The Oscillation That Never Stops: The Time Crystal

Bölüm metni

If you look at a salt crystal you see the atoms repeating at fixed intervals in space. But can a piece of matter do the same thing in time — can it oscillate in the same rhythm forever, consuming no energy at all? The physicist who asked that question proved his own answer impossible; then others shifted it one step sideways and realised it in the laboratory.

Take a single grain of salt and blow it up large enough, and what you see has a patience about it that is almost unsettling. Sodium and chlorine atoms are laid out like a three-dimensional chessboard, in rows that follow one another without ever losing step. The distance from one atom to its neighbour is always the same; measure it once and you have measured the distance at the far end of the grain as well. And yet this arrangement is written down nowhere. There is no such plan in the water the salt was dissolved in; there every point is like every other, no position enjoys any privilege. As the water evaporates, the atoms pick a place. The moment they pick it, the uniformity of space is broken: from then on, here and half a nanometre away are different from each other.

Physicists call this spontaneous symmetry breaking, and the word spontaneous carries the whole weight of the thing. The laws of nature favour no point in space. But matter does. The law stays neutral while matter takes sides, picks a pattern and locks onto it. That is exactly what a crystal is: an order that repeats itself through space and that nobody dictated. It also has a measurable signature. Shine X-rays on a crystal and on the detector behind it you see not a blurred smudge but needle-sharp spots. That sharpness is the proof of how meticulous the repetition is.

In the autumn of 2012, Frank Wilczek, working at the Massachusetts Institute of Technology and one of the winners of the 2004 Nobel Prize in Physics, put an uncomfortable question to this very familiar picture. Ever since relativity we have not thought of time as a separate ledger from space; the two are threads of the same fabric. And just as the laws favour no point in space, they favour no moment in time. An equation that holds today holds tomorrow. What this symmetry corresponds to in physics is nothing ordinary: it is the conservation of energy itself. Wilczek's question was this. If matter can break the uniformity of space on its own, why should it not break the uniformity of time? Could a piece of matter, by itself, keep a rhythm in time the way it keeps a pattern in space?

Something has to be separated out here, because the description of a thing that repeats in time sounds boringly ordinary on first hearing. A pendulum repeats. A heart beats. A planet turns. But all of these were either set going by somebody's first push or are slowly burning through a store of energy; leave them alone and they all stop. What Wilczek wanted was far harder. Imagine a piece of matter that has sunk to the lowest state physics allows, the lowest energy state. Nothing left to give, nowhere left to be cooled to. And this matter still does not stop. It oscillates. Its motion, in other words, is not the last writhing of some leftover energy; it is stillness itself. For such matter the calmest possible condition is not to sit motionless, but to stir with a definite period.

This sounds like something forbidden, and it matters to understand why it was thought forbidden. It has nothing to do with perpetual motion machines. You cannot draw energy out of such matter; you cannot run a wire from it and light a bulb. Nothing flows out, nothing is used up inside. In fact we already know something that resembles this from a distance: the current circulating in a superconducting ring goes round, with no battery at all, in principle forever. What Wilczek proposed was to carry that idea one step further. His concrete example was close to it as well, charged particles confined to a ring with magnetic flux threading through it, a system whose lowest-energy arrangement is not sitting still but travelling around the ring. Together with Alfred Shapere he also designed a cousin of this in classical physics: a system in which having zero speed at minimum energy is forbidden.

The name of the idea was irresistible. Time crystal. But physicists are drawn less to beautiful names than to ideas that look like contraband. And this idea looked exactly like contraband. If there is a rhythm even in the lowest state, that rhythm is a clock. And a clock nobody wound and nobody feeds is the kind of thing that makes a physicist's nose itch.

The most reliable sign that an idea is being taken seriously is that someone sets out to demolish it. The time crystal sat that examination both quickly and harshly.

The first blow came in 2013. Patrick Bruno, working at the European Synchrotron Radiation Facility in Grenoble, went after the proposed ring model and showed that the state with the circulating current was not in fact the lowest-energy state. Then he generalised his objection: the ground state of a system cannot rotate of its own accord. But the real blow came in 2015, from Haruki Watanabe and Masaki Oshikawa at the University of Tokyo, and this time the objection was aimed not at a single model but at the definition of the idea. The heart of the theorem they proved will feel maddeningly familiar to anyone who knows a little quantum mechanics: the ground state is an energy eigenstate, and the measurable properties of an energy eigenstate do not change with time. It is stationary because it is stationary. The same holds for a system in thermal equilibrium. So in a closed system, under time-independent laws, and provided the interactions are not very long-ranged, you cannot see a persistent oscillation in equilibrium spread across the whole system. Measure what you like, what you get is a flat line.

Within three years the idea seemed to have been buried by a sentence sharper than the one that gave birth to it. Impossibility theorems usually work like this in physics; they close a door. But read carefully, they also tell you which door they closed. Watanabe and Oshikawa's theorem did not say that the idea was silly. It said that it could not happen in equilibrium. And equilibrium is a very small room in the universe. Everything we live through, from our cells to the stars, happens outside that room.

The way around the ban came out of this. What happens if, instead of leaving the system alone, you start prodding it without pause? You send it laser or microwave pulses at a definite period, at regular intervals. Such a system is no longer in its ground state, so it has fallen outside the scope of the theorem. But something more interesting also happens: the symmetry to be broken changes. You no longer have the unbroken uniformity of time; the laws now repeat themselves only once every period. And breaking that discrete symmetry means something very precise. You prod the system a thousand times a second, and it answers five hundred times. It settles into a rhythm at twice the period of the drive.

And here the resemblance to space becomes tighter than it was at the start. A crystal takes continuous space and leaves a discrete lattice behind it. A driven time crystal takes the discrete tempo of the drive and leaves a coarser tempo behind it.

But there were two serious obstacles. The first is this. Doubling a period is not a rare event. A child on a swing, a dripping tap, certain electronic circuits all do it. What it takes for something to count as a phase of matter is rigidity. Shift the period of the drive a little, make the pulses imperfect, and an ordinary period doubler will either drift or start following you. A time crystal locks on; even if you move the parameters within a certain range, it goes on standing at exactly twice. That is what separates a coincidence from an order. The second obstacle was more insidious. If you prod a piece of matter without pause, you are pumping energy into it without pause, and a many-particle system swallows that energy, heats up, and in the end turns into an indescribable mush at what is effectively infinite temperature. No order survives there.

The rescue came from a phenomenon known as many-body localisation. If there is enough disorder and the right kind of interaction, the system cannot spread energy around inside itself. It refuses to heat up. In 2016, Vedika Khemani and three colleagues worked out the phase diagram of a disordered, periodically driven spin chain and found a phase in which the spins stubbornly turned in a period-doubled pattern. Another team that same year gave this state the name Floquet time crystal. Now there was a recipe.

The recipe was turned into a meal in two separate laboratories in March 2017, and the two were published side by side in the same issue of the same journal. At the University of Maryland, Christopher Monroe's team used 10 ytterbium ions lined up in an electromagnetic trap. With alternating laser pulses they flipped the spins and forced them to interact with one another, added disorder on purpose, and saw the system's magnetisation oscillating at twice the period of the drive, staying locked to that tempo even when they shifted the drive slightly. At Harvard University, Mikhail Lukin's team caught the same signature in an entirely different setting, in the crowded and disordered world of nitrogen-vacancy defects scattered through a diamond lattice, using microwaves. The same order, in two systems, one a chain of ions hanging almost in vacuum, the other a solid stone.

Even so, the case was not closed. These oscillations lived only a few dozen cycles, which is to say milliseconds. The critics' objection was fair: perhaps these were not a real phase, but an intermediate state stuck for a long while on the way to thermal equilibrium. You cannot prove that something lasts forever with an experiment that lasts a millisecond.

The step that largely closed the argument came in November 2021. Teams from Stanford, Google and Max Planck used 20 qubits as spins on Google's quantum processor, Sycamore. The advantage of a programmable machine was this: you can start the system from hundreds of different initial states, sweep the parameters one by one, and use echo techniques to separate the hardware's own decay from the physics itself. What they showed was that this order was not the trick of one lucky initial state, but was soaked through the system's entire energy spectrum. Even so, the lifetime was limited by the machine's own decay time; hundreds of cycles.

So at the end of 2021 this is what was in hand. A way around the theorem had been found, the signature had been seen on very different platforms, most of the objections had been answered. The remaining problem was embarrassingly simple. These things lived a very short time. A millisecond can be a lifetime for physics; for a material it is nothing.

That wall came down in February 2024, and by several times over. At TU Dortmund, Alex Greilich's team was working on a special semiconductor made of indium gallium arsenide. They illuminated the sample with continuous light. The light was aligning the electron spins, and the electrons were passing that alignment on to the spins of the atomic nuclei around them. The nuclear spins thus became a kind of reservoir, and this electron and nucleus pair began to oscillate on its own, with nobody dictating a period to it. The system chose the frequency.

This was a different kind of thing from the earlier experiments. In driven time crystals the rhythm is half the tempo coming from outside; you set the beat and the matter halves it. Here there was no tempo coming from outside, only uninterrupted light; the symmetry being broken was not the discrete but the continuous uniformity of time. This is called a continuous time crystal. The system is open, energy enters as light and leaves as heat, so it too is outside the room of the old theorem. And the lifetime: at least 40 minutes. About 10 million times longer than anything shown until that day. What is more, 40 minutes was not where the crystal ended but where the measurement ended; it could probably have gone on much longer. When the lifetime of something goes from milliseconds to minutes, the question being asked changes too. It is no longer whether the thing exists, but what it is good for.

The second threshold was about dimension, and it was crossed in January 2026. A team led by Eric Switzer observed a two-dimensional discrete time crystal, using 144 qubits of IBM's 156-qubit Heron r2 processor in the shape of a decorated hexagonal lattice. To see why this matters you have to go back to that heating problem. It was seriously doubtful whether many-body localisation, which saves the day in one dimension, could survive in two; in two dimensions there are far more routes by which energy can escape and disperse. The team did not only see the oscillation, they mapped the entire phase diagram: a spin glass region on one side, an ergodic region where everything races to equilibrium on the other, and the time crystal phase in between. They compared their results with tensor network calculations. So this order is not a by-product of a one-dimensional model.

And in March 2026, 14 years after the idea was born, the first real job arrived. Ashok Ajoy of the University of California, Berkeley, Paul Schindler of the Max Planck Institute for the Physics of Complex Systems and their colleagues turned a time crystal, built from carbon nuclear spins inside diamond, into a measuring instrument. The logic is this. If the strongest feature of this phase is its rigidity, then use the rigidity. You place the crystal inside a weak, oscillating magnetic field. When the frequency of the field matches the crystal's own frequency the crystal is not disrupted; on the contrary, its lifetime grows longer, it makes food of the signal. The result is a very narrow-band detector that burns only at the frequency it is tuned to. The range it works in is between half a kilohertz and 50 kilohertz. That band is precisely the awkward region for many other quantum sensors. And what sets the sensitivity is not the fine tuning of the interactions but the lifetime of the crystal; and it is robust against imperfections in the pulses and inconsistencies in the sample, because the order itself is rigid. The thing that was considered a nuisance for years, the crowd interacting with itself, here becomes the operating principle of the instrument.

A similar door was opened at Aalto University in Finland. Jere Mäkinen's team sent radio waves into superfluid helium-3 at temperatures very close to absolute zero, creating excitations called magnons; these excitations organised themselves and set up a time crystal. What the team did that was new was to connect this crystal to a mechanical oscillator, making it talk for the first time with an apparatus outside itself, and to tune its properties by that route. The analogy they draw is with the optomechanical techniques used in gravitational wave detectors. Sensitive detectors and quantum memory are the possibilities this work points towards; not devices already built, but roads opened.

There is also an example you can see. In September 2025, Hanqing Zhao and Ivan Smalyukh of the University of Colorado Boulder showed that when liquid crystals squeezed between glass plates coated with dye molecules are driven with light, thousands of defects that behave like particles form, and their pattern repeats for hours. This pattern can be seen with a microscope, and under the right conditions even with the naked eye. The uses the researchers propose are light-activated anti-counterfeiting marks and data storage; these are at the level of proposals for now.

The result that best describes where the field stands today comes from this very month. The Dortmund team, in a gallium arsenide sample doped with indium and silicon, at around minus 270 degrees Celsius, set up separate time crystals; at the centre of each is a localised electron and roughly 1 million nuclear spins talking to it. What they see is this. Two of these oscillators lock to the same frequency despite being as much as 40 micrometres apart. That distance is more than a thousand times the length of a single oscillator. The bond between them is not mechanical; it is carried by the drift through the crystal of electrons whose spins are aligned. The researchers draw the comparison themselves: in 1665 Christiaan Huygens noticed that two pendulum clocks hanging from the same beam fell into step with each other.

In this 14-year story nobody refuted a theorem. Watanabe and Oshikawa's proof still stands; rather than demolish it, physicists changed the room, stepping out of equilibrium into driven and open systems. Nor does free energy come out of a time crystal; they all eat either light or pulses. What comes out is subtler. It is that, just as there is repetition in space, a repetition in time can be chosen spontaneously by matter, which is to say that there is one more kind of order in the world. Whether what follows turns into real laboratory instruments or stays a beautiful demonstration is still an open question. But what is argued about now is no longer whether matter that keeps a beat when nobody is counting is possible; it is what we are going to lock it to.

Nöron Science'de bugünkü bölümün sonuna geldik. Yeni bir hikâyede yeniden buluşmak üzere.