Almost all of what I write about in this blog concerns only 1% of the earth’s volume. All crust, all sedimentary rocks, the glories of mountain building, all occupy an insignificant portion of the earth. It’s the only bit we can get to – in geology, we are the 1%. This post is all about the 99%, the earth’s interior, the mantle and core, which is remarkably different to here. The inner core reaches 5430°C (as hot as the surface of the sun) and pressure is multiple millions times atmospheric pressure. Forget your Jules Verne, we are never going to reach the centre of the earth*.
We aren’t isolated from the interior though, it affects us in many ways – from the earth’s magnetic field to plate tectonics and volcanic islands. The influence goes the other way too. Oceanic crust is constantly being pushed down into the mantle (maybe 20% by volume over time) and it goes a surprisingly long way. Rocks that form part of the familiar 1% start on the surface but travel down into the extreme conditions of the deep interior.
Deeper and deeper
In a previous post, we traced the journey oceanic crust takes, from its birth in a mid-ocean ridge down into the mantle in a subduction zone. We left the story at 250km depth.
Here our crust has been transformed. Its minerals have changed into new ones, stable at these depths. Metamorphism and partial melting have taken away most of its water and some other elements that have large ions. It consists largely of garnet and pyroxene. The material surrounding it, called peridotite may in addition contain olivine. As it descends further, at around 2cm a year, our plate continues to change.
A huge amount of what we know about the deep earth comes from listening to earthquakes. Seismologists can identify the location of earthquakes within the sinking slab. This pattern of earthquakes (called the Benioff zone) shows the shape of the subduction zone. Seismic waves can be used in multiple ways to identify the properties of the materials they pass through. The liquid outer and solid inner core were identified this way in the 1930s.
We can also now spot sudden changes in the speed at which the waves pass (seismic velocity). This implies a sudden change in the properties of the rock. In the mantle, there are multiple places where we see such a change. They are known as discontinuities and they occur at specific depths.
The shallowest mantle discontinuity is at 300km depth. It’s only seen intermittently and is most common beneath continents and island arcs. It’s been interpreted as a phase change in minerals made of SiO2 (silica). At the surface, the stable form of silica is quartz, but at depth it is transformed, first into coesite (discovered 1953) and then, at 300km depth into a mineral called stishovite (1961). Stishovite is dramatically denser than coesite so the transformation from one into the other changes the rock properties, making it visible to seismologists. Nearly all rocks contain SiO2, but in the mantle it is all locked up into other minerals. Only subducted oceanic crust has enough SiO2 to contain ‘free silica’. For this reason, the existence of the 300 km discontinuity in a particular place is evidence for the presence of subducted oceanic crust .
There is not much free silica even in our sinking slab of crust, so as it passes through the 300km discontinuity the transformation is relatively minor. Bigger changes await.
Into the lower mantle
At 410km we are leaving the upper mantle and entering the transition zone, marked by several seismic discontinuities. Here were are reaching pressures of 24GPa (250,000 atmospheres) and temperatures of 1600°C. Now olivine is no longer stable. It transforms itself into other minerals with the same composition (polymorphs), first wadsleyite (1966) and then ringwoodite (1969).
These minerals with the funny names aren’t found in crustal rocks, aren’t stable under conditions humans can reach. The first line of evidence came from meteorites. Most meteorites are pieces of a rocky planet that formed a core and mantle but was then smashed into pieces by the hurly-burly of the early solar system. Both wadsleyite and ringwoodite were both first recognised in meteorites and are named after Twentieth Century scientists. Samples from meteorites give us an insight into high pressure conditions and a rare opportunity to find actual samples of them. The only other place is tiny pieces inside diamonds.
At a depth of 650km there is a significant discontinuity that marks the beginning of the lower mantle. This is the point where most of the minerals mentioned above are no longer stable. Under the intense pressure minerals with extremely dense, tightly-packed mineral structure are formed. Here Ferropericlase (aka Magnesiowüstite) and Silicate perovskites become the dominant minerals. Note the silicate perovskites have the same mineral structure as perovskite (a calcium titanium oxide) but a different composition.
It used to be thought that sinking crust couldn’t get through the 650km discontinuity. Maybe some slabs don’t, but ours does, continuing its journey into the lower mantle.
Much of what we know about the lower mantle comes from experimental work. In order to simulate the crushing pressures and high temperatures within the deep earth, experimenters have to take tiny samples and squeeze them between two diamonds and blast them with frickin’ lasers. These ‘diamond anvils’ allow scientists to peer through to materials under deep-earth conditions, studying the minerals that form there. Only recently have scientists managed to create diamond anvils able to achieve conditions within the deepest parts of the earth. This brings home quite how extreme and alien the deep earth is.
Research on the deepest earth is cutting edge and not yet settled so the story of our plate’s progress becomes a little vague. A recent addition to the menagerie of extreme minerals is post-perovskite (2004) a mineral stable in the lowest 200 km of the mantle, a material so exotic it doesn’t yet have a proper name (being just ‘stuff perovskite turns into’). It’s been linked to the deepest seismic discontinuity, the D” (D double prime) which is a layer just above the core-mantle boundary. This mysterious boundary zone may be linked to the fate of our oceanic slab. As our crust ends its long journey it encounters a boundary it cannot cross. The core is made of metal, which doesn’t mix with our silicate crust, so the slab ends up lying flat on the base of the mantle, forming the D” layer.
Everything is connected
In one of earth science’s many dizzyingly lovely links between different domains, the very deepest layers of the mantle have dramatic influences on the earth’s surface. Firstly, large igneous provinces – big areas of vulcanism not obviously linked to plate tectonics, seem to originate in mantle plumes that start at the core-mantle boundary (these in turn may cause mass extinctions). One theory is that the presence of old oceanic crust leads to the instabilities that drive plume formation . Here geochemistry, specifically isotope geochemistry, is our tool for understanding places we can’t reach.
Another link between the surface and the deep relates to the magnetic field that helps homing pigeons and old-fashioned geologists find their way. Some researchers speak of hot dense “thermochemical piles” created by deep subduction which influence heat loss from the outer liquid core. This core creates earth’s magnetic field and a recent paper links patterns of magnetic polarity reversals with these piles and so with patterns of deep subduction and plume formation.
The reality of deep subduction seems clear, even if the details are still coming into focus. A recent comparison of numerical modelling of subduction with cutting edge seismic imaging (seismic tomography) shows a good match  – oceanic plates really are going all the way down.
Although the deep earth seems an alien place, turning familiar minerals into substances we are only starting to understand, still it is intimately related to our surface world. A lump of basalt that forms at a mid-ocean ridge, that gets subducted and transformed into new minerals again and again, may ultimately rise up as part of a mantle plume and get melted, ending up as a new form of basalt. Crust and mantle are intimately linked – the 1% and the 99% together make up a single planet, after all.
* I say never, but it may be possible, provided you got hold of a nuclear weapon and a hundred thousand of tonnes of molten iron.
 Evidence for 300km discontinuity (abstract only)
 Link between oceanic-island basalts and deep mantle (pdf of paper)
 Deep subduction – comparison of geodynamic and tomographic models (open source paper)
Q. Williams, & J. Revenaugh (2004). Ancient subduction, mantle eclogite, and the 300 km seismic discontinuity Geology DOI: 10.1130/G20968.1
W. M. White (2010). Oceanic Island Basalts and Mantle Plumes: The Geochemical Perspective Annual Review of Earth and Planetary Sciences DOI: 10.1146/annurev-earth-040809-152450
B. Steinberger, T. H. Torsvik, & T. W. Becker (2012). Subduction to the lower mantle – a
comparison between geodynamic and
tomographic models Solid Earth Discuss DOI: 10.5194/sed-4-851-2012