cryo-FIB followed by TEM tomography essentially supersedes cryo- electron
microscopy/tomography of vitreous sections (CEMOVIS/CETOVIS) because it does
not suffer from the unavoidable artefacts that plague knife-cut sections. These include
significant compression (>30%), and further deformations that occur during cutting or
attachment to a support grid. Nevertheless, tomography of vitreous sections led to key
breakthroughs including the first demonstration of ATP synthases in mammalian cells
[64], and the discovery of the mycobacterial outer membrane [70, 71].
1.4 Radiation Damage
An electron beam represents a form of ionising radiation, collimated into a small area
such that all specimens, whether frozen or plastic-embedded, are prone to significant
and irreversible damage. In TEM tomography, the need to acquire all of the projections from the identical region of interest places an ultimate limit on the amount of
information that can be acquired. It is therefore critical to know how to measure and
control the electron dose, and to keep this dose at subcritical levels. In frozen-hydrated
specimens, radiation (colloq., ‘beam’) damage can be seen in the form of gas bubbles
that are generated as a result of radiolysis. Perhaps more critically, this is preceded by
more subtle damage that can erase fine detail. The consequence of an upper limit to the
cumulative electron dose is that it must be estimated before exposure of the specimen.
Too little dose may lead to difficulty in acquisition (failed position tracking, for
example) but more importantly, it results in a poor yield of information. The cumulative dose during tilt-series acquisition should not exceed 100 e
− /Å
2 [3].
For plastic sections, deformations in the form of shrinkage and mass loss are
profound [72]. It is common practice to pre-irradiate the specimen, which allows it
to undergo a rapid shrinkage phase before acquisition of the tilt series. Here,
successful alignment of tomograms is dependent on comparatively minor changes
between subsequent projections. This practise is addressed systematically in
Chap. 4. In the author’s experience, a tomographic reconstruction of an Epon
section with nominal (microtome) section thickness of 150 nm will approximate
110 nm, even when using low-dose acquisition techniques and cooling the specimen in a cryo- sample holder. The final thickness also appears to be quite variable.
Quantitation from plastic sections is not possible without bold assumptions concerning distortions. Furthermore, these distortions are not isotropic. Even for
pre-irradiated plastic sections, it is advisable to use low-dose acquisition techniques
as described below. This helps to facilitate subsequent alignment.
1.5 Acquisition of Projections
We have already noted that tomograms comprise multiple views of the same object.
We will first discuss the considerations for acquisition of cryo tomograms. Gordon
et al. [73] summarise the electron tomography problem as the necessity to “reconstruct
1 Electron Tomography: A Primer
9
microscopy/tomography of vitreous sections (CEMOVIS/CETOVIS) because it does
not suffer from the unavoidable artefacts that plague knife-cut sections. These include
significant compression (>30%), and further deformations that occur during cutting or
attachment to a support grid. Nevertheless, tomography of vitreous sections led to key
breakthroughs including the first demonstration of ATP synthases in mammalian cells
[64], and the discovery of the mycobacterial outer membrane [70, 71].
1.4 Radiation Damage
An electron beam represents a form of ionising radiation, collimated into a small area
such that all specimens, whether frozen or plastic-embedded, are prone to significant
and irreversible damage. In TEM tomography, the need to acquire all of the projections from the identical region of interest places an ultimate limit on the amount of
information that can be acquired. It is therefore critical to know how to measure and
control the electron dose, and to keep this dose at subcritical levels. In frozen-hydrated
specimens, radiation (colloq., ‘beam’) damage can be seen in the form of gas bubbles
that are generated as a result of radiolysis. Perhaps more critically, this is preceded by
more subtle damage that can erase fine detail. The consequence of an upper limit to the
cumulative electron dose is that it must be estimated before exposure of the specimen.
Too little dose may lead to difficulty in acquisition (failed position tracking, for
example) but more importantly, it results in a poor yield of information. The cumulative dose during tilt-series acquisition should not exceed 100 e
− /Å
2 [3].
For plastic sections, deformations in the form of shrinkage and mass loss are
profound [72]. It is common practice to pre-irradiate the specimen, which allows it
to undergo a rapid shrinkage phase before acquisition of the tilt series. Here,
successful alignment of tomograms is dependent on comparatively minor changes
between subsequent projections. This practise is addressed systematically in
Chap. 4. In the author’s experience, a tomographic reconstruction of an Epon
section with nominal (microtome) section thickness of 150 nm will approximate
110 nm, even when using low-dose acquisition techniques and cooling the specimen in a cryo- sample holder. The final thickness also appears to be quite variable.
Quantitation from plastic sections is not possible without bold assumptions concerning distortions. Furthermore, these distortions are not isotropic. Even for
pre-irradiated plastic sections, it is advisable to use low-dose acquisition techniques
as described below. This helps to facilitate subsequent alignment.
1.5 Acquisition of Projections
We have already noted that tomograms comprise multiple views of the same object.
We will first discuss the considerations for acquisition of cryo tomograms. Gordon
et al. [73] summarise the electron tomography problem as the necessity to “reconstruct
1 Electron Tomography: A Primer
9