1. Plans for ions in the injector complex D
Plans for ions in the injector complex D.Manglunki with the help of I-LHC and LIU-PT teams Special acknowledgements to T.Bohl, C.Carli, E.Carlier, H.Damerau, L.Ducimetière, R.Garoby, S.Gilardoni, S.Hancock, J.Jowett, D.Küchler
How can we reach the requested beam parameters for high luminosity operation in I-LHC after LS2 & LS3?

2. Previously on the LHC ion injector chain…
“Nominal beam” in Design report
L = 1027cm-2s-1 at 7 TeV/c/charge
~600 bunches of 7x107 Pb82+ ions
eH,V= 1.2mm
b* = 0.5m
To combat IBS and space charge on SPS flat bottom, Complicated gymnastics in PS & SPS (splitting in bunchlets in PS, recombining in SPS using 100MHz system).
Scheme questioned in Chamonix XII (2003)
Decision to start with “EARLY” scheme, single bunch from LEIR > PS > SPS
Assumed possibility of up to 13 PS injections into SPS

3. Experience in 2010 (EARLY) and 2011
Linac3 delivers ~30-40% of design (15-20mA vs 50)
Additional injections in LEIR
Better beam lifetime at low energy in PS thanks to excellent vacuum
IBS and DQ on SPS flat bottom less harmful than foreseen
“EARLY” run in 2010:
1.2×108 ions/bunch; eH,V= 1.2mm ; L = 3x1025cm-2s-1 at 3.5TeV/c/charge
Provisionally demonstrated that bunchlets are unnecessary
Raised the bar high for 2011 run
(Terribly named) “INTERMEDIATE” run in (Beam designed over coffee in Evian)
Keep bunch intensity as high as possible (no splitting)
Insert as many bunches as possible into LHC
“quick’n’dirty”, reversible 200 ns bunch spacing in PS
Batch spacing 200ns thanks to shorter SPS injection kicker rise time (only 3 modules at 17GeV/c/charge)

4. Present scheme (“intermediate” in 2011)
LEIR
7 multiturn injections of Linac3 pulse (~15-20mA)
2 bunches of 4.5x108 Pb54+
PS (similar gymnastics as nominal, minus splittings)
Batch expansion h = 16 -> 14 -> 12
Rebucketing (was splitting) h = 12 -> 24
Batch expansion h = 24 ->21
Rebucketing h = 21 -> 169
After stripping, 2 bunches of ~3x108 Pb82+ bunch spacing 200 ns
SPS
12 injections of PS batches, batch spacing 200 ns
24 bunches of ~1.4x108 Pb82+ (0.9x108 design)
Transverse emittances ~0.85mm (1.2 design)

5. Present issues (which have limited 2011 performance to only 2x design lumi…)
Low current from Linac3
Forces LEIR to inject 7 times instead of 4, decreasing cooling time
LEIR losses after capture and at the beginning of the ramp
Currently not understood
Longitudinal emittance budget in PS
Not an issue when splitting but creating satellites otherwise
RF Noise, IBS & DQ on SPS flat bottom
First batch suffers 40 more seconds on flat bottom: lower intensity/bunch, transverse emittance blowup

6. Users requirements 2012 p-Pb 2015-17 Pb-Pb (+ p-Pb)
Pb-Pb, p-Pb and Ar-Ar
“Ar-Ar if and when Pb-Pb accumulated more than 1 nb-1 ” while ~0.15 nb-1 accumulated in 2011
After LS3 Pb-Pb
Goal: 10 nb-1 … needs ~4-5 luminosity increase
Also after LS3 … d-Pb?
ALICE will tell at the end of 2012
No other species requested as of now

7. Luminosity increase Note: current brightness performance is already 2 x design
b functions at IP
Transverse emittances
Intensity per bunch
Number of bunches -> reduce spacing
Nominal was 100 ns / 225 ns
Present scheme is 200 ns
ALICE ask for 50ns …but number of bunches will not be x4
Influence of LHC injection kicker (0.9ms) & abort gap (3ms)
Effect of SPS kicker rise time:
12 injections in SPS
Scaling Luminosity as NB x IB2

8. Lumi scaling (%) vs SPS injection kicker rise time (ns)

9. Possible route to 50ns Design current from Linac 3 (~50mA) LEIR
Produce 2 bunches of ~109 Pb54+ in same emittance (i.e. twice today)
PS gymnastics
Batch compression to 100ns h = 16 -> 18 -> 21 (no need for new cavities, 10MHz system exists)
Splitting h = 21 -> 42 (20MHz system exists but VRF acceptance to be checked)
4 bunches > 1.4 x108 Pb82+ into SPS
12 SPS injections spaced by 50ns
Similar bunch quality as present beam
48 bunches of ~1.4x108 Pb82+
Transverse emittances ~0.85mm
But with 50ns spacing and hopefully less spread in bunch population
Note: longer LHC injection time

10. Necessary studies/upgrades
Linac 3 (~50mA)
New ECR source?
Multiple charge acceleration?
Faster Linac rep rate (10Hz)?
Construction of Linac 5?
LEIR
loss at acceleration / limits to be understood
ECOOL?
Transverse damper? PS
Tranverse damper?
RF gymnastics
Revive bunchlets scheme?
SPS
Can Q20 help on flat bottom?
RF noise?
Upgrade of MKP?
Increase cycle length?
Reinstall 100 MHz system (cavities, amplifiers, beam control)?

11. What can we do today (i.e.2015) ?
Same beam from Linac 3 (20mA) into LEIR into PS (2 bunches…)
PS gymnastics
Batch compression h = 16 -> 18 -> 21 (100 ns)
12 SPS injections
Spaced by 200 ns (resp. 150 ns)
Resulting beam
SPS train: 24 bunches of 1.4 x108 Pb82+
Transverse emittances ~0.85mm
Spacing 3x100 ns + 1x200 (resp. 150) ns
460 (resp. 530) bunches per LHC ring in 19 (resp. 22) injections from SPS
Luminosity increase ~25% (resp. 47%)

12. Other species
Ar / Xe
Will be studied/produced starting in 2013 in ECR/RFQ/Linac3, in view of fixed target runs for NA61 in 2014/15.
Could be available to LHC before LS2 (but no request)
The present baseline is to use the same ECR for all ion species
Deuterons
Cannot be produced in present ECR -> Need a new source
Cannot be accelerated in Linac3 -> Need a new pre-accelerator
Uranium (No official request)
Many safety/handling issues

13. Conclusions
With the present injector complex, increasing the number of bunches seems to be the only route for a marginally higher luminosity, and at the expense of a longer LHC filling time
If we are to implement the suggested improvements in order to reach the required Pb-Pb luminosity (provided the LHC can digest it), it is more than time to start the RnD on all parts of the injector chain.
Ar and Xe will be available after LS1 (parameter list still to be defined and optimised) but other species, if desired, would come in addition, and require more studies, in particular a new source & pre-accelerator for deuterons, or safety and handling issues for Uranium.