Introduction

Spreading rates of oceanic ridges are commonly considered as superfast (> 140 mm yr−1), fast (80–140 mm yr−1), intermediate (55–80 mm yr−1), slow (20–55 mm yr−1), or ultraslow (< 20 mm yr−1) (Dick et al. 2003 and references therein; Beaulieu et al. 2013). Since longitudinal profiles of ridges are closely associated with the intensity of magmatism (Standish and Sims 2010), and hydrothermal activity is significantly correlated with spreading rate (Baker et al. 1996; Baker 2009), previous studies considered that hydrothermal fields in oceanic ridges are predominantly controlled by the eruption and/or intrusion of basaltic magmas (Hannington et al. 2005; Fouquet et al. 2010). This results in most sulfide deposits being hosted in basalts, especially at fast-spreading ridges, with few being hosted in ultramafic rocks. However, in fast-spreading ridges, hydrothermal circulation may be blocked by such eruptions, resulting in abundant small-scale sulfide deposits (Fornari et al. 1998). Contrarily, sulfide deposits and vent sites in slow-spreading ridges tend to be much larger than those found in classic magmatic settings (German et al. 2016); in most cases, they are hosted in long-lived detachment faults (leading to oceanic core complex formation with exhumation of ultramafic rocks). This allows the formation of relatively stable and long-lived fluid pathways, and also the involvement of serpentinization reactions within the hydrothermal circulation cells (Hannington et al. 2005; McCaig et al. 2007; Ildefonse et al. 2007; Escartin et al. 2008; Fouquet et al. 2010).

Hydrothermal activity in slow-spreading ridges may be extensive (German and Parson 1998; Münch et al. 2001; Baker et al. 2004; Nayak et al. 2014), even in ultraslow ridges, where it was considered to be absent because of scarce magmatic activity (Baker et al. 1996). Tao et al. (2012) found an active black smoker chimney (Longqi field; 37.8° S, 49.6° E) in the ultraslow-spreading Southwest Indian Ridge (SWIR), showing that sulfide deposits develop in such environments, with hydrothermal activity being associated with localized magmatic activity and stable permeability. More hydrothermal fields have since been discovered near Longqi, including the Yuhuang (49.3° E), Xilongjing (49.7° E), Duanqiao (50.4° E), Changbai (~ 51.0° E), and Zhanqiao (~ 51.0° E) fields (Tao et al. 2014; Yang et al. 2017b; Liao et al. 2018). An important feature of ultraslow-spreading ridges is that they can be considered in terms of magmatic and amagmatic segments (Sauter et al. 2009; Standish and Sims 2010; Li et al. 2015; Yang et al. 2017a). The extensive occurrence of basalts suggests that SWIR fields, including the previously discovered inactive Mt. Jourdanne field at 63° 56′ E, are most likely associated with magmatic activity (Münch et al. 2001; Nayak et al. 2014; Tao et al. 2014; Yang et al. 2017b; Liao et al. 2018; Yuan et al. 2018).

Sulfide deposits associated with ultramafic rocks along the slow-spreading Mid-Atlantic Ridge (MAR) appear to be more abundant than in ophiolites on land (Fouquet et al. 2010), with ultramafic-hosted volcanogenic sulfide deposits being a specific marine mineralization type occurring along the MAR in the Logatchev (Bogdanov et al. 1997), Rainbow (German et al. 1996; Lein et al. 2001), Lost City (Kelley et al. 2001), Ashadze (Cherkashev et al. 2013), Menez Hom (Fouquet et al. 2002), Nibelungen (Melchert et al. 2008), Saldanha (Dias and Barriga 2006; Cherkashev et al. 2013), and Semyonov (Cherkashev et al. 2010) hydrothermal fields. In most cases, on-land ultramafic-hosted sulfide deposits, if not clearly of primary magmatic origin, are thought to be linked to serpentinization that remobilizes metals from primary silicates (e.g., olivine) into hydrothermal sulfides, if enough H2S is available (Barrie et al. 1999; Marques et al. 2007 and references therein). However, analysis of vent fluids in active seafloor hydrothermal systems has shown that the concentrations of most elements (e.g., Mn, Zn, Cu, Ag, Ba, Cd, Rb and As) are comparable in ultramafic and basaltic systems, with the exception that fluids from ultramafic-hosted systems have higher concentrations of CH4, H2, Co, Ni, Fe and lower Si, Pb, Al, Mg concentrations (Wetzel and Shock 2000; Charlou et al. 2002; Douville et al. 2002; Schmidt et al. 2007). This implies that even in hydrothermal fields hosted in tectonic-related hydrothermal systems, fluids circulate probably also through mafic basements. Allen and Seyfried Jr (2004) further proposed that even for the low-temperature ultramafic-hosted Lost City system, serpentinization likely played an insignificant role accounting for hydrothermal circulation compared with hot lithospheric units and/or near magmatic heat sources, as the chemical compositions (such as Cl, K/Cl, and Na/Cl ratios) of vent fluids from this hydrothermal field are virtually unchanged from the seawater values. In addition, most modern seafloor sulfide deposits are not affected by metamorphism, meaning that the study of modern ultramafic-hosted seafloor hydrothermal systems may provide new insights into the formation of ancient volcanogenic massive sulfide deposits that are spatially related to ultramafic rocks.

Abundant sulfur isotope work has been conducted to investigate the sulfide mineralization mechanisms for modern seafloor hydrothermal systems in a variety of geological settings, such as subduction zones, sediment-covered oceanic ridges near continental margins, and sediment-starved mid-ocean ridges (Shanks III 2001; Hannington et al. 2005; Seal 2006; Peters et al. 2010). Sulfide sulfur in mafic-hosted mid-ocean ridge hydrothermal vents is derived from leaching of basaltic-sulfide and seawater-derived sulfate that is reduced during high-temperature water rock interaction (Ono et al. 2007). However, microbial sulfate reduction during low-temperature serpentinization can provide significant sulfur for sulfide-mineralized systems when they are hosted by mantle peridotites as commonly exposed at slow- and ultraslow-spreading mid-ocean ridges (Alt and Shanks III 1998, 2011; Alt et al. 2007). The possible sulfur reservoirs for ultramafic-hosted hydrothermal fields therefore have distinct sulfur isotopic compositions, i.e., seawater-derived sulfate with δ34S values ~ 21‰ (Rees et al. 1978), sulfide derived from mafic and ultramafic rocks with δ34S values ~ 0‰ (Sakai et al. 1984), and sulfide from microbial sulfate reduction in open environments with significantly 34S depleted sulfur (O'Hanley 1992). Therefore, sulfur isotopes can provide valuable information to identify the sulfur sources and possible reactions during the migration and precipitation of sulfides within hydrothermal systems.

The recently discovered Tianzuo hydrothermal field (27° 57′ S, 63° 32′ E) is an inactive field hosted in ultramafic rocks associated with detachment faults along the SWIR. The sulfides from the Tianzuo hydrothermal field can be generally subdivided into two stages based on their mineral assemblages, with late low-temperature sulfides (pyrite and covellite) replacing or growing around early high-temperature sulfides (isocubanite, pyrrhotine, and sphalerite) (Cao et al. 2018). On the basis of detailed mineralogical studies, this study provides precise sulfur isotopic compositions of sulfides obtained by in situ LA-MC-ICP-MS in order to further understand how sulfide mineralization can develop within ultramafic rocks along amagmatic segments.

Geological setting

SWIR geology

The SWIR separates the African and Antarctic plates from the Bouvet Triple Junction (BTJ; 0° E) in the west to its eastern end at the Rodrigues Triple Junction (RTJ; 70° E) (Fig. 1a). It extends approximately 8000 km with an almost constant spreading rate of ~ 14 mm yr−1, being one of the slowest-spreading ridges globally (Dick et al. 2003; Horner-Johnson et al. 2005). The ridge is characterized by very rugged topography with an axial rift valley and water depths of > 5000 m. It is cut by a series of N–S striking transform faults, which divide it into individual segments with distinct geological characteristics. Bathymetric data reveal a shallow central region between the Prince Edward (35° E) and Gallieni fracture zones (52° 20′ E), with an average depth of ~ 3200 m (Fig. 1a; Sauter et al. 2001; Cannat et al. 2008). The thick oceanic crust indicates a robust magma supply since 10–8 Ma, associated with the Crozet and Marion hotspots (Sauter et al. 2004, 2009; Zhou and Dick 2013; Yang et al. 2017a). The SWIR region between the Melville Fracture Zone (61° E) and RTJ is oriented at ~ 60° to the spreading direction, with no transform faults (Fig. 1b). The easternmost SWIR is characterized by a common departure from isostatic compensation of seafloor topography, pronounced asymmetry in seafloor relief, and a smooth seafloor between the two ridge flanks (Cannat et al. 2003; Sauter et al. 2013). This indicates an anomalously low melt supply and strong tectonism. There has been no obvious magmatic activity for at least 26 Ma in this region (Cannat et al. 2006), distinct from the region between the Prince Edward and Gallieni fracture zones. Compared with other segments of the ridge, the easternmost SWIR has the greatest average water depth (4730 m), thinnest (or missing) crust, and highest basalt Na8.0 content, with these features being attributed to lower mantle temperatures and lower degrees of mantle melting (Minshull et al. 2006; Cannat et al. 2008). However, several magmatically robust areas with high relief (e.g., at ~ 61.5° E, 64° E, and 65.5° E) exist along the easternmost SWIR (Fig. 1c), displaying volcanic structures several tens of meters high, with lateral extension for several tens of kilometers.

Fig. 1
figure 1

(a) The geotectonic setting and topography of the Southwest Indian Ridge. (b) The area between the Melville fracture zone and Rodrigues Triple Junction. Ridge segments and non-transform faults are based on Cannat et al. (1999) and Sauter et al. (2001). (c) Along-axis bathymetric profile between 61° E and 69° E (modified from Cannat et al. 2003). (d) Topography of the Tianzuo hydrothermal field and the adjacent Tiancheng and Mt Jourdanne fields from multibeam sonar data. (e1–e2) The red-brown sediments in the Tianzuo hydrothermal field; weathered massive sulfides and opal are commonly observed on the seafloor. (e3–e4) Samples obtained from the Tianzuo hydrothermal field. The altered ultramafic rock in (e3) indicates that the pyroxene phenocrysts are surrounded by serpentine minerals. The sulfides in (e4) are generally cracked and covered by red iron oxide

Full size image

Geology of the Tianzuo hydrothermal field

The Tianzuo hydrothermal field was discovered during the COMRA DY115-20 cruise by the R/V Dayangyihao in 2009 (Tao et al. 2014), and is located in Segment 11 of the SWIR, near the inactive Mt. Jourdanne and active Tiancheng fields (Fig. 1d; Münch et al. 2001; Nayak et al. 2014; Cao et al. 2018; Chen et al. 2018). Further detailed investigations of this hydrothermal site were carried out by the Chinese manned submersible Jiaolong during the COMRA DY115-35 cruise conducted by the R/V Xiangyanghong 9 in 2014–2015. Segment 11 has an hourglass shape extending N–S for > 20 km, with an axial volcanic ridge bounded to the east and west by a non-transform discontinuity (NTD). Along-axis topography in Segment 11 varies within ~ 2700 m, and the crustal thickness is > 6 km at the center with high relief, with an average thickness of ~ 4 km. Maximum crustal thicknesses in adjacent low-relief segments are 2–3 km, indicating focused and variable magmatic accretion (Cannat et al. 2003). The inactive Mt. Jourdanne and adjacent active Tiancheng (which shows low-temperature diffuse flows) hydrothermal fields are located at a volcanic structure along the rift axis, at water depths of 2750 and 2950 m, respectively (Fig. 1c, d). Outcrops in both these fields comprise predominantly basaltic rocks (Münch et al. 2001; Tao et al. 2014; Chen et al. 2018). The Tianzuo hydrothermal field is situated on top of a dome-shaped structure typically associated with oceanic detachment faults (Fig. 1d; Standish and Sims 2010) and outcrops comprise ultramafic rocks, which are distinct from the adjacent fields (Cao et al. 2018; Chen et al. 2018). The Tianzuo field is thus the first confirmed sulfide field hosted by ultramafic rocks and controlled by detachment faults in the SWIR.

The Tianzuo hydrothermal field is located ~ 14 km southeast of the Tianzuo seamount at a depth of 3630 m (Fig. 1d). Hydrothermal activity is likely controlled by the detachment fault, as in the Logatchev and Rainbow fields (German et al. 1996; Schmidt et al. 2007). The DY115-20 cruise and 88th Jiaolong dive found that the main outcrops of the Tianzuo field are altered ultramafic rocks and hydrothermal precipitates (Fig. 1e), with no significant evidence of magmatic activity or typical vent fauna, indicating hydrothermal activity is ancient. Adjacent areas are covered by gray sediments, distinct from the Tianzuo field where opal and red-brown sediments occur over an area of about 800 × 530 m (Fig. 1e1, e2) and where massive sulfide deposits observed on the surface are strongly weathered (Fig. 1e3, e4). Samples analyzed in the present study were collected from the surface of this field by the manned submersible Jiaolong during the DY115-35 cruise and 88th Jiaolong dive, and mainly include massive sulfide deposits and serpentinized peridotites (Figs. 1e3, e4 and 2a–o). The manned submersible Jiaolong dived at the coordinates of 27° 57′ S, 63° 32′ E, exactly located at the eastern part of the Tianzuo hydrothermal field. Four fresh serpentinized peridotite samples were collected from the margin of this field. Jiaolong traveled through this field westwards with a total distance of several hundred meters. The field was generally covered by red-brown sediments which most probably resulted from the weathering and oxidization of sulfides. Some strongly weathered ancient sulfide chimneys have been observed standing above these sediments with distinct shapes. Faults and collapses have been observed, with relatively fresh massive sulfides occurring along the fault surfaces. The massive sulfides most likely represent a mound with the diameter of several hundred meters. Three relative fresh massive sulfide samples were collected from the sulfide mound.

Analytical methods

Preliminary SEM and EDS analyses

Thin sections of samples from the Tianzuo hydrothermal field were prepared for mineralogical examination by reflected and transmitted light under a petrographic microscope. Scanning electron microscope (SEM) imaging and energy dispersive system (EDS) element mapping were performed with a Zeiss Supra 55 at the State Key Laboratory for Mineral Deposits Research, Nanjing University, Nanjing, China, with an accelerating voltage of 15 kV and a beam spot size of 60 aperture.

In situ LA-MC-ICP-MS sulfur isotope analysis

Sulfur isotopic compositions of sulfide minerals were determined in situ using a 193-nm ArF excimer laser ablation system (RESOlution S-155) connected to a multi-collector inductively coupled plasma mass spectrometer (Nu Plasma II) at the State Key Laboratory for Geological Processes and Mineral Resources, China University of Geosciences (Wuhan), Wuhan, China. Secondary and back-scattered electron images were used to select spots for analysis, avoiding defects and contamination from different sulfide minerals. The in situ LA-MC-ICP-MS analytical procedure follows the protocol defined in Zhu et al. (2016), and the detailed description of this technique is reported by Mason et al. (2006), Craddock et al. (2008), and Bühn et al. (2012). In brief, helium carrier gas was used at 0.4 L min−1 and mixed with Ar (0.8 L min−1) and N2 (~ 4 mL min−1) (addition of N2 eliminates polyatomic interferences such as 16O2+; Fu et al. 2016). The laser beam diameter was 33 μm, with a repetition rate of 10 Hz for single spot analyses, an energy of 3–4 J cm−2, and the ablation process was set to last for 40 s. An in-house pyrite standard WS-1, consisting of a natural pyrite crystal from the Wenshan polymetallic skarn deposit (Yunnan Province, South China), was used to calibrate the mass bias for S isotopes. The δ34S composition of WS-1 (1.1 ± 0.2‰) was determined by secondary ion mass spectrometry at the Chinese Academy of Geochemistry, Guangzhou, China (Zhu et al. 2016). Secondary reference material WS-2 pyrite (δ34S = 2.2‰) was used to monitor the accuracy of the analysis results. All sulfur isotopic compositions are given in the common δ34S notation as permil difference to the V-CDT-reference (Vienna Canyon Diablo Troilite). The true sulfur isotope ratio of unknown samples was calculated by correction for instrumental mass bias by linear interpolation between the biases calculated from two neighboring standard analyses (WS-1). The internal analytical precision is given as 2 standard deviations (2σ), which is better than ± 0.5‰ (2σ).

Results

Mineralogy of the Tianzuo hydrothermal field

The serpentinized peridotite samples collected from the Tianzuo hydrothermal field are characteristically light-colored showing, in hand specimens, pyroxene crystals surrounded by dark serpentine minerals (Fig. 1e3). Microscopic textures indicated that pyroxene is dominated by clinopyroxene and is cut by veinlets of later serpentine, magnetite, and chrysotile (Fig. 2a). Serpentine minerals, mainly including lizardite and chrysotile, generally retain the pseudomorphic textures after olivine as mesh texture forming a network of veinlets between the relict clinopyroxene, olivine, and minor spinel grains (Fig. 2a–c). These indicate that the source rocks of serpentinized ultramafic rocks were most likely clinopyroxene peridotite. In addition to the abundant magnetite veinlets that crosscut serpentine minerals (Fig. 2a), many subhedral to euhedral magnetite grains occur with a pseudomorphic texture after olivine (Fig. 2d, e), contemporaneous subhedral to euhedral shape hematite is observed coexisting with these magnetite grains (Fig. 2d, e). Minor quartz veinlets also occur within the serpentine minerals, possibly a product of serpentinization.

Fig. 2
figure 2

Microscopic images of samples collected from the Tianzuo hydrothermal field. a Partially serpentinized peridotite showing clinopyroxene phenocrysts surrounded by serpentine minerals, both of which are crosscut by late magnetite veinlets (cross-polarized light). b Olivine altered as serpentine minerals with the mesh texture consisting of relict olivine surrounded by a network of serpentine veinlets between the olivine grains (reflected light). c Spinel surrounded by serpentine minerals (reflected light). d Magnetite coexisting with hematite and occurring as fillings in holes of serpentine minerals (reflected light). e Subsection of Fig. 2 d, showing hematite with subhedral to euhedral shape, or occurring as an exsolution within magnetite (back-scattered electron image). f Isocubanite with typical grid structure disseminated in serpentine minerals (reflected light). g Isocubanite replaced by pyrite-II; sphalerite replaced by late covellite (reflected light). h Sphalerite with isocubanite exsolution (reflected light). i, j Pyrite-II rims retaining a pseudomorphic texture after pyrrhotite (i reflected light, j back-scattered electron image). k Pyrite-I replacing pyrrhotite, showing pseudomorphic textures after pyrrhotite, which was further replaced by pyrite-II; marcasite can be observed coexisting with pyrite-I or included in pyrite-II as relicts (reflected light). l–n Fine-grained pyrite-I filling the outline of early pyrrhotite with quartz, and replaced by late pyrite-II (l reflected light, m back-scattered electron image, n secondary electron image). o Covellite occurred in quartz as lamellae or fibrous (secondary electron image). Mag, magnetite; Serp, serpentine; Cpx, clinopyroxene; Ol, olivine; Spl, spinel; Hem, hematite; Iso, isocubanite; Py, pyrite; Sp, sphalerite; Cv, covellite; Po, pyrrhotite; Mar, marcasite; Qtz, quartz

Full size image

Sulfide samples collected from the Tianzuo field are yellow–brown in color in hand specimen, and are usually covered by red iron oxide as a result of sulfide weathering (Fig. 1e4). Microscopic observation revealed that sulfides are mainly characterized by major pyrite and isocubanite; and minor sphalerite, pyrrhotite, and covellite, a mineral assemblage of sulfides similar to those described for other hydrothermal fields hosted in ultramafic rocks, such as Logatchev-1, -2, and Rainbow (Lein et al. 2001; Mozgova et al. 2005). These sulfide minerals contain anhedral grains with rare subhedral to euhedral crystals, which are generally disseminated in serpentine-bearing ultramafic rocks, with the exception of covellite and partial pyrite, hosted in quartz veinlets (Fig. 2 and ESM Fig. 1).

Isocubanite disseminated in serpentinite is easily distinguishable from other sulfide minerals in the Tianzuo field. It shows an anhedral shape, brown color, and usually displays grid structures typically resulting from exsolution within other Cu-Fe-sulfides (Fig. 2f, g), such as chalcopyrite, which is further evidenced by the Cu-rich rims around isocubanite, as illustrated by ESM Fig. 1a1-a4. There is no obvious crosscutting relationship between isocubanite, sphalerite, and pyrrhotite (or pyrite with pseudomorphic texture after pyrrhotite), and so precipitation of isocubanite was likely contemporaneous with pyrrhotite and sphalerite crystallization. Isocubanite displays obvious replacement by late pyrite-II (Fig. 2g and ESM Fig. 1a1-a6), and its precipitation must pre-date pyrite-II. The matrix seems to present weathering of serpentine minerals due to the higher concentrations of Si and Mg, as illustrated by EDS element mapping (ESM Fig. 1a5-a6). Experimental studies of the phase relationships of reactions in the Cu-Fe-S and Fe-S systems indicate that at temperatures of > 335 °C, isocubanite begins to crystallize with chalcopyrite, pyrite, and pyrrhotite (Lusk and Bray 2002). Isocubanite was thus likely precipitated during the early high-temperature phase of the Tianzuo field.

Sphalerite is rare in the Tianzuo samples, as indicated by SEM imaging and EDS element mapping (ESM Fig. 1b-c). It is generally anhedral, gray in color, and usually contains abundant isocubanite or, conversely, it is included in isocubanite as exsolution lamellae (Fig. 2h and ESM Fig. 1b1-b6). Fewer subhedral sphalerite grains can also be found contained in serpentine minerals in this field (ESM Fig. 1c1-c3). This mineral is usually replaced by late pyrite-II similar to isocubanite, indicating that it precipitated with isocubanite during an early high-temperature phase.

Pyrrhotite is rare in Tianzuo sulfide samples, and is usually included in pyrite-II, with an anhedral shape, as a relict, which was replaced by late pyrite-II (Fig. 2i, j and ESM Fig. 1d1-d3). In addition, pyrrhotite plates can also be completely replaced by pyrite-I, resulting abundant pseudomorphic textures after pyrrhotite. Sometimes, pyrite-I crystals were absent due to strong weathering, resulting in abundant holes surrounded by pyrite-II rims with pseudomorphic textures after pyrrhotite (Fig. 2i–n and ESM Fig. 1d-e). Pyrite-I and -II replacing or growing around pyrrhotite displays pseudomorphic textures after pyrrhotite, indicating that Tianzuo field was originally enriched in pyrrhotite that was almost completely replaced by late pyrite-I and -II during the late low-temperature phase, resulting in few pyrrhotite relicts in pyrite-II and abundant pseudomorphic textures (Fig. 2j–n and ESM Fig. 1e1-e6).

Pyrite is the most abundant sulfide phase in Tianzuo field and it is either hosted in serpentine minerals or coexists with quartz (Fig. 2g–n and ESM Fig. 1e1-e6). Based on its morphology, pyrite can be subdivided into two types, i.e., pyrite-I coexisted with marcasite as colloform, retaining pseudomorphic textures after pyrrhotite and creating a turbid appearance in thin section (Fig. 2k–n); pyrite-II with crystals that have typically grown around and replace early isocubanite, sphalerite, pyrrhotite, and pyrite-I, or are present as individual grains in quartz veinlets, with no inclusions visible in microscopic examination (Fig. 2g–n and ESM Fig. 1a, d, e). The matrix containing pyrite-I is also abundant in serpentine minerals, while that containing pyrite-II is most likely quartz as indicated by EDS element mapping (ESM Fig. 1e5-e6). Marcasite relicts can be also observed retained in pyrite-II (Fig. 2k). Marcasite is unstable and precipitates at relatively low temperatures (< 200 °C) and low pH (< 4.5). It undergoes transformation to pyrite at temperatures of > 350 °C (Fleet 1970). Pyrite-I and -II in the Tianzuo hydrothermal field usually replaced early minerals, including isocubanite, sphalerite, and pyrrhotite, suggesting that they precipitated after these minerals and crystallized during a low-temperature phase. This is particularly evident for pyrite-II, which also replaced pyrite-I, indicating that this mineral precipitated later than pyrite-I (Fig. 2k–n).

Covellite is another low-temperature sulfide mineral found in Tianzuo sulfide samples. The atom ratios of Cu/S obtained by EMPA are 1:1 (Cao et al. 2018). It is blue in color, usually replacing sphalerite with an anhedral morphology (Fig. 2g–h) or occurring in quartz veinlets as lamellae or fibers (Fig. 2o and ESM Fig. 1f1-f6).

On the basis of these observations, it is concluded that sulfide mineralization in the Tianzuo hydrothermal field occurred in two stages, i.e., an early high-temperature stage (> 335 °C; isocubanite, sphalerite, and pyrrhotite) and a late low-temperature stage (< 200 °C; pyrite-I, -II, and covellite).

Sulfur isotopic composition of sulfide phases

Sulfur isotopic compositions of 81 individual sulfide minerals were determined, including 26 isocubanite, 8 sphalerite, 32 pyrite-I, and 15 pyrite-II samples. Results are presented in ESM Table 1 and illustrated in Fig. 3 a. The grain sizes of pyrrhotite and covellite were too small for such analyses. Sulfide minerals from the Tianzuo hydrothermal field display a large variation in δ34S values. Isocubanite has the least variability in δ34S, with values of 9.6 to 12.2‰; sphalerite has the highest value, ranging from 9.1 to 14.1‰; pyrite-I has variable and generally positive δ34S values of − 0.1 to 12.0‰; and pyrite-II has the lowest and most variable values of − 23.8 to − 3.6‰.

Fig. 3
figure 3

a Frequency distribution of δ34S values for sulfides from the Tianzuo hydrothermal field. b Range of δ34S values for sulfides from modern sediment-starved ridges, showing that sulfides hosted in ultramafic rocks (cross) have wider ranges than those hosted in basaltic rocks (square). Data sources: Seawater (Rees et al. 1978), sulfides in MORB (Sakai et al. 1984), magmatic volatiles based on the sulfides (native sulfur) from Lau Basin (Herzig et al. 1998), East Pacific Rise (EPR) 21° S and EPR 13° N (Ono et al. 2007), EPR 9°–10° N (Ono et al. 2007; Rouxel et al. 2008), EPR 21° N (Woodruff and Shanks III 1988); Galapagos Rift (Knott et al. 1995); Juan de Fuca Ridge (Shanks III and Seyfried Jr 1987; Hannington and Scott 1988); Edmond and Kairei (Wang 2012; Zeng et al. 2016); Logatchev (Bogdanov et al. 1997; Rouxel et al. 2004; Zeng et al. 2016); Broken Spur (Duckworth et al. 1995; Butler et al. 1998); Snake Pit (Kase et al. 1990); Trans-Atlantic Geotraverse (TAG; Herzig et al. 1998; Knott et al. 1998; Zeng et al. 2016); Lucky Strike (Rouxel et al. 2004; Ono et al. 2007); Rainbow (Lein et al. 2001; Rouxel et al. 2004); Yuhuang (Liao et al. 2018); Duanqiao (Yang et al. 2017b); Longqi (Ye 2010). Spreading rates of mid-ocean ridges are based on Baker et al. (1996) and Hannington et al. (2005, 2010)

Full size image

Discussion

Potential sulfur sources for modern seafloor hydrothermal fields

The δ34S values of sulfide minerals in modern sediment-starved ridge hydrothermal systems typically range from 1 to 7‰ (Shanks III et al. 1995; Shanks III 2001; Hannington et al. 2005; Seal 2006; Peters et al. 2010). In mid-ocean ridge hydrothermal systems, sulfur forming hydrothermal sulfides are derived from different sources, i.e., (1) magmatic (including degassing processes), (2) leaching from the ocean basement rocks and sediment covers, (3) thermochemical reduction of seawater sulfate, and (4) remobilization of sulfur produced by microbial sulfate reduction. Each of these sources has distinctive sulfur isotopic signatures (Ohmoto and Rye 1979; Seal 2006 and references therein). At magmatic temperatures (higher than 400 °C), sulfur at deeper zones of the hydrothermal system has δ34S values close to 0‰, as SO2 is totally reduced to H2S and sulfides forming from these fluids (Herzig et al. 1998). However, at the temperatures below 350–400 °C, the δ34S values will disproportionate rapidly according to the disproportionation reaction: 4SO2(aq) + 4H2O(g) = H2S(aq) + 3H+ + 3HSO4−, the δ34S values of the sulfides precipitated from these fluids will be < 0‰ and the δ34S values of the sulfate derived from the same fluids will be > 0‰ and generally lower than seawater (Ohmoto and Rye 1979 and references therein). For example, sulfides (native sulfur) and barite from the Hine Hina hydrothermal field, Lau Basin, have δ34S values ranging from − 7.7 to − 2.4‰ and 16.1 to 16.7‰, respectively (Herzig et al. 1998). According to the model of Janecky and Shanks III (1988), simple adiabatic mixing associated with seawater-basalt interactions produces H2S with a maximum δ34S value of 4.5‰, whereas H2S produced through thermochemical reduction of modern seawater sulfate, through interaction with magnetite and fayalite, has δ34S values which can fractionate by a scale up to 21‰ (Ohmoto et al. 1983). Anhydrite precipitated from the pristine seawater has the δ34S value as the same as seawater (~ 21‰; Rees et al. 1978; Chiba et al. 1998; Herzig et al. 1998; Zeng et al. 2016). In contrast, H2S produced through dissolution of biogenic sulfides in sedimentary rocks has negative δ34S values reflecting microbial sulfate reduction (Canfield and Teske 1996; Ohmoto and Goldhaber 1997; Alt and Shanks III 1998, 2011; Allen and Seyfried Jr 2004; Alt et al. 2007), although questions remain about the magnitude of sulfur isotope fractionation during microbial sulfate reduction (Rouxel et al. 2008); this microbial sulfate reduction may cause sulfate–sulfide fractionation of 15 to 60‰ (Goldhaber and Kaplan 1975; Canfield and Teske 1996; Detmers et al. 2001; Wortmann et al. 2001; Canfield 2002).

The δ34S values of hydrothermal sulfides precipitated in sediment-starved ridges with a range of spreading rates and hosted in basaltic rocks varies in a narrow range (1 to 7‰) (Fig. 3b; Shanks III et al. 1995; Shanks III 2001; Hannington et al. 2005; Seal 2006; Peters et al. 2010; Zeng et al. 2016) and mainly represent sulfur leached from basaltic substrate, with little contribution from inorganically reduced seawater sulfate. Increasing contributions of reduced seawater sulfate and higher δ34S values occur as ridge spreading rate decreases (Fig. 3b), reflecting equilibration of hydrothermal fluids with rocks that have already undergone a long history of seawater–rock interaction (Shanks III 2001). Sulfides hosted by ultramafic rocks, as in the Logatchev, Rainbow, and Tianzuo hydrothermal fields, or closely associated with these rocks as in the Yuhuang field (Liao et al. 2018), have greater δ34S variability than those hosted in basaltic rocks (Fig. 3b). This feature applies especially to the Tianzuo field, where sphalerite has the highest (14.1‰) and pyrite-II has the lowest (− 23.8‰) δ34S values found to date in sediment-starved ridges, with these minerals representing high- and low-temperature mineralization phases, respectively. Neither basaltic magmatism nor seawater-basalt interactions can generate reduced H2S with such extremely positive and negative δ34S values as mentioned above, thus, we favor that thermochemical reduction of early precipitated anhydrite and leaching of sulfides in the wall rocks produced by microbial sulfate reduction most likely have isotopically played an important role for the sulfide mineralization in this Tianzuo field.

Possible early anhydrite precipitation in the Tianzuo hydrothermal field

In the Tianzuo hydrothermal field, sulfide minerals are hosted in serpentine-bearing ultramafic rocks where fractures are widely developed, in accordance with the SWIR being controlled mainly by tectonism rather than magmatism (Standish and Sims 2010; Tao et al. 2020). Early stage sulfide minerals, including isocubanite and sphalerite, are isotopically heavier than those from other sediment-starved systems, especially for sulfides hosted in basaltic rocks (ESM Table 1; Fig. 3). H2S from the deep reaction zone in the Tianzuo field is thus isotopically heavier than that from other sediment-starved systems. δ34S values for the two early sulfide minerals are as high as 14.1‰ (ESM Table 1), with such heavy values requiring the addition of sulfur from seawater sulfate, and limited volcanogenic and/or microbial reduction contributions (Lein et al. 2001; Shanks III 2001; Hannington et al. 2005; Seal 2006; Ono et al. 2007; Peters et al. 2010). It seems, therefore, that reduction of aqueous sulfate and/or reduction of anhydrite must have occurred in the deep reaction zone of this hydrothermal system. Seawater contains ~ 10 mmol/kg Ca2+ and ~ 28 mmol/kg SO2–4 and anhydrite precipitates from pristine seawater upon heating above 150 °C in the low-temperature recharge zone, only a small proportion (< 1 mmol/kg) sulfate enters to the deep high-temperature (> 250 °C) reaction zone (Woodruff and Shanks III 1988; Sleep 1991), this indicates that if the temperature in the pathway between recharge and reaction zones was low enough, seawater sulfate could reach the reaction zone, leading to anhydrite precipitation at depth (Lowell and Yao 2002; Lowell et al. 2003; Ono et al. 2007).

The dissolution of anhydrite precipitated during an early low-temperature phase may explain the heavy isotopic composition of sulfide minerals formed during the subsequent high-temperature sulfide mineralization, such as isocubanite and sphalerite, in the Tianzuo hydrothermal field. Although the formation of oceanic core complexes means that at least part of the oceanic lithosphere formed at slow- and ultraslow-spreading centers comprises peridotite with gabbroic intrusions in varying proportions (Ranero and Reston 1999; Lowell and Rona 2002; Ildefonse et al. 2007; Escartin et al. 2008), there is no evidence that earlier sulfides existed before isocubanite and sphalerite in this Tianzuo hydrothermal field. In addition, δ34S values of isocubanite and sphalerite from this field fall within a narrow range (9.1 to 14.1‰), meaning that the formation of these two sulfide minerals most likely dominated by one major gabbroic intrusion event rather than multiple magmatic activities, as the latter most likely will result in multi-stage sulfide mineralization overlapped with different δ34S values (Woodruff and Shanks III 1988), especially with the involvement of seawater sulfate, so we favor that there were probably no early high-temperature gabbroic intrusions which have significantly contributed to the sulfide mineralization in this Tianzuo hydrothermal field. It appears likely, therefore, that hydrothermal circulation was initially of low-temperature, dominated by seawater, and driven more notoriously by the exothermic serpentinization of ultramafic rocks. In fact, the decreasing solubility of anhydrite and other sulfate minerals as temperature increases leads to their removal from seawater by precipitation during heating associated with downwelling (Bischoff and Seyfried Jr 1978; Seyfried Jr and Bischoff 1981; Shanks III et al. 1995), as illustrated in Fig. 4a. Although the studies about the faults in the Tianzuo field were limited, fractures with extreme depths (13 ± 2 km), which are even deeper than the other mafic- and ultramafic-hosted hydrothermal fields, such as Trans-Atlantic Geotraverse (TAG) and Logachev-1 on the MAR, can well developed in this ultraslow-spreading SWIR (Tao et al. 2020). The anhydrite precipitated in the deep fractures during this early low-temperature phase can be dissolved into the subsequent high-temperature hydrothermal circulation, which has an isotopically significant contribution for the formation of isocubanite and sphalerite, as discussed below.

Fig. 4
figure 4

Cartoon showing the sulfide formation processes in the Tianzuo hydrothermal field. a Anhydrite precipitated in the reaction zone during an earlier phase of low-temperature hydrothermal circulation. b Gabbroic intrusion provided sufficient heat to drive high-temperature hydrothermal circulation; earlier precipitated anhydrite is available for high-temperature sulfate reduction; all high-temperature and isotopically heavier sulfides precipitated, including isocubanite, sphalerite, and pyrrhotite. c Seawater-dominated, low-temperature, hydrothermal circulation while the magma chamber is cooling; sulfur from microbial sulfate reduction is the predominant sulfur source for late sulfide minerals, including pyrite-II and most probably covellite

Full size image

High-temperature phase sulfide mineralization in the Tianzuo hydrothermal field

Previous studies have shown that venting in modern seafloor hydrothermal systems, such as the TAG, Edmond, Kairei, and Duanqiao fields, has been intermittent in the past (Lalou et al. 1998; Chiba et al. 1998; Rouxel et al. 2008; Wang 2012; Yang et al. 2017b), implying that hydrothermal discharge through major fracture systems may be reactivated. If high-temperature activity was reactivated after a previous low-temperature phase, seawater-derived sulfate (in the form of anhydrite) would be available as a source of sulfur, in addition to basaltic sulfur (Janecky and Shanks III 1988). With abundant sulfate, sulfate sulfur would be reduced by reaction with magnetite and/or fayalite in ultramafic rocks and incorporated into the hydrothermal fluid, such as 4H+ + 2SO2–4 + 12Fe3O4 = 18Fe2O3 + FeS2 + 2H2O (Gow et al. 1994). Abundant hematite and magnetite coexisting in the serpentine minerals further support this hypothesis for Tianzuo hydrothermal field (Fig. 2d, e). Potential sources of heat include magmatism, mantle upwelling, and serpentinization (German and Lin 2004; Lowell and Rona 2002), with serpentinization previously being considered the major source (McCaig et al. 2007). However, heat balance models predict that heat released during serpentinization of peridotites may result in a wide range of hydrothermal venting temperatures, while temperatures of hydrothermal fluids triggered by this process are up to only 40–75 °C (Lowell and Rona 2002). Furthermore, serpentinization cannot explain the abundance of rare earth elements, Ba, and Si in hydrothermal fluids (Douville et al. 2002). Gabbroic intrusions are therefore likely sources of heat and elements key for the formation of sulfides, even in hydrothermal fields with sulfides hosted in ultramafic rocks (Wetzel and Shock 2000). Movement along detachment faults may exhume gabbros and even peridotites, both of which are features of modern oceanic core complexes (Ranero and Reston 1999). However, in the Tianzuo field, sulfide minerals precipitated during a high-temperature phase, including isocubanite and sphalerite, have the highest and most uniform δ34S values (Fig. 3a), indicating that sulfur was derived from a homogeneous source with high δ34S values before precipitation as Fe, Cu, and Zn sulfides.

For the seafloor hydrothermal systems, precipitation of anhydrite is probably important (Sleep 1991; Lowell and Yao 2002; Lowell et al. 2003), and numerous studies have shown the presence of anhydrite in the serpentine ultramafic rocks (Alt and Shanks III 1998, 2011; Allen and Seyfried Jr 2004; Alt et al. 2007). As mentioned above, the gabbroic intrusion most likely has provided sufficient heat driving the hydrothermal circulation where high-temperature phase (> 335 °C) isocubanite, sphalerite, and pyrrhotite precipitated in Tianzuo hydrothermal field; this means that in the high-temperature reaction zone, the hydrothermal fluids can gain reduced sulfur (H2S) from reduction of seawater-derived anhydrite in addition of the leaching of sulfides from the wall rocks (e.g., gabbros and peridotite). The models previously used to interpret the various δ34S values of sulfides in the seafloor hydrothermal systems including (1) the anhydrite buffer model (Ohmoto et al. 1983), (2) disproportionation of SO2 (Ohmoto and Rye 1979; Herzig et al. 1998), and (3) two-component mixing (Janecky and Shanks III 1988; Shanks III 2001). The anhydrite buffer model suggests that the majority of hydrothermal sulfides ultimately derived from seawater-derived sulfate, the δ34S values of these sulfides essentially reflect the fractionation factor between sulfate (SO2–4) and sulfide (H2S). The isotope fractionation factors between sulfate and sulfide are 21 and 15‰ at 300 °C and 400 °C (Ohmoto et al. 1983), respectively, which can explain why the hydrothermal sulfides in the seafloor hydrothermal system have the δ34S values generally range from 1 to 7‰ (seal 2006 and the references therein). However, this model cannot explain the extremely high δ34S values for the isocubanite and sphalerite from the Tianzuo hydrothermal field (9.1 to 14.1‰). The disproportionation of SO2 means that the SO2 in the volcanic gases will disproportionate rapidly into reduced H2S and oxidized SO2–4 when temperatures below 300–400 °C, with the sulfides precipitated in the hydrothermal fluids have low δ34S values (< 0‰) and the sulfates have relative high δ34S values (> 0‰). This model was generally used to explain the formation of those sulfides occurred in the arc/back arc environments with low δ34S values and characteristic acid-sulfate alteration assemblages (e.g., alunite/barite-pyrite in the Lau Basin; Herzig et al. 1998). The isocubanite and sphalerite have extremely high δ34S values and sulfates are absent in the Tianzuo hydrothermal field; the influence of disproportionation of SO2 therefore can be removed from this field. Thus, for the Tianzuo hydrothermal field, the δ34S values of isocubanite and sphalerite most likely reflect two-component mixing between seawater-derived sulfate (δ34S ≈ 21‰) and wall rocks/gabbroic intrusion-derived sulfide (δ34S ≈ 0‰). Based on this two component mixing model, the highest δ34S value of sulfides form the Tianzuo hydrothermal field (δ34S = 14.1‰) indicates that about 33% of H2S was derived from leaching of sulfides from wall rocks/gabbroic intrusion, and the seawater-derived sulfates have provided the other 67% of H2S; the gabbroic intrusion most likely has a significant contribution to the sulfur forming these high-temperature phase sulfides in addition of the abundant heat as mentioned before.

The Tianzuo hydrothermal field is located about 35 km to the southwest of Mt. Jourdanne hydrothermal field (Fig. 1d). Previous studies have shown that abundant galena occurred in the sulfide assemblages within the Mt. Jourdanne field, and the sulfides from this field have high concentrations for the elements Pb (up to 3.54 wt%), As (up to 0.57 wt%), Cd (up to 0.22 wt%), and Ag (up to 0.13 wt%) (Münch et al. 2001; Nayak et al. 2014). Since the enrichment of galena and those elements Pb, As, Cd, and Ag are generally observed in the arc/back arc environments associated with felsic magmas rather than sediment-starved oceanic ridges associated with basalts, Nayak et al. (2014) proposed that within the Mt. Jourdanne seamount structure, isolated bodies of more felsic rocks may provide a source for the abundant Pb, As, Cd, and Ag found in its sulfides. Therefore, another explanation for the isotopically heavier sulfides in the Tianzuo field may be the occurrence of granitic rocks in the area. Such as rocks of granitic composition have already been found in the Central Indian Ridge near the Agro Fracture Zone (Engel and Fisher 1975) and a range of volcanic rocks, from andesite to rhyodacite, has been recovered from the Galapagos Spreading Center (Shibata et al. 1979). Granitoids have an average δ34S value of 1.0 ± 6.1‰, with a range of − 11 to 14.5‰, reflecting variable assimilation or partial melting of either pyritic sedimentary rocks with low δ34S values or evaporites with high δ34S values (Ishihara and Sasaki 1989). Granites and syenites from southern India have δ34S values of 2.5 to 14.5‰ (Santosh and Masuda 1991), which are approximatively close to those of isocubanite and sphalerite from Tianzuo field. However, sulfides such as isocubanite and pyrite are depleted in these elements, including Pb, As, Cd, and Ag (Cao et al. 2018), and the Tianzuo field lacks galena (abundant in the Mt. Jourdanne field), with no significant evidence of granitic rocks. The influence of granitic rocks is thus precluded, with reduction of anhydrite (deposited during early low-temperature hydrothermal circulation through deep fractures) in the high-temperature phase being favored as the main source of isotopically heavy sulfur in isocubanite and sphalerite. Hydrothermal fluids would have been mixed in the reaction zone and heated by the gabbroic intrusion; hot fluids with homogeneous sulfur isotopic compositions would then ascend and be exhaled on the seafloor, resulting in sulfide deposition during the high-temperature phase (Fig. 4b). When high-temperature activity declined, low-temperature seawater-derived fluid may have eventually dominated hydrothermal fluid circulation again. Pyrite-I, -II, and covellite would have precipitated during this low-temperature phase.

Low-temperature phase sulfide mineralization in the Tianzuo hydrothermal field

Pyrite-II has the lowest and most variable δ34S values (− 23.8 to − 3.6‰) of sulfide minerals in the Tianzuo field, with significant signs of sulfate-reducing microbial activity. Such low δ34S values have not previously been reported for sulfides in sediment-starved ridges, but are commonly observed in sediment-covered ridges and back arc basins (e.g., Shanks III 2001; Hannington et al. 2005; Seal 2006; Ono et al. 2007; Peters et al. 2010), such as the Kebrit, Shaban and Atlantis II Deeps, Red Sea (Zierenberg and Shanks III 1988), Lau Basin (Herzig et al. 1998), Guaymas Basin (Peter and Shanks III 1992), and Middle Valley (Zierenberg 1994). Pyrite-II δ34S values indicate some influence of magmatic sulfur, and a larger contribution of microbial action than in basalt-dominated systems. In addition to detachment faults, the widely developed fractures in the Tianzuo field may have also resulted from expansion during serpentinization of minerals in ultramafic rocks such as olivine, where a 40 to 50% volume increase may occur (O'Hanley 1992). With the stable permeability of the ridge, hydrothermal fluid pathways may exist long after high-temperature hydrothermal circulation has ended. Furthermore, compared with the initial low-temperature stage when the anhydrite precipitated, the high-temperature circulation provided enough materials for life to flourish in the hydrothermal field, probably including sulfate-reducing bacteria, with numerous sulfides containing microbially reduced sulfur, which precipitated when temperatures declined.

As mentioned above, disproportionation of SO2 can also generate reduced H2S enriched in 32S (Ohmoto and Rye 1979; Herzig et al. 1998). However, this reaction most likely happened under arc/back arc environment with felsic magmatism, and the reported δ34S values of H2S range from − 2.4 to − 7.7‰ in the Lau Basin (Herzig et al. 1998), which is much higher than those of the pyrite-II (− 23.8 to − 3.6‰) from this Tianzuo field. In addition, disproportionation of SO2 will result in acid-sulfate alteration assemblages which are absent in the Tianzuo field (Herzig et al. 1998), this reaction therefore can be removed when pyrite-II precipitated. Although significant microbial sulfate reduction has been generally observed in the sediments covered hydrothermal fields, previous studies have shown that microbial sulfate reduction was also widespread in the altered basalts and mantle exposed at the seafloor along the sediment-starved ridges, as indicated by that sulfides contained in these rocks have significantly negative δ34S values (Canfield and Teske 1996; Alt and Shanks III 1998, 2011; Allen and Seyfried Jr 2004; Alt et al. 2007). In fact, the uptake of sulfur by serpentinites in oceanic basement is comparable to or greater than that by comparable volumes of mafic oceanic crust, as the microbial communities can be supported by the hydrogen and methane generated by both the alteration of basalts and serpentinization reactions of ultramafic rocks with the temperature up to 120 °C (Canfield 2002; Schwarzenbach et al. 2012). Pyrite-I, -II, and covellite in the Tianzuo field are low-temperature phase sulfides (< 200 °C), meaning that the precipitation of these three sulfides were not significantly influenced by high-temperature gabbroic intrusions, this enabled the microbial reduction of seawater sulfate and addition of low-δ34S sulfide to the wall rocks. Therefore, it is likely that the leaching of sulfide minerals produced by microbial sulfate reduction was the main source of sulfur for pyrite-II during the late low-temperature phase, with the influence of isotopically heavy residual sulfate and volcanogenic sulfur being limited (Fig. 4c).

Pyrite-I has similar variability, but much higher δ34S values (− 0.1 to 12.0‰) than pyrite-II (Fig. 3 and ESM Table 1), and displays obvious mixing of sulfur sources. As indicated by the mineralogy of sulfides, pyrite-I has obvious pseudomorphic texture after pyrrhotite (Fig. 2k–n). The influence of pyrite-II on pyrite-I is difficult to evaluate as it usually grows around pyrite-I (Fig. 2k–n and ESM Fig. 1e1-e6), but the generally positive δ34S values of pyrite-I indicate a limited contribution of sulfur from microbial sulfate reduction. One mixing end-member of pyrite-I is likely dominated by volcanogenic sulfur leaching from the wall rocks, and the other by thermochemical reduction of earlier formed sulfate during the high-temperature phase when pyrrhotite precipitated. As with isocubanite and sphalerite, the high δ34S values of pyrite-I were likely inherited from sulfur isotopic compositions of original pyrrhotite.

Conclusions

Sulfides hosted in serpentinized ultramafic rocks in the Tianzuo hydrothermal field, in an amagmatic segment of the ultraslow-spreading SWIR, include minerals from different formation phases. These include high-temperature isocubanite, sphalerite, and pyrrhotite, and low-temperature pyrite and covellite. Pyrite may be subdivided into pyrite-I and -II based on morphology, with the former displaying an obvious pseudomorphic texture after pyrrhotite. Isocubanite and sphalerite have high and uniform δ34S values, with thermochemical reduction of sulfate precipitated during early phase low-temperature hydrothermal circulation contributing sulfur for these sulfide minerals, and probably for the original pyrrhotite. Gabbroic intrusions provided heat for driving the subsequent high-temperature hydrothermal fluid circulation, creating a homogeneous sulfur environment in the reaction zone. Pyrite-I, -II, and covellite precipitated when temperature declined. Pyrite-I has variable and generally positive δ34S values, and its sulfur was inherited from the original pyrrhotite through thermochemical reduction of sulfate, mixed with some additional volcanogenic sulfur. Pyrite-II has the lowest and most variable δ34S values, and displays significant signs of microbial sulfate-reducing activity, with the leaching of sulfide minerals from wall rocks with seawater being the main sulfur source for this mineral. Intermittent magmatism, together with persistent high permeability, probably contributed to the formation of sulfides in the Tianzuo hydrothermal field, and this may also be the mechanism for the formation of sulfides hosted in other ultramafic rocks at slow- and ultraslow-spreading ridges.