An Introduction to the

Physics and Technology

of e+e- Linear Colliders

Lecture 7: Beam-Beam Effects

Nick Walker (DESY)

DESY Summer thStudent Lecture

USPAS, Santa Barbara, 16 -27
31st June,
th
2003
July 2002
Introduction
• Beam-beam interaction in a linear collider is
basically the same Coulomb interaction as in a
storage ring collider. But:
– Interaction occurs only once for each bunch (single
pass); hence very large bunch deformations permissible.
– Extremely high charge densities at IP lead to very
intense fields; hence quantum behaviour becomes
important
• Consequently, can divide LC beam-beam
phenomena into two categories:
– classical
– quantum
 Introduction (continued)
• Beam-Beam Effects
– Electric field from a “flat” charge bunch
– Equation of motion of an electron in flat bunch
– The Disruption Parameter (Dy)
– Crossing Angle and Kink Instability
– Beamstrahlung
– Pair production (background)

Ebeam x y x y z Ne

GeV mm mm nm nm m 1010
NLC 250 8 0.1 245 2.7 110 0.75
TESLA 250 14 0.4 550 5 300 2.0
CLIC 1500 8 0.15 43 1 30 0.42
 Storage Ring Collider Comparison

re N b  x, y
Linear beam-beam tune shift  x, y 
2  x , y ( x   y )

Putting in some typical numbers (see previous table) gives:

 x 0.54
 y 1.44

Storage ring colliders try to keep  x , y  0.05

 Electric Field from a Relativistic Flat Beam

• Highly relativistic beam E+vB  2E

• Flat beam x  y (cf Beamstrahlung)
• Assume
• infinitely wide beam with constant density per
unit length in x ( (x))
• Gaussian charge distribution in y:
• for now, leave (z) unspecified

1  1  y 
2
 1
 ( y)  exp       ( x) 
2  y  2   y   2  x
 
 Electric Field from a Relativistic Flat Beam

q
Use Gauss’ theorem: E ds 
s 0

y
qN  ( x)  ( z )xz
E y ( y, z )xz 
0   ( y ')dy '
y '0

qN  y 
E y ( y, z )  Erf    ( z)
2 2  0 x  2 
 y 

Assuming Gaussian distribution for z, the peak field is given by

qN
Eˆ y 
4 0 x z
 Electric Field from a Relativistic Flat Beam

E y MV/cm
 x 500 nm 2000

 y 5 nm
 z 300 μm 1000

N 1010
 10 5 5 10
y / y
 1000

 2000
z 0

Note: 2Ey plotted

Assuming a Gaussian distribution for (z)
 Electric Field from a Relativistic Flat Beam

E y MV/cm
 x 500 nm
 y 5 nm
1500
 z 300 μm
N 1010 1000

effect of x width
500

z 0
10 20 30 40 50
y / y

Assuming a Gaussian distribution for (z)

 Equation Of Motion
F = ma: 2qE y ( y, t )
y (t ) 
m0

Changing variable to z: y (t ) c 2 y( z )

2qE y ( y, z )
y( z ) 
m0 c 2
 y( z) 
2
q N Erf    z (2 z ) why
 2 
  y 
z(2z)?
2  0 x mo c 2

q2 2 2 Nre  y( z) 
re  y( z )  Erf    (2 z )
4 0 m0 c 2  x  2  z
 y 
Linear Approximation and the Disruption Parameter
4 Nre  (2 z )
Taking only the linear part of y( z )  y( z)
 x y
the electric field:   
k2 (z)

Take ‘weak’ approximation: 4 Nre 

y  y0   z (2 z )dz
 x y 
y(z) does not change during
interaction y(z) = y0 2 Nre 1
 y0  y0
 x y f
1 2 Nre
Thin-lens focal length: 
f  x y

Define Vertical Disruption

Parameter  2 Nre z 2 Nre z
Dy  z  exact: Dy 
  x   y  y
f  x y
 Number of Oscillations
Equation of motion re-visited:
2 Dy
y( z )   z (2 z ) y ( z )
z
Approximate (z) by rectangular distribution with same RMS
as equivalent Gaussian distribution (z)
half-length!
Dy 3
0.4 y( z )  y ( z ); z  y
0.3
1 3 2
z
2
2 3 z
Dy 3k z
0.2 2
k  
0.1 3 z2 2
0 1
3 2 1 0 1 2 3 3 4 Dy
 3 z 3 z   0.21 Dy
2
 Example of Numerical Solution
10 y / nm
N 2 10 40 green: rectangular approximation
 x 500 nm
20
 y 5 nm
 z 300 μm 0

E 250 GeV
20 black: gaussian

Dy 27.7 40

 1.1 500 0
z (ct )/ μm
500 1000 1500
Pinch Enhancement
• Self-focusing (pinch) leads to higher luminosity
for a head-on collision.

 D 3
  0.8  x , y 
H Dx , y 1/ 4
1  Dx , y 
x, y
 1  D 3  
x, y  
ln  Dx , y 
 1  2 ln 
 z




‘hour glass’ effect

Empirical fit to beam-beam simulation results

Only a function of disruption parameter Dx,y

The Luminosity Issue: Hour-Glass
3 3

2 2

1 1

0 0
Y

Y
1 1

2 2

3 3
2 1 0 1 2 2 1 0 1 2
Z Z

 = “depth of focus”
reasonable lower limit for
 is bunch length z
Luminosity as a function of y

L (cm 2s 1 )
51034

 z 100 m  BS  1
z
41034

31034  z 300 m

21034 500 m nb N 2 f
L
4 x y
700 m
11034 900 m

200 400 600 800 1000

 y ( m)
Beating the hour glass effect

Travelling focus (Balakin)

• Arrange for finite chromaticity at IP (how?)

• Create z-correlated energy spread along the bunch (how?)
*
z   y
Beating the hour glass effect

Travelling focus

• Arrange for finite chromaticity at IP (how?)

• Create z-correlated energy spread along the bunch (how?)

 z   y*
Beating the hour glass effect

Travelling focus

• Arrange for finite chromaticity at IP (how?)

• Create z-correlated energy spread along the bunch (how?)

ct
Beating the hour glass effect

foci ‘travel’ from z = 0 to z =  3 z

chromaticity:
3 z 3 z f y  y

travelling focus:
t= 0
3  2ct NB: z correlated!
z

z
 RMS 
2 y
t> 0 z 2 y RMS
2
f =  ct
 z2

2 y
 1 2 3

6
4 5

The arrow shows position of focus for the read

beam during travelling focus collision
 Kink Instability
1
Simple model: ‘sheet’ beams with:  x,z 
2 3 x , z
Linear equation of motion becomes

2 Dy 3
y( z )  y ( z ); z  z
6 z2
2
Need to consider relative motion of both beams in t and z:
2
 
  c y
 1 (t , z )   2
0 ( y1  y2 )
 t z  2 2 Dy
2
 
2
0 c 2
  6 z
  c y
 2 (t , z )  2
0 ( y1  y2 )
 t z 

Classic coupled EoM.

 Kink Instability
Assuming solutions of the form

y1(2) a1(2) exp  i kz   t 

and substituting into EoM leads to the dispersion relation:

 2 c 2 k 2   02  4 02 c 2 k 2   04

Motion becomes unstable when 2 0, which occurs when

2 0
k 
c
Kink Instability
Exponential growth rate:
1
 4 2 c 2 k 2   4   2  c 2 k 2 
2

 0 0 0

3 0 
Maximum growth rate when k  ;  i 0
2 c 2
3 z 3 z
Remember!  t   t  3 z / c
2 c 2 c

Thus ‘amplification’ factor of an initial offset is:

1  
1
  0 t  4
0.6 Dy
exp   exp    Dy  e
 2   2  2  

For our previous example with Dy ~28, factor ~ 3

 Kink Instability
y 0.1 y
Dy 20
 z 300 μm
 Pinch Enhancement

L Lgeom H H= enhancement factor

results of
simulations:

  2y 
exp  
 4 2 
 y 
 High Dy example: TESLA 500
Dy 24
 Disruption Angle
z
Remembering definition of Dy Dy 
f
The angles after collision are characterised by
Dx x Dy y 2 Nre 2 Nre
0    
z z  ( x   y )  x
Numbers from our previous example give  0 467 μrad
OK for horizontal plane where Dx<1
For vertical plane (strong focusing Dy > 1), particles oscillate:
previous linear approximation:
Dy 3 Dy 0
y( z )  y ( z ); z  y  y  67 μrad
3 2
2 1 1 1
z 3 z 4
3 Dy4 2
Disruption Angle: simulation results

250
1000

200
800

150 600

100 400

50 200

0 0
- 400 - 200 0 200 400 - 400 - 200 0 200 400

horizontal angle (rad) vertical angle (rad)

Important in designing IR (spent-beam extraction)

Beam-Beam Animation Wonderland

Animations produced by A. Seryi using the

GUINEAPIG beam-beam simulation code
(written by D. Schulte, CERN).
Examples of GUINEAPIG Simulations

NLC
parameters
Dy~12

Nx2
Dy~24
Examples of GUINEAPIG Simulations

NLC parameters
Dy~12

Luminosity
enhancement
HD ~ 1.4

Not much of an
instability
Examples of GUINEAPIG Simulations

Nx2
Dy~24

Beam-beam
instability is
clearly pronounced

Luminosity
enhancement is
compromised by
higher sensitivity
to initial offsets
 Beam-beam deflection

Sub nm offsets at IP cause large well detectable offsets (micron

scale) of the beam a few meters downstream
Beam-beam deflection
allow to control collisions
Examples of GUINEAPIG Simulations
Examples of GUINEAPIG Simulations
Beam-Beam Kick
Beam-beam offset gives rise to an equal and opposite mean
kick to the bunches – important signal for feedback!
For small disruption ( D1) and offset (y/y<<1)
1 y
 bb   0
2 y
F ()  

For large disruption or

offset, we introduce the
form factor F:

1
 bb   0 F  y /  y 
2
Long Range Kink Instability & Crossing Angle

To avoid parasitic bunch interactions in the IR, a horizontal

crossing angle is introduced:

ctb c
X angle c
2

ctb
l
2

2 Nre
parasitic beam-beam kick: r   r  x , y
r
Long Range Kink Instability & Crossing Angle
y IP
1
   0 F (     )

ip
2
small vertical offset
e 
( = y/y) gives
e+
rise to beam-beam
kick
y
l
 resulting vertical
offset at parasitic
crossing gives next
incoming bunches
 additional vertical

 2 Nr l ip Nr  F (  
  
) kick: IP offset
  e
 e 0
increases
 X 2
 lc2
instability
 Long Range Kink Instability & Crossing Angle
offset at IP of k-th bunch:  k  k   k
(k  l )tb c
distance from IP to encounter with l-th bunch (l < k) llk 
2
contribution from encounter with l-th bunch:

 k  k ,0  llk,lk /  y
 Nre  0
 k ,0  F (l )
  y
2
c

 k  k   k
2 Nre 0 NB. independent
 k ,0  F ( l )
c2 y of llk
2
k1
2 Nre 0   x / z 
total offset:  k  k ,0  C  F ( i ); C  Dx Dy  
i 1 c2 y   c 
Long Range Kink Instability & Crossing Angle

Assuming all bunches have same initial offset:  k ,0  0

For small disruption and small offsets, F ( i )  i

 k (1  C ) k  1  0
2
  x / z  k
since C Dx Dy   1 1  (k  1)C
 c  0

thus (mb  1)C 1 where mb = number of interacting bunches

example:
Dx,y = 0.1, 10
x = 220 nm C = 0.012  mb < 80 for factor of 2 increase
z = 100 m
c = 20 mrad
 Crab Crossing
x
 x , projected   x2  c2 z2
c z
20mr 100μm 2μm
factor 10 reduction in L!

use transverse (crab) RF

x cavity to ‘tilt’ the bunch at IP

 2 z  2 Vˆ z
V ( z ) VˆRF sin    RF

RF kick    
VˆRF
c 4  cav  ip z
RF
Beamstrahlung
Magnetic field of bunch B = E/c
2 Emax qN
Peak field: Bmax   1160Tesla
c 2 0 c x z

From classical theory, power radiated is given by

cC E 4 5 -3
P  ; C 8.85 10 m..GeV
2  2

For E = 250 GeV and z = 300m:

1 cB
 1.4 m -1
 E
z
Erad P t P 32 GeV E/E ~ 12%
c
note:  z /  200 μrad 1/  2 μrad
 Beamstrahlung

Most important parameter is 

2  c e 2 2B e

3 E


  2
Bs m0
F
 2
p 

c critical photon frequency

e Compton wavelength
 local bending radius
B beam magnetic field
Bs Schwinger’s critical field (= 4.4 GTesla)
F em field tensor
p electron 4-momentum
 Beamstrahlung: photon spectrum
NOT classical synchrotron radiation spectrum!
Need to use Sokolov-Ternov formula:
10

y2  10
5

FBS  K 5 3 ( )d   K 2 3 ( )
 1 y 1
 1

0.5
classical quantum theoretical
  1/ 3
0.1
  0.1
y 0.05

E
2  y     1  0.01
       0.001 0.005 0.01 0.05 0.1 0.5 1

3  1  y    c   1  y  y  / E
 Beamstrahlung Numbers
5 Nre e
average and maximum   avr 
6  z ( x   y )
 max 2.4  avr

  z   avr
photons per electron: n 2.54  
 
 e  1   avr
2/3

E   z   2avr
average energy loss:  BS   1.24  
E   2/3 2
 e   1  (1.5 avr ) 

BS and n (and hence ) talk directly to luminosity spectrum and

backgrounds
Beamstrahlung energy loss
In lecture 1, we used the following equation to derive our
luminosity scaling law

ere3  Ecm  N2
 BS 0.86 2  
2m0c   z  ( x   y )2

Now we have this:

  z   2avr
 BS 1.24  
  2/3 2
 e   1  (1.5 avr ) 

5 Nre e
with  avr 
6  z ( x   y )

under what regime is our original expression valid?

 Luminosity scaling revisited
 y*  z

 BS
low beamstrahlung regime 1: L  Pbeam
 y ,n

 BS
3
2

high beamstrahlung regime 1: L  Pbeam

 z  y ,n

homework: derive high BS scaling law

Beamstrahlung Numbers: example

E 500GeV  avr 0.28 E

 ,c 210 GeV
 x 220 nm  max 0.66
 y 2 nm n 1.3
 x 110μm  BS 7.5%
N 0.75 1010

TESLA
 Why Beamstrahlung is bad
• Large number of high-energy photons interact with electron
(positron) beam and generate e+e pairs
– Low energies (0.6), pairs made by incoherent process
(photons interact directly with individual beam particles)
– High energies (0.6100), coherent pairs are generated by interaction of
photons with macroscopic field of bunch.
– Very high energies (100), coherent direct trident production
e  e  e+ e

TESLA 0.5TeV  ~ 0.06

NLC 1TeV  ~ 0.28
CLIC 3TeV  ~ 9

• Beamstrahlung degrades Luminosity Spectrum

 Pair Production
• Incoherent e+e pairs 0.6
– Breit-Wheeler:   e e 
– Bethe-Heitler: e   e e  e 
– Landau-Lifshitz: ee  ee ee
• Coherent e+e pairs 0.6100
– threshold defined by  2 B
 2
10
m0 c Bs
5
 10

 1   1
E
1

0.5

0.1
  1/ 3
for >1  / E ~ O(1)   1
 0.1
0.05

0.01
for intermediate colliders (Ecm<1TeV),
incoherent pairs dominate
0.001 0.005 0.01 0.05 0.1 0.5 1

y  / E
 Pair Production
e+e pairs are a 1

potential major source

of background

PT (GeV/c)
0.1
Most important: angle
with beam axis () and
transverse momentum
0.01
PT.
0.1 1
ld  (rad)

rd d
pairs curl-up (spiral) in
PT
vertex solenoid field of r  rd
detector
detector cBz
Summary
• Single pass collider allows us to use very strong
beam-beam to increase luminosity
• beam-beam is characterised by following
important parameters:
– Dy = z/f defines pinch effect (HD), kink
instability, dynamics
  QM effects, backgrounds,
– BS [=f(av)] energy loss, lumi spectrum
• strong-strong regime requires simulation
(e.g. GUINEAPIG). Analytical treatments limited.