Calculating the Loss Factor of the Lcls Beam Line Elements for Ultra Short Bunches*

The Linac Coherent Light Source (LCLS) is a SASE 1.5-15 Å x-ray Free-Electron Laser (FEL) facility [1]. Since an ultra-short intense bunch is used in the LCLS operation one might suggest that wake fields, generated in the vacuum chamber, may have an effect on the x-ray production because these fields can change the beam particle energies thereby increasing the energy spread in a bunch. At LCLS a feedback system precisely controls the bunch energy before it enters a beam transport line after the linac. However, in the transport line and later in the undulator section the bunch energy and energy spread are not under feedback control and may change due to wake field radiation, which depends upon the bunch current or on a bunch length. The linear part of the energy spread can be compensated in the upstream linac; the energy loss in the undulator section can be compensated by varying the K-parameter of the undulators, however we need a precise knowledge of the wake fields in this part of the machine. Resistive wake fields are known and well calculated [2]. We discuss an additional part of the wake fields, which comes from the different vacuum elements like bellows, BPMs, transitions, vacuum ports, vacuum valves and others. We use the code “NOVO” together with analytical estimations for the wake potential calculations. INTRODUCTION An ultra-short bunch is used in the LCLS operation. The bunch is prepared, compressed and accelerated in the main part of the machine, which includes an injector, two bunch compressors and a linac (Fig. 1). Many feedback loops are used in the LCLS operation. The most important loop uses 6 RF phases and amplitudes to control the energy at the four bend systems and the bunch length after each compressor. The bunch length measurement is based on coherent edge radiation from the last dipole of each bunch compressor chicane and each is calibrated in amperes of peak current with the transverse RF cavities. The peak current is stabilized to about 10% rms [1]. The bunch charge is normally held at 0.25 nC and peak bunch current is about 3000 A. A feedback system precisely controls the final bunch energy (better than 0.04%) up to the region DL2, where the bunch is translated in horizontal direction (Fig.1). DL2 is situated at the beginning of the 300 m long beam transport beam line. This beam line is used to transport the electron beam from the end of the SLAC linac to the FEL undulator (LTU beam line). In this line and later in the undulator section (130 m long) the bunch may lose energy due to wake field radiation. Additionally, wake fields may add a significant energy spread to the bunch. If we know this additional energy spread we can compensate linear part of it by changing the phase and the amplitude of an accelerator field in the last linac sections, introducing an opposite energy spread. For better performance the energy loss in the undulator section has to be compensated by varying the K-parameter of the undulators. In LCLS the undulator magnet gap has a small cant angle, which allows remote tuning of K by ±0.6% by small horizontal translations of the undulator magnets [1]. For these reasons we need a precise knowledge of the wake fields in this part of the machine. Additionally we need to know the wake field losses in the 50 m long beam line from the end of the undulator section to the beam dump for a better comparison with the loss factor measurements. These measurements were done using BPMs near the dump. BEAM LINE ELEMENTS The total length of the beam line from the BPM located between the two pairs of magnets in DL2 to the BPMs and YAG screen before the beam dump is approximately 400 m. The LTU beam line is approximately 225 m long and composed of two horizontal dipole magnets (the second pair of magnets of DL2), and a series of quadrupole magnets and correctors. The LTU vacuum chamber contains BPMs, OTRs, and collimators, four types of bellows, vacuum valves, vacuum ports and different kind of transitions. Fig. 2 shows a photograph of a corrector, a bellows and a chamber step. These elements are located after the DL2. Fig. 3 shows photographs of LTU strip-line BPM and a vacuum cross. The main part of the LTU stainless steel vacuum pipe is copper plated. This part is approximately 200 m long. The remaining 25m are a collection of small sections of stainless steel pipe of different cross sections. The undulator section is 130 m long and contains another type of bellows and BPMs. Fig. 4 shows a photograph of an undulator shielded bellows together with an undulator RF BPM. The vacuum chamber within each undulator segment is highly polished aluminium Figure 1: LCLS layout. ___________________________ *Work supported by Department of Energy contract DE-AC02-76SF00515 #novo@slac.stanford.edu (<0.2-μm surface finish) with a 5-mm height and 11-mm width. Figure 2: LTU beam line: corrector, bellows, quad, chamber step and bellows. Figure 3: LTU strip-line BPM. “Lesker” bellows and a vacuum cross. The dump beam pipe is approximately 50m long and has a stainless steel pipe of larger size. Figure 4: Undulator bellows and RF BPM Not all vacuum elements produce significant wake fields; however we tried to include each element in our wake field simulation. The beam line elements are specified in Table 1. WAKE FIELD SIMULATIONS We use the code “NOVO” [4] for the wake potential calculations. This code can simulate wake fields of very short bunches. We calculate the wake potentials of bellows and BPMs for a bunch length of 2 micron. Using additional analytical estimations [5] we can construct the wake field Green’s function. Wake fields from the resistive wall pipe with an oxide dielectric layer were calculated using a direct solution of the Maxwell’s equations [6]. Surface roughness wake fields were calculated using synchronous mode model [7]. Table 1: Beam pipe elements from DL2 to dump Element Quantity Description LTU pumping cross 30 ø34.8 mm