May_EDFA_Digital

edfas.org 27 ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 20 NO. 2 Fig. 1, VDW forces are more than one order of magnitude large than capillary forces, more than eight orders of magnitude larger than electrostatic forces, andmore than 13 orders of magnitude larger than the force of gravity. VDW forces are directly proportional to the Hamaker constant, which varies by two orders of magnitude across the periodic table. [6] Thus, a change in material is essen- tially irrelevant for the EXLO method and van der Waals forces dominate the adhesion process of the smooth FIB specimenmanipulatedwith a smooth probe surface. As a rule, excessive repulsive or attractive electrostatic forces should be avoided whenever possible. An easy way to avoid electrostatic forces, relying solely on VDW forces, is to metallize only the glass probe. This may be accom- plished by sputter coating a conductive layer on the glass probe [6] or by dipping the probe into permanent marker ink (e.g., carbon or conductive particles in a solvent). [9,10] Figure 2 shows the solid glass probes used in EXLO where the probe labeled (a) has been dipped into and painted with permanent ink, and the probe labeled (b) has no coating applied. Thus, theEXLOprocessworksbyoptimizing the surface area. The specimen is easily extracted from its trench because the probe-to-specimen surface area is greater than the surface area created by the specimen edges touching the trench sidewalls. The specimen releases from the probe and adheres to the carrier because the surface area between the specimen and carrier becomes larger than the surface area between the probe and specimen. Therefore, the specimen will remain either in the trench, on the probe, or on the carrier, unless acteduponby larger VDW or other forces. REPRODUCIBILITY AND THROUGHPUT WITH EXLO EXLO is easy to learn and master, and as previously mentioned, is highly reproducible and fast. A series of images illustrating this point is shown in Fig. 3, which includes still images acquired from a real-time video during the liftout andmanipulation of two specimens to a carbon-coatedgrid.Thespecimensare ~12-15µmlongand thegridsquareopenings are~40µmwide. Note the images include time stamps in minutes:seconds:milliseconds showing that two lift outs may be manipulated in < 5 minutes. [5,6] This translates to the achievable manipu- lation of 24 samples or more per hour. Ease of use aided by a semi-automated instrumentation processes helps achieve this speed. DEVELOPMENT OF SLOTTED HALF- GRIDS: NO CARBON FILM NEEDED Over the years, a perceived disadvantage to EXLO arose due to supposed limitations in S/TEM analysis and the ability for further FIB or ion milling due to the pres- ence and interference of the thin carbon film support. However, much S/TEM work can be performed through the carbon film or with a FIB specimen manipulated and centered over a hole in holey carbon film. In addi- tion, recent advances in slotted half-grids allow electron transparent FIB specimens to bemanipulated, supported, and protected directly by the smooth recessed regions provided by the grids. [6] The half-grid design also allows FIB specimens to be processed after EXLO manipulation, if desired. This allows for easy EXLO manipulation of (continued on page 30) (a) (b) Fig. 2 Solid glass probes used in EXLO. (a) Dipped into and painted with permanent ink and (b) as-pulled glass probe with no coating. Fig. 3 Image of the EXLO process of two specimens in < 5minutes. Still images with time stamps (minutes: seconds:milliseconds) acquired from a real-time video. [5]

RkJQdWJsaXNoZXIy MjA4MTAy