Tuning the third dot and tracking slow frequency drift produced a 99.56% single-spin gate in 4.05 nanoseconds.
Fast electrical control of a quantum-dot spin can make the qubit more vulnerable to charge-related fluctuations. In a triple-dot device, the researchers tuned the third dot to flatten how the qubit frequency changes with detuning in the target double-dot pair, while retaining strong spin-charge hybridization and Rabi frequencies above 100 MHz.
They then used a feedforward neural network to track slow changes in the qubit frequency. The feedback reduced the standard deviation of those fluctuations to 388 kHz. Randomized benchmarking measured a π/2 gate fidelity of 99.56% with a 4.05-nanosecond gate time.
What the extra dot changed
The researchers demonstrated a single-electron spin qubit in a GaAs triple-quantum-dot structure. Adjusting the detuning of the third dot reshaped the qubit’s orbital band, making its frequency less sensitive to detuning in the target double-dot system without giving up strong spin-charge hybridization. The device supported electrically driven Rabi frequencies above 100 MHz.
A feedforward neural network used results from earlier Ramsey measurements to estimate and correct slow qubit-frequency drift. With this feedback, the frequency fluctuations had a standard deviation of 388 kHz. Randomized benchmarking then measured a π/2 gate fidelity of 99.56% for a 4.05-nanosecond gate.
Why the added control matters
In a double-dot flopping-mode qubit, the same orbital splitting affects both the speed of electrical spin rotations and the width of the low-sensitivity, or “sweet-spot,” region. That links fast operation and protection from detuning noise. The third dot adds an electrical control that can tune these properties more independently.
The result identifies orbital engineering in a triple-dot structure as an electrical way to adjust this speed–coherence trade-off. It is presented as complementary to approaches based on a material’s intrinsic spin-orbit coupling or on fixed micromagnet geometry.
Evidence and boundaries
This is an experimental demonstration in a GaAs/AlGaAs device containing a triple-dot qubit, with gate performance evaluated by randomized benchmarking. The reported fidelity includes feedforward feedback for slow frequency drift and applies to a π/2 gate at the engineered operating point.
The evidence comes from one device and one demonstrated qubit configuration; the excerpts do not establish performance across a larger array or other materials. The researchers discuss further optimization, including a possible two-dimensional search for a second-order sweet spot, but that optimization was not demonstrated here.
// Source
npj Quantum Information · 2026 · DOI: 10.1038/s41534-026-01400-1
Authors: Y. Matsumoto, X.-F. Liu, Arne Ludwig, Andreas Dirk Wieck, K. Koike, T. Miyoshi, T. Fujita, A. Oiwa
Institutions: The University of Osaka, Ruhr University Bochum, Spintronics Research Network of Japan, Trend Micro (Japan), Center for Quantum Information and Quantum Biology, The University of Osaka