INSIDE THE EARTH
A briefing document prepared for the Royal Society
and the Association of British Science Writers
by Dr Martin Redfern
June 1995
Introduction We have intimate knowledge of the surface of the Earth,
scientists have been to the Moon, robot craft have explored most of
the planets and astronomers can look at stars and galaxies directly
- yet geologists only have indirect evidence of what conditions are
like more than a few tens of kilometres beneath our feet. Boreholes
simply cannot penetrate more than about a dozen kilometres without
encountering rocks too hot to drill. The lavas spewed out of
volcanoes are almost certainly not representative of the bulk of the
Earth, although they sometimes contain lumps that may be.
Ironically, some of the most representative samples of rocks with
the likely composition of the deep interior of the Earth are
probably some of those found in meteorites. But, in the last few
years, there have been important developments in the scientific
understanding of the interior of the Earth which are now coming
together:
* Seismologists are using earthquake waves to build up
observations of properties of the Earth's mantle and core in a
process known as seismic tomography.
* Experimental scientists are simulating the high temperatures
and incredible pressures of the deep Earth in the laboratory and
observing the properties of rocks under such conditions.
* Theoreticians are building models of the Earth in their
computers and using calculations based on them to predict conditions
in the real Earth.
There have been significant advances in all three of these areas and
the results are now combining to suggest answers to important
questions about the dynamics of the Earth's interior, how the Earth
works and ultimately how it was put together in the first place.
The structure and composition of the Earth Our planet is rather like
an onion - a series of concentric shells or layers. Put very simply,
there is a thin crust of hard, cold rock - about seven kilometres
thick under the ocean and typically 35 km thick under continents.
The bulk of the Earth comprises a rocky mantle which, though solid,
is hot and slightly plastic. Beneath that, is a core of molten iron
with a small, solid inner core of iron, about the size of the Moon.
In fact, the Earth is only 99% like a perfectly concentric onion.
The 1% variation from that perfect structure - horizontal density
differences, temperature gradients, compositional variations -
represent the frontier of geophysics and the key to the dynamic
processes within our planet.
The crust and upper few kilometres of the mantle comprise the rocky
lithosphere which is divided into a series of plates. Some are as
big as continents, others, mere splinters. They are pushed and
pulled by movements in the semi-plastic mantle beneath. Where they
are pulled apart, oceans form; where they grate alongside one
another, there are regular earthquakes; where they collide with one
another, mountain ranges form to the accompaniment of both
earthquakes and volcanoes. These are forces that can destroy lives
and change the face of the planet, but these plates are but the scum
on the surface compared to the vast bulk of the mantle.
The mantle extends to a depth of about 2 900 km but there are
various layers within it that reflect earthquake waves. One, at
about 420 km, represents what is known as a phase change: below it
minerals with the same chemical composition are forced into a denser
crystal lattice structure by pressure. The commonest mineral in the
top of the mantle is a magnesium silicate called olivine. Below 420
km the structure readjusts to form spinel. There is another feature
about 670 km down. That marks the boundary between the upper and
lower mantle. Once again, there is a phase change across it, with
the spinel structure above and an even more tightly packed one
called perovskite below. (Together with olivine in the upper mantle
are various calcium-rich minerals called pyroxenes.) As olivine
changes into spinel, pyroxene gives way to garnet. But below 670 km,
both are replaced by perovskite-structured minerals. In addition,
there is magnesium iron oxide with a dense crystal structure similar
to that of rock salt. Magnesium silicate perovskite is probably the
most abundant mineral in the Earth. Its tight-packed structure makes
it of special interest to materials scientists, but at surface
pressures it's unstable and a struggle to make, even in milligram
quantities. It's ironic that no further beneath our feet than the
horizontal distance between London and Edinburgh, there must exist
millions of tons of the stuff.
The phase change at 670 km represents a more substantial barrier
than the one at 420 km. The rocks of the mantle are heated by
radioactive decay from within and by the core beneath. This causes
the rocks to circulate slowly or convect, a bit like very thick
porridge on the stove. Although the rocks are indeed solid, they
can still move in rather the same way that ice moves in a glacier.
Like porridge, circulation takes place with hot plumes rising and
colder material sinking. Where a mantle plume comes up underneath
the crust, it can form a so-called hot spot with prolific volcanoes
such as those of Hawaii. If the upwelling takes place along a line,
it can open a rift valley, as in East Africa, and ultimately an
ocean such as the Atlantic. The basalt that flows out of volcanoes
above mantle plumes probably only represents about 4% of the bulk
composition of the mantle - the fraction that melts completely as
the pressure drops. It carries with it clues to the depth from
whence it came in the form of traces of gases such as helium.
The newly-formed ocean crust moves away from the ocean ridge on both
sides. Eventually, it is dense enough to sink back down into the
mantle. Typically, it does so in an ocean trench and an arc of
island volcanoes springs up as volatile components including water
begin to assist the melting and return some of the material to the
surface. The path of the slab, as it continues, can be traced by the
epicentres of deep earthquakes. They map its descent down to 670 km
where it seems to grind to a halt, at least temporarily. One of the
major issues in modern geology is whether, after absorbing
sufficient heat to undergo a further phase change, it then continues
through the lower mantle so that the entire mantle takes part in the
circulation, or whether the mantle is like a great double boiler
with separate circulations in the upper and lower mantle and little
or no chemical mixing between them. Evidence from simulations is now
beginning to suggest a compromise. The phase change from spinel to
perovskite takes up a lot of heat, so the descending slab cannot
undergo the phase change and cross the boundary until it has had
time to warm up - probably millions of years. So it tends to spread
out into a sort of holding reservoir that forms a pronounced layer
in the seismic images. Eventually, when it is hot enough, computer
models predict that vast slabs can break through quite quickly, like
a slow motion avalanche through the lower mantle.
What was probably the most powerful earthquake since the 1964
Alaskan earthquake took place in June 1994 beneath Bolivia. Its
vibrations could be felt through cold, hard continental rocks as far
apart as Canada and West Africa. But it caused comparatively little
damage above the epicentre. That was because it was extremely deep -
about 640 km. That is so deep that, for a long time, geologists did
not believe earthquakes were possible so deep because the rocks were
too soft to crack. What they now believe may happen is that a phase
change suddenly runs through a whole layer of rock in a sort of
anti-crack. An alternative theory involves a zone of weakness,
introduced in the plate long before it has reached such depths. Even
so, the Bolivian quake was bigger and deeper than anyone had
expected. But it proved a marvellous tool for seismologists.
Just as medical radiographers analyse the variation in absorption of
x-rays coming from different directions through a patient's body to
build up a three dimensional picture of that person's insides, a
process called computerised tomography, so geologists use variations
in earthquake wave velocities for seismic tomography, studying the
insides of the Earth. Each time there is an earthquake, it sets the
planet ringing like a bell. But like a real bell, it is not a clear,
perfect tone. There are harmonics and even discords due to the
internal structure and flaws. Seismic tomographers attempt to
subtract the effects of the primary tone and harmonics from the
multiple signals they build up from many earthquakes so as to leave
the effects due to flaws in the Earth. These are principally density
differences due to the convection processes. The hot, soft
low-density rocks in a mantle plume will not transmit the seismic
waves as quickly as the cold, hard dense material in a descending
slab. The images that the seismic tomographers get still depend to
some extent on how well they estimate the simple tone and harmonics,
but at most levels in the mantle they are beginning to agree with
one another and produce a complex picture of the dynamic circulation
patterns. Since it was so deep, the Bolivian earthquake gave some
particularly clear signals that were comparatively unaffected by the
complexities of continents and mountains on the surface.
Simulating the centre of the Earth It may sound a simple matter to
take a sample of rock and see what happens when you subject it to
high temperatures and pressures. But the temperatures, and
particularly the pressures, of the interior of the Earth are so high
that it becomes very difficult indeed. The pressure at the boundary
between the core and the mantle is about 1.3 million times the
pressure on the Earth's surface, and that at the centre of the Earth
is 3.6 million times atmospheric pressure. The first experiments to
simulate pressures of the planet's interior were done using big
hydraulic presses. They had two great advantages: they could work on
samples of rock a centimetre or more across and they could maintain
conditions for hours or days at a time until equilibrium was
established. However, such presses cannot simulate more than a few
hundred kilometres depth and even for that, they push materials
technology to the limit in the design of sample holders that will
not themselves break, melt or interfere with the experiments.
Professor Thomas Ahrens has used a very different approach at the
California Institute of Technology. His basement laboratory here
boasts the fastest gun in the west: a giant two-stage gun that first
uses explosives to compress hydrogen to pressures of about a
thousand atmospheres and then allows that to break through a
diaphragm and accelerate a 15 gram (1 ounce) bullet to speeds of 7-8
km per second. That is about the speed of a spacecraft orbiting the
Earth. When the projectile hits a mineral specimen, it creates a
shock wave that briefly matches the pressures and temperatures of
the core of the Earth. But only for less than a millionth of a
second, so there is no chance to establish equilibrium and a host of
microscopes and detectors must analyse the flash of light in the
same way as astronomers analyse starlight - and without getting
blown up in the process.
The most popular technique today seems far simpler. One material
that will withstand the temperatures and pressures involved is
diamond. Unfortunately, the researchers cannot afford fist-sized
diamonds, so they have become adept at handling microscopic
specimens, typically weighing only a few millionths of a gram. They
are placed between the flattened points of two sparkling, gem
quality, diamonds to form what is called a diamond anvil. The
diamonds are so-called brilliant cuts, with the flattened faces at
their points perfectly parallel to concentrate the pressure so that
nothing more than a precision thumbscrew is needed to reach the
pressure of the centre of the Earth in the sample.
Conveniently, the properties of ruby change with increasing pressure
so a tiny speck of ruby next to the sample acts as a pressure meter.
Also conveniently, diamond is transparent, so a laser beam can be
shone through one face of the diamond to heat the sample and a
microscope can see in to record what happens. But that does call for
rare, white, nitrogen-free diamonds. In other experiments, powerful
x-rays from a synchrotron radiation source can shine in to reveal
the atomic structure of the mineral sample. In spite of the cost and
complexity, the anvils are compact and teams in Britain, Germany and
the USA are using them.
One of the principal results that has come both from the diamond
anvil work and from the shock-wave gun has been to estimate the
temperature at the centre of the Earth. The researchers have a
pretty good idea of the chemical composition of the lower mantle and
the core from density estimates and they know the pressure from the
density and the depth. Seismic waves tell them that the mantle is
solid and the outer core is liquid. Therefore, the lower mantle can
be no hotter than the melting point of its constituents and the
liquid outer core can be no colder than the melting point of iron at
the pressure (with allowances for certain impurities). Most
specifically, the temperature at the boundary between the outer and
inner core must be at precisely the melting point since both have
the same composition. Original estimates differed quite widely
between experimental groups. But one of the principal advances
revealed at the Royal Society meeting was that most of the
scientists are approaching a consensus.
The centre of the Earth: 5 500 - 6 500oK
Inner core/outer core boundary: 5 000 - 5 500oK
Top of outer core: 5 000 - 5 500oK
Base of mantle: 4 000 - 4 500oK
These temperatures are hotter than many expected. Indeed, the centre
of the Earth seems to be hotter than the surface of the Sun. That,
in turn, raises the question of how it became so hot in the first
place; a question with implications for how the Earth formed.
The Earth's core The boundary between the base of the mantle and the
Earth's core is not a simple one. Seismic profiles imply that, at
least in some places, there is a layer there up to 200 km thick of
material of a slightly different density to the bulk of the mantle.
This layer is known as the D" (pronounced D double prime). It
appears to be a discontinuous layer. It is simply not present under
some areas of the globe. It could be the densest material of the
mantle that has accumulated there, or a sort of solid scum that has
formed above the core.
Either way, it is likely to be a mixture of iron and silicates.
There are also indications of physical unevenness at the base of the
mantle. Very precise measurements of the rotation rate of the Earth,
made from space, reveal slight irregularities of the order of a
millionth of a second in a day. Some of these are believed to be due
to atmospheric circulation, blowing on mountain ranges; the
mountains acting like sales in the wind. But there is also evidence
that the liquid core of the Earth has a similar effect on what may
be ridges and valleys in the solid base of the mantle, rather like a
ship's keel.
The bulk of the core is made of molten iron. There is almost
certainly nickel present as well and, from the density, there may be
up to 10% of other impurities which may include sulphur, silica and
oxygen. The entire liquid outer core is slowly churning about like
cement in a cement mixer, moving at several millimetres a second. It
is in this region that the Earth's magnetic field originates.
Although the overall effect of the field is rather as if there was a
giant bar magnet there, as early geologists thought, it is far too
hot for permanent magnetism (above the Curie point). But the liquid
metal conducts electricity and as it churns about it acts like a
great dynamo, generating the magnetic field. It is not an even
circulation and it is not a smooth nor unchanging magnetic field.
Even over short periods of a few decades the magnetic poles at the
surface can wander over several degrees. Mathematical simulations of
the Earth's core suggest that there are many more rapid fluctuations
in the dynamo but that these are to some extent screened and
smoothed out by the presence of the solid inner core and the mantle.
As a result, the magnetic envelope around the Earth, the
magnetosphere, which deflects charged particles from the Sun, has
remained in place through historical time. There is clear evidence
from the alignment of magnetic particles in rocks, however, that the
entire field has reversed many times in the past, typically at
intervals of 100 000 years or so. On that basis, we are overdue for
a reversal now and since none has been observed, we do not know if
it is a sudden event with little effect beyond disorientating our
compasses, or if the field fades away completely for days or years
during the reversal, in which case harmful radiation could reach the
ground.
The inner core of the Earth, although probably at 5 000oK or more,
is under such pressure that it is solid. It is almost entirely iron,
perhaps alloyed with some nickel. As the Earth cools, the inner core
slowly grows. It is actually freezing at a rate of about a thousand
tons per second. That seems fast but, after 4.5 billion years, still
only 4% of the total core has solidified, so the Earth is likely to
have a magnetic field for billions of years to come. It is this
freezing process which keeps the outer core in motion. In addition
to heat released from the decay of radioactive elements, heat is
released by the freezing process. Also, lighter impurities in the
outer core are left behind as the iron freezes and these float
upwards through the outer core, stirring it up. All this results in
eddy-like motions in the outer core, and where eddies converge and
diverge, the magnetic field is intensified or rarefied. In spite of
the screening effect of the mantle, such magnetic anomalies can
sometimes be detected at the surface.
The deep earthquake under Bolivia in 1994 produced some particularly
clear signals from the inner core as the seismic waves were
reflected off and refracted through it. In particular, it confirmed
suspicions that the inner core is not a simple, even sphere. It
seems that the seismic waves travel 3%-4% quicker through the inner
core when going north-south compared with when they cross between
opposite points on the equator. Professor Guy Masters of the
University of California at San Diego says the effect is similar to
what might be seen if the inner core was a single crystal of iron
the size of the Moon. He emphasises that that does not mean that it
is a single crystal, but that is one explanation. Another is that it
could be made of many crystals all aligned the same way, as iron
compass needles might be in a magnetic field. Although we know very
little about the core of the Earth, Professor Masters points out
that it is so inaccessible and its conditions so extreme that it is
pretty amazing we know anything at all about it!
The formation of the Earth We know from several lines of evidence
that the Solar System is about 4 600 million years old. That figure
is supported by the ages of meteorites based on their radioactive
elements and the decay products of them. It is backed up by radio
isotope measurements on moon rocks and estimates of the bulk
composition of the Earth. There are no rocks of this age that have
survived on the surface of the Earth due to the continual
reprocessing of crustal rocks as the continents drift. The oldest
that have survived are about 3 800 million years old.
It is likely that the Sun formed as a result of density fluctuations
in a cloud of gas and dust rather like ones in the Orion Nebula
today. Gravitational contraction pulled molecules, mostly of
hydrogen, towards the forming Sun until nuclear fusion could begin
within it. That began to blow the remaining clouds away, leaving a
disc of material destined to form planets. The Sun's radiation meant
that the inner planets were mostly rocky of the refractory iron and
silicates that could stand the heat. Further out, the bulk of the
mass of the disc collected and formed the gas giants such as
Jupiter. Volatile ices, such as water and methane, were driven
further still to condense far from the Sun in the so-called Kuiper
Belt and as future comets. There are two main theories as to how the
fine dust in the solar nebula formed solid planets. Dr John Jones of
the NASA Goddard Space Flight Centre supports the idea that the
Earth slowly condensed out of the nebula, with denser materials
working their way deeper before the whole thing solidified under
gravity. But the more widely accepted theory is one in which small
particles coalesced into bigger ones and crashed into each other
time after time, sometimes fragmenting, sometime fusing, until there
were proto-planets sufficiently big to attract other passing objects
by gravity so that they crashed into them. One result of this
process was that each impact released heat. Once the Earth
approached it present mass, it would have attracted bodies of a
kilometre or more with such force that they would vaporise on impact
and melt a sizeable crater in the Earth. If this was frequent
enough, the entire surface of the planet might have turned into a
deep ocean of molten magma. The biggest impact of them all may have
been with a body as big as Mars; the resulting splash could have
thrown up enough material to create the Moon. Add to the heat
released by the impacts that coming from radioactive decay of
elements formed perhaps only a short time before in a supernova
explosion, and there may have been enough heat to melt the whole
Earth. So it is easy to see how the layered structure of our planet
might have formed and why it is still so hot in the middle. No ocean
and little of the original atmosphere could have survived such heat,
so what we have today may have come partly from volcanoes but
largely from later comet impacts. But this hot beginning means that
there can have been little time between the Earth cooling
sufficiently for seas to form and the origin of life.
Contacts
Professor Keith O'Nions F.R.S.
Department of Earth Sciences
University of Cambridge
Downing Street
Cambridge tel: 01223 333400
CB2 3EQ fax: 01223 333450
Dr Andrew Jephcoat
Department of Earth Sciences
University of Oxford
Parks Road
Oxford tel: 01865 272000
OX1 3PR fax: 01865 272072
Professor Guy Masters
Institute of Geophysics and Planetary Physics
University of California, San Diego
La Jolla
California 92093-0225 tel: 00 1 619 534 3120
UNITED STATES OF AMERICA fax: 00 1 619 534 8090
Professor Peter Olson
Department of Earth and Planetary Science
Johns Hopkins University
Baltimore MD21218 tel: 00 1 410 516 8000
UNITED STATES OF AMERICA fax: 00 1 410 516 7933
Dr Ross Angel
Department of Geological Science
University College London
Gower Street
London tel: 0171 387 7050
WC1E 6BJ fax: 0171 387 1612
Professor Raymond Jeanloz
Department of Geology and Geophysics
University of California at Berkeley
Berkeley CA94720 tel: 00 1 415 642 6000
UNITED STATES OF AMERICA fax: 00 1 510 643 9980
Professor Thomas Ahrens
California Institute of Technology
MS 252-21
Pasadena CA 91125 tel: 00 1 818 395 6906
UNITED STATES OF AMERICA fax: 00 1 818 564 0715
Professor Dan McKenzie F.R.S.
Bullard Laboratories
University of Cambridge
Madingley Rise
Madingley Road
Cambridge tel: 01223 337177
CB3 0EZ fax: 01223 60779
Enquiries to: Ref: PR 25 (95)
Miss Anna Link
Science Promotion Section
The Royal Society
6 Carlton House Terrace
London
SW1Y 5AG
Telephone: 0171 839 5561 ext 315 20th June 1995
INFORMATION NOTE
Inside the Earth
On 24th and 25th January 1995, the Royal Society held a scientific
meeting on Developments in high pressure, high temperature research
and the study of the Earth's deep interior. The enclosed document
was prepared afterwards to summarize key issues raised by the
speakers and to provide a list of helpful contacts for future
reference.
This document does not necessarily constitute the views of the Royal
Society, and views expressed in it should not be attributed to the
Society. The document is free of copyright and may be used without
reference to source.