A banshee comes up through the cloud deck at dawn, and the first thing you see is wrong. Not strange-beautiful, not alien-pretty — wrong, in the specific way that makes a physicist sit up. There is a mountain ahead of the rider, several kilometres of grey rock and dripping forest, and it is not attached to anything. It hangs. Below it is air, and below the air is more air, and somewhere far down through the mist is the ground it should be lying on.
Your eye keeps trying to find the trick. A pillar, a cable, a matte line where the model meets the plate. There isn't one. The Hallelujah Mountains of Pandora — Ayram alusìng in Na'vi, "the floating mountains" — simply float, the way a cork floats, except there is no water, only sky.
The easy thing to do with an image like that is to call it magic and move on. I want to do the harder thing, which is to take it seriously as a specimen. Because if you ask the patient, slightly obsessive question — what would actually have to be true for a mountain to hang in the air and stay there — you do not end up in fantasy. You end up in one of the strangest, best-verified corners of real physics, in a Leiden laboratory in 1911, and eventually inside a phenomenon that lets a coin-sized disc lock itself rigidly into empty space as if the air had quietly turned to glass around it.
The mountains are the bait. Superconductivity is the meal. Let me earn it.
What canon actually claims
Let us start with what the films and the official material commit to, because that is the part we are not allowed to invent.
Canon is specific about the cause: the mountains float because of unobtanium, a mineral concentrated in their cores. Pandora's magnetic field is unusually strong — the moon's iron core does the ordinary planetary dynamo thing, but buried unobtanium is described as amplifying that baseline field more than a hundredfold.
Where Pandora's field overlaps the field of the gas giant it orbits, Polyphemus, you get a region of violent, concentrated magnetic flux: the Flux Vortex, centred over the mountains themselves.
Inside that vortex, canon says, the unobtanium responds to the field by expelling it, generating a repulsive force strong enough to tear ore-bearing rock off the crust and hold it up. The phrase the early material reaches for is the Meissner effect — a real piece of physics, which we will get to. The vortex is also why the region is a graveyard for instruments: it jams radio, blinds radar, and forces pilots through on visual flight rules with unguided rockets, because nothing electronic survives the field gradients intact.
So far, so confident. But here is where reading canon carefully pays off, because the official story quietly contradicts itself, and the contradiction is the most interesting thing in the chapter.
The crack in the official story
If you push on the word "Meissner," it gives way.
The Meissner effect, as we will see, is a purely repulsive phenomenon. A magnet floating above a superconductor by Meissner repulsion alone is in exactly the situation of a marble balanced on top of a bowling ball: technically up, in practice doomed. The smallest nudge sends it sliding off sideways. There is a theorem about this — Earnshaw's theorem — and it is unforgiving. Pure magnetic repulsion cannot give you stable levitation. It can give you a brief, twitchy hover before everything tips over and falls.
A mountain held up by the Meissner effect alone would not serenely drift. It would flip, slide off the field lines, and come down.
The later canonical material seems to have noticed. The newer companion books quietly introduce two new phrases to explain how the mountains stay — "flux pinning" and "quantum locking" — without ever announcing that they are correcting the earlier account. It is the kind of silent retcon you only catch if you are reading for the physics. And it is the right correction, because flux pinning is precisely the thing that turns an unstable hover into a rigid, locked suspension. Canon stumbled into the right answer on the second try.
To understand why that correction matters — why "pinning" and not "repulsion" is the word that holds a mountain up — we have to leave Pandora for a while and go to Leiden.
Where resistance went to die
In 1911 the Dutch physicist Heike Kamerlingh Onnes had something almost nobody else on Earth had: the ability to make things extraordinarily cold. He had liquefied helium, which boils at about four degrees above absolute zero, and he was using it to ask a simple question — what happens to the electrical resistance of a metal as you cool it toward the bottom of the temperature scale?
The expected answer was it gets smaller, smoothly. Electrons carry current through a metal by drifting through a lattice of atoms, and they are constantly bouncing off the lattice's thermal jitter, losing energy as heat. That bouncing is resistance. Cool the metal, quiet the jitter, and the resistance should ease down toward some small leftover value.
That is not what mercury did. As Onnes cooled it past about four kelvin, the resistance did not ease toward a small value. It fell off a cliff — straight to zero, abruptly, completely, as far as any instrument could measure.
Not nearly zero. Zero. A current started in a closed loop of such a material will, as far as anyone has ever been able to detect, run forever — no battery, no voltage, no loss. Onnes had found superconductivity, and the temperature at which a given material falls off that cliff is its critical temperature, its T꜀.
Superconductor cool-down
It took another forty-six years to explain why.
Why electrons stop bouncing
The explanation, when it came in 1957 from John Bardeen, Leon Cooper, and Robert Schrieffer — BCS theory — is one of those ideas that sounds like it cannot possibly be right.
Electrons repel each other. They are all negatively charged; pushing two of them together costs energy. And yet the whole of superconductivity rests on electrons pairing up. How?
The trick is that the electrons are not alone. They move through a lattice of positive ions, and a passing electron tugs those heavy positive ions very slightly toward it as it goes. The ions are sluggish — they lurch inward after the electron has already moved on, leaving behind, for a fleeting instant, a small pocket of concentrated positive charge in the lattice. A second electron, arriving in that instant, is pulled toward the pocket. The first electron has, in effect, left a faint positive wake, and the second electron rides it.
That indirect, lattice-mediated attraction is enough — barely — to bind the two electrons into a Cooper pair. And the pairing changes everything, because of a deep rule about how nature sorts its particles.
A lone electron is a fermion: antisocial, forbidden by the Pauli exclusion principle from sharing a quantum state with another like it. But a Cooper pair has equal-and-opposite spins that cancel to zero, and a composite object with zero net spin behaves as a boson — and bosons are the opposite of antisocial. They are gregarious. They will happily pile into the same quantum state, all of them, at once.
That is the first miracle: no resistance. But a floating mountain needs the second one.
The mirror that pushes back
In 1933, Walther Meissner and Robert Ochsenfeld found that a superconductor does something beyond merely conducting perfectly. Cool it below T꜀ in a magnetic field, and it actively throws the field out. The magnetic flux that had been threading through the metal is expelled; inside the superconductor, the field drops to nothing. This is the Meissner effect, and it is true perfect diamagnetism — the material becomes a perfect magnetic mirror.
The mechanism is elegant. The moment a field tries to enter, lossless currents spring up on the superconductor's surface — and because they are lossless, they do not fade. They circulate exactly as needed to generate an opposing field that cancels the intruder inside the bulk. A permanent magnet brought near such a surface sees its own mirror image pushing back, and the push can be strong enough to hold the magnet up against gravity.
This is the part early Avatar canon got hold of, and you can see why it is tempting. A material that expels magnetic fields and floats magnets — point it at a mountain full of the stuff sitting in a giant magnetic field, and surely the mountain floats.
Except for Earnshaw.
Why repulsion alone always falls
Here is the problem the films' first explanation walks straight into. Pure magnetic repulsion cannot hold anything stably. Samuel Earnshaw proved it in 1842: you cannot trap an object in a stable equilibrium using forces that fall off the way magnetic and electric forces do. There is always an escape direction — a way to slide "downhill" off the field and out.
A magnet floating on Meissner repulsion is balanced, but it is balanced like a pencil on its point. Nudge it and there is no restoring force pulling it back to centre; instead it tips, slides off the field line, and falls. Lab demonstrations of pure Meissner levitation are accordingly fussy and twitchy — the magnet wants to escape, and it will.
So a mountain held up by Meissner repulsion alone is not a floating mountain. It is a falling mountain that hasn't fallen yet. The thing that turns "hover" into "lock" is what canon reached for on its second attempt.
The materials that let the field in
Superconductors come in two kinds, and the difference is everything here.
Type-I superconductors — most pure elemental metals, like Onnes's mercury — are absolutists about the Meissner effect. They expel magnetic field completely, right up until the field gets too strong, at which point superconductivity collapses entirely and the material goes normal. All or nothing. Their tolerance for magnetic field is tiny, which makes them useless for anything involving big fields.
Type-II superconductors — alloys like niobium-titanium, ceramics like YBCO — are subtler, and far more useful. They have two thresholds. Below the first, they behave like everyone else: full Meissner expulsion. Above the second, superconductivity dies. But in the wide region between, they do something no Type-I material will: they let the field in partway, on the field's own terms, without giving up superconductivity.
The field does not flood in uniformly. It pierces the material in a forest of discrete, quantized threads — Abrikosov vortices, each a tiny tube carrying exactly one quantum of magnetic flux, each with a normal (non-superconducting) core, each wrapped in its own whirlpool of circulating supercurrent. The bulk between the threads stays superconducting. The material is sewn through with magnetic needles, and it does not mind.
And those threads are the hooks the mountain hangs from.
Quantum locking: the beat the whole chapter turns on
Here is where it stops being repulsion and starts being something with no everyday analogue.
Those Abrikosov vortices threading a Type-II superconductor would, if left free, drift around under any push — and a drifting vortex dissipates energy, which spoils the superconductivity. So a good Type-II material does not leave them free. Its crystal lattice is full of imperfections — missing atoms, impurities, grain boundaries, deliberately introduced voids — and each imperfection is a place where superconductivity was already going to be weak. A vortex core, which is also a spot of suppressed superconductivity, finds its lowest-energy home sitting exactly on one of those defects. It drops into the defect like a key into a lock and stays there. This is flux pinning.
Now follow what that does. The magnetic flux threads are locked to the superconductor's lattice. The flux threads are also anchored in the external magnet's field. So the superconductor is, in effect, sewn to the field itself — not resting on it, threaded through it. Try to lift it and the pinned threads pull back. Try to push it down and Meissner repulsion shoves up. Try to slide it sideways and the threads resist being dragged off their defects. Every direction has a restoring force. Earnshaw's escape route is gone.
This is quantum locking, and if you have seen the lab demonstration you know it is uncanny in a way the words undersell. A small disc of YBCO, cooled with liquid nitrogen, is held above a magnetic track — and it just stops, fixed in the air at whatever height and tilt you left it. Tip the whole track upside down and the disc hangs underneath it, suspended against gravity, refusing to fall. Give it a track shaped in a loop and the disc glides around the loop without friction, locked at its hovering distance the whole way, hands-off.
It is not balancing. It is locked. And that is the difference between a mountain that hovers for an instant and a mountain that has hung over the Iknimaya for longer than the Na'vi have had names for it.
Back to the mountain, now with the right tool
Return to Pandora holding the correct idea, and the floating mountains snap into a surprisingly coherent picture — the one the later companion books were groping toward.
The Flux Vortex is the magnet track, scaled to a planet: the overlapping fields of Pandora and Polyphemus, dense and structured, doing what the neodymium rail does under the YBCO disc. The unobtanium in the mountains' cores is the superconducting puck. And the canon detail that suddenly matters is the description of unobtanium's structure as a quasicrystal riddled with microscopic voids — because voids are pinning centres. A material described as full of billions of tiny voids is, whether the writers intended it or not, a material described as an excellent flux-pinner.
So the planetary flux threads pierce the voids in the unobtanium, pin there, and lock the ore-bearing rock into Pandora's field. The mountain is not resting on a magnetic cushion that it might slide off. It is pinned — threaded onto the field lines of a gas giant and held, the way the upside-down disc is held beneath its track.
There is even a tidy mechanical bonus the companion material gestures at, and it is sound. The non-magnetic lower rock of each mountain is dead weight, pulling straight down. The pinned superconducting core resists any change of position at all. Gravity on the ballast and the restoring force of the pinned vortices act like two opposed springs, and the mountain settles into a stable, self-correcting equilibrium — passive, powerless, permanent. A keel of pinned flux; a ballast of ordinary stone.
It is, frankly, better physics than the films promised. The first explanation was wrong in an instructive way; the corrected one is not just plausible but elegant.
Where the story still cheats
Honesty requires flagging what canon cannot pay for, because a specimen is only worth reading if you read all of it.
The biggest cheat is temperature. Every superconductor we have ever made needs to be cold — liquid helium for niobium-titanium, liquid nitrogen for the ceramic high-temperature ones. Unobtanium is asserted to superconduct at room temperature, on a warm moon, with no cooling at all. That is not a small extrapolation. A genuine room-temperature, ambient-pressure superconductor would be one of the most consequential discoveries in the history of materials science, and on our Earth it does not yet exist.
So unobtanium's defining property is, by our physics, speculation — a wish dressed as a mineral. There are smaller cheats too. Canon can't decide whether unobtanium is a "compound" or "Element 120" (those are mutually exclusive, and Element 120 would be a wisp that exists for milliseconds). And canon is conspicuously silent on a lethal detail: fields strong enough to levitate kilometres of rock would induce ruinous currents in the bodies of anything flying through — the banshee, the rider, the Na'vi. The films never explain why the Iknimaya isn't a graveyard. I don't think there is an explanation. It is the price of the image.
None of this spoils the chapter. It sharpens it. The mountains are where Pandora's storytelling borrowed a real, beautiful piece of physics, got the mechanism slightly wrong, quietly corrected it, and then asked for one big impossible thing — room temperature — to make it work. Knowing which parts are real is the difference between being fooled and being taught.
The payoff is on Earth, and it is already moving
Here is the part that should land hardest, because it is not speculative at all.
Flux pinning is not exotic. It is infrastructure.
The same quantum locking that would hold up a Pandoran mountain is what holds up a train. A superconducting maglev vehicle carries no wheels and touches no rail; it is pinned above its magnetic guideway exactly as the YBCO disc is pinned above its track. Japan's L0 — niobium-titanium magnets, liquid helium — has carried passengers at 603 kilometres per hour. A high-temperature design in Chengdu uses flux pinning so directly that the vehicle passively locks above a permanent-magnet track with no active control system at all, the same hands-off stability as the lab demo, scaled to something you can ride.
And the loss-free currents — the first miracle — are why an MRI machine can sit in a hospital basement holding an enormous, perfectly steady magnetic field with no power feeding the magnet, and why the magnets that bottle the plasma in a fusion reactor can exist at all. Every one of these is the floating mountain, disassembled into parts we can build.
Lab YBCO disc
~10 mm
hover gap, hands-off
L0 maglev
603 km/h
manned record
Hallelujah Mountains
16 km
canon, the same idea scaled up
That is the trick the floating mountains pull on you. They look like the most fantastical thing on Pandora — pure spectacle, pure invention. And the physics underneath them is the most real thing in the whole film: a phenomenon we discovered in 1911, explained in 1957, and now ride to work. Cameron's artists reached for an impossible image and, almost by accident, drew a faithful portrait of one of the truest strange things our own universe permits.
What stays open
Not yet, as far as anyone has shown reproducibly. It would be civilisation-changing, which is exactly why every claim draws intense scrutiny — and why LK-99 and the Rochester hydrides collapsed under replication. The honest answer is: profoundly wanted, not yet real.
Canon never says. Fields that levitate kilometres of rock would induce dangerous currents in nervous systems and blood. There is no in-universe explanation; it is the price paid for the image.
The official record says both, and they are mutually exclusive. 'Element 120' would be a superheavy atom existing for milliseconds; a stable quasicrystalline rock cannot be a single such element. The contradiction is unresolved canon.
Unknown, and probably far less than canon needs. A superconductor expels and pins field locally, but amplifying a planetary dynamo a hundredfold is well beyond any mechanism we can point to — this sits firmly in speculation.
The next time a banshee crests the cloud and a mountain hangs there where it has no right to be, you can keep the wonder. Just move it to the right place. The marvel was never that the rock breaks the rules. It is that, to hold a mountain in the sky, the universe only needed a rule it already had — and that, somewhere colder and quieter than Pandora, we found the same rule first.


