Geology , Mineralogy , Fluid Inclusion , and H – O – S – Pb Isotope Constraints on Ore Genesis of the Keyue Sb – Pb – Zn – Ag Deposit in Southern Tibet

State Key Laboratory of Geological Processes and Mineral Resources, School of Earth Sciences and Resources, China University of Geosciences, Beijing 100083, China No. 10 Detachment, Division of Gold, Chinese People’s Armed Police Force, Kunming 650001, China State Key Laboratory of Geological Processes and Mineral Resources, and Faculty of Earth Resources, China University of Geosciences, Wuhan 430074, China China Three Gorges Projects Developments Co., Led, Beijing 100038, China


Introduction
The North Himalayan Metallogenic Belt (NHMB), located in the most representative orogenic belt worldwide, is an important part of the Tethys-Himalaya metallogenic domain [1].The NHMB is located in the eastern section of the North Himalayan Tectonic Belt (NH), and the main ore-bearing formations are a set of Late Triassic-Early Cretaceous flysch formations formed by turbidity sediment and carbonaceous-siliceous-argillaceous rock series related to hydrothermal sedimentation [1,2].According to previous research [1][2][3][4], there are mainly three mineralization events within NHMB that generate a series of Sb, Sb-Au, Au, Pb-Zn (Ag), and Sb-Pb-Zn-Ag deposits (Figure 1).The first mineralization event is related to multiple seafloor volcanic events during a synsedimentary period (220~130 Ma; e.g., the first pulse of mineralization in the Zhaxikang Sb-Pb-Zn-Ag deposit); the second mineralization event is associated with a

Geological Setting
2.1.Regional Geology.The Himalayan Terrane, located between the Lhasa Terrane and the Indian Continent, is an important tectonic unit of the Cenozoic Himalayan-Tibetan orogen (Figure 1(a)) [21].From north to south, the Himalayan Terrane is divided into the NH, the High  [5] and Wang et al. [25]).(b) Regional geological map of the North Himalayan Metallogenic Belt (modified from Zheng et al. [1] and Wang et al. [25]).
2 Geofluids Himalayan Crystalline Rock Belt (HH), the Low Himalayan Fold Belt (LH), and the Sub-Himalayan Tectonic Belt (SH), respectively (Figure 1(a)) [22,23].These four tectonic belts are separated by four nearly EW-trending faults with the name of the South Tibet Detachment System (STDS), the Main Central Thrust (MCT), the Main Boundary Thrust (MBT), and the Main Frontal Thrust (MFT; Figure 1(a)) [5,24].The sedimentary sequence that includes the Late Precambrian to Devonian pre-rift, Carboniferous to Early Jurassic syn-rift, and Middle Jurassic to Cretaceous passive continental margin sediments, are recorded in NH and crop out in an EW and NWW trend [20].The strata are composed of the Precambrian Laguigangri Group and a series of Upper Triassic, Jurassic, Lower Cretaceous, and Quaternary sediments [25].A set of Late Triassic-Early Cretaceous flysch formations formed by turbidite deposits in neritic-bathyal environments is in dominance and crops out across the NH.This set of flysch formations host the majority of the Au-Sb-Pb-Zn-Ag deposits in the NHMB and consist of weak metamorphic slate that is intercalated with metamorphosed fine-grained sandstone, argillaceous limestone, micrite, and siliceous rock that is intercalated with volcanic rocks [2].In addition, the Laguigangri Group containing schist, gneiss, and migmatite units crops out in the core of the Yelaxiangbo dome (Figure 1(b)), as well as some Quaternary sediments composed of gravel, sand gravel, sandy loam, clay, and ice boulder spread in the central and southern areas of the NH [2].
Two sets of faults both with multiple episodes of motion occur in the NH, and they are EW-trending and NStrending, respectively.The older EW-trending faults cover a larger area and control the distribution of intermediate-acid magmatic rocks and ore deposits in the NH [2].The Lazi-Qiongduojiang, Rongbu-Gudui, and Luozha faults, as well as the STDS and numerous metamorphic core complexes, are all representative EW-trending faults (Figure 1(b)).In comparison, the younger NS-trending faults are also the important ore-controlling structures in the NH [19] and are considered as the result of east-west extension of the Qinghai-Tibet Plateau [26,27].These NS-trending faults are deemed to form from 25 Ma to now, especially during 18 to 4 Ma [28,29].Among these NS-trending faults, a series of rifts is considered to be associated with the older EW-trending faults [21,22,30,31], which mainly include the Sangri-Cuona, Yadong-Gulu, Shenzha-Xietongmen, and Dangreyongcuo-Gucuo rift zones from east to west [32].
The Mesozoic and Cenozoic magmatism is recorded in the NH.From Late Triassic to the Early Cretaceous, the Mesozoic magmatism generated multiple suites of maficintermediate igneous rocks that include basaltic volcanic interlayers, dyke swarms, and subvolcanic dykes.The previous geochronological data reveal that the basic dyke swarms from different areas in the NH have SHRIMP U-Pb ages of 134.9 ± 1.8 Ma, 135.5 ± 2.1 Ma [33], and 138.0 ± 3.5 Ma [34], respectively.The SHRIMP U-Pb age of the gabbro is 155.8Ma [35].However, the genesis of these magmatic rocks is controversial.Tong et al. [34], Pan et al. [36], and Zhong et al. [37] regarded these basic dyke swarms as the result of the late-stage massive expansion of the Neo-Tethys Ocean under the structural environment of the Himalayan passive continental margin intensive stretching and breaking-off, lithosphere extension-thinning, and asthenosphere upwelling.On the contrary, Zhu et al. [38] and Qiu et al. [39] suggested that these basic dyke swarms are the result of the interaction between mantle plume and lithospheric mantle material and form in the continental rift environment.The Cenozoic magmatism, considered to be the result of crustal thickening [40] related to the collision of the Indian Plate and the Eurasian Plate during the postcollision period (25 to 0 Ma) [41,42], is characterized by the formation of monzogranite, leucogranite, diorite, porphyritic diorite, and aplite units [43,44].These Cenozoic intermediate-acidic intrusive masses distribute in the core of Ranba, Kangma, and Yelaxiangbo domes as well as the EW-trending faults in the form of batholith, laccolith, and dykes (Figure 1(b)).
2.2.Ore Deposit Geology.The Keyue deposit is located ~43 km west from Longzi County Town in NHMB, where it is near the Zhaxikang deposit (Figure 1(b)).This deposit has a reserve of 232,100 t Pb with an average grade of 3.00%, 172,648 t Zn with an average grade of 2.44%, 81,261 t Sb with an average grade of 1.15%, and 711.8 t Ag with an average of 100.72 g/t [45].The majority of mineralization in the orefield is hosted by the Lower Jurassic Ridang formation that dips shallowly to the north with the dip angle of 10 °~20 °.This formation is dominated by a set of epi-metamorphic marine clastic rocks that mainly consists of slate and some shale (Figure 2(a)) with few quartzose sandstone, calcareous sandstone, limestone, and tuff [45].Some Middle Jurassic Zhela formations composed of limestone and dolomite with few phyllite rock, slate, and quartz siltstone that is intercalated with altered intermediate-basic volcanic rock, as well as Quaternary sediments distributed along valleys, also crop out in the orefield (Figure 2(c)) [10].The intense magmatic and volcanic activities have generated a series of intermediate-basic and acidic magmatic rocks hosted by the Ridang Formation.Separately, the diabase is identified by drillhole and footrill in the central part of the orefield as dykes that emplaced into the Jurassic Ridang Formation (Figure 2(a)); the basalt usually occurs near the orebody in the form of consequent layer or shear layer distributing in slate and the contact position of slate and diabase [10,45].
A series of faults and some folds developed in the Keyue orefield.These folds make the rocks within the strata deformed and metamorphic in different degrees [10].Engineering and geological mapping projects of Huayu Mining Company [45] have identified two dominating faults (F1 and F2) that control the distribution of 7 orebodies within the orefield (Figure 2).The NS-striking F1 fault, with the length of 1600 m, width of 100 m, altitude of 265 °~295 °, and dip angle of 50 °~70 °, cut the Ridang Formation.In comparison, the smaller F2 fault also cut the Ridang Formation and is 500 m long and 1~15 m wide with the altitude of 110 °~120 °and dip angle of 55 °~70 ° [45].The orebodies are usually veined, stratiform, and lenticular with ramification in the depth.Combining with the observation of drillhole and 3 Geofluids footrill, the orebodies have an obvious crack related to hydrothermal activity [10,46].Among these orebodies, orebody 1 controlled by fault F1 is the largest and richest one within the orefield, which hosts an overwhelming majority of the reserves (Figure 2) [45].
From altitude 4422 to 4722 m in orebody 1, the oreforming elements of the Keyue Sb-Pb-Zn-Ag deposit are zoned from a lowermost Pb + Zn zone through the central Ag-Sb (Pb + Zn) zone to an uppermost Sb (Hg) zone (Figure 2(c)), although there is no horizontal zoning [45,46].Additionally, various types of alteration associated with mineralization have occurred in the orefield [10,45,46], including (1) the carbonatization that is associated with Pb-Zn mineralization in the form of Mn-Fe carbonate veins (Figures 3(a

Materials and Analytical Methods
We have done the field work in the Keyue orefield for more than 120 days in recent years, including the mapping, drillhole and footrill record compiling, and sampling.Around 550 ore and related rock samples have been collected from the drillhole, footrill, and dressing plant.Meanwhile, we also took more than 800 pictures for the geological phenomenon and samples.Besides the hand specimens' observation, 165 thin sections were also made for microscopic observation (around 300 photomicrographs have been taken) on a Zeiss AX 10 microscope with an AxioCam MRc 5 microscope camera system under plane polarized light, perpendicular polarized light, and reflected light.Combining these observations, the representative samples are selected for electron probe microanalysis (EPMA), scanning electron microanalysis (SEM), energy-dispersive spectroscopy (EDS), fluid inclusion, and H-O-S-Pb isotopic analyses.

EPMA and SEM-EDS.
To determine the chemical compositions of sulfides and sulfosalt minerals, 105 finely polished sections were prepared for EPMA.The EPMA for 260 points in these sections was conducted on a JEOL JXA-8100 electron microprobe at the State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences, Wuhan.The electron microprobe was operated under the condition of a 20 kV accelerating voltage, a 20 nA beam current, and a 5 μm spot diameter.The natural and synthetic standards were used to control the analytical quality, and the tested accuracy is better than 0.01%.The correction program supplied by the manufacturer is used for matrix corrections [47].Based on the microscopic observation and EPMA, 20 points in five representative sections were selected for SEM-EDS analyses to study the mineral compositions, textures, coexisting relationships, and occurrence state of Ag.The SEM-EDS analyses were performed at the State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences, Wuhan, using a Quanta 450 FEG-SEM equipped with an SDD Inca X-Max 50 EDS system.

Fluid Inclusion.
The petrography, microthermometry, and laser Raman spectrometry analyses of fluid inclusions (FIs) from 4 sphalerite, 2 Mn-Fe carbonate, and 9 quartz samples in different assemblages were carried out in the State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences, Wuhan.The sample preparation for microthermometric measurement of FIs followed the procedures described in Shepherd et al. [48].The microthermometry data were obtained using a Linkam THMS600 heating-freezing stage attached with a Zeiss microscope and related standard procedures, which has an enabled measurement within the range of −196 to 600 °C.Liquid nitrogen and a thermal resistor were used for freezing and heating runs, respectively.The ice-melting temperatures (T m−ice ) and total homogenization temperatures (T h−total ) of fluid phases in FIs have been recorded during the measurements.During the initial stages of each heating run, the heating rate was maintained at 5 °C per minute and then reduced to 0.1 °C per minute when the temperature is approached to the phase change points.Usually, the heating measurements were performed after freezing measurements.The densities of FIs were calculated using the FLINCOR procedure as in Brown and Lamb [49], and the salinities of aqueous (NaCl-H 2 O) inclusions were calculated using the final melting temperatures of ice points [50].In addition, a Renishaw MK1-1000 Laser Raman microspectrometer is used to determine the compositions of individual FIs in minerals, including vapor and liquid.The instrumental settings were kept constant during the analyses.An argon laser with a wavelength of 532 nm was used as the laser source at the power of 1000 MW.The spectral range fell between 1000 and 4000 cm −1 for the analyses, and the spectral resolution was ±2 cm −1 with a beam size of 1 μm.

H-O Isotopic
Analyses.The H-O isotopic compositions of 12 quartz and 2 calcite samples from the Keyue deposit were measured on a MAT 253EM mass spectrometer at the Institute of Mineral Resources, Chinese Academy of Geological Sciences, Beijing.For H isotopic analyses, the water from the fluid inclusion separates was collected by thermal decrepitation under vacuum at 500 °C and then reduced over zinc at 400 °C to generate H 2 that is released for mass spectrometry [51,52].As for O isotopes, oxygen was generated by the BrF 5 method the same as in Clayton and Mayeda [53], and then the resultant oxygen was converted to CO 2 by reaction with a graphite rod for the following measurement.The H-O isotopic values were reported relative to the V-SMOW standard, and the analytical errors are ±2‰ for δD and ± 0.2‰ for δ 18 O.

S-Pb Isotopic
Analyses.Nine sphalerite, five galena, three boulangerite, two chalcopyrite, two pyrite, and one stibnite samples were chosen for S isotopic analyses, and five galena, two boulangerite, and one stibnite samples were selected for Pb isotopic analyses.Separates of these minerals (40~80 mesh) were prepared by careful handpicking under a binocular microscope to achieve a purity more than 99.8%.Both the S-Pb isotopes were analyzed at the Analysis and Testing Research Center of the Nuclear Industry, Beijing Institute of Geology, Beijing, China.
For S isotopic analyses, a mixture of approximately 15 mg mineral separates and 150 mg Cu 2 O was crushed into 200 mesh and heated to 980 °C under a vacuum pressure of 0.02 Pa to convert sulfur to sulfur dioxide (SO 2 ) [54].Then the generated SO 2 was measured on a Thermo Scientific Delta V Plus mass spectrometer.The S isotopic values were reported relative to the Canyon Diablo Troilite (V-CDT) standard, and the analytical uncertainties were better than ±0.2‰.
For Pb isotopic analyses, approximately 20 mg mineral separates were placed in 15 ml Teflon jars and the solids were dissolved in a mixture of HCl-HNO 3 .After 24 h, the solutions were dried and 1 ml 6 N HCl was added, then the samples were dried again and dissolved in 1 ml 0.5 N HBr.After being centrifuged, lead was purified using anion exchange chromatography and collected in 1 ml 6 N HCl.At last, the dried samples were coated on a single rhenium filament by the classic H 3 PO 4 -silica gel method, and the Pb isotopic compositions were measured on a GV IsoProbe-T thermal ionization mass spectrometer (TIMS).All the measured Pb isotopic values were corrected according to the values of NBS SRM 981 standards, and the measurement accuracy was better than 0.005% (2σ).
Chalcopyrite is the prime Cu-bearing mineral within the orefield.The xenomorphic fine-to coarse-grained chalcopyrite (0.3~6 mm) coexisting with pyrite, galena, sphalerite, and arsenopyrite is hosted by Mn-Fe carbonate and quartz to form the disseminated, veined, and massive ores (Figures 4(c 1), which reveals the substitution between Fe-Zn and Cu-Zn, as well as positive correlation between Ag-Fe and Ag-Cu.In addition, andorite is another Ag-and Sb-bearing mineral but is rarely found in the Keyue orefield; we can only find the stumpy andorite distributed in freibergite (Figure 6(d)).
Stibnite is the main Sb-bearing mineral within the orefield and usually occurs in the upper part of the orebody.Stibnite coexists with quartz to form the veined and massive ores (Figures 4(r)-4(t) and 5(f)), and the stibnite-quartz veins sometimes cut the earlier boulangerite (Figure 5(e)).Stibnite contains 70.179%~70.3530%Sb, 28.400%~28.5340%S, and trace amounts of As, Pb, Cd, Sn, and Cu (Table 1).Meanwhile, cinnabar is the most important Hg-bearing mineral in Keyue orefield, and the fine-grained cinnabar (0.5~1 mm) aggregates coexist with stibnite and quartz (Figure 4(t)).
The drusy and radial valentinite with yellow color and geyserite usually coexist with stibnite and cinnabar (Figures 4(t)), the green massive malachite is intergrown with chalcopyrite, the brown-black massive ferrohydrite occurs in the surface overgrowing Mn-Fe carbonate and pyrite, and the white-grey sardinianite usually shows massive and nodular textures and coexist with galena.Meanwhile, the BSE and X-ray mapping images from the EPMA, SEM, and EDS analyses reveal the following facts: (1) the border between the Ag-bearing and other minerals is very clear without transition region; (2) around the main Ag-bearing minerals, there are tiny Ag-bearing inclusions distributing in galena, with Ag-bearing inclusions differing in size and shape; and (3) there are no Ag-bearing inclusions present in other coexisting minerals, such as sphalerite and sulfosalt minerals (Figure 6).These facts suggest that the Ag-bearing inclusions are another occurrence state of visible Ag and relates to galena.
Furthermore, in some tiny areas (<1 μm) where there are no Ag-bearing inclusions observed, the EPMA shows that there is Ag enrichment, which indicates that there is invisible Ag within sulfides in the Keyue deposit.There are two following invisible Ag types: (1) isomorphism lattice Ag: under high-temperature conditions, Ag and Sb enter into the mineral lattice by isomorphic substitution with Pb, and (2) submicroscopic Ag-bearing inclusions: under low- temperature conditions, Ag and Sb together form the sulfosalt minerals [55,56].In view of microthermometry research of fluid inclusions, the invisible Ag in galena should be mainly submicroscopic Ag-bearing inclusions.
Overall, the occurrence state of Ag in the Keyue deposit are prevailingly in the form of visible Ag that can be divided into independent Ag-bearing minerals (>50 μm; e.g., tetrahedrite, freibergite, and andorite) and Ag-bearing inclusions (1~10 μm).By contrast, invisible Ag is much less and principally in the form of submicroscopic Ag-bearing inclusions.

4.2.
Ore Paragenetic Sequence.Based on the detailed hand specimen and microscopic observations, six ore-forming assemblages of ore and gangue minerals and one supergene assemblage have been recognized (Figure 7).The details are as follows.
The supergene assemblage prevailingly includes ferrohydrite, sardinianite, valentinite, malachite, and siliceous sinter (Figures 4(t        (3) pseudosecondary FIs: the shape of pseudosecondary FIs is similar to that of secondary FIs, and their compositions are consistent with primary FIs yet usually distribute in mineral crystal in the form of filiform assemblages (Figure 8(b)).
(2) Phase Types.Based on the observation of FI compositions and phases at room temperature (21 °C), as well as phase transitions during heating and cooling, there are four phase types of FIs which have been recognized in the Keyue deposit: (1) water FIs (W-type): the W-type FIs are the most common type in the Keyue deposit and are widespread in every oreforming assemblages, which accounts for more than 90% of the total amount.The W-type FIs are usually black and consist of vapor and liquid phase water with minor reducing gases (e.g., CO 2 ).At room temperature (21 °C), the W-type

Microthermometry.
In view of the following facts, (1) the L-type FIs are in dominance, (2) the amounts and sizes of the FIs in assemblages 1 and 2 Mn-Fe carbonate and assemblage 5 quartz are relatively less and tiny, and (3) the primary FIs can constrain the characteristic of oreforming fluid better; the microthermometry research are mainly trapped on 306 primary L-type FIs from 13 representative assemblage 2 sphalerite and assemblage 3 and 4 quartz samples.The results are listed in Table 2.

Discussion
5.1.Geology, Mineralogy, and Element Characteristics.Based on the detailed field and petrological observations, as well as the cross-cutting relationships between minerals, the six ore-forming assemblages of ore and gangue minerals can represent six stages of ore formation from early to late  21 Geofluids (Figure 7).These six stages are assigned to three clear pulses, and this ore paragenetic sequence in the Keyue deposit is similar to that in the Zhaxikang deposit [20,25].
The first pulse is Pb-Zn mineralization which consists of stages 1 and 2. Stage 1 is the initial stage of ore formation, and stage 2 hosts majority of the Pb-Zn mineralization in the Keyue deposit.Although the net-veined, brecciated, concentric annular, and globular textures that are present in the Zhaxikang deposit are rarely found in the Keyue deposit, the lamellar (Figures 4(a) and 4(b)), disseminated (Figure 4(c)), banded, and massive (Figures 4(d)-4(i) and 5(b)) ore textures; the assemblages and mineralogy of Mn-Fe carbonate and sulfides; and the hydrothermal sedimentary ore-bearing rocks (Ridang Formation) [2] still reveal the submarine hydrothermal sediment (metasomatism) genesis of the first pulse of mineralization in the Keyue deposit [1,2].Meanwhile, in view of that the Ridang Formation also contains some marine volcanic rocks [1,2], Pb-Zn mineralization may relate to the seafloor volcanism.The second pulse is Sb-Ag mineralization which includes stages 3 and 4. Stage 3 is the transitional period from the first pulse to the second pulse of mineralization and overprints the first pulse mineralization.Stage 4 hosts majority of the Sb-Ag mineralization and yields the ores with relatively high average Ag grades.Comparing with the Zhaxikang deposit, the Keyue deposit contains a higher amount of chalcopyrite in stage 3 and fewer kinds of sulfosalt minerals in stage 4. The third pulse is Sb-Hg mineralization which is composed of stages 5 and 6.Stage 5 hosts part of Sb and majority of Hg mineralization, and stage 6 represents the youngest stage of mineralization in the Keyue deposit.The veined, massive, needle-like, elongated, and radial ore textures; the assemblages and mineralogy of quartz-calcite-sulfide-sulfosalt minerals; and the visible Ag in tetrahedrite, freibergite, andorite, and Ag-bearing inclusions and the invisible Ag in the form of submicroscopic Ag-bearing inclusions all evidence the epithermal genesis of the second and third pulses of mineralization in the Keyue deposit, which may be related to the later magmatism [1,2,10,19].
Comparing the geology, mineralogy, and ore paragenetic sequence of the Keyue and Zhaxikang deposits, the orebodies are all controlled by faults, the ore-bearing rocks are both Ridang formations, the mineralogy and ore paragenetic sequence of ore and gangue minerals are almost the same, and there is only one regional fault that separates them.We infer that there is a genetic relationship between the Keyue and Zhaxikang deposits, and they have the same origins and experiences and the same mineralization events.On the other hand, both the orebodies in the Keyue and Zhaxikang deposits are controlled by faults, which do not present the typical pattern of SEDEX deposits that the bedded orebody occurs in the upper part and the veined orebody occurs in the lower part.Zheng et al. [1,2] explained this phenomenon in Keyue and Zhaxikang deposits as follows: the location of ore-bearing faults is ancient hydrothermal vents, and due to the collision of the Indian and Eurasian plates, the Plateau Uplift makes the bedded orebody to be eroded, and the later tectogenesis forms the faults and destroys the orebodies to the present form.The fact that orebodies have an obvious crack related to hydrothermal activity is also a good evidence.
Additionally, the elemental characteristics of sulfides can also be used to distinguish the deposit type.For instances, Wen et al. [59] investigated several Pb-Zn deposits with different geneses in China and divided them into three systems according to Zn/Cd ratios: (1) high-temperature systems (porphyry, magmatic hydrothermal, skarn, and volcanic hosted massive sulfide-(VMS-) type deposits): 155~223; (2) low-temperature systems (Mississippi Valleytype (MVT) deposits):17~201; and (3) exhalative systems: ① SEDEX-type deposits: 316~368 and ② seafloor hydrothermal sulfides: 211~510.The results indicate that the exhalative system usually has the highest Zn/Cd ratios.Calculated from the EPMA data, the Zn/Cd ratios of sphalerite range from 296 to 4402 in the Zhaxikang deposit [9,20] and 1226 to 12,846 in the Keyue deposit (Table 1), which are both much higher than those of high-and low-temperature systems (Figure 12).However, these Zn/Cd ratios are also higher than those of exhalative systems, which could be related to the Cd element dispersion resulting from the overprint by second and third pulses of mineralization.Meanwhile, the Co/Ni ratios of pyrite are also an effective indicator for ore genesis: the Co/Ni ratio of pyrite related to volcanism is usually >1 and concentrates on 5~10, the Co/Ni ratio of sedimentogenic pyrite is <1, and the magmatic hydrothermal pyrite has Co/ Ni ratios ranging from 1 to 5 [60][61][62][63].As for the Keyue deposit, the Co/Ni ratios of pyrite range from 3.7 to 7.1 according to the EPMA data (Table 1), which indicates that Note: Abbreviations are the same as in Table 4.
22 Geofluids pyrite could be influenced by volcanism and magmatism.The inferences from geology, mineralogy, and elemental characteristics are consistent.

FIs
Evidence Implications for the Nature and Origin of Ore-Forming Fluids 5.2.1.Epithermal Genesis for the Keyue Deposit.Previous literatures [64,65] suggest that FIs can provide a record of the ore-forming fluid system and trace the nature and genesis of original fluids.In the "temperature versus salinity" plot (Figure 13), the FI data almost fall into the epithermal deposits area.Meanwhile, when assuming a simple H 2 O-NaCl system, the ore-forming pressure can be estimated by the "temperature versus density" plot [66].Plotting the FI data of the Keyue deposit in the "temperature versus density" plot (Figure 14

Geofluids
These FI data suggest that the ore-forming fluid for the first pulse of mineralization is the low-temperature, lowmedium salinity, and low-density H 2 O-NaCl system containing CO 2 .This is similar to the Red Dog (100~200 °C and 14%~19%) and Tom (234~274 °C and 2%~18%) SEDEX deposits, the ore-forming fluids of which are also of low temperature and low-medium salinity [69].From the above, the FI data also support the submarine hydrothermal sediment (metasomatism) genesis of the first pulse of mineralization.
By comparison with the first pulse of mineralization, the second pulse of mineralization mainly contains the W-type FIs with few C-type, G-type, and S-type FIs.Meanwhile, according to laser Raman spectroscopy and micropetrography of FIs, the vapor and fluid phase of FIs are also dominated by H 2 O with some daughter minerals and minor CO 2 (Figures 8 and 10), and the FIs show a more variable higher homogenization temperature (Figures 9(c) and 9(e)), slightly lower salinities (Figures 9(d) and 9(f)), and approximate densities (Figure 14 and Table 2).These FI data reveal that the ore-forming fluids for the second pulse of mineralization are the low-medium temperature, low-salinity, and lowdensity H 2 O-NaCl system containing CO 2 .Comparing the Keyue deposit with other regional deposits in NHMB, the stage 3 and 4 minerals from the Zhaxikang deposit (Figure 14) show little higher homogenization temperature (212~338.0,concentrate on 220~310 °C), a little lower salinities (3.9~13.2wt%, concentrate on 6.0~10.0wt%), and similar densities (0.73~0.91 g/cm 3 ) [68], whereas the Shalagang Sb deposit that is dominated by the stibnitequartz ores has different FI characteristics (homogenization temperature: 134.9 °C~221.9°C, concentrate on 160~190 °C; salinities: 1.65~7.25 wt%, concentrate on 5.0~6.0 wt%) [68].These comparisons suggest that Sb mineralization for stibnite in the Shalagang deposit distinguishes from that for sulfosalt minerals in Keyue and Zhaxikang deposits.
As discussed in Section 5.1, there is a genetic relationship between the Keyue and Zhaxikang deposits, so when there is a lack of analytical data for minerals in the Keyue deposit, we can take the Zhaxikang deposit as an analogy.In the Zhaxikang deposit, from stages 3 and 4 to stages 5 and 6 (homogenization temperature: 210~317.0°C, concentrate on 230~270 °C; salinities: 2.4~9.9 wt%, concentrate on 3.0~9.9wt%), the homogenization temperature and salinities also show a decreasing trend and is more approximate to the Shalagang deposit.Meanwhile, the stibnite-quartz veins in Shalagang (ESR: 18.0 ± 1.8; zircon U-Pb: 23.6 ± 0.8) [1] and Zhaxikang (ESR: 18.3 ± 1.8 and 23.3 ± 2.3; 40 Ar/ 39 Ar: 17.9 ± 0.5) [1] deposits have similar mineralization ages.These FIs and geochronology data indicate that the stage 5 and 6 stibnite-cinnabar-quartz-calcite assemblages in Zhaxikang and Keyue deposits may share the same origin with the stibnite-quartz veins in Shalagang.Nevertheless, we come to a conclusion in the previous paragraph that Sb mineralization for stibnite in the Shalagang deposit distinguishes from that for sulfosalt minerals in Keyue and Zhaxikang deposits, so the second and third pulses of mineralization may have different Sb origins.
On the other hand, the FI characteristics of the second and third pulses of mineralization in Keyue and Zhaxikang deposits are similar with those of quartz-stibnite-sulfosalt vein systems in the Variscan orogen in northwest Germany that have salinities of 1~11 wt% and homogenization temperatures of 133~389 °C concentrated on 150~220 °C [13].Thus, the second and third pulses of mineralization of Keyue and Zhaxikang deposits may have similar genesis with deposits in the Variscan orogen in northwest Germany, which result from the overprinting of preexisting Pb-Zn veins by later mesothermal Sb veins related to later orogeny.
Combining with the fact that the regional orogenic Sb-Au deposits mainly formed during the syn-collision period (60~42 Ma) [1], the second pulse of mineralization in Keyue and Zhaxikang deposits must relate to the regional orogeny during the syn-collision period, yet the third pulse of mineralization is most likely associated with the magmatichydrothermal activity during the postcollision period (25 Ma to now) [1].Overall, the FI data indicate that the Keyue deposit is an epithermal deposit with three pulses of mineralization that may be associated to three regional mineralization events, which is in keeping with the inferences from geological, mineralogical, and element characteristics in Section 5.1.The calculated results also exhibit a decreasing trend in δ 18 O H2O values from stage 3 (2.86‰~11.34‰,average: 7.78‰, n = 7) to stage 4 (1.59‰~2.69‰,average: 2.29‰, n = 3).In the "δ 13 D V-SMOW versus δ 18 O H2O " plot, stage 3 and 4 quartz from the Keyue deposit is similar to stage 3 and 4 quartz and calcite from the Zhaxikang deposit (Figure 15), which suggests that there may be a genetic relationship of these two deposits again.However, the H-O isotopic characteristics of stage 3 and 4 quartz and calcite from the Keyue and Zhaxikang deposits are significantly different from stage 1 and 2 Mn-Fe carbonate, as well as stage 5 and 6 quartz from the Zhaxikang deposit (Figure 15).This difference further proves the three pulses of mineralization in Keyue and Zhaxikang deposits as mentioned above in Sections 5.1 and 5.2.

Multiple Isotopes
Based on the geochronology evidence, the Shalagang Sb [1], Mazhala Au-Sb (ESR: 24.2 ± 2.4) [1], Zhegu Sb-Au (ESR: 18.4 ± 1.8 and 17.6 ± 1.8) [1], and Langkazi Au deposits [1] are all formed by the third regional mineralization event that relates to the magmatic-hydrothermal activity during the postcollision period (25 Ma to now) [1].In the "δ 13 D V-SMOW versus δ 18 O H2O " plot, the H-O isotopic compositions of these deposits are also influenced by the mixture of formation and magmatic water in different ratios (Figure 15).However, the H-O isotopic compositions of stage 5 and 6 quartz from the Zhaxikang deposit are obviously influenced by the mixture of formation and geothermal water; this may relate to the ore-forming depth.The stage 5 and 6 stibnite-cinnabar-quartz-calcite assemblages formed near the earth's surface, so the magmatic water mixed with meteoric water to form the geothermal water.Meanwhile, the FI evidence shows that the second pulse of mineralization (stages 3 and 4) may relate to the regional orogeny during the syn-collision period, and the H-O isotopic characteristics of  25 Geofluids stage 3 and 4 quartz from the Keyue deposit exhibit the influence by mixture of formation and magmatic and geothermal water, which not only demonstrates the magmatic activity contribution to the regional orogenic mineralization event during the syn-collision period (the second pulse of mineralization) but also confirms the overprint by the later geothermal water during the postcollision period (the third pulse of mineralization).

S-Pb
Isotopes.There are mainly three sulfur sources in nature: (1) mantle sulfur: the δ 34 S values are 0 ± 3‰, (2) seawater sulfur: the δ 34 S values are around 20‰ (the values vary during different geological history periods), (3) the reduced sulfur in sediments: the δ 34 S values are usually relatively light [72].The stage 2 sulfides in the Keyue deposit have δ 34 S values of 5.4‰~11.2‰,which are similar to those of stage 2 sulfides in the Zhaxikang deposit (7.8‰~12.2‰)[3,9,18,73] and regional Ridang Formation (6.4‰~9.2‰; Figure 16(a)) [9].Thus, the sulfur source for the first pulse of mineralization in Keyue and Zhaxikang deposits is compatible with mainly seawater sulfur mixing with some mantle sulfur.Meanwhile, the S isotopic compositions of stage 2 sulfides in Keyue and Zhaxikang deposits overlap the range of many SEDEX-type deposits that all have relatively heavy δ 34 S values (Figure 16(b)).Nevertheless, the Red Dog (−37.2‰~−16.4‰),Sullivan, HYC, and Dugald River SEDEX-type deposits show much lighter δ 34 S values (Figure 16(b)), which may be due to the contribution of reduced sulfur in sediments and biogenic sulfur.Taking all of this evidence together, the first pulse of mineralization in the Keyue and Zhaxikang deposits could be related to the multiple seafloor volcanic events during the synsedimentary period (220~130 Ma) with the submarine hydrothermal sediment (metasomatism) genesis.Furthermore, the δ 34 S values of sulfides in Keyue and Zhaxikang deposits show a decreasing trend with time (from stage 2 through stages 3 and 4 to stage 5; Figure 16(a)) and exhibit more contribution from mantle sulfur associated with magmatism [74] in stages 3 to 5, which may relate to the overprint by later mineralization events.This inference is in keeping with the fact that the deposits (Shalagan, Mazhala, Zhegu, and Langkazi deposits) may have formed in relation with the third regional mineralization event which relates to the magmatic-hydrothermal activity during the postcollision period having much lighter δ 34   17), most of the data points fall into the domain of the upper crust, rarely plotting in the orogenic belt area in the tectonic setting discrimination diagrams (Figures 17(a) and 17(b)); similarly, all the data points are located around the upper crust and orogenic belt evolution curves in the Pb isotopic growth curve diagrams (Figures 17(c) and 17(d)).To be specific, the Pb isotopic compositions of stage 3 and 4 galena and boulangerite from the Keyue deposit are similar with those of stage 3 and 4 sulfides from the Zhaxikang deposit, which show that Pb is mainly sourced from the crust for the second pulse of mineralization in Keyue and Zhaxikang deposits.Meanwhile, most of the data points for stage 2 sphalerite from the Zhaxikang deposit fall into the Ocean island volcanic rock area (Figure 17(a)) and located below the orogenic belt curve (Figure 17(c)) in Pb isotopic diagrams, which proves the inference from S isotopes again that the first pulse of mineralization in Keyue and Zhaxikang deposits could be related to the multiple seafloor volcanic events during the synsedimentary period.Meanwhile, the Pb isotopic compositions of stage 5 stibnite in Keyue and Zhaxikang deposits are slightly different with those of earlier sulfides.Therefore, the Pb isotopic data also support the existence of three pulses of mineralization in Keyue and Zhaxikang deposits.Moreover, it is worth noting that sulfides share similar Pb isotopic characteristics with regional magmatic rocks and Ridang Formation, which also demonstrate the contribution from magmatism to the mineralization events in Keyue and Zhaxikang deposits.
In a word, the H-O-S-Pb isotopic evidence confirms the existence of three pulses of mineralization related to the three regional mineralization events again, which are in accord with the inferences from geology, mineralogy, elements, and FI evidence.
5.4.Ore-Forming Model.The ore-forming model synthesizes the geology, mineralogy, ore paragenetic sequence, elements, FIs, and H-O-S-Pb isotopes; the Keyue and Zhaxikang deposits have experienced three pulses of mineralization that may be associated with three regional mineralization events.The ore-forming model is as follows (Figure 18): (1) the first pulse of mineralization: during the synsedimentary period (220~130 Ma), under the influence of multiple seafloor volcanic events, the reducing environment in rifts and the adsorption of organic matter lead to the enrichment of metallogenic elements (e.g., Pb, Zn, Au, Ag, Mn, Fe, Ba, Ga, In, Tl, and B) in the Late Triassic-Early Cretaceous flysch formations formed by turbidity sediment and carbonaceous-siliceous-argillaceous rock series related to hydrothermal sedimentation, and the seafloor hydrothermal system circularly leached the metallogenic elements to form the first pulse of mineralization (stages 1 and 2) in Keyue and Zhaxikang deposits.(2) The second pulse of mineralization: during the syn-collision period (60~42 Ma) of the Indian-Eurasian plate, influenced by orogeny, the deep source mantle fluid formed by mantle exhaust and the CO 2 -bearing fluid formed by Lower crustal dehydration mixed and raised along the shear zone, with the contribution for ore-forming elements from magmatism to form a series of regional orogenic Au-Sb deposits.This mineralization event is related to the second pulse of mineralization (stages 3 and 4) and overprints the first pulse of mineralization in Keyue and Zhaxikang deposits.(3) Third pulse of mineralization: during the postcollision period (25 Ma to now), with the development of STDS, the metamorphic core complex, and the extensive emplacement of leucogranite, a series of W-Sn, Pb-Zn (Ag), and Sb-Au deposits developed in the region.This magmatic-hydrothermal mineralization event is associated with the third pulse of mineralization and also overprints the first and second pulses of mineralization in Keyue and Zhaxikang deposits.Moreover, the Plateau Uplift and complicated tectogenesis related to the collision of the Indian and Eurasian plates during the syn-collision and postcollision periods erode and destroy the early bedded Pb-Zn orebodies.

Conclusions
The geology, mineralogy, ore paragenetic sequence, elements, FIs, and H-O-S-Pb isotope evidence allow us to make the following conclusions: (1) There are six ore-forming assemblages of ore and gangue minerals that can represent six stages of ore  [18], and Zhang [73]), Shalagang, Mazhala, Zhegu, and Langkazi deposits (data cited from Yang et al. [1]) in NHMB.(b) The S isotopic compositions of stage 2 sulfide minerals from the Keyue and Zhaxikang Sb-Pb-Zn-Ag deposits; the S isotopic compositions of sulfide minerals from typical SEDEX deposits all over the world are also plotted for comparison (modified from Leach et al. [78]).
27 Geofluids formation from early to late, and these six stages are assigned to three clear pulses.
(2) The occurrence states of Ag in the Keyue deposit are prevailingly in the form of visible Ag that can be divided into independent Ag-bearing minerals and Ag-bearing inclusions.The invisible Ag is much less and principally in the form of submicroscopic Ag-bearing inclusions.
(3) The FIs in the Keyue deposit can be divided into three genetic (primary, secondary, and pseudosecondary)  (4) There is an increasing trend in δD V-SMOW values, a decreasing trend in δ 18 O V-SMOW values, and a decreasing trend in δ 34 S values from early to late stages.However, the Pb isotopic data have relatively uniform ranges.
(5) The Keyue deposit is an epithermal deposit and has mainly experienced three pulses of mineralization that relates to the three regional mineralization events in NHMB.
(6) There is a genetic relationship between the Keyue and Zhaxikang deposits, they have the same origins, and they experience the same mineralization events.

Figure 2 :
Figure 2: (a) Geological map of the Keyue Sb-Pb-Zn-Ag deposit (modified from Yang [10]); (b) cross-section along Exploration Line 44; (c) the schematic illustration of the internal structure of the Keyue deposits, with the vertical zonation, different alteration associations, and mineralization.

4. 1 . 2 .
Occurrence State of Ag in Ag-Bearing Minerals.The microscopic observation, EPMA, and back-scattering electron (BSE) microscopic research show that the main Agbearing minerals are tetrahedrite, freibergite, and andorite (Figures 4(n) and 6), which indicate that visible Ag is the dominating occurrence state of Ag in the Keyue deposit.

Figure 8 :( 3 )
Figure 8: Photomicrographs showing the types and characteristic features of fluid inclusions (FIs) at the Keyue Sb-Pb-Zn-Ag deposit.(a) Coexistence of the primary and secondary FIs that linearly distribute in assemblage 3 quartz.(b) The pseudosecondary FIs linearly distribute in assemblage 4 quartz.(c) The elliptic V-type FIs in assemblage 3 quartz.(d) The long-columnar C-type FIs in assemblage 3 quartz.(e) The irregular halite-bearing FIs in assemblage 4 quartz.(f) The elliptic G-type FIs in assemblage 3 quartz.(g) The rectangular L-type FIs in assemblage 1 Mn-Fe carbonate.(h) The rhombic L-type FIs in assemblage 2 Mn-Fe carbonate.(i) The L-type FIs isolatedly distribute in assemblage 5 quartz.(j) The long-columnar L-type FIs in assemblage 2 sphalerite.(k) The long-columnar L-type FI inclusions in assemblage 2 sphalerite.(l) The S-type FIs in assemblage 2 sphalerite.Abbreviations are the same as in Figure 5.

Figure 10 :
Figure 10: Representative laser Raman spectra of FIs in the Keyue Sb-Pb-Zn-Ag deposit.(a) CO 2 spectra of vapor phase in the W-type FIs of assemblage 2 Mn-Fe carbonate; (b) H 2 O spectra of liquid phase in the CO 2 -bearing W-type FIs of assemblage 2 Mn-Fe carbonate; (c) CO 2 spectra of vapor phase in the CO 2 -bearing W-type FIs of assemblage 3 quartz; (d) H 2 O spectra of liquid phase in the CO 2 -bearing W-type FIs of assemblage 3 quartz; (e) CO 2 spectra of vapor phase in the G-type FIs of assemblage 3 quartz; (f) CO 2 spectra of vapor phase in the C-type FIs of assemblage 3 quartz; (g) CO 2 spectra of vapor phase in the C-type FIs of assemblage 3 quartz; (h) H 2 O spectra of liquid phase in the C-type FIs of assemblage 3 quartz; (i) CO 2 spectra of vapor phase in the CO 2 -bearing W-type FIs of assemblage 4 quartz; (j) H 2 O spectra of liquid phase in the CO 2 -bearing W-type FIs of assemblage 4 quartz.
(a)), stages 2, 3, and 4 show the ore-forming pressure of 800~2000 bars, 500~1000 bars, and 300~1000 bars, respectively.To regard the average pressure gradient as 270 × 105 Pa/km, the calculated mineralization depths of stages 2, 3, and 4 are 3.0~7.4km, 1.9~5.6 km, and 1.1~3.7 km, respectively.The first pulse of mineralization has an obviously higher ore-forming pressure and deeper mineralization depth.From the above, the ore-forming fluids of the first and second pulses of mineralization show significantly different characteristics, which further support the existence of multiple pulses of mineralization in the Keyue deposit.Nevertheless, both the first and second pulses of mineralization belong to the epithermal category according to the calculated mineralization depths.

Figure 13 :
Figure 13: Diagram of homogenization temperature versus salinity for the fluid inclusions at the Keyue Sb-Pb-Zn-Ag deposit (modified from Wilkinson [75]).

Figure 14 :
Figure 14: (a) The NaCl-H 2 O system ρ-T-D diagram of the different stage fluids in the Keyue Sb-Pb-Zn-Ag deposit (modified from Chen et al. [67]).(b) The NaCl-H 2 O system ρ-T-D diagram of the different stage fluids in the Zhaxikang Sb-Pb-Zn-Ag deposit (modified from Yu [68]).
Constrain the Ore Genesis 5.3.1.H-O Isotopes.According to the H-O isotopic data, there is an increasing trend in δD V-SMOW values and a decreasing trend in δ 18 O V-SMOW values from early to late stages.Meanwhile, combining with the measured δ 18 O V-SMOW values and homogenization temperatures of FIs (T h−total ), the δ 18 O H2O values of ore-forming fluids are calculated by the theoretical equation of 1000 ln α quartz−H2O = 3 38 × 10 6 T −2 − 3 40 [70, 71].Due to the lack of the T h−total data of stage 3 calcite and stage 5 quartz, the calculations are only aimed at stage 3 and 4 quartz samples.

Figure 18 :
Figure 18: The ore-forming model for three pulses of mineralization in the Keyue deposit.

11 Geofluids Table 1 :
EPMA data for minerals from the Keyue deposit.

Table 2 :
Microthermometric results of fluid inclusions from minerals in the Keyue deposit.

Table 3 :
The H-O isotopic data for minerals from the Keyue deposit.

Table 4 :
The S isotopic data for sulfides from the Keyue deposit.

Table 5 :
The Pb isotopic data for sulfides from the Keyue deposit.
5.2.2.Comparison with Regional and Other Similar Deposits.For the Keyue deposit, minerals from the first pulse of mineralization mainly contain L-type FIs with few S-type S values.With regards to Pb isotopes, the Pb isotopic data of sulfides from the Keyue deposit have relatively uniform ranges, with 206 Pb/ 204 Pb ratios of 18.530~19.780, 207Pb/ 204 Pb ratios of 15.674~15.939,and 208 Pb/ 204 Pb ratios of 38.618~40.559.In the " 207 Pb/ 204 Pb versus 206 Pb/ 204 Pb" and " 208 Pb/ 204 Pb versus 206 Pb/ 204 Pb" diagrams (Figure