Summary of PITOS-3 point contacts on 400nm Bi2Se3 film (2

Summary of PITOS-3 point contacts on 400nm Bi2Se3 film (2

Conductance spectra of NbN-Bi2Se3-Au junctions:
Tunneling gaps & ZBCPs
Gad Koren & Tal Kirzhner
Faculty of Physics, Technion – Israel Institute of Technology
Haifa, 32000, Israel
Online: Condmat arXiv:1207.5352v1 (July 24, 2012)
http://lanl.arxiv.org/abs/1207.5352
Some basics & what is done here
Topological Insulators (TOI) are band-gap insulators with conduction by
surface states due to a strong spin-orbit interaction
• The hallmark of Topological Superconductivity is the presence of
Majorana fermions (MF), which are a special sub-group of surface Andreev
bound states (SABS) at zero energy
In the experiments, they will appear as ZBCPs
•
• We use proximity induced superconductivity in a TOI in NbN-Bi2Se3-Au junctions,
similar to Yang & Li Lu, PE in Sn/Bi2Se3, PRB 85, 104508 (2012)
& G. Koren et al., PE in Bi/Bi2Se3, PRB 84, 224521 (2011)
• We characterize the films structure & growth orientation,
& measure transport (R vs T), and conductance spectra vs H & T
• The observed ZBCPs are ~0.15mV & ~0.9mV wide (FWHM). The narrower one
is kT limited (~2K) & the broader one might be a result of thermally smeared
split ZBCP (Yamakage & Tanaka, PRB 85, 180509(R) (2012) or
other non zero bound states
(200)STO
(0,0,15)BiSe
1000
W
(0,0,30)BiSe
??
W
100nm Bi2Se3 /70nm NbN/(100)STO
c=2.88 nm hexagonal Bi2Se3
c=0.253 hexagonal NbN
a=0.433 nm cubic NbN
d=0.3904 nm (100) STO
rombohedral (0 0 3) Bi2Se3
(400)NbN
10000
(0,0,21)BiSe
(300)STO
(002)NbN
(0,0,24)BiSe
X-ray intensity (counts)
100000
(200)NbN
1000000
(001)NbN
(0,0,6)BiSe
(100)STO
(0,0,9)BiSe
X-ray diffraction of Bi2Se3 /NbN bilayer on (100)STO
Shows hexagonal Bi2Se3 & mostly cubic but some hexagonal NbN
100
110
100
10
0
10
20
30
40
50
2θ (deg)
60
70
80
90
120
An AFM image of a
100nm thick Bi2Se3 on 70nm thick NbN bilayer on (100) SrTiO3
Crystallized well,
in laterally disordered
hexagonal form
1x1 µm2
Junctions layout
• 10 base electrodes were
patterned first in a NbN film on
half the wafer
• Then a cover bilayer of
Au/Bi2Se3 was deposited on the
other half of the wafer with the
aid of a shadow mask
• No further patterning was done,
only Ag paste was added on the
NbN contact pads
Junction with in-situ interface
(Full patterning)
Cover (Au)
• Start with a Bi2Se3/NbN
bilayer (base)
• Then add the Au cover electrode
Junctions with ex-situ interface
Junctions' layout
overlap junctions'
area is 100x150 µm2
NbN base
electrodes
Au on Bi2Se3
cover electrode
Barrier (Bi2Se3)
I
Substrate (SrTiO3)
Base (NbN)
AFM images
NbN films
As deposited
virgin films (a & b)
A large via-hole
After patterning (c)
of the base electrode
(d) A small junction
with full patterning
A schematic cross section near a via hole
Via-hole in a trilayer junction
100nm Au - N
Proximity-induced
Topological superconductivity
100nm Bi2Se3 - TOI
(*) N'
N
S
N'
70 nm NbN - S
(100) SrTiO3 substrate
(*) Since the Bi2Se3 film is heavily hole doped by
Se vacancies, it can also be considered as N'
R vs T of a 100x150 µm2 area ex-situ NbN-Bi2Se3-Au junctions
40
10000
R (Ω)
8000
6000
4000
30
R (Ω)
J1
J2
J3
J4
J5
J6
J7
J8
J9
J10
J1
J2
J3
J4
J5
J6
J7
J8
20
10
2000
0
0
0
50
100
150
T (K)
•
•
•
•
200
250
300
2
3
4
5
6
7
T (K)
Insulating vs T due to the low Tc NbN film (deposited in excess N2)
Superconductive transition only at Tc~8K (onset)
Junctions’ resistance at low T is of about 10Ω
Junctions with a metallic behavior at low T have ZBCPs (#3, #5 & #6)
8
Conductance spectra of all working ex-situ junctions
Narrow (kT- limited)
and broad ZBCPs
0.4
J1 1.81 K
J2 1.81 K
J3 1.86 K
J4 1.81 K
J5 1.89 K
J6 1.82 K
J7 1.81 K
J8 1.85 K
0.3
dI/dV (1/Ω)
All junctions have
coherence peaks
of the NbN
SC electrode at
1-1.2 mV, consistent
with the BCS ratio:
2Δ/kTc~3.5
0.2
0.1
The broad one might
be a thermally smeared
split one, or with a narrow
ZBCP superimposed
0.0
-4
-3
-2
-1
0
V (mV)
1
2
3
4
Conductance spectra with s-wave BTK fits of a few NbN-Au/Bi2Se3 junctions
J4 1.813K 0T
Z=0.79 ∆=1.2 mV Γ=0.2 mV
3.5
• Andreev
3.0
1.2
2.5
2.0
• tunneling
1.5
Normalized dI/dV (arb. u.)
Normalized dI/dV (arb. u.)
J8 1.85K 0T
Z=1.3 ∆=1.25 mV Γ=0.2 mV
1.3
1.0
1.1
1.0
0.9
-6
-5
-4
-3
-2
-1
0
1
2
3
4
5
6
-4
V (mV)
-2
0
2
4
V (mV)
• Broad ZBCP
J6 1.82K 0T
0.9mV FWHM
Z=3 ∆=1 mV Γ=0.24 mV
Normalized dI/dV (arb. u.)
3.0
2.5
2.0
1.5
• Narrow ZBCP
0.15mV FWHM
1.0
0.5
J3 1.80K 0T
Z=1.55 ∆=1.1 mV Γ=0.22 mV
3.5
3.0
Normalized dI/dV (arb. u.)
3.5
2.5
2.0
1.5
1.0
0.5
0.0
-3
-2
-1
0
V (mV)
1
2
3
-4
-3
-2
-1
0
V (mV)
1
2
3
4
Conductance spectra at various temperatures
J4 1.81 K
J4 2.51 K
J4 3.13 K
J4 4.41 K
J4 4.86 K
J4 5.29 K
J4 5.65 K
J4 5.78 K
J4 6.04 K
J4 6.16 K
J4 6.22 K
J4 6.35 K
0.4
BCS-like gap Δ
0.3
dI/dV (1/Ω)
Coherence peaks
(CP) & SC gap are
suppressed
vs T (as usual)
∆ (mV)
0.2
1.0
J4 0 T
0.5
0.0
1 2 3 4 5 6
T (K)
0.1
0.0
-5
-4
-3
-2
-1
0
V (mV)
1
2
3
4
5
Conductance spectra under various magnetic fields
Coherence peaks
& SC gap are
suppressed
vs H (also as usual)
J4 1.84 K 0 T
J4 1.84 K 0.5 T
J4 1.84 K 1.5 T
J4 1.85 K 2.5 T
J4 1.87 K 4 T
J4 1.87 K 6 T
J4 1.83 K 0 T
0.4
A small ZBCP at
Low T, suppressed
under 6 T, and
recovers when
field is removed
dI/dV (1/Ω)
0.3
J4 1.86K 0 T
J4 1.87 K 6 T
J4 1.82 K 0 T
0.22
0.20
0.18
0.16
-0.4 -0.2 0.0 0.2 0.4
V (mV)
0.2
1.0
∆ (mV)
0.1
Need higher fields
to test BCS behavior
of the gap Δ
dI/dV (1/Ω)
0.5
0.0
0.0
-5
-4
-3
-2
-1
J4 1.85 K
0
2
4
0
V (mV)
6
1
8 H (T)
2
3
4
5
Conductance spectra at various T, CP & broad ZBCP
Coherence peaks,
SC Gap and ZBCP
height, width & area
are suppressed
vs H
Yamakage & Tanaka,
PRB 85, 180509(R) (2012)
J6 1.85 K
ZBCP HWHM (mV)
ZBCP height (1/Ω )
ZBCP area (mV/Ω )
0.20
0.3
0.20
2σ=0.2 mV
0.15
dI/dV (1/Ω)
Broad ZBCP can be
due to a split ZBCP,
with superimposed
narrow ZBCP
0.4
0.2
J6 1.82K 0T
Gauss fit
0.1
-0.4 -0.2 0.0 0.2 0.4
0.15
J6 1.83 K 0T
J6 1.83 K 0.5 T
J6 1.84 K 1.5 T
J6 1.85 K 2.5 T
J6 1.86 K 4 T
J6 1.87 K 6 T
J6 1.82 K 0 T
0.10
0.0
0 1 2 3 4 5 6
H (T)
0.05
-3
-2
-1
0
V (mV)
1
2
3
Conductance spectra at various T, narrow ZBCP
0.4
Narrow ZBCP with
kT limited width
(thermal smearing
at 1.8 K)
Can be due to
Majorana fermions
Yang & Li Lu PRB 85,
104508 (2012)
Mourik et al., Science 336,
1003 (2012).
Das et al., arXiv:1205.7073
0.03-0.05 mV widths at 0.1K
0.2
J3 1.81 K 0 T
J3 1.81 K 0 T
J3 1.81 K 0.05 T
J3 1.81 K 0.1 T
J3 1.81 K 0.2 T
J3 1.81 K 0.3 T
J3 1.81 K 0.5 T
J3 1.83 K 1.5 T
J3 1.83 K 1.5 T
J3 1.85 K 3 T
J3 1.85 K 3 T
J3 1.87 K 5 T
ZBCP height (1/Ω)
0.3
dI/dV (1/Ω)
Its’ ZBCP height is
suppressed
differently than the
height of the
broad ZBCP
0.3
J3 1.85K
J6 1.85 K
0.2
0.1
0.0
0 1 2 3 4 5 6
H (T)
~kT=0.15 mV
0.1
-0.4
-0.3
-0.2
-0.1
0.0
V (mV)
0.1
0.2
0.3
0.4
Smaller 10x5 µm2 area ex-situ NbN-Bi2Se3-Au junctions
R(junction) went down!
Se vacancies doping by milling.
⇒ Get a low R shunt resistance,
Strongly shunted junction.
2.75
R (Ω)
10000
R (Ω)
J1
J2
J3
J4
J5
J7
3.00
12000
8000
2.50
6000
2.25
1
2
3
4
T (K)
5
6
7
4000
2000
0.378
-6
-4
-2
V (mV)
0 2
4
6
0.38
J1 1.83 K
0
0.37
dI/dV (1/Ω)
dI/dV (1/Ω)
J1 1.83 K
J1 3.75 K
J1 4.43 K
0.376
0
50
100
150
200
T (K)
0.36
0.35
0.374
0.34
One weak ZBCP survived
No coherence peaks now
(must be a milling effect)
0.372
0.370
0.368
-1.5
-1.0
-0.5
0.0
V (mV)
0.5
1.0
1.5
250
300
10x5 µm2 area ex-situ NbN-Bi2Se3-Au junctions under different fields
0.378
The ZBCP is suppressed
with increasing
magnetic field
0.376
J1 1.94 K 0T
J1 1.83 K 0T
J1 1.81K 0.1T
J1 1.81K 0.5T
J1 1.82K 1T
J1 1.83K 2T
J1 1.85K 4T
Also the critical current
goes down with increasing
field (the “dips”)
dI/dV (1/Ω)
0.374
0.372
0.370
0.368
-1.5
-1.0
-0.5
0.0
V (mV)
0.5
1.0
1.5
Layout of a fully lithographically-patterned, in-situ junction
• 10 base electrodes are
patterned first in a
Bi2Se3/NbN bilayer on half
the wafer
• Then a cover gold film is
deposited on the whole
wafer, followed by further
patterning of this cover
electrode
• The resulting junction area
is A~5x5 µm2
A fully lithographically-patterned junction layout
overlap junction
area is 5x5 µm2
Bi2Se3/NbN
base electrode
Top view
Au cover electrode
R vs T of bridges in a 70nm thick NbN & of 100nm Bi2Se3/70nm NbN bilayer
R2(Ohm)
R4(Ohm)
R6(Ohm)
R7(Ohm)
R8(Ohm)
R9(Ohm)
R10(Ohm)
1000
R (Ω)
• ρ of bare NbN bridges at RT
is similar to that of YBCO,
but is not metallic vs T.
• Tc of the bare NbN film is ~12.5K
(onset)
• Jc(2K) of NbN bridges with
0.07x10µm2 cross-section area is
~1.4MA/cm2, which is equivalent
to 10mA/10µm width
June 7 2012, 2012, NbN8 on (100)STO, ~10J/cm2
2kpls on Nb target 600 C, 34-38mT N2 flow,
µ-bridges 0.07x10x100 µm3 for IC measurements
100 µA bias, on Ag paste pads.
ρ(RT)~500x0.07x10-4x10/100=350 µΩcm
1500
500
0
0
50
100
150
200
250
300
T (K)
May 29, 2012, NbN7-BiSe bilayer on (100)STO
base ~1x10-6Torr, 3.5Wx10Hz on ~2.2x2 mm2 area, φ~10J/cm2
2kpls on Nb target 600 C, 40-38mT N2 flow,
Then 500pls on Bi2Se3 target under 2.5x10-6Torr, 400C, 0.4W at 3.33Hz
100 µA bias (weakly pressed contacts)
• An un-patterned bilayer. When
patterned, its’ resistance R would
probably be similar to that of the
patterned NbN bridges.
R (Ω)
12
8
R2(Ohm)
R3(Ohm)
R4(Ohm)
R5(Ohm)
R6(Ohm)
R7(Ohm)
R8(Ohm)
R9(Ohm)
R10(Ohm)
4
0
0
50
100
150
T (K)
200
250
300
R versus T of 100nm Bi2Se3/70nm NbN-100nm Au in-situ junctions
June 5, 2012, NbN7-BiSe-Au junctions on (100)STO
5x5 µm2 junctions' area, 100nm Au-100nm Bi2Se3/70nm NbN
1 mA bias J1 & J2 are shorted since J2 is disconnected, reversed order
June 4, 2012, NbN7-BiSe-Au junctions on (100)STO
5x5 µm2 junctions' area, 100nm Au-100nm Bi2Se3/70nm NbN
1 mA bias J1 & J2 are shorted since J2 is disconnected, reversed order
1.0
0.3
0.8
R1(Ohm)
R2(Ohm)
R3(Ohm)
R4(Ohm)
R5(Ohm)
R6(Ohm)
R7(Ohm)
R8(Ohm)
R9(Ohm)
R10(Ohm)
0.4
0.2
0.0
0
50
100
150
T (K)
200
250
300
R (Ω)
R (Ω)
0.6
0.2
R1(Ohm)
R2(Ohm)
R3(Ohm)
R4(Ohm)
R5(Ohm)
R6(Ohm)
R7(Ohm)
R8(Ohm)
R9(Ohm)
R10(Ohm)
0.1
0.0
0
2
4
6
8
10
12
14
16
18
20
T (K)
• Metallic vs T. Very low ρ even when compared to that of the unpattern bilayer
(x10 lower). This seems to be due to additional Se vacancies in the Bi2Se3 layer
(hole doping), possibly created by the direct Ar ion beam on the Bi2Se3 when
milling the gold cover electrode.
Should use a shadow mask instead
• Tc of the NbN electrode is ~12.5K (onset) & ~8K of the junction
• Junctions’ resistance at low T is very low, about 0.1Ω (cleanest interface due to
in-situ deposition of the bilayer). In the previous ex-situ junctions with the x600
larger area, this resistance was x100 larger!!!
Conductance spectra of J7 at different T & H
dI/dV(C7 2.003K 0T June 5 2012)
dI/dV(C7 3.104K 0T June 5 2012)
dI/dV(C7 4.173K 0T June 5 2012)
dI/dV(C7 5.537K 0T June 5 2012)
dI/dV(C7 6.862K 0T June 5 2012)
dI/dV(C7 7.794K 0T June 5 2012)
dI/dV(C7 8.566K 0T June 5 2012)
dI/dV(C7 9.009K 0T June 5 2012)
15 Cann't reach Andreevgap at 2K (~1.5mV),
reach Ic of NbN first
(~0.35mV & ~5mA)
14
20x20µm2
13
Bi2Se3
/NbN
12
Au
A=5x5µm2
11
10
-1.5
Same 70nm thick NbN film
has Ic(2K)~ 10mA/10µm width
Ic dips
-1.0
-0.5
0.0
0.5
1.0
1.5
15
Same 70nm thick NbN film
has Ic(2K)~ 10mA/10µm width
dI/dV(C7 1.895K 0.1T June 4 2012)
dI/dV(C7 1.909K 0.5T June 5 2012)
dI/dV(C7 1.93K 1T June 4 2012)
dI/dV(C7 1.873K 2T June 5 2012)
dI/dV(C7 1.977K 3T June 4 2012)
dI/dV(C7 1.88K 4.5T June 5 2012)
dI/dV(C7 2.007K 6T June 4 2012)
dI/dV(C7 2.003K 0T June 5 2012)
dI/dV(C7 1.829K 0T June 5 2012)
V (mV)
14
13
• 6T is not sufficiently high
to suppress the low bias
conductance
dI/dV (1/Ω)
dI/dV (1/Ω)
• Looks like s-wave Andreev,
but gap can’t be at 0.35mV.
• Reach Ic first via the 5+10 µm length
of the bilayer-Au overlap boundary.
See Ic dips.
Hc2 of NbN
is 19 T
12
11
10
Ic dips
9
-1.5
-1.0
-0.5
0.0
V (mV)
0.5
1.0
1.5
Conductance spectra of J1, an in-situ junction at different T
Ic dips at ~±1, 1.5 & 2mV
9
Same seemingly
“emergent” ZBCP in
reference NbN-Au junction
with no Bi2Se3 in between
dI/dV (1/Ω)
No CP, Ic is reached first
No ZBCP, just decreasing
critical current with
increasing T (and H)
J1
J1
J1
J1
J1
J1
J1
1.99 K
3.05 K
4.10 K
7.22 K
7.82 K
8.40 K
8.71 K
5x5 µm2
Au-Bi2Se3/NbN
junction
8
20
100x150 µm2
15 Au-NbN
junction
7
J2 9.75 K
J2 10.18 K
J2 10.38 K
10
5
Need a strong tunneling
Barrier to observe ZBCPs
0
-1.0
6
-2
-1
-0.5
0.0
0
V (mV)
0.5
1.0
1
2
Reference NbN-Au junctions
Reference wafer: May 17, 2012, NbN-6-Au on (100)STO
Base fingers: 70nm thick NbN, 600C, 46mT N2 flow
3.5 Wx10Hz on ~2.3x2 mm2 area, 2kpls, φ~10J/cm2
Cover 1/2 wafer: 100nm Au - 150C, vac ~6e-6 Torr, 3.3kpls 0.96Wx10Hz
On Ag paste pads & reversed # order, 1 mA bias
0.30
Tconset ~ 12.5K
0.25
R2(Ohm)
R4(Ohm)
R6(Ohm)
R7(Ohm)
R8(Ohm)
R9(Ohm)
R10(Ohm)
0.20
R (Ω)
Rjunction ~ 50 mΩ
0.15
0.10
Maybe NbN is not fully milled
(shorted junctions) & resistive
I-V with 50mΩ is due to the Au
Or junctions are too large 0.025mm2
--> resistance is very low!
0.05
35
30
Similar to junctions with Bi2Se3 barrier
==> has nothing to do with TOI here!
dI/dV (1/Ω)
10
dI/dV(C10 9.751K 0T May 17 2012)
dI/dV(C10 9.754K 0T May 17 2012)
dI/dV(C10 10.18K 0T May 17 2012)
dI/dV(C10 10.393K 0T May 17 2012)
NbN-Au
junctions
dI/dV(C9 9.751K 0T May 17 2012)
dI/dV(C9 10.179K 0T May 17 2012)
dI/dV(C9 10.375K 0T May 17 2012)
25
20
0.00
-10mA --> 10mA
dI/dV(C6 10.096K 0T May 17 2012)
15
12
13
14
T (K)
• Same as in Au-Bi2Se3/NbN junctions !!
• The “ZBCP” seems to be an Ic effect
due to heating.
10
5
0
-0.8
11
heating
starts transition
-0.6
-0.4
-0.2
0.0
V (mV)
0.2
0.4
0.6
0.8
Conclusions
• Proximity superconductivity is induced by NbN in Topological Bi2Se3 films
• Tunneling and Andreev bound states (ZBCP) were observed
• The broad ZBCP (~0.9mV) can’t originate in simple Andreev since coherence peaks
exist. It may be due thermally smeared, split ZBCP.
• The narrow ZBCP (~0.15mV) width is kT limited (~1.8K)
• Both types of ZBCP could originate in Majorana fermions of a Topological SC