The technology underlying the ability to focus MeV ion beams to spot sizes ranging from 100 nm to 1 μm using nuclear microprobes has continually evolved over the last 30 years to a high state of sophistication. However, it is only in the last few years that powerful analytical techniques have been developed that fully exploit its capabilities for materials analysis. This has opened up new and exciting applications for the characterization of crystallograph-ic and electronic properties of materials and increased our basic understanding of ion-solid interactions.

The operation of a nuclear microprobe is analogous to that of a scanning electron microscope, in which a focused beam of charged particles is swept over the sample surface to produce spatially resolved images using an analytical signal produced by the ion-solid interactions. Nuclear microprobes were originally developed to image variations in sample composition by measuring x-rays emitted from the sample, backscattered ions, or nuclear-reaction products, as are widely used with unfocused beams. Accounts of these analytical techniques using focused ion beams for applications in biomedicine, geology, and archaeology, and the micro-probe hardware required to accomplish them, are given in References 2–4.

Their high magnetic rigidity means that MeV ions require much stronger magnetic fields to focus them than those required for keV electrons in a scanning electron microscope.

Injection, transport, and recombination of charge in semiconductors are important phenomena for governing the operation of microelectronic devices, the function of radiation detectors, and for fundamental studies of the properties of advanced materials. A new technique for the study of these important phenomena is ion-beam induced charge (IBIC) microscopy, which uses a focused, scanned beam of MeV ions. The information provided by this technique is used to study single-event effects in operating integrated circuits (ICs), charge-collection efficiency in radiation detectors that contain defects or other nonuniformities, and radiation damage in semiconductor materials. Detailed information can also be provided for physics-based simulation codes.

As with the related and more widely known technique of electron-beam induced current (EBIC) microscopy, an energetic ion slowing in a sample creates charge in the form of electron-hole pairs. These may drift from their origin under the influence of an internal electric field or diffuse through field-free regions of the sample. In the case of IBIC microscopy, however, a single MeV ion creates a dense path of electron-hole pairs, displaces sample atoms, and may induce sample-atom spallation or fission. In addition to these atomic effects, the other major differences between IBIC and EBIC are that an MeV ion will typically reach depths an order of magnitude deeper in the sample than the electrons used in EBIC and will also undergo relatively minimal sideways scattering. Thus deeply buried structures can be probed with relatively little loss of spatial resolution. The collection of the charge primarily depends on the electric-field distribution from, for example, p-n junctions in an irradiated device and the material properties in the immediate vicinity of the induced charge. The ability to scan a focused MeV ion beam over the area of an integrated circuit, for example, enables the mapping of the combined effect of material properties and field distribution through the measurement of charge-collection efficiencies.

The passage of ions through a crystalline material can be affected by the regular and repeating arrangement of atoms within the crystal structure; this is called ion channeling. Channeling occurs when an ion beam is aligned with a planar or axial direction of a crystal lattice. The particles are shepherded through the crystal by the gentle steering effects of the atomic rows and planes so that they travel in the regions of reduced electron density between the atoms. This leads to a significantly reduced energy-loss rate compared with nonchanneled (or “random”) ions: MeV protons channeled in the {111} and {110} planes of silicon suffer an average energy-loss rate of 45% and 60% of the nonchanneled value, respectively. Channeling also lowers the close-encounter probability between the channeled ions and the crystal's atoms, reducing the rate of ion backscattering and ion-induced x-ray production. Comparison of the yields of these signals between channeled and nonchanneled alignment gives information on crystal quality, epitaxy, and the lattice position of interstitial elements.

Table I gives some useful parameters for 3-MeV protons channeled along axial and planar directions of silicon. The channeling critical angle ψ is a measure of how close in angle a channeled ion must be to an axial or planar direction in order for it to become channeled. This is of the order of 0.3° for ⟨110⟩ directions and 0.1° for the planar directions. Even when an ion beam is aligned to within the critical angle of a channeling direction, not every ion becomes channeled, as some are scattered by close interactions with the first few monolayers of the crystal. The fraction of non-channeled ions is given by the minimum yield Xmin and this is ~30% for planar directions and a few percent for axes. Those ions that become channeled eventually leave their privileged trajectories and revert to nonchanneled paths through the crystal. This dechanneling occurs naturally, due to scattering from electrons within the channel and perturbations from the atoms of the channel walls. For planar channeling directions, the fraction of ions remaining channeled decreases exponentially with depth into the crystal and is described by a half-thickness z1/2, the depth within the crystal at which 50% of the initially channeled ions have become dechanneled. This is of the order of 5 μm for 3-MeV protons channeled in the major planes of silicon. Finally, the energy-loss ratio shows the reduction in energy loss for a channeled ion compared with that of a random ion.

This article describes a new technique for nuclear microprobes that has been developed to produce angle-resolved channeling information from small areas of crystalline material without any rotation or translation of the sample. This “beam rocking” system, with a focused MeV ion beam, has been used to detect and quantify small interface rotation angles in strained Si1−xGex/Si samples with 0.015 < x ξ 0.175, as well as samples with graded germanium composition. By eliminating possible rotation errors caused by translation of the sample stage or backlash in the gears of a goniometer, smaller rotation angles can be measured than with standard ion-channeling analysis.

Channeling is frequently used in ion-beam analysis to characterize crystalline materials (see the previous article by King). If the crystal is positioned near axial or planar alignment and then tilted incrementally with respect to the beam, as the sample is brought into channeling alignment, the yield of interactions with the host lattice will be reduced as the ions become channeled, since they are less likely to interact closely with atomic nuclei. Angular scans can be used to measure strain in heteroepitaxial layers by measuring the rotation of off-normal axes using this process.

This article describes the production of different types of microstructures using high-energy heavy-ion beams, a process called single-ion micromechanics. The formation of high-resolution “radiation shadows” using unfocused, high-energy heavy ions is first described, and the rest of the article discusses the production of microstructures using precisely aimed single ions and the technology underlying this process using a microprobe.

The process of using heavy ions for microlithography was initially developed at GSI (Gesellschaft für Schwerionen-forschung, Darmstadt, Germany) because of their property of traveling in straight lines within matter. Owing to their very high momentum, they tend to undergo very little lateral scattering. This is an even more pronounced characteristic for heavy ions than for light ions, such as MeV protons (as were discussed for lithography in the article by Watt et al.). Figure 1 shows an example of this behavior for 7.5-MeV per nucleon xenon ions in mica. To make the tracks visible, the irradiated mica was broken along the tracks. The length of the individual tracks from top to bottom is about 25 μm.

Such heavy, high-energy ions cause greater radiation damage than lighter ions. Their large rate of nuclear energy loss results in their displacing lattice nuclei with enough imparted energy that these nuclei create more displacements in subsequent collisions of their own. Therefore, high-energy heavy ions are very efficient for creating microstructures, compared with other types of radiation, because each individual ion can produce an observable irradiation effect, while it always takes more than one light ion, electron, x-ray, or light photon to render developable one volume element of the sample. Figure 2 shows the lattice disorder created by a single high-energy uranium ion in germanium sulfide.