Vacuum System Performance and Specifications

A good vacuum system is an essential requirement of any SR facility in order to guarantee adequate beam lifetime, however the cost is also an important factor. Three possible ways for reducing vacuum system costs were discussed: no pre-baking, no in-situ bakeout and reducing the number of pumps. Recent experiments now confirm previous observations that there is no difference in desorption yield for pre-baked and not pre-baked stainless steel. In-situ bakeout reduces pressure rapidly and gives a "clean" vacuum, but at the expense of thermal stress, possible leaks, and non-negligible cost and effort. Experience at ELETTRA suggests that in-situ bakeout is not necessary, and preliminary conclusions from experiments at DCI support this. The elimination of bake-out would also ease the construction of the vacuum chamber, for example bellow sections could be reduced in number or shortened. However, long ID chambers and r.f systems may still have to be baked-out to guarantee performance. Also, it was pointed out that the chamber geometry can strongly influence the desorption rate because of restrictions to the photo-electron paths. Finally, experience at ELETTRA and DESY suggests that many fewer pumps are needed to maintain UHV conditions than are often assumed during the design stage.

Storage Ring FEL s

Six storage ring FELs have so far been operated, however the lowest wavelength reached is still 240 nm obtained at VEPP3 in 1988. Progress in reducing the wavelength has not therefore been very rapid, one of the major problems being tackled is damage to the cavity mirrors. Nevertheless, it is generally believed that storage ring FELs will make advances in the near future, particularly with the advent of new dedicated rings such as DELTA and Duke University - the first lasing of the latter at 350 nm was announced at the Workshop. It was speculated that the 150-200 nm region should soon be a accessible using normal incidence optics, 30 nm using ring cavities, and eventually possibly wavelengths as low as 4 nm using harmonics etc.. In the present context the main issues are the compatibility of the ring design and FEL operation with the needs of a high brightness synchrotron radiation facility. For this reason the SASE mode, requiring generally a very long undulator, possibly in a pulsed by-pass, is not very attractive compared to an oscillator. SuperACO can support FEL and SR operation at the same time at 800 MeV, however higher energy rings involve some compromise, for example SOLEIL will operate at 1.5 GeV for the FEL, however the ring is optimized at 2.5 GeV for SR users. An alternative possibility in the case of the SLS project is the use of the extracted low emittance beam from the booster.

prepared by C.J. Bocchetta and R.P. Walker