ETH Zurich Geometric-Phase Swap Gates Achieve 99.91% Fidelity Across 17,000 Neutral-Atom Qubit Pairs

Overview

A team at ETH Zurich has demonstrated a new class of quantum gate for neutral-atom systems that operates with 99.91% fidelity across 17,000 qubit pairs simultaneously, in under a millisecond. The research, published April 8, 2026 in Nature under the title “Protected quantum gates using qubit doublons in dynamical optical lattices,” uses geometric phases rather than the conventional dynamical phase approach, producing swap gates that are substantially more resistant to the laser noise and intensity fluctuations that have long degraded neutral-atom gate performance.

The lead researcher is Professor Tilman Esslinger of ETH Zurich’s Institute for Quantum Electronics. Key contributors include postdoctoral researcher Yann Kiefer and junior group leader Konrad Viebahn.

What We Know

Neutral-atom quantum computers trap individual atoms — in this case, ultracold fermionic potassium atoms — inside optical lattices, which are artificial crystals of light formed by intersecting laser beams. Swap gates are among the most fundamental operations in such systems: they exchange the quantum states of neighboring qubits, enabling information to be routed through the processor.

Previous swap gate designs relied on the tunnel effect, a quantum-mechanical phenomenon in which atoms pass through energy barriers. As Phys.org notes, this approach is “extremely sensitive to laser light intensity,” meaning that even small fluctuations in beam power can cause errors. The ETH Zurich team instead built swap gates around geometric phases — a property of quantum systems where the phase acquired by a particle depends on the overall path it traces through the lattice rather than on the speed or intensity of the driving fields.

According to The Quantum Insider, junior group leader Konrad Viebahn explained: “Unlike dynamical phases, this geometric phase is largely independent of the speed with which we manipulate the atoms.” That independence is what gives the gates their noise resistance: because the geometric phase is determined by topology rather than by precise laser timing or power, the gate outcome does not drift when experimental conditions fluctuate.

As BioEngineer.org summarizes it: “Unlike dynamical phases, which are highly sensitive to operational speed and laser stability, the geometric phase acquired is remarkably robust.”

The team demonstrated two gate variants. Full swap gates exchange the complete quantum state between two qubit sites. Half-swap gates perform a partial exchange that simultaneously generates quantum entanglement between the two atoms, an operation relevant to quantum communication and measurement-based computing protocols. Both variants were realized in the same optical lattice apparatus.

The scale of the demonstration is notable. The ETH Zurich apparatus ran swap gates on 17,000 qubit pairs in parallel, all within a single sub-millisecond operation, at 99.91% fidelity. According to BioEngineer.org, “The ability to perform synchronized swap gates on this many qubits opens promising pathways for constructing large-scale quantum processors.”

Postdoctoral researcher Yann Kiefer offered a plain-language account of the trapping mechanism to Interesting Engineering: “Laser light is nothing but monochromatic electromagnetic radiation. If a neutral atom is placed inside this electric field, a dipole moment is induced, which leads to a force that enables us to hold atoms in place.”

Professor Esslinger summarized the team’s position in measured terms, as quoted by The Quantum Insider: “We can now make lots of swap gates with neutral atoms, but of course we still need a few other ingredients to build a working quantum computer.”

What We Don’t Know

The ETH Zurich demonstration was conducted in an optical lattice system that, while capable of addressing thousands of atom sites, does not yet include the quantum gas microscope capability needed to selectively read out and manipulate individual qubit pairs. The team has identified combining these geometric-phase gates with a single-site-resolution microscope as a next step, but has not published a timeline for doing so.

The 99.91% fidelity figure was measured across the large parallel array; it is not yet established whether the same fidelity holds for gates applied to individually addressed qubit pairs under realistic error-correction conditions. That distinction matters because fault-tolerant algorithms generally require both high gate fidelity and the ability to selectively operate on specific logical qubits — two requirements that are easier to satisfy independently than together.

It also remains open how the geometric-phase swap gate approach compares with the highest two-qubit gate fidelities reported for other neutral-atom platforms and for competing qubit technologies such as trapped ions. The ETH Zurich result does not include direct benchmarking against systems from QuEra, Pasqal, or other neutral-atom operators.

Analysis

Neutral-atom platforms have emerged as a leading contender for fault-tolerant quantum computing in part because their qubit positions can be reconfigured mid-computation — an advantage that superconducting and trapped-ion architectures cannot replicate as readily. The persistent challenge has been that the same laser fields used to trap and address atoms are also a source of noise. A gate design that is topologically protected from that noise, without requiring additional hardware overhead, directly addresses one of the platform’s most persistent engineering constraints.

The parallel operation across 17,000 qubit pairs is also significant beyond the fidelity figure alone. Quantum processors useful for real-world computation will require the ability to run many gate operations in parallel rather than serially, and the ETH Zurich result demonstrates that geometric-phase swap gates are compatible with that requirement at a scale already large enough to be relevant for near-term error-corrected systems.

ETH Zurich described the outlook from the Live Science perspective: the new technique is “bringing large-scale, stable quantum processors a step closer to reality.”