Lecture 10: TEM sample preparation

Contents
1 TEM sample holders 1

2 Parts of a TEM holder 1

2.1 Types of TEM holder . . . . . . . . . . . . . . . . . . . . . . . 4
2.2 In-situ TEM holders . . . . . . . . . . . . . . . . . . . . . . . 7

3 TEM sample preparation 9

3.1 Electrolytic polishing . . . . . . . . . . . . . . . . . . . . . . . 9
3.2 Ion milling technique . . . . . . . . . . . . . . . . . . . . . . . 12
3.3 Cross section sample preparation . . . . . . . . . . . . . . . . 17
3.4 Replica technique . . . . . . . . . . . . . . . . . . . . . . . . . 18

4 Focused ion beam TEM sample preparation 20

1 TEM sample holders

The specimen holder is used to insert the sample into the TEM from outside
for imaging. The sample is loaded onto the TEM stage for imaging. Now,
TEM holder and stage are essentially integrated. The TEM holder plays
and important role since it allows for introduction of a sample in ambient
conditions into an instrument that is held in vacuum. Usually the holder
region is pumped separately before the sample is fully introduced in the
TEM. A schematic diagram of some of the pumps used in the TEM is showed
in figure 1.
There are 2 main types of holder in the TEM
1. Top loading - the sample has no connection to the outside. After de-
livery of the sample to the stage the holder is withdrawn. Top loading
holders were used in earlier instrument version, especially for high res-
olution since this minimizes external vibrations. A schematic of a top
loading holder is shown in figure 2.

1
 MM3030: Materials Characterisation

Figure 1: Pumping system in a TEM. Taken from Transmission Electron

microscopy - Williams and Carter.

2
 MM3030: Materials Characterisation

Figure 2: Top entry loader in (A) cross section and (B) top view. Taken
from Transmission Electron microscopy - Williams and Carter.

3
 MM3030: Materials Characterisation

Figure 3: Schematic of a side entry holder. Taken from Transmission Electron

microscopy - Williams and Carter.

2. Side loading - these are the most commonly used holders now. The
advantage of the side loading holders is that it is possible to provide
external stimuli (heat, current, load) while imaging.

2 Parts of a TEM holder

The focus will be on the side loading holder since it is ubiquitous in use. A
schematic of the holder is shown in figure 3. Various parts of the holder can
be identified from the figure

1. O-ring - Provides the mechanical link to the TEM column. The o-ring
separates the portion of the holder in the ambient from the portion
in vacuum. Typically a viton o-ring is used and some holders have
more than one. The o-rings are usually greased but the excess grease
should be removed before inserting in the TEM since this can affect the
vacuum. The way in the holder is inserted in the instrument depends
on the specific TEM.

2. Jewel bearing - This provides the other mechanical link of the holder
to the TEM. This is used for mechanical translation of the sample (in
the x-y plane).

3. The cup - this actually holds the sample. Typical TEM samples are 3
mm discs though bulk holders are available (not very common).

4. Clamping ring - this clamps and holds the sample in place. Both the
cup and the clamping ring are exposed to the electron beam and hence

4
 MM3030: Materials Characterisation

Figure 4: Some types of side entry holders. From top a rotation holder,
heating holder, cooling holder, double-tilt, and single-tilt holder. Taken from
Transmission Electron microscopy - Williams and Carter.

should be made of a material with low x-ray fluorescence. Be is the

material most commonly used.

2.1 Types of TEM holder

There are various designs for the side entry holder. Some of them are shown
in figure 4.

1. Single tilt holder - most commonly used holder. Tilting around the
axis of the holder is possible. Shown in figure 4.

2. Quick change holder - this is a single tilt holder with a clamp instead of
screw. Useful for quick change of samples. Philips CM 12 series TEMs
have a quick change holder.

3. Multiple specimen holder - some holders can hold more than one sample.

5
 MM3030: Materials Characterisation

Figure 5: Single tilt holder with 2 and 5 specimen cups. Taken from Trans-
mission Electron microscopy - Williams and Carter.

This is again for multiple sample preparation. These are usually single
tilt holders. Multiple tilt holders are shown in figure 5.

4. Bulk specimen holders - These are used for bulk specimens that cannot
be made into 3 mm disks. This is not very common. The area of interest
should still be thin enough for the electron beam to pass through. A
bulk specimen holder is shown in figure 6.

5. Double tilt holder - two tilt angles are possible - in-plane and out-of-
plane. This is most commonly used for diffraction and imaging.

6. Tilt-rotate holder - this allows to select the tilt axis and then rotate
the specimen.

7. Low-background holder - the cup and clamping ring are made of Be.
Mainly used for x-ray studies (EDAX) since Be has low x-ray fluores-
cence.

2.2 In-situ TEM holders

One of the advantages of a side loading holder is that it is possible to have
signals from the outside to interact with the sample while imaging. This gives
added functionality to the TEM and provides for direct structure-property
correlations. Some of the in-situ TEM holders are

6
 MM3030: Materials Characterisation

Figure 6: Bulk specimen holder. Taken from Transmission Electron mi-

croscopy - Williams and Carter.

1. Heating holder - for heating samples in-situ. Usually a resistive heater

is used with temperatures as high as 1000 ◦ C possible. Care should be
taken not to weld the sample to the holder. A heating holder is shown
in figure 4.

2. Cooling holder - holders can go up to liquid nitrogen (77 K) or helium

temperatures (4 K). This is mainly used for biological samples, where
beam damage is minimized by cooling the sample. A cooling holder is
shown in figure 4.

3. Cryo-transfer holder - cooling holder with cryo-transfer so that the

sample is never heated to room temperature.

4. Straining holder - holder is used for applying external strain to the

TEM sample. Usually piezoelectric actuators are used for precise strain
application. They can be combined with other in-situ holders, like
heating holder, like figure

5. EBIC and CL holders - these can supply electrical current to the sam-
ple.

There are a number of in-situ holders developed by specific research groups.

For example, in-situ thermal evaporation is possible in the TEM. Similarly
dynamic TEM holders are available for driving in-situ reactions at high heat-
ing rates combined with diffraction measurement.

7
 MM3030: Materials Characterisation

Figure 7: Strain holder with heating. Taken from Transmission Electron

microscopy - Williams and Carter.

8
 MM3030: Materials Characterisation

3 TEM sample preparation

Sample preparation is an important part of TEM analysis. There are two
main criteria for TEM sample preparation

1. The sample should be electron transparent. If not the whole sample

at least the ROI should be thin. Typical thickness values for metallic
samples should be 30 - 50 nm. Usually, 100 nm is a upper limit for the
sample thickness.

2. The sample should be mechanically robust for handling.

Usually, the sample to be investigated is directly prepared from the under-

lying specimen. Sometimes for bulk specimens replica samples are prepared.
This is mainly for studying surface topographies and precipitates. TEM sam-
ples are either self-supported or mounted on a grid for analysis. Copper grids
are the most commonly used, though for high temperature work Mo grids
are used. For nanoparticles and thin films a-C film is used as support. a-C
has low contrast in the TEM and will not obscure the contrast arising from
the specimen. Some typical TEM grids are shown in figure 8.
The TEM preparation technique depends on the type of sample and also
on the property of interest. Mechanical damage during sample preparation
might be of an issue for studying dislocation densities in a specimen. Simi-
larly cross sectional TEM preparation might be necessary for some samples,
which can be a time consuming process. For most samples we need to pre-
thin the sample to an initial thickness of 100 - 200 µm. Cut a 3 mm disk
from the sample (for bulk specimens). Thin the central disk to a few µm and
then thin further (by different techniques) till sample is electron transparent.
Sample sizes in the different stages of TEM preparation are shown in figure
9.

3.1 Electrolytic polishing

Electrolytic polishing is used for conducting samples like metals/alloys in
order to produce samples that are electron transparent. The initially sheet
thickness can be around a few hundred µm. This can be prepared by rolling
or grinding bulk specimens. Similarly, metal coatings on substrates can be
peeled off and used for the final thinning. Thin discs can also be cut from
bulk specimens. This process is called coring. Some of the different equip-
ment for coring is shown in figure 10. There will be some mechanical damage
at the surface due to coring but these can be removed during polishing.
These discs are thinned by electrolytic polishing. The most common elec-

9
 MM3030: Materials Characterisation

Figure 8: Typical TEM grids. Taken from Transmission Electron microscopy

- Williams and Carter.

Figure 9: Stages in the TEM sample preparation from a bulk specimen.

Taken from TEM sample preparation - Sridhara Rao et al, Microscopy: Sci-
ence, Technology, Applications and Education.

10
 MM3030: Materials Characterisation

Figure 10: Different coring tools (A) mechanical punch (B) Abrasive-slurry
disc cutter (B) Ultrasonic cutter (D) Spark-erosion cutter. Taken from Trans-
mission Electron microscopy - Williams and Carter.

11
 MM3030: Materials Characterisation

trolytic polishing technique is the window technique. The sample is made

the anode and a thin stainless sheet is made the cathode. The sample is
immersed in the electrolyte, which is usually cooled by water or liquid nitro-
gen. Perchloric acid is usually used as the electrolyte. The sample edges are
covered by lacquer to expose a ’window’, hence the name. The experimental
setup and the hole generation are shown in figure 11.
When a current is applied the material is dissolved from the anode (sample)
and deposits on the cathode. The rate of dissolution depends on the cur-
rent and applied voltage. The I − V characteristics are shown in figure 12.
Depending on the current and voltage, there are three regimes - etching, pol-
ishing, and pitting. The edges are coated so that material removal will start
within the window. Once a hole is formed within the window the sample is
pulled out. The region around the hole is usually electron transparent and
can be mounted on a TEM grid. The problem with this technique is that it
is hard to control the location of the hole in the window.
A modification to the window technique is the Bollmann technique. A
pointed cathode is used to form an initial hole in the location of choice and
this is then replaced by flat electrodes to expand the electron transparent re-
gion. A variation of the Bollmann technique, which is very commonly used,
is the Jet polishing technique. A schematic of the jet polishing technique is
shown in figure 13. Here the electrolyte is locally sprayed on the sample,
using pointed electrodes. The jet helps in polishing from both ends of the
sample and hence it is faster. The method also reduces electrolyte use and
can produce large electron transparent areas.
The electrolyte polishing technique can be used for samples that are conduct-
ing - mostly metals and alloys. For non-conducting samples like ceramics,
semiconductors, deposited films, other TEM sample preparation techniques
are needed.

3.2 Ion milling technique

For non-conducting samples usually grinding and polishing steps are used in
order to reduce sample thickness. Some an ultramicrotome is used in order
to generate thin samples. These can be either electron transparent or can
be used as the starting material for further thinning. The schematic of the
technique is shown in figure 14. For samples, where ultramicrotome cannot
be used then a standard tripod polisher is used in order to thin the sample.
This produces samples that are a few µm thick. The final polishing step is
done by an ion beam miller.
The schematic of the ion beam miller and an actual instrument are shown in
figure 15. The sample is bombarded by high energy ions or neutral atoms.

12
 MM3030: Materials Characterisation

Figure 11: Window polishing technique. Taken from Transmission Electron

microscopy - Williams and Carter.

13
 MM3030: Materials Characterisation

Figure 12: I − V characteristics during polishing. Taken from Transmission

Electron microscopy - Williams and Carter.

Figure 13: Schematic of the jet polishing technique. Taken from Transmis-
sion Electron microscopy - Williams and Carter.

14
 MM3030: Materials Characterisation

Figure 14: Schematic of ultramicrotome technique. Taken from Transmission

Electron microscopy - Williams and Carter.

15
 MM3030: Materials Characterisation

Figure 15: Schematic of ion beam milling technique. Taken from Transmis-
sion Electron microscopy - Williams and Carter.

16
 MM3030: Materials Characterisation

Figure 16: Penetration vs. thinning as a function of beam incidence. Taken

from Transmission Electron microscopy - Williams and Carter.

Usually Ar ions are used and they are formed by passing the Ar gas though
a high voltage (4 - 6 keV ). The sample is held in vacuum and also usually
cooled by liquid nitrogen. The ions are incident on the sample and sputter
material away. To minimize ion penetration the beam is usually incident
at a low angle (v 20◦ ) though if the angle is very small the sputter rate is
small. The trade off between ion penetration and thinning rate is shown in
figure 16. Ion beam is highly controlled and a localized process but it is time
consuming. Sputter rates are usually a few Å per second so that creating an
electron transparent sample can take hours, especially if the initial thickness
is high.

3.3 Cross section sample preparation

Cross sectional samples are routinely interrogated in the TEM, especially for
studies of interface. The schematic of the cross section process is shown in
figure 17. Slices from the sample are cut using a diamond slicer. These slices
are placed between spacer layers and then glued on to a grid. The slices
are glued in such a way that the interface is parallel to the slot in the grid.
This sample is then thinned by standard tripod polishing until it is a few
µm thick. The final sample is thinned using a ion beam miller to create an
electron transparent sample.

17
 MM3030: Materials Characterisation

Figure 17: Cross sectional sample preparation. Taken from Transmission

Electron microscopy - Williams and Carter.

3.4 Replica technique

Replica technique is used for studying bulk specimens which cannot be de-
stroyed to prepare electron specimens. It is also useful for studying surface
topography features and precipitates though SEM techniques have gradually
replaced replica sample preparation. The schematic of the replica technique
is shown in figure 18. A replica of the sample surface is prepared using a
plastic mold. The mold is then removed from the surface and the surface of
the specimen is replicated by the surface of the plastic. A thin film of carbon
or metal like Cr, Pt is evaporated on the surface of the plastic. Sometimes
the evaporation is done from an oblique angle, shadow evaporation, to en-
hance the contrast. The plastic is removed by dissolving in a suitable solvent
and the film is then floated on to a grid for analysis.
A variation of the replica technique is the extraction replica technique. As
the name implies, it is used to extract and study precipitates embedded in a
matrix. The schematic of the process is shown in figure 19. The bulk sample
is etched in order to expose the precipitates. A C film is then evaporated on
the sample and then etching is continued until the bulk material is removed
leaving behind the precipitates. A TEM image of precipitates removed by
the extraction technique is shown in figure 20.

18
 MM3030: Materials Characterisation

Figure 18: Replica technique for sample preparation. Taken from Transmis-
sion Electron microscopy - Williams and Carter.

Figure 19: Extraction replica technique for sample preparation. Taken from
Transmission Electron microscopy - Williams and Carter.

19
 MM3030: Materials Characterisation

Figure 20: TEM image of sample prepared by extraction replica technique.

Taken from Transmission Electron microscopy - Williams and Carter.

4 Focused ion beam TEM sample prepara-

tion
Focused ion beam (FIB) instrument along with SEM is becoming increas-
ingly important in TEM sample preparation. The advantage of FIB-SEM
is that it is possible to prepare TEM specimens from precise locations in
the sample. This is not possible in other techniques since there is always
some uncertainty from where the final TEM sample will be obtained. One
area where the FIB is used extensively is in the microelectronics industry for
evaluation of defects in parts of the micro-chip. Here the area to be investi-
gated is of the order of tens of nms, so that conventional TEM preparation
techniques cannot be used.
The schematic of the 2 beam FIB-SEM is shown in figure 21. The FIB can
be considered as an ion beam miller except that Ga ions are used instead of
Ar ions in a conventional ion beam miller. The Ga ions are originated from
a liquid metal source (Ga melting point is 30 ◦ C). The beam is accelerated
and scanned over the specific portion of the specimen for milling. Modern
FIB-SEM also have gas-injection systems (GIS) for Pt deposition and a mi-
cromanipulator (Omniprobe) for removing the specimen from the sample and
attaching to the TEM grid. The steps involved in sample preparation are
1. Identification of the ROI.

2. Deposition of a Pt layer on top of this region. Pt deposition is done

20
 MM3030: Materials Characterisation

Figure 21: Schematic of the 2 beam FIB-SEM. Taken from Transmission

Electron microscopy - Williams and Carter.

using the GIS. The Pt deposition is done over a region few µm long
and wide and a few nm thick.

3. The Ga ions are used to ion mill the region around the Pt layer to create
a trench. Sample can be manipulated in-situ with three dimensional
translation and rotation possible. The Pt layer is also cut from below
to create a lamellar.

4. The Omniprobe is used to lift this lamellar from the sample and then
attach it to the TEM grid. The Pt metal is used to weld the lamellar
to the grid.

5. The final cleaning of the lamellar is done using the Ga ions. Final
thinning to electron transparency is also done.

The various stages in making the lamellar are shown in figure 22. The entire
TEM preparation is done in-situ and the SEM is used for imaging the sample
preparation. The FIB can also be used for imaging. The FIB-SEM instru-
ment and the electron and ion beam column is shown in figure 23. The TEM
sample before lift-off and the final sample attached to the grid is shown in
figure 24. The final sample has dimensions of the order of µm but its thick-
ness is of the order of nm, so that it is electron transparent. The FIB-SEM
technique is the most versatile of the TEM sample preparation techniques
since it can be used for both conducting and insulating samples. It also al-
lows for the preparation of TEM samples from specific areas of the sample.
But sample preparation and manipulation using the FIB-SEM is technically
challenging.

21
 MM3030: Materials Characterisation

Figure 22: Stages in creation of the lamella from a specimen are shown.
Taken from Transmission Electron microscopy - Williams and Carter.

22
 MM3030: Materials Characterisation

Figure 23: Dual beam FIB-SEM instrument with the close-up of the elec-
tron and ion beam column. Taken from Transmission Electron microscopy -
Williams and Carter.
23
 MM3030: Materials Characterisation

Figure 24: TEM lamellar before lift-off and the final sample attached to the
TEM grid. From P. Swaminathan - unpublished.

24