There are a variety of sample stages used in SEMs; pretty much every manufacturer has one or more unique sample stage designs, and most of those stages will take more than one type of sample mount. That said, the most common type of sample mount, or stub, consists of a circular disk with a 3mm/0.125in cylindrical rod coming off the bottom. Other stub types I've used have included simple cylinders (JEOL, CamScan), and thickened disks with a tapped hole (Hitachi). Once you learn what the sample stage accepts, it is sometimes convenient to even make your own holders. The figure below shows a variety of home-built sample mounts.

When mounting (that is, physically attaching) your sample to the stub/holder there are a variety of options. A very common sample mounting approach is to use conductive tape. The ubiquitous 'carbon tape' is essentially a graphite felt impregnated with adhesive. The graphite makes it conductive, while the adhesive adheres the sample to the mount. As might be imagined, this is an extremely easy mounting method: just cut and stick, and it will work with almost any sample geometry. Unfortunately, the disadvantages include contamination, leaving a mess on the sample/mount if either is to be re-used, and the possibility of your sample 'creeping', or slowly sliding downhill if the sample/stage is tilted. For many single-use samples in a 'scope of mediocre cleanliness requirements, however, this is the best choice.

For a re-usable sample mount, often a clamping system is a good solution (see figure). The advantages of this approach are much lower contamination levels introduced to the microscope and sample by the mounting media, quick cleaning, and quick sample swapping. There is also the major advantage of being able to design the holder for the job. The tilted clamp in the right of the figure is a case in point, providing a reproducible, pre-determined sample tilt quickly.

A few of the sample holders I have made: several single clamp mounts, a TEM foil holder for SEM [front center]; one multi-clamp mount [left], and a tilted, bend-loading holder [right]. For instructions on how to make these vices, click here

It is often a good idea to plan ahead for SEM work when considering whatever other experimental work is going to be done with the samples eventually destined for the SEM. The ideal sample is one that is conductive, will not degrade from sitting around, is robust, is small enough to easily mount and fit in the chamber, will not outgas in the vacuum system, and will not suffer beam degradation; everyone else just has to make do. It is a good idea, however, to try and knock a few problems off in the planning stages, before your imaging headaches begin. Ask yourself if you really need to mount it in Bakelite before polishing (outgassing from hell), or if your tensile samples simply must be 24 inches long (it simply won't fit).

SEM chamber sizes vary considerably, so you need to know the peculiarities of the SEM in question before loading, and even better, while fabricating your sample. Most any SEM will accept a 1cm x 1cm x 1/2 cm sample, but often it is desirable to load a much larger sample so as to avoid preparation artifacts, or to load multiple samples in a single holder so as to avoid swapping out of multiple samples.

After the sample is prepared and attached to a stub with a standard interface, the sample must be loaded. Older SEMs, or those with simple chamber designs (RJ Lee Personal SEM) typically have only one chamber door, and are loaded by venting, opening the sample chamber, and loading the sample. This has the advantage of being the most flexible loading scheme, but pump down times are considerable due to the exposure of the chamber surfaces to air, from which water is adsorbed. Desorption of the water tends to dominate the pump down time. Most modern SEMs also have a sample introduction airlock, through which smaller mounted samples can be introduced into the chamber without breaking vacuum. These airlocks tend to be small and pumped only by a roughing pump, so samples introduced this way need to be small. Sample loading times are shorter through an airlock. Because of the airlock rough vacuum, outgassing samples will still prove a problem in the main chamber, and act as the limiting factor in pump down time.

Pure thermionic emitters are typically kept either at room temperature or at some temperature below the operational temperature. This is because the evaporation, diffusional effects, and reaction with residual gas species increases with temperature. Gas adsorption to the emitter is significantly reduced or eliminated by keeping it 'warm'. Many of the hexaboride-based electron guns are maintained this way, both to reduce total start-up time, as well as to reduce vacuum pressures in the gun during start-up.

Standard start-up involves applying the accelerating potential to the gun, then warming up the emitter. A typical beam current-emitter temperature plot for a self-biasing electron gun is shown in fig [heating]. The total emission is essentially nil until a little below the filament operating temperature. It is in this range that the 'warm' emitters are kept. A little above the temperature where emission really gets going there is typically a sudden increase in the beam current, followed by a sharp drop at a slightly higher temperature. This 'false peak' is followed by an increase in current at higher temperatures, until a plateau is reached. Normal operation with good filament life is typically found at a temperature a little below the plateau temperature.

Many field emission systems actually involve a lengthy start-up procedure for the initial start-up of the electron gun, followed by leaving the emitter in a standby mode. The standby mode is essentially a condition from which the emitter can be brought up to full emission quickly by a built-in routine in the microscope. The transfer from standby to full operation for these machines is then quixoticly the fastest, on the order of a few seconds.

Once the basic operating conditions are established (accelerating voltage, working distance), then the gun and column need to be aligned. If the accelerating voltage is changed during operation, then some (or all) alignment steps may need to be repeated. Learning to recognize the characteristics of the various misalignments and beam distortions allows an operator to quickly fix what needs fixing and not disturb what's OK.

For the thermionic emitters, gun tilt is most simply set by viewing the SEM image and adjusting the gun tilt controls until the highest image brightness is obtained. Some SEM manufacturers will also have special scan modes so that the gun tilt can be adjusted in a manor analogous to in a TEM, but the 'brightest image' criterion will always work, and in many cases is simpler.

Field emitters are another story altogether. I don't have enough experience with different field emission SEMs to even try and make any generalizations about this; all I can say is that every SEM worth the money paid for it comes with a manual that describes the start-up procedures, and the gun tilt adjustment should be included in there. Sound like a cop-out? You betcha. Also sound advice, unless you have a spare half million kicking around to buy an extra 'scope to experiment with.

To align the aperture, first find a reasonably small, roundish feature in good contrast to the background on your sample, and get it in focus. If the aperture is off-axis from the condenser lens, then as the focus is adjusted the image of the feature will appear to move across the screen as it goes in and out of focus. Most SEMs will have a control that automatically strengthens and weakens the objective in a periodic fashion, typically labeled 'image wobbler', 'focus wobbler', 'aperture align', or some variant on these. With the 'wobbler' on, the aperture position is adjusted so that the feature image goes in and out of focus around the same point on the screen. This corresponds to an aligned aperture.

Some (OK, only one that I can think of) SEMs (the RJ Lee Personal SEM in particular) have only one aperture, so aperture alignment is not included in the normal operation of the SEM, but rather in the start-up after filament change.

Gun shift:
Gun horiz
Gun X-Y

Astigmatism is when a lens has different strengths as a function of the rotation about the lens axis. This ends up giving the lens two line foci instead of a single point focus. The line foci of the astigmatic lens are perpendicular to each other, and lie at different focal lengths from the lens. When forming an image in the SEM, astigmatism is seen when going through 'focus' where the image seems to be 'streaked' in one direction, then at a different focus to be streaked in a perpendicular direction. There are two different schools to stigmation: the first (my favorite) is to find a line focus, use the stigmators to 'squash' the streaks until they are simply out of focus, re-check the focus/astigmatism condition, and repeat. The second method is to find the focus point equally between the two line foci, 'focus' the image using the stigmators, re-check the focus/astigmatism condition, and repeat. Some people can use one method easily and the other only with difficulty, so if one approach isn't successful, the second might be. In order to the use the first method, near-simultaneous control over both stigmators and the focus control is needed, which makes it more difficult to use for SEMs that have gone to pure computer control (RJ Lee P-SEM). For both methods, independent control of each stigmation direction is very useful, leaving the computerized operator interface rather lacking.