Focused ion beam scanning electron microscopy in biology
Summary
Since the end of the last millennium, focused ion beam scanning electron microscopy (FIB-SEM) has progressively found use in biological research. This instrument is a scanning electron microscope (SEM) with an attached gallium ion column and the 2 beams, electrons and ions (FIB) are focused on one coincident point. The main application is the acquisition of three-dimensional data, FIB-SEM tomography. With the ion beam, some nanometres of the surface are removed and the remaining block-face is imaged with the electron beam in a repetitive manner. The instrument can also be used to cut open biological structures to get access to internal structures or to prepare thin lamella for imaging by (cryo-) transmission electron microscopy. Here, we will present an overview of the development of FIB-SEM and discuss a few points about sample preparation and imaging.
Introduction
The first experiments with a focussed ion beam were reported in 1974 by Seliger and Fleming (Seliger & Fleming, 1974) who were able to focus a beam of boron ions to about 3.5 micrometers. A number of different ions were tested and the liquid gallium turned out to be the most suited (reviewed in Sugiyama & Sigesato, 2004). Further developments in the FIB technology were pioneered by the group of Dr. Ishitani at Hitachi High-Technologies Corporation (Ishitani et al., 1995; Ishitani & Yaguchi, 1996). The main application of the instrument was sample preparation, preparing thin lamella, for transmission electron microscopy (TEM) for defect analysis, especially in the semi-conductor industry (Overwijk et al., 1993; Ishitani & Yaguchi, 1996; Giannuzzi et al., 1998; Giannuzzi & Stevie, 1999; McGuinness, 2003).
Later the FIB column was integrated into a scanning electron microscope. It is inclined at a specific angle in the range of 52 to 55 degrees depending on the manufacturer (Fig. 1). An SEM with a built-in gallium ion source is called a focused ion beam scanning electron microscope, FIB-SEM. Thus the sample can be milled with Ga+ ions and imaged by electrons in the same instrument, opening the door to FIB-SEM tomography and controlled micromachining (Yonehara et al., 1989).
FIB for lamella for difficult-to-cut biological specimen
The first single beam instruments were used in the semi-conductor industry for defect analysis. A thin lamella is cut out at a defined location from a wafer, thinned and then analysed with a transmission electron microscope. With the introduction of the FIB-SEM instruments, the process could be better monitored with the electron beam. Early biological applications of the FIB-SEM were to prepare TEM lamella from material that is difficult or impossible to cut with a diamond knife in an ultramicrotome, e.g. titan implants in bone (Giannuzzi et al., 2007) or dentin (Nalla et al., 2005). With the electron beam, using secondary or backscattered electrons, the area of interest is identified. Then with the ion beam a lamella of that area is cut and transferred onto a special grid where the lamella is attached with platinum deposition. On the grid, the lamella is further thinned in the ion beam until it is electron transparent. The grid is transferred from the FIB-SEM to a TEM to analyse the lamella at high resolution.
Cryo-sectioning in the FIB-SEM
Cryo-electron microscopy of vitrified sections (CEMOVIS; Frederik et al., 1982; Zierold, 1982; Frederik et al., 1984; Dubochet & McDowall, 1984; Al-Amoudi et al., 2004) is becoming more and more popular for imaging biological structures in their natural, water-containing surroundings. Next to the cryo-fixation process, the limitations of CEMOVIS are cutting artefacts such as compression of the section, crevasses and a limited thickness of about 100 nm (Al-Amoudi et al., 2005). Thin sections of 100 nm are perfectly suited for high-resolution analysis of cellular proteins, e.g. the molecular structure of cadherins in desmosomes (Al-Amoudi et al., 2007) but for meaningful cryo-tomography thicker sections of up to 500 nm would be preferred. Here the focussed ion beam might provide an answer. As a first approach, Michael Marko thinned bacteria, plunge-frozen on an electron microscope grid, with the ion beam to a thickness of about 500 nm (Marko et al., 2006; Marko et al., 2007). This technology has been and still is being perfected at the Max-Planck-Institute in Munich (Rigort et al., 2010; Rigort et al., 2012a, b; Lučić et al., 2013). Up to now there are attempts to use high-pressure frozen material in copper tubes or membrane carriers (Studer et al., 2001; Vanhecke et al., 2003; Vanhecke & Studer, 2008) to prepare a lamella for cryo-electron microscopy (Hayles et al., 2010; De Winter et al., 2013). As a proof of principle, single cells were used and meaningful tissue samples still need to be done. To cryo-fix larger samples such as tissue, the copper tubes are less practical, therefore a sandwich of open holders, the so-called planchettes, is preferred (Moor et al., 1980). A new method has been developed for a cryo-liftout (Rubino et al., 2012), similar to the preparation of a TEM lamella done at room temperature. This method could have a great impact on cryo-sectioning by FIB. More recently, the first publication on cryo-FIB-SEM tomography on the mouse optic nerve has appeared; the images were recorded with the in-lens SE detector (Schertel et al., 2013). The lipid-rich area of myelin appears dark while the water-rich cytoplasm white. The authors are still arguing about how the contrast is generated but these experiments are encouraging for cryoFIB tomography.
FIB to get insight into biological objects
The first biological application a FIB was digging into a hair or an eye of a fly (Ishitani et al., 1995) to make internal surfaces available for imaging by scanning electron microscopy (SEM). No significant biological data were presented in this publication but the method was taken up by Damjana Drobne. She used the ion beam to manipulate unfixed yeast cells and oocytes of human and isopod origin. The opened-up structures were imaged in the secondary electron mode (Milani & Drobne, 2006). Later, the techniques were improved by using standard SEM preparation technology based on critical point drying (Robards & Sleytr, 1985). Here the internal structures, especially the digestive gland of Porcellio scaber were investigated (Drobne et al., 2004; Drobne et al., 2005; Lešer et al., 2010) (see review in Drobne, 2013). The FIB-SEM also allows for direct analysis of cell–substrate interfaces, e.g. neurons grown on microelectrodes, that is not possible with traditional ultramicrotomy (Greve et al., 2007).
FIB tomography
The most promising technology at the moment, however, is FIB-SEM tomography (Inkson et al., 2001; Holzer et al., 2004; Heymann et al., 2006; Knott et al., 2008; De Winter et al., 2009; Hekking et al., 2009; Heymann et al., 2009; Schroeder-Reiter et al., 2009; Lamers et al., 2011). From a sample embedded in resin, a cutting plane from the surface of the resin block is opened at a preselected region of interest. A thin section, as thin as 3 nm, is milled off with the ion beam and then with the electron beam the freshly generated plane is imaged. By repeating this process, a 3D data set of the object under investigation is generated. With this process, volumes usually in the range of 20–40 micrometres in all three dimensions can be imaged with a voxel resolution down to 3 nm. Up to now FIB-SEM tomography resides between TEM tomography, with a volume of 2–5 micrometres in X,Y and 500 nm in Z, at around 1 nm resolution and serial block face tomography (SBF-SEM), using a microtome built into a scanning electron microscope (Denk & Horstmann, 2004; Rouquette et al., 2009; Pellettieri et al., 2010) that provides much larger volumes at 10–15 nm pixel resolution and about 30–50 nm Z resolution (Fig. 2).
Here the different techniques to acquire three-dimensional information by electron microscopy are compared with each other in terms of resolution and volumes to be covered. The values are approximate and depend on the usage of the machines and the skills of the operator
In this chapter, we will concentrate on FIB-SEM tomography and the corresponding preparation methodology. Biological samples require an extensive preparation protocol to stabilize the sample for electron microscopy and preserve the delicate ultrastructure. Further care has to be taken that the entire sample is well contrasted as no on-section staining can be applied. Presently, standard preparation methods based on glutaraldhyde fixation are predominantly used. After the initial fixation the sample is treated with osmium tetroxide and also frequently with potassium ferricyanide (McDonald, 1984) to add enough metal for contrasting. A final uranyl acetate block staining is an additional advantage.
For FIB-SEM tomography, high contrast is needed to enhance the signal-to-noise ratio and to reduce the acquisition time of the electron image. Further, a high metal ion concentration can reduce charging effects. To enhance staining, especially of membranes, variations of a protocol to image internal structures by SEM (Tanaka & Mitsushima, 1984; Tanaka, 1989) are used that apply several steps of osmium tetro-xide in combination with tannic acid or thiocarbohydrazide. The sample undergoes a primary fixation with glutaraldehyde. Then it is postfixed with a mixture of osmium tetroxide and potassium ferrocyanide that is followed by a treatment of tannic acid and once more with osmium tetroxide (Jiménez et al., 2009). The other approach is postfixation with osmium tetroxide followed by thiocarbohydrazide and again osmium tetroxide (Tapia et al., 2013; Wilke et al., 2013). Finally the sample can be en bloc stained with uranyl acetate, lead aspartate or both (Hayat, 2000). These protocols are well suited to enhance membrane contrast but risk loosing proteins due to the extensive osmium tetroxide treatment (Behrman, 1980; Behrman, 1984). Further effort has to be made to combine high contrast with high-quality sample preparation.
After chemical fixation and contrasting, the sample is embedded in an epoxy resin. Durcupan (Stäubli, 1963; Kushida, 1964) is traditionally used in neuroscience (Knott et al., 2008; Wilke et al., 2013) whereas others use one of the numerous Epon 812 (Luft, 1961) substitutes (De Winter et al., 2009; Hekking et al., 2009; Tapia et al., 2013). With a microtome, the block is trimmed to expose the sample. Then the block is mounted on an SEM support stub with conductive carbon or resin and is sputter-coated with a thin layer of 5 nm to 10 nm of platinum to increase conductivity.
Often there are only a few areas of interest and they need to be identified in the large resin block. Sometimes those areas can be located by secondary electron imaging directly from the relief of the block surface (De Winter et al., 2009; Hekking et al., 2009; Bittermann et al., 2012) or with high acceleration voltages the heavy metal stained structures beneath the resin surface can be seen in the backscattered imaging mode. For a precise location of specific cellular events, the use of cover slips with an engraved reference grid (Loussert et al., 2012) or placing landmarks on the sample (Lucas et al., 2012) and correlative light and electron microscopy protocols are needed.
In the FIB-SEM, the sample is positioned at the coincident point of the ion and the electron beam. Traditionally, the stage is tilted so that the block surface becomes normal to the ion beam. This geometry often results in a serious brightness gradient from top to bottom of the imaged surface that is reduced when the sample is kept at 0o tilting and using oblique ion cutting (De Winter et al., 2009; Kizilyaprak et al., 2014). New software and higher mechanical accuracy of the instruments allow for tilting the block normal to the ion beam for cutting and the cut surface normal to the electron beam for imaging. This strategy results in higher resolution images, as cutting and imaging can be done under optimum conditions but is much more time consuming. For cutting, the ion beam is used with low currents in the range of 800 pA at an acceleration voltage of 30 kV. Imaging is done with 1–3 kV acceleration voltage also with low currents of about 800 pA using a back-scatter detector (Fig. 3). In one publication, secondary electron images were successfully used (Villinger et al., 2012).
An example of a micrograph of the liver was obtained in our Helios 600 FIB-SEM. The image was taken at 2 kV with an 800 pA electron current. The dwell time was 20 μs, pixel resolution 3 nm and the scan frame was 4096 px x 3536 px. Under these conditions, small structures like glycogen A, and the 67 nm banding pattern of collagen C can easily be seen. The typical bilayer of membranes could not be resolved, only the double membranes, e.g. around mitochondria (arrow; B). The bar of insets represents 500 nm.
Future
There are a few challenges for FIB-SEM tomography. Firstly, specimen preparation has to be adapted to meet the demands of SEM imaging, which are high contrast and electrically conductive without compromising the quality. The cellular details have to be preserved to a level that is at the resolution power of the instrument. Secondly, detection systems have to be developed and milling and imaging conditions found that allow much faster acquisition. Presently, in our hands acquisition of images (4096 px x 3536 px) at 5 nm voxel resolution takes about 10 min per picture and covers a surface of 20.5 × 17.7 μm2. Thirdly, data sets of several tens GB are not unusual asking for large computer power. Fourthly, last but not least, automated segmentation software (Andres et al., 2012) needs to be developed that can discriminate and segment structures in complicated grey-value background.
References
Al-Amoudi, A., Casta ̃no D ́ıez, D., Betts, M.J. & Frangakis, A.S. (2007) Themolecular architecture of cadherins in native epidermal desmosomes.Nature450, 832–837.Al-Amoudi, A., Norlen, L.P.O. & Dubochet, J. (2004) Cryo-electron mi-croscopy of vitreous sections of native biological cells and tissues.J.Struct. Biol.148, 131–135.Al-Amoudi, A., Studer, D. & Dubochet, J. (2005) Cutting artefacts andcutting process in vitreous sections for cryo-electron microscopy.J.Struct. Biol.150, 109–121.Andres, B., Koethe, U., Kroeger, T., Helmstaedter, M., Briggman, K.L.,Denk, W. & Hamprecht, F.A. (2012) 3D segmentation of SBFSEM im-agesofneuropilbyagraphicalmodeloversupervoxelboundaries.Med.Image Anal.16, 796–805.Behrman, E.J. (1980) On the oxidation state of osmium in fixed tissues.J.Histochem. Cytochem.28, 285–286.Behrman, E.J. (1984) The chemistry of osmium tetroxide fixation.TheScience of Biological Specimen Preparation 1983(ed. by J. P. Revel, T.Barnard & G. H. Haggis), pp. 1–5. SEM Inc., AMF O’Hare, IL.C©2014 The AuthorsJournal of MicroscopyC©2014 Royal Microscopical Society,254, 109–114FOCUSED ION BEAM SCANNING ELECTRON MICROSCOPY113Bittermann, A.G., Schaer, D., Mitsi, M., Vogel, V. & Wepf, R. (2012) Thinlayer plastification vs. block embedding: two alternative preparationstrategies for 3D-imaging of cultured cells and biofilms by FIB/SEM.European Microscopy Conference 2012(ed. by E. M. Society). John Wiley& Sons Ltd, Manchester.De Winter, D.A.M., Mesman, R.J., Hayles, M.F., Schneijdenberg,C.T.W.M., Mathisen, C. & Post, J.A. (2013) In-situ integrity controlof frozen-hydrated, vitreous lamellas prepared by the cryo-focused ionbeam-scanning electron.J. Struct. Biol.183, 11–18.De Winter, D.A.M., Schneijdenberg, C.T.W.M., Lebbink, M.N., Lich, B.,Verkleij,A.J.,Drury,M.R.&Humbel,B.M.(2009)Tomographyofinsu-lating biological and geological materials using focused ion beam (FIB)sectioning and low kV BSE imaging.J. Microsc.233, 372–383.Denk,W.&Horstmann,H.(2004)Serialblock-facescanningelectronmi-croscopy to reconstruct three-dimensional tissue nanostructure.PLoSBiol.2, 1900–1909.Drobne, D. (2013) 3D imaging of cells and tissues by focused ion beam /scanning electron microscopy (FIB/SEM).Meth. Mol. Biol.950, 275–292.Drobne, D., Milani, M., Ballerini, M., Zrimec, A., Zrimec, M.B., Tatti, F. &Draˇslar, K. (2004) Focused ion beam for microscopy and in situ samplepreparation: application on a crustacean digestive system.J. Biomed.Opt.9, 1238–1243.Drobne, D., Milani, M., Zrimec, A., Zrimec, M.B., Tatti, F. & Draslar, K.(2005)Focusedionbeam/scanningelectronmicroscopystudiesofPor-cellioscaber(Isopoda,Crustacea)digestiveglandepitheliumcells.Scan-ning27, 30–34.Dubochet, J. & McDowall, A.W. (1984) Frozen hydrated sections.Scienceof Biological Specimen Preparation, 1983(ed. by J.P. Revel, T. Barnard &G.H. Haggis), pp. 147–152. SEM Inc., AMF O’Hare.Frederik,P.M.,Busing,W.M.&Persson,A.(1982)Concerningthenatureof the cryosectioning process.J. Microsc.125, 167–175.Frederik,P.M.,Busing,W.M.&Persson,A.(1984) Surfacedefectsonthincryosections.Scanning Electron Microsc.1984, 433–443.Giannuzzi,L.A.,Drown,J.L.,Brown,S.R.,Irwin,R.B.&Stevie,F.A.(1998)ApplicationsoftheFIBlift-outtechniqueforTEMspecimenpreparation.J. Electron Microsc. Tech.41, 285–290.Giannuzzi, L.A., Phifer, D., Giannuzzi, N.J. & Capuano, M.J. (2007) Two-dimensional and 3-dimensional analysis of bone/dental implant inter-faces with the use of focused ion beam and electron microscopy.J. OralMaxillofac. Surg.65, 737–747.Giannuzzi,L.A.&Stevie,F.A.(1999)Areviewoffocusedionbeammillingtechniques for TEM specimen preparation.Micron30, 197–204.Greve, F., Frerker, S., Bittermann, A.G., Burkhardt, C., Hierlemann, A. &Hall, H. (2007) Molecular design and characterization of the neuro-microelectrode array interface.Biomaterials28, 5246–5258.Hayat, M.A. (2000)Principles and Techniques of Electron Microscopy. Bio-logical Applications. Cambridge University Press, Cambridge.Hayles, M.F., de Winter, D.A.M., Schneijdenberg, C.T.W.M., Meeldijk,J.D., Luecken, U., Persoon, H., de Water, J., de Jong, F., Humbel, B.M. &Verkleij, A.J. (2010) The making of frozen-hydrated, vitreous lamellasfrom cells for cryo-electron microscopy.J. Struct. Biol.172, 180–190.Hekking, L.H.P., Lebbink, M.N., De Winter, D.A.M., Schneijdenberg,C.T.W.M., Brand, C.M., Humbel, B.M., Verkleij, A.J. & Post, J.A. (2009)Focused ion beam-scanning electron microscope: exploring large vol-umes of atherosclerotic tissue.J. Microsc.235, 336–347.Heymann, J.A.W., Hayles, M., Gestmann, I., Giannuzzi, L.A., Lich, B. &Subramaniam, S. (2006) Site-specific 3D imaging of cells and tissueswith a dual beam microscope.J. Struct. Biol.155, 63–73.Heymann, J.A.W., Shi, D., Kim, S., Bliss, D., Milne, J.L.S. & Subramaniam,S. (2009) 3D Imaging of mammalian cells with ion-abrasion scanningelectron microscopy.J. Struct. Biol.166,1–7.Holzer, L., Indutnyi, F., Gasser, P.H., M ̈unch, B. & Wegmann, M. (2004)Three-dimensional analysis of porous BaTiO3 cer