Fluid and metal sources in the Fäboliden hypozonal orogenic gold deposit, Sweden

To model the formation of orogenic gold deposits, in a global perspective, it is important to understand the ore-forming conditions not only for deposits hosted in greenschist facies rocks but also in amphibolite facies. The Paleoproterozoic Fäboliden deposit in northern Sweden belongs to the globally rare hypozonal group of orogenic gold deposits and, as such, constitutes a key addition to the understanding of amphibolite facies orogenic gold deposits. The Fäboliden deposit is characterized by auriferous arsenopyrite-rich quartz veins, hosted by amphibolite facies supracrustal rocks and controlled by a roughly N-striking shear zone. Gold is closely associated with arsenopyrite-löllingite and stibnite, and commonly found in fractures and as inclusions in the arsenopyrite-löllingite grains. The timing of mineralization is estimated from geothermometric data and field relations at c. 1.8 Ga. In order to constrain the origin of gold-bearing fluids in the Fäboliden deposit, oxygen, hydrogen, and sulfur isotope studies were undertaken. δ18O from quartz in veins shows a narrow range of + 10.6 to + 13.1‰. δD from biotite ranges between − 120 and − 67‰, with most data between − 95 and − 67‰. δ34S in arsenopyrite and pyrrhotite ranges from − 0.9 and + 3.6‰ and from − 1.5 and + 1.9‰, respectively. These stable isotope data, interpreted in the context of the regional and local geology and the estimated timing of mineralization, suggest that the sulfur- and gold-bearing fluid was generated from deep-crustal sedimentary rocks during decompressional uplift, late in the orogenic evolution of the area. At the site of gold ore formation, an 18O-enriched magmatic fluid possibly interacted with the auriferous fluid, causing precipitation of Au and the formation of the Fäboliden hypozonal orogenic gold deposit.


Introduction
Gold has been one of the most sought-after commodities for thousands of years. The global historical production is estimated to be approximately 180,000 tons and about one-third of that gold comes from orogenic gold deposits (Frimmel 2008). Orogenic gold deposits are considered hydrothermal in character, having formed from gold-bearing fluids that are focused into structural traps during metamorphism and deformation associated with accretional and collisional regimes at continental margins (Groves 1993;Groves et al. 1998Groves et al. , 2003McCuaig and Kerrich 1998;Hagemann and Cassidy 2000;Goldfarb et al. 2001Goldfarb et al. , 2005. In recent years the Crustal Continuum model outlined during the late 1980s to early 1990s (Colvine 1989;Groves 1993) has been challenged Powell 2009, 2010;Tomkins and Grundy 2009). It has been suggested that orogenic gold deposits cannot form at metamorphic conditions beyond midamphibolite facies due to the inhibited fluid flow that occurs at temperatures above approximately 600-650°C, and gold deposits hosted by rocks subjected to higher metamorphic conditions are assumed to either have formed pre-peak metamorphism and subsequently overprinted by peak metamorphic conditions (Tomkins and Grundy 2009), or to have formed under retrograde PT conditions (e.g., Kolb et al. 2015). This has led to a debate on whether all orogenic gold deposits in amphibolite and granulite facies (hypozonal group) are in fact metamorphically overprinted greenschist deposits (Phillips Editorial handling: K. Kelley andPowell 2009, 2010;Tomkins and Grundy 2009;Tomkins 2010;Kolb et al. 2015). More recently, evidence suggests that the hypozonal group is indeed post-peak metamorphic orogenic gold deposits (Kolb et al. 2015;Groves et al. 2019). In relation to this debate, the estimation of the timing of mineralization relative to peak metamorphism is vital to the understanding of the formation of orogenic gold deposits.
In northern Sweden, the Lycksele-Storuman ore province (commonly referred to as the Gold Line) is presently being explored (Fig. 1). The Gold Line is manifested by a roughly north-west trending linear anomaly of Au in the till overburden. During the past twenty years, a number of promising gold prospects with many features similar to orogenic gold deposits have been discovered in this area, and only a few papers have been published on the metallogeny of the Gold Line (e.g., Hart et al. 1999; Weihed 2003, 2007;Bark et al. , 2013; Bark and Weihed 2012;Schlöglova et al. 2013). To date, only one of the gold deposits in the Lycksele-Storuman area, the Svartliden gold deposit (3 Mt at 4.5 ppm Au; Dragon Mining 2005) has been mined (Fig. 1). Approximately 30 km from the Svartliden mine is the hypozonal Fäboliden orogenic gold deposit Weihed 2003, 2007). Fäboliden is presently the largest of the known gold prospects in the area, with previous measured and indicated mineral resources at 55 Mt at 1.0 ppm Au (Lappland Goldminers 2012). The current owner is planning to mine the richer parts of the deposit (10.6 Mt at 3 ppm Au, Dragon Mining 2016).
The geology and geochemistry of unaltered and altered lithologies of the Fäboliden deposit is described in detail by  as arsenopyrite-bearing quartz veins that occur within a high-angle shear zone in amphibolite facies rocks, mainly metagreywacke. Based on these observations, they considered the deposit to be a hypozonal orogenic gold deposit. In this contribution, stable isotope data (S, O, and H) of the mineralized quartz veins at Fäboliden are used to characterize the oreforming fluid, to delineate potential fluid source rocks, and to put the Fäboliden deposit into the larger Fig. 1 Bedrock map of the Lycksele-Storuman area. Coordinates in Swedish National Grid (RT90). Modified after  perspective in order to refine genetic models for the Fäboliden and similar deposits that will be valuable for future exploration. This is the first study of sulfur isotopes undertaken in the Lycksele-Storuman area. Stable isotope data, here interpreted in the context of the regional and local geology, and the estimated timing of mineralization, in combination with previously published fluid inclusion data , suggest that the Fäboliden gold deposit was formed from mixing of metamorphic and 18 O-enriched magmatic fluids during rapid uplift and erosion, after peak metamorphism.

Fäboliden orogenic gold deposit
The steeply dipping sheet-like deposit at Fäboliden extends for about 1.3 km along strike (Fig. 2), is up to 50 m wide, and has been observed at depths of over 350 m. The mineralized zone is open at depth.

Geology of the Fäboliden area
The main rock type in the Fäboliden area (Fig. 2) is the late-to post-orogenic, 1.80-1.77 Ga, Revsund granitoid (Claesson and Lundqvist 1995;Billström and Weihed 1996;Weihed et al. 2002a). This medium-to coarse-grained porphyritic, isotropic granitoid surrounds a narrow belt of Bothnian Basin metagreywackes intercalated with mafic metavolcanic rocks. The poorly exposed fine-grained and biotite-rich metagreywackes are strongly foliated. In a few outcrops, the metagreywackes are more coarse-grained (< 1 cm grain size) and less deformed. In less deformed areas, primary sedimentary textures such as stratification and bedding are observed. The fine-grained and banded metavolcanic rocks dominate the northern parts of the Fäboliden area.
The granitic and supracrustal rocks in the Fäboliden area are crosscut by a set of c. 1.26 Ga dolerites (Söderlund et al. 2006). The dolerites are not mineralized, and clearly post-date the gold mineralization at Fäboliden.

Metamorphic conditions
The metasedimentary rocks in the northern part of the Bothnian Basin, the Lycksele-Storuman area ( Fig. 1), have been metamorphosed to amphibolite facies (Weihed et al. 1992), and locally to granulite facies (Lundström 1998). Regional metamorphic PT conditions of 3-5 kbar and 550-700°C were suggested by Lundqvist (1990). Geothermometric data from supracrustal rocks indicate peak metamorphic temperatures of 570-640°C (garnetbiotite) for the Fäboliden area Weihed 2007, 2012), suggesting amphibolite facies conditions. Peak metamorphism and known ductile deformation events in the northern part of the Bothnian Basin occurred between 1.84 and 1.80 Ga (Weihed et al. 1992;Billström and Weihed 1996;Bergman Weihed 2001;Rutland et al. 2001;Weihed et al. 2002b). Based on aeromagnetic data that show discrete shear zones, denoted as D 3 -structures, and their field relations with intrusive rocks, Rutland et al. (2001) inferred that the last of the ductile events occurred prior to 1.81 Ga. However, in the Fäboliden deposit area, located west of the study area of Rutland et al. (2001), discrete ductile shear zones interpreted as regional D 3structures host the gold within the supracrustal rocks. These shear zones at Fäboliden are seen to affect the margin of the 1.80-1.77 Ga (Claesson and Lundqvist 1995;Billström and Weihed 1996;Weihed et al. 2002a) Revsund granite in a ductile manner (Bark and Weihed 2003). The ductile fabric, which is sub-parallel to the contact between the metagreywacke and the granite, progresses laterally across the contact where it gradually weakens and disappears a few meters into the granite. These relationships contradict Rutland et al. (2001) and clearly show that the margin of the granite was affected by ductile deformation, requiring the last known ductile event to be c. 1.80 Ga.  discussed the relationships between garnet-biotite and graphite geothermometers from the Fäboliden area, representing peak metamorphic temperatures (garnet-biotite) and temperatures of gold-associated hydrothermal alteration (graphite grains spatially associated with arsenopyrite in the gold-bearing quartz veins). Graphite is a good indicator of temperature conditions, since the process of graphitization of carbonaceous material (CM) in metasedimentary rocks during hydrothermal alteration is considered irreversible and shows no known effects on the crystal structure during retrograde cooling of the rocks (Pasteris and Wopenka 1991; Beyssac et al. 2002). Graphite in the Fäboliden veins is suggested to have originated from the reaction CO 2 + CH 4 -> 2C + 2H 2 O, during mineralization . Although graphite thermometry studi e s t y p i c a l l y f o c u s o n a n a l y s e s o f C M f r o m metasedimentary rocks and not vein-hosted graphite, Beyssac et al. (2002) have shown that graphite thermometry is less sensitive to different CM precursor types at temperatures between 330 and 650°C. Thus, since graphite data from Fäboliden suggest a narrow range of peak temperatures (520-560°C) for the graphitization and associated hydrothermal alteration, considerably lower than peak metamorphic temperatures ( 5 7 0 -6 4 0°C ) i n d i c a t e d f r o m g a r n e t -b i o t i t e geothermometers ) the use of graphite thermometry in veins at Fäboliden is justified. Assuming that the geological environment at Fäboliden did not act as a closed system after the mineralizing event, the graphite temperature range suggests that gold mineralization occurred post-peak metamorphism; otherwise, the graphite geothermometer would have indicated temperatures matching peak metamorphic conditions. The field observations of the ductile fabric (shearing) that passes from the metagreywacke into the granite, together with the relationship between the different geothermometers, suggests that the gold mineralization at Fäboliden is considered post-peak metamorphic and contemporaneous with the emplacement of the 1.80-1.77 Ga Revsund granite.

Gold mineralization
The gold at Fäboliden is contained in quartz and sulfide veins, which parallel the main foliation within the shear zone in the supracrustal host rocks (Fig. 2). The metagreywacke sequences are the main host for the veins, which are variably boudinaged. In places, the metavolcanic rocks are also mineralized. Gold has not been observed in the Revsund granite.  Björk and Kero (2002) The sulfides are situated in semi-ductile structures, i.e. in thin veins parallel to the foliation planes and in the necks of boudinaged quartz veins indicating that the timing of sulfide crystallization is syn-to late-deformation Weihed 2003, 2007).
Sulfides proximal to the mineralization comprise mainly arsenopyrite, löllingite (commonly as inner cores in arsenopyrite; Fig. 3b) and pyrrhotite with accessory chalcopyrite, sphalerite, stibnite, and galena, whereas distal to the mineralization, pyrrhotite is the dominant sulfide and Asbearing minerals are virtually absent. Pyrrhotite occurs throughout the metagreywacke sequence at Fäboliden and is interpreted to have formed prior to the mineralization event. However, in the proximal alteration zone, pyrrhotite commonly occurs as fracture fillings within the arsenopyrite, suggesting that arsenopyrite crystallized prior to pyrrhotite. This paragenetic relationship is also indicated from fluid inclusion analysis where PT conditions are significantly different between fluid inclusions associated with arsenopyrite, at 4 kbar, compared with 0.3 kbar for pyrrhotite ). For example, these contradictory textural relationships between arsenopyrite and pyrrhotite might be an effect of re-mobilization of early-stage pyrrhotite during gold mineralization, reflected by late pyrrhotite textures.
Gold is typically very fine-grained (below 10 μm), and occurs mainly as inclusions and as fissure fillings in the arsenopyrite-löllingite (Fig. 3), but also as free grains in the silicate matrix of the host rocks Weihed 2003, 2007). Gold typically occurs as electrum (Au/Ag 2:1), but is also closely associated with stibnite, as auriferous stibnite.

Hydrothermal alteration
The supracrustal rocks in the northern parts of the Bothnian Basin constitute metagreywacke sequences with subordinate, mainly mafic, metavolcanic rocks, and are metamorphosed to amphibolite facies (Weihed et al. 1992;Lundström 1998). The mineral assemblage in the metagreywackes is characterized by Ca-and Fe-Mg-rich amphiboles together with hedenbergite, biotite, quartz, plagioclase, and potassium feldspar. Pyrrhotite is a common constituent together with trace amounts of chalcopyrite, sphalerite, and galena.
The transition from regional metamorphic mineral assemblages to hydrothermal alteration assemblages is gradual. Lower amphibole contents in the distal alteration zone compared with those in regional metamorphic rocks is the only discernible feature that distinguishes these rocks. This style of vague distal alteration is a common feature for hypozonal orogenic gold deposits globally (Eilu et al. 1999).
Proximal to the mineralized zones, the metagreywacke displays an intense compositional banding due to variations in the abundances of biotite, amphibole, and pyroxene. Quartz veins 1-5 cm thick are common in these proximal alteration zones. Alteration zones enveloping the veins are comprised of diopside, calcic amphibole, biotite, potassium feldspar, plagioclase, minor andalusite, and very fine-grained tourmaline . Compared with regional metamorphic mineral assemblages, proximal alteration zones contain Electron backscatter images of gold closely associated with arsenopyrite-löllingite. a Gold and probable stibnite in fractures within arsenopyrite-löllingite. b Electron microprobe mapping image (intensity of As) showing arsenopyrite with common löllingite cores. Apy, arsenopyrite; Au, gold; Löll, löllingite; Sb, probable stibnite more calcium-rich minerals, manifested by a higher Ancontent of plagioclase. Higher An-content plagioclase, together with the presence of diopside and calcic amphiboles, is positively correlated with higher Au grade. There are also thinner mineralized lenses parallel to higher grade veins, termed satellite veins, that show intense diopside-amphibolebiotite alteration (Fig. 2, see cross section).

Analytical methods
Quartz and biotite from the quartz vein system were analyzed for oxygen and hydrogen isotopes. Arsenopyrite and pyrrhotite from the vein system were analyzed for sulfur isotopes. In total, six drill cores were sampled at 35 different sites (for distribution of samples, see Fig. 2). Fifty-two arsenopyrite, 17 pyrrhotite, 14 quartz, and 7 biotite grains were analyzed in this study. Samples were crushed and sieved to 75 μm. Minerals were separated using magnetic separation in two steps, followed by handpicking of targeted minerals under stereographic microscope. Isotopic analysis was subsequently performed at the Scottish Universities Environmental Research Centre (S.U.E.R.C.) in Glasgow, Scotland.

Oxygen isotope analysis
All separates were analyzed using a laser fluorination procedure, involving total sample reaction with excess ClF 3 using a CO 2 laser as a heat source (in excess of 1500°C; following Sharp 1990). All combustion resulted in 100% release of O 2 from the silica lattice. This O 2 was then converted to CO 2 by reaction with hot graphite, then analyzed online by a VG SIRA 10 spectrometer. Reproducibility is better than ± 0.3‰ (1σ). Results are reported in standard notation (δ 18 O) as per mil (‰) deviations from the Standard Mean Ocean Water (V-SMOW) standard, in Table 1.

Hydrogen isotope analysis
Pure biotite samples were heated to 150°C overnight under high vacuum to release labile volatiles after loading into thoroughly outgassed Pt crucibles. Samples were then gradually heated by radiofrequency induction in an evacuated quartz tube, to temperatures in excess of 1200°C. The released water was then reduced to H 2 in a chromium furnace at 800°C (Donnelly et al. 2001), with the evolved gas measured quantitatively in a Hg manometer, before collecting using a Toepler pump. The gas was subsequently analyzed on a VG 602D mass spectrometer with a manual Hg, high gas-compression inlet system. Replicate analyses of water standards (international stds V-SMOW and GISP, and internal standard Lt Std) gave a reproducibility of ± 2‰. Replicate analyses of international mineral standard NBS-30 (biotite) of δD = − 65‰ (V-SMOW) also gave reproducibility around ± 2‰. Analytical results are summarized, as δD notation as per mil (‰) variations from the international V-SMOW standard, in Table 1.

Sulfur isotope analysis
Sulfide separates (arsenopyrite and pyrrhotite) were analyzed by conventional techniques (Robinson and Kusakabe 1975) in which SO 2 gas was liberated by combusting the sulfides under vacuum with excess Cu 2 O at 1075°C. In addition to the conventional analysis, a number of in situ laser combustion analyses were carried out on polished blocks of the respective sample, following the technique described in Wagner et al. (2004). This allowed a spatial resolution of sulfur isotopes to be undertaken (McConville et al. 2000). All values in per mil Liberated gases from both techniques were analyzed on a VG Isotech SIRA II mass spectrometer. Calculation of δ 34 S values from raw machine δ 66 SO 2 data was carried out by calibration with international standards NBS-123 (+ 17.1‰) and IAEA-S-3 (− 31.5‰), as well as SUERC's internal lab standard CP-1 (− 4.6‰). Reproducibility was better than ± 0.2‰ (1σ). Data are reported in δ 34 S notation as per mil (‰) variations from the Vienna Cañon Diablo Troilite (V-CDT) standard. The analytical results, for conventional and laser-combusted sulfur analysis, are summarized in Tables 2 and 3. The majority (about 80%) of sulfur V-CDT Vienna Cañon Diablo Troilite 1 All values were corrected from raw δ 34 S values, using the laser fractionation factor determined by Wagner et al. (2004) whereby; δ 34 S true = δ 34 S measured + 0.4, which is the same as that calculated for pyrrhotite by Maynard et al. (1997). Shaded rows indicate core and rim analysis in one and the same grain isotope analyses were made on arsenopyrite since the arsenopyrite is considered to be associated with the precipitation of gold at Fäboliden ).

Oxygen and hydrogen isotope data
Results of oxygen and hydrogen isotope analysis are presented in Table 1. The oxygen isotope compositions of quartz from the auriferous veins at Fäboliden are homogeneous and range between + 10.6 and + 13.1‰. The mean value for the quartz data is + 11.8 ± 0.6‰ (1σ, n = 14). Oxygen isotope compositions of biotite samples show a narrow δ 18 O range of + 7.3 to + 8.0‰. Hydrogen isotope (δD) compositions range from − 120 to − 67‰ (Table 1), with the majority of compositions between − 95 and − 67‰. The mean δD value is − 88 ± 18‰ (1σ, n = 7).
A higher spatial resolution compared with conventional methods was gained by analyzing individual arsenopyrite and pyrrhotite grains in situ using laser combustion techniques. Most analyses were of arsenopyrite because it is temporally associated with gold at Fäboliden ), but some pyrrhotite was also analyzed. In total, 23 arsenopyrite and 6 pyrrhotite samples were analyzed. In each sample, one mineral grain was analyzed, but the core and rims of 7 arsenopyrite and 1 pyrrhotite grains were also analyzed. The rims and cores of the minerals were analyzed to test within-grain sulfur isotope compositional variations ( Table 2). The arsenopyrite cores yielded δ 34 S values from −0.8 to +2.3‰ (n = 19), whereas the rims ranged between 0 and + 3.6‰ (n = 11; Fig. 5b). The mean and 1σ standard Fig. 4 Diagram showing the range in calculated fluid oxygen (quartz and biotite) and hydrogen (biotite) isotope compositions. The bold red bar in the center of the symbol indicates mean oxygen and hydrogen values, in the temperature range 520 and 560°C. Fractionations for δD after Suzuoki and Epstein (1976), and for δ 18 O quartz-fluid after Matsuhisa et al. (1979). Metamorphic and magmatic fields after Taylor (1974). Orogenic gold field from Ridley and Diamond (2000) deviation of these data are 1.1 ± 0.9‰ and 1.9 ± 1.0‰ for core and rim, respectively. Thus, there is no statistically significant isotopic variation between rim and core in the arsenopyrite data. A similar lack of variation is suggested by analysis of pyrrhotite, yielding a δ 34 S range in cores of + 0.8 to + 1.5‰ (n = 4), and analysis of the rims ranging between 0 and + 1.9‰ (n = 3). The marked isotopic homogeneity of the dataset is evident in the combined vein sulfide data, which gives a mean of 1.1 ± 1.0‰ for arsenopyrite, and 0.4 ± 0.9‰ for pyrrhotite.
In an attempt to characterize background values of δ 34 S, three weakly altered pyrrhotite samples from about 100 m outside the proximal alteration zone yielded a δ 34 S range of − 3.3 to + 0.8‰. These results overlap the δ 34 S values from the proximal alteration zone but may indicate a slight shift towards more negative values for pyrrhotite in regional metagreywacke. However, further analysis is required to define a more representative sulfur isotope range for the unaltered metagreywackes in the Fäboliden area.

Discussion
Metamorphic fluids are widely accepted to be important in the formation of orogenic gold deposits. For example, metamorphic fluids leached metals from the country rocks and were enriched in gold deposits in the Otago Schist Belt, New Zealand, indicating metal transport by metamorphic fluids (Pitcairn et al. 2006). Metamorphic fluids have also been implicated in forming many other gold deposits (Kerrich et al. 2000;Goldfarb et al. 2005;Pitcairn et al. 2006Pitcairn et al. , 2010Large et al. 2011;Phillips and Powell 2010;Tomkins 2010Tomkins , 2013Gaboury 2013;Groves and Santosh 2016;Goldfarb et al. 2017), but orogenic gold deposits have also been suggested to have formed from magmatichydrothermal fluids (Burrows et al. 1986;Pattrick et al. 1988;Burrows and Spooner 1989;de Ronde et al. 2000;Xue et al. 2013), from deep-convection of meteoric water (Nesbitt et al. 1989;Hagemann et al. 1994;Jenkin et al. 1994), from mantle-derived CO 2 -rich fluids (Cameron 1988(Cameron , 1989Colvine 1989), or from fluids expelled during subduction of oceanic crust (Kerrich and Wyman 1990;Goldfarb et al. 1991a;Jia et al. 2003). Mixing of different fluids has also been suggested to be important in causing precipitation of gold in orogenic deposits (Anderson et al. 1992;Bateman and Hagemann 2004;Hill et al. 2013;Molnar et al. 2016;Yardley and Cleverley 2015;Shen et al. 2016;Spence-Jones et al. 2018), whereas some authors dismiss fluid mixing as a significant process in the formation of orogenic gold systems (Goldfarb and Groves 2015). The ultimate fluid source for orogenic gold deposits remains controversial and the fluid may have multiple origins (Goldfarb and Groves 2015;Yardley and Cleverley 2015;Groves and Santosh 2016;Groves et al. 2019).

Origin of fluid
The calculated fluid oxygen and hydrogen isotope compositions of quartz in the Fäboliden deposit plot in the overlapping fields of metamorphic and magmatic fluids (Fig. 4), suggesting that the hydrothermal fluid at Fäboliden originated from a deep-seated metamorphic and/or magmatic source. The range in oxygen and hydrogen compositions is similar to many other orogenic gold deposits (i.e. McCuaig and Kerrich 1998;Hagemann and Cassidy 2000;Ridley and Diamond 2000;Groves et al. 2003).
Oxygen isotopes from the Fäboliden deposit range from + 10.6 to + 13.1‰ (δ 18 O quartz ). Oxygen isotope compositions greater than + 8‰ are interpreted as being the result of surface or near-surface processes during sedimentation, diagenesis, or low-temperature hydrothermal alteration (Taylor 1980), suggesting that the fluid cannot be solely magmatic (McCuaig and Kerrich 1998). Also, if the source rocks contain 18 Oenriched metasedimentary rocks, a metamorphic fluid with δ 18 O values greater than + 8‰ is possible (Böhlke and Kistler 1986;McCuaig and Kerrich 1998  values from the Fäboliden gold-bearing metasedimentary host rocks may thus suggest a purely metamorphic fluid source, or a mixed fluid source. In the Lycksele-Storuman area (Fig. 1), few stable isotope studies have been undertaken. Wilson et al. (1985) presented oxygen isotope data (n = 5, δ 18 O SMOW ranging from + 9.2 to + 11.8‰) from two different granitic suites in the northern part of the Bothnian Basin, the Revsund type and the less common Skellefte-Härnö type, and concluded that there is assimilation of pelitic components in both granites but the Skellefte-Härnö type typically has the highest degree of incorporated metasedimentary material. Based on geothermometry and field relations, the mineralizing event at Fäboliden is estimated to be broadly coeval with emplacement of the spatially proximal 1.80-1.77 Ga Revsund granite (Claesson and Lundqvist 1995;Billström and Weihed 1996;Weihed et al. 2002a) . The Skellefte-Härnö type granite is not known to occur in the Fäboliden area.
Orogenic gold mineralization at about 1.8 Ga is a global phenomenon (Goldfarb et al. 2001) and an alignment of orogenic gold deposits of assumed similar age occurs in the Fennoscandian Shield both in Sweden and Finland (Saalmann et al. 2009;Bark and Weihed 2012). This suggests that widespread gold-bearing fluids percolated through the Fennoscandian crust at this time. Owing to the fact that the oxygen isotope values of the Revsund granite are interpreted to be the result of assimilation of metamorphosed sedimentary country rocks (Wilson et al. 1985), a Revsund-related magmatic fluid would likely resemble the isotopic oxygen signature of the metamorphosed sedimentary host rocks. It is common that magmatic fluids are present in hydrothermal systems, but the magmatic isotope signature may be strongly overprinted or erased during subsequent events, obscuring the original compositions (Rye 1993;Hedenquist and Lowenstern 1994;Hoefs 2009;Pili et al. 2011;Yardley and Cleverley 2015;Lüders et al. 2015;Goldfarb et al. 2017).
Therefore, a metamorphic origin of the gold-mineralizing fluid at Fäboliden seems plausible but a magmatic component to the fluid cannot be ruled out.

Origin of sulfur
Gold is likely transported as bisulfide complexes in metamorphic fluids that produced orogenic gold deposits (Groves et al. 2003;Goldfarb et al. 2005), especially at high pressure and temperature conditions (550-725°C and 1-4 kbar; Loucks and Mavrogenes 1999). Therefore, the source of sulfur is an important factor in developing a genetic model.
The remarkable homogeneity of sulfur isotope compositions for all vein sulfide samples, showing a mean value of + 1.1 ± 1.0‰ for arsenopyrite, and + 0.4 ± 0.9‰ for pyrrhotite, suggests that the sulfur likely originated from one uniform source reservoir. A narrow range of sulfur isotope values that cluster around 0‰ was previously considered consistent with a magmatic source for the sulfur (Ohmoto and Rye 1979). However, the local host rock can have a significant effect on the isotope signature (Ohmoto and Goldhaber 1997).
The dominant host rock at Fäboliden is a thick sequence (> 10 km) of schistose metagreywacke, with carbonate-rich parts, intercalated with metavolcanic units . The role of black shales, diagenetic pyrite, and orogenic gold formation in sedimentary host rocks has been illustrated in a number of studies (Large et al. 2007(Large et al. , 2009Thomas et al. 2011;Gaboury 2013). The model assumes that diagenetic pyrite is later transformed into pyrrhotite during metamorphism and deformation, releasing the gold from the pyrite crystal structure and concentrating the gold during orogenic processes, forming gold deposits. Carbonaceous shales are a suitable source rock for elements that are commonly enriched in orogenic gold deposits, such as S, Au, As, and Sb Pitcairn 2011;Steadman et al. 2014;Lisitsin and Pitcairn 2015).
Sedimentary rocks commonly show a wide range (+ 50 to − 40‰) in initial sulfur isotopic composition (Fig. 6) (Ohmoto and Rye 1979;Ohmoto 1986;Ohmoto and Goldhaber 1997;Seal 2006;Hoefs 2009). A wide range of compositions suggests either a mixture of sulfur from different sources under reduced conditions, or precipitation of sulfur from a single source under more oxidizing conditions (Ohmoto and Rye 1979). However, subsequent geological processes can modify the initial isotopic signatures. Metamorphism is commonly assumed to homogenize isotope compositions of sulfides through recrystallization, alteration of pyrite to pyrrhotite and sulfur, and diffusion at increased temperatures (Cook and Hoefs 1997). However, isotopic homogenization is commonly restricted to specific parts of the deposit and is controlled by local conditions such as focused fluid flow and tectonic framework (Cook and Hoefs 1997). Chang et al. (2008) showed that variable sulfur isotope compositions in rocks hosting the Sukhoi Log orogenic gold deposit in Russia were homogenized during metamorphism and peak deformation.
Compilations of sulfur isotope data from a large number of deposits show a common range of δ 34 S compositions from − 3 to + 9‰ for orogenic lode gold deposits globally (Steed and Morris 1997;McCuaig and Kerrich 1998;Ramsay et al. 1998;Ridley and Diamond 2000;Hodkiewicz et al. 2009). Fäboliden sulfur isotope data clearly fall within this range, with δ 34 S between −1.5 and + 3.6‰ (Fig. 5), a narrow range that is similar to other hypozonal orogenic gold deposits, such as the giant Champion deposit in the Kolar Schist Belt, India (Hagemann and Cassidy 2000) and the Wiluna, Morning Star, Princess Royal, and Mt. Charlotte deposits in Australia (McCuaig and Kerrich 1998). This range of δ 34 S in the gold-associated sulfide minerals has been interpreted to originate from common sulfur reservoirs such as mantle-derived magmatic rocks, average crustal sulfur, or metamorphosed sedimentary sulfur (Ridley and Diamond 2000;Chang et al. 2008).
A narrow range of sulfur isotope data clustering around 0‰ may suggest a single uniform source for the sulfur (and by inference gold). Based on fluid inclusion data,  showed that an extremely CO 2 -rich fluid (75-95 mol%) was involved in the formation of the Fäboliden deposit. Large amounts of CO 2 are generated during metamorphism (Goldfarb et al. 2017), and carbonate-rich sedimentary rocks, such as the metagreywacke sequence at Fäboliden would result in CO 2 -rich ore-related fluids. This would indicate that the fluid originated from sedimentary rocks beneath the site of ore formation, as is suggested for other hypozonal orogenic gold deposits elsewhere (Kolb et al. 2015). Gaboury (2019) concludes that the fluid, ligand, and metal (gold) involved in orogenic gold mineral systems are all sourced from the same carbonaceous and pyrite-rich sedimentary rock, at conditions transitional between greenschist and amphibolite facies. The Fäboliden data are consistent with this model.
Overall, sulfur, oxygen, and hydrogen isotopic data interpreted in the context of the regional and local geology suggests that the mineralizing fluid at Fäboliden originated from metamorphic devolatilization of deep-seated sedimentary rocks, but there is a possibility that a magmatic fluid triggered precipitation of metals through mixing with the goldbearing metamorphic fluid.

Proposed genetic model
Models for orogenic gold systems suggest that the fluid is generated during prograde metamorphism at the greenschistamphibolite facies transition. Fluids and sulfur are released from the metamorphosed rocks during the breakdown of hydrous silicates and pyrite (Pitcairn et al. 2006;Tomkins 2010;Finch and Tomkins 2017) and, if the elemental composition of the source rock is favorable (e.g., carbonaceous sedimentary rocks; Vilor 1983;Large et al. 2007Large et al. , 2009Large et al. , 2011Gaboury 2013Gaboury , 2019, the fluid has the potential of forming orogenic gold deposits, assuming the fluid is focused through a structural framework in the crust with subsequent precipitation of gold over a geologically short time period. The amount of time required to generate an orogenic gold deposit obviously varies with size and gold concentration of the deposit, but the actual duration of the mineralizing event is suggested to be geologically short (<1 m.y.; Goldfarb et al. 1991b; 10,000 to 100,000 years; Weatherley and Henley 2013;Cox 2016), compared with the overall process of regional metamorphism (Yardley and Cleverley 2015). Exceptions are suggested for giant deposits, which may take several tens of millions of years to form (e.g., Golden Mile; Bateman and Hagemann 2004).
The generation of an ore fluid during prograde metamorphism as described above seems like a valid scenario for deposits in greenschist facies, and possibly also at lower amphibolite facies, but the model has difficulty explaining orogenic gold formation in metamorphic terranes at higher PT conditions (Kolb et al. 2015;Groves et al. 2019), or in deposits that clearly post-date peak metamorphism. Based on geothermometric data, mineralization at Fäboliden took place post-peak metamorphism, during retrograde conditions , requiring an alternative model for fluid generation.
The average gold concentration in the upper crust is typically between 0.5 and 5 ppb (Wedepohl 1995;Laznicka 1999;Rudnick and Gao 2003;Pitcairn 2011), and thus, scavenging of gold by metamorphic fluids requires a large volume of crust to generate a hydrothermal gold deposit. Yardley and Cleverley (2015) argued that fluid production during regional metamorphism is a slow process because heat transfers through insulating materials, and there are two possible ways of generating a metamorphic fluid post-peak metamorphism: Fig. 6 Sulfur isotope ranges of some common reservoirs. Data compiled from Steed and Morris 1997, McCuaig and Kerrich 1998, Ramsay et al. 1998, Ridley and Diamond 2000, Hodkiewicz et al. 2009, and Hoefs 2009 (1) the introduction of post-orogenic magmas that supply heat, generating a metamorphic fluid from dehydration of the metamorphosed rocks, or (2) fast dehydration of the crust during post-orogenic decompression due to rapid uplift and erosion. Metamorphic fluids and associated gold would, in the latter case, be generated as consequences of crustal decompression during uplift that is commonly associated with a shift in the far-field stress regime due to a transition from compressional to transpressional settings (Groves et al. 1987;Goldfarb et al. 1991a;Elmer et al. 2006;Craw et al. 2010;Vry et al. 2010;Yardley and Cleverley 2015;White et al. 2015).
At Fäboliden, a significant decrease in pressure, from 4 to 0.3 kbar, during late stages in the fluid evolution at Fäboliden is suggested by . This pressure drop may indicate rapid uplift during rotation of the regional stress regime (Weihed et al. 2005;Bark and Weihed 2012) and subsequent erosion during post-orogenic tectonic settings.
However, coeval to the mineralization at Fäboliden, large volumes of magmatic rocks were emplaced proximal to the site of gold deposition , supplying heat and possibly fluids. Our isotopic data suggest that an input of a magmatic source enriched in 18 O from assimilation of sedimentary rocks during emplacement cannot be ruled out. However, the sulfur isotope data coupled with fluid inclusion data ) suggests significant input of metamorphic fluid.  suggested that gold formation at Fäboliden was the result of possible mixing of two fluids (a CO 2 ± CH 4 fluid and a H 2 S ± CH 4 fluid). Phase separation is a widespread mechanism for precipitating gold from the hydrothermal fluid in orogenic gold systems (Ridley and Diamond 2000;Heinrich 2007), and the mechanism is likely triggered by fluctuations in lithostatic pressure during progressive deformation (e.g., the fault-valve model; Sibson et al. 1988;Cox et al. 1995). At Fäboliden, no evidence for phase separation (such as boiling) has been detected ). This may be due to the limited data set in the fluid inclusion study (n = 14). However, the CH 4 -component in the fluid is a strong reductant and would likely cause precipitation of gold from the fluid in case of fluid mixing (Heinrich 2007).

Conclusions
The Fäboliden orogenic gold deposit was formed from a metamorphic fluid, that provided the sulfur (and gold), generated during retrograde metamorphism of deep-crustal sedimentary rocks during decompressional uplift. The metamorphic fluid is suggested to have mixed with an 18 O-enriched magmatic fluid, causing precipitation of Au and the formation of the Fäboliden hypozonal orogenic gold deposit. Since the Fäboliden deposit is hosted by amphibolite facies rocks, and formed post-peak metamorphism, this contribution adds important data to the understanding of the formation of orogenic gold deposits in high-metamorphic terrains.
The authors obviously recognize that using only stable isotope data in constraining an ore model is limiting; however, all interpretations made from isotope data in this paper are corroborated with already published data on geochemistry, mineralogy, petrography, and fluid inclusion data from the Fäboliden deposit.
Acknowledgments Karen Kelley and Bernd Lehmann are thanked for constructive editorial work, and two anonymous reviewers significantly improved the manuscript. The authors would also like to thank the lab staff at S.U.E.R.C. for excellent support.
Funding information Open access funding provided by Lulea University of Technology. This study is funded by Georange, being part of the Georange research project 89126: Genesis and tectonic setting of Aulode in the new Lycksele-Storuman ore province. S.U.E.R.C. is funded by NERC and the consortium of Scottish Universities. AJB is funded by the NERC support of the Isotope Community Support Facility at S.U.E.R.C.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.