Skip to main content

SEM Adventures Pt. 3: Oil Change, New Filament and Testing

 

Completing the initial overhaul of 30+ year old scanning electron microscope.

No car service would be complete without an oil change and the same could be said to be true with a scanning electron microscope, or rather it's rotary vane vacuum pump. In this post we look at changing the oil in this, together with fitting and aligning a new tungsten electron gun filament, before finally testing and capturing the first high resolution images.

Backing pump oil change

The electron column plinth can be seen above and immediately below this, since the cover has been removed, we can just about see the turbomolecular vacuum in the background, behind the white panel. This is responsible for creating a high vacuum in the sample chamber, but it cannot operate with the exhaust port at atmospheric pressure and must be backed by another vacuum pump.

To the left of the column cabinet is pictured a rotary vane vacuum pump, which serves as the “backing” or “roughing” pump, sitting between the turbomolecular pump and atmosphere.

Upon inspection the oil visible in the sight glass was of a fairly dark colour and it wasn't clear when this had last been changed, so as a precaution this was drained and the pump filled with new oil.

Aside from the potential for causing increased wear and tear on the pump itself, it's also possible for contaminated oil to result in a lower ultimate vacuum.

At this point it should be noted that the turbomolecular pump should also be serviced at regular intervals. However, it was decided against changing more than one thing at a time in the vacuum system, in case of problems arising. Also the turbo was spinning up to full speed and getting to grips with removing this, dismantling and servicing it, is an adventure that can wait for another day.

Vibration is an enemy of microscopy at higher magnifications and given that the rotary vane pump is located just beside the column, it was decided to stick foam pads to a piece of wood and sit the pump on this, in a bid to help reduce vibration being transmitted through the floor.

New filament

The cap at the top of the electron column is hinged and, providing it's not under vacuum, this can be swung open to provide access to the electron gun, pictured above on the left.

Unfortunately, a tungsten filament has a finite lifetime and while this may be affected by things such as the magnification level and beam current, it's probably only going to be about 20 hours or so of use. Filaments are consumables and are obviously designed to be changed without too much trouble, but it's not exactly a fun process and you do worry about losing or damaging parts of the gun — particularly when sourcing replacement parts could prove next to impossible!

First the top section is removed from the gun and then an inner assembly with the filament secured is removed from within this, both of which can be seen above.

It's just about possible to observe that the old filament had a break at the top-right.

Once the filament had been replaced the assembly was put back together and it was time to centre the filament in the aperture, via the aid of grub screws. This was achieved using a stereo microscope — yes, a light microscope to fix a scanning electron microscope! — and a little patience.

It's important to note that nitrile gloves must be worn whenever handling parts of the electron, so as to avoid contamination, which can easily affect operation.

New filament current potentiometer

If any potentiometer was going to degrade and become noisy, just about the worst for this to happen to is the one for filament current control, which would otherwise allow you to gently increase and decrease current so as to avoid thermal shock which may reduce filament lifetime.

The old, somewhat jumpy, potentiometer can be seen above with the main control panel slid out of the cabinet ready for it to be replaced.

As luck would have it, RS stocked a replacement which would fit and with the appropriate resistance and power rating, albeit based on much more modern cermet technology instead of being wirewound. This should hopefully give many years reliable service.

Testing!

With a refurbish PSU, oil change, new filament and replacement potentiometer, it was time to fire up the machine and this can be seen in the above video.

First high resolution images

A friend kindly leant us a Canon DSLR, which meant that we could attempt to capture the first high resolution images by clamping this onto the 4,000 line CRT (all our cameras are Nikon fit). Following which switching the SEM from video mode into “normal raster” mode, whereupon a record frame duration of between 15 and 2,000 seconds can be selected and a button pressed to capture an image. Initially experimenting with capture times of 15, 30 and 60 seconds/frame.

Watch gear tooth detail

There's room for a lot of improvement, but these first high resolution results have been encouraging — and there is no shortage of controls to endlessly fiddle with in an effort to improve the quality.

Next steps

Clearly, in this day and age capturing images by pointing a digital camera at a CRT is far from optimal. So the next job will be to see if we can achieve direct digital image acquisition. Fortunately for us the video signal, along with signals for line and frame sync, are all easily accessible.

For easier prototyping we put together a laser cut acrylic enclosure with BNCs on the back, binding posts on the front and a breadboard just below these.

As mentioned in a previous post, the machine did also come with a primitive single frame store and image processor, which may be faulty and isn't really worth the effort trying to get to work.

However, the image processor interfaced with the machine via a control box fitted inside the main cabinet, which also has hooks into other parts of the SEM, e.g. for record/capture control. As such this may help with enabling more complete integration, such as for triggering capture.

Fingers crossed we'll have some initial direct capture images to show in the next post!

Andrew Back

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.