Introduction

The eastern block in China contains the eastern part of the Central Asian orogenic belt (CAOB), North China craton (NCC) (Fig. 1a), Sulu–Dabie orogenic belt, Yangtze craton, and southeastern coastal orogenic belt of China (Fig. 1b) (Sun et al. 2022); this area is characterized by widespread Mesozoic–Cenozoic volcanic rocks, subvolcanic rocks, and intrusive rocks. The late Cenozoic volcanic rocks are mainly composed of alkaline basalts that erupted in an intraplate environment (Chen et al. 2017). They are mainly distributed in the eastern margin of the CAOB and Northeast China, with the Changbaishan volcanic field (CVF) rocks being the most widely distributed. The CVF consists of basalt shields, monogenic vents, and composite central volcanoes (Li et al. 2021a, b). In the last ten years, seismic and three–dimensional magnetotelluric imaging has revealed that the CTV is located above the 500 km deep stagnant plate of the western part of the subducting Pacific plate (Yang et al. 2021; Song and Lei 2023). Three geodynamic mechanisms have been proposed to explain large-scale basaltic magmatic eruptions in the CVF: upwelling mantle plumes (Chen et al. 2007), intraplate continental rifting (Gilder et al. 1991; Dong et al. 2023) and Pacific plate subduction (Zhang et al. 2018). More studies on volcanism have focused on the Millennium eruption (ME) (Zou et al. 2014; Pan et al. 2020), and minimal attention has been given to the chronology, single mineral geochemistry and related magmatic stages or genesis of basalts in the CVF. There are different views on the nature or origin of shield-formed and postshield basaltic magmas in the CVF: 1) partial melting of asthenospheric mantle, with contamination by crustal material (Li et al. 2021a, b); 2) partial melting of the subducting Pacific plate (Zou et al. 2008; Ma et al. 2019); 3) involvement of depleted mantle (DM) and enriched mantle I-type (EM I) (Basu et al. 1991) or involvement of oceanic island basalt (OIB) and EMI materials (Choi et al. 2014; Chen et al. 2021); 4) an asthenospheric mantle source that is involved in subducted sediments or originated from subcontinental lithospheric mantle or has the characteristics of mid-ocean ridge basalt (MORB) (Choi et al. 2020); and 5) similar to the EMI source, derived from the contamination of OIB-type mantle by Paleoproterozoic lithospheric crust (Lee et al. 2021).

Fig. 1
figure 1

a Map showing the tectonic framework of the main subdivisions of central and eastern Asia, as well as eastern China (modified after Zhou et al. 2018); b simplified tectonic units of eastern China and the location of the CVF (modified according to Xu et al. 2012); c regional geological map of the CVF (Li et al. 2021a, b)

Full size image

Apatite (Ca5[PO4]3[F, Cl, OH]) is a low-temperature accessory mineral that is widely developed in the volcanic rocks of the CTV, and its U–Pb sealing temperature (450 °C–550 °C) is lower than that of zircon (700 °C–900 °C). The mineral lattice of apatite is often rich in Sr, Mn, Mg, rare earth elements (REEs), Hf, Pb, U, Th and other elements. This information, combined with isotope geochemical data, can also provide specific rock ages and genetic information (Cherniak 2005). Palma et al. (2019) and Broom et al. (2017) suggested that apatite could be used as a powerful tool for resolving complex magmatic activity and that apatite that crystallizes from magmas with different components does not undergo isotopic changes in the host rock. The geochemistry of associated apatite effectively records the magma source and possible magmatic activity (Chu et al. 2009; Laurent et al. 2017). The CVF is composed of massive late Cenozoic basaltic, trachytic and alkaline rhyolitic volcanic rocks. The following are unresolved issues regarding the preshield and postshield basaltic magmatic eruption of the CTV: 1) whether they were products of continuous or multiple intermittent eruptions from the same magma chamber; 2) whether the eruption of magma was caused by the evolution of the magma itself (such as the fractional crystallization of magma, volatiles and water content in magma, and magma mixing) or by other factors (such as regional deformation, the presence of faults, and the uneven thickness of the overlying crust); and 3) whether there were single or multiple magma sources. Furthermore, volcanic gases (such as CO2, H2S, and SO2) are emitted from the CVF, all the rocks have undergone varying degrees of gas‒liquid alteration (Aleinikoff et al. 2015; Mathieu et al. 2016), and the genesis/magmatic processes of these rocks cannot be clearly elucidated by only whole–rock geochemistry (Warren et al. 2007). Therefore, it is necessary to use new methods to evaluate the role of these basaltic magmatic evolutionary processes in the formation of basalts in the CVF.

The 22–18 Ma Naitoushan basalt and 0.23–0.22 Ma Heishigou dikes are located in the northeastern CTV, were emplaced together and are easily discoverable, providing us with an opportunity to study the above issues. In this study, we present new Rb–Sr dating results for alkaline feldspar and plagioclase, zircon and apatite U–Pb geochronological data and comprehensive in situ geochemical (main oxides, volatiles, trace elements and Sr–Nd isotopic compositions) analyses of apatite in combination with whole–rock geochemical data from preshield and postshield basalts in the CTV in Jilin Province, NE China. These results improve our understanding of basaltic magma sources and the role of magmatic processes in the petrogenesis of basalts.

Geological background and samples

The CTV is an active volcano located on the boundary between China and North Korea. Tectonically, it is the eastern margin of the CAOB, the northern margin of the NCC, and near the NE-striking regional Dunhua–Mishan fault (Fig. 1b). Several monogenetic scoria cones and dikes surround the CTV, and they align approximately in a NW–SE or NE–SW direction. This direction is the same as that of the main faults (Baishan–Jince or Tianchi–Zengfengshan fault) affecting the Tianchi caldera (Fig. 1c). The basement of the CTV consists of Archean and Proterozoic metamorphic rocks (Wang et al. 2003; Li et al. 2023).

In the preshield stage of the CTV, scattered monogenetic basaltic volcanoes widely developed across the Gaima Plateau (22–13 Ma), with limited parasitic basaltic monogenic volcanism occurring in the Tianchi cone, such as occurred with the early Miocene Naitoushan basalt (Figs. 1c and 2). The shield-forming eruptive materials of the CTV mainly comprise a series of basaltic volcanic rocks, from lower to upper: Quanyang basalt (5.02–4.26 Ma), Toudao (Baishan) basalt (2.77–1.48 Ma) and Laofangzixiaoshan basalt (1.17–0.75 Ma) (Wei et al. 2021). The construction of the CTV cone is composed mainly of trachytic and comenditic volcanic rocks, and their eruptions lasted throughout the Pleistocene (0.61–0.019 Ma), with small numbers of basaltic volcanic rocks, such as Heishigou basaltic dikes, forming in the postshield stage (Fig. 2). Before the formation of the Tianchi volcanic cone, small-scale intermediate–felsic magma erupted between 1.49 and 1.00 Ma. Postcaldera-forming and postcaldera eruptions mostly occurred between 900 and 1100 BP. The ME had several potential eruption ages, including 928, 939 and 946 AD. All the ME ages confirmed a likely tenth century eruption age, and the eruption resulted in mainly pumiceous ignimbrite deposits. The records of the 1903 eruption were limited to phreatomagmatic eruptions (Yatsuzuka et al. 2010).

Fig. 2
figure 2

Simplified geological map of the study area (after Yan et al. 2023)

Full size image

Due to the certainties in the petrological relationship between the Naitoushan basalt and Heishigou dikes, we aim to better understand their magma sources and magmatic petrogenetic processes. Our samples were collected from those rocks (Fig. 2), and their key petrographic features and rock-forming mineral assemblages are summarized as follows:

  1. 1.

    The Naitoushan basalt constitutes the main body in the study area (Fig. 2) and is located in the northeastern part of the Tianchi volcanic area (Fig. 1c). The basalt has massive structures (Fig. 3a and c) and intergranular textures, with few pores. The major mineral assemblage includes plagioclase (60–65%), clinopyroxene (10–15%), olivine (5–10%), and alkali feldspar (5%), with accessory minerals such as apatite, Fe–Ti oxides, zircon, and monazite (5%) (Fig. 3e).

    Fig. 3
    figure 3

    a and b Representative field photographs showing outcrops of the Naitoushan basalt (CBS–12a) and Heishigou dike (CBS–8a). c and d Hand specimen photographs of representative samples. e, f and f’ Photomicrographs of the Naitoushan basalt and Heishigou dike. g and h SEM–CL images of representative apatite grains from e and f. The circles indicate the locations of the EPMA and LA–ICP–MS analysis. Ap: apatite, Cpx: clinopyroxene, Ol: olivine, Phl: phlogopite, Pl: plagioclase

    Full size image
  2. 2.

    The Heishigou dikes are distributed in the northeastern part of the Tianchi volcanic cone (Figs. 1c and 2) and occur as dikes or as intrusions within the ignimbrite deposits and the Naitoushan basalt (Figs. 2 and 3b). The strike is NE, the dip angles are 70–85°, and the widths vary from 2.2–6.7 m. The pore content is approximately 15 vol%, and the sizes are 0.05–0.5 mm, indicating the degassing of magma near the surface (Fig. 3d). They are composed mainly of plagioclase (70–75%), alkali feldspar (5–10%), clinopyroxene (5–10%), olivine (5%) and accessory apatite, magnetite, titanite, phlogopite, zircon and monazite (5%) (Fig. 3f, f’).

Analytical methods

Sampling and representativity

Apatite laser ablation–inductively coupled plasma–mass spectrometry (LA–ICP–MS) U–Pb dating samples were obtained from the Naitoushan basalt. We used an in situ sampling method and collected samples from homogeneous and weakly weathered parts of the rock (Fig. 3a, c). To ensure sufficient samples for testing, we collected 10.5 kg of fresh rock samples and cleaned them to prevent contamination. The zircon laser ablation–multicollector–inductively coupled plasma–mass spectrometry (LA–MC–ICP–MS) U–Pb and Rb–Sr dating samples were obtained from the Heishigou basaltic dike. We also used an in situ sampling method, and the samples were all from homogeneous and weakly weathered rocks (Fig. 3b, d). We collected a total of 15 kg of fresh rock samples; cleaned them to prevent contamination; and then cleaned, cut, broke, roughened, overmagnetized, electromagnetically treated and fine-panned the samples for in situ geochemical analysis of apatite and zircon.

Thin section preparation and optical petrography

The Naitoushan basalt and Heishigou basaltic dike samples used for probe testing were cut and numbered, heated in an oven with volatile water (85 ℃), mixed with resin adhesive and triethanolamine (5:1), and this mixture was applied evenly to the heated samples; the samples were subsequently heated for ten hours (85 ℃), followed by rough grinding and fine grinding and polishing to a thicknesses of 0.04–0.05 mm. The zircon and apatite were picked out under a microscope, the samples with good self-shapes and no fragmentation were attached, and the resin was fixed and polished until the apatite and zircon were exposed internally for analysis. Sample pretreatment and scanning electron microscopy cathodoluminescence (SEM–CL) imaging were performed at Nanjing Hongchuang Geological Exploration Technology Service Co., Ltd. (NHEXTS), Nanjing, China.

Mineral composition by EPMA

The current and beam spot used by the electron probe to test different minerals were different, i.e., 10 nA and 10 μm for feldspar, and 2 nA–5 nA and 20 μm for apatite. The same mineral was affected by the size, ring and other factors, and the current and beam spot used in different situations were different. The concentrations of apatite were determined using a JEOL JXA–8230 electron probe microanalysis (EPMA) instrument. The accelerating voltage, beam current and beam diameter were 15 kV, 5 nA and 20 µm, respectively. The peak counting times were 10 s for Si, Ti, Al, Fe, Mn, Mg, Ca, Na, K, P, F, and Cl and 20 s for S. The background counting time was half of the peak counting time. All the data were corrected for matrix effects using the ZAF correction method (Donovan and Tingle 1996; Cao et al. 2013). The following standards were used: jadeite (Si), rutile (Ti), yttrium Al garnet (Al), hematite (Fe), rhodonite (Mn), olivine (Mg), apatite (Ca), albite (Na), orthoclase (K), apatite (P), fluorite (F), tugtupite (Cl), and celestite (S). A 1–10 μm electron beam was used according to the size of the mineral shards, and the analytical uncertainties were less than 2%. The data are listed in Table 1.

Table 1 Main element oxides and volatile contents of CBS–12a and CBS–8a apatites by EMPA (wt%). FeOT represents total iron (oxide) content
Full size table

Trace element contents of apatite by LA–ICP–MS (including Sr and Nd)

Trace element analysis and Sr–Nd isotope analysis were performed at Nanjing Hongchuang Geological Exploration Technology Service Co., Ltd. (NHEXTS), Nanjing, China. In situ trace element analysis of apatite was conducted using LA–ICP–MS. The analysis was completed in the super purification laboratory. The Resolution SE model laser ablation system (Applied Spectra, USA) was equipped with an ATL (ATLEX 300) excimer laser and a two-volume S155 ablation cell. The laser ablation system was coupled to an Agilent 7900 ICP–MS (Agilent, USA). The detailed tuning parameters are provided in Thompson et al. (2018). LA–ICP–MS tuning was performed using a 50 micron-diameter line scan at 3 μm/s on a NIST 612 instrument at ~ 3.5 J/cm2 with a repetition rate of 10 Hz. The gas flow was adjusted to obtain the highest sensitivity (238U ~ 6 × 105 cps) and the lowest oxide ratio (ThO/Th < 0.2%). P/A calibration was conducted with NIST 610 using a 100 micron-diameter line scan. Preablation was conducted for each spot analysis using 5 laser shots (~ 0.3 μm in depth) to remove potential surface contamination. The mass was analyzed with a total sweep time of ~ 0.26 s. The analysis was performed using a 38 μm-diameter spot at 5 Hz with a fluence of 3 J/cm2. The Iolite software package was used for data reduction (Paton et al. 2011). NIST 610 and 43Ca were used to calibrate the trace element concentrations as an external reference material and internal standard element, respectively. The analytical uncertainties were less than 2%. The data are presented in Tables 2 and 3, Supplementary Tables 1 and 2.

Table 2 Trace element and rare-earth element contents in CBS–12a and CBS–8a apatites by LA–ICP–MS in parts per million (ppm, or μg·g-1)
Full size table
Table 3 REE data and calculated values from apatite trace element concentrations (ppm)
Full size table

In situ Sr isotope measurements were performed on a Neptune Plus MC–ICP–MS (Thermo Fisher Scientific, USA) instrument in combination with a Resolution SE 193 nm laser ablation system (Applied Spectra, USA). The analysis was completed in the super purification laboratory. The MC–ICP–MS interface was modified using a high-efficiency dry pump to increase the ion transmission efficiency. The JET sample and X skimmer cones were used along with the guard electrode (GE). The Faraday cup configuration of the ion detection system was composed of an array from L4 to H3 aligned to monitor Kr, Rb, Er, Yb and Sr (Tong et al. 2016). The determined 87Sr/86Sr of the NBS 987 standard was 0.710245 ± 0.000025 (2σ, n = 12), which is consistent with the results determined via thermal ionization mass spectrometry (TIMS) with an 87Sr/86Sr of 0.710236 (Jochum et al. 2006; Paton et al. 2007; Yang et al. 2014a). The Durango apatite standard was analyzed every 10–15 samples to monitor the instrument stability.

In situ Nd isotope measurements were performed on a Neptune Plus MC–ICP–MS (Thermo Fisher Scientific, Germany) instrument in combination with a J–200 343 nm femtosecond laser ablation system (Applied Spectra, USA). The analysis was completed in the super purification laboratory. The JET sample and X skimmer cones were used along with the GE instrument. The instrumental mass bias for Nd and Sm isotopes was corrected using an exponential law function based on a 146Nd/144Nd value of 0.7219 and a 147Sm/149Sm value of 1.0868 (Yang et al. 2014b). The interferences of 143Sm to 143Nd were corrected based on the measured signal intensities of 147Sm, mass bias coefficient and natural isotopic composition. The Durango apatite standard was analyzed for every 10 samples to monitor the instrument stability and external correction of 147Sm/144Nd ratios. The values of (87Sr/86Sr)i, (143Nd/144Nd)i and εNd(t) were calculated from Lugmair and Marti (1978). The Sr–Nd isotope values are summarized in Table 4 and Supplementary Table 3.

Table 4 Contents and isotopic compositions of Sr and Nd for the apatites from CBS–12a and CBS–8a
Full size table
Table 5 LA–ICP–MS U–Pb apatite data for the Naitoushan basalt (CBS–12) samples from the CTV
Full size table

U–Pb dating of apatite and zircon

Apatite LA–ICP–MS U–Pb dating was completed by Yanduzhongshi Geological Analysis Laboratories Ltd. (YGALL). The U‒Pb dating of apatite in this study was performed in the super purification laboratory via an Analytik Jena M90 quadrupole laser ablation multicollector inductively coupled plasma‒mass spectrometry (LA–ICP–MS) instrument with a 193 nm NWR193 Ar–F excimer laser. The downhole fractionation, instrument drift and mass bias correction factors for the Pb/U ratios in apatite were calculated using 2 analyses of the primary (MAD2 Cochrane et al. 2014) and secondary standard apatites. McClure Mt (Schoene and Bowring 2006), Otter Lake (Barfod et al. 2005) and Durango (McDowell et al. 2005) were analyzed at the beginning of the session and after every 10 unknown apatites using the same spot size and conditions as those used for the samples. The trace element contents of apatite were quantified using SRM610 as an external standard and 44Ca as the internal standard element, assuming stochiometric proportions. Each analysis of apatite began with a 20 s blank gas measurement followed by a further 40 s analysis time when the laser was switched on. A flow of He carrier gas at a rate of 0.55 L/minute carried particles ablated by the laser out of the chamber to be mixed with Ar gas and to the plasma torch. The measured isotopes were 31P, 44Ca, 206Pb, 207Pb, 208Pb, 232Th, 235U and 238U; each element was measured every 0.18 s, with a longer counting time for Pb isotopes than for other elements. The data reduction that was used was based on the method outlined in detail in Meffre et al. (2008), with an additional modification to correct for the small amount of common Pb present in the primary standard using the 207Pb correction of Chew et al. (2014). The results of the analyses are listed in Table 5 and Supplementary Tables 4 and 5.

Zircon U–Pb dating was conducted via LA–MC–ICP–MS. The Resolution SE model laser ablation system (Applied Spectra, USA) that was used was equipped with an ATL (ATLEX 300) excimer laser and a two-volume S155 ablation cell. The experiment was completed in the super purification laboratory. The LA system was coupled to a Neptune Plus MC–ICP–MS (Thermo Fisher, USA). The analysis was performed using a 30 μm-diameter spot at 5 Hz and a fluence of 2 J/cm2. For every 10–12 analyses of the sample, 91,500 was analyzed twice, and GJ–1 was analyzed once. Typically, 35–40 s of sample signals were acquired after 20 s of gas background measurement. The 91,500 and GJ–1 zircons were used as primary and secondary reference materials, respectively. The exponential function was used for calibrating the downhole fractionation, the Iolite software package was used for data reduction (Paton et al. 2010), and age correction was performed by adopting the method of Sakata et al. (2017). NIST 610 and 91Zr were used to calibrate the trace element concentrations as an external reference material and internal standard element, respectively. The measured ages of the reference materials in the batch were 1063 ± 0.9 Ma, 2σ (91,500) and 604 ± 1.9 Ma, 2σ (GJ-1), which agree well with the reference values within a given uncertainty. The CBS–8 data are listed in Table 6.

Table 6 LA–MC–ICP–MS zircon U–Pb data for the Heishigou dike (CBS–8) from the CTV
Full size table

Rb–Sr dating of feldspars

Rb–Sr isochron dating of alkaline feldspar and plagioclase in the analysis of basalt was performed at the South Tai Geological Test Institute (STGTI) in Nanjing, China. Before the Rb–Sr isochron dating analysis, we selected fresh test samples with a perfect crystal morphology under a binocular microscope and then accurately measured the contents and isotope ratios of Rb and Sr. The values of 87Rb/86Sr and 87Sr/86Sr were determined by isotope mass spectrometry (VG354, UK). For the detailed experimental analytical process, please refer to Xu et al. 2020. The isotope standard sample was NBS987 (87Sr/86Sr = 0.710236 ± 0.000007), and 87Sr/86Sr = 0.1194 was used for standardization calibration (± 1%, λ(Rb) = 1.42 × 10–11 a–1). All operations were carried out in the super purification laboratory, with process blanks for Rb and Sr of 2 × 10–10 g and 2 × 10–10 g, respectively. The analytical accuracy was better than × 10–5. Isochron ages were calculated according to Brown et al. (1989). The data fitting for the age calculation was performed using Isoplot software by Ludwig (2001). The isochronal regression calculation showed a 2% error for the 87Sr/86Sr ratio and a 0.02% error for the Sr isotope ratio. The data for the alkaline feldspar and plagioclase samples are listed in Table 7.

Table 7 Rb–Sr dating results for alkaline feldspar and plagioclase from CBS–8
Full size table

Results

Summary of the whole–rock geochemistry

In this study, the whole-rock geochemistry of the Naitoushan basalt (CBS–12a) and Heishigou dike (CBS–8a) was comprehensively reported by Ye et al. (2022) and Li et al. (2021a, b). The Naitoushan basalt (SiO2 = 50–52 wt%) shows an alkaline affinity and is characterized by high Al2O3 and MgO contents, with trace element features similar to those of intraplate basaltic magma. They also have moderately enriched Sr–Nd isotope features, with (87Sr/86Sr)i = 0.70491–0.70520 and εNd(t) = − 0.9 to − 2.7. The Heishigou dikes (SiO2 = 48–49 wt%) with OIB affinity are characterized by Na2O > K2O and moderately enriched Sr–Nd isotopic signatures ((87Sr/86Sr)i = 0.70469–0.70450, εNd(t) = − 1.2 to − 1.7). Both have Sr and Nd isotopic compositions similar to those of basalts in the CVF (Li et al. 2021a, b).

Apatite texture

Representative CL images for apatites from CBS–12a and CBS–8a are presented in Fig. 3g and h, respectively. Apatite grains from the Naitoushan basalt (CBS–12a) are euhedral or subhedral and mostly long and columnar in shape (up to 120 μm, Fig. 3g). They commonly occur as inclusions hosted in other major phases, such as plagioclase, indicating that apatite is in an early crystallization phase. Only a few apatite grains have micropores and fractures and are light gray in color.

Apatite grains from the Heishigou dike (CBS–8a) are mostly subhedral crystals (up to 130 μm) and show homogeneous textures (Fig. 3h). They commonly occur as columnar inclusions hosted in plagioclase and phlogopite (Fig. 3f’), with a light gray color in CL images (Fig. 3f’ and h).

Chemical composition of apatite

The apatites in Naitoushan basalt (CBS–12a) are low-F magmatic apatites (Fig. 5c; Sun et al. 2021) that have values from 54.13 to 54.28 wt% CaO and 40.22 to 42.87 wt% P2O5. They have high K2O (0.01–0.06 wt%) and MgO (0.15–0.34 wt%) contents and low FeO (0.19–0.35 wt%) contents (Fig. 5a and b; Table 1). The apatites in the Heishigou dike (CBS–8a) have higher F contents and similar CaO and P2O contents to those in the Naitoushan basalt. Compared to those in the Naitoushan basalt, the Heishigou dike apatites have higher F and lower Al2O3 contents. The Cl contents in the apatites of CBS–12a and CBS–8a (Table 1) are equivalent to those in the lunar high-Ti basalts (Ji et al. 2022).

The apatites from CBS–12a and CBS–8a show large ranges for both halogens (Fig. 5c) and Sr contents (566.40–1390.00 ppm) (Fig. 6a; Table 2, Supplementary Tables 1 and 2). The CBS–12a apatites have high Sr (1247.00–1390.00 ppm). As shown in Figs. 6a and 7, Sr shows negative correlations with La, U, ∑REE and La/Yb but a positive correlation with Eu/Eu*. The CBS–12a apatite also has a narrow range of ∑REEs from 3504.75 to 5022.14 ppm (Fig. 7) and is characterized by light REE/heavy REE (LREE/HREE) fractionation (Fig. 6b), with weak negative Eu anomalies (Eu/Eu* = 0.63–0.69). Compared with CBS–12a apatites, CBS–8a apatites have low Sr contents (566.40–1193.00 ppm) and a wide range of ∑REE values from 5833.76 to 13,020.30 ppm; they also have high La, Yb, and U contents and La/Yb ratios (38.37–46.17) and strong negative Eu anomalies (Eu/Eu* = 0.34–0.69) (Figs. 7 and 8). The CBS–8a apatites show REE distribution patterns that are similar to those of apatites from CBS–12a (Fig. 6b).

All apatite grains from the studied Naitoushan basalt and Heishigou dike are enriched in light REEs (LREEs) relative to heavy REEs (HREEs), with (La/Yb)N ratios ranging from 19.20–20.82 and 25.87–31.81, respectively (Table 3 and Supplementary Table 2). In the primitive mantle-normalized trace element and chondrite-normalized REE diagrams (Fig. 6), all apatite grains are enriched in Th, U, and LREEs and depleted in high field strength elements (HFSEs), with strongly negative Rb, Ba, Nb, Ta, K, Zr, Hf and Ti anomalies. They show positive Sr and negative Pb anomalies relative to those of the whole-rock samples of host rocks. Moreover, the CBS–12a and CBS–8a apatite grains have similar Th/U (3.91–5.27), Zr/Hf (165.89–370.00), La/Sm (3.47–4.25) and Nd/Tb (45.59–52.23) ratios, with no correlation with the total REE contents (Fig. 7a, b, c and d).

In situ Sr–Nd isotopic compositions of apatite

The apatite grains of CBS–12a and CBS–8a have highly variable Sr–Nd isotopic compositions (Supplementary Tables 3 and 4); they have relatively high values of 143Nd/144Nd (0.512396–0.512879) and 147Sm/144Nd (0.11127–0.123507) and low ratios of 87Rb/86Sr (0.00015–0.03994) and 84Sr/86Sr (0.05506–0.05720), with no strong relationship with total REE contents (Fig. 7e, f, g and h). CBS–12a apatite grains generally have variable initial 87Sr/86Sr ratios and εNd(t) values of 0.70485–0.70599 and − 4.72 to 4.70, which give an age of 22.2 Ma (Fig. 9a, b and c); these values are significantly distinct from those of CBS–8a apatite grains, which have relatively low initial 87Sr/86Sr ratios (0.70482–0.70533) and positive εNd(t) values (0.46 to 3.80), with an age calculated at 0.218 Ma.

Apatite U–Pb geochronology

The U–Pb data of thirty–two apatite grains from sample CBS–12 (Supplementary Tables 4 and 5) plotted on a Tera–Wasserburg diagram (Fig. 4a) show discordant ages with rather high proportions of common Pb and 207Pb/206Pb ratios between 0.77 and 0.85. They yield a lower intercept age of 18.7 ± 2.6 Ma with an initial 207Pb/206Pb value of 0.8837 (Fig. 4a). The weighted average 207Pb-corrected ages are consistent at 22.2 ± 2.9 Ma (Fig. 4b).

Fig. 4
figure 4

a Tera–Wasserburg concordia and b weighted average 207Pb-corrected age plots for LA–ICP–MS U–Pb apatite analyses from the Naitoushan basalt. c Tera–Wasserburg concordia age plot for LA–MC–ICP–MS U–Pb zircon analyses from the Heishigou dike. d Rb–Sr isochron age diagram for alkaline feldspar and plagioclase from the Heishigou dike

Full size image

Zircon U–Pb dating

Zircons from the Heishigou dike (CBS–8) are light gray with few inclusions. They are euhedral columnar crystals with sizes of 55–180 μm. They exhibit homogeneous luminescence and occasionally display oscillatory zoning (Fig. 4c). Th/U ratios range from 0.7 to 2.0, and the zircons have a magmatic origin (Hoskin and Schaltegger 2003). The 8 analyses of zircons from CBS–8 provide 230Th disequilibrium-corrected ages of 0.325–0.360 Ma (Table 6). Tera–Wasserburg concordia diagrams are used to calculate the lower intersection age (0.2301 ± 0.0049 Ma, Fig. 4c).

Rb–Sr dating of alkaline feldspar and plagioclase

The Rb and Sr concentrations and isotopic compositions from alkaline feldspars and plagioclases in the eight CBS–8 samples are presented in Table 7. All analyzed samples have relatively low Rb and Sr contents of 12.89–227.8 ppm and 2.139–449.6 ppm, respectively. The 87Rb/86Sr ratios vary from 0.0954 to 295.3, the (87Sr/86Sr)i values range from 0.70354 to 0.70361, and the initial 87Sr/86Sr ratio is 0.703573 ± 0.000044. In the 87Rb/86Sr vs. 87Sr/86Sr diagram (Fig. 4d), the data yield a Rb–Sr isochron age of 0.218 ± 0.021 Ma (mean standard weighted deviation (MSWD) = 0.24, n = 8).

Discussion

Ages of the Naitoushan basalt and Heishigou dike

All the apatite grains from the Naitoushan basalt and Heishigou dike are euhedral or subhedral, and most are long and columnar in shape (Fig. 3g and h). Their Th/U ratios range from 1.9–5.2, indicating a magmatic origin. Compared with the previously published age of the Naitoushan basalt (K − Ar age of 13 − 16 Ma, alkaline feldspars; Liu 1988), our LA–ICP–MS U–Pb age for apatite is considerably more accurate because apatite has a low crystallization temperature that represents the final cooling age of the magma (Chew et al. 2014). The apatite U–Pb analyses yield an age of 18.7 ± 2.9 Ma for the lower intercept (n = 32, Fig. 4a) and a weighted mean 207Pb-corrected age of 22.2 ± 2.9 Ma for CBS–12a (Fig. 4b), implying that the large-scale eruption of magma occurred in the Miocene Aquitanian stage (22.2 Ma) and that the final cooling time of the magma occurred in the Burdigalian stage (18.7 Ma). All dated zircons from the Heishigou dike (CBS–8a) are euhedral columnar crystals and display oscillatory zoning, and their Th/U ratios range from 0.7 to 2.0, implying a magmatic origin. The single-grain LA–MC–ICP–MS zircon U–Pb intercept age is 0.2301 ± 0.0049 Ma (n = 8; Fig. 4c). Our data for alkaline feldspar and plagioclase Rb–Sr isochron ages of 0.218 ± 0.021 Ma (n = 8; Fig. 4d) correspond well with the emplacement age of the zircon; thus, we conclude that the studied Heishigou dike was emplaced in the Chibanian stage of the Pleistocene. The age of CBS–12a is definitely older than the age of CBS–8a, indicating that the Heishigou dike was emplaced after the Naitoushan basalt. This conclusion is consistent with the geological relationships because the Heishigou dike intruded the earlier and larger irregular Naitoushan basalt (Fig. 2).

As previously reported, the last three main stages of CTV volcanism are the shield stage (2.8 to 0.3 Ma), cone construction stage (1.5 to 0.01 Ma), and caldera-forming stage (0.2 Ma to present). The preshield stage mainly resulted in Gaima Plateau basalts (such as the Naitoushan basalt). Therefore, our new geochronological data also show that the Heishigou dike was emplaced before the pre-shield/shield/cone-forming stage. Furthermore, the NE–SW strike of the Heishigou dike also reveals the structural discontinuity affecting the CTV, i.e., the Tianchi–Zengfengshan fault, and the preferred alignment of the scoria cones (Figs. 1c and 2).

Origin of apatite from the Naitoushan basalt and Heishigou dike

CBS–12a apatite is in a late crystallization phase in the Naitoushan basalt; it occurs as euhedral or subhedral grains between plagioclase grains. This apatite generally displays higher Sr and lower Cl concentrations and shows negative correlations with La, Yb, U, REEs, La/Yb and Eu/Eu* (Fig. 8a, b, c, d, e and f), indicating simultaneous plagioclase, zircon and monazite accumulation during magmatic evolution (Zhang et al. 2019; Kendrick et al. 2014). Our previous study based on whole-rock Sr–Nd–Pb isotopes indicated that the magma with OIB attributes played an insignificant role in the formation of the Naitoushan basalt (Li et al. 2021a, b). Accordingly, the variations in the Th/U (4.05–4.31), Zr/Hf (191.93–370.00), La/Sm (3.59–3.98), Nd/Tb (45.59–47.78), 143Nd/144Nd (0.512396–0.512879), and 147Sm/144Nd (0.11127–0.121146) ratios in CBS–12a apatite are small, indicating limited crustal contamination of the parental magma. The ratio of Sr isotopes in apatite can better reflect the non-contaminated original initial magma (such as by alteration or source variation), or they can indicate a lower degree of contamination than shown by whole-rock geochemistry (Charlier et al. 2006). For CBS–12a, the 87Sr/86Sr (0.70485–0.70599) and 84Sr/86Sr (0.05591–0.05720) ratios are lower than those of the whole-rock samples (Fig. 9a, b, and c), indicating that these apatites were produced through fractional crystallization of magma.

The negative relationships of U, La, Yb, REEs, La/Yb and Sr in the CBS–12a apatite indicate that olivine + pyroxene + plagioclase ± monazite ± zircon-dominated fractionation contributed to magmatic evolution (Pan et al. 2016; Fig. 8a, b, c, d and f). The Eu/Eu* values in the CBS–12a apatite increased with increasing Sr (Fig. 8e), reflecting the influence of fractional crystallization of feldspar (Ladenburger et al. 2016). The experimental results also showed that the partition coefficients of REEs between apatite and melt were easily affected by the composition of the magma, i.e., an increase in the SiO2 content (Prowatke and Klemme 2006). In a magmatic system, feldspar fractionation could lead to decreasing Sr and increasing SiO2 and REEs during the magmatic evolutionary process, thereby increasing the partition coefficients of REEs. Correspondingly, apatite crystallizing from the magma could contain high ∑REEs (Fig. 6b) and low Sr contents (Fig. 6a). The CBS–12a apatites have lower ∑REE contents and Sr/Y values and higher Sr/Y and lower Y ratios than those from CBS–8a (Figs. 7, 8 and 9d), indicating that the basaltic magma underwent fractional crystallization to a limited degree (Sun et al. 2022). The results also reveal significant decreases in La, Yb and La/Yb with Sr but increases in F/Cl (Fig. 5c), Th/U, Zr/Hf, La/Sm and Nd/Tb with REEs (Fig. 7a, b, c and d), which is consistent with feldspar fractionation (Zhang et al. 2020).

Fig. 5
figure 5

a Al2O3–Na2O–K2O, b TiO2–MnO–MgO and c Cl–F diagrams for the apatite samples from the Naitoushan basalt and Heishigou dike. Other data for apatites from trachytes in the CVF are from Wei (2014)

Full size image

The CBS–8a apatites have Th/U (3.91–5.27), Zr/Hf (165.89–316.51), La/Sm (3.47–4.25), Nd/Tb (47.73–52.23), 143Nd/144Nd (0.512653–0.512824), 147Sm/144Nd (0.119867–0.123507), 87Sr/86Sr (0.70482–0.70533) and 84Sr/86Sr (0.05506–0.05649) ratios and almost identical (87Sr/86Sr)i (0.70482–0.70533) and εNd(t) (0.46–3.80) values to those of the OIB–EMI (Armienti and Gasperini 2007) (Fig. 9a, b and c), further indicating a weak fractional crystallization relationship between the Heishigou dike and the Naitoushan basalts with increasing REEs. According to the primitive mantle-normalized trace element patterns and chondrite-normalized REE patterns, the whole-rock composition of CBS–8a has the same trend and is slightly lower than that of the Naitoushan basalt, suggesting that the Heishigou dike and Naitoushan basalt may have originated from the same magma with a continuous weak fractional crystallization process; this interpretation is also supported by the right-leaning distribution of rare earth elements (Yan et al. 2023), as shown in Fig. 6. The volume of Heishigou dike rocks in outcrop is quite small relative to that of the Naitoushan basalt (Fig. 2), which is also consistent with a fractionation model, where the mafic magma generally has a much smaller dike volume than basalt (Chandrasekharam et al. 2000; Petcovic and Grunder 2003). The continuous compositional variations in apatite between the Heishigou dike and the Naitoushan basalt (Figs. 6, 7, 8 and 9d) are also consistent with the trends defined by weak fractional crystallization. In general, CBS–8a apatites show a closer affinity to CBS–12a apatites, indicating their similar origin. The CBS–8a apatites show lower Al2O3 and MgO contents and higher F contents (Fig. 5); these features are potentially attributed to fractionation of Al-, Mg- and F-bearing minerals, such as clinopyroxene and olivine (Aoki et al. 1981; Busa et al. 2002; Beyer et al. 2012). Alternatively, the relationships of Sr with La, Yb, REEs, U and La/Yb indicate simultaneous crystallization of other REE- and U-rich minerals, accompanied by the formation of apatite. Generally, plagioclase in magma generally has a high Sr content and low element diffusion rate (Costa et al. 2003), so the Sr isotope ratios in the analyzed apatite can reflect mineral crystallization in the parental magma, and the 87Sr/86Sr and 84Sr/86Sr values in apatite (Fig. 8g and h) slightly decrease with increasing REEs in response to feldspar fractionation during magmatic evolution (Belousova et al. 2001; Bruand et al. 2017). To produce the whole-rock geochemical results of the Heishigou dike and the trace element and Nd isotopic compositions of apatite, the Heishigou dike is likely derived from an enriched mantle with εNd(t) values (Fig. 9a, b and c) similar to those of the Naitoushan basalt.

Fig. 6
figure 6

a Primitive mantle-normalized trace element patterns and b chondrite-normalized REE patterns (McDonough and Sun 1995) for CBS–12a and CBS–8a apatite grains and host rocks. Other whole-rock data are from Yan et al. (2023)

Full size image
Fig. 7
figure 7

Plots of a La, b Yb, c REEs, d La/Yb, e Eu/Eu*, and f U vs. Sr contents for the apatite grains from CBS–12a and CBS–8a

Full size image
Fig. 8
figure 8

Plots of a Th/U, b Zr/Hf, c La/Sm, d Nd/Td, e 143Nd/144Nd, f 147Sm/144Nd, g 87Sr/86Sr, and h initial 84Sr/86Sr ratios vs. REE values for the apatite grains from CBS–12a and CBS–8a

Full size image
Fig. 9
figure 9

a and b Plots of εNd(t) vs. (87Sr/86Sr)i for apatite grains from CBS–12a and CBS–8a (Sun et al. 2014). c (143Nd/144Nd)i vs. (87Sr/86Sr)i plot of apatite grains and volcanic basalts from the CVF (Zhang et al. 2018). d Sr/Y vs. Y plot of apatites (Sun et al. 2022)

Full size image

Implications for the petrogenesis of basalts

Apatite grains in the host basalts have similar occurrences, geochemical features, and Sr–Nd isotopic compositions, indicating that they have the same origins. The apatites of the Naitoushan basalt (CBS–12a) crystallized from magma under equilibrium or near-equilibrium conditions, with no other late crystallization phases (Webster and Piccoli 2015). They have narrow initial 87Sr/86Sr and 143Nd/144Nd ratios (0.70485–0.70599 and 0.512396–0.512879, respectively) and variable εNd(t) values (− 4.72 to 4.70), and their Sr–Nd isotopic compositions are similar to those of host basalts (Fig. 9a, b and c; Ye et al. 2022), indicating that these apatites crystallized from the host magma. The apatites from the Heishigou dike (CBS–8a) have relatively variable (87Sr/86Sr)i and (143Nd/144Nd)i ratios (0.70482–0.70533 and 0.512639–0.512810, respectively) and εNd(t) values (0.46–3.80) and an age of 0.218 Ma. Their Sr–Nd isotopic compositions are generally similar to those of the host magmatic apatite grains of the Heishigou dike and Naitoushan basalt (Supplementary Tables 3 and 4), indicating that they potentially crystallized from the same magma source. Furthermore, apatite can also be inherited from source materials, thus providing insight into the origins of magma or petrogenesis (Watson and Green 1981; Wolf and London 1994; Dempster et al. 2003). The Sr–Nd isotopic compositions of the CBS–12a and CBS–8a apatites, which are similar to those of the host Naitoushan basalt and Heishigou dikes, agree with an inherited origin of the Rb–Sr and Sm–Nd isotope systematics of magmatic melt (Chappell et al. 1987; White et al. 1999). Therefore, we infer that the CBS–12a and CBS–8a apatite grains may represent host magma originating from a possibly depleted mantle source but modified by crustal contamination or the involvement of enriched components in the source.

Bruand et al. (2020) and Piccoli and Candela (2002) suggested that accessory minerals not only affect the trace element chemistry of host rocks but also record changes in the magmatic composition during magmatic evolution, i.e., magmatic differentiation. The relatively narrow range of initial 87Sr/86Sr ratios and εNd(t) values (Fig. 9a, b and c) in the CBS–12a and CBS–8a apatites indicates that there was limited crustal assimilation and that no subducted sedimentary material was involved in the genesis of apatite. Apatites from CBS–12a and CBS–8a display strong correlations between Th/U, Zr/Hf, La/Sm, Nd/Tb, 143Nd/144Nd, 147Sm/144Nd, 87Sr/86Sr, 84Sr/86Sr and REE contents (Fig. 7); these characteristics of magmatic apatite are attributed to the crystallization of minerals, with progressive depletion of the remaining magmatic melt in trace elements (Miles et al. 2013). The most striking geochemical features of apatites in CBS–12a and CBS–8a that crystallized from magma are their variable Sr anomalies (Fig. 8), with contents ranging from 1390.00 to 566.40 ppm. The Sr content is generally controlled by various factors, such as the crystallization of feldspars or the oxidation state of the host magma (Ramos et al. 2004; Yuan et al. 2018). For variable oxidation states, Ce and Eu are usually presumed to exist in large quantities in apatite (Piccoli and Candela 2002); however, correlations between Ce or Eu anomalies and the oxidation state of their host magmas have not been reported. There are no significant Ce anomalies (Ce/Ce* = 1.06–1.00, in apatite, Supplementary Tables 4 and 5) and no MnO content variations (~ 0.7 wt%; Table 1) in the apatites from CBS–12a and CBS–8a. The crystallization of plagioclase effectively reduces Sr and Eu contents from magma (Ching–Oh et al. 1974), and the Sr contents and Eu anomalies decrease with increasing amounts of plagioclase (Figs. 6 and 7). The apatites in CBS–12a and CBS–8a that crystallized from the host magma have variable medium REE (MREE) contents (1157.56–456.68), with Nd/Tb ratios weakly increasing with increasing REE contents (Fig. 7d). Titanite and zircon have high mineral/melt partition coefficients for MREEs (Sm + Eu + Gd + Tb + Dy; Sun et al. 2022), and during crystallization, they are also enriched in U; thus, their crystallization in host basaltic magmas can potentially explain the increasing Nd/Tb ratios of apatites.

Apatites in CBS–12a and CBS–8a have decreasing Th, U and REE contents and Sr/Y ratios; weakly increasing Th/U, La/Sm and Nd/Tb ratios; and relatively uniform Zr/Hf and Sr–Nd isotope ratios (Figs. 6, 7 and 8). Therefore, these results can potentially indicate coeval crystallization of titanite and other minerals, such as feldspar and monazite (Bingen et al. 1996; Lloyd et al. 1996; Bruand et al. 2016), and depletions in Hf, Zr, Nb, and Ta (Fig. 6a) can potentially be caused by the cocrystallization of zircon (Nardi et al. 2013). This interpretation is also consistent with the petrographic characteristics of euhedral–subhedral apatites enclosed in plagioclase (Fig. 3e, f and f’). Their variations in Eu anomalies (0.69–0.34) are attributed to weak plagioclase crystallization or fractionation in the magma chamber, with an EMI-type nature. Furthermore, the whole-rock apatite samples from the Naitoushan basalt and Heishigou dikes have relatively high REE and trace element contents (Fig. 6b) and relatively evident negative Eu anomalies, further indicating fractional crystallization of feldspars and accessory minerals. In addition, a slab of the Pacific plate in the MTZ (mantle transition zone) and a large mantle wedge (BMW) below NE China have been confirmed by seismic imaging data (Tian et al. 2016). Beginning in the middle Cenozoic, the subducted plate gradually subsided into the MTZ, resulting in water-rich melt, and slab-derived fluids may have metasomatized the overlying mantle (Li et al. 2021a, b). Combined with the geochemistry of the gas emitted in NE China (Xu et al. 2013) and of the Cenozoic basalts on the continental margin of NE China (Zhang et al. 2009), the EM1 source magma could have originated in the subcontinental lithospheric mantle (SCLM), so the EMI signature of the studied dikes and basalts in the CVF can be explained by the back-arc extension and rifting caused by the rollback of the Pacific plate (Kuritani et al. 2009; Zheng et al. 2018). The tomography data in NE China (Zhu et al. 2019) also clearly show a thickening of the Moho due to the accumulation of mantle melts at the base of the crust below the CVF. Therefore, we propose that the dikes and basalts in the CVF occur at the Moho depth.

Based on these results, these crystal features of minerals, geochemical and isotopic characteristics of apatites and whole-rock compositions of the Naitoushan basalt and Heishigou dike represent a continuous phase that crystallized from mafic magmas that came from a magma source with OIB–EMI affinity.

Conclusions

Based on the integrated in situ oxide, volatile element and Sr–Nd isotope analyses of apatite and geochronology of the Naitoushan basalt and Heishigou dikes from the CTV, the following conclusions can be drawn:

  1. 1.

    The ages of the Naitoushan basalt and Heishigou dike in the Changbaishan area are 22.2 ~ 18.7 Ma and 0.231 ~ 0.218 Ma, respectively. This indicates that the basaltic magmatism between the preshield and postshield stages of the CTV potentially lasted for 22 Ma until the cone formation stage.

  2. 2.

    The apatites from the Naitoushan basalt and Heishigou dike are characterized by low FeO, MnO, TiO2, F, and Cl contents and high MgO and K2O + Na2O concentrations, indicating their crystallization from alkaline mafic host magmas with limited volatile contents.

  3. 3.

    The chemical and Sr–Nd isotopic compositions from whole-rock and apatite samples of the Naitoushan basalt and Heishigou dike collectively indicate that the preshield and postshield magmas were derived from an enriched mantle source with EMI-type attributes, related to the rollback of the Pacific plate. The host magmas of apatites were controlled mainly by weak crystallization of olivine, pyroxene, plagioclase, and other cocrystallizing accessory minerals.