Evolution of Ore-Forming Fluids and Gold Deposition of the Sanakham Lode Gold Deposit, SW Laos: Constrains from Fluid Inclusions Study

: The Sanakham gold deposit is a newly discovered gold deposit in the Luang Prabang (Laos)–Loei (Thailand) metallogenic belt. It consists of a series of auriferous quartz-sulﬁde veins, which is distinguished from the regional known porphyry-related skarn and epithermal gold deposits. There are four mineralization stages identiﬁed in Sanakham, with native gold grains mainly occurring in stages II and III. Evolution of ore-forming ﬂuids and gold deposition mechanisms in Sanakham are discussed based on ﬂuid inclusion petrography, microthermometry, and Laser Raman spectroscopy. The original ore-forming ﬂuids belong to a medium-high temperature (>345 ◦ C) CH 4 -rich CH 4 –CO 2 – NaCl–H 2 O system. In stages II and III, the ore ﬂuids evolve into a NaCl–H 2 O–CO 2 ± CH 4 system characterized by medium temperature (~300 ◦ C), medium salinity (~10 wt% NaCl eq.), and CO 2 -rich (~10% mol). They might ﬁnally evolve into a NaCl–H 2 O system with temperature decreasing and salinity increasing in stage IV. Two ﬂuid immiscibility processes occurred in stages II and III, which created high-CH 4 & low-CO 2 and low-CH 4 & high-CO 2 end-members, and CO 2 -poor and CO 2 -rich endmembers, respectively. Gold-deposition events are suggested to be associated with the ﬂuid immiscibility processes, with P–T conditions and depth of 236–65 MPa, 337–272 ◦ C, and 8.7–6.5 km, respectively.


Introduction
The mainland of Southeast Asia consists of a series of continental blocks or terranes together with related sutures and volcanic arcs (Figure 1a). It is endowed with a diversity of mineral resources distributed in several important metallogenetic belts, such as the Truong Son Fe-Cu-Au-W-Sn belt, Vientiane-Kon Tum K-Al-Cu-Au belt, Da Lat Au-Fe-Cu belt, Gulf of Thailand-East Malaysia Au-Sb-Cu belt, and the Luang Prabang-Loei Cu-Au-Ag belt [1]. The Luang Prabang (Laos)-Loei (Thailand) metallogenic belt lies along the western periphery of the Indochina Block, to the east of Changning-Menglian-Chieng Mai Suture zone (Figure 1a). The belt, formed during the subduction-accretionary orogeny of the Paleo-Tethys Ocean and the subsequent collisional orogeny between the Sibumasu Block and the Indochina Block, is traditionally recognized as a metallogenic belt dominated by porphyry-related skarn and epithermal deposits [2,3] (Figure 1b).  [4][5][6][7][8][9][10][11][12][13][14][15][16], modified from [5,10]  The Sanakham gold deposit, a newly discovered gold deposit in this belt with a gold resource of 10.6 t @ 3.05 g/t Au, is located about 3 km north to the Muang Sanakham County, southwest Laos (Figure 1b). The gold orebodies consist of a series of quartz-sulfide veins that are controlled by secondary structures of the regional NNE-trending brittle-ductile shear zone, which is different from the regional porphyry-related skarn and epithermal gold deposits. Given it is such a particular lode gold deposit in the Luang Prabang-Loei belt, studying it should be meaningful in both prospecting and theory. Nevertheless, there are no reports of studies on the Sanakham gold deposit up to now. We therefore carried out detailed petrography, microthermometry, and Laser Raman spectroscopy on fluid inclusions in gold-related quartz from different mineralization stages, identified by field and microscopic observation. This study aims to discuss the evolution of ore-forming fluids and gold deposition mechanisms in Sanakham, and to provide basic data for further study on regional gold metallogenic regularity.
The northwest side of the Luang Prabang tectonic belt belongs to the Simao-Phitsanulok Block (Figure 1a), which is regarded as a back-arc foreland basin [1]. It mainly consists of Late Paleozoic sandstone, shale, and epimetamorphic rocks, Triassic clastic rocks and carbonate rocks, and Middle Jurassic red clastic rocks [10]. Only a few copper occurrences are found in the basin.
The southeast side of the Luang Prabang tectonic belt is the Loei Fold Belt (Figure 1a), with Late Paleozoic and Mesozoic sediments and volcanic rocks [6,13], as well as Early-Middle Triassic magmatic rocks exposed ( Figure 1b). The representative porphyry-skarn copper-gold deposits (Phu Lon, Puthep, and Phu Thap Fah) develop to the north and northeast of Loei in Thailand and are genetically related with the felsic-intermediate intrusive rocks [14,15]. A set of Late Permian to Early Triassic andesitic-rhyolitic volcanic and pyroclastic rocks develops in the south of the metallogenic belt, associated with several epithermal Au-Ag deposits [16].

Local Geological Setting
The main stratum in the mining area is the Upper Carboniferous Nanpo Formation (C 2 n), which consist of purple-red thick-layered metamorphic siltstone, silty slate, slate, and brown thin-layered feldspar-debris sandstone ( Figure 2). It is N to NNE trending, generally dipped to the west with a dip angle of 60-85 • , and is more than 600 m in thickness. The Quaternary sediments (Q), which mainly consists of fragments that came from slate, carbonaceous slate, and metamorphic siltstone, are widespread on gentle slopes and platforms in various areas of the mining area.
The structures in the mine are mainly brittle-ductile shear zone, fault, and fracture ( Figure 2). The brittle-ductile shear zone trends northeast, consistent with the regional Luang Prabang structural belt, and generally controls the overall output of the orebodies. The faults are secondary structures of the regional brittle-ductile shear zone and trend NNE or nearly S-N. They control the main intrusions and geological boundaries in the mining area. Gold orebodies normally occur within the faults and their secondary fractures. The Middle Triassic intermediate-acid magmatic rocks are well developed and intruded into the Upper Carboniferous low-grade metamorphic rocks as an NNE-trending intrusive stock ( Figure 2). The intrusions mainly consist of quartz monzodiorite with a hypidiomorphic granular texture and massive structure. It is composed of plagioclase, amphibole, alkaline feldspar, a small amount of quartz, and trace amounts of pyrite in size of 0.1-4 mm, and is gray-white in color because of weathering.
Ninety-one gold orebodies have been identified in the Sanakham deposit, and they can be divided into more than ten orebody groups ( Figure 2). Most of the orebodies, vein or veinlet in shape, are included in the Au9 orebody group and account for 3/4 of the total gold resources. The several main orebodies in the Au9 orebody group are about 80-400 m in length, strike towards the NE, and dip 10 • -40 • SE with a depth of 300-600 m. They range in thickness from 0.8 to 4 m, averaging about 1.2-1.5 m. The gold grade ranges from 1.0 to 12.3 g/t, with an average gold grade of 3.05 g/t. The wallrocks are generally quartz monzonite, slate, and, locally hornfel.

Wall-Rock Alterations and Ore Mineralization
The wall-rock alteration at Sanakham mainly developed in the quartz monzodiorite and is controlled by the NE-NNE trending faults. The hydrothermal alterations are primarily silicification, sulfidation, sericitization, and carbonation, among which silicification and sulfidation have the closest relationship with gold mineralization. Silicification is the most widely distributed alteration, normally occurred as veins and networks, with a small amount disseminated in the intrusions. Sulfidation is dominated by pyrite, and lesser pyrrhotite, chalcopyrite, arsenopyrite, galena, and sphalerite. They co-exist with quartz to form quartz-sulfide veins (Figure 3a-d), generally as banded sulfide assemblages or massive aggregates (Figure 3e-f). The plagioclase, biotite, and potassium feldspar are locally subjected to sericitization around the orebodies, showing a coexistence relationship with disseminated silicification. Carbonation is normally in the form of calcite veinlets, 0.1 to 1 cm in width, interspersed in the wall rock. K-feldspar and chlorite are widely distributed in the intrusions and formed earlier than gold-related mineralization. Hydrothermal metasomatic alteration develops along the contact zones between quartz monzonite and slate, with a width of less than 20 m. It is characterized by skarnization, marbleization and hornfel alteration, but has little spatial relation with gold mineralization.
The lode gold orebodies primarily occur in the fault zones and fractures in the quartz monzodiorite and locally extend into Upper Carboniferous strata along with the structures. Primary sulfide ores mainly present vein stockwork and massive structures followed by disseminated structures. The ore minerals are primarily pyrite (~10 vol%), with lesser chalcopyrite and pyrrhotite (each 2-3 vol%), and a small amount of arsenopyrite, galena, and sphalerite (each~1 vol%); the gangue minerals mainly consist of quartz (60-70 vol%) with minor calcite (5-10 vol%) and lesser sericite (1-3 vol%). Based on cross-cutting relationships and mineral assemblage characteristics (Figure 3), four mineralization stages have been identified in Sanakham ( Figure 4): quartz-pyrite-arsenopyrite (stage I), quartzpyrite-pyrrhotite-chalcopyrite (stage II), quartz-pyrite-base-metal sulfide (stage III), and quartz-carbonate (stage IV). Native gold occurs mainly in the main ore stages, i.e., stages II and III.
Stage I is normally quartz veins that comprise milky white quartz (Figure 3a), minor sericite, and euhedral to subhedral coarse-grained disseminated pyrite ( Figure 3g) and arsenopyrite ( Figure 3i). Stage II is characterized by quartz-sulfide veins and veinlets in grey color. The minerals are primarily subhedral to anhedral fine-grained pyrite, with minor pyrrhotite and chalcopyrite, which normally occurs around the boundary of stage I pyrite and arsenopyrite (Figure 3h,i). Native gold appears generally as irregular inclusions in pyrrhotite or chalcopyrite that fill in fractures of pyrite (Figure 3j,k). Similarly, stage III is characterized by quartz-sulfide veins and sulfide massive aggregations that consist of quartz, pyrite, pyrrhotite, chalcopyrite, and galena and sphalerite in particular (Figure 3l). They normally crosscut or overprint the stage II veins. Stage IV is characterized by quartzcalcite veinlets that crosscut the quartz-sulfide veins (Figure 3d).

Sampling and Analytical Methods
A total of nine samples of gold-bearing ore were selected for fluid-inclusion analyses. Twenty doubly polished sections (about 0.1 mm thick) were chosen for petrographical study; of these, twelve sections were selected for microthermometry and Laser Raman study on hydrothermal quartz grains at the Fluid Inclusion Laboratories, Chengdu Center of China Geological Survey.
Fluid inclusion microthermometry was made with a Linkam THMSG 600 heatingcooling stage (−198 to 600 • C). The reproducibility of measurements was ±1 • C at temperatures >30 • C and ± 0.1 • C at temperatures <30 • C. The measurements of melting temperatures of ice (T m ice ), melting temperatures of the carbonic phase (mainly consist of CO 2 and CH 4 , T m carbon ), clathrate-melting temperatures (T m clath ), and partial homogenization of carbonic phase (T h carbon ) were made at the heating rate of 0.2 • C/min, and for final homogenization temperature (T h TOT ), the heating rate was 1.0 • C/min.
Laser Raman spectroscopy of individual inclusions was performed using a Renishaw Raman spectrometer equipped with a 514 nm Ar ion laser as the source of excitation. The instrument records peaks in the range of 1200 to 3800 cm −1 with a resolution of ±1 cm −1 and a spectral repeatability of less than 0.1 cm −1 .

Fluid-Inclusion Petrography
Based on micro-observation, microthermometry and Laser Raman analyses, the fluid inclusions were classified and grouped according to textural criteria using the FIA (fluid inclusion assemblage) method [20,21]. Three gold-related primary fluid inclusion types were observed in quartz: type M, CH 4 -rich fluid inclusion; type C, CO 2 -rich fluid inclusion; and type W, aqueous inclusions. Type M inclusions normally comprise a carbonic vapor phase and an aqueous phase at room temperature ( Figure 5a). The carbonic phase consists primarily of CH 4 , less of CO 2 , and occasionally of C 2 H 6 and H 2 S, and separates into vapor and liquid carbonic phases through cooling ( Figure 5b). These inclusions have a continuum V/(L + V) ratios range from 0.3 to 0.8 and homogenize to either liquid or vapor. Daughter minerals can be occasionally found in individual type M inclusions (Figure 5c). Type C inclusions contain two (aqueous + carbonic) or three (aqueous + liquid carbonic + vapor carbonic) phases at room temperature, and can be further divided into two subtypes based on their V/(L + V) ratios: type C1 has V/(L + V) ratios mostly between 0.1 and 0.3 and homogenizes into liquid (the aqueous phase, Figure 5d), whereas type C2 has V/(L + V) ratios of 0.6 and 0.9 and normally homogenizes into vapor (the carbonic phase, Figure 5e).
The carbonic phase of type C inclusions consists of major CO 2 and less CH 4 , which is different from type M inclusions. Type W aqueous inclusions comprise an aqueous phase and a vapor phase with V/(L + V) ratios of 0.05 to 0.25, and homogenize to liquid (Figure 5f). The type M and C inclusions are normally 5-15 µm in diameter, have predominately regular to negative crystal shape or partly irregular shape. The type W inclusions are relatively smaller (2-10 µm) and have regular to irregular shape.
Six groups of FIA occur in hydrothermal quartz grains and are summarized in Table 1.

Results of Microthermometry
A total of 23 FIAs selected from hydrothermal quartz from all four stages were analyzed. The microthermometric data and calculated parameters (min, max, mean and number) for every FIA are listed in Tables 2-5. Of which, mean values of each type of fluid inclusions in every FIA are used for assessment and math statistics.

FIA in Stage I
Four Group 1 FIAs, each comprising 3-5 inclusions, were measured to gain exact microthermometry data (Table 2 and Figure 6). The T m carbon of type M inclusions occurs between −74.2 and −69.5 • C, with a mean value of −72.2 • C. The T m clath and T h carbon of type M inclusions range from −5.7 to −2.9 • C (mean −4.2 • C), and 4.9 to 9.5 • C (mean 6.8 • C), relatively. These inclusions generally homogenize into liquid phase finally, with T h TOT between 326 • and 345 • C (mean 337 • C).

FIA in Stage II
Two Group 2 FIAs, which are comprised of four type C1 inclusions each, were measured successfully (Table 3 and Figure 6). The T m carbon , T m clath , T h carbon , and T h TOT of C1 inclusions in Group 2 FIA are −57.9 to −57.5 • C (mean −57.7 • C), 4.3 to 4.8 • C (mean 4.6 • C), 18.2 to 20.3 • C (mean 19.3 • C), and 305 to 318 • C (mean 312 • C), respectively. The microthermometric data are similar to those in Group 1 FIA in stage I quartz.
Six Group 3 FIAs, which are widespread in stage II quartz and consist of several coexisting type M and C1 inclusions, were chosen for microthermometry (Table 3 and   (Table 4 and Figure 6), the T m carbon , T m clath , T h carbon , and T h TOT of type C1 inclusions are, respectively, −57.  (Table 5 and Figure 6). The type W inclusions have melting temperatures of ice (T m ice ) range from −14.5 to −8.6 • C (mean −11.3 • C), and total homogenization into liquid occurs between 196 and 235 • C (mean 218 • C).

Results of Laser Raman Spectroscopy
Laser Raman spectroscopy indicates that both type M and type C fluid inclusions have carbonic component-dominate vapor phase (Figure 7). The gas phases of these inclusions consist of CH 4 + CO 2 ± C 2 H 6 ± H 2 S, with no detectable N 2 . The type C inclusions normally have high CO 2 spectroscopic peaks (Figure 7a), while most spectroscopic analyses show high CH 4 peaks in type M inclusions (Figure 7b-d). Type M inclusions contain about 70-85 mole % CH 4 in carbonic phase, much higher than that of C1 inclusions (CH 4 contents normally lower than 20%). The occasional daughter minerals in type M inclusions are normally chalcopyrite (Figure 7d) and calcite.      The salinity (wt% NaCl equiv (eq.)), bulk density, and composition of the above fluid inclusions were estimated according to the program of Bakker [22] using the software MacFlincor [23]. The CH 4 content in the carbonic phase of type C inclusions was estimated from the V-X phase diagram of the CO 2 -CH 4 system [24]. The bulk CH 4 content, thus can be calculated combined with X CO2 + X CH4 and carbonic X CH4 . For the type M inclusions, CH 4 content is too high to get its accurate data based on the V-X phase diagram, thus evaluated by Laser Raman spectroscopy here. The salinities of type M inclusions are calculated using T m clath based on the equation from Darling et al. [25]. The bulk density of type M inclusions are hard to assess because of the high CH 4 content. The results of FIA assessment and math statistics are presented in Tables 2-5

Salinity
Stage I: Salinities in type M inclusions in Group 1 FIA vary from 18.6 to 21.0 wt% NaCl eq. (mean 9.8 wt% NaCl eq.).
Stage II: Salinities in type C1 inclusions in Group 2 FIA vary from 9.4 to 10.1 wt% NaCl eq. (mean 9.8 wt% NaCl eq.). The type C1 inclusions in Group 3 FIA that coexist with type M inclusions, similarly, have salinities ranging from 7.6 to 11.3 wt% NaCl eq. (mean 9.6 wt% NaCl eq.). Type M inclusions in Group 3 FIA, however, have higher salinities ranging from 19.3 to 24.1 wt% NaCl eq. (mean 21.3 wt% NaCl eq.), which are even higher than those in type M inclusions in Group 1 FIA.
Stage III: The range of salinities of types C1 and C2 inclusions in Group 5 FIAs are 6.8-10.6 wt% and 7.6-12.6 wt% NaCl eq. (mean 10.3, and 9.2 wt% NaCl eq.), respectively. The salinities of inclusions show a decreasing trend along with the increasing of CO 2 content in Group 5 FIAs. Salinities of type C1 inclusions in Group 4 FIA are 9.1-10.1 NaCl eq. (mean 9.6 wt% NaCl eq.), basically equal to C1 inclusions in Group 5 FIA in stage III, as well as C1 inclusions in Group 2 FIA in stage II.
Stage IV: Type W inclusions in Group 6 FIA have relatively higher salinities of 12.4-18.4 NaCl eq. (mean 15.2 wt% NaCl eq.) than C1 inclusions from previous stages.

Bulk Density
Stage II: The bulk densities of the type C1 inclusions in both Groups 2 and 3 FIA are nearly the same, from 0.856-0.877 g/cm 3 (mean 0.867 g/cm 3 ) and 0.825-0.925 g/cm 3 (mean 0.868 g/cm 3 ), respectively.
Stage III: For Group 5 FIA, the bulk densities of the type C1 and C2 inclusions are quite different, with ranges of 0.886-0.974 g/cm 3 (mean 0.922 g/cm 3 ) and 0.802-0.872 g/cm 3 (mean 0.845 g/cm 3 ), respectively. Nevertheless, bulk densities of type C1 inclusions in Group 4 FIA are 0.865-0.885 g/cm 3 (mean 0.875 g/cm 3 ), lower than the C1 inclusions in Group 5 FIA in stage III but generally equal to C1 inclusions in Group 2 FIA in stage II.
Stage IV: Bulk densities of W inclusions in Group 6 FIA are 0.935-1.012 g/cm 3 (mean 0.963 g/cm 3 ), obviously higher than the type C inclusions from the quartz of previous stages.

Content of CO 2 and CH 4
Both type M and C inclusions are able to assume the contents of CO 2 and CH 4 . For type M inclusions, the content of CO 2 + CH 4 varies from 30 to 80 mol%, carbonic X CH4 of are generally 70-85 mole% based on Raman data.

Nature of Ore-Forming Fluids
According to fluid-inclusion petrography, Laser Raman spectroscopy, and microthermometry studies, there are big differences in composition among each fluid inclusion type and FIA group. Each of the Group 1, 2, 4, and 6 FIAs comprise only one type of fluid inclusion, and they represent a homogeneous fluid (Table 1). Group 1 FIA, consists of type M CH 4 -rich fluid inclusions and represents a CH 4 -rich CH 4 -CO 2 -NaCl-H 2 O system in stage I, is characterized by medium temperatures (326-345 • C) and is CH 4 -rich (~20-70% mol). Group 2 and 4 FIAs consist of type C1 inclusions in stages II and III, respectively, belong to a NaCl-H 2 O-CO 2 -CH 4 system, with medium temperatures (296-318 • C) and medium salinity (9.1-10.1 wt% NaCl eq.), and are CO 2 -rich (~10% mol) with minor CH 4 (~5% in carbonic phase). While Group 6 FIAs in stage IV belong to a NaCl-H 2 O system, which is characterized by medium-low temperatures (196-239 • C) and relatively higher salinity (12.4-18.4 wt% NaCl eq.). Thus, the ore-forming fluids of Sanakham gold deposit might primarily belong to a medium temperature CH 4 -rich CH 4 -CO 2 -NaCl-H 2 O system, which evolved into a NaCl-H 2 O-CO 2 -CH 4 system in the main ore stage, and finally transformed into a medium-low temperature, medium-high salinity NaCl-H 2 O fluid.

Fluid Immiscibility Processes
As mentioned above, both Group 3 FIAs in stage II and Group 5 FIAs in stage III consist of fluid inclusions with significantly different composition. The coexistence of the CH 4 -rich and CO 2 -rich inclusions (Group 3 FIA) and C1 and C2 inclusions with obviously different contents of CO 2 (Group 5 FIA) can be explained by the following three processes [26]: (1) fluid immiscibility caused by phase separation of a homogeneous fluid; (2) fluid mixing caused by heterogeneous trapping of multi-source hydrothermal fluids; or (3) post-entrapment modification. In this study, undeformed hydrothermal quartz grains with euhedral to subhedral shape are preferentially selected for fluid inclusion analysis. The inclusions for microthermometry, moreover, have no evidence of necking down [27]. The post-entrapment modification process thus can be excluded in Sanakham.
Ramboz et al. [28] propose three lines of evidence which may indicate a fluid immiscibility process for the two contrasting types of inclusions: (i) the two types of inclusions must occur in the same regions and there must be good evidence of contemporaneous trapping; (ii) the two types of inclusions must have similar total homogenization temperature ranges, with one type of inclusion homogenizing to the aqueous phase, and the others homogenizing to the vapor phase; (iii) if one inclusion type decrepitates before homogenization, then the other type must behave similarly. With reference to the above principles, Group 5 FIA meets the criteria of fluid immiscibility, including (1) type C1 and C2 inclusions contemporaneously trapped in the same FIA (Figure 5k,l); (2) type C1 and C2 inclusions, respectively, homogenize to the aqueous phase (liquid) and CO 2 vapor phase, and have a similar T h TOT (272-322 • C and 278-339 • C; Table 4); (3) some coexisting C1 and C2 inclusions generally have moderate V/(L + V) ratios, decrepitate before homogenization, and are not recorded in the data sheet. In addition, type C1 inclusions have slightly higher salinities (7.6-12.6 wt% NaCl eq.) than type C2 inclusions (6.8-10.6 wt% NaCl eq. Figure 7c), because salt is preferentially fractionated into the aqueous phase during phase separation [29]. It is inferred that the C1 and C2 inclusions in Group 5 FIA are two endmembers of unmixed fluids of a medium temperature, medium salinity, and CO 2 -rich parent fluid (Figure 9). Considering the widespread Group 5 FIA in stage III quartz, fluid immiscibility was a dominant process for stage III mineralization. For Group 3 FIA in stage II quartz, the contemporaneously trapped type M and C1 inclusions also have similar homogenization temperature ranges (Figure 5i and Table 3), which are 288-355 • C and 290-337 • C, respectively. Except for the liquid phase, furthermore, type M and type C1 inclusions have entirely different compositions in their carbonic phase. According to the Raman results, CH 4 contents in the carbonic phase of type M inclusions are~80%, and the CO 2 contents are~20%. Meanwhile, the type C inclusions contain contrary amounts of CO 2 and CH 4 ( Figure 6), which can also be proven by their distinctly different T m carbon (Table 3). Thus, we cannot attribute the Group 3 FIA to a heterogeneous trapping of multi-source fluids led by fluid mixing [31], which prefers a continuum of compositions in diverse fluid inclusions. The type M and C1 inclusions are better suggested to be two endmembers of unmixed fluids that belong to a CH 4 -rich CH 4 -CO 2 -NaCl-H 2 O system. The fluid immiscibility process should also take part in stage II mineralization in Sanakham.

Implications for Gold Deposition
Given that the native gold in Sanakham is closely associated with sulfides such as pyrite, pyrrhotite, and chalcopyrite, the ore fluids are considered to be near-neutral to weakly acidic [32,33]. The ore-forming fluids are < 400 • C and enriched in CO 2 , indicating that the gold-transporting species were probably bisulfide complexes, mainly Au(HS) 2− [34,35]. Gold deposition, led by the destabilization of gold-bisulfide complexes, was likely associated with fluid immiscibility, fluid mixing, and other chemical changes [36,37].
Taking into consideration the fluid immiscibility process in stages II and III, the Group 3 and 5 FIAs are reliable for estimating the P-T conditions of gold mineralization. The final homogenization temperatures of type C1 inclusions, respectively, in Group 3 FIA (290-337 • C) and Group 5 FIA (272-322 • C) are considered to be close to the temperature of entrapment [20]. The P-T phase diagram ( Figure 10) shows a representative solvus for H 2 O-CO 2 fluids containing 10 wt% NaCl eq. and 10 mol% CO 2 (calculated after [30]), which is close to the composition of ore-forming fluids in stages II and III. The trapping pressures can be calculated from their isochores (0.825-0.925 g/cm 3 for stage II and 0.886-0.974 g/cm 3 for stage III) at the given homogenization temperatures. The estimated P-T conditions of gold mineralization in stages II and III ranged, respectively, from 134 to 65 MPa and 337 to 290 • C, and 236 to 75 Mpa and 322 to 272 • C ( Figure 10). During fluid immiscibility caused by multiple hydraulic fracturing and pressure pulsation, gold deposition is believed to occur in conditions between lithostatic and hydrostatic pressure [38]. The highest and lowest pressures can be used to determine the depth considering that they are produced by the lithostatic and hydrostatic pressures, respectively. The densities of host quartz monzodiorite and ore-forming fluids are assumed, respectively, to be 2.7 g/cm 3 and 1.0 g/cm 3 . Thus, for stage III, the calculated maximum and minimum depths are 8.7 km and 7.5 km, respectively. However, the highest pressure (134 MPa) in stage II indicates a depth of < 5.0 km, while the lowest pressure (65 MPa) indicates a depth of > 6.5 km. Taking the depth assumed from stage III into consideration, the depths for gold deposition in Sanakham are suggested to be 8.7-6.5 km. Fluid immiscibility in stage III occurred at relatively higher pressures than that in stage II, which might be attributed to a large release of CO 2 and subsequent pressure increase in the fracture system. Figure 10. Estimated P-T conditions for the formation of the quartz veins in stages II and III in the Sanakham deposit. Representative isochores for mean bulk densities, respectively, for type C1 inclusions, and the solvus for H 2 O-CO 2 fluids containing 10 wt% NaCl eq. and 10 mol% CO 2 (calculated after reference [30]). The blue and red dashed lines define the range of T h TOT and bulk density of type C1 inclusions in Groups 3 and 5 FIA, respectively. The dotted lines represent the range of trapping pressures. Except for Group 3 and 5 FIAs, which show evidence for fluid immiscibility, the homogenization temperatures of other primary fluid inclusions could only be regarded as the minimum temperature of entrapment. Combined with the above discussion, thus, the evolution of ore-forming fluids and related gold-deposition processes can be summarized below (as shown in Figure 11). The primary ore-forming fluids in stage I belonged to a CH 4 -rich CH 4 -CO 2 -NaCl-H 2 O system (Group 1 FIA), which is characterized by mediumhigh temperature (>345 • C) and a high content of CH 4 (~20-70% mol). The first fluid immiscibility process occurred in stage II and is represented by two coexisting endmembers (Group 3 FIA) including both those with high-CH 4 and low-CO 2 and the opposite. The ore fluids gradually evolved into a NaCl-H 2 O-CO 2 ± CH 4 fluid (Group 2 FIA) with medium temperature (~300 • C), medium salinity (9.1-10.1 wt% NaCl eq.), and a high content of CO 2 (~10% mol) during stages II and III. The second fluid immiscibility process occurred widely in stage III and therefore created two new coexisting endmembers (Group 5 FIA), which are the CO 2 -poor (~7% mol) and relatively salinity-rich (7.6-12.6 wt% NaCl eq.) ones, as well as the opposite. Due to the volatile loss caused by fluid unmixing, the ore fluids finally evolved into a medium-low temperature (>239 • C), medium-high salinity (12.4-18.4 wt% NaCl eq.) NaCl-H 2 O system in stage IV (Group 6 FIA) accompanied by simply cooling. The two important gold-deposition events were generated, respectively, along with fluid immiscibility processes during stages II and III mineralization. Compared with important gold deposits in the Luang Prabang-Loei metallogenic belt, the epithermal gold class can be excluded in Sanakham because there is neither continental volcanic rock exposure, nor characteristic kaolinite, alunite, or porous silicification. Even though quartz monzodiorite predominates in the wall rock, the hydrothermal metasomatic alteration, such as skarnization and horny alteration, has little spatial relation with gold mineralization, different from the regional representative porphyry-skarn copper-gold deposits [15,17]. Actually, the lode-gold orebodies in Sanakham are strictly controlled by the NE-NNE trending faults and crosscut the lithological interfaces between quartz monzodiorite and slate (Figure 2), which is similar to the Phapon epizonal orogenic gold deposit in the northern part of the Luang Prabang-Loei belt [13,14].
In most gold deposits, including porphyry and orogenic systems, the ore-forming fluids normally comprise remarkable contents of CO 2 which are closely related to the fluid immiscibility system. Additionally, it is not uncommon for methane to be included in the ore fluids, e.g., gold deposits from Jiaodong [39][40][41][42][43][44][45], Ailaoshan [46], West Qinling [47,48], and Southwest Guizhou [49] in China, the Abitibi gold belt in Canada [50], and Hamadi gold deposit in Sudan [51], as well as some porphyry Cu-Mo and rare-metal deposits [52][53][54][55][56][57][58][59][60][61]. Based on high P-T experimental results, Naden et al. [62] pointed out that the CH 4 -bearing fluid system has a much wider immiscible area than the CO 2 -bearing fluid system under the same P-T conditions. At 300 • C for example, the immiscibility process could occur when CO 2 content is higher than 15% mol, while the CH 4 content only needs to be above 5% mol [63]. Thus, a small amount of CH 4 in a NaCl-H 2 O-CO 2 system could greatly expand the immiscibility area and result in gold deposition, which is identical to the stage II mineralization in Sanakham. However, the CH 4 -rich CH 4 -CO 2 -NaCl-H 2 O ore fluids cannot be simply interpreted as products of alternative magmatic, metamorphic, or mantlederived fluids. It is generally assumed that the extremely high content of CH 4 has an organic source, including partial decomposition of organic matter, or a reaction between water and graphite or organic carbon in the formations [64]. The Upper Carboniferous carbonaceous slate, therefore, might be a possible source of the abundant CH 4 , even though more refined studies on the source and evolution of ore-forming fluids are further needed.

Conclusions
The Sanakham gold deposit, located in the Luang Prabang-Loei metallogenic belt, predominately consists of auriferous quartz vein-type orebodies that are controlled by NE-NNE trending faults within Middle Triassic quartz monzodiorite and secondary Upper Carboniferous low-grade metamorphic rocks. Detailed study through petrography, microthermometry, and Laser Raman spectroscopy on fluid inclusions in gold-related quartz were carried out in the Sanakham gold deposit. The following conclusions are drawn from this study on the evolution of ore-forming fluids and gold deposition mechanisms in Sanakham: (1) Four mineralization stages have been distinguished in Sanakham, of which the quartzpyrite-pyrrhotite-chalcopyrite stage (II) and the quartz-pyrite-base-metal sulfide stage (III) are identified as the main gold deposition stages. (2) The initial ore-forming fluids belonged to a medium-high temperature CH 4 -rich CH 4 -CO 2 -NaCl-H 2 O system, which gradually evolved into a medium temperature, medium salinity, and high CO 2 content NaCl-H 2 O-CO 2 ± CH 4 system during the main ore stages, and then a medium-low temperature, medium-high salinity NaCl-H 2 O system in stage IV. (3) Gold-deposition events were generated along with the two fluid immiscibility processes during stage II and III mineralization, respectively. The former created two coexisting high-CH 4 and low-CO 2 and low-CH 4 and high-CO 2 endmembers, while the latter created two coexisting CO 2 -poor and CO 2 -rich endmembers.