High Field Magnets. Lucio Rossi CERN. CAS Intermediate level Course 2 October 2009

Size: px

Start display at page:

Download "High Field Magnets. Lucio Rossi CERN. CAS Intermediate level Course 2 October 2009"

Herbert Hopkins
5 years ago
Views:

1 High Field Magnets Lucio Rossi CERN CAS Intermediate level Course 2 October 2009

2 Content Definition and historic Basic of Sc magnets for accelerator Superconductivity and Nb Ti review Reasons to pursue HFM LHC lum up (high grad quads) LHC energy upgrade; upgrade of any Current Densities progress Issues: brittleness, insulation, radiation Structural design Fabrication technologies The perspectives and the roadmap Acknowledgements: A. Devred & E. Todesco CERN, G. Ambrosio FNAL, G.L. Sabbi LBNL and many others colleague of US LARP 2nd Oct 2009 L. Rossi CAS Darmstadt 2

quality is iron dominated PM configuration may reach 4 6 T peak (small bore)

3 When a magnetic field is high? 1 Iron dominated magnets < 2 T The coils are giving a minor contribution, field quality is iron dominated PM configuration may reach 4 6 T peak (small bore) Stored energy in iron is very small E d H V 2nd Oct 2009 B V dbd CERN LHC MQW Twin resistive! L. Rossi CAS Darmstadt B H 3

4 When a field is high? 2 Superconductor may reach more The field that can be reached depends on the temperature Sc not only reach high field but also they do not dissipate! 2nd Oct 2009 L. Rossi CAS Darmstadt 4

Resistive coils: 40 T in LN, 20 mm 77 K bore.

5 When a field is high? 3 However simple copper coil cryogenically cooled (typically Liquid 77 K) can give more Resistive coils: 40 T in LN, 20 mm 77 K bore. It relies on heat capacity of the conductor reinforced copper reaching 300 K each shot. 4 milliseconds of field duration. A little longer ( ms ms), 10 mm bore the 60 T long pulse of NHMFL in Florida 2nd Oct 2009 L. Rossi CAS Darmstadt 5

When a field is high? 4 And, surprisingly as it may be, the highest d.c.

Hybrid system (SC outsert with a resistive insert) provide the record 45 T

6 When a field is high? 4 And, surprisingly as it may be, the highest d.c. field are also obtained by copper coils! Field in excess of 30 T in steady state! However an enormous cooling power of MW is needed for mm bore Hybrid system (SC outsert with a resistive insert) provide the record 45 T field. 900 MHz NMR NHMFL SCH Keck 45T Hybrid 13.8 T Nb 3 Sn Superconducting Outsert 22.3 T Resistive Insert 2nd Oct 2009 L. Rossi CAS Darmstadt 6

7 When a field is high? 5 In practice : High Field Magnet (HFM) is what is beyond LHC, 10 T! 2nd Oct 2009 L. Rossi CAS Darmstadt 7

8 High field: accelerator vs. detector magnets Magnetic fields needed for - electric charge identification - momentum spectrometry - p = mv = q B; = q/p B L BL is often the comparison parameter If momentum analysis is done by tracking inside the field volume: - p/p 1/BL 2 large volume better than high field - Field homogeneity appreciated but NOT critical (field knowledge of 0.1% usually suffices) 2nd Oct 2009 L. Rossi CAS Darmstadt 8

9 Accelerator magnets : basic 1 Field shape DIPOLE Half the field of a solenoid for same J and coil thickness J = J 0 cos Field shape QUADRUPOLES J = J 0 cos2 - - w r + + B 2nd Oct 2009 Jd 2 0 e y B y 0 a Jbd b L. Rossi CAS Darmstadt B B x y 0J ( a b) y a b 0J ( a b) x a b 9

It can be shown that if the cos(n ) is approximate by step function, there is a magic angle that makes nil the

10 Accelerator magnets : basic 2 In practice the above current distributions are approximate, so the field contains also higher order harmonics (see later). It can be shown that if the cos(n ) is approximate by step function, there is a magic angle that makes nil the first higher order harmonics. Approximation of cos with coil blocks (left) and multiple shells (centre) and of intersecting ellipses (from Wilson book). 2nd Oct 2009 L. Rossi CAS Darmstadt 10

40 0-40 0 40-40 Accelerator magnets : basic 3 The basic shape : mix between cos and shell Shells with const J is a very good approximation

11 Accelerator magnets : basic 3 The basic shape : mix between cos and shell Shells with const J is a very good approximation Field expansion y - - B y ib x 10 4 B 1 n 1 ( b n ia n ) x iy R ref n R ref r + + x - - 2nd Oct 2009 L. Rossi CAS Darmstadt 11

12 Accelerator magnets : basic 4 Transverse field: no self supporting w.r.t. stress J overall 500 A/mm 2! e.m. forces are not kept by conductors but tend to torn apart the winding. Midplane 2nd Oct 2009 L. Rossi CAS Darmstadt Area of Maximum Displacements 12

13 Accelerator magnets : basic 5 All in series The worst performing determines the accelerator energy They must be all equal (within few units, 1 u=10 4 ) Stand alone magnets (DS, MS, low beta triplets) they also must ha The field quality is typically controlled at a fraction of unit over a wide range. Sigma of coil waviness (mm) Coil positioning and e.m. forces 30 micron target Firm 1 Firm 2 Firm 3 Collared coil Magnet progressive number AT-MAS updated on Firm 3: Inner layers coil size and E-modulus Collared coil number Inner layer coil size (average of left & right sides) Inner layer E-modulus at 80 MPa Coil E-modulus at 80 MPa [GPa] 2nd Oct 2009 L. Rossi CAS Darmstadt 13

14 SC and critical parameters: Critical Temperature 2nd Oct 2009 L. Rossi CAS Darmstadt 14

15 SC and critical parameters: Critical Field 2nd Oct 2009 L. Rossi CAS Darmstadt 15

16 SC and critical parameters: Critical Current Densities 2nd Oct 2009 L. Rossi CAS Darmstadt 16

17 (Overall) Engineering J 4.2 K YBCO B Tape Plane YBCO B Tape Plane J E (A/mm²) MgB 2 Nb-Ti 18+1 MgB 2 /Nb/Cu/Monel Courtesy M. Tomsic, 2007 Domain of iron dominated magnets Maximal J E for entire LHC Nb Ti strand production (CERN T. Boutboul '07) RRP Nb 3 Sn Bronze Nb 3 Sn Complied from ASC'02 and ICMC'03 papers (J. Parrell OI ST) 4543 filament High Sn Bronze 16wt.%Sn 0.3wt%Ti (Miyazaki MT18 IEEE 04) Applied Field (T) filament strand with Ag alloy outer sheath tested at NHMFL YBCO Insert Tape (B Tape Plane) YBCO Insert Tape (B Tape Plane) MgB 2 19Fil 24% Fill (HyperTech) 2212 OI-ST 28% Ceramic Filaments NbTi LHC Production 38%SC (4.2 K) Nb 3 Sn RRP Internal Sn (OI-ST) SuperPower tape used in record breaking NHMFL insert coil 2007 Nb 3 Sn High Sn Bronze Cu:Non-Cu 0.3 2nd Oct 2009 L. Rossi CAS Darmstadt 17

30 years of large projects 300 500 tonnes the yearly

18 Nb Ti the almost invariable choice Used in 95% of the Sc magnets Ductile, robust, relatively easy to manufacture, 30 years of large projects tonnes the yearly production (70% for MRI) 2nd Oct 2009 L. Rossi CAS Darmstadt 18

and boosted by use of HEII 1985 1990: LHC conceptual design two routes, one in

19 and boosted by use of HEII : LHC conceptual design two routes, one in Nb 3 Sn alternative to Nb Ti. 1 model magnet successful at about K but Extra cost of HEII vs. LHe: 15% of the cryomagnet system Extra cost of the Nb 3 Sn vs. NbTi : at least 50%... 2nd Oct 2009 L. Rossi CAS Darmstadt 19

20 Good reasons to pursue HFM : 1 LHC HigherG LargerA low triplet L f rev ( N b ) 4 n * 2 F( ) nb * Q1 Q2a Q2b Q3 low triplet D2 Sc magnet in LHC 2nd Oct 2009 L. Rossi CAS Darmstadt 20

21 What we can get Order zero: G /2 critical field 13T for Nb-Ti, 25T for Nb 3 Sn This is a bad approximation! Results relative to a sector coil for 100 mm Nb-Ti: G /2 10 T Nb 3 Sn: G /2 15 T Nb 3 Sn: 50% more than Nb-Ti Some dependence on aperture Better for large apertures A safety margin of 20% is then subtracted on both cases Gradient* /2 (T) Operational gradient [T /m] 80% of Nb 3 Sn at 1.9 K 80% of Nb-Ti at 1.9 K Magnet aperture (mm) 500 Nb 3 Sn at 1.9 K % 200 Nb-Ti at 1.9 K +47% % Aperture [mm] 2nd Oct 2009 L. Rossi CAS Darmstadt 21

22 Scaling laws for LHC triplets Solutions can be found for both materials. What we need is a given focusing strength: G lt Large apertures: is this possible? Increase of lt, triplet length < 50 m (today 20, phase 1 lt=40 m) Stresses, difficult beyond =120 mm. We need mm, R&D needed. aberrations? 2nd Oct 2009 Gradient (T/m) Nb-Ti 1.9 K l*=23 m Nb3Sn 1.9 K lt=20 m lt=25 m lt=30 m lt=40 m lt=50 m *=55 cm *=28 cm *=14 cm *=7 cm Aperture (mm) Stress [MPa] Nb 3 Sn 1.9 K 2r=40 mm 2r=80 mm 2r=120 mm 2r=160 mm 2r=200 mm 2r=240 mm L. Rossi CAS Darmstadt Critical gradient [T/m] 22

23 Luminosity up: more than low beta quads Separation dipoles Today 1.5 and 3 T In 2014 upgrade 3.5 and 4.5 T In final lum up: may be 7 10 T needed at 4.2 K Nb3Sn or HTS (here high temperature may be an invaluable asset Matching section: large quads with samse G*L are really an asset (for LHC) Dispersion suppression zone to make room for collimators: LHC 8.3 T x 14.2 m. 12 T x 10 m!!! Possible 2nd Oct 2009 L. Rossi CAS Darmstadt 23

24 Good reasons to pursue HFM : 2 Energy upgrade with same size Evolution of the gluon spectrum Assumptions: Luminosity grows x3 with adiabatic damping Luminosity needed to produce a given number of particles of mass m (assuming gauge couplings constant) scales with m 2 So twice the mass scale requires 4/3 the luminosity. Triple the energy double the mass reach P. McIntyre Texas A&M Acc. center 2nd Oct 2009 L. Rossi CAS Darmstadt Dutta

Extend to 24 Tesla: Bi 2212 in inner (high field) windings, Dual dipole (ala LHC) Bore field Nb 3 Sn in outer (low field) windings 24 Tesla Max stress in superconductor 130 MPa

25 Extend to 24 Tesla: Bi 2212 in inner (high field) windings, Dual dipole (ala LHC) Bore field Nb 3 Sn in outer (low field) windings 24 Tesla Max stress in superconductor 130 MPa Superconductor x section: Cable current Beam tube dia. P. McIntyre Texas A&M Acc. center Nb 3 Sn 26 cm 2 Bi cm 2 25 ka 50 mm Beam separation 194 mm 2nd Oct 2009 L. Rossi CAS Darmstadt 25

26 More modestly: 20 T.E 50 mm aperture 20 Tesla operational field Inner layers: High Tc superconductor Outer layers: Nb 3 Sn Operational current: 18 KA Operational current density: 400 A/mm 2 20% operational margin: Max B = 24 T 15 m 500 HTS Nb 3 Sn y(mm) HTS Nb 3 Sn N W 0 y (mm) N W LHC size Km Km x (mm) x (mm) 2nd Oct 2009 L. Rossi CAS Darmstadt 26

27 Good reasons to pursue HFM : 3 Medical accelerators : hadron teraphy 2nd Oct 2009 L. Rossi CAS Darmstadt 27

28 Protons Current Beam Delivery System Proton Cyclotron- 250 tonnes Treatment Gantry- 100 tonnes

29 One room plus equipment: >$50 million Four rooms plus equipment: >$100 million Big problem to manage project, staff, patients Price: much more than $100 million Problem Proton Beam Radiation Therapy costs prohibitively are too high 2nd Oct 2009 L. Rossi CAS Darmstadt 29

30 High Field (8 12 T) Superconducting magnets reduce size of accelerator Accelerator on conventional X ray gantry Allows retrofit into existing space Solution High Field Weak Focusing Cyclotron Price: less than $15 million 2nd Oct 2009 L. Rossi CAS Darmstadt 30

Good reasons to pursue HFM : 4 Undulators and Wigglers SC

280 mm 8 NbTi coils placed horizontally Operating at 4.

Superconducting Undulator for the LHC Synchrotron Radiation

31 Good reasons to pursue HFM : 4 Undulators and Wigglers SC undulator in LHC NbTi Undulator for the LHC beam monitor Period = 280 mm 8 NbTi coils placed horizontally Operating at 4.4 K with Current=450 A Design magnetic flux density = 5 T (gap) T= 280 mm B gap = 5 T / B coil 5.9 T / B pole 7 T Manufacture and Test of the Prototype 5 T Superconducting Undulator for the LHC Synchrotron Radiation Profile Monitor Maccaferri, R ; Bettoni, S ; Tommasini, D ; Venturini Delsolaro, W 2nd Oct 2009 L. Rossi CAS Darmstadt 31

32 Upgrade of existing LHC undulators for LHC ion beam Flexible operation: undulator period 280 mm/140 mm with 60 mm gap B GAP = 5 T for 280 mm period which mean B COIL : 8 T(590 A/mm K B GAP =>3 T for 140 mm period which mean B COIL = 9T (550 A/mm K 2nd Oct 2009 Courtesy R. Maccaferri, N. Duarte CERN L. Rossi CAS Darmstadt 32

33 What about CLIC (and ILC)? 2nd Oct 2009 L. Rossi CAS Darmstadt 33

34 Test at CERN from Nb Ti to Nb 3 Sn Nb Ti winding trial NbSn single coil manufactured and tested: B > 10 T Courtesy R. Maccaferri, D. Schoerling CERN 2nd Oct 2009 L. Rossi CAS Darmstadt 34

35 Is Nb3Sn used in large scale? Yes about 100 HF NMR solenoids per year HF: the next large project ITER 900 MHz ( K) ITER: 13 T coils, 400 tons of Nb3Sn 2nd Oct 2009 L. Rossi CAS Darmstadt 35

36 Nb3Sn magnets : a long history W. Sampson Nb 3 Sn 76 mm quad BNL, K in a 50 mm bore, 1 m long (Twente University, CERN, 1995) 2nd Oct 2009 L. Rossi CAS Darmstadt 36

37 Attained field follow progress in J c Progress on non Cu JC (at 4.2 K and 12 T) of multifilament composite wires Progress on maximum quench field of dipole magnet models Non-EU EU IGC IT IGC IT TWCA MJR ECN PIT IGC IT OST MJR EM IT SMI PIT OST MJR IGC IT SMI PIT IGC IT OST MJR&RRP MELCO IT Alstom IT Time (year) ANSALDO K K ELIN MTACERN MTAJS LBNL D19 LBNL D20 (1.8 K) MSUT MFISC MSA4KEK MFRESCA LBNL RD-3 LBNL HD Time (year) 2nd Oct 2009 L. Rossi CAS Darmstadt 37

38 Where is the issue? Of course F B 2 brittleness Degradation of critical magnetic field as a function of strain (J.W. Ekin, 1984) F x F y Resulting Force per Quadrant (kn/m) LHC NbTi (56mm 8.3 T) MSUT/Twente Nb3 Sn (50mm 11 T) MFRESCA/CERN NbTi (88mm 10 T) D20/LBNL Nb3 Sn (50mm 13 T) Fx Fy Stored Energy (kj/m) 2nd Oct 2009 L. Rossi CAS Darmstadt 38

at 650 700 C x 5 20 days Internal Tin diffusion is best for J c but

Dia. Before reaction copper Sn pool Nb filaments copper Nb barrier

39 Nb3Sn wires Nb3Sn is a compound. Formed in solid state reaction in inert atm. at C x 5 20 days Internal Tin diffusion is best for J c but gives large effective filaments size This materila has 110 m eff. Fil. Dia. Before reaction copper Sn pool Nb filaments copper Nb barrier copper Nb 3Sn filaments tin reach bronze bronze Nb barrier (partially Nb 3 Sn) After reaction 2nd Oct 2009 L. Rossi CAS Darmstadt 39

40 After reaction Nb3Sn is very brittle Decrease in Ic upon longitudinal stress in PIT conductor Behaviour of Ic on Rutherford cable vs. transverse stress 2nd Oct 2009 L. Rossi CAS Darmstadt 40

41 And we need to use ka CABLE 2nd Oct 2009 L. Rossi CAS Darmstadt 41

42 2 nd issue 2nd Oct 2009 L. Rossi CAS Darmstadt 42

43 Flux jump: sudden movement of fluxoids Fluxoid array interact with current. When f = J x B > f pinning the flux moves and generates heat Criteria to avoid this avalanche effect: 2 2 0J c a 3 C( T T ) c op First photo of the fluxoid array In practice this means filament diameter of less than 100 m Embedding fine filaments inside a good conductor (copper for LTS) to achieve both goals 2nd Oct 2009 L. Rossi CAS Darmstadt 43

44 2nd Oct 2009 L. Rossi CAS Darmstadt 44

45 2nd Oct 2009 L. Rossi CAS Darmstadt 45

46 More conductor development Wire with only 50 m filam. Dia. And with more Cu (58%), 1 mm Powder intube (Europe under developemnt Building comprhensive 3 D model to understand deeformation and its dynamics 2nd Oct 2009 L. Rossi CAS Darmstadt 46

47 Intrinsic diamagnetism of Sc and magnegtization loop G. Ambrosio, FNAL, CERN Ac. Training nd Oct 2009 L. Rossi CAS Darmstadt 47

48 Correction works b3 LHC range G. Ambrosio, FNAL, CERN Ac. Training nd Oct 2009 L. Rossi CAS Darmstadt 48

49 Coil technology issues: W&R vs. R&W Despite few attempt R&W has to be abandoned (good for large radius jacketed coils like ITER 2nd Oct 2009 L. Rossi CAS Darmstadt 49

50 Insulation: a further issue 2nd Oct 2009 L. Rossi CAS Darmstadt 50

51 Heat treatment (steps 1 3) Vacuum impegnation (step 4) nd Oct 2009 L. Rossi CAS Darmstadt 51

Mechanical structure: the classical route The long tradition from BNL and Fermilab:

quality is in good part determined by collar shape If the coils size is not so well

is applied because of stress losses during cooldown: for very high field tends to be

52 Mechanical structure: the classical route The long tradition from BNL and Fermilab: collaring concept Pros and cons The coils is well contained in a fixed cavity Field quality is in good part determined by collar shape If the coils size is not so well controlled, the stress can be too high or too low A prestress 2 3 times what is need is applied because of stress losses during cooldown: for very high field tends to be too high Since NbSn is stress sensitive 2nd Oct 2009 L. Rossi CAS Darmstadt 52

Mechanical structure The new route (from LBNL) Suppose the inner blue ring (iron pads) just

Then you pump in the bladders. Alu shell expands and goes in tension. The keys become loose.

Then depressurize the bladders and remove.

53 Mechanical structure The new route (from LBNL) Suppose the inner blue ring (iron pads) just in contact wit key with yoke. No prestress. Then you insert the bladders in the interstice. Then you pump in the bladders. Alu shell expands and goes in tension. The keys become loose. Remove the keys and put new keys as thick as the new interstice. Then depressurize the bladders and remove. The Al shell, not anymore pushed by the bladders exerts all its pressure on the coils through yoke keys iron pads. Inflatable bladders 2nd Oct 2009 L. Rossi CAS Darmstadt 53

54 3d_lay2_top A lof of 3 D FEM 2nd Oct 2009 L. Rossi CAS Darmstadt 54

55 The last product HD2 a real magnet with bore (LBNL) 2nd Oct 2009 L. Rossi CAS Darmstadt 55

56 And it works as first model 2nd Oct 2009 L. Rossi CAS Darmstadt 56

57 Roadmap for HG quadrupole 2nd Oct 2009 L. Rossi CAS Darmstadt 57

58 Fist long (4 m) HF magnet test Nov 09: 200 T/m in 90 mm bore 2nd Oct 2009 L. Rossi CAS Darmstadt 58

59 Next step: near our needs: G> 200 T/m in 120 mm aperture 2nd Oct 2009 L. Rossi CAS Darmstadt 59

60 And for dipole? NbSn cannot go beyond T 15 Nb 3 Sn at 4.2 K Bc (T) Nb-Ti at 4.2 K r=30 mm r=60 mm r=120 mm Cos theta equivalent width (mm) 2nd Oct 2009 L. Rossi CAS Darmstadt 60

61 Can HTS made to work for us? YBCO B Tape Plane YBCO B Tape Plane J E (A/mm²) MgB 2 Nb-Ti 18+1 MgB 2 /Nb/Cu/Monel Courtesy M. Tomsic, 2007 Domain of iron dominated magnets Maximal J E for entire LHC Nb Ti strand production (CERN T. Boutboul '07) RRP Nb 3 Sn Bronze Nb 3 Sn Complied from ASC'02 and ICMC'03 papers (J. Parrell OI ST) 4543 filament High Sn Bronze 16wt.%Sn 0.3wt%Ti (Miyazaki MT18 IEEE 04) Applied Field (T) filament strand with Ag alloy outer sheath tested at NHMFL YBCO Insert Tape (B Tape Plane) YBCO Insert Tape (B Tape Plane) MgB 2 19Fil 24% Fill (HyperTech) 2212 OI-ST 28% Ceramic Filaments NbTi LHC Production 38%SC (4.2 K) Nb 3 Sn RRP Internal Sn (OI-ST) SuperPower tape used in record breaking NHMFL insert coil 2007 Nb 3 Sn High Sn Bronze Cu:Non-Cu 0.3 2nd Oct 2009 L. Rossi CAS Darmstadt 61

62 Part of R&D in Magnet Labs Cable fabrication YBCO SUPERPOWER Record field (25 T), adding 3 T NHMFL Florida Bi 2212 Coil Winding 2nd Oct 2009 L. Rossi CAS Darmstadt 62

63 Their technology is more difficult than Nb 3 Sn Nb3Sn Limit of T (acc. mag.) Very high Jc Low Tc (average stability, easy to protect) Available in long wire Developed ka cables Brittle but we can work out Th. Treatment: difficult HTS Can go to 25 T and beyond Realtively low Jc High Tc (better stability, more difficult to protect) Wire uniformity is an issue Cabling possible for Biscco, not yet for Ybco Very brittle, still to learn Th. Treatment: much more difficult, with leakage 2nd Oct 2009 L. Rossi CAS Darmstadt 63

64 Where can go with HF dipoles? 2nd Oct 2009 L. Rossi CAS Darmstadt 64

65 Where we can go? Design to fit HTS? To the farthest energy frontier! Coil #1 # m a Be 1 m# a e B 2 Coil #2 Main Coils of the Common Coil Design Courtesy R. Gupta BNL 2nd Oct 2009 L. Rossi CAS Darmstadt 65

High Field Magnets Perspectives from High Energy Physics. Dr. Glen Crawford Director, Research and Technology R&D DOE Office of High Energy Physics

High Field Magnets Perspectives from High Energy Physics Dr. Glen Crawford Director, Research and Technology R&D DOE Office of High Energy Physics What is High Energy Physics? The High Energy Physics (HEP)