The FLASH free electron laser facility at DESY in Hamburg  generates short and very intense light pulses in the Vacuum Ultra Violet (VUV) region of the electromagnetic spectrum. The elementary physical process behind the generation of these pulses is that of Self Amplified Stimulated Emission (SASE) in the large undulator section (ca. 30 m, 1000 periods) of the FLASH accelereator.
The SASE process  depends critically on the parameters of the electron bunches travelling down the accelerator, such as the bunch length, the slice emittance and the slice energy spread. For a proper operation of the machine it is therefore of utmost importance that these parameters can be controlled in detail. A first step in the control is, of course, an accurate measurement of the properties of the electron bunches.
The electron bunches are small, ca. 0.2 mm in diameter and a few tens of microns long, and they travel at nearly the speed of light. Consequently they are rather elusive and their properties are not easily determined. Several ingenious methods had to be devised to accomplish this task. Among them are the transverse deflecting cavity , a kind of streak camera) and the use of electro-optically active crystals . Both have been implemented at FLASH and, as is so often the case, each method has its own advantages and disadvantages.
The Optical Replica Synthesiser (ORS) is an additional method for determining the properties of the electron bunches. The principle behind the ORS is quite straightforward. We have seen that a direct measurement of the properties of the electron bunches is difficult. On the other hand, in recent years several techniques have been developed to fully characterise short (femtosecond or picosecond) light pulses. If one would somehow succeed to code the properties of the electron bunch in a light pulse, in other words make an optical replica, one would have made a big step forward in characterising the electron bunch. This is precisely what the ORS is designed to do. The original idea of the ORS was put forward in a recent paper by E.L. Saldin, E.A. Schneidmiller and M.V. Yurkov in Nuclear Instruments and Methods in Physics Research (Nucl. Instr. Meth. A 539, 499-526, 2005).
The ORS uses a seed laser, two short undulators: a modulator and a radiator, a chicane and a device to characterise the optical replica using Frequency Resolved Optical Gating (FROG) . The undulators and the chicane are installed between the dog-leg, where there is a vacuum window to let in the seed laser beam, and the main VUV-FEL undulator of FLASH, i.e. between roughly 166 m and 182 m in the FLASH co-ordinate system where the electron injector (upstream) is at 0 m and the electron beam dump (downstream) is at ca. 250 m. The seed laser is placed in a newly constructed laser building just besides the accelerator tunnel. The seed laser beam is transported to the vacuum window at the dog-leg via a laser beam transport line specifically build for that purpose.
In the first undulator, the modulator, the electron beam is overlapped with a seed laser pulse, both in space and in time. Due to the interaction of the (oscillating) electrons in the bunch with the electromagnetic field of the laser pulse the energy of the electrons become modulated, i.e. some electrons in the bunch gain energy and others loose. The modulator is a short undulator with 7 periods and is in resonance with the wavelength of the seed laser of about 780 nm. The magnets are of electromagnetic design, the period is 200 mm, and the maximum field strength is about 0.5 T. We have modified a Clark-MXR CPA2001 Ti:sapphire laser to act as the seed laser. The modification was necessary in order to be able to synchronise the laser to the electron bunches. In order to achieve a sufficiently deep energy modulation the intensity of the seed laser beam has to be rather high, about 5 GW/cm2. The output of the CPA2001 of about 1 mJ /pulse is, with proper focussing, sufficient to achieve this value. The operating frequency will be between 1 and 10 Hz.
In the following chicane this modulation in energy space is converted into a density modulation in real space. To put it simply: electrons with different energies will take different paths through the arrangement of dipole magnets of the chicane which means that they will leave the chicane at different times, i.e. one has achieved a density modulation in ‘real’ space. Given the energy modulation achieved in the modulator, the depth of the density modulation can be controlled by changing the R56 (dispersion) value of the chicane. The analysis of Saldin et al. shows that the optimum R56 value is determined by the energy spread in the original electron bunch.
The final undulator, the radiator, is almost identical to the modulator. When the now ‘pre-bunched’ electron bunch enters the radiator the bunch will radiate strongly at the seed laser wavelength. Provided all parameters are properly set the radiator light will contain the desired information about the electron bunch. The intensity of the radiator light is much less than that of the original seed laser. We have chosen to separate them using their polarisation properties. To this end the radiator is rotated by 90 degrees with respect to the modulator and the radiator light is polarised orthogonal with respect to the seed laser beam. The seed laser beam can now be easily rejected by a high quality polariser.
The radiator light is finally injected into a ‘Grenouille’. The Grenouille is a commercial device  which analyses a short light pulse by means of second harmonic FROG . This measurement provides information about the amplitude and the phase of the radiator light. From these two parameters the desired properties of the electron bunch can be calculated.
As described above the ORS design is based on the interaction between the electromagnetic field of a laser beam and the oscillating electron bunch in an undulator. Here, this interaction is used to make a so-called optical copy of the electron bunch, but this interaction can also be used for other purposes. For example, the laser – electron bunch interaction can be applied to impose certain properties on the electron bunch which can lead to the production of attosecond VUV pulses .
The ORS experiment was completely setup at FLASH during the summer of 2007. Via gradual improvements of the setup and interesting results along the way , we could eventually verify that the ORS is experimentally realizable . Spatial and temporal overlap between the electron- and laser pulse was achieved and optical replica pulses were generated. The optical replica pulses were analyzed with the FROG technique to extract the bunch profile. Due to the shutdown of FLASH in September 2009, for the build up of the seeded FEL, sFLASH , further investigations are being postponed. At the restart of the experiments in 2010 a new high energy laser will be employed that will make measurements easier and e.g. enable characterization of longer electron bunches than has been possible so far.
 S.V. Milton et al., Science 292, 2037-2041, 2001 and J.W. Lewellen et al., Nucl. Instr. Meth. A 483, 40-45, 2002
 R. Trebino et al., Rev. Sci. Instr. 68, 3277-3295, 1997
 Swamp Optics, http://www.swampoptics.com/
 E. L. Saldin et al., DESY 04-045 and E. L. Saldin et al., preprint DESY 06-051
 G. Angelova et al., Observation of two-dimensional longitudinal-transverse correlations in an electron beam by laser-electron interactions, Physical Review Special Topics – Accelerators and Beams 11 (2008) 070702