Evolution of Nanowire Transmon Qubits and Their Coherence in a Magnetic Field

F. Luthi, T. Stavenga, O. W. Enzing, A. Bruno, C. Dickel, N. K. Langford, M. A. Rol, T. S. Jespersen, J. Nygård, P. Krogstrup, and L. DiCarlo

Phys. Rev. Lett. 120, 100502 – Published 9 March 2018

Abstract

We present an experimental study of flux- and gate-tunable nanowire transmons with state-of-the-art relaxation time allowing quantitative extraction of flux and charge noise coupling to the Josephson energy. We evidence coherence sweet spots for charge, tuned by voltage on a proximal side gate, where first order sensitivity to switching two-level systems and background $1 / f$ noise is minimized. Next, we investigate the evolution of a nanowire transmon in a parallel magnetic field up to 70 mT, the upper bound set by the closing of the induced gap. Several features observed in the field dependence of qubit energy relaxation and dephasing times are not fully understood. Using nanowires with a thinner, partially covering Al shell will enable operation of these circuits up to 0.5 T, a regime relevant for topological quantum computation and other applications.

Received 22 November 2017

DOI:https://doi.org/10.1103/PhysRevLett.120.100502

Physics Subject Headings (PhySH)

Quantum computation Topological quantum computing

Nanowires Superconductors

Josephson junctions SQUID Superconducting qubits

Quantum Information, Science & TechnologyCondensed Matter, Materials & Applied Physics

Authors & Affiliations

F. Luthi^1,2, T. Stavenga^1,2, O. W. Enzing^1,2, A. Bruno^1,2, C. Dickel^1,2, N. K. Langford^1,2, M. A. Rol^1,2, T. S. Jespersen³, J. Nygård^3,4, P. Krogstrup³, and L. DiCarlo^1,2,*

¹QuTech, Delft University of Technology, Lorentzweg 1, 2628 CJ Delft, Netherlands
²Kavli Institute of Nanoscience, Delft University of Technology, Lorentzweg 1, 2628 CJ Delft, Netherlands
³Center for Quantum Devices, Niels Bohr Institute, University of Copenhagen, DK-2100 Copenhagen, Denmark
⁴Nano-Science Center, Niels Bohr Institute, University of Copenhagen, DK-2100 Copenhagen, Denmark

^*Corresponding author. l.dicarlo@tudelft.nl

Article Text (Subscription Required)

Click to Expand

Supplemental Material (Subscription Required)

Click to Expand

References (Subscription Required)

Click to Expand

Issue

Vol. 120, Iss. 10 — 9 March 2018

Reuse & Permissions

Access Options

Author publication services for translation and copyediting assistance advertisement

Images

Figure 1
Characterization of flux-tunable split-junction qubit at $B_{∥} = 0$ . (a) False-colored SEM micrograph of the SQUID-loop area. The current $I$ in the flux-bias line (yellow) threads a magnetic flux $Φ$ through the SQUID loop (green), tuning $E_{J} (Φ)$ . (b) The joint fit (black) of the first three transitions (orange, dark red, and purple symbols) versus flux yields the transmission probability and the induced gap of each junction. (c) $T_{1}$ (blue symbols) limits $T_{2}^{Echo}$ (green symbols) at the qubit flux sweet spot $Φ = 0$ . A fit to $T_{2}^{Echo}$ that includes the measured $T_{1}$ limit allows extraction of flux-independent (cyan line), $1 / f$ (pink line) and white-noise (gold line) contributions to the dephasing. $T_{2}^{*}$ is typically below $4 μ s$ . Top axis indicates the frequency detuning from the flux sweet spot. (d) $Γ_{φ}^{Echo}$ versus flux sensitivity, extracted from (c), with the different contributions to the fit.
Reuse & Permissions
Figure 2
Frequency stability analysis of the flux-tunable qubit at the flux sweet spot and at $B_{∥} = 0$ . (a) Ramsey experiment (dots) with artificial detuning (12 MHz). The strongly coupled TLS yields a beating between two decaying sinusoids (purple line, $τ^{A} = 2.2 μ s$ , and $τ^{B} = 2.0 μ s$ ). (b) The extracted detunings of repeated Ramsey experiments show a constant frequency spacing and drift of the center frequency. The gray vertical line indicates the trace in (a). (c) Pulse sequence of the Ramsey-type TLS-state detection scheme. The free evolution time $τ$ is chosen as $τ_{wait} = 1 / 2 Δ f$ for maximal contrast. (d) The PSD (red line) of the TLS is computed from qubit state traces obtained by monitoring the qubit frequency real time using the pulse sequence in (c). The PSD is fitted using RTN models with (blue line) and without (green line) $1 / f$ noise.
Reuse & Permissions
Figure 3
Gatemon characterization at $B_{∥} = 0$ . (a) False-colored SEM micrograph of the nanowire Josephson junction (light red) with a side gate (yellow) enabling $V_{G}$ tuning. (b) Deviation of $f_{R}$ , $Δ f_{R}$ , from the bare resonator frequency $f_{bare} = 6.732 GHz$ for a triangle sweep in $V_{G}$ . Note the change in direction of the $V_{G}$ sweep, indicated by the dashed line. On return to the same $V_{G}$ , $f_{R}$ is roughly reproduced. (c) $f_{01}$ versus $V_{G}$ . Random, but sometimes reproducible jumps of $f_{01}$ occur (at light gray lines). (d) Plot of $f_{01}$ against $f_{R}$ (orange dots) and dressed state fit (black curve) with coupling strength $g / 2 π = 60.8 MHz$ , allowing a prediction of $f_{01}$ . (e) Gatemon $T_{1}$ (blue symbols), $T_{2}^{Echo}$ (green symbols), and $T_{2}^{*}$ (red symbols) versus $V_{G}$ . Both $T_{2}^{Echo}$ and $T_{2}^{*}$ are strongly correlated with the $V_{G}$ sensitivity (black line). (f) $Γ_{φ}^{Echo}$ against $V_{G}$ sensitivity, extracted from (e). The fitted $1 / f$ noise (blue line) is above the setup-imposed dephasing limit (purple line), indicating additional on-chip noise.
Reuse & Permissions
Figure 4
Evolution of the gatemon of Fig. 3 in $B_{∥}$ . (a) Qubit $f_{01}$ and $f_{02} / 2$ (orange and red dots) are described by a closing BCS gap (curves) with $B_{c} = 83.9 mT$ . (b) At each value of $B_{∥}$ , the gatemon is tuned to a $V_{G}$ sweet spot (SS) to measure $T_{1}$ and $T_{2}^{Echo}$ (blue and green symbols). At low $B_{∥}$ ( $f_{01}$ near the resonator), $T_{1}$ is mainly Purcell limited (red line). At $B_{∥}$ close to $B_{c}$ the superconducting gap becomes so weak that quasiparticle tunneling dominates $T_{1}$ (brown, assumes 100 mK effective quasiparticle temperature). In between, the $T_{1}$ evolution can be attributed to a step in $Q_{b}$ , see below. (c) $T_{1}$ versus frequency at different $B_{∥}$ . Accounting for ${T_{1}}^{P}$ , we fit a $Q_{b}$ at each $B_{∥}$ (inset), finding a steplike drop from 6 to 10 mT. (d) Keeping $B_{∥}$ fixed [same color scale as (c)], $V_{G}$ scans are performed to extract $Γ_{φ}^{Echo}$ , similar to Fig. 3. Inset: averaged extracted $1 / f$ voltage noise amplitude. (e) Pure dephasing rates at $V_{G}$ sweet spots versus $B_{∥}$ from data in (b) and (d). Stars are the interpolated minimal dephasing rates from (d).
Reuse & Permissions

Physical Review Letters