World record DC (45.5 T) magnetic field using a REBCO insert inside a 31 T resistive magnet

Transcription

1 World record DC (45.5 T) magnetic field using a REBCO insert inside a 31 T resistive magnet Seungyong Hahn1, Kwanglok Kim, Kwangmin Kim, Xinbo Hu, Thomas Painter, Iain Dixon, Seokho Kim2, Kabindra Bhattarai, So Noguchi3, Jan Jaroszynski, and David Larbalestier September 11, 2018 *Support by NHMFL NSF core grant, DOE-High Energy Physics (HEP), CERN, NIH, DOE-SBIR pass through awards, and KBSI to the group of Seungyong Hahn. 1Also Seoul National University, Seoul, Korea, 2on leave from Changwon University Korea, 3on leave from Hokkaido University Japan World Record DC Magnetic Field using an REBa2Cu3Ox (RE = Y,Gd) Superconducting Magnet Seungyong Hahn, Kwanglok Kim, Kwangmin Kim, Xinbo Hu, Thomas Painter, Iain Dixon, Seokho Kim, Kabindra Bhattarai, So Noguchi, Jan Jaroszynski, and David Larbalestier Manuscript under review

2 History of >30 T test coils Explorations of test magnets with fields above 30 T in the 50 mm warm bore 31 T Bitter magnet at the NHMFL (He cryostat bore 37 mm) Markiewicz et al T (AIP Conf. Series 1218, 225 (2010) Pancake, insulated SuperPower REBCO with 50 um Hastelloy Trociewitz et al T (Appl. Phys. Letts. 99, (2011) Layer wound insulated SuperPower REBCO with 50 um Hastelloy Larbalestier et al T (Nature Materials 13, (2014) 10 bar over-pressure processed Bi-2212 Hahn et al. 40 T, 42.5 T and 45.5 T (Under review) NI pancake construction using 30 um substrate SuperPower REBCO tape Increasing fields demand ever increasing Coil winding current density J W : NI REBCO is the most practical path to this goal with minimum Cu and substrate thickness CCA Vienna, September 11-14, 2018 David C Larbalestier 2

3 The NI path to the highest possible field B~ µ 0 J W d (long solenoid approximation) For J W = 1000 A/mm 2, winding thickness d = 10 mm (OD 34 mm, ID 14 mm) B~ 15 T How to maximize J W? Minimize substrate to that required to support the stresses 30 µm cold-rolled Hastelloy C276 being offered by SuperPower in 2016 Use NI technique to avoid insulation, minimize Cu thickness (7.5 µm each side used) and maximize REBCO fraction (1.5 µm out of 45 µm total, ~3%) Avoid impregnation with epoxy to avoid any delamination problem hoop stresses then given by σ θ = J.B.r, increasing with turn radius in the pancake Pancake wind to be compatible with NI technique and to get maximum REBCO packing factor CCA Vienna, September 11-14, 2018 David C Larbalestier 3

4 Very LBC1 compact, (April 2016): very NI high REBCO field NI with (No developmental Insulation) REBCO SuperPower magnets 30 are coming µm substrate to NHMFL gave 40 T inside 31 T Seungyong Hahn et al. : m of 7.5Zr tape in 12 single pancakes (14 ID/34 OD/53mm long) Large helium gas bubble produced by joint dissipation Safe quench at 18 K T at J E = 905 A/mm 2 CCA Vienna, September 11-14,

5 LBC: A very compact, powerful magnet test bed! (April 11, 2016) CCA Vienna, September 11-14, 2018 David C Larbalestier 5

6 LBC2 generated 42.5 T in January 2017 Tape current density, JJ ee : 1151 A/mm 2 ; coil current density: 1040 A/mm 2. Coil temperature: >10 K due to the helium bubble; survived quench without burn-out JJ ee : 1151 A/mm 2 JJ ww : 1040 A/mm 2 CCA Vienna, September 11-14,

7 LBC2 post mortem Hall probe fixing screw indented ~67 turns of bottom pancake (#1) and 42.5 T quench initiated from this pancake Large mechanical damage seen in many turns G10 sheet piece No denting away from screw CCA Vienna, September 11-14, 2018 David C Larbalestier 7

8 LBC3 planning Full YateStar analysis of the LBC3 tape: As received After winding and unwinding After test unwind each pancake and look for damage Active attempt to minimize charging losses by slow ramp and removal of He gas with localized gas pumping CCA Vienna, September 11-14, 2018 David C Larbalestier 8

9 YateStar transport in perpendicular and parallel field (or variable angle) and remanent field Hall probe array Transport device invented by Yates Coulter, LANL, NHMFL Further developed with magnetization at ASC, NHMFL Reel #1 I ʘ Electro-magnet, H c Voltage taps LN 2 bath Hall probe array Rotating permanent magnet Voltage taps Reel #2 I Transport Current goes from one pulley to the other biangular vs. position or vs. angle, field Magnetization Remnant field Hall probe array with 7 sensors Transport sensitivity with ~2 cm resolution with B parallel and perpendicular Magnetization sensitivity with mm resolution in remanent field Yates [1] Coulter, J.Y., et al., IEEE Trans. Appl. Supercond., 17, (2007) [2] Coulter, J.Y., et al., IEEE Trans. Appl. Supercond., 19, (2009) Transport at T with 2 orthogonal measurements (parallel to c and to ab-plane) reveals fluctuations of vortex pinning that are not visible in standard remanent magnetization TapeStar devices CCA Vienna, September 11-14, 2018 David C Larbalestier 9

10 YateStar transport does not degrade tapes Xinbo Hu Measurements on very thin 30 μm substrate x 4 mm wide tape used in NI LBC3 coil that achieved 45.5 T Tests before and after test are most revealing of response to strong mechanical stresses CCA Vienna, September 11-14, 2018 David C Larbalestier 10

11 Slitting is a trivial (but common) source of Ic(x) variation -Tape is in most cases slit from wider tape, typically 12 mm as made - Slitting wheel can wobble, giving rise to width fluctuations on scale of wheel circumference (28 cm here) -appeared in early 32 T production and was significant, leading to decision to use only 4 mm wide MS tape for 32 T pancakes 12 mm Markiewicz and Jaroszynski Vintage 2012 tape for 32 T: width variation due to slitting wheel wobble has almost disappeared today CCA Vienna, September 11-14, 2018 David C Larbalestier 11

12 LBC3: Design outline CCA Vienna, September 11-14, 2018 David C Larbalestier 12

13 LBC3 achieved 14.5 T inside 31 T: 45.5T World Record DC Magnetic Field using an REBa 2 Cu 3 O x (RE = Y,Gd) Superconducting Magnet Seungyong Hahn 1,2, Kwanglok Kim 1, Kwangmin Kim 1, Xinbo Hu 1, Thomas Painter 1, Iain Dixon 1, Seokho Kim 1,3, Kabindra Bhattarai 1, So Noguchi 1,4, Jan Jaroszynski 1, and David Larbalestier 1 Coil suffered one trip of the 31 T outer magnet without current in LBC3 no apparent damage LBC3 driven to quench at 45.5T Some added joint resistance noted after quench so further test was stopped Magnet was unwound and conductor run through Ic(77K) measurement device (YateStar) J e : 1419 A/mm 2 J w : 1262 A/mm 2 This coil operated at 5 times the conductor current density (J E ) of our 32 T insulated REBCO magnet where no current sharing is possible CCA Vienna, September 11-14, 2018 David C Larbalestier 13

14 Pre- Very and compact, post mortem very high testing field NI of (No LBC3 Insulation) in YateStar (at 77K) REBCO magnets are coming to NHMFL P1 P2 P3 P4 P5 P6 P7 Before test After test P8 Outer Inner P9 P10 P11 T=77 K, Bǁc, 0.6T P12 P2 and P11 are UNDAMAGED and have resisted both hoop tensile stresses of >400 MPa and the unpeeling stresses of the radial end fields Can 50 T be possible by orienting slit edges so that cracks do not propagate? No visible damage after test Black traces show Ic(x) before test Blue traces after test Paul Hu PhD Damage of two types occurred in 10 of 12 pancakes: a. cracks with increasing severity at larger radius and larger hoop stress (up to 440 MPa) b. wrinkling of outer tape surfaces where radial field is high and producing out of plane forces plastic deformation of whole conductor and REBCO cracks propagate from the slit side of the conductor

15 A closer look across the top half of the coil after post-test unwinding 7 Hall probe array across 4 mm width and along the full conductor length in each pancake O.D. P1 Slitting I.D. Wavy, BS Top P2 Slitting Flat, BS Undamaged PC has slit edge pointing in (PC2 and PC11) P3 P4 Slitting Slitting Wavy, BS Wavy, BS P5 Flat, MS P6 Wavy, MS Paul Hu PhD Tapes are made 12 mm wide and slit to 4 mm both edge and middle slit tapes are delivered P 2 (and P11) tape almost undamaged: its slit edge faces to the top of the coil and resists unpeeling forces 15

16 Macroscopic rippling is often seen after high-field test when a slit edge faces outwards towards the magnet ends Rippling on outer end winding of LBC2 Marked PLASTIC deformation of the tape occurs even though the REBCO is wound inside and should be supported by the Hastelloy substrate the plastic deformation is inducing delamination Cu μm μm 500 nm Buffer Substrate Good edge Wavy edge FIB cut 10 μm away from the wavy edge Internal damage appears to be triggered by macroscopic plastic deformation of the conductor CCA Vienna, September 11-14,

17 Many REBCO failure modes: lowest stress mode is unpeeling The failure stress of a REBCO tape is very highly anisotropic Tensile (axial) Tensile (transverse) Copper stabilizer F Hastelloy substrate F YBCO layer F F >700 MPa [1-3] MPa [7,10] Unfortunately end windings of solenoids favor unpeeling, the weakest mode Compressive (transverse) Shear F Cleavage Peel F F F F F F >100 MPa [15,16] >19 MPa <1 MPa (Tensile force / bonded area) Yanagisawa and Maeda (RIKEN) CCA Vienna, September 11-14, 2018 David C Larbalestier 17

18 Lessons of LBC so far We significantly exceeded the Kamerlingh Onnes dream of 1000 A/mm 2 for design of a 100 kilogauss magnet (IIR Chicago 1913) using Pb or Hg wires We survived quench (T max from model ~70 K) We observed significant mechanical damage: Macroscopic rippling of the conductor, especially end pancakes When slit edge was outward, major crack damage occurred in tapes, especially outer windings with largest hoop strains CCA Vienna, September 11-14, 2018 David C Larbalestier 18

19 Can proper orientation of the slit edge suppress cracking damage? Test coil geometry simulates the most vulnerable end windings Variation of orientation of the slit edge and elimination of slit edges are principal variables Central line of the magnet 100 mm Test coil B r ~1.6T 31 T magnet 17 mm CCA Vienna, September 11-14, 2018 David C Larbalestier 19

20 Parameters of off-center test coils (all in 31 T) Parameters Coil A Coil B Coil C Slit edge information Slit edge faces magnet center Slit edge faces magnet end Unslit tapes Substrate and thickness [μm] Hastelloy 30 Hastelloy 30 Stainless Steel 100 I.R.; O.R.; Height [mm] 15.88; 16.93; ; 16.98; ; 17.08; 4 Number of turns REBCO tape length [m] Inductance [μh] Magnet constant [mt/a] R c (measured) [mω] R ct [mω cm 2 ] Charging time constant [s] I op range [A] Tape current density at 250 A [A/mm 2 ] Total strain at r = O.R., 250 A [%] ; (500A); B r at r = a 1 and max. I op [T] ~1.6 ~1.6 ~1.6 T CCA Vienna, September 11-14, 2018 David C Larbalestier 20

21 Coil A: high field tests Coil A: Off-centered, slit-edge inside, without voltage tap for inner joint Total 8 cycles at I op = A Current ramp rate: 1 A/s I op = A CCA Vienna, September 11-14, 2018 David C Larbalestier 21

22 Coil A post mortem No obvious test damage Current was cycled 8 times between A. Inner Before After Slit edge Inner Outer groove up Damage concentrates on turn7-10, slit edge. No plastic deformation seen in tapes small cracks believed due to winding or local stress concentrations at the groove in the winding mandrel: maximum strain 0.27% Slight Ic drops may be because of the groove. CCA Vienna, September 11-14, 2018 David C Larbalestier 22

23 Coil B Never quenched but clear tape deformation and degradation Rippling degradation over whole top (slit) edge of the coil buckling Inner Before Slit edge Outer After Periodic damage matches the circumferences of the coil turns ~50 mm David C Larbalestier CCA Vienna, September 11-14,

24 Another look at P1 and P2 of LBC3 P1 Outer edge: Delamination on the edge Inner edge: No delamination observed. Outer edge: Localized delamination observed P2 Inner edge: Mostly are cracks due to slitting. Delamination also observed. CCA Vienna, September 11-14, 2018 David C Larbalestier 24

25 Post mortem of not-slit tape Due to solder joint Operation current reached ~600 A. The inner terminal failed and caused the quench. ~220 mm Interesting result: tape was plastically deformed significantly and stiff, hence poor Hall probe signal but not damaged in transport evaluation CCA Vienna, September 11-14, 2018 David C Larbalestier 25

26 Summary thoughts Thin, strong substrates of cold-rolled Hastelloy made into extremely high J E and high J w coils can now double the maximum field of any Nb 3 Sn magnet Such coils challenge multiple facets of the conductor design that we are just beginning to explore: Stress Quench Artifacts of tape manufacture e.g. slitting method or absence We hope to approach 50 T soon but the territory shows new challenges for each LBC CCA Vienna, September 11-14, 2018 David C Larbalestier 26

27 Thanks We offer many thanks to Drew Hazelton and the SuperPower team for their unique 30 mm substrate tape Following the tests of LBC3 and understanding the importance of slitting damage, we are very grateful to Bruker HTS/ Alexander Usoskin for provision of a short test length to evaluate the potential of never-slit tape for such a magnet Many thanks to colleagues at the NHMFL, especially Dmytro Abraimov, Hongyu Bai, Mark Bird, Greg Boebinger, Lance Cooley, Eric Palm, Huub Weijers, and many others for helpful discussions. CCA Vienna, September 11-14, 2018 David C Larbalestier 27