Melting experiments on mantle vein assemblages have been undertaken in order to provide major and trace element compositions of melts from non-peridotitic mantle assemblages. The specific goal of these experiments is to determine the melting temperatures and details of the melting mechanisms of likely vein assemblages so that later reaction between these melts and their surrounding peridotitic host rocks can be quantitatively assessed.
The starting compositions consisted of mixtures of a third each of clinopyroxene (DI), phlogopite (PHL) and
K-richterite (KR) separated from natural ultramafic rocks, with the addition of 5% of an accessory phase ilmenite (ILM), rutile (RU) or apatite (AP) in some cases. These mixtures were chosen because of the commonness of DI and PHL in near-liquidus experiments on mantle-derived alkaline volcanics Foley (1992) and because of the known stability of KR under high-pressure conditions (Foley, 1991) and in K-rich melt compositions (Mitchell and Bergman, 1991).
Two series of experiments were performed: the first consisted of melting the mineral mixtures in sealed Pt capsules with internal graphite capsules at pressures of 15 and 50 kbar in order to determine the exact melting temperature. The second series used similar starting compositions, but included a melt extraction trap consisting of glassy carbon, which maintains its pore space under high-pressure conditions and serves to extract the melt into pools large enough for later trace element analysis by Laser Ablation Microprobe (LAM).
KR melts completely in all experiments within 50°C of the solidus, and all three accessory phases investigated also melt completely over about the same temperature range. This leads to major element compositions of melts from the basic DI+PHL+KR assemblage at 15 kbar which are rich in SiO2 (>58wt%) and K2O (>11%) but poor in Al2O3 (5.8%) and CaO (3.9%). This chemistry reflects the high content of potassic phases, and the amplification of the K-content of melts by incongruent melting reactions for KR (and, to a lesser extent, PHL). Melts from accessory mineral-bearing assemblages are additionally enriched in those elements which are major constituents only in the accessory phases (Ti for ILM and RU, and P for AP).
Trace element analyses of experimental glasses by the Newfoundland LAM, a Laser-ICP-MS technique specifically optimized for the analysis of minute experimental charges (Jackson et al., 1992; Jenner et al., 1993), imply that the presence of accessory phases rich in incompatible
elements in the vein assemblage is essential if lamproites are to be explained by melting of veined mantle. Melts from the DI+PHL+KR assemblage show enrichments of more than 100 relative to primitive mantle only for Cs, Rb and Ba. The other elements show enrichments mostly between 1 and 10 which are more reminiscent of ocean island basalts than of strongly enriched lamproites. A lamproite-like trace element pattern for either olivine lamproites or leucite lamproites can only be produced from the vein assemblage alone if both apatite and a titanate mineral are present in addition to DI+PHL+KR. Amongst the 22 incompatible trace elements measured, this five-phase vein assemblage would leave a discrepancy for only Ba, Zr and Hf, which are more enriched in natural lamproites. The discrepancy for Zr and Hf may indicate the presence of additional accessory phases or be due to the effect of major element composition on PHL/Lq partitioning, which is known to have a strong effect on the partitioning of barium (Schmidt et al., 1996).
Foley, S.F., Lithos 28, 187-204 (1992).
Foley, S.F., Geochim. Cosmochim. Acta 55, 2689-2694 (1991).
Jackson, S.E., Longerich, H.P., Dunning, G.R. & Fryer, B.J., Can. Mineral. 30, 1049-1064 (1992).
Jenner, G.A., Foley, S.F., Jackson,S.E., Green, T.H., Fryer, B.J. & Longerich, H.P., Geochim. Cosmochim. Acta 57, 5099-5103 (1993).
Mitchell, R.H. & Bergman, S.C., Petrology of Lamproites, Plenum Press, New York, 447pp. (1991).
Schmidt, K.H., Bottazzi, P., Vannucci, R., Mengel, K. & Foley, S.F., J. Conf. Abs. 1, (1996).