Experimental Constraints on the Possible Mantle Origin of Natrocarbonatite

Abstract

The phase relations of a “primary” sodic dolomitic carbonatite (CM1), determined by Wallace and Green (1988) to be in equilibrium with an amphibole lherzolite assemblage, have been investigated from 5 to 27 kb over a temperature range of 650–1225 °C with 2 and 4wt% added water in order to test the model, first proposed by Wallace and Green (1988), that natrocarbonatite may result from the closed system crystal fractionation of such a “primary” composition. Our experiments suggest that if Oldoinyo Lengai magmas evolve from a primary carbona-titic liquid at depth by crystallization of dolomite, olivine and spinel, then this liquid must be more silicic, less FeO-rich and have a greater K/Na ratio than the Wallace and Green (1988) “primary” carbonatite composition. These suggested differences may be consistent with differences in the mantle source region from which the Oldoinyo Lengai composition was derived and the peridotite composition (Hawaiian Pyrolite) with which the Wallace and Green (1988) composition was equilibrated.

The phase relationships of CM1 also demonstrate the existence of a decarbonation reaction to lower pressures and higher temperatures which results in the formation of a CO₂-rich fluid phase. This fluid phase contains alkali elements and chlorine and may be responsible for the fenitization of wall-rocks associated with many carbonatite complexes. Additionally, a late-stage fractionate would be extremely sodic (approximately 40wt% Na₂O) and may also fenitize wall-rocks.

A twofold increase in the added water content (from 2 to 4wt%) depressed the liquidus temperature by <30°C and had little effect on the observed liquidus phase relationships. No evidence for liquid immiscibility was observed in the P-T region studied.