EXPERIMENTAL STUDY ON MELTING PHASE RELATIONS AND DIAMOND FORMATION IN THE CARBONATE-RICH KIMBERLITE FROM MAJUAGAA,
Butvina V.G.*, Nielsen T.F.D.**, Safonov O.G*., Litvin Yu.A.*
*Institute of Experimental Mineralogy,
Carbonate-rich hypabyssal kimberlites of the Majuagaa dyke, Maniitsoq,
In order to further estimate a possible mantle source for the Majuagaa kimberlites, melting relations of the texturally homogeneous sample are experimentally studied at pressure 8.5 GPa and temperatures 1200 1750oC using the anvil-with-hole HP apparatus and a thermogradient cell. The studied interval of temperatures includes conditions of subsolidus, subliquidus, and complete melting conditions for the kimberlite. Liquids forming both during partial and complete melting of the sample are homogeneous carbonate-silicate melts (LCS). The quenched carbonate-silicate melt only (i.e. complete melting) is observed at temperatures above 1730OC. Compositional characteristics of this melt and its MgO/CaO (~1.4) are very similar to the model Majuagaa kimberlite melt calculated from the rock bulk composition after subtraction of olivine composition . It means that this melt would coexist with olivine. However, olivine Fo90Fa10 was firstly identified only at 1730oC. Garnet and Mg-bearing ilmenite (?) appear at about 1710oC forming olivine + garnet + ilmenite + LCS cotectic assemblage. The melt at this temperature is more carbonatitic and shows lower MgO/CaO ratio (~0.8). Such evolution of the melt is consistent with that identified for kimberlite-related systems at high pressures [2-5]. Perovskite, apatite, and Ti-magnetite (?) subsequently join this assemblage at 1700-1680oC. Appearance of carbonate phase, presumably, dolomite, at 1680OC manifests the solidus of the system. At lowest temperature (down to 1200OC), the subsolidus includes the assemblage olivine + garnet + ilmenite (Ti-magnetite) + Mg-aragonite(?) + Ca-Fe-magnesite + perovskite + apatite. No pyroxenes were observed in the solidus run products. Their absence is, probably, related to low activity of CO2 in the experiments, stabilizing olivine with calcic carbonate (for example, 2olivine + 4CaCO3 + 2CO2 = 3dolomite + diopside).
Thus, the present experiments indicate that the Majuagaa carbonate-rich kimberlite has originated from the carbonated garnet peridotite enriched in titanium. Experiments indicate very narrow temperature interval for formation of the kimberlite melt, i.e. 50-70oC (~1760 - 1680oC). It is consistent with experiments by Dalton & Presnall  in the model CMAS-CO2 system, but strongly disagrees with the results of experiments on natural peridotite-carbonate systems [e.g. 3-5]. This narrow solidus-to-liquidus interval implies that the Majuagaa kimberlite, probably, is a near-eutectic (olivine + garnet + ilmenite + dolomite + LCS) melt separated from the above carbonated garnet peridotite at very low degree of partial melting. Addition of carbonate to the peridotite is, possibly, related to metasomatic processes via CO2-rich fluids or complex carbonatite melts. The presence of the second agents is supported by recent observations of chloride-bearing carbonatite inclusions in olivines from the Majuagaa kimberlites geochemically similar to the host kimberlites .
High efficiency of the Majuagaa carbonate-silicate melts for diamond spontaneous nucleation and growth in the case of oversaturation of this melt with dissolved carbon with respect to diamond is demonstrated with additional experiments at 6.5 8.5 GPa and 1650 1780oC with the homogeneous mixture (3:2) of the kimberlite and spectral graphite. Formation of diamond crystals of octahedral habit were observed within the 5 12 min period.
The study is supported by the INTAS project 05-1000008-793,8 RFBR grants 08-05-00110-a and 07-05-00499-a, the RF President grants MK-194.2008.5 and MD-130.2008.5.
 Nielsen T.F.D., Jensen S.M.
The Majuagaa calcite-kimberlite
dyke, Maniitsoq, Southern West Greenland //
Geological Survey of Denmark and
 Girnis A.V., Bulatov V.K., Brey, G.P. Transition from kimberlite to carbonatite melt under mantle parameters: an experimental study // Petrology. 2005. V. 13. P. 1-15.
 Brey G.P., Bulatov V.K., Girnis A.V., Lahaye Y. Experimental melting of carbonated peridotite at 6 - 10 GPa. Journal of Petrology. 2008. V. 49. P. 797-821.
 Kamenetsky V.S., Kamenetsky
M.B., Weiss Y., Navon O., Nielsen T.F.D., Mernagh T.P. Alkali carbonates and chlorine in kimberlites from