collection [70], and through-focal tilt-series approach for keeping thick volume
entirely in focus, ameliorating the beam-spreading phenomenon [35, 36].
2.3.3 Cryo-STEM Tomography for Vitrified,
Unstained Cells
The advantages of STEM tomography for stained, embedded thick sections leads to
the obvious question of whether there is an advantage for vitrified, unstained
samples as well. Until recently, there was a commonly held belief that STEM
imaging for cryo-preserved, unstained biological specimens would not be possible
due to insufficient scattering contrast from the light elements of biological specimens, given the dose limits for sample damage by the focused scanning probe
beam. We recently showed that these concerns can be overcome [20].
As discussed in Sect. 2.2.3, cells and tissue consist almost entirely of organic
material and water, which constitute small atoms with low atomic number Z:
hydrogen, carbon, nitrogen, oxygen. Specific objects in the cells are enriched in
heavier atoms. Examples include ribosomes and poly-phosphate bodies, with
enriched phosphorus content, and mineralization precursors based on calcium
deposits [71, 72]. In addition, the cell has varying mass density due to inherent
density differences in specific organelles macromolecules (lipids vs proteins, for
example). The contrast observed in cryo-STEM tomography (CSTET) images will
depend on the detector used. The DF image represents the elastic scattering, with an
extra sensitivity to heavier atoms. The BF contrast shows the missing signal, i.e.,
the scattering to all angles beyond the detector outer cutoff. BF contrast represents
mainly the mass density variation in different parts of the cell, as projected down the
path of the electron probe beam. Choosing a sufficiently large upper collection
angle for the BF signal, similar or larger than the probe convergence angle, will
suppress phase contrast.
Damage to vitrified biological material is a major concern, and much literature
has documented the dose (usually measured in electrons/Å
2 ) that is tolerable in
order to prevent loss of detail and resolution [19, 73–75]. Dose tolerance is based
on the effect of specific interactions between the electron and the specimen (as
described in Sect. 2.2.1). Indeed, Dose tolerance of vitrified specimens was shown
to actually be improved in cryo-STEM with respect to cryo-TEM imaging (as
measured by observation of bubbling in the sample) [20]. Beam scanning was
recognized long ago as a means to reduce sample damage in electron crystallography, and the STEM probe can be considered an extreme form of this approach
[62]. As for preservation of high resolution information, it was shown that with
judicious selection of scan rate and electron flux, resolution reaching 2 Å could be
obtained for STEM images of SrTiO 3 at 15 electron/Å
2 electron doses, similar to
those used for single particle macromolecular TEM imaging [76]. Due to the
46
S. G. Wolf et al.
entirely in focus, ameliorating the beam-spreading phenomenon [35, 36].
2.3.3 Cryo-STEM Tomography for Vitrified,
Unstained Cells
The advantages of STEM tomography for stained, embedded thick sections leads to
the obvious question of whether there is an advantage for vitrified, unstained
samples as well. Until recently, there was a commonly held belief that STEM
imaging for cryo-preserved, unstained biological specimens would not be possible
due to insufficient scattering contrast from the light elements of biological specimens, given the dose limits for sample damage by the focused scanning probe
beam. We recently showed that these concerns can be overcome [20].
As discussed in Sect. 2.2.3, cells and tissue consist almost entirely of organic
material and water, which constitute small atoms with low atomic number Z:
hydrogen, carbon, nitrogen, oxygen. Specific objects in the cells are enriched in
heavier atoms. Examples include ribosomes and poly-phosphate bodies, with
enriched phosphorus content, and mineralization precursors based on calcium
deposits [71, 72]. In addition, the cell has varying mass density due to inherent
density differences in specific organelles macromolecules (lipids vs proteins, for
example). The contrast observed in cryo-STEM tomography (CSTET) images will
depend on the detector used. The DF image represents the elastic scattering, with an
extra sensitivity to heavier atoms. The BF contrast shows the missing signal, i.e.,
the scattering to all angles beyond the detector outer cutoff. BF contrast represents
mainly the mass density variation in different parts of the cell, as projected down the
path of the electron probe beam. Choosing a sufficiently large upper collection
angle for the BF signal, similar or larger than the probe convergence angle, will
suppress phase contrast.
Damage to vitrified biological material is a major concern, and much literature
has documented the dose (usually measured in electrons/Å
2 ) that is tolerable in
order to prevent loss of detail and resolution [19, 73–75]. Dose tolerance is based
on the effect of specific interactions between the electron and the specimen (as
described in Sect. 2.2.1). Indeed, Dose tolerance of vitrified specimens was shown
to actually be improved in cryo-STEM with respect to cryo-TEM imaging (as
measured by observation of bubbling in the sample) [20]. Beam scanning was
recognized long ago as a means to reduce sample damage in electron crystallography, and the STEM probe can be considered an extreme form of this approach
[62]. As for preservation of high resolution information, it was shown that with
judicious selection of scan rate and electron flux, resolution reaching 2 Å could be
obtained for STEM images of SrTiO 3 at 15 electron/Å
2 electron doses, similar to
those used for single particle macromolecular TEM imaging [76]. Due to the
46
S. G. Wolf et al.