A new geological environment for rodingite formation: Metasomatism in kimberlite volcanoes

Rodingite is a calc-silicate rock containing garnet, diopside, and chlorite that develop via metasomatic replacement of ultramafic to felsic rocks. Rodingites are found in ophiolites subjected to low-grade metamorphism or hydrothermal serpentinization of the surrounding ultramafic rocks. Here we report a previously unrecognized volcanic environment for rodingite formation. Rodingites can result from skarn-like reactions between kimberlite and xenoliths of silicate country rocks.

We studied hypabyssal and pyroclastic kimberlites from Renard and Gahcho Kue clusters (Canada) and Orapa (RSA). The kimberlites entrain 20-90 vol.% of granitoid and gneiss (Renard 65, Gahcho Kue) and basalt (Orapa) xenoliths collectively called silicate country rocks. Skarn-like reactions triggered by gradients in the chemical potentials of Si, Al, Ca, and Mg across the xenolith–kimberlite contacts produce concentric reaction zones within the xenoliths and a reaction halo in the surrounding contaminated kimberlite. The original mineralogy of the unreacted xenoliths is replaced by prehnite, diopside, pectolite, wollastonite, serpentine, garnet, calcic hydrosilicates (hydrogarnet, xonotlite, amphiboles). In the kimberlite halo, diopside and phlogopite form, carbonate is leached out, olivine is completely serpentinized. Rodingites develop by moderate degrees of reaction, and are replaced with monomineral serpentine and chlorite if metasomatism advances further.

Petrographic evidence for post-emplacement, metasomatic development of rodingite assemblages in kimberlite matches the Perple_X phase equilibria calculations that model the formation of the reactive mineralogy in the subsolidus, at T<600^oC. Mass transfer of major elements is demonstrated by bulk compositions, thermodynamic modelling, and conserved element ratio analysis. The xenoliths experience Si loss and gain of Ca and Mg; the opposite trends are observed in the adjacent kimberlite. A metasomatic rather than hydrothermal origin of the reactive mineralogy is suggested by the inability to model hydrogarnet and pectolite under fluid saturated conditions. The observed mineralogy is accurately reproduced only when H₂O and CO₂ are modelled as components controlled by the kimberlite composition.

Rodingite patches in kimberlites have many common characteristics with rodingites in ophiolites and serpentinites, i.e. 1) a wide range of protoliths; 2) a juxtaposition of different lithologies; 3) synchronicity with the proximal serpentinization; 4) local sources of elemental mass transfer; 5) a combination of desilication and Ca gain; 6) a low-T character; 7) a common replacement of garnet by hydrous sheet silicates in subsequent de-rodingization.

Rodingite formation in kimberlites is distinct from the classic rodingization in several aspects. In kimberlites, rocks of contrasting compositions are juxtaposed on the smaller scale creating centimeter-sized domains of rodingite mineralogy. An influx of H₂O is not required, as the deuteric hydrous fluid is already in excess in the cooling kimberlite. In this environment, serpentinization is driven by the Si flux. Rodingites in kimberlites form at surface pressure. De-rodingization is not a result of gradually decreasing P-Ts, but is due to the continued mass transfer and cooling. Our observations suggest that hydration, serpentinization of clinopyroxene, decarbonation of marine sediments and moderate P= 3-6 kb are not pre-requisites for rodingite formation. A new geological environment of rodingite formation in kimberlites expands the range of parameters for rodingization.