New Identity of the Kimberlite Melt: Constraints from Unaltered Diamondiferous Udachnaya – East Pipe Kimberlite, Russia - Advances in Data, Methods, Models and Their Applications in Geoscience - page 200

Krafft, 1990; Keller & Spettel, 1995) and experimental evidence that alkali carbonatite
magmas “will persist only if the magma is dry” (Cooper et al., 1975) we conclude that the
parental magma of the studied kimberlite was essentially anhydrous and carbonate-rich.
This is indirectly supported by the spectroscopic study of micro-inclusions in Udachnaya
cubic diamonds that showed that their parental media was a H 2 O-poor carbonatitic melt
(Zedgenizov et al., 2004).
Chlorine and H 2 O show opposing solubilities in fluid-saturated silicate melts, as they
apparently compete for similar structural positions in the melt. Although Cl does not form
complexes with Si in a melt, it may complex with network modifier cations, especially the
alkalies, Ca and Mg (Carroll & Webster, 1994). General “dryness” of carbonatites and
enrichment of natrocarbonatites in halogens (Gittins, 1989; Jago & Gittins, 1991; Keller &
Krafft, 1990) suggest that Cl and H 2 O decouple which can be an intrinsic feature of
carbonate-rich kimberlite magmas. If this is the case, the conventional role of H2O in
governing low temperatures and low viscosities of kimberlite magmas can be readdressed
to Cl. Furthermore, the data on carbonate-chloride compositions of melt inclusions in
diamonds (Bulanova et al., 1998; Izraeli et al., 2001; Izraeli et al., 2004; Klein-BenDavid et al.,
2004), nucleation and growth of diamonds in alkaline carbonate melts (Pal'yanov et al., 2002)
and catalytic effect of Cl on the growth of diamonds in the system C-K 2 CO 3 -KCl
(Tomlinson et al., 2004) concur with the proposed mantle origin of chloride and alkali
carbonate components in the Udachnaya-East kimberlite.
7.3 Liquid immiscibility and crystallisation of residual kimberlite magma
Liquid immiscibility is observed in the olivine-hosted melt inclusions at ~600 o C on cooling
(Fig. 8d). The immiscible liquids are recognized as the carbonate and chloride on the basis
that these minerals are dominantly present in the unheated melt inclusions (Golovin et al.,
2003; Kamenetsky et al., 2004). Remarkable textures, observed in melt inclusions at the exact
moment of melt unmixing (Fig. 8d), is governed by the carbonate crystallographic
properties. The presence of similar textures in the chloride-carbonate nodules (Fig. 10 c-e) is
the first “snapshot” record of the unambiguous chloride-carbonate melt immiscibility in
rocks. The previous natural evidence was based on melt and fluid inclusions in the skarn
minerals of Mt Vesuvius (Fulignati et al., 2001) and kimberlitic diamonds (Bulanova et al.,
1998; Izraeli et al., 2001; Izraeli et al., 2004; Klein-BenDavid et al., 2004). However, the
extensive review of experimental studies (Veksler, 2004) points to the lack of data for
chloride-carbonate systems.
Given the analogy with the texture of melt inclusions at the onset of immiscibility, the
boudin-like shape of the carbonate sheets and their subparallel alignment (Fig. 10c, e),
argues for preservation of primary (instantaneous) immiscibility texture. This means that
post-immiscibility (< 600 o C) cooling and crystallisation were fast enough to prevent
aggregation of one of the immiscible liquids into ovoid or spherical globules that are more
typical of steady-state immiscibility. Occurrence of the chloride-rich veinlets in the
carbonate sheets (Fig. 11) testifies to later solidification of the chloride liquid relative to
carbonate crystallisation. The round and ameboid-like bleb textures of sylvite in halite (Fig.
11) are also reminiscent of liquid immiscibility. In theory this contradicts the fact of
complete miscibility in the system NaCl-KCl above the eutectic point of ~660 o C. However,
the separation of the Na-K chloride melt from the carbonatitic melt, in the case of
Udachnaya-East residual melt pockets, occurred at temperatures below the eutectic, and