Quantum optimal control

The primary focus of this project will be on the design, optimal characterization and control of multi-qubit superconducting devices based on transmon qubits in a circuit QED architecture. We will use measurement in a closed-loop way to optimize the tune-up of the system to obtain high-fidelity quantum gates. The project also addresses the question how to tailor control and measurements of a complex multi-qubit quantum processor in order to obtain targeted information in the most efficient and robust way. We will develop the tools to make best use of the retrievable information in our measurements, including statistical accuracies, backgrounds and imperfections, to find an optimal model of the system by comparing experimentally measured results with numerical/analytical predictions. The project is part of the EU training network QuSCo in which we are closely collaborating with the group of Frank Wilhelm-Mauch.

(c) Egger & Wilhelm-Mauch.

Team members

Max Werninghaus (PhD student working at IBM Research – Zurich)
Federico Roy (PhD student working IBM Research – Zurich)


Daniel Egger (IBM Research – Zurich)
Shai Machnes & Frank Wilhelm-Mauch (University Saarland)


Recent results

Leakage reduction in fast superconducting qubit gates via optimal control – M. Werninghaus et al. arXiv:2003.05952 (2020).

In this artice we demonstrate a sevenfold reduction of the leakage rate in fast superconduction qubits gates via optimal control. We have implemented a closed-loop optimization that simultaneously adapts all control parameters based on measurements of a cost function built from Clifford gates. By parameterizing pulses with a piecewise-constant representation that matches the capabilities of the control hardware we create a 4.16 ns single-qubit pulse with 99.76% fidelity and 0.044% leakage.

Single-qubit gate fidelity estimated via Randomized Benchmarking for different pulse shapes.
Pulsed Reset Protocol for Fixed-Frequency Superconducting Qubits – D. Egger et al. Phys. Rev. Applied 10, 044030 (2018).

Increasing coherence times of quantum bits is a fundamental challenge in the field of quantum computing. With long-lived qubits it is, however, inefficient to wait until the qubits have relaxed to their ground state after completion of an experiment. Moreover, for error-correction schemes it is important to rapidly reinitialize syndrome qubits. We present a simple pulsed qubit reset protocol based on a two-pulse sequence. A first pulse transfers the excited-state population to a higher excited qubit state and a second pulse transfer it into a lossy environment provided by a low-Q transmission-line resonator, which is also used for qubit read-out. We show that the remaining excited-state population can be suppressed to (1.7±0.1)% and that this figure may be reduced by further improving the pulse calibration.

Figure 5
Population in the qubit for different trigger rates. (a) Pulse sequence. After an initial pi pulse used to identify |e⟩ in the I-Q read-out plane, the qubit is repeatedly measured at each successive trigger pulse. (b) Population in the qubit with (triangles) and without (circles) the pulsed reset sequence.