Conventional synchrotron light sources and Free-Electron Lasers (FELs) utilize permanent magnet undulators with periods on the order of a few centimeters, and to generate X-rays they need GeV scale electron beam energies. Such facilities are very large and expensive. Inverse Compton scattering sources use a laser beam as an undulator with micrometer periods and produce X-ray energies on the order of tens of keV. These sources operate with MeV scale beam energies, and therefore they are much more compact. However, their average photon flux is typically small, especially in the EUV and soft X-ray regime. We present a novel compact linac-driven light source, which could produce both incoherent and FEL radiation depending on its configuration. This source is based on a mm-period RF undulator. The RF undulator is a mm-wave cavity resonating at a deflecting mode. The source operates as follows: a train of electron bunches is generated in a thermionic X-band RF injector. These bunches are accelerated in an X-band linac and then interact with the RF undulator. The RF power that feeds the undulator is extracted from the electron beam in a decelerating RF structure, located downstream of the undulator. As an example, a light source with a 91.392 GHz RF undulator and a 129 MeV electron beam can generate incoherent EUV radiation at 13.5 nm. Such a light source could be less than 6 m long, and potentially be used for EUV mask metrology. Similar approach will enable soft X-Ray imaging.
4th Generation Light Sources will have brilliance performances which will exceed those of the 3rd Generation Light Sources by 10 orders of magnitude in peak value. 3rd Generation Light Sources were based on storage ring. Those sources had improved the quality of the X-Ray produced with respect to the 2nd generation of machines by reducing the emittance and by increasing the current of the stored beam. The great stability and small emittance were intrinsically obtained by the damping produced by the radiation emission itself. In 4th generation sources, the X-Ray source quality (brilliance, peak brilliance, coherence) will directly depend on the quality of the injected beam (emittance, peak current, average current). Also, its stability will primarily depend on that of their injector. 4th generation sources include both X-Ray FELs and ERL based sources. The technological challenges of injectors for X-Ray FELs include small emittances, high peak current, high stability and reliability. ERL based sources aim at the same type of performances, but in addition average current as high as those of the 3rd generation light sources is desired. The focus of this review will be on the technological challenges of X-Ray FELs sources but for solutions proposed by the ERL injector community which could benefit X-Ray FELs sources. Photoinjectors are the primary source choice for many of the X-Ray FELs under design and construction. Critical issues for this technology include optimum laser pulse shaping and high quality of emission from the photocathode. Beam performances obtained from photoInjectors up to date just fulfill the requirements of X-Ray FEL drivers, but adequate stability and reliability remain to be demonstrated. Alternate technologies to X-Ray FEL sources will also be briefly discussed.
For storage ring based synchrotron radiation sources which deliver high flux coherent X-ray pulses at a high repetition rate, the figure of merit is the brilliance. Third generation synchrotron radiation sources provide a brilliance in the 1020 range, in the usual units, for 10 keV photon beams. Such a performance approaches the ultimate limit imposed by the diffraction limit of synchrotron radiation emission. A gain of up to 6 orders of magnitude, with respect to the brilliance of second generation sources, was obtained by reducing transverse electron beam emittances and by optimizing undulator radiation. The horizontal emittance was dramatically decreased by reducing the dispersion function in the bending magnets. As a consequence, the dispersion of the electron revolution time around the ring, as a function of energy, was reduced. Under this quasi-isochronous condition, electron bunch lengths were naturally shortened and reached the few tens of ps range. The quasi-isochronous tuning of storage rings was further tested on third generation rings to evaluate the possibilities of reaching the subpicosecond range and improving the present peak brilliance of 1023. Unfortunately, the nature of the electromagnetic environment, with which the electron bunch interacts, makes the bunch lengthen with increasing current. Best predictions of peak brilliance performances achievable on storage rings will leave them far behind linac driven FEL sources if the principle of Self Amplified Spontaneous Emission works as predicted.