Home
Electron bunch time pattern of FLASH with 5 Hz repetition rate and up to 800 bunches in an 800-μs-long bunch train. The separation of electron bunches within a train is 1 μs. To a certain extent, the bunch distance in a train can be varied; for instance to 2, 10, or 100 μs and some other distances. The duration of the electron bunches is 20 to 100 fs. The nonlinear FEL process reduces the duration of the photon pulses down to 10 - 50 fs.
Timing the Laser and the FEL
Two optical laser systems, both with 100-femtosecond pulses, are used. One mimics the bunch train structure of the FEL with 50 μJ per pulse and up to 800 pulses per train. The other laser has a pulse energy of 25 mJ with one pulse per bunch train, i.e. a repetition rate of 5 Hz. It is mandatory that the pump and probe pulses overlap each other in time and space. Accuracy of the order of the pulse length of both pulses is desired, and tremendous effort is made to enable stable synchronization and precise measurement of the remaining temporal jitter and drift.
The determination of the jitter between the optical laser and the FEL pulse is essential. Therefore part of the optical laser pulse is reflected to a streak camera recording the arrival time of the pulse. For the timing reference of the FEL pulse the optical portion of a synchrotron radiation pulse is used. This synchrotron radiation pulse is produced when the electrons are deflected in a dipole magnet into the dump. From the image of the streak camera, the relative jitter between both pulses is determined from the peak positions.
An even better temporal resolution on a shot-to-shot basis is achieved by a timing electro-optical sampling system which determines the jitter between the optical laser pulse and the electron bunch directly. In this system some of the optical laser pulses are sent into the accelerator tunnel guided by a glass fiber. Here, the laser pulses pass through an electro-optically active crystal located only a few millimeter away from the electron beam. The laser passes through the crystal with an angle of incidence of 45 degrees.
Image of the optical laser pulse and the synchrotron radiation pulse as a reference for the FEL arrival time on the streak camera. The shown time frame is about 140 ps. |
Scheme of the single-shot Electro-Optical (EO) sampling. Due to the 45-degree angle of incidence, the optical laser pulses cross different parts of the crystal at different times. If the laser pulse is later or earlier than the electron bunch, the image on the camera is shifted. |
Peak brilliance of FLASH and future FELs (XFEL at DESY, LCLS (USA)) compared with selected 3rd generation synchrotron radiation sources (PETRA III at DESY, Spring-8 (Japan), ESRF (France), APS (USA), SLS (Switzerland), and BESSY (Germany). DESY facilities are displayed in red, others in green. Blue dotes are used for measured values. – The European XFEL will be ready for commissioning in 2013; PETRA III will deliver photons for user operation starting in 2009.
When the electron bunch comes close to the crystal, its electric field induces a change of the birefringence inside the crystal, thus producing an electro-optical effect. If the laser pulse reaches the crystal at exactly the moment when the electron bunch passes the crystal, the polarization of the laser pulse will be turned slightly. Using a polarizer behind the crystal, this phase effect is transformed into an intensity signal, which is monitored by a CCD camera. Due to the mentioned angle of incidence, the optical laser pulses cross different parts of the crystal at different times; hence a mapping of time to space is achieved. If the laser pulse arrives later or earlier than the electron bunch, the image on the camera is shifted. The current time resolution is better than 100 femtoseconds.
The idea of this diagnostic tool is to record the jitter data and provide it to the experimentalists to sort their measured results after the experiment. When using data from timing experiments, the time resolution of pump-and-probe experiments is not determined by the timing jitter anymore, but rather by the accuracy of the jitter measurements. Thus the experiments can be improved by providing a temporal resolution beyond the “jitter limit”.
The next chapter on the first experimental results at FLASH presents an overview of the experiments that have been performed in 2005 and 2006 and describes in detail some of the most important results achieved so far.
