High Brightness Beams for LHC: Needs and Means

1
High Brightness Beams for LHC Needs and Means

• Performance of the LHC injectors chain
• Needs of the LHC upgrade
• Possible improvements
• Batch compression in the PS
• Linac 4
• SPL
• Summary

2

• PERFORMANCE OF THE LHC INJECTORS CHAIN

3
Baseline design (1994) of the LHC injectors
chain

• Main limitations before PS for LHC upgrade
  (1994)

excessive incoherent space charge tune spreads
DQSC at injection in the PSB (50 MeV) and PS
(1GeV) because of the high required beam
brightness N/e.

• Þ PS for LHC project designed to enable
  production of ultimate beam, by fighting the
  space charge limit
• at PSB injection, filling the PS with two PSB
  batches to halve N/e and therefore the tune
  spread in the PS Booster.
• Þ DQ from 0.7 to 0.35 for nominal beam and to
  0.55 for ultimate beam.
• at PS injection, increasing the PSB - PS transfer
  energy from 1 GeV to 1.4 GeV.
• Þ DQ from 0.3 to 0.2 for nominal beam and to 0.32
  for ultimate beam.

4
Limitations of the 1994 baseline design of the
LHC injectors chain

• Longitudinal emittance blow-up during
  debunching/rebunching (40 MHz) at 26 GeV in the
  PS
• Þ too long bunches (gt 5 ns at PS ejection)
• Beam loss at ejection because of the 100 ns
  rise-time of the ejection kicker
• Risk of partial kick of a transmitted bunch
• No clean possibility to send less than 80
  bunches to the SPS

5
Improved longitudinal procedure (1999)in the PS
(1)
0.35 eVs (bunch) 1.1 1011 ppb
40 Blow-up (pessimistic)
1 eVs (bunch) 4.4 1011 ppb
115 Blow-up (voluntary)
1.4 eVs (bunch) 13.2 1011 ppb
6
Improved longitudinal procedure (1999)in the PS
(2)
Triple splitting at 1.4 GeV experimental result
7
Improved longitudinal procedure (1999)in the PS
(3)
Quadruple splitting at 26 GeV experimental
result
8
Improved longitudinal procedure (1999)in the PS
(4)
Final beam at 26 GeV experimental result
Bunch train 1.1 1011 p/bunch Modulation 15
Bunch length 4.0 ns 0.2 ns
Status in 2001. Reduced to 10 in 2002
9
Improved longitudinal procedure (1999)in the PS
(5)
Nominal LHC filling scheme

• 2808 bunches/ring.

10
Evolution of requirements in beam intensity and
brightness

• Since 1994, modifications of the LHC parameters
  and changes of the longitudinal procedure in the
  PS had the following consequences
• for a given luminosity in the LHC, the PSB and PS
  must deliver a higher intensity / brightness.
• this increased intensity / brightness implies
  higher space charge tune spreads at low energy.
• The following contributions have been recorded
• LHC (only for the nominal beam)
• Crossing angle change 200 to 285 mrad
  (1995) factor 1.1
• b change 0.5 to 0.55 m (2003) factor 1.05
• PS change of bunch production scheme (debunching
• rebunching to multiple splitting (1999)
• 12 instead of 10.5 LHC bunches per PSB
  bunch factor 1.14

11
Aggravating effect measured transmission
efficiencies along chain

• 100 transmission from capture in the PSB to SPS
  extraction was assumed in 1994. ? Low emittance
  and no intrinsic loss mechanisms.
• Experience has proved this assumption to be
  wrong, even with a good matching in all planes.
• Transmission efficiencies for nominal and
  ultimate 25 ns LHC beams

• Major part of losses at start of acceleration in
  SPS (possible improvement ?)

12
Overall performance summary
25 ns LHC beam intensity requirements 1994 - 2003
LHC changes
Transmission eff.
PS process change
Þ Ultimate 25 ns beam is far out of reach of the
PSB () with the standard production scheme DQ
at injection in the PSB 0.8 and in the PS
0.45 Þ Nominal 25 ns beam can be produced but
there is no longer a comfortable emittance
budget. (Close to 1994 ultimate requirements) Þ
All other beam variants (75 ns, single bunch
physics beams, pilot, etc.) can be produced by
the PS complex.
13

• NEEDS OF THE LHC UPGRADE

14
Needs of various LHC luminosity upgrade scenarios
Phase Luminosity 1034 cm-2s-1 Comment Brightness factor Protons in 25 ns (en3.75 mm) Protons ejected from the PS (en3 mm)
0 3.6 Large crossing angle 1.5 2.6 1011 3.1 1011
1a 4.6 Low b Large crossing angle 1.0 1.7 1011 2.0 1011
1b 7.8 (e clouds ?) Phase 1a 15 ns bunch spacing 1.7 2.9 1011 3.3 1011
1c 9 1 long bunch _at_ 1ADC 1.2 2.0 1011 2.4 1011
1d 10 80 long bunches _at_ 1.6 ADC 1.9 3.2 1011 3.7 1011
2 10 ? Beam beam compensation gt 1 gt 1.7 1011 gt 2.0 1011
w.r.t. ultimate in LHC
Transmission PS-gt LHC 0.85
LHC Project report 626
15

• POSSIBLE IMPROVEMENTS

16
Increasing brightness in the PS batch
compression (1)

• Proposed procedure
• inject 7 (43) or possibly 8 (44) bunches from
  two PSB batches into the PS operating on harmonic
  9,
• accelerate this beam up to an intermediate energy
  where space charge is sufficiently reduced,
• compress the 7 (8) bunches into 7 (8)/14 of the
  PS circumference by adiabatically increasing from
  h9 to 10,11, 12, 13, 14,
• accelerate the beam on harmonic 14 up to 25 GeV,
• triple split the bunches using rf on h14, 28 and
  42 (similar process than used at 1.4 GeV for the
  25 ns bunch train),
• double split bunches, changing the harmonic from
  42 to 84, and rotate them before ejection, as in
  the present 25 ns bunch train scheme.
• Þ Finally, a train of 42 or 48 bunches,
  spaced by 25 ns
• is sent to the SPS every 3.6 s.
• Best expected performance assuming that the
  space-charge limit (with a 1.2 s flat porch at
  1.4 GeV) is attained in the PS with 841.71011
  protons over the circumference
• Þ 2.61011 ppb _at_ PS ejection

17
Batch compression (7 bunches case) (2)
Bucket Height (arb. units)
Time
18
Batch compression (7 bunches case) (3)
Bucket parameters during the process (arb. units)
Extreme edge Buckets (9 bunches)
19
Batch compression (7 bunches case)
(4)Ultimate filling scheme for 42 PS bunches
(i)

• 2604 bunches/ring only 7 fewer than for
  nominal 72 bunch scheme.
• 6 injections and 18 s SPS flat bottom problems
  with high brightness?

20
Batch compression (7 bunches case)
(5)Alternative filling scheme for 42 PS bunches

• 2436 bunches/ring 13 fewer than for nominal 72
  bunch scheme.
• The 48 PS bunch scheme gives fewer bunches in LHC
  in all cases.

21
Batch compression in the PS (6)Main
characteristics
Stage Advantages Limitations drawbacks
Implementation Fast Low cost (low level RF) Manpower intensive preparation Need for much machine time
Operation Delicate operation (manpower intensive prone to imperfection) Lower LHC filling factor ( -7 ) Longer LHC filling time ( 1.35) Reduced availability for other users 1.2 s flat porch in the PS with high space-charge SPS capability ?
Potential (ppb at PS ejection) 42 bunches every 3.6 s with 2.61011 ppb DQ 0.25
22
Increasing brightness in the PSB withLinac 4
(scheme 1)

• Procedure 1 (72 bunches)
• inject 12 (43) bunches from one PSB batch into
  the PS operating on harmonic 14,
• accelerate this beam on h14 up 25 GeV,
• triple split the bunches using rf on h14, 28 and
  42 (similar process than used at 1.4 GeV for the
  25 ns bunch train),
• double split bunches, changing the harmonic from
  42 to 84, and rotate them before ejection, as in
  the present 25 ns bunch train scheme.
• Þ Finally, a train of 72 bunches, spaced by
  25 ns
• is sent to the SPS every 2.4 s.
• Best expected performance assuming that the PSB
  can deliver 3.61012 protons/ring within 2.5 mm
  emittances (bg2 x 2 at injection) and that
  space-charge in the PS at 1.4 GeV can be
  increased by 1.17 (14/12) (no flat porch)
• Þ 2.01011 ppb _at_ PS ejection

23
Increasing brightness in the PSB with Linac 4
(scheme 2)

• Procedure 2 (48 bunches)
• inject 4 (41) bunches from one PSB batch into
  the PS operating on harmonic 7,
• accelerate this beam on h7 up to an intermediate
  energy,
• double split bunches, changing the harmonic from
  7 to 14,
• accelerate on h14 up to 25 GeV,
• triple split the bunches using rf on h14, 28 and
  42 (similar process than used at 1.4 GeV for the
  25 ns bunch train),
• double split bunches, changing the harmonic from
  42 to 84, and rotate them before ejection, as in
  the present 25 ns bunch train scheme.
• Þ Finally, a train of 48 bunches, spaced by
  25 ns
• is sent to the SPS every 2.4 s.
• Best expected performance assuming that the PSB
  can deliver 3.61012 protons/ring within 2.5 mm
  emittances (bg2 x 2 at injection) and that
  space-charge in the PS at 1.4 GeV can be
  increased by 1.75 (7/4) (no flat porch)
• Þ 3.01011 ppb _at_ PS ejection

24
Increasing brightness in the PSB with Linac 4
Main characteristics
Stage Advantages (25 ns bunch spacing) Limitations drawbacks
Implementation Cost (PM 70 MCHF) Construction time ( 3 years)
Operation Reliability (simple robust operation for the 72 bunches scheme) Short dwelling time at high space charge Reduced LHC filling time ( 0.82 with 72 bunches) Increased beam availability for other users Need for similar RF gymnastics than today in the PS Capability to accept a DQ of 0.44 for a short duration at 1.4 GeV in the PS ? SPS capability ?
Potential (ppb at PS ejection) 72 bunches every 2.4 s with 21011 ppb DQ 0.3 48 bunches every 2.4 s with 31011 ppb DQ 0.44 Safe Possibly OK
25
Replacing the PSB by an SPL (1)Main features

• Injection energy in the PS (SPL CDR-1) 2.2 GeV
• Þ bg2 1.9 DQ 0.53
• For a given DQ, brightness can be doubled.
• Þ 4 1011 ppb at 26 GeV DQ 0.31
• Longitudinal bunch population can easily be
  tailored to the needs (number of bunches,
  distance between bunches)
• Acceleration takes place with the harmonic number
  used at injection
• No need for RF gymnastics (except rebucketing
  inside 40 or 80 MHz buckets)
• Cycling rate is determined by the PS magnetic
  cycle
• A train of 1 to 80 bunches, spaced by 25 ns
  (typical) can be sent to the SPS every 2.4 s

26
Replacing the PSB by an SPL (2)Main
characteristics
Stage Advantages Limitations drawbacks
Implementation Cost (PM 500 MCHF) Construction time ( 5-6 years)
Operation Renewed modern PS injector No need for RF gymnastics in the PS Reliability (simple robust operation) Short dwelling time at high space charge Reduced LHC filling time ( 0.82 with 80 bunches) Increased beam availability for other users SPS capability ?
Potential (ppb at PS ejection) Train of 1-80 bunches every 2.4 s with up to 41011 ppb DQ 0.31
27
Replacing the PSB by an RCS (1)Main features

• To improve space charge in the PS, transfer
  energy has to be significantly above todays
  value (1.4 GeV). Typically 2.2 GeV
• Þ bg2 1.9 DQ 0.53
• For a given DQ, brightness can be doubled.
• Þ 4 1011 ppb at 26 GeV DQ 0.31
• Space charge in the RCPSB must match the PS
  capability. Assuming it has the size of the PSB
• Þ bg2 3.8 w.r.t. 50 MeV
• Þ Injection energy 360 MeV
• To minimize the duration of the PS flat porch, it
  must either have 4 rings or cycle fast (typically
  50 Hz)
• To match the SPL potential w.r.t. the LHC, it
  must deliver gt 8.41012 protons/pulse within
  en2.5 mm
• A train of 72 bunches (4 1011 ppb) , spaced by
  25 ns (typical) can be sent to the SPS every 2.4 s

28
Replacing the PSB by an RCS (2)Main
characteristics
Stage Advantages Limitations drawbacks
Implementation Cost (PM xy0 MCHF) Construction time ( 3-4 years)
Operation Renewed modern PS injector Reliability (simple robust operation for the 72 bunches scheme) Medium dwelling time at high space charge Reduced LHC filling time ( 0.82 with 80 bunches) Increased beam availability for other users Need for similar RF gymnastics than today in the PS SPS capability ?
Potential (ppb at PS ejection) Train of 72 bunches every 2.4 s with up to 41011 ppb DQ 0.31
29
Comparative summary(25 ns bunch spacing)
Batch compression in PS Linac 4 SPL RCS
Cost (PM) Low 50 70 MCHF 500 MCHF gt 150 MCHF
Delay Fast 3 years 5-6 years 3-4 years
Implementation MD intensive Setting-up period during start-up Setting-up period during start-up Setting-up period during start-up
Operation Delicate Limited reliability Comfort Reliability Comfort Reliability Comfort Reliability
Number of bunches / PS pulse 42 72 (48) 1-80 72
Potential intensity per PS bunch 2.6 1011 ppb 2 1011 ppb (3 1011) 4 1011 ppb 4 1011 ppb
Repetition period 3.6 s 2.4 s 2.4 s 2.4 s
BEST USE Exploratory tests Reliable operation 50 of upgrades Reliable operation all upgrades Reliable operation all upgrades
30

• FINAL WORD

31
SUMMARY

• The schemes presently used in the
  injectorscomplex can provide the nominal beam
  for LHC with 25 ns bunch spacing, as well as the
  75 ns bunch train and various test beams.
• The ultimate beam is not feasible today.
• Space charge is the primary limitation, both in
  the PSB and in the PS, and solutions are
  proposed.
• Beam experiments will be necessary to discover
  and address further limitations in the existing
  machines.
• Þ work is needed to understand and minimise the
  beam losses,
• Þ improvements are mandatory to prepare for the
  ultimate beam and the future LHC upgrades
• Þ beam experiments are necessary to investigate
  the potential of the existing accelerators, and
  especially the SPS ( needs early upgrade of the
  lower energy injectors)
• A solution based on RF gymnastics in the PS can
  be considered, but it will never cover the full
  range of possibilities envisaged for the LHC
  upgrade,
• The solution based on a new PSB injector (linac
  4) gives the potential to investigate most
  possibilities, and to operate with high
  reliability.
• Ultimately, if a higher energy accelerator like
  the SPL replaces the PSB, very high performance
  beam would be available at the PS exit.

Facts
Consequences
Actions
32
We hope to have contributed to the analysis of
the situation
Thank you for your attention !
33

• ANNEXES

34
Demonstrated Performance of the PSB for Single
Bunch Beams
Ultimate transverse beam characteristics
()
Data from M. Giovannozzi
35
Batch compressioncase of 8 bunches (1)
Bucket Height (arb. units)
Time
36
Batch compressioncase of 8 bunches (2)
Bucket parameters during the process (arb. units)
Edge Buckets
37
Increasing brightness in the PSB with Linac 4
(scheme 3)

• Procedure 3 (24 bunches)
• inject 4 (41) bunches from one PSB batch into
  the PS operating on harmonic 7,
• accelerate this beam on h7 up to an intermediate
  energy,
• compress the 4 bunches into 4/14 of the PS
  circumference by adiabatically increasing from
  h7 to 9,11, 14,
• accelerate on h14 up to 25 GeV,
• triple split the bunches using rf on h14, 28 and
  42 (similar process than used at 1.4 GeV for the
  25 ns bunch train),
• double split bunches, changing the harmonic from
  42 to 84, and rotate them before ejection, as in
  the present 25 ns bunch train scheme.
• Þ Finally, a train of 24 bunches, spaced by
  25 ns
• is sent to the SPS every 2.4 s.
• Best expected performance assuming that the PSB
  can deliver 3.61012 protons/ring within 2.5 mm
  emittances (bg2 x 2 at injection) and that
  space-charge in the PS at 1.4 GeV can be
  increased by 1.75 (7/4) (no flat porch)
• Þ 6.01011 ppb _at_ PS ejection