1. beam-beam compensation schemes
Frank Zimmermann
HHH-2004, 09/11/2004
perspective
alternating crossing
electron lenses
electromagnetic wires
super-bunches
crab cavities

2. (1) perspective

3. historical attempts of beam-beam ‘simulation’ & compensation
4-beam collisions at DCI ~1970 charge neutralization; limited by collective instabilities (J.Augustin et al., Y.Derbenev)
nonlinear lens at ISR, 1975 ‘simulation’ of head-on collision, copper bars with 1000 A current (E. Keil, G. Leroy)
octupoles compensating BB tune spread for LEP, 1982
(S. Myers)
octupoles at VEPP-4/-2M & DAFNE, 1986, (A. Temnykh, M. Zobov)
plasma compensation for linear & muon colliders, (D. Whittum, A. Sessler et al.)
Tevatron Electron Lens, FNAL, operational since 2001 compensating intra-&inter-bunch tune spread (E.Tsyganov, 1993
V. Shiltsev, 1998)
octupoles in LHC?

4. (Courtesy A. Verdier)

5. 30 long-range collisions per IP
LHC: 4 primary IPs
and
30 long-range collisions per IP
120 in total
compensation can act on either head-on or long-range collisions, or on both

6. long-range beam-beam collisions
perturb motion at large betatron amplitudes, where particles come close to opposing beam
cause ‘diffusive aperture’ (Irwin), high background, poor beam lifetime
increasing problem for SPS, Tevatron, LHC,...
that is for operation with larger # of bunches
#LR encounters
SPS 9
Tevatron Run-II 70
LHC
120

7. ey experience from Tevatron Run-II time
“long-range beam-beam interactions in Run II at the Tevatron
are the dominant sources of beam loss and lifetime limitations of
anti-protons …” (T. Sen, PAC2003) ey
LR collisions reduce the
dynamic aperture by about
3s to a value of 3-4s;
little correlation between
tune footprint and dynamic
aperture
drop in ey for first 4 pbar
bunches after injection; asymp-
totic emittance is measure
of their dynamic aperture
time

8. (2) alternating crossing

9. crossing scheme D. Neuffer, S. Peggs:
“BEAM-BEAM TUNE SHIFTS AND SPREADS IN THE SSC: HEADON, LONG RANGE, AND PACMAN CONDITIONS,” SSC-63, Apr Proposal of alternating crossing at two IPs, which cancels the linear part of the long-range beam-beam tune shift and reduces bunch-to-bunch tune spread.
Alternating planes of crossing at IP1 & 5 and IP2 & IP 8 were
adopted as LHC baseline (W. Herr, CERN/SL/90-06 (AP) (1990)).
Since a beam screen is to be installed in the final triplet quadrupoles
for vacuum reasons, the orientation of the LHC crossing planes will
be frozen for each IP.
Alternative crossing schemes would offer other advantages, like
vanishing vertical dispersion (xx), easier long-range beam-beam com-
pensation (yy), or equal background in the different experiments (xx
or yy). They might also prove intrinsically more stable?!
Another option would be inclined hybrid crossing (K. Takayama et al.).

10. LHC tune “footprint” due to head-on & long-range collisions
in IP1 & IP5
(Courtesy
H. Grote)
LR vertical crossing
head-on
LR horizontal crossing

11. LR collisions ‘fold’ the footprint!
total LHC
tune
“footprint”
for regular
and
“PACMAN”
bunch
(Courtesy
H. Grote)
tune footprints of nominal & PACMAN bunches
similar thanks to alternating crossing
LR collisions ‘fold’ the footprint!

12. diffusion rate (log scale) as a function of amplitude
LHC
centre
of other
beam
here tunes w/o
beam-beam were
held constant
diffusive aperture
with xx or yy
crossing
diffusive aperture
with alternating
crossing
comparing xy, xx and yy crossing for two working points

13. weak-strong beam-beam simulations for different crossing schemes in LHC keeping the same 0-amplitude tunes

14. simulation parameters
symbol
value
horizontal tune at A=0 Qx
vertical tune at A=0 Qy
bunch population Nb
1.15x1011
beta function at IP
b*x,y
0.55 m
relativistic gamma g
7461
normalized emittance (1s) eN
3.75 mm
full crossing angle q
285 mrad
no. of IPs
NIP 2
no. of parasitic crossings / IP
Npar 30
rms beam size at IP
s*x,y
16.6 mm
rms beam divergence at IP
q*x,y
30.2x10-6 mrad

15. xy xy w/o HO xx w/o HO, xx yy, yy w/o HO
Simulated diffusion rate as a function of start amplitude for XX, XY and YY
crossing with LR only and with the combined effect of LR and SR collisions,
for the same 0-amplitude tune , ; start amplitudes x=y.

16. xx, yy, and xy crossing in LHC IPs 1 & 5 w & w/o HO @ different y tunes
varied in steps of 0.005

17. qc=285 mrad
qc=200 mrad
[Qy]=64
[Qy]=64
simulated LHC diffusive aperture for nominal & reduced crossing angle vs. Qy, x=y
qc=285 mrad
qc=200 mrad
[Qy]=63
[Qy]=63

18. qc=285 mrad
qc=285 mrad
[Qy]=64
[Qy]=63
simulated LHC diffusive aperture for nominal & reduced crossing angle vs. Qy, y=0 (only x amplitude nonzero), in this case xx and yy crossing are always stable, and diffusive aperture found only
for xy crossing

19. xx xy frequency maps for nominal LHC tunes yy thanks to Yannis
Papaphilippou
for his help in
calculating
frequency
maps!

20. xx xy frequency maps for QY lower by 0.01 (~on coupling resonance) yy
thanks to
Yannis
Papaphilippou
for his help in
calculating
frequency
maps!

21. result of these simulations:
in all cases diffusive aperture is larger
for equal-plane crossing
possible explanation:
different ‘folding’ since xy crossing
cancels dodecapole and 20-pole terms;
&
(2) twice the number of resonances
for xy crossing

22. attempt at an experimental verification in the SPS

23. 2nd generation BBLR in the CERN SPS
can model various LHC crossing schemes
G. Burtin, J. Camas, J.-P. Koutchouk, et al.

24. xx “xy” “xy-2” yy approximations of XY, XX, and YY crossing
in the SPS experiment
BBLR2x on
(strength x2)
beam xx
BBLR1 off
BBLR2x on
beam
x bump -23 mm
BBLR1 on
“xy”
beam
x bump -23 mm
BBLR2x off
&
“xy-2”
(strength x2)
BBLR1 on
(strength x2) yy
x bump -23 mm

25. experiment simulation experimental result: beam lifetime
measured for the three SPS wire
configurations and w/o BBLRs
diffusive aperture simulated for
the three SPS wire configurations
experiment
simulation
simulated diffusive aperture for XX crossing is 10% larger than for
‘quasi-XY’ or ‘quasi-YY’ crossing
measured beam lifetime is best
for XX crossing, second best for
‘quasi-YY’ crossing, lowest for
‘quasi-XY’ crossing; the case
w/o BBLR is similar to ‘quasi-XY’ ,
- so the equal-plane crossing
actually improve the lifetime
lifetimes in this experiment are
comparable to lifetimes w/o BBLR
we need higher wire current or shorter
distances for better resolution (tonight!)

26. SPS MD tonight beam = inclined hybrid crossing! (K. Takayama et al.)
BBLR1 (rotated) & BBLR2 (45 degree)
beam
J.-P. Koutchouk
= inclined hybrid crossing!
(K. Takayama et al.)

27. (3) electron lenses

28. V. Shiltsev, EPAC’04

29. Fermilab Tevatron Electron Lens
V. Shiltsev
designed
to compensate
beam-beam
tune shift
of pbars
current: ~1-3 A
solenoid field:
3.5 T
rms radius:
0.66 mm
length: 2 m
in operation
since 2001

30. V. Shiltsev, EPAC’04

31. beam-beam compensation at Tevatron assumes installation
of two electron lenses with pulsed and variable currents
multiple purposes:
reduce bunch-by-bunch tune variation
compensation also at injection
reduction of intra-bunch tune spread
spring 2001 installation of 1st TEL
1st studies
tune shifts of DQ= with 3A e- beam
10-kV e- gun, 3.4 mm diameter, 2 m interaction length
critical factors: a) transverse profile of e- beam (various guns)
uniform, Gaussian, flat-top + Gaussian tails
b) electron-beam alignment

32. 2002: TEL was found to be effective for dc beam cleaning,
in routine operation (avoids quenches when beam is dumped)
2003: 1st demonstration of beam-beam compensation
TEL slowed down emittance growth of pbar bunch
when ‘scallops’ were present
effect of TEL not well controlled
precise centering of pbar and e- beam difficult since different
e- and pbar bunch length yield different BPM response
pbar orbits can migrate by 0.5 mm over 12 hours
fabrication of 2nd TEL in collaboration with IHEP (Protvino) is
underway, completion end of 2004

33. V. Shiltsev, EPAC’04

34. V. Shiltsev, EPAC’04

35. V. Shiltsev, EPAC’04

36. (4) electromagnetic wires

37. Long-Range Beam-Beam Compensation for the LHC
To correct all non-linear effects correction must be local.
Layout: 41 m upstream of D2, both sides of IP1/IP5
current-carrying
wires
Phase difference between BBLRC &
average LR collision is 2.6o
(Jean-Pierre Koutchouk)

38. simulated LHC tune footprint with & w/o wire correction
Beam
separation
at IP
(Jean-Pierre Koutchouk, LHC Project Note 223, 2000)

39. motivation SPS wire LHC beam diffusive aperture
long-range collisions cause a large diffusion and proton losses - effect of SPS BBLR wire resembles LHC long-range collisions -hence we can study LHC long-range collisions and their compensation in the SPS
simulation with WSDIFF
simulation with WSDIFF
SPS wire
LHC beam
diffusive aperture

40. early in 2002 a single 1-wire BBLR was installed in the SPS
each BBLR consists
of 2 units, total length:
2x =1.85 m

41. 2002/2003 MDs
the first BBLR allowed us to model the effect of SPS long-range collisions in the SPS and to benchmark the simulations
6 MDs were performed in 2002 and 2003
we observed lifetime reduction, beam losses, and emittance shrinkage, tune shift, orbit distortion, enhanced decoherence; we scanned both the beam-wire distance and the wire current,…
results were published in two reports: CERN-AB ABP and LHC-Project-Report-777

42. y orbit change x tune change y tune change
measured vs. predicted changes in orbit & tunes
induced by SPS beam-beam compensator
y orbit
change
J. Wenninger
x tune change
y tune
change

43. emittance shrinkage , dynamic aperture, lifetime drop
& diffusion induced by SPS beam-beam compensator
WEPLT045
J. Wenninger, J.-P. Koutchouk, et al.
aperture vs.
wire current
aperture vs. separation
diffusion measurement
lifetime vs. separation

44. for 2004 two novel 3-wire BBLRs were built; one of these is installed in LSS5 close to the 1st BBLR; the 2nd new device is being refurbished, after a vacuum leak, for installation in LSS2

45. LSS5
G. Burtin
remotely
movable
in Y by
5 mm!
installed since July

46. SPS BBLR MDs in 2004 29./30.07.04 compensation of BBLR1 by BBLR2,
mismatched emittance to 4-6 mm,
scan compensator current and position,
tune scans around SPS & LHC tunes
test of crossing schemes HH, VV,
and “pseudo-VH”, tune scan around
LHC tunes
compensation of BBLR1 by BBLR2,
“scaled” with original emittance at
smaller distance

47. compensate BBLR1 by BBLR2
measured BBLR compensation efficiency vs. working point
- scan around LHC tunes
we scanned QY w/o BBLRs, with BBLR1 only, and with BBLR1 & BBLR2
3rd
10th
7th
4th
nearly perfect
compensation
what happens here?
compensate BBLR1 by BBLR2

48. ~69 min. ~36 min. ~61 min. LHC tunes
‘scaled’ experiment: measured beam lifetime
~69 min.
~36 min.
~61 min.
LHC tunes
J.-P. Koutchouk

49. (5) super-bunches

50. superbunch hadron collider
K. Takayama et al.,
PRL 88, (2002)
x-y crossing or 45/135 degree crossing

51. for large Piwinski parameter qsz /(2s
for large Piwinski parameter qsz /(2s*)>>1, with alternating crossing at two collision points
SuperLHC
F. Ruggiero, F. Zimmermann, G. Rumolo, Y. Papaphilippou,
“Beam Dynamics Studies for Uniform (Hollow) Bunches or Super-bunches in the LHC : Beam-beam effects, Electron Cloud, Longitudinal Dynamics, and Intrabeam Scattering”, RPIA2002
“Beam-Beam Interaction, Electron Cloud and Intrabeam Scattering for Proton Super-bunches”, PAC2003

52. luminosity increase in the long-bunch regime
factor
10 gain
possible
for LHC
upgrade
constant DQ
LHC, L0=2.3x1034 cm-2 s-1, q0=300 mrad
relative increase in LHC luminosity vs sz (qc) for Gaussian
or hollow bunches, maintaining DQtot=0.01 with alternating x-y
crossing at 2 IPs [F. Ruggiero & F.Z., PRST-AB 5, (2002)]

53. (6) crab cavities

54. Crab Cavity KEK variable symbol KEKB SuperLHC beam energy Eb 8 GeV
Courtesy K. Hosoyama, K. Ohmi
Crab Cavity
R. Palmer, 1988
K. Oide, K. Yokoya
combines all advantages of head-on
collision and crossing angle
KEK
variable
symbol
KEKB
SuperLHC
beam energy Eb
8 GeV
7 TeV
rf frequency
fcrab
508 MHz
1.3 GHz
crossing angle Qc
11 mrad
8 mrad
IP b b*
0.33 m
0.25 m
cavity b
bcav
100 m
2 km
kick voltage
Vcrab
1.44 MV
46 MV
phase tolerance Df
0.06 mrad

55. conclusions

56. alternating crossing cancels linear tune shift, but
might act de-stabilizing
active beam-beam compensation programme
in progress for Tevatron and LHC
TEL promising, but conditions difficult to control
wire compensation of long-range collisions at the LHC will allow for smaller crossing angles and/or higher bunch charges; experimental demonstration in the SPS;
pulsed wire desirable for selective correction of PACMAN bunches
flat long or super-bunches may yield higher luminosity for the same bb tune shift
crab cavities alternative option for large qc

57. thank you for your attention!