Ion beams from laser-cooled gases

High-brightness ion sources are important for several applications such as focussed ion beam (FIB) systems, in which an ion beam is focussed on a sample to a spot size of a few nanometer. These systems have become an important tool in a wide array of material science and technological applications because they offer both high-resolution imaging as well as micromachining capabilities. The resolution in the most common FIB system, the Gallium-Liquid Metal Ion Source (Ga-LMIS) based FIB, is limited by chromatic aberrations in the lens column due to the energy spread of o2 eV. To reach a higher resolution, a high-brightness ion source is needed with a lower energy spread. In this thesis such a new source is presented, the ultra-cold ion source (UCIS) based on the photo-ionization of a laser-cooled atomic gas confined in a magneto-optical trap. The low temperature (_ 150 µK) of the source makes it possible to reach high-brightness with an extended source size. The charge density at the source is thereby reduced compared to sources with a sharp tip, such as the LMIS, and thus also the Coulombic interactions that deteriorate the beam and increase the longitudinal energy spread. Additionally many different atomic species can be laser-cooled, making the UCIS a versatile source for both heavy and light ion species. Light ion species such as Li are very suitable for imaging purposes without directly damaging the sample. Heavy ion species on the other hand, can be used for manipulation of a sample by sputtering away material at a small scale. A analytical model has been developed to estimate the initial source performance. The extractable current from the ionization volume is estimated based on the influx of atoms, which depends on temperature and atomic density. According to the model, the energy spread oU is proportional to the finite length of the ionization volume, limited by the size of the laser focus waist, and the acceleration field E0. To study the influence of the ionic interactions in the beam in more detail, particle tracking simulations have been performed which take all mutual Coulombic interaction into account under realistic experimental conditions. From these simulations we learned that when the current is increased, the brightness is lowered due to increased stochastic heating. Beam currents between 1 and 100 pA are feasible, with an energy spread much lower than that of the LMIS. At 1 pA the source has a peak brightness up to Br = 8 × 104 Am-2sr-1 with an energy spread _U = 0.08 eV. At 10 pA it has a brightness up to Br = 2 × 104 Am-2sr-1 with an energy spread oU = 0.14 eV. An experimental setup has been built, consisting of a rubidium magneto-optical trap inside a cylindrically symmetric accelerator structure. By using a two-step photo-ionization process, a small fraction of the cold atoms in the cross-section of two laser beams is ionized. A pulsed ionization laser was used to produce ion bunches. A short beam line with a multichannel plate detector was used to obtain both the temporal and the transverse spatial profileof the bunches. Time-of-flight measurements have been performed to characterize the longitudinal energy spread of the beam as function of the beam energy U. We have demonstrated, as predicted by the model, that the energy spread oU is indeed due to the length of the ionization volume in the acceleration field. By lowering the extraction field, well defined beams with an average energy of only 1 eV have been produced with an energy spread as low as 20 meV, more than two orders of magnitude lower than the LMIS. A transition from ballistic behavior at high beam energies, where oU / U, to a space-charge dominated regime at low beam energies, where oU / pU, has been observed. In the cross-over region an interesting new effect has been observed in which the space charge forces reduce the energy spread. The experimental results agree well with detailed particle tracking simulations. The combination of well-defined ion beams at low energy and time-dependent acceleration fields opens new possibilities. Precise manipulation of the bunch distribution in both longitudinal and transverse phase space has been demonstrated. Reduction of the longitudinal energy spread has been realized by switching the accelerator field off before the ions have had a chance to leave the accelerating structure. A reduction by a factor 3 is achieved, which is limited by the inhomogeneity of the field. Calculations show that much stronger reduction is achievable in a more homogenous acceleration field. This would make low-energy spread possible without the need to go to small _z0 and E0, enabling even lower energy spread. We also demonstrated that the accelerator structure, which normally acts as a fixed negative lens, can be transformed into a versatile lens by using time-dependent fields. Control over the focal length and sign, as well as the sign and strength of the spherical aberrations can be obtained by tailoring the applied time-dependent acceleration voltage. The possibility to create a negative spherical aberration may be used to correct for the spherical aberrations in the focal columns which are presently limiting the spatial resolution of FIBs. The dependence of the properties of this lens on the parameters of the applied tri-polar pulse has been studied in detail, both experimentally and with particle tracking simulations, which are in good agreement. Finally, based on lessons learned, the design of a next generation UCIS is briefly discussed, based on photo-ionization of a laser cooled and intensified atomic beam. The principle is similar as an UCIS based on a MOT, but with a higher current density. Calculations show that this source should be able to produce orders of magnitude brighter beams in a more compact configuration. With this new design it should be possible to beat the LMIS not only in energy spread but now also in brightness