Unit 3 Basics of Superconductivity: Soren Prestemon Paolo Ferracin and Ezio Todesco

Download as pdf or txt
Download as pdf or txt
You are on page 1of 41

Unit 3


Basics of superconductivity

Soren Prestemon
Lawrence Berkeley National Laboratory (LBNL)
Paolo Ferracin and Ezio Todesco
European Organization for Nuclear Research (CERN)

USPAS, June 2015, Rutgers University Superconducting accelerator magnets


Scope of the course

Basics of superconductivity
1. History
2. General principles
3. Diamagnetism
4. Type I and II superconductors
5. Flux pinning and flux creep
6. Critical surfaces for superconducting materials

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 2


References

Wilson, “Superconducting Magnets”


Mess, Schmueser, Wolff, “Superconducting Accelerator
Magnets”
Arno Godeke, thesis: “Performance Boundaries in Nb3Sn
Superconductors”
Alex Guerivich, Lectures on Superconductivity
Roberto Casalbuoni: Lecture Notes on Superconductivity:
Condensed Matter and QCD

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 3


History of Cryogenics

Cryogenics is the science of producing temperatures below


~200K

Faraday (~1820’s) demonstrates ability to liquify most known gases


by first cooling with a bath of ether and dry ice, followed by
pressurization
he was unable to liquify oxygen, hydrogen, nitrogen, carbon monoxide,
methane, and nitric oxide
The noble gases, helium, argon, neon, krypton and xenon had not yet
been discovered (many of these are critical cryogenic fluids today)
In 1848 Lord Kelvin determined the existence of absolute zero:
0K=-273C (=-459F)
In 1877 Louis Caillettet (France) and Raoul-Pierre Pictet
(Switzerland) succeed in liquifying air
In 1883 Von Wroblewski (Cracow) succeeds in liquifying Oxygen
In 1898 James Dewar succeeded in liquifying hydrogen (~20K!); he
then went on to freeze hydrogen (14K).
Helium remained elusive; it was first discovered in the spectrum of
the sun
1908: Kamerlingh Onnes succeeded in liquifying Helium

USPAS
USPAS,June
June 2009, Superconducting
2015, Rutgers Universityaccelerator magnets
Superconducting accelerator magnets Basic Cryogenics 12.
History

1911: Kamerlingh Onnes discovery of mercury


superconductivity: “Perfect conductors”
- A few years earlier he had succeeded in liquifying Helium,
a critical technological feat needed for the discovery
1933: Meissner and Ochsenfeld discover perfect Kamerlingh Onnes,
diamagnetic characteristic of superconductivity Nobel Prize 1913

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 5


History - theory


A theory of superconductivity took time to evolve:


1935: London brothers propose two equations for E
and H
results in concept of penetration depth λ

1950:Ginzburg and Landau propose a macroscopic


theory (GL) for superconductivity, based on Heinz and Fritz London
Landau’s theory of second-order phase transitions
Results in concept of coherence length ξ

Ginzburg and Landau (circa 1947)


Nobel Prize 2003: Ginzburg,
Abrikosov, Leggett

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 6


History - theory

1957: Bardeen, Cooper, and Schrieffer publish microscopic theory (BCS) of


Cooper-pair formation that continues to be held as the standard for low-
temperature superconductors
1957: Abrikosov considered GL theory for case κ=λ/ε>>1
Introduced concept of Type II superconductor
Predicted flux penetrates in fixed quanta, in the form of a vortex array

Bardeen, Cooper and Schrieffer


Nobel Prize 1972

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 7


History - theory

1957: Abrikosov considered GL theory for case κ=λ/ε>>1


Introduced concept of Type II superconductor
Predicted flux penetrates in fixed quanta, in the form of a vortex array

Nobel Prize 2003: Ginzburg, Abrikosov, Leggett (the GLAG members)

Abrikosov with
Princess Madeleine

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 8


History – High temperature superconductors

1986: Bednorz and Muller discover superconductivity at


high temperatures in layered materials comprising copper
oxide planes

Discovery of
superconductors

39K Jan 2001 MgB2

George Bednorz and Alexander Muller


Nobel prize for Physics (1987)

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 9


General Principals

Superconductivity refers to a material


state in which current can flow with no
resistance
Not just “little” resistance - truly ZERO
resistance
Resistance in a conductor stems from scattering
of electrons off of thermally activated ions
Resistance therefore goes down as temperature
decreases
The decrease in resistance in normal metals
reaches a minimum based on irregularities and
impurities in the lattice, hence concept of RRR
(Residual resistivity ratio)
RRR is a rough measure of cold-work and
impurities in a metal

RRR=ρ(273K)/ ρ(4K))
USPAS, June 2015, Rutgers University Superconducting accelerator magnets 10
Aside: Maxwell’s equations

~ ⇢
r·E = Gauss’ law

~
r·B = 0
@B ~
~
r⇥E = Faraday’s law
@t
@ ~ Ampere’s law
E
~
r⇥B = µ0 J~ + µ0 ✏0
@t (corrected by Maxwell)
Permeability of free space
Permittivity of free space

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 11


Some reminders of useful formulas

∇ ⋅ (∇ × F )= 0 ∀F ∇ × (∇ × F )= ∇(∇ ⋅ F )− ∇ 2 F ∀F

∇ × (∇u ) = 0 ∀u or ∇× F = 0 ⇔ F = ∇u
(F is conservative if curl F is zero)

Volume Integral

∫ F ⋅ n dS = ∫ ∇ ⋅ F dV Divergence Theorem
S V

Surface Integral (Flux)

∫ F ⋅ dl = ∫ (∇ × F )⋅ n dS
l S
Curl Theorem (Stoke’s Theorem)

Line Integral (Circulation)


USPAS, June 2015, Rutgers University Superconducting accelerator magnets 12
Some direct results from Maxwell

Electric and magnetic fields are fundamentally linked


dB/dt induces voltage (Faraday)
Moving charge generates B (Ampere)
Amperes law applied to DC fields and flowing currents:

Gauss’s law: no magnetic monopoles


Magnetic field lines cannot emanate from
a point; they “curl” around current

Equations admit wave solutions


Take the curl of Faraday’s and Ampere’s laws; E and B admit waves with
velocity 1
v= = c = speed of light
µ 0ε
USPAS, June 2015, Rutgers University Superconducting accelerator magnets 13
Magnetization

From a macroscopic perspective, critical insight can be gleaned from


magnetization measurements
Magnetization is the magnetic (dipole) moment generated in a material by
an applied field

∇ × B = µ0 J Amperes law

J = J free + J bound Arbitrary but useful distinction


J bound = ∇ × M
Results in a practical
1 definition: we know and
⇒H = B−M
µ0 control free currents We do notNote:
need M; every
calculation could be
⇒ ∇ × H = J free ⇒ !∫ H idl = I enclosed free current performed using B and
H

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 14


magnetization in superconductors

Example: iron is ferromagnetic – it has a strong


paramagnetic moment (i.e. the magnetization is parallel and
additive to the applied field)
Most materials are either diamagnetic or paramagnetic, but the
moments are extremely small compared to ferromagnetism
In diamagnetic and paramagnetic materials, the magnetization is a
function of the applied field, i.e. remove the field, and the
magnetization disappears.
In ferromagnetic materials, some of the magnetization remains
“frozen in” => hysteretic behavior

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 15


Basics of superconductivity

In a superconductor, when the temperature descends below


the critical temperature, electrons find it energetically
preferable to form Cooper pairs
The Cooper pairs interact with the positive ions of the lattice Alex Guerivich,
lecture on
Lattice vibrations are often termed “phonons”; hence the
superconductivity
coupling between the electron-pair and the lattice is referred to as
electron-phonon interaction
The balance between electron-phonon interaction forces and
Coulomb (electrostatic) forces determines if a given material is
superconducting
Electron-phonon interaction can occur over BCS breakthrough:
long distances; Cooper pairs can be Fermi surface is unstable to
separated by many lattice spacings bound states of electron-pairs

~meV~10-22J!!

% 1 & k =Boltzmann constant =1.38x10-23


Δ 0 ≅ 2!ωD exp * − + B
*, λep +- ω =Debye frequency
D
e γ
Tc ≅ Δ0 λep =electron-phonon coupling
π kb
USPAS, June 2015, Rutgers University Superconducting accelerator magnets
γ=euler constant=0.577 16
Diamagnetic behavior of superconductors

What differentiates a “perfect” conductor from a


diamagnetic material?

Cool Apply Remove B


B

A perfect conductor apposes


any change to the existing
Apply B Cool Remove B magnetic state

Apply B Cool Remove B Superconductors exhibit diamagnetic


behavior: flux is always expulsed -
Meissner effect

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 17


The London equations

Derive starting from the classical Drude model, but adapt to account for
the Meissner effect:
o The Drude model of solid state physics applies classical kinetics to electron
motion
✓ Assumes static positively charged nucleus, electron gas of density n.
✓ Electron motion damped by collisions
“Frictional drag” on “normal” conduction electrons
d~v ~
m = eE ~v
dt
J~s = ens~v
✓ ◆
~ = 1 B
2
@ m ~s + B
~ = 0 =) r B
2 ~ = µ0 n s e ~
) 2
r ⇥ J B 2
@t ns e m L

The penetration depth λL is the characteristic depth of the supercurrents


on the surface of the material.

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 18


Concept of coherence length

The density of superconducting electrons ns decreases to zero near a


superconducting /normal interface, with a characteristic length ξ
(coherence length, first introduced by Pippard in 1953). The two length
scales ξ and λL define much of the superconductors behavior.
The coherence length is proportional to the mean free path of conduction
electrons; e.g. for pure metals it is quite large, but for alloys (and
ceramics…) it is often very small. Their ratio, the GL parameter, determines
flux penetration:
κ = λL / ξ
B g
ns
t in
uc
From “GLAG” theory, if: m
al
rc
on
d
or p e
N su
κ < 1/ 2 Type I superconductor
κ > 1/ 2 Type II superconductor
Note: in reality ξ and λL are functions λL
of temperature ξ
USPAS, June 2015, Rutgers University Superconducting accelerator magnets 19
Thermodynamic critical field

The Gibbs free energy of the superconducting state


is lower than the normal state. As the applied field
B increases, the Gibbs free energy increases by
B2/2µ0.
The thermodynamic critical field at T=0
corresponds to the balancing of the
superconducting and normal Gibbs energies:
Hc
Gn = Gs +
2

The BCS theory states that Hc(0) can be calculated


from the electronic specific heat (Sommerfeld
coefficient): γ 1/ 2T
H c (0) = 7.65 ×10−4 c

µ0

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 20


Type I and II superconductors

Type I superconductors are characterized by the Meissner effect, i.e. flux


is fully expulsed through the existence of supercurrents over a distance
λL.
Type II superconductors find it energetically favorable to allow flux to
enter via normal zones of fixed flux quanta: “fluxoids” or vortices.
The fluxoids or flux lines are vortices of normal material of size ~πξ2
“surrounded” by supercurrents shielding the superconducting material.

First photograph of vortex lattice,


U. Essmann and H. Trauble
Max-Planck Institute, Stuttgart
Physics Letters 24A, 526 (1967)

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 21


Fluxoids

Fluxoids, or flux lines, are continuous thin tubes characterized by a


normal core and shielding supercurrents.
The flux contained in a fluxoid is quantized:
φ0 = h /(2e)
h = Planck's constant=6.62607 ×10−34 Js
e = electron charge=1.6022 ×10−19 C

The fluxoids in an idealized material are uniformly distributed in a


triangular lattice so as to minimize the energy state
Fluxoids in the presence of current flow (e.g. transport current) are subjected
to Lorentz force:

⇒ Concept of flux-flow and associated heating


Solution for real conductors: provide mechanism to pin the fluxoids
See flux flow movies…
USPAS, June 2015, Rutgers University Superconducting accelerator magnets 22
Critical field definitions

T=0
Hc1: critical field defining the transition from the Meissner state

φ0 % 1&
H c1 ≈ Ln ( κ + ) ; κ >> 1
4 2πµ0 λ 2
* 2+
-M
Hc: Thermodynamic critical field
Hc=Hc1 for type I superconductors

φ0
Hc = 2
Hc1 Hc Hc2
2 2 µ0κπξ

Hc2: Critical field defining the transition to the normal state


φ0
Hc2 =
2πµ0ξ 2

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 23


Examples of Superconductors

Many elements are superconducting


at sufficiently low temperatures
None of the pure elements are useful
for applications involving transport
current, i.e. they do not allow flux
penetration
Superconductors for transport
applications are characterized by
alloy/composite materials with κ>>1

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 24


Aside – uses for type I superconductors

Although type I superconductors cannot


serve for large-scale transport current
applications, they can be used for a variety of
applications
Excellent electromagnetic shielding for
sensitive sensors (e.g. lead can shield a sensor
from external EM noise at liquid He
temperatures
Niobium can be deposited on a wafer using
lithography techniques to develop ultra-
sensitive sensors, e.g. transition-edge sensors
Using a bias voltage and Joule heating, the
superconducting material is held at its
transition temperature;
Mo/Au bilayer TES detector
absorption of a photon changes the circuit
Courtesy Benford and Moseley, NASA Goddard
resistance and hence the transport current,
which can then be detected with a SQUID
(superconducting quantum interference device)
See for example research by J. Clarke, UC Berkeley;

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 25


Flux Flow

The Lorentz force acting on a fluxoid will, in the absence of


pinning, result in motion of the fluxoid
Fluxoid motion generates a potential gradient (i.e. voltage) and
hence heating
This can be modeled using Faraday’s law of induction:

⇒ E=ρeffJ => “ideal” superconductors can support no transport current


beyond Hc1!
Real superconductors have defects that can prevent the flow of
fluxoids
The ability of real conductors to carry transport current depends on the
number, distribution, and strength of such pinning centers
USPAS, June 2015, Rutgers University Superconducting accelerator magnets 26
Flux pinning

Fluxoids can be pinned by a wide


variety of material defects
Inclusions
Under certain conditions, small inclusions
of appropriate materials can serve as
pinning site locations; this suggests
tailoring the material artificially through
manufacturing
Lattice dislocations / grain boundaries
These are known to be primary pinning
sites. Superconductor materials for wires
are severely work hardened so as to
maximize the number and distribution of
grain boundaries.
Precipitation of other material phases
In NbTi, mild heat treatment can lead to
the precipitation of an a-phase Ti-rich
alloy that provides excellent pinning
strength.

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 27


Pinning strength

The distribution and pinning of fluxoids


depends on the operating regime:
At low field (but>Hc1) the distribution is
governed mainly by interaction between flux-
lines, i.e. the fluxoids find it energetically
advantageous to distribute themselves “evenly”
over the volume (rather weak)
fp(h)
At intermediate fields, the pinning force is
provided by the pinning sites, capable of
hindering flux flow by withstanding the Lorentz
force acting on the fluxoids. Ideally, the pinning
sites are uniformly distributed in the material
(very strong)
At high field, the number of fluxoids significantly 1
h=H/Hc2
exceeds the number of pinning sites; the effective
pinning strength is a combination of defect
pinning strength and shear strength of the
fluxlines (rather weak)

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 28


High-Temperature superconductors

Much of HTS behavior can be understood in terms of the


BCS and GLAG theory parameters
The new features of HTS have to do with:
1) highly two-dimensional domains of superconductor, separated by
regions of “inert” material
▪ Macroscopic behavior is therefore highly anisotropic
▪ Different layers must communicate (electrically) via tunneling, or incur
Joule losses
2) a much larger range of parameter space in which multiple effects
compete
▪ The coherence lengths for HTS materials are far smaller than for LTS
materials
▪ Critical fields are ~10 times higher
=> Thermal excitations play a much larger role in HTS behavior

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 29


Modeling pinning

Precise first-principles physical descriptions of overall pinning strength (and


hence critical current) of real superconductors is difficult due to ambiguities
intrinsic in pinning
Nevertheless, models based on sound physics minimize free parameters needed
to fit measured data and provide reliable estimates for classes of materials
One of the most cited correlations is that of Kramer:
H νc 2
Fp = Fmax f (h) ∝ γ
f ( h) “Kramer plot”
κ Fp = JcB ∝ b1/2(1–b)2
2
f (h) = h1/ 2 (1 − h ) ; h = H / H c 2 FK = Jc1/2B1/4 ∝ (1–b)

From L. Cooley, USPAS


The fitting coefficients ν and γ depend
on the type of pinning. Furthermore, it
is experimentally verifies that

! # T $2 "
H c (T ) ! H c (0) %1 − ( ) &
%, * Tc + &-
USPAS, June 2015, Rutgers University Superconducting accelerator magnets 30
Scaling of critical current:

field dependence
The Kramer formulation provides excellent fits in the region 0.2<h<0.6 for Nb3Sn; it is
appropriate for regimes where the number of fluxoids exceeds the number of pinning sites
Outside this region, a variety of effects (e.g. inhomogeneity averaging) can alter the pinning
strength behavior, so the pinning strength is often fitted with the generalization
q
f p (h) ∝ h p (1 − h ) ; h = H / H c 2

It is preferable to stay with the Kramer formulation, yielding:

J c1/ 2 B1/ 4 !
1.1×105
µ 0 (H c 2 − H )
κ

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 31


Scaling of critical current: 

temperature dependence
The temperature dependence of Jc stems from the term
ν
$& µ0 H c 2 (T ) %'
κ γ (T )
Scalings are typically generated by considering the normalized
thermodynamic critical field and the the normalized GL parameter
(here t=T/Tc):

H c (T )
= 1− t2
H c (0)
#
$1 − 0.31t 2 (1 − 1.77 ln(t ) ) Summers
κ (T ) $$
=% 1 − 0.33t Summers (reduced)
κ (0) $ 1.52
$ 1 − t Godeke / De Gennes
2
$& 1− t
USPAS, June 2015, Rutgers University Superconducting accelerator magnets 32
Scaling of critical current, Nb3Sn

Empirical Strain dependence
The critical current of Nb3Sn is strain dependent, particularly at high
field
The strain dependence is typically modeled in terms of the normalized
critical temperature:
3
H c 2 (4.2, ε ) " Tc (ε ) #
!$ % = s (ε )
H c 2 m (0) T
& cm '

The term Tcm and Hc2m refer to the peaks of the strain-dependent
curves
A “simple” strain model proposed by Ekin yields
1.7
s (ε ) = 1 − a ε axial
# 900 ε axial < 0
a=$
%1250 ε axial > 0
USPAS, June 2015, Rutgers University Superconducting accelerator magnets 33
Strain dependence of Jc in Nb3Sn:

physics-based model
A physics-based model of strain dependence has been developed using the
frequency-dependent electron-phonon coupling interactions (Eliashberg;
Godeke , Markiewitz)
Phonon density of states
α 2 (ω ) F (ω )
λep (ε ) = 2 ∫ dω
ω
From the interaction parameter the strain dependence of Tc can be derived
Experimentally, the strain dependence of Hc2 behaves as
H c 2 (4.2, ε ) Tc (ε )

H c 2 m (4.2) Tcm
The theory predicts strain dependence of Jc for all LTS materials, but the
amplitude of the strain effects varies (e.g. very small for NbTi)
The resulting model describes quite well the asymmetry in the strain
dependence of Bc2, and the experimentally observed strong dependence on the
deviatoric strain

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 34


Strain dependence of Jc in Nb3Sn

The strain dependence is a strong function of the applied


field and of temperature

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 35


Critical surface:

Example fit for NbTi

NbTi parameterization
Temperature dependence of BC2 is provided by Lubell’s formulae:
& , T )1.7 #
BC 2 (T ) = BC 20 $1 − ** '' !
$% + TC 0 ( !"
where BC20 is the upper critical flux density at zero temperature
(~14.5 T)
Temperature and field dependence of Jc can be modeled, for example,
by Bottura’s formula
γ NbTi
J C (B, T ) CNbTi $ B %
α NbTi
$ B %
β NbTi
$ & T '1.7 %
= ( ) (1 − ) (1 − + , )
J C ,ref B - BC 2 (T ) . - BC 2 (T ) . T
(- / C 0 0 ).

where JC,Ref is critical current density at 4.2 K and 5 T (e.g. ~3000 A/


mm2) and CNbTi (~30 T), αNbTi (~0.6), βNbTi (~1.0), and γNbTi (~2.3) are
fitting parameters.

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 36


Scaling Jc for NbTi & Nb3Sn

(Courtesy Arno Godeke)

1.0

0.8

0.6
Fits for NbTi

fHtL
0.4

0.2

0.0
0.0 0.2 0.4 0.6 0.8 1.0
t=TêTc0

Godekeet al., Supercond.


Sci. Technol. 19 (2006)

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 37


Nb3Sn: Strain and temp. dependence

(Courtesy Arno Godeke)


"
(ε axial − ε shift ) + (ε 0,a ) #& − Ca2ε axial
2 2 2 2
Ca1 % (ε shift ) + (ε 0,a ) −
s (ε axial ) = ' ( + 1,
1 − Ca1ε 0,a
Ca2ε 0,a
ε shift =
2 2
(Ca1 ) − (Ca2 )

C1 2
J c (H , T , ε ) ≅
µ0 H
( )( )
s (ε ) 1 − t1.52 1 − t 2 h 0.5 (1 − h )

H c2* (T , ε ) ≅ H c2m
*
(
(0 )s (ε ) 1 − t1.52 )

, T - ,1- , 1 D µ0 H c2 (T ) -
ln . / = ψ . / − ψ .. + //
. Tc (0 ) / 0 2 1 2 2 φ k T
0 1 0 0 B 1
USPAS, June 2015, Rutgers University Superconducting accelerator magnets 38
Using magnetization data
We have seen that the Meissner state corresponds to perfect diamagnetic behavior
We have seen that beyond Hc1, flux begins to penetrate and can be pinned at defects =>
hysteretic behavior
⇒ Much can be understood by measuring the effective magnetization of superconducting material

The measured magnetization provides insight into flux pinning and flux motion,
key concepts governing the performance of superconducting materials.

ΔM
M · B / Fp (T, B)
Often used to evaluate Jc(B,T)!

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 39


Example material: Nb3Sn

Phase diagram, A15 lattice…

BSCCO2223

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 40



Final comments

Recent developments in Tc and Jc are quite impressive


Improvements in material processing has lead to
enhanced pinning
Enhanced Tc
Smaller superconducting filaments

Expect, and participate in, new and dramatic developments as fundamental


understanding of superconductivity evolves and improvements in nanoscale
fabrication processes are leveraged
A basic theory of superconductivity for HTS materials has yet to be formulated!

Some understanding of the fundamentals of superconductivity are critical to


appropriately select and apply these materials to accelerator magnets
Superconductors can be used to generate very high fields for state-of-the-art
facilities, but they are not forgiving materials – in accelerator applications they
operate on a precarious balance of large stored energy and minute stability
margin!

USPAS, June 2015, Rutgers University Superconducting accelerator magnets 41

You might also like