1. 1 LHC : construction and operation Jörg Wenninger CERN Beams Department / Operation group LNF Spring School 'Bruno Touschek' - May 2010 Part 1: Introduction to accelerator physics LHC magnet and layout Luminosity and interaction regions Injection and filling schemes J. Wenninger LNF Spring School, May 2010

2. Outline 2 The LHC challenges Introduction to magnets and particle focusing LHC magnets and arc layout LHC luminosity and interaction regions Injection and filling schemes Machine protection Incident 19 th Sept. 2008 and consequences LHC operation Part 1 Part 2 J. Wenninger LNF Spring School, May 2010

3. LHC History 3 1982 : First studies for the LHC project 1983 : Z0/W discovered at SPS proton antiproton collider (SppbarS) 1989 : Start of LEP operation (Z/W boson-factory) 1994 : Approval of the LHC by the CERN Council 1996 : Final decision to start the LHC construction 2000 : Last year of LEP operation above 100 GeV 2002 : LEP equipment removed 2003 : Start of LHC installation 2005 : Start of LHC hardware commissioning 2008 : Start of (short) beam commissioning Powering incident on 19th Sept. 2009 : Repair, re-commissioning and beam commissioning 2010 : Start of a 2 year run at 3.5 TeV/beam J. Wenninger LNF Spring School, May 2010

4. 17.03.2010 Der LHC Beschleuniger - DPG - Bonn The Large Hadron Collider LHC 4 CMS, Totem ATLAS, LHCf LHCbLHCb ALICEALICE Lake of Geneva Installed in the 26.7 km LEP tunnel Depth of 70-140 m Control Room LHC ring SPS ring

5. 5 Tunnel circumference 26.7 km, tunnel diameter 3.8 m Depth : ~ 70-140 m – tunnel is inclined by ~ 1.4% J. Wenninger LNF Spring School, May 2010

6. 6 IR6: Beam dumping system IR4: RF + Beam instrumentation IR5:CMS IR1: ATLAS IR8: LHC-B IR2:ALICE Injection ring 2 Injection ring 1 IR3: Momentum collimation (normal conducting magnets) IR7: Betatron collimation (normal conducting magnets) Beam dump blocks LHC Layout  8 arcs.  8 straight sections (LSS), ~ 700 m long.  The beams exchange their positions (inside/outside) in 4 points to ensure that both rings have the same circumference ! J. Wenninger LNF Spring School, May 2010 Beam1 Beam2

7. 7 LHC – yet another collider? The LHC surpasses existing accelerators/colliders in 2 aspects :  The energy of the beam of 7 TeV that is achieved within the size constraints of the existing 26.7 km LEP tunnel. LHC dipole field8.3 T HERA/Tevatron ~4 T  The luminosity of the collider that will reach unprecedented values for a hadron machine: LHC pp ~ 10 34 cm -2 s -1 Tevatron pp3x10 32 cm -2 s -1 SppbarSpp6x10 30 cm -2 s -1 The combination of very high field magnets and very high beam intensities required to reach the luminosity targets makes operation of the LHC a great challenge ! A factor 2 in field A factor 4 in size A factor 30 in luminosity J. Wenninger LNF Spring School, May 2010

8. Luminosity challenges 8 The event rate N for a physics process with cross-section  is proprotional to the collider Luminosity L: To maximize L: Many bunches (k) Many protons per bunch (N) A small beam size  * u = (  *  ) 1/2  * : the beam envelope (optics)  : is the phase space volume occupied by the beam (constant along the ring). k = number of bunches = 2808 N = no. protons per bunch = 1.15×10 11 f = revolution frequency = 11.25 kHz  * x,  * y = beam sizes at collision point (hor./vert.) = 16  m High beam “brillance” N/  (particles per phase space volume)  Injector chain performance ! Small envelope  Strong focusing ! Optics property Beam property J. Wenninger LNF Spring School, May 2010

9. Basics of accelerator physics 9J. Wenninger LNF Spring School, May 2010

10. Accelerator concept 10 Charged particles are accelerated, guided and confined by electromagnetic fields. - Bending:Dipole magnets - Focusing:Quadrupole magnets - Acceleration:RF cavities In synchrotrons, they are ramped together synchronously to match beam energy. - Chromatic aberration:Sextupole magnets J. Wenninger LNF Spring School, May 2010

11. Bending 11 Magnetic rigidity Force Lorentz force LHC: ρ = 2.8 km given by LEP tunnel! → → →→ J. Wenninger LNF Spring School, May 2010 To reach p = 7 TeV/c given a bending radius of  = 2805 m:  Bending field : B = 8.33 Tesla  Superconducting magnets To collide two counter-rotating proton beams, the beams must be in separate vaccum chambers (in the bending sections) with opposite B field direction.  There are actually 2 LHCs and the magnets have a 2-magnets-in-one design!

12. Bending Fields 12 Two-in-one magnet design B field B p p F F force I I I B J. Wenninger LNF Spring School, May 2010

13. Focusing 13 13 N NS S ByBy FxFx FyFy Transverse focusing is achieved with quadrupole magnets, which act on the beam like an optical lens. Linear increase of the magnetic field along the axes (no effect on particles on axis). Focusing in one plane, de-focusing in the other! x y J. Wenninger LNF Spring School, May 2010

14. Accelerator lattice 14 horizontal plane vertical plane Focusing in both planes is achieved by a succession of focusing and de-focusing quadrupole magnets : The FODO structure

15. Alternating gradient lattice 15 s x y One can find an arrangement of quadrupole magnets that provides net focusing in both planes (“strong focusing”). Dipole magnets keep the particles on the circular orbit. Quadrupole magnets focus alternatively in both planes. The lattice effectively constitutes a particle trap! J. Wenninger LNF Spring School, May 2010

16. LHC arc lattice 16  Dipole- und Quadrupol magnets –Provide a stable trajectory for particles with nominal momentum.  Sextupole magnets –Correct the trajectories for off momentum particles (‚chromatic‘ errors).  Multipole-corrector magnets –Sextupole - and decapole corrector magnets at end of dipoles –Used to compensate field imperfections if the dipole magnets. To stabilize trajectories for particles at larger amplitudes – beam lifetime ! One rarely talks about the multi-pole magnets, but they are essential for good machine performance ! J. Wenninger LNF Spring School, May 2010

17. Beam envelope J. Wenninger LNF Spring School, May 201017 Betatron functions in a simple FODO cell QF QD  The focusing structure (mostly defined by the quadrupoles: gradient, length, number, distance) defines the transverse beam envelope.  The function that describes the beam envelope is the so-called ‘  ’-function (betatron function): In the LHC arcs the optics follows a regular pattern – regular FODO structure. In the long straight sections, the betatron function is less regular to fulfill various constraints: injection, collision point focusing… The envelope peaks in the focusing elements ! Horizontal Vertical

18. Beam emittance and beam size J. Wenninger LNF Spring School, May 201018  For an ensemble of particles: The transverse emittance, ε, is the area of the phase-space ellipse. Beam size = projection on X (Y) axis.  The beam size  at any point along the accelerator is given by (neglecting the contribution from energy spread): For unperturbed proton beams, the normalized emittance  n is conserved:  = Lorentz factor The beam size shrinks with energy:

19. Why does the transverse emittance shrink? J. Wenninger LNF Spring School, May 201019  The acceleration is purely longitudinal, i.e the transverse momentum is not affected:  The emittance is nothing but a measure of.  To maintain the focusing strength, all magnetic fields are kept proportional to E (  ), including the quadrupole gradients.  With constant and increasing quadrupole gradients, the transverse excursion of the particles becomes smaller and smaller !

20. LHC beam sizes J. Wenninger LNF Spring School, May 201020  Beta-function at the LHC ARC  Nominal LHC normalized emittance : Energy (GeV)  (nm)  (mm) 4507.21.14 35000.930.41 70000.470.29 Example LHC arc, peak  = 180 m

21. Acceleration 21 s  Acceleration is performed with electric fields fed into Radio-Frequency (RF) cavities. RF cavities are basically resonators tuned to a selected frequency.  To accelerate a proton to 7 TeV, a 7 TV potential must be provided to the beam:  In circular accelerators the acceleration is done in small steps, turn after turn.  At the LHC the acceleration from 450 GeV to 7 TeV lasts ~20 minutes, with an average energy gain of ~0.5 MeV on each turn. J. Wenninger LNF Spring School, May 2010

22. LHC RF system 22  The LHC RF system operates at 400 MHz.  It is composed of 16 superconducting cavities, 8 per beam.  Peak accelerating voltage of 16 MV/beam. For LEP at 104 GeV : 3600 MV/beam ! Synchrotron radiation loss LHC @ 3.5 TeV0.42 keV/turn LHC @ 7 TeV6.7 keV /turn LEP @ 104 GeV~3 GeV /turn The nominal LHC beam radiates a sufficient amount of visible photons to be actually observable ! (total power ~ 0.2 W/m) J. Wenninger LNF Spring School, May 2010

23. Visible protons ! J. Wenninger LNF Spring School, May 201023  Some of the energy radiation by the LHC protons is emitted as visible light. It can be extracted with a set of mirrors to image the beams in real time.  This is a powerful tool to understand the beam size evolution. Protons are very sensitive to perturbations, keeping their emittance small is always a challenge. Flying wire SPS (injector) Synch. light Flying wire LHC

24. Cavities in the tunnel J. Wenninger LNF Spring School, May 201024

25. RF buckets and bunches 25 EE time RF Voltage time LHC bunch spacing = 25 ns = 10 buckets  7.5 m 2.5 ns The particles are trapped in the RF voltage: this gives the bunch structure RMS bunch length 12.8 cm 5.8 cm RMS energy spread 0.031%0.02% 450 GeV 3.5 TeV The particles oscillate back and forth in time/energy RF bucket 2.5 ns J. Wenninger LNF Spring School, May 2010

26. Magnets & Tunnel 26J. Wenninger LNF Spring School, May 2010

27. Superconductivity 27 Bc Tc  The very high DIPOLE field of 8.3 Tesla required to achieve 7 TeV/c can only be obtained with superconducting magnets !  The material determines: Tc critical temperature Bc critical field  The cable production determines: Jc critical current density  Lower temperature  increased current density  higher fields.  Typical for NbTi @ 4.2 K 2000 A/mm2 @ 6T  To reach 8-10 T, the temperature must be lowered to 1.9 K – superfluid Helium ! J. Wenninger LNF Spring School, May 2010

28. The superconducting cable 28  1 mm  6  m Typical value for operation at 8T and 1.9 K: 800 A Rutherford cable width 15 mm A.Verweij J. Wenninger LNF Spring School, May 2010

29. Coils for dipoles 29 Dipole length 15 m I = 11’800 A @ 8.3 T The coils must be aligned very precisely to ensure a good field quality (i.e. ‘pure’ dipole) J. Wenninger LNF Spring School, May 2010

30. 30 Rüdiger Schmidt30 Beam tube Superconducting coil Non-magnetic collars Ferromagnetic iron Steel cylinder for Helium Insulation vacuum Supports Vacuum tank Weight (magnet + cryostat) ~ 30 tons, length 15 m J. Wenninger LNF Spring School, May 2010

31. 31 J. Wenninger - ETHZ - December 2005 31 1232 main dipoles + 3700 multipole corrector magnets (sextupole, octupole, decapole) 392 main quadrupoles + 2500 corrector magnets (dipole, sextupole, octupole) Regular arc: Magnets J. Wenninger LNF Spring School, May 2010

32. 32 J. Wenninger - ETHZ - December 2005 32 Regular arc: Cryogenics Supply and recovery of helium with 26 km long cryogenic distribution line Static bath of superfluid helium at 1.9 K in cooling loops of 110 m length Connection via service module and jumper J. Wenninger LNF Spring School, May 2010

33. 33 J. Wenninger - ETHZ - December 2005 33 Insulation vacuum for the cryogenic distribution line Regular arc: Vacuum Insulation vacuum for the magnet cryostats Beam vacuum for Beam 1 + Beam 2 J. Wenninger LNF Spring School, May 2010

34. Tunnel view (1) J. Wenninger LNF Spring School, May 201034

35. Tunnel view (2) 35J. Wenninger LNF Spring School, May 2010

36. Complex interconnects 36 CERN visit McEwen Many complex connections of super-conducting cable that will be buried in a cryostat once the work is finished. This SC cable carries 12’000 A for the main quadrupole magnets J. Wenninger LNF Spring School, May 2010

37. Magnet cooling scheme 37 Courtesy S. Claudet  He II: super-fluid o Very low viscosity o Very high thermal conductivity J. Wenninger LNF Spring School, May 2010

38. Cryogenics J. Wenninger LNF Spring School, May 201038 8 x 18kW @ 4.5 K 1’800 SC magnets 24 km & 20 kW @ 1.8 K 36’000 t @ 1.9K 130 t He inventory Courtesy S. Claudet Grid power ~32 MW

39. Cool down J. Wenninger LNF Spring School, May 201039 Cool-down time to 1.9 K is nowadays ~4 weeks/sector [sector = 1/8 LHC]

40. Vacuum chamber 40 Beam envel (  4  ) ~ 1.8 mm @ 7 TeV 50 mm 36 mm  The beams circulate in two ultra-high vacuum chambers, P ~ 10 -10 mbar.  A Copper beam screen protects the bore of the magnet from heat deposition due to image currents, synchrotron light etc from the beam.  The beam screen is cooled to T = 4-20 K. Cooling channel (Helium) Beam screen Magnet bore J. Wenninger LNF Spring School, May 2010

41. Luminosity and interaction regions 41J. Wenninger LNF Spring School, May 2010

42. Let us look at the different factors in this formula, and what we can do to maximize L, and what limitations we may encounter !!  f : the revolution frequency is given by the circumference, f=11.246 kHz.  N : the bunch population – N=1.15x10 11 protons - Injectors (brighter beams) - Collective interactions of the particles - Beam encounters  k : the number of bunches – k=2808 - Injectors (more beam) - Collective interactions of the particles - Interaction regions - Beam encounters   * : the size at the collision point –  * y =  * x =16  m - Injectors (brighter beams) - More focusing – stronger quadrupoles Luminosity 42 For k = 1: J. Wenninger LNF Spring School, May 2010

43. Collective (in-)stability 43  The electromagnetic fields of a bunch interact with the vacuum chamber walls (finite resistivity !), cavities, discontinuities etc that it encounters:  The fields act back on the bunch itself or on following bunches.  Since the fields induced by of a bunch increase with bunch intensity, the bunches may become COLLECTIVELY unstable beyond a certain intensity, leading to poor lifetime or massive looses intensity loss.  Such effects can be very strong in the LHC injectors, and they will also affect the LHC – in particular because we have a lot of carbon collimators (see later) that have a very bad influence on beam stability !  limits the intensity per bunch and per beam ! J. Wenninger LNF Spring School, May 2010

44. ‘Beam-beam’ interaction 44 Quadrupole lens Beam(-beam) lens  When a particle of one beam encounters the opposing beam at the collision point, it senses the fields of the opposing beam.  Due to the typically Gaussian shape of the beams in the transverse direction, the field (force) on this particle is non-linear, in particular at large amplitudes. focal length depends on amplitude !  The effect of the non-linear fields can become so strong (when the beams are intense) that large amplitude particles become unstable and are lost from the machine:  poor lifetime  background THE INTERACTION OF THE BEAMS SETS A LIMIT ON THE BUNCH INTENSITY! J. Wenninger LNF Spring School, May 2010

45. From arc to collision point 45 Fits through the hole of a needle! ARC cells CMS collision point  Collision point size @ 7 TeV,  * = 0.5 m (=  -function at the collision point): CMS & ATLAS : 16  m  Collision point size @ 3.5 TeV,  * = 2 m: All points : 45  m J. Wenninger LNF Spring School, May 2010

46. Limits to  * 46  The more one squeezes the beam at the IP (smaller  *) the larger it becomes in the surrounding quadrupoles (‘triplets’): J. Wenninger LNF Spring School, May 2010 Small size Huge size !! Smaller the size at IP:  Larger divergence (phase space conservation !)  Faster beam size growth in the space from IP to first quadrupole ! Aperture in the ‘triplet’ quadrupoles around the IR limits the focusing !

47. Combining the beams for collisions 47  The 2 LHC beams must be brought together to collide.  Over ~260 m, the beams circulate in the same vacuum chamber. They are ~120 long distance beam encounters in total in the 4 IRs. J. Wenninger LNF Spring School, May 2010

48. Crossing angles 48 IP  Since every collision adds to our ‘Beam-beam budget’ we must avoid un-necessary direct beam encounters where the beams share a common vacuum: COLLIDE WITH A CROSSING ANGLE IN ONE PLANE !  There is a price to pay - a reduction of the luminosity due to the finite bunch length and the non-head on collisions: L reduction of ~17% Crossing planes & angles ATLAS Vertical 280  rad CMS Horizontal 280  rad LHCb Horizontal 300  rad ALICE Vertical 400  rad 7.5 m J. Wenninger LNF Spring School, May 2010

49. Separation and crossing : example of ATLAS 49 Horizontal plane: the beams are combined and then separated Vertical plane: the beams are deflected to produce a crossing angle at the IP ~ 7 mm 194 mm ATLAS IP Not to scale ! ~ 260 m Common vacuum chamber J. Wenninger LNF Spring School, May 2010

50. Tevatron 50J. Wenninger LNF Spring School, May 2010

51. CDF D0 Tevatron

52. Tevatron I 52  The Tevatron is presently the ‘energy frontier’ collider in operation at FNAL, with a beam energy of 980 GeV and a size of ~ ¼ LHC (about same size than SPS).  It is the first super-conducting collider ever build.  It collides proton and anti-proton bunches that circulate in opposite directions in the SAME vacuum chamber.  One of the problems at the TEVATRON are the long-distance encounters of the bunches in the arc sections. A complicated separation scheme with electrostatic elements has to be used: Tricky to operate !! E E J. Wenninger LNF Spring School, May 2010

53. Tevatron II 53  The Tevatron has undergone a number of remarkable upgrades and it presently collides 36 proton with 36 anti-proton bunches (k=36), with bunch populations (N) similar to the ones of the LHC (but there are always fewer anti-protons !).  Compare LHC and Tevatron: f Tevatron  4 f LHC Tevatron gets a factor 4 ‘for free’ due to ring size !! k LHC  100 k Tevatron L LHC  30 L Tevatron N 2 /(  x  y ) ~ equal Luminosity gain of LHC comes basically from the number of bunches (k) !! J. Wenninger LNF Spring School, May 2010

54. Injection and injector complex 54J. Wenninger LNF Spring School, May 2010

55. 55 Top energy/GeV Circumference/m Linac 0.05 30 PSB 1.4 157 CPS 26 628 = 4 PSB SPS 450 6’911 = 11 x PS LHC 7000 26’657 = 27/7 x SPS LEIR CPS SPS Booster LINACS LHC 3 4 5 6 7 8 1 2 Ions protons Beam 1 Beam 2 TI8 TI2 Note the energy gain/machine of 10 to 20. The gain is typical for the useful range of magnets.

56. J. Wenninger LNF Spring School, May 2010 Principle of injector cycling 56 PS Booster PS SPS time B field B The beams are handed from one accel. to the next or used for its own customers ! Beam transfer SPS waits at injection to be filled by PS SPS ramp SPS top energy, prepare for transfer …

57. J. Wenninger LNF Spring School, May 2010 Principle of injection (and extraction) 57 Septum magnet Kicker magnet Injected beam Circulating beam B B  A septum dipole magnet (with thin coil) is used to bring the injected beam close to the circulating beam.  A fast pulsing dipole magnet (‘kicker’) is fired synchronously with the arrival of the injected beam: deflects the injected beam onto the circulating beam path.  ‘Stack’ the injected beams one behind the other.  At the LHC the septum deflects in the horizontal plane, the kicker in the vertical plane (to fit to the geometry of the tunnels).  Extraction is identical, but the process is reversed ! Kicker B-field time Injected beam Circulating beam

58. J. Wenninger LNF Spring School, May 2010 58 Linac2 Delivered beam current:~150mA Beam energy:90 keV (source) → 750 keV (RFQ) → 50 MeV Repetition rate:1 Hz Radio-frequency system:202 MHz Radio-frequency quadrupole (RFQ) Alvarez’s drift-tube

59. J. Wenninger LNF Spring School, May 2010 59 PS Booster  Constructed in the 70ies to increase the intensity into the PS  Made of four stacked rings  Acceleration to E kin =1.4 GeV  Intensities > 10 13 protons per ring.

60. J. Wenninger LNF Spring School, May 2010 60 Filling the PS with LHC beams x 3  Rings 2,3 & 4 are filled with 2 bunches per ring.  The 6 bunches are transferred to the PS.

61. J. Wenninger LNF Spring School, May 2010 61 Proton Synchrotron Recently celebrated its first 50 years!!

62. J. Wenninger LNF Spring School, May 2010 Bunch Splitting at the PS 62  The bunch splitting in the PS is probably the most delicate manipulation for the production of LHC beams – multiple RF systems with different frequencies: from 6 injected to 72 extracted bunches  The quality of the splitting is critical for the LHC (uniform intensity in all bunches…).

63. J. Wenninger LNF Spring School, May 2010 63 Super-Proton Synchrotron

64. J. Wenninger LNF Spring School, May 2010 64 Courtesy of J. Uythoven SPS-to-LHC transfer lines

65. Collision schemes 65  The 400 MHz RF system provides 35’640 possible bunch positions (buckets) at a distance of 2.5 ns along the LHC circumference.  A priori any of those positions could be filled with a bunch…  The smallest bunch-to-bunch distance is fixed to 25 ns, which is also the nominal distance: max. number of bunches is 3564.  In practice there are fewer bunches because holes must be provided for the fast pulsed magnets (kickers) used for injection and dump.  But the LHC and its injectors are very flexible and can operate with many bunch patterns: from isolated bunches to trains. 2.5 ns 25 ns = bunch position= filled position … J. Wenninger LNF Spring School, May 2010

66. Collision point symmetry 66 LHC CMS Atlas Alice LHCb displaced by c x 37.5 ns = 11.25 m LHCb = collision point  ATLAS, ALICE and CMS are positioned on the LEP symmetry axis (8 fold sym.)  LHCb is displaced from the symmetry axis by 11.25 m > 37.5 ns.  For filling patterns with many bunches this is not an issue, but it becomes a bit tricky with few bunches. Symmetry axis J. Wenninger LNF Spring School, May 2010

67. Filling pattern example: 1x1 67 LHC CMS Atlas Alice LHCb  With 1 bunch/beam, there are 2 collision points at opposite sides of the ring.  Depending on their position along the circumference, the 2 bunches can be made to collide: in ATLAS and CMS, OR in ALICE, OR in LHCb, but never in all experiments at the same time !! J. Wenninger LNF Spring School, May 2010

68. (Some) LHC filling patterns 68 SchemaNominal bunch distance (ns) No. bunchesComment 43x43202543No crossing angle required 156x156525156No crossing angle required 25 ns252808Nominal p filling 50 ns501404 2010-2011 run target Ion nominal100592Nominal ion filling Ion early135062No crossing angle required  With 43x43 and 156x156, some bunches are displaced (distance  nominal) to balance the ALICE and LHCb luminosities.  In the multi-bunch schemes (25, 50, 100 ns) there are larger gaps to accommodate fast injection magnets (‘kickers’) rise times.  There is always a ≥ 3  s long particle free gap for the beam dump kicker. J. Wenninger LNF Spring School, May 2010

69. Nominal filling pattern 69  The nominal pattern consists of 39 groups of 72 bunches (spaced by 25 ns), with variable spacing to accommodate the rise times of the injection and extraction magnets (‘kickers’). 11    72 bunches b=bunch, e=empty J. Wenninger LNF Spring School, May 2010

70. Spare slides 70J. Wenninger LNF Spring School, May 2010

71. 71 PS - bunch splitting

72. J. Wenninger LNF Spring School, May 2010 Injection elements 72 12 mrad 0.8 mrad TED From the LHC Page1

73. J. Wenninger LNF Spring School, May 2010 Role of the TDI collimator 73 The TDI is one of the key injection protection collimators: Protects the machine in case of (1) missing kicks on injected beam and (2) asynchronous kicker firing on the circulating beam. It must be closed around the circulating beam trajectory when the kicker is ON.