As one peculiar area and the most stable region of a continent, Archean cratons contain light, cold, and very thick (>200km) lithosphere. While there is no evidence of tectono-magmatic activity, mineralization, or seismicity in most old cratons (Van der Hilst and McDonough, 1999; Sleep, 2003, 2005; Carlson et al., 2005; King, 2005), the North China Craton (NCC), with a thickness of ca. 60~80km (Menzies et al., 1993; Griffin et al., 1998; Rudnick et al., 2004; Xu, 2007) and an age of >3.8Ga (Liu et al., 1992). The most striking features of the NCC are the gravity anomalies within the Taihang-Da Hinggan Mountain gravity lineament and the Tan-Lu Fault.

In the Late Mesozoic, large-scale volcanic rocks in the Taihang-Da Hinggan Mountain belt and Tan-Lu magmatic belts have been reported (Ma, 1997; Su et al., 1999; Xie et al., 2007; Tang et al., 2008; Cao, 2009; Fu, 2013; Wu et al., 2013; Cao et al., 2014). Moreover, the NCC was within an extensional setting, as indicated by presence of large-scale extensional basins (e.g., the Haogou, Guzhen, Pingyi, Mengyin, Jiyang, and Liaoxi basins; Meng, 2003), detachment fault zones (e.g., the Dayingzi Fault zone), and metamorphic core complexes (e.g., Hohhot, Yunmeng mountain, Waziyu in West Liaoning Province, South Liaoning Province, Lesser Qinling, and Song-Liao; Yang and Li, 2008). Mesozoic mafic dykes (e.g., diorite, lamprophyre, and diabase porphyry dykes) were produced in an extensional setting, with more than 300 mafic dykes dispersed throughout the NCC. It is generally accepted that these dykes yield important information on the evolution of the lithosphere including its extensional characteristics, mantle compositions, and temporal and spatial evolution. Despite this, few studies have examined these dykes. Nevertheless, the mafic rocks have only rarely been reported from two regions.

The mafic dykes are dominantly dolerite and lamprophyre, along with minor porphyritic diabase, and they generally occur in NE-SW, E-W, and N-S trending dyke swarms. Here, we report new representative zircon U-Pb ages, and geochemical and Sr-Nd-Pb isotopic data for mafic dykes from the Taihang-Da Hinggan Mountain gravity lineament (Fig. 1a-c) and the Tan-Lu Fault zone (Fig. 1d). These data provide insights into the timing, source, and origin of the dykes, and enable an assessment of the geodynamic processes that led to this magmatism. They also provide evidence of the tectonic setting of the magmatism within the magmatic Taihang-Da Hinggan Mountain belt and Tan-Lu Fault zone.

图 1

Fig. 1

Figure 1 Simplified tectonic map of the area surrounding the Taihang-Da Hinggan Mountain belt, containing anomalous gravity gradients, and the Tan-Lu Fault zone Mafic dykes are concentrated at the anomalous gravity gradients and along the fault

The NCC is located in northern China and covers an area of ~1.7 million km² (Zhai and Santosh, 2011, 2013; Li et al., 2013; Zheng et al., 2013). It is bounded by the Yinshan-Yanshan orogenic belt to the north and the Qingling-Dabie orogenic belt to the south. The most commonly employed model of the NCC comprises uniform Precambrian (Archaean-Sinian) crystalline basement overlain by a variety of Cambrian-Quaternary cover rocks. The NCC can be divided into the Eastern, Western, and Central blocks (Zhao et al., 2001), and the Eastern and Western blocks can be further subdivided into micro-continental blocks and active belts (Zhai et al., 2000). The NCC is gravity anomalies within the Taihang-Da Hinggan Mountain belt are located between the NCC and the Erdos Plateau. They span both the NCC and the Hercynian Xingmeng orogenic belt, and are connected to the Yanshanian belt in the north and the Qinling-Dabie tectonic belt in the south. The Taihang-Da Hinggan Mountain belt is also proximal to the Pacific tectonic belt and its numerous associated major NNE-SSW trending Faults.

Another striking feature of the NCC is the Tan-Lu Fault zone of the Eastern Block. This Fault begins under the western Pacific Ocean and extends into eastern China (Zhao, 2014). It trends NNE-SSW and stretches for more than 2400km, passing through Hubei, Anhui, Jiansu, and Shandong provinces, through the Gulf of Bohai, and into northeastern China. The Tan-Lu Fault zone also cuts the Yangtze Craton and the Xingmeng-Jihei-Dabie-Sulu orogenic belts. Movement on the Fault was mainly strike-slip during the Late Jurassic to Early Cretaceous, but has changed to compression since the Neogene (Wang et al., 2006). The Tan-Lu Fault zone has been the subject of long-term (ca. 40 years) study, reflecting its importance in the regional geological framework (Zhu et al., 2010).

Mafic dykes are widespread throughout the 12 counties that form the focus of this study (i.e., Laiyuan, Luxian, Quyang, Lingshou, Yangquan, Qixian, Changyi, Tancheng, Sijing, Dingyuan, Hefei, Lujiang counties), in Hebei, Shanxi, Shandong, and Anhui provinces of the NCC (Fig. 1a-d). This area contains granite, monzonite, gabbro, and gneissic country rocks, all intruded by mafic dykes. However, the majority of the dykes are hosted in the granite and monzonite units. Individual mafic dykes are vertical and strike NE-SW, E-W, and N-S (Fig. 1a-d). They are 8~15m wide, 3.0~40km long (Fig. 1a-d). The studied dykes can been divided into NE (the Taihang-Da Hinggan Mountain gravity lineament)-and NW (the Tan-Lu Fault)-trending two categories.

The NE-trending mafic dykes (Laiyuan, Luxian, Quyang, Lingshou, Yangquan, Qixian) (Fig. 2a, c, d) from the Taihang-Da Hinggan Mountain gravity lineament are hosted in monzonites and granites, are vertical, and strike from E-W to NW-SE. They are 10~130m wide and 5~60km long. They contain 30%~36% micro-phenocrysts (0.5~1.3mm across) of clinopyroxene and plagioclase, alkali feldspar, and plagioclase, along with minor biotite, within a groundmass of pyroxene, plagioclase, magnetite, and chlorite.

图 2

Fig. 2

Figure 2 The mafic dyke swarm from Laiyuan and Sijing (a, b), and the monoclinic pyroxene plots from Laiyuan, Lingshou, Changyi, and Tancheng (c-f)

In contrast, the NW-trending mafic dykes (Changyi, Tancheng, Sijing, Dingyuan, Hefei) (Fig. 2b, e, f) from the Tan-Lu Fault are hosted in Proterozoic granites, sedimentary rocks and granitic complexes, and Mesozoic volcanic and sedimentary rocks, are vertical, and strike NE-SW. They are 10~130m wide and 5~60km long. They contain 30%~35% micro-phenocrysts (0.5~1.3mm across) of clinopyroxene and plagioclase, along with minor biotite, within a groundmass of pyroxene, plagioclase, magnetite, and chlorite.

Zircons from 12 mafic dyke samples from the Taihang-Da Hinggan Mountain gravity gradient belt and the Tan-Lu Fault zone of Hebei, Shanxi, Shandong, and Anhui provinces were separated using conventional heavy liquid and magnetic techniques. Representative zircons were then handpicked under a binocular microscope before being mounted in an epoxy resin disc, polished, and coated with gold prior to analysis. These zircons were imaged using transmitted and reflected light microscopy, and cathodoluminescence (CL) to highlight external and internal structures. The CL imaging and U-Pb analyses were undertaken at the State Key Laboratory of Continental Dynamics, Northwest University, Xi'an, China. The analytical procedures used are described in detail in Harris et al. (2004) and Campbell et al. (2006). U-Th-Pb ratios and absolute abundances were determined by reference to multiple measurements of a standard TEMORA zircon and a NIST 610 glass standard.

The whole-rock and Sr-Nb-Pb isotopic geochemistry of 26 mafic dyke samples was determined during this study. Prior to analysis, these samples were trimmed to remove altered surface material before being cleaned with deionized water, crushed, and powdered in an agate mill.

Major element compositions were determined using a PANalytical Axios-Advanced X-ray fluorescence (XRF) spectrometer at the State Key Laboratory of Ore Deposit Geochemistry (SKLODG), Institute of Geochemistry, Chinese Academy of Sciences, Guiyang, China, using fused lithium tetraborate glass pellets. These analyses have an analytical precision of better than 5%. Trace element compositions were determined by ICP-MS using a Perkin-Elmer ELAN DRC-e instrument at SKLODG. Prior to analysis, powdered samples (50mg) were dissolved in high-pressure Teflon bombs using a HF+HNO₃ solution for 48 hours at ca. 190℃ (Qi et al., 2000). Signal drift during analysis was monitored using Rh as an internal standard, and the GBPG-1, OU-6, GSR-1, and GSR-3 standards were used for analytical quality control, indicating an analytical precision generally better than 5% for trace elements.

The Rb-Sr and Sm-Nd isotopic analyses began by spiking the sample powders with mixed isotope tracers before dissolution in Teflon capsules using a HF+HNO₃ acid solution and separation using conventional cation-exchange techniques. Isotopic measurements were undertaken using thermal ionization mass spectrometry (TIMS) at the Isotopic Geochemistry Laboratory of the Yichang Institute of Geology and Minerals Resources, Yichang, China. This analysis yielded procedural blanks of < 200pg for Sm and Nd, and < 500pg for Rb and Sr. The isotopic ratios of Sr and Nd were corrected for mass fractionation by normalizing to ⁸⁶Sr/⁸⁸Sr=0.1194 and ¹⁴⁶Nd/¹⁴⁴Nd=0.7219, respectively. Analysis of the NBS987 standard yielded a mean ⁸⁷Sr/⁸⁶Sr value of 0.710246±16 (2σ), and analysis the La Jolla standard yielded a mean ¹⁴³Nd/¹⁴⁴Nd value of 0.511863±8 (2σ). Prior to Pb isotopic analysis, Pb was separated and purified by conventional cation-exchange techniques (i.e., micro-columns filled with AG1.8, 200~400 mesh resin) using diluted HBr as an eluent. Analysis of the NBS981 standard yielded a mean ²⁰⁴Pb/²⁰⁶Pb value of 0.0896±15, a mean ²⁰⁷Pb/²⁰⁶Pb value of 0.9145±8, and a mean ²⁰⁸Pb/²⁰⁶Pb value of 2.162±2.

Zircon is relatively abundant in the mafic dykes (Table 1), and all zircons show oscillatory or planar zoning in CL images (Fig. 3a-l), indicating a magmatic origin. None of the zircons show evidence of inherited cores and all have relatively high Th/U ratios (0.78~2.49), which also indicate a magmatic origin.

表 1

Table 1

表 1(Table 1)

Table 1 Zircon LA-ICP-MS U-Pb isotopic data for samples from dolerites in the study area of the Taihang-Da Hinggan Mountain belt and Tan-Lu Fault zone

Table 1 Zircon LA-ICP-MS U-Pb isotopic data for samples from dolerites in the study area of the Taihang-Da Hinggan Mountain belt and Tan-Lu Fault zone

图 3

Fig. 3

Figure 3 Zircon LA-ICP-MS U-Pb concordia diagrams for mafic dykes in the study area of the Taihang-Da Hinggan Mountain belt and the Tan-Lu Fault zone

Results from ²⁰⁶Pb/²³⁸U age analyses of samples from dykes in the 12 counties considered in this study are presented in Table 1 and summarized here. Zircons from NE-trending mafic dykes yielded the weighted mean age between 128.1±1.2Ma and 121.4±1.1Ma (Fig. 3a-f). And the zircons from NW-trending mafic dykes yielded the weighted mean age between 125.2±1.0Ma and 115.0±0.8Ma (Fig. 3g-l).

The whole-rock compositions of the mafic dykes are presented in Tables 2 and Table 3. These dykes are characterized by only slight variations in SiO₂ (51.28%~51.86%), Na₂O (2.45%~3.43%), and MnO (0.13%~0.17%), but have more variable concentrations of TiO₂ (0.76%~1.33%), Al₂O₃ (14.53%~16.36%), K₂O (2.62%~3.63%), Fe₂O₃ (7.03%~9.63%), MgO (5.06%~7.36%), CaO (6.85%~9.12%), P₂O₅ (0.43%~1.24%), and Mg^# values (62~70) (Table 2). All dykes plot within the alkaline and shoshonitic fields on a TAS diagram (Fig. 4).

表 2

Table 2

表 2(Table 2)

Table 2 Major element compositions (wt%) of dolerites within the study area of the Taihang-Da Hinggan Mountain belt and Tan-Lu Fault zone

Table 2 Major element compositions (wt%) of dolerites within the study area of the Taihang-Da Hinggan Mountain belt and Tan-Lu Fault zone

表 3

Table 3

表 3(Table 3)

Table 3 Trace element compositions (×10-6) of dolerites within the study area of the Taihang-Da Hinggan Mountain belt and Tan-Lu Fault zone

Table 3 Trace element compositions (×10-6) of dolerites within the study area of the Taihang-Da Hinggan Mountain belt and Tan-Lu Fault zone

图 4

Fig. 4

Figure 4 Total alkali vs. silica (TAS) plot for mafic dykes within the study area of the Taihang-Da Hinggan Mountain belt and Tan-Lu Fault zone (after Middlemost, 1994; Le Maitre, 2002) All major element data have been recalculated to 100% on an anhydrous basis

The mafic dykes have near-identical chondrite-and primitive-mantle-normalized compositions (Table 3; Fig. 5) that are enriched in light rare earth elements (LREEs), some large ion lithophile elements (LILEs; e.g., Rb, Ba, and Sr), Pb, and Th. They also have negative Nb, Ta, and Ti anomalies, and contain generally negative Eu anomalies (Eu/Eu^*=0.64~1.02). The samples have Nb/Ta ratios of 12.4~23.5 and Zr/Hf ratios of 30.3~95.3.

图 5

Fig. 5

Figure 5 Chondrite-normalized REE pattern (a) and primitive-mantle-normalized multi-element variation diagram (b) of the mafic dykes in the study area of the Taihang-Da Hinggan Mountain belt and Tan-Lu Fault zone (normalization values after Sun and McDonough, 1989)

The Sr, Nd, and Pb isotopic compositions of representative dolerite samples are presented in Tables 4 and Table 5. These dykes have a wide range of initial ⁸⁷Sr/⁸⁶Sr ratios (0.7056~0.7057, the Taihang-Da Hinggan Mountain gravity lineament; 0.7102~0.7105, the Tan-Lu Fault) and negative ε_Nd(t) values (-15.5 to -12.4) (Table 4) that are indicative of a common source region. In addition, analysis of these dykes yields Nd model ages of 1.82~2.69Ga (Table 4), suggesting that these samples have EM1-like Sr-Nd isotopic ratios (Hart, 1984; Zindler and Hart, 1986), similar to other Mesozoic mafic dykes from the NCC (Liu et al., 2008a, b, 2009a, b). This is also reflected in the position of these dykes within the mantle array and shoshonitic field of an ε_Nd(t) vs. (⁸⁷Sr/⁸⁶Sr)_i diagram (Fig. 6).

表 4

Table 4

表 4(Table 4)

Table 4 Sr-Nd isotopic compositions of representative dolerites from the study area of the Taihang-Da Hinggan Mountain belt and Tan-Lu Fault zone

Table 4 Sr-Nd isotopic compositions of representative dolerites from the study area of the Taihang-Da Hinggan Mountain belt and Tan-Lu Fault zone

表 5

Table 5

表 5(Table 5)

Table 5 Pb isotopic compositions of representative dolerites from the study area of the Taihang-Da Hinggan Mountain belt and Tan-Lu Fault zone

Table 5 Pb isotopic compositions of representative dolerites from the study area of the Taihang-Da Hinggan Mountain belt and Tan-Lu Fault zone

图 6

Fig. 6

Figure 6 Initial 87Sr/86Sr vs. εNd(t) diagram for mafic dykes in the study area of the Taihang-Da Hinggan Mountain belt and Tan-Lu Fault zone

Also shown are the compositions of Mesozoic mafic rocks from the NCC (Liu et al., 2008a, b, 2009a) and mafic rocks from the Yangtze Craton (Chen et al., 2001; Li et al., 2004)

The dolerites have relatively constant Pb isotopic ratios similar to other of mafic dykes of the NCC: (²⁰⁶Pb/²⁰⁴Pb)_i=16.45~16.49, (²⁰⁷Pb/²⁰⁴Pb)_i=15.44~15.51, and (²⁰⁸Pb/²⁰⁴Pb)_i=36.49~36.53 (Zhang et al., 2004; Xie et al., 2006; Liu et al., 2008a, b, 2009a, b) (Table 5). They also show an EM1-like isotopic signature (Fig. 7b), but differ from the composition of mafic dykes from the Yangtze Craton.

图 7

Fig. 7

Figure 7 208Pb/204Pb and 207Pb/204Pb vs. 206Pb/204Pb diagrams for mafic dykes from the study area of the Taihang-Da Hinggan Mountain belt and Tan-Lu Fault zone

Also shown are I-MORB (Indian MORB) and P&N-MORB (Pacific and North Atlantic MORB), OIB, and NHRL fields after Zou et al. (2000), and a 4.55Ga geochron from Hart (1984). The NCC data are from Zhang et al. (2004) and Xie et al. (2006), and data of the Yangtze Craton mafic rocks are from Yan et al. (2003)

Mesozoic intrusions are widespread in the Taihang-Da Hinggan Mountain gravity lineament, and include 140~125Ma diorite-monzonite complexes (Dong et al., 2003), Late Mesozoic gabbros and intermediate-felsic rocks (Cai et al., 2003, 2004; Yang et al., 2004; Qin, 2005), magmatic rocks in the Taihang mountains (Chen et al., 2005, 2007; Ying et al., 2010), alkaline rocks, granitoid complexes, and granites (Cai et al., 2006), 125~127Ma monzonites and diorites (Wang et al., 2006), 115~135Ma diorites, 134~149Ma volcanic rocks, 135~145Ma complexes (Cai et al., 2003; Chen et al., 2003; Chen and Zhai, 2003), and ca.120Ma calc-alkaline lamprophyres (Chen et al., 2003). Mesozoic magmatism is also widespread in the area of the Tan-Lu Fault zone, including high-K calc-alkaline and volcanic rocks (Zhu et al., 2010; Li et al., 2012), monzonites and granites (Cao et al., 2010), and alkali-rich intrusive rocks. Both the magmatism in the Taihang-Da Hinggan Mountain gravity lineament and in Tan-Lu Fault zone is the result of lithospheric extension. In addition, Mesozoic mafic dykes (e.g., lamprophyre and dolerite dykes) are widespread in the Tan-Lu Fault zone, and these formed in an extensional setting. However, until now, few studies have examined the Mesozoic mafic dykes in the Taihang-Da Hinggan Mountain gravity gradient belt and the Tan-Lu Fault zone (Guo et al., 2001; Huang et al., 2012).

The dykes considered in this study are characterized by low SiO₂ contents (50.3%~51.8%; Table 2), suggesting they were derived from an ultramafic source (Liu et al., 2008a, b, 2009a, 2013a, b, c, d). Crustal rocks can therefore be ruled out as possible sources, as partial melting of crustal rocks (Hirajima et al., 1990) and lower-crustal intermediate granulites (Gao et al., 1998) in the deep crust would produce liquids with high Si and low Mg contents (i.e., granitoid liquids; Rapp et al., 2003). In addition, the high initial ⁸⁷Sr/⁸⁶Sr ratios (0.7056~0.7105) and negative ε_Nd(t) values (-15.4 to -12.4) of the mafic dykes are consistent with derivation from an enriched lithospheric mantle source, rather than an asthenospheric mantle source which would contain a depleted Sr-Nd isotopic composition, such as MORB.

Crustal contamination may cause a significant depletion in Nb-Ta and enrichment in Sr-Nd isotopic signatures in basaltic rocks (Guo et al., 2004). The mafic dykes studied here are characterized by negative Nb-Ta anomalies (Table 3; Fig. 4b), which implies a crustal component in their origin. In addition, plots of SiO₂, CaO, TiO₂, and P₂O₅ vs. MgO (not shown) show only a weak linear correlation, suggesting magma mixing or contamination played an important role during magma ascent. This inference is also supported by relatively low Ni (6.65×10^-6~225×10^-6), low Ta/La ratios (0.01~0.02 cf. Ta/La=0.06 for primitive mantle; Wood et al., 1979), lack of correlations between Mg^#, Ni, and the initial Sr ratio (not shown), depletion in high field strength elements (Nb, Ta, and Ti), positive Pb anomalies (Fig. 5b; Zhang et al., 2005), and high Ba/Nb ratios (140~438; Table 3; Jahn et al., 1999). However, these dykes lack classical signatures of crustal assimilation including variations in Sr-Nd isotopes, a positive correlation between MgO and ε_Nd(t) values, and a negative correlation between MgO and (⁸⁷Sr/⁸⁶Sr)_i ratios. These observations suggest that the mafic dykes have not, in fact, been significantly affected by crustal contamination.

All the dykes examined in this study are distributed along a partial melting trend line on a plot of La vs. La/Sm (not shown). This result, combined with their Sr-Nd-Pb isotopic compositions (Tables 4, Table 5), indicates the dykes were derived from partial melting of an EM1-like mantle source (Fig. 6, Fig. 7). This view is also supported by their relatively high Ti/Y ratios (213~477; Johnson, 1998). Plots of La/Sm vs. La and Sm/Yb vs. Sm (Fig. 8a, b) indicate that the dykes were derived from 1.0%~5.0% partial melting of a garnet-lherzolite mantle source. In addition, the distinctive negative Nb, Ta, and Ti anomalies on a primitive-mantle-normalized trace element diagram (Fig. 4b) indicate the involvement of components from the Proto-Tethys oceanic or ancient continental crust (Zhang et al., 2005). There are positive correlations between MgO and Fe₂O₃, and CaO, CaO/Al₂O₃, and Ni, and negative correlations in plots of MgO vs. Al₂O₃ and Sr (not shown), suggesting the fractionation of olivine, clinopyroxene, hornblende, and plagioclase. This inference is also supported by the correlations between Sr and each of Ba and Rb. In addition, the high La/Nb ratios (3.6~9.7) in these rocks (Table 3) differ from those of most intra-plate volcanic rocks, including OIB, alkali basalt, and kimberlite (typically 0.5~2.5; Jahn et al., 1999). These data suggest that continental materials (granitoids, granulites, sediments, etc.) were involved in the origin of the mantle-derived magma's, which is supported by the low ε_Nd(t) values (-15.4 to -12.4) and high initial ⁸⁷Sr/⁸⁶Sr values (0.7056~0.7105). Therefore, we propose the involvement of crustal components already incorporated into the mantle source. However, it is important to identify the mechanism by which these crustal materials were incorporated.

图 8

Fig. 8

Figure 8 Variations in La/Sm vs. La (a) and Sm/Yb vs. Sm (b) compositions of samples analyzed during this study, and melt curves and lines derived from the non-modal batch melting equations of Shaw (1970) Melt curves are shown for spinel lherzolite (with mode and melt mode compositions of ol0.600+opx0.200+cpx0.100+gt0.100 and ol0.030+opx0.160+cpx0.880+sp0.110, respectively; Walter, 1998). Mineral/matrix partition coefficients and the composition of the DMM are from McKenzie and O'Nions(1991, 1995). Primitive mantle, and N-MORB and E-MORB compositions are from Sun and McDonough (1989). Tick marks on each curve or line correspond to changing degrees of partial melting of a given mantle source

Previous research has suggested the destruction of the NCC lithosphere was controlled by a number of factors, such as collision between the Yangtze Craton and the NCC, subduction of the Paleo-Pacific Plate and closure of the overlying ocean, India-Eurasia collision, and a mantle plume (Gao et al., 2004). However, as described above, these factors are still debated, meaning that the origin of the Mesozoic igneous rocks remains a topic of controversy. Here we provide a discussion on key aspects of the various models that describe the origin of the Mesozoic igneous rocks.

It is generally believed that the final collision between the NCC and the Yangtze Block occurred during the Triassic (Zhang et al., 2005). Additionally, the studied mafic dykes have Pb isotopic characteristics that are distinct from those of the Yangtze Craton lithosphere mantle (Fig. 7a, b), which rules out the involvement of this mantle in their origin (Xie et al., 2006; Liu et al., 2009a). In contrast, the distinctive Pb isotopic data suggest that the dykes were derived from the overlying NCC during the Late Mesozoic (Xie et al., 2006).

During subduction, the Paleo-Pacific Plate most likely released melt and/or fluids during its descent into the mantle, and the melt/fluids from the subduction of the Pacific Plate (i.e., the Izanagi Plate) metasomatized and modified the lithospheric mantle beneath the NCC. However, during the Late Mesozoic, the Izanagi Plate primarily moved towards the N or the NNE (Maruyama and Send, 1986; Kimura et al., 1990), and there was no westward subduction of an ancient Pacific Plate below the NCC prior to the Early Cretaceous, meaning that this plate had little influence on the origin of the dykes considered in this study. Moreover, a recent U-Th disequilibrium study of Cenozoic potassium basalt from northeastern China also argue against contributions from the Paleo-Pacific Plate (Zou et al., 2003). Research into early Cretaceous mantle-derived rocks from the western NCC also indicates that the origin of the enriched lithospheric mantle sources for the Late Mesozoic rocks was unrelated to subduction of the Paleo-Pacific Plate (Wang et al., 2006; Ying et al., 2007). Additionally, subduction of the Paleo-Pacific Plate cannot explain the ubiquitous compositional grading of the Mesozoic igneous rocks in eastern China. If the adjacent ocean (e.g., the Asia or Tethys Ocean) was subducted, the destruction of the NCC should have been complete; however, this did not occur.

The presence of a mantle plume is one of the least discussed mechanisms that may explain the destruction of the NCC. There are three similar models: 1) upwelling of asthenosphere and destruction of the NCC occurred as a result of a mantle plume (Xu, 2007); 2) giant mantle plumes (e.g., Ontong Java) ascended in the SW Pacific Ocean, which finally resulted in destruction of the NCC (Zhao et al., 2004); and 3) fracturing of the Gondwana mainland and destruction of the NCC occurred due to the activity of mantle plumes in the Late Mesozoic (Wilde et al., 2003). However, based on the petrologic, geochemical, and geophysical evidence, there have been no mantle plumes in the NCC since the Paleozoic. Therefore, none of the above mantle plume models can explain the origin of the mafic dykes.

In general, it is accepted that lithospheric mantle beneath NCC was progressively enriched due to successive hybridization of foundered lower crust (Liu et al., 2008a, b, 2009a). Eclogite can be recycled into the mantle (Arndt and Goldstein, 1989; Kay and Mahlburg-Kay, 1991; Jull and Kelemen, 2001; Gao et al., 2004), primarily as eclogite has a higher density (0.2~0.4g·cm^-3) than lithospheric mantle peridotite (Rudnick and Fountain, 1995; Jull and Kelemen, 2001; Anderson, 2006; Levander et al., 2006). Eclogites also have lower melting temperatures than mantle peridotites (Yaxley, 2000; Kogiso et al., 2003; Sobolev et al., 2005), indicating that foundered silica-saturated eclogites can melt to produce silicic tonalite to trondhjemite melts that may variably hybridize with overlying mantle peridotite material. These reactions can produce an olivine-free pyroxenite that, if subsequently melted, generates basaltic melts (Kogiso et al., 2003; Sobolev et al., 2005). The foundering model is supported by the voluminous coeval magmatism that is well documented in the Taihang-Da Hinggan and Tan-Lu tectonic belts (Deng et al., 1996, 2000; Zhu et al., 2010; Cai et al., 2003, 2004, 2006; Chen et al., 2003, 2005, 2006, 2007; Liu et al., 2004, 2005, 2006, 2008a, b, c, 2009a, b), as well as the large-scale mineralization and adakitic lavas (Xiong et al., 2011; Gu et al., 2013) in this region.

We therefore propose a model in which lower crustal delamination coincided with mafic magmatism. The Triassic collision (240~185Ma; Zhang et al., 2005; Liu et al., 2008a, b, 2009a) between the Yangtze Craton and the NCC generated thickened crust, causing the eclogitization of its lower parts (Liu et al., 2008a, b, 2009a). Foundering of this eclogite at ca. 185~165Ma beneath the Taihang-Da Hinggan orogenic belt and the eastern NCC (Liu et al., 2008a, b, 2009a), combined with ongoing subduction of the Paleo-Pacific Plate (Chen et al., 2005), triggered asthenospheric upwelling, orogenic collapse, and lithospheric extension and thinning in the study area. The silicic melts generated by the melting of the foundered eclogites reacted extensively with the overlying mantle peridotite, with the subsequent (128~115Ma) decompression melting of this hybridized lithospheric mantle producing primary basaltic melts that fractionated to produce the mafic dyke swarms of the Taihang-Da Hinggan Mountain belt and Tan-Lu Fault zone.

New geochronological, geochemical, and Sr-Nd-Pb isotopic data for the mafic dykes of the northern Taihang-Da Hinggan Mountain belt and Tan-Lu Fault zone allow us to reach the following conclusions.

(1) The mafic dykes were intruded during the Early Cretaceous, as indicated by 12 new zircon U-Pb ages of that range between 128.1±1.2Ma and 115.0±0.8Ma.

(2) All of the dykes are dolerites and have doleritic textures. They are all alkaline and shoshonitic, are enriched in the LREE, some LILE (e.g., Rb, Ba, and Sr), Pb, and Th, and are depleted in Nb, Ta, and Ti. They have high initial ⁸⁷Sr/⁸⁶Sr ratios (0.7056~0.7105), negative ε_Nd(t) values (-15.4 to -12.4), relatively constant initial Pb isotopic ratios (16.45~16.49, 15.44~15.51, and 36.49~36.53 for ²⁰⁶Pb/²⁰⁴Pb, ²⁰⁷Pb/²⁰⁴Pb, and ²⁰⁸Pb/²⁰⁴Pb, respectively), and relatively old Nd model ages (1.82~2.69Ga). These data suggest that the magma that formed these dykes were generated by a certain extent partial melting (1.0%~5.0%) of EM1-like garnet-lherzolite mantle material. The magmas fractionated olivine, clinopyroxene, and hornblende during ascent while undergoing negligible crustal contamination.

(3) The 128~115Ma mafic dyke swarms of the Taihang-Da Hinggan Mountain belt and Tan-Lu Fault zone formed as a result of asthenospheric upwelling associated with foundering of eclogite from lower thickened crust in a tectonic setting influenced by ongoing collision of the Paleo-Pacific Plate.