The Geometry of Quantum Memory: How Anyons Rewrite the Rules of Remembering and Computation in the Quantum Realm

Insider BriefAn ordinary qubit remembers too much and listens too well. Its delicate phase is the hidden part of a qubit’s state. It carries much of the qubit’s information and is very easy to disturb. The phase takes account of every stray field, lattice vibration, and fluctuating charge. In turn, a qubit lives a fragile existence and decoheres due to unwanted interactions with its environment.But what if we could curb a qubit’s ability to listen to the environment, giving it a selective “amnesia”? What if a qubit could store information somewhere local disturbances cannot see? And what if its operations depended on how particles are tangled, rather than exactly how they move?This is the core idea behind topological quantum computation and it all begins with the two-dimensional loophole that brings to life the strange particles at its heart.Topology is the study of the properties of shapes and spaces that survive smooth, local deformations.Take, for example, a donut and a coffee mug. They look different and serve completely different purposes, but topologically they are equivalent: each has a single “hole.”A bite out of the donut or a chip in the mug changes the object “locally.” Its overall topology remains unchanged. Cutting the donut in half or breaking the mug’s handle, however, removes the hole and permanently alters its topology.Reality is full of local disturbances. To store and process quantum information safely, we must block outside noise and ensure every operation performs exactly as intended. This is one of the central challenges of quantum computation.A big part of the field has evolved to address this. Quantum error correction, for one, is an algorithmic way to deal with errors. It spreads information across many physical qubits, the individual pieces of hardware. Together, they act as a single, more reliable, logical qubit. Meanwhile, repeated checks catch errors in any single member. With today’s hardware, one high-quality logical qubit can take around a thousand physical qubits, which is costly to scale.Now, imagine storing quantum information in the overall topology of something rather than in its fine details, and computing in a way that depends only on that topology. Like the donut and the mug, such a system would shrug off small, local disturbances.Topological quantum computation aims to turn the situation around. It does so by designing each physical qubit to be far more robust from the start. With stability built into the hardware, each logical qubit would need far fewer physical qubits.So, what does a topological qubit look like? The story of the topological qubit begins with a simple rule: the local physics of a system should remain unchanged when two identical particles swap places.In quantum mechanics, identical particles fall into one of two families: bosons and fermions.Exchange two bosons and the wavefunction describing the system stays the same; exchange two fermions and it flips sign.In 1977, Jon Magne Leinaas and Jan Myrheim from the University of Oslo made a key discovery. They showed that this familiar rule depends on the number of dimensions. In a two-dimensional plane, swapping two particles could do more than the usual. Instead of the fixed +1 and −1 outcomes of bosons and fermions, their phase could take any value.The reason lies in how particles can move around one another in two and three dimensions. Imagine two particles sitting in space. Take one and walk it in a loop around the other, back to where it started. That loop is like a piece of string tied in a circle. Swapping these two identical particles twice would then equal taking one of them on a full loop around the other.In three-dimensional space, you can reshape this loop smoothly by stretching, shrinking, or sliding it. You never need to cut it or pass it through the other particle. For example, you can lift it up and slide it over the top of the other particle, like lifting a hula hoop off someone’s head. Once it’s off, you can shrink it to a point. So, in 3D, circling the other particle is topologically the same as doing nothing.On the other hand, suppose two particles could move only in a 2D space, like a flat sheet. There would then be no “up” to lift into. The loop around the other particle stays stuck, unless you cut through the particle. So, in 2D, circling a particle is genuinely different from doing nothing, and the system can “remember” that it happened.That memory is what allows a swap in two dimensions to leave behind any phase. In 1982, American theoretical physicist Frank Wilczek, showed how such particles might arise and gave them a name: anyons.In real materials, anyons appear as quasiparticles: collective behaviors of many electrons that act like single particles.Leinaas and Myrheim showed that mathematics allows anyons but not what one might look like. Wilczek supplied a picture with a toy model. It was a composite particle built from a magnetic flux confined inside a small solenoid with an electric charge circluing around it. Move one of these composite particles in a loop around another. Its charge then travels around the other’s flux, and vice versa.Quantum mechanics has a strange feature here, known as the Aharonov–Bohm effect. Here, a charged particle passing around a magnetic field picks up a subtle imprint on its quantum state. This happens even though it never touches the field. The imprint depends only on whether the particle went around. It does not depend on the exact route, how far away it was, or how fast it moved. By adjusting the charge and field, that imprint can take any value, exactly the behavior Leinaas and Myrheim had allowed. Wilczek’s model was a thought experiment, not a recipe. It showed that nothing in physics forbids anyons, but pointed to no real material containing them.A real home soon appeared in the fractional quantum Hall effect, discovered experimentally in 1982. In 1983, Robert Laughlin, then a physicist at Lawrence Livermore National Laboratory in California, explained how electrons in a thin layer under a strong magnetic field act collectively to form excitations carrying a fraction of an electron’s charge. In 1984, Wilczek and two colleagues at Santa Barbara’s Institute for Theoretical Physics, Daniel Arovas and Robert Schrieffer, calculated that these excitations should behave as anyons.Anyons had gone from a possibility to an expected feature of nature, though still only a prediction. So far, every anyon proposed was of one type, now called Abelian. Swapping two leaves an imprint on their shared quantum state, but it remains the same state. The order of swaps doesn’t matter, just as adding numbers gives the same total in any order.Then, in 1991, Gregory Moore and Nicholas Read, two theoretical physicists at Yale University, proposed something stranger. Their theoretical model showed how electrons in such a quantum Hall system might organize into a new collective state.Its excitations, they showed, would be a new kind of anyon, now called non-Abelian. A group of them can share several possible collective states. Which one they are in belongs to the group as a whole, not to any single anyon. Swapping two of them doesn’t just leave an imprint; it moves the group from one collective state to another. Like turns on a Rubik’s cube, the order now matters. The same moves in a different order give a different result.Step back, and the arc is clear. In three dimensions, a swap only changes who is where, the paths they took leave no trace. Flatten the world to two dimensions and the paths start to matter. Particles can now wind around one another, and the system remembers it. Mathematicians describe this difference as the step from the permutation group, which tracks only final positions, to the braid group, which tracks how the paths were woven together.Abelian anyons use part of that memory. Each braid leaves a single imprint, and the order of swaps makes no difference. As a result, much of the braid’s structure is lost. Non-Abelian anyons use all of it. Their collective states store information where no local disturbance can reach it. Every braid, its direction and its order, acts as a distinct operation on that information. A sequence of braids becomes a calculation. Because the result depends only on how the strands are woven, not the exact route each took, small wobbles along the way don’t spoil the answer.Russian theoretical physicist Alexei Kitaev connected these ideas to comuputing in a 1997 preprint, published in 2003. He saw that certain quantum error-correcting codes behave like spin systems whose excitations are anyons. He also saw that moving those anyons around one another performs operations that depend only on topology. His toric code is the forerunner of today’s surface codes. Its anyons are Abelian, so they protect stored information but cannot compute by braiding alone. For that, Kitaev pointed to the non-Abelian kind.The key word is local. A topological qubit is not immune to everything. In principle, it is blind only to disturbances that act in a small region. Its information sits in a global property shared by widely separated quasiparticles. Only winding those particles around one another can change it. The windings are recorded as a braid in spacetime. Any any two braids that can be smoothly deformed into each other perform the same operation. Such a qubit has amnesia for local detail and perfect recall for global topology.Whether nature will supply such qubits is still open. In 2020, experiments directly observed the exchange behavior of Abelian anyons. Later, in 2023, researchers engineered non-Abelian anyons on conventional quantum processors, though without built-in protection. We are yet to conclusively observe non-Abelian anyons in real materials. That search remains among the most ambitious bets in quantum computing.TopicsShareGet the latest research, company news, and market intelligence every week.MENTIONED IN THE ARTICLEMore in Education