RIKEN Team Keeps Silicon Qubits Near 99.9% Accuracy With Five Operating at Once

Five silicon spin qubits, wired to one shared microwave control line, were switched on together in a RIKEN experiment, and they kept their accuracy at roughly 99.9 percent. The result, published in Nature Physics, comes from researchers at the RIKEN Center for Emergent Matter Science and their collaborators. It speaks to a problem that stands in the way of building larger quantum computers: running many qubits at once tends to make each of them less reliable.

Silicon spin qubits hold information in the spin of individual electrons confined within silicon. Where an ordinary bit is a 0 or a 1, a qubit can exist in combinations of both. Controlling those electron spins takes extremely precise microwave signals, and when several qubits are driven at the same time, those signals produce unwanted side effects that degrade accuracy.

The team traced much of the trouble to a phenomenon called the AC Stark shift. Microwave control signals alter a qubit's effective energy levels, which changes the phase of its quantum state. Those phase shifts can disrupt the intended operation of a qubit, especially when several are being controlled together.

To counter it, the researchers built a calibration method that compensates for the shifts. The key choice was to avoid calibrating every possible combination of qubits separately. Their technique relies instead on calibrations between pairs of qubits, which makes it far more practical for systems with larger numbers of qubits. Yi-Hsien Wu, the first author, who carried out the experiments during his PhD research, described it this way: rather than needing a wholly separate procedure for every combination, the method uses pairwise calibrations to compensate for the unwanted phase shifts.

Combined with an improved device design and carefully adjusted microwave pulses, the method produced strong numbers. Tested one at a time, each of the five qubits reached single-qubit gate fidelity well above 99.99 percent, with the best approaching 99.999 percent. Fidelity measures how closely a quantum operation matches its intended result. With up to three qubits running simultaneously, fidelity stayed above 99.99 percent. With all five controlled at once, it held at about 99.9 percent, which the team describes as approaching the level needed for practical fault-tolerant quantum computing.

The errors were reduced, not eliminated. The researchers linked the remaining loss of accuracy to drive-induced decoherence, caused by the greater microwave power needed to operate several qubits together. Unlike the phase shifts the new method corrects, decoherence is a loss of the delicate quantum coherence that reliable operations depend on. Leon Camenzind, a RIKEN research scientist and co-first author who supervised the experiments, called this a limitation but said further improvements to device design could help mitigate it.

The stakes are about scale. Giving every qubit its own independent control system would make larger devices hard to manage, whereas shared control lines offer a more practical route. Camenzind said the study supports shared lines as a practical approach to scaling up, and noted that silicon's compatibility with established semiconductor manufacturing strengthens the case. Seigo Tarucha, project manager of a silicon quantum computing project under JST's Moonshot Goal 6, said the long-term aim is a fault-tolerant quantum computer built on silicon spin qubits, with reliable parallel control an essential step. The next task is the one the team has already named: redesigning devices so that drive-induced decoherence stops being the limit as more qubits are added.