I've always thought the earth's layers were as follows:

  • Crust, a few dozen KM think in most places
  • Mantle, completely composed of liquid rock and molten metals, a few thousand KM thick
  • Outer Core, molten iron
  • Inner core, solid iron

The mantle being composed of a liquid made a lot of sense. The idea of plate tectonics floating on top of magma and bumping against each other made sense.

However, I've been told by Michael that the mantle is completely solid, which seems to not work as well with the ideas of

  1. Floating/moving plate tectonics
  2. Convection currents

If the earth's mantle is solid, why don't we say the plate tectonics are "resting on" the mantle instead of "floating on"?

How can we have convection currents through something that doesn't flow? Even if the solid rock in the mantle can why doesn't the heat just distribute itself evenly through conduction, and negate any convection tendencies?

I realize that entire books can be written on this topic, so a generalized answer is fine.

  • $\begingroup$ The answer to this question would be a duplicate of that answer. However, the [question itself ](earthscience.stackexchange.com/questions/7888/…) was hard to find if one did not read it accidentally beforehand and this question is far more advanced (hence I am not voting to close this question as ). $\endgroup$ Jul 4, 2016 at 7:24
  • 1
    $\begingroup$ I'd like to make a quick point that the mantle has no molten metals. Is it only rock, overwhelmingly solid and occasionally liquid. There may be small pockets (centimetre scale and lower) of molten metal somewhere in the mantle, but this can occur only in extreme chemical conditions which are simply not present in the mantle. Even if there is any molten metal, the amount of it is completely insignificant. $\endgroup$
    – Gimelist
    Jul 6, 2016 at 11:40

3 Answers 3


Is the mantle solid?
It's all a matter of timescales. The mantle is undoubtedly solid (except locally in the uppermost regions where melting can occur) but on a long enough timescale it can display plastic behaviour under high temperatures and pressures, with a viscosity on the order of $\text{~}\mathrm{10^{21}~Pa~s}$ (based on estimates from isostatic rebound after melting of ice sheets). Think of plasticine - it is obviously solid but it deforms easily under stress. This is how the mantle behaves on long timescales at high temperatures and pressures. On short timescales the mantle is a brittle solid and this is what allows it to transmit seismic S-waves.

Why does the crust 'float' on top of the mantle?
The answer to this is quite simple - the crust is less dense than the mantle and so on the long timescales, where the mantle is ductile, the crust does not sink into the mantle (except at subduction zones but that's another story). Granitic continental crust has an average density of around $\mathrm{2.6~g~cm^{-3}}$, basaltic oceanic crust is around $\mathrm{3~g~cm^{-3}}$, and the upper mantle is around $\mathrm{3.5~g~cm^{-3}}$.

How does convection work in the mantle?
The details of mantle convection are a subject of active research but all geologists agree that the mantle does convect vigorously and that this is the major form of heat transport from the core to the crust. If you calculate the Rayleigh number for the mantle you come out with $R_a~\text{~}~10^6$ which indicates that convection should be vigorous. On the other hand, the thermal diffusivity of the mantle is very low ($\text{~} \mathrm{10^{-6}~m^2~s^{-1}}$) and so conduction is not the dominant form of heat transport.

  • 2
    $\begingroup$ I like this answer but it would benefit from some links or references. $\endgroup$
    – Matt Hall
    Jul 6, 2016 at 15:40

Humans have difficulty visualizing how rocks flow because the required conditions are outside our everyday experience. We are familiar with surface temperatures and pressures, and geologically short time-frames, where the distinction between solid and liquid is obvious. Under very high temperatures and pressures, and especially over longer time periods, solids can assume plasticity, almost like a fluid, albeit a fluid with ultra-high viscosity. You can see the outcome of these changed rock properties when rocks, which have undergone deformation at depth, are brought back to the surface. Drive or trek through any mountainous area where the mountain roots have been exposed by deep erosion, and you will see dramatic folding - examples of rock plasticity. The mantle is much deeper, hotter and under immense confining pressure, so yes, the mantle can have convection cells without being a liquid.


I'd like to add to the correct answers already given.

Solids can flow and deform without being a liquid. You can bend iron bars. You can dent plastic. You can squeeze nylon bags. The cotton on your t-shirt bends as you move you hand. These are all solids. The ability of this to happen depends on temperature and pressure. Iron bars bend easier when heated. I'm sure you know that freezing (already solid) objects with liquid nitrogen makes them much more brittle.

The most obvious difference between solids and liquids is their ability to deform, but that does not mean that only liquids do it. Liquids do not have crystal structures, whereas (some) solids do. Take H2O: water is liquid, but ice is solid. However, glaciers can deform and move. All in the solid state. Metamorphic rocks deform, in the solid state.

Once we accept that solids can deform without being liquid, the ability to sustain convection cells and density-based stratification are not surprising. Fill in a box with gummi bears, then drop a huge block of steel on to it. Eventually the candy will flow upwards while the steel block will sink. All of this will happen in the solid state.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.