Miniaturized RF technology for femtosecond electron microscopy

In this thesis we have taken many of the RF ideas of Ura, Oldfield, and Hawkes from the 1960’s, coupled with modern technology and simulating power to investigate the usage of miniaturized RF technology for femtosecond electron microscopy without the mandatory use of femtosecond lasers. To achieve femtosecond time resolution, a TM110 cavity is used to streak a DC electron beam across a slit, creating ultrashort electron bunches at a regular repetition rate. To create a miniaturized cavity, the TM110 cavity was loaded with ZrTiO4. This reduced the size of the cavity by nearly a factor of 6, and the power was reduced by nearly a full order of magnitude. Characterization of the cavity found that the dielectric loaded cavity operated as it was designed to do, matching both simulations and theory. To test the RF technology developed in our group, a 30 keV SEM beam line was built on top of an optical table. This allows for the beam line to be extended ~2.5 meters beyond the end of the SEM, giving space and room for experiments that would not otherwise be available in a standard electron microscope. This allows for multiple cavity systems to be tested, as well as emittance measurements of a high quality beam, requiring long distances to resolve the small angles. Implementation of the TM110 cavity into the beam line proved quite successful, with the beam behaving as expected. Chapter 5 builds an analytical model which predicts the beam’s behavior through the cavity, and goes a step further to predict the growth in transverse normalized emittance and the energy spread of a beam as it traverses the cavity. The analytical model is then compared to particle tracking simulations, proving to be accurate for high quality beams. The measurements of the emittance demonstrate the behavior predicted by the analytical model, and match with the numerical simulations. An important point realized from the model is that focusing in the cavity makes all the difference, minimizing the emittance and energy spread growth. This is an important detail for using the TM110 cavity, since a high beam quality is always desired to obtain spatial resolution. Chapter 6 detailed a few of the myriad of ideas of which can be done with the RF technology studied within the scope of this project. First, as the project is part of a larger industrial partnership program with the private company FEI, a full simulated study of implementing the TM110 cavity into an existing FEI electron microscope was carried out in cooperation with the company. In the parameter space presented by FEI, the energy spread growth from the cavity in generating 100 fs bunches appears too high, reaching ~10.5 eV FWHM. However, by relaxing a few of the constraints, and changing the parameters of the setup, it was shown that the energy spread growth can be compensated for. Because of the compact size and low power consumption of the dielectric cavity, this makes the cavity an ideal candidate for femtosecond electron bunch generation in a microscope. Following the implementation study, the idea of reducing the repetition rate of the bunches by using a dual mode cavity was presented. This was an idea that stemmed from a "Huh, that is strange…" moment in lab. The moment is seen in Fig. 6.4; it is when the two formerly degenerate modes were being operated simultaneously in one cavity. If the cavity was re-designed to use the two modes, running with two different frequencies, the repetition rate of the electron bunches could then be tailored to match repetition rates in the MHz range, rather than GHz. The original designs of Ura and Oldfield included multiple cavities synchronized to manipulate the beam. However, Oldfield commented in his Ph.D. thesis that the phase control was of the utmost importance. In Sec. 6.3, we demonstrated a high precision of phase control between two cavities, a TM110 streak cavity and a TM010 compression cavity, and supported the results with numerical simulations. With phase control between multiple cavities demonstrated, both synchronized to the same RF signal, the door to time-dependent beam manipulation has flown wide open. Section 6.4 walks right through the multiple cavity door, and presents a two cavity setup to compete with Zewail’s photocathode driven ultrafast electron microscope (UEM). Using two TM110 cavities, femtosecond electron energy loss spectroscopy (FEELS) was simulated using particle tracking software. It was shown that with the two cavities, appropriately phased, an energy resolution of