Lens geometry in STEM mode on a TEM

In normal TEM mode, we have two cross-overs below C2, the lower one occurring within the objective pre-field, as shown in the elementary guide. One way of forming a probe would be to de-excite C2, as we did in the very first experiment in first section of the TEM guide, but then we still have a second cross-over which is happening somewhere odd within the pre-field.

Greater flexibility and an optimised probe focus can be obtained by avoiding the cross-over within the objective pre- field. In practice, this is achieved by either lowering the overall objective excitation or (depending on the make of microscope) using another small lens within the upper part of the objective twin lens, which is called either a mini-condenser or mini-lens. When the machine is being run in normal TEM mode, this lens is run in the same sense as the objective, making the pre-field very strong. When STEM mode is selected, the mini-lens is reversed (or switched off), and so it cancels out the pre-field to a certain extent.

The next Figure illustrates conventional TEM imaging:

where the top lens is C2, and the two lower lenses represent the objective prefield (and/or the objective lens pre-field boosted by a condenser mini-lens or objective mini-lens).

Now when we go into STEM mode, most microscopes reduce the influence of the pre-field, so now the ray-diagram looks like this:

Note that both C2 and the objective lens affect the focus of the electron probe. Because C2 is so weakly excited, you can see from the figure above that it will have a huge effect on the convergence angle of the illumination at the specimen plane. For this reason, many manufacturers fix the value of C2 in STEM mode. To obtain control of it, the user has to override the computer.

Remember: In STEM, the only electron optics that matter all happen before the specimen. Alignments you made relating to TEM imaging, especially the objective focus and stigmation, are useless in STEM mode. We must alter the objective lens (thus altering the pre-field above the speciment) and C2 to control the alignment. The relevant stigmators for STEM mode are the condenser stigmators, not the objective stigmators.

Now that we know that two lenses are involved in the forming probe, we can work how to line them up with one another. The most sensitive mis-alignment arises from the double- deflection coils being on the wrong tilt setting. If the beam tilt is wrong, the beam from C2 passes into the objective pre-field off-axis and at an angle. It will come out below the specimen also at the wrong angle: this is most common reason for losing the beam in STEM or nanoprobe mode at high camera lengths.

The first thing to get right in STEM mode is therefore the objective rotation centre. The alignment is almost certainly not the same as for TEM, because the objective is in a completely different state, and we are aligning with respect to the pre-field, not the main body of the lens below the specimen. Similarly, the condenser aperture is almost certainly off-line, even if you lined it up in TEM mode.

All the corrections are best done in the electron Ronchigram.

Experiment: Wobble the objective lens, and try to get the image moving in and out symmetrically with the condenser (we assume here that this is on line). As you alter the beam tilt (this should be the correction you are adjusting on the multi-function knobs when aligning in STEM mode), the whole Ronchigram will at the same time bodily move across the phosphor screen. That's because the conical beam cast by the condenser lens is rocking like a lever through the specimen, and moves all over the far-field plane, like the sweep of a torch beam. Start with a large condenser aperture. The aim is to get a stationary shadow image at the very centre of the disc you can see on the screen.

Just get the rotation centre roughly right, stop the objective wobbling, and then change C2. (Remember that on most machines you may have over-ride the C2 setting). Is the Ronchigram going in and out symmetrically around a point at the centre of the condenser aperture? If not, shift the condenser aperture onto that axis. If in any doubt at all, leave the condenser aperture where it was correct for TEM imaging. You will not get a perfect STEM image, but neither will you lose the beam, which is very easy at this stage.

Now go back and wobble the objective lens again. When both these centres are roughly co-incident with the condenser aperture, set C2 at the value you are going to use it in STEM mode.

Focus the objective to get the Ronchigram to 'blow up' into infinitive magnification at its centre. Adjust the condenser stigmators so that the Ronchigram changes symmetrically as a function of defocus. This takes some practice. (Remember that the objective stigmators, which are below the specimen, are useless in STEM mode, although you should try to set them roughly right in TEM mode before switching to STEM mode.)

If in doubt, wait until you are in scanning mode and can see an image on the scanning monitors before you attempt to correct astigmatism. However, note that if the stigmators are wrong and you change them later, all the above alignments should be re-iterated if you want the best probe possible. Misalignment, tilt, astigmatism, defocus and spherical aberration (which is the thing that makes the Ronchigram look the way it does) are all just lens aberrations that add up and affect the focus of the probe. Change any one of these variables, and another one will no longer be optimal.

Whatever you do, don't try turn C2 so high that what you see on the phosphor screen reaches a focus. This is bound to look astigmatic - but don't correct the condenser stigmators in this condition. Set C2 at its STEM mode setting. Now focus the objective lens: in other words, change the focus until the Ronchigram magnification is a maximum (i.e. the intensity distribution 'blows up' at its centre). If there is astigmatism present, the Ronchigram will not 'blow up' into a uniform infinite magnification blob. Instead, it will look stretched in one direction at one setting of objective defocus, and stretched roughly at right-angles at another setting of defocus. Set the objective between these two extremes, and then correct the condenser stigmators, turning them so that whatever you see spreads more uniformly over the Ronchigram. If the stretching of the image gets more extreme, you're turning the stigmator the wrong way.

As a general rule, for the best scanning probe resolution, you should select a condenser aperture size that just selects the middle 'flat' area of intensity in the Ronchigram when it is just in focus. 'Just in focus' means that as we decrease focus from above the beam cross-over setting, the very centre of the Ronchigram has just exploded into a blob, but has not yet become a clear inverted shadow image. Because changing spot size changes the value of C2 as well as C1 (in STEM mode), and this affects the angle of convergence at the specimen, the size of the condenser aperture needed to fulfil this condition may also change.

Copyright J M Rodenburg