Superconductors for Superconducting Magnets

1
Superconductors for Superconducting Magnets

• David Larbalestier
• Applied Superconductivity Center
• National High Magnetic Field Laboratory
• Florida State University
• MagLab Summer School, Tallahassee FL
• June 22-26, 2009

2
Thousands of superconductors but only 6 useful
conductors.

• Nb47wtTi (Nb-Ti)
• Nb3Sn
• Bi-2223
• YBCO
• Bi-2212
• MgB2

3
Superconductor Phase Diagram

• Critical Parameters
• Critical Temperature, Tc
• Critical Magnetic Field, Hc2
• Critical Current Density, Jc
• For HTS Hc2 is not the phase boundary
• Thermal fluctuations make the dissipation line
  lie a long way below Hc2 at an irreversibility
  field H

Tc, Hc2, H relatively fixed for given
material, Jc highly dependent on specific sample!
4
A historical perspective.Onnes in Chicago 1913
(IIR)
5
(No Transcript)
6
(No Transcript)
7
Onnes in 1913..!

• The conception of a 10 T magnet
• The impossibility of doing this with Cu cooled by
  liquid air (as expensive as a warship)
• The possibility of doing it with superconductor
  (1000 A/mm2 with a Hg wire, 460 A/mm2 with a Pb
  wire
• A few problems
• Resistance developed at 0.8 A, not 30 A
• Silk insulation allowed easy He permeation
• Sn coated on a strong constantan wire
• 48 years had to go by before the path to high
  field superconducting magnets was cleared

8
The decisive experiment -1961
Phys Rev Letts 6, 89 (1961), submitted January 9,
published February 1, 1961!
9
Superconducting MRI Magnets made the
Superconducting Industry

• Closed (1-3 Tesla) and open (0.3T) MRI magnets
  both use Nb-Ti with a transition temperature (Tc)
  of only 9K, -450F.
• Nb-Ti might be replaced by MgB2 (ASG-Columbus)

10
LHC at CERN LTS enabled by HTS
Mont Blanc
1500 tonnes of LTS SC cables
1232 SC Dipoles
3286 HTS Leads
Lake Geneva
Switzerland

• Nb-Ti at 1.9 K at CERN France/Switzerland
• 5000 Superconducting Magnets in 27 km tunnel
• Beam-steering dipole magnets reach 8.36 T (1.9 K)

Large Hadron Collider 15000 MJ of magnetic energy
27 km Tunnel
France
11
Onnes view Superconductivity Zero
Resistivity

• Non-Superconducting Metals
• r ro aT for T gt 0 K
• r ro Near T 0 K
• Recall that r(T) deviates from linearity near T
  0 K
• Superconducting Metals
• r ro aT for T gt Tc
• r 0 for T lt Tc
• Superconductors are more resistive in the normal
  state than good conductors such as Cu

12
The Meissner Perfect Diamagnetic State (1933)

• c m -1
• Means
• B mo(H M)
• B mo(H cm H)
• B 0

Flux is excluded from the bulk by supercurrents
flowing at the surface to a penetration depth (l)
200-500 nm
13
Shubnikov (1937) Type I and Type II

• Type I
• Material Goes Normal Everywhere at Hc

• Type II
• Material Goes Normal Locally at Hc1, Everywhere
  at Hc2

Complete flux exclusion up to Hc1, then partial
flux penetration as vortices Current can now
flow in bulk, not just surface
Complete flux exclusion up to Hc, then
destruction of superconductivity by the field
14
Type II field penetrates as quantized vortices

• Two characteristic lengths
• coherence length x, the pairing length of the
  superconducting pair
• penetration depth l, the length over which the
  screening currents for the vortex flow
• Vortices have defined properties in
  superconductors
• normal core dia, 2x
• each vortex contains a flux quantum f0 currents
  flow at Jd over dia of 2l
• vortex separation a0 1.08(f0/B)0.5

Hc2 f/2px2 f0 h/2e 2.07 x 10-15 Wb B/Bc2
(b) 0.2
15
Nb-Ti
Optimized Nb-Ti strands have 25 a-Ti More
precipitates than fluxons (full summation) Very
strong flux pinning 5-10 Jd Laminar,
proximity-coupled N pins Highly heterogeneous
nano-structure
TEM by Peter Lee
16
FP tuned by nanostructure
For conventionally processed Nb-Ti, Fp increases
with drawing strain after the last heat treatment
ef. The increase occurs at all fields as the
precipitate size and spacing are reduced to less
than a coherence length (x) in thickness. The
refinement of the microstructure with increasing
strain for the same strand is shown schematically
in transverse cross-sections with the a-Ti
precipitates in black.
Meingast, Lee and Larbalestier JAP 1989
17
Nb3Sn

• Twice the Tc and Hc2 of Nb-Ti
• Brittle
• Must be reacted at final wire size to make Nb3Sn
  from a Sn and Nb mixture to avoid damage to
  conductor

18
Nb3Sn compositional Tc and Hc2 effects
Devantay et al. J. Mat. Sci., 16, 2145 (1981)
Thesis work of Jewell, Adv Cryo Eng 2004
Charlesworth et al. J. Mat. Sci., 5, 580 (1970)
Filaments of Nb3Sn must be made by diffusion
under non-equilibrium conditions Sn, Tc, Hc2 and
Jc gradients!
19
Grain boundaries pin the vortices
a) 57 nm
b) 70 nm
c) 77 nm
A15
A15
d) 89 nm
e) 113 nm
20
Low Temperature Superconductors
Type I 1911-1940
Bc
Bc2
Type II 1961-1980
21
How high can Nb go? Only 23T!
Plan A
Plan B
The NHMFL 900-MHz Ultra-Wide Bore (2004)
State of the art Nb3Sn 900 MHz NMR magnet,
operating persistent at 21T, 950 MHz now
achieved, 1 GHz (23T) is the likely limit
Higher fields require HTS or MgB2
Bi-2212 data Chen, Halperin et al. Nat Phys 2007
22
Summary - I

• The vision of 10 T magnets in Chicago in 1913, 2
  years after discovery
• Bad places in the wire were NOT the cause of
  Onnes wire leaving the superconducting state
• Glimpses of the emergence of negative surface
  energy superconductivity (Shubnikov 1936, Landau
  1940s, Abrikosov 1957)
• Experimental validation by Kunzler et al. in 1961
  made finally superconductivity technological

23
1986, the 75th Anniversary.

• POSSIBLE HIGH-TC SUPERCONDUCTIVITY IN THE
  BA-LA-CU-O SYSTEM BEDNORZ JG, MULLER KA Z FUR
  PHYSIK B-CONDENSED MATTER   64, 189-193   1986 ,
  Times Cited 7,656

• Superconductivity induced by doping carriers into
  an insulating anti-ferromagnetic state
• Non-Fermi liquid behavior, but strong
  correlations that still prevent any generally
  accepted model for superconductivity in the
  cuprates

24
Higher Tc greater complexity
Nb-Ti
Nb3Sn
18-23 K
9 K
39 K
92-95 K
110 K
MgB2
YBCO
Bi-2223
25
Single crystal architecture imposed by the strong
depression of superconductivity at GBs

• Jb(?) J0exp(-?/?0), ?0 ? 4-5o.
• (D. Dimos, P. Chaudhari, J. Mannhart, PRB 41,
  4038 (1990) R. Gross (1994)
• N.F. Heinig et al., APL, 69, 577 (1996))

HRTEM
c
c
?
J
a
a
Conclusion strong texture is needed!
26
Electronic State of Low-Angle GBs
AJ vortices
Insulating dislocation cores
A vortices
Current channels
Hole-depleted layer

• GB dislocations enable the misorientation
• but produce strains which destroy
  superconductivity
• GB dislocations cause charge imbalance, thus
• suppress the superconducting gap in the current
  channels

HRTEM image of 8001 tilt GB in Bi2Sr2CaCu2Ox
A Gurevich and E.A. Pashitskii, PRB 57, 13875
(1998) J. Mannhart and H. Hilgenkamp, APL 73,
265 (1998)
GBs could be improved by increasing the hole
density on the GB.
27
8-10º GBs force current to flow through lower
angle GBs
0.42 mV
-0.04 mV
J
OIM, E field map
J
GB dissipation visualized by LT laser imaging
Abraimov, Li et al. FSU
28
HTS greatly extends the capability at 4K
Courtesy Peter Lee www.asc.magnet.fsu.edu
29
Nb47Ti (OST)
Internal Sn Nb3Sn (OST)
Bi-2212 (OST)
Bi-2223 (AMSC)
YBCO coated conductors next
Preferred conductor features Multifilament Round
or lightly aspected shape with no Jc
anisotropy Capability to wind in unreacted form
while conductor fragility is minimized
MgB2 (Hypertech)
30
And coated conductors of YBCO which approximate
single crystals by the mile.

• The IBAD approach ion-beam-assisted deposition
  of the textured template

Copper Stabilizer 50-75 mm
Ag (lt1 mm)
CeO2 (75 nm)
Pilot production of 100-500 m lengths
YSZ (75 nm)
Y2O3 (75 nm)
Metallurgical Texture introduced here (RABiTS)
Ni-W alloy (50-75 mm)
31
Summary II - HTS Conductor Issues

• Hc2(T) much larger than for Nb3Sn
• 100-120T versus 30T (30T for YBCO at 55K)
• But, thermal fluctuation effects depress the
  irreversibility field at which Jc 0 well below
  Hc2, except at low T
• Grain boundaries easily acquire depressed
  properties and degrade Jc even for small
  misorientations of 3-5º
• Conductors must be fabricated with extreme texture

HTS are extremely interesting for point 1, rather
bad from point 2 can we get some project pull
to help HTS?
32
National Magnet Lab User Facility

• Provides the worlds highest DC magnetic fields
• 45T in hybrid, 32 mm warm bore
• Purely resistive magnets 35T in 32 mm warm bore,
  31 T in 50 mm bore and 19T in 195 mm warm bore
• 20 MW resistive magnets are costs on average
  1000/hr in electricity (full power cost is
  2400/hr)
• Long-time, full-field experiments are very
  expensive
• Quantum oscillation, quantum Hall effect, low
  noise, large signal averaging experiments could
  run 7 days a week

33
2007 12 pancake SuperPower CC coil in NHMFL
20cm bore, 20 MW, 19T background field yielded
26.8 T
2G HF Insert Coil Showing Terminals, Overbanding
and Partial Support Structure. Flange OD is 127
mm.
Hazelton et al. July 2007
34
Almost 27 T in stable fashion, even though some
damage and dissipative pancakes
Peak hoop stress 215 MPa, well below tape limit
26.8 T _at_ 175 A
9.81 T _at_ 221 A
Drew Hazelton et al. MT20
35
BSCCO Technology

• How can 2212 and 2223 be so different as
  conductors when they are so similar as structures?

RW - 2212
Charge reservoir layer
Charge reservoir layer
Charge reservoir layer
Charge reservoir layer
Tape 2223
36
The doping state matters

• High doping means higher current density and less
  degradation of superconductivity at the GB
• Bi-2223 and YBCO can hardly be overdoped 2212
  can be strongly overdoped

37
Round wire Bi2212
RW
TapeHc
Tape H ab

• Hirr of the round wires lies between H//c and
  H//ab of the tapes
• Much less anisotropy in RW More versatile for
  magnets
• Ag-BSCCO interface is irregular and untextured
• Orientation of Bi2212 grains is randomly
  distributed, although ab-planes are almost
  parallel to the longitudinal direction

Jianyi Jiang ASC-NHMFL
38
Bi-2212 round wire coil (Trociewitz, Weijers, DCL
on Oxford 2212) conductor reached 32 T in 31 T
before HT
after HT

• coil specs
• 15 mm ID, 38 mm OD
• 100 mm high
• 10 layers, 750 turns, 66 m
• DB 2.2 T at 31 T
• L 1 mH

• slight discoloration of braid at enclosed
  feed-through
• regular HT, no visible leaks

15mm spiral results

• First HTS wire-wound coil to go beyond 30 T

39
32 T Superconducting User Magnet Designs
Design 1 2 3 Cold bore YBCO
(mm) 40 40 40 Cold bore outer
(mm) 250 200 200 YBCO conductor Length
(m) 10145 7347 4620 Cost at 50/m
(k) 507 367 231
Denis Markiewicz design
40
Summary - III

• HTS cuprates have taken 15-20 years to become
  attractive for magnets
• Principal problem is the need for extreme texture
  (single crystals by the mile)
• Suddenly there is a new discovery of
  superconductivity in layered Fe-As compounds Tc
  so far up to 55 K in SmFeAsO, where O is the
  doping site, either e.g. O0.85 or F0.15
• Several classes of compound with Fe layers
• What next?
• Is a RT superconductor possible and if so will
  it carry any current?