Kiglapait mineralogy III: Olivine compositions and Rayleigh fractionation models

Abstract

Olivines, present throughout the layered cumulates of the Kiglapait Intrusion, record extreme iron enrichment during fractional crystallization. Mn is less compatible than Fe; the exchange coefficient KD for Mn/Fe (OL-LIQ) has values near 0·95 but they drop to 0·7 near the Lower Zone-Upper Zone boundary. Ca is depleted by more than a factor of 30 relative to experimental values at 5 kbar pressure. Attempts to find a sink for Ca in the Lower Zone without invoking liquid as a reactant are unconvincing, and questions of metastability arise. Cumulus olivines vary systematically to lower Fo contents with stratigraphic height. Using equilibrium values for the exchange coefficient KD (Fe/Mg, OL-LIQ), and the observed mode of olivine and residual porosity, a Rayleigh fractionation calculation reproduces the observed Lower Zone trend without recourse to multiple injections of magma. An anomalously Fe-rich region above the Main Ore Band in the Upper Zone is ascribed to ponding of Fe-rich magma at the floor of the intrusion until later, buoyant residual magma entrained the Fe-rich residua. Such Fe-rich ponds probably account for the abnormal Fe enrichment shown by some other Fenner-trend layered intrusions. Summation over the observed crystal compositions yields values of XMg that do not represent the equilibrium liquid, but that reduce to it by a simple multiplier p = 0·78. This result means that if KD is known, the high-temperature, liquidus olivine compositions can be retrieved, or vice versa. Because the observed olivine compositions are supported by realistic forward models, there is no need to invoke evolved boundary-layer liquids to explain the in situ crystallization of Kiglapait cumulates. Classical cumulate theory suffices.