ion–solid interactions that lead to the var- ious functionalities of FIBs. In the topical articles that follow, the major subspecial- Focused ties of FIB research are discussed. The FIB Instrument The basic functions of the FIB, namely, Microscopy and imaging and with an ion beam, require a highly focused beam. A consis- tent tenet of any focused beam is that the smaller the effective source size, the more Micromachining current that can be focused to a point. Unlike the broad ion beams generated from plasma sources, high-resolution C.A. Volkert and A.M. Minor, Guest Editors ion beams are defined by the use of a field ionization source with a small effective source size on the order of 5 nm, therefore Abstract enabling the beam to be tightly focused. The fairly recent availability of commercial focused ion beam (FIB) microscopes has The type used in all com- led to rapid development of their applications for . FIB instruments mercial systems and in the majority of have both imaging and micromachining capabilities at the nanometer–micrometer scale; research systems designed with microma- thus, a broad range of fundamental studies and technological applications have been chining applications in mind is the liquid- 6,7 enhanced or made possible with FIB technology. This introductory article covers the metal ion source (LMIS). Of the existing basic FIB instrument and the fundamentals of ion–solid interactions that lead to the ion source types, the LMIS provides the many unique FIB capabilities as well as some of the unwanted artifacts associated with brightest and most highly focused beam (when connected to the appropriate FIB instruments. The four topical articles following this introduction give overviews of optics). There are a number of different specific applications of the FIB in materials science, focusing on its particular strengths types of LMIS sources, the most widely as a tool for characterization and transmission electron microscopy sample preparation, used being a Ga-based blunt needle as well as its potential for ion beam fabrication and prototyping. source. Ga has decided advantages over other LMIS metals such as In, Bi, Sn, and Au because of its combination of low melting temperature (30∞C), low volatility, Introduction and low vapor pressure. The low melting The focused ion beam (FIB) microscope ment the FIB column with an additional temperature makes the source easy to has gained widespread use in fundamen- SEM column so that the instrument design and operate, and because Ga does tal materials studies and technological becomes a versatile “dual-beam” platform not react with the material defining the applications over the last several years (FIB–SEM, see Figure 1) for imaging, needle (typically W) and evaporation is because it offers both high-resolution material removal, and deposition at negligible, Ga-based LMISs are typically imaging and flexible micromachining in a length scales of a few nanometers to hun- more stable than other LMIS metals. single platform. dreds of microns. The FIB instrument During operation, Ga flows from a reser- The FIB instrument is similar to a scan- becomes a powerful tool for nanomanipu- voir to the needle tip (with an end radius ning (SEM), except lation and fabrication through the aug- of about 10 mm), where it is extracted by that the beam that is rastered over the mentation of an FIB instrument with field emission. A large negative potential sample is an ion beam rather than an elec- micromanipulators and gas injection for between the needle and an extraction tron beam. Secondary electrons are gener- local chemical vapor deposition (CVD). electrode generates an electric field of ated by the interaction of the ion beam The first FIB instruments evolved from magnitude 1010 V/m at the needle tip. The with the sample surface and can be used advances in field ion microscopes1 and balance between the electrostatic forces to obtain high-spatial-resolution images. through the development of high- and the Ga surface tension wetting the In most commercially available systems, resolution liquid metal ion sources tapered W needle geometry results in Ga ions are used, and their sputtering (LMISs).2–4 In the 1980s, FIB instruments the formation of a single Taylor cone at the action enables precise machining of sam- were embraced by the needle tip. For typical emission currents ples. In conjunction with the gas-injection industry as offline equipment for mask or used in FIB microscopes (~2 mA), a cusp capabilities on these systems, which circuit repair. It was not until the 1990s forms at the tip of the Taylor cone with a enable ion-beam-activated deposition and that FIB instruments began to be used in tip radius of approximately 5 nm. enhanced etching, a range of sample fabri- research laboratories, and today there are The simplest and most widely used ion cation schemes are possible. commercial instruments available from beam columns consist of two lenses (a During the last 25 years, FIB instrumen- multiple manufacturers.5 With the popu- condenser and objective lens) to define tation has become an important technol- larity of FIB instruments for TEM sample the beam and then focus it on the sample, ogy for a wide array of materials science preparation, microstructural analysis, and beam-defining apertures to select the applications, from circuit editing and nanomachining, dual-beam FIB instru- beam diameter and current, deflection transmission electron microscopy (TEM) ments are becoming a versatile and pow- plates to raster the beam over the sample sample preparation to microstructural erful tool for materials researchers. surface, stigmation poles to ensure a analysis and prototype nanomachining. This introductory article focuses on spherical beam profile, and a high-speed Most modern FIB instruments supple- the FIB instrument itself and the basic beam blanker to quickly deflect the beam

MRS BULLETIN • VOLUME 32 • MAY 2007 • www/mrs.org/bulletin 389 Focused Ion Beam Microscopy and Micromachining

gases, leading to local deposition of the conducting material (W, Pt, or C) or insu-

lating material (SiO2); see the article by MoberlyChan et al. in this issue. The local deposition of material also enables sophis- ticated micromanipulation within the FIB chamber, made possible through micro- manipulation accessories and the ability of the FIB to cut (sputter), paste (deposit), and watch (image) during a manipulation process within the chamber. The result is a system that can image, analyze, sputter, and deposit material all with very high spatial resolution and controlled through one software program. In a large part, it is this multifunctional versatility that has made FIB instruments popular among materials researchers. Ion–Solid Interactions Ion–solid interactions play an important role in many different endeavors, ranging from fabrication of microelectronic devices to understanding distributions of cosmic gases. This brief introduction will be limited to the processes and conditions Figure 1. Schematic illustration of a dual-beam FIB–SEM instrument. Expanded view relevant to the use of FIB systems in mate- shows the electron and ion beam sample interaction. rials science. For more detailed descrip- tions of ion–solid interactions, the reader is referred to books and review articles off the sample and onto a beam stop such to have a eucentric point (i.e., a well- on the subject 6,10–12 as well as to literature as a Faraday cup. Because the focusing centered point such that the field of view specifically on FIB.13–15 strength of an electromagnetic lens is is maintained when tilting the specimen) When an ion impinges on a solid, it directly related to the charge/mass ratio at the location where the two beams cross loses kinetic energy through interactions of a particle, it is impractical to build elec- (or at the working distance of the ion with the sample atoms. This transfer tromagnetic lenses for ions (which would beam, in the case of a single-beam FIB). of energy from the ion to the solid results weigh thousands of kilograms); thus, The region of interest on the sample is in a number of different processes (see focusing and steering are performed using moved to the eucentric point using trans- Figure 2): electrostatic components rather than the lation and rotation and then tilted for the ᭿ ion reflection and backscattering electromagnetic components used for desired angle of beam incidence. The total ᭿ electron emission electrons. current on the sample (sum of the incom- ᭿ electromagnetic radiation The size and shape of the beam intensity ing ion or electron beam and all emitted ᭿ atomic sputtering and ion emission profile on the sample determines the basic charged particles) is measured at the ᭿ sample damage imaging resolution and micromachining stage. ᭿ sample heating precision. Generally, the smaller the beam Depending on the application, the vari- The ion typically comes to rest in the diameter, the better the achievable resolu- ous emitted particles or radiation can be solid, leading to implantation of the ion. tion and milling precision, although the detected with appropriate detectors in the With the possible exception of electromag- requirements for the two applications are sample chamber. Traditional detectors netic radiation generation, all of these not exactly the same.8 For the energies, such as those in an SEM can be used to processes are important to FIB and currents, and acceptance angles used in detect the electrons or x-rays created by FIB–SEM system applications and are typical FIB systems, the beam spot size is the interaction of the ion beam with the described in the next sections. limited mostly by the chromatic aberration sample. The ions sputtered from the sam- that results primarily from the energy ple can also be detected using a variety of Collision Cascade spread of the beam due to space charge detectors such as charge electron multipli- Ion kinetic energy and momentum are effects at the ion source and secondarily ers, and mass selection of the sputtered transferred to the solid through both from the spherical aberration of the lenses. charged particles is also possible (second- inelastic and elastic interactions. In inelas- However, the ultimate spatial resolution ary ion mass spectrometry). FIBs derive tic interactions (called electronic energy for FIB imaging is, in fact, limited by sput- an important additional functionality loss), ion energy is lost to the electrons in tering and is thus sample-dependent.9 In through the use of gas-injection sources to the sample and results in ionization and modern FIB systems, the imaging resolu- deliver gas locally to either enhance the the emission of electrons and electromag- tion determined by the sputter-limited sig- etching rate or result in site-specific CVD. netic radiation from the sample. In elastic nal/noise usually is about 10 nm. Secondary electrons generated by the inci- interactions (called nuclear energy loss), The sample is mounted on a grounded dent ion beam (or, alternatively, the ion energy is transferred as translational stage with three-axis translation, rotation, incident electron beam in dual-beam sys- energy to screened target atoms and can and tilt capabilities. The stage is designed tems) can crack hydrocarbon precursor result in damage (displacement of sample

390 MRS BULLETIN • VOLUME 32 • MAY 2007 • www/mrs.org/bulletin Focused Ion Beam Microscopy and Micromachining

the collisions are nonadiabatic, because of the very short time scale. After approximately 10–11 s, the 5–30 keV Ga ion comes to rest in the solid, and the energies of all particles participating in the cascade have decreased below the dis- placement energy. At this point, the colli- sion cascade has ended. What remains are the emitted particles and radiation, and ion beam damage such as lattice defects, incorporated Ga, and heat, all of which may continue to interact and evolve. Molecular dynamics calculations are ideally suited for simulating collision cascades, because of the short length and time scales. Monte Carlo calculations are also well suited to simulating ion–solid interactions by including the “frictional” electronic stopping and stochastic elastic collisions. The most widely used Monte Carlo simulation is the program TRIM or SRIM (transport, or stopping range, of ions in matter).18 Such calculations for 30 keV Ga into elements from Li to Bi show that roughly two times as much of the ion energy is lost to nuclear energy losses than to ionization energy losses (the former from interactions with the nucleus of an atom and the latter from interactions with the electrons). Of the nuclear energy losses, the majority is lost through sample atom vibrations or heating rather than through vacancy formation. The projected and lateral ranges of the 30 keV Ga scale inversely with the sample density and are between 10 nm and 100 nm (projected) and between 5 nm and 50 nm (lateral). TRIM sputtering yields (sputtered target atoms per incoming ion) are between 1 and 20 for normal- incidence 30 keV Ga and increase some- what with atomic number.6 However, sputtering yield predictions depend criti- cally on surface binding energies, which are not well known and are sensitive to surface structure and chemistry. TRIM vacancy generation predictions between 300 and 1000 vacancies per incoming Figure 2. Schematic illustration of a collision cascade generated by a 30 keV Ga+ ion incident on a crystal lattice, showing the damage created in the collision cascade volume, ion are overestimates because defect diffusion and interactions are ignored. In and the projected range Rp and lateral range Rl of the implanted ion. addition, discrepancies between experi- mental and simulated ranges and colli- atoms from their initial sites) and sputter- example, creating an interstitial–vacancy sion cascade shapes in crystals can be ing from the sample surface. pair in a crystalline sample. This primary expected, because TRIM samples are The most widely accepted concept for recoil atom may have sufficient energy to isotropic and cannot capture channeling ion–solid interactions is the collision cas- displace further sample atoms (secondary effects. Despite these limitations, such cal- cade model16,17 (Figure 2). For the case of recoils), thus generating a volume where culations are invaluable in predicting 5–30 keV Ga impinging on most solids, large numbers of atoms have excess trends and in estimating the effects of the collision cascade involves a series of kinetic energy. If a displacement collision ion–solid interactions. independent binary collisions (the linear occurs near the surface, the recoil atom collision cascade regime). If the transla- may be emitted from the solid and lead to Ion Beam Imaging tional energy transferred to a target atom sputtering. The displacement energy (typ- In the same manner that images are during a collision exceeds a critical value ically on the order of 20 eV) is much larger generated in an SEM, the ion beam can be called the displacement energy, the atom than the binding energy for the atoms (of rastered over a sample surface and the will be knocked out of its original site, for the order of 1 eV), reflecting the fact that emitted electrons, particles (atoms and

MRS BULLETIN • VOLUME 32 • MAY 2007 • www/mrs.org/bulletin 391 Focused Ion Beam Microscopy and Micromachining

ions), and electromagnetic radiation can ion beam imaging always results in some Ion Beam Sputtering be detected. Conventional SEM imaging is Ga implantation and sputtering of the Because of the sputtering action of the based on detecting the secondary elec- sample surface. ion beam, the FIB can be used to locally trons (SEs). To date, most imaging in a FIB remove or mill away material. For exam- is based on detecting the low-energy elec- ple, a Sn sphere is progressively sputtered trons, often referred to as ion-induced sec- over a defined area in Figure 5. For direct ondary electrons (ISEs). Typically, 1–10 milling, the limiting feature size is typi- electrons with energies below 10 eV are cally about 10 nm9,19,20 (see the articles by generated per incoming 5–30 keV Ga ion. MoberlyChan et al. and Langford et al. in These electrons are created by both kinetic this issue). Quantitative aspects of sputter- and potential emission from the top few ing are complicated and depend on the atomic layers where the primary ion material, crystal orientation, ion beam impacts the solid as well as where incidence angle, and the extent of redepo- backscattered or sputtered particles exit sition. the sample. The total low-energy electron As the incidence angle of the ion beam yield depends strongly on surface oxida- is increased, the intersection of the colli- tion and contamination and thus will sion cascade with the sample surface change as the surface is sputter-cleaned increases, and the number of sputtered and Ga is incorporated. atoms per collision cascade increases Ion beams are not as finely focused as (a similar effect of geometry on electron electron beams and, partly for this reason, emission is shown in Figure 4d). they generally offer lower resolution. However, at the same time, the fraction However, the contrast mechanisms for ISE of reflected or backscattered Ga ions generation are different from those for increases. The combination of these two SE generation and can offer complemen- effects leads to a maximum in sputtering tary information about a sample surface. yield at an incidence angle of approxi- SE and ISE images of the same sample are mately 75–80∞. This effect has been con- shown in Figure 3. Both the SE (Figure 3a) firmed for 25–30 keV Ga into a variety of and the ISE (Figure 3b) images show con- materials, including single-crystal Si,19–22 21,22 trast due to surface topography and mate- amorphous SiO2, and polycrystalline rial differences. However, ISE imaging Au and W21 and shows good agreement typically delivers stronger channeling con- between experiment and theory. Si or trast from crystals than SE imaging. The amorphous solids are ideal for such a contrast due to crystal orientation is easily study because the effects of crystal chan- distinguished from material contrast, Figure 3. (a) Secondary electron (SE) neling are avoided (the surface region of because crystal contrast changes with the and (b) ion-induced secondary electron Si amorphizes under the Ga beam7). The incidence angle of the ion beam and mate- (ISE) images of an FIB-cut cross behavior is more complicated in crys- rial contrast does not. In ideal samples section in brass. Surface topography talline material where both incident angle generated during the FIB milling and 21 such as Cu or Au, ISE channeling contrast phase contrast are visible in both and channeling effects are present. can reveal twin lamellae as narrow as 20 images. The heavier second-phase The sputtering yield at a given inci- nm and grains as small as 50 nm. precipitates are bright in the SE image dence angle can vary by as much as a fac- The different contrast mechanisms are and dark in the ISE image.37 Channeling tor of 10 for strongly channeling crystal illustrated in Figure 4. A comparison of contrast showing the grain structure is orientations in materials such as Cu.23 This Figures 4a and 4b shows that when the visible only in the ISE image. is because for easy channeling orienta- crystal is oriented so that the ion “chan- nels” along crystal planes, there are fewer ion interactions with sample atoms near the surface and thus fewer electrons are emit- ted. Figure 4c illustrates that heavier sam- ples typically result in more ISEs (and SEs). Figure 4d shows that surface topography can lead to increases in the number of ISEs (and SEs), because of the increase in the number of ion–solid interactions near the sample surface. The SE and ISE images are also often distinguished by the amount of charging generated in insulating samples. Presumably because of differences in the low-energy/secondary electron yields and to the fact that the Ga implantation creates a thin conducting layer at the sam- ple surface, the FIB can often be used to Figure 4. Schematics showing the influence of (a), (b) crystal orientation, (c) atomic mass, image uncoated samples that are difficult and (d) surface geometry on 30 keV Ga+ collision cascades and ISE image contrast to image even with low-voltage SEM. formation. The orange atoms in (c) are more massive than the yellow atoms in (a), (b), and However, it is important to remember that (d). Similar concepts influence sputtering yields.

392 MRS BULLETIN • VOLUME 32 • MAY 2007 • www/mrs.org/bulletin Focused Ion Beam Microscopy and Micromachining

tions, the ion experiences only inelastic rastered over the surface.22 For instance, “curtain effect,”13 is seen in the lower half glancing-angle collisions with the atoms the sputter profile of a ring cut by rapid of the images in Figure 3. Surface rough- lying in a crystal plane and travels deeper and repetitive scanning (“multi-pass” ening, specifically ripple formation, is into the crystal before causing elastic colli- scanning) differs from that obtained by widespread during ion bombardment 9,26 sions, so that fewer atoms are sputtered slowly scanning the beam just once over and is attributed to competition between from the surface. This is analogous to the the same area (“single-pass” scanning) smoothing by surface diffusion or viscous effect of crystal orientation on low-energy (Figure 6). Despite an identical total flow and roughening because of surface- electron yields illustrated in Figures 4a ion dose (ions per unit area) for both rings curvature-dependent sputter yields. (The and 4b. Channeling effects on sputtering in Figure 6, the slow single-pass scan, sputter yield depends on local curvature at a vertical grain boundary in Cu are spiraling from outside to inside, results in for the same reasons it depends on angle shown in Figure 6. The grain with the a deeper cut, because of effects of ion of incidence.) smaller electron yield appears dark in focusing, incident angle effects, and rede- Even during normal-incidence sputter- Figure 6. Its lower sputter yield results in position. ing, surface roughening can occur and is a shallower sputter profile than that Redeposition decreases the effective dependent on the crystal orientation, as observed for the brighter neighboring sputter yield and changes sputter profiles. shown in the multi-pass ring cut across grain. The decreased yield comes about because two differently oriented grains in Figure The sputter profiles also depend on the redeposited material lands in the area 6. Such crystal-orientation-dependent exact sequence in which the ion beam is being sputtered and must be sputtered a rippling is attributed to anisotropic sur- second time. Redeposition is also given as face diffusion.27 Other unusual effects a reason why completely vertical side- of sputtering and redeposition on surface walls cannot be cut with the FIB without evolution continue to be discovered.28,29 over-tilting the sample,13,24 but it is cer- Nonetheless, progress is being made tainly also partly because of the intensity toward models that can predict surface tails of the ion beam profile and of the evolution during sputtering and can be decrease in sputter yield at high incidence used to achieve desired sputter profiles angles.19,20,25 Many details of redeposition (see the article by Langford et al. in this effects remain open, such as the develop- issue). ment of crystal orientation and channeling effects seen in the redeposited material Ga Incorporation (e.g., Figure 6, single-pass ring). Imaging and milling with Ga ions In addition to redeposition, surface always result in Ga incorporation near the roughening and shadowing effects are sample surface. As the sample surface is prevalent during sputtering. An example sputtered away at a rate proportional to of a shadowing effect on topology during the sputtering yield and the ion flux (ions cross-sectioning with the FIB, the so-called per area per time), the Ga is implanted

Figure 5. (a)–(c) Secondary electron images during sputtering of a large Sn Figure 6. Ion-induced secondary electron image of rings milled with 30 keV Ga at a grain sphere surrounded by Sn droplets, boundary in Cu. The single-pass ring shows an enhanced sputtering yield (deeper trough) using a 30 keV Ga beam rastered and redeposition (sloped, thicker sidewalls); the multi-pass ring shows channeling effects across a 10 mm ¥ 10 mm area. and surface roughening.

MRS BULLETIN • VOLUME 32 • MAY 2007 • www/mrs.org/bulletin 393 Focused Ion Beam Microscopy and Micromachining

further into the sample, and a steady-state ability to obtain high-quality images from profile of Ga is reached. The maximum Ga certain materials. concentration occurs at steady state after the removal of target material to a depth Ion Beam Damage roughly equal to the ion range and is con- A major drawback of FIB imaging and stant over a depth also roughly equal to machining, particularly for TEM samples, the ion range. Ignoring the effects from is the damage created by the ion beam possible diffusion of Ga in the target mate- (again, see Mayer et al. in this issue). rial, differences in molar volume, and As the ion dose increases, the individual preferential sputtering, the Ga atom frac- disordered cascade regions overlap and tion at steady state is a damaged surface layer is formed. Depending in particular on the sample a material and temperature, the ion beam fGa = 1/( Y), (1) damage can take the form of sample sur- where a is the fraction of ions that are face amorphization, point defect creation, not reflected or backscattered from the dislocation formation, phase formation, sample surface, Y is the sputtering yield grain modification, or other unusual (number of sputtered atoms/ions per effects. With the exception of Si amor- incoming ion)10 and aY is the sputtering phization, systematic investigations of FIB yield per implanted ion. Based on this damage are just beginning. Nonetheless, equation and typical TRIM sputtering several trends can be identified based on yields (from 1 to 20 sputtered atoms literature results for broad beam ions and per incoming ion), atom fractions from on anecdotal FIB observations. 1 at.% to 50 at.% Ga are expected near the Ion beam amorphization is a well- sample surface. Because of the increase in known phenomenon and has been exten- sputter yield with incident angle, the sively studied for covalently bound steady-state Ga concentration is expected materials such as Si, Ge, GaAs, and C to be roughly 5–10 times smaller during (diamond). Because of the highly direc- glancing-angle sputtering than normal- tional nature of the atomic bonds in incidence sputtering. In contrast, channel- covalent materials and in certain alloys, ing is expected to raise the steady-state Ga the atomic rearrangements necessary to concentration. heal the disorder created by the ion beam Most studies report Ga concentrations are often hindered. In contrast, pure met- Figure 7. (a) Transmission electron in good agreement with simulations and als have nondirectional bonding and do microscopy (TEM) image of a Cu film, showing that the existing dislocation theory, although a few claim to see almost not amorphize. Thus, thin amorphous structure was modified by FIB milling. no Ga. Unless the Ga diffuses away or layers sometimes observed at the edge (b) Schematic illustration of the milling is reflected, it must end up in the surface of pure metal TEM samples made by geometry. (TEM image courtesy of G.P. region of the sample. Some studies were FIB presumably contain impurities such Zhang.) performed with energy-dispersive spec- as Ga, C, or O. Some oxygen gets in troscopy in an SEM, which for typical elec- because of the relatively poor quality of tron beam parameters will look through the vacuum. tures, can lead to chemical changes in the the thin Ga-doped layer at the surface and Point defects and dislocation loops surface region and influence the ease of underestimate the actual Ga concentration. can also be created during FIB imaging damage formation and amorphization. Other studies have used surface Auger and machining. Systematic studies of FIB- analysis (without sputtering) that may induced defects have not been under- Ion Beam Heating also underestimate the Ga concentration, taken, although there are many isolated During , almost all of because of the presence of carbonaceous or observations, both published and unpub- the ion kinetic energy is eventually oxide surface layers. Carbon-based surface lished. For example, Cu is prone to exten- converted to heat, with only a small frac- layers are particularly prevalent in sam- sive FIB damage,31 as shown in the TEM tion stored as defects in the sample or ples that have been imaged in the SEM image in Figure 7, but Al is not and emitted as energetic particles or radia- after FIB milling.30 In contrast, Auger even provides reasonable-quality high- tion.14 For times longer than approxi- depth profiles or chemical analysis in resolution TEM samples.32 Tooth enamel mately a nanosecond and distances larger the TEM give reasonable agreement (hydroxyapatite) is resistant to FIB- than around 100 nm, the ion beam can with predictions for both Ga concentra- induced damage,33 but other apatite be approximated as a continuous heat tion and range (allowing for channeling crystals that decompose at lower temper- source. At shorter times, there are large effects). atures amorphize easily. Several addi- temporal variations in heating, and at All of these considerations become tional and unusual types of damage times of less than 10–12 s, the atoms barely much more complicated in alloys or if Ga have been observed in FIB-milled sam- have time to interact with each other, and diffusion or reactions take place. One ples. These include the formation of Ga- the temperature of the solid is not well unusual example is the formation of Ga- containing surface phases (see Mayer defined.15 The maximum temperature containing surface phases during imaging et al., this issue) as well as ion-beam- reached in a sample depends on the beam of certain fcc metals (see the article by induced grain growth in fine-grained Ni power P, sample thermal conductivity k, Mayer et al. in this issue). This phenome- and Ni alloys.34 Preferential sputtering, sample geometry, and contact to a heat non, which is coupled with sample crystal which is prevalent during FIB milling of reservoir.15,35 Beam powers in commercial structure and channeling effects, limits the materials that decompose at low tempera- systems have maximum values of 1 mW.

394 MRS BULLETIN • VOLUME 32 • MAY 2007 • www/mrs.org/bulletin Focused Ion Beam Microscopy and Micromachining

When the ion beam is incident on a flat it also facilitates the fabrication of lamellae 10. M. Nastasi, J.W. Mayer, J.K. Hirvonen, Ion- surface, the transfer of heat away from the from composite samples consisting of Solid Interactions: Fundamentals and Applications incidence point is so effective that even in materials with very different properties. (Cambridge University Press, Cambridge, UK, the absence of a heat reservoir (e.g., for a The strengths of the method as well as the 1996). 11. E. Chason et al., Appl. Phys. Rev. 81, 6514 semi-infinite sample), a finite steady-state problems, such as FIB-induced damage 36 (1997). temperature increase is reached given by and Ga contamination, are illustrated with 12. J.S. Williams, J.M. Poate, Ion Implantation examples. The second article, by Uchic and Beam Processing (Academic, Sydney, T = P/(pak), (2) et al., describes how the FIB can be used as 1984). a tool for three-dimensional characteriza- 13. J. Orloff, M. Utlaut, L. Swanson, High where a is the radius of the circular ion tion by complementing 2D imaging or Resolution Focused Ion Beams: FIB and its beam profile on the sample surface. For mapping with serial sectioning at a Applications (Kluwer Academic, Dordrecht, values of P/a available in commercial submicron level. The third article, by 2002). FIBs, between 1 W/m and 1000 W/m, the Langford et al., is organized according to 14. L.A Giannuzzi, F.A. Stevie, Introduction to Focused Ion Beams: Instrumentation, Theory, the general technological area, such as temperature rise predicted by Equation 2 Techniques, and Practice (Springer, New York, can range from entirely negligible for sam- microelectronics, photonics, microelectro- 2005). ples with good thermal conductivity to mechanical systems, and rapid prototyp- 15. J. Melngailis, J. Vac. Sci. Technol., B 5, 469 huge for poor conductors. For example, ing. The strengths and limitations of FIB (1987). for Si (k = 148 W/m K) the temperature micro- or nanostructuring are illustrated 16. P. Sigmond, Phys. Rev. 184, 383 (1969). increase is <2∞C even for the most extreme through representative examples, and the 17. P. Sigmond, J. Mater. Sci., 8, 1545 beam conditions. In contrast, for poly- article ends with a brief summary and a (1973). meric or biological materials (k ~ 0.1 W/m look to the future of FIB nanostructuring. 18. J.F. Ziegler, J.P. Biersack, U. Littmark, The K), such small temperature rises are only In the fourth and final article, by Stopping Range of Ions in Solids (Pergamon Press, P a MoberlyChan et al., advanced topics such New York, 1984). The SRIM code is available achieved with / values of <2 W/m, online at www.srim.org (accessed February which is at the low end of what is avail- as single-ion implantation, surface mor- 2007). able in commercial FIBs. phology during ion-induced erosion, and 19. C. Lehrer, L. Frey, S. Petersen, H. Ryssel, Beam heating can be diminished by ion-induced CVD are discussed. J. Vac. Sci. Technol., B 19, 2533 (2001). placing samples in good contact with a We hope that this issue of MRS Bulletin 20. L. Frey, C. Lehrer, H. Ryssel, Appl. Phys. heat reservoir. On the other hand, when will inform readers about the fundamental A 76, 1017 (2003). imaging or machining TEM lamellae, science and current applications of focused 21. X. Xu, A.D. Della Ratta, J. Sosonkina, membranes, or other structured samples, ion beam microscopy and micromachin- J. Melngailis, J. Vac. Sci. Technol., B 10, 2675 much higher temperatures may be ing and give them a sense of the future (1992). reached when the sample geometry limits potential of FIB in materials science. 22. D. Santamore, K. Edinger, J. Orloff, J. Melngailis, J. Vac. Sci. Technol., B 15, 2346 the transfer of heat.35 An extreme example Acknowledgments (1997). of this is when the heat transfer is reduced 23. B.W. Kempshall et al., J. Vac. Sci. Technol., B to one dimension, such as for a cylindrical A.M. Minor was supported by the 19, 749 (2001). pillar with a diameter equal to the beam Director, Office of Science, Office of Basic 24. B.I. Prenitzer et al., Metall. Trans. A 29, 2399 diameter. In this case, the temperature rise Energy Sciences, of the U.S. Department (1998). is given by Equation 2 multiplied by the of Energy under contract DE-AC02– 25. T. Ishitani, H. Tsuboi, T. Yaguchi, H. Koike, height-to-diameter ratio of the pillar. 05CH11231. J. Electron Microsc. 43, 322 (1994). Thus, by determining the length-to-width 26. E. Chason, M.J. Aziz, Scripta Mater. 49, 953 References (2003). ratio of the thermal path in a structured 1. E.W. Mueller, T.T. Tsong, Field Ion Microscopy 27. H.X. Qian et al., Appl. Surf. Sci. 240, 140 sample, a worst-case estimate of the actual Principles and Applications (American Elsevier, (2005). temperature rise can be obtained and used New York, 1969). 28. H.H. Chen et al., Science 310, 294 to select appropriate beam conditions. 2. V.E. Krohn, G.R. Ringo, Appl. Phys. Lett. 27, (2005). However, imaging or cutting of high- 479 (1975). 29. M. Castro, R. Cuerno, L. Vázquez, R. Gago, aspect-ratio features in low-conductivity 3. R.L. Seliger, J.W. Ward, V. Wang, R.L. Phys. Rev. Lett. 94, 016102 (2005). materials may lead to unacceptably high Kubena, Appl. Phys. Lett. 34, 310 (1979). 30. C.A. Volkert, E.T. Lilleodden, Philos. Mag. temperatures, even for the most mild 4. L.W. Swanson, Nucl. Instrum. Methods Phys. 86, 5567 (2006). beam conditions in commercial systems. Res., Sect. 218, 347 (1983). 31. J. Marien, J. Plitzko, R. Spolenak, R. Keller, 5. For example, FEI Co., Carl Zeiss Inc., Seiko J. Mayer, J. Microscopy 194, 71 (1999). Instruments Inc., Hitachi Inc., JEOL Ltd., Orsay 32. W. Sigle, private communication (2002). In This Issue Physics. 33. C.A. Volkert, S. Busch, B. Heiland, G. Four articles follow this introduction 6. A. Benninghoven, F.G. Rüdenauer, H.W. Dehm, J. Microscopy 214, 208 (2004). and are intended to cover the most widely Werner, Secondary Ion Mass Spectrometry: Basic 34. C.M. Park, J.A. Bain, J. Appl. Phys. 91, 6380 used FIB applications as well as the cur- Concepts, Instrumental Aspects, Applications and (2002). rent state of the art in FIB materials Trends (John Wiley, New York, 1987). 35. T. Ishitani, H. Kaga, J. Electron Microsc. 44, 7. P.D. Prewitt, G.L.R. Mair, Focused Ion Beams 331 (1995). research. In the first article, by Mayer from Liquid Metal Ion Sources (John Wiley, New 36. H.S. Carlslaw, J.C. Jaeger, Conduction of Heat et al., one of the most important applica- York, 1987) p. 291. in Solids (Oxford University Press, Oxford, UK, tions of the FIB is presented. The FIB not 8. J. Orloff, Rev. Sci. Instrum. 64, 1105 (1993). ed. 2, 1959) p. 264. only enables the preparation of site- 9. J. Orloff, L.W. Swanson, M. Utlaut, J. Vac. Sci. 37. T. Ishitani, Y. Madokoro, M. Nakagawa, K. specific samples of uniform thickness, but Technol., B 14, 3759 (1996). Ohya, J. Electron Microsc. 51, 207 (2002). ■

MRS BULLETIN • VOLUME 32 • MAY 2007 • www/mrs.org/bulletin 395 Focused Ion Beam Microscopy and Micromachining

Cynthia A. Volkert Andrew M. Minor David P. Adams Michael J. Aziz Yongqi Fu

Cynthia A. Volkert, engineering from Yale ogy techniques, investiga- MA 02138 USA; tel. 617- Fu can be reached at Guest Editor for this issue University and MS and tions of surface morphol- 495-9884 and e-mail Precision Engineering of MRS Bulletin, is a PhD degrees in materials ogy during ion [email protected]. and research scientist and science and engineering irradiation, and surface Center, School of group leader at from the University of micromachining. He has Yongqi Fu is a research Mechanical and Forschungszentrum California, Berkeley. His authored or co-authored fellow and leader of the Aerospace Engineering, Karlsruhe in Germany, research group focuses on more than 50 publications focused ion beam group Nanyang Technological where she specializes in the mechanical properties and is a member of the in the Precision University, 50 Nanyang microstructure and of small volumes, in situ Materials Research Engineering and Ave., Singapore 639798; mechanical properties TEM technique develop- Society. Nanotechnology (PEN) tel. 65-67906336, fax studies of small metal ment, and sample manip- Adams can be reached Center at Nanyang 65-67904674, e-mail structures. She received a ulation and preparation at Sandia National Technological University [email protected], and bachelor’s degree in methods for electron Laboratories, 1515 (NTU) in Singapore. He URL www.ntu.edu.sg/ physics from McGill microscopy investigations Eubank Blvd. SE, MS received his BEng home/yqfu/fyqpage.htm. University and her PhD of both hard and soft 1245, Albuquerque, NM, (mechanical engineering, degree from Harvard materials. 87123 USA; tel. 505- 1988), MSEng (optoelec- Lucille A. Giannuzzi is a University. Volkert then Minor can be reached 844-8317 and e-mail tronics, 1994), and PhD product marketing engi- spent 10 years working as at Lawrence Berkeley [email protected]. (optical engineering, neer for focused ion beam a staff scientist at Bell National Laboratory, 1 1996) degrees from Jilin and DualBeam™ systems Laboratories before mov- Cyclotron Rd., MS 72, Michael J. Aziz is the University, Changchun at FEI Co. in Hillsboro, ing to Germany, first to Berkeley, CA 94720 USA; Gordon McKay Professor University of Science and Oregon. She received her the Max Planck Institute tel. 510-495-2749, fax 510- of Materials Science at Technology, and the BE and MS degrees from for Metals Research in 486-5888, and e-mail Harvard University. Aziz Chinese Academy of SUNY–Stony Brook and Stuttgart and then to [email protected]. has made significant con- Sciences, respectively. Fu her PhD degree from the Karlsruhe. tributions to the self- worked in the State Key Pennsylvania State Volkert can be reached David P. Adams is a dis- organization and kinetics Laboratory of Applied University. Giannuzzi at Institut für tinguished member of of formation, thermal sta- Optics from 1996 to 1998 joined the University of Materialforschung II, technical staff at Sandia bility, and decay of as a postdoctoral fellow. Central Florida in 1994 Forschungszentrum National Laboratories. nanoscale structures; thin- He then worked in the and is founding director Karlsruhe, Postfach 36 Adams received his bach- film growth by pulsed PEN Center at NTU as of the Materials 40 76021, Karlsruhe, elor’s degree in physics laser melting and pulsed a research fellow from Characterization Facility. Germany; tel. 49-7247- from the University of laser deposition; growth, 1998 to 2001. Fu also She joined FEI Co. in 82-3577, fax 49-4247- Virginia and his PhD diffusion, and viscous worked in the 2003. 82-2347, and e-mail degree in materials sci- flow in stressed solids; Innovation in Her research interests cynthia.volkert@imf. ence and engineering and the kinetics of rapid Manufacturing Systems include ion–solid interac- fzk.de. from the University of alloy solidification. and Technology (IMST) tions, grain-boundary dif- Michigan. Aziz is a recipient of Program of the fusion and segregation, Andrew M. Minor, Guest Throughout his techni- the Presidential Young Singapore–MIT Alliance and structure–property Editor for this issue of cal career, Adams has Investigator Award, the as a research fellow from relationships of materials MRS Bulletin, is a staff sci- been involved with thin- ONR Young Investigator 2001 to 2005. Fu is a using ion and electron entist and principal inves- film deposition, studies of Award, and the Sauveur guest professor at the beam microscopy. She has tigator at the National microstructure and mor- Memorial Lectureship. He Institute of Optics and received an NSF CAREER Center for Electron phology evolution during also is a fellow of APS Electronics, Chinese Award and is a member Microscopy at Lawrence film growth, and materi- and AAAS. Academy of Sciences, of numerous societies Berkeley National als analysis. His current Aziz can be reached at and has authored more including the Materials Laboratory in Berkeley, activities involve focused the School of Engineering than 80 peer-reviewed Research Society. California. He received a ion beams, including and Applied Sciences, journal papers and 21 Giannuzzi has authored bachelor’s degree in eco- rapid prototyping, devel- Harvard University, 29 conference proceeding more than 100 publica- nomics and mechanical opment of in situ metrol- Oxford St., Cambridge, papers. tions, is an instructor at

396 MRS BULLETIN • VOLUME 32 • MAY 2007 • www/mrs.org/bulletin Focused Ion Beam Microscopy and Micromachining

Lucille A. Giannuzzi Jacques Gierak Gerhard Hobler Lorenz Holzer Beverley Inkson the Lehigh Microscopy 33-1-69-63-60-75, fax which deals mainly with current research interests Technologies School, and is co-editor of 33-1-69-63-60-06, 3D microscopy and are centered on the Corporation, 11-1, the book Introduction to e-mail jacques. image-based modeling nanomanipulation and Ishikawa-cho, Focused Ion Beams. [email protected], and for quantitative mechanical properties of Hitachinaka-shi, Ibaraki- Giannuzzi can be URL www.lpn.cnrs.fr/fr/ microstructure analysis. nanostructured materials, ken, 312-0057, Japan; reached at FEI Co., 5350 PHYNANO/NanoFIB. For the last three years, he including in situ FIB tel. 81-29-354-2970, fax NE Dawson Creek Dr., php. worked on FIB tomogra- and TEM nanotesting, 81-29-354-1971, and Hillsboro, OR 97124 USA; phy methods for the 3D tomography, NEMS, e-mail kamino-takeo@ tel. 321-663-3806, fax 503- Gerhard Hobler is an topological characteriza- and physics. nak.hitachi-hitec.com. 726-7509, and associate professor of tion of complex granular Inkson can be reached e-mail lucille.giannuzzi@ semiconductor electronics textures and porous net- at the Department of Joachim Mayer is a pro- fei.com. at the Vienna University works in cementitious Engineering Materials, fessor of physics at of Technology. He and ceramic materials. Mappin St., Sheffield, RWTH Aachen Univer- Jacques Gierak is respon- received his Dipl-Ing, Holzer can be reached S1 3JD, United Kingdom; sity, where he is head of sible for focused ion beam PhD, and venia docendi at 3D-Mat Group, Empa tel. 44-114-2225925, and the Central Facility for research activity at degrees from Vienna Materials Science and e-mail beverley. Electron Microscopy. He Laboratoire de University of Technology Technology, Dübendorf, [email protected]. studied physics at the Photonique et de in 1985, 1988, and 1997, Switzerland; tel. 41-44- University of Stuttgart Nanostructures (LPN) respectively. From 1996 to 823-44-90, fax 41-44- Takeo Kamino is techni- and received his PhD in Marcoussis, France. 1998, Hobler was a visit- 823-40-35, and e-mail cal advisor of the Naka degree in 1988. Mayer He graduated from the ing scientist at Bell [email protected]. Application Center at spent two years as a post- Physics–Electronics Laboratories, Lucent Hitachi High- doctoral researcher in the Department at Technologies. His current Beverley Inkson is a sen- Technologies Corp. in Materials Department at Conservatoire National research interests include ior lecturer and head of Japan. He received his the University of des Arts et Métiers in atomistic modeling of the NanoLAB group PhD degree in materials California, Santa Barbara. Paris, where he earned ion–solid interactions at the Department of science and engineering Afterward, he returned to his DEA and PhD degrees and thermal processes fol- Engineering Materials at Ibaraki University in Max-Planck-Institut für in instrumentation. lowing ion implantation. at the University of 1997. Kamino’s current Metallforschung in Gierak has been involved He has authored more Sheffield. Inkson work focuses on the Stuttgart. In 1999, Mayer in FIB research since 1984 than 100 papers in this received her MA degree development of a tech- joined RWTH Aachen and was the coordinator field. in physics and her PhD nique for 3D structure University. In addition to of the European Hobler can be reached degree in materials sci- observation of nanomate- his work at RWTH Commission–funded at the Institute of Solid ence from the University rials using a combination Aachen, Mayer is co- NanoFIB project. In 2004, State Electronics, Vienna of Cambridge. Her PhD of FIB and high-resolu- director of the newly Gierak was awarded the University of Technology, studies were followed by tion STEM or TEM. In founded Ernst Ruska CNRS Crystal, a prize for Floragasse 7/362, A-1040 an Alexander von addition, Kamino’s Center for Microscopy exceptional engineering Vienna, Austria; tel. Humboldt fellowship at research includes the and Spectroscopy with and technical achieve- 43-1-58801-36233, fax the Max Planck Institute development of TEM Electrons at the Research ment. He is the main 43-1-58801-36291, and in Stuttgart, Germany, techniques allowing Center Jülich, Germany. author of several interna- e-mail gerhard.hobler@ and a Royal Society observation of atoms Mayer can be reached tional CNRS patents tuwien.ac.at. research fellowship in during sintering, at the Central Facility for related to ion sources, ion nanomechanics at the reaction, and deposition. Electron Microscopy optics, and their control. Lorenz Holzer is a senior University of Oxford. He has been a member (GFE), RWTH Aachen Gierak can be reached scientist at Empa, the She is director of the of the board of directors University, Ahornstr. at Laboratoire de Federal Institute for U.K. NanoFIB Network of the Japanese Society of 55, D-52074 Aachen, Photonique et de Materials Testing and and director of the Microscopy since 2005. Germany; tel. 49-241-80- Nanostructures–CNRS, Research, in Dübendorf, RCUK Basic Technology Kamino can be reached 24345, fax 49-241-80- Route de Nozay, 91460 Switzerland. He is leader Programme in at Naka Application 22313, and e-mail mayer@ Marcoussis, France; tel. of the 3D-Mat Group, Nanorobotics. Her Center, Hitachi High- gfe.rwth-aachen.de.

MRS BULLETIN • VOLUME 32 • MAY 2007 • www/mrs.org/bulletin 397 Focused Ion Beam Microscopy and Micromachining

Takeo Kamino Joachim Mayer Richard M. Langford Joseph Michael Warren J. MoberlyChan

Richard M. Langford Recently, Michael has scopy, FIB, and spec- Munroe can be reached tions manager within the works in the Materials been interested in the troscopy to study their at the Department of Nanotechnology Systems Science Centre at the use of focused ion beam chemical and mechanical Materials Science and Division of Carl Zeiss University of Manchester. tools for the preparation properties. Engineering, University SMT. He received his PhD He received his PhD of samples for electron MoberlyChan can be of New South Wales, degree in engineering sci- degree from the Electrical backscatter diffraction, reached at Lawrence Sydney, NSW 2052 ence from the Engineering Department and for determining the Livermore National Australia; tel. 61-2- Pennsylvania State at Imperial College. For 3D structure of materials. Laboratory, MS L372, 9385-4435 and e-mail University with research the last decade, He has received the 7000 East Ave., [email protected]. in the development of Langford’s main Burton Medal from the Livermore, CA corrosion-resistant thin research interests have Microscopy Society of 94550–9234 USA; tel. Philipp M. Nellen works films. Before joining Carl been in using focused America and the 925-424-2721 and e-mail at RUAG Aerospace in Zeiss, Principe was a ion beams for nano- Heinrich Award from [email protected]. Wallisellen, Switzerland. member of technical staff engineering and nanofab- the Microbeam Analysis He received his PhD at Applied Materials, rication and developing Society. Michael also has Paul Munroe is a profes- degree in physics from working on process methods for sample published many papers sor of materials science ETH Zurich, where he development and materi- preparation for electron on the application of and engineering and also worked in the Optics als characterization microscopy. electron microscopy to director of the Electron Laboratory at the Institute across a broad range of Langford can be materials. Microscope Unit at the of Quantum Electronics semiconductor processes. reached at the Materials Michael can be reached University of New South on the subject of inte- He also worked as a staff Science Centre, University at Sandia National Wales in Sydney, grated optical chemo- and scientist at Charles Evans of Manchester, Laboratories, MS 0886, Australia. He received his biosensors. Afterward, he and Associates, specializ- Manchester, United PO Box 5800, PhD degree from the joined Empa Switzerland. ing in XPS and Auger Kingdom; tel. 44-161- Albuquerque, NM 87185 University of As a senior scientist at spectroscopy. His 3065914, fax 41-161- USA; tel. 505-844-9115 Birmingham in 1987, and Empa, his interests current work includes 3065912, and e-mail and e-mail jrmicha@ for his thesis topic he started in the field of the development of 3D richard.langford@ sandia.gov. characterized the fiber-optical sensor sys- metrology, automated manchester.ac.uk. microstructure of tems for materials and sample processing, and Warren J. MoberlyChan advanced titanium alloys. structures, and the relia- 3D nanofabrication. Joseph Michael is a dis- is a microscopist at His principal research bility of such systems. Principe received the tinguished member of Lawrence Livermore interests include Following this research, ISTA/EDFAS technical staff in the National Laboratory in structure–property rela- he worked in the field of Outstanding Paper Materials and Process the Chemistry, Materials, tionships in thin-film and FIB micro- and nanostruc- Award in 2005 and the Sciences Center at Sandia and Life Sciences coated materials, and in turing of photonic and Microscopy and Micro- National Laboratories. He Division. He received his the behavior of inter- other devices. analysis Best Paper received his BS, MS, and PhD degree in materials metallic alloys. Munroe The published work Award in 2003. In addi- PhD degrees in materials science and engineering received the Cowley– was done at Electronics/ tion, he has authored two science and engineering from Stanford University Moodie Award from the Metrology/Reliability text chapters on FIB from Lehigh University. in 1991, and his ScB Australian Society for Laboratory, Empa, Swiss applications and more Before joining Sandia, degree from Brown Microscopy and Federal Laboratories for than 20 journal publica- Michael was employed as University. MoberlyChan Microanalysis for Materials Testing and tions. a senior research engineer has worked in the infor- “outstanding physical sci- Research, Überlandstrasse Principe can be at Bethlehem Steel’s mation storage industry ences electron 129, CH-8600 Dübendorf, reached by e-mail at Homer Research at ReadRite, Komag, microscopy.” He has pub- Switzerland. Nellen can [email protected]. Laboratory. His research Quantum, MPI, and SSL. lished more than 160 jour- be reached by e-mail at interests involve the His research interests nal papers and serves on [email protected]. Thomas Schenkel is a application of advanced include thin-film inter- the editorial board of staff scientist in the electron microscopies to faces and surfaces and the Microscopy Research and Edward L. Principe is the Accelerator and Fusion the study of materials. application of micro- Technique. focused ion beam applica- Research Division of

398 MRS BULLETIN • VOLUME 32 • MAY 2007 • www/mrs.org/bulletin Focused Ion Beam Microscopy and Micromachining

Paul Munroe Philipp M. Nellen Edward L. Principe Thomas Schenkel Michael D. Uchic

Lawrence Berkeley Schenkel joined LBNL by e-mail at PhD degree from ment of new experimen- National Laboratory. He in 2000. His current [email protected]. Stanford University, tal methods to rapidly earned his PhD degree in research is focused on where he characterized assess both the micro- physics from the Johann the development and Michael D. Uchic is a the low-temperature structure and mechanical Wolfgang Goethe– testing of silicon-based materials research engi- mechanical properties of properties of aerospace University of Frankfurt quantum computer neer in the Metals the intermetallic alloy metals. am Main in 1997. schemes. Development Group of Ni3Al. Uchic started Uchic can be reached at Following his postdoc- Schenkel can be the Air Force Research working with the AFRL/MLLM, 2230 10th toral research in the reached at Lawrence Laboratory, Materials Metals Development St., Wright Patterson Chemistry and Materials Berkeley National and Manufacturing Group in 1998, and for AFB, OH 45433 USA; tel. Science Department at Laboratory, 1 Cyclotron Directorate, at Wright the past five years his 937-255-4784 and e-mail Lawrence Livermore Rd., 05R12, Berkeley, Patterson Air Force research efforts have michael.uchic@wpafb. National Laboratory, CA 94720 USA; and Base. He received his focused on the develop- af.mil. ■

MRS BULLETIN • VOLUME 32 • MAY 2007 • www/mrs.org/bulletin 399