IS it possible to walk through walls? Can solid objects really pass through each other? Moses Chan thinks they can, and he says he has the proof. Chan and his colleagues at Pennsylvania State University have created the world's first "supersolids", bizarre crystals that slide through each other like ghosts. It is a finding that promises to revolutionise the way we think about matter. "It really changes one's concept of solids," says Jason Ho, a solid-state theorist at Ohio State University in Columbus.

The idea that one solid object can flow through another contradicts all our everyday experiences: no one has ever seen a teacup dissolve through a saucer. And when you prop up the bar on Friday nights there is no danger of you slowly melting into the surface and falling out the other side.

Solids get their reassuring rigidness from the orderly way their atoms are arranged. Unlike liquids and gases, the atoms in a crystal are fixed, a bit like the squares on a chequerboard. While they may quiver a little due to their thermal energy, this normally isn't enough to dislodge the atoms and cause them to flow over each other.

However, physicists have long suspected that this rule could be broken. In 1969, Russian theorists Alexander Andreev and Ilya Lifshitz were studying the properties of solids and found that there was indeed a way for them to flow. In certain crystals, the bonds between atoms are so weak that you can squeeze the solids like a sponge. Such weak bonds give a crystal another property. Close to absolute zero its atoms have barely enough thermal energy to shiver. But for some crystals, even this tiny movement is thought to be enough for an atom to break free. This leaves the crystal latticework full of gaps called "zero-point vacancies" and these vacancies are mobile, even at absolute zero.

It might sound odd that a gap in a crystal can have any physical properties. But physicists think of these vacancies as having energy and mass just as atoms do. And they can even move around the crystal. Andreev and Lifshitz predicted even more bizarre behaviour: the vacancies can move through the solid in synchrony as if one solid were flowing through another.

This happens because of the peculiar nature of quantum mechanics. Close to absolute zero, quantum theory says that groups of atoms can lose their individual identities and start behaving like a single, giant atom. Instead of dancing around randomly in a gas or liquid, these atoms can condense into a single quantum state and start moving in perfect lock-step. This happens because their quantum wave function, the region of space in which a particle is found, spreads out and grows larger than the distance between the atoms.

The details of the process depend on a basic property of particles called spin. Atoms divide into two families called bosons and fermions depending on the nature of their spin. Quantum mechanics dictates that no two identical fermions can share the same quantum state, so they cannot move as a single entity. But there is no such restriction on bosons. They can crowd together in the same energy state when the wave functions begin to overlap. It is at this point, they start behaving like one massive atom, dubbed a Bose-Einstein condensate. In principle at least, this means that solids made of bosons can flow through each other.

In many solids this cannot happen because the atoms are locked in their individual fixed positions like people sitting in their assigned seats in a theatre. But Andreev and Lifshitz recognised that atoms have more freedom in crystals that have weak bonds. And they realised that the same thing should apply to vacancies, so they would have the freedom to move around too and condense into one, giant vacancy. Like a ghost walking through a wall, they predicted that this "supersolid" made of vacancies would pass eerily though the rest of the crystal.

Andreev and Lifshitz were thinking of solid helium-4 when they did their calculations because helium-4 atoms are bosons and the bonds between them are so weak. Helium is also the second lightest element in the periodic table, so its quantum behaviour should become more apparent when it cools than with heavier elements. Supersolids should therefore show up in helium and hydrogen first.

But they could form in any solid under the right conditions. "If this is what quantum mechanics tells us about the ground state of solid hydrogen and solid helium, it is true of every solid in the universe," says Philip Anderson of Princeton University, an expert on the quantum theory of solids and winner of the 1977 Nobel prize for physics.

Since Andreev and Lifshitz made their prediction, however, many researchers have looked in vain for supersolids. Yet there are some tantalising hints that supersolids do indeed exist. In the late 1990s, John Goodkind of the University of California in San Diego and his colleagues saw something strange in experiments with crystals made from helium's isotopes helium-3 and helium-4.

Creating helium crystals is not easy. You need more than just freezing temperatures to transform liquid helium-4 into a solid: you also have to crush it to at least 25 times normal atmospheric pressure.

Goodkind's team was bombarding a solid helium-4 crystal with ultrasound. As they cooled the crystal close to absolute zero, the researchers noticed the ultrasound waves speeding up. This could have been down to the formation of a supersolid. When sound travels through a solid, it causes the atoms to vibrate. If part of the crystal had turned into a supersolid, it would have decoupled from the rest of the crystal and allowed the sound waves to travel faster. But the team needed to do more experiments to be sure.

Enter Chan and his graduate student Eunseong Kim. Inspired by Goodkind's results, they decided to take a look for themselves in 2001. They began by putting some solid helium-4 in a bucket, hanging it from a string, and spinning it first clockwise then counter clockwise at 1000 times a second while freezing it. This experiment would enable them to see what happens to solid helium as it oscillates rapidly at very low temperatures.

In reality the device is more complicated, surrounded by tubes and wires that deliver helium and record measurements. The "bucket" is actually smaller than a sewing thimble and rotates on a shaft rather than a string. To compress the helium into a crystal Chan and Kim ratcheted up the pressure to 60 times atmospheric pressure.

Success finally came last year. As they cooled the crystal down below 2 K, the experimenters carefully monitored the rate at which the thimble oscillated. The frequency of this oscillation is governed by the shaft's stiffness and the thimble's inertia, which is determined by the mass of helium inside.

Supersolid debut

At about 0.2 K, the thimble began to oscillate faster, as if some of the helium had escaped from the bucket. Except that no escape was possible since Chan and Kim had already checked for leaks in the system. They concluded that roughly 1 per cent of the solid had stopped moving with the rest of the helium crystal and was standing stock still. The rest of the crystal appeared to be passing through it.

It was an electrifying moment, but Chan and Kim had to make sure that what they were seeing wasn't the result of a flaw in their set-up. So they replaced helium-4 with helium-3, whose atoms are fermions and should not form a supersolid. If they saw no effect with helium-3, they could be confident that the original result was real.

Sure enough, there was no sign of supersolid flow with helium-3. But still the researchers weren't completely sure of what they saw. They published the results in Nature (vol 427, p 225) under the rather hesitant title "Probable observations of a supersolid helium phase".

Chan's fellow physicists were intrigued. "It is hard to believe that a solid can behave that way," says Wayne Saslow, a condensed matter physicist at Texas A&M University in College Station. "If it is right, it means that this particular solid is like a localised fluid rather than a solid."

While their fellow physicists scratched their heads, Kim and Chan did more experiments to check their results. They published their later findings in Science (vol 305, p 1941) but this time they dropped "probable" from the title.

The most logical explanation for his results, thinks Chan, is that 1 per cent of the atoms or vacancies condensed into a single unit that then played by the rules of quantum mechanics while the rest of it continued to live in the classical world.

"You think of solids as very reliable and boring, with all the atoms fixed in their expected positions," says Chan. "We are saying no, there is actually flow within a solid."

Yet Chan's results are proving controversial. While everyone agrees that the experiment was carried out carefully, other groups of researchers are still trying to replicate it. A Japanese group led by Masaru Suzuki of the University of Electro-Communications in Tokyo and Keiya Shirahama at Keio University in Yokohama, has recently conducted a similar experiment to Goodkind's, using sound waves to study solid helium-4. So far they have found no evidence of supersolid-like behaviour.

Meanwhile theorists are arguing about what Chan has really seen. Nikolai Prokof'ev, a theorist at the University of Massachusetts in Amherst and his colleague Boris Svistunov think what Chan spotted were myriad microscopic crystals slipping around in a sea of liquid helium, rather than a supersolid passing through a single crystal. "Think of chunks of packed ice with water in between," says Prokof'ev. Yet Chan argues that such behaviour would have shown up in his experiments.

Theorists are also going back to basics and re-examining Andreev and Lifshitz's theory. Some, like David Ceperley, a theorist at the University of Illinois in Urbana-Champaign, thinks vacancies cannot be present at the high pressures in Chan's experiment. He argues that rising pressure would squeeze the vacancies out of the crystal. He expects them to migrate to the end of the crystal and disappear like bubbles escaping from a fizzy drink.

Others, including Tony Leggett, also at Illinois and winner of the 2003 physics Nobel for his work on superfluid helium, argue that vacancies might not even be needed for supersolid behaviour. Leggett says it might be down to atoms switching places like in a game of musical chairs. Chan is convinced vacancies and other types of crystal defects are involved, but he is leaving it to the theorists to work out the details. "The explanation is probably a bit more subtle than what was proposed by Andreev and Lifshitz," he says.

But many researchers simply want to skip the controversy and go straight to the question of how the stuff would act if it really was a supersolid. One reason they are so keen is that if Chan is correct, that means that quantum mechanics can operate on large scales in all three states of matter - solids, liquids, and gases. "Quantum mechanics is supposedly the theory that explains nature, yet we don't see it existing in our environment," says Seamus Davis, a physicist at Cornell University in Ithaca, New York. "We know of liquids and gases that can be macroscopic quantum systems. It would be very exciting if we now have an example in a solid."

Towards absolute zero

It is too early to say what supersolids could be used for. But the very fact that Chan may have uncovered an entirely new state of matter is reason enough to get excited. So could supersolids turn up in any other elements? In March, Chan and his students Anthony Clark and Xi Lin reported similar behaviour in solid hydrogen at even lower temperatures. Hydrogen molecules are bosons, so the result ties in nicely with their previous experiments. And Chan is confident that other elements could follow.

Researchers have already coaxed over half a dozen different gases into Bose-Einstein condensates using special traps that employ magnetic fields and lasers to hold and cool the atoms. And it is possible to make superfluids from fermions, such as helium-3. Here the atoms first pair up to form bosons before crowding into the same quantum state. So who knows what elements might follow next.

But don't get too excited just yet: other supersolids may be hard to detect. As you move down the periodic table, the atoms act less like quantum particles because they are more massive and bond more strongly to neighbouring atoms. So the supersolid effect is likely to be smaller. To see it, you would need extremely low temperatures, perhaps as low as a few thousandths or a few millionths of a degree above absolute zero. Yet it is still possible. "It is just that the supersolid fraction will be smaller and we will have a difficult time measuring it," says Anderson.

So it is unlikely that the supersolid phase will help you walk through walls, at least not yet. But studying supersolid behaviour could usher in a new way of thinking about solids, says Anderson. The textbook picture of a solid as a chequerboard array of atoms with one atom per space could be replaced by something a whole lot weirder. "It really shakes your confidence in what a solid is," says Anderson.