Transcript
MID‐INFRARED ASTRONOMY WITH T RECS: THEORY WITH T‐RECS: THEORY, TECHNIQUES, AND DATA
James Radomski Gemini South Observatory UFGRS March 19, 2009
Acknowledgments Gemini North (MICHELLE) • • • •
Kevin Volk Rachel Mason Tom Geballe R Scott Fisher R. Scott Fisher
Gemini South (T-ReCS) James Radomski Michelle Edwards Fredrik Rantakyro T H Tom Hayward d
Also Thanks to Chris Packham (UF) and Jim De Buizer (SOFIA)
Mid‐IR Astronomy id (See Mid‐IR Resources (See d esou es o on Gemini web page) Ge e page)
Why Mid‐ Why Mid‐IR Astronomy? Mid-IR Mid IR (8 (8-25 25µm) suffers 25 25-75X 75X less extinction than optical and traces many features including: • Dust re re--radiation •Fine structure Lines ([ArIII ([ ArIII], ], [SIV], [Ne II II]] • Silicate emission/absorption • PAH features
Image Courtesy Charles Telesco
The Atmosphere •Strongly affected by WV and clouds (water emits AND absorbs mid-IR pphotons)) •Overwhelmed by emission from the sky and telescope (Background limited regime S/N~sqrt(time)) •Not affected by Moon (or Sun ☺)
MIR Seeing Seeing and N and Q band much more stable S i d N d Q b d h bl than visible λ (~1/2 of visible) Typically obtain images ~0.35” at N, ll b diffraction limited at Q (~0.5”) In good seeing and telescope image quality, 2‐3 Airy rings around a point source can be observed b d
Chopping & Nodding T‐ReCS Source Frame (typically can’t see object due to background) Off‐source beam unguided (small affect of 15% in S/N)
Chopping & Nodding Redux T‐ReCS data structure 6 dimensional fits [320, 240, 2, S, 2, N]
Why do we chop & Nod? “Because of the sky Because of the sky variation, of course!” Time variable sky Time‐variable sky background Telescope thermal emission So‐called 1/f detector noise i
Image Courtesy of TIMMI2 team
Array Noise Inputs Note the closed cycle cooler noise For T‐ReCS it’s ~1.2Hz
Note well the 1/f noise For T‐ReCS F T R CS this thi necessitates chopping to avoid resonance with this frequency
The chop freq enc is frequency (typically) determined by the 1/f noise of the array
T‐ReCS (Thermal Region Camera Spectrometer)
T‐ReCS Imaging - Filters •320 x 240 pixels •Pixel size = 0.089" (fixed) •Field Fi ld off view i = 28.8"×21.6“ 28 8" 21 6“ •Diff. Limit ~0.3” (10um)
S t Spectroscopy - Modes M d •R~ 100 (Low) •R R~1000 1000 (Med)
T ReCS T‐ReCS Almost all data taken in chop nod mode Almost all data taken in chop‐nod mode T‐ReCS can only nod ABAB:ABAB:… Only guide in one beam, (off source beam unguided) Only guide in one beam, (off‐source beam unguided) Chop and Nod Throw ~15” (hope to improve to 30”)
T‐ReCS T ReCS Exposure Times 4 “exposure” times “ ” ti Frame time – optimized in software (typical is 25 ms) Saveset time – time optimized in software, (default is 10 s) optimized in software (default is 10 s) Nod Time – optimized in software (~40 sec) Total exposure time – user selected, tweaked in Total exposure time software (300 sec ~ 317.0874 sec)
Total exposure time is defined parameter Clock time takes account of efficiency losses (factor of 3~4) 6 f 60s of exposure time is ~200s of clock time i i f l k i
Filter Selection Need to consider Sensitivityy Diffraction limit Source color N‐band very wide good for detection (but other problems) Qa best for 20um imaging Si2 best combination of resolution and sensitivity Si5 great for highly reddened objects Si1 & all Q band filters are highly dependant on water vapor Si3 & Si4 strongly affected by O3
T-ReCS Sensitivity ii i
Imaging i C lib ti Calibration
Calibration (Imaging) Flux standard u sta da d ~15% uncertainty typical
PSF standard May need to take separate flux and PSF standard as Flux standard too bright or radically different flux to object Flux standard could to too distant from object Change of gravity vector to telescope can significantly affect delivered PSF Large telescope slew forces large pupil rotation
Flux Standard (Imaging) Fl Flux standards are often drawn from d d f d f ‘Cohen’ standards TIMMI /MIRAC li TIMMI2/MIRAC lists Flux calibration u ca b at o in mid‐IR (for imaging) relatively simple d (o ag g) e at e y s p e Aperture on star to determine ADU’s Ratio by flux of star in mJy (cal value 10um ~0.02, 18um ~0.2) Final image in mJy/pixel g
Cohen Standards (Imaging) Cohen has modeled a continuous spectra of many stars carefully tied to observational data Spectra can be used for imaging [and spectroscopic flux calibration] by integrating over the filter’s bandpass [or smoothing to the appropriate resolution] i l i ] Best reference is Cohen et al. 1999, AJ, 117, 1864 C lib ti h d t t i Calibration anchored to two primary standards t d d Alpha Lyr (AO V) Alpha CMa (A1 V)
Flux Standard Selection (Imaging) ( g g) Gemini has a tool to facilitate this search
PSF Calibration (Imaging) Some flux standards can also be used for PSF calibration Often not the case as flux standards are too bright
Best PSF standards are stars of spectral type K or M giants (<10 deg from source if possible) M supergiants should be avoided due to possible extended dust shell emission Hi Hipparcos provides ideal catalogue for PSF searches id id l t l f PSF h
Stars that are very bright should be avoided due to array effects
Spectroscopy C lib ti Calibration
Calibration (Spectroscopy) C b Can be a star or asteroid id Asteroids preferable if in high spectral resolution as stars can show resolvable spectral details at hi‐ show resolvable spectral details at hi res Stars preferable at low res as can be used as flux Stars preferable at low‐res as can be used as flux calibrators
Telluric Standards (Spectra) For object observations lasting longer than ~30 minutes (2hrs clock time), Gemini suggest two standards, prior and post observing For Gemini in queue mode, both are ‘free’
At low resolution, B, A, F and G stars have a smooth spectrum in MIR region Early K maybe O.K., but Late K and M stars should be avoided
Gemini List of B‐G stars [not well flux calibrated] http://www.gemini.edu/sciops/instruments/mir/MIRSpecStdBAFG.html
Cohen Standards (Spectra) Many Cohen spectrophotometic standards are early K dwarfs Fine for Low‐Res spectra but not High‐res mode Many have accurate IRAS mid‐infrared flux densities, making them potentially decent spectrophotometric calibrators Providing telluric line removal and flux calibration
At low resolution the fundamental vibration‐rotation band of SiO significantly depresses the spectrum at 7.5 significantly depresses the spectrum at 7 5‐10 microns in 10 microns in stars later than K0III ‐ K2III Affects ratio’ing Can use Cohen model template specta p p to correct effect Accounted for by Gemini IRAF spectral reduction task msabsflux, which can also be called from within msreduce
Other h C lib ti Calibration
Airmass Correction Airmass calibration difficult due to rapid changes in Sk Sky transmission Sky emission
Flux calibrators should be at similar airmass to source If long source observation and high photometric required, may want to have pre‐ and post‐flux standard observations where pre‐ has RA
Airmass correction that is analogous to NIR world is unusual due to intrinsic photometric uncertainty
Seeing Changes vs. time (& g g ( airmass) (Radomski et al. 2008) ( )
Diffraction & Temperature Effects
Flat Fielding When chopping, flats, biases, bad pixel masks Wh h i fl t bi b d i l k are not needed Act of chopping and fullness of pixel wells effectively removes need to flat field as multiplictive gain change across array minimal and ell acco nted b chopping and well accounted by chopping Bias is removed through chopping process Array has very few (~2) bad pixels Array has very few ( 2) bad pixels, therefore bad therefore bad pixel masks irrelevant
DATA
DATA Reduction IRAF/IDL The Gemini supported IRAF DR package is Th G i i t d IRAF DR k i the official one Other methods possible include IDL (meftools provided by Jim De Buizer)
IRAF DR provided and continually supported p y pp by Gemini Will be evolved to pyraf
IRAF Tasks TBACKGROUND C G O TPREPARE TVIEW MISTACK or MIREGISTER
IDL Tasks MEFHEAD MEFGET Or in one command, MEFREDUCE
Or in one command, MIREDUCE
Other tasks for removing noise miclean.cl miclean cl (IRAF), noise_mask (IDL)
T‐ReCS T ReCS Source Image
T ReCS Bright Object T‐ReCS Bright Object
T ReCS Chop Correction T‐ReCS Chop Correction
T‐ReCS Chop Correction: T ReCS Chop Correction: With Cross‐Talk
T ReCS Spectra: Object T‐ReCS Spectra: Object
T ReCS Spectra: Sky T‐ReCS Spectra: Sky
T ReCS Spectra: Object T‐ReCS Spectra: Object
Typical Noise (OK) Telescope NOD (Dif1, Dif 2)
Window (Src1, Src2)
Crosstalk (Any)
Typical Noise (BAD) Clouds (Sig)
High Frequency Striping (Any)
Chop elongation (Any)
Questions?
