Early Jurassic monzogranite-tonalite association from the southern Zhangguangcai Range: Implications for paleo–Pacific plate subduction along northeastern China

The Lesser Xing’an–Zhangguangcai Range is located in the eastern part of the Central Asian Orogenic Belt (CAOB) (Fig. 1A). The CAOB is considered to result from collision between the North China block and Siberia block during the Late Permian (Li, 2006) or Early Triassic (Xu et al., 2009). Unlike other parts of the CAOB, northeastern China was significantly affected by the subduction of the paleo–Pacific plate in the Jurassic (Xu et al., 2009; Wu et al., 2011). Phanerozoic granitic rocks in northeastern China are exposed over an area of ∼200,000 km², and they display depleted Sr-Nd isotopic compositions, which indicate significant Phanerozoic crust growth (Han et al., 1997; Jahn et al., 2000; Jahn, 2010; Wu et al., 2000, 2002, 2003, 2011; Guo et al., 2010). However, the crustal growth model and petrogenesis of these granitic rocks remain controversial. The granitic rocks with depleted Sr-Nd isotopic compositions were considered to be formed by fractional crystallization of mantle-derived melt (Han et al., 1997), mixing of crust-derived and mantle-derived melts (Jahn et al., 2000), or remelting of preexisting juvenile basaltic rocks (Wu et al., 2002).

The Lesser Xing’an–Zhangguangcai Range is considered to result from collision between the Songneng block and Jiamusi block in the Paleozoic or early Mesozoic (Zhou et al., 2009; Xu et al., 2009; Wu et al., 2011; Zhou and Wilde, 2013; Shao et al., 2013). Detailed geochronology works (Wu et al., 2011) indicate that the granitoids in the Lesser Xing’an–Zhangguangcai Range were mostly formed during the Jurassic (190–150 Ma), with a small amount in the Paleozoic, and the Jurassic granitic rocks are considered to have been caused by subduction of the paleo–Pacific plate beneath the Eurasian continent (Wu et al., 2011). The following two geological problems in this region are still debated. (1) Initiation time of the subduction of the paleo–Pacific plate. Based on zircon U-Pb dating of the high-pressure metamorphic mafic rocks, Zhou et al. (2009) argued that the Heilongjiang Complex records the time when northward movement of the combined Mongolia–North China block toward Siberia was waning, and was surpassed by the onset of Pacific accretion from the east. (2) The transformation mechanism from the paleo-Asian tectonic regime to the paleo-Pacific regime. Xu et al. (2009) reported Early Jurassic volcanic rocks from the eastern part of the Jilin-Heilongjiang area, and detailed geochemistry indicates that these volcanic rocks were derived from juvenile crust in an extensional setting. It is intriguing as to whether the early Mesozoic granitic rocks and related volcanic rocks in the Zhangguangcai Range were caused by subduction of the paleo–Pacific oceanic slab.

In this paper we report zircon U-Pb ages, geochemistry, Sr-Nd-Pb isotopic compositions, and zircon Lu-Hf isotopic composition of an Early Jurassic tonalite-monzogranite suite from the Shihe area, in the southern part of the Zhangguangcai Range. Detailed geochemical investigations indicate that the Na-rich tonalite was derived from juvenile basaltic crust, while the monzogranite was formed by partial melting of sediment–mid-oceanic ridge basalt (MORB) mélanges in the subduction accretion complex. These results can promote our understanding of the early Mesozoic subduction process of the paleo–Pacific oceanic slab along northeastern China.

Northeastern China is located in the eastern part of the CAOB (Şengör et al., 1993; Jahn et al., 2000; Wu et al., 2011), where previous work revealed two stages of evolution under different tectonic regimes (Wu et al., 2002, 2011; Li, 2006; Xu et al., 2009). Northeastern China consists of the Erguna terrane in the northwest, the Xing’an and Songliao terranes in the center, and the Liaoyuan terrane in the southeast (Fig. 1B). The Lesser Xing’an–Zhangguangcai Range is located at the boundary between the Songliao and Jiamusi terranes. The Songliao terrane is overlain by the Mesozoic Songliao basin, and most of the basement beneath the Songliao basin is composed of Paleozoic–Mesozoic granitoids and Paleozoic strata (Wu et al., 2000, 2011; Xu et al., 2013) with minor Proterozoic granitoids (Pei et al., 2011). The Jiamusi massif consists of the Mashan and Heilongjiang complexes, and the deformed granitoids in the Mashan complex have a peak metamorphic age of ca. 500 Ma (Wilde et al., 2000). Detailed zircon U-Pb dating of the blueschist facies metamorphosed pillow basalts (Zhou et al., 2009) indicates that these basaltic rocks are Late Triassic, and are considered to be a mélange along the suture between the Jiamusi massif and Songliao terrane (Wu et al., 2011; Zhou et al., 2009; Zhou and Wilde, 2013).

Voluminous granites and some Proterozoic to Paleozoic metamorphic rocks occur in the Zhangguangcai Range. Proterozoic strata in the Zhuangguangcai Range include the Dongfenshan and Xingdong groups. Paleozoic strata include the Early Devonian Heilonggong Formation. Mesozoic strata consist of the Triassic Fengshantun Formation, the Middle Jurassic Taiantun Formation, the Late Jurassic Maoershan Formation, and the Early Cretaceous Banfangzi, Guanghua, Ningyuancun, and Taoqihe Formations (Zhou and Wilde, 2013).

The Shihe pluton is located in the eastern part of the Shangzhi area in the southern Zhangguangcai Range (Fig. 2), which was considered to result from collision between the Songneng and Kiamusze block (Zhou et al., 2009; Zhou and Wilde, 2013; Shao et al., 2013). The granites and associated volcanic rocks intruded into the Permian Tangjiatun formation, which consists of deformed intermediate-felsic volcanic rocks. According to field observations, monzogranite and tonalite were in the Shihe pluton, while the mafic enclaves were mainly hosted in the tonalities (Fig. 3). The sampling locations are shown in Figure 2; the monzogranite samples were collected from the western Shihe village (44°52′22″Ν, 128°40′26″Ε), while the tonalite samples were collected from the eastern Shihe Village (44°51′43″N, 128°42′30″E). The monzogranite display medium- to coarse-grained textures, and mainly consist of alkali feldspar (35–40 vol%), plagioclase (25–35 vol%), biotite (10–15 vol%), quartz (25–30 vol%), and accessory minerals including zircon, apatite, and magnetite. The plagioclase is 1.0–3.0 mm long and exhibits well-developed twinning and concentric zoning (Fig. 3C). Biotite is the predominant mafic mineral in the monzogranite; most of the biotites were corroded and display subeuhedral shapes (Fig. 3C). The tonalite samples from the Shihe area are medium grained, and consist of plagioclase (35–45 vol%), alkali feldspar (15–20 vol%), quartz (5–15 vol%), amphibole (10–15 vol%), and accessory minerals including zircon, apatite, and magnetite (Fig. 3D). Plagioclase in the tonalite has oligoclase composition; they display the albite twin law, and amphibole fragments were included in the large oligoclase crystals. Mafic enclaves hosted in the tonalite range from 10 to 60 cm (Fig. 3B), are round, and have transitional contacts with the host tonalite. They display hypidiomorphic texture, with large amount of biotite clusters and magnetite (Figs. 3E, 3F); considering these petrographic features, we argue that these mafic enclaves may represent a mafic mineral cumulate during the crystallization process of the tonalitic magma.

All of the analyses for this paper were performed at the State Key Laboratory of Continental Dynamics, Northwest University, China. For major and trace element analysis, fresh chips of whole-rock samples were powdered to 200 mesh using a tungsten carbide ball mill. Major and trace elements were analyzed using X-ray fluorescence (Rikagu RIX 2100) and inductively coupled plasma–mass spectrometry (ICP-MS) (Agilent 7500a), respectively. Analyses of U.S. Geological Survey and Chinese national rock standards (BCR-2, GSR-1, and GSR-3) indicate that both the analytical precision and accuracy for major elements are generally better than 5%. For trace element analysis, sample powders were digested using an HF + HNO₃ mixture in high-pressure Teflon bombs at 190 °C for 48 h. The analytical precision was better than 10% for most of the trace elements.

Whole-rock Sr-Nd-Pb isotopic data were obtained using a Nu Plasma HR multicollector (MC) mass spectrometer. The Sr and Nd isotopic fractionation was corrected to ⁸⁷Sr/⁸⁶Sr = 0.1194 and ¹⁴⁶Nd/¹⁴⁴Nd = 0.7219, respectively. During the analysis period, the NIST SRM 987 standard yielded an average value of ⁸⁷Sr/⁸⁶Sr = 0.710250 ± 12 (2σ, n = 15), and the La Jolla standard gave an average of ¹⁴³Nd/¹⁴⁴Nd = 0.511859 ± 6 (2σ, n = 20). Whole-rock Pb was separated by an anion exchange in HCl-Br columns, and the Pb isotopic fractionation was corrected to ²⁰⁵Tl/²⁰³Tl = 2.3875. Within the analytical period, 30 measurements of NBS 981 gave average values of ²⁰⁶Pb/²⁰⁴Pb = 16.937 ± 1(2σ), ²⁰⁷Pb/²⁰⁴Pb = 15.491 ± 1 (2σ), and ²⁰⁸Pb/²⁰⁴Pb = 36.696 ± 1 (2σ). The BCR-2 standard gave ²⁰⁶Pb/²⁰⁴Pb = 18.742 ± 1 (2σ), ²⁰⁷Pb/²⁰⁴Pb = 15.620 ± 1 (2σ), and ²⁰⁸Pb/²⁰⁴Pb = 38.705 ± 1 (2σ). Total procedural Pb blanks were in the range of 0.1–0.3 ng.

The zircon grains were separated using conventional heavy liquid and magnetic techniques. Representative zircon grains were hand-picked and mounted on epoxy resin discs, then polished and coated with carbon. Internal morphology was examined using cathodoluminescent (CL) prior to U-Pb and Lu-Hf isotopic analyses. Laser ablation (LA) ICP-MS zircon U-Pb analyses were conducted on an Agilent 7500a ICP-MS equipped with a 193 nm laser, following the method of Yuan et al. (2004). The ²⁰⁷Pb/²⁰⁶Pb and ²⁰⁶Pb/²³⁸U ratios were calculated using the GLITTER data reduction software program (http://www.glitter-gemoc.com/), which was corrected using Harvard zircon 91500 as external calibration. These correction factors were then applied to each sample to correct for both instrumental mass bias and depth-dependent elemental and isotopic fractionation. The detailed analytical technique was described in Yuan et al. (2004). Common Pb contents were evaluated using the method described by Andersen (2002). The age calculations and plotting of concordia diagrams were performed using ISOPLOT (version 3.0; Ludwig, 2003). The errors quoted in the tables and figures are at the 2σ levels. In situ zircon Hf isotopic analyses were conducted using a Neptune MC-ICP-MS equipped with a 193 nm laser. During analyses, a laser repetition rate of 10 Hz at 100 mJ was used and spot sizes were 44 μm. The ¹⁷⁶Yb/¹⁷²Yb value of 0.5887 and mean Yb value obtained during Hf analysis on the same spot were applied for the interference correction of ¹⁷⁶Yb on ¹⁷⁶Hf (Iizuka and Hirata, 2005). The detailed analytical technique was described by Yuan et al. (2008). During analyses, the ¹⁷⁶Hf/¹⁷⁷Hf and ¹⁷⁶Lu/¹⁷⁷Hf ratios of the standard zircon (91500) were 0.282294 ± 15 (2σ, n = 20) and 0.00031, similar to the commonly accepted ¹⁷⁶Hf/¹⁷⁷Hf ratio of 0.282302 ± 8 and 0.282306 ± 8 (2σ) measured using the solution method. The notations of ε_Hf(t) value, f_Lu/Hf (f—different constant of Lu and Hf), single-stage model age (T_DM1—depleted mantle), and two-stage model age (T_DM2) are as defined in Yuan et al. (2008).

Zircons from the monzogranite (STH-1), tonalite (STH-2), and mafic enclaves (STH-2E) that were hosted in the tonalite from the Shihe pluton were selected for LA-ICP-MS U-Pb dating and Lu-Hf isotopic analysis. Zircon cathodoluminescence (CL) images and U-Pb isotopic compositions of the zircon from the Shihe tonalite and monzogranite are presented in Figure 4, and the U-Th-Pb isotope data are listed in Table DR1 in the GSA Data Repository¹.

Zircons from monzogranite (STH-1) are euhedral, light yellowish-brown to colorless, long prismatic crystals, with aspect ratios of 3:1–5:1. Most of the grains display well-developed oscillatory zoning (Fig. 4A). Of 36 spots, 7 (4, 8, 12, 16, 21, 28, and 29) display disconcordant U-Pb ages. Other 29 spots have Th = 117–1144 ppm and U = 163–1507 ppm, with Th/U ratios of 0.29–1.45. These grains have ²⁰⁶Pb/²³⁸U ages of 184 ± 6 Ma to 212 ± 6 Ma (Fig. 4D), yielding a weighted mean age of 201 ± 2 Ma (mean square of weighted deviates, MSWD = 1.2, 2σ), which represents the crystallization age of the monzogranite in the Shihe area.

Zircons from the tonalite (STH-2) are subhedral to euhedral, light yellowish-brown to colorless, with crystal lengths of 100–150 μm and aspect ratios of 1:1–2:1. Most of the grains are gray and display well-developed oscillatory zoning in CL images (Fig. 4B). Of 36 spots, 11 display disconcordant U-Pb ages. The other 25 spots have U contents of 273–1404 ppm and Th contents of 172–1173 ppm, with Th/U ratios of 0.28–1.20. They also have concordant ²⁰⁶Pb/²³⁸U ages of 188 ± 3 Ma to 213 ± 6 Ma, yielding a weighted mean age of 198 ± 3 Ma (MSWD = 3.2, 2σ), representing the crystallization age of the tonalite in the Shihe area.

Zircons (STH-2E) from the mafic enclaves hosted in the tonalite display subeuhedral to euhedral, light yellowish-brown to colorless crystals; as shown in the CL images, most grains display oscillatory zoning (Fig. 4C). Of 36 spots, 14 (1, 2, 8, 12, 13, 16, 17, 18, 22, 29, 30, 32, 33, and 35) display disconcordant U-Pb ages. The other 22 spots display concordant U-Pb ages, have Th = 133–3031 ppm, U = 180–1880 ppm, and high Th/U ratios of 0.51–1.13, suggesting a magmatic origin. These 22 spots have ²⁰⁶Pb/²³⁸U ages of 186 ± 3 Ma to 230 ± 6 Ma (Fig. 4F), yielding a weighted mean age of 198 ± 3 Ma (MSWD = 1.1, 2σ). The geological significance of this age should be carefully considered; according to the field and microscope photo of the mafic enclaves (biotites clusters in the mafic enclaves), we argued that the zircons in the mafic enclaves were from the host tonalite.

Major and trace element analyses of monzogranite and tonalite from the Shihe pluton are listed in Table 1.

Three monzogranite samples display high SiO₂ (73.21–74.68 wt.%) contents, and low TiO₂ (0.19–0.21 wt.%) and Al₂O₃ (13.90–14.40 wt.%) contents, low A/CNK values (1.03 to 1.04). The monzogranite samples are sodic, with Na₂O = 4.30–4.53 wt.%, K₂O = 3.57–3.83 wt.%, and Na₂O/K₂O = 1.12–1.27, all the samples plot in the calc-alkaline field. As shown in the trace element ratios versus SiO₂ or Sr_N diagrams, the monzogranite samples display a significant compositional gap with the tonalite (Fig. 5). In the chondrite-normalized rare earth element (REE) patterns (Fig. 6A), the monzogranite samples show slight enrichment in light REEs, with low (La/Yb)_N ratios of 6.6–11.4, and negative Eu anomalies of 0.66–0.75; in addition, they have low total REE contents (ΣREE) of 91.22–122 ppm. Furthermore, the monzogranite samples display flat heavy REE patterns, (Dy/Yb)_N = 0.84–0.99. In primitive mantle–normalized element diagrams (Fig. 6C), the monzogranite samples show troughs in Nb, Ta, Sr, P, and Ti, and spikes in Rb, Ba, Th, and K, with K/Rb = 296–355, Ba/Th = 81–97, Rb/Sr = 0.45–0.51, and Nb/Ta = 8.27–12.23, similar to granites derived from continental crust (Pitcher, 1997).

Eight tonalite samples from the Shihe pluton have relatively low SiO₂ (63.78–67.39 wt%) contents, high TiO₂ (0.35–0.44 wt%), Al₂O₃ (17.18–19.48 wt%), and high A/CNK values of 1.02–1.04. The samples have Na₂O = 5.92–6.72 wt%, K₂O = 1.66–2.22 wt%, and Na₂O/K₂O = 2.90–3.54. They have MgO contents of 0.69–0.83 wt% and Fe₂O_3total = 2.32–2.75 wt% with Mg# ranging from 36.7 to 40.1. In the chondrite-normalized REE diagram, the tonalite samples display enrichment in light REE (Fig. 6C), with (La/Yb)_N ratios of 7.4–10.5. Most of the samples (except STH-2-1) display positive Eu anomalies (Eu/Eu* = 1.05–1.40). The tonalite also have flat heavy REE patterns with low (Dy/Yb)_N ratios of 0.62–1.07; their ΣREE contents (117–216 ppm) are slightly higher than those of the monzogranite samples. As shown in the primitive mantle–normalized spider diagrams (Fig. 6D), the tonalite samples are enriched in Rb, Ba, and Th, and depleted in Nb, Ta, Sr, and Ti, displaying the typical geochemical features of crustal-derived rocks or island-arc volcanic rocks (Wilson, 1989). Compared to the monzogranite, the tonalite samples have high Sr (479–617 ppm) contents, variable Y contents of 16.9–35.5 ppm, K/Rb ratios of 214–307, variable Nb/Ta ratios of 8.27–24.61, and Ba/Th = 32.88–128.58.

Whole-rock Sr-Nd-Pb isotopic compositions are given in Tables 2 and 3. Initial isotopic values were calculated according to the LA-ICP-MS zircon U-Pb dates for the monzogranite and tonalite. Whole-rock Nd model ages were calculated using the model of DePaolo (1981).

The monzogranite (STH-3, STH-4) samples display higher Rb (89.7–100 ppm) and lower Sr (176–203 ppm) contents; they have (⁸⁷Sr/⁸⁶Sr)_i = 0.704118–0.704207, high ¹⁴³Nd/¹⁴⁴Nd ratios of 0.512905–0.512929, ε_Nd(t) values of +8.3 to +11.7, and single-stage Nd model ages of 0.05–0.30 Ga (Table 2). The tonalite samples (STH-2-3, STH-2-5) have high Sr contents (580–583 ppm), with (⁸⁷Sr/⁸⁶Sr)_i = 0.704625–0.704631, ¹⁴³Nd/¹⁴⁴Nd ratios of 0.512572–0.512878, ε_Nd(t) values of +1.9 to +3.3, and two-stage Nd model ages of 0.65–0.75 Ga.

As shown in the ε_Nd(t) versus (⁸⁷Sr/⁸⁶Sr)i diagram (Fig. 7A), the monzogranite and tonalite samples display Sr-Nd isotopic compositions similar to mid-oceanic ridge basalt (MORB) (Tribuzio et al., 2004; Xu and Castillo, 2004), adakite rocks from Cenozoic subduction zones (Defant, 1990; Sajona et al., 2000), and adakite derived from partial melting of newly underplated basaltic crust (Petford and Atherton, 1996) and clearly differ from the Triassic igneous rocks in the northern margin of the North China craton (Yang et al., 2007, 2012), suggesting a juvenile source region.

In the ²⁰⁶Pb/²⁰⁴Pb–²⁰⁷Pb/²⁰⁴Pb and ²⁰⁶Pb/²⁰⁴Pb–²⁰⁷Pb/²⁰⁴Pb diagrams (Figs. 7B, 7C), the monzogranite and tonalite plot in the transitional zone between lower continental crust and MORB, also indicating a juvenile source region. The monzogranite samples have (²⁰⁶Pb/²⁰⁴Pb)i = 18.463–18.525, (²⁰⁷Pb/²⁰⁴Pb)i = 15.558–15.562, and (²⁰⁸Pb/²⁰⁴Pb) = 37.924–37.928. The tonalite samples have (²⁰⁶Pb/²⁰⁴Pb)i = 18.401–18.641, (²⁰⁷Pb/²⁰⁴Pb)i = 15.555–15.561, and (²⁰⁸Pb/²⁰⁴Pb) = 37.945–37.958 (Table 3).

Zircons from the tonalite and monzogranite dated by U-Pb were also analyzed for Lu-Hf isotopes of the same domain, and the results are listed in Table 4. Initial ¹⁷⁶Hf/¹⁷⁷Hf ratios and ε_Hf(t) values of the magmatic Jurassic zircons were calculated according to their ²⁰⁶Pb/²³⁸U ages. Figure 8A shows ε_Hf(t) values versus crystallization ages of the monzogranite and tonalite in the Shihe pluton.

We selected 30 zircon grains from the monzogranite (STH-1) for Lu-Hf isotope analysis. Of 30 spots, 7 (4, 8, 12, 16, 21, 28, and 29) display disconcordant U-Pb ages, and their Lu-Hf isotopic compositions have no geological significance. The other 23 grains display depleted Hf isotopic compositions with ε_Hf(t) values ranging from +7.8 to +18.2, with corresponding single-stage Hf model ages of 120–542 Ma. The #26 spot has the highest ε_Hf(t) values (+18.2) with a single-stage Hf model age of 120 Ma, which is significantly younger than the zircon U-Pb age. Furthermore, this spot has a higher ¹⁷⁶Yb/¹⁷⁷Hf ratio of 0.00028; therefore, the data should be used with caution.

Zircons from the tonalite (STH-2) also display depleted Lu-Hf isotopic compositions (Table 4). Of 30 spots, 9 display disconcordant U-Pb ages (Table DR1), and the other 21 grains display depleted Hf isotopic compositions, with ε_Hf(t) values of +9.4 to +17.8 and single-stage Hf model ages ranging from 463 to 133 Ma. Spots #1, #2, and #16 have higher ε_Hf(t) values and younger single-stage Hf model ages of 160–133 Ma, significantly younger than their zircon U-Pb ages; the geological significance of these three spots should be interpreted with caution. As shown in the U/Yb versus Y discrimination diagram (Fig. 8B), the zircons from the monzogranite and tonalite plotted in the transition zone between the oceanic crust and continental crust (Grimes et al., 2007).

Geochemical variations are a common feature in granitic rocks, and the following models have been applied to explain the chemical variations in granitic rocks suites: (1) different partial melting conditions, resulting in different melting reactions from a homogeneous source region; (2) two distinct source regions; and (3) fractional crystallization and assimilation of wall rocks (Pitcher, 1997). Clemens and Stevens (2012) argued that when granitic melts segregate from their source region, the melt may carry small crystals of the peritectic phase assemblage formed in the melting reaction, and this mechanism can be responsible for most of the primary elemental variations in granitic magmas.

The tonalite and monzogranite from the Shihe area have identical formation ages but different geochemical features. The tonalite samples display low SiO₂ and K₂O contents and high Al₂O₃, TiO₂, CaO, and MgO contents, while the monzogranite samples have higher SiO₂, K₂O, and Rb contents and low TiO₂, Fe₂O_3total, and MgO contents (Table 1). The tonalite and monzogranite thus have a distinct compositional gap (Fig. 5); the tonalite samples display higher (Ba/Rb)_N, (Sr/Ce)_N, Eu_N, Sr_N values than those of the monzogranite samples. Plagioclase and alkaline feldspar are the predominant minerals that host Eu and Sr in granitic melt (Rollinson, 1993; Słaby and Martin, 2008); considering the high Ba_N values in the monzogranite, we argue that the geochemical variation between tonalite and monzogranite indicates that the tonalite has a more plagioclase component. This phenomenon indicates that normal fractional crystallization from a homogeneous source cannot account for the chemical variations between the tonalite and monzogranite. Furthermore, the monzogranite samples display more depleted whole-rock Sr-Nd-Pb isotopic compositions and lower total REE contents (91.2–122.3 ppm), suggesting that assimilation of crustal rocks cannot result in higher SiO₂, K₂O, and Rb contents. Therefore, the geochemical variations between the tonalite and monzogranite indicate that they were derived from two distinct source regions.

The geochemistry of granitic rocks provides us with windows into the partial melting process and conditions and materials at unseen depths within the continents (Clemens, 2014). The monzogranite samples from the Shihe area display high SiO₂ (73.21–74.68 wt%), K₂O (3.57–3.83 wt%), and Rb (89.7–100 ppm) contents and low Al₂O₃ contents (13.90–14.40 wt%), with A/CNK values of 1.03–1.04, low Sr (176–203 ppm), and high Rb/Sr (0.45–0.51) ratios, as well as significant negative Eu anomalies of 0.62–0.70. As shown in the Rb/Sr versus Rb/Ba diagram (Fig. 9A), the monzogranite samples display high Rb/Sr and Rb/Ba ratios, suggesting that they were derived from partial melting of graywackes (Sylvester, 1998). The biotites have high K₂O and Rb, and low K/Rb ratios (Clemens, 2014), and the high K/Rb ratios (296–355) of the monzogranite samples suggest that biotite is the predominant residue mineral in their source region. In the Al₂O₃ + CaO + Na₂O + K₂O versus A/CNK diagram (Fig. 9B), all the samples plot in the field of melts derived from graywackes (Patiño Douce, 1999). These geochemical features are identical with the I-type granites that formed in the collisional orogenic belt (Chappell and White, 1992; Chappell et al., 2000). Calc-alkaline to high-K calc-alkaline I-type granites are usually considered to have resulted from partial melting of low to middle continental crust (Chappell and White, 1992; Pitcher, 1997).

However, the monzogranite samples display low ΣREE contents (91.22–122.34 ppm) and (La/Yb)_N ratios of 6.6–11.4, which indicate the low content of accessory minerals in their source region (Bea, 1996; Bea and Montero, 1999; Hoskin et al., 2000). Compared to the tonalite, the monzogranite samples display more depleted Sr-Nd isotopic compositions: (⁸⁷Sr/⁸⁶Sr)_i = 0.704118–0.704207, high ¹⁴³Nd/¹⁴⁴Nd ratios of 0.512905–0.512929, ε_Nd(t) values of +8.3 to +11.7, and single-stage Nd model ages of 0.30–0.05 Ga, as shown in the (⁸⁷Sr/⁸⁶Sr)_i-ε_Nd(t) diagram (Fig. 7A). The monzogranite samples display Sr-Nd isotopic compositions identical to those of the Cenozoic adakite derived from young and hot oceanic crust (Defant, 1990); zircons from the monzogranite also display depleted Lu-Hf isotopic compositions (Fig. 8A), and all the grains display positive ε_Hf(t) values of +7.8 to +18.2, with single-stage Hf model ages of 542–120 Ma. These features also suggest that the monzogranite samples were derived from a juvenile source region; like the zircons from the tonalite, the zircons from the monzogranite also have transitional chemical features between oceanic crust zircon and continental crust zircon (Fig. 8B). In summary, the monzogranite samples display some contradictory geochemical features: their high SiO₂, K₂O, and Rb contents suggest the sediment in their source region, but the depleted Sr-Nd-Hf isotopic compositions and low ΣREE contents indicate a juvenile and depleted source region. These contradictory geochemical features have also been found in Cordilleran-type batholiths (Wyllie, 1977) that formed in the active continental margin. Castro et al. (2010) proposed an alternative view of the genesis of Cordilleran-type batholiths; they argued that melting of the mixture of basaltic rocks (MORB-derived amphibolite) and sedimentary components (biotite-rich metagraywacke) in the mantle wedge can produce Cordilleran-type granodioritic magmas. The depleted Sr-Nd-Hf isotopic compositions and low ΣREE contents of the monzogranite indicate that the primitive melts were derived from melting of juvenile basaltic rocks (MORB), and the subsequent mixing with granitic melts derived from sedimentary rocks can account their high K₂O and Rb contents. Considering the high SiO₂ contents and low Sr contents, it can be inferred that the primitive melts were formed by low-degree melting of basaltic rocks with plagioclase residue in the source region (Rapp and Watson, 1995; Clemens, 2014). Overall, the monzogranite samples from the Shihe area display geochemical features identical to Cordilleran-type batholiths, i.e., the primitive trondhjemitic melts derived from juvenile basaltic rocks mixed with the granitic melts derived from sedimentary rocks formed the monzogranite.

The geochemistry of the granitic rocks was controlled by the temperature-pressure conditions of the source rocks, with or without the involvement of fluids during the partial melting process and subsequent magmatic process, i.e., wall-rock assimilation, magma mixing, and fractional crystallization (Rapp and Watson, 1995; Pitcher, 1997; Brown, 2010, 2013). Experiments have revealed that partial melting of both oceanic and continental crust (including basaltic lower crust and granitic upper crust) can produce granitic melt (Brown, 2001, 2013; Castro et al., 2010). Therefore, the first step is to determine whether the granitic rocks were derived from oceanic crust or continental crust.

The tonalite samples from the Shihe area display high Na₂O/K₂O ratios of 2.84–3.57 and are characterized by high Na₂O (5.92–6.72 wt%) and Al₂O₃ (17.18–19.34 wt%) contents, low K₂O (1.66–2.22 wt%) contents, and relatively depleted Sr-Nd isotopic compositions, with (⁸⁷Sr/⁸⁶Sr)_i = 0.704625–0.704631 and ¹⁴³Nd/¹⁴⁴Nd ratios of 0.512572–0.512878, with ε_Nd(t) values of +1.9 to +3.3. These features are identical to the Na-rich intermediate to silicic rocks from active continental margins (Defant, 1990; Petford and Atherton, 1996; Petford and Gallagher, 2001). Na-rich adakitic rocks that formed in active continental margins are considered to result from partial melting of hot and young oceanic slabs (Kay, 1978; Defant, 1990; Drummond et al., 1996) or melting of newly underplated basaltic crust (Petford and Atherton, 1996). Some have argued that partial melting of mafic lower crust in postcollisional settings can also produce Na-rich adakitic melts (Coldwell et al., 2011; Qian and Hermann, 2013). Compared to the Triassic oceanic crust derived from a depleted asthenosphere (White and Klein, 2014), the tonalite samples display lower ε_Nd(t) values of +1.9 to +3.3, with corresponding model ages of 0.65–0.75 Ga, suggesting that they were derived from Neoproterozoic juvenile basaltic crust. These features are consistent with other Mesozoic granites from the Zhangguangcai Range (Guo et al., 2010; Wu et al., 2011). Furthermore, the tonalite samples have low to intermediate K/Rb ratios of 217–307, suggesting hornblende rather than biotite residue in their source region (Peacock et al., 1994). In combination with their high Al₂O₃/TiO₂ and low CaO/Na₂O ratios (Table 1), it can be inferred that there were limited sedimentary rocks in their source region (Sylvester, 1998). Zircons from the tonalite also display depleted Lu-Hf isotopic compositions (Table 4); some grains even have ε_Hf(t) values >+10. The corresponding single-stage model ages are identical with the zircon U-Pb ages, suggesting depleted mantle materials in their source region (Chauvel et al., 2008). The trace element geochemistry of zircons has been applied to distinguish the zircons that crystallized from continental or oceanic crust (Grimes et al., 2007); zircons from oceanic crust usually have higher Y contents and lower U/Yb ratios. As shown in Figure 8B, zircons from the tonalite mainly plot in the transition zone between continental crust zircon and oceanic crust zircon.

Compared to the typical Na-rich adakites from the subduction zone (Kay, 1978; Defant, 1990; Drummond et al., 1996; Condie, 2005) and trondhjemite-tonalite-granodiorite suites from the Archean craton (Smithies and Champion, 2000; Moyen and Martin, 2012), the tonalite samples from the Shihe area also have high Sr (479–617 ppm) and Ba (481–675 ppm), low Y (17.1–28.2 ppm), and negligible Eu anomalies (Eu*/Eu = 1.05–1.30). However, their lower (La/Yb)_N, (Ce/Yb)_N, and Dy/Yb ratios (Fig. 9C) suggest a hornblende rather than garnet source region (Foley et al., 2002). The tonalite samples also have A/CNK values of 1.02–1.04, displaying some features of S-type granites (Chappell and White, 1974). In summary, according to the geochemical features listed here, we argue that the Early Jurassic tonalite from the Shihe area may have been formed by partial melting of Neoproterozoic juvenile basaltic crust in intermediate- to high-pressure conditions. Their relatively low SiO₂ contents (63.78–67.39 wt%) also require temperatures >1000–1100 °C (Rapp and Watson, 1995).

The Mesozoic to Cenozoic westward subduction of the paleo–Pacific plate beneath the Eurasian continent caused significant geological phenomena in eastern China, i.e., widespread I-type granites and associated volcanic rocks in northeastern China (Wu et al., 2002, 2011; Xu et al., 2009, 2013; Zhou and Wilde, 2013; Wang et al., 2015; Guo et al., 2015), the Mesozoic decratonization process in the North China craton (Zhai et al., 2011), and Mesozoic granites and Cenozoic basalts in southeast China (Zhou et al., 2006; He and Xu, 2012). In northeastern China, the initiation time and tectonic style of the paleo-Pacific subduction are controversial (Wu et al., 2011; Xu et al., 2013; Zhou and Wilde, 2013; Guo et al., 2015). The Late Triassic to Early Jurassic (210–155 Ma) I-type granites in northeastern China were considered to result from the westward subduction of the paleo–Pacific plate (Wu et al., 2011). The Early Triassic (228–201 Ma) bimodal igneous suites from the eastern Heilongjiang province also support the subduction of the paleo–Pacific plate beneath Eurasia (Wang et al., 2015). However, the detailed subduction process and origin of the subduction-related igneous rocks are still not well understood. Guo et al. (2015) reported an Early Jurassic mafic intrusive complex from the Tumen area, and the detailed mineral chemistry and whole-rock geochemistry indicate that these mafic rocks crystallized in a water-saturated parental magma that was comparable to that of the arc mafic cumulates, in combination with other north-south–trending mafic rocks and related I-type granites (Wu et al., 2011; Yu et al., 2012). Guo et al. (2015) also proposed Early Jurassic subduction of the paleo–Pacific oceanic plate beneath northeastern China.

The monzogranite was considered to be derived from partial melting of MORB-sediment mélanges in the subduction zone, while the Na-rich tonalite was formed by partial melting of juvenile basaltic crust in the active continental margin. We propose that this Early Jurassic monzogranite-tonalite association in the Shihe area was formed by the Early Jurassic westward subduction of the paleo–Pacific oceanic crust beneath northeastern China. In the case of oceanic subduction (Fig. 10), melting of sediment in the subduction channel produces leucogranitic melts, and these melts assimilate some basaltic rocks (i.e., oceanic crust in the subduction zone) and induce further melting of the basaltic rocks (Otamendi et al., 2009; Castro et al., 2010). Mixing between the leucogranitic melts and the melts derived from the oceanic crust formed the monzogranite samples that have high SiO₂ contents and depleted Sr-Nd-Hf isotopic compositions. In addition, melts or fluids derived from the subducted crust induced partial melting of the basaltic rocks situated in the mantle wedge, and melts from these basaltic rocks formed the Na-rich tonalite in the Shihe area (Fig. 10).

1. The Early Jurassic monzogranite and tonalite from the Shihe area display nearly identical ²⁰⁶Pb/²⁰⁸U ages of 201 ± 2 (MSWD = 1.2, 2σ) and 198 ± 3 Ma (MSWD = 3.2, 2σ), respectively; these ages are contemporaneous with the volcanic rocks from the eastern Heilongjiang Province.

2. The monzogranite samples display high SiO₂, K₂O, and Rb contents, as well as depleted whole-rock Sr-Nd-Pb isotopic compositions, i.e., ε_Nd(t) = +8.3 to +11.7, with single-stage Nd model ages of 0.30–0.05 Ga. Zircons from the monzogranite also have depleted Lu-Hf isotopic compositions. These contradictory geochemical features suggest that the monzogranite may be derived from melting of MORB-sediment mélanges in the subduction zone.

3. The Na-rich tonalite samples display lower SiO₂ and higher TiO₂ contents and depleted zircon Lu-Hf isotopic compositions. Considering their relatively evolved Sr-Nd-Pb isotopic compositions, it can be considered that the tonalite was derived from juvenile basaltic crust in the active continental margin.

4. The Early Jurassic monzogranite-tonalite association from the Shihe area was caused by westward subduction of the paleo–Pacific plate beneath northeastern China.

This work was supported by the Chinese Geological Survey (project 1212011121085), the Foundation for the Author of National Excellent Doctoral Dissertation of the People’s Republic of China (201324), the National Natural Science Foundation of China (41102037, 41190072), and the Ministry of Science and Technology (MOST) Special Fund from the State Key Laboratory of Continental Dynamics, Northwest University. We thank Kurt Stuewe, Ewa Slaby, and an anonymous reviewer for their constructive comments and kind suggestions.