Electron Microscopy Techniques, Strengths, Limitations and Applications
In this article, we explore how electron microscopy works, some of the common techniques and applications and ways the field continues to push the boundaries
Complete the form below to unlock access to ALL audio articles.
What is electron microscopy?Before answering this question, we need to take a step back in history and acknowledge the work of the Dutch scientist Antoni van Leeuwenhoek (1632-1723) and Englishman Robert Hooke (1635-1703). Both played instrumental roles in the identification of microorganisms using glass lenses and essentially launched the fields of optical microscopy and microbiology. Yet with advances in the field, German physicist Ernst Abbe recognized that optical microscopy would be limited by a fundamental law of optical physics - the diffraction of light - that would limit the resolution of optical microscopes. This was called the "Abbe diffraction limit".1 Abbe deduced that a microscope could not resolve two objects located closer than λ/2NA, where λ is the wavelength of light and NA is the numerical aperture of the imaging lens. Central to this equation defining the resolution limit is the wavelength, λ. It is directly proportional to the achievable resolution - the shorter the wavelength, the better the resolution. This is the link to electron microscopy.
What's the resolution of electron microscopy?
Types of electron microscope and how they work
- Transmission electron microscopy (TEM)
- Scanning electron microscopy (SEM)
- Scanning transmission electron microscopy (STEM)
- Reflection electron microscopy (REM)
- Freeze fracture electron microscopy
Recent developments in the field: Cryo-EM
Recent developments in the field: In-situ TEM
Without the work of German physicist Hans Busch (1884-1973), development of electron microscopy would have been impossible. He was the first to show that a magnetic field could focus a beam of electrons in a manner similar to how glass lenses can focus visible light. Fellow German physicist Ernst Ruska (1906-1988) took note of the work of Busch and applied it to the development of the electron microscope, for which he won the 1986 Nobel Prize in Physics.2 He was also well aware of the impact his discovery could have if the wavelength of the electrons could be reduced. French physicist Louis de Broglie (1892-1987) had already taken care of this using the theory of wave particle duality showing that beams of electrons acted as waves with a wavelength that could be predicted based on the velocity of the electron. This could be adjusted by accelerating the electrons through an electrical potential (in practice this is known as the accelerating voltage of the electron microscope). Electron wavelengths at different accelerating voltages are given in Table 1.
Accelerating voltage (kV)
The impact on microscopy resolution is now clear. The wavelength of the electrons listed in the table is about 5 orders of magnitude lower than that of the lowest wavelength of visible light (380,000 pm). Based on the Abbe diffraction limit equation given above, this will have an enormous effect on the spatial resolution. This was the theoretical driving force for the development of electron microscopy.
So, after all that, we can answer our original question. Electron microscopy uses a beam of electrons focused by electromagnetic lenses to image materials of all types at spatial resolutions that far exceed those obtainable by standard optical microscopy. There are two common types of electron microscopes: transmission electron microscopes (which was the type developed by Ruska) and scanning electron microscopes. These two types have also been hybridized to produce scanning transmission electron microscopes and scanning electron microscopes equipped with transmission detectors. Yet despite the maturation of the technologies, new technological developments continue to push the limits of resolution forward.3
What's the resolution of electron microscopy?
The resolution of electron microscopy depends upon several factors, central to which is the accelerating voltage as noted above. But other factors are important as well, such as the magnetic lenses in the microscopes and the effect of aberrations in electromagnetic lenses.4 Resolution limits for different types and sophistication of electron microscopes are listed in Table 2. Note that these modern TEM designs can achieve atomic resolution.
Type of electron microscope
Typical spatial resolution
Tabletop SEM (compact version can sit on a desk - thermal emission source)
~3 -15 nm
Thermal emission electron source SEM
Schottky field emission SEM
120 kV TEM
200 kV TEM
300 kV TEM
300 kV TEM with aberration correction
Types of electron microscope and how they work
Transmission electron microscopy (TEM)
In TEM, a beam of accelerated electrons is transmitted through the sample, interacting with the sample in various ways to obtain different types of information, before being detected by a phosphor screen, film or semiconductor-based detectors below the sample. As the beam is transmitted through the sample, there are two basic requirements in TEM. First, the accelerating voltage must be sufficiently high that the beam of electrons can pass through the sample without being fully absorbed, and secondly, to assist with this requirement, the samples must be thin, typically 100 nm in thickness. This latter requirement is a function of the average atomic number of the sample. Heavier elements that make up metals and alloys will be stronger electron absorbers and less lenient in terms of sample thickness. On the other hand, biological samples, which consist mainly of C, H, O, N and these low atomic number elements, do not readily absorb the electrons and can thus accommodate thicker samples. Biological samples are usually prepared by ultramicrotomy, where samples are embedded into a plastic resin and then sectioned using a microtome with either a glass or diamond knife. Inorganic materials may also be prepared this way, but more often they are cut into 3 mm disks, mechanically polished and then finally thinned to perforation with an ion beam or with an electrolytic solution. It is the areas around the perforation that will be thin enough for electron transmission. A slight disadvantage is that these samples will be gradually thicker as you move further from the point of perforation, so sample thickness effects will reduce transmission. Alternatively, one can use a focused ion beam instrument to pick out a specific area of the sample, remove it, attach to a TEM sample carrier and thin it down to ~100 nm using a focused Ga+ ion beam. This has the advantage of precision sample site selection and producing relatively uniform thickness samples. However, the instrumentation is expensive and a great deal of operator skill is required.
A schematic diagram depicting the major electron optical elements of a TEM is shown in Figure 1. In (a), operation of the TEM in bright field imaging mode is shown. This begins at the top of the instrument with a source of electrons, most commonly either a W-filament (where electrons are thermally emitted) or a field emission source (where high potentials are applied to extract the electrons from the source tip). Field emission sources may be thermally assisted (called Schottky field emission sources) or not (referred to as cold field emission sources). A set of condenser lenses are then used to shape the beam onto and through the sample. Focusing of the image takes place after the beam has passed through the specimen using a collection of lenses (objective, intermediate and projector lenses). The final image is formed on a fluorescent screen (the impact of electrons causing the emission of light), photographic film or semiconductor-based detectors (e.g., charge-couple devices (CCDs)).
Figure 1: Schematic diagram of a TEM in (a) bright field imaging mode and (b) electron diffraction mode. Credit: Technology Networks
However, images are not of much use unless they show some form of contrast. There are different ways in which contrast is formed in a TEM. For example, many types of samples imaged in a TEM will be crystalline in nature, and subject of the laws of electron diffraction given by the Bragg equation:
Where λ is the wavelength of the electrons, d is the spacing between lattice planes of a particular orientation, θ is the Bragg diffraction angle and n is the order of reflection. Using a TEM at 200 kV accelerating voltage, the wavelength is (Table 1) 0.00251 nm. Using the crystallographic planes of atoms in the metal Cu with the largest spacing between planes (the (111) plane of Cu) as an example, with a d-spacing of 0.207 nm, we can solve for sinθ.
What this tells us is that the Bragg angle for diffraction is nearly parallel to the electron beam. This means that crystallographic planes will diffract intensity away from the otherwise transmitted beam when they are aligned nearly parallel to the electron beam. This manifests itself as a darker area on the image screen or film. This type of contrast is referred to as diffraction contrast, and one can learn a great deal about the crystallographic make-up of the samples using this form of contrast. Furthermore, as samples can be tilted in the TEM, one can produce a series of images where different crystallographic planes are brought into the Bragg condition for diffraction to provide even more information.
The electron diffraction pattern can in fact be recorded by the TEM by operating it in a slightly different manner. The objective aperture that is inserted just below the objective lens in Figure 1a is removed to allow the diffracted beams to be transmitted. The lenses below are configured in a slightly different manner to allow the diffracted beams to be projected onto the image screen along with the intense transmitted beam centered in the diffraction pattern. The distance between the sample and the screen is known, as is the wavelength. These patterns can be indexed to gain crystallographic information and also help with compound identification, as all compounds will have specific crystal structures and lattice spacings that can be determined from the patterns and matched to data in readily available crystallographic libraries5 for which most academic institutions will have a license or at least the old school card catalogues.
A second type of contrast formation that commonly occurs is that of atomic number contrast. In its simplest form, this refers to the case where the electrons are absorbed by some high atomic number element in the sample, converting their energy to heat. Once absorbed, these electrons cannot transmit through the sample, and we are left with a darker area on the image screen below. However, because electrons accelerated with high voltages and thin samples are used, this form of atomic number contrast is not the most prevalent. More likely to occur is an interaction of the negatively charged electron beam with the positive potential of the nucleus which will scatter the trajectory of the electron and again lead to an intensity deficit on the imaging screen.
Neither of these mechanisms are likely to occur in a biological sample, as they typically do not show high degrees of crystallinity, and nor do they typically show a great deal of atomic number contrast, being primarily composed of low atomic number elements such as C, N, O and H. To aid contrast, lower accelerating voltages are often used, but more importantly the samples undergo an additional step during preparation called staining. Though there are many different recipes for staining biological samples, a commonly used combination is uranyl acetate and lead citrate. Note both contain the heavy metals, U and Pb which will be the contrast formers. These two stains will bind to different structures in the cell and their use provides the contrast required to identify the different organelles within cells and tissues. The result is a TEM image such as that shown in Figure 2 of a human mast cell.
Figure 2: Example of a biological sample, high pressure frozen, stained with uranyl acetate and lead citrate and visualized by TEM. Wild type Human mast cell line 1 (HMC-1); Image shows the nucleus of the cells and the black arrow indicates nuclear envelope budding events. Credit: Johanna Höög, Dimitra Panagaki and Jacob Croft, reproduced under the Creative Commons Attribution 3.0 Unported (CC BY 3.0) license.
The key information that can be gained from TEM and sample requirements are summarized below:
● Images with very high spatial resolution (~ 0.04 nm on high end modern instruments, easily achieving atomic resolution).
● Crystallographic information including direct imaging of crystal lattice defects and faults in crystalline materials.
● Electrons need to be accelerated at high voltages (usually 100-200 kV but can range from 40-300 kV on commercial instruments).
● Samples need to be thin to allow for electron transmission through the samples (~100 nm).
● Biological samples need some form of heavy metal staining, with protocols using uranyl acetate and lead citrate being the most common.
Scanning electron microscopy (SEM)
The scanning electron microscope is the type of electron microscope with which most of us are probably more familiar. Those images we see of the compound eye of a fly were acquired using an SEM. But interestingly, their development lagged a little compared to TEM. A wonderful history of the early development of SEM is provided by McMullan.6
There are two chief differences between an SEM and a TEM. In an SEM, the focused electron beam is scanned in the x-y directions across the sample. The beam in the TEM is not scanned unless used in scanning transmission electron microscopy (STEM - see next section). Secondly, there is no requirement that the sample be thin. The surface of the sample is what is examined and perhaps the only limitation of the sample is that it must be able to be inserted or fit within the analysis chamber.
A schematic diagram of an SEM is shown in Figure 3 and consists of many of the same elements as the TEM. It starts with a source of electrons at the top of the column which, again, may be a W filament thermal emission source or a Schottky thermally assisted field emission source. Some instruments do use a cold field emission source, but these are the exception more than the norm. And just like the TEM there is a series of condenser lenses that assist in shaping the beam and adjusting the beam current that will eventually impinge on the sample. A strip of apertures also assists in this - smaller apertures will lead to smaller beam currents and beam diameters (which ultimately determines the spatial resolution). The scan coils raster the beam across the sample in dimensions chosen by the user through selection of the magnification, while the chosen signal is detected in synchronicity with the raster. A higher magnification scans the beam across increasingly smaller regions. The objective lens focuses the beam to the finest diameter possible. The smaller the beam diameter, the finer the features of the sample that can be resolved. With modern instruments, the area of the column around the objective lens has been modified and we will return to this in the “Detectors in SEM” section.
Figure 3: Schematic diagram of an SEM. Credit: Technology Networks
An example of the 3D nature of images provided by SEM is illustrated by the stereoocilia bundle of the inner ear shown in Figure 4.
Figure 4: Example of an SEM image. An SEM image of the sensory hair bundle of a single hair cell from a terrapin's hearing organ in the inner ear. Vibrations made by sound cause the hairs to be moved back and forth, alternately stimulating and inhibiting the cell. When the cell is stimulated it causes nerve impulses to form in the auditory nerve, sending messages to the brain. Credit: Dr. David Furness reproduced under the Creative Commons Attribution 3.0 Unported (CC BY 3.0) license.
SEM facilities are found in academic, commercial and industrial centers worldwide. There are probably no materials that have not been examined in an SEM before. Applications include those that one would expect - biology, materials science and nanotechnology - but also in fields such as forensics,7 art conservation8 and cosmetics.9
The electron beam/sample interaction volume
When an energetic electron beam strikes a sample, many different signals originating from different depths from the sample surface are produced. These can be detected and exploited to provide a wealth of information about the sample. This beam-sample interaction, and the main signals it produces, are shown in Figure 5.
- Auger electrons are low energy electrons emitted from the near surface region of the sample and have energies characteristic of the elements from which they are emitted. They are used exclusively in Auger electron spectroscopy instruments rather than SEM’s.
- Secondary electrons originate from deeper within the sample and are a result of inelastic scattering of the primary beam. The primary beam interacts with electrons in the sample and imparts some of its energy to them. These secondary electrons may exit the sample and be detected. These electrons are also of generally low energy and provide a wealth of information about the topography of the samples.
- Finally, there is the case where again the primary electron interacts with the nucleus of atoms in the sample and its trajectory is reversed back away from the specimen. These are called backscattered electrons, emanate from greater depths within the sample and are of much higher energy, being effectively elastically scattered. As a result, backscattered electron images can give qualitative information about the relative atomic number of the elements present in the sample.
- Still deeper into the beam/sample interaction volume, characteristic X-rays are also generated which may be used to quantitatively determine the chemical composition of the sample.
Figure 5: Electron beam/sample interaction produces various signals from different depths within the sample. Auger electrons, secondary electrons, backscattered electrons and characteristic X-rays. Credit: Technology Networks
The depths of emission of these different signals is not constant and depends very highly on the accelerating voltage of the electron beam and the sample composition. The typical pear- shaped volume from which these signals originate will be confined closer to the sample surface the lower the accelerating voltage and the higher the atomic number. This is particularly crucial for characteristic X-ray emission. When the electron beam strikes the sample, it can knock electrons out of inner shell orbitals. These vacancies are then filled by electrons from the outer shells and in the process, emit X-rays that are characteristic of the element from which it originated. This is the basis of an ancillary microbeam analysis technique that is very common on SEM’s - energy dispersive X-ray analysis (EDS or EDX). However, the volume where X-ray generation occurs is much larger than the diameter of the electron beam, so if we were to detect and map the distribution of the characteristic X-rays of the elements in the sample, the spatial resolution will depend on the size of that interaction volume and not the diameter of the electron beam. One way to improve this is to lower the accelerating voltage. This reduces the size of that interaction volume and improves the spatial resolution of the X-ray maps.
Detectors in SEM
Every SEM will have an Everhart-Thornley secondary electron detector, which uses a positively biased Faraday cage to attract secondary electrons. These electrons are accelerated towards a scintillator where they produce photons, which are funneled into a photomultiplier tube where the signal is amplified and measured. By applying a small positive voltage to the Faraday cage, one can filter out the low energy electrons and detect only the higher energy backscattered electrons. However, this practice is not as common as it once was due to the development of solid-state backscattered detectors, which are typically ring-type detectors fitted directly under the objective lens or located on a piston that can be inserted and retracted to that same location (see Figure 6 - retractable backscattered electron detector). Note that these backscattered electrons are also subject to Bragg’s law of diffraction and with careful sample preparation, crystallographic information can also be obtained. This is the basis for the technique known as electron backscattered diffraction, or EBSD.
Most modern SEM’s will have one or more “in-lens” detectors. This type of configuration is shown in Figure 6. In this case, the magnetic field resulting from the electromagnetic objective lens and the electrostatic lens is utilized to collect and funnel the low energy secondary electrons back up the column for detection. The magnetic field is highly efficient at collecting these low energy electrons and so this detector, much like the backscatter in-lens detector, is very useful when operating at very low accelerating voltages (100 V to 3-5 kV) where the sample must be within a few mm of the objective lens to maintain a finely focused electron beam. However, this may also be one of the disadvantages of the in-lens detectors, as it is exactly these very low energy electrons that will be most susceptible to sample charging artefacts, which occur when the sample is not grounded and builds up charge during analysis. This is common when insulating samples are analyzed and resolved by the application of a thin (5-10 nm) film of carbon or some conducting metal (Au, Pd and Pt are commonly used) using an evaporation or sputter coater that is a common accessory in every SEM laboratory.
Figure 6: Schematic diagram of the region around the objective lens of a modern SEM, showing dual electromagnetic and electrostatic lenses and the trajectory of the secondary electrons collected by the actions of these two lenses back up the column to the upper secondary electron and backscattered electron detectors. Credit: Technology Networks
The key information that we can gain from an SEM and any special sample requirements are summarized below:
● High spatial resolution of sample surfaces (~ 0.5 nm) providing 3-dimensional topographic information with secondary electrons.
● Qualitative compositional differences in samples using backscattered electrons (one may also obtain crystallographic information with EBSD detectors).
● Can be very surface sensitive when used at low voltages owing to the reduction of the electron beam/sample interaction volume.
● In-lens detectors are very effective under these conditions.
● Quantitative elemental analysis possible using EDS.
● There is really no limit to the sample size, other than it must be able to be introduced into and fit in the analysis chamber.
Scanning transmission electron microscopy (STEM)
STEM is pretty much exactly what it sounds like - a combination of TEM and SEM. In fact, most TEM’s today are combination systems that can be operated in either TEM or STEM mode. Like TEM, the samples still need to be electron transparent, but the addition of the ability to raster the beam in a TEM allows the use of additional signals that cannot be spatially coordinated in conventional TEM. These include scattered primary beam electrons, characteristic X-rays and electron energy loss events. The exceptional spatial resolution of the TEM is maintained, with the electron beam focused to a small diameter at the sample surface. Characteristic X-rays can be detected and mapped just as they are in an SEM. However, the issue of the size of the interaction volume disappears because the sample is thin. When electrons produce an X-ray, they lose the equivalent amount of energy and when an electron energy loss spectrometer (EELS) is attached to the instrument, the loss events can also be mapped as a function of position on the sample. EELS has two main strengths over EDS - the maps will have a slightly better spatial resolution and it is a technique sensitive to the chemical environment and can therefore provide information about bonding and oxidation states similar to X-ray photoelectron spectroscopy (XPS), but at significantly higher spatial resolution.
Reflection electron microscopy (REM)
Reflection electron microscopy is a form of electron microscopy that also has its origins from Ernst Ruska.12 REM is generally performed in a TEM but by tilting the sample such that the electron beam is at near grazing incidence to the sample surface. The sample therefore no longer needs to be thin and electron transparent. As a result of this, the interaction of the beam with the sample is far less and the resultant information obtained is from the very sample surface. It is therefore a technique used to study crystal surfaces. Examples of applications include the study of surface topography, observation of surface structures, surface adsorption and oxidation processes.13 It is somewhat of a specialized method that is not nearly as commonly used as SEM, TEM and STEM.
Freeze fracture electron microscopy
Freeze-fracture electron microscopy is a TEM-based method used in biological studies and is particularly valuable as a means of imaging membrane structure. When biological samples are in the frozen state, membranes have a plane of weakness in their hydrophobic interior. If the sample is then fractured, it will do so along that plane of weakness and often split the membrane into two halves, each corresponding to a phospholipid monolayer with associated proteins. This produces a three-dimensional perspective of the membranous organization of the cell, along with views of the membrane interior. These details are made visible in the electron microscope by evaporating Pt on to the specimen at an angle, effectively making a Pt-C replica of the fracture plane.
There are four main steps in making a standard freeze-fracture replica:
i) Rapid freezing of the specimen
ii) Fracturing the specimen at low temperature (-100 °C or lower)
iii) Make the replica of the newly exposed frozen surface by vacuum-deposition of Pt and C
iv) Clean the replica using bleach or acids to remove the biological material
A full procedure is given by Severs.14 A typical image showing the distributions of lipid droplets surrounding the endoplasmic reticulum in a macrophage is shown in Figure 7.
Figure 7: Example of freeze fracture electron microscopy image showing lipid droplets (LD) and endoplasmic reticulum in a macrophage. Credit: From Robenek and Severs,15
Recent developments in the field: Cryo-EM
Cryo-EM is a workflow/microscopy that has been developed to address a longstanding issue in biological electron microscopy — how to look at biological samples in their native state. The impact of the development of cryo-EM was recognized in 2017 by being awarded the Nobel Prize in Chemistry.16 This has always been an issue as biological samples contain a large fraction of water, which is removed during typical sample preparation protocols such as chemical fixation or freeze drying. In cryo-EM, the sample is maintained at cryogenic temperatures throughout all stages of sample preparation and TEM analysis. The method is especially useful for studying protein structure in the native state, and perhaps most famously, at least recently, determining the structures of the proteins in the spikes17 of the novel coronavirus.18An example of how cryo-EM can be used to acquire images of samples such as bacteria, and also its use in cryo-EM tomography to produce 3D reconstruction images is shown in Figure 8.19 This is achieved by simply acquiring numerous images of the same section at different angles of incidence to the electron beam.
Figure 8: Use of cryo-EM tomography to image the interior structure of bacterial cells. (a,b) Illustration of spiral architecture of the nucleoid in bacteria showing (a) a 21 nm thick tomographic slice through the 3D volume of a cell and (b) a 3D surface rendering of the same cell, with the spiral nucleoid highlighted (yellow). (c) Higher magnification view of a tomographic slice through the cell, showing well-separated nucleoid spirals and ribosomes (dark dots) distributed at the edge of the nucleoid. (d) Expanded views of 21 nm thick tomographic slices, showing top-views of polar chemoreceptor arrays. A schematic model (inset) illustrates the spatial arrangement of the chemoreceptor arrays in the plane of the membrane. Credit: Milne et al. 19
Recent developments in the field: In-situ TEM
Another area of research that is garnering much attention recently is that of in-situ TEM, where dynamic processes are being followed at a near-atomic scale20 through the development of specialized environmental holders for TEM. Biomineralization processes can be followed in-situ which may provide insight in bone healing and repair of hard tissue.21 The role of bacteria in the cycling of metals in the environment22 has also been followed using in-situ TEM, the results of which may have important implications in environmental science. Host-pathogen interactions23 have also been studied which provided new information on the rules of engagement. An oxidation phenomenon of the Ni-based alloy 600, used in pressurized water reactor nuclear power24 systems, has also been studied and new processes that may affect the stress corrosion cracking behavior were observed.
1. Beyond the diffraction limit. Nat Photonics. 2009;3(7):361-361. doi:10.1038/nphoton.2009.100
2. The Nobel Prize. https://www.nobelprize.org/prizes/physics/1986/summary/. Accessed August 25, 2021.
3. Courtland R. The microscope revolution that’s sweeping through materials science. Nature. 2018;563(7732):462-464. doi:10.1038/d41586-018-07448-0
4. Rose HH. Historical aspects of aberration correction. J Electron Microsc (Tokyo). 2009;58(3):77-85. doi:10.1093/jmicro/dfp012
5. International Centre for Diffraction Data. https://www.icdd.com/. Accessed August 25, 2021.
6. McMullan D. Scanning electron microscopy 1928–1965. Scanning. 1995;17(3):175-185. doi:10.1002/sca.4950170309
7. Basu S. Scanning Electron Microscopy in Forensic Science. In: Encyclopedia of Analytical Chemistry. John Wiley & Sons, Ltd; 2006. doi:10.1002/9780470027318.a1122
8. Zagora J. SEM-EDX Pigment Analysis and Multi-Analytical Study of the Ground and Paint layers of Francesco Fedrigazzi’s painting from Kostanje. CeROArt. 2013;(EGG 3). doi:10.4000/ceroart.3248
9. Kaliyadan F, Gosai B, Al Melhim W, et al. Scanning electron microscopy study of hair shaft damage secondary to cosmetic treatments of the hair. Int J Trichology. 2016;8(2):94. doi:10.4103/0974-7753.188035
10. Sousa AA, Leapman RD. Development and application of STEM for the biological sciences. Ultramicroscopy. 2012;123:38-49. doi:10.1016/j.ultramic.2012.04.005
11. van den Bos KHW, Altantzis T, De Backer A, Van Aert S, Bals S. Recent breakthroughs in scanning transmission electron microscopy of small species. Adv Phys X. 2018;3(1):1480420. doi:10.1080/23746149.2018.1480420
12. Ruska E. Die elektronenmikroskopische Abbildung elektronenbestrahlter Oberflachen. Zeitschrift fur Phys. 1933;83(7-8):492-497. doi:10.1007/BF01338960
13. Yagi K. Reflection electron microscopy. J Appl Crystallogr. 1987;20(3):147-160. doi:10.1107/S0021889887086916
14. Severs NJ. Freeze-fracture electron microscopy. Nat Protoc. 2007;2(3):547-576. doi:10.1038/nprot.2007.55
15. Robenek H, Severs NJ. Recent advances in freeze-fracture electron microscop: the replica immunolabeling technique. Biol Proced Online. 2008;10(1):9-19. doi:10.1251/bpo138
16. The Nobel Prize. https://www.nobelprize.org/prizes/chemistry/2017/press-release/. Accessed August 25, 2021.
17. Wrapp D, Wang N, Corbett KS, et al. Cryo-EM structure of the 2019-nCoV spike in the prefusion conformation. Science (80- ). 2020;367(6483):1260-1263. doi:10.1126/science.abb2507
18. Walls AC, Park Y-J, Tortorici MA, Wall A, McGuire AT, Veesler D. Structure, Function, and Antigenicity of the SARS-CoV-2 Spike Glycoprotein. Cell. 2020;181(2):281-292.e6. doi:10.1016/j.cell.2020.02.058
19. Milne JLS, Borgnia MJ, Bartesaghi A, et al. Cryo-electron microscopy - a primer for the non-microscopist. FEBS J. 2013;280(1):28-45. doi:10.1111/febs.12078
20. Pu S, Gong C, Robertson AW. Liquid cell transmission electron microscopy and its applications. R Soc Open Sci. 2020;7(1):191204. doi:10.1098/rsos.191204
21. He K, Sawczyk M, Liu C, et al. Revealing nanoscale mineralization pathways of hydroxyapatite using in situ liquid cell transmission electron microscopy. Sci Adv. 2020;6(47):eaaz7524. doi:10.1126/sciadv.aaz7524
22. Couasnon T, Alloyeau D, Ménez B, Guyot F, Ghigo J-M, Gélabert A. In situ monitoring of exopolymer-dependent Mn mineralization on bacterial surfaces. Sci Adv. 2020;6(27):eaaz3125. doi:10.1126/sciadv.aaz3125
23. Dearnaley WJ, Schleupner B, Varano AC, et al. Liquid-cell electron tomography of biological systems. Nano Lett. 2019;19(10):6734-6741. doi:10.1021/acs.nanolett.9b01309
24. Burke MG, Bertali G, Prestat E, Scenini F, Haigh SJ. The application of in situ analytical transmission electron microscopy to the study of preferential intergranular oxidation in Alloy 600. Ultramicroscopy. 2017;176:46-51. doi:10.1016/j.ultramic.2016.11.014