They say nothing lasts forever – but, hey, what do they know? Sure, the passing years soften the carvings on statues, pigments on paintings flake with age, and even the most formidable fortresses collapse eventually. But those are all features of the human-scale world. Things tend to work a touch differently in the quantum realm.

For nigh on 70 years, physicists have been chasing the dream of quantum eternity: an arrangement of atoms positioned so that the quantum states between them are frozen forever, like light bouncing in a never-ending hall of mirrors. Proving that such a thing could exist wouldn’t just be an incredible scientific milestone, it would also be very handy indeed. Quantum states that last forever – or even just for a very long time – could enable us to create completely new states of matter, some of which could be the basis of powerful new quantum computers. “It would open up a whole new class of phases that are otherwise impossible,” says mathematical physicist Wojciech De Roeck at KU Leuven in Belgium.

There is, however, a good reason why eternity has always seemed elusive. Thermodynamics, one of the central pillars of modern theoretical physics, insists that the fine details of things are always, eventually, smudged away. And until recently, physicists’ efforts to study quantum eternity had only served to underscore this seemingly unavoidable truth. Now, however, things seem to be changing and powerful experiments are hinting that eternity may not be out of reach.

One of the rules that governs reality is that things tend to get messier over time unless energy is expended to intervene. This is a truism in life generally, but it is also a key assumption that underpins thermodynamics, the physics of heat, work and energy. It explains why pouring milk into coffee turns our cup a creamy, uniform beige, even without stirring. More broadly, the theory says that all systems eventually thermalise, meaning the different parts of things mix into averages.

Based on this, you would think that nothing could endure forever. But in 1958, physicist Philip Anderson suggested a striking possible exception in the world of materials. To get your head around it, picture the inside of a material as a grid of different kinds of atoms that can be more or less ordered. Crystals, for example, tend to have a strictly repeating three-dimensional pattern of atoms, whereas in other materials, such as glass, the atoms are highly disordered. Materials can also have particles and waves, like electrons or light, move through them.

Anderson imagined taking a crystal and introducing some disorder – impurities, say, or atoms knocked slightly out of place. He thought there would be certain such arrangements in which an electron travelling as a wave would scatter off the disorder again and again. Those scattered ripples would cancel out so completely that the electron couldn’t get anywhere. The particle would be trapped, a quantum state frozen in time, at least until the material itself ceased to exist. Anderson theoretically showed that this could happen, winning himself a share of a Nobel prize in physics in 1977. Decades later, experiments proved him right, albeit in simplified sets of atoms where the particles didn’t tug, shove and exchange energy as they would in the real world.

Yet Anderson speculated that his effect would occur in the messy world of real materials, where particles very much do shove and tug on one another. The idea that a trapped particle could avoid being jostled loose and instead enter a frozen state became known as many-body localisation (MBL). “Chaos should be everywhere. It should be more or less inescapable,” says De Roeck. “However, these many-body-localised systems do not exhibit chaos.”

The seductive promise of eternity caught the attention of generations of physicists, but it wasn’t until 2006 that we got the first serious hint that MBL might really be possible. Physicists Denis Basko at Princeton University and Igor Aleiner and Boris Altshuler, both at Columbia University in New York, built a mathematical description of a conducting material that electrons can glide through easily. They then proved mathematically that changing the structure of the material so it was more disordered could trap those electrons, creating an MBL and transforming the conductor into an electrical insulator.

This wasn’t an experimental demonstration, but the star of MBL still rose. More importantly, the work suggested that if we could conjure up an MBL for real, it would fundamentally change the properties of materials – perhaps even generate whole new states of matter. One example is the “time crystal” originally envisaged in 2012 by physicist Frank Wilczek, but there are several others that might prove useful for all sorts of technologies, from quantum information storage to super-precise clocks (see “Strange states”).

The thing about these states, though, is that although many have been glimpsed, they don’t last long. For them to endure, we would need to fully realise an MBL. “So you have all these avenues that have already been explored, but there is a doubt hanging over them,” says mathematical physicist François Huveneers at King’s College London.

Thermal avalanches

In the past decade, doubts over MBL have only deepened as a result of findings that touch on two fundamental objections to the idea. The first has to do with scale: you might coax MBL in a tiny corner of a material, but thermodynamics already allows for its rules to be broken in tiny patches, just not overall. For an MBL to defy thermodynamics, it would have to survive throughout a large piece of material, and there is huge uncertainty over whether that is possible. The second objection is about time: you might watch a localisation for 10 minutes, an hour, a day – but can you guarantee it wouldn’t vanish if you watched a second longer?

In 2018, De Roeck and Huveneers made a discovery concerning the scale side of things. To get MBL, disorder has to permeate a material. But they pointed out that, in reality, there will always be small, unusually neat patches where particles don’t freeze, but remain free to move and share energy around. De Roeck and Huveneers then mathematically proved that the neat patches could feed energy into the frozen regions, disrupting the MBL. This effect would rapidly spread until no frozen quantum states were left. They named this mechanism a thermal avalanche, and it seemed to cast grave doubts on whether MBL could survive in real materials.

Then there is the separate issue of time. Could there be MBL-disrupting mechanisms that only emerge if they are left long enough? In 2024, Nicolas LaFlorencie, Fabien Alet and Jeanne Colbois at the University of Toulouse, France, identified a phenomenon known as resonances that could fit the bill. Atoms in a material with MBL usually stay locked in a specific set of quantum properties. Even so, this lock is rarely perfect – those properties still vary ever so slightly over time, and in doing so, the material could stumble on a completely different arrangement that happens to have exactly the same energy as its starting point. When that happens, the two states can resonate or meld together, and this process can undermine the pristine MBL regions.

Testing eternity

It might seem like all of this hacks away at the foundations of quantum eternity. But experiments from the past few years are now beginning to tell a more hopeful story. Much of what physicists know about MBL comes from computer models, but such models can only get so big before they get too complex to handle. The obvious workaround would be to carry out real experiments, but this has always been out of reach. To test MBL properly, physicists need to prepare delicate quantum states, subject them to precisely controlled disorder and observe their evolution without washing the effect away. It is only recently that experiments with ultracold atoms, trapped ions and superconducting qubits, the building blocks of quantum computers, have become precise enough for the job.

That shift is beginning to reshape the field. Take, for example, a 2025 study led by Junhyeok Hur at the Korea Advanced Institute of Science and Technology, which probed MBL in real arrays of ultracold atoms that were as large as a 24-by-24 grid. “This is an experiment on a system that has a size and timescale larger than what we can achieve in simulations,” says Alet, who wasn’t involved in the work. Alet and LaFlorencie are both working at the cutting edge of computer simulations, but these begin to struggle when they get up to around two dozen particles.

Hur and his team compared two kinds of disorder that could affect their array. In one, each atom was assigned a completely random energy. This produced a mottled landscape: regions of strong disorder interspersed with unusually clean stretches, precisely the kind of rare patches thought to act as seeds of thermalisation. In the other, the disorder followed a quasi-random pattern, meaning it was completely free of any ordered patches.

As the team tested successively larger arrays, the contrast between the two kinds of disorder became stark. With random disorder, larger systems needed ever-stronger disorder to remain localised – a sign that MBL would inevitably get washed out as the systems grew. But for quasi-periodic disorder, the threshold barely shifted as the simulations got larger. The implication is that structuring a material’s disorder could help stabilise MBLs as scales get larger, says Alet.

Amos Chan at the University of Warwick, UK, one of those who worked with Hur on the study, says it isn’t yet clear what drives thermalisation in the randomly disordered case, whether avalanches, resonances or something else. But newer data that the researchers have measured since their paper only builds confidence that localisation can persist in two-dimensional arrays when disorder is carefully controlled, he says. Though it isn’t definitive proof of MBL, it shows why this new generation of experimental quantum systems matters: they can test whether MBL survives the thing that may destroy it – scale.

A second experiment, published soon after Hur and his team’s by Google’s quantum computing group, looked at systems with up to 70 superconducting qubits. They found that, at moderate disorder, the system settled into a strange in-between state – not quite MBL, but something that resisted thermalisation nonetheless, known as a quantum glass.

Other physicists are working on another line of attack that might allow us to finally pin down the existence of MBL states. To understand this, it helps to think of how we define states of matter. No one proves that a lump of iron is a magnet by watching every atom in it for the rest of time. Instead, we look at a property of atoms in a material called quantum mechanical spin. When enough of these spins align, we say a material has “changed phase” and is a magnet. Recently, Alet, LaFlorencie and others have been trying to find a similar way to define an MBL.

To do so, they enter the realm of abstract quantum states. They consider every quantum property of each atom in a material and plot these over a multidimensional map. At any given moment, the material’s overall state would be represented by one point on this map and, over time, it would vary, exploring the terrain. But, crucially, Alet and LaFlorencie found in 2019 that a material exhibiting MBL wouldn’t do this. It would be confined to patches of the map, frozen in little islands with no way to explore beyond them. This behaviour is known as multifractality, and if a material behaved in this way, it would signify that it was exhibiting an MBL state.

Since 2019, theorists have worked to turn this idea into something measurable. In 2025, a paper by David Logan at the Tata Institute of Fundamental Research in Mumbai and Sthitadhi Roy at the University of Oxford set out one possible test in a system of quantum spins. Prepare the spins in a simple pattern, such as alternating up and down. Let the system evolve, then ask how much of that pattern is still visible later on. In an ordinary thermalising system, the pattern should wash away, but if MBL is there, some of the pattern should stubbornly survive. “Multifractality is appealing in that it’s rather directly connected to what experiments would be able to look at,” says theorist David Huse at Princeton University. LaFlorencie says his group is beginning to search for this fingerprint this year using ultracold atoms. “The idea is to begin probing experimentally this multifractal dimension,” he says. “It’s very exciting.”

For his part, Huse still doubts these experiments will be the be-all and end-all. “We’re probably still waiting on big mathematical physics theorems that finally resolve all these disagreements,” he says. But perhaps that’s no big surprise. No one said the road to quantum eternity would be short.