Diagnostics for Damping Rings 2000:
Transverse and longitudinal beam size monitors.
Theo Kotseroglou, Patrick Krejcik
We have identified issues associated with the measurement and monitoring of the electron beam transverse and longitudinal size in the current SLC damping rings. Specifically, there is data that show that the transverse beam size is not gaussian in the first hundred turns after injection into the Damping rings . Fourier analysis of this data exhibited peaks at the betatron frequency and its second harmonic indicating the presence and size of the dispersion and b -mismatches. Observation of higher betatron harmonics could also point in nonlinearities in the ring lattice.
We propose two diagnostics that will aid in studying this effect utilizing the existing synchrotron light optical transport: the first is based on a fast commercial CCD and the second on fiber-optically coupling the synchrotron light into fast photodiodes which are then sampled by fast electronics. We also suggest a technique based on electro-optically scanning a laser onto the electron beam in order to measure the emittance in a long store mode, where the synchrotron light measurement is diffraction limited.
Furthermore we suggest that we implement the interferometric technique used successfully at the damping ring at ATF, KEK in order to complement and check the gated camera results.
Measuring the bunch length of the electron beam at the sub-picosecond level is still a difficult problem and a streak camera is still the only proved diagnostic. In order to use it for turn by turn measurements a synchro-lock and a dual axis scan unit are needed, which make he streak camera very expensive . We suggest here two bunch length monitors: the first utilizes a CW laser and the RF pickup in an electro-optic crystal and a Fabry Perot spectrometer to diagnose the polarization modulation induced on the laser beam and is on an R&D stage. The second is optical gating of the synchrotron light through a Kerr medium using a femtosecond laser and has never been tried before with synchrotron radiation.
2. Fast, turn by turn profile monitors
Currently the measurement of the electron beam size and subsequently of the damping time and emittance is done by imaging the synchrotron light into a gated camera . The state of the art gated cameras can resolve images within a gate of 2 ns which is an issue for the ILC damping rings since the bunch spacing is 1.4 or 2.8 ns. Furthermore, the data can be transferred no faster than few milliseconds from the camera to the data acquisition system, which means that the spot size of the electron beam can not be studied within one storage cycle and that is an issue for any damping ring since one would like to study the beam in the first few injection cycles.
We are proposing two methods that will allow us to disentangle the above issues:
2.1 A turn by turn, single bunch profile monitor.
The system is shown in Fig. 1. It is an extension of the current measurement methods and its basis are a stage of bunch selection from a pulse train turn by turn and a new state-of- the-art CCD camera with a speed adequate for measuring turn by turn for 16 consecutive turns.
Fig. 1. Synchrotron light imaging system for turn by turn single electron bunch monitoring.
The synchrotron light is filtered to allow green light with a bandwidth of a few nm. Then it is imaged through the pulse selection which consists of a 1:100000 extinction ratio polarizer and analyzer coated for green and a Pockels cell stage with a 1 ns rise and fall time with a resonant circuit at 2.5 MHz . The optical modulator Model no. 4104 from New Focus is an example of such device.
The light is then imaged onto a multi-channel plate (MCP) in order to increase the light intensity from the single bunch. The MCP is operated with a long enough gate ( microseconds ) since the gating of ~2 ns width is already done by the Pockels cell. The bias to the MCP is ramped appropriately in order to account for the varying light intensity associated with the damping of the electron beam size.
The backbone of the profile monitor is the CCD camera model SMD-64K1M from Silicon Mountain Design. It is a 256 x 256 pixel array with each "pixel" occupying a 56m m x 56m m area. From this area only the 14m m x14m m is active, i.e. there is a large gap ( considering the usual CCD camera design) in between the real pixels. This area, in simple words, consists of 16 capacitors. So the image data from each pixel are stored parallel onto the space next to them allowing the camera to collect 16 frames at a speed of 2.5 MHz, i.e. a modified model SMD-64K4M. Then the unloading of the 16 frames takes 30 ms onto a DAQ system using frame grabbing software ( that we expect to modify in order to be compatible with appropriate Damping Ring control system).
The cost of the system is dominated by the following items: CCD $50k, optical modulator $7k, polarizers $5k, MCP $2k, imaging system $2k.
The relevant issues of this diagnostic are:
The quantum efficiency (QE) of the CCD is 17% @ 750 nm. The read noise is 55 e- and the fill factor 1/26. A 256x256 monitor is expected in June ı98.
The design of the imaging system is interesting since has to be insensitive to Pockels cells small aperture ( ~2 mm). Also issues regarding imaging the synchrotron light through such crystal should be studied.
The aliasing due to the pixel vs. dead area size is not of concern even if the pulse is not exactly gaussian, since we believe that there are no high frequency components in the synchrotron light or that the smaller necessary image we will acquire is larger than a few pixels.
Future designs of the CCD camera can be improved further in the context of SBIR projects ( the camera itself in its present state is the product of an SBIR contract that won a 96 award).
2.2 A turn by turn multiple bunch monitor.
Figure 2. A fiber-optically coupled turn by turn and bunch by bunch synchrotron light monitor
This system is geared more towards an ILC damping ring with many bunches in one bunch train, separated by 1.4 ns. It is more complicated than the CCD based version but can provide us with measurements for many bunches at the same time within the same or multiple bunch trains. The system is shown in Fig. 2.
The system is an improved version of an idea to monitor beam beam interactions . It is based on a fiber-optic bundle that acts as the front end of a similar to a CCD light collection system and fast photodiodes, whose signal is sampled with a fast sample and hold clocked by the DR RF. By using a trigger input one can sample any of the bunch trains or a few bunches within one bunch train, depending on the S/H speed of the DAQ system. The signal is then directed into the digitizer and the computer.
Although the fiber wavelength can be chosen to be as short as possible in order to improve the diffraction limit of the apparatus, the photodiodes response falls off drastically at approximately 900 nm ( for an InGaAs type diode ).
The number of synchrotron radiation photons per electron bunch within a 100 nm bandwidth ( in order for the diode to have a flat response ) around a wavelength of 1 m m for the SLC ring is approximately N = 125 photons and the fluctuations are large . These are not very many! Currently averaging of 8 pulses is done in order to improve the statistics and fit the data with the emittance damping effect, but we believe that the injection data in  were taken using a single bunch and no averaging. We possibly need to improve the existing intensifier too.
A rough cost estimate of the diagnostic follows:
The fiber-optic unit with connectors is custom made from Dolan Jenner and will be $5k
The diodes without pigtail cost $500 each. The electronics are fairly cheap as long as we do not use faster chips than the DATEL S/H
The digitizer is still under investigation, although most of the electronics are duplicate or along the lines of BPM signal DAQ for the ILC design and can be reproduced with very small cost here .
We propose to do this test in two stages:
First instead of diodes we plan to use a multi-anode PMT from Hamamatsu. Models with 64 and 100 anodes in a grid of 8x8 and 10x10 are available. The rise time of this PMT is measured 1.2 ns and is adequate for the SLC DR. In this stage of the experiment we will have tested also the DAQ system. Then we propose to substitute the PMT with fast diodes and in that way built an ILC system.
3. Beam Size Monitors for the low emittance ( damped ) beam
3.1 Interferometric technique for spot size measurements
This has been applied by Mitsuhashi et al.  in the ATF ring with results better than the gated camera due to diffraction limit of the latter method.
It seems an easy enough diagnostic to be incorporated in the monitors for the improved SLC DR.
Laser-wire pulsed monitor for long store mode without scanning of the electron beam
This is more difficult than the above idea with potential to measure much smaller emittances in the ring but may also be unnecessary for the SLC ring since the beam can be measured at the extraction line with wire scanners. Nevertheless we mention this here since some installation in the DR will be needed if such measurement proved to be interesting. It is, again, geared more towards the ATF DR.
The difference of this laser-wire module and of the existing ones is that the laser beam is scanned onto the electron beam. This is especially attractive for the damping ring due to the constrains on the DR motion. This also makes the measurement limited only by the laserıs repetition rate and/or the repetition rate of the beam deflector module. The laser rep. Rate can be up to 100 MHz if an oscillator is used , with of course the appropriate energy and pulse width.
The experimental - 6 - setup is shown in Fig. 3.
Figure 3. A laser-wire monitor based on an electro-optic beam deflector beam that scans the laser onto the electron beam.
The calculations of Compton scattered photons from the interaction allow for a choice of mainly 3 high rep rate lasers: An oscillator from Quantronix with 300 microJ per pulse and rep rate of 100 MHz which though is fairly expensive ( $100K) and a Q-switched laser from Spectra Physics in the green with rep rate 50 kHz and 750 microJ per pulse. ( $60K)
The electro-optic beam deflector is a Conoptics module that also requires a fast driver( $10K).
The Interaction point optics can be much simpler than the existing SLC laser-wire.
4. Bunch length measurements
Apart from the well known streak camera measurements there is still not a reliable method for measuring sub-picosecond electron beams ( or rather in the DR case, 10 picosecond beam but with, possibly, a head to tail variation at t he sub-picosecond range). Here we present two promising ideas for such measurement. The first technique is based on conversion of the beam diffracted field in optical modulation and has not been tried before and the second is a cross-correlation technique used recently in measuring sub-picosecond laser beams.
4.1 Electro-optic measurement
A new monitor for electron beam bunch-lengths at the ps level was suggested by . The pickup from the electron beam modulates the birefringence of an electro-optic crystal that is located close the beam line. A CW laser is then used to detect this modulation by setting the electro-optic crystal between two crossed polarizers. Since the electron beam bunch-length is of the 10 ps level, a fast diode and a sampling scope can measure this modulation. Up to this point there is not much interest since the diode can be instead installed on the synchrotron light on the optical table upstairs. But it is interesting to detect fast longitudinal instabilities and possibly this can be done by using a spectrometer in the frequency domain as explained in . This technique will be tested at LLNL in an electron injector within the next month and is of interest for the LCLS bunch length diagnostics.
The bunch length measurement equipment will be positioned close to a 1" ceramic gap of the electron beam pipe. The test stand is shown in Fig. 4.
Figure 4. Experimental Setup
Preliminary radiation damage tests of a KD*P crystal proved promising. No optical degradation of the crystal was observed. Some relevant experiments on radiation on non-linear crystals have observed the decrease of the nonlinear conversion efficiency of such crystals at the few percent level and its recovery within 1 h of the irradiation termination . We speculate that the radiation level of this beamline is lower than the sources used at these tests. The setup also employs fibers that are prone to radiation damage unless they are chosen to be quartz. We plan to do more damage tests on these components. Probably shielding in the DR will be required.
The RF field power emitted from the electron beam from the dielectric gap, if we assume that this is not metal coated , was found to be adequate to create a birefringence rotation of a few degrees in the electro-optic crystal .
The laser pulse will be modulated at the 300 GHz level for 1 ps electron pulse, so the linear dispersion in the return optical fiber matters. This was found to be of the order of 1 ps  which is adequate for the 10 ps electron beam of the DR and head to tail electron beam intensity distribution that varies at the ps level.
The electro-optic technique will be tested in the LLNL electron injector in the near future. The experiment will be performed in two stages:
In the first stage we will only use the fast photodiode as a detector of long electron beam pulse widths ( of the order of 15 30 ps ) and we plan to make a first measurement of the beam bunch length vs. the compressor and also optimize the injector performance.
In the second stage we will use the spectrometer as discussed in  and in that way commission a bunch length diagnostic capable of measuring LCLS type beam sizes .
A cost estimate of the components of the first stage experiment is presented in table 1.
Table 1. Component description and cost estimate for stage1.
|Item Description||Vendor||Part No.||Num.||Price||Total|
|Fiber Optic U-bracket||Princeton Optics 609-771-4370||50-820-2||1||$600||$600|
|Polarization maintaining fiber @820 nm||Thorlabs
|FS-PM-4611||2 ea 10 m||$8.40/m||$168|
|FC Fiber connector
Locking of fibers
|Princeton Optics 609-771-4370||Service||2 ea||$70||$140|
|LiNBO3 electro-optic crystal||New Focus||2x2x20 mm||On Loan|
With FC-PC chuck
20 GHz with pigtail FC-PC
( SLAC )
( LLNL or SLAC)
4.2 Optical Gating
Auto-correlation techniques have been performed in the IR and 532 nm regime on short pulse lasers for many years and recently can be done also in the UV and also in the Fs regime. Most of the techniques need a high intensity laser beam to be monitored but using a new idea  one could envision performing optical gating of the synchrotron light through a Kerr medium with a pump beam provided by a fs laser.
The main limitation would be the phase locking of the laser to the electron beam at the subpicosecond level.
The technique will be studied as part of the LCLS laser system, where instead of the synchrotron beam a low intensity laser beam will be detected.
1. M.Minty et al., Using a fast-gated camera for measurements of transverse beam distributions and damping times, SLAC-PUB-5993
2. A. Fisher, Private communication.
3. T. Chen et. al. An apparatus for measuring turn by turn transverse beam profile in electron storage rings, submitted to EPACı96.
4. B. Siemann, Intro to accelerator physics, SLAC SLUO lecture #4, Feb. 20 , 1998.
5. S. Smith, private communication.
6. T. Mitsuhashi, Spatial Coherency of te Synchrotron Radiation at the visible Light Region and its Application for the Electron Beam Profile Measurement., KEK-PREPRINT-97-56
Using Optical techniques", unpublished.