Spin-orbit interaction in silicon quantum dots is weak compared with that in some other semiconductor systems, but it can still influence how spin qubits behave. Its effects include shifts in electron spin resonance (ESR) frequency, changes in g-factor and mixing between spin states during electron tunneling. These effects can be useful for qubit control, or they can contribute to noise sensitivity and unwanted transitions. Measurements on a silicon metal-oxide-semiconductor (SiMOS) double quantum dot show that changing the external magnetic-field direction provides a way to tune these competing outcomes.
Why spin-orbit interaction matters in silicon qubits
Silicon is attractive for spin-qubit research in part because its comparatively weak spin-orbit interaction can support long coherence times. Yet the silicon interface and the structure of the quantum dot create a measurable interaction that affects qubit operation. For engineers and system designers, the key point is that this is not simply an effect to eliminate: its impact depends on the operation being performed.
Spin-orbit coupling can alter the electron’s effective g-factor and the Stark shift, the change in ESR frequency associated with an electric field. Those effects can support electrical control or help distinguish neighboring qubits under a shared microwave drive. Conversely, electric-field sensitivity can make the resonance frequency more vulnerable to charge noise, while spin-state mixing can contribute to leakage during tunneling or transport.
Magnetic-field direction tunes the response
The study characterized two electron-spin qubits in a SiMOS double quantum dot while varying the direction of the applied magnetic field. Measurements included the g-factor of one dot, the g-factor difference between dots, the derivative of the Stark shift, and singlet-to-polarized-triplet mixing. Several of these quantities varied with the in-plane field angle, demonstrating that the magnetic-field orientation is an operational control parameter.
The analysis distinguishes two contributions to the interaction. Changes in the g-factor and Stark shift were associated mainly with the vector-potential contribution to spin-orbit coupling. The singlet-triplet mixing observed during electron tunneling, by contrast, was attributed primarily to a spin-orbit field associated with the electron’s motion between dots. These mechanisms matter in different ways, so optimizing one measured property does not automatically optimize every aspect of qubit performance.
Balancing qubit addressability and coherence
Differences in the g-factors of neighboring dots can help address qubits individually using a global microwave field. Reducing that difference, however, can make it possible to drive multiple qubits at a shared frequency. Magnetic-field direction changes the g-factor difference, allowing the device’s operating point to be selected according to the control strategy.
The same choice affects sensitivity to electric noise. When the derivative of the Stark shift is small, fluctuations in the electric field produce less change in the resonance frequency. In the reported device, the derivative approached zero near a particular in-plane field orientation, and the measured dephasing time T2* increased by 80% compared with the value near the [110] direction. The improvement was observed in the three-electron charge configuration used for that coherence measurement; it should not be treated as a universal result for every dot or operating condition.
This result illustrates a practical trade-off. The field orientation associated with reduced Stark-shift sensitivity can also reduce the g-factor difference that enables individual addressability. A design decision therefore needs to account for both coherence and control requirements rather than optimizing either in isolation.
Controlling singlet-triplet mixing during tunneling
When an electron moves between dots, spin-orbit coupling can mix the singlet state with the spin-polarized triplet state known as T−. Such mixing is relevant to operations based on singlet-triplet states and to schemes that shuttle electron spins through a quantum-dot array. Unwanted transitions can move population out of the intended operating states.
In the measured double dot, the coupling varied substantially with magnetic-field orientation: the reported minimum was 1.85 MHz, while the maximum was 27.55 MHz. Its strong angular dependence was inconsistent with residual nuclear-spin effects as the main explanation. The measurements also did not support differences in off-diagonal g-tensor terms as the dominant mechanism. Instead, the observed behavior was explained primarily by the spin-orbit field generated as charge moves between the dots.
Identifying an orientation that minimizes this mixing offers a potential way to reduce leakage during relevant operations. In a larger array, however, dot pairs may differ in their local interaction strengths and geometry. The best field direction for one pair may therefore not be the best for every pair.
Implications for scalable quantum-dot design
The results point to a system-level tuning problem. Magnetic-field orientation influences qubit frequency, addressability, electric-noise sensitivity and spin-state mixing, while the preferred setting depends on how the qubits are operated. In an array, variation in the spin-orbit interaction from dot to dot can further complicate the selection of a single global operating condition.
Gate-voltage tuning may provide another degree of control: the study notes that the Rashba interaction is tunable through the top-gate voltage. This suggests that field direction and electrostatic settings can be considered together when balancing array-level requirements. The measurements do not establish one universally optimal setting; rather, they show why device characterization must consider multiple performance parameters.
Conclusion
Spin-orbit interaction in silicon quantum dots is both a source of operational sensitivity and a resource for control. Varying the magnetic-field direction can change g-factor differences, Stark-shift sensitivity and singlet-triplet mixing, with measurable consequences for addressability, coherence and leakage. For scalable quantum-dot systems, the practical task is to characterize these effects and choose operating conditions that match the intended qubit operations—not to assume that a single field orientation will optimize every metric.