Complete Characterization of sub-pixel Response of Near-Infrared Detectors ( Spot s -O-Matic )

October 14, 2009 Detectors for Astronomy, Garching

Complete Characterization of sub-pixel Response of Near-Infrared Detectors (Spots-O-Matic)Tomasz Biesiadzinski, Greg Tarlé, Michael Howe, Curtis

Weaverdyck, Michael Schubnell, Wolfgang Lorenzon

October 14, 2009 Detectors for Astronomy, Garching

Outline Why do we care about sub-pixel response? The Spot-O-Matic

Our first look at intra-pixel response Barron et al., “Subpixel Response Measurement of Near-Infrared

Detectors”, PASP (2007) Spots-O-Matic

Concept Effects of sub-pixel structure

Simulation of errors on point source photometry and simulated Spots-O-Matic correction

Weak lensing considerations Possible errors Shape projection capabilities

Spots-O-Matic Progress Lens characterization

October 14, 2009 Detectors for Astronomy, Garching

Under-sampling for Survey Telescopes

Modern survey telescopes employ under-sampling to improve survey speed. Precision photometry or galaxy shape measurements (for Weak Lensing) in

under-sampled telescopes requires dithering and/or well-characterized intra-pixel response.

For under-sampled NIR survey telescopes, sub-pixel detector properties become important Charge Diffusion (~1.87m) Capacitive Coupling (~2%) Sub-pixel structure (pixel geometry, defects…)

Well SampledUnder-sampled (Assuming Perfect Pixel)

October 14, 2009 Detectors for Astronomy, Garching

Looking Inside a Pixel - the Pixel Response Function (PRF) as measured by the Spot-O-Matic

Single ~1 m spot projector (Spot-o-Matic)

2 dimensional scan over several pixels mapping the internal response

Objective: Determine the largest plate scale a telescope can have while still delivering 1% photometric precision for point sources (e.g. SNe).H2RG-102 one dimensional scan (and

fit to data points) showing effects of charge diffusion and capacitive coupling

October 14, 2009 Detectors for Astronomy, Garching

Spot-O-Matic Results

For detectors with high quantum efficiency, better than 1% photometry is achieved with PSF’s > ½ the size of a pixel

Note small random defects (~5%) that could affect galaxy shape reconstruction for weak lensing.

2 dimensional scan of an individual pixel response

Summation over multiple pixels showing uniform photometry

October 14, 2009 Detectors for Astronomy, Garching

Spots-O-MaticConcept

Simultaneously scan anarray of 160000 spots (400 x 400) to rapidly characterize the sub-pixel response of an entire detector Standard 17.5 cm x 17.5cm

photolithography mask Illuminated by NIR laser Each spot scans a 5 by 5 pixel array

6 axis computer controlled stage X and Y axis to perform a 2D scan of the entire detector Z axis to sample focal “plane” over depth of focus tip, tilt, rotation stages to make sure the image and detector are co-planar and scan is

aligned with rows/columns Use a commercial 50mm lens to demagnify and focus the image

Zeiss Planar T* 1.4/50ZF IR, optimized for NIR light (optics, AR coatings) 28 cm object distance, 6.2 cm image distance, -1/4.5 magnification Cold laser line filter inside dewar blocks out-of-band light

October 14, 2009 Detectors for Astronomy, Garching

Simulation of the use the Spots-o-Matic to Improve Photometry Simulated Spots-o-

Matic signal obtained by convolving Spot-o-Matic Scan with 6m PSF

October 14, 2009 Detectors for Astronomy, Garching

Point Source Photometry (SNe) withoutSpots-o-Matic correction

All fitted fluxes are within ±0.3% of the true values

3.6% of SNe have fluxes with more than 1% error 0.4% of SN have fluxes with

more than 2% error

Large plate Scale 0.81 arcsec/pixel = 0.11 Pixels

Small plate scale 0.23 arcsec/pixel = 0.39 Pixels

October 14, 2009 Detectors for Astronomy, Garching

Errors are now at the sub 0.1% level

All fitted fluxes are within ±0.2% of the true values

Point Source Photometry (SNe) withSpots-o-Matic correction

Large plate Scale 0.81 arcsec/pixel = 0.11 Pixels

Small plate scale 0.23 arcsec/pixel = 0.39 Pixels

October 14, 2009 Detectors for Astronomy, Garching

Spots-O-Matic & Weak Lensing Dithering

Reconstructs diffraction limited seeing in under-sampled telescopes at the expense of survey speed

Compensates for intrapixel structure A complete Spots-o-Matic scan can reduce the number of

dithers required to achieve a given level of shape discrimination

Simulations have yet to be performed to quantify this Can large plate scale HgCdTe detectors be used for

shape reconstruction? Spots-o-Matic data will provide the answer Elliptical “galaxies” and point sources (PSF calibration

“stars”) will be projected onto real detectors and shapes will be extracted

Sub-pixel features mapped by the Spots-O-Matic will be used to correct the shapes for known intrapixel response and determine the errors after correction

October 14, 2009 Detectors for Astronomy, Garching

Spots-O-Matic Lens Characterization Questions

Can a lens produce small enough spots? What is the f-stop that results in minimum

spot size? Lens quality vs. diffraction limit How does the spot size change with

location in the field of view? Knife edge technique

Scaned the spot repeatedly across a knife edge1 (razor blade) while focusing in z.

The spatial derivative of the signal at best focus gives a one dimensional profile of the lens PSF

1Firester, A. H., Heller, M. E., & Sheng, P. 1977, Appl. Opt., 16, 1971

October 14, 2009 Detectors for Astronomy, Garching

Lens Characterization RunsFocusing in Z AxisRaw Signal

Obtained Scans at f / # 5.6, 2.8 and 1.4 on lens axis

At f / # 5.6, 7.5 cm off axis Planned Spots-O-Matic mask range

October 14, 2009 Detectors for Astronomy, Garching

Measured PSF’s of the Lens at Different F-stops

October 14, 2009 Detectors for Astronomy, Garching

Lens Characterization Results Obtained spot sizes (Demagnified by

factor of 4.5): f / # 1.4: on axis

FWHM = 5.4 m Fitted = 6.6 m

f / # 2.8: on axis FWHM = 4.7 m Fitted = 2.9 m

f / # 5.6: on axis FWHM = 5.8 m Fitted = 2.8 m

f / # 5.6: off axis Fitted = 3.4 m

Manufacturer indicated diffraction limited psf at f/# 5.6 Complicated structure at lower f-stops likely due to lens defects Can’t measure FWHM in the off-axis scan (due to low signal to noise) but the fit

indicates ~17% deterioration of from on axis Plan to scan at f/# 2.8 which appears to be a better choice for the Spots-O-Matic Spots-o-Matic spot size will be m (as determined by fwhm/2.35)

October 14, 2009 Detectors for Astronomy, Garching

Conclusion The Spots-O-Matic will provide ~ 2 m resolution

scans of an entire HgCdTe detector in ~ 1 day. With a standard lens it can be used on visible CCDs with

likely better resolution The spot size is not significantly degraded at the

periphery of the field of view Now that we have characterized the optics we are

proceeding with final design and construction of the Spots-o-Matic.

We expect first scans by Spring 2010 in time to influence JDEM instrument design.