SEARCHING FOR PARENTAL KIMBERLITE MELT

Abstract

Constraining the composition of primitive kimberlite magma is not trivial. This study reconstructs a kimberlite melt composition using vesicular, quenched kimberlite found at the contact of a thin hypabyssal dyke. We examined the 4 mm selvage of the dyke where the most elongate shapes of the smallest calcite laths suggest the strongest undercooling. The analyzed bulk compositions of several 0.09-1.1 mm2 areas of the kimberlite free from macrocrysts were considered to be representative of the melt. The bulk analyses conducted with a new "chemical point-counting" technique were supplemented by modal estimates, studies of mineral compositions, and FTIR analysis of olivine phenocrysts. The melt was estimated to contain 26-29.5 wt% SiO2, ~7 wt% of FeOT, 25.7-28.7 wt% MgO, 11.3-15 wt% CaO, 8.3-11.3 wt% CO2, and 7.6-9.4 wt% H2O. Like many other estimates of primitive kimberlite magma, the melt is too magnesian (Mg# = 0.87) to be in equilibrium with the mantle and thus cannot be primary. The observed dyke contact and the chemistry of the melt implies it is highly fluid (η = 101-103 Pa s at 1100-1000 °C) and depolymerized (NBO/T = 2.3-3.2), but entrains with 40-50% of olivine crystals increasing its viscosity. The olivine phenocrysts contain 190-350 ppm of water suggesting crystallization from a low SiO2 magma (aSiO2 below the olivine-orthopyroxene equilibrium) at 30-50 kb. Crystallization continued until the final emplacement at depths of few hundred meters which led to progressively more Ca- and CO2-rich residual liquids. The melt crystallised phlogopite (6-10%), monticellite (replaced by serpentine, ~10%), calcite rich in Sr, Mg and Fe (19-27%), serpentine (29-31%) and minor amounts of apatite, ulvöspinel-magnetite, picroilmenite and perovskite. The observed content of H2O can be fully dissolved in the primitive melt at pressures greater than 0.8-1.2 kbar, whereas the amount of primary CO2 in the kimberlite exceeds CO2 soluble in the primitive kimberlite melt. A mechanism for retaining CO2 in the melt may require a separate fluid phase accompanying kimberlite ascent and later dissolution in residual carbonatitic melt. Deep fragmentation of the melt as a result of volatile supersaturation is not inevitable if kimberlite magma has an opportunity to evolve. © 2007 Elsevier Ltd. All rights reserved.