Silicon spin qubits have an obvious appeal: they can be built with the same manufacturing methods that turn out ordinary computer chips. The catch, as ever, is in the detail. Drive several of them at once and their gates get sloppier, which is awkward when a useful machine needs many qubits working in parallel with errors kept low.
Taming the AC Stark shift
Researchers at the RIKEN Center for Emergent Matter Science and collaborators now report in Nature Physics that they have controlled several silicon spin qubits at once with high accuracy. Their device was an array of five qubits sharing a single microwave control line. Under the hood, they traced much of the trouble to the AC Stark shift: the microwave pulses that drive one qubit nudge its effective energy levels, which shifts the phase of its quantum state. With several qubits being driven together, those phase errors pile up and derail the intended operations.
The team’s fix is a calibration scheme that compensates for the shifts. It avoids calibrating every possible combination of qubits and instead relies on calibrations between pairs. “Our method uses calibrations between pairs of qubits to compensate for the unwanted phase shifts,” said first author Yi-Hsien Wu, who did the experiments as part of his PhD work.
Combined with an improved device design and tuned microwave pulses, it produced results worth reading closely. Tested alone, each of the five qubits reached single-qubit gate fidelity well above 99.99%, the best approaching 99.999%. With up to three running together, fidelity stayed above 99.99%. With all five active, it held at about 99.9%.
That last figure is not perfect, and the team says so. The remaining loss comes from drive-induced decoherence, caused by the extra microwave power needed to run several qubits at once. Leon Camenzind, a RIKEN research scientist and co-first author, called it “a limitation, but further improvements to device design could help mitigate this source of error.” His wider claim is that the work supports shared control lines as a practical route to scaling, since giving every qubit its own control system would become unmanageable. Seigo Tarucha, who manages a silicon quantum computing project under JST’s Moonshot Goal 6, said the long-term goal is a fault-tolerant machine built on silicon spin qubits.
Error correction on a 156-qubit chip
Fault tolerance is the other half of the problem. Qubits are fragile, and heat or motion in their surroundings can corrupt a calculation. A separate study in Nature Physics, from US researchers including teams at Rutgers, IBM, the University of Florida and Louisiana State University, tested a scheme on a 156-qubit IBM Quantum Heron processor, linking as many as 100 qubits. The system alternated two tasks: scrambling, which spreads chaos through the processor, and controlling, which resets qubits where needed. That is the same basic mechanism fault tolerance will demand.
Across nearly 5,000 qubit operations and almost 5,000 checks and resets, order returned once the checks and resets made up more than half of all operations. As more qubits were added, that tipping point sharpened in the data, which the authors read as a genuine phase transition, like water turning to ice. “The agreement between the experiment and the theory showed that this was not an accidental feature of a small circuit,” said Haining Pan of the University of Florida. “It was collective behavior emerging across the quantum system.”
IBM’s Maika Takita stressed the engineering point: the hard part was measuring a qubit repeatedly while the rest of the processor kept running. Rutgers physicist Jedediah Pixley called the experiment “a landmark on that road”. It remains a stress test, and the scheme would need more complex rules to be useful in real computation. Governments have put about $10 billion into quantum computing as of April 2025, and for now demonstrations like these are best read as scientific milestones, not deployment plans.
Sources
- Quantum computing
- Scientists overcome a hidden obstacle to more powerful quantum computers
- We're a Step Closer to Quantum Computers That Can Correct Their Own Mistakes
Photo: Department for Science, Innovation & Technology (CC BY 2.0)
