Jump to ContentJump to Main Navigation
Electron CrystallographyElectron Microscopy and Electron Diffraction$

Xiaodong Zou, Sven Hovmöller, and Peter Oleynikov

Print publication date: 2011

Print ISBN-13: 9780199580200

Published to Oxford Scholarship Online: January 2012

DOI: 10.1093/acprof:oso/9780199580200.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2017. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 25 February 2017

(p.316) Appendix 9: Sample preparation and data collection

(p.316) Appendix 9: Sample preparation and data collection

Electron Crystallography
Oxford University Press

Sample preparation is a very important part of TEM studies. Many different methods are used for sample preparation. The choice of the methods depends on the sample itself and the problem we want to study by TEM. Ideally, the final samples for TEM should be very thin, uniform and without artefacts.

Several examples of different crystalline samples are shown in Figure A9.1, as SEM images.

A9.1 Are we looking at the compound we think?

Electron microscopy and especially TEM is unique in that we only look at extremely small particles. The advantage of this is obvious and has been mentioned in previous chapters; accurate structure analysis is possible on very tiny samples. But there is also a severe disadvantage; we may look at something else than we think we look at. Consider, for example, a sample consisting of 99% of rather thick lumps of the majority phase and 1% very thin plates of an impurity. The crystals of the major phase will all be too thick for HRTEM and ED studies. An example of how such a sample may look is shown in Figure A9.2. There is an obvious risk that all attention is then on the thin impurity, from which nice data can be collected. Note that in cases like this, it doesn't help to look at 10 or 20 particles to check for reproducibility! Since we know only thin samples are useful, we may well select all 10 or 20 particles from this minority phase.

It is always a good idea to check the sample also by SEM, powder X‐ray diffraction and EDS. Look in the SEM: is the sample uniform or do we have two or more phases? Check the chemical composition of the crystallites of different shapes and sizes – are they similar? The ultimate test is to compare the d‐values of the strongest peaks in electron diffraction with the strongest peaks in the X‐ray powder diffraction pattern. Only if they match well can we be confident that we know which phase is the majority phase. Sometimes, of course, we want to study not only the majority phase but also the other phase(s), but at least we must be sure to avoid mistaking a single small crystal from being representative of the whole sample. Somebody estimated that the (p.317)

Appendix 9: Sample preparation and data collection

Fig. A9.1 Some examples of different shapes, sizes and uniformity of crystals. The images are taken by SEM, with scale bars included. These are all zeolite‐related porous materials made in Xiaodong Zou's group at Stockholm University.

total weight of all the crystals looked at by TEM in the whole world until now is only a few micrograms!

A9.2 Nanocrystals

If the specimen is made up of tiny crystals, with one or two dimensions in the range 1–20 nm, the crystals can be looked at without any special treatment at all. They just have to be put on an EM‐grid with a carbon film. This is sometimes done by just pouring the powder over a grid and shaking off what does not stick to the grid. Another possibility is to make a sludge of the crystals in a solvent that does not dissolve the sample, put a drop on the grid and then shake off the excessive liquid or dry it off with a filter paper. (p.318)

Appendix 9: Sample preparation and data collection

Fig. A9.2 An example of a two‐phasic sample.

A9.3 Agglomerates: Crushing and grinding

Samples prepared by high‐temperature synthesis, for example alloys or metal oxides, end up as solid agglomerates of thousands of crystallites. Such samples are typically crushed in a mortar until there are many pieces that are thin enough, at least close to their edges.

For brittle materials such as ceramics and many minerals, the simplest way to make TEM samples is crushing and grinding them into small particles in an agate mortar with a clean pestle, preferably in an inert liquid, for example ethanol or butanol. A drop of the liquid containing the particles is placed on a holy carbon film supported by a grid. Most inorganic samples can be prepared in this way. One example is shown in Figure A9.3.

Appendix 9: Sample preparation and data collection

Fig. A9.3 A TEM image of a crushed metal oxide sample, the tantalum oxide Li2NaTa7O19. The crystal is quite thin near the edge, but within a few nanometres from the edge it increases rapidly in thickness, at irregular steps. For structure determination, the thinnest area should be used.

(p.319) The thinner the sample, the closer it is to kinematical conditions, so HRTEM images and ED patterns should normally be collected from the thinnest areas near the edge.

A9.4 Metals and alloys: Ion milling

Metals and alloys may not be brittle enough for successful crushing in a mortar. Then, the best sample preparation may be to thin a foil by mechanical polishing or even ion milling. Ion milling is use of a beam of accelerated atoms or ions (often Ar gas) to knock out the atoms in the specimen. It is time consuming and thus normally requires pre‐thinning of the sample. This can be done by first polishing the sample to about 100–200 μm thickness either mechanically, chemically or by electropolishing, then cut a 3‐mm diameter disk from the thin slice and finally pre‐thinning the central region to a few μm. Then, ion milling is performed on the pre‐thinned sample. This technique can be used for most materials, for example ceramics, composites, polyphasic semiconductors and alloys. It is especially useful if we want to study the cross‐section of an interface.

A9.5 Ultramicrotomy

The ultramicrotome is a slicing instrument for cutting samples into thin sections. In an ultramicrotome a firmly mounted specimen is moved against a fixed knife of diamond or glass. The final slices are collected in a liquid‐filled trough and are mounted on a grid. It is routinely used for biology and for polymers (soft materials), but it can also be used for hard materials.

A9.6 Other methods for sample preparation

Electropolishing The principle of electropolishing is that the specimen is made as the anode in an electrolytic cell. When a current is passed, the specimen (which has to be a conducting material such as a metal or alloy) is dissolved from the anode and deposited on the cathode.

Chemical polishing For chemical polishing a mixture of acids is used to dissolve the specimen. It can be used for non‐conducting materials.

Cleaving Layered materials, for example graphite, mica, etc. can be cleaved by attaching tape to both side of the sample and then pulling the two pieces of tape apart. The procedure is repeated until the specimen is thin enough. Then the glue is dissolved. Replication is to put a thin layer of carbon on the surface of the sample and then remove the sample and look at the carbon film in the TEM. It is used for studying surface topography.

Extraction is used for studying small particles extracted from the matrix samples. It is useful if one is only interested in these small particles and wants to make chemical analysis of them. (p.320)

Appendix 9: Sample preparation and data collection

Fig. A9.4 SEM images of the preparation of a TEM specimen by FIB. The sample is a core–shell composite, consisting of relatively large (~40 μm diameter) core crystals of zeolite X on which much smaller (~1 μm) shell crystals of zeolite A grow. a) The composite has the shape of an octahedron, while the shell crystals are needle‐ or plate‐shaped. With FIB it is possible to cut out a thin layer of sample at any desired place. Pt is first deposited on the surface area of interest to prevent damaging from the ion beam. A thin sample is then made by removing materials from its two sides by a focused ion beam. b) A tungsten needle is soldered to the thin layer for lift‐out. This image was taken after rotating a) by 90° around a horizontal axis. c) The thin sample was then transferred to a copper TEM holder. d) The tungsten needle was detached by FIB cutting after the thin layer was soldered onto the holder. After the final thinning, the sample is ready to be studied in the TEM. Images taken by Wilder Carrillo‐Cabrera in Wang et al. (2011). with permission.

A9.7 FIB – focused ion beam

A very sophisticated modern technique of specimen preparation is focused ion beam, FIB. It is possible to select a very tiny part of the sample inside a SEM and then cut out a thin slice (〈 1 μm) and transfer it onto a grid for analysis in the TEM. Usually, the sample is further thinned by for example ion milling to become transparent for electrons. An example of the different steps is shown in Figure A9.4.

A9.8 References

Dorset, D.L. (1995) Structural electron crystallography, Plenum Press, New York. Wang, Z., Gröner, D., Wan, W., Sun, J., Carrillo‐Cabrera, W., Zhu, G. and Zou, X.D. (2011) Epitaxial growth of zeolite X/A core‐shell composites Manuscript in preparation. Introduction to Focused Ion Beams: Instrumentation, Theory, Techniques and Practice, (2005) eds. Lucille A. Giannuzzi, and Fred A. Stevie, Springer, New York.