Theory of the Earth
Don L. Anderson
Chapter 2. Earth and Moon
Boston: Blackwell Scientific Publications, c1989
Copyright transferred to the author September 2, 1998.
You are granted permission for indi
vidual, educational, research and
noncommercial reproduction, distribution, display and performance of this work
in any format.
Recommended citation:
Anderson, Don L. Theory of the Earth.
Boston: Blackwell Scientific Publications,
1989.
http://resolver.calt
ech.edu/CaltechBOOK:1989.001
A scanned image of the entire book may
be found at the following persistent
URL:
http://resolver.caltech.edu/CaltechBook:1989.001
Abstract:
Recent compilations of cosmic abundances
(Table 2-1) agree fairly closely for the
more significant rock-forming elements. From these data a simple model can be
made (Table 2-2) of the Earth's bulk chem
ical composition. The mantle is nearly
completely oxidized and is therefore composed of compounds primarily of MgO,
SiO
2
, Al
2
O3, CaO, Na
2
O and Fe-oxides.
Earth and
Moon
Strange all this
dgerence
should
be
Twixt
tweedle-durn
and
tweedle-dee.
-
JOHN
BYROM
BULK
COMPOSITION
OF
THE
EARTH
Recent compilations
of
cosmic abundances
(Table
2
-
1)
agree fairly closely for the
more
significant rock
-
forming
elements. From these data a simple model can
be
made
(Table 2
-
2)
of
the
Earth's bulk
chemical composition.
The mantle is nearly completely oxidized
and
is there
-
fore composed
of
compounds primarily
of
MgO,
SiO,,
A1,0,,
CaO,
NazO
and
Fe
-
oxides. The major minerals in
rocks from the upper mantle are
MgO
+
SiO,
=
MgSiO,
(enstatite)
and
2Mg0
+
SiO,
=
Mg,SiO,
(forsterite)
Iron
occurs in various oxidation states,
FeO,
Fe,O,,
Fe,O,,
and typically replaces about 10 percent
of
the
MgO
in the
above
minerals:
(Mg,.,Feo.,)
SiO,
(orthopyroxene)
(Mgo.,Feo.,),
SiO,
(olivine)
These minerals can
be
considered as solid solutions
between
MgSiO,
and
FeSiO,
(ferrosilite)
and
Mg,SiO,
and
Fe,SiO,
(fayalite). Another important solid
-
solution series is
be
-
tween
MgO
(periclase)
and
FeO
(wustite) giving
(Mg,Fe)O
(magnesiowiistite
or
ferropericlase). These oxides
usually
combine with
SiO,
to
form silicates,
but they
may
occur
as
separate phases in the
lower
mantle.
The approximate composition
of
the
mantle
and
the
relative sizes
of
the
mantle and
core can
be
determined from
the
above
consideration. Oxygen
is combined with
the ma
-
jor
rock
-
forming elements,
and then
the relative
weight
fractions can
be
determined from the
atomic weights.
The
preponderance
of
silica
(SiO,)
and
magnesia
(MgO)
in man
-
tle
rocks
gave
rise to the
term
"
sima,
"
an archaic
term
for
the chemistry
of
the mantle. In contrast, crustal rocks
were
called
"
sial
"
for silica
and
alumina.
The
composition in the
first column
in
Table 2
-
2 is
based
on
Cameron's
(1982) cosmic abundances. These are
converted to
weight
fractions via the molecular
weight and
renormalization. The
Fe,O
requires
some
comment.
Based
on cosmic abundances, it is plausible
that
the
Earth's
core
is
mainly
iron;
however,
from seismic data
and
from the
total
mass and moment
of inertia
of
the Earth,
there must
be
a light
alloying
element
in
the core.
Of
the candidates
that have been
proposed
(0,
S,
Si,
N,
H,
He
and
C),
only
oxygen
and
silicon are likely to
be brought
into a planet
in
refractory solid
particles
-
the
others are all
very volatile
elements
and
will
tend
to
be
concentrated
near
the surface
or in the atmosphere. The hypothetical high
-
pressure
phase
Fe,O
has
about
the right density to
match
core values. There
is also the possibility
that
the oxidation state
of
iron
de
-
creases
with
pressure from
Fe
3
+ to
Fez+
to
Fe+
to Fe
O
. At
high pressure
FeO
is soluble in molten iron,
giving
a com
-
position similar
to
Fe,O
at core
pressures
(Ohtani
and
Ring-
wood, 1984).
If most of the
iron
is in
the
core,
in
Fe,O
proportions,
then
the mass
of
the core will
be 30
to
34 weight
percent
of
the planet. The actual
mass
of
the core is 33 percent,
which
is excellent agreement. There
may
also,
of
course,
be some
sulfur, carbon,
and
so on
in
the core,
but
little
or
none
seems
necessary. The
FeO
content
of
the mantle
is less
than
10 percent
so
that does
not
change the
results
much. The
mass
of
the core
will
be
increased
by
about
5
percent if
it
TABLE
2
-
1
Short
Table
of Cosmic Abundances
(AtomsISi)
Element
Cameron
Anders and Ebihara
(1982)
(1982)
has
a cosmic
NiIFe
ratio,
and will be
larger
still if the newer
estimates
of
solar
iron
abundances are used,
unless
the iron
content
of
the
mantle
is greater
than
is evident from
upper-
mantle
rocks. The
new
solar
values
for
FeISi
give
15%
weight percent
FeO
for
the
mantle
if the core
is
32%
with
an
Fe20
composition.
This
FeO
content is similar to esti
-
mates
of
the
composition
of
the
lunar
and
Martian mantles.
Note
that
the
cosmic atomic
ratio
MgISi
is 1.06.
This
means that the mantle
(the iron
-
poor oxides or silicates) is
predicted
to
be
mainly pyroxene,
MgSiO,,
rather
than
oli
-
vine,
Mg2Si0,,
which
has
a
MgISi
ratio
of
2
(atomic). The
presence
of
FeO
in
the
mantle means
that slightly more
olivine can
be
present. It is
the
(Fe
+
Mg)/Si
ratio
that
is
important.
Garnet also
has an
MgISi
ratio near
1,
although garnet
usually
has
more
FeO
than
do
olivine
and
pyroxene. This
suggests that
much
of
the mantle
is composed
of
pyroxene
and
garnet, rather
than
olivine,
which
is the
main
constitu
-
ent
of
the
shallow
mantle. The
CaO
and
A1203
are contained
in garnet
and
clinopyroxene
and
the
Na20
in
the
jadeite
molecule
at upper
mantle
pressures. Seismic data are con
-
sistent
with
a mainly
"
perovskite
"
lower mantle.
MgSiO,
transforms
to
a perovskite structure,
an
analog
of
CaTiO,,
at lower
-
mantle pressures.
MgSiO,
"
perovskite
"
is there
-
fore the dominant mineral
of
the mantle.
TABLE
2
-
2
Simple Earth
Model
Based
on
Cosmic Abundances
Oxides
MolecuPes
Molecular
Weight
Grams
Weight
Fraction
MgO
1
.06
SiO,
1
.OO
A1zO,
0.0425
CaO
0.0625
Na,O
0.03
Fe,O
0.45
Total
The crust is rich
in
SiO,
and
Al,O,
(thus
"
sial
"
),
CaO
and
Na,O.
The
Al,O,
and
CaO
and
Na20
content
of
the
"
cosmic
"
mantle
totals
to
5.8
percent. The crust represents
only
0.5
percent
of
the
mantle. Thus, there
must be
consid
-
erable potential crust
in
the
mantle. The thinness
of
the ter
-
restrial crust,
compared to the Moon and
Mars
and
relative
to the potential crust, is partially related
to
the presence
of
plate tectonics on Earth,
which
continuously recycles
crustal material.
It is also
due
to high
pressures in the Earth,
which
convert crustal materials
such
as basalt to
high-
density eclogite.
Estimates
of
upper
-
mantle compositions
based
on
per-
idotites are deficient
in
SO2,
TiO,,
A120,
and
CaO
relative
to cosmic abundances. Basalts,
picrites
and
eclogites are
enriched
in
these oxides. Cosmic ratios
of
the refractory
oxides
can
be
obtained
with
a mixture
of
approximately
80
percent peridotite
and
20
percent eclogite.
If
the
lower
mantle
is
mainly MgSiO,
(
"
perovskite
"
),
then
a much
smaller eclogitic, or basaltic, component
is required to
bal
-
ance the
MgISi
ratio
of
upper
-
mantle peridotites. A
chon-
dritic Earth contains
about
6
to
10 percent
of
a basaltic
component.
EVOLUTION OF
THE
EARTH'S
INTERIOR
The evolution
of
Earth,
its
outgassing
and
its
differentiation
into a crust,
mantle
and
core depend
on
whether
it accreted
hot or cold
and
whether it
accreted
homogeneously
or
in-
homogeneously.
It
now
appears
unavoidable
that a large
fraction
of
the gravitational energy
of
accretion
was
trapped
by
the Earth,
and
it therefore started life
as
a hot body
rela
-
tive to
the
melting point. The processes
of
differentiation,
including core formation,
were
probably occurring while
the
Earth
was
accreting,
and
it is therefore misleading
to
talk
of
a later
"
core
formation event.
"
Melting near the sur
-
face leads to a zone
refining
process
with
light
melts
rising
to the surface
and
dense
melts and
residual crystals sinking
toward
the center.
What
is not
so clear
is
whether
most
of
the
Earth accreted homogeneously or inhomogeneously. A
certain
amount of
inhomogeneity in the chemistry
of
the
accreting material is required, otherwise reactions
between
H20
and
free
iron
would
oxidize
all
the iron
and
no water
would
exist at the surface. This
can be
avoided
by
having
more
iron accrete
in
the early stages
and
more
H,O
in
the
later stages. Iron metal,
of
course, condenses earlier in a
cooling nebula,
and
because
of
its density
and
ductility
may
have formed
the
earliest planetesimals
and
perhaps the
ini
-
tial nuclei
of
the
planets.
In earlier theories
of
Earth evolution, it
was
assumed
that cold volatile
-
rich material similar to type
I
carbona
-
ceous chondrites accreted
to
form a homogeneous planet,
perhaps
with
some
reduction
and
vaporization at the surface
to form reduced iron,
which
subsequently warmed
up and
differentiated into
an
iron
-
rich core
and
a crust.
In
these
theories it
proved difficult to
transport
molten
iron to the
center
of
the planet because
of
the
effect
of
pressure
on
the
melting point:
Molten iron
in
the
upper mantle
would
freeze
before it reached the
lower
mantle. A central iron
-
rich
nu
-
cleus
mixed with
or surrounded
by
refractory
-
rich material
including aluminum
-
26, uranium
and
thorium
would
alle
-
viate this thermal
problem
as
well
as
the iron
-
oxidation
problem. The presence
of an
ancient magnetic field,
as
re
-
corded
in the oldest rocks, argues for a sizeable molten iron
-
rich core early in
Earth's history.
The abundance
of
sidero-
phile elements
in
the
upper mantle also suggests that
this
region
was
not
completely stripped
of
these elements
by
molten
iron draining to the core.
The early
and
extensive
melting and
differentiation
of
the
Moon and
some meteorite parent
bodies
attests to the
importance
of
melting and
differentiation
of
small bodies in
the early solar system. This
melting
may
have
been
caused
by
the energy
of
accretion or the presence
of
extinct, short
-
lived radioactive nuclides.
In
some cases, as
in
present
-
day
10,
tidal pumping is
an
important
energy
source. Isotopic
studies indicate
that
distinct geochemical reservoirs formed
in
the
mantle early
in
its
history. Magmas from these res
-
ervoirs retain their isotopic identity,
and this proves
that the
mantle has not
been homogenized
in spite
of
the fact that it
has
presumably
been
convecting throughout
its
history. A
chemically stratified mantle is one
way
to
keep reservoirs
distinct and separate until partial
melting
processes
allow
separation
of
magmas,
which
rise quickly
to
the
surface.
Zone
refining
during accretion
and
crystallization
of
a
deep
magma ocean
are possible
ways
of
establishing a
chemically
zoned planet
(Figure
2
-
1).
At
low
pressures
ba
-
saltic
melts
are less dense
than
the residual refractory crys
-
tals,
and they
rise to the surface,
taking with
them
many
of
the trace elements. The refractory crystals
themselves
are
also less dense
than
undifferentiated
mantle and tend
to con
-
centrate
in
the
shallow
mantle.
At
higher pressure there is a
strong likelihood that melts
become
denser
than
the crystals
they
are
in
equilibrium with. Such melts, trapped at depth
and insulated
by
the overlying rock,
may
require consider
-
able time to crystallize.
Most scientists agree that simplicity is a desirable at
-
tribute
of
a theory. Simplicity,
however,
is in
the eye
of
the
beholder. The
end
results
of
natural processes can
be in
-
credibly complex,
even
though the underlying principles
may be
very simple. A uniform,
or
homogeneous, Earth
is
probably the
"
simplest
"
theory, but
it violates the simplest
tests.
Given
the
most
basic observations about the
Earth,
the next simplest theory is a three
-
part Earth: homogeneous
crust, homogeneous mantle
and
homogeneous core.
How
the Earth might
have achieved this
simple state,
however,
involves a complex series
of
ad
hoc
mechanisms to separate
core
and
crust from the primitive mantle
and then
to
ho
-
mogenize the separate products.
On
the other hand, one can assume that the processes
PRIMITIVE
MANTLE
EVOLI
RESIDUE
n
JTION
OF
'I
rHE
E
A
R
T
H
'
S
INTERIOR
29
CFB
MORB
010
DEPLETED
I
FIGURE
2
-
1
A
model
for
the evolution
of
the mantle. Primitive mantle
(1)
is
partially molten either during accretion or
by
subsequent
whole-
mantle
convection,
which brings the
entire
mantle across the
soli-
dus
at shallow depths. Large
-
ion lithophile
(LIL)
elements are
concentrated
in
the melt.
The
deep magma ocean
(2)
fractionates
into a
thin plagioclase
-
rich surface layer and deeper olivine
-
rich
and garnet
-
rich cumulate layers
(3).
Late
-
stage melts in the
eclo-
gite
-
rich cumulate
are
removed
(4)
to
form
the continental crust
(C.C.),
enrich the peridotite layer
and
deplete
MORBS,
the
source region
of
oceanic crust
(O.C.)
and lower oceanic litho
-
sphere. Partial melting
of
the plume
source
(5)
generates conti
-
nental flood basalts (CFB), ocean
-
island basalts (IOB) and other
enriched
magmas,
leaving
a depleted residue (harzburgite) that
stays in
the
upper mantle.
The
MORB source may be the main
basalt reservoir
in
the
mantle,
and enriched or hot
-
spot magmas
(OIB,
CFB) may be MORB contaminated
by
interaction with
shallow mantle.
that formed
and
differentiated
the
Earth
were
uniform
and
ask
what the
resulting Earth
would look
like.
For
example,
the crust is
made
of
the lighter
and
more
fusible materials
and
was
formed
as
a result of
igneous
differentiation involv
-
ing
the
upward
migration
of
light melts. The core is
an
iron
-
rich
alloy
that also
melts
at
low
temperature
but
drains to
the interior because
of
its high density.
This
suggests
a
simple hypothesis: The stratification
of
the
Earth is a
result
of
gravitational separation
of
materials
according
to their
melting
points
and
density.
The materials accreting to form
the Earth
may
have
been
uniform,
but
the
high
temperatures
associated
with
accretion,
even
a violent
accretion,
would
result
in
a chemically differentiated planet. Thus, a simple,
even
obvious, process
gives
a complex result.
The
mantle
itself
is neither expected
to be
nor
observed
to
be
homogeneous.
Let
us
follow this
single, simple
hy
-
pothesis further.
As
the
Earth grows,
the
crustal elements
are continuously concentrated into the
melts
and
rise
to the
surface.
When
these
melts
freeze,
they
form the crustal
minerals that are rich
in
silicon, calcium, aluminum, potas
-
sium
and
the large
-
ion lithophile (LIL) elements.
Melts
generally are also rich in
FeO
compared
to primitive mate
-
rial.
This plus the high
compressibility
of
melts means
that
the densities
of
melts and
residual crystals converge, or
even
cross, as
the
pressure increases. Melt separation is
therefore
more difficult
at depth,
and melts
may
even
drain
downward
at very
high
pressure.
However,
during accretion
the
majority
of
the
melt
-
crystal separation occurs at
low
pressure.
All
of
the material in the deep interior
has
passed
through this
low
-
pressure
melting stage
in
a sort
of
continu
-
ous zone
refining.
The magnesium
-
rich minerals,
Mg2Si04
and
MgSiO,,
have high
melting temperatures
and
are
fed
through
the
melting zone
into the interior.
Even
if the
ac-
creting material
is
completely
melted
during assembly
of
the
Earth, these
minerals will be the
first
to freeze,
and they
will
still separate from the
remaining
melt.
(Mg,Fe),SiO,
has
a slightly higher
melting
or freezing point than
(Mg,Fe)SiO,
and
a slightly
higher density, so that olivine
may
even
separate
from
orthopyroxene. This separation,
however,
is expected to
be
much
less
effective than the
separation
of
olivine
and
orthopyroxene from
the melts rich
in
SiO,,
A1203,
CaO,
Na20
and
K20
which
differ substan
-
tially from
these minerals
in
both
density
and melting
point.
The
downward
separation
of
iron
-
rich melts,
along with
nickel, cobalt, sulfur
and
the trace siderophile elements,
strips these elements out
of
the crust
and
mantle. The
A120,,
CaO
and
Na20
content in chondritic
and
solar
ma
-
terial
is adequate
to
form a
crust
some 200
km
thick. The
absence
of
such
a massive
crust
on
the Earth might suggest
that the Earth
has
not
experienced a very
efficient
differen
-
tiation.
On
the other hand, the size
of
the core
and
the ex
-
treme concentration
of
the large
-
ion, magmaphile elements
into
the crust suggest
that
differentiation
has been
extremely
efficient.
The solution
to this
apparent
paradox
does not require
special pleading.
At
pressures
corresponding to depths
of
the order
of
50
km, the low
-
density minerals
of
the crust
convert
to
a mineral
assemblage denser
than
olivine
and
orthopyroxene.
Most
of
the
original crust therefore
is
unstable
and
sinks into the mantle.
Any melts below
200
-
400
km
may
suffer
the same fate.
Between
50
and
500 km
the
A1203-CaO-Na20-rich
materials crystallize
as
clinopyroxene
and
garnet, a dense eclogite assemblage that
is
denser
than
peridotite. Eclogite transforms
to
a garnet
solid solution,
which
is still denser
and which
is stable
be
-
tween about
500
and
800
km.
Peridotite also undergoes a
series
of
phase
changes that
prevent
eclogite from sinking
deeper
than about
650 km.
However, when
the Earth
was
about
Mars' size
and
smaller, the eclogite could sink to the
core
-
mantle
boundary.
The
base
of
the
present mantle, a
region
called
D
"
,
is
anomalous
and
may
be
composed
of
high
-
pressure eclogite.
The earliest stages
of
accretion
probably involved
the
most
refractory materials
under nebular
conditions. These in
-
clude compounds
rich
in Fe and
CaO,
A1203
and
TiO,.
The
refractory lithophiles
would
have been
excluded from a
molten
Fe
-
rich
core,
and
these
may
also
be
concentrated
in
D
"
. In either case
D
"
would
be
enriched
in
A120,
and
CaO.
Equilibration
between
material
in
D
"
and
the core
may
also
result
in
a high
FeO
content for
this
region.
If D
"
is intrin
-
sically denser
than
the rest
of
the
lower mantle,
it would be
gravitationally stable at the
base
of
the mantle.
On
the other
hand, it is embedded in the thermal boundary
layer between
mantle
and
core
and
therefore has a high temperature that
may
locally permit
D
"
material to rise into the lower mantle
until
it becomes neutrally buoyant.
As
it cools, it
will
sink
back to
D".
The
end
result for a
planet
experiencing partial melt
-
ing, gravitational separation
and
phase changes is chemical
stratification. The possibility
of
three
"
basaltic
"
regions
(high
CaO
and
A1,0,
and
possibly
FeO,
relative
to
MgO)
has been
identified. These regions are the crust, the transi
-
tion
region (between
upper
and lower mantle) and D
"
. The
latter
two
may
be the
result
of
solid subduction
or
sinking
of
high
-
density melts.
The
bulk
of
the upper mantle
and
lower mantle
may
therefore
be sandwiched between
basalt
-
rich layers.
If melt
-
ing during accretion extended
below
some 300
km,
the
composition
of
the
melt
and
residual refractory phases
changes. Orthopyroxene transforms to majorite, a
garnet-
like phase that replaces olivine on the liquidus. Melts,
therefore, are
MgO
rich,
and
we
have
another mechanism
for
separating major elements in the mantle
and
concentrat
-
ing olivine
(Mg2Si04)
in
the
shallow
mantle. Giant impacts
in
early Earth history
have
the potential for melting mantle
to great depth
and
for concentrating dense refractory re
-
sidual phases
such
as
orthopyroxene
-
majorite in the lower
mantle.
The subsequent cooling
and
crystallization
of
the Earth
introduces additional complications. A chemically stratified
mantle cools
more slowly than
a homogeneous Earth. Phase
change boundaries are
both
temperature
and
pressure depen
-
dent,
and
these migrate
as
the Earth cools. A thick basalt
crust, stable
at high
temperature, converts to eclogite
at its
base
as it cools through the basalt
-
eclogite phase boundary.
The initial crust
of
the Earth,
or
at least
its
deeper portions,
therefore can
become
unstable
and
plunge into the mantle.
This is
an
effective
way
to cool the mantle
and
to displace
lighter
and
hotter material
to
the
shallow
mantle where
it
can
melt
by
pressure release, providing a continuous
mecha
-
nism
for bringing
melts
to the surface.
The separation
of
melts and
crystals is a process
of
differentiation. Convection
is
often thought
of
as
a ho
-
mogenization process, tantamount to stirring. Differenti
-
ation,
however, can be
irreversible. Melts that are sepa
-
rated from the
mantle when
the Earth
was
smaller or from
the present
upper
mantle crystallize
to
assemblages
that
have
different phase relations
than
the residual crystals or
original mantle material.
If
these rocks are returned to the
mantle,
they will not
in general
have
neutral
buoyancy,
nor
are
they
necessarily denser
than
"
normal
"
mantle
at all
depths. Eclogite, for example, is denser
than
peridotite
when
the latter is
in
the olivine,
P-spinel
and
y
-
spinel
fields
but is less dense
than the
lower mantle,
and
it transforms
to dense perovskite
-
bearing assemblages at higher pressure
than
peridotite.
ORIGIN
OF
THE
CRUST
Although the crust represents less
than
0.5
percent
of
the
mass
of
the Earth,
it contains a large fraction
of
the ele
-
ments
that
preferentially enter the
melt when
a silicate
is
melted
-
the
large
-
ion lithophile elements (LIL). For ex
-
ample,
it can
be
estimated
that
the continental crust contains
58 percent
of
the rubidium, 53 percent
of
the cesium, 46
percent
of
the potassium, 37 percent
of
the barium
and
35
percent
of
the uranium
and
thorium
in
the crust
-
mantle sys
-
tem. Other highly concentrated elements include
bismuth
(34 percent),
lead
(32 percent), tantalum (30 percent), chlo
-
rine (including
that
in seawater,
26
percent), lanthanum
(19
percent), and strontium
(13
percent). These
high
concentra
-
tions in
such
a small volume mean that
the Earth is an ex
-
tensively differentiated
body. Apparently, most or
all
of
the
mantle has
been processed
by
partial melting and
upward
melt extraction.
It has also
been
estimated
that
77 percent
of
the argon
-
40
produced
by
the
decay
of
potassium
-
40
in
the mantle
and
crust resides
in
the
atmosphere. This also
points
toward
a well
-
differentiated,
and
outgassed, Earth. It
is
possible
that most of the
H20
is
in
the
ocean and
the
crust.
The
age
of
the
oldest continental crust is
at
least
3.8
x
lo9
years,
and
perhaps
as
old
as
4.2
X
lo9
years;
isotopic results from mantle rocks are also consistent
with
ancient differentiation.
Yet
the
mean
age
of
the
continental
crust is
1.5
x
lo9
years
(Jacobsen
and
Wasserberg, 1979).
The energy
of
accretion
of
the Earth
is great enough to melt
a large fraction
of
the incoming material, therefore the pro
-
cesses
of
melting, crust formation
and
outgassing
were
probably contemporaneous
with
accretion. The absence
of
older crust
may reflect
extensive bombardment
in
the later
stages
of
accretion
and
a high
-
temperature crust
and upper
mantle rather
than
a late onset
of
the crust
-
forming process.
A stable crust
may
have been delayed by
the freezing
of
a
deep
magma
ocean.
Most of the continental crust formed
between
2.5
and
3 Ga ago.
In more recent times a small amount
of
continen
-
tal crust
has
been
added
by
accretion
of
island arcs, oceanic
islands
and
plateaus
and
by
continental
flood
basalts. Some
crustal material is eroded, subducted
and
recycled into the
mantle, but the
net effect
is nearly
constant crustal
volume
over
the past 1
-
2 Ga.
Estimates
of
growth
rates
of
the
continental crust are
0.1
-
0.5
km3/yr
at present,
0.3-
0.7
krn
3
/yr
over
the
past
2.5
Ga
and
5.7
-
6.6
km
3
/yr
at the
end
of
the
Archean,
or approximately
ten
times the present
rate
(McLennan and
Taylor,
1980).
Isotopic
ages of
the crust suggest that crustal growth
has been
episodic
with
major
additions
in
the time intervals
3.8
-
3.5, 2.8
-
2.5, 1.9
-
1.6, 1.6
-
1.2,
1.2
-
0.9
and0.5
-
0
Ga.
The
composition
of
the crust
has
varied
with
time,
most
abruptly at the Archean
-
Proterozoic
boundary. In
particular
there are decreases in chromium
and nickel
and increases in
REE,
thorium, uranium,
Th/U
and
87Sr/86Sr
(see
Weaver
and
Tarney,
1984,
and
McLennan
and Taylor,
1980, for
reviews).
The
formation
of
the original crust is
probably
linked
to the
formation
and
evolution
of
a magma
ocean.
However,
it is harder to form a crust
on
a
magma ocean than
on a
watery ocean because a silicate crust is
probably denser
than
the liquid
it freezes from. Feldspar
crystals
of
sufficient
size might
be
able
to break away
from
convection and
float,
but
they
only form after extensive freezing
and
grow
only
slowly.
The
density contrast
is
also relatively
low.
Gas-
crystal packets,
or
foams,
may
provide sufficient buoyancy
even though
the individual crystals are small.
The
intersti
-
tial
fluids
formed
in
deep olivine
-
orthopyroxene cumulates
may
be
the source
of
the
protocrust
-
the
crust
may
have
formed
by
the freezing
of
light
fluids
that
rose to
the surface
of
the magma ocean. Certainly, the crust
is extremely en
-
riched
in
the incompatible elements,
and
forming it from a
liquid from
which
the refractory crystals
have been
re
-
moved
is one
way
to account for this enrichment.
The
major additions to the continental crust in
recent
times are due to the lateral accretion
of
island arcs. The
average composition
of
the continental crust
may
therefore
be
similar
to the average
island
-
arc andesite
(Taylor
and
McLennan, 1981). The andesite
model can
be used
to
esti
-
mate the composition
of
the
lower
crust
if it is
assumed
that
andesite
melts
during orogenesis to form a granodioritic
up
-
per
crust
and
a refractory, residual lower crust. The total
average crust, in these models, is
average
island
-
arc
andes-
ite,
and the upper
-
crust composition
is approximated
by
the
average post
-
Archean sediment. There
is
some
evidence,
however,
that island arcs are more
basic
than
andesite.
Seis
-
mic
data also
suggest
that the
average
crust
is
closer
to
quartz diorite
and
granodiorite
than to
andesite; that
is,
it is
more
siliceous. Heat
flow
suggests that the
lower
crust
is
rather depleted
in
the
radioactive heat
-
producing elements
(K,
U,
Th).
An
alternative
model
for continental crust
chemistry
was
derived
by
Weaver
and
Tarney
(1984)
using
a variety
of
xenolith
and
geophysical
data.
Both models
are
listed
in Table
2
-
3.
ORIGIN OF
THE
MANTLE
AND CORE
Larimer
(1967)
has
outlined the condensation history
of
a
cooling gas
of
cosmic composition. Compounds
such
as
CaTiO,,
MgA120,,
Al,SiO,,
and
CaAl,Si,O,
condense
first
at temperatures
between 1740
and
1620
K
(see Figure 1
-
6).
Iron condenses next at 1620
K.
Magnesium
-
rich
pyroxenes
and
olivines
condense between 1740
and
1420
K;
FeS
con
-
denses
at
680
K
and
H,O
at 210
K.
All
the
above
tempera
-
tures were calculated
on
the assumption
of
a total pressure
of
6.6
X
atmosphere.
Larimer
and Anders
(1970)
concluded
that
the fractionation patterns in meteorites
oc
-
curred
in
the solar
nebula
as it
cooled from high tempera
-
tures
and could not
be
produced
in
the meteorite parent