The following are the high lights of his
research (the numbers in brackets correspond to the the number of the publication
in the bibliography):
Since the Earth’s interior is inaccessible,
the only information on the composition of the deep Earth is indirect.
Therefore, it is necessary to consider the cosmochemical constraints on
planetary chemistry. The basic tenet is that if it is not in the solar gas, it
cannot be in the terrestrial planets. The concentration of a chemical species in
the Earth, depending on the accretion history, should have some proportional
relationship to its solar abundance. During the 1980’s, we developed programs
and data bases to compute equilibrium condensation of solids from solar gas.
This was a substantial improvement over previous studies because the technique
of Gibbs free energy minimization was used (52) and solid solutions were
included in the data base (63,64). These studies clearly mapped the
pressure-temperature stability of several solids found in different classes of
meteorites. For the first time, the study covered the condensation of both
hydrous and anhydrous species (63). The data bases and the calculations have
acquired a new significance because of the current interest in the formation and
composition of the asteroidal bodies and their exploration for space resources.
Using the density data on various terrestrial planets, the possibility of the
formation of planetary bodies, entirely by condensation from the solar gas, was
explored and it was demonstrated that this could have happened in the primitive
solar cloud (71). In these models, the Earth continues to remain chondrititc in
composition which has been the geochemical consensus for several decades.
2.
Computation of phase equilibrium
Computation of phase equilibrium to understand
the chemical composition of the Earth’s core and mantle has always remained as
one of our important goals. Earth’s outer core has liquid properties and its
density is substantially less than that of iron. Several dilutents have been
proposed to explain this property. One early study (51),
showed that it was thermodynamically possible to dissolve sufficient Si
in iron to lower the density to an appropriate level.
Computation of multicomponent phase
equilibrium by using the minimization of Gibbs free energy in a given chemical
system (62) has been one of the mainstays of our work. In these studies, the
chondritic model composition has been used to simulate the mantle (65,113) and
core densities both under hydrous and
anhydrous conditions. These studies have led to various interesting results. The
mantle density can be modeled in the ternary system (MgO-FeO-SiO2) for a
chondritic mantle but the iron content has a strong influence on the density
variation (113). The mineralogical model, while not requiring any seismic data
for its construction, reproduces the seismic discontinuities and the density
variation in the mantle which are almost exact replica of those produced by the
Preliminary Earth Reference Model (PREM). Calculated adiabatic geothermal
gradient starting at 6 GPa and 1500 K reaches a temperature of 2046 K at the
core/mantle pressure (135 GPa) in a pyrolite mantle. The model Earth parameters
in the lower mantle are (PREM parameters in bracket): Ks = 308 (306) to
687 (656) GPa; f
= 70 (69) to 121 (118) km2s-2.
3.
Fluids in the Earth’s interior
The question whether there is any significant
amount of fluid phase in the deep mantle has always occupied our minds. It is
quite evident that a certain amount of trapped primitive fluid in the mantle
would facilitate the geodynamical processes. With this view in mind and
particularly for modeling processes in the proto-planetary bodies, we have
worked on a program to model high pressure fluids. This work started with formal
thermodynamic applications (76-80,99) and then continued with calculations using
molecular dynamics methods (90,91,96,98) at ultra-high pressures. The models
have now been incorporated in the Gibbs free energy program (Chemsage) and can
be used for systems of as many as 13 different fluids in the system (C-H-O-S-N-Ar)
(100). With such multicomponent non-ideal fluid models, it is possible to study
the role that fluids must have played in the early history of the planet forming
processes. Our work (81,82) showed that the fluid in growing primitive mantle of
carbonaceous chondritic composition was dominantly methane with subordinate
amounts of hydrogen and water. Such a composition is stable over a broad range
of pressure and temperature. The solids in equilibrium with such fluids would be
olivine and pyroxene (or their high pressure equivalents), graphite or diamond
and iron. This mixture is quite
appropriate for an undifferentiated Earth. At upper mantle pressures, the fluid
composition is strongly influenced by presence or absence of free iron. A fluid
with as much as 75% methane could be in equilibrium with olivine (13% fayalite)
without metallic Fe as a coexisting phase. The oxygen fugacity of the primitive
mantle with such fluid composition would be several log units below that of the
quartz-fayalite-magnetite buffer.
4. Thermodynamic
data bases
Thermodynamic data bases are crucial to all
our applications requiring the calculation of the planetary processes. We have
devoted considerable efforts to systematize and maintain a thermodynamic data
base which is both geological as well as metallurgical. Our early efforts were
in the field of solid solution modeling which resulted in the publication of two
monographs (1,3). The question of the choice and implementation of the solid
solution models has now reached a stage that we do not have to worry about the
details which are all computerized. The Chemsage (GTT,Aachen) and Thermocalc
(Stockholm) programs permit the use of a variety of solutions and we are happy
to have contributed to these programs which can be used for all types of
geophysical and geochemical computations. The improvements in solution models
followed the attempt to obtain internally systematized data bases. We worked on
improving the analytical expressions for heat capacity (4,78), thermal expansion
and bulk modulus (4,85-97) and making all these internally consistent
(4,72,89,125). Since all individual parameters, e.g. heat capacity, thermal
expansion and compressibility are constrained to vary with one another and with
temperature, our data base can be used with confidence for extrapolations
(113,125). The use of the data base permits the calculations on the variations
of planetary densities, thermal gradients and
the nature of the Earth’s core and its interaction with the lower mantle.
5.
High pressure and high temperature experiments
Iron
Phase Diagrams and Earth’s Core:
Earth
has a large core reaching to a depth of nearly 2900 km from its center; this
core stores a substantial part of the planet's energy and, therefore, exercises
significant influence on the dynamic processes. Iron has always been considered
as well-suited to form a major part of the core; it is sufficiently abundant and
seems to have the right density. Seismic data require that the core contain a
solid inner core and a liquid outer core. The study of iron at Uppsala has been
focused on three aspects. The first one concerns melting to ultra-high pressures
(103,104,105), the second with finding new iron phases (104, 105, 108, 110, 117,
121)
and the third with systematizing a thermodynamics data base (104, 113, 114) with
which to calculate the phase equilibrium relations in the proto-Earth, the deep
mantle and the core. Iron melting
has been done to a pressure of 140 GPa (1.4 megabar) reaching the outer core of
the Earth (104). These data along with those of
R. Boehler’s to 195 GPa have been used to calculate the melting curve
of iron to pressures reaching the center of the Earth for which a temperature of
6150 K is estimated (104). We have generated a considerable debate after
proposing that iron may occur as a new polymorph beta, now recognized to be a
double layer DHCP structure. We have confirmed this structure to pressures as
high as 130 GPa or 1.3 megabar. Several other interesting polymorphic
transitions are possible at high pressure and temperature. Our latest result
have reached pressures as high as 200 GPa .
Stability
of perovskite in the deep Earth: Perovskite (MgSiO3) has always
been regarded as the stable phase in the mantle. It is generally considered as
occurring with magnesiowustite and the iron is distributed between the two solid
solutions. We asked the question: what if the silicate was not stable with
respect to the oxide mixture MgO + SiO2. While it may or may not make
a major difference in the density profile of the deep Earth, it would surely
constrain all the iron to be contained in the magnesiowustite causing important
redistribution of species in the deep mantle. We followed this research both
experimentally (119) and theoretically (121,122). It was discovered that pure
MgSiO3 perovskite broke down to MgO+SiO2 between 60 and 70
GPa in the diamond-anvil cell. Such a reaction could be stress related.
Therefore, we studied the problem theoretically and found that an oxide mixture
does become more stable than perovskite at pressures between 115 (15) GPa.
Finally, we have repeated the experiment again with a differently designed
diamond-anvil cell with electrical heating and and studied the
dissociation reaction: MgSiO3 (perovskite) = MgO (periclase) +
SiO2 (high pressure silica phase) at 82 (3) GPa by in-situ heating to
temperatures between 300 and 1780 (50) kelvin and x-ray diffraction. The
orthorhombic perovskite changed to a pseudo-cubic phase between 1280 to 1485
kelvin. The oxide mixture was observed to grow at
temperatures between 1600 to 1700 kelvin and orthorhombic perovskite was
recovered upon cooling at 1140 K. Compositionally, the instability of perovskite
makes the oxide mixture as the most probable mantle material which can provide a
variety of possible mantle compositions. Dynamically, the lighter SiO2 component
would facilitate differentiation of the lower mantle.
The new mineral physics data has some very important consequences for
modeling Earth’s early history. Hot condensing
material in a nebular setting, where differentiated planetesimals with their
iron cores are plentiful, forms the solid protocore (Fe-Ni-S-C) which reacts
with FeO available in magnesiowustite and due to dissociation of perovskite at
the core-mantle interface. The interface grows to form
the liquid core at the expense of the solid proto-core. The model is
consistent with geochemical data and has important implications for the dynamo
and for the rotation of the inner core.