FNAL Conductor Development for the 11 T Dipole Program
E. Barzi
03/06/2012 1
Magnet Seminar, CERN
March 6, 2012
Outline
Brief Summary of the Nb3Sn Magnet R&D at FNAL
since 1998
Introduction to the 11 T Dipole Program
Cable and Conductor Requirements
Properties of Available Conductors in the US
Cable Development:
o Feasibility
o Uncored Cable Technology
o Cored Cable Technology
Strand and Cable Modeling
Summary
03/06/2012 2
3
Main Activities of the Nb3Sn Magnet R&D at FNAL
43.5 mm Nb3Sn dipoles for VLHC with operation fields
up to 10-11 T (1998-2007)
90 mm Nb3Sn quadrupoles for LHC IR’s with operation
gradients up to 200 T/m (2005-2011)
Nb3Sn magnet technology scale-up (2007-2011)
Rutherford cables development (since 2000)
Contributions to VLHC, LARP, MCTF-MAP, CDP,
VHFSMC studies
4
Dipole Magnets for Hadron Colliders
Development of a series of 43.5 mm Nb3Sn dipoles
with Bnom ~ 10 T based on a collar-free structure
First demonstration of quench performance and field
quality reproducibility for Nb3Sn accelerator magnets
Development and demonstration of effective passive
correction scheme based on iron strips
5
Technology Development
Development of Coil Test Structure (i.e. Mirror) for single
short and long dipole and quadrupole coils
o Lower cost, shorter turnaround time, advanced
instrumentation
Experimental studies:
o Nb3Sn strand (PIT, MJR, RRP) performance at 1.9-4.5K
o Cable with SS core to suppress eddy currents
o Cable insulation based on ceramic, E-glass and S2-glass tapes
o Coil structural materials (bronze vs. Ti poles) and processing
o Effect of coil pre-stress on its quench performance
6
Nb3Sn Magnet Technology Scale-up
4 m long coil test structure
4 m long Nb3Sn dipole (left) and quadrupole
LHC Collimation Upgrade
For CERN planned upgrade to the LHC collimation system:
o Additional collimators in DS regions around points 2, 3, and 7
o IR 1 and 5 as part of the HL-LHC
The collimator space ~3.5 m requires stronger and shorter dipoles
o These dipoles will be operated at 1.9 K, powered in series with the
main dipoles and deliver the same integrated strength at 11.85 kA
o MB: Bnom=8.35 T, Lmag=14.3 m, ∫BdL = 119.2 Tm @ Inom = 11.85 kA
o Lmag=14.3-3.5=10.8 m, Bnom=11 T => Nb3Sn technology
03/06/2012 7
General Design Approach
Coil aperture 60 mm
o To accommodate the beam sagitta and avoid the additional
complication of curved Nb3Sn coils
Coil length 5.5 m
o Present tooling length limitations at Fermilab
Modified 550 mm iron yoke from the LHC main dipole
o Compatibility with LHC main systems
11 m long cold mass combines two 5.5 m long cold masses
o For arrangement flexibility
03/06/2012 8
Magnet Development Plan
CERN and FNAL have started in October 2010 a joint development
program to demonstrate feasibility of 11 T twin-aperture, 5.5 m
long Nb3Sn dipoles by 2014.
03/06/2012 9
Possible Production Phase 2014-17
03/06/2012
10
Single-Aperture Demonstrator
Modified structure of FNAL Nb3Sn
dipole (HFDA)
14.85 mm wide and 1.3 mm thick
(high aspect ratio) 40 strand
Rutherford cable
0.7 mm Nb3Sn strand.
Two-layer 6-block coil design.
Stainless steel collar.
400 mm vertically split iron yoke.
Al clamps to control yoke gap.
12 mm thick stainless steel skin.
50-mm thick end plates.
03/06/2012 11
Maximum stress during assembly (collaring, skin welding, and
cool-down) ~130 MPa to keep coil under compression up to 12 T
bore field.
Mechanical structure optimized to maintain the coil stress below
165 MPa - safe for brittle Nb3Sn coils.
Coil Technology
03/06/2012
12
Twin-Aperture Design Studies
Modified MB structure with separate
collared coils – to simplify the
process and reduce the press force
during collaring
Electromagnetic design challenges
o Matching the MB transfer function
o Control the magnetic cross-talk
between apertures
o Minimization of the unwanted
multipoles in the current cycle
03/06/2012 13
Mechanical design challenges
o First twin-aperture Nb3Sn dipole
o Coil prestress and Lorentz force management inside
the LHC iron yoke
o Poles under compression and coil stress <165 MPa
Demonstrator Dipole Parameters
03/06/2012 14
Parameter Single-aperture Twin-aperture
Aperture [m] 60
Nominal current Inom
[A] 11850
Yoke outer diameter [mm] 400 550
Nominal bore field [T] 10.86 11.25
Short-sample bore field at 1.9 K [T] 13.6 13.9
Margin Bnom
/Bmax
at 1.9 K [%] 80 81
Maximum design field [T] 12.0 12.0
Inductance at Inom
[mH/m] 5.6 11.98
Stored energy at Inom
[kJ/m] 473 969
Fx per quadrant at I
nom [MN/m] 2.89 3.16
Fy per quadrant at I
nom [MN/m] -1.57 -1.59
Coil length [m] 1.9 1.9
Magnetic length [m] 1.79 1.68
Cable Specs v.1
03/06/2012 15
16 03/06/2012
Cable Requirements
17 03/06/2012
What Happens to Round Strands in a Cable
Examples of Internal Damage in Transversally
Deformed (Flat-rolled) Strands
Broken Only
Broken and Possibly Merged Broken and Merged
Possibly Broken
03/06/2012 18
19 03/06/2012
Good PIT (by SMI) vs. old RRP
0
0.2
0.4
0.6
0.8
1
1.2
0 0.2 0.4 0.6
Strand Relative Deformation
I c(1
2T
)/I c
0(1
2T
)
RRP1
PIT
0.4
0.6
0.8
1
1.2
1.4
1.6
0 0.1 0.2 0.3 0.4 0.5 0.6
Strand Relative Deformation
M
(12T
)/
M0(1
2T)
RRP1
PIT
Def.=(d0-t)/d0 t
“Study of Effects of
Deformation in Nb3Sn
Multifilamentary Strands”,
IEEE Trans. Appl. Sup., V.
17, No. 2, p. 2710 (2007).
B
20 03/06/2012
Solutions to Damage and Merging
Because of the stringent requirements on the deff
for magnet field quality and for low field stability, a
multi-year effort was devoted to looking for
solutions to improve the RRP deff performance
under deformation.
A first obvious solution to minimize the effect of
merging is to reduce the subelement size.
But it was also found that increasing subelement
spacing in the copper matrix reduced merging.
Experiments indicate that this is due in part to
mechanical reasons, but mostly to the longer Sn-
Nb diffusion paths in the Cu.
21
Summary of RRP Strand Development with OST
127 restack with spaced SE’s
127 restack 61 restack
61 restack with spaced SE’s
Increase Subelement Number
Inc
rease
Sub
element
Spa
cing
22
Cross-section
Strand diameter, mm 0.700±0.003 0.700±0.003
Cross-section design 108/127 150/169
Cu fraction, % 54 51
Effective sub-element diameter, μm 52-75 36-52
Ic(12T, 4.2K), A >470 >470
Jc(12T, 4.2K), A/mm2 >2650 >2500
RRR(after heat treatment) >60 >60
Twist pitch, mm 14±2 14±2
41 36
RRP® Strands for 11 T Dipole
Ds, mm
“Studies of Nb3Sn Strands based on the Restacked-Rod Process for High Field
Accelerator Magnets”, MT-22, accepted in IEEE Trans. Appl. Sup. (2012).
0
200
400
600
800
1000
1200
1400
1600
1800
2000
0.00 0.10 0.20 0.30 0.40 0.50 0.60
J C, A
/mm
2
Wire Deformation
150/169
108/127
23 03/06/2012
Transport Properties at 4.2 K
Jc (14T) RRR
1 mm original size
• Ic(14 T) degrades similarly under
increasing deformation as in the
108/127 design.
• The deformed 1 mm 150/169
strand shows RRR values that
are consistently larger than for
the 108/127.
For this first study the 108/127 and 150/169 wires were given two
different heat treatments, which were tailored to produce Jc’s as close as
possible for the two round wires.
Def.=(d0-t)/d0 t
24 03/06/2012
Behavior of Deformed Wires
0
1000
2000
3000
4000
5000
6000
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
J, A
/mm
2
Magnetic Field, T
108/127
0% def V-I20% def V-I30% def V-I40% def V-I50% def V-I20% def V-H30% def V-H
0
1000
2000
3000
4000
5000
6000
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
J, A
/mm
2
Magnetic Field, T
150/169
10% def V-I30% def V-I40% def V-I50% def V-I40% def V-H50% def V-H
The deformed 150/169 strand has a systematically better J
performance over most of the deformation range.
Def.=(d0-t)/d0 t
25 03/06/2012
Transport Properties at 1.9 K
0.7 mm Round
At 1.9 K the better stability of the 169 restack design
is apparent.
-300
-200
-100
0
100
200
300
0 0.5 1 1.5 2 2.5 3
M [
kA/m
]
B [T]
150/169
108/127
Magnetization at 4.2 K
0.7 mm Round
• For this study, samples from the two wires were given the same heat
treatment.
• Magnetization m0M(4.2 K, 12 T) per total strand volume was 37.541.73 mT
for the 108/127 RRP sample and 29.331.06 mT for the 150/169 RRP
sample.
• The max. Jc(4.2 K, 12 T) was 2890 A/mm2 for the 108/127 RRP sample and
2660 A/mm2 for the 150/169 RRP sample. The resulting effective filament
size deff in the round filament approximation was 54.73.4 mm for the
108/127 RRP sample and 42.92.1 mm for the 150/169 RRP sample.
27
Strand Summary
This study compared the effect of increasing deformation
on the 150/169 and the baseline 108/127 RRP design:
o The 150/169 design used in this study performed as well as
the 108/127 baseline design and is more stable, providing
some margin on the deff to account for the subelement
merging that occur in cables.
The effective filament size (in the round filament
approximation) is larger than the flat-to-flat dimension of
the hexagonal subelement, which at 0.7 mm of strand size is
41 mm for a 127 restack, 36 mm for a 169 restack and 32 mm for a
217 restack.
It takes a number of full scale billets to reach production
level. The 169 restack is ~1 year away from production.
The company will need to fabricate enough full scale
billets to develop production procedures for the 217
restack.
28
FNAL Cable Technology
“Nb3Sn Cable Development for the 11 T Dipole Demonstration Model”, CEC/ICMC
2011, accepted for publication in Advances in Cryogenic Engineering (2011). 29
• 4 rectangular Cu practice cables: • Keystoned cables 14.7 mm and 15.1 mm wide, with and
without SS core.
• => 15.1 mm required 41 strand for stability, therefore the
14.7 mm width was chosen.
• 2 rectangular Nb3Sn cable short samples at 14.7 mm
width, with and without SS core: • Average Ic (4.2 K, 12 T) degradation was 3.5% for uncored
cable and 2% for cored one.
• At 1.9 K, Ic(12T)=602 A and Is=951 A for uncored cable, and
Ic(12T)=592 A and Is=812 A for cored one.
• => Feasibility of a cored cable technology to suppress eddy
currents.
• A 250 m long Cu cable for the first practice coil.
• A 250 m long Nb3Sn keystoned cable for the second
practice coil made of RRP 114/127 wire
Cable Development – First Phase
Focused on feasibility and long length production
(using Cu and older RRP wires)
30
The second phase of cable development was based on the
important feedback obtained from winding the practice
coils, which revealed insufficient cable mechanical stability
for a production oriented winding process:
• The thickness of the rectangular (first pass) cable was
reduced from 1.32 mm to 1.30 mm.
• The width of the rectangular cable (first pass) was also
reduced and optimized at 14.5 mm to allow for any width
expansion due to spring back and intermediate
annealing.
• Then short samples of keystoned (second pass) cables
were made within a range of mid-thicknesses to study
sensitivity of electrical properties to compaction.
Cable Development – Second Phase
“Development and Fabrication of Nb3Sn Rutherford Cable for the 11 T DS Dipole
Demonstrator Model”, MT-22, accepted in IEEE Trans. Appl. Sup. (2012).
Focused on Uncored Cable Technology using 108/127 RRP
31
Test Results – Ic retention
Reducing keystoned cable thickness of the uncored
cable from 1.27 mm to 1.25 mm improves cable
mechanical stability and reduces the risk of cable
collapsing during fabrication with an Ic degradation still
within specs.
80%
85%
90%
95%
100%
105%
1.22 1.23 1.24 1.25 1.26 1.27 1.28
Ic_E
xtra
cted
/ Ic
_Ro
un
d @
12
T
Average Cable Thickness (mm)
108/127 Annealed
Specs for Ic Degradation
32 03/06/2012
Cable Development – Third Phase
40-strand (RRP-150/169) cable with stainless steel core 11 mm wide
Thickness Width PF
mm mm %
1.32 14.48 84.8
1.3 14.5 86
1.28 14.55 87.1
1.26 14.6 88.1
Rectangular Cable Compaction Study
Mid-thickness Width PF
mm mm %
1.27 14.66 87.1
1.253 14.66 88.3
1.23 14.68 89.8
Keystoned Cable Compaction Study
Keystoned Cable Compaction Study
Effect of Annealing
No Annealing
Annealing Mid-thickness Width PF
mm mm %
1.27 14.68 85.7
1.251 14.68 87.1
1.232 14.69 88.5
Focused on Cored Cable Technology using 150/169 RRP
Test Results – Ic retention and RRR
33
80%
85%
90%
95%
100%
105%
1.22 1.23 1.24 1.25 1.26 1.27 1.28
Ic_E
xtra
cted
/ Ic
_Ro
un
d @
12
T
Average Cable Thickness (mm)
150/169 Annealed
150/169 Non annealed
Specs
100
120
140
160
180
200
1.22 1.23 1.24 1.25 1.26 1.27 1.28
RR
R
Average Cable Thickness (mm)
150/169 Annealed
150/169 Non annealed
A mid-thickness spec of 1.25 mm meets the Ic
degradation requirements also in the case of a cored
cable. This allows using the same insulation thickness
and preserving the same magnet design when using
cored or uncored cable.
Specs v.2 Cable Production
Spec v.2 Cable production
234 m + 167 m (RRP-108/127) – practice cable
440 m piece (RRP-108/127)
o Two ~210 m long unit lengths for demonstrator model
o ~20 m for short sample studies
60 m + 120 m + 230 m (RRP-150/169) with 25 mm SS core
230 m (RRP-108/127) made at CERN for coil #3
03/06/2012 34
1.25 1.27
Coil Fabrication Status
PC#1 – rectangular Cu cable, two types of end parts
PC#2 – Nb3Sn cable (RRP-114/127) – coil was reacted and
impregnated. Is being used as mechanical coil.
Coil # 1 – Being instrumented for test.
Coil #2 – Being prepared for epoxy impregnation,
Coil #3 – Inner layer was wound, waiting for parts from
CERN to wind outer layer.
03/06/2012 35
36 9/15/2011
Failure Mechanisms of RRP Nb3Sn
“A Model to Study Plastic Deformation in Nb3Sn Wires”, IEEE Trans. Appl. Sup., V. 21, No. 3, p. 2588 (2011).
The outer walls of the Nb-Sn bundles break and merge together only after the Cu channels between them becomes thin enough. The Cu channel thickness goes to zero, i.e. breaks, at 22% deformation. At 26% deformation the Nb starts breaking in the innermost bundles, and propagating outward. At 30% deformation the merging has encompassed the whole thickness of the superconducting area.
Close-up of Von Mises strain distribution at 18% deformation. For this model, a principal tensor strain was defined in the Cu channels.
37 9/15/2011
Multi-Filamentary Composite Strand Modeling
Strand ID RRP1 RRP2
Stack design 150/169 108/127 Strand diameters, mm 0.7 1.0 0.7 1.0
Nb hexagon apothem, mm 18.8 27 21.6 30.4
Sn rod radius, mm 10.0 14.0 12.5 17.6
Cu spacing, mm 4.5 6.25 6.4 9.3
38 9/15/2011
Strain Modeling vs. Data (1)
26% Def.
1st 2nd diagonal
26% Def.
MODEL DATA
“Studies of Nb3Sn Strands based on the Restacked-Rod Process for High Field Accelerator Magnets”, MT-22, accepted in IEEE Trans. Appl. Sup. (2012).
39 9/15/2011
Strain Modeling vs. Data (2)
MODEL DATA
20% Def.
In the original work, the very good correlation between model and data had allowed identifying a critical criterion for RRP wires. For the Nb-Sn bundles not to merge and start breaking, the principal tensile strain in the Cu should not exceed 0.48 ± 0.10.
40 03/06/2012
Simplified Composite Strand Modeling
“FEM Analysis of Nb-Sn Rutherford-type Cables”, MT-22, accepted in IEEE Trans. Appl. Sup. (2012).
To build Rutherford cable models, the central hexagonal area was replaced by a homogeneous region with average properties weighted on the area percentage of each component. Such percentages were then corrected in order to obtain maximum lateral displacements within ~2% of those produced by the original
detailed model.
Cu Sn
Nb
Cu
Cu+Nb+Sn
Cux
y
The simplified model reproduces displacements very well, but does not represent local effects in detail.
Plastic Work (J/mm3)
41 03/06/2012
Rutherford Cable Modeling (1)
Vertical (dy) and lateral (dx1) displacements were imposed to reproduce the first rectangular forming stage
1 2 3 4 5 6 7 8 9 10 11 12 13 14
y
𝑥𝑢𝑛𝑑𝑒𝑓 = 2𝑥 + 𝑁 − 1 𝑟′
𝑥 = 𝑟′3
2 𝑟′ =
𝑟
𝑐𝑜𝑠(𝛼𝑝𝑖𝑡𝑐ℎ) ,
𝑦𝑢𝑛𝑑𝑒𝑓 = 4𝑟
The keystoning stage was modeled by imposing an additional vertical displacement varying linearly on the strands (represented by dy-mid and a), and another lateral displacement (dx2)
Undeformed geometry for a cable with odd number of strands
42 03/06/2012
Rutherford Cable Modeling (2)
Keystoning increases the number of areas with higher values of plastic energy, and therefore the average values of plastic work in the cable, but does not substantially increase the maximum value of plastic work in the edge strand, which is primarily determined by the rectangular step of the deformation.
Rectangular Stage
Keystoned Stage
Plastic Work (J/mm3)
43 03/06/2012
Detailed Modeling in Rutherford Cable
To identify local strains and critical locations inside the composite strands in the cable, the most critical strands in the cable cross section were modeled in detail by using surface displacements obtained through the simplified cable model. The goal ideally is to predict local damage whenever the failure mechanisms of a specific strand technology are known.
ba
ba
44 03/06/2012
Model of Edge Strand in 40-strand Cable
Values shown are of the principal tensile strain in the most critical Cu channels of the strand at cable edge. These values are consistent with the critical strain threshold of 0.48 ± 0.1 that had been observed in the single strand analysis.
45
Acknowledgments
The CERN-FNAL Collaboration:
B. Auchmann, B. Holzer, M. Karppinen, L. Oberli,
L. Rossi, D. Smekens (CERN)
N, Andreev, G. Apollinari, R. Bossert, F. Nobrega,
I. Novitski, A. Zlobin (FNAL)
My students and group members who participated
in the work shown:
G. Gallo, V. Lombardo, P. Neri, D. Turrioni, A.
Rusy, T. Van Raes, M. Bossert
My host E. Todesco for the kind invitation
Backup Slides
03/06/2012 46
Field Quality Correction
• The magnetization effect depends on the current pre-cycle: – Using Ires=100 A, b3 can
be reduced to stay within ±20 units between Iinj and Inom
• Passive correction – Deff_coil=45 µm;
– Deff_cor=55 µm
– Pre-cycle 0.8T-12.7 T
– No effect on b5
– Need more optimization
03/06/2012 -20
-10
0
10
20
30
40
50
60
70
80
0 2 4 6 8 10 12 14
b3
B0, T
No correction
2 cables/pole/midplane
V. Kashikhin
Stack 0.5mm
0.6mm
0.7mm 0.8mm 1.0mm
61 42mm 51mm 59mm 68mm 85mm
91
34mm 41mm 48mm 55mm 69mm
127
29mm 35mm 41mm 47mm 59mm
169
26mm 31mm 36mm 42mm 52mm
217 23mm 27mm 32mm 36mm 45mm
From Oxford Instruments
49 10/4/2011
Finally, a long length sample of more compacted keystoned cable obtained from a more compact and also narrower rectangular cable to allow for any width expansion due to spring back and annealing was made. This determined the narrower width to use in the rectangular stage.
Cable Width Optimization
The Ic (4.2 K, 12 T) degradation of the narrower rectangular cable was ~0%. The Ic degradation of the keystoned cable with 1.256 mm mid-thickness was ~4%. The single 440 m long piece for two unit lengths for the coils was made using the same cable geometry.
0
200
400
600
800
1000
1200
1400
1600
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Cu
rren
t, A
Magnetic Field, T
Round (12292-3) V-I
Round (12292-3) V-H
KS 1.256 V-I
KS 1.256 V-H
Keystoned stage
0
200
400
600
800
1000
1200
1400
1600
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Cu
rren
t, A
Magnetic Field, T
Round (13062-1) V-I
Round (13062-1) V-H
Rect. Ann. V-I
Rectangular stage
50 03/06/2012
Witness Sample Test Results
0
200
400
600
800
1000
1200
1400
1600
1800
2000
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Cri
tica
l Cu
rren
t (A
)
Magnetic Field (T)
MBH02_Witness_VI_4.2K
MBH03_Witness_VI_4.2K
MBH03_Witness_VH_4.2K
MBH02_Witness_VH_4.2K
Damage Analysis • Is performed on 6 cross sections for each cable
geometry. • Samples are polished to a finish of 0.05 microns. • The internal damage analysis counts the following:
– Broken Subelements – Merged Subelements – Damaged Subelements (Broken or Merged)
• Damage is normalized on the number of triangular ends at the cable edges, since independently of the edge being the thick or thin one, data and modeling show that the triangular configuration is that of maximum strain.
03/06/2012 51
Cable Geometries Analyzed
1.320mm RC
1.300mm RC
1.280mm RC
1.260mm RC
1.270mm KS Non-annealed
1.270mm KS Annealed 03/06/2012 52
DM-CF-02-0 Results
• All damage levels of the new cored cable made with the 150/169 wire were similar to or better than the nominal cables made with either 108/127 or 114/127 (in green).
• One damaged subelement was found in the middle of an annealed cable cross section, which is an abnormal location. We have only seen this kind of damage in other annealed cables.
Traveler
Keyston-
ing
RC
Thickness
KS
Thickness
Anneal
-ing
Strands
Used
No. Cross
Sections
Analyzed
No.
Triangular
Ends
No of
Strands
w/possible
damage
Total No.
Broken
Subelements
Total No.
Merged
Subelements
No. Damaged
Subelements
No. Damaged
Subelements/
No. Triangular
Ends
R&DT_110315
_40_1_0 Y 1.316 1.265 N 114/127 6 12 4 15 5 15 1.25
DM-CF-01-0a
(13) Y 1.297 1.272 Y 108/127 6 11 2 13 8.5 13 1.18 DM-CF-01-0b
(18) Y 1.294 1.256 Y 108/127 6 8 0 0 0 0 0 DM-CF-02-0
(A1) N 1.320 N/A N 150/169 6 9 1 5 5 5 0.56 DM-CF-02-0
(A2) N 1.300 N/A N 150/169 6 9 1 4.5 3 4.5 0.5 DM-CF-02-0
(A3) N 1.280 N/A N 150/169 6 8 1 2 0 2 0.25 DM-CF-02-0
(A4) N 1.260 N/A N 150/169 6 9 2 9.5 5 9.5 1.06 DM-CF-02-0
(B1) Y 1.320 1.270 N 150/169 6 9 2 4.5 4 4.5 0.5 DM-CF-02-0
(C1) Y 1.320 1.270 Y 150/169 6 8 4 12 9 12 1.5
03/06/2012 53
Location Map for Deformed Wire