Suppression of Crosstalk in Superconducting Qubits Using Dynamical Decoupling

Currently available superconducting quantum processors with interconnected transmon qubits are noisy and prone to various errors. The errors can be attributed to sources such as open quantum system effects and spurious interqubit couplings (crosstalk). Static ZZ coupling between qubits in transmon architectures is always present and contributes to both coherent and incoherent crosstalk errors. Its suppression is therefore a key step toward enhancing the fidelity of quantum computation using transmons. Here, we propose the use of dynamical decoupling to suppress the crosstalk and we demonstrate the success of this scheme through experiments performed on several IBM quantum cloud processors. In particular, we demonstrate improvements in quantum memory as well as the performance of single-qubit and two-qubit gate operations. We perform open quantum system simulations of the multiqubit processors and find good agreement with the experimental results. We analyze the performance of the protocol based on a simple analytical model and elucidate the importance of the qubit drive frequency in interpreting the results. In particular, we demonstrate that the XY4 dynamical decoupling sequence loses its universality if the drive frequency is not much larger than the system-bath coupling strength. Our work demonstrates that dynamical decoupling is an effective, practical, and scalable way to suppress crosstalk and open-system effects, thus paving the way toward higher-fidelity logic gates in transmon-based quantum computers.

Automated summary

The current available superconducting quantum processors are noisy and prone to errors, which can be suppressed by using dynamical decoupling to suppress crosstalk. We demonstrated the success of this scheme through experiments on several IBM quantum cloud processors and achieved improvements in quantum memory and gate operations. Our work paves the way for higher-fidelity logic gates in transmon-based quantum computers.