Apatite compositions and groundmass mineralogy record divergent melt/fluid evolution trajectories in coherent kimberlites caused by differing emplacement mechanisms

Abstract

Kimberlites are pipe-like igneous bodies, consisting of a pyroclastic crater and diatreme, commonly underlain by coherent root-zone rocks, and with associated dyke/sill complexes. The processes that control the different modes of coherent kimberlite emplacement remain uncertain. In addition, late evolution of kimberlite melts during emplacement into the upper crust remains poorly constrained. Therefore, it is unclear whether there is a link between melt composition/evolution and the emplacement mechanism of coherent kimberlites (i.e. planar dykes/sills vs. irregular bodies in the root zone). An absence of comparative studies on late-stage magmatic phases across the different emplacement modes of coherent kimberlite from the same locality hamper resolution of these issues. Therefore, we report petrographic and mineral chemical data for groundmass apatite in samples of dyke, sill, and root-zone kimberlites from the Kimberley cluster (South Africa). Early crystallised phases (olivine, spinel, Mg-ilmenite) in dyke/sill and root-zone kimberlites have indistinguishable compositions, and hence crystallised from similar primitive melts. Conversely, apatite compositions are generally distinct in dyke/sill (low Sr, high and variable Si) and root-zone kimberlites (high and variable Sr, low Si). The Si enrichment of apatite in dykes/sills is attributed to the coupled incorporation of CO32− and SiO44− for PO43−, reflecting higher CO2 contents in their parental melts, and potentially higher Si contents due to the preferential crystallisation of carbonates over mica/monticellite. The low Sr contents of apatite in dyke/sill kimberlites reflect equilibrium with a (kimberlite) melt (i.e. DSr is close to unity for carbonate and silicate melts), whereas the higher Sr contents of apatite in root-zone kimberlites require crystallisation from, or overprinting by a H2O ± CO2 fluid (significantly higher DSr). The relative enrichment of CO2 in kimberlite dykes/sills is evident from the abundance of carbonates, the presence of mesostasis dolomite and calcite phenocrysts in some samples, and concomitant reduced proportions of other groundmass phases (e.g. serpentine, mica, monticellite). During late alteration of kimberlite dykes/sills, monticellite is typically replaced by carbonates, whereas olivine and pleonaste are relatively stable, indicating the melts which form dykes/sills evolve to higher CO2/H2O ratios. It is unlikely that these two distinct evolutionary paths were caused by crustal contamination before or during near surface magma emplacement, because crustal assimilation is not recorded in the O and Sr isotopic composition of late crystallising olivine rinds or carbonates, respectively. We suggest that higher concentrations of CO2 are retained in kimberlite dykes/sills due to higher confining pressures (i.e. lack of breakthrough to the surface). In contrast, exsolution of CO2 from root-zone kimberlites increased melt H2O/CO2 ratios and promoted the crystallisation of mica and monticellite at the expense of dolomite and calcite. Apatite compositions have the potential to aid in the discrimination of kimberlites from lamproites (higher LREE, Sr, F, and S, lower Si contents) and carbonatites (higher LREE, F, Cl and S, lower Fe contents). However, the compositions of kimberlitic apatite overlap those from aillikites, probably due to similar late-stage melt compositions.