Multicontact measurements of a superconducting Sn nanowire

Article abstract of DOI:10.1063/1.2759471

Multicontact transport measurements were performed on a superconducting tin nanowire to probe its local properties and especially the formation of phase-slip centers (PSCs). We find that normal metal contacts strongly disturb the behavior of the nanowire. In particular, they provide an efficient escape way for the energy dissipated at the PSC and thereby limit its size. We were able to observe PSC at 30 mK with an associated relaxation time typical to that of the inelastic scattering time for bulk normal tin.

Full text of DOI:10.1063/1.2759471

Multicontact measurements of a superconducting Sn nanowire
D. Lucot, F. Pierre, D. Mailly, K. Yu-Zhang, S. Michotte, F. de Menten de Horne, and L. Piraux
Citation: Applied Physics Letters 91, 042502 (2007); doi: 10.1063/1.2759471
View online: http://dx.doi.org/10.1063/1.2759471
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/91/4?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Jitter analysis of a superconducting nanowire single photon detector
AIP Advances 3, 072135 (2013); 10.1063/1.4817581
Photon-number resolving detector based on a series array of superconducting nanowires
Appl. Phys. Lett. 101, 072602 (2012); 10.1063/1.4746248
Direct-write, focused ion beam-deposited, 7 K superconducting C–Ga–O nanowire
Appl. Phys. Lett. 96, 262511 (2010); 10.1063/1.3458863
Timing jitter of cascade switch superconducting nanowire single photon detectors
Appl. Phys. Lett. 95, 132503 (2009); 10.1063/1.3237172
Development of phase-slip centers in superconducting Sn nanowires
Appl. Phys. Lett. 85, 3175 (2004); 10.1063/1.1804608
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
134.176.129.147 On: Sun, 14 Dec 2014 13:55:23

APPLIED PHYSICS LETTERS 91, 042502 ͑2007͒
Multicontact measurements of a superconducting Sn nanowire
D. Lucot,a͒ F. Pierre, and D. Mailly
Laboratoire de Photonique et de Nanostructures (LPN-CNRS), Route de Nozay, 91460 Marcoussis, France
K. Yu-Zhang
Laboratoire de Miscroscopies et d’Étude de Nanostructures (LMEN), Université de Reims,
21 rue Clément Ader, BP 138 51685 Reims Cedex 2, France
S. Michotte, F. de Menten de Horne, and L. Piraux
Unité de Physico-Chimie et de Physique des Matériaux (PCPM), Research Center on Microscopic and
Nanoscopic Electronic Devices and Materials (CERMIN), Université Catholique de Louvain, Place Croix du
Sud 1, B-1348 Louvain-la-Neuve, Belgium
͑Received 13 May 2007; accepted 22 June 2007; published online 23 July 2007͒
Multicontact transport measurements were performed on a superconducting tin nanowire to probe
its local properties and especially the formation of phase-slip centers ͑PSCs͒. We find that normal
metal contacts strongly disturb the behavior of the nanowire. In particular, they provide an efficient
escape way for the energy dissipated at the PSC and thereby limit its size. We were able to observe
PSC at 30 mK with an associated relaxation time typical to that of the inelastic scattering time for
bulk normal tin. © 2007 American Institute of Physics.
͓DOI: 10.1063/1.2759471͔
Destruction of superconductivity by a large current or
magnetic field in a quasi-one-dimensional geometry ͓diam-
eter less or of the order of the superconducting coherence
length ␰͑T͔͒ has been studied a few decades ago on
whiskers1,2 or microbridges3 with diameter of ϳ0.5 ␮m,
therefore restricting these studies to temperature only a few
tens of millikelvins below the critical temperature Tc. In par-
ticular, the formation of phase-slip centers ͑PSCs͒ at current
above the critical one has attracted much interest. A PSC
consists of time periodic oscillations of the superconducting
order parameter in a core of size ϳ2␰ at the Josepson fre-
quency accompanied by a 2␲ change in the phase. This dis-
sipative dynamical process gives rise to a nonequilibrium
distribution of quasiparticles. These quasiparticles diffuse out
from the dynamical region and the phase-slip core is accom-
panied by larger regions on either side, where a stationary
nonequilibrium distribution gradually decreases to zero with
a characteristic distance ⌳Q,4 associated with the quasiparti-
cle relaxation time ␶ ͑⌳Q=͑D␶ ͒1/2, with D the diffusion
segments of the wire, with different probes for the injection
of the current and voltage measurements. This enabled us to
highlight PSCs with a spatial extention much smaller than
with two contact measurements.
The nanowires were grown by electrodeposition ͑−0.5 V
vs Ag/AgCl reference electrode͒ from a tetrafluoroborate
bath ͓41.8 g/l Sn͑BF4͒2 in water solution͔ into
a
22-␮m-thick nanoporous track-etched polycarbonate
membranes11 with average pore diameter near the surface of
56 nm. Chloroform was used to dissolve the membrane. Mi-
croscopic characterizations have also revealed the presence
of polycarbonate residues and oxide layer8,12 on the surface
of the wire, which reduced the conducting section. This re-
sidual insulating layer prevents from obtaining good Ohmic
contacts with the probe by a simple evaporation of metal
directly on the wire. In order to make reliable transport mea-
surements, the contact resitance has to be improved
͑Ͻ10 ⍀͒ by reducing the thickness of the insulator layer
using an ion beam etching light enough to avoid damaging
the conducting part of the wire.
The lithography process is based on previous works per-
formed by Naud et al.13 on carbon nanotubes. Electron beam
lithography ͑EBL͒ was performed using a JEOL 5DIIU
nanowriter. The first step of the process is the deposition of
Ti/Au alignment marks on the Si substrate to pinpoint the
wires. A few drops of freestanding nanowires suspended in
chloroform were then spilled onto the substrate. After drying
the solution, the wires are strongly stuck on the substrate
surface probably by van der Walls forces. This allows us to
perform all the nanofabrication steps ͑resist spinning, lithog-
raphy, lift-off, etc.͒ with a very high accuracy. Optical and
scanning electron microscopy investigations are then used to
select and locate isolated nanowires. We then design the
probe setup adapted to the geometry of a particular nano-
wire. Prior to the evaporation of the Ti/Au ͑10/100 nm thick
and 250 nm wide͒ metallic pad, an Argon ion bombardment
at 500 V is performed in situ in the evaporator chamber to
destroy the insulating layer. Micrograph on Fig. 1 shows the
Q
Q
coefficient͒. ⌳Q is temperature dependent and diverge as
5
−1/4
͑Tc−T͒
.
Recently, thanks to the improvements in fabrication
techniques, superconducting nanowires with much smaller
diameter have been realized by many groups6–10 allowing to
probe their properties far below Tc while still being quasi-
one-dimensional. Suspended carbon nanotubes have been
used as templates for an evaporation process which resulted
in amorphous structures.6 In the case of electrodeposited
nanowires, a single wire can be contacted by evaporating
metallic electrodes on both sides of the template.7 However,
up to now, only two contacts at the extremities of the nano-
wires were realized. Here we present a way to attach many
low-resistance Ohmic contacts along electrodeposited Sn
nanowires and perform electrical measurements on different
a͒
Also at Laboratoire de Miscroscopies et d’Étude de Nanostructures
͑LMEN͒, Universite de Reims, 21 rue Clément Ader, BP 138 51685 Reims
Cedex 2, France; electronic mail: damien.lucot@lpn.cnrs.fr
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
0003-6951/2007/91͑4͒/042502/3/$23.00 91, 042502-1 © 2007 American Institute of Physics
134.176.129.147 On: Sun, 14 Dec 2014 13:55:23

042502-2
Lucot et al.
Appl. Phys. Lett. 91, 042502 ͑2007͒
FIG. 1. Scanning electron micrograph of the Sn nanowire and the Ti/Au
pads attached to it. Inset is a zoom of the defect in segment S1. A bias
current was applied between the ends of the nanowire.
final multiprobe connection of a Sn nanowire. Because of the
small diameter of the nanowires and the low melting point of
Sn ͑232 °C͒, special care is taken to prevent the wires from
being burnt out during the bonding and by unwanted electri-
cal discharges during the manipulations. For this purpose
bonding pads are attached to the measurement pads through
a chip 1 k⍀ Cr resistances, deposited during a second EBL
step. By using a four-point measurement method, this resis-
tive protection does not affect the measurements. All the re-
sults presented here have been obtained in a dilution refrig-
erator with a base temperature of 30 mK, using current-
biased lock-in techniques.
One of the major advantage of this technique compared
to measurements7,8 performed with the wire inside the nano-
porous template is the possibility to check the morphology of
the nanowire. In particular, we can notice the presence of a
defect in segment S1 ͑inset of Fig. 1͒. The length of the three
segments S1, S2, and S3 are 1.20, 3.10, and 1.15 ␮m, re-
spectively. The average diameters of S2 and S3 are 90 and
93 nm, including the surrounding insulator layer of unknown
thickness. The difference ⌬R=R300 K−R4 K between ambient
resistance R300 K and T=4 K resistance R4 K is essentially
FIG. 2. ͑a͒ Low temperature resistance of the three segments S1, S2, and
S3. ͑b͒ Evolution of the resistance of S1, S2, and S3 when sweeping up and
down the external perpendicular magnetic field at 30 mK. Resistance mea-
surements were made with a 10 nA bias current.
Tϳ3.75 K and then continuously decreases until it reaches
“zero” ͑if fall below our voltage detection limit, which is
about 10 nV͒, interpreted as a “local” Tc below the contacts
of about 2.2 K. A superconducting area appears in the middle
of S2 at the Tc of bulk Sn because the segment is long
enough such that its center is not sensitive to the presence of
N electrodes. On the other hand, S3 which is smaller is more
influenced by IPE and needs a lower temperature in order to
establish superconductivity. Note that the local Tc is the same
for both segments as expected. We also observed IPE on S1
but superconductivity is strongly suppressed in the defect
region, with a Tc of only ϳ1 K. We found no evidence of
quantum phase slip6 at low T, probably because the diameter
is too large to observe them.
Figure 2͑b͒ shows the suppression of superconductivity
by a magnetic field perpendicular to the axis of the wire at
30 mK for the three segments. Nonzero resistance reappears
at about the same field for all segments ͑Hϳ0.29 T͒. This
onset is due to the transition of Nc around the contacts where
superconductivity is strongly weakened. We recover the nor-
mal resistance at the critical magnetic field Hcϳ0.4 T for
S2, which is about 12 times larger than the bulk value
͑Hbculk=0.03 T͒. For S3, Hc is smaller since IPE is more im-
portant. In dirty superconductors at T=0 K, the penetration
due to the phonon contribution. For Sn, ⌬␳ph=␳300 K−␳
4 K
Х8.5 ␮⍀ cm.14 Given the length L of a segment, we can
determine its conducting cross section S=⌬␳phL/⌬R. For S2
and S3, having ⌬R=134 and 54.7 ⍀, we found diameters of
51 and 49 nm. The resistivity above the transition was
␳
4 Kϳ2.1 ␮⍀ cm. Using the free-electron model and the
Fermi velocity of Sn F=0.7ϫ106 m/s,5 we find the elastic
v
mean free path ᐉϳ50 nm, presumably limited by the wire
diameter. In the dirty limit, we estimate ␰͑T=0 K͒
Ϸ͑␰0ᐉ͒1/2Ϸ100 nm, where ␰0=230 nm is the coherence
length for bulk Sn. In the case of S1 we measure
⌬R=27.6 ⍀ and a relatively large R4 K value ͓Fig. 2͑a͔͒.
This is
a
signature of the observed constriction
͑inset of Fig. 1͒ which completely modifies the electrical
properties.
Figure 2͑a͒ presents the low temperature behavior of the
three segments. Their wide superconducting transition is
principally due to the inverse proximity effect ͑IPE͒ induced
by the normal ͑N͒ metal electrode which strongly weakens
the superconducting ͑S͒ order near the N-S interfaces. In-
deed, the Ti/Au voltage probes15 introduce normal areas Nc
in the S nanowire around the contacts. However, contrary to
a two-point measurement,16 the resistance goes to zero with a
supercurrent carried by Andreev bound states across the nor-
mal areas such as in a mesoscopic SNS junctions.17 For S2,
length ␭ may be expressed as a function of ␭ ͑0͒=34 nm for
L
Sn, in the local approximation given by:5 ␭=␭ ͑0͒͑1
L
+␰0/ᐉ͒1/2. For S2 and S3, we estimate ␭Ϸ80 nm. Since ␭ is
larger than the diameter of the wire, the Meissner effect is
incomplete and the critical magnetic field is increased
compared to the bulk value. For a cylinder ͑diameter dӶ␭͒
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
upon decreasing temperature the resistance drops at
are finds5 Hc/Hbulk=8␭/d. Hcϳ0.4 T corresponds to
c
134.176.129.147 On: Sun, 14 Dec 2014 13:55:23

042502-3
Lucot et al.
Appl. Phys. Lett. 91, 042502 ͑2007͒
PSC. Surprisingly, we observe about the same number of
PSCs as in almost identical Sn nanowires but 25 to 50 times
longer and kept inside the membrane with only contacts at
the extremities.7 This small extension of PSCs in our nano-
wire is most probably due to the presence of the nearby
normal pads which help to limit the low temperature exten-
sion of ⌳Q.
In conclusion, we deposited normal multicontacts on a
superconducting Sn nanowire, which allows us to probe the
local electrochemical potentiel of the quasiparticles. We have
observed the signature of PSC at very low temperature
whose extension are much smaller than in previous studies
on similar wires but with different connecting geometries.
These results also highlight the important role played by the
multicontacts that provide rapid relaxation of quasiparticles
created by these PSC.
FIG. 3. Current-voltage characteristics at 30 mK of the three segments
when the current is swept up and down ͑open and full markers, respec-
tively͒. Arrow for S1 indicates a 3 ⍀ jump.
The authors thank J. P. Maneval and F. R. Ladan for
fruitful discussions and the Laboratoire des Hauts Polymères
of the UCL for providing the membrane. One of the authors
͑S.M.͒ is a postdoctoral researcher of the FNRS. Another
authors ͑F.M.H.͒ acknowledges financial support from the
FRIA. This work has been supported by the EU Commission
FP6 NMP-3 Project No. 505457-1 ULTRA-1D and by the
Interuniversity Attraction Poles Program ͑P6/42͒, Belgian
State, Belgian Science Policy.
dϳ50 nm which is the diameter of our wire. Large increase
of Hc to ϳ0.6 T results from the diameter reduction of S1 in
the constriction.
Figure 3 shows the destruction of superconductivity by
sweeping the current up and down at 30 mK. The presence
of a defect in S1 depresses notably its critical current with a
transition to the normal state already at 23 ␮A ͓the arrow at
23 ␮A points out a 3 ⍀ jump that achieves the return to
normal state resistance for S1 ͑see Fig. 3͔͒. At around 8 ␮A,
there is a common jump in S1 and S2 corresponding to the
return to the normal state of the region below the contact
they share. The same is true at around 20 ␮A but this time
between S2 and S3. The second jump in S3 at 29 ␮A has no
influence on the other segments and corresponds to the tran-
sition of the right contact at the end of the nanowire. Except
for the constriction, we observe, as for the application of a
magnetic field, that superconductivity is first suppressed in
the regions around and under the contacts by the application
of a current. Because the wire is well thermalized by normal
leads, the extent of normal areas in the remainder of the
segment is very limited. Therefore, except for the very last
jump at high current in the respective voltage-current char-
acteristics all the other steps are a signature of PSCs.
Due to the small length of these segments relative to the
quasiparticule diffusion length ⌳Q, once a PSC appears in a
segment, it also affects the segment nearby ͑see dotted lines͒,
probably by a larger number of incoming quasiparticles. Us-
1W. W. Webb and R. J. Warburton, Phys. Rev. Lett. 20, 461 ͑1968͒.
2J. Meyer and G. von Minnigerode, Phys. Lett. 38, 529 ͑1972͒.
3W. J. Skocpol, M. R. Beasley, and M. Tinkham, J. Low Temp. Phys. 16,
145 ͑1974͒.
4For a review, R. Tidecks, Current-Induced Nonequilibrium Phenomena in
Quasi-One-Dimensional Superconductors, Springer Tracts in Modern
Physics Vol. 121 ͑Springer, Berlin, 1990͒.
5M. Tinkham, Introduction to Superconductivity, 2nd ed. ͑Dover,
New York, 2004͒.
6A. Bezryadin, C. N. Lau, and M. Tinkham, Nature ͑London͒ 404, 971
͑2000͒.
7S. Michotte, S. Mátéfi-Tempfli, and L. Piraux, Appl. Phys. Lett. 82, 4119
͑2003͒.
8M. Tian, J. Wang, J. S. Kurtz, Y. Liu, and M. H. W. Chan, Phys. Rev. B
71, 104521 ͑2005͒.
9F. Altomare, A. M. Chang, M. R. Melloch, Y. Hong, and C. W. Tu,
Phys. Rev. Lett. 97, 017001 ͑2006͒.
10M. Zgirski, K.-P. Riikonen, V. Touboltsev, and K. Arutyunov, Nano Lett.
5, 1029 ͑2005͒.
11See, for example, E. Ferain and R. Legras, Nucl. Instrum. Methods Phys.
Res. B 131, 97 ͑1997͒, and references therein.
12D. Bouchet, E. Roy, K. Yu-Zhang, and Y. Leprince-Wang, Eur. Phys. J.:
Appl. Phys. 30, 193 ͑2005͒.
ing the model of Skocpol et al.,3 the voltage across a PSC of
13C. Naud, G. Faini, D. Mailly, and H. Pascard, C. R. Acad. Sci., Ser. IIb
Mec. Phys. Astron. 327, 945 ͑1999͒.
1
¯
¯
extent 2⌳Q is given by VϷ2͑I−Is͒␳⌳Q/S, with IsϷ 2Ic the
14International Critical Table, edited by the National Research Council of
the USA ͑McGraw-Hill, New York, 1929͒, Vol. VI, p. 124.
15The nonsuperconducting properties of Ti/Au reservoirs are confirmed by
measuring the resistance of a 10 nm Ti/100 nm Au bilayer down to
30 mK; the thick Au layer destroys the superconductivity of Ti by inverse
proximity effect.
time-averaged supercurrent through the core. From the first
PSC of S2 we estimate ⌳Qϳ1.25 ␮m ͑=length of S2 8
,
20
since this PSC in S2 corresponds to an 8 ⍀ jump over a
20 ⍀ transition͒. With D=ᐉ F/3=115 cm2/s, this gives
v
␶Q=͑⌳Q͒2/Dϳ1.3ϫ10−10 s which is what we expect for in-
elastic electron-phonon relaxation processes in Sn.5 In S3
͑length Ͻ2⌳Q͒ the voltage jump of the PSC is much smaller
because there is not enough space to accommodate an entire
16G. R. Boogaard, A. H. Verbruggen, W. Belzig, and T. M. Klapwijk,
Phys. Rev. B 69, 220503 ͑2004͒.
17T. Hoss, C. Strunk, T. Nussbaumer, R. Huber, U. Staufer, and C.
Schönenberger, Phys. Rev. B 62, 4079 ͑2000͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
134.176.129.147 On: Sun, 14 Dec 2014 13:55:23