Sulfur solubility in a deep magma ocean and implications for the deep sulfur cycle
Affiliations | Corresponding Author | Cite as | Funding information- Share this article
-
Article views:504Cumulative count of HTML views and PDF downloads.
- Download Citation
- Rights & Permissions
top
Abstract
Figures and Tables
Figure 1 Backscattered electron images of runs ESS-5-DAC and ESS-7-DAC. Line in ESS-5-DAC is a decompression crack. | Figure 2 Newly derived SCSS values and FeO-normalised SCSS values (normalised to 8.1 wt. % FeO or xsil meltFeO = 0.05, using the model of Steenstra et al. (2018) (Supplementary Information). (a) SCSS versus FeO content of the silicate melt. (b) Comparison between measured and predicted SCSS values calculated using previous models. The SCSS values of the peridotitic L16 model are compared with our SCSS data normalised to the same FeO content as used for that model (8.1 wt. % FeO). The measured and compared values of other SCSS models were based on measured FeO contents. Light green symbols indicate other FeO-normalised literature data which were compared with the model of Equation 2 (Table S-2). (c–d) The FeO-normalised SCSS as a function of P-T. Grey and red lines indicate the P-T dependencies of the L16 peridotite SCSS model and our new model (Eq. 2), respectively. Literature data and/or previous models from Kiseeva and Wood (2013, 2015); Vogel et al. (2015); Laurenz et al. (2016); Smythe et al. (2017); Ding et al. (2018); Blanchard et al. (2021) (Table S-2). | Figure 3 (a) Variation of the SCSS during terrestrial magma ocean crystallisation along a geotherm ranging approximately midway between the peridotite solidus and liquidus (Rubie et al., 2015; Eqs. S-2,3). Plotted for comparison are previous peridotite SCSS models (Laurenz et al., 2016; Blanchard et al., 2021). (b) Calculated SCSS values for various average effective pressures of sulfide saturation (kS; Rubie et al., 2016; Eq. S-4) as a function of accreted mass. | Table 1 Experimental run conditions and results. |
Figure 1 | Figure 2 | Figure 3 | Table 1 |
top
Introduction
Sulfur (S) plays a key role in planetary geochemistry because of its ability to act as a major sink for elements when S is present as a sulfide (Kiseeva and Wood, 2015
Kiseeva, E.S., Wood, B.J. (2015) The effects of composition and temperature on chalcophile and lithophile element partitioning into magmatic sulphides. Earth and Planetary Science Letters 424, 280–294. https://doi.org/10.1016/j.epsl.2015.05.012
). Whether sulfide liquid precipitates from a silicate, magma is controlled by the S content at sulfide saturation (SCSS) of that magma. The SCSS is a function of composition, most notably FeO, pressure (P) and temperature (T), and has been extensively studied at lower pressures (<24 GPa; O’Neill and Mavrogenes, 2002O’Neill, H.St.C., Mavrogenes, J.A. (2002) The sulfide capacity and the sulfur content at sulfide saturation of silicate melts at 1400°C and 1 bar. Journal of Petrology 43, 1049–1087. https://doi.org/10.1093/petrology/43.6.1049
; Namur et al., 2016Namur, O., Charlier, B., Holtz, F., Cartier, C., McCammon, C. (2016) Sulfur solubility in reduced mafic silicate melts: Implications for the speciation and distribution of sulfur on Mercury. Earth and Planetary Science Letters 448, 102–114. https://doi.org/10.1016/j.epsl.2016.05.024
; Smythe et al., 2017Smythe, D.J., Wood, B.J., Kiseeva, E.S. (2017) The S content of silicate melts at sulfide saturation: New experiments and a model incorporating the effects of sulfide composition. American Mineralogist 102, 795–803. https://doi.org/10.2138/am-2017-5800CCBY
; Ding et al., 2018Ding, S., Hough, T., Dasgupta R. (2018) New high-pressure experiments on sulfide saturation of high-FeO* basalts with variable TiO2 contents - Implications for the sulfur inventory of the lunar interior. Geochimica Cosmochimica Acta 222, 319–339. https://doi.org/10.1016/j.gca.2017.10.025
; Steenstra et al., 2020aSteenstra, E.S., Berndt, J., Klemme. S., Snape, J.F., Bullock. E.S., van Westrenen, W. (2020a) The fate of sulfur and chalcophile elements during crystallization of the lunar magma ocean. Journal of Geophysical Research: Planets 125, e2019JE006328. https://doi.org/10.1029/2019JE006328
,bSteenstra, E.S., Berndt, J., Klemme, S., Rohrbach, A., Bullock, E.S., van Westrenen, W. (2020b) An experimental assessment of the potential of sulfide saturation of the source regions of eucrites and angrites: implications for asteroidal models of core formation, late accretion and volatile element depletions. Geochimica et Cosmochimica Acta 269, 39–62. https://doi.org/10.1016/j.gca.2019.10.006
; Blanchard et al., 2021Blanchard, I., Abeykoon, S., Frost, D.J., Rubie, D.C. (2021) Sulfur content at sulfide saturation of peridotitic melt at upper mantle conditions. American Mineralogist 106, 1835–1843. https://doi.org/10.2138/am-2021-7649
).Experimental investigation of metal-silicate partitioning of S during core formation in the Earth suggests mildly siderophile behaviour of S (Dmet–silS = 10–55; Boujibar et al., 2014
Boujibar, A., Andrault, D., Bouhifd, M.A., Bolfan-Casanova, N., Devidal, J.-L., Trcera, N. (2014) Metal-silicate partitioning of sulphur, new experimental and thermodynamic constraints on planetary accretion. Earth and Planetary Science Letters 391, 42–54. https://doi.org/10.1016/j.epsl.2014.01.021
; Suer et al., 2017Suer, T.-A., Siebert, J., Remusat, L., Menguy, N., Fiquet, G. (2017) A sulfur-poor terrestrial core inferred from metal-silicate partitioning experiments. Earth and Planetary Science Letters 469, 84–97. https://doi.org/10.1016/j.epsl.2017.04.016
). Depending on the core formation scenario, considered S abundances may therefore have been relatively high after core segregation in the magma ocean (Rubie et al., 2016Rubie, D.C., Laurenz, V., Jacobson, S.A., Morbidelli, A., Palme, H., Vogel, A.K., Frost, D.J. (2016) Highly siderophile elements were stripped from Earth’s mantle by iron sulfide segregation. Science 353, 1141–1144. https://doi.org/10.1126/science.aaf6919
). It has been hypothesised that at some stage the S content of the magma ocean reached the SCSS, due to the incompatible behaviour of S (Callegaro et al., 2020Callegaro, S., Geraki, K., Marzoli, A., de Min, A., Maneta, V., Baker, D.R. (2020) The quintet completed: the partitioning of sulfur between nominally volatile-free minerals and silicate melts. American Mineralogist 105(5), 697–707. https://doi.org/10.2138/am-2020-7188
) and the strongly negative dependence of the SCSS on temperature, resulting in segregation of sulfide matte (‘’the Hadean matte’’; O’Neill, 1991O’Neill, H.St.C. (1991) The origin of the Moon and the early history of the Earth – A chemical model. Part 2: the Earth. Geochimica et Cosmochimica Acta 55, 1159–1172. https://doi.org/10.1016/0016-7037(91)90169-6
). Due to the importance of the Hadean matte for the deep S cycle and chalcophile element abundances, constraints on the SCSS and relative timing of sulfide segregation (during magma ocean crystallisation) are required. Currently, there are no SCSS measurements available at the P-T conditions that are directly relevant for a deep(er) magma ocean (>25 GPa; Huang et al., 2020Huang, D., Badro, J., Siebert, J. (2020) The niobium and tantalum concentration in the mantle constraints the composition of Earth´s primordial magma ocean. Proceedings of the National Academy of Sciences 117, 27893–27898. https://doi.org/10.1073/pnas.2007982117
), requiring significant extrapolations of lower pressure data. In addition, the sulfide liquids of many available higher-pressure datasets contain high (>5–15 %) amounts of other elements in addition to Fe-S-O, which will decrease the SCSS (Smythe et al., 2017Smythe, D.J., Wood, B.J., Kiseeva, E.S. (2017) The S content of silicate melts at sulfide saturation: New experiments and a model incorporating the effects of sulfide composition. American Mineralogist 102, 795–803. https://doi.org/10.2138/am-2017-5800CCBY
).To determine the SCSS in a deep terrestrial magma ocean, FeS-rich sulfide liquids and silicate MORB melts were equilibrated in 3 experiments at 43–53 GPa and 3925–4600 K by laser heating in a diamond anvil cell at the University of Bristol, UK (Table 1). A MORB composition was chosen to ensure that the silicate melt could be quenched to a glass. Run products were analysed using a JEOL JXA 8530F field emission electron microprobe at the Carnegie Institution for Science, USA. The reader is referred to the Supplementary Information file for additional details on experimental and analytical techniques.
Table 1 Experimental run conditions and results.
Experiment | P(GPa)a | P(GPa)c | T(K) | log FeO | SCSS (ppm) |
ESS-DAC-4 | 40 ± 2b | 53 ± 2b | 4605 ± 117 | 1.08(1) | 6979 ± 350 |
ESS-DAC-5 | 35 ± 2 | 43 ± 2 | 4300 ± 129 | 1.26(1) | 10837 ± 2124 |
ESS-DAC-7 | 38 ± 2 | 44 ± 2 | 3927 ± 37 | 1.42(1) | 11806 ± 780 |
aDefined as the average of the pre- and post-heating measured pressures. bPressure uncertainties are based on Walter et al. (2015)Walter, M.J., Thomson, A.R., Wang, W., Lord, O.T., Ross, J., McMahon, S.C., Baron, M.A., Melekhova, E., Kleppe, A.K., Kohn, S.C. (2015) The stability of hydrous silicates in Earth’s lower mantle: Experimental constraints from the systems MgO-SiO2-H2O and MgO-Al2O3-SiO2-H2O. Chemical Geology 418, 16–29. https://doi.org/10.1016/j.chemgeo.2015.05.001. cPost-heating pressures corrected upwards for thermal pressure effects and subsequently used in this study (Supplementary Information).
top
Results
The heated spots of the run products were characterised by homogeneous quenched silicate melts with abundant sub-micron quenched FeS droplets (Fig. 1; Supplementary Information). Sulfur contents of the silicate melts varied between 0.70 and 1.18 wt. % (Fig. 2) and FeO contents significantly increased relative to the starting composition, consistent with previous studies on basaltic melts (Blanchard et al., 2017
Blanchard, I., Siebert, J., Borensztajn S., Badro, J. (2017) The solubility of heat-producing elements in Earth’s core. Geochemical Perspective Letters 5, 1–5. https://doi.org/10.7185/geochemlet.1737
; Suer et al., 2021Suer, T.-A., Siebert, J., Remusat, L., Day, J.M.D., Borensztajn, S., Doisneau, B., Fiquet, G. (2021) Reconciling metal-silicate partitioning and late accretion in the Earth. Nature Communications 12, 2913. https://doi.org/10.1038/s41467-021-23137-5
). This is likely a result of small differences in the ratio of FeS to silicate within the heated region and variable degrees of perovskite crystallisation on the edges of the heated spot. The data reproduce a positive dependency between FeO content and the SCSS as thermodynamically and experimentally predicted from low P-T experiments (Wykes et al., 2015Wykes, J.L., O’Neill, H.St.C, Mavrogenes, J.A. (2015) The effect of FeO on the sulfur content at sulfide saturation (SCSS) and the selenium content at selenide saturation of silicate melts. Journal of Petrology 56, 1407–1424. https://doi.org/10.1093/petrology/egv041
), strongly suggesting sulfide saturation of the melts at high P-T (Supplementary Information, Fig. 2a). After normalisation of the SCSS to a common FeO content (xsil meltFeO = 0.05 or 8.1 wt. % FeO, corresponding to the present day terrestrial primitive mantle; Palme and O’Neill, 2014Palme, H., O’Neill, H.St.C (2014) 3.1 - Cosmochemical estimates of mantle composition. In: Holland, H.D., Turekian, K.K. (Eds.) Planets, Asteroids, Comets and The Solar System, Treatise of Geochemistry (Second Edition). Elsevier, Oxford, 1–39. https://doi.org/10.1016/B978-0-08-095975-7.00201-1
) using the FeO term of an existing SCSS model (Supplementary Information, section S.3), the effects of P-T on the SCSS at the conditions of a deep magma ocean are assessed (Fig. 2c–d). Our results confirm the increase of the SCSS with increasing T and the decrease of the SCSS with P (O’Neill and Mavrogenes, 2002O’Neill, H.St.C., Mavrogenes, J.A. (2002) The sulfide capacity and the sulfur content at sulfide saturation of silicate melts at 1400°C and 1 bar. Journal of Petrology 43, 1049–1087. https://doi.org/10.1093/petrology/43.6.1049
; Blanchard et al., 2021Blanchard, I., Abeykoon, S., Frost, D.J., Rubie, D.C. (2021) Sulfur content at sulfide saturation of peridotitic melt at upper mantle conditions. American Mineralogist 106, 1835–1843. https://doi.org/10.2138/am-2021-7649
), as seen in previous low P-T data (Ding et al., 2018Ding, S., Hough, T., Dasgupta R. (2018) New high-pressure experiments on sulfide saturation of high-FeO* basalts with variable TiO2 contents - Implications for the sulfur inventory of the lunar interior. Geochimica Cosmochimica Acta 222, 319–339. https://doi.org/10.1016/j.gca.2017.10.025
; Steenstra et al., 2018Steenstra, E.S., Seegers, A.X., Eising, J., Tomassen, B.G.J., Webers, F.P.F., Berndt, J., Klemme, S., Matveev, S., van Westrenen, W. (2018) Evidence for a sulfur undersaturated lunar interior from the solubility of sulfur in lunar melts and sulfide-silicate partitioning of siderophile elements. Geochimica et Cosmochimica Acta 231, 130–156. https://doi.org/10.1016/j.gca.2018.04.008
). Our FeO-normalised SCSS values are consistently higher than predicted by existing high-P peridotite models (Laurenz et al., 2016Laurenz, V., Rubie, D.C., Frost, D.J., Vogel, A.K. (2016) The importance of sulfur for the behavior of highly-siderophile elements during Earth’s differentiation. Geochimica et Cosmochimica Acta 194, 123–138. https://doi.org/10.1016/j.gca.2016.08.012
; Blanchard et al., 2021Blanchard, I., Abeykoon, S., Frost, D.J., Rubie, D.C. (2021) Sulfur content at sulfide saturation of peridotitic melt at upper mantle conditions. American Mineralogist 106, 1835–1843. https://doi.org/10.2138/am-2021-7649
), with an offset of up to +6700 ppm (Fig. 2b). The differences between our FeO-normalised SCSS values and modelled SCSS values are significantly larger when implementing the other published SCSS models. These models are based on a wide range of silicate compositions and, when calculated using raw FeO contents, the differences between our measured and calculated SCSS values are at least 5000 ppm and as high as 1.07 wt. % (Fig. 2b). This shows that the models are not reliable at the conditions of a deep magma ocean. However, because previous parameterisations (Laurenz et al., 2016Laurenz, V., Rubie, D.C., Frost, D.J., Vogel, A.K. (2016) The importance of sulfur for the behavior of highly-siderophile elements during Earth’s differentiation. Geochimica et Cosmochimica Acta 194, 123–138. https://doi.org/10.1016/j.gca.2016.08.012
; Blanchard et al., 2021Blanchard, I., Abeykoon, S., Frost, D.J., Rubie, D.C. (2021) Sulfur content at sulfide saturation of peridotitic melt at upper mantle conditions. American Mineralogist 106, 1835–1843. https://doi.org/10.2138/am-2021-7649
) were derived for a peridotitic melt, an assessment of the potential effects of non-FeO silicate melt variation on the SCSS is required. Using the Smythe et al. (2017)Smythe, D.J., Wood, B.J., Kiseeva, E.S. (2017) The S content of silicate melts at sulfide saturation: New experiments and a model incorporating the effects of sulfide composition. American Mineralogist 102, 795–803. https://doi.org/10.2138/am-2017-5800CCBY
model at 1873 K and 1 GPa, calculated SCSS values for a peridotitic melt (Palme and O’Neill, 2014Palme, H., O’Neill, H.St.C (2014) 3.1 - Cosmochemical estimates of mantle composition. In: Holland, H.D., Turekian, K.K. (Eds.) Planets, Asteroids, Comets and The Solar System, Treatise of Geochemistry (Second Edition). Elsevier, Oxford, 1–39. https://doi.org/10.1016/B978-0-08-095975-7.00201-1
) are ≈860–1250 ppm higher than for our experimental silicate melt compositions for 8.1 wt. % FeO. A higher SCSS calculated for peridotite is also generally consistent with the results of Laurenz et al. (2016)Laurenz, V., Rubie, D.C., Frost, D.J., Vogel, A.K. (2016) The importance of sulfur for the behavior of highly-siderophile elements during Earth’s differentiation. Geochimica et Cosmochimica Acta 194, 123–138. https://doi.org/10.1016/j.gca.2016.08.012
. The large offset of our measured SCSS values compared to predicted values confirms that either the positive T effects on SCSS were underestimated and/or negative P effects were overestimated in previous SCSS models.top
Discussion
Modelling the solubility of S in a terrestrial magma ocean requires knowledge of the variation of the SCSS with P-T (e.g., Boukare et al., 2019
Boukare, C.-E., Parman, S.W., Parmentier, E.M., Anzures, B.A. (2019) Production and Preservation of Sulfide Layering in Mercury´s Mantle. Journal of Geophysical Research: Planets 124, 3354–3372. https://doi.org/10.1029/2019JE005942
). High-P data for the SCSS of FeS-rich liquids are relatively scarce in the literature and predominantly derived for peridotite liquids. This prohibits a quantitative assessment of the effects of silicate melt composition at high P, which could be very different at such conditions, and constraining this would require tens, if not hundreds, of additional experiments at such conditions. It is also very likely that strong correlations exist between fitted melt composition parameters and P-T effects, given that such regressions contain up to 11 terms. Instead, we fitted our new data together with all previous SCSS data that was obtained for very similar compositions (Table S-2; Supplementary Information) to Equation 1:Prior to fitting, all data were normalised to a common FeO value of 8.1 wt. % (xsil meltFeO = 0.05) (Supplementary Information, section S.3). The SCSS does not vary significantly (200–300 ppm) within the FeO range relevant for terrestrial magma ocean crystallisation (2 to 8.1 wt. % FeO; Tagawa et al., 2021
Tagawa, S., Sakamoto, N., Hirose, K., Yokoo, S., Hernlund, J., Ohishi, Y., Yurimoto, H. (2021) Experimental evidence for hydrogen incorporation into Earth’s core. Nature Communications 12, 2588. https://doi.org/10.1038/s41467-021-22035-0
; Fig. 2a) and no FeO term is required for the parameterisation.Fitting FeO-normalised literature SCSS data obtained for silicate melts with very similar compositions (N = 42; Supplementary Information, section S.5) and assuming asulfideFeS = 1, yields:
The regression results suggest that the negative pressure effect on the SCSS is (significantly) smaller than previously reported (e.g., Blanchard et al., 2021
Blanchard, I., Abeykoon, S., Frost, D.J., Rubie, D.C. (2021) Sulfur content at sulfide saturation of peridotitic melt at upper mantle conditions. American Mineralogist 106, 1835–1843. https://doi.org/10.2138/am-2021-7649
), whereas the derived negative temperature term is considerably lower than the high-pressure models (Laurenz et al., 2016Laurenz, V., Rubie, D.C., Frost, D.J., Vogel, A.K. (2016) The importance of sulfur for the behavior of highly-siderophile elements during Earth’s differentiation. Geochimica et Cosmochimica Acta 194, 123–138. https://doi.org/10.1016/j.gca.2016.08.012
; Smythe et al., 2017Smythe, D.J., Wood, B.J., Kiseeva, E.S. (2017) The S content of silicate melts at sulfide saturation: New experiments and a model incorporating the effects of sulfide composition. American Mineralogist 102, 795–803. https://doi.org/10.2138/am-2017-5800CCBY
). It is, however, larger than currently available low-pressure models (Ding et al., 2018Ding, S., Hough, T., Dasgupta R. (2018) New high-pressure experiments on sulfide saturation of high-FeO* basalts with variable TiO2 contents - Implications for the sulfur inventory of the lunar interior. Geochimica Cosmochimica Acta 222, 319–339. https://doi.org/10.1016/j.gca.2017.10.025
; Steenstra et al., 2018Steenstra, E.S., Seegers, A.X., Eising, J., Tomassen, B.G.J., Webers, F.P.F., Berndt, J., Klemme, S., Matveev, S., van Westrenen, W. (2018) Evidence for a sulfur undersaturated lunar interior from the solubility of sulfur in lunar melts and sulfide-silicate partitioning of siderophile elements. Geochimica et Cosmochimica Acta 231, 130–156. https://doi.org/10.1016/j.gca.2018.04.008
) as well as the high-pressure model of Blanchard et al. (2021)Blanchard, I., Abeykoon, S., Frost, D.J., Rubie, D.C. (2021) Sulfur content at sulfide saturation of peridotitic melt at upper mantle conditions. American Mineralogist 106, 1835–1843. https://doi.org/10.2138/am-2021-7649
. Our modelling results thus suggest significantly higher SCSS values for the terrestrial magma ocean at high P-T conditions (Fig. 3a).top
Implications for the Terrestrial Sulfur Cycle
In Figure 3b, the new SCSS model is incorporated in Earth accretion models from previous studies (Rubie et al., 2016
Rubie, D.C., Laurenz, V., Jacobson, S.A., Morbidelli, A., Palme, H., Vogel, A.K., Frost, D.J. (2016) Highly siderophile elements were stripped from Earth’s mantle by iron sulfide segregation. Science 353, 1141–1144. https://doi.org/10.1126/science.aaf6919
; Tagawa et al., 2021Tagawa, S., Sakamoto, N., Hirose, K., Yokoo, S., Hernlund, J., Ohishi, Y., Yurimoto, H. (2021) Experimental evidence for hydrogen incorporation into Earth’s core. Nature Communications 12, 2588. https://doi.org/10.1038/s41467-021-22035-0
), while exploring different average effective pressures of sulfide saturation or kS (Eq. S-4). The range of kS considered here is based on the preferred value of Rubie et al. (2016)Rubie, D.C., Laurenz, V., Jacobson, S.A., Morbidelli, A., Palme, H., Vogel, A.K., Frost, D.J. (2016) Highly siderophile elements were stripped from Earth’s mantle by iron sulfide segregation. Science 353, 1141–1144. https://doi.org/10.1126/science.aaf6919
(kS = 0.44) while exploring the sensitivity of the results to different kS values. Figure 3b shows that for the mantle S contents modelled by Rubie et al. (2016)Rubie, D.C., Laurenz, V., Jacobson, S.A., Morbidelli, A., Palme, H., Vogel, A.K., Frost, D.J. (2016) Highly siderophile elements were stripped from Earth’s mantle by iron sulfide segregation. Science 353, 1141–1144. https://doi.org/10.1126/science.aaf6919
sulfide saturation in the magma ocean will most certainly occur at approximately 55 % of magma ocean crystallisation. This conclusion is virtually independent of the assumed effective pressure of FeS saturation and would imply a major S reservoir in the deep mantle in addition to the core itself. This is also consistent with mantle HSE systematics (Laurenz et al., 2016Laurenz, V., Rubie, D.C., Frost, D.J., Vogel, A.K. (2016) The importance of sulfur for the behavior of highly-siderophile elements during Earth’s differentiation. Geochimica et Cosmochimica Acta 194, 123–138. https://doi.org/10.1016/j.gca.2016.08.012
; Rubie et al., 2016Rubie, D.C., Laurenz, V., Jacobson, S.A., Morbidelli, A., Palme, H., Vogel, A.K., Frost, D.J. (2016) Highly siderophile elements were stripped from Earth’s mantle by iron sulfide segregation. Science 353, 1141–1144. https://doi.org/10.1126/science.aaf6919
). In contrast, the modelled mantle S evolution curves from Suer et al. (2017)Suer, T.-A., Siebert, J., Remusat, L., Menguy, N., Fiquet, G. (2017) A sulfur-poor terrestrial core inferred from metal-silicate partitioning experiments. Earth and Planetary Science Letters 469, 84–97. https://doi.org/10.1016/j.epsl.2017.04.016
are much lower relative to modelled SCSS values. Here, the mantle S content is insufficient to yield sulfide saturation at any stage of magma ocean crystallisation, especially if one considers the slightly higher SCSS values expected for a peridotite liquid (Fig. 3b). The large differences between the mantle S evolution models of Rubie et al. (2016)Rubie, D.C., Laurenz, V., Jacobson, S.A., Morbidelli, A., Palme, H., Vogel, A.K., Frost, D.J. (2016) Highly siderophile elements were stripped from Earth’s mantle by iron sulfide segregation. Science 353, 1141–1144. https://doi.org/10.1126/science.aaf6919
and Suer et al. (2017)Suer, T.-A., Siebert, J., Remusat, L., Menguy, N., Fiquet, G. (2017) A sulfur-poor terrestrial core inferred from metal-silicate partitioning experiments. Earth and Planetary Science Letters 469, 84–97. https://doi.org/10.1016/j.epsl.2017.04.016
are due to the fact that they considered very different accretion models. The first model considers accretion of a fraction of undifferentiated planetesimals (i.e. fully oxidised) with low degrees of terrestrial core-mantle re-equilibration, whereas in the second model all accreted bodies are differentiated and equilibrated at low P-T conditions and their cores experienced further partial re-equilibration in the deep terrestrial magma ocean. If modelled S abundances for the terrestrial magma ocean of Suer et al. (2017)Suer, T.-A., Siebert, J., Remusat, L., Menguy, N., Fiquet, G. (2017) A sulfur-poor terrestrial core inferred from metal-silicate partitioning experiments. Earth and Planetary Science Letters 469, 84–97. https://doi.org/10.1016/j.epsl.2017.04.016
are correct, our results imply that sulfide saturation could not have occurred during magma ocean crystallisation, or perhaps only very late (>99.9 %) when the very last residual liquid is extremely enriched in S. The absence of, or very late, sulfide saturation of the residual magma ocean is problematic in terms of transporting sulfide liquid to the deep mantle as proposed to explain HSE systematics (Rubie et al., 2016Rubie, D.C., Laurenz, V., Jacobson, S.A., Morbidelli, A., Palme, H., Vogel, A.K., Frost, D.J. (2016) Highly siderophile elements were stripped from Earth’s mantle by iron sulfide segregation. Science 353, 1141–1144. https://doi.org/10.1126/science.aaf6919
), given the limited percolation of FeS liquid through a crystalline upper mantle (Terasaki et al., 2008Terasaki, H., Frost, D.J., Rubie, D.C., Langenhorst, F (2008) Percolative core formation in planetesimals. Earth and Planetary Science Letters 273, 132–137. https://doi.org/10.1016/j.epsl.2008.06.019
). Although interconnection of FeS liquid occurs at lower mantle conditions (Shi et al., 2013Shi, C.Y., Zhang, L., Yang, W., Liu, Y., Wang, J., Meng, Y., Andrews, J., Mao, W. (2013) Formation of an interconnected network of iron melt at Earth´s lower mantle conditions. Nature Geoscience 6, 971–975. https://doi.org/10.1038/ngeo1956
), it is unlikely that such late segregated FeS liquid would be transportable to the lower mantle and that global HSE depletions would be established at such late stages of magma ocean crystallisation (Fig. 3b). The S evolution models of Suer et al. (2017)Suer, T.-A., Siebert, J., Remusat, L., Menguy, N., Fiquet, G. (2017) A sulfur-poor terrestrial core inferred from metal-silicate partitioning experiments. Earth and Planetary Science Letters 469, 84–97. https://doi.org/10.1016/j.epsl.2017.04.016
do reproduce the current S content of the bulk silicate Earth, and given the highly chalcophile affinities of the HSE (Laurenz et al., 2016Laurenz, V., Rubie, D.C., Frost, D.J., Vogel, A.K. (2016) The importance of sulfur for the behavior of highly-siderophile elements during Earth’s differentiation. Geochimica et Cosmochimica Acta 194, 123–138. https://doi.org/10.1016/j.gca.2016.08.012
), only very minor amounts of sulfides would be required to establish primitive mantle HSE depletions. Overall, our results show that either the magma ocean must have been very rich in S to achieve sulfide saturation as proposed to satisfy HSE abundance constraints (Rubie et al., 2016Rubie, D.C., Laurenz, V., Jacobson, S.A., Morbidelli, A., Palme, H., Vogel, A.K., Frost, D.J. (2016) Highly siderophile elements were stripped from Earth’s mantle by iron sulfide segregation. Science 353, 1141–1144. https://doi.org/10.1126/science.aaf6919
) or that, instead, iron disproportionation affected HSE systematics in the early Earth (Suer et al., 2021Suer, T.-A., Siebert, J., Remusat, L., Day, J.M.D., Borensztajn, S., Doisneau, B., Fiquet, G. (2021) Reconciling metal-silicate partitioning and late accretion in the Earth. Nature Communications 12, 2913. https://doi.org/10.1038/s41467-021-23137-5
).top
Acknowledgements
E.S.S. was funded through H2020 Marie Skłodowska-Curie Postdoctoral Fellowship 101020611 and by a Carnegie Postdoctoral Fellowship. O.T.L. acknowledges support from the Royal Society in the form of a University Research Fellowship (UF150057). We thank two anonymous reviewers for their useful comments which greatly improved the quality of the manuscript and would like to thank Maud Boyet for editorial handling.
Editor: Maud Boyet
top
References
Blanchard, I., Siebert, J., Borensztajn S., Badro, J. (2017) The solubility of heat-producing elements in Earth’s core. Geochemical Perspective Letters 5, 1–5. https://doi.org/10.7185/geochemlet.1737
Show in context
Sulfur contents of the silicate melts varied between 0.70 and 1.18 wt. % (Fig. 2) and FeO contents significantly increased relative to the starting composition, consistent with previous studies on basaltic melts (Blanchard et al., 2017; Suer et al., 2021).
View in article
Blanchard, I., Abeykoon, S., Frost, D.J., Rubie, D.C. (2021) Sulfur content at sulfide saturation of peridotitic melt at upper mantle conditions. American Mineralogist 106, 1835–1843. https://doi.org/10.2138/am-2021-7649
Show in context
The SCSS is a function of composition, most notably FeO, pressure (P) and temperature (T), and has been extensively studied at lower pressures (<24 GPa; O’Neill and Mavrogenes, 2002; Namur et al., 2016; Smythe et al., 2017; Ding et al., 2018; Steenstra et al., 2020a,b; Blanchard et al., 2021).
View in article
Our results confirm the increase of the SCSS with increasing T and the decrease of the SCSS with P (O’Neill and Mavrogenes, 2002; Blanchard et al., 2021), as seen in previous low P-T data (Ding et al., 2018; Steenstra et al., 2018).
View in article
Our FeO-normalised SCSS values are consistently higher than predicted by existing high-P peridotite models (Laurenz et al., 2016; Blanchard et al., 2021), with an offset of up to +6700 ppm (Fig. 2b).
View in article
However, because previous parameterisations (Laurenz et al., 2016; Blanchard et al., 2021) were derived for a peridotitic melt, an assessment of the potential effects of non-FeO silicate melt variation on the SCSS is required.
View in article
Grey and red lines indicate the P-T dependencies of the L16 peridotite SCSS model and our new model (Eq. 2), respectively. Literature data and/or previous models from Kiseeva and Wood (2013, 2015); Vogel et al. (2015); Laurenz et al. (2016); Smythe et al. (2017); Ding et al. (2018); Blanchard et al. (2021) (Table S-2).
View in article
The regression results suggest that the negative pressure effect on the SCSS is (significantly) smaller than previously reported (e.g., Blanchard et al., 2021), whereas the derived negative temperature term is considerably lower than the high-pressure models (Laurenz et al., 2016; Smythe et al., 2017).
View in article
It is, however, larger than currently available low-pressure models (Ding et al., 2018; Steenstra et al., 2018) as well as the high-pressure model of Blanchard et al. (2021).
View in article
Plotted for comparison are previous peridotite SCSS models (Laurenz et al., 2016; Blanchard et al., 2021).
View in article
Boujibar, A., Andrault, D., Bouhifd, M.A., Bolfan-Casanova, N., Devidal, J.-L., Trcera, N. (2014) Metal-silicate partitioning of sulphur, new experimental and thermodynamic constraints on planetary accretion. Earth and Planetary Science Letters 391, 42–54. https://doi.org/10.1016/j.epsl.2014.01.021
Show in context
Experimental investigation of metal-silicate partitioning of S during core formation in the Earth suggests mildly siderophile behaviour of S (Dmet–silS = 10–55; Boujibar et al., 2014; Suer et al., 2017).
View in article
Boukare, C.-E., Parman, S.W., Parmentier, E.M., Anzures, B.A. (2019) Production and Preservation of Sulfide Layering in Mercury´s Mantle. Journal of Geophysical Research: Planets 124, 3354–3372. https://doi.org/10.1029/2019JE005942
Show in context
Modelling the solubility of S in a terrestrial magma ocean requires knowledge of the variation of the SCSS with P-T (e.g., Boukare et al., 2019).
View in article
Callegaro, S., Geraki, K., Marzoli, A., de Min, A., Maneta, V., Baker, D.R. (2020) The quintet completed: the partitioning of sulfur between nominally volatile-free minerals and silicate melts. American Mineralogist 105(5), 697–707. https://doi.org/10.2138/am-2020-7188
Show in context
It has been hypothesised that at some stage the S content of the magma ocean reached the SCSS, due to the incompatible behaviour of S (Callegaro et al., 2020) and the strongly negative dependence of the SCSS on temperature, resulting in segregation of sulfide matte (‘’the Hadean matte’’; O’Neill, 1991).
View in article
Ding, S., Hough, T., Dasgupta R. (2018) New high-pressure experiments on sulfide saturation of high-FeO* basalts with variable TiO2 contents - Implications for the sulfur inventory of the lunar interior. Geochimica Cosmochimica Acta 222, 319–339. https://doi.org/10.1016/j.gca.2017.10.025
Show in context
The SCSS is a function of composition, most notably FeO, pressure (P) and temperature (T), and has been extensively studied at lower pressures (<24 GPa; O’Neill and Mavrogenes, 2002; Namur et al., 2016; Smythe et al., 2017; Ding et al., 2018; Steenstra et al., 2020a,b; Blanchard et al., 2021).
View in article
Our results confirm the increase of the SCSS with increasing T and the decrease of the SCSS with P (O’Neill and Mavrogenes, 2002; Blanchard et al., 2021), as seen in previous low P-T data (Ding et al., 2018; Steenstra et al., 2018).
View in article
It is, however, larger than currently available low-pressure models (Ding et al., 2018; Steenstra et al., 2018) as well as the high-pressure model of Blanchard et al. (2021).
View in article
Grey and red lines indicate the P-T dependencies of the L16 peridotite SCSS model and our new model (Eq. 2), respectively. Literature data and/or previous models from Kiseeva and Wood (2013, 2015); Vogel et al. (2015); Laurenz et al. (2016); Smythe et al. (2017); Ding et al. (2018); Blanchard et al. (2021) (Table S-2).
View in article
Huang, D., Badro, J., Siebert, J. (2020) The niobium and tantalum concentration in the mantle constraints the composition of Earth´s primordial magma ocean. Proceedings of the National Academy of Sciences 117, 27893–27898. https://doi.org/10.1073/pnas.2007982117
Show in context
Currently, there are no SCSS measurements available at the P-T conditions that are directly relevant for a deep(er) magma ocean (>25 GPa; Huang et al., 2020), requiring significant extrapolations of lower pressure data.
View in article
Laurenz, V., Rubie, D.C., Frost, D.J., Vogel, A.K. (2016) The importance of sulfur for the behavior of highly-siderophile elements during Earth’s differentiation. Geochimica et Cosmochimica Acta 194, 123–138. https://doi.org/10.1016/j.gca.2016.08.012
Show in context
Our FeO-normalised SCSS values are consistently higher than predicted by existing high-P peridotite models (Laurenz et al., 2016; Blanchard et al., 2021), with an offset of up to +6700 ppm (Fig. 2b).
View in article
However, because previous parameterisations (Laurenz et al., 2016; Blanchard et al., 2021) were derived for a peridotitic melt, an assessment of the potential effects of non-FeO silicate melt variation on the SCSS is required.
View in article
A higher SCSS calculated for peridotite is also generally consistent with the results of Laurenz et al. (2016).
View in article
Grey and red lines indicate the P-T dependencies of the L16 peridotite SCSS model and our new model (Eq. 2), respectively. Literature data and/or previous models from Kiseeva and Wood (2013, 2015); Vogel et al. (2015); Laurenz et al. (2016); Smythe et al. (2017); Ding et al. (2018); Blanchard et al. (2021) (Table S-2).
View in article
The regression results suggest that the negative pressure effect on the SCSS is (significantly) smaller than previously reported (e.g., Blanchard et al., 2021), whereas the derived negative temperature term is considerably lower than the high-pressure models (Laurenz et al., 2016; Smythe et al., 2017).
View in article
Plotted for comparison are previous peridotite SCSS models (Laurenz et al., 2016; Blanchard et al., 2021).
View in article
This is also consistent with mantle HSE systematics (Laurenz et al., 2016; Rubie et al., 2016).
View in article
The S evolution models of Suer et al. (2017) do reproduce the current S content of the bulk silicate Earth, and given the highly chalcophile affinities of the HSE (Laurenz et al., 2016), only very minor amounts of sulfides would be required to establish primitive mantle HSE depletions.
View in article
Kiseeva, E.S., Wood, B.J. (2013) A simple model for chalcophile element partitioning between sulphide and silicate liquids with geochemical applications. Earth and Planetary Science Letters 383, 68–81. https://doi.org/10.1016/j.epsl.2013.09.034
Show in context
Grey and red lines indicate the P-T dependencies of the L16 peridotite SCSS model and our new model (Eq. 2), respectively. Literature data and/or previous models from Kiseeva and Wood (2013, 2015); Vogel et al. (2015); Laurenz et al. (2016); Smythe et al. (2017); Ding et al. (2018); Blanchard et al. (2021) (Table S-2).
View in article
Kiseeva, E.S., Wood, B.J. (2015) The effects of composition and temperature on chalcophile and lithophile element partitioning into magmatic sulphides. Earth and Planetary Science Letters 424, 280–294. https://doi.org/10.1016/j.epsl.2015.05.012
Show in context
Sulfur (S) plays a key role in planetary geochemistry because of its ability to act as a major sink for elements when S is present as a sulfide (Kiseeva and Wood, 2015).
View in article
Grey and red lines indicate the P-T dependencies of the L16 peridotite SCSS model and our new model (Eq. 2), respectively. Literature data and/or previous models from Kiseeva and Wood (2013, 2015); Vogel et al. (2015); Laurenz et al. (2016); Smythe et al. (2017); Ding et al. (2018); Blanchard et al. (2021) (Table S-2).
View in article
Namur, O., Charlier, B., Holtz, F., Cartier, C., McCammon, C. (2016) Sulfur solubility in reduced mafic silicate melts: Implications for the speciation and distribution of sulfur on Mercury. Earth and Planetary Science Letters 448, 102–114. https://doi.org/10.1016/j.epsl.2016.05.024
Show in context
The SCSS is a function of composition, most notably FeO, pressure (P) and temperature (T), and has been extensively studied at lower pressures (<24 GPa; O’Neill and Mavrogenes, 2002; Namur et al., 2016; Smythe et al., 2017; Ding et al., 2018; Steenstra et al., 2020a,b; Blanchard et al., 2021).
View in article
O’Neill, H.St.C. (1991) The origin of the Moon and the early history of the Earth – A chemical model. Part 2: the Earth. Geochimica et Cosmochimica Acta 55, 1159–1172. https://doi.org/10.1016/0016-7037(91)90169-6
Show in context
It has been hypothesised that at some stage the S content of the magma ocean reached the SCSS, due to the incompatible behaviour of S (Callegaro et al., 2020) and the strongly negative dependence of the SCSS on temperature, resulting in segregation of sulfide matte (‘’the Hadean matte’’; O’Neill, 1991).
View in article
O’Neill, H.St.C., Mavrogenes, J.A. (2002) The sulfide capacity and the sulfur content at sulfide saturation of silicate melts at 1400°C and 1 bar. Journal of Petrology 43, 1049–1087. https://doi.org/10.1093/petrology/43.6.1049
Show in context
The SCSS is a function of composition, most notably FeO, pressure (P) and temperature (T), and has been extensively studied at lower pressures (<24 GPa; O’Neill and Mavrogenes, 2002; Namur et al., 2016; Smythe et al., 2017; Ding et al., 2018; Steenstra et al., 2020a,b; Blanchard et al., 2021).
View in article
Our results confirm the increase of the SCSS with increasing T and the decrease of the SCSS with P (O’Neill and Mavrogenes, 2002; Blanchard et al., 2021), as seen in previous low P-T data (Ding et al., 2018; Steenstra et al., 2018).
View in article
Palme, H., O’Neill, H.St.C (2014) 3.1 - Cosmochemical estimates of mantle composition. In: Holland, H.D., Turekian, K.K. (Eds.) Planets, Asteroids, Comets and The Solar System, Treatise of Geochemistry (Second Edition). Elsevier, Oxford, 1–39. https://doi.org/10.1016/B978-0-08-095975-7.00201-1
Show in context
After normalisation of the SCSS to a common FeO content (xsil meltFeO = 0.05 or 8.1 wt. % FeO, corresponding to the present day terrestrial primitive mantle; Palme and O’Neill, 2014) using the FeO term of an existing SCSS model (Supplementary Information, section S.3), the effects of P-T on the SCSS at the conditions of a deep magma ocean are assessed (Fig. 2c–d).
View in article
Using the Smythe et al. (2017) model at 1873 K and 1 GPa, calculated SCSS values for a peridotitic melt (Palme and O’Neill, 2014) are ≈860–1250 ppm higher than for our experimental silicate melt compositions for 8.1 wt. % FeO.
View in article
Rubie, D.C., Jacobson, S.A., Morbidelli, A., O’Brien, D.P., Young, E.D., de Vries, J., Nimmo, F., Palme, H., Frost, D.J. (2015) Accretion and differentiation of the terrestrial planets with implications for the compositions of early-formed Solar System bodies and accretion of water. Icarus 248, 89–108. https://doi.org/10.1016/j.icarus.2014.10.015
Show in context
(a) Variation of the SCSS during terrestrial magma ocean crystallisation along a geotherm ranging approximately midway between the peridotite solidus and liquidus (Rubie et al., 2015; Eqs. S-2,3).
View in article
Rubie, D.C., Laurenz, V., Jacobson, S.A., Morbidelli, A., Palme, H., Vogel, A.K., Frost, D.J. (2016) Highly siderophile elements were stripped from Earth’s mantle by iron sulfide segregation. Science 353, 1141–1144. https://doi.org/10.1126/science.aaf6919
Show in context
Depending on the core formation scenario, considered S abundances may therefore have been relatively high after core segregation in the magma ocean (Rubie et al., 2016).
View in article
(b) Calculated SCSS values for various average effective pressures of sulfide saturation (kS; Rubie et al., 2016; Eq. S-4) as a function of accreted mass.
View in article
In Figure 3b, the new SCSS model is incorporated in Earth accretion models from previous studies (Rubie et al., 2016; Tagawa et al., 2021), while exploring different average effective pressures of sulfide saturation or kS (Eq. S-4).
View in article
The range of kS considered here is based on the preferred value of Rubie et al. (2016) (kS = 0.44) while exploring the sensitivity of the results to different kS values.
View in article
Figure 3b shows that for the mantle S contents modelled by Rubie et al. (2016) sulfide saturation in the magma ocean will most certainly occur at approximately 55 % of magma ocean crystallisation.
View in article
This is also consistent with mantle HSE systematics (Laurenz et al., 2016; Rubie et al., 2016).
View in article
The large differences between the mantle S evolution models of Rubie et al. (2016) and Suer et al. (2017) are due to the fact that they considered very different accretion models.
View in article
The absence of, or very late, sulfide saturation of the residual magma ocean is problematic in terms of transporting sulfide liquid to the deep mantle as proposed to explain HSE systematics (Rubie et al., 2016), given the limited percolation of FeS liquid through a crystalline upper mantle (Terasaki et al., 2008).
View in article
Overall, our results show that either the magma ocean must have been very rich in S to achieve sulfide saturation as proposed to satisfy HSE abundance constraints (Rubie et al., 2016) or that, instead, iron disproportionation affected HSE systematics in the early Earth (Suer et al., 2021).
View in article
Shi, C.Y., Zhang, L., Yang, W., Liu, Y., Wang, J., Meng, Y., Andrews, J., Mao, W. (2013) Formation of an interconnected network of iron melt at Earth´s lower mantle conditions. Nature Geoscience 6, 971–975. https://doi.org/10.1038/ngeo1956
Show in context
Although interconnection of FeS liquid occurs at lower mantle conditions (Shi et al., 2013), it is unlikely that such late segregated FeS liquid would be transportable to the lower mantle and that global HSE depletions would be established at such late stages of magma ocean crystallisation (Fig. 3b).
View in article
Smythe, D.J., Wood, B.J., Kiseeva, E.S. (2017) The S content of silicate melts at sulfide saturation: New experiments and a model incorporating the effects of sulfide composition. American Mineralogist 102, 795–803. https://doi.org/10.2138/am-2017-5800CCBY
Show in context
The SCSS is a function of composition, most notably FeO, pressure (P) and temperature (T), and has been extensively studied at lower pressures (<24 GPa; O’Neill and Mavrogenes, 2002; Namur et al., 2016; Smythe et al., 2017; Ding et al., 2018; Steenstra et al., 2020a,b; Blanchard et al., 2021).
View in article
In addition, the sulfide liquids of many available higher-pressure datasets contain high (>5–15 %) amounts of other elements in addition to Fe-S-O, which will decrease the SCSS (Smythe et al., 2017).
View in article
Using the Smythe et al. (2017) model at 1873 K and 1 GPa, calculated SCSS values for a peridotitic melt (Palme and O’Neill, 2014) are ≈860–1250 ppm higher than for our experimental silicate melt compositions for 8.1 wt. % FeO.
View in article
Grey and red lines indicate the P-T dependencies of the L16 peridotite SCSS model and our new model (Eq. 2), respectively. Literature data and/or previous models from Kiseeva and Wood (2013, 2015); Vogel et al. (2015); Laurenz et al. (2016); Smythe et al. (2017); Ding et al. (2018); Blanchard et al. (2021) (Table S-2).
View in article
The regression results suggest that the negative pressure effect on the SCSS is (significantly) smaller than previously reported (e.g., Blanchard et al., 2021), whereas the derived negative temperature term is considerably lower than the high-pressure models (Laurenz et al., 2016; Smythe et al., 2017).
View in article
Steenstra, E.S., Seegers, A.X., Eising, J., Tomassen, B.G.J., Webers, F.P.F., Berndt, J., Klemme, S., Matveev, S., van Westrenen, W. (2018) Evidence for a sulfur undersaturated lunar interior from the solubility of sulfur in lunar melts and sulfide-silicate partitioning of siderophile elements. Geochimica et Cosmochimica Acta 231, 130–156. https://doi.org/10.1016/j.gca.2018.04.008
Show in context
Our results confirm the increase of the SCSS with increasing T and the decrease of the SCSS with P (O’Neill and Mavrogenes, 2002; Blanchard et al., 2021), as seen in previous low P-T data (Ding et al., 2018; Steenstra et al., 2018).
View in article
Newly derived SCSS values and FeO-normalised SCSS values (normalised to 8.1 wt. % FeO or xsil meltFeO = 0.05, using the model of Steenstra et al. (2018) (Supplementary Information).
View in article
It is, however, larger than currently available low-pressure models (Ding et al., 2018; Steenstra et al., 2018) as well as the high-pressure model of Blanchard et al. (2021).
View in article
Steenstra, E.S., Berndt, J., Klemme. S., Snape, J.F., Bullock. E.S., van Westrenen, W. (2020a) The fate of sulfur and chalcophile elements during crystallization of the lunar magma ocean. Journal of Geophysical Research: Planets 125, e2019JE006328. https://doi.org/10.1029/2019JE006328
Show in context
The SCSS is a function of composition, most notably FeO, pressure (P) and temperature (T), and has been extensively studied at lower pressures (<24 GPa; O’Neill and Mavrogenes, 2002; Namur et al., 2016; Smythe et al., 2017; Ding et al., 2018; Steenstra et al., 2020a,b; Blanchard et al., 2021).
View in article
Steenstra, E.S., Berndt, J., Klemme, S., Rohrbach, A., Bullock, E.S., van Westrenen, W. (2020b) An experimental assessment of the potential of sulfide saturation of the source regions of eucrites and angrites: implications for asteroidal models of core formation, late accretion and volatile element depletions. Geochimica et Cosmochimica Acta 269, 39–62. https://doi.org/10.1016/j.gca.2019.10.006
Show in context
The SCSS is a function of composition, most notably FeO, pressure (P) and temperature (T), and has been extensively studied at lower pressures (<24 GPa; O’Neill and Mavrogenes, 2002; Namur et al., 2016; Smythe et al., 2017; Ding et al., 2018; Steenstra et al., 2020a,b; Blanchard et al., 2021).
View in article
Suer, T.-A., Siebert, J., Remusat, L., Menguy, N., Fiquet, G. (2017) A sulfur-poor terrestrial core inferred from metal-silicate partitioning experiments. Earth and Planetary Science Letters 469, 84–97. https://doi.org/10.1016/j.epsl.2017.04.016
Show in context
Experimental investigation of metal-silicate partitioning of S during core formation in the Earth suggests mildly siderophile behaviour of S (Dmet–silS = 10–55; Boujibar et al., 2014; Suer et al., 2017).
View in article
In contrast, the modelled mantle S evolution curves from Suer et al. (2017) are much lower relative to modelled SCSS values.
View in article
The large differences between the mantle S evolution models of Rubie et al. (2016) and Suer et al. (2017) are due to the fact that they considered very different accretion models.
View in article
If modelled S abundances for the terrestrial magma ocean of Suer et al. (2017) are correct, our results imply that sulfide saturation could not have occurred during magma ocean crystallisation, or perhaps only very late (>99.9 %) when the very last residual liquid is extremely enriched in S.
View in article
The S evolution models of Suer et al. (2017) do reproduce the current S content of the bulk silicate Earth, and given the highly chalcophile affinities of the HSE (Laurenz et al., 2016), only very minor amounts of sulfides would be required to establish primitive mantle HSE depletions.
View in article
Suer, T.-A., Siebert, J., Remusat, L., Day, J.M.D., Borensztajn, S., Doisneau, B., Fiquet, G. (2021) Reconciling metal-silicate partitioning and late accretion in the Earth. Nature Communications 12, 2913. https://doi.org/10.1038/s41467-021-23137-5
Show in context
Sulfur contents of the silicate melts varied between 0.70 and 1.18 wt. % (Fig. 2) and FeO contents significantly increased relative to the starting composition, consistent with previous studies on basaltic melts (Blanchard et al., 2017; Suer et al., 2021).
View in article
Overall, our results show that either the magma ocean must have been very rich in S to achieve sulfide saturation as proposed to satisfy HSE abundance constraints (Rubie et al., 2016) or that, instead, iron disproportionation affected HSE systematics in the early Earth (Suer et al., 2021).
View in article
Tagawa, S., Sakamoto, N., Hirose, K., Yokoo, S., Hernlund, J., Ohishi, Y., Yurimoto, H. (2021) Experimental evidence for hydrogen incorporation into Earth’s core. Nature Communications 12, 2588. https://doi.org/10.1038/s41467-021-22035-0
Show in context
The SCSS does not vary significantly (200–300 ppm) within the FeO range relevant for terrestrial magma ocean crystallisation (2 to 8.1 wt. % FeO; Tagawa et al., 2021; Fig. 2a) and no FeO term is required for the parameterisation.
View in article
In Figure 3b, the new SCSS model is incorporated in Earth accretion models from previous studies (Rubie et al., 2016; Tagawa et al., 2021), while exploring different average effective pressures of sulfide saturation or kS (Eq. S-4).
View in article
Terasaki, H., Frost, D.J., Rubie, D.C., Langenhorst, F (2008) Percolative core formation in planetesimals. Earth and Planetary Science Letters 273, 132–137. https://doi.org/10.1016/j.epsl.2008.06.019
Show in context
The absence of, or very late, sulfide saturation of the residual magma ocean is problematic in terms of transporting sulfide liquid to the deep mantle as proposed to explain HSE systematics (Rubie et al., 2016), given the limited percolation of FeS liquid through a crystalline upper mantle (Terasaki et al., 2008).
View in article
Vogel, A.K. (2015) Siderophile element partitioning at high pressures and temperatures: implications for core formation processes. PhD thesis, Universität Bayreuth, Germany. http://nbn-resolving.org/urn:nbn:de:bvb:703-epub-2039-3.
Show in context
Grey and red lines indicate the P-T dependencies of the L16 peridotite SCSS model and our new model (Eq. 2), respectively. Literature data and/or previous models from Kiseeva and Wood (2013, 2015); Vogel et al. (2015); Laurenz et al. (2016); Smythe et al. (2017); Ding et al. (2018); Blanchard et al. (2021) (Table S-2).
View in article
Walter, M.J., Thomson, A.R., Wang, W., Lord, O.T., Ross, J., McMahon, S.C., Baron, M.A., Melekhova, E., Kleppe, A.K., Kohn, S.C. (2015) The stability of hydrous silicates in Earth’s lower mantle: Experimental constraints from the systems MgO-SiO2-H2O and MgO-Al2O3-SiO2-H2O. Chemical Geology 418, 16–29. https://doi.org/10.1016/j.chemgeo.2015.05.001
Show in context
Defined as the average of the pre- and post-heating measured pressures. b Pressure uncertainties are based on Walter et al. (2015).
View in article
Wykes, J.L., O’Neill, H.St.C, Mavrogenes, J.A. (2015) The effect of FeO on the sulfur content at sulfide saturation (SCSS) and the selenium content at selenide saturation of silicate melts. Journal of Petrology 56, 1407–1424. https://doi.org/10.1093/petrology/egv041
Show in context
The data reproduce a positive dependency between FeO content and the SCSS as thermodynamically and experimentally predicted from low P-T experiments (Wykes et al., 2015), strongly suggesting sulfide saturation of the melts at high P-T (Supplementary Information, Fig. 2a).
View in article
top
Supplementary Information
The Supplementary Information includes:
- S.1 Experimental and Analytical Details
- S.2 Evidence for Sulfide-saturation of Experimental Silicate Melts
- S.3 Correcting SCSS values for Variable Silicate Melt FeO Contents
- S.4 Non FeO-compositional Effects on Derived SCSS values
- S.5 Dataset used for Parameterisations
- S.6 Modelling Approach
- Tables S-1 and S-2
- Figures S-1 to S-3
- Supplementary Information References
Download Table S-2 (Excel).
Download the Supplementary Information (PDF).