Current performance of the self-extracting cyclotron
S. Lucas, F. Swoboda,
IBA Radio-Isotopes, Av de l'Espérance, 1, 6220 Fleurus, Belgium
W. Kleeven, J.L. Delvaux, Y. Jongen.
IBA, Chemin du Cyclotron, B-1348 Louvain-La-Neuve, Belgium
Abstract. The self-extracting cyclotron is a 14MeV multi-mA H+ machine from which the beam extracts without a
deflector. The development of this prototype has started in 1998, and has now reached a point such that IBA
considers to use it as a production machine. It is now installed in an irradiation facility and is equipped with two
beam lines and two high power target-system.
Beams of more than 1 mA have been extracted and transported to targets Further development is ongoing in order
to increase the current on target to at least 2 mA in the coming months. Commercial isotope production will start at
the end of this year.
This paper will describe the current configuration of the cyclotron and the associated performances. Emphases will
be put on reliability and associated problems, beam optics and performances of sub-systems.
extended radius; a groove (figure 1-1) machined on the
extended hill sectors along the extracted orbit creates a
sharp dip in the magnetic field where the field index is
<-1. That groove is shaped so that a strong separation
gradient (septum action) between the last internal turn
and the extracted beam is obtained. The extraction of
the beam is obtained by creating a turn–separation at
the entrance of the groove by the use of harmonic
extraction coils (precessional extraction, figure 1-2).
This differs from what has been presented in [3].
INTRODUCTION
In 1995 IBA proposed a method to extract positive
ions from a cyclotron without the use of an
electrostatic deflector [1]. It relies on a very fast
transition of the average magnetic field near the pole
radius from the internal isochronous region to the
region where the field index is smaller than -1 and the
bending strength of the field is too low to keep the
beam in the machine. Self-extraction was already
experimentally observed on the IBA 230 MeV
protontherapy cyclotron, where there was some beam
intensity present in the extracted beam line even when
the deflector was removed from the machine.
Encouraged by these experiences and their agreement
with computer simulations of the self-extraction
principle, IBA started in 1998 the construction of a
high intensity self-extracting cyclotron [2, 3, 4].
The beam exiting from the cyclotron is horizontally
as well as vertically diverging. A magnetic gradient
corrector is placed in the valley behind the groove in
order to adopt the beam phase space to the external
beam line (figure 1-3).
When passing the return yoke the beam is
horizontally focused by a permanent magnet
quadrupole. It is built up of layers of 2.0 cm and 3.0
cm thick allowing the total length of a quadrupole to
be varied with a step of 1.0 cm (figure 1-4).
BASIC DESIGN
The design has several unconventional features: the
hill gap has a quasi-elliptical shape that creates an
average magnetic field which remains isochronous up
to the pole radial edge; the hill–sector guiding the
extracted beam and the opposite hill–sector have an
The cavities design is classical (figure 1-5) with
vertical symmetry and dee stem located close to the
central region. The RF chain involves a 3 steps
CP680, Application of Accelerators in Research and Industry: 17th Int'l. Conference, edited by J. L. Duggan and I. L. Morgan
© 2003 American Institute of Physics 0-7354-0149-7/03/$20.00
956
amplifier with a final power of 200 kW as described in
[3].
BEAM LINE LAYOUT
An experimental beam line described in [3] was put
in place for the first tests.
The extraction principle allows for multi-turn
extraction. Of course, some particles fall in between
the inner limit of the extracted beam and the outer
limit of the internal beam. This beam loss is catched
by a special beam dump called beam separator (figure
1-6) as described in section 5 (Beam separator)
6
Moroever, the industrial configuration involves an
intermediate beam line, a switching magnet and two
target beam lines as shown in figure 2.
3
Fa r a d a y
Co l l ima t o r s
Qu a d r u p o l e
Be r g o z
4
Swit c h in g
St e e r in g
7
St e e r in g
Qu a d r u p o l e
Fa r a d a y
Be r g o z
Sc a n n in g ma g n e t
X&Y BPM
Co l l ima t o r s
Ta r g e t
5
2
8
1
Figure 2: General layout of the installation
The total length of a beam line from the cyclotron
to both targets is about 9 meters. The main optical
elements in the cyclotron vault are a permanent
magnet quadrupole located in the exit port of the
cyclotron return yoke, an active doublet at about 2.5
meters from the cyclotron and the switching magnet.
Each irradiation vault contains an active doublet and a
scanning magnet that scans the beam over the target.
There is a XY-steering magnet in each of the vaults.
Furthermore in each of the three vaults there is an
interceptive beam stop for current measurement
(Faraday cup) and a non-interceptive current
measurement (Bergoz). At about 1 meter from the
cyclotron there is a pair of horizontal and a pair of
vertical drum collimators, that cut away the halo of the
beam. Just before the target there are two BPM’s for
measurement of the horizontal and vertical beam
profile at high beam intensity.
Figure 1: Inside view of the cyclotron. 1:groove,
2:extraction coils, 3:gradient corrector, 4:permanent magnet
quadrupole, 5:cavities, 6:beam separator, 7:radial probe,
8:ion source
Current beam characteristics are presented in table 1.
Table 1: Phase space parameters at the exit of the vacuum
chamber. All quantities are rms defined. If the phase space
would have elliptical symmetry with a Gaussian distribution,
then the used definition would correspond with 86 % of the
beam.
Emittance (π mm-mrad)
Half beam size (mm)
Max. divergence (mrad)
Beam divergence (mrad)
Correlation
Twiss α
Twiss β (mm/mrad)
Twiss γ (mrad/mm)
Converging/Diverging
Horizontal
250.1
26.3
59.3
58.5
0.987
-6.168
2.77
14.1
Diverging
Vertical
40.3
6.2
9.79
-7.29
-0.745
1.123
0.954
2.38
Converging
These BPM have been designed to be easy to
maintain, reliable and able to perform at high beam
current. One of these is shown in figure 3.
957
EXTRACTION EFFICIENCY
Cu p ip e
The extraction efficiency is defined as follow:
η (%) =
I Faraday + I Collimators
I Faraday + I Collimators + I Beam Separator
Figure 3: High Current Beam Profile Monitor
The water-cooled pipe oscillates back and forth at
180° and its position is measured with a potentiometer
located at the opposite side of the motor. Two identical
systems are installed in a row, the last one oriented at
90 degree in order to measure the profile in both the
horizontal and the vertical plane. The system is used to
setup the machine at each run, but also to assess the
beam quality and stability during the run. Therefore
beam profiles are recorded at periodic intervals
without the need to reduce the beam current.
1.10
where IFaraday and ICollimator denotes the current
measured on the Faraday cup and on the collimators
attached to the cyclotron, and IBeam Separator is the current
measured on the beam separator.
Figure 4 shows the evolution of η versus current
extracted from the cyclotron. This is automatically
logged at every run and the increase of the current
takes place in two steps: a fast slope for current <= 500
µA, followed by a slower slope up to the presetted
value (1310 µA in this case). The data in the inset
shown the variation of the extraction yield during 7
consecutive hours at 1310 µA setpoint.
0.80
0.79
1.05
7 hours run with a
setpoint of 1.31 mA
0.78
0.77
1.00
0.76
0.75
Extraction yield (%)
0.95
0.74
0.73
0.90
0.72
0.71
0.85
0.70
1280
1290
1300
1310
1320
1330
1340
0.80
0.75
0.70
Ripple < 100 V
0.65
Ripple > 100 V
0.60
0
100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400
Beam current (µA)
Figure 4: Extraction yield versus current extracted from the cyclotron
958
For current <= 500 µA, the extraction yield
decreases from 80 % to 73 %. This is to be related to
the dee-voltage ripple that increases with the
accelerated current. The extraction is obtained by
creating a turn-separation at the entrance of the groove
with the harmonic coils. Unfortunately there exists still
some modulated turn pattern structure at the entrance
of the groove. However, if there is a dee-voltage
ripple, then this turn pattern will be oscillating in time
around some average and the extraction is no longer
fully optimized. Therefore, the extraction efficiency
slightly drops with increasing dee-voltage ripple.
The measured extraction yield at 1310 µA during 7
hours is fairly stable and has an average value of 73 %.
Thanks to our main-coil and ion-source regulation
software, one can see in the graph-inset that the
maximum beam current variation on the target is kept
within 2.3 %
BEAM SEPARATOR
Typical extraction efficiency is 73 %. Therefore 27
% of the beam is not extracted and is collected on a
special device called beam separator. That device is
water cooled (4 bars, 40 l/min) to handle beam current
up to 1 mA on a surface of 5.6 cm2 (2.5 kW/cm2).
Figure 5: Beam separator assembly
Another configuration will be soon tested: a special
quality Rh sheet (pinhole free, rolling in a preferential
direction) that will act as a target and will allow us to
produce several Ci of 103Pd per week in addition to our
conventional targets.
In order to reduce as much as possible the power
delivered to the beam separator, thin (100 µm) Ta
sheet is used as shown in figure 5.
This sheet is simply clamped on a part that includes
a thin (2 mm) water channel, and the sealing is done
by EPDM O'Ring. Mean Time Before Failure of this
system has not yet been assess, but up to 25 mA.H has
been assess so far.
BEAM PROFILE AND BEAM STABILITY
Figure 6 and 7 show typical horizontal and vertical
profiles of the 7 hours run at 1310 µA.
Vertical beam profile at BPM
1.0
Intensity (a.u.)
0.8
0.6
0.4
0.2
0.0
-60
-40
-20
0
20
Position (mm)
Figure 6: Vertical profile
959
40
60
near future to act as a target for additional
production capabilities.
Horizontal beam profile at BPM
End of the run
Beginning of the run
3.5
3.0
Intensity (a.u.)
Pd
Several runs of one shift duration (7-8 hours) have
demonstrated that this prototype can be used in the
near future for 103Pd production. Unfortunately, these
runs also revealed that the beam size/shape changes
with time. Work is in progress to solve that issue.
4.5
4.0
103
2.5
2.0
REFERENCES
1.5
1.0
[1] Y. Jongen, D. Vandeplassche, P. Cohilis, High
Intensity Cyclotrons for Radioisotope Production or
the Comeback of the Positive Ions, Proc. 14th Int.
Conf. on Cyclotrons and their pplications, Cape Town,
South Africa, 1995, World Scientific Publisher, pp.
115–119.
[2] W. Kleeven, M. Abs, J.C. Amelia, W. Beeckman,
J.L. Bol, V. Danloy, Y. Jongen, G. Lannoye, S. Lucas,
J. Ryckewaert, D. Vandeplassche, S. Zaremba, SelfExtraction in a Compact High-Intensity H+ Cyclotron
at IBA, Proc. 7th European Part. Accel. Conf.
(EPAC2000), Vienna, Austria, 2000, pp. 2530-2532.
[3] S. Lucas, W. Kleeven, M. Abs, E. Poncelet, Y.
Jongen, Status Report of the development of a multimA Self-Extracted H+–beam Cyclotron, Proceedings
of the 16th Int. Conf. on the Applications of
Accelerators in Research and Industry CAARI 2000.
[4] W. Kleeven, S. Lucas, S. Zaremba, W. Beeckman,
D. Vandeplassche, M. Abs, P. Verbruggen, Y. Jongen,
The Self-Extracting Cyclotron, Proceedings of the 16th
Int. Conf. On Cyclotrons and their Applications
(CYC2001), East Lansing 2001, pp. 69-73.
[5] W. Kleeven, S. Lucas, J.-L. Delvaux, F. Swoboda,
S. Zaremba, W. Beeckman, D. Vandeplassche, M. Abs
and Y.Jongen, The IBA Self-Extracting Cyclotron
Project, Presented at the 23rd European Cyclotron
Progress Meeting, Warsow 2002, to be published in
Nukleonika.
0.5
0.0
-0.5
-60
-40
-20
0
20
40
60
80
Position (mm)
Figure 7: Horizontal profiles
One can see that for the beam line setting
considered, the vertical profile is gaussian like with a
FWHM of about 1 cm. The horizontal beam profiles
recorded at the beginning and at the end of the 7 hours
run exhibit higher intensities at both ends. This is due
to the multi-turn extraction as described in [5]. In
addition, the profile is symmetrical at the beginning of
the run but becomes asymmetrical with time. We do
not have yet clear explanations for that effect.
CONCLUSIONS.
Four years ago IBA started the development of a new
platform technology for the extraction of high intensity
beam. A prototype has been build and a subsidiary of
IBA (IBA Radio-Isotopes) has been created with the
main goal to complete the development, commission
the cyclotron and setup an irradiation center for the
production of 103Pd.
Today, currents up to 1.3 mA on target have been
accelerated for several hours. At low current the
extraction efficiency is 80 %. It drops when the current
is increased but remain stable for current larger than
500 µA.
The cyclotron stability associated with the software
control allow to keep the beam current fluctuations on
target below 2.3 %.
A thin foil technology has been developed to
accommodate the power being dissipated into the
beam separator. That last one will be modify in the
960