Nonvolatile Cryogenic Phase Slip Memory with Single-Shot Readout

Abstract

The demand for cryogenic memory components is driven by the need for ultra-fast, low-power, and highly reliable computing systems. Phase slip-based devices promise to fulfill all these requirements, with potential applications in both classical and quantum information processing. However, previous implementations have faced challenges due to inefficient writing and readout schemes. In this work, we address these limitations with a simple device design and measurement techniques inspired by circuit quantum electrodynamics. We present a memory element that stores information in the winding of a high-kinetic inductance superconducting loop, inductively coupled to a coplanar waveguide resonator. Using single-shot measurements, we achieve a readout fidelity of 99.698\% with an active measurement time of just 25 ns.

Figures (4)

• Figure 1: a) A schematic representation of the memory device comprising an Al loop (indicated in orange) coupled to a NbTiN $\lambda$/4 resonator (light gray). A current bias line (control line) supplies a local magnetic field in the vicinity of the Al loop, which is positioned at the Si trench (dark gray). The orange arrows indicate the path of the bias current. A complete view of the device is found in the Supplementary Information. The insets show SEM images of the loop and constriction. b) Spectroscopy of the transmission through the feedline near the resonator's $f_r$ for two distinct energy states of the loop, labeled 0 and 1. c) The variation in resonance frequency following ZFC as $I_{bias}$ is altered following the values represented in panel d as a function of time. $f_0$ = 6.139074 GHz is the resonance frequency of state 0 at zero bias. Both phase slip events (from 0 to 1 and from 1 to 0) manifest as discontinuous jumps. The inset shows a magnified view of the 0 to 1 phase slip event. e) A schematic representation of the absolute value of the circulating current inside the Al loop during the $I_{bias}$ cycle. At the values indicated in panel d, a phase slip (PS) event occurs, altering the loop's energy state.
• Figure 2: a) A schematic representation of the measurement protocol. After each write current pulse of $\pm 28$ mA, the VNA performs a frequency sweep. The resonance frequency is determined by fitting the obtained data, as described in the text. b)The resulting $f_r-f_0$ values at 300 mK, with $f_0$ = 6.139074 GHz, obtained after 375 alternating 160 $\mu$s write pulses. Orange and blue shaded regions correspond to states 0 and 1, respectively.
• Figure 3: a) Transmission measurements at 13 mK are represented in the $IQ$ plane. The data is collected using the VNA and $t_{int}$ = 3.7 $\mu$s (light markers) and 320.5 $\mu$s (darkmarkers). The orange markers refer to data for state 0, while the blue ones refer to state 1. b) Representative bivariate normal distribution fitting for one of the distributions presented in panel a. c) SNR as a function of $t_{\text{int}}$ for data obtained using the VNA (dark grey, diamond) and the AWG (light grey, circle).
• Figure 4: a) The output signal amplitude ($V_{AWG}$) of the readout pulse as a function of time. The result highlights the resonator response. The active measuring time is represented by the black vertical lines for $t_{int}$ = 25 ns. b) The acquired transmission values at 13 mK represented in the $IQ$ plane with $t_{int}$ = 25 ns (light markers) and 50 ns (dark markers) after preparing the system in states 0 and 1. State 0 is represented in orange and state 1 in blue.