Transcript High Contrast Imaging with the Hubble Space Telescope

Hubble Space Telescope
Coronagraphs
John Krist
Space Telescope Science Institute
Why Use HST?
• High resolution with wide field of view anywhere
in the sky
• Wavelength coverage from l = 0.2 - 2.2 mm
• Its stability allows significant PSF subtraction
High Contrast Imaging Techniques
Used on HST
• Direct observation with PSF subtraction
• Coronagraphic observation with PSF
subtraction
• Spatial filtering
• Spectral+spatial filtering
Choice of Cameras
for High Contrast Imaging
Direct imagers:
• WFPC2: 160” x 160”, l = 0.2-1.0 mm
• STIS: 52” x 52”, l = 0.2-1.0 mm
• ACS Wide Field Camera: 200” x 200”, l = 0.4-1.0 mm
• ACS High Res Camera: 26” x 29”, l = 0.2-1.0 mm
• NICMOS: 11” x 11” to 51” x 51”, l = 0.9–2.2 mm
Coronagraphs:
• ACS High Res Camera
• STIS
• NICMOS Camera 2: 19” x 19”
Components of the HST PSF
• Diffraction from obscurations
– Rings, spikes
•
•
•
•
Scatter from optical surface errors
Stray light & ghosts
Diffraction from occulter (coronagraph)
Electronic & detector artifacts
– CCD red scatter,
detector blooming
Diffraction from Obscurations
HST Entrance Pupil
PSF
V band (no aberrations)
Model
Scatter from Optical Surface Errors
Midfrequency Error Map
Phase retrieval derived
PSF
18 nm RMS wavefront error
V band (ACS/HRC)
Krist & Burrows (1995)
Observed
ACS Surface
Brightness
Plots
ACS
V band (F606W)
Observed PSF
Model PSF
No surface errors
Electronic & Detector Artifacts
Electronic
banding
NICMOS
WFPC2
No Halo (model)
Observed (I band)
CCD Red Halo
ACS/HRC shown.
Also in STIS and
WFPC2 F1042M
Stray Light & Ghosts
NICMOS (direct) F110W
“Grot”
Defocused
ghost
PSF Subtraction
Stability of HST allows diffracted and
scattered light to be subtracted
Reference PSF Subtraction
Roll Subtraction
Beta Pictoris
WFPC2
WFPC2 Science Team
(Unpublished)
Alpha Pic
Beta - Alpha Pic
ACS coronagraph
ACS Science Team
(work in progress)
Sources of PSF Mismatches
• Focus changes caused by thermal variations
– “Breathing” = 3-5 mm primary-secondary separation
change within an orbit = 1/18-1/30 wave RMS change
– Attitude changes (0 – 1/9 wave change)
– Internal changes in camera
•
•
•
•
•
Color differences
Field position variations (WFPC2)
Star-to-occulter alignment (coronagraphs)
Lyot stop shifting (NICMOS)
Jitter
Direct Observation with
PSF Subtraction
• Primarily used for WFPC2, but also ACS
and NICMOS on occasion
• PSF is subtracted using an image of another
star (or roll self-subtraction)
• Deep exposures saturate the detector, but
bleeding is confined to columns (for CCDs)
or just the saturated pixels (NICMOS)
Direct Observations – WFPC2
GG Tauri Circumbinary Disk
Science results in Krist, Stapelfeldt, & Watson (2002)
V band
Log stretch Unsubtracted
- PSFs
• Disk around
binary T Tauri
system
I band
• Inner region
cleared by
tidal forces
• Integrated
ring flux = 1%
of stellar flux
@ I band
Direct Observations – ACS/HRC
HD 141569
Disk around a Herbig Be star
at d = 99 pc
Disk flux = ~0.02% of stellar
flux
HD 141569 - PSF
PSF is 2.5x brighter
than disk here
Reference PSF
7”
ACS Science Team observations (unpublished)
Using a Coronagraph
• Suppresses the perfect diffraction structure
• Does not suppress scatter from surface
errors prior to occulter
• Reduces sensitivity to PSF mismatches
caused by focus changes & color
differences
• Occulting spot prevents detector saturation,
ghosts, and scattering by subsequent
surfaces
• Deeper exposures possible
NICMOS Coronagraph
• 0.076” pixels, l = 0.9 - 2.2 mm
• Spot and Lyot stop always in-place
• Occulting spot is r = 0.3” hole drilled in mirror
– Contains 2nd dark Airy ring at l=1.6 mm (spot
diameter = 4.3l/D, 83% of light)
– Rough edge scatters some light (“glint”)
– Useful inner radius ~0.5”
– Spot in corner of field
0.6”
NICMOS Coronagraph Pupil
Models
Pupil after spot
With an Aligned
Lyot Stop
With a Misaligned
Lyot Stop
• Stop does not block spiders, secondary, edge
• Stop “wiggles” causing PSF variations
• Too-small spot causes “leakage” of light into pupil
Effects of NICMOS Lyot Stop Misalignment
F110W (~J band)
Aligned Lyot Stop
Misaligned Lyot Stop
Model
Model
Observed
Misalignment results in 2x more light in the wings + spikes
NICMOS PSF Mean Brightness Profiles (F110W)
Normal PSF
3x reduction
200x reduction
Coronagraph
│Coronagraph - PSF│
(Roll subtraction)
500x reduction
NICMOS Image of HD 141569
F110W (~J band)
Science results in Weinberger et al. (1999)
HD 141569
Image1 – PSF1
Image1 – PSF2
Reference Star
Image2 – PSF1
Image2 – PSF2
NICMOS Coronagraph Advantages
• Only HST camera to cover near-IR
• Small spot allows imaging fairly close to star
• Lower background compared to groundbased telescopes
NICMOS Coronagraph Problems
• Poorly matched spot/Lyot stop sizes result
in low diffracted light suppression
• Small spot results in sensitivity to offsets &
focus changes
• Lyot stop position “wiggles” over time
• Numerous electronic artifacts and blocked
pixels (“grot”)
STIS Coronagraph
• Primarily a spectrograph
• CCD, 0.05” pixels, PSF FWHM = 50 mas,
52” x 52” field
• Unfiltered imaging: l = 0.2 - 1.0 mm
• Occulters are crossed wedges: r = 0.5”-2.8”
(21l/D – 110l/D @ V)
• Lyot stop always in the beam
• “Incomplete” Lyot stop
STIS Occulters
STIS Coronagraph Pupil
Models
After Occulter,
Before Lyot Stop
After Lyot Stop
STIS PSF Mean Brightness Profiles
Direct
Wings high due
to red halo, UV scatter
2x reduction
Coronagraph
6x reduction
1200x reduction
│Coronagraph - PSF│
(Roll subtraction)
5000x reduction
STIS Image of HD 141569
HD 141569
HD 141569 - Reference Star
7”
Reference Star
Science results in Mouillet et al. (2001)
STIS Coronagraph Advantages
• Smallest wedge widths allow imaging to
within ~0.5” of central source
• Occulter largely eliminates CCD red halo
and ghosts seen in direct STIS images
STIS Coronagraph Problems
• Incomplete Lyot stop results in low
diffracted light supression
• Unfiltered imaging
• Wedge position not constant
ACS/HRC Coronagraph
• Selectable mode in the HRC: the occulting spots and
Lyot stop flip in on command
• CCD, 25 mas pixels, PSF FWHM=50 mas @ 0.5 mm
• Multiple filters over l = 0.2 - 1.0 mm
• Two occulting spots: r = 0.9” and 1.8” (38l/D –
64l/D @ V)
• Occulting spots in the aberrated beam from HST,
before corrective optics
ACS Coronagraph
1st (Aberrated) Image Plane
Model
r =1.8”
(96%)
r = 0.9”
(86%)
ACS Coronagraph Pupil Models
Pupil After Spot
Pupil After Lyot Stop
ACS Coronagraph PSF
V band, r = 0.9” spot, Arcturus (500 sec)
Scattered light
from surface errors
Shadows of large
occulting spot &
finger
Spot interior
filled with
corrected light
Rings caused
by spot diffraction
Scattered light
streak from
unknown source
29”
ACS PSF Mean Brightness Profiles (V)
Star outside
Surface
scatter
of spot
dominated
7x reduction
Coronagraph
6x reduction
1200x reduction
1500x reduction
│Coronagraph - PSF│
(Roll subtraction)
ACS Coronagraph Image of HD 141569
V band (F606W)
Disk is 2.4x brighter
than PSF here
7”
Science results in Clampin et al. (2003)
ACS Coronagraph Images of HD 141569
B
V
I
• Disk is redder than the star
• No internal color variations
• Moderate forward scattering
• g = 0.25 – 0.35
• Integrated disk flux is ~0.02%
of stellar flux
ACS Coronagraph Image of HD 141569
3.3x fainter
than PSF here
Hard stretch
Deprojected
Density Map
Deprojected
Density Map
ACS Coronagraph Point Source Detection
Limits
ACS Coronagraph Advantages
• Greatest supression of diffracted light
– Only coronagraph in which residual PSF is
dominated by surface error scatter
• Highest resolution & sampling
• Variety of filters
ACS Coronagraph Problems
• Large spots (inner working radius ~1.2”)
• Spots move over time
• Occulting spot interior begins to saturate in
short time on bright targets (~2 sec for Vega)
Sources of PSF Mismatches
• Focus changes caused by thermal variations
– “Breathing” = 3-5 mm primary-secondary separation
change within an orbit = 1/18-1/30 wave RMS change
– Attitude changes (0 – 1/9 wave change)
– Internal changes in camera
•
•
•
•
•
Color differences
Field position variations (WFPC2)
Star-to-occulter alignment (coronagraphs)
Lyot stop shifting (NICMOS)
Jitter
Sensitivity to PSF Mismatches:
ACS Coronagraph+Disk at V (Models)
Color Difference
Focus Difference
Occulting Spot
Shift
A0V-A5V
DfocusSM = 0.5 mm
Shift = 6 mas
K7V-K4V
DfocusSM = 3 mm
Shift = 25 mas
ACS Coronagraph Sensitivity to Breathing
(dZ4 = 1/36 wave)
(dZ4 = 1/120 wave)
ACS Coronagraph Sensitivity to Color
ACS Coronagraph Sensitivity to Decentering
HST Midfrequency Wavefront Stability
• Stability derived from subtraction of ACS
coronagraph B-band images of Arcturus
separated by 24 hrs
• Modeling used to estimate residual errors
due to focus and star-to-spot alignment
differences
• Measured 40-100 cycles/diameter (lower
value limited by occulting spot)
• Midfrequency wavefront varies by <5Å
(conservative), <2Å (likely)
HST vs. Ground: HD 141569
ACS Direct (V)
STIS Coronagraph (U→I)
Palomar AO
Coronagraph (2.2 mm)
Boccaletti et al. 2003
(Their image)
ACS Coronagraph (V)
NICMOS Coronagraph (J)
HST can image disks in the visible – AO can’t
Spectral Deconvolution
Sparks & Ford (2002)
Images courtesy of Bill Sparks
HD 130948 (ACS Coronagraph)
After Spectral Deconvolution
What Might Have Been: CODEX
• Proposed optimized HST coronagraph with
– High density deformable mirror (140 actuators/D)
– Active focus and tip/tilt sensing and control
– Selection of Lyot stops & Gaussian occulting spots
• DM optimization algorithm corrects wavefront &
amplitude errors over ½ of r = 5” field at a given
wavelength
• Was one of two proposed instruments considered
selectable, but COS spectrograph chosen
• Would have easily detected nearby Jovian planets
• PI = Bob Brown (STScI)
CODEX: Our Solar System at 4 pc
Medium band filter, lc = 0.5 mm
Raw CODEX Image
PSF Subtracted Image
S
S
J
J
5”
CODEX Azimuthal profile plot
The Future of HST High Contrast Imaging
• WFC3(?): UV-Vis & near-IR cameras
– No coronagraphs or occulters
• WFPC2: Cumulative radiation damage taking its
toll (WFPC2 would be replaced by WFC3)
• STIS & ACS: Can continue for years
• NICMOS: Can continue, but may need to be turned
off if power system (battery) begins to deteriorate
• Gyroscope failure:
– Would result in increased jitter (3 mas now, perhaps up
to 30 mas on 2 gyros)
– NICMOS & small-diameter STIS coronagraphic
observations probably discontinued
– ACS coronagraph might possibly continue, but depends
on jitter repeatability