18 10 2010 LHC Status ATLAS France LHC

18. 10. 2010 LHC Status - ATLAS France LHC achievements, status and plans J. Wenninger BE OP group 1

Outline Machine protection (driven) commissioning 18. 10. 2010 LHC Status - ATLAS France Setting up for trains Train operation The last week of protons Ion run 2

18. 10. 2010 LHC Status - ATLAS France The target 32 >10 -2 -1 cm s -1 fb for 2011) (>1 3

Machine protection driven commissioning q MP phase 1: low intensity MP commissioning. o Commissioning of the protection systems. o Low intensity single bunch commissioning of the systems, including beam tests (manually triggered failures). 18. 10. 2010 LHC Status - ATLAS France q q MP phase 2: MP running in with gradual intensity increase. o Intensity increase in steps, factor 2 – 4, up to ~ MJ stored energy. o Stability run of a few weeks around 1 -3 MJ. MP Phase 3: intensity increase to 10’s MJ regime. o Intensity increase in steps of 2 -3 MJ ( 1 TEVATRON beam). o Initially o With planned one step every 1 -2 weeks. the good MPS performance, agreed to reduce the step to: Ø 3 fills and 20 hours of stable beams. This span is also driven by operational considerations, as conditions can change drastically with large number of trains. 4

Machine protection: some statistics q More than 270 dumps above injection energy. q Between March and end of August over 70% of the beams above injection energy were dumped by the MPS ! q The MPS has caught all events perfectly, even some weird operational mistakes. 18. 10. 2010 LHC Status - ATLAS France q Remaining worries arise from COMBINED failures, most scenarios involve an injection or a dump error (asynchronous dump). o That is why we carefully track anomalies in the present period of rapid intensity increase. Beam dumps > 450 Ge. V ‘False’ dumps Beam monitoring HW surveillance Operator MPS test 5

UFOs 7 th July we observed the first occurrence of fast beam loss events in the super-conducting regions of the ring: 18. 10. 2010 LHC Status - ATLAS France q On o Beam loss at a SC magnet. o Fast loss over ~0. 5 -2 ms. o Most events during stable beams: no power converter changes, orbit rock-stable, no lifetime issue before the event… o Loss at regions of very large aperture > 40 beam sigma (collimators between 6 and 15 sigma). q The hypothesis quickly emerged that it is not the beam that moves to the aperture, but rather the opposite ! o ‘Dust’ particles ‘falling’ into the beam, estimated size ~100 mm think Carbon-type object. o Two events in perfect coincidence (time & space) with TOTEM roman pot movements make this hypothesis rather convincing. 6

Example of a 152 bunches UFO Beam loss monitor post-mortem LHCb LHC Status - ATLAS France IR 7 Arc IR 1 s 0. 5 ms Time evolution of loss 1 bin = 40 ms 18. 10. 2010 Dump trigger 7

A worrying correlation… 18. 10. 2010 LHC Status - ATLAS France T. Baer One of the reasons why we want to observe the beams for 20 hours before increasing intensity! 8

UFO near threshold About 50% of the UFOs lead to dumps while the loss is decaying… IR 1 ALICE IR 3 18. 10. 2010 LHC Status - ATLAS France Arc Dump trigger The dump is triggered on the loss integral ! Time evolution of loss 1 bin = 40 ms 9

UFO and BLM thresholds q 2 weeks ago we had accumulated 12 UFO events ( beam dumps). q But o there was no quench – the BLMs always triggered first. In many cases the signal was just above threshold of the BLMs… we decided 2 weeks ago to increase the dump thresholds of most BLMs at super-conducting elements by a factor 3. LHC Status - ATLAS France q Therefore o Initial thresholds were set to 30% of quench level – we are now essentially at the estimated quench level. o New models of the magnet cooling indicate more margin than initially estimated, and the BLM response is tuned on a different loss scenario. q Since then there was only one UFO dump in the IR 8 triplet, but from LHCb BCM (machine BLMs far below threshold). q The rate of UFOs (the ones that dump + the ones that are below dump threshold) increases with intensity: 18. 10. 2010 o With 250 bunches there are ~0. 8 UFOs/hour 10

Outline Machine protection commissioning 18. 10. 2010 LHC Status - ATLAS France Setting up for trains Train operation The last week of protons Ion run 11

Crossing angles q Until o end of August the large bunch spacing did not require Xing angles. A Xing angle was used in stable beams to avoid some parasitic encounters from the ‘democratic’ filling scheme. q To operate with closely spaced bunches (trains) a Xing angle is required to avoid parasitic encounters in the common vacuum chamber. 18. 10. 2010 LHC Status - ATLAS France o Xing angle value is a compromised between maximizing aperture and minimizing beam-beam effects. ‘long-range beam-beam’ For 150 ns spacing, the first encounter is at 22. 5 m from the IP ½ crossing angles (sign B 1) IR Plane Injection & ramp Squeeze & collisions 1 V -170 -100 2 V 170 110 5 H 170 100 8 H -170 -100 12

Separation and crossing in ATLAS Horizontal plane: the beams are combined and then separated 194 mm ATLAS IP ~ 260 m 18. 10. 2010 LHC Status - ATLAS France Common vacuum chamber During injection, ramp & squeeze there is an additional parallel separation of 2 mm in the horizontal plane ! Vertical plane: the beams are deflected to produce a crossing angle at the IP Not to scale ! ~ 7 mm 13

Crossing angles : injection ATLAS ALICE CMS LHCb B 1 Hor 10 mm LHC Status - ATLAS France B 1 Vert B 2 Hor 18. 10. 2010 B 2 Vert 14

150 ns beam structure PS produces trains of m bunches (spacing 150 ns). o m = 8 or 12. o m is fixed for a given filling sequence. q SPS assembles n PS trains of m bunches o n = 1, 2, 3 or 4. So far we use only 1 and 2. o train spacing defined at injection to SPS, presently 300 ns. o n may change for each cycle. q LHC requests p SPS train groups o n may vary from one injection to the next. 18. 10. 2010 LHC Status - ATLAS France q Very flexible – can produce a large variety of patterns in the LHC m SPS m m or m m m m Circumference 15

Overview of proposed 150 ns “structure” n PS trains of each m bunches q N = n x m = tot nr of bunches q M = N – n/4 = nr of collisions in each of IP 1, IP 5 and IP 8 q example (N=48): q one train of m bunches -3 shifted by -3 x 25 ns slots 18. 10. 2010 LHC Status - ATLAS France n = 12 & m = 4 => M = 45 n = 4 & m = 12 => M = 47 – NB: m=12 is max for 150 ns +3 -3 Add k bunches for ALICE, typically k = ~ N/16 & k ~16 q NB: the N bunches will give parasitic encounters at q -3 -3 m k IP 2: +/-11. 25 m, +/-33. 75 m, … IP 1, 5, 8: +/-22. 5 m, +/-45 m, … m m 16

A possible evolution 18. 10. 2010 LHC Status - ATLAS France 2/3 equalitarian 4 -fold symmetric private IP 2 bunches per LHC injections x bunches ( per beam) Total bunches per LHC beam LHC inj. Collisions in IP 1, 5, 8 Collisio ns in IP 2 8 3 x 8 24 3 16 16 8 4 x 8 +3 x 8 56 7 47 16 8 or : 8 & 8+8 12 x 8 + 1 x 8 or : 4 x 8+4 x(8+8) + 1 x 8 104 13 9 93 8 8 or : 8 & 8+8 16 x 8 + 3 x 8 or : 8 x(8+8) + 3 x 8 152 19 11 140 16 8 & 8+8 4 x 8+10 x(8+8) + 1 x 8 200 15 186 8 8 & 8+8 or : 8 & 8+8+8 4 x 8+12 x(8+8) + 3 x 8 4 x 8+8 x(8+8+8) + 3 x 8 248 19 15 233 16 etc… etc 312 19 290 16 M. Ferro-Luzzi 17

18. 10. 2010 LHC Status - ATLAS France 312 bunches Beam 1 18

LHC aperture and collimator settings 18. 10. 2010 12. 0 s 15. 0 s ± 8. 5 σ Primary (robust) Secondary (robust) Absorber (W metal) Tertiary (W metal) TCP TCSG TCLA TCT ARC b* 3. 5 m IP & Triplets Beam halo ➠ ARC IP & Triplets 15 s ARC 8. 5 s 12. 5 s ARC 6. 7 s Beam halo ➠ 5. 7 s 3. 5 Te. V LHC Status - ATLAS France Injection 5. 7 s ± 18 σ Courtesy S. Redaelli 19

Machine protection setup/tests for trains q Collimator setup. o IR 7 (betatron [transverse] cleaning) and IR 3 (off-momentum cleaning) untouched wrt pre-train period. Exception for beam 2 in IR 3 due to a collimator hierarchy issue (relative retraction). o Tertiary collimator setup at injection, at 3. 5 Te. V un-squeezed and for collisions. Positions are interpolated in ramp and for the squeeze steps. LHC Status - ATLAS France o 8 q collimators per beam (2 per IR and per beam). Collimator and dump setup validation. o Large losses induced by crossing resonance to verify collimator setup (loss maps). o De-bunch beam to fill abort gap and trigger dump to verify protection by absorbers (and collimators). o Very time (beam) consuming exercise. Minimum of 10 cycles. 18. 10. 2010 o Used up a large part of the train setup. 20

Good setup - hierarchy respected IP 6 IP 7 IP 8 an Ts cle s LA TC TC Ps IP 7 SG s TC Ts TC β Du m p cle an ing The collimator hierarchy is verified with dedicated loss maps induced by artificially high loss rates: record beam losses around the ring while crossing betatron resonances. TC LHC Status - ATLAS France Beam 1 18. 10. 2010 IP 5 ing IP 4 /p TC IP 3 Δp IP 2 Ts IP 1 Normal cond. magnet cleaning insertion 21

Cleaning inefficiency evolution Courtesy D. Wollman LHC Status - ATLAS France Example of the loss leakage on the horizontal tertiary collimators (sum over all IRs) over 2 months. 18. 10. 2010 q Leakage q Stable into cold aperture around 2 e-4. over 2 months. 22

Dump protection and b* q b* at the IR is limited by the beam size in the triplet. o Beam size in triplet 1/ b*. o b* reduction is limited by the aperture in the triplet and the tolerances for collimators, protection devices (dump) and orbit movements. are not too aggressive to avoid constant OP interruptions for alignment verification and checks. 18. 10. 2010 10. 5 s 15 s b* 3. 5 m IP & Triplets LHC Status - ATLAS France q Tolerances Beam halo ➠ Dump protection TCDQ Tertiary (W metal) TCT ± 18 σ = orbit tolerances 1 -2 s s 0. 3 mm at TCT. 23

18. 10. 2010 LHC Status - ATLAS France What is an Asynchronous Beam Dump? TCSG 24

What is an Asynchronous Beam Dump? LHC Status - ATLAS France Abort gap =3 ms Bunched beam and perfect synchronization with RF Bunches TCSG 18. 10. 2010 Empty region 25

What is an Asynchronous Beam Dump? 18. 10. 2010 LHC Status - ATLAS France Abort gap =3 ms Loss of synchronization with RF or RF off Unbunched beam filling the abort gap Unbunched beam Unbunched bean TCSG 26

Protection in case of an asynchronous beam dump LHC Status - ATLAS France Estimated occurrence : at least once per year, 0 events up to now! 18. 10. 2010 TCDQ = 6 m long one-sided collimator TCSG = 1 m long two-sided collimator TCSG to TCDQ + TCSG protect downstream SC magnet (Q 4) 27

Asynchronous dump test q De-bunch a low intensity beam and fire the dump. q Verify that losses are contained in dump and collimation regions. q This is re-checked for EVERY dump using the small amount of beam present in the abort gap. ALICE CMS dump betatron cleaning LHCb 18. 10. 2010 LHC Status - ATLAS France ATLAS dp/p cleaning 28

Outline Machine protection commissioning 18. 10. 2010 LHC Status - ATLAS France Setting up for trains Train operation The last week of protons Ion run 29

Injection q Injection is becoming more critical: o o 18. 10. 2010 LHC Status - ATLAS France o q Injected beams have now some damage potential. Losses at injection collimators become more critical – LHCb BCMs can tell some stories… De-bunching of the already circulating beam can lead to beam dumps during injection. o Abort gap cleaning by exciting particles in the gap ( collimators) may soon become mandatory. o Frequent cause of dumps at injection. Injection was going rather well until an aperture restriction was suddenly observed some 10 days ago near the injection septum of beam 1. o Led to excessive losses and beam dumps during injection. 30

Injection region The injection septum (MSI) bends the injected beam parallel to the circulating beam in the horizontal plane. q The injection kicker (MKI) is deflecting the injected beam into the plane of the LHC (vertical deflection). q The TDI injection absorber is protecting the machine from damage in case of MKI ‘failures’ (not rare !!). o The TDI also intercepts the circulating bunch during over-injecting. 18. 10. 2010 LHC Status - ATLAS France q 31

Injection area investigations at the MSI q Suspected aperture restriction at the transition of injection septa magnets (MSIB-MSIA). 18. 10. 2010 LHC Status - ATLAS France Beam 1 : circulating & injected beam Beam 2 32

Injection area investigations at the MSI There is a problem with the RF fingers at the transition. Not clear how did it get worse suddenly. q Must be repaired >> technical stop advanced to this Tuesday (18. 10). 18. 10. 2010 LHC Status - ATLAS France q Circulating b 1 Injected b 1 33

Injection septum loss mapping Scan of losses versus beam position (injected beam) clearly show obstruction. Steered beam down and towards left sort of OK again. q Unfortunately this lasted only about one day, then the beam to be moved further down and more to the left (last Saturday). 18. 10. 2010 LHC Status - ATLAS France q High losses old new No losses High losses -4 mm Vertical +1 mm Horizontal old new No losses 34

Ramp rate At the start of the run the ramp rate had to be limited to 2 A/s (1. 2 Ge. V/s) for magnet protection reasons. o Ramp duration 0. 45 -3. 5 Te. V: 46 minutes q Since mid-July the rate for down-ramps and magnet pre-cycles (magnetic history reset) was increased to nominal value of 10 A/s (6 Ge. V/s). q Ramp speed with beam now to 10 A/s (6 Ge. V/s). o Pure ramp duration 0. 45 -3. 5 Te. V: 16 minutes. LHC Status - ATLAS France q 2 A/s 3500 Ge. V 10 A/s 18. 10. 2010 450 Ge. V 35

From injection to collisions TCT = Tertiary Collimator 18. 10. 2010 LHC Status - ATLAS France 100/110 q 4 stops on the flat top for feedbacks and collimators. 36

Emittance q Injected emittance can be reduced to less than 1. 5 mm – almost a factor 3 below nominal value (3. 5 mm). o 18. 10. 2010 LHC Status - ATLAS France o q Emittances for 50, 75 and 150 ns are lower than for 25 ns (injectors). Rather strong beam-beam was observed in one fill with emittances at 3. 5 Te. V below 2 mm. Losses lead to beam dump – curable since thresholds were too low on some normal conducting elements in IR 7. Presently we aim for/inject beams with emittances of ~2 mm. o Emittance increase to collisions under control (transverse damper) – routinely start collisions with emittances around 2. 5 mm (better for B 1 than for B 2). o Since Luminosity ~ 1/emittance ~30% gain of luminosity. 37

The good news : beam-beam, lifetimes q Beam current lifetimes in collisions now ≥ 25 hours. No or very small lifetime dips when bringing beams into collisions. o Excellent news is that the beam-beam effects (both head-on and long -long range) seem much less critical than anticipated. >> Can think of more bunch current, smaller emittance ! 18. 10. 2010 LHC Status - ATLAS France o q Luminosity decay dominated o Current decay ~30 -40% o Emittance growth ~60 -70% by emittance growth, 38

High intensity issue: vacuum activity IR 1 q Vacuum pressure increase observed around the 4 experiments since LHC switched to train operation – issue becomes more and more critical as the intensity increases. 18. 10. 2010 LHC Status - ATLAS France q Local pressure bump around 60 m from the IP. o In the region of an uncoated segment of vacuum chamber at the warm-cold transition (after triplet). q Pressure o rise driven by the presence of both beams. No significant effect with a single beam. q Signs of cleaning by beam, and dependence on intensity (bunch/total). Suspicion that this might be electron clouds ! 39

Intensity and vacuum over 2 weeks 18. 10. 2010 LHC Status - ATLAS France 3. e 13 p+ 3. 5· 10 -7 mbar 40

Electron clouds … affect high intensity beams with positive charge and closely spaced bunches. q Electrons are generated at the vacuum chamber surface by beam impact, photons… q If the probability to emit secondary e- is high (enough), more e- are produced and accelerated by the field of a following bunch(es). Multiplication starts… o LHC Status - ATLAS France q Electron energies are in the 10 -100 e. V range. The cloud of e- can drive the beam unstable, and at the LHC, overload the cryogenic system by the heat deposited on the chamber walls ! The cloud can ‘cure itself’ because the impact of all those electrons cleans the surface, reduces the electron emission probability and eventually the cloud disappears ! Bunch N+2 accelerates the e-, more multiplication… Bunch N+1 accelerates the e-, multiplication at impact Bunch N liberates an e- e- ++++++ N+2 e- ++++++ N+1 ++++++ N 18. 10. 2010 e- 41

Electron cloud q In principle no electron cloud expected with 150 ns beams. Room temperature vacuum chambers are coated with a NEG that kills/reduces the likelihood of electron clouds. o But not the few pieces at the transition after the triplet… o fact that the pressure increases with two beams, is close to a parasitic encounter and in a region without coating makes e-cloud a possible a candidate… 18. 10. 2010 LHC Status - ATLAS France q The vacuum group installed small solenoids around the chamber in IR 1 during a cryo stop to test the hypothesis of electron cloud build-up (standard cure). The presence of e-clouds could be demonstrated: with solenoid the vacuum is orders of magnitude better. o Brought a significant improvement of the vacuum and background for ATLAS. o The fact that the effect is much less visible in CMS could be due to the CMS solenoid STRAY field (to be confirmed). o 42

Solenoids between DFBX and D 1 in IR 1 L 18. 10. 2010 LHC Status - ATLAS France M. Jimenez 43

Solenoid A 4 R 1 - ON 18. 10. 2010 LHC Status - ATLAS France Solenoid A 4 L 1 - ON 44

Peak luminosity performance Peak luminosity = 1. 3 1032 cm-2 s-1 18. 10. 2010 LHC Status - ATLAS France (312 bunches/beam, 295 colliding bunches) 45

Stored energy Peak ~20 MJ (TEVATRON ~2 MJ) 18. 10. 2010 LHC Status - ATLAS France The present beam intensity will slice open a vacuum chamber even at injection. 46

Integrated luminosity ~23 pb-1 (17. 10. 2010) 18. 10. 2010 LHC Status - ATLAS France A fill with 250 bunches delivers 2 pb-1 in ~7 -8 hours 47

Outline Machine protection commissioning 18. 10. 2010 LHC Status - ATLAS France Setting up for trains Train operation The last week of protons Ion run 48

Plans 18. 10. 2010 LHC Status - ATLAS France Advanced technical stop 104 200 296 152 248 344 312 q We are only at 312 bunches due to the injection problems. But with the smaller emittance we exceeded 1032. q If in the next 2 weeks, we make 6 fills per week it is possible to collect ~40 pb-1 more since for ≥ 350 b fills will deliver ~3 -4 pb-1 / 12 hour. 49

The next 2 weeks q Continue intensity increase towards ~400+ bunches. q Determine o Consolidate recent aperture measurements at injection confirming that we have more space than ‘designed’ – much better orbit, better alignment. o May reduce b* to 2 m in 2011 while keeping the same margins/tolerances. LHC Status - ATLAS France q Test 18. 10. 2010 limits on b* for 2011: physics fills with 50 ns trains. o Start with ~50 bunches, then increase in few steps to 300+ bunches. o Precious experience if we want to push intensity further and anticipate train effects (vacuum…). q Quench q Test tests. b* = 90 m optics (TOTEM, total cross section). q Feedback improvements for tune and orbit (could also be done during ion period). 50

Outline Machine protection commissioning 18. 10. 2010 LHC Status - ATLAS France Setting up for trains Train operation The last week of protons Ion run 51

Early ion scheme 18. 10. 2010 LHC Status - ATLAS France J. Jowett q Peak luminosity ~1025 cm-2 s-1 , integrated L ~ few mb. q Bunch charge corresponds ~6 E 9 – equivalent to a pilot proton bunch. o Visible on beam position system only down to ~2 -3 E 9 (with good quality !) 52

Ions are like protons at the LHC q At the LHC the difference between Pb ions and protons is very small because of the high energy. o Same q orbit, tunes, optics, geometrical emittance… Main difference between ions and protons is the RF frequency (revolution frequency): 18. 10. 2010 LHC Status - ATLAS France o RF frequency swing 5 k. Hz instead of 800 Hz (wrt 400 MHz). o Difference in frequency is vanishing at 3. 5 Te. V : ~ 10 Hz. q It is possible to reuse almost all proton settings for ions ! 53

Ions are almost like protons at the LHC Some changes wrt protons: q We may remove the crossing angle – partly your choice. o ALICE wants to run with 0 effective angle. They need the external (machine) crossing angle bump to compensate their spectrometer crossing angle. q Collimation does not work well for ions due to fragmentation. 18. 10. 2010 LHC Status - ATLAS France o Basically a single stage system. Details of settings not finalized yet. o Collimation setup must be completely redone, but it will be simpler. o ALICE has a problem with the vertical TCT that shadows spectator neutrons in their ZDC. o q Request to retract TCT as no fast loss (asynch. dump) in vertical plane. This will probably be accepted. MPS issues are minor – no special setup. o But we will keep our eyes open ! 54

18. 10. 2010 LHC Status - ATLAS France A 128 bunches/ring scheme Basic building block: – Four bunches spaced by 500 ns – Injection kicker: 975 ns q 9 such blocks fill exactly one LHC quarter. . q 55

Commissioning of ions Commissioning should be feasible in one week. 3 weeks of good running… q Intensity increase should be fast – over a couple of fills. q Clearly no serious MP issue here. o Possibly driven by operational aspects (injection etc). 18. 10. 2010 LHC Status - ATLAS France o 56

Summary q. L = 1032 cm-2 s-1 achieved ! q So o far possible obstacles have been solved: UFOs – the BLM threshold increase helped – we were bloody lucky! >> But at 7 Te. V this could become a serious issue !! 18. 10. 2010 LHC Status - ATLAS France o Vacuum – solenoids help against e-cloud. q Major test in the last 2 weeks: 50 ns operation test. q Injection q Switch problem to be fixed during the technical stop. over to ion should be fast – 1 week. o Some details to be finalized (Xing angle or not). o Possibility to increase bunches from 62 to 140 is on the table. 57

Outlook 2011 18. 10. 2010 LHC Status - ATLAS France Possible gains in luminosity: q 50 ns trains x 3 q b* = 2 m x 1. 7 q Lower emittance – already used (x 1. 3) q Bunch charge to 1. 3 x 1011 p x 1. 4 Total x 7 ! Total intensity may be limited by collimation or lifetime considerations – we may not gain the full factor ! 58

18. 10. 2010 LHC Status - ATLAS France Spares 59

Lifetime versus crossing angle 18. 10. 2010 LHC Status - ATLAS France d mra 100 d m ra 20 d mra 30 d mra 40 d mra 50 d m ra 60 d mra 70 d m ra 80 mra 90 170 100 mrad d Test with 3 batches of 8 bunches each, spacing 150 ns at injection up to 6 long range interactions per bunch. At injection the minimum Xing angle with 150 ns trains is 100 mrad. It was decided to use the nominal value of 170 mrad to gain experience. Aperture was found not to be a problem. 60

Beam loss in collisions – 104 bunches Luminosity driven ! ALICE dp/p cleaning CMS betatron cleaning LHCb 18. 10. 2010 LHC Status - ATLAS France ATLAS One can clearly determine the collimation efficiency online (not IRs !) 61

Example of global aperture measurement 18. 10. 2010 LHC Status - ATLAS France Beam 2, V plane Q 4. R 6 TCPV = 15 sig Q 4. R 6 TCPV = 13 sig Q 4. R 6 TCPV = 12 sig Loss at Q 4 higher than at TCP Loss at Q 4 about same as TCP Loss at Q 4 lower than at TCP 62

Measured 450 Ge. V Aperture 18. 10. 2010 LHC Status - ATLAS France Beam / plane Limiting element Aperture [s] Beam 1 H Q 6. R 2 12. 5 Beam 1 V Q 4. L 6 13. 5 Beam 2 H Q 5. R 6 14. 0 Beam 2 V Q 4. R 6 13. 0 No aperture bottlenecks in triplets. q On momentum aperture ≥ 3 sigma larger than ‘predicted’ Main reason: much better orbit (≥ 2 sigma gain) q 63