Trapped Mode Study for a Rotatable Collimator Design for the Lhc Upgrade*

A rotatable collimator is proposed for the LHC phase II collimation upgrade. When the beam crosses the collimator, it will excite trapped modes that can contribute to the beam energy loss and power dissipation on the vacuum chamber wall. Transverse trapped modes can also generate transverse kicks on the beam and may thus affect the beam quality. In this paper, the parallel eigensolver code Omega3P is used to search for all the trapped modes below 2 GHz in two collimator designs, one with rectangular and the other with circular vacuum chamber. It is found that the longitudinal trapped modes in the circular vacuum chamber design may cause excessive heating. Adding ferrite tiles on the circular vacuum chamber wall can strongly damp these trapped modes. We will present and discuss the simulation results. INTRODUCTION SLAC proposed a rotatable jaw collimator design for the LHC phase II collimation upgrade through the US LHC accelerator research program (LARP) [1]. Figure 1 illustrates two collimator vacuum chamber designs. The rectangular vacuum chamber design was first proposed. The circular vacuum chamber design is currently considered for its easier fabrication and better vacuum pumping. Figure 1: 3D solid models of the SLAC rotatable collimators (Courtesy of Steven Lundgren). Left: Rectangular vacuum chamber design; Right: Circular vacuum chamber design. There are 20 facet faces on each cylindrical jaw surface and two jaws are rotatable and will move in and out with a 2mm to 42 mm gap during operation. When a beam happens to hit a jaw, the jaw can be rotated to introduce a clean surface for continued operation. The rotatable collimator is designed for 20 year life-time. Four thin flexible EM foils are used to connect the moving jaw ends to the rigid vacuum chamber. A beam can excite trapped modes in the vacuum chamber of the collimator. These trapped modes can cause beam energy loss and heating on the vacuum chamber wall as well as coupled bunch instabilities. In this paper, the parallel finite-element frequency domain code Omega3P is used to calculate the trapped modes below 2 GHz [2]. For modes above 2 GHz, their wakefield effects are negligible for the LHC long bunch length ( = 7.55 cm). Furthermore 42mm beampipe radius gives a cutoff frequency of 2.1GHz for the TE11 mode, and 2.7GHz for the TM01 mode. Therefore, the trapped modes below 2 GHz cannot propagate out of the two end beampipes.