Abstract
Amphiboles are rather rare in the volcanics of the whole Etnean succession and commonly are represented by kaersutites to titanian pargasites, mostly found in differentiated products. Titanian Mg-hastingsites have been found in lavas and tephra from the 2001 eruption at Mt. Etna. New major (EMPA) and trace element (LAM-ICP/MS) data on amphiboles from this eruption have been compared with reference data for kaersutites from prehistoric eruptions. The two amphibole groups significantly differ from each other in their AlIV, AlVI, K and Mg# values, which are higher in Mg-hastingsite than in kaersutite. Ti and Na are lower in Mg-hastingsite than in kaersutite. REE and trace element patterns for all the analysed Mg-hastingsite crystals are quite homogeneous. Kaersutite patterns generally conform to those of Mg-hastingsite but display higher concentrations for most of the trace elements.
The exceptional occurrence, exclusively in tephra, of some crystals under equilibrium conditions with the coexisting residual glass has made it possible to calculate the partition coeffcients between amphibole and melt (Amph/meltD) for trace elements. A new set of partition coeffcients is then provided, deriving from analyses on five amphibole/melt pairs at equilibrium. These data highlight the effects of amphibole crystallization in controlling some trace element ratios (e.g. Th (or U)/Ta, Th/Nb, La/Nb) in residual melts of alkali basaltic systems, and suggest new hints for interpreting possible geochemical anomalies of these magmas.
In addition, the comparison between the calculated Amph/meltD of Mg-hastingsite and those from literature relative to kaersutite from prehistoric eruptions shows that they are generally lower in the former than in the latter one for most of the trace elements. All of the available data provide constraints on the physical growth conditions for the 2001 Mg-hastingsite. Temperatures around 980 °C and volatile pressures in the range of 200–300 MPa have been estimated by integrating geophysical and petrological data. The highest pressure values are however larger than the lithostatic pressure alone acting on the magma reservoir (~ 6 km b.s.l.), as defined on the grounds of the hypocentres depth of the seismic events associated to the magma rise. This implies that Mg-hastingsite was probably crystallizing in a closed reservoir under overpressure conditions. Finally, microchemical data and trace element partitioning suggest that the differences between Mg-hastingsites and kaersutites in the Etnean products are mainly due to the less differentiated character of the magmas emitted after the 1971, and particularly after the 2001 eruption, compared to the compositions that characterize products of the prehistoric events. Furthermore, also a higher pressure of the system where Mg-hastingsite crystallized would account for its compositional differences respect to prehistoric kaersutite.