In this article, I’m going to talk about something that has nothing to do with software development, but something I enjoyed working on and would like to put down on paper before I forgot everything.
Indeed, I will talk about Electron Microscopy, a topic I worked on when I was at Thermo Fisher company. All the information here are taken from the excellent playlist of Professor Grant Jensen: Getting Started in Cryo-EM with Professor Grant Jensen that I strongly encourage you to watch if you work in electron microscopy field, or if you have a strong appetite for knowledge that is useless in everyday life.
What is an Electron Microscope ?
This is an electron microscope:
It is a big box that measure about 3.5 meter high, connected to an army of computing cluster, a big power plug, some water and air tubes and sometimes a liquid Nitrogen tank. And it can do SCIENCE ! You can literally count the atom composing a transistor or see a COVID-19 virus entering a cell with this machine. For example, below is a spike of coronavirus reconstructed from CryoEM acquisition. This is basically what is inside the vaccine and that teach your cells how to recognize the virus:
So, is it just a microscope with electronic parts ?
Yes, there are electronics in Electron microscope. But it is called Electron Microscope because, instead of using Photon like classical microscope you use in biology course, it uses Electrons. To understand why Electron are better than Photons, let’s have a look at how a classical microscope works:
Looking to an object is measuring how photons interact with this object. You light up the object from below and use a lens to scale up the resulting image. This works as long as photons are smaller than this object, because the size of a pixel of the resulting image cannot be smaller than the size of the photo that create this pixel.
The problem with photons is that they are much bigger than atoms. An Atom is ~1 Å, which is about 0.1 nm. A photon is ~4000 Å, so 4000 time bigger. On the other hand, an electron is ~0.01 Å, a hundred time smaller. And if you think about it, an atom is just a nucleus with electrons spinning around. So yes, an electron IS much smaller than an atom.
In the end we just have to build an electron flashlight to illuminate our object, an electron magnifying glass to magnify the image and an electron camera to take a picture, and we have an electron microscope 😎
Parts of an Electron Microscope
If we open our electron microscope, we have the Column, which is the heart of the system:
This column look like an optical microscope, except that it is upside down. The light is at the top, then you have a few lenses to diverge/converge the light, then the sample, another bunch of lenses and finally the sensor at the bottom.
2 parts compose the Electron gun :
- The cathode, which generates the electron.
- The accelerator, which accelerate the electron to the good direction.
Cathode is a very simple object. It is basically a tungsten wire bent into a V shape:
So we apply a negative tension, typically -300 kV. As an electron has a negative charge, it will be pushed back. On the other end, the ground of the Electron Microscope is kept at 0V potential, so electron will be attracted by the bottom of the microscope. Furthermore, to properly accelerate electron, they go through an accelerator which maintain a gradient of potential that smoothly go from -300kV to 0V:
By the way, an electron emitted with a 300Kv tension has an energy of 300 kilo electron volts, or 300keV. And current microscopes are able to generate electron one by one.
An optical lens converges and diverge light rays. It is the same for electron: it converges and diverge electrons “rays” except that electron lens are not made in glasses or transparent material. They are made with magnetic field generate by electromagnet:
These magnets heat when in use. And heating a copper wire change its resistivity, which change its magnetic field, which is not a good thing for us because it’s like changing the focus while taking a picture with a camera. So to avoid this, we have to cool down all these elements with water. Remember there is a 300,000 Volt electrical device operating a few inches away…
But how does an electron lens work ? If we compare to optical lens, you probably have seen in college that rays emitted from a point converge on a single point to form an image, with some stories of focal point, focal distance, etc… :
Actually, there is a reason why rays converge in this particular points. To understand why, we have to remember that light is a wave, so the diagram above should look like:
The interesting thing about waves is that they can add themselves if they have the same phase:
or subtract themselves if their phases are opposite:
It is the same for lights. Traversing a transparent material change the light frequency, and traversing different thickness change the light wave phase when it comes out of the lens. Also, when a wave comes out of a surface, it is emitted into every direction. So we can consider that every single point of the universe potentially have an image of the initial object, but for most of these points, waves arrive with the wrong phase, and they destroyed each others.
But there is a single point where all the light rays arrive with the same phase, and all contribute together to form the image:
It is the same for electron. Electrons are also waves, and when they pass through a magnetic field, we can change their phase like a lens do for light.
One thing to know about Electron microscope column is that they are maintained in a nearly perfect vacuum. Any “air” molecule can interact with the electron beam and decrease acquisition quality. In that case, how do you place the sample inside the acquisition chamber ? With an airlock of course 😉
On the side of the microscope, there is a tube with a valve where we can insert the sample holder:
The sample holder also contains a liquid nitrogen tank to keep the sample frozen during the whole acquisition.
EM Acquisition energy is measure in kilo-electron volt (keV). A typical CryoEM acquisition is done at 300keV. For various reason, it is very important that electrons that reach the detector really are at 300keV. We don’t really care if electron traversing the sample are not at the good energy, but electrons reaching the detector must be at 300keV.
To enforce this, there is a system placed before the detector called Energy filter, which goal is to filter electrons that are not perfectly at 300keV.
To do so, the energy filter will generate (yet another) magnetic field to bend the electron beam up to a diaphragm. Electrons with an energy of 300keV will pass through a hole in this diaphragm, whereas other electron will not follow the same trajectory and will crash into the shield:
Today’s detectors are able to detect electron one by one, and a raw electron microscope acquisition is far far away from beautiful colored images you can see on scientific papers. Indeed, a raw image looks like this:
Each white dot is an electron impact. But by accumulating lots and lots of images, we can obtain this:
And finally, with lots of mathematics and processing, we can obtain this kind of images that, in addition to being beautiful, really helps scientist and biologist to understand the deep behavior of nature: