Chapter
I 0
The
bowels
of
the
Earth
The
lower
mantle
I
must
be
getting
somewhere
near
the
centre
of
the
earth.
Let
me
see:
that
would
be
four
thousand
miles
down.
I
think-
Alice
The
traditional
lower
mantle
starts
near
800-
1000
km
where
the
radial
gradient
of
the
seismic
velocities
becomes
small
and
smooth
.
This
is
Bullen's
Region
D.
The
1000-km
depth
region
appears
to
be
a
fundamental
geodynamic
inter-
face,
perhaps
a
major-element
chemical
and
a
viscosity
interface.
Some
authors
take
the
lower
mantle
to
start
just
below
the
major
mantle
dis-
continuity
near
650
km.
The
depth
of
th
is dis-
continuity
varies,
perhaps
by
as
much
as
40
km
and
is
variously
referred
to
as
the
'660
km
dis-
continuity'
or
'670
km
discontinuity';
the
average
depth
is
650
km.
In
detailed
Earth
models
there
is
a
region
of
high
velocity
gradient
for
another
50-100
lan
below
the
discontinuity
.
This
is
prob-
ably
due
to
phase
changes,
but
it
could
represent
a
chemical
gradient.
Plate
reconstructions
show
that
past
subduction
zones
correlate
with
high-
velocity
regions
of
the
mantle
near
800-1000
km
depth.
The
'lower
mantle
proper'
therefore
does
not
start
until
a
depth
well
below
the
650
Ian
boundary
,
more
in
agreement
with
the
classical
definition.
Below
this
depth
the
lower
mantle
is
relatively
homogenous
until
about
300
km
above
the
core-mantle
boundary
.
If
there
are
chemical
discontinuities
in
the
mantle
the
boundaries
will
not
be
at
fixed
depths
.
These
clarifications
are
needed
because
of
the
controversy
about
whether
slabs
penetrate
into
the
lower
mantle
or
whether
they
just
push
down
a
discontinuity,
and
where
the
boundary
of
the
lower
mantle
really
is.
The
midmant!e,
or
mesosphere,
extends
from
1000
to
2000
km.
This
is
the
blandest
part
of
the
mantle.
Estimates
of
its
composition
range
from
pure
MgSi0
3
perovskite
to
chondritic
-
in
major
refractory
elements
-
to
pyrolite,
within
the
uncertainties.
It
can
be
Si-
and
Fe-rich
com-
pared
with
the
upper
mantle,
or
identical
to
the
upper
mantle.
Geochemical
models
assume
that
it
is
undegassed
and
unfractionated
cosmic
material,
but
there
is
no
support
for
this
con-
jecture
.
The
lower
mantle
must
be
quite
differ-
ent
from
that
which
appears
in
the
standard
model
of
mantle
geochemistry,
and
the
geodynamic
models
that
are
based
on
it.
Composition
of
the
lower
mantle
The
probable
mineralogy
of
the
deep
mantle
is
known
from
high-pressure
mineral
physics
-
squeezing
-
experiments
.
The
mineralogy
is
sim-
ple
in
comparison
to
that
of
the
crust
and
upper
mantle,
consisting
of
(Mg,Fe)Si0
3
-A}z0
3
orthorhombic
perovskite,
CaSi0
3
cubic
perovskite,
and
(Mg,Fe)O
magnesiowi.istite
. No
Al-rich
phases,
except
for
perovskites,
are
considered
to
exist
in
the
main
part
of
lower
mantle
.
The
total
basaltic
content
of
the
mantle
is
only
about
6 %
and
most
of
this
is
probably
in
the
upper
mantle
and
tran-
sition
region.
If
most
of
the
crustal
and
basaltic
materi
al
has
been
sweated
out
of
the
lower
mantle
,
then
it
will
be
low
in
Ca
. Al
,
U,
Th
and
K.
amon
g
many
other
things
.
Th
e
lower
mantle
would
then
be
mainly
oxides
of
Si,
Mg
and
Fe.
The
uncertain
spin-state
and
oxidation-state
of
Fe
introduces
a
bit
of
spice
into
lower
mantle
min-
eralogy
.
The
composition
of
the
lower
mantle
is
another
s
tor
y.
Most
plausible
compositions
have
similar
properties
.
Candidates
for
the
dominant
rock
types
in
various
deep-mantle
layers
are
so
sim-
ilar
in
seismic
properties
that
standard
methods
of
se
ismic
petrology
fail.
Small
differences
in
den-
sity,
however
,
can
irreversibly
stratify
the
mantle
so
it
is
methods
based
on
density,
impedance,
anisotropy,
dynamic
topo
g raphy,
pattern
recog-
nition,
scattering
and
convective
style
that
must
be
used,
in
addition
to
seismic
velocity.
Visual
inspection
of
color
tomographic
cross
-
sections
cannot
reveal
subtle
chemical
contrasts.
Several
methods
have
been
used
to
estimate
the
composition
of
the
lower
mantle
from
seis-
mic
data
but
they
are
all
non-unique
and
require
assumptions
about
temperature
gradients,
tem-
perature
and
pressure
derivatives
,
equations
of
s
tate
and
homogeneity
.
Perhaps
the
most
direct
method
is
to
compare
shock-wave
densities
at
hi
gh
pressure
of
various
silicates
and
oxides
with
seismically
determined
densities
.
There
is a
trade-
off
betw
e
en
temperature
and
composition
, so
this
exercise
is
non-unique.
Materials
of
quite
differ-
ent
compositions
, say
(M
g .Fe)Si0
3
(p
erovskite)
and
(Mg
,Fe)O
,
can
have
identical
densities.
and
mix-
tures
involving
different
proportions
ofMgO,
FeO
and
Si0
2
can
satisfy
the
density
constraints.
In
addition
,
the
density
in
the
Earth
is
not
as
well
determined
as
such
parameters
as
the
compres-
sional
and
shear
velocities.
The
mineralogy
and
composition
of
the
lower
mantle
are
hard
to
determine
since
plausible
combinations
of
per-
ovsldte
and
magnesiowiistite
ranging
from
chon-
dritic
to
pyrolite
have
similar
elastic
properties
when
FeO
and
temperature
are
taken
as
free
parameters
.
But
they
can
differ
enough
in
den-
sity
to
allow
chemical
stratification
that
is
stable
against
overturn.
Oxide
mixtures,
such
as
MgO
+
Si0
2
(stishovite)
.
can
have
densities
,
at
high
pres-
sure,
similar
to
compounds
such
as
perovskite
having
the
same
stoichiometry.
C
OMP
OSITION
OF
THE
LOWER
MANTLE
117
It
can
be
shown
that
a
chondri
tic
composition
for
the
lower
mantle
gives
satisfactory
agreem
e
nt
between
shockwave,
equation
of
state
and
seis-
mic
data,
for
the
most
plausible
lower
mantle
temperature.
The
Si0
2
content
of
the
lower
man-
tle
may
be
closer
to
chondritic
than
pyrolitic.
If
the
lower
mantle
falls
on
or
above
the
1400
oc
adiabat,
then
chondritic
or
pyroxenitic
compo-
sitions
are
preferred.
If
temperatures
are
below
the
1200
oc
adiabat
,
then
more
olivine
(p
erovsk
ite
plus
(MgFe)O)
can
be
accommodated
. A
variety
of
evidence
su
g
gests
that
the
higher
tempera-
tures
are
more
appropriate
.
The
temperature
gra-
dient
in
the
lower
mantle
can
be
subadiabatic
or
superadiabatic
.
Attempts
to
estimate
composi-
tion
assume
chemical
and
mineralo
g ical
homo-
geneity
and
adiabaticity
but
the
problem
is
still
indeterminate.
A
variety
of
chemical
models
can
be
made
consistent
with
the
geophysical
data
but
the
actual
chemical
composition
of
the
lower
mantle
is
unknown
,
except
within
very
broad
limits
.
Equation-of-state
modeling
is
much
too
blunt
a
tool
to
'prove'
that
the
lower
mantle
has
the
same
,
or
different
,
chemistr
y
as
the
upper
mantle
.
Internal
chemical
boundaries
in
the
mant
le ,
in
contrast
to
phase
boundaries,
and
the
sur-
face,
Moho
and
core-mantle
boundaries,
mu
s
t-
exhibit
enormous
variations
in
depth,
because
of
the
low
density
contrast.
This
plus
the
low
predicted
seismic
impedance
means
that
compo-
sitional
boundaries
are
difficult
to
detect
,
even
if
they
are
unbreachable
by
mantle
convection
.
They
are
s
te
alth
boundarie
s.
Low
-s pin F
e
2
+
Fe
u n d
ergoes
a s
pin
-tra
n
sitio
n
at
hig
h
press
u
re
with
a
large
reduction
in
ionic
radius
and
a
probable
increase
in
the
bulk
modulus
and
seismic
velocities
.
The
transition
may
be
spread
out
over
a
large
depth
interval
The
major
minerals
in
the
deep
mantle
are
predicted
to
be
almost
Fe-free
perovskite
[M
g
Si0
3
]
and
P
e-
rich
magnesiowiistite,
(Mg
,Fe)O.
This
has
several
important
geodynamic
implications
.
Over
tim
e ,
the
dense
FeO-rich
material
may
accumulat
e ,
irreversibly,
at
the
base
of
the
mantle
, a
nd
,
in
addition,
may
interact
with
the
core
.
The
lattice
conductivity
of
this
iron-rich
layer
will
be
high
118
THE
BOWELS
OF
THE
EARTH
and
the
radiative
term
should
be
low.
A
thin
layer
convects
sluggishly
(because
of
the
h
3
term
in
the
Rayleigh
number)
but
its
presence
slows
down
the
cooling
of
the
mantle
and
the
core.TI1e
overlying
FeO-poor
layer
may
have
high
radiative
conductivity,
because
of
high
T
and
transparency,
and
have
high
viscosity
and
low
thermal
expan-
sivity,
because
of
P
effects
on
volume.
TI1is
part
of
the
mantle
will
also
convect
sluggishly.
If
it
repre-
sents
about
one-third
of
the
mantle
(by
depth)
it
will
have
a Rayleigh
number
about
30
times
less
than
Rayleigh
numbers
based
on
whole
mantle
convection
and
orders
of
magnitude
less
than
Ra
based
on
P
=
0
properties.
It
is
likely
than
some
of
the
Fe
in
the
lower
mantle
is
low-spin
and
some
is
high-spin
with
the
proportions
changing
with
depth.
TI1e
oxi-
dation
state
of
Fe
in
the
lower
mantle
is
also
likely
to
be
different
than
in
the
shallow
man-
tle
. TI1ese
considerations
complicate
the
inter-
pretation
of
lower
mantle
properties
and
the
geodynamics
and
melting
point
of
the
deep
man-
tle.
It
is
certainly
dangerous
to
fit
a
single
equa-
tion
of
state
to
the
whole
lower
mantle
or
to
argue
that
seismic
data
requires
the
lower
man-
tle
to
be
homogenous,
or
the
same
as
the
upper
mantle.
Reg
io n D
"
The
lowermost
mantle,
Bullen's
Region
D" is
a
region
of
generally
low
seismic
gradient
and
increased
scatter
in
seismic
travel
times
and
amplitudes.
Lay
and
Heimberger
(1983)
found
a
shear-velocity
jump
of
2.8%
in
this
region
that
may
vary
in
depth
by
up
to
40
km.
They
con-
cluded
that
a
large
shear-velocity
discontinuity
exists
about
280
km
above
the
core,
in
a
region
of
otherwise
low
velocity
gradient.
There
appears
to
be
a
lateral
variation
in
the
velocity
increase
and
sharpness
of
the
structure,
but
the
basic
character
of
the
discontinuity
seems
to
be
well
established.
Because
the
core
is
a
good
conductor
and
has
low
viscosity,
it
is
nearly
isothermal.
Lateral
temperature
variations
can
be
maintained
in
the
mantle,
but
they
must
converge
at
the
base
ofD
" .
This
means
that
temperature
gradients
are
vari-
able
in
D" .
In
some
places,
in
hotter
mantle,
the
gradient
may
even
be
negative.
Regions
of
neg-
ative
shear
velocity
gradient
in
D"
are
probably
regions
of
high
temperature
gradient
and
high
heat
loss
from
the
core.
It
is
plausible
that
the
layers
at
the
base
of
the
mantle
interact
with
the
core
and
therefore
differ
in
composition
from
the
rest
of
the
mantle
.
Lateral
heterogeneity
D"
may
represent
a
chemically
distinct
region
of
the
mantle.
If
so
it
will
vary
laterally,
and
the
dis-
continuity
in
D"
will
vary
considerably
in
radius,
the
hot
regions
being
elevated
with
respect
to
the
cold
regions.
A
chemically
distinct
layer
at
the
base
of
the
mantle
that
is
only
marginally
denser
than
the
overlying
mantle
would
be
able
to
rise
into
the
lower
mantle
when
it
is
hot
and
sink
back
when
it
cools
off.
The
mantle-core
bound-
ary,
being
a
chemical
interface,
is a
region
of
high
thermal
gradient,
at
least
in
the
colder
parts
of
the
lower
mantle.
Seismic
observations
suggest
the
presence
of
broad
seismic
velocity
anomalies
in
the
deep
mantle.
The
nature
of
these
anomalies
is
incon-
sistent
with
purely
thermal
convection
and
suggests
the
existence
of
large-scale
chemical
heterogeneities
in
the
lower
mantle
.
The
anti-
correlation
between
bulk
sound
speed
and
shear
wave
velocity
anomalies
in
the
lowermost
mantle
suggests
the
presence
of
chemical
density
heterogeneities.
The
co
re
In
the
first
place
please
bear
in
mind
that
I
do
not
expect
you
to
believe
this
story.
Nor
could
you
wonder
had
you
witnessed
a
recent
experience
of
mine
when
,
in
the
armor
of
blissful
and
stupendous
ignorance,
I
gaily
narrated
the
gist
of
it
to
a Fellow
of
the
Royal
Geological
Society
....
The
erudite
gentleman
in
whom
I
confided
congealed
before
I
was
half
through!-
it
is
all
that
saved
him
from
exploding-
and
my
dreams
of
an
Honorary
Fellowship,
gold
medals,
and
a
niche
in
the
Hall
of
Fame
faded
into
the
thin
, cold
air
of
his
arctic
atmosphere.
But
I believe
the
story
,
and
so
would
you,
and
so
would
the
learned
Fellow
of
the
Royal
Geological
Society,
had
you
and
he
heard
it
from
the
lips
of
the
man
who
told
it
to
me
Edgar
Rice
Burroughs
A
molten
iron-rich
core
appeared
early
in
Earth
history,
the
evidence
being
in
the
remnant
mag-
netic
field
and
isotopic
record
of
ancient
rocks.
This
in
turn
implies
a
short
high
temperature
accretion
for
the
bulk
of
the
Earth,
with
per-
haps
a
drawn
out
accretionary
tail
to
bring
in
noble
gases
and
other
volatile
elements
and
to
salt
the
upper
mantle
with
siderophile
ele-
ments
that
would
otherwise
be
in
the
core.
The
long-standing
controversy
regarding
a
drawn-out
(100
milllion
years)
versus
a
rapid
(~1
Myr)
ter-
restrial
accretion
appears
to
be
resolving
itself
in
favor
of
the
shorter
time
scales
and
a
high-
temperature
origin.
The
core
is
approximately
half
the
radius
of
the
Earth
and
is
about
twice
as
dense
as
the
mantle.
It
represents
32%
of
the
mass
of
the
Earth.
A
large
dense
core
can
be
inferred
from
the
mean
density
and
moment
of
inertia
of
the
Earth,
and
this
calculation
was
per-
formed
by
Emil
Wiechert
in
1891.
The
exis-
tence
of
stony
meteorites
and
iron
meteorites
had
earlier
led
to
the
suggestion
that
the
Earth
may
have
an
iron
core
surrounded
by
a
silicate
man-
tle.
The
first
seismic
evidence
for
the
existence
of
a
core
was
pres
en
ted
in
19
0
6
by
01
dharn,
although
it
was
some
time
before
it
was
real-
ized
that
the
core
does
not
transmit
shear
waves
and
is
therefore
probably
a
fluid.
It
was
recog-
nized
that
the
velocity
of
compressional
waves
dropped
considerably
at
the
core-mantle
bound-
ary.
Beno
Gutenberg
made
the
first
accurate
determination
of
the
depth
of
the
core,
2900
km,
in
1912,
and
this
is
remarkably
close
to
current
values.
The
core-mantle-boundary
is
referred
to
as
the
Gutenberg
discontinuity
and
as
the
CMB.
Although
the
idea
that
the
westward
drift
of
the
magnetic
field
might
be
due
to
a
liq-
uid
core
goes
back
300
years,
the
fluidity
of
the
core
was
not
established
until
19
2 6
when
J
e
f-
freys
pointed
out
that
tidal
yielding
required
a
smaller
rigidity
for
the
Earth
as
a
whole
than
indicated
by
seismic
waves
for
the
mantle.
It
was
soon
agreed
by
most
that
the
transition
from
mantle
to
core
involves
both
a
change
in
com-
position
and
a
change
in
state.
Subsequent
work
has
shown
that
the
boundary
is
extremely
sharp.
There
is
some
evidence
for
variability
in
depth,
in
addition
to
hydrostatic
ellipticity.
Variations
in
14
13
"'E
~
.i!'
12
'iii
c:
Q)
0
11
10
THE
CORE
119
Pressure
(M
bar)
•~11.11•
Estimated
densities
of
Fe,
Ni
and
some
Fe-rich
alloys
compared
with
core
densities.
The
estimated
reduction
in
density
due
to
melting
is
shown
(dashed
line)
for
one
of
the
alloys.
11
~
E
~
10
.i!'
'(}
0
~
9
Pressure
(M
bar
)
Compressional
wave
velocities
in
the
outer
core
and
compressional
and
bulk
sound
speeds
in
th
e
inner
core
compared
to
estimates
for
iron
and
nickel.
Values
are
shown
for
Pois
son
ratios
in
the
inner
core.
lower-mantle
density
and
convection
in
the
lower
mantle
can
cause
at
least
several
kilometers
of
relief
on
the
core
-m
antle
boundary.
The
outer
core
has
extremely
high
Qand
transmits
P-waves
with
very
low
attenuation.
The
elastic
properties
and
density
of
the
core
are
consistent
with
an
iron-rich
alloy
(Figures
10
.1
and
10.2).
Evidence
that
the
outer
core
is
mainly
an
iron-rich
fluid
also
comes
from
the
magnetohydrodynamic
120
THE
BOWELS
OF
THE
EARTH
requirement
that
the
core
be
a
good
electrical
conductor.
Although
the
outer
core
behaves
as
a fluid,
it
does
not
necessarily
follow
that
temperatures
are
above
the
liquidus
throughout.
It
would
behave
as
a
fluid
even
if
it
contained
30%
or
more
of
sus-
pended
particles.
All
we
know
for
sure
is
that
at
least
part
of
the
outer
core
is
above
the
solidus
or
eutectic
temperature
and
that
the
outer
core,
on
average,
has
a
very
low
rigidity
and
low
viscos-
ity.
Because
of
the
effect
of
pressure
on
the
liq-
uidus
temperature,
a
homogenous
core
can
only
be
adiabatic
if
it
is
above
the
liquidus
through-
out.
An
initially
homogenous
core
with
an
adi-
abatic
temperature
profile
that
lies
between
the
solidus
and
liquidus
will
contain
suspended
par-
ticles
that
will
tend
to
rise
or
sink,
depending
on
their
density.
The
resulting
core
will
be
on
the
liq-
uidus
throughout
and
will
have
a
radial
gradient
in
iron
content.
The
core
will
be
stably
stratified
if
the
iron
content
increases
with
depth.
The
inner
core
In
193
6
Inge
Lehmann
used
seismic
data
from
the
core
shadow
to
infer
the
presence
of
a
higher
velocity
inner
core.
Although
no
waves
have
yet
been
identified
that
have
traversed
the
inner
core
unambiguously
as
shear
waves,
indirect
evi-
dence
indicates
that
the
inner
core
is
solid
(Birch,
1952)
.
Julian
and
others
(1972)
reported
evidence
for
PKJKP,
a
compressional
wave
in
the
man-
tle
and
outer
core
that
traverses
the
inner
core
as
a
shear
wave.
Observation
of
PKJKP
is
difficult
and
claimed
observations
are
contro-
versial.
Early
free-oscillation
models
gave
very
low
shear
velocities
for
the
inner
core,
2
to
3 kmfs;
more
recent
models
give
shear
velocities
in
the
inner
core
ranging
from
3.46
to
3.7
kmfs,
in
the
range
of
crustal
values.
The
boundary
of
the
inner
core
is
also
extremely
sharp.
The
Q
of
the
inner
core
is
relatively
low
,
and
appears
to
increase
with
depth.
The
high
Poissons
ratio
of
the
inner
core,
0.44,
has
been
used
to
argue
that
it
is
not
a
crystalline
so
lid
,
or
that
it
is
near
the
melting
point
or
partially
molten
or
that
it
involves
an
electronic
phase
change
. However,
Poissons
ratio
increases
with
both
temperature
and
pressure
and
is
expected
to
be
high
at
inner
core
pressures
,
particularly
if
it
is
metallic
.
Some
metals
have
Poisson's
ratios
of
0.43
to
0.46
even
under
laboratory
conditions.
The
solid
inner
core
is
the
most
remote
and
enigmatic
part
of
our
planet,
and
except
for
the
crust,
is
the
smallest
'official'
subdivision
of
Earth's
interior.
Only
a
few
seismic
waves
ever
reach
it
and
return
to
the
surface
.
The
inner
core
is
a
small
target
for
seismologists
and
seis-
mic
waves
are
distorted
by
passing
though
the
entire
Earth
before
reaching
it.
The
inner
core
is
isolated
from
the
rest
of
Earth
by
the
low
vis-
cosity
fluid
outer
core
and
it
can
rotate,
nod
,
wobble,
precess
,
oscillate,
and
even
flip
over,
only
loosely
constrained
by
the
surrounding
shells.
Its
existence,
size
and
properties
constrain
the
temperature
and
mineralogy
near
the
center
of
the
Earth.
Among
its
anomalous
characteristics
are
low
rigidity
and
viscosity
(compared
to
other
solids),
bulk
attenuation,
extreme
anisotropy
and
super-rotation
(or
deformation)
.
The
inner
core
has
a
radius
of
1222
km
and
a
density
about
13
gfcm
3
.
Because
of
its
small
size,
it
is
difficult
to
determine
a
more
accurate
value
for
density
. It
represents
about
1.7
%
of
the
mass
of
the
Earth.
The
density
and
velocity
jumps
at
the
inner-core-outer-core
boundary
are
lar
ge
enough,
and
the
boundary
is
sharp
enough,
so
that
the
boundary
is
a
good
reflector
of
short-
period
seismic
energy.
The
inner
core
is
seismi-
cally
anisotropic;
compressional
wave
speeds
are
3-4%
faster
along
the
Earth's
spin
axis.
The
main
constraint
on
composition
and
structure
is
the
compressional
velocity
and
anisotropy.
From
seismic
velocities
and
cosmic
abundances
we
know
that
it
is
mainly
com-
posed
of
iron-nickel
crystals,
and
the
crystals
must
exhibit
a
large
degree
of
common
orien-
tation
.
The
inner
core
is
predicted
to
have
very
high
thermal
and
electrical
conductivity,
a
non-
spherica
l
shape,
frequency-dependent
properties
and
it
may
be
partially
molten.
It
may
be
essen-
tial
for
the
existence
of
the
magnetic
field
and
for
polarity
reversals
of
this
field.
Freezing
of
the
inner
core
and
expulsion
of
imp
uri
ties
is
likely
responsible
for
powering
the
geodynamo.
Within
the
uncertainties
the
inner
core
may
be
simply
a
frozen
version
of
the
outer
core,
Fe
2
0
or
FeNiO,
pure
iron
or
an
iron-nickel
alloy.