Paleo‐Pacific Plate Subduction and Basement Mobilization Triggered Large‐Scale Formation of Mesozoic Gold Deposits in the Northern Margin of the North China Craton

The northern margin of the North China Craton (NNCC) hosts Mesozoic gold deposits of significant economic importance. The metal sources of these gold systems have long been debated with mainly magmatic‐hydrothermal versus metamorphic fluid models. Mercury (Hg) isotopes, which undergo unique mass‐independent fractionation, can provide important insights into the source of metals in gold deposits due to the close association between Hg and Au in such systems. Here, we investigated the Hg isotopic composition of six gold deposits of the Middle Jurassic to Early Cretaceous age and potential source rocks in the NNCC. Variable ∆199Hg values were observed in bulk ore and pyrite samples (−0.28‰ to 0.34‰, n = 51) and coeval granites (−0.21‰ to 0.13‰; n = 25). The negative ∆199Hg values of bulk ore and pyrite samples mostly agree with that observed for metamorphic basement rocks (−0.37‰ to 0.11‰; n = 32). The positive ∆199Hg values of bulk ore and pyrite samples agree with those reported in marine sediments (0‰–0.3‰). So the data suggest binary mixing of ocean‐recycled Hg originated from the subducted oceanic slab, and terrestrial‐recycled Hg from the Precambrian basement in Mesozoic gold deposits. The contribution of Hg (and Au, by analogy) from both reservoirs varies depending on active continental arc versus intracontinental setting and is ultimately controlled by translithospheric heat flow driven by paleo‐Pacific plate subduction.

mass-dependent fractionation (MDF, expressed as δ 202 Hg) and unique mass-independent fractionation (MIF, expressed as Δ 199 Hg). A variety of biological, chemical, and physical processes can trigger Hg-MDF (Blum et al., 2014); however, Hg-MIF mainly occurs during photochemical reactions with limited interference from other processes (Bergquist & Blum, 2007). Magmatic, metamorphic, and hydrothermal processes do not induce Hg-MIF, allowing the use of Hg-MIF signals for tracing metal sources in geological reservoirs Deng et al., 2021;Deng, Hong, et al., 2022;Tian et al., 2022;Yin et al., 2022). The primitive mantle has nearzero ∆ 199 Hg values (0.00‰ ± 0.10‰, Moynier et al., 2021), whereas photochemical processes on Earth's surface result in negative and positive ∆ 199 Hg values in terrestrial (soil and vegetation) and oceanic (seawater and marine sediments) reservoirs, respectively (Blum et al., 2014). The recent observation of positive ∆ 199 Hg values in global basalts and arc-related epithermal gold systems at convergent plate margins highlight marine Hg recycling into the mantle and crust via oceanic plate subduction Moynier et al., 2021;Yin et al., 2022). The observation of negative ∆ 199 Hg values in intracontinental epithermal gold deposits in south China and NE China  suggests terrestrial-derived Hg was sourced from the Precambrian metamorphic basement.
Here, we present Hg isotope data for gold ore samples collected from six Mesozoic hydrothermal gold deposits in the NNCC. Combined with analyses of potential source rocks for these gold deposits (e.g., Precambrian basement rocks, Mesozoic granites). This study aims to gain an understanding of the genesis of these Mesozoic gold systems from the perspective of Hg isotope geochemistry.
Mesozoic gold deposits in the NNCC were formed during two major metallogenic stages (Table 1): Middle Jurassic (e.g., Songjianghe, Jiapigou, Yu'erya and Jinchangyu deposits) and Early Cretaceous (e.g., Paishanlou and Jinchanggouliang deposits). These gold deposits are located in brittle-ductile shear zones and fractures ( Figure 2). Gold occurs as disseminated or quartz veins within either the Precambrian metamorphic basement (e.g., Songjianghe, Jiapigou, Paishanlou, Jinchanggouliang, and Jinchangyu; L. Liu et al., 2020;Song et al., 2016;X. H. Zhang et al., 2005;X. T. Zhang et al., 2019) or Mesozoic granites (e.g., Yu'erya; Kong et al., 2015). The sulfides in the ore include pyrite, chalcopyrite, galena, and sphalerite. Pyrite is the major gold-bearing mineral. Fluid inclusion studies indicate that the ore-forming fluids belong to the H 2 O-NaCl-CO 2 -CH 4 system, with low to intermediate salinity (1.1-15.3 wt.% NaCl eq.) and temperatures in the range of ∼200-∼400°C (Table 1 and references therein). A total of 51 ore samples were collected from the above-mentioned six gold deposits in three major gold districts named Jiapigou, Western Liaoning, and Jidong ( Figure 1b). The ore samples are mainly quartz vein type and disseminated altered rock type.
Potential source rocks, including Mesozoic granites (n = 25) and Precambrian basement rocks (n = 30), were also collected in the vicinity of the studied deposits (Figure 2, Tables 3 and 4). These rocks are fresh without influences of hydrothermal alteration. The collected granites can be divided into two ages: Early-Middle Jurassic (Huangniling and Wudaoliuhe granite from Songjianghe and Jiapigou, Yu'erya granite from Yu'erya, Qinshankou granite from Jinchangyu) and Early Cretaceous (Yilvshan granite from Paishanlou, Duimiangou granite from Jinchanggouliang). The age of the granite samples is broadly coeval with that of the corresponding gold deposit    (2011), Kong et al. (2015), and B. W. Guo et al. (2017); The information of Jinchangyu deposit come from Song et al. (2016) and H. Q. Wang et al. (2020). Table 1 Detailed Information of the Gold Deposits Studied (Tables 1 and 3). These granites intruded into the Precambrian basement ( Figure 2). The collected basement rocks (granulite, gneiss, amphibolite, mica schist, andesite, and andesitic tuff) cover a time span from Archean to Mesoproterozoic (Table 4).

Methods
Twenty-five of the collected ore samples (n = 51) were used for bulk analyses, and the 26 remaining samples were used to select pyrite separates by handpicking under a binocular microscope. We failed to acquire enough chalcopyrite separates for analyses due to their small grain sizes and small amounts in the samples. Prior to chemical analyses, the ore samples and pyrite separates were powdered to 200 mesh size and homogenized in an agate mortar, in the same way as that used for granites (n = 25) and basement rocks (n = 30).
The samples and pyrite separates were then measured for total Hg concentration and Hg isotopic composition at the State Key Laboratory of Environmental Geochemistry, Institute of Geochemistry, Chinese Academy of Sciences, Guiyang, China, using a Lumex RA 915+ Hg analyzer and a Neptune Plus multi-collector inductively coupled plasma mass spectrometer, respectively (Yin et al., 2016). The collected samples were washed with deionized water, air-dried, crushed, ground to 200 mesh powder, and homogenized, prior to total Hg (THg) concentration and Hg isotopic composition analysis.
THg concentrations were measured by a Lumex RA 915+ Hg analyzer (Lumex Ltd. Russia), which yielded THg recoveries of 90%-110% for standard reference material GSS-4 (soil) and uncertainty of <10% for sample replicates. The Lumex RA 915+ Hg analyzer has an analytical uncertainty of ±10% (SD) during THg concentration analyses. A double-stage tube furnace coupled with 40% reverse aqua regia (HNO 3 /HCl = 3/1, v/v) trapping solutions was used for Hg preconcentration (Zerkle et al., 2020). The trapping solutions were diluted to 1.0 ng/mL Hg in 10% HCl (v/v) prior to Hg isotope analysis by Neptune Plus multi-collector inductively coupled plasma mass spectrometry (Yin et al., 2016). Mercury isotope composition was reported following the convention (Bergquist & Blum, 2007). Hg-MDF is expressed in δ 202 Hg notation in units of ‰ referenced to the NIST-3133 Hg standard (analyzed before and after each sample analysis): Hg-MIF is reported in Δ notation, which describes the difference between the measured δ xxx Hg and the theoretically predicted δ xxx Hg value, in units of ‰, with xxx = 199, 200 or 201: Δ xxx Hg ≈ xxx Hg − 202 Hg × β is equal to 0.2520 for 199 Hg, 0.5024 for 200 Hg, and 0.7520 for 201 Hg (Bergquist & Blum, 2007).
GSS-4 soil standard reference material, which contains silicate substances and has Hg contents in the same order of magnitude as the rock samples, was prepared in the same way and measured at the same time as the samples. To ensure accurate measurement of the ore samples, we also conducted replicated analyses of NIST-3177 secondary Hg standard (diluted to 1.0 ng/mL Hg in 10% HCl (v/v)) at the same analytical session, given that the NIST-3177 standard is a digest of cinnabar collected from the Almaden Hg mine, Spain (Bergquist & Blum, 2007).

Results
Analytical results of this study are given in Tables 2-4. Despite published data of three basement rocks (Table 4), the results of gold ores and rocks are novel and not published elsewhere. In brief, Precambrian metamorphic basement and granite in the NCC show variable Hg concentrations of 0.46-23.6 and 0.50-12.1 ppb (Tables 3 and 4), with mean values of 4.18 ± 10.6 (n = 32, 2SD) and 1.83 ± 4.58 ppb (n = 25, 2SD), respectively. Gold-bearing

Table 4
Information, Hg Concentration, and Hg Isotopic Composition of Basement Rock Samples bulk ore and pyrite samples show elevated Hg concentrations ranging from 0.89 to 221 and 1.52 to 1,230 ppb, with mean values of 23.1 ± 91.0 (n = 25, 2SD) and 87.6 ± 466 ppb (n = 26, 2SD), respectively (Table 2). No independent Hg mineral was observed in any of the samples. Like that observed in other gold deposits (Rytuba, 2003), it is believed that Hg is present mainly as Hg(II) in substitution to Fe(II) in sulfides of the samples investigated, due to their similar ionic radius. Compared with the results of arc-related and intracontinental gold deposits , the Hg concentration of gold deposits in NNCC is 2-3 orders of magnitude lower (Figures 5a-5c). This could be due to the fact that NNCC gold deposits were formed at relatively higher temperatures (Table 1), which favor the loss of Hg from the hydrothermal fluids (Smith et al., 2005).

Precambrian Basement as a Major Hg Source for Mesozoic Granites
Large variations of δ 202 Hg can be observed for the studied granite samples and basement rocks (Figure 3a). The variation of δ 202 Hg is resulted from  MDF during a large number of processes and cannot provide enough information on Hg sources. A large variation of δ 202 Hg (∼4‰) due to Hg-MDF during hydrothermal processes (e.g., hydrothermal boiling and redox reaction) has been previously reported in epithermal Hg deposits in the USA (Smith et al., 2005(Smith et al., , 2008. However, hydrothermal processes are unlikely to produce Hg-MIF (Sherman et al., 2009;Smith et al., 2005Smith et al., , 2008. It is also reported that metamorphic processes could trigger Hg-MDF but do not produce obvious Hg-MIF Deng, Hong, et al., 2022;Moynier et al., 2021). Hg-MIF is commonly found in surface reservoirs and is mainly produced via photochemical processes (Blum et al., 2014). The lack of Hg-MIF during geological processes allows the use of ∆ 199 Hg to trace surface Hg recycling into Earth's interior reservoirs (e.g., mantle and crust).
Hg-MIF mainly occurs during Hg photochemical reactions on Earth's surface with no interference from other processes . The ∆ 199 Hg/∆ 201 Hg ratio of 1.06 ± 0.13 (2SD) agrees well with that observed for Hg(II) photoreduction (∆ 199 Figure 3a). As Mesozoic granites studied here are intruded into the Precambrian basement (Figure 2), inheriting of Hg from partial melting of the Precambrian basement by granites may explain the congruency of ∆ 199 Hg values of the studied basement rocks and granites. This is supported by previous studies based on multiple geochemical indicators that (e.g., REE, trace element, Hf and Pb isotopes) suggest that Mesozoic granites in NNCC were derived from Precambrian basement melting in the lower crust (Fu et al., 2012;Han et al., 2022;Kang et al., 1998;Lu et al., 2019;Song, 2011;S. K. Sun et al., 2012).

Dual Hg Sources for the NNCC Gold Deposits
The formation of the Mesozoic gold deposits of the NNCC was interpreted as related to either magmatic-hydrothermal (N'dri et al., 2021;Yang et al., 2021) or metamorphic-fluid models (Goldfarb & Pitcairn, 2022;Goldfarb et al., 2001Goldfarb et al., , 2019Tomkins, 2010), which would suggest that Hg was sourced from the Mesozoic granites or the Precambrian metamorphic basement, respectively. As shown in Figure 3b, bulk ore and pyrite samples show ∆ 199 Hg values of −0.28‰ to 0.34‰, which cover a larger range than that of the Mesozoic granites (−0.21‰ to 0.13‰) and the Precambrian basement rocks (−0.37‰ to 0.11‰), suggesting that the gold deposits studied cannot be explained exclusively by any of the above-mentioned models. Instead, we suggest that the sources of Hg (and Au, by analogy) are variable among different gold deposits, as discussed below.
The two Middle Jurassic (∼160 Ma) gold deposits (Jiapigou and Songjianghe) in the Jiapigou gold district mainly show positive ∆ 199 Hg values with a mean of 0.14‰ ± 0.26‰ (n = 18, 2SD) ( Figure 5a). The Jiapigou gold district was located in an active continental arc in the Jurassic and formed during the subduction of the paleo-Pacific plate at ∼160 Ma. The lithospheric mantle is greatly modified by fluid released from the subducted oceanic plate during this stage (F. Guo et al., 2015). The mostly positive ∆ 199 Hg values of the Jiapigou gold district (−0.1‰ to 0.34‰) are consistent with those observed in ocean sediments and metasomatized mantle materials (∆ 199 Hg: −0.08‰ to 0.37‰; X. Y. Wang et al., 2021) rather than Precambrian basement rocks in the area (∆ 199 Hg: −0.30 to 0.11; This study). In addition, as shown in Figure 5a, the ∆ 199 Hg values of the Jiapigou gold district almost overlap with those of arc-related epithermal gold systems, suggesting that Hg was derived from a marine source via oceanic plate subduction . We hereby suggest that the gold deposits of Jiapigou district should have received predominantly ocean-recycled Hg from the subducted paleo-Pacific slab.
Contrary to the Jiapigou gold district, the two ∼160 Ma gold deposits (Jinchangyu and Yu'erya) in the Jidong district mainly show negative ∆ 199 Hg values with a mean value of −0.07‰ ± 0.24‰ (n = 19, 2SD) ( Figure 5b). The predominantly negative ∆ 199 Hg values in the Jinchangyu gold district are consistent with that observed for the Precambrian basement rocks in the area (∆ 199 Hg: −0.21‰ to 0.04‰). Previous studies on intracontinental Au deposits in SW China and NE China also observed negative ∆ 199 Hg values (∆ 199 Hg: −0.24‰ to −0.02‰; Figure 5b), which were attributed to the addition of mobilized Hg from the Precambrian basement . The ∆ 199 Hg values of Jidong gold deposits mostly overlap with those of intracontinental Au deposits . Therefore, we suggest the Precambrian metamorphic basement is the most probable Hg source for the Jidong gold deposits. This is supported by the fact that the Jidong gold district is located on the inner continent behind the active continental arc (Yang et al., 2021 and references therein), with the lack of significant influence from the subducted paleo-Pacific plate.
Both positive and negative ∆ 199 Hg values with a mean of 0.01‰ ± 0.32‰ (n = 14, 2SD) were observed in the ∼130 Ma gold deposits (Paishanlou and Jinchanggouliang) in the Western Liaoning gold district (Figure 5c). The Precambrian basement rocks and Mesozoic granites in this area show negative ∆ 199 Hg values. The negative ∆ 199 Hg values in some of these gold deposits may be explained by the contribution of Hg from the Precambrian basement, similar to the Jidong gold district. Meanwhile, the Western Liaoning gold district was formed during the destruction of the NCC destruction at ∼130 Ma . Prior to 130 Ma, the lithospheric mantle beneath the NCC had been extensively metasomatized by fluids released by the subducted oceanic slab . Therefore, the positive ∆ 199 Hg values observed in the Paishanlou and Jinchanggouliang deposits suggest that some Hg is also sourced from the metasomatized lithosphere mantle. The occurrence of both positive and negative ∆ 199 Hg values in the Western Liaoning gold district indicates binary mixing of ocean-recycled Hg (from the subducted paleo-Pacific slab) and terrestrial-recycled Hg (from the Precambrian basement) with opposing ∆ 199 Hg signals. We infer that fluids both from the subducted oceanic slab and continental basement contributed Hg to the ∼130 Ma gold deposits in the Western Liaoning gold district.

Genetic Model for the Mesozoic Gold Deposits in the NNCC
Results of this study allow us to develop a genetic model for the Mesozoic gold deposits in the NNCC (Figure 6). During the Middle-Late Mesozoic, two major stages of magmatism occurred in the NNCC at ∼160 and ∼130 Ma,  respectively (Yang et al., 2003(Yang et al., , 2008. Subduction of the paleo-Pacific plate is believed to be responsible for the magmatism occurring at ∼160 Ma (Yang et al., 2003(Yang et al., , 2008; Figure 6a), whereas the rollback of the paleo-Pacific plate associated with deep and steep subduction is proposed as the cause of NCC destruction that occurred at ∼130 Ma (Zhu and Sun., 2021; Figure 6b). During the Jurassic, the Jiapigou area was located in the active continental arc, and the involvement of subducted oceanic slab could contribute abundant marine-derived volatiles and fluids (e.g., H 2 O, CO 2 , and Hg) into the lithospheric mantle, causing positive ∆ 199 Hg values of the lithospheric mantle and inducing arc magmatism as well. During this process, the ascension of mantle-derived magmas and hydrothermal solutions into the upper crust could have released sufficient amounts of fluids containing recycled marine Hg, favoring the formation of gold deposits in the Jiapigou gold district with positive ∆ 199 Hg values (Figure 6c). The negative ∆ 199 Hg values in the Jidong gold district, which was located in a relatively intercontinental environment, could be formed by fluid released from the remobilized Precambrian basement due to heat flow and crustal compression thickening driven by paleo-Pacific plate subduction (Figure 6d).
Since the Early Cretaceous, the rollback of the paleo-Pacific plate triggered the destruction of the NCC, as represented by lithospheric thinning and upwelling of the asthenosphere . Destruction of the NCC not only drove the melting of the metasomatized mantle to release magmatic-hydrothermal fluids characterized by positive ∆ 199 Hg values but also could benefit remobilization of the Precambrian basement to release ore-forming fluids containing Hg with negative ∆ 199 Hg values (Figure 6e). These two kinds of fluids favored the formation of the ∼130 Ma gold deposits in the Western Liaoning gold district, which could explain the existence of both positive and negative ∆ 199 Hg in these deposits.
The distinct Hg-MIF signals between the Jiapigou gold district (mostly positive ∆ 199 Hg values) and Jidong gold district (mostly negative ∆ 199 Hg values) may be related to their distinct tectonic environment. Combined with the ∆ 199 Hg characteristics and tectonic evolution of the three regions, whether positive MIF signals occur in gold deposits depends on the amount of subducted oceanic plate-derived fluids that affected the lithosphere. In analogy to Hg, the source of gold lies both in the metasomatized lithospheric mantle (positive ∆ 199 Hg) and the continental basement (negative ∆ 199 Hg), with variable contributions from both reservoirs depending on the regional geotectonic situation or, ultimately, the translithospheric heat flow driven by subducting oceanic plate.

Conclusion
In summary, our results on Hg isotopes show that the source of ore-forming materials of the Mesozoic gold deposits in the NNCC can be both Precambrian metamorphic basement and lithospheric mantle metasomatized by subduction oceanic plates. Mesozoic gold deposits show variable Δ 199 Hg values. Gold deposits with negative Δ 199 Hg receive Hg mainly from the Precambrian metamorphic basement, while those with positive Δ 199 Hg receive Hg from the lithospheric mantle metasomatized by the subducted paleo-Pacific slab. This study shows a novel application of Hg isotopes as a source tracer for hydrothermal gold deposits and highlights their potential for metallogenetic tracing in other hydrothermal systems.

Data Availability Statement
The data set of this research (Yin, 2022) can be found in a public domain repository.