This is intended as an assistance for any microscopist interested in doing ECCI. Despite the apparent difficulty involved in early ECCI work [1], for a properly equipped microscope, ECCI is remarkably easy to do. Note the important phrase, 'properly equipped'; this (I suspect) is the major reason ECCI is not as common an experimental method as Z-contrast in the SEM. The essentials for ECCI include:

• A high brightness electron gun. (Some type of field emitter is recommended, although for lower resolution work a LaB₆ gun is sufficient.[2])
• A small beam convergence angle. I have typically used a probe convergence angle of very approximately 6 mrad (2a). The contrast doesn't drop off sharply for larger a, but channeling contrast does get worse for higher a. Joy [3] recommends a<10 mrad from a simplified theoretical treatment, which seems a decent approximation.
• A backscattered electron detector with a large solid angle of collection. Much of my work so far has used a polepiece mounted Si diode detector with ~0.6p steradians.
• A clean, smooth sample surface free of preparation artifacts. This means minimal layers of surface contamination, and no alteration of the near-surface crystallinity through polishing, etc. An surface that is (optically) mirror-smooth will not work if there are sub-surface dislocations that have been introduced by polishing. [4] (At least if you are interested in the base material instead of the polishing artifacts.)
• Some way of collecting channeling patterns to align your sample to the channeling contrast conditions. If your sample is a large single crystal, than the regular scanning action of the electron beam at low magnification will be enough to yield channeling patterns, however most polycrystals will require selected area channeling capability to collect selected area channeling patterns (SACPs).
• A sufficiently high probe current to resolve the channeling contrast above the detector noise. (A rule of thumb I use is at least 1.5 nA.)

Some things which are also useful are:

• A sample stage that is eucentric for at least one tilt axis. (This allows you to tilt to the channeling contrast conditions without chasing the area of interest all over the sample, which gets old fast.)
• Fine control over the stage tilt angle.

Also, I recommend:

• Avoid carbon tape, samples mounted in Bakelite/epoxy, and other sources of hydrocarbon contamination around the sample. The contamination that builds up on the sample surface reduces the coherency of the beam and obscures the channeling contrast. The scan square build-up due to the high probe currents is often considerable. I recommend mechanically mounting the sample in some sort of vice fixture. [5] This has the added advantage of being easily reusable.

Although historically people have used a high sample tilt angle for ECCI, this involves some problems including foreshortening of the image, greater constraints on sample size and maneuverability, and some sort of custom BSE detector. The constraints come from operating near the end of the tilt range available in most microscopes, and the custom BSE detector because of the need to get high collection efficiencies at high sample tilts. Typically, this involves some sort of retractable detector that is snuggled in close to the sample surface, which in turn restricts sample motion even further.

In contrast, I would recommend the low tilt configuration [6], which places the sample approximately normal to the beam axis (allowing sample sizes limited more by the chamber and stage size) and uses commercially available polepiece mounted BSE detectors.

For whatever sample geometry chosen, the procedure is as follows:

1. Prepare your sample in such a way that there is minimal disruption of the near-surface crystallinity. This usually means electropolishing or chemical attack. A further caveat is that topography must be suppressed so as not to overwhelm the comparatively weak channeling contrast with topographic contrast.
2. Mount your sample in the microscope so as to maximize your BSE detector collection efficiency, minimize your beam convergence angle, and still allow room to move your sample + stage without destroying anything.
3. After aligning the microscope and adjusting your sample to the eucentric height, choose a grain or area of interest. Collect the channeling pattern/SACP from this region. Assuming your sample 1) isn't heavily deformed, 2) isn't dirty, and 3) isn't overwhelmed by topographic or Z contrast, then you will get something like figure 1.
4. Tilt your crystal so that the microscope beam axis (the center of the channeling pattern) is on the edge of the band for your desired channeling contrast condition, as is marked by the cross in fig. 1.
5. Return to image mode (if using SACP), raise magnification, and focus. If there are channeling contrast features to be seen under these channeling conditions (and your contrast is sufficient), you will have an ECC image, as exemplified in figure 2. (Note: often you will need to use a very long frame time, or recursive filtering to bring the image out of the noise. A very high BSE gain is essential!)

Figure 1 shows a SACP from polycrystalline g-TiAl, covering roughly 15^o of tilt; the central 10^o or so is from a single grain, while the remainder is from the surrounding sample.

Because of the way the SACP is captured, the microscope beam axis can be thought of as the orientation described by the center of the SACP (indicated by the cross in the figure), so a BSE image of this crystal formed by a beam parallel to the microscope axis would be expected to have a greyscale intensity equivalent to the point appearing at the center of the SACP. This is seen in the channeling contrast image (which is simply the BSE image) taken under the channeling conditions of figure 1.

Figure 1: SACP from polycrystalline g-TiAl. The cross indicates the approximate microscope axis.	Figure 2: ECC images of a deformed grain of g-TiAl taken under the contrast conditions in fig. 1. Note the dislocations appearing as bright and dark specks in b.

The grey level of the background in the image corresponds to the grey level along the beam direction in the SACP. Note that for the low magnification image (fig. 2a), the beam rock due to the scanning action is sufficient to move the beam position far enough off the microscope axis to change the background grey level. At low enough magnification this is seen as a small portion of the channeling pattern. (In fig. 2a the channeling pattern fragment is inverted and rotated relative to the SACP. This is due to the difference in electron optics used between the two modes.)

Copyright 2001 Benjamin Andrew Simkin

REFERENCES

1.	Electron-channelling imaging in scanning electron microscopy P.Morin, M.Pitaval,D.Besnard,G.Fontaine, Philosophical Magazine A, 40, 4, pp. 511-524 (1979)
2.	Electron channelling contrast imaging of interfacial defects in strained silicon-germanium layers on silicon A.J. Wilkinson, G.R. Anstis, J.T. Czernuszka, N.J. Long, P.B. Hirsch, Philosophical Magazine A, 68, 1, pp. 59-80 (1993)
3.	Direct Defect Imaging in the High Resolution SEM D.C. Joy Mat. Res. Soc. Symp. Proc., 183, p. 199 (1990)
4.	Use of electron channelling contrast imaging to assess near-surface mechanical damage B.A. Simkin, M.A. Crimp, Philosophical Magazine Letters, 80, 6, pp. 395-400 (2000)
5.	Re-Usable Vice-Style SEM Sample Holder for Pin-Mount Stages B.A. Simkin Microscopy Today, 01-4, p. 3 (2001)
6.	An experimentally convenient configuration for electron channelling contrast imaging B.A. Simkin, M.A. Crimp, Ultramicroscopy, 77, pp. 65-75 (1999)