Superconducting integrated on-demand quantum memory with microwave pulse preservation

Abstract

Microwave quantum memory represents a critical component for quantum radars and resource-efficient approaches to quantum error correction. Superconducting microwave resonators provide highly efficient storage, long coherence times, on-demand reading and even in memory pulse engineering, but it is still challenging to overcome design and materials induced loss channels for on-chip realization. In this work, we present a novel architecture of integrated superconducting quantum memory with a dynamically controlled RF-SQUID coupling element in pulse regime, thus ensuring high efficiency storage and cycling storage time. It demonstrates a memory cycle time of 1.51 $μs$ and 57.5% storage fidelity with preservation of the stored pulse shape during the retrieval at single-photon level excitations. We establish that while the proposed active coupler realization introduces no measurable fidelity degradation, the primary limitation arises from impedance matching and materials imperfections. Still the device was used only for storing finite-duration near-single-photon classical microwave pulses, we assert that it operates as a linear device when the photon population in the common resonator remains low so it should be compatible with quantum state storage.The proposed architecture highlights a disruptive potential for on-chip qubit and memory integration for scalable quantum error correction, while identifying specific avenues for near-unity storage fidelity.

Figures (3)

• Figure 1: (a) Image of the fabricated quantum memory device. All structures are embedded in vortex-pinning hole array that helps achieve high internal Q-factor by magnetic vortices trapping McRae2020(b) Principal scheme of the quantum memory device with the relative positions of the common resonator and internal resonators. The device is mounted in a sample holder and electrically connected to a copper printed circuit board (PCB) via three parallel aluminum wire bonds, introducing an additional parasitic inductance of $L_{wb}\approx 0.5\ nH$. (c) Modeled dependencies of $\kappa$ and common resonator frequency on external magnetic flux in RF-SQUID loop. Designed operating points with common resonator frequency of 6 GHz and $\kappa$ of 20 MHz (red dashed lines) correspond to magnetic flux of $\approx0.33$ and are marked by red stars. (d) Concept of storage stage implementation. Green pulse - device response without coupling strength tuning. At the moment of the dark state in the common resonator, the coupling between the external waveguide and the common resonator should be switched off for the required period of time (red shaded rectangular). At the n-th period of dark state in common resonator (purple stars) coupling strength could be tuned to impedance-matching value to effectively release quantum pulse from memory cell.
• Figure 2: Device operation point calibration. (a) Device single-tone spectroscopy. Switching-off points marked by black arrows. The voltage and pulse frequency fine range used in the next calibration stages is highlighted by a dashed blue frame. McRae2020(b) initial reflection intensity dependence on voltage and pulse frequency in a fine range highlighted in (a). A Frequency of 5.992 GHz and a voltage of 0V are chosen as operating points for record and release stages, such that they provide minimal initial reflection intensity. (c) Impulse response intensity on pulse frequency at constant voltage of 0V. (d) Reference (black) and two cross-sections marked by dashed lines at (c). We define the initial reflection intensity as the maximal value (marked by corresponding stars) in the red shaded area.
• Figure 3: Device fidelity measurements. a) Pulse sequence for fidelity measurements. The storage stage duration is varied to demonstrate on-demand retrieval capability. b) Quantum memory responses for the first three storage periods (red, green and black, respectively. Blue graph corresponds to reference). We find memory cycle time equal to $T_{stor}=1.51\ \mu s$ and fidelity of first storage period $F=57.5(4)\%$. Inset: Fidelity on storage time. Due to internal losses, it has exponential decay behavior with effective decay time $T_{decay}=11.44(3.58)\ \ \mu s$, which corresponds to effective quality factor of $Q_{i,eff}=4.3\cdot10^5$.