TRASCO RFQ DESIGN

M. Comunian, A. Pisent, INFN-LNL, Legnaro, Italy

G.V. Lamanna, INFN-Bari, Bari, Italy.

kept constant, while the average aperture R0 is increased

Abstract
so to allow a higher electrode modulation keeping the
In this paper, we describe the design of the 5 MeV 30 necessary aperture a. This choice allows to safe about half
mA proton CW RFQ, of the Italian research program a meter in RFQ length (i.e. 40 kW RF power) without a
TRASCO (TRAsmutazione SCOrie). The first third of ramp of the voltage.
this structure is under construction. The main parameters are listed in Tab. I, and their
evolution along the structure are shown in Fig.1.
1 INTRODUCTION
We describe the design of the RFQ (Radio Frequency
Quadrupole) of the Italian research program TRASCO
(TRAsmutazione SCOrie) [1]. This research program is
aimed to test the critical components of a high intensity
proton linac to be used in an ADS (Accelerator Driven
System) for nuclear wastes transmutation or energy
production following the Energy Amplifier scheme.
Other CW proton RFQs at 352 MHz have been built or
designed recently, and in particular, LEDA RFQ at LANL
is necessarily a point of comparison [2]. The main
difference respect to LEDA design is the lower beam
current (30 mA instead of 100 mA), dictated by the lower
beam power requirement for our ADS applications.
Moreover, in our linac architecture, keeping the same
frequency in the accelerator (DTL or ISCL) that follows
the RFQ, we can accept 5 MeV as RFQ output energy [3] Figure 1: RFQ parameters as a function of length.
As a result, respect to LEDA, we could relax some
parameters. Most relevant in our opinion are the Table II:RFQ parameters
possibilities of using a single klystron and of keeping a
Emittance T RMS in/out 0.2/0.2 mmmrad
constant intervane voltage along the structure. The first
normalised
point has a considerable impact in the cost of the RFQ,
Emittance L RMS 0.18 MeVdeg
while the second point allows a “reasonable” power
RFQ Length 7.13 m
dissipation density of 850 W per structure cm length.
Intervane voltage 68 kV
Table 1 RFQ specifications Transmission 96 %
Particle p Modulation 1 ÷ 1.943
Input Energy 80 KeV Average aperture R0 2.93÷ 3.19 mm
Output Energy 5 MeV Minimal aperture 2.05 mm
Frequency 352.2 MHz Synchronous phase -90 ÷ -29 deg
Current 30 mA Dissipat power (SF*1.2) 0.579 MW
Max Surface Field 33 MV/m Beam loading 0.1476 MW
RF Power consumption <800 kW Q 8261
Duty factor 100 %
The RFQ has been extensively simulated with
PARMTEQM [4] (Fig. 2), while crosschecks have been
2 RFQ BEAM DYNAMICS done with TOUTATIS [5] made by CEA/Saclay and
The RFQ is divided in the four traditional sections, the LIDOS.RFQ made by lidos team [6]. The results are:
radial matching section, the shaper, the gentle buncher 97.41 % accelerated and 98.52 % transmitted protons by
and the accelerator. In the last section, that corresponds to TOUTATIS and, 98.66 % accelerated and 98.66 %
more than ¾ of the total structure length, the voltage is transmitted particles by LIDOS.RFQ, all runs are at

Proceedings of EPAC 2000, Vienna, Austria 927

50 mA. The transmission as obtained by PARMTEQM is 3 THE RESONATOR
96 % at 50 mA (Fig. 3), with losses mainly located below
The resonator, of four vanes kind (Fig.5), is divided in
2 MeV (Fig. 4), so to minimize the activation problems of
three segments, resonantly coupled, following LANL
the copper structure.
technique. In this way, the operating mode is about 2
MHz distant respect to the closest quadrupole modes, and
the dipole modes are outside the range of the main
quadrupole band (Fig. 6). This approach was checked with
extensive simulations (SUPERFISH, MAFIA and HFSS
codes) and by comparison with the measurements on the
cold model, performed with bead pulling system.
The modulation will be milled with a single tool, with
transverse radius 2.93 mm. The homogeneous voltage
along the RFQ is obtained by increasing the transverse
section dimensions as R0 increases. It results in a constant
local cut-off frequency due to inductance compensation.

Figure 2: RFQ Beam dynamics in the RFQ for a current

of 50 mA, 10000 particles.

100 0.32
T [% ]
98 0.3
E L RM S [M eVdeg]
96 0.28
Transm ission [% ]

E lR M S [M eVdeg]

94 0.26

92 0.24

90 0.22
Figure 5: Mafia representation of one sector of the RFQ
88 0.2
transverse section with a tuner.
86 0.18

84 0.16

82 0.14
Dipoles Quadrupoles
0 10 20 30 40 50 60 70 80 90 100
C urrent[m A ]

Figure 3: Transmission and longitudinal emittance in the

RFQ as function of beam current.

3.50 350.0

3.00 300.0

Relative Losses [% ]
2.50 250.0 Dipole cut-off Operating mode
D eposited P ower[W ]
R elative Losses [% ]

Deposited Power[W ]
(quadrupole cut-off)
2.00 200.0

1.50 150.0

1.00 100.0
Figure 6: The eigenfrequencies correspond to the zeros
0.50 50.0 the plotted functions for quadrupole and dipole modes for
0.00 0.0 the stabilized RFQ. For comparison, the thick dots and
0.5 1 1.5 2 2.5 3
Energy [M eV]
3.5 4 4.5 5 crosses correspond to the quadrupole and dipole modes of
the unstabilized RFQ.
Figure 4: Particles losses and power deposited in the RFQ
as function of energy; the total power deposited is 1 kW
and above the 2 MeV is 210 Watts.

928 Proceedings of EPAC 2000, Vienna, Austria

 4 RFQ ERROR STUDY corresponding to the first third of the structure, will be
built during 2001.
A 50 mA proton beam has been simulated under
various conditions, in order to test the tolerances to errors;
in all cases we plot the beam transmission and the final ACKNOWLEGMENTS
longitudinal emittance, that are both sensitive deign This study has been possible thanks the fruitfull
parameters. discussions with many colleagues. We thank A. Lombardi
and G. Bisoffi from LNL, D. Schrage from LANL, C.
Pagani and D. Barni from LASA, J.M. Lagniel, F. Simeon
and R. Ferdinand from CEA/Saclay.
dx [cm]
-0.06 -0.05 -0.04 -0.03 -0.02 -0.01 0.00 0.01 0.02 0.03 0.04 0.05 0.06

β [cm/rad]
x
4.17 4.47 4.77 5.07 5.36 5.66 5.96 6.26 6.56 6.85 7.15 7.45 7.75

RMS Transverse Emittance [mm.mrad]

0.14 0.15 0.16 0.17 0.18 0.19 0.20 0.21 0.22 0.23 0.24 0.25 0.26
98.0 0.20
LE β /β0
97.2 0.18
LE dx/am in
96.4 0.16
LE ε/ε0

LE:RMS Longitudinal Emittance [MeVdeg]

95.6 0.14

94.8 0.12

T:Transmission [%]
T ε/ε0
T V/V0
94.0 0.10

93.2 0.08
LE V/V0
T dx/am in
92.4 0.06
T β/β0
91.6 0.04

90.8 0.02

90.0 0.00
-30 -25 -20 -15 -10 -5 0 5 10 15 20 25 30
Mism atch (βx/β x0)[%]
Misalign (dx/am in) [%]
Emittance/E 0 [%]
Voltage/V0 [%]

Figure 8: Errors sensitivity for a 50 mA proton beam

Figure 7: Transmission and Longitudinal emittance in a
three segments RFQ (50 mA). The Voltage in REFERENCES
PARMTEQM is shaped as indicated in the upper figure [1] “http://trasco.lnl.infn.it”
(blue line).
[2] L.M. Young, L.J. Rybarcyk “Tuning the LEDA RFQ
6.7 MeV Accelerator” (LANL) Proceedings of the 1998
In fig.7 is plotted the effect of field dishomogeneity. In
International Linac Conference, Argonne, USA p. 270
the upper part we show the characteristic field pattern of a
segmented RFQ, where construction errors make non null [3] A. Pisent et al. ‘’TRASCO 100 MeV high intensity
the component of the two neighborhood quadrupolar Proton Linac” this conference
modes. This determines a tolerance in field homogeneity [4] Kenneth R. Crandall, James H. Billen, Rene S. Mills,
of about +1%, that is compatible with our mechanical Dale L. Schrage, Richard H. Stokes, George H.
tolerances and field adjustment procedure. Neuschaefer, Thomas P. Wangler, and Lloyd M. Young
In fig. 8 we plot the calculated sensitivity to initial beam ”RFQ Design Codes”, LA-UR-96-1836
alignment, mismatch, emittance and vane voltage [5] B. Bondarev, A. Durkin, S. Vinogradov, J-M. Lagniel,
variation. For example a beam misalignment of 1 has R. Ferdinand “CW RFQ designing using the LIDOS.RFQ
almost no effect. From the previous results we can codes” LINAC 1998, pp. 502-504.
conclude that a high beam transmission is preserved in
case of different error sources and that the beam dynamics [6] R. Ferdinand, R. Duperrier “IPHI-RFQ reliability
is well estabilished. approach”, proceedings of Second Workshop on
At present the main choices for the machining and Utilizations and Reliability of High Power Accelerators
welding are taken, and we are completing the construction Aix-en-Provence, France, from 22 to 24 November 1999.
drawings. A technological model of this RFQ,

Proceedings of EPAC 2000, Vienna, Austria 929