|
Spin transitions and
thermodynamics of lower mantle materials
Amy Bengtson,
Dane Morgan, Materials Science Program
The lower mantle is approximately 650-2800 km
below the Earth’s surface and exists under
extremely high pressures and temperatures,
making experimental investigation of its
properties very challenging.
For studying this region, computational modeling
is an extremely powerful and informative tool.
At pressures similar to those in the lower
mantle, iron is known to undergo a spin
transition in both (Mg1-x,Fex)O
ferropericlase and (Mg1-x,Fex)SiO3
perovskite, the two most prevalent materials in
the lower mantle.
The goal of my research is to use ab initio
calculations to study these spin transitions and
the different properties of these two phases.
These spin transitions may have an important
impact on lower mantle thermodynamics, phase diagrams,
and iron partitioning, and we use our
calculations on the low-, intermediate-, and
high-spin phases to better understand the spin
transition in the lower mantle. The overall
impact of this research is to lead to more
accurate models of thermal transport, density,
rheology, and thermoelasticity within the Earth.

Figure:
High- to low-spin
transition pressure as a function of Fe
composition. The transition pressure trend is
opposite in ferropericlase and perovskite,
suggesting the Fe spin transition has a strong
structural dependence.
We find that the
spin transition as a function of iron
composition has a very different trend in the
two main lower mantle materials: ferropericlase
and perovskite. Our results show that at high
Fe content, Fe in perovskite will be low-spin
under lower mantle pressure, but Fe in
ferropericlase will remain high-spin. These
calculations agree with experimental trends in
composition for ferropericlase, while
experimental studies of perovskite are still
uncertain due to the more complex structure and
possible presence of intermediate spin.
We gratefully
acknowledge financial support from the National
Science Foundation (NSF), Earth Sciences (EAR)
division, award number 0738886. We gratefully
acknowledge computing support from the National
Science Foundation (NSF) National Center for
Supercomputing Applications (NCSA), award number
DMR060007.
|