EUROPEAN ORGANIZATION FOR NUCLEAR RESEARCH CERN-AT DIVISION QERM - RT QQ - at F1 A

he `·t I L

CERN AT/93-46 (MA) caew Lteeeeias. eamave LHC Note 255 ||U||\\||||\ll||||l||I\|\l||\!\|ll|I||lIl|||\1\\\I| P®@BEl413

Magnets for the Large Hadron Collider

presented by R. Perin and ]. Vlogaert for the LHC Magnet Team

Abstrut

CERN is finalising the design of the Large Hadron Collider (LHC), a proton-proton accelerator/collider to be installed in the existing 27 km long LEP tunnel.

The principal component of the projected machine is a high-field superconducting magnet system embracing two beam pipes in common yokes and cryostats. The dipole magnets, to guide the beams in their quasi circular orbit, will provide a field of 8.65 T for a 14 TeV centre of mass energy in proton-proton collisions. The corresponding field gradient of the main focusing quadrupole magnets will be 220 T/m. To reach these high performances the magnets will be cooled by superfluid helium at 1.9 K. Dipoles and quadrupoles will be of the two-in—one type. There will be about 1300 thirteen metre long dipoles, 380 three metre long quadnxpoles and a large number of other superconducting magnets for correction and special purposes, for a total of about 10'000 superconducting magnetic umts.

The magnet R&D programme, which was started several years ago, in centred on the main dipoles and quadrupoles. Prototype short dipoles have reached the record field of 10.5 T.

European Conference on Applied Superconductivity, EUCAS'93, Gottingen, Germany, 4-8 October 1993

Geneva, Switzerland 02/17/94 OCR Output

OCR OutputMamrets for the Large Hadron Colhder presented by R. Perin and J. Vlogaert for the LHC Magnet Team, CERN, European Organization for Nuclear Research, CH-1211 Geneva 23, Switzerland

Abstract CERN is finalising the design of the Large Hadron Collider (LHC), a proton-proton accelerator/collider to be installed in the existing 27 km long LEP tunnel.

The principal component of the projected machine is a high-field superconducting magnet system embracing two beam pipes in common yokes and cryostats. The dipole magnets, to guide the beams in their quasi circular orbit, will provide a field of 8.65 T for a 14 TeV centre of mass energy in proton-proton collisions. The corresponding field gradient of the main focusing quadrupole magnets will be 220 T/m. To reach these high performances the magnets will be cooled by superfluid helium at 1.9 K. Dipoles and quadrupoles will be of the two—in—one type. There will be about 1300 thirteen metre long dipoles, 380 three metre long quadrupoles and a large number of other superconducting magnets for correction and special purposes, for a total of about 10’O00 superconducting magnetic units.

The magnet R & D programme, which was started several years ago, is centred on the main dipoles and quadrupoles. Prototype short dipoles have reached the record field of 10.5 T.

A general description of the system is given together with the main design features and the state of the R & D programme.

1. The LHC Collider The new machine, a high—energy accelerator collider consists essentially of a double ring of high field superconducting magnets installed, together with the other machine components, in the same 27 km long underground tunnel, which houses the existing LEP collider [1].

Particles circulating in counter sense in the rings can be brought to collide in intersecting points.

The main performance parameters of the LHC for proton-proton and for Pb ion collisions are given in Table 1.

2. The magnet svstem. Constraints and basic choices As the circumference of the LHC is fixed by the LEP tunnel and the available cross-sectional space is limited, there has been from the start a quest for compact magnets operating at the strongest possible field to reach the highest collision energies. The dimensions of single magnets operating at the projected field level exclude the possibility to install two separate magnetic rings on top of the LEP machine. This constraint has led to the development of the two—in—one concept in which two sets of magnetic excitation windings are placed in a common yoke and cryostat assembly, leading to smaller cross-sectional dimensions and to more economic solution [2].

It was also decided that, for the LHC, coils wound from NbTi alloy operating at superfluid helium temperatures will be used.

C.m. energy for | No. of bunches | Particles per | Luminosity B = 8.7 T beam cm•2 $-1 TeV PP 14 2835 4_7 IQI4 1034 Pb ions 1150 560 5_2 1()10 1_g 1027 OCR Output Table 1: Main parameters of the LHC The Nb3Sn superconductor route at 4.2 K along which an industrially made single bore model magnet [3] was successfully tested is still being investigated with a twin-aperture dipole model under construction, but further development would be necessary for such an option to become technically and economically valid for the whole machine.

3. Lb About 24 km of the LHC ring will be occupied by superconducting magnets of various types. The long arcs, covering approximately 20 km of the circumference are composed of "standard cells", a bending focusing configuration which is periodically repeated 192 times around the ring.

One half of an arc cell has a length of 51 m and contains three 13.15 m long bending magnets and a 6.5 m long short straight section. In the straight section are installed: the 3.05 m main quadrupole, a beam position monitor and two corrector magnets, one with a combined dipole/sextupole corrector and the other one with a tuning quadrupole and octupole corrector. A set of sextupole and decapole correctors is attached to each main dipole to correct the field errors introduced by these magnets. Fig. l shows the layout of the standard half-cell.

Dipoles and main quadrupoles of the two rings are combined into "two—in-one" units, each pair having a common yoke and cryostat. Multipole correctors are assembled concentrically on a common beam pipe.

In addition to the regular arcs, there are other magnets in the dispersion suppressor section, and on either side of each crossing point.

Table 2 shows the types and numbers of superconducting magnets.

The LHC magnets will be one of the most massive applications of superconductivity. The quantity of conductor will be about 1400 t of which ~ 400 t will be N bTi alloy. The mass to be kept at 1.8 K temperature will be about 30'000 t distributed over a 27 km circumference.

3.1 Dipoles The design of the dipoles is made to meet the following requirements: Operational field: 8.65 T. Magnetic length: 13.145 m. Coil inner diameter: 56 mm. Distance between the axes of the apertures: 180 mm. Overall diameter of cold mass: S 580 mm.

The resulting main parameters are listed in Table 3, and the corresponding cross-section is shown in Fig. 2. Table 4 shows the dipole strand and cable characteristics.

The coils are formed of two winding layers made with keystoned cables of the same width but of different thickness, resulting from the wanted grading of current density for optimum use of the superconducting material. The proposed filament size allows the fabrication of superconducting wires by single stacking process. The persistent current sextupole component is - 3.56 x 10'4 and the decapole component is 0.18 x 10*4 for these filament diameters.

The force containment structure consists of the collars, the iron yoke and the shrinking cylinder which all contribute to produce the necessary azimuthal pre-compression in the coils. The shrinking cylinder is at the same time the outer shell of the helium tank, while the inner wall forms the beam vacuum chamber. The assembly between these two cylindrical walls, the "cold mass", is kept at 1.9 K in superfluid helium at atmospheric pressure and cooled by two—phase low—pressure helium circulating in a heat exchanger tube installed in an axial hole of the iron yoke. The magnet will be curved with a radius of 2700 m to match the beam paths. OCR Output 51.021

§§ 13.145 13.145 13.145 §

MB MB MB MQ

8 2 E 2 MB: Dipole magnets, MQ: Lattice quadrupolcs, MOQ: Combined octupoles and tuning quadrupolcs, MSB: Combined sextupoles and dipole corrector, BPM: Beam position monit0r,, - 0 -: Local sexwpole or decapole corrector Fig. 1: Layout of standard half cell

Tvne Nominal strength | Magnetic length (m) | Number Rezular lattice Main dipoles B0 = 8.65 T 13.15 1280 Main quadrupoles G = 220 T/m 3.10 384 Tuning quadrupoles G = 120 T/m + octupoles Bm = 1 X 105 T/ms 0.72 768 Combined correctors Dipole B = 1.5 T 6-pole Bu = 2000 T/m2 1.0 768 Sextupole correctors BH = 57(X) T/m2 ~0,1O 1280 Decapole correctors BIV = 4_6 X 107 T/m4 ~ 0.10 1280 Interaction rezions Twin—apertu1·e quads G = 220 T/m 3.7 to 8.0 130 Single-aperture quads G = 225 T/m 6.9 24 Separation dipoles B0 = 5.2 T 9.0 Table 2: Type and number of superconducting magnets

rational Held 8.65 T Coil aperture 56 mm Magnetic length 13.14 m

rating current 11'470 A Inner I Outer Lavcr I Laver ratmg temperature 1.9 K Coil mms per beam channel Str. inner shell 30 Diameter (mm) 1.065 I 0.825 outer shell 52 cu/sc ratio 1.6 I 1.9 Distance between aperture axes | 180 mm Filament size (um) I 7 I 6 Outer diameter of cold mass I 560 mm Twist pitch (mm) 25 I 25 Overall length of cold mass 14085 mm Critical current (A) Outer diameter of cryostat 980 mm 10 T, 1.9 K/9 T, 1.9 K 2 510 I z 370 Overall mass of cryomagnet | 29 t Cable Stored energy for both channels | 7.2 MJ Number of strands 28 I 36 Self—inductance for both channels I 110 mH Cable dimension Resultant of e-magn. forces in the lst width (mm) 15.0 I 15.0 coil quadrant 2Fx (1.80 MN/m) I 23.7 thin/thick edge (mm) 1.72/2.06 I 1.34/1.60 inner layer 2Fy (- 0.15 MN/m) | - 2.0 Transposition pitch (mm) 110 I 100 outer layer 2Fy (- 0.62 MN/m | - 8.2 Critical current (A) Axial e-magn. force on magnet ends I 0.55 10 T. 1.9 K/9 T. 1.9 K 2 13750 I 2 12880 Table 3: Dipole parameters Table 4: Dipole strand and cable characteristics OCR Output 3.2 Lattice Main Quadmpol The main quadrupoles are designed to provide a 220 T/m field gradient over a magnetic length of 3.01 m, on the basis of the two-in-one configuration with 0 56 mm coil aperture and distance between aperture axes of 180 mm.

The main parameters and the cable characteristics are listed in Tables 5 and 6. Their constructional features are very similar to that of the already built prototypes. The two layer coils will be wound from the same superconducting cable in the double pancake style. This technique avoids the inter layer splices which, being numerous in the quadrupole, would be a significant load to the cryogenic system.

3.3. Insertion Magnet There are a total of 166 special quadrupoles in the LHC insertions and straight sections differing in type, length and integrated gradient. Of this total, 114 magnets are of the lattice quadrupole type having the same cross-section but with three different lengths. The other ones are special magnets, based on a 70 mm aperture coil which is being developed for the low-beta quadrupoles. In the experimental insertions, these are single aperture units, while for other insertions a two-in-one version is being designed, which may, however, operate at 4.5 K.

Operational iield gradient I 220 T/m I I Strand diameter (mm) 0.825 Coil aperture 56 mm 28 Magnetic length N f d 3.01 m I I umbcr 0 Stian S Operating current 11470 A I I Cable dimensions Operating temperature 1.9 K Width (mm) 11_6() Tums per coil inner layer 4+6 . . thin/thick edge (mm) I 1.348/1.60 Outer layer 10 Transposition pitch (mm) I 95 Distance between aperture axes I 180 mm I I Critical current at 9 T, 1.9 K (A) I 2 10060 Yoke <>¤t¢rd1&m¢t¤r 444 mm Table Q: Characteristics of the cable for the

StoredSelf-inductance energy (both (both apertures) apertures)I 580 8.8 kJ mH I quadrupgleg Table 5: Main Parameters of LHC lattice quadrupole magnets

The low-beta quadrupoles are single—bore units 6.1 and 6.9 m long. The LHC performance depends critically on their field quality, especially on the higher order random multipole errors at low field. In order to obtain this field quality and to provide additional space for the cone of secondary particles emanating from the interaction, a novel design based on a graded coil with an aperture of 70 mm [4], wound from NbTi keystoned cables has been proposed. The design gradient of the magnet is 250 T/m for an excitation current of 5200 A. The design concept is presently being verified on a 1.3 m model magnet which is being developed in collaboration with an industrial firm.

Other magnets in the long straight sections and insertions include beam separation/recombination dipoles, skew quadrupoles, dipole correction magnets, and special dipoles for the dump insertion. Whereas most of the correcting elements are of the same design as that used for the lattice, twelve of the 192 skew quadrupoles require a 70 mm aperture, as do 40 correcting dipoles. The remaining skew quadrupoles are regular tuning quadrupoles rotated through 459 OCR Output 180

K0 .a<§?°"'i

*b"“ 4 E E r “' {2* I ;—0ll9

•¤!• \ Y / • I •

$56-/ \•r•/A¤\*.t• L we a 4

asso

Fig. 2: LHC dipole cross-section

3.4 Other Magnets The strengths of all other magnets, i.e. tuning quadrupoles and octupoles, sextupoles and correction dipole, sextupole and decapole correctors spools are lower than those of the previous design for which prototypes have been successfully built and tested or are in the construction phase. Therefore the further R & D work on them will be aimed at simpler and more economical solutions.

4. Status of the LHC magnet development The programme was mainly focused on the standard cell components and in particular on the dipoles, but included study and construction of several other magnets and items.

4.1 Dipoles The R & D effort covers: Superconductin g cables. Short single—ape1ture and twin—aperture models (~ 1.3 m long). Long twin-aperture prototypes (~ 10 m long).

4. 1.1 Superconducting cables Twenty-four kilometres of cable for the irmer coil shell made of 26, Z 1.29 mm strands, and 42 kilometres of cable for the outer coil shell, made of 40, 0 0.84 mm strands, have been developed and produced by the five European manufacturers and at present satisfy all main technical requirements. The process of fabrication of the wires and cables with NbTi filaments of 0 5 mm has been optimized for high fields. The current densities in the non-copper part of the strand cross-section, taken from the finished cables at the 3 G limit in the distribution curve of production, is 980 A/mm4 at 8 T, 4.2 K, for the inner layer and 2000 A/mm! at 6 T, 4.2 K, for the outer layer. The results of the models show that the short sample quenching field of the dipole (BSS) is close (within ~ 2%) to the cable short sample limit determined from strands extracted from the cables. OCR Output The total quantity produced, 14 t, is a significant amount and gives confidence that the required quality can be maintained in mass production.

4.1.2 Model mam Ten models of the main dipoles (see Table 7) have been made and tested [5, 6, 7]. Two magnets, identified by KEK in the table, were built in Japan by KEK in collaboration with industrial companies. All other magnets were entirely built by European companies except the MTA3/CERN model which was assembled at CERN using industry made coils and other components.

The most recent model, MTA3/CERN, was assembled at CERN using industry made coils basically identical to those of the previous models but wound from better performing last generation cables and with more compact ends and a support structure made with separated austenitic steel collars and a yoke vertically split into three parts.

A very high azimuthal compressive pre-stress was applied, on average 12 daN/mmi in the coil inner layer and 9 daN/mm! in the outer layer. The vertical gaps of the yoke were closed at assembly by applying a force of 2.8 MN/m by means of the stainless steel (316LN) shrinking cylinder. In those conditions the radial interference between the stainless steel collars and the yoke is 0.2 mm, and occurs over a 1 230 angle around the horizontal axis. At room temperature, the design average tensile stress in the 10 mm thick shrinking cylinder should be 14 daN/mmL· In the MTA3/CERN model the azimuthal stress went up to 28 daN/mm{ after the longitudinal welding of the cylinder two halves. At 1.9 K, the stress reached 46 daN/mm

At 4.3 K, the conductor short sample limit was attained at the second quench and no retraining occurred after thermal cycles.

At 2 K the magnet had the first quench above 9 T central field, reached 9.5 T in five quenches and finally attained the record field of 10.5 T. After thermal cycles to room temperature all quenches were above 9.75 T. The conductor limit was measured by fixing the excitation at 10.5 T and gradually raising the superfluid helium temperature: quench started at 2.13 K at the first turn of the coil inner layer in the straight part, where the field is at its peak value.

4.1.3 Ten meg lgng magnets [8, 9] A magnet with a CERN twin structure but coils identical to those of the HERA dipole was built in industry and successfully tested at CEA-Saclay at the nominal temperature of 1.9 K. This magnet, called TAP, exhibited excellent behaviour reaching the conductor short sample field at 4.2 K with no quench, and showing exactly the same properties as the single-aperture HERA magnets. When cooled down at 1.9 K, it was possible to increase the current from 6600 A to 9500 A thus reaching the conductor short sample limit of 8.3 T. It should be noted that the HERA coils have a larger aperture than the LHC coils (75 mm instead of 56 mm).

Subsequently, the construction of seven higher field (cable width 17 mm), 10 m long magnets was launched in four firms or consortia. All these magnets have the same type of coils, but three slightly different mechanical structures (Table 8).

The present status of manufacturing is as follows Several sets of coils have been made and precompressed in their collars, all mechanical components are nearing completion. The first fully assembled cold mass was finished in September 1993, and the magnet inserted in its cryostat should be delivered to CERN before the end of the year. This is the first of two magnets which are being supplied to CERN by INFN, Italy. OCR Output Outer Cable Collars BSSI B0 I N0.0f I reachedl quenches | quench tield after Strand I Widt smc I width mat. to pass | thermal N0. x G No. x 0 9T | cycle [mm] [mm] [mm] [mm] mm Single aperture models 8 TM1 50 30 x 0.84 12.6 30 x 0.84 12.6 9.3 9.3 3 I 8.9 8 TM2 50 30 x 0.84 12.6 30 x 0.84 12.6 9.4 9.4 3 I 8.9 MSA 1E I 50 26 x 1.29 17 40 x 0.84 17 9.8 9.7 1 1 I not tested MSA 2 KEK I 50 22 x 1.39 15 37 x 0.79 15 Mn/Al 10.2 9.8 4 I 9.5 Twin-apenure_m0dels MTA 1/E I 50 26 x 1.29 17 40 x 0.84 17 A1 (S) 9.8 9.0 12 I not tested MTA1/A | 50 26 x 1.29 17 40 x 0.84 17 Al (C) 9.8 9.7 9 I 8.3 MTA1/H | so 26 x 1.29 17 40 x 0.84 17 Al (C) 10.0 10.0 5 I 8.9 MTA1/JS I 50 26 x 1.29 17 40 x 0.84 17 A1 (C) 10.0 10.0 10 I 8.4 MTA2/KEK | 50 22 x 1.39 15 37 x 0.79 15 MnA1 10.2 under test M1‘A2pm~2·= I 50 26 x 1.29 17 40 x 0.84 17 SS (S) 10.5 10.5 I 0 I 9.75 MBTRA I 56 30x0.808 12.4 36xO.648 11.7 SS (S) 9.4 under constmction MFISC I 56 30 x 1.10 16.7 38 x 0.87 16.7 SS (S 9.9 under construction With industrially-made coils. Table 7: Characteristics of LHC magnet models (L = 1.3 M)

Name | No. | AperI11re| Inner Cable | Outer Cable I Collars I BSS I BO material reached uml G I smc Iwmml smc Iwmm No.x® No. x 0 [mm] I [mml I [mm mml I Imm TAP 1 I 75 I 2410.84 I 10 24 1 0.84 I 10 Al (C) 8.3 I 8.3 MT1>1/A I 3 I 50 I 2611.29 I 17 40 X 0.84 I 17 Al (C) 10.0 I Under MTP1/N I 2 I 50 I26x1.29 I 17 40 x 0.84 I 17 Al (C) 10.0 I cons MTP2/AJS I 1 I 50 I 26x1.29 I 17 40 x 0.84 I 17 A1 (S) 10.0 I uuc MTP3/EH I 1 I 50 I26x1.29I 17 40 x 0.84 I 17 SS (S 10.0 I tion (S): Separate collars. (C): Combined collars. Al: Aluminium alloy, SS: Austenitic steel, Mn: Mn steel. BO = Field in tl1e centre of the aperture. BSS = Field in the aperture centre at which the conductor placed in the magnet peak field position reaches the short sample limit. Table 8: Characteristics of LHC Twin-aperture magnet prototypes (L = 10 M)

4.2 Ouadnmoles Two full Size quadrupole magnets of f`ma1 aperture (56 mm) and length (3.5 m) have been designed and built by CEA-Saclay, using s.c. cables and structural parts made in industry. Their design and construction are described in [10].

The first one is under testing at CEA-Saclay, and will be installed in the string test.

5. Conclusions The first phase of the R & D programme has confirmed the soundness of the proposed technical choices for the LHC magnets: OCR Output The 10.5 T field has been passed for the first time in an accelerator dipole and in the twin aperture configuration foreseen for the LHC. This proves that an adequate margin with respect to the operational Held can be obtained. It has been demonstrated both in short and long magnets that twin-aperture (two-in-one) magnets behave as well as single-aperture magnets with respect to quench/training performance. It has been verified by measurements on model magnets that the required Held quality can be achieved.

Acknowledgment Several National Institudons and Laboratories have contributed in an important way to the reported R & D work, which received a very substantial support from industry. The LHC Magnet Team gratefully thanks them all.

References: [1] The LHC Study Group, "Design Study of the Large Hadron Collider (LHC), CERN 91-03, May 1991. [2] D. Leroy, R. Perin, G. de Rijk, W. Thomi, "Design of a High-field Twin-aperture S.C. Dipole Model, IEEE Trans. on Magn., vol. 24, March 1988, pp. l373—76. [3] A. Asner, R. Perin, S. Wenger, F. Zerobin, "First Nb3Sn, 1 m Long Superconducting Dipole Model Magnets for LHC Break the 10 T Field Threshold", Proc. llth Int. Conf. on Magnet Techn., Tsukuba, 1989, PP. 36-41. [4] R. Ostojic, T.M. Taylor, "Design and Construction of a One—Metre Model of the 70 mm Aperture Quadrupole for the LHC Low-b Insertions, presented at MT13, Victoria, Canada, 20-24 September 1993. [5] M. Bona, D. Leroy, R. Perin, P. Rohmig, B. Szeless, W. Thomi, "Design, Fabrication Variants and Results of LHC Twin-Aperture Models", IEEE Trans. on Magn., vol. 28, pp. 338-34l,Jan. 1992. [6] D. Leroy et al., "Test Results on 10 T LHC Superconducting One-Meter Long Dipole Models", IEEE Trans. on Applied Superconductivity, March 1993, vol. 3, No. 1, pp. 614 621. [7] A. Yamamoto et al., "Development of 10 T Dipole Magnets for the Large Hadron Collider", IEEE Trans. on Applied Superconductivity, March 1993, vol. 3, No. 1, pp. 769-772. [8] M. Gramer et al., "Performance of the Twin-Aperture Dipole for the CERN LHC, Proc. EPAC92, Berlin, March 1992. [9] E. Acerbi, M. Bona, D. Leroy, R. Perin, L. Rossi, "Development and Fabrication of the First 10 Metre Long Superconducting Dipole Prototype for the LHC", presented at MT13, Victoria, Canada, 20-24 September 1993. [10] J.M. Baze et al., "Design and Fabrication of the Prototype Superconducting Quadrupole for the CREN LHC Project", IEEE Trans. on Magn., Jan. 1992, vol. 28, pp. 335-337. J.M. Rifflet et al., "Status of the Fabrication and Test of the Prototype LHC Lattice Quadrupole Magnets", presented at MT13, Victoria, Canada, 20-24 September 1993.