[meteorite-list] Squeezing Meteorites to Reveal the Martian Mantle

From: Ron Baalke <baalke_at_meteoritecentral.com>
Date: Tue, 19 Dec 2006 16:32:25 -0800 (PST)
Message-ID: <200612200032.QAA28383_at_zagami.jpl.nasa.gov>

http://www.psrd.hawaii.edu/Dec06/Y-980459.html

Planetary Science Research Discoveries

Citation: Taylor, G. J. (December, 2006) Squeezing Meteorites to Reveal
the Martian Mantle. Planetary Science Research Discoveries.


Squeezing Meteorites to Reveal the Martian Mantle
December 19, 2006

--- Experiments at high temperature and pressure give clues to the
composition of the interior of Mars.

Written by G. Jeffrey Taylor
Hawai'i Institute of Geophysics and Planetology

A piece of a Martian lava flow, Antarctic meteorite Yamato-980459,
appears to represent the composition of a magma produced by partial
melting of the Martian interior. That's the view of researchers Don
Musselwhite, Walter Kiefer, and Allan Treiman (Lunar and Planetary
Institute, Houston) and Heather Dalton (Arizona State University).
Musselwhite and his colleagues determined that this basaltic Martian
meteorite represented a primary melt from the mantle. This was an
important discovery because magma produced inside a planet contains
significant clues to the composition of the region of the interior in
which it formed. The lava flows that decorate the surface of planets
tell us about the mantle, the rocky region beneath the crust and above
the metallic core.

The researchers used apparatus at the Johnson Space Center to determine
what minerals are present when samples with the composition of Y-980459
are heated to a range of temperatures and squeezed to a range of
pressures like those that planetary scientists expect to exist in the
interior of Mars. The results indicate that the magma represented by
this special meteorite formed at a depth of about 100 kilometers and a
temperature of about 1540 C. From the
high temperature and high ratio of magnesium to iron in the magma,
Musselwhite and his colleagues infer that the amount of melting to
produce the Y-980459 parent magma was high, which suggests that the
temperature at the boundary between the metallic core and the rocky
mantle was higher than previous estimates. This work gives us clues to
the composition and dynamics of the Martian interior--all from a rock
chipped off a lava flow on Mars and flung to Earth by an impact.

Reference:

    * Musselwhite, D. S., H. A. Dalton, W. S. Kiefer, and A. H. Treiman
      (2006) Experimental petrology of the basaltic shergottite
      Yamato-980459: Implications for the thermal structure of the
      Martian mantle. Meteoritics and Planetary Science, v. 41, p.1271-1290.

------------------------------------------------------------------------

A Special Rock from Mars

All meteorites from Mars are special, of course. Some, such as the
shergottites, are pieces of more-or-less familiar lava flows. Others are
bits of unusual lava flows, such as the nakhlites. One is ancient, ALH
84001, and represents an accumulation of the mineral orthopyroxene deep
inside the crust of Mars, but later modified by water and possibly
containing evidence for past life on Mars (see the first PSRD article:
Life on Mars? <http://www.psrd.hawaii.edu/Oct96/LifeonMars.html>). But
Y-980459 has the added virtue of having the composition of a magma
that has not been modified much since it formed by partial melting of
the Martian interior a few hundred million years ago. The age of
Y-980459 is 472 +/- 47 million years based on Rb-Sr and Sm-Nd isotopic
measurements.

Samples of a partial melt of a planetary interior are like the Holy
Grail. They were at one time in equilibrium with minerals in the
interior, so they contain a record of the chemical and mineralogical
composition of their place of formation. They are messengers from the
interior.

The shapes and sizes of the minerals in Y-980459 and the way the
minerals are intergrown indicate that Y-980459 is a piece of a lava
flow. It consists of large grains of olivine embedded in a matrix of
smaller crystals, including lath-shaped plagioclase feldspar. There is
no question that the minerals crystallized in a lava flow.

[Thin section of Y-980459 Two views of a thin slice of Y-980459.]

TOP: Photomicrograph of a thin slice of the meteorite as viewed in
polarized light in a microscope. The large grains are olivine. They are
surrounded by a finer-grained intergrowth of plagioclase feldspar and
pyroxene. The straight edges of the olivine suggest that they formed in
a magma.

BOTTOM: Map of the magnesium (Mg) concentration in the minerals in
Y-980459, showing the same area as in the photograph on the top. Many of
the grains have high Mg concentrations, indicative of olivine.

[Mg concentration map of Y-980459]

What Don Musselwhite and other experimentalists want is to know the
composition of the magma (also called "melt" by experimental
petrologists) produced by partial melting. The problem is that many
magmas begin to crystallize and the early-formed minerals accumulate in
some portions of the magma (see illustration below). Olivine is usually
the first major mineral to crystallize, so often a lava flow will
contain large olivine crystals dragged up with the magma from a storage
area (magma chamber) beneath the ground. If olivine has accumulated, the
chemical composition of the lava does not represent a melt because it
has been modified by addition of crystals.

[crystals accumulate]

In many cases, one mineral forms first as a magma
begins to crystallize. Crystals of this mineral can accumulate inside
the magma chamber. If a batch of the magma containing the extra crystals
erupted, its total composition (crystals plus liquid magma) would not
represent the composition of the original crystal-free magma. Y-980459,
however, appears not to have accumulated extra olivine crystals, hence
its chemical composition is the same as it was when it formed by partial
melting of the Martian mantle.

There is a way to determine if olivine has been accumulated. If olivine
crystallized directly from the lava it is found in, it will have the
composition (specifically the ratio of magnesium to iron) predicted from
experiments to form from a magma of the composition of the lava rock
(large olivine crystals plus the finer-grained groundmass surrounding
them) in which they occur. Many lava flows contain a olivine crystals
with a large range of compositions, which indicate that they accumulated
in the magma. Their presence affects the bulk chemical composition and
the lava does not represent a pristine liquid derived from the mantle.
The composition of Y-980459 has been measured by Japanese researchers E.
Koizumi and his coworkers. Musselwhite's calculations indicate that the
olivine crystals in Y-980459 would have formed from a melt with the
composition of the meteorite. Thus, it is likely that Y-980459
represents an unmodified melt from the Martian mantle. It is a probe of
the interior of Mars.

------------------------------------------------------------------------

Revealing Mantle Mineralogy

Once planetary scientists know that a rock is a primary melt from the
interior of a planet, they want to find out as much as they can about
the composition of the mantle rock in which the magma formed and the
pressure (hence depth) at which it formed. To do this, Musselwhite and
Dalton made a homogeneous glass with a chemical composition equal to
that measured by other investigators for Y-980459. Making starting
compositions for experiments is part of the art of experimental
petrology. In this case, the investigators made a mixture of oxides and
carbonates and ground them together in an agate mortar and pestle, with
the powdery mixture immersed in acetone to prevent contamination and
loss of the finest powder. The mixed powder was then melted in a
high-temperature furnace and ground up again to ensure that it was
homogeneous. Because multiple experiments were going to be done on the
same starting material, it was essential that it be homogeneous.

Experiments were done at high temperature and pressure in what PSRD
calls a Squeeze-O-Matic. Experimental petrologists call this particular
piece of gear a "Quickpress non-end-loaded piston cylinder apparatus."
The sample was placed inside a graphite capsule that in turn was placed
inside a pressure cell made of barium carbonate and magnesium oxide.
This was squeezed to a pressure between 7 and 15 kilobars (one bar is
the pressure of air at sea level) and heated to temperatures ranging
from 1410 to 1615 oC. After holding the sample at a given pressure and
temperature for up to 24 hours, it was quenched to room temperature
first, then the pressure was released. Musselwhite and his colleagues
took great care to ensure that the measurements of pressure and
temperature were accurately calibrated, and include in their paper an
extensive Appendix explaining their procedure. Products of the
experiments were made into polished pieces and analyzed with an electron
microprobe at the Johnson Space Center.

[Quickpress apparatus]
The Quickpress apparatus used in the experiments on
Y-980459. This device generates very high pressures at high temperatures.

olivine and glass in BSE image Electron microscope image of a polished
sample of an experimental run product. The dark material at the top is
the graphite capsule the sample was placed in. The light gray is glass
(which was molten during the high temperature and pressure at which the
experiment was run), and the gray area contains olivine crystals.
Electron microprobe analysis of the glass and olivine gives their
chemical compositions.

A central purpose of the experiments was to determine the pressure and
temperature at which the Y-980459 magma formed. But how do we figure
that out? The experiments show what minerals are present at different
compositions, what their compositions are, and what the composition of
the co-existing melt is. At some pressure the experimental product might
be olivine and melt (which is preserved as glass in the experimental
samples). At another pressure and temperature pyroxene might be present
with glass. This means that if the Y-980459 magma formed at those
conditions the left over solid rock would contain just one mineral,
olivine or pyroxene. This is unlikely. For a typical mantle rock inside
a planet, it takes about 50% melting to leave just one mineral behind.
Hot magmas are so mobile (low viscosity) that they readily move up in
the planet when the amount of melting is as low as a few percent, and
quite rapidly at 20%. Thus, experimental petrologists look for the
pressure and temperature at which a melt co-exists with two or more
minerals. For Y-980459 that pressure is 12 kilobars (plus or minus 0.5
kilobars), equivalent to a depth of about 100 kilometers inside Mars.

    * A quick reminder: When most substances melt they do not go from
    solid to liquid at a single temperature. Pure materials like ice do
    melt at a single temperature, but complex ones like rocks melt over
    a range of temperatures and the chemical composition of the molten
    material and the minerals present change as the amount of melting
    increases.

[graph of experimental results]

This plot shows results of the experiments
done by Don Musselwhite and his colleagues. Above the uppermost diagonal
line, a magma with the composition of Y-980459 would be completely
molten; this line is called the "liquidus," which means that everything
is molten above it. In the lefthand field labeled "olivine + melt" it
would consist of olivine and molten silicate; this corresponds to
conditions as the magma was nearing eruption at low pressure. At high
pressure the magma consists of pyroxene and melt. There is a region in
between where melt is accompanied by olivine and pyroxene. The liquidus
at this point is the likely formation pressure and temperature of the
Y-980459 magma, 12 kilobars (equivalent to a depth of 100 kilometers on
Mars).

This approach to determining the depth of origin and mineralogy of the
mantle was applied to basalt lava flows from the lunar maria, too.
Investigators (mostly in the 1970s and 1980s) sought primary magmas like
Musselwhite and his colleagues did for Martian meteorites. Experiments
at high pressure and temperature showed that the pressure at which a
melt co-existed with two or more solids varied, corresponding to depths
of 100 to 400 kilometers.

------------------------------------------------------------------------

Amount of Melting and Composition of the Mantle

The experiments suggest that when the Y-980459 magma formed the leftover
solid minerals were olivine and low-calcium pyroxene. Neither contains
much aluminum or calcium, so the amount of melting must have been high
enough to completely dissolve minerals that contain those elements
(plagioclase feldspar, oxides, or garnet). This implies that the amount
of melting was at least 15-20%. The experiments tell us what the olivine
composition was when the magma separated. It contained 86 mole percent
forsterite (Mg2SiO4) and 14 mole percent fayalite (Fe2SiO4). This
composition can be abbreviated as the Mg#, simply the ratio of magnesium
to magnesium plus iron [Mg/(Mg+Fe)], in this case 0.86. This means that
the leftover olivine in the mantle after the Y-980459 magma had migrated
away had an Mg# of 0.86.

Knowing the composition of the residual minerals is useful, but we
really want to know the composition of the mantle before it melted. That
gets us a step closer to figuring out the composition of the entire
Martian mantle. Unfortunately, this cannot be determined uniquely from
knowing the composition of the residual minerals because different
amounts of melting give different calculated Mg#. Musselwhite and his
colleagues calculated the Mg# of the initial, pre-melting Y-980459
mantle source rock as a function of the amount of melting. To do this,
they used well-established geochemical equations that describe the
partial melting process. As shown in the diagram below, the amount of
melting could range from an infinitesimal amount (corresponding to Mg#
of 0.86) to over 70% (Mg# of 0.75 in the original mantle rock).

[graph of experimental results]

This graph shows calculations of the Mg#
of the mantle for different amounts of partial melting to produce
Y-980459. The common view is that the average Martian mantle has a Mg#
of about 0.75. If so, then formation of Y-980459 required over 70%
melting, an unreasonably large value in a single event. In fact, if the
amount of melting was much over 40% all the pyroxene would have been
melted, leaving only olivine in the residual solid. Musselwhite and his
colleagues suggest that the source in the mantle had already experienced
one or more melting events before the one that produced the magma in
which Y-980459 crystallized.

Percentages of melting higher than 50% are not reasonable, except in the
early stages of a planet's life when it could have been mostly molten.
This means that either the Martian mantle has a higher Mg# than we think
(say 0.80 instead of 0.75), or that multiple episodes of melt extraction
occurred. Either is reasonable and we have no unique way of determining
which is correct. It is possible that there was an initial large melting
event, say the Martian magma ocean (see PSRD article: A Primordial and
Complicated Ocean of Magma on Mars
<http://www.psrd.hawaii.edu/Mar06/mars_magmaOcean.html>),
resulting in formation of a region with relatively high Mg#. Subsequent
melting could have formed the Y-980459 magma, possibly after an
intermediate melting event or events that produced other, unsampled magmas.

------------------------------------------------------------------------

Thermal State of the Interior

Results of these experiments and calculations have implications for the
temperature inside Mars as a function of time. Co-author Walter Kiefer
used a complex geophysical computer model to estimate temperature inside
Mars, from the core-mantle boundary to the base of the lithosphere (the
upper, relatively cool, rigid part of the planet that extends from the
surface to a depth of about 200 kilometers). In this case he updated the
approach he took previously, taking into account the high temperature
reached to melt the mantle that gave rise to Y-980459, and including
some other tricks of the geophysics trade. The fact is that these
calculations are both informative and uncertain. The uncertainty stems
from the large number of variables that must be accounted for, including
temperature, viscosity of the solid mantle, the volume of magma produced
over time and over specific time periods, the percentage of melting, how
many melting episodes, to name a few.

[figure of potential temperature]

This shows the temperature inside Mars
at about 280 million years ago (the middle of the range in ages of the
shergottie Martian basalts), using Walter Kiefer's previous
calculations. The central hot feature (red) is a plume of hot mantle
rock that flowed upwards from the core-mantle boundary. When shallow
enough, it began to melt, indicated by the white region. The updated
calculations are similar, but tying them to the high melting temperature
needed to make Y-980459 raises the temperature at the core-mantle
boundary. ("Potential temperature" means the physical temperature minus
the effects that pressure has on temperature caused by the work done to
squeeze solid rock.)

The new calculations suggest that the temperature at the boundary
between the mantle and the core is hotter than calculated previously.
This might be caused by the core containing less sulfur than thought, a
higher viscosity than used in the calculations, or by a high
concentration of radioactive elements (potassium, thorium, and uranium)
at the base of the mantle. These elements are important because they
release heat when they decay, raising the local temperature. The last
hypothesis may be most likely and is consistent with geophysical models
of what could have happened in a Martian magma ocean. This is itself a
very complicated process, but it seems likely that it would have
overturned, depositing dense rock at its base. Some of the dense
minerals, such as garnet, would have concentrated thorium and uranium.
See PSRD article: A Primordial and Complicated Ocean of Magma on Mars
<http://www.psrd.hawaii.edu/Mar06/mars_magmaOcean.html> for details.

------------------------------------------------------------------------

Just the Beginning . . .

Identifying a primary melt from the mantle of Mars is a another step on
the road to unraveling the detailed composition of the Martian interior.
The shergottite Martian meteorites have already told a complicated story
of the mantle (see PSRD article: The Multifarious Martian Mantle
<http://www.psrd.hawaii.edu/June04/martianMantle.html>). Nevertheless, we
need more samples that represent magma compositions, and preferably
magmas that represent formation over a range of time back to more than
four billion years ago. To find them we need more samples. The search
for Antarctic meteorites and meteorite finds in desert regions on Earth
are helping fill in gaps, but they are biased towards younger samples.
Returning samples from volcanic areas and the ancient Martian surface
would help enormously. There might even be mantle rocks excavated when
huge basins such as Hellas (2500 kilometers in diameter) formed. The
sample returns could be supplemented by installation of a geophysical
network that could use Mars quakes to probe the interior. The chemical
composition of Mars is an important piece of the puzzle of planet
formation, and Martian samples help us put those pieces together.

------------------------------------------------------------------------

ADDITIONAL RESOURCES

    * Kiefer, W. S. (2003) Melting in the Martian mantle: Shergottite
      formation and implications for present-day mantle convection on
      Mars.Meteoritics and Planetary Science, v. 38, p. 1815-1832.
    * Musselwhite, D. S., H. A. Dalton, W. S. Kiefer, and A. H. Treiman
      (2006) Experimental petrology of the basaltic shergottite
      Yamato-980459: Implications for the thermal structure of the
      Martian mantle. Meteoritics and Planetary Science, v. 41, p.1271-1290.
    * Taylor, G.J. (1996) Life on Mars? Planetary Science Research
      Discoveries. http://www.psrd.hawaii.edu/Oct96/LifeonMars.html
    * Taylor, G.J. (2004) The Multifarious Martian Mantle. Planetary
      Science Research Discoveries.
      http://www.psrd.hawaii.edu/June04/martianMantle.html
    * Taylor, G.J. (2006) A Primordial and Complicated Ocean of Magma on
      Mars. Planetary Science Research Discoveries.
      http://www.psrd.hawaii.edu/Mar06/mars_magmaOcean.html
Received on Tue 19 Dec 2006 07:32:25 PM PST


Help support this free mailing list:



StumbleUpon
del.icio.us
reddit
Yahoo MyWeb