Origin of Picritic Green Glass Magmas by Polybaric Fractional Fusion

Calculations show that it is possible to generate picritic mare green glass compositions by inefficient polybaric fractional fusion of a low-Al2O3 source region. In these calculations it is assumed that a rising source region undergoes small amounts of pressure-release melting (1-1.2%/kbar) and that some of this melt escapes from the source into a reservoir that remains in close proximity to die source. At the end of the calculations, the pooled melt in the reservoir is a composite of small degrees of melt drawn off over a range of pressures from a progressively depleting source. The apparent pressure of melt segregation of the green glasses (greater-than-or-equal-to 20 kbar), derived from high-pressure melting experiments, is lower than the initial pressure of melting (approximately 40 kbar) and higher than the final pressure of pooling (4-5 kbar). Polybaric fractional fusion thus overcomes some of the troublesome aspects of simpler static models, such as the need for high degrees of partial melting at lower pressures or modest degrees of partial melting at higher pressures followed by transit of the melt to the surface without fractionation. Varying degrees of efficiency of fractional fusion allow for a wide range of source compositions; however, the calculations show that in order for adiabatic decompression to supply sufficient heat to balance the heat of fusion, the melt extraction efficiency can be no more than approximately 0.4. In order to produce pooled melts in the Al2O3 concentration range of the green glasses (6-9 wt%) with inefficient fractional fusion plus small degrees of melting of the source, it is necessary to start with a very-low-Al2O3 source (less-than-or-equal-to 1.5 wt%). Such a source is consistent with cumulate models of the mare basalt source region, but is lower in Al2O3 content than some estimates of the average lunar mantle (e.g., Taylor, 1982). Small variation of the ilmenite content of the source can produce pooled melts with a wide range of TiO2 concentration; however, the high-Ti picritic glasses require a different or perhaps additional melting mechanism because they arc too dense to have formed at pressures greater than 20-25 kbar and then risen to the surface (Delano, 1990). Two physical models of the melt reservoir seem possible. One is an initially barren olivine layer above the source into which the melt percolates. Calculations of the gravitational stability of melt in a porous layer suggest that the thickness of layer must be on the order of hundreds of meters to tens of kilometers in order for convective mixing (necessary for the success of the polybaric model) to take place. The other is a simple melt layer or bubble atop the rising source. Calculations of ascent velocities indicate that for a substantial range of density contrasts and partial melting the greater size of the source more than compensates for the lower density of the melt. Thus, melt may percolate out of the source, but free bodies of melt will not rise faster than the source, until the volume of the melt becomes a significant fraction of the source. The integrated degrees of melting required by the polybaric model (approximately 12%) fall well within the no-escape range for reasonable choices of density contrasts. Although the high pressures of initial melting imply differentiated material at great depths (>1000 km) within the Moon, a shallower magma ocean is possible if radioactive heating of the layer beneath the magma ocean eventually leads to convective overturn.