School of Life Sciences

The Sussex Centre for Advanced Microscopy

Why Do We Use Electron Microscopes?

(link to Alternative Version for the Younger Reader)

You are probably familiar with light microscopes (LMs) , but what are electron microscopes (EMs), why were they developed and why are they now used so widely? The answer is all about resolution (the discrimination of fine detail in the image). The latter is dependent upon the wavelength of the illumination used in the microscope. The LM - of course - utilises light as the illuminating source. The highest effective magnification of a LM is about 1,000X; at magnifications beyond this no further detail in the image is acquired. The best possible theoretical resolution of a LM is about 0.25 microns, but in practice (allowing, for example, for lens defects and section thickness) is rather more than this; in other words, for example, a mitochondrion in the cytoplasm of a cell may just be distinguished. The EM, however, has a beam of electrons as its source of illumination. As the effective wavelength of this electron beam is shorter than that of light then resolution is improved; in modern transmission EMs this is typically around a few tenths of a nanometre (i.e. 0.0002 microns, though again - as with the LM - limitations of the specimen [section thickness usually] usually preclude the attainment of maximum theoretical resolution). Therefore the image in an EM may be magnified up to around 500,000X to elucidate much finer detail in the specimen observed. For example, whereas with a LM it is difficult even to resolve a mitochondrion, in an EM fine structure within a mitochondrion is discernible.

The Transmission EM

Basic principles and structure

The transmission EM (TEM) - as its name suggests - transmits electrons through the specimen (as light is transmitted through a specimen in a light microscope). The source of electrons is a filament (a thin piece of tungsten wire which is drawn out to a point). This filament is heated up by the application of a high voltage and electrons are literally boiled off the tip of the tungsten wire; this is termed thermionic emission. The filament is housed within an assembly which acts as a cathode gun and below this is an anode plate (which has a potential difference with respect to the cathode) so that electrons are attracted to it. A small hole (aperture) in this anode plate allows the electrons to pass through and down through the rest of the TEM column.

The TEM column stands vertically (up to a height typically of 7-8 feet) and at the top is the electron gun. Below this is a series of electromagnetic lenses which serve the same function as glass lenses in a LM. The layout of the lenses in an EM may be compared to that of a LM, but upside-down. Below the gun are the condenser lenses which condense the electron beam on the specimen and then an objective lens is at this specimen level where focusing takes place. The difference from hereon between an EM and a LM is that instead of an eyepiece lens in a LM there are projector lenses which project the focused image onto a fluorescent screen; as the eye cannot see electrons they are projected onto a (phosphor-coated) fluorescent screen so that the resultant fluorescence (when an electron hits the phosphor) is visible.

The whole TEM column is maintained under vacuum because the electrons do not have sufficient energy to pass through gas or water molecules. The electromagnetic lenses diffract the electrons by creating an electromagnetic field (at the pole pieces of the lens) which interacts with the charged electrons. When operating the TEM the currents in the electromagnetic lenses are altered (by turning control buttons) to focus the image or magnify it, etc.

Specimen preparation

There are two principle considerations when preparing specimens for TEM. Firstly, the resolution of the instrument and, secondly, the finite energy of the electrons. Because of the former, great care must be taken in 'fixing' a specimen to preserve its content in fine detail; there is little point in examining structure at high magnifications if there is no fine structure to discern. The limited energy of the electron beam requires that the specimen sections have to be very thin (typically around 100nm) and are therefore ultimately processed through into a medium (a resin) which facilitates this.

Fixation involves the use of a cross-linking agent (glutaraldehyde and/or formaldehyde) so that the internal in vivo content - and thus structure - is preserved. This stabilisation of intracellular content also prevents any extraction during subsequent processing. The fixative solution is buffered very carefully at the correct pH and ionic concentration so that no osmotic shock (swelling or shrinking) occurs. There is usually a secondary fixation in osmium tetroxide also as this fixes lipids efficiently and also imparts electron-density (and hence contrast) to the sample (as it is a heavy metal). After fixation specimens are dehydrated (usually in an ethanol series). Dehydration is necessary for two reasons: firstly, because the TEM is under a high vacuum there must be no water present in the sample, and, secondly, because the resins devised for ultra-thin sectioning are water-immiscible. Routinely-used resins are epoxy in nature (like araldite) and are added in liquid form in place of the final (100%) ethanol. This liquid resin is allowed to infiltrate and 'embed' the sample over a few days and is then polymerised (by heat). The final hard, 'cured' specimen 'block' is then ready for sectioning.

A typical preparative procedure would be:

1. Primary fix in 2-5% glutaraldehyde in a phosphate or sodium cacodylate buffer for 2-3h at room temperature.

2. Thorough buffer rinse.

3. Secondary fixation in 1% (buffered) osmium tetroxide for 2-4h at room temperature.

4. Thorough rinse in distilled water (samples are now fixed, so require no buffering).

5. Dehydration through an ethanol series (50%, 75% & 3 X 100% 20-30min each).

6. Infiltration with liquid epoxy resin for several days.

7. Resin (heat-)polymerisation.

The (cured) resin-embedded specimen is then trimmed down and very thin (ultrathin) sections are cut on a specialised section-cutter (an ultramicrotome). These sections are floated off onto water and are picked up on special TEM support grids (which are 3mm in diameter, circular with a very fine mesh of Cu bars). The sections are then stained in solutions of heavy metals (usually uranyl acetate and lead citrate) which produce contrast in the observed image by causing scattering of the electrons in the beam.

The above procedure is used for all tissue and cell samples. However if your specimen is particulate in nature then a much more rapid and simple method of preparation may be used. Particulate samples (such as viruses, ribosomes, DNA) may be pipetted directly onto a coated grid (a grid with a very thin layer of carbon cast over it). A heavy metal stain (usually uranyl acetate) is then added and allowed to dry down. This procedure is called negative staining because the specimen appears electron-luscent (light) against the dark background of the heavy metal stain which is dried down around it.

Specimen imaging

Because electrons have a limited energy, very thin sections of a specimen need to be cut (in a specialised section-cutting machine: the 'ultramicrotome') so that the electrons of the electron beam may pass through (these 'ultra-thin' sections are typically 50-100nm thick, which is equivalent to say 1/10th the diameter of a mitochondrion within a cell). These sections are collected onto support grids (equivalent to glass slides in the LM), which are small (c. 3mm), circular, fine meshes of copper routinely. The idea of these is that the sections are supported by the very fine mesh ('grid bars') and in-between there is only the sectioned specimen itself for the electron beam to pass through.

For image detail to be properly visible it must have contrast. In a TEM this is achieved as follows. Sections are stained with heavy metal solutions (normally uranium and lead); heavy metals are required because stains in TEM work have to be able to scatter the electrons. When the electron beam passes through a section then areas which have stained will diffract electrons and, dependent on what proportion of the electrons are scattered, these area will appear light grey to black in the final image. Conversely, totally unstained areas will appear white as the electrons will transmit straight through the sample and hit the phosphor screen to produce fluorescence. This contrast is enhanced in the TEM by the positioning of apertures (metal 'foils' or disks with very small [20-100µm] holes pierced through them) at the level of the objective lens. The smaller the aperture then the greater will be the ultimate contrast; this is so because a lower amount of scattering will be required to divert the electrons away from the hole within the aperture so that they will not be focused in the final image plane.

Research applications

Because of the resolution of the TEM very fine detail within a specimen may be resolved. Thus ultrastructural changes in samples with different treatments (e.g. drug effects) or at different stages of development may be evaluated (see Interpretation of Cellular Ultrastructure in the TEM below). Any such changes may be quantified by subsequent morphometric analysis (or 'stereology'): for example, changes in the numbers or amount (relative volume fraction within a cell) of a particular organelle such as mitochondria or secretory vesicles. The acquisition of digital images from the TEM greatly facilitates this process as there is numeric data within each image ('grey levels' ascribable to different cellular components). For more detail, see digital imaging and morphometric analysis.

The TEM may also be used to elucidate the precise subcellular location of specific proteins under study. This is achieved by a process known as immunogold labelling. In short, this involves a slightly modified sample preparation routine which aims to preserve the integrity (antigenicity) of the proteins (by lowered temperature and milder fixation). Subsequently, sections are incubated in specific antibodies directed against the protein of interest. These antibodies are not visible in the TEM, so their sites of binding are made electron-dense (and thus visible) by incubating in a second antibody (which binds to the primary antibody) to which a very small gold particle is bound. For more detail, see immunogold labelling.

The Scanning EM

Basic principles and structure

The scanning EM (SEM) is somewhat different to the TEM in many ways, but relies on the same basic principles. The source of illumination is identical to the TEM (tungsten filament) and there is an evacuated column. The latter in the SEM is much smaller than in a TEM; this is because there is not the same array of electromagnetic lenses within a SEM. As its name suggests, in the SEM the electron beam is scanned across the specimen surface (instead of being transmitted through). In this way the surface structure of a sample may be imaged. The SEM is therefore used for so-called 'bulk' samples (i.e. intact, unsectioned specimens such as whole, small insects, plant parts, etc.).

Specimen preparation

SEM samples are routinely fixed and dehydrated in similar fashion as for TEM preparation (see above). After this dehydration step the preparation is very different for SEM. The samples are routinely critical-point dried after dehydration. This step involves the replacement of the final (absolute) ethanol with liquid CO2 under pressure in a specialised pressure vessel (the critical-point dryer). The temperature of the liquid CO2 is then raised to its critical-point (around 32deg C) where it sublimes instantaneously from a liquid to a gaseous form. In this way the sample is perfectly dry but its turgidity is maintained. The dried sample is then stuck onto a special SEM sample holder (stub) and then sputter-coated with a very thin layer of gold in a coating device. This serves to give the sample a conductive layer which serves to earth (conduct charge away from) the specimen when the electron beam hits it (charging of a sample results in poor imaging quality).

Specimen imaging

Scanning coils within the SEM serve to scan the electron beam across the sample surface in a 'raster' fashion (as a TV image is built up for example by the fast scanning left-to-right and then downwards of hundreds of lines to give a complete image). When the electron beam hits a specimen a number of different interactions occur with the specimen; these interactions give rise to electron emission from the sample surface. Secondary electrons are used for routine, morphological imaging in the SEM. They are collected by a secondary electron detector, which very simply comprises a disk of phosphor (where electrons are converted to photons) behind which is a photomultiplier (light tube) through which the signal is transmitted. This light signal is converted to electrons which are subsequently amplified before being used to modulate the intensity of a cathode ray tube display. In this way the image is built up pixel by pixel (the individual 'dots' of information which make up the image) as the beam is scanned across the sample and is then displayed on the display monitor of the PC which controls the SEM instrumentation. The result is a 3-dimensional-looking image of the sample.

X-rays are also emitted from the sample under the electron beam. Information concerning the elemental composition of the specimen may be derived from their collection by an x-ray detector and its associated analytical software. This is because each element has its own 'fingerprint' of energies of the x-rays it emits. Therefore, when a sample's total x-ray emission is collected and the spectrum is displayed, peaks at particular energies can be related to specific elements.

The resolution of a SEM is much less than that of a TEM (typically around several nanometres). This is because the resolution is limited by the size of the electron beam 'spot' being scanned over the specimen surface (not the effective wavelength of the electron beam). When using a SEM for high resolution work the current to the beam ('probe current') is reduced to produce the smallest spot size possible. A simple analogy would be as follows: imagine yourself in a darkened room trying to discern the pattern of a raised sculpture on a wall with your clenched hand. If you were to unclench your hand and use a finger instead (equivalent to a smaller spot size) you would be able to feel (resolve) much finer detail immediately.

Research Applications

An SEM is often used purely to visualise detailed surface morphology. Because of its good resolution it may be used at high magnifications (typically 50-100,000X for biological samples) to reveal fine detail in structure (e.g. hair cells of the inner ear). Relatively low magnification images can also provide a lot of detail. It may be used in this way, for example, by taxonomists to verify specific anatomical traits of a genus (family) of animal or plant. Similarly, geneticists may wish to clarify any changes in external structure in particular mutants (such as the cockroach or Drosophila [fruitfly], for example).

Information concerning the elemental composition of a specimen can be revealed by x-ray microanalysis (see above). For biological samples elemental analysis is often used in conjunction with a cryo stage (which can be cooled down to liquid nitrogen temperatures, facilitating the examination of rapidly-frozen, hydrated samples); this methodology is thus termed cryo-analysis. An example within this Department is the subcellular localisation of certain key elements in plant tissue. (Biological samples are cryo-prepared for elemental analysis because the usual preparation using chemical fixation, dehydration, etc. will result in movement and/or extraction of the elements under study.)