I have to admit that, prior to owning an SEM I had only a basic understanding of how they worked; I knew it involved scanning an electron beam across a sample and this in turn necessitated a high vacuum, but that was about it. Fortunately, it turns out that the fundamental concept is not that much more complex, assuming that you don't need to concern yourself with the associated mathematics.
The sample is placed into the sample chamber and above this sits the electron column, at the top of which is an electron gun. The source of electrons is typically a heated tungsten filament, although it can also be a more exotic material that has a lower work function and will therefore emit electrons with reduced energy input, such as lanthanum hexaboride. One key benefit of this being that there is then less energy in the beam, which is good news when imaging more sensitive subjects.
The electron beam is focused by apertures and condenser lenses in the column, which instead of being glass as you might find in a light microscope, are instead electromagnets. Together these determine the size of the resulting spot which will be scanned across the surface of the sample.
It turns out that there are a number of different interactions which can be analysed as the electron beam scans the sample, with the main ones being:
Secondary electrons ejected from sample atoms due to interaction with the beam
Electrons from the beam that upon hitting the sample are backscattered
X-rays produced by the interaction of the beam with the sample
The first two are used for imaging physical features, whereas the X-rays can be analysed to ascertain the abundance and distribution of elements in the sample.
The beam of a display and/or record CRT is synchronised with the beam scanning the sample, the intensity of which is determined by the signal from the detector(s). Of course, there are lots of variables associated with the control of the beam and configuration of detectors, which determine things such as resolution and focus, along with optimising for the particular sample or task.
Main system components
The main cabinet can be seen above, with no shortage of controls! The front panel has been removed from the lower half, exposing the electron gun EHT power supply at the top-left, with a large smoothing capacitor PCB for the main multi-rail linear PSU hidden behind a cover below this. To the bottom-right and at the back it's just about possible to make out a very large toroidal transformer, above which is a rack shelf with power supplies for a CRT and photomultiplier tube.
The main controls are towards the right-hand side of the control panel, just below the two CRTs. The left-hand CRT is the “CCTV” display used for setting up the instrument and locating the area of interest, displaying live video with a TV rate, but of a lower quality. The round CRT to the right is surrounded by a hood onto which a camera can be attached. With the machine switched into a slow scan mode this can be used to capture images with a resolution of up to 4,000 lines.
The panel to the left of the round CRT, with the quadrant illustration and two rotary controls, is for configuring the backscattered electron detector. The controls below the CRTs are for aligning and focusing the beam, configuring the resolution, setting the filament and beam currents, EHT voltage, operating mode and so forth.
Above we can see the electron column and below this the sample chamber. The arm sticking out from the left-hand side of the chamber holds the X-ray detector, which in use is cooled with liquid nitrogen in the dewar vessel above marked “Cambridge”. On the front of the sample chamber door we can see three manual rotary controls for sample stage X, Y & Z position. Just above which are two smaller controls for setting the stage tilt and rotation angles.
Just visible, sticking out the back of the chamber, is the secondary electron detector.
Directly beneath the chamber and connected to it via a large port is a turbomolecular pump, which is used to achieve a high vacuum in the chamber. However, this in turn needs to be “backed” by another pump, which in this case is a rotary vane vacuum pump located to the left.
The photo above shows the sample chamber with the door open, onto which is affixed the sample stage and mechanisms that position this, with the vacuum port to the aforementioned turbomolecular pump being clearly visible in the bottom of the chamber itself.
This particular machine had also been fitted with an electronic stage positioning facility, the joystick control and main electronics units for which can be seen above. This is used to control stepper motors connected to the X, Y & Z mechanisms — up to a certain resolution, beyond which it can be switched into a mode whereby it steers the electron beam electronically, since at the highest magnification levels mechanical sample positioning is insufficiently fine grained.
There are also other add-ons, but these will be covered in future posts.
A good starting position
After receiving the SEM from its previous owners, a few weeks of serious tinkering and a steep learning curve followed — aided, thankfully, by plentiful documentation and the generous remote help of a veteran SEM engineer, who also happened to know this particular machine.
Once we'd learnt what the controls are for and which we needed to concern ourselves with, and also where the unlabelled cables should be connected, this was followed by lots of removing PCBs, cleaning tarnished edge connectors and then reinserting them, plus the odd less obvious procedure.
Those interested can find further details in a previous blog post. However, although we then had a (mostly) working machine, this was really just the start of the journey, since:
There was an intermittent fault whereby vertical scanning would cease to work properly and text would no longer be displayed on the CRT
The image was fuzzy at higher magnifications
The PSU was filled with large electrolytic capacitors which were way past their best
There were numerous cooling fan related issues
We had no idea when the oil had last been changed in the vacuum pump
A number of controls were loose or faulty
There was no digital image output
The first three are quite possibly related and we'll come on to this later. The final point is really an enhancement, given that the SEM is for all intents and purposes an analogue machine. Nevertheless, recording an image by clamping a camera onto a CRT simply isn't acceptable these days, even if using a digital SLR; the requisite signals are available and can be digitised, resulting in greater convenience and a solution that doesn't involve needlessly turning these into light and back again.
In the next post we'll take a look at refurbishing the cooling system and PSU capacitor PCB.
Open source (hardware and software!) advocate, Treasurer and Director of the Free and Open Source Silicon Foundation, organiser of Wuthering Bytes technology festival and founder of the Open Source Hardware User Group.