LHC status and perspectives

LHC status and perspectives LHC performance in 2011 - PAF - La Londe J. Wenninger CERN Beams Department Operation group / LHC

Outline Proton operation 2011 High intensity issues Ions 2011 Outlook for 2012 LHC performance in 2011 - PAF - La Londe

Luminosity Recall the formula for the luminosity (head-on collision): (Round beams) LHC performance in 2011 - PAF - La Londe • f is the revolution frequency (11.25 kHz), g = E/m, • k is the number of colliding bunch pairs, • N is the bunch population, • F is a geometric (loss) factor (≤ 1) from the crossing angle, F  0.95 in 2011, • e is the normalized emittance, b* the betatron (envelope) function at the IP, • s is the beam size at IP:

Injector beams • 50 ns spacing is used as operationally since the successful vacuum condition for electron clouds (‘beam scrubbing’) in April. • Progressive increase of the bunch number  June. • The 50 ns beam has the highest luminosity potential. • 25 ns beam was anyhow not ready in the LHC ( later). LHC beam parameters (SPS extraction) LHC performance in 2011 - PAF - La Londe Limit Present param.

Limits on b* • b* is constrained by aperture and beam size in the triplet quadrupoles, space is needed for: • the beam envelope (12sfor tertiary collimators TCT), • a 2s margin from TCT to triplet , • the crossing angle –separation of the beams at the parasitic encounters. LHC performance in 2011 - PAF - La Londe TCT TCT Triplet Triplet ≥14s 12s ~ 7 mm

Limits on b* • Interpolating aperture measurements performed at INJECTION in 2010/11, and taking into account the 2010 experience lead to select: • Refined aperture measurements at 3.5 TeV end of August paved the way for smaller b*: b* = 1.5 m for startup 2011 LHC performance in 2011 - PAF - La Londe b* = 1 m after August technical stop (commissioned in less than a week)

From 2010 to 2011 • Main changes in 2011: • No change of the beam energy: 3.5 TeV. • Reduction of b* better understanding of the triplet aperture. • Faster ramp, faster squeeze. • 50 ns bunch spacing. LHC performance in 2011 - PAF - La Londe

Luminosity 2011 Peak luminosity b* 1 m 3.3×1033 cm-2s-1 1380 bunches Reduce e, increase N LHC performance in 2011 - PAF - La Londe 50 ns increase k LHCb luminosity limited to ~3.5×1032 cm-2s-1 by leveling (beams collide with transverse offsets) 75 ns

Luminosity and energy • The reach in b* depends on beam size in the triplets where we have the aperture limits: (Approximate) luminosity scaling with energy LHC performance in 2011 - PAF - La Londe Scaling the 3.5 TeV performance to 7 TeV yields a luminosity of 1.2×1034 cm-2s-1 -design !

Luminosity 2011 Integrated proton luminosity 2011 now > 4 fb-1 But we do not transform all our gain in peak luminosity into gain in integrated luminosity – radiation effects on electronics ( later) LHC performance in 2011 - PAF - La Londe ~0.4 fb-1/week 0.2 fb-1/week

Efficiency • ‘Scheduled loss’ of days: • 18 days of technical stops, • 10 days of scrubbing run, • ~16 days of MD. LHC performance in 2011 - PAF - La Londe Lost ~ 11 days due to cryogenics issues (frequently knock on from another issue, like electric network glitch) between July and September • Effective no. of days for physics : 201 – 44 = 157 days. • >> efficiency for stable beams : 29%

Fill length • Luminosity lifetime is typically 16-25 hours. • Optimum fill length ~12-15 hours. • but beams are frequently dumped before due to HW issues. • Ideally we could produce up to ~ 120 pb-1/day ~ 800 pb-1/week • our best is 520 pb-1. LHC performance in 2011 - PAF - La Londe Av. ~6 hours

Projections • ~3 weeks of p operation left in 2011 (some days also for tests !). • Present production: • ~0.4 fb-1 / week • > 5 fb-1 LHC performance in 2011 - PAF - La Londe

High intensity issues • With 50 ns operation and with stored intensity of ~1.8x1014 protons (1380 bunches) a number of issues related to high intensity have started to surface: • Vacuum pressure increases, • Radiation induced failures of critical tunnel electronics, • Heating of the beam screen temperature, • Heating of injection kickers, collimators, • Losses due to (supposed) dust particles, • RF beam loading, • Beam instabilities leading to emittance blow-up. • Those effects have slowed down the pace of the intensity increase, and affect the machine availability. LHC performance in 2011 - PAF - La Londe

Vacuum - electron clouds • Since LHC switched to trains, electron cloud are with us. • E-clouds induced pressure rise and beam instabilities ( large emittance). • Can be conditioned away… by e-clouds ! • In April high intensity beams of 50 ns (up to 1080b) were used at injection to condition the vacuum chamber over a ~10 day period. • Provided adequate conditions for operation (vacuum, beam stability) at 3.5 TeV, with gradual increase of intensity / number of bunches. LHC performance in 2011 - PAF - La Londe Bunch N+2 accelerates the e-, more multiplication… Bunch N+1 accelerates the e-, multiplication at impact Bunch N liberates an e- e- e- N+2 N+1 N ++++++ ++++++ ++++++ e-

Vacuum cleaning with beam • Pressure decrease (normalized to intensity) as a function of effective beam time in April – very effective vacuum cleaning. • Gain one order of magnitude/15 hours. • Common vacuum chamber • regions around experiments • are most critical due to • the overlap of the beams. • During intensity ramp up, additional (and progressive) cleaning occurred in the background. LHC performance in 2011 - PAF - La Londe J.M. Jimenez

Example of vacuum issue – IR2 • When ALICE insisted on flipping the polarity of the solenoid and spectrometer: • Strong local pressure – prevented ALICE from switching ON for some time. Still difficult 3 weeks after the change. • ALICE had to be patient until vacuum was conditioned, plus install new solenoids to reduce multi-pacting from the electrons (see below) LHC performance in 2011 - PAF - La Londe

Vacuum spikes • Vacuum conditions in IR2 and IR8 have been periodically so poor that they lead to beam dumps from beam losses. • Causes of the vacuum spikes are not well understood, could be linked to gas at the warm-cold transitions of the stand-alone cryostats, e-clouds… • This issue is coming and going… has also appeared around CMS recently. 10-6 Example of vacuum issue in IR8 LHC performance in 2011 - PAF - La Londe 10-7

Radiation induced problems • With the increasing luminosity tunnel electronics starts to suffer from SEE (Single Event Errors). • In particular the quench protection ad cryogenics systems. Collisions points Collimators LHC performance in 2011 - PAF - La Londe Loss rate S

Mitigation of radiation effects • As the peak luminosity increased, we have suffered from increasing rates of radiation induced failures (SEE) leading to premature beam dumps – dominated by ‘luminosity radiation’ in IR1 and IR5. • SEE failure are now the dominant cause of beam dumps ! • Some mitigation of radiation effects in 2011: • Relocation of equipment away from the tunnel (cryo and interlock PLC) during the technical stops. • Improvement of codes (FPGAs, PLCs) to cope with errors and avoid reset involving tunnel access (Quench protection system - QPS, cryogenics). Radiation effect were expected to affect the LHC performance in above1033 cm-2s-1 – we clearly hit this problem in 2011 ! LHC performance in 2011 - PAF - La Londe

SEE failure rate evolution M. Brugger QPSPatch Slope: ~1 SEE dump per 60 pb-1 The Number to remember : ~1 SEE dump per 60 pb-1

2011 operation: Identified the most critical equipment – dominated by QPS, cryo. Mitigation through better firmware (QPS), special reset procedures (cryo), signal filtering (RF) etc. SEEs rates from cryo have been reduced a lot. Presently the fill length is basically defined by R2E effects ! 2011/12 Christmas Break (and Technical Stops): Relocation of most critical elements. Additional shielding of most critical areas (UJ’s in Pt1). It seems we cannot do much for QPS equipment located in the dispersion suppressors areas – source of a large fraction of dumps – for LS1. Next long-shutdown: Relocation & Shielding for all critical areas. R2E Situation LHC performance in 2011 - PAF - La Londe

UFO status • Very fast beam loss events (~ millisecond) in super-conducting regions of the LHC were THE surprise of 2010 – nicknamed UFOs (Unidentified Falling Object). • Beam dumps triggered by UFO events: • 18 beams dumps in 2010, • 11 beam dumps in 2011, last beam dump mid-July 2011. • All but one dump at 3.5 TeV. • Things are ‘calming down’ at 3.5 TeV, but the situation is worrying for future 7 TeV operation: • Extrapolation to 7 TeV predicts ~ 100 dumps / year. • Due to lower quench thresholds and larger deposited energy density. LHC performance in 2011 - PAF - La Londe

UFO rate in 2011 5827 candidate UFOs in cell 12 or larger during stable beams for fills longer than 1 hour. Techical Stop (09. – 13.05.2011) Techical Stop (04. – 08.07.2011) Techical Stop (29.08. – 02.09.2011) LHC performance in 2011 - PAF - La Londe Fill number UFO Rate ~ slowly decreasing to ~ 3-4 / hour

Dust particles BLM loss distribution for 3670 arc UFOs LHC performance in 2011 - PAF - La Londe • Most likely hypothesis for UFO: small ‘dust ‘ particle falling into the beam (1-100 mm). • UFO loss amplitude distribution is consisted with measured dust particle distributions in the assembly halls…

UFO distribution in ring 3.5 TeV3686 candidate UFOs. Signal RS05 > 2∙10-4Gy/s. Red: Signal RS01 > 1∙10-2Gy/s. 450 GeV486 candidate UFOs. Signal RS05 > 2∙10-4Gy/s. LHC performance in 2011 - PAF - La Londe The UFOs are distributed around the machine. About 7% of all UFOs are around the injection kickers. . Mainly UFOs around injection kickers (MKI) >> we are focusing on the understanding of UFO at the MKIs

Studies of injection kicker UFOs LHC performance in 2011 - PAF - La Londe • Detailed FLUKA model of the injection region to reproduce UFO losses and help localizing the source(s). • Spare MKI that was removed from the LHC last year will be opened for dust analysis.

Ion beams • The Pb ion beam details are still under discussion. • Most likely scenario: • Bunch spacing of 200 ns(nominal = 100 ns – also possible). • 358 bunches (24 per injection), ~350 colliding pairs per experiment. • Intensity ~1.2x108 ions / bunch. • b* = 1 m in ATLAS, CMS and ALICE (if aperture OK !). • ATLAS & CMS standard crossing angles (120 mrad). LHC performance in 2011 - PAF - La Londe

P-ion test • Issue with mixed p-ion mode is the difference in revolution frequencies at injection (4.5 kHz). At 3.5 TeV the frequencies can be forced to be equal (differences very small, few 10’s Hz). • The beams slip one wrt the other: locations of (parasitic) collisions move longitudinally. Possible source of poor lifetime. • At 3.5 TeV, the frequencies are locked together and the beams must be cogged longitudinally to collide at the right place. • An injection test of protons into ring 1 and Pb into ring 2 is foreseen during the next MD (before the ion run). • Then – depending on the ‘smoothness’ – tests of: • Ramp and cogging at 3.5 TeV. • Test of very low intensity collisions – but unlikely to provide stable beams conditions (time constraints). LHC performance in 2011 - PAF - La Londe

Ion run schedule • Success oriented planning given that the p-ion test is taking place more or less at the same time. • Duration of p-ion not well defined… • We should be able to reach 50-80 mb-1 (2010 : 8 mb-1) LHC performance in 2011 - PAF - La Londe p-ion test (tentative)

Lower beta* • b* could be lowered further in 2012 if the collimators are set tighter around the beams: • b* ~ 0.7 m looks feasible > 30% higher luminosity…. • Tight settings were tested in 2011 (MDs) and operation should be possible – but more delicate since a significant beam halo is then touching the primary collimators. Settings in sigma (e = 3.5 mm) LHC performance in 2011 - PAF - La Londe 2.5 sigma extra margin in the triplet for b* At some point it is only worth pushing the peak luminosity if we can improve the situation of the SEEs. Or fills will become shorter and shorter…

The diode story • Quench propagation tests have been performed during the technical stops. Very positive outcome: • Quenches did not reach the joint at 3.5 TeV. Operation at 3.5 TeV is safer than had been estimated. • But there were some ‘worrying’ side results on the DIODEs: • Resistances of the diode leadsare up to 15 times larger than measured during cold tests in SM18, and seem to strongly increase with the current. • The observed spread in the resistance of 12 diodes leads is very large (factor 20), much larger resistances are likely to be present in some of the other 4000 diode leads of the machine. • The results are irreproducible, and correct simulation is presently not possible due to the large number of unknowns. LHC performance in 2011 - PAF - La Londe

The diode Diode is used as current bypass in case of a quench ‘Half moon’ contact Lower diode busbar Main busbars Upper diode busbar (partially flexible) Upper heat sink ‘Half moon’ contact towards diode Lower heat sink Diode box, Helium contents : 5 l

High current quench simulation Comsol output for the final temperature after a 6 kA quench with Rc,moon=40 mW (adiabatic conditions) A. Verweij 90 K LHC performance in 2011 - PAF - La Londe 180 K 95 K If the anomalous resistance is located at the half-moon connection, there is a risk of melting down at 7 TeV

Diode studies • Cold test in SM18 on several diodes (2011). • Warm test of diodes (ongoing). • Tests in SM18 on magnet+diode (2012). • Proposal to warm up a short section and remove a diode from the machine. • The CSCM project for splice measurements (that is described next) could probe the resistances of all the installed diodes. LHC performance in 2011 - PAF - La Londe

Energy reach and CSCM • Energy reach is so far limited to 3.5 TeV due to risk of bad (= high resistance) splices in the busbar connections of the main circuits. • The CSCM project (Copper Stabilizer Continuity Measurements)aims to develop a method for measuring the resistance of all busbar connections with limited risk and to determine the maximum safe energy sector by sector. • Conditions similar to a quench, but no stored energy in the magnets – thermal run away can be stopped ‘easily’. • Measurements are performed at 20 K when the magnets are not superconducting – current bypasses magnets via the diodes. • Requires modifications to power converters and protection systems. LHC performance in 2011 - PAF - La Londe

CSCM current cycle T=20 K (DT to be defined) H. Thiesen TE/EPC I Trip by mQPS tplateau Iplateau 4-6 kA Fast ramp down if V>Vthr LHC performance in 2011 - PAF - La Londe Fast Power Abort if V>Vthr dI/dt 500 A/s dV/dt to open the diodes 500 A H. Thiesen – 16 August 2011 – TE-TM t1 t2 t 60 s PC in voltage mode PC in current mode

CSCM measurements • All main circuits (RB, RQD, RQF) • All interconnection splices • All current lead-busbar connections at the DFB • All bypass diode paths A. Siemko TE/MPE LHC performance in 2011 - PAF - La Londe H. Thiesen – 16 August 2011 – TE-TM

CSCM and energy status • Reviews of the CSCM will take place this week to analyze the risks and readiness of the project. • Measurement of the entire ring will take ~2-3 months: a full campaign would delay the startup with beam to May 2012. • For the moment the most likely scenario is a measurement of 1-2 sectors as tests in shutdown 2011-2012. The remaining sectors would be measured after the end of the 2012 run. • The decision about the energy at the startup of 2012 may be taken before the CSCM would provide data, either 3.5 or 4 TeV. • Overhead for beam setup of 4 TeV is negligible if we startup with that energy. Some additional hardware commissioning needed (order of 1 week). LHC performance in 2011 - PAF - La Londe

25 ns beams • 25 ns beam was injected in batches of 12 (nominal 278). • Severe electron cloud instability issues – 48 bunches are heavily unstable. • Required beam scrubbing time ~ 5-10 x longer than for 50 ns. • This week a tests with collisions at 3.5 TeV of short trains (12 or 24 bunches) is foreseen. • Injector performance: LHC performance in 2011 - PAF - La Londe

50 versus 25 • Assuming similar emittance blow-up in the LHC (twice as many bunches with 25 ns !): LHC performance in 2011 - PAF - La Londe • We expect 25 ns to be more difficult: • Larger emittance : injection. • Smaller spacing : e-cloud ( longer scrubbing run), vacuum. • Long range beam-beam (twice as many encounters). • Larger stored energy (UFO amplitude and rate?). • But it would of course half the no. events per crossing….

2012 run • LHC startup with beam : 7th March 2012. • 3 weeks for startup with beam (to first moderate intensity). • End of the run ~ mid-November. • 1 month Pb-Pb or p-Pb run ? • No schedule available to date. • Integrated luminosity projection: • 50 ns beam, • same performance of 0.5 fb-1 / week, • assuming 20 effective weeks of high intensity • >> ~ 10 fb-1 integrated L LHC performance in 2011 - PAF - La Londe or even more if we increase L or efficiency !

Summary • The peak performance in 2011 exceeded our most optimistic expectations – we are now in routine ~3x1033 cm-2s-1 regime. • Operation of the beams is smooth, yet we have trouble to achieve long fills and highest integrated performance. • Limited 0.4-0.5 fb-1 / week • The main issue are SEEs that now trigger the majority of the beam dumps. The Quench Protection System is in the first line… • The energy discussion for 2012 is open (3.5 or 4 TeV) – wait for Chamonix Workshop in January. • Extrapolated performance for 2012 is ~ 10 fb-1 assuming similar performance in integrated luminosity. • The machine favors 50 ns over 25 ns, but a common request from the experiments could change the balance towards 25 ns. • Thank you for your attention ! LHC performance in 2011 - PAF - La Londe

Spares LHC performance in 2011 - PAF - La Londe

b* limits • The focusing at the IP is defined by b* which relates to the beam size s • b* is limited by the aperture of the triplet quadrupoles around the collision point and by the retraction margins between collimators. s2 = b* e • Smaller size s at the IP implies: •  Larger divergence (phase space conservation !) •  Faster beam size growth in the space from IP to first quadrupole ! LHC performance in 2011 - PAF - La Londe 33 mm e = 2.8 mm b* = 11 m 1.5 m Squeeze 90 mm

Separation and crossing: example of ATLAS Horizontal plane: the beams are combined and then separated ATLAS IP 194 mm ~ 260 m LHC performance in 2011 - PAF - La Londe Common vacuum chamber Vertical plane: the beams are deflected to produce a crossing angle at the IP to avoid undesired encounters in the region of the common vac. chamber. ~ 7 mm a Not to scale ! 2011 !

1380 bunches with 50 ns spacing Beam 1 Beam abort gap LHC performance in 2011 - PAF - La Londe LHC circumference

LHC status and perspectives