MODELING S DEGASSING DURING MAGMA ASCENT

The degassing of CO₂ and S from arc volcanoes is fundamentally important to global climate, eruption forecasting, and cycling of volatiles through subduction zones. While the behavior of CO₂-H₂O is relatively well understood from thermodynamic/empirical models¹, significant discrepancies exist in predicting the behavior of sulfur. In this study, we aim to develop a degassing model to track the evolution of volatiles and redox state in melt and co-existing vapor in ascending mafic-intermediate magma.

This degassing model takes an approach similar to previous studies², by combining an existing H₂O-CO₂ degassing model¹ with sulfur partitioning between vapor and silicate melt (Kd_S^fluid-melt), assuming the magma is not sulfide or sulfate saturated. With known H₂O, CO₂ in the melt and vapor, and initial S in the melt, S in the vapor can be calculated from Kd_S^fluid-melt. Mass fractions of vapor, melt and crystals are constrained using mass balance of H₂O or CO₂, and K₂O as a perfectly incompatible element in both crystals and vapor. Sulfur evolution in the melt can then be tracked along the degassing path by mass balance.

However, because sulfur exists as S^2- and S⁶⁺ in the melt and H₂S and SO₂ in the vapor, it is difficult to develop a bulk Kd_S^fluid-melt that applies over the range of fO₂ encountered in arc magmas. We thus consider separate Kds for the four relevant reactions: I. S^2-(m)→H₂S (v), Ia. S^2- (m)→SO₂ (v), II. SO₄^2-(m)→SO₂ (v), and IIa. SO₄^2-(m)→H₂S (v). We examined experimental studies relevant to parental arc magmas and parameterized Kd(I), Kd(Ia), and Kd(II) as functions of T, P, fO₂, H₂O and melt composition, and also used constraints on H₂S/SO₂ calculated from the equilibrium gas reaction H₂S + (3/2)O₂ → SO₂ + H₂O. There are no experiments that constrain Kd(IIa), but it can be calculated from the three other Kds, as can the Kd_{total S^fluid-melt}. Applying the Kds to the 1974 eruption of Fuego gives between 20-80, agreeing with that derived from fitting the Fuego melt inclusion CO₂-S-H₂O data trend³. The advantage of considering all four S reactions is that we can track fO₂, S isotopes and Fe³⁺/Fe_T using the S⁶⁺ and S^2- in the melt⁴. We plan to develop such an open source model that can be applied broadly to degassing magmas.

1. Ghiorso and Gualda (2015), CMP; 2. Sisson and Layne (1993) EPSL; 3. Rasmussen et al. (2020), Am.Min., in press. 4. Nash et al. (2019) EPSL.