A quantum computer’s power depends not only on how many qubits it contains, but also on how accurately they can work together.
Now, researchers have demonstrated a way to operate five silicon-based qubits simultaneously through a single microwave control line, maintaining operation accuracy at or above 99.9%. This approach could help make future quantum processors easier to scale without requiring a separate control line for every qubit.
However, controlling several qubits at once is not as simple as sending them microwave signals together. Quantum computers need to perform operations in parallel to run error-correction routines, which help protect fragile quantum information.
Yet the signals used to control one qubit can interfere with its neighbors, making their operations less accurate. In earlier experiments, this problem caused gate fidelity—a measure of how accurately a quantum operation is performed—to fall from 99.95% to around 99.5%.
“Simultaneous control remains a major challenge in silicon devices, as fidelities typically degrade under parallel operation,” the researchers note in their study.
Researchers from Japan’s RIKEN and collaborating institutions have now identified the main source of this interference and developed a way to compensate for it.
The hidden frequency shift throwing qubits off course
The team worked with five qubits made from electron spins trapped in tiny structures called quantum dots within a silicon/silicon-germanium device.
“Silicon spin qubits, in particular, offer a promising path to scalable quantum computing due to their compatibility with industrial semiconductor manufacturing,” the researchers added.
Microwave signals control these qubits, but a signal aimed at one can also affect its neighbors. This unwanted influence, known as crosstalk, can make quantum operations less accurate even when the neighboring qubits are not the intended targets.
The researchers traced the main source of this interference to a phenomenon called the AC Stark shift. A microwave signal can change a neighboring qubit’s effective energy levels, altering how its quantum state evolves.
This produces an unwanted phase shift—a change in the relative timing of the waves that describe a quantum state—which can throw off the intended operation.
A pair-by-pair fix keeps quantum operations accurate
The team found that these phase shifts could be predicted by measuring how individual pairs of qubits affected one another. They used those measurements to compensate for the unwanted changes when several qubits operated together, avoiding the need to calibrate every possible combination separately.
As study first author Yi-Hsien Wu explained, “Rather than requiring an entirely separate calibration procedure for every possible combination of qubits, our method uses calibrations between pairs of qubits to compensate for the unwanted phase shifts.”
The results showed how effective the approach could be. When operated individually, the five qubits achieved gate fidelities above 99.99%, with the best approaching 99.999%.
When up to three qubits operated simultaneously, their fidelities remained above 99.99%. Even when all five operated in parallel, their gate fidelities stayed at or above 99.9%.
This level of accuracy meets a widely used benchmark for fault-tolerant quantum computing, where error-correction techniques help protect calculations from mistakes. It does not, however, mean the researchers have already built a fully fault-tolerant quantum computer.
Faster quantum operations, with less power
Using a shared microwave line could reduce the number of control connections needed by larger processors, saving space and potentially reducing the heat generated by control hardware.
The researchers see the approach as a possible foundation for modules containing around ten qubits per shared line, although this larger configuration remains a future goal rather than a demonstrated result.
Other challenges remain. Stronger microwave signals can speed up operations but also introduce unwanted disturbances that disrupt qubits’ delicate quantum states. The team proposes improving the tiny magnets used to control electron spins by integrating them into the chip’s wiring layers.
These improved magnets could produce stronger local magnetic fields, potentially allowing faster operations at lower microwave power.
For now, the achievement addresses one obstacle to scaling silicon quantum computers: maintaining accurate control when multiple qubits operate together. Extending the approach to larger arrays while controlling other sources of error will be essential before it can support practical, fault-tolerant quantum computing.
The study is published in the journal Nature Physics.