Aligning the detectors

Once we have corrected the Ronchigram and the pivots, then all we have to do is get the detectors lined up with the beam coming out of the bottom of the specimen. In a TEM/STEM, the STEM detectors are usually mounted below the phosphor screen, and normally you can't actually see them. Quite often, because of space constraints, they are not mounted on the centre line of the microscope, but are off to one side.

Ask the demonstrator: Please tell me roughly where the detectors are relative to the centre of the phosphor screen.

There are normally two detectors. A solid-state circular disc is used to collect the bright-field signal. Around it is mounted another solid-state detector in the shape of a circular annulus, which is used to collect all the dark- field electrons: that is, all the electrons which have scattered to a large angle outside the cone of the Ronchigram. Looking down on the two detectors, they look like this:

The STEM imaging will only work properly if we get three variables correctly adjusted: the camera-length (i.e. the effective distance or magnification between the specimen and the detector plane) and the x- and y-shift of the detector plane. It sounds simple, but there are a lot of things that can go wrong.

Suppose for a moment that the detectors are on the optic axis. What effect does changing the camera length have? Look at the next diagram.

At very short camera-lengths, the disc of the Ronchigram will be much smaller than the disc of the bright-field detector. This condition is good for dark-field imaging, because the annular detector is effectively lying at a high angle which is what we want for easy image interpretation. The image contrast is roughly proportional to the Rutherford scattering of electrons from the atomic nuclei.

As the camera length increases, the Ronchigram gets bigger and bigger, until it reaches a point where it covers all the bright-field detector. In this condition, the signals on both the dark-field and bright-field detectors are at a maximum, although the contrast on both images will be rather poor.

At even longer camera lengths, both the bright-field detector and the annular dark field detector is covered by the Ronchigram. In this condition, the bright-field image is very noisy and has low intensity, but it will have much more contrast and will be more like a conventional bright- field image in TEM mode.

However, for any one camera length, think of all the things that can go wrong if the detector is not properly aligned, as shown in the next figure:

Clearly, if the Ronchigram misses both detectors (which, by Sod's Law, is the most likely occurrence) then we will see nothing at all on the STEM monitors, because very few electrons will be hitting either of the detectors.

If the detector is almost aligned correctly, the central disc of the Ronchigram will hit the dark-field detector, but miss the bright-field detector. Under these circumstances, the signal displayed on the so-called dark-field monitor will be bright, and will look like a bright-field image, because it is collecting all the electrons which have passed through the specimen. Even worse, the so-called bright-field detector will be detecting high-angle scattered dark-field electrons, and so it will look like a dark-field image. Because the annular dark-field detector is so much larger than the bright-field detector, there are many more ways to get this inversion of signals to occur than to get the signals the right way round. For this reason, people often spend many happy hours thinking the monitors on the STEM are the wrong way round.

Remember: If the detectors are mis-aligned in STEM, their signals will appear to be the wrong way round, and there are more ways for this happen than to get the signals the right way round.

Experiment: Line up the Ronchigram as well as you can, focussing it with the objective lens. Adjust the camera length so that the Ronchigram is a small disc (ask the demonstrator what 'small' means in this context - ideally about the size of the bright-field detector). Adjust the detector align so that, roughly speaking, the beam is hitting the phosphor at the place where the demonstrator told you the detectors are mounted. De-select the alignment page. Lift the phosphor screen and start the beam scanning. Select the very lowest STEM magnification possible.

Ask the demonstrator: How do I adjust the monitors to get their grey-levels, gain and offset roughly right? Which detector corresponds to the bright-field image and which is the dark- field image. How do adjust the gain and offset of the detectors?

Note that you should the gains and offsets of the display unit right first, before you begin playing the gains and offsets of the detectors themselves

The chances are that when you see the STEM images, their contrasts will be the wrong way round, or they will both be bright or both dark. As a first step, get the dark-field detector lined up. To get a good annular dark-field image, you must have a sufficiently short camera length so that the cone of the Ronchigram is small enough to fit inside the dark field detector. Start off with a small camera length and a small condenser aperture (which is best aligned, as before, in the Ronchigram). Observe the dark-field STEM image at low magnification.

Ask the demonstrator: How do I adjust the detector alignment?

Adjust the two knobs that control the x-y displacement of the detector plane, keeping your eyes on the dark-field image. As you go through the full range of offsets on one of the knobs, one of three things will happen:

• the image stays dark;
• it starts dark gets bright and then gets dark;
• it goes dark-bright-dark-bright-dark.

The second two situations correspond to the Ronchigram moving over the detector as follows:

If you never get the dark-bright-dark-bright-dark combination, it means the camera length is too long, so the disc of electrons can't fit inside the inner diameter of the detector. What we want is to be in the central dark region of the dark-bright-dark-bright-dark range of the adjustments. Doing this blind - i.e. by just watching the dark-field image - sounds difficult, but is easy to learn when you have a picture in your mind of what's going on. Of course, the bright-field image will become bright when you the dark-field detector is on line, which is a useful hint.

Aside: You can actually form an image of the STEM detectors on the scanning screens by doing one of three things (in order of drasticness):

• put the objective out of focus,
• under-focus the diffraction lens
• wreck the pivot points.

Draw the ray diagrams to work out why all these things will cause the Ronchigram to scan across the detectors, and thus form an image of them.

To perfect the image, you should now focus with the objective pre-field (i.e. focus the objective) and stigmate the image by adjusting the condenser stigmators. You can also re-adjust the objective rotation centre using the dark- field image, but this correction will only be approximate if the condenser aperture is not exactly on line. The best place to get the probe-forming system lined up is in the Ronchigram.

To get the bright-field detector accurately on-line at high camera length, wobble the objective lens (i.e. align the rotation centre) and adjust the detector shift until the bright-field image is stationary. Be warned that this only works if the condenser aperture, condenser lens and objective lens is already accurately aligned in the Ronchigram. Some computer-controlled microscopes will not let you do this, and so you may have to wobble the objective by manually defocussing it back and forth.

The ideal resolution can only be achieved if you have the correct convergence angle relative to condenser aperture size. In STEM, the biggest influence on very high resolution is the demagnification of source. Even in our very first experiment, we saw that when we form a probe in the specimen plane, what we actually see is an image of the source. If source is big, then we have no hope of achieving good STEM resolution. We can make the effective source size small by demagnifying its image, which we know we can do by increasing the spot size. But as spot size goes up, intensity - or the number of electrons going through the probe - goes down. STEM images are pretty noisy and hard to see at the best of times. Good STEM resolution can therefore only be obtained with a very bright source, i.e. by using a field-emission gun.

Copyright J M Rodenburg