Copper skarn mineralization related to reduced monzodioritic magma at the Huanren deposit, northeastern China Open Access

Copper skarns are the world's most abundant skarn type and represent one of the most important sources of Cu worldwide (Meinert et al., 2005). They are common in oceanic and continental subduction and intracontinental extensional settings (Meinert et al., 2005; Li et al., 2008; Xie et al., 2015, 2019; Pirajno and Zhou, 2015; Chang et al., 2019). Such deposits result from the interaction between magmatic-hydrothermal fluids and carbonate rocks within kilometer-scale convective hydrothermal systems that develop around dioritic to granitic intrusions (Meinert et al., 2005; Chang et al., 2019). It is widely accepted that Cu skarns are typically associated with oxidized fluids and intrusions (fayalite-magnetite-quartz [FMQ] buffer = 1.0–1.5; Candela, 1992; Meinert et al., 2005; Richards, 2015). Proof of this high magmatic oxygen fugacity (fO₂) comes from the abundant primary magnetite, hematite, and anhydrite in equilibrium with Cu-bearing minerals and the association of skarn Cu deposits with elevated fO₂ intrusions (e.g., Chelle-Michou et al., 2015; Shen et al., 2015; Cao et al., 2018; Chang et al., 2019). These oxidized intrusions are generally considered to be the source of the metals and volatiles in the ore-forming fluids (Meinert et al., 2005).

In silicate melts, the partitioning of Cu between sulfide liquid (or monosulfide solid solution) and melt depends on the speciation and solubility of S, which in turn are sensitive to melt redox conditions (Jugo et al., 2010; Kiseeva and Wood, 2013; Rezeau and Jagoutz, 2020). For example, experiments have demonstrated a sharp transition from sulfide (S²⁻) to sulfate (SO₄²⁻) within a narrow range between FMQ +1 and +2, resulting in a 10-fold increase in total S solubility in the melt (e.g., Jugo et al., 2005, 2010). This constitutes favorable conditions for effective Cu transport through the crust and for later Cu skarn mineralization (Meinert et al., 2005; Richards, 2015). Furthermore, the oxidized magmatic fluids that emanate from such oxidized magmas are able to effectively extract Cu from the magma (e.g., Richards, 2015; Rezeau and Jagoutz, 2020) and/or remobilize any Cu-sulfides that may have precipitated previously in shallow magma reservoirs (e.g., Halter et al., 2002; Nadeau et al., 2010). Thus, fO₂ has traditionally been regarded as a fundamental factor affecting Cu skarn formation (e.g., Ishihara, 1981; Meinert et al., 2005; Li et al., 2008; Xie et al., 2015; Chelle-Michou et al., 2015). However, the potential for Cu skarn association with reduced magmas remains overlooked.

In this study, we documented the geological features, including magmatism, alteration, and mineralization, of a Cu skarn deposit at Huanren, northeastern China, which contains 0.41 Mt of economic Cu. We were particularly interested in obtaining quantitative fO₂ estimates of the causative intrusion at Huanren given the importance of this parameter in Cu skarn formation. Our results highlight the presence of a Cu skarn deposit associated with reduced monzodioritic magma in an intracontinental extensional setting, which is at odds with the classic genetic model for Cu skarn formation associated with oxidized magmas in arc settings, thereby offering a new perspective for the formation of skarn deposits in non-arc settings associated with reduced magmas. Furthermore, this recognition of reduced magma in a Cu skarn system is crucial for a more accurate understanding of the formational processes associated with Cu skarns and should prompt exploration in non-arc settings deemed favorable for the genesis of this new class of Cu skarn.

The Huanren deposit is located in the Liaodong Peninsula, North China Craton (Fig. 1A). Tectonically, the Liaodong Peninsula can be subdivided into three tectonic units: the Archean Liaonan block in the south, the central Paleoproterozoic Liaoji orogenic belt, and the Archean Liaobei block in the north (Fig. 1B). The basement consists of the Liaonan and Liaobei blocks, which are composed of Archean tonalite-trondhjemite-granodiorite (TTG) suites, whereas the Liaoji orogenic belt is composed of Paleoproterozoic Liaohe Group sedimentary and volcanic rocks, both of which were metamorphosed during a ca. 1.85 Ga orogenic event (Zhao et al., 2001). From the late Paleoproterozoic to the early Paleozoic, the Liaodong Peninsula received frequent deposition of shallow-marine carbonate platform sediments (Yang et al., 1986). Since the late Paleozoic, the Liaodong Peninsula has been affected by multiple orogenic events, especially during the late Permian to late Mesozoic. These events included (1) thrusting and crustal thickening around the northern margins of the North China Craton during the closure of the Paleo-Asian Ocean in the Late Permian to Early Jurassic (Xiao et al., 2003) and (2) the destruction of the North China Craton during the Middle Jurassic to Early Cretaceous (Menzies and Xu, 1998; Zhu et al., 2011). These events coincided with a major change in the subduction direction of the paleo-Pacific plate (Sun et al., 2007), and with the formation of intracontinental rift basins, metamorphic core complexes, and bimodal volcanic rocks in the North China Craton (Ren et al., 2002; Zhu et al., 2011).

Accompanying the multiple orogenic events, the Liaodong Peninsula recorded episodic magmatism expressed by voluminous mafic to felsic intrusions (Fig. 1B). Three pulses of magmatism have been identified in the Liaodong Peninsula (Wu et al., 2005a): (1) minor Late Triassic (233–212 Ma) syenite, diorite, and monzogranite; (2) minor Jurassic (180–156 Ma) tonalite, granodiorite, and monzogranite; and (3) voluminous Early Cretaceous (131–117 Ma) mafic to felsic intrusions. The Late Triassic intrusions were emplaced during crustal extension associated with the final closure of the Paleo-Asian Ocean (Yang et al., 2007a). The Early Jurassic intrusions, which typically occur as batholiths, were formed by mixing of crustal- and enriched mantle–derived magmas (Wu et al., 2005b). In contrast, the Early Cretaceous intrusions were formed in an extensional setting related to rollback of the paleo-Pacific plate (Liu et al., 2020; Wu et al., 2021) and have associated important Cu, Au, and Pb-Zn mineralization.

Several medium- to large-sized Pb-Zn, Au, and Cu-polymetallic ore fields/deposits exist in the Liaodong Peninsula, including the intermediate-sulfidation Qingchengzi Pb-Zn ore field, the low-sulfidation Baiyun, Jingchanggou, and Wulong Au deposits, and the porphyry-skarn Wanbaoyuan and Huanren Cu-polymetallic deposits (Fig. 1B). Orebodies in the Qingchengzi Pb-Zn ore field are hosted in the Paleoproterozoic Liaohe Group sedimentary and volcanic rocks and have no clear association with intrusive units (Yu et al., 2009). Recent xenotime and rutile U-Pb dates indicate that Au mineralization in the Baiyun deposit occurred ca. 230 Ma (Feng et al., 2022), whereas molybdenite Re-Os and monazite U-Pb ages suggest that Au mineralization at Wulong occurred ca. 127 Ma (Yu et al., 2020). By contrast, the porphyry-skarn Cu-polymetallic deposits in the Liaodong Peninsula are either hosted within, or are clearly associated with, the Early Cretaceous dioritic to felsic intrusions, which were generated during a relatively narrow time interval of 130–128 Ma (Ouyang et al., 2013; Zhang et al., 2018).

The Huanren Cu skarn deposit is situated in the Liaoji orogenic belt, in the Liaodong Peninsula, ~40 km east of the China–North Korea boundary (Fig. 1B). It is one of China's well-known deposits and contains 0.41 Mt Cu, 0.50 Mt Zn, and 0.14 Mt Pb, accompanied by recoverable concentrations of Fe, Mo, and Ag (P.J. Zhang, personal commun., 1953; Y.B. He and Z.L. Jiang, personal commun., 1975; Y.W. Zhang, personal commun., 2016). The mining history of the Huanren deposit dates back to the end of the Qing Dynasty, when Pb-Zn-Ag minerals were mined from weathered carbonate-replacement orebodies. Systematic exploration was initiated in 1956 and ended in 1974 and led to the discovery of concealed skarn Cu-Zn orebodies at the contact zones between the dioritic intrusion and the carbonate rocks in the Songlan and Xiangyang ore blocks (Figs. 2 and 3). The deposits yielded a resource of 0.15 Mt Cu at an average grade of 0.72 wt%, 0.50 Mt Zn at an average grade of 0.45 wt%, and 0.14 Mt Pb. More recently (2000 to present), exploration to alleviate the resource crisis has led to the discovery of skarn-type Cu-(Zn-Pb-Mo) and Fe-(Cu-Zn) mineralization, and porphyry- to skarn-type Mo mineralization at depth, which yields an additional resource of 0.26 Mt Cu at 0.74 wt%. The Huanren deposit thus provides an ideal example of lessons learned, which may help to locate and assess mineral resources in other skarn mineral systems. However, to date, the geology of the Huanren deposit remains poorly constrained, particularly at depth.

In this study, we reexamined relationships among intrusion types, hydrothermal alteration, and mineralization style at Huanren through a comprehensive field study. This field study formed the basis for interpreting the whole-rock geochemistry, U-Pb and Re-Os geochronology, and mineralogical estimation of the magmatic oxidation state. Our observations, combined with exploration data, permit a new detailed reconstruction of the Huanren deposit geology, which indicates that the deposit genesis is more complex than previously considered (e.g., Song, 2010; Zhang et al., 2018).

The oldest exposed strata in the Huanren deposit are Cambrian shale and carbonate rocks, which are unconformably overlain by Silurian sandy shale and Jurassic andesitic to rhyolitic volcanic rocks. The Cambrian carbonate rocks are well bedded, composed of (1) limestone and dolomite with lenses of evaporite composed of gypsum, anhydrite, and halite, and (2) purple to reddish calcareous shale, calcareous siltstone, sandstone, chert, and nodules, which represent a peritidal carbonate succession (Yang et al., 1986; Zhao et al., 2001; Wu et al., 2014). They occur as relics surrounded by Jurassic andesitic to rhyolitic volcanic rocks, which in turn are intruded by Cretaceous mafic to intermediate intrusions associated with Cu skarn mineralization (Fig. 2). Only a single set of faults trends NE-NNE in the Huanren deposit, which displaced the Precambrian granite and Paleozoic strata and controlled the emplacement of Cretaceous mafic to intermediate intrusions.

Based on crosscutting relationships and mineral modal abundances (International Union of Geological Sciences recommendation), four types of intrusions are identified in the Huanren deposit, which include, chronologically, porphyritic granite, monzodiorite, syenite, and dioritic porphyry. The porphyritic granite occurs within the central and eastern zones of the mining area (Fig. 2). Luo et al. (2019) performed sensitive high-resolution ion microprobe (SHRIMP) zircon U-Pb dating of the porphyritic granite, yielding a weighted ²⁰⁶Pb/²³⁸U mean age of 2490 ± 8.0 Ma (1σ, mean square of weighted deviates [MSWD] = 0.38). As the porphyritic granite bears no genetic relationship with mineralization, it is not described further. The monzodiorite, syenite, and dioritic porphyry are by far the most voluminous types in the Huanren deposit, with an outcrop area of over 18 km² (Fig. 2). Cambrian shale and carbonate rocks and Jurassic andesitic to rhyolitic volcanic rocks surround the intrusions.

In plan view, the monzodiorite is exposed as irregularly shaped, elongated stocks >10 km² intruding into Cambrian carbonate rocks and Jurassic volcanic rocks (Fig. 2). Our field observations showed that the monzodiorite has many irregular contacts and apophyses into the Cambrian carbonate rocks and even hosts some (dolomitic) limestone xenoliths, especially near the contact. Two textural varieties were identified: one coarse-grained variety (Fig. 4D), commonly observed in the inner part of the monzodiorite intrusion, and the other medium- to fine-grained variety (Fig. 4E), usually found within the outer margins of the monzodiorite intrusion. The contact between them is commonly gradational, and no crosscutting relationships were observed in this study. The coarse-grained monzodiorite is composed of 70–75 vol% plagioclase, 10–26 vol% amphibole, ~5 vol% biotite, and 3–5 vol% alkaline feldspar, with minor amounts of quartz (<2 vol%) and clinopyroxene (<2 vol%) (Figs. 4F–4H); accessory minerals include apatite, zircon, and ilmenite, with rare magnetite (Figs. 4H–4I). The medium- to fine-grained monzodiorite contains major and accessory minerals similar to the coarse- to medium-grained monzodiorite, but with higher proportions of alkaline feldspar (5 vol%) and quartz (<5 vol%), accompanied by a decrease in the abundance of plagioclase (~70 vol%) and amphibole (~20 vol%). At Huanren, the monzodiorite is the most significant intrusion in terms of volume of quartz-molybdenite or molybdenite veining (Figs. 5E–5F and 6F–6H), and it is the principal host for economic Mo mineralization.

Swarms of syenite dikes and dioritic porphyry dikes are well developed in the Huanren deposit. They are more than 4 km in length and range from 2 m to 200 m in width (Fig. 2). Field observations showed that the syenite dikes crosscut the monzodiorite stock and, in turn, are crosscut by the dioritic porphyry dikes (Figs. 4A and 4B). The syenite and dioritic porphyry dikes are fresh, bear no relationship with mineralization, and are interpreted as post-ore intrusions.

Three main styles of hydrothermal alteration were identified within the Huanren deposit in this study: potassic, epidote-chlorite-sericite, and skarn alteration. Potassic alteration was only observed in the monzodiorite. It is represented by the replacement of plagioclase by secondary K-feldspar and/or the replacement of amphibole by secondary biotite in and around high-temperature barren quartz veinlets or veins (Figs. 5A–5D), and/or by the presence of hydrothermal biotite veinlets (Figs. 5D, 6D, and 6E) and quartz–K-feldspar veinlets (Figs. 6A–6C). Hence, the intensity of potassic alteration mainly relies on the abundance of quartz, quartz–K-feldspar, and biotite veinlets or veins in the monzodiorite. Our field observations showed that the potassic alteration is more intense in areas proximal to the contacts between the monzodiorite and Cambrian carbonate rocks, in which quartz veinlets or veins are relatively well developed in the monzodiorite, becoming less common with distance from the contacts. In extreme circumstances, potassic alteration almost completely overprints the monzodiorite (Fig. 5C). No (molybdenite and chalcopyrite) mineralization was observed in this alteration stage.

Epidote-chlorite-sericite alteration was only observed in the monzodiorite. It is characterized by the replacement of amphibole and biotite by epidote and chlorite or the replacement of plagioclase by fine-grained sericite, accompanied by a small amount of disseminated pyrite (Figs. 4D, 4F, and 5C). It commonly overprints earlier potassic alteration (Figs. 5C and 5D). In the Huanren deposit, epidote-chlorite-sericite alteration displays a spatial association with Mo and Cu mineralization in the form of molybdenite ± quartz ± pyrite veinlets (Figs. 5E and 6F–6H), quartz-epidote-molybdenite ± pyrite veins (Fig. 5F), and disseminated chalcopyrite (Fig. 6I).

Skarn alteration in the Huanren deposit is extensive and represents the most important mineralization-related alteration type. Generally, this type of alteration is pervasive in areas peripheral to the contacts between the monzodiorite and the Cambrian carbonate rocks in the form of exoskarn (Fig. 4C). The exoskarn hosts the bulk of the Huanren orebodies, but the ratios of garnet and pyroxene, and of economic minerals (e.g., chalcopyrite, sphalerite, galena, molybdenite, and magnetite), in the exoskarn assemblages vary with carbonate rock protolith and proximity to the contacts between the monzodiorite and the carbonate rocks. Two exoskarn assemblages are recognized: Mg and Ca exoskarns.

The Mg exoskarns are characterized by prograde minerals, including pyroxene and minor garnet, and retrograde minerals, including phlogopite, magnetite, actinolite, and calcite, with minor amounts of chalcopyrite, sphalerite, pyrite, and chlorite. They are invariably associated with Fe-(Cu-Zn) mineralization in the Huanren system and tend to be massive (Figs. 7B and 7F). Retrograde alteration, consisting dominantly of magnetite, calcite, and/or actinolite, occurs along the grain boundaries of pyroxene crystals (Figs. 8A–8B). Locally, chalcopyrite, sphalerite, arsenopyrite, pyrite, pyrrhotite, and rare opaque phases are also present, and these generally postdate magnetite mineralization (Figs. 7E–7F).

The Ca skarns consist of prograde minerals, including garnet, pyroxene, and wollastonite, and retrograde minerals, including chalcopyrite, sphalerite, pyrrhotite, galena, arsenopyrite, actinolite, and epidote, with minor amounts of quartz, calcite, molybdenite, magnetite, and pyrite. They host the bulk of the Huanren Cu-(Zn-Pb-Mo) and Pb-Zn-(Ag) orebodies. The Ca skarns are most widely developed proximal to the contact between the monzodiorite unit and the carbonate rocks in the form of massive skarns (Fig. 4C), laterally extending for variable distances (up to 500 m) into the Cambrian carbonate rocks as skarn veins (Fig. 2). In the Ca skarns, garnet is commonly altered to or rimmed by epidote, calcite, chalcopyrite, pyrrhotite, sphalerite, galena, arsenopyrite, and/or magnetite; pyroxene is altered to dark-green actinolite (Fig. 7A), and, in extreme cases, individual garnet and pyroxene grains are completely replaced by actinolite.

As with many skarn systems, garnet/pyroxene abundance ratios in the Ca skarns in the Huanren deposit are spatially zoned with distance to the monzodiorite. The Ca skarns in zones proximal to the monzodiorite are generally massive, coarse-grained, and garnet-rich skarns (Fig. 4C). Retrograde alteration (mostly actinolite and epidote) and sulfide minerals (mostly chalcopyrite, pyrrhotite, and sphalerite, with minor molybdenite) are abundant in proximal Ca skarns (Figs. 7A, 7C–7D, and 8C–8F). Outward or upward from the monzodiorite-Cambrian carbonate rock contact, the garnet content in the Ca skarns decreases, whereas the pyroxene abundance generally increases (Fig. 9).

In the Huanren deposit, chalcopyrite, sphalerite, magnetite, galena, and molybdenite are typically present in economic proportions, although pyrrhotite, arsenopyrite, bornite, pyrite, native gold, and Ag-bearing minerals are also present. Interestingly, in plan view, skarn mineralization mainly occurs at the northeastern margins of the monzodiorite; Pb, Zn, and Ag all show enrichment outward from the center to the contacts between the monzodiorite and the Cambrian carbonate rocks (Fig. 2). Cross-section data show that the bulk of skarn mineralization at Huanren developed around the monzodiorite in the form of skarn (Fig. 3) down to depths >1000 m below the current surface. In addition, skarn Cu-(Zn-Pb-Mo) and Fe-(Cu-Zn) mineralization is relatively proximal, and skarn Pb-Zn-(Ag) mineralization is more distal to the intrusion-carbonate rock contact (Figs. 3 and 9). In this study, two main styles of mineralization were observed: porphyry and skarn (subtypes are detailed in the following text).

Porphyry-style Mo mineralization consists of veinlet- and vein-hosted Mo and represents the main types of Mo mineralization at Huanren. Veinlet-hosted Mo mineralization typically occurs as molybdenite ± pyrite veinlets (Fig. 5E) or quartz-molybdenite veinlets (Figs. 6F–6H) with epidote-sericite-chlorite halos in the monzodiorite. The vein-hosted Mo mineralization is characterized by quartz-epidote-molybdenite ± pyrite veins 1–4 cm in width with epidote-sericite-chlorite envelopes (Fig. 5F). In general, the veinlet-hosted Mo mineralization is commonly developed at depth in the monzodiorite. At shallow levels, Mo mineralization in the monzodiorite is dominantly the vein-hosted type. Porphyry-style Cu mineralization occurs as disseminated chalcopyrite in the monzodiorite with epidote-sericite-chlorite alteration (Fig. 6I). However, the average Cu grade in the monzodiorite is generally lower than the economic value (≥0.5 wt%). At Huanren, the porphyry-style Mo and Cu mineralization is surrounded by zones of skarn Mo, Fe-(Cu-Zn), Cu-(Zn-Pb-Mo), and Pb-Zn-(Ag) mineralization.

Skarn-style mineralization in the Huanren deposit is dominantly developed in the Cambrian carbonate rocks in the form of exoskarn and represents the most important mineralization type at Huanren. It can be further subdivided into four subtypes: (1) Mo, (2) Fe-(Cu-Zn), (3) Cu-(Zn-Pb-Mo), and (4) Pb-Zn-(Ag), based on the components of sulfide minerals. The exoskarn Mo mineralization is commonly developed at the contacts between the monzodiorite and the carbonates rocks and is characterized by Mo mineralization overprinting massive Ca exoskarn (Fig. 7A). Comparatively, this type of Mo mineralization is economically less important than the porphyry-style Mo mineralization. The exoskarn Fe-(Cu-Zn) mineralization is commonly developed adjacent to the monzodiorite as massive exoskarn and is characterized by magnetite replacing massive Mg and less commonly Ca skarns (Fig. 7B), which in turn are replaced by chalcopyrite, sphalerite, and calcite (Figs. 7E–7F), with variable amounts of pyrrhotite, arsenopyrite, and pyrite. The exoskarn Cu-(Zn-Pb-Mo) mineralization occurs in the proximal to intermediate Ca skarn (Fig. 9) in the form of stratiform skarn or pockets of massive Cu-rich skarn (Figs. 7C and 7D). In addition to chalcopyrite, the exoskarn Cu-(Zn-Pb-Mo) mineralization has variable sphalerite, galena, molybdenite, bornite, and pyrite contents (Figs. 8C–8F). Gangue minerals include epidote, pyroxene, garnet, calcite, quartz, chlorite, sericite, and actinolite. Chalcopyrite is commonly intergrown with bornite, galena, sphalerite, and molybdenite (Figs. 8E and 8F). It also occurs as chalcopyrite-pyrite veins crosscutting sphalerite (Fig. 8C) or as interstices in sphalerite (Fig. 8D). At Huanren, this subtype of exoskarn mineralization represents the most important type of Cu mineralization.

The exoskarn Zn-Pb-(Ag) mineralization commonly occurs in the intermediate and distal pyroxene-garnet or pyroxene skarns (Fig. 9), where the carbonate rocks are only partially replaced by skarn minerals. It consists mainly of pyroxene, sphalerite, and galena with minor amounts of garnet, chalcopyrite, and pyrite, and it is commonly narrow and elongated. Moreover, the abundance of sphalerite generally correlates positively with the abundance of pyroxene.

To better constrain the timing of the causative intrusion at Huanren, the least-altered coarse-grained and medium- to fine-grained monzodiorite samples from the underground mine were selected for laser ablation–inductively coupled plasma–mass spectrometer (LA-ICP-MS) zircon U-Pb dating (Table S1¹). Zircon grains were handpicked from crushed samples and then mounted in epoxy resin. The sample mounts were polished to expose grain interiors. The grains were photographed under transmitted and reflected light and then examined using scanning electron microscope–cathodoluminescence (SEM-CL) imaging to reveal internal textures of the zircons. Imaging revealed no apparent inherited zircon cores or magmatic rim textures (Fig. S1). Only clear regions in individual zircon grains that lacked any opacity and inclusions were selected for analysis (Fig. S1). LA-ICP-MS zircon U-Pb dating was carried out using an Agilent Technologies 7700x quadrupole ICP-MS equipped with a 193 nm ArF excimer laser at Nanjing FocuMS Technology Co. Ltd., China. The laser was focused on the sample with a fluence of 5.0 J/cm² and a spot of 40 μm diameter at a repetition rate of 5 Hz for 50 s. Zircon 91500 (primary standard), GJ-600, and Ple-337 were used as external calibration standards to correct for instrumental mass bias and elemental fractionation (Wiedenbeck et al., 1995) and to verify data accuracy. Measured results were within error of recommended values (Table S2). The Pb, U, and Th contents of zircon were externally calibrated against SRM NIST610 with ²⁹Si (15.32 wt%) as internal standard (Hu et al., 2011). Raw-data reduction was performed offline using the ICPMSDataCal software (Liu et al., 2010). The uncertainties on the external standard and decay constant were propagated into the uncertainties of unknowns during raw-data reduction following the methods described by Liu et al. (2010). Uncertainties on individual analyses are reported at the 2σ level.

Zircon trace elements were measured simultaneously during the zircon U-Pb dating using an Agilent Technologies 7700x quadrupole ICP-MS equipped with a 193 nm ArF excimer laser at Nanjing FocuMS Technology Co. Ltd. The laser was focused on the sample with a fluence of 5.0 J/cm² and a spot of 40 μm diameter at a repetition rate of 5 Hz for 50 s. Raw-data reduction and concentration calculations were performed offline using the ICPMSDataCal software (Liu et al., 2010). SRM NIST610 glass was used as the calibration standard, and trace-element concentrations were calculated using ²⁹Si (15.32 wt%) as the internal standard. The zircons 91500, GJ-600, and Ple-337 were used to verify data accuracy. The accuracy of trace-element concentrations was better than 5% for most elements based on repeat analyses of reference materials (Table S3). Uncertainties on individual analyses are quoted at the 1σ level.

Hafnium isotope analyses of zircon grains from the least-altered coarse-grained and medium- to fine-grained monzodiorite samples were carried out using a Nu Plasma II multicollector (MC) ICP-MS (Nu Instruments) equipped with a 193 nm ArF excimer laser at Nanjing FocuMS Technology Co. Ltd. Analyses were acquired proximal to spots ablated for zircon U-Pb dating. The analysis was conducted on a circular 50 μm laser spot with a fluence of 3.5 J/cm², a repetition rate of 9 Hz, and a duration of 40 s. Mass fractionation corrections for Hf and Yb isotopic ratios were based on ¹⁷⁶Lu/¹⁷⁵Lu = 0.02656 (Blichert-Toft et al., 1997) and ¹⁷⁶Yb/¹⁷³Yb = 0.7876 (McCulloch et al., 1977), respectively. To monitor the accuracy of this correction, every five unknowns were bracketed by analyses of reference zircons, including Penglai, GJ-1, Plešovice, 91500, and Mud Tank. The measured results were within the error of recommended values (Table S4). A decay constant for ¹⁷⁶Lu of 1.867 × 10⁻¹¹ yr⁻¹ was adopted (Söderlund et al., 2004). Initial ¹⁷⁶Hf/¹⁷⁷Hf ratios, denoted as ε_Hf(t), were calculated relative to the chondritic reservoir with a ¹⁷⁶Hf/¹⁷⁷Hf ratio of 0.282772 and ¹⁷⁶Lu/¹⁷⁷Hf of 0.0332 (Blichert-Toft and Albarède, 1997).

Two molybdenite-bearing massive skarn samples, one molybdenite veinlet sample, and two quartz-epidote-molybdenite-pyrite vein samples from the underground mine (Table S1), encompassing the majority of Mo mineralization types in the Huanren deposit, were sampled for isotope dilution–inductively coupled plasma–mass spectrometer (ID-ICP-MS) molybdenite Re-Os dating. The molybdenite Re-Os dating was conducted at the National Research Center of Geoanalysis, Chinese Academy of Geological Sciences. The detailed analytical procedures are described in Li et al. (2012) and are summarized here. The powdered (≤200 mesh) molybdenite sample was weighed and transferred to a Carius tube for dissolution. Enriched ¹⁹⁰Os and ¹⁸⁵Re were used as spikes to correct for common Os and mass fractionation. Osmium was separated using in situ distillation equipment, and Re was then isolated using an anion exchange column. The determination of the Re-Os isotope ratio was performed on a TJA X-series ICP-MS. The ¹⁸⁷Re decay constant of 1.6689 ± 0.0031 × 10⁻¹¹ yr⁻¹ (Selby et al., 2007) was used to calculate the molybdenite model ages. During the calculation, the uncertainty of the ¹⁸⁷Re decay constant was propagated into the uncertainty of the calculated molybdenite Re-Os model ages systematically following the methods described in Li et al. (2012). Uncertainties on each individual age determination are reported at the 2σ level. The GBW04435 standard was used to verify data accuracy, and the measured average was within error of the recommended value (Table S5).

Whole-rock major- and trace-element analyses were carried out at Nanjing FocuMS Technology Co. Ltd., except for whole-rock FeO contents, which were analyzed at Guangzhou Tuoyan Testing Technology Co. Ltd., China. Relatively fresh coarse-grained and medium- to fine-grained monzodiorite samples were first broken into centimeter-sized crush; fresh pieces were selected, washed with deionized water, dried, and then ground to less than 200 mesh for geochemical analysis. For major-element analysis, sample powders were fluxed with Li₂B₄O₇ (1:8) to make homogeneous glass disks at 1250 °C using a V8C automatic fusion machine. Major elements were then analyzed using Spectro Xepos X-ray fluorescence spectrometry. BHVO-2 and AGV-2 standards were run as unknowns to evaluate accuracy and precision. Analytical errors were better than 1% for all elements. Results for secondary standards are listed in Table S6. The FeO content of the samples was determined by conventional wet chemical titration methods. The analytical procedures have been described in detail by Gao et al. (1995).

For trace-element analysis, sample powders were first dissolved using distilled HF + HNO₃ in screw-top Teflon beakers for 4 d at 100 °C. Trace elements were analyzed using an Agilent solution ICP-MS. BHVO-2 and AGV-2 standards were run as unknowns to evaluate data quality. Trace-element results for the standards agreed within 5% of recommended values, except for Mo, which yielded errors of 12%–20%. Results for standards are included in Table S6.

Major-element compositions of magmatic Fe-Ti–oxide pairs and apatite were analyzed on a JEOL JXA-8230 electron microprobe (EPMA) with five wavelength-dispersive spectrometers at Wuhan Microbeam Analysis Technology Co. Ltd., China, calibrated against a range of mineral, oxide, and metal standards. Coexisting Fe-Ti–oxide pairs from the coarse-grained monzodiorite samples (Fig. S2) were measured with an accelerating voltage of 15 kV, a beam current of 20 nA, and a 1 µm beam diameter. Peak counting times were 10 s for Fe, Si, and Ti, and 20 s for Ca, Al, Mg, Cr, V, and Mn. The background counting time was half of the peak counting time on the high- and low-energy background positions. The following standards were used: pyrope garnet (Al), diopside (Ca and Mg), rhodonite (Mn), olivine (Si), rutile (Ti), hematite (Fe), chromium (Cr), and vanadium (V). Data were corrected online using a ZAF (atomic number, absorption, fluorescence) correction procedure. All cations were analyzed as metals. To obtain accurate ZAF values, the oxygen content was calculated by difference and then incorporated into the ZAF correction procedure. The final oxide results (oxide mass percentage; Table S7) were recalculated based on the measured cations and added oxygen.

Prior to magmatic apatite EPMA analyses, apatite grains were handpicked from crushed, least-altered coarse-grained and medium- to fine-grained monzodiorite samples and then mounted in epoxy resin. The sample mount was polished to expose grain interiors. The grains were photographed under transmitted and reflected light and then examined using optical microscopic cathodoluminescence (OP-CL) imaging and EPMA mapping to identify any igneous (e.g., zoning) or hydrothermal features (e.g., alteration) of the apatites. Only clear, subhedral to euhedral magmatic apatite crystals or regions in individual apatite grains that lacked any opacity and inclusions were selected for analysis (Fig. S3). Imaging revealed no chemical zoning features and no unequivocal evidence for hydrothermal alteration of the selected grains (Table S8). An accelerating voltage of 15 kV, a beam current of 5 nA, and a 20 µm beam diameter with a peak counting time of 10 s for Na, P, Cl, S, Ca, and F were used for analysis (e.g., Goldoff et al., 2012; Stock et al., 2015). The background counting time was half of the peak counting time on the high- and low-energy background positions. The following standards were used: jadeite (Na), apatite (P and Ca), barite (S), sodium chloride (Cl), and barium fluoride (F). An NHNH 104021 fluorapatite standard was analyzed five times as an unknown at the beginning of the analytical session to determine analytical accuracy and precision. Results agreed within 4% of recommended values (Table S8).

In situ laser sulfur isotope analyses of chalcopyrite samples collected from the Cu-(Zn-Pb-Mo) orebody (Table S1) were performed using a Nu Plasma II MC-ICP-MS equipped with a Resonetics-S155 excimer ArF laser-ablation system at the Ministry of Natural Resources (MNR) Key Laboratory of Metallogeny and Mineral Assessment, Institute of Mineral Resources, Chinese Academy of Geological Sciences, Beijing. A circular 50 μm laser spot with a fluence of 5.0 J/cm², a repetition rate of 6 Hz, and a duration of 40 s were used for analysis. Full details of the analytical procedures are provided by Chen et al. (2017). Reproducibility of the analytical results and mass spectrometer calibration were monitored through replicate measurements of standard HTS4-6 (Li et al., 2020) and an in-house chalcopyrite standard (CuFeS2-2). The measured average was within error of the recommended value. All sulfur isotope compositions are reported in Table S9 using standard delta notation relative to Vienna Canyon Diablo Troilite (VCDT).

Zircon U-Pb dating results for the coarse-grained and medium- to fine-grained monzodiorite samples are listed in Table S2 and graphically present in Figure 10. Zircon grains from the medium- to fine-grained monzodiorite contained variable Th and U concentrations ranging 89–1447 ppm and 96–1281 ppm, respectively, corresponding to Th/U values of 0.6–1.3. The 24 zircon grains analyzed here yielded ²⁰⁶Pb/²³⁸U ages between 129.2 Ma and 124.2 Ma and a weighted mean ²⁰⁶Pb/²³⁸U age of 125.4 ± 0.8 Ma (2σ; MSWD = 0.27; Fig. 10). Zircon grains from the coarse-grained monzodiorite had 99–383 ppm Th and 112–366 ppm U, with Th/U ratios of 0.7–1.2. Individual zircon grains (n = 29) yielded ²⁰⁶Pb/²³⁸U ages ranging from 128.3 Ma to 124.5 Ma and a weighted mean ²⁰⁶Pb/²³⁸U age of 125.4 ± 0.9 Ma (2σ; MSWD = 0.10; Fig. 10), similar to that of the medium- to fine-grained monzodiorite. In summary, the 53 zircon grains analyzed were concordant, yielding a weighted mean ²⁰⁶Pb/²³⁸U age of 125.4 ± 0.6 Ma (2σ; MSWD = 0.16) for the monzodiorite (Table S2).

The ID-ICP-MS results of Re-Os molybdenite dating are presented in Table S5. Using the decay constant of Selby et al. (2007), the five dated samples yielded Re-Os model ages between 126.0 ± 1.8 Ma and 124.4 ± 1.7 Ma (2σ). Pooled together, they yielded a weighted mean date of 125.3 ± 0.8 Ma (2σ; Fig. 10). A regression through our data on a ¹⁸⁷Re-¹⁸⁷Os isochron plot yielded an initial ¹⁸⁷Os of 0.24 ± 0.42 ppb (Table S5), providing evidence for the absence of common Os in the molybdenite samples and supporting the hypothesis that all five samples were related to the same hydrothermal event.

Major- and trace-element compositions for the coarse-grained and medium- to fine-grained monzodiorite samples are listed in Table S6. The monzodiorite samples contained 52.5–55.3 wt% SiO₂ with Mg# values of 52–66 [Mg# = molar Mg/(Mg + Fe²⁺)]. On a total alkali versus silica plot, samples plot in the field of basaltic trachybasalt (Fig. 11A). On a SiO₂ versus K₂O plot, they show a high-K calc-alkaline to shoshonite affinity (Fig. 11B). In the primitive mantle–normalized spider diagram, they are moderately enriched in the large ion lithophile elements (LILEs) Rb, Ba, K, Pb, and Sr, and they are depleted in the high field strength elements (HFSEs) Nb, Ta, and Ti (Fig. 12A). These features are distinct from normal mid-oceanic-ridge basalt (N-MORB) and oceanic-island basalt (OIB) (Sun and McDonough, 1989) but similar to arc-related igneous rocks (e.g., Thirlwall et al., 1994). In terms of rare earth elements (REEs), the monzodiorite samples show listric-shaped patterns (La/Yb_N = 9–15), moderate fractionation between light REEs (LREEs) and middle REEs (MREEs; La/Sm_N = 3.2–4.5), slight fractionation between MREEs and heavy REEs (HREEs; Dy/Yb_N = 1.1–1.4), and slight to negligible negative Eu anomalies (Eu/Eu* = 0.7–1.0) (Fig. 12B).

The zircon Hf isotopic data for the coarse-grained and medium- to fine-grained monzodiorite samples are reported in Table S4. The zircon grains from the coarse-grained and medium- to fine-grained monzodiorite samples had similar initial ¹⁷⁶Hf/¹⁷⁷Hf ratios ranging from 0.282329 to 0.282401, with corresponding initial ε_Hf(t) values restricted to −13.4 to −10.9 (−12.2 ± 0.6 on average, 1σ, n = 52). These values are comparable to those of coeval enriched lithospheric mantle–derived mafic intrusions from the North China Craton (−11.7 ± 2.6 on average) but are distinct from those of coeval asthenospheric mantle– and ancient lower crust–derived magmas from the North China Craton (0.5 ± 1.8 and −22.2 ± 2.3 on average, respectively; Fig. 13; Wu et al., 2005b; Yang et al., 2007a, 2007b; Zhang et al., 2010; Ma et al., 2014a, 2014b).

Within the monzodiorite samples, ilmenite is a common accessory mineral and occurs as inclusions within amphibole, biotite, and plagioclase phenocrysts (Fig. 4I), whereas magnetite is rare and is commonly in contact with ilmenite or occurs as inclusions in ilmenite (Fig. 4I; Fig. S2). Compositions of magnetite and ilmenite pairs from the coarse-grained monzodiorite sample are given in Table S7. Ilmenites are nearly pure FeTiO₃ with low amounts of MgO (<0.12 wt%) and MnO (<2.85 wt%). Magnetites are nearly pure Fe₃O₄ with less than 5.83 wt% TiO₂.

The major-element compositions of the apatite grains from the coarse-grained and medium- to fine-grained monzodiorite samples are reported in Table S8 and Figure 14. Apatite grains are euhedral to subhedral, occurring as inclusions within biotite and amphibole or intergrown with biotite and amphibole (Fig. 4H). Under OP-CL light, apatite grains from the medium- to fine-grained monzodiorite displayed blue luminescence, whereas apatite in the coarse-grained monzodiorite sample displayed brown-yellow, green, blue, and gray luminescence (Fig. S3). EPMA maps show that the analyzed apatite grains are compositionally homogeneous (Table S8), indicating our measurements accurately capture the distribution of major-element contents (especially S, Cl, and F) in apatite. The average SO₃ contents of the apatite grains from the coarse-grained and medium- to fine-grained monzodiorite samples were similarly low (<0.05 wt%; Fig. 14B), while the average Cl content of the apatite grains from the coarse-grained monzodiorite sample was 0.83 ± 0.27 wt% (1σ, n = 12), i.e., significantly higher than Cl content from the medium- to fine-grained monzodiorite sample (0.08 ± 0.04 wt%, 1σ, n = 26). Moreover, the apatite grains from the coarse-grained monzodiorite samples generally displayed lower F (2.23 ± 0.53 wt%, 1σ, n = 12) concentrations than those from the medium- to fine-grained monzodiorite samples (3.37 ± 0.20 wt%, 1σ, n = 26). In the Cl-F-OH ternary diagram, all apatite grains fall along an F-OH exchange line (Fig. 14A). In addition, most apatite grains from the medium- to fine-grained monzodiorite samples and limited numbers of apatite grains from the coarse-grained monzodiorite samples define an F-rich group (fluorapatite proportion >80%; Fig. 14A).

In this study, magmatic fO₂ values of the coarse-grained and medium- to fine-grained monzodiorite samples were qualitatively estimated using the whole-rock Fe₂O₃/FeO ratios of Ishihara (1981) and were further calculated using the magmatic Fe-Ti–oxide and zircon trace-element oxybarometers of Ghiorso and Evans (2008) and Loucks et al. (2020), respectively. Details and results are presented in Tables S3 and S7 and Figures 15–16.

Relative to oxidized intrusions, which generally have whole-rock Fe₂O₃/FeO ratios above 0.5 (Ishihara, 1981), the coarse-grained and medium- to fine-grained monzodiorite samples at Huanren were characterized by similarly low Fe₂O₃/FeO ratios (ranging from 0.27 to 0.52, 0.39 ± 0.08 on average; Table S6). On a TFe₂O₃ versus log₁₀Fe₂O₃/FeO plot, the monzodiorite samples plot in the moderately oxidized to moderately reduced fields (Fig. 15A). On a plot of oxidation state of intrusions (expressed as Fe₂O₃/FeO) versus degree of compositional evolution (expressed as Rb/Sr), contoured for dominant metal associations in related mineralization, most of the monzodiorite samples are restricted to the reduced field (Fig. 15B).

Using the zircon trace-element oxybarometer of Loucks et al. (2020), the calculated fO₂ values of the coarse-grained monzodiorite ranged from FMQ −1.41 to FMQ −0.73 with an average value of FMQ −1.14 ± 0.14 (1σ, n = 29), which is in good agreement with those of the medium- to fine-grained monzodiorite (ranging from FMQ −1.51 to FMQ −0.53, FMQ −1.04 ± 0.22 on average, 1σ, n = 24; Fig. 16). Using the Fe-Ti–oxide oxybarometer of Ghiorso and Evans (2008), the calculated fO₂ values for the coarse-grained monzodiorite ranged from FMQ +0.25 to FMQ +3.14, with an average value of FMQ +1.01 ± 0.60 (1σ, n = 22). This value is significantly higher than those derived from the zircon trace-element oxybarometer.

In situ laser sulfur isotope analysis results from chalcopyrite are listed in Table S9 and illustrated in Figure 17. The analyzed chalcopyrites gave a relatively uniform δ³⁴S composition, ranging between 2.78‰ and 5.65‰ (4.05‰ ± 0.78‰, 1σ, n = 15), consistent with isotopic data produced by bulk analysis of chalcopyrite separates, which cluster around 4.6‰–5.8‰ (Zhang et al., 2018). A striking feature of the sulfur isotopic data at Huanren is that the chalcopyrite δ³⁴S values (4.34‰ ± 0.88‰ on average, 1σ, n = 20) are significantly higher than those from most of the porphyry-skarn Cu ± Mo ± Au deposits (Fig. 17).

To date, the ages of magmatism and mineralization in the Huanren deposit have been poorly documented. One low-precision whole-rock K-Ar age of 116 Ma from an uncharacterized monzodiorite sample was the primary constraint on the age of magmatism at the Huanren deposit (Song, 2010). Later, Zhang et al. (2018) performed LA-ICP-MS zircon U-Pb dating of the monzodiorite intrusion from the Songlan and Xiangyang ore blocks at Huanren, yielding a mean weighted ²⁰⁶Pb/²³⁸U age of 128.6 ± 3.0 Ma (2σ). In this study, we dated two textural varieties of the monzodiorite, one coarse-grained and the other medium- to fine-grained (Figs. 4D and 4E), from the Songlan and Xiangyang ore blocks, respectively, which show spatial and temporal relationships with both porphyry Cu and Mo and skarn Cu-Mo-Zn-Fe-Pb-Ag mineralization. Our LA-ICP-MS zircon U-Pb ²⁰⁶Pb/²³⁸U age data show that the intrusive ages of the coarse-grained and medium- to fine-grained monzodiorite samples are comparable, with a weighted mean age of 125.4 ± 0.9 Ma (2σ) and 125.4 ± 0.8 Ma (2σ), respectively (Fig. 10). Our new zircon U-Pb ages are consistent with previous dates, but they provide more systematic and precise constraints on the timing of crystallization of the monzodiorite.

Our field observations indicate that molybdenite mineralization in the Huanren deposit mainly occurs as molybdenite-bearing veinlets and veins hosted in the monzodiorite intrusion (Figs. 5E–5F and 6F–6H) or as disseminations in proximal massive Ca skarns (Fig. 8F). In this study, molybdenite separates selected from these Mo mineralization styles were successfully dated. The five dated samples yielded comparable molybdenite Re-Os model ages (126.0 ± 1.8 Ma to 124.4 ± 1.7 Ma, 2σ) and a weighted mean age of 125.3 ± 0.8 Ma (2σ; Fig. 10), consistent with the zircon U-Pb age of the monzodiorite (within uncertainty), providing the first age constraints of mineralization at Huanren. These new geochronology data indicate that porphyry Cu and Mo and skarn Cu-(Zn-Pb-Mo), Fe-(Cu-Zn), and Pb-Zn-(Ag) mineralization at Huanren may have been coeval with emplacement of the monzodiorite magma (i.e., porphyry- and skarn-type mineralization at Huanren was caused by the emplacement of the monzodiorite). This interpretation is supported by field observations: (1) the monzodiorite hosts high-temperature biotite and quartz–K-feldspar veinlets (Figs. 5A–5D and 6A–6E) and porphyry-style Cu and Mo mineralization (Figs. 6F–6I); (2) the skarns have decreasing garnet/pyroxene abundances with increasing distance from the monzodiorite (Fig. 9); and (3) Cu, Mo, and Fe mineralization increases toward the monzodiorite (Fig. 9). All these observation indicate that the ore-forming fluids and metals at Huanren were sourced from the monzodiorite magma.

The monzodiorite samples from the Huanren deposit are characterized by low zircon ε_Hf(t) values [ε_Hf(t) = −13.4 to −10.9], which are similar to those of coeval enriched lithospheric mantle–derived mafic intrusions in the North China Craton (Fig. 13). It should be noted that, on the plot of intrusion age versus ε_Hf(t), data for the monzodiorite samples also generally follow the magma mixing model between the asthenospheric mantle (represented by coeval asthenospheric mantle–derived mafic rocks) and ancient lower crust (represented by the coeval ancient lower crust–derived granitoids) of the North China Craton (Fig. 13). Thus, there are two possible models to explain the isotopic signatures of the monzodiorite samples: (1) assimilation of ancient crustal materials or mixing with ancient crust–derived melt during the emplacement of asthenospheric mantle–derived magma and (2) fractional crystallization of enriched lithospheric mantle–derived magma without significant crustal assimilation. Our data show that the Huanren monzodiorite does not contain inherited ancient zircons and has homogeneous zircon Hf isotope compositions regardless of the domain analyzed (cores or rims; Fig. S1; Fig. 13). Moreover, the monzodiorite samples exhibited relatively constant Nb/La ratios irrespective of Mg# value (Fig. S4). Such features cannot be produced by crustal assimilation of asthenospheric mantle–derived magma or mixing between asthenospheric mantle and ancient crust derivatives. Therefore, we propose that the geochemical characteristics of the monzodiorite are most consistent with fractional crystallization of enriched lithospheric mantle–derived magma without significant crustal assimilation. The negative Nb, Ta, and Ti anomalies of the monzodiorite samples are more likely an indication of subduction-related metasomatism in the source region. This inference is supported by the La/Sm_N values of the monzodiorite samples (3.2–4.5; Table S6), which are higher than the OIB average of 2.4 (Sun and McDonough, 1989), and by the similarity of primitive mantle–normalized trace-element distribution patterns of the monzodiorite samples to arc-related igneous rocks (e.g., Thirlwall et al., 1994).

The monzodiorite samples exhibited high Rb/Sr (0.06–0.19) and low Ba/Rb (3–15) values, similar to those derived from a phlogopite-bearing magma source (e.g., Furman and Graham, 1999; Fig. 18A), implying that phlogopites could have been involved in magma generation. This inference is consistent with the high K₂O contents and significant LILE enrichment of the monzodiorite samples (Fig. 12A), and with the high F contents of the apatite grains from the monzodiorite (up to 3.73 wt%; Table S8), as phlogopite is one of the major host phases for LILEs and volatile elements in the lithospheric mantle (McInnes and Cameron, 1994). Based on the above discussion, we conclude that the parental magma of the monzodiorite was most likely derived from partial melting of a phlogopite-bearing lithospheric mantle source, which was previously metasomatized by subduction processes.

To better constrain the magmatic fO₂ of the causative intrusion at Huanren, we employed whole-rock Fe₂O₃/FeO ratios (Ishihara, 1981), a magmatic Fe-Ti–oxide oxybarometer (Ghiorso and Evans, 2008), and a zircon trace-element oxybarometer (Loucks et al., 2020). The fO₂ estimates from these methods varied markedly. The Ishihara (1981) qualitative classification indicated that the causative intrusion of the Huanren Cu skarn was moderately oxidized to moderately reduced, and the Fe-Ti–oxide oxybarometer of Ghiorso and Evans (2008) suggested it was oxidized (FMQ +1.01 ± 0.60), whereas the zircon trace-element oxybarometer of Loucks et al. (2020) implied that it was reduced (FMQ −1.14 ± 0.14 to FMQ −1.04 ± 0.22).

The oxidation state of Fe (commonly expressed as the Fe₂O₃/FeO ratio) is a widely used index for magmatic redox state. Thus, wet-chemical measurements of Fe₂O₃/FeO ratios in whole-rock samples can be used to estimate the fO₂ of igneous rocks. Ishihara (1981) first proposed a whole-rock Fe₂O₃/FeO criterion to discriminate oxidized versus reduced igneous rocks whereby oxidized igneous rocks have whole-rock Fe₂O₃/FeO ratios greater than 0.5. Such relationships have subsequently been validated by theoretical and experimental investigations, which have served to emphasize the importance of magmatic fO₂ in determining metal ratios in magmatic-hydrothermal systems (e.g., Blevin et al., 1996; Blevin, 2004). A limitation of this method, however, is that Fe₂O₃/FeO ratios in igneous rocks are highly susceptible to alteration, especially for samples from mineral deposits. This is also the case at Huanren. As shown in Figure 19A, the Fe₂O₃/FeO ratios of the Huanren monzodiorite samples exhibited obvious variation with increasing loss on ignition, implying that their Fe₂O₃/FeO ratios have been modified by alteration. Accordingly, using whole-rock Fe₂O₃/FeO ratios (Ishihara, 1981) to estimate the magmatic fO₂ values of the monzodiorite samples at Huanren may not be valid.

The Fe-Ti–oxide oxybarometer has been improved by calibration in a wider range of conditions and has been widely used to estimate redox conditions in igneous rocks (e.g., Devine et al., 2003; Gilmer et al., 2018). A successful application of such an approach requires equilibrium conditions between Fe-Ti oxides at given temperature-pressure-composition-redox state. In this study, the Fe-Ti–oxide oxybarometer of Ghiorso and Evans (2008) was applied to adjacent ilmenite and magnetite pairs in the monzodiorite samples to estimate the oxidation state of the magmas (Table S7; Fig. S2). However, the estimated values were significantly higher than those derived from the zircon trace-element oxybarometer (FMQ +1.01 ± 0.60 and FMQ −1.04 ± 0.22 on average, respectively). The difference may be due to subsolidus disequilibration of Fe-Ti–oxide compositions during slow cooling (e.g., Hou et al., 2021). This inference is supported by the log (Mg/Mn) (atomic) values of the analyzed Fe-Ti–oxide pairs, which deviate significantly from the linear regression of equilibrated Fe-Ti–oxide pairs given in Bacon and Hirschmann (1988; Fig. 19B). The estimated fO₂ values of the monzodiorite based on the Fe-Ti–oxide oxybarometer varied widely (from FMQ +0.25 to FMQ +3.14; Table S7), further indicating that the Fe-Ti–oxide oxybarometer is problematic.

Zircon is usually resistant to compositional disturbance during hydrothermal alteration and postcrystallization slow cooling of intrusive rocks. Recently, several oxybarometers based on zircon composition have been proposed, e.g., zircon-melt partitioning cerium oxybarometer of Smythe and Brenan (2016) and zircon trace-element oxybarometer of Loucks et al. (2020). The former was established on the basis of experimental results and is sensitive to equilibrium melt composition, especially H₂O content (e.g., Zou et al., 2019), whereas the latter is empirically calibrated (only involves measured zircon Ti, Ce, and age-corrected initial U contents) and is independent of crystallization temperature, pressure, and equilibrium melt composition. The zircon trace-element oxybarometer of Loucks et al. (2020) has proven to be an effective method to quantify the fO₂ values of mafic to felsic intrusive rocks (e.g., Jara et al., 2021; Meng et al., 2021; Loader et al., 2022; Ge et al., 2023). In this study, using the zircon trace-element oxybarometer of Loucks et al. (2020), the estimated fO₂ value for the monzodiorite from the Huanren Cu skarn deposit is FMQ −1.09 ± 0.19 on average, which is dramatically lower than the fO₂ values of intrusions associated with porphyry-skarn Cu ± Mo ± Au deposits globally (FMQ +1.5 to FMQ +2.0 worldwide; e.g., Richards, 2015; Fig. 16). A reduced magmatic environment for the Huanren Cu skarn deposit is consistent with the predominance of magmatic ilmenite over magnetite in the Huanren monzodiorite (estimated petrographically; Fig. 4I). Moreover, typical oxidized porphyry-skarn Cu ± Mo ± Au systems precipitate hypogene chalcopyrite, pyrite, and bornite in equilibrium with primary hematite and anhydrite (e.g., Streck and Dilles, 1998; Seedorff et al., 2005; Sinclair, 2007). However, neither anhydrite nor hematite is present at Huanren, and the abundance of hypogene pyrrhotite (Fig. 8D) is consistent with a reducing magmatic-hydrothermal system. The low SO₃ contents of the magmatic apatite grains from the monzodiorite may also suggest a reduced magmatic environment. As shown in Figure 14, compared with magmatic apatite from classic oxidized porphyry-skarn Cu ± Mo ± Au intrusions, those from the Huanren deposit have negligible SO₃. A caveat is that low SO₃ in magmatic apatite can exist even in oxidized magmas if considerable S is lost or degassed. Although we have no independent record of the S content in the original Huanren monzodiorite magma, experiments under reducing conditions suggest at least 200 ppm S in silicic melts in equilibrium with hydrothermal fluid (Scaillet and Macdonald, 2006; Zajacz et al., 2012). This would result in easily detectable S in apatite (0.10 wt%) even in a degassed magma, assuming a D_S^{apatite/liquid} of 4.5–14 (Parat and Holtz, 2004), which is inconsistent with our results.

The evidence presented above indicates that the Huanren Cu skarn deposit represents a magmatic-hydrothermal system associated with a reduced monzodioritic intrusion, in contrast to other Cu (porphyry-)skarn systems worldwide (e.g., Dilles, 1987; Li et al., 2008; Xie et al., 2015; Chelle-Michou et al., 2015). Exactly how a reduced Cu skarn magma was generated at Huanren is not clear at this stage. Notably, however, as the coarse-grained and medium- to fine-grained monzodiorites at Huanren exhibit whole-rock geochemistry and zircon Hf isotopic compositions that are comparable to coeval enriched lithospheric mantle–derived mafic rocks from the North China Craton (Figs. 11–13), the most likely scenario is that their reduced character was inherited from a reduced metasomatized lithospheric mantle source rather than achieved via reduction of oxidized magmas during their ascent and emplacement upon assimilation of graphitic metasedimentary rocks. This inference is also supported by the homogeneous zircon Hf isotope composition of the monzodiorite samples (Fig. 13).

Oxygen fugacity is commonly considered to be an important thermodynamic parameter in the formation of magmatic-hydrothermal deposits because it exerts a first-order control on the speciation of sulfur in magmas and thus the partitioning of ore-forming metals between minerals and melts (e.g., Jugo et al., 2005, 2010; Kiseeva and Wood, 2013). In reduced magmas, most sulfur exists as S²⁻, and so it cannot be accommodated in the apatite structure because apatite only accommodates oxidized S species (e.g., Jugo et al., 2005, 2010). The low SO₃ content of the apatites from the reduced Huanren monzodioritic samples (Fig. 14B) is in agreement with this observation. In these magmas, sulfur will likely form sulfides that can strip the magma of metals such as Cu (Candela, 1992; Blevin and Chappell, 1992), a process that was proposed to be unfavorable for effective Cu transport through the crust and for later Cu skarn mineralization (e.g., Meinert et al., 2005; Richards and Şengör, 2017). A paradox of the Huanren copper skarn is why the reduced magma at Huanren was still fertile for Cu mineralization. Our geochemical data show that, compared to oxidized porphyry-skarn Cu ± Mo ± Au intrusions, the causative intrusion at Huanren was significantly less evolved, being trachybasaltic in composition (Fig. 11A). This characteristic may indicate that, although sulfur saturation in the form of sulfides likely occurred at Huanren, this process may not have significantly stripped substantial Cu from the magma, as it was relatively mafic. Other potential factors include: (1) the melt Cu content is not crucial for magma fertility (e.g., Richards, 2015; Chelle-Michou and Chiaradia, 2017; Rezeau and Jagoutz, 2020) or (2) the melt Cu content is not appreciably affected by oxidation state (e.g., Meinert, 1993).

Because the magmatic-hydrothermal system at Huanren was reduced, another question is: How was Cu transported from the magma and into the aqueous phase? According to Williams-Jones and Migdisov (2014), Cu has higher solubility in oxidized fluids compared to reduced fluids, but the difference decreases with increasing temperature. For example, in the temperature range from 450 °C to 560 °C, the solubility of Cu in a fluid buffered by the oxidized assemblage of pyrite-magnetite-hematite is 400–2300 ppm, whereas the solubility of Cu in a fluid buffered by the reduced assemblage of pyrite-pyrrhotite-magnetite is 250–2000 ppm (Williams-Jones and Migdisov, 2014). Moreover, for the salinity and temperatures typically encountered in porphyry-skarn systems, Cu is transported dominantly as an aqueous chloride complex (e.g., CuCl⁰ and CuCl²⁻), even under reducing conditions (e.g., Hemley et al., 1992; Williams-Jones and Migdisov, 2014). Taken together, although Cu in most porphyry-skarn systems is interpreted to have been transported under oxidizing conditions, reduced aqueous phases are also capable of transporting appreciable quantities of Cu if the magmatic-hydrothermal systems are not depleted in Cl. This interpretation is supported by the magmatic apatites of the monzodiorite, which contain 0.83 ± 0.27 wt% Cl (Table S8), indicating the Huanren magmatic-hydrothermal system was not Cl poor. Audétat (1999, 2019) demonstrated that the fluids in pure Sn and W deposits (commonly related to reduced intrusions) actually contain even higher concentrations of Cu than in porphyry-skarn Cu ± Mo ± Au deposits, further demonstrating that reduced fluids can transport appreciable concentrations of Cu if the fluids contain substantial Cl.

The sulfur isotopic compositions of chalcopyrite in the Huanren deposit range from 2.8‰ to 5.8‰ (4.34‰ ± 0.88‰ on average; Fig. 17), i.e., higher than the chalcopyrite δ³⁴S values from the majority of porphyry-skarn Cu ± Mo ± Au deposits (−3.0‰ to +1.5‰; Fig. 17) and the δ³⁴S values of mantle (−1.0‰; Labidi et al., 2013), primitive basalt (−3.0‰–0‰; Labidi et al., 2013), and average lower and upper crust intrusions (0‰–3.0‰; Rezeau et al., 2023). Our chalcopyrite sulfur isotope data indicate that addition of external sulfur with a high δ³⁴S signature probably occurred at Huanren during the precipitation chalcopyrite, most likely by the interaction of reduced magma-derived ore fluids with fluids equilibrated with evaporites (e.g., subsurface saline fluids). This explanation is consistent with the fact that the Cambrian carbonate rocks in the Huanren area contain lenses of evaporite composed of gypsum, anhydrite, and halite (Zhao et al., 2001; Wu et al., 2014). In addition, the presence of Na-Ca alteration in the monzodiorite (e.g., epidote; Fig. 4D) indicates that external fluids may also have contributed Cl to the Huanren magmatic-hydrothermal system, probably by facilitating the transportation of Cu from reduced monzodioritic magma into hydrothermal fluids. Notably, compared to apatites from typical porphyry-skarn Cu ± Mo ± Au intrusions, the apatite Cl contents of the Huanren monzodiorite are at the lower limit (Fig. 14B). The relatively low Cl content of the Huanren magmatic system may account for the intermediate size of the Huanren Cu deposit.

Based on the discussion above, we propose a model for ore-forming fluid evolution and ore deposition at Huanren in Figure 20. Following intrusion of the reduced monzodioritic magma, cooling and crystallization led to fluid saturation, hydrofracturing, and upward movement of reducing magmatic fluids (containing Fe²⁺, Cu²⁺, Mo⁴⁺, Pb²⁺, Zn²⁺, S²⁻, Cl⁻, Na⁺, and K⁺) to produce the porphyry-type Cu and Mo mineralization in the monzodiorite. Fracturing allowed oxidized fluids that were equilibrated with evaporite (containing SO₄²⁺, Cl⁻, Ca²⁺, Mg²⁺, Na⁺, and K⁺) to move into the cooling monzodiorite, where they were heated, causing Na-Ca alteration in the monzodiorite. As this fluid circulated upward through the top of the intrusion and mixed with reduced magmatic fluids, deposition of magnetite began following the reaction: 12Fe²⁺ + SO₄²⁻ + 12H₂O = 4Fe₃O₄ + H₂S + 22H ⁺ . According to this reaction, magnetite deposition was accompanied by an increase in H₂S in the fluids, which may have effectively driven the precipitation of sulfides following reactions such as Cu⁺ + Fe²⁺ + 2H₂S = CuFeS₂ + 3H ⁺ .

It is widely accepted that skarn Cu deposits commonly form from oxidized magmas and fluids (e.g., Candela, 1992; Meinert et al., 2005; Richards, 2015). Our study at Huanren shows that Cu skarns may also be genetically related to reduced magmatic-hydrothermal systems, suggesting that not all Cu skarn deposits are associated with oxidized magmas and fluids. This defines a new type of Cu skarn and thereby opens new potential for Cu skarn exploration. Moreover, our study shows that the Cl contents of the magmatic system and the availability of external S and/or Cl were as critical as fO₂ in facilitating the actual ore-forming event in the Huanren Cu skarn system. This is because Cl is an important control on the complexing and transport of Cu in fluids, and S is required for the precipitation of economically important sulfide minerals such as chalcopyrite from the fluid. Consequently, we propose that, in addition to magmatic fO₂, the availability of Cl and S in magmatic-hydrothermal systems may also be crucial in determining the genesis of Cu skarn deposits.

With only the Huanren example known for the occurrence of Cu skarn related to a reduced monzodioritic magma, it is likely too early to summarize the common features into a broadly applicable genetic model. However, based on the present study, we propose the following paradigm for the exploration of Cu skarns associated with reduced intrusions: (1) the presence of ilmenite-series intrusions, with (2) a relatively high magmatic apatite Cl content, and (3) the presence of evaporite layers in carbonate host rocks. Notably, because coeval mafic rocks in the North China Craton have many similarities in whole-rock major- and trace-element compositions (Figs. 11–12 and 18), zircon Hf isotopic compositions (Fig. 13), oxygen fugacities (e.g., Huang et al., 2016), and volatile contents (e.g., Zheng et al., 2023) with the Huanren monzodiorite, further Cu skarn exploration in the North China Craton should focus on these intrusions, especially in areas where evaporite-bearing carbonate rocks have developed (e.g., the Liaodong Peninsula; Fig. 1A).

It is known that most Cu skarns are typically formed in subduction settings (Meinert et al., 2005). The Huanren Cu skarn, however, was formed in an intracontinental extensional setting and is associated with a lithospheric mantle–derived intrusion that has many arc-like chemical characteristics. From the perspective of Cu skarn formation, this indicates that a chemically preconditioned lithospheric mantle source might be more important than the tectonic setting in which it melts. Moreover, magmas formed in intracontinental extensional settings are generally characterized by having low oxygen fugacities (lower than arc-related magmas by 2 units of FMQ; Cottrell et al., 2021) and are commonly interpreted to be infertile for Cu mineralization (Meinert et al., 2005; Richards, 2015). The recognition of the reduced monzodioritic intrusion–related Huanren Cu skarn indicates that reduced monzodioritic intrusions should not be ignored as hosts for Cu deposits, especially in intracontinental extensional settings that are known as repositories for relatively reduced magmas (Cottrell et al., 2021), and in areas that contain evaporite-bearing carbonate country rocks. However, the importance of the reduced Cu skarn deposits with regard to their metal endowments in these geological settings remains unclear and requires further investigation.

Our field investigations showed that skarn Cu mineralization at Huanren was spatially and temporally associated with monzodiorite intrusion. LA-ICP-MS U-Pb zircon geochronology indicates the monzodiorite crystallized at 125.4 ± 0.6 Ma (2σ). ID-ICP-MS Re-Os molybdenite geochronology indicates that skarn mineralization at Huanren occurred at 125.3 ± 0.8 Ma (2σ). The Re-Os age for the mineralizing event overlaps within uncertainty with the LA-ICP-MS U-Pb zircon age of the monzodiorite, confirming a genetic link between skarn mineralization and monzodiorite intrusion at Huanren. Geochemical data indicate that the parental magma responsible for the generation of the Huanren deposit was sourced from a subduction metasomatized lithospheric mantle without significant crustal assimilation. Zircon trace-element and magmatic apatite major-element compositions of the monzodiorite indicate that the ore-forming magmas at Huanren were reduced. Together, these results indicate that the Huanren Cu skarn system was formed by the emplacement of a reduced monzodioritic intrusion. The Cl-bearing nature of the Huanren magmatic-hydrothermal system, possibly aided by the interaction of reduced magmatic fluids with external oxidized fluids that were equilibrated with evaporite, increased the ability of fluids to scavenge Cu from the reduced magma and subsequently precipitate Cu in the carbonate rocks. The recognition of Cu skarn mineralization related to a reduced monzodioritic magma at Huanren suggests that the availability of Cl and S in magmatic-hydrothermal systems may be as critical as fO₂ in facilitating the actual ore-forming event in Cu skarn systems. The recognition of reduced Cu skarn deposits does not contradict the current understanding of Cu skarn formation; rather, it adds yet another dimension to Cu skarn formation.

This research was supported by the Scientific Research Fund of the China Central Non-Commercial Institute (KK2203) and the National Natural Science Foundation of China (41925011, 43272096). We are grateful to three anonymous reviewers for their constructive reviews, which greatly improved an earlier version of this manuscript, and Chengshuai Zhang for fieldwork. We extend special thanks to Hervé Rezeau and editors Peter Luffi and Brad Singer for particularly thoughtful reviews that helped to improve the clarity and scope of the manuscript.

Erratum Erratum: Copper skarn mineralization related to reduced monzodioritic magma at the Huanren deposit, northeastern China

Hegen Ouyang, John Caulfield, Guiqing Xie1, Chao Duan, Jingwen Mao, and Xin Li

An error was made in the pre-issue publication version of this paper, first published online on 9 February 2024. In the Abstract, “(fayalite-magnetite-quartz [FMQ] buffer = 1.09 ± 0.19)” should have read “(fayalite-magnetite-quartz [FMQ] buffer = −1.09 ± 0.19).” The correct text now appears in the final published version of the paper to which this erratum is appended.