Garnets - Rare Earth Elements in Ultramafic and Mafic Rocks and their Minerals - page 63

found in the diamond fields from Liqhobong, Lesotho, Jagersfontein and Premier
(South Africa) provinces [Nixon et al ., 1987; Hoal et al ., 1994; Shimizu et al ., 1994;
Pearson et al ., 1998], in xenoliths of megacrystalline peridotites from Udachnaya,
Mir, Aykhal pipes and others (Yakutia) [Pokhilenko et al ., 1993; Pearson et al ., 1993;
Hoal et al ., 1994; Shimizu & Sobolev, 1995; Shimizu et al ., 1997a, b; Pearson &
Milledge, 1998], in xenoliths of harzburgites and lherzolites from Roberts Victor pipe
(South Africa) [Stachel et al ., 1998], in xenoliths of eclogites from Mbuji-Mayi
deposit (Congo) [Demaiffe et al ., 1998], in xenocrysts of kimberlites from Camsell
Lake (Canada) [Pokhilenko et al ., 1998]. As mentioned above, the sinusoidal REE
patterns have been also ascertained in many garnets, which are micro-inclusions in
diamond crystals. However, there are examples when sinusoidal patterns, which are
typical for inner zones of garnet crystals from micro-inclusions in diamonds, moving
towards the peripheral zones were replaced by those more common for garnet pat-
terns of simple form with a steep positive slope [Shimizu et al ., 1997b]. Apart from
sinusoidal shape of patterns in the garnets associating with diamonds, other kinds of
anomalous patterns were found: in particular, arched up or with an almost flat shape
in the range between Sm and Lu.
The nature of sinusoidal shape of the REE patterns of garnets is still a subject
of debate. Originally it was assumed that the origin of such patterns in garnets was
caused by metasomatic change of mantle substrate that preceded the formation of
garnets and diamonds in peridotite parageneses [Shimizu, Richardson, 1987]. Later it
was suggested that peridotites, containing garnets with a sinusoidal REE distribution,
right after its formation have undergone a metasomatic recycling under the influ-
ence of deep-seated melts that were geochemically similar to carbonatite and that this
process was a really time-consuming one [Pokhilenko et al ., 1993]. It is the proximity
of the REE compositions of garnets from micro-inclusions in diamond crystals and
their varieties, which are parts of peridotite xenoliths from Roberts Victor and Akwa-
tia pipes, that allowed to assume that the garnets from harzburgites and lherzolites,
presented in the xenoliths, were formed from protolith, which had previously been
depleted during partial melting in stability field of spinel [Stachel et al ., 1998]. There
is also a hypothesis that subcalcium pyropes from deep-seated xenoliths were formed
during the metasomatic transformation of spinel harzburgites in the stability field
of diamond, and the process involved carbon-bearing fluids or melts [Jacob et al .,
1998]. It was also assumed that zoning in the distribution of REE and other impuri-
ties in subcalcium garnets, associating with diamonds, could be due to the continuous
growth of grains on the background of the changes in PT- conditions of crystallization
[Shimizu et al ., 1994]. It was taken into account that the inner zones of garnet grains,
apparently, were originally crystallized under nonequilibrium conditions, which have
been replaced by the equilibrium conditions till the moment of crystallization of the
peripheral zones, and it was considered that the growth of zonal garnet crystals and
associating diamonds occurred shortly before they got into the kimberlite matrix.
According to Shimizu et al. [1997b], a similar mechanism of formation of garnet
can explain the transformation from the sinusoidal patterns of REE distribution to
ordinary patterns as we move from the inner zones of garnet crystals towards the
peripheral.
Apparently, for a more accurate study of the proposed mechanisms of arising
for sinusoidal patterns of REE distribution in garnets, associating with diamonds,