Upload
phamthuan
View
230
Download
0
Embed Size (px)
Citation preview
Introduction to Quantum Mechanics of Superconducting
Electrical Circuits
• What is superconductivity?
• What is a Josephson junction?
• What is a Cooper Pair Box Qubit?
• Quantum Modes of Superconducting Transmission Lines•See R.-S. Huang PhD thesis on Boulder 2004 web page
What is superconductivity?
See: Boulder Lectures 2000
Cooper pairs are like bosons in a BEC -- except --size of Cooper pair spacing between electrons
F 1k ξ >>
k
k−Complex order parameter like BEC
( ) k kr c cψ↑ − ↓
∼( )( ) i rr e ϕψ ψ∼
Excitation Gap for Fermions
broken pair
k
In a semiconductor the gap is tied to the spatial lattice(only occurs at one special density commensurate with lattice)
In a superconductor, gap is tied to fermi level. Compressible.
WHY SUPERCONDUCTIVITY?
E few electrons N ~ 109
total numberof electrons
N even“forest” of states
~ 1 eV 2∆ ~ 1 meVsuperconducting gap
SUPERCONDUCTINGNANOELECTRODE
ATOM
Energy vs. Particle Number on Grain
…. N-1 N N+1 N+2 N+3 N+4 ....
∆
Josephson tunneling between two grains
2∆
1 2,N N 1 21, 1N N− + 1 22, 2N N− +
Josephson Tunneling ICoherent tunneling of Cooper pairs
Restrict Hilbert space to quantum ground states of the form
1 22 , 2 , n ... 3, 2, 1,0, 1, 2, 3...N n N n− + = − − − + + +
…. –4 -3 -2 -1 0 +1 +2 +3 +4 ….. n
Josephson Tunneling IIRestrict Hilbert space to quantum ground states of the form
1 22 , 2 , n ... 3, 2, 1,0, 1, 2, 3...nN n N n− + ⇒ = − − − + + +
…. –4 -3 -2 -1 0 +1 +2 +3 +4 …..
J 1 12
n
n
ET n n n n=+∞
=−∞
= − ⎡ + + + ⎤⎣ ⎦∑
Tight binding model: single ‘particle’ hopping on a 1D lattice
28 /t
JGEe h
∆=
normal state conductance
sc gap
N.B. not 1/JE ∝ ∆ ∆
Josephson Tunneling III
…. –4 -3 -2 -1 0 +1 +2 +3 +4 …..
J 1 12
n
n
ET n n n n=+∞
=−∞
= − ⎡ + + + ⎤⎣ ⎦∑Tight binding model
‘position’ n‘wave vector’ (compact!)
‘plane wave eigenstate’
ϕ
" "n
i n ikx
ne n eϕϕ
=+∞
=−∞
= =∑'
J'
' 1 ' ' ' 1n n
i n
n n
T E n n n n e nϕϕ=+∞ =+∞
=−∞ =−∞
= − ⎡ + + + ⎤⎣ ⎦∑ ∑
Josephson Tunneling IV
…. –4 -3 -2 -1 0 +1 +2 +3 +4 …..'
J'
' 1 ' ' ' 1n n
i n
n n
T E n n n n e nϕϕ=+∞ =+∞
=−∞ =−∞
= − ⎡ + + + ⎤⎣ ⎦∑ ∑
J 1 12
ni n
n
ET e n nϕϕ=+∞
=−∞
= − ⎡ + + − ⎤⎣ ⎦∑
J cos( )T Eϕ ϕ ϕ= −
ϕ
Supercurrent through a JJ
…. –4 -3 -2 -1 0 +1 +2 +3 +4 …..
‘position’ n‘momentum’ ϕ J cos( )T Eϕ ϕ ϕ= −
c c
1 sin
current:2(2 ) sin
2sin
J
J
J
Edn dTdt d
dn eI e Edt
eI I I E
ϕϕ
ϕ
ϕ
= =
= =
= ≡
Wave packet group velocity
ϕ
Charging Energy
1 22 , 2 , ... 3, 2, 1,0, 1, 2, 3...nN n N n n− + ⇒ = − − − + + +
2
2
(2 ) 2 2c2
c 2
4eC
eC
U n E n
E
= =
≡
n
U
‘Second’ Quantization
" "n
i n ikx
ne n eϕϕ
=+∞
=−∞
= =∑
Number Operator: n̂ iϕ∂
≡ −∂
2c J
2
c J2
ˆ4 cos( )
4 cos( )
( ) ( ) ( )j j j
H E n E
H E E
H E
ϕ
ϕϕ
ϕ ϕ ϕ
= −
∂= − −
∂Ψ = Ψ
ϕ Now viewed as coordinate
( )j ϕΨ Describes quantum amplitude for (quantum) phase
Weak Charging Limit: Josephson Plasma Oscillations
c JE E 2c J
2
c J2
2 2
c J2
ˆ4 cos( )
4 cos( )
4 12
H E n E
H E E
H E E
ϕ
ϕϕ
ϕϕ
= −
∂= − −
∂
⎡ ⎤∂≈ − + − +⎢ ⎥∂ ⎣ ⎦
Simple harmonic oscillator
spring constant JK E=
J c J8E Eω =2c
18
ME
=mass
Weak Charging Limit: Josephson Plasma Oscillations
c JE E (heavy ‘mass’, small quantum fluctuations)
J c J8E Eω =
ϕ
spring constant JK E=
2c
18
ME
=mass
Weak anharmonicity can be usedto form qubit and readout levels
Strong Charging Limit: Cooper Pair Box
ϕ Fluctuates too wildly;better to use number (position) representationc J4E E≥
n
U
-1 0 +1 +2
C
500aF2K
CE∼∼
Random offset charge due to built in asymmetry
THE BARE JOSEPHSON JUNCTION: SIMPLEST SOLID STATE ATOM
Josephson junction with no wires to rest of world:superconductor
tunnel oxide layersuperconductor
( )2ˆ4 / 2 cos .....C r JH E n Q e E ϕ= − − +Hamiltonian:
2
2
8 /
2
r
tJ
CJ
QGEe heEC
∆=
=
negligibleterms atsmall T
residual offset charge3 parameters:
Josephson energy
Coulomb energy
IMPERFECTIONS OF JUNCTION PARAMETERS
TUNNEL CHANNEL ELECTRIC DIPOLESCHARGED IMPURITY
_+
-1nm
CAN IN PRINCIPLE MAKE JUNCTION PERFECT, BUT MEANWHILE.....
JUNCTION PARAMETER FLUCTUATIONS
( )( )( )
statr r r
statJ J J
statC C C
Q Q Q t
E E E t
E E E t
= + ∆
= + ∆
= + ∆
param. dispers.
Qrstat random!
EJstat 10%
ECstat 10%
noise sd@1Hz
∆Qr/2e ~10-3Hz-1/2
∆EJ /EJ 10-5-10-6Hz-1/2
∆EC /EC <10-6Hz-1/2?
Oth order problem: get rid of randomness of static offset charge!
Cooper pair box (charge qubit)2 solutions: - control offset charge with a gateor
- make EJ /EC very large RF-SQUID (flux qubit)Cur. Biased J. (phase qubit)
SENSITIVITY TO NOISE OF VARIOUS QUBITS
junction noise port noise
qubits ∆Qr ∆EJ ∆EC technical quantum back-action
charge
flux
phase
chargequbit
fluxqubit
phasequbit
COOPER BOX STRATEGY
gate charge can be controlledNg=CgU/2e
U
compensatestatic
offset chargewith gate
( )2
gˆ4 cos .....C JH E n N E ϕ= − − +
n
U
n
U
Jcos 1 12
n
Jn
ET E n n n nϕ=+∞
=−∞
= − = − ⎡ + + + ⎤⎣ ⎦∑
charge deg. point
Mapping to two level system
n
U
↑ ↓
0 1
1ˆ2
z
n σ−=
( )2
gˆ4 cos .....C JH E n N E ϕ= − − +
JJ
2cos 1 1
2
n
Jx
n
ET E n n En nϕ σ=+∞
=−∞
= − = − ⎡ + + + −⎤ ⇒⎣ ⎦∑
Mapping to two level system
n
U
↑ ↓
0 1
1ˆ2
z
n σ−=
( )g2 1 2 constant.....2
z xJC
EH E N σ σ= − − − +
2ˆ ˆ[ ]n n=
Rotate:
x zz x
→ −→ + x
z( )g2 1 2
2constant.....x zJ
CEH E N σ σ += − − +
Mapping to two level system
( )g2 1 22
constant.....x zJC
EH E N σ σ += − − +
2 2x z
H B
B B
σ
ε±
=
= ± +
i
EJ
( )g2 1N −
B2
JE
0
n
U
n
U
SPLIT COOPER PAIR BOX
φIcoil
Ng
Q
Ugate
I
maxJ J
0
cos 2E E φπφ
⎛ ⎞= ⎜ ⎟
⎝ ⎠
JJ is actually a SQUID
Split Cooper Pair Box
( )g2 1 22
constant.....x zJC
EH E N σ σ += − − +
Vg
JJ is actually a SQUID
maxJ J
0
cos 2E E φπφ
⎛ ⎞= ⎜ ⎟
⎝ ⎠
(Buttiker ’87; Bouchiat et al., 98)
Superconducting Tunnel Junction as aCovalently Bonded Diatomic ‘Molecule’
(simplified view at charge deg. point)
1 pairsN +
pairsN1 pairsN +
pairsN810N ∼ 1 mµ∼tunnel barrier
aluminum island
aluminum island
Cooper Pair Josephson Tunneling Splits the Bonding and Anti-bonding ‘Molecular Orbitals’
bondinganti-bonding
Bonding Anti-bonding Splitting
( )12
ψ ± = ±810 1+
810 1+
810
810
anti-bonding bonding J 7 GHz 0.3 KE E E− = ∼ ∼Josephson coupling
J
2zEH σ= −
bonding
anti-bonding
↑ =
↓ =
Dipole Moment of the Cooper-Pair Box(determines polarizability)
2
1 2 3
1/(2 )1/ 1/ 1/
Cd e LC C C
=+ +
Vg
1 nm
L = 10 µm
+ ++ +
+ +
- -- -
- -
/gE LV=
~ 2 - md e µ
Vg
0
1C
2C
3C
J
2z x
gd VL
EH σσ −= − bonding
anti-bonding
↑ =
↓ =
Energy, Charge, and Capacitance of the CPB
↑
↓J
2z x
gE dH V
Lσ σ= − −JE
Ene
rgy
dEQdV
=
Cha
rge
Cap
acita
nce
no chargesignal
↑
↑
↓
↓
charge
dQCdV
= polarizability is state dependent
/g gC V e0 1 2deg. pt. = coherence sweet spot
sweet spot
E1
Ene
rgy
kK
)(
B
Ng=CgU/eφ/Φ0=δ/2π
E0
hν01
RESPONSES OF BOXIN EACH LEVEL
kk
EQU
∂=
∂k
kEIφ
∂=
∂island chargeloop current
12
2k
kELφ
−⎛ ⎞∂
= ⎜ ⎟∂⎝ ⎠
2
2k
kEC
U∂
=∂
box capacitancebox inductance
Dephasing, decoherence, decay
150.658 10 eV-s−≈ ×
151 nV acting on charge for 1 s 10 eV-se µ −=
Spins very coherent but hard to couple and very hard to measure.
Avoiding Dephasing1
0 1
0Ebox
2ggC V
e
→
←
1/2
2'sweet spot' for T
ν
( ) ( )( )( ) cos / 2 1 sin / 2 0i tt e ϕψ θ θ= +0
( ) ( )t
t i dt tϕ ν′ ′= − ∫ g
0Q QVν∂
= → → − ← ← =∂
Vion et al. Science 2002
16460 16462 16464 16466 16468
30
32
34
36
38
sw
itchi
ng p
roba
bilit
y (%
)
ν (MHz)
Lorentzian fit:ν01= 16463.64 MHzQ ~20 000
Ng = 0.5φ = 0τ = 4 µs
‘Atomic spectral line’
Courtesy M. Devoret
0 1 2 4 5 6 730
35
40
45
50
switc
hing
pro
babi
lity
(%)
time tw (µs)
RELAXATION TIME AT OPTIMAL POINT
Exponential fit:T1 = 1.84 µs
Q1 ~ 90 000
tw
PRINCIPLE OF RAMSEY EXPERIMENT
prepa-ration
freeevolution
measu-rement
90ºpulse
90ºpulse
time
prepa-ration
freeevolution
measu-rement
90ºpulse
90ºpulse
0.0 0.1 0.2 0.3 0.4 0.5 0.6
30
35
40
45
sw
itchi
ng p
roba
bilit
y (%
)
time between pulses ∆t (µs)
RAMSEY FRINGES
νRF = 16409.50 MHz
0.0 0.1 0.2 0.3 0.4 0.5 0.6
30
35
40
45
sw
itchi
ng p
roba
bilit
y (%
)
time between pulses ∆t (µs)
Qϕ ~ 25000 (8000 fringes)
∆t
Courtesy M. Devoret