A HIGH CURRENT SUPERCONDUCTING PROTON LINAC

FOR AN ACCELERATOR DRIVEN TRANSMUTATION SYSTEM

C. Pagani, G. Bellomo, P. Pierini, G. Travish, D. Barni, A. Bosotti, R. Parodi*
INFN Sezione di Milano, LASA, Via Fratelli Cervi 201, 20090 Segrate (MI), Italy
*
INFN Sezione di Genova, Via Dodecaneso 33, 16146 Genova, Italy
Abstract cyclotrons possess the advantage of being compact and so
they could be candidates for a demonstration plant.
High current CW proton linac accelerators have been
Room temperature linacs offer a mature technology,
recently proposed for a number of applications based on
but suffer from a much lower RF efficiency and high heat
the use of a large flux of spallation neutrons. In this con-
loads in CW operation.
text, an R&D program on an accelerator driven system
One approach, which relies on well tested technology
for nuclear waste transmutation has been recently approv-
and provides efficient use of wall plug power, combines a
ed in Italy. Our specific task is to develop, together with
low energy section based on a room temperature linac,
the national industry, a design of the high energy part of
and a high energy section based on b-graded
the proton accelerator, along with prototype development
superconducting structures.
for the most critical components. In this paper we present
The low energy section of such a system poses a
a revised version of the design proposed at Linac’96,
number of challenges; however, a number of existing
using five cell cavities, rather than the original four cells.
technologies address the regime of interest. For instance,
This modification, together with a new criterion for using
proton sources in excess of 100 mA have been
the transit time factor curve for non resonant proton
demonstrated, RFQs operating with more than 100 mA
velocities, results in a more modular and efficient design.
are operating and DTL structures at high current, up to
A 1.6 GeV linac, operated at 25 mA, allows to reach
200 MeV exist, albeit at low duty cycle.
40 MW beam power. The beam power upgrade is
Here we consider systems with energy greater than
achievable using additional couplers per cavity.
1 GeV and average beam power in excess of 30 MW.
1 PROJECT OVERVIEW Moreover, we address only the high energy section
(above 100 MeV) of the system. In the design presented
An R&D program has started in Italy on an accelerator
in Section 4 we choose a final beam energy of 1.6 GeV,
driven system for nuclear waste transmutation. In the
very close to the Los Alamos APT design [3]. The final
recently approved two year program, starting in January
choice of beam energy and current is heavily influenced
1998, two Italian research agencies (INFN and ENEA)
by the specific application and is outside of the scope of
will be supported to study, together with other institutions
the present paper.
and the national industry, critical components of the
The following sections of this paper illustrate the
accelerator driven system. In the following we present a
approach followed to design the high energy section from
revised version of the reference design, originally
100 MeV to the target energy of 1.6 GeV.
proposed at Linac’96 [1], and the rationale behind the
parameter choices. 2 THE 350 MHZ SC LINAC OPTION
More generally, a number of applications have been
The motivation for turning to superconducting
recently proposed for high current proton linacs,
technology is twofold: high RF efficiency is readily
including nuclear waste transmutation, energy production
achieved, which is important for CW operations, and
by sub-critical reactors, spallation neutron sources and
large apertures are possible, which are relevant to
trit-ium production [2]. These applications share in
minimizing particle loss.
common a number of requirements on the accelerator. A
The choice of operating frequency is set by a few
high average current, high energy proton beam is needed.
general considerations. CERN has operated the LEP2
Additional requirements include high wall plug efficiency
cavities at 350 MHz, and has added extensive experience
and low particle losses to guarantee the hands-on
at this frequency to the accelerator community. A number
maintenance of the system. While these systems differ in
of cavity production tools and test stands are available in
the detailed parameters desired, they generally operate in
European companies as a result of the LEP2
the space charge dominated regime.
development. Additionally, the 350 MHz choice matches
A number of accelerator configurations for the system
existing RFQ and DTL designs.
in question have been considered in the literature. While
Frequencies higher than the LEP frequency imply smal-
this paper focuses on the superconducting linac option, it ler physical dimensions for the cavities and associated
is worth mentioning alternatives. components, but require a more complex 2 K cryogenics.
Cyclotrons are presently limited to beams up to 1 MW, The injection energy of 100 MeV is mainly determin-
and can conceivably be extended up to a few MW. Still,

0-7803-4376-X/98/$10.00  1998 IEEE 1126

ed by elliptical superconducting cavity constraints. Ellip- beam acceleration, which must be within the limits of the
tical cavities for low energies (matched b<0.5) have a available coupler technology [4].
short active length and a low accelerating gradient, result- Five cell cavities represent the best compromise to
ing in a poor “real estate” gradient from an expensive preserve a three section linac design, while the maximum
structure. Alternative schemes for the low energy section number of cavities per cryomodule has been limited
may allow the extension of the initial section to higher respectively to two, three and four, on the basis of the
energy (>100 MeV). Nevertheless, elliptical cavities can focusing scheme outlined above. The considerations
capture efficiently the 100 MeV beam and higher presented in this section, combined with basic beam
injection energies have little impact on performance. dynamics, and details of the specific system being
Due to the wide proton velocity variation from 100 considered, yield the present reference design.
MeV to over 1 GeV, and to the limited velocity
acceptance of any resonant accelerating structure, the 4 A DESIGN FRAMEWORK
linac has to be sectioned. Analysis indicates three sections For our design, we assume the parameters and limits
are sufficient to cover the energy range from 100 MeV to given in Table 1. The assumed values for the peak
1.6 GeV. electric field and RF power transmitted through the
couplers have been previously achieved CERN test bench
3 OPTIMIZATION CONSIDERATIONS values [4]. The nominal current of 25 mA is reachable
There are a number of constraints and practical with a single coupler per RF cavity. To provide a possible
arguments which limit the usable parameter space. A beam current upgrade, the cavities will be designed from
central need is to make the machine cost effective in the beginning with multiple coupler ports, despite the fact
terms of both capital and operational costs. In general, the that the present design parameters are based on a one
capital cost is reduced by decreasing the length of the coupler solution.
accelerator, while the operational cost is reduced by The minimum cavity matched (“effective”, not
having “efficient” acceleration — good RF to beam geometric) b has been set to 0.5 based on preliminary
coupling. Additionally, the RF is a major component of electromagnetic and mechanical studies. The reference
the system cost, and care must be taken to optimally use design of such a cavity, which features an appropriate
the klystron’s power rating, while saving a proper margin stiffening structure and an iris radius of 10 cm, has a peak
for controls and mismatching. surface to accelerating field ratio of 3.2. For the assumed
On the basis of the existing 1.3 MW CW LEP kly- peak surface field of 15 MV/m and considering the poor
stron, a nominal power of 1 MW is used for beam accel- packing factor of the active cavity length to the lattice
eration in most of the linac structures. This power feeds period the average “real estate” accelerating gradient of
four cavities in the high energy section (one cryomodule) the low b section is 1 MeV/m.
and six (two cryomodules) in the intermediate one. For Table 1: Parameters and limits used for the design of the
general machine reliability, the first section uses smaller superconducting linac.
klystrons (500 kW) feeding four cavities (two cryo-
modules). This scheme is compatible with further current Parameter Value
Initial Energy 100 MeV
upgrading, based on multiple RF coupler operation. Final Energy 1600 MeV
Efficient use of the cryostats is also demanded because Beam Current 25 mA
of the capital cost, however beam dynamics consider- Beam Power 40 MW
Peak Surface Field £ 15 MV/m
ations affect the arrangement of the accelerating cavities. Minimum Cavity b 0.5
A major consideration is the need to keep the beam size RF Coupler Max. Power 250 kW
well under control, and this implies a limit on the Available Power/Klystron 1 MW
focusing lattice period length. In turn, the focusing
The procedure for determining the b values and the
period, in combination with the need for warm diagnostic
transition energies of the sections depends strongly on the
boxes in each cell, sets limits on the cryomodule length.
operating range allowed by the cavities. Initial studies
The cost and the added length of the cold/warm
utilized an operating range which was symmetric in the
transitions require a compact focusing scheme, thus a
velocity acceptance curve [1], while present work relies
singlet scheme seems less desirable than a doublet
on an asymmetric approach.
focusing structure.
The present method sacrifices some RF efficiency in
It is standard practice to use multicell cavities in order
the first part of each linac section for improved average
to increase the ratio between the active and physical
(real estate) acceleration gradient. Since the particle vel-
cavity length. The velocity acceptance of the cavities
ocity changes most rapidly at low energies, the initial few
decreases with the number of cells in the structure, and
cavities in each section are used for velocity matching by
that, in turn, can affect the number of sections needed in
operating them at lower RF power than the subsequent
the linac for efficient acceleration. Increasing the active
cavities. Such an approach allows the totality of a sec-
cavity length also increases the RF power needed for the
tion’s cavities to operate closer to the maximum acceler-

1127
ating gradient (see Fig. 1), while the best klystron effic- The linear beam dynamics with space charge has been
iency is set for the majority of a section’s cavities which, studied; and a procedure for matching the beam between
in the last part of each section, operate at a constant the linac sections has been devised. The matching of the
energy gain, independently of the proton velocity. short period FODO structure of the DTL to the long
Moreover, the smooth ramping of the effective energy period doublet structure in the SC linac is still under
gain at the entrance of each linac section is expected to investigation.
help the beam matching between the different lattice A lattice, consistent with the above machine parameters
periods. In the high energy section the energy gain per is shown in Fig. 2. The maximum needed quadrupole
cavity of 10 MeV and a better packing factor allows a integrated strength is approximately 1.6 T, which is
real estate gradient of 2.5 MeV/m. achievable within the 60 cm reserved for each quadrupole
12
in the 3 m intermodule section. A normalized emittance
of 1 p mm-mrad (transverse) and 1 p deg-MeV
Energy Gain/Cavity [MeV]

10 (longitudinal) has been assumed [3], and a constant

8 synchronous phase of -30° has been used along the linac.
6 5 R&D PROGRAM
4 On the basis of the present reference design, the objective
of the two year program, for the high energy part of the
2
100 500 900 1300 1700 linac, is to develop the technology of the most crucial
Beam Energy [MeV]
components, while finalizing the design of the machine.
Figure 1: Maximum (upper line) and actual (lower line) energy gain We plan to develop, together with the industry, Nb
along the accelerator as a function of the beam energy, assuming a single cell cavities at the lowest beta, to be measured at
peak surface field of 15 MV/m and synchronous phase js= -30°.
CERN, and a complete five cell copper structure for
In Table 2 we list parameters for the three sections. Note mechanical and RF warm tests. A complete copper, Nb
that a shorter lattice period is used at the lower energy sputtered, cavity at the highest beta is also planned; the
ranges, as is required to compensate the effect of R&D on single cells being in progress at CERN [5].
transverse RF defocusing.
ACKNOWLEDGEMENTS
Table 2: Machine and lattice parameters. We are grateful to Simone Visonà for the cavity studies
Parameter Value and Massimo Bonezzi for the technical drawings.
Total length ~ 720 m
Number of cells/cavity 5 REFERENCES
Number of couplers/cavity 1
Section 1 2 3 [1] C. Pagani, G. Bellomo, P. Pierini, Proc. of the XVIII Int. Linear
Acc. Conf., eds. C. Hill, M. Vretenar, CERN 96-07, (1996) p.107.
Section total length [m] 96 146 475 [2] M. Prome, Proc. of the XVIII Int. Linear Acc. Conf., eds. C. Hill,
Injection energy [MeV] 100 190 428 M. Vretenar, CERN 96-07, (1996), p. 9.
Section period [m] 8 11.2 15.3 [3] J. Schneider, Proc. of the XVIII Int. Linear Acc. Conf., eds. C.
# of cavities/section 24 39 124 Hill, M. Vretenar, CERN 96-07, (1996) p. 22; and APT
# of cavities/cryomodule 2 3 4 Conceptual Design Report, LANL.
[4] G. Geschonke, Proc. of the XVIII Int. Linear Acc. Conf., eds. C.
# of cavities/klystron 4 6 4 Hill, M. Vretenar, CERN 96-07, (1996), p. 910.
Klystron rating (kW) 500 1300 1300 [5] E. Chiaveri, Private communication.
Cavity matched b 0.5 0.65 0.85
Accel. Gradients [MV/m] 4.3 5.4 6.4
Power/coupler [kW] / 25 mA 100 160 250

Figure 2: Layout of the basic cell in the high energy section (the low and intermediate energy sections differ only in the cryomodule len gth). The
3 m intermodule distance accomodates the quadrupole doublet, steering magnets (not shown), and a diagnostic and pumping box. Th e beam pipe
aperture is maintained at 20 cm or greater throughout the linac.

1128