The Accelerator Side of Interaction Regions

In the first part of the talk the principles of low-beta insertions were reviewed. The first low-beta insertion was proposed by Voss and Robinson in 1965. The simplest configuration uses quadrupole doublets on each side of the interaction point (IP). The typical cross section reduction factor of the beams is in the range of 50 to 150 for modern ring colliders.

Next the layout of the interaction regions (IR) of PEP-II, KEK-B and HERA was compared. All three machines are asymmetric in energy with an asymmetry factor of roughly 3 for the B-factories and more than 30 for HERA. One of the important differences between the machines is the way they deal with the synchrotron radiation (SR) background. KEK-B has a crossing angle of 11mrad at the IP. This allows to separate the beams without deflecting the incoming beam close to the IP. Consequently only small amounts of SR are produced and collimation is relatively easy. The other two machines have the concept to pass the SR through the detector without loosing too much power in the detector. The typical power levels that have to pass the detector are several 10\,kW, however, acceptable levels on the detector beam pipe are less than a mW.

One essential ingredient for a modern IR are specialized magnets. Often magnets are needed that have to accommodate tight space requirements inside a detector, have a large aperture to avoid background from SR tails, o  have a special geometry to allow for close passage of a second beam or SR fan. As an example the superconducting HERA magnets were shown which provide a large aperture with a small outer diameter at the same time. Another example were the HERA septum quadrupoles with a triangular cut-out in the iron mirror plate that allows a very close orbit for the electron beam without disturbing it too much. Further points of discussion were field errors of the IR magnets and alignment sensitivity. For the new HERA IR a stretched wire alignment system has been installed that allows for transverse position and roll angle measurements of all IR quadrupoles.

The next topic discussed was the control of detector backgrounds which are distinguished into SR and particle background. The particle background is usually controlled by momentum collimation in a dispersive section upstream of the IP. Simulation results were shown for HERA. Beside that one has to provide good vacuum conditions to reduce the rate of beam-gas scattering. The common problem in all modern layouts is that the beam separation magnets are installed close to the IP and they tend to sweep low energy particles from the beam directly into the detector. The other important background source is SR. Numerical calculations that allow for precise predictions of the SR levels at certain locations in the lattice were presented. A semi-analytic method is used that computes analytical radiation distributions for thin slices of combined function magnets and sums them up on a plane of interest, for example inside the detector. One problem are non-Gaussian tails in the beam distribution that affect the width of the radiation fan. It was shown how such effects can be taken into account if the beam distribution has been measured by scraper experiments. Another concern for the location of the radiation fan is its sensitivity to orbit errors.

Finally some specific aspects of the vacuum system were discussed. Often very specific vacuum chambers are needed in order to adopt to uncommon magnet designs. Often excellent vacuum is required which can be achieved with parallel NEG pumps. Another topic are higher order mode losses (HOM's) in certain chamber geometries. The HOM losses in high Q arrangements can easily reach several kW's of deposited power, which is usually by far not acceptable. One has to avoid cavity like vacuum chambers and all critical chambers have to be checked with field computation codes like MAFIA. Another subject are SR absorbers which are critical in view of heat load and backscattered radiation that causes background in the detector.