UFOs Observations Statistics and Extrapolations Evian Workshop 2012

UFOs Observations, Statistics and Extrapolations Evian Workshop 2012 Tobias Baer December, 19 th 2012 Acknowledgements: V. Baglin, M. J. Barnes, S. Bart Pedersen, F. Cerutti, F. Day, B. Dehning, L. Ducimetière, E. Effinger, A. Ferrari, N. Fuster Martinez, N. Garrel, A. Gerardin, B. Goddard, E. Griesmayer, M. Hempel, E. B. Holzer, S. Jackson, M. J. Jimenez, V. Kain, A. Lechner, V. Mertens, M. Misiowiec, R. Morón Ballester, E. Nebot del Busto, A. Nordt, M. Sapinski, R. Schmidt, J. Uythoven, V. Vlachoudis, J. Wenninger, C. Zamantzas, F. Zimmermann, … December, 19 th 2012 Evian Workshop 2012 1

Content 1. UFO Observations and Statistics 2. MKI UFO Observations 3. Extrapolation for after LS 1 4. Mitigation Strategies and Outlook December, 19 th 2012 Evian Workshop 2012 2

Content 1. UFO Observations and Statistics 2. MKI UFO Observations 3. Extrapolation for after LS 1 4. Mitigation Strategies and Outlook December, 19 th 2012 Evian Workshop 2012 3

Introduction In 2012: 20 beam dumps due to (Un)identified Falling Objects. B 1 B 2 200 m Pt. 4 UFO location 2011: 17 dumps, 2010: 18 dumps. 14 dumps at 4 Te. V, 4 Te. V 3 during ramp, 3 at 450 Ge. V. 8 dumps by MKI UFOs, Diamond BLM in IR 7 4 by UFOs around collimators during movement (TCL. 5 L 5. B 2, TCSG. 4 L 6. B 2) 4 by ALICE Ufinos. ≈ 17, 000 candidate UFOs below BLM thresholds found in 2012 2011: about 16, 000 candidate UFOs. Spatial and temporal loss profile of UFO at BSRT. B 2 on 27. 08. 2012 at 4 Te. V. December, 19 th 2012 Evian Workshop 2012 4

Loss Duration BLM Study Buffer provides for all BLMs temporal loss profile with 80µs resolution (350 ms buffer length). Average temporal width of 4 Te. V arc UFOs has maximum at about 80µs. May become faster with smaller emittance at higher energies. December, 19 th 2012 Evian Workshop 2012 Risetime: 212µs Falltime: 92µs 683 arc (≥ cell 12) UFO events at 4 Te. V with 1374/1380 50 ns bunches. Signal RS 04 > 2∙ 10 -4 Gy/s. Peak loss (fit) > 1∙ 10 -3 Gy/s. Only datasets with R² ≥ 0. 97. Data is corrected for BLM time response. 5

Arc UFO Rate 2011: Decrease from ≈10 UFOs/hour to ≈2 UFOs/hour. 2012: Initially, about 2. 5 times higher UFO rate than in October 2011. UFO rate decreases since then. Up to 10 times increased UFO rate with 25 ns December, 19 th 2012 Evian Workshop 2012 6

Content 1. UFO Observations and Statistics 2. MKI UFO Observations 3. Extrapolation for after LS 1 4. Mitigation Strategies and Outlook December, 19 th 2012 Evian Workshop 2012 7

MKI UFOs • In 2012: 8 dumps by MKI UFOs (5 at 4 Te. V). 2011: 11 dumps by MKI UFOs (2010: 2) 2012: 6 dumps from MKI. D 5 L 2 (since 2010: 17). • During MKI inspection in 2011 many macro particles were found, which are believed to originate from Al 2 O 3 ceramic tube Typical macro particle diameter: 1 -100µm. Al O A. Gerardin, N. Garrel EDMS: 1162034 • Temporal distribution: Mainly within 30 min after last injection. Many events within a few hundred ms after MKI pulse. (T. Baer et al. , CERN-ATS-Note-2012 -018 MD) pulse • Positive correlation between MKI UFO rate and pressure in MKI tank at 450 Ge. V. 1024 MKI UFOs until 6. 12. 2012. in 1374 b/1380 b physics fills with at least 3 hours after the last injection. Signal RS 04 > 2∙ 10 -4 Gy/s. (T. Baer et al. , CMAC #6) December, 19 th 2012 Evian Workshop 2012 8

MKI UFO rate Decrease from about 8 MKI UFOs per fill with 1380 b in 2011 to about 2 MKI UFOs per fill No increased activity for 25 ns 4 Te. V fills (with intermediate intensities). December, 19 th 2012 Evian Workshop 2012 9

MKI UFO Distribution Beam 1 145/52 UFOs around MKIs in IR 2 in 159/77 fills with stable beams and at least 1000 bunches before/after TS#3. Signal in RS 4>5· 10 -4. replaced in TS#3 - 32% Beam 2 -72% 194/52 UFOs around MKIs in IR 2 in 159/77 50 ns fills with stable beams and at least 1000 bunches before/after TS#3. Signal in RS 4>5· 10 -4. Slightly different BLM positions since TS#3 may slightly perturb the analysis. Most MKI UFOs occur at MKI. D 5 L 2. UFO activity at MKI. D 5 R 8 (replaced during TS#3: improved cleaning, reduced electrical field) after TS#3 (72 ± 11)% lower. In the other IR 8 MKIs, UFO activity decreased by (32 ± 12)%. December, 19 th 2012 Evian Workshop 2012 10

Content 1. UFO Observations and Statistics 2. MKI UFO Observations 3. Extrapolation for after LS 1 4. Mitigation Strategies and Outlook December, 19 th 2012 Evian Workshop 2012 11

25 ns Observations from 2012 operation with 25 ns: • During high intensity fills at 4 Te. V about 5 -10 times increased arc UFO activity. Particularly high arc UFO rate in first fills (3427 -3429). • Between 13. 12. 2012 and 17. 12. 2012: 10 UFOs above 10% of BLM dump threshold. Largest UFO 67% of BLM dump threshold. • During 450 Ge. V scrubbing run: In total 221 MKI UFOs in 13 fills with > 1014 protons ≈ 17 MKI UFOs per fill. For 50 ns ≈2 MKI UFOs per fill. • No increased MKI UFO rate in 25 ns 4 Te. V fills. December, 19 th 2012 Evian Workshop 2012 12

Quench Margin 2012 Largest arc UFO in 2012: 0. 67 m. Gy at BLM 2 (5. 10. 2012 06: 19: 41). Assume that UFO occurred at Pos #2 (i. e. close to the MQ): From FLUKA simulations (for 3. 5 Te. V): ≈1. 3· 108 inelastic interactions needed to explain BLM signal. (A. Lechner at QTSWG, 04. 05. 2012) Peak energy density in MB. C: ≈3. 9 m. J/cm³. Simulations by A. Lechner and the FLUKA team December, 19 th 2012 Evian Workshop 2012 13

Quench Margin at 7 Te. V For UFO at Pos #1 with inelastic interactions: ≈1. 3· 108 Simulations by A. Lechner and the FLUKA team x 4. 2 at 3. 5 Te. V: ≈7. 8 m. J/cm³. at 7. 0 Te. V: ≈32. 5 m. J/cm³. For 3. 5 Te. V: About a factor 4 below QP 3 quench margin. For 7 Te. V: About a factor 10 - 20 (QP 3/Note 44) above quench margin. December, 19 th 2012 M. Sapinski Evian Workshop 2012 Note 44 ZERODEE M. Sapinski QP 3 Note 44 ZERODEE FLUKA simulations for UFO at Pos #1. 14

Energy Extrapolation to 7 Te. V: BLM signal/threshold (based on Note 44) is for arc UFOs about 20 times larger than at 3. 5 Te. V. Based on 2012 arc UFOs: 91 UFO related beam dumps. (based on 2011: 112 dumps) Additionally, 21 beam dumps by MKI UFOs (2012 data, full cycle). (based on 2011: 27 dumps) Additionally (not considered): UFOs around IRs until cell 11, at collimators/movable devices and Ufinos in experiments. December, 19 th 2012 Based on the applied threshold table from 10. 12. 2012. For MKI UFOs, only the BLMs at Q 4 and D 2 are considered. The energy scaling applies only to events at flat top, but (for MKI UFOs) the full cycle is taken into account for the extrapolation. Apart from the beam energy, identical running conditions as in 2011/2012 are assumed. In particular not included are: margin between BLM thresholds and actual quench limit, 25 ns bunch spacing, intensity increase, beam size, scrubbing for arc UFOs, deconditioning after LS 1. Evian Workshop 2012 15

Content 1. UFO Observations and Statistics 2. MKI UFO Observations 3. Extrapolation for after LS 1 4. Mitigation Strategies and Outlook December, 19 th 2012 Evian Workshop 2012 16

MKI UFO Mitigations Improvements of all MKIs: • NEG coating of interconnects, by-pass-tubes and BI equipment Mitigation of electron-cloud. Improved vacuum in MKI. • Improved cleaning procedure Reduced initial dust contamination. Additional cleaning after insertion of metal wires. • Up to 24 screen conductors on all MKIs Reduced electric field during MKI pulse. • Coating of ceramic tube (under investigation, max. 1 MKI). Reduced SEY and surface charging. Two options under consideration: carbon and Cr 2 O 3. December, 19 th 2012 Evian Workshop 2012 (cp. EDMS Doc. No. 1235378) 17

Arc UFO Mitigations • Increase BLM thresholds towards the quench limit. A. Lechner • UFO scrubbing. • Optimized BLM distribution for better protection against arc UFOs. 12. 10. 2012) • (M. Sapinski at MPP, ADT fast-losses quench test. -> see presentation by Eduardo Better understanding of quench margin. December, 19 th 2012 Evian Workshop 2012 18

Summary • 20 beam dumps due to UFOs in 2012. • Temporal width typically 50 -200µs May be too fast for active protection with smaller emittance at higher energy. • Arc UFO rate at beginning of 2012 ≈2. 5 times higher than in October 2011. Arc (and MKI) UFO rate decreases since then. • Energy extrapolation to 7 Te. V: 2011 arc and MKI UFOs would have caused 139 beam dumps. 2012 arc and MKI UFOs would have caused 112 beam dumps. • About 5 -10 times increased UFO activity with 25 ns. • Mitigations: For MKI UFOs, different mitigations are in preparation. Observations with improved MKI. D 5 R 8 look promising. For Arc UFOs, optimized BLM distribution allows a better UFO protection. December, 19 th 2012 Evian Workshop 2012 19

Thank you for your Attention Further information: • E. Nebot del Busto et al. , “Detection of Unidentified Falling Objects at the LHC”, HB 2012, TUO 1 C 04. Tobias Baer CERN BE/OP Tobias. Baer@cern. ch • T. Baer et al. , “UFOs in the LHC: Observations, Studies and Extrapolations”, IPAC’ 12, THPPP 086. • B. Goddard et al. , “Transient Beam Losses in the LHC Injection Kickers from Micron Scale Dust Particles”, IPAC’ 12, TUPPR 092. • A. Lechner et al. , “FLUKA Simulations of UFO-Induced Losses in the LHC Arc“, Quench Test Strategy Working Group. • T. Baer et al. , “UFOs in the LHC after LS 1”, Chamonix Workshop 2012. • T. Baer et al. , “UFOs in the LHC”, IPAC’ 11, TUPC 137. • N. Fuster et al. , “Simulation Studies of Macroparticles Falling into the LHC Proton Beam”, IPAC’ 11, MOPS 017. December, 19 th 2012 Evian Workshop 2012 20

Content Backup slides December, 19 th 2012 Evian Workshop 2012 21

Spatial UFO Distribution • Distribution in 2011 and 2012 is very similar. MKI Additional BLMs in cell 19 R 3 • In 2012 larger fraction of MKI UFOs. MKI • Some arc cells with significantly increased number of UFOs: 19 R 3 B 1, 25 R 3 B 2, BSRT B 2, 31 L 7 B 2, 28 R 7 B 2, 28 L 8 B 1… 407 BSRT 332 No direct correlation with sector 2011: 347409 repairs identified. UFOs at 3. 5 Te. V. 2012: 6922 UFOs at 4 Te. V until 6. 12. 2012. Signal RS 04 > 2∙ 10 -4 Gy/s. Gray areas around IRs are excluded from the analysis. December, 19 th 2012 Evian Workshop 2012 22

UFOs in cell 19 R 3 • Additional BLMs in cell 19 R 3 to determine UFO location. 19 R 3 UFOs with different spatial loss pattern were observed in cell 19 R 3, suggesting that the UFOs originate from various position across the cell. 19 R 3 December, 19 th 2012 cp also A. Lechner et al. , Quench Test Strategy Working 19 R 3 Group, May 2012 Evian Workshop 2012 23

Intrafill UFO rate No significant change of UFO rate throughout a fill. December, 19 th 2012 Evian Workshop 2012 24

Below Threshold UFOs 4513 arc UFOs (≥cell 12) at 3. 5 Te. V with signal RS 01 > 1∙ 10 -3 Gy/s. courtesy of J. M. Jimenez December, 19 th 2012 Evian Workshop 2012 25

Intensity Dependency 500 candidate UFOs during stable beams with a signal in RS 04 > 2∙ 10 -4 Gy/s. 28 fills with at least 1 hour in stable beams in the first quarter of 2012 are considered. The beam intensity is computed as the maximum intensity per fill, averaged over both beams. December, 19 th 2012 Evian Workshop 2012 26

Asymmetry vs. Loss Amplitude Simulation Measurement Simulated loss rate for different macroparticle masses. Beam intensity: 1. 6· 1014 protons. courtesy of F. Zimmermann, N. Fuster IPAC’ 11: MOPS 017 683 arc (≥ cell 12) UFO events at 4 Te. V with 1374/1380 50 ns bunches. Signal RS 04 > 2∙ 10 -4 Gy/s. Peak loss (fit) > 1∙ 10 -3 Gy/s. Only datasets with R² ≥ 0. 97. Data is corrected for BLM time response. Falltime becomes faster than risetime for large loss amplitues. December, 19 th 2012 Evian Workshop 2012 27

UFO Duration vs. Amplitude Risetime Falltime No significant correlation of UFO duration with loss amplitude. December, 19 th 2012 Evian Workshop 2012 28

UFO Rate vs. Emittance from luminosity at start of stable beams. courtesy of F. Day UFO rate increases with normalized emittance. December, 19 th 2012 Evian Workshop 2012 29

UFO rate vs Bunch Intensity No dependency of UFO rate on bunch intensity. Data for 3336 candidate arc UFOs during stable beams in 2011 and 2012 (until 19. 07. 2012). Fills with 1374/1380 bunches and at least 1 hour of stable beams are considered. Signal RS 04 > 2∙ 10 -4 Gy/s. December, 19 th 2012 Evian Workshop 2012 30

Peak Signal courtesy of E. Nebot • No clear dependency of peak loss on intensity. Dec. 2010) (cf. E. B. Holzer at Evian • No clear dependency of peak loss on bunch intensity. December, 19 th 2012 Evian Workshop 2012 31

4 Te. V Arc UFOs with 25 ns • With 50 ns, 1374 b: ≈ 1. 3 arc UFOs per hour Fill # bunches Time @ 4 Te. V [h: m] # arc UFOs/hour 3425 84 3: 21 8 2. 38 (± 0. 84) 3427 156 1: 02 7 6. 73 (± 2. 55) 3428 372 2: 37 36 13. 68 (± 2. 28) 3429 804 9: 49 87 8. 85 (± 0. 95) 3436 804 1: 19 8 6. 02 (± 2. 13) 3441 108 2: 11 5 2. 29 (± 1. 02) 3442 204 2: 31 5 1. 98 (± 0. 88) 3443 396 0: 38 3 4. 73 (± 2. 73) 3453 396 8: 43 11 1. 26 (± 0. 38) 3457 780 0: 28 2 4. 17 (± 2. 95) • 13. – 17. 12. 2012: 10 UFOs > 10% of BLM dump threshold Largest UFO: 67% of BLM dump threshold December, 19 th 2012 Evian Workshop 2012 32

Identified/Possible UFO Sources • Movable Devices (roman pots, collimators). • Ceramic particles from MKI tube. • Distributed ion pumps (PF-AR, HERA), not in LHC arc. • Macro particles frozen or condensated at cold elements. • Sparking/Electrical Discharges (PF-AR). December, 19 th 2012 Evian Workshop 2012 33

UFO Model • Implemented in dust particle dynamics model, which predicts (among others): Loss duration of. Detaching a few ms. stimulated by vibration, electrical field Losses become faster for larger during MKI pulse and/or electrical beam potential. beam intensities. Al 2 O 3 fragment of vacuum chamber. Size: 1 -100µm. Interaction with beam Potentially charged leads to positive charging by electron cloud of UFO. Particle could be repelled by beam courtesy of F. Zimmermann, N. Fuster IPAC’ 11: MOPS 017 Beam loss rate as a function of time for different macroparticle masses. Beam intensity: 1. 6· 1014 protons. December, 19 th 2012 Evian Workshop 2012 ceramic tube e- e- e 19 mm Metal strips for image currents Beam Local beam losses due to inelastic nuclear interaction. 34

Arc UFO FLUKA Simulations • UFO amplitude: At 7 Te. V about 4 times higher than at 3. 5 Te. V. For UFO at Pos #1 At indicated longitudinal position for UFO at 7 Te. V. Beam direction: out of screen. x 4 Courtesy of A. Lechner and the FLUKA team. December, 19 th 2012 Evian Workshop 2012 35

Vacuum Correlation 5. 1σ statistical significance of positive correlation. 141 MKI UFOs in Pt. 8 between last injection of beam 2 and beginning of ramp for 178 fills with 1374/1380 bunches until 23. 07. 2012. Positive correlation between pressure at MKI and MKI UFO rate. Similar indications also from scrubbing runs and 2012 UFO MD. December, 19 th 2012 Evian Workshop 2012 36

UFOs after MKI Pulse • Many events within a few hundred ms after MKI pulse. T. Baer et al. , CERN-ATS-Note-2012 -018 MD • First event 3 ms after MKI pulse. Compared to 62 ms for free fall from aperture. Could be explained by negatively charged particles. (F. Zimmermann at LIBD, 29 th Nov. 2011) T. Baer et al. , IPAC‘ 11, TUPBC 137 December, 19 th 2012 Evian Workshop 2012 37

Macro Particle Size t s u m s O F U e g r a l f o. s m u µ i 0 d 4 a R Particle mass t least ≈ UFO event on 16. 07. 2011 14: 09: 18 be a I=1. 02· 1014 protons, E=3. 5 Te. V, with ԑn=2. 5µm·rad, βx=158. 5 m, βy=29. 5 m, σx=325µm, σy=140µm. Nuclear interaction length Material Resulting mass (A) Radius of spherical object Al (l=39. 7 cm, ρ=2700 kg/m 3) 5. 53· 1017 43µm Al 2 O 3 (l=24. 8 cm, ρ=3970 kg/m 3) 5. 08· 1017 37µm December, 19 th 2012 Evian Workshop 2012 38

Vibration Measurements • Measurements carried out on spare MKI with kicker pulsing at full voltage under vacuum using accelerometers and laser vibrometers. • Many issues of electrical noise and spurious vibration (e. g. pumps) • When the kickers fire, a mechanical vibration in 60 -300 Hz range is measured. The amplitudes are but very small (≈10 nm). nm courtesy of R. Moron Ballester, S. Redaelli EDMS: 1153686 December, 19 th 2012 Evian Workshop 2012 39

Lead MKI UFOs IP 8 MQML. 10 L 8 (highest loss) MKI UFO at MKI. D 5 R 8. TCTH MKI (UFO location) 10 % of threshold at MQML. 10 L 8. Losses are not as localized as for protons. Highest loss is in the dispersion suppressor downstream of the IR (due to ion fragmentation). Horizontal dispersion December, 19 th 2012 Evian Workshop 2012 40

Plans for 2012 • Better localization of arc UFOs by mobile BLMs in cell 19 R 3. • FLUKA simulations for arc UFOs. • Better Mad. X simulations temporal resolution for UFOs. of UFO events (dust particle dynamics). 80µs time resolution of BLMof study buffer. • Better temporal resolution UFO events (dust particle dynamics). UFO location • Bunch-by-bunch diagnostics with diamond BLMs. 80µs time resolution of BLM study buffer. Study. Bunch-by-bunch impact of 25 ns operation. diagnostics with diamond BLMs. 25 ns high intensity (several 100 - 1000 bunches) beam for few hours at 4 Te. V. • Study impact of 25 ns operation. • MKI 25 ns UFO MD. high intensity (several 100 - 1000 bunches) beam for few hours at 4 Te. V. Courtesy of 25 ns, dynamics. 25 ns e-cloud correlation, UFO production mechanism, mechanism particle M. Hempel • MKI UFO MD. cp. LBOC, 24. 07. 2012 25 ns, 25 ns e-cloud correlation, UFO production mechanism, mechanism particle dynamics. scheduled, but due to many (unrelated) technical problems with limited 2 success. x 201 i n o m a m Ch December, 19 th 2012 Evian Workshop 2012 Fro 41