Does the Core Leak?
Richard J. Walker1,*and David Walker2
1 - Department of Geology, University of Maryland, College Park, MD 20742, USA
2 – Lamont Doherty Earth Observatory and Department of Earth and Environmental Sciences of
Columbia University, Palisades, NY 10968, USA
* - Correspondence author
([email protected])
Text + References Word Count
2039
Submitted To:
EOS
February 18, 2005
1
Within the past ten years, appreciable geophysical evidence has accumulated to suggest that
some slabs subduct into the deep mantle, and that certain mantle plumes originate in the lower
mantle. Despite this progress in understanding geodynamics on the geophysical front, there is
still much debate regarding the existence of corroborating geochemical evidence. If some plumes
rise from the core-mantle boundary and there is chemical interaction between the core and
mantle, it is possible that such plumes could contain a unique fingerprint that is characteristic of
the core. Unequivocal identification of a core component in a plume-derived rock could
potentially settle major questions in geodynamics. Here we review geochemical tools that would
likely be sensitive to such interactions, and consider possible causes of, and mechanisms for
core-mantle interaction.
Consideration of Tools for Detecting Core-Mantle Interaction
Probably the most sensitive suite of elements for geochemically identifying possible core
contributions to plumes are the highly siderophile elements (HSE = Pt, Re, Os, Ir, Pd, Ru, Rh, Au), and
certain moderately siderophile elements (MSE) such as Ag and W. The HSE are so termed because of their
extreme affinity for metal relative to silicate, with bulk distribution coefficients >104. Because of the
extreme preference of the HSE for metal relative to silicates, the formation of the core nearly quantitatively
sequestered the Earth’s HSE, and also likely dominates the budget of MSE. Mass balance is, therefore,
potentially optimal for detection of core additions to the mantle. The high affinity for iron also means that
the relative abundances of these elements in the bulk core are probably little fractionated compared to
chondritic meteorites, or the bulk earth. Formation of the inner core, however, may have fractionated
the HSE, as occurred in differentiates of asteroidal cores sampled by iron meteorites .
Discovering a unique isotopic fingerprint of the outer core using HSE and MSE is a
promising means of identifying a core component in rocks collected from Earth’s surface. The
coupled 187Re-187Os and 190Pt-186Os long-lived isotope systems (187Re t½ = 42 b.y.; 190Pt t½ = 460
b.y.) may be useful in identifying the presence of an evolved outer core component in mantlederived rocks [Walker et al., 1995]. This hypothesis is based on conclusions from the study of
asteroidal core crystallization and both low and high-pressure experimental studies indicating that
Pt/Os and Re/Os ratios may be substantially higher in the outer core than in chondritic meteorites or
the silicate mantle, as a consequence of inner core crystallization.
If an inner core with substantial mass formed within ~2 b.y. of planetary formation, the
outer core could exhibit coupled 187Os and 186Os enrichments of >7% and >0.01%, respectively,
relative to non-radiogenic isotopes of Os, as compared to chondrites. The degree of isotopic
enrichment would reflect both the cumulative formation age of the inner core, and the magnitude
of partitioning of these elements between solid and liquid metal. It is now possible to measure
Pt-Re-Os partitioning between liquid and solid Fe metal alloys at relatively high pressures,
although still well below core conditions. Work at 100 kbars indicates that the relative and
absolute partitioning characteristics of these elements are not sensitive to pressure [Walker,
2000]. The magnitude of partitioning, however, is strongly correlated with S and P content.
Variable mixing between Os contained in an outer core component and “normal” Os present in a
mantle plume could lead to formation of a suite of rocks that define a linear trend on a plot of
187
Os/188Os versus 186Os/188Os (Figure 1). Coupled enrichments in 186Os-187Os similar to those
predicted have been detected in putative plume-derived rocks related to the 251 Ma Siberian
flood basalt event, the 89 Ma Gorgona Island (Colombia) komatiites, and picritic rocks of the
Hawaiian plume [e.g. Brandon et al., 2003].
The 186Os-187Os “test” for core material, however, is not without its weaknesses. It
requires relatively early crystallization of the inner core, which is contrary to some current
models for the timing of inner core crystallization. Furthermore, processes other than coremantle interaction can be envisioned to explain some observations [Smith, 2002]. Consequently,
Os isotopes cannot serve as the sole arbiter of core additions to plumes, and the addition of other
isotopic tracers to the study of core-mantle interaction may be critical to unambiguously identify
an outer core component. For this task, so called “short-lived” isotope systems are likely the best
hope to confirm a core signature implicated by Os isotopes. The decay of short-lived nuclides
during the first few tens of Ma of solar system history could have led to the generation of
isotopic compositions of HSE or MSE in the core that are resolvable from the silicate Earth.
Most germane to studies of core-mantle interaction is the 182Hf-182W (182Hf t½ = 8 m.y.) system.
Hafnium is a lithophile element, so there is little Hf in the core. Tungsten is a MSE, with ~90%
of the Earth’s W residing in the core. Consequently, the Hf/W of the core is extremely low.
Recent work on the 182Hf-182W isotopic system has shown that the silicate Earth has a 182W
isotopic composition that is approximately 2 parts in 10,000 higher than chondritic meteorites.
Assuming the bulk Earth has a chondritic W composition, mass balance requires that the core
must have a W isotopic composition that is similar to chondritic, or slightly more than 2 parts in
10,000 lower in 182W than mantle rocks. Thus, sufficient transfer of core W to a plume could
generate a resolvable depletion in 182W relative to the bulk silicate Earth. Initial study of W in
Hawaiian lavas examined previously for Os found no deviation from the silicate Earth [Schersten
et al., 2004], however the magnitude of 182W depletion may be less than current mass
spectrometric resolution.
Investigation of physical mechanisms of core to mantle transfer.
Whether or not isotopic signals from the core are recognizable in rocks at the Earth’s
surface, it is interesting to consider how core material could be reincorporated into the mantle.
Physical entrainment of metal into a rising plume via capillary action has previously been
considered [Walker et al., 1995]. This mechanism seemingly would link HSE and MSE content
with Fe and Ni. Other possible mechanism that may not result in such linkages need also be
explored. The materials from Earth’s surface we examine for their core signature have been
oxidized so they no longer contain metal. Mechanisms in which the oxidative step is intimately
connected to the process of escape from the core may be important. Such processes include
oxidative exsolution, oxidative titration and electrochemical transfer.
Secular cooling of the Earth with core solidification will eventually precipitate buoyant
saturating oxides or oxide liquids This oxidative exsolution from the core as it cools could cause
transfer of siderophile material from core to mantle. The cause of the transfer would be the
decrease in O2 solubility in very high pressure metallic liquid with falling temperature, as the
core cools. This putative transfer mechanism supposes as a base-line assumption that O2 is
appreciably soluble in the liquid metal of the outer core. An important question then, is whether
oxygen is soluble in liquid Fe at high pressure? It is well known that FeO is only sparingly
soluble in Fe liquid near its melting point at 1 bar. Experimental information on the pressure
variation of oxygen solubility has been highly controversial.
Increasing oxygen solubility with pressure for several 10s of kbars pressure and projected
high and increasing solubilities for oxygen with pressure to several hundreds of kbars has been
previously noted [Ohtani and Ringwood, 1984]. When pressure levels to 250 kbars were
eventually examined with multi-anvil (MA) techniques, however, oxygen solubilities seemed to
decrease with increasing pressure [O’Neill et al., 1998].
In contrast to the multi-anvil studies, diamond anvil cell (DAC) studies at significantly
higher pressures [Knittle and Jeanloz, 1991] inferred a nearly quantum upward leap in oxygen
solubility above 300 kbar. The onset of high solubility was correlated with the initiation of a
reaction between Mg-Fe perovskite and liquid Fe that proceeds only at P > 300 kbar – beyond
the P range of the multi-anvil observations.
Quantitative determination of the solubility of oxygen in liquid metal has not yet been
satisfactorily resolved experimentally. The present state of the analysis does not permit a
quantitative prediction of how large FeO solubility becomes, nor do quenching problems with
the MA and DAC experiments allow meaningful direct determinations of the oxygen solubility
at lower mantle pressures and temperatures. Alternate experimental strategies to determine
oxygen solubility in the outer core have been proposed [Walker, 2005], but await action. At the
present state of knowledge, it is reasonable to suppose that secular cooling and exsolution of
oxide liquid/crystals from the core contribute to core/mantle transfer, however, if the megabar
solubility of oxygen in liquid metal remains modest, then this mechanism is probably not
significant in generating core to mantle exchange.
Another mechanism to consider for driving transfer to the mantle is to simply inject
oxygen-rich material into the outer core, causing oxidative titration. Morse (2000) imaginatively
explored scenarios involving highly ferrous magma cupolas at the base of the mantle. Such basal
fluxing baths are an attractive locus for transfer. Should highly oxidized slab tops find their way
to the base of the mantle and be processed through one of these cupolas on top of the core, or the
core-mantle boundary (CMB), oxidative titration with transfer of siderophile material from the
core to the mantle would be a plausible consequence. Oxide-rich magma chambers thought to
exist at the CMB, atop the outer core, would provide an ideal transfer medium for digestion of
oxidized input, causing oxidative corrosion of the outer core’s surface. The corrosion product
would be the physical manifestation of the geochemical kick the oxidized material may give to
core-to-mantle transfer, irrespective of core cooling. This would eliminate the need to have
substantial temperature dependence to the solubility, a requirement if cooling is to generate much
precipitate. In the titration scenario, the limit on the yield is no longer the solubility or its
temperature dependence, but the delivery efficiency of reactants to the CMB. Low oxygen
solubility would make the titrative precipitation of core material more responsive to the input
stream of oxidized slab material. It is interesting to imagine that the biosphere could have some
impact on the Earth’s core.
One final mechanism to be considered for driving core-to-mantle transfer of matter is
electrochemical transfer. The dynamo produces a voltage across the CMB interface and chemical
transport must occur. Electrochemical transfer mechanisms can send material in either direction
depending on the strength and polarity of the field, which changes in time and space (Kavner and
Walker, in review). Very little data currently exist with regard to the transfer of trace elements.
So this is a topic that will require much future investigation.
REFERENCES CITED
Brandon A.D., R.J. Walker, I.S. Puchtel, H. Becker, M. Humayun, S. Revillon (2003) 186Os-187Os
systematics of Gorgona Island komatiites: implications for early growth of the inner core. Earth
Planet. Sci. Lett. 206, 411-426.
Kavner A. and D. Walker. Core/mantle-like interactions in an electric field. Earth Planet. Sci. Lett.,
in review.
Knittle E and R. Jeanloz (1991) Earth’s core-mantle boundary: results of experiments at high pressures and
temperatures. Science 251, 1438-1443.
Morse SA (2000) A double magmatic heat pump at the core-mantle boundary, American Miner.
85, 1589-1594.
Ohtani E., A.E. Ringwood, W. Hibberson (1984) Composition of the core, II. Effect of high
pressure on solubility of FeO in molten iron. Earth Planet. Sci. Lett. 71, 94-103.
O’Neill H. St.C., D. Canil D, D.C. Rubie (1998) Oxide-metal equilibria to 2500oC and 25 GPa:
implications for core formation and the light component in the Earth’s core. J. Geophys. Res. 103
12239-12260.
Schersten A., T. Elliott, C. Hawkesworth, M. Norman (2004) Tungsten isotope evidence that mantle
plumes contain no contribution from the Earth’s core, Nature 427, 234-237.
Smith A.D. (2003) Critical evaluation of Re-Os and Pt-Os isotopic evidence on the origin of
intraplate volcanism. Journ. Geodyn. 36, 469-484.
Walker D. (2000) Core participation in mantle geochemistry: Geochemical Society Ingerson Lecture, GSA
Denver, October 1999. Geochim. Cosmochim. Acta 64, 2897-2911.
Walker D. (2005) Core issues. Canadian Mineralogist. Fleet volume, in press.
Walker R.J., J.W. Morgan, M.F. Horan (1995) 187Os enrichment in some plumes: evidence for core-mantle
interaction. Science 269, 819-822.
Figure 1. The D" magma chamber complex at the CMB is a crucible for digestion of oxidized
slab material, potentially leading to reactive extraction of core material into the source region of
some plumes. The signature of recycled crust in those plumes is a natural consequence of that
crustal material providing the chemical impetus to extract core material. Also shown are predicted
effects on evolution of 186Os/188Os versus 187Os/188Os in chondrites and the outer core. Trend for outer core
is based on iron meteorite analogy and assuming early growth of inner core. Crosses represent compositions
at 1 billion year intervals. Data for Gorgona, Hawaiian and Siberian systems are from Brandon et al. (2003)
and references cited therein.
186Os/188Os
ng
0.119860
Mi
xi
0.119870
Mauna Loa
Loihi
Koolau
Mauna Kea
Evolution of OC
Kilauea
Hualalai
Kohala
Siberia
Gorgona
0.119850
N2 + O2
CoreCoreflavored
plume
Liquid metal
outer core
Oxidized
slab
Inner
core
solid
mantle
0.119840
0.119830
Chondritic Evolution
0.119820
0.100
0.120
187
D" with magmatic
titration chambers
0.140
188
Os/ Os
Walker and Walker Figure 1