Surface and material analytics based on Dresden-EBIS platform technology

High current liquid metal ion sources are well known and found their first application as field emission electric propulsion thrusters in space technology. The aim of this work is the adaption of such kind of sources in broad ion beam technology. Surface patterning based on self-organized nano-structures on, e.g., semiconductor materials formed by heavy mono- or polyatomic ion irradiation from liquid metal (alloy) ion sources (LMAISs) is a very promising technique. LMAISs are nearly the only type of sources delivering polyatomic ions from about half of the periodic table elements. To overcome the lack of only very small treated areas by applying a focused ion beam equipped with such sources, the technology taken from space propulsion systems was transferred into a large single-end ion implanter. The main component is an ion beam injector based on high current LMAISs combined with suited ion optics allocating ion currents in the μA range in a nearly parallel beam of a few mm in diameter. Different types of LMAIS (needle, porous emitter, and capillary) are presented and characterized. The ion beam injector design is specified as well as the implementation of this module into a 200 kV high current ion implanter operating at the HZDR Ion Beam Center. Finally, the obtained results of large area surface modification of Ge using polyatomic Bi2+ ions at room temperature from a GaBi capillary LMAIS will be presented and discussed.

UltraCold Ion Beam Source. Focused ion beam (FIB) machines are largely used in the semiconductor industry mainly for imaging, milling and deposition of material, at the nanometer scale. To keep up with the trend of making smaller and smaller components, FIBs need to be continuously improved. A FIB is mainly made of an ion source and a focusing column. In this thesis, two concepts of ion sources are presented, modeled and discussed. The ultra-cold ion source (UCIS) is based on creating very cold ion beams (T <1 mK) by near-threshold photo-ionization of a laser-cooled and trapped atomic gas. In our experiment, 85Rb atoms are confined in a magneto optical trap built directly inside an accelerator structure. A spherical portion of the atom cloud is ionized with a 2-step ionization process and the bunched ions are accelerated by either a DC or a pulsed electric field. The liquid-metal ion source (LMIS) is the current state-of-art for focused ion beam technology, having a reduced brightness of 106 A/m2 sr eV. The reduced brightness characterizes the source and is proportional to the ion current and inversely proportional to the beam emittance (closely related with the source temperature). A high brightness is of importance as much as a low energy spread of the ion beam, since the latter is related with the focus ability of the beam. We have already shown that the UCIS can provide much lower energy spread (20 meV) than LMIS (˜ 4.5 eV), see Reijnders et al. (PRL, 2009), but not yet in combination with a high brightness as well. The energy spread is the limiting factor for the final achievable spot size in a FIB based on a LMIS. Chapter 3 and 4 present measurements of the source temperature and of the extracted current to complete the characterization of the UCIS. The ultra low temperature of the source permits collimated bunches to be created at low kinetic energy (down to few eV), which allows using time-dependent fields for accelerating and focusing. With this lens, a transversal ion temperature of (3±2) mK has been measured, as shown in chapter 3. This temperature is an upper limit since it is ten times higher than the expected value. The novel technique of focusing with time-dependent electric field is successfully applied to an ultra cold ion beam. In chapter 4, a dynamic model of the source describing its properties under pulsed operation has been developed and experiments validated it. A maximum average ion current of (13 ± 1) pA has been extracted from the UCIS and this is enough for imaging (1-100 pA are used). Those results correspond to a source reduced brightness of (8 ± 1) × 104 A/m2 sr eV with 0.9 eV energy spread. This result is 12.5 times lower than the brightness of a LMIS. An interesting pushing effect, due to the laser used to excite the atoms in the cloud, is described as well. Particle tracking simulations show that disorder induced heating has an important effect on the brightness, lowering its value by 1-2 orders of magnitude depending on the extracted current. Calculations and simulations of a different ion source, the atomic beam laser-cooled ion source (ABLIS), are presented in chapter 6. The follow-up project is based on a similar idea as an UCIS but aims for production of ion beams from a much higher initial atomic flux. The atoms originate from a Knudsen cell and, after being cooled and compressed transversally with laser cooling and trapping techniques, are photo-ionized and accelerated. The source size is reduced as well with respect to the UCIS. Realistic particle tracking simulations with 87Rb show that disorder induced heating is still very important, but in certain conditions, the so-called pencil-beam regime is reached and the disorder-induced heating does not deteriorate the brightness anymore. Depending on the extracted current, it is possible to reach a reduced brightness of 2 × 107 A/m2 sr eV with an ion current of 22 pA (corresponding to a final aperture diameter df = 1.6 µm) and an energy spread of 0.7 eV can be achieved. This source is likely capable of provide a high brightness and a low energy spread at the same time. The calculations are done assuming a purified 87Rb source. Eventually, it is possible to reach very high current (1 nA), making the ABLIS ideal also for milling applications with FIBs.

This paper discusses recent items of progress in understanding liquid-metal ion source (LMIS) behaviour, in the more general context of electrohydrodynamics. (I) Like the Taylor-Gilbert cone, the phenomenon of the vena contracta has been known for several hundred years. It is argued that the cusp-on-a-cone shape of an operating LMIS is a similar phenomenon, except that the forces acting are electrical rather than gravitational, and the pressure in the jet is negative rather than positive. (2) In discussing LMIS stability in engineering terms, mechanisms affecting the jet shape can be classified as giving negative feedback (stabilising), or positive feedback (destabilising). Effects associated with field evaporation and/or space charges give negative feedback, effects associated with pressure changes positive feedback. Attention is drawn to the arguments of Ganan-Calvo about the role of 'supercritical' flow as a stabilising factor against hydrodynamic disturbances. (3) Electrohydrodynamic spraying is superficially very similar to LMIS operation, and both sprayers and ion sources adopt a cusp-on-a-cone (or 'cone-jet') shape. But the accepted theoretical explanations of the driving mechanisms are different. Confirmation is offered that the LMIS story is the correct one for very highly conducting liquids, and that a difference in driving mechanism is plausible.

High electric fields are used for ion generation in gas field ion sources and liquid metal ion sources. With the preparation of stable microprotrusions on refractory metals, stably emitting gas field ion sources for hydrogen and helium ions with angular intensities of 35 and 20 μA sr−1, respectively, are available with a high chromatic angular intensity. Liquid metal ion sources can stably deliver a variety of ions with a source life in some cases over more than 1000 h. They are an essential element for pattern generation in the nanometer region and other focused ion beam applications.

The properties of liquid-metal ion sources are very sensitive to their tip structure. Because the impregnated-electrode-type liquid-metal ion source has a porous tip, it generates more than 300-μA ion currents of various metals including relatively high vapor-pressure metals such as Li, Cu, Ga, Ge, Ag, In, and Au, from which ions can be stably extracted. The large beam divergence, a common disadvantage in liquid-metal ion sources, can be overcome by designing new lens systems to efficiently transport the ion beam for both high- and low-energy applications. In addition, the ion beam current can be multiplied by using multiple emission cusps; thus, a germanium ion current of 1.3 mA was obtained by a three-point emission source. Therefore, the impregnated-electrode-type liquid-metal ion source can be utilized as a metal ion source for general applications.

Advances in metal ion sources

A high brightness plasma ion source has been developed to address focused ion beam (FIB) applications not satisfied by the liquid metal ion source (LMIS) based FIB. The plasma FIB described here is capable of satisfying applications requiring high mill rates (&amp;gt;100μm3∕s) with non-gallium ions and has demonstrated imaging capabilities with sub- 100-nm resolution. The virtual source size, angular intensity, mass spectra, and energy spread of the source have been determined with argon and xenon. This magnetically enhanced, inductively coupled plasma source has exhibited a reduced brightness (βr) of 5.4×103Am−2sr−1V−1, with a full width half maximum axial energy spread (ΔE) of 10eV when operated with argon. With xenon, βr=9.1×103Am−2sr−1V−1 and ΔE=7eV. With these source parameters, an optical column with sufficient demagnification is capable of forming a sub-25-nm spot size at 30keV and 1pA. The angular intensity of this source is nominally three orders of magnitude greater than a LMIS making the source more amenable to creating high current focused beams, in the regime where spherical aberration dominates the LMIS-FIB. The source has been operated on a two lens ion column and has demonstrated a current density that exceeds that of the LMIS-FIB for current greater than 50nA. Source lifetime and current stability are excellent with inert and reactive gases. Additionally, it should be possible to improve both the brightness and energy spread of this source, such that the (βr∕ΔE2) figure-of-merit could be within an order of magnitude of a LMIS.

The practical lateral spatial resolution of scanned ion microprobes has been limited until recently to ≳ 1μm by the relatively low brightness and large source size of duoplasmatron ion sources [1]. During the last decade liquid metal field-ion sources (LMFIS) have been developed with much higher brightness (~106 A/cm2.sr) and small source size (<100nm) [2,3] : it was recognised early on that these sources were potentially very suitable for ion probes with sub-micron resolution [1, 31. SELIGER and coworkers [4] showed that probe diameters of less than 100nm at 55keV were possible using gallium ion.KeywordsQuadrupole Mass FilterHigh Resolution SemiSmall Source SizeLarge Source Size200nm ResolutionThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Andromede project: Surface analysis and modification with probes from hydrogen to nano-particles in the MeV energy range

Recent studies on the use of Taylor cones of ionic liquids as ion sources have demonstrated a considerable effect of source geometry on the characteristics of the resulting ion beam. Of particular interest is the recent discovery by Lozano and Martinez-Sanchez of the excellent performance of externally wetted and electrochemically etched and sharpened tungsten tips. These sources are termed ionic liquid ion sources (ILIS), in analogy to conventional liquid metal ion sources (LMIS). Even more noteworthy is the fact that certain ionic liquids that had in a variety of prior reports always operated in the mixed ion-drop regime from capillary tips, do run exclusively in the purely ionic regime (PIR) on ILIS type sources. The present study uses time of flight mass spectrometry to compare full ion beams emitted from the two types of sources for a diversity of ionic liquids. We confirm earlier reports that internally fed capillary sources achieve purely ionic emissions more readily for liquids having high surface tension (γ > 40 dyn/cm) and high electrical conductivity (K ~ 1 S/m), and produce ion beam currents of several hundred nA with non-negligible energy spread. In contrast, externally wetted ILIS sources produce purely ionic emissions at tens of nA with liquids with substantially smaller γ and K, and much narrower energy distributions.

The cusp length of an operating liquid-metal ion source (LMIS) is of interest, since this may affect source stability with respect to the emission of liquid droplets. Some years ago, Mair and Forbes (1991,92) derived a formula that relates the length of the cusp of a liquid metal ion source to the ion emission current. Since our theory was originally published there have been several new experiments on a wider range of materials than we originally considered. So it has seemed useful to compare experiment and theory for most of the data now available.

Production of isotope-resolved beams of highly charged xenon ions

A joint research and development effort whose ultimate goal is to develop an intense, high charge state, ion source for mega-electron-volt ion implanters has been initiated. Present day high-energy ion implanters utilize low charge state (usually single charge) ion sources in combination with radio frequency (rf) accelerators. Usually, a MeV Linear Accelerator (MV LINAC) is used for acceleration of a few milliamperes. It is desirable to have instead an intense, high charge state ion source on a relatively low energy platform [direct current (dc) acceleration] to generate high-energy ion beams for implantation. This endeavor is a continuation of earlier research, which resulted in generating ions like Pb+7 and Bi+8 and ion currents exceeding 200 mA. The natural next step is to convert and optimize ion charge state enhancement techniques to generate B, P, As, and Sb ions, and adapt them to a dc implanter. A number of schemes are to be pursued simultaneously. The most promising approach is to be developed into a commercial ion source.

Secondary Ion Mass Spectrometry (SIMS) is a powerful technique for high spatial resolution chemical mapping and characterization of native surfaces. The use of massive cluster projectiles has been shown to extend the applicable mass range of SIMS and improve secondary ion yields 100 fold or beyond. These large projectiles however, present a challenge in terms of focusing due to the initial spatial and kinetic energy spreads inherent to their generation. In the present work, we describe the development and construction of a novel primary ion (PI) column employing a gold nanoparticle – liquid metal ion source (AuNP-LMIS) and the coupling to ultrahigh resolution mass spectrometers (e.g., Fourier Transform Ion Cyclotron Resonance Mass Spectrometer, FT-ICR MS) for accurate chemical characterization of complex biological surfaces. This work describes the ion dynamics, development and the experimental characterization of the AuNP-LMIS.

Summary form given only, as follows. Vacuum arc ion sources are used to generate high current broad beams of metal ions. The attraction of increasing the mean ion charge state in the beam is because of the possibility of thereby increasing the ion beam energy without applying higher extraction voltage. This is important both for heavy ion accelerator injectors and for ion implantation technologies. We have explored three different methods to enhance the degree of multiple ionization in vacuum arc plasmas. These methods are: application of a strong axial magnetic field in the arc region, generation of current spikes, and injected electron beam. A strong magnetic field increases the arc burning voltage because of improved confinement of the plasma column and leads to higher plasma electron temperature, which is primarily responsible for ionization. The B-field also increases the length of the ionization zone, beyond which there is no significant change in the ion charge state distribution (the distribution is frozen). Another approach to increase the arc voltage and electron temperature is to use an arc current spike of pulse duration comparable to the arc current relaxation time or less. This leads to a somewhat elevated arc voltage during the spike. Finally, additional ionization in the vacuum arc plasma can be brought about by injection into the plasma of an externally-produced electron beam. In this case the most important problem is to obtain optimum electron beam parameters and e-beam transport through the entire arc region. Here we describe these approaches and the experimental results of our investigations of ion charge state distribution. Mean ion charge states are presented, compared and discussed.