Materials aspects of quantum metrology sensored · • To build a high-throughput microwave loss...
Transcript of Materials aspects of quantum metrology sensored · • To build a high-throughput microwave loss...
Materials aspects of quantum metrology
Alexander Tzalenchuk
445th WE-Heraeus SeminarQuantum Measurement and Metrology with Solid State Devices1-5 November 2009 Physikzentrum Bad Honnef, Germany
Scope
Real materials and precise measurements: friends or foes?
• Superconducting resonators vs. dielectrics• Quantum Hall effect vs. graphene
Superconducting resonatorsat low-temperatures
Tobias Lindström(NPL)
Joanne E. Healey, Chris M. Muirhead(University of Birmingham)
Yuichi Harada, Yoshiaki Sekine(NTT BRL)
Motivation
• To understand and quantify the loss mechanisms in superconducting resonators at low temperatures
• To build a high-throughput microwave loss measurement system based on the resonators
• To compare and quantify losses in practical materials coupled to this system
Astronomy instrumentation communityQubit community
In-line λ/2 co-planar resonator
f0
QIn Out
d
C
L… …
CcCc
f0 ∝ 1/d
d
In-line λ/2 co-planar resonator
superconductor
Where it all started: KIDs
•Photons break Cooper pair•Kinetic contribution to total inductance increases•Resonance frequency decreases
CLLf
kg )(1
0 +∝
In-line λ/2 co-planar resonator
W=10 μmS=5 μm
G=4 μm
In Out
-0.50 -0.25 0.00 0.25 0.500
300000
600000
900000
1200000
-0.5 -0.4 -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 0.4 0.5-200
-150
-100
-50
0
50
100
150
200
-0.50 -0.25 0.00 0.25 0.50-1.00
-0.75
-0.50
-0.25
0.00
0.25
0.50
0.75
1.00
QMagnetic flux density, mT
Freq
uenc
y, k
Hz
Magnetic flux density, mT
25 mK 50 mK 100 mK 100 mK reverse 280 mK300 mK, -21 dBm300mK, -27 dBm 500 mK
df0/d
B (G
Hz/
T)
Magnetic flux density, mT
Magnetic tuning of superconducting resonators
•Frequency tuned by magnetic field•No deterioration of Q – no additional losses
Magnetic tuning of superconducting resonators
More details in: J. Healey et al., APL 93, 043513 (2008)
Non-linear London equations
( )
( )( )
λ
βλλ
ββ
β
∝
⎥⎦
⎤⎢⎣
⎡+=
Δ−+=
⎟⎟
⎠
⎞
⎜⎜
⎝
⎛⎟⎟⎠
⎞⎜⎜⎝
⎛−=+−=
k
c
B
c
sssqpsss
LTH
HTTTH
TkT
vvTvTenJvenJ
2
2''
)()(1
),0(),(
exp1)0()(
1)(
( ) ( ) ( )( ) ( )( )224/0,10,, CTK HHTLTLTfHTf β−=
First conclusions
• Losses in superconducting resonators at low temperature are not determined by the superconductor
• Small magnetic field (current in a control line) can tune the resonance frequency of high quality resonators by hundreds of linewidths
Temperature dependence of resonance frequency
0 500 1000 1500 20005712.24
5712.26
5712.28
5712.30
5712.32
5712.34
5712.36
5712.38
Centre Freq Fr
eque
ncy,
MH
z
Temperature, mK
Superconductor Mattis-Bardeen
Shunting λ/4 co-planar resonator
Shunting λ/4 co-planar resonator
In Out
Frequency multiplexing
5.9850 5.9855 5.9860 5.9865 5.9870
-8
-6
-4
-2
0
Mag
nitu
de(d
B)
Frequency (GHz)
4000 4500 5000 5500 6000 6500 7000 7500 8000-10
-8
-6
-4
-2
0
Fre que ncy(MHz)
S21
(dB
)
• Can multiple on-chip resonators be made more compact while maintaining high Q?
Shunting lumped-element resonator
Shunting lumped-element resonator
Transmission line
L C
Shunting lumped-element resonator
Fitting
yedcxxjQg
xjQR
QCZg
fffx
j
u
u
uc
−⎥⎦
⎤⎢⎣
⎡++
+++
=
=
−=
θ
πϖ
)21(2)21(2
:Minimize
)2( coupling
shift frequency normalised
200
0
0
Resonator3 parameters
Skewingparasitics
Magnitudeoffset
Phaserotation
Ideal Smith chart
Fitting
• Qu = 345514
• f0=6958658487 Hz
• g = 3.9074
• θ = −0.86441
• c = 17281
• d = 0.08
Low temperature frequency shift
Low-temperature frequency shift
0 1 2 3 4 5 6 7 8
-10
-5
0
5
x 10-7
hf/kBT
δ f
0 2 4 6 8 10 12 14
0
2
4
6
8x 10-5
hf/kBT
δ f
Frequency shift as a function of the normalized inverse temperature at several different powers
- λ/4- LE
SiO2 / Si
sapphire
- LE
Resonant process
TLF
⎟⎟⎠
⎞⎜⎜⎝
⎛
⎟⎟⎠
⎞⎜⎜⎝
⎛+
=Tk
PPc
nd
B
c
r 2tanh
13
4
0
22 ϖ
ε
ϖπα h
αr –resonant absorptionω -measurement frequencyn -density of TLFd -dipole momentc0 -speed of lightPc -critical powerT1, T2 -relaxation and dephasingΨ -digamma function significant when thermal energy is smaller than photon energy
212
02
23 where
and 1
TTdcP
PPconstPPP
c
ccr
ε
α
h=
<←>←∝
( )
⎭⎬⎫
⎩⎨⎧
+Ψ=
⎥⎦
⎤⎢⎣
⎡−−⎟⎟
⎠
⎞⎜⎜⎝
⎛=
−=
Δ
TikTg
TgTgTTnd
TTT
Bπϖϖ
ϖϖεε
εεεε
221Re),(
),(),(ln3
2)(
)()(0
0
2
0
0
hεεδ Δ
−=Δ
≡20
0 Ffff
F –filling factor
critical power
Permittivity from αr & Kramers-Kronig relations
0 1 2 3 4 5 6 7 8
-10
-5
0
5
x 10-7
hf/kBT
δ f
0 2 4 6 8 10 12 14
0
2
4
6
8x 10-5
hf/kBT
δ f
Low-temperature frequency shift
)(Tg⎟⎟⎠
⎞⎜⎜⎝
⎛
0
lnTT
SiO2 / Si
sapphire
Substrate losses
6
52
108.2106
tan3−
−
××
== πδε iFnFd
All data on a particular substrate is fitted using a single parameter:
SiO2 / Si
sapphire
More details in: T. Lindström et al., PRB 80, 132501 (2009)
Comparison of losses in different substrates
10-6 10-4 10-2 10010-6
10-5
10-4
10-3
tan
δ
W/W0
Sapphire50 mK
SiO2 / Si3 LE resonators
x4 temperatures
(50-340 mK)∝W-0.4
212
02
23 where
and 1
TTdcP
PPconstPPP
c
ccr
ε
α
h=
<←>←∝
W – energy in the resonator at f0W 0~10-16 J
02121)1(2 ϖinu PSSQW −=
A bit more subtle
• Can resonators probe thin films?
Current distribution in a lumped element resonator
μ-sensitive
ε -sensitive
Atomic Layer Deposition
•ALD involves the use of a pair of reagents.• each reacts with the surface completely• each will not react with itself
•Application of this AB Scheme•Reforms the surface•Adds precisely 1 monolayer
•Pulsed Valves allow atomic layer precision in growth•20 nm•Al(CH3)3 alternating H2O cycles•~100 C deposition temperature
ALD0 2 4 6 8
-2
0
2
4
6
8
10
12
14 x 10-6
hf/kB T
δ f
Larger nd2
Conclusions:
• Low-temperature losses in high quality superconducting microwave resonators are determined by the surrounding dielectrics
• The non-monotonic temperature dependence of the resonance frequency can in most cases be accurately described by resonant absorption of microwaves by two-level fluctuators (TLF), which couple to the resonator via their electric-dipole moments.
• Quantitative information about the distribution of two-levelfluctuators can be extracted from such measurements and different materials can be compared.
• This approach can be used for high-throughput (frequency-multiplexed) diagnostics of dielectric materials for quantum circuits at the temperatures, microwave frequencies and power levels relevant to their operation.
Quantum Resistance StandardBased on Epitaxial Graphene
Sergey Kubatkin, Samuel Lara-Avila, Alexei Kalaboukhov, Sara Paolillo (Chalmers)Mikael Syväjärvi, Rositza Yakimova (Linkoping)Olga Kazakova, T.J.B.M. Janssen (NPL)Vladimir Fal'ko (Lancaster)
Some unpublished figures have been deleted.More details in: A. Tzalenchuk et al., arXiv:0909.1220
On electrical metrology1
Dear Editor,
The formulation of the Ohm’s law has to be modified in the following way: “When carefully selected and perfectly prepared materials are used, then on a good day and with some skill one can construct an electric circuit for which the ratio of voltage to current measured over a finite time gives a constant value after introduction of relevant corrections.”
Signed
1 Physicists Make Jokes. Moscow, 1968, p. 218
Consistency of the Josephson effect
The voltage position of the microwave-induced Shapiro steps is independent of the composition and geometry of the junctions to a part in 1016
This is used in the SI realisation of the voltage standard.
Tsai et al., 1983
Junctions
1000
500
0
-500
-1000-10 0 10Array Voltage (mV)
Bia
s Cur
rent
(μA
)
The most precise law of physics (?)
Resistance metrology
Hall Effect
Edwin Hall, 1879
Uy = B Ix / neeRxy = UH / Ix = B /nee
Von Klitzing, 1980
Quantization of transverse (Hall) resistance in rational fractions of the resistance quantum
Ω== 557(18) 812.807 252K ehR
© D.R. Leadley, Warwick University 1997
Quantum Transport in 2DEG in Strong Magnetic Fields
L.Landau 1930© D.R. Leadley, Warwick University 1997
Materials
New materials are sought
MOSFET III-V
K. von Klitzing, Nobel lecture, December 9, 1985
Dirac Fermions in quantizing magnetic field
2nehRxy ±=
14/ 2
+±=
nehRxy
2/14/ 2
+±=
nehRxy
Landau level spacing
BBBcBLL
e
BBFBLL
F
kBkmBekkEmm
kBkBevkE
smv
/20)*(//067.0*
/420/2/
/106
≈==Δ=
≈=Δ
=
hh
h
ϖ
Graphene flakes
Pablo Jarillo-Herrero,
Experiment
Epitaxial graphene
LEEMphotoemission
Is SiC graphene sick?
C - terminated face: good mobility, stacked layers kill QHESi - terminated face: poor quality and mobility
Growth
AFM images of the sample reveal large flat terraces on the surface of Si-face of 4H-SiC(0001) substrate with graphene after high-temperature annealing in Ar atmosphere.
Patterning
Twenty Hall bar devices of different sizes, from 160 μm x 35 μm down to 11.6 μm x 2 μm were produced on 0.5 cm2 wafers. AFM shows that the Hall bars are patterned across many substrate terraces
Magnetotransport
300 K
μ≈ 2400 cm2/Vs at RT
4.2 K
μ≈ (4000-7500) cm2/Vs
ns ≈ (5-10)x1011 cm-2
Contact resistance
Rc=1.5 Ω
cf. kΩ in contacts to exfoliated graphene
Breakdown current
Non-dissipative transport
300 mKRxx<0.2 mΩ at Isd= 12 μA4.2 KRxx<2.4 mΩ at Isd= 2.5 μA
Cryogenic current comparator
B=0
Quantisation accuracy
Accuracy of resistance quantization:
3 ppb at 300 mK
Allan deviation from RK/2 vs. measurement time τ. The square root dependence indicates purely white noise.
0
12
1
2)]()([
)(ˆ0
PyyP
k kky
∑ = + τ−τ=τσ
Comparison
Quantisation 10000 times more accurate than in graphene flakes,comparable to semiconductor QH standards at 300 mK,still accurate at 4.2 K.
Conclusions
• Graphene Quantum Hall Resistance standard potentially superior to its seminconducting counteparts; optimization is routine!
• Scalable graphene technology!
• The role of substrate in doping and scattering processes of epi-graphene?
More details in: A. Tzalenchuk et al., arXiv:0909.1220
How about friends and foes?
Real materials and precise measurements: friends or foes?
The question is irrelevant.Important:
• Understanding real materials by precise measurements:– Superconducting resonators vs. dielectrics
• Making use of novel materials in quantum applications:– Quantum Hall effect vs. graphene