Scanning Electron Microscope

Design Process

From my research, I have gathered that a SEM is generally composed of an electron gun (thermionic), focusing lenses, scanning lenses that control scanning of the beam, and a detector for various energies of electrons. In my design, I will only focus on detecting the lower energy (<50eV) secondary electrons for simplicity. I have decided to use magnetic lenses, as the focusing effect scales with the velocity of the electrons (and hence, the energy), in contrast to electric deflection plates.

​

I have attempted to create a computation simulation in Python of a simple quadrupole and sextupole magnetic focusing lenses with limited success. I assumed that electrons that are thermionically released from the tungsten maintain their Fermi energy (I approximated this as ~7eV) and allowed them to have an initial velocity with magnitude that reflects this (the true vector values are randomly determined). The position of these electrons is randomly determined to allow the extent of the focusing of the lenses to be explored. In the simulation I present, I ignore the Fermi energy and only account for the energy the electron picks up in the electric field of the accelerator mesh. In some simulations I managed to produce a semi-focusing effect, but I was limited by the computational power of my computer and my the inability to simulate magnetization of iron cores. I will experimentally determine the beam properties when the prototype is constructed. I would like to have access to professional electromagnetic simulation software like CST to do this properly. The simulation code may be found in the Files section. Results can be viewed below.

​

My electron gun will be composed of a tungsten wire bent into a sharp "V" shape (vertex pointed towards specimen stage) and an accelerator mesh biased with ~+10kV to begin. I will adjust this voltage as needed.

​

The focusing lenses will be magnetic quadrupoles composed of four coils attached to an iron core to direct the magnetic field properly. The appropriate current will be determined experimentally. However, I do estimate (from the journals I have researched) that the current in each coil should be no more than 10A. Windings for each coil will also be experimentally adjusted. The current prototype calls for ~2500 turns, as this presents a small enough form factor to be easily moved and adjusted. The scanning coils will be controlled by adjusting the current in such a way to control the x-y coordinates of the focal point (See figure below). The current control scheme involves DAC controlled MOSFETs to precisely adjust the current each coil. More details will come when the prototype is constructed

​

The secondary electron detection system is composed of an external mesh biased to ~+200V attached to a photomultiplier biased to 10kV. The signal from the photomultiplier will be interpreted as an "intensity" and will displayed in a rastering manner similar to the scanning electron beam on a CRT TV.

​

In the prototype, all components will be mounted in an adjustable "slide lock" (See below) so that I may easily adjust each component (with particular care taken to find the focal point of the focusing lens) to find the optimal positions for image resolution. This scheme will not be present in the final design, as adjustable components introduce imprecision.

​

And of course, the entire assembly will be under vacuum with care taken to ensure the outgassing of the tungsten and plastic mounts does not interfere with the beam. I will not discuss the vacuum chamber construction in this report.

​

For a visual overview, see SEM MKI Overview in Files. For specific details, see the ipt files in Parts RAR.