Coherent oscillations in weakly anharmonic NbSe2 qubit

TL;DR

The aluminum-based superconducting qubits face material-related coherence and noise challenges that limit performance in demanding tasks such as dark-matter sensing. The authors address this by fabricating a NbSe$_2$-based qubit embedded in an aluminum cavity and characterizing it as a weakly anharmonic Duffing oscillator using spectroscopy, two-tone experiments, Rabi dynamics, and $T_1$ measurements. Key contributions include the demonstration of the first NbSe$_2$ qubit with $T_1 = 6.5 \pm 0.4 \mu s$, photon-noise resilience up to $5$–$10$ photons, a dispersive coupling to the readout mode of $g_{110}/2\pi = 67 \pm 17$ MHz, and an anharmonicity $\alpha/2\pi = -1.3$ MHz, alongside a drive-induced sign change of the Kerr nonlinearity. The work establishes NbSe$_2$ and vdW materials as viable platforms for high-coherence, magnetic-field–resistant quantum devices and points toward applications in quantum sensing and dark-matter detection in high-field environments.

Abstract

The functionalization of quantum devices to increase their performance and extend their fields of application is an extremely active research area. One of the most promising approaches is to replace aluminum with more performant materials. Within this context, van der Waals (vdW) materials are ideal candidates since they would allow to embed their unique properties into qubits. However, the realization of qubits based on vdW materials other than graphene is yet to be achieved. In this work we present a weakly anharmonic NbSe2 qubit. Our device exhibits a relaxation time T1 = 6.5 +\- 0.4 us which is roughly 2 orders of magnitude larger of other vdW qubits in addition to robustness to photon noise up to 5-10 thermal photons. Our work serves as a demonstrator of the advantage of integration of vdW materials into quantum technologies as well as serving as the first step toward the application of quantum non demolition photon detection protocols in the challenging field of dark matter search.

Figures (11)

• Figure 1: Experimental scheme. The upper part shows the room temperature instrumentation and wiring, the lower part in the 300 K dashed black box represents the setup in the dilution refrigerator. The red line is a superconducting NbTi cable for low-loss transmission. The I--Q mixers serving as vector-modulator and down-converter are of the same model. The small dashed lines with open circles at their edges represent mechanical switches. More details in the main text.
• Figure 2: a): Optical image of the NbSe$_2$-NbSe$_2$ JJ. The overlap area (88 $\mu\hbox{m}^2$) is reported in false color for visibility. b) Dressed cavity transmission evolution as a function of the readout power (Read out frequency expressed as difference with $\nu_{r}=7.1873$ GHz). The appearance of a step-like resonance is a fingerprint that our system behaves as a Duffing oscillator. The spectra are vertically shifted for better visualization. c): Hysteresis of the Duffing oscillator, depending if the frequency sweep spans from low to high frequency value (ascending) or viceversa (descending). The transmission curves are acquired with readout power at resonator input, P$= -114$ dBm (the x axis is expressed as detuning from $\nu_{r}=7.18723$ GHz). d): 2D map displaying the Duffing oscillator hysteresis where we plotted the difference of cavity transmission measured from ascending and descending sweeps. The x axis is expressed as detuning from $\nu_r=7.187$ MHz
• Figure 3: a): Switch from softening (pump off) to hardening (pump on: $-75$ dBm) Duffing oscillator (the $x$ axis is expressed as detuning from $\nu_{c}=7.1873$ GHz).b): Two-tone spectroscopy of the device as a function of the drive tone frequency and power. The readout tone is set at $\nu_{ro}$ =7.1873 GHz with power P$_{ro}= -132$ dBm. c): Dressed resonator frequency $\Delta\nu_{r}'$ (expressed as detuning from $\nu_c=7.187$ GHz) as a function of the inverse qubit frequency $\nu^{-1}_q$. d) Rabi oscillation of the qubit for a drive power P$=-95$ dBm as a function of the time length of the driving pulse $\Delta t$.
• Figure 4: a) Rabi frequency as a function of the drive amplitude expressed as squared root of the effective cavity occupation-number. The deviation from the linear behavior points to a multi-level process. Error bars are dominated by systematics and the horizontal ones are correlated. The data are reported along with the simulated curve obtained solving the time evolution of eq. \ref{['simulatio_hamil']} in appendix \ref{['cavityattenuationII']} for different values of $g/2\pi$. b) Decay time $T_1$ expressed as a function of the delay between the $\pi$ excitation pulse and the readout, acquired with power P$=-101$ dBm. The data are expressed as the probability to find the qubit in the ground state.
• Figure 5: Simulated transmission of the four main modes of the Al cavity.