Oxford physicists reproduce the Aharonov-Bohm effect on a single loop of a simulated gauge theory

In a lattice gauge theory, matter sits on the points of a grid and the fields sit on the links between them ^[6]. Oxford's design put both on the same trapped ions ^[9]. One loop of the grid, two matter sites joined by two links, is the theory's elementary building block ^[10], so the entire simulation ran on four encoded components ^[18]. The work dates to 2022, when lead author Sebastian Saner, Oana Bazavan and colleagues at Oxford began building it with Alejandro Bermudez of the Instituto de Fisica Teorica in Madrid ^[8].

The experiment has its own control. The same loop was run with and without the flux ^[15]. Without it, a matter particle tunneled freely around the loop. With it, the two paths interfered destructively and the system stayed frozen in its starting state, according to Phys.org ^[15]. That suppression is the Aharonov-Bohm signature: a phase picked up by going around a flux without ever entering a region of magnetic field ^[4]. Aharonov and Bohm predicted it in 1959, and experiments with real electrons later confirmed it ^[5].

The interesting physics came out of a hardware limit. Producing the flux as a fixed background would have needed an interaction the hardware could not provide, so the team prepared the two qubits in an entangled state that the theory reads as a flux piercing the loop ^[11][12]. "For us, the exciting step was to encode the magnetic flux in a gauge field that was itself dynamical," Saner said ^[13]. "Rather than having matter evolve in a fixed background, the matter and gauge field become part of the same quantum dynamics," he said ^[14].

Phys.org presents the Nature Physics result as showing how hybrid machines can simulate matter-field interactions that become increasingly hard to model on classical computers ^[3][1]. In the same account, that difficulty arrives as the systems grow ^[7]. This system is the smallest the theory has ^[10]. We think the paper establishes two things: the encoding works, and a dynamical field produces the predicted interference. Whether the method holds on lattices large enough to strain a classical computer is untested here ^[10][7]. The account does not report error bars, or how close the measured state came to the complete suppression it describes ^[15].

The Oxford paper runs alongside one from a University of Maryland group led by Norbert Linke. That group simulated the Yukawa potential, an interaction that bears on nuclear and particle physics, on a hybrid quantum system ^[16]. In our view the pairing is better evidence for hybrid hardware than either result alone. The two groups developed their approaches independently before coordinating their submissions to Nature Physics ^[17].