A Hybrid Quantum Computer Freezes a Particle With a Force It Never Touches

Physicists at the University of Oxford have used a quantum computer built from trapped atomic ions to recreate one of the oldest and strangest predictions in quantum mechanics: that a charged particle can be steered by a magnetic field it never physically enters. The experiment, described this month in Nature Physics, reproduces the effect inside a miniature working model of a lattice gauge theory, the mathematical framework used to describe how matter and force-carrying fields interact at the subatomic scale. Those theories underpin the Standard Model of particle physics but are notoriously difficult to simulate on ordinary computers.

The team, led by researcher Sebastian Saner of Oxford's ion trap quantum computing group in the Department of Physics, working with theorist Alejandro Bermúdez of the Instituto de Física Teórica in Madrid, built what is known as a hybrid qubit-oscillator quantum computer. Rather than encoding every piece of physics into strings of identical qubits, as most quantum computers do, the device uses two different kinds of quantum hardware for two different jobs: qubits, held in the internal electronic states of trapped ions, stand in for the gauge field that carries a force, while quantum oscillators, held in the ions' own vibrational motion, stand in for the matter the field acts on.

A loop built from a handful of ions

To recreate the Aharonov-Bohm effect, the group assembled the simplest possible closed circuit for it: two "matter" oscillators, representing a particle that could sit at either of two locations, connected by two qubits representing the gauge field between them. In the absence of any magnetic flux through that small loop, a simulated particle placed at one site tunneled freely back and forth to the other, exactly as standard quantum mechanics predicts for a particle with two possible paths available to it.

The key design choice, and the one the researchers describe as the most consequential, was to encode the magnetic flux itself as a dynamical quantum state of the two qubits, rather than as a fixed external parameter bolted onto the simulation. The two qubits were prepared in an entangled state representing a fixed unit of flux threading the loop. That let the gauge field and the matter evolve together under the same quantum dynamics, rather than treating the field as a static backdrop — closer to how gauge fields actually behave in nature, where they are not fixed but respond to and shape the matter moving through them.

With flux encoded this way, the result reversed completely. The two possible tunneling paths available to the simulated particle interfered destructively, cancelling each other out, and the particle's tunneling was suppressed entirely: it stayed frozen at its starting site instead of oscillating between the two. That outcome is the real-time, dynamical signature of the Aharonov-Bohm effect, first predicted by Yakir Aharonov and David Bohm in 1959 and long confirmed in electron-optics experiments, but reproduced here for the first time inside an interacting lattice gauge theory rather than free space. According to coverage of the work in ScienceDaily's summary of the findings, the Oxford group began developing the experiment in 2022, and first circulated its results in a preprint posted in mid-2025 before the paper's formal publication this month.

"For us, the exciting step was to encode the magnetic flux in a gauge field that was itself dynamical," Saner said of the design choice that let matter and field evolve together rather than treating the flux as a fixed background.

A problem classical computers cannot crack

Lattice gauge theories describe particles as sitting at the points of a mathematical grid, with the forces between them — the strong and electroweak interactions that bind atomic nuclei and govern particle decays — represented as fields living on the links connecting those points. Solving the real-time dynamics of such theories, how particles actually move and interact moment to moment rather than settling into a static equilibrium, is exponentially hard for conventional computers, which run into a notorious obstacle known as the sign problem when they try to track quantum interference numerically. For decades that has left physicists largely unable to simulate processes such as how matter behaved an instant after the Big Bang or how particles collide inside an accelerator, except through indirect, approximate methods.

Quantum simulators, physical quantum systems purpose-built to mimic another quantum system rather than run generic calculations, have been pursued as a way around that barrier because a quantum device does not have to fight the same interference effects a classical algorithm does. Trapped-ion, superconducting-qubit and neutral-atom platforms have each demonstrated small toy versions of lattice gauge theories over roughly the past decade. The Oxford result adds a hybrid qubit-oscillator approach to that list, and its authors argue it is a more natural fit for theories, including the real Standard Model, in which some of the fundamental objects are naturally wave-like bosons rather than qubit-like two-level systems.

A second group, a coordinated result

The Oxford paper did not appear alone. A separate team at the University of Maryland, led by physicist Norbert Linke, used an analogous hybrid trapped-ion system — qubits paired with the ions' vibrational "phonon" modes — to simulate the dynamics of a Yukawa-type interaction, the kind of force, mediated by a massive boson, originally invoked to explain how protons and neutrons bind inside a nucleus. That companion study, whose methods first circulated among physicists in September 2025, was developed independently of the Oxford work, but the two groups coordinated their submissions so that the papers were published together, underlining that hybrid qubit-oscillator hardware is emerging as a genuine subfield rather than a one-off demonstration. Coverage of the work, including a report on the tunneling-suppression result, noted that the two teams view their platforms as complementary routes toward simulating theories with both matter and force-carrying fields in the same device.

Neither result has an immediate practical application; the simulated system is a single loop, far short of the large two- and three-dimensional lattices needed to represent anything resembling the full Standard Model. The immediate audience is theoretical and experimental physicists working on quantum simulation and lattice field theory, along with the broader quantum-computing field, for whom a working hybrid qubit-oscillator architecture is itself a hardware milestone, independent of which physics problem it is pointed at first. The original method for encoding the flux, the researchers have noted, began as a workaround forced by a limitation of their ion-trap hardware before the team recognized it was the more physically interesting approach to take.

Scaling up from a single loop

The immediate next step for both groups is to add more ions and more plaquettes — the individual loops of the lattice — to move from a single two-site system toward small two-dimensional grids, where qualitatively new physics, such as how gauge fields mediate interactions between multiple separated particles, can appear. Researchers in the field also want to test how such hybrid devices perform as they add noise, more realistic matter fields that carry spin, and eventually the more complicated, non-Abelian gauge symmetries that describe the real strong and electroweak forces, as opposed to the simplified "Z2" toy gauge theory used in this experiment. Whether hybrid qubit-oscillator hardware can be scaled that far without errors overwhelming the signal remains an open, and closely watched, question for the field.