Changing hot-filament SEM electron source into field-emission one MOTIVATION: The resolution of scanning electron microscope (SEM) images is limited by multiple factors, one of them being relatively low electron-beam current density emitted by cheap tungsten filament sources. 
Of course, one can reduce the numerical aperture of the SEM to get somewhat better resolution (20~50 nm), but then the e-beam current drops down to picoampere level, images get noisy and it takes minutes to take them.
Much better scanning speeds and/or image quality can be acquired when extremely bright e-beam is formed by a field-emission gun (FEG) using a cold sharp tungsten tip. 
Since I have an access to an older SEM (which needs to replace the filament anyway), to a spot welder and to an inexpensive source of etched tungsten tips with radius well below 50nm, I can quite simply weld the sharp tip to the filament holder and place it in the Wehnelt assembly where normally a hot filament is found. 
QUESTIONS: 
1) Has anybody tried this before? Is there any fundamental change in the electron source geometry that prevents converting thermal-emission source to FEG source?
2) Field-emission SEM are known to run at ultrahigh vacuum (<10⁻⁸ mbar). Does any FE-SEM user have an experience what happened to the electron gun when this vacuum was compromised to roughly 10⁻⁶ mbar? (This is lower-quality vacuum typically used in a filament SEMs.) 
3) Is there some hidden pitfall that could permanently damage the microscope if I do this simple experiment myself? Is there any other risk (except risking the loss of my time)?
 A: 1) I have discussed this specific question with vendor of my SEM. I was also considering upgrading to ether Lab6/CeB6 or FE source. Both these options offer improvement in resolution with similar disadvantages.
While LaB6/CeB6 could be an easier upgrade (you might not need to change electronics at all, and compatible sources are manufactured for many tungsten-source SEMs), FE source might require more mechanical/electronic changes. 
In ether case, column of the SEM has to be upgraded to add ion pumps to reach and maintain UHV for the whole lifetime of the machine.
2) I did not found good answers on what is happening at poor vacuum. I would assume emission will drop significantly, but I found no references. I see that field source could be flashed to de-contaminate, and you might need to do that (significantly) more often, with reduced lifetime. 
3) You might not be able to install FE source (mechanically and electronically). FE source might be damaged early (i.e. 1-10 hours instead of 1000 hours of operation). You might get less resolution increase than expected: High-resolution in-lens FESEM use different and more complicated lens arrangement.
As a side note, good SEMs with tungsten source reach resolution of 3nm. If you are talking about improving resolution from 20-50nm - I would say SEM is very far from reaching the limits of electron optics with tungsten filament, and your FE/*B6 source will be mainly limited by the aberrations of the electron optics of the SEM. 
