How DRAG Pulses Reduce Leakage in Superconducting Qubits
Pulse shaping helps superconducting transmon qubits perform fast control operations without inadvertently populating states outside the computational basis.
By The Quantum Lab desk
Written in-house by the The Quantum Lab desk — an explainer, not a report of a news event.
- Transmons have additional energy levels that can absorb control pulses intended for the qubit transition.
- DRAG control adds a calibrated quadrature component that suppresses leakage and phase errors.
- Pulse shaping improves gate fidelity but does not eliminate leakage from every source.
A transmon is often treated as a two-level system whose lowest energy states represent 0 and 1. Physically, however, it is a weakly anharmonic oscillator with higher states such as 2 and 3. Their presence creates a risk that a microwave control pulse will drive population out of the computational subspace.
The risk increases when gates are made faster. A short pulse has a broader frequency spectrum, so some of its energy can overlap the nearby transition between states 1 and 2. This unwanted excitation is called leakage and differs from an ordinary bit or phase error because the qubit is no longer confined to the two states assumed by the computation.
Derivative removal by adiabatic gate, or DRAG, is a pulse-shaping technique designed to limit this problem. Alongside the main microwave envelope, the controller applies a second component in the orthogonal quadrature, typically proportional to the time derivative of the main pulse. With suitable amplitude and phase, this component counteracts the dynamics that would otherwise populate the higher level.
The correction also helps manage phase shifts caused by off-resonant coupling to nearby transitions. In practice, engineers calibrate the pulse amplitude, duration, frequency and DRAG coefficient for each qubit. Too little correction leaves leakage, while too much can introduce different rotation or phase errors.
DRAG is not a complete solution to leakage. Strong interactions during two-qubit gates, frequency crowding, pulse distortion in control electronics and coupling to other modes can create additional pathways out of the computational subspace. Systems therefore combine pulse shaping with careful device design, calibration, leakage-sensitive characterization and, in some error-correction schemes, procedures that detect or remove leaked population.
This story started as a thread in our Discord. Join the discussion with the researchers and engineers who work on it.
Related
Surface-code processor stays below threshold as it scales to 1,000 qubits
The logical error rate keeps falling as the code distance grows — the clearest sign yet that adding qubits now buys reliability instead of noise.
Trapped-ion shuttling record pushes two-qubit fidelity past 99.9%
Moving ions between zones used to cost coherence. A revised trap geometry makes transport nearly free.
Open-source compiler cuts logical-qubit overhead by a third
A new lattice-surgery scheduler squeezes routing overhead out of fault-tolerant circuits, and it ships under a permissive license.
