Experimental insights into the origin of crystal-poor phonolitic magmas

Crystal-poor phonolitic magmas have been commonly erupted during the quaternary explosive volcanism of Central Italy. The origin of crystal-poor magmas represents a complex issue of igneous petrology and since the first studies on crystal fractionation by settling, many alternative mechanisms of crystal-melt separation have been proposed [see Bachmann and Bergantz, 2004 and references therein]: i) convective fractionation in a crystallizing boundary layer; ii) gas-driven filter press; iii) thermal gradient responsible for mass transport (thermal migration) resulting in segregation of melt from the mushy, boundary zone of magma chambers; iv) melt migration induced by crystal compaction; v) instability of “solidification front”. In the frame of the highly explosive volcanism of Central Italy, that was fed by differentiated and thermally zoned pre-eruptive systems, not all of the abovementioned mechanisms would operate efficiently. Moreover, these mechanisms are mainly based on theoretical models and natural evidences, but poorly constrained by experiments. In this work, in order to shed light on the origin of large volumes of highly differentiated, crystal-poor magmas in thermally zoned systems, we have experimentally investigated crystallization, differentiation, and crystal-melt separation in the presence of a thermal gradient. Melt differentiation has been investigated under isothermal conditions (i.e. phase equilibria experiments) as well. As a case study, we have used the Sabatini Volcanic District, one of the main volcanic districts of the Roman Province characterized by several explosive eruptions producing large volumes of crystal-poor phonolitic magma [Masotta et al. 2010]. Phase equilibria experiments constrained the liquid line of descent leading to phonolitic magmas but gave no insights into the origin of crystal-poor textures. On the contrary, thermal gradient experiments produced structures clearly showing crystal-rich textures overlaid by crystal-poor batches of differentiated melt. The structures pictured at the cool top of the charges remind the so-called solidification front [Marsh, 1996]. Thermal gradient experiments, being performed under controlled conditions, allowed us to determine parameters necessary to model the processes responsible for the formation of these batches of differentiated, crystal-poor melt. We advocate that these experiments represent an important tool to model crystal-melt separation in thermally zoned natural systems as well.