Introduction: The presence of
alternately magnetized sources on Mars indicates that at one time Mars had a
magnetic field, presumably generated by an internal dynamo, and that its
polarity reversed at least once before shutting off1. Mars does not
currently possess a magnetic field, so the anomalies are attributed to remanent
magnetization (RM) of minerals in the crust. The initiation and cessation of
the inducing field, its timing, and the implications for Mars thermal evolution,
compositional differentiation, and volatile budget are critical components of
understanding the planetary dynamics of Mars. Isobaric (1 bar) constant-rate cooling experiments were run
using Fe-rich (i.e., Martian) basalt to ascertain whether conditions necessary
for crystallization of minerals retaining intense RM are consistent with any of
the models proposed for generation of the Martian magnetic anomaly.
Cooling experiments: The material investigated is synthetic A*, a liquid
composition in equilibrium with olivine and pyroxene during crystallization of
the Shergottites, Chassigny, and other cumulate SNC meteorites2. The compositional and rheological
characteristics of A* are shared by other proposed SNC parent melts (low in Al2O3
and FeO-rich)3,4.
The first series of experiments determined the
near-liquidus phase relations for the A* composition over a range of fO2 values corresponding to the Fe-wŸstite
(IW) to MnO-Mn3O4 (MNO) solid buffer assemblages (Figure
1). Care was taken to minimize Fe
loss during the experiments: runs at the most reducing conditions were executed
in Mo foil, and the intermediate to oxidizing fO2 experiments were run in Pt capsules
pre-saturated with Fe by running A* at the experimental temperature for 12-24
h. Sample capsules and solid
buffer assemblages were loaded into fused quartz tubes, evacuated to a moderate
vacuum (²10 Pa), then sealed shut.
The solid buffer assemblage was inspected at the termination of each run
to confirm or bracket the experimental fO2.
A second matrix of 24 experiments were run to examine
the kinetics of low-pressure oxide crystallization as a function of fO2 and cooling rate. Evacuated fused quartz capsules
containing sample + buffer (IW, QFM, NNO, MNO) were bundled together, brought
to 1210 oC, then cooled to ~300 oC. Five sets of constant-rate cooling
experiments (230, 72, 19, 6, and 3 oC h-1) were run using
a programmable temperature controller.
Rapid (highly nonlinear, ~1.5E5 oC h-1) cooling
was imposed by immersion of the capsules in water.
Textural characterization: A dazzling variety of titanomagnetite (TiMt) and
clinopyroxene (Cpx) morphologies were produced that broadly correlate with
cooling rate (Figure 2).
Texturally, there was no discernible effect of experiment fO2, except on the abundance of
titanomagnetite in the charge. For
both TiMt and Cpx, increasing experiment duration (lower cooling rate) correlates
with increasing crystal content, grain size coarsening, and morphological
transition from highly skeletal, cruciform (TiMt), and dendritic (Cpx) forms
toward euhedral forms. Except for
the MNO charge, which started below the TiMt liquidus at 1210 oC, no
crystallization was observed in the most rapidly cooled (10 E5 oC h-1)
series. Long strands of collinear
TiMt crystals evident in 2D images appear to represent single crystals linked
outside the plane of the cut surface.
These morphologies complicate the determination of crystal nucleation
rate by standard image processing techniques5. Instead, the ratio of the mineral
surface area to mineral volume is considered a measure of the surface free
energy controlling nucleation kinetics in these experiments. We calculated a nondimensional index of
surface energy by finding the ratio of the glass-TiMt interfacial boundary
length, including internal holes, to the circumference of a single circle
having the equivalent area.
Preliminary measurements indicate that the index ranges over several
orders of magnitude, consistently increasing in proportion to cooling rate.
Magnetic characterization: Samples
from the cooling experiments were characterized for concentration of
ferromagnetic material, magnetic domain state, magnetic mineralogy, ability to
carry anhysteretic remanent magnetization and thermoremanent magnetization (ARM
and TRM, respectively) using a suite of applied field- and
temperature-dependent methods. The
range of laboratory cooling rates used was sufficient to generate magnetic
mineral assemblages with variable composition and magnetic domain states.
Several broad trends are evident. The samples prepared at the MNO buffer
have the highest saturation magnetization and the highest magnetic susceptibility
of any given cooling rate applied.
The samples prepared at the IW buffer have the lowest saturation magnetization
and the lowest magnetic susceptibility for any given cooling rate. The QFM and NNO sample sets are very
similar with respect to the abundance of magnetic material. Consistent with the petrographic
observations, we observe a magnetic grain size coarsening with slower cooling
rates for the QFM, NNO, and MNO samples.
Using in-field measurements and high-temperature methods, the IW sample
set appears dominantly paramagnetic with the exception of the 
sample synthesized at the slowest cooling rate. However, low temperature methods and
induced remanences (ARM and TRM) indicate the presence of a remanence-bearing
phase in the IW samples. We did
not observe pure end-member
magnetite or hematite. Curie temperatures range from ~300-510 oC,
consistent with Ti-rich titanomagnetite or Ti-rich titanohematite. Samples prepared at the MNO buffer had
the highest Curie temperatures (475-495 oC), with no apparent
relationship to cooling rate.
Samples prepared under the QFM and NNO buffers had lower Curie temperatures,
although there appears to be a significant shift in mineralogy at the slowest
cooling rates. Nearly all samples
in the QFM, NNO, and MNO sets displayed a magnetic order/disorder transition in
the 30-65K range, with no apparent relationship to cooling rate.
We imparted TRM at 550oC in
ambient fields of 10, 20, 50 and 100 mT (Figure 3). At the 50 mT step, we observe TRM intensities of 0.1-70 x10-3 Am2kg-1.
Summary: Magnetic
characteristics of rapidly cooled A* basalt crystallized under moderate to
highly oxidizing conditions are comparable to those described for rapidly
cooled terrestrial basalts6 and modeled as potential source
materials for the Martian anomalies7,8.
References: [1] Stevenson D. J. (2001) Nature 412, 214-219.
[2] Johnson M.C., Rutherford M.J., and Hess P.C. (1991) Geochim.
Cosmochim. Acta 55 349-366.
[3] McSween H.Y., Eisenhour D.D., Taylor L.A., Wadhwa M., and Crozaz G.
(1996) Geochim. Cosmochim. Acta 60 4563-4569.
[4] Wadhwa M. and Crozaz G. (1995) Geochim. Cosmochim. Acta 59
3629-3645. [5] Hammer, J.E. and Rutherford M.J. (2002) J. Geophys. Res. 107 10.1029/2001JB000281. [6] Zhou, W et al., (2000) Earth
Planet Sci Lett 179 9-20. [7] Nimmo,
F., (2000) Geology 28 391-394. [8] Connerney J.E.P., (1999) Science 284 794-798.