You are on page 1of 32

Fundamentals of Microwave Superconductivity

Short Course Tutorial


Superconductors and Cryogenics in Microwave Subsystems
2002 Applied Superconductivity Conference
Houston, Texas

Steven M. Anlage
Center for Superconductivity Research
Physics Department
University of Maryland
College Park, MD 20742-4111 USA
anlage@squid.umd.edu

Objective
To give a basic introduction to superconductivity,
superconducting electrodynamics, and microwave
measurements as background for the Short Course
Tutorial Superconductors and Cryogenics in
Microwave Subsystems

Outline
Superconductivity
Microwave Electrodynamics of
Superconductors
Experimental High Frequency
Superconductivity
Current Research Topics
Further Reading

Superconductivity
The Three Hallmarks of Superconductivity
Superconductors in a Magnetic Field
Where is Superconductivity Found?
BCS Theory
High-Tc Superconductors
Materials Issues for Microwave Applications

The Three Hallmarks of Superconductivity


Zero Resistance

Complete Diamagnetism Macroscopic Quantum Effects


Flux

Tc
Temperature

Magnetic Induction

DC Resistance

I
T>Tc

T<Tc

Tc
Temperature

Flux quantization = n0
Josephson Effects

Zero Resistance
R = 0 only at = 0 (DC)

R > 0 for > 0


E
Quasiparticles

The Kamerlingh Onnes resistance


measurement of mercury. At 4.15K the
resistance suddenly dropped to zero

Energy
Gap

Cooper Pairs

Perfect Diamagnetism
Magnetic Fields and Superconductors are not generally compatible
The Meissner Effect

Superconductor


B = 0 H + M = 0
(T)
magnetic
penetration
depth

superconductor

H = H 0 e z / L

T<Tc

Spontaneous exclusion of magnetic flux

=0

surface
screening
currents

(0)
is independent of frequency ( < 2)

H
T>Tc

H (z )

vacuum

Tc

T
The Yamanashi MLX01 MagLev
test vehicle achieved a speed of
343 mph (552 kph) on April 14, 1999

Macroscopic Quantum Effects


= e i

Superconductor is described by a single


Macroscopic Quantum Wavefunction

Flux

Consequences:
Magnetic flux is quantized in units of 0 = h/2e (= 2.07 x 10-15 Tm2)
R = 0 allows persistent currents
Current I flows to maintain = n 0 in loop
n = integer, h = Plancks const., 2e = Cooper pair charge

Magnetic vortices
have quantized flux
A vortex

|(x)|

superconductor

B(x)

Line cut
Type II
<<

vortex
core

vortex
lattice
screening
currents
Sachdev and Zhang, Science

Macroscopic Quantum Effects

Continued

Josephson Effects (Tunneling of Cooper Pairs)

1 = 1 e

i1

2 = 2 e

i 2

I = I c sin (1 2 )

DC Josephson
Effect

I
(Tunnel barrier)

e
VDC

1 2 =

eVDC

I = I c sin
t + 0

VDC

AC Josephson
Effect

e* 1
MHz
=
= 483.593420
Quantum VCO:
h 0
V

Superconductors in a
Magnetic Field

0
FL = J

The Vortex State

Hc2(0)

Normal
State

Lorentz
Force

B 0, R 0
vortex

Abrikosov
Vortex Lattice
Hc1(0)

B = 0, R = 0

Meissner State
Type II SC

T
Tc

Vortices also experience


a viscous drag force:

FDrag = vvortex

Moving vortices
create a longitudinal voltage

I
V>0

10

What are the Limits of Superconductivity?


Phase Diagram

Jc
Normal
State

Superconducting
State

Tc

Ginzburg-Landau
free energy density
11

Temperature
dependence

Currents

0Hc2

Applied magnetic field

BCS Theory of Superconductivity


Bardeen-Cooper-Schrieffer (BCS)
Cooper Pair
s-wave ( = 0) pairing
+
+

+
+

Spin singlet pair

+
+

+
+

S
First electron polarizes the lattice

Tc Debye e 1/ NV

Second electron is attracted to the


concentration of positive charges
left behind by the first electron

Debye is the characteristic phonon (lattice vibration) frequency


N is the electronic density of states at the Fermi Energy
V is the attractive electron-electron interaction

A many-electron quantum wavefunction made up of Cooper pairs is constructed


with these properties:
An energy 2(T) is required to break a Cooper pair into two quasiparticles (roughly speaking)
Cooper pair size: = vF
12

http://www.chemsoc.org/exemplarchem/entries/igrant/hightctheory_noflash.html

Where do we find Superconductors?

Also:
Nb-Ti, Nb3Sn, A3C60, NbN, MgB2, Organic Salts ((TMTSF)2X, (BEDT-TTF)2X),
Oxides (Cu-O, Bi-O, Ru-O,), Heavy Fermion (UPt3, CeCu2Si2,),
Electric Field-Effect Superconductivity (C60, [CaCu2O3]4, plastic),
Most of these materials, and their compounds, display spin-singlet pairing

13

The High-Tc Cuprate Superconductors


Layered structure quasi-two-dimensional
Anisotropic physical properties
Ceramic materials (brittle, poor ductility, etc.)
Oxygen content is critical for superconductivity

YBa2Cu3O7-

Tl2Ba2CaCu2O8

Two of the most widely-used HTS materials in microwave applications

14

Spin singlet pairing


d-wave ( = 2) pairing

HTS Materials Issues Affecting Microwave Applications


film
grain
boundaries
substrate

Most HTS materials made as epitaxial


thin films for use in planar microwave
devices

High-Tc small Cooper Pair size ( correlation length)

1
= vF vF

Tc

~ 1 2 nm for HTS materials used in microwave applications

Superconducting pairing is easily disrupted by defects:


grain boundaries
cracks
Josephson weak links are created, leading to:
nonlinear resistance and reactance
intermodulation of two microwave tones
harmonic generation
power-dependence of insertion loss, resonant frequency, Q
15

Microwave Electrodynamics
of Superconductors
Why are Superconductors so Useful at Microwave Frequencies?
The Two-Fluid Model
London Equations
BCS Electrodynamics
Nonlinear Surface Impedance

16

Why are Superconductors so Useful


at Microwave Frequencies?

Low Losses:

Filters have low insertion loss Better S/N, can be made small
NMR/MRI SC RF pickup coils x10 improvement in speed
High Q Steep skirts, good out-of-band rejection

Low Dispersion:
SC transmission lines can carry short pulses with little distortion
RSFQ logic pulses 1 ps long, ~2 mV in amplitude: V (t ) dt = 0 = 2.07 mV ps

Superconducting Transmission Lines


microstrip

Kinetic
Inductance
Lkin

Geometrical
Inductance
Lgeo

(thickness t)

E
propagating TEM wave
attenuation ~ 0

17

ground
plane

Lkin ~

2
t

v phase =

1
LC

L = Lkin + Lgeo is frequency independent

Electrodynamics of Superconductors
In the Meissner State
Normal Fluid channel

Quasiparticles
(Normal Fluid)

Energy
Gap

Ls

Cooper Pairs
(Super Fluid)

Superfluid channel

Current-carrying superconductor

J
J=E
= n i 2
18

Js

J = Js + Jn

Jn
n = nne2/m
2 = nse2/m

ns(T)

nn(T)

Tc

nn = number of QPs
ns = number of SC electrons
= QP momentum relaxation time
m = carrier mass
= frequency

Surface Impedance
Z s = Rs + iX s =

J (z ) dz

-z

H
E
J

x
conductor

Surface Resistance Rs: Measure of Ohmic power dissipation


PDissipated

1
2
1 2
1
Rss H dA ~ I Rs
= Re J E dV = R
2 Volume
2
2 Surface

Surface Reactance Xs: Measure of stored energy per period


WStored

{ }

2

1

= H + Im J E
2 Volume

Xs = Ls =
19

2
1 2
dV = 1 X
~
H
dA
LI
X
ss

2
2
Surface

Two-Fluid Surface Impedance


Normal Fluid channel

1
Rs = 2 0 3 n
2
Z s = Rs + iX s
0
10
X s = 0

-2

10

Rn ~ 1/2
Cu(77K)

Superfluid channel

Surface resistance R ()

Ls

-1

10

2:

Because Rs ~
The advantage of HTS over Cu
diminishes with increasing frequency

-3

10

-4

poly

10

YBCO

T=0.85T
-5

10

epitaxial
-6

10

Rs ~ 2
Nb Sn

-7

Rs crossover at f ~ 100 GHz at 77 K

10

T=0.5T
-8

10

10

-1

10
10
10
frequency f (GHz)

M. Hein, Wuppertal
20

The London Equations

mv
dv
m
= eE
dt

Newtons 2nd Law for


a charge carrier
Superconductor:
1/ 0
1st London
Eq. and

B yield:
E =
t

= momentum relaxation time


Js = ns e vs

dJ s ns e 2
1
E=
E
=
2
dt
m
0 L
ns e 2
d
B = 0
J s +
dt
m

1st London Equation

London
surmise

ns e 2
Js +
B=0
m

2nd London Equation

These equations yield the Meissner screening


vacuum

H (z )

superconductor
z/
H = H 0e

1
H= 2 H
2

L is frequency independent ( < 2)


21

m
0 ns e 2

L ~ 20 200 nm

The London Equations continued


Normal metal
E is the source of Jn
Lenzs Law

Jn = nE

1
d

+
J

n
dt
0 L 2

Superconductor

dJ s
1
=
E E=0: Js goes on forever
2
dt
0 L

2
B = 0 0 L J s = B B is the source of Js,
spontaneous flux

exclusion

1st London Equation E is required to maintain an ac current in a SC


Cooper pair has finite inertia QPs are accelerated and dissipation occurs

22

BCS Microwave Electrodynamics


Low Microwave Dissipation

Full energy gap -> Rs can be made arbitrarily small

Rs,residual ~ 10-9 at 1.5 GHz in Nb

ky

Filled
Fermi
Sea

node

HTS materials have nodes in


d
the energy gap. This leads
to power-law behavior of
(T) and Rs(T) and residual losses

90 40
10

kx

ky

Filled
Fermi
Sea

kx

(T ) = (0) + a T
Rs = Rs ,residual + b T
Rs,residual ~ 10-5 at 10 GHz in YBa2Cu3O7-
23

20

15 T (K) 10
YBa Cu O

-1

7-x

@ f=87 GHz
s

Rs = RBCS (T ) + Rs ,residual

for T < Tc/3 in a


fully-gapped SC

Surface resistance R ()

Rs e

( 0 ) / k BT

10

-2

sputtered
LaAlO
3

10

-3

coevaporated
MgO
10

-4

Nb Sn on
3

10

log{R (T)} /kT T /T


s

sapphire

-5

2
4
6
8
10
Inverse reduced temperature T /T
c

M. Hein, Wuppertal

Nonlinear Surface Impedance of Superconductors


The surface resistance and reactance values depend on the rf current
level flowing in the superconductor

10

YBaCuO 1, PLD, 2"


YBaCuO 2, sputt, 1"

100

surface resistance, R (m)

surface resistance, R (m)

1000

YBa2Cu3O7- Thin Film


Made by Pulsed Laser Deposition
and sputtering
19 GHz

1
0.1
0

2
77K
1.5
71K
1

30K
0.5
0

20
40
60
80
temperature, T (K)

100

61K

4K
0

15

20

25

30

field amplitude, B (mT)

Similar results for Xs(Bs)


24

10

Data from M. Hein, Wuppertal

Why are Superconductors so Nonlinear?


Granularity

Josephson
weak links

Small ~ grain boundary thickness

JJs have
a strongly
nonlinear
impedance

Superconducting
grains

Edge-Current Buildup

McDonald + Clem
PRB 56, 14 723 (1997)

120

Heating

LTLSM Response, a.u.

+ Vortex Entry and Flow

Jrf2 (a.u.)

100
80

Scanning Laser Microscope image


YBCO strip at T = 79 K
f = 5.285 GHz, Laser Spot Size = 1 m

60
40
20

See poster 1EG08

0
0

100

200

300

400

Distance,

500

600

700

Microstrip (Longitudinal view)

Intrinsic Nonlinear Meissner Effect


rf currents cause de-pairing convert superfluid into normal fluid
2
2
J
(0, T )

JNL(T) calculated by theory (Dahm+Scalapino)

= 1
(
)
(
)
J
T
J
T
,

NL

Nonlinearities are generally strongest near Tc and weaken at lower temperatures


25

How to Model Superconducting Nonlinearity?


(1) Taylor series expansion of nonlinear I-V curve (Z. Y. Shen)
= 0 if I(-V) = -I(V)
1 d 2I
1 d 3I
dI
2
3
4

(
)

I (V ) = I (0 ) +
V
V
O
V
V +
2
3

2! dV V =0
3! dV V =0
dV V =0
1/R linear term

3rd order term dominates

V = V0 sin(t) input yields ~ V03 sin(3t) + output

(2) Nonlinear transmission line model


I
R

I
V
= C
z
t

L
C

L and R are nonlinear:


I

L = L0 + L
I
NL
26

V
I
= L RI
z
t

R = R0 + R
I
NL

(Dahm and Scalapino)

3rd harmonics and 3rd order IMD result

I
L I
R
2I
2I
C
C
I
LC
RC
+
+
=
+
2
2
z
t
t
t t
t
additional terms

Experimental Microwave Superconductivity


Cavity Perturbation
Measurements of Nonlinearity
Topics of Current Interest
Microwave Microscopy

27

Cavity Perturbation
Objective: determine Rs, Xs (or 1, 2) from f0 and Q measurements
of a resonant cavity containing the sample of interest
Input

~ microwave
wavelength

Microwave
Resonator

Output

transmission

Sample at
Temperature T

frequency

f0

f0

f = f0 f0 (Stored Energy)
(1/2Q) (Dissipated Energy)

EStored
f
= 0
EDissipated f
Rs =

28

T2

Quality Factor
Q=

T1

Cavity perturbation means f << f0

X s =

is the sample/cavity geometry factor

Measurement of Nonlinearities
Intermodulation is a practical problem

signal

20
15
10

linear
1

Bandwidth of
passband
15

3rd-order
intercept
Point (TOI)

Pout (dB)

Nonlinear (i. e., signal strength


dependent) microwave response
induces undesirable signals within
the passband by intermodulation.

10

21 - 2, 22 - 1

10

15

Pin (dB)

Pin

10

frequency
2f1-f2 f1

f2 2f2-f1

Intermodulation
29

2f1
harmonic generation

15

SC Pout
Device

20
25

M. Hein, Wuppertal

Topics of Current Interest


In Microwave Superconductivity Research
Identifying and eliminating the microscopic sources of extrinsic nonlinearity
Increase device yield
Allows further miniaturization of devices
Will permit more transmit applications
Identify the additional Drude term now seen in (,T)
under-doped cuprates show 2>0 above Tc
pseudo-gap electrodynamics

Nonlinear and Tunable Dielectrics


MgO substrates have a nonlinear dielectric loss at low temperatures
Ferroelectric and incipient ferroelectric materials as tunable
microwave dielectrics/capacitors

30

Microwave Microscopy of Superconductors

200m
loop probe

500
YBCO film
J

SrTiO3

P2f & P3f (dBm) at 60K

Use near-field optics techniques to obtain super-resolution images of:


1) Materials Properties: Nonlinear response
2) RF fields in operating devices -55

30 misorientation
Bi-crystal grain boundary

Image Jrf2(x,y) in an operating


superconducting microwave device
Image JIMD
31

60 K
P3f

P2f

See poster
2MC10
Noise Level of P2f

Noise Level of P3f

reflectance

Position (mm)

20 m

See poster
1EG08

Jrf2

Scanning Laser Microscopy

-60
-65
-70
-75
-80
-85
-90
-95
-100
-105

20 m

References and Further Reading


Z. Y. Shen, High-Temperature Superconducting Microwave Circuits,
Artech House, Boston, 1994.
M. J. Lancaster, Passive Microwave Device Applications,
Cambridge University Press, Cambridge, 1997.
M. A. Hein, HTS Thin Films at Microwave Frequencies,
Springer Tracts of Modern Physics 155, Springer, Berlin, 1999.
Microwave Superconductivity,
NATO- ASI series, ed. by H. Weinstock and M. Nisenoff, Kluwer, 2001.
T. VanDuzer and C. W. Turner, Principles of Superconductive Devices and Circuits,
Elsevier, 1981.
T. P. Orlando and K. A. Delin, Fundamentals of Applied Superconductivity,
Addison-Wesley, 1991.
R. E. Matick, Transmission Lines for Digital and Communication Networks,
IEEE Press, 1995; Chapter 6.

32