🚀 go-pugleaf

In article <15563@suns3.crosfield.co.uk>, rak@crosfield.co.uk (Richard Kirk) writes:
|> My first reaction was to do away with the lens altogether, and put the CCD
|> straight into the TEM (it is a TEM isn't it?).  I'm not sure how far your
|> electrons would go in Si, as I used to do SEMs instead, but it ought to be
|> quite a bit less than the ten microns pitch of the devices.  They might also
|> make considerable IR, and the window of the CCD would probably have to go.
|> Has anyone tried this? :)

I'd asked this of everyone I could find about a year ago and found that the
overwhelming consensus was that this was a Bad Idea.  

First, depending on the electron voltage, one could torch the CCD.  Also, 
some sort of shutter mechanism will be needed to handle exposure timing 
(one could use the beam on/off/intensity, but the change in voltage produces
a blooming that would distort), so you'll be bombarding metal with electrons,
which is also a Bad Idea for a sensitive CCD/electronics.  In addition, if 
any gold is used in the CCD (contacts, whatever), then you'll get some small
x-ray emissions to deal with as well.  Also, thermal expansion of the CCD 
could eventually be a problem.  

I don't want to be a downer on this and would actually be really excited for
someone to look into it, but I do want people to realize the implications.
I'd suggest calling some of the manufacturers of TEMs (Philips, Zeiss, JEOL)
and finding out what the details are and then contacting the CCD fabs and
asking if they can handle those kinds of conditions.  I'd really love this
to work someday...

							kevin