BESSY VSR: Enabling the study of the dynamics of matter
To fully characterize materials and their function, and perhaps even influence their function, modern synchrotron light sources must enable the determination of matter’s structure and its dynamics. The latter requires the creation of very short photon pulses and hence, in storage rings, electron bunches. But in standard x-ray and UV storage rings the natural pulse length is limited to some 10 ps.
To circumvent this limit, BESSY II generally operates a few days of the year in a so-called low-alpha mode where the machine optics are modified to provide a low momentum compaction and as a result bunch lengths of the order 1 ps. However, this bunch shortening comes at a price: The total current that can be stored in the machine is reduced significantly due to beam instabilities. While the low-alpha mode enables studies of material dynamics, the photon intensity is low. Other, “photon-hungry” users, who may not be interested in dynamics experiments are literally left in the dark. This is the reason why the low-alpha mode is offered on a very limited basis.
Rather than live with this limitation, HZB is developing a completely new scheme called BESSY VSR which is able to provide short (1 ps) and long pulses to (17 ps) simultaneously at high current so that users can pick those pulses which best suit their experiments. In a nutshell, short bunches are generated by installing new, CW high-voltage superconducting RF cavities that operate at a harmonic of the 500-MHz accelerating system. The combination of the high voltage and the higher frequency means that the bunches see a large voltage gradient when they pass these cavities, which compresses their length down to the low ps range. At the same time the high gradient also suppresses the beam instabilities that are encountered during normal low-alpha operation.
However, if all bunches in the ring were only 1 ps long, then their interaction with accelerator components via wake fields would result in a prohibitive amount of heating. Fortunately, users interested in dynamic experiments only require very few bunches in the ring’s fill pattern to be so short (for repetition rates of order MHz). If one can find a scheme that only shortens a few of the bunches, while leaving the others long, one could thus satisfy all users: Those who require short bunches and those who require a high photon flux but do not care about the bunch length.
BESSY VSR does exactly this. The trick is to use two SRF cavity units, one operating at the 3rd harmonic of 500 MHz (1.5 GHz) and the other operating at the 3½ harmonic (1.75 GHz). These two frequencies cause a beating of the total voltage. By adjusting the individual voltages and their phases precisely, the beating cancels the voltage at the position of every second bunch (thus generating long bunches as dictated by the 500 MHz system). At the other bunches the voltages add and short bunches are generated. Not all of these buckets must be populated, but the current in the long bunches can be increased to always provide the maximum possible average current (300 mA for BESSY II) and hence average photon flux. More details on the scheme are provide here and 
For BESSY VSR CW SRF systems at two frequencies will be installed. The beating between the two creates RF buckets for short and long pulses.
The SRF system is one very challenging aspect of BESSY VSR, pushing SRF to its limits. Operating in the L-band, these systems must able to handle at least 300 mA beam current and operate at the maximum possible voltage (> 15 MV/m) to limit the valuable space in the accelerator required for the SRF system. A completely new cavity design must be developed. The excitation of higher-order modes (HOMs) and their extraction from the cavities, the RF field stability, as well as the RF losses in the cavities thus move into the foreground of the technology development. All these aspects are being studied by G-ISRF in the ARD Topic of the Matter and Technology Helmholtz Program. Fortunately, these stringent cavity requirements are shared by the BERLinPro project so that much of the development can go hand in hand. The focus of the current activities is to
- Develop cell-shape and cavity designs with low loss factors and as few trapped modes as possible, optimizing the number of cells. These designs must also enable high-field operation and careful attention must be paid to the peak surface electric and magnetic fields.
- Develop waveguide damping schemes to reduce the HOM quality factors below instability thresholds. The impedance spectrum is used by G-IA to perform beam stability analyses.
- Develop schemes to provide sufficient RF field stability for stable beam operation. The microphonics spectrum is also used by G-IA for beam stability analyses.
- Construct prototype systems for testing on vertical test stands and in HoBiCaT.
 G. Wüstefeld, A. Jankowiak, J. Knobloch, M. Ries, Simultaneous Long and Short Electron Bunches in the BESSY II Storage Ring, Proc. IPAC 2011, San Sebastian, Spain, 2011