2011

News	Registration	Abstract submission	Deadlines	Excursions	Accommodation	Organizing committee
First circular	Second circular	Abstracts	Seminar History	Program	Travel	Contact us

Experimental study of near-liquidus phase relations in kimberlites and estimation of CO₂ and H₂O contents in primary magmas

Bulatov V.K.*, Girnis A.V.**, Brey G.P.***

*Vernadsky Institute of Geochemistry and Analytical Chemistry, Moscow, Russia; **Institute of Geology of Ore deposits, Petrography, Mineralogy, and Geochemistry, Moscow, Russia; ***J.-W. Goethe University, Frankfurt, Germany

bulatov@bk.ru

Experimental melting of mantle rocks has demonstrated that kimberlite-like melts can be derived by low degrees of melting of carbonated peridotites at 5–10 GPa. However, a detailed comparison of experimental and natural compositions showed that primitive kimberlites are significantly different from partial melts (Brey et al, 2008). The compositions of the latter strongly depend on the conditions of melting and volatile contents in the source. The estimation of the composition of primary kimberlite magma is a difficult problem owing to the considerable amounts of xenogenous materials and strong secondary alteration. Nevertheless, the estimates of primary magmas for group I (low-K) kimberlites on the basis of comprehensive petrographic, mineralogical, and geochemical investigations (e.g. Kopylova et al., 2007; Becker, LeRoex, 2006) converge to a narrow compositional range with ca. 30 wt % MgO and 30 wt % CO₂ (calculated on a volatile-free basis). The contents of volatile components (CO₂ and H₂O) are most uncertain, because they are always incorporated in secondary minerals (primarily, calcite and serpentine). Therefore, previous experimental investigations of kimberlites were carried out at different proportions of volatiles. Experimental models of kimberlite genesis assumed different contents of volatiles in primitive kimberlite magmas ranging from >30 wt % CO₂ and 0% H₂O (Brey et al., 1991) to 0% CO₂ and >10 wt % H₂O (Kawamoto, Holloway, 1998).

We assumed that the primary kimberlite melt is derived by the partial melting of peridotite material enriched in volatile and incompatible elements. Then, the parameters of its generation must correspond to the conditions of equilibrium with the residual mantle assemblage. The composition of model primary melt was chosen on the basis of the ratios of nonvolatile components estimated from natural observations. Our experimental study aimed at determining the P–T conditions and CO₂ and H₂O contents at which kimberlite melt can be saturated simultaneously in olivine and low-Ca pyroxene (orthorhombic at <8 GPa and monoclinic at higher pressures) within the pressure range 4–10 GPa. Four synthetic mixtures were used in our experiments with identical proportions of major silicate components (wt %): ~32 SiO₂, 3 Al₂O₃, 10 FeO, 16 CaO, 0.5 Na₂O, 1 K₂O, and varying CO₂ and H₂O contents: 0 H₂O + 33 CO₂, 0 H₂O + 18 CO₂, 10 H₂O + 10 CO₂, and 12 H₂O + 5 CO₂. The starting mixtures were loaded into Pt capsules lined with Re foil for the prevention of Fe loss. The experiments were carried out using a multianvil apparatus described in detail by Brey et al. (2008). The compositions of quenched melts and crystalline phases were determined using a Jeol Superprobe 8900 electron microprobe. The contents of CO₂ and H₂O in melts were estimated on the basis of the analyses of phases and mass balance equations.

Experiments were carried out near the kimberlite liquidus at 6–12 GPa and 1200–1800°C. Experimental run duration varied from 8 to 48 h depending on temperature. At high temperatures such durations were sufficient for the formation of homogeneous (equilibrium) minerals and melts and were short enough to eliminate significant hydrogen loss and sample oxidation.

The experimental results indicate strong dependence of near-liquidus phase relations on the contents of fluid components. At high pressures CO₂ solubility in kimberlite melt is very high. Saturation with respect to fluid was detected only in experiments with the mixture containing 32 wt % CO₂ at temperatures of about 1700°C. The liquidus crystalline phase of this composition at pressures higher than 8 GPa is coesite, which is joined by garnet, low-Ca pyroxene, and magnesite at decreasing temperature. Coesite was never found in experiments with other mixtures containing smaller amounts of CO₂ at pressures up to 12 GPa, and the main near-liquidus phases in these compositions are low-Ca pyroxene and garnet. The addition of water decreases the liquidus temperatures, although they remain high even in water-rich compositions. In addition, the field of pyroxene crystallization is reduced in hydrous systems, and garnet becomes the main liquidus phase. At 6 GPa, olivine appears on the liquidus of the mixture with 12 wt % H₂O and 5 wt % CO₂. At higher pressure, it is changed by garnet.

The obtained experimental results led us to the following conclusions important for the estimation of the conditions and mechanisms of kimberlite magma generation. (1) Despite the relatively high CaO contents in the starting compositions, high-Ca pyroxene was observed only in the lowest temperature experiments in equilibrium with strongly evolved melts. This implies that kimberlite melts containing ~15 wt % CaO can be derived only from clinopyroxene-free (harzburgitic) materials. (2) An increase in CO₂ content (activity) results in a significant increase in SiO₂ activity owing to the occurrence of carbonation reactions in the melt, e.g. MgSiO₃ + CO₂ = MgCO₃ + SiO₂. This results in a change in liquidus phases at constant proportions of major components from coesite at the highest CO₂ content, through low-Ca pyroxene, and to olivine in the CO₂ poorest system. The composition considered here can be simultaneously saturated with respect to olivine and low-Ca pyroxene at relatively moderate contents of volatile components in melt (15 wt % CO₂ + H₂O) and various CO₂/H₂O proportions. Higher contents of volatiles are possible, if the primary magmas were poorer in SiO₂ than the composition studied by us. In such a case, the CO₂ content in the primary melt is limited by the conditions of magnesite appearance and cannot be much higher than 20 wt %. The suggestion on a relatively low SiO₂ content (composition transitional between kimberlites and carbonatites) in primitive kimberlite magmas is reasonable, because an increase in SiO₂ in the magma can result from interactions with peridotite wallrocks during magma ascent. Indeed, if primary magma was in equilibrium with olivine and low-Ca pyroxene in the source region, the melt will be supersaturated in olivine and undersaturated in pyroxene at decreasing pressure. Hence, interaction with wallrock peridotites will lead to pyroxene dissolution and olivine precipitation, i.e. an increase in SiO₂ content in the melt without significant modifications of the proportions of other components. (3) In all cases, the crystallization field of garnet expands significantly at pressures higher than 8 GPa, and it usually becomes the liquidus phase. The crystallization of other silicates begins only after considerable removal of alumina. This indicates that the melting of carbonated garnet peridotite at a pressure of 10 GPa and higher produces melts significantly depleted in Al compared with the estimated compositions of primary kimberlite magmas. The same conclusion was derived on the basis of peridotite partial melting experiments (Brey et al., 2008). (4) Even at high H₂O contents (>10 wt %), the liquidus temperatures of kimberlite melts are very high (>1500°C at 8 GPa) and exceed model cratonic geotherms for surface heat flows of 40–44 mW/m² (Pollack, Chapman, 1977). This implies that the formation of kimberlite magmas require a thermal event, whereas only carbonate-rich melts with <10 wt % SiO₂ can be formed in a thermally undisturbed cratonic lithosphere.

This study was supported by grants of the Russian Foundation for Basic Research and the Presidium of the Russian Academy of Sciences.

References:

Becker M., Le Roex A.P. Geochemistry of South African on- and off-craton, group I and group II kimberlites: Petrogenesis and source region evolution // Journal of Petrology, 2006. Vol. 47, P. 673–703.

Brey G.P., Kogarko L.N., Ryabchikov I.D. Carbon dioxide in kimberlitic melts // N. Jahrb., Mineral. Monatsh. 1991. No. 4. P. 159-168.

Brey G.P., Bulatov V.K., Girnis A.V., Lahaye Y. Experimental melting of carbonated peridotite at 6-10 GPa // Journal of Petrology, 2008. Vol. 49. P. 797-821.

Kawamoto T., Holloway J.R. (1997) Melting Temperature and Partial Melt Chemistry of H₂O-Saturated Mantle Peridotite to 11 Gigapascals // Science, 1997. Vol. 276. P. 240-243.

Kopylova M.G., Matveev S., Raudsepp M. Searching for parental kimberlite melt // Geochim. Cosmochim. Acta, 2007. Vol. 71. P. 3616-3629.

Pollack H.N., Chapman D.S. On the regional variation of heat flow, geotherms, and lithospheric thickness // Tectonophysics, 1977. Vol. 38. P. 279–296.