COSY-SCL, the superconducting injector linac for COSY

Abstract

The superconducting injector linac COSY-SCL is being designed and constructed at the Forschungszentrum Juelich. The train goal of the new injector is to fill the cooler synchrotron COSY with polarized protons as well as with polarized deuterons up to the space-charge limit at injection energy. COSY-SCL is characterized by a base frequency of 160 MHz, 25 kV ion-source extraction voltage, a pulse length of up to 500 μs, a maximum repetition rate of 2 Hz, injection into the linac at β=0.073 and injection into COSY at kinetic energies of 52 MeV for protons and 56 MeV for deuterons, respectively. The injector configuration is presented, and its main subsystems-ion source (CIPIOS from IUCF for polarized H^- and D^-), RFQ (built in co-operation with the University of Frankfurt), linac (based on half-wave resonators operating at 160 and 320 MHz)- are described and discussed. The present status of the project is reported.

Full text

Page 1

COSY-SCL, THE SUPERCONDUCTING INJECTOR LINAC FOR COSY
R.Toelle*, U.Bechstedt, N.Bongers, J.Dietrich, R.Eichhorn, F.M.Esser, O.Felden, R.Gebel, K.Henn,
H.Jungwirth#,A.Lehrach, R.Maier, U.Pfister, D.Prasuhn, P.v.Rossen,A.Schnase, H.Schneider,
Y.Senichev, R.Stassen, H.Stockhorst, E.Zaplatin, Forschungszentrum Juelich GmbH, Postfach 1913,
D-52425 Juelich, Germany;
A.Schempp , IAP, Johann Wolfgang Goethe-Universitaet, Frankfurt am Main, Germany
Abstract
The superconducting injector linac COSY-SCL is being
designed and constructed at the Forschungszentrum
Juelich. The main goal of the new injector is to fill the
cooler synchrotron COSY with polarized protons as well
as with polarized deuterons up to the space-charge limit at
injection energy. COSY-SCL is characterized by a base
frequency of 160 MHz, 25 kV ion-source extraction
voltage, a pulse length of up to 500 µs, a maximum
repetition rate of 2 Hz, injection into the linac at β=0.073
and injection into COSY at kinetic energies of 52 MeV
for protons and 56 MeV for deuterons, respectively. The
injectorconfigurationis
subsystems -ion source (CIPIOS from IUCF for polarized
H-and D-), RFQ (built in co-operation with the University
of Frankfurt), linac (based on half-wave resonators
presented,anditsmain
operating at 160 and 320 MHz)- are described and
discussed. The present status of the project is reported.
INTRODUCTION
Several independent review committees have identified
a much enhanced future scientific potential for the COSY
accelerator facility if intensities for polarized ions could
be raised to about 2*1011, which roughly corresponds to
the limitation by space charge. The evaluations also
showed that a different injector is needed for this purpose
because the present injector, the 30 years old isochronous
cyclotron JULIC, is constituting a bottle neck that would
foil any attempt to reach that goal. The most suitable new
injector for COSY turned out to be a superconducting (sc)
linac with independently phased single-cell cavities [1].
Figure 1: Detailed view of the injector linac area with (i) ion source section, (ii) RFQ section,
(iii) sc linac section, (iv) transfer line to COSY and the diagnostics line (from right to left).
___________________________________________
* r.toelle@fz-juelich.de
# present address:Accel Instruments, Bergisch Gladbach, Germany
0-7803-7739-9 ©2003 IEEE 598
Proceedings of the 2003 Particle Accelerator Conference

Page 2

LINAC CONCEPT
The new injector should be able to deliver polarized
proton and deuteron beams with sufficient intensities to
fill COSY to the space-charge limit at the chosen injection
energy of about 50 MeV. In addition, the new injector
should fit into the existing building of COSY to keep civil
engineering costs low. Furthermore, construction should
not interfere with the ongoing experimental program and
allow a fast switching from the cyclotron to the new
injector once it has been completed.
A normal conducting linac, due to its excessive length,
turned out to be incompatible with the existing space
limitations. In addition, a beam dynamics analysis showed
that superconducting quarter-wave resonators such as in
use at other facilities to accelerate heavy ions are not
suitable in our energy range for light ions. Hence, a new
design in form of superconducting half-wave resonator
(HWR) structures was worked out.
For establishing this design, a large number of obstacles
had to be overcome and new solutions for many crucial
details had to be found. An international review
committee scrutinized the proposed design and found that
this project could make significant contributions to the
development in this field of accelerator technology. Their
encouragement and recommendations paved the way
from design to construction.
Figure 1 shows the injector area and gives a detailed
layout of the injector linac components. The cryogenic
plant is not included in this picture, because the
compressors will be located outside the building to
sufficiently suppress mechanical vibrations.
ION SOURCES
The ion source complex located in the small annex will
be able to deliver polarized and unpolarized H-or D-with
an energy of 25 keV and pulses of up to 500 µs width at a
repetition rate up to 2 Hz. Two ion sources are foreseen, a
commercial multi-cusp source that is already delivered
and the adapted CIPIOS source of Indiana University
(IUCF). The contract for delivery has already been signed
and also includes the electrostatic low energy beam-
transport system. In addition, a collaboration with the
INR-institute (Troitzk) was established to further improve
the performance of this source. First results of this work,
still obtained at IUCF, succeeded in significantly reducing
the electron current emission and reached a much higher
peak current as well as an increased pulse width. The
product of these values will determine the number of ions
that can be injected into COSY.
RFQS
The
section will be used to bunch the beam, accelerate the
ions to 2.5 MeV/nucleon, and shape the phase space
according to the requirements of the superconducting
linac. As the velocity profiles are different, each ion
species needs its own RFQ. The RFQs will be mounted
subsequentradio-frequencyquadrupole(RFQ)
alongside each other on a mechanical rail system to allow
switching the ion species in less than one working day.
Phase-space matching is achieved using 2 quadrupole
doublets in combination with an rf-cavity, acting as a
rebuncher [2]. The RFQs and the rebuncher cavity have
been designed in collaboration with the University of
Frankfurt/Main [3].
HALF-WAVE RESONATORS
The superconducting linac section comprises 11 unit
cells plus an option for two further ones. Each cell is 1.7
m long. It contains at the outgoing end a quadrupole
doublet for transverse focussing and a diagnostic box in
between. The quadrupoles are designed for a maximum
gradient of 45 T/m and will be operated in a pulsed mode
to save electric power as well as space.
Although quarter-wave resonators have a proven record
for the acceleration of heavy ions, the sensitivity of
lighter ions to magnetic fields mandated the use of half-
wave resonators. Four HWRs, made of ultra pure niobium
and cooled close to 4 K, are grouped inside one cryostat
without focusing elements. The resonators are arranged in
two families. The first 20 resonators (βstr=0.11) operate at
160 MHz, and the following 24 resonators (βstr=0.2)
operate at 320 MHz.
160MHz
0.11
249
10.4
4.5
∫
−
=
Eacc
320MHz
0.2
244
9.4
4
βstr
R/Q
Bpeak/Eacc
Epeak/Eacc
βλ
∞
∞
dzzE
)(
Table 1: Characteristic data of the two HWR types with
definition of Eacc.
The resonators have to provide very high acceleration
gradients (6-8 MV/m) that are limited by the maximum
magnetic fields tolerable inside the structure without
quenching. To ensure reliable operation a design was
chosen that limits the magnetic field to 80 mT in normal
operation. The given acceleration gradient corresponds to
8 MV/m and will result in an average energy gain per
resonator of slightly more than 1 MeV. The calculated
distribution of the magnetic field in the resonator is
shown in figure 2.
In close co-operation with industry, concepts for the
manufacturing of such resonators have been developed.
As no precedent for this type of rf-resonator exists, two
companies have been selected to manufacture one proto-
type each. They use different manufacturing techniques
for the critical 80 mT region, which ultimately will
determine the performance. The final decision for the type
of resonator to be used depends on the outcome of
performance tests.
One of the cleaning ports at the resonator is used for rf-
coupling. The coupler design has been developed in our
599
Proceedings of the 2003 Particle Accelerator Conference

Page 3

institute and the manufacturing of a prototype is in
progress. The design takes into account the pulsed
operation of the linac and tries to minimize cryogenic
losses for this operational mode. The coupling strength
can be adjusted for the acceleration of ions or rf-
conditioning of the resonators. Each resonator will be
driven individually in amplitude and phase by a
commercial pulsed 4 kW solid state amplifier. A
prototype of this amplifier has been successfully tested.
Figure 2: The magnetic field distributuion computed
inside the cavity of the half-wave resonator. Sections
perpendicular to (left) and in beam direction are shown
with details of the most critical regions as indicated.
It is well known that superconducting resonators are
very susceptible to microphonics due to their extremely
narrow resonance curve. The mechanical vibration modes
of the resonator have been studied with extensive
simulations to arrive at a design with the necessary rf-
stability for acceleration.
Another great concern in high performance super-
conducting acceleration resonators for pulsed operation is
the Lorentz force detuning. This effect becomes relevant
as a result of the extremely high gradients reached inside
such cavities, leading to a sizeable mechanical wall
deformation and hence to a change of the resonance
frequency well outside the allowed bandwidth. This
change is proportional to the square of the peak rf-
amplitude and has to be minimized by a careful choice of
the structure. An extensive analysis has been carried out
to establish the magnitude of this effect for the chosen
design and for making appropriate design optimizations.
The remaining detuning effect will be dealt with by a
sophisticated adaptive feed-forward control system that
employs piezo-actuators counteracting the deformation.
Also the development of an optimized cryostat is an
issue of great concern as it affects to a high degree the
operating costs of a superconducting injector. For a
pulsed machine static losses and dynamic losses have to
be carefully balanced. A further limitation exists in terms
of the cryostat size, as the unit cell length may not exceed
1.7 m. This constraint is a consequence of the beam
behaviour that can be expected from extensive beam
dynamics simulation results. Protons and deuterons must
move on stable trajectories in the required phase space
during the entire acceleration process.
To conform to space requirements and at the same time
keep thermal losses low, a design was developed for the
cold-warm junctions of the beam ports of the cryostat,
which had only a length of 60 mm. The calculated
cryogenic loss of such a junction is only around 0.1 W.
Figure 3 shows the results of the calculated thermal
energy flow for such a cold-warm junction.
Figure 3: Calculated thermal energy flow for a cold-warm
junction of the cryostat taking into account heat radiation
and thermal conduction [4].
OUTLOOK
At the very end of the year 2002, due to unprecedented
budget cuts imposed by the Federal Government, the
Board of Directors of the Forschungszentrum Jülich has
been left with no other choice but suspending the COSY-
SCL project for at least two years. However, prototype
work may proceed and first results can be expected during
2003/2004. The conviction that this new injector linac
will be an important scientific gain for the COSY
accelerator facility has not been diminished by this
decision.
ACKNOWLEDGEMENT
The authors would like to thank especially A. Facco
(INFN-LNL),A. Schempp
K. W. Shepard (ANL) and S. O. Schriber (MSU) for
fruitful discussions.
(Univ.Frankfurt),
REFERENCES
Design
Juelich,
[1] COSY-SCL
Forschungszentrum
(http://www.fz-juelich.de/ikp/linac-public/)
[2] A. Lehrach, H. Jungwirth, R. Maier, R. Toelle,
“Beamdynamic
Superconducting Injector Linac”, this conference
[3] K.-U. Kuehnel, A. Schempp, A. Schnase, R. Toelle,
C. Welsch, “The RFQ-Injector for COSY-SCL”, this
conference
[4] H.Cords,“Berechnung
Kuehlleistung für den Kryostaten des geplanten
LINAC zum Cooler Synchrotron COSY“, Internal
Report, FZJ-ZAT-378
Update,Internal
March
Report,
2002
Calculationsforthe
derbereitzustellenden
3.4 W
3.3 W
0.1 W
600
Proceedings of the 2003 Particle Accelerator Conference