Silicate melt properties and volcanic eruptions

Abstract

[1] Knowledge about the properties of silicate melts is needed by volcanologists and petrologists to evaluate the dynamics of volcanic eruptions and magmatic processes. These properties include the solubility and diffusivity of volatile components in silicate melts, silicate melt viscosity, and the fragmentation condition. Data and models of each property are reviewed and assessed. For rhyolitic melts many properties are sufficiently well known to allow realistic modeling of volcanic and magmatic processes. One interesting example is the role of speciation in the solubility and diffusivity of H 2 O and CO 2 . Even though both H 2 O and CO 2 are present in silicate melts as at least two species, the complexity in the solubility and diffusion behavior of H 2 O and the simplicity of CO 2 are due to differences in the speciation reaction: For the H 2 O component the stoichiometric coefficient is one for one hydrous species (molecular H 2 O) but is two for the other hydrous species (OH) in the species interconversion reaction, whereas for CO 2 the stoichiometric coefficients for all carbon species are one. The investigation of the species reaction not only helps in understanding the solubility and diffusion behavior, but the reaction among the hydrous species also serves as a geospeedometer (cooling rate indicator) for hydrous rhyolitic pyroclasts and glass and provides a method to infer viscosity. For melts other than rhyolite, a preliminary description of their properties is also available, but much more experimental and modeling work is necessary to quantify these properties more accurately.

NanoSIMS results from olivine-hosted melt embayments: Magma ascent rate during explosive basaltic eruptions

“…The parameterization we developed here leads to diffusivities that are within the +/-0.2 log units uncertainty of the viscosity-based model of Persikov et al (2010). This is no surprise, as their model in the region of interest is also highly constrained by the same experimental diffusivities we use here (Zhang et al 2007;and Behrens et al, 2004), and melt viscosity is highly related to NBO/T.…”

“…The diffusivity is well characterized for Fuego melts with a pressure-and H 2 O-dependent Arrhenius equation (Fig. 7; Zhang et al 2007):…”

“…We use an Arrhenius equation derived from experiments conducted on Etna and Stromboli basalts for temperature, pressure, and H 2 O concentrations that are relevant to the Fuego magma (Fig. 7, Zhang et al, 2007): eq.6 where D S is in m 2 /s. In these experiments the oxygen fugacity was 3 log units below QFM and sulfur was present primarily as sulfide.…”

Magma decompression rates during explosive eruptions of Kīlauea volcano, Hawaii, recorded by melt embayments

“…For embayments with bubbles attached to the embayment mouth, this diffusion, while a consequence of the drop in solubility during decompression, is driven by the concentration of dissolved volatiles at the melt-bubble interface, a direct function of vapor pressure inside the bubble. The diffusivities of H 2 O, CO 2 , and S in silicate melts are sufficiently small to prevent the attainment of equilibrium concentrations within embayments that have lengths on the order of 100 microns for a wide range of decompression rates (e.g., Lensky et al, 2004;Gonnerman and Manga, 2005;Pichavant et al, 2013;Zhang et al, 2007Zhang et al, , 2010. Consequently, distinct diffusion-controlled concentration profiles may be preserved within an embayment, allowing the estimation of decompression rates via diffusion modeling (Liu et al, 2007).…”

“…Magma temperature was found using the olivine-liquid thermometer of Helz and Thornber (1987) and was assumed constant. Although the magma may undergo cooling of about 100℃ upon ascent (e.g., Sahagian and Proussevitch, 1996;Mastin and Ghiorso, 2001), the corresponding change in volatile diffusivity is approximately a factor of 1/2 for H 2 O, CO 2 and S each (Freda et al, 2005;Zhang et al, 2007Zhang et al, , 2010. Because cooling will be most pronounced in the upper few hundred meters of the conduit, where most of the H 2 O exsolves and vapor expansion is largest, the effect of cooling is likely no more or less than a factor of two on decompression-rate estimates.…”