Geoscience Reference
In-Depth Information
a growing inner core and further gravitational
separation in the outer core.
In the
inhomogenous accretion model
the early
condensates, calcium--aluminum-rich silicates,
heavy refractory metals, and iron accreted to
form the protocore. The early thermal history is
likely to be dominated by aluminum-26, which
could have produced enough heat to raise the
core temperatures by 1000 K and melt it even
if the Earth accreted 35 My after the Allende
meteorite, the prototype refractory body. Melting
of the protocore results in unmixing and the
emplacement of refractory material (including
uranium, thorium and possibly 26 Al) into the
lowermost mantle. Calculations of the physical
properties of the refractory material and normal
mantle suggest that the refractories would be
gravitationally stable in the lowermost mantle
but would have a seismic velocity difference of
a few percent.
the core. Various scenarios have been invented
to explain the siderophile abundances in the
mantle; these include rapid settling of large iron
blobs so that equilibration is not possible or a
late veneer of chondritic material that brings
in siderophiles after the core is formed. The
siderophiles are not fractionated as strongly as
one would expect if they had been exposed to
molten iron. Some groups of siderophiles occur
in chondritic ratios in upper mantle rocks.
The highly siderophile elements (Os, Re, Ir,
Ru, Pt, Rh, Au, Pd) strongly partition into any
metal that is in contact with a silicate. These ele-
ments are depleted in the crust--mantle system
by almost three orders of magnitude compared
with cosmic abundances but occur in roughly
chondritic proportions. If the mantle had been
in equilibrium with an iron-rich melt, which was
then completely removed to form the core, they
would be even more depleted and would not
occur in chondritic ratios. Either part of the melt
remained in the mantle, or part of the mantle,
the part we sample, was not involved in core
formation and has never been in contact with
the core. Many of the moderately siderophile ele-
ments (including Co, Ni, W, Mo and Cu) also
occur in nearly chondritic ratios, but they are
depleted by about an order of magnitude less
than the highly siderophile elements. They are
depleted in the crust--mantle system to about
the extent that iron is depleted. These elements
have a large range of metal--silicate partition
coefficients, and their relatively constant deple-
tion factors suggest, again, that the upper man-
tle has not been exposed to the core or that
some core-forming material has been trapped
in the upper mantle. It is not clear why the
siderophiles should divide so clearly into two
groups with chondritic ratios occurring among
the elements within, but not between, groups.
The least depleted siderophiles are of intermedi-
ate volatility, and very refractory elements occur
in both groups.
Mantle--core equilibration
Upper mantle rocks are extremely depleted in
the siderophile elements such as cobalt, nickel,
osmium, iridium and platinum, and it can
be assumed that these elements have mostly
entered the core. This implies that material in
the core had at one time been in contact with
material currently in the mantle, or at least
the upper mantle. Alternatively, the siderophiles
could have experienced preaccretional separa-
tion, with the iron, from the silicate material
that formed the mantle. In spite of their low con-
centrations, these elements are orders of magni-
tude more abundant than expected if they had
been partitioned into core material under low-
pressure equilibrium conditions. The presence of
iron in the mantle would serve to strip the side-
rophile elements out of the silicates. The mag-
nitude of the partitioning depends on the
oxidation state of the mantle. The 'overabun-
dance' of siderophiles in the the upper mantle
is based primarily on observed partitioning
between iron and silicates in meteorites. The con-
clusion that has been drawn is that the entire
upper mantle could never have equilibrated with
metallic iron, which subsequently settled into
Light element in the core
The core's density is about 10% less than that
of Fe (or Fe--Ni alloy) at core conditions (Figure
10.1) and thus there is a significant amount of
an element or element mixture having a lower
