T. Kamps / Proceedings of the 2004 FEL Conference, 381-384 381 COLLIMATION SYSTEM FOR THE BESSY FEL∗ T. Kamps Berliner Elektronenspeicherring-Gesellschaft fur¨ Synchrotronstrahlung BESSY, D-12489 Berlin, Germany Abstract magnet undulators. Each undulator is composed of mod- ules with length ranging from 1.6 m to 3.9 m, the total Beam collimation is an essential element for the success- length of the high energy FEL line is 70 m. A FODO lat- ful running of a linear accelerator based free electron laser. tice is superimposed onto the undulator structure with the The task of the collimation system is to protect the undu- quadrupole magnets located at the intersections between lator modules against mis-steered beam and dark-current. the undulator modules. For the present analysis in total 14 This is achieved by a set of apertures limiting the succeed- undulator modules are taken into account for the simulation ing transverse phase space volume and magnetic dogleg studies. The gap between the undulator poles is 10 mm, the structure for the longitudinal phase space. In the follow- beampipe radius is 4 mm. The undulator magnets are made ing the design of the BESSY FEL collimation system is of NdFeB in order to obtain large undulator K parameters. described together with detailed simulation studies. This material is very sensitive to irradiation, especially to the reduced dose deposited by high-energy electrons (> MOTIVATION 20 MeV). From studies with NdFeB under electron irra- Experiences from linac driven FEL operation (for exam- diation [4] the limit for acceptable beam loss in the undula- ple at TTF1 [1]) indicate that beam collimation is essential tor chamber has been derived. It can be shown that for 1% in order to provide protection of the undulator modules en- demagnetization the acceptable beam loss in the vacuum suring safe FEL operation. chamber is 300 nC [3]. It is evident that a collimator sys- The collimator section serves as transverse and longitu- tem together with a fast-acting machine protection system dinal phase space filter which lets particles within a cer- must be included in the FEL design. tain energy bandwidth and transverse phase space volume pass. Particles outside this region are blocked and cannot Machine Protection System be lost in the undulator modules. In the transverse plane this is achieved by a set of apertures limiting the succeeding The three FEL lines will initially operate at a bunch rep- transverse phase space volume. In addition the collimator etition frequency of 1 kHz each. With a bunch charge of is designed to remove halo and dark-current created and 2.5 nC and a beam energy of 2.3 GeV the beam power is 3 6 kW. At no location in the≤ accelerator the structures transported through the linac upstream. The various halo × sources include the bunch compressor, wakefield generated can withstand the power load for 6 successive incidents, the tails in the linac, mis-matched and mis-steered beam, and vacuum chamber wall will be destroyed within 2 ms. The particles produced by Coulomb Scattering. In the longitu- goal of the machine protection system will be to limit the 5 dinal plane off-energy particles are dark current electrons relative beam loss to 10− of the maximum beam current. and energy tails produced in the bunch compressors. Dark This can be achieved by comparing the signals of a series of current is mainly created in the gun and through field emis- fast current monitors that allow one to measure the bunch sion in the super-conducting cavities [2]. Another impor- charge at a precision of better than 0.2 pC. Together with tant role for the collimator is set during the commissioning two other systems based on NaI beam loss monitors and of the machine as linac and collimator tune-up dumps are optical fibers this safety chain of triple redundancy will be part of the system. able to enable machine operation in case of sudden beam losses. Undulator Protection The BESSY FEL facility [3] will provide light from SYSTEM OVERVIEW three different lines running at different energies. Each In Fig. 1 the general layout of the collimator section FEL line will have its own collimator section directly in is depicted. After the last accelerating module the beam front of the undulator modules. The collimator is designed enters first the transverse collimation section containing with the high energy beamline running at 2.3 GeV elec- a matching quadrupole triplet and the two transverse col- tron beam energy in mind but is without any modifications limator apertures. From there the beam passes another adaptable for all other FEL lines. The modulators and ra- matching section before going through a dogleg structure diators of the FEL are realized as planar pure permanent made up with two dipole magnets separated by a drift re- gion with a quadrupole triplet. The two energy collimators ∗ Funded by the Bundesministerium fur¨ Bildung und Forschung, the state of Berlin and the Zukunftsfonds Berlin are located before the first and after the third quadrupole. Available online at http://www.JACoW.org FEL Technology 382 T. Kamps / Proceedings of the 2004 FEL Conference, 381-384 TUNE−UP DUMP COLL ENERGY COLLIMATION TRANSVERSE LINAC MODULE COLLIMATION FEL MATCHING APERTURE TUNE−UP SEXT BLOCK BEND DUMP LINAC QUAD Figure 1: Collimator beamline including the last accelerating module and the beginning of the FEL undulator section. An additional set of sextupoles is placed close to these blocks separated by π/2 phase advance. Each block is fit- quadrupoles. In order to minimize the flux of photons ted with a small round aperture limiting the passage for escaping the transverse collimation section the dogleg is the horizontal and vertical direction at the same position. placed between undulator and transverse collimation sec- This scheme requires that the twiss parameters for both di- tion. The collimator beamline ends with the last matching mensions are the same as well as the phase advance. This section preparing the twiss parameters according the spec- is achieved by producing a waist with the beam coming ification of the undulator lattice. The section contains two from the linac by means of a quadrupole triplet as shown in tune-up dump beam exits. One directly after the linac, the Fig. 2. The beam enters the transverse collimator from left other just before the entrance of the undulator. The whole coming out of the last accelerating section. The quadrupole section is less than 43 m long. triplet focuses the beam in both dimension to waist, which is located at the center between the two apertures CT1 and Transverse Collimation CT2. The horizontal phase space portrait at the center be- tween CT1 and CT2 is illustrated in Fig. 3. The inner For transverse collimation many configurations with aperture slots or holes are possible [5]. The most common 0.4 UNDULATOR solution is to place a thin ( 1/2 radiation length) spoiler 0.3 CHAMBER with a small aperture in front∼ of a long ( 20 radiation ∼ 0.2 CT1 length) absorber block with a slightly larger aperture. The CT2 spoiler produces a large transverse momentum spread in 0.1 σ the intercepted particles. The increased transverse momen- 0 x’ [mrad] σ tum spread is converted into a large radial spread which −0.1 10 is completely attenuated by the following absorber block. Two or more pairs are located in an optics channel with ad- −0.2 equate twiss functions, which could be a FODO lattice with −0.3 π/2 phase advance between two pairs. −0.4 −4 −3 −2 −1 0 1 2 3 4 x [mm] Figure 3: Phase space portrait at the center between CT1 β x π /2 and CT2. Also shown is the inner boundary of the back- PHASE ADVANCE transformed undulator vacuum chamber. βy boundary of the undulator vacuum chamber is completely BETAFUNCTION shaded by the two apertures. The area inside the parallel- CT1 CT2 ogram is the upper limit for the collimator acceptance and s is given by the beam stay clear distance ∆xclear and the CENTER maximum beta function in the undulator βmax according to Figure 2: Transverse collimation scheme with two aper- 2 (rpipe ∆xclear) tures CT1 and CT2. acoll − (1) ≤ 2βmax In order to keep the BESSY FEL collimator system as and for the given case the collimator acceptance is acoll = compact as possible and the total length under 50 m the 0.3575 µm. From this the aperture radius can be calculated scheme with two spoiler-absorber pairs was dropped in fa- with g = √a L where the transverse collimator coll coll · coll vor of a setup with two single absorbers. Recent studies for length Lcoll is the distance between the two apertures CT1 the TTF2 collimator [6] show that with this scheme a high and CT2 which is equal to the value for the betafunction collimation efficiency can be achieved with just two aper- at these location. The collimator aperture radius is then tures working in parallel for both transverse dimensions. 2.6 mm. In the following copper is assumed for the ab- The transverse collimation section contains two absorber sorber material. More studies are underway to compare the TUPOS02 T. Kamps / Proceedings of the 2004 FEL Conference, 381-384 383 performance for copper and titanium. The length of the tures. absorber is around 10 radiation length and thus 140 mm. 1 2 1 2 CT CT In order to minimize wakefield effects and to keep the to- CE CE ACC FEL tal block length within limits a step-in plus taper combi- nation was chosen for the transition between gap radius to beampipe radius.