CCD pixel-to-pixel variations boundary="part0_921735424_boundary"

[MYPIXEL at aol.com]

***********************************************jj
Paul. . . .

below . .  . 

Jim
***************************************************jj


In a message dated 3/17/99 1:22:54 AM Pacific Standard Time,
Paul.Jorden at eev.com writes:

<< >
  > Has anyone any knowledge on how pixel-pixel QE variations
  > changes as a function of
  >
  > 1) Wavelength
  > 2) Time >>


**************************jj
Might throw in my two cents worth . . . . . not already mentioned . . .

Pixel nonuniformity is dependent on clock bias. For example, two barrier
phases that are exactly equal in potential will produce the greatest
nonuniformity (charge has a hard time knowing how to divide). If these
voltages vary slightly a new nonuniformity pattern is seen. Sometimes it is
necessary to slightly offset barrier phases to help direct charge to the
collecting phase consistently. Integrating under two phases (as opposed to
one) helps this problem. Attached find a plot of pixel nonuniformity as a
function of negative bias to barrier phases 2  and 3 leaving the collecting
phase at 7 V. Note that pixel nonuniformity is greatest when the two barrier
phases are equal in potential. Taking a CCD into inversion also reduces pixel
nonuniformity by pinning and flattening small potential barriers under a phase
(how most pixel nonuniformity originates). 

Residual Bulk Image (RBI) also adds instability to pixel nonuniformity. All
frontside CCDs built on epitaxial material show QE instability (i.e., traps at
epi interface). The problem is a red one where absorption length is longer
than the epi thickness. For the Hubble, Galileo, SXT and Cassini CCDs the QE
varied as much as 10 % depending on signal level and operating temperature.
For Galileo and Cassini an IR light flood was required before each exposure to
stabilize QE to 1%.

RBI becomes obvious for operating temperatures below -60 C and wavelengths
longer than 7000 A. Backside CCDs do not exhibit RBI in that epi interface
(where traps are located) is thinned away. Excess dark current from the same
region is seen when a dark frame is taken after a high level exposure (hence
the name residual bulk image). The time constant associated with RBI is
exponentially dependent on operating temperature (from usec to hours).

Dark current nonuniformity. .  . another interesting subject to get into   ..
. . . 

Jim

**********************************************jj
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/tiff
Size: 58256 bytes
Desc: not available
Url : http://www.ctio.noao.edu/pipermail/ccd-world/attachments/19990318/513bfd65/attachment.tif