Modeling Post-Entrapment Modification of Volatile Contents in Olivine-Hosted Melt Inclusions from Mid-Ocean Ridges

Abstract:

The presence of small amounts of H₂O in the mantle beneath mid-ocean ridges (MORs) critically affects melting dynamics, resulting in increased depth of melting, total melt production, and crustal thickness. One method for estimating the amount of H₂O and other volatiles in the oceanic mantle is analysis of olivine-hosted melt inclusions. Recent studies, however, demonstrate the efficiency of post-entrapment diffusive modification of H₂O, so that caution must be taken when interpreting volatile concentrations of included melts. In addition, diffusive loss of H₂O results in a decrease in the internal pressure of the melt inclusion, exsolution of CO₂ from the included melt, and formation of a mixed-volatile vapor bubble. Although studies examining diffusive modification of melt inclusion volatile contents have focused on high-H₂O melt inclusions from subduction zones, low-H₂O melt inclusions from MORs are also potentially susceptible to post-entrapment diffusive modification if the external magma experiences degassing upon ascent. In order to quantitatively assess this process, we present results from a holistic numerical model, which incorporates the effects of post-entrapment modification on H₂O and CO₂ within the melt inclusion. The numerical model incorporates time-dependent external boundary conditions (e.g., H₂O contents of the host melt), temperature, and pressure to reproduce conditions during magma ascent, cooling, and crystallization in the oceanic crust. In contrast to initially high-H₂O melt inclusions from subduction zones, we demonstrate that low-H₂O melt inclusions tend to preserve their pre-entrapment H₂O and CO₂ concentrations and provide reliable estimates for the depth of entrapment. The presence of a vapor bubble, however, may indicate that either diffusive H₂O loss or post-entrapment crystallization has occurred, in which case the included melt must be corrected for exsolved CO₂. Even small vapor bubbles can result in significant decreases in the CO₂ contents of the melt inclusion, yielding low entrapment pressure estimates. We compare model results with available data from melt inclusions from MORs and demonstrate the necessity of correcting for the presence of vapor bubbles to obtain correct estimates of pressures and volatile contents at the time of entrapment.