How Did We Learn the Earth Has a Core We Have Never Seen?

We know the Earth has a core because earthquakes ring the planet like a bell, and the ringing tells us what the inside is like. Waves from a quake travel through the whole planet. Careful instruments around the world record when they arrive, how strong they are and which ones fail to turn up at all. Those patterns only make sense if the Earth is layered, with a liquid outer core and a solid inner core. Nobody has seen or touched either.

I find this one of the most satisfying ideas in science, partly because it is so indirect. It's like working out what's in a wrapped present by shaking it, except the shaking is done by the planet and the listening takes a hundred years.

Let me walk through how it fits together. I'll start with the humble question of why we can't just go and look.

Why can't we just dig down and look?

The deepest hole humans have drilled goes down about 12 kilometres. That sounds a lot until you remember the Earth's radius is about 6,371 kilometres. Our deepest borehole has barely scratched the skin of an apple.

The core begins roughly 2,900 kilometres down. Pressures there are enormous and temperatures run to thousands of degrees. No drill, probe or clever robot is going to get there.

My first guess, when I started reading about this, was that someone must have found the core by some sort of deep sampling, perhaps from volcanoes bringing bits of the interior to the surface. Volcanoes do bring up useful clues, but they come from the upper mantle at best. The core is a different story. Everything we know about it is inferred from things that reach the surface: waves, gravity and magnetism.

That's the detective part. We never see the culprit. We see footprints, and we ask what kind of thing could have made them.

First clue: the Earth is heavier than its rocks

The earliest clue has nothing to do with earthquakes. It's about weight.

By the end of the eighteenth century, Newton's law of gravitation made it possible to work out the Earth's total mass, if you could measure the strength of gravity between ordinary objects. Henry Cavendish's famous experiment, using small and large lead balls on a delicate twisting balance, did exactly that. From it came the Earth's average density, which is about 5.5 times that of water.

Here's the thought experiment. Pick up a typical rock from a New Zealand beach or hillside. Most common surface rocks, such as granite, sandstone and basalt, are about 2.5 to 3 times as dense as water. If the whole planet were made of that sort of rock, its average density would come out at around 3 or less.

But the real average is nearly double. So something inside must be much denser than anything we can pick up from the surface. Squeezing helps a bit, since rock compresses under pressure, but not by enough to explain the gap. The simplest reading is that the deep interior contains a lot of something heavy.

This didn't tell anyone it was a core, a metal or a particular size. It only said that the Earth is not the same stuff all the way through. A clue, not a conclusion. (I like that about good detective stories: the first clue rarely solves anything, it just narrows the field.)

Earthquake waves: a torch for the dark interior

Now for the main evidence. When an earthquake happens, it releases energy that spreads out as waves. Some run along the surface. Others, called body waves, dive through the planet's interior. There are two kinds that matter here:

• P waves (the "primary" ones) are push-pull waves. Think of a sound wave, or a squashed Slinky where the compression pulse runs along the coils. They can travel through solids, liquids and gases.
• S waves (the "secondary" ones) are side-to-side waves. Think of flicking one end of a rope so a wiggle runs down it. They can travel through solids, but not through liquids.

P waves are faster, so they arrive first at a seismometer. S waves come later, and they usually shake harder.

Why can't S waves go through a liquid?

This is the "why" that unlocks the whole story, so it's worth slowing down.

A shear wave works by shoving material sideways and relying on it to spring back, which then shoves the next bit along. A solid does spring back. If you push the top of a block of jelly sideways, it pushes back and returns to shape.

A liquid doesn't. Stir your tea and the liquid just flows past itself. Nothing springs back, so there's no restoring force to pass the wiggle along. (If you want a formula, the speed of an S wave depends on the material's resistance to shearing, and in a liquid that resistance is zero. So the speed is zero. The wave simply can't travel.)

P waves, on the other hand, work by squeezing, and liquids resist squeezing very well. Water in a sealed syringe won't compress when you push the plunger. So P waves go through liquids happily, though more slowly than through the stiff rock around them.

This difference is the key. It means that if you could track whether S waves make it through a region, you'd be testing whether that region is solid or liquid.

(A small tip if you like hands-on checks: spin a hard-boiled egg and a raw egg on the kitchen bench. Stop each with a finger and let go. The hard-boiled one stays stopped, while the raw one creeps back into motion because the liquid inside keeps swirling. Same shell, very different inside, and you can tell without cracking either. Seismologists do something rather like this at planetary scale.)

How do waves reveal layers?

Here is how the evidence is collected. When a big earthquake happens somewhere, seismometers all over the world record it. Each one logs when the P and S waves arrive. Because the stations sit at different distances from the quake, you can plot arrival time against distance.

If the Earth were a uniform ball of one material, that plot would be smooth and predictable. Even in a ball where the rock gets stiffer with depth, you'd expect a smooth curve, with paths that bow gently upwards.

Real Earth doesn't give a smooth curve. Sudden changes in speed make waves bend, reflect and refract at certain boundaries. Early in the twentieth century, Andrija Mohorovičić noticed this kind of kink in records from nearby quakes and used it to identify the boundary between the crust and the mantle (it now carries his name, the Moho). The same logic, applied to waves that went much deeper, became the way to find the core.

A handy way to picture bending is a shopping trolley rolling diagonally from smooth pavement onto grass. The wheel that hits the grass first slows down, and the trolley swings round. Waves behave similarly when they cross into a material where they travel at a different speed. So a change in direction or timing tells you there's been a change in material.

The shadow that gave the liquid away

In 1906, Richard Oldham published an analysis of earthquake records showing that waves passing through the deep interior arrived later than they should have if the Earth were uniform. He argued that there was a distinct central region where waves slow down. That was the first clear evidence of a core. Later, in the early 1910s, Beno Gutenberg worked out how deep its boundary lies, at roughly 2,900 kilometres, and that figure has held up well since.

Two patterns in the records point to what that core is like.

Pattern one: S waves vanish. For seismometers on the far side of the planet from a quake, S waves simply don't arrive. Roughly speaking, beyond about 100 degrees of arc from the earthquake (a quarter of the way round the globe, and a bit more), there are no S waves. Since S waves can't pass through liquid, something in the middle must be liquid, and big enough to block those paths. The mantle can't be what blocks them, because S waves clearly do travel through the mantle (we see them nearer the quake). So the blocking must happen deeper, in the core.

Pattern two: P waves are bent into a shadow. P waves do get through the core, but they slow down sharply when they enter it, because liquid is less stiff than the mantle rock above. That slowing bends them steeply inward, like the trolley swinging toward the grass. The result is a ring-shaped band on the far side of the planet, from about 103 to 142 degrees away from the quake, where direct P waves barely arrive. This is the P-wave shadow zone.

The two patterns agree, and that matters. The P-wave pattern says the core is a region where waves slow down. The S-wave pattern says it can't be solid. Different observations, one story.

One common misconception is worth clearing up here. It's easy to hear "the mantle is solid, the outer core is liquid" and then picture the mantle as a rigid, immovable shell. In fact, the mantle is solid rock (that's why S waves cross it) but over millions of years it slowly creeps. Solid on the timescale of a wave, flowing on the timescale of continents. I tripped over this myself and I'm still getting used to the idea that one material can be both.

A bit of arithmetic helps with scale. The core's radius is about 3,480 kilometres, which is 55% of the Earth's radius. Volume goes with the cube of radius, so the core takes up about 16% of the planet's volume (0.546 cubed is roughly 0.163). It's a lot smaller than half the planet by volume, yet because it's so dense it holds about a third of the Earth's mass. That fits nicely with the weight clue from earlier, which is the sort of thing that makes me grin when I notice it.

The surprise inside the shadow: a solid inner core

If the story ended there, we'd have a liquid core and a mantle. But the shadow zone wasn't quite empty.

In 1936, Inge Lehmann studied records of faint P waves that were showing up inside the shadow zone, where, according to the simple one-liquid-core picture, none should arrive. Her explanation was that the core isn't one uniform liquid. There's a distinct boundary deeper inside, and some P waves reflect or bend off it back toward the surface, landing in the shadow. In other words, there is a separate inner core, with different properties from the outer core around it.

Later work, over the following decades, showed that this inner core is solid. The evidence builds on the same principle. Waves speed up as they enter it, which is what you'd expect from stiffer material. And studies of the Earth's slower, longer vibrations, along with S-type waves that seem to be generated within it by converting from P waves at the boundary, fit with a solid centre. (This last part gets technical and I'm still learning the details, so I'll leave it at "the later evidence agrees".)

You might wonder how the middle can be solid when it's hotter than the layer around it. The answer is pressure. Squeeze a material hard enough and its melting point rises. At the very centre, the pressure is so extreme that the iron-rich metal stays solid even at temperatures that would melt it near the surface. The outer core is slightly less squeezed, so at nearly the same heat it stays liquid.

By the numbers, the inner core has a radius of about 1,220 kilometres. That's roughly 19% of the Earth's radius, so only around 0.7% of its volume. A very small thing at the centre of a very large thing, found by listening to faint echoes.

What is the core made of, and how do we know?

Here I have to be careful, because the honest answer has two parts: what is well established, and what is still being worked out.

Well established: the core is mostly iron, with some nickel. The reasoning goes back to that density clue. Iron is abundant in the solar system and dense enough, under core pressures, to match what the waves and the Earth's mass require. Meteorites that look like pieces of broken-up planetary cores, which are made of iron and nickel, support the idea that this is how planets sort themselves out when they form.

Still being worked out: the core appears to be slightly less dense than pure iron-nickel would be under those conditions, which suggests some lighter elements are mixed in. Which ones, and in what amounts, is something scientists are still debating. I'd be making things up if I named a winner.

There's also a nice link to a completely different field. A compass works because the Earth has a magnetic field. The most widely accepted explanation is that the field comes from the liquid outer core. Molten iron is a good electrical conductor, and as it churns (carried by heat escaping from the core and by the planet's rotation), it generates electric currents, which in turn make a magnetic field. This is called the geodynamo. So the same liquid layer revealed by vanishing S waves is, in all likelihood, the reason a compass needle points north. It's one story told by two entirely different instruments, and I think that's quietly wonderful.

What this kind of detective work can and can't tell us

It's worth stepping back and noticing how the method works, because it's the real hero of the story.

No single scientist found the core in a flash. The picture built up as many people collected records, compared them, noticed things that didn't fit, and proposed explanations that could be checked against the next earthquake. Each answer raised a sharper question. The Earth is denser than its rocks. Why? The waves arrive late. Why? S waves vanish. Why? Faint signals appear in the shadow. Why?

It also means the picture has limits. Seismic waves tell us about speeds, and we turn speeds into guesses about materials, temperature and state. Different combinations of ingredients can sometimes produce similar speeds, so there's room for honest disagreement. When scientists say the inner core is solid iron-rich metal, they mean it's the best explanation for a lot of independent evidence, not that anyone has held a sample.

A few things are still open, and I'd rather say so than gloss over them:

• Exactly which lighter elements are in the core.
• How old the solid inner core is, and how fast it has been growing as the planet slowly cools.
• Whether the inner core has finer structure inside it. Some waves travel at slightly different speeds depending on direction through it, and scientists are still working out what that means.

Something to wonder about

Next time you feel a rumble (and in this part of the world, most of us do now and then), it might be worth knowing that some of that shaking is also heading downward, through rock and then through liquid metal, and a bit of it is bouncing off a ball of solid iron the size of a large moon, thousands of kilometres below your feet. Somewhere, an instrument is listening.

The question I keep turning over is this: if a few wiggly lines on paper were enough for people to work out the structure of a place no one can visit, what else is hidden in signals we've already recorded and haven't yet learned to read?

If you've got a hard-boiled egg and a raw one in the kitchen, give them a spin. It's a surprisingly good way to feel the idea in your hands. And thank you for reading along. I'm still learning this too.