NASA Technical Reports Server (NTRS) 20060056239: Early Planetary Differentiation: Comparative Planetology

0.03 MB·English

by  NASA Technical Reports Server (NTRS)

EARLY PLANETARY DIFFERENTIATION: COMPARATIVE PLANETOLOGY. J.H. Jones, KR,
NASA/JSC, Houston, TX 77058 ([email protected]).

Introduction: We currently have extensive data Even if this separation occurred without the aid of
for four different terrestrial bodies of the inner solar a magma ocean, the chemical complementarity be-
system: Earth, the Moon, Mars, and the Eucrite Parent tween the lunar highlands and the depleted mare
Body [EPB]. All formed early cores; but all(?) have source regions is undeniable. Melt was clearly ex-
mantles with elevated concentrations of highly sidero- tracted from the lunar interior early in its history, de-
phile elements, suggestive of the addition of a late “ve- pleting the lunar mantle in incompatible elements.
neer”. Two appear to have undergone extensive dif- These incompatibles were concentrated into a reservoir
ferentiation consistent with a global magma ocean. known as KREEP. The uniformity of elemental ratios
One appears to be inconsistent with a simple model of in KREEP is itself an argument for a magma ocean —
“low-pressure” chondritic differentiation. Thus, there i.e., a single event.
seems to be no single, simple paradigm for understand- Model ages of KREEP are nearly concordant be-
ing early differentiation. tween several different chronometer systems: 4.35-
EPB. Currently, there is no evidence that eucrites 4.45. This suggests that the magma ocean lasted about
are other than simple partial melts of a chondritic (CO- 200 m.y. Perhaps coincidentally, this is the age of
like) parent body. The agreement between phase equi- lunar differentiation inferred from 142Nd systematics.
librium experiments on natural eucrites and de- Therefore, there are hints that early lunar differentia-
volatilized Murchison (CM) is remarkable. The most tion lasted for quite some time.
compelling reason for believing that eucrites might be Unlike the EPB, 182W anomalies in lunar samples
the result of equilibrium or fractional crystallization of are small to nonexistent. This suggests that, even
a more primitive magma is the desire to relate eucrites though lunar differentiation was prolonged, it was not
and diogenites (i.e., howardites). But at this juncture, initiated until 182Hf was nearly extinct. These observa-
no model has been able to produce such a relationship tions would also be consistent witth a Moon that dif-
in a quantitatively satisfying way. Both partial melting ferentiated quickly at ~4.35 b.y. and that inherited a
experiments of chondrites and constraints from eucrite small 182W anomaly.
Sc/La ratios imply that eucrites were produced by It is now generally believed that the Moon has a
~20% partial melting of a CO-like source. Such a small core, consistent with the redox conditions in-
source is depleted in low-Ca pyroxene by 1200°C, ferred earlier.
with a maximum MgO content of ~En , whereas al- Experiments show that lunar basalts formed near
70
most all diogenites have Mg#’s > 70. Experiments IW-1, similar to eucrites.
show that eucrites formed at IW-1, consistent with a Unlike the EPB(?), there are enclaves of the lunar
chondritic ol-opx-metal assemblage. mantle that are probably quite fertile. For example, the
The EPB differentiated very early while 182Hf was REE pattern of the A15 Green Glass is rather flat. It is
still present (t = 9 m.y.). Therefore, eucrite petro- not chondritic and has a small negative Eu anomaly,
1/2
genesis occurred much less than 50 m.y. after CAI but it is not strongly depleted in the LREE. This
formation. Since the EPB appears to have experienced would make it more fertile than modern-day MORB
core formation, this means that core formation oc- mantle. Therefore, models of lunar differentiation
curred at this time, or earlier. must accommodate a wide range of incompatible ele-
Lithophile incompatible elements were quantita- ment depletions. Note that the A15 Green is not
tively removed from the eucrite source regions. In- KREEPy; so REE abundances were not established by
compatible element ratios in eucrites are generally a late-stage mantle overturn.
chondritic. Mars. Mars differentiated early into core, mantle,
The Moon. The Moon is the type locality for the and crust. This is established by the significant 182W
magma ocean hypothesis. Mare basalts (and their anomalies in martian meteorites. Like mare basalts
source regions) are depleted in alumina and have nega- and unlike eucrites, martian basalts all come from de-
tive Eu anomalies. Conversely, the highlands crust is pleted source regions and it is inferred that this deple-
enriched in plagioclase and has a positive Eu anomaly. tion also occurred very early.
Therefore, there is a natural complementarity between There are at least two varieties of martian mantles:
these two reservoirs, suggesting separation of a plagio- one that produced the nakhlites and chassignites and
clase component by floatation from a magma ocean. another that produced the shergotttites. In particular,
the shergottitte mantle is so depleted that it is difficult
to understand how shergottite basalt generation occurs. the mare basalt source regions by ~100 m.y. However,
In one model-dependent (but internally consistent) such an age may be consistent with the discovery of a
model, the nakhlite source region is twice as depleted small (~2ε) 182W anomaly in terrestrial rocks. Al-
in incompatibles as the terrestrial MORB mantle; and though if the core formed early, it is unclear why the
the shergottite mantle is three times more depleted still. W anomaly is not larger. Presumably either the core
There is great similarity between the Lu-Hf and formed later than 4.54 b.y. or nonradiogenic chondritic
Sm-Nd systematics of shergottites on the one hand and W was added after core formation. In the latter case, it
high-Ti mare basalts and KREEP on the other. This is unclear why this event is not recorded in the Pb sys-
suggests that, if the Moon went through a magma tem (i.e., why isn’t the Pb age of the Earth younger?).
ocean stage, so did Mars. The redox state of the Earth is much more compli-
The redox state of the martian mantle is not known cated than any other planet and is much beyond the
precisely. However, Fe-Ti oxide assemblages in the scope of this abstract. There is an apparent paradox in
most primitive shergottites [i.e., those with the highest that terrestrial rocks have less FeO than their extrater-
initial ε(Nd)] indicate an oxygen fugacity of ~IW. restrial counterparts (~8 wt.% FeO). Taken at face
And the similarity of FeO contents in basalts from the value this means that core formation occurred under
Moon, Mars and the EPB indicates that all three of more reducing conditions (~IW-2.5). Yet, mantle
these bodies were at ~IW-1 at the time of core forma- rocks have oxygen fugacities of ~QFM. How can
tion. And the shergottite mantle is near that value. these observations be reconciled?
Because there are correlations between redox state Venus and the Earth have similar FeO contents, so
and ε(Nd) within the shergottite suite, I believe the there is a strong possibility that planet size (i.e., pres-
most oxidized shergottites have been influenced by sure) plays a role. Recent experiments, in fact, suggest
crustal contamination. If so, then the martian crust is that FeO may disproportionate into Fe metal and Fe3+
more oxidized than the martian mantle (~QFM). at pressures as low as ~250 kbar. If metal produced in
The redox state of the nakhlite/chassignite mantle this way segregates to the core, this could explain both
may also be more oxidizing. Wadhwa found no Eu the high-fo2 of the upper mantle and its low FeO con-
anomaly in Nakhla clinopyroxenes; but Nakamura did. tent. Further, thermodynamic calculations indicate that
Wadhwa also found Eu anomalies in clinopyroxene low pressure (< 30 kbar) pyrolite mineral assemblages
from Chassigny. Therefore, the redox state of the at QFM would become more consistent with IW at
nakhlite/chassignite mantle is uncertain. Even so, the higher pressures. Again, the main difference between
similarity between the FeO contents of nakhlites and Earth/Venus and the other terrestrial planets is size.
shergottites suggests that at the time of core formation, The extent to which the Earth has been differenti-
differences in redox states could not have been large. ated is unclear. Oceanic basalts often have elevated
Part of our uncertainty about nakhlites stems from 3He/4He ratios and solar Ne, suggestive of primoridial
the complex history of the nakhlite mantle. Compari- signatures. In addition, fertile spinel lherzolites from
son of the short- and long-lived chronometers in the continental lithospheres may show little evidence of
Sm-Nd system indicates that, at some time in the past substantial melt removal (i.e., they are still lherzolites).
(~4 b.y.), the nakhlite mantle was refertilized. This is On the other hand, no mantle reservoir yet sampled has
tricky and difficult to do, suggesting a differentiation a chondritic Nb/U ratio, as expected for undifferenti-
event internal to the nakhlite mantle. ated mantle. These observations, taken as a whole,
In summary, Mars differentiated very early into argue for extensive but incomplete differentiation. A
reservoirs that still exist today. The petrology of the corollary of these observations is that there is no ter-
martian mantle is complex but maybe less so than that restrial evidence for a magma ocean (unlike the case of
of the Moon. The variation in redox state on Mars is, the Moon/Mars) or for a Moon-forming giant impact.
however, more complex than on the Moon. Another peculiarity of the Earth is that material has
Earth. Of course, the most complex planet of all been returned to the mantle via subduction tectonics.
is the Earth. And because of that, a great deal of its This has allowed the transport of hydrous minerals to
earliest history has been erased. Therefore, oddly, the mantle, which would otherwise be dry (like the
though we know more about the Earth than any other other terrestrial planets).
planet, it’s earliest conditions are murky. And because the Earth has abundant surficial wa-
The Earth differentiated early to form a core. Tera ter, it may be the only planet with granitoid continents.
has argued that this event occurred at 4.54 b.y. How- Tonalites, the main component of ancient continental
ever, the arguments are complex to the non- shields, are formed by the partial melting of altered
plumbologist. But, if so, this predates the formation of (i.e., hydrated basalts).