Simulations and experiments on external electron injection for laser wakefield acceleration

Laser wake field acceleration is a technique that can be used to accelerate electrons using electric fields that are several orders of magnitude higher than those available in conventional accelerators. With these higher fields, it is possible to drastically reduce the length of accelerator needed to reach a desired energy. This technique also allows the generation of (trains of) very short electron bunches with a duration of only a few femtoseconds. The electric fields are created by focusing a high power (10_12 - 10_15 Watt) laser pulse into a plasma. As the laser pulse moves through the plasma, it will push the plasma electrons away from the ions, creating a charge separation and accompanying electric field. Once the laser pulse is passed, the electrons will start oscillating around their original position, leading to a moving density fluctuation or plasma wave behind the laser pulse. If electrons are injected into this wave with enough speed/energy, they can 'surf' on the plasma wave like a surfer does on an ocean wave. While 'surfing', the electrons will be accelerated until they start surfing ahead of the wave and have reached their maximum speed. If the electrons continue on, they will have to climb the next wave, loosing energy. First we investigated the effects of the main experimental parameters (plasma density and length, initial energy of the injected electrons and the intensity of the laser pulse) on the acceleration of electrons. For this we used the General Particle Tracer-framework, a simulation suite that can be used to study the effects of electro-magnetic fields on the motion of charged particles. Within this framework, we introduced a simple plasma wave model consisting of a harmonic wave with the period and amplitude based on the experimental parameters. The plasma was assumed to have a parabolic profile such that it acts akin to an optical fibre, preventing the focused laser pulse from defocusing. The resulting accelerated bunches were evaluated based on their final energy, energy spread, charge and emittance. The goals was to create electron bunches with a high energy and charge and a low energy spread and emittance; such electron bunches will experience little bunch lengthening and are easier to focus down to small spots. From the results, the following has been concluded: A higher plasma density will generally lead to a lower final energy and a higher energy spread, emittance and accelerated charge. While a lower plasma density has the opposite effect. The total plasma length will affect the final energy, energy spread and emittance but has little to no effect on the trapped charge. The lengths at which the energy is at its maximum and the emittance and energy spread are minimal are very close to the optimal length predicted by 1-D linear theory. This means that we do not have to optimize the plasma length while studying the effects of the other parameters, but can use a simple analytical expression. When the energy of the injected electrons is too low, they will not be trapped by the plasma wave and will be expelled, leading to no accelerated electrons. Above the threshold for trapping, increasing the initial electron energy leads to higher energy spread and a higher number of accelerated electrons. The laser intensity shows a similar influence: too low and there is no trapping by the plasma wave. A higher intensity will lead to a higher energy spread and number of accelerated electrons. Additionally, a higher intensity gives rise to higher electric fields and as a result a higher final energy. With the results of this parameter study it was possible to create a design for an experimental setup to study the effects of external injection in a laser wake field accelerator within the constrains of the available infrastructure. The setup consists of three main components: A 2.6-cell photogun that serves as the source of the electrons that will be injected into the plasma wave (the 'surfer'). A 70 mm long aluminiumoxide tube with an inner radius of 100 µm, filled with hydrogen and two cathodes for the creation of the plasma channel (the 'sea'). And a 2 TW, 50-fs pulsed laser system to provide the needed laser power to generate the plasma wave (the 'tide') and provide the UV pulses needed to create electrons inside the photogun. The injection beamline has been optimized to keep the created electron bunches as short as possible while having sufficient energy and a small enough focal size to be efficiently trapped. The photogun and injection beamline are designed to produce an injected electron bunch of 10 pC at an energy of 6:7 MeV with a length of 100 µm and a radius of 30 µm. Acceleration by the plasma wave transforms this bunch into a bunch train of three bunches with a duration of 10 fs, an average energy of 90 MeV and a total charge of 1 pC. A more detailed design of the photogun and accompanying infrastructure was presented in Chapter 4 of this thesis. After the machining of the accelerating cavity of the photogun, it was tested at low power to compare the characteristics with those of the design. There were differences in both the length of the antenna needed for optimal coupling of the field into the cavity and the generated electric field inside the cavity. The observed difference in the needed antenna-length for optimum coupling is most likely due to the simulation method used in the design as it has been known to give similar, though smaller deviations in the past. The differences in the measured field profile are unlikely to be caused by any machining errors, but might be due to small slits between the different parts of the cavity. Though the photogun, has not yet been operated at the power levels set forth in the design, some initial measurements have been done on the produced electron bunches. These first measurements show that electron bunches can be produced with a charge of 73 pC with a 10% deviation, corresponding to fluctuations in the UV-laser used to generate the electrons. The average energy of the electrons was 3:4 MeV, due to the lower RF-power used to accelerate the electrons. In Chapter 5 of this thesis, we introduced a more advanced model to describe the plasma wave using relativistic hydrodynamic equations of motion for the plasma electrons and the Maxwell equations. The laser pulse was modeled by its intensity profile, averaging over the laser oscillations. This fluid model of the plasma was used to investigate the possible effects of various fluctuations in the experimental parameters when injecting electrons into a plasma wave: Synchronization between the laser and the electrons, plasma density and laser pulse energy. This was done for two different methods of injecting electrons: either behind the laser pulse into the plasma wave as is the case in the designed experiment or in front of the plasma wave as has been suggested as an alternative method. For injection in front of the laser, we studied three different regimes of laser power (10 - 30 - 100 TW). Especially, the synchronization between the laser used to generate the plasma wave and the injected electrons was identified as a critical parameter: Even the lowest timing jitter thus far reported can lead to a large increase in energy spread and a decrease of the final energy. For injection in front of the laser, it was shown that in the regime of the highest laser intensity (a0 = 1:02), the scheme was least sensitive to fluctuations while in the regime of lower laser power, it was more sensitive than injection behind the laser pulse. Finally, we introduced a new application of a plasma wave, namely that of a bunch compressor for electron bunches. This so-called plasma compressor uses the fact that when injected in front of a laser pulse, the plasma does not only accelerate, but also compress the injected electron bunch. By adjusting the original acceleration scheme, it is possible to make this compression both far less sensitive to synchronization jitter and allow for the compression of larger initial electron bunches. The compression factor of the plasma compressor can be varied by adjusting the plasma density. The bunch compressor is also cylindrically symmetrical, avoiding various sources of beam quality degradations due to symmetry breaking in a conventional bunch compressor. Concluding, we have looked at laser wake field acceleration using an external electron source as a method to create highly energetic, short electron bunches. We have looked at the influence of the different parameters of the process in order to design a proof of concept experiment. This experiment is currently partially implemented and tested. Also, we have studied the effects on the acceleration of fluctuations in several of the experimental parameters that can occur during the course of experiments. Finally, we have proposed a new application of a laser wake field accelerator; The plasma wave as an electron bunch compressor