High-Kinetic-Inductance Superconducting Nanowire Resonators
High-Kinetic-Inductance Superconducting Nanowire Resonators
High-Kinetic-Inductance Superconducting Nanowire Resonators
DOI: 10.1103/PhysRevApplied.5.044004
a magnetic field, resulting in Qi > 2 × 105 up to B∥ ¼ 6 T. essentially distributed resonators with a negligible direct
We also investigate the evolution of these resonators with capacitance between the nanowire ends. In the configura-
perpendicular magnetic field B⊥ , finding a clear depend- tion of Fig. 1(a), the coupling of the fundamental (half-
ence of the magnetic field resilience of Qi on the nanowire wave) mode of the resonator to the feedline is inductive,
width w. The narrowest nanowires (w ≈ 100 nm) achieve which for our high impedance resonators is extremely weak
Qi > 3 × 104 at B⊥ ≈ 350 mT. [Fig. 1(a)]; therefore, we focus on the full-wave mode,
leaving the discussion of the fundamental to Appendix A.
Device fabrication begins with sputtering of a NbTiN
II. METHODS film (thickness t ∼ 8 nm) on a high-resistivity Sih100i
The resonators consist of NbTiN nanowire loops inter- substrate [16,27]. A CPW feedline and several (four or
rupted by a small gap (Fig. 1) and coupled to a common five) nanowire resonators are next defined in a single
CPW feedline. To minimize C, the nanowire is detracted electron-beam lithography step followed by reactive ion
as far as possible from the ground planes, with the etching in a SF6 =He plasma. The completed devices are
distance being limited by the requirement of sufficient cooled in a 3 He refrigerator with 280-mK base temperature
coupling strength to the feedline. Figure 1(d) shows the and 70-dB cold attenuation between room temperature and
simulated feedline transmission for the device shown in the feedline input. Each resonator is characterized by
Fig. 1(a). The ratio between the resonance frequencies measuring the complex-valued feedline transmission near
of the two lowest modes extracted from the simulation is its resonance (Fig. 2). A fitting of the model from Ref. [25]
2.01, demonstrating that the nanowire resonators are to the data allows us to extract the resonance frequency and
the coupling and intrinsic quality factors [16,28]. Each
resonator is designed to have the coupling quality factor
(a)
(b)
QC ≈ 105 , the same order of magnitude as the intrinsic
Pin quality factor.
The highly disordered nature of NbTiN and the
extremely small cross-sectional area of the nanowires make
106
1.0
Normalized |S21|
(c)
0.5 105
Qi
200 µ m 500 nm
Data
-0.3 (a) Fit (b)
0.0 104
3.7188 3.7190 3.7192 0.5 1.0 1.5 2.0
f (GHz) T (K)
-0.6 (d)
106
2.20 2.25 2.30 fr (GHz)
|S21| (dB)
4.54
|J| (arb. units)
-0.9 4.06
3.72
2.98
Qi
-1.2 2.77
0
(c)
2.0 2.5 3.0 3.5 4.0 4.5 5.0 105
f (GHz) -130 -120 -110 -100 -90
Pin (dBm)
FIG. 1. Resonator design. (a) Dark-field optical image of a
typical nanowire resonator. (b) Schematic (not to scale) showing FIG. 2. Power and temperature dependence of intrinsic
the nanowire resonator, CPW feedline, and patterned ground quality factors of five nanowire resonators. (a) Normalized
plane. Here, NbTiN is shown in gray and Si substrate is white. absolute transmission around a typical resonance. The curve is
The ground plane is patterned in a square grid shape to enhance constructed from the best fit to the complex-valued feedline
the visibility for wire bonding. We see no evidence of ground transmission data [16,28]. (b) Temperature dependence of
plane patterning affecting performance of the resonators in the intrinsic quality factors measured at a fixed input power
magnetic field. (c) Scanning-electron-microscope enlargement of Pin ≈ −110 dBm. The symbols correspond to the legend in
the gap of a typical resonator. (d) Simulated feedline transmission (c). Two distinct regimes are observed for T < 1 K and
for the device in (a). The insets show (absolute) current T > 1 K, in which dominant loss is expected from TLS and
distributions along the nanowire for the fundamental and second quasiparticle dissipation, respectively. (c) Power dependence of
resonance modes, as well as an enlargement of the feedline intrinsic quality factors measured at 280 mK. The positive slope
transmission near the fundamental resonance. is consistent with TLS-dominated loss.
044004-2
HIGH-KINETIC-INDUCTANCE SUPERCONDUCTING … PHYS. REV. APPLIED 5, 044004 (2016)
the kinetic inductance the dominant contribution to the total superconducting resonators [22,23]. Moreover, we do not
inductance of the resonators. From the measured critical observe hysteretic behavior or abrupt jumps in f r with
temperature T c ≈ 9.3 K and room-temperature resistivity increasing B∥ . These effects plague standard CPW reso-
ρ ¼ 200 μΩ cm of the film, we estimate a sheet kinetic nators and are usually attributed to unstable magnetic-flux
inductance LS ≈ 35 pH=□ [29], close to the value vortices in the superconducting film [19,22,33,34]. These
38 pH=□ needed in a Sonnet simulation to match the findings suggest that vortex nucleation does not take place
resonance in Fig. 1(d) to the measurements. For a resonator in the nanowires. Vortices may still be created in the ground
of length l ¼ 2.9 mm and w ¼ 100 nm (2.77-GHz full- plane. However, due to the large separation between the
wave mode), this corresponds to a total in-line induct- nanowires and the ground planes, we expect only minimal
ance Ll ∼ 1 μH. current densities to be induced in the ground plane, thus,
weakly contributing to dissipation.
III. RESULTS AND DISCUSSION Further insight into the effect of a magnetic field on
the resonators is gained by orienting the field perpendicular
Figure 2(c) shows Qi of five resonators (w ¼ 100 nm) as
to the device plane. Figure 4(a) shows the dependence
a function of input power, Pin . We find Qi > 105 at Pin ≈
of Qi in seven nanowire resonators (widths ranging from
−130 dBm corresponding to an average occupation of the
w ¼ 100 to 700 nm) on B⊥ . The magnetic field resilience
resonator by hnph i ≈ 10 photons. The observed increase of
depends strongly on the nanowire width, and the narrowest
Qi with Pin indicates dominant loss by coupling to spurious
resonators show superior performance. We observe Qi >
two-level systems (TLSs) which saturate at high power
3 × 104 for the narrowest resonator (w ¼ 100 nm) for
[16,30,31]. This conclusion is further supported by the
B⊥ ≤ 350 mT [Fig. 4(a)]. This field range is one order
temperature dependence of Qi at Pin ≈ −110 dBm, corre-
sponding to hnph i ≈ 1000 [Fig. 2(b)]. Thermally excited
quasiparticles dominate loss only above 1 K ∼ T c =10,
consistent with previous studies of quasiparticle-induced
w (nm) fr (GHz)
dissipation in highly disordered thin-film resonators [32]. 5 100 2.94
10
150 3.42
200 3.86
A. Performance of the resonators in a magnetic field
250 4.22
The resilience of the nanowire resonators to a magnetic 500 4.00
field is seen in Fig. 3, which shows the typical dependence 600 4.33
700 4.63
Qi
5 × 105
0 10
Fit Fit
-20
fr/fr (10-3)
k (T-2)
-40
0
T
105 1
fr /fr (10-3)
fr (GHz) -60
Qi
-5
4.68
4.19 -80 (b) (c)
3.84 -10
Fit
3.05 0.0 0.1 0.2 0.3 0.4 0.5 100 400 700
2.85 0 1 2 3 4 5 6 B (T)
T w (nm)
B|| (T)
104
0 1 2 3 4 5 6 FIG. 4. Evolution of nanowire resonator characteristics with a
B|| (T) perpendicular magnetic field, B⊥ . (a) Qi as a function of B⊥ for
various nanowire widths w. The dips in Qi at low field suggest
FIG. 3. Evolution of nanowire resonator characteristics with coupling to magnetic impurities, similar to the case for B∥ in
an in-plane magnetic field B∥ (w ¼ 100 nm, T ¼ 280 mK, Fig. 2. The narrowest resonator retains Qi > 3 × 104 up to
Pin ≈ −110 dBm). The intrinsic quality factor Qi remains un- B⊥ ¼ 350 mT. (b) Fractional shift of the resonance frequencies
affected in the range 400 mT ≲ B∥ ≤ 6 T. The maximum B∥ is with B⊥ . Same symbols as in (a). The red curves are best fits of
limited by our experimental setup. (Inset) All fractional frequency Δf r =f r ¼ −k⊥ ðwÞB2⊥ to the data. (c) Best-fit coefficient k⊥
shifts fit to the same simple quadratic curve. versus w and best quadratic fit.
044004-3
N. SAMKHARADZE et al. PHYS. REV. APPLIED 5, 044004 (2016)
5 × 105 with applied field [Fig. 3 inset and Fig. 4(b)]. Fitting the
fractional shifts with the expression Δf r =f r ¼
−k∥ð⊥Þ B2∥ð⊥Þ , we extract the coefficients k∥ and a width-
dependent k⊥ ðwÞ [35,36]. These coefficients reflect the
increase in kinetic inductance of the superconducting
nanowire due to the Cooper-pair-breaking effect of the
Qi
fr (GHz)
105 4.68 external magnetic field. Taking into account that the
4.19 dominant contribution to the nanowire inductance is
3.84 kinetic, we have f r ∝ L−1=2k , where Lk is the kinetic
3.05 inductance of the resonator. Further, for T ≪ T c we have
2.85 (a) Lk ∝ T −1
c [29], and for small changes in frequency
0.0 0.1 0.2
Δ f r =fr ¼ − 12 ΔLk =Lk ¼ 12 ΔT C =T C .
B || (T) The applied magnetic field splits the time-reversal
0
degeneracy of the paired electrons, giving rise to an
effective depairing energy 2α [37]. In the dirty limit and
02
(b) (c)
0.
for small α, the change in T c due to this pair-breaking effect
±
00
fr /fr (10-3)
0.15
2.
-1 is linear in α: kB ΔT c ¼ −ðπ=4Þα. The penetration depth in
Bmin (T)
=
in
m
-2
B min
|| make use of the expression for α valid in the “thin film in
B min
0.10
parallel field” approximation, α ¼ 16 ðDe2 B2⊥ w2 =ℏÞ, where
T
-3
0.0 0.1 0.2 3 4 5 D is the electronic diffusion constant [37]. Thus, we
B|| (T) fr (GHz)
recover the experimentally observed scaling Δf r =f r ¼
−ðπ=48Þ½De2 =ðℏkB T c ÞB2⊥ w2 and extract the diffusion
FIG. 5. Signatures of electron-spin resonance near the
Zeeman field for five nanowire resonators. (a),(b) Data from
constant D ≈ 2 cm2 s−1 . This value is consistent with an
Fig. 3, expanded for clarity around 100 mT. The minima of the earlier estimate [38] of the electronic diffusion constant in
quality factors of the resonators occur at different values of the NbTiN thin films.
magnetic field. (c) Dependence of the magnetic field positions Furthermore, extending this geometrical scaling to the
of quality factor minima on the resonator frequencies. Black case of a parallel field yields an effective thickness of the
points correspond to B∥ measurements [Fig. 3(a)], and blue superconductor teff ≈ 3.5 nm. The reduced effective thick-
points to B⊥ measurements [Fig. 4(a)]. The straight line is the ness of the film in the context of a magnetic field expulsion
best fit to the data. is likely a combined effect of surface oxidation and the
suppression of shielding currents within a coherence length
from the edge.
of magnitude higher than the highest at which Qi ≈ 104 has
been previously reported [22].
Figures 3 and 4 show sharp dips in the quality factors of IV. SUMMARY
the resonators around B∥;⊥ ¼ 100 mT. Upon closer inspec- In summary, microwave resonators based on NbTiN
tion, it is evident that the magnetic field values, at which nanowires with extremely small cross section are highly
these dips occur, scale with the frequency of the resonators insensitive to a parallel magnetic field, with Qi remaining
[Figs. 5(a) and 5(c)]. This suggests that the resonators unaffected up to B∥ ¼ 6 T. Because of the high kinetic
couple with magnetic impurities in the silicon substrate or inductance of the nanowires, the resonators are expected to
at one of the interfaces. Moreover, the magnetic field produce an order of magnitude higher vacuum voltage
dependence of the frequency shifts of the resonators fluctuations compared to standard CPW resonators. Our
shows an incipient avoided crossing [Fig. 5(b)]. Fitting next experiments will focus on achieving strong coupling
the frequency dependence of the magnetic field positions of between these nanowire resonators and spin qubits in gate-
the quality factor minima with the condition for spin defined quantum dots, which have small electric dipole
resonance hfr ¼ gμB B, we extract the value for the moments and require a magnetic field.
Landé g factor: g ¼ 2.00 0.02 [Fig. 5(c)].
ACKNOWLEDGMENTS
B. Resonance frequency shift in a magnetic field We thank L. P. Kouwenhoven and his team for sputtering
Turning our attention to the shift of resonance frequency of NbTiN thin films and G. de Lange, T. M. Klapwijk,
induced by the magnetic field, we observe for both field and A. Wallraff for fruitful discussions. We acknowledge
orientations a quadratic shift of the resonance frequency funding by an ERC Synergy Grant, the Dutch Organization
044004-4
HIGH-KINETIC-INDUCTANCE SUPERCONDUCTING … PHYS. REV. APPLIED 5, 044004 (2016)
fr (GHz)
Fit 105
fundamental frequencies of five nanowire resonators 4.45
Qi
3.5 3.84
(w ¼ 100 nm) on the inverse of their total length, l. The 3.38
fr /fr (10-3)
fractional frequency shifts for all resonators follow the same -0.1
curve, demonstrating that the contribution from any out-of
-0.2
plane component due to field misalignment is negligible.
-0.3
0.0 0.2 0.4 0.6 0.8 1.0
APPENDIX C: ZERO-POINT VOLTAGE 104 B|| (T)
FLUCTUATIONS AT THE ENDS OF THE 0.0 0.2 0.4 0.6 0.8 1.0
NANOWIRE RESONATOR B|| (T)
Figure 6 demonstrates that the nanowire resonator acts FIG. 7. B∥ evolution of Qi in the four narrowest nanowire
as a distributed half-wavelength resonator. Thus, in the resonators shown on Fig. 4. (Inset) Fractional shifts of the four
lowest mode, current distribution on the resonator can be resonance frequencies as a function of the applied field. Symbols
expressed as correspond to those in the main figure.
044004-5
N. SAMKHARADZE et al. PHYS. REV. APPLIED 5, 044004 (2016)
044004-6
HIGH-KINETIC-INDUCTANCE SUPERCONDUCTING … PHYS. REV. APPLIED 5, 044004 (2016)
[18] C. Song, T. W. Heitmann, M. P. DeFeo, K. Yu, R. McDermott, transmission applied to superconducting devices, J. Appl.
M. Neeley, J. M. Martinis, and B. L. T. Plourde, Microwave Phys. 111, 054510 (2012).
response of vortices in superconducting thin films of Re and [29] A. J. Annunziata, D. F. Santavicca, L. Frunzio, G. Catelani,
Al, Phys. Rev. B 79, 174512 (2009). M. J. Rooks, A. Frydman, and D. E. Prober, Tunable super-
[19] D. Bothner, C. Clauss, E. Koroknay, M. Kemmler, T. Gaber, conducting nanoinductors, Nanotechnology 21, 445202
M. Jetter, M. Scheffler, P. Michler, M. Dressel, D. Koelle, (2010).
and R. Kleiner, Reducing vortex losses in superconducting [30] J. Gao, M. Daal, A. Vayonakis, S. Kumar, J. Zmuidzinas,
microwave resonators with microsphere patterned antidot B. Sadoulet, B. A. Mazin, P. K. Day, and H. G. Leduc,
arrays, Appl. Phys. Lett. 100, 012601 (2012). Experimental evidence for a surface distribution of two-
[20] D. Bothner, T. Gaber, M. Kemmler, D. Koelle, and R. level systems in superconducting lithographed microwave
Kleiner, Improving the performance of superconducting resonators, Appl. Phys. Lett. 92, 152505 (2008).
microwave resonators in magnetic fields, Appl. Phys. Lett. [31] A. D. O’Connell, M. Ansmann, R. C. Bialczak, M. Hofheinz,
98, 102504 (2011). N. Katz, E. Lucero, C. McKenney, M. Neeley, H. Wang,
[21] J. I. Vestgården, V. V. Yurchenko, R. Wördenweber, and E. M. Weig, A. N. Cleland, and J. M. Martinis, Microwave
T. H. Johansen, Mechanism for flux guidance by micro- dielectric loss at single photon energies and millikelvin
metric antidot arrays in superconducting films, Phys. Rev. B temperatures, Appl. Phys. Lett. 92, 112903 (2008).
85, 014516 (2012). [32] P. C. J. J. Coumou, M. R. Zuiddam, E. F. C. Driessen, P. J.
[22] S. E. de Graaf, A. V. Danilov, A. Adamyan, T. Bauch, and de Visser, J. J. A. Baselmans, and T. M. Klapwijk, Micro-
S. E. Kubatkin, Magnetic field resilient superconducting wave properties of superconducting atomic-layer deposited
fractal resonators for coupling to free spins, J. Appl. Phys. TiN films, IEEE Trans. Appl. Supercond. 23, 7500404
112, 123905 (2012). (2013).
[23] S. E. de Graaf, D. Davidovikj, A. Adamyan, S. E. Kubatkin, [33] V. Ranjan, G. de Lange, R. Schutjens, T. Debelhoir, J. P.
and A. V. Danilov, Galvanically split superconducting Groen, D. Szombati, D. J. Thoen, T. M. Klapwijk, R.
microwave resonators for introducing internal voltage bias, Hanson, and L. DiCarlo, Probing Dynamics of an
Appl. Phys. Lett. 104, 052601 (2014). Electron-Spin Ensemble via a Superconducting Resonator,
[24] A. Ghirri, C. Bonizzoni, D. Gerace, S. Sanna, A. Cassinese, Phys. Rev. Lett. 110, 067004 (2013).
and M. Affronte, YBa2 Cu3 O7 microwave resonators for [34] D. I. Schuster, A. P. Sears, E. Ginossar, L. DiCarlo, L.
strong collective coupling with spin ensembles, Appl. Phys. Frunzio, J. J. L. Morton, H. Wu, G. A. D. Briggs, B. B.
Lett. 106, 184101 (2015). Buckley, D. D. Awschalom, and R. J. Schoelkopf, High-
[25] P. K. Day, H. G. DeLuc, B. A. Mazin, A. Vayonakis, and Cooperativity Coupling of Electron-Spin Ensembles to
J. Zmuidzinas, A broadband superconducting detector Superconducting Cavities, Phys. Rev. Lett. 105, 140501
suitable for use in large arrays, Nature (London) 425, (2010).
817 (2003). [35] J. E. Healey, T. Lindstrm, M. S. Colclough, C. M. Muirhead,
[26] J. Luomahaara, V. Vesterinen, L. Grönberg, and J. Hassel, and A. Y. Tzalenchuk, Magnetic field tuning of coplanar
Kinetic inductance magnetometer, Nat. Commun. 5, 4872 waveguide resonators, Appl. Phys. Lett. 93, 043513 (2008).
(2014). [36] D.-H. Wu, C. A. Shiffman, and S. Sridhar, Field variation of
[27] The NbTiN films are deposited in an ATC-180 UHV custom the penetration depth in ceramic Y1 Ba2 Cu3 Oy , Phys. Rev. B
sputtering system from AJA International Inc., with dc- 38, 9311 (1988).
magnetron sputtering from a Nbð70%Þ=Tið30%Þ target, in a [37] M. Tinkham, Introduction to Superconductivity, 2nd ed.
reactive atmosphere of argon (90%) and nitrogen (10%) at (McGraw-Hill, New York, 1996).
pressure 3 mTorr and 250 W power. Typical values for [38] C. Tong, J. Stern, Megerian, H. LeDuc, T. Sridharan, H.
substrate temperature and voltage bias are 20 °C and −45 V, Gibson, and R. Blundell, in A low-noise NbTiN hot electron
respectively. bolometer mixer, Proceedings of the 12th International
[28] M. S. Khalil, M. J. A. Stoutimore, F. C. Wellstood, and Symposium on Space THz Technology, California, 2001
K. D. Osborn, An analysis method for asymmetric resonator (JPL Publication, Pasadena, California, 2001), p. 253.
044004-7