Anti-Blooming Strategies for LSST

1
Anti-Blooming Strategies for LSST

• or
• What we can do about all these bloomin pixels?

2
LSST Design Parameters
3
Stellar Flux, Charge Accumulation vs. V-band
Magnitude
4
Stellar Density vs. V-band Magnitude
5
Flux, Charge, and Blooming in 10 Second LSST
Exposures
6
I-band vs. V-band Stellar Densities
There are 3 5x as many bright stars in I-band
7
So, how big IS the problem of bloomed charge for
LSST?

• Simple model predicts a small portion (0.11)
  of a typical LSST
• field will be affected by bloomed charge,
  based on V-band magnitudes

• I-band stellar brightness suggests a
  multiplication of this amount by
• as much as 5x, so thats maybe 0.5

• Corrupted PSFs suggest a multiplication by
  another 3x to 5x, so thats
• maybe 2.5, thats starting to be a problem

So, what can we do about it?

• Structural solutions built-in anti-blooming
  drains

• Operational solution clocked recombination
  anti-blooming

8
Anti-Blooming Structures Vertical and Lateral
Drains
(Graphics stolen shamelessly, err, borrowed, from
Scientific Charge-Coupled Devices by James
Janesick)

• Vertical anti-blooming structures are
  incompatible with backside illumination

• Lateral anti-blooming structures sacrifice too
  much fill factor

9
Lincoln Laboratories Blooming Control

• Drain hidden in the channel stop

• 100 fill factor

• gt1000x blooming control

• Similar design available from Sarnoff
  Laboratories

10
Other Possible Anti-Blooming Structures

• Column end excess charge drain

• Absorbs charge as columns saturate completely
• Also serves as a fast clear charge sink

• Lateral drain on serial register

• Prevents charge from smearing during readout
• Also serves as a fast clear charge sink

11
Clocked Anti-Blooming
P1 collecting charge, P2 and P3 inverted
P1 saturated, excess charge at the Si/SiO2
interface at the surface
Excess charge trapped at surface while the charge
is moved in the buried channel
Trapped charge annihilated by holes flooding
surface states when P1 is clocked into inversion
Diagram borrowed from Gary Sims
12
Spurious Charge Generated by Clocked Anti-Blooming
Plot stolen (lets just say it) from Sims and
McCurnin

• Spurious charge generation rises with larger
  clock swing

• Spurious charge generation rises with shorter
  clock rise times

LSST cannot afford the noise from spurious charge
13
Conclusions, Options, Further Research

• Without anti-blooming, bloomed charge will
  affect between 0.5 and 2.5
• of a typical LSST field, is this a problem?

• Most built-in anti blooming structures are
  incompatible with LSST

• MIT/LL, Sarnoff have compatible, effective
  anti-blooming structures

• Can MIT/LL produce the required number of CCDs?

• Can the design be licensed to another vendor?

• How about Sarnoff Laboratories?

• Clocked anti-blooming works, but has a LOT of
  potential problems

• Spurious charge, jim-dots, other effects are
  not acceptable

• Further experiments required before considering
  this option