STScI/IDTL Near-IR Detector Simulations

Transcription

1 STScI/IDTL Near-IR Detector Simulations Anand Sivaramakrishnan Ernie Morse, Russ Makidon, Eddie Bergeron, Stefano Casertano, Don Figer Space Telescope Science Institute with Scott Acton, Paul Atcheson Ball Aerospace Marcia Rieke Steward Observatory Sivaramakrishnan STScI July 2004 JIM

2 Why? Wavefront sensing relies on nir detectors in NIRCam nir detectors worse than CCDs HST NICMOS photometry good to 3-5% Wavelength-dependent flat fields Intra-pixel sensitivity, crosstalk between output channels Persistent after-images Electronics settling time of cabling to ADC CRs worse for JWST cf HST Flat errors, 1/f component of read noise Amp glow Science data simulations Sivaramakrishnan STScI July 2004 JIM

3 Sivaramakrishnan STScI July 2004 JIM Modelling science data First cut Python prototype code IDTL darks, flats, settling time: real read noise, dark currents Poisson photon noise in 1000s Simulated CR s up the ramp 10 read NICMOS CALNIC A pipeline software CR Images Data Apply Darks (w/bias) Background Sky Image Background Image with Stellar PSFs Exposure Time Image with Poisson Noise Image with Poisson Noise and Cosmic Rays Data MuxIR Read Input PSFs at GSC2 Positions (IPS + QE), Telescope Area, Exposure Time, Poisson Noise IR Detector Image CALNIC A Final Science Image Sivaramakrishnan, Makidon, Figer, Bergeron, Rauscher, Bushouse, Jedzrejewski, Stockman, Im, SPIE Kona 2002

4 Sivaramakrishnan STScI July 2004 JIM Cosmic Ray Model Offenberg C++ 18 x 18 x 7 um pixels 100 e per 0.1 um liberated (+/- 10) NGST sun shield blocks all CR s 10% of CR s are He nuclei (4x effect) Neighbouring pixels affected No pixel crosstalk No persistence Uncertainties CR spectrum at L2 CR effects through titanium shielding (of detectors)

5 Sivaramakrishnan STScI July 2004 JIM E S0 Galaxy Background (recycled Im & Stockman fortran) 1hour exp Pec Sp Hi z

6 Stars + Galaxies: no shot noise (same image, different stretches) Sivaramakrishnan, Makidon, Figer, Bergeron, Rauscher, Stockman, SPIE Kona 2002 Star on postage stamp brighter than galaxies Sivaramakrishnan STScI July 2004 JIM

7 1 hour exposure w/cr s: raw data Sivaramakrishnan STScI July 2004 JIM

8 Modelling WFSC data second cut IDL O-O code based on python Detector & Amp objects (Morse) Read patterns & timing for HAWAII-2RG (4 on-chip outputs used) IDTL darks, flats, settling time: real read noise, dark currents 60s exposure Simulated CR s Double-correlated sample (two reads) Out-of-focus images as per Ball WFSC algorithms (Acton, Atcheson) No intra-pixel sensitivity modelled (image structures >> pixel on nircam) Offenberg CR model improved After-market Poisson photon (ARTDATA in IRAF) Simple IRAF cl script data reduction Three undithered exposures, median-filtered for CRs Sivaramakrishnan, Morse, Makidon, Bergeron, Casertano, Figer, Acton, Atcheson, Rieke SPIE Glasgow 2004 Sivaramakrishnan STScI July 2004 JIM

9 Sivaramakrishnan STScI July 2004 JIM JWST OPD 160 nm rms JWST PM figure and PSF below spec of 80% JWST 2micron PSF Strehl 78%, log stretch Highly oversampled cf NIRCam Sivaramakrishnan, Morse, Makidon, Bergeron, Casertano, Figer, Acton, Atcheson, Rieke SPIE Glasgow 2004

10 Three exposures NOISELESS RAW DATA 1 3 waves defocus 6 waves defocus RAW DATA 2 RAW DATA 3 Sivaramakrishnan, Morse, Makidon, Bergeron, Casertano, Figer, Acton, Atcheson, Rieke SPIE Glasgow 2004 Sivaramakrishnan STScI July 2004 JIM

11 Simple data reduction Sivaramakrishnan, Morse, Makidon, Bergeron, Casertano, Figer, Acton, Atcheson, Rieke SPIE Glasgow 2004 Sivaramakrishnan A. Median filtered B.Dark-subtracted C.Double-correlated (zero-read subtraction) D. Flattened (bad pixels high) STScI July 2004 JIM

12 Sivaramakrishnan STScI July 2004 JIM Algorithm for routine JWST WFS Misell-Gerchberg-Saxton (MGS) phase retrieval from focus-diverse, known pupil support, data (Acton, Atcheson (Ball)) Local expert: John Krist Sivaramakrishnan, Morse, Makidon, Bergeron, Casertano, Figer, Acton, Atcheson, Rieke SPIE Glasgow 2004

13 End product after processing by Scott Acton s MGS implementation Sivaramakrishnan STScI July 2004 JIM JWST 2micron PSF Applying the tip-tilt-piston and radius of curvature segment corrections determined by this data reduction and analysis perfectly to each segment will result in a 99% Strehl ratio image as shown here Strehl >99% Sivaramakrishnan, Morse, Makidon, Bergeron, Casertano, Figer, Acton, Atcheson, Rieke SPIE Glasgow 2004

14 Sivaramakrishnan STScI July 2004 JIM Is this good enough? JWST 2micron PSF Strehl >99% 11.3 nm rms on wavefront Detector noise added 10nm rms Need to halve this noise More/deeper observations Dither for better flat Look at CR noise more carefully Use reference pixels well Develop good cal pipeline (IDT?) Sivaramakrishnan, Morse, Makidon, Bergeron, Casertano, Figer, Acton, Atcheson, Rieke SPIE Glasgow 2004