XFEL Project accelerator Overview and recent developments

1
XFEL Project (accelerator)Overview and recent
developments

• R. Brinkmann, DESY
• TESLA Collaboration Meeting, Frascati May 2003

2
TESLA TDR, March 2001 integrated XFEL
3
Considerations leading to XFEL TDR-update, autumn
2002

• Avoid strong coupling of XFEL and LC parts of
  TESLA project during construction, commissioning
  and operation stages (and approval)
• Gain flexibility in operation parameter space
• ? XFEL driver linac in separate tunnel
• Limitation of additional cost
• Reduce linac length energy to 1.5km, 20 GeV
• Difference in accelerator cost
• (sep. linac TDR2001) 196 Mio.
• Reduce of photon beam lines 10 ? 5, of
  experiments
• 30 ? 10

4
Comparison of parameters (1Å, fixed-gap undulator)
5
Injector TTF-II
6
Simulations (60MV/m cathode field) indicate ? lt 1
mmmrad possible Including BCs (CSR!), phase
space structure non-trivial ? start-to-end
simulations Linac wakefields unlikely a serious
problem Exploration of bunch parameter space
(e.g. charge vs. emittance bunch length) ?
7
3.9km
8
(No Transcript)
9
Proposed modifications in recent discussions

• Shortening of the accelerator (Injector linac
  collimation/diagnostic etc.) tunnel to 2km
• XFEL site layout not necessarily linked to LC
  site (except dont exclude LC construction at
  foreseen Hamburg/S.H. site)
• TTF-like modules with 8 cavities per module of
  modules per Klystron ?
• Flexibility in duty cycle ? (see below)

10
Duty cycle limitations I Cryogenics

• TDR layout for module/He distribution (GRP ? ,
  pressure drop) allows for upgrade to 800 GeV
• ? At 23 MV/m, 1.5km linac could be operated up
  to about 20 Hz rep rate (2 duty cycle)
• Required cryogenic plant would have approx. the
  size of one of the six TESLA-500 LC plants
• From cryogenics point of view, could scale duty
  cycle as 1/energy2

11
How about CW operation??

• At 23 MV/m, Q0 1010, rf losses at 2K are 55W/m
  CW ? 20 GeV linac would require 3 times total 2K
  capacity for TESLA-500 LC (forget it)
• At half gradient (linac length ? 3km!), Q0
  2?1010, rf losses at 2K are 7W/m CW ? required 2K
  capacity still 45 times one 5km unit of
  TESLA-500 LC (? considerable additional
  investment and operating costs, modification of
  He distribution?)
• A big issue for all (near-) CW considerations at
  present, no suitable beam source available

12
Duty cycle limitations II RF system

• present design of modulator/klystron station can
  operate at max. 10 Hz, 10MW, 1.4ms pulse length,
  65 efficiency (average power into klystron
  220kW)
• Higher duty cycle at lower peak power possible as
  long as average power from modulator/into
  klystron gun is kept ? above limit (careful
  DC?RF efficiency drops at lower power!) concerns
  IGCTs at high rep rate, RF drive power
• Scale beam pulse current with acc. Gradient
  (beam energy) ? loaded Qext constant, or
  optimise for constant beam current, variable
  Qext
• Assume 33 RF power overhead Prf 1.33 Pbeam

13
Max rep rate and beam power, four vs. six
modules/klystron
Remark av. Beam power is maximum possible -
because of beam dump (solid absorber option) we
wanted to limit Pav? 600kW
14

• What can we gain from variable Qext? Example (6
  modules)
• Keep Ib 5 mA const., scale Qext ? Eacc
• Attractive option shorter pulses/higher frep
  (RF gun!)

15
XFEL Linear Collider Synergies
Working towards getting ready for start of
construction of 20 GeV s.c. linac in 2 years
from now is a big step forward for making TESLA
technology available for large-scale projects The
issues in common for developing the 500-800 GeV
LC s.c. linac and getting ready for constructing
the XFEL linac (by far) outweigh those issues
which may be different and may require potential
priority conflict discussions There is also
overlap between LC and XFEL for a number of other
design issues and sub-systems (e.g. failure
handling/operational reliability, beam size and
profile monitoring, fast orbit feedback, LLRF)
16
Example 1 Tunnel Layout
E.g. Electronics in tunnel/radiation environment
(? test in DESY-LINAC-II) Handling of RF and
cavity failures Stray fields? Supports and
alignment
17
Example 2 fast kicker systems

• Damping ring re-distribute the train of bunches
  in time (compress/de-compress at
  injection/ejection)
• XFEL user beam lines distribute bunches within a
  train to different beam lines (possibly
  extraction points at different energies, etc)
• Technology for both applications may also be
  similar to fast orbit feedback requirements

18
As for the LC, DESY will work out a site proposal
for the XFEL Not yet decided green field vs.
near-DESY
19
Expected orbit vibration from linac FODO lattice
0.1? (rms) from noisy (workday) HERA data
not a problem (?)
HERA tunnel
Ellerhoop (barn)
20
Concluding remarks

• Sequence of projects on scales of time
  realisation probability TTF-II (scTech.VUV-FEL)
  XFEL LC
• XFEL is a great project for DESYPartners and a
  great opportunity for DESY to host the project
• It is also a great opportunity for all of us