TIPS-JIM Meeting
19 May 2005, 10am, Auditorium
1.
ACS short annealing test and
hot pixel behavior
Marco Sirianni
2.
FGS Tunable Filter Imager: Updates
from PDF
Alex Fullerton
3.
Drizzling the ACS mosaic of M51
Max Mutchler
Next TIPS Meeting will be held on 16 June 2005.
Results from the
ACS “Short Annealing” test
and the life of hot pixels
Marco Sirianni
Special thanks to WWV
Alan Welty, Tom Walker and Alison Vick
5/25/05
Marco Sirianni
Dark and hot pixels:
•
Radiation produces displacement damage
(creation of defects with intermediate energy levels)
Ec
Et
Ev
The majority of the defects have
Et right in the middle of the
band gap -> maximum efficiency
for dark rate increase
the density of the defects increase linearly with time
and as consequence the mean dark current increase with time…
Observed in ACS, STIS and WFPC2 (until summer 1998)
Marco Sirianni
Hot pixels
Besides the “Uniform’ damage there is a
less frequent but more severe local damage
• Impact with very energetic particles can create centers
of very high e- production
• The same defect that creates the “normal dark”
Can generate much more current in presence of an
enhanced field (field enhanced emission)Ἶ
Marco Sirianni
Evolution of the dark rate distribution
• Is very well predictable
• The histogram can be be fitted with a Gaussian main
peak whose Half-width increase ~ sqrt (fluence) and an
exponential tail whose amplitude ~ fluence
Only 2 possible solutions:
reduce the T
anneal the hpix
Marco Sirianni
HPIX flagging
> 5 sigma
Replaced in
Daily dark
0.02
0.04
> 0.08
Flagged in DQ
0.06
Marco Sirianni
0.08
0.10
The usual 12 hr Anneal
• A 12hr anneal takes ~16 hrs:
–
–
–
–
10 min - put ACS on hold
12 hr - anneal
10 min - put ACS on hold
30 min - Allow HRC to reach set temperature
• WFC needs extra 60 min to do the same
– 3 hr
- CEB temperature stabilization
Ἶ
After STIS departure it has been difficult to schedule nonACS SI activities during these long periods…
-> loss of HST time
Marco Sirianni
Shorter anneal
•
We still do not know how exactely the anneal works
we only know that at ~ +20 C we should not see any
improvement in a short period (< weeks) of time
(The majority of defects need +160 C)
•
Initially ACS used 24 hr anneals….
–
Going from 24 to 12 hrs did not seem to create any problem
WFC3 ground testing (E. Polidan):
1. Anneal occurs also at very low temp
2. Most of the anneal seems to occur early on
(90% during warm-up period)
Still, each chip is different, (doping and impurities)
Marco Sirianni
Temperature Profile
HRC: from -81 to +23 C in ~ 2hr
WFC: from-76 to +18 C in ~ 4hr
(after 2 hr is ~ 7 C)
Removed 6 hrs
12 hours
Marco Sirianni
Anneal rate: calculation
Anneal day
A
Daily
Hot Pixel
growth
B
Permanent
Hot pixels
growth
C
Annealing
rate
(A - B) / ( A - C)
Marco Sirianni
Anneal rate - comparison
Marco Sirianni
Short vs long anneal
WFC
HRC
Mean
12hr
+/-
6hr
+/-
Mean
12hr
+/-
6hr
+/-
>0.04
0.70
0.07
0.78
0.24
0.84
0.10
0.69
0.13
>0.06
0.78
0.04
0.83
0.22
0.87
0.07
0.79
0.09
> 0.08
0.82
0.03
0.86
0.15
0.87
0.07
0.81
0.08
> 0.10
0.84
0.02
0.89
0.12
0.85
0.08
0.80
0.05
Marco Sirianni
Anneal Rate
wfc1
wfc2
HRC
Series4
1
0.9
0.8
anneal rate
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0.02
0.04
0.06
0.08
0.1
1
signal (e-/pix/sec) (>)
•The anneal rate change significantly with the signal level
• no anneal for the “normal dark” pixels
Marco Sirianni
Permanent hot pixel growth
HRC [>0.08]]
WFC [>0.08]]
Marco Sirianni
Number of permanent hot pixels..
years
years
years
%
%
WFC
Marco Sirianni
WFC
%
HRC
The life of HPIX-1
• Randomly selected hpix above the 0.08
e-/pix/sec threshold in May 2004…
in a random section of the frame..
• Monitored the evolution in time since
launch
Marco Sirianni
Hot Pixel “Donald”
Marco Sirianni
Hot pixel “Gladstone”
Marco Sirianni
The “fly low” hot pixel
Marco Sirianni
The “restless” hot pixel
Marco Sirianni
Their life in a flash
Marco Sirianni
Conclusion
• 6hr annealing could be a solution…
– Started in April, monitored closely in the next few
months to check for longer term effects
• Hot pixel growth is as expected by theory
• The anneal rate depends on the signal
level… impact only field-enhanced
defects…
• Complete anneal is quite rare
• Many hot pixels behave erratically
• Particular programs may need to use
different HPIX / WPIX flagging
Marco Sirianni
Anneal Rate vs frequency
Anneal cycle is usually 27-29 days
Shorter cycles are linked to gyro safing…
HRC [> 0.08 e-/pix/sec]]
WFC [> 0.08 e-/pix/sec]]
No clear trend visible, same result at all threshold levels
Marco Sirianni
Daily rate of HPIX formation
HRC [>0.08]]
WFC [>0.08]]
Marco Sirianni
Hot Pixel Growth impact
Permanent hot pixel growth
(% of total number of pixels / year)
Threshold
e-/pix/sec
WFC
HRC
STIS
WFPC2
temp
- 77 C
- 80 C
- 83 C
- 88 C
Dark curr.
0.003
0.004
0.006
0.008
> 0.02
1.60
1.54
2.99
(0.30--0.11)
>0.04
0.78
0.52
>0.06
0.46
0.29
>0.08
0.30
0.21
>0.10
0.23
0.17
0.36
>1
0.03
0.02
0.08
Hot pixels are not fully stable, noise > shot noise
Solution: monitor their population and dither the observations
Marco Sirianni
Annealing rate
Anneal Rate
100
Anneal Rate (%)
90
WFC (-77/+20)
HRC (-81/+20)
80
70
60
50
c
40
30
20
10
0
> 0.02
> 0.04
> 0.06
> 0.08
> 0.1
Hot pixel Threshold (e-/pix/sec)
Instrument
Temp
(CCD/ann.)
Threshold
(e-/pix/sec)
Anneal rate
STIS
-83 / +5
> 0.1
~ 80 %
~ 75 %
Hayes et al.1998
Kim Quijano et al. 2003
WFPC2
-88 / +22
> 0.02
variable
~ 80 %
Koekemoer et al. 2003
WFC3
ground
-83 / +30
>0.01
>0.04
~ 80 %
~ 97 %
Polidan et al. 2004
Marco Sirianni
Source
FGS Tunable Filter Imager:
Updates From PDR
Alex Fullerton
STScI / UVic
TIPS/JIM
May 19, 2005
FGS PDR: May 4/5, 2005
01-1
Design Updates to TF Imager at PDR
•
Maturing Instrumental Design
•
•
•
•
•
Coronagraph moved
Revised distortion map
Progress on coatings for etalons and dichroic beamsplitter
Much more work on ghosts
Mechanisms and Etc.
•
•
•
Dual filter wheel design is well in hand
Some elaboration of onboard calibration units
Preferred material for etalon launch lock identified
FGS PDR: May 4/5, 2005
01-2
FGS-TFI Solid Model
FGS PDR: May 4/5, 2005
01-3
Current JWST FOV Layout
•
•
The FGS Optical Assembly supports two
pickoff mirrors at or near the OTE focal plane
Packaging is tight but current design meets
the allocation
FGS-TF Coronagraphic
Mask locations
80”
20”
7λ/D
(1’’)
FGS PDR: May 4/5, 2005
10λ/D
(1.4’’)
15λ/D
(2.2’’)
20λ/D
(2.9’’)
01-4
Coronagraphic slide design
7λ/D (1’’)
10λ/D (1.4’’)
15λ/D (2.2’’)
20λ/D (2.9’’)
RMS WFE versus Coronagraph Slide Thickness
40
LW Full Field Performance
LW Coronagraph
SW Full Field Performance
RMS nanometers
35
SW Coronagraph Performance
30
25
20
•
Occulting spots have a graded intensity profile
(Bessel^2, Sinc^2)
•
•
Best performance at long wavelengths (~4 µm)
•
Bars optimized for short wavelengths (~2 µm)
Circular spots optimized for long wavelength
operation
15
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Slide Thickness (mm)
FGS PDR: May 4/5, 2005
01-5
FGS-Tunable Filter Imager Distortion
as Plate Scale Variations
Plate Scale vs Field
Variation over image
Pos 1
Plate Scale vs Field
MAS
Variation over image
Pos 2
MAS
65.558
65.520
64.913
64.914
64.307
64.267
Plate scale units = mas/pixel (FOV each pixel sees)
Longwave
Shortwave
Average (x/y)
64.96
64.95
Minimum (x/y)
64.27
64.31
Maximum (x/y)
65.56
65.52
FGS PDR: May 4/5, 2005
01-6
FGS TFI Plate Scale Anisotropy
X and Y Values
Longwave
Shortwave
X
Y
X
Y
Average
64.59
64.96
65.27
65.77
Minimum
63.91
64.63
63.94
64.66
Maximum
65.30
65.82
65.27
65.77
FGS PDR: May 4/5, 2005
01-7
Preliminary Dichroic Design
•
•
Germanium Substrate
72 layer a-Si / SiO2
FGS PDR: May 4/5, 2005
01-8
Fabry-Perot Etalons
Optical path difference = 2µl cosθ
Phase difference = (2π/λ) 2µl cosθ + δr
mλ = 2µl cosθ + δr /2π
Tuning: For fixed m,
on-axis λ(θ=0) ∝ l
µ=index of refraction
FGS PDR: May 4/5, 2005
Detuning: For fixed m, l,
off-axis λ(θ) ∝ cos θ ≈ 1 − θ2/2
01-9
a-Si
SiO2
a-Si
SiO2
a-Si
SiO2
a-Si
SiO2
a-Si
SiO2
a-Si
SiO2
a-Si
Substrate (SiO2)
FGS PDR: May 4/5, 2005
18.5
72.6
231.9
79.5
382.8
18.3
90.7
26.9
437.3
18.2
34.7
166.6
75.4
18.4
18.1
74.4
242.3
79.3
381.6
16.9
93.4
26.5
459.0
16.6
38.2
173.9
71.9
20.1
5.0
0.875
4.5
0.850
4.0
0.825
3.5
0.800
3.0
2.5
0.775
Reflectance
Phase
0.750
800
1000
1200
1400
1600
1800
2000
2.0
2200
Wavelength (nm)
Design Update
130
4000
120
3500
110
3000
100
2500
90
2000
80
1500
Gap (nm)
Material
Air / Vacuum
SiO2
"In Process"
Original Design Design Update
Thickness (nm)
0.900
Reflectance
•
•
Deposit first 8-layers of design &
measure layer properties
Re-optimize design of remaining layers
Deposit remaining 6 layers
5.5
Resolution
•
0.925
Phase (radian)
SW Etalon Mirror Coating
Deposition Process
Resolution
Gap
70
Measured
800
1000
1200
1400
1600
1800
2000
1000
2200
Wavelength (nm)
01-10
•
•
•
•
Material
Air / Vacuum
SiO2
a-Si
SiO2
a-Si
SiO2
a-Si
SiO2
a-Si
SiO2
a-Si
SiO2
a-Si
SiO2
a-Si
SiO2
a-Si
SiO2
a-Si
SiO2
a-Si
SiO2
Substrate (Si)
4.5
0.875
4.0
0.850
3.5
0.825
3.0
0.800
2.5
Thickness (nm)
20.92
201.34
484.53
210.5
562.99
30.47
240.21
400.57
62.14
120.2
91.47
521.62
102.62
139.75
96.24
536.97
86.27
140.4
109.85
150.08
46.22
0.775
2.0
Reflectance
0.750
2000
Phase
2500
3000
3500
4000
4500
1.5
5000
Wavelength (nm)
112
9000
110
8500
108
8000
106
7500
104
7000
102
6500
100
6000
98
5500
96
5000
94
4500
92
4000
3500
90
88
86
2000
Resolution
Gap
2500
3000
3500
4000
4500
3000
2500
5000
Wavelength (nm)
FGS PDR: May 4/5, 2005
( di
)
0.900
Ph
Design updated to use optical
parameters from INO’s high
temperature deposition process
21 layer design on
silicon substrate
Candidate silicon
substrate from
Lattice Materials
Polished by BMV
Technologies
Discussed in
Etalon Design
section
5.0
Resolution
•
0.925
Reflectance
LW Etalon Status
01-11
Ghost Analysis FGS-TFI Pupil Region
Coronagraph
Plane
Dichroic
BeamSplitter
SW Pupil
Region
•
3 elements create 15 ghost images in each
channel
•
•
3 intra-element ghosts
12 inter-element ghosts
LW Pupil
Region
FGS PDR: May 4/5, 2005
From document 843183 FGS TFI Ghost Image Analysis
01-12
FGS-TFI Ghost image analysis
•
FGS PDR: May 4/5, 2005
Example: Tilting pupil mask one degree
translates ghost image location
01-13
FGS-TFI Ghost Image Elimination
•
•
•
Tilt Pupil Mask and Blocking Filter by 3 degrees across plane of instrument
Prevent redirected light from propagating along unwanted paths along TMA components.
One ghost image from bottom edge of detector to Etalon presents potential ghost of
relative strength of 0.0004, using conservative reflectance values
FGS PDR: May 4/5, 2005
01-14
Dual Wheel Configuration
Filter Wheel
8 position Filter Wheel allows for 6
blocking filters, an open for
calibrations and one spare
Open
8 position Pupil Wheel allows for a Lyot
mask, up to 4 apodization masks,
2 calibration source positions and a neutral
density position for target acquisition.
B6
Spare
B5
B1
B4
B2
B3
Pupil Wheel
Lyot
Apo4
FFCal
Apo3
Note: Calibration source positions on Pupil
Wheel also serve as closed
positions for dark calibrations
FGS PDR: May 4/5, 2005
Apo1
WCal
Apo2
ND
01-15
Calibration Optics (PFlat & LFlat)
•
•
•
•
Diffuser & Cavity mounted in Pupil Wheel
Light sources located adjacent to pupil wheel
on main optical bench
Multiple sources feeding a small (<6 cm
diameter) sphere can provide some
redundancy
Pupil Wheel:
FFCal position
Note: Difference between PFlat & LFlat is in
blocking filter & etalon settings
•
•
PFlat: Filter wheel set to open position, etalon
scanned during integration
LFlat: Blocking filter selected & etalon set during
integration
FGS PDR: May 4/5, 2005
01-16
Calibration Optics (Etalon Monitor)
Pupil Wheel: WCal position
Option 1
Option 2
Side view:
LW WCal
Si
Cut Away Side View
Front View
Pupil
10.00
•
•
MM
Two options for WCal
Relative Intensity of images on
detector will allow etalon set point
and parallelism to be monitored
FGS PDR: May 4/5, 2005
01-17
Flight Tie-Down Design
The Question: How to provide 890 N preload (sufficient by analysis) under
launch conditions, and only 20 N under operational conditions?
•
High-Modulus / High-Strength / Negative CTE tie-down using Zylon HM.
•
•
•
•
•
Low elastic strain @ 890 N, < 40 µm.
Creep is an issue that can be designed for,
•
•
Since stress level is only 2.4% of the UTS.
Three-week creep test performed,
needs to be confirmed with flight tie-down.
Fiber grows as it cools to operational temperature.
•
•
> 100 µm growth.
Positive or 0 CTEs of other materials in the load path
will aid this mechanism.
This mechanism has been tested more than 30 times at various preload
levels, by tightening a representative loaded assembly with a Zylon tiedown, subjecting it to LN2 and feeling it go slack when cold, and
measurably recovering the preload after the return to room temperature.
Loop of fiber fed through a hole in the screw and around a steel dowel,
then the nut is tightened with a torque wrench, preloading the assembly.
FGS PDR: May 4/5, 2005
01-18
Hubble Space Telescope
ACS multi-color mosaic of M51,
the Whirlpool Galaxy
Max Mutchler
TIPS Meeting
19 May 2005
Press release for Hubble’s
15th Anniversary,
with public “unveiling”
events at ~100 museums.
Data release
via MAST,
and DVD
High-Level
Science Products
(HLSP):
data processed well beyond
standard pipeline processing
Approved follow-up proposals for M51
GO proposal 10501, PI Rupali Chandar
Extending the Heritage: Clusters, Dust, and Star Formation in M51
(will add WFPC2 U-band and NICMOS H-band, parallel Paschen-α data)
AR proposal 10662, PI Benne Holwerda
An dust extinction map of M51 from counts of distant galaxies
AR proposal 10666, PI Chris Impey
The ACS Mosaic of M51 and the Intersection of Research and Education
AR proposal 10669, PI Roy Kilgard
A Comparative X-ray and Optical Study of M51
AR proposal 10684, PI Paul Scowen
Multiwavelength Analysis of the Star Formation Process in M51
Observations
1
2
4
3
5
•
•
•
•
•
6
24 orbits (DD)
Filters: B, V, I, Hα
6 pointings, 2x3 mosaic
Dithers: gap, 2-point sub-pixel
96 total exposures
Data reduction
• Pipeline calibration with “best” reference files
• Registration with imexam/geomap (tinkered
with tweakshifts): intertile only
• Masking: satellite trail, high background
• MultiDrizzle: final combination with rejection
of cosmic rays and detector artifacts
• Output mosaics ~420 MB each; but ~26 MB
block-averaged versions also available
Raw
Distortion
Calibrate
Drizzle
Color composite
Cosmic rays and detector artifacts
Registration objects in tile overlaps
Mosaic registration (intertile)
51
61
31
41
11
21
Data quality
• Optimal registration: tweakshifts,
align/superalign?
• Optimal scale, pixfrac (sub-pixel resolution)?
• Optimal drizzle kernel, CR rejection, bits?
• Cosmic ray contamination in Hα tile 4 gap,
due to excluded high-background frame
• Inspection of weight maps
• Make version 2.0 mosaics?
Weight map
Weight map for Hα
B mosaic
V mosaic
I mosaic
Hα mosaic
Currently on display on the VizWall in 331B
Documentation:
README now
superseded by…
Mutchler et al., 2005,
AAS, Vol 37, No 2