THE ESS SC REFERENCE LINAC
K.Bongardt 1, A.Letchford 2 for the ESS accelerator team
FZ Juelich, Germany RAL, UK
Abstract applications where the integrated intensity is the important
parameter. The Short Pulse (SP) target station also
This note is a short summary of the ESS SC reference receives 5 MW of beam power but from 1.4 µsec proton
linac with its innovative chopper/collector system and the pulses arriving at a frequency of 50 Hz (100 kJ/pulse) for
1120 MHz high frequency SC part .More details can be applications where the peak intensity in the pulse is the
found in  and in the three adjacent papers. key parameter.
The ESS SC reference linac starts with a low frequency The high total beam power (10 MW), the demand for low
front end, housing the innovative chopper/ collector loss in the accelerator and the combination of short and
system. High frequency SC elliptical cavities are used long pulses put rather stringent requirements on both the
from 400 MeV on. The accelerator layout and the main accelerator and the target stations.
parameters are presented. Outlined are two dedicated
test-stands necessary to construct the ESS facility within The ESS SC reference linac
the new time schedule .
In the ESS technical proposal  both NC (normal
Introduction conducting) and SC (super conducting) solutions, with
different frequencies, were described. All the described
The aim of the ESS project is to design an affordable, proposals were feasible and estimated to result in almost
technically feasible next generation neutron source that on the same cost. A specific SC reference design has now
completion will provide World-leading performance for all been finally selected by the ESS accelerator team and
classes of instrumentation. The result of a close dialogue approved by the ESS Council.
between users, instrument designers, target and accelerator
experts is a facility with two complementary target The layout of the accelerator system is shown in Fig 2, and
stations. This is a unique feature of the ESS . Fig 1 the main parameters summarised in Tab 1. About
show for the scientific most important applications the 10 MW power can be saved compared to a ESS NC linac.
performance of the ESS SP&LP facility compared to a The time structure for the ESS SP&LP scheme is shown
1.4 MW SNS SP facility only. Clearly can be seen the in Fig 3.Not shown are the two other SP pulses. The main
advantages of a 5 MW LP target station : only scaling in RF parameters of the ESS linac are given in Tab 2.
power would give the ESS facility only a factor 7 better
performance than the SNS SP one. The main difference between the ESS accelerator and the
accelerators currently under construction for SNS  and
The Long Pulse (LP) target station receives 5 MW of beam J-PARC  is the requirement of simultaneously
power from 2 ms long proton pulses with a frequency of delivering both short and long pulses .
16 2/3 Hz (300 kJ/pulse). This is ideal for broad bandwidth
Fig 1 : Normalized performance of the ESS SP&LP facility compared with a 1.4 MW SNS SP facility only
H Ion Sources : Achromat SP
280 MH z 1120 M Hz SC linac and rings Target
65 m A each 560 MHz
RFQ DTL 228 m A
2 x 57 m A 114 m A 560 M Hz
CCDTL CCL ß=0.8 CCL,
SP , LP
Funnel Energy R am ping /
2.5 M eV 20 M eV 100 M eV 400 M eV 1334 M eV
262 m 308 m 78 m
Fig2: The ESS 1120 MHz superconducting (SC) reference Linac.
PRF (pulses per second) 50 16.67
Beam pulse length( ms) 2x0.48 2.0
Beam duty factor 4.8% 3.3%
Non-chopped beam current (mA) 114 114
Chopping factor 70% 70% 100%
Final energy (MeV) 1334 1334
Peak beam power (MW) 107 107 152
Mean beam power (MW) 5.1 3.5 5.1
Pulse gaps, ring separation (ms) 0.1
280/560 MHz NC-Linac
Energy range ( MeV ) 400
NC linac length (m) 262
Peak RF power (nominal)(MW) 64 78 (100%)
RF pulse: length (msec) / duty cycle (d.c.) 1.4/7.0% 2.3/3.83%
Wall plug RF power (MW) 12 8
(30 % RF control included)
1120 MHz SC-Linac
Energy range ( MeV ) 400 –1334
SC linac length (m) 308
Accel. gradient in SC cells ( MV/m) 10.2
Peak RF power (nominal) (MW) 75 107 (100 %)
RF pulse: length (ms) / d.c. 1.4/7.0% 2.3/3.83%
Wall plug RF power (MW) 15 11
(30 / 40 % RF control included) (40 % ) (30 % )
AC Cryo power (MW) 2.4 1.6
Tab 1: Main parameters for the ESS SC reference linac with its simultaneous SP&LP operation.
During commissioning the LP beam will also be chopped.
In order to deliver 5 MW beam power in about 1.4 µs to Figure 3. This makes the ESS facility unique in its neutron
the SP target, the ESS facility needs 2 accumulator rings scattering performance, but is challenging for some ESS
with 35 m mean radius in a shared tunnel. Ring injection linac components : the front end with its chopper/collector
utilises H- stripping injection with painting in the system and layout / RF control of pulsed SC cavities.
horizontal, vertical and momentum dimensions. Each ring
is filled sequentially and injection is limited to 0.48 ms and The LP target station needs a 2 ms linac pulse every 60 ms
600 turns per ring in order to limit the temperature rise in or at 16.67 Hz repetition rate with 114 mA pulse current.
each stripping foil. The linac pulse is chopped to 70 % of This can be achieved with two H- ion sources at 65 mA
the 800 ns ring revolution time at the ring revolution each, funneled together at about 20 MeV. No beam
frequency to leave a gap for the ring extraction kicker chopping is required here - see Fig 2. The RF control
magnets resulting in low ring losses. A 100 µs gap is system for pulsed SC cavities has to be very carefully
required for vertical deflection of the linac beam between designed as the system is matched only for the 2 msec un-
the rings. The pulse structure in the linac is shown in chopped LP pulse, but quite heavily mismatched for the
70% chopped SP. Operating at high frequencies and / or elements with a rise time of less than 2 ns to avoid beam
small accelerating gradients is a possible solution here. loss further down the accelerator. The beam collection
system must be able to cope with up to 10 kW power,
The chopping line for the ESS linac must be able to switch since both the SP and LP beam will be chopped initially.
the beam on and off between RF bunches resulting in
S h o r t p u ls e U n c h o p p e d lo n g p u ls e
50H z 5 0 /3 H z
P G e
I b ea
0 .3 0 .4 8 0 .1 0 .4 8 0 .3 2 .0
Fig3: Pulse sequence for the ESS SP&LP scheme ( not shown are the two other SP pulses)
Vca= Cavity voltage, Ibea = beam current relative to a chopped beam, PGe power from the RF generator
for the SC cavities. Between the 2 ring pulses, the RF generator power as to be reduces to about 25 %
in order to keep the accelerating voltage unchanged in the SC cavities.
The ESS reference linac with 10 MW of beam power, In tab 2 , the main parameters for the RF systems along
shared between the SP and the LP target stations, cannot the ESS SC reference linac are shown. The funnel section
be a direct copy of any current or planned linear at 20 MeV needs 840 MHz bunching cavities  . Multi
accelerator. The ESS accelerator team therefore had to find beam klystrons could be a choice for the CCL section, as
a linac design that is cost effective and that will provide even at 5 MW they only needs voltages below 100 kV.
the 10 MW of beam
power with a high degree of certainty. The 280/560 MHz Operation of both long and short pulses may require two
normal conducting linac provides the 10 MW of beam H- ion sources in each leg of the front
power with a high degree of certainty. The 280/560 MHz end. Neither the H- ion-sources nor the chopper /collector
normal conducting (NC) linac design described in ESS system will be overloaded, but both
Volume 3  is a technically feasible and physically beams must be combined at 20 MeV in the funnel.
robust design with a reasonable cost estimate. Selecting Progress in high intensity H- ion-sources indicates that the
1120 MHz elliptical SC cavities above 400 MeV and using ESS SP & LP requirements may be achieved with two H-
the original 280/560 MHz NC linac below 400 MeV was sources only, if the beams are separated by 10 ms.
found to promise both good beam quality with low losses
and competitive construction and operating costs. The To replace the warm parts of the ESS reference linac up to
resulting main ESS linac parameters are shown in Table 1 400 MeV by SC low or medium ß structures is not
and, as shown in Figure 1, the ESS reference linac starts considered to be a valid alternative due to the ESS linac
with a low frequency front end, houses an innovative RF duty cycle of only 12 % and the expected time scale for
chopper/ collector system, combines two H- beams at 20 ESS, even if it is delayed by a few years. The ESS
MeV and uses high frequency SC cavities for beam accelerator team regards SC low and medium ß structures
acceleration above 400 MeV. Placing 78 m behind the ESS as an ongoing long-term R&D programme.
linac NC ß=0.91 ,560 MHz cavities for bunch rotation &4
MeV energy ramping, see Fig 1, leads to a much easier RF
control system than by using for SC cavities.
The 1120 MHz SC linac is 308 m in length and 172
cavities are required with only one SC main coupler per
cavity designed for 0.85 MW peak power. Although cavity
and cryostat can be scaled from the J-PARC 972 MHz SC
proton linac test-stand, R&D is required for the SC main
coupler. As the cavity bandwidth and stiffness is increased
at the higher frequency, an 1120 MHz SC linac is well
suited to guarantee loss free injection into both ESS
compressor rings whilst not being hindered by the ESS
Accelerator Freq. (MHz) Peak power (nominal) (MW) Number of klystrons &
structure (100 % chopped beam) Peak power ( with RF control)
Structure Beam Total
2 × RFQ 280 1.6 0.45 2.1 2 × 1.3 MW
2 × Chopper 280 0.15 ~0 0.15 10 × ~0.04 MW
2 × DTL 280 1.3 2.0 3.3 4 × 1.3 MW
Funnel 840/560/280 0.7 ~0 0.7 12 × ~0.08 MW
CCDTL 560 5 9 14 15 × 1.3 MW
CCL 560 22 35 57 22 × 5 MW
SCL 1120 ~0 107 ~107 172 × 0.85 MW
BR cavity 560 0.6 0.46 1.10 1x1.45 MW
Σ ~31 152 ~185 238
Tab 2: Main parameters for the RF systems along the ESS 1120 MHz SC reference linac
Status of the ESS accelerator system – the way ahead
To have a decision on the ESS in 4-5 years time requires a
project following the proposed time schedule associated by
competence centres to look into key technological  “ The ESS linac & front end “ in :
questions. The ESS linear accelerator needs two dedicated The ESS Project Volume III Update Report :
test-stands for prototyping and development: a 2.5 MeV http://www.neutron-eu.net/en/index.php?cat=93&id=117
chopping line as test-bed for front end components and a  K.N.Clausen , Proc. ICANS –XVI , Düsseldorf-Neuss,
complete 1120 MHz cryomodule as full power test-bed for May 2003 ,p.61 and : http://www.ess-europe.de
SC couplers.  “The ESS Project Volume III – Technical Report”,
http://www.ess-europe.de ISBN 3-89336-303-3 (2002)
Acknowledgements 438 pages.
This short summary for the ESS SC reference linac is  The SNS project http://www.sns.gov/
based on the considerable work of many colleagues in the  Y.Oyama, “Present Status of J-PARC-High Intensity
last 6 month. Without all these contributions it would have Proton Accelerator Project in Japan “ , Proc. ICANS –
been impossible to describe the ESS SC reference linac as XVI , Düsseldorf-Neuss, May 2003, p.7
it is up to now. and http://j-parc.jp/T