CRAB CROSSING IN A LARGE HADRON COLLIDER
Jie Wei Brookhaven National Laboratory, Upton, New York 11973, USA
ABSTRACT efits of a relatively small bunch spacing, the beams must cross at an angle of about 70 r to avoid more than one bunch-bunch Since its invention by Palmer[1] in 1988, crab crossing has collision in each experimental straight section. Such a non- been explored by many people for both linear and storage ring zero crossing angle causes a degradation in luminosity. Con-
colliders to allow for an angle crossing without a loss of lumi-
2 sequently, it demands a short bunch length ( cm) which nosity. Various crab crossing scenarios have been incorporated can only be achieved with a large rf voltage (100 MV) when op- in the design of newly proposed linear colliders and B-factory erating at a frequency of 379 MHz. The problem can be solved projects. For a hadron collider, this scheme can also be em-
by crab crossing the two counter-circulating proton beams to ployed to lower at the interaction point for a higher luminos- make them collide head-on. ity. In Section II of this paper, we first summarize the princi- In this paper, we first review the principle and operational re- ple and operational requirements of three crabbing schemes for quirements of various crab crossing schemes for storage ring storage ring colliders. A Hamiltonian formalism is developed in colliders. A Hamiltonian formalism is developed to study the Section III to calculate the emittance growth and crabbing qual- dynamics of crab crossing and the related synchro-betatron cou- ity degradation produced by the errors in voltage and betatron- pling. Requirements are obtained for the operational voltage phase matching of the crab cavities. The results are applied to a and frequency of the crab cavities, and for the accuracy of volt- conceptual design of angle crossing in the proposed high-field age matching and phase matching of the cavities. hadron collider. Conclusions and a discussion are given in Sec- For the recently proposed high-field hadron collider,[2, 3] a
tion IV.
0:1 deflection crabbing scheme can be used to reduce from m to 0.05 m and below, without a loss of luminosity due to angle II. CRAB CROSSING SCHEMES crossing. The required voltage of the storage rf system is re- The goal of crab crossing in a storage ring collider is to make duced from 100 MV to below 10 MV. With the same frequency the two counter-circulating angle-crossing beams collide head- of 379 MHz operating in a transverse mode, the required voltage on at the IP without sacrificing beam quality and luminosity of the crab cavities is about 3.24.4 MV. The required accuracy lifetime. In this section, we present three schemes: deflection
of voltage and betatron-phase matching is about 1%.
crabbing,[1] dispersive crabbing,[4] and crabbing.[5] The z I. INTRODUCTION angle crossing is assumed to occur in the horizontal (x– )plane. The subject of angle crossing at the interaction point (IP) of A. Deflection Crabbing a storage ring collider has been studied for many years with the With the deflection crabbing scheme, as shown in Fig. 1, two realization that the synchro-betatron resonance induced by the transverse deflectors (rf cavities operating at their transverse crossing angle severely limits the luminosity. In 1988, a beam- beam collision scheme was invented by Palmer[1]to allow a large crossing angle for a linear collider without a loss of lumi- p p nosity. The Palmer scheme (or deflection crabbing scheme) em-
ploys transverse rf deflectors placed at locations where the beta-
90 tron phase advance is from the IP. Both colliding bunches are tilted by the cavities by half the crossing angle at the IP α so that they collide head-on. Subsequently, several alternative schemes[4, 5, 6, 7] were also introduced to apply crab crossing to storage ring colliders. Recently, various crab crossing sce- IP narios have been incorporated in the design of newly proposed linear colliders[8] and B-factory projects.[9] The design goal of a high-field hadron collider[2, 3, 10] is
a 50 TeV storage ring that can achieve a peak luminosity of
34 2 1 10 cm s with a small number of interactions per bunch- Figure 1: Schematic view of deflection crab crossing of two
bunch collision. With an experimental drift space of 25 m and counter-circulating beams. The crossing angle between the two a focusing strength of 360 T/m at the triplet quadrupoles, a of beam trajectories is . 0.1 m can be achieved at the energy of 50 TeV. To enjoy the ben-
346 modes) are positioned on each side of the IP, preferably at high-
Table I: Comparison between nominal and crabbing operations.
locations with betatron phase advances of 90 from the IP.
s 90
At an azimuthal location 0 with a betatron phase of from
the IP, the particle receives a kick in the horizontal direction x, Quantity Unit Nominal Crab I Crab II
p=p
along with a change in momentum ( ), m 0.1 0.1 0.05
r707097
K R h
0 c
V 0
rf MV 100 10 10
x = z ; sin
h R
V c
cr ab MV 0 3.2 4.4
(1)
f
cr ab
MHz – 379 379
h
c
z mm 22 41 41
= K x cos z ;
0
34 2 1
R L
ini 10 cm s 1.1 1.1 2.2
0
dx=ds z
where x , is the longitudinal displacement from the
= E=c C =2R rf bucket center, p is the momentum, is the
which results in an offset in the betatron closed orbit, K
circumference, and the strength 0 of crabbing is related to the
V 0
peak voltage c of the cavity by
x = ;
0 0
(7)
qeV h
c0 c
V h
= ; 0 0 K where and are the peak voltage and harmonic number of
0 (2)
RE the rf cavity. At the IP where the dispersion is zero, the bunch
=2
qe h
with the electric charge of the particle, and c ,aninteger,the is tilted by an angle , with
harmonic number of the cavity. If the crabbing wavelength is s
x 0
much larger than the bunch length, i.e.,
:
tan (8)
2 z
0
2R
;
z (3)
h
c V
The voltage 0 required is thus
0 s
the kick in x is approximately linearly proportional to the dis-
2
RE
0 z
placement z . At the IP, this kick results in a -dependence of
qeV = tan :
0 (9)
h 2
0 0
the horizontal displacement x. Thus, the bunch can be tilted by
p
=2 x z tan( =2) K 0 an angle in the – plane with 0 ,
A second cavity located at a place with a betatron phase of
where and 0 are the functions at the IP and the cavity
+180 from the IP operates at the same voltage as the first one s
( 0 ), respectively. The voltage required is thus
to restore the particle motion. The dispersion and function at
RE the second cavity needs to be the same as at the first one.
p
qeV = tan : 0
c (4)
2
h The dispersive scheme usually requires a large dispersion at
c 0
the cavity locations along with a large operating rf voltage. For
Obviously, for given and , a high- location and a high the high-field collider, the rf cavities are assumed to be posi- h
operational frequency (or harmonic number c ) is preferred for
200
tioned at places with 0 m. Even with a high disper-
the cavity, provided that the condition Eq. 3 is satisfied. The
=10
sion of 0 m, an impractically high voltage of more than
s +90
second cavity located at 1 with a betatron phase of from
=70 1 GV is required for a r crossing. Furthermore, since the IP needs to operate at a voltage the required dispersions at the two cavities are the same, addi-
s tional dipoles are needed to make the dispersion at the IP zero.
0
= V
V The longitudinal slippage between the two cavities produced by
1 c0
c (5)
1 these dipoles will inevitably degrade the crabbing accuracy.
to restore the particle motion to its unperturbed state. C. Crabbing
For the high-field hadron collider, the crab cavities can be Another scheme for crab crossing is not to employ any ded-
50 positioned at places with 0 km. With a frequency of icated cavities. Instead, the storage rf cavity is placed near the
379 MHz, the required voltage for the transverse cavities is be- IP,and the dispersion at the IP is made non-zero. With a peak
tween 3.2 and 4.4 MV for a crossing angle between 70 and
V h
voltage rf and harmonic number , the rf cavity changes the
97 r,asshowninTableI. momentum of the particle,
B. Dispersive Crabbing qeV h
rf
= sin z : (10)
An alternative scheme for crab crossing is to employ two reg- 2
E R
ular (instead of transverse deflecting) cavities located at disper-
sive regions, where the betatron phase advances are 180 from Due to the dispersion , the horizontal displacement at the
IP resulting from this momentum change produces a tilt in the
the IP. At the first cavity where the dispersion is 0 , the particle
tan( =2) =z
receives a kick in momentum, bunch with . The voltage required is thus
2
qeV h
RE
0 0
= sin z ;
qeV = tan :
(6) rf (11)
2
E R
h 2
347 V
For a moderate voltage rf , this scheme often requires a large average the contribution of the storage rf cavities over the cir-
dispersion at the IP. Such a dispersion effectively enlarges cumference, and define
2
the horizontal beam size at the IP, which inevitably causes a 2
h ! qeV
s
; and C ; C W
degradation in luminosity. Furthermore, the dispersion mod- (15)
2
2E
ulates the beam-beam interaction to give nonlinear synchro- h
2
= h =i 1= betatron coupling. where x is the slippage factor. The Hamil-
For the high-field collider, we assume that an rf cavity of tonian (Eq. 13) can be expressed in terms of the action-angle
J J J s
x y y z z
50 MV is located near each one of the two IPs. To achieve variables ( x , , , , , ; ) using a canonical transfor-
=70
an angle of r, the needed is more than 4 m. With a mation. The new variables are related to the old ones as
s