Mark Utlaut

Fundamental limits to imaging resolution for focused ion beams

This article investigates the limitations on the formation of focused ion beam images from secondary electrons. We use the notion of the information content of an image to account for the effects of resolution, contrast, and signal‐to‐noise ratio and show that there is a competition between the rate at which small features are sputtered away by the primary beam and the rate of collection of secondary electrons. We find that for small features, sputtering is the limit to imaging resolution, and that for extended small features (e.g., layered structures), rearrangement, redeposition, and differential sputtering rates may limit the resolution in some cases.

The focused ion beam as an integrated circuit restructuring tool

One of the capabilities of focused ion beam systems is ion milling. The purpose of this work is to explore this capability as a tool for integrated circuit restructuring. Methods for cutting and joining conductors are needed. Two methods for joining conductors are demonstrated. The first consists of spinning nitrocellulose (a self‐developing resist) on the circuit, ion exposing an area, say, 7×7 μm, then milling a smaller via with sloping sidewalls through the first metal layer down to the second, e‐beam evaporating metal, and then dissolving the nitrocellulose to achieve liftoff. The resistance of these links between two metal levels varied from 1 to 7 Ω. The second, simpler method consists of milling a via with vertical sidewalls down to the lower metal layer, then reducing the milling scan to a smaller area in the center of this via, thereby redepositing the metal from the lower layer on the vertical sidewall. The short circuit thus achieved varied from 0.4 to 1.5 Ω for vias of dimensions 3×3 μm to 1×1...

Computer simulation of current density profiles in focused ion beams

We have made a computer simulation of the target current density in our single‐lens focused ion beam column. This column images at unity magnification the virtual source of a liquid metal ion source. We simulate the nonlaminar properties of the emitted ions with a two‐dimensional function f (r,r’) that represents the phase space density distribution of the beam, where r and r’(dr/dz) are the radial position and slope, respectively, in the aperture plane. This analysis further shows how f (r,r’) can be described in terms of the experimentally measured target current density profile.

Transverse thermal velocity broadening of focused beams from liquid metal ion sources

Experiments have shown that the target current density in focused ion beam columns have long ‘‘tails’’ outside the central submicron region. We show that these tails result from a transverse velocity distribution which has a Holtsmark probability density. Both theory and experiment show that the tails are reduced as the system magnification and source current are reduced.

Long‐lifetime, reliable liquid metal ion sources for boron, arsenic, and phosphorus

Operation of liquid–metal ion sources based on palladium alloys that contain boron, arsenic, and phosphorus (singly or in combination) was studied. These sources, when run on refractory metal needles and heater ribbons, have exhibited high angular intensity (1.5–5 μA/sr), long lifetime (>150 h), low energy spread (<15 eV), and stable operation with extracted currents down to 2 μA.

Gallium-Induced Milling of Silicon: A Computational Investigation of Focused Ion Beams

Molecular dynamics simulations are performed to model milling via a focused ion beam ~FIB! .T he goal of this investigation is to examine the fundamental dynamics associated with the use of FIBs, as well as the phenomena that govern the early stages of trench formation during the milling process. Using a gallium beam to bombard a silicon surface, the extent of lateral damage ~atomic displacement! caused by the beam at incident energies of both 2 and 30 keV is examined. These simulations indicate that the lateral damage is several times larger than the beam itself and that the mechanism responsible for the formation of a V-shaped trench is due to both the removal of surface material, and the lateral and horizontal migration of subsurface silicon atoms toward the vacuum/crater interface. The results presented here provide complementary information to experimental images of trenches created during milling with FIBs.

Focused ion beam methods of nanofabrication: room at the bottom

Various focused ion beam (FIB) processes, which can generate two- or three-dimensional (2D or 3D) features on surfaces by removing or depositing material, are reviewed for their efficiency. Processes for removing material include direct ion milling, chemical etching with a reactive gas, reactive ion etching, ion implantation with post chemical etch, deposition (simultaneous FIB and gas jet), FIB induced nucleation, and ion resist lithography. Surface material removal and deposition efficiencies (cubic microns per nano- coulomb of ion beam current) are examined, and their ability to form 2D and 3D surface structures is analyzed. In general, the ion lithography processes are the most efficient, whereas direct ion deposition is very inefficient. FIB instrument capabilities are examined including their focused ion beam size and beam current characteristics. These FIB instruments commonly employ field emission LMISs (liquid metal ion sources), can focus beams to less than 10 nm diameter, and can rapidly mill in the 100 nm beam range. The removal and addition rates of material (cubic microns per second) are then examined for the various surface modification processes using these instruments. Since the milling throughput of FIB Instruments has increased by two orders of magnitude in the last decade, new nanofabrication applications are rapidly unfolding. Direct FIB milling, FIB chemical etching, and FIB deposition are now common processes in the semiconductor industry for circuit modification, defect analysis, and process control. Production applications include thin film magnetic head trimming and aperture fabrication.

Fabrication of bipolar transistors by maskless ion implantation

The first focused ion beam (FIB) arsenic ion implants are reported. A shallow junction, vertical npn bipolar transistor fabricated by maskless implantation of B and As is described. For comparison, devices on the same wafer were also processed with conventional, broad‐beam B and/or As implants. Good transistor performance is obtained for each type of implanted transistor. Device characteristics for FIB and conventional implants are generally the same. However, initial results indicate that diode quality and junction leakage appear somewhat degraded (excess generation–recombination) for FIB arsenic implanted devices. Characteristics of FIB boron implanted devices obtained over an extended period have been measured. These data indicate that wafer‐to‐wafer dose uniformity and quality (diode ideality and leakage currents) is equal to that for conventional implants (standard deviations <10%). Device‐to‐device quality on a single wafer is also equal for the two techniques, while the device reproducibility is so...

A focused ion beam secondary ion mass spectroscopy system

This article describes a Ga+ focused ion beam secondary ion mass spectroscopy system, and measures several quantities of interest to aid in interpreting secondary ion mass spectroscopy results. We have measured sputter yields and rates, estimated the instrument efficiency, and calculated useful yields and practical sensitivities for a variety of elements used in the semiconductor industry. We have performed measurements at the system base pressure, and have also introduced oxygen and iodine to determine any enhancement effects.

High‐gain lateral pnp bipolar transistors made using focused ion beam implantation

We report the fabrication of lateral pnp bipolar transistors using focused ion beam (FIB) implants of boron and phosphorus for the collector and base, respectively. The implants of B+, P+, and P++ were all at a dose of 1×1013 /cm2 and a beam voltage of 75 kV. These implants defined spaces between the emitter and collector regions of 0.5–1.50 μm; which, after diffusion and zero voltage depletion width effects were considered, produced effective on‐wafer device basewidths of ∼0.2 μm. For the best devices, values of hFE near 100 were obtained with good junction characteristics and at peak collector currents of 10 μA/μm of device width.