Significance of halogens (F, Cl) in kimberlite melts: Insights from mineralogy and melt inclusions in the Roger pipe (Ekati, Canada)
Abersteiner, AB and Kamenetsky, VS and Kamenetsky, MB and Goemann, K and Ehrig, K and Rodemann, T, Significance of halogens (F, Cl) in kimberlite melts: Insights from mineralogy and melt inclusions in the Roger pipe (Ekati, Canada), Chemical Geology: An International Journal ISSN 0009-2541 (In Press) [Refereed Article]
The abundance and distribution of halogens (F, Cl) are rarely recorded in kimberlites and therefore their petrogenetic significance is poorly constrained. Halogens are usually present in kimberlite rocks in the structure of phlogopite and apatite, but their original concentrations are never fully retained due to the effects of alteration. To provide new constraints on the origin and evolution of halogens in kimberlites and their melts, we present a detailed study of the petrography and geochemistry of the late-Cretaceous Group-I (or archetypal) Roger kimberlite (Ekati cluster, Canada). The studied samples contain abundant anhedral-to-euhedral olivine which is set in a crystalline groundmass of monticellite, phlogopite, apatite, spinel (i.e. magnesian ulvöspinel-magnetite (MUM), Mg-magnetite, pleonaste, Cr-spinel), and perovskite along with abundant secondary alteration phases (i.e. serpentine, garnet (andradite-schlorlomite), amakinite ((Fe2 +, Mg, Mn)(OH)2), calcite). The Roger kimberlite is characterised by the highest recorded F-content (up to 2688 ppm) of the Ekati cluster kimberlites, which is reflected by the preservation of F-rich phases, where bultfonteinite (Ca4(Si2O7)(F, OH)2) and fluorite commonly replace olivine.
In order to examine the composition and evolution of the kimberlite melt prior to post-magmatic processes, we studied melt inclusions in olivine, Cr-spinel, monticellite and apatite. Primary multiphase melt inclusions in Cr-spinel, monticellite and apatite and secondary inclusions in olivine are shown to contain a diversity of daughter phases and compositions that are dominated by alkali/alkali-earth (Na, K, Ba, Sr)-enriched Ca-Mg-carbonates ± F, Na-K-chlorides and sulphates, phosphates ± REE, spinel, silicates (e.g. olivine, phlogopite, (clino)humite), and sulphides. Although alkali/alkali-earth- and halogen-bearing phases are abundant in melt inclusions, they are generally absent from the kimberlite groundmass, most likely due to ubiquitous effects of syn- and/or post-magmatic alteration (i.e. serpentinisation).
Comparisons between halogens and other trace elements of similar compatibility (i.e. F/Nd and Cl/U) in the Roger kimberlite and their respective estimated primitive mantle abundances show that halogens should be a more significant component in kimberlites than typically measured. We propose that fluorine in the Roger kimberlite was magmatic and was redistributed during hydrothermal alteration by Ca-bearing serpentinising fluids to produce the observed bultfonteinite/fluorite assemblages. Based the compositions and daughter mineral assemblages in primary melt inclusions and reconstructed halogen abundances, we suggest that Cr-spinel, monticellite and apatite crystallised from a variably differentiated Si-P-Cl-F-bearing carbonate melt that was enriched in alkalis/alkali-earths and highly incompatible trace elements.