Introduction

The well-known Dongping gold deposit, with a gold reserve of about 100 tons, is located on the northern margin of the North China Craton (NCC) (Fig. 1a, b). It is thought to be genetically related to alkaline magmatism (Mao et al. 2003) as auriferous quartz veins containing disseminated sulfides are hosted mainly by the Late Paleozoic Shuiquangou syenite complex and surrounding Archean metamorphic rocks (Fig. 2; Cook et al. 2009; Cisse et al. 2017; Li et al. 2018a). Dongping has been referred to as an “oxidized intrusion-related gold deposit” that is rich in Te, poor in Cu and Zn, and associated with oxidized minerals and potassic alteration (Helt et al. 2014). Several similar gold deposits have been found in this region (e.g., Hougou and Huangtuliang) and are referred to as “Dongping-type” gold deposits (Bao et al. 2016). Previous researchers have identified two stages of mineralization (~ 380 Ma and ~ 140 Ma; Miao et al. 2002; Bao et al. 2014; Cisse et al. 2017; Li et al. 2018a; Wang et al. 2019b), but the sources of ore-forming metals and the history of fluid evolution of this deposit are still controversial (Nie 1998; Fan et al. 2001; Mao et al. 2003; Bao et al. 2016; Wang et al. 2019a). Furthermore, the relationship of gold mineralization to the host rock is also unclear (Nie et al. 2004; Bao et al. 2014). Previous studies claimed that the Dongping gold deposit has a close relationship with the Shuiquangou syenite (Bao et al. 2014; Jiang and Nie 2000; Li et al. 2000; Luo et al. 2001; Miao et al. 2002), and trace element data in pyrite have been studied in this context (Cook et al. 2009). However, more recent studies speculated that other granitic rocks, such as the Shangshuiquan granite and Archean metamorphic rocks, also provided ore-forming materials at different stages of ore deposit evolution (Li et al. 2018a). Previous work mainly used bulk-rock analysis for sulfur and lead isotopic studies in sulfides (Nie 1998; Bao et al. 2016), which cannot clearly distinguish different pyrite generations and thus is difficult to compare the fluid features of the two mineralization stages.

Fig. 1
figure 1

a Location of the study area (after Zhou and Wang 2012); b simplified geological map of the Dongping region in Chongli County, Hebei Province (after Song and Zhao 1996; Li et al. 2012a, b, c; Gao et al. 2015)

Fig. 2
figure 2

Simplified geological map of the Dongping gold deposit (modified after Li et al. 2010)

Pyrite is one of the most ubiquitous and useful of auriferous minerals in gold deposits. It can provide information on ore-forming processes and ligand and metal sources, and its association with galena, sphalerite, and other sulfides can be used to assess the geochemistry of coeval hydrothermal fluids (Barker et al. 2009; Koglin et al. 2010; Deditius et al. 2014; Tanner et al. 2016). Laser ablation inductively coupled plasma mass spectrometry (LA–ICP–MS) is widely used for in situ analysis to accurately measure the trace element contents in pyrite (Cook et al. 2009; Large et al. 2009; Sung et al. 2009; Bi et al. 2011;Zhang et al. 2014; Yang et al. 2016). Pyrite usually contains a diverse assemblage of trace elements (Co, Ni, Se, Te, Hg, Tl, Au, Ag, Cu, Pb, Zn, and Bi) and isotopes (Pb and S) at concentrations that are within the detection limits of LA–ICP–MS (Reich et al. 2005, 2013; Large et al. 2007; Deditius et al. 2009, 2014; Cook et al. 2013). Gold in pyrite usually occurs in the form of sub-microscopic (“invisible”) gold (< 100 nm; nanoparticle or lattice-bound gold), as micron-scale mineral inclusions, distributed within pyrite fractures, or between grains in the form of native gold (Morey et al. 2008; Large et al. 2009; Pokrovski et al. 2009; Deditius et al. 2011). Native gold may precipitate directly from hydrothermal fluids, or it may form by re-mobilization of earlier-formed pyrite within hydrothermal veins during later mineralization stages (Cook et al. 2009; Sung et al. 2009; Morishita et al. 2018). Pyrite appears to be the dominant gold carrier in the Dongping gold deposit in both mineralization stages, suggesting that pyrite geochemistry can provide useful information about the genesis of this deposit.

In this study, the textural and geochemical characteristics of pyrite that formed during the two main mineralization stages were documented using optical petrography, electron probe micro analysis (EPMA), LA–(MC)–ICP–MS trace element, and in situ S and Pb isotopic analyses. The results show that systematic differences in texture and geochemistry exist among the four different types of pyrite at Dongping, which implies variable sources of ore-forming materials and a complicated fluid evolution history. We aim to use pyrite compositions to infer gold mineralization sources/types, identify gold occurrence and precipitation mechanisms, and gain insights into the genesis of the Dongping gold deposit.

Regional geology

The Dongping gold deposit is situated near the northern margin of the North China Craton (NCC; Fig. 1a), about 10 km to the south of the Shangyi–Chongli–Chicheng fault (Fig. 1b). The exposed stratigraphic units in the Dongping region include the Archean Sanggan Group, Paleoproterozoic Hongqiyingzi Group, Mesoproterozoic Changcheng Series, Cretaceous Zhangjiakou Formation, and Quaternary sediments (Fig. 1b). The Archean Sanggan Group consists mainly of amphibolite, granulite, and gneiss derived from metamorphism of mafic to felsic volcaniclastic rocks and clastic rocks. The Paleoproterozoic Hongqiyingzi Group is composed of marble, quartzite, amphibolite, and gneiss. The Mesoproterozoic Changcheng Series consists of shallow-marine siliciclastic and carbonate rocks and terrestrial siliciclastics that are overlain by minor basalt and felsic volcanic rocks that contain stromatolite and microplant fossils. The Cretaceous Zhangjiakou Formation is composed of rhyolite, sandstone, conglomerate, and pyroclastic rocks.

Igneous rocks ranging from Archean to Cretaceous in age are widely distributed in the Dongping region (Figs. 1b and 2). They include Neoarchean to Paleoproterozoic granodiorite and monzonite (Liu et al. 2006; Li et al. 2012a, b, c), the 380-Ma Devonian Shuiquangou syenite (Luo et al. 2001; Bao et al. 2014), the Cretaceous 140-Ma Shangshuiquan granite, and the Zhangjiakou Formation volcanic rocks (Jiang et al. 2007; Li et al. 2012a, b, c).

The study region was affected by Late Jurassic–Early Cretaceous contraction related to the collision and amalgamation of the Siberia Block and Mongolia-North China Block and subsequent extension caused by rollback of the northwest-dipping Pacific slab beneath North China (Zhang et al. 2010, 2011). Structures in the Dongping region include E–W-, NW-, WNW-, NE-, and N-S-trending faults and parallel folds (Figs. 1b and 2). One of the most significant tectonic discontinuities is the E–W-trending Shangyi-Chongli-Chicheng fault zone, a mantle-penetrating fault (Zhang et al. 2007) that controlled emplacement of the Shuiquangou syenite complex and the distribution of gold mineralization (Jiang and Nie 2000; Li et al. 2000). It is dominated by multistage, ductile to brittle thrusting and is seismically active (Ma and Zhao 1999).

Deposit geology

The Dongping gold deposit is located at 40°50′15″–40°53′45″N latitude and 115°18′15″–115°25′30″W longitude, in the southeastern part of Chongli County, Hebei Province. It consists of ore bodies distributed over an area of about 40 km2 (Fig. 2). The Jianhegou Formation of the Archean Sanggan Group is exposed in the southwestern corner of the mining district (Fig. 2) and is composed mainly of hornblende-bearing gneiss and amphibolite with minor biotite schist and metamorphosed granitoids. The Devonian Shuiquangou syenite complex occupies an area of about 400 km2 and is distributed along an E-W axis with a length of ~ 56 km and a width of 6–8 km (Fig. 1b; Song 1991). It was emplaced into Jiangouhe Formation metamorphic rocks of the Archaean Sanggan Group and is unconformably overlain by Lower Cretaceous Zhangjiakou Group volcanic tuffs in the southeastern corner of the mining district (Fig. 2). The syenite complex is dominated by K-feldspar syenite, hornblende syenite, and quartz syenite (Fig. 2). Magmatic zircons from the syenite complex yielded a concordant age of 382.8 ± 3.3 Ma by LA–ICP–MS U–Pb dating (Li et al. 2010). The Cretaceous Shangshuiquan granite, which has been classified as an alkali granite (Jiang et al. 2009; Cisse et al. 2017), crops out in the southeastern part of the mining district (Fig. 2). It was emplaced into the Devonian Shuiquangou syenite in the north and west, Archean Sanggan Group metamorphic rocks in the south, and Cretaceous Zhangjiakou Formation volcanic rocks in the east. At depth, the Shangshuiquan pluton extends towards the northwest beneath the Dongping ore bodies, as revealed by drilling (Cisse et al. 2017). LA–ICP–MS and SHRIMP zircon U–Pb dating yielded ages of 136–142 Ma for the Shangshuiquan granitic pluton (Miao et al. 2002; Cisse et al. 2017; Li et al. 2018a).

Gold mineralization was mainly controlled by several generations of faults and fractures (Song et al. 1996; Jiang and Nie 2000), particularly NW-striking dextral conjugate faults and NE- to NNE-trending sinistral faults (Fig. 2). It is commonly believed that dilatant faults provided channels for the migration of hydrothermal fluids in the Dongping deposit (Li 1999; Bao et al. 2016).

There are more than 70 gold veins (ore bodies) in the district that vary in size and geometry. The dominant NE- and NNE-trending veins can be subdivided into nine vein systems (mineralization zones), with gaps between them ranging from 400 to 800 m (Fig. 2). Among them, the No. 1 and No. 70 vein systems host > 60 ore bodies and ~ 80% of the total gold reserves of the Dongping deposit (Bao et al. 2014). The gold ores in the entire deposit have an average grade of 6 g/t Au. The No. 1 and No. 70 vein systems partially overlap and are more than 750 m long and 130–300 m wide and extend about 800 m downward. Individual ore bodies (veins) are 200–400 m long and 0.5–40 m wide and extend 100–600 m downward (Fig. 3a, b). They strike 10–40° and dip 20–55° to the NW.

Fig. 3
figure 3

(a) Plan view of the No. 70 ore body; b cross-section of the No. 70 ore body (after Zijin Mining in Chongli, 2011)

Two gold mineralization events

There are various ore types in the Dongping gold deposit as well as quartz veins and silicified and calcitized rocks (Figs. 4a–i). Quartz veins occur mainly in the upper part of the ore bodies and consist of quartz, pyrite, and native gold with minor galena and sphalerite. Silicified and calcitized rocks occur mainly in the lower part of the ore bodies and contain quartz, chalcopyrite, and pyrite veins, in which galena and sphalerite can also be observed. In places, silicified and potassically altered rocks occur in contact with quartz veins. The ores are composed mainly of quartz and feldspar (68.6–98.7%), with only subordinate metallic minerals. Sphalerite and chalcopyrite account only for 1.3–1.4% of the ore, and the total content of sulfur is less than 5%, making the quartz veins sulfur-poor. Minor sericite, calcite, barite, epidote, and clay minerals are present locally (Gao et al. 2017).

Fig. 4
figure 4

Field and underground photographs of two stages of auriferous quartz veins–associated wall rocks in the Dongping gold deposit. a Stage-1 Gy quartz veins in syenite; b Stage-2 Gy quartz veins in syenite and Archean metamorphic rocks; c disseminated pyrite in Stage-1 quartz veins; d–e two stages of cross-cutting quartz veins in syenite; f Stage-2 quartz veins in Shangshuiquan granite; g Stage-2 milky gold-rich quartz-polymetallic sulfide veins in Archean metamorphic rocks; h–i Stage-2 gold-bearing quartz-polymetallic sulfide veins in Shangshuiquan granite. Qtz: quartz; Kfs: K-feldspar; Py: pyrite; Gn: galena; Au: gold

Two mineralization stages can be identified from field observations. Stage 1 produced gray quartz veins in potassically altered wall rocks associated with emplacement of the Shuiquangou syenite at ca. 380 Ma, as dated by hydrothermal zircon U–Pb geochronology (Miao et al. 2002; Cisse et al. 2017; Li et al. 2018a). These gray, narrow (mostly from 5 to 50 cm in width) auriferous quartz veins (Fig. 4a–e) contain coarse-grained cubic pyrite (Fig. 5a–d). These veins show characteristics of pervasive K-feldspathization and silicification. The gold is generally fine-grained and rarely visible in hand specimens.

Fig. 5
figure 5

Microphotographs showing textures and features of Stage-1 and Stage-2 pyrite. a, b Coarse-grained, euhedral to subhedral Py1a that is homogeneous and lacking pores; c, d porous Py1b develops on the edge of Py1a, suggesting replacement of massive pyrite (Py1a) by porous pyrite (Py1b); e, f Py2 shows similar textural relation between Py2a (massive and homogeneous) and Py2b (porous); g Py2a contains native gold and is contemporaneous with galena, chalcopyrite, and other sulfide minerals; h native gold in fractures and pores of Py2b, which is often associated with telluride (e.g., calaverite); i porous Py2b contains more gold than homogeneous Py2a. Qtz: quartz; Kfs: K-feldspar; Py: pyrite; Gn: galena; Ccp: chalcopyrite; Cav: calaverite; Au: gold. Au and As contents shown in d and i are in ppm

Stage 2 polymetallic sulfide–quartz precipitation formed milky quartz veins (Fig. 4f–i) related to intrusion of the Shangshuiquan granite dated at 140 Ma, as dated by hydrothermal zircon and garnet U–Pb geochronology (Miao et al. 2002; Cisse et al. 2017; Li et al. 2018a; Fan et al. 2021). These veins cut through (Fig. 4d, e) or parallel to the first-stage quartz veins, containing pyrite, galena, chalcopyrite, and high-grade gold (Fig. 5e–i). These veins generally have higher gold contents with visible gold grains (Fig. 5g, h). Close to these veins, abundant sulfides are found in the Shangshuiquan granite and Archean metamorphic rocks (Li et al. 2018a). This stage hosts the vast majority of gold in the mining district (Miao et al. 2002; Cook et al. 2009; Bao et al. 2016; Fan et al. 2021).

Sampling and analytical methods

Sample preparation

Ten representative samples were collected from the Stage-1 quartz-pyrite veins (dated at ~ 380 Ma; Li et al. 2018a), Stage-2 quartz-polymetallic sulfide veins (dated at ~ 140 Ma; Fan et al. 2021), and wall rocks in four mine levels (at depths of − 1144 m, − 1184 m, − 1344 m, and − 1390 m; Fig. 3). Thin sections were made from these samples for petrographic, textural, and in situ geochemical, isotopic, and geochronological analyses.

EPMA analysis

Electron microprobe analysis of pyrite was carried out using a Shimadzu EPMA-1720H equipped with an energy dispersive spectroscopy (EDS) system at the Key Laboratory of Metallogenic Prediction of Nonferrous Metals and Geological Environment Monitoring, Ministry of Education based at School of Geosciences and Info-Physics, Central South University (China). The operating conditions of the electron microprobe were an accelerating voltage of 15 kV, a beam current of 20 nA, and an electron beam diameter of ~ 5 µm that was adjusted based on the size of pyrite grains. The following eight elements were analyzed: Fe (Kα), S (Kα), As (Lα), Co (Kα), Ni (Kα), Pb (Mα), Zn (Kα), and Ag (Lα). Mineral and metal standards used for elemental calibrations included chalcopyrite (Fe, S), arsenopyrite (As), metallic cobalt (Co), pentlandite (Ni), galena (Pb), sphalerite (Zn), and metallic silver (Ag) (Liu et al. 2017). All data were corrected using a standard ZAF routine, and the minimum detection limits of the elements are ≤ 0.01 wt.%.

LA–ICP–MS trace element analysis

Trace element analysis of pyrite was conducted by laser ablation inductively coupled plasma mass spectrometry (LA–ICP–MS) at the State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences (Wuhan). Detailed operating conditions of the laser ablation system, the ICP–MS instrument, and data reduction are given in Zong et al. (2017). Laser sampling was performed using a GeolasPro laser ablation system that consists of a COMPexPro 102 ArF excimer laser, with a wavelength of 193 nm and a maximum energy of 200 mJ, and a MicroLas optical system. An Agilent 7700e ICP–MS instrument was used to acquire ion-signal intensities. Helium was used as the carrier gas and argon as the make-up gas, with mixing via a T-connector before entering the ICP-MS. A “wire” signal smoothing device was present in this laser ablation system (Hu et al. 2015). The spot size and frequency of the laser were set at 40 µm and 10 Hz, respectively. Trace element composition of sulfides was calibrated against various reference materials (NIST 610 and NIST 612), and no internal standard was used (Liu et al. 2008). The sulfide reference material of MASS-1 (USGS) was used as the unknown sample to verify the accuracy of the calibration method. Each analysis incorporated a background acquisition of approximately 20–30 s followed by 50 s of data acquisition for the sample. An Excel-based software ICP–MS DataCal was used to perform off-line selection and integration of background and analyzed signals, as well as time-drift corrections and quantitative calibrations of trace element data (Liu et al. 2008). Details of LA-ICP-MS data quality control and related matters are presented in Electronic Supplementary Material (ESM 1).

In situ S isotope analysis

In situ sulfur isotope analysis of pyrite was performed on a Neptune Plus MC–ICP–MS (Thermo Fisher Scientific, Germany) equipped with a Geolas HD excimer ArF laser ablation system (Coherent, Germany) at the Sample Solution Analytical Technology Co., Ltd. (Wuhan, China). In the laser ablation system, helium was used as the carrier gas in the ablation cell and subsequently mixed with argon (the makeup gas). The single-spot ablation mode was used. A large spot size (44 µm) and slow pulse frequency (2 Hz) were used to avoid down-hole fractionation effects (Fu et al. 2016). During each analysis, 100 laser pulses were completed. A new signal-smoothing device was used downstream from the sample cell to efficiently eliminate short-term variation of the signal, especially for the slow-pulse frequency condition (Hu et al. 2015). The laser fluence was kept constant at ~ 5 J/cm2. The Neptune Plus was equipped with nine Faraday cups fitted with 1011Ω resistors. Isotopes 32S, 33S, and 34S were collected in Faraday cups in static mode. The newly designed X skimmer cone and Jet sample cone in the Neptune Plus were used to improve signal intensity. Nitrogen (4 ml/min) was added to the central gas flow to reduce polyatomic interferences. All measurements were performed using medium resolution with a revolving power (as defined by a peak edge width ranging from 5 to 95% of the full peak height) greater than 5000 × . A standard-sample bracketing method (SSB) was employed to correct for instrumental mass fractionation. To avoid matrix effects, a pyrite standard PPP-1 was chosen as reference material for correcting the natural pyrite samples. The reference values for this standard relative to Vienna Canyon Diablo Troilite (VCDT) were reported by Fu et al. (2016). In addition, in-house standards, including a pyrrhotite SP-Po-01 (δ34SVCDT =  + 1.4 ± 0.4‰), a chalcopyrite SP-CP-01 (δ34SVCDT =  + 5.5 ± 0.3‰), and two synthetic Ag2S standards IAEA-S-2 (δ34SVCDT =  + 22.6 ± 0.39‰) and IAEA-S-3 (δ34SVCDT =  − 32.2 ± 0.45‰), were analyzed repeatedly as unknowns to verify the accuracy of the method. Details of the in situ S isotopic ratio analytical procedures were described by Fu et al. (2016).

In situ Pb isotope analysis

In situ lead (Pb) isotope analysis of pyrite was performed on the same MC-ICP-MS instrument at the State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences (Wuhan). During laser ablation, helium carrier gas in the ablation cell was mixed with argon after the ablation cell. The spot diameter ranged from 44 to 90 µm dependent on Pb signal intensity. The pulse frequency was 4–10 Hz, but the laser fluence was kept constant at ~ 5 J/cm2. A new signal-smoothing and mercury-removing device was used downstream from the sample cell to efficiently eliminate the short-term variation of the signal and to remove mercury from the background and sample aerosol particles (Hu et al. 2015). The Neptune Plus was equipped with nine Faraday cups fitted with 1011 Ω resistors. Isotopes 208Pb, 207Pb, 206Pb, 204Pb, 205Tl, 203Tl, and 202Hg were collected in Faraday cups using static mode. The mass discrimination factor for Pb was determined using a Tl solution, nebulized at the same time as the sample, and an Aridus II desolvating nebulizer. The mass fractionation of Pb isotopes was exponentially corrected by 205Tl/203Tl. The optimized values of 205Tl/203Tl were calibrated from measuring two Pb isotope standards, MASS-1 (USGS) and Sph-HYLM (sphalerite, in-house), and were used to replace the natural Tl isotopic composition for the mass fractionation correction of Pb isotopes. The 202Hg signal was used to correct the remaining 204Hg interference on 204Pb, using the natural 202Hg/204Hg ratio (0.230). In addition, the mass fractionation of 204Hg/202Hg was corrected by the 205Tl/203Tl normalization. In this case, we assumed identical mass fractionation factors for 204Hg/202Hg and 205Tl/203Tl. Standard Sph-HYLM was used to monitor the precision and accuracy of the measurements after ten sample analyses, over the entire period of analysis. The obtained accuracy is estimated to be equal to or better than ± 0.2‰ for 208Pb/204Pb, 207Pb/204Pb, and 206Pb/204Pb compared to the solution value determined by MC–ICP–MS, with a typical precision of 0.4‰ (2σ). More details of the in situ Pb isotopic ratio analysis are described in Zhang et al. (2016).

Analytical results

Texture and stages of pyrite

Based on field observations, petrographic, and mineralogical study, pyrite from mineralized quartz veins and adjacent rocks can be assigned to two stages (Py1 and Py2), and the pyrite of each stage can be further assigned to one of two substages (Py1a and Py1b; Py2a and Py2b).

The early-stage pyrite (Py1) is associated with gray-colored, gold-poor pyrite–quartz veins (Stage-1 quartz veins; Fig. 4a–d). It is medium- to coarse-grained and euhedral to subhedral, with cubic or pyritohedral shapes (Fig. 5a, b), and in some cases occurs as large individual crystal in the quartz veins. The Py1 crystals vary from < 10 to 800 µm in size, and some are larger than 2 mm (Fig. 5a). They commonly contain microfractures and holes on their surfaces, and some are homogeneous (Fig. 5a, b). Py1a and Py1b are distinguished based on textural relationships: Py1a is massive and euhedral and possesses a homogeneous characteristic (Fig. 5a, b), whereas Py1b is porous and developed on the edge of Py1a, showing a replacement texture (Fig. 5c, d).

Late-stage pyrite (Py2) occurs in milky, gold-rich, polymetallic sulfide–quartz veins and adjacent wall rocks (Fig. 4g, h, i). Py2 is anhedral to subhedral and exhibits grain sizes ranging mostly from 1 to 200 µm and occasionally up to 300 µm (Fig. 5e–i). The Py2 crystals have a close paragenetic relationship with galena, chalcopyrite, and other sulfide minerals (Fig. 5g). Free gold (Fig. 5g, h) occurs in some disseminated, porous, coarse-grained, and anhedral pyrite grains. Py2a and Py2b are also distinguished based on textural relationships: Py2a is relatively larger, homogeneous, and associated with sulfide minerals, whereas Py2b is porous and has a smaller grain size and higher gold content (Fig. 5e–i).

Major element composition of pyrite analyzed by EPMA

The major element composition of pyrites (Py1a, Py1b, Py2a, and Py2b) was analyzed in 10 representative samples by EPMA (ESM 2 Table 1). The abundance and correlations between these elements are shown in Fig. 6. Sulfur, Fe, Co, and As were uniformly above the detection limit (dl), whereas elements such as Ni, Ag, Cu, Zn, and Pb were sometimes below the dl (ESM 2 Table 1).

Fig. 6
figure 6

Major element compositions analyzed by using EPMA. a Fe–S plot shows a negative correlation from Py1a to Py2b, and higher Fe contents for Py2a and Py2b relative to Py1a and Py1b; b As–S plot shows no distinction among the four types of pyrite; c the Fe–Cu and d the Fe–Ni plots do not show any correlations, and the distributions of the four types of pyrite are similar

Although Stage-1 and Stage-2 pyrites have similar chemical compositions, there are some significant differences. The Fe content range of Py1 (44.7–46.9 wt%) is lower than that of Py2 (45.4–47.6 wt%). Average Fe contents generally increase through the sequence (Py1a 45.8 wt%, Py1b 46.1 wt%, Py2a 46.6 wt%, Py2b 46.4 wt%) (Fig. 6a). The average S contents of Py1 and Py2 are effectively identical (Py1a 53.4 wt%, Py1b 53.4 wt%, Py2a 53.3%, Py2b 53.3 wt%) (Fig. 6b). The average As contents of Py1 and Py2 are similar (Py1a 0.19 ± 0.02 wt%, Py1b 0.2 ± 0.03 wt%, Py2a 0.2 ± 0.02 wt%, Py2b 0.19 ± 0.02 wt%) (Fig. 6b). The average Cu contents of Py1 (Py1a 0.028 ± 0.020 wt%, Py1b 0.030 ± 0.030 wt%) are slightly lower than those of Py2 (Py2a 0.031 ± 0.020 wt%, Py2b 0.032 ± 0.020wt%) (Fig. 6c). The average contents of Ni in Py1 (range: dl–0.085 wt%; average: Py1a 0.02 ± 0.01 wt%, Py1b 0.01 ± 0.02 wt%) are also lower than those of Py2 (range: dl–0.21 wt%; average: Py2a 0.03 ± 0.04 wt%, Py2b 0.02 ± 0.01 wt%) (Fig. 6d). Similarly, the average contents of Ag in Py1 (range: dl–0.04 wt%; Py1a 0.01 ± 0.01 wt%, Py1b 0.01 ± 0.01wt%) are lower than those of Py2 (range: dl–0.052 wt%; Py2a 0.02 ± 0.01%, Py2b 0.02 ± 0.01 wt%). The contents of other trace metals (e.g., Zn and Pb) measured by EPMA are similar in Py1 and Py2.

Trace element compositions of pyrite analyzed by LA–ICP–MS

Twenty-seven trace and rare earth elements (Co, Ni, Cu, Zn, As, Se, Mo, Ag, Sb, Au, Tl, Bi, Pb, La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, and Lu) were measured by LA–ICP–MS (ESM 2 Table 2). Although none of these elements show a systematic change in content from Py1a to Py2b, some are distinct (Figs. 7 and 8).

Fig. 7
figure 7

Trace element compositions of pyrite analyzed by using LA–ICP–MS. a Ag–Au plot shows a trend of increasing Ag and Au contents from Py1a to Py2b. Most Py1a and Py2a plot in the Au/Ag < 1 field, whereas Py2b plots in the Au/Ag > 1 field; b Au–Tl plot shows a positive correlation; c Au–Bi plot shows a positive correlation. Py2b has the highest Au contents among the four types of pyrite; d, e Au–Pb and Ag–Pb plots show positive correlations; (f) Pb–Bi plot indicates that Py2a and Py2b have higher Bi and Pb contents than Py1a and Py1b; (g) Cu–Bi plot of the four types of pyrite; (h) Au–As plot shows no correlation; (i) Ni–Co plot shows that Py2a and Py2b plot mainly below, and Py1a and Py1b plot mainly above, the Co/Ni = 1 line

Fig. 8
figure 8

Chondrite-normalized REE patterns of Dongping gold deposit samples: a Py1a and Py1b; b Py2a and Py2b; c Shuiquangou syenite, altered syenite adjacent to auriferous quartz veins, and Archean country rocks (data from Bao et al. 1996; Bao and Zhao 2003; Zhao et al. 2010); d hydrothermal and magmatic zircons from Shuiquangou syenite and Shangshuiquan granite (data from Li et al. 2018a). REE normalization values are from Sun and McDonough (1989)

The gold contents of Py1 range from 0.003 to 82.4 ppm, and the average gold content increases significantly from 0.06 ppm in Py1a to 9.3 ppm in Py1b (Fig. 7a–e). The As contents of Py1 range from 1.93 to 1031 ppm, and the average As content shows no regular trend between Py1a (66.4 ppm) and Py1b (27.6 ppm) (Fig. 7h). However, the average gold contents of Py2 increase from 1.1 ppm in Py2a to 473.4 ppm in Py2b, higher than those in Py1 (0.1 to 10.0 ppm) (Fig. 7a–e).

The trace element compositions of Py1 and Py2 exhibit both similarities and differences. Specifically, they have similar Zn, Sb, and Tl contents, although Py1b and Py2b have relatively higher concentrations of Au and Ag than Py1a and Py2a (Figs. 7b, 9). Py2 is commonly enriched in chalcophile elements as shown by average concentrations of Ni (85.6 ppm), Cu (446.3 ppm), Se (100.1 ppm), Mo (17.9 ppm), Ag (158.9 ppm), Pb (592.6 ppm), and Bi (2.88 ppm), whereas these elements have lower average concentrations in Py1: Ni (63.9 ppm), Cu (35.9 ppm), Se (23.0 ppm), Mo (0.26 ppm), Ag (9.4 ppm), Pb (28.7 ppm), and Bi (1.13 ppm) (Fig. 7c–g). However, the average contents of Co (783.2 ppm) and Zn (7.0 ppm) in Py1 are higher than those in Py2 (Co: 247.3 ppm; Zn: 4.6 ppm) (Fig. 7i). The remaining trace elements were not compared because of their low concentrations and limited variation.

Fig. 9
figure 9

Spider diagram illustrating trace element compositions of Py1a, Py1b, Py2a, and Py2b

The REE contents for Py1a, Py1b, Py2a, and Py2b are given in ESM 2 Table 3, and chondrite-normalized REE patterns are shown in Fig. 8. The REE patterns are generally similar but also show some differences. Py1a and Py1b have average (La/Yb)N values of < 0.28 and < 0.25, respectively. Py1b has a wider range of total REEs than Py1a (Fig. 8a; ESM 2 Table 3). The HREE distribution shows an obvious interelemental fractionation pattern. A pronounced positive Eu anomaly can be seen, and average Eu/Eu* values of Py1a and Py1b are < 1.16 and < 1.20, respectively. In addition, their average Ce/Ce* values are < 0.45 and < 0.49 for Py1a and Py1b, respectively. In contrast, Py2a and Py2b have relatively higher ΣREE and are HREE-enriched (Fig. 8b). The (La/Yb)N values of Py2a (< 0.38) and Py2b (0.55) are different, and their average Eu/Eu* values are also variable (Py2a: < 1.33; Py2b: 1.04), but their average Ce/Ce* values (Py2a: < 0.51; Py2b: < 0.55) are similar.

In-situ S isotope compositions of pyrite

Pyrite from nine representative samples was selected for S isotope analysis by LA–MC–ICP–MS (ESM 2 Table 4; Fig. 10). Each type of pyrite has a distinct range of S isotope compositions. The average δ34S values of Py1a (4 spots, − 4.3 ± 0.5‰) and Py1b (4 spots, − 4.9 ± 1.6‰) are higher than those of Py2a (12 spots, − 7.0 ± 0.2‰) and Py2b (4 spots, − 6.4 ± 0.5‰).

Fig. 10
figure 10

Histograms of sulfur isotope compositions of a four types of pyrite in the Dongping gold deposit; b sulfide minerals in the Dongping gold deposit; c pyrite from the Dongping, Xiaoyingpan, Huangtuliang, Hougou, and Shuijingtun gold deposits hosted by the Shuiquangou syenite complex. Data are from this study and Wang et al. (1990), Yin (1994), Song and Zhao (1996), Nie (1998), Li et al. (2000), Xing et al. (2011), and Bao et al. (2016)

In-situ Pb isotope composition of pyrite

The Pb isotope composition of pyrite is given in ESM 2 Table 5 and shown in Fig. 11. Py1 is characterized by narrow ranges of 206Pb/204Pb (17.38–17.58), 207Pb/204Pb (15.43–15.50), and 208Pb/204Pb values (37.30–37.58). Py2 has higher 206Pb/204Pb (17.52–17.83) and 208Pb/204Pb values (37.51–37.71) and almost identical 207Pb/204Pb values (15.47–15.57).

Fig. 11
figure 11

Lead isotope compositions of Py1 and Py2, galena, K-feldspar, syenite, and Archean rocks in the Dongping gold deposit. a 207Pb/204Pb vs. 206Pb/204Pb; b 208Pb/204Pb vs. 206Pb/204Pb. Data are from this study and Nie (1998) and Bao et al. (2016). The evolution curves are from Zartman and Haines (1988)

Discussion

Origin of hydrothermal fluids

The concentrations and ratios of trace elements (e.g., Co, Ni, and Se) in pyrite have been used to assess the origin of hydrothermal fluids (Huston et al. 1995; Belousov et al. 2016). For example, Co/Ni ratios are widely used to assess the origin of pyrite and to elucidate the genesis of hydrothermal ore deposits (Bajwah et al. 1987; Li et al. 2014a, b). Pyrite with Co/Ni ratios < 1 are generally considered to have a sedimentary origin, whereas pyrite with Co/Ni ratios > 1 (particularly between 1 and 5) are regarded as of hydrothermal origin (Loftus-Hills and Solomon 1967; Large et al. 2014). In some cases, hydrothermal pyrite may have Co/Ni ratios of < 1, in which case it is necessary to consider geological evidence such as the mineralogical characteristics and genetic type of the ore deposit (Bralia et al. 1979). Excluding abnormally high values caused by micro-inclusions, the Co/Ni ratios of Py1 and Py2 are 0.05‒47.47 (average = 4.37) and 0.02‒73.67 (average = 12.43), respectively (ESM 2 Table 2). The Co/Ni ratios of most pyrite are between 1 and 10 (Fig. 7i), indicating that pyrite precipitated from hydrothermal fluids. We speculate that hydrothermal fluids provided the ore-forming materials during both the early and late mineralization stages.

The Se contents of the four pyrite types vary considerably from those of Carlin-type to those of porphyry Cu-Au deposits (Keith et al. 2018). The Se contents in most pyrite range from a few to several dozens of ppm, whereas Py2a grains in sample DP23-3 possess the highest Se contents, ranging from 528 to 1657 ppm (ESM 2 Table 2). At the Bendigo orogenic gold deposit, the average Se contents of diagenetic and hydrothermal pyrite are 67 ppm and 32 ppm, respectively (Thomas et al. 2011). The Se-rich (mean = 319 ppm) hydrothermal pyrite of some Au deposits is thought to result from leaching of sedimentary rocks and transport by metamorphic fluids (Brill 1989). The high Se contents of the late-stage pyrite at Dongping indicate that sedimentary Se participated in the Au mineralization process.

In hydrothermal systems, rare earth elements (REEs) can be used to trace the source of hydrothermal fluids and the degree of water–rock interaction (Henderson 1984; Wang et al. 2012; Li et al. 2018b). The ΣREEs of pyrite from the Dongping gold deposit are low but increase progressively from Py1a to Py2b (Fig. 8a, b). The chondrite-normalized REE patterns of early- and late-stage pyrite are different from those of the Archean country rocks and intensely altered syenite adjacent to the auriferous quartz veins (Fig. 8a–c), which suggests that the pyrite REEs were derived from another source. Both early- and late-stage pyrites are relatively HREE-enriched, possibly due to discrimination against LREEs in the pyrite crystal lattice (Fan et al. 2000; Li et al. 2003; Wang et al. 2004), whereas the syenite and Archean country rocks are LREE-enriched. Moreover, the Ce and Eu anomalies of Py1a and Py1b are similar to those of magmatic and hydrothermal zircons in the Shuiquangou syenite (Fig. 8a, d), whereas those of Py2a are similar to those of hydrothermal zircons in the Shangshuiquan granite (Fig. 8b, d). These observations may indicate a close genetic relationship between Py1 and the Shuiquangou syenite, as well as between Py2 and the Shangshuiquan granite.

Sulfur isotope compositions are commonly useful tools for defining both the source and genesis of sulfides (Ohmoto and Rye 1979; Ulrich et al. 2011; Li et al. 2013; Zhang et al. 2014; Laflamme et al. 2018; Zhai et al. 2018). The δ34S values of Py1a and Py1b (− 4.3‰ and − 5.0‰, respectively) are ~ 2‰ higher than those of Py2a and Py2b (− 7.0‰ and − 6.4‰, respectively), suggesting that S sources changed during each mineralization stage. Negative δ34S values in sulfides can be interpreted as indicating a biogenic or sedimentary sulfur source, or the incorporation of oxidized magmatic fluids, or isotopic fractionation during the evolution of the mineralizing fluid (Ohmoto and Rye 1979; Phillips et al. 1986; Cameron and Hattori 1987; Oberthuer et al. 1996; Hodkiewicz et al. 2009). For both Py2a and Py2b, S isotope compositions are heterogeneous within single grains, as shown by systematically higher δ34S values in the core of a grain relative to its rim (ESM 2 Table 4). Song and Zhao (1996) measured the S isotopic compositions of co-occurring pyrite, sphalerite, and galena at Dongping, yielding similar negative δ34S values for all mineral phases. Total S isotopic compositions for the early mineralization stage were calculated using equations from Ohmoto and Rye (1979), yielding estimates ranging from − 1 to + 2‰. This result is similar to the S-isotopic compositions of pyrite in the Shuiquangou syenite (δ34S =  + 1.8 to + 3‰; Nie 1998; Bao et al. 2016). We speculate that the Shuiquangou syenite provided sulfur and perhaps metals during the early stage of mineralization at Dongping. The δ34S values of pyrite from regional gold deposits hosted in the Shuiquangou syenite are all negative values (− 17 to − 1; Fig. 10c; Song and Zhao 1996; Nie 1998; Li et al. 2000; Bao et al. 2016), which indicates that the Shuiquangou syenite may have provided sulfur to these deposits. Compared with Py1, the δ34S values of Py2 are more negative (down to − 7), indicating a different sulfur source may have involved in this stage, e.g., metamorphic rocks. Previous studies have shown that the Precambrian metasedimentary rocks of the North China Croton are enriched in isotopically light sulfur (32S) (Jo et al. 2021; Liu et al. 2022). Thus, we suggest that early-stage hydrothermal fluids in the Dongping gold deposit were derived mainly from the Shuiquangou syenite, and late-stage ore-forming materials mainly from the Shangshuiquan granite with a larger contribution from Archean metamorphic rocks.

Our Pb isotope compositions of pyrite (ESM 2 Table 5; Fig. 11) were combined with previous Pb isotopic data from the Dongping gold deposit (Nie 1998; Fan et al. 2001; Bao and Zhao 2006; Bao et al. 2016) to evaluate Pb sources. Py1 has lower isotopic ratios and therefore contains less radiogenic Pb than Py2. On a 207Pb/204Pb versus 206Pb/204Pb diagram (Fig. 11a), Py2 and Py1 generally plot between the mantle and orogenic reservoir compositions, although Py1 also plots between the lower continental crust and mantle reservoir compositions. We infer that early-stage ore fluids obtained Pb from both the lower crust and mantle, whereas late-stage fluids derived more radiogenic Pb mainly from the lower crust. On a 208Pb/204Pb vs. 206Pb/204Pb diagram (Fig. 11b), most samples plot between the mantle and lower crust lines, indicating multiple sources of Pb. Previous studies reported that sulfides, disseminated ores, and auriferous quartz veins plot along an array defined by the Pb isotope compositions of K-feldspar, syenite, and Archean metasedimentary rocks, which is indicative of mixing between two Pb sources (Nie 1998; Bao et al. 2000, 2016). Moreover, Py1 and the Shuiquangou syenite have similar Pb isotopic compositions (Fig. 11a, b), which suggests that the syenite provided Pb and perhaps other metals to hydrothermal fluids during early-stage mineralization. In contrast, Py2 has a greater radiogenic Pb content than Py1, indicating that the Shuiquangou syenite was not involved in late-stage mineralization, and that a radiogenic source of Pb was involved. We surmise that the Shangshuiquan granite was the major source of Pb during late-stage mineralization, and that the Pb contribution from Archean rocks became more significant. The Shuiquangou syenite was produced by fractional crystallization of mantle-derived melts (Li et al. 2018a), whereas the Shangshuiquan granite was derived from the lower continental crust by partial melting (Jiang et al. 2009), making the latter a likelier source of radiogenic Pb.

Fluid evolution during the two mineralization stages

Siderophile and chalcophile elements (e.g., Co, Ni, As, Se, and Te) are commonly accommodated in pyrite via substitution for Fe (Zhang et al. 2014). However, EPMA results (Fig. 6a, c, d) reveal that Fe shows no correlation with any of these elements. The low trace element contents of Py1a and Py2a may have been due to rapid pyrite growth rates, causing weak adsorption of cations from solution onto the growing crystal surfaces.

The REE composition of pyrite is controlled not only by its crystal texture but also by the REE characteristics of the hydrothermal fluid (Yuan et al. 2017). Differences in the ionic radii of REE3+ and Fe2+ make accommodation of REEs in Fe structural sites of the pyrite lattice difficult (Mao et al. 2009), and thus REEs tend to build up in fluid inclusions (Li et al. 2003; Fan et al. 2000; Wang et al. 2004; Shannon 1976; Mao et al. 2009). In the Dongping gold deposit, Py2a and Py2b have higher ΣREE contents than Py1a and Py1b. The high ΣREE contents of Py2a may have been caused by release of REEs from HREE-rich inclusions. In addition, water–rock interaction may have played a role in uptake and transport of REEs, particularly the HREEs, which preferentially enter into solution relative to the LREEs (Mao et al. 2009). The more variable and higher REE contents of Py2 relative to Py1 suggest intense fluid interaction of the pyrite with the Shangshuiquan granite and Archean metamorphic rocks. In addition, some Py1a grains show an obvious positive Ce anomaly due to formation in an oxidizing environment (Fig. 8b; Li et al. 2014a, b; Kong et al. 2016).

The more negative δ34S values of Stage-2 pyrite likely resulted from sulfur isotope fractionation in a high-fO2 hydrothermal fluid (Zhang et al. 2022). Fluid oxidation is evidenced by a general decreasing trend of δ34S values from the parent euhedral pyrite to porous pyrite (Fig. 10). Dissolution of the parent pyrite and precipitation of the later pyrite may have occurred at low pH in an Fe2+-rich fluid caused by partial oxidation of aqueous H2S and/or S2− in ore fluids (Wu et al. 2019).

Textural and chemical differences among the multiple generations of pyrite allow the migration of gold and other metals to be tracked throughout the formation process of gold deposits (Large et al. 2009). Py1a and Py1b are enriched in Co, Zn, and S and contain only minor Au and As (ESM 2 Table 2), which indicates that early-stage hydrothermal fluids were Au-poor and contributed relatively little to ore formation. In contrast, Py2a and Py2b are strongly enriched in Au and As and moderately enriched in Ag, Tl, Bi, and Pb (Fig. 8), which indicates that Au-As transport and deposition occurred during late-stage mineralization (Cook et al. 2009; Large et al. 2009, 2013). During the replacement of euhedral pyrite, depletion of gold and other trace elements in porous pyrites (Py1b and Py2b) relative to their precursors (Py1a and Py2a) indicates remobilization of these metals from the crystal lattice of the original pyrite (Fig. 5d, i; Wu et al. 2019). The variable but in general more radiogenic Pb isotope ratios in late, porous, and inclusion-rich pyrite (Py2b) can be attributed to modification of and precipitation from a highly evolved crustal fluid during the re-precipitation of native gold (Ma et al. 2022).

In summary, the trace element redistribution sequence during precipitation of Py1a, Py1b, Py2a, and Py2b plus or minus associated sulfides can be described as follows: (1) Py1a + fluids (Tl+, Au+, Ag+, Zn2+, Cu2+, Pb2+, Ni2+, Co2+, Bi2+) (aq) → Py1b + fluids (As2+, Se2+) (aq) and (2) Py2a + Ccp + Gn + Sp + fluids (Tl+, Au+, Ag+, Pb2+, Bi2+, Sb) (aq) → Py2b + fluids (As2+, Se2+, Zn2+, Cu2+, Ni2+, Co2+) (aq).

Gold occurrence and precipitation

In the Dongping gold deposit, pyrite is the main host mineral of gold. Gold occurs within pyrite as lattice-substituted gold and in micro- to nanograins, in the form of native gold, electrum, or Au–Ag tellurides (Fig. 5g–h). Our LA–ICP–MS results show a systematic increase in the Au and trace element content from Py1 to Py2 (Fig. 7). Gold may occur in submicroscopic inclusions of discrete Au-bearing phases (Large et al. 2007, 2009, 2013; Cook et al. 2009; Fougerouse et al. 2016). Py2a and Py2b contain significant amounts of Au and Ag, e.g., spot analyses of Py2b reveal concentrations up to 1000–2000 ppm (Fig. 5i). However, there is no correlation between the Au and As contents, as seen in previous studies (e.g., Morishita et al. 2018).

The four types of pyrite have similar As contents, and most data plot below the “gold solubility in pyrite” line (Fig. 7h; Reich et al. 2005), although some analyses plot above the line, indicating that Au exists as nano-inclusions. We propose that most gold was incorporated into the pyrite lattice as Au+, with a small fraction present as nanoparticles of native gold (Au0) within cavities, pores, and interstices in Py2b (Fig. 5h, i; Simon et al. 1999; Reich et al. 2005).

The substitution of S for As in pyrite reduces the symmetry around the iron atom and results in structural distortion and formation of defects that can facilitate the incorporation of gold into pyrite (Simon et al. 1999; Reich et al. 2005; Deditius et al. 2014; Li et al. 2019). The positive correlations between Au and Ag, Ti, Pb, and Bi indicate that gold was co-precipitated with these elements in pyrite, perhaps as inclusions (Fig. 7a–e).

In the Dongping gold deposit, Cook et al. (2009) observed a good correspondence between high gold values and two key textural criteria: areas of clustered inclusions, and microshearing and fracturing/brecciation within pyrite grains. In this study, we found that porous Py1b has a higher Au content relative to Py1a (Fig. 5d). High Au contents have also been obtained from spots within porous Py2b grains (Fig. 5i). The enrichment of Au in the late-stage pyrite was a result of leaching of earlier-formed Au from microfractures during pyrite recrystallization (Fig. 5c, d, f, i) (Large et al. 2009; Bi et al. 2011; Velasquez et al. 2014).

The Shuiquangou syenite, Shangshuiquan granite, and Archean metamorphic rocks are all thought to contain substantial amounts of ore-forming materials. However, ore-forming metals, e.g., Au, As, Ag, Cu, Pb, and Zn, were not rich in early intrusive rocks and magmatic fluids (those forming Py1) but were remarkably enriched in later hydrothermal fluids (those forming Py2). The high gold content in Py2b suggests a contribution from magmatic-hydrothermal fluids derived from the contemporary Shangshuiquan granite, which may have also been circulated in the Archean metamorphic rocks. According to the S and Pb isotope data, ore-forming metals were derived in part from the mantle during the early mineralization stage, but mainly from the lower continental crustal and Archean metamorphic rocks during the late mineralization stage.

A two-stage mineralization model: pyrite perspective

The texture, structure, and trace element compositions of pyrite can provide evidence for overprinting of mineralization events and their differing impacts on Au distribution (Baker et al. 2006; Cook et al. 2009). The LA–ICP–MS data indicate that only minor Au was deposited from hydrothermal fluids in Py1a, whereas Py2b has the highest Au contents. The variable Co and Ni contents of pyrite support an origin from a magma-related hydrothermal system derived from multiple sources. Mineralization may have been related to the proximity of the Devonian Shuiquangou syenite complex and the Cretaceous Shangshuiquan granite, an inference that is consistent with the results of earlier studies. For instance, Nie (1998), Bao et al. (2016), and Nie et al. (2004) suggested that pyrite with positive δ34S values was derived from the Shuiquangou syenite complex; and Cisse et al. (2017) and Miao et al. (2002) inferred that mineralization in the Dongping region was of magmatic hydrothermal origin, with a mixture of mafic magmas derived from the mantle and felsic magmas derived from partial melting of ancient Archean metamorphic rocks. The Shangshuiquan granite served as a potential source of ore-forming materials owing to its close relationship with the late-stage, milky colored, and gold-bearing quartz veins.

Based on our textural, mineralogical, multi-element, and isotopic studies of pyrite as well as the results of previous research (Nie 1998; Miao et al. 2002; Cook et al. 2009; Bao et al. 2016; Cisse et al. 2017; Li et al. 2018a), we propose that the Dongping gold deposit was the result of two episodes of mineralization (Fig. 12), with most of the gold emplaced during Stage 2.

Fig. 12
figure 12

Schematic diagrams illustrating the two stages of gold mineralization in the Dongping gold deposit. a Early-stage mineralization was associated mainly with intrusion of Devonian Shuiquangou syenite generated by Paleozoic subduction of the paleo-Asian Ocean Plate beneath the North China Craton; b late-stage mineralization was related to emplacement of the Cretaceous Shangshuiquan granite generated by Mesozoic subduction of the paleo-Pacific Plate beneath the North China Craton, with significant contributions from Archean metamorphic rocks

The first stage of mineralization, which occurred during the Devonian (~ 380 Ma), was related to intrusion of the Shuiquangou syenite generated by Paleozoic subduction of the paleo-Asian Ocean Plate beneath the North China Craton (Fig. 12a, Zhang et al. 2010). During post-collision extension, mantle-derived magma melted pre-existing lower continental crust, forming a mixed magma for the Shuiquangou syenite (Shao et al. 1999; She et al. 2006; Chen et al. 2008; Zhang et al. 2009). The gray-colored, auriferous quartz veins in the Dongping gold deposit formed after emplacement of the Shuiquangou syenite. The hydrothermal fluids derived from syenite intrusions had high F and K contents (Fan et al. 2001; Gao et al. 2015, 2017; Wang et al. 2019a) and resulted in strong potassic metasomatism (Figs. 4e, 5b, c). Gold derived from the syenite may have been transported as sulfide and/or telluride complexes in the hydrothermal fluids (Bao et al. 2016). The cooling caused by extensive silicification and potassic metasomatism contributed to the precipitation of gold.

The second stage of mineralization was related to emplacement of the Cretaceous (~ 140 Ma) Shangshuiquan granite, which was generated by subduction of the paleo-Pacific Plate beneath the North China Craton (Fig. 12b; Li et al. 2018a). The Shangshuiquan granite originated from partial melting of the ancient lower continental crust (Ma et al. 2012), and its parent magma may have been I-type and experienced extensive fractional crystallization (Jiang et al. 2009). Lithospheric thinning accompanied by asthenospheric upwelling and lower crustal degassing produced large amounts of magma that provided abundant fluids and heat for gold mineralization (Mao et al. 2008; Li et al. 2012a, b, c; Li et al. 2018a). Fluids exsolved from the Shangshuiquan granite leached S and Pb from Archean metamorphic rocks. Hydrothermal fluids migrating along faults and fissures in Archean basement rocks were the main source of late-stage gold deposited during the Cretaceous.

Applicability to other gold deposits

The northern margin of the North China Craton is well-known for gold deposits, with reserves of about 900 tons of gold. The gold district occurs as an E-W-trending belt (~ 1500 km long) in which metamorphosed Archean and Proterozoic strata were episodically uplifted during Paleozoic and Mesozoic deformation events. Most gold deposits are hosted by uplifted Precambrian metamorphic rocks, and ~ 30% of the deposits are hosted by Paleozoic and Mesozoic syenite and granite (Qiu et al. 1993; Nie 1997; Deng et al. 2014; Fu et al. 2020). The gold deposits in the northern NCC, such as Dongping, possess common features to other intrusion-related gold deposits worldwide: (1) spatial and temporal associations with igneous rocks; (2) high Au/Ag ratios and anomalous Te contents; and (3) CO2-rich ore fluids associated with the degassing of alkalic igneous rocks (Mutschler et al. 1985; Thompson et al. 1985; Richards and Kerrich 1993; Helt et al. 2014; Bao et al. 2016). Richards and Kerrich (1993) suggested that the ore-forming materials of these deposits were remobilized from early crystallized granitic bodies, whereas other studies inferred that the metals may have come directly from magmas (Werle et al. 1984; Anderson et al. 1987). With regard to the “Dongping-type” gold deposits on the northern margin of the NCC, these possible genetic associations deserve further study.

The multistage mineralization events of the gold deposits on the northern margin of the NCC are consistent with episodic tectonic activation and associated magmatism. Four major magmatic episodes have been recognized on the northern margin of the NCC: Devonian (390–353 Ma), Triassic (236–218 Ma), Jurassic (199–161 Ma), and Cretaceous (143–125 Ma) (Deng and Wang 2016). The Devonian magmatic episode, represented by the Shuiquangou syenite and the Dahuabei granite in the study region, was related to arc accretion. The Triassic magmatic episode, represented by some monzogranite and igneous dikes, was a post-collisional event following closure of the Central Paleo-Asian Ocean. The Jurassic magmatic episode, represented by granite and quartz diorite, may represent the post-orogenic stage. The Cretaceous event, represented by the Shangshuiquan granite at Dongping, has been related to lithospheric thinning, coinciding with widespread Cretaceous magmatic activity in eastern China (Miao et al. 2002; Deng et al. 2004, 2009). Recent geochronological studies of these deposits support a multi-stage mineralization process, ranging from the Devonian to the Cretaceous (Zhang et al. 2013; Deng et al. 2014; Gao et al. 2014; Jia et al. 2018; Fu et al. 2020).

In summary, we propose that gold deposits on the northern margin of the NCC underwent multistage mineralization events, and that Cretaceous hydrothermal fluids had the largest influence on formation of large-scale gold deposits. Therefore, regional gold exploration may need to pay more attention to late-stage superposition of gold in mineralized deposits. The Cretaceous magmatic episode may have directly enhanced the ore-forming process and contributed large amounts of ore to these gold deposits, which should be regarded as the most favorable prospecting direction in the region.

Conclusions

1. Two stages of gold mineralization in the Dongping gold deposit have been differentiated based on pyrite texture and geochemistry. Early-stage hydrothermal fluids were mainly derived from the Devonian Shuiquangou syenite with a larger mantle contribution, whereas late-stage hydrothermal fluids were mainly derived from the Cretaceous Shangshuiquan granite with significant material input from Archean Sanggan Group metamorphic rocks.

2. The trace element redistribution process during precipitation of Py1a, Py1b, Py2a, and Py2b plus or minus associated sulfides can be described as follows: (1) Py1a + fluids (Tl+, Au+, Ag+, Zn2+, Cu2+, Pb2+, Ni2+, Co2+, Bi2+) (aq) → Py1b + fluids (As2+, Se2+) (aq) and (2) Py2a + Ccp + Gn + Sp + fluids (Tl+, Au+, Ag+, Pb2+, Bi2+, Sb) (aq) → Py2b + fluids (As2+, Se2+, Zn2+, Cu2+, Ni2+, Co2+) (aq).

3. Gold occurs as macroscopic native gold and microscopic gold in solid solution and within the crystal lattice of pyrite. High Au contents occur mainly in the rims, fractures, and pores of Stage-2 pyrite grains. The enrichment of Au in the late-stage pyrite was a result of leaching of earlier-formed microscopic Au from microfractures during pyrite recrystallization.