Electron source for sub-relativistic single-shot femtosecond diffraction

In this thesis we show unprecedented 100-fold compression of space-charge dominated electron bunches to sub-100 fs durations. Thereby we show for the first time full control of the fourth dimension (time) of sub-relativistic electron bunches. With 95 keV the kinetic energy of the electrons is within the optimum range of 30 { 300 keV for diffraction experiments. Furthermore, our bunches carry sufficient charge to record a diffraction pattern using only a single bunch. Compared to state-of-the-art diffraction setups we have increased the bunch charge by two orders of magnitude, combined with a factor 5 improved temporal resolution. With the work presented in this thesis we pave the way for the study of structural dynamics by means of single-shot, femtosecond electron diffraction. To realize extremely short, highly charged bunches the problem to be overcome is irreversible expansion due to the repelling Coulomb force. A uniformly charged ellipsoidal bunch, or `watterbag' bunch, is the only distribution that has space-charge fields which are linear functions of position. Its expansion is therefore reversible with external linear electro-magnetic fields. We have introduced the use of waterbag bunches into the sub-relativistic regime. In Ch. 2 we present analytical equations in closed form that describe the space-charge induced expansion of waterbag bunches. We create electron bunches by femtosecond photoemission in a 100 kV DC photogun. The necessary transverse shaping of the laser pulses and the robust design of the DC photogun are described in detail in Ch. 4. To compress the electron bunches we use an oscillatory electric field, sustained in a 3 GHz resonant radio-frequency (RF) cavity. This cavity thus acts as a temporal lens, a novelty in the sub-relativistic regime (see Ch. 3). Its shape has been optimized for power efficiency, saving about 90% power compared to a regular pillbox design. To measure the bunch length we use a 3 GHz streaking camera, in which the detector is another power efficient RF cavity. The strategy for power optimization of both cavities and the resulting designs are presented in Ch. 5. The resonant frequencies and the on-axis field profiles are in excellent agreement with the results of the numerical Poison solvers that we used to design the cavities. At optimum settings of the RF field amplitude and phase offset a shortest bunch duration of 67 fs has been measured for a 0:1 pC, 95 keV bunch (see Ch. 6). Bunch duration measurements as a function of the RF amplitude and the phase offset are in good agreement with state-of-the-art particle tracking simulations that include all Coulomb interactions of the electrons in the bunch, and utilize the detailed fieldmaps of the accelerator and the RF cavity. To show that our bunches are of sufficient quality for diffraction experiments, in terms of angular and energy spread, we have recorded single-shot diffraction patterns of a polycrystalline gold nanolayer (see Ch. 7). The four lowest-order diffraction peaks are easily resolved and their positions are in excellent agreement with theoretical values. Finally, preliminary measurements of the transverse phase-space and the transverse density profile of our bunches indicate that we have realized 95 keV, 0:1pC waterbag-like bunches (see Ch. 8). Further measurements on the longitudinal phase-space are desirable to confirm the realization of true waterbag bunches. With the work presented in this thesis we show that sub-relativistic single-shot femtosecond electron diffraction is possible. Thereby we provide an important analytical tool for the study of structural dynamics in, e.g., phase transitions, chemical reactions, and conformation changes with both atomic spatial and temporal resolution, i.e., 1ºA and 100 fs. With the present 100 kV photogun, the transverse coherence length is on the order of 1 nm. This allows the study of dynamics in a wide range of samples, that consist of crystals of atoms or small molecules. Our temporal charged-particle lens (i.e., the RF compression cavity) may be used as well in combination with the extraction of electron bunches from an ultracold plasma (another development in our research group), instead of extraction from a photocathode. This would lead to an increase of the transverse coherence length of the electron bunches by an order of magnitude. Such a development would enable the study of dynamics of relatively large (bio-)molecules at the atomic spatio-temporal scale, which will undoubtedly lead to new insight into the building-blocks of life