In this study, systematic sampling was conducted from various ore bodies and different wall rocks. Due to the strong weathering and oxidation of the outcrops, representative samples were primarily collected from the drill core. Only a limited number of relatively fresh samples were obtained from outcrops. A total of 92 samples were used to select representative fluid inclusions for laser Raman spectroscopy and temperature measurements. A total of 14 samples from the main ore-forming phase were selected for H–O isotope analysis. Individual minerals such as pyrite, arsenopyrite, and antimonite were selected from 12 ore samples, four wall-rock samples, and two rock-body samples for S isotope analysis. The 14 samples selected for H–O isotope analysis are characterized by siliceous metasomatism associated with auriferous quartz or net-like veins. A summary of sample locations, rock types, and selected individual minerals is provided in Table 1. Representative field photographs are shown in Figure 4, and photomicrographs of typical mineral assemblages are presented in Figure 5.

Doubly polished wafers of representative samples were prepared for inclusion petrography, which was conducted at the University of the Chinese Academy of Sciences using an optical polarizing microscope. Microthermometry studies involving heating and freezing runs of the fluid inclusions were performed using a LINKAM THMS600 housed at the fluid inclusion laboratory of the North China University of Science and Technology. This equipment operates within a temperature range of −180°C to +600°C. Calibration was conducted at −56.6°C and 0°C using synthetic fluid inclusion standards and pure water to ensure measurement accuracy. The samples were first frozen and then heated gradually to prevent inclusion bursting. The heating rate varied from 0.1°C to 10 C/min, with a slower rate of 0.1 C/min when approaching phase transitions. Repeated measurements were conducted to ensure the accuracy of the measured phase transition temperatures.

Laser Raman spectroscopy of fluid inclusions from doubly polished wafers of representative samples was performed using the LABHR–VIS LabRAM HR800 machine at the Solid Mineral Resource Key Laboratory under the Institute of Geology and Geophysics, Chinese Academy of Sciences. The laser source was a yttrium aluminum garnet (YAG) laser with a wavelength of 532 nm and a power of 1,000 mW. Each spectral line involved an integration time of 10 s, with 10 integrations performed per line. The spectral resolution was ±2 cm^-1 with a beam size of 2 μm.

The H–O–S isotopes were analyzed at the Research Center for Analysis and Measurement, affiliated with the Beijing Uranium Geology Research Institute. For the extraction of H₂O from fluid inclusion in quartz, a decrement method was employed. The extracted H₂O was prepared for the H isotope component measurement. The measurement procedure was as follows: individual quartz samples were first baked low temperature to remove the adsorbed water. The samples were then heated until they fractured, releasing the volatiles. Water vapor was extracted and subjected to a reduction reaction with zinc at 400°C to produce hydrogen. The resulting hydrogen was frozen with liquid nitrogen and collected in a sample bottle containing active carbon. Finally, its isotope component was analyzed using a MAT-253 mass spectrometer, calibrated with the international standard V-SMOW, with an analytical accuracy of ±2%.

For the analysis of the O isotope in quartz, the BrF₅ method was used as follows: a 12 mg pure quartz sample was reacted with BrF₅ under vacuum and high-temperature conditions for 15 h. Oxygen was extracted, and a resistance graphite rod was burnt with oxygen to produce CO₂ at 700°C. Its O isotope component was analyzed using a Delta V Advantage mass spectrometer, calibrated with the international standard V-SMOW, with an analytical accuracy of ±0.2%.

The S isotopic composition was measured according to the following procedure: first, individual metal sulfides were uniformly mixed with cuprous oxide—ground to 200 mesh—using a precision scale. The mixture was then heated under vacuum condition at 2.0 Pa. Subsequently, the mixture was subjected to an oxidizing reaction at 980°C to produce SO₂. SO₂ was collected using the freezing method under vacuum conditions. The S isotope component was analyzed using the Delta V Plus mass spectrometer, calibrated with the CDT standard, with an analytical accuracy better than ±0.2%.

The ore characteristics of the various stages are listed below:

Stage I: During this stage, mineralization is either absent or occurs to a minimal extent. Well-developed quartz veins and quartz lenses are present, characterized by coarse brownish-white quartz grains ranging in size from 0.5 to 1 mm (Figure 4e). The average thickness of quartz veins is 1–3 cm, with the thickest reaching 30 cm. Small amounts of cubic-type pyrite occur within the quartz veins at this stage. The characteristics of quartz veins are similar to those observed in the earliest stage of the Muruntau gold deposit (Wilde et al., 2001).

Stage II (quartz polymetallic sulfide stage): This is the main stage of gold mineralization. In contrast to the barren quartz veins, mineralized quartz veins have experienced intense structural deformation, fracturing, and hydrothermal alteration. These mineralized quartz veins exhibit fine-grained or net-like textures and are commonly found within altered and fractured zones. The quartz grains range in color from deep gray to light gray, with grain sizes typically between 0.05 and 0.2 mm. The sulfides within these veins are predominantly euhedral or subhedral pyrite, accompanied by minor amounts of arsenopyrite, stibnite, galena, sphalerite, and chalcopyrite (Figures 5c–i). Gold is primarily hosted within pyrite and arsenopyrite, which are distributed along fractures in early-stage mineralized quartz veins or quartz breccias (Figures 4f,g). Notably, pyrite exhibits a zoning structure (Figures 5c,e), indicating that the mineralization process underwent a transition from compressional shear to tension stress conditions and involved multiple episodes of hydrothermal activity. Stibnite ores lack significant gold mineralization and occur mainly as veinlets, net veins, breccia, and disseminations.

Stage III (quartz-carbonatization stage): This is the late mineralization stage, characterized by the development of quartz–carbonate or net-like veins with widths ranging from 0.2 to 3 cm. The quartz–carbonate assemblages are metasomatic, replacing the sulfide quartz veins formed in stage II or filling the crushed alteration zones (Figure 4G). Quartz veins exhibit geode and pectinate structures. Carbonate minerals primarily consist of calcite and ferrodolomite. Sulfide occurrences are minimal at this stage, and gold mineralization is weak.

Orogenic gold deposits form during compressional to transpressional deformation processes at convergent plate margins in accretionary and collisional orogens. These deposits exhibit temporal and spatial association with orogeny (Groves et al., 1998). They display similar geodynamic settings and geochemical characteristics, forming at various depths ranging from the earth’s surface to 2–20 km, including the superficial zone (<6 km), middle zone (6–12 km), and deep zone (>12 km) (Groves et al., 1998). As an important mineralization type, 23 gold deposits worldwide, each with reserves exceeding 500 t, belong to the orogenic gold deposit category (Bierlein et al., 2006). Since the concept of orogenic gold deposits was introduced, extensive research has been conducted on their tectonic settings, mineralization characteristics, formation ages, ore-forming fluid features, and mineralization mechanisms (Bierlein and Maher, 2001; Wilde et al., 2001; Goldfarb et al., 2001; Groves et al., 2003; Mao et al., 2002; Chen YJ et al., 2004; Chen YJ, 2006; Bodnar et al., 2014; Derek et al., 2016; Damien, 2019). Concurrently, numerous orogenic-type gold deposits have been discovered.

A comprehensive comparison (Table 5) between the Ashawayi gold deposit and the typical orogenic gold deposit was carried out regarding geological and mineralization fluid characteristics (Groves et al., 1998; Kerrich et al., 2000; Goldfarb et al., 2001; Chen YJ, 2006; Chen Y. J. et al., 2007). The Ashawayi gold deposit shares many similarities with the typical orogenic gold deposit, including structural environment, host rock features, ore-controlling structures, ore body shapes, ore types, mineral assemblages, fluid inclusion types, fluid properties, original fluid characteristics, main mineralization temperature, fluid pressure system, and fluid evolution. Therefore, we propose that the genetic type of the Ashawayi gold deposit is defined as an orogenic gold deposit.

The heterogeneous capture of fluid inclusions is prevalent in the Ashawayi gold deposit, with types I, II, and III inclusions coexisting within the same sample (Figure 6c). Two distinct types of fluid inclusions with differing characteristics occur in close association within the same sample, indicating simultaneous trapping. The simultaneous capture of fluids with distinct properties by minerals is most likely caused by heterogeneous trapping, which is an important indicator of fluid immiscibility or mixing (Ramboz et al., 1982; Roedder, 1984). The evolution of salinity and homogenization temperature of fluid inclusions during each metallogenic stage further supports the occurrence of fluid immiscibility: ① from the early to the middle stage of mineralization, fluid salinity increases and disperses as temperature decreases, contradicting the general trend that solubility decreases with decreasing temperature. High salinity predominantly occurs in type I inclusions, while type II inclusions are mainly characterized by low salinity (Figures 8, 9). This phenomenon may result from fluid concentration due to the escape of the gas phase, i.e., boiling. ② Some fluid inclusions exhibit varying salinities within approximately the same temperature range. Thus, the metallogenic process of the Ashawayi gold deposit resembles that of most vein-type mesothermal hydrothermal deposits worldwide, where fluid immiscibility plays a significant role in gold deposit formation (Drummond SE and Ohmoto H, 1985; Cox et at., 1995; McCuaig and Kerrich, 1998; Sibson et al., 1988; Jia et al., 2000; Chen YJ et al., 2005; Chen et al., 2006; Chen et al., 2012a; Chen et al., 2012b; Chen B et al., 2019; Skinner, 1997; Wilkinson, 2001). Between 320°C and 360°C, the fluid underwent its first immiscibility event, leading to the formation of various types of fluid inclusions through heterogeneous trapping. During this stage, initial gold mineralization occurred, although it was generally low, forming quartz veins with little to no mineralization. In the temperature range of 220°C–300°C, primarily between 220°C–250°C, secondary fluid immiscibility occurred after the early fluid evolution stage. This resulted in higher salinity in the liquid phase and the formation of low-salinity and gas-rich inclusions, specifically high-salinity type I inclusions and low-salinity type II inclusions. Continuous CO₂ escape from the fluid ore-forming system led to fluid concentration and supersaturation, promoting the rapid precipitation of ore-forming materials (Sui et al., 2000; Chen H. Y. et al., 2007; Chen et al., 2012a; Chen et al., 2012b). Meanwhile, the escape of CO₂ will disrupt the stability of the Au(HS)₂ ^- complex in the fluid, prompting the decomposition of the Au(HS)₂ ^- complex and leading to the enrichment and precipitation of Au. The reaction equations are as follows:

①: 4Au(HS)₂ ^- + C + 4H⁺ + 2H₂O = 4Au + CO₂ + 8H₂S (Craw, 2002; Craw et al., 2007; Zoheir et al., 2008; Hu et al., 2015; 2017; Damien, 2019).

In conclusion, secondary fluid immiscibility is thus a critical mechanism for gold enrichment and precipitation. Additionally, the water–rock reaction between the fluid and the surrounding rock may contribute to gold precipitation to some extent. These reactions not only extract ore-forming substances from the surrounding rock to promote fluid saturation but also consume H₂O during the reaction, further enhancing fluid concentration. For example, C + 2H₂O = CO₂ + 2H₂ (Cox et al., 1995; Naden and Shepherd, 1989). In the late stage, at temperatures between 180°C and 200°C, fluid immiscibility becomes less pronounced, characterized by temperature decrease. This stage is marked by occasional type II inclusions, absence of type III inclusions, homogenization of the fluid to the liquid phase, and decreased fluid salinity. Quartz–carbonate veins develop during this stage, with very weak or nominal mineralization.

The homogenization temperature (T) and pressure (P) of fluid inclusions trapped in the fluid boiling state are theoretically equal to the temperature and pressure at the time of trapping (Pichavant et al., 1982; Ramboz et al., 1982; Brown, 1998). However, this hypothetical condition does not apply to the Ashawayi gold deposit. As previously discussed, the fluids in the Ashawayi gold deposit underwent a complex inhomogeneous trapping process. Nevertheless, the mineralization depth was estimated using the homogenization pressure of fluid inclusions as the minimum trapping pressure.

After excluding homogenization temperatures from the main mineralization stage that were influenced by late-stage fluids (homogenization T < 200°C), the depth is estimated using all remaining Type I and Type II inclusions, based on the minimum capture pressure. The pressures of types I and II inclusions are estimated based on the methods described by Bodnar and Vityk (1994) and the FLINCOR Program proposed by Brown (1989), respectively. The H₂O–CO₂–NaCl system was selected to estimate the mineralization depth. The results show two distinct pressure ranges: a low-pressure range (16.2–36.7 MPa) and a high-pressure range (75.0–103.6 MPa). The uncertainty in estimating the fluid capture depth (mineralization depth) makes it difficult to determine whether the estimated capture pressure is lithostatic or hydrostatic, which depends on the degree of opening and closing of the fault (Roedder, 1984). Fluid immiscibility indicated by fluid inclusions has special implications for the hydrodynamics of mineralization. One of the major causes of fluid immiscibility is the decrease in fluid pressure, which is directly related to fluid flow driving forces (Guoxiang and Steele-Macinnis, 2025). The co-existence or alternative appearance of lithostatic and hydrostatic pressure results from oscillations in fault healing and fracturing in orogenic gold deposits, as described using the Fault Valve Model (Sibson et al., 1988; Kerrich et al., 2000; Cox et al., 2001). When faults heal, fluids are under (super)lithostatic pressure due to tectonic overpressure, and when fracturing or hydrofracturing occurs, fluids are under hydrostatic pressure. The disappearance of vapor components removes heat, causing a decrease in the mineralization temperature and an increase in the salinity of the remaining fluid—sometimes to the point of oversaturation, thereby leading to the rapid deposition of ore-forming materials and subsequent fracture healing (Deng et al., 2008). Determining the values of these two types of pressures is crucial for accurate estimates. The mineralization depth remains relatively constant during the fluid deposition process, with pressure fluctuations triggered by changes between different pressure types. Lithostatic and hydrostatic pressures frequently change during fault closure and opening. If trapped fluid inclusions record both pressure types, the end members of the two pressure ranges can be considered. Since lithostatic pressure is higher than hydrostatic pressure at the same depth, the low-value pressure endmember is closer to hydrostatic pressure, while the high-value pressure endmember is closer to lithostatic pressure (Xu et al., 1995). The Ashawayi gold deposit is located in a thrust fault system. Ores’ fabric features indicate that the mineralization process experienced a transition from compressive stress to intense stress with multistage action characteristics, suggesting multiple fracture and closure processes that cause consistent fluctuations in fluid pressure.

Accordingly, we assume that the high- and low-pressure range fluids are under the pressure systems of static rock pressure and static hydrostatic pressure, respectively, and estimate them using the general formula for the pressure–depth relationship, P = ρgh, taking ρ as the average rock density of the mining area at 2.6 g/cm³ when the fluids are under the static rock pressure system and ρ as the density of various inclusions calculated at the time when the fluids are under the static hydrostatic pressure system. The corresponding fluid depth ranges are 1.8–4.2 km for the high-pressure range and 2.9–4.2 km for the low-pressure range. The high values of the depth ranges of the two are close, so the fluid may be in the depth range of 1.8–4.2 km due to the complex non-miscible process caused by the co-existence and alternation of static rock pressure and static hydrostatic pressure, resulting in the precipitation of gold. The following three points also directly or indirectly support the above viewpoint: (1) the estimated pressure range of Type II inclusions falls entirely within the low-pressure range, and the peak of the low-pressure range is located in the temperature range of 220°C–250°C, which is consistent with the result that Type I and Type II fluid inclusions are captured simultaneously by the fluid inclusions in the temperature range of 220°C–250°C due to the immiscibility of the fluid inclusions in this temperature range. (2) No Type II inclusions are found in the high-pressure range, only a part of Type I inclusions, i.e., no Type II inclusions are captured in the high-pressure peak range, which is consistent with the fact that the closed-system hydrofluids in the static rock pressure system do not undergo obvious immiscibility. (3) Type I inclusions account for the majority of the total inclusions in the temperature range of 220°C–250°C, which is considered representative of the mineralization conditions.

It is noteworthy that the homogenization temperature range (220°C–250°C) under low-pressure conditions is comparable to the peak values observed in well-known mesothermal gold deposits such as the Muruntau, Kumtor, and Sawayaerdun deposits (Ivanove, 2000; Wilde et al., 2001; Graupner et al., 2001; Chen et al., 2012a) and other typical mesothermal vein-type gold deposits (Spooner et al., 1987; Goldfarb et al., 1989; Fan et al., 2003; Zhang et al., 2009; Hagemann and Luders, 2003). However, the corresponding mineralization depth (1.8–4.2 km) is significantly shallower than that of the Sawayaerdun gold deposit in Chinese Tianshan (Chen et al., 2012a).

The δ¹⁸O values of mineralized quartz from the Ashawayi gold deposit range from 18.6‰ to 23.6‰ (Table 3; Figure 10). These values are higher than those of most lode gold deposits (δ¹⁸O = 10‰–18‰) (Kyser et al., 1986; Kerrich, 1987; Goldfarb et al., 1991; 1997; McCuaig and Kerrich, 1998; Ivanove, 2000; Kerrich et al., 2000; Jia et al., 2001; Chen et al., 2012b) but are similar to those of the Sawayaerdun gold deposit in southwestern Tianshan, China (Chen et al., 2012b). The elevated δ¹⁸O_quartz values compared to those of other vein-type gold deposits (commonly in the temperature of 300°C–350°C; Ridley and Diamond, 2000) were likely caused by strong isotope fractionation under low-mineralization temperature conditions. The calculated δ¹⁸O values of the mineralization fluid range from 8.1‰ to 13.1‰, which is consistent with those of most lode deposits.

The δ¹⁸D values of the mineral-forming fluid range from −84.8 to −69.2‰ (Table 3; Figure 11). This range falls within that of most vein-type gold deposits (Kerrich, 1987; McCuaig and Kerrich, 1998; Ridley and Diamond, 2000; Chen et al., 2012b), but the interval is narrower than that of the Juneau gold mineralization belt in Alaska, United States (Goldfarb et al., 1991), and the Sawayaerdun gold deposit in China (Chen et al., 2012b). The peak value of δ¹⁸D_water in the range of −68 to −72‰ is close to the δ¹⁸D_water value of meteoric water (-70‰) in the Tianshan region, suggesting that part of the mineralization fluid may be derived from atmospheric precipitation. However, according to the low δ¹⁸D_water value (<−80%) obtained from lode gold deposits hosted by carbonaceous turbidite, the fluid is also sourced from reactions between deep-source non-atmospheric water and organic substances in the host rock, leading to depleted δ¹⁸D_water (Goldfarb et al., 1989; McCuaig and Kerrich, 1998; Jia et al., 2001). As shown in Figure 12, the mineralization fluid of the Ashawayi gold deposit plots to the right of the metamorphic and igneous water fields, consistent with most vein-type gold deposits. It is comparable to the Juneau orogenic gold deposit belt (Goldfarb et al., 1991) and the Sawayaerdun gold deposit (Yang et al., 2006; Chen et al., 2012b). In conclusion, the fluids of the Ashawayi gold deposit are likely derived from a mixture of metamorphic fluid and meteoric water.

The amount of sulfides increases significantly from the early to the main mineralization phase, while the ratio of CO₂–H₂O inclusions to total inclusions decreases markedly. This indicates a transition in the mineralized fluid from high oxygen fugacity during the early stage to low oxygen fugacity during the main stage. This feature is consistent with the low SO₂/H₂S redox level observed in most orogenic gold deposits (Phillips and Groves, 1983; Mikucki and Ridley, 1993; McCuaig and Kerrich, 1998; Jia et al., 2000; Kerrich et al., 2000). At low oxygen fugacity and moderate temperature (220°C–250°C), S isotope fractionation between sulfide and fluid is generally minimal (<2%, Ohmoto and Goldhaber, 1997). Therefore, the δ³⁴S_V-CDT values of sulfide and fluid in the main mineralization stage are similar.

The sulfides from orogenic gold deposits exhibit a wide range of δ³⁴S_V-CDT values, but the range of the S isotope in the S-complexion of mineralization fluid remains inconclusive (McCuaig and Kerrich, 1998; Goldfarb et al., 2005). For example, the δ³⁴S_V-CDT values of sulfides in Archean orogenic gold deposits range from 0‰ to 9‰ (Kerrich, 1987; Golding et al., 1990; McCuaig and Kerrich, 1998), whereas those in Phanerozoic orogenic gold deposits range −20‰–25‰ (Peters and Golding, 1989; Kontak et al., 1990). The δ³⁴S_V-CDT values of pyrite, arsenopyrite, and stibnite in the ore bodies range from +13.8‰ to 24.4‰ (Table 4). Specifically, pyrite and arsenopyrite, as gold-carrying minerals, have δ³⁴S_V-CDT values ranging from +16.8‰ to 24.4‰, with an average of +18.3‰. These values fall within the ranges typical of metamorphic and sedimentary rock and are significantly higher than those of basaltic and granitic rocks. They are consistent with the δ³⁴S_V-CDT values (17.1‰–18.5‰) of pyrite and arsenopyrite in phyllite and siltstone, which are the wall rocks of the ore bodies. In contrast, the δ³⁴S_V-CDT values (−6.9‰–1.5‰) of the rock body in the diggings (Figure 13) indicate that the ore components are derived from wall rocks. Compared with that in other well-known orogenic gold deposits, such as the Juneau gold belt (Goldfarb et al., 1989), the Bendigo in Australia (Jia et al., 2001), the Kumtor in Kyrgyzstan (Ivanove, 2000), and the Sawayaerdun in China (Chen et al., 2012b), the positive anomaly of δ³⁴S_V-CDT in the Ashawayi gold deposit suggests that the sulfur in the ores originates primarily from a shallow position within the orogenic belt and upper crust.

Numerous giant and ultra-large orogenic gold deposits have been discovered in the Southern Tianshan region of Central Asia. In contrast, only one ultra-large deposit, Sawayaerdun, has been identified in Southern Tianshan in Xinjiang Province. Although the Ashawayi gold deposit ranks second in scale, it is only moderately large and has relatively low gold grades. To understand the significant disparity in gold scale and ore grade between deposits formed within the same geological evolution history, we compare factors such as the ore-hosting formation background, structural superposition, magmatic activity, and ore denudation. The reasons for the lower size and grade of the Southern Tianshan deposits in Xinjiang Province than those in Central Asia can be summarized as follows: (1) the ore-hosted strata in the Southern Tianshan of Xinjiang Province are relatively young and thin and have a low carbon content, lacking a material basis for large-scale gold mineralization. (2) The crust in the Southern Tianshan of Xinjiang Province is relatively thick, leading to shallow fracture depths, limited denudation, weak deformation, and low metamorphism. (3) Magmatic activity in the Southern Tianshan of Xinjiang Province is relatively weak, resulting in less intense crust–mantle interaction and a relative lack of magmatic activation and re-enrichment processes for gold.

Based on this analysis, future prospecting efforts at Ashawayi should focus on strata with high carbon content formations and areas subject to strong structural deformation and metamorphism, such as the intersection of low-angle and high-angle faults to the northeast of the mining area.