Synchrotron Based Proton Drivers
Weiren Chou
Fermi National Accelerator Laboratory
P.O. Box 500, Batavia, IL 60510, USA
Abstract. Proton drivers are the proton sources that produce intense short proton bunches. They have a wide range of
applications. This paper discusses the proton drivers based on high-intensity proton synchrotrons. It gives a review of the
high-intensity proton sources over the world and a brief report on recent developments in this field in the U.S. highenergy physics (HEP) community. The Fermilab Proton Driver is used as a case study for a number of challenging
technical design issues.
difficult to design, build and operate than an
accumulator. The hardware is more challenging and
the reliability is not as high.
INTRODUCTION
Intense proton sources have been around for years.
At this time, the highest beam power from a
synchrotron is 160 kW at the ISIS (0.8 GeV) at the
Rutherford Appleton Laboratory (RAL) in England,
from an accumulator is 64 kW at the PSR (0.8 GeV) at
the Los Alamos National Laboratory (LANL) in the
U.S., and from a cyclotron is 1 MW at the PSI (0.59
GeV) in Switzerland. A proton driver differs from
these sources in the following aspects: (1) The beam
energy E is higher; (2) The bunch length a is shorter;
(3) The beam power P is larger. These differences
come from the requirements of physics experiments, in
particular, the neutrino oscillation experiments.
Typical parameters of a proton driver are: E > 4 GeV,
a < 3 ns (rms), P > 1 MW. When the proton beam
energy is below 4 GeV, n/[i yield from a carbon
target would be too low to be useful. When the proton
bunch length is longer than 3 ns, the production rate
(i.e., number of TC/JI particles per unit proton beam
power) and n+/ri or ji+/|^t" polarization ratio would be
uneconomical. Because the physics case is strong and
the capital cost is modest (less than 1/10 of the cost of
a linear collider), proton drivers have attracted
worldwide attention. A recently issued U.S. HEPAP
Sub-Panel Report identified such a facility as a
possible candidate for a construction project in the
U.S. starting in the middle of this decade.1
There are many similarities between the two types
of proton drivers, in particular in the linac and linac
front-end part. However, the design of a synchrotron
and an accumulator is quite different. This paper will
focus on the synchrotron-based proton drivers. A
paper discussing linac-based ones can be found in Ref.
2.
OVERVIEW OF HIGH INTENSITY
PROTON SOURCES
Table 1 is a survey carried out during the
Snowmass 2001 Workshop. It gives the major
parameters of high intensity proton sources over the
world, including machines existing, under construction
and proposed. In addition to the ISIS and PSR, several
other existing machines also provide considerable
beam power: AGS, IPNS, Fermilab Booster, Main
Injector and SPS. There are two big accelerator
projects currently under construction. One is the
Spallation Neutron Source (SNS) at the Oak Ridge
National Laboratory (ORNL) in the U.S. It consists of
a 1 GeV superconducting linac and an accumulator.
The beam power is 1.4 MW. Another is the JHF at the
JAERI/KEK in Japan. It has a 400 MeV linac, a rapid
cycling 3 GeV synchrotron at 1 MW, and a slow
ramping 50 GeV synchrotron at 0.75 MW. There are a
number of proton driver proposals from several labs,
including Fermilab and BNL in the U.S., and CERN
and RAL in Europe. There are also various proposals
of high intensity proton sources for applications other
than a proton driver, e.g., nuclear waste transmutation
and plutonium production (AAA), spallation neutron
sources (ESS, KOMAC), proton radiography (AHF),
and accelerator driven system for multiple purposes
(CONCERT, IHEP/China).
There are two types of proton drivers: one is
synchrotron-based, like the ISIS; another is linacbased (a linac plus an accumulator), like the PSR.
Each type has its pros and cons. Compared with a
linac-based system, for given beam power, a
synchrotron has the advantage of lower cost, higher
beam energy and lower beam current. Its injection
beam power is lower. Hence, the stripping foil is less
demanding and larger injection loss could be tolerated.
On the other hand, however, a synchrotron is more
CP642, High Intensity and High Brightness Hadron Beams: 20th ICFA Advanced Beam Dynamics Workshop on
High Intensity and High Brightness Hadron Beams, edited by W. Chou, Y. Mori, D. Neuffer, and J.-F. Ostiguy
© 2002 American Institute of Physics 0-7354-0097-0/02/$ 19.00
29
TABLE 1. High Intensity Proton Sources
Flux1"
(10 /pulse)
Rep Rate
(Hz)
(1020/year)
Energy
(GeV)
Power
(MW)
2.5
7
2.5
0.3
0.5
3
4.8
50
0.5
20
30
7.5
0.54
0.17
125
3.5
50
9
3.8
1.6
0.8
0.8
24
0.8
0.45
8
120
400
0.16
0.13
0.064
0.0065
0.05
0.3
0.5
14
32
8
60
0.3
25
840
10
200
1
50
3
1.4
0.75
1
3
2.5
10
15
10
20
23
6.6
10
15
15
15
0.65
2.5
5
50
25
50
45
37.5
150
9.8
25
100
1100
165
500
16
8
8
120
24
24
2.2
15
5
1.2
0.5
2
1.9
1
4
4
4
4
Europe ESS (**)
46.8
Europe CONCERT
234
LANLAAA
LANLAHF
3
KOMAC/Korea
IHEP/China
1.6
T
1 year = 1 x 107 seconds.
(*) Including planned improvements.
(**) Based on 2-ring design.
50
50
CW
0.04
CW
25
2340
12000
62500
0.03
12500
40
1.334
1.334
1
50
1
1.6
5
25
100
0.003
20
0.1
Machine
Flux
13
Existing:
RAL ISIS
BNLAGS
LANL PSR
ANLIPNS
Fermilab Booster (*)
Fermilab Main Injector
CERN SPS
Under construction:
ORNLSNS
JHF 50 GeV
JHF 3 GeV
Proton Driver proposals:
Fermilab Study I
Fermilab Study n
Fermilab Study n Upgrade
Fermilab MI Upgrade
BNL Phase I
BNL Phase E
CERN SPL
RAL 15 GeV (**)
RAL 5 GeV (**)
Other proposals:
In the meantime, based on the U.S. HEPAP SubPanel recommendation, the directors of Fermilab and
BNL had, respectively, initiated proton driver design
studies at the two labs. The reports of these studies
have either been published or will be released
soon.4,5
RECENT DEVELOPMENT IN THE
U.S. HEP COMMUNITY
In July 2001, about 1,200 physicists gathered at
Snowmass, Colorado, U.S.A. for a 3-week
workshop. The topic was the future of the highenergy physics program in the U.S. One of the
working groups was focused on high-intensity proton
sources. After a 3-week intensive study, this group
issued a 33-page report.3 This report emphasized that
the U.S. high-energy physics program needs an
intense proton source, a 1-4 MW Proton Driver by
the end of this decade. It also identified areas of
accelerator R&D needed to achieve the required
performance of a Proton Driver, i.e., a
comprehensive and prioritized 26-point plan. This
plan serves as the basis for future research and
development of high-intensity proton machines
including both linacs and synchrotrons.
TECHNICAL CHALLENGES
The design of a high-intensity proton synchrotron
involves a number of technical challenges. It requires
a careful balance on the performance of various
technical systems. It also calls for trade-offs between
performances and costs. In the following we will list
the major design issues and highlight the critical
ones. The Fermilab 8-GeV Proton Driver, which is
shown in Figure 1, will be used as an example in
these discussions.
30
FIGURE 1.
1. The
The layout
layout of
of the
the Fermilab
Fermilab 8-GeV
8-GeV Proton
Proton Driver
FIGURE
Driver (the
(the ring
ring of
of the
the racetrack
racetrack shape).
shape).
which is
which
is basically
basically aa singlet
singlet 3-cell
3-cell modular
modular structure
structure
6,7,8
with aa missing
missing (or
with
(or short)
short) dipole
dipole in
in the
the mid-cell.
mid-cell.6'7'8 (b)
(b)
with
aa missing
(or
aa doublet
doublet 3-cell
3-cell modular
modular structure
structure
with
missing
(or
9
short)
short) dipole
dipole in
in the
the mid-cell.
mid-cell.9 Figure
Figure 22 is
is an
an example
example
of (b),
of
(b), which
which is
is designed
designed for
for the
the Fermilab
Fermilab 8-GeV
8-GeV
Proton
Driver.
Proton Driver.
The choice
The
choice of
of phase
phase advance
advance per
per module
module is
is of
of
critical
importance
in
this
type
of
lattice.
There
are
critical importance in this type of lattice. There are two
two
reasons. (i)
reasons,
(i) The
The chromaticity
chromaticity sextupoles
sextupoles are
are placed
placed in
in
the
mid-cell,
where
the
beta-function
peaks
the mid-cell, where the beta-function peaks and
and
available space
available
space exists.
exists. In
In order
order to
to cancel
cancel the
the higher
higher
order
effects
of
these
sextupoles,
they
need
order effects of these sextupoles, they need to
to be
be
paired properly,
properly. (ii)
paired
(ii) The
The phase
phase advance
advance per
per arc
arc in
in the
the
horizontal plane
plane must
horizontal
must be
be multiple
multiple of
of 2π
2n in
in order
order to
to get
get
zero dispersion
zero
dispersion in
in the
the straights
straights without
without using
using
dispersion suppressors
dispersion
suppressors (which
(which are
are space
space consuming).
consuming).
Other
requirements in
Other requirements
in the
the lattice
lattice design
design include:
include:
ample
space
for
correctors
(steering
ample space for correctors (steering magnets,
magnets, trim
trim
quadrupoles, chromaticity
quadrupoles,
chromaticity and
and harmonic
harmonic sextupoles,
sextupoles,
etc.), ample
etc.),
ample space
space for
for diagnostics,
diagnostics, low
low beta
beta and
and
1. Lattice
Lattice Design
Design
1.
Lattice isis the
the foundation
foundation of
of aa synchrotron.
synchrotron. It
It is
Lattice
is
worth
every
effort
to
design
the
best
lattice
as
one
can.
worth every effort to design the best lattice as one can.
would be
be aa mistake
mistake to
to pick
pick aa lattice
lattice in
in hurry
hurry due
due to
to
ItIt would
other
factors
(e.g.,
pressed
by
the
project
schedule,
other factors (e.g., pressed by the project schedule,
which did
did happen
happen in
in the
the past).
past). A
A poorly
poorly chosen
chosen lattice
lattice
which
will have
have adverse
adverse effects
effects for
for the
the life
life of
of the
the machine.
machine. A
A
will
proton driver
driver has
hastwo
two basic
basic requirements
requirements on
on the
the lattice:
lattice:
proton
(1) transition
transition free,
free, (2)
(2) zero-dispersion
zero-dispersion in
in the
the rf
rf straight
straight
(1)
sections.
The
former
is
to
avoid
particle
loss
and
sections. The former is to avoid particle loss and
emittance
dilution
during
transition
crossing;
the
latter
emittance dilution during transition crossing; the latter
toavoid
avoid the
the synchro-betatron
synchro-betatron coupling
coupling resonance.
resonance.
to
For aa medium-energy
medium-energy synchrotron
synchrotron (above
(above 6
6 GeV),
GeV),
For
the
regular
FODO
lattice
(in
which
γ
∝
√R,
the
t
the regular FODO lattice (in which yt oc V#, RR the
machine
radius)
is
ruled
out
because
it
would
use
too
machine radius) is ruled out because it would use too
many bending
bending magnets
magnets in
in order
order to
to achieve
achieve (1).
(1). There
There
many
are
several
lattices
that
have
been
investigated
to
are several lattices that have been investigated to
obtain
either
a
high
or
an
imaginary
γ
.
For
example,
t
obtain either a high or an imaginary yt. For example,
(a) aa flexible
flexible momentum
momentum compaction
compaction (FMC)
(FMC) lattice,
lattice,
(a)
31
dispersion functions (to make the beam size small),
small),
dispersion
functions
(to(to
make
the beam size
large
dynamic
aperture
accommodate
beamsmall),
halo),
accommodate beam
beam halo),
halo),
large dynamic aperture (to accommodate
and large momentum acceptance (to allow for bunch
for bunch
and large momentum acceptance (to allow for
compression). Table 2 lists the lattice parameters of
compression). Table 2 lists the lattice parameters of
the Fermilab 8-GeV Proton Driver.
the Fermilab 8-GeV Proton Driver.
2. Space Charge
2. Space
Space Charge
Charge
2.
Amongst numerous beam physics issues, the space
Amongst numerous
numerous beam
beam physics
physics issues,
issues, the
the space
space
Amongst
charge is a major concern. It is usually the bottleneck
charge is
is aa major
major concern.
concern. ItIt isis usually
usually the
the bottleneck
bottleneck
charge
limiting the beam intensity in an intense proton source.
limiting the
the beam
beam intensity
intensity in
in an
an intense
intense proton
proton source.
source.
limiting
A useful scaling factor is the Laslett tune shift ∆ν ∝
A useful
useful scaling
scaling
factor isis the
the Laslett
Laslett tune
tune shift
shift ∆ν
Av ∝<*
A
2 factor
(N/ ) × (1/βγ22 ), in which N is number of particles per
(N/ 8NNN)) ×x (1/βγ
(1/py ),), in
in which
which N
N isis number
number of
ofparticles
particlesper
per
(N/
bunch, N the normalized transverse emittance, β and γ
bunch, EN
the normalized
normalized transverse
transverse emittance,
emittance, βp and
and γy
bunch,
N the
the relativistic factors. It shows the space charge effect
the relativistic
relativistic factors.
factors. ItIt shows
shows the
the space
space charge
charge effect
effect
the
is most severe at injection because the beam energy is
is most
most severe
severe at
at injection
injection because
because the
the beam
beam energy
energy isis
is
low. The situation becomes worse for high-intensity
low.
The
low.
The situation
situation becomes
becomes worse
worse for
for high-intensity
high-intensity
machines
not only
only because
because the
the intensity
intensityisisishigh
highbut
but
machines
machines not
not
only
because
the
intensity
high
but
also
because
the
injection
time
is
long.
Numerical
also
also because
because the
the injection
injection time
time isis long.
long. Numerical
Numerical
simulation is
is the
the main
main tool
tool to
to study
study this
this effect.
effect. A
simulation
simulation
is
the
main
tool
to
study
this
effect.
AA
number
of
1D,
2D
and
3D
codes
have
been
or
are
number
number of
of 1D,
ID, 2D
2D and
and 3D
3D codes
codes have
have been
been or
or are
are
being written
written at
at many
many institutions.
institutions. An
An example
example isisis
being
being
written
at
many
institutions.
An
example
shown in
in Figure
Figure3.
3.These
Thesecodes
codesare
areparticularly
particularlyuseful
useful
shown
shown
in
Figure
3.
These
codes
are
particularly
useful
to
the
design
of
the
injection
kicker
current
waveform
to
to the
the design of
of the
the injection
injection kicker
kicker current
current waveform
waveform
for achieving
achievinguniform
uniformparticle
particledistribution
distributionin
thebeam,
beam,
for
for
achieving
uniform
particle
distribution
ininthe
the
beam,
reducing emittance
emittance dilution
dilution and
and minimizing
minimizingaverage
average
reducing
reducing
emittance
dilution
and
minimizing
average
number of
ofhits
hitsper
perparticle
particleon
onthe
thestripping
strippingfoil
foilduring
during
number
number
of
hits
per
particle
on
the
stripping
foil
during
the
phase
space
painting
process.
Several
other
the
the phase
phase space
space painting
painting process.
process. Several
Several other
other
measures,
e.g.,
tune
ramp,
inductive
inserts,
measures,
measures, e.g.,
e.g., tune
tune ramp,
ramp, inductive
inductive inserts,
inserts,
quadrupole mode
mode damper
damper and
and electron
electron beam
beam
quadrupole
quadrupole
mode
damper
and
electron
beam
compensation
are
under
investigation
for
possible
compensation
compensation are
are under
under investigation
investigation for
for possible
possible
cures of
of the
the space
space charge
charge effects.
effects. This
This isisisan
an active
active
cures
cures
of
the
space
charge
effects.
This
an
active
research
field.
research
field.
research field.
TABLE 2. Lattice Parameters of the Fermilab
TABLE 2. Lattice Parameters of the Fermilab
Fermilab
8-GeV Proton Driver
8-GeV Proton Driver
ReesGarren
ReesGarren Racetrack
Racetrack Rbend
Rbend Lattice,
Lattice, No
No Trims
Trims
SUN
SUNSunOS
SunOS 5.X
5.X version
version 8.21/0
8.21/0
25.0
25.0
ββxx
ββyy
03/05/02 16.53.47
16.53.47
03/05/02
D
Dxx
3.0
3.0
DD
x (m)
x (m)
(m)
ββ (m)
Circumference (m)
474.2
Circumference (m)
474.2
Super-periodicity
2
Super-periodicity
2
Number
Number of
of straight
straight sections
sections
222
sections
Length
161.66
Length of
of each
each arc
arc (m)
(m)
161.66
161.66
Length
of
each
straight
section
(m)
75.44
Length of each straight section
section (m)
(m) 75.44
Injection
(MeV)
600
Injection kinetic
kinetic energy
energy (MeV)
600
Extraction
kinetic
energy
(GeV)
8
Extraction kinetic energy (GeV)
88
Injection
0.2
Injection dipole
dipole field
field (T)
(T)
0.2
Peak
dipole
field
(T)
1.5
Peak dipole field (T)
1.5
1.5
Bending
19.77
Bending radius
radius (m)
(m)
19.77
19.77
Peak
(T/m)
10
Peak quadrupole
quadrupole gradient
gradient (T/m)
10
10
Good
field
region
4"
6"
Good field region
4" ×
x× 6"
Max
β
,
β
(m)
15.14,
20.33
Max βpxx, βpyy (m)
15.14,
20.33
15.14,20.33
Min
β
,
β
(m)
4.105,
4.57
yy (m)
Min βpxx, βPy
4.105,
4.57
4.105,4.57
Max
D
in
the
arcs
(m)
2.52
Max Dxx in the arcs (m)
2.52
Dispersion
sections 00
Dispersion in
in the
the straight
straight sections
Transition
13.8
Transition γγytt
13.8
13.8
Horizontal,
vertical
tune
ν
,
ν
11.747,
8.684
Horizontal, vertical tune νvxx, νvyy
11.747,
11.747, 8.684
8.684
Natural
ξ£yyy
-13.6, -11.9
-11.9
xx,, ξ
Natural chromaticity
chromaticity ξξ^,
-13.6,
-11.9
Momentum
±1%
Momentum acceptance
acceptance ∆p/p
∆p/p
±1%
Ap/p
±1 %
Dynamic
120
Dynamic aperture
aperture
>> 120
120 πnπ
22.5
22.5
FIGURE3.
3. Space
Spacecharge
chargesimulation
simulationusing
usingTrack2D
Track2D(by
(byC.
C.
FIGURE
FIGURE
3.
Space
charge
simulation
using
Track2D
(by
C.
Prior). ItItIt shows
shows the
the particle
particle distribution
distribution after
after 45
45 turns
turns
Prior).
Prior).
shows
the
particle
distribution
after
45
turns
injection in
in the
the Fermilab
Fermilab Proton
Proton Driver
Driver with
with (left)
(left) and
and
injection
injection
in
the
Fermilab
Proton
Driver
with
(left)
and
without(right)
(right)the
thespace
spacecharge
chargeeffects.
effects.
without
without
(right)
the
space
charge
effects.
2.5
2.5
20.0
20.0
2Q.O17.5
17.5
17J-
2.0
2.0
15.0
15.0
15JB1.5
1.5
12.5
12.5
123-
10.0
10.0
10J>-
3. Electron
ElectronCloud
CloudEffect
Effect
3.
3.
Electron
Cloud
Effect
1.0
1.0
7.5
7.5
The
Theelectron
electroncloud
cloudeffect
effect(ECE)
(ECE)has
hasbeen
beenthe
theNo.
No.111
The
electron
cloud
effect
(ECE)
has
been
the
No.
problem
limiting
the
PSR
beam
intensity
for
many
problem
limiting
the
PSR
beam
intensity
for
many
problem limiting the PSR beam intensity for many
years.
years. Recent
Recentobservations
observationsand
andsuccessful
successfulcures
curesof
ofthis
this
Recent
observations
and
successful
cures
of
this
effect
on
the
CERN
SPS,
PEP-II
and
KEK
B-factory
effect
on
the
CERN
SPS,
PEP-II
and
KEK
B-factory
effect on the CERN SPS, PEP-II and KEK B-factory
have
have stimulated
stimulated worldwide
worldwide interest.
interest. At
Atthis
thismoment,
moment,
have
stimulated
worldwide
interest.
At
this
moment,
there
there are
are six
six proton
proton machines
machines that
that have
have reported
reported
there
are
six
proton
machines
that
have
reported
observations
observations of
of ECE.
ECE. They
Theyare:
are:ISR,
ISR,CERN
CERNPS,
PS,SPS
SPS
observations
of
ECE.
They
are:
ISR,
CERN
PS,
SPS
with
with LHC
LHC beams,
beams, SPS
SPS with
with fixed
fixedtarget
targetbeams,
beams,PSR
PSR
with
LHC
beams,
SPS
with
fixed
target
beams,
PSR
and
and RHIC.
RHIC. A
A key
keyparameter
parameterfor
forthe
theECE
ECEseems
seemsto
be
and
RHIC.
A
key
parameter
for
the
ECE
seems
totobe
be
the
volume
density
of
particles.
It
is
interesting
the volume
volume density
density of
of particles.
particles. ItIt isis interesting
interestingto
the
toto
notice
notice that,
that, despite
despiteenormous
enormousdifferences
differencesamong
amongthese
these
notice
that,
despite
enormous
differences
among
these
machines
machines in
in beam
beam energy,
energy, number
number of
of particles
particles per
per
machines
in
beam
energy,
number
of
particles
per
bunch
bunch and
and bunch
bunch size,
size, the
the volume
volume density
density takes
takes aaa
bunch
and
bunch
size,
the
volume
density
takes
5.0
5.0
0.5
0.5
2.5
2.5
0.0
0.0
0.0
0.0
5.
5.
/p00cc == 00..
δδEE/p
Tablename
name== TWISS
TWISS
Table
10.
10.
15.
20.
25.
25.
30.
30.
35.
35.
0.0
0.0
40.
40.
ss(m)
(m)
Table Hone = TWISS
FIGURE
FIGURE
functions of
of the
the
FIGURE 2.
2. Arc
Arc module and
and lattice
lattice functions
functions
the
Fermilab
Fermilab
Fermilab 8-GeV
8-GeV Proton Driver.
Driver. Each
Each module
module has
has three
three
doublet
doublet
short. The
The phase
phase
doublet cells.
cells. The dipole in the
the mid-cell
mid-cell is
is short.
advance
advance
advance per
per module is 0.8 and
and 0.6
0.6 in
in the
the hh- and
and v-plane,
v-plane,
respectively. There
respectively.
arc.
respectively.
There are
are five
five modules
modules in
in each
each arc.
32
remarkably similar value (about 0.2 ± 0.1 x 108 /mm3)
when reaching the ECE threshold.10 This is called the
critical mass phenomenon.
hands-on maintenance can be performed. This number
is based on the operation experiences of many
machines in many years as well as on numerical
simulations at many labs. It is now widely accepted as
a design criterion for high-intensity machines. For a 1MW, 100-m machine, this would mean the loss had to
be below 10"4, a mission impossible! To solve this
problem, collimators are introduced to localize the
beam loss. A 2-stage (i.e., primary + secondary)
collimator system can absorb more than 99% lost
particles and leave most of the enclosure below 1
W/m. A well-designed collimator system not only has
high efficiency, but also is not susceptible to parameter
changes (tunes, closed orbit, different stages during the
cycle, etc.).
It is believed that the ECE is mainly due to
secondary electron yield from the wall. Reducing
primary electrons (which come from beam loss and
stripping foil in proton machines) does not seem to be
helpful. It should be pointed out that, by far all
reported ECE are either in DC machines (accumulators
and storage rings) or AC machines in DC operation
(i.e., on flat top or flat bottom). No ECE has been seen
in AC machines during ramping. This implies that AC
machines could be immune to ECE.
However, all these claims are based on empirical
observations or numerical simulations. Lack of a
reliable theory for understanding and analyzing the
ECE is a loophole that urgently needs to be filled.
The beam power deposited onto the components
near the collimator area can reach as high as ~kW/m.
It is a difficult but also critical problem how to handle
these components in case they need to be repaired or
replaced. Invaluable experiences can be learned from
LANSCE (LANL) and PSI. n These machines have
been handling MW beams for years and have designed
several remote-handling systems that work reliably.
4. Other Beam Dynamics Issues
In addition to the space charge and ECE, there are
several other beam dynamics issues important to the
proton driver design.
• Microwave instability of bunched beam below
transition. Because the machine will always operate
below transition, the negative mass instability due to
space charge would not occur. Would then this
machine be immune to the microwave instability?
6. Negative Ion Sources
Modern high-intensity circular proton machines
almost universally adopt the charge exchange
injection. The main requirements of the negative ion
sources are high intensity (~100 mA), high brightness
(rms normalized emittance < 0.2n mm-mrad), high
duty factor (several percent) and long lifetime (> 2
months). Low noise surface plasma sources with
Cesium catalysis and volume sources are widely used
to achieve these goals.
• Bunch rotation with path length dependence on
momentum spread Ap/p and space charge tune shift
Av. This is a new problem for proton drivers. Bunch
rotation is necessary for obtaining short bunch length
(a basic feature of a proton driver). However, due to
large momentum spread (a few percent) and large tune
shift (a few tenth), the dependence of the path length
AL on Ap/p and Av can no longer be ignored. In other
words, the momentum compaction factor a = (AL/L) /
(Ap/p) cannot be treated as a constant during bunch
rotation. It is dependent upon the momentum and
amplitude of each particle. This results in a longer
bunch after rotation.
7. Chopper
In order to reduce the injection loss during rf
capture, chopping the beam at low energy is crucial.
The function of the chopper is to create a macrostructure in the linac beam so that it can fit properly
into the rf buckets in the ring. The requirements of a
chopper are: fast rise- and fall-time (tens of nsec),
short physical length (to minimize the space charge
effect, which is dominant at low energies), and flat top
and bottom in the current waveform (to minimize the
energy jittering in a beam). The ideal place to chop the
beam is at the ion source, because the beam energy is
the lowest. But the rise- and fall-time would be long
(hundreds of nsec) due to the slow response of the
plasma. The next best place is in the LEBT (low
energy beam transfer) between the ion source and the
RFQ. There are two designs. One is the LBL design
for the SNS, which places the chopper (made of split
electrodes) right after the Einzel lenses. 12 Another is
• A split between the horizontal and vertical tunes
is required in order to avoid the strong resonance 2vx 2vy = 0 that could be excited by the space charge.
However, it is not clear how big the split needs to be.
Does it have to be an integer? Or would a half-integer
suffice?
5. Beam Loss, Collimation and Remote
Handling
The rule of thumb for allowable uncontrolled beam
loss in an accelerator enclosure is 1 W/m so that
33
the Fermilab-KEK design, which places the chopper in
front of the RFQ. 13 The latter is a pulsed beam
transformer made of Finemet cores and does the
chopping by using the narrow energy window of the
RFQ. It is now installed on the fflMAC linac in Chiba,
Japan. A schematic drawing of this chopper is shown
in Figure 4. There are also choppers made of traveling
wave deflectors placed after the RFQ, which have
been in use at the LANL and BNL.
obtain a uniform distribution of particles in a bunch so
that the spaces charge and transient beam loading
effects can be reduced. Painting also helps minimize
the average number of hits per particle on the foil. The
emittance dilution due to Coulomb scattering needs to
be controlled. Carbon foil is widely used. R&D on
diamond foil and laser stripping is being pursued.
9. Slow Extraction
Although the efficiency of one-turn fast extraction
can exceed 99%, it is much lower for multi-turn slow
extractions. At high-intensity operation, the beam loss
in existing machines during slow extraction is usually
around 4-5%. This is not acceptable for the next
generation of high-intensity machines, in which the
beam power will be 1 MW or higher and one percent
loss would mean 10 kW or higher. This is a serious
problem when physics programs require slow
extractions (which is the case for KAMI and CKM at
the Fermilab Main Injector, and for kaon and nuclear
physics at the JHF). Workshops and beam experiments
are planned for tackling this problem.
10. Hardware
Although a number of proton synchrotrons have
been built in the past half-century, hardware for MW
machines presents particular challenges.
10.1. Magnets
;p
;;:|v:
Magnets are one of the most expensive technical
systems of a synchrotron. A critical parameter in the
magnet design is the vertical aperture of the main
bending magnets. The magnet cost is essentially
proportional to the aperture.
It should be large
enough to accommodate a full size beam and its halo.
The following criterion was adopted in the Fermilab
Proton Driver design:
fliif
FIGURE 4. A schematic drawing of an rf chopper made of
three Finemet cores. It is installed on the HIMAC linac.
8. H" Injection
A={3e N xp m a x /py} 1 / 2 + £> max xAp/p + c.o.d.
This is a complicated part in the proton driver
design and has many technical issues involved. Most
of particles losses in a synchrotron usually occur at
this stage as well as during the rf capture immediately
following it.
in which A is the half aperture, eN the normalized
100% beam emittance, pmax the maximum betafunction, Dmax the maximum dispersion, c.o.d. the
closed orbit distortion. The parameter 3 is the
estimated size of the beam halo relative to the beam
size.
H" particles are injected into the ring via a charge
exchange process, in which the electrons are stripped
by a foil and dumped, and the If (proton) particles
stay in the ring. This process takes hundreds or even
thousands turns. The stripping foil must be able to
stand high temperatures and large shock waves, and
must have high efficiency and reasonable lifetime. The
unstripped H", H° and electrons should be collected.
Lorenz stripping of H" ions in a magnetic field must be
avoided. Phase space painting in the transverse and
longitudinal planes needs to be employed in order to
Because this is an AC machine, field tracking
between the dipoles and quadrupoles at high field is an
important issue. Trim quads or trim coils are needed.
The peak dipole field should not exceed 1.5 Tesla. The
peak quadrupole gradient is limited by the saturation at
the pole root (not pole tip).
34
The
of the
the coil
coil turn
turn number
number per
per pole
pole isis aa
The choice
choice of
Thebetween
choice ofthe
thecoil
coil AC
turn loss
number
per
pole is a
tradeoff
and
voltage-totradeoff between the coil AC loss and voltage-totradeoff
between
the
coil
AC
loss
and
voltage-toground.
former requires
requires the
the use
use of
of many
many small
small
ground. The
The former
ground.
The
formertherequires
the use of
small
size
coils,
whereas
latter requires
requires
themany
opposite,
size
coils,
whereas
the
latter
the
opposite,
size
coils,
whereas
the
latter
requires
the
opposite,
namely,
small number
number of
of turns.
turns. There
There are
are two
two ways
ways to
to
namely,
small
namely,
small
number
turns. There
are two
ways to
compromise.
One
is to
toofemploy
employ
stranded
conductor
compromise.
One
is
stranded
conductor
compromise. One is to employ stranded conductor
coils,
as
shown in
in Figure 5,
5, which was
was adopted in
in the
coils,
as as
shown
coils,
shown inFigure
Figure 5,which
which wasadopted
adopted inthe
the
JHF
3-GeV
ring
design.
Another
is
to
connect
several
JHF
3-GeV
ring
design.
JHF
3-GeV
ring
design.Another
Anotherisistotoconnect
connectseveral
several
coils in parallel at the magnet ends,
ends, as done in
in the
coils in parallel at the magnet ends,asasdone
done inthe
the
ISIS.
The
ratio
DC
coil
of
the
AC
vs.
loss
should
ISIS. The ratio of the AC vs. DC coil loss shouldbebe
kept
2-3.
The
voltage-to-ground
shouldnot
not
The
keptaround
around2-3.
2-3.
Thevoltage-to-ground
voltage-to-groundshould
should
not
14
14 14
exceed
a
few
kV.
kV.
exceed a few kV.
The
aperture
region
should
include
and
good
The
aperture
and
goodfield
fieldregion
regionshould
shouldinclude
includeaaa
rectangular
of
an
elliptical
area).
This
area
(instead
of
an
elliptical
area).
rectangular area (instead of an elliptical area).This
Thisisisis
because
there
number
ofofparticles
particles
will
bebea significant
because
there
will
asignificant
significantnumber
numberof
particles
residing
in inthethecorners
rectangle.
The
Fermilab
residing
cornersofofthe
therectangle.
rectangle.The
TheFermilab
Fermilab
8-GeV
chose
inin×x× 666 in
in
8-GeVProton
ProtonDriver
Driverdesign
designchose
choseaaa444in
in
rectangle.
rectangle.
FIGURE
6.Waveform
Waveformof
ofthe
thetime
timederivative
derivativeofofthe
theB-field
B-field
FIGURE
FIGURE6.6.
Waveform
of
the
time
derivative
(dB/dt
T/s)generated
generatedby
byaaadual-harmonic
dual-harmonicpower
powersupply
supply
(dB/dt
(dB/dtinin T/s)
T/s)
generated
by
dual-harmonic
system.
Compared
toaaasingle
singleharmonic
harmonicsystem,
system,the
thepeak
peak
system.
system. Compared
Compared to
to
single
harmonic
value
value
duringup-ramp
up-rampisisisdecreased
decreasedby
by25%.
25%.
valueduring
up-ramp
decreased
by
25%.
FIGURE
Stranded conductorcoils
coils forreducing
reducing coilAC
AC
FIGURE
5. 5.
Stranded
FIGURE
5.
Stranded conductor
conductor coils for
for reducing coil
coil AC
losses.
losses.
losses.
10.2. PowerSupplies
Supplies
10.2.
10.2. Power
Power Supplies
This
anotherexpensive
expensivetechnical
technicalsystem.
system.There
There
This
is is
another
This
is
another
expensive
technical
system.
There
are
several
choices
for
the
power
supplies
in
a
rapid
are several
several choices
choices for
the
power
supplies
in
are
for
the
power
supplies
in aresonant
a rapid
rapid
cycling
machine.
(1)
A
single
harmonic
cycling machine.
machine. (1)
harmonic resonant
cycling
(1) A
A single
single
resonantat
system,
e.g.,
theFermilab
Fermilab
Boosterharmonic
whichresonates
resonates
system,
e.g.,
the
Booster
which
at
system,
thedual-harmonic
Fermilab Booster
which
resonates
at
15 Hz.e.g.,
(2) A
resonant
system,
e.g., the
15
Hz.
dual-harmonic
resonant
system,
e.g.,
the
15Fermilab
Hz. (2)
(2) A
AProton
dual-harmonic
resonant
system,
e.g.,
the
Driver design which uses a 15 Hz
Fermilab Proton
Proton Driver design
which uses
Hz
Fermilab
design
uses aaas15
15
Hz
component plusDriver
a 12.5%
30 Hzwhich
component
shown
component
plus
a
12.5%
30
Hz
component
as
shown
15
component
plus
a
12.5%
30
Hz
component
as
shown
below:
15
below: 15
below:
I(t) = I0 - I cos(2πft) + 0.125 I sin(4πft)
I(t) =
7(t)
= II0 --1I cos(2πft)
cos(2nft) ++ 0.125
0.125 II sin(4πft)
sin(4nft)
in which f =0 15 Hz, I0 and I are two constants
in determined
which f =by15the
Hz,injection
I0 and and
I are
two
constants
peak
The
in which f = 15 Hz, I0 and 7 are
twocurrent.
constants
determined
by
the
injection
and
peak
current.
The
advantage by
of this
is that
peak value
of dB/dt
determined
the system
injection
andthepeak
current.
The
advantage
of this system
that
the peak
value of
dB/dt
is decreased
25%, is
shown
in Figure
which
advantage
of thisbysystem
isasthat
the peak
value 6,
of dB/dt
is leads
decreased
by
25%,
as
shown
in
Figure
6,
which
to
a
saving
of
the
peak
rf
power
by
the
same
is decreased by 25%, as shown in Figure 6, which
leads
to a saving
of the peak rframp
power
by thee.g.,
same
amount.
(3) A programmable
system,
the
leads
to a saving
of the peak rf power
by the same
amount.
(3) A programmable
ramp system,
the
AGS Booster
and AGS. Although
this ise.g.,
a most
amount.
(3) A programmable
ramp system,
e.g.,
the
AGS
Booster
and(e.g.,
AGS.
Although
is a most
versatile
system
allowing
for a this
front
and a
AGS
Booster
and AGS.
Although
this porch
is a most
versatile
system
(e.g.,
allowing
for a front porch and a
flat
top),
it
is
also
most
expensive.
versatile
system (e.g., allowing for a front porch and a
flat top), it is also most expensive.
flat top), it is also most expensive.
FIGURE7.7.AA7.5
7.5MHz
MHz Finemet rf
rf cavity installed
installed in the
FIGURE
FIGURE
7. A Injector.
7.5 MHzFinemet
Finemet rfcavity
cavity installedininthe
the
Fermilab
Main
Fermilab
FermilabMain
MainInjector.
Injector.
10.3. RF
10.3.
103.RF
RF
The rf system is demanding, because it must
The
system
isis demanding,
because itit must
The arfrflarge
system
must
deliver
amount
ofdemanding,
power to thebecause
beam in a short
deliver
a
large
amount
of
power
to
the
inina ashort
deliver
amount
power
to thebeam
beam
short
period. aInlarge
addition,
it of
must
be tunable,
because
the
period.
In addition,
must
because
the
period.
addition, itit frequency
mustbebetunable,
tunable,
because
the
particle Inrevolution
increases
during
particle
revolution
frequency
increases
during
particle
revolution
frequency
increases
during
acceleration.
Cavities with
ferrite tuners
have been
in
acceleration.
Cavities
with
tuners
inin
use for decades.
Recently
the development
of
the
acceleration.
Cavities
withferrite
ferrite
tunershave
havebeen
been
use
for
decades.
Recently
the
development
of
the
Finemet
cavities at
the KEK
aroused strong
use
for decades.
Recently
the has
development
of the
Finemet
cavities
at the KEKThanks
strong
interest at
many laboratories.
to
a US-Japan
Finemet
cavities
at the KEK has
has aroused
aroused
strong
interest
at many
laboratories.
Thanks
a US-Japan
collaboration,
Fermilab
has built
a 7.5toto
MHz,
15 kV
interest
at many
laboratories.
Thanks
a US-Japan
collaboration,
has built
a 7.5
MHz,
15 for
kV
Finemet cavityFermilab
and installed
in the
Injector
collaboration,
Fermilab
has itbuilt
a Main
7.5 16
MHz,
15 kV
Finemet
cavity
and
installed
it
in
the
Main
Injector
for
bunch coalescing,
shown in
Figure
7. 16 Injector
The main
Finemet
cavity andasinstalled
it in
the Main
for
bunch
coalescing,
as
shown
in
Figure
7.
The
main
16
advantages
of the Finemet
cores
are high7.accelerating
bunch
coalescing,
as shown
in Figure
The main
advantages
ofwide
the Finemet cores
high is
accelerating
gradient and
Theare
former
especially
advantages
of the bandwidth.
Finemet cores
are
high accelerating
gradient
andforwide
bandwidth.small
The former
is especially
important
high-intensity
size
rings,
in
which
gradient
and
wide
bandwidth.
The
former
is
especially
important for high-intensity small size rings, in which
important for high-intensity small size rings, in which
35
space
isisits
its
spaceis
precious.The
Themain
mainconcern,
concern,however,
however,is
its
space
isisprecious.
precious.
The
main
concern,
however,
high
power
consumption.
For
example,
the
Fermilab
highpower
powerconsumption.
consumption.For
Forexample,
example,the
theFermilab
Fermilab
high
Finemet
amplifier
toto
Finemetcavity
cavityneeds
needsaa a200
200kW
kWpower
power amplifier
amplifier to
Finemet
cavity
needs
200
kW
power
drive
alloys
are
under
driveit.
Newtypes
typesof
magnetic alloys
alloys are
are under
under
drive
it.it.New
New
types
ofofmagnetic
magnetic
investigation
investigationfor
forperformance
performanceimprovement.
improvement.
investigation
for
performance
improvement.
10.4.
Vacuum
10.4.Vacuum
Vacuum
10.4.
Vacuum
cycling
machine
isis
Vacuumpipe
pipe for
for aa a rapid
rapid cycling
cycling machine
machine is
Vacuum
pipe
for
rapid
probably
items.
Ceramic
probablyone
oneof
themost
mostchallenging
challengingitems.
items.Ceramic
Ceramic
probably
one
ofofthe
the
most
challenging
pipe
successfully
pipewith
withaa ametallic
metalliccage
cageinside
insidehas
hasbeen
beensuccessfully
successfully
pipe
with
metallic
cage
inside
has
been
employedatatthe
theISIS.
ISIS. However,
However, this
this isis aa costly
costly
employed
solution,because
becauseit itoccupies
occupiesa asignificant
significantportion
portion ofof
solution,
themagnet
magnetaperture.
aperture.Assuming
Assumingthe
theceramic
ceramicwall
walland
and
the
the
magnet
thecage
cageneed
needa a 1-in
1-in vertical
vertical space,
space, the
the magnet
magnet
the
the
cage
aperturewould
wouldhave
havetotobebeincreased
increasedfrom
from4-in
4-intoto5-in
5-in
aperture
FermilabProton
ProtonDriver,
Driver,a a25%
25%increase.
increase.This
This
ininthetheFermilab
Fermilab
willdirectly
directlybebetranslated
translatedtotoa a25%
25%increase
increase inin the
the
will
magnet
and
power
supply
costs,
equivalent
to
tens
magnet
magnet and power supply costs, equivalent to tens ofof
millionsdollars.
dollars.Therefore,
Therefore, itit was
was rejected
rejected by
by the
the
millions
millions
FermilabProton
ProtonDriver
Driverdesign.
design.
Fermilab
Fermilab
FIGURE 8.
8. Corner
Corner section
canned
dipole
with
FIGURE
8.
Corner
section of
of aa canned
canned dipole
dipole with
with aaa
FIGURE
perforatedmetallic
metallicliner.
liner.
metallic
liner.
perforated
Inductive inserts:
are
made
of
ferrite
rings
Inductive
inserts: They
They are
are made
made of
offerrite
ferrite rings
rings
•••Inductive
andalso
also can
can have
have bias
bias current
current for
for
impedance
tuning.
and
also
can
have
bias
current
for impedance
impedance tuning.
tuning.
and
Their inductive
inductive impedance
impedance would
would fully
fully
or
partially
inductive
impedance
would
fully or
or partially
partially
Their
compensate the
the space
space charge
charge impedance,
impedance, which
which
is
compensate
the
space
charge
impedance,
which is
is
compensate
capacitive.
The
first
successful
experiment
was
at
the
capacitive.
The
first
successful
experiment
was
at
the
capacitive.
The
first
successful
experiment
was
at
the
18
18
LANL.18
Twoferrite
ferrite modules
modules made
made
by
Fermilab
have
LANL.
Two
ferrite
modules
made by
byFermilab
Fermilabhave
have
LANL.
Two
been
installed
in
the
PSR.
They
help
increase
the
e-p
installed in
in the
the PSR.
PSR. They
They help
help increase
increase the
the e-p
e-p
been installed
instability
threshold,
which
is
a
major
bottleneck
of
threshold, which
which isis aa major
major bottleneck
bottleneck of
of
instability threshold,
that machine.
machine. Another
Another experiment
experiment at
at
the
Fermilab
machine.
Another
experiment
at the
the Fermilab
Fermilab
that
Boosteris
beingplanned
planned(Figure
(Figure 9).
9).
Booster
isisbeing
being
planned
(Figure
9).
Thinmetallic
metallicpipe
pipeisisananalternative.
alternative. However,
However, itit
Thin
must
be
very
thin
(several
mils)
in
order
minimize
must
must be very thin (several mils) in order totominimize
the
eddy
current
effects
(pipe
heating
and
induced
the
eddy
current
effects
(pipe
heating
and
induced
the eddy
effects
magneticfield).
field). Such
Such a a thin
thin pipe
pipe isis mechanically
mechanically
magnetic
magnetic
field).
unstableunder
undervacuum.
vacuum.Several
Severalmethods
methodshave
have been
been
unstable
unstable
tried
to
enhance
its
stability,
including
ceramic
shields,
tried
tried to enhance its stability, including ceramic shields,
metallicribs
ribsand
andspiral
spirallining.
lining.The
Thefirst
firsttwo
twodo
donot
not
metallic
metallic
look
promising.
The
third
one
is
under
investigation.
look promising. The third one is under investigation.
FermilabProton
ProtonDriver
Driverdesign,
design,aadifferent
different
InInthetheFermilab
approach
was
adopted.
The
magnets
employ
external
approach was adopted. The magnets employ external
vacuum
skins
like
those
in
the
Fermilab
Booster.
vacuum skins like those in the Fermilab Booster.
vacuum
Perforated metallic liners are usedininthe
themagnet
magnetgap
gaptoto
Perforated
Perforated
metallic liners are used
provide a low-impedance environment for the beam as
provide a low-impedance
provide
environment for the beam as
shown in Figure 8.1717
shown in Figure 8. 17
shown
10.5. Diagnostics
10.5. Diagnostics
A system that can diagnose beam parameters
beam parameters
A system
that injection
can diagnose
during
multi-turn
is highly desirable. The
during multi-turn injection
injection is highly desirable. The
during
method for fast, accurate non-invasive tune
method for fast,
fast, accurate non-invasive tune
method
measurement is being developed. A circulating beam
measurement
is
being developed. A circulating beam
measurement
profile monitor covering a large dynamic range with
profile monitor covering a large dynamic range with
profile
turn-by-turn speed will be crucial for studying beam
turn-by-turn
fordeveloped
studying for
beam
turn-by-turn
speedinstrument
will be crucial
halo. (A similar
has been
the
halo.
for
the
2
halo.
(A
similar
instrument
has
been
developed
linac beam halo experiment at LANL.22 )
linac beam halo experiment at LANL. )
FIGURE 9. Inductive inserts in the Fermilab Booster.
FIGURE
FIGURE 9.
9. Inductive
Inductive inserts
insertsin
inthe
theFermilab
FermilabBooster.
Booster.
• Induction synchrotron: This is a longitudinally
•• Induction
synchrotron:
aa longitudinally
Induction
synchrotron:
This
longitudinally
separated
function
machine.This
In isis
other
words, the
separated
function
machine.
In
other
separated
function
machine.
In
other
words,
the
longitudinal focusing and acceleration arewords,
carried the
out
longitudinal
focusing
and
acceleration
are
carried
longitudinal
focusing
and acceleration
out
by two separate
rf systems.
The former are
usescarried
barrierout
rf
by
two
rf
The
uses
barrier
rfrf
by
two separate
separate
rf systems.
systems.
Therfformer
former
uses
barrier
buckets,
the latter
a constant
voltage
curve.
One
buckets,
the
aa type
constant
rf
curve.
One
buckets,
the latter
latter
constant
rf voltage
voltage
curve.bunch
One
useful feature
of this
of machine
is tunable
useful
feature
of
this
type
of
machine
is
tunable
bunch
useful
feature
of
this
type
of
machine
is
tunable
bunch
lengths. So the so-called superbunch acceleration
lengths.
So
lengths.
So the
the1919so-called
so-called superbunch
superbunch acceleration
acceleration
could be possible.
could
could be
be possible.
possible.19
• Barrier rf stacking: The application of Finemet
Barrier
rf
application
of
Barrier
rf stacking:
stacking:
The
application
of Finemet
Finemet
and•• other
magnetic
alloysThe
makes
it possible
to
build
and
other
magnetic
alloys
makes
it
possible
to
and
other magnetic
makes
possible
to build
build
broadband
barrier rfalloys
cavities
withithigh
voltage
(~10
broadband
barrier
rf
cavities
with
broadband
barrier
rf can
cavities
with
high voltage
voltage
(~10
kV or higher).
They
be used
to high
stack
beams in(~10
the
kV
or
They
can
used
in
kV
or higher).
higher).phase
They space.
can be
be This
used to
to stack
stack beams
beamsuseful
in the
the
longitudinal
is
particularly
longitudinal
longitudinal phase
phase space.
space. This
This isis particularly
particularly useful
useful
11. New Ideas
11. New Ideas
In the past several years, there are a number of new
Inrevitalized
the past
past several
several
years,
there to
are the
aa number
number
new
the
there
are
ofdriver
new
orIn
ideasyears,
proposed
protonof
or
revitalized
ideas
proposed
to
the
proton
driver
orstudy.
revitalized
ideas proposed to the proton driver
For example:
study. For
For example:
example:
study.
36
when the beam intensity of a synchrotron is limited by
its injector (e.g., the intensity of the Fermilab Main
Injector is limited by its Booster). Compared to the
slip stacking, an advantage of barrier rf stacking is the
greatly reduced beam loading effects due to a much
lower peak beam current. 20
REFERENCES
1. U.S. HEPAP Sub-Panel Report on Long Range Planning
for U.S. High Energy Physics, January 2002, http://doe-
hep.hep.net/HEPAP/lrp_report0102.pdf
2. Wangler, T.P., "Linac Based Proton Drivers," these
proceedings.
• Fixed field alternating gradient (FFAG)
accelerator: Although MURA first proposed this idea
about 40 years ago, it was almost forgotten. Only the
recent activities at the KEK brought it back to the
world's attention. KEK has successfully built a 1 MeV
proton FFAG and is building a large 150 MeV one.21
FFAG is an ideal machine for high intensity beams. Its
repetition rate can be much higher than a rapid cycling
synchrotron (in the range of kHz). One problem of the
FFAG, though, is that it is difficult (if not impossible)
to fit it into an existing accelerator complex, which
usually consists of a linac and a cascade of
synchrotrons.
3. Chou, W. and Wei J., editors, "Report of the Snowmass
M6 Working Group on High Intensity Proton Sources,"
FERMILAB-Conf-01/396, BNL-52639, August 10,
2001.
4. "The Proton Driver Design Study," FERMILAB-TM2136, December 2000.
5. "Proton Driver Study E - Part 1," FERMILAB-TM2169, May 2002.
6. Lee, S.Y. et al., Phys. Rev. E48, 3040 (1993).
7. Wienands, U. et al., Proc. 1992 HEACC (Hamburg,
Germany), p. 1070.
8. See Chapter 3 of Ref. 4.
• Repetition rate increase in existing synchrotrons:
This is a brute force approach but can be appealing
because it is straightforward. For example, in the BNL
Proton Driver design, one proposal is to increase the
AGS repetition rate from 0.5 Hz to 2.5 Hz. 22 The
Fermilab Main Injector upgrade also includes a rep
rate increase (from 0.53 Hz to 0.65 Hz). 5
9. See Chapter 3 of Ref. 5.
10. Chou, W., "Summary Report of Session VI," Proc.
ECLOUD'02 Workshop, CERN, Geneva, April 15-18,
2002, CERN Yellow Report 2002-001.
11.Wagner, E., "Remote Handling and Shielding at PSI,"
these proceedings.
12. Staples, J. et al., Proc. 1999 PAC (New York, USA),
p.1961.
CONCLUSIONS
Proton drivers are a hot topic in today's accelerator
community. Because of their versatile applications,
modest costs and great potentials to serve future big
projects (a neutrino factory, a muon collider or a
VLHC), the designs are being pursued in numerous
laboratories over the world. There are many technical
challenges. But there are no showstoppers towards the
construction of such a facility. The world needs more
than one proton driver. International collaborations on
a number of R&D items have been formed. Steady
progress and fresh ideas can be expected in the coming
years in this dynamic field.
13. Chou, W. et al., "Design and Measurements of a Pulsed
Beam Transformer as a Chopper," KEK Report 98-10
(September 1998); Chou, W. et al., Proc. 1999 PAC
(New York, USA), p. 565; Shirakabe, Y. et al., Proc
2000 EPAC (Vienna, Austria), p. 2468.
14. Ostiguy, J.-F. and Mills, F., Proc. 2001 PAC (Chicago,
USA), p. 3248.
15.
Jach, C. and Wolff, D., Proc. 2001 PAC (Chicago,
USA), p. 3248.
16. Wildman, D. et al., Proc. 2001 PAC (Chicago, USA), p.
882.
17. See Chapter 8 of Ref. 4.
ACKNOWLEDGMENTS
18. Ng, K-Y. et al., Proc. 2001 PAC (Chicago, USA), p.
2890.
A large portion of this paper is based on the
Fermilab Proton Driver Study conducted by a group of
physicists and engineers from Fermilab and several
other institutions.
Yakayama, K. et al., Proc. 2002 EPAC (Paris, France),
p. 998.
Ng, K-Y., "Doubling MI Beam Intensity using RF
Barriers," these proceedigns.
This work is sponsored by Universities Research
Association Inc. under Contract No. DE-AC0276CH03000 with the United States Department of
Energy.
Yoshimoto, M. et al., Proc. 2001 PAC (Chicago, USA),
p. 51.
37