New method speeds up key quantum computing operations 1,000-fold

Researchers at Chalmers University of Technology in Sweden have developed a method that can carry out certain advanced quantum computing operations more than a thousand times faster than before. The work, led by Lei Du with co-author Tangyou Huang, addresses a long-standing slowdown in how quantum computers protect and manipulate their information. It was published as a theoretical study in Physical Review Letters, alongside collaborator Lingzhen Guo.

Quantum computers are prone to errors because their basic units of information, called qubits, are extremely sensitive to disturbance. Even small environmental effects, such as electrical noise, cosmic radiation, or overheating, can knock a qubit off course and cause it to lose its information. Ordinary computers face similar problems but can fix them quickly using well-established error-correction methods. For quantum computers, correcting errors before they pile up and derail a calculation has proven far harder.

One strategy for making quantum information more robust is to avoid storing it in single qubits altogether. Instead, so-called bosonic quantum codes store the information in the microwave fields inside superconducting circuits, an approach that has been shown to guard against certain kinds of errors. Building and controlling the quantum states used by these codes has traditionally required guiding the system through thousands of repeated control cycles, a slow process during which errors have more chances to creep in.

The Chalmers team's method instead completes these operations within a single control cycle rather than several thousand. It relies on a recently proposed set of building blocks called quantum lattice gates, which the same research team had introduced previously. Tangyou Huang compared the approach to building with pre-built modules rather than individual bricks: instead of assembling a structure piece by piece and risking mistakes along the way, the gates snap together quickly as ready-made units.

The technique is a form of Floquet control, in which a quantum system is steered by a periodic, repeating control signal. Earlier Floquet-based approaches typically needed many of these repeating cycles to build up the desired state, which is the slow process the new method replaces. By reaching the same result within a single cycle, the method cuts the exposure time during which disturbances could corrupt the operation.

The longer a quantum operation takes, the more time environmental noise has to disrupt it, so shrinking a process that took thousands of cycles down to one directly reduces the chance of error. The researchers describe this as addressing one of the major bottlenecks in building quantum computers that can correct their own errors reliably, known as fault-tolerant quantum computing. Quantum computers with this kind of reliability are expected to eventually contribute to fields including drug discovery, energy systems, cryptography, artificial intelligence, and logistics, though those applications depend on the broader technology maturing.

The method is designed for superconducting quantum computers, one of the leading hardware platforms being pursued for large-scale quantum computing. Chalmers is currently developing a 100-qubit quantum computer using this same superconducting platform. Because the new approach can be implemented using existing superconducting quantum circuit designs, the researchers say they are already discussing possible experimental demonstrations with colleagues at Chalmers.

The Chalmers result is a theoretical study, described by the researchers as a computational simulation, and an experimental demonstration has not yet been carried out. The researchers say they hope to see the method tested in the near future, but the thousand-fold speed-up applies specifically to building and controlling bosonic quantum states, not to every part of running a quantum computer.

One separate bottleneck is reading out a qubit's final state, a step distinct from the operations the Chalmers method speeds up. Other research groups have reported progress there: one team demonstrated single-shot readout fidelity up to 99.5% for distinguishing two states within 140 nanoseconds without a specialized amplifier, using a technique that shelves qubit population in higher energy levels to slow unwanted decay during measurement. A separate group achieved single-shot readout of a quantum dot electron spin in 3 nanoseconds with 95.2% fidelity using an optical microcavity, which they describe as the fastest single-shot readout of a quantum state achieved across any material platform to date. A third approach, called Generalized Syndrome Measurement, aims to reduce the number of readouts needed per error-correction cycle by extracting error information from a single ancilla measurement rather than multiple measurements, an idea so far tested through numerical analysis rather than physical hardware. Together these efforts illustrate that speeding up quantum operations and speeding up reliable measurement remain distinct problems, each still being worked on separately.