Source and Evolution of Ore-Forming Fluids in the Dulanggou Gold Deposit, Danba, Sichuan, China: Constraints from Fluid Inclusions and C–H–O Isotopes

Abstract

The Danba–Dadu River gold belt on the western Yangtze Craton margin is a major gold province in China. The Dulanggou gold deposit is a large quartz-vein-type deposit recently discovered in this belt. Ore bodies are fault-controlled veins hosted in high-grade metamorphic rocks of the Devonian Weiguan Formation. Mineralization includes three stages: early (quartz–minor sulfide), main (quartz–abundant sulfide–native gold–Te–Bi minerals), and late (quartz–minor sulfide–calcite). Fluid inclusion studies show the following. Early-stage inclusions are mainly CO₂–H₂O-type (homogenization temperature 307–388 °C, salinity 0.4–7.1 wt.% NaCl eqv.) with minor NaCl–H₂O-type. Main-stage inclusions are dominated by CO₂–H₂O and NaCl–H₂O types, with minor pure CO₂ inclusions (homogenization temperature 207–307 °C, salinity 0.2–11.2 wt.% NaCl eqv.). Late-stage inclusions are mainly NaCl–H₂O-type (168–223 °C, 4.6–10.1 wt.% NaCl eqv.). Laser Raman analysis detects CH₄ in the fluid. The ore-forming fluid is a reducing, medium–low temperature, low-salinity H₂O–CO₂–NaCl–CH₄ system. Thermodynamic calculations of CO₂–H₂O inclusions yield total densities of 0.94–1.03 g/cm³ and total homogenization pressures of 170–276 MPa for the early stage, and slightly lower densities (0.94–1.01 g/cm³) with pressures of 170–246 MPa for the main stage, indicating a progressive pressure decrease during fluid evolution. Hydrogen and oxygen isotopes (early stage: δD −96.4‰ to −78.9‰, δ¹⁸OH₂O 6.1‰ to 6.5‰; main stage: δD −104.3‰ to −75.1‰, δ¹⁸OH₂O 5.3‰ to 7.1‰) indicate that the ore-forming fluid was mainly derived from primary magmatic water. Immiscible CO₂–H₂O and NaCl–H₂O inclusion assemblages in the main stage suggest that fluid immiscibility was the key mechanism for gold precipitation. The Dulanggou deposit resembles classic orogenic gold deposits in host rocks, ore-controlling structures, mineral assemblages, and low-salinity CO₂-rich fluids. However, its H–O isotopes and thermodynamic data point to a magmatic water source, distinct from the metamorphic water source of typical orogenic gold deposits. This highlights the diversity of fluid sources in orogenic gold systems along the western Yangtze Craton margin.

1. Introduction

The Danba–Daduhe gold metallogenic belt, located along the western margin of the Yangtze Craton, is an important gold-producing region in southwestern China. The Dulanggou gold deposit in Danba, Sichuan Province, is a typical quartz-vein-type gold deposit newly identified in this belt in recent years. It contains proven gold resources of 48 t with an average grade of 6.71 g/t, and is therefore classified as a large high-grade gold deposit. Previous studies on this deposit have mainly focused on its geological characteristics, sources of ore-forming materials, mineralization age, and prospecting indicators [1,2,3,4,5,6], providing a solid basis for further investigation. However, key issues concerning the sources and evolution of ore-forming fluids and the mineralization mechanism remain controversial. In particular, the relative contributions of metamorphic fluids, anatectic magma, and mantle-derived fluids have not yet been systematically constrained.

The root cause of this controversy lies in the existence of two end-member models for the fluid sources of global orogenic gold deposits: the crustal metamorphic fluid model [7,8] and the subduction zone-mantle-derived fluid model [9]. The former emphasizes that the ore-forming fluids originate from metamorphic devolatilization during crustal thickening, whereas the latter argues that the fluids are mainly derived from subducting slab dehydration or from mantle source regions metasomatized by slab-derived fluids. Therefore, systematic fluid source tracing is a key approach to resolving this type of controversy.

Fluid inclusions are the most direct and best-preserved samples of ore-forming fluids trapped during mineral growth, and they provide important information on the physicochemical properties, sources, and evolution of hydrothermal fluids [10,11,12,13,14,15]. In addition, the C–H–O stable isotope system is an effective tool for tracing the sources and evolution of ore-forming fluids and for constraining mineralization processes [16].

Therefore, based on detailed field geological investigations, this study carried out fluid inclusion petrography, microthermometric analyses, and C–H–O isotope measurements on ores from different mineralization stages of the Dulanggou gold deposit. The main objectives are to systematically constrain the characteristics, evolution, and sources of the ore-forming fluids, and to further discuss the mechanism of gold precipitation in the context of the regional geological setting.

2. Geological Setting

The Dulanggou gold deposit in Danba is tectonically located at the junction of the Yangtze, North China, and Indian plates (Figure 1a) [1,4,17]. It lies on the northwestern side of the Y-shaped tectonic framework formed by the NW-trending Xianshuihe fault zone, the NE-trending Longmenshan fault zone, and the Sichuan–Yunnan tectonic belt along the southern segment of the western margin of the Yangtze Craton (Figure 1b) [1,4,17]. The northern branch of this Y-shaped structure corresponds to the Jintang arcuate structure within the Songpan–Ganzi fold belt.

The regional strata consist of a lower Neoproterozoic crystalline basement overlain by Paleozoic–Cenozoic sedimentary cover sequences. Lithologies of the cover are diverse and include Silurian, Devonian, and Permian marble, quartzite, quartz schist, schist, and gneiss, as well as Triassic sandstone, carbonate rocks, and flysch. Cambrian strata are completely absent in this area, and the Neoproterozoic crystalline basement is in unconformable contact with the overlying strata. Regional tectonic activity was intense and produced a variety of geological structures, including faults, domes, and folds, forming three principal fault sets trending NW–SE, NE–SW, and nearly N–S. Igneous rocks of different ages are exposed in the region. The Jinning–Chengjiang stage is represented mainly by plagiogranite, which commonly occurs as batholiths and is concentrated in the core of the Gezong dome. Regional metamorphism is strong and metamorphic rocks are widespread, with metamorphic grades mainly ranging from greenschist facies to amphibolite facies. In terms of metamorphic origin, rocks in the area mainly experienced regional metamorphism and dynamic metamorphism, whereas thermal contact metamorphism is relatively weak. The products of regional metamorphism are dominated by medium- to high-grade amphibolite-facies rocks that generally underwent recrystallization. The region is rich in mineral resources, including Au, Ag, Cu, Pb–Zn, and Mo deposits.

3. Deposit Setting

3.1. Orebody Characteristics

The Dulanggou gold deposit is exposed in the third (Dwg3) and fourth (Dwg4) members of the Devonian Weiguan Group (Figure 2). The third member (Dwg3) is the main exposed stratum in the study area and also the principal ore-hosting horizon in the mining district (Figure 2 and Figure 3). It is composed mainly of garnet–sillimanite two-mica quartz schist intercalated with quartzite, with dolomitic marble locally developed at the top. A total of 29 economic orebodies have been delineated in the mining district, all of which are quartz-vein-type orebodies (Figure 3a and Figure 4a). Four major orebodies are present. The orebodies are strictly controlled by structures. The ore-controlling faults are three secondary faults derived from the Huodi fault, namely F₁, F₂, and F₃ (Figure 2). Fault F1 strikes 325–360°, generally dips west to northwest, and has dip angles of 65–82°. Fault F2 strikes 336–30°, generally dips 221–275°, with dip angles of about 45–90°. Fault F3 is approximately parallel to F2, dipping 232–264° with dip angles of 50–90°. Among them, F2 is the principal ore-controlling fault. The orebodies mostly occur as veins, lenses, and en echelon bodies. They overall dip westward, with dip directions mostly between 230° and 310°, and are steeply dipping, generally at 70–90°.

3.2. Ore Characteristics and Wall–Rock Alteration

The non-metallic minerals in the ore are dominated by quartz. According to their paragenetic sequence, they are divided into early-stage hydrothermal quartz veins (QtzI) (Figure 4b), main-stage quartz veins (QtzII) (Figure 4b,c), and late-stage hydrothermal quartz veins (QtzIII) (Figure 4c). The metallic minerals in the ore mainly consist of three types: sulfides, gold minerals, and tellurium–bismuth minerals. The sulfides primarily include pyrrhotite, pyrite, and chalcopyrite (Figure 4b–e). Native element minerals are typically represented by native gold (Figure 4d,f) and native bismuth. Tellurium–bismuth minerals are diverse, mainly comprising a series of such minerals, including tetradymite.

In the Dulanggou gold mining district, wall–rock alteration is intense, but alteration zoning is indistinct. The wall rocks mainly develop silicification, biotitization, and hornblendization, overprinted by sericitization, with minor carbonatization and chloritization at the late stage.

3.3. Division of Hydrothermal Mineralization Stages

Based on the field occurrence of quartz veins, hand specimen textural and structural characteristics, and detailed microscopic mineralogical studies, the mineralization stages of the Danba Dulanggou gold deposit are divided into the hydrothermal mineralization stage and the supergene oxidation stage. According to the paragenetic relationships among minerals, the hydrothermal mineralization stage is further subdivided into three substages: the early-stage quartz–minor sulfide substage (I), the main-stage quartz–abundant sulfide–native gold–tellurium–bismuth mineral substage (II), and the late-stage quartz–minor sulfide–calcite substage (III).

Early hydrothermal ore stage (quartz + minor sulfides): this stage is characterized mainly by quartz together with minor pyrite, pyrrhotite, and trace chalcopyrite, whereas Te–Bi minerals are absent. Pyrite grains are generally coarse, mostly larger than 100 μm, and relatively euhedral. Early-stage pyrite can be observed to have been replaced by main-stage pyrrhotite, forming replacement-corrosion textures.

Main hydrothermal ore stage: this stage is characterized by abundant sulfides together with visible native gold and abundant Te–Bi minerals. Te–Bi minerals are diverse and include mainly pilsenite and joséite. The sulfides are represented mainly by pyrrhotite, chalcopyrite, and pyrite. The main stage can be further divided into early and late substages (Figure 5).

Late hydrothermal ore stage: this stage consists mainly of quartz with minor pyrite and calcite.

The supergene stage is manifested mainly by oxidation of pyrrhotite, pyrite, and other Fe-bearing minerals to limonite.

4. Sampling and Analytical Methods

Samples were systematically collected from different elevations and exploration lines in this study. Systematic sampling was carried out at different elevations along the No. 1, No. 7, and No. 9 main orebodies of the Dulanggou gold deposit, mainly along exploration line 0 (3830, 3930, 4075, 4113, and 4115 m), exploration line +5 (3830, 3930, and 4075 m), and exploration line 7 (3730, 3830, and 3930 m) (Figure 3). More than 300 quartz inclusion wafers were prepared for this work in order to conduct systematic research on ore-forming fluids. These samples were used for detailed inclusion petrography, followed by fluid inclusion microthermometry (heating–freezing stage) and laser Raman spectroscopy. In addition, single-mineral quartz separates were handpicked from representative ores for Carbon–Hydrogen–Oxygen isotope analysis.

Preliminary fluid inclusion petrography and microthermometric measurements were completed at Chengdu University of Technology using a Nikon 100POL polarizing microscope (Nikon Corporation, Tokyo, Japan) and a Linkam THM SG600 heating–freezing stage (Linkam Scientific, Redhill, UK). The microscope was equipped with objective lenses of 2×, 5×, 10×, 20×, 50×, and 100×. The heating–freezing stage uses a platinum resistance thermometer as the temperature sensor, with a working range of −196 to 600 °C, a display precision of 0.1 °C, and a temperature stability better than ±0.1 °C. During homogenization and freezing measurements, the heating/cooling rate was generally controlled between 0.2 and 5.0 °C/min; when approaching phase-transition points, the rate was reduced to 0.2 °C/min. To avoid mechanical deformation of inclusions during freezing, the experiments followed the sequence of measuring homogenization temperature (Th) first and ice-melting temperature (Tm-ice) afterward.

Laser Raman analyses of fluid inclusions were carried out at the Institute of Geochemistry, Chinese Academy of Sciences. A LabRAM HR Evolution micro-confocal laser Raman spectrometer (Horiba, Glasgow, UK) was used, with an excitation wavelength of 532 nm and a spectral resolution of ±0.65 cm⁻¹.

H–O isotope analyses were completed at Beijing Zhongke Kuangyan Testing Technology Co., Ltd. Quartz samples for analysis were crushed and sieved to a 40–60 mesh fraction, and high-purity single-mineral grains (>99.9%) were handpicked under a binocular stereomicroscope. For hydrogen isotope analysis of fluid inclusions, 10–20 mg of single-mineral quartz of 40–60 mesh grain size was weighed into tin sample cups and dried at 90 °C for 12 h in a constant-temperature oven to completely remove surface-adsorbed water. After drying, the tin-wrapped samples were introduced through an autosampler into a high-temperature pyrolysis furnace (Flash EA, Thermo Fisher Scientific, Waltham, MA, USA). At 1420 °C, water released from the inclusions reacted with glassy carbon filling material in the reactor to generate H₂ and CO. These gases were carried by high-purity helium (99.999%; 5N), separated by gas chromatography, and introduced into an isotope ratio mass spectrometer (253 Plus, Thermo Fisher Scientific, Waltham, MA, USA) for determination of the δD value of H₂. International polyethylene standard IAEA-CH-7 (δD_V-SMOW = −100.3‰) was used for quality control, and analytical precision was better than ±0.1‰. Detailed analytical procedures follow Gong et al. [18].

For oxygen isotope analysis of quartz, single-mineral quartz samples were ground to 200 mesh. About 6 mg of pure quartz or an equivalent amount of silicate mineral containing the same oxygen content was dried at 105 °C for 12 h. Oxygen was extracted using the conventional bromine pentafluoride (BrF5) method [19]: samples were reacted with BrF5 under vacuum at 580 °C to generate O₂, which was collected in 5 Å molecular sieve sample tubes and measured relative to the V-SMOW standard. Isotopic measurements were conducted on a 253 Plus gas isotope ratio mass spectrometer, and data reduction followed the procedures of Clayton et al. [20]. The δ¹⁸O_H₂O value of fluid inclusions in quartz was calculated according to the classical isotope fractionation equation, using the δ¹⁸O_quartz value of quartz and the corresponding homogenization temperature data (Equation (1)) [20,21].

1000lnαquartz-water = 3.38 × 10⁶/T² − 3.40

(1)

In the calculation, T is the homogenization temperature of quartz-hosted fluid inclusions.

The carbon isotopic composition of CO₂ in mineral fluid inclusions is an important tool for constraining the sources and evolution of ore-forming fluids. In this study, the δ¹³C and δ¹⁸O values of CO₂ entrapped in fluid inclusions hosted by single quartz separates were determined. All analyses were carried out by Beijing Zhongke Kuangyan Testing Technology Co., Ltd. using a MAT 253 stable isotope ratio mass spectrometer.

For carbon isotope analysis of fluid inclusions, approximately 0.1 g of single-mineral quartz separates with a grain size of 40–60 mesh were selected and dried in a constant-temperature oven at 105 °C for more than 4 h to remove volatile components adsorbed on the grain surfaces. The dried samples were then loaded into the reaction tube of an online continuous-flow extraction system. Prior to analysis, the elemental analyzer (Flash EA) and extraction lines were thoroughly purged with high-purity helium (He, 99.999%) to minimize the background CO₂ signal of the system. The samples were then decrepitated online at 550 °C for 5 min to ensure complete release of CO₂ from the fluid inclusions. The liberated gases were passed through a water trap filled with magnesium perchlorate [Mg(ClO₄)₂] to remove moisture, after which CO₂ was collected in a trapping tube packed with a Ni catalyst, whereas CH₄ and other gas components were flushed out of the system. When the temperature of the Flash EA reactor reached 960 °C and the baseline CO₂ signal had stabilized at below 10 mV, the collected CO₂ was introduced into the mass spectrometer for isotopic measurement. The analyses were calibrated against the national reference material GBW04417 (δ¹³CV-PDB = −6.06‰) to ensure analytical accuracy and inter-laboratory comparability.

5. Result

5.1. Petrographic Characteristics of Fluid Inclusions

This thesis presents a systematic fluid inclusion microthermometric study of quartz from the three mineralization stages of the Dulanggou gold deposit: the early-stage quartz–minor sulfide substage (I), the main-stage quartz–abundant sulfide–native gold–tellurium–bismuth mineral substage (II), and the late-stage quartz–calcite substage (III). According to the criteria for distinguishing primary and secondary inclusions proposed by Van den Kerkhof and Hein [22], most inclusions occur as isolated or densely clustered forms (Figure 6a,b,d,e,i), indicating that they are primary inclusions. Inclusions distributed along healed fractures are secondary (Figure 6c). Following the room-temperature phase classification scheme of Roedder [10], the observed primary inclusions are divided into the following two assemblage types:

CO₂–H₂O inclusions (Type C): This type is the most dominant and widely distributed fluid inclusion type in the Dulanggou gold deposit, accounting for the vast majority in all stages. The inclusions are mainly elliptical, negative crystalline, or irregular in shape, with long-axis lengths ranging from 5 to 25 μm, and the volume fraction of the CO₂ phase varies between 30% and 70%. Based on the phase state of CO₂, Type C inclusions are further divided into three subtypes: (1) C-1 three-phase inclusions: LH₂O + LCO₂ + VCO₂, showing a typical “double-eyelid” appearance (Figure 6d,e); (2) C-2 two-phase inclusions: LH₂O + VCO₂, in which the volumetric proportion of the CO₂ phase is 10%–35% (Figure 6f); when cooled to around −10 °C, liquid CO₂ appears around the gaseous CO₂; and (3) C-3 two-phase inclusions: LH₂O + LCO₂, in which the CO₂ phase proportion is relatively high, mostly 40%–70% (Figure 6g); during cooling (around −10 °C), gaseous CO₂ appears within the liquid CO₂.

CO₂ inclusions (PC-type): these are nearly pure CO₂ inclusions (Figure 6b,c,h), mainly elliptical, and secondarily negative-crystal or irregular in shape, with sizes not exceeding 15 μm. At room temperature, they commonly occur as pure liquid CO₂ and develop a gaseous CO₂ phase during cooling (−10 °C or lower). PC-type inclusions may occur alone or coexist with C-type inclusions within the same field of view (Figure 6h).

NaCl–H₂O inclusions (W-type): these are saline aqueous inclusions, mostly elongate or elliptical in shape, with long axes ranging from 3 to 12 μm and gas-phase H₂O proportions generally not exceeding 25%. W-type inclusions may occur alone (Figure 6i) or coexist with C-type inclusions in the same field of view (Figure 7c).

In quartz grains of the early ore stage, the dominant inclusion type is C-type (CO₂–H₂O), followed by W-type (NaCl–H₂O). Most CO₂-bearing inclusions occur as tightly associated small clusters or isolated inclusions and are negative-crystal or elliptical in shape, with sizes of 10–25 μm (Figure 7a). A small number of saline aqueous inclusions are typically elliptical, occur in clusters, and are 10–20 μm in size (Figure 7b). In the same microscopic field of view, C-type (CO₂–H₂O) inclusions coexist with W-type (NaCl–H₂O) inclusions (Figure 7c). According to the definition of fluid inclusion assemblages (FIA) proposed by Goldstein and Reynolds [23], these two types are interpreted as contemporaneously trapped products formed during fluid immiscibility in the evolution of the same ore-forming fluid, thus constituting a typical immiscible fluid inclusion assemblage (Figure 7c).

In quartz grains of the main ore stage, C-type (CO₂–H₂O), PC-type (CO₂), and W-type (NaCl–H₂O) inclusions are all present. C-type and W-type inclusions are the principal types. They are commonly negative-crystal to elliptical in shape, 10–20 μm in size, and occur as tightly associated small clusters within the same field of view (Figure 7e–g). PC-type inclusions are elliptical to subrounded, 5–10 μm in size, and occur either isolated or in clusters, commonly together with C-type inclusions in the same field of view (Figure 7h). The coexistence of these three inclusion types within a single field of view indicates that they were trapped contemporaneously from the same fluid and together define an immiscible fluid inclusion assemblage [23].

In the late ore stage, W-type inclusions dominate, and C-type and PC-type inclusions are absent. Most W-type inclusions are elliptical and 8–15 μm in size. They occur either linearly along healed fractures in quartz grains or as small clusters within grains (Figure 7j–l). The observed inclusions show highly uniform vapor/liquid ratios. According to the criteria for recognizing fluid inclusion assemblages proposed by Goldstein and Reynolds [23], this feature indicates that these inclusions were trapped simultaneously from the same pulse of ore-forming fluid.

5.2. Characteristics of Homogenization Temperatures and Salinities of Fluid Inclusions

Based on petrographic observations, representative fluid inclusions in quartz from the early, main, and late ore stages of the Dulanggou gold deposit were selected for microthermometric analysis. The analytical results are listed in

Table 1. Microthermometric data of fluid inclusions from the three mineralization stages.

Stage	Type	Host Mineral	n	Tm, h (°C)	Tm, Ice (°C)	Salinity (wt.% NaCl eq.)	Pressure (MPa)
I	C	Qtz	36	307~388	−7.6~−1.4	0.4~5.9	171~276
I	W	Qtz	10	313~366	-	3.1~7.1	230~296
II	C	Qtz	55	220~307	−7.6~−0.1	0.2~3.1	170~246
II	W	Qtz	66	207~306	−6.7~−2.8	0.2~11.2	166~326
III	W	Qtz	24	168~234	−7.6~−1.4	4.6~10.1	144 ~218

. The distributions of homogenization temperatures and salinities for inclusions from each stage are shown in the histograms (Figure 8) and scatter plot (Figure 9).

Studies of fluid inclusions from different mineralization stages of the Dulanggou gold deposit, combined with salinity calculations using the corresponding empirical equations, show the following characteristics of homogenization temperature and salinity in the ore-forming fluids:

(1) Early ore stage (Stage I)

Both C-type (CO₂–H₂O) and W-type (NaCl–H₂O) inclusions occur in this stage. The partial homogenization temperatures of C-type inclusions range from 11.8 to 27.3 °C, whereas complete homogenization temperatures range from 307 to 388 °C, with a mean of 341 °C (Figure 8a). The initial melting temperatures of solid CO₂ are −58.9 to −56.8 °C, and clathrate melting temperatures range from 6.9 to 9.8 °C, with a mean of 8.9 °C. Salinity of this inclusion type must be calculated from clathrate dissociation temperatures using the following empirical equation (Equation (2)) [7]:

WNaCl = 15.52022 − 1.02342 × T − 0.05286 × T2

(2)

where WNaCl is the mass fraction of NaCl in the aqueous solution, and T is the clathrate melting temperature in CO₂–H₂O inclusions. This equation is applicable only when −9.6 °C ≤ T ≤ 10 °C. Calculated salinities of the Stage I C-type inclusions range from 0.4 to 5.9 wt.% NaCl eqv. (Figure 8b).

The coexisting W-type inclusions have homogenization temperatures mainly ranging from 313 to 366 °C, with a mean of 326 °C (Figure 8a). Their salinities were calculated from ice-melting temperatures using the following equation (Equation (3)):

$$\mathrm{S}=0.00+1.78T_m-0.0442T_m^2+0.000557T_m^3\left(\mathrm{for}\mathrm{NaCl}\mathrm{solutions}\mathrm{with}\mathrm{salinities}\mathrm{of}0\text{–}23.3\mathrm{wt}.\%\right)$$

(3)

where S is salinity (wt.% NaCl eqv.), and Tm is the ice-melting temperature (°C).

This equation is applicable to NaCl solutions with salinities of 0–23.3 wt.%. The measured ice-melting temperatures of W-type inclusions in this stage are −4.5 to −3.2 °C, corresponding to calculated salinities of 3.1–7.2 wt.% NaCl eqv. (Figure 8b). The overall mean homogenization temperature of early-stage inclusions (C-type + W-type) is 341 °C.

(2) Main ore stage (Stage II)

This stage also contains C-type (CO₂–H₂O) and W-type (NaCl–H₂O) inclusions. The partial homogenization temperatures of C-type inclusions are 13.1–26.7 °C, and complete homogenization temperatures range from 220 to 307 °C, with a mean of 279 °C (Figure 8c). Initial melting temperatures of solid CO₂ are −58.6 to −56.7 °C, and clathrate melting temperatures range from 8.4 to 9.9 °C. Using Equation (1), calculated salinities range from 0.2 to 3.2 wt.% NaCl eqv., with a peak value of 1.5 wt.% NaCl eqv. (Figure 8d).

The coexisting liquid-rich W-type inclusions have homogenization temperatures of 207–306 °C, with a mean of 269 °C (Figure 8c). Their salinities were calculated using Equation (2) based on ice-melting temperatures. Measured ice-melting temperatures range from −7.6 to −0.1 °C, corresponding to salinities of 0.2–11.2 wt.% NaCl eqv. (Figure 8d). The overall mean homogenization temperature of inclusions from the main stage is 274 °C.

(3) Late ore stage (Stage III)

This stage is dominated by W-type inclusions (NaCl–H₂O), whose homogenization temperatures range from 168 to 223 °C, with a mean of 199 °C (Figure 8e). Measured ice-melting temperatures range from −6.7 to −2.8 °C, and calculated fluid salinities based on Equation (2) range from 4.6 to 10.1 wt.% NaCl eqv. (Figure 8f).

5.3. Inclusion Density and Pressure

In the Dulanggou gold deposit, the dominant fluid inclusion type developed in different mineralization stages is the CO₂–H₂O type, and some CO₂–H₂O inclusions coexist with NaCl–H₂O inclusions. The specific steps for calculating the homogenization pressure of the CO₂–H₂O inclusions are as follows:

(1) Calculate the CO₂ phase density at partial homogenization of the CO₂–H₂O inclusions [24,25]:

Density of liquid CO₂ (Equation (4)):

$$d_{\mathrm{C}\mathrm{O}2\left(liquid\right)}=0.4683+0.001442\times \left(31.35-t_c\right)+0.1318\times {\left(31.35-t_c\right)}^{{\displaystyle \frac{1}{3}}}$$

(4)

Density of vapor CO₂ (Equation (5)):

$$d_{\mathrm{C}\mathrm{O}2\left(vapor\right)}=0.4683+0.001442\times \left(31.35-t_c\right)-0.1318\times {\left(31.35-t_c\right)}^{{\displaystyle \frac{1}{3}}}$$

(5)

where dCO₂(liquid) is the density (g/cm³) of liquid CO₂ when vapor CO₂ gradually homogenizes into liquid CO₂ upon partial homogenization of the CO₂–H₂O inclusions; dCO₂(vapor) is the density (g/cm³) of vapor CO₂ when liquid CO₂ gradually homogenizes into vapor CO₂ upon partial homogenization; and t_c is the partial homogenization temperature (°C) of the CO₂ phase upon partial homogenization of the CO₂–H₂O inclusions.

(2) Calculate the density of liquid water upon partial homogenization of the CO₂–H₂O inclusions [25,26] (Equations (6) and (7)):

$$m=(1000\times w)/[58.4428\times \left(100-w\right)]$$

(6)

$$\begin{gathered} d_{aq}=0.999839\times \left(1000+58.4428\times m\right)÷\left[1000+0.999839\times \left(12.43\times m+3.07\times m^{1.5}-0.02\times m^2\right)\right]+ \\ 5.2777\times {10}^{-5}\times t_c-1.0113\times {10}^{-5}\times t_c^2+9.3537\times {10}^{-8}\times t_c^3 \end{gathered}$$

(7)

where m is the molality of NaCl in the liquid water upon partial homogenization (mol/kg H₂O); w is the salinity of the liquid water (wt.% NaCl equiv); and d_aq is the density of the liquid water upon partial homogenization (g/cm³).

(3) Calculate the mole fraction in the liquid water upon partial homogenization of the CO₂–H₂O inclusions [27] (Equations (8)–(10)):

$$S_{\mathrm{C}{\mathrm{O}}_2\left(aq\right)}=7.2754-0.056634\times t_c+7.877\times 10^{-4}\times t_c^2-2.037\times 10^{-5}\times t_c^3$$

(8)

$$w_{\mathrm{C}{\mathrm{O}}_2\left(aq\right)}=S_{\mathrm{C}{\mathrm{O}}_2\left(aq\right)}÷\left(S_{\mathrm{C}{\mathrm{O}}_2\left(aq\right)}+100\right)\times 100$$

(9)

$$X_{\mathrm{C}{\mathrm{O}}_2\left(aq\right)}=18.0152\times W_{\mathrm{C}{\mathrm{O}}_2\left(aq\right)}÷[18.0152\times W_{\mathrm{C}{\mathrm{O}}_2\left(aq\right)}+44.0098\times (100-W_{\mathrm{C}{\mathrm{O}}_2\left(aq\right)}\left)\right]$$

(10)

where $S_{\mathrm{C}{\mathrm{O}}_2\left(aq\right)}$ is the solubility of CO₂ in the liquid water upon partial homogenization (g); $w_{\mathrm{C}{\mathrm{O}}_2\left(aq\right)}$ is the mass fraction of CO₂ in the liquid water upon partial homogenization (g/100 g); and $X_{\mathrm{C}{\mathrm{O}}_2\left(aq\right)}$ is the mole fraction of CO₂ in the liquid water upon partial homogenization.

(4) Calculate the mole fraction of CO₂ (XCO₂) and the mole fraction of NaCl (XNaCl) in the CO₂–H₂O inclusions, as well as the molar volume of the inclusions (V) [28] (Equation (11)–(19)):

$$X_{\mathrm{C}{\mathrm{O}}_2}={\displaystyle \frac{m_{\mathrm{C}{\mathrm{O}}_2}}{m_{\mathrm{C}{\mathrm{O}}_2}+m_{\mathrm{N}\mathrm{a}\mathrm{C}\mathrm{l}}+m_{{\mathrm{H}}_2\mathrm{O}}}}$$

(11)

$$X_{\mathrm{N}\mathrm{a}\mathrm{C}\mathrm{l}}={\displaystyle \frac{m_{\mathrm{N}\mathrm{a}\mathrm{C}\mathrm{l}}}{m_{\mathrm{C}{\mathrm{O}}_2}+m_{\mathrm{N}\mathrm{a}\mathrm{C}\mathrm{l}}+m_{{\mathrm{H}}_2\mathrm{O}}}}$$

(12)

$$V={\displaystyle \frac{1}{m_{\mathrm{C}{\mathrm{O}}_2}+m_{\mathrm{N}\mathrm{a}\mathrm{C}\mathrm{l}}+m_{{\mathrm{H}}_2\mathrm{O}}}}$$

(13)

$$m_{\mathrm{C}{\mathrm{O}}_2}=m_{\mathrm{C}{\mathrm{O}}_2\left(\mathrm{C}{\mathrm{O}}_2\right)}+m_{\mathrm{C}{\mathrm{O}}_2\left(aq\right)}$$

(14)

$$m_{\mathrm{C}{\mathrm{O}}_2\left(\mathrm{C}{\mathrm{O}}_2\right)}=d_{\mathrm{C}{\mathrm{O}}_2}\times F_{\mathrm{C}{\mathrm{O}}_2}÷44.0098$$

(15)

$$\begin{gathered} m_{\mathrm{C}{\mathrm{O}}_2\left(aq\right)}=d_{aq}\times (1-{\mathrm{F}}_{\mathrm{C}{\mathrm{O}}_2})÷44.0098\times (1-w)÷18.0152\times K\times X_{\mathrm{C}{\mathrm{O}}_2\left(aq\right)}\times 44.0098÷ \\ \left[1+(1-w)÷18.0152\times K\times X_{\mathrm{C}{\mathrm{O}}_2\left(aq\right)}\times 44.0098\right] \end{gathered}$$

(16)

$$m_{{\mathrm{H}}_2\mathrm{O}}=d_{aq}\times (1-F_{\mathrm{C}{\mathrm{O}}_2})÷18.0152\times (1-w)÷\left[1+(1-w)÷18.0152\times K\times X_{\mathrm{C}{\mathrm{O}}_2\left(aq\right)}\times 44.0098\right]$$

(17)

$$m_{\mathrm{N}\mathrm{a}\mathrm{C}\mathrm{l}}=d_{aq}\times (1-F_{{\mathrm{C}\mathrm{O}}_2})÷58.4428\times w÷[1+(1-w)÷18.0152\times K\times X_{{\mathrm{C}\mathrm{O}}_2\left(aq\right)}\times 44.0098]$$

(18)

$$K=1.0-(11.0÷3.0)\times w$$

(19)

where m_CO2, m_NaCl, and m_H2O are the numbers of moles of CO₂, NaCl, and H₂O, respectively; $m_{\mathrm{C}{\mathrm{O}}_2\left(\mathrm{C}{\mathrm{O}}_2\right)}$ and $m_{\mathrm{C}{\mathrm{O}}_2\left(aq\right)}$ are the numbers of moles of CO₂ dissolved per-unit volume in the CO₂ phase and in the aqueous phase upon partial homogenization, respectively; F_CO2 is the filling degree of the CO₂ phase upon partial homogenization; and K is the correction coefficient (0 ≤ K ≤ 10 ≤ K ≤ 1).

(5) Calculate the homogenization pressure of the CO₂–H₂O inclusions upon complete homogenization [26] (Equations (20) and (21)):

$$T=Th+273.15$$

(20)

$$P=83.143\times {\displaystyle \frac{T}{V-b}}-{\displaystyle \frac{\mathrm{a}}{\left[T^{0.5}\times V\times \left(V+b\right)\right]}}$$

(21)

where Th is the complete homogenization temperature (°C); P is the homogenization pressure upon complete homogenization of the CO₂–H₂O inclusions (bar); a and b are calculation parameters.

Fluid inclusion thermodynamic calculations show that the thermodynamic parameters of CO₂–H₂O-type fluid inclusions in the Dulanggou gold deposit exhibit regular changes with respect to mineralization stages. A total of 30 liquid-rich CO₂–H₂O inclusions were measured in stage I, and 37 in stage II (

Table 2. Thermodynamic calculation results of CO₂–H₂O inclusions in the Dulanggou gold deposit.

Mineralization Stage	d/(g/cm3)	V/(cm3·mol−1)	X (CO2)	FCO2/%	P (MPa)
I (n = 30)	0.94~1.03	19.14~28.14	0.04~0.26	0.05~0.50	171~276
II (n = 37)	0.94~1.01	19.16~28.14	0.04~0.26	0.05~0.50	170~246

Note: d is the density of the fluid upon inclusion homogenization (g·cm⁻³); V is the molar volume of the inclusion (cm³·mol⁻¹); X(CO2) is the mole fraction of CO₂; FCO₂ is the calculated filling degree of CO₂ (%); Th is the complete homogenization temperature of the inclusion; P is the complete homogenization pressure of the inclusion (MPa).

). For stage I, the total density (d) of liquid-rich CO₂–H₂O inclusions ranges from 0.94 to 1.03 g/cm³, the molar volume (V) ranges from 19.14 to 28.14 cm³·mol⁻¹, the CO₂ mole fraction (X(CO₂)) ranges from 0.04 to 0.26, the CO₂ filling degree (FCO₂) ranges from 5% to 50%, and the complete homogenization pressure (P) ranges from 171.6 to 275.7 MPa. For stage II, the total density is slightly lower than that of stage I, ranging from 0.94 to 1.01 g/cm³; the molar volume is similar to that of stage I, ranging from 19.16 to 28.14 cm³·mol⁻¹; the CO₂ mole fraction and CO₂ filling degree are consistent with those of stage I (0.04–0.26 and 5%–50%, respectively); whereas the complete homogenization pressure is generally lower than that of stage I, ranging from 169.7 to 245.9 MPa. The inclusions in both stages exhibit typical thermodynamic characteristics of the CO₂–H₂O binary fluid system, and the differences in parameter distribution reflect changes in the physicochemical conditions of the ore-forming fluid during different evolution stages.

The density calculation of NaCl–H₂O inclusions in different mineralization stages of the Dulanggou gold deposit is based on previous empirical formulas [27,29,30] (Equation (22)):

$$\rho =A+Bt+C{t}^2(\text{Salinity }S=0\text{–}3030\text{ wt}.\%)$$

(22)

where ρ is the density of the saline solution (g/cm³), and t is the homogenization temperature (°C). The applicable conditions of the above formula are:

Homogenization temperature ≤ 500 °C, where A (Equation (23)), B (Equation (24)), and C (Equation (25)) are all functions of salinity:

$$\mathrm{A}=0.993531+8.72147\times 10^{-3}\times \mathrm{S}-2.43975\times 10^{-5}\times {\mathrm{S}}^2$$

(23)

$$\mathrm{B}=7.11652\times 10^{-5}-5.2208\times 10^{-5}\times \mathrm{S}+1.26656\times 10^{-6}\times {\mathrm{S}}^2$$

(24)

$$\mathrm{C}=-3.4997\times 10^{-6}+2.12124\times 10^{-7}\times \mathrm{S}-4.52318\times 10^{-9}\times {\mathrm{S}}^2$$

(25)

Calculated results show that the densities of NaCl–H₂O inclusions in the stage I mineralization of the Dulanggou gold deposit range from 0.60 to 0.75 g/cm³, with an average of 0.71 g/cm³; and those in the stage II mineralization range from 0.69 to 0.96 g/cm³, with an average of 0.84 g/cm³. The homogenization pressure of these NaCl–H₂O inclusions can be roughly estimated using the previous empirical formula [31].

The formula for the homogenization pressure of the NaCl–H₂O fluid inclusion system [31] is as follows (Equations (26)–(28)):

$$P_1=P_0\times T_1/T_0$$

(26)

$$P_0=219+2620\times S$$

(27)

$$T_0=374+920\times S$$

(28)

where P₁ is the homogenization pressure, P₀ is the initial pressure, T₁ is the ore-forming temperature (the homogenization temperature of immiscible inclusions is approximately equal to this temperature), T₀ is the initial temperature, and S is the salinity.

Calculation results show that for the stage I mineralization of the Dulanggou gold deposit, the homogenization pressures of NaCl–H₂O inclusions range from 230 to 296 MPa, with an average of 273 MPa; for the stage II mineralization, the homogenization pressures range from 166 to 326 MPa, with an average of 215 MPa.

5.4. Laser Raman Spectroscopic Characteristics of Fluid Inclusions

Laser Raman spectroscopy was performed on typical fluid inclusions from different mineralization stages in the Dulanggou gold deposit. The results show the following. In the early mineralization stage, the gas phase of W-type inclusions is mainly H₂O (Figure 10a). Correspondingly, the C-type CO₂–H₂O inclusions from the same stage exhibit distinct CO₂ Raman peaks at 1285 cm⁻¹ and 1387 cm⁻¹ (Figure 10b). In the main mineralization stage, C-type CO₂–H₂O inclusions show clear CO₂ peaks at 1285 cm⁻¹ and 1387 cm⁻¹, as well as a clear CH₄ peak near 2900 cm⁻¹ (Figure 10c–e). In the late mineralization stage, the gas phase of the liquid-rich aqueous inclusions is dominated by H₂O, with no CO₂ or CH₄ peaks observed (Figure 10f). These results indicate that the ore-forming fluids in the early and main mineralization stages of the Dulanggou gold deposit are CO₂-dominated and contain CH₄ as a reducing gas component. In contrast, the fluids in the late mineralization stage are H₂O-dominated. In summary, the ore-forming fluid of the Dulanggou gold deposit is a reduced CO₂–H₂O–NaCl–CH₄ system.

5.5. H–O Isotopes

Hydrogen and oxygen isotopic compositions were determined for quartz from the early and main ore stages of the Dulanggou gold deposit. The results show that, for fluids in quartz from the early ore stage, δD_V-SMOW values range from −96.4‰ to −78.9‰, with a mean of −85.2‰, and δ¹⁸O_V-SMOW values range from 11.6‰ to 13.6‰, with a mean of 12.4‰. Based on homogenization temperature data from fluid inclusions and using the quartz–water oxygen isotope fractionation equation of Clayton et al. [20], the calculated δ¹⁸O_H2O-V-SMOW values of fluids in equilibrium with early-stage quartz range from 6.1‰ to 6.5‰, with a mean of 6.3‰. For quartz from the main ore stage, δD_V-SMOW values range from −104.3‰ to −75.1‰, with a mean of −88.4‰, and δ¹⁸O_V-SMOW values range from 12.4‰ to 14.2‰, with a mean of 13.4‰. The calculated δ¹⁸O__H2O-V-SMOW values of fluids in equilibrium with main-stage quartz, based on the quartz–water isotope fractionation equation, range from 5.3‰ to 7.1‰, with a mean of 6.3‰ (

Table 3. H–O isotopic compositions of gangue minerals, Dulanggou gold deposit (‰, SMOW).

Deposit	Sample	Host Mineral	Stage	th/°C	δ18OV-SMOW/‰	δ18OH2O/‰	δDV-SMOW/‰	Reference
Dulanggou	4150-1-007	Qtz	Early	339.4	13.4	6.19	−78.9	This study
	4150-1-008				13.7	6.53	−80.4
	3830-16-14				13.6	6.43	−96.4
	3830pm-2	Qtz	Main	269.6	14.0	6.91	−75.1	This study
	4075pm16-4			269.6	13.9	6.83	−104.3
	3830j-1-2			269.6	12.8	5.75	−76.7
	3930pml-8			269.6	12.4	5.32	−85.1
	3830-1-7			269.6	14.2	7.13	−80.5
	4150-1-13			269.6	13.5	6.45	−70.6
	3730PM14-2			269.6	14.0	6.95	−91.3
	4113-1			269.6	13.6	6.55	−79.9
Pianyanzi	P-4	Qtz	/	/	12.0	2.66	−61.2	[32]
	W-15		/	/	13.9	4.80	−59.6
Xiaoshandun	/	Grt	/	/	/	8.32	−82.94	[33]
Basement	H1	Qtz	/	/	11.0	6.37	−89.434	[34]
	H2		/	/	11.6	5.91	−49.887
	H3		/	/	12.1	5.32	−67.079
Cover	Df-1	Qtz	/	/	7.9	0.04	−110	[35]
	Df-2		/	/	15.8	7.94	−103
Jintaizi	/	Qtz	/	/	9.5	0.26	−99	[33]
	/		/	/	9.5	−0.32	−99
	/		/	/	9.4	0.84	−92
	/		/	/	9.4	0.25	−92
Heijintaizi	HJ-1	Qtz	/	/	11.1	1.14	−90.4	[36]
	HJ-8		/	/	11.0	1.04	−53.78
	HJ-9		/	/	10.3	0.34	−68.06
	HJ-10		/	/	10.7	0.74	−57.76
	HJ-16		/	/	11.9	1.94	−44.07
Baijintaizi	BJ-1	Qtz	/	/	7.9	−2.61	−47.79	[36]
	BJ-2		/	/	10.6	0.09	−49.73
	BJ-3		/	/	12.3	1.79	−39.13
	BJ-4		/	/	14.7	4.19	−54.78
Huangjinping	HJP-3	Qtz	/	/	12.0	3.04	−52.61	[36]
	HJP-14		/	/	12.5	3.54	−108.23
	HJP-20		/	/	12.6	3.64	−63.59
Yanzigou	B10	Qtz	III	/	/	5.07	−46.78	[37]
	B13		III	/	/	3.95	−48.44
	B14		I	/	/	9.28	−40.32
	B15		I	/	/	9.95	−43.02
	B41		II	/	/	6.97	−42.86
	B46		II	/	/	6.12	−43.55
Xinjintaizi	/	Qtz	II	/	/	8.89	−66.5	[38]
	/		II	/	/	8.49	−68.3
	/		II	/	/	7.29	−63.5
	/		II	/	/	10.89	−67.7
	/		V	/	/	9.16	−85.5
Lianhuataizi	/	Qtz	II	/	/	9.81	−66.8	[38]
	/		II	/	/	10.61	−64.8
	/		II	/	/	9.01	−68.3

Equation for calculating δ18O_H2O: 1000lnαquartz-water = 3.38 × 106/(273.15 + th)2 − 3.40 [20]. (Base diagram after Clayton et al. [20]; data for other deposits from Luo et al. [32], Ge and Chen [33], Chen et al. [34], Teng et al. [35], Li et al. [36], Hou et al. [37], and Li HB et al. [38]).

).

5.6. C–O Isotopes

Carbon and oxygen isotope analyses were carried out on quartz-hosted fluid inclusions from the early mineralization stage and the main mineralization stage of the Dulanggou Te–Bi-rich gold deposit. The analytical results (

Table 4. Summary of C-H-O isotopic compositions of the Dulanggou gold deposit, Danba, Sichuan Province.

Sample	Host Mineral	Stage	th/°C	δ18OV-SMOW/‰	δ18OH2O/‰	δDV-SMOW/‰
4150-1-007	Qtz	Early	339.4	13.4	6.19	−78.9
4150-1-008				13.7	6.53	−80.4
3830-16-14				13.6	6.43	−96.4
3830-1-7	Qtz	Main	269.6	14.0	6.91	−75.1
3830-1-6			269.6	13.9	6.83	−104.3
4150-1-13			269.6	12.8	5.75	−76.7
3730PM14-2			269.6	12.4	5.32	−85.1
4113-1			269.6	14.2	7.13	−80.5
4113-2			269.6	13.5	6.45	−70.6

) show that quartz from the early mineralization stage yields δ¹³C_V-PDB values ranging from −7.7‰ to −5.9‰, with an average of −6.9‰, and δ¹⁸O_V-SOMW values ranging from 13.4‰ to 13.6‰, with an average of 13.5‰. The calculated δ¹⁸O_H2O-V-SMOW values of the fluid in equilibrium with early-stage quartz range from 6.1‰ to 6.5‰, with an average of 6.3‰. Quartz from the main mineralization stage yields δ¹³C_V-PDB values ranging from −8.7‰ to −5.8‰, with an average of −7.3‰, and δ¹⁸O_V-SOMW values ranging from 13.5‰ to 14.3‰, with an average of 13.9‰. The calculated δ¹⁸O_H2O-V-SMOW values of the fluid in equilibrium with main-stage quartz range from 6.4‰ to 7.2‰, with an average of 6.8‰.

6. Discussion

6.1. Characteristics and Evolution of Ore-Forming Fluids

Detailed petrographic analysis of the three stages shows the following. In the early stage, fluid inclusions are mainly CO₂–H₂O type, with total homogenization temperatures ranging from 307 to 388 °C. NaCl–H₂O type inclusions are subordinate, with homogenization temperatures mainly from 313 to 366 °C. The combined homogenization temperature range is 307–388 °C, and salinities range from 0.4 to 7.1 wt.% NaCl eqv. The early-stage ore fluid is characterized by medium–high temperature, low salinity, and CO₂–H₂O richness.

In the main stage, the inclusions are predominantly CO₂–H₂O and NaCl–H₂O types, with minor pure CO₂ inclusions. The CO₂–H₂O inclusions have total homogenization temperatures of 220–307 °C, and NaCl–H₂O inclusions have temperatures of 313–366 °C. Salinities range from 0.2 to 11.2 wt.% NaCl eqv. The CO₂-rich feature of the early and main stages is similar to that of most orogenic gold deposits [39,40]. However, laser Raman analysis shows the presence of CH₄ in the fluid, and the relative CO₂ content is slightly lower than that in typical orogenic gold deposits, indicating a certain uniqueness of the fluid system.

In the late stage, inclusions are mainly NaCl–H₂O type, with homogenization temperatures of 168–234 °C (average 199 °C) and salinities of 4.6–10.1 wt.% NaCl eqv. Overall, from the early to the late stage, the ore-forming fluid shows a gradual decrease in temperature and an increase in salinity, indicating that fluid evolution is related to immiscibility.

Laser Raman spectroscopy of fluid inclusions reveals that in the early and main stages, the fluid is dominated by CO₂, with the presence of CH₄ as a reducing gas component. In the late stage, the fluid is dominated by H₂O, also with CH₄. In summary, the ore-forming fluid of the Dulanggou gold deposit is a reducing, medium–high temperature, low salinity H₂O–CO₂–NaCl–CH₄ system.

The CO₂–H₂O and NaCl–H₂O inclusions in the early and main stages have broadly similar homogenization temperatures, indicating fluid immiscibility [41]. This is consistent with the petrographic observation that these three types of inclusions occur together in the same field of view, forming an immiscible inclusion assemblage. Previous studies have shown that fluid immiscibility can cause phase separation of a formerly homogeneous fluid into two or more heterogeneous fluids, destroying the original equilibrium of the system and promoting the precipitation and enrichment of useful components. For example, coexisting quartz formed during the main stage of gold precipitation often contains immiscible fluid inclusions [14,42,43]. Physicochemical experiments have also confirmed that fluid immiscibility favors gold precipitation [44].

6.2. Sources of the Ore-Forming Fluids

Fluid inclusions provide critical evidence for deposit classification and for constraining the sources and evolution of ore-forming fluids [39]. Low-salinity, CO₂-rich inclusions are widely regarded as a characteristic feature of orogenic gold deposits [40,45].

In general, CO₂ in ore-forming fluids may be generated through two main mechanisms: (1) dissolution or decarbonation of carbonate rocks. Direct dissolution of carbonate rocks does not cause carbon isotope fractionation [46,47], whereas decarbonation can increase the δ¹³C value of the fluid by approximately 3–5‰ [48]; and (2) oxidation or hydrolysis of organic matter. At temperatures of 350–600 °C, organic carbon may be converted through the reaction 2C + 2H₂O → CO₂ + CH₄, and this process may increase the δ¹³C value of the fluid by 3–12‰ [46].

δ¹³C_V-PDB values of the ore-forming fluids in the Dulanggou Te–Bi-rich gold deposit range from −8.77‰ to −5.85‰, showing a relatively wide variation. Therefore, the carbon in the ore-forming fluids was not derived entirely from the direct dissolution or decarbonation of marine carbonate rocks in the Devonian host strata (δ¹³C ≈ 0‰; [46,49]), nor can it be attributed solely to an organic matter source, because sedimentary organic carbon generally has δ¹³C values of −30‰ to −20‰ [50,51]. When the carbon and oxygen isotopic compositions of the ore-forming fluids are plotted on the C–O isotope diagram (Figure 11), all data points fall within the fields of igneous carbonatites and mantle xenoliths, and closely match the typical carbon isotopic composition of igneous carbonatites (−8‰ to −5‰; [52]). This feature indicates that the principal carbon source was most likely closely related to mantle-derived carbonatitic magma. Therefore, we suggest that the dominant carbon source of the deposit was mantle-derived magmatic carbon, although minor contamination from marine limestone or organic sedimentary materials cannot be excluded.

H–O isotope studies are one of the most effective methods commonly used to trace the sources of ore-forming fluids in hydrothermal deposits [54,55,56].

A comparison of hydrogen and oxygen isotopic compositions between the Dulanggou gold deposit and typical gold deposits in the Danba–Dadu River metallogenic belt is presented. On the H–O isotopic composition diagram (Figure 12), data points of the Dulanggou gold deposit mainly plot below the field of Phanerozoic typical orogenic gold deposits (i.e., metamorphic water) and fall within the primary magmatic water field. This indicates that the ore-forming fluids were primarily derived from primary magmatic water. This feature differs from that of typical orogenic gold deposits, where metamorphic water is the dominant fluid source [57]. Some deposits differ from the Dulanggou gold deposit, such as Lianhuataizi, Xinjintaizi, and Yanzigou. Their data points plot within the fields of Phanerozoic and Archean orogenic gold deposits (Figure 12). This indicates that their ore-forming fluids are characterized by both metamorphic water and primary magmatic water.

In summary, the ore-forming fluid of the Dulanggou gold deposit is predominantly derived from primary magmatic water. The ore-forming system was relatively closed, and the carbon in the fluid originated from mantle-derived magmatic carbon. Compared with typical orogenic gold deposits in the region, which are dominated by metamorphic water, the fluid source of the Dulanggou deposit is distinctive.

6.3. Discussion of the Ore-Forming Mechanism

Previous studies have shown that fluid immiscibility is one of the most effective mechanisms for gold precipitation. In the Dulanggou gold deposit, the phase types and distribution assemblages of fluid inclusions in quartz veins from the early and main mineralization stages indicate that gold mineralization occurred during fluid immiscibility [58,59]. Microthermometric data show that the ore-forming fluid belongs to the H₂O–CO₂–NaCl–CH₄ system [60]. Petrographic analysis further reveals that in the main-stage quartz veins, CO₂-bearing inclusions (C) and aqueous-rich inclusions (W) abundantly coexist. Because isochores can constrain the minimum trapping pressure relative to the measured homogenization temperature [12,61], the homogenization temperatures of inclusions in main-stage quartz veins yield an estimated ore-forming pressure range of 166–326 MPa [61,62].

The main-stage quartz veins contain abundant immiscible inclusion assemblages of CO₂-bearing (C) and aqueous-rich (W) inclusions (Figure 7). These two types have similar homogenization temperatures but significantly different salinities (Figure 8 and Figure 9), indicating that fluid immiscibility occurred during the main mineralization stage. This immiscibility was likely a key factor for gold precipitation [10,14,63,64,65,66].

The temperature of the early-stage fluid (307–388 °C, average 341 °C) is higher than that of the main stage and also exceeds the eutectic point of the Au–Bi–Te ternary melt system (235–266 °C) [4,67]. This suggests that during the main stage, the initially Au-rich Te–Bi melt underwent fluid–melt phase separation as the temperature dropped below its melting point, accompanied by fluid immiscibility as revealed by fluid inclusions. This process formed Te–Bi-rich metal melt droplets. Native gold adsorbed in the melt was then precipitated sequentially according to the melting points of the tellurium–bismuth minerals [4].

The Dulanggou gold deposit resembles classic orogenic gold deposits in terms of host rocks, ore-controlling structures, ore mineral assemblages, and low-salinity CO₂-rich ore-forming fluids. However, its C–H–O isotopic characteristics indicate that the fluid is dominated by mantle-derived magmatic components. This is different from typical orogenic gold deposits, where metamorphic water is the main fluid source.

7. Conclusions

(1) The mineralization of the Dulanggou gold deposit process can be divided into three stages: the early mineralization stage (quartz + minor sulfides), the main mineralization stage (quartz + abundant sulfides + native gold + telluride–bismuthide minerals), and the late mineralization stage (quartz + minor sulfides + calcite).

(2) In the early stage, fluid inclusions are dominated by the CO₂–H₂O type, with subordinate NaCl–H₂O type. They show medium–high temperature and low salinity. In the main stage, the inclusions are mainly CO₂–H₂O and NaCl–H₂O types, with minor pure CO₂ inclusions. They exhibit wide ranges of homogenization temperature and salinity. In the late stage, inclusions are mainly the NaCl–H₂O type, with low homogenization temperatures. Raman spectroscopy indicates that, in addition to CO₂, CH₄ is present in the vapor phase of inclusions from all three stages. Therefore, the ore-forming fluid of the Te-Bi-rich Dulanggou gold deposit is a reducing, medium–low temperature, low-salinity H₂O–CO₂–NaCl–CH₄ system.

(3) Thermodynamic calculations of CO₂–H₂O inclusions show that from the early stage to the main stage, the total density, molar volume, CO₂ mole fraction, and CO₂ filling degree have similar ranges. However, the total homogenization pressure is generally lower in the main stage than in the early stage. This reflects a progressive decrease in pressure during fluid evolution.

(4) The C–O and H–O isotopic compositions of fluid inclusions in quartz are consistent with each other. They indicate that the ore-forming fluid is genetically related to magmatic fluid and contains mantle-derived components modified by carbonate metasomatism. H–O isotopic data plot in the field of primary magmatic water. This suggests that the ore-forming fluid was mainly derived from primary magmatic water in a relatively closed deep environment. This feature is distinctly different from that of typical orogenic gold deposits, where metamorphic water is the dominant fluid source.

(5) The coexistence of CO₂–H₂O and NaCl–H₂O inclusions in the early and main stages indicates fluid immiscibility. The decrease in temperature and pressure led to phase separation of the fluid. This destabilized gold–sulfur complexes and promoted gold precipitation. In the main stage, abundant tellurium–bismuth minerals are developed. The ore-forming temperature is higher than the eutectic point of the Au–Bi–Te melt system. A melt phase may have participated in the extraction and enrichment of gold [4]. The synergistic effect of fluid immiscibility and melt separation is the key control on efficient gold precipitation in the Dulanggou gold deposit. This study improves the understanding of gold mineralization mechanisms on the western margin of the Yangtze Craton. It reveals the diversity of fluid sources in orogenic gold systems. It also provides a basis for regional metallogenic regularity and exploration targeting.