WEOAA1
Proceedings of IPAC2017, Copenhagen, Denmark
COMMISSIONING OF SPIRAL2 CW RFQ AND LINAC
R. Ferdinand, P.E. Bernaudin, P. Bertrand, M. Di Giacomo, A. Ghribi, H. Franberg, O. Kamalou,
J-M. Lagniel, G. Normand, A. Savalle, F. Varenne, GANIL, Caen, France
D. Uriot, CEA/DRF/IRFU, Saclay, France
Abstract
The SPIRAL2 88 MHz CW RFQ is designed to accelerate light and heavy ions with A/Q from 1 to 3 at
0.73 MeV/A. The nominal beam intensities are up to
5 mA CW for both proton and deuteron beams and up to
1 mA CW for heavier ions. The design foresees almost
100% transmission for all ions at nominal beam current
and emittance. Beam commissioning of the RFQ and
linac cool down started already. The specifications have
been achieved within the measurement precision for the
different ions accelerated yet. This paper describes the
beam commissioning strategy, the measurement results in
both transverse and longitudinal planes and the successfully first cryogenic tests of the linac.
INTRODUCTION
Copyright © 2017 CC-BY-3.0 and by the respective authors
GANIL is significantly extending its facility with the
new SPIRAL2 project based on a multi-beam Superconducting CW linac driver [1, 2].
The layout of the SPIRAL2 driver takes into account a
wide variety of beams to fulfill the physics requests. It is
a high power CW superconducting linac delivering up to
5 mA proton and deuteron beams or 1 mA ion beams for
Q/A > 1/3 (Table 1). Our major challenges are to handle
the large variety of different beams due to their different
characteristics (in terms of particle type, beam currents –
from a few µA to a few mA - and/or beam energy), a high
beam power (200 kW, CW) and to answer correctly to the
safety issues, especially with the deuteron beam.
Table 1: Beam Specifications
Particles
H+
D+ ions
A/Q
1
2
3
Max I (mA)
5
5
1
Max energy (MeV/A) 33
20
15
Max beam power (kW) 165 200
45
option
6
1
8.5
51
PROJECT STATUS
Figure 1: Cryomodules in the linac tunnel.
All superconducting cryomodules are installed (Fig. 1).
HEBT installation is ongoing. Cryogenic valves boxes
manufacturing defects resulted in more than a year delay
for the repairs. Small isolation vacuum leaks are still
observed on several valve boxes and cryomodules when
cold, therefore a new dynamic pumping system has been
designed and installed.
BEAM COMMISSIONING
Reference particles were selected related to an increasing stress for the RFQ cavity (increasing vane voltage).
We started with the proton beam (A/Q=1). This validated
the light ion source, its LEBT and the RFQ at 50 kV vane
voltage. The second beam, 4He2+ beam, up to 2mA
(A/Q=2), is chosen to mimic the future deuteron beam
and validate the RFQ vane voltage at 80kV. It also allowed us to start validating the heavy ion source performances. The third beam is chosen to demonstrate the
ultimate performances of the injector: 1 mA, CW, A/Q=3
ion beam. For this, the 18O6+ ion beam is chosen as the
more convenient to produce up to 1 mA. The RFQ has to
work at its maximum vane voltage of 113.6 kV. The 5-mA
deuteron beam or RF injection in the cryomodules requires the final authorization from the French nuclear
safety authorities. We are still waiting for this, and ready
to proceed. All the other A/Q particle tunings will be
extrapolated from these reference beams.
The heavy ion source was damaged during the installation work, making it at the moment impossible to measure
other particle beams before the linac injection.
INJECTOR RESULTS
ECR Source Results
Up to 11 mA H+ beam current can be extracted from
the light ion source (70% proton fraction). The permanent
magnet positions in the ion source have been adapted
online in order to optimize and stabilize the tuning performances and repeatability of the ion source. Argon,
helium and oxygen beams have been extracted from the
heavy ion source..
Especially with the heavy ion source, the LEBT emittance may show some strong filamentations. Fortunately,
three pairs of H and V slits are located in the common
LEBT to define the emittances. We usually optimize the
line transport with the transverse emittances to get the
highest beam current on the final LEBT Faraday cup, then
cut the halo (few % of the total intensity) to get a 100%
transmission through the RFQ.
The beam performances measured at the end of the
LEBT are given in Table 2.
Emittances have been measured both in CW and pulsed
mode operation, to measure and optimize the neutralization time. For example, the characteristics of a 5.8 mA
proton beam are stabilised after about 400 µs with a residual pressure of 10-6 mbar (uncorrected value).
ISBN 978-3-95450-182-3
04 Hadron Accelerators
2462
A08 Linear Accelerators
Proceedings of IPAC2017, Copenhagen, Denmark
Table 2: Measured Performances at the LEBT End
Particle
H+
4
He2+
18 6+
O
Beam current
(mA)
5.2
1.35
0.75
Emit X
(π.mm.mrad)
0.18
0.54
0.44
Emit Y
(π.mm.mrad)
0.2
0.43
0.41
WEOAA1
closing of the LEBT slits.
On December 2016 the oxygen beam could be accelerated in the RFQ, but not at the project frequency (see
above). Only the beam current transmission could be
measured.
RFQ RF Conditioning
Injector Diagnostic Plate
The D-Plate is installed in the Medium Energy Beam
Transport Line (MEBT, Fig. 2) in order to validate the
RFQ performances, to develop and qualify the diagnostics
and to measure the following beam characteristics:
• Intensity with Faraday cups, ACCT and DCCT
• Transverse profiles with classical multi wire profilers and ionisation gas monitor (MIGR)
• H and V transverse emittance with Allison type
scanners
• Energy with a Time of Flight (TOF) monitor
• Phase with the TOF and the BPM
• Longitudinal profile with a Fast Faraday Cup (FFC),
and a Beam Extension Monitor (BEM)
• Beam position and ellipticity (ߪ௫ଶ െ ߪ௬ଶ , with σx and
σy the standard deviations of the beam transverse
sizes) with the BPM.
The diagnostics performances are given in [6,7,8]
Figure 2: Injector scheme up to Diagnostic Plate.
RFQ Beam Commissioning
On December 3, 2015, the first proton beam was accelerated at 0.73 MeV (200 µA of proton, 200 µs/250 ms,
50 kV vane voltage law). By noon the same day, 100%
transmission was demonstrated and within a few days, a
5.2 mA CW proton beam was successfully accelerated.
On February 2016, a 1.34 mA, CW 4He2+ beam was accelerated with up to 98.5% transmission in spite of an
input transverse emittance bigger than expected (see Table 2). The 100% transmission was obtained with a slight
Figure 3: Comparison between measurement and TraceWin/Toutatis simulation (p, He and O beams).
The beam transmission as a function of RF vane voltage and the beam characteristics were measured. There is
a very good agreement between these measurements and
the beam dynamics simulations performed using the
TraceWin/Toutatis code (Fig. 3). RFQ output horizontal
emittance could be also validated with proton and Helium
beam.
The RFQ beam energy is measured using 3 ToF pick up
electrodes [8]. The proton beam was measured from
10 µA to 5 mA (pulsed and CW), helium beam from
10 µA to 1.5 mA. (see Table 3 below)
Table 3: RFQ Measured Beam Energy
Energy
(keV/nucleus)
Proton
Helium
Toutatis
TOF buncher TOF buncher
simulation off
on
730
727.2
729.3
728.1
727.3
The longitudinal bunch parameters were characterized
using two tools: a Fast Faraday Cup (FFC) and a Beam
Extension Monitor (BEM). The BEM is composed of a
150 µm tungsten wire interacting with the beam (limited
beam power), the measurement is done analyzing the
emitted X-rays using µchannel plates coupled with a fast
readout anode [8]. The estimated temporal resolution σ =
47 ps corresponds to 1.5° of phase resolution at 88 MHz.
Using the rebuncher, the 3-gradient method has been
used to measure a longitudinal emittance of
0.27 π.deg.MeV for the Helium beam (0.19 expected).
The bunch profile measurements done with helium
beam current from 0.1 to 1 mA showed very interesting
behaviors. The longitudinal bunch shapes are quasi
Gaussian at high intensity but have a thin structure at low
intensity (Fig. 4). The same behavior is observed with
protons (quasi Gaussian shape from 0.5 mA to 5 mA).
This behavior was explained through TraceWin simulations (Fig. 5): at low beam current the S-shaped particle
distribution in the longitudinal phase-space is not scram-
04 Hadron Accelerators
ISBN 978-3-95450-182-3
A08 Linear Accelerators
2463
Copyright © 2017 CC-BY-3.0 and by the respective authors
TOUTATIS [3] simulation with the measured voltage
law [4,5] AND the manufacturing errors showed that the
expected transmission up to the MEBT faraday cup is
100% for proton, 99.97% for 4He2+ and 99.77% for 18O6+.
The cavity voltage measurement was calibrated using
an X-ray energy measurement technique. See [5] for the
details of the voltage law in operation.
Up to now, various technical difficulties did not allow
us to use the RFQ cavity at its ultimate performance (CW,
113.6 kV). The RF consumption at nominal voltage is
38kW above the expected value (200kW), and has not
been explained so far. LLRF, amplifier and cooling circuit
do not allow us to 'lock' the cavity at the right frequency.
The data (power meas., beam trans.) are obtained using a
loop that follows the frequency of the cavity [5].
WEOAA1
Proceedings of IPAC2017, Copenhagen, Denmark
bled by the space charge force.
Table 4: Cryomodules LHe Consumption
Static @4K
Saclay/Orsay meas.
GANIL meas.
Figure 4: 4He2+ longitudinal bunch shape for 1 and
0.1 mA.
CMA01
5.73 W
4.95 W
CMA02
3.98 W
2.99 W
CMB07
19 W
12.33 W
The gain observed in the static helium consumption is
due to a lower copper shielding temperature (less radiation losses, 60K vs 80K). No degradation during transport
from Paris area to GANIL and installation was recorded.
The pressure stability could be optimized within
± 3 mBar for the 3 cryomodules cooled together (fig. 7).
The dynamic RF losses were successfully simulated on
pressure stability with the CMA heaters (adding 24W per
cryomodule, well above the expected RF consumption).
Steps of 4W had no impact on pressure stability.
Figure 5: TraceWin simulation of a 0.15mA He beam at
the BEM location (see green projection on x-axis).
LINAC First Cool Down
Copyright © 2017 CC-BY-3.0 and by the respective authors
In July 2016 all the conditions were gathered to allow a
partial cool down of the SC Linac. The tests consisted of
cooling down three cavities in two different types of cryomodules (one high and one low β). This stage allowed
testing cryodistribution, behavior of cryoplant when connected to the LINAC cryo lines, as well as the preliminary
version of cryo PLCs and Local Control system [9-11].
However, thermal acoustic oscillations (TAO) in the main
liquid helium phase separator prevented the validation of
the cryoplant and the cavities pressure/level control. The
TAO problem was solved in March 2017, allowing a
second cool down with three cryomodules (4 cavities).
This time, the objective was to validate the helium pressure stability requirements of ± 3mbar.
Figure 7: Pressure and liquid helium regulation for 3
cryomodules A01, A02, B07.
The next stage aims to validate several cavities with
RF. This is programmed to begin as soon as we obtain the
safety authorities authorization. Upon system’s performance approval, the whole LINAC will be cooled down
and all cavities shall be commissioned at nominal RF
gradient.
CONCLUSION
We are facing exciting days, with the first accelerated
beams in the injector, and a great 100% transmission
through the RFQ. The preliminary results are very similar
to the expected theoretical ones, illustrating the good
design of the machine, and giving us confidence for the
next phases. We are working to solve the technical difficulties in order to validate the A/Q=3 beam at the RFQ
exit (Source and RF), hopefully before the end of 2017.
We are still waiting for the safety authority authorisation to allow us to inject RF in the cryomodules.
ACKNOWLEDGEMENT
Figure 6: Cool down of the 4 cavities.
Once the cryomodules were cold (cavity at 4K and LHe
level stabilized, fig. 6), several successive measurements
showed that the thermalization required 5 days for a CMA
and 6 days for the CMB. Once stabilized, the following
static cryogenic consumptions at 4K were measured (Table 4).
The authors wish to thank all the GANIL and partner
labs staff for their deep involvement in the SPIRAL2
project. The presented results are a great reward after so
many years of their involvement.
ISBN 978-3-95450-182-3
04 Hadron Accelerators
2464
A08 Linear Accelerators
Proceedings of IPAC2017, Copenhagen, Denmark
[1] E. Petit, “Status of SPIRAL 2 project”, in Proc. HIAT’12,
Chicago, USA, June 2012.
[2] P. Anger, P. Bisson, O. Danna, X. Hulin, J. M. Lagniel,
S. Montaigne, F. Perocheau, E. Petit, L. Roupsard,
"SPIRAL2 project: integration of the accelerator processes,
construction of the buildings and process connections",
in Proc. LINAC’16, paper THPLR024.
[3] D. Uriot, N. Pichoff, "Status of TraceWin code",
in Proc. IPAC’15, paper MOPWA008, Richmond, USA.
[4] O. Piquet "RFQ developments at CEA-IRFU",
in Proc. IPAC’16, paper MOOCA02, Busan, Korea.
[5] O. Piquet, M. Desmons, A. France. P. Galdemard,
M. Di Giacomo, R. Ferdinand, J. M. Lagniel,
"RF CONDITIONNING OF THE SPIRAL2 CW RFQ",
presented at IPAC’17, Copenhagen, Denmark, May 2017,
this conference.
[6] C. Jamet, W. Le Coz, G. Ledu, S. Loret, C. Potier de Courcy, "Energy and Longitudinal Bunch Measurements at the
SPIRAL2 RFQ Exit", in Proc. IBIC’16, paper WEPG42,
Barcelona, Spain.
[7] P. Ausset, M. B. Abdillah, G. Joshi, P. D. Motiwala,
S. K. Bharade, G. Randale, R. Ferdinand, D. Touchard,
"Operation of the Beam Position Monitor for the Spiral 2
Linac on the Test Bench of the RFQ", in Proc. IBIC’16,
paper WEPG11, Barcelona, Spain.
[8] R. Revenko and JL Vignet, "Bunch Extension Monitor for
LINAC of SPIRAL2 Project", Proc IBIC 2016, TUPG59,
Barcelona, Spain
[9] R. Ferdinand, P. Bertrand, M. Di Giacomo, H. Franberg,
O. Kamalou, J. M. Lagniel, G. Normand, A. Savalle,
F. Varenne, D. Uriot, J. L. Biarrotte, "Status of SPIRAL2
and RFQ beam commissioning", in Proc. LINAC’16,
paper WE1A06, East Lansing, USA.
[10] A. Ghribi et al., "Status of the Spiral 2 Cryogenic System"
(Accepted for publication), in: Cryogenics (2017).
[11] A. Ghribi, P. E. Bernaudin, Y. Bert, C. Commeaux,
M. Houeto, G Lescalié, “Spiral 2 Cryogenic System for the
Superconducting LINAC”, in: IOP Conference Series:
Materials Science and Engineering, 171.1 (2017),
p. 012115,
http://stacks.iop.org/1757-899X/171/i=1/a=012115.
04 Hadron Accelerators
ISBN 978-3-95450-182-3
A08 Linear Accelerators
2465
Copyright © 2017 CC-BY-3.0 and by the respective authors
REFERENCES
WEOAA1