Granitic magmatism associated with gold mineralization: evidence from the Baizhangzi gold deposit, in the northern North China Craton

Abstract

The relationship between magmatism and gold mineralization has been a topic of interest in understanding the formation of ore deposits. The Baizhangzi gold deposit, located in the northern margin of the North China Craton, is hosted by the Baizhangzi granite (BZG) and provides a case to evaluate the relation between granite and gold mineralization in Late Triassic. In this study, we present petrography, bulk geochemistry, zircon U-Pb isotope and trace elements data, as well as major elements of biotite and plagioclase for the BZG to evaluate the petrogenesis and link with gold mineralization. The BZG comprises biotite monzogranite, biotite-bearing monzogranite and monzogranite (BZGs). Zircon U-Pb geochronology shows that all the granitoids of BZGs were coeval with a formation age of 232 Ma. The granitoids, with high SiO₂, Al₂O₃ and Sr, while low Y and Yb, show adakitic affinity. They are enriched in LILFs (e.g., Rb, Ba, Th, U and Sr) and LREEs, while depletion in HFSEs (e.g., Nb, Ta, P and Ti). The geochemical and mineral chemical data suggest that the granitoids have experienced the fractional crystallization of biotite + plagioclase + K- feldspar + apatite. Crystallization temperature is estimated as ca. 700°C, and pressure is between 0.71 kbar and 1.60 kbar. The monzogranite shows higher values of logfO₂, △FMQ and △NNO than the biotite-bearing monzogranite, ranging from −19.76 to −11.71, −4.93 to +3.67 and −5.48 to +3.11, respectively. The fractional crystallization, together with high fO₂, K-metasomatism and low evolution degree, provided favourable conditions for gold mineralization.

1. Introduction

The intrusion-related gold deposits (IRGDs) exhibit that the gold mineralization is spatially and temporally associated with the granitoids (Goldfarb et al. 2001; Lang & Baker, 2001). They mainly contain intrusion-hosted, proximal and distal deposit styles (Dressel et al. 2018). The characteristics of the latter two types are the relationship between the structural setting and their ore-forming process (Cepedal et al. 2013; Tuduri et al. 2018), for instance, the Fuwen Au-dominated Au-Ag deposit (Xu et al. 2017) and Num˜ao gold deposit (Leal et al. 2022). However, the intrusion-hosted deposits are characterized by the direct effect of magmatism on gold mineralization (Lang & Baker, 2001), such as the Linares deposit (Cepedal et al. 2013), Bilihe gold deposit (Yang et al. 2016) and the Julia deposit (Soloviev et al. 2022). The IRGDs have been intensively studied, and the role of magmatism in gold mineralization is a subject of debate. However, there is a consensus on the basically consistent time of magmatism and mineralization (e.g. Yang et al. 2016; Wang et al. 2020; Liu et al. 2021; Soloviev et al. 2022). In summary, four different aspects have been stressed: (1) the origin of magma, such as reworking of an ancient crust is related to the formation of gold in the Argun massif (Liu et al. 2022); (2) fractional crystallization, for instance, highly fractionated quartz diorite porphyry is related to the Bilihe gold deposit (Yang et al. 2016); (3) composition (type, volatiles, K₂O content and alkalinity), for example adakitic granitoids have genetic links to the Fuwen Au-Ag ore deposit (e.g. Zorin et al. 2001; Xu et al. 2017); ilmenite-series igneous rocks are related to the Bilihe gold deposit (Yang et al. 2016); higher volatiles, K and alkaline are more conducive to the gold mineralization in the northern margin of the North China Craton and Xiaoqinling (e.g. N’dri et al. 2021; Liu et al. 2022); and (4) high oxygen fugacity. Furthermore, Blevin (2004) has assessed the relationships of magmatic-hydrothermal mineralization associated with an igneous suite, suggesting that (1) silica distributions ‘skewed’ to high values are typically associated with Au, (2) K/Rb ratios greater than 200 are associated with Cu-Au systems, (3) the degree and type of fractionation determine the potential for mineralization and the type of mineralization and (4) the relative oxidation state control the compatible or incompatible nature of the ore. These investigations suggest that the composition, magmatic process and physiochemical conditions are effective parameters to assess the relationship between magmatism and gold mineralization.

The North China Craton (NCC) is one of the major gold-producing regions in China, and over 60% of the gold deposits are hosted by or are related to Phanerozoic intrusions (Nie et al. 2004; Deng & Wang, 2016; Yang & Santosh, 2020). These gold deposits have been classified as either IRGDs (Sillitoe & Thompson, 1998) or orogenic gold deposits (Goldfarb et al. 2001; Miao et al. 2005; Goldfarb & Groves, 2011; Groves & Santosh, 2016; Santosh & Groves, 2022). It is, therefore, of great significance to study the relationship between magmatism and gold mineralization.

There are many gold deposits defining gold ore concentration in the eastern Hebei – western Liaoning area in the northern margin of the NCC (Kong et al. 2015). These deposits are mostly medium to large scale in the ore reserves and are distributed as ‘satellites’ around the Dushan batholith (DSB) (Fig. 1b), such as Yu’erya, Jinchangyu, Huajian, Baizhangzi, etc. They are closely related to the Mesozoic alkaline to calc-alkaline intrusive rocks, directly hosted by, or located in the vicinity of igneous rocks with almost the same mineralization age as that of magmatism (Chen et al. 2019; Zhang et al. 2020), indicating a typical IRGD characteristic according to Lang and Baker (2001) and Goldfarb et al. (2001). These gold deposits have long been studied (e.g. Sillitoe & Thompson, 1998; Miao et al. 2008; Chen et al. 2014; Song et al. 2016; Bai et al. 2019), with the understanding that the gold mineralization is related to the magmatism. For example, the Jinchangyu gold deposit was regarded as IRGD type by Sillitoe and Thompson (1998) and further considered to be related to the DSB (Luo et al. 2001), while the Yu’erya gold deposit together with its hosted intrusion is identified as the product of partial melting in the lower crust during early Yanshan Movement (Chen et al. 2014), and the Baizhangzi gold deposit is thought to be related to the Late Triassic magmatism (Miao et al. 2008).

Figure 1. (a) Tectonic map of the North China Craton (Jiang et al. 2013), CAOB – Central Asian Orogenic Belt, SLS – Solonker suture, QDOB – Qinling-Dabie Orogenic Belt. (b) Sketch geological map of the LX district showing the distribution of gold deposits and igneous rocks with ages (modified from Xiong et al. 2017; Zhang et al. 2020). (c) Sketch map of BZG at the 30m level with photos of contact relationship. Intrusive rocks: BJ – Baijiadian, DS – Dushan, DSZ – Dashizhuzi, GJ – Gaojiadian, GSZ – Gushanzi, HK – Hekanzi, JG – Jinbaogou, JJ – Jiajiashan, LW – Luowenyu, MJ – Maojiagou, NX – Niuxinshan, QS – Qingshankou, SJ – Sanjia, TZ – Tangzhangzi, XY – Xiaoyingzi, YE – Yu’erya, YZ-Yangzhangzi, ZJ-Zhaojiazhuang.

The Baizhangzi gold deposit, one of the typical IRGDs, is hosted by the Baizhangzi granite (BZG). Previous studies reported zircon SHRIMP (222 ± 3 Ma) (Luo et al. 2004) and LA-ICP-MS age (233 ± 3 Ma) (Xiong et al. 2017) of the BZG. The ore formation mechanism and metallogenic prognosis (Wang, 1989; Wei et al. 2016; Wang et al. 2018) were also reported. However, the factors which are conducive to the gold mineralization during the magmatic process have not been clarified. In this study, we present detailed petrography, bulk geochemistry, zircon U-Pb isotope and trace element data, as well as EPMA mineral compositions, with a view to evaluate the magmatism associated with gold mineralization to provide guidelines for the future mineral exploration in the area.

2. Geological setting and petrology

2.a. Geological setting

The NCC, as one of the oldest cratons in the world with a basement rocks as old as 3.8 Ga (e.g. Zhai & Santosh, 2011), is bordered by the Central Asian Orogenic Belt (CAOB) in the north, the Qinling-Dabie Orogenic Belt (QDOB) in the south and the Sulu Orogen in the east (Fig. 1a). The basement of the NCC is composed of metamorphosed Archaean and Paleoproterozoic rocks, including tonalite-trondhjemite-granodiorite (TTG) gneisses, amphibolite and a few supracrustal rocks (Lu et al. 2008; Kong et al. 2015). These are overlain by post-Mesoproterozoic and Cambrian-Ordovician marine clastic and carbonate rocks, Middle and Late Carboniferous marine and continental facies, Permian to Triassic fluvial and delta facies sedimentary rocks and Jurassic-Cretaceous sedimentary and volcanic rocks. After the final cratonization of the NCC during the late Paleoproterozoic at about 1.85 Ga (Wilde et al. 2002; Zhao & Zhai, 2013), the craton remained mostly stable prior to Triassic. During Mesozoic, the NCC witnessed intense magmatic activity during six stages: Early Triassic, Middle-Late Triassic, Early Jurassic-earliest Middle Jurassic, Middle-Late Jurassic, Early Cretaceous and Late Cretaceous (Zhang et al. 2014). Among them, Permian-Triassic magmatic rocks are widely distributed in the Yinshan and Yanshan belts along the northern margin of the NCC (Zhang et al. 2010). Middle–Late Triassic intrusive rocks are mainly distributed in the northern and eastern NCC, Liaodong and Korean Peninsula (Wu et al. 2005a, 2007). Early Jurassic–earliest Middle Jurassic igneous rocks are mainly distributed in the southern Yanbian area (Wu et al. 2011) and Western Hills of Beijing in the northern NCC (Liu et al. 2012). Middle–Late Jurassic magmatic rocks are widely distributed in the Yanshan Belt in the northern NCC, the Liaodong area and Jiaodong Peninsula in the eastern NCC, the southern Yanbian-Liaobei area in the northeastern NCC and in the northern part of North Korea (Wu et al. 2011). Early Cretaceous igneous rocks are widely distributed throughout the eastern and central NCC and are considered to be a ‘giant igneous event’ in eastern China (Wu et al. 2005b). Late Cretaceous magmatic rocks are relatively minor and distributed in the eastern part of the NCC (Zhang et al. 2011).

The eastern Hebei – western Liaoning district, located in the northern margin of the NCC, is underlain by gneiss, amphibolite and granulites of the Mesoarchean Badaohe Group (Fig. 1b), with Proterozoic carbonate and Quaternary cover (Miao et al. 2008; Zhang et al. 2020). Dozens of intrusive rocks occur in this area, including the Dushan, Dashizhu, Hekanzi, Xiaoyingzi, Qingshan and BZGs, etc. (Fig. 1b), which belong to three periods of magmatism (Miao et al. 2008): ∼220 Ma, 190 Ma–200 Ma and ∼170 Ma. Among them, the DSB is the largest with nearly 30 km in NE-SW and ∼17 km in E-W and an elliptic shape area of over 450 km² (Miao et al. 2008). Previous zircon U-Pb dating reported ages of 215.3 Ma (Rao et al. 2002), 223 Ma (Luo et al. 2003), 221–222 Ma (Ye et al. 2014), 218 Ma (Jiang et al. 2018), 162 Ma and 170 Ma (Jiang et al. 2018), indicating magmatism during Late Triassic and Middle-Late Jurassic. There are a series of E-, NE- and NNE-trending faults and folds (Fig. 1b, e.g. EW-trending Xinglong-Xifengkou-Qinglong fault, Lingyuan fault extending NE-SW and Malanyu anticline) in the western Liaoning region (Zhang et al. 2020). The E-trending faults are generally parallel to the northern margin of the NCC, which is believed to be related to the southward subduction of the Paleo-Asian Ocean (Davis et al. 2001). The NE- to NNE-trending faults are in general parallel to the Tanlu fault, considered to be related to the subduction of the Pacific plate (Cox et al. 1989; Miao et al. 2008).

2.b. The Baizhangzi granitoid

The BZG, located 3 km north of the Dashizhu (DSZ) granite (Fig. 1b), is the main ore-bearing rock of the Baizhangzi gold deposit. It occurs as a stock or dike along the NNE-trending fault, with an exposed area of 1 km², intruding the Proterozoic clastic rocks and Triassic diorite (Luo et al. 2004; Xiong et al. 2017). Our field investigation and microscopic observation revealed that the Baizhangzi stock intrusion is composed of medium to fine biotite monzogranite, biotite-bearing monzogranite and monzogranite (BZGs) (Fig. 1c). These granites are separated by transitional boundaries (Fig. 1c).

Biotite monzogranite occurs as medium to fine-grained, dark red rocks (Fig. 2a), composed of plagioclase (35–40%), K- feldspar (30–35%), quartz (ca. 20%) and biotite (10–15%) (Fig. 2b). Zircon, apatite and magnetite are accessory minerals. Plagioclase is often euhedral with prismatic to lath-like form, and a few crystals are selectively sericitized (Fig. 2c). K- feldspar is present as sub-euhedral to anhedral. Quartz commonly is anhedral. Biotite occurs as euhedral to sub-euhedral with strong pleochroism from yellowish brown to dark reddish brown, and a few of them are altered into chlorite.

Figure 2. Photographs of BZGs. (a) and (b) Biotite Monzogranite. (c) Sericitization. (d) Biotite-bearing monzogranite. (e) and (f) Monzogranite. (g) Silicification (quartz metasomatic K- feldspar). (h) Potassium (K- feldspar metasomatic plagioclase). Ser – sericite; Bi – biotite; Kf – K- feldspar; Pl – plagioclase; Q – quartz.

Biotite-bearing monzogranite contains much less plagioclase (30–35%) and biotite (5 - 10%), while little more quartz (20–25%) (Fig. 2d). The accessory mineral types and alteration characteristics are the same as those of biotite monzogranite.

Monzogranite is constituted mainly by plagioclase (30–35%), K- feldspar (35–40%) and quartz (25%±), with <5% biotite (Fig. 2e, f). The accessory minerals are zircon, apatite and magnetite. These rocks occur as white or red colours, which is interpreted as silicification (Fig. 2g), and potassium alteration (Fig. 2h).

3. Analytical methods

3.a. LA-ICP-MS Zircon trace elemental and U-Pb dating

Zircon separations were performed at the Xinhang Institute of Surveying and Mapping in Hebei Province through panning, heavy liquid and magnetic separation. Around 180 zircons, with complete crystal shape, with no cracks and no inclusions, were selected under binocular and mounted with epoxy resin on a glass plate and polished. The transparent, reflected light and cathodoluminescence (CL) images were obtained at the Beijing Zhongke Mineral Research Testing Technology Co., Ltd. Zircon trace element and U-Pb analyses were conducted in the Beijing Zirconian Pilot Technology Co., Ltd., Beijing, China, using the UP 213 nm laser ablation system and Agilent 7500A ICP-MS produced by New Wave. Helium gas was used as carrier gas, and the laser beam diameter was 30 μm. One 91500 standard sample was used for calibration of test data at every 10 sample points. Andersen (2002) was used for the calibration of ordinary lead. ICPMSDataCal program (Liu et al. 2010) was used to process the data, and U-Pb age concordance plots were plotted using Isoplot 3.0 (Ludwig, 2003).

3.b. Whole-rock major and trace elements

The least altered samples were selected and crushed to less than 200 mesh. The whole-rock major and trace element compositions were analysed at the Wuhan SampleSolution Analytical Technology Co., Ltd., Wuhan, China. Major element oxides were measured by X-ray fluorescence spectrometer (XRF). Trace elements were determined using an Agilent 7700a ICP–MS. The standards for major and trace elements are GBW07103, GBW07105, GBW07111, and BHVO-2, BCR-2, RGM-2, respectively, and the relative standard deviation is <5% for both of them.

3.c. Electron probe microanalyses

The mineral compositions of the studied rocks were analysed using an electron microprobe analyser (EPMA; JXA-8100, JEOL) with a 15 kV accelerating voltage, 20 nA probe current and 5 μm beam diameter, at the Institute of Geology, Chinese Academy of Geological Science. SiO₂, Al₂O₃, MgO, MnO, CaO, Na₂O, K₂O, FeO, TiO₂, Cr₂O₃ and NiO were analysed. Line and spectrometer crystal used for major element are as follows: Si (Kα, TAP), Na (Kα, TAP), K (Kα, PETJ), Al (Kα, TAP), Mg (Kα, TAP), Mn (Kα, LIFH), Fe (Kα, LIFH), Cr (Kα, LIFH), Ni (Kα, LIFH), Ti (Kα, PETJ) and Ca (Kα, PETJ). Count times were 10s for peak and 5s for background per element. Natural and synthetic minerals used for standardization are as follows: (1) for plagioclase: albite (Na and Al), K- feldspar (K), rutile (Ti), NiO (Ni), Cr₂O₃ (Cr), Fe₂O₃ (Fe), spessartite (Mn), diopside (Ca and Mg), plagioclase/K- feldspar (Si); (2) for biotite: albite (Na and Al), K- feldspar (K), rutile (Ti), NiO (Ni), Cr₂O₃ (Cr), Fe₂O₃ (Fe), spessartite (Mn), diopside (Ca), Biotite (Si and Mg). ZAF corrections were carried out. The estimated precisions for major elements are ±2%.

4. Results

4.a. Zircon trace elemental and U-Pb ages

One sample each from the biotite-bearing monzogranite (No. 0-8) and monzogranite (No. -30-7) were selected for LA-ICP-MS zircon U-Pb dating and trace elements. The analytical data are listed in Supplementary Tables S1, S2.

The zircon grains from both the biotite-bearing monzogranite and monzogranite are colourless and euhedral to subhedral. They show oscillatory zoning, without inherited cores in the CL images (Fig. 3a, c). The abundances of thorium and uranium of analysed spots range from 280 ppm to 1648 ppm, 245 ppm to 1390 ppm, respectively, corresponding to Th/U ratios between 0.92 and 1.35 (>0.4). Zircon shares light rare earth element (LREE)-inclined patterns, with varying contents of LREE, characteristics of the Ce-positive anomalies (Ce/Ce*) and Eu-negative anomalies (Eu/Eu*) (Fig. 3b, d), which are consistent with the characteristics of magmatic zircons (Hoskin & Ireland, 2000). All of such indicate a magmatic origin (Christopher et al. 2003; Jan & Sylvester, 2003; Randall & Stephen, 2003).

Figure 3. U–Pb concordia diagrams with cathodoluminescence (CL) images of representative zircons analysed and zircon chondrite-normalized REE patterns from the BZGs. (a) and (b) Biotite-bearing monzogranite (sample No. 0-8). (c) and (d) Monzogranite (sample No. -30-7). Chondrite values are from Sun and McDonough (1989).

Fifteen zircons from sample 0–8 yielded ²⁰⁶Pb/²³⁸U ages from 230 to 238 Ma, with a weighted mean age of 231.7 ± 3.6 Ma (MSWD = 0.69) (Fig. 3a). Fifteen zircons in sample -30-7 yielded ²⁰⁶Pb/²³⁸U ages from 229 to 241 Ma, with a weighted mean age of 231.7 ± 3.3 Ma (MSWD = 2.9) (Fig. 3c). The two ages are almost identical to each other.

4.b. Whole-rock major and trace element compositions

Two biotite monzogranites, four biotite-bearing monzogranites and seven monzogranites of the BZGs were analysed in this work. Whole-rock major and trace element compositions of these samples are listed in Supplementary Table S3. The following data of major elements are recalculated after removing the LOI.

The SiO₂ content of biotite monzogranite samples ranges from 67.57 to 67.81 wt%, and Na₂O +K₂O ranges from 9.77 to 9.99 wt%. The data fall in the monzogranite field in the QAP diagram (Fig. 4a) and plot to the granite field in the An-Ab-Or diagram (Fig. 4b). The values of Na₂O +K₂O −CaO range from 8.03 to 8.18 wt%, and the two data plot along with the alkaline series in a (Na₂O+K₂O −CaO) versus SiO₂ diagram (Fig. 4c). The content of Al₂O₃ is between 15.75 and 16.38 wt%, with A/CNK values ranging from 0.96 to 0.98, and A/NK values between 1.20 and 1.22, identifying with the metaluminous field in an A/CNK versus A/NK diagram (Fig. 4d). The content of MgO ranges from 0.87 to 1.18 wt%, with the Mg values (nMg²⁺/ (nMg²⁺ +nFe²⁺)*100) from 40.48 to 42.36. The biotite monzogranite shows similar chondrite-normalized rare earth element (REE) patterns (Fig. 5a), with being enriched in LREE (La_N = 356.03–357.66, Yb_N = 6.21–6.53 and La_N/Yb_N = 54.8–57.4) and insignificant Eu anomaly (Eu/Eu* = Eu_N/[(Sm_N) ×(Gd_N)]1/2 = 0.94–1.01). A primitive-mantle-normalized trace element diagram (Fig. 5b) shows negative anomalies for high field strength elements (HFSEs) (such as Nb, Ta, P and Ti) and variable enrichment in large ion lithophilic elements (LILEs) (for instance, Rb, Ba, Th, U and Sr), with high Sr (793–855 ppm).

Figure 4. Classification diagrams of samples from the BZGs. (a) QAP diagram (Le Maitre, 2002). (b) An-Ab-Or diagram (Barker & Millard, 1979). (c) (Na₂O +K₂O-CaO) vs. SiO₂ (Frost et al. 2001). (d) A/NK vs. A/CNK diagram (Manilar & Piccoli, 1989).

Figure 5. Chondrite-normalized rare earth element (REE) patterns (a, c, and e) and primitive-mantle-normalized trace element spider diagrams (b, d, and f) are shown. Chondrite values and primitive-mantle values are from Sun and McDonough (1989).

The monzogranite exhibits higher SiO₂ (from 68.70 to 70.51 wt%), and Na₂O +K₂O (9.87–10.25 wt%). In a QAP diagram (Fig. 4a), they all fall into the monzogranite field, which is consistent with plotted in the granite field in the An-Ab-Or diagram (Fig. 4b). The Na₂O +K₂O −CaO contents vary widely, from 8.15 to 9.02, and the data mainly plot along with the alkalic series in a (Na₂O+K₂O −CaO) versus SiO₂ diagram (Fig. 4c). The Al₂O₃ content ranges from 15.16 to 15.56, with A/CNK and A/NK values 0.88- 0.96, 1.11- 1.16, respectively, corresponding to the metaluminous field in an A/CNK versus A/NK diagram (Fig. 4d). The MgO content ranges from 0.46 to 0.82 wt% (average of 0.66 wt%), with the Mg values (nMg²⁺/ (nMg²⁺ +nFe²⁺) *100) from 29.23 to 47.31. The monzogranite shows similar characteristics in the chondrite-normalized REE diagram (Fig. 5c) and primitive-mantle-normalized trace element diagram (Fig. 5d) with the biotite monzogranite, while the lower Sr (420–612 ppm) and Eu/Eu* (0.77–0.89) relatively, suggesting that the crystallization of plagioclase may have occurred.

The biotite-bearing monzogranite displays a coherent behaviour of major and trace elements with the other rock types (Figs. 4, 5) suggesting that the biotite monzogranite, biotite-bearing monzogranite and monzogranite are likely cogenetic, and magma evolution may have played an important role in their formation.

For the BZGs, we used the samples of Xiong et al. (2017) and Luo et al. (2004).

4.c. Mineral geochemistry

Representative grains of plagioclase and biotite were analysed by microprobe for mineral chemistry.

4.c.1. Plagioclase

The following samples were analysed: thirteen plagioclases from the three biotite monzogranite samples (20BZ125, 20BZ126, 20BZ07), fourteen plagioclases from the three biotite-bearing monzogranite samples (120-1, 19BZ38, 20BZ16) and thirty-five plagioclases from the nine monzogranite samples (150-1, 0-12, 20BZ31, 20BAZ36, 108-12, -30-5, 20BZ01, 19BZ41, 19BZ35). Plagioclase in the BZGs is generally fresh to slightly altered and euhedral to subhedral tabular crystals, exhibiting obvious polysynthetic twinning under cross-polarized light. In this study, unaltered regions of plagioclase were selected for chemical composition analysis; therefore, the analysis results are little or not affected by alteration. Some representative data are listed in Supplementary Table S4.

The major element contents of plagioclase from the biotite monzogranite are as follows: SiO₂ (63.5–67.9 wt%, average 65.7%), Al₂O₃ (19.5–23.0 wt%, average 20.8%), CaO (0.42–3.07 wt%, average 1.67%), FeO (0.08–0.58 wt%, average 0.23%), Na₂O (9.2–11.8 wt%, average 10.6%) and K₂O (0.1–0.8 wt%, average 0.36%), while An values range from An_2.01 to An_14.06, with an average of An_7.84, corresponding to four of them belong to oligoclase and the other nine belong to albite in Ab-An-Or diagram (Fig. 6).

Figure 6. Ternary plots of plagioclase compositions (after Smith, 1974), and the right is a 5× magnification of the projection area of the left.

The plagioclase from the monzogranite has higher SiO₂ (ranges from 64.9 to 68.9wt%, average 67.3%), Na₂O (9.3–12.3 wt%, average 11.3%) and K₂O (0.06–3.12 wt%, average 0.41%); however, lower Al₂O₃ (19.0–21.5 wt%, average 19.9%), CaO (0.006–1.7 wt%, average 0.4%), FeO (0.003–0.66 wt%, average 0.17%) and An values (ranged from An_0.03 to An_7.78, with an average of An_2.02), corresponding to all of them belong to albite (Fig. 6).

The average major element contents of plagioclase from the biotite-bearing monzogranite are as follows: SiO₂ (66.8%), Al₂O₃ (20.2%), CaO (0.9%), Na₂O (10.9) and An_4.22, and all belong to albite (Fig. 6).

The composition and chemical formula of plagioclase listed in Supplementary Table S5 show that the number of Na, Ca, Al, Si from the biotite monzogranite ranges from 0.7922 to 1.0030 (average 0.9095), from 0.0198 to 0.1466 (average 0.0791), from 1.0098 to 1.2033 (average 1.0857), from 2.8205 to 2.9817 (average 2.9087), respectively. There are more Na and Si (average 0.9667, 2.9650, respectively), while less Ca and Al (average 0.0207, 1.0323, respectively) of the monzogranite. The number of cations (Na, Ca, Al, Si) lie between biotite monzogranite and monzogranite (average 0.9370, 0.0420, 1.0534, 2.9467, respectively) from the biotite-bearing monzogranite. The mean chemical formula of plagioclase from three types of rock are (Na_0.9095, Ca_0.0791) [Al_1.0857Si_2.9087O₈], (Na_0.9370, Ca_0.0420) [Al_1.0534Si_2.9467O₈] and (Na_0.9667, Ca_0.0207) [Al_1.0323Si_2.9650O₈], respectively.

On the An vs. SiO₂ Al₂O₃, CaO, Na₂O, Ab, Or diagram (Fig. 7), plagioclase from the biotite-bearing monzogranite mostly plot between biotite monzogranite and monzogranite, exhibiting the same characteristic as the major and trace elements of the BZGs, which also suggests that the BZGs are products from cogenetic magma.

Figure 7. Correlation diagram of plagioclase compositions.

4.c.2. Biotite

The EPMA data and chemical formula of biotite are presented in Supplementary Table S6. The number of cations and related parameters is calculated using the Li (Li et al. 2020) method, based on the general formula of A₁M₃T₄O₁₀W₂.

The biotite from the biotite monzogranite shows SiO₂ from 34.61 wt% to 36.32wt% (average 35.67%), Al₂O₃ in the range of 13.54–14.59 wt% (average 13.86 wt%), FeO of 18.73–23.88 wt%, (average 20.97 wt%), MgO of 9.80–12.15 wt% (average 11.55 wt%), TiO₂ of 3.60–4.23 wt% (average 3.92 wt%) and K₂O of 7.98–8.96 wt% (average 8.58 wt%).

The biotite from the biotite-bearing monzogranite shows SiO₂ (35.27wt%), Al₂O₃ (13.29 wt%,), FeO (21.23 wt%), MgO (12.92 wt%), TiO₂ (4.56 wt%) and K₂O (7.59 wt%).

The biotite from the monzogranite has the highest SiO₂ (36.92wt%) and MgO (13.16 wt%), while the lowest FeO (18.24 wt%) and TiO₂ (4.56 wt%).

Fe /(Fe+Mg) ratio ranges from 0.44 to 0.58, suggesting that they were not been altered by fluids (Stone, 2000), consistent with falling in the primary biotite field in the 10×TiO₂-FeO-MgO diagram (Fig. 8a). The Mg/(Fe+Mg) ratio is relatively high, corresponding to magnesium biotite in the Mg-(Al^VI+Fe³⁺+Ti) –(Fe²⁺+Mn) diagram (Fig. 8b). The Mg/(Fe+Mg) ratio increases gradually from the biotite monzogranite (average 0.49) to the monzogranite (0.56), indicating that the biotite became more Mg-rich as the magma evolved.

Figure 8. Genesis (a, after Nachit et al. 2005) and classification (b, after Foster, 1960) diagram of biotite.

5. Discussion

5.a. Magmatic evolution

The contents of TiO₂, Al₂O₃, FeO, CaO, P₂O₅ and MgO (Fig. 9a–f) decrease with increasing SiO₂ from the biotite monzogranite to the monzogranite, attributed to the separation of some minerals. As previously mentioned, the biotite monzogranite is composed of plagioclase, K- feldspar, quartz, biotite, minor zircon, apatite and magnetite. Among them, biotite in these rocks has high FeO, MgO, Al₂O₃ and TiO₂ (Table S6). Significant fractionation of biotite can cause the variation as observed in Fig. 9. The relatively stable concentration of TiO₂ with decreasing Nb/Ta ratios (Fig. 10a) precludes any major role of magnetite, as the separation of magnetite would lead to a negative correlation between Ti and Nb/Ta (Green & Pearson, 1987; Wang et al. 2017). Apatite fractionation was weak during the magmatic evolution, because the P₂O₅ content shows only slight decrease from the biotite monzogranite (average 0.25%) to the monzogranite (average 0.14%). The P₂O₅ also shows decrease with respect to the REE (Fig. 10b). Plagioclase is a major mineral in the monzogranite and has high Al₂O₃, CaO and Na₂O (Table S4). The obvious separation of plagioclase, supported by lower Sr, Eu/Eu* for the monzogranite, the weak decrease in Rb concentrations and decrease of Ba concentrations with decreasing Sr concentrations (Fig. 10c, d), the increasing Rb/Ba and Rb/Sr ratios with decreasing Eu and Ba concentrations (Fig. 10e, f), are also typical. It has been known that fractional crystallization of plagioclase could result in decreasing of Sr of the residual magma, while the declining Ba is due to the separation of K- feldspar, attributed to the high partition coefficient of Sr for plagioclase and Ba for K- feldspar. The lower Ba (average 1284 ppm) and Sr (average 510 ppm) concentrations for the monzogranite than the biotite monzogranite (average 1637 ppm and 824 ppm, respectively) also indicate fractional crystallization of plagioclase and K- feldspar. Collectively, the BZGs are the products of magmatic evolution, and fractional crystallization of biotite, plagioclase, K- feldspar and apatite is the major magmatic process.

Figure 9. Variation of ω(SiO₂) % vs. ω(TiO₂) %, ω(Al₂O₃) %, ω(FeO_T) %, ω(CaO) %, ω(P₂O₅) %, ω(MgO) %.

Figure 10. Diagrams of (a) TiO₂ (wt%) versus Nb/Ta ratios, (b) P₂O₅ (wt%) versus REE (ppm) (after Wang et al. 2017), (c) Sr (ppm) versus Rb (ppm), (d) Sr (ppm) versus Ba (ppm), (e) Eu (ppm) versus Rb/Ba ratios and (F) Ba (ppm) versus Rb/Sr ratios, showing the fractional crystallization of different minerals. See Supplementary Table S7 for the partition of those minerals. Symbols are: Mt = magnetite; Sp = spinel; Hb = hornblende; Pl = plagioclase; Kf = K- feldspar; BI = biotite; Ap = apatite.

5.b. Petrogenetic model

The BZGs have high SiO₂(≥56 %), Al₂O₃ (≥14.5%) and Sr(>400 ppm), low Y(<18 ppm) and Yb (<1.9 ppm), corresponding with the high Sr/Y (>35) and La/Yb (>60) features of characteristics of adakites (Defant & Drummond, 1990; Sajona et al. 1993). Also, they plot in the adakite area in the Sr/Y-Y and (La/Yb)_N-Yb_N diagram (Fig. 11). Both of them indicate that our studied rocks have the properties of adakite. Various models have been proposed for the genesis of adakites including: (1) partial melting of a subducting slab (e.g. Kay, 1978; Saunders et al. 1987); (2) partial melting of the thickened lower crust (e.g. Yang et al. 2016; Xu et al. 2018); (3) mixing of mantle-derived and crust-derived magma (e.g. Xu et al. 2012; Zhang et al. 2013); and (4) fractional crystallization and assimilation (AFC) of mantle-derived basaltic magma (e.g. Chiaradia et al. 2004; Macpherson et al. 2006; Foley et al. 2013; Dai et al. 2017; Azizi et al. 2019; Temizel et al. 2020; Wang et al. 2020).

Figure 11. (a) Variation of Sr/Y vs. Y. Fields for adakites and island arcs are from Defant and Drummond (1990). (b)Variation of (La/Yb)_N vs. Yb_N. Data are normalized (N) to C1 (chondrite) values of Sun & McDonough (1989). Fields for adakites and island arc magma are from Martin (1999).

In general, adakites formed by partial melting of the subducted slab are characterized by high Na₂O and low K₂O contents (Na₂O >3.5%, K₂O/Na₂O≈0.4) (Richards & Kerrich, 2007), which is not consistent with our samples because K₂O/Na₂O ratios range from 1.12 to 1.13 (average 1.12), 0.99 to 1.53 (average 1.18) and 0.90 to 1.57 (average 1.12), respectively. As a consequence, partial melting of subducted slab is unlikely. Second, there is no evidence for the required regional crustal uplift in response to the delamination of a thickened lower crust (Li et al. 2013). Furthermore, the BZGs are massive and homogeneous, and no direct evidence has been found in the field for magma mixing, such as the presence of mafic enclaves. The limited Sr-Nd isotopic composition variation with (⁸⁷Sr/⁸⁶Sr(t) ranging from 0.70445 to 0.70524, and ϵNd(t) from −7.3 to −1.7) (Xiong et al. 2017), and the lack of any correlation between Sr-Nd and SiO₂ also precludes the magma mixing model (Li et al. 2013).

In summary, we infer that the adakitic BZGs were products of fractional crystallization and assimilation (AFC) of mantle-derived basaltic magma.

5.c. Magmatic physiochemical conditions

5.c.1. Temperature

The distribution of some trace elements in minerals is a function of temperature and is used as a geological thermometer (Albuquerque, 1973). Henry et al. (2005) proposed the thermometers for titanium content in biotite: T = {[ln (Ti)-a-c (X _Mg)³]/b}^0.333. The procedure follows T = temperature (°C); Ti = the number of Ti cations in biotite based on 22 oxygen atoms, calculated using the Luhr method (Luhr et al. 1984); X _Mg = Mg/ (Mg+Fe); a = −2.3594, b = 4.6484×10⁻⁹, c = −1.7283. The calculation results are listed in Supplementary Table S8, indicating that the crystallization temperature of the biotite ranges from 700 °C to 744 °C (average 723 °C). Besides, zircon crystallization temperatures of biotite-bearing monzogranite and monzogranite were estimated using the equations of Waston et al. (2006): [log₁₀Ti in zircon (ppm)] = (6.01 ± 0.03)–(5080 ± 30)/T(K), listed in Table S2. The biotite-bearing monzogranite shows zircon temperature in the range of 679–732°C (n = 15, average 705°C). The zircon temperatures of monzogranite are relatively concentrated (except for the -30-7-6 of 783 °C), ranging from 693 to727°C (n = 14, average 709 °C), suggesting that they have similar high crystallization temperature. Both estimates are consistent with the features shown by the biotite, indicating that the BZGs were formed at a temperature of ca. 700°C.

5.c.2. Pressure

The Al content of biotite is positively correlated with the pressure of granitoids according to Uchida et al. (2007). They proposed the empirical formula: P(kb) = 3.03×T ^Al−6.53 (±0.33) (T ^Al = the number of Al cations in biotite based on 22 oxygen atoms, calculated using the Luhr method) (Luhr et al. 1984). The results show that the BZG emplacement pressure is between 0.71 kbar and 1.60 kbar (average 0.97 kbar), corresponding to paleodepths of 2.7 to 6.0 km under lithostatic pressure.

5.c.3. Oxygen fugacity

The concentration of Fe³⁺, Fe²⁺, Mg²⁺ in biotite can be used to estimate fO₂ (e.g. Dymek, 1983; Wones, 1989). The Fe²⁺-Fe³⁺-Mg²⁺ ternary diagram (Fig. 12) shows that the biotite from BZGs plot in the compositional fields between the Ni-NiO and Fe₂O₃-Fe₃O₄ buffer, and closer to Fe₂O₃-Fe₃O₄ line, indicating a high fO₂ in the environment of biotite crystallization from the BZGs parent magma. Moreover, zircon fO₂ was calculated using the Ti in zircon of Geo- fO₂ software (Li et al. 2019), listed in Table S2. LogfO₂, △FMQ, and △NNO for the biotite-bearing monzogranite vary from −22.39 to −13.44 (average of −17.02), −6.31 to +1.30 (average of −1.67), and −6.88 to +0.74 (average of −2.22), respectively. The monzogranite shows higher values of logfO₂, △FMQ and △NNO, ranging from −19.76 to −11.71 (average of −15.14), −4.93 to +3.67 (average of −0.03) and −5.48 to +3.11 (average of −0.58), respectively. These data suggest that the magma may have evolved to a more oxidized state from the biotite-bearing monzogranite to monzogranite.

Figure 12. Fe²⁺-Fe³⁺-Mg²⁺ ternary diagram (after Wones, 1989) of the biotite from the BZGs.

5.d. Potential link between magmatism and mineralization

5.d.1. Conductive factors

Fractional crystallization is considered to be a critical component for mineralization (e.g. Blevin, 2004; Chiaradia et al. 2012; Hronsky et al. 2012). As discussed in previous sections, fractional crystallization played a dominant role during magma differentiation of the BZGs.

Besides, other three parameters have been proposed to evaluate the magmatic-hydrothermal associated with gold-bearing granitoids by Blevin (2004): (1) redox state; (2) compositions, for example, K₂O and alkalinity; (3) degree of magma evolution.

On the redox state classification diagram of Fe²⁺-Fe³⁺-Mg²⁺ (Fig. 12), the studied rocks are closer to Fe₂O₃-Fe₃O₄, indicating that they belong to the magnetite series. The magnetite-bearing granitoids are generally related to Au, Cu and Mo deposits (Ishihara, 1977). Furthermore, the BZGs have a high fO₂, as evidenced by EPMA data of biotite and trace element characteristics of zircon, which are favourable for gold mineralization (Park et al. 2015; Xu et al. 2021).

The BZGs are alkali-rich, with Na₂O + K₂O values ranging from 9.77 to 9.99% (average 9.88%), 9.93 to 10.16% (average 10.07%), 9.87 to 10.25% (average 10.07%), respectively. On the (Na₂O+K₂O −CaO) versus SiO₂ diagram, they all plot as alkaline series (Fig. 5c). On the other hand, potassium alteration is also observed (Fig. 2h), which is a significant feature of gold deposits related to alkaline magmatism (Jensen & Barton, 2000), and is favourable for the transport of gold (N’dri et al. 2021).

The Rb/Sr and K/Rb ratios are two useful parameters to evaluate the degree of compositional evolution. The Rb/Sr ratio of the studied rocks is low (0.16–0.18, 0.14–0.22 and 0.22–0.32, respectively), indicating that these rocks are less evolved (Blevin, 2004). Previously studied Au-Cu deposits are related to low differentiation granites (Thompson et al. 1999; Blevin et al. 2003). In addition, granites associated with Cu and Cu-Au deposits have a K/Rb ratio >200 (Blevin, 2004). The K/Rb ratio is all greater than 200 of the BZGs, ranging from 296 to 318 (average 307), 275 to 326 (average 301) and 306 to 350 (average 324), respectively. The above features suggest that the BZGs are less evolved and conducive to gold mineralization according to Blevin (2004).

5.d.2. Timing of magmatism and gold mineralization

The northern margin of the NCC is an important gold mineralization belt, and there are multiple magmatic events associated with numerous gold mineralization (e.g. N’dri et al. 2021). These gold ores are hosted by both Precambrian metamorphic rocks and Variscan, Indosinian and Yanshanian granites (Zhou et al. 2002). Many gold deposits, related to the Mesozoic intrusive rocks, are distributed in the eastern Hebei – western Liaoning area in the northern margin of the NCC (Kong et al. 2015). These deposits have almost the same mineralization age as that of magmatism (Chen et al. 2019; Zhang et al. 2020), such as Yuerya and Jinchangyu (Wang et al. 2020).

In this study, the zircon U-Pb dates of 231.7 ± 3.6 Ma (biotite-bearing monzogranite sample 0–8) and 231.7 ± 3.3 Ma (monzogranite sample -30-7) have been obtained from the BZGs. They are very similar to those reported previously, where LA-ICP-MS zircon U-Pb ages show 233 ± 3 Ma (Xiong et al. 2017) and SHRIMP zircon U-Pb ages yielded 222 ± 3 Ma (Luo et al. 2004). All these data indicate that the time of magmatism is Late Triassic. We therefore speculate that the Baizhangzi gold deposit, one of the typical IRGDs, was also formed in the Late Triassic. It was also demonstrated by a number of investigated gold deposits hosted by igneous rocks, including Bilihe (Yang et al. 2016), Daxiyingzi (Liu et al. 2021), Julia (Soloviev et al. 2022) and Xiajinbao (Wang et al. 2020).

In summary, the BZGs experienced fractional crystallization, together with high fO₂, alkali-rich, K-metasomatism and low evolution degree, which are conductive for gold mineralization. Furthermore, the monzogranite probably has high metallogenic potential due to its more oxidized state. We expect that this investigation will provide an important guide for regional gold exploration in the area.

6. Conclusions

1. (1) Zircon U−Pb dating shows that the biotite-bearing monzogranite and monzogranite were formed at 231.7 ± 3.6 Ma and 231.7 ± 3.3 Ma, respectively. The dating, supporting previous geochronology, represents the age of magmatism and gold mineralization in the area.

2. (2) Bulk and mineral geochemistry data indicate that the BZGs are cogenetic and fractional crystallization is the dominant magmatic process.

3. (3) Crystallization temperature is ca. 700°C, and pressure is between 0.71 kb and 1.60 kb of the BZGs. Zircon trace elements suggest that the magma may have evolved to a more oxidized state.

4. (4) The fractional crystallization, together with high fO₂, K-metasomatism and low evolution degree, are considered favourable conditions for gold mineralization.