Serial sectioning methods for 3D investigations in materials science

G Model

ARTICLE IN PRESS

JMIC 2052 1–13

Micron xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Micron journal homepage: www.elsevier.com/locate/micron

1

Review

2

Serial sectioning methods for 3D investigations in materials science

3 4 5

Q1

Armin Zankel ∗ , Julian Wagner, Peter Poelt Institute for Electron Microscopy and Nanoanalysis, Graz University of Technology & Graz Centre for Electron Microscopy, Steyrergasse 17, 8010 Graz, Austria

6

7 21

a r t i c l e

i n f o

a b s t r a c t

8 9 10 11 12 13

Article history: Received 30 December 2013 Received in revised form 4 March 2014 Accepted 4 March 2014 Available online xxx

14

20

Keywords: Tomography 3D SBEM AFM FIB

22

Contents

15 16 17 18 19

23 24 25

1. 2. 3.

26 27 28 29 30 31

4.

32 33 34 35 36 37

5.

38 39 40 41 42 43 44

6. 7.

A variety of methods for the investigation and 3D representation of the inner structure of materials has been developed. In this paper, techniques based on slice and view using scanning microscopy for imaging are presented and compared. Three different methods of serial sectioning combined with either scanning electron or scanning ion microscopy or atomic force microscopy (AFM) were placed under scrutiny: serial block-face scanning electron microscopy, which facilitates an ultramicrotome built into the chamber of a variable pressure scanning electron microscope; three-dimensional (3D) AFM, which combines an (cryo) ultramicrotome with an atomic force microscope, and 3D FIB, which delivers results by slicing with a focused ion beam. These three methods complement one another in many respects, e.g., in the type of materials that can be investigated, the resolution that can be obtained and the information that can be extracted from 3D reconstructions. A detailed review is given about preparation, the slice and view process itself, and the limitations of the methods and possible artifacts. Applications for each technique are also provided. © 2014 Published by Elsevier Ltd.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Methods used for serial sectioning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Serial block-face scanning electron microscopy (SBEM) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Sample preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Imaging with electrons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Imaging with X-rays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5. Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3D AFM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Setup (RT AFM tomography) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Sample preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3. Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4. Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5. Cryo conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3D FIB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1. Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2. Sample preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3. Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discussion and comparison of the methods SBEM, 3D AFM and 3D FIB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00

∗ Corresponding author. Tel.: +43 316 873 8832; fax: +43 316 873 108831. E-mail address: [email protected] (A. Zankel). http://dx.doi.org/10.1016/j.micron.2014.03.002 0968-4328/© 2014 Published by Elsevier Ltd.

Please cite this article in press as: Zankel, A., et al., Serial sectioning methods for 3D investigations in materials science. Micron (2014), http://dx.doi.org/10.1016/j.micron.2014.03.002

G Model JMIC 2052 1–13

A. Zankel et al. / Micron xxx (2014) xxx–xxx

2 45

46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72

ARTICLE IN PRESS

1. Introduction Microscopy classically delivers only two-dimensional (2D) images, without providing direct quantitative information about the three-dimensional (3D) internal structure of a material. However, in the life sciences and materials science this knowledge is extremely important, and has fostered interest in 3D reconstructions of all types of materials. The 3D morphology of biological objects determines their functionality (Denk and Horstmann, 2004). The properties of materials are strongly dependent on their design, including micro- and nanostructures. There are numerous examples of steric features such as anisotropic grains, precipitates, intergranular phases, crack distributions in deformation zones (Möbus and Inkson, 2007), for which a 3D analysis of the structure of the material is fundamental. There were several attempts early on in the history of microscopy to gather 3D models of microstructures. In 1876, Born published a 3D reconstruction of anatomical parts of amphibians based on serial sectioning combined with light microscopy (Born, 1876). Since that time, various microscopic techniques have been developed for both sectioning and imaging, additionally featuring an increase in microscopic resolution. The reader is referred to review articles such as (Midgley et al., 2007) concerning 3D methods by angular tomography. The discussion here will be restricted to 3D investigations using serial sectioning methods in combination with scanning microscopy techniques such as serial block-face scanning electron microscopy (SBEM), 3D atomic force microscopy (3D AFM) at room and cryogenic temperature and 3D focused ion beam (3D FIB). Fig. 1. 3D renderings of 56 axons tracked through 500 slices (voxelsize 26 × 26 × 50 nm3 , 900 × 500 × 500 voxels). (a) Entire volume and (b) close-up.

73

74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105

2. Methods used for serial sectioning Serial sectioning at the micro/nanoscale can be performed by microtomy, ion milling or consecutive polishing steps. Microtomy is an established sample preparation method that was initially invented for light and transmission electron microscopic applications, but is now also used for preparing specimens for scanning electron microscopy (SEM), ion microscopy with an FIB, and AFM. However, ultramicrotomy is restricted to soft materials and metals ductile enough to be cut by a diamond knife. Still, many soft materials need to undergo a series of preparative steps before they can be finally cut by a microtome. An ideal tool for serial sectioning of hard materials, albeit only for small volumes, is milling by FIB. For imaging, both electrons and ions can be used. However, milling needs much more time than cutting by a knife. Thus, the number of sections will be generally limited. Additionally, cutting and imaging generally happen at different orientations of the specimen surface. Therefore, specimen movements between cutting and imaging are necessary. Polishing and etching is typically applied to brittle and hard materials (e.g., ceramics, hard metals and alloys) where microtomy cannot be used, and larger areas and volumes are necessary than those that can be prepared by FIB. Here, material is removed from a block-shaped specimen with a thickness of typically 1 ␮m or more, enabling the analysis of large sample volumes of up to a few mm3 . This method is typically realized using light microscopy in reflection (e.g. Hopkins and Kraft, 1965; Lee et al., 2006), SEM (Mangan et al., 1997) and an electron probe microanalyzer (Marschallinger, 1998; Spowart, 2006). Furthermore, analytical information can be obtained using energy dispersive X-ray spectrometry (EDS) and electron backscatter diffraction (EBSD) (Yokomizo et al., 2003; Lewis and Geltmacher, 2006; Sharma et al., 2010). This particular technique, which lacks excellent depth resolution of the other methods, will not be discussed here.

Jurrus et al. (2009); with permission from Elsevier.

3. Serial block-face scanning electron microscopy (SBEM) This method was developed by Denk and published in a landmark paper in 2004 (Denk and Horstmann, 2004). It enables the reconstruction of the internal structures of soft materials over hundreds of microns providing 3D information with an imaging resolution typical for SEM. Thus, the gap between high-resolution tomography in a transmission electron microscope (TEM) and light microscopy is bridged. A prerequisite for this method is an electron microscope that enables imaging of electrically non-conductive samples without any coating, e.g. a variable-pressure SEM (VPSEM) or an environmental scanning electron microscope (ESEM) facilitating a low vacuum mode (Danilatos, 1988; Stokes, 2008). However, in special cases samples can also be investigated in high vacuum mode, e.g. metals that are soft enough to be sliced with a diamond knife (Mancuso et al., 2010) or biological samples if staining with heavy metals delivers sufficient intrinsic conductivity (Denk and Horstmann, 2004; Briggman et al., 2011). A combination of a conventional SEM and a microtome was already published in 1981 by Leighton (1981). However, this idea was not pursued because VPSEMs were not available on the market and computer capabilities were too low (Denk and Horstmann, 2004). Therefore, this method had to be redeveloped according to new technological possibilities in 2004, using VPSEM for the compensation of charges on the sample surface. Representative of the significant number of publications in the field of life sciences (Denk and Horstmann, 2004; Macke et al., 2008; Lang et al., 2011; Briggman et al., 2011; Helmstaedter and Mitra, 2012) a 3D reconstruction of 56 axons tracked through 500 slices (Jurrus et al., 2009) can be seen in Fig. 1. This method is now spreading in materials science for the investigation of polymers and composite

Please cite this article in press as: Zankel, A., et al., Serial sectioning methods for 3D investigations in materials science. Micron (2014), http://dx.doi.org/10.1016/j.micron.2014.03.002

106

107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135

G Model JMIC 2052 1–13

ARTICLE IN PRESS A. Zankel et al. / Micron xxx (2014) xxx–xxx

3

Fig. 2. Setup of SBEM. (a) Schematic of the ultramicrotome mounted on a special door on the ESEM. In addition to the backscattered electron detector (BSED), secondary electron detectors for high vacuum (ETD) and low vacuum (LFD) can be seen. Furthermore, an EDS detector can be used for recording elemental maps (slightly modified sketch from Zankel et al., 2009; with permission from Wiley). (b) Photograph showing the specimen holder (1) and the diamond knife (2). Zankel et al. (2009); with permission from Wiley.

148

materials (Zankel et al., 2009), membranes and foam-like structures (Reingruber et al., 2011; Muellner et al., 2012; Koku et al., 2011), materials including particles (Koch et al., 2012; Trueman et al., 2013) and metals (Zankel et al., 2011; Hashimoto et al., 2013; Thompson et al., 2013). Recently, the combination of a conventional high vacuum SEM with an ultramicrotome was realized, facilitating an automated in-chamber specimen coating. Here, the sample is coated after each cutting cycle with a 1–2 nm metallic film, using an electron beam evaporator integrated into a microscope chamber. This method eliminates charging effects and leads to a better signal-to-noise ratio (SNR) for BSE imaging under typical imaging conditions (e.g. 2 keV, 150 pA) compared with the conventionally used low-vacuum mode (Titze and Denk, 2013).

149

3.1. Principle

136 137 138 139 140 141 142 143 144 145 146 147

150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169

A thin slice of a specimen is cut off by a diamond knife, and then the block face of the sample is imaged. Afterwards, it is fed with a motor where cutting and imaging then repeats. As this process is automated, it can be repeated as often as necessary within a range of about 600 ␮m. The slices themselves are not imaged but accumulate at the back of the knife as debris. The series of images is typically well aligned, which eliminates one source of error, which was one main reason for establishing this method (Denk and Horstmann, 2004). The achieved slice thickness depends on the type of material, the preparation and the signal used for imaging. If the signal should originate from within a single slice only, the signal depth determines the z-resolution. The electron interaction volume and the depth from which backscattered electrons emerge are dependent on the electron energy and the density of the material and can be estimated by Monte Carlo simulations (Drouin et al., 2007). However, typical slice thicknesses published in the field of life sciences are between 25 and 50 nm (Rouquette et al., 2009; Briggman et al., 2011; Peddie and Collinson, 2014) and in the field of materials science at about 50 nm (Koku et al., 2011; Reingruber et al., 2011; Koch

et al., 2012; Trueman et al., 2013). Results down to 15 nm have been achieved by Hashimoto et al. (2013) and Thompson et al. (2013) cutting metal samples. In the context of EDS the electron energy cannot fall below the limit necessary to excite the respective Xray lines and thus the corresponding slice thickness may even be at around 100 nm (Zankel et al., 2011). While the newest generation of in situ ultramicrotomes are specified for a minimum slice thickness of 15 nm, a slice thicknesses of 10 nm was published by Mancuso et al. (2010). Section thickness control at the SBEM system is performed by the regulation of the servomotor which determines the feed rate in the z direction. The calculated precision of 1 nm of the z-value is determined by the precision of the motor and the gear ratio, but in literature unlike to 3D FIB (Jones et al., 2014) no measurement procedure for the evaluation of the actual slice thickness can be found. Useful specimens for such work could be resins with dispersed nanoparticles or sliceable metals with precipitates in the nanometer range. In Fig. 2, a schematic of the whole instrumentation can be seen. Contrary to a conventional microtome where the specimen is moved relative to a fixed knife, here the knife is moved parallel to the surface of the sample. Several detectors can be used for image recording: a backscattered electron detector (BSED), which is mounted at the end of the final lens, a secondary electron detector for the low vacuum mode (LFD, large field detector) and one for high vacuum mode (ETD, Everhart Thornley-detector). Additionally, energy dispersive X-ray spectrometry (EDS) can be performed. Due to mechanical cutting providing smooth surfaces, backscattered electrons are generally used for imaging (Denk and Horstmann, 2004) as they deliver compositional contrast (Goldstein et al., 2003). The recent systems on the market have an enhanced cutting concept, where the cutting direction of the knife is the same as the retracting direction of a modified knife holder (i.e. the knife is oriented 180◦ to the knife of the schematic in Fig. 2). This altered detail delivers the same cutting quality, but an even better removal of potential debris.

Please cite this article in press as: Zankel, A., et al., Serial sectioning methods for 3D investigations in materials science. Micron (2014), http://dx.doi.org/10.1016/j.micron.2014.03.002

170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205

G Model JMIC 2052 1–13

A. Zankel et al. / Micron xxx (2014) xxx–xxx

4 206

207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237

ARTICLE IN PRESS

3.2. Sample preparation A sample for SBEM should be sliceable but still hard enough to avoid smearing, and it should deliver enough contrast to image structures of interest. In addition, cutting performance should be stable during the automated process. Thus, samples are trimmed to an appropriate geometry, typically a cuboid of about 0.6 × 0.6 × 0.6 mm3 in size, which is glued onto a rivet using, e.g., superglue. If the sample has to be stabilized either to avoid rounding of the edges or because the structure of the specimen is not solid enough for a direct cutting (foams, thin foils, powders, etc.), it first needs to be embedded in resin. The choice of appropriate resin depends on the material and the preparation process. Polymers can typically be cut with a microtome, but smearing may occur at room temperature (RT). Hardening of the material can be achieved either by staining with e.g. ruthenium tetroxide, osmium tetroxide or uranyl acetate (Michler et al., 2004; Sawyer and Grubb, 1996), or by resorting to cryo-ultramicrotomy. On the one hand, SBEM can currently only be performed at RT. On the other hand, staining has the severe disadvantage that the diffusion depth of the stain is often very limited, giving rise to contrast changes during automated slicing and imaging. In some cases, specimen hardening can also be caused by electron irradiation itself, causing cross-linking of the molecules as was observed in the case of ethylene propylene rubber (EPR)-modified polypropylene (Zankel et al., 2009). If a material is sufficiently stiff, producing a rod-like sample may be feasible. From an aluminum sample, a rod with a length of about 10 mm and a diameter of 2 mm was carved (Zankel et al., 2011). At the top of the rod, a cylinder was prepared with a diameter of 3 mm and a height of 2 mm, which served as a basis of another cylinder with a diameter of 1 mm and a height of 0.5 mm. With the glass knife of a conventional microtome, the final cylinder was tailored

to a mesa shape with a square base of 0.5 × 0.5 mm2 . Finally, a block face with an area of about 200 × 300 ␮m2 was created using a 45◦ diamond knife (Diatome, Switzerland) in a conventional ultramicrotome. The sample was then mounted at the specimen holder of the microtome in the ESEM. Besides the stability of the sample during the serial cutting and imaging procedure, sufficient phase or topography (e.g. in case of cracks) contrast is a prerequisite for subsequent image processing. In the life sciences, heavy metal staining of tissues using techniques that are routine for TEM are used (Denk and Horstmann, 2004). These problems do not arise for composite materials with both organic and inorganic components such as paper, where mere stabilization of the specimen in resin may be sufficient (Zankel et al., 2009). In backscattered electron images, such materials show enough intrinsic compositional contrast because of large differences in the mean atomic number of different phases. 3.3. Imaging with electrons Because a smooth block face in general does not exhibit topographic features, backscattered electrons (BSE) are used for imaging in the majority of cases because the respective images show compositional contrast. Nevertheless, imaging with secondary electrons (SE) may be the choice for special materials. Muellner et al. (2012) used SE for imaging a resin-embedded hyper-crosslinked poly(styrene–divinylbenzene) monolith because the interior structure of the specimen showed topography but insufficient compositional contrast. In addition, an interface between a homogeneous polymer and wood could be imaged using SE (Zankel et al., 2009). Another example could be the 3D-representation of cracks inside a material, e.g. after a tensile test. The energy of primary electrons in ESEM determines both lateral and depth resolution (Thompson et al., 2013). Above all, the

Fig. 3. 3D reconstructions of different layers of the DuraPES® 450 microfiltration membrane. (a) Support layer, (b) intermediate layer, (c) separation layer and (d) air side of the membrane (section thickness 50 nm). Reingruber (2011); with permission from the author.

Please cite this article in press as: Zankel, A., et al., Serial sectioning methods for 3D investigations in materials science. Micron (2014), http://dx.doi.org/10.1016/j.micron.2014.03.002

238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253

254

255 256 257 258 259 260 261 262 263 264 265 266 267 268

G Model JMIC 2052 1–13

ARTICLE IN PRESS A. Zankel et al. / Micron xxx (2014) xxx–xxx

information depth should not be larger than the slice thickness. Otherwise, an image of the block face comprises features from at 271 least two slices. For the enhancement of resolution newly devel272 oped BSE detectors enable imaging even at or below an electron 273 energy of 1.0 keV. Simultaneous imaging with a secondary electron 274 detector for low vacuum (an LFD) may help to control charging 275 Q2 effects (Reingruber, 2011). In Hashimoto et al. (2013), imaging of 276 electrically conductive aluminum was performed at very low ener277 gies in high vacuum mode. 269

5

3.5. Applications

291

270

278

279 280 281 282 283 284 285 286 287 288 289 290

3.4. Imaging with X-rays Any signal available in an SEM can be used for imaging. One example is the combination of SBEM and energy dispersive X-ray spectrometry (EDS). After each cut, an elemental map of the block face is recorded. The resolution of the elemental maps is generally lower than that of BSE- or SE-images and again depends on the electron energy. However, in this case the electron energy cannot be freely chosen but has to be higher than the energy of the X-ray lines used for elemental mapping. With spectrum imaging, even a full X-ray spectrum can be recorded at every pixel. First results using a combination of SBEM and EDS were published in Zankel et al. (2011) delivering a 3D-reconstruction of the distribution of various precipitates in aluminum.

SBEM has been mainly used in the field of life sciences. However, it can be also a very valuable tool for investigations in materials science, as will be demonstrated at two examples. Microfiltration membranes and aluminum with precipitates. Microfiltration is an important process in beverage industry, wastewater treatment or in medical applications (Reingruber et al., 2011). Very desirable properties of membranes are a high flow rate and a low fouling rate, which are strongly dependent on the microstructure of the material. Because these parameters are determined by the 3D structure, SBEM may help to learn more about these parameters and further help to optimize products. Recent designs of microfiltration membranes feature strongly asymmetric pore structures, consisting of a very thin separation layer and a thick backup layer. The separation layer inside the membrane is responsible for the filtration process. The backupor support layer is necessary for mechanical stability and protects the separation layer from outside damage (Ulbricht et al., 2007; Ziel et al., 2008). In Fig. 3, the 3D reconstructions of different layers of the DuraPES® 450 membrane (Membrana GmbH, Wuppertal, Germany) are outlined showing the support layer (a), intermediate layer (b), separation layer (c) and the air side of the membrane (d). These four reconstructions are part of a series of systemtically reconstructed regions, which give the datum

Fig. 4. (a) SEM (BSE) image of a cross-section of the DuraPES® 450 membrane. (b) Pore diameter (the color bar gives the normalized probability of the occurrence of the pore diameters). (c) Mean pore size and pore size distribution as a function of the position at the cross-section of the respective membrane and (d) number of connections between pores as a function of the position at the cross-section. Reingruber et al. (2011), with permission from Elsevier.

Please cite this article in press as: Zankel, A., et al., Serial sectioning methods for 3D investigations in materials science. Micron (2014), http://dx.doi.org/10.1016/j.micron.2014.03.002

292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314

G Model JMIC 2052 1–13

ARTICLE IN PRESS A. Zankel et al. / Micron xxx (2014) xxx–xxx

6

342

points for the respective parameter profiles, which can be found in Fig. 4. SBEM can also be used for 3D-reconstruction of metal specimens if they are ductile enough to be sliceable by a diamond knife (Furneaux et al., 1978; Thompson et al., 2013). Mancuso et al. (2010) reported the first results of SBEM imaging of an aluminum specimen. Aluminum is important in many branches of industry, from the wrapping industry to aerospace applications, and so acquiring 3D information may be important for product optimization, monitoring of corrosion processes, crack formation and propagation and investigation of precipitate distributions to determine special properties. Hashimoto et al. (2013) used SBEM for the investigation of dealloying in an aluminum specimen during corrosion, revealing details of a copper-rich sponge-like structure. Zankel et al. (2011) reported the first results of investigations of an aluminum–copper alloy specimen of the type EN AW-2024 T351 (by AMAG, Ranshofen, Austria; supplied by Dr. Thomas Koch, Vienna University of Technology) using a combination of SBEM and EDS. EDS spectrum maps were recorded at an electron energy of 5 keV with a resolution of 512 × 416 pixels by use of an X-Max Silicon Drift Detector (SDD) from Oxford Instruments Analytical (Abingdon, UK) featuring an 80 mm2 active area. A stack of 200 slices with a thickness of 100 nm was cut, representing a volume of 42.7 × 34.5 × 20.0 ␮m3 . Quantification of elemental maps was performed with INCA software (Oxford Instruments) at a reduced resolution of 256 × 208 pixels. The region was 3D reconstructed, and the color-coded 3D elemental reconstruction can be found in Fig. 5.

343

4. 3D AFM

315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341

360

Another method for 3D reconstruction using serial sectioning is AFM combined with ultramicrotomy. The method can be applied for serial section tomography of typically biological and polymeric materials. There are two basic instrumentation setups that perform AFM 3D reconstruction both at RT and at cryo conditions. The RT instrument is automated and allows an acquisition of up to 10 sections per hour. The block face is imaged using a scanning probe microscope (SPM) NTEGRA (NT-MDT, Moscow, Russia). It allows a 3D reconstruction made of several tens of block face images. The thickness of the removed layer can vary between 10 and 2000 nm being measured by its interference color in water. The cryo AFM SNOTRA (SNOTRA, Moscow, Russia) is not automated and the scanning parameters in terms of acquisition time per image are considerably lower than at RT condition. However, cryo AFM is the only method that enables non-destructive serial sectioning tomography of soft materials compared to methods that would cause beam damage.

361

4.1. Setup (RT AFM tomography)

344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359

362 363 364 365 366 367 368 369 370 371 372 373 374 375

Here analogous to SBEM a sample is cut by a diamond knife and afterwards the block face is imaged by a scanning system. The technical implementation is slightly different from the former as is sketched in Fig. 6. Here a special AFM head was designed which is placed directly on the knife holder of an ultramicrotome (Leica UC6NT series). The freshly cut block face of the sample is imaged when the ultramicrotome arm is in the highest position. In Fig. 7 photographs of the AFM head in the position of AFM measuring and microtome operating can be seen. The region of cutting and imaging are laterally shifted contrary to SBEM where the lateral and the vertical position of the imaging and cutting plane are the same. Here, image alignment is an issue compared with SBEM. There is an x–y displacement of the relative position between probe and sample, which is determined by several factors:

Fig. 5. Color-coded 3D-reconstruction of the precipitates of a part of the sliced volume. The region numbers correspond to the colors shown in the legend (1: Mg2 Si, 2: Al2 CuMg). Phases of different elemental composition (Al2 Cu, Al3 Mn) in the inhomogeneous precipitate (3) can be distinguished by different colors. Zankel et al. (2011); Originally published in Imaging & Microscopy issue 2/2011, courtesy of Wiley-VCH.

mechanical repeatability of the system, positioning, thermal and other drifts and the relaxation of the block face after sectioning. The total misalignment of the images after each cycle reaches several microns. Efimov et al. (2007) solved this by using the post-processing software ImagePro Plus 5.1 by Media Cybernetics, Inc. (Rockville, MD, USA) with the options of automatic image alignment and visualization of 3D voxel images. A very important parameter in high-resolution 3D imaging via ultramicrotomy is the control of section thickness. Efimov et al. (2007) handled this with an established technique for the control of section thickness (Hayat, 2000), where interference colors of sections are measured. A systematic approach was realized by acquiring a video image from the optical microscope after each section and storing it in a computer. Because an AFM measurement takes a certain amount of time (e.g. 5–10 min), errors comparable with the thickness of a

Fig. 6. Schematic of the combination of an atomic force microscope (AFM) with an ultramicrotome. (a) Sample, (b) sample holder, (c) ultramicrotome arm, (d) ultramicrotome knife, (e) AFM scanner, (f) holder for AFM and (g) AFM probe. Efimov et al. (2007); with permission from Wiley.

Please cite this article in press as: Zankel, A., et al., Serial sectioning methods for 3D investigations in materials science. Micron (2014), http://dx.doi.org/10.1016/j.micron.2014.03.002

376 377 378 379 380 381 382 383 384 385 386 387 388 389 390

G Model JMIC 2052 1–13

ARTICLE IN PRESS A. Zankel et al. / Micron xxx (2014) xxx–xxx

7

Fig. 7. Head of the AFM NTEGRA Tomo. (a) Position of measuring and (b) position of sectioning. 1: AFM probe; 2: sample; 3: ultramicrotome knife; 4: spring locks; 5: support platform; 6: stepper motor for probe approach; 7: positioning microscrews; 8: polysapphire plate; 9: support pivots; 10: motorized rear screw; 11: ultramicrotome arm support. Efimov et al. (2007); with permission from Wiley.

392

slice (20–50 nm) can occur because of thermal drift and block face relaxation.

393

4.2. Sample preparation

391

Fig. 8. (a) 3D AFM reconstructed ABS/PA6 (Acrylonitrile-butadienestyrenestyrene/polyamide 6) polymer blend structure and (b) orthogonal slices of the ABS/PA6 for 3D visualization. Efimov et al. (2007); with permission from Wiley.

409

Preparation for this method is generally the same as that carried out for microtomy. The sample has to be fixed and embedded. However, contrary to SBEM no staining for the sake of contrast enhancement has to be done because AFM image contrast formation is based on variations of topography or electrical properties (Alekseev et al., 2009, 2012) and/or phase shift. In some cases, staining may be an option to enhance cutting stability at RT. However, Efimov et al. (2007) emphasized that the preservation of the natural state of specimens can be realized by this method. Furthermore, an interesting complementary idea was realized (Efimov et al., 2007) whereupon several slices of biological samples were cut and imaged without staining, and then the last slice underwent the conventional preparation procedure and was imaged with a TEM. This gives complementary information between AFM micrographs and TEM imaging. This could reveal different artifacts and decisively improve interpretations.

410

4.3. Imaging

394 395 396 397 398 399 400 401 402 403 404 405 406 407 408

411 412 413 414

Phase imaging is used to investigate ultrastructure of polymeric and biological samples. AFM phase contrast is generated by local variations of viscoelastic properties of the sample. For polymer samples, the AFM technique can be regarded as a non-destructive

method because no damage can be quoted compared with that due to electrons (Egerton et al., 2004). Furthermore, AFM tomography can work at different scanning regimes besides phase imaging: contact and semi-contact feedback-error modes, magnetic force microscopy (MFM) and several electrical measuring techniques enabling an appropriate choice of imaging capabilities. 4.4. Application In Fig. 8, the 3D reconstruction of an ABS/PA6 (Acrylonitrilebutadiene-styrene/polyamide 6) polymer blend structure is shown. Here, only conventional ultramicrotome trimming of the sample was done without any further special treatment. A series of phase images was recorded automatically. The average acquisition time per cycle for sectioning and imaging was around 6 min. Twentyfive frames were imaged with a resolution of 512 × 512 pixels. The thickness of a section was 40 nm. A 3D reconstruction performed with part of these data is shown where a resulting volume of 8.75 × 5.0 × 1.0 ␮m3 corresponding to 350 × 200 × 25 voxels is displayed. With this 3D representation, spherical clusters of ABS in the PA6 matrix are found. The smooth circular shape of these clusters in sections of the z–x and z–y planes of the reconstructed 3D volume additionally verifies proper alignment of the consecutive cuts (Efimov et al., 2007; Mittal and Matsko, 2012).

Please cite this article in press as: Zankel, A., et al., Serial sectioning methods for 3D investigations in materials science. Micron (2014), http://dx.doi.org/10.1016/j.micron.2014.03.002

415 416 417 418 419 420

421

422 423 424 425 426 427 428 429 430 431 432 433 434 435 436

G Model JMIC 2052 1–13

ARTICLE IN PRESS A. Zankel et al. / Micron xxx (2014) xxx–xxx

8

first cryo-sectioned and immediately afterwards imaged at −80 ◦ C (a). For comparison, the specimen was sectioned and scanned at RT (b). In (b), it is clearly visible that small PA6 domains are displaced from the matrix during sectioning at RT, whereas cryo conditions enable stabilization. Thus, by working at cryo temperatures a more authentic information about the distribution of the polymer phases in the specimen can be gained. Therefore, the 3D reconstructions in Fig. 10 were done by using images recorded with the cryo system (Efimov et al., 2012). 5. 3D FIB

Fig. 9. Schematic of the setup of the cryo-AFM SNOTRA. (a) Nickel support plate for the fixation of the specimen, (b) horizontally oriented piezotube scanner, (c) moveable carriage on a base platform with guide notches, (d) AFM probe and (e) diamond knife. Efimov et al. (2012); with permission from RSC Publishing.

437

438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478

4.5. Cryo conditions A further development of 3D AFM is the cryo AFM system SNOTRA (Efimov et al., 2012). Here, a specially designed cryo-AFM is mounted directly within the cryogenic chamber of an ultramicrotome. With this system, soft and hydrated materials can be analyzed after serial sectioning at RT or cryo conditions. Polymers can be cut below their glass transition temperature, thus staining to get hardening of the material is no longer necessary. This method can be used for the investigation of thermotropic dynamic processes in soft materials and for 3D tomography at different temperatures (−120 ◦ C to 50 ◦ C). The investigation can even be performed stepwise at different temperature levels, e.g., in the case of samples with different polymer phases of different Tg or specimens having temperature dependent dynamic structural changes (Efimov et al., 2012). As published in Efimov et al. (2012), this method uses chemically etched tungsten tips attached to a quartz tuning fork as an AFM probe, enabling several AFM measuring modes. The resolution of the cryo-AFM, which is mainly determined by the sharpness of the probe tip, lies between 5 and 10 nm. In Fig. 9, a schematic of the setup of the cryo-AFM, which replaces the conventional sample holder in a cryo-ultramicrotome is shown. The sample investigated is glued or frozen to a nickel support plate (a), which is attached magnetically to a horizontally oriented piezotube scanner (b) with a scan range of 50 × 50 × 5 ␮m3 at RT. The scanner is mounted on a base platform with guide notches on which a moveable carriage (c) with an AFM probe (d) is installed. The carriage may be moved along the guide notches to approach the sample. The base platform of the cryo-AFM unit has an opening below the sample for the diamond knife (e) of the cryo-ultramicrotome with which a sectioning of the sample can be made. Contrary to the former system, the AFM and the arm of the specimen remain motionless in x and y directions relative to each other, which delivers better image alignment down to a level below 1 ␮m (Efimov et al., 2012). Even hydrated samples, e.g. nanoliquids and biological objects, can be fixed by high-pressure freezing and then transferred to SNOTRA by a cryo transfer system. In Fig. 10, an example of a 3D reconstruction of a polyamide 6 (PA 6) and styrene-acrylonitrile (SAN) polymer blend can be seen basing on a series of AFM phase images acquired at −80 ◦ C. In Fig. 10(a) and (b), a comparison between topographical AFM images of this blend recorded both at RT and under cryo conditions is presented. The block face was

The previous two instruments are typically applicable for soft materials and specimens that are sufficiently ductile that they can be cut with a diamond knife. For most other materials, a dual beamfocused ion beam microscope (DB-FIB), which combines an SEM and a precisely focused ion beam, may be the best choice. With the ion beam milling of hard materials, serial sectioning can thus be performed. The advent of DB-FIB systems compared to a single focused ion beam instrument (Inkson et al., 2001) enables imaging and analyzing using secondary, backscattered and diffracted electrons and, furthermore, X-rays. Typical terms for this method, which is represented by the common principle of serial acquisition of 2D data from equidistant sections of a specimen’ volume, are “FIB tomography”, “FIB serial sectioning” and “3D-FIB analysis” (Schaffer, 2008). The use of electron beam induced SE/BSE imaging became a wide spread technique (Schaffer, 2008), enabling investigations of crack-forming mechanisms (Park et al., 2004), alloys (Bansal et al., 2006), and pores (Tomutsa and Silin, 2004). The first commercially available automated system for the combination of 3D FIB and electron backscatter diffraction (3D-FIB/EBSD) was published by Mulders and Fraser (2005), but later a great variety of such systems were also developed (Groeber et al., 2005; Konrad et al., 2006; Xu et al., 2007). The first use of EDS in 3D-FIB was published by Kotula et al. (2003) and the combination of SE/BSE imaging and EDS was published by Lifshin et al. (2004). The first fully automated system using X-rays (EDS) for imaging was developed by Schaffer et al. (2007), and enhancements of sample preparation were shown in (Schaffer et al., 2007) and (Schaffer and Wagner, 2008). Today 3D-FIB methods can be regarded as an established technique delivering new insight into the morphology, texture and chemism of a great variety of materials such as ceramics, metals, alloys, especially for the investigation of interphases in these materials (Lasagni et al., 2008; Holzer and Münch, 2009; Kotula, 2009; Anderson and McCarron, 2011; West and Thomson, 2009; Burdet et al., 2013). 5.1. Principle A specimen is placed at the eucentric point, where the electron beam and the ion beam in a conventional FIB system converge at an angle of 52◦ (Holzer et al., 2004). The ion beam removes material of the volume of interest slice by slice and after each cut the exposed surface is imaged. In Fig. 11 the principal geometry is shown. During milling the stage tilt is kept constant at 52◦ in order to realize an ion beam perpendicular to the original surface of the sample, which should be smooth (e.g. by mechanical pretreatment). The ion beam erodes a series of layers for imaging with electrons at an angle of 52◦ . If additionally EDS-mapping is performed, no shadowing of the EDS signal should happen, which requires a special preparation preceding the slice and view process. A slice and view process is realized by periodically milling with the ions, imaging with electrons and e.g. EDS mapping. Even if the aim is mere 3D elemental mapping a periodical imaging with electrons is required in order to get information for drift correction and differentiation between the bulk and potential voids (Schaffer et al., 2007).

Please cite this article in press as: Zankel, A., et al., Serial sectioning methods for 3D investigations in materials science. Micron (2014), http://dx.doi.org/10.1016/j.micron.2014.03.002

479 480 481 482 483 484 485 486 487

488

489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521

522

523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539

G Model JMIC 2052 1–13

ARTICLE IN PRESS A. Zankel et al. / Micron xxx (2014) xxx–xxx

9

Fig. 10. Topographical AFM images of the PA6/SAN polymer blend (30/70, w/w). (a) Section and scan at −80 ◦ C, (b) section and scan at RT. (The scale bars in (a) and (b) are 2000 nm.) 3D reconstructions of PA6/SAN obtained at −80 ◦ C, (c) dimensions of 7.9 × 6.2 × 0.75 ␮m3 and (d) dimensions of 2.0 × 2.0 × 0.75 ␮m3 . (6 sections each, 125 nm section thickness). Efimov et al. (2012); with permission from RSC Publishing.

540 541 542 543 544 545 546 547 548 549 550 551

The choice of the slice thickness depends on the one hand on the size of the structures that have to be reconstructed, on the other hand on the type of signal used for imaging. The signal depth ranges between around 5 nm for secondary electrons (SE), around 20 nm–5 ␮m for backscattered electrons (BSE) and around 50 nm to several ␮m for X-rays (EDS). For imaging of organic materials with secondary electrons an acceleration voltage of 2 kV can be used and a mean slice thickness of 7.8 nm is applicable and was performed by Maco et al. (2013). By use of an acceleration voltage of 1.5 kV and additional staining of the material (osmium tetroxide, lead) a slice thickness down to 3.6 nm is feasible (Wei et al., 2012). However if chemical characterization of a material is required (EDS),

e.g. the analysis of precipitates in an alloy, an acceleration voltage of at least 5 kV is advisable, resulting in a signal depth and thus a slice thickness larger than 50 nm. In literature slice thicknesses of 300 nm (Schaffer et al., 2007), 250 nm (Kotula, 2009), 100 nm (Uchic et al., 2006), 60 nm (Lasagni et al., 2008; Holzer et al., 2006), 50 nm (Munroe, 2009), 10 nm (Balach et al., 2012) and 5–8 nm (Gaiselmann et al., 2013) were reported. Especially using 3D elemental mapping a slice thickness according to the interaction volume is advised. A physically optimized combination was given by Burdet et al. (2013) where the slice thickness for SE imaging was 12.5 nm and an EDS map was acquired at every eighth slice. 5.2. Sample preparation

Fig. 11. Schematic of the slice and view process in the FIB. The ion beam erodes a series of layers for imaging with electrons at an angle of 52◦ . Additionally EDS mapping can be performed. A Pt layer protects the volume of interest. Preceding the slice and view process a U-pattern is prepared (Holzer et al., 2004; Schaffer et al., 2007).

Before serial sectioning and imaging with an FIB can be started, the volume of interest has to be separated. It is typically located below the surface of a flat sample. On the top of the volume of interest, a layer of platinum/carbon with a thickness of 1–2 ␮m has to be deposited. This can be realized by electron-beam deposition or by ion-beam deposition using a Pt-precursor gas of a gas injection system (GIS). The Pt-layer is important for surface protection of the volume of interest, and for the prevention of the rounding of top edges of the cross-sections during milling (Schaffer, 2008). After this pretreatment, the milling of a starting pattern is necessary because the region should be visible; shading of the signal used for imaging is prevented, and aspects such as redeposition of material have to be taken into account. In Kotula et al. (2003) this was firstly done by milling a trench to expose an initial analysis surface and then milling an additional perpendicular trench in order to avoid obstacles between the surface and the EDS detector. In Holzer et al. (2004) and Schaffer et al. (2007) a so called U-pattern is suggested, where the volume of interest is exposed by three trenches. In Fig. 11 a U-pattern can be seen, which minimizes the total time of the slice and view process, avoids problems caused by material redepositon and shadowing effects from the surrounding sample and is advantageous especially for an automated process. While the dimensions of the front and right side trenches are determined by the position of the EDS detector and the electron source relative to the cross-section, the left side trench serves as a gap to

Please cite this article in press as: Zankel, A., et al., Serial sectioning methods for 3D investigations in materials science. Micron (2014), http://dx.doi.org/10.1016/j.micron.2014.03.002

552 553 554 555 556 557 558 559 560 561 562

563

564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588

G Model JMIC 2052 1–13 10

ARTICLE IN PRESS A. Zankel et al. / Micron xxx (2014) xxx–xxx

Fig. 12. 3D-reconstructions of elemental distributions in a ceramic after the first fully automated experiment. (a) Calcium, (b) magnesium and (c) titanium. Schaffer et al. (2007); with permission from Elsevier.

612

prevent charging effects and material redeposition at the sidewall. The width of the left side trench has to be greater than a third of the intended slice height to avoid effects of redeposited material on the outcome (Schaffer et al., 2007). In Schaffer and Wagner (2008) a block lift-out sample preparation was published as a substantial improvement for automated 3D sectioning experiments in a dual beam FIB. As a first step a Pt protection layer is put onto the surface of the sample, which was deposited from a pre-cursor gas with the ion beam. Then several ion milling patterns are supplied in order to expose a block of the interesting part of the sample, which is still connected to the bulk by two bridges. As a final step the block is lifted out using an omniprobe micromanipulator and then transferred to a TEM specimen grid. The block lift out method definitely lasts longer than the previously described methods. For a sample volume of approximately 20 × 20 × 20 ␮m3 the U-pattern method needs about 3 h and the block lift out method about 6 h. However, block lift out offers a strong advantage with respect to EDS-measurements. While at the U-pattern the X-ray signal is partially overlaid by signal generated from the surrounding of redeposited material, which is hit by backscattered electrons, the block lift-out method avoids artificial signal, since no material is close to the region of investigation. This leads to EDS analysis with improved accuracy (Schaffer and Wagner, 2008).

613

5.3. Application

589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611

614 615 616 617 618 619 620

For the wide area of FIB tomography using electrons for imaging the reader is referred to articles as are listed in the introduction of this chapter. Here, the process of 3D FIB combined with EDS is discussed, which was published in Schaffer et al. (2007) as the first fully automated 3D elemental mapping in the FIB. Nowadays companies deliver fully automated systems and versatile user interfaces to establish slice and view in the FIB with enhanced voxel resolution

and shorter acquisition times of bigger volumes of interest compared to the years of development. But when the pioneering labs were dealing with the first experiments most of the crucial aspects were analyzed and largely solved (Kotula et al., 2003, 2006; Schaffer et al., 2007). The investigated sample was a CaMgTiOx ceramic with a MgTiO3 matrix containing grains of different chemistry (CaTiO3 , Mg2 TiO4 ) with a diameter of one to several microns. As a first step, the sample was mechanically polished and then coated with amorphous carbon to reduce charging of regions surrounding the volume of interest. Before slicing, a 1-␮mthick Pt layer was deposited and the appropriate trenches (U-pattern) were established according to the necessary geometry. FIB tomography was performed with an FEI Nanolab Nova200 dual-beam FIB (FEI Company, Eindhoven, The Netherlands) equipped with an energy dispersive Si(Li) X-ray detector (10 mm2 ) system from EDAX (Mahwah, USA). For BSE imaging, a standard through-the-lens detector (TLD-BSE) was used. Thirty slices (nominal slice thickness 300 nm) were cut by a Ga ion beam with a beam current of 3.0 nA at an energy of 30 keV (Schaffer et al., 2007). The scripting language of the Nova200 software (product version 1.06) was used for automation of the milling process. The Genesis software (version 3.60) for automated EDS acquisition was controlled and synchronized with FIB-milling software by use of the program AutoIT v3.1 (by J. Bennett). For imaging and mapping, an electron beam current of 6.2 nA (electron energy: 15 keV) was used. The EDS map size was 256 × 200 pixels (pixel dwell time 25 ms) leading to a total time per slice of about 1 h. As a result, a reconstructed volume of 12 × 12 × 9 ␮m3 was gained. After data correction using customized scripts within the DigitalMicrograph software (GATAN, Pleasanton, CA, USA), visualization was realized with Amira 3.1 software (Mercury Computer Systems) (Schaffer et al., 2007).

Please cite this article in press as: Zankel, A., et al., Serial sectioning methods for 3D investigations in materials science. Micron (2014), http://dx.doi.org/10.1016/j.micron.2014.03.002

621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653

G Model

ARTICLE IN PRESS

JMIC 2052 1–13

A. Zankel et al. / Micron xxx (2014) xxx–xxx

11

Table 1 Comparison of the three methods. 3D method

SBEM

3D AFM

3D FIB

Lateral resolution (x, y)a,b Slice thickness (z)b Time per slicec Sample volumeb Sample prerequisites

>1 nm >10 nm 0.5–5 min Up to 0.6 × 0.6 × 0.6 mm3 Mainly soft materials, suitable for microtome cutting Embedding, staining, block face preparation by conventional microtomy EDS Not realized yet

>0.1 nm >10 nm 5–60 min Up to 100 × 100 × 20 ␮m3 Mainly soft materials, suitable for microtome cutting Embedding, block face preparation by conventional microtomy Phase analysis possible Yes

>1 nm >2 nm 10–30 min Up to 50 × 50 × 50 ␮m3 Solid materials, suitable for ion milling U-pattern or block lift-out, Pt deposition

Sample preparation

Elemental analysis Cryo a b c

654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671

672 673

674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700

EDS, EBSD Yes

The values depend on the material. Values from (Schaffer, 2008). i.e. cutting and imaging, estimated typical values.

Fig. 12 shows the 3D distributions of different elements of the ceramic (a: calcium, b: magnesium c: titanium) (Schaffer et al., 2007). According to Schaffer (2008) these phases can be correlated to CaTiO3 , Mg2 TiO4 and MgTiO3 respectively. However, several fundamental problems are discussed and solved in Schaffer et al. (2007) like the redeposition of eroded material and shadowing effects due to surrounding material, which could be handled by applying a U-pattern preparation. Because of the long acquisition time of the experiment (32 h) drift effects during both milling and analysis are an issue. Since the standard online drift correction is not accurate enough especially for electrically non-conductive samples and necessary high current ion milling a post-acquisition drift correction method was provided, which was realized by intermediate BSE imaging in an “overview” and a “detail” scope. Furthermore the determination of voids in a porous material was enhanced or even realized by a combination of BSE imaging and EDS maps since EDS maps alone may be erroneous since X-rays may have their origin of deeper regions exposed by voids (Schaffer et al., 2007). 6. Discussion and comparison of the methods SBEM, 3D AFM and 3D FIB As shown in Table 1, the three methods presented here differ especially in lateral resolution and sample volume that can be processed. SBEM and 3D AFM are very similar as far as serial sectioning (cutting with a diamond knife) is concerned. How serial sectioning is performed determines the possible slice thickness, which in both cases can be quoted as greater 10 nm. Concerning SBEM Mancuso et al. (2010) used a slice thickness of 10 nm and Thompson et al. (2013) 15 nm. In Alekseev et al. (2009, 2012) 3D AFM applications on nanocarbon/polymer composites with 12 nm section thickness are reported which were achieved by the use of an oscillating diamond knife. Both lateral and depth resolution are determined by the signal used for imaging, but AFM definitely provides in both cases the best resolution. The instrument used for sectioning and imaging also determines which materials can be processed with the respective method. The use of a microtome requires samples that can be cut by a diamond knife, so mainly soft materials and materials sufficiently ductile for cutting (e.g. aluminum) can be used. With an FIB, especially if equipped with a cryo stage, nearly all types of materials can be investigated. The advantages of sectioning soft materials with a diamond knife are the larger area of the block face that can be processed, and the lack of both irradiation damage and spurious materials. By milling with an FIB, not only ions are implanted in the material, but also selective sputtering takes place, thus changing the chemical composition at the block face. Additionally, amorphization of a crystalline material can take place and stress can be induced, which is especially important in combination with 3D

EBSD. The SBEM method has the great advantage that, in general, all images that were recorded are well aligned. Only in case of strong specimen drift a correction is necessary. Sample preparation for SBEM and 3D AFM involves fixation and embedding in resin and further tailoring of typical microtomy geometry. For SBEM, often staining of the specimen has to be performed to attain phase contrast. It is possible to proceed without staining in cases where either the different phases contain different inorganic elements, e.g. inorganic filler particles in polymers or paper, or where morphological details are imaged, e.g. cracks in a material. For biological samples, conventional preparation procedures for TEM samples, including fixation and staining, can be used. For imaging with cryo-AFM, most of these steps are not necessary. For polymers, a temperature below the glass transition temperature can be chosen. In biological specimens after high-pressure freezing, the water remains in a frozen amorphous state in the specimen. The different phases are resolved by intrinsic properties of the materials, e.g. the difference in elastic parameters. This is a great advantage of this method because changes of the inner structure and composition of the specimen can be avoided. Furthermore, different scanning regimes besides phase imaging can be facilitated like contact and semi-contact feedback-error modes, MFM and several electrical measuring techniques. Cooling of samples may also be necessary to avoid smearing during slicing with a diamond knife, or to prevent irradiation damage in case of milling with a FIB. However, for SBEM a cryo-stage for specimen cooling has not yet been developed. Additional analytical methods can be applied with a 3D FIB using EBSD and EDS. Results gained with the combination of SBEM and EDS were presented in this paper and in principal the combination of SBEM with EBSD would be possible, but samples that deliver channeling contrast and are sliceable by a diamond knife are rather rare. 7. Conclusions With the methods SBEM, 3D AFM and 3D FIB a 3D representation of the inner morphology of a material can be gained with high resolution. In images recorded of the block face, the resolution is limited by resolution of the microscope. In the direction perpendicular to the block face, the resolution is limited by minimum slice thickness and generally is lower than the resolution in the image plane. This may have to be taken into account if a material is not isotropic, but its inner structures have a preferential direction. All three techniques, which were developed in the first decade of our century, provide correlations between the microstructure of a material, the distribution of chemical and crystallographic phases and the material properties. In case of porous materials, parameters such as volume porosity, specific surface area and tortuosity can be calculated. The respective results can also help in the design

Please cite this article in press as: Zankel, A., et al., Serial sectioning methods for 3D investigations in materials science. Micron (2014), http://dx.doi.org/10.1016/j.micron.2014.03.002

701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732

733

734 735 736 737 738 739 740 741 742 743 744 745 746 747

G Model JMIC 2052 1–13 12 748 749 750

751

752 753 754 755 756

757

758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 Q3 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823

ARTICLE IN PRESS A. Zankel et al. / Micron xxx (2014) xxx–xxx

and development of new materials with tailored properties. Thus the described techniques have the ability to contribute to materials science and technological progress.

Acknowledgements The authors would like to thank Prof. Dr. Ferdinand Hofer and Dr. Nadejda B. Matsko for fruitful suggestions. We are grateful to Manuel Paller for generating the schematics of SBEM and FIB. This work was financially supported by Austrian Cooperative Research (ACR), Vienna.

References Alekseev, A., Efimov, A., Lu, K., Loos, J., 2009. Three-dimensional electrical property mapping with nanometer resolution. Adv. Mater. 21, 4915–4919. Alekseev, A., Chen, D., Tkalya, E.E., Ghislandi, M.G., Syurik, Y., Ageev, O., Loos, J., de With, G., 2012. Local organization of graphene network inside graphene/polymer composites. Adv. Funct. Mater. 22, 1311–1318. Anderson, A.J., McCarron, T., 2011. Three-dimensional textural and chemical characterization of polyphase inclusions in spodumene using a dual focused ion beam – scanning electron microscope (FIB-SEM). Can. Mineral. 49, 541–553. Balach, J., Miguel, F., Soldera, F., Acevedo, D.F., Mücklich, F., Barbero, C.A., 2012. A direct and quantitative image of the internal nanostructure of nonordered porous monolithic carbon using FIB nanotomography. J. Microsc. 246 (3), 274–278. Bansal, R.K., Kubis, A., Hull, R., Fitz-Gerald, J.M., 2006. High-resolution threedimensional reconstruction: a combined scanning electron microscope and focused ion-beam approach. J. Vac. Sci. Technol. B 24 (2), 554–561. Born, G., 1876. Ueber die Nasenhöhlen und den Thränennasengang der Amphibien. Morph. Jb 2, 577–646. Briggman, K.L., Helmstaedter, M., Denk, W., 2011. Wiring specificity in the directionselectivity circuit of the retina. Nature 471, 183–188. Burdet, P., Vannod, J., Hessler-Wyser, A., Rappaz, M., Cantoni, M., 2013. Threedimensional chemical analysis of laser-welded NiTi–stainless steel wires using a dual-beam FIB. Acta Mater. 61, 3090–3098. Danilatos, G.D., 1988. Foundations of environmental scanning electron microscopy. Advances in Electronics and Electron Physics, vol. 71. Academic Press, pp. 109–250. Denk, W., Horstmann, H., 2004. Serial block-face scanning electron microscopy to reconstruct three-dimensional tissue nanostructure. PLoS Biol. 2 (11), 1900–1909. Drouin, D., Réal Couture, A., Joly, D., Tastet, X., Aimez, V., Gauvin, R., 2007. CASINO V2.42: a fast and easy-to-use modeling tool for scanning electron microscopy and microanalysis users. Scanning 29, 92–101. Efimov, A.E., Tonevitsky, A.G., Dittrich, M., Matsko, N.B., 2007. Atomic force microscope (AFM) combined with the ultramicrotome: a novel device for the serial section tomography and AFM/TEM complementary structural analysis of biological and polymer samples. J. Microsc. 226 (Pt 3), 207–217. Efimov, A.E., Gnaegi, H., Schaller, R., Grogger, W., Hofer, F., Matsko, N.B., 2012. Analysis of native structures of soft materials by cryo scanning probe tomography. Soft Matter 8, 9756–9760. Egerton, R.F., Li, P., Malac, M., 2004. Radiation damage in the TEM and SEM. Micron 35, 399–409. Furneaux, R.C., Thompson, G.E., Wood, G.C., 1978. The application of ultramicrotomy to the electronoptical examination of surface films on aluminium. Corros. Sci. 18, 853–881. Gaiselmann, G., Neumann, M., Holzer, L., Hocker, T., Prestat, M.R., Schmidt, V., 2013. Stochastic 3D modeling of La0.6Sr0.4CoO 3-␦ cathodes based on structural segmentation of FIB-SEM images. Comput. Mater. Sci. 67, 48–62. Goldstein, J.I., Newbury, D.E., Echlin, P., Joy, D.C., Lyman, C.E., Lifshin, E., Sawyer, L., Michael, J.R., 2003. Scanning Electron Microscopy and X-ray Microanalysis, third ed. Kluwer Academic/Plenum Publishers, New York. Groeber, M., Bhandari, Y., Uchic, M.D., Dimiduk, D.M., Ghosh, S., 2005. Quantitative characterization of grain structure and orientation using electron back-scattered diffraction patterns collected by serial sectioning. Microsc. Microanal. 11 (Suppl. 2), 1626–1627. Hashimoto, T., Curioni, M., Zhou, X., Mancuso, J., Skeldon, P., Thompson, G.E., 2013. Investigation of dealloying by ultra-high-resolution nanotomography. Surf. Interface Anal. 45, 1548–1552. Hayat, M.A., 2000. Principles and Techniques of Electron Microscopy: Biological Applications, fourth ed. Cambridge University Press, New Jersey. Helmstaedter, M., Mitra, P.P., 2012. Computational methods and challenges for large-scale circuit mapping. Curr. Opin. Neurobiol. 22, 162–169. Holzer, L., Indutnyi, F., Gasser, P.H., Münch, B., Wegmann, M., 2004. Threedimensional analysis of porous BaTiO3 ceramics using FIB nanotomography. J. Microsc. 216 (Pt 1), 84–95. Holzer, L., Muench, B., Wegmann, M., Gasser, P., Flatt, R.J., 2006. FIB-nanotomography of particulate systems – part I: particle shape and topology of interfaces. J. Am. Ceram. Soc. 89 (8), 2577–2585.

Holzer, L., Münch, B., 2009. Toward reproducible three-dimensional microstructure analysis of granular materials and complex suspensions. Microsc. Microanal. 15, 130–146. Hopkins, R.H., Kraft, R.W., 1965. A rapid technique for observation of threedimensional microstructures: application to analysis of fault structure in eutectic alloy. Trans. Metall. Soc. AIME 233, 1526–1532. Inkson, B.J., Mulvihill, M., Möbus, G., 2001. 3D determination of grain shape in a FeAl-based nanocomposite by 3D FIB tomography. Scripta Mater. 45, 753–758. Jones, H.G., Mingard, K.P., Cox, D.C., 2014. Investigation of slice thickness and shape milled by a focused ion beam for three-dimensional reconstruction of microstructures. Ultramicroscopy 139, 20–28. Jurrus, E., Hardy, M., Tasdizen, T., Fletcher, P.T., Koshevoy, P., Chien, C., Denk, W., Whitaker, R., 2009. Axon tracking in serial block-face scanning electron microscopy. Med. Image Anal. 13, 180–188. Koch, T., Salaberger, D., Zankel, A., Reingruber, H., Steiger-Thirsfeld, A., Voronko, Y., Seidler, S., 2012. Methods for characterizing the 3-D morphology of polymer composites. Macromol. Symp. 315, 115–124. Koku, H., Maier, R.S., Czymmek, K.J., Schure, M.R., Lenhoff, A.M., 2011. Modeling of flow in a polymeric chromatographic monolith. J. Chromatogr. A 1218, 3466–3475. Konrad, J., Zaefferer, S., Raabe, D., 2006. Investigation of orientation gradients around a hard Laves particle in a warm-rolled Fe3 Al-based alloy using a 3D EBSD-FIB technique. Acta Mater. 54, 1369–1380. Kotula, P.G., Keenan, M.R., Michael, J.R., 2003. Tomographic spectral imaging: comprehensive 3D X-ray microanalysis. Microsc. Microanal. 9 (Suppl. 2), 1004–1005. Kotula, P.G., Keenan, M.R., Michael, J.R., 2006. Tomographic spectral imaging with multivariate statistical analysis: comprehensive 3D microanalysis. Microsc. Microanal. 12, 36–48. Kotula, P.G., 2009. FIB-SEM 3D spectral imaging for materials characterization. Microsc. Microanal. 15 (Suppl. 2), 604–605. Lang, S., Drouvelis, P., Tafaj, E., Bastian, P., Sakmann, B., 2011. Fast extraction of neuron morphologies from large-scale SBFSEM image stacks. J. Comput. Neurosci. 31, 533–545. Lasagni, F., Lasagni, A., Engstler, M., Degischer, H.P., Mücklich, F., 2008. Nanocharacterization of cast structures by FIB-tomography. Adv. Eng. Mater. 10 (1-2), 62–66. Lee, S.G., Gokhale, A.M., Sreeranganathan, A., 2006. Reconstruction and visualization of complex 3D pore morphologies in a high-pressure die-cast magnesium alloy. Mater. Sci. Eng. A 427, 92–98. Leighton, S.B., 1981. SEM images of block faces, cut by a miniature microtome within the SEM-A technical note. Scan. Electron Microsc. 2, 73–76. Lewis, A.C., Geltmacher, A.B., 2006. Image-based modeling of the response of experimental 3D microstructures to mechanical loading. Scripta Mater. 55, 81–85. Lifshin, E., Evertsen, J., Principe, E., Friel, J., 2004. Three dimensional imaging of microelectronic devices using a crossbeam FIB. In: Proceedings from the 30th International Symposium for Testing and Failure Analysis, pp. 429–435. Maco, B., Holtmaat, A., Cantoni, M., Kreshuk, A., Straehle, C.N., Hamprecht, F.A., Knott, G.W., 2013. Correlative in vivo 2 photon and focused ion beam scanning electron microscopy of cortical neurons. PLoS ONE 8 (2), e57405, http://dx.doi.org/10.1371/journal.pone.0057405. Macke, J.H., Maack, N., Gupta, R., Denk, W., Schoelkopf, B., Borst, A., 2008. Contourpropagation algorithms for semi-automated reconstruction of neural processes. J. Neurosci. Methods 167, 349–357. Mancuso, J., Hashimoto, T., Thompson, G., 2010. Serial block face imaging of aluminium. Microsc. Microanal. 16 (Suppl. 2), 1864–1865. Mangan, M.A., Lauren, P.D., Shiflet, G.J., 1997. Three-dimensional reconstruction of Widmanstätten plates in Fe–12.3Mn–0.8C. J. Microsc. 188 (Pt 1), 36–41. Marschallinger, R., 1998. Three-dimensional reconstruction and modeling of microstructures and microchemistry in geological materials. Scanning 20, 65–73. Michler, G.H., Lebek, W., Godehardt, R., Galetzka, W., Gnaegi, H., Vastenhout, B., 2004. Ultramikrotomie in der Materialforschung. Carl Hanser Verlag, Munich. Midgley, P.A., Ward, E.P.W., Hungría, A.B., Thomas, J.M., 2007. Nanotomography in the chemical, biological and materials sciences. Chem. Soc. Rev. 36, 1477–1494. Mittal, V., Matsko, N.B., 2012. Analytical Imaging Techniques for Soft Matter Characterization. Springer, Berlin, Heidelberg. Möbus, G., Inkson, B.J., 2007. Nanoscale tomography in materials science. Mater. Today 10, 18–25. Muellner, T., Zankel, A., Mayrhofer, C., Reingruber, H., Höltzel, A., Lv, Y., Svec, F., Tallarek, U., 2012. Reconstruction and characterization of a polymerbased monolithic stationary phase using serial block-face scanning electron microscopy. Langmuir 28, 16733–16737. Mulders, J.J.L., Fraser, H.L., 2005. Automated three-dimensional EBSD analysis of materials. Microsc. Microanal. 11 (Suppl. 2), 506–507. Munroe, P.R., 2009. The application of focused ion beam microscopy in the material sciences. Mater. Charact. 60 (1), 2–13. Peddie, C.J., Collinson, L.M., 2014. Exploring the third dimension: volume electron microscopy comes of age. Micron, http://dx.doi.org/10.1016/j. micron.2014.01.009. Park, K.H., Kishimoto, H., Kohyama, A., 2004. 3D analysis of cracking behaviour under indentation in ion-irradiated ␤-SiC. J. Electron Microsc. 53 (5), 511–513. Reingruber, H., Zankel, A., Mayrhofer, C., Poelt, P., 2011. Quantitative characterization of microfiltration membranes by 3D reconstruction. J. Membr. Sci. 372, 66–74. Reingruber, H., 2012. Ph.D. thesis, University of Technology, Graz. Q4

Please cite this article in press as: Zankel, A., et al., Serial sectioning methods for 3D investigations in materials science. Micron (2014), http://dx.doi.org/10.1016/j.micron.2014.03.002

824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909

G Model JMIC 2052 1–13

ARTICLE IN PRESS A. Zankel et al. / Micron xxx (2014) xxx–xxx

910 911 912 913 914 915 916 917 918 919 920 921 922 Q5 923 924 925 926 927 Q6 928 929 930 931 932 933 934 935 936 937

Rouquette, J., Genoud, C., Vazquez-Nin, G.H., Kraus, B., Cremer, T., Fakan, S., 2009. Revealing the high-resolution three-dimensional network of chromatin and interchromatin space: a novel electron-microscopic approach to reconstructing nuclear architecture. Chromosome Res. 17, 801–810. Sawyer, L.C., Grubb, D.T., 1996. Polymer Microscopy, second ed. Chapman and Hall, London. Schaffer, M., Wagner, J., Schaffer, B., Schmied, M., Mulders, H., 2007. Automated three-dimensional X-ray analysis using a dual-beam FIB. Ultramicroscopy 107, 587–597. Schaffer, M., Wagner, J., 2008. Block lift-out sample preparation for 3D experiments in a dual beam focused ion beam microscope. Microchim. Acta 161, 421–425. Schaffer, M., 2008. Ph.D. thesis, University of Technology, Graz. Sharma, H., van Bohemen, S.M.C., Petrov, R.H., Sietsma, J., 2010. Three-dimensional analysis of microstructures in titanium. Acta Mater. 58, 2399–2407. Spowart, J.E., 2006. Automated serial sectioning for 3-D analysis of microstructures. Scripta Mater. 55, 5–10. Stokes, D.J., 2008. Principles and Practice of Variable Pressure/Environmental Scanning Electron Microscopy (VP–ESEM). John Wiley & Sons Ltd. Thompson, G.E., Hashimoto, T., Zhong, X.L., Curioni, M., Zhou, X., Skeldon, P., Withers, P.J., Carr, J.A., Monteith, A.G., 2013. Revealing the three dimensional internal structure of aluminium alloys. Surf. Interface Anal. 45, 1536–1542. Titze, B., Denk, W., 2013. Automated in-chamber specimen coating for serial blockface electron microscopy. J. Microsc. 250 (Pt 2), 101–110. Tomutsa, L., Silin, D., 2004. Nanoscale pore imaging and pore scale fluid flow modeling in chalk. In: Proceedings of the Conference: 25th Annual Workshop and Symposium Collaborative Project on Enhanced Oil Recovery, pp. 1–15. Trueman, A., Knight, S., Colwell, J., Hashimoto, T., Carr, J., Skeldon, P., Thompson, G., 2013. 3-D tomography by automated in situ block face ultramicrotome imaging

13

using an FEG-SEM to study complex corrosion protective paint coatings. Corros. Sci. 75, 376–385. Uchic, M.D., Groeber, M.A., Dimiduk, D.M., Simmons, J.P., 2006. 3D microstructural characterization of nickel supperalloys via serial-sectioning using a dual beam FIB-SEM. Scripta Mater. 55, 23–28. Ulbricht, M., Schuster, O., Ansorge, W., Ruetering, M., Steiger, P., 2007. Influence of the strongly anisotropic cross-section morphology of a novel polyethersulfone microfiltration membrane on filtration performance. Sep. Purif. Technol. 57, 63–73. Wei, D., Jacobs, S., Modla, S., Zhang, S., Young, C.L., Cirino, R., Caplan, J., Czymmek, K., 2012. High-resolution three-dimensional reconstruction of a whole yeast cell using focused-ion beam scanning electron microscopy. Biotechniques 53, 41–48. West, G.D., Thomson, R.C., 2009. Combined EBSD/EDS tomography in a dual-beam FIB/FEG-SEM. J. Microsc. 233 (Pt 3), 442–450. Xu, W., Ferry, M., Mateescu, N., Cairney, J.M., Humphreys, F.J., 2007. Techniques for generating 3-D EBSD microstructures by FIB tomography. Mater. Charact. 58, 961–967. Yokomizo, T., Enomoto, M., Umezawa, O., Spanos, G., Rosenberg, R.O., 2003. Threedimensional distribution, morphology, and nucleation site of intragranular ferrite formed in association with inclusions. Mater. Sci. Eng. A344, 261–267. Zankel, A., Kraus, B., Poelt, P., Schaffer, M., Ingolic, E., 2009. Ultramicrotomy in the ESEM, a versatile method for materials and life sciences. J. Microsc. 233 (Pt 1), 140–148. Zankel, A., Reingruber, H., Schroettner, H., 2011. 3D elemental mapping in the ESEM. Imaging & Microscopy 2, 35–37. Ziel, R., Haus, A., Tulke, A., 2008. Quantification of the pore size distribution (porosity profiles) in microfiltration membranes by SEM, TEM and computer image analysis. J. Membr. Sci. 323, 241–246.

Please cite this article in press as: Zankel, A., et al., Serial sectioning methods for 3D investigations in materials science. Micron (2014), http://dx.doi.org/10.1016/j.micron.2014.03.002

938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966