QE and MTF

[Michael P. Lesser]

John, Paul, Ringo, and others:

Paul is correct that changing the peak QE in the blue will effect the
red response.  We can easily get 90% QE at 400nm, but the red QE (800nm) falls 
from 80% to about 70%.

As Paul alludes to, there can be significant QE differences between RT and
-100C depending on backside charge, backside coatings, and most likely many 
other parameters.  Compare at the same operational temperature to be sure!

In general the devices should be produced on the correct thickness epitaxial
silicon.  If the epi is too thick, as was used for many foundry lot
runs which astronomers now have, then the additional thinning required to 
move well into the epilayer enhanced silicon defects.  
We can imporve the MTF pretty 
much as predicted by models, but the device quality suffers.  This appears to
be a function of the silicon quality, since it varies greatly from lot to lot.
If we leave the devices too thick, MTF is noticably degraded.  The solution is
to start with the correct material (and to get high quality silicon!).  If the
silicon quality is very good (and consistant) then ultrathinning is 
straightforward and MTF and QE can be tailor for specific applications.  

Of course, higher resistivity material can be used thicker with good MTF as the
internal electric fields penetrate farther into the silicon.

-Mike Lesser