Previous page     Contents     Next page

The mantle source region of potassic and ultrapotassic magmas (molar K₂O/Na₂O > 1) has been the subject of considerable controversy. Most models suggest that these magmas cannot be derived from melting of a homogeneous phlogopite- or amphibole-rich peridotite source. Rather, present models require a heterogeneous source region such as in the vein - wallrock melting model. In this model veins that are rich in phlogopite and/or pargasite melt and this melt subsequently reacts with the surrounding peridotitic wallrock.

Melting experiments on such a system at pressures of 1.5 and 2.5 GPa and temperatures between 950-1325°C have been carried out in the piston cylinder apparatus. The starting materials were extracted from a natural veined xenolith from the West Eifel volcanic field, Germany.

The melts produced at 1.5 GPa are all silica-undersaturated. At low degrees of partial melting (1050°C) the melts are relatively sodic since they result from the breakdown of pargasite to clinopyroxene + melt. Phlogopite melts completely over the temperature range 1200-1225°C resulting in a melt that is much more potassic. Reaction of the vein melt with the wallrock component, particularly clinopyroxene and orthopyroxene over the entire melting range results in its enrichment in Ca, Mg and to a lesser extent Si; however the degree of enrichment is dependent on the time available for reaction.

Experiments at 2.5 GPa, where amphibole breaks down to garnet, give a very different melt composition. In this case the initial melt at 1100°C is Si- saturated and is in equilibrium with pyrope garnet. At 1150°C pyrope is no longer present as a residual phase and appears to have dissolved in the evolving melt, which at this stage is Si-undersaturated. Continued melting results in breakdown of phlogopite and formation of a more potassic melt. The main effect of wallrock dissolution is to enrich the melt in Ca and Mg and to a lesser extent in Na. In general, the melts produced at 2.5 GPa are more Si-rich than those formed at 1.5 GPa. However, interpretation of these data is complicated by the effects of quench crystallisation; further experiments are underway to eliminate this effect.

The final composition of the melt in the vein wallrock system depends strongly on the starting composition of the vein and on the wallrock composition. Initial results using different wallrock compositions suggest that orthopyroxene exerts a strong control on the final melt composition since it is unstable in Si-undersaturated melts. Clinopyroxene does not dissolve completely but rather undergoes solid solution melting leaving an Al- and Na-poor and Ca-rich residual clinopyroxene. Dissolution of garnet may play a major role in determining the trace element composition of melts. If melt extraction is rapid and occurs immediately after amphibole melting the resulting magma will show a trace element signature consistent with residual garnet in the source. If, however, melt extraction is slow and does not take place until after phlogopite breakdown (when garnet has dissolved), the trace element signature of the resulting magma may not reflect its generation in the garnet stability field even though garnet was a residual phase early in the melting history.

The results of this preliminary study suggest that melting of a heterogenous mantle assemblage of hydrous veins and anhydrous wallrock can result in formation of a variety of potassic and ultrapotassic melts that resemble those found as volcanic rocks.