Abstract:
The effect of compression on noble gas solubility in silicate melts is still badly understood due to a lack of theoretical guidance. In the experimental literature, noble gases dissolving in liquid silicates are found to concentrate almost linearly with increasing pressure up to several tens of kbar, suggesting that Henry’s law could be valid up to very high pressures, although this law stipulates that the gaseous phase in contact with the liquid must be ideal. Recently, new experiments dealing with the dissolution of argon in synthetic and natural silicate melts have pointed out that the evolution of concentration with pressure exhibits a departure from linearity in the 50-100 kbar range, leading either to a levelling off or to a sudden collapse of the argon concentration above 50 kbar. Here, we investigate by means of liquid state physics how volatile species dissolve into silicate melts under pressure. We use a hard sphere model (the reference fluid in liquid state physics) to describe silicate melts and gas at high pressures. One of our main results is that, when pressure increases, the concurrent compaction of gas and melt explains the almost-linear behaviour of the noble gas concentration up to several tens of kbars, before melt compaction dominates and concentration either levels off or decreases gradually in the 50-100 kbar range. In spite of the existence of a quasi-linear regime over a large pressure range, our work disqualifies the use of the Henry law when dealing with high pressures. The implication of these findings to provide an understanding of degassing at mid-ocean ridges is next investigated. Applying our model to the scenario where CO2 vesicle generation occurs in the magma at mantle depths during its ascent from melting regions, we evaluate magma vesicularity as well as noble gas concentrations in the basalt melt and in vesicles as a function of pressure at depth. It is stressed that the variable and usually strong noble gas elemental fractionation observed in mid-ocean ridge basalts can be explained by assuming a sequence of several vesiculation stages interrupted by vesicle loss during magma ascent. ? 2005 Elsevier Inc. All rights reserved.
