Journal of Petrology

The Highly Compatible Trace Element Paradox - Fractional Crystallisation Revisited.

M.J. O'Hara, N. Fry

Abstract

Field relations in dissected volcanic terrains and the internal evidence of persistent low-pressure cotectic character in erupted basalts point to the frequent and substantial modification of liquid compositions by some form of partial crystallisation within the crust. In contrast, the highly compatible trace elements do not display the marked variations and extreme depletions which are predicted to result from perfect fractional crystallisation (PFC). Imperfect fractional crystallisation, refilling of magma chambers during fractionation and in situ crystallisation are important factors which can help to explain this apparent paradox. This paper explores another effect, the integration of residual liquids from differing extents of partial crystallisation, which can help to resolve this paradox, even while still permitting perfect fractional crystallisation at all points in the magma chamber. Integration of such residual liquids through the thickness of the crystallisation zone is explicit, although not implemented, in the model of in situ crystallisation proposed by Langmuir (1989). It may be separated as a process for purposes of mathematical modelling from the basic concept of partial crystallisation of small packets of magma with remixing of the residual liquids into the main body of magma. Integration of melts from differing extents of partial crystallisation might in principle also be applied to the case of lateral variations in the mass fraction crystallised with position in the magma chamber.

Integrated PFC itself can develop residual liquids which differ little from products of equilibrium (batch) crystallisation (EPC) at the same average mass fraction of liquid remaining in both incompatible and compatible trace element concentrations. For one specific combination of parameters these integrated liquids are identical in composition at all values of the distribution coefficient to the EPC liquid. At other values of the parameters the integrated liquids may even - a new paradox - have higher relative concentrations of highly compatible elements than the EPC products. Any integration of residual liquids from different mass fractions of PFC rapidly eliminates what have in the past been taken to be the diagnostic differences between PFC and EPC. Integration of EPC liquids (towards which the products of imperfect fractional crystallisation processes will tend) produces even more pronounced effects, with highly compatible elements less depleted even than in EPC and far less depleted than would be predicted by simple models. When interpreted according to simplistic models, sequences of residual liquids produced in such processes might appear to be inconsistent with products of a partial crystallisation process and to require a process of progressively smaller mass fractions of melting of inhomogeneous and progressively more refractory (higher Mg) source regions.

Keywords

highly compatible elements;in situ crystallisation;boundary layer;integrated crystallisation

Return to header

Current Contents ................

J. Petrology Editor's Page