heavy metals (such as osmium, uranium and lead) into the samples allowing their
attachment to structures to give the sample conductivity. The work of JoAnn
Buchanan has also been invaluable in designing SBEM protocols and she discusses
many useful protocols for SBEM [43]. On the other hand, preparations for
FIB-SEM do not require as many heavy metal steps as for SBEM since the ion
beam introduces positive Ga
þ ions on/into the sample that help to neutralize
surface charging [29, 30]. Thus the conductivity of a FIB-SEM sample itself is less
critical and it is possible to image samples with a large variety of preparations
(Fig. 5.3).
Utilizing a microwave oven for sample processing has been around for some
time. The microwaves aid in the processing by enhancing certain reactions and
increasing the diffusion of chemicals into the sample. It can speed up processing
substantially by reducing processing protocols from days to hours [44]. The wattage
of the microwave energy is kept low and is cycled so that the specimens do not
overheat and until now no changes to the ultrastructure have been attributed to
using this method. In addition, a low vacuum is applied during the processing to
enhance diffusion. It has even been shown that microwave processing can reduce
artefacts induced during standard processing [45]. Later sections, 5.3.1.2 and
5.4.1.2 give examples of microwave assisted processing protocols showing their
usefulness for SBEM and FIB-SEM.
As with TEM, biological samples need to be embedded in resins. Unfortunately,
these resins are not at all conductive which is a serious issue particularly for SBEM
[18, 46]. Standard formulas for resins used in TEM work will not work well in
SBEM or FIB-SEM because they need to be more resistant to beam damage. As a
result, resins that are more highly cross-linked have been used. The epoxy resin
Durcupan ACM gives the most reliable results with greater stability in the beam.
However, other resins such as hard formulas of Epon and Spurrs have also been
successfully used.
Specific to SBEM, attempts have been made to improve the conductivity by
including conductive material in the resins. Wanner et al. [31], removed the
pre-polymerised resin from around the sample and re-embedded the sample in
Epo-Tek EE129-4, an epoxy glue that contains silver particles. Nguyen et al. [47]
have used a fine conductive carbon powder, Ketjenblack EC600JD, and included it
in the final embedding resin to reduce charging. In both of these cases the conductive material in the resin does not penetrate the sample but sits around the
outside. There is a critical need for a truly conductive resin that could be used in
SBEM as the benefits would be enormous.
5.2.2 Imaging
As with SEM imaging in general, the parameters responsible for good image
quality are the voltage, the probe current and the dwell time. The aim is to get a
124
R. I. Webb and N. L. Schieber
attachment to structures to give the sample conductivity. The work of JoAnn
Buchanan has also been invaluable in designing SBEM protocols and she discusses
many useful protocols for SBEM [43]. On the other hand, preparations for
FIB-SEM do not require as many heavy metal steps as for SBEM since the ion
beam introduces positive Ga
þ ions on/into the sample that help to neutralize
surface charging [29, 30]. Thus the conductivity of a FIB-SEM sample itself is less
critical and it is possible to image samples with a large variety of preparations
(Fig. 5.3).
Utilizing a microwave oven for sample processing has been around for some
time. The microwaves aid in the processing by enhancing certain reactions and
increasing the diffusion of chemicals into the sample. It can speed up processing
substantially by reducing processing protocols from days to hours [44]. The wattage
of the microwave energy is kept low and is cycled so that the specimens do not
overheat and until now no changes to the ultrastructure have been attributed to
using this method. In addition, a low vacuum is applied during the processing to
enhance diffusion. It has even been shown that microwave processing can reduce
artefacts induced during standard processing [45]. Later sections, 5.3.1.2 and
5.4.1.2 give examples of microwave assisted processing protocols showing their
usefulness for SBEM and FIB-SEM.
As with TEM, biological samples need to be embedded in resins. Unfortunately,
these resins are not at all conductive which is a serious issue particularly for SBEM
[18, 46]. Standard formulas for resins used in TEM work will not work well in
SBEM or FIB-SEM because they need to be more resistant to beam damage. As a
result, resins that are more highly cross-linked have been used. The epoxy resin
Durcupan ACM gives the most reliable results with greater stability in the beam.
However, other resins such as hard formulas of Epon and Spurrs have also been
successfully used.
Specific to SBEM, attempts have been made to improve the conductivity by
including conductive material in the resins. Wanner et al. [31], removed the
pre-polymerised resin from around the sample and re-embedded the sample in
Epo-Tek EE129-4, an epoxy glue that contains silver particles. Nguyen et al. [47]
have used a fine conductive carbon powder, Ketjenblack EC600JD, and included it
in the final embedding resin to reduce charging. In both of these cases the conductive material in the resin does not penetrate the sample but sits around the
outside. There is a critical need for a truly conductive resin that could be used in
SBEM as the benefits would be enormous.
5.2.2 Imaging
As with SEM imaging in general, the parameters responsible for good image
quality are the voltage, the probe current and the dwell time. The aim is to get a
124
R. I. Webb and N. L. Schieber