Last Author

Credit: M. LEROUGE

The boundary between Earth's molten iron core and the solid mantle is the place to look for clues to how heat from the core is convected to the surface, says geophysicist Patrick Cordier. He and his colleagues at the University of Lille in Villeneuve d'Ascq, France, want to find out what is happening in the D″ layer just above the core–mantle boundary, where temperature and density are dramatically higher than in the rest of the mantle. On page 68 they report the use of quantum mechanics to model how flow occurs in a solid in these conditions.

What is the biggest misconception about the mantle?

That it is liquid. The mantle is rock. The lower mantle is quite homogeneous and dull — until you get to the D″ layer, where seismic waves reveal properties that differ according to the direction of measurement.

Why is the D″ layer the most important?

The density of the molten iron core is double that of the solid mantle, a bigger jump than between the surface of Earth and the air. The core is also immensely hot — around 5,000–6,000 K. Some hypotheses suggest that the mantle may be melting in this layer or maybe that subducted plates stop there. Two years ago, scientists found a new phase transformation in the mantle — from perovskite to a post-perovskite found only at high pressures — which may help explain what is happening in the D″ layer.

Why is this boundary so difficult to model?

Flow of a solid implies the motion of a large number of crystal defects, or dislocations, involving several thousand atoms. This is very complicated to describe quantitatively at the atomic scale. The best tool is quantum mechanics, but in practice modelling even simple materials is difficult. And here we're dealing with an inaccessible mineral under extreme temperature and pressure.

How does your model improve on past efforts?

At this scale, matter behaves like a continuous medium. There's only a thin layer of atoms close to the core of the defect for which we need quantum mechanics. The problem was to link the quantum mechanisms with the bulk of the material.

What insights are now within reach?

My dream is to create a complete numerical model of the deformation of the mantle materials. We've described dislocations, and next we need to describe their behaviour, motion and interactions. The big problem then is to jump from the atomic scale to the grain — and then to the rock.