; Rapp et al. 1999; Wang et al. 2007; Yin et al. 2017). The porphyritic-like monzogranite samples have weakly Sr positive anomalies without obvious Eu anomalies (Fig. 5a, b), which is not consistent with plagioclase fractional crystallization (Wilson 1989; Wang et al. 2007).Moreover, they lack any evident depletion in middle REE (Fig. 5a), suggesting that amphibole fractionation did not play an important role in their genesis (Gromet and Silver 1987). In addition, lacking clear trends of fractional crystallization (Fig. 8d) also suggests that the fractional crystallization could not be main factor responsible for genetic models of porphyritic-like monzogranite. Likewise, magma mixing also cannot account for the genesis of the porphyritic-like monzogranite. On the one hand, the porphyritic-like monzogranite samples have limited variations in Lu-Hf isotopic compositions, which are different from granodiorite (magma-mixing) (Fig. 6). On the other hand, there are not mafic microgranular enclaves within porphyritic-like monzogranite.

Fig. 9
figure 9

a Nb/Ta vs. Zr/Sm diagram (after Foley et al. 2002 and Guan et al. 2014); b Sm/Yb vs. Sm diagram (after Aldanmaz et al. 2000); c Th/Yb-Ba/La diagram (after Hanyu et al. 2006)

Experimental petrology has indicated that adakitic magmas can be produced by the partial melting of a thickened region of the crust (> 50 km) (Petford and Atherton 1996; Chung et al. 2003; Xiong et al. 2005). The porphyritic-like monzogranite samples display a compositional trend of partial melting (Fig. 8d), indicating that this process was responsible for their genesis (Xiao and Clemens 2007; Wang et al. 2007) (Fig. 9). Furthermore, they provide evidence of the presence of residual garnet (high Sr and low Yb contents and HREE depletions), suggesting that they formed under high-pressure conditions. These suggest that the formation of the porphyritic-like monzogranite involved lower-crustal material within a thickened region of the crust. In addition, the Nb/Ta ratios of porphyritic-like monzogranite (15.51–19.23, average of 17.7) are higher than the values expected for the primitive mantle (16.7) (Condie 2005). Adakitic magmas with high Nb/Ta ratios are generally associated with the partial melting of rutile-bearing eclogite, as Ta has higher rutile–melt distribution coefficient values compared with Nb (Foley et al. 2002; Rapp et al. 2003; Xiong et al. 2005). In the Nb/Ta vs. Zr/Sm diagram, their high Nb/Ta ratios and negative Nb-Ta anomalies suggest that porphyritic-like monzogranite was probably generated by partial melting of rutile-bearing eclogites like Tibetan Plateau thickened lower crustal-derived adakitic rocks (Fig. 9a). As a consequence, we have an opinion that the porphyritic-like monzogranite was produced through partial melting of thickened eclogitic lower crust (Figs. 8a, b and c and 9a and b).

The gabbro samples have high Mg# (74–75) values and relatively low Cr (42–54 ppm) and Ni (17–19 ppm) content (typical upper mantle mineral association, Mg# >70, Cr > 400 ppm, and Ni > 1000 ppm (Wilson 1989)), indicating that they were not derived from pure mantle peridotite. However, the low FeOT/MgO and high MgO or Mg# values suggest that their genesis must include a certain portion of mantle peridotite, as partial melting of metabasaltic and eclogite only generates the mafic-felsic melts with high FeOT/MgO and low MgO or Mg# values (Wang et al. 2007). Moreover, the gabbro samples are characterized by LREE enrichment and pronounced negative Nb and Ti anomalies, which are comparable to those of E-MORB, different from OIB and N-MORB (Fig. 5a, b). Considering these features along with the εHf(t) values of -14.3 to -9.3 and old TDM2 model ages (1.77 to 2.08 Ga) of gabbro samples, we suggest that enriched lithospheric mantle is one of the possible sources. The rare element ratios, such as Th/Yb, Ba/La, La/Sm, Ba/Th, Nb/Zr and Th/Zr are widely used to identify metasomatic agents and estimate the flux from the subducted slabs (Kepezhinskas et al. 1997; Tatsumi 2006). The gabbro samples have low La/Sm (3.78–4.59), Th/Zr (3.78–4.59), and Nb/Zr (0.039–0.043) ratios, indicating that sediment-derived melts and aqueous fluids did not play an important role in magma generation (Kepezhinskas et al. 1997; Tatsumi 2006).Furthermore, in the Ba/La vs. Th/Yb diagram(Fig. 9d), they also show normal arc-magma characteristics without significant fluid and melt involvement. In the Sm/Yb vs. Sm diagram, the gabbro and porphyritic like monzogranite plot near the garnet-lherzolite mantle source (Fig. 9b), implying that they have a garnet lherzolite mantle source because garnet has a high partition coefficient for Yb (Dgarnet/melt = 6.6) relative to Sm (Dgarnet/melt = 0.25) (Johnson 1994). In conclusion, we propose a genetic model, which is similar to the model of mafic–ultramafic rocks in the Dabieshan Orogen, the gabbro was generated by partial melting of a source formed by the mixing of ancient enriched lithospheric mantle and eclogitic lower crust mobilized by delamination or foundering (Wang et al. 2007).

The granodiorite

It is difficult to differentiate A-type granites from highly fractionated I-type granites (Chappell and Stephens 1988; Chappell and White 1992; Wu et al. 2003). However, the following points indicate that the granodiorite can be identified as A-type granites. First, according to Watson and Harrison’s (1983) formula, the zircon saturation temperatures (TZr) values of the granodiorite (813–822 °C) are higher than those of other granitoids in Wulong region(742–798 °C) and highly fractionated I-type granites (average of 764 °C), which is in accordance with the range for A-type granites (Table 1, King et al. 1997; Watson and Harrison 1983; Xin et al. 2018). Second, the characteristics of granodiorite, including the enrichment in HFSEs (Zr, Nb, and Y) and HREEs as well as the significant depletions in P, Ti, and Eu, are features of A-type granites (Loiselle and Wones 1979; Collins et al. 1982; Whalen et al. 1987). In addition, the granodiorite samples have high contents of Zr + Nb + Ce + Y (377–393 ppm) and therefore all samples fall into the A-type granite field (Fig. 7a, c).

Recently, several petrogenetic mechanisms have been proposed for the origin of A-type granite, including the crystallization differentiation of a mantle basaltic magma (Eby 1992), partial melting of crust materials (Patiño Douce 1997; Skjerlie and Johnston 1992), and mixing of mantle and crust materials (Yang et al. 2006). However, it is unlikely that the granodiorite in this study was generated by the direct melting of a mantle source, given that their SiO2 contents are as high as 67.8 wt% and their Mg# values are as low as 50.

The high SiO2 and high-K calc-alkaline geochemical characteristics of the granodiorite, together with significant LILE and LREE enrichment and HFSE depletion imply that the primary magma may be derived by the partial melting of crustal material. In addition, the limited variations in the geochemical compositions of εHf(t) (-21.0 to -17.7) and TDM2(2.29 to 2.50 Ga) values also suggest it was mainly derived from a crustal source (Xin et al. 2018). However, mafic microgranular enclaves (MMEs) were widely found within the granodiorite (Fig. S1a in the ESM). Moreover, a large number of igneous rocks, whose origins are crust–mantle interactions, have been found in the NCC during the Early Cretaceous, such as south Taihang magmatic complex (132 Ma) (Chen et al. 2008), Zijinshan alkaline intrusive complex (127 Ma) (Ying et al. 2007), Longbaoshan alkaline complex (129–132 Ma) (Lan et al. 2011) and Laiwu quartz monzonite (130 Ma) (Lan et al. 2013). In particular, the A-type granites in Qianshan, Liaodong Peninsula were generated by the mixing of enriched lithospheric mantle-derived magma with upper crustal-derived magmas (Yang et al. 2006). Furthermore, the granodiorite has almost the same εHf(t) values and crystallization age as the above igneous rocks, which were generated by magma-mixing (Fig. 7).

MMEs are widely used to understand the petrogenesis and source of magmas Barbarin 2005; Didier 1973; Yang et al. 2015; Zhao et al. 2018). In previous research, many different models have been proposed to explain their origin, including: (1) refractory restites, (2) xenoliths (Bonin 2004; Vernon 1984), (3) mineral cumulates from a mafic magma (Didier 1973), and (4) the mixing of mafic with felsic magmas(Chen et al. 2009; Feeley et al. 2008; Holden et al. 1987; Vernon 1984). We found that there were no inherited zircons for granodiorite (Table 1), which is inconsistent with the refractory restites model and xenoliths model. The igneous textures of the MMEs (Fig. S1c, d in the ESM) also indicate that their formation was not xenoliths or refractory restites. In addition, the lack of cumulate textures (Fig. 3c, d) suggests that the origin is not mafic mineral cumulates. Therefore, the magma-mixing model is likely the origin of the MMEs because: (1) They underwent plastic deformation in a partially crystallized convective magma reservoir, with elongated shapes (Fig. S1a) (Zhao et al. 2018). (2) The presence of chilled margins against the granodiorites directly suggests the mixing of mafic melts with felsic crustal melts (Fig. S1a, c). (3) The disequilibrium textures, such as the reverse zoning of plagioclase, indicate that magma mixing occurred in the MMEs (Fig. S1c in the ESM). (4) Acicular apatites in the MMEs indicate that quenching occurred as a result of mixing of hot mafic melts with cooler felsic crustal melts (Fig. S1d in the ESM; Vernon 1984). Therefore, we believe that the origins of the MMEs and the granodiorite are magma-mixing.

In addition, the granodiorite has high CaO, high Sr, low Zn, low total alkali contents, and relatively low 10,000 Ga/Al ratios, which are higher than the porphyritic-like biotite monzogranite, porphyritic-like syenogranite and syenogranite ratios (1.70–2.06) (Table 2; Fig. 8a, b). These unfavorable characteristics of A-granite can be explained by the magma–mixing mechanism. The gabbro in the Wulong region, which may represent the mafic end–member, has low 10,000 Ga/Al ratios (1.60–1.80), Zr concentrations (32–40 ppm), and high Cao and Sr contents (754–784 ppm). Therefore, we propose that the mixing of mafic melts weakens the A-type granite characteristics.

The porphyritic‐like biotite monzogranite, porphyritic‐like syenogranite and syenogranite

The porphyritic-like biotite monzogranite, porphyritic-like syenogranite, and syenogranite in this study are different from the granodiorite; therefore, we can distinguish them from A-type granites. All of the samples fall in the field of I- and S-type granites (Fig. 7a, b). They contain typical minerals of I-type granites, such as amphibole and biotite (Fig. S1g,h,k,i in the ESM). Moreover, as previously discussed, the S-type granite affinity is excluded. Therefore, they can be identified as I-type granites. In the FeOT/MgO vs. (Zr + Nb + Ce + Y) diagram (Fig. 7c), they exist in the highly fractionated I-type granite field, which is further supported by the striking depletions in Nb, P, and Ti (Fig. 5b, d). The negative Nb–Ti anomalies may be related to the fractionation of the Ti-bearing phases (such as ilmenite, titanite) and the negative P anomalies might result from apatite separation (Wu et al. 2003). Based on this examination, the porphyritic-like biotite monzogranite, porphyritic-like syenogranite, and syenogranite located in the NCC can be classified as highly fractionated I-type granites.

Generally, I-type granites are proposed to form by the partial melting of mafic to intermediate igneous rocks (Chappell et al. 2012; Jiang et al. 2018; Li and Li 2007). As can be observed in the Nb/Ta vs. Zr/Sm diagram (Fig. 9a), the porphyritic-like biotite monzogranite, porphyritic-like syenogranite, and syenogranite were originated from the amphibolite melt. Dehydration melting experiments suggest that melts from amphibolitic/basaltic rocks usually have Mg# values lower than 40, regardless of the degree of partial melting (Rapp and Watson 1995; Jiang et al. 2018). The porphyritic-like biotite monzogranite, porphyritic-like syenogranite, and syenogranite in this study have low Mg# values and high K2O and Na2O contents and K2O/Na2O ratios (0.94–0.98), which are similar to that in the experimentally obtained melts for medium- to high-K basaltic rocks (Sisson et al. 2005). Moreover, the porphyritic-like biotite monzogranite, porphyritic-like syenogranite and syenogranite are characterized by low εHf(t) values of -17.4 to -11.2 (Fig. 7) and old Hf model ages (TDM2) of 1.88 Ga to 2.27 Ga (Table 2), indicating that they were derived from Paleoproterozoic medium- to high-K basaltic lower crust.

Tectonic setting

Cretaceous igneous rocks are widespread throughout the NCC. Many scholars believe that Cretaceous igneous rocks were all emplaced in an extensional setting, as indicated by the occurrence of A-type granites, mafic rocks and metamorphic core complexes in NCC (Wu et al. 2005a). However, Sun et al. (2007) proposed that the subduction direction of the Paleo-Pacific Plate in this region changed by approximately 80° in approximately 125 − 120 Myr ago and that the major tectonic mechanism changed from extension to transpression (Wessel and Kroenke 1997; Koppers et al. 2001, 2003). The zircon LA-ICP-MS U-Pb chronology shows that the crystallization ages of Group 1 and Group 2 are 129–126 Myr ago and 124–123 Myr ago, respectively. We conducted a detailed study of the tectonic setting to help resolve this debate.

The gabbro and granitoids samples belong to medium- to high-K calc-alkaline series, and are characterized by their enrichments in LILEs and LREEs, and depletions in HREEs and HFSEs, indicating that they formed in a subduction-related setting. In Fig. S2 in the ESM, all granites fall into an arc-related setting, which is in line with their above geochemical characteristics. Therefore, these igneous rocks are related to the subduction of the Paleo-Pacific Plate. As previously stated, the gabbro, porphyritic-like biotite monzogranite, and porphyritic-like monzogranite belong to the same period of magmatism in the Wulongbei pluton. Moreover, the presence of a ‘Daly gap’ between mafic (the gabbro) and felsic rocks (the porphyritic-like biotite monzogranite and porphyritic-like monzogranite) suggests that they are bimodal magmatic rocks. A-type granites and bimodal magmatic rocks are generally indicative of an extensional tectonic setting (Eby 1992; Whalen et al. 1996; King et al. 1997; Förster et al. 1997; Bonin 2004; Ikeda and Yuasa 1989; Pin and Paquette 1997). Therefore, the Group 1 rocks, inciuding Sanguliu granodiorite (A-type granite,129 Ma) and Wulongbei gabbro, porphyritic-like biotite monzogranite and porphyritic-like monzogranite (bimodal magmatic rocks, 126–127 Myr ago), were emplaced in an extensional setting associated with subduction.

However, the porphyritic-like syenogranite (123 Ma) and syenogranite (124 Ma) are typically I-type granites, which can emplace in extensional, compressional, or transpressional settings. Moreover, I-type granites were widely found in the Liaodong Peninsula during 125–120 Myr ago (Fig. 1c; Table 3). To solve this discordant problem, we summarized the geological ages of the A-type granites, mafic rocks, and extensional structures based on previously published data for the NCC.

Table 3 Rock types and geochronological data of the plutons in the North China Craton

It can be clearly seen in Table 3; Fig. 10 that the A-type granites are widely developed during 140–125 Myr ago and 120–108 Myr ago, while no A-type granites are discovered during 125–120 Myr ago. As mafic rocks are rarely found, the trend is not clear, but no mafic rocks have been found, which emplaced at 125–120 Myr ago (Table 3; Fig. 10). Moreover, a series of extensional structures were discovered, including the Liaonan metamorphic and Wanfu metamorphic core complexes, as well as several detachment faults (e.g. Dayingzi detachment fault system) and half grabens (Tongyuanpu, Dandong, and Benxi half grabens) in a 40,000 km2 area (Fig. 1c). However, the Yuhuangding andesite from the Pulandian-Wafangdian basin, andesite from the Guiyunhua Formation in the Dayingzi basin, rhyolite from the Huanghuadian basin, Guiyunhua Formation andesite in the Dandong basin, granitic dike intruding the andesite in the Tongyuanpu basin, and rhyolite in the Guiyunhua Formation of the Benxi basin formed at 126 ± 6 Ma, 132 ± 5 Ma, 135 ± 1 Ma, 106 ± 3 Ma, 116 ± 2 Ma, and 126 ± 3 Ma, respectively (Liu et al. 2011). Note that the A-type granites, mafic rocks, and extensional structures are almost non-existent during 125 − 120 Myr ago in the NCC (Table 3; Fig. 10). Therefore, according to our results, we propose that the tectonic setting changed from extension to transpression during 125 − 120 Myr ago.

Fig. 10
figure 10

Plot of recent high-precision dating results for the Cretaceous in NCC, data are listed in Table 3 (after Sun et al. 2007)

Mechanism

The destruction of the NCC is a continually debated issue, as its processes remain undetermined. The two main processes proposed in previous studies are delamination (Gao et al. 2004) and thermal/chemical erosion (Menzies and Xu 1998). A key criterion in distinguishing delamination from thermal/chemical erosion is whether continental components are present in the lithospheric mantle (Xu et al. 2013). The characteristics of Mesozoic basalts in the eastern NCC provide indirect evidence for the recycling of continental crust (Gao et al. 2004). Moreover, xenolith studies indicate that olivines are the residue of ancient lithospheric mantle that was modified intensively by a melt derived from recycled continental crust (Xu et al. 2013). These results imply that delamination was the main mechanism for the destruction of the NCC.

Gao et al. (2004, 2009) suggested four possibilities for the tectonic driving force of the delamination of the NCC: (1) collision between North China and the Yangtze Cratons; (2) subduction of the Paleo-Pacific Plate; (3) collision between the North China-Mongolia Plate and Siberian Plate; and (4) global tectonics (mantle plume). The natures of adakite in the Wulong region is similar to that of the early Cretaceous adakitic granites in Dabie Orogen. However, collision between the North China and Yangtze Cratons (Wang et al. 2007) and the North China-Mongolia Plate and Siberian Plate seem unlikely to play an important role in the Liaodong Peninsula because of the subduction-related setting. In addition, due to the lack of an asthenospheric mantle or superplume-derived mafic igneous rocks in the study area, the mantle plume should be considered carefully. Therefore, we propose the tectonic driving force of the delamination of the NCC as a result of thePaleo-Pacific Plate subduction in the NCC (Wu et al. 2005a; Sun et al. 2007).

The thick (approximately 200 km) Archean or Proterozoic lithosphere (Fan and Menzies 1992; Menzies et al. 1993; Menzies and Xu 1998; Griffin et al. 1998) results in a high- to ultrahigh-pressure setting in the crustal root, which can transform the mafic lower crust into eclogite. As eclogite has a density that is higher than that of peridotite by 0.2–0.4. g cm− 3 (Rudnick 1995), the eclogitic thickened lower crust can be recycled into the mantle (Gao et al. 2004).The delamination of the eclogitic crust probably caused it to sink into the underlying hotter and more plastic lithosphere (Xu et al. 2002; Wang et al. 2006) or the asthenospheric mantle (Arndt and Goldstein 1989; Kay and Mahlburg-Kay 1991; Rudnick 1995; Lustrino 2005). This process would give rise to a variety of magmatic rocks, such as the granites, adakites, and gabbro with the subduction of the Paleo-Pacific Plate (Fig. 11a). When the subduction direction of the Paleo-Pacific Plate changed by approximately 80° during 125–120 Myr ago, the regional tectonic stress fields changed from extensional to transpressional, subsequently forming the I-type granites (Fig. 11b).

Fig. 11
figure 11

Synoptic block diagrams. a 140 − 125 Myr ago; b 125 − 120 Myr ago

Conclusions

The results of tis present study have led to four main conclusions. (i) The Wulong intrusive rocks emplaced during 129–123 Myr ago. They comprise A-type granite, K-rich adakite, highly fractionated I-type granites and gabbro. (ii) The granodiorite was formed by the mixing of crust materials and mafic melts. The gabbro was generated by the partial melting of a source formed by the mixing of enriched lithospheric mantle and delaminated eclogitic lower crust. The porphyritic-like monzogranite was produced through partial melting of thickened eclogitic lower crust. The porphyritic-like biotite monzogranite, porphyritic-like syenogranite and syenogranite originated from medium- to high-K basaltic lower crust. (iii) The tectonic setting of NCC changed from extension to transpression during 125 − 120 Myr ago. (iv) The Early Cretaceous delamination resulted from the Paleo-Pacific Plate subduction in NCC.