RETHINKING THE MAGMA OCEAN

Since the early 1970s our understanding of lunar formation has been dominated by the idea of a magma ocean, which has been borrowed to explain terrestrial bodies from Mercury to Vesta. However, O'Hara [2000, J. Pet. 41, 1545] has attacked this model on a number of grounds. We concentrate here on three specific questions:

Is the lunar magma ocean consistent with the giant impact origin for the Moon? An impact-formed Moon depleted in siderophiles must have been well mixed and probably completely melted, not to just a depth of 400 km, during the initial impact and / or core formation. Geophysical constraints on the composition of the Moon are difficult to reconcile with the standard model.

Did a chilled anorthosite crust exist? Recent lunar orbiters indicate that the average crustal composition is significantly different from the limited Apollo samples on which this hypothesis was based. Lunar anorthosites may have been formed over too long a period to be consistent with a rapidly chilled crust. And magma ocean models of Vesta do not need such a crust for HED formation. Perhaps modifications indicated by HED models -- including more pervasive equilibrium during melting -- might be applied to the Moon.

Is the depletion of volatiles in both the Moon and the HED parent body primordial? O'Hara suggests it could have occurred during fire-fountain volcanism; large-scale shock reduction and dehydrogenation are also possible. We note the abundant vesicles seen in pristine lunar basalts and eucrites, and our new Mossbauer and micro-XANES evidence for ferric iron in lunar anorthosites.

Until we resolve these basic issues we will not be able to make progress in understanding the other terrestrial planets. If the magma ocean hypothesis survives these challenges, it will be a theory we can use with greater confidence; if it fails, everything we know about the origin of the terrestrial planets will need to be rethought.