Transcript Modeling of Optical Ghosts in WFC3

WFC3 Filter Testing, Modeling, Designing
TEAM
Ghosts (spurious reflections) in some UVIS filters were discovered during
first ambient calibration of WFC3
Ambient Cal Filter Ghost Characterization – T. Brown, O. Lupie
GSFC Lab Setup (spare and proto-type filters): Randy Telfer (Orbital, GSFC),
Ray Boucarut (GSFC)
Filter Modeling: Dave Kubalak, Randy Telfer, (Orbital), Bill Eichhorn (GSFC)
GSFC Lab Data Reduction , Analysis: Sylvia Baggett, Olivia Lupie
Vendors: Barr Associates, Omega Optical
UVIS Ambient Nano Calibration – G. Hartig, N. Reid, S. Baggett, T. Brown,
H. Bushouse, B. Hilbert, O. Lupie
March 18, 2004
Tips – Olivia Lupie
1
Parameters Used to Spec a Filter to Vendors
Parameters Used to Spec a Filter
Spectral Requirements
Wavelength tolerances
Central wavelength
Slopes of bandpass sides/wings
Out of Band Rejection Longward of Passband
Out of Band Rejection Shortward of Passband
Ripples in Passband
Scattered Light
Angle of Incidence
Focus shift and Filter thickness
Anti Reflection Coatings
Operational Temperature
Dimensions
mechanical size, shape
Clear Aperture size
Wedge - for transmitted wavefront - substrates
Optical Figure
Surface quality
Transmitted Wavefront
Scratch-Dig-blemishes
Adhesion of coating
Hardness of coating
Humidity
Construction
Adhesive bondline
Coating materials - general
Adhesives
Edge Sealants
Environmental Requirements
not usually tested before install March 18, 2004
instrument specific, costly, schedule prohibitive
Pre-Install Test
x
x
x
x
x
x
x
x
x
x
x
x
x
x
Vendor uses specs to design a filter:
•Determine substrates, coatings, coating
thicknesses, deposition process.
•Provide model of throughput, out of band
rejection, spatial uniformity
•Model is accepted or rejected
•Vendor builds the filter
x
x
x
x
x
x
x
x
x
x
Tips – Olivia Lupie
2
WFC3 Filter Testing, Modeling, Designing
WFPC-1’s SOFA – 12 wheels
• Converging instantaneous beam footprint
• F31 beam
• +/- 3 degree range for field angles
Filter Wheel Filter Wheel Filter Wheel
1
6
12
UVIS CCD
.
10 mm
14 mm
~3o
.
.
Edge Rays Define
Field-Of-View
March 18, 2004
Tips – Olivia Lupie
Instantaneous Beam
Footprint
Image Rays
Image
3
F225 Trans m is s ion The ore tical vs M e as ure d
1.1
1
0.9
Transmission
0.8
0.7
0.6
0.5
sht1
0.4
desired bandshape
0.3
0.2
0.1
0
150
1.1
200
250
is s ion The ore tical
300 F225 Trans m350
400 vs M e as ure d
450
500
550
600
Wave le ngth (nm )
1
0.9
Transmission
0.8
0.7
sht1
0.6
sht2
0.5
desired bandshape
0.4
0.3
0.2
0.1
0
150
200
250
300
1.1
350
400
450
500
550
600
Wave le ngth (nm )
1
0.9
"Transmission"
0.8
0.7
sht1
0.6
sht2
0.5
AR Coating
0.4
met block
0.3
desired bandshape
0.2
Theoretical Transmission
0.1
0
150
200
March 18, 2004
250
300
350
400
Tips
– Olivia Lupie
Wave le ngth (nm )
450
500
550
4
600
Spectral Shaping of the Filter
1.1
1
0.9
"Transmission"
0.8
0.7
sht1
0.6
sht2
0.5
AR Coating
0.4
met block
0.3
desired bandshape
0.2
Theoretical Transmission
0.1
0
150
200
250
300
350
400
450
500
550
600
Wave le ngth (nm )
1.1
Theoretical Reflectivity
1.0
0.9
Reflectivity
0.8
0.7
0.6
sht1 refl
0.5
sht2 refl
0.4
AR Coat
met block refl
0.3
0.2
0.1
0.0
150
200
March 18, 2004
250
300
350
400
Wavelength
(nm )
Tips
– Olivia Lupie
450
500
550
600
5
Modeling Status
Air-Gap
Construction
Possible Model of F225W Ghosting (modelers: Randy Telfer, Dave Kubalak)
Aberrations result from reflections from metal blocker
Anti-reflection
Substrate #1
Metal blocker
(aberrations –
astigmatism)
1
Substrate 1.1 mm
2
Bond &
GAP (0.38 mm)
Spacer
3
Shortpass 1
Substrate 3.0 mm
Substrate #2
4
Shortpass 2
Ideally – all surfaces
perfectly parallel
2nd order ghosts
doubly aberrated
1st order ghosts
Transmitted
aberrated
Airgap replaced adhesive – adhesive reduced throughput and introduce
major spatial dependence across filter
March 18, 2004
Tips – Olivia Lupie
6
Flight Filter Ghosts
(worst cases)
Some wide band UV air-gap filters exhibited large amplitude ghost images:
> 10% in white light
> 10% in white light
F218W
F225W
F218W
F225W
<1% in white light
F300X
<1% in white light
F606X
(analysis T. Brown with ICAL team)
March 18, 2004
Tips – Olivia Lupie
7
Lab Measurements of Spare F606
Flight – F606W in WFC3
White light – 5 micron fiber
Faint point
0.02%
ghosts at
0.3%
~0.1% of the
0.06%
primary image
0.08%
test artifact intensity, moving
0.08%
little with field
0.13
position.
%
Larger donut
ghosts at 0.3%,
moving
significantly.
March 18, 2004
Faint point
ghosts at
~0.01% of the
primary image
Intensity. Field
angle check in
work..
*white light
ghosts 10x
fainter than
Flight –
however more
testing is
needed to
verify.
Tips – Olivia Lupie
Spare – F606W – lab
Cohu, 10 micron fiber
Xenon Lamp
8
F225W
Primary images
ghosts
200 nm
275 nm
Low level
ghosts
400 nm
1100 nm
From Nano-Cal :
First Order Ghost Strength as a Percentage of Primary Image F225W
80%
70%
strong ghost
From Models : surfaces 4-3
strong ghost
surfaces 4-2
60%
Transmission
strong ghost
surfaces 3-2
50%
40%
Ratio: Meas ured
ghos t flux/ prime flux
Meas ured
Filter Throughput
30%
20%
10%
March 18, 2004
0%
180
200
220
240
260
280
300
340 360
380
Tips320– Olivia
Lupie
Wavelength (nm)
400
420
440
460
480
500
9
Spectral Modeling of Ghosts (D. Kubalak)
First Order Ghost Strength as a Percentage of Primary Image F225W
80%
70%
Surfaces (3-2)
Note - Surface (4-2)
curve overlaps with
extended wing of in band
transmission
Surfaces (4-3)
Surfaces (4-2)
Surfaces (4-1)
60%
meas ghost ratio
meas trans
Surface (3-1)
Transmission
50%
40%
30%
Surface (2-1)
Model Transmission after two reflections. To
compare models to measured ratio of ghost strength to in-band
transmission (black curve with open circles), scale surface
curves by transmission (wfc3 + filter + ota + stimulus) at the
wavelength. Models do not yet produce as high
a ghost transmission ecause of complexities but they
indicate which surfaces are most responsible.
20%
F225
10%
0%
180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 500 520 540 560 580 600
Wavelength (nm)
Ghost spectral modeling- D. Kubalak
March 18, 2004
Tips – Olivia Lupie
10
Phase Retrieval and Spectral Ghost Models
Flight Filter
Strange Morphology 
combination of Astigmatism,
overall curvature, local surface
ripples
Phase Retrieval from Focus Sweeps
(R. Telfer)
March 18, 2004
Tips – Olivia Lupie
11
Air-Gap Ghost Mechanisms
Two-surface reflection modeling indicates the metal blocker is the likely origin of the ghost
behavior. Vendors also say that the metal coating is the least “controllable”.
The observed wavelength dependence is understood. Red and Blue near-band wings are not as steep
as desired. This excess light occurs at the wavelengths where ghosts could be produced by the airgap construction. Model ghost fluxes (10-12%) are comparable to measured in white light.
Phase Retrieval reveals ghost images are astigmatic for 218W and 225W, that the coatings are tilted
w/r to one another, and filter has a slow, slight curvature  possibly consistent with a
shrinkage/distortion at the spacer/metal blocker interface.
None of these issues have any effect on the transmitted beam and throughput – both were excellent.
March 18, 2004
Tips – Olivia Lupie
12
Status Filter-Ghost
Mitigation Plan
1.
PLAN
Adopt a 3-option approach for Air-Gaps:
Barr to proto-type new F218, 225, (and 300X) filters
– single substrate
IPT tested image quality using special lab setup – flight
Spare filters to see if they exhibit less ghosting.
IPT is investigating a wedge fix – original design but
with increased wedge to deflect reflected light;
IPT is investigating dual-wheel air gap – achieved
wedge by stacked-SOFA wheel approach;
requires two coated substrates and loss of a filter(s).
2. Mechanism for F606w (laminated) ghosts
is being discussed with OMEGA.
March 18, 2004
STATUS
Barr sent thin prototype single substrate – testing
in the GSFC – only one ghost present, white
light 0.6%
F218 and F225 spares  same problem.
F606 spare exhibits similar type ghosts but
greatly reduced ghost amplitude (~0.03%)
Modeling shows you cannot tilt filter enough
and still stay with bounds of the filter housing..
By “tilting” the spare air-gap, we can determine
how much relative tilt of the two substrates
is needed to move the ghosts out of the fov.
Data taken last night.
OMEGA is devising a new design.
Tips – Olivia Lupie
13
GSFC Lab Testing Facility
Optics Team: R. Telfer, R. Boucarut, D. Kubalak, B. Eichhorn, J. Kirk, B. Greeley
Science IPT: O. Lupie, S. Baggett, B. Hilbert, T. Brown, G. Hartig
Goals – last few weeks: 1) Prove that the GSFC Lab Test setup accurately simulates
the WFC3, i.e., measurements are true representations of the filter imaging quality,
and 2) Measure the flight spares.
•UV Sensitive CCD.
•Cover Structure for
uniform/dark background.
Off-axisParabola
Fiber
Filter
CCD
F/31Beam
•Automated Castle and
CCD data take system.
Castle Cart
Double
Mono
chrometer
March 18, 2004
•Mechanical stage mount
for filters.
presents F31 light beam
to the filters as they would
see in the WFC3+OTA
Tips – Olivia Lupie
•Semi-Automated data
reduction and analysis.
14
Lab Measurements of Spare F225W
SPARE F225
Cohu Video CCD,
200 micron fiber
FLIGHT F225
5 micron fiber,
WFC3
(sum ghosts=15%)
Saturated
prime
Saturated
prime
10%
10%
SPARE F225
SBIG CCD,10 micron fiber,
9%
0.5%
2%
UV00
0.5%
Saturated
prime
UV00
UV14
UV14
Note – rotation and stretch
are different.
March 18, 2004
Tips – Olivia Lupie
Relative positions and fluxes
of ghosts in the spares are
comparable to those in the
flight  similar mechanisms.
15
Example Monochromatic Results for
Spare F225W
Spare F225, SBIG-CCD, 200 micron fiber, 13nm bandpass, double UVIS, ,
with ND1 (removed in later imaging).
220nm
240nm
260nm
300 nm
320 nm
340 nm
280nm
Figure from S.Baggett
March 18, 2004
Tips – Olivia Lupie
16
Establish Setup Sensitivity and Repeatability
Spare F225W
Ran many tests to
establish sensitivity to
ghosts, setup alignment
accuracy, and
experimented with
several different filter
orientations: rotation,
back to front, tilts,
translation, wedge
orientation and detector
tilts.
nominal
-1d
-2d
+1d
+2d
Xenon Lamp, 10 mic Fiber
Rotate Filter a few degrees from nominal and compare
Ghost morphology
March 18, 2004
Tips – Olivia Lupie
17
Establish Setup Sensitivity and Repeatability
Also Helps modelers to see all the ghosts
saturated unsaturated
Primary, Secondary
ghosts emerge from
behind the primary and
each other when large
translations or rotations of
the filter are introduced:
ie different locations on
filter and differing field
angles.
center
saturated unsaturated
center - shows repeatability
+0.5 in
+1.0 in
-0.5 in
-1.0 in
Xenon Lamp, 10 mic Fiber
v. Large tilt –
30 deg to corner
Translating the filter
Figure from S.Baggett
March 18, 2004
Tips – Olivia Lupie
18
Prototype F225 – Single Substrate
Setup artifact
Prototype thickness is
smaller than that of a flight
filter. Thicker filters result
in ghosts at a larger radial
distance from the primary.
But tilting the thin filter, we
can see when the ghost
emerges and use a simple
Model to derive the ghost
position with a thicker
filter.
nominal
-3d
-9d
-12d
-6d
Ghost 0.6%
-15d
Prototype Single Substrate – large tilts
(Xenon Lamp, 10 micron fiber, CCD SBIG)
Figure from S.Baggett
March 18, 2004
Tips – Olivia Lupie
19
8.0E-01
trial
7.0E-01
th trans
th refl
6.0E-01
flight 225-302 trans
orig theor
5.0E-01
Transmission
Proto
Type
F225
From
Barr
"Feb_19 Trial F225W"
4.0E-01
3.0E-01
2.0E-01
1.0E-01
0.0E+00
190
200
210
220
230
240
250
260
270
280
290
300
"Wavelength"
7.0E-05
Feb_19 Trial F225W" - OOB Blocking
trial
6.0E-05
theoretical trans
meas 225-302 OOB
Transmission
5.0E-05
4.0E-05
3.0E-05
2.0E-05
1.0E-05
0.0E+00
200
March 18, 2004
300
400
500
600
700
Tips
– Olivia Lupie
"Wavelength"
800
900
1000
1100
20
Prototype F225 – Single Substrate
Setup artifact
Ghosts as a function of
wavelength
200nm
250nm
300nm
400nm
450nm
500nm
600nm
650nm
700nm
750nm
800nm
850nm
900nm
900nm
350nm
Setup
SBIG CCD, 200 mic fiber
Castle Modes
<250nm double UV
250-310 double UVIS
310-760 double VIS
>760 double IR
March 18, 2004
225nm
Ghost#1
Ghost#2
Tips – Olivia Lupie
Figure from S.Baggett
21
March 18, 2004
Tips – Olivia Lupie
22
• 53 of 63 filters exhibit excellent performance, consistent with spec.
•
•
47 filters < 0.2% ghosts
6 filters 0.2-0.5% ghosts
- multi-substrate - 410M, 689M, 814M
- air-gap – 656N, 665N, 673N
• 2 filters 0.7% ghosts:
•
single substrate+Al blocker - 275W  can calibrate
•
air-gap - 658N
• 2 UV high priority air-gap (with Aluminum blocker) 10-15% ghosts - 218W, 225W  unsuitable for
flight
• 1 UV air-gap 1% with strange morphology - 300X  marginal, tough to calibrate
• 1 3-substr laminated, < 0.5% “point-like” ghosts - 606W  most used filter, concern
(other filters with very low level “point ghosts”: 625W, 775W, 410M, 467M, 547M, 621M, 689M)
• 1 UV single subst.+Al block, possible surface flaw – 280N  serious but low priority filter
• 2 UV Quad filters single substrate, 5% ghosts:– 232N, 243N  low priority filters, can calibrate
• Grism – data reduction in work
March 18, 2004
Tips – Olivia Lupie
23