edfas.org 21 ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 27 NO. 3 This concept of dose is also used within the FIB community. However, the more relevant concept for FIB is the ion dose rate, which is the dose per unit time, C/(s*m2) or simply A/m2. In terms of FIB parameters, it is the dose per ion beam dwell time that really matters; alternatively, this can be thought of as the instantaneous ion beam current density. Figure 1 shows this concept for the Thermo Fisher Scientific Helios 5 Ga+ FIB. Figure 1a indicates the amount of ion beam defocus that is necessary to create a defined spot size on the sample’s surface, known as blur diameter in Thermo Fisher’s software. For example, Fig. 1a shows that it is necessary to defocus a 30 kV, 26 pA ion beam by approximately 1.5 mm to create a circular spot diameter of 5 µm at the sample’s surface, i.e., the well-focused beam at 10 mm should be focused at either 8.5 mm or 11.5 mm in order to create a 5 µm spot on the surface. In Fig. 1b, the instantaneous ion beam current density is calculated by assuming a circular area as a function of the spot diameters in Fig. 1a. It is not intuitive that the instantaneous current density of a well-focused ion beam (no defocus) is approximately the same value regardless of the beam current. This happens because the beam spot diameter with no defocus increases as the current increases. In fact, for this microscope, a low current beam (30 kV, 0.09 nA) has an instantaneous current density of approximately 286 nA/µm2, while a high current beam (30 kV, 20 nA) has an instantaneous current density of 129 nA/µm2. Figure 1b also shows that it is possible to systematically decrease the instantaneous current density with beam blurring and therefore, in theory, control whether the ion-induced thermal spikes overlap or not. ENHANCED MATERIAL DEPOSITION In applying this concept of beam defocus to material deposition with the ion beam, it is common to deposit FIB “Pt,” the precursor of which is (methylcyclopentadienyl)trimethyl platinum metalorganic gas molecules. This protects the sample’s surface prior to cross-sectioning or when attaching a TEM lamella to a micromanipulator. The standard rule of thumb used to optimize the deposition rate is to adjust the beam current depending on the size of the total deposition area so that the total current density is equal to 6 pA/µm2 for this Pt precursor.[3] For example, to deposit a Pt box with dimensions 1 µm wide by 15 µm long, a well-focused ion beam of 90 pA (90 pA/(1 µm * 15 µm) = 6 pA/µm2) is suggested by the rule of thumb. Setting aside this rule for a moment, let’s turn to the FIBID and FEBID communities that have the best understanding of this very complex deposition process, which is essentially room temperature, beam-assisted, sitespecific chemical vapor deposition (CVD). For full details of this process please refer to this FEBID tome[2] and the references therein. Briefly, to construct the very complex, layer-by-layer depositions that FIBID and FEBID are known for, the deposition condition must be in a beam-limited regime (excess precursor at the deposition location) and not a precursor-limited regime (excess ions and/or electrons at the deposition location) as is guaranteed when using a well-focused ion beam. Note that in the Pt deposition example above, the ion beam spot size has a diameter of approximately 20 nm, giving an instantaneous current density, J, of 286 nA/µm2 (90 pA/(π (20 nm/2)2) = 286 nA/µm2). According to the FEBID literature, the organo-metallic precursor gas creates a monolayer on the sample’s surface. It is cracked most efficiently by low energy secondary electrons generated from the primary (ion or electron) beam that exit the surface both inside and outside of the main beam spot. This cracked molecule then adheres to the substrate and becomes deposited material. When using a well-focused ion beam, there are too many Ga+ ions within the focused beam that generate too many secondary electrons, which overwhelm the monolayer of precursor. This results in a precursor-limited condition, leading to a combination of both deposition and milling which results in inefficient deposition rates. EXPERIMENTAL COMPARISON Figure 2 compares the traditional precursor-limited deposition of Pt as described above to the proposed beamlimited deposition of Pt using a defocused ion beam where Fig. 2 Comparison of Pt deposition methods. The deposition on the left is a traditional Pt deposition recipe that creates a precursor-limited condition with a deposition rate of approximately 2.5 nm/s. This is compared to the defocused ion beam deposition that creates a beamlimited condition and has an order of magnitude increase in deposition rate to approximately 30 nm/s.
RkJQdWJsaXNoZXIy MTYyMzk3NQ==