The NLC 2001 Beam Delivery Configuration

 Post Linac Collimation System

Peter Tenenbaum

2/15/2001.

Requirements:

The collimation system must satisfy the following requirements:

==> Collimate a non-Gaussian beam halo assumed to contain a maximum of 1e+09 particles per bunch train, removing all particles which fall outside a window defined by particle physics requirements (presently, 240 urad x 1000 urad x' vs y' at IP, set by quad SR requirements on vertex detector); said collimation may not produce a flux of secondaries at the detector which would be unacceptable (present limit believed to be N x 10 muons per bunch train per side, where N is of order unity); transmission of collimated halo particles must be on the order of 10 parts per million. ==> Provide protection of final focus and detector from bunch trains which would potentially cause damage ==> Resist serious damage to protection elements during one typical run-year ==> Permit operation over the full range of CM energies envisioned (presently 90 GeV to 1 TeV in main IR) ==> Minimize damage to quality of beams which are close enough to design conditions to produce luminosity (ie, a bunch train with a 1 sigma oscillation cannot be drastically degraded, but a train with a 3 sigma oscillation would not have produced any luminosity anyway so we can destroy it if necessary).

Technical description:

Present architecture is logically divided into an energy collimation region, a betatron collimation region, a post-collimation "dogleg," and matching between these regions. System length is approx. 1460 meters per side. All magnets are envisioned as electromagnet to ease energy adjustment.

Contains two dispersive regions (eta_max = 20 cm in each) in which off-energy particles are collimated. Presently we envision removing particles which are more than 1% from the nominal energy. The dispersive regions are separated by a -I transform in order to obtain certain useful symmetries between the dispersive regions. Since energy errors may occur on a single-train basis with no prior warning, vertical betatron function is very large (approx. 22 km for 1 TeV CM, can be reduced at lower energies) to ensure that beam spoilers are not damaged by a single bunch train. System contains 2 sets of horizontal spoilers (adjustable jaws) and 2 sets of horizontal absorbers (adjustable jaws); half-gaps are 1.2 mm for spoilers, 2.0 mm for absorbers. In addition to long weak bends and large-bore (2.54 cm half-aperture) quads, system contains several sextupoles (2.54 cm half-aperture), octupoles (1.5 cm half aperture), an energy-diagnostic section with high-performance BPM and wire scanner for linac tuneup, and a pair of pulsed bend magnets which can produce a 9 kG pole-tip field and pulse at 120 Hz for dumping bunch trains after the energy diagnostic section. Note that SR from DC bends and pulsed extraction bends is a serious issue for this region.

Contains a triplet lattice with modest betatron functions (270 m x 500 m); four sets of betatron spoilers (adjustable rectangular gap), four sets of absorbers (fixed round gap), and a large number of large-aperture protection collimators (fixed round gap). No exotic magnets or vacuum chambers (all quads have 1 cm half-aperture) are used here. Spoilers are assumed to be "consumable" wheel design presently in prototype; half-gaps are typically 150 um to 350 um at 500 GeV to 1 TeV CM; at lower energies, gaps are enlarged as 1/sqrt(E_CM). Note that these gaps correspond to approx. 4.8 sig_x and 17 sig_y at 500 GeV CM and below. Absorbers have 1 mm radius beam aperture, protection collimators are 6 mm radius. Absorbers and PCs are water-cooled.

Contains a combined-function (bend/quad) FODO lattice which has a large eta/beta ratio to permit pre-FF cleanup of low-energy beam particles which would otherwise escape collimation region. Magnet apertures are 1 cm radius. Collimators are not present in the deck yet. There are several sextupoles in the deck, which are at zero strength; don't know yet whether a use will be found for these.

Wakefields of tapered collimators of various materials (near-wall and near-center); single-pulse damage thresholds for spoilers made from various materials; long-term fatigue damage to collimators (mainly determines minimum half-gap attainable); production of muons and attenuation of primary particles; high-order optical effects in high-beta systems; tolerances in high-beta systems; synchrotron radiation; fast high-field magnets; high-radiation environments.

LCC-0052, which will be released before 2/27/01 FNAL meeting

SLAC-PUB-8511 (2000)

ecol_250gev.xsif, in NLC-2001 web space

Virtually every technical issue listed above contains significant unknowns in the context of the collimation system. Most notably:

==> We have assumed a halo population of approx. 1e9 per bunch train based on SLC experience. It could be much less (esp. since halo is collimated upstream of main linac), if it is much more the design will have problems with muons. ==> Halo transmission is 40 parts per million if betatron spoilers are 0.5 X0, goest to 10 ppm for 1.0 X0 but heating of spoilers might become an issue. ==> At present no model exists for the geometric wakefield of a moderately-tapered collimator, and resistive wake model remains to be fully tested. ==> Damage thresholds rely on several assumptions, and beam tests indicate thresholds may be higher than expected (good news). ==> System tolerances will be tight, though not as tight as in ZDR system or final focus. ==> Dynamic aperture of system may not be adequate, although full simulations not available at this time. ==> Gaps are relatively conservative based upon fatigue considerations.

Discussion of configuration choices:

The collimation system design is primarily driven by requirements of machine protection, and only secondarily by the expected beam halo or other issues. It is clear that large lattice functions (betatron and eta functions) improve the machine protection aspects of the design but make the nonlinear optics harder to manage. The present design is a compromise.

A further compromise is that the energy collimation is fully protected against a bunch train impact, while the betatron collimators are "consumable." This drove us to separate the energy and betatron systems as described. We believe that the betatron functions in the energy collimation section are a decent compromise between gap size (we want larger gaps, hence requiring larger beta functions) and nonlinear optics (we want weaker nonlinear optics, hence requiring smaller beta functions). The large betay in the energy region drove us to add sextupoles and octupoles to that part of the system. Also, the energy collimation achieves its protection through linear optics rather than a nonlinear magnet of some form. Attempts to use such a nonlinear magnet were not successful, though not in principle precluded. The single beam dumper pusled bends are horizontal bends; this allows the spoiler and absorber in the second energy collimation module to protect most downstream elements from a pulsed magnet failure.

A collimation system of "2 phases, 2 planes, 2 times" seems prudent and conservative, but does increase the length of the collimation system by about 200 meters and doubles the jitter amplification. Note that the second energy collimation dispersive region is there primarily to cancel the wakefield kick of the first one (the two systems have a +I in betatron phase and a -I in dispersion to achieve this). Removal of the second module would save 300 meters but would probably blow the horizontal jitter budget rather severely, since the wake cancellation would be lost. The symmetry also allows the use of sextupoles and octupoles, which would be precluded by removal of the second region.

Return to NLC 2001 Beam Delivery Configuration

Return to Beam Delivery Home page

Last modified 01/14/04
Tom Markiewicz