Previous page     Contents     Next page

Calc alkaline magmas, which form at most of the currently-active volcanoes, originate from partial melting of the mantle. This source mantle is geochemically modified through interaction with fluids or melts produced during the metamorphic evolution of subducted oceanic lithosphere (i.e. oceanic crust which is buried into the mantle). The geochemical cycle of potassium and related trace elements such as Rb, Be, and B has been investigated experimentally in order to define which minerals transport potassium to large depths and which conditions are favorable for the generation of potassium-rich fluids or melts.

The most important potassium-bearing mineral in crustal lithologies at high pressures is phengite (a white mica). It forms in the shallow portion of the subducted slab at blueschist conditions (10 - 50 km depth) and remains stable to 300 km depth (at 800 - 1000 ^oC). In this depth interval phengite continuously changes composition towards the Si-rich celadonite-endmember of the muscovite-celadonite solid solution. At the same time, omphacitic clinopyroxene increasingly incorporates a KAlSi₂O₆ component - for example, at 300 km depth, 1 wt% K₂O is soluble in clinopyroxene. As a result of the reaction which produces K₂O-bearing clinopyroxene, fluid is continuously produced and released to the mantle.

Mass balance calculations reveal that in basalts potassium initially incorporated in phengite can be completely disolved in omphacitic clinopyroxene. Thus, at the phengite destabilization pressure, a K-rich fluid pulse originating from basaltic rock compositions is unlikely to occur. Instead, in metasediments (greywackes and pelites) with much higher potassium contents and lower abundances of clinopyroxene and in which 10 - 20 % phengite are present the destabilization of phengite results in the release of a K-rich fluid with K-hollandite (KAlSi₃O₈) remaining in the residue.

The experimental results also show that the presence of phengite lowers the solidus of metabasalts by 100 °C (at 150 km depth) to 250 ^oC (at 300 km depth). Thus, in relatively warm subduction zones, melting of phengite-bearing metabasalts will occur and melts rich in potassium as well as in the trace elements Rb, Be, and B (which are concentrated in phengite) will geochemically alterate the overlying mantle. It is concluded that a distinct depth of dehydration through a single reaction does not exist and that the depth of complex dehydration/melting processes depends on the thermal structure of subduction zones.