Elsevier

Geochimica et Cosmochimica Acta

Volume 304, 1 July 2021, Pages 305-326
Geochimica et Cosmochimica Acta

High sulfur solubility in subducted sediment melt under both reduced and oxidized conditions: With implications for S recycling in subduction zone settings

https://doi.org/10.1016/j.gca.2021.04.001Get rights and content

Highlights

  • In hydrous rhyolitic melt, S is dissolved as H2S/HS (reduced) or SO42− (oxidized).

  • H2S and SO2 dissolution in rhyolitic melt follows the Fincham-Richardson relationship.

  • SCSS in hydrous rhyolitic melt increases with decreasing fO2.

  • Subducted sediment melts have high SCSS values when FeO contents are low.

Abstract

The relative enrichment of sulfur (S) observed in arc magmas when compared to MORB, reflects the addition of slab-derived S to the mantle wedge source region. However, the mechanisms and efficiency of such S recycling remain poorly constrained. In this study, sediment melting experiments have been conducted using a synthetic pelite starting composition containing ∼7 wt% H2O and ∼1.9 wt% S, at 3 GPa, 1050 °C and variable oxygen fugacity (fO2), to investigate the effect of fO2 on S solubility in sediment melts. To assess temperature and concentration effects, selected experiments were repeated either at lower temperatures of 950 °C and 1000 °C, or with a higher bulk S content of ∼4 wt%. All experiments produced hydrous rhyolitic melts, saturated with either pyrrhotite under reduced conditions or anhydrite under oxidized conditions. For 3 GPa, 1050 °C experiments, the sulfur content at sulfide saturation (SCSS) in melt is found to increase with decreasing fO2, from ∼200 ppm at FMQ-0.5 to ∼1900 ppm at FMQ-7.5. The highest S solubility is achieved at FMQ + 1.6 where melt is saturated with both anhydrite and pyrrhotite. The sulfur content at sulfate saturation (SCAS) decreases from ∼2600 ppm at FMQ + 1.6 to ∼2000 ppm at FMQ + 7. Increasing either bulk S content or temperature produces a positive effect on SCSS and SCAS. Raman spectra of our experimental melts show that S exists as H2S/HS under reduced conditions and as SO42− under oxidized conditions. The solubility minimum, i.e., the onset of transition from S2− to S6+ is estimated to occur at ∼FMQ, with full transition to S6+ by ∼FMQ + 2. While the SCAS values are in good agreement with previous reports, the distinct increase of SCSS with decreasing fO2 (when fO2 < FMQ) has not been observed in previous slab melting experiments. Furthermore, we report for the first time that dissolution of H2S and SO2 in hydrous rhyolitic melt follows the Fincham-Richardson relationship; and propose the definition of a hydro-sulfide capacity as CHS = [HS]*(fO2/fS2)1/2; where [HS] is the concentration of S in melt (in ppm) dissolved as HS and H2S. SCSS for H2S dissolution can therefore be modeled using CHS in an analogous fashion to modeling SCSS for anhydrous melts using the sulfide capacity (CS2−), with the relation ln[HS]SCSS=-ΔGFes - FeO°/RT+lnCHS-lnaFeOmelt+lnaFeSsulfide. As predicted by such a model framework, we indeed observe a linear correlation between logSCSS and logXFeO (the mole fraction of FeO in melt) with a slope close to −1, i.e., SCSS experiences a sharp increase when FeO in melt falls below 1 wt%. Therefore, both our experimental results and model predictions suggest hydrous low-Fe rhyolitic melt produced by sediment melting under reduced conditions has the required S solubility to account for the relative enrichment of S observed in arc magmas.

Introduction

Sulfur is one of the major volatiles released during volcanic processes. It controls the partitioning, transport and enrichment of ore metals during magmatic evolution and ore deposit formation. The subduction related S cycle has gained research attention in recent years due to its potentially important role in the evolution of mantle redox state (e.g., Kelley and Cottrell, 2009, Evans, 2012) and the formation of porphyry copper deposits (e.g., Richards, 2015, Canil and Fellows, 2017). Although estimates of S fluxes indicate a significant percentage of S is returned to the deep mantle (Wallace, 2005), the higher S content in arc basalt compared to MORB (Wallace, 2005) and high δ34S values in arc magma (e.g., De Hoog et al., 2001) suggest recycling of seawater-derived S from subducted slab to the mantle wedge. The mechanism through which such S transfer takes place remains poorly constrained. If S transfer was achieved via fluids or melts produced by slab devolatilization or fluxed-melting, the efficiency of S recycling would be dependent upon the S solubility,1 which, in turn, is a function of pressure (P), temperature (T), fluid/melt composition and oxygen fugacity (fO2).

Experimental studies of sulfur solubility in silicate melts, namely the sulfur content at sulfide saturation (SCSS) and the sulfur content at anhydrite saturation (SCAS), span several decades (Baker and Moretti, 2011), with a general focus on mantle melting and volcanic processes. More recently, a small number of studies (Prouteau and Scaillet, 2013, Jégo and Dasgupta, 2013, Jégo and Dasgupta, 2014, Canil and Fellows, 2017, Chowdhury and Dasgupta, 2019) have been devoted to the behavior of S during subduction processes.

Prouteau and Scaillet (2013) reported S solubilities in hydrous felsic melts produced by melting of hydrous MORB and pelitic compositions at 2–3 GPa, 700–950 °C and fO2 from NNO-2 to NNO + 1, with values falling in a large range of 100 ppm to 1 wt%, showing positive correlations with P, T, fO2 and bulk S content (1–4 wt%). Jégo and Dasgupta, 2014, Jégo and Dasgupta, 2013 studied the melting of hydrated basaltic crust with 1 wt% S in the PT range of 2–3 GPa, 700–1050 °C, and fO2 conditions from FMQ-3 to FMQ + 3.5. They reported consistently low SCSS values at reduced conditions (fO2 < FMQ) averaging ∼100 ppm, with higher values up to 518 ppm at intermediate conditions (FMQ < fO2 < FMQ + 2), and SCAS values of 700–3000 ppm at oxidized conditions (fO2 > FMQ + 2). While Prouteau and Scaillet (2013) reported S solubilities for hydrous felsic melts high enough to efficiently transfer S from the slab to the mantle wedge, particularly under relatively oxidized conditions, Jégo and Dasgupta, 2014, Jégo and Dasgupta, 2013 suggested SCSS values for rhyolitic melts are too low for them to be suitable carriers of S, and that aqueous fluids present a more efficient medium for S transfer. While the solubility data from the above two studies seem comparable at first order, their contrasting conclusions reflect the difficulty in constraining S recycling mechanisms based on the sparse experimental data currently available for limited P-T-composition-fO2 space. Generally speaking, S recycling in subduction zone settings has the potential to occur at different depths via differing mechanisms (Tomkins and Evans, 2015), accomplished through the mediums of aqueous fluid, silicate melt and supercritical fluids at different stages.

Further discrepancies arise when comparing experimental data from these studies with predictions using the published S solubility models of Clemente et al., 2004, Liu et al., 2007. Based on S solubility data in hydrous rhyolitic melts at low pressure (2 kbar), Clemente et al. (2004) developed a thermodynamic model relating S content in melt to fH2S and fSO2, in addition to an empirical model for S solubility in hydrous rhyolitic melts as a function of T, fO2 and fS2 when fO2 < NNO + 1.5. Liu et al. (2007) developed an empirical SCSS model based on experimental data from the literature covering a wide range of PT and melt compositions, however, this model was unable to reproduce the experimental data of Clemente et al. (2004). Empirical SCSS models suffer from sampling bias, as the majority of experimental data were obtained for anhydrous mafic melts. Wykes et al. (2015) demonstrated that the Liu et al. (2007) model underestimates the SCSS values for melt compositions with FeO contents <5 wt%. The dependence of SCSS on FeO content in melt can be depicted with a U-shaped curve, with SCSS positively correlated with FeO content (>5 wt%) in the right limb, while negatively correlated with FeO (<5 wt%) in the left limb (Haughton et al., 1974, O’Neill and Mavrogenes, 2002, Wykes et al., 2015). Prouteau and Scaillet (2013) reported that while their solubility data can be reproduced by the model of Clemente et al. (2004), it is significantly underestimated by the Liu et al. (2007) model. Conversely, Jégo and Dasgupta, 2014, Jégo and Dasgupta, 2013 report SCSS data that fit with model predictions of Liu et al. (2007).

In order to resolve such a discrepancy and explore the suitability of different S solubility models for high pressure melts produced by the fluxed melting of subducted slab, we performed a series of subducted sediment melting experiments at 3 GPa, 1050 °C with varying fO2 to observe the variations of SCSS and SCAS. The choice of PT conditions was intended for comparison with the 3 GPa, 1050 °C experimental series from Jégo and Dasgupta (2014). While 1050 °C is at the upper limit for sediment melting at sub-arc depth (e.g., Syracuse et al., 2010, Cooper et al., 2012), this high T enables the generation of large mineral-free melt pools in the experimental charge and facilitates S speciation analysis using Raman spectroscopy. Repetitive experiments were performed both at lower T and with higher bulk S content to assess related T and concentration effects.

Section snippets

Experimental starting materials

A synthetic experimental pelite starting material (EPSM, Table 1), with a major element composition similar to both the “Global Subducting Sediment” (GLOSS, Plank and Langmuir, 1998) and upper continental crust (Rudnick and Gao, 2003), with the addition of S, was chosen for the sediment melting experiments conducted in this study. Starting compositions were synthesized following the procedure previously outlined in Li and Hermann, 2015, Li and Hermann, 2017 for S-free EPSM compositions. Major

Phase relations and compositions

For 1050 °C reduced experiments (fO2 < FMQ), a simple phase assemblage of quenched melt, pyrrhotite and small fractions of the liquidus phases garnet and coesite were observed (Table 2; Fig. 2a); while the degree of melting was much lower for oxidized experiments (fO2 ≥ FMQ + 1.6), for which the mineral assemblage is garnet + coesite + anhydrite + kyanite + rutile (Table 2; Fig. 2b, c). For exp. LMD704 (FMQ + 1.6), we observed the coexistence of pyrrhotite and anhydrite (Fig. 2c). A small

Experimental equilibrium and the effect of carbon

Within such an experimental study, assuring the attainment of both compositional and fO2 equilibrium remains a primary consideration. Without conducting reversed or bracketed experiments, an equilibrium state is demonstrated by the homogeneity of phase relations and compositions. All experiments, other than LMD719, displayed phase homogeneity. With a 24-hour run time, the IrIrO-buffered exp. LMD719 preserved a two-layer structure (Fig. 2f), with the outer layer in fO2 equilibrium with the

Implications for S recycling in subduction zones

The average S content in arc basalt is estimated to be 1500 ± 500 ppm (Jégo and Dasgupta, 2013), higher than the average S content of ∼1000 ppm in MORB. The average H2O content in arc basalt is estimated to be ∼4 wt% (Plank et al., 2013). As a rough estimate, this 4 wt% H2O content and apparent excess of 500 ppm S in arc basalt, may be sourced from a subduction component with a H2O/S ratio of 80. The required composition of such a component may theoretically span a range from aqueous fluid

Conclusions

We have conducted sediment melting experiments at 3 GPa, 1050 °C and variable fO2 to investigate the variation of SCSS and SCAS with fO2 in hydrous rhyolitic melt at high pressure. We observe an increase of SCSS with decreasing fO2 from ∼200 ppm at FMQ-0.5 to ∼1800 ppm at FMQ-7.5, which correlates with the decrease in melt FeO from 3.5 to 0.3 wt%. The S content in melt increases sharply during the S2− to S6+ transition beginning at ∼FMQ, reaching a maximum value of ∼2600 ppm at FMQ + 1.6, when

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This work benefited from discussions with Hugh St.C. O’Neill from the Australian National University and Joerg Hermann from the University of Bern. The authors would also like to thank Yunlu Ma, Yichuan Wang and Wei Yan from the High Pressure and High Temperature laboratory, Peking University for their assistance with the experimental program; Xiaoli Li from Peking University and Quan Wei from the Chinese Academy of Geological Sciences for their help with electron microprobe analyses. Comments

References (80)

  • J.C.M. De Hoog et al.

    Sulfur isotope systematics of basaltic lavas from Indonesia: Implications for the sulfur cycle in subduction zones

    Earth Planet. Sci. Lett.

    (2001)
  • S. Jakobsson et al.

    The system C-O in equilibrium with graphite at high pressure and temperature: An experimental study

    Geochim. Cosmochim. Acta

    (1994)
  • S. Jégo et al.

    Fluid-present melting of sulfide-bearing ocean-crust: Experimental constraints on the transport of sulfur from subducting slab to mantle wedge

    Geochim. Cosmochim. Acta

    (2013)
  • P.J. Jugo et al.

    Sulfur K-edge XANES analysis of natural and synthetic basaltic glasses: Implications for S speciation and S content as function of oxygen fugacity

    Geochim. Cosmochim. Acta

    (2010)
  • H. Keppler

    The distribution of sulfur between haplogranitic melts and aqueous fluids

    Geochim. Cosmochim. Acta

    (2010)
  • K. Klimm et al.

    The dissolution mechanism of sulphur in hydrous silicate melts. II: Solubility and speciation of sulphur in hydrous silicate melts as a function of fO 2

    Chem. Geol.

    (2012)
  • K. Klimm et al.

    The dissolution mechanism of sulphur in hydrous silicate melts. I: Assessment of analytical techniques in determining the sulphur speciation in iron-free to iron-poor glasses

    Chem. Geol.

    (2012)
  • H. Li et al.

    Apatite as an indicator of fluid salinity: An experimental study of chlorine and fluorine partitioning in subducted sediments

    Geochim. Cosmochim. Acta

    (2015)
  • J.-L. Li et al.

    Redox processes in subducting oceanic crust recorded by sulfide-bearing high-pressure rocks and veins (SW Tianshan, China)

    Contrib. to Mineral. Petrol.

    (2016)
  • Y. Liu et al.

    Sulfur concentration at sulfide saturation (SCSS) in magmatic silicate melts

    Geochim. Cosmochim. Acta

    (2007)
  • Z. et al.

    Omphacite-bearing calcite marble and associated coesite-bearing pelitic schist from the meta-ophiolitic belt of Chinese western Tianshan

    J. Asian Earth Sci.

    (2013)
  • M. Masotta et al.

    Anhydrite solubility in differentiated arc magmas

    Geochim. Cosmochim. Acta

    (2015)
  • D.A. McKeown et al.

    Raman studies of sulfur in borosilicate waste glasses: Sulfate environments

    J. Non. Cryst. Solids

    (2001)
  • O. Namur et al.

    Sulfur solubility in reduced mafic silicate melts: Implications for the speciation and distribution of sulfur on Mercury

    Earth Planet. Sci. Lett.

    (2016)
  • H.S.C. O’Neill et al.

    Activity coefficients at low dilution of CrO, NiO and CoO in melts in the system CaO-MgO-Al2O3-SiO2 at 1400°C: Using the thermodynamic behaviour of transition metal oxides in silicate melts to probe their structure

    Chem. Geol.

    (2006)
  • T. Plank et al.

    Why do mafic arc magmas contain ∼4wt% water on average?

    Earth Planet Sci. Lett.

    (2013)
  • T. Plank et al.

    The chemical composition of subducting sediment and its consequences for the crust and mantle

    Chem. Geol.

    (1998)
  • J.P. Richards

    The oxidation state, and sulfur and Cu contents of arc magmas: implications for metallogeny

    Lithos

    (2015)
  • R.L. Rudnick et al.

    Composition of the Continental Crust

  • C. Spandler et al.

    Element recycling from subducting slabs to arc crust: A review

    Lithos

    (2013)
  • E.M. Syracuse et al.

    The global range of subduction zone thermal models

    Phys. Earth Planet. Inter.

    (2010)
  • R. Tao et al.

    Formation of abiotic hydrocarbon from reduction of carbonate in subduction zones: Constraints from petrological observation and experimental simulation

    Geochim. Cosmochim. Acta

    (2018)
  • T.B. Thomsen et al.

    Melting of carbonated pelites at 2.5–5.0 GPa, silicate–carbonatite liquid immiscibility, and potassium–carbon metasomatism of the mantle

    Earth Planet. Sci. Lett.

    (2008)
  • A.G. Tomkins et al.

    Separate zones of sulfate and sulfide release from subducted mafic oceanic crust

    Earth Planet. Sci. Lett.

    (2015)
  • P. Toulmin et al.

    A thermodynamic study of pyrite and pyrrhotite

    Geochim. Cosmochim. Acta

    (1964)
  • K. Tsuno et al.

    The effect of carbonates on near-solidus melting of pelite at 3GPa: Relative efficiency of H2O and CO2 subduction

    Earth Planet. Sci. Lett.

    (2012)
  • P.J. Wallace

    Volatiles in subduction zone magmas: concentrations and fluxes based on melt inclusion and volcanic gas data

    J. Volcanol. Geotherm. Res.

    (2005)
  • Z. Zajacz et al.

    An accurate model to predict sulfur concentration at anhydrite saturation in silicate melts

    Geochim. Cosmochim. Acta

    (2019)
  • D.R. Baker et al.

    Modeling the solubility of sulfur in magmas: A 50-year old geochemical challenge

    Rev. Mineral. Geochemistry

    (2011)
  • I. Barin

    Thermochemical Data of Pure Substances

    (1995)
  • Cited by (9)

    • Favorable P–T–ƒO<inf>2</inf> conditions for abiotic CH<inf>4</inf> production in subducted oceanic crusts: A comparison between CH<inf>4</inf>-bearing ultrahigh- and CO<inf>2</inf>-bearing high-pressure eclogite

      2022, Geochimica et Cosmochimica Acta
      Citation Excerpt :

      Previous studies have shown that prograde fluid-induced dissolution of carbonate and carbonaceous materials may represent an essential process by which C-bearing fluids are released from subducting slabs (Frezzotti et al., 2011; Ague and Nicolescu, 2014; Facq et al., 2014; Kelemen and Manning, 2015; Menzel et al., 2020; Tumiati et al. 2020; Vitale et al., 2020a). However, whether slab fluids are dominated by reduced (e.g., CH4, H2S, HS−) or oxidized (e.g., CO2, HSO4−, SO42−) species remains controversial, and this dynamic has significant implications for constraining the redox evolution of subducting slabs, mantle wedges, and related volcanic arcs (e.g., Debret and Sverjensky, 2017; Gerrits et al., 2019; Walters et al., 2019; 2020; Vitale et al., 2017; Piccoli et al., 2019; Li et al., 2020; Li et al., 2021, 2022). Determining the mobility and speciation of carbon-bearing phases in subduction zone fluids remains an active and challenging research frontier (e.g., Debret et al., 2021).

    • Sulfur dissolution capacity of highly hydrated and fluid-saturated dacitic magmas at the lower crust and implications for porphyry deposit formation

      2022, Geochimica et Cosmochimica Acta
      Citation Excerpt :

      High H2O and fO2 can shift reactions (2) and (3) to the right, thus elevating the S content in the melts and fluids. Developing a unique model to depict the behavior of S in the magma-hydrothermal system is challenging because sulfide/anhydrite solubility in melt and fluid is a complex function of multiple variables, including temperature, pressure, melt composition, sulfide composition, S species, and valence states (e.g., Chowdhury and Dasgupta, 2019; Fortin et al., 2015; Jugo et al., 2010; Li et al., 2021; Masotta and Keppler, 2015; Matjuschkin et al., 2016; Nash et al., 2019; O’Neill and Mavrogenes, 2002; Smythe et al., 2017; Xu and Li, 2021; Zajacz, 2015; Zajacz and Tsay, 2019). Our experiments at a fixed pressure (1.0 GPa) and temperature (950 °C), but variable H2O and fO2 obtained three important results (Figs. 7 and 8): (1) At a given fO2, the S content in the dacitic melt increases with increasing melt hydration.

    • A thermodynamic model for sulfur content at sulfide saturation (SCSS) in hydrous silicate melts: With implications for arc magma genesis and sulfur recycling

      2022, Geochimica et Cosmochimica Acta
      Citation Excerpt :

      As a safe measure, the database used for model calibration is comprised of experiments with fO2 conditions of FMQ and below. Additional datasets are included from Scaillet and MacDonald (2006), Matjuschkin et al. (2016), D’Souza and Canil (2018) and Li et al. (2021). In the database of Moretti and Baker (2008), data for hydrous silicate melts are from Bradbury (1983), Carroll and Rutherford (1985), Carroll and Rutherford (1987), Luhr (1990) and Clemente et al. (2004).

    View all citing articles on Scopus
    View full text