Phlogopite-rich and phlogopite-poor kimberlite intrusions within the Du Toitspan kimberlite pipe, South Africa: Petrogenetic relationships and localised source heterogeneity

Abstract

Samples from three petrographically distinct, intrusive kimberlite bodies and associated kimberlite dykes from the eastern lobe of the Du Toitspan kimberlite pipe, Kimberley, South Africa, have been analysed for their bulk rock major and trace element compositions and their olivine and phlogopite compositions. The two dominant intrusive bodies (D13, D14) are distinguished by the one (D13) being phlogopite-rich and best classified as a macrocrystic hypabyssal phlogopite kimberlite, and the other (D14) being phlogopite-poor and best classified as a macrocrystic hypabyssal monticellite kimberlite. The minor D17 intrusive body is classified as a macrocrystic transitional hypabyssal serpentinized phlogopite kimberlite. The associated kimberlite dykes range texturally from aphanitic to macrocrystic and are classified as calcite kimberlites.The major kimberlite intrusions and their associated dykes show no evidence of crustal contamination and are characterised by broadly overlapping geochemistry except for distinctly higher K2O (> 2 wt%) and Al2O3 (>3 wt%) and flattening HREE patterns (Gd/YbN = 6.5–7.0) in the D13 – phlogopite kimberlite compared to the D14 – monticellite kimberlite and the calcite kimberlite dykes (Gd/YbN = 9.6–12.1). These distinguishing geochemical features of the D13 – phlogopite kimberlite are comparable to typical Group II kimberlites in southern Africa. However, their diagnostic incompatible trace element ratios (for example, Th/Nb, La/Nb, Ce/Pb, and Ba/Nb) are instead comparable to other kimberlite intrusions analysed in this study and to southern African Group I kimberlites in general.Semi-quantitative modelling suggests that these kimberlite intrusions could have derived by low (<1%) degrees of partial melting of a source region that is enriched in LREE (Lan = ~6.1; Ybn ~ 1.47) comparable to metasomatised peridotites from the underlying lithospheric mantle. The composition of the D13 phlogopite kimberlite is consistent with a partial melt of a modally metasomatised source containing a higher proportion of residual clinopyroxene relative to garnet (compared to that giving rise to the D14 monticellite kimberlite and calcite kimberlite dykes), as well as accessory amounts of phlogopite, i.e. a garnet phlogopite peridotite (GPP). The absence of K-anomalies on primitive mantle normalized diagrams for the D13 phlogopite kimberlite requires that phlogopite was not a residual phase during partial melting and was exhausted shortly before or at the moment of melt segregation. The higher Gd/Yb ratios and lower K2O in the D14 monticellite kimberlite and calcite kimberlite dykes can be explained by partial melting of a cryptically metasomatized, phlogopite – free, garnet peridotite (GP) source, containing a higher proportion of garnet relative to clinopyroxene. The low absolute K and strong negative K-anomaly on primitive mantle normalized diagrams for the D14 monticellite kimberlite were inherited from a source region that previously experienced cryptic metasomatism by a differentiated fluid already carrying a negative K-anomaly.