PSI experience with high power beam handling, activation and radiation protection. Mike Seidel (PSI).

1. PAUL SCHERRER INSTITUT
PSI experience with high power beam handling,
activation and radiation protection
M.Seidel, PSI
ESS - Bilbao Initiative Workshop, March 16-18, 2009

2. Outline
the PSI facility and its performance
[overview, cyclotrons, availability, resonators, target]
intensity limiting effects
[space charge and beam blow up, scaling law for cyclotrons, round
beam, beam dynamics simulations]
high power related accelerator aspects
[special controls/diagnostics, interlock systems, shielded beamlines,
activation]
specific infrastructure
[rad.safety, disposal, legal requirements and all that …]
conclusion
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

3. Overview PSI Facility
Injector II Cyclotron 72 MeV Cockcroft Walton
isotope production 870 keV transfer channel
µ
(Ib <100µA)
Ring Cyclotron 590 MeV
72 MeV transfer channel
2.2 mA /1.3 MW
µ/π secondary beamlines
π
target M (d = 5mm)
target E (d = 4cm) 1.5 mA /0.9 MW
CW operation
SINQ
spallation source
proton therapie center
[250MeV sc. cyclotron]
SINQ transfer channel
SINQ
experiments
[Markus Lüthy]
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

4. Spallation Source Experimental Area K. Clausen
– 13 Beamlines
NEUTRA
TRICS HRPT POLDI MORPHEUS AMOR SANS-I
CNR
MARS
DMC SANS-II
RITA-II TASP FOCUS
Eiger M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

5. CW Acceleration using a
Sector Cyclotron
590 MeV Ring Cyclotron
(magnets) in operation for 30+ years
- 8 Sector Magnets 1T
- Magnet weight ~250 tons
- 4 Accelerator Cavities850kV (1.2MV)
- Accelerator frequency: 50.63 MHz
- harmonic number: 6
72 → 590MeV
- beam energy:
- beam current max.: 2.2 mA
- extraction orbit radius: 4.5m
⋅
- relative Losses @ 2mA: ~1..2⋅10-4
- transmitted power: 0.26-0.39 MW/Res.
Pro: Con:
- CW operation is inherently stable - inj./extr. difficult, interruptions, losses!
- efficient power transfer with only 4 resonators - large and heavy magnets (therm. equiliblium!)
- cost effective, compact - energy limited ~1GeV
[- no pulsed stress in target] [- no pulsed structure for neutrons]
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

6. history max. current of the PSI accelerator
4 Cu Resonators
license temporary operation complete
with 2.2mA given
1.3MW !
beam current is limited
by beam losses;
upgrade path foresees
constant absolute
losses by improvements
of the accelerator
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

7. reliability and trips
typical: short (30sec) interrupts; trips of electrostatic elements; loss interlocks
average availability: 90%; from June to December 2008: 94%(!)
on average: 25 trips per day
in the discussion
on application of
cyclotrons for
duration of run period (this case: 21hours!)
ADS systems
the frequency of
interruptions is of
major interest
duration of interruption ~30sec
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

8. major component: RF Resonators for Ring Cyclotron
• the shown Cu Resonators have replaced the original Al resonators [less
wall losses, higher gap voltage possible, better cooling distribution, better
vacuum seals]
⋅
• f = 50.6MHz; Q0 = 4⋅104; Umax=1.2MV (presently 0.85MV→186 turns in
cyclotron, goal for 3mA: 165 turns)
• transfer of up to 400kW power to the beam per cavity
• deformation from air pressure ~20mm; hydraulic tuning devices in feedback
loop → regulation precision ~10µm
resonator
inside
hydraulic tuning
devices (5x)
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

9. Cyclotron Facility Upgrade Path
• keep absolute losses constant; increase acceleration
voltage and beam quality, better turn separation at
extraction
∝ [turns]3 ∝ [charge density (sector model)]
• losses
× [accel. time] / [turn separation] (W.Joho)
planned turn numbers and voltages
turns Ring turns Injector
now 202 81
(2.0mA) (Upeak≈3.0MV) (Upeak≈1.12MV)
inter. step ~180 ~73
(Upeak≈3.3MV) (Upeak≈1.25MV) historical development of turn
(2.6mA)
numbers in PSI Ring Cyclotron
upgrade ~165 ~65
(Upeak≈3.6MV) (Upeak≈1.40MV)
(3.0mA)
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

10. critical: electrostatic elements
parameters
extraction chan.:
principle of extraction Ek= 590MeV
channel E = 8.8 MV/m
θ = 8.2 mrad
ρ = 115 m
U = 144 kV
injection
element in Ring
Tungsten stripes
beam pattern on
outer turns in Ring
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

11. extraction losses reduced by faster acceleration
absolute loss (nA) in Ring Cyclotron
as a function of current
achieved in 2008:
gap voltage increase:
780kV → 850kV
turn number reduction:
202 → 186
figure shows absolute
losses for optimized
machine setup
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

12. avoid tail generation with short bunches
numerical study of beam dynamics in Ring Cyclotron
→ behavior of short bunches, generated by 10’th harmonic buncher
→ optimum parameters of flat-top cavity at these conditions
after 100 turns!
-multiparticle
simulations
-106 macroparticles
[on supercomputer
simulation of all protons in
bunch feasible!]
- precise field-map
- bunch dimensions:
radial/long. bunch distribution, varying initial bunch length
σz ~ 2, 6, 10 mm;
σxy ~ 10 mm
→ operation with short
bunches and reduced
flattop voltage seems
possible
court.: J.Yang CAEA
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

13. “round beam” – space charge in cyclotrons
qualitative picture: coordinate frame
moves with bunch
protons in the field of a round,
short bunch + vertically
oriented magnetic field
(neglect relativistic effects and
focusing)
[Chasman & Baltz (1984)]
though the force is repulsive
a “bound motion” is
established
→ for short bunches a round
beam shape is formed
→ a round beam is observed
in the Injector II cyclotron
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

14. High Power Meson Production Target
TARGET CONE
3.0mA o.k., limit: sublimation
Mean diameter: 450 mm
Graphite density: 1.8 g/cm3
Operating Temp.: 1700 K
Irrad. damage rate: 0.1 dpa/Ah
Rotation Speed: 1 Turn/s
Target thickness: 60 / 40 mm
10 / 7 g/cm2
Beam loss: 18 / 12 %
Power deposit.: 30 / 20 kW/mA
SPOKES
To enable the thermal expansion p-beam
of the target cone
BALL BEARINGS *)
Silicon nitride balls
Rings and cage silver coated
Lifetime 2 y
G.Heidenreich et. al.
*) GMN, Nürnberg, Germany
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

15. high power beams – technical safety
beam hits material (e.g. steel vacuum chamber):
beam power
∆T P0
≈ = 250.000 K/sec
∆t 2π σ xσ y λI [g/cm ] c p
2
melts after
8ms !
heat capacity
beam size material specific
interaction length
→ fast and reliable interlock systems
are necessary to avoid damage !
[even more so in pulsed machines
with higher P0]
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

16. Diagnostics for Machine Protection
System based on ca. 150 interconnected very fast (<100µs) hardware CAMAC
and VME modules treating about 1500 signals provided by the equipment:
Ionisation chambers as beam loss monitors with fixed warning and
1.
interlock limits; critical ones also with limits as function of the beam current.
Simple and reliable device
Warning level
log scale
Dyamic window
Actual losses
Permanent display of losses
→ losses outside margins are interlocked (including low values)
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

17. Diagnostics (cont.)
2. Segmented Collimators measuring the balance of right and left, up and down
scraped beam currents
interlock is generated in case currents or asymmetry of currents
exceeds margin
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

18. Diagnostics (cont.)
3. Beam current transmission monitors compare the beam current at different
spots for detecting loss of beam, normally 100% of transmission except at
the targets and when beams are splitted.
100 % transmission in main cyclotron
transmission monitoring
100 % transmission in beampipes,
through Meson
except for splitted beams
production target E
97 % transmission of thin target M
70 % transmission of thick target E
Integration time:
110 ms at 0 µA down to 10 ms above 1.5 mA
Window:
± 5 µA at 0 µA and ± 90 µA at 2 mA
4. Many other signals: validity window on magnet settings, cavity voltages, …
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

19. practical experience: overheating of Lead filled steel
tubes in spallation target because of wrong beam optics
cause: wrong beam optics installed, caused narrow wrong beam optics (pink), matched for
beam width → power density too high 6cm Meson target but accidentally
applied for 4cm target
usual current density <=40µA/cm2, in this case:
70µA/cm2 note: βγεx = 7.5/40 mm⋅mrad
⋅
before/after 4cm graphite target
Pb filled stainless steel
tubes
Window Pb filled Zy-2
cooling tubes
damaged
Target
tubes
cooling
neutron
radiography
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

20. special interlock system: direct current density monitoring
beam projection on a glowing mesh observed with an optical camera;
problem: lifetime of camera; new version: glass-fiber optics to
shielded camera position
metallic mesh
(tungsten)
vertical cut through
beamline and target
showing installation
of VIMOS camera
2d display and
projection
VIMOS, K.Thomsen
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

21. Shielding
strategy: use standardized (mobile) concrete/iron blocks for shielding purposes
in PSI experimental hall: totally ~10.000 shielding blocks; in sum 32.000 tons of
weight; all documented in CATIA CAD system
iron is 30% fraction by the number
14 different standardized shapes
20% per the number of the blocks have special shapes with odd dimensions
PSI
experimental
hall
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

22. shielding (2) – target / 590MeV transport chan.
elaborate shielding required
reliability of activated components!; water cooling; few electrical connections
Component activation in beamline up to ~300Sv/h!
secondary shielding
platform for electronics, pumps etc.
4cm Target
Beam Dump
collimators
BEAM DUMP
vacuum chimneys for inserts,
primary shielding pumping connections
Iron ! side view
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

23. predicted activation at Meson production target
30% beam loss; target (1..3Sv/h) exchange via special flask
top view
D.Kiselev, M.Wohlmuther
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

24. measured component activation – Ring Cyclotron
activation level allows for necessary service/repair work
• personnel dose for typical repair mission 50-300µSv
• optimization by adapted local shielding measures; shielded service boxes for
exchange of activated components
• detailed planning of shutdown work
activation map of Ring
Cyclotron
(EEC = electrostatic
extraction channel)
personal dose for 3
month shutdown (2008):
57mSv, 188 persons
max: 2.6mSv
cool down times for
service:
2000 → 1700 µA for 2h
0 µA for 2h
map interpolated from ~30
measured locations
EEC
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

25. overview: auxiliary expertise for High Power Operation
Facility Licensing
Personnel Safety
[documentation and
Systems
paperwork, communication Radiation Safety
[access to accelerator with legal authorities]
bunkers, experimental [personnel dose
areas] monitoring, radiation
monitoring etc.]
Radiative Waste
Disposal
High Power
[declare nuclide technical safety systems
Accelerator
inventory, predict [interlock systems]
disposal costs,
Facility
simulate temporary
storage decay etc]
technical infrastructure specific infrastructure
[vacuum, cooling, RF, magnets, [HotLab, mobile shielding
power supplies, survey, devices, radioanalytics etc.]
diagnostics, controls etc.]
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

26. special infrastructure – hot cell facility
view through lead-glass window:
Meson production target wheel
irradiated SINQ spallation target,
taken out of the cover
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

27. special infrastructure – mobile shielding / exchange flask
• mobile and specifically adapted shielding devices are used for
targets and critical components as extraction elements or septum
magnets
• target exchange flasks are complicated and expensive devices
[heavy, motors, instrumentation, SPS controls]
picture: exchange flask for
Meson production target E
(4cm graphite wheel)
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

28. numerical modelling in the context of high power beams
particle generation/transport without/with MCNPX,
electromagnetic fields MARS
activation of components PWWMBS,
MCNPX
radioactive waste disposal, declaration of PWWMBS,
nuclide inventory ISRAM
design/estimation of shielding MCNPX,
ATTILA
beam dynamics
codes omitted thermo hydraulic / cooling problems ANSYS,
here!
CFD-ACE
Activity [Bq]
example:
1.E+14
predicted nuclide 1.E+12
inventory for old 1.E+10
beam dump 1.E+08
Bq
1.E+06
1.E+04
M.Wohlmuther 1.E+02
S.Teichmann 1.E+00
D.Kiselev
C5
3
N9
Fe 6
Be 0
60
Zn 2
54
M4
93
m
-3
M4
-1 5
0
-5
-6
i-6
i-5
2
l-3
-1
-4
Ag -6
-1
08
H
o-
a-
n-
o-
Fe
Ti
C
N
N
C
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

29. Summary
• the PSI accelerator delivers 1.3MW beam power
in continuous mode; average reliability is 90%;
~25 trips per day
• performance is limited by uncontrolled losses →
upgrade path through faster acceleration (higher
resonator voltages) + bunch compression; max.
acceptable activation of serviced accelerator
components: ~ several mSv; total loss ~ 200W
• infrastructure for dealing with activated
components and formal aspects connected to
safety and radiation issues are challenging and
are often underestimated!
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)

30. Thank you for your attention!
M.Seidel, ESS Bilbao Initiative Workshop, March 16-18 (2009)