Carbon dioxide in silicate melts: A molecular dynamics simulation study

Abstract

The distribution, recycling and storage of carbon in the Earth are of fundamental importance to understand the global carbon cycle between the deep Earth and near surface reservoirs. Degassing of CO2 at mid-ocean ridges may give information on the source region but the very low solubility of CO2 in tholeitic basalts has for consequence that near all Mid-Ocean Ridge Basalts glasses exsolve their CO2 rich vapor at shallow depth as they approach the ocean floor. Hence their CO2 contents mostly represent the pressure at eruption and not the source region. Recent petrological investigations have shown that the presence of carbonates at depth in the upper mantle has a large effect on the solidus of carbonated silicates by inducing incipient melting at much lower temperature. So the role of carbon-rich melts at great depth is now becoming a credible scenario to explain the extraction of CO2 from the source region to the surface. During the last three decades many studies have been devoted to measure the solubility of CO2 in silicate melts of various composition. But due to experimental difficulties these studies were generally restricted to low and moderate pressures (below ~20 kbar). By performing a series of molecular dynamics simulation where a supercritical CO2 phase is in contact with a silicate melt of various composition (from felsic to ultrabasic) at different temperatures (1473-2273K) and pressures (20-150kbar), we have been able to evaluate the solubility of CO2, the population of molecular and carbonate species, their diffusivity through the melt and the local structure. We show that this kind of molecular simulation is a useful theoretical guide to better understand the behavior of CO2 in magmas at depth.Comment: 109 pages, 14 figures. To be published in Geochimica et Cosmochimica Act

An H2O–CO2 mixed fluid saturation model compatible with rhyolite-MELTS

“…Stolper (1985, 1986 and thereafter many others: see the excellent summary in Guillot and Sator 2011) suggest, from FTIR studies of quenched glass experimental run products, that carbon dioxide exists in the melt as two distinct species: molecular CO 2 and carbonate ions, the latter presumably associated principally with alkali metal and alkali earth cations. Theirs and subsequent studies also infer the predominance of molecular CO 2 in rhyolitic and more siliceous compositions with increasing importance of carbonate at decreasing SiO 2 concentration.…”

“…Indeed, that assumption has proved successful in previous models of CO 2 solubility in silicate liquids (Spera and Bergman 1980;Papale 1997Papale , 1999Papale et al 2006;Duan 2014), although alternate formulations that embody speciation (e.g., Dixon et al 1995;Newman and Lowenstern 2002) are equally effective. The recent molecular dynamics study of Guillot and Sator (2011) has explored molecular and carbonate speciation as a function of T, P, and composition. They find that increasing temperature stabilizes the carbonate species as does pressure, but that in simulations for kimberlitic, basaltic, and rhyolitic composition liquids below 2 GPa (all at a temperature of 2000 K), molecular CO 2 is the predominant melt species-essentially controlling the solubility of dissolved carbon dioxide.…”

“…Recently, Duncan and Agee (2011) determined, using the sink-float experimental method, a value for v o CO 2 for molten peridotite at 2123 K in the pressure range 4.3-5.5 GPa to be 23.71 cc/mol to 22.06 cc/mol, respectively, which yields their model estimate for v o CO 2 at 1 bar as 36.57 cc/mol. Guillot and Sator (2011) compute from first principles a value of ~25.5 cc/mol for kimberlitic liquid and ~30.5 cc/mol for MORB liquid (both at 2000 K).…”

The compressibility of CaCO3–Li2CO3–Na2CO3–K2CO3 liquids: application to natrocarbonatite and CO2-bearing nephelinite liquids from Oldoinyo Lengai

“…More recently, however, it has been shown that at high temperature depolymerized melts contain substantial molecular CO 2 (Nowak et al 2003;Ni and Keppler 2013) and that at high pressure polymerized melts contain significant amounts of the carbonate ion (e.g., Mallik and Dasgupta 2014). Although numerous experimental studies show that the solubility of the CO 2 component is highest in depolymerized melts and increases with the abundance of non-bridging oxygens (NBO) relative to T (tetrahedral) sites in the melt (e.g., Mysen et al 1975;Brey and Green 1975;Eggler 1978;Spera and Bergman 1980;Pan et al 1991;Thibault and Holloway 1994;Behrens et al 2004Behrens et al , 2009Vetere et al 2014), a more recent molecular dynamic simulation (Guillot and Sator 2011) shows that at high pressure (~10 GPa) and high temperature (~2000 °C), there is only a weak dependence of silicate melt composition on bulk CO 2 solubility.…”

“…b Same as a except converted to density by dividing V CO 2 and V H 2 O into their respective molecular weights (44.01 and 18.015 g/mol) 3 Page 14 of 18 ions, although there are conflicting results on the order of importance. There is widespread agreement that Mg 2+ , Ca 2+ , Na + , and K + have the strongest effect on increasing the solubility of carbonate in silicate melts, although their relative effect is debated (e.g., Dasgupta et al 2007;Moore 2008;Behrens et al 2009;Lesne et al 2011;Guillot and Sator 2011;Iacono-Marziano et al 2012;Vetere et al 2011Vetere et al , 2014Morizet et al 2010Morizet et al , 2014Dasgupta 2013, 2014). It is inferred from solubility studies (e.g., Brooker et al 2001a) and phase-equilibrium results (e.g., Dasgupta 2013, 2014) that Fe 2+ is less important than the other cations, and the reason may be found in spectroscopic evidence that Fe 2+ undergoes coordination change (from sixfold to fivefold and fourfold) and enters tetrahedral (T) sites in highly depolymerized (high NBO/T) silicate melts (e.g., Brooker et al 2001b).…”