© 1993 IEEE. Personal use of this material is permitted.

However, permission to reprint/republish this material

for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers
or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.

Longitudinal Beam-Beam Effects in Circular Colliders

M. Hogan and J. Rosenzweig
Departmentof Physics
University of California, Los Angeles, CA 90024

Abstract derived from a form of the Panofsky-WenzelTheorem which is

The longitudinal beam-beaminteraction, which can lead generalized to include fields arising from free chnrges [see
to incoherent heating, synchrobctatron coupling, and coherent Appendix],
longitudinal instabilities in circular colliders, is examined. @@,.> .
%(APr)=~ (2)
This analysis discussestwo types of energy kicks, those due
to the transverse particle motion coupling to the electric We will employ the Panofsky-Wenzel theorem method in this
portion of the transverse kick, and those derived from the paper, as it is simpler and more powerful than doing the
inductive electric field induced near the interaction, which is straight forward calculation, especially for off-axis pnrticles.
obtained from the transverse kick through use of a
generalization of the Panofsky-WenzelTheorem. Implications II. THE LONGITUDINAL BEAM-BEAM INTERACTION
for low energy e+e- colliders (cp& B factories) with beams
crossing head-on, and at finite angles, with and without crab The longitudinal kick due to transversemotion is the sum
crossing,are discussed. of the individual kicks a particle receives traversing the
opposing bunch, calculated by integrating over the phasespace
of the opposing bunch. The resulting differential equation for
I. INTRODUCTION the transverse acceleration involves complex error functions
for elliptical beams.
Table 1: Notation usedin this paper. A focusedgaussinnbeamis describedby a time dcpendcnt
e electronchnrge ch‘argedensity and has associatedtransversecurrents given by
r radial position the continuity equation. By integrating the charge and current
X,Y transversecoordinates densities we obtain the corresponding scalar and vector
z longitudinal coordinate potentials, which intern describe associated electric and
c speedof light magnetic fields. The Lorentz force law may then be used to
s=z-ct beamcoordinate compute the instantaneous acceleration felt by a test particle
P particle momentum traversing these fields. The resulting equations contain very
, d/ds tedious integrals that may be solved numerically [2].
‘i rms. beam size in i dim. Both casesyield results that are not intuitive and do not
P Beta-function allow for an estimate of the size of these kicks which could
a Courant-Snyderamplitude determinethe importance of this analysis for circular colliders.
N # electrons in bunch Therefore we will now consider certain limiting cases.
* denotesevaluation at IP
A. Routd Becms
An ultra-relativistic particle with longitudinal
coordinates, has E-M fields which are ne,arly normal to the The energychange (A,!?) an off-axis test particle which travels
direction of motion, and may be approximatedas[l] al a nonzero angle with respect to the axis receives passing
E, =%(s-so) through an opposing beam is given by projecting the orbit of
r the particle onto the field of the opposing beam p,articles,as
where the notation is given in Table 1. Since particles do not given by Eq. (1). Assuming a round beam with a gaussian
all collide head on there is some longitudinal kick given distribution, Eq. (1) can be integrated to give the total energy
during the beam-beaminteraction. The net longitudinal kick changeper plunge:
can be obtained for a single particle by taking the projection of
the transverse fields from the opposing beam onlo the design -=-~[i-exp[-g)]~.,‘+y~~. (4)
orbit of the particle and integrating over the betntron phase
spaceof the opposing beam. The resulting energy kick can be Averaging over a betatron oscillation (which is assumedto be
thought of as arising from two sources:a) longitudinal fields, much shorter than a synchrotron oscillation), and expanding
and b) the work done by the transversemotion of the particle equation(4) for r < 0, reducesthis to
againstthe transversefields.
For beams that do not collide at the nominal intcrnction
point (IP), there is a time dependentbeta funcrion,p(s), and
the beam size variation gives rise to an inductive longitudinal
*E = Nc’P’(f)
2/q+)
ax + a,
2E ’
where we have introduced Ihe Courant-Snyder amplitudes of
field. The longitudinal momentumhasbeen previously de&cd the particle. Note that if rhe beams do not collide at the
(5)
[ 1
from a straight forward ret,ardedrelativistic calculation[2]. In nominal IP, implying that0 f 0, the energy kick averaged
the limit that the beams are shorter than p* (this limit is
over a betalron oscillation is non-vanishing.
assumedEd throughout this paper), these kicks can also be

0-7803-1203-l/93$03.00 0 1993 IEEE 3494

PAC 1993
 The energy kick due to the inductive field is computed Again, the transversedependenceis proportional to the current
using the Panofsky-WenzelTheorem: density. a variation which is canceled exactly by the energy
kick due to transversemotion derivable from Eq. 12 for small
4h) = 4*d (G) amplitude (y < 0;) particles. Note that this expression is
f3r as *
For round beamsthe transversekick is, from equation (4), smaller th‘anthe equivalent round beamformula (Eq. (10)) by a
factor of R = cry/o,. This factor is due to larger average
Ap, =2fff-[l-exp(-$)J. (7) distancesbetweenparticles (R/2) and weaker focusing in the x-
dimension (2).
Since
III. COHERENT BEAM-BEAM OSCILLATIONS
(8) Longitudinal beam-beam effects can drive a coherent
longitudinal oscillation. While this subject has been analyzed
has an s dependence,the longitudinal kick is obtained using before[l], it has never been understood that the longitudinal
Eqs. (7) and (8), and integrating the Panofsky-Wenzel beam-beamkick is nearly independentof x and y. The coupled
expression (taking the boundary condition Al>lr=- = 0) to equationsof motion for the be,amcentroids (s,,~)are
yield s,” + c&s, = (k)k,,(s, -s*)
(1%
AE = Ap,c = N;;w:i’ exp(-$1 (9) s*” + w,2s2 = (+-)k[,b& -s,)
; r
Here, &, E VL, / V;, the rf gradient V; = krfVrf, and V&,
Averaging over a betatron oscillation for r < 0, we have
(the effective beam-be‘amgradient) includes componentsdue to
LiE = Ne*P (f> [i-y]. both pnrallel and transversemotion of the beamparticles:
- 2/q+) Ne*
“G, = m. w-3
and the amplitude dependenceof this kick is canceled by the
term found in Eq. (5). Previous ,analysis[l] has not included The + (-) sign rcfcrs to operation above (below) transition.
this cancellation, which makes the energy kick more uniform Above transition, we obtain the dispersion relation
as a function of position. The resulting total energy kick for w = w,(l- 2k,;,)f.
round beamsis (17)
Thus the instability threshold, which occurs when w=O, is
Ne*p’ ($)
AE = Apzc = (10) given by 2V&, = V’,,-. If the vertical beta function is lowered
W($) ’
by a factor TJ,then VLh will increnseby a factor of q*. This
This energy kick, which is now correlated only to the relative
longitudinal position of the oncoming bunch, can contributeis a strong dependenceand may indicate trouble with higher
to a coherent instability, as discussedin section III. luminosity designs.
If the machine is run below transition, the frequency of
B. Flat Beams the coupled modebecomes
w= w,(1+2k,,)+ (18)
Other relevant aspectsof the beam-beaminteraction can and there is no possibility of this coherent longitudinal
be analyzed in the limit that 0, >> oy. In this quasi-one instability.
dimensionaJcasethe Panofsky-Wenzeltheoremreads
444 4~~~1 IV. IMPLICATIONS FOR LOW ENERGY COLLIDERS
dy =-r- (11)
A. (p Factories
The transversekick in a gaussianbenm is given by
Ap For the UCLA rp-factory design parameters,the expected
Y (12) energy kick for a one CT,particle is - 1 keV. The rf voltage
where gradient V,) = 1 MeV/m is quite small due to the quasi-
1 isochronous condition being employed. With N=1.6 x loll,
R=7, /3;= 4 mm, and emittancesof 1.1 x 10m6m-rad in both
(13)
x and y, the effective beam-beamgradient is
The contribution to the energy kick is now due to transverse Ne*
eV&, = - = 1 MeV/m . (19)
motion can be found in analogy to Eq. 4. Also, the inductive 2RP**
energy kick is, using Eq. (1 l), If the machine is run above transition, this gradient is
Ne*o,P;
AE=A&c= ($) longitudinally defocusing, and one would expect seriousbunch
(14) lengthening, since to fist order there is complete longitudinal
2crxpy(“i) defocusing,

3495
PAC 1993
 -3 Appendix: Generalized Partofsky-Wenzel Theorem
a,= &!!b?. =M (20)
o-20 i Tf 1 The Panofsky-Wenzel theorem gives a relationship between
This system is also above threshold for the longitudinal the integrated longitudinal and transverse momentum kicks a
instability by a factor of two in beam charge. p&articlereceives as it traversesa medium or device excited in
If the machine is designedto operatebelow transition, the the wnke of another particle [4]. This appendix will generalize
beams have stable coherent longitudinal motion. The benm- this theorem to include fields arising from free charges.
beameffectsreinforce the rf focusing because assuming that fields and potentials vanish at infinity, and that
the particle receiving the kick travels p,arallel to the z axis. In
-4
o* - l-v,:, =&. (21)
general. an electric field E may be described in terms of a
020 t ?f 1 scalnrpotential (0 ) and vector potential ( 2 ):
This implies that the bunches could be shortened allowing &-I&-pj)
shorterbunchesthan presentdesignsindicate. c ar
Inserting (22) into the Lorentz force equation
B. B Factories with Crab Crossing

Crab crossing schemes (Figure 1) may bc necessnry to

and noting that for a particle traveling parallel to the z-axis
provide the high luminosity’s (-3 x 103j) required for B-
factories. ‘~l)xB=Poix(Vx~)=~(PbA;)-Pb~ (24)
1 I
I we obtain the following expression for W. the force per unit
electrons positrons
chargeq:
w=-f~-~~-[p(B~a,)+ip,~]. (25)
By noting that
~[~+vb$p$$$, (26)

equation (25) can be rewritten as

c I
*p=;;
[
Figure 1: Schematic representation of crab crossing. Crab Outside of region R. i = 0, leaving
pb$+r@bA;-@)
z 1
dz. (27)

cavities apply time dependentrf kicks which tilt the bunches.

After the collision, another set of crab cavities kick the beams Ap= t[s$l”; -.!,]. (28)
back to there original orientations.
The longitudinal beam-beam interaction could become Since it can be derived from a potential, Ap satisfies the
important in this type of scheme.Recall that the energy kick relation
due to transverse motion for round beams contains terms vx(Ajj)=O
proportional to xx’,yy’. For crab crossing the angle X’ is (29)
In shorthand notation, this can be written as
now essentially to the crossing angle. The resulting energy
kick may again lead to large longitudinal effects. This issue ?,(A/,Z) = y. 00)
needsto be investigatedfurther. i

V. CONCLUSION
This paper has analyzed two types of longitudinal kicks
arising from the longitudinal benm-beaminteraction: thosedue
to the transverse particle motion coupling to the transverse
portion of the electric kick, and those derived from the
inductive electric field induced near the interaction point.
These effects may become important in low energy or high
luminosity colliders (cp& B factories) since they may lead to
coherent longitudinal instabilities. These effects can be
minimized by the use of flat beams. In addition, if coherent
instabilities becomea problem, it may be necessaryto operate
below transition. The effect on low energy colliders such as B-
factories which may utilize crab crossing to improve
luminosity ncedr to be addressedfurther.
REFERENCES
[l] V.V.Danilov, et. al., Longitudind Effects in Beam-Beam
Itltercrctiorl for an Ultru-High Luminosity Regime, in Be&am-
Beamand Beam-RadiationInteractions p.l-10, Eds. C. Pelleg-
rini, J. Rosenzweigand T. Katsouleas,(World Scientific
Publishing Co., Singapore, 1991)
[2] G. Jacksonand R. H. Seimann, Nucl. Instr. and Meth.
A286 (1990) p. 17-31.
[3] W. K. H. Panofsky and W. A. Wcnzcl, Rev. Sci. Inst.,
27. 967, 1956.

PAC 1993