Effect of sulfur on the melt structure of reduced melts

Mercury’s extremely low oxygen fugacity (~IW-3 to IW-7) produces silicate melts with anomalously high sulfur contents. These sulfur-rich melts exhibit significantly lower viscosities than sulfur-free melts of the same composition [1]. This has important implications for magmatic processes including magma ocean dynamics, crystal settling, and volcanic eruption style. However, a molecular/structural mechanism responsible for this viscosity reduction has not been fully characterized. The observation of Si-S bonds first identified in reduced mercurian melts by Pommier et al. [2] using ²⁹Si nuclear magnetic resonance (NMR) presents an interesting target that suggests that changes in the Si network bonding environment might be a powerful control on the changes in transport properties of such melts. Here we build on this work with a systematic suite of experiments designed to isolate the effects of sulfur content and oxygen fugacity (fO₂) on melt structure.

Experiments were synthesized at 1 GPa across two fO₂ conditions (~IW-5.5 and IW-4) with varying sulfur contents, using a Na-rich composition approximating the Northern Smooth Plains. This design allows us to examine the effects of sulfur content and fO₂ independently, providing a systematic view of the melt structure in highly reduced conditions.

²⁹Si NMR spectra reveal a prominent peak near -93 ppm corresponding to Si-O bonds in silica tetrahedra units. The specific frequency indicates predominantly Q³ species (three bridging oxygens and one non-bridging oxygen). With increasing sulfur content, a broad shoulder near -56 ppm emerges and grows in intensity. This feature is consistent with the presence of Si-S bonding, arising from S^2- substituting for O^2- within the tetrahedral unit. The proportion of Si-S bonds scales with bulk sulfur content, suggesting that most or all dissolved sulfur in these melts is accommodated through bonding with Si rather than existing as a free metal sulfide phase.

Raman spectroscopy provides complementary constraints on sulfur speciation. Spectra identify multiple sulfide species within the melt, including evidence for Si-S bonding consistent with the NMR results. The distribution of sulfide species varies with both sulfur content and fO₂. Comparison across the two fO₂ suites allows us to disentangle the effects of sulfur abundance from those of redox states on the structural configuration of sulfur in the melt.

The structural implications of Si-S bond formation are significant. Elementally, the substitution of S^2- for O^2- in tetrahedral units is not radical, as such substitution is broadly recognized in solid state and materials chemistry [3]. However, one would not expect S^2- to behave identically to O^2-. Sulfur may favor forming complexes with network-modifying metals such as Mg, Ca, and Fe as readily as forming Si-S bonds. Where Si-S bonds do form, non-bridging configurations charge-balanced with these metals may be preferred over Si-S-Si bridging. This is supported by S K-edge XANES analyses on similar compositions, which show that at the relevant oxygen fugacities, sulfur preferentially forms complexes with Ca and Mg in the melt, favoring Si-S-Mg/Ca configurations [4]. In any case, any bond formed with S^2- will be weaker than its O^2- equivalent, and substituting S^2- for O^2- is therefore likely to have profound effects on transport properties such as viscosity.

The shift in NMR peak position from Si-O to Si-S environments is larger than would be expected from a simple change in bridging versus non-bridging oxygen speciation. This indicates that sulfur dissolution drives substantial reorganization of the melt network. Comparison between the IW-5.5 and IW-4 sample suites will test whether fO₂ alone, independent of sulfur content, produces measurable changes in Si coordination, and whether higher fO₂ suppresses Si-S bond formation even at equivalent bulk sulfur contents.

This work contributes to a broader understanding of how highly reducing planetary environments produce melts with distinct physical properties. Viscosity is a first-order control on magma transport and volcanic behavior, and characterizing the structural basis for reduced-melt viscosities is essential for interpreting Mercury’s volcanic history and the dynamics of its early magma ocean. 

 

[1] Mouser, M. D. et al. (2021) JGRP, 126. [2] Pommier, A. et al. (2023) GCA, 363, 114-128. [3] Asahi, T. et al. (1999) Korean J. Ceramics, 5, 178-182. [4] Anzures, B. A. et al. (2020) GCA, 286, 1-18.